Solid phase synthesis of globomycin and SF-1902 A5

DOI: 10.1021/jo1025145

Source and publish data:

Journal of Organic Chemistry p. 2132 - 2144 (2011)

Update date:2022-07-30

Topics:

Authors:

Sarabia, Francisco

Chammaa, Samy

Garcia-Ruiz, Cristina

Read Full Text PDF DownLoad Join now for total 90,000,000 free articles

Article abstract of DOI:10.1021/jo1025145

The syntheses of the natural lipocyclodepsipeptide-type antibiotics globomycin and SF-1902 A₅ are reported, utilizing solid phase technology for the construction of the peptidic fragment and a new asymmetric methodology of epoxidation for the p

Full text of DOI:10.1021/jo1025145

ARTICLE

pubs.acs.org/joc

Solid Phase Synthesis of Globomycin and SF-1902 A₅

Francisco Sarabia,* Samy Chammaa, and Cristina García-Ruiz

Department of Organic Chemistry, Faculty of Sciences, University of Malaga, Campus de Teatinos s/n 29071, Malaga, Spain

S Supporting Information

ABSTRACT: The syntheses of the natural lipocyclodepsipep-

tide-type antibiotics globomycin and SF-1902 A₅are reported,

utilizing solid phase technology for the construction of the

peptidic fragment and a new asymmetric methodology of

epoxidation for the preparation of the lipidic chain. The linkage

between both fragments was successfully achieved in solid phase

to complete the syntheses via a macrolactonization reaction

executed prior to the cleavage of the acyclic precursors from the

solid support. These syntheses provide access to the rapid

generation of a library of analogues via modiﬁcation of the aminoacid residues as well as the lipidic chain, thus facilitating the

identiﬁcation of new antibiotics with interesting mechanisms of action based upon the inhibition of the enzyme signal peptidase II.

’ INTRODUCTION

structure of globomycin, established by Haneishi et al. in 1980,¹⁴

was recently conﬁrmed with its ﬁrst total synthesis in 2000 by the

Kogen group.¹⁵This same group has investigated the initial

chemistry in the area with the synthesis of SF-1902 A₅¹⁶and the

preparation of a series of analogues with modiﬁcations of various

constituents of the molecule including the amino acids, lipidic

chain, stereochemistry, and heteroatoms.^12,17This impressive

chemical study led to the generation of a set of 15 analogues

whose biological evaluation allowed the establishment of a

preliminary structure-activity relationship describing the struc-

tural elements essentials for biological activity (Figure 1, part B).

On the basis of these chemical precedents and prompted by

the potential that these compounds may represent in the ﬁeld of

antibiotics, as well as their unique mechanism of action, we

decided to embark on a research program directed to the design

of an eﬃcient and readily accessible route toward this class of

natural products and analogues. With this purpose in mind, we

initially planned the synthesis of the more relevant natural

members, globomycin (1) and SF-1902 A₅(2), utilizing solid

phase synthesis as a suitable technology to reach our goals. To

this aim, resins 7 and 8 would contain the acyclic precursors of 1

and 2, which would be prepared via a macrolactonization

reaction after the cleavage of such precursors from the resin.

Functionalized resins 7 and 8 would be prepared via coupling of

the acids 9 or 10 with the peptidic chain linked in a solid support

in the form of resin 11 (Scheme 1).

The lipocyclodepsipeptides¹represent a rich and intriguing

class of natural products that possess a broad range of biological

properties including antitumoral, antibiotic, antifungal, and anti-

inﬂammatory activities.²Among them, globomycin (1)³and its

congeners 2-6⁴(Figure 1, part A), isolated from four diﬀerent

strains of actinomycetes in 1978, are representative and inter-

esting examples, especially distinguished by their striking anti-

biotic activities against Gram-negative bacteria.⁵Recently proven

as speciﬁc inhibitors of signal peptidase II,⁶an enzyme respon-

sible for transforming acylated lipoproteins into apoliproteins,⁷

these compounds trigger the accumulation of acylated forms of

lipoproteins in the cytoplasmatic membrane, resulting in cell

death. Therefore, this enzyme represents an attractive biological

target for the development of new antibiotics⁸that has not been

fully exploited. In fact, despite the discovery of this class of

natural products being reported more than 30 years ago,

globomycin (1), the major component of this family, has been

commonly used as a biochemical tool for the identiﬁcation of

new lipoproteins amenable to acylation⁹and for biosynthetic

studies of them.^10,11Comparatively, globomycin and SF-1902 A₅

display greater antibiotic activities when compared to other

antibiotics,¹²such as ampicillin or streptomycin, against E. coli,

Salmonella enteriditis, and Enterobacter cloacae. On the other

hand, SF-1902 A₅is more active than globomycin (MIC against

E. coli NIHJ JC-2: 6.25 μg/mL for 1, 1.56 μg/mL for 2), revealing

that the lipidic chain plays an important role in the biological

activity. More recently, it was discovered that globomycin

exhibited antibacterial activity against Mycobacterium tuberculosis,

showing that the biological action was not dependent on its

inhibition of signal peptidase II, thus suggesting an alternative

mechanism of action in this bacteria.¹³All of these biological

properties render these cyclodepsipeptides as attractive synthetic

endeavors in the search for new antibiotics. Chemically, the

In the present article we describe the total synthesis of these

natural products based on the delineated strategy to provide

access to analogues with modiﬁcation of the amino acids and the

lipidic chain in an eﬃcient and rapid way, taking advantage of

solid phase methodology.

Received: December 19, 2010

Published: March 02, 2011

2011 American Chemical Society

2132

dx.doi.org/10.1021/jo1025145 J. Org. Chem. 2011, 76, 2132–2144

The Journal of Organic Chemistry

ARTICLE

Scheme 2. Synthesis of the Peptidic Fragment in Solid Phase

Figure 1. (A) Molecular structures of globomycin and its congeners.

(B) SAR studies of analogues.

Scheme 1. Synthetic Plan for Globomycin and SF-1902 A₅

amino acid and to ensure the eﬀectiveness of the chosen coupling

procedure, we decided to cleave the dipeptide by treatment of 13

with CH₂Cl₂/AcOH/CF₃CH₂OH (TFE) (7:2:1), which gave

pure dipeptide 14. The synthesis was continued from 13,

repeating the procedure of coupling and Fmoc deprotection

steps for each amino acid being loaded in the following order: (1)

Fmoc-^L-Ser(Bzl)-OH, (2) Fmoc-L-allo-Ile-OH, and (3) Fmoc-

N-Me-^L-Leu-OH to obtain resin 15. At this stage, we again

checked the overall loading of the resin, and thus 15 was treated

with CH₂Cl₂/AcOH/TFE (7:2:1) to provide pure Fmoc pro-

tected pentapeptide 16 in an amount that revealed a load of 0.6

mmol/g. Finally, resin 15 was treated with piperidine to remove

the Fmoc group and prepare the polymer-bound pentapeptide

11 for the coupling with the fatty acid derivative 9 or 10

(Scheme 2).

For the synthesis of the lipidic fragments, we planned to apply

the recent methodology of asymmetric epoxidation developed in

our laboratories based on the use of cyclic sulfonium salts derived

from R-amino acids²¹as a means of stereoselectively construct-

ing an oxirane ring.²²In fact, our experience in this ﬁeld has

allowed us to demonstrate the utility and eﬃciency of the

resulting epoxy amides in reactions with nucleophiles with

complete C-2 regioselectivity.²³Thus, treatment of commercially

available aldehydes 18 and 19 with the sulfur ylide, generated

from its corresponding sulfonium salt 17, provided epoxy amides

’ RESULTS AND DISCUSSION

We began the synthesis of globomycin and SF-1902 A₅with

the construction of the linear peptide on a 2-chlorotrityl chloride

(CTC) resin¹⁸using the Fmoc strategy.¹⁹Thus, Fmoc-Gly was

linked onto the CTC resin by esteriﬁcation with N,N-diisopro-

pylethylamine (DIPEA) to obtain resin 12. After removing the

Fmoc (9-ﬂuorenylmethoxycarbonyl) group by treatment with

20% of piperidine in N,N-dimethylformamide (DMF), the

following Fmoc aminoacid derivative, Fmoc-^L-allo-Thr(TBS)-

OH,²⁰was loaded onto the resulting resin by the action of N,N⁰-

diisopropylcarbodiimide (DIC) in the presence of 1-hydroxy-

benzotriazole (HOBt) in DMF. To check the loading of the

2133

dx.doi.org/10.1021/jo1025145 |J. Org. Chem. 2011, 76, 2132–2144

The Journal of Organic Chemistry

ARTICLE

Scheme 3. Synthesis of the Lipidic Chains of Globomycin

and SF-1902 A₅

considered the use of LiNH₂BH₃, which has been proven to be

an eﬃcient reducing agent for the conversion of amides to

alcohols.²⁸Thus, when 24 and 25 were treated with this reagent

at room temperature, a 1.5:1 mixture consisting of the desired

alcohols²⁹together with the desilylated derivatives, diols 26 and

27, were obtained in 75% combined yields. In light of these

promising results, we carried out this reaction under reﬂux in

THF ﬁnding that, after 4 h, 24 and 25 were transformed into the

corresponding diols 26 and 27 in 85% and 92% yields, respec-

tively. Unable to preserve the silyl protecting group during the

reduction process, we decided to attempt this transformation

directly from the hydroxy amides 22 and 23 by use of a large

excess of LiNH₂BH₃in reﬂuxing THF. To our delight, these

reactions aﬀorded very good results, providing the corresponding

diols 26 and 27 in very high yields (90% and 93%, respectively).

Finally, chemoselective oxidations of 26 and 27 to the targeted

fatty hydroxy acids 9 and 10 were accomplished by the oxidative

system 2,2,6,6-tetramethyl-1-piperidinyloxy free radical/bisace-

toxyiodobenzene (TEMPO/BAIB) in the presence of water³⁰

(Scheme 3, part A). Comparatively, Sharpless asymmetric

epoxidation³¹was similarly exploited for allylic alcohol 28 to

obtain epoxy alcohol 30³²in 75% yield. However, the opening

reaction of the oxirane ring with lithium dimethylcuprate furn-

ished an inseparable mixture of 2-methyl and 3-methyl opened

products³³in a 3:1 ratio. The mixture was subjected to the action

of sodium periodate, and the desired diol 26 was isolated in 41%

yield from epoxy alcohol 30 after puriﬁcation by ﬂash column

chromatography. Oxidation of 26 provided the desired hydroxy

acid 9 as described above. In order to circumvent the lack of

regioselectivity observed during the opening process, we pre-

pared epoxy alcohol 31,³⁴via Sharpless asymmetric epoxidation

of allylic alcohol 29 in 70% yield, which was then smoothly

oxidized to the corresponding epoxy aldehyde with DMSO/

pyr SO₃.³⁵With this epoxy aldehyde in hand, we planned to use

Bode’s carbene-catalyzed epoxide-opening reaction³⁶utilizing

the thiazolium salt 32,³⁷which has proven to be very eﬃcient in

stereoselective oxidative openings of epoxy aldehydes. In our

case, this reaction aﬀorded the corresponding 2-methyl-3-hydro-

xy ester 33 with an anti relative conﬁguration accompanied with

its syn isomer in a 7:3 ratio and in 70% overall yield from 31.

Finally, hydrolysis of 33 provided hydroxy acid 9 contaminated

with its C-2 epimer, which was not separable by chromatographic

methods (Scheme 3, part B). In summary, we concluded that our

approach using sulfur ylides was more eﬃcient than the Sharpless

epoxidation approach in terms of chemical yields as well as in

regio- and stereoselectivities.

20 and 21 in 80% and 73% yields, respectively, and excellent

diastereoselectivities (>98% according to NMR and GC-MS

analyses). This excellent diastereoselectivity was conﬁrmed via

reduction of epoxyamide 20 to epoxy alcohol 30 (see part B of

Scheme 3) by the action of Super-H in 85% yield, displaying

spectroscopic and physical properties, especially its optical rota-

tion ([R]²⁵_D= þ39.5 (c 1.0, CHCl₃); lit.²⁴[R]²⁰_D= þ38.7 (c

0.03, CHCl₃)), which matched with the described in the

literature.²⁴

Therefore, we proceeded with the introduction of the methyl

groups at the C-2 position. To this end, epoxy amides 20 and 21

were exposed to the action of lithium dimethylcuprate²⁵to

obtain hydroxy amides 22 and 23 in excellent yields (82% and

96%, respectively). At this stage, protection of the hydroxyl

groups of both hydroxy amides was required prior to the

reduction of the amide functionality to the corresponding

alcohol. Thus, 22 and 23 were transformed into their corre-

sponding silyl ethers 24 and 25 and then subjected to the action

of lithium triethylborohydride (Super-H).²⁶Unexpectedly, these

reactions did not proceed as desired, recovering starting material

in both cases. The lack of reactivity of compounds 24 and 25

toward Super-H was ascribed to steric factors, thus prompting us

to consider alternative methods to achieve the amide reduction.

Among the various alternatives described in the literature,²⁷we

Having extensively explored the synthesis of the lipidic chain

via epoxide chemistry, with concomitant manipulation of the

amide function by reduction to the alcohol and reoxidation to the

carboxylic acid, we did not wish to discard the possibility of a

direct hydrolysis of the amide group despite its recognized

robustness. To this aim, we initially hydrolyzed the acetal

function installed at the amino alcohol unit by treating hydroxy

amide 22 with 1 N aqueous HCl in dioxane³⁸to provide

compound 34 in 98% yield. Attempts at direct amide hydrolysis

of 22 or from 34 required harsh acidic conditions (H₂SO₄in

reﬂux)³⁹and were unsuccessful as a result of decomposition of

starting material for the ﬁrst case and formation of hydroxy acid 9

for the second, albeit in poor yield (30%). Thus, we proceeded to

investigate the hydrolysis of compound 34 under mild condi-

tions. This task required activating the amide group for subse-

quent hydrolysis. For this objective, we considered diﬀerent

2134

dx.doi.org/10.1021/jo1025145 |J. Org. Chem. 2011, 76, 2132–2144

The Journal of Organic Chemistry

ARTICLE

Scheme 4. Studies on Hydrolysis of Hydroxy Amide 22

Scheme 5. N-Nitrosations of Amides 36 and 45. Synthesis of

Hydroxy Acids 9 and 10

tetrabutylammonium ﬂuoride (TBAF) to provide the hydroxy acid

9 in 95% yield (Scheme 5). This interesting result encouraged us to

try direct N-nitrosation of 34, obtaining in this case acetate 43,

which under basic conditions aﬀorded hydroxy acid 9 in a lower

31% overall yield. Then, we extended this synthetic path to amide

23 to get acid 46, through compounds 44 and 45 obtained in very

similar yields compared with the globomycin series, which was

ﬁnally desilylated to acid 10 in 98% yield (Scheme 5).

With all this chemistry spread around hydroxy amides 34 and

44 and with both peptidic and lipidic key fragments prepared, we

proceeded to completion of the synthesis of the natural cyclo-

depsipetides. For this purpose, we initially coupled hydroxy acids

9 and 10 with peptidic derivative 11 linked onto the resin by

treatment with diethyl cyanophosphonate (DEPC).⁴⁶This treat-

ment was repeated once more in order to obtain a complete

loading of acids 9 and 10 onto the solid support through an

amide bond to give resins 7 and 8. Once the acyclic precursors

were prepared in solid phase, we then achieved the correspond-

ing cleavages by the action of AcOH/CF₃CH₂OH in CH₂Cl₂to

obtain compounds 47 and 48 in high purity as determined by

their NMR spectra and HPLC analyses, verifying the eﬃciency of

the solid phase synthesis. The syntheses of globomycin (1) and

SF-1902 A₅(2) were then completed following the procedure

described by Kogen et al. using a Yamaguchi macro-

lactonization⁴⁷and ﬁnal deprotection of the silyl and benzyl

ether groups via compounds 49 and 51 for globomycin (1) and

50 and 52 for SF-1902 A₅(2) (Scheme 6).

possibilities. One of them was the formation of an oxazolidinone

ring, amenable to hydrolysis under basic conditions.⁴⁰For this

purpose, dihydroxy amide 34 was transformed into its mono-

carbonate 35 by reaction with 1.0 equiv of ethyl chloroformate in

good yield (87%), which was subjected to basic conditions

provided by diﬀerent bases such as LDA, LHMDS, or NaH.

Unfortunately in all attempted cases, we were not able to

construct the oxazolidinone ring, compound 38, recovering

starting material or obtaining 34 via hydrolysis of the carbonate

group. Considering the possibility of an interference of the free

hydroxyl group, we decided to prepare bis(silyl ether) 36 by

treatment of 34 with excess of tert-butyldimethylsilyl triﬂuor-

omethanesulfonate (TBSOTf), which was selectively desilylated

by the action of (1S)-(þ)-10-camphorsulfonic acid (CSA) to

obtain hydroxy amide 37. Reaction of 37 with ethyl chlorofor-

mate and subsequent basic treatment of the resulting carbonate

or reaction of 37 with triphosgene⁴¹or N,N⁰-carbonyldiimida-

zole (CDI)⁴²proved similarly unsuccessful in the formation of

the coveted oxazolidinone 39. In an alternative way, activation of

the amide group by introduction of Boc,⁴³Ac or Ms⁴⁴groups via

amide deprotonation of 36 by the action of base (LDA, BuLi,

LHMDS or NaH), followed by the addition of the corresponding

electrophilic agent, did not provide the expected products 40a-c,

recovering starting bis(silyl ether) 36 instead (Scheme 4).

In light of the lack of reactivity of 36 or related amides, as

described before, we considered the introduction of a N-nitroso

functional group as a way of activation of the amide group.⁴⁵

Interestingly, N-nitrosation of 36 under conventional conditions⁴⁵

aﬀorded the silyloxy acid derivative 42 in 75% yield, probably

formed through N-nitroso derivative 41a, which surprisingly was

not detected. Silyloxy acid 42 was subjected to treatment with

’ CONCLUSIONS

In conclusion, we have described new syntheses of the

antibiotics globomycin (1) and SF-1902 A₅(2) that incorporate

2135

dx.doi.org/10.1021/jo1025145 |J. Org. Chem. 2011, 76, 2132–2144

The Journal of Organic Chemistry

ARTICLE

Scheme 6. Synthesis of Globomycin (1) and SF-1902 A₅(2)

via a Yamaguchi Macrolactonization

0.78 mmol, 2.0 equiv) and DIC (155 μL, 1.0 mmol. 2.5 equiv) in dry

DMF (3 mL). The resulting suspension was shaken at 280 rpm for 24 h,

and then, the solution was unloaded and the resin washed with dry DMF

(5 ꢀ 3 mL). The resulting swelled resin was used in the next step.

2-Chlorotrityl-Gly-L-allo-Thr(TBS)-L-Ser(Bzl)-L-allo-Ile-N-

Me-^L-Leu-Fmoc Resin 15. The polypropylene syringe loaded with

the swelled resin 13 was treated with 20% piperidine in DMF (3 ꢀ

3 mL ꢀ 10 min). After the last run, the resin was washed with dry

DMF (5 ꢀ 3 mL) and loaded with a solution of Fmoc-^L-Ser(Bzl)-OH

(559 mg, 1.18 mmol, 3.0 equiv), HOBt (161 mg, 1.17 mmol, 3.0

equiv) and DIC (216 μL, 1.37 mmol, 3.5 equiv) in dry DMF (3 mL).

The resulting suspension was shaken at 280 rpm for 24 h, the solution

was unloaded, and the resin was washed with dry DMF (5 ꢀ 3 mL).

The resulting swelled resin was used in the next step. This sequence

was repeated with Fmoc-^L-allo-Ile-OH (276 mg, 0.78 mmol, 2.0

equiv), HOBt (107 mg, 0.78 mmol, 2.0 equiv) and DIC (155 μL,

1.0 mmol, 2.5 equiv), followed after treatment with 20% piperidine in

DMF by loading of Fmoc-N-Me-^L-Leu-OH (358 mg, 0.98 mmol, 2.5

equiv), HOBt (135 mg, 0.98 mmol, 2.5 equiv) and DIC (195 μL, 1.25

mmol, 3.2 equiv). Finally, the resin was washed with DMF (5 ꢀ

3 mL), DCM (3 ꢀ 3 mL), MeOH (3 ꢀ 3 mL) and Et₂O (3 ꢀ 3 mL).

The resulting resin was dried under vacuum to recover 626 mg of

polymer-bound N-Fmoc protected pentapeptide 15.

Fmoc-N-Me-L-Leu-L-allo-Ile-L-Ser(Bzl)-L-allo-Thr(TBS)-Gly-

OH (16). Release of a small amount of peptide from resin 15 (17.5 mg)

by treatment with CH₂Cl₂/AcOH/TFE (1.0 mL, 7:2:1) gave the N-

Fmoc protected pentapeptide 16 (9.6 mg), which revealed a load of 0.6

mmol/g for resin 15. Data for 16: ¹H NMR (400 MHz, DMSO-d₆, two

rotamers in a 2.9:1 ratio) Major rotamer: δ 0.03 (s, 3 H), 0.01 (s, 3 H),

0.71-0.89 (m, 9 H), 0.80 (s, 9 H), 0.88 (d, J = 6.6 Hz, 3 H), 0.95-1.07

(m, 1 H), 1.05 (d, J = 6.2 Hz, 3 H), 1.21-1.36 (m, 2 H), 1.47-1.61 (m, 2

H), 1.73-1.84 (m, 3/4 H), 2.68 (s, 9/4 H), 2.79 (s, 3/4 H), 3.52-3.59

(m, 3 H), 3.77 (dd, J = 17.5, 6.2 Hz, 1 H), 4.01 (q, J = 6.5 Hz, 1 H), 4.25-

4.45 (m, 4 H), 4.38 (dd, J = 9.0, 7.2 Hz, 1 H), 4.48 (m, 2 H), 4.65-4.70

(m, 2 H), 7.24-7.33 (m, 7 H), 7.41 (t, J = 7.5 Hz, 2 H), 7.62-7.70 (m, 1

H), 7.63 (d, J = 7.1 Hz, 2 H), 7.89 (d, J = 7.5 Hz, 2 H), 7.99 (d, J = 9.0 Hz,

1 H), 8.05 (d, J = 7.6 Hz, 1 H), 8.16 (t, J = 5.0 Hz, 1 H); FAB HRMS

(NBA) m/e 930.5043, M þ H^þcalcd for C₅₀H₇₁N₅O₁₀Si 930.5049.

Epoxy Amide 20. To a solution of sulfonium salt 17 (3.66 g, 11.59

mmol, 1.2 equiv) in ^tBuOH (20.0 mL) was added a solution of NaOH

3.0 M in H₂O (3.86 mL, 11.59 mmol, 1.2 equiv). After 1 h at 25 °C, a

two important novel methodologies for the total syntheses of

these cyclodepsipeptides: (1) the use of solid phase for the

assembly of the peptidic and lipidic fragments that are found in

these molecules, and (2) extension of our new asymmetric

methodology of epoxidation for the stereoselective synthesis of

the lipidic chains. The implementation of this new methodology

of epoxidation oﬀers the advantage of constructing the lipidic

chain through an oxirane ring, a versatile functional group that

allows the generation of 1,2-bifunctional groups. This in turn

provides access to a broad variety of modiﬁed-lipidic chains that

appear to play an essential role in the antibiotic properties of

these compounds. This feature combined with the use of a solid

phase-based synthesis allows rapid and easy access to a wide array

of globomycin analogues in the quest for new antibiotics with

novel mechanism of action. The generation of a broad library of

globomycin analogues as well as their biological evaluations

represent our priorities in current and future investigations.

solution of heptanal 18 (1.34 mL, 9.66 mmol, 1.0 equiv) in BuOH

(5.0 mL) was added and the reaction mixture was vigorously stirred

overnight at 25 °C. After this time, reaction mixture was diluted with

Et₂O and H₂O, both phases were separated, and the aqueous layer

extracted with Et₂O twice. Combined organic extracts were then washed

with water and brine, filtered and concentrated. Purification by flash

column chromatography (silica gel, 30% AcOEt in hexanes) provided

epoxy amide 20 (2.55 g, 80%) as a colorless oil: R_f= 0.48 (silica gel, 50%

AcOEt in hexanes); [R]²⁵= þ11.9 (c 1.8, CH₂Cl₂); H NMR (400

’ EXPERIMENTAL SECTION⁴⁸

MHz, CDCl₃) δ 0.86 (t, J = 6.9 Hz, 3 H), 1.26-1.30 (m, 4 H), 1.32-

1.36 (m, 2 H), 1.40-1.55 (m, 2 H), 1.51 (s, 3 H), 1.62 (s, 3 H), 1.67-

1.74 (m, 2 H), 1.77-1.81 (m, 1 H), 2.01-2.07 (m, 1 H), 2.10 (s, 3 H),

2.45 (ddd, J = 13.1, 8.6, 7.0 Hz, 1 H), 2.58 (ddd, J = 12.7, 7.3, 5.2 Hz, 1

H), 3.15 (ddd, J = 6.0, 4.3, 1.9 Hz, 1 H), 3.32 (d, J = 1.9 Hz, 1 H), 3.88 (d,

J = 9.2 Hz, 1 H), 3.99 (ddd, J = 9.2, 5.3, 1.3 Hz, 1 H), 4.28 (ddd, J = 10.3,

5.0, 3.2 Hz, 1 H); ¹³C NMR (100 MHz, CDCl₃) δ 14.0, 16.0, 22.5, 22.9,

25.8, 26.3, 29.0, 30.9, 31.5, 31.7, 34.5, 53.9, 55.8, 58.5, 67.0, 95.8, 164.2;

FAB HRMS (NBA) m/e 330.2093, M þ H^þcalcd for C₁₇H₃₁NO₃S

330.2103.

Fmoc-Gly-OH Loaded 2-Chlorotrityl Resin 12. A 5 mL poly-

propylene syringe fitted with polyethylene porous disk charged with

2-chlorotrityl chloride resin (300 mg, L = 1.3 mmol/g, 0.39 mmol, 1.0

equiv), was loaded with a solution of Fmoc-Gly-OH (348 mg, 1.17

mmol, 3.0 equiv) and DIPEA (235 μL, 1.37 mmol, 3.5 equiv) in dry

DMF (3 mL). The resulting suspension was shaken at 280 rpm for 30 h,

the solution was unloaded, and the resin was washed by shaking with dry

DMF (5 ꢀ 3 mL). The resulting swelled resin was used in the next step.

2-Chlorotrityl-Gly-L-allo-Thr(TBS)-Fmoc Resin 13. The poly-

propylene syringe loaded with the swelled resin 12 was treated with 20%

piperidine in DMF (3 ꢀ 3 mL ꢀ 10 min). After the last run, the resin was

washed with dry DMF (5 ꢀ 3 mL) and loaded with a solution of Fmoc-^L-

allo-Thr(TBS)-OH (355 mg, 0.78 mmol, 2.0 equiv), HOBt (107 mg,

Epoxy Amide 21. To a solution of sulfonium salt 17 (2.51 g, 7.95

mmol, 1.2 equiv) in ^tBuOH (12.0 mL) was added a solution of NaOH

3.0 M in H₂O (2.65 mL, 7.95 mmol, 1.2 equiv). After 1 h at 25 °C, a

solution of nonanal 19 (1.14 mL, 6.63 mmol) in ^tBuOH (5.0 mL) was

2136

dx.doi.org/10.1021/jo1025145 |J. Org. Chem. 2011, 76, 2132–2144

The Journal of Organic Chemistry

ARTICLE

added and the reaction mixture was vigorously stirred overnight at

25 °C. After this time, reaction mixture was diluted with Et₂O and H₂O,

both phases were separated, and the aqueous layer extracted with Et₂O

twice. Combined organic extracts were then washed with water and

brine, filtered and concentrated. Purification by flash column chroma-

tography (silica gel, 30% AcOEt in hexanes) provided epoxy amide 21

(1.72 g, 73%) as a colorless oil: R_f= 0.51 (silica gel, 50% AcOEt in

29.5, 29.6, 30.9, 31.8, 33.5, 35.8, 43.1, 56.9, 66.5, 74.8, 95.3, 173.6; FAB

HRMS (NBA) m/e 374.2735, M þ H^þcalcd for C₂₀H₃₉NO₃S

374.2729.

Silyl Ether 24. A solution of hydroxy amide 22 (2.92 g, 8.44 mmol,

1.0 equiv) in CH₂Cl₂(20 mL) was treated with tert-butyldimethylsilyl

trifluoromethanesulphonate (TBSOTf) (3.9 mL, 16.88 mmol, 2.0

equiv) at 0 °C in the presence of 2,6-lutidine (2.5 mL, 21.10 mmol,

2.5 equiv). After 0.5 h at 0 °C, the reaction mixture was quenched by

addition of MeOH (2.0 mL), followed by addition of aqueous saturated

NH₄Cl solution and dilution with Et₂O (50 mL). After separation of

both phases, the aqueous phase was extracted with Et₂O (2 ꢀ 30 mL),

and the combined organic layers were washed with brine and dried with

MgSO₄. After filtration, the solvents were removed by reduced pressure

to obtain a crude product, which was purified by flash column

chromatography (silica gel, 20% EtOAc in hexanes) to afford silyl ether

24 (3.70 g, 95%) as a colorless oil: R_f= 0.40 (silica gel, 20% AcOEt in

hexanes); [R]²⁵= þ19.4 (c 0.8, CH₂Cl₂); ¹H NMR (400 MHz,

CDCl₃) δ 0.88 (t, J = 7.0 Hz, 3 H), 1.20-1.37 (m, 12 H), 1.43-1.49 (m,

1 H), 1.53 (s, 3 H), 1.64 (s, 3 H), 1.69-1.78 (m, 1 H), 1.79-1.84 (m, 1

H), 2.03-2.08 (m, 1 H), 2.12 (s, 3 H), 2.47 (ddd, J = 13.1, 8.7, 7.0 Hz, 1

H), 2.60 (ddd, J = 12.8, 7.4, 5.2 Hz, 1 H), 3.17 (ddd, J = 6.5, 4.6, 2.0 Hz, 1

H), 3.33 (d, J = 2.0 Hz, 1 H), 3.89 (d, J = 9.2 Hz, 1 H), 4.01 (ddd, J = 9.2,

5.3, 1.5 Hz, 1 H), 4.30 (ddd, J = 10.3, 4.9, 3.1 Hz, 1 H); ¹³C NMR (100

MHz, CDCl₃) δ 14.1, 16.0, 22.6, 22.9, 25.9, 26.3, 29.2, 29.3, 29.4, 30.9,

31.5, 31.8, 34.6, 53.9, 55.8, 58.5, 67.0, 95.8, 164.2; FAB HRMS (NBA)

m/e 358.2386, M þ H^þcalcd for C₁₉H₃₅NO₃S 358.2416.

hexanes); [R]²⁵= -14.3 (c 1.6, CH₂Cl₂); ¹H NMR (400 MHz,

Hydroxy Amide 22. To a suspension of CuI (925 mg, 4.84 mmol,

4.0 equiv) in THF (20 mL) was added dropwise MeLi (1.6 M in Et₂O,

6.10 mL, 9.68 mmol, 8.0 equiv) at 0 °C. Then, a solution of epoxy amide

20 (400 mg, 1.21 mmol, 1.0 equiv) in THF (5 mL) was added to the

resulting colorless solution of Me₂CuLi at 0 °C. The reaction mixture

was stirred for 8 h at this temperature and quenched by careful addition

of aqueous saturated NH₄Cl solution, followed by dilution with Et₂O.

After separation of both phases, the aqueous phase was extracted with

Et₂O and the combined organic layers were sequentially washed with

aqueous saturated NH₄Cl solution, water and brine. After treatment

with MgSO₄, the solvents were removed by reduced pressure to obtain

crude 2-methyl-3-hydroxy amide, which was subjected to purification by

flash column chromatography (silica gel, 50% AcOEt in hexanes) to

CDCl₃) δ -0.05 (s, 3 H), -0.02 (s, 3 H), 0.80 (s, 3 H), 0.83 (t, J = 7.1

Hz, 3 H), 0.95 (d, J = 6.9 Hz, 3 H), 1.18-1.23 (m, 12 H), 1.36-1.50 (m,

2 H), 1.51 (s, 3 H), 1.53 (s, 3 H), 1.62-1.72 (m, 1 H), 1.94-2.02 (m, 1

H), 2.07 (s, 3 H), 2.30-2.41 (m, 1 H), 2.48-2.57 (m, 1 H), 2.68 (dq, J =

13.7, 6.9 Hz, 1 H), 3.74 (d, J = 9.1 Hz, 1 H), 3.84 (ddd, J = 9.0, 4.8, 1.5

Hz, 1 H), 3.94 (dt, J = 9.0, 3.5 Hz, 1 H), 4.10 (ddd, J = 10.8, 4.8, 2.3 Hz, 1

H); ¹³C NMR (100 MHz, CDCl₃) δ = -4.9, -4.2, 13.9, 14.8, 17.8, 21.7,

22.5, 23.6, 25.8, 26.3, 29.6, 31.8, 32.9, 43.4, 56.4, 66.5, 73.7, 94.7, 172.6;

FAB HRMS (NBA) m/e 460.3308, M þ H^þcalcd for C₂₄H₄₉NO₃SSi

460.3281.

Silyl Ether 25. A solution of hydroxy amide 23 (350 mg, 0.94 mmol,

1.0 equiv) in CH₂Cl₂(10 mL) was treated with tert-butyldimethylsilyl

trifluoromethanesulphonate (TBSOTf) (0.43 mL, 1.88 mmol, 2.0

equiv) at 0 °C in the presence of 2,6-lutidine (0.27 mL, 2.35 mmol,

2.5 equiv). After 0.5 h at 0 °C, the reaction mixture was quenched by

addition of MeOH (0.5 mL), followed by addition of aqueous saturated

NH₄Cl solution and dilution with Et₂O (20 mL). After separation of

both phases, the aqueous phase was extracted with Et₂O (2 ꢀ 10 mL),

the combined organic layers were washed with brine and dried with

MgSO₄. After filtration, the solvents were removed by reduced pressure

to obtain a crude product which was purified by flash column chroma-

tography (silica gel, 10% EtOAc in hexanes) to afford silyl ether 25 (450

mg, 98%) as a colorless oil: R_f= 0.28 (silica gel, 10% AcOEt in hexanes);

[R]²⁵_D= -9.6 (c 0.8, CH₂Cl₂); ¹H NMR (400 MHz, CDCl₃) δ 0.04 (s,

3 H), 0.05 (s, 3 H), 0.86 (s, 9 H), 0.87 (t, J = 6.8 Hz, 3 H), 1.00 (d, J = 6.8

Hz, 3 H), 1.21-1.32 (m, 12 H), 1.42-1.53 (m, 2 H), 1.57 (s, 3 H), 1.58

(s, 3 H), 1.70-1.78 (m, 1 H), 1.99-2.08 (m, 1 H), 2.12 (s, 3 H), 2.43

(ddd, J = 13.3, 9.0, 6.9 Hz, 1 H), 2.58 (ddd, J = 12.3, 6.8, 5.0 Hz, 1 H),

2.74 (dq, J = 8.9, 6.8 Hz, 1 H), 3.80 (d, J = 9.0 Hz, 1 H), 3.89 (ddd, J = 9.0,

4.8, 1.3 Hz, 1 H), 3.99 (dt, J = 9.0, 3.3 Hz, 1 H), 4.16 (ddd, J = 10.7, 4.6,

2.2 Hz, 1 H); ¹³C NMR (100 MHz, CDCl₃) δ -4.7, -4.1, 14.1, 14.9,

15.9, 17.9, 21.9, 22.7, 23.8, 25.9, 26.5, 29.3, 29.7, 30.1, 31.2, 31.9, 33.1,

33.3, 43.6, 56.6, 66.7, 73.9, 94.8, 172.8; FAB HRMS (NBA) m/e

488.3605, M þ H^þcalcd for C₂₆H₅₃NO₃SSi 488.3594.

obtain 22 (341 mg, 82%) as a colorless oil: R_f= 0.38 (silica gel, 33%

AcOEt in hexanes); [R]²⁵= þ19.7 (c 0.9, CH₂Cl₂); H NMR (400

MHz, CDCl₃) δ 0.88 (t, J = 6.9 Hz, 3 H), 1.26 (d, J = 7.0 Hz, 3 H), 1.27-

1.32 (m, 8 H), 1.37-1.47 (m, 2 H), 1.55 (s, 3 H), 1.64 (s, 3 H), 1.72-

1.78 (m, 1 H), 2.00-2.08 (m, 1 H), 2.12 (s, 3 H), 2.42 (ddd, J = 13.3, 9.4,

6.5 Hz, 1 H), 2.59 (ddd, J = 13.3, 6.7, 4.8 Hz, 1 H), 2.68 (dq, J = 7.0, 5.1

Hz, 1 H), 3.32 (d, J = 7.7 Hz, 1 H), 3.57-3.63 (m, 1 H), 3.84 (d, J = 9.2

Hz, 1 H), 3.98 (ddd, J = 9.1, 5.2, 1.5 Hz, 1 H), 4.14 (ddd, J = 10.6, 5.0, 2.6

Hz, 1 H); ¹³C NMR (100 MHz, CDCl₃) δ 14.1, 14.2, 15.8, 16.0, 22.6,

22.9, 25.8, 26.5, 29.3, 30.9, 31.8, 33.5, 43.1, 57.0, 66.5, 74.8, 95.3, 173.6;

FAB HRMS (NBA) m/e 346.2428, M þ H^þcalcd for C₁₈H₃₅NO₃S

346.2416.

Hydroxy Amide 23. To a suspension of CuI (2.13 g, 11.20 mmol,

4.0 equiv) in THF (50 mL) was added dropwise MeLi (1.6 M in Et₂O,

14.0 mL, 22.40 mmol, 8.0 equiv) at 0 °C. Then, a solution of epoxy

amide 21 (1.00 g, 2.80 mmol, 1.0 equiv) in THF (10 mL) was added to

the resulting colorless solution of Me₂CuLi at 0 °C. The reaction

mixture was stirred for 8 h at this temperature and quenched by careful

addition of aqueous saturated NH₄Cl solution, followed by dilution with

Et₂O. After separation of both phases, the aqueous phase was extracted

with Et₂O and the combined organic layers were sequentially washed

with aqueous saturated NH₄Cl solution, water and brine. After treat-

ment with MgSO₄, the solvents were removed by reduced pressure to

obtain crude 2-methyl-3-hydroxy amide, which was subjected to pur-

ification by flash column chromatography (silica gel, 30% AcOEt in

Diol 26 from Hydroxy Amide 22. A freshly prepared solution of

LDA [Diisopropylamine (0.41 mL, 2.90 mmol, 10.0 equiv) was added to

a solution of n-BuLi (1.6 M solution in hexanes, 1.8 mL, 2.90 mmol, 10.0

equiv) in THF (5.0 mL) at 0 °C] was treated with borane-ammonia

complex (90 mg, 2.90 mmol, 10.0 equiv) at 0 °C. After 15 min at this

temperature, the resulting suspension was stirred at room temperature

for additional 30 min before cooling again at 0 °C. Then, a solution of

hydroxy amide 22 (100 mg, 0.29 mmol, 1.0 equiv) in THF (3.0 mL) was

added. The reaction mixture was warmed to room temperature and then

heated under reflux for 8 h. After this time, reaction mixture was allowed

to warm to room temperature and hydride excess was quenched by

careful addition of MeOH (2.0 mL). The crude mixture was then diluted

with Et₂O and treated with a saturated aqueous NH₄Cl solution. The

hexanes) to obtain 23 (1.0 g, 96%) as a colorless oil: R_f= 0.42 (silica gel,

33% AcOEt in hexanes); [R]²⁵= þ19.3 (c 1.9, CH₂Cl₂); H NMR

(400 MHz, CDCl₃) δ 0.88 (t, J = 7.1 Hz, 3 H), 1.26 (d, J = 6.7 Hz, 3 H),

1.23-1.32 (m, 12 H), 1.41-1.48 (m, 1 H), 1.56 (s, 3 H), 1.64 (s, 3 H),

1.72-1.77 (m, 1 H), 1.79-1.88 (m, 1 H), 2.00-2.08 (m, 1 H), 2.13 (s, 3

H), 2.42 (ddd, J = 15.7, 9.4, 6.5 Hz, 1 H), 2.56-2.64 (m, 1 H), 2.68 (dq, J

= 6.8, 5.1 Hz, 1 H), 3.58-3.62 (m, 1 H), 3.83 (d, J = 9.2 Hz, 1 H), 3.98

(ddd, J = 9.2, 5.2, 1.6 Hz, 1 H), 4.14 (ddd, J = 11.2, 5.1, 2.6 Hz, 1 H); ¹³C

NMR (100 MHz, CDCl₃) δ 14.0, 15.7, 15.9, 22.6, 22.8, 25.8, 26.5, 29.2,

2137

dx.doi.org/10.1021/jo1025145 |J. Org. Chem. 2011, 76, 2132–2144

The Journal of Organic Chemistry

ARTICLE

organic phase was separated, and the aqueous phase was extracted with

Et₂O (twice). The combined organic solution was sequentially washed

with water and brine, dried (MgSO₄), filtered and concentrated under

reduced pressure. The resulting crude product was purified by flash

column chromatography (silica gel, 20% f 50% AcOEt in hexanes) to

obtain to obtain diol 26 (45 mg, 90%) as a colorless oil: R_f= 0.45 (silica

gel, 50% AcOEt in hexanes); [R]²⁵_D= þ16.1 (c 0.7, CH₂Cl₂); ¹H NMR

(400 MHz, CDCl₃) δ 0.87 (t, J = 7.0 Hz, 3 H), 0.88 (d, J = 6.6 Hz, 3 H),

1.24-1.35 (m, 7 H), 1.39-1.58 (m, 3 H), 1.67(dsext, J = 7.2, 3.7 Hz, 1

H), 3.29 (s, 2 H), 3.53 (dt, J = 7.8, 3.2 Hz, 1 H), 3.59 (dd, J = 10.8, 7.3 Hz,

1 H), 3.75 (dd, J = 10.8, 3.7 Hz, 1 H); ¹³C NMR (100 MHz, CDCl₃) δ

13.9, 14.1, 22.6, 25.1, 29.4, 31.8, 35.3, 39.7, 67.6, 77.3; FAB HRMS

(NBA) m/e 175.1708, M þ H^þcalcd for C₁₀H₂₂O₂175.1698.

heated under reflux for 4 h. After this time, reaction mixture was allowed

to warm to room temperature and worked up as described in Procedure

A to obtain a crude product that was purified by flash column

chromatography (silica gel, 50% AcOEt in hexanes) to obtain diol 27

(80 mg, 92%) that exhibited identical physical and spectroscopic

properties as that obtained from 23 as described above.

Hydroxy Acid 9. Diol 26 (100 mg, 0.57 mmol) was dissolved in a

mixture of CH₃CN/H₂O (8.0 mL, 1:1) and the resulting solution was

treated with BAIB (918 mg, 2.85 mmol, 5.0 equiv) followed by TEMPO

(44 mg, 0.29 mmol, 0.5 equiv) at 25 °C. After 6 h, the crude mixture was

diluted with AcOEt and quenched by the addition of a saturated aqueous

Na₂S₂O₃solution, and after separation of both layers, the aqueous phase

was then extracted with AcOEt. The combined organic solution was

washed with saturated aqueous Na₂S₂O₃solution again and dried over

anhydrous MgSO₄, and the solvent was evaporated under reduced

pressure. Purification of the crude acid by flash column chromatography

(silica gel, 3% f 5% MeOH in CH₂Cl₂) provided hydroxy acid 9 (77

mg, 72%) as a colorless oil: R_f= 0.25 (silica gel, 5% MeOH in CH₂Cl₂);

[R]²⁵_D= þ5.0 (c 0.54, CH₂Cl₂); ¹H NMR (400 MHz, CDCl₃) δ 0.86 (t,

J = 6.8 Hz, 3 H), 1.20-1.37 (m, 7 H), 1.23 (d, J = 7.2 Hz, 3 H), 1.40-

1.57 (m, 3 H), 2.54 (quint, J = 6.9 Hz, 1 H), 3.64-3.70 (m, 1 H); ¹³C

NMR (100 MHz, CDCl₃) δ 14.0, 14.2, 22.6, 25.4, 29.2, 31.8, 34.6, 45.2,

67.6, 77.3, 180.8; FAB HRMS (NBA) m/e 189.1504, M þ H^þcalcd for

C₁₀H₂₀O₃189.1491.

Hydroxy Acid 10. Diol 27 (70 mg, 0.37 mmol) was dissolved in a

mixture of CH₃CN/H₂O (6.0 mL, 1:1) and the resulting solution was

treated with BAIB (602 mg, 1.84 mmol, 5.0 equiv) followed by TEMPO

(30 mg, 0.18 mmol, 0.5 equiv) at 25 °C. After 6 h, the crude mixture was

diluted with AcOEt and quenched by the addition of a saturated aqueous

Na₂S₂O₃solution, and after separation of both layers, the aqueous phase

was then extracted with AcOEt. The combined organic solution was

washed with saturated aqueous Na₂S₂O₃solution again and dried over

anhydrous MgSO₄, and the solvent was evaporated under reduced

pressure. Purification of the crude acid by flash column chromatography

(silica gel, 3% f 5% MeOH in CH₂Cl₂) provided hydroxy acid 10 (49

mg, 65%) as a colorless oil: R_f= 0.29 (silica gel, 5% MeOH in CH₂Cl₂);

[R]²⁵_D= þ1.9 (c 2.5, CH₂Cl₂); ¹H NMR (400 MHz, CDCl₃) δ 0.88 (t, J

= 6.9 Hz, 3 H), 1.24 (d, J = 7.2 Hz, 3 H), 1.25-1.39 (m, 11 H), 1.42-

1.57 (m, 3 H), 2.56 (quint, J = 6.9 Hz, 1 H), 3.67-3.72 (m, 1 H), 6.22

(bs, 1 H); ¹³C NMR (100 MHz, CDCl₃) δ 14.1, 14.2, 22.6, 25.4, 29.2,

29.5, 31.8, 34.6, 45.2, 77.3, 180.9; FAB HRMS (NBA) m/e 217.1824, M

þ H^þcalcd for C₁₂H₂₄O₃217.1804.

Diol 26 from Silyl Ether 24. A freshly prepared solution of LDA

[Diisopropylamine (0.44 mL, 3.2 mmol, 4.0 equiv) was added to a

solution of n-BuLi (1.6 M solution in hexanes, 2.0 mL, 3.2 mmol, 4.0

equiv) in THF (5.0 mL) at 0 °C] was treated with borane-ammonia

complex (100 mg, 3.2 mmol, 4.0 equiv) at 0 °C. After 15 min at this

temperature, the resulting suspension was stirred at room temperature

for additional 30 min before cooling again at 0 °C. Then, a solution of the

silyl ether 24 (367 mg, 0.80 mmol, 1.0 equiv) in THF (5.0 mL) was

added. The reaction mixture was warmed to room temperature and then

heated under reflux for 4 h. After this time, reaction mixture was allowed

to warm to room temperature and worked out as described in Procedure

A to obtain a crude product that was purified by flash column

chromatography (silica gel, 50% AcOEt in hexanes) to obtain diol 26

(119 mg, 85%) that exhibited identical physical and spectroscopic

properties than the obtained from 22 as described above.

Diol 27 from Hydroxy Amide 23. A freshly prepared solution of

LDA [Diisopropylamine (0.36 mL, 2.60 mmol, 10.0 equiv) was added to

a solution of n-BuLi (1.6 M solution in hexanes, 1.6 mL, 2.60 mmol, 10.0

equiv) in THF (5.0 mL) at 0 °C] was treated with borane-ammonia

complex (80 mg, 2.60 mmol, 10.0 equiv) at 0 °C. After 15 min at this

temperature, the resulting suspension was stirred at room temperature

for additional 30 min before cooling again at 0 °C. Then, a solution of

hydroxy amide 23 (97 mg, 0.26 mmol, 1.0 equiv) in THF (3.0 mL) was