A New Approach to Models of the 4,5-Dihydroxycyclopentenone Core of the Kodaistatins A–D: Elucidation of the Diol Configuration in Kodaistatin A

Article abstract of DOI:10.1002/chem.201701185

The kodaistatins A–D are strongly anti-diabetic natural products from Aspergillus terreus that hold some promise of a novel diabetes cure. However, considerations of that kind face two drawbacks: 1) The kodaistatins A–D contain a heavily substituted pulvinone/cyclopentenone combination; 2) they are 1,2-diols, the 3D structures of which have not been assigned yet. However, we can exclude two of the four possible stereostructures. We conclude that kodaistatin A is a trans-, not a cis-diol from NMR comparisons with a pair of cis, trans-isomeric kodaistatin models, which we synthesized in 11 and 12 steps, respectively. The stereocenters of the diol moiety arose from stereocomplementary, highly diastereoselective aldol additions of a lithium enolate or the corresponding silyl ketene acetal. The cyclopentenone moieties stemmed from intramolecular aldol additions and ensuing dehydrations. The requisite enolates were obtained by the reduction of α-bromoketones with samarium diiodide.

Full text of DOI:10.1002/chem.201701185

A Journal of
Accepted Article
Title: A New Approach to Models of the 4,5-Dihydroxycyclopentenone
Core of the Kodaistatins A-D. Elucidation of the Diol
Configuration in Kodaistatin A
Authors: Reinhard Brückner and David Peter
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201701185
Link to VoR: http://dx.doi.org/10.1002/chem.201701185
Supported by

10.1002/chem.201701185
Chemistry - A European Journal
COMMUNICATION
Entry for the Table of Contents
COMMUNICATION
2nd generation
kodaistatin models:
David Peter, Reinhard Brückner*
Li enolate
O
Ph
O
or
Page No. – Page No.
O
O
Ph
Mukaiyama
aldol addition;
SmI2-mediated
aldol addition;
O
O
HO
CO2Me
O
A New Approach to Models of the 4,5-
Dihydroxycyclopentenone Core of the
Kodaistatins A–D. Elucidation of the
Diol Configuration in Kodaistatin A
+
*
*
3 follow-
up steps
4 follow-
up steps
TBSO
HO
O
Br
O
Cl
syn or anti
(choosable)
cis or trans
(choosable)
Aldols are the key: We determined the relative configuration in the diol moiety of
the natural product kodaistatin A, an anti-diabetic lead, being trans. This follows
from NMR comparisons with cis,trans-isomeric models, which we synthesized in
11-12 steps. Early-stage aldolizations ensured stereocomplementarity by using
either a lithium enolate or an enol silane. Cyclopentanone formations by intra-
molecular aldol additions of a samarium enolates ensued.
1
This article is protected by copyright. All rights reserved.

10.1002/chem.201701185
Chemistry - A European Journal
COMMUNICATION
A New Approach to Models of the
4,5-Dihydroxycyclopentenone Core of the Kodaistatins A–D.
Elucidation of the Diol Configuration in Kodaistatin A
David Peter[a] and Reinhard Brückner*[a]
Dedicated to Professor Lutz F. Tietze at the occasion of his 75th birthday
Abstract: The kodaistatins are natural products from Aspergillus
terreus. Being strongly anti-diabetic they hold some promise of a
novel diabetes cure. However, considerations of that kind face two
drawbacks. (1) The kodaistatins contain a heavily substituted pulvi-
the side-chain stereocenter of kodaistatin A (1).[6] They ozono-
lyzed the natural product, worked up reductively, and identified
the resulting alcohol S-5 (Figure 1b) by chiral GLC. Wüster im-
plemented this finding in a synthesis of the “first generation ko-
none/cyclopentenone combination. (2) The kodaistatins are 1,2-diols, daistatin model” cis-6 (Figure 1b).[6] Its cis-dihydroxycyclopen-
whose 3D structure have not been assigned yet. However, we can
exclude two of the four possible stereostructures. We conclude that
kodaistatin A is a trans-, not a cis-diol, from NMR comparisons with
a pair of cis,trans-isomeric kodaistatin models, which we synthesized
in 11 and 12 steps, respectively. The stereocenters of the diol moiety
arose from stereocomplementary, highly diastereoselective aldol ad-
ditions of a lithium enolate or the corresponding silyl ketene acetal.
The cyclopentenone moieties stemmed from intramolecular aldol
additions and ensuing dehydrations. The requisite enolates were ob-
tenone moiety displayed 13C NMR resonances, which hinted at
kodaistatin A being trans-dihydroxylated. Unfortunately, this in-
ference was negative evidence because Wüster´s route failed to
provide a “correctly” configured model trans-6 besides the “in-
correctly” configured cis-isomer. Here we make up for that deficit
by synthesizing a “second generation kodaistatin model” 7
(Figure 1c) both as the cis- and the trans-isomer. This succeed-
ed by a completely different strategy.
tained by the reduction of α-bromoketones with samarium diiodide.
Diabetes mellitus has become a major health threat. It affects
415 million people around the world.[1] Patients suffering from
type 2 diabetes exhibit reduced insulin activity.[2] As a conse-
quence, gluconeogenesis in their livers is increased.[ 3 ] This
leads to an excessive release of glucose into the blood stream.[3]
Drugs promoting lower glucose concentrations there might alle-
viate the resulting disorders. The natural products kodaistatin A–
D (1–4; Figure 1a) are potential leads for such a medication.[4,5]
They inhibit the transport protein glucose-6-phosphate T1 trans-
locase in the nM range. This protein is indispensable for trans-
porting glucose-6-phosphate from the cytoplasm into the endo-
plasmatic reticulum of hepatocytes. This is the site of the last
step both of gluconeogenesis and glycogenolysis, the conver-
sion glucose-6-phosphate → glucose + Pi.[2]
The kodaistatins A–D (1–4; Figure 1a) were isolated by a group
from Hoechst Marion Roussel Deutschland GmbH (now Sanofi-
Aventis) from the soil mold Aspergillus terreus.[4,5] They com-
prise a pulvinone unit, a dienone side-chain with one stereocen-
ter, and a dihydroxycyclopentenone core with two stereocenters.
Originally, all stereocenters were left unassigned. Kaczybura
and Wüster from our group determined the S-configuration of
Figure 1. a) Structures of the kodaistatins A–D (1–4).[4,5] b) Degradation prod-
uct S-5[6] obtained from the side-chain of kodaistatin A (1) by ozonolyis and
NaBH4 reduction;[6] our 1st generation kodaistatin model cis-6.[6] c) The 2nd
generation kodaistatin model targets cis- and trans-7 of the present study.[7]
Our route to the second generation kodaistatin models cis- and
trans-7 should entail a cyclopentenone-forming Wittig reaction. It
is sketched as the transformation 8 → 7 in our retrosynthetic
analysis (Scheme 1, at left). Such ring-closures are known.[8]
The phosphonium salt 8 should result from tributyl- or triphenyl-
phosphane and an α-haloketone 9 (Hal = Cl, Br). We deemed to
obtain the latter by a brominating hydrolysis[9] of the chloroolefin
[a]
D. Peter (M. Sc.), Prof. Dr. R. Brückner
Institut für Organische Chemie, Albert-Ludwigs-Universität
Albertstraße 21, 79104 Freiburg (Germany)
E-mail: reinhard.brueckner@organik.chemie.uni-freiburg.de
Homepage: http://www.brueckner.uni-freiburg.de
Supporting information for this article is given via a link at the end of
the document.
2
This article is protected by copyright. All rights reserved.

10.1002/chem.201701185
Chemistry - A European Journal
COMMUNICATION
reactions reported in ref.[10b]
:
retrosynthetic analysis of this work:
Li enolate
aldol addition
5
O
O
O
R
R
R
R
+
4
O
O
CO2Bn
Cl
O
Ph
O
a)
O
CO2Bn
R = Me
Cl
1'
O
HO
HO
*
*
17a
b
12
19a: 29% (d.r. unknown)
b: 46% (d.r. unknown)
HO
R = (CH2)5
13
syn-19[19]
O
7
ylide formation
and Wittig reaction
in situ
O
Li enolate
aldol addition
Ph
O
O
R
R
5
O
O
O
R
R
R
*
O
+
4
O
O
O
CO2Me
*
*
a)
R
O
CO2Me
Cl
1'
Hal
PR3
PBu3 Br
PGO
HO
R = Me
O
17c
d
e
12
19c: 83% (d.r. 95:3:2)
O
Cl
8
14
R = (CH2)
d:
81% (d.r. 92:6:2)
90% (d.r. 94:4:2)
R = (CH2)45
syn-19[19]
SN2 reaction with PR3
e:
O
Ph
O
O
R
R
b)
*
O
O
O
Mukaiyama
aldol addition
*
*
5
O
O
O
R
R
R
R
Br
PGO
Hal
+
4
O
CO2Me
OMe
c)
O
O
O
Cl
1'
15
9
HO
OSiMe3
NaOHal, AcOH
12
Cl
18c R = Me
19c: 52% from ester
d: 52% from ester
e: 43% from ester
(pure diastereomers)
O
Ph
d
e
R = (CH2)
O
O
R
R
R = (CH2)45
anti-19[19]
*
O
1
O
O
2
3
*
*
5
4
PGO
Scheme 2. Aldol additions of esters 17a-e via their lithium enolates (top and
middle) or the derived silyl ketene acetals 18c-e (bottom). a) LDA, THF, −78°C,
15−60 min; 12, −78°C, 3–4 h. b) LDA, THF, −78°C, 30 min; Me3SiCl, −78°C →
RT (within 4 h); product not purified. c) BF3·OEt2, CH2Cl2, −78°C → RT (within
5 h).
Cl
Cl
10
aldol addition (precedents[15, 16]);
then O-protection
16
O
O
R
R
*
*
O
+
Cl
CO2R'
12
As explained above we needed to add an 11-derived C-nucleo-
phile to (Z)-2-chlorobut-2-enal (12[13]) with the highest possible
simple and induced diastereoselectivity. 11-based C-nucleo-
philes, which would react with the opposite simple diastereo-
selectivity should exhibit high induced diastereoselectivity, too.
Scheme 2 shows how we accomplished both goals. We depro-
tonated the benzyl esters 17a (acetonide-protected) and b (cy-
clohexylidene-protected), the corresponding methyl esters 17c
and e, and their analog 17d (cyclopentylidene-protected) with
LDA at −78°C in THF.[18] The resulting enolates were combined
with (Z)-2-chlorobut-2-enal (12[13]). This delivered the aldols 19a-
11
Scheme 1. Left: Joint retrosynthesic analysis of the cis- and the trans-isomer
of the 2nd generation kodaistatin model 7. Right: A related 4-chloropent-4-en-1-
one → cyclopentenone transformation by Stoltz et al.[10b] − Eventually, we
could not progress from compound 9 (Hal = Br) to phosphonium salts 8. This
reduces the overlap with Stoltz´ route to a single step, i. e., the brominating
hydrolysis[9] 10→9. It is akin to 16→15.
10. Related transformations are known.[10] Up to this point, our
retrosynthetic simplification 7 ⇒ 8 ⇒ 9 ⇒ 10 parallels a 3-step
route of Stoltz et al. to the cyclopentenone 13[10b] (Scheme 1, at
right). They treated the chloroolefin 16 with aqueous NaOBr in
acetic acid, allowed the crude bromoketone 15 to react with
PBu3, and cyclized the resulting phosphonium salt 14 in situ by
adding base. This gave rise to their cyclopentenone (13). We
were not sure whether our envisaged sequence chloroolefin (10)
→ bromoketone (9) → phosphonium salt (8) → [ylide (in situ) →]
cyclopentenone (symbolized by 7) would proceed as smoothly,
too. This is because in each step our reaction center would be
more hindered than theirs – by an extra methyl group.[11]
O
O
O
O
O
O
H
O
H
Cl
O
O
O
H
Br
O
Cl
Br
H
O
(p-bromobenzoyl)-syn-19c
(= syn-20)
(p-bromobenzoyl)-anti-19e
(= anti-21)
Finishing the retrosynthetic analysis of Scheme 1, we realized
that the chloroolefin 10 represents a protected β-hydroxyketone.
It therefore looked accessible by an aldol addition.[12] The re-
quired aldehyde was (Z)-2-chlorobut-2-enal (12), a known com-
pound.[13] Pertinent enolates should stem from dioxolane-4-es-
ters of generic structure 11.[14] For the present study we used
those as racemic mixtures. Aldol additions of related dioxolane-
4-esters to achiral electrophiles revealed a chance of steering
their simple diastereoselectivity by moving from lithium eno-
late[15] to Mukaiyama-type aldol additions.[16] However, only very
few of the reported reactions exhibited high simple and induced
diastereoselectivities[17] simultaneously.[16]
Figure 2. Crystal structures of esters syn-20[21] and anti-21[24] derived from the
aldols syn-19c (yield: 72%) and anti-19e (yield: 78%), respectively, as configu-
rational proofs. Ellipsoids shown at 50% probability; black: C, red: O, gray: H
(omitted for clarity unless bound to a stereocenter), green: Cl, purple: Br.
3
This article is protected by copyright. All rights reserved.

10.1002/chem.201701185
Chemistry - A European Journal
COMMUNICATION
e overwhelmingly 5,4trans,4,1’syn-configured.[ 19 - 22 ] Yields were
lower starting from the benzyl (→ 19a: 29%; b: 46%) than from
the methyl esters (→ 19c-e: 81-90%).[23] The syn-preference
complies with precedence.[15] For obtaining the diastereomeric
aldols anti-19 we converted the methyl esters 17c-e into the silyl
ketene acetals 18c-e. BF3·OEt2 let them undergo Mukaiyama al-
dol additions to (Z)-2-chlorobut-2-enal (12[13]). In line with earlier
observations[16] this occurred 5,4trans- and 4,1’anti-selectively[19,20]
and provided the aldols 19c-e diastereopure.[24] All assignments
were based on X-ray analyses (Figure 2) and NMR analogy.[21,24]
We then silylated the OH group of the cyclohexylidene-protected
aldol syn-19e and the cyclopentylidene-protected aldol anti-19d
(Scheme 3). Next we elaborated the CO2Me groups of the re-
sulting ethers syn-22 and anti-24 into phenyl ketone. Surprising-
ly, the Weinreb amide derived from syn-22 did not react with
PhLi at −78°C or with PhMgBr at +60°C.[25] In contrast, PhLi re-
acted with the esters syn-22 and anti-24 already at −78°C. This
gave the respective phenyl ketones syn-23 and anti-25 in 86%
and 90% yield, respectively, over the 2 steps.[26 ] The corre-
sponding tertiaryl alcohols were not observed.[27]
Scheme 3. Preparation of silyl-protected phenyl ketones. a) TBSOTf, 2,6-lutid-
ine, CH2Cl2, 0°C → RT, 18 h; 95%. b) PhLi, THF, −78°C, 1 h; 91%. c) Same
as (a); 97%. d) Same as (b); 93%.
NaOBr, H2O,
AcOH, acetone,
0 °C, 90 min
Ph
Ph
Ph
Ph
O
O
O
O
PBu3, toluene, 2 h
O
O
O
O
O
O
O
O
88%
TBSO
TBSO
TBSO
TBSO
Br
Cl
syn-23
O
O
O
Ph
O
28
syn-26 (d.r. 90:10)
29
BrPBu3
O
citric acid- or
DHQ-promoted
dihydroxylation
O
workup
no conversion
TBSO
PBu3
X
NEt3, toluene, 110 °C, 16 h
O
27
Ph
Ph
O
O
O
O
Ph
O
O
O
O
O
TBSO
TBSO
TBSO
OH
OMs
O
O
30
31
cis-32
Scheme 4. Attempted conversions of the phenyl ketone syn-23 into the cyclopentenenone cis-32 via the bromoketone syn-26 or the corresponding mesylate 31.
The chloroolefin moiety of compound syn-23 was hydrolyzed ox-
idatively with NaOBr in a mixture of aqueous acetone and AcOH
(Scheme 4). The bromoketone syn-26 was isolated as a 90:10
mixture of diastereomers (88% yield). It was treated with PBu3 at
room temp. for 2 h; we then added NEt3 and heated at 110°C,
establishing the conditions of Stoltz’ conversion of the bromoke-
tone 15 into the cyclopentenone 13[10b] (Scheme 1). Unfortunate-
ly, the greater steric hindrance of bromoketone syn-26 let PBu3
attack at Br (→ phosphonium enolate 28) rather than at C. This
changed the chemoselectivity just like when moving from an Ar-
busov to a Perkow reaction. Protic workup gave the debromoke-
tone 29 (68%). We hoped to circumvent this obstacle with the
(mesyloxy)ketone 31 as a non-reducible alkylant. However, its
putative precursor, the alcohol 30, proved inaccessible by ex-
posing the chloroolefin syn-23 to several dihydroxylation cock-
tails.[28]
Ph
O
O
Ph
O
Ph
O
OH
a)
b)
O
O
O
O
94%
47%
TBSO
TBSO
TBSO
Br
O
O
syn-26 (d.r. 90:10)
cis-33
cis-32
c) 99%
Scheme 5 shows that we got around that hurdle by engaging the
bromoketone syn-26 in an SmI2-mediated cyclization.[29] It repre-
sents a Reformatsky reaction or the aldol addition of an Sm(III)
enolate (if so, it is a far better nucleophile than the phosphonium
enolate 28; cf. Scheme 4).[30] The hydroxycyclopentanone cis-33
resulted diastereopure[31] in 94% yield. It was dehydrated with
the Burgess reagent[32] giving the cyclopentenone cis-32 in 47%
yield (after separation from 37% recovered starting material).
Reaching the 1,2-diol cis-7, our first 2nd generation kodaistatin
model, required 3 more steps: acetal cleavage to the diol cis-34
(99%); Swern oxidation to the (tert-hydroxy)ketone cis-35 (82%);
desilylation (92%).
O
HO
HO
Ph
O
Ph
Ph
O
e)
92%
d)
82%
HO
HO
HO
TBSO
TBSO
cis-34
O
O
cis-7
cis-35
Scheme 5. Finishing our 2nd generation kodaistatin model cis-7. a) SmI2, THF,
−78°C → RT (within 2 h); 94%. b) Burgess reagent, toluene, 100°C, 12 h; 47%
(74% accounting for 37% recovered cis-33). c) CF3CO2H, H2O, CH2Cl2, RT, 1
h; 99%. d) (COCl)2, DMSO, CH2Cl2, −78°C → −45°C (within 1 h); NEt3, −45°C
→ RT (within 2 h); 82%. e) HCl, MeOH, RT, 18 h; 92%.
4
This article is protected by copyright. All rights reserved.

10.1002/chem.201701185
Chemistry - A European Journal
COMMUNICATION
Ph
Ph
O
O
C
O
a)
b)
C
O
O
O
O
HO
HO
HO
HO
HO
92%
C
TBSO
TBSO
Br
HO
H
Cl
anti-25
O
H
H
O
O
trans-40
39
cis-40
anti-36 (d.r. 95:5)
Figure 3. The uniquely substituted cyclopentenone core 39 targeted in this
study.
O
Ph
O
O
O
Ph
Ph
OH
OH
c)
The 3D structure of compound trans-7 was ascertained by X-ray
crystallography.[35] This and the similarity of the 13C-NMR shifts
of this compound and kodaistatin A (Table 1) support the cor-
rectness of the connectivity assigned[5] to the latter. This is reas-
suring considering the uniqueness of the dihydroxycyclo-
pentenone motif 39 (Figure 3). In isolated rings it is non-existent
except in compounds 1-4, cis-6, cis- and trans-7); fused to an
aliphatic ring it occurs only in compounds cis- and trans-40.[36]
Table 1 juxtaposes 13C NMR shifts of the acetylcyclopentenone
motifs of our 1st generation model cis-6,[6] 2nd generation models
cis- and trans-7, and kodaistatin A (1). Cis-6 having shown dis-
tinctive deviations,[6] it was no surprise retrieving them in com-
pound cis-7, no matter the different side-chains. Still, this re-
newed negative evidence (“kodaistatin A should be configured
oppositely than cis-7“) fades besides the positive one: The
13C NMR shift devations of our model trans-7 vs. the natural
product are as small as ≤2 ppm. This allows to deduce that ko-
daistatin A is configured the same as trans-7. The common
origin of the kodaistatins A-D suggests they share this trans-
diol moiety.
In conclusion, we synthesized a pair of cis,trans-isomeric mod-
els of the kodaistatin cyclopentenone core, beginning with dia-
stereodivergent and highly stereocontrolled aldol additions. It is
noteworthy that the inducing stereocenter was in a motif, which,
overall, served as a latent acetyl group. Hence, this stereocenter
was given up in the end. Our routes comprised 11 steps (→ cis-
7, 28% total yield) and 12 steps (→ trans-7, 11% total yield), re-
spectively. Stoltz´ chloroolefin → acylphosphorane → enone
strategy was inapplicable to the spiroannulated precursors of our
hindered targets. Accordingly, we developed a high-yielding by-
pass through a brominating hydrolysis,[9] an Sm(II)-induced cy-
clization, and a dehydration. It should provide other densely
functionalized cyclopentenones,[37] too. 13C NMR comparisons of
cis- and trans-7 with kodaistatin A (1) revealed that the latter is a
trans-1,2-diol.
+
O
O
O
88%
TBSO
TBSO
TBSO
O
O
49%
trans-37a
27%
trans-37b
trans-38
d)
incomplete conversion into trans-38;
trans-37b and trans-38 inseparable
e) - g) 74%
(3 steps)
O
Ph
HO
HO
H
HO
O
O
HO
O
trans-7
Scheme 6. Completion of the synthesis of our second kodaistatin model trans-
7 [crystal structure (ref.[35]) with ellipsoids at 50% probability; black: C, red: O,
gray: H (omitted for clarity unless bound to a stereocenter)]. a) Ca(OBr)2 in
H2O, AcOH, CH3CN/CH2Cl2, 0°C, 6 h; 92%. b) SmI2, THF, −78°C → −10°C
(within 4 h); 49% trans-37a + 27% trans-37b (separable). c) Burgess reagent,
toluene, 100°C, 17 h; 88% (97% accounting for 6% recovered trans-37a). d)
Burgess reagent, toluene, 100°C, 18 h; 29% (inseparable from 58% un-
changed trans-37b). e) CF3CO2H, H2O, CH2Cl2, 0°C, 2 h; 89%. f) (COCl)2,
DMSO, CH2Cl2, −78°C → −60°C (in 1 h); NEt3, −78°C → RT (in 3 h); 95%. g)
HCl, MeOH, RT, 3.5 d; 88%.
Our second 2nd generation kodaistatin model, the 1,2-diol trans-
7, was synthesized from the aldol anti-25 (Scheme 6) essentially
like we had gained its isomer cis-7 from the aldol syn-23
(Scheme 4-5). One difference was that the brominating hydrol-
ysis[9] of the chloroolefin anti-25 required using Ca(OBr)2, which
was more reactive than NaOBr. The bromoketone anti-36 re-
sulted as a 95:5 mixture of diastereomers (92%). Its SmI2-in-
duced cyclization[29] was non-diastereoselective. The hydroxycy-
clopentanones trans-37a[33] and trans-37b[34] resulted in 49%
and 27% yield, respectively, after separation by column chroma-
tography. Treatment with Burgess’ reagent[32] provided the cy-
clopentenone trans-38. It formed in 88% yield from trans-37a but
was inseparably contaminated with unchanged substrate start-
ing from trans-37b. Acetal cleavage, Swern oxidation, and desi-
lylation completed the synthesis of trans-7.
Acknowledgements
We thank Dr. Manfred Keller (Institut für Organische Chemie)
and Dr. Daniel Kratzert (Institut für Anorganische und Analyti-
sche Chemie) for the X-ray structure analyses and Sabine Bog-
nar and Stephanie Keller for experimental contributions.
Keywords: aldol reaction • α-bromoketones • cyclopentenone •
natural products • samarium enolate • structure elucidation
5
This article is protected by copyright. All rights reserved.

10.1002/chem.201701185
Chemistry - A European Journal
COMMUNICATION
Table 1. 13C NMR chemical shift comparisons (DMSO-d6 solutions): kodaistatin A (1; 151 MHz),[5] model compounds cis-6 (125.6 MHz; ref.[6]), cis-7 (100.6 MHz;
this study), and trans-7 (100.6 MHz, this study). Structure-revealing shift differences ∆δ (≡ δin model − δin 1) on grey background.
2' Ph
2' Ph
1'
2' Ph
1'
2'
Ar
2
O
O
O
O
1'
1'
3
3
HO
3
HO
3
HO
HO
4
4
4
4
2
2
1
2
1
5
5
5
5
1
1
O
O
HO
HO
HO
HO
O
O
O
O
(2nd generation models)
kodaistatin A (1[5]
)
1st generation model (cis-6[6]
)
trans-7
cis-7
C
1
δ/ppm
δ/ppm
∆δ/ppm
δ/ppm
∆δ/ppm
δ/ppm
∆δ/ppm
200.0
137.2
161.6
89.7
204.4
135.3
165.3
85.5
4.4
−1.9
3.7
205.0
136.4
162.7
85.1
5.0
−0.8
1.1
202.0
137.0
160.5
88.6
2.0
−0.2
−1.1
−1.1
−1.6
0.1
2
3
4
−4.2
−9.0
3.2
−4.6
−9.0
3.6
5
84.5
75.5
75.5
82.9
1’
2’
207.7
27.7
210.9
27.3
211.3
27.1
207.8
26.9
−0.4
−0. 6
−0.8
J. Yamaguchi, K. Sato, S. Yamaguchi, T. Mukaiyama, K. Sakai, Y. Asami, H.
Kakeya, H. Osada, J. Am. Chem. Soc. 2002, 124, 12078.
[17] For an authoritative definition of these terms see Section 1.2.2.3 (pp. 53-55) of
G. Helmchen in Houben-Weyl Methods of Organic Chemistry, Vol. E 21a
(Eds.: G. Helmchen, R. W. Hoffmann, J. Mulzer, E. Schaumann), Thieme,
Stuttgart 1995, pp. 1-74.
[1]
[2]
IDF Diabetes Atlas - 7th edition, http://www.diabetesatlas.org/, July 14, 2016.
L. K. Charkoudian, B. P. Farrell, C. Khosla, Med. Chem. Commun. 2012, 3,
926.
a) A. Consoli, N. Nurjhan, F. Capani, J. Gerich, Diabetes 1989, 38, 550; b) C.
Meyer, M. Stumvoll, V. Nadkarni, J. Dostou, A. Mitrakou, J. Gerich, J. Clin.
Invest. 1998, 102, 619.
N. V. S. Ramakrishna et al. (Hoechst Marion Roussel Deutschland GmbH),
WO1998047888A1, 1998.
L. Vértesy, H.-J. Burger, J. Kenja, M. Knauf, H. Kogler, E. R. Paulus, N. V. S.
Ramakrishna, K. H. S. Swamy, E. K. S. Vijaykumar, P. Hammann, J. Antibiot.
2000, 53, 677.
T. Wüster, N. Kaczybura, R. Brückner, M. Keller, Tetrahedron 2013, 69, 7785.
Wedged bonds identify enantiomerically pure, straight bonds racemic com-
pounds.
a) B. M. Trost, D. P. Curran, J. Am. Chem. Soc. 1980, 102, 5699; b) C. Kuhn,
J.-C. Florent, Tetrahedron Letters 1998, 39, 4247; c) T. Kanger, K. Raudla, R.
Aav, A.-M. Müürisepp, T. Pehk, M. Lopp, Synthesis 2005, 3147; e) T. Werner,
M. Hoffmann, S. Deshmukh, Eur. J. Org. Chem. 2014, 6630.
The term “brominating hydrolysis“ describes the overall change of substruc-
ture C(−Cl)C=C into substructure C(=O)C−C−Br. Formally (not mechanisti-
[3]
[18] Starting from crotonic acid or its methyl ester, we prepared the esters 17a,b
and 17c-e, respectively, as detailed in the Supporting Information. Ester 17b is
new while 17a,c-e were known: a: C. G. Boojamra, et al. (Gilead Sciences
Inc.), US20110288053A1, 2011; c: ref.[14]; d: W. R. Roush, R. A. Hartz, D. J.
Gustin, J. Am. Chem. Soc. 1999, 121, 1990; e: J. H. Dodd, R. S. Garigipati, S.
M. Weinreb, J. Org. Chem. 1982, 47, 4045.
[19] The stereodescriptor 5,4trans refers to the orientation of the C5−Me vs. C4−C1’
bond in the dioxolane ring of aldols 19a-e {as imposed by induced diastereo-
control (cf. ref. [17])}. The stereodescriptor 4,1’syn – and 4,1’anti likewise – de-
scribes the orientation of the C4−O vs. the C1’−O bond {as imposed by simple
diastereocontrol (cf. ref. [17])} if drawn as in Scheme 2 (cf. ref. [20]).
[20] An aldol addition subject to simple diastereocontrol creates two vicinal ste-
reocenters. Their relative configurations were distinguished first as erythro and
threo [a) H. E. Zimmerman, M. D. Traxler, J. Am. Chem. Soc. 1957, 79, 1920;
b) W. A. Kleschick, C. T. Buse, C. H. Heathcock, , J. Am. Chem. Soc. 1977, 99,
247; c) D. A. Evans, J. Bartroli, T. L. Shih, J. Am. Chem. Soc. 1981, 103,
2127] and later as syn and anti [d) footnote 7 in S. Masamune, S. A. Ali, D. L.
Snitman, D. S. Garvey, Angew. Chem., Int. Ed. Engl. 1980, 19, 557; Angew.
Chem. 1980, 92, 573]. Our aldols 19 evade either nomenclature since a quater-
nary carbon separates the C−O from the C=O bond.
[4]
[5]
[6]
[7]
[8]
[9]
cally!), this transformation results from
a hydrolysis C(−Cl)C=C →
C(−OH)C=C, a tautomerization C(−OH)C=C → C(=O)C−C−H, and an α-
bromination C(=O)C−C−H → C(=O)C−C−Br. Hence the designation.
[10] Selected precedents: a) C.-N. Hsiao, M. R. Leanna, L. Bhagavatula, E. de Lara,
T. M. Zydowsky, B. W. Horrom, H. E. Morton, Synthetic Comm. 1990, 20,
3507: NBS, H2O; b) R. A. Craig, J. L. Roizen, R. C. Smith, A. C. Jones, B. M.
Stoltz, Org. Lett. 2012, 14, 5716: NaOBr, aq. AcOH; c) V. Pace, L. Castoldi,
M. J. Hernáiz, A. R. Alcántara, W. Holzer, Tetrahedron Lett. 2013, 54, 4369:
Ca(OBr)2, aq. AcOH.
[21] Plausibly, the aldols syn-19a,b and d,e are configured like the aldol syn-19c,
whose 5,4trans,4,1’syn-configuration was proved by an X-ray structural analysis
of its bromobenzoate syn-20 (Figure 2). The respective data are contained in
CCDC 1508542 and available free of charge from the Cambridge Crystallo-
graphic Data Centre via the link www.ccdc.cam.ac.uk/data_request/cif.
[22] The steric course of the underlying aldol addition is rationalized in the Sup-
porting Information.
[11] As would turn out, our plan from Scheme 1 gave the α-bromoketone 9 alright
but did not work out beyond (Scheme 4)
[12] Selected reviews: a) R. Mahrwald (Ed.), Modern Methods in Stereoselective
Aldol Reactions, Wiley-VCH, Weinheim, New York, 2013; b) G. L. Beutner, S.
E. Denmark, Angew. Chem. Int. Ed. 2013, 52, 9086; Angew. Chem. 2013,
9256; c) S. B. J. Kan, K. K. H. Ng, I. Paterson, Angew. Chem. Int. Ed. 2013, 52,
9097; Angew. Chem. 2013, 9267; d) J.-i. Matsuo, M. Murakami, Angew. Chem.
Int. Ed. 2013, 52, 9109; Angew. Chem. 2013, 125, 9280.
[13] Prepared as described by L. Skattebøl, J. Org. Chem. 1966, 31, 1554.
[14] Exemplary preparation: Y. Yang, X. Fu, J. Chen, H. Zhai, Angew. Chem. Int.
Ed. 2012, 51, 9825; Angew. Chem. 2012, 124, 9963.
[15] Previous additions of Li enolates of type-11 esters to achiral electrophiles: a) R.
Naef, D. Seebach, Angew. Chem. Int. Ed. Eng. 1981, 20, 1030; Angew. Chem.
1981, 93, 1113; b) W. Ladner, Chem. Ber. 1983, 116, 3413; c) M. Pohmakotr,
T. Junpirom, S. Popuang, P. Tuchinda, V. Reutrakul, Tetrahedron Lett. 2002,
43, 7385; c) M. S. M. Timmer, B. L. Stocker, P. H. Seeberger, J. Org. Chem.
2006, 71, 8294; d) E. Bette, A. Otto, T. Dräger, K. Merzweiler, N. Arnold, L.
Wessjohann, B. Westermann, Eur. J. Org. Chem. 2015, 2357; e) K. C.
Nicolaou, Q. Cai, H. Sun, B. Qin, S. Zhu, J. Am. Chem. Soc. 2016, 138, 3118.
[16] Previous additions of silyl ketene acetals of type-11 esters to achiral electro-
philes: a) D. A. Evans, W. B. Trotter, J. C. Barrow, Tetrahedron 1997, 53,
8779; b) V. K. Aggarwal, S. J. Masters, H. Adams, S. E. Spey, G. R. Brown, A.
J. Foubister, J. Chem. Soc., Perkin Trans. 1 1999, 155; c) Y. Hayashi, M. Shoji,
[23] Deprotonating the benzyl ester 17b as before (LDA, THF, −78°C) but quench-
ing it with aq. NH4Cl 45 min later caused mainly decomposition.
[24] The aldol anti-19c should be identically configured as aldols anti-19d and e.
The 5,4trans,4,1’anti-configuration of aldol anti-19d emerged from an X-ray
structural analysis of the final product of the follow-up sequence anti-19d →
anti-24 → anti-25 → anti-36 → trans-37a → trans-38 → trans-7. The
5,4trans,4,1’anti-configuration of aldol anti-19e was proved by X-raying its
bromobenzoate anti-21 (Figure 2). The crystallographic data are contained in
CCDC 1508543 and available free of charge from the Cambridge Crystallo-
graphic Data Centre via the link www.ccdc.cam.ac.uk/data_request/cif.
[25] Weinreb amide 41, too poor an acylating agent for reacting with phenyl organ-
ometallics:
PhMgBr, THF,
0°C - 60°C
OMe
HNMe(OMe)•HCl,
BuLi, THF, RT;
O
Ph
O
O
O
CO2Me
N
O
O
O
O
Me
then syn-22, THF,
−78 °C RT (in 3 h)
99%
or
TBSO
TBSO
TBSO
PhLi, THF,
−78 °C
Cl
Cl
Cl
syn-22
41
syn-23
[26] Processing the CO2Me-containing aldol adducts anti-19c,e of Scheme 2 analo-
gously was less efficient – either until reaching the respective phenyl ketones
or later on.
[27] This selectivity may have either of two reasons: (1) The ketones syn-23 and
anti-25 are sterically too hindered for reacting with PhLi under the reaction
6
This article is protected by copyright. All rights reserved.

10.1002/chem.201701185
Chemistry - A European Journal
COMMUNICATION
conditions. (2) The tetrahedral intermediate preceding the respective ketone is
stabilized through chelation by the dioxolane ring until all PhLi is consumed.
If the ketone forms after this point in time, it cannot encounter PhLi any more.
[28] In spite of that, a few chloroolefin → hydroxyketone conversions are known,
e. g.: a) T. Satoh, K. Onda, K. Yamakawa, J. Org. Chem. 1991, 56, 4129; b) B.
M. Trost, A. B. Pinkerton, J. Am. Chem. Soc. 2002, 124, 7376.
[29] Prior to our work SmI2-induced inter- and intramolecular aldol additions of α-
haloketones have been used by a) Z. Yang, D. Shannon, V.-L. Truong, P.
Deslongchamps, Org. Lett. 2002, 4, 4693; b) D. Chapdelaine, J. Belzile, P.
Deslongchamps, J. Org. Chem. 2002, 67, 5669; c) B. A. Sparling, R. M.
Moslin, T. F. Jamison, Org. Lett. 2008, 10, 1291; d) E. J. Horn, J. S. Silverston,
C. D. Vanderwal, J. Org. Chem. 2016, 81, 1819.
[30] A referee wondered whether a radical-mediated cyclization could represent yet
another explanation. We do not think so because the bromoketone syn-26 is
reduced instantaneously when added, as a solution in THF, to 2.0 equiv. of
SmI2 at −78°C. As soon as the last drop was added, the reaction mixture turned
from dark blue [Sm(II)] to yellow [Sm(III)]. Aqueous quenching of an aliquot
of the reaction mixture a good hour later followed by TLC analysis indicated
that the bromoketone had disappeared completely and the debrominated ketone
29 – plausibly formed from the Sm(III) enolate – and the aldol adduct cis-33
were present instead. Thus (at least some) C−Br bond breakage occurs well
ahead of C−C bond formation.
[31] The cis-orientation of the 3-Me, 4-OH, and 1-OTBS bond in compound cis-33
follows from pertinent crosspeaks in the NOESY spectrum (CDCl3):
[32] a) Original report: G. M. Atkins, E. M. Burgess, J. Am. Chem. Soc. 1968, 90,
4744; summaries of applications: b) S. Santra, Synlett 2009, 328, c) M. M.
Heravi, T. Ahmadi, A. Fazeli, N. M. Kalkhorani, Curr. Org. Synth. 2015, 12,
328.
[33] The cis-orientation of the 3-Me, 4-OH, and 1-H bond in compound trans-37a
follows from pertinent crosspeaks in the NOESY spectrum (CDCl3):
[34] The cis-orientation of the 3-Me, 4-OH, and 1-OTBS bond in compound trans-
37b follows from pertinent crosspeaks in the NOESY spectrum (CDCl3):
[35] The crystallographic data of compound trans-7 are contained in CCDC
1508544 and available free of charge from the Cambridge Crystallographic
Data Centre via the link www.ccdc.cam.ac.uk/data_request/cif.
[36] G. B. Mitre, N. Kamiya, A. Bardón, Y. Asakawa, J. Nat. Prod. 2004, 67, 31-
36).
[37] Reviews: a) S. P. Roche, D. J. Aitken, Eur. J. Org. Chem. 2010, 5339; b) D. J.
Aitken, H. Eijsberg, A. Frongi, J. Ollivier, P. P. Piras, Synthesis 2014, 46, 1; il-
lustrative examples: c) ref.[15d]; d) Z. Zhou, D. D. Dixon, A. Jolit, M. A. Tius,
Chem. Eur. J. 2016, 22, 15929.
7
This article is protected by copyright. All rights reserved.