Stereoselective total synthesis of etnangien and etnangien methyl ester

pubs.acs.org/joc

Stereoselective Total Synthesis of Etnangien and Etnangien Methyl Ester

Pengfei Li,^†Jun Li,^‡,§Fatih Arikan,^‡,Wiebke Ahlbrecht,^†Michael Dieckmann,^†and

Dirk Menche*^,†,‡

†

€

Institut fu€r Organische Chemie, Ruprecht-Karls-Universitat Heidelberg, Im Neuenheimer Feld 270,

D-69120 Heidelberg, Germany, and ^‡Helmholtz-Zentrum fu€r Infektionsforschung, Medizinische Chemie,

Inhoffenstrasse 7, D-38124 Braunschweig, Germany. ^§Present address: Institut fu€r Pharmazeutische

Wissenschaften, ETH Zu€rich.. Present address: ABX GmbH, Radeberg.

dirk.menche@oci.uni-heidelberg.de

Received February 5, 2010

A highly stereoselective joint total synthesis of the potent polyketide macrolide antibiotics etnangien

and etnangien methyl ester was accomplished by a convergent strategy and proceeds in 23 steps

(longest linear sequence). Notable synthetic features include a sequence of highly stereoselective

substrate-controlled aldol reactions to set the characteristic assembly of methyl- and hydroxyl-

bearing stereogenic centers of the propionate portions, an efficient diastereoselective Heck macro-

cyclization of a deliberately conformationally biased precursor, and a late-stage introduction of the

labile side chain by means of a high-yielding Stille coupling of protective-group-free precursors.

Along the way, an improved, reliable protocol for a Z-selective Stork-Zhao-Wittig olefination of

aldehydes was developed, and an effective protocol for a 1,3-syn reduction of sterically particularly

hindered β-hydroxy ketones was devised. Within the synthetic campaign, a more detailed under-

standing of the intrinsic isomerization pathways of these labile natural products was elaborated. The

expedient and flexible strategy of the etnangiens should be amenable to designed analogues of these

RNA-polymerase inhibitors, thus enabling further exploration of the promising biological potential

of these macrolide antibiotics.

Introduction

cellulosum, strain So ce750 and later from So ce1045, by the

1,2

€

groups of Hofle and Reichenbach. Etnangien together

The polyketide natural product etnangien (1, Figure 1)

was originally isolated from the myxobacterium Sorangium

with the more readily available methyl ester 2³present highly

potent antibiotics, with average IC₅₀values in the submicro-

molar range, both in vitro and in vivo.^1-3They are effective

against a broad panel of Gram-positive bacteria, especially

those belonging to the actinomycetes.¹On a molecular level,

they constitute inhibitors of RNA-polymerase. To date,

the rifamycins are the sole class of clinically used RNA-

polymerase inhibitors which qualifies RNA polymerase as

one of the rare validated, but underexploited, targets for

€

(1) (a) Hofle, G.; Reichenbach, H.; Irschik, H.; Schummer, D. German Patent

€

DE 196 30 980 A1: 1-7, May 2, 1998. (b) Irschik, H.; Schummer, D.; Hofle, G.;

Reichenbach, H.; Steinmetz, H.; Jansen, R. J. Nat. Prod. 2007, 70, 1060.

(2) For a recent review on polyketides from myxobacteria, see: Menche,

D. Nat. Prod. Rep. 2008, 25, 905.

(3) Menche, D.; Arikan, F.; Perlova, O.; Horstmann, N.; Ahlbrecht, W.;

€

Wenzel, S. C.; Jansen, R.; Irschik, H.; Muller, R. J. Am. Chem. Soc. 2008,

130, 14234.

DOI: 10.1021/jo100201f

r

Published on Web 03/24/2010

J. Org. Chem. 2010, 75, 2429–2444 2429

2010 American Chemical Society

Li et al.

JOCFeatured Article

SCHEME 1

FIGURE 1. Etnangiens: potent macrolide antibiotics of myxobac-

terial origin.

broad-spectrum antibacterial therapy.⁴However, resistance

to the rifamycins has been increasingly spreading, which

renders the development of structurally novel inhibitors an

important research goal. Importantly, the etnangiens appear

to exhibit no cross-resistance to rifampicin.^1,3Furthermore,

they also retain activity against retroviral DNA polymerase

and show only low cytotoxicity against mammalian cell

cultures, which adds with their attractiveness for further

development. However, preclinical advancement is severely

hampered by the notorious instability of etnangien (1), which

has been associated to the polyene side chain.^1,3

The unique architecture of the etnangiens is characterized

by a 22-membered macrolactone ring with two alkenes and

a polyunsaturated side chain with seven E-configured double

bonds. In total, they contain 12 methyl- and hydroxyl-

bearing stereogenic centers. Although originally reported

as a planar structure, the absolute and relative stereochem-

istry has been determined in our group in cooperation with

development and execution of our synthetic strategy, which

culminated in the first and joint total synthesis of etnangien

(1) and its methyl ester (2).

Results and Discussion

In a rationale to confirm the stereochemistry of the natural

product as well as to enable access to simplified analogues

lacking the labile polyene side chain for SAR studies, we

initiated our synthetic campaign with an effort to first target

the macrocyclic core 3 of etnangien. As outlined in Scheme 1,

the fully protected precursor 4 may be disconnected at the

lactone bond and the (31,32)-diene moiety, revealing two

fragments 5 and 6 of similar complexity. For cyclization,

various cross-coupling strategies to set the characteristic

(Z,E)-diene may be envisioned as an alternative to a more

conventional⁷macrolactonization strategy. While a terminal

Z-vinyl iodide was planned for the southern fragment 6, a

terminal alkyne on the northern part 5 should serve as a

flexible precursor to vinylstannanes, boronates, or alkenes,

allowing for a very high degree of flexibility in our synthetic

plan. In order to mimic the solution structure of the natural

product,³which is characterized by a hydrogen bond be-

tween OH-36 and OH-38, an acetonide protection group was

chosen for these two hydroxyls in a rationale to potentially

facilitate macrocyclization. Acid 5 may be further discon-

nected by a 1,4-syn-aldol reaction to ethyl ketone 8 and

Roche ester derived aldehyde 7, while the southern fragment

6 may likewise be derived by an aldol disconnection here

from methyl ketone 10 and aldehyde 9. Introduction of the

presumably labile vinyl iodide would then be effected by

means of a Stork-Zhao-Wittig olefination.

€

the group of Rolf Muller by a combination of J-based

configurational analysis with molecular modeling, chemical

derivatization, and an innovative method based on gene-

cluster analysis.³First SAR data suggest that a free acid is

not essential for biological activity and that it is possible to

modify and stabilize the etnangien structure yet still retain

activity. Truncation of the side chain, however, leads to

significant loss of activity, which suggests this region to be

part of the pharmacophore region.⁵

The important biological properties of etnangien together

with its natural scarcity and unique and challenging structure

render it an attractive synthetic target, to support further

biological evaluation, as well as to enable more elaborate

structure-activity relationship (SAR) studies. Recently, the

first total synthesis of etnangien was accomplished in our

group,⁶which also unambiguously confirmed our stereo-

chemical assignment. Herein, we provide a full account of the

As shown in Scheme 2, two main strategies for construc-

tion of the polypropionate C32-C42 fragment were evalu-

ated.⁸Both routes were based on a boron-mediated aldol

(4) (a) Darst, S. Trends Biochem. Sci. 2004, 29, 159. (b) Chopra, I. Curr.

Opin. Invest. Drugs 2007, 8, 600. (c) Haebich, D.; von Nussbaum, F. Angew.

Chem., Int. Ed. 2009, 48, 3397.

(5) Menche, D.; Li, P.; Irschik, H. Bioorg. Med. Chem. Lett. 2010, 20, 939.

(6) Li, P.; Li, J.; Arikan, F.; Ahlbrecht, W.; Dieckmann, M.; Menche, D.

J. Am. Chem. Soc. 2009, 131, 11678.

(7) Parenty, A.; Moreau, X.; Campagne, J. M. Chem. Rev. 2006, 106, 911.

(8) For a review on polypropionate syntheses, see: Li, J.; Menche, D.

Synthesis 2009, 2293.

2430 J. Org. Chem. Vol. 75, No. 8, 2010

Li et al.

JOCFeatured Article

SCHEME 2

Unexpected difficulties were then encountered in the

hydroxyl-directed 1,3-syn reduction of 20. As shown in

Scheme 2, a wide variety of commonly used reagents failed

almost completely, resulting in either very low conversion

and/or low degrees of selectivity, suggesting a very high

degree of steric hindrance of propionate 20. Initially, useful

levels of 1,3-syn induction could only be obtained with

Zn(BH₄)₂as reducing agent. However, yields remained

moderate and variable, particularly on large scale. After

considerable optimization, a two-step procedure was deve-

loped, which first involved chelation of β-hydroxy ketone

with (c-Hex)₂BCl (1.5 equiv) and NEt₃(1.5 equiv) at -30 °C

in Et₂O and subsequent reduction with LiBH₄, (1 equiv) at

-78 °C, resulting in the formation of 21 with high yield and

selectivity, allowing for a large-scale access to this fragment.

As shown in Scheme 3, the diol 21 was then protected

as the acetonide 22. At this stage, the relative configuration

of the stereohexad was rigorously confirmed by NMR

methods.¹²Selective cleavage of the primary TBS-ether with

NaIO₄using a procedure recently developed in our group¹³

€

and attachment of the tosylate by a method of Hunig and

Hartung¹⁴proceeded in good yields. For homologation to

alkyne 24 by substitution of tosylate 23, the use of sodium

acetylide proved beneficial as compared to the more con-

ventional lithium salt, giving the desired alkyne 24 in useful

yields (72%). The carboxylic acid was then introduced by

deprotection of the PMB group with DDQ and two-step

oxidation (three steps, 71%). In total, the sequence to 5

proceeded in 12 steps from 11 and 12% yield and allowed

access to multigram quantities of the key northern building

block.

For further functionalization of alkyne 5, we first targeted

E-vinylstannane 26. Based on unfavorable precedents of

direct E-selective hydrostannylation of terminal alkynes,¹⁵

a two-step method was applied,¹⁶which first involves

generation of acetylenic bromide 25 by a silver nitrate

mediated bromination, which proceeded in good yield. The

derived bromoalkyne was then directly hydrostannylated

with Pd₂(dba)₃/PPh₃/Bu₃SnH according to the published

procedure.¹⁶Although E-selection was almost complete (dr

>19:1), the yield remained low (30-35%) and not repro-

ducible. The destannylated analogue was obtained as the

main side product, possibly due to an unfavorable influence

of the carboxylic acid group. Furthermore, model Stille

coupling reactions of stannane 26 with simple vinyl iodides

were not successful under various catalytic conditions.

Consequently, we turned our attention to alternatives.

First, Lindlar hydrogenation of 5 was applied to give the

terminal alkene 27. The reaction proceeded in good yields,

based on careful monitoring of the conversion, to avoid

overreduction to the alkane. Alternatively, hydroboration of

the terminal alkyne was studied. Since in situ hydroboration

with catechol borane and immediate Suzuki macrocycliza-

tion failed to afford the macrocyclic core (Scheme 6, vide

infra), a potentially more robust and more easily handable

reaction of lactate-derived methyl ketone 11⁹with aldehydes

12 and 13, which proceeded with high stereoselectivity and

yield to give the anti-propionates 14 and 16, respectively. The

derived TBS ether 15 was then further elaborated to ethyl

ketone 8 by cleavage of the benzoyloxy group with samarium

diiodide, while 17 was transformed to aldehyde 18 by a two-

step procedure involving reductive removal of the benzoate

and subsequent diol cleavage. Best results in the pivotal 1,4-

syn aldol coupling of ketone 8 with aldehyde 7 were obtained

with a tin(II) triflate-mediated coupling following a slightly

modified procedure of Paterson¹⁰to afford the desired

β-hydroxy ketone 20 in good stereoselectivity (dr 7:1) and

yield. This procedure proved to be more robust and reliable

also on large scale as compared to a likewise evaluated

titanium-mediated aldol coupling.¹¹The likewise tested

alternative on an inverse aldol coupling of ethyl ketone 19

with aldehyde 18 proceeded with much lower yields and

selectivities. Accordingly, this route was no longer pursued.

(12) (a) Rychnovsky, S. D.; Skalitzky, D. J. Tetrahedron Lett. 1990, 31,

945. (b) Evans, D. A.; Dale, L. R.; Gage, J. R. Tetrahedron Lett. 1990, 31,

7099.

(13) Li, J.; Menche, D. Synthesis 2009, 1904.

€

(14) Hartung, J.; Hunig, S.; Kneuer, R.; Schwarz, M.; Wenner, H.

(9) Paterson, I.; Wallace, D. J.; Cowden, C. J. Synthesis 1998, 639.

(10) Paterson, I.; Tillyer, R. D. Tetrahedron Lett. 1992, 33, 4233.

(11) Arikan, F.; Li, J.; Menche, D. Org. Lett. 2008, 3521.

Synthesis 1997, 1433.

(15) Trost, B. M.; Ball, Z. T. Synthesis 2005, 853.

(16) Zhang, H. X.; Guibe, F.; Balavoine, G. J. Org. Chem. 1990, 55, 1857.

J. Org. Chem. Vol. 75, No. 8, 2010 2431

Li et al.

JOCFeatured Article

SCHEME 3

pinacol borane was used to target the presumably more

stable¹⁷boronate 29. First, the direct hydroboration of

alkyne 5 was studied. Interestingly, to the best of our knowl-

edge there appears to be no precedence of such a hydrobora-

tion with pinacol borane in the presence of a free carboxylic

acid group. Recently, Shirakawa et al. have developed an

efficient hydroboration method with pinacol borane in the

presence of catalytic amounts of dicyclohexylborane.¹⁸Pre-

sumably, the reaction involves initial hydroboration by

dicyclohexylborane, followed by an alkenyl B-to-B transfer

to pinacolborane. In our case, a slight modification of the

original procedure was applied, due to the presence of a free

acid group leading to rapid protonation of dicyclohexylbor-

ane. Accordingly, the substrate was first treated with 3.0

equiv of pinacol borane for 15 min at 0 °C to consume the

acidic proton of the terminal carboxylic acid, followed by

addition of dicyclohexylborane. However, only very low

degrees of conversion could be observed for substrate 5

under conventional conditions, and extended reaction times

or elevated temperatures led to complex mixtures. Conse-

quently, introduction of the boronate prior to the acid group

formation was studied. Hydroboration of alkyne 24 with

pinacolborane and removal of the PMB group with DDQ

proceeded in good yields. Despite initial concerns of the

known lability of boron compounds to oxidations,¹⁹oxida-

tion of the derived primary alcohol with Dess-Martin

periodinane proceeded smoothly to the corresponding alde-

hyde in good, albeit to a certain degree variable yields

(71-84%). The final oxidation to the desired acid 29, how-

ever, proved very capricious. Best results were obtained with

the Pinnick protocol,²⁰giving the acid 29 in low and variable

yields (20-47%), presumably due to oxidative decomposi-

tion pathways of the boronate.²¹

Accordingly, a more efficient approach was required,

which was finally devised by a cross metathesis²²of alkene

27 with commercially available vinyl boronate 28. After

considerable optimization, it was found that preparatively

useful yields (64%) may be obtained, with the first-genera-

tion catalyst of Grubbs, involving high catalyst loading

(30 mol %) and a large excess of 28 (20 equiv).

As in Scheme 4, our synthetic approach for the southern

fragment 6 was based on an asymmetric allylation of 1,5-

pentanediol-derived aldehyde 30,^23-25which was initially

studied using Brown’s methodology.²⁶However, even under

low reaction temperatures (-100 °C), the enantioselectivity,

as determined by advanced Mosher esters analysis,²⁷re-

mained low (<80% ee). In conjunction with the tedious

isolation procedure due to coelution of the reagent-derived

byproducts during column chromatography, we turned our

attention to Leighton’s reagent 31.^28,29Gratifyingly, a solu-

tion of aldehyde 30 and (S,S)-31 afforded at -10 °C the

homoallylic alcohol 32 in high enantioselectivity (92% ee),

albeit with moderate yields (62%), presumably due to gra-

dual decomposition of the material under the Lewis acidic

conditions over the required prolonged reaction times (48 h).

Isolation of the allylation product 32 was straightforward by

(22) Morrill, C.; Grubbs, R. H. J. Org. Chem. 2003, 68, 6031.

(23) Marshall, J. A.; Shearer, B. G.; Crooks, S. L. J. Org. Chem. 1987, 52,

1236.

(24) Kozikowski, A. P.; Shum, P. W.; Basu, A.; Lazo, J. S. J. Med. Chem.

1991, 34, 2420.

(25) The corresponding homoallylic alcohol has previously been pre-

pared in racemic form before: (a) Aiguade, J.; Hao, J.; Forsyth, C. J. Org.

Lett. 2001, 3, 979. (b) Bonini, C.; Campaniello, M.; Chiummiento, L.;

Videtta, V. Tetrahedron 2008, 64, 8766.

(17) Roy, C. D.; Brown, H. C. Monatsh. Chem. 2007, 138, 879.

(18) Shirakawa, K.; Arase, A.; Hoshi, M. Synthesis 2004, 1814.

(19) Kabalka, G. W.; Hedgecock, H. C. J. Org. Chem. 1975, 40, 1776 and

references cited therein.

(20) Bal, B. S.; Childers, W. E.; Pinnick, H. W. Tetrahedron 1981, 37,

2091.

(26) Racherla, U. S.; Brown, H. C. J. Org. Lett. 1991, 56, 401.

(27) (a) Dale, J. A.; Dull, D. L.; Mosher, H. S. J. Org. Chem. 1969, 34,

2543. (b) Ohtani, I.; Kusumi, T.; Kashman, Y.; Kakisawa, H. J. Am. Chem.

Soc. 1991, 113, 4092. (c) Seco, J. M.; Quinoa, E.; Riguera, R. Chem. Rev.

2004, 104, 17.

(28) Kubota, K.; Leighton, J. L. Angew. Chem., Int. Ed. 2003, 42, 946.

(29) Zhang, X.; Houk, K. N.; Leighton, J. L. Angew. Chem., Int. Ed. 2005,

44, 938.

(21) NMR analysis of the crude product suggested the presence of formyl

protons.

2432 J. Org. Chem. Vol. 75, No. 8, 2010

Li et al.

JOCFeatured Article

SCHEME 4

TBS deprotection could be done in one-pot fashion; i.e., after

disappearance of the starting material, as monitored by TLC

analysis, 40% aqueous HF was directly added dropwise to

the reaction mixture. The Dess-Martin oxidation of alcohol

34 proceeded in a straightforward manner in the presence of

sodium bicarbonate, giving aldehyde 9 essentially in quanti-

tative yield after flash chromatography.

For the pivotal aldol coupling, we followed a slightly

modified procedure of Paterson,³²which involved conversion

of the methyl ketone into the corresponding enol borinate

with (þ)-Ipc₂BCl and triethylamine (TEA). In situ reaction

with aldehyde 9 afforded the desired 1,4-syn aldol product 35

in 75% yield and high stereoselectivity (dr >19:1).³³

For construction of the C-22 stereocenter by means of a 1,3-

anti reduction of β-hydroxy ketone 35, the Evans protocol³⁴

using tetramethylammonium triacetoxyborohydride [Me₄NBH-

(OAc)₃] in acetonitrile and anhydrous acetic acid (1:1 v/v) was

efficiently applied, proceeding with excellent diastereoselectivity

(dr >19:1) and yield (99%) at -40 °C (20 h).

For further advancement toward the etnangien macro-

cycle, a differentiation of the two hydroxyl groups of diol 36

was required. We anticipated that the steric environments

might be sufficiently different to allow for selective protec-

tion of the less hindered 24-OH. However, treatment with 1.2

equiv of TBSOTf at -78 °C for 1 h gave almost exclusively

the bis-protected product (64% isolated yield), suggesting

little steric bias toward selective monoprotection. No pro-

duct formation was observed when substrate 36 was pre-

36

treated with organotin reagents,³⁵i.e., Bu₂SnCl₂or Bu₂-

SnO³⁷(entries 2 and 3), presumably because both complexed

hydroxyl groups were not reactive enough. In both cases,

only decomposition was observed. Finally, use of TBSCl/

imidazole/DMAP in DMF proved relatively effective (entry

4). At 0 °C to room temperature a certain degree of regiose-

lection was observed, giving the desired TBS ether 37 as the

main product (47%), together with 32% reisolated starting

material and 10% bis-protected diol. The diol could easily be

recycled. At this stage the configuration of the stereocenter at

C-22 was confirmed by Mosher ester analysis, which also

confirmed the regioselective protection at C-24.

As shown in Scheme 5, protection of the remaining free

hydroxyl at C-22 proceeded smoothly with triethylsilyl tri-

flate in the presence of 2,6-lutidine. Subsequent treatment of

the derived TES ether³⁸with potassium carbonate in metha-

nol removed the benzoyl group cleanly. The good but not

chromatography and the reaction proved well-scalable, without

deterioration of efficiency. The corresponding methyl ether 33

was then prepared with NaH/MeI. Initial attempts to directly

homologate the corresponding C24-aldehyde of alkene 33,

readily obtained by deprotection of the primary TBS group

(HF/CH₃CN) and subsequent oxidation using Swern conditions

by aldol coupling³⁰with methyl ketone 10,³¹resulted, however,

in only low yields (<10%), presumably due to practical difficul-

ties associated with the high volatility of the aldehyde.

Accordingly, the corresponding Bz-protected aldehyde 9

was prepared in good yields (81%) from 32 by methylation,

ozonolysis with reductive workup, Bz-protection, subse-

quent TBS deprotection, and DMP oxidation of intermedi-

ate alcohol 34. Notably, it was found that Bz protection and

(32) Paterson, I.; Coster, M. J.; Chen, D. Y. K.; Oballa, R. M.; Wallace,

D. J.; Norcross, R. D. Org. Biomol. Chem. 2005, 3, 2399.

(33) For practical purposes, it proved beneficial to rely on a stock solution

of (þ)-Ipc₂BCl in pentane. Prior to the reaction, the solvent was completely

removed and the reagent dried for more than 2 h under high vacuum to

remove any trace of HCl. The absolute configuration of the newly generated

stereocenter and the degree of stereoselectivity were confirmed by Mosher

ester analysis

(34) (a) Evans, D. A.; Chapman, K. T. Tetrahedron Lett. 1986, 27, 5939.

(b) Evans, D. A.; Chapman, K. T.; Carriera, E. M. J. Am. Chem. Soc. 1988,

110, 3560.

(35) For tin-based regioselective protections of hydroxyl groups in car-

bohydrate chemistry, see: David, S.; Hanessian, S. Tetrahedron 1985, 41, 643.

(36) Pitsch, S.; Weiss, P. A.; Jenny, L. US Patent 5,986,084, 1999.

(37) (a) Leigh, D. A.; Martin, R. P.; Smart, J. P.; Truscello, A. M.

J. Chem. Soc., Chem. Commun. 1994, 1373. (b) Reginato, G.; Ricci, A.;

Roelens, S.; Scapecchi, S. J. Org. Chem. 1990, 55, 5132.

(38) The TES ether was so unpolar that removal of the byproduct

Et₃SiOH by flash chromatography was sometimes not complete. However,

after Bz deprotection, alcohol 39 was obtained in pure form after flash

chromatography.

(30) Paterson, I.; Goodman, J. M.; Lister, M. A.; Schumann, R. C.;

McClure, C. K.; Norcross, R. D. Tetrahedron 1990, 46, 4663.

(31) Ketone 10 was prepared by a slightly modified procedure of the

literature: Smith, A. B., III; Doughty, V. A.; Sfouggatakis, C.; Bennett, C. S.;

Koyanagi, J.; Takeuchi, M. Org. Lett. 2002, 4, 783. In our case, PMB-

protection of (R)-(-)-^D-Roche’s ester with p-methoxybenzyl imidate was

best done in the presence of a catalytic amount of 10-camphorsulphonic acid

in dichloromethane. Notably, the reaction scale seems important for the

outcome. While small-scale (86 μmol) reactions were slow (24 h for comple-

tion, ∼60% yield), large-scale (4.86 mmol) reactions proved fairly fast (<4 h,

95% yield).

J. Org. Chem. Vol. 75, No. 8, 2010 2433

Li et al.

JOCFeatured Article

SCHEME 5

SCHEME 6

excellent overall yield (77%, two steps) may be attributed to

partial removal of the labile TES group in the Bz deprotec-

tion step. The primary alcohol 39 was then oxidized by

Dess-Martin periodinane in the presence of sodium bicar-

bonate to aldehyde 40 uneventfully.

For introduction of the required Z-vinyl iodide, a Stork-

Zhao-Wittig olefination³⁹was envisioned. The required

phosphonium reagent [Ph₃PCH₂I]^þI^-was prepared by a

modified procedure,⁴⁰which involved heating a solution of

triphenylphosphine and methylene iodide in dry toluene to

50 °C and stirring at this temperature for 96 h in the dark.

After the solution was washed with diethyl ether and dried

under vacuum, the desired phosphonium salt was obtained

1

in 56% yield. Importantly, following these conditions, H

NMR and ¹³C NMR analysis indicated the presence of a

single compound and not mixtures as claimed in the litera-

ture.³⁹The olefination step was then conducted under “salt-

free” conditions.⁴¹Accordingly, the required phosphorylide

was first prepared by deprotonation of the phosphonium salt

with KHMDS. After standing for 30 min, the yellow super-

natant solution of the “salt-free” ylide was cannulated into

the other flask containing the aldehyde. The reaction was

complete within 30 min at -78 °C and produced the desired

vinyl iodide 41 in excellent selectivity (Z/E = 50:1) but,

however, only moderate yield (52%). After considerable

experimentation, an improved protocol was applied (to

compare, see Scheme 8, vide infra), giving the desired iodide

in 83% yield and Z/E = 19:1 selectivity. Finally, selective

deprotection of the TES protecting group in the presence of a

less hindered but more bulky TBS group was best achieved

by using a catalytic amount of PPTS (0.25 equiv) in ethanol

giving alcohol 6 in85% yield. Intotal, the southernfragment6

was prepared in 7.4% overall yield (9.8% with recycling once

diol 36 during the regioselective protection step) and 12 steps

from 12 in the longest linear sequence from aldehyde 30.

As shown in Scheme 6, two main strategies were then

pursued for advancing to the macrocyclic core of etnangien.

Initial attempts relying on a Stille coupling of 26 resulted in

only low and not reproducible yields on model compounds,

giving mainly destannylated product, possibly due to the

presence of the free carboxylate. Attempts were then directed

toward an esterification, hydroboration, and Suzuki macro-

cyclization approach. Thus, under standard Yamaguchi

esterification conditions (2,4,6-trichlorobenzoyl chloride,

TEA, and DMAP in toluene), ester 42 was prepared from

5 and 6 in 73% yield. Subsequently, an in situ procedure for

hydroboration and Suzuki macrocyclization without inter-

mediate isolation of the boronate was evaluated.⁴²Accord-

ingly, after complete hydroboration of alkyne 42 with

catechol borane (10.0 equiv) in the presence of a catalytic

amount of dicyclohexylborane (0.2 equiv), the crude product

was directly submitted to catalytic amounts of Pd(PPh₃)₄

and a slight excess of TlOEt. However, only complex product

mixtures could be obtained. The successful strategy finally

involved an intermolecular esterification between alkene 27

and alcohol 6, followed by a Heck macrocyclization. Again,

the Yamaguchi protocol was effectively employed, giving

ester 43 in 81% yield. For complete conversion and ease of

(39) Stork, G.; Zhao, K. Tetrahedron Lett. 1989, 30, 2173.

(40) Seyferth, D.; Heeren, J. K.; Singh, G.; Grim, S. O.; Hughes, W. B.

J. Organomet. Chem. 1966, 5, 267.

(41) Davison, E. C.; Fox, M. E.; Holmes, A. B.; Roughley, S. D.; Smith,

C. J.; Williams, G. M.; Davies, J. E.; Raithby, P. R.; Adams, J. P.; Forbes,

I. T.; Press, N. J.; Thompson, M. J. J. Chem. Soc., Perkin Trans. 1 2002, 1494.

(42) Nicolaou, K. C.; Nold, A. L.; Milburn, R. R.; Schindler, C. S.; Cole,

K. P.; Yamaguchi, J. J. Am. Chem. Soc. 2007, 129, 1760.

2434 J. Org. Chem. Vol. 75, No. 8, 2010

Li et al.

JOCFeatured Article

SCHEME 7

relative stereochemistry of the macrocyclic core of the nat-

ural product.

With these findings in hand, a synthetic strategy toward

the authentic natural product was devised. As shown in

Scheme 7, critical disconnections in the macrocyclic part

were chosen in an analogous fashion to the model system,

i.e., an ester formation and a cross coupling to forge the diene

system. However, in contrast to the model system, the side

chain was dissected in the middle of the conjugate hexaene

subunit in a rationale to set the labile hexaene system at a

late stage of the synthesis. Accordingly, a base-free Stille

coupling⁴⁵was envisioned for a late-stage attachment of the

side chain, revealing the trienyl stannane building block 47

(C1-C13 fragment). The macrocycle would be disconnected

into the same northern fragment 27 (C32-C42) and a

homologated southern fragment 48 (C15-C31) in a ratio-

nale to increase the overall convergence. Furthermore, it was

hoped that the more elaborate macrocycle 46 might be less

prone to isomerization processes, which have not been

reported for the authentic natural products.¹Compound

48 was planned to arise from an aldol connection along the

C23-C24 bond. Notably, a 3,4-dimethoxybenzyl (DMB)

protection group was chosen for the 20-OH to allow for

useful levels of a favorable 1,5-anti induction.⁴⁶On the basis

of the observed stability of vinyl iodide 6, introduction of the

vinyl iodide portion was planned before the aldol coupling to

render the strategy more convergent, which revealed the

coupling partners 49 and 50.

As shown in Scheme 8, the synthesis of vinyl iodide

fragment 49 involved the same intermediate 33 as used before

(see Scheme 4). However, in contrast to the route above, a

revised synthetic plan based on a Jacobsen hydrolytic kinetic

resolution (HKR) of epoxide 52⁴⁷was implemented, due to

the high expense of the chiral reagent 31 and the require-

ments for even larger amounts of 33. Accordingly, the chiral

epoxide was obtained in two steps from readily available

TBS ether 51 following a published procedure,^47awhich

proved well-scalable. Regioselective ring-opening, methyla-

tion, and ozonolysis of 33 gave aldehyde 53 in essentially

quantitative yield. Stork-Zhao-Wittig olefination was first

tested under the “salt-free” conditions, as evaluated above

(see Scheme 5).³²However, in contrast to the observations

with substrate 40, here a mixture of the expected Z-vinyl

iodide 55 (25% yield) together with the geminal diiodide 54

was obtained which could only be separated by normal-

phase preparative HPLC.

separation, it proved beneficial to use a slight excess of acid

27 (1.2 equiv), which was easily recovered after the reaction.

Subsequently, the Heck macrocyclization reaction could be

effected under slightly modified Jeffery conditions,⁴³invol-

ving a degassed mixture (3.9 mM) of the substrate, tetra-

butylammonium chloride (3.0 equiv), anhydrous potassium

carbonate (8.0 equiv), and palladium acetate (1.2 equiv) in

DMF at 60 °C, giving the desired macrolactone in 47% yield,

with complete E-selectivity. Presumably, this remarkable

diastereoselectivity⁴⁴may be caused by a favored conforma-

tion controlled by the rigid acetonide protection group. For

further elaboration of 4, removal of the protecting groups

was studied. In agreement with more elaborate experiments

(vide infra, Scheme 17), the TBS groups were completely

removed by a freshly prepared DMF solution of TBAF (2.0 M)

and acetic acid (0.2 M) at 0 °C. The expected product 44 could

be isolated by preparative normal-phase HPLC in 35% to-

gether with the contracted analogue 45 in 53% yield.

In summary, the synthesis of 44 was accomplished in 16

steps from ketone 11 with 1.4% overall yield. A sequence of

highly stereoselective aldol reactions was developed to set the

characteristic sequence of hydroxyl- and methyl-bearing

stereogenic centers with high degrees of efficiency and con-

vergence while a Heck macrocyclization was effectively used

for closing the macrocyclic core. Importantly, at this stage,

the close similarity of the NMR data of 44 and the corres-

ponding etnangien acetonide³allowed us to confirm the

A similar observation has only been reported by the group

of Bestmann.⁴⁸To explain the generation of the unwanted

diiodide, they suggested an iodine-hydrogen exchange pro-

cess, as shown in Scheme 9. According to their model, the

normal ylide 56 and the undeprotonated salt 57 might

exchange a hydrogen and iodide, leading to an iodine-free

(45) Paterson, I.; Findlay, A. D.; Noti, C. Chem. Asian J. 2009, 41, 594.

(46) (a) Paton, R. S.; Goodman, J. M. Org. Lett. 2006, 8, 4299.

(b) Paterson, I.; Gibson, K. R.; Oballa, R. M. Tetrahedron Lett. 1996, 37,

^ ꢀ

8585. (c) Evans, D. A.; Coleman, P. J.; Cote, B. J. Org. Chem. 1997, 62, 788.

^ ꢀ

(d) Evans, D. A.; Cote, B.; Coleman, P. J.; Connell, B. T. J. Am. Chem. Soc.

2003, 125, 10893.

(47) (a) Myers, A. G.; Lanman, B. A. J. Am. Chem. Soc. 2002, 124, 12969.

(43) Jeffery, T. Tetrahedron 1996, 52, 10113.

(b) Tokunaga, M.; Larrow, J. F.; Kakiuchi, F.; Jacobsen, E. N. Science 1997,

277, 936.

(48) Bestmann, H. J.; Rippel, H. C.; Dostalek, R. Tetrahedron Lett. 1989,

30, 5261.

(44) (a) Menche, D.; Hassfeld, J.; Li, J.; Rudolph, S. J. Am. Chem. Soc.

2007, 129, 6100. (b) Menche, D.; Hassfeld, J.; Li, J.; Mayer, K.; Rudolph, S.

J. Org. Chem. 2009, 74, 7220.

J. Org. Chem. Vol. 75, No. 8, 2010 2435

Li et al.

JOCFeatured Article

SCHEME 8

SCHEME 9

to the suspension of the salt and the deprotonation processes

is much faster than the I-H exchange process. In our case,

however, we could not add large amount of the base in a few

seconds, due to reasons of reproducibility and scalability,

whereas prolonging exposal of the salt 57 to ylide 56 might

lead to unfavorable I-H exchange processes.

ylide 58 and a new salt (59) with two covalently attached

geminal iodine atoms (i.e., H to I exchange). The resulting

two species may then further undergo proton transfer, lead-

ing to phosphor ylide 62, resulting in the formation of the

geminal diiodide 64.

Based on these thoughts, we adopted the following proce-

dure for our preparation of the Z-vinyl iodide. In the absence

of light, a suspension of a fine powder of the phosphonium

iodide (1.5 equiv) in anhydrous HMPA was slowly added to a

solution of NaHMDS (1.5 equiv) in dry THF at 10-15 °C.

After 1 min, the mixture was cooled to -78 °C and a solution

of the aldehyde (1.0 equiv) in THF was added dropwise. After

20 min at -78 °C, the reaction was stopped and a usual

aqueous workup procedure produced the Z-vinyl iodide.

Following this procedure, the Z-vinyl iodidewas reproducibly

obtained in good yield (70-77%) and high Z/E selectivity

(15:1 to 27:1), without any observable formation of the vinyl

diiodide (see the table in Scheme 8). Notably, when the

insoluble salt was added in solid form batchwise and under

otherwise identical conditions, a small amount of the diiodide

was still formed (Z-55/E-55/diiodide 54 = 10:1.0:1.3), pre-

sumably as the adding rate could not be rigorously controlled

under these conditions.

Subsequently, the primary TBS protecting group was

removed with a catalytic amount of 10-camphorsulfonic

acid (0.3 equiv) in methanol, giving the derived alcohol in

essentially quantitative yields. The subsequent oxidation was

most effectively performed with the Parikh-Doering proce-

dure.⁵¹

As shown in Scheme 10, construction of the desired methyl

ketone 50 was initiated by a previously described Wittig

reaction of commercial aldehyde 65 with ylide 66.⁵²In our

hands, however, the reported procedure in toluene resulted

in only moderate degrees of conversion and yields. In con-

trast, it was observed that the reaction proceeds much faster

in DCM, resulting in higher degrees of conversion and yields.

Optimum results were obtained with 1.0 equiv of the Wittig

reagent and 3.0 equiv of the aldehyde at room temperature,

In order to understand the side reaction and ultimately

improve the reaction outcome, a possible mechanism of this

exchange process may be proposed as shown in Scheme 9

(right part). Accordingly, the salt 57 might be attacked by

ylide 56 in a nucleophilic-substitution fashion to give a

hetereodimerized salt (60) as an intermediate. A second

substitution, now by an iodide, might then lead to C-C

bond cleavage forming the neutral ylide 58.⁴⁹Possibly, such

an I-H exchange process might be useful for the synthesis of

1,1-diiodoalkenes in a HWE variant.⁵⁰However, to the best

of our knowledge, a detailed study and mechanistic explana-

tion for this process appear not to have been reported.

Based on this proposal, the I-H exchange is initiated by

two species, 56 and 57, which are present in the deprotona-

tion stage. We hypothesized that if there is little undeproto-

nated salt 57 present, formation of the diiodide side product

should be avoided. As the deprotonation process with strong

bases such as NaHMDS should be a rapid process, slow

addition of the salt to a solution of the base would be critical

to ensure the all-over excessive presence of NaHMDS.

On the other hand, excess of base might be detrimental to

the following reaction with the aldehyde 63. Accordingly,

exactly a 1:1 molar ratio of the phosphonium salt and the

base was considered to be ideal.

Conventionally, the phosphonium salt is added to the

reaction flask before addition of the base, presumably due

to the insolubility of the salt in most solvents and due to ease

of conduction. Surprisingly, however, in many cases, I-H

exchanges appear not to have been observed. One possible

explanation could be that normally the base is rapidly added

(49) An analogous Br-H exchange process has also been observed:

(a) Fischer, H.; Fischer, H. Chem. Ber. 1966, 99, 658. (b) Bestmann, H. J.

Chem. Ber. 1962, 95, 58.

(51) Parikh, J. R.; Doering, W. v. E. J. Am. Chem. Soc. 1967, 89, 5505.

(52) Famer, L. J.; Marron, K. S.; Koch, S. S. C.; Hwang, C. K.; Kallel,

E. A.; Zhi, L.; Nadzan, A. M.; Robertson, D. W.; Bennani, Y. L. Bioorg.

Med. Chem. Lett. 2006, 16, 2352.

(50) Bonnet, B.; Gallic, Y. L.; Ple, G.; Duhamel, L. Synthesis 1993, 1071.

2436 J. Org. Chem. Vol. 75, No. 8, 2010

Li et al.

JOCFeatured Article

SCHEME 10

SCHEME 12

SCHEME 11

shown in Scheme 11. Unexpectedly,⁸the subsequent Abiko-

Masamune aldol reaction with propionate 70 proceeded in

only moderate selectivity and yield (76%, dr >5:1). Also,

conversion of aldol adduct 71 to Weinreb amide 74 using

our previously developed method with a slight excess of

i-PrMgCl and O,N-dimethylhydroxylamine hydrochloride⁵⁴

was accompanied with variable degrees of elimination and

protection of the newly generated hydroxyl of 71 as a DMB-

ether failed to provide the desired ether 73.⁵⁵

Due to theses unsatisfactory results, we turned our atten-

tion to a Paterson anti-aldol reaction with lactate-derived

ethyl ketone 11,⁹which proceeded with excellent selectivity

and yield (97%, dr >20:1). However, again DMB protection

of the generated hydroxyketone 72 proved capricious. Despite

considerable experimentation, yields remained low. Best re-

sults were obtained with PPTS giving 75 with 37% yield. Also,

subsequent addition of MeLi to the ketone functionality with

concomitant benzoylcleavageprovidedthe desired 1,2-diol 76

also only in low yield (22%). Finally, conversion to the desired

methyl ketone 50 by oxidative cleavage with NaIO₄in THF/

H₂O also proved difficult to control, giving variable degrees of

methyl ketones 50 and 77.⁵⁶

On the basis of these unsatisfactory results, an alternative

protective group strategy had to be implemented. For con-

venience, a TBS group was chosen, despite expected unfa-

vorable effects on the envisioned 1,5-anti induction in the

subsequent aldol coupling with aldehyde 49. As shown in

Scheme 12, TBS protection of aldol product 72 was best

carried out with TBSOTf in the presence of Proton-sponge,

while alternative bases (2,6-lutidine, 2,4,6-collidine, 2,3,5-

collidine, 2,4-di-tert-butyl pyridine) resulted in lower yields

and were accompanied by various degrees of elimination.

giving the desired aldehyde 67 in 80% yield. The homolo-

gated compound 68 was observed as the main side product.⁵³

With 67 in hand, a series of functional group operations,

involving reduction of the aldehyde, TBS protection of

the derived primary alcohol, DIBAL reduction of the ester,

and MnO₂oxidation of the derived allylic alcohol gave

aldehyde 69 in high yield (82%) over these four steps, as

(54) Li, J.; Li, P.; Menche, D. Synlett 2009, 2417.

(55) While TBS protection was possible at this stage (TBSOTf, 2,6-

lutidine, DCM, 98%, see the Supporting Information), conversion of the

derived TBS ether to the Weinreb amide failed, possibly due to steric reasons.

(56) In one instance, 7 equiv of NaIO₄was added in three portions over a

period of 3.5 h, giving the desired methyl ketone 49 in high yields as the main

product (86%), while under similar conditions (6 equiv of NaIO₄, addition in

one portion), the deprotected analogue 77 was obtained in high yields, see the

Supporting Information.

(53) The amount of this side product may be further minimized by

increasing the excess of the starting aldehyde, which may readily be recycled

after the reaction.

J. Org. Chem. Vol. 75, No. 8, 2010 2437

Li et al.

JOCFeatured Article

Presumably, the increased steric hindrance of the base

suppresses these unwanted side reactions. For conversion

of 78 to the desired methyl ketone 82, it proved beneficial to

first remove the benzoate reductively (LiBH₄), followed by

oxidative cleavage of the derived diol 80, which was best

carried out with Pb(OAc)₄in toluene to avoid removal of the

allylic TBS ether, as observed before (see Scheme 11). Alde-

hyde 81 was then converted in two steps and high yields

(94%) to the desired methyl ketone 82 by addition of MeLi

and oxidation of the derived alcohol. This strategy proved to

be superior to a likewise tested alternative of direct addition

of methyl lithium to benzoate 78, leading to extensive

decomposition, presumably by direct nucleophilic attack to

the electron-rich allylic TBS ethers.⁵⁷In summary, ketone 82

was prepared in 11 steps and 47% yield.

SCHEME 13

Gratifyingly, the pivotal coupling of ketone 82 with alde-

hyde 48 proceeded with high stereoselectivity and yield follow-

ing an Ipc-boron-mediated aldol reaction (76%, dr >14:1). As

shown in Scheme 13, optimum conditions involved enolization

of methyl ketone with (þ)-DIPCl and Et₃N in Et₂O at 0 °C,

which was allowed to react with aldehyde 48 at -78 °C for 1 h

and then at -20 °C for 14 h. Subsequent 1,3-anti reduction of

the aldol product was straightforward following the procedure

as established above, giving 84 in high yield (92%) and

stereoselectivity (dr >19:1). In agreement with the results for

the model compound, selective protection of diol 84 was

optimal with TBSCl/imidazole in dichloromethane, giving

the desired monoprotected TBS ether 85 in 77% yield. The

slightly improved result as compared to the previous substrate

might be attributed to the increased steric hindrance around

the C22 hydroxyl group.

As shown in Scheme 14, efforts to advance to the elabo-

rated macrocyclic core 87 of etnangien were initiated by a

Suzuki coupling strategy. Since hydroboration with catechol

borane and immediate Suzuki macrocyclization failed to

afford the macrocyclic core before (compare Scheme 6), we

turned our attention to first implement the cross coupling.

However, coupling of vinyl iodide 85 with 29 proceeded in

only low yield (21%), giving seco-acid 86, which was further

carried on to the macrocycle uneventfully under standard

Yamaguchi esterification conditions in 51% yield.

evaporation after chromatography, mainly by scrambling

of olefin geometry. Consequently, the esterification was con-

ducted prior to Stille homologation. Accordingly, Yamaguchi

coupling of alcohol 85 and acid 27 produced the ester in good

yield (97%). At this stage, Stille coupling with vinyl stannane

90 was straightforward (rt, 40 min, 82%). The corresponding

diene product 89 was then immediately used for RCM reac-

tions in order to prevent decomposition processes due to any

residual palladium species. However, as shown in Table 2,

reaction conditions could not be optimized to allow for an

effective access to the desired macrolactone 87. The main side

product proved to be the contracted 20-membered ring.

Furthermore, a mixture of 89 and the smaller congener could

not be resolved by various chromatographic methods.

In a parallel fashion, Heck macrocyclization of alkene 88,

readily available from 85 and 27 by esterification, was

studied.^60,61Accordingly, ester 88 was subjected to various

Heck conditions to close the 22-membered ring, as shown in

Table 1 of Scheme 14. After some optimizations, it was found

that the reagent combination Pd(OAc)₂/Bu₄NCl/K₂CO₃/

DMF (70 °C, 50 min)⁶²provided the desired macrocycle 87

in high yield (70%) as the only isolable product with exclu-

sively C32-C33 E-configuration.

As an alternative, a ring-closing metathesis (RCM)⁵⁸of

precursor 89 was evaluated. Initially, the required terminal

diene subunit was synthesized by C2-homologation of iodide

85 by Stille coupling with commercially available tributyl

vinylstannane giving diene 91 in high yield (64%). Surpris-

ingly, however,⁵⁹upon storage even at 0 °C, this (Z)-diene

gradually isomerized to a mixture of (E)- and (Z)- isomers,

presumably because of the presence of trace of palladium

species even after column purification. Therefore, diene 91

had to be used immediately for the next transformation.

However, esterification with 27 was unexpectedly inefficient

under Yamaguchi conditions. Possibly, traces of residual

palladium species led again to decomposition during solvent

With the fully protected macrocyclic core in hand, efforts

could then be directed toward construction of the side chain.

As shown in Scheme 15, its preparation starts from PMB-

protected hydroxypentanone 92, which was homologated by

an HWE reaction with trimethyl phosphonoacetate to pro-

vide the R,β-unsaturated ester 93 in preparatively useful

yields (62%) as an E/Z mixture of 3:1, following a procedure,

recently developed in our group on a related substrate.^44b

(60) There are only sporadic known successful examples of Heck macro-

cyclization in natural product synthesis; in particular, the reaction between

two simple double bonds (without electron-withdrawing substituents on

either alkene) appears to be nonexistent. In this context, our synthesis of

etnangien macrocycle utilizing Heck macrocyclization is remarkable with

respect to the efficiency (yield, E/Z selectivity, atom- and step-efficiency); see

also: (a) Yokoyama, H.; Satoh, T.; Furuhata, T.; Miyazawa, M.; Hirai, Y.

Synlett 2006, 2649. (b) Stocks, M. J.; Harrison, R. P.; Teague, S. J.

Tetrahedron Lett. 1995, 36, 6555. (c) Geng, X.; Miller, M. L.; Lin, S.; Ojima,

I. Org. Lett. 2003, 5, 3733.

(57) Ahmed, M.; Atkinson, C. E.; Barrett, A. G. M.; Malagu, K.;

Procopiou, P. A. Org. Lett. 2003, 5, 669.

(58) (a) Grubbs, R. H.; Chang, S. Tetrahedron 1998, 54, 4413. (b) Mulzer,

€

J.; Ohler, E.; Enev, V. S.; Hanbauer, M. Adv. Synth. Catal. 2002, 344, 573.

(c) Nicolaou, K. C.; Bulger, P. G.; Sarlah, D. Angew. Chem., Int. Ed. 2005, 44,

ꢀ

4490. (d) Gradillas, A.; Perez-Castells, J. Angew. Chem., Int. Ed. 2006, 45,

(61) For intramolecular Heck reactions for formation of medium size

rings, see: (a) Dounay, A. B.; Overman, L. E. Chem. Rev. 2003, 103, 2945.

(b) Shibasaki, M.; Boden, C. D. J.; Kojima, A. Tetrahedron 1997, 53, 7371.

(62) Jeffery, T. Tetrahedron 1996, 52, 10113.

6086.

(59) Paterson, I.; Florence, G. J.; Gerlach, K.; Scott, J. P. Angew. Chem.,

Int. Ed. 2000, 39, 377.

2438 J. Org. Chem. Vol. 75, No. 8, 2010

Li et al.

JOCFeatured Article

SCHEME 14

Reduction of this ester to the corresponding alcohol with

DIBAL-H and reoxidation of the allylic alcohol with MnO₂

gave aldehyde 94 in a straightforward fashion. Brown allyla-

tion under “salt-free” conditions at -100 °C afforded the

homoallylic alcohol 95 in 80% yield and good enantioselec-

tivity (90% ee). After protection with TBSOTf, the E/Z

isomers could be separated by ordinary column chromato-

graphy with silica gel. Finally, the E isomer was transformed

to methyl ester 96 over four steps by PMB deprotection,

Dess-Martin oxidation, Pinnick oxidation, and methyl

esterification with diazomethane.

For introduction of the stannane, we initially conceived

an approach via iodide 102 by mono-Stille coupling with

trans-1,2-bis(tri-n-butylstannyl)ethylene 103.⁶³Accord-

ingly, the required iodide 102 was prepared by cross-me-

tathesis, Takai olefination, and TBS deprotection from the

terminal alkene 96. Cross coupling was best performed with

Grubbs II catalyst (5% mol) and an excess of crotonalde-

hyde (10 equiv) at 60 °C in dry toluene, giving the desired

aldehyde 99 in high yields (90%). Subsequently, Takai

olefination^64,65with iodoform (8.8 equiv) and chromium-

(II) chloride (14.0 equiv) in a mixed solvent system

(dioxane/THF = 6:1)⁶⁶gave dienyl iodide 102 in 65% yield

and acceptable selectivity (E/Z = 6:1), which proved to be

highly labile to light (E/Z isomerization) and traces of

iodoform.⁶⁷

The final critical step, the Stille monocoupling with trans-

1,2-bis(trin-butylstannyl)ethylene (103), was first tried under

standard Stille conditions.⁶⁸However, treatment of iodide

102 with PdCl₂(CH₃CN)₂(10% mol) and 10 equiv of trans-

1,2-bis(trin-butylstannyl)ethylene in degassed DMF resulted

in complex mixtures. Also, with Pd(PPh₃)₄as catalyst, for-

mation of complex mixtures was observed, and the deiodi-

nated congener was observed as the main product (40%). We

therefore resorted to an alternative protocol utilizing a one-

pot Sn-Li-Zn transmetalation and subsequent Negishi

coupling.⁶⁹Accordingly, the required reagent 104 was pre-

pared in situ from 103, as previously described,^69aand

directly added to a cooled (0 °C) THF solution of iodide

102 and catalyst Pd(PPh₃)₄. However, only low degrees of

conversions were observed under conventional conditions,

and after extended reaction times formation of various side

products was observed. Moreover, the low and similar

(67) Because of the low polarity of iodide 102, complete removal of the

excess of iodoform or its derivatives was not practical. Column-purified

compound also decomposed upon storage even under argon atmosphere at

-20 °C

(68) Stille, J. K.; Groh, B. L. J. Am. Chem. Soc. 1987, 109, 813.

(69) (a) Pihko, P. M.; Koskinen, A. M. P. Synlett 1999, 1966. (b) Sorg, A.;

(63) trans-1,2-Bis(tri-n-butylstannyl)ethylene was synthesized from com-

mercially available tributylethynylstannane following a reported procedure:

Renaldo, A. F.; Labadie, J. W.; Stille, J. K. Org. Syn. 1989, 67, 86.

(64) Takai, K .; Nitta, K.; Utimoto, K. J. Am. Chem. Soc. 1986, 108, 7408.

(65) Hodgson, D. M. J. Organomet. Chem. 1994, 476, 1.

€

Bruckner, R. Angew. Chem., Int. Ed. 2004, 43, 4523.

(66) Evans, D. A.; Black, W. C. J. Am. Chem. Soc. 1993, 115, 4497.

J. Org. Chem. Vol. 75, No. 8, 2010 2439

Li et al.

JOCFeatured Article

SCHEME 15

SCHEME 16

extremely challenging, resulting in very low yields in all

cases. After a series of further unsuccessful attempts includ-

ing a tin-Takai olefination,⁷⁰we turned our attention to the

introduction of the stannane by a HWE olefination with

stannylated phosphonate 101, a related work of which had

previously been used by the Smith group with limited

success.^71,72As shown in the table in Scheme 15, various

reaction conditions were applied, resulting, however, initi-

ally only in very low yields of the right product which was

also contaminated by a few other side products. It seemed

that sodium as counterion and lower temperatures appeared

to be beneficial (entries 1-4), suggesting the size (K^þ(18-c-6)

> K^þ> Na^þ> Li^þ) or coordinative capability (K^þ(18-c-6)

< K^þ< Naþ < Li^þ) of the counterion to be important for

the reaction outcome. Indeed, deprotonation of phospho-

nate 101 with KHMDS in the presence of 18-crown-6

followed by addition of the aldehyde 99 at -78 °C resulted

in the formation of the expected triene 46 in 38% yield after

preparative HPLC (entry 5). Subsequent removal of the

secondary TBS was cleanly effected by TBAF, giving the

desired alcohol 105 in useful yields (69%).⁷³

To test the projected late-stage generation of the all-(E)

hexaene subunit of etnangien, the Stille coupling of 46 was

subsequently attempted with simplified model vinyl iodide

107, which was prepared from 78 by allylic oxidative depro-

tection with DDQ.⁷⁴Notably, the secondary allylic TBS

ether remained untouched. Takai olefination of aldehyde

106 gave the labile trienyl iodide 107 (56% yield, E/Z = 4:1).

Gratifyingly, the subsequent Stille coupling between iodide

107 and stannane 46 proceeded smoothly using PdCl₂-

(CH₃CN)₂in the DMF system, giving the desired product

in high yield (85%) (Scheme 16).

With the introduction of the side chain established, efforts

were then directed to access the macrocyclic vinyl iodide 110.

Initially, a similar sequence as before was applied, involving

allylic oxidation and Takai reaction. However, direct treat-

ment of 87 with DDQ/pH7 buffer gave only low yields (36%)

of the expected aldehyde 109. Alternatively, a two-step

process, involving first liberation of the primary allylic

alcohol, best done with aqueous 65% acetic acid in THF,⁷⁵

and subsequent allylic oxidation with manganese(IV) diox-

ide, proved beneficial giving aldehyde 109 which was directly

used for the Takai olefination, which proceeded in excellent

polarity of starting material and product and slow decom-

position of stannane 46 on silica gel in the absence of

triethylamine rendered isolation of the desired stannane 46

(70) Hodgson, D. M.; Foley, A. M.; Lovell, P. J. Tetrahedron Lett. 1998,

39, 6419.

(71) Smith, A. B.; Wan, Z. J. Org. Chem. 2000, 65, 3738.

(72) We slightly modified the preparation of 101. Direct hydrostannyla-

tion of progargylic alcohol has been reported to produce a mixture of three

possible isomers: E, Z, and geminal bistannylated products and separation of

E/Z mixture was not straightforward (see: Jung, M. E.; Light, L. A.

Tetrahedron Lett. 1982, 23, 3851. In contrast, radical hydrostannylation of

TBS protected propargyl alcohol gave little Z product (see: Hayes, M. P.;

Hatala, P. J.; Sherer, B. A.; Tong, X.; Zanatta, N.; Borer, P. N.; Kallmerten,

J. Tetrahedron 2001, 57, 1515). Therefore, propargyl alcohol was protected

as the TBS ether. The crude ether was directly mixed with tributyltin hydride

(1.2 equiv) and a catalytic amount of AIBN (3% mol) and heated at 80-85°C

for 2 h and then diluted with THF and TBAF (2.0 equiv) to access pure (E)-

stannane in 64% yield over three steps. Subsequent bromination under Appel

reaction conditions with catalytic amounts of 2,6-lutidine (0.2 equiv) gave the

corresponding allylic bromide which was used immediately for displacement

by sodium salt of diethyl phosphate (HPO(OEt)₂, NaH), giving the required

phosphonate 101 in 65% yield over two steps.

(73) At this stage, the all-E configuration of alcohol 105 was confirmed by

¹H NMR (600 MHz, C₆D₆) analysis, revealing the characteristic coupling

constants.

(74) Paterson, I.; Cowden, C. J.; Rahn, V. S.; Woodrow, M. D. Synlett

1998, 915.

(75) With TBAF in THF, the primary TBS could be removed in variable

yields (50-85%) together with some over-deprotected product even at 0 °C.

2440 J. Org. Chem. Vol. 75, No. 8, 2010

Li et al.

JOCFeatured Article

SCHEME 17

SCHEME 18

obtained using a variety of conditions, including CSA, HF

and HCl in various concentrations and solvents and no

traces of the desired compound could be detected by

LC-MS analysis (entries 1-5).

Due to these difficulties, we turned our attention to the

macrocycle 87. Furthermore, to understand the difficulties of

the deprotection process, we first evaluated concomitant removal

of the acetonide and TBS groups with the northern fragment 113.

For this compound, both hydrochloric acid and hexafluorosilicic

acid (H₂SiF₆) proved effective with complete removal, resulting

in selective formation of the desired tetraol 114. However, when

these or related conditions were transplanted onto macrocycle 87

(entries 6-9), again only complex mixtures were obtained, and

no traces of the desired product were found. Similar results were

obtained with various Lewis acids, including TMSOTf or

BF₃-OEt₂(entries 10 and 11).

At this stage we decided to apply nonacidic conditions and

attempted to first remove the TBS groups. While some

reagents (TASF, Zn(NO₃)₂], entries 12 and 19) were too

unreactive, giving mainly removal of the primary TBS group,

other likewise mild reagents resulted in extensive decomposi-

tion (silica gel supported FeCl₃, BF₃OEt₂/1,3-propane-

dithiol, BF₃OEt₂, TBSOTf, entries 11, 17, 18, and 20).

3

Finally, TBAF (1.0 M in THF) emerged as a promising

reagent, resulting in a limited number of products. However

by conventional aqueous workup, the corresponding open-

chain seco-acid was obtained as the main product. Accord-

ingly, to avoid potential saponification during workup,

which may be caused by basic fluorides, we applied an

alternative workup procedure, recently developed by the

Kishi group, by means of Dowex 50w, CaCO₃, and MeOH.⁷⁷

Indeed, applying this procedure (entry 14) resulted in the

formation of three distinct compounds, which were even-

tually confirmed to be 116 together with translactonized 115

and 117 (see Scheme 18), together with smaller amounts of

the corresponding seco-acid. After further optimization

(entries 20-22), optimum results were obtained with use of

yield and acceptable stereoselectivity (92% over two steps,

E/Z = 4:1).⁷⁶

For deprotection, we initially planed to remove both the

acetonide as well as the TBS groups of 110 in a one-pot

fashion under acidic conditions. However, as shown in the

table in Scheme 17, only complex product mixtures could be

(76) The olefination product isomerizes rapidly to an E/Z = 1:1 mixture

upon exposure to light.

(77) Kaburagi, Y.; Kishi, Y. Org. Lett. 2007, 9, 723.

J. Org. Chem. Vol. 75, No. 8, 2010 2441

Li et al.

JOCFeatured Article

SCHEME 19

FIGURE 2. Stability of etnangien methyl ester at various pH

values.

TBAF buffered with HOAc in combination with short

reaction times, in combination with the Kishi workup,

resulted in the formation of three distinct isomers 115, 116,

and 117 with complete loss of TBS groups, which were

isolated after preparative HPLC separation on silica gel

(DCM/i-PrOH = 91:9) in preparatively acceptable yields,

with only negligible traces of the seco-acid.

With these TBS-deprotected macrocycles in hand, we

again studied removal of the acetonide. No reaction was

observed when lactone 115 was subjected to a suspension of

Dowex 50w in methanol for 8 h or PPTS in methanol,

suggesting a high degree of steric hindrance exerted by the

adjacent methyl groups. While acetonide removal may be

effected by treatment of 116 with CSA/MeOH or HCl/

MeOH, in both cases several products were obtained.

Furthermore, LC-MS analysis suggested the incorporation

of an additional methylene-unit, presumably by addition of

methanol, suggesting alternative solvents to be more appro-

priate. Acetonide removal was possible with 65% aqueous

acetic acid; however, again a number of side products were

concomitantly formed. Other methods which were evaluated

included silica gel supported FeCl₃, leading to dehydration

or PdCl₂(CH₃CN)₂, which resulted in a complex mixture.

We also wondered if the ring size makes difference upon

deprotection of the acetonide. These results suggested that

also the primary allylic C-15 alcohol might be too labile

under these acidic conditions. Together with the propensity

of isomerization of the terminal vinyl iodide 110, we re-

evaluated the endgame strategy and decided to remove the

acetonide after the Stille coupling.

To further evaluate the endgame strategy, the stability of a

solution of etnangien methyl ester (2)⁷⁸in methanol under

various pH values was studied, as shown in Figure 2. While, as

expected,³the methyl ester was stable under neutral conditions

and can be conveniently stored at pH 7 in dark vessels in

solution, considerable degrees of decomposition were observed

at very low (pH < 3) or high pH values (pH > 10), in agree-

ment with the observed lability of 87, 111, and 115-118 in

acidic media. However, despite this pronounced tendency for

decomposition, it was observed that even at pH 2, around 60%

of the initial product could be retained after 2 h. Accordingly,

it was envisioned that final acetonide removal might be possible

within this pH range and short reaction times, in agreement

with promising results obtained above with HOAc.

Accordingly, as shown in Scheme 19, the macrocyclic vinyl

iodide 110 was submitted to the previously established

conditions (TBAF/AcOH) for complete removal of all TBS

groups giving the expected three isomers 118, 120, and 121

together with partially deprotected 119 as the major pro-

ducts, after purification by preparative HPLC on silica gel.

Subsequent Stille coupling of 120 with iodide 105 proceeded

smoothly using our previously established conditions, giving

the desired coupling product 122 in good yields (70%), which

was then submitted to 65% acetic acid at 0 °C. Gratifyingly,

the desired etnangien methyl ester could then be obtained

in acceptable yield (35%) after 80 min. Finally, cleavage

of the methyl ester was effectuated with porcine liver esterase

following our previously established procedure,³giving

etnangien, which was identical in all aspects (¹H and ¹³C

NMR, HRMS, R_f, and optical rotation) to an authentic

sample.

Conclusions

In conclusion, a highly stereoselective total synthesis of

etnangien was accomplished by a convergent strategy. It

proceeds in 23 steps (longest linear sequence) and 0.25%

yield starting from commercially available aldehyde 65. This

work establishes unambiguously the relative and absolute

configuration. Notable synthetic features include an array of

highly stereoselective aldol reactions to set the characteristic

methyl- and hydroxyl-bearing stereogenic centers, an effi-

cient conformation controlled Heck macrocyclization, by

use of an acetonide protecting group, mimicking solution

(78) Etnangien methyl ester was available from previous isolation proce-

dures; see ref 3.

2442 J. Org. Chem. Vol. 75, No. 8, 2010

Li et al.

JOCFeatured Article

structure of the natural products and a late-stage introduc-

tion of the labile side chain by an efficient Stille coupling

reaction, in the presence of four nonprotected hydroxyls.

Along this synthesis, considerable time and effort had to

be invested in a few critical steps. The best method to

construct the macrocyclic core proved to be the most direct

approach, a Heck reaction involving a nonfunctionalized

alkene. With regard the side-chain stannane 105, after

several unsuccessful attempts to introduce the stannane from

a terminal alkene, an effective approach involving an HWE

reaction with a metalated allylic phosphonate (101) was

developed. However, the most demanding part appeared

very late. Global deprotection of all TBS groups and the

acetonide was extremely demanding, resulting in a plethora

of unexpected side reactions, reminding us many times that a

much deeper understanding of this kind of natural products

and related chemistry is vital. Ultimately, a two-step method

(TBAF/AcOH removal of all TBS groups and 65% AcOH

deprotection of the acetonide) enabled us to attain the goal,

however, in only very low yield. This problem should be

addressed in future synthetic studies toward the etnangiens.

The presented modular synthesis of etnangien should be

instructive and amenable to designed analogues of this novel

RNA-polymerase inhibitor, thus enabling further explora-

tion of the promising biological potential of this macrolide

antibiotic.

0.25 mmol, 1.30 equiv) in 250 mL of toluene was heated to 50 °C

and stirred for 96 h in the absence of light. The mixture was

cooled to room temperature and filtered. The precipitate was

washed with dry toluene and dry diethyl ether subsequently, and

then dried under vacuum to yield a colorless solid (56.8 g, 0.11

mol, 56%): ¹H NMR (300 MHz, DMSO-d₆) δ = 5.08 (d, J = 8.9

Hz, 2 H), 7.73-7.96 (m, 15 H); ¹³C NMR (75 MHz, DMSO-d₆)

δ = 117.9, 119.1, 130.4, 134.1, 135.5; MS calcd for C₁₉H₁₇IP^þ

[M]^þ= 529.9, found 529.5.

Stork-Zhao-Wittig Reaction to Vinyl Iodide 55. In the

absence of light, to a solution of NaHMDS (1.0 M in THF,

2.96 mL, 2.96 mmol, 1.5 equiv) in absolute THF (58 mL) was

added a suspension of the phosphonium iodide (1.57 g, 2.96

˚

mmol, 1.5 equiv) in anhydrous HMPA (dried over 4 A mole-

cular sieves, 4.0 mL and 2 ꢀ 2.5 mL washes) at 10-15 °C

(ice-water bath). After 1 min, the mixture was cooled to -78 °C.

and a solution of aldehyde 53 (541 mg, 1.97 mmol, 1.0 equiv) in

THF (1.0 mL and 2 ꢀ 0.5 mL washes) was added dropwise. After

being stirred at -78 °C for 20 min, the reaction was quenched

by addition of aqueous saturated NH₄Cl (25 mL) and water

(10 mL). When the mixture was warmed to room temperature,

the insoluble solids were filtered off and the layers of the filtrate

were separated. The aqueous layer was extracted with diethyl

ether (3 ꢀ 25 mL). The combined organic phases were washed

withbrine (40mL), dried overMgSO₄, filtered, andconcentrated.

Flash chromatography on silica gel (80 g, hexanes/Et₂O = 35: 1

to 12:1) afforded the vinyl iodide 55 (607 mg, 1.52 mmol, 77%,

Z/E = 27:1): R_f= 0.68 (hexanes/EtOAc = 9:1); [R]²³_D= -9.9

(c = 2.0, CHCl₃); ¹H NMR (400 MHz, CDCl₃) δ = 0.06 (s, 6H),

0.90 (s, 9H), 1.36-1.60 (m, 6H), 2.30-2.43 (m, 2H), 3.30 (m, 1H),

3.36 (s, 3H), 3.61 (d, J = 6.6 Hz, 1H), 3.63 (d, J = 6.1 Hz, 1H),

6.24-6.32 (m, 2H);¹³C NMR (100 MHz, CDCl₃) δ = -5.2, 18.4,

21.8, 26.0, 32.9, 33.5, 38.6, 56.7, 63.1, 79.6, 84.1, 137.7; HRMS

calcd for C₁₅H₃₁IO₂SiNa [M þ Na]^þ421.1036, found 421.1039.

Wittig Reaction to Aldehyde 67. To a flask containing

Ph₃PCHCHO (146 mg, 0.481 mmol, 1.0 equiv) was added a

solution of (E)-ethyl 3-methyl-4-oxobut-2-enoate (207 μL, 1.52

mmol, 3.0 equiv) in absolute DCM (100 μL). The reaction

mixture was stirred for 15 h at room temperature and purified

direct by flash chromatography (hexanes/Et₂O = 6:1 to 1:1) to

give aldehyde 67 (64.6 mg, 0.384 mmol, 80%) and aldehyde 68

(2.5 mg, 13 μmol, 2.7%) as amorphous solids. Aldehyde 67:

R_f= 0.37 (hexanes/EtOAc = 9:1); ¹H NMR (300 MHz, CDCl₃)

δ = 1.29 (t, J = 7.2 Hz, 3H), 2.29 (s, 3H), 4.20 (q, J = 7.2 Hz,

2H), 6.11 (s, 1H), 6.43 (dd, J = 15.6, 7.4 Hz, 1H), 7.08 (d, J =

15.6 Hz, 1H), 9.65 (d, J = 7.5 Hz, 1H); ¹³C NMR (75 MHz,

CDCl₃) δ = 13.8, 14.2, 60.5, 127.0, 132.7, 148.6, 154.2, 165.8,

193.3; HRMS calcd for C₉H₁₂O₃[M]^þ168.0786, found

168.0788. Aldehyde 68: R_f= 0.37 (hexanes/EtOAc = 9:1); ¹H

NMR (300 MHz, CDCl₃) δ = 1.28 (t, J = 7.2 Hz, 3H), 2.31 (s,

3H), 4.18 (q, J = 7.0 Hz, 2H), 6.94 (s, 1H), 6.26 (dd, J = 15.3, 7.7

Hz, 1H), 6.65 (d, J = 15.4, 1H), 6.74 (dd, J = 15.3, 10.4 Hz, 1H),

Experimental Section

1,3-Syn Reduction of 20 to 21. To a solution of chlorodicy-

clohexylborane (27 μL, 0.125 mmol) in 300 μL of Et₂O was

added Et₃N (18 μL, 0.125 mmol). The mixture was cooled to

-20 °C before a solution of the ketone 20 (50 mg, 0.083 mmol) in

50 μL of Et₂O was added. The mixture was stirred at -30 °C for

75 min and then cooled to -78 °C before 200 μL of a solution of

LiBH₄in THF (2 M, 0.415 mmol) was added. The mixture was

stirred again for 60 min at -78 °C and then diluted with Et₂O

(1 mL) and quenched with 1 mL of satd aq NH₄Cl solution. The

layers were separated, and the aqueous layer was extracted with

Et₂O (3 ꢀ 2 mL). The combined organic extracts were washed

with brine and concentrated in vacuo to an oil. The crude oil was

dissolved in MeOH (330 μL), 10% NaOH solution (100 μL), and

30% H₂O₂(150 μL), and stirred for 60 min at rt before the

reaction was diluted with 2 mL of H₂O and 2 mL of DCM. The

layers were separated, and the aqueous layer was extracted with

CH₂Cl₂(3 ꢀ 3 mL). The combined organic phases were dried

with MgSO₄and concentrated to an oil. The reaction was

purified by column chromatography (light petroleum ether/

EtOAc 85:15) to afford the diol 21 (42.6 mg, 0.071 mmol,

87%, dr 19:1) as a clear oil: R_f= 0.28 (light petroleum ether/

7.16 (dd, J = 15.3, 10.4 Hz, 1H), 9.60 (d, J = 7.9 Hz, 1H); ¹³

C

EtOAc 90:10); [R]²²= þ2.5 (c = 1.1, CHCl₃); ¹H NMR

D

NMR (75 MHz, CDCl₃) δ = 13.6, 14.3, 60.2, 123.9, 131.1,

133.5, 144.8, 150.0, 150.4, 166.4, 193.3; HRMS calculated for

C₁₁H₁₄O₃[M]^þ194.0943, found 194.0940.

(300 MHz, CDCl₃) δ ppm 0.02 (s, 3H), 0.03 (s, 3H), 0.05 (6H),

0.87 (s, 9H), 0.88 (s, 9H), 0.98 (d, J = 6.8 Hz, 3H), 1.00 (d, J =

7.0 Hz, 3H), 1.01 (d, J = 6.8 Hz, 3H), 1.73 (m, 2H), 1.81 (m, 1H),

1.88 (m, 1H), 1.96 (m, 1H), 3.55 - 3.47 (m, 2H), 3.51 - 3.60 (m,

3H), 3.75 (s, 3H), 3.85 (dd, J = 6.7, 2.9 Hz, 1H), 3.87 (dd, J =

7.4, 2.4 Hz), 4.39 (d, J = 11.9 Hz), 4.40 (d, J = 11.9 Hz), 6.87

(dd, J = 8.7, 1.9 Hz, 2H), 7.24 (dd, J = 8.7, 1.9 Hz, 2H); ¹³C

NMR (75 MHz, CDCl₃) δ ppm -5.5, -5.4, -4.6, -4.5, 8.5,

11.4, 13.7, 17.9, 18.3, 23.8, 25.8, 25.9, 29.7, 34.2, 37.3, 37.9, 38.2,

55.5, 66.2, 66.7, 72.7, 73.8, 74.7, 75.9, 113.8, 129.5, 130.2, 159.9;

HRMS calcd for C₃₂H₆₂O₆NaSi₂[M þ Na]^þ621.3983, found

621.3990.

Heck Macrocyclization to 87. To a solution of ester iodide 88

(88.0 mg, 71.2 μmol, 1.0 equiv), Pd(OAc)₂(16.0 mg, 71.2 μmol,

1.0 equiv), Bu₄NCl (49.5 mg, 178 μmol, 2.5 equiv), and K₂CO₃

(78.5 mg, 569 μmol, 8.0 equiv) was added DMF (18.0 mL) at

room temperature. The resulting yellow suspension was stirred

at 70 °C for 50 min. The mixture was cooled to room tempera-

ture, diluted with Et₂O (40 mL), and filtered through a Celite

plug (3 ꢀ 10 mL Et₂O). After removal of the solvent, the residue

was purified by column chromatography (hexanes/Et₂O =

20:1) to give diene 87 (55.1 mg, 49.7 μmol, 70%) as a colorless

oil: R_f= 0.55 (hexanes/Et₂O = 9:1); [R]²⁰_D= þ6.3 (c = 0.57,

CHCl₃); ¹H NMR (600 MHz, CDCl₃) δ = -0.04 (s, 3H),

Preparation of Stork-Zhao-Wittig Reagent. A solution of

PPh₃(50.0 g, 0.19 mol, 1.00 equiv) and diiodomethane (20.0 mL,

J. Org. Chem. Vol. 75, No. 8, 2010 2443

Li et al.

JOCFeatured Article

0.005 (s, 3H), 0.02 (s, 3H), 0.04 (s, 3H), 0.07 (s, 9H), 0.08

(s, 3H), 0.83 (d, J = 7.1 Hz, 3H), 0.84 (s, 9H), 0.87 (s, 9H),

0.89 (s, 9H), 0.90 (d, J = 6.6 Hz, 3H), 0.91 (s, 9H), 0.91 (d, J =

7.0 Hz, 3H), 0.93 (d, J = 6.6 Hz, 3H), 1.34 (s, 3H), 1.37 (s, 3H),

1.39-1.53 (m, 6H), 1.61 (m, 1H), 1.68-1.75 (m, 2H), 1.73 (s,

3H), 1.80-1.85 (m, 2H), 1.89-1.93 (m, 2H), 2.11 (m, 1H), 2.20

(m, 2H), 2.34 (dd, J = 16.1, 8.8 Hz, 1H), 2.46 (m, 1H), 3.20 (m,

1H), 3.31 (s, 3H), 3.31-3.36 (m, 2H), 3.78 (m, 1H), 4.24-4.26

(m, 3H), 4.40 (dd, J = 9.2, 6.6 Hz, 1H), 5.17 (m, 1H), 5.36 (d,

J = 9.5 Hz, 1H), 5.38 (m, 1H), 5.64 (m, 1H), 5.69 (dt, J = 15.8,

5.3 Hz, 1H), 6.04 (t, J = 10.8 Hz, 1H), 6.20 (d, J = 15.8 Hz, 1H),

6.23 (dd, J = 14.7, 11.0 Hz, 1H); ¹³C NMR (100 MHz, CDCl₃)

δ = -5.1, -4.7, -4.5, -4.4, -4.3, -4.1, -3.8, 5.8, 11.3, 13.3,

16.3, 18.0, 18.1, 18.5, 19.6, 25.9, 26.0, 26.1, 30.1, 31.1, 31.8, 35.0,

37.2, 38.3, 40.1, 44.1, 56.5, 64.1, 70.7, 71.0, 72.9, 80.3, 99.1,

126.0, 127.1, 127.9, 130.2, 132.2, 133.1, 133.6, 134.4, 171.0;

HRMS calcd for C₆₁H₁₁₈O₉Si₄Na [M þ Na]^þ1129.7751, found

1129.7743.

Stannane 46. To a cooled (-78 °C) solution of phosphonate

101 (74.9 mg, 160.3 μmol, 1.2 equiv) and 18-crown-6 (70.6 mg,

267.2 μmol, 1.2 equiv) in dry THF (0.8 mL) was added dropwise

a solution of KHMDS (0.5 M in toluene, 320.6 μL, 1.2 equiv).

After the solution was stirred for 5 min, a solution of aldehyde 99

(45.5 mg, 133.6 μmol, 1.0 equiv) in dry THF (0.2 and 0.2 mL

washing) was added dropwise. The resulting mixture was stirred

at -78 °C for 1 h followed by dilution with n-hexane (5.0 mL)

and quenching with pH 7 phosphate buffer (4.0 mL). After the

mixture was warmed to room temperature, the layers were

separated and the aqueous layer was extracted with n-hexane

(2 ꢀ 2.0 mL). The combined organic phases were washed with

brine (4.0 mL), dried over anhydrous sodium sulfate, filtered,

and concentrated under vacuum. The residue was chromato-

graphed by preparative HPLC (column: LiChrospher Si 60

(10 μm) 250 ꢀ 10 mm, n-hexane/EtOAc = 97:3, 6.0 mL/min,

254 nm) to produce the pure trienylstannane (retention time =

8.08 min, 33.1 mg, 50.6 μmol, 38% yield).

Etnangien Methyl Ester 2. A cooled (0 °C) solution of the

acetonide 122 (0.3 mg, 0.34 μmol) in AcOH/water (65/35 v/v,

100 μL) was stirred at this temperature for 80 min. Aqueous

saturated NaHCO₃(1.5 mL) and EtOAc (1.5 mL) were added,

and the layers were separated. The aqueous layer was extracted

with EtOAc (9 ꢀ 2.0 mL). The combined organic phases were

dried over anhydrous sodium sulfate, filtered, and concentrated

under reduced pressure. The residue was flash chromato-

graphed (silica gel, dichloromethane/methanol = 15:1 to 10:1)

to afford the product 2 (0.1 mg, 0.12 μmol, 35% yield): R_f= 0.31

(dichloromethane/methanol =10:1); [R]²²_D= þ17.0 (c = 0.20,

MeOH); ¹H NMR (600 MHz, acetone-d₆) δ ppm 0.84 (d, J = 7.2

Hz, 3H), 0.92 (d, J = 6.8 Hz, 3H), 0.97 (d, J = 7.0 Hz, 3H), 0.98

(d, J = 6.8 Hz, 3H), 1.31 (m, 2H), 1.41 (m, 1H), 1.42 (m, 1H),

1.46 (m, 1H), 1.65 (m, 1H), 1.66 (d, J = 0.7 Hz, 1H), 1.75 (m,

1H), 1.82 (d, J = 0.7 Hz, 1H), 1.84 (m, 1H), 1.87 (m, 1H), 1.95

(m, 1H), 1.98 (m, 1H), 2.22 (m, 1H), 2.23 (m, 1H), 2.27 (m, 1H),

2.33 (ddd, J = 14.9, 7.5, 6.8, 1H), 2.37 (m, 1H), 2.38 (m, 1H),

2.40 (d, J = 5.5 Hz, 1H), 2.44 (m, 1H), 3.25 (m, 1H), 3.29 (s, 3H),

3.44 (dd, J = 6.8, 3.8 Hz, 1H), 3.51 (m, 1H), 3.56 (m, 1H), 3.61

(s, 3H), 3.61 (dd, J = 7.9, 2.6 Hz, 1H), 3.72 (m, 1H), 3.81 (d, J =

2.6 Hz, 1H), 3.89 (m, 1H), 3.98 (d, J = 4.4 Hz, 1H), 4.17 (dt, J =

5.5, 5.5 Hz, 1H), 4.37 (ddd, J = 8.4, 6.8, 6.8 Hz, 1H), 4.39 (dd,

J = 8.7, 8.3 Hz, 1H), 5.25 (d, J = 8.3 Hz, 1H), 5.35 (m, 1H), 5.44

(ddd, J = 8.5, 4.4, 4.4 Hz, 1H), 5.53 (d, J = 9.0 Hz, 1H), 5.71

(ddd, J = 14.9, 8.5, 6.2 Hz, 1H), 5.76 (ddd, J = 14.9, 6.8, 7.5 Hz,

1H), 6.06 (dd, J = 10.9, 10.9 Hz, 1H), 6.16 (dd, J = 14.9,

10.4 Hz, 1H), 6.24 (m, 1H), 6.25 (m, 1H), 6.31 (m, 2H), 6.33 (d,

J = 17.9 Hz, 1H), 6.35 (m, 3H); ¹³C NMR (100 MHz, acetone-d₆)

δ ppm 7.2, 10.9, 11.1, 13.2, 15.1, 16.7, 22.2, 30.7, 33.0, 33.2, 35.2,

36.5, 37.3, 37.7, 38.6, 38.7, 39.2, 42.4, 42.8, 43.8, 51.7, 56.4,

67.9, 68.4, 69.0, 69.1, 73.5, 76.9, 79.4, 81.1, 126.0, 128.3, 129.5,

130.0, 130.9, 132.1, 132.6, 133.3, 133.6, 133.7, 133.9, 134.0, 134.1,

134.3, 135.5, 135.9, 136.2, 138.3, 173.6, 174.2; HRMS calcd for

C₅₀H₇₈O₁₁Na [M þ Na]^þ877.5444, found 877.5442.

Etnangien 1. A solution of etnangien methyl ester 2 (1.00 mg,

1.17 μmol) in 500 μL of DMSO/water = 1:3 was treated under

argon at room temperature with 2 mg of porcine liver esterase

(obtained from Sigma) and stirred for 24 h. The crude product

was purificated by HPLC (Nucleosil 100-7 C18, 250/21; MeOH/

H₂O, 0.01 M phosphate buffer =75:25, flow: 18 mL/min; UV

detection 355 nm) to give etnangien (1, 0.65 mg, 0.71 μmol, 61%)

as a colorless oil: [R]²⁰_D= þ16.2 (c = 0.65 mg/mL, MeOH) [lit.⁷

[R]²¹= þ18.7 (c = 0.9, MeOH)]; ¹H NMR (600 MHz,

D

acetone-d₆) δ ppm 0.84 (d, J = 7.2 Hz, 3H), 0.92 (d, J =

6.4 Hz, 3H), 0.97 (d, J = 7.0 Hz, 3H), 0.98 (d, J = 6.8 Hz,

3H), 1.31 (m, 2H), 1.41 (m, 1H), 1.42 (m, 1H), 1.46 (m, 1H), 1.65

(m, 1H), 1.66 (d, J = 1.1 Hz, 1H), 1.75 (m, 1H), 1.82 (d, J =

0.7 Hz, 1H), 1.84 (m, 1H), 1.87 (m, 1H), 1.95 (m, 1H), 1.98 (m,

1H), 2.22 (m, 1H), 2.23 (m, 1H), 2.27 (m, 1H), 2.33 (ddd, J =

14.9, 7.5, 6.8, 1H), 2.37 (m, 1H), 2.38 (m, 1H), 2.40 (m, 2H), 2.44

(m, 1H), 3.25 (dddd, J = 6.0, 6.0, 5.6, 4.5 Hz, 1H), 3.29 (s, 3H),

3.44 (dd, J = 6.8, 3.8 Hz, 1H), 3.51 (tt, J = 8.3, 4.1 Hz, 1H), 3.56

(m, 1H), 3.61 (dd, J = 7.9, 2.6 Hz, 1H), 3.73 (m, 1H), 3.84 (d, J =

2.6 Hz, 1H), 3.89 (m, 1H), 3.98 (m, 1H), 4.17 (dt, J = 5.4, 6.2 Hz,

1H), 4.37 (ddd, J = 8.4, 6.4, 6.8 Hz, 1H), 4.39 (dd, 9.2, 7.8 Hz,

1H), 5.25 (d, J = 8.3 Hz, 1H), 5.35 (ddd, J = 10.7, 7.4, 7.4 Hz,

1H), 5.44 (m, 1H), 5.53 (m, 1H), 5.71 (ddd, J = 14.9, 8.5, 6.2 Hz,

1H), 5.76 (ddd, J = 14.9, 6.8, 7.5 Hz, 1H), 6.06 (dd, J = 11.2, 11.2

Hz, 1H), 6.16 (dd, J = 14.9, 10.4 Hz, 1H), 6.24 (m, 1H), 6.25 (m,

1H), 6.31 (m, 2H), 6.33 (d, J = 17.9 Hz, 1H), 6.35 (m, 3H); ¹³

C

NMR (100 MHz, acetone-d₆) δ ppm 7.2, 10.9, 11.1, 13.2, 15.1,

16.7, 22.0, 30.7, 33.1, 33.2, 35.1, 36.5, 37.3, 37.8, 38.71, 38.72, 39.2,

42.3, 42.8, 43.9, 56.4, 68.1, 68.4, 69.0, 69.6, 73.5, 76.8, 79.4, 81.1,

126.0, 128.3, 129.5, 129.9, 130.9,132.1, 132.6, 133.4, 133.5, 133.7,

133.8, 134.0, 134.2, 134.3, 135.5, 135.9, 136.4, 138.6, 173.5, 174.3;

HRMS calcd for C₄₉H₇₅O₁₁[M - H] 839.5309, found 839.5312.

All data were identical to those previously reported for etnangien

from S. cellulosum.¹

Acknowledgment. This work was supported by the

“Deutsche Forschungsgemeinschaft”, the “Wild-Stiftung”,

€

and the “Fonds der Chemischen Industrie”. We thank Bjorn

Wiegmann for support and Tanja Lau for early exploratory

studies. Excellent NMR assistance by Markus Enders and

Beate Termin is most gratefully acknowledged.

Supporting Information Available: Full experimental details

and copies of the ¹H and ¹³C NMR spectra for all new

compounds. This material is available free of charge via the

Internet at http://pubs.acs.org.

2444 J. Org. Chem. Vol. 75, No. 8, 2010

Article Doi

Stereoselective total synthesis of etnangien and etnangien methyl ester

DOI: 10.1021/jo100201f

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jo100201f

Full text of DOI:10.1021/jo100201f

Products guided by the article

R&D Labs maybe for 1221294-85-4

Relevant to this article

Hot Product