Total synthesis of nominal gobienine A

Article abstract of DOI:10.1002/chem.201300827

The lichen-derived glycoconjugate gobienine A is structurally more complex than most glycolipids isolated from higher plants by virtue of the all-cis substituted γ-lactone substructure embedded into its macrocyclic frame. A concise entry into this very epimerization-prone and hence challenging structural motif is presented, which relies on an enantioselective cyanohydrin formation, an intramolecular Blaise reaction, a palladium-catalyzed alkoxycarbonylation, and a diastereoselective hydrogenation of the tetrasubstituted alkene in the resulting butenolide. This strategy, in combination with ring-closing olefin metathesis for the formation of the macrocyclic perimeter, allowed the proposed structure of gobienine A (1) to be formed in high overall yield. The recorded spectral data show that the structure originally attributed to gobienine A is incorrect and that it is not the epimerization-prone ester site on the butanolide ring that is the locus of misassignment; rather, the discrepancy must be more profound. Copyright

Full text of DOI:10.1002/chem.201300827

FULL PAPER
DOI: 10.1002/chem.201300827
Total Synthesis of Nominal Gobienine A
Azusa Kondoh, Alexander Arlt, Barbara Gabor, and Alois Fꢀrstner*[a]
Abstract: The lichen-derived glycocon-
Blaise reaction, a palladium-catalyzed
alkoxycarbonylation, and a diastereose-
lective hydrogenation of the tetrasub-
stituted alkene in the resulting buteno-
lide. This strategy, in combination with
ring-closing olefin metathesis for the
formation of the macrocyclic perime-
jugate gobienine A is structurally more
complex than most glycolipids isolated
from higher plants by virtue of the all-
cis substituted g-lactone substructure
embedded into its macrocyclic frame.
A concise entry into this very epimeri-
zation-prone and hence challenging
structural motif is presented, which
relies on an enantioselective cyanohy-
drin formation, an intramolecular
ter, allowed the proposed structure of
gobienine A (1) to be formed in high
overall yield. The recorded spectral
data show that the structure originally
attributed to gobienine A is incorrect
and that it is not the epimerization-
prone ester site on the butanolide ring
that is the locus of misassignment;
rather, the discrepancy must be more
profound.
Keywords: butanolides · carbonyl-
AHCTUNGTRENNUNG
Introduction
Lichens are symbiotic alliances between fungal and algal (or
cyanobacterial) partners, which are so closely interwoven
that they appear as a single organism.[1] The success of such
partnerships is evident from the fact that lichens are able to
subsist even under very hostile conditions, including expo-
sure to extreme temperatures, prolonged darkness, UV irra-
diation, drought and/or saline substrates. In contrast to this
unusual robustness, they are very sensitive to air pollution
and therefore serve as bio-indicators. Their sometimes color-
ful appearance is a visible sign of a potentially interesting
secondary metabolism, which is the likely reason why differ-
ent lichens have served traditional medicine since ancient
times.[1,2]
The most common class of lichen-derived natural prod-
ucts are the so-called paraconic acids (4-carboxy-g-butyro-
lactones), which exist in nature in all possible stereoisomeric
formats as well as in partly unsaturated forms.[3] A particu-
larly complex incarnation is found in the gobienines (1–3),
which were isolated in very small amounts from samples of
the lichen Acarospora gobiensis collected in the Central
Asian Tian-Shan mountains (<0.01% of the dry mass).[4,5]
The structures of these highly unusual glycoconjugates were
elucidated by NMR spectroscopy in combination with chem-
ical and enzymatic degradation (Figure 1).
Figure 1. Proposed structures of the gobienines and the lipidic aglycone
constituent of putative gobienine A.
It is of particular note that 1–3 supposedly embody hy-
droxylated paraconic acids in their composite macrocyclic
frames, which differ from each other not only in the relative
stereochemistry of the butanolide ring but even in the abso-
lute configuration of the remote hydroxylated site at C18,
which is supposedly (R)-configured in gobienine A (1) but
(S)-configured in its relatives 2 and 3.[4–6]
[a] Dr. A. Kondoh, Dipl.-Chem. A. Arlt, B. Gabor, Prof. A. Fꢀrstner
Max-Planck-Institut fꢀr Kohlenforschung
45470 Mꢀlheim/Ruhr (Germany)
Fax : (+49)208306-2994
No biological data were reported for 1–3 by the isolation
team, but the diverse spectrum of biological and pharmaco-
logical activities of paraconic acids on the one hand[2,3,7] and
E-mail: fuerstner@kofo.mpg.de
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201300827.
Chem. Eur. J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

of macrocyclic glycolipids de-
rived from higher plants on the
other hand[8–10] encouraged us
to pursue the total synthesis of
these challenging targets. We
focused our efforts on gobie-
nine A (1) as the arguably most
demanding of the three mem-
bers of this family. The all-cis
configured trisubstituted bu
lide ring contained in its li
ACHTUNGTRENNUNG
ic aglycone 4[11] is very epimeri-
ACHTUNGTRENNUNG
A
N
ACHTUNGTRENNUNG
zation-prone and therefore con-
stitutes a serious challenge.[12]
This notion is reflected in the
conspicuous lack of selective
syntheses of paraconic acids
with this obviously unfavorable
substitution pattern, whereas
many elegant approaches to the
thermodynamically more stable
isomers have been developed
over the years.[3,7,13]
Outlined below is a stereoselective and largely catalysis-
based entry into all-cis configured 4-carboxy-g-butyrolac-
tones, which laid the ground for a concise approach to what
is nominal gobienine A (1). Unfortunately, however, the
spectroscopic data of synthetic 1 exclude that the originally
assigned structure is correct; the more stable isomer 3-epi-1,
which has also been prepared during this campaign, does
not represent the natural product either.[14]
Scheme 1. Diastereoselective synthesis of an all-cis configured paraconic acid derivative in the racemic series:
a) i) TMSCN, Yb(OTf)3·nH2O (5 mol%), CH2Cl2; ii) aq. HCl, MeCN, 52%; b) 2-bromopropionyl bromide,
Et3N, CH2Cl2, 96%; c) i) Zn, MeSO3H (1 mol%), THF, reflux; ii) aq. HCl, 908C; d) Tf2O, Et3N, CH2Cl2, 08C,
69% (over both steps); e) see Table 1; f) H2 (15 atm) Rh/Al2O3 cat., EtOAc, 87% (11), 9% (12); NOE = nu-
clear Overhauser effect; Tf = trifluoromethanesulfonyl, TMS = trimethylsilyl.
AHCTUNGTRENNUNG
form,[21] we believe that the intramolecular Blaise chemistry
constitutes a valuable gateway to tetronic acids and deriva-
tives thereof.
A subsequent palladium-catalyzed methoxycarbonylation
of triflate 9 was expected to afford the corresponding
methyl ester 10. Applications of this chemistry to electron-
deficient alkenyl triflates, however, are surprisingly rare.[22]
Indeed, reaction of 9 under standard conditions (1 atm of
CO, MeCN/MeOH) using either [PdACHTNUGTRENNUG(PPh3)4] or [PdACHTUNGTRENNUNG(OAc)2]/
PPh3 in the presence of an amine base furnished only disap-
pointingly low yields of methyl ester 10 (Table 1). A short
screening, however, showed that the use of chelating bis-
phosphines with large bite angles leads to better outcomes,
Results and Discussion
The all-cis substituted paraconic acid head group: Prior to
the actual synthesis of the target natural product, the con-
ceived approach towards all-cis configured 4-carboxy-g-bu-
tyrolactones was tested by the model study shown in
Scheme 1. To this end, the cyanohydrin 6, derived from oc-
tanal on treatment with TMSCN and ytterbium triflate,[15]
was esterified with 2-bromopropionyl bromide under stand-
ard conditions. Reaction of compound 7 with activated zinc
dust in refluxing THF generated the corresponding Refor-
matsky reagent,[16] which cyclized smoothly onto the nitrile
to give the tetronic acid 8 upon acidic work up; to facilitate
product isolation, the crude material was transformed into
triflate 9, which was obtained in 69% yield over both steps.
Even though unoptimized, the outcome of this intra-molecu-
lar Blaise reaction[17,18] compares favorably to related chem-
istry previously reported by other groups: specifically, inter-
molecular condensations of zinc ester enolates with cyano-
hydrins are generally less effective;[19] likewise, the few re-
ported examples of intramolecular reactions of bromoalka-
noates with tethered esters (rather than nitriles) mediated
by magnesium (rather than zinc) are much lower yielding.[20]
Since cyanohydrins are readily available in optically pure
with DPEphos being optimal. Replacement of Pd
ACHTUNGRTEN(NUNG OAc)2 by
Pd(OCOCF3)2 further improved the results, allowing prod-
AHCTUNGTRENNUNG
uct 10 to be isolated in well reproducible 89% yield.
Not unexpectedly, the projected diastereoselective hydro-
genation of the tetrasubstituted and, at the same time, very
electron-deficient double bond in 10 required optimization.
Whereas an assortment of homogeneous and heterogeneous
catalysts essentially failed to afford the desired product,[23]
Table 1. Methoxycarbonylation of alkenyl triflate 9.[a]
Entry
[Pd] (mol%)
Ligand (mol%)
T [8C]
Yield [%]
1
2
3
4
5
6
7
Pd
Pd
N
–
80
80
80
80
RT
RT
RT
16
25
0
27
53
61
89
Ph3P (35)
dppe (15)
dppb (15)
Xantphos (10)
DPEphos (10)
DPEphos (10)
Pd
Pd
[a] All reactions were performed under CO (1 atm) in MeCN/MeOH 5:1
in the presence of iPrNEt2 (2–2.5 equiv); dppb bis(diphenylphos-
phino)butane; dppe = bis(diphenylphosphino)ethane.
=
A
ACHTUNGTRENNUNG
&
2
&
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Total Synthesis of Nominal Gobienine A
FULL PAPER
Rh/Al2O3 effected the reduc-
tion quite cleanly.[24] When the
reaction was performed with an
initial hydrogen pressure of
ꢀ15 bar in EtOAc as the sol-
vent at ambient temperature,
product 11 was isolated in re-
spectable 87% yield; small
amounts (ꢁ9%) of isomer 12
were also formed but could be
separated by flash chromatogra-
phy. The recorded NOE data
are in line with an all-cis orien-
tation of the substituents in 11,
thus confirming that the two
hydrogen atoms were delivered
trans to the heptyl side-chain
present in the substrate. Fur-
thermore, the characteristic
coupling constants of 11 and 12
concur with those of paraconic
acid derivatives of known con-
Scheme 2. a) NaH, THF/DMF, 08C, then BnBr, RT, 49%; b) PCC, silica, CH2Cl2, 75%; c) TMSCN, TiACHTUNGTRENNUNG(OiPr)4
figuration (see the Supporting
(10 mol%), 24 (10.5 mol%), THF, ꢂ408C, then aq. HCl, 88% (89% ee); d) 2-bromopropionyl bromide, Et3N,
Information).[7,13,25]
CH2Cl2, 97%; e) Zn, CH3SO3H (1 mol%), THF, reflux, then aq. HCl, 958C; f) Tf2O, Et3N, CH2Cl2, 08C, 77%
(over both steps); g) Pd
TMSCH2CH2OH (5 equiv), 82%; h) H2 (1 atm), Pd/C, EtOH, 79%; i) H2 (11 atm), Rh/Al2O3 cat., EtOAc,
81% (20) and 13% (21); j) PhI(OAc)2, TEMPO (10 mol%), CH2Cl2, 84%; k) CH2I2, Cp2ZrCl2, Zn, THF,
71%; l) TBAF, CH2Cl2, 08C, 92%; Bn = benzyl; Cp = cyclopentadienyl; PCC = pyridinium chlorochromate;
TBAF = tetra-n-butylammonium fluoride; TEMPO = 2,2,6,6-tetramethyl-1-piperinyloxy radical.
ACHTUNGTRENN(NUG OCOCF3)2 (10 mol%), DPEphos (10 mol%), iPrNEt2, CO (1 atm), MeCN,
Nominal gobienine A and 3-
epi-gobienine A: The favorable
results of the model study en-
couraged us to translate this ap-
proach into a total synthesis of
gobienine A (1). Our previous
AHCTUNGTRENNUNG
experiences with glycolipids from higher plants suggested
that ring-closing olefin metathesis (RCM) should allow the
macrocyclic frame of the target to be forged with ease.[26–29]
As the resulting cycloalkene has to be hydrogenated
anyway, benzyl ethers seemed optimal as the protecting
groups for the hydroxyl functions on the disaccharide back-
bone, as they can be removed concomitantly. Since an ade-
quate benzyl-protected sugar constituent is easily attainable
(see below),[27b] the route to the final target was expected to
be straightforward.
treatment of the crude product with Tf2O and Et3N. The
subsequent alkoxycarbonylation reaction also proceeded
smoothly under the conditions set forth in the model study.
Apprehensive that the saponification of a methyl ester
might be problematic once the double bond in 18 is reduced
and the all-cis substitution pattern set,[12] we opted to pre-
pare the trimethylsilylethyl ester 18; in this case, epimeriza-
tion seems less likely, since the ester cleavage can be effect-
ed under mild conditions with the aid of fluoride. For the
high-yielding preparation of 18, it sufficed to replace the
methanol used as a co-solvent in the model study outlined
above by trimethylsilylethanol (5 equiv) to trap the acylpal-
ladium species formed in the actual carbonylation step.
Next, the terminal benzyl ether was hydrogenolytically
cleaved over Pd/C, because direct exposure of 18 to Rh/
Al2O3 would saturate the arene ring. After this minor
detour, compound 19 was reduced under the described con-
ditions to give product 20 in 81% yield. Since the minor cis-
trans-configured isomer 21 could be separated by flash chro-
matography, the stage seemed set for the completion of the
total synthesis.
In the event, the readily available aldehyde 14 was sub-
jected to an enantioselective catalytic hydrocyanation with
TMSCN (Scheme 2). A mixture of TiACTHNUTRGNENUG(OiPr)4 (10 mol%) and
the bifunctional BINOL-derived ligand 24 (10.5 mol%)
served this purpose well,[30] which had previously been
shown to give good results with unbranched aliphatic alde-
hydes that are quite challenging substrates otherwise.[21] The
desired cyanohydrin 15 was obtained in respectable 88%
yield and 89% ee. The minor isomer was carried through
until after the esterification of the paraconic acid sector
with the sugar fragment and RCM; at this stage, all isomeric
impurities could be removed by chromatography (see
below).
In preparation for the projected macrocyclization by
RCM,[31,32] the terminal OH group of 20 had to be elabo-
ꢂ
As expected, the a-bromopropionate 16 derived from 15
underwent the intramolecular Blaise reaction with ease, fur-
nishing alkenyl triflate 17 in an overall yield of 77% after
rated into an alkene. While the oxidation to the correspond-
ing aldehyde with PhI
ACHTUNGERTN(NUNG OAc)2 and catalytic TEMPO was un-
eventful,[33] the subsequent olefination was far from trivial.
Chem. Eur. J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
3
&
ÞÞ
These are not the final page numbers!

A. Fꢀrstner et al.
The standard Wittig reagent Ph3P=CH2 turned out to be too
basic, resulting in significant epimerization. This problem
could be largely avoided by recourse to the Takai olefina-
tion,[34] although the yield of the desired product 22 was
only moderate (ꢁ50%). Better results were obtained with
the reagent combination CH2I2/Cp2ZrCl2/Zn in THF,[35]
which furnished alkene 22 in 71% yield.
At this stage, the choice of the trimethylsilylethyl ester
moiety paid dividends, because it could be cleaved with
TBAF in CH2Cl2 at 08C to release the required acid in read-
iness for attachment to the disaccharide core.[36] The 1H
NMR data of 23 display the characteristic fingerprint of the
paraconic acids containing an all-cis configured head-group
(see the Supporting Information).
The inherent bias of the all-cis configured butanolide
head group toward epimerization, foreshadowed by the dif-
ficulties in the olefination step, became a very serious issue
during the seemingly trivial esterification of 23 with the
sugar segment. The required disaccharide component 29 was
obtained by adapting a route previously described by our
group,[27b] using the same monosaccharide building block
25[37] twice (Scheme 3). However, all attempts to condense
29 and the paraconic acid 23 with the aid of carbodiimides,
(benzotriazol-1-yloxy)-tripyrrolidinophosphonium hexa
ACHUTGTNRENNUGfluACHTUNGTERNoNUGN -
ACHTUNGTRENNUNGroACHTUNGTRENNUNGphosphate (PyBOP), bis-(2-oxo-3-oxazolidinyl)phosphinic
acid chloride (BOPCl), Mukaiyamaꢂs reagent,[38] 1-hydroxy-
1H-benzotriazole (HOBt), or under Shiina esterification
conditions,[39] in the absence or presence of different basic
or acidic promoters, either met with failure or resulted in
rapid epimerization at the ester site with formation of the
more stable trans,trans-isomer 3-epi-30 as the only product.
The characteristic change in the J pattern of the protons on
the butanolide rim allows this change to be readily
diACHTUNGTRENNUNG
Scheme 3. a) 3-Butenyl magnesium bromide, CuCN (10 mol%), THF,
ꢂ788C ! RT, 84 %; b) BF3·Et2O, CH2Cl2/pentane 1:1, ꢂ208C, 76%;
c) NaOMe, MeOH, 89%; d) 25, BF3·Et2O, CH2Cl2/pentane 1:1, ꢂ208C;
e) NaOMe, MeOH, 83% (over both steps); f) 23, N,N’-diisopropylcarbo-
diimide, DMAP, CH2Cl2, 97%; g) i) 23, (COCl)2, DMF cat., CH2Cl2; ii)
Et3N, CH2Cl2, 08C, 66% (30) and 21% (3-epi-30), see text; DMAP = 4-
dimethylaminopyridine.
3
reaction of 29 with the acid chloride derived from 23 in the
presence of Et3N in CH2Cl2 at 08C afforded a product mix-
ture, in which the desired isomer prevailed. Optimization of
this screening hit allowed ester 30 to be isolated in 66%
yield, together with 21% of 3-epi-30, which is separable by
flash chromatography. It is of note that bases other than
Et3N led to far inferior outcomes: even a slight increase of
the size of the tert-amine (e.g., Bu3N) resulted in unaccepta-
bly low conversions,[40] and more bulky amines ((iPr)2NEt,
2,6-lutidine, 2,6-di-tert-butylpyridine) shut the reaction com-
pletely down. In contrast, the use of a mixture of Et3N/
DMAP furnished 3-epi-30, virtually as the only product.
This surprisingly narrow window is tentatively explained by
the in situ formation of acylammonium cations as the actual
acylating agents.[41] If the chosen amine base R3N is too
bulky, the reactive encounter between the strongly shielded
With 30 and 3-epi-30 finally in hand, the projected macro-
cyclization could be tested using the ruthenium indenylidene
complex 31 previously developed by our group as a cost ef-
fective alternative to the first generation Grubbs catalyst
(Scheme 4).[42–44] As expected, both RCM reactions proceed-
ed smoothly, furnishing the desired products 32 and 3-epi-32
in 83 and 91% yield, respectively. Whereas the E/Z isomers
were inseparable in both cases, routine flash chromatogra-
phy allowed the minor diastereomers to be removed at this
stage, which derived from the fact that cyanohydrin 15 uti-
lized for the preparation of the paraconic acid had only had
an ee of 89%. Finally, all peripheral benzyl ethers of isomer-
ically pure 32 and 3-epi-32 were hydrogenolytically cleaved
while saturating the double bond over palladium on char-
coal. For solubility reasons, the reaction was best performed
in a mixture of MeOH/EtOAc 1:1 as the medium, supple-
mented by a small amount of formic acid.
ꢂ
acyl center and the hindered secondary OH of the sugar is
impeded; slim bases, in contrast, have unhindered access to
the acidic proton H3 on the open b-side of the tetrahydro-
furan-2-one ring, causing rapid epimerization (see
Scheme 3).
The detailed analysis of the recorded 1D and 2D NMR
spectra of our synthetic samples leaves no doubt that their
&
4
&
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Total Synthesis of Nominal Gobienine A
FULL PAPER
Table 2. Comparison of the shift and 3J pattern of the synthetic samples
of 1 and 3-epi-1 with the data of the gobienine A aglycone reported in
ref. [4]; for the numbering scheme, see Scheme 4.
Position Gobienine A agly-
cone
Synthetic 1
3-epi-1
H2
H3
H4
2.88 ppm (dq, J =
7.0, 7.5 Hz)
3.24 ppm (dd, J =
7.0, 8.0 Hz)
4.57 ppm (ddd, J =
8.0, 8.0, 5.0 Hz)
3.04 ppm (dq, J =
7.0, 7.0 Hz)
3.49 ppm (dd, J =
6.9, 5.1 Hz)
4.47 ppm (ddd, J =
7.0, 7.0, 5.2 Hz)
3.16 ppm (dq, J =
11.4, 7.1 Hz)
2.87 ppm (dd, J =
11.4, 9.4 Hz)
4.94 ppm (td, J =
9.1, 3.4 Hz)
Table 3. Comparison of the 1H NMR data (600 MHz) (d [ppm], multi-
plicity, J [Hz]) of the disaccharide unit of synthetic 1 with the data of
gobie
scheme, see Scheme 4.
nine A reported in the literature (ref. [4]).[a] For the numbering
ACHTUNGTRENNUNG
Position
“Gobienine A”
Synthetic 1
1’
2’
3’
4’
5.78 d (7.1)
4.54 d (7.0)
3.79 m
3.81 t (9.1)
3.64 dd (9.5, 8.8)
3.44 m
4.09 dd (11.9, 2.3)
3.91 dd (12.0, 5.5)
5.33b
5.07 dd (9.4, 8.2)
3.67 t (9.2)
3.72 t (9.1)
3.44 m
4.06 dd (11.9, 2.3)
3.88 dd (12.0, 5.2)
4.18 dd (7.1, 8.1)
4.26 dd (8.1, 8.2)
4.10 dd (8.2, 8.1)
3.91 ddd (8.1, 6.5, 2.9)
4.21 dd (11.0, 6.5)
4.34 dd (11.0, 2.9)
5.61 d (7.5)
5.73 dd (7.5, 8.0)
4.23 dd (8.0, 8.0)
4.07 dd (8.0, 8.1)
3.88 ddd (8.1, 6.1, 3.2)
4.19 dd (10.8, 6.1)
4.30 dd (10.8, 3.2)
5’
6a’
6b’
1’’
2’’
3’’
4’’
5’’
6a’’
6b’’
Scheme 4. Completion of the total synthesis of nominal gobienine A (1)
and 3-epi-1: a) 31 (10 mol%), CH2Cl2, reflux, 83% (32), 91 % (3-epi-32);
b) H2 (1 atm), Pd/C cat., MeOH/EtOAc/HCOOH, 93% (1), 91%
(3-epi-1).
[a] The spectra of synthetic 1 were recorded in [D4]-MeOH/[D5]-pyridine
1:1 (v/v), cf. ref. [47]; for selected data recorded in CDCl3, see ref. [48]
and the Supporting Information. [b] Signal partly hidden under the sol-
vent peak.
constitution and stereostructure is properly depicted as
drawn in 1 and 3-epi-1, respectively. This includes NOE data
to re-confirm the relative all-cis configuration of the sub-
stituents on the g-lactone ring of 1; the HRMS (ESI+) and
IR data are also consistent with these formulae. Very much
to our dismay, however, neither compound corresponds to
the natural product, albeit the comparison is impeded by
the fact that the reported dataset of gobienine is incom-
plete.[4,45–48] Yet, the differences are so significant that we
firmly believe our conclusion to be definite.
Table 4. Comparison of the 13C NMR shifts (150 MHz) (d [ppm]) of the
disaccharide unit of synthetic 1 and 3-epi-1 with the data of gobienine A
reported in the literature (ref. [4]).[a] For the numbering scheme, see
Scheme 4.
Position
“Gobienine A”
Synthetic 1
3-epi-1
1’
2’
3’
4’
5’
6’
1’’
2’’
3’’
4’’
5’’
6’’
103.4
80.3
77.4
72.1
76.1
63.4
103.8
80.2
78.1
72.6
77.9
64.1
101.5 (br)[b]
ꢀ 80 (br)[b]
78.2 (br)[b]
71.7
78.5
62.7
100.9
77.1
77.0
72.2
101.3
82.9
77.5
72.3
78.2
63.3
102.0
77.4
76.4
71.4
78.7
62.1
A comparison of the NMR shifts of the sugar segments of
the synthetic samples of nominal gobienine A (1) and 3-epi-
1 with those of the natural product shows a striking mis-
match of the data (Tables 2–4);[46,47] this is particularly true
for the proton- and carbon resonances of either anomeric
center (Tables 3 and 4). In addition, the 13C NMR spectrum
of synthetic 1 features a characteristic line broadening, par-
ticularly for the C-atoms neighbouring the glycosidic bond
between the paraconic acid aglycone and the first glucopyra-
nose (C1’–C3’ and C17–C19). The isolation paper does not
mention this very distinctive phenomenon.[49,50]
78.7
62.6
[a] The spectra of synthetic 1 and 3-epi-1 were recorded in [D4]-MeOH/
[D5]-pyridine 1:1 (v/v), cf. ref. [47]. [b] br = broad.
3
Furthermore, the pattern of the J coupling constants for
aglycone). We like to emphasize that this spectral finger-
print is very characteristic (Figure 2) and was consistently
observed throughout the entire series of butanolides with an
all-cis substitution pattern prepared during our synthesis
the protons on the g-lactone ring is not matching (Table 2).
3
In particular, the JH3,H4 significantly deviates from the re-
ported value (5.1 Hz in 1, but 8.0 Hz in the gobienine A
Chem. Eur. J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
5
&
ÞÞ
These are not the final page numbers!

A. Fꢀrstner et al.
1
Figure 2. Comparison of the low-field region of the H NMR spectra of 3-epi-1 (top, green) and 1 (bottom, brown), which shows the very diagnostic dif-
ferences in the shifts and multiplicities caused by the different substitution patterns on the paraconic acid aglycone.
NMR data of the gobienine A aglycone with paraconic acid derivatives
of known configuration, and copies of spectra of new compounds.
campaign, independent of whether they carry an ester or a
free carboxylic acid at C3. Likewise, this coupling pattern is
basically unaffected by macrocyclization and largely inde-
pendent of whether the spectra were recorded in CDCl3 or
[D4]-MeOH/[D5]-pyridine (for details, see the Supporting
Information).[46–48] Furthermore, our spectral data map well
Acknowledgements
onto other literature reports concerning paraconic acid de-
rivatives of known configuration.[7,25]
Generous financial support by the Max-Planck-Gesellschaft and the
Finally we note that compounds representing all four pos-
sible diastereomers of the proposed 4-carboxy-g-butyrolac-
tone entity embedded into gobienine A were eventually pre-
pared and characterized during our program. It is puzzling
that the literature data of the gobienine A aglycone corre-
spond to none of them really well (for details, see the Sup-
porting Information).
Fonds der Chemischen Industrie (Kekulꢃ stipend to A.A.) is gratefully
acknowledged. We thank the analytical departments of our Institute for
the excellent support of our program.
[1] Lichen Biology (Ed.: T. H. Nash), Cambridge University Press,
Cambridge (UK), 1996.
[2] S. Huneck, Symbiosis 1991, 11, 225–248.
[3] Review: R. Bandichhor, B. Nosse, O. Reiser, in Natural Product
Synthesis I (Ed.: J. Mulzer), Top. Curr. Chem., Vol. 243, Springer,
Berlin, 2005, pp. 43–72.
Conclusion
ˇ
[4] T. Rezanka, I. A. Guschina, Phytochemistry 2001, 58, 1281–1287.
[5] For closely related but non-macrocyclic glycoside esters of paraconic
ˇ
ˇ
acids isolated from the same or closely related lichens, see: a) T. Re-
We hence conclude that the structure originally attributed
to the complex lichen-derived glycoconjugate gobienine A is
incorrect. We can ascertain that it is not the very epimeriza-
tion-prone ester site on the butanolide ring that is the locus
of misassignment. Rather, the spectral data of synthetic 1
and 3-epi-1 suggest that the discrepancy is more profound.[51]
Speculation aside, the magnitude and scatter of the mis-
match, the incomplete data set, the lack of an authentic
sample, and the unavailability of the original spectra make
it impossible for us to answer the question as to the correct
structure of this intriguing lichen-derived metabolite at this
point.[45] This specific issue notwithstanding, the concise
entry into tetronic acids as well as all-cis configured para-
conic acids developed during this campaign closes a synthet-
ic gap in the coverage of these classes of compounds and
might facilitate future investigations into their chemistry
and biology.
zanka, I. A. Guschina, Phytochemistry 2001, 58, 509–516; b) T. Re-
zanka, I. A. Guschina, Phytochemistry 2001, 56, 181–188.
[6] The assignments were based on the Mosher analysis of the aglycone
and on correlations of its ORD spectra with earlier data.
[7] For leading references and extensive literature compilations, see:
a) P. A. Jacobi, P. Herradura, Tetrahedron Lett. 1996, 37, 8297–8300;
b) A. Forster, J. Fitremann, P. Renaud, Tetrahedron Lett. 1998, 39,
7097–7100; c) J. Mulzer, N. Salimi, H. Hartl, Tetrahedron: Asymme-
try 1993, 4, 457–471; d) J. Mulzer, U. Steffen, H. J. Martin, L. Zorn,
Eur. J. Org. Chem. 2005, 1028–1043; e) M. Bella, R. Margarita, C.
Orlando, M. Orsini, L. Parlanti, G. Piancatelli, Tetrahedron Lett.
2000, 41, 561–565; f) P. A. Jacobi, P. Herradura, Can. J. Chem. 2001,
79, 1727–1735; g) M. Amador, X. Ariza, J. Garcia, Heterocycles
2006, 67, 705–720; h) M. T. Barros, C. D. Maycock, M. R. Ventura,
Org. Lett. 2003, 5, 4097–4099; i) C. Le Floch, E. Le Gall, E. Lꢃonel,
T. Martens, T. Cresteil, Bioorg. Med. Chem. Lett. 2011, 21, 7054–
7058; j) R. B. Chhor, B. Nosse, S. Sçrgel, C. Bçhm, M. Seitz, O.
Reiser, Chem. Eur. J. 2003, 9, 260–270; k) M. P. Sibi, P. Liu, J. Ji, S.
Hajra, J. Chen, J. Org. Chem. 2002, 67, 1738–1745; l) T. Martꢄn,
C. M. Rodrꢄguez, V. S. Martꢄn, J. Org. Chem. 1996, 61, 6450–6453;
m) M. Amador, X. Ariza, J. Garcia, J. Ordiz, J. Org. Chem. 2004, 69,
8172–8175; n) S. Drioli, F. Felluga, C. Forzato, P. Nitti, G. Pitacco,
E. Valentin, J. Org. Chem. 1998, 63, 2385–2388; o) Y.-S. Hon, C.-H.
Hsieh, H.-F. Chen, Synth. Commun. 2007, 37, 1635–1651; p) S. Shi-
mada, Y. Hashimoto, K. Saigo, J. Org. Chem. 1993, 58, 5226–5234;
q) A. M. Jacobine, W. Lin, B. Walls, C. K. Zercher, J. Org. Chem.
2008, 73, 7409–7412.
Experimental Section
All experimental details can be found in the Supporting Information.
The material includes compound characterization, comparison of the
&
6
&
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Total Synthesis of Nominal Gobienine A
FULL PAPER
[8] R. Pereda-Miranda, D. Rosas-Ramꢄrez, J. CastaÇeda-Gꢅmez, in
Prog. Chem. Org. Nat. Prod., Vol. 92 (Eds.: A. D. Kinghorn, H.
Falk, J. Kobayashi), Springer, Wien, 2010, pp. 77–153.
[9] a) R. Pereda-Miranda, M. Bah, Curr. Top. Med. Chem. 2003, 3, 111–
131; b) V. M. Dembitsky, Chem. Biodiversity 2004, 1, 673–781; c) J.
Furukawa, N. Sakairi, Trends Glycosci. Glycotechnol. 2001, 13, 1–
10.
[31] a) T. M. Trnka, R. H. Grubbs, Acc. Chem. Res. 2001, 34, 18–29;
b) A. Fꢀrstner, Angew. Chem. 2000, 112, 3140–3172; Angew. Chem.
Int. Ed. 2000, 39, 3012–3043; c) K. C. Nicolaou, P. G. Bulger, D.
Sarlah, Angew. Chem. 2005, 117, 4564–4601; Angew. Chem. Int. Ed.
2005, 44, 4490–4527; d) A. H. Hoveyda, A. R. Zhugralin, Nature
2007, 450, 243–251.
[32] A. Fꢀrstner, Chem. Commun. 2011, 47, 6505–6511.
[33] A. De Mico, R. Margarita, L. Parlanti, A. Vescovi, G. Piancatelli, J.
Org. Chem. 1997, 62, 6974–6977.
[10] A. Fꢀrstner, Eur. J. Org. Chem. 2004, 943–958.
[11] The isolation team called the aglycone 4 “ACTHNUTRGENUG(N 18R)-hydroxy-dihydro-
[34] a) T. Okazoe, K. Takai, K. Utimoto, J. Am. Chem. Soc. 1987, 109,
951–953; b) review: A. Fꢀrstner, Chem. Rev. 1999, 99, 991–1045.
[35] J. M. Tour, P. V. Bedworth, R. Wu, Tetrahedron Lett. 1989, 30, 3927–
3930.
[36] Attempted saponification of the analogous methyl ester under a va-
riety of conditions either resulted in no conversion or in the forma-
tion of complex product mixtures.
[37] A. Fꢀrstner, I. Konetzki, Tetrahedron Lett. 1998, 39, 5721–5724.
[38] T. Mukaiyama, M. Usui, E. Shimada, K. Saigo, Chem. Lett. 1975,
1045–1048.
[39] I. Shiina, M. Kubota, H. Oshiumi, M. Hashizume, J. Org. Chem.
2004, 69, 1822–1830.
[40] In line with the proposed mechanism, the use of Bu3N instead of
Et3N results in higher diastereoselectivity (ca. 7:1) but a very low
yield.
[41] a) P. Hubbard, W. J. Brittain, J. Org. Chem. 1998, 63, 677–683;
b) J. A. King, Jr., G. L. Bryant, Jr. J. Org. Chem. 1992, 57, 5136–
5139; c) J. V. Paukstelis, M. Kim, J. Org. Chem. 1974, 39, 1503–1507.
[42] A. Fꢀrstner, O. Guth, A. Dꢀffels, G. Seidel, M. Liebl, B. Gabor, R.
Mynott, Chem. Eur. J. 2001, 7, 4811–4820.
allo-protolichesterinic acid.” We note, however, that this terminolo-
gy is ambiguous since dihydro-allo-protolichesterinic acid, as previ-
ously defined in ref. [13], has a side chain of only 13-carbon atoms
branching off the all-cis-configured butanoide ring, whereas com-
pound 4 contains a side chain of 15-carbon atoms.
[12] As one might expect, trans-trans configured paraconic acid deriva-
tives are the most stable isomers, cf.: E. E. van Tamelen, S. R. Bach,
J. Am. Chem. Soc. 1958, 80, 3079–3086.
[13] For proto- and allo-protolichesterinic acid, see: S. Huneck, R.
Takeda, Z. Naturforsch. B Chem. Sci. 1992, 47, 842–854.
[14] For a review on natural products of mistaken identity, see: K. C.
Nicolaou, S. A. Snyder, Angew. Chem. 2005, 117, 1036–1069;
Angew. Chem. Int. Ed. 2005, 44, 1012–1044.
[15] Y. Yang, D. Wang, Synlett 1997, 1379–1380.
[16] a) A. Fꢀrstner, Synthesis 1989, 571–590; b) R. Ocampo, W. R. Dol-
bier, Jr., Tetrahedron 2004, 60, 9325–9374.
[17] H. S. P. Rao, S. Rafi, K. Padmavathy, Tetrahedron 2008, 64, 8037–
8043.
[18] P. Pçchlauer, P. Riebel, H. Mayrhofer, I. Wirth, Patent, US 6417377
B1, 2002.
[19] a) L. R. Krepski, L. E. Lynch, S. M. Heilmann, J. K. Rasmussen, Tet-
rahedron Lett. 1985, 26, 981–984; b) T. Kitazume, J. Fluorine Chem.
1987, 35, 287–294; c) J. J. Duffield, A. C. Regan, Tetrahedron:
Asymmetry 1996, 7, 663–666; d) J. R. Anderson, R. L. Edwards,
A. J. S. Whalley, J. Chem. Soc. Perkin Trans 1 1982, 215–221.
[20] a) T. R. Kelly, N. S. Chandrakumar, J. D. Cutting, R. R. Goehring,
F. R. Weibel, Tetrahedron Lett. 1985, 26, 2173–2176; b) J. Chiarello,
M. M. Joulliꢃ, Synth. Commun. 1989, 19, 3379–3383.
[21] W. T. Wang, L. L. Lin, X. H. Liu, X. M. Feng, in Science of Synthesis,
Knowledge Updates, Vol. 2011/4, Thieme, Stuttgart, 2012, pp. 377–
444.
[22] Review: A. Brennfꢀhrer, H. Neumann, M. Beller, Angew. Chem.
2009, 121, 4176–4196; Angew. Chem. Int. Ed. 2009, 48, 4114–4133.
[23] The use of the following catalysts was unsuccessful: Pd/C, Pd(OH)2,
[43] Ruthenium-indenylidene based metathesis catalysts are sold under
the tradename “Neolyst” (Umicore, Hanau) and “catMETium”
(Degussa).
[44] Selected applications in total synthesis: a) A. Fꢀrstner, J. Grabowski,
C. W. Lehmann, J. Org. Chem. 1999, 64, 8275–8280; b) A. Fꢀrstner,
O. R. Thiel, J. Org. Chem. 2000, 65, 1738–1742; c) A. Fꢀrstner, M.
Schlede, Adv. Synth. Catal. 2002, 344, 657–665; d) B. Scheiper, F.
Glorius, A. Leitner, A. Fꢀrstner, Proc. Natl. Acad. Sci. USA 2004,
101, 11960–11965; e) A. Fꢀrstner, K. Radkowski, C. Wirtz, R. God-
dard, C. W. Lehmann, R. Mynott, J. Am. Chem. Soc. 2002, 124,
7061–7069; f) A. Fꢀrstner, A. Korte, Chem. Asian J. 2008, 3, 310–
318; g) V. Hickmann, A. Kondoh, B. Gabor, M. Alcarazo, A. Fꢀrst-
ner, J. Am. Chem. Soc. 2011, 133, 13471–13480.
[45] The corresponding author (T.R.) of the isolation paper informed us
that he neither holds samples nor the original NMR spectra any
longer.
[46] For the proper natural product, only the NMR data of the sugar seg-
ment are reported in the isolation paper, whereas the data of its
[(Ph3P)3RhCl], [(Ph3P)3RuCl2], [Rh
(PCy3)]BF4; the use of PtO2 led to partial reduction but suffered
from poor reproducibility.
ACHTUNGTRENGU(N nbd)2]BF4/PPh3, [IrACHTUNGTERN(NUGN cod)(py)-
ACHTUNGTRENNUNG
[24] For a related reduction of tetrasubstituted olefins lacking the car-
boxylate branch, see: M. K. Gurjar, J. Cherian, C. V. Ramana, Org.
Lett. 2004, 6, 317–319.
[25] S. Huneck, T. Tønsberg, F. Bohlmann, Phytochemistry 1986, 25,
453–459; for the correction of the assignment of one of the com-
pounds contained in this paper, see ref. [7p].
[26] a) A. Fꢀrstner, T. Mꢀller, J. Am. Chem. Soc. 1999, 121, 7814–7821;
b) A. Fꢀrstner, T. Mꢀller, J. Org. Chem. 1998, 63, 424–425; c) C. W.
Lehmann, A. Fꢀrstner, T. Mꢀller, Z. Kristallogr. 2000, 215, 114–117.
[27] a) A. Fꢀrstner, F. Jeanjean, P. Razon, Angew. Chem. 2002, 114,
2203–2206; Angew. Chem. Int. Ed. 2002, 41, 2097–2101; b) A.
Fꢀrstner, F. Jeanjean, P. Razon, C. Wirtz, R. Mynott, Chem. Eur. J.
2003, 9, 307–319; c) A. Fꢀrstner, F. Jeanjean, P. Razon, C. Wirtz, R.
Mynott, Chem. Eur. J. 2003, 9, 320–326.
paraACTHNUGRTENUNGconic acid sector are given only for downstream products de-
rived from gobienine by enzymatic and/or acid catalyzed hydrolysis.
[47] A further complication arises from the fact that it is not entirely
clear in which solvent the spectra of the gobienines were recorded.
We suppose that a 1:1 mixture of [D4]-MeOH/[D5]-pyridine was
used, which had been routinely employed by the isolation team in
other studies on related lichen-derived glycosylated paraconic acid
derivatives (ref.[5]). Moreover, the spectra of the paraconic acid
aglycone derived from gobienine A by chemical and enzymatic deg-
radation were recorded in [D4]-MeOH/[D5]-pyridine (the dataset
contained in the gobienine isolation paper (ref. [4]) corresponds to a
previous dataset reported in ref. [5], wherein the NMR solvent is
defined).
[48] We point out that our synthetic samples of 1 and 3-epi-1 are only
sparingly soluble in CDCl3 or CD2Cl2. Although the quality of the
recorded 1H, COSY and HSQC spectra was therefore not optimal,
some indicative data could be extracted. For 1, this includes the
complete set of shifts and 3J coupling constants of the protons on
the butanolide ring, which are compiled in the Supporting Informa-
tion. Moreover, the signals of H-1’’ (5.03 ppm, d, J = 7.5 Hz), H-2’’
(4.72 ppm, t, J = 8.6 Hz), C-1’’ (99.0 ppm) and C-2’’ (75.4 ppm)
could be tentatively assigned. A comparison with the corresponding
[28] A. Fꢀrstner, K. Radkowski, J. Grabowski, C. Wirtz, R. Mynott, J.
Org. Chem. 2000, 65, 8758–8762.
[29] a) A. Fꢀrstner, T. Nagano, J. Am. Chem. Soc. 2007, 129, 1906–1907;
ˇ
b) T. Nagano, J. Pospꢄsil, G. Chollet, S. Schulthoff, V. Hickmann, E.
Moulin, J. Herrmann, R. Mꢀller, A. Fꢀrstner, Chem. Eur. J. 2009,
15, 9697–9706.
[30] F. Yang, S. Wei, C.-A. Chen, P. Xi, L. Yang, J. Lan, H.-M. Gau, J.
You, Chem. Eur. J. 2008, 14, 2223–2231.
Chem. Eur. J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
7
&
ÞÞ
These are not the final page numbers!

A. Fꢀrstner et al.
data recorded in [D4]-MeOH/[D5]-pyridine makes clear that solvent
effects are rather small and hence unlikely the reason for the severe
mismatch between synthetic 1 and gobienine A, as reported in ref.
[4].
fore we are inclined to believe in cumulative misassignments within
the structural framework.
[50] Signals of the carbons next to the second glycosidic linkage (C1’’,
C2’’) as well as of C3 on the butanolide ring and even of some C-
atoms within the aliphatic tether are also somewhat broadened, see
the Supporting Information.
[51] For a related case, see: A. Larivꢃe, J. B. Unger, M. Thomas, C.
Wirtz, C. Dubost, S. Handa, A. Fꢀrstner, Angew. Chem. 2011, 123,
318–323; Angew. Chem. Int. Ed. 2011, 50, 304–309.
[49] The significant line broadening in vicinity to the C18 stereocenter of
1 might suggest that the configuration of this site in putative gobie-
nine A was misassigned; in this context we reiterate that C18 in
gobienine A is supposedly R-configured, whereas the sister com-
pounds gobienine B and C are supposedly S-configured, see Intro-
duction. However, such a mistake alone would neither easily explain
the dramatic shift difference of the anomeric proton H-1’ of synthet-
ic 1 versus that of putative gobienine A (DdH
=
1.24 ppm, cf.
Received: March 4, 2013
Table 3) nor the observed mismatch in the butanolide ring. There-
Published online: && &&, 0000
&
8
&
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Total Synthesis of Nominal Gobienine A
FULL PAPER
Enigmatic beauty: The putative lichen-
derived glycolipid gobienine A is struc-
turally rather unique, not least because
of its thermodynamically unfavorable
all-cis paraconic acid (4-carboxy-g-
butyrolactone) aglycone. A largely cat-
alysis-based and broadly applicable
entry into this unusual motif was
developed, and the total synthesis of
the target glycoconjugate completed,
just to find out that the originally pro-
posed structure must have been pro-
foundly misassigned.
Total Synthesis
A. Kondoh, A. Arlt, B. Gabor,
A. Fꢀrstner* ................... &&&& — &&&&
Total Synthesis of Nominal Gobienine
A
Chem. Eur. J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
9
&
ÞÞ
These are not the final page numbers!