Synthesis and structural analysis of sterically hindered chiral 1,4-diol ligands derived from the lignan hydroxymatairesinol

Article abstract of DOI:10.1016/j.tetlet.2012.12.066

The readily available natural lignan hydroxymatairesinol was transformed into sterically hindered and optically pure diphenyl, di-2-naphthyl, and tetramethyl 1,4-diol derivatives via arylation/alkylation of the aryltetralinbutyrolactone lignan (-)-conidendrin. In addition, the diastereoselective formation of stable hemiketals from the highly substituted butyrolactone was studied in detail. The conformations of the molecules prepared were studied computationally at molecular mechanics (MM), Hartree-Fock (HF)/6-31G*, and (DFT/B3LYP/TZVP) levels including entropy contributions and by NMR-spectroscopy. The conformations adopted showed that these novel chiral 1,4-diols may be suitable as chiral ligands for the development of new chiral transition metal and organo catalysts.

Full text of DOI:10.1016/j.tetlet.2012.12.066

Tetrahedron Letters 54 (2013) 1112–1115
Contents lists available at SciVerse ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Synthesis and structural analysis of sterically hindered chiral 1,4-diol
ligands derived from the lignan hydroxymatairesinol
Yury Brusentsev a, Thomas Sandberg b, Matti Hotokka b, Rainer Sjöholm a, Patrik Eklund a,
⇑
a Laboratory of Organic Chemistry, Åbo Akademi University, Biskopsgatan 8, 20500 Åbo (Turku), Finland
b Laboratory of Physical Chemistry, Åbo Akademi University, Porthansgatan 3-5, 20500 Åbo (Turku), Finland
a r t i c l e i n f o
a b s t r a c t
Article history:
The readily available natural lignan hydroxymatairesinol was transformed into sterically hindered and
optically pure diphenyl, di-2-naphthyl, and tetramethyl 1,4-diol derivatives via arylation/alkylation of
the aryltetralinbutyrolactone lignan (ꢀ)-conidendrin. In addition, the diastereoselective formation of sta-
ble hemiketals from the highly substituted butyrolactone was studied in detail. The conformations of the
molecules prepared were studied computationally at molecular mechanics (MM), Hartree–Fock (HF)/6-
31G⁄, and (DFT/B3LYP/TZVP) levels including entropy contributions and by NMR-spectroscopy. The con-
formations adopted showed that these novel chiral 1,4-diols may be suitable as chiral ligands for the
development of new chiral transition metal and organo catalysts.
Received 5 October 2012
Revised 3 December 2012
Accepted 13 December 2012
Available online 25 December 2012
Keywords:
Lignans
Hydroxymatairesinol
Chiral 1,4-diol
Ó 2013 Elsevier Ltd. All rights reserved.
Asymmetric catalysis
Ab initio calculations
Chiral 1,4-diols have been shown to be useful starting materials
for the preparation of chiral catalysts, chiral reagents, resolving
agents, etc.1 The numerous reports of TADDOLs or BINOLs in the
literature show the versatility of this type of chiral diol. Tradition-
ally, TADDOLs or BINOLs have been used as ligands for transition
metal based asymmetric catalysts, but more recently and in partic-
ular phosphoric acid derivatives have shown excellent results as
chiral Brønsted acids in the field of organocatalysis.2 We herein
report the synthesis and conformational study of chiral 1,4-diols
derived from the lignan (ꢀ)-conidendrin, obtained from 7-
hydroxymatairesinol.
(ꢀ)-Conidendrin is a natural product of the lignan family. In the
early reports of Erdtman et al.3 it was referred to as ‘sulfitlaugen
lacton’. Later it was shown that (ꢀ)-conidendrin could be easily ob-
tained in quantitative yields from the lignan hydroxymatairesinol.4
Large-scale methods for the isolation of hydroxymatairesinol from
spruce knotwood have been developed.5 Today, a mixture of
hydroxymatairesinol diastereomers can be produced on a kilogram
scale, which after treatment with acid and recrystallization gives
access to optically pure (ꢀ)-conidendrin in quantitative yield.4
The enantiomerically pure butyrolactone ring of (ꢀ)-coniden-
drin prompted us to perform alkylation (arylation), oxidation,
and a second alkylation (arylation) in order to obtain sterically hin-
dered 1,4-diol structures resembling those of TADDOLs (Fig. 1). We
wanted to evaluate both the synthetic pathways and the possibility
to utilize the backbone of conidendrin as a chiral ligand. The choice
of substituents for the target diols was based on simple TADDOL
derivatives (phenyl, naphthyl, and methyl). The target diols were
considered as suitable lead compounds for further investigation
of their structural properties by NMR spectroscopy and by compu-
tational molecular modeling, as well as for evaluation of their
properties as chiral ligands in asymmetric catalysis (proof of
principle).
The compounds presented, are semisystematically named
according to the trivial names of the parental-lignan structure.
The structures are based on the aryltetralin 20,7-cyclolignan skele-
ton (Fig. 1), where the lactone structures are derived from dimeth-
ylconidendrin 1 and dimethylretrodendrin 6 (R = H), and the diol
structures are derived from dimethylcyclolariciresinol 5 (R = H)
(Scheme 2). The derivatization of the fundamental parental struc-
tures, that is, introduction of R-groups at carbons 9 and 90 is
R
H
6'
7'
R
MeO
MeO
5'
9'
8'
8
1'
2'
OH
OH
9
7
4'
3'
R'
H
1
R'
2
6
3
5
MeO
4
OMe
Figure 1. General structure and numbering of the sterically hindered 1,4-diol
structures, represented by the 9 and 90 substituted aryltetralin 20,7-cyclolignan
dimethylcyclolariciresinol.
⇑
Corresponding author. Tel.: +358 2 215 4720.
E-mail address: paeklund@abo.fi (P. Eklund).
0040-4039/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tetlet.2012.12.066

Y. Brusentsev et al. / Tetrahedron Letters 54 (2013) 1112–1115
1113
O
H
O
O
H
H
H
H
MeO
MeO
MeO
MeO
MeO
HO
OMe
OMe
O
O
1) Br2/NaOH
2) MeI/K2CO3
1) HCOOH
HO
H
O
2) MeI/K2CO3
OMe
OMe
OMe
(40%)
2
MeO
OMe
(90%)
OH
1
Hydroxymatairesinol
PhMgBr
H
H
H
H
MeO
MeO
MeO
MeO
OH
OH
O
OH
MeO
MeO
OMe
3a
OMe
4a
Scheme 1. The diester synthetic route applied on dimethylconidendrin (13).
Scheme 2. Stepwise synthesis of the diols.
indicated in the prefix. The carbons are numbered as depicted in
Figure 1.
worked better than the corresponding Grignard reagents. In the
reactions of 1 with Grignard reagent, the yields of 5a–c were usu-
ally around 65%, and with alkyl- or aryllithium yields of 85% or
higher were obtained.
In the literature, only a few methods for the synthesis of TAD-
DOL-like structures are described. Usually, the synthesis is based
upon the one-pot addition of four alkyl or aryl groups with an
appropriate Grignard reagent to the diester of a 1,4-dicarboxylic
acid. To evaluate this methodology on the conidendrin structure,
we prepared the diester 2 (see the Supplementary data).
However, the Grignard reaction of the diester with excess of
phenylmagnesium bromide gave only the undesired hemiketal
3a (Scheme 1). This problem has not been previously reported
for TADDOLs or similar structures, although the reported yield of
the target diol in most cases seldom exceeded 50%.6
The oxidation of the primary alcohols 5a–c with pyridinium
chlorochromate (PCC) proceeded smoothly as a stepwise process.
In reactions performed with 1 equiv of PCC the hemiacetal was
the major product and could be isolated in 60% yield. This hemiac-
etal was shown to be very stable and the corresponding aldehyde
could not be detected in aqueous solutions of the hemiacetal.
When an excess of PCC was used the oxidation gave excellent
yields (ꢁ90%) of the lactone 6a or 6b.
Phenylation of 90,90-diphenyl-dimethylretrodendrin (6a) with
either phenylmagnesium bromide or with phenyllithium gave only
the hemiketal 3a as the major product and the other diastereomer
as a minor product (less than 15%).
In other attempts to obtain the target structures, stepwise Grig-
nard reactions and oxidations (Scheme 2) were performed. In this
study the arylation and alkylation with organolithium reagents

1114
Y. Brusentsev et al. / Tetrahedron Letters 54 (2013) 1112–1115
The stereochemical configurations of the hemiketals were
Ph
Ph
Ph
H
H
H
H
Ph
H
Ph
Ph
determined by NMR and assigned as 8R,80R,7S,9R for 3a and
8R,80R,7S,9S for 3b. No signals from the corresponding ketone or
the tetrasubstituted diol were found in the NMR spectra. To ex-
plain this phenomena phenylation of 90,90-dimethyl-dimethylret-
PCC
PhMgBr
PhMgBr
OMe
OMe
OMe
OH
O
OH
H
Ph
9a
10 (61%)
11 (63%)
PCC
rodendrin
(6b)
and
methylation
of
90,90-diphenyl-
dimethylretrodendrin (6a) were performed. Both reactions led to
the formation of hemiketal and no 1,4-diol was formed. Only the
methylation of 90,90-dimethyl-dimethylretrodendrin (6b) gave a
mixture of the target 9,9,90,90-tetramethyl-dimethylcyclolariciresi-
nol (4b) and the trimethyl-hemiketal (3b). The mixture formed had
a constant ratio of 3b and 4b (1:1.3), and only one diastereomer of
the hemiketal was formed. According to these results, we assume
that in this reaction the 90,90-disubstituted lactone structure (6a,
6b) forms one diastereomer of the hemiketal, which is extremely
stable and there is no equilibrium between this hemiketal and
the ketone. The other hemiketal seemed to react further to give
the diol product 4b.
Ph
Ph
H
H
Ph
Ph
OMe
OMe
O
OH
Ph
H
H
13
Ph
12 (68%)
Ph
Scheme 4. Attempt to prepare a tetraphenyl substituted diol from the methyl
protected compound 9a.
To study the possibility to further convert the intermediate
hemiketals (3a, 3b) into the diols, the hemiketals were isolated
and treated under different nucleophilic and reductive conditions.
However, in most cases, only the starting material was isolated and
no desired reactivity was observed (see Supplementary data).
Computational analysis of the hemiketal 3 (see Supplementary
data) supports the suggestion that the thermodynamically favored
isomer should be 9S-isomer of 3b (the Gibbs energy of the stable
hemiketal isomer of 3b was lower by 24–28 kJ/mol compared to
the other diastereomer). Also information on the atomic charges
showed that the reactivity of 3b should be lower. However, a com-
plete explanation for the lack of equilibrium with the correspond-
ing ketone as well as the extremely poor reactivity would require
additional studies.
To overcome the problem with the extreme stability of the
hemiketals, we investigated a strategy in which no ring could be
formed during the oxidation, that is, protection of the tertiary
OH-group (PG2, Scheme 3). We therefore set out to selectively pro-
tect the primary alcohol (PG1) followed by the protection of the
tertiary alcohol and then deprotection of PG1 (Scheme 3). A num-
ber of different protecting groups and conditions were tested
(see Supplementary data). Different PG1 groups were successfully
introduced but only a methyl group could be introduced as PG2
(Scheme 3).
Figure 2. Energetically favored conformer of 9,9,90,90-tetramethyl-dimethyl-cyclol-
ariciresinol (4b). The distance is given in Ångströms.
diol 4b which corresponded to the calculated structures: in the
four minimum energy conformers the torsion angle H7–C7–C8–
H8 was close to 80°, which was in agreement with the small cou-
pling constant. The most energetically favored conformers showed
an H8–C8–C80–H80 torsion angle of around 115° and a coupling
Since
90O-methyl-90,90-diphenyl-dimethyl-cyclolariciresinol
(9a) was obtained, the free hydroxymethyl group was oxidized to
the aldehyde 10, phenylated, and oxidized again to give the ketone
12 (Scheme 4). However, it turned out that this ketone did not re-
act with phenylmagnesium bromide or with phenyllithium. This
suggested that the desired tetraphenyl structure 13 was too steri-
cally hindered to be prepared by the routes described.
The conformation of 4b was studied computationally at molec-
ular mechanics (MM), Hartree–Fock (HF)/6-31G⁄, and (DFT/B3LYP/
TZVP) levels including entropy contributions. The computational
results showed good correlation with the conformational studies
by NMR spectroscopy.7 For example, JH7–H8 was almost 0 Hz for
0
constant of JH8–H8 = 3.5 Hz in the NMR spectrum. The torsion an-
gles H80–C80–C70–H70 were 39° and 154° and correlated with
0
0
JH8 –H7 = 7 and 12 Hz. Both computational and NMR studies (see
Supplementary data) showed that the most stable conformers
adopted a half-boat conformation of the six-membered ring,8
which was the major contribution to the overall conformation. In this
conformation the hydroxy-containing arms are in a trans-equatorial
Ph
Ph
H
H
H
H
Ph
Ph
OH
OPG1
MeO
MeO
MeO
MeO
Ph
OH
Ph
H
H
Ph
H
Ph
OH
OPG2
OH
I
OPG2
II
III
OPG1
H
MeO
MeO
OMe
5a
OMe
7a-e
8a-e
9a-e
Scheme 3. Attempts to protect the more hindered OH group (O90) (Experimental conditions in Table 2 in Supplementary data).

Y. Brusentsev et al. / Tetrahedron Letters 54 (2013) 1112–1115
1115
R
H
H
R
MeO
OH
OH
cat.
OH
PhCHO
ZnEt2
cat. =
MeO
Ti(OiPr)4
R'
R'
MeO
OMe
Scheme 5. Diethylzinc addition to benzaldehyde (test reaction).
conformation which is crucial for the formation of metal com-
plexes and catalytic activity (Fig. 2).
Acknowledgements
To evaluate these novel 1,4-diol structures the well-known
TADDOL-catalyzed diethylzinc addition to benzaldehyde9 was em-
ployed as a preliminary model reaction.
However, diol 4b showed reverse selectivity in comparison to
the corresponding (ꢀ)-R,R-TADDOL (Scheme 5). The major isomer
in the reaction catalyzed by our diols was shown to be R in contrast
to the S-isomer formed in the (ꢀ)-TADDOL-catalyzed reaction.9
Despite the fact that the selectivity was only moderate, for
example: ꢁ20% ee with 4b as the catalyst, the observed phenom-
ena are of considerable interest.
The unsymmetrical structure of 4b is, however, not directly
comparable with TADDOLs and it can be assumed that the aro-
matic substituent at position seven has a major influence on the
catalyst structure and hence on the induction of enantioselectivity.
However, these results could be a good starting point for further
investigation of lignan-based chiral ligands.
The authors would like to thank the Academy of Finland for the
financial support, UPM Ltd for support with starting materials and
the Finnish IT Center for Science for providing computational re-
sources. Mr. Markku Reunanen is thanked for recording HRMS
spectra.
Supplementary data
Supplementary data (full experimental and spectral data) asso-
ciated with this article can be found, in the online version, at
http://dx.doi.org/10.1016/j.tetlet.2012.12.066.
References and notes
1. Seebach, D.; Beck, A. K.; Heckel, A. Angew. Chem., Int. Ed. 2001, 40, 92–138.
2. Terada, M. Synthesis 2010, 12, 1929–1982.
3. Erdtman, H. Liebigs Ann. 1934, 513, 229–239.
4. Freudenberg, K.; Knof, L. Chem. Ber. 1957, 90, 2857–2869.
5. Holmbom, B.; Eckerman, Ch.; Eklund, P.; Hemming, J.; Nisula, L.; Reunanen, M.;
Sjöholm, R.; Sundberg, A.; Willför, S. Phytochem. Rev. 2003, 2, 331–340.
In conclusion, the TADDOL-like chiral 1,4-diols (5a–c and 4b)
were synthesized from the natural lignan hydroxymatairesinol.
Furthermore, the addition reaction to the highly substituted lac-
tones 6a and 6b was studied in detail. This showed that substituted
butyrolactones form stable hemiketals, which has not been previ-
ously discussed in the literature. The steric hindrance and the for-
mation of a stable hemiketal structure prevented the synthesis of
the tetra-substituted 1,4-diol derivatives, unless a methyl was used
as the alkyl group. The structures and the conformation of the diols
were studied both by NMR spectroscopy and computational meth-
ods. The structural properties of these diols were mostly deter-
mined by the conformation of the six-membered non-aromatic
ring and the 3,4-dimethoxyphenyl substituent at position seven.
The 1,4-diol 4b was proven to work as a chiral ligand and there-
fore the study of these diols as well as the synthesis of other lig-
nan-based diols, phosphines, and phosphoric acids is ongoing in
our laboratory.
ˇ
6. Ito, Y. N.; Ariza, X.; Beck, A. K.; Bohác, A.; Ganter, C.; Gawley, R. E.; Kühnle, F. N.
M.; Tuleja, J.; Wang, Y. M.; Seebach, D. Helv. Chim. Acta 1994, 77, 2071–2109.
7. Analytical data for compound 4b: 1H NMR (600 MHz, CDCl3) d ppm 0.50 (s, 3H,
9Me), 0.65 (s, 3H, 9Me), 1.26 (s, 3H, 9Me), 1.58 (s, 3H, 9Me), 2.29 (ddd, J = 12.1,
6.8, 3.0 Hz, 1H, H8), 2.38 (dd, J = 14.0, 12.1 Hz, 1H, H7), 2.60 (dd, J = 14.0, 6.8 Hz,
1H, H7), 2.64 (d, J = 3.0 Hz, 1H, H8), 3.78 (s, 3H, 3OMe), 3.81 (s, 3H, 3OMe), 3.83
(s, 3H, 4OMe), 3.92 (s, 3H, 4OMe), 4.06 (s, 1H, H7), 6.40 (dd, J = 8.3, 1.9 Hz, 1H,
H6), 6.56 (d, J = 1.9 Hz, 1H, H2), 6.59 (s, 1H, H6), 6.71 (d, J = 8.3 Hz, 1H, H5), 6.75
(s, 1H, H3). 13C NMR (150 MHz, CDCl3) d ppm 23.86 (9Me), 25.7 (9Me), 30.72
(C7), 30.78 (9Me), 31.62 (9Me), 44.96 (C8), 46.87 (C7), 52.69 (C8), 55.86 (OMe),
55.91 (OMe), 55.94 (OMe), 55.94 (OMe), 72.83 (C9), 73.8 (C9), 110.94 (C5),
111.02 (C3), 111.62 (C2), 112.38 (C6), 120.07 (C6), 130.32 (C1), 131 (C2), 136.84
(C1), 147.15 (C4), 147.69 (C5), 147.78 (C4), 148.9 (C3). HRMS (ESI): calcd for
C
26H36NaO6 [M+Na]+ 467.2410; found 467.2430. ½a 2D0
+67° (c = 1, CHCl3).
ꢂ
Values for the H–H coupling constants for 4b were calculated from the 1H NMR
spectral data using the PERCH software.
8. Sandberg, T.; Brusentsev, Y.; Eklund, P.; Hotokka, M. Int. J. Quantum Chem. 2011,
111, 4309–4317.
9. Schmidt, B.; Seebach, D. Angew. Chem., Int. Ed. Engl. 1991, 30, 1321–1323.