Exploiting the vicinal disubstituent effect on the diastereoselective synthesis of γ and δ lactones

Article abstract of DOI:10.1039/c8ob02715c

Trifluoroacetic acid catalysed lactonization of vicinal disubstituted γ-hydroxyesters was investigated in different solvents. The reaction kinetics, monitored by NMR spectroscopy, showed that: (i) the vic-disubstituent effect is stereoselective since the anti diastereoisomer ring closes substantially more rapidly than the syn isomer ring; (ii) the anti-vic effect is much stronger than the classical Thorpe-Ingold effect (known also as the gem-disubstituent effect), instead the syn diastereoisomers have rate constants comparable to that of the gem-disubstituted ester; (iii) the vic-effect can be enhanced by increasing the steric hindrance of one of the two substituents or carrying out the reaction in non-polar solvents. DFT computations of energy barriers (ΔG^?) were in good agreement with the experimental data. The distortion/interaction-activation strain model together with the Winstein-Holness kinetic scheme gave more insights into the origin of the vic-effect. An application of this effect consists of the diastereomeric resolution of disubstituted γ and δ lactones, among which are the naturally occurring Nicotiana t. lactone, the whisky and cognac oak lactones, and the Aerangis lactone. Both cis and trans diastereoisomers of these lactones were isolated in good yield and with high diastereomeric excess (de >92%). The selectivities of the diastereomeric resolution process, determined by NMR spectroscopy, are reported as well.

Full text of DOI:10.1039/c8ob02715c

Organic &
Biomolecular Chemistry
View Article Online
View Journal
PAPER
Exploiting the vicinal disubstituent effect on the
diastereoselective synthesis of γ and δ lactones†
Cite this: DOI: 10.1039/c8ob02715c
Elisabetta Brenna,a Francesco Dalla Santa,a Francesco G. Gatti, *a Giuseppe Gattib
a
and Davide Tessaro
Trifluoroacetic acid catalysed lactonization of vicinal disubstituted γ-hydroxyesters was investigated in
different solvents. The reaction kinetics, monitored by NMR spectroscopy, showed that: (i) the vic-disub-
stituent effect is stereoselective since the anti diastereoisomer ring closes substantially more rapidly than
the syn isomer ring; (ii) the anti-vic effect is much stronger than the classical Thorpe–Ingold effect
(known also as the gem-disubstituent effect), instead the syn diastereoisomers have rate constants com-
parable to that of the gem-disubstituted ester; (iii) the vic-effect can be enhanced by increasing the steric
hindrance of one of the two substituents or carrying out the reaction in non-polar solvents. DFT compu-
tations of energy barriers (ΔG‡) were in good agreement with the experimental data. The distortion/inter-
action-activation strain model together with the Winstein–Holness kinetic scheme gave more insights
into the origin of the vic-effect. An application of this effect consists of the diastereomeric resolution of
disubstituted γ and δ lactones, among which are the naturally occurring Nicotiana t. lactone, the whisky
and cognac oak lactones, and the Aerangis lactone. Both cis and trans diastereoisomers of these lactones
were isolated in good yield and with high diastereomeric excess (de >92%). The selectivities of the diaster-
eomeric resolution process, determined by NMR spectroscopy, are reported as well.
Received 1st November 2018,
Accepted 16th November 2018
DOI: 10.1039/c8ob02715c
rsc.li/obc
phenyl propionic acid derivatives. The placement of a gem-
Introduction
dimethyl group accelerated the cyclization rate constant 400
The rate of intramolecular reactions affording small to times3 with respect to the unsubstituted precursor and further
medium rings is often accelerated by introducing a geminal substitution (trimethyl lock effect) speeded-up the ring closure
disubstituent group (C–CR2–C) into
(Fig. 1a). These different kinetic behaviours, often referred to
a
linear precursor by 5 orders of magnitude.3b,c
Incidentally, although the gem-disubstituent effect is con-
as gem-disubstituent effects, were first studied by Thorpe and sistently adopted in many fields of organic chemistry includ-
Ingold.1,2 They concluded that, by increasing the R–C–R angle
compression β arising from the mutual steric repulsion of the
geminal substituents of the precursor, the terminal reactive
groups should be brought into a more favourable orientation
(decreasing of the α angle) for ring formation, and consequen-
tially to an increase of both reaction rate and equilibrium con-
stant (Fig. 1a).
An impressive example of this effect was found in the com-
parative rates of lactone formation from various 2-hydroxy
aDipartimento di Chimica, Materiali ed Ingegneria Chimica “G. Natta”,
Politecnico di Milano, P.zza Leonardo da Vinci 32, 20133 Milano, Italy.
E-mail: Francesco.gatti@polimi.it
bDipartimento di Scienze Biomolecolari Scuola di Farmacia,
Università degli Studi di Urbino, P.zza del Rinascimento 6, 61029 Urbino, Italy
†Electronic supplementary information (ESI) available: Experimental pro-
cedures, compound characterization, copies of 1H-NMR and 13C-NMR spectra of
all compounds and Cartesian coordinates of all computed structures. See DOI:
10.1039/c8ob02715c
Fig. 1 (a) Explanation by internal angle reduction of the Thorpe–Ingold
effect; (b) cyclization rates of succinate derivatives.
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
ing metal-catalysed cyclizations,4 asymmetric catalyses,5
peptide chemistry6 and polymerizations,7 its full comprehen-
sion is still largely debated, as shown by the numerous expla-
nations that have been put forward in the last fifty years.2
In contrast, and quite surprisingly, the effect produced by
two substituents in a vicinal stereochemical relationship
(the vic-disubstituent effect) has only been marginally
considered.2,8,9
In the seventies Eberson et al. investigated the base cata-
lysed cyclization of succinic acid derivatives: the reaction rate
of the vic-dimethyl substituted succinates was found to be
slightly greater than that of the gem-dimethyl isomers, but, it
was not possible to determine which of the two diastereo-
isomers underwent faster ring closure, due to the syn/anti equi-
librium (Fig. 1b).9
Table 1 Kinetics of TFA catalysed ring closure of γ-hydroxyesters 1a–d
in CDCl3 at 303 K, and the relative experimental and computed energy
barriers
δΔG‡r;Exp
δΔG‡r;Comp
Hydroxyester
kr
(kcal mol−1
)
(kcal mol−1
)
1.0a
0.00
0.00
4.4a
−0.88
−1.08
5.3
−1.00
−0.93
However, concerning the vic-effect, some issues ought to be
clarified. More precisely, is the ring closure rate acceleration
(i) larger or smaller than that produced by a gem-substitution?
(ii) Is it depending on the relative stereochemistry of the two
vicinal substituents (syn or anti) of the linear precursor?
(iii) How influential is the steric hindrance of the substituents?
(iv) Has the solvent any effect on points (i) and (ii)?
Prompted by the above questions we decided to study
the trifluoro acetic acid (TFA) catalysed lactonization of
γ-hydroxyesters 1a–d in different solvents (Chart 1), either by
kinetic measurements or by DFT computations.
99.4 (18.9)b
5.3
−2.77 (−1.77)c
−1.00
−2.88 (−1.95)c
d
—
415.0 (78.4)b
−3.63 (−2.63)c
—
e
a Data ref. 10. b Selectivity, s = kanti/ksyn c δΔG‡anti–syn = ΔG‡anti − ΔGs‡yn
. .
d Not calculated. e Computations have been carried out only for the
(1Re,3S,4R,g+,E) isomer of anti-1d.
(δΔG‡r;Exp = −0.607 ln kr) relative to the ring closure of the
unsubstituted ester, i.e. 1a.
Results and discussion
Kinetic study
The inspection of Table 1 shows that the rate constants kr
of the vic-disubstituted hydroxyesters 1c–d are strongly influ-
enced by the relative stereochemistry, being that the anti dia-
stereoisomers cyclize always much more rapidly than the syn
isomers. Evidently, the effect has to be related to the steric hin-
drance of the substituent, since by replacing the methyl at C(4)
The ring closure reaction rates of hydroxyesters 1a–b (in CDCl3
at 303 K) have been recently determined by 1H-NMR spec-
troscopy.10 In the present work the kinetics of esters 1c–d were
analogously measured (see the ESI† for the preparation of
esters). The data were analysed on the basis of a pseudo first-
order reaction in the hydroxyester concentration [E], according
to the following equation:
carbon with an i-Pr group, the selectivity (s = kobs,anti/kobs syn)
,
dramatically increases from 19 for 1c up to an astonishing
value of roughly s = 80 for 1d.11 On the other hand, the kr of
the syn-diastereoisomers are surprisingly very similar being
unaffected by the steric hindrance of the substituent. Finally,
the acceleration produced by a syn disubstitution is compar-
able to that of a gem-dimethyl group (the difference between
the δΔG‡r;Exp of 1b and 1c or 1d is just of 0.12 kcal mol−1).
Alternatively, the anti-vic disubstitution induces a much higher
reaction rate enhancement; in the case of anti-1d it reaches the
awesome value of 415.
ꢀd½Eꢁ=dt ¼ k2½Eꢁ½TFAꢁ ¼ kobs½TFAꢁ:
The observed reaction rate constants kobs were obtained by
linear regression of ln([E]/[E]0) vs. time (for linear regression
plots see the ESI†). In Table 1 are reported the rate constants
(kr = k2,ester/k2,1a) and the Gibbs free energies of activation
Concerning point (iv), in Table 2 are reported the selectivity
s of 1c (s = kobs,anti-1c/kobs,syn-1c) and the corresponding differen-
tial energy barriers in different solvents, s changes signifi-
cantly from aprotic solvents to water. First, in water the ring
closure of hydroxyesters occurs by specific acid catalysis (SAC),
whereas, for aprotic solvents the cyclization goes through a
general acid catalysis (GAC),10 thus, it is not surprising that
Chart 1 γ-Hydroxyesters examined for the study of the vicinal disubsti-
tuent effect.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Organic & Biomolecular Chemistry
Paper
Table 2 Kinetic data of TFA catalysed ring closure of 1c at 303 K in
different solvents
complete catalytic cycle see the ESI†). Moreover, we concluded
that the different reaction rates are mainly related to enthalpy
factors, and were well rationalized in terms of energy strain.
In the present study we adopted the same model chemistry
previously used: each species of the RDS exists as a mixture of
different stereoisomers that we classified in relation to the con-
formational and stereochemical configurations (Fig. ESI3†).
For instance, syn-(3R,4R)-CC-1c is distributed into 8 energeti-
cally different stereoisomers: (1Re,g+,E), (1Re,g+,Z), (1Si,g+,E),
(1Si g+,Z), (1Re,g−,E), (1Re,g−,Z), (1Si,g−,E), and (1Re,g−,Z).16
The stereo descriptors g+/g− are related to the sign of the tor-
sional angle τ[OC(1)C(2)C(3)], while E or Z to the angle τ[EtO
(12)C(1)O(11)H(10)]. Finally, the 1Re/1Si face is relative to the
CO carbonyl group prochirality: for achiral esters 1a–b was
chosen arbitrarily 1Re, whereas for chiral 1c we arbitrarily ana-
lysed all the stereoisomeric subsets of the syn-(3R,4R)-1c and
the anti-(3S,4R)-1c diastereoisomers. For all esters, we found
that the E conformers of CC are thermodynamically more
stable than the Z isomers, for instance the Boltzmann distri-
bution analysis of all stereoisomers of anti-(3S,3R)-CC-1c indi-
cates that (1Re,g−,E) and (1Re,g+,E) have a molar fraction of x =
49.1% and x = 32.9%, respectively.
Selectivity
(s = kobs,anti-1c/kobs,syn-1c
δΔG‡r;Exp
a
Solvent
)
(kcal mol−1
)
εr
C6D6
26.2
18.9
12.6
10.5
−1.97
−1.77
−1.52
−1.41
2.28
4.81
36.64
78.54
CDCl3
d3-MeCN
D2O
a Dielectric constant.
the type of mechanism might have an influence on the extent
of the vic-effect. Moreover, we observed that in less polar sol-
vents, such as benzene, the anti-vic-effect is substantially stron-
ger than in more polar solvents, e.g. MeCN. This behaviour is
likely due to the strength of the H-bond interactions occurring
between the ester (E) and the TFA in the catalytic complex CC
(Fig. 2). In reverse, it is well known that by increasing the sol-
vent’s polarity the H-bond becomes weaker. At the most, in an
H-bonding disrupting solvent such as DMSO we did not
observe any reactivity.
For each stereoisomer of the catalytic complex we found the
corresponding TS1‡ (checked by IRC analysis, see the ESI†)
and TI1, and then each rate constant k was calculated with
Eyring’s equation. Thus, it was possible to determine the
mean rate constant kw of each stereoisomeric subset by the
Winstein–Holness equation (kw = ∑xi ki).17 Finally, the ensem-
ble free energy barrier ΔG‡Comp: was calculated by Eyring’s
equation. Using this approach, the differential computed
Gibbs energy barriers δΔG‡r;Comp correlated very well with the
experimental values δΔG‡r;Exp (coefficient of determination R2 =
0.99, Fig. ESI4†).
Computational study
In our previous article10 we performed a DFT computational
study (model chemistry B3LYP/6-31G(d)/SCRF = SMD,12–14
Gaussian09 software15), which established that the rate deter-
mining step (RDS) of the GAC ring closure of γ-hydroxyesters is
the conversion of the catalytic complex CC into the tetrahedral
intermediate TI1 through a transition state TS1‡ (Fig. 2, for the
In Table 3 we report the computed Gibbs energy barriers
ΔG‡ of each conformer of esters 1b and anti-1c, together with
the molar fractions x of the CC, the rate constants k and the
weighted rate constants (x × k); remarkably the ratio between
the two kw was very close to the experimental value (20 vs. 23).
However, the anti-vic disubstituent effect takes its origin by
the fact that for the hydroxyester anti-1c the CC conformer
with the lowest energy barrier, i.e. (1Si,g+,E), is also the one
mostly populated (x = 49.1%). Inversely, even if one of the con-
formers of the gem-dimethyl ester 1b has in absolute the
lowest energy barrier, i.e. (1Re,g−,E) with ΔG‡ = 11.95 kcal
mol−1, its reacting ground state is scarcely populated (x =
0.2%). In addition, the rotamer of gem-dimethyl ester 1b with
the highest molar fraction, i.e. (1Re,g+,E), is associated with a
chemical path with a much higher energy barrier (ΔG‡
=
15.22 kcal mol−1).
Interestingly, by the joined effect of these molar fractions
and rate constants, the lactonization of anti-1c goes mainly
through the (1Si,g−,E) stereoisomer (over 98%), whereas for 1b,
among all possible chemical paths, the ring closure occurs pre-
dominantly through the conformers with an E conformation.
Thus, in this case, the advantage of the anti-vic effect over the
classical Thorpe–Ingold effect arises primarily from a more
Fig. 2 Rate determining step of the TFA catalysed ring closure of
γ-hydroxyester 1a, and D/I-A decomposition of ΔE‡ for unimolecular
reactions.
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Table 3 Computed Gibbs energy barriers, molar fractions and rate constants relative to the RDS of lactonization of 1b and anti-1c
Ester
Configuration
xa (%)
ΔG‡ (kcal mol−1
)
kb (s−1
)
x × k (s−1
)
x × k/kw (%)
1b
(1Re,g+,E)
(1Re,g+,Z)
(1Re,g−,E)
(1Re, g−,Z)
85.2
5.8
0.2
8.8
15.22
15.98
11.95
16.19
66.8
19.0
57.0
1.1
25.7
1.18
67.0
1.3
30.3
1.4
kw 85.0 (s−1
)
1.53 × 104
13.4
anti-1c c
(1Re,g+,Z)
(1Re,g−,E)
(1Re,g−,Z)
(1Si,g+,Z)
(1Si,g−,E)
(1Si,g−,Z)
0.2
16.09
15.17
15.63
14.40
12.87
14.87
16.0
72.8
33.9
262.6
3341.7
120.2
0.04
24.0
4.9
2.1
1640.2
3.2
<0.1
1.4
0.3
0.1
98.0
0.2
32.9
14.4
0.8
49.1
2.6
kw 1647 (s−1
)
a By Boltzmann’s distribution of CC. b By Eyring’s equation. c For (3R,4S)-1c we found six out of eight possible chemical paths, and trajectories
associated with (g+,E) conformations appear to be forbidden for steric reasons.
favourable synergy between high population and low energy each class of conformers has its own ring closure activation
barriers, in agreement with the stereo population control strain mode (vide infra).
hypothesis of Cohen and Milstien.3a Such explanation is con-
Thenceforth, in order to gain more insights into the substi-
firmed by the perceived lower efficacy of the syn-vic effect of tuent effect, we carried out the analysis of the energy barrier
syn-1c, which can be compared to that of 1b (Table ESI4†).
by the distortion/interaction activation strain model (D/I-ASM)
Last but not least, if the kw are calculated taking into of Houk and Bickelhaupt,20 an approach inspired by the pio-
account exclusively the enthalpy component of the ΔG‡ the neering work of Morokuma and Ziegler on the activation
ratio kw,anti-1c/kw,1b decreases to 6.1 (δΔH‡ = 1.09 kcal mol−1
,
energy decomposition analysis21 (EDA).
Table ESI5†), and this result shows that the entropy too plays
Essentially, in this model ΔE‡ is decomposed as a sum of
two contributions: ΔE‡ = ΔEd‡ist + ΔEi‡nt. The distortion or strain
an important role in the anti-vic effect.18
However, at this stage, we were more interested to rational- energy ΔE‡dist takes into account all structural deformations
ize the relation between the reactivity and substituent effects needed to transform the reagents, from their lowest energy
(either geminal or vicinal) in terms of energy without thermal conformations, into the structures adopted in the transition
corrections. To begin with, the Brønsted–Evans–Polany (BEP) state. Instead, the interaction energy ΔE‡int is more related to
plots of compounds 1a–d showed a rough correlation between the electronic reorganization of the reagents needed to give
the energy barrier ΔE‡ and the reaction energy ΔErxn the product. This activation energy barrier decomposition,
(Fig. ESI5†);19 however, when they were repeated for classes of originally conceived for bimolecular reactions, was ingeniously
conformers, i.e. (g+,E), (g−,E), (g+,Z) and (g−,Z), the R2 improved extended by Bickelhaupt to unimolecular reactions through
significantly (Fig. 3 for (g−,E) and Fig. ESI6 and ESI8–9† for the the formal splitting of the reactant into two fragments.22 Since
other conformational subsets).
our RDS is a unimolecular reaction, we followed that pro-
This result suggests that, independent of the degree of sub- cedure by breaking the CC into the ester and TFA fragments
stitution and of the regio-/stereochemistry of hydroxyester, (Fig. 2). Therefore, the total distortion energy is:
ΔEd‡ist ¼ ΔEd‡ist;E þ ΔEd‡ist;TFA
‡
‡
TS1
CC
TFA
where, ΔEdist;E = ETES1 − EECC and ΔEdist;TFA = E
− E
.
TFA
In Fig. 4 we show the correlation of ΔE‡ with the ester dis-
tortion energy ΔE‡dist;E and with the TFA deformation energy
ΔE‡dist;TFA for the (g−,E) subset of esters. We were particularly
interested in this subset because it contains two of the most
reactive conformers, i.e. (1Re,g−,E)-1b (ΔE‡ = 11.17 kcal mol−1
and (1Re,3R,4R,g−,E)-1c (ΔE‡ = 11.90 kcal mol−1).
)
It appears that ΔE‡ is well correlated with the ester defor-
mation energy ΔE‡dist;E (R2 > 0.98), while the energy spent to
distort the TFA seems to have less influence on the different
reaction rates. Indeed, the percent difference of ΔE‡dist;TFA of
the slowest ester, i.e. 1a, with the fastest one, i.e. 1b, is much
lower than that of ΔE‡dist;E, 3% and 12%, respectively. Since the
interaction energy ΔE‡int, similarly to the ΔE‡dist;TFA, is poorly
correlated with the ΔE‡, we can conclude that the GAC ring
Fig. 3 Correlation between ΔE‡ and ΔErxn for the (g−,E) conformational
subset; the energies are in kcal mol−1
.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Organic & Biomolecular Chemistry
Paper
inside of each conformational subset, the substituent effect
cannot be related in a simple way to changes of individual geo-
metrical parameters.23 For instance, although (1Re,g−,E)-1b
and (1Re,3R,4R,g−,E)-1c have energy barriers very similar, the
differential dihedral angles Δτ (Δτ = τ CC − τ TS1‡) are quite
different, whereas the differential nucleophile–electrophile dis-
tance of 1.06 Å is more or less the same for all esters of this
conformational subset. Reasonably, the molecular rearrange-
ment of the ester from the unstrained state in the CC to the
strained state in the TS1‡ is the result of several distortions,
which overall cooperates with the scissoring process.
Diastereomeric resolution
Fig. 4 Correlation between ΔE‡ and distortion energies involved in the
lactonization of (g−,E)-1a–c subset of conformers; the energies are in
As a final point, we show a practical application of the vic-dis-
ubstituent effect consisting of the diastereoselective synthesis
of some of the most important naturally occurring cyclic esters
such as the Nicotiana tabacum lactone, i.e. 2c,24 the whisky and
cognac oak lactones, 2e and 2f respectively,25 and the Aerangis
lactone,26 i.e. 2g, and in addition the valero lactones 2h–i
(Fig. 6).
kcal mol−1
.
closure of γ-hydroxyesters is primarily an ester strain con-
trolled process. The lactonization of the other conformational
families of hydroxyesters showed analogous trends (see the
ESI†).
Interest on these compounds goes beyond the typical
natural products’ synthesis, since they possess unique olfac-
We reckon that the ester distortion energy ΔE‡dist;E is mainly
employed to bring the nucleophilic hydroxyl group closer to
the electrophilic CvO through a complex scissoring move-
ment during the transition state formation. To better illustrate
this point in Fig. 5 we show both CC and TS1‡ ester fragments
of the most reactive conformers of the (g−,E) subset together
with some geometrical parameters (we omitted the TI1 ester
fragments representation, since the RDS goes through a late
transition state). Certainly, this type of conformation somehow
determines a different variation of the ring strain exerted by
the substituents during the TS1‡ formation. However, even
Fig. 6 Lactones subjected to the resolution based on the vicinal disub-
stituent effect.
Fig. 5 Selected catalytic complex and transition state ester fragments
of the (g−,E) subset.
Scheme 1 Diastereomeric resolution of γ- and δ-vic-disubstituted
lactones.
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Table 4 Diastereomeric excesses and yields of the kinetic resolution of vic-disubstituted γ and δ lactones carried out at −35 °C, together with the
selectivity measured at 303 K
Trans
Cis
Ester
n
R′
R
Selectivitya (s = kanti/ksyn
)
Yieldb (%)
Dec (%)
Yieldd (%)
Dec (%)
1c
1d
1e
1f
1g
1h
1i
1
1
1
1
2
2
2
Me
i-Pr
n-Bu
n-Pent
n-Pent
Me
Et
Et
i-Pr
i-Pr
i-Pr
i-Pr
i-Pr
17.8 (42)e
78.4 (>250)e
21.7 (50)e
80
—
94
—
69
—
99
—
f
78
75
79
83
90h
98
96
94
92
92h
70
72
70
70
83h
99
98
98
98
98h
g
—
g
—
16.2 (34)e
t-Bu
142 (>250)e
a The selectivity was determined at 303 K. b Yields are based upon isolated products relative to anti-1. c By 1H-NMR. d Yields are based upon iso-
lated products relative to syn-1, over three steps: (i) O-silylation, (ii) protective group cleavage and (iii) ring closure. e Selectivity calculated at
238 K. f No resolution has been carried out. g No kinetics have been carried out. h The resolution was accomplished without O-silylation.
tory properties. Indeed, these γ and δ lactones are key-ingredi-
For the resolution of lactones 2e–i we found more con-
ents of the flavours & fragrances industry. Moreover, the venient to work with the isopropyl hydroxyl esters, i.e. 1e–i,
odour’s perception of chiral odorant molecules often depends since the latter are less reactive than the corresponding ethyl
on the absolute stereochemical configuration,27 however, it is esters, which instead undergo cyclization too easily.24 Thus,
not uncommon to find fragrances in which the odour lactones 2e–i (see the ESI† for the preparation of these sub-
threshold difference between the enantiomers is almost negli- strates) were opened with KOH in MeOH to give the potassium
gible, while for the diastereoisomers it is much more pro- carboxylate salt, which was then transformed into the isopro-
nounced. For instance, the two diastereoisomers of Aerangis pyl esters 2e–i with i-PrI in DMF (Scheme 1).31 The afforded
lactone 2g have very different odour thresholds, and the esters were subjected directly to kinetic resolution without any
perfume of the cis diastereoisomer is much more intense than purification; in Table 4 we report the selectivity (for the kinetic
the trans isomer.27b However, even if many asymmetric synth- study of 2e and 2g–i see the ESI†), and both yield and diaster-
eses of these lactones have been reported,26,27a,b,28 they are eomeric excess of cis and trans lactones.
still produced at the industrial scale as racemic mixtures of the
Table 4 shows that the des are in all cases very high (at
two diastereoisomers; this is likely due to the relatively low worst de = 92%), and even when the resolution was applied to
commercial value of these fragrances/flavours, which is still the synthesis of the less strained δ lactones (2g–2i) the results
economically unfeasible with the actual costs of most of the were still very satisfactory. The best des have been achieved for
stereoselective synthetic methodologies. Besides, chromato- the cis lactones, since most of the hydroxyester mixtures were
graphic separations on a large scale are impractical as well.29 unbalanced in favour of the syn-isomer. Particularly significant
Thus, a very cheap and efficient method for the resolution of is the resolution of 1i, indeed, despite the hydroxyester
diastereomeric mixtures of these lactones would be a highly mixture being enriched with the syn diastereoisomer (syn-1i/
1
desirable goal.
anti-1i, 78 : 22 by H-NMR), it was still possible to isolate the
First, with the ethyl hydroxyester 1c (syn/anti ≈ 1 : 1) in our trans-lactone, i.e. 2i, with a good de of 92%.
hands, we tested the TFA catalysed diastereoselective ring
Not surprisingly, the selectivity improved by increasing the
closure in CH2Cl2 on a preparative scale (Scheme 1). In order steric hindrance of one of the two substituents, and emble-
to improve the diastereomeric excess of lactone trans-2c and of matic is the case of 2h and 2i: by changing a methyl with a
the unconsumed syn-1c, the reaction was carried out at t-butyl group the selectivity improved by almost one order of
−35 °C.
magnitude, s = 16 and 142, respectively. Noteworthily, the
After that most of the hydroxyester anti-1c converted into hydroxyester syn-1i ring closes so slowly with respect to the
the lactone trans-2c,30 the reaction was quenched with imid- anti-1i that the resolution was accomplished without the need
azole and the unreacted isomer syn-1c was quantitatively for the O-silylating step. Indeed, in this case the unconsumed
O-silylated with SiEt3Cl affording the silyl ether 3. Then, the syn hydroxyester, i.e. syn-1i, and the trans-lactone, i.e. trans-2i,
latter was easily separated from the very polar lactone, i.e. were directly separated by column chromatography.
trans-2c, by a column chromatography technique, or alterna-
tively the two products could be separated by bulb-to-bulb dis-
tillation. Finally, the silyl ether 3 was quantitatively trans-
formed into the lactone cis-2c by treatment with an excess of
Conclusion
TFA in CH2Cl2. By adopting this procedure both trans and cis To sum up, we found that the GAC lactonization rate of vic-di-
isomers of Nicotiana tabacum lactone were isolated with a high substituted γ and δ-hydroxyesters is strongly influenced by the
diastereomeric excess, de
=
94% and 99%, respectively relative stereochemistry: the syn-vic effect is comparable to the
(Table 4).
classical gem disubstituent effect, whereas the anti-vic effect
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Organic & Biomolecular Chemistry
Paper
overwhelms the latter by one order of magnitude at the worst. 1H-NMR),30 the reaction mixture was quenched with imidazole
The rate acceleration can be amplified simply by increasing (2.0 eq.) and left to reach r.t., then the solvent was removed
the steric hindrance of one of the two substituents: for under reduced pressure to give a crude material, which was
instance, in the case of 1d the rate is enhanced by a factor of left for at least 15 min under high vacuum.34 To a solution of
almost two orders of magnitude relative to the ring closure of the crude material in CH2Cl2 (30 mL) were added Et3SiCl
the gem-dimethyl hydroxyl ester. Computations suggest that (1.1 eq. with respect to anti-1) and a catalytic amount of
the observed rate accelerations may be attributed, primarily, to DMAP. After 12 hours, the reaction mixture was concentrated
a synergy between the population and the conformer reactivity, under reduced pressure and subjected to the column chrom-
which for the anti diastereoisomers is much more favourable atography purification procedure using a gradient eluent
with respect to the syn or the gem isomers. These conclusions (SiO2, n-hexane/AcOt, 95 : 5 → 6 : 4) affording in order of
are in agreement with the stereo population hypothesis of elution the silyl derivative 3 and then trans-2; the latter was
Cohen and Milstien, and even though, entropic effects cannot distilled with a bulb to bulb apparatus. Alternatively, the
be neglected. Moreover, we found a strong correlation between organic phase was diluted with CH2Cl2 (30 mL), washed with
the ester strain energy and the activation barrier.
HCl (0.2 M, 2 × 30 mL) and brine (sat., 1 × 30 mL), dried over
Clearly, the vic-effect opens a new paradigm in the field of Na2SO4 and concentrated under vacuum. Finally, the lactone
generalized substituent effect, since it offers the possibility of was isolated by distillation with a bulb to bulb apparatus.
controlling the ring closure rate through the relative stereo-
Trans-4,5-dimethyldihydrofuran-2(3H)-one (trans-2c). Yield =
chemistry of the substituents; and this is the real novelty with 80% (0.94 g as a colourless liquid); b.p. 70–80 °C, 0.5 mmHg;
respect to the classical Thorpe–Ingold effect. We reckon that, purity 99% by GC (tR = 6.58 min); de 94% by 1H-NMR;
as the gem disubstituent effect has undoubtedly played an 1H-NMR (400 MHz, CDCl3): δ = 4.15 (tt, J = 7.1, 5.5, 3H),
important role for over 100 years in several fields of organic 2.76–2.57 (m, 1H), 2.27–2.08 (m, 2H), 1.41 (d, J = 6.2, 3H), 1.15
chemistry, the vicinal effect should provide further benefits. In (d, J = 6.2, 3H). 13C-NMR (101 MHz, CDCl3): δ = 176.4, 83.5,
this regard, the reported highly efficient diastereomeric resolu- 38.3, 37.4, 19.2, 16.9. GC-MS: m/z (%) 114 [M]+ (10), 99 (20), 70
tion of the very important Quercus flavours and Aerangis (60), 55 (50), 42 (100).24
l. fragrance is a preliminary example of the potentiality of the
Procedure B: achievement of cis-lactones with high de. This
vic-disubstituent effect. Computational studies and further procedure is similar to that of procedure A; however a higher
applications are ongoing in our laboratory and will be the amount of TFA (0.15 eq.) was used, and the conversion of 1
1
subject of upcoming reports.
into 2 was pushed over 50% (checked by H-NMR),29 and the
O-silylation was carried out under the same conditions
described for procedure A.
(syn)-Ethyl 3-methyl-4-((triethylsilyl)oxy)pentanoate (3c).
Experimental
General procedure for the diastereomeric resolution of
Yield 78% (2.8 g as a colorless liquid); purity 98% by GC (tR
18.91 min); de 99% by H NMR; H NMR (400 MHz, CDCl3):
=
1
1
lactones
δ = 4.14 (q, J = 7.1, 2H), 3.81 (qd, J = 6.3, 3.5, 1H), 2.59–2.46 (m,
Ring opening of lactones to give hydroxyesters. To an ice 1H), 2.14–2.00 (m, 2H), 1.27 (t, J = 7.2, 3H), 1.09 (d, J = 6.3,
cooled and well stirred solution of lactone 2 (17.0 mmol) in 3H), 1.04–0.86 (m, 15H), 0.60 (td, J = 8.3, 7.7, 7H). 13C-NMR
MeOH (30 mL) was added a solution of KOH (19.0 mmol) in (101 MHz, CDCl3): δ = 173.7, 70.8, 60.1, 37.5, 37.4, 20.0, 14.7,
H2O (3 mL). The heterogeneous mixture was stirred for 14.3, 6.9, 5.2. GC-MS: m/z (%) 273 [M − 1]+ (1), 259 (2), 245
5 hours at room temperature and then concentrated under (100); HRMS (quadrupole) calcd for C12H25O3Si+ [M − Et]+
vacuum to give a viscous oil. The latter was treated with Et2O 245.1573, found 245.1572.
(4 × 10 mL) and concentrated under reduced pressure. This
procedure was repeated at least 4 times in such a way to elim-
General procedure for the cleavage of O-silyl protective group
inate all traces of H2O and MeOH. The crude material was left To a solution of 3 (1.0 mmol) in CH2Cl2 (5 mL) was added TFA
under high vacuum for 6 hours. Then, to a solution of the (100 μL). After complete conversion of 3 into cis-2, checked by
crude mixture in anhydrous DMF (45 mL) was added 2-iodo- TLC, the reaction mixture was diluted with Et2O (15 mL) and
propane (5.8 g, 34 mmol). After 14 hours, the reaction mixture washed with a solution of NaHCO3 (sat., 1 × 30 mL) and brine
was diluted with brine (sat., 70 mL) and then extracted with (sat., 1 × 20 mL). The organic phase was dried over Na2SO4,
Et2O (5 × 40 mL). The combined organic phase was washed concentrated under reduced pressure and the crude material
with brine (sat., 50 mL), dried over Na2SO4 and concentrated was subjected to the bulb to bulb distillation apparatus
under reduced pressure to give the corresponding hydroxyester affording cis-2.
1,32 which was of sufficient purity33 for the next step.
cis-4,5-Dimethyldihydrofuran-2(3H)-one (cis-2c). Yield = 89%
Procedure A: achievement of trans-lactones with high de. To (1.0 g as a colourless liquid); b.p. 73–81 °C, 0.5 mmHg; purity
a mixture of syn/anti hydroxyester 1 (23.4 mmol) in CH2Cl2 98% by GC (tR = 7.34 min); de 99% by 1H-NMR; 1H-NMR
(60 mL) at – 35 °C was added a precooled solution of TFA (400 MHz, CDCl3): δ = 4.65 (m, 1H), 2.75–2.52 (m, 2H), 2.21
(0.01 ÷ 0.1 eq.) in CH2Cl2 (2 mL) under a N2 atmosphere. After (dd, J = 16.6, 5.1, 1H), 1.29 (d, J = 6.5, 3H), 1.03 (d, J = 6.9, 3H).
the conversion of most of anti-1 hydroxyester (checked by 13C-NMR (101 MHz, CDCl3): δ = 176.8, 79.7, 37.0, 33.5, 15.5,
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
14.0. MS: m/z (%) 114 [M]+ (10), 99 (20), 70 (60), 55 (50), 42
9 (a) T. Higuchi, L. A. Eberson and K. Herd, J. Am. Chem.
Soc., 1966, 88, 3805–3808; (b) L. Eberson and
L.-A. Svensson, Acta Chem. Scand., 1972, 26, 2631–2641.
(100).24
Kinetics. The reaction progress was monitored by measuring
1
a series of H-NMR (500 MHz) spectra recorded as pseudo 2D 10 E. Brenna, F. Distante, F. G. Gatti and G. Gatti, Catal. Sci.
of a sample of hydroxyesters in CDCl3 at 303 K allowing a fixed
time delay between successive spectra. A pulse width of 30°
and a relaxation delay of 2 s were used in order to avoid errors
Technol., 2017, 7, 1497–1507. Table 2, last column, row 5
contains a misprint for the ensemble ΔG‡1a, ensemble =
16.06 kcal mol−1. The correct value is 16.16 kcal mol−1
.
due to the relaxation. An exponential filter was applied to FID 11 R. E. Gawley, J. Org. Chem., 2005, 71, 2411–2415.
for optimization of the signal-to-noise ratio. Baseline correc- 12 (a) A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652;
tion was applied to spectra before integration. The bias and
slope of integrals were carefully adjusted. The sample was pre-
pared by adding directly, with a gastight Hamilton syringe, a
solution of TFA in CH2Cl2 (20 ÷ 60 μL, 10−2 M) to a solution of
hydroxyester (typically ≈8–10 mg each) in CDCl3 (500.0 μL)
into a NMR tube.
(b) C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens.
Matter Mater. Phys., 1988, 37, 785–789; (c) S. Vosko,
H. L. Wilk and M. Nusair, Can. J. Phys., 1980, 58, 1200–
1211; (d) P. J. Stephens, F. J. Devlin, C. F. Chabalowski and
M. J. Frisch, J. Phys. Chem., 1994, 98, 11623–11627.
13 R. Ditchfield, W. J. Hehre and J. A. Pople, J. Chem. Phys.,
1971, 54, 724–728.
14 V. Marenich, C. J. Cramer and D. G. Truhlar, J. Phys. Chem.
B, 2009, 113, 6378–6396.
Conflicts of interest
15 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone,
B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato,
X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng,
J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda,
J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, T. Vreven, J. A. Montgomery Jr., J. E. Peralta,
F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers,
K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand,
K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar,
J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene,
J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo,
R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin,
R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,
K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador,
J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö Farkas,
J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox,
Gaussian 09: R. A, Gaussian, Inc., Wallingford, CT, 2009.
16 The overall number of stereoisomers is twice with respect
to the number of energetically different isomers, for
instance for syn-1c there are eight couples of enantiomers.
17 J. I. Seeman, Chem. Rev., 1983, 83, 84–134.
There are no conflicts to declare.
Acknowledgements
We thank POR-Lombardia, VIPCAT project (Value Added
Innovative Protocols for Catalytic Transformations-ID 228775)
for the financial support. We acknowledge CINECA for the
high performance computing resources and support (Class C
Research Projects ISC21, ISC29, ISC37). We are grateful to
Dr Douglas J. Fox of Gaussian Inc. for support. We thank Mr
Alberto Betti of Lombardia Informatica for technical
assistance.
Notes and references
1 R. M. Beesley, C. K. Ingold and J. F. Thorpe, J. Chem. Soc.
Trans., 1915, 107, 1080–1106.
2 M. E. Jung and G. Piizzi, Chem. Rev., 2005, 105, 1735–1766.
3 (a) S. Milstien and L. A. Cohen, Proc. Natl. Acad. Sci. 18 R. Salvio, L. Mandolini and C. Savelli, J. Org. Chem., 2013,
U. S. A., 1970, 67, 1143–1147; (b) S. Milstien and 78, 7259–7263.
L. A. Cohen, J. Am. Chem. Soc., 1972, 94, 9158–9165; 19 (a) R. A. Marcus, Annu. Rev. Phys. Chem., 1964, 15, 155;
(c) M. Caswell and G. L. Schmir, J. Am. Chem. Soc., 1980, (b) G. S. Hammond, J. Am. Chem. Soc., 1955, 77, 334–338.
102, 4815–9165; (d) M. N. Levine and R. T. Raines, Chem. 20 (a) D. H. Ess and K. N. Houk, J. Am. Chem. Soc., 2007, 129,
Sci., 2012, 3, 2412–1420.
10646–10647; (b) F. M. Bickelhaupt and K. N. Houk, Angew.
Chem., Int. Ed., 2017, 56, 10070–10086; (c) W.-J. Van Zeist
and F. M. Bickelhaupt, Org. Biomol. Chem., 2010, 8, 3118–
3127; (d) I. Fernández and F. M. Bickelhaupt, Chem. Soc.
Rev., 2014, 43, 4953–4967.
4 D. J. Nelson, I. W. Ashworth, I. H. Hillier, S. Kyne,
H. S. Pandian, J. A. Parkinson, J. M. Percy, G. Rinaudo and
M. A. Vincent, Chem. – Eur. J., 2011, 17, 13087–13097.
5 H. Chen, L. G. Wang, F. Wang, L. P. Zhao, P. Wang and
Y. Tang, Angew. Chem., Int. Ed., 2017, 56, 6942–6945.
6 C. Toniolo, M. Crisma, F. Formaggio and C. Peggion,
Biopolymers, 2001, 60, 396–419.
21 (a) K. Morokuma, Acc. Chem. Res., 1977, 10, 294–300;
(b) S. Nagase and K. Morokuma, J. Am. Chem. Soc., 1978,
100, 1666–1672; (c) T. Ziegler and A. Rauk, Theor. Chim.
Acta, 1977, 46, 1–10.
7 S. Lee, E.-H. Kang, H. Park and T.-L. Choi, Chem. Sci., 2012,
3, 761–765.
22 (a) I. Fernández, F. M. Bickelhaupt and F. P. Cossío, Chem.
8 M. E. Jung, Synlett, 1999, 843–846.
–
Eur. J., 2012, 18, 12395–12403; (b) I. Fernández,
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Organic & Biomolecular Chemistry
Paper
F. M. Bickelhaupt and F. P. Cossío, Chem. – Eur. J., 2014, 28 C. Kumru, T. Classen and J. Pietruszka, ChemCatChem,
20, 10791–10801. 2018, 10, 1–11.
23 F. Liu, R. S. Paton, S. Kim, Y. Liang and K. N. Houk, J. Am. 29 H. G. Schmarr, M. Mates, K. Wall, F. Metzner and
Chem. Soc., 2013, 135, 15642–15649. M. Frefel, J. Chromatogr. A, 2017, 1516, 135–141.
24 E. Brenna, F. G. Gatti, D. Monti, F. Parmeggiani, 30 The reaction progress was monitored by H-NMR analysis.
1
A. Sacchetti and J. Valoti, J. Mol. Catal. B: Enzym., 2015,
114, 77–85.
25 A. Mosandl, W. Deger, M. Gessner, C. Gunther,
G. Heusinger and G. Singer, Lebensmittelchem. Gerichtl.
Chem., 1987, 41, 35.
26 E. Brenna, C. Dei Negri, C. Fuganti and S. Serra,
Tetrahedron: Asymmetry, 2001, 12, 1871–1879.
27 (a) E. Brenna, C. Fuganti and S. Serra, Tetrahedron:
Asymmetry, 2003, 14, 1–42; (b) A. Abate, E. Brenna,
A sample (0.1 mL) was quickly added to a precooled solu-
tion of NEt3 (50 mg) in CH2Cl2 (0.5 mL), and then the
solvent was removed under reduced pressure and subjected
to 1H-NMR analysis.
31 C. Günther and A. Mosandl, Liebigs Ann. Chem., 1986, 12,
2112–2122.
32 The hydroxyesters can be stored at −20 °C for several
months, in the presence of a small amount of NEt3, other-
wise at rt they undergo spontaneous cyclization.
C. Fuganti, F. G. Gatti and S. Serra, Chem. Biodiversity, 33 The purity of hydroxyesters cannot be detected by GC-MS,
2004, 1, 1888–1898; (c) E. Brenna, C. Fuganti, F. G. Gatti because the esters partially cyclize into the GC-injector.
and S. Serra, Chem. Rev., 2011, 111, 4036– 34 It is important to remove the alcohol (i-PrOH or EtOH)
4072.
before the O-silylation.
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.