Modular total synthesis of rhizopodin: A highly potent G-actin dimerizing macrolide

FULL PAPER

DOI: 10.1002/chem.201302197

Modular Total Synthesis of Rhizopodin: A Highly Potent G-Actin Dimerizing

Macrolide

Manuel Kretschmer ,^{[a, c]}Michael Dieckmann,^[b]Pengfei Li,^{[a, d]}Sven Rudolph,^{[a, e]}

Daniel Herkommer,^[b]Johannes Troendlin,^[a]and Dirk Menche*^[b]

Abstract: A highly convergent total

synthesis of the potent polyketide mac-

rolide rhizopodin has been achieved in

29 steps by employing a concise strat-

egy that exploits the molecule’s C₂

symmetry. Notable features of this con-

vergent approach include a rapid as-

sembly of the macrocycle through

pling strategy and the bidirectional at-

tachment of the side chains by means

course of this endeavor, scalable routes

for synthesis of three main building

blocks of similar complexity were de-

veloped that allowed for their stereo-

controlled construction. This modular

route will be amenable to the develop-

ment of syntheses of other analogues

of rhizopodin.

of

Horner–Wadsworth–Emmons

(HWE) coupling reactions. During the

Keywords: macrocycles

·

natural

products · polyketides · rhizopodin ·

total synthesis

a

site-directed sequential cross-cou-

Introduction

Actin polymerization not only plays a crucial role in cy-

toskeleton formation, but also in various other vital cell pro-

cesses, such as adhesion, cell motility, and intracellular trans-

Myxobacteria are an extremely rich source of structurally

unique and highly diverse natural products with a broad

range of important biological properties.^[1,2]In 1993, the

groups of Hçfle and Reichenbach introduced rhizopodin as

a new compound from the myxobacterium Myxococcus stip-

itatus (strain Mx f164), which was obtained from a soil

sample that was collected on San Andrꢀs Island in the Ca-

ACHTUNGTRENNUNG

phological changes on treatment with rhizopodin in low

nanomolar concentrations, owing to the formation of a dis-

torted cytoskeleton.^[4]In detail, rhizopodin affects the poly-

merization process of monomeric G-actin to afford filamen-

tous F-actin.

[a] Dr. M. Kretschmer ,⁺Prof. Dr. P. Li, Dr. S. Rudolph, Dr. J. Troendlin

Institut fꢁr Organische Chemie, Universitꢂt Heidelberg

INF 270, 69120 Heidelberg (Germany)

[b] Dr. M. Dieckmann,⁺Dipl.-Chem. D. Herkommer,

Prof. Dr. D. Menche

Kekulꢀ-Institut fꢁr Organische Chemie und Biochemie

Universitꢂt Bonn, Gerhard-Domagk-Str. 1, 53121 Bonn (Germany)

E-mail: dirk.menche@uni-bonn.de

[c] Dr. M. Kretschmer ⁺

Current address: Columbia University, New York, NY (USA)

[d] Prof. Dr. P. Li

Current address: Xi’an Jiaotong University

Xi’an, Shaanxi Province (China)

[e] Dr. S. Rudolph

Current address: ASM Research Chemicals, Burgwedel (Germany)

[⁺] These authors contributed equally to this work.

Supporting information for this article is available on the WWW

under http://dx.doi.org/10.1002/chem.201302197.

Figure 1. Originally reported monomeric structure (1) and revised dimer-

ic structure of rhizopodin (2).

Chem. Eur. J. 2013, 19, 15993 – 16018

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

15993

portation.^[5]A more fundamental understanding of the vari-

ous processes that are involved in actin dynamics may lead

to a more profound knowledge regarding cell organization,

which might eventually be exploited for pharmaceutical pur-

poses, for example, in cancer therapy.^[6]In addition, it has

been shown that the treatment of yeast cells with rhizopodin

results in a dramatically decreased phagocytosis efficiency.^[7]

Intrigued by this impressive biological profile, programs

that were directed toward the determination of the absolute

structure of rhizopodin were initiated. During their X-ray

diffraction studies of actin-bound rhizopodin in 2008, in

combination with advanced HRMS analysis, Jansen, Schu-

bert, and co-workers realized that the originally proposed

structure of rhizopodin needed to be revised to a C₂-sym-

metric dimer (2, Figure 1).^[8]In a parallel study, the full rela-

tive and absolute stereochemistry of rhizopodin was as-

signed by ourselves by exploiting a variety of advanced

NMR spectroscopic methods and molecular modeling, as

well as chemical-derivatization experiments.^[9]Simultaneous-

ly, the complete stereochemistry of rhizopodin was inde-

pendently assigned by the group of Schubert.^[10]

Altogether, the unique architecture of rhizopodin is char-

acterized by a central 38-membered macrolactone core,

which is comprised of two oxazole rings and two diene units,

together with two stereotetrads adjacent to geminal dimeth-

yl centers. Notably, the relative configuration within this

specific type of hindered subunit appears to be unprecedent-

ed in nature. Attached to the central core are two side

chains with N-vinyl-formamide termini. It became clear

from the X-ray structure of actin-bound rhizopodin that

these side chains were responsible for the formation of

a very stable ternary complex with two G-actin units.

More recently, Pistorius and Mꢁller proposed a biosynthet-

ic pathway to rhizopodin by analyzing the respective gene

clusters that are responsible for the synthesis of rhizopodin

in Stigmatella aurantiaca.^[11]Interestingly, they found a trans-

polyketide synthase^[12]that was involved in these bacteria,

which meant that all of the methyl branches on the carbon

skeleton must have resulted from methyl-transferase do-

mains, such as in the biosynthesis of bryostatin. By applying

the empirical models that were proposed by McDaniel and

co-workers and Caffrey for the stereochemical outcome of

ketoreductase activity,^[13]they found a complete match be-

tween all of the hydroxy-bearing stereocenters and the ini-

tially proposed structure.

The intriguing and synthetically challenging architecture

of rhizopodin, together with its sparse natural supply,

piqued our interest in developing a total synthesis of this un-

precedented dimeric macrolide, not only to enable the un-

ambiguous confirmation of its initially uncertain stereo-

chemistry, but also to support further biological evaluations

and to enable structure–activity studies, as well as the devel-

opment of simplified analogues to further understand and

potentially utilize the unique biological profile of rhizopo-

din.

ments,^[14]as well as a preparation of the originally proposed

monomeric structure, had been reported^[15]before a first

total synthesis was accomplished by our group.^[16]During

the preparation of this manuscript, Paterson and co-workers

described an alternative synthesis of rhizopodin,^[17]which

relied on similar synthetic fragments and an endgame strat-

egy that was originally discussed in our total synthesis.^[16]

Herein, we report the full details of the various strategies

that we pursued, which eventually culminated in the first

total synthesis of rhizopodin.^[18]

Results and Discussion

Retrosynthetic analysis: As shown in Figure 2, our synthetic

approach relied on the late-stage attachment of the side

chains in a bidirectional manner to enable a modular con-

nection of this critical part of the pharmacophore. We plan-

ned to introduce the side chains either through an aldol/

elimination/reduction sequence (i.e., through the coupling

of compound 3 with compound 4)^[19]or by applying a HWE

coupling of compound 3 with compound 5 and subsequent

1,4-reduction.

Because we were concerned about the stability of com-

pound 5, we decided to use the OTES group as a synthetic

equivalent of the enamide terminus, in contrast to a fully

elaborate side chain (4). In both cases, macrocyclic dialde-

hyde 3 would serve as the coupling partner.^[20]At this point,

we planned to exploit the inherent symmetry in compound 3

and, thus, we dissected both ester motifs, thereby leading to

their respective monomeric fragments (6 and 7). A distinc-

tive structural feature in fragments 6 and 7 is the diene

moiety, which could be forged by means of a suitable cross-

coupling method. Then, ring-closure may either be achieved

by means of a macrolactonization reaction or by employing

a cross-coupling reaction, thereby rendering considerable

flexibility into our synthetic plan. An inspection of these

strategic considerations revealed that four fragments,

namely C1–C7 building blocks 8 and 9 and C8–C22 building

blocks 10 and 11, could be suitable for the assembly of the

macrocyclic core.

As shown in Figure 3, we considered various strategies for

the synthesis of this central building block. To enable

a short route, we intended to form the oxazole ring through

a convergent cyclodehydration sequence, which required

aminoalcohol 13 and an appropriate carboxylic acid as the

key components. Our first approach envisaged a synthesis of

compound 12 with all of the stereocenters between atoms

C16 and C21 already in place before the ring-closure reac-

tion.

As mentioned above, no method for the construction of

the sterically hindered stereotetrad of compound 11 with

the required relative configuration had been reported at the

beginning of our campaign. Therefore, we planned a stepwise

strategy, which relied on an Evans aldol reaction for the

preparation of the C20/C21-syn-relationship, an asymmetric

Sharpless epoxidation to install the stereocenter at the C18

Rhizopodin has attracted considerable synthetic efforts

from various groups and several syntheses of elaborate frag-

15994

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 15993 – 16018

Modular Total Synthesis of Rhizopodin

FULL PAPER

Figure 2. Retrosynthetic analysis of rhizopodin (2). Pin=pinacol, PMB=para-methoxybenzyl, TBS=tert-butyldimethylsilyl, TES=triethylsilyl.

position, and either an asymmetric Mukaiyama aldol reac-

tion or an iridium-catalyzed reverse-prenylation reaction as

reported by Krische to gain access to the C16 stereocenter.

In parallel, we also evaluated a shorter approach to com-

pound 11, which relied on an early-stage oxazole ring-clo-

sure from chiral acid 23 (Figure 4). All of the remaining ste-

reocenters would be installed by either applying an umpo-

lung approach (by using epoxide 20) or an aldol-coupling re-

action with a methyl ketone of type 22. This sequence

would either require dithiane 19 or aldehyde 21, which may

be derived from compound 13 and acid 23, which, in turn,

should be accessible from (À)-pantolactone (24). This latter

approach would be more convergent than the previous one

described in Figure 3 and should allow a more rapid access

to the desired fragment (11).

reported by Noyori and co-workers were found to work best

on this substrate, thereby allowing access to compound 28 in

80% yield and 95% ee.^[22]This result represents the first ex-

ample in which a b-keto ester is used as a substrate for this

powerful method. However, this reaction was difficult to

scale-up, presumably owing to the presence of trace

amounts of deprotected alkyne, which is known to inhibit

the reduction reaction.

Despite these problems, compound 28 was efficiently ac-

cessible by using this sequence and was subsequently trans-

Synthesis of the C1–C7 building block: To set the stereocen-

ters at the C3 and C5 positions, we envisioned an asymmet-

ric reduction of b-keto compounds, according to the proce-

dure reported by Noyori and co-workers. Our first attempt

to introduce the vinyl iodine started from an alkyne by

using a hydrometalation/iodination sequence.

Accordingly, commercially available alkyne 25 smoothly

underwent a Claisen condensation reaction with lithiated

ethyl acetate to furnish compound 26 in 72% yield

(Scheme 1).^[21]Then, various conditions for the stereoselec-

tive reduction step were tested and the standard conditions

Scheme 1. Transfer hydrogenation reaction of b-keto ester 26 under the

conditions reported by Noyori and co-workers. LDA=lithium diisopro-

pylamide, Ts=para-toluenesulfonyl.

Chem. Eur. J. 2013, 19, 15993 – 16018

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

15995

D. Menche et al.

Figure 4. Alternative retrosynthetic analysis of compound 11: early-stage

oxazole formation. PG=protecting group.

Figure 3. Further dissection of fragment 11: late-stage oxazole formation.

Boc=tert-butoxycarbonyl.

for the desired isomer (32), as did other conventional reduc-

ing agents (NaBH₄, LiBH₄, Zn

ingly, our last option, namely an asymmetric hydrogenation

reaction with [RuCl₂A(benzene)₂] as a catalyst and (S)-

ACHTUGNTRENUN(GN BH₄)₂, DIBAL). Disappoint-

formed into methyl ether 29, which could then be subjected

to the same condensation/reduction sequence as before

(Scheme 2). Accordingly, compound 29 was treated with the

lithium enolate that was derived from tert-butyl acetate (30),

thereby enabling access to compound 31 in 83% yield, to-

gether with the formation of a byproduct that resulted from

the double addition of the enolate (6%).^[23]We found that

the use of tert-butyl acetate instead of ethyl acetate was ben-

eficial for the suppression of this undesired byproduct. Next,

a second asymmetric reduction of compound 31 was exam-

ined. However, on applying Noyori’s catalyst anew, the reac-

tion was sluggish and only gave a moderate selectivity of 2:1

CHTUNGTRENNUNG

BINAP as a chiral ligand, resulted in over-reduction of the

triple bond.^[24]

These results prompted us to revise our synthetic plan.

Because the alkyne motif seemed to be responsible for the

problems that we encountered in the asymmetric reduction

reactions, we chose an alcohol as a synthetic equivalent of

the desired vinyl iodine, which would then be installed at

a later stage by means of a Takai olefination reaction. The

initial steps in this successful strategy are shown in

Scheme 3. Starting from known ester 33, which was avail-

able in three steps from d-malic acid,^[25]ether 34 was ob-

tained by treatment with methyl iodine in almost quantita-

tive yield, which was then directly exposed to lithiated tert-

butyl acetate.

This Claisen condensation reaction gave an inseparable

mixture of compounds 35 and 36 in a 2.8:1 ratio, in which

compound 36 was again the result of a double addition of

the enolate.^[26]Next, various attempts were made to improve

this ratio, such as reversing the order of addition of the reac-

tion partners to minimize the concentration of the enolate

or employing other bases. In detail, the use of NaHMDS led

to the formation of a considerable amount of a side-product

that resulted from the elimination of the methoxy group of

compound 34. Alternatively, more LDA was added once the

Scheme 2. Attempted diastereoselective reduction of b-keto ester 31.

15996

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 15993 – 16018

Modular Total Synthesis of Rhizopodin

FULL PAPER

Scheme 4. Synthesis of the C1–C7 building block. DMP=Dess–Martin

periodinane.

anticipated esterification reactions and, secondly, the pro-

tecting group at the C1 terminus had to be chosen carefully

to allow for its selective removal at a later stage of the syn-

thesis under mild conditions. During the assembly of the

macrocycle of rhizopodin, we realized that selective cleav-

age of the tert-butyl ester was problematic, which motivated

us to also synthesize the corresponding methyl ester. Ac-

cordingly, acid 8 was obtained by the treatment of com-

pound 41 with TMS triflate and hydrolysis of the intermedi-

ary silyl ester during work-up (96%). Then, transformation

into methyl ester 9 was effected through the in situ genera-

tion of diazomethane from TMSCHN₂in the presence of

MeOH.^[29]The whole sequence was found to be reliable and

allowed access to compounds 8 and 9 on a gram-scale in 11

and 12 steps, respectively.

Scheme 3. Successful route to the C3/C5 stereocenters of the C1–C7 frag-

ment. BINAP=2,2’-bis(diphenylphosphino)-1,1’-binaphthyl.

reaction had been judged to be complete by TLC to trigger

a retro-aldol process to convert compound 36 back into

compound 34; however, this attempt was unsuccessful and

no increase in the formation of compound 35 was observed.

Another idea was the addition of a second, weaker base,

such as DBU, to transform compound 35 in situ into the cor-

responding enolate, which should be unreactive towards

a second attack of the nucleophilic enolate, but this was also

unsuccessful. At this point, we decided to directly use the

mixture in the subsequent asymmetric hydrogenation reac-

tion and, much to our delight, the application of H₂(4 bar)

Synthesis of the C23–C29 building block: For the construc-

tion of the anti-propionate motif in the C25/C26 region of

rhizopodin, we used an auxiliary-based approach, as de-

scribed by Abiko, Masamune, and co-workers.^[30,31]As

shown in Scheme 5, the treatment of ephedrine derivative

43 with dicyclohexyl boron triflate and triethylamine and

coupling of the resulting E enolate to aldehyde 42 (derived

in two steps from 1,4-butanediol) gave the desired C25/C26

anti isomer in high yield and good selectivity (95%, d.r.=

15:1). Following careful chromatography on silica gel, re-

moval of the undesired diastereomer was possible and com-

pound 44 now served as a common intermediate for the

preparation of both compounds 4 and 5.

in the presence of [RuCl₂ACTHNUTRGNEUNG

(benzene)₂] and (S)-BINAP^[23]

gave a separable mixture of compound 37 with the concomi-

tant loss of the TBS groups (40% over two steps) as a single

diastereomer, together with compound 38 (10% over two

steps). At this stage, the stereochemistry at the C3 position

was confirmed by analysis of the Mosher’s ester derivatives.

Then, we proceeded with the elaboration of compound 37

into the desired vinyl iodine motif (Scheme 4). First, we in-

vestigated a selective primary oxidation reaction,^[27]but un-

fortunately it resulted in the formation of the corresponding

À

lactone from an intramolecular attack of the C3 OH group

onto the newly formed aldehyde and a second oxidation

process. To circumvent this problem, we transformed com-

pound 37 into bis-TBS-ether 39, which was then selectively

deprotected in high yield at the primary position by using

carefully controlled conditions (CSA in MeOH at À108C).

Oxidation to the aldehyde was uneventful and subsequent

treatment with iodoform in the presence of chromium(II)

chloride, following the original procedure reported by Takai

et al., gave compound 41 as an inseparable mixture of iso-

mers (E/Z=7:1, 76% upon scale-up).^[28]

To this end, compound 44 was methylated and the auxil-

AHCTUNGTRENNUNG

iary moiety was elaborated into Weinreb amide 45^[32]follow-

ing a method developed in our group, and was then easily

transformed into the desired methyl ketone (46) in almost

quantitative yield (Scheme 5, left). Although the formation

of the Weinreb amide was accompanied by small degrees of

elimination along the methoxy group and the formation of

the corresponding enone (about 10%), it was still found to

be more effective than a likewise-tested sequence that in-

volved the reductive removal of the ephedrine ester and

a subsequent oxidation/addition/oxidation sequence. Next,

the TBS group was removed under acidic conditions by

treatment with CSA (93%), because basic media led to con-

With regard to the upcoming assembly of the macrocycle,

two aspects were considered at this stage. Firstly, the corre-

sponding acid (8) was required as a coupling partner for the

Chem. Eur. J. 2013, 19, 15993 – 16018

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

15997

D. Menche et al.

instable aldehyde, to which compound 48 was smoothly

added. The obtained mixture of diastereomers was directly

oxidized with PDC in an efficient manner. This sequence

proceeded with a typical yield of 40% over three steps. To

allow for the selective liberation of the primary alcohol at

a late stage of the synthesis, we decided to convert the pri-

mary TBS group of compound 49 into its more labile TES

congener (5), which was accomplished by treatment with

CSA and re-protection with TESCl (88% over two steps).

In principle, the TES analogue of compound 42 could have

been used directly, but in this case the first aldol reaction

(Scheme 5) only gave a yield of 64%, which prompted us to

rely on this somewhat longer sequence. Ultimately, attempts

were made to prepare phosphonate 51 with a fully elaborat-

ed enamide terminus (Scheme 6). Not unexpectedly, all of

Scheme 6. Failed attempts towards fully functionalized b-keto phoshonate

51. IBX=2-iodoxybenzoic acid.

Scheme 5. Preparation of side chains 4 and 5. PDC=pyridinium dichro-

mate, PPTS=pyridinium para-toluenesulfonate.

these attempts were unsuccessful, presumably owing to the

harsh reaction conditions that were required. First, the inter-

mediate alcohol that resulted from the removal of the TBS

group from compound 49 was oxidized by treatment with

DMP and the aldehyde was heated in the presence of PPTS

and N-methylformamide. However, this reaction resulted in

extensive decomposition of the starting material. Alterna-

tively, the sequence for the conversion of compound 46 into

compound 4 (Scheme 5, left) was applied to compound 45

and successfully furnished compound 52. Disappointingly,

treatment of the latter compound with the appropriate lithi-

ated phosphonate (48) only led to decomposition.

siderable formation of the elimination product. The by-

product that resulted from elimination could be easily re-

moved at this stage, which secured access to compound 4

after oxidation with DMP (96%) and Brønsted-acid-mediat-

ed condensation with N-methyl formamide. Notably, no reli-

able conditions for this condensation step could be estab-

lished, that is, neither variation of reaction time, solvent, or

acid gave satisfactory results. Finally, moderate yields (34–

57%) were accepted, which allowed the isolation of com-

pound 4 as a 2:1 mixture of rotamers.

In a rationale for the preparation of compound 5, the

direct replacement of the auxiliary with lithiated compound

48 was also tested, but met with failure.^[31]Instead, treat-

ment of compound 44 with MeI and Ag₂O, followed by re-

duction with LiAlH₄, gave alcohol 47 in 98% yield over two

steps (Scheme 5, right). Oxidation with DMP furnished an

These results made us aware of the instability of the

vinyl-formamide motif and we integrated this knowledge in

our future synthetic considerations.

Synthesis of the C8–C22 building block: Next, our efforts

were directed towards the preparation of the C8–C22 build-

ing block. The oxazole ring should be closed in a convergent

manner, following a cyclodehydration strategy. To enable

15998

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 15993 – 16018

Modular Total Synthesis of Rhizopodin

FULL PAPER

a simultaneous deprotection at a late stage of the synthesis,

a TBS group was chosen to protect the C16 alcohol group,

in analogy to the protecting group on the C3-hydroxy func-

tion. The C22 terminus needed to be transformed into the

desired aldehyde, so our initial idea was to orthogonally pro-

tect it as a PMB-ether. Likewise, a TBS-protected congener

was considered (10 and 11, Figure 2).

ed carefully. Reaction time and temperature were crucial for

preventing either N-methylation or the formation of oxazoli-

dinone 57. Under these optimized conditions, the synthesis

of methyl ether 56 was possible in reproducibly high yield

(87%). Finally, both protecting groups could be simulta-

AHCTUNGTREGnNNNU eously cleaved with aqueous TFA, with special attention to

the high polarity of the target compound (13) during work-

up and isolation. In summary, gram-quantities of compound

13 were accessed in 49% yield over five steps.

Synthesis of amino alcohol 13: Common to all of these

routes was the requirement of amino alcohol 13 for the for-

mation of the oxazole ring, the preparation of which is sum-

marized in Scheme 7. Starting from commercially available

Late-stage oxazole ring-closure: As mentioned earlier

(Figure 3), we next focused on the synthesis of carboxylic

acid 12, with all of the stereocenters between atoms C16

and C21 already in place, which should all be installed in

a stepwise, stereocontrolled manner.

Kiyooka aldol approach: First, we adapted a strategy devel-

oped by Meyers and co-workers.^[35]Starting with an asym-

metric Mukaiyama aldol reaction reported by Kiyooka

et al.^[36]between crotonaldehyde (14) and compound 15,^[37]

silylacetal 58 was afforded in high yield in an enantiomeric

excess of 85% with respect to the C16 stereocenter

(Scheme 8).

Scheme 7. Preparation of aminoalcohol 13 and verification of the 1,2-

anti-aminoalcohol relationship of 55.

N-Boc-d-serine, TBS protection of the hydroxy group pro-

ceeded in good yield (79%, not shown), thus allowing the

formation of the lithium carboxylate of compound 53 and

the subsequent addition of allylmagnesium chloride, as de-

scribed by Knudsen and Rapoport.^[33]This reaction afforded

homoallylic ketone 54, which was prone to isomerization

and, hence, was directly reduced to the corresponding alco-

hol (55). According to a report by Evano and co-workers,

this transformation proceeded smoothly by using sodium

borohydride at À788C.^[34]Presumably owing to internal che-

lation of the amino group, the shown diastereomer was

formed in a ratio of 11:1. The undesired isomer was easily

removed by column chromatography on silica gel and gave

access to compound 55 on a multigram scale. The stereo-

chemical outcome of this reaction was confirmed by trans-

forming compound 55 into the corresponding oxazolidinone

Scheme 8. Kiyooka aldol reaction and further elaboration of compound

58 to obtain the desired C16/C18 configuration. (À)-d-DIPT=(À)-d-di-

AHCTUNGTREGiNNUN sopropyltartrate.

Upon treatment with base in the presence of methyl di-

ethylphosphonoacetate, a 1,5-O-silyl migration took place to

afford an aldehyde intermediate, which was converted with-

out isolation into the corresponding unsaturated ester in

79% yield.^[38]Reduction of the ester with DIBAL proceed-

ed smoothly and the resulting allylic alcohol was subjected

to Sharpless asymmetric epoxidation conditions^[39]to furnish

epoxide 59 in 92% yield in a diastereomeric ratio of 8:1 in

favor of the desired isomer.

Next, our efforts were directed towards the transforma-

tion of compound 59 into a 1,3-diol equivalent (Scheme 9).

To this end, various hydride sources were examined for the

regioselective opening of the epoxide, including RedAl,

LiAlH₄, DIBAL, and AlH₃, but all of them resulted in

either decomposition of the starting materials or afforded

mixtures of diols as a result of silyl migration under the

basic reaction conditions, with the formation of only trace

amounts of the desired compound (60). Ultimately, we oxi-

(57), which showed

a characteristic nOe correlation

(Scheme 7, bottom).

After having verified the correct stereochemistry of com-

pound 55, selective O-methylation of compound 55 was re-

quired, which required some optimization. Although a varie-

ty of bases (NaHMDS, 2,6-lutidine, etc.) and methylation

agents (MeOTf, Me₃O·BF₄) were tested, NaH/methyl iodide

gave superior results if the deprotonation step was conduct-

Chem. Eur. J. 2013, 19, 15993 – 16018

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

15999

D. Menche et al.

tions for a selective copper-mediated opening of the epoxide

with one of the many reported methyl nucleophiles to gain

access to compound 67.^[42]Thus, we examined various cop-

per(I) sources (CuI, CuCN) with different nucleophiles

(MeLi, MeMgCl, MeMgBr), as well as the reagent combina-

tion nBuLi/AlMe₃, which was only recently described.^[43]

However, disappointingly, all of these conditions either re-

sulted in decomposition (CuI/MeMgBr) or no conversion,

so we had to revise our strategy.

Alternatively, Bode and co-workers described an organo-

catalytic internal redox-isomerization reaction of 2,3-epoxy-

AHCTUNGTREGaNNUN ldehydes, such as compound 61 (Scheme 9), to afford the

corresponding b-hydroxyesters.^[44]Application of these con-

ditions allowed the preparation of compound 69 in high

yield, without losing any stereochemical information, even

on a gram scale (Scheme 11). Following the routine adjust-

Scheme 9. Attempted regioselective opening of epoxide 59.

dized compound 59 to the respective epoxyaldehyde (61) by

treatment with DMP, for which a method to convert it to

a b-hydroxyaldehyde has been described.^[40]Indeed, treat-

ment of compound 61 with NaBACHTNUGRTENUNG(OEt)₃SePh furnished com-

pound 62; however, the hydroxy group in the product could

not be orthogonally protected under a variety of conditions

(Ac₂O/pyridine, PivCl/NEt₃, TESOTf/2,6-lutidine, etc.).

In contrast, compound 62 smoothly underwent a Still–

Gennari olefination reaction^[41]to afford the pure Z isomer

(63) in 77% yield after chromatography on silica gel

(Scheme 10), which was subsequently protected as a TES-

ether and reduced by treatment with DIBAL to afford com-

pound 65 in excellent yield.

Scheme 11. Organocatalytic redox-isomerization of compound 61 and

Evans aldol reaction to install the full C16–C21 stereotetrad. DIPEA=

N,N-diisopropylethylamine, py=pyridine, Bn=benzyl.

ment of the oxidation states and protecting groups, aldehyde

72 was finally obtained in high yield over these three steps

and also allowed the partial removal of the undesired diaste-

reomer by careful chromatography on silica gel. From a stra-

tegic point of view, compound 72 was a well-suited precur-

sor for installing the missing stereochemical C20/C21-syn re-

lationship by using an auxiliary-controlled aldol reaction, as

described by Evans et al.^[45]Consequently, compound 72 was

treated with different enolates that were derived from com-

pound 74, of which the titanium enolate as described by

Crimmins and co-workers was most reliable and allowed the

formation of compound 73 in 81% yield as a single

isomer.^[46]

Scheme 10. Still–Gennari olefination of compound 62 and attempted se-

lective epoxide opening to afford compound 67.

In agreement with previous reports that Z-allylic alcohols

are poor substrates for asymmetric epoxidation reactions,^[38]

the standard conditions described by Sharpless did not allow

complete conversion, but compound 66 could be isolated in

61% yield as a single diastereomer. It should be noted that

protection of the C18-alcohol group was crucial, otherwise

a 4:1 mixture of both epoxides was observed. With com-

pound 66 in hand, we were confident of establishing condi-

Next, the auxiliary was removed upon treatment with

LiBH₄, as reported by Soai and Ookawa,^[47]and the resulting

diol was selectively protected as a primary TBS-ether before

16000

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 15993 – 16018

Modular Total Synthesis of Rhizopodin

FULL PAPER

the remaining hydroxy group at the C20 position was trans-

formed into the corresponding methyl ether (76,

Scheme 12). This strategy was found to be superior to an in-

verted order of these steps, that is, methylation/reduction/

TBS protection.

access to acid 12. Our revised strategy is shown in

Scheme 13 and commenced with an asymmetric reverse-

AHCTUNGTREGpNNUN renylation reaction, following a procedure reported by Kri-

sche and co-workers.^[52]

Scheme 13. Reverse-prenylation approach.

We identified PMB-protected alcohol 16^[53]as a suitable

precursor for the installation of the required carboxylic acid,

which was treated with 1,1-dimethyl allene (17) in the pres-

ence of propionaldehyde and a catalytic amount of Ir-cata-

lyst 79 to afford compound 80 in quantitative yield. The

enantiomeric excess could be improved from 82 to 90% if

compound 79 was purified by column chromatography prior

to use^[54]and the configuration of the new stereogenic

center was validated by analysis of the Mosher esters. Next,

the free OH group was protected as a TBS-ether (96%)

before oxidative cleavage of the double bond afforded an al-

dehyde, which was directly homologated by means of

a HWE reaction and reduced to the corresponding allylic al-

cohol (81) in high yield. Whereas an ozonolysis reaction was

feasible on a small scale (Scheme 13), problems arose

during scale-up, with partial cleavage of the PMB group

upon prolonged exposure to ozone. However, this problem

could be easily circumvented by applying a two-step proce-

dure of bis-hydroxylation, followed by glycol cleavage with

NaIO₄.^[55]

Inspection of compound 81 unambiguously revealed its

similarity to an intermediate in the route presented above

and we hoped that the sequence that we had developed for

the installation of the missing stereocenters could be easily

adapted. As shown in Scheme 14, this was indeed the case.

Epoxidation of compound 81 under the previously described

conditions was performed to afford an epoxyalcohol as an

8:1 mixture in favor of the expected C18 diastereomer. Oxi-

dation with DMP furnished aldehyde 82, which then under-

went an internal redox-isomerization^[43]into the respective

b-hydroxyester (83), which was elaborated as described

above. Aldehyde 85 now served as a starting material for an

Evans-aldol reaction. Treatment with the titanium enolate

Scheme 12. Attempts to synthesize acid 12.

To advance compound 76 into compound 12, our synthetic

plan now demanded the elaboration of the vinyl unit into

a carboxylic acid. To this end, the olefin was oxidatively

transformed into aldehyde 77. At this stage, several methods

were considered. Because compound 77 was extremely steri-

cally hindered and, hence, was unreactive toward a variety

of nucleophiles, we took note of a method that was used by

Kçrner and Hiersemann to solve a related problem, which

allowed the formation of an alkyne from sterically congest-

ed aldehydes.^[48]Thus, the treatment of compound 77 with

an ylide that was derived from chloromethyltriphenylphos-

phonium chloride smoothly resulted in the formation of

a mixture of E/Z-vinyl chlorides, which, upon treatment

with LDA, collapsed in a Fritsch–Buttenberg–Wiechell-like

rearrangement into compound 78. With alkyne 78 in hand,

numerous conditions for its transformation into acid 12

were examined (most likely through a two-step process that

involved the initial formation of an aldehyde). Among these

conditions were an anti-Markovnikov hydration,^[49]various

hydroboration/oxidation methods, and a hydrosilylation re-

action followed by a Tamao–Fleming oxidation.^[50,51]Disap-

pointingly, all of these efforts were unsuccessful and only

gave trace amounts of compound 12 at best. Mostly, decom-

position of starting materials was observed.

Reverse-prenylation approach: Despite this highly unsatisfy-

ing outcome, we had established a sequence for the con-

struction of the C16–C21 stereocenters and we felt that

a slight variation of the starting material should enable

Chem. Eur. J. 2013, 19, 15993 – 16018

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

16001

D. Menche et al.

Scheme 15. Completion of the late-stage oxazole formation approach to

building block 11. DEPBT=3-(diethoxyphosphoryloxy)-1,2,3-benzotria-

zin-4(3H)-one.

Scheme 14. Final successful route to acid 12.

ing the complete C8–C22 fragment (11) in a longest linear

of compound 74 provided compound 86 as a sole product in

95% yield, which was found to be the C20/C21-syn isomer,

as expected. With the lessons that were learned from our

previous efforts, reductive removal of the auxiliary (LiBH₄,

88%), introduction of the primary TBS-ether (TBSCl), and

methylation (LiHMDS, MeI, 88% over two steps) proceed-

ed uneventfully and afforded compound 87. In sharp con-

trast to the problems discussed above, the PMB group was

removed under standard conditions (DDQ in buffered aque-

ous solution) and the liberated alcohol was oxidized to the

required carboxylic acid (12) in two steps with DMP and

NaClO₂.

With efficient access to both compounds 12 and 13 in

hand, we proceeded with their assembly. A few coupling

conditions were examined, of which treatment with

DEPBT^[56]in the presence of NEt₃was found to be most re-

liable and allowed the formation of amide 88 in 73% yield

(Scheme 15).

Next, we successfully realized the planned convergent

ring-closure of the oxazole by applying slightly modified

Robinson–Gabriel conditions reported by Wipf and

Miller,^[57]which involved oxidation of the hydroxyamide and

treatment of the intermediate aldehyde with triphenylphos-

phine/iodine in the presence of triethylamine. Much to our

delight, these reagents resulted in the formation of com-

pound 89 in 87% yield over both steps. Our final task was

the selective removal of the TES group. Whereas HF/pyri-

dine cleaved both the primary TBS- and the TES-ether, sur-

prisingly, treatment with MgBr₂·OEt₂and SMe₂exclusively

deprotected the TES group as desired (80%),^[58]thus afford-

sequence of 23 steps in 8.1% yield.

Early-stage oxazole formation: The route described in the

previous section allowed access to compound 11, but it was,

nevertheless, lengthy and difficult to scale-up. This drawback

prompted us to address this problem by developing an alter-

native approach that was more convergent and, thus, al-

lowed a more practical synthesis of the C8–C22 fragment.^[16]

As discussed in the retrosynthetic analysis (Figure 4), acid

23 was required, which we anticipated could be available

from (À)-pantolactone (24). In the forward sense, com-

pound 24, which already contained the stereocenter at the

C16 position, was first protected as the TBS-ether and, sub-

sequently, reduced to the corresponding lactol (90,

Scheme 16).

Next, we examined conditions to convert compound 90

into olefin 91. Although all efforts to perform a Wittig olefi-

nation failed, in accordance with a literature precedent,^[59]

the desired methylene unit could successfully be installed by

using the Tebbe reagent.^[60]However, this reaction was very

expensive when performed on a large scale. Therefore, we

employed a two-step Peterson olefination procedure as de-

scribed by Ley and co-workers.^[61]The addition of a commer-

cially available Grignard reagent and subsequent treatment

with BF₃·OEt₂triggered the desired elimination reaction

and afforded compound 91 on a multigram scale. Moreover,

only a single chromatographic purification step was needed.

Diol 91 was conveniently protected with two TBS groups

and the double bond could be elaborated into the carboxylic

acid in a straightforward manner. Hydroboration/oxidation

16002

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 15993 – 16018

Modular Total Synthesis of Rhizopodin

FULL PAPER

route in terms of step count and reproducible overall yield

(Scheme 18).

In 2009, Krische reported the facile and scalable synthesis

of building block 93, which was prepared by an iridium-cata-

Scheme 16. Elaboration of acid 23 from (À)-pantolactone (24). 9-BBN=

9-borabicycloACHTUNGTRENNUNG[3.3.1]nonane.

using mild conditions developed by Trost et al. provided an

alcohol (90%, not shown),^[62]which was oxidized to the re-

spective acid (23) by using IBX at reflux in EtOAc (96%)^[63]

and a subsequent Pinnick oxidation (92%). It is noteworthy

that the intermediate aldehyde was susceptible to decompo-

sition upon chromatography on silica gel. This decomposi-

tion could be omitted if the aforementioned conditions were

used, which only required simple filtration.

Scheme 18. Optimized strategy leading to aldehyde 21. DBU=1,8-

diazabicycloACTHNUTRGNEU[GN 5.4.0]undec-7-ene.

With acid 23 and amino alcohol 13 (Scheme 7) in hand,

these two fragments were coupled together by using

DEPBT/NEt₃through our previously developed sequence

(Scheme 17). Oxidation with IBX, followed by treatment

lyzed allylation reaction within three steps in high optical

purity and excellent overall yield.^[64]In our hands, this

method was confirmed as a powerful tool for the allylation

of extremely sterically hindered alcohols or aldehydes. We

oxidized compound 93 by means of ozonolysis and Pinnick

oxidation. Amide formation was again best performed by

using DEPBT as a coupling reagent in 93% yield. N,N’-dicy-

clohexylcarbodiimide (DCC), O-(7-azabenzotriazol-1-yl)-

N,N,N’,N’-tetramethyluronium

hexafluorophosphate

(HATU), and the corresponding acid chlorides that were

generated by heated acid 94 at reflux in thionyl chloride or

oxalyl chloride gave lower yields. After oxidation, oxazole

building block 95 was most efficiently obtained under slight-

ly different cyclodehydration conditions, as also described

by Wipf and co-workers.^[56a]The intermediate aldehyde (not

shown) was treated with 1,2-dibromo-1,1,2,2-tetrachoro-

ethane, PPh₃, and NEt₃and subsequently exposed to DBU

to give rise to compound 95 in good yield. It should be men-

tioned that the procedure as described above (I₂/PPh₃/NEt₃)

suffered from highly varying yields. As depicted in

Scheme 17, the selective deprotection of the primary TBS

group was also problematic in terms of material throughput,

because incomplete conversion and varying yields had to be

accepted. Within the optimized route shown in Scheme 18,

the orthogonal PMB protection of compound 95 was found

to be advantageous, owing to the possibility of selectively

deprotecting compound 95 in almost quantitative yield by

employing mild Lewis acidic conditions (MgBr₂·Et₂O/SMe₂).

After oxidation with IBX, the desired aldehyde (21) was

now accessible in only 10 steps and an overall yield of 30%.

Scheme 17. Early-stage oxazole formation and synthesis of aldehyde 21.

with PPh₃/I₂and NEt₃, triggered the desired cyclodehydra-

tion and resulted in the smooth formation of oxazole 92

(72% over three steps).^[56a]Because our synthetic plan en-

visaged the formation of aldehyde 21 as an intermediate for

the application of either umpolung or aldol-coupling strat-

egies, compound 92 needed to be selectively deprotected at

the primary position. After considerable experimentation,

we eventually found that HF/pyridine in buffered THF were

the optimal conditions, which afforded the respective alco-

hol in various yields (40–90%, based on 75% conversion).

Oxidation to aldehyde 21 was unproblematic and proceeded

in quantitative yield.

We used this route (early-stage oxazole formation) to

gain access to multigram quantities of aldehyde 21. Never-

theless, during further investigations, we could optimize this

Umpolung approach: For the desired umpolung reaction, we

required compound 20 as a coupling partner, whose mirror

image was already literature-known (Scheme 19).^[65]We

Chem. Eur. J. 2013, 19, 15993 – 16018

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

16003

D. Menche et al.

entry 3). We anticipated that compounds 98 and 99 arose

from E1cb mechanisms, which implied that protons adjacent

to the heterocyle possessed pK_avalues within the same

range as the dithiane proton, which indeed might be the

case. In addition, treatment of compound 19 with various

bases and quenching with D₂O did not result in the forma-

tion of deuterated compound 19, which gave rise to the con-

clusion that no deprotonation at the desired position oc-

curred.

For a related system, Smith and co-workers extensively

studied several sterically encumbered dithianes, which they

were able to metalate without problem.^[68]However, a relat-

ed substrate to compound 19, which contained an allyl-ether

functionality, could not be deprotonated either. This result

prompted us to further study this process. From literature

precedents, we concluded that the equatorial proton in a di-

thiane was abstracted much more quickly;^[69]this conclusion

was rationalized by the assumption that the formed anion

showed no disturbing interactions with the electron lone

pair located on the sulfur atom, and the anion exhibited a fa-

Scheme 19. Preparation of the precursors for the envisaged umpolung ap-

proach.

adapted the described sequence starting from commercially

available diol 96 by using the enantiomeric tartrate for the

initial Sharpless epoxidation, which furnished compound 20

in 8 steps. Then, aldehyde 21 was transformed into the de-

sired dithiane (19) by treatment with 1,3-propanedithiol and

catalytic amounts of TiCl₄.^[66]

With compounds 19 and 20 in hand, we were poised to

study the key umpolung step.^[67]To this end, various bases

and conditions were applied to form the anion that was de-

rived from compound 19, which should open epoxide 20

through a nucleophilic attack (Scheme 20).

À

vorable interaction with the s* orbital of the carbon sulfur

bond.^[70]This delocalization would result in a lengthening of

À

the indicated carbon sulfur bond. Although Lehn and Wipff

confirmed this theory for thioacetals in

a theoretical

study,^[71]to the best of our knowledge, no data for (cyclic)

1,3-dithianes were available. Consequently, we calculated

the bond lengths in both neutral and anionic 2-tert-butyl-1,3-

dithiane as models for compound 19, which showed the an-

ticipated result (Figure 5). The anion located at the C1 posi-

Scheme 20. Umpolung approach.

Much to our disappointment, we observed either no con-

version or the formation of several side-products that result-

ed from elimination processes. With nBuLi as a base

(Table 1, entry 1), both starting materials were almost fully

recovered. With sBuLi, the reaction proceeded sluggishly

(Table 1, entry 2), that is, compound 19 decomposed and

low yields of compound 99 were observed. In contrast, treat-

ment of compound 19 with tBuLi resulted in the smooth for-

mation of undesired compound 98 after 10 min (Table 1,

Figure 5. Orbital interactions in 2-tert-butyl-1,3-dithiane; bond lengths

[ꢄ] were calculated by using DFT at the B3LYP, 6-31G** level of theory.

tion transferred electron density into the s* orbital, which

resulted, on the one hand, in a substantial contraction of the

indicated bond and, on the other hand, in an enlargement of

À

the C S bond of 0.03 ꢄ, owing to partial occupancy of the

antibonding orbitals. This result strongly underpins the im-

portance of the s* orbital for the stabilization of the anionic

species, which is only possible in this conformation with the

tert-butyl substituent being pseudo-axial.

Table 1. Attempted umpolung reaction of dithiane 19.^[a]

According to the report by Smith et al., two conforma-

tions of compound 19 might, in principle, be involved. Pre-

sumably owing to the Thorpe–Ingold effect, compound 19 is

expected to reside in either an axial or an equatorial ar-

rangement with respect to the proton of the dithiane, as

shown in Figure 6.^[68]A second DFT study revealed that, un-

Entry

Base

Additive

Result (yield [%])

1

2

3

nBuLi

sBuLi

tBuLi

HMPA

TMEDA

HMPA

no conversion

99 (15)

98 (96)

[a] HMPA=hexamethylphosphoramide,

methylethane-1,2-diamine.

TMEDA=N,N,N’,N’-ACTHNGUTERNtUNG etra-

N

ACHTUNGTRENNUNG

16004

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 15993 – 16018

Modular Total Synthesis of Rhizopodin

FULL PAPER

Figure 6. Possible conformations of dithiane 19.

Scheme 21. Aldol coupling reaction of aldehyde 21 with methyl ketone

100.

surprisingly, structure 19-ax (left) was energetically favored,

although it would only allow slow axial deprotonation. Con-

formation 19-eq (right) indeed possessed the desired ar-

rangement for effective proton abstraction, but it was less

stable. Moreover, it could exhibit an interaction between the

p system of the heterocycle and the empty s* orbital. Al-

though this explanation was originally devised for an olefin

system, the heteroaromatic ring system has, in principle, the

ability to exhibit a comparable stereoelectronic effect,

which, in turn, might diminish the ability of this orbital to

stabilize the required carbanion and, thus, impede deproto-

nation. These findings allow the conclusion that neither con-

formation allows an effective deprotonation at the desired

position. Because compound 19 contained several protons

of comparable acidity, namely those at the benzylic posi-

tions, these protons were preferentially abstracted instead of

the desired dithiane proton, thus leading to the observed

side-products.

C18 epimer. Initially, the coupling reaction between the cor-

responding Li-, Na-, and K-enolates of methylketone 100

and aldehyde 21 only resulted in low to moderate yields of

compound 101, always in favor of the non-desired diastereo-

mer, (epi-18)-101. To subvert this preference, we attempted

(+)-/(À)-Ipc-boron

aldol

reactions,

as

well

as

Mukaiyama-type aldol reactions, but the yields of the isolat-

ed products were very low and the undesired diastereomer

still remained the main product. Table 2 shows representa-

tive aldol reactions amongst a large variety of tested condi-

tions within this project.^[72]In terms of chemical yield, the

Li-enolates that were formed from PMB-protected methyl-

ketone 100 were found to be the best after optimization

(Table 2, entries 1 and 2). However, only mixtures of diaste-

reomers in favor of the undesired (epi-18)-101 (determined

by analysis of the Mosher esters) could be obtained. The use

of chiral (+)-/(À)-Ipc-boron enolates could not overrule the

preference for the formation of the ’’undesired’’ diastereo-

mer (Table 2, entries 5 and 6 and Scheme 22), presumably

owing to an inherent 1,4-syn substrate control of methylke-

tone 100. In all of these “closed-transition-state” aldol reac-

tions,^[73]the stereochemical outcome was clearly directed

toward the undesired diastereomer. In contrast, a Mukaiya-

ma aldol reaction of the in situ formed TMS-silyl enol ether

of compound 100 proceeded with some preference for the

desired diastereomer (Table 2, entry 7), which could be ex-

plained in contrast to earlier aldol attempts through an

“open-transition-state” model. However, this Mukaiyama

Despite these unfruitful efforts, aldehyde 21 could serve

as a coupling partner in an aldol-coupling reaction, which

would allow an alternative option to install the missing ste-

reocenters in a convergent manner.

Aldol-coupling approach for the construction of the C8–C22

building block: After the disappointing results with the um-

polung approach, we decided to pursue an aldol-coupling

strategy between aldehyde 21 and methyl ketone 100, which

was derived from the (S)-Roche ester. This coupling process

would open up a highly convergent route toward the desired

C8–C22 building block and, hence, was promising for gain-

ing access to multigram quantities of this compound. For

this purpose, our efforts were

aimed at establishing appropri-

ate conditions that would deliv-

er the desired coupling product

(101) in acceptable yield and

with the desired stereochemis-

try at C18 (Scheme 21).

However, implementation of

this strategy was extremely

challenging, owing to the high

steric hindrance of aldehyde 21

and its strong preference for

the formation of the undesired

Table 2. Conditions for the aldol-coupling reaction between compounds 21 and 100.

Entry

Base

Additive

Solvent

T [8C]

Yield [%]

d.r.

(101/

G

ACHTUNGTREN(NUNG epi-18)-101)

1

2

3

4

5

6

7

LiHMDS

NaHMDS

KHMDS

(+)-Ipc₂BCl

TMSCl

THF

DMF

THF

À40

À78

95

19

25–45

43

1:1.7

1:1.1

1:2.4

1:2.6

1:4.0

1:4.8^[a]

THF

NEt₃

Et₂O

CH₂Cl₂

À78 to À20

À78 to +25

À78

17

95

5

NEt₃then BF₃·OEt₂

2.2:1

[a] Cyclic boronate 102 was obtained as the product (also see Scheme 22). LiHMDS=lithium bis(trimethyl-

silyl)amide, (+)-Ipc₂BCl=(+)-chlorodiisopinocampheyl borane, TMSCl=trimethylsilyl chloride.

A

ACHTUNGTRENNUNG

Chem. Eur. J. 2013, 19, 15993 – 16018

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

16005

D. Menche et al.

mize the initial aldol reaction between aldehyde 21 and

methylketone 100, before developing independent routes for

the preparation of compound 10 from both compounds 101

and (epi-18)-101 after their separation. As shown in

Scheme 23, this reaction could indeed be achieved. The use

Scheme 22. Observed aldol-coupling/1,3-anti-reduction route to diol 103.

reaction only proceeded in very low yield (5%), which

could not be further improved.

During the course of these investigations, we also discov-

ered a new type of domino aldol-coupling/1,3-anti-reduction

sequence (Scheme 22), which has been described in detail

elsewhere.^[74]Because compound 21 only showed moderate

reactivity upon treatment with the (+)-Ipc-boron enolate of

compound 100 at À78 to À208C, the temperature was in-

creased to 258C, which, surprisingly, resulted in the smooth

formation of a single product, which was revealed to be

compound 102 by using advanced NMR spectroscopic meth-

ods and chemical derivatization experiments (95% yield

and a 5:1 ratio in favor of the epi-18/epi-20 isomer). The

outcome of this reaction was rationalized to arise from the

Ipc moiety acting as an internal reducing agent after the ini-

tially performed aldol coupling. Moreover, conditions for

the effective removal of the boronate unit to afford com-

pound 103 could be disclosed (basic H₂O₂in THF at 258C).

The scope of this newly developed domino process of an

aldol coupling between a methyl ketone and an aldehyde

and subsequent Ipc-mediated reduction to 1,3-anti diols was

thoroughly explored and the stereochemical outcome could

be shown to be independent from both substrate control

and from the nature of the employed Ipc reagent. This

method is expected to be efficiently applied in a diverse

range of syntheses of polyfunctional substrates with 1,3-diol

motifs.

Scheme 23. Aldol-coupling approach and preparation of compound 10.

of the Li-enolate of compound 100, which was generated by

using LiHMDS at À788C, followed by increasing the reac-

tion temperature to À408C, and the subsequent addition of

aldehyde 21 with a prolonged reaction time of 40 min ena-

bled us to perform this reaction in almost quantitative yield

(Table 2, entry 1). Following this procedure, we obtained

a mixture of diastereomers that was slightly in favor of the

undesired stereoisomer (epi-18)-101, 1:1.7 desired/unde-

sired), which were separable by preparative HPLC. By using

this route, we could obtain multigram quantities of both

compounds 101 and (epi-18)-101 (Scheme 23).

Next, the desired diastereomer (101) was transformed

into building block 10 through a stereoselective 1,3-anti re-

duction reaction, according to the method described by

Evans, Carreira, and co-workers (95%, d.r.>19:1),^[75]and

a selective methylation of the less hindered C20-hydroxy

group, which could be achieved in good yield after optimiza-

tion. This mono-methylation reaction was best performed in

a THF/MeI mixture (8:1, v/v) with an excess of sodium hy-

dride (10 equiv) as a base and by careful adjustment of the

reaction time and temperature (+108C, approximately

60 min with close monitoring by TLC). Within this two-step

sequence, we could secure access to compound 10 in 76%

yield. Alternatively, we tested the reduction of compound

101 by following the Evans–Tishenko procedure (SmI₂/acet-

aldehyde),^[76]but this was not reliable in our hands and only

led to a moderate yield of the expected product.

Disappointingly, this new method did not help us to

devise a route for the selective preparation of compound

101 either, which forced us to accept the undesired diaste-

reoselectivity at this stage. However, due to its high conver-

gence, this aldol approach proved to be superior in terms of

overall material throughput, which prompted us to further

pursue this strategy. To this end, efforts were made to opti-

16006

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 15993 – 16018

Modular Total Synthesis of Rhizopodin

FULL PAPER

Table 3. Substrate-controlled diastereoselective reduction of ketone

To likewise convert compound (epi-18)-101 into com-

pound 10, a stereoselective 1,3-syn reduction of the b-hy-

droxy ketone moiety of compound (epi-18)-101 was applied

to first install the desired C20-hydroxy group (Scheme 24).

This installation was best achieved by means of internal che-

lation with Cy₂BCl prior to the reduction with LiBH₄(95%,

d.r.>19:1).^[77]Subsequently, the C20-hydroxy group was se-

lectively methylated under slightly modified conditions, as

described above (MeI, NaH, THF, 08C, 40 min, 65% yield,

89% brsm), to furnish compound 104 in good yield and se-

lectivity.

105.^[a]

Entry Reductant

Solvent T [8C]

Yield d.r.

[%] (10/104)

ACHTUNGTRENNUNG

1

2

3

4

5

6

NaBH₄

DIBAL

THF

À78 to 25 n.c.

À78 n.c.

–

1.3:1

8:1

(R)-Me-CBS, BH₃·SMe₂CH₂Cl₂

0 to 25 n.c.

À20 50

LiAlH₄

Red-Al

THF

À90 80

À78 5–10

1:2

[a] n.c.=no conversion. (R)-Me-CBS=(R)-methyl oxazaborolidine,

DIBAL=diisobutylaluminum hydride, Red-Al=sodium bis(2-methoxy-

AHCTUNGERTGeNNUN thoxy)aluminumhydride.

conversion (Table 3, entries 1–3). Pleasingly, by using lithi-

um aluminum hydride at À208C, we were glad to observe

ketone reduction in slight favor of the desired C18 diaste-

reomer (10; Table 3, entry 4). This substrate control could

be improved upon lowering the reaction temperature to

À908C, thereby allowing the selective formation of com-

pound 10 (d.r.=8:1) in good yield (Table 3, entry 5). Inter-

estingly, application of Red-Al favored the undesired diaste-

reomer (d.r.ACTHNUTRGNEUNG(10/104)=1:2; Table 3, entry 6).

In summary, the routes outlined in Scheme 18 and

Scheme 23, which relied on an early-stage oxazole forma-

tion, allowed the effective preparation of building block 10

in a scalable and highly reproducible manner (13 steps, 20%

overall yield).

Synthetic routes towards the macrocycle of rhizopodin:

Heck macrocyclization strategy: Figure 7 outlines our initial

retrosynthetic plan to close macrocycle 106 of rhizopodin.

We envisaged a chemoselective cross-coupling strategy by

using sequential Suzuki and Heck coupling reactions. This

plan required the C1–C7 fragment (8), the C8–C22 oxazole

fragment (10), and boronate 107, which was derived from

oxazole 10.

We planned to synthesize boronate 107 through the cross-

metathesis of compound 10 with compound 108

(Scheme 25).^[80]Given the complexity of compound 10,

some effort was necessary to obtain good results for this

transformation. The results are summarized in Table 4.

Whereas Hoveyda–Grubbs 2nd-generation catalyst al-

lowed the formation of compound 107 in good yield

(Table 4, entry 1), the E/Z ratio varied over a broad range,

whereas it remained more constant by applying Grubbs

2nd-generation catalyst (Table 4, entries 2–4). However, we

found that an increased catalyst loading (10 mol%) was es-

sential for ensuring reproducibly high yields and full conver-

sion (Table 4, entry 4).

The next objective was the assembly of macrocycle 106.

As shown in Scheme 26, sterically hindered secondary alco-

hol 10 was first esterified with acid 8 by using Yamaguchi’s

reagent in excellent yield.^[81]The obtained ester (109), which

contained both vinyl iodine and terminal olefin groups, was

connected to compound 107 by applying a Suzuki coupling

Scheme 24. Conversion of the undesired (epi-18)-104 into building block

10: brsm=based on recovered starting material, PCC=pyridinium chlo-

ACHTUNGTRENNUNGroACHTUNGTRENNUNGchromate.

With compound 104 in hand, we aimed to invert the ste-

reocenter at the C18 atom. Because Mitsunobu inversions

are known to be hampered if the alcohol is sterically too

burdened,^[78]we envisaged an oxidation/asymmetric reduc-

tion sequence. The oxidation to ketone 105 was best

achieved by heating compound 104 at reflux with pyridini-

um chlorochromate in CH₂Cl₂(97%), whilst other oxidation

methods, such as Swern or IBX, afforded lower yields. Next,

the diastereoselective reduction of ketone 105 was exam-

ined, which was found to be difficult in the beginning

(Table 3). Reduction with sodium borohydride, DIBAL, or

[79]

(R)-Me-CBS oxazaborolidine and BH₃·SMe₂

showed no

Chem. Eur. J. 2013, 19, 15993 – 16018

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

16007

D. Menche et al.

Table 4. Conditions for the cross-metathesis reaction between com-

pounds 10 and 108.

Entry

Catalyst

Yield of

E/Z

loading [mol%]

107 [%]

ratio

1

2

3

4

Hoveyda–Grubbs-II

Grubbs-II

5

2

5

64–77

11

60–80

80–95

5:1–24:1

8:1–16:1

Grubbs-II

10

reaction with [PdACTHNUTRGNENUG(dppf)Cl₂] and Ba(OH)₂·8H₂O as a base to

furnish diene 110 in high yield. The E/Z selectivity of this

coupling process corresponded to the E/Z ratio of the em-

ployed vinyl boronate (107, E/Z=16:1), as judged by

¹H NMR spectroscopy. Next, another Yamaguchi esterifica-

tion reaction with acid 8 under the same conditions fur-

nished cyclization precursor 111 in very good yield. Whilst

these events proceeded without the need for optimization,

the envisaged final intramolecular Heck macrocyclization

was difficult in the beginning.

Figure 7. Retrosynthetic analysis of macrocycle 106.

Scheme 26. Heck macrocyclization strategy to prepare macrocycle 106.

DMAP=4-dimethylaminopyridine, TCBCl=2,4,6-trichlorobenzoyl chlo-

ride.

Scheme 25. Preparation of vinyl boronate 107.

16008

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 15993 – 16018

Modular Total Synthesis of Rhizopodin

FULL PAPER

A variety of conventional conditions^[82,83]resulted in ex-

tensive decomposition of the starting material (Table 5, en-

tries 1 and 2) and only trace amounts of the desired macro-

cycle (106) could be detected when the conditions described

by Fu et al.^[84]were used (Table 5, entry 3). Much to our de-

light, we found that optimized conditions described by Jeff-

ery^[85]in terms of temperature and reaction time allowed the

formation of compound 106 in 77% yield and E/Z selectivi-

ty of 5:1 in favor of the desired product (106; Table 5,

entry 4). In addition, the undesired Z isomer could be re-

moved by careful column chromatography on silica gel. No-

tably, too-long reaction times or higher temperatures result-

ed in decomposition of the material.

The overall strategy towards compound 106 relied on an

early-stage oxazole formation by employing Krische allyla-

tion, a highly convergent aldol-coupling reaction with fur-

ther elaboration to building block 10, and a sequential

cross-coupling strategy, which provided macrocyclic com-

pound 106 in 17 linear steps and an overall yield of 11%.^[18b]

Moreover, because only acid 8 was required for the C1–C7

building block, selective cleavage of the ester group was un-

necessary, which added to the overall effectiveness of this

entry to the macrocyclic core of rhizopodin. This route was

scalable and considerable amounts of compound 106 could

be prepared (100 mg).

Scheme 27. Attempted deprotection of the PMB groups on model system

10 and macrocycle 109.

Table 5. Conditions for the Heck macrocyclization reaction of precursor 111.

Not only was the deprotec-

Solvent^[86]

T [8C]

t

Yield

106 [%]

tion of macrocycle 106 prob-

lematic, additional difficulties

arose in the reproducibility of

good yields in the Heck macro-

cyclization step, depending on

the source of the chemicals (Pd-

Entry

“PdL_n”

Additive(s)

1

2

3

4

N

(CH₃CN)₂]

NEt₃, HCO₂H

K₂CO₃, NBu₄Cl

cHex₂NMe

MeCN

DMF

1,4-dioxane

DMF

RT

70

RT

60

2 h

1 h

24 h

50 min

dec.

N

trace

N

K₂CO₃, NBu₄Cl

77^[87]

[a] dba=dibenzylideneacetone.

AHCTUNGTRENNUNG

However, at this stage, all of our efforts to further ad-

vance compound 106 were thwarted by a seemingly impossi-

ble removal of the PMB group. In detail, we examined com-

pound 10 as a simple model system to find suitable condi-

tions for PMB deprotection, which could then be applied to

macrocycle 106 (Scheme 27).

As shown in Table 6, PMB ether 10 could be easily depro-

tected under standard conditions by using oxidative (DDQ;

Table 6, entry 1) or reductive methods (lithium naphtha-

of PMB-macrocycle 106 could be obtained without any rea-

sonable explanation, which prompted us not only to re-eval-

uate the protection-group strategy, but also to slightly adopt

our sequence to construct the macrocycle.

Suzuki macrocyclization strategy: To circumvent the depro-

tection problems, a terminal TBS group was chosen, which

should be more readily removable. As shown in Scheme 28,

we cleaved the PMB protecting group and transformed com-

pound 10 into its TBS congener (11). The conditions of

choice involved MgBr₂·OEt₂and dimethylsulfide and subse-

quent selective TBS protection of the primary C22 alcohol,

which was best achieved by using TBSOTf and 2,6-lutidine

at À788C in high yield. The corresponding vinyl boronate

(114) could again be prepared from compound 11 by cross-

metathesis with boronate 108. This metathesis step required

the same catalyst loading (10 mol%) as described for com-

pound 10; however, Grubbs 2nd-generation catalyst gave

lower yields (53–65%) and some side-reactions, such as

C=C isomerization, occurred. In contrast, Hoveyda–Grubbs

2nd-generation catalyst in the presence of 2,6-dichloroben-

ACHTUNGTRENNUNGlenide, entry 13). High yields of compound 112 were also

obtained by using mild Lewis acidic conditions (Table 6,

entry 2). However, all of these methods, as well as many

others, failed in the case of compound 106.

We realized that the difficulty in deprotecting both PMB

groups of macrocycle 106 resulted from the higher reactivity

of the C5–C9 diene moiety, which was substituted with a ho-

moallylic methoxy-ether group, as discussed elsewhere.^[18b]

In principle, various procedures exist for the deprotection of

PMB-containing compounds in the presence of diene moie-

ACHTUNGTRENNUNG

ties.^[81,88]However, these examples were performed with

alkyl-substituted diene substrates, which might possibly have

rendered them more stable.

Chem. Eur. J. 2013, 19, 15993 – 16018

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

16009

D. Menche et al.

Table 6. Attempted conditions for the deprotection of the PMB groups on model system 10 and macrocycle

106.^[a]

acid 8 to give compound 115

under the above-described con-

ditions in excellent yield

(Scheme 29).

We employed methyl ester 9

in this first Suzuki coupling re-

action because earlier investiga-

tions had shown that the corre-

sponding tert-butyl ester could

not be cleaved under Lewis

acidic conditions (TMSOTf,

TESOTf, TBSOTf) or by using

Entry

Reagent(s)

Solvent

T [8C]

Yield

10 [%]

106 [%]

1

2

3

4

5

6

7

8

DDQ

CH₂Cl₂/pH-buffer (1:1)

CH₂Cl₂

0

68

95

dec.

n.c.

dec.

n.c.

dec.

n.c.

MgBr₂·OEt₂, SMe₂

TMSOTf, lutidine

SnCl₄

SnCl₄, PhSH

SnCl₂, TMSCl, anisole

CAN

RT

À78

RT

0

CH₂Cl₂

MeCN/H₂O (1:1)

MeCN/pH-buffer (1:1)

CH₂Cl₂

1,4-dioxane

MeOH,

then CH₂Cl₂

CH₂Cl₂

THF

CAN

0

9

10

11

TFA (10% aq solution)

TfOH, tolSO₂NH₂

CSI, Na₂CO₃,

then NaOH

Ph₃CBF₄

RT

40

30

Brønsted

acids

(HCO₂H,

AcOH) without extensive de-

composition, again presumably

owing to the instability of the

12

13

RT

À78 to RT

90

75

n.c.

dec.

Li/C₁₀H₈

[a] n.c.=no conversion. DDQ=2,3-dichloro-5,6-dicyano-1,4-benzoquinone, TMSOTf=trimethylsilyltrifluoro-

methanesulfonate, CAN=ceric ammonium nitrate, TFA=trifluoroacetic acid, TfOH=trifluoromethanesul-

fonic acid, CSI=chlorosulfonyl isocyanate.

a-methoxy-substituted

diene

moiety. With methyl ester 115

in hand, selective ester cleavage

was achieved under basic condi-

zoquinone^[89]seemed to suppress side reactions and gave

compound 114 in almost quantitative yield.

tions. Good yields were obtained by using lithium hydroxide

in a mixture of THF/MeOH/H₂O 4:4:1 at ambient tempera-

With regards to the good results that we had previously

obtained for the Suzuki and Yamaguchi reactions, the

second strategy for the construction of the macrocyclic core

of rhizopodin now relied on an envisaged Suzuki macrocyc-

lization step. For this purpose, vinyl boronate 114 was cross-

coupled with vinyl iodine 9 and subsequently esterified with

ture

overnight.^[90]

Milder

conditions

by

using

Ba(OH)₂·8H₂O gave comparable results, but required

Scheme 28. Transformation of compound 10 into its TBS-congener (11)

and cross-metathesis to afford vinyl boronate 114.

Scheme 29. Suzuki macrocyclization strategy to access compound 117.

16010

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 15993 – 16018

Modular Total Synthesis of Rhizopodin

FULL PAPER

longer reaction times (75%, 7 days).^[87]Afterwards, the ob-

tained acid was esterified with boronate 114 in acceptable

yield (60%) to afford macrocyclization precursor 116. After

some optimization in terms of the catalyst and base loadings

(Table 7), the desired macrocyclic compound (117) was re-

ceived in reproducibly good yields (60–68%).

Endgame strategies towards rhizopodin: Having accom-

plished the preparation of TBS-protected macrocycle 117,

we sought to attach both side chains in a bidirectional

manner. The most convergent approach would rely on the

introduction of an entire side chain that already contained

the labile N-vinyl-formamide motif. We aimed to achieve

this connection by means of an aldol-condensation reaction.

As discussed above, a less convergent alternative would in-

Table 7. Optimization of the Suzuki macrocyclization reaction of com-

pound 116.^[a]

volve

(Figure 2).

a HWE reaction with b-keto phosphonate 5

Entry

ACHTUNGTRENNUNG[PdCl₂ACHTUNGTRENNUNG

(dppf)] [equiv] Conditions ([equiv])^[91]Yield of 120 [%]

1

2

3

0.3

0.5

1.0

Ba(OH)₂·8H₂O (3.0)

Ba(OH)₂·8H₂O (5.0)

Ba(OH)₂·8H₂O (10.0) 35–50

20–43

60–68

Model study of an aldol-condensation approach to attach the

side chains: In view of the relatively small amount of com-

pound 117 that was initially available, we decided to first

use a truncated aldehyde (119, available from compound 73

in three steps) as a model substrate to examine the pivotal

aldol coupling with side-chain fragment 4 (Scheme 31). By

[a] dppf=1,1’-bis(diphenylphosphino)ferrocene.

It should be mentioned that we also tried to synthesize

the respective macrocyclic analogue of compound 117 bear-

ing terminal TES protecting groups. To this end, the C22-

TES-congener of compound 114 was prepared and an analo-

gous strategy was pursued (Scheme 29). However, the labili-

ty of primary TES-ethers under basic conditions made the

selective methyl-ester cleavage of the TES-analogue of com-

pound 115 impossible in our hands and, thus, we discarded

this approach.

Yamaguchi macrolactonization strategy: Likewise, an alter-

native Yamaguchi macrolactonization strategy was also eval-

uated. As shown in Scheme 30, vinyl iodine 114 was coupled

Scheme 31. Model system to investigate the attachment of side chain 4

through an aldol-coupling strategy. Burgess reagent=methyl N-(triethyl-

Scheme 30. Yamaguchi macrolactonization strategy to access compound

117.

A

ACHTUNGTREN[NUNG a,a-bis(trifluoro-

with vinyl boronate 115 to give methyl ester 118 in high

yield. However, it turned out that cleavage of the methyl

ester by using either lithium hydroxide or barium hydroxide

gave consistently low to moderate yields. Essentially neutral

conditions by using bis(tributyltin) oxide^[92]resulted in de-

composition of the material. The macrolactonization itself

proceeded in good yield (47%).^[93]However, this route was

not pursued further, owing to the higher reliability of the

Suzuki macrocyclization strategy as described above, which

allowed the preparation of sufficient amounts of TBS-pro-

tected macrocycle 117 (120 mg).

using conditions that had been previously described by Pa-

terson and co-workers, we smoothly prepared compound

120 as a single diastereomer in 66% yield.^[18,94]Likewise,

compound 119 was treated with the lithium enolate of com-

pound 4, which also resulted in the formation of compound

120, albeit in lower yield (47%, 75% over two cycles) and

only modest diastereoselectivity.

Next, compound 120 needed to be dehydrated to afford

the corresponding enone. First, we successfully applied a pro-

cedure that involved acetate formation with acetic anhy-

Chem. Eur. J. 2013, 19, 15993 – 16018

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

16011

D. Menche et al.

dride and DBU-mediated elimination.^[81]Alternatively, we

examined the feasibility of Burgess’ reagent^[95]for this trans-

formation. For successful application, it was of vital impor-

tance to use a freshly prepared reagent that could be stored

at À208C for a maximum of about three weeks whilst the

commercially available material resulted in decomposition

of the starting material. Finally, we found that Martinꢅs sul-

furane^[96]was also suitable for performing the desired elimi-

nation process.

For the subsequent selective 1,4-reduction of the respec-

tive enone, the application of Stryker’s reagent^[97]was the

method of choice and worked uneventfully. Because the sep-

aration of residual PPh₃was tedious, we aimed to perform

the direct deprotection of all of the remaining silyl groups.

Because we expected the “real” system to be labile under

basic conditions, we tested TAS-F, which is known

to particularly tolerate base-labile substrates.^[98]

synthesis (Scheme 32). To this end, we needed to selectively

deprotect both primary TBS groups of macrocyclic com-

pound 117. Based on our earlier experiences with PMB de-

protection, we mostly evaluated non-basic and non-Lewis

acidic conditions, owing to the labile diene moiety, and

turned our attention to Brønsted acidic methods. Pyridine-

buffered HF/pyridine solution only allowed the detection of

trace amounts of the desired macrocyclic diol (122) by using

ESI-HRMS (Table 8, entry 1). Moreover, treatment with

CSA, both in catalytic amounts and in excess, only resulted

in very small quantities of the desired product besides de-

composition of the material (Table 8, entries 2 and 3). Less

basic TBAF buffered with acetic acid^[99]gave compound 122

in low yield, besides decomposition (Table 8, entry 4). Grati-

fyingly, the use of 3m HCl in MeOH at À208C resulted in

Table 8. Selective deprotection of the terminal TBS groups of macrocycle 117.

Treatment of the crude reduction product with

excess TAS-F resulted in the clean formation of

diol 121 in 72% yield over two steps.

Entry Reagent ([equiv])

Solvent

T [8C]

t

Yield

122 [%]

1

2

3

4

5

6

HF-pyridine

CSA (0.3)

CSA (3.0)

pyridine/THF (1:1)

MeOH

0

24 h

1.5 h

5 h

trace

5

Aldol condensation strategy towards rhizopodin:

The successful development of an effective se-

quence to afford model compound 121 prompted us

to apply this strategy to the final steps of the total

TBAF/AcOH (1:1, 10.0) THF

RT

0

À20

24 h

30 min 10–20

60 min 50–60

30–40

3m HCl/MeOH

THF

Scheme 32. Attempted aldol-condensation strategy with macrocyclic compound 117.

16012

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 15993 – 16018

Modular Total Synthesis of Rhizopodin

FULL PAPER

the formation of compound 122 in acceptable yield (50–

60%), following careful adjustment of the amount of that

was HCl added (Table 8, entry 6). If the temperature was in-

creased to 08C, undesired deprotection of presumably the

C3-OTBS groups and decomposition were observed, with

the formation of only small amounts of the desired com-

pound (122; Table 8, entry 5). As mentioned above, we also

tried to prepare the corresponding terminal TES-protected

macrocycle. However, several of our established synthetic

methods failed or gave significant lower yields, presumably

owing to the lability of primary TES groups.

Then, macrocyclic diol 122 was oxidized to bis-aldehyde

3, which was best accomplished by using buffered Dess–

Martin periodinane, which proceeded in excellent yield. No-

tably, to obtain good and reproducible yields, the reagent

had to be freshly prepared and used within about four

weeks.^[100]Parikh–Doering oxidation or treatment with

TPAP/NMO only gave low yields of the desired compound

(3).

In our model study (Scheme 31), this transformation had

been effected in very good yields with several reagents.

However, much to our disappointment, we were unable to

successfully apply the dehydration conditions that we devel-

oped for compound 120 in a reliable fashion to compound

123. The two-step procedure (Ac₂O then DBU), Burgess’ re-

agent, and Martinꢅs sulfurane all failed to give rise to com-

pound 124. The Paterson group has reported a successful ap-

plication of this strategy to a very similar substrate (the

C16/C16’-bis-TES analogue).^[101]After considerable efforts,

compound 123 had to be considered as a dead end, which

prompted us to revise our endgame strategy.

HWE strategy and completion of the total synthesis: Thus,

we turned our attention to an alternative HWE reaction,

which utilized side-chain fragment 5 (Figure 2). This route

required the introduction of the labile N-methyl-vinyl-for-

mamide moiety at a late stage in the entire synthesis. Grati-

fyingly, the double-HWE reaction could be successfully im-

plemented by treating dimethyl phosphonate 5 under anhy-

drous conditions with BaO and the subsequent addition of

With bis-aldehyde 3 in hand, the bidirectional aldol cou-

pling reaction with compound

4

could be tackled

(Figure 2).^[18]However, the boron-mediated coupling of

compound 3 with compound 4 under the above-evaluated

conditions was extremely unreliable and we were only able

to perform this coupling reaction in high yield on one occa-

sion, despite considerable efforts (Scheme 32). Indeed, the

use of freshly prepared Cy₂BCl,^[89]performing the reaction

over CaH₂as a drying reagent, co-evaporating the reactants

with dry toluene under an argon atmosphere to remove

trace amounts of water, and employing completely fresh

chemicals did not render this transformation more reliable

and, typically, only trace amounts of the desired bis-aldol

product (123) could be detected by ESI-HRMS. It should be

mentioned, that Paterson et al. described the successful im-

plementation of this aldol-coupling strategy with the same

side-chain compound (4) on a very similar macrocyclic bis-

aldehyde that only differed in the nature of the silyl-protect-

ing groups. However, the same aldol reaction was also unre-

liable in our model system (119, Scheme 31). After examin-

ing several different model systems, we assumed that the N-

vinyl-formamide moiety was responsible for the inconstant

outcome in the boron-mediated aldol-coupling reactions. By

applying the previously developed conditions with the Li-

enolate of compound 4, we obtained the desired bis-aldol

product (123) in 25% yield, together with the respective

mono-aldol product, which only had one side chain attached

(50% yield). Owing to the high polarity of these com-

pounds, their separation was difficult, but could be achieved

by preparative TLC. The a-methyl aldehyde moiety of com-

pound 3 and the respective mono-aldol product were prone

to racemization at the C21-methyl stereocenter within about

24 h, thus impeding the recovery of these compounds and

their reuse in another lithium-aldol reaction. However, de-

spite these difficulties, we were eventually capable of obtain-

ing useful amounts of material to investigate the envisaged

b-hydroxy elimination at the C22 position to form the re-

spective bis-enone (124).

bis-aldehyde

3

in an THF/H₂O mixture (40:1,

Scheme 33).^[102]The obtained bis-enone was subsequently re-

Scheme 33. HWE strategy to attach the side chain and further elabora-

tion to afford diol 125.

Chem. Eur. J. 2013, 19, 15993 – 16018

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

16013

D. Menche et al.

duced in quantitative yield by using Stryker’s reagent, as

previously evaluated with model compound 120

(Scheme 31). Treatment of the respective bis-TES ether

with catalytic amounts of CSA in MeOH furnished diol 125

in very good yield (67%, 3 steps).

Next, our strategy relied on first introducing the N-vinyl-

formamide motifs and then removing the TBS groups.

During the preparation of side-chain building block 4, which

contained the N-vinyl-formamide unit, we became aware of

the difficulties in preparing this structural motif in accepta-

ble reproducible yields. With the rationale of not squander-

ing any material of diol 125, we used several model systems

to further investigate this transformation (126, 128, and 129;

Scheme 34). The commonly described literature procedure

Scheme 35. Model study to evaluate various conditions for global depro-

tection.

As shown in Table 9, a variety of reagents were investigat-

ed and the products were analyzed by crude NMR spectros-

copy and by ESI-HRMS. However, these reactions mostly

Table 9. Deprotection of the TBS groups in macrocycle 106.^[a]

Entry Reagent ([equiv]) Solvent T [8C] t [h] Result

1

2

3

4

5

6

7

8

HCl/MeOH

HF-pyridine

HF

H₂SiF₆

CsF, MeOH

THF

0

RT

0

RT

60

RT

70

1

72

4

14

72

24

2

dec.

n.c.

dec.

n.c.

dec.

n.c.

dec.

MeCN

THF

Bu₄N

TBAF, AcOH

TBAF(tBuOH)₄

A

Scheme 34. Model systems to investigate the formation of the N-vinyl

formACHTUNGTRENNUNGamide.

A

THF

9

10

11

TBAF, SiO₂

TBAF

TAS-F (25)

THF

DMF

RT

60

20

72

n.c.

131 (À4TBS)

À2TBS, trace

amounts of 131

dec.

for this reaction requires heating a mixture of the aldehyde

and N-methyl formamide (NMF) in the presence of catalytic

amounts of Brønsted acids, such as PPTS or p-toluenesul-

fonic acid (PTSA). An important issue was the efficient re-

moval of the water that was generated within this condensa-

tion process, either by using molecular sieves or by azeo-

tropic distillation.^[103]However, on applying these proce-

dures to our model systems, we obtained varying results.

Even in the case of simple aldehyde 126, only low yields or

even no conversion were observed. Whereas the results with

compounds 128 and 129 were somewhat more promising,

small-scale reactions remained problematic. Furthermore,

we were concerned about the stability of the macrocycle

under these rather harsh and acidic conditions. In contrast,

another procedure that used phosphorus(V) oxide as a dehy-

drating agent and a large excess of NMF^[104]gave reliable

yields of 50–70% of the respective desired N-vinyl form-

12

13

14

TAS-F (100)

dry TAS-F (20)

DMF

THF

MeCN

RT

72

24

72

À1/À2TBS

À2TBS

[a] n.c.=no conversion; dec.=no defined compound(s) of reasonable

molecular weight could be detected. TBAF=tetrabutylammonium fluo-

ride,

difluoro

AcOH=acetic

acid,

TAS-F=tris(dimethylamino)sulfonium

AHCTUNGTRENNUNG

resulted in either no conversion (Table 9, entries 2, 5, 7, and

9) or decomposition without any distinct product formation

(Table 9, entries 1, 3, 4, 6, 8, and 12). When TAS-F was em-

ployed (Table 9, entry 11), trace amounts of the desired fully

deprotected macrocycle could be detected; however, the

main product was a compound that lacked two TBS groups.

We assumed that the product was a C3/C3’-TBS-deprotected

compound, in which the sterically hindered C16/C16’-TBS-

ethers had essentially remained untouched. Modification of

the reaction conditions in terms of the solvent and the

number of equivalents of TAS-F remained fruitless. Eventu-

ally, we found that TBAF was suitable for liberating all of

the OH groups, with the formation of the fully TBS-depro-

tected macrocycle (131; Table 9, entry 10).^[105,106]

ACHTUNGTRENNUNGamides, accompanied by undefined byproducts that were de-

rived from the P₂O₅/NMF mixture.

In a likewise manner, we decided to first carry out

a model study to establish successful conditions for the re-

moval of the four TBS groups; however, this transformation

was found to be extremely challenging. Thus, we chose

PMB-protected macrocycle 106 as a model substrate for

evaluating the promising reagents (Scheme 35).

The results from these model studies prompted us to

tackle the remaining steps of the total synthesis. Thus, first,

16014

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 15993 – 16018

Modular Total Synthesis of Rhizopodin

FULL PAPER

we oxidized diol 125 (Scheme 36) by using DMP into the

corresponding bis-aldehyde in quantitative yield, which was

highly unstable and, hence, was directly used without further

Much to our delight, application of the aforementioned

procedure (P₂O₅in neat NMF) to the bis-aldehyde that was

derived from compound 125 reliably furnished the desired

bis-N-vinyl formamide (132, Scheme 36) in acceptable con-

version, as judged by crude NMR spectroscopic analysis, as

well as undefined products that were derived from P₂O₅/

NMF. However, compound 132 showed some “tailing”

during TLC analysis, thereby impeding its clean purification

without the considerable loss of material. However, it was

possible to isolate small quantities of pure compound 132

1

for its characterization by H NMR spectroscopy and ESI-

HRMS. To not squander any of the obtained material, we

decided to employ the crude mixture that was obtained

from the condensation reaction with NMF/P₂O₅in the final

global deprotection step.

The silyl-deprotection conditions that were developed in

our model study with macrocyclic compound 106 had shown

that TBAF could be employed to liberate all of the TBS

protecting groups and, gratifyingly, as shown in Scheme 36,

the treatment of crude compound 132 with TBAF in THF at

room temperature allowed the isolation of rhizopodin (2)

after HPLC purification in 14% yield over the final two

steps.^[101,107]The cleavage of the extremely hindered C16/

C16’-OTBS groups was definitely the most challenging

transformation within these final steps and was accompanied

by partial decomposition. However, the obtained characteri-

zation data (¹H and ¹³C NMR spectroscopy and ESI-

HRMS) of the isolated synthetic material were in excellent

accordance with the reported data for natural rhizopodin.

Furthermore, the optical rotation value was also in full

agreement with the literature data. Therefore, the structural

assignment of the natural product was fully confirmed.

Conclusion

In summary, a highly convergent synthesis of the polyketide

macrolide rhizopodin has been accomplished in a longest

linear sequence of 29 steps with an overall yield of 0.25%,

which has unequivocally established its absolute and relative

configurations. Along the way, we have developed two inde-

pendent routes for the synthesis of the challenging central

C8–C22 fragment and an efficient synthesis of the C1–C7

building block. These fragments could be successfully used

to forge the macrocyclic core of rhizopodin through Heck or

Suzuki macrocyclization reactions, as well as through a Ya-

maguchi macrolactonization reaction. Two different C23–

C29 side chains were synthesized, which could be introduced

by using either aldol or HWE coupling approaches. The

HWE strategy was found to be superior and the resulting

material could be successfully elaborated into the authentic

natural product by the late-stage introduction of labile N-

vinyl formamide motifs. The outlined modular synthesis of

rhizopodin should be instructive and amenable towards the

development of potent analogues of this actin-polymeri-

zation inhibitor. Studies concerning the design and synthesis

Scheme 36. Introduction of the N-vinyl formamide units and final depro-

tection.

characterization. The oxidizing reagent had to be freshly

prepared to obtain good results. Treatment with tetrapropyl-

ACHTUNGTRENNUNG

methylpiperidin-1-yl)oxidanyl (TEMPO)/PhI

much less effective.

ACHTUNGRTENN(UNG OAc)₂were

Only decomposition was observed when the macrocyclic

bis-aldehyde that was derived from compound 125 was

heated in the presence of N-methyl formamide and PPTS,

presumably owing to the instability of the diene moiety

under these conditions. These conditions had been success-

fully applied by Nicolaou et al. in their synthesis of mono-

rhizopodin.^[15]

Chem. Eur. J. 2013, 19, 15993 – 16018

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

16015

D. Menche et al.

relied on a stepwise and more flexible strategy for the construction

of compound 2.

[21] A. F. Salit, C. Meyer, J. Cossy, B. Delouvrie, L. Hennequin, Tetra-

hedron 2008 , 64 , 668 4 –6 697.

of such analogues are currently under investigation in our

laboratories.

[22] a) K. Matsumura, S. Hashiguchi, T. Ikariya, R. Noyori, J. Am.

Chem. Soc. 1997 , 119 , 873 8 –8 739; b) R. Noyori, S. Hashiguchi,

Acc. Chem. Res. 1997 , 30 , 9 7 –1 02.

Acknowledgements

[23] a) V. I. Tararov, G. Kçnig, A. Bçrner, Adv. Synth. Catal. 2006 , 348 ,

263 3 –2 644; b) A. Fꢁrstner, M. Albert, J. Mlynarski, M. Matheu, E.

DeClercq, J. Am. Chem. Soc. 2003 , 125 , 1313 2 –1 3142.

[24] a) M. Kitamura, M. Tokunaga, T. Ohkuma, R. Noyori, Org. Synth.

1993, 71, 1–7; b) M. Kitamura, M. Tokunaga, T. Ohkuma, R.

Noyori, Tetrahedron Lett. 1991 , 32 , 416 3 –4 166.

We thank the DFG (Me 2756/4-1), the DFG Graduiertenkolleg 850

(scholarships to M.K. and M.D.), and the “Fonds der Chemischen Indus-

trie” for their generous funding and Wiebke Ahlbrecht, Carolin Lang,

and Nathalie-Desirꢀe Costa Pinheiro for performing the exploratory

studies. Excellent assistance with the NMR spectroscopic analysis by

Prof. M. Enders and B. Termin is gratefully acknowledged, as are the

helpful discussions and assistance with the HPLC analysis by Prof. G.

Helmchen.

[25] M. E. Fox, M. Jackson, I. C. Lennon, R. McCague, J. Org. Chem.

2005 , 70 , 122 7 –1 236.

[26] Only a few examples of the formation of related side-products

have been reported; see: ref. [22].

[27] J. Einhorn, C. Einhorn, F. Ratajczak, J. L. Pierre, J. Org. Chem.

1996 , 61 , 745 2 –7 454.

[28] K. Takai, K. Nitta, K. Utimoto, J. Am. Chem. Soc. 1986 , 108 , 740 8 –

7410.

[1] a) H. Reichenbach, J. Ind. Microbiol. Biotechnol. 2001 , 27 , 14 9 –

156; b) S. C. Wenzel, R. Mꢁller, Nat. Prod. Rep. 2007 , 24 , 121 1 –

1224; c) G. Rꢁckert, Z. Pflanzenernaehr. Bodenk. 1979 , 142 , 33 0 –

343.

[29] a) N. Hashimoto, T. Aoyama, T. Shioiri, Chem. Pharm. Bull. 1981 ,

29 , 147 5 –1 478; b) E. Kꢁhnel, D. D. P. Laffan, G. C. Lloyd-Jones, T.

Martꢆnez del Campo, I. R. Shepperson, J. L. Slaughter, Angew.

Chem. 2007 , 119 , 720 5 –7 208; Angew. Chem. Int. Ed. 2007 , 46 ,

707 5 –7 078.

[2] K. Weissman, R. Mꢁller, Nat. Prod. Rep. 2010 , 27 , 127 6 –1 295.

[3] F. Sasse, H. Steinmetz, G. Hçfle, H. Reichenbach, J. Antibiot. 1993 ,

46 , 74 1 –7 48.

[4] T. M. A. Gronewold, F. Sasse, H. Lꢁnsdorf, H. Reichenbach, Cell

Tissue Res. 1999 , 295 , 12 1 –1 29.

[5] a) J. S. Allingham, V. A. Klenchin, I. Rayment, Cell. Mol. Life Sci.

2006 , 63 , 211 9 –2 134; b) K.-S. Yeung, I. Paterson, Angew. Chem.

2002 , 114 , 482 6 –4 847; Angew. Chem. Int. Ed. 2002 , 41 , 463 2 –4 653.

[6] a) J. Rao, N. Li, Curr. Cancer Drug Targets 2004, 4, 345–354;

b) M. A. Jordan, L. Wilson, Curr. Opin. Cell Biol. 1998 , 10 , 12 3 –

130.

[30] For a comprehensive review, see: J. Li, D. Menche, Synthesis 2009,

2293–2315.

[31] A. Abiko, J.-F. Liu, S. Masamune, J. Am. Chem. Soc. 1997 , 119 ,

258 6 –2 587; b) T. Inoue, J.-F. Liu, D. C. Buske, A. Abiko, J. Org.

Chem. 2002 , 67 , 525 0 –5 256; c) A. Abiko, J.-F. Liu, J. Org. Chem.

1996 , 61 , 259 0 –2 591.

[7] N. Klippel, U. Bilitewski, Anal. Lett. 2007 , 40 , 140 0 –1 411.

[8] R. Jansen, H. Steinmetz, F. Sasse, W.-D. Schubert, G. Hagelꢁken,

S. C. Albrecht, R. Mꢁller, T e trahedron Lett. 2008 , 49 , 579 6 –5 799.

[9] N. Horstmann, D. Menche, Chem. Commun. 2008 , 517 3 –5 175.

[10] G. Hagelueken, S. C. Albrecht, H. Steinmetz, R. Jansen, D. W.

Heinz, M. Kalesse, W.-D. Schubert, Angew. Chem. 2009 , 121 , 60 3 –

606; Angew. Chem. Int. Ed. 2009 , 48 , 59 5 –5 98.

[11] D. Pistorius, R. Mꢁller, ChemBioChem 2012 , 13 , 41 6 –4 26.

[12] J. Piel, Nat. Prod. Rep. 2010 , 27 , 99 6 –1 047.

[13] a) R. Reid, M. Piagentini, E. Rodriguez, G. Ashley, N. Viswana-

than, J. Carney, D. V. Santi, R. Hutchinson, R. McDaniel, Biochem-

istry 2003 , 42 , 7 2 –7 9; b) P. Caffrey, ChemBioChem 2003 , 4 , 65 4 –

657.

[14] a) Z. Cheng, L. Song, Z. Xu, T. Ye, Org. Lett. 2010 , 12 , 203 6 –2 039;

b) T. Chakraborty, K. K. Pulukuri, M. Sreekanth, Tetrahedron Lett.

2010 , 51 , 644 4 –6 446; c) T. K. Chakraborty, M. Sreekanth, K. K. Pu-

lukuri, Tetrahedron Lett. 2011 , 52 , 5 9 –6 1; d) K. K. Pulukuri, T. K.

Chakraborty, Org. Lett. 2012 , 14 , 285 8 –2 861.

[15] K. C. Nicolaou, X. Jiang, P. J. Lindsay-Scott, A. Corbu, S. Yama-

shiro, A. Bacconi, V. M. Fowler, Angew. Chem. 2011 , 123 , 117 1 –

1176; Angew. Chem. Int. Ed. 2011 , 50 , 113 9 –1 144.

[16] M. Dieckmann, M. Kretschmer, P. Li, S. Rudolph, D. Herkommer,

D. Menche, Angew. Chem. 2012 , 124 , 576 5 –5 768; Angew. Chem.

Int. Ed. 2012 , 51 , 566 7 –5 670.

[17] S. M. Dalby, J. Goodwin-Tindall, I. Paterson, Angew. Chem. 2012,

125, 6645–6649; Angew. Chem. Int. Ed. 2013 , 52 , 651 7 –6 521.

[18] Part of these results have been communicated previously, see

ref. [16] and: a) M. Kretschmer, D. Menche, Org. Lett. 2012 , 14 ,

38 2 –3 85; b) M. Dieckmann, S. Rudolph, S. Dreisigacker, D.

Menche, J. Org. Chem. 2012 , 77 , 1078 2 –1 0788.

[32] J. Li, D. Menche, Synlett 2009, 2417–2420.

[33] C. G. Knudsen, H. Rapoport, J. Org. Chem. 1983 , 48 , 226 0 –2 266.

[34] M. Toumi, F. Couty, G. Evano, Angew. Chem. 2007 , 119 , 57 8 –5 81;

Angew. Chem. Int. Ed. 2007 , 46 , 57 2 –5 75.

[35] M. C. Hillier, A. T. Price, A. I. Meyers, J. Org. Chem. 2001 , 66 ,

603 7 –6 045.

[36] S. Kiyooka, Y. Kaneko, M. Komura, H. Matsuo, M. Nakano, J.

Org. Chem. 1991 , 56 , 227 6 –2 278.

[37] K. Mikami, S. Matsumoto, A. Ishida, S. Takamuku, T. Suenobu, S.

Fukuzumi, J. Am. Chem. Soc. 1995 , 117 , 1113 4 –1 1141.

[38] M. C. Hillier, A. I. Meyers, Tetrahedron Lett. 2001 , 42 , 514 5 –5 147.

[39] a) T. Katsuki, K. B. Sharpless, J. Am. Chem. Soc. 1980 , 102 , 597 4 –

5976; b) R. M. Hanson, K. B. Sharpless, J. Org. Chem. 1986 , 51 ,

192 2 –1 925; c) Y. Gao, J. M. Klunder, R. M. Hanson, H. Masa-

mune, S. Y. Ko, K. B. Sharpless, J. Am. Chem. Soc. 1987 , 109 ,

576 5 –5 780.

[40] M. Miyashita, T. Suzuki, M. Hoshino, A. Yoshikoshi, Tetrahedron

1997 , 53 , 1246 9 –1 2486.

[41] W. C. Still, C. Gennari, Tetrahedron Lett. 1983 , 24 , 440 5 –4 408.

[42] a) M. A. Tius, A. H. Fauq, J. Org. Chem. 1983 , 48 , 413 1 –4 132;

b) M. R. Johnson, T. Nakata, Y. Kishi, Tetrahedron Lett. 1979 , 20 ,

434 3 –4 346.

[43] M. Sasaki, K. Tanino, M. Miyashita, Org. Lett. 2001 , 3 , 176 5 –1 767.

[44] K. Y.-K. Chow, J. W. Bode, J. Am. Chem. Soc. 2004 , 126 , 812 6 –

8127.

[45] a) D. A. Evans, J. Bartroli, T. L. Shih, J. Am. Chem. Soc. 1981 , 103 ,

212 7 –2 129; b) D. A. Evans, J. V. Nelson, E. Vogel, T. R. Taber, J.

Am. Chem. Soc. 1981 , 103 , 309 9 –3 111.

[46] a) M. T. Crimmins, B. W. King, E. A. Tabet, K. Chaudhary, J. Org.

Chem. 2001 , 66 , 89 4 –9 02; b) M. T. Crimmins, J. She, Synlett 2004,

1371–1374.

[47] K. Soai, A. Ookawa, J. Org. Chem. 1986 , 51 , 400 0 –4 005.

[48] M. Kçrner, M. Hiersemann, Org. Lett. 2007 , 9 , 497 9 –4 982.

[49] a) D. B. Grotjahn, D. A. Lev, J. Am. Chem. Soc. 2004 , 126 , 1223 2 –

12234; b) A. Labonne, T. Kribber, L. Hintermann, Org. Lett. 2006 ,

8 , 585 3 –5 856.

[19] a) I. Paterson, K. Ashton, R. Britton, G. Cecere, G. Chouraqui,

G. J. Florence, J. Stafford, Angew. Chem. 2007 , 119 , 627 9 –6 283;

Angew. Chem. Int. Ed. 2007 , 46 , 616 7 –6 171; b) I. Paterson, K.

Ashton, R. Britton, G. Cecere, G. Chouraqui, G. J. Florence, H.

Knust, J. Stafford, Chem. Asian J. 2008 , 3 , 36 7 –3 87.

[20] Because examples of similar dimerization strategies were scarce,

we did not consider such an approach to be promising and, hence,

16016

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 15993 – 16018

Modular Total Synthesis of Rhizopodin

FULL PAPER

[50] R. Takeuchi, S. Nitta, D. Watanabe, J. Org. Chem. 1996, 61, 3045–

3051.

[77] I. Paterson, M. Donghi, K. Gerlach, Angew. Chem. 2000 , 112 ,

345 3 –3 457; Angew. Chem. Int. Ed. 2000, 39, 3315–3319.

[78] S. F. Martin, J. A. Dodge, Tetrahedron Lett. 1991 , 32 , 301 7 –3 020.

[79] E. J. Corey, C. J. Helal, Angew. Chem. 1998 , 110 , 209 2 –2 118;

Angew. Chem. Int. Ed. 1998 , 37 , 198 6 –2 012.

[80] C. Morril, R. H. Grubbs, J. Org. Chem. 2003, 68, 6031–6034.

[81] J. Inanaga, K. Hirata, H. Saeki, T. Katsuki, M. Yamaguchi, Bull.

Chem. Soc. Jpn. 1979 , 52 , 198 9 –1 993.

[82] a) D. Menche, J. Hassfeld, J. Li, S. Rudolph, J. Am. Chem. Soc.

2007 , 129 , 610 0 –6 101; b) D. Menche, J. Hassfeld, J. Li, K. Mayer, S.

Rudolph, J. Org. Chem. 2009 , 74 , 722 0 –7 229.

[83] F. E. Ziegler, U. R. Chakraborty, R. B. Weisenfeld, T e trahedron

1981 , 37 , 403 5 –4 040.

[51] For a review on the scope of this transformation, see: a) G. R.

Jones, Y. Landais, Tetrahedron 1996 , 52 , 759 9 –7 662; b) K. Tamao,

M. Kumada, Tetrahedron Lett. 1984 , 25 , 32 1 –3 24; c) K. Tamao, N.

Ishida, T. Tanaka, M. Kumada, Organometallics 1983 , 2 , 169 4 –

1696.

[52] a) S. B. Han, I. S. Kim, H. Han, M. J. Krische, J. Am. Chem. Soc.

2009 , 131 , 691 6 –6 917; b) for a correction, see:S. B. Han, I. S. Kim,

H. Han, M. J. Krische, J. Am. Chem. Soc. 2010 , 132 , 12517.

[53] S. Shibahara, M. Fujino, Y. Tashiro, N. Okamoto, T. Esumi, T. Ta-

kahashi, J. Ishihara, S. Hatakeyama, Synthesis 2009, 2935–2953.

[54] X. Gao, I. A. Townsend, M. J. Krische, J. Org. Chem. 2011 , 76 ,

235 0 –2 354.

[84] G. C. Fu, A. F. Littke, J. Am. Chem. Soc. 2001, 123, 6989–7000.

[85] T. Jeffery, Tetrahedron 1996 , 52 , 1011 3 –1 0130.

[86] The reaction was performed at a concentration of 0.005m.

[87] The yield and outcome of this reaction strongly depended on the

[55] W. Yu, Y. Mei, Y. Kang, Z. Hua, Z. Jin, Org. Lett. 2004 , 6 , 321 7 –

3219.

[56] H. Li, X. Jiang, Y. Ye, C. Fan, T. Romoff, M. Goodman, Org. Lett.

1999 , 1 , 9 1 –9 3.

batch of PdACTHNUGRTENUNG(OAc)₂that was used; a reason for this observation

[57] a) P. Wipf, C. P. Miller, J. Org. Chem. 1993 , 58 , 360 4 –3 606;

b) M. C. Bagley, K. E. Bashford, C. L. Hesketh, C. J. Moody, J. Am.

Chem. Soc. 2000 , 122 , 330 1 –3 313.

could not be determined.

[88] M. Penner, V. Ranniyar, L. T. Kaspar, D. G. Hall, J. Am. Chem.

Soc. 2009 , 131 , 1421 6 –1 4217.

[58] Further examination of this transformation revealed that com-

pound 89 was a special case, because the treatment of similar com-

pounds (e.g., compound 76) under identical conditions only result-

ed in recovery of the starting materials (>95%).

[89] S. H. Hong, D. P. Sanders, C. W. Lee, R. H. Grubbs, J. Am. Chem.

Soc. 2005 , 127 , 1716 0 –1 7161.

[90] a) T. W. Greene, P. G. M. Wuts, Protective Groups in Organic Syn-

´

thesis 4th ed., Wiley, Stuttgart, 2006; b) P. J. Kocienski, Protecting

[59] U. Klar, B. Rçhr, F. Kuczynski, W. Schwede, M. Berger, W. Skubal-

la, B. Buchmann, Synthesis 2005 , 30 1 –3 05.

[60] F. N. Tebbe, G. W. Parshall, G. S. Reddy, J. Am. Chem. Soc. 1978 ,

100 , 361 1 –3 613.

Groups, 3rd ed, Thieme, New York, 2005.

[91] The reaction was performed at a concentration of 0.009m.

[92] C. J. Salomon, E. G. Mata, O. A. Mascaretti, J. Org. Chem. 1994 ,

59 , 725 9 –7 266.

[61] a) R. I. Storer, T. Takemoto, P. S. Jackson, S. V. Ley, Angew. Chem.

2003 , 115 , 262 5 –2 629; Angew. Chem. Int. Ed. 2003 , 42 , 252 1 –2 525;

b) R. I. Storer, T. Takemoto, P. S. Jackson, D. S. Brown, S. V. Ley,

Chem. Eur. J. 2004 , 10 , 252 9 –2 547.

[62] B. M. Trost, H. Yang, O. R. Thiel, A. J. Frontier, C. S. Brindle, J.

Am. Chem. Soc. 2007 , 129 , 220 6 –2 207.

[93] The reaction was performed at a concentration of 0.017m.

[94] Following an empirical model, we assumed that the configuration

at the C22 position arose from inherent substrate control. In this

case, both 1,4-syn and 1,5-anti induction would lead to the product

as shown, see: C. J. Cowden, I. Paterson, Org. React. 1997, 51, 1–

200.

[63] J. D. More, N. S. Finney, Org. Lett. 2002 , 4 , 300 1 –3 003.

[64] A. Hassan, Y. Lu, M. J. Krische, Org. Lett. 2009 , 11 , 311 2 –3 115.

[65] J. D. White, P. R. Blakemore, N. J. Green, E. B. Hauser, M. A. Hol-

oboski, L. E. Keown, C. S. Nylund-Kolz, B. W. Phillips, J. Org.

Chem. 2002 , 67 , 775 0 –7 760.

[66] V. Kumar, S. Dev, Tetrahedron Lett. 1983 , 24 , 128 9 –1 292.

[67] For a review of the application of this reaction in the synthesis of

natural products, see: M. Yus, C. Nꢇjera, F. Foubelo, Tetrahedron

2003 , 59 , 614 7 –6 212.

[95] a) E. M. Burgess, H. R. Penton, Jr., E. A. Taylor, J. Org. Chem.

1973 , 38 , 2 6 –3 1; for the preparation of the reagent see: b) G. M.

Atkins, E. M. Burgess, J. Am. Chem. Soc. 1968 , 90 , 474 4 –4 745.

[96] J. C. Martin, R. J. Arhart, J. Am. Chem. Soc. 1971 , 93 , 233 9 –2 341.

[97] W. S. Mahoney, D. M. Brestensky, J. M. Stryker, J. Am. Chem. Soc.

1988 , 110 , 29 1 –2 93.

[98] K. A. Scheidt, H. Chen, B. C. Follows, S. R. Chemler, D. S. Coffey,

W. R. Roush, J. Org. Chem. 1998 , 63 , 643 6 –6 437.

[99] B. Zhu, J. S. Panek, Eur. J. Org. Chem. 2001 , 170 1 –1 704.

[100] R. K. Boeckman, P. Shao, J. J. Mullins, Org. Synth. 2004, 10, 696–

698.

[101] The Paterson group did not comment on the reproducibility of this

sequence. We had previously reported our difficulties in accessing

compound 124 by using this route (see ref. [17]).

[102] I. Paterson, K.-S. Yeung, J. B. Smaill, Synlett 1993 , 77 4 –7 76.

[103] K. Suenaga, T. Kimura, T. Kuroda, K. Marsui, S. Miya, S. Kuri-

bayashi, A. Sakagura, H. Kigoshi, Tetrahedron 2006 , 62 , 827 8 –

8290.

[104] a) I. Paterson, C. Watson, K.-S. Yeung, P. A. Wallace, R. A. Ward,

J. Org. Chem. 1997 , 62 , 45 2 –4 53; b) I. Paterson, C. Cowden, C.

Watson, Synlett 1996 , 20 9 –2 11.

[105] The Paterson group reported that, under similar conditions, struc-

turally related macrocyclic halves of rhizopodin that contained

OTBS groups at the C3 and C16 positions, an OTES group at the

C18 position, and an OTIPS group at the C22 position could be de-

protected in useful yields (52%), whereas, for a macrocyclic ana-

logue with the attached side chain with O-TBS groups at the C16

and C16’ positions and O-TES groups at the C3 and C3’ positions,

an elimination product was observed in approximately 20% yield.

The yield of this transformation could only be approximated based

on the experimental data given, because the final product was only

obtained as a mixture with TBAF in a ratio of approximately 1:10.

Because TBAF was not removed during the reaction workup, but

rather was carried through the complete isolation procedure (in-

[68] A. B. Smith, G. K. Friestad, J. Barbosa, E. Bertounesque, K. G.

Hull, M. Iwashima, Y. Qiu, B. A. Salvatore, P. G. Spoors, J. J.-W.

Duan, J. Am. Chem. Soc. 1999 , 121 , 1046 8 –1 0477.

[69] For a kinetic study, see: a) E. L. Eliel, A. H. Hartmann, A. G.

Abatjoglou, J. Am. Chem. Soc. 1974 , 96 , 180 7 –1 816; evidence for

the exclusive formation of a single anionic conformation was ob-

tained from an X-ray structural analysis, see: b) R. Amstutz, D.

Seebach, P. Seiler, B. Schweizer, J. D. Dunitz, Angew. Chem. 1980 ,

92 , 5 9 –6 0; Angew. Chem. Int. Ed. Engl. 1980 , 19 , 5 3 –5 4.

[70] D. Seebach, J. Gabriel, R. Hꢂssig, Helv. Chim. Acta 1984 , 67 , 108 3 –

1099.

[71] J. M. Lehn, G. Wipff, J. Am. Chem. Soc. 1976 , 98 , 749 8 –7 505.

[72] M. Dieckmann, S. Rudolph, C. Lang, W. Ahlbrecht, D. Menche,

Synthesis 2013, 45, 2305–2315.

[73] a) B. M. Kim, S. F. Williams, S. Masamune, in Comprehensive Or-

ganic Synthesis Vol. 2 (Eds.: B. M. Trost, I. Fleming, C. H. Heath-

cook), Pergamon, Oxford, 1991, p.229; b) K.-S. Yeung, I. Paterson,

Chem. Rev. 2005 , 105 , 423 7 –4 313; c) B. Schetter, R. Mahrwald,

Angew. Chem. 2006 , 118 , 766 8 –7 687; Angew. Chem. Int. Ed. 2006 ,

45 , 750 6 –7 525.

[74] M. Dieckmann, D. Menche, Org. Lett. 2013 , 15 , 22 8 –2 31.

[75] D. A. Evans, K. T. Chapman, E. M. Carreira, J. Am. Chem. Soc.

1988 , 110 , 356 0 –3 578.

[76] D. A. Evans, A. H. Hoveyda, J. Am. Chem. Soc. 1990 , 112 , 644 7 –

6449.

Chem. Eur. J. 2013, 19, 15993 – 16018

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

16017

D. Menche et al.

cluding evaporation of the solvents, etc.), it was not clear when the

elimination has occurred. A higher yield for the final deprotection

step was reported by the Paterson group in the global deprotection

of rhizopodin that contained OTES groups at the C16 and C16’ po-

sitions but OTBS groups at both the C3 and C3’ positions (67%

yield).

during the reaction workup, but rather was carried through the

complete isolation procedure (including evaporation of solvents,

etc.), it was not clear when the elimination has occurred.

[107] This reaction was independently performed five times by two dif-

ferent persons in our group; in all cases, rhizopodin was obtained,

albeit in low yields, thus demonstrating the feasibility of this ap-

proach in principal. A HPLC trace of the crude product is provid-

ed in the Supporting Information. A more efficient protecting-

group strategy was reported by the Paterson group; see ref. [101].

[106] It should be stressed that no products that resulted from elimina-

tion of the C16/C16’-OTBS groups were found, as reported by Pa-

terson and co-workers (see ref. [17]) for a structurally related com-

pound with the complete side chains but containing OTES groups

at the C3 and C3’ positions, neither by NMR spectroscopy nor by

HRMS (also see ref. [101]). Because TBAF was not removed

Received: June 10, 2013

Published online: October 10, 2013

16018

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 15993 – 16018

Article Doi

Modular total synthesis of rhizopodin: A highly potent G-actin dimerizing macrolide

DOI: 10.1002/chem.201302197

Source and publish data:

Authors:

Article abstract of DOI:10.1002/chem.201302197

Full text of DOI:10.1002/chem.201302197

Products guided by the article

R&D Labs maybe for 1568949-40-5

Relevant to this article

Hot Product