Callipeltosides A, B and C: Total Syntheses and Structural Confirmation

DOI: 10.1002/chem.201501877

Full Paper

&

Natural Products |Very Important Paper|

Callipeltosides A, B and C: Total Syntheses and Structural

Confirmation

James R. Frost,*^[a]Colin M. Pearson,^[a]Thomas N. Snaddon,^{[a, b]}Richard A. Booth,^[a]

Richard M. Turner,^[a]Johan Gold,^[a]David M. Shaw,^[a]Matthew J. Gaunt,^[a]and Steven V. Ley*^[a]

Abstract: Since their isolation almost 20 years ago, the calli-

peltosides have been of long standing interest to the syn-

thetic community owing to their unique structural features

and inherent biological activity. Herein we present our full

research effort that has led to the synthesis of these mole-

cules. Key aspects of our final strategy include 1) synthesis

of the C1–C9 pyran core (5) using an AuCl₃-catalysed cyclisa-

tion; 2) formation of C10–C22 vinyl iodide (55) by sequential

bidirectional Stille reactions and 3) diastereoselective union

of these advanced fragments by means of an alkenylzinc ad-

dition (d.r.=91:9 at C9). The common callipeltoside aglycon

(4) was completed in a further five steps. Following this, all

three sugar fragments were appended to provide the entire

callipeltoside family. In addition to this, d-configured callipel-

tose B was synthesised and appended to the callipeltoside

1

aglycon. The H NMR spectrum of this molecule was found

to be significantly different to the natural isolate, further

supporting our assignment of callipeltoside B (2).

Introduction

Natural product synthesis continues to provide an attractive

platform for the discovery of new synthetic methods and fur-

ther elaboration of novel synthesis pathways. In doing so, this

effort importantly not only provides material for biological

evaluation but also serves as a tool enabling unambiguous

confirmation of structure. This can, in many examples, lead to

structural refinement or even complete re-evaluation. A case in

point concerns callipeltosides A, B and C (1–3) (Figure 1,

shown in their finally corrected form). These fascinating com-

pounds were first isolated by Minale in 1996 in vanishingly

small quantities from the marine sponge Callipelta sp.^[1]Prelimi-

nary biological assays indicated a degree of cytotoxicity

against human bronchopulmonary non-small-cell lung carcino-

ma (NSCLC-N6 and P388 cell lines).^[1]However, it was their un-

usual structural features: a 14-membered macrolide incorporat-

ing a tetrahydropyran hemiacetal together with a di-ene-yne

attached to a trans-configured chlorocyclopropane ring that in-

Figure 1. The callipeltoside family of natural products.

[a] Dr. J. R. Frost, Dr. C. M. Pearson, Prof. T. N. Snaddon, Dr. R. A. Booth,

Dr. R. M. Turner, Dr. J. Gold, Dr. D. M. Shaw, Prof. M. J. Gaunt, Prof. S. V. Ley

Department of Chemistry, University of Cambridge

Lensfield Road, Cambridge, CB2 1EW (UK)

trigued the synthesis community. The trans-configured chloro-

cyclopropane ring was of particular interest, since this feature

is extremely scarce even now, with the phorbasides^[2]and the

recently isolated muironolide A^[3]being the only other exam-

ples. Given the low availability of callipeltosides (0.8–3.5 mg)

and their lack of suitable crystallinity for X-ray studies their

complete structural assignment has been challenging.

E-mail: james.frost@chem.ox.ac.uk

svl1000@cam.ac.uk

[b] Prof. T. N. Snaddon

Current address: Department of Chemistry, Indiana University

800 E. Kirkwood Ave., Bloomington, IN 47405 (USA)

Supporting information for this article is available on the WWW under

htt p ://dx.doi.org/10.1002/chem.201501877.

ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.

This is an open access article under the terms of the Creative Commons At-

tribution License, which permits use, distribution and reproduction in any

medium, provided the original work is properly cited.

While Minale correctly deduced the connectivity of the calli-

peltosides, several stereochemical features remained unclear

and could not be resolved. Although the relative stereochemis-

Chem. Eur. J. 2015, 21, 13261 – 13277

13261 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

try of the C1ÀC19 and C1’ÀC8’ fragments of these molecules

could be deduced, the main concerns centred on the relative

configuration of the trans-chlorocyclopropane ring with re-

spect to the rest of the molecule, the stereochemical assign-

ment of each glycoside moiety (d or l) and the stereochemis-

try of the glycosidic linkage. To solve these structural uncer-

tainties and unambiguously deduce the absolute stereochem-

istry of the callipeltosides, total synthesis of these molecules

was clearly necessary.

The synthesis of callipeltoside A was first achieved by the

groups of Trost^[4]and Evans^[5]in 2002, following Paterson’s^[6]

enantiomeric synthesis of the callipeltoside aglycon in 2001.

Further syntheses of callipeltoside A were completed thereafter

by the groups of Paterson,^[7]Panek,^[8]and Hoye^[9]as well as the

preparation of several advanced fragments by others. In 2008,

the synthesis of callipeltoside C was disclosed by MacMillan,^[10]

and in so doing confirmed the absolute structure and provided

evidence suggesting that the sugar moiety was l-configured

as in callipeltoside A.

Prior to our recent report,^[11]the synthesis of callipeltoside B

had not been achieved, whilst the glycosidic linkage of callipel-

toside C had only been tentatively assigned on the basis of

¹H NMR coupling constants.^[10]Whilst there has clearly been

a considerable amount of effort dedicated to the syntheses of

these molecules,^[12]we sought to develop an approach that

would allow access to not only one, but all three of the calli-

peltosides in a highly convergent manner. Our synthetic strat-

egy to complete various fragments of these molecules has

evolved considerably over time, with certain methodologies,

often developed in our own lab, superceded by more efficient

and scalable alternatives. Herein we present our full research

effort which has resulted in the realisation of our goal. We also

provide further evidence to aid in the stereochemical assign-

ment of the glycosidic linkages present in callipeltosides B and

C.

Scheme 1. Retrosynthesis of callipeltosides A, B and C.

more efficient, alternative methodology, they nicely illustrate

how a complex synthesis evolves.

Results and Discussion

Synthesis of pyran 5

a) Double dithiol conjugate addition approach to C1–C9

pyran 5: Our synthetic effort began with the preparation of

the pyran aldehyde 5 from (R)-configured Roche ester 7. This

was initially protected as its TBS-ether and converted to the

corresponding aldehyde in a further two steps. Reaction with

mesylate 9 using the conditions developed by Marshall^[17]then

enabled formation of the C6 and C7 stereocentres in good

yield (70%) and high diastereoselectivity (d.r.=94:6). The C7 al-

cohol was protected as the p-methoxybenzyl (PMB) ether and

the scaffold further elaborated in two steps to give alkyne 11 a.

This provided the opportunity to perform a double dithiol con-

jugate addition which, as anticipated, smoothly furnished 12a

containing a masked ketone at the C5-position.^[16]Deprotec-

tion of the PMB group using DDQ resulted in spontaneous cyc-

lisation and simultaneously gave the desired functionalised

pyran system as a single diastereoisomer. Subsequent ketal for-

mation and removal of the dithiane provided deprotected

ketone 14a in 29% overall yield over 11 steps (Scheme 2).

At this stage a diastereoselective reduction of the ketone

was required to set the C5 stereocentre. However, under a vari-

ety of common conditions it was not possible to obtain the

desired configuration with good control. Further investigation

suggested that the C3-ketal functionality was responsible, ap-

parently influencing the trajectory of the incoming hydride

source and hence the observed stereochemistry.^[8]This being

the case, we chose to remove the troublesome C3-ketal

through elimination to the pyranone 17a.^[18]Following this,

Synthetic plan

Identical to previous approaches to the callipeltosides, we

chose to firstly disconnect the glycosidic linkage to reveal the

common callipeltoside aglycon 4 (Scheme 1). With the knowl-

edge that callipeltosides A and C both contained l-configured

sugars it was considered likely that callipeltose B was also of

the same configuration and, as such, we set ourselves the ad-

ditional target of synthesising each callipeltoside sugar from

a common, easily accessible precursor. In order to assemble

the callipeltoside aglycon, we chose to use a Yamaguchi pro-

cess^[13]to form the macrocyclic ring but also committed our-

selves to a bold coupling strategy using similar-sized frag-

ments 5 and 6. At the start of the synthetic campaign, we ini-

tially considered the use of an asymmetric organocatalytic cy-

clopropanation method,^[14]developed in our laboratory, to pre-

pare the trans-chlorocyclopropane ring. We anticipated that

the C13 stereocentre could be set by use of a pyrrolidine tetra-

zole catalysed oxyamination,^[15]and a double dithiol conjugate

addition strategy would provide access to the core pyran frag-

ment (5).^[16]While these were later abandoned in favour of

Chem. Eur. J. 2015, 21, 13261 – 13277

www.chemeurj.org

13262 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

Scheme 4. a) TfOH, CH₂Cl₂, RT, 17b=88%, (for 17a (ii) TBSCl, imidazole,

CH₂Cl₂, RT, 83% over 2 steps); b) NaBH₄, CeCl₃·7H₂O, MeOH, À788C; c) TBSCl,

imidazole, DMAP, CH₂Cl₂, RT, 19a=90% over 2 steps, 19b=97% over

2 steps; d) (Æ)-CSA, MeOH, RT, 20a=20–68%, 20b=20–74%; e) TBAF, THF,

RT, TBS=67% (+11 % recovered 20a), for Bn: H₂, Pd/C (10 wt.%), EtOAc, RT,

96%; f) Dess–Martin periodinane, CH₂Cl₂, RT, 87%. (Æ)-CSA=(Æ)-camphorsul-

fonic acid; TBAF=tetrabutylammonium fluoride.

Scheme 2. a) TBSCl, Et₃N, DMAP, CH₂Cl₂, RT, 99%; or benzyl trichloroacetimi-

date, TfOH, CH₂Cl₂, RT, 81%; b) LiBH₄, Et₂O, MeOH, 08C, TBS/Bn=99%;

c) oxalyl chloride, DMSO, Et₃N, CH₂Cl₂, À788C, 8a/8b=quant.; d) Pd(OAc)_2,

PPh₃, Et₂Zn, THF, À78!À258C, 10a=70%, 10b=71%; e) NaH, PMBBr,

DMF/THF (1:1), 08C, TBS=82%, Bn=86%; f) nBuLi, ethyl chloroformate, THF,

À788C!RT, TBS=95%, Bn=95%; g) LDA, methyl acetate, THF, À788C!RT;

h) 1,2-ethanedithiol, NaOMe, CH₂Cl₂/MeOH (1:1), À108C!RT; i) DDQ, pH 7

phosphate buffer, CH₂Cl₂, RT, 13a=68% over 3 steps, 13b=60% over

3 steps; j) PPTS, trimethylorthoformate, MeOH, RT, TBS=92%, Bn=81%;

k) [bis(trifluoroacetoxy)iodo]benzene, MeCN, H₂O, 08C, 14a=86%,

14b=87%. DDQ=2,3-dichloro-5,6-dicyano-1,4-benzoquinone; DMAP=4-di-

methylaminopyridine; LDA=lithium diisopropylamide; PMB=4-methoxy-

benzyl ether. PPTS=pyridine p-toluenesulfonate.

cious, with parallel batches under apparently identical reaction

conditions providing variable yields ranging between 20 and

68%. Unfortunately the variable nature of this reaction prohib-

ited the synthesis of sufficient quantities of material. Our frus-

tration was then further compounded by the fact that we

were also unable to cleanly differentiate between the primary

and secondary TBS protecting groups in the next step of the

sequence. Optimal conditions required the use of TBAF as the

limiting reagent (0.9 equiv), with separation and several recy-

cles to process the material. For this reason, we chose to re-

place the primary TBS-ether with the benzyl protecting group

at an early stage (see Schemes 2 and 3 for yields). This then re-

moved the selectivity issue during the final deprotection step

and provided pyran aldehyde 5 in a marginally improved step-

count (17 steps) and yield (10.4% overall yield) relative to the

analogous TBS-protected sequence (18 steps, 8.6% overall

yield). However, the issues associated with the elimination and

reinstallation of the ketal functionality remained, and therefore

an alternative preparation of this fragment was sought.

b) Gold-catalysed approach to C1–C9 pyran 5: In 2009, we

established a method whereby five and six-membered cyclic

acetals could be synthesised by a AuCl₃-catalysed hydroalkoxy-

lation of appropriate conjugated alkynoates.^[20]This approach

resulted in the formation of pyran motifs that were very similar

to the desired C1ÀC9 pyran aldehyde 5. It was therefore antici-

pated that this new approach would replace the dithiol conju-

gate addition methodology^[16]and could be adapted for the

scale-up of the C1ÀC9 pyran aldehyde 5. In doing so, the im-

portant contiguous C5ÀC8 stereocentres would also be imple-

mented prior to cyclisation so as to overcome the problematic

installation of the C5 stereocentre.

Scheme 3. Installation of the C5 stereocentre.

the desired C5 stereochemistry could be set by Luche reduc-

tion^[19]to give 18a as a single diastereoisomer (Schemes 3

and 4).

As before, aldehyde 8b was produced using a high-yielding

three-step sequence of protection, reduction and oxidation

(Scheme 5). With this in hand, we decided to set the C6 and

C7 stereocentres by means of a diastereoselective crotylation

process. While there have been several methods developed to

This slight detour in our sequence therefore required that

the C3-ketal had to be re-installed in order to complete the

preparation of the key pyran aldehyde 5. This seemingly trivial

transformation was, however, found to be extremely capri-

Chem. Eur. J. 2015, 21, 13261 – 13277

www.chemeurj.org

13263 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

and also importantly required only eight chromatographic pu-

rifications allowing for the synthesis of multigrams of material.

With a high-yielding and scalable route to pyran aldehyde 5 in

place, we turned our attention to the preparation of the di-

ene-yne containing vinyl iodide fragment 6.

Formation of the C16ÀC17 bond: Horner–Wadsworth–

Emmons and Julia–Kocienski approaches to vinyl iodide 6

Our principal retrosynthetic disconnection of the callipeltosides

required union of the fully elaborated vinyl iodide fragment 6

(shown with a generic protecting group) with pyran aldehyde

5 by a diastereoselective alkenylmetal addition to form the

C9ÀC10 bond, allowing for maximum convergence. A clear pri-

ority in the synthesis of vinyl iodide 6 was the construction of

the embedded di-ene-yne system as a single E,E-isomer. In

order to achieve this, we initially chose the C16ÀC17 bond as

the key disconnection, for which there was little literature prec-

edence.^[25]This analysis revealed two potential pathways

(Scheme 6, disconnection A and B) from which the fragment

Scheme 5. a) Benzyl trichloroacetimidate, TfOH, CH₂Cl₂, RT, 81%; b) LiBH₄,

Et₂O, MeOH, 08C, 99%; c) oxalyl chloride, DMSO, Et₃N, CH₂Cl₂, À788C,

quant.; d) crotylborane 21, 4 MS, PhMe, À788C, 70%; e) OsO₄, NMO, ace-

tone/H₂O (2:1), RT; f) NaIO₄, THF/H₂O (10:1), 08C!RT; g) Zn, propargyl bro-

mide, THF, 08C!À1008C, 72% over 3 steps; h) 2,2-dimethoxypropane, (Æ)-

CSA, acetone, RT; i) nBuLi, THF, À40!À788C, then ClCO₂Me, 73% over

2 steps; j) QP-SA, MeOH, RT, 95%; k) AuCl₃(2 mol%), MeOH, RT, 96%; l) 2,6-

lutidine, TBSOTf, CH₂Cl₂, À788C, 91%; m) H₂, Pd/C (10% wt), EtOAc, RT, 96%;

n) Dess–Martin periodinane, K₂CO₃, CH₂Cl₂, RT, 87%. NMO= N-methylmor-

pholine-N-oxide.

achieve related transformations, both the Brown^[21]and

Roush^[22]procedures have been frequently used in the synthe-

sis of complex natural products. Both of these methods result-

ed in good levels of diastereoselectivity (Brown=86:14 c.f.

Roush=88:12), although the Roush crotylation was found to

be higher yielding (70 vs. 37%).^[23]The remaining crucial C5

stereocentre was then set using a three-step sequence involv-

ing dihydroxylation, diol cleavage and propargylzinc addition,

which resulted in good diastereoselectivity (85:15) and yield

(72% yield over 3 steps). Protection as the acetonide, followed

by reaction with methyl chloroformate then provided ynoate

24. At this stage all minor diastereoisomers could be removed

by column chromatography and importantly we could confirm

the relative stereochemistry of the four contiguous C5ÀC8 ster-

eocentres by X-ray crystallography (see the Supporting Infor-

mation). Finally removal of the acetonide afforded the diol 25,

ready for the key AuCl₃-catalysed cyclisation. Treatment of diol

25 with only 2 mol% AuCl₃in MeOH at room temperature re-

sulted in clean formation of pyran 26 in an impressive 96%

yield, with product purification requiring simple filtration

through celite^ꢁto remove metal impurities.^[24]Our previous

route could then be intercepted by TBS-protection of the C5

alcohol (Scheme 5), with hydrogenolysis of the benzyl group

and oxidation of the resulting primary alcohol completing the

synthesis of pyran aldehyde 5.

Scheme 6. Horner–Wardsworth–Emmons and Julia–Kocienski approaches to

vinyl iodide 6.

could be obtained, namely by a Horner–Wadsworth–Emmons

(HWE)^[26]or Julia–Kocienski^[27]coupling partners 27 and 30,

along with their respective aldehydes 28 and 29. As men-

tioned earlier, it was anticipated that the organocatalytic meth-

ods developed in our laboratory would be key to the forma-

tion of the trans-chlorocyclopropane ring^[14]and also the C10À

C15 sub-fragments.^[15]

a) Synthesis of C10–C15 vinyl iodide 37: For the C10ÀC15

unit, we chose to make use of known aldehyde 31, prepared

in two steps from 4-pentyl-1-ol (see the Supporting Informa-

tion). In a similar fashion to the synthesis of callipeltoside C dis-

closed by MacMillan,^[10]we chose to apply an enantioselective

organocatalytic oxyamination reaction^[28]to set the C13 stereo-

centre. This was achieved by treatment of aldehyde 31 with ni-

trobenzene and proline-derived tetrazole catalyst 32 to deliver

33 in excellent enantioselectivity (e.r. determined to be >99:1

following NaBH₄reduction). Practically, we found that aldehyde

33 was not easily isolated in pure form, and therefore it was

conveniently reacted directly with phosphorane 34 in a one-

pot procedure to install the C14/C15 E-configured double

This third-generation sequence provided a much-improved

overall yield of 20% in just 14 steps (from (R)-Roche ester 7),

Chem. Eur. J. 2015, 21, 13261 – 13277

www.chemeurj.org

13264 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

tramolecular organocatalytic asymmetric cyclopropanation pro-

cess developed in our laboratory. Previous studies had shown

that these methods could deliver trans-configured cyclopro-

panes of this type in relatively high yield and in good to excel-

lent diastereo- and enantioselectivity.^[14]

The synthesis began by reaction of a-bromo amide 42 with

acrylphenone (43) to generate trans-configured cyclopropane

45 (e.r.=98.5:1.5) in good yield (82%) (Scheme 9). The use of

these coupling partners importantly provided synthetic han-

dles to allow differentiation between the C20- and C21-posi-

tions of the cyclopropane ring. In order to introduce the chlor-

ine substituent at C21 by a modified Hunsdiecker^[32]reaction

(in analogy to Trost^[4,33]), an oxidation state change was re-

quired to allow formation of a carboxylic acid precursor.

Scheme 7. a) 32 (10 mol%), nitrosobenzene, DMSO, RT; b) 34, THF, RT, 66%

over 2 steps; c) CuSO₄, iPrOH/THF (10:1), 408C, 52%; d) TBSOTf, 2,6-lutidine,

CH₂Cl₂, À788C, 93%; e) DIBAL-H, THF, À788C, quant. DIBAL-H=diisobutyla-

luminium hydride.

bond and provide 35 as a single isomer in 66% yield over two

steps (Scheme 7). Cleavage of the NÀO bond, TBS protection

and reduction provided allylic alcohol 37, from which both the

HWE or Julia–Kocienski coupling partners could be realised.

The allylic alcohol 37 was converted either to ethyl phos-

phonate 39 by using a standard two-step Appel^[29]/Arbuzov^[30]

procedure, or directly by a Mitsunobu^[31]reaction to afford a va-

riety of sulfides (40a–c) (Scheme 8). Although we could oxidise

each TBS-containing sulfide smoothly to the corresponding

sulfone (41a–c) using ammonium molybdate tetrahydrate and

H₂O₂, the analogous TES-protected compound (not depicted)

resulted in multiple side reactions and decomposition products

under the conditions as well as with other oxidants (mCPBA,

MMPP, oxone).

Scheme 9. a) 44 (20 mol%), Cs₂CO₃, MeCN, 808C, 82%; b) Urea·H₂O₂, HFIP,

TFAA, 08C!RT; c) NaOH, H₂O, MeCN, 508C, 48% over 2 steps; d) SOCl₂, RT,

98%; e) 2-mercaptopyridine N-oxide sodium salt, TBAI (20 mol%), DMAP

(20 mol%), CCl₄, RT, then AIBN, 808C, 57%. AIBN=2,2’-azobisisobutyronitrile;

HFIP=hexafluoroisopropanol; TBAI=tetrabutylammonium iodide.

Synthesis of the ester could be achieved using the Baeyer–

Villiger reaction;^[34]however, this itself caused a number of

issues. The oxidation was often found to be exceptionally slug-

gish and gave variable reaction times (2–14 days), often with

incomplete conversion. Furthermore, following ester saponifi-

cation, removal of the phenol generated was difficult and tedi-

ous.^[35]In an attempt to circumvent these issues, organocatalyt-

ic cyclopropanation using benzyl acrylate (correct oxidation

level, ester removable by hydrogenolysis) was investigated. Un-

fortunately, cyclopropanation resulted in significantly reduced

yield and enantioselectivity (not depicted).^[14c]

Undeterred, acid chloride formation and Hunsdiecker reac-

tion pleasingly gave the desired trans-configured chlorocyclo-

propane 47 in excellent diastereoselectivity (d.r.=97:3) in mod-

erate yield over two steps. Although reliable, scale-up of this

sequence was impacted by the requirement for the high dilu-

tion conditions employed (0.02m, CCl₄). Having set the desired

stereochemistry of this key fragment in high diastereo- and

enantioselectivity, we attempted to reduce the amide function-

ality in preparation for the synthesis of versatile dibromoolefin

48, which had been used in previous approaches to the calli-

peltosides.^[4–10]Application of a variety of different reducing

conditions ([Cp₂Zr(H)Cl], LiBHEt₃, DIBAL-H, DIBAL-H/nBuLi and

NH₃·BH₃/nBuLi) gave either no desired product, mixtures, or

could not be reproduced in a reliable fashion.^[36]Despite fur-

ther experimentation, work-arounds, and attempts to deliver

Scheme 8. a) PPh₃, CBr₄, CH₂Cl₂, À408C; b) P(OEt)₃, 1008C, 80% over 2 steps;

c) 40a: 1-tert-butyl-1H-tetrazole-5-thiol, PPh₃, DIAD, THF, RT, 85%; 40b: 1-

phenyl-1H-tetrazole-5-thiol, PPh₃, DIAD, THF, RT, 96%; 40c: mercaptobenzo-

thiazole, PPh₃, DIAD, THF, RT, quant.; d) 41a: Mo₇O₂₄(NH₄)₆·4H₂O, EtOH, H₂O₂,

08C!RT, 76%; 41b: Mo₇O₂₄(NH₄)₆·4H₂O, EtOH, H₂O₂, 08C!RT, 82%; 41c:

Mo₇O₂₄(NH₄)₆·4H₂O, EtOH, H₂O₂, 08C!RT, 85%. DIAD=diisopropylazodicar-

boxylate.

b) Organocatalytic approach to the trans-chlorocyclopro-

pane: As mentioned previously, it was hoped that the trans-

chlorocyclopropane unit could be accessed by means of an in-

Chem. Eur. J. 2015, 21, 13261 – 13277

www.chemeurj.org

13265 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

significant quantities of material, we were forced to abandon

this organocatalytic approach to the trans-chlorocyclopropane

unit and consider a more scalable alternative.

Our attention turned to the study of the Julia–Kocienski ole-

fination. These investigations began with sulfone 41a

(Scheme 8) which, in combination with Cs₂CO₃in THF/DMF

(3:1) at room temperature gave vinyl iodide 55 in moderate

yield (54%), but disappointingly as a mixture of Z/E isomers fa-

vouring the Z form (6:1) (Table 1, entry 1). Changing the sol-

c) Alternative approach to the trans-chlorocyclopropane

unit: Paterson^[6,7]and Panek^[8]have both shown that the asym-

metric Simmons–Smith reaction developed by Charette^[37]pro-

vided rapid access to dibromoolefin 48 in only four steps from

epi-chlorohydrin. This being the case, we also chose to assess

this sequence for the preparation of this fragment.^[38]As ex-

pected, the methodology was robust, delivering the desired

cyclopropyl alcohol in excellent levels of enantioselectivity

(e.r.=97.5:2.5). This could then elaborated to the desired di-

bromoolefin (48) in a further two steps (Scheme 10).

Table 1. Conditions for the formation of the C16ÀC17 bond.^[a]

Entry Sulfone Conditions

Z/E^[b]Yield [%]^[c]

1

2

3

4

5

6

7

41a

41b

41c

41b

Cs₂CO₃, THF/DMF (3:1), RT, 15 h

Cs₂CO₃, DMF, RT, 6 h

6:1

5:1

1:1

54

n.d.^[d]

44

Cs₂CO₃, DMF, RT, 15 h

Cs₂CO₃, DMF, RT, 6 h

1.4:1 n.d.^[d]

KHMDS, 18-crown-6, DMF, À508C, 6 h 1.3:1 n.d.^[d]

Cs₂CO₃, DMF, 08C, 6 h

1.4:1 40^[e]

1:1

Scheme 10. a) nBuLi, TMEDA, THF À788C, 70%; b) Et₂Zn, CH₂I₂, CH₂Cl₂, 08C,

then 50, 51, CH₂Cl₂, 08C!RT, 74%; c) PCC, celite^ꢁ, CH₂Cl₂, RT; d) PPh₃, CBr₄,

CH₂Cl₂, 08C!RT, 70% over 2 steps. PCC=pyridinium chlorochromate; TME-

DA=N,N,N’,N’-tetramethylethylenediamine.

Cs₂CO₃, DMPU, RT, 15 h

52

[a] Reagents and conditions: a) i) nBuLi, THF, À788C, (CHO)_n, 71%; ii)

MnO₂, CH₂Cl₂, RT, used directly. [b] Determined by ¹H NMR spectroscopy

of the crude reaction mixture. [c] Yield based on sulfone 41. [d] Not de-

termined (n.d.). [e] 41b recovered (15%).

After having gained access to reasonable quantities of dibro-

moolefin 48 and also having implemented a versatile approach

to ethyl phosphonate 39 and several Julia–Kocienski coupling

partners (41a–c), we sought to install the C17ÀC19 portion of

key vinyl iodide 6 and investigate the stereoselective formation

of the C16ÀC17 E-configured alkene.

vent to solely DMF afforded a slight improvement in the Z/E

ratio (5:1) (entry 2). In order to improve the observed selectivi-

ty further we reasoned that the synthesis of sulfones enabling

the stabilisation of the negative charge on the tetrazole unit

would lower the energy of the zwitterionic intermediate reac-

tion pathway more than the non-zwitterionic reaction pathway.

Therefore we employed phenyltetrazole sulfone 41b (entry 3).

This led to a much improved, but still unacceptable Z/E (1:1)

ratio. Further modification by use of sulfone 41c offered no

improvement in terms of selectivity (entry 4) whilst additional

experimentation using 41b by altering the base, solvent and

temperature (entries 5–7) used in the reaction also did not

favour the desired E-isomer.^[39]

d) Initial approach to vinyl iodide 6 (PG=TBS (55)): Prelimi-

nary studies towards the construction of the C16ÀC17 bond

using phosphonate 39 quickly showed this to be an unproduc-

tive route to the fragment. Treatment of phosphonate 39

using LiHMDS, perhaps unsurprisingly only led to the elimina-

tion of the C13ÀOTBS group to provide conjugated diene 54

(Scheme 11). In order to make the addition of phosphonate 39

to aldehyde 28 competitive with elimination of the C13ÀOTBS

group, 28 and 39 were pre-mixed prior to the slow addition of

LiHMDS. However, this process resulted in recovered phospho-

nate 39 and decomposition of the aldehyde. Use of a weaker

base such as Cs₂CO₃also proved ineffective in producing the

desired product 55.

In an attempt to further improve the E/Z selectivity, we de-

cided to reverse the phosphonate and sulfone coupling part-

ners such that they were derived from the cyclopropane unit

(58 and 60), with the aldehyde prepared from the C10ÀC16

fragment (37) (Scheme 12). However, although the aldehyde

56 could be prepared easily from the corresponding alcohol

(37) using MnO₂, we found that phosphonate 58 (in which R=

Me or Et) could not be isolated. In addition, the preparation of

the sulfone 60 by oxidation of sulfide 59 proved problematic

(mCPBA, oxone, H₂O₂/Mo₇O₂₄(NH₄)₆·4H₂O) and could not be

achieved. These factors, coupled with the poor selectivity ob-

served in the Julia–Kocienski coupling between aldehyde 28

Scheme 11. a) LiHMDS, THF, À788C; proposed decomposition of phospho-

nate 38.

Chem. Eur. J. 2015, 21, 13261 – 13277

www.chemeurj.org

13266 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

The required aldehyde 64 was synthesised from (S)-glycidol in

6 steps. Epoxide opening, followed by carboalumination/iodi-

nation yielded diol 62.^[42]Following protection of the primary

alcohol (PivCl), TBS protection and pivolate reduction, alcohol

63 was produced in 80% yield over three steps.^[43]Swern oxi-

dation^[44]followed by HWE olefination using the literature con-

ditions (LiHMDS, THF, À788C^[8]), yielded the C10ÀC22 vinyl

iodide in good yield as an inseparable 4:1 E/Z mixture. Un-

fortunately attempts to improve this ratio (i.e. alternative bases

and lower temperature) gave no improvement in selectivity.

Enrichment of the isomeric ratio by isomerisation was also at-

tempted (similar to Evans^[5]and MacMillan^[10]); however, this

generally lead to extensive decomposition (including proto-de-

iodination) under a range of conditions employed. Although

this approach was found to be reasonably scalable and provid-

ed significant quantities of vinyl iodide 55 for initial fragment

union studies, focus remained on the development of a route

to provide the vinyl iodide as a single E,E-isomer.

Scheme 12. a) MnO₂, CH₂Cl₂, RT, used directly; b) 1-phenyl-1H-tetrazole-5-

thiol, PPh₃, DIAD, THF, RT, 89% (based on 1-phenyl-1H-tetrazole-5-thiol).

and sulfone(s) 41a–c dictated that we reassess our approach

to vinyl iodide 55.

Cross-coupling approach to vinyl iodide 6 (55)

Formation of the C14ÀC15 bond

After having observed poor selectivity by disconnection of the

C16ÀC17 bond, we chose therefore to disconnect vinyl iodide

55 at both the C15ÀC16 and C17ÀC18 junctions with the idea

of performing sequential cross-coupling reactions (Scheme 14).

Horner–Wadsworth–Emmons (HWE) approach: As our investi-

gations into olefination-based methods to form the C16ÀC17

bond yielded only poor results (low yields, poor selectivity) we

chose to briefly investigate the formation of the C14ÀC15

alkene using

a

Horner–Wadsworth–Emmons olefination

(Scheme 13). This route required the use of phosphonate 66,

common to several of the published callipeltoside synthe-

ses.^[4,9]In a novel approach, phosphonate 66 was synthesised

in two steps from the dibromoolefin, by TBAF-induced elimina-

tion and Stille cross-coupling^[40]using stannane 65^[41]in moder-

ate yield. This reaction was found to be robust and scalable,

and was not optimised to improve the yield of the product 66.

Scheme 14. Cross-coupling approach to vinyl iodide 55 (6).

Such an approach should provide vinyl iodide 55 in a stereo-

specific manner. To facilitate this approach it was necessary to

mask the C10 vinyl iodide functionality as a vinyl silane, there-

by preventing side reactions during the cross-coupling process.

This design feature, as well as the previously noted issues con-

cerning the isolation of aldehyde 33 using the organocatalytic

oxyamination procedure, encouraged us to explore an alter-

nate route to the C10ÀC15 unit.

a) Synthesis of the C10ÀC15 vinyl silane fragment: Assembly

began from commercially available TMS-propyne 72 which, in

the presence of 5 mol% of molybdenum catalyst 71,^[46]could

be regioselectively hydrostannylated to give a single isomer.^[45]

Subsequent iodine-tin exchange gave vinyl iodide 73 in 62%

yield over two steps on a large scale (ca. 45 g).

Scheme 13. a) Lithium acetylide, ethylene diamine complex, HMPA, THF, RT,

50%; b) i) [(Cp)₂ZrCl₂], Me₃Al, 1,2-dichloroethane, RT; ii) I₂, THF, À308C, 60%;

c) PivCl, pyr, 08C; d) TBSOTf, 2,6-lutidine, À788C; e) DIBAL-H, CH₂Cl₂, À788C,

80% over 3 steps; f) oxalyl chloride, DMSO, Et₃N, CH₂Cl₂, À788C, quant.;

g) TBAF, DMF, 658C; h) 65, [Pd₂(dba)₃] (2 mol%), AsPh₃(8 mol%), THF, 608C,

56% over 2 steps; i) LiHMDS, THF, À788C, 80%. dba=dibenzylideneacetone;

Cp=cyclopentadienyl; HMPA =hexamethylphosphoramide; LiHMDS=lithi-

um hexamethyldisilazide.

With this material in hand, we elected to form the corre-

sponding alkenylmetal species, with which opening of an ap-

Chem. Eur. J. 2015, 21, 13261 – 13277

www.chemeurj.org

13267 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

propriate (S)-configured epoxide could be achieved. This reac-

tion was first carried out by addition of the corresponding

higher-order heterocuprate (of 73) to the lithium alkoxide of

(S)-glycidol, affording diol 75 in 72% yield (Table 1, entry 1).^[47]

However, in the first attempt epoxide 61 was used as the limit-

ing reagent, with epoxide ring-opening requiring a large

excess (5.0 equiv) of vinyl iodide 73.^[48]Although substrate 73

could be produced in multigram quantities, we sought to de-

crease further the number of equivalents of 73 such that

greater quantities of 75 could be produced. Unfortunately, re-

ducing the excess of vinyl iodide 73 resulted in lower isolated

yields of diol 75 (Table 2, entries 2 and 3), while reaction with

Following ring opening of PMB-protected epoxide 74, a two

step TBS protection/PMB deprotection afforded alcohol 79.

After some optimisation, aldehyde 80 was synthesised using

Parikh–Doering conditions,^[51]with subsequent Ramirez dibro-

moolefination^[52]providing compound 81 in 89% yield over

two steps.^[53]Finally, Corey–Fuchs reaction^[54]delivered terminal

alkyne 82, with which hydrometallation studies could be con-

ducted (Scheme 15).

Table 2. Conditions for epoxide ring opening.^[a]

Scheme 15. a) TBSOTf, 2,6-lutidine, CH₂Cl₂, À788C, 53% over 2 steps from

73; b) DDQ, pH 7 phosphate buffer, CH₂Cl₂, 08C, 89%; c) SO₃·pyr, DMSO,

Et₃N, CH₂Cl₂, 08C!RT; d) PPh₃, CBr₄, 2,6-lutidine, 08C, 89% over 2 steps;

e) nBuLi, THF, À788C!RT, then H₂O, 85%.

Entry

1

Conditions

Epoxide

[equiv]

Vinyl iodide

73 [equiv]

Yield

[%]

Firstly, hydrozirconation of alkyne 82 using the Schwartz re-

agent ([(Cp)₂Zr(H)Cl])^[55]followed by iodine exchange was in-

vestigated. Unfortunately, this approach provided a mixture of

desired (E)-vinyl iodide 84 as well as protodesilylated com-

pound 85 in a 1:2 ratio (by ¹H NMR spectroscopy). In a different

approach, aldehyde 80 was found to be poorly reactive to-

wards Takai olefination^[56]conditions and only starting material

was isolated or trace conversion to a mixture of 84 and 85.

Pleasingly, regioselective Pd⁰-catalysed hydrostannylation of

alkyne 82^[57]could be achieved which, following the dropwise

addition of iodine, gave the desired (E)-vinyl iodide 84 in 49%

yield over two steps.

thiophene, nBuLi, CuCN,

tBuLi, THF, BF₃·OEt₂,

À78!À90!À408C

as for entry 1

61 (1.0)

5.0

72

2

3

4

5

61 (1.0)

74 (1.0)

3.5

2.0

5.0

1.2

55

33

60^[a]

–

as for entry 1

iPrMgCl, CuI,^[b]

THF, À788C

6

7

tBuLi, BF₃·OEt₂,

PhMe, À788C

74 (1.0)

74 (2.5)

2.0

1.0

11

tBuLi, BF₃·OEt₂,

PhMe, À788C

53^[c]

[a] Reagents and conditions: a) allyl bromide, MeCN, PhH, 858C, 44%;

b) 71 (5 mol%), Bu₃SnH, THF, RT; c) I₂, CH₂Cl₂, 08C, 62% over 2 steps.

[b] Compound could not be purified fully. [c] 25 mol% of CuI was used.

[d] Following TBS protection.

Synthesis of vinyl iodide 84 by this method was advanta-

geous since both (E)-vinyl iodide 84 and (E)-vinyl stannane 83

could be used as coupling partners in the Stille reaction

(Scheme 16). This protocol was used for the synthesis of multi-

PMB-protected (S)-glycidol afforded a mixture of products

(entry 4). Interestingly, lowering the number of equivalents of

vinyl iodide 73 resulted in increasing amounts of byproduct

77. The formation of compound 77 represented a rather un-

usual case whereby a tert-butyl group (presumably from tBuLi)

had instead reacted with the lithium alkoxide of (S)-glycidol.

Although there is literature precendent for the opening of an

epoxide with both tBuLi and its corresponding lower-order ho-

mocuprate,^[49]it is unknown which species is responsible for

the formation of 77 in this instance. Unable to prevent the for-

mation of byproduct 77 (or 78), we turned to the direct open-

ing of epoxide 74 using the organolithium species of 73.^[50]Ad-

justment of the stoichiometry of the reaction indicated that an

excess of PMB-protected epoxide 74 (2.5 equiv) was optimal,

producing the desired product in 53% yield (following TBS

protection) and in multigram quantities (entry 7).

Scheme 16. a) [(Cp)₂Zr(H)Cl], THF, RT, then I₂; b) Bu₃SnH, [Pd(PPh₃)₂Cl₂]

(3 mol%), THF, 08C; c) I₂, CH₂Cl₂, À788C, 49% over 2 steps; d) CrCl₂, THF,

08C, then 80, THF, CHI₃, RT!508C.

Chem. Eur. J. 2015, 21, 13261 – 13277

www.chemeurj.org

13268 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

gram quantities of (E)-vinyl stannane 83 and (E)-vinyl iodide

84, which could be straightforwardly isolated and purified.

agent^[61]with HBF₄·OEt₂at 08C afforded the desired E-isomer

exclusively in an impressive 96% yield (Scheme 17).

At this point attention switched to the synthesis of the alter-

nate (E)-vinyl stannane coupling partner (93, Table 3). The syn-

thesis of vinyl stannane 93 relied on the Stille coupling be-

tween monobromide 86 and bis-stannane 92.^[62]However, de-

spite the investigation of a number of different literature

known Stille conditions using a variety of different catalyst

b) Synthesis of the C16–C22 coupling partners: With knowl-

edge that both (E)-vinyl stannane 83 and (E)-vinyl iodide 84

could be accessed straightforwardly, the synthesis of their re-

spective (E)-vinyl iodide (89) and (E)-vinyl stannane (93)

(shown in Table 3, below) coupling partners began. Initial at-

tempts at generating vinyl iodide 89 focused on a Takai olefi-

nation^[56]of propargylic aldehyde 28 (shown in Table 1). Al-

though this strategy delivered the desired vinyl iodide 89, no

selectivity between the E- and Z-isomers was observed (1:1).

Reassessment of the approach led us to target (E)-vinyl silane

88, since silicon-iodine exchange would be expected to pro-

ceed with retention of configuration (Scheme 17). Construction

systems,

[Pd₂(dba)₃]/AsPh₃

(1:4),^[63]

[Pd(MeCN)₂Cl₂],^[64]

[Pd(PPh₃)₄],^[65][Pd(PFur₃)₂Cl₂],^[66]vinyl stannane 93 could only

be produced in a disappointing 6–10% yield over 2 steps

(from dibromide 48). On each occasion only desired vinyl stan-

nane 93 and bis-stannane (92) could be isolated from the reac-

tion mixture, with substantial amounts of unidentifiable de-

composition products observed. Since no set of conditions

had proven superior for this transformation, Farina’s catalyst

system^[63]([Pd₂(dba)₃], AsPh₃) was chosen as a basis for further

optimisation, with the amount of bis-stannane 92 maintained

at two equivalents throughout (Table 3).

Table 3. Conditions for C17ÀC18 bond formation.^[a]

Entry

[Pd₂(dba)₃]

[mol%]

AsPh₃

[mol%]

Additive

[mol%]

T

[8C]

Yield^[a]

[%]

1

2

3

4

5

6

7

8

2

10

8

40

–

60

25

0

8

6

7

15

32

23

35

45^[b]

Scheme 17. a) TBAF, DMF, 658C; b) [Pd(MeCN)₂Cl₂] (18 mol%), DMF, RT, 23%

over 2 steps; c) nBuLi, Et₂O, À788C then H₂O; d) [(PPh₃)₂PdCl₂] (7 mol%), CuI

(21 mol%), MeCN, RT then Et₃N, 08C, 82% (from 91); e) Py₂IBF₄, HBF₄·OEt₂,

CH₂Cl₂, 08C, 96%.

0

À10

À20

À10

Ag₂CO₃(25)

Ag₂CO₃(100)

of this compound began with the formation of volatile mono-

bromide 86 following elimination using TBAF. Bromoalkyne 86

was then coupled with trimethyl[(E)-2-(tributylstannanyl)ethe-

nyl]silane 87^[58]to give the desired (E)-vinyl silane 88 in poor

yield (23%) (Scheme 17, route A). Since this particular Stille

coupling was low yielding, a second route to enable the syn-

thesis of compound 88 was explored (Scheme 17, route B).

This involved a Sonogashira reaction^[59]between known vola-

tile alkyne 90^[6,7]and vinyl iodide 91; itself derived by iodina-

tion of 87. Reassuringly, this gave a reproducible yield of 82%

and was optimised such that only 2.5 equivalents of alkyne 90

was required.^[60]This proved a much more effective way to

access vinyl silane 88. To deliver the desired vinyl iodide 89, sil-

icon-iodine exchange was initially conducted by treatment of

vinyl silane 88 with either I₂(CH₂Cl₂and THF solutions) or with

NIS. On both occasions the use of I₂resulted in the generation

of complex mixtures, with NIS surprisingly favouring the forma-

tion of the undesired Z-isomer (3:1). However, further experi-

mentation revealed that a combination of the Barleunga re-

[a] Reagents and conditions: a) TBAF, DMF, 658C. [b] Yield over 2 steps.

[c] Yield reproduced on 9.85 mmol scale.

Initial investigations revealed that the reaction time and

temperature had little effect on the isolated yield of (E)-vinyl

stannane 93 (Table 3, entries 1–3) when [Pd₂(dba)₃] and AsPh₃

were used in 2 and 8 mol% quantities. However, the isolated

yield doubled when a five-fold increase in both [Pd₂(dba)₃] and

AsPh₃were administered at 08C (entry 4). Further cooling of

the reaction to À108C proved optimal, producing a much-im-

proved yield of 32% (entry 5). We then chose to add Ag₂CO₃

to the reaction mixture since it could potentially act as both

an acid and halide scavenger.^[67]Pleasingly, addition of one

equivalent (entry 8) of Ag₂CO₃resulted in an acceptable isolat-

ed yield of 45% over two steps, which could be reproduced

on a gram-scale to provide vinyl stannane 93. In order to

assess the effect of alternative additives in the reaction several

different bases (organic and inorganic, Table 4, entries 1 and 2),

Chem. Eur. J. 2015, 21, 13261 – 13277

www.chemeurj.org

13269 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

Union of pyran 5 with vinyl iodide 554 and completion of

the callipeltoside aglycon (4)

Table 4. Further conditions for C17ÀC18 bond formation.

Entry

Additive

Amount [equiv]

Yield [%]^[a]

With routes providing gram-scale quantities of both pyran 5

and vinyl iodide 55 in place, our next major challenge was to

address the diastereoselective union of these fragments to

form the C9 stereocentre. In the first instance, the coupling of

C10ÀC22 vinyl iodide 55 and pyran aldehyde 5 was attempted

in the absence of any chiral additives. Reaction of the corre-

sponding alkenylzinc species of 55 (formed by iodine–lithium

exchange followed by transmetallation with ZnBr₂) with pyran

aldehyde 5 afforded a 1:1 mixture of C9 epimeric products.

This important result was encouraging as it established for the

first time, that a vinyl metal addition of the entire C10ÀC22

fragment to aldehyde 5 could be achieved. Encouraged by this

result, we sought a method that would enable this key cou-

pling to be conducted in a stereocontrolled manner. In order

to achieve this, we investigated the elegant work by Oppolzer

and Radinov,^[68]as well as the studies of Marshall,^[69]who had

previously shown that the stereochemical information present

in an appropriate enantioenriched lithio-N-methylephedrine

alkoxide could be transferred to reactions of this type. Analysis

of the proposed model revealed (1R,2S)-(À)-N-methylephedrine

96 to be the reagent of choice; but in practice, disappointing

diastereoselectivity at C9 was observed (34:66). Undeterred,

we chose to perform this reaction with the enantiomeric

ligand, (1S,2R)-(+)-N-methylephedrine 94, this time observing

a much-improved diastereomeric ratio of 91:9 at the crucial C9

stereogenic centre. Methylation of both diastereomeric mix-

tures resulted in known compound 95; which was previously

described in the MacMillan synthesis of callipeltoside C.^[10]

Comparison with the literature therefore provided rapid deter-

mination of the stereochemical outcome of these reactions

(Scheme 19).

1

2

3

4

5

DIPEA

Cs₂CO₃

AgOAc

Ag₂O

1.0

2.0

25

20

28

36

Ag₂CO₃

[a] Yield over 2 steps. DIPEA=N,N-diisopropylethylamine.

and silver sources (entries 3 and 4) were evaluated, whilst the

stoichiometry of Ag₂CO₃was also doubled (entry 5). Disap-

pointingly, however, the isolated yield was found to be re-

duced in all cases. At this stage the role of the Ag₂CO₃in these

reactions remains unclear, and has not been investigated in

detail.

c) Stereospecific synthesis of (E)-vinyl iodide 55 (6): With

both vinyl iodide 89 and vinyl stannane 93 in hand, their Stille

reactions with respective coupling partners 83 and 84 were ex-

plored Scheme 18. Attempts to construct the C15ÀC16 bond

by a Stille reaction between vinyl stannane 83 and vinyl iodide

89 under several different conditions ([Pd₂(dba)₃]/AsPh₃(1:4),

[Pd(PFur₃)₂Cl₂]) led to either no reaction or only trace conver-

sion to the desired product. As a result of the poor observed

reactivity, we were prompted to investigate the Stille cross-

coupling between reversed coupling partners 84 and 93.

Indeed reaction between vinyl iodide 84 and vinyl stannane

93 could be achieved in the presence of freshly prepared

[Pd(PFur₃)₂Cl₂] (15 mol%).^[66]This gave C10ÀC22 vinyl silane 67

in 63% yield and importantly as the desired single (E,E)-isomer.

Having successfully obtained the complete di-ene-yne unit in

the required configuration, we were pleased to find that the

requisite silicon–iodine exchange occurred with complete re-

tention of configuration (confirmed by no nOe correlation, see

the Supporting Information), delivering the fully elaborated

C10ÀC22 vinyl iodide fragment 55 (6) as a single isomer. This

route provided C10ÀC22 vinyl iodide 55 in a scalable longest

linear sequence of 12 steps (from TMS-propyne (72)] and in

5.7% overall yield.

Pleasingly, the major product resulting from the matched

addition of (1S,2R)-(+)-N-methylephedrine 94 resulted in the

correct C9 stereochemistry for the callipeltosides. This result

was welcome, but gave the opposite result to that predicted

by the model’s suggested by Oppolzer^[68]and Noyori.^[70]A simi-

lar reversal in the selectivity was also observed by Myers^[71]in

his synthesis of the tetracycline antibiotics. In order to account

for this observation, Myers suggested that the aldehyde (in

Myers case 3-benzyloxy-5-isoxazolecarboxaldehyde (98), shown

in Scheme 19) formed a bis-chelate with the active metal com-

plex, resulting in the exposure of the opposite enantiotopic

face to reaction. We therefore speculate, but without evidence,

that the presence of the pyran oxygen atom results in a similar

chelation effect (for comparison, see Scheme 19).^[72]

With the advanced fragment 95 in hand, selective TBS de-

protection was conducted using TBAF, with subsequent sapo-

nification of the ester functionality affording seco-acid 100,

ready for Yamaguchi macrolactonisation.^[13]Cyclisation to give

the desired macrocycle was achieved, but also produced varia-

ble amounts of C3-acetal eliminated product (as observed by

others), which could not be separated by flash column chro-

matography. However, this was inconsequential since treat-

ment of the mixture with TFA in THF/H₂O (5:1) reinstalled the

Scheme 18. a) [Pd(PFur₃)₂Cl₂] (15 mol%), DMF, RT, 63%; b) NIS, MeCN, RT,

84%. NIS=N-iodosuccinimide.

Chem. Eur. J. 2015, 21, 13261 – 13277

www.chemeurj.org

13270 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

termediate from which both of these sugars could be derived.

In a similar manner, callipeltose C was to be accessed from C4’-

epimeric compound 105, instead containing a hydroxyl group

at this position. Pyranone 106 was thereafter considered to be

an appropriate common building block for the preparation of

104 and 105 (Scheme 21).

Scheme 19. a) i) 55, tBuLi (2.8 equiv), Et₂O, À788C; ii) ZnBr₂(0.9m in Et₂O,

1.35 equiv), 08C; iii) 94 (or 96) (1.1 equiv), PhMe, 08C; iv) 5, PhMe, 08C, 48%;

b) MeOTf, 2,6-di-tert-butylpyridine, CH₂Cl₂, RT, 73%.

Scheme 21. Retrosynthesis of callipeltose A, B and C.

hemi-ketal functionality and removed the C5-TBS group in

a one-pot process to deliver the callipeltoside aglycon 4 in

58% over two steps (Scheme 20).

a) Synthesis of common callipeltose A and B precursor 110:

Pyranone 106 was readily obtained from commercially avail-

able 3,4-di-O-acetyl-6-deoxy-l-glucal 107, following deprotec-

tion using polymer supported Na₂CO₃, and allylic oxidation in

74% over two steps without column chromatography required

(Scheme 22). In keeping with Nicolaou’s synthesis of d-callipel-

tose A,^[73]we chose to activate the C4’ hydroxyl as its nosylate,

and invert the stereocentre by displacement with nBu₄NN₃.

This gave pyranone 108 as a single diastereoisomer (>95:5 by

¹H NMR spectroscopy). Following this, addition of MeLi at

À1008C occurred with complete diastereoselectivity to provide

desired compound 104 in excellent yield (79%). With the cor-

rect C3’ and C4’ stereochemistry in place, the remaining C1’

Scheme 20. a) TBAF, THF, RT, 74%; b) Ba(OH)₂·8H₂O, MeOH, RT, quant.;

c) 2,4,6-trichlorobenzoyl chloride, Et₃N, PhMe, RT, then added to DMAP,

PhMe, 808C; d) TFA, THF/H₂O (5:1), RT, 58% over 2 steps. TFA=trifluoroace-

tic acid.

Synthesis of the callipeltoside sugar fragments

Given the similarity of these fragments, we considered the pos-

sibility of beginning our studies from a common, readily avail-

able precursor. Since callipeltose A and B both contained nitro-

gen at the C4’-position, azido sugar 104 was chosen as an in-

Scheme 22. a) PS-Na₂CO₃, MeOH, RT; b) MnO₂, CH₂Cl₂, RT, 74% over 2 steps;

c) NsCl, pyr, CH₂Cl₂, 08C!RT, 95%; d) nBu₄NN₃, CH₂Cl₂, 08C, 72%; e) MeLi,

THF, À1008C, 79%; f) mCPBA, NaHCO₃, MeOH, 08C!RT, 52%; g) KOtBu, THF,

08C then MeI, 08C, 79%. mCPBA=meta-chloroperoxybenzoic acid.

Chem. Eur. J. 2015, 21, 13261 – 13277

www.chemeurj.org

13271 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

and C2’ stereocentres were installed by an epoxidation/metha-

nolysis sequence to provide compound 109. Finally, selective

methylation of the C2’OH was carried out by careful control of

KOtBu stoichiometry (1.05 equiv) and addition of MeI using the

conditions described by Panek^[74]to afford versatile azido sugar

110 in 79% yield. At this point the approach could diverge to

either the callipeltose A or B sugars.

b) Synthesis of the callipeltose A and B thioglycosides: In

order to synthesise the cyclic carbamate scaffold present in cal-

lipeltose A, the azide was reduced under hydrogenation condi-

tions with Pearlman’s catalyst to form the corresponding pri-

mary amine. This was then reacted with triphosgene in pyri-

dine to afford known bicyclic compound 101 in 72% yield.^[4]

Protection then provided callipeltose methoxyacetal 111,

ready for manipulation as an appropriate glycosyl donor

(Scheme 23). Since previous reports had indicated that the cor-

Scheme 24. a) H₂, Pd(OH)₂/C (27 mol%), EtOAc, RT, 87%; b) 113, CHCl₃, RT,

75% over 2 steps; c) TMSSPh, ZnI₂, TBAI, 1,2-dichloroethane, 658C, 60%.

sion, with the NMR spectroscopic signals coalescing between

393 and 413 K (see the Supporting Information).

With formylated methoxyacetal 102 in hand, we attempted

to convert this compound to its corresponding thioglycoside

(114) (Scheme 24). However, for unknown reasons, this could

not be achieved (PhSH, BF₃·OEt₂and TMSSPh, ZnI₂, TBAI^[76]),

with re-isolation of starting material observed. As a result, we

decided that azido sugar 115 would have to be converted to

its thioglycoside and further manipulated to provide the

formyl sugar following attachment to the callipeltoside agly-

con (4). In contrast to formylated sugar 102, azido precursor

115 could be easily converted into the thioglycoside by treat-

ment with a combination of TMSSPh, ZnI₂and TBAI to once

again provide a single anomer.^[76]In doing so, the C3’ hydroxyl

was advantageously protected as its TMS-ether in readiness for

its forthcoming attachment to the callipeltoside aglycon.

Scheme 23. a) H₂, Pd(OH)₂/C (27 mol%), EtOAc, RT, 87%; b) triphosgene, pyr,

CH₂Cl₂, À788C!RT, 72%; c) TIPSCl, 2,6-lutidine, CH₂Cl₂, RT, 97%; d) PhSH,

BF₃·OEt₂, CH₂Cl₂, 08C!RT, 80%.

c) Synthesis of the callipeltose C thioglycoside: As was the

case for callipeltose A and B, work began from common pyra-

none 106. Initial investigations focused on the formation of

the C3’ stereocentre by means of a diastereoselective methyl

addition to pyranone 106. Early attempts at this transformation

were conducted by the addition of MeLi·LiBr at À788C to sub-

strates bearing a protected C4’ hydroxyl (pivolyl, not shown).

This unfortunately resulted in the undesired stereochemistry,

with only the C3’ (R)-diastereoisomer observed. After some in-

vestigation, we found that the selectivity could be completely

reversed by leaving the C4’ hydroxyl group unprotected to

afford diol 116 in good yield (78%) and as a single diastereo-

isomer. As a result, we postulate that this excellent diastereo-

control is the result of a complex-induced proximity effect ex-

erted by the neighbouring a-hydroxyl group (Scheme 25).^[77]

Having set the C3’ stereocentre, our next challenge was to se-

lectively protect the C4’ secondary alcohol in preference to the

tertiary C3’ alcohol functionality.

responding trichloroacetimidate of callipeltose C was unstable

to chromatography,^[10]we chose to convert each callipeltose

sugar to its thioglycoside. Reaction of methoxyacetal 111 with

PhSH and BF₃·OEt₂resulted in thioglycoside 112 as a single

anomer. The relative and absolute stereochemistry of this com-

pound was confirmed by X-ray crystallography (Scheme 23).

In an identical fashion to callipeltose A, azide 110 was re-

duced to the primary amine and the formyl group then instal-

led. Since we were concerned by the sensitivity of the methox-

yacetal to acidic conditions, we sought a mild, acid-free proce-

dure to install the requisite formyl group. Kisfaludy^[75]had pre-

viously shown that pentafluorophenyl formate (113) was

a mild and selective reagent for the formylation of amines,

with no reaction observed in the presence of alcohol function-

ality. Pleasingly, application of this reagent delivered callipelto-

se B methoxyacetal 102 in good yield (75% over two steps)

and as a 3.2:1 rotameric mixture (Scheme 24). This was in

keeping with Minale’s original isolation paper,^[1b]since it is

mentioned that callipeltoside B exists as a mixture of ‘two in-

separable conformers’. The resulting variable temperature

¹H NMR spectroscopic studies conducted on callipeltose B me-

thoxyacetal 102 indicated a very high barrier to interconver-

Although a selective protection was ultimately not realised,

the unwanted mono- or bis-silylated material could be easily

separated from the required C4’-protected product and recy-

cled to reclaim the original diol.^[78]

Following protection of the C4’-hydroxyl, epoxidation with

concomitant methanolysis provided product 117. Methylation

and thioglycoside formation using conditions analogous to

Chem. Eur. J. 2015, 21, 13261 – 13277

www.chemeurj.org

13272 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

Completion of callipeltoside C

Glycosidation reaction between bis-protected callipeltose C

sugar 119 and the callipeltoside aglycon (4) provided the de-

sired product (120) in good yield (80%; Scheme 27). However,

deprotection of bis-protected callipeltoside C (120) proved

problematic, with the TBS group being resilient to a number of

Scheme 25. a) MeLi·LiBr, Et₂O, À788C, 78%; b) TBSOTf, 2,6-lutidine, DMF,

48%; c) MMPP, NaHCO₃, MeOH, 08C; d) MeI, Ag₂O, DMF, RT, 80% over

2 steps; e) TMSSPh, ZnI₂, TBAI, 1,2-dichloroethane, 658C, 64%. MMPP=mag-

nesium monoperoxyphthalate.

that described for azido sugar 110 successfully delivered TBS-

protected callipeltose C thioglycoside 119. With sufficient

quantities of the callipeltoside aglycon and all three callipelto-

side sugars in hand efforts began to assemble all three natural

products.

Completion of callipeltoside A

The callipeltoside aglycon was successfully coupled with thio-

glycoside donor 112 using the conditions described by Evans

in his synthesis of callipeltoside A.^[5d]Pleasingly, treatment of

the TIPS-protected material with TBAF then afforded callipelto-

side A in 83% over two steps (Scheme 26). This synthetic mate-

Scheme 27. a) 4, 119, 4 MS, CH₂Cl₂, DTBMP, RT, then À158C, NIS, TfOH,

À158C!RT, 80%.

different deprotection conditions (TBAF, HF·pyr, TASF,^[79]THF/

HCO₂H/H₂O (6:3:1)]. Although the TMS protecting group could

be readily cleaved under each set of conditions, in most cases

the TBS-ether remained intact (121), resulting in either the re-

isolation of TBS-protected material 120 or products whereby

unidentified side reactions had occurred. While the number of

deprotection conditions attempted was not extensive, the de-

cision was made to revisit the preparation of callipeltose C, this

time protecting the C4’ hydroxyl moiety as its TES-ether.

In the absence of the callipeltoside aglycon, thioglycoside

119 could be easily deprotected using TBAF to provide diol

122 in 89% yield. The secondary alcohol was then selectively

protected as its TES-ether and the TMS-group installed in an

efficient two-step process (Scheme 28).

Scheme 26. a) 4, 112, 4 MS, CH₂Cl₂, DTBMP, RT, then À158C, NIS, TfOH,

À158C!RT; b) TBAF, THF, 83% over 2 steps. DTBMP=2,6-di-tert-butyl-4-

methylpyridine.

With TES-protected thioglycoside 123 in hand, the assembly

of callipeltoside C was revisited. Once again the glycosidation

reaction proceeded without incident, to afford the bis-protect-

ed substrate, ready for deprotection. On this occasion, treat-

ment with TASF resulted in the removal of both protecting

groups to provide callipeltoside C (3) (Scheme 29) in 57% yield

(2 steps).

1

rial was found to match both the H and ¹³C NMR spectra pro-

vided for the natural isolate, whilst also exhibiting near-identi-

cal optical rotation ([a]²_D^3:5=À17.5 (c=0.33 in MeOH) com-

pared to [a]²_D^5:0=À17.6 (c=0.04 in MeOH) for the natural prod-

uct). Since the structure of callipeltoside A had been rigorously

determined following the initial efforts of Trost,^[4]Evans^[5]and

Paterson,^[7]it is assumed that the stereochemistry of the glyco-

syl linkage is that depicted in Scheme 26.

1

This material was identical to both the H NMR spectra of

the natural isolate and the synthetic material disclosed by Mac-

Millan. However, the small amounts of natural product isolated

meant that the ¹³C NMR spectra disclosed by Minale had been

extrapolated from the HMQC spectrum, and so prevented fur-

ther accurate ¹³C NMR spectroscopic comparison. In addition

Chem. Eur. J. 2015, 21, 13261 – 13277

www.chemeurj.org

13273 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

Scheme 28. a) TBAF, THF, RT, 89%; b) TESCl, pyr, DMAP, RT, 85%; c) TMSOTf,

2,6-lutidine, CH₂Cl₂, À788C, 85%.

Figure 2. Key NOESY correlation observed for callipeltoside C.

mation). Therefore, based on this finding, we have tentatively

assigned the glycosidic linkage of callipeltoside C to be that

depicted in Figure 2. This is in full agreement with the stereo-

chemical assignment by MacMillan.^[10]

Scheme 29. a) 4, 123, 4 MS, CH₂Cl₂, DTBMP, RT, then À158C, NIS, TfOH,

À158C!RT; b) TASF, DMF, 408C, 57% over 2 steps.

After having completed the synthesis of callipeltoside C by

means of a protecting group change (from TBS to TES), we de-

cided it was appropriate to also complete the forward synthe-

sis of TES-protected callipeltose C sugar 123 from common

pyranone 106. This was of course performed in an analogous

manner to that described for the TBS-protected thioglycoside

119 (Scheme 25). However, to our delight, in this instance se-

lective protection of the C4’ hydroxyl could be easily achieved

using TESCl in combination with pyridine and DMAP

(Scheme 30).

to this, no optical rotation was recorded for the natural prod-

uct. Therefore, our structural assignment of callipeltoside C^[80]

1

rests only on the comparison of the recorded H NMR spectra

with the natural isolate.

Analysis the glycosidic linkage of callipeltoside C

In order to complete the stereochemical assignment of callipel-

toside C, we attempted to determine the configuration of the

1

glycosidic linkage by analysis of the J_C-Hcoupling constant

and NOESY data.

a) ¹J_C-Hcoupling constant: Early empirical observations have

1

shown that measurement of the J_C-Hcoupling constant de-

rived from a HSQC (Heteronuclear Single Quantum Coherence)

experiment without ¹³C decoupling provides an indication of

whether the proton at the anomeric centre of a sugar moiety

is axial or equatorial. This method has been shown to be

a useful technique for the assignment of the configuration of

the glycosidic linkage for a wide variety of carbohydrates. A

value of ~170 Hz typically suggests an equatorial proton at

C1’H, whilst ~160 Hz indicates an axial proton.^[81]Unfortunate-

ly, despite the good literature precedent for this technique,

Scheme 30. a) MeLi·LiBr, Et₂O, À788C, 78%; b) TESCl, pyr, DMAP, RT;

c) mCPBA, NaHCO₃, MeOH, 08C!RT, 45% over 2 steps; d) KOtBu, THF, 08C,

then MeI, 08C, 81%; e) TMSSPh, ZnI₂, TBAI, 1,2-dichloroethane, 658C, 86%.

1

measurement of the J_C-Hcoupling constant in this system

Structural elucidation and completion of callipeltoside B

1

gave a value of J_C-H=166.5 Hz, therefore providing an incon-

clusive result. Undeterred, we therefore chose to analyse the

NOESY spectra of callipeltoside C in the hope that one confor-

mation (assumed chair) of callipeltose C would be favoured

over others.

With the syntheses of both callipeltosides A and C achieved,

our attention switched to callipeltoside B in order to complete

the series. As previously mentioned, there had been no prior

total synthesis of callipeltoside B and therefore it was assumed

that it also contained an l-configured sugar. As expected, at-

tachment of 115 to the aglycon proceeded without incident,

delivering 126 in moderate yield (56%) (Scheme 31). At this

point we faced the rather daunting prospect of having to

reduce the azide moiety to the corresponding amine in the

b) NOESY data: As a result, we assessed all possible chair con-

formations (Figure 2a–d). Analysis of the averaged data re-

vealed that structure (a) was the only conformer that account-

ed for all observed nOe interactions (see the Supporting Infor-

Chem. Eur. J. 2015, 21, 13261 – 13277

www.chemeurj.org

13274 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

Scheme 31. a) 4, 115, 4 MS, CH₂Cl₂, DTBMP, RT, then À158C, NIS, TfOH, À158C!RT, 56%; b) 1,3-propanedithiol, Et₃N, pyr/H₂O (10:1), RT; c) 113, CHCl₃, RT;

1

d) TASF, DMF, 408C, 52% over 3 steps, 4:1 rotameric mixture by H NMR spectroscopy; e) 4, ent-115, 4 MS, CH₂Cl₂, DTBMP, RT, then À158C, NIS, TfOH,

1

À158C!RT, 41%; f) 1,3-propanedithiol, Et₃N, pyr/H₂O (10:1), RT; g) 113, CHCl₃, RT; h) TASF, DMF, 408C, 57% over 3 steps, 4:1 rotameric mixture by H NMR

spectroscopy.

presence of multiple unsaturated functional groups. As antici-

pated, a hydrogenation-based approach was found to be un-

successful, and therefore we expected that the Staudinger re-

action^[82]would be an ideal method in order to achieve this

transformation. However, in practice poor isolated yields and

mixtures of unidentified products were obtained. Undeterred,

we attempted the reduction using a combination of 1,3-pro-

panedithiol and Et₃N in aqueous pyridine.^[83]To our delight this

afforded the requisite amine, which was immediately formylat-

ed using the aforementioned methodology developed by Kis-

faludy (Scheme 24).^[75]Final treatment of this material with

TASF to remove the TMS protecting group then gave callipel-

Although it was clear that synthetic callipeltoside B had

been synthesised, the configuration of the glycosidic linkage

still needed to be ascertained. Our previous strategy of infer-

ring the linkage configuration by NOESY experiments was

again employed (Figure 3).

Once again, the NOESY spectra revealed that a single chair

conformation (d) accounted for all observed correlations, sug-

gesting that the glycosidic linkage is that depicted in Figure 3.

It should be noted that whilst the configuration of the glycosi-

dic linkage of callipeltoside B is identical to callipeltoside A, it

is the opposite of callipeltoside C. Since exactly the same gly-

cosidation conditions were used to attach each callipeltose

sugar, the stereochemical course of the reaction must be influ-

enced by the C4’ substituent present on the sugar moiety;

however, additional studies to further study this effect have

not been conducted.

1

toside B (2) as a mixture of conformers (4:1 by H NMR spec-

1

troscopy) for the first time. The H NMR spectrum of the syn-

thetic material was found to be in complete agreement to that

disclosed by Minale and his isolation team, and therefore pro-

vided evidence to validate our assumption that the attached

sugar unit was also l-configured. However, since the structural

determination of these molecules rests on only a single

¹H NMR spectrum, we chose to synthesise the corresponding

d-configured callipeltoside B sugar (ent-115, achieved in identi-

cal fashion to that shown in Scheme 22 and Scheme 24) and

also attach this to the callipeltoside aglycon. Further elabora-

tion would then allow direct comparison of ¹H NMR spectra

and provide additional evidence relating to the structure of

callipeltoside B. This was straightforwardly achieved using the

glycosidation, reduction, formylation and deprotection se-

quence described previously to deliver 128. Comparison of the

1

resulting H NMR spectrum revealed significant deviations from

structure 2, and therefore we confidently assign callipeltoside B

to be that depicted in Scheme 31.

Figure 3. Key NOESY correlation observed for callipeltoside B.

Chem. Eur. J. 2015, 21, 13261 – 13277

www.chemeurj.org

13275 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

[7] a) I. Paterson, R. D. M. Davies, A. C. Heimann, R. Marquez, A. Meyer, Org.

Lett. 2003 , 5 , 447 7 –4 480; b) I. Paterson, A. D. Findlay, Pure Appl. Chem.

2008, 80, 1773–1782.

Conclusion

The synthesis of the entire callipeltoside family of natural prod-

ucts has been disclosed. At the beginning of our study we

committed to the ambitious union of pyran aldehyde 5 and

vinyl iodide 6 (55 with a TBS protecting group) by a diastereo-

selective alkenylmetal addition with subsequent Yamaguchi

macrocyclisation to complete the common callipeltoside agly-

con. Attachment of callipeltoses A, B and C at a late stage then

led to the synthesis of each natural product. Although we

never deviated from these primary disconnections, the efficient

synthesis of various key fragments provided a significant chal-

lenge and led us to re-evaluate our synthesis on the basis of

selectivity, practicality and scale-up on a number of occasions.

Still, perseverance, endeavour and determination have resulted

in the successful completion of this research programme. We

hope that this full account adequately highlights the trials and

tribulations often encountered when designing, carrying out

and completing a complex natural product synthesis.

[8] H. Huang, J. S. Panek, Org. Lett. 2004 , 6 , 438 3 –4 385.

[9] a) T. R. Hoye, H. Zhao, Org. Lett. 1999 , 1 , 16 9 –1 72; b) T. R. Hoye, M. E.

Danielson, A. E. May, H. Zhao, Angew. Chem. Int. Ed. 2008 , 47 , 974 3 –

9746; Angew. Chem. 2008 , 120 , 988 9 –9 892; c) T. R. Hoye, M. Danielson,

A. E. May, H. Zhao, J. Org. Chem. 2010 , 75 , 705 2 –7 060.

[10] J. Carpenter, A. B. Northrup, D. Chung, J. J. M. Wiener, S. Kim, D. W. C.

MacMillan, Angew. Chem. Int. Ed. 2008 , 47 , 356 8 –3 57 2; Angew. Chem.

2008 , 120 , 362 4 –3 628.

[11] J. R. Frost, C. M. Pearson, T. N. Snaddon, R. A. Booth, S. V. Ley, Angew.

Chem. Int. Ed. 2012 , 51 , 936 6 –9 371; Angew. Chem. 2012 , 124 , 950 0 –

9505.

[12] For the syntheses of fragments of the callipeltosides, see: a) G. R. Smith,

J. J. Finley, R. M. Giuliano, Carbohydr. Res. 1998 , 308 , 22 3 –2 27; b) M. K.

Gurjar, R. Reddy, Carbohydr. Lett. 1998, 3, 169–172; c) F. Velµzquez, H. F.

Olivo, Org. Lett. 2000 , 2 , 193 1 –1 933; d) H. F. Olivo, F. Velazquez, H. C.

Trevisan, Org. Lett. 2000 , 2 , 405 5 –4 05 8; e) M. Romero-Ortega, D. A.

Colby, H. F. Olivo, T e trahedron Lett. 2002 , 43 , 643 9 –6 441; f) A. Tóth, J.

Remenyik, I. Bajza, A. Liptak, ARKIVOC (Gainesville, FL, U.S.) 2003, 28–45;

g) J. S. Yadav, A. Haldar, T. Maity, Eur. J. Org. Chem. 2012 , 206 2 –2 071.

[13] J. Inanaga, K. Hirata, H. Sacki, T. Katsuki, M. Yamaguchi, Bull. Chem. Soc.

Jpn. 1979 , 52 , 198 9 –1 993.

[14] a) C. D. Papageorgiou, S. V. Ley, M. J. Gaunt, Angew. Chem. Int. Ed. 2003 ,

42 , 82 8 –8 31; Angew. Chem. 2003 , 115 , 85 2 –8 55; b) N. Bremeyer, S. C.

Smith, S. V. Ley, M. J. Gaunt, Angew. Chem. Int. Ed. 2004 , 43 , 268 1 –2 684;

Angew. Chem. 2004 , 116 , 273 5 –2 738; c) C. D. Papageorgiou, M. A. Cubil-

lo de Dios, S. V. Ley, M. J. Gaunt, Angew. Chem. Int. Ed. 2004 , 43 , 464 1 –

4644; Angew. Chem. 2004 , 116 , 474 1 –4 744; d) C. C. C. Johansson, N.

Bremeyer, S. V. Ley, D. R. Owen, S. C. Smith, M. J. Gaunt, Angew. Chem.

Int. Ed. 2006 , 45 , 602 4 –6 028; Angew. Chem. 2006 , 118 , 617 0 –6 175.

[15] D. A. Longbottom, V. Franckevicius, S. Kumarn, A. J. Oelke, V. Wascho-

lowski, S. V. Ley, Aldrichimica Acta 2008, 41, 3–11.

Experimental Section

A complete description of all experimental procedures and charac-

terisation data relating to compounds described in this manuscript

can be found in the Supporting Information. Additional data relat-

ed to this publication is available at the University of Cambridge

institutional data repository (https://www.repository.cam.ac.uk/

handle/1810/248743).

[16] a) M. J. Gaunt, H. F. Sneddon, P. R. Hewitt, P. Orsini, D. F. Hook, S. V. Ley,

Org. Biomol. Chem. 2003 , 1 , 1 5 –1 6; b) H. F. Sneddon, M. J. Gaunt, S. V.

Ley, Org. Lett. 2003 , 5 , 114 7 –1 15 0; c) H. F. Sneddon, A. van den Heuvel,

D. M. Shaw, M. J. Gaunt, A. Hirsh, R. A. Booth, S. V. Ley, J. Org. Chem.

2006 , 71 , 271 5 –2 725; d) B. C. Ranu, S. Bhar, R. Chakraborti, J. Org.

Chem. 1992 , 57 , 734 9 –7 352.

[17] a) J. A. Marshall, H. R. Chobanian, J. Org. Chem. 2000 , 65 , 835 7 –8 360;

b) J. A. Marshall, G. M. Schaaf, J. Org. Chem. 2001 , 66 , 782 5 –7 831.

[18] The primary TBS group was reinstalled prior to Luche reduction.

[19] a) J. L. Luche, J. Am. Chem. Soc. 1978 , 100 , 222 6 –2 227; b) A. L. Gemal,

J. L. Luche, J. Am. Chem. Soc. 1981 , 103 , 545 4 –5 459.

Acknowledgements

We thank Novartis for a research studentship (J.R.F) and also

gratefully acknowledge the EPSRC (Award numbers: EP/

F06985/1; EP/K009494/1; EP/K039520/1) for financial support

(C.M.P., T.N.S., R.A.B., J.G. and D.M.S).

[20] A. DiØguez-Vµzquez, C. C. Tzschucke, J. Crecente-Campo, S. McGrath,

S. V. Ley, Eur. J. Org. Chem. 2009 , 169 8 –1 706.

[21] a) H. C. Brown, K. S. Bhat, J. Am. Chem. Soc. 1986 , 108 , 591 9 –5 923;

b) H. C. Brown, K. S. Bhat, R. S. Randad, J. Org. Chem. 1989 , 54 , 157 0 –

1576; c) A. Vulpetti, M. Gardner, C. Gennari, A. Bernardi, J. M. Goodman,

I. Paterson, J. Org. Chem. 1993 , 58 , 171 1 –1 718.

[22] a) W. R. Roush, A. D. Palkowitz, M. J. Palmer, J. Org. Chem. 1987 , 52 ,

31 6 –3 18; b) W. R. Roush, K. Ando, D. B. Powers, A. D. Palkowitz, R. L.

Halterman, J. Am. Chem. Soc. 1990 , 112 , 633 9 –6 34 8; c) W. R. Roush,

A. D. Palkowitz, K. Ando, J. Am. Chem. Soc. 1990 , 112 , 634 8 –6 359.

[23] The minor diastereoisomer was difficult to separate and therefore was

progressed to the next step of the synthesis.

Keywords: callipeltosides · cross-coupling · glycosidation ·

gold catalysis · organocatalysis

[1] a) A. Zampella, M. V. D’Auria, L. Minale, C. Debitus, C. Roussakis, J. Am.

Chem. Soc. 1996 , 118 , 1108 5 –1 108 8; b) A. Zampella, M. V. D’Auria, L.

Minale, Tetrahedron 1997 , 53 , 324 3 –3 248.

[2] a) C. K. Skepper, J. B. MacMillan, G. Zhou, M. N. Masuno, T. F. Molinski, J.

Am. Chem. Soc. 2007 , 129 , 415 0 –4 151; b) J. B. MacMillan, G. Zhou, C. K.

Skepper, T. F. Molinski, J. Org. Chem. 2008 , 73 , 369 9 –3 706; c) D. S. Dali-

say, T. F. Molinski, J. Nat. Prod. 2010 , 73 , 67 9 –6 82.

[3] D. S. Dalisay, B. I. Morinaka, C. K. Skepper, T. F. Molinski, J. Am. Chem.

Soc. 2009 , 131 , 755 2 –7 553.

[24] This reaction was attempted with a variety of different Bronsted acid

and bases with either no reaction or decomposition observed.

[25] Previous syntheses of the callipeltosides (refs. [4–10]) rely on disconnec-

tion of the C14ÀC15 bond.

[4] For deschlorocallipeltoside: a) B. M. Trost, J. L. Gunzner, J. Am. Chem.

Soc. 2001 , 123 , 944 9 –9 45 0; b) B. M. Trost, O. Dirat, J. L. Gunzner, Angew.

Chem. Int. Ed. 2002 , 41 , 84 1 –8 43; Angew. Chem. 2002 , 114 , 86 9 –8 71;

c) B. M. Trost, J. L. Gunzner, O. Dirat, Y. H. Rhee, J. Am. Chem. Soc. 2002 ,

124 , 1039 6 –1 0415.

[5] a) D. Evans, J. D. Burch, Org. Lett. 2001 , 3 , 50 3 –5 05; b) D. A. Evans, E.

Hu, J. S. Tedrow, Org. Lett. 2001 , 3 , 313 3 –3 13 6; c) D. A. Evans, E. Hu,

J. D. Burch, G. Jaeschke, J. Am. Chem. Soc. 2002 , 124 , 565 4 –5 655;

d) D. A. Evans, J. D. Burch, E. Hu, G. Jaeschke, Tetrahedron 2008 , 64 ,

467 1 –4 699.

[26] a) L. Horner, H. Hoffmann, H. G. Wippel, Chem. Ber. 1958 , 91 , 6 1 –6 3;

b) W. S. Wadsworth Jr., W. D. Emmons, J. Am. Chem. Soc. 1961 , 83 ,

173 3 –1 738; c) D. H. Wadsworth, I. O. E. Schupp, E. J. Sous, J. J. A. Ford,

J. Org. Chem. 1965 , 30 , 68 0 –6 85.

[27] a) P. J. Kocienski, B. Lythgoe, S. Ruston, J. Chem. Soc. Perkin Trans.

1 1978 , 82 9 –8 34; b) P. J. Kocienski, B. Lythgoe, D. A. Roberts, J. Chem.

Soc. Perkin Trans. 1 1978 , 83 4 –8 37; c) P. J. Kocienski, B. Lythgoe, S.

Ruston, J. Chem. Soc. Perkin Trans. 1 1979 , 129 0 –1 293; d) P. J. Kocienski,

B. Lythgoe, I. Waterhouse, J. Chem. Soc. Perkin Trans. 1 1980 , 104 5 –

1050.

[6] I. Paterson, R. D. M. Davies, R. Marquez, Angew. Chem. Angew Chem.

2001 , 113 , 62 3 –6 27; Angew. Chem. Int. Ed. 2001 , 40 , 60 3 –6 07; Angew.

Chem. 2001, 113, 623–627.

Chem. Eur. J. 2015, 21, 13261 – 13277

www.chemeurj.org

13276 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

[28] a) G. Zhong, Angew. Chem. Int. Ed. 2003 , 42 , 424 7 –4 250; Angew. Chem.

2003 , 115 , 437 9 –4 382; b) S. P. Brown, M. P. Brochu, C. J. Sinz, D. W. C.

MacMillan, J. Am. Chem. Soc. 2003 , 125 , 1080 8 –1 0809.

[29] R. Appel, Angew. Chem. Int. Ed. Engl. 1975 , 14 , 80 1 –8 11; Angew. Chem.

1975 , 87 , 86 3 –8 74.

[30] a) A. Michaelis, R. Kaehne, Ber. Dtsch. Chem. Ges. 1898 , 31 , 104 8 –1 055;

b) A. J. Arbuzov, J. Russ. Phys. Chem. Soc.1906, 38, 687.

[31] a) O. Mitsunobu, Y. Yamada, Bull. Chem. Soc. Jpn. 1967 , 40 , 238 0 –2 382;

b) O. Mitsunobu, Synthesis 1981 , 1 –2 8.

[57] a) N. D. Smith, J. Mancuso, M. Lautens, Chem. Rev. 2000 , 100 , 325 7 –

3282; b) S. Tang, Z. Xu, T. Ye, Tetrahedro n : A symmetry 2009 , 20 , 202 7 –

2032.

[58] P. Prinz, A. Lansky, B. Knieriem, A. de Meijere, T. Haumann, R. Boese, M.

Noltemeyer, Angew. Chem. Int. Ed. Engl. 1997 , 36 , 128 9 –1 29 2; Angew.

Chem. 1997 , 109 , 134 3 –1 346.

[59] K. Sonogashira, J. Organomet. Chem. 2002 , 653 , 4 6 –4 9.

[60] S. Birkett, D. Ganame, B. C. Hawkins, S. Meiries, T. Quach, M. A. Rizzaca-

sa, Org. Lett. 2011 , 13 , 196 4 –1 967.

[32] H. Hunsdiecker, C. Hunsdiecker, Chem. Ber. 1942 , 75 , 29 1 –2 97.

[33] D. H. R. Barton, D. Crich, W. B. Motherwell, Tetrahedron Lett. 1983 , 24 ,

497 9 –4 982.

[61] a) J. Barluenga, L. J. Alvarez-García, J. M. Gonzµlez, Tetrahedron Lett.

1995 , 36 , 215 3 –2 156; b) J. Barluenga, Pure Appl. Chem. 1999, 71, 431–

436.

[34] A. Baeyer, V. Villiger, Ber. Dtsch. Chem. Ges. 1899 , 32 , 362 5 –3 633.

[35] The phenol was removed by evaporation at high temperature and low

pressure.

[36] The analogous route using a Weinreb amide could have been investi-

gated but wasn’t pursued due to the several other problems/issues as-

sociated with this sequence.

[62] A. F. Renaldo, J. W. Labadie, J. K. Stille, Org. Synth. Coll. 1993, 8, 268.

[63] V. Farina, B. Krishnan, J. Am. Chem. Soc. 1991 , 113 , 958 5 –9 595.

[64] F. Suzenet, J-L. Parrain, J.-P. Quintard, Eur. J. Org. Chem. 1999 , 295 7 –

2963.

[65] S. Brückner, E. Abraham, P. Klotz, J. Suffert, Org. Lett. 2002 , 4 , 339 1 –

3393.

[37] a) A. B. Charette, H. Juteau, J. Am. Chem. Soc. 1994 , 116 , 265 1 –2 652;

b) A. B. Charette, S. Prescott, C. Brochu, J. Org. Chem. 1995 , 60 , 108 1 –

1083; c) for DFT computational studies: T. Wang, Y. Liang, Z.-X. Yu, J.

Am. Chem. Soc. 2011 , 133 , 934 3 –9 353.

[38] Although previous reports by Charette had indicated that the cyclopro-

panation reaction resulted in ‘violent explosions’ on a larger scale

(8 mmol of substrate), we found that we could perform the transforma-

tion regularly on a scale exceeding 8 mmol, without problems provid-

ing that the internal temperature did not exceed 08C during formation

of the zinc carbenoid species.

[66] a) C. M. Hettrick, W. J. Scott, J. Am. Chem. Soc. 1991 , 113 , 490 3 –4 910;

b) S. V. Ley, M. N. Tackett, M. L. Maddess, J. C. Anderson, P. E. Brennan,

M. W. Cappi, J. P. Heer, C. Helgen, M. Kori, C. Kouklovsky, S. P. Marsden,

J. Norman, D. P. Osborn, M. . Palomero, J. B. J. Pavey, C. Pinel, L. A. Rob-

inson, J. Schnaubelt, J. S. Scott, C. D. Spilling, H. Watanabe, K. E.

Wesson, M. C. Willis, Chem. Eur. J. 2009 , 15 , 287 4 –2 914.

[67] J-M. Weibel, A. Blanc, P. Pale, Chem. Rev. 2008 , 108 , 314 9 –3 173.

[68] W. Oppolzer, R. N. Radinov, Tetrahedron Lett. 1991 , 32 , 577 7 –5 780.

[69] a) J. A. Marshall, P. Eidam, Org. Lett. 2004 , 6 , 44 5 –4 48; b) J. A. Marshall,

P. M. Eidam, Org. Lett. 2008 , 10 , 9 3 –9 6; c) For a further application in

natural product synthesis see: M. E. Layton, C. A. Morales, M. D. Shair, J.

Am. Chem. Soc. 2002 , 124 , 77 3 –7 75.

[70] For a comprehensive review, see: a) L. Pu, H.-B. Yu, Chem. Rev. 2001 ,

101 , 75 7 –8 24; b) R. Noyori, S. Suga, K. Kawai, S. Okada, M. Kitamura, N.

Oguni, M. Hayashi, T. Kaneko, Y. Matsuda, J. Organomet. Chem. 1990 ,

382 , 1 9 –3 7.

[71] J. D. Brubaker, A. G. Myers, Org. Lett. 2007 , 9 , 352 3 –3 525.

[72] The MacMillan group also invoke a pyran oxygen chelation argument

to explain the diastereoselectivity observed at C9 in their synthesis of

callipeltoside C (ref. [10]).

[39] C. A¨ıssa, Eur. J. Org. Chem. 2009 , 183 1 –1 844.

[40] a) D. Azarian, S. S. Dua, C. Eaborn, D. R. M. Walton, J. Organomet. Chem.

1976 , 117 , C5 5 –C 57; b) M. Kosugi, K. Sasazawa, Y. Shimizu, T. Migita,

Chem. Lett. 1977 , 30 1 –3 02; c) D. Milstein, J. K. Stille, J. Am. Chem. Soc.

1978 , 100 , 363 6 –3 638; d) P. Espinet, A. M. Echavarren, Angew. Chem. Int.

Ed. 2004, 43, 4704–4734; Angew. Chem. 2004, 116, 4808–4839.

[41] P. Li, J. Li, F. Arikan, W. Ahlbrecht, M. Dieckmann, D. Menche, J. Am.

Chem. Soc. 2009 , 131 , 1167 8 –1 1679.

[42] J. D. White, S. Jana, J. Org. Chem. 2014 , 79 , 70 0 –7 10.

[43] K. Suenaga, H. Hoshino, T. Yoshii, K. Mori, H. Sone, Y. Bessho, A. Saka-

kura, I. Hayakawa, K. Yamada, H. Kigoshi, Tetrahedron 2006 , 62 , 768 7 –

7698.

[73] A. J. Pihko, K. C. Nicolaou, A. M. P. Koskinen, Tetrahedro n : Asymmetry

2001 , 12 , 93 7 –9 42.

[44] K. Omura, D. Swern, Tetrahedron 1978 , 34 , 165 1 –1 660.

[45] a) K. C. Nicolaou, T. K. Chakraborty, A. D. Piscopio, N. Minowa, P. Bertina-

to, J. Am. Chem. Soc. 1993 , 115 , 441 9 –4 420; b) K. C. Nicolaou, A. D. Pis-

copio, P. Bertinato, T. K. Chakraborty, N. Minowa, K. Koide, Chem. Eur. J.

1995 , 1 , 31 8 –3 33.

[46] a) H. T. Dieck, H. Friedel, J. Organomet. Chem. 1968 , 14 , 37 5 –3 85;

b) H. X. Zhang, F. GuibØ, G. Balavoine, J. Org. Chem. 1990 , 55 , 185 7 –

1867.

[47] R. J. K. Taylor, Organocopper Chemistry: An Overview. In Organocopper

Reagents; (Ed. R. J. K. Taylor), University Press, Oxford, 1994.

[48] T. M. Baker, D. J. Edmonds, D. Hamilton, C. J. O’Brien, D. J. Procter,

Angew. Chem. Int. Ed. 2008 , 47 , 563 1 –5 633; Angew. Chem. 2008 , 120 ,

571 3 –5 715.

[74] H. B. Huang, J. S. Panek, Org. Lett. 2003 , 5 , 199 1 –1 993.

[75] L. Kisfaludy, L. Ötvçs, Synthesis 1987 , 510.

[76] a) R. U. Lemieux, K. B. Hendriks, R. V. Stick, K. James, J. Am. Chem. Soc.

1975 , 97 , 405 6 –4 062; b) S. Hanessian, Y. Guindon, Carbohydr. Res. 1980 ,

86 , C 3 –C 6; c) K. C. Nicolaou, S. P. Seitz, D. P. Papahatjis, J. Am. Chem.

Soc. 1983 , 105 , 243 0 –2 434; d) I. Paterson, T. Paquet, Org. Lett. 2010 , 12 ,

215 8 –2 161.

[77] P. Beak, A. I. Meyers, Acc. Chem. Res. 1986 , 19 , 35 6 –3 63.

[78] Initial attempts focused on installation of aTIPS protecting group at the

C4’ position. However, this could not be achieved and the alcohol func-

tionality was instead protected as the corresponding TBS ether.

[79] 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.

[49] a) S. D. Rychnovsky, G. Griesgraber, S. Zeller, D. J. Skalitzky, J. Org. Chem.

1991 , 56 , 516 1 –5 169; b) J. M. Swarbrick, P. R. J. Gaffney, J. Org. Chem.

2010 , 75 , 437 6 –4 386.

[50] D. Zurwerra, J. Gertsch, K.-H. Altmann, Org. Lett. 2010 , 12 , 230 2 –2 305.

[51] J. R. Parikh, W. v. E. Doering, J. Am. Chem. Soc. 1967 , 89 , 550 5 –5 507.

[52] N. B. Desai, N. McKelvie, F. Ramirez, J. Am. Chem. Soc. 1962 , 84 , 174 5 –

1747.

[80] This is also true of the MacMillan synthesis.

[81] a) N. Muller, D. E. Pritchard, J. Chem. Phys. 1959 , 31 , 76 8 –7 71; b) G. E.

Maciel, J. W. McIver Jr, N. S. Ostlund, J. A. Pople, J. Am. Chem. Soc. 1970 ,

92 , 1 –1 1 ; c) K. Bock, I. Lundt, C. Pedersen, Tetrahedron Lett. 1973 , 14 ,

103 7 –1 040; d) K. Bock, L. Wiebe, Acta Chem. Scand. 1973 , 27 , 267 6 –

2678; e) K. Bock, C. Pedersen, J. Chem. Soc. Perkin Trans. 1 1974 , 29 3 –

299.

[53] The use of 2,6-lutidine was essential to the success of the reaction, pre-

venting protodesilylation and decomposition under the reaction condi-

tions.

[82] H. Staudinger, J. Meyer, Helv. Chim. Acta 1919 , 2 , 63 5 –6 46.

[83] H. Bayley, D. N. Standring, J. R. Knowles, Tetrahedron Lett. 1978 , 19 ,

363 3 –3 634.

[54] E. J. Corey, P. L. Fuchs, T e trahedron Lett. 1972 , 13 , 376 9 –3 772.

[55] J. Schwartz, J. A. Labinger, Angew. Chem. Int. Ed. Engl. 1976 , 15 , 33 3 –

340; Angew. Chem. 1976 , 88 , 40 2 –4 09.

Received: May 13, 2015

[56] K. Takai, K. Nitta, K. Utimoto, J. Am. Chem. Soc. 1986 , 108 , 740 8 –7 410.

Published online on July 31, 2015

Chem. Eur. J. 2015, 21, 13261 – 13277

www.chemeurj.org

13277 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Article Doi

Callipeltosides A, B and C: Total Syntheses and Structural Confirmation

DOI: 10.1002/chem.201501877

Source and publish data:

Authors:

Article abstract of DOI:10.1002/chem.201501877

Full text of DOI:10.1002/chem.201501877

Products guided by the article

R&D Labs maybe for 105859-44-7

Relevant to this article

Hot Product