Enantioselective Organocatalytic Cascade Approach to Different Classes of Benzofused Acetals

Article abstract of DOI:10.1002/chem.201602992

A novel enantioselective organocatalytic strategy is presented for the synthesis of tetrahydrofurobenzofuran and methanobenzodioxepine natural product core structures. The strategy is based on a pair of divergent reaction pathways in which hydroxyarenes react with γ-keto-α,β-unsaturated aldehydes, catalyzed by a chiral secondary amine. One reaction pathway, which leads to chiral 5,5-fused acetals with two stereocenters—the tetrahydrofurobenzofuran scaffolds—proceeds in moderate yields and up to 96 % ee. The other reaction pathway provides 5,6-bridged methanobenzodioxepine scaffolds with three stereocenters in moderate to good yields and up to 95 % ee. The reaction is remarkable as it can proceed with catalyst loadings as low as 0.25 mol %, providing one of the highest known turnover numbers in iminium ion catalysis. Furthermore, the hemiacetal tetrahydrofurobenzofuran can undergo functionalizations including reduction, oxidation, and allylation. Finally, the effects involved in the substrate control for the divergent pathways, based on both experimental and computational studies, have been investigated. A model involving steric, electronic and stereoelectronic interactions is discussed to rationalize the observed selectivities.

Full text of DOI:10.1002/chem.201602992

DOI: 10.1002/chem.201602992
Full Paper
&
Organocatalysis |Hot Paper|
Enantioselective Organocatalytic Cascade Approach to Different
Classes of Benzofused Acetals
Bruno Matos Paz, Lydia Klier, Line Næsborg, Vibeke Henriette Lauridsen, Frank Jensen, and
Karl Anker Jørgensen*[a]
Chem. Eur. J. 2016, 22, 1 – 10
1
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Abstract: A novel enantioselective organocatalytic strategy
is presented for the synthesis of tetrahydrofurobenzofuran
and methanobenzodioxepine natural product core struc-
tures. The strategy is based on a pair of divergent reaction
pathways in which hydroxyarenes react with g-keto-a,b-un-
saturated aldehydes, catalyzed by a chiral secondary amine.
One reaction pathway, which leads to chiral 5,5-fused acetals
with two stereocenters—the tetrahydrofurobenzofuran scaf-
folds—proceeds in moderate yields and up to 96% ee. The
other reaction pathway provides 5,6-bridged methanoben-
zodioxepine scaffolds with three stereocenters in moderate
to good yields and up to 95% ee. The reaction is remarkable
as it can proceed with catalyst loadings as low as
0.25 mol%, providing one of the highest known turnover
numbers in iminium ion catalysis. Furthermore, the hemiace-
tal tetrahydrofurobenzofuran can undergo functionalizations
including reduction, oxidation, and allylation. Finally, the ef-
fects involved in the substrate control for the divergent
pathways, based on both experimental and computational
studies, have been investigated. A model involving steric,
electronic and stereoelectronic interactions is discussed to
rationalize the observed selectivities.
Introduction
Chiral fused, bridged, and spirocyclic acetals are widespread in
nature.[1] The bioactivity of these compounds is often depen-
dent on the substituent pattern and stereochemistry of the
acetal functionality, which has motivated numerous efforts to-
wards the development of concise stereoselective syntheses of
such acetals.[2]
Restricting attention to benzofused acetals, some of these
natural compounds and their synthetic analogues possess
a
wide range of biological activities. Whereas aflatoxins
(Figure 1, A), widespread food contaminants, are toxic and car-
cinogenic mycotoxins,[3] other highly desirable antimicrobial,
antifungal, anti-HIV, and anticancer properties are found in re-
lated benzofused acetal scaffolds (Figure 1, C).[4]
Since pioneering works by the groups of Bꢁchi and Rob-
erts,[5] several racemic synthetic endeavors have been devel-
oped for the construction of the tetrahydrofurobenzofuran
system present in aflatoxins (Figure 1, A).[6] However, more re-
cently, enantioselective approaches aiming for the natural
product synthesis of the tetrahydrofurobenzofuran scaffold
have been put forward. These strategies have relied on the use
of chiral Lewis acid catalysis for the activation of para-qui-
nones[7] or palladium-catalyzed acetalization followed by re-
ductive Heck coupling.[8] The first strategy restricted the substi-
tution pattern of the aromatic moiety, whereas the second
relied on prefunctionalization of the hydroxyarenes used, and
both were mainly limited to hydrogen substituents on the
acetal moiety.[9]
Figure 1. Biologically active compounds containing benzofused acetal scaf-
folds.
In striking contrast to the tetrahydrofurobenzofuran core,
the methanobenzodioxepine scaffold, present in bullataketals
(Figure 1, B), has received little attention and no enantioselec-
tive synthesis of this scaffold has been reported to date.[10] Fur-
thermore, the tetrahydrofurobenzofuran and methanobenzo-
dioxepine scaffolds are both present in butyrylcholinesterase
inhibitors with potential use in the treatment of neurodegener-
ative diseases such as Alzheimer’s disease.[11]
Driven by the intriguing structural complexity and remark-
able biological properties of the tetrahydrofurobenzofuran and
methanobenzodioxepine scaffolds, we decided to investigate
the feasibility of enantioselective organocatalytic divergent
pathways for the synthesis of these scaffolds. The present
strategy is based on the hypothesis that an iminium ion-cata-
lyzed reaction between hydroxyarenes and g-keto-a,b-unsatu-
rated aldehydes (g-keto-enals) could give access to both scaf-
folds in an enantioselective fashion, by means of a Friedel–
[a] B. M. Paz, Dr. L. Klier, L. Næsborg, V. H. Lauridsen, Prof. Dr. F. Jensen,
Prof. Dr. K. A. Jørgensen
Department of Chemistry, Aarhus University
DK-8000 Aarhus C (Denmark)
E-mail: kaj@chem.au.dk
Supporting information for this article can be found under:
http://dx.doi.org/10.1002/chem.201602992.
&
&
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
2
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Table 1. Optimization of solvent and acid cocatalyst for the formation of
tetrahydrofurobenzofurans.[a,b]
Entry Cat. [mol%] Cocat. [mol%] Acid cocat.
t [h][b] ee[c] [%]
1
2
3
4
5
6
7
8
9
10
10
10
10
10
2.5
1
0.5
0.5
0.5
0.5
–
10
10
10
10
1
0.5
2.5
5
–
2
<1
<1
60
62
76
94
94
87
85
92
93
94
tBuCO2H
PhCO2H
o-NO2C6H4CO2H <1
o-NO2C6H4CO2H
o-NO2C6H4CO2H
o-NO2C6H4CO2H 24
o-NO2C6H4CO2H 15
o-NO2C6H4CO2H 15
o-NO2C6H4CO2H 15
Scheme 1. Iminium ion-catalyzed formation of benzofused acetals: tetrahy-
drofurobenzofuran (6) and methanobenzodioxepine (7) scaffolds.
3
6
Crafts reaction[12] followed by an acetalization cascade
(Scheme 1).
10
The two different reaction products are formed through
a pair of divergent pathways from the same classes of starting
compounds. The first path (Scheme 1, I; black arrows) leads to
a chiral 5,5-fused acetal bearing two stereogenic centers, in-
cluding a quaternary center, whereas the second (Scheme 1, I;
green arrows) provides a 5,6-bridged product with three ste-
reocenters. Given the usually low turnover numbers of organo-
catalyzed reactions, a further development presented is the
use of low catalyst loadings, and the use of commercially avail-
able hydroxyarenes as nucleophiles. Finally, we have also inte-
grated a series of experimental and computational studies in
an attempt to account for the origins of a remarkable influence
of substrate control in the product distribution.
[a] Reactions were performed on a 0.1 mmol scale, with 1.2 equivalents of
1-naphthol; [b] full conversion (>95%) was determined by 1H NMR spec-
troscopy of the crude reaction mixture; [c] determined by chiral stationary
phase chromatography (UPC2) for 6a.
(Table 1, entry 4). For each reaction screened, 5 equivalents of
water were applied since, in the absence of water, no product
formation was observed.
The short reaction times (<1 h) prompted us to investigate
the influence of catalyst and acid cocatalyst loadings (Table 1,
entries 5–10). By reducing the catalyst loading to 2.5 mol%,
full conversion into 6a was obtained in 3 h, without loss of
enantioselectivity (Table 1, entry 5; 94% ee). However, a drop in
enantioselectivity of 6a was observed when catalyst and acid
loadings were simultaneously reduced (Table 1, entries 6 and
7). The optimal compromise between reaction time and selec-
tivity was obtained by applying a catalyst loading of 0.5 mol%
and an acid loading of 10 mol% (Table 1, entry 10; 94% ee).
Lower catalyst loadings led to very long reaction times
(>48 h). Higher acid loadings did not reduce the reaction time
or improve the enantioselectivity, but provided an increased
amount of unidentified byproducts (results not shown). With
the final conditions in hand, we decided to explore the two re-
action pathways.
Results and Discussion
The reaction between g-keto-enal 1a and 1-naphthol 2a was
used as model to test our hypothesis. When diarylprolinol silyl
ether 3a[13] was applied as the catalyst (10 mol%), to our de-
light, full conversion was rapidly obtained (Table 1). However,
the products were formed as a mixture of epimers of hemiace-
tals 4a and 5a. Upon reduction with BF3·OEt2/HSiEt3, product
6a was obtained as the major product, together with 7a as
the minor product. To optimize the reaction conditions for the
formation of tetrahydrofurobenzofuran 6a (5,5-reaction path,
Scheme 1), the influence of various reaction parameters was in-
vestigated. Full conversion was observed in 2 h in the absence
of acid; however, the isolated product 6a was formed in lower
enantioselectivity (Table 1, entry 1; 60% ee). By applying pivalic
acid, full conversion was observed in <1 h, albeit with an
almost equally low enantioselectivity (Table 1, entry 2; 62% ee).
When benzoic acid was applied as a cocatalyst, the enantiose-
lectivity was increased to 76% ee. By testing benzoic acids
with different substituent patterns,[14] ortho-nitrobenzoic acid
proved to be the best and 6a was formed with 94% ee
To understand the nature and extension of substrate control
for the 5,5 and 5,6 reaction pathways (Scheme 1) leading to
the tetrahydrofurobenzofuran 6 and methanobenzodioxepine
7 scaffolds, respectively, the influence of the R1 substituent on
the g-keto-enals was investigated. The ratio between 6 and 7
is postulated to be dependent on the relative amounts of
hemiacetals 4 and 5, which are in equilibrium via intermediate
I (Scheme 1). As the mixture of four isomers rendered impracti-
cal the direct measurement of the equilibrium ratio between 4
and 5 prior to reduction, an indirect measurement was used
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
3
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
methanobenzodioxepines 7 by exploring the reactivity of aro-
matic substituted g-keto-enals (Scheme 2). By using aldehyde
1h, which bears a phenyl side-chain, acetal 7h was obtained
in 69% yield and 94% ee. 4-Chloro-1-naphthol 2b also reacted
smoothly. However, 7k was formed in slightly lower enantiose-
lectivity (87% ee), which could be solved by applying 1 mol%
of the bulkier aminocatalyst 3b (92% ee).
Table 2. Substrate control for the formation of 6 and 7.[a]
Entry
R1
6:7
1
2
CH3 (1b)
CH2CH3 (1a)
2.3:1
9:1
3
4
5
CH(CH3)2 (1c)
C(CH3)3 (1d)
CH2Ph (1e)
>20:1
>20:1
9:1
6
7
8
9
CH2CH2Ph (1 f)
CH2(CH2)3CH3 (1g)
Ph (1h)
p-CNC6H4 (1i)
p-NO2C6H4 (1j)
>20:1
3.3:1
1:>20
1:4.5
10
1.2:1
[a] Ratios were determined by 1H NMR spectroscopy of the crude reaction
mixture after the reduction step. For product 6, the ratio also includes the
amount of a benzofuran byproduct.[14]
instead, based on the ratios of the reduced products 6 and 7
(Table 2).[14]
The product distribution providing the tetrahydrofurobenzo-
furan 6 and methanobenzodioxepine 7 scaffolds is highly de-
pendent on the R1 substituent in the g-keto-enal 1 (Table 2).
For R1 =CH3, the reaction shows a small preference towards
the formation of 6b (Table 2, entry 1), whereas for R1 =CH2CH3
product 6a is formed with a 9:1 selectivity relative to 7a
(Table 2, entry 2). With more sterically demanding groups in
the aldehyde side-chain (R1 =CH(CH3)2 and C(CH3)3), the selec-
tivity of the reaction changes exclusively to 6c or 6d (Table 2,
entries 3 and 4; >20:1). In the presence of a benzyl group
(R1 =CH2Ph), the selectivity towards 6e is 9:1 and, by elongat-
ing the side-chain with one more methylene unit (R1 =
CH2CH2Ph), the reaction course is completely shifted towards
Scheme 2. Scope for the enantioselective formation of methanobenzodioxe-
pine. Reactions were performed on a 0.25 mmol scale, with 1.5 equivalents
of g-keto-enal. Full conversion (>95%) and ratios were determined by
1H NMR spectroscopy of the crude reaction mixture. [a] Reaction performed
with 1 mol% of catalyst 3b; [b] reaction performed with 0.25 mol% of cata-
lyst 3a.
6 f. However, with
a
longer linear alkyl chain (R1 =
CH2(CH2)3CH3), a lower selectivity of 3.3:1 is found towards 6g.
This is in striking contrast to the substrate with R1 =C6H5,
which forms exclusively 7h (Table 2, entry 8). However, this
change in selectivity is diminished in the presence of electron-
withdrawing groups in the para position of the aromatic
system; for R1 =p-CNC6H4, the 6i/7i ratio is 1:4.5, whereas for
R1 =p-NO2C6H4 the reaction towards 6j and 7j is almost com-
pletely unselective (Table 2, entries 9 and 10; 1.2:1).
For the reaction of the more electron-rich 4-methoxy-1-
naphthol 2c, only 0.25 mol% of catalyst 3a was required,
which afforded 7l in 63% yield and 89% ee. To our delight,
not only naphthols but also an electron-rich phenol reacted.
When sesamol 2d was applied as nucleophile, product 7m
was obtained in 56% yield and 78% ee. Unfortunately, increas-
ing the size of the catalyst protecting group did not improve
the enantioselectivity. 1-Naphthol 2a reacted smoothly with ar-
omatic g-keto-enals bearing different substituent patterns, af-
fording methanobenzodioxepines 7 in moderate to high yields
(54–79%) and high enantioselectivities (90–95% ee). The pres-
ence of a p-methyl group in the aromatic moiety of the g-
5,6 Reaction pathway: Scope of methanobenzodioxepine
scaffolds
Based on our observations of the substrate control (Table 2),
we started the investigation of the scope for the formation of
&
&
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
4
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
keto-enal 1k afforded 7n in 79% yield and 90% ee. Since the
incorporation of a fluorine atom can change the biological
properties and metabolic pathway of organic compounds,[15]
we tested p-fluorinated g-keto-enal 1l, which afforded 7o in
74% yield and 95% ee. When g-keto-enals 1m and 1i were ap-
plied, bearing a p-bromo and a p-cyano group, respectively,
only 0.25 mol% of catalyst 3a was required. Product 7p was
isolated in 65% yield and 94% ee, whereas 7i was formed in
54% yield and 93% ee. An aldehyde bearing a p-methoxy
group also reacted under the reaction conditions. However,
only decomposition was observed after the reduction step.
The reaction between g-keto-enal 1h and 2a was scaled up
to 5 mmol without observable loss in yield or enantioselectivi-
ty. The methanobenzodioxepine 7h was isolated in 67% yield
and 94% ee.
plied. Product 6c was obtained in 60% yield and 94% ee.
When the benzyl-substituted aldehyde 1e was used, 2 mol%
of 3a was also needed to afford product 6e in 42% yield and
96% ee. For aldehyde substrates 1 f and 1g, product 6 f was
obtained in higher yield and enantioselectivity (59% yield,
95% ee) than 6g (40% yield, 91% ee). g-Keto-enals 1 f and 1g
also reacted with 4-chloro-1-naphthol 2b, forming 6q in
higher yield (62% yield, 90% ee) than 6r (41% yield, 89% ee).
The reaction of benzyl-substituted g-keto-enal 1e with other
nucleophiles also required 2 mol% of catalyst 3a. When 4-
chloro-1-naphthol 2b was used, 6s was obtained in 51% yield
and 88% ee. 4-Methoxy-1-naphthol 2c allowed the formation
of 6t in 41% yield and 92% ee. To our delight, sesamol 2d
also reacted, forming 6u in 47% yield, albeit with lower enan-
tioselectivity (84% ee).
2-Naphthol was also observed to react under our reaction
conditions. To our surprise, this substrate seems to have
a greater preference than 1-naphthol to form the 5,5-fused
system over the 5,6-bridged system. However, low enantiose-
lectivities were observed for both aldehydes bearing methyl
(60% ee) and phenyl (44% ee) substituents.[14]
5,5 Reaction pathway: Scope of tetrahydrofurobenzofuran
scaffolds
The scope of the formation of tetrahydrofurobenzofurans 6
from aliphatic g-keto-enals, was also explored (Scheme 3). The
chiral products were obtained in moderate to good yields (40–
62%) and high enantioselectivities (84–96% ee).
The absolute configuration of the benzofused acetals were
unambiguously assigned by X-ray analysis of crystals of tetra-
hydrofurobenzofuran 6s and methanobenzodioxepine 7k
(Figure 2).[14]
1-Naphthol 2a reacted smoothly with g-keto-enals 1a,c,e–g.
The tetrahydrofurobenzofuran 6a was isolated in 40% yield
and 94% ee. By applying the bulkier aldehyde 1c, longer reac-
tion times were required, so 2 mol% of catalyst 3a was ap-
Figure 2. X-ray structures of 6s and 7k.
Transformations
In the previous sections, the hemiacetal moiety of intermediate
4 was reduced in all cases. To further explore the application
of this functionality, g-keto-enal 1g and 4-chloro-1-naphthol
2b were subjected to the optimized organocatalytic conditions
and then further functionalized in
a one-pot fashion
(Scheme 4).[16] With BF3·OEt2 as Lewis acid and allylsilane as the
nucleophile, the highly substituted tetrahydrofurobenzofuran 8
was obtained as a single stereoisomer in 56% yield and
90% ee. Furthermore, applying pyridinium chlorochromate
(PCC) as oxidant led to the formation of lactone 9, which was
isolated in 37% yield and 90% ee.
Scheme 3. Scope for the enantioselective formation of tetrahydrofurobenzo-
furan. Reactions were performed on a 0.25 mmol scale, with 1.5 equivalents
of g-keto-enal. Full conversion (>95%) and ratios were determined by
1H NMR spectroscopy of the crude reaction mixture. [a] Reaction performed
with 2 equivalents of g-keto-enal and 2 mol% of catalyst 3a.
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
5
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Scheme 4. Functionalizations on the tetrahydrofurobenzofuran scaffold. Re-
actions were performed on a 0.25 mmol scale, with 1.5 equivalents of g-
keto-enal 1g. Full conversion (>95%) and ratios were determined by
1H NMR spectroscopy of the crude reaction mixture. [a] Compound 3a
(0.5 mol%), oNO2C6H4CO2H (10 mol%), H2O (5 equiv.), CHCl3 (0.5m); [b] MS
3 ꢂ (300 mg), À78 to À208C, BF3·OEt2 (2.0 equiv.), allylsilane (2.0 equiv.);
[c] MS 3 ꢂ (300 mg), PCC (2.5 equiv.).
Mechanistic insight
In Scheme 5, we propose a catalytic cycle that takes into ac-
count the observed experimental results. Assuming that water
assists or accelerates the hydrolysis step, the lack of reactivity
in the absence of added water suggests that the catalyst
might be trapped during the course of the reaction. To sup-
port this hypothesis, stoichiometric amounts of catalyst 3a
were mixed with 4-chloro-1-naphthol 2b and g-keto-enals 1 f
and 1h, respectively, in CDCl3 in the presence of molecular
sieves. Mass spectrometry analysis of the crude reaction mix-
ture identified the mass peaks corresponding to 10 and 11
(Scheme 5). 13C NMR spectroscopy of the crude reaction mix-
tures revealed the presence of a carbonyl group in 10, whereas
this was not the case for 11.[14] These observations support the
proposed pathway for the formation of products 7k and 6q,
indicating that selectivity of the reaction pathways might pre-
cede the release of the catalyst.[17] Having outlined a mechanis-
tic proposal, our attention turned towards the investigation of
the low catalyst loading required for the reaction.
With some remarkable exceptions,[18] aminocatalytic transfor-
mations often rely on relatively high catalyst loadings (10–
20 mol%). The observed results for the present synthesis of
benzofused acetals imply turnover numbers up to 260
(Scheme 2, 7p), which is among the highest turnover numbers
found for iminium ion-catalyzed transformations.[19] With
a view to accounting for the low catalyst loadings (down to
0.25 mol%), we carried out some computational studies of the
LUMO of the systems studied (Figure 3).
Scheme 5. Proposed catalytic cycle.
intermediate. It appears from Figure 3 that the LUMO coeffi-
cient for the b-carbon atom of the iminium ion of the g-keto-
enal is not significantly different compared to the iminium ion
of pentenal. Based on frontier molecular orbital considerations,
we assume that the major contribution(s) to the increased re-
activity of the iminium ion system derived from the g-keto-
enal might be due to the lower LUMO energy for this system
compared to the corresponding enal (DLUMO =0.84 eV) and/or
an acceleration of the condensation step of the g-keto-enal
with the aminocatalyst compared to that of the a,b-unsaturat-
ed aldehyde (DLUMO =1.36 eV).
A possible explanation for the selectivity of the reaction (5,5-
vs. 5,6-pathway) could arise from the relative stability of the
oxocarbenium ion intermediates leading to the different prod-
ucts. To investigate this, we calculated the relative energies, by
a composite CCSD(T) method described below, of the oxocar-
benium ion intermediates (12–14) involved in the reduction
steps (Table 3). The results show that the oxocarbenium ion 14
has a higher stability than the other intermediates 12 and 13
for both R1 =C2H5 and R1 =Ph. The effect is most prominent
for R1 =Ph, owing to the higher stabilization of a positive
charge in the benzylic position on intermediate 14.
Although the b-carbon atom of each carbonyl group in the
g-keto-enal has equal partial positive charge, the LUMO coeffi-
cients show a polarization upon formation of its iminium ion
The calculations showed that the oxocarbenium ion 14 is
the most stable intermediate and thus if oxocarbenium ion for-
mation was the rate-determining step, benzofused acetal 7
should be the only observed product for both aliphatic and ar-
Figure 3. LUMO energies and atomic contributions (wB97XD/pcseg-1).
&
&
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
ble continuum model).[20] Addition of the zero-point energy
differences from wB97XD/pcseg-1 frequency calculations and
Boltzmann averaging over the two diastereoisomers provided
the DG(theo) values. The DG values relative to those R1 =CH3
(Table 4, entry 1) are also given in Table 4 (columns 3 and 4).
Table 3. Relative energies of the oxocarbenium ion intermediates 12–
14.[a]
Table 4. Theoretical and experimental relative free energies of hemiace-
tals 4 and 5 [DG=G(5)ÀG(4)].[a]
Column
Entry
1
2
3
4
R1
DG(exp) DG(theo)
DG(exp, rel.) DG(theo, rel.)
1
2
3
4
5
6
7
8
CH3 (1b)
CH2CH3 (1a)
CH(CH3)2 (1c)
C(CH3)3 (1d)
CH2Ph (1e)
2.1
5.4
8.7
9.7
5.4
10.5
11.0
14.4
18.3
16.3
3.5
0
3.3
6.6
7.6
0
0.5
3.9
7.8
5.8
À7.0
0.3
1.6
3.3
Ph (1h) À7.4
À9.5
À5.8
À1.6
p-CNC6H4 (1i)
p-NO2C6H4 (1j)
À3.7
10.8
12.1
0.5
[a] Ratios were estimated based on the product distributions given in
Table 2. For product 6, the ratio also includes the amount of the open
acetal form.[13]
The experimental energy differences DG(exp) (Table 4) be-
tween isomeric hemiacetals
4 and 5 are small (0.5–
9.7 kJmolÀ1). Figure 4 shows that the variation with substitu-
ents is reproduced with errors of a few kJmolÀ1 (Table 4, col-
umns 3 and 4), but the absolute values are overestimated by
approximately 10 kJmolÀ1 (columns 1 and 2). This systematic
error could be due to using only the cc-pVDZ basis set for esti-
mating electron correlation effects beyond MP2 and/or defi-
ciencies in the IEFPCM solvent model that treats the interac-
tion of alkyl and aryl groups with the solvent as alike.
Entry
R1
12
13[b]
14
1
2
CH2CH3
Ph
12.3
48.0
57.7
69.2
0.0
0.0
[a] Relative energies in kJmolÀ1 obtained by
method; [b] upon rotation of the carbonyl group, this intermediate ring-
a composite CCSD(T)
closes to 14 without an energy barrier.
omatic g-keto-enals. However, as we observe both benzofused
acetal scaffolds 6 and 7, these results indicate that the equilib-
rium between the hemiacetal intermediates 4 and 5 is much
slower than the reduction. Furthermore, by dissolving a crystal
of 4a in [D6]benzene, we found that >40 h were necessary to
re-establish the original equilibrium ratio between 4 and 5,
while the reduction takes 2 h at À208C.[14]
In an attempt to shed further light on the effects governing
the equilibrium between the hemiacetal intermediates 4 and
5, calculations of the ground-state energies of these intermedi-
ates were performed and compared with the experimentally
observed product distribution of 6 and 7 (Table 4). The experi-
mental free energies DG(exp) (Table 4, column 1) are estimated
based on the product distribution given in Table 2. The theo-
retical free energies DG(theo) (Table 4, column 2) are calculated
by a composite CCSD(T) method consisting of wB97XD/pcseg-
1 geometry optimization of all conformations generated by
the MMFF force field. Improved relative energies of the lowest
conformation for each species were obtained by extrapolating
MP2/cc-pVTZ and cc-pVQZ energies to the basis set limit and
addition of the CCSD(T)-MP2 energy difference with the cc-
pVDZ basis set. All calculations employed the CHCl3 IEFPCM
solvent model (IEFPCM=integral equation formalism polariza-
Figure 4. Correlation between experimental and computational differences
in free energies (DG) between 4 and 5.
To account for the different reaction pathways leading to
the tetrahydrofurobenzofuran 6 and methanobenzodioxepine
7 scaffolds, a simplified explanation is suggested in Scheme 6.
An initial consideration is that any effect that destabilizes 5 will
shift the equilibrium towards hemiacetal 4 and vice versa. We
propose a simplified model to try to account for the difference
in reaction pathways. The experimentally observed conforma-
tional restrictions in propanal (Scheme 6a) are applied to try to
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
7
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
account for the influence of different R-alkyl substituents on
the relative stability of hemiacetal 5 (Scheme 6b). For propa-
nal, the eclipsed conformations eclipsed-I and eclipsed-II are
more stable than the bisected conformations bisected-I and
bisected-II. Counterintuitively, the conformation eclipsed-I was
found to be 3.3–3.7 kJmolÀ1 more stable than eclipsed-II.[21]
These weak conformational preferences for eclipsed-I arise
from stereoelectronic interactions between the sCÀH and the
p*C orbitals of the carbonyl group (Scheme 6c).[22] The same
action path. We suggest that 5e behaves in a similar manner
to that proposed for the alkyl-substituted substrates discussed
above. However, the present simplified model seems to under-
estimate the selectivity for 5 f and overestimate that for 5g.
The lower predictability for 5 f and 5g might be due to the
higher degrees of freedom of the longer alkyl side-chain,
which are not accounted for by the simplified model.
For the aromatic side-chains, conjugation with the carbonyl
group strongly stabilizes hemiacetal 5 compared to 4, leading
to a change in the selectivity towards the 5,6-reaction pathway,
providing 7 as the major product (Scheme 6d). However, the
presence of an electron-withdrawing group in the aromatic
ring seem to partially counter this effect as the increased elec-
trophilicity of the keto-group shifts the equilibrium slightly to-
wards hemiacetal 4.[23] This effect is weakly observed for p-CN
(sp =0.66),[24] but for p-NO2 (sp =0.78) the reaction becomes
unselective as a result of the competing effects (Table 2, en-
tries 9 and 10).
=
O
eclipsed conformers are also calculated as the lowest energy
conformation of hemiacetal 5 bearing different R-alkyl sub-
stituents (Scheme 6b).[14]
The combined experimental and computational studies
seem to account for the organocatalytic divergent 5,5- and
5,6-pathways observed for the reactions of g-keto-enals with
hydroxyarenes. These studies indicate that the distribution of
the products 6 and 7 is determined prior to the reduction step
by the equilibrium between 4 and 5. This is rationalized by the
reduction step being much faster than the equilibrium process.
For the g-keto-enal bearing alkyl R1 substituents in the hemia-
cetal intermediate, the observed 5,5-pathway can be account-
ed for by a steric repulsion between the alkyl substituents and
the carbonyl moiety in 5, shifting the equilibrium towards 4.
The preference for the 5,6-pathway for the g-keto-enal bearing
aryl R1 substituents is based on a preferred conjugation be-
tween the aromatic substituent and the p*C orbital in 5,
=
O
forming 7 as the product. The calculations provided useful in-
sights into the relatively low required catalyst loadings and the
origins of the observed product distributions.
Conclusions
Scheme 6. Simplified models for the hemiacetal equilibrium.
We have developed the first organocatalytic enantioselective
syntheses of tetrahydrofurobenzofuran and methanobenzo-
dioxepine scaffolds. The development is based on a pair of di-
vergent pathways from hydroxyarenes and g-keto-enals. One
reaction path leads to chiral 5,5-fused tetrahydrofurobenzofur-
an scaffolds bearing two stereocenters, whereas the other
pathway provides 5,6-bridged methanobenzodioxepine scaf-
folds containing three stereocenters. The tetrahydrofurobenzo-
furans are formed in moderate yields and up to 96% ee,
whereas the methanobenzodioxepines are afforded in moder-
ate to good yields and up to 95% ee. The reaction proceed in
the presence of catalyst loadings as low as 0.25 mol%, provid-
ing one of the highest turnover numbers found in iminium ion
catalysis. We applied various reactions to effect further func-
tionalizations of the hemiacetal tetrahydrofurobenzofuran,
such as reduction, oxidation and allylation. Furthermore, we
studied the effects involved in the substrate control of the re-
action paths, relying on both experimental and computational
investigations. A simplified model based on steric, electronic,
This implies that, when moving along the alkyl series—
methyl to tert-butyl—in 5, an increase in steric repulsion be-
tween the alkyl side-chain and the carbonyl group is observed.
Although the repulsion between the carbonyl and each addi-
tional methyl group might be approximately the same, the
number of repulsive eclipsed conformers increases along the
series. Thus, for a methyl side-chain, no such repulsion takes
place, ethyl gives rise to one, isopropyl to two, and tert-butyl
to three such repulsions, thereby increasing the overall steric
interaction with the carbonyl group. This should increase the
relative energy of the hemiacetal 5 when the R substituent in-
creases in size, leading to a shift in equilibrium towards 4. Con-
sequently, an increased tendency towards the 5,5-reaction
pathway leading to 6 should be observed, in accordance with
the experimental results.
For the benzyl substituted g-keto-enal 1e (Table 2, entry 5),
the 5,5-reaction pathway is also observed as the preferred re-
&
&
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
8
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
123, 4370–4371; b) J. F. Austin, D. W. C. MacMillan, J. Am. Chem. Soc.
2002, 124, 1172–1173; c) N. A. Paras, D. W. C. MacMillan, J. Am. Chem.
Soc. 2002, 124, 7894–7895; for representative examples of iminium
ion-catalyzed Friedel–Crafts with naphthols, see: d) Ł. Albrecht, L. K.
Ransborg, V. Lauridsen, M. Overgaard, T. Zweifel, K. A. Jørgensen,
Angew. Chem. Int. Ed. 2011, 50, 12496–12500; Angew. Chem. 2011, 123,
12704–12708; e) L. Hong, L. Wang, W. Sun, K. Wong, R. Wang, J. Org.
Chem. 2009, 74, 6881–6884; f) E. Paradisi, P. Righi, A. Mazzanti, S. Ra-
nieri, G. Bencivenni, Chem. Commun. 2012, 48, 11178–11180; P. H. Poul-
sen, K. S. Feu, B. M. Paz, F. Jensen, K. A. Jørgensen, Angew. Chem. Int. Ed.
2015, 54, 8203–8207; Angew. Chem. 2015, 127, 8321–8325.
and stereoelectronic interactions was proposed to rationalize
the observed selectivities.
Acknowledgements
Thanks are expressed to Aarhus University and the Carlsberg
Foundation for financial support, the CAPES Foundation and
the Ministry of Education of Brazil for a predoctoral fellowship
to B.M.P. (no. 9525-13-0), and the German Research Foundation
(DFG) for a postdoctoral fellowship to L.K. (KL2984/1-1). Thanks
are expressed to Rasmus Mose for producing the artwork.
[13] For seminal works using diarylprolinol silyl ethers as organocatalysts,
see: a) M. Marigo, T. C. Wabnitz, D. Fielenbach, K. A. Jørgensen, Angew.
Chem. Int. Ed. 2005, 44, 794–797; Angew. Chem. 2005, 117, 804–807;
b) Y. Hayashi, H. Gotoh, T. Hayashi, M. Shoji, Angew. Chem. Int. Ed. 2005,
44, 4212–4215; Angew. Chem. 2005, 117, 4284–4287; for a recent
review, see e.g: c) B. S. Donslund, T. K. Johansen, P. H. Poulsen, K. A. Jør-
gensen, Angew. Chem. Int. Ed. 2015, 54, 13860–13874; Angew. Chem.
2015, 127, 14066–14081.
Keywords: asymmetric catalysis · computational chemistry ·
fused-ring systems · organocatalysis · reaction mechanisms
[14] For full details, see the Supporting Information.
[1] For representative examples, see: a) F. Perron, K. F. Albizati, Chem. Rev.
1989, 89, 1617–1661; b) J. E. Aho, P. M. Pihko, T. K. Rissa, Chem. Rev.
2005, 105, 4406–4440; c) J. A. Palmes, A. Aponick, Synthesis 2012,
3699–3721.
[2] For a review, see: a) L. Cala, F. J. FaÇanꢃs, F. Rodrꢄguez, Org. Biomol.
Chem. 2014, 12, 5324–5330; for representative examples of enantiose-
[15] J. Wang, M. Sꢃnchez-Rosellꢆ, J. L. AceÇa, C. del Pozo, A. E. Sorochinsky,
S. Fustero, V. A. Soloshonok, Chem. Rev. 2014, 114, 2432–2506.
[16] For example, see: a) Y. Hayashi, Chem. Sci. 2016, 7, 866–880; b) L. Al-
brecht, H. Jiang, K. A. Jørgensen, Angew. Chem. Int. Ed. 2011, 50, 8492–
8509; Angew. Chem. 2011, 123, 8642–8660.
[17] For reviews about mechanistic studies in aminocatalysis, see: a) M. Niel-
sen, D. Worgul, T. Zweifel, B. Gschwend, S. Bertelsen, K. A. Jørgensen,
Chem. Commun. 2011, 47, 632–649; b) C. Moberg, Angew. Chem. Int. Ed.
2013, 52, 2160–2162; Angew. Chem. 2013, 125, 2214–2216; c) J. Burꢅs,
A. Armstrong, D. G. Blackmond, Acc. Chem. Res. 2016, 49, 214–222.
[18] For a review on low-loading organocatalysis, see: a) F. Giacalone, M.
Gruttadauria, P. Agrigento, R. Noto, Chem. Soc. Rev. 2012, 41, 2406–
2447; for representative examples using enamine catalysis, see: b) M.
Wiesner, G. Upert, G. Angelici, H. Wennemers, J. Am. Chem. Soc. 2010,
132, 6–7; c) R. Martꢄn-Rapffln, X. Fan, S. Sayalero, M. Bahramnejad, F.
Cuevas, M. A. Pericàs, Chem. Eur. J. 2011, 17, 8780–8783; d) B. Wang, X.-
W. Liu, L.-Y. Liu, W.-X. Chang, J. Li, Eur. J. Org. Chem. 2010, 5951–5954;
e) X. Fan, E. Alza, M. A. Pericàs, RSC Adv. 2012, 2, 6164–6166; f) S. Chan-
drasekhar, C. P. Kumar, T. P. Kumar, K. Haribabu, B. Jagadeesh, J. Lakshmi,
P. S. Mainkar, RSC Adv. 2014, 4, 30325–30331.
[19] For examples with low-loading iminium ion catalysis, see: a) S. Shiomi,
E. Sugahara, H. Ishikawa, Chem. Eur. J. 2015, 21, 14758–14763; b) M.-Y.
Han, H.-Z. Wang, W.-K. An, J.-Y. Jia, B.-C. Ma, Y. Zhang, W. Wang, Chem.
Eur. J. 2013, 19, 8078–8081; c) L. Bernardi, M. Fochi, R. Carbone, A. Mar-
tinelli, M. E. Fox, C. J. Cobley, B. Kandagatla, S. Oruganti, V. H. Dahanu-
kar, A. Carlone, Chem. Eur. J. 2015, 21, 19208–19222.
[20] a) J.-D. Chai, M. Head-Gordon, Phys. Chem. Chem. Phys. 2008, 10, 6615–
6620; b) F. Jensen, J. Chem. Theory Comput. 2014, 10, 1074–1085; c) J.
Tomasi, B. Mennuci, E. Canc›s, J. Mol. Struct. THEOCHEM 1999, 464,
211–226; d) T. H. Dunning, Jr., J. Chem. Phys 1989, 90, 1007–1023.
[21] a) S. S. Butcher, E. B. Wilson, J. Chem. Phys. 1964, 40, 1671–1678; b) G. J.
Karabatsos, H. Hsi, J. Am. Chem. Soc. 1965, 87, 2864–2870.
ˇ
´
lective organocatalytic synthesis of acetals, see: b) I. Coric, B. List,
Nature 2012, 483, 315–319; c) M. F. Polat, L. Hettmancyk, W. Zhang, Z.
Szabo, J. Franzꢅn, ChemCatChem 2013, 5, 1334–1339; d) H. Huang, S.
Konda, J. C.-G. Zhao, Angew. Chem. Int. Ed. 2016, 55, 2213–2216;
Angew. Chem. 2016, 128, 2253–2256.
[3] G. J. Gnonlonfin, K. Hell, Y. Adjovi, P. Fandohan, D. O. Koudande, G. A.
Mensah, A. Sanni, L. Brimer, Crit. Rev. Food Sci. Nutr. 2013, 53, 349–365.
[4] a) F. M. Talontsi, B. Dittrich, A. Schuffler, H. Sun, H. Laatsch, Eur. J. Org.
Chem. 2013, 3174–3180; b) M. El Amrani, D. Lai, A. Debbab, A. H. Aly, K.
Siems, C. Seidel, M. Schnekenburger, A. Gaigneaux, M. Diederich, D.
Feger, W. Lin, P. Proksch, J. Nat. Prod. 2014, 77, 49–56.
[5] a) G. Bꢁchi, D. M. Foulkes, M. Kurono, G. F. Mitchell, J. Am. Chem. Soc.
1966, 88, 4534–4536; b) G. Bꢁchi, D. M. Foulkes, M. Kurono, G. F. Mitch-
ell, R. S. Schneider, J. Am. Chem. Soc. 1967, 89, 6745–6753; c) J. C. Rob-
erts, A. H. Sheppard, J. A. Knight, P. Roffey, J. Chem. Soc. C 1968, 22–24.
[6] a) A. J. Castellino, H. Rapport, J. Org. Chem. 1986, 51, 1006–1011; b) G.
Weeratunga, S. Horne, R. Rodrigo, J. Chem. Soc. Chem. Commun. 1988,
721–722; c) S. Horne, G. Weeratunga, R. Rodrigo, J. Chem. Soc. Chem.
Commun. 1990, 39–41; d) M. Koreeda, L. A. Dixon, J. D. His, Synlett
1993, 555–556; e) M. C. Pirrung, Y. R. Lee, Tetrahedron Lett. 1996, 37,
2391–2394.
[7] G. Zhou, E. J. Corey, J. Am. Chem. Soc. 2005, 127, 11958–11959.
[8] a) B. M. Trost, F. D. Toste, J. Am. Chem. Soc. 1999, 121, 3543–3544;
b) B. M. Trost, F. D. Toste, J. Am. Chem. Soc. 2003, 125, 3090–3100.
[9] During the preparation of this manuscript, an method capable of over-
coming the substitution pattern limitation by palladium-catalyzed car-
boetherification was reported: G. M. Borrajo-Calleja, V. Bizer, C. Mazet, J.
Am. Chem. Soc. 2016, 138, 4014–4017.
[10] W. Luo, Q.-S. Yu, S.-S. Kulkarni, D. A. Parrish, H. W. Holloway, D. Tweedie,
A. Shafferman, D. K. Lahiri, A. Brossi, N. H. Greig, J. Med. Chem. 2006, 49,
2174–2185.
[22] O. Takahashi, K. Yasunaga, Y. Gondoh, Y. Kohno, K. Saito, M. Nishio, Bull.
Chem. Soc. Jpn. 2002, 75, 1777–1783.
[23] H.-J. Buschmann, H. H. Fꢁldner, W. Knoche, Ber. Bunsen-Ges. 1980, 84,
41–44.
[24] C. Hansch, A. Leo, R. W. Taft, Chem. Rev. 1991, 91, 165–195.
[11] M. A. Kamal, X. Qu, Q.-S. Yu, D. Tweedie, H. W. Holloway, Y. Li, Y. Tan,
N. H. Greig, J. Neural Transm. 2008, 115, 889–898.
[12] For representative examples of iminium ion-catalyzed Friedel–Crafts re-
actions, see: a) N. A. Paras, D. W. C. MacMillan, J. Am. Chem. Soc. 2001,
Received: June 23, 2016
Published online on && &&, 0000
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
9
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
FULL PAPER
&
Organocatalysis
B. M. Paz, L. Klier, L. Næsborg,
V. H. Lauridsen, F. Jensen, K. A. Jørgensen*
&& – &&
Different class: A novel enantioselec-
tive organocatalytic strategy is present-
ed for the synthesis of tetrahydrofuro-
benzofuran structures, based on a pair
of divergent reaction pathways in which
hydroxyarenes react with g-keto-a,b-un-
saturated aldehydes, catalyzed by
a chiral secondary amine. The reaction
is remarkable as it proceeds with cata-
lyst loadings as low as 0.25 mol%.
Enantioselective Organocatalytic
Cascade Approach to Different Classes
of Benzofused Acetals
Cascade Synthesis
In their Full Paper on page && ff., K. A. Jørgensen et al.
present a novel enantioselective strategy for the synthesis
of tetrahydrofurobenzofuran and methanobenzodioxepine
structures, based on a pair of divergent reaction pathways
in which hydroxyarenes react with g-keto-a,b-unsaturated
aldehydes, catalyzed by a chiral secondary amine. One
reaction pathway leads to chiral 5,5-fused acetals, the
tetrahydrofurobenzofuran scaffolds, with two stereocenters,
whereas the other reaction pathway provides 5,6-bridged
methanobenzodioxepine scaffolds with three stereocenters.
The reaction can proceed with catalyst loadings as low as
0.25 mol%, giving among the highest known turnover
numbers in iminium ion catalysis.
&
&
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
10
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!