Salinosporamide A

A Total Synthesis of Salinosporamide A

Léo B. Marx,[a] and Jonathan W. Burton*[a]

Abstract: Salinosporamide A is a b-lactone proteasome inhibitor currently in clinical trials for the treatment of multiple-myeloma. Herein we report a short synthesis of this small, highly functionalized, biologically important natural product that uses an oxidative radical cyclization as a key step and allows the preparation of gram quantities of advanced synthetic intermediates.

Introduction

In 2003, Fenical and co-workers reported the isolation, structure determination and cancer cell cytotoxicity of the marine-derived
[1, 2]
natural product salinosporamide A 1 (Figure 1). Salinosporamide A, is structurally closely related to omuralide 2, the ring-closed form of lactacystin 3, in that they both contain a pyrrolidinone fused to a b-lactone which is key to their biological activity. Both 1 and 2 are small molecule proteasome inhibitors whose mechanism of action involves esterification of an N- terminal threonine residue of the 20S proteasome by the electrophilic β-lactone.[3] Given the higher proteasome activity exhibited by cancer cells, proteasome inhibition is an active area of research for cancer chemotherapy. Indeed, salinosporamide A has entered clinical trials for the treatment of multiple-myeloma,
[4, 5]
solid tumors or lymphoma.

product in enantiopure form.[6-9] It is surprising, given all of the advances in synthetic methodology over the previous decades, that only three of the reported total syntheses of salinosporamide A (1) have fewer than 20 steps and a number have greater than
30steps. These step counts, subjective as they may be, demonstrate the synthetic challenge that such a densely functionalized, stereochemically rich molecule presents. Herein we report a novel 16 step (9 chromatographed intermediates) stereocontrolled synthesis of salinosporamide A (1) that features an oxidative radical cyclization and a selenolactonization as key steps and provides gram quantities of key synthetic intermediates.

Results and Discussion

Retrosynthetic Analysis
Our retrosynthetic analysis of salinosporamide A (1) is delineated in Scheme 1. In the first synthesis of salinosporamide A,6a Corey reported an elegant method to install the C-5 and C-6 stereocenters of 1 that involved addition of a cyclohexenylzinc reagent to a [4.3.0]-bicyclic aldehyde. Variations of this method have been used by the majority of researchers to set the C-5 and C-6 stereocenters of 1 and we elected to investigate this method for our synthesis. We therefore decided to synthesize salinosporamide A (1) from the lactone-lactam 4 which would, in turn, be made from the fused bicyclic aldehyde 5 on addition of the appropriate cyclohexenyl organometallic. Taking our cue

1
2

H
N
4
3
H

OH
O

O
Me
H
N
Me

OH
O

O
Me
H
N
Me

OH
S

NHAc
from Danishefsky’s synthesis of salinosporamide A6c (the second synthesis of the natural product), we reasoned that the C-3 tertiary alkoxy-bearing stereocenter could be installed by

Me
O
Me
H
O
Me
HO
O
selenolactonization onto the 1,1-disubsituted alkene contained

Cl
CO
H
2
within 6. The alkene 6 would itself be prepared from the selenide

salinosporamide A; 1
omuralide; 2
lactacystin; 3
7 with the selenide being formed by nucleophilic opening of the [3.3.0]-bicyclic γ-lactone 8. A key step in our synthesis of

Figure 1. Salinosporamide A, omuralide and lactacystin.

This biologically important natural product presents a significant challenge to synthetic chemists as it contains a high concentration of both electrophilic and nucleophilic functional groups within a small [3.2.0]-bicyclic core containing five contiguous stereocenters including adjacent quaternary centers. The biological activity exhibited by 1 coupled with its interesting structure has resulted in nine total syntheses of the natural
salinosporamide A was to be the oxidative radical cyclization of the amidomalonate 9 to give 8. We have recently shown that N- PMB-protected amidomalonates bearing pendent alkenes form [3.3.0]-bicyclic γ-lactones under oxidative radical conditions and used this methodology in our formal synthesis of 1.8g All of the previous syntheses of salinosporamide A (1) have, at some point involved, protection of the lactam NH (frequently as an N-PMB group)[6-9] and it appeared that such lactam protection was crucial to the successful synthesis of salinosporamide A (1); our first target was therefore the tertiary amide 9a (with the choice of ester to be determined by experiment). In the event, we found that the secondary amide 9b (R = Me or tBu, R’ = H) was a competent

[a] Dr L. B. Marx, Dr J. W. Burton
Department of Chemistry, University of Oxford, Chemistry Research Laboratory, Mansfield Road, Oxford OX1 3TA UK
E-mail: [email protected]
Supporting information for this article is given via a link at the end of the document.
substrate for oxidative radical cyclization to form the [3.3.0]- bicyclic γ-lactone 8b (R = Me or tBu, R’ = H) and we carried the unprotected amide/lactam through the complete synthetic sequence to the natural product itself (vide infra).

Chemistry – A European Journal

FULL PAPER
10.1002/chem.201800046

H
N

R’
N

withdrawing malonate functionality. Coupling of deactivated amines such as 16 would require a highly electrophilic activated carboxylic acid component; however, activation of β-carboxy

O
H
CO
R
2
H
CO
R
2
carboxylic acids (such as 14) can result in the formation of the

Me
O
O
Me
O
Me corresponding anhydrides (e.g. 17).[12, 13] Both the formation of

Cl
1
O
4
O
5
17 coupled with the low nucleophilicity of 16 results in the formation of 15 being highly challenging. Acid fluorides have been reported to prevent anhydride formation allowing coupling of

O
R’
N

CO
R
2
O
O
R’
N

H
2
O
R’
N
OH
β-carboxy carboxylic acids substrates,[13] however, the corresponding acid fluoride derived from 14 (νmax = 1836 cm-1)

tBuO

C
2

O
8a; R’ = PMB 8b; R’ = H

CO2R

tBuO

C
2
CO R
2

SePh
7

tBu
2

CO2R
CO

6
R
2
failed to undergo amide bond formation with 16 (or derivatives). Indeed, under a wide variety of amide bond forming conditions we failed to form any of the desired product 15.[14] We reasoned that reducing the steric bulk of the amine partner might result in successful amide bond formation. We therefore investigated

NCO R
2
R’ CO tBu
2
9a; R’ = PMB 9b; R’ = H

10
OH CO R
2
+
R’HN CO R
2

11a; R’ = PMB 11b; R’ = H
amide formation between dimethyl aminomalonate 18, with reduced steric hindrance both at nitrogen and at the malonate esters. In the event amide coupling was readily achieved between the acid 14 and dimethyl aminomalonate 18 in the presence of

Scheme 1. Retrosynthetic analysis of salinosporamide A (1).

Synthesis
We initially prepared the carboxylic acid 14 in 73% yield by alkylation of the known oxazolidinone 128g using t-butyl bromoacetate followed by hydrolysis (Scheme 2).[10] We aimed to convert 14 into the amidomalonate 15 by coupling of the acid chloride derived from 14 with the readily prepared aminomalonate
HATU (Scheme 3). Oxidative elimination of the phenylselanyl group from 19 occurred in good yield using a modification of the procedure reported by Kocienski[15] to give cyclization substrate 20. We have previously shown that N-PMB protected amidomalonates bearing terminal alkenes readily undergo cyclization to form the corresponding [3.3.0]-bicyclic γ-lactones and had found that cyclization of simple N-unprotected amidomalonates bearing terminal alkenes was capricious.8g

16 (ESI) under Shotten-Baumann
conditions.[6e],[11]
Thus,
O
a)dimethyl aminomalonate 18,
O
CO
Me
2

exposure of the acid 14 to oxalyl chloride followed by addition of the amine 168g resulted in quantitative recovery of 16 after workup
PhSe

OH
HATU
84%
PhSe

N
H

CO
Me
2

and formation of the corresponding dicarboxylic acid derived from
tBuO
C
2
14
tBuO
C
2
19

14 (mass spectrometry evidence).
b)NaIO
, then heat
4

79%

a)NaHMDS,

O
CO
Me
2

Me
2

PhSe
N
O
BrCH CO
2
85%
tBu
2
PhSe
N
O
H
N
2
CO
18
Me
2

N
H

CO
Me
2

Bn
12
tBuO
C
2
Bn
13
tBuO
C
2
20

PMBHN
CO2Bn

CO2Bn 16
O

PhSe

CO2Bn

b)BnOLi, then LiOH 85%

O
Scheme 3. Synthesis of amide 20. Reagents and conditions: a) dimethyl aminomalonte 18, HATU, i-Pr2NEt, DMF, 0 °C to RT, 84%; b) NaIO4, NaHCO3, MeOH, THF, water, then heat in toluene, 79%.

PhSe

N
PMB

CO
Bn
2
see text
PhSe

OH
Given these previous results we were delighted to find that exposure of amidomalonate 20 to our standard conditions for

tBuO
C
2
tBuO
C
2
oxidative radical cyclization[16] resulted in the formation of the

15
14
desired [3.3.0]-bicyclic γ-lactone 23a in 48% yield along with the uncyclized but oxidized products 21 and 22 (Scheme 4). 1H NMR

Scheme 2. Synthesis of acid 14. Reagents and conditions: a) NaHMDS, THF,
-78 °C, 10 min, then BrCH2CO2tBu, -78 °C, 45 min, 85%; b) BnOH, nBuLi, THF, 0 °C, 45 min, add 13, -78 to 0 °C, then LiOH, MeOH, water, THF, 0 °C to RT, 16 h, 85%.

The failure of this seemingly simple amide bond forming reaction is most likely attributable to the amine 16 being both sterically encumbered, and of reduced nucleophilicity due to the electron
analysis of the crude reaction mixture indicated the [3.3.0]-bicyclic γ-lactone 23a was formed as an 8:1 mixture along with what was assumed to be diastereomer 23b (based on previous experience).8g

Chemistry – A European Journal

FULL PAPER
10.1002/chem.201800046

minimise

tBuO

C
2
O

N
H
E

[O]

E
E
•

N
H
O

tBu
2

HN
•
E
E
O

tBu
2

O
HN
H
E
•
E

H
A1,3 -strain

O
HN

E
H

H
•

H
[O] and hydrolysis

tBuO
O

C
2

H
H
N
CO

O
Me
2
O

20; E = CO
Me
2

E
[O]

N
H

O
25a; s-cis

CO2tBu

AcO or water
25b; s-trans

N
H

CO Me
2
CO Me
2
OAc
26; R = CH

or
CO
2
tBu
2

N
H

Me
2
CO Me
2
27; R = CH2CO2tBu

tBuO2C

H
N

Me
2
O
23a

tBuO
C
2
tBuO
C
2
H H
O

28 22 21 H H

23a

Scheme 5. Plausible mechanism for formation of products 21, 22 and 23a under oxidative radical conditions and 1H NMR nOes for compound 23a.
= nOe

The configuration of the major diastereomer 23a was assigned by 1H NMR nOe experiments (Scheme 5). The formation of the products 21, 22 and 23, most likely arises by oxidation of the substrate 20 with MnIII to generate the captodative C-centered

The next step in the synthetic pathway required differentiation of the γ-lactone carbonyl group from the remaining carbonyl groups in 23a. Our synthetic strategy was to open the γ-lactone with the highly nucleophilic phenylselanyl anion. Unsurprisingly exposure

radical 25 (Scheme 5).[17, 18] In order for 5-exo-trig cyclization to of the [3.3.0]-γ-lactone 23a to the anion formed on reduction of

occur from 25 it is necessary for this radical to adopt an s-trans conformation 25b. Secondary amides are well known to adopt an s-cis geometry preferentially and it is likely that the corresponding radical s-cis 25a will also be the lower energy conformation compared with s-trans 25b. The s-trans radical 25b can undergo 5-exo-trig cyclization from the pre-transition state assembly 26 with the carboxymethylene side chain occupying a pseudo- equatorial position in the chair-like transition state.[19] The adduct radical 27 then undergoes oxidation and lactonization to give the product [3.3.0]-bicyclic γ-lactone 23a. Alternatively, competitive oxidation of the captodative radical 25 to the corresponding iminium ion (or imine) 28 may occur followed by trapping with acetate anion or water giving 21 and 22.[20] The stability of the “tetrahedral intermediates” 21 and 22 is undoubtedly a result of
diphenyl diselenide with sodium borohydride,[22] resulted in substitution at the methyl ester to give the lactone acid 24 (Scheme 4);[23] clearly a more sterically demanding ester group was required.

Synthesis – Second Generation
The synthesis of the cyclization substrate 20, and successful cyclization to give the bicyclic γ-lactones 23 had demonstrated the feasibility of oxidative radical methodology to form [3.3.0]-bicyclic γ-lactones from terminal alkene-containing secondary
amidomalonates. However, it was clear that a change of malonate ester was required and we elected to use a di-tert-butyl malonate as Danishefsky had demonstrated that tert-butyl esters were compatible with lactone opening by the phenylselanyl

them being the formal addition products to the tricarbonyl anion.[6c],[23] Additionally, we sought a shorter synthesis of the

compound dimethyl ketomalonate.[21]
acid coupling partner 10. As shown in the retrosynthetic analysis (Scheme 1) we intended to synthesize the cyclization substrates

N
H
E

a)Mn(OAc) Cu(OTf)
2

,
3
O

N
H
E

+
O

N
H
E

E
OAc
by direct amide bond formation between the enantiopure β,γ- unsaturated acid 10 and aminomalonates represented by 11b. We were fully aware that this might be a challenging amide bond

tBuO
C
2
20; E = CO
Me
2
tBuO
C
2
21

+
tBuO
C
2
22
formation due to the reduced nucleophilicity of the amine 11b and the distinct possibility of epimerization of the stereocenter in 10

O
H
N
CO
H
2
O

b)PhSeNa
O
H
N

4
CO
Me
2
O

+
O
H
N
CO
Me
2
O
during amide bond formation. Our first task was to synthesize the acid, represented by 10, in enantiopure form.

H
O
H
O
H
OOppolzer has reported that treatment of the enolate derived from

tBuO
C
2
tBuO
C
2
tBuO
C
2
the β,γ-unsaturated
sultam 29 (Scheme 6) with benzyl bromide

24
23a
23b
gave the corresponding α-benzylated product in good yield (80%)[24] and excellent diastereoselectivity. Additionally, the

Scheme 4. Initial cyclizations. Reagents and conditions: a) Mn(OAc)3, Cu(OTf)2, MeCN, reflux; 23a, 48%; 23b, 5% (NMR yield); 21, 5-10% (NMR yield); 22, 5- 10% (NMR yield); b) (PhSe)2, NaBH4, DMF, 100 °C then add lactone 23a, 50 °C, 46%.
deconjugative methylation of the enolate derived from the sultam
31has been reported by Golec.[25] Following these precedents, treatment of the crotonyl substituted sultam 31 with LiHMDS followed by the addition of tert-butyl bromoacetate gave the sultam 32 in 88% yield and >20:1 diastereomeric ratio as a white crystalline solid. Hydrolysis using hydroperoxide anion[26] gave the desired carboxylic acid 33. The auxiliary 30 was separated

Chemistry – A European Journal

FULL PAPER

from 33 by trituration with hexane giving the acid in 91% yield (90% purity, 98.8% ee). The next challenge was to develop a racemization free amide bond forming reaction between the chiral non-racemic acid 33 and the amine 34. Initial trials using HATU with either DIPEA or N-methylmorpholine as added base resulted in the product 35 being isolated in 70-75% ee. Ringling recently advocated the use of T3P (propylphosphonic anhydride) as a mild reagent for racemization free amide bond formation.[27] Following Ringling’s procedure, with a reaction temperature of 0 °C, gave the amide 35 with 96% ee, while lowering the temperature to -40 °C gave the desired product 35 in 65% yield (98.4% ee) from the sultam 32. This short synthetic sequence allowed for the preparation of multigram quantities of the cyclization substrate 35 with high yield and enantiomeric purity.
10.1002/chem.201800046

5). Although the influence of water was not profound, the best yield of bicyclic lactone (±)-36 was found using a 50:3 ratio of acetonitrile to water (entry 4). Pleasingly, these conditions were transferable to gram scale with ready separation of the C-2 lactone diastereomers of 36 being achieved (entry 6). During optimization of the cyclization the reaction was also conducted on 5 g of enantiopure material using a 9:1 ratio of acetonitrile:water which gave the bicyclic lactone in 67% isolated yield (entry 7). Although it is not immediately clear as to the reason for water influencing the product distribution in the reaction of 35, we postulate that the amount of water affects the oxidation potential of the copper(II) salt which in turn could influence the product distribution.

tBuO
C
2
O
1Mn(OAc) ,
3
Cu(BF ) •6H
4 2
O
2
O
H
N
CO
tBu
2
O
O
H
N
CO
tBu
2

Me Me
Me Me
tBuO
C
2
N
H
2
+
CO
tBu
2

a)NaH then crotonyl chloride
O

(±)-35
CO
tBu
2
tBuO
C
2
H
(±)-36
O
tBuO
C
2

(±)-37

S O2
NH
82%
S O2
N
Me
+
tBuO
C
2
O

30 31

b)LiHMDS, BrCH CO tBu
2 2
88%
tBuO
C N
2
AcO H
(±)-38

tBu
2

Me Me
Table 1. Optimization of the cyclization of substrate (±)-35 with respect to water content.

O O
c)LiOH, H O
2 2
HO N S
91%
O2
33 CO tBu CO tBu
2 32 2
tBuO C
2
d)34
tBuO C NH
2 2
T3P, pyridine
Me Me
65% (2 steps)
O
tBuO C O
2
N tBuO C N S
2
H O2 29
35 CO tBu
2

Scheme 6. Synthesis of amine 35. Reagents and conditions: a) NaH, toluene, 0 °C to RT then add crotonyl chloride, 0 °C to RT, 2 h, 82%; b) LiHMDS, THF, HMPA, -78 °C, 30 min then BrCH2CO2tBu, THF -78 °C, 7 h, 88%; c) 30% H2O2, LiOH, THF, water, 5 min, 91%; d) 34, T3P {[CH3CH2CH2OP(O)]3}, pyridine, EtOAc, -40 to -10 °C, 6 h, 65% from 32.

Cyclization Optimization
Having developed an efficient synthesis of the cyclization substrate optimization of the oxidative radical cyclization was required. Initial studies using racemic substrate (±)-35 with 2 equivalents of Mn(OAc)3 and 1 equivalent of Cu(OTf)2 in acetonitrile at 80 °C gave the desired [3.3.0]-bicyclic γ-lactones (±)-36 in 50-60% yield as a >8:1 mixture of C-2 diastereomers along with the oxidized and uncyclized material (±)-38. Changing the amount of CuII salt had little influence on the outcome of the reaction. We moved to using cheaper Cu(BF4)2 and found that the amount of water in the reaction mixture had an influence on the product distribution of the reaction (Table 1). With no added water in the reaction the desired lactones (±)-36 were formed as a 9:1 mixture along with the methylene pyrrolidinone (±)-37 (Table 1, entry 1). Addition of water lead to increasing quantities of the oxidized, but uncyclized malonate (±)-38 being formed (entries 2-
[a] Mn(OAc)3•2H2O (3.0 equiv.), Cu(BF4)2•6H2O (0.3 equiv.), MeCN/H2O (see column 2 for proportions), 105 °C (oil bath temperature), 30 min, 500 mg and 0.25 M in substrate. [b] mol% from crude 1H NMR spectrum. [c] Total isolated yield of an 8.5-9.5:1 mixture of C-2 diastereomers (natural product numbering, major diastereomer shown). [d] Reaction conducted on 4.06 g, 9.82 mmol of material of 82% ee. [e] Reaction conducted on 5.93 g, 14.4 mmol of material of
>98% ee. [f] n.d. = not determined. [g] Yield of isolated pure single (2R)- diastereomer (+)-36.

Optimization of the cyclization had given us access to gram quantities of the enantiopure [3.3.0]-bicyclic γ-lactone (+)-36. The next step in the synthesis involved opening of the g-lactone by phenylselanyl anion. Attempted opening of the γ-lactone in (+)- 36 by exposure to phenylselanyl anion, formed by the reduction of diphenyldiselenide with sodium borohydride, in DMF at 100 ºC led to decomposition whereas no reaction was observed at 50 ºC. Liotta showed that a non-complexed, and hence more nucleophilic anion is formed by reduction of diphenyl diselenide by sodium metal, or deprotonation of benzeneselenol with sodium

Chemistry – A European Journal

FULL PAPER
10.1002/chem.201800046

hydride.[28] Pleasingly exposure of the enantioenriched γ-lactone (+)-36 to phenylselanyl anion prepared by reduction of diphenyl diselenide with sodium hydride[29] allowed reaction at room

H HO H
O
NCO2tBu N
OH a) PhSeBr, AgBF4 H

H
2

temperature to give the fully substituted pyrrolidinone 39 after acid
tBuO
C
2
O
SePh

base extraction (Scheme 7). Reduction of the carboxylic acid in 39 to the corresponding primary alcohol 40 proved problematic with direct reduction using borane resulting in substrate
41
O
42

b) PMBCl 84% from 41

decomposition and attempted reduction of the corresponding mixed anhydride with sodium borohydride resulting in numerous

H
N
HO

uncharacterized products being formed.

OH tBu O HN
N CO a) PhSeNa
2
O

tBu
2
H

44
CO
Me
PMB
2
c) Bu3SnH 93%
H

43
CO2PMB

SePh

tBuO

C
2

H
(+)-36
O

tBuO

C
2

39
CO2H SePh

Scheme 8. Synthesis of bicyclic lactone 44. Reagents and conditions: a) PhSeBr, AgBF4, CH2Cl2, CH3CN, RT; b) 4-CH3O(C6H4)CH2Cl, K2CO3, DMF,

b)(COCl)
, DMF
2
81%
40 °C, 84% from 41; c) Bu3SnH, AIBN, toluene, 105 °C, 93%.

then LiAlH(Ot-Bu)3 [from (+)-36]

As part of optimizing the synthetic route we had prepared the tert-

O
H
N
CO
tBu
2
OH
c)NaIO ,
4
then heat quant.
O
H
N
CO
tBu
2
OH
butyl ester (±)-45 (inset Scheme 9) in racemic form and we used this material to investigate the oxidation and cyclohexenylation

tBuO
C
2

40
SePh
reaction following the excellent precedent from Corey.6a,b We initially investigated the formation of the aldehyde derived from oxidation of (±)-45. Using standard reagents such as the Dess-

Scheme 7. Synthesis of pyrrolidinone 41. Reagents and conditions: a) (PhSe)2, NaH, THF, 65 °C, 90 min, then (+)-36, 18-crown-6, 0 °C then RT, 3 h; b) (COCl)2, DMF, then LiAlH(OtBu)3, THF, MeCN, -78 °C to RT, 81% from (+)-36; c) NaIO4, NaHCO3, THF, MeOH, water, RT, then CHCl3, reflux, quant.

Ultimately, we found that conversion of the acid 39 into the corresponding acid chloride followed by reduction with lithium tri- tert-butoxy aluminum hydride[30] gave the primary alcohol 40 in 81% yield from the bicyclic lactone (+)-36. Elimination of the phenylselanyl group from 40 was achieved by oxidation and thermal treatment to give the alkene 41 in quantitative yield. All that remained for the synthesis of the fully functionalized pyrrolidinone core of salinosporamide A 1 was installation of the C-3 tertiary hydroxyl/alkoxy group.
The lactonization of the ester alkene 41 (for example to give 42 Scheme 8) was crucial to our synthetic strategy. After extensive experimentation, we found that lactonization could be achieved using the phenylselanyl cation as reported by Danishefsky for a closely related selenocyclo-acetalization.6c Thus, exposure of the alkene 41 to phenylselanyl bromide and silver(I) tetrafluoroborate led to lactonization with concomitant loss of the tert-butyl ester to give the γ-lactone 42 (νmax 1780 cm-1) in 90% yield (Scheme 8); the free carboxylic acid was readily converted into the corresponding PMB ester 43 and the phenylselanyl group was reduced under radical conditions to give 44. Stereoselective introduction of the cyclohexenyl side-chain, conversion of the γ- lactone into the C-2 chloroethyl group and β-lactone formation were now required to complete the synthesis of salinosporamide A (1).
Martin periodinane, pyridinium chlorochromate, or DMSO with pyridinesulfur trioxide, followed by the usual aqueous workup, either led to no reaction (PCC) or to decomposition. This was surprising given that in a number of previous syntheses of salinosporamide A (1), oxidation of related primary alcohols to the corresponding aldehydes had been readily achieved using DMP; however, in all of these cases, the pyrrolidinone nitrogen atom was protected with either a benzyl-type protecting group or a carbamate protecting group. Interestingly we found that 1H NMR analysis of the oxidation of (±)-45 by DMP in d2-dichloromethane showed clean conversion to the corresponding aldehyde (not shown) in under one hour. Addition of water to the reaction mixture followed by 1H NMR analysis showed clear decomposition of the aldehyde, as did direct elution of the reaction mixture through silica or basic alumina. In a similar manner, 1H NMR analysis of the DMP oxidation of alcohol 44, in both d2- dichloromethane and d8-THF demonstrated clean conversion to
the corresponding aldehyde (not shown). We therefore developed a one-pot oxidation, alkylation procedure. Treatment of the alcohol 44 with 2 equivalents of DMP in THF at ambient temperature followed by cooling of the reaction mixture to -78 °C and addition of 10 equivalents of cyclohexenylzinc bromide 467c gave the desired product 47 as a single diastereomer in 56% yield from the alcohol 44 (Scheme 9). This reaction was conducted with
>900 mg of alcohol 44 giving >600 mg of the cyclohexenyl- substituted product 47. Reduction of the γ-lactone in 47 to the corresponding diol 48 in the presence of the lactam and PMB- ester was the next challenge. Lam8b had shown that reduction of the γ-lactone in the N-PMB protected, methyl ester analogue of 47, occurred in 60% yield on treatment with two equivalents of

Chemistry – A European Journal

FULL PAPER
10.1002/chem.201800046

Me Me

tBuO

C
2

a)NaH then
O
b)LiHMDS,
O
c)LiOH,
O
tBuO
C
2
NH2
tBuO
C
2
O

crotonyl chloride
BrCH
CO
2
tBu
2
H
O
2
2
T3P, pyridine

S
NH
82%
S
N
Me
88%
S
N
HO
65% (2 steps)
tBuO
C
2
N
H

O2 30
O2 31
O2 32
CO
tBu
2

33
CO
tBu
2
e)Mn(OAc) Cu(BF )
4 2
71%
,
3

35
CO
tBu
2

O
H

H
N

H
2

i) PhSeBr, AgBF
4

H
N

tBu
2
OH

h) NaIO ,
4
then heat

H
N

tBu
2
OH

g) (COCl) , DMF
2
then LiAlH(Ot-Bu)

H
N

CO tBu
2
CO H
2

f)PhSeNa

H
N

tBu
2
O

O
SePh
tBuO
C
2
quant.
tBuO
C
2
SePh
81% (2 steps)
tBuO
C
2
SePh
tBuO
C
2
H
O

O
42
j)PMBCl 84% (2 steps)
41
40
39; 4.87 g prepared
(+)-36; 9:1 dr*

H H H
l) DMP then

O
H
H
N
OH

PMB
2

k)Bu

SnH
3
O
H
H
N
HO

PMB
2

ZnBr
46
O
H
H
N

OH CO PMB
2
m) DIBAl-H, then NaBH
4
O
H
N

OH CO PMB
2
n)BCl
3
o)BOPCl
p)Ph PCl
3

2
O
H
N

OH
O

SePh
93%

O
Me
56%

O
Me
69%
Me
OH
61% (3 steps)
Me
O

O
43; 2.31 g prepared
O
44
O
47; 645 mg prepared
HO
48; 140 mg prepared
Cl
1; 30 mg prepared

Scheme 10. Total synthesis of salinosporamide A 1. Reagents and conditions: a) NaH, toluene, RT then add crotonyl chloride, 0 °C to RT, 2 h, 82%; b) LiHMDS, THF, HMPA, -78 °C, 30 min then BrCH2CO2tBu, THF -78 °C, 7 h, 88%; c) 30% H2O2, LiOH, THF, water, 10 min, 91%; d) 34, T3P {[CH3CH2CH2P(O)O]3}, pyridine, EtOAc, -40 to -10 °C, 6 h, 65% from 32; e) Mn(OAc)3•2H2O (3.0 equiv.), Cu(BF4)2•6H2O (0.3 equiv.), MeCN, water, 71% (desired diastereomer); f) (PhSe)2, NaH, THF, 65 °C, 90 min, then (+)-36, 18-crown-6, 0 °C then RT 3 h; g) (COCl)2, DMF, then LiAlH(OtBu)3, THF, MeCN, -78 °C to RT, 81% from (+)-36; h) NaIO4, NaHCO3, THF, MeOH, water, RT, then CHCl3, reflux, quant; i) PhSeBr, AgBF4, CH2Cl2, CH3CN, RT; j) 4-CH3O(C6H4)CH2Cl, K2CO3, DMF, 40 °C, 84% from 41; k) Bu3SnH, AIBN, toluene, 105 °C, 93%. l. Dess-Martin periodinane, THF, RT, then cyclohexenylzinc bromide 46, -78 °C, 56%; m) DIBAl-H, THF, -10 °C, then MeOH, then NaBH4, MeOH/THF, RT, 69%; n) BCl3, CH2Cl2, 0 °C. o. bis(2-oxo-3-oxazolidinyl)phosphinic chloride (BOPCl), CH2Cl2, pyridine, then further BOPCl, RT; p) Ph3PCl2, CH3CN, pyridine, RT, 61% (3 steps).

sodium borohydride in methanol to give the corresponding triol. Using Lam’s procedure on 47 gave incomplete reduction of the γ- lactone and a number of other products were formed as evidenced by TLC analysis.
We found that exposure of the γ-lactone to DIBAl-H led to rapid reduction (<20 min) of the γ-lactone to corresponding lactol (LRMS analysis), which on addition of methanol and excess sodium borohydride gave the triol 48. The PMB ester in 48 was readily removed by treatment with boron trichloride and the total synthesis was completed using the method of Corey,6a namely β-

O
H

H
N

CO
Me

PMB
2

a)DMP then
ZnBr
46
56%

O
H

H
N

O
H

OH CO PMB
2
Me
lactone formation using BOPCl and chlorination with triphenylphosphine dichloride to give salinosporamide A (1). The analytical data for our synthetic salinosporamide A were in excellent agreement for that of both the natural and previously synthesized material.

O
44
O
47
O
H
N
HO
A summary of our synthesis is shown in Scheme 10. A number

b)DIBAl-H, then NaBH4
69%
H
CO
Me
tBu
2
of points are worthy of comment. The synthesis allows the production of good quantities of key intermediates. We have

O
O

(±)-45
prepared gram quantities of the carboxylic acid 39 which were transformed into gram quantities of the [3.3.0]-bicyclic γ-lactone

O
H
N

OH
c)BCl
3
d)BOPCl
e)Ph PCl
3

O
H
N

OH
43. From intermediate 43 we have prepared 645 mg of the fully elaborated pyrrolidinone 47. We ultimately synthesized 30 mg of

O
O
61%

OH
CO
Me
PMB
2
the natural product although our route would undoubtedly allow us to prepare significantly more. The route proceeds in 16 steps

Cl
1
HO
48
(5% yield) from the commercially available sultam 30 (15 steps and 6% from commercially available sultam 31). For comparison,

Scheme 9. Completion of the synthesis. Reagents and conditions: a) Dess- Martin periodinane, THF, RT, then cyclohexenylzinc bromide 46, -78 °C, 56%; b) DIBAl-H, THF, -10 °C, then MeOH, then NaBH4, MeOH/THF, RT, 69%; c) BCl3, CH2Cl2, 0 °C; d) bis(2-oxo-3-oxazolidinyl)phosphinic chloride (BOPCl), CH2Cl2, pyridine, then further BOPCl, RT; e) Ph3PCl2, CH3CN, pyridine, RT, 61% (3 steps).
Corey’s synthesis of 1 proceeds in 17 steps (9.8%) from threonine, Fukuyama’s synthesis of 1 proceeds in 14 steps (19%) from 4- pentenoic acid, and Romo’s bioinspired synthesis of 1 in 90% ee, proceeds in 9 steps (3.6%) from serine. The key features of the synthesis reported above include: the use of an oxidative radical

Chemistry - A European Journal

FULL PAPER

cyclization for the synthesis of the key [3.3.0]-bicyclic

γ-lactone
10.1002/chem.201800046

Fukuda, M. Tomizawa, T. Masaki, M. Shibuya, N. Kanoh, Y. Iwabuchi,

(+)-36 with good diastereocontrol, a selenolactonization to set the required C-3 tertiary-alkoxy stereocenter giving 42, and the use of Corey’s method for diastereoselective introduction of the cyclohexenyl side chain to give 47. Other notable aspects of our synthesis include the scalability of the route and the limited use of protecting groups as demonstrated by the unprotected amide/lactam NH being carried through the whole synthetic sequence.

Conclusions

In conclusion, we have developed a short enantioselective synthesis of the potent proteasome inhibitor salinosporamide A. Work is ongoing to synthesize more complex, biologically active, pyrrolidinone natural products using our oxidative radical cyclization methodology.

Experimental Section
Heterocycles 2010, 81, 2239-2246; h) H. Nguyen, G. Ma, D. Romo, Chem. Commun. 2010, 46, 4803-4805; i) N. Satoh, S. Yokoshima, T. Fukuyama, Org. Lett. 2011, 13, 3028-303; j) Y. Kaiya, J. Hasegawa, T. Momose, T. Sato, N. Chida, Chem. Asian. J. 2011, 6, 209-219; k) H. Nguyen, G. Ma, T. Gladysheva, T. Fremgen, D. Romo, J. Org. Chem. 2011, 76, 2-12.
[7]For total syntheses of 1 in racemic form see: a) N. P. Mulholland, G. Pattenden, I. A. S. Walters, Org. Biomol. Chem. 2006, 4, 2845-2846; b) G. Ma, H. Nguyen, D. Romo, Org. Lett. 2007, 9, 2143-2146; c) N. P. Mulholland, G. Pattenden, I. A. Walters, Org. Biomol. Chem. 2008, 6, 2782-2789.
[8]For formal syntheses of 1 see: a) V. Caubert, J. Massé, P. Retailleau, N. Langlois, Tetrahedron Lett. 2007, 48, 381-384; b) I. V. Margalef, L. Rupnicki, H. W. Lam, Tetrahedron 2008, 64, 7896-7901; c) T. Momose, Y. Kaiya, J. Hasegawa, T. Sato, N. Chida, Synthesis 2009, 2983-2991; d) R. A. Mosey, J. J. Tepe, Tetrahedron Lett. 2009, 50, 295-297; e) J. R. Struble, J. W. Bode, Tetrahedron 2009, 65, 4957-4967; f) T. T. Ling, B. C. Potts, V. R. Macherla, J. Org. Chem. 2010, 75, 3882-3885; g) A. W. J. Logan, S. J. Sprague, R. W. Foster, L. B. Marx, V. Garzya, M. S. Hallside, A. L. Thompson, J. W. Burton, Org. Lett. 2014, 16, 4078-4081.
[9]For recent reviews containing synthetic routes to salinosporamide A see:
a)M. Shibasaki, M. Kanai, N. Fukuda, Chem. Asian. J. 2007, 2, 20-38;
b)T. A. M. Gulder, B. S. Moore, Angew. Chem. Int. Ed. 2010, 49, 9346- 9367; Angew. Chem. 2010, 122, 9734-9556; c) B. C. Potts, K. S. Lam,

Supporting Information. Experimental procedures; spectroscopic and analytical data for all new compounds including copies of NMR spectra.

Acknowledgements

We thank the EPSRC for funding this work. The John Fell Oxford University Press (OUP) Research Fund is gratefully acknowledged for an equipment grant. We thank the analytical sections of our Department for their excellent support. We are grateful to Prof. Danishefsky for providing spectroscopic data.

Keywords: Total synthesis • Natural products • Radical reactions

[1]R. H. Feling, G. O. Buchanan, T. J. Mincer, C. A. Kauffman, P. R. Jensen, W. Fenical, Angew. Chem. Int. Ed. 2003, 42, 355-357; Angew. Chem. 2003, 115, 369-371.
[2]Throughout the manuscript solid and broken bold lines will be used for racemic compounds with two or more stereocenters and solid and broken wedges will be used for enantiopure compounds, see: H. Maehr, J. Chem. Educ. 1985, 62, 114-120.
[3]M. Groll, R. Huber, B. C. M. Potts, J. Am. Chem. Soc. 2006, 128, 5136- 5141.
[4]W. Fenical, P. R. Jensen, M. A. Palladino, K. S. Lam, G. K. Lloyd, B. C. Potts, Biorg. Med. Chem. 2009, 17, 2175-2180.
[5]T. A. M. Gulder, B. S. Moore, Angew. Chem. Int. Ed. 2010, 49, 9346- 9367.
[6]a) L. R. Reddy, P. Saravanan, E. J. Corey, J. Am. Chem. Soc. 2004, 126, 6230-6231; b) L. R. Reddy, J. F. Fournier, B. V. S. Reddy, E. J. Corey,
Marine Drugs 2010, 8, 835-880; d) A. Rentsch, D. Landsberg, T. Brodmann, L. Bulow, A. K. Girbig, M. Kalesse, Angew. Chem. Int. Ed. Engl. 2013, 52, 5450-5488; Angew. Chem. 2013, 125, 5560-5599.
[10]D. A. Evans, M. D. Ennis, D. J. Mathre, J. Am. Chem. Soc. 1982, 104, 1737-1739.
[11]W. Hess, J. W. Burton, Chem. Eur. J. 2010, 16, 12303-12306.
[12]B. Barlaam, T. G. Bird, C. Lambert-van der Brempt, D. Campbell, S. J. Foster, R. Maciewicz, J. Med. Chem. 1999, 42, 4890-4908.
[13]L. A. Carpino, D. Sadat-Aalaee, H. G. Chao, R. H. DeSelms, J. Am. Chem. Soc. 1990, 112, 9651-9652.
[14]A number of amide bond forming conditions were trialled including: TBTU, i-Pr2NEt; BOPCl, i-Pr2NEt; DCC, DIPEA; T3P, Et3N; HATU, i-Pr2NEt; DPPA, Et3N; PivCl, Et3N.
[15]P. Kocienski, P. Raubo, J. K. Davis, F. T. Boyle, D. E. Davies, A. Richter, J. Chem. Soc., Perkin Trans. 1 1996, 1797-1808.
[16]For reviews of manganese(III) acetate in organic synthesis see: a) B. B. Snider, Chem. Rev. 1996, 96, 339-363; b) G. G. Melikyan, Org. React. 1997, 49, 427-675; c) A. S. Demir, M. Emrullahoglu, Curr. Org. Synth. 2007, 4, 321-351; d) J. W. Burton (2012) in Encyclopedia of Radicals in Chemistry, Biology and Materials, C. Chatgilialoglu and A. Studer (eds). John Wiley & Sons Ltd, Chichester, UK, pp 901-942; e) M. Mondal, U. Bora, RSC Advances 2013, 3, 18716-18754.
[17]For a review of the mechanisms of manganese(III) acetate-mediated reactions see: B. B. Snider Tetrahedron, 2009, 65, 10738-10744.
[18]For some examples of the use of manganese(III) acetate in systems reacting with intermediate captodative radicals, see: (a) A. Citterio, A. Marion, A. Maronati, M. Nicolini, Tetrahedron Lett., 1993, 34, 7981–7984; (b) P. J. Crocker, U. Karlsson-Andreasson, B. T. Lotz, M. J. Miller, Heterocycles, 1995, 40, 691–716; (c) P. J. Crocker, M. J. Miller, J. Org. Chem., 1995, 60, 6176-6179.
[19]The pre-transition state assembly is a modified version of the Beckwith- Houk model for the cyclization of 5-hexenyl radicals: a) A. L. J. Beckwith, C. H. Schiesser, Tetrahedron Lett. 1985, 26, 373-376; b) K. N. Houk, M.

Org. Lett. 2005, 7, 2699-2701; c) A. Endo, S. J. Danishefsky, J. Am. Chem. Soc. 2005, 127, 8298-8299; d) T. T. Ling, V. R. Macherla, R. R. Manam, K. A. McArthur, B. C. M. Potts, Org. Lett. 2007, 9, 2289-2292; e) K. Takahashi, M. Midori, K. Kawano, J. Ishihara, S. Hatakeyama, Angew. Chem. Int. Ed. 2008, 47, 6244-6246; Angew. Chem. 2008, 120, 6340-6342; f) T. Fukuda, K. Sugiyama, S. Arima, Y. Harigaya, T. Nagamitsu, S. Omura, Org. Lett. 2008, 10, 4239-4242; g) Y. Sato, H.
N. Paddonrow, D. C. Spellmeyer, N. G. Rondan, S. Nagase, J. Org. Chem. 1986, 51, 2874-2879.
[20]If the reaction is conducted at 40 °C the oxidized, uncyclized products predominate, in keeping with the s-cis radical 25a being more stable than the s-trans radical 25b.
[21]H. Asahara, K. Inoue, S. Tani, K. Umezu, N. Nishiwaki, Adv. Synth. Catal. 2016, 358, 2817-2828.

Chemistry - A European Journal

FULL PAPER

[22]R. M. Scarborough, B. H. Toder, A. B. Smith, J. Am. Chem. Soc. 1980, 102, 3904-3913.
[23]Attempted chemoselective lactone opening with phenylselanyl anion of the ethyl, i-propyl and benzyl esters corresponding to 23a was also unsuccessful.
[24]W. Oppolzer, R. Moretti, S. Thomi, Tetrahedron Lett. 1989, 30, 5603- 5606.
[25]M. J. Batchelor, R. J. Gillespie, J. M. C. Golec, C. J. R. Hedgecock, S. D. Jones, R. Murdoch, Tetrahedron 1994, 50, 809-826.
10.1002/chem.201800046

[26]D. A. Evans, T. C. Britton, J. A. Ellman, Tetrahedron Lett. 1987, 28, 6141- 6144.
[27]J. R. Dunetz, Y. Xiang, A. Baldwin, J. Ringling, Org. Lett. 2011, 13, 5048- 5051.
[28]D. Liotta, U. Sunay, H. Santiesteban, W. Markiewicz, J. Org. Chem. 1981, 46, 2605-2610.
[29]V. Rodeschini, J.-G. Boiteau, P. Van de Weghe, C. Tarnus, J. Eustache, J. Org. Chem. 2004, 69, 357-373.
[30]H. C. Brown, R. F. McFarlin, J. Am. Chem. Soc. 1958, 80, 5372-5376.

Chemistry - A European Journal

FULL PAPER

Entry for the Table of Contents (Please choose one layout) Layout 1:

FULL PAPER

Text for Table of Contents

((Insert TOC Graphic here: max. width: 5.5 cm; max. height: 5.0 cm))

Layout 2:

FULL PAPER

Herein we report a scalable total synthesis of the potent proteasome inhibitor salinosporamide A that proceeds in 16 steps and 5% overall yield and features an oxidative radical cyclization as a key step.
10.1002/chem.201800046

Author(s), Corresponding Author(s)*

Page No. – Page No. Title

Léo B. Marx and Jonathan W. Burton*

Page No. – Page No.
A Total Synthesis of Salinosporamide A

Peptide Solubility

The prohormones are then packaged into membrane-bound secretory vesicles.

Leave a Reply Cancel reply