On the effect of tether composition on cisltrans selectivity in intramolecular diels-alder reactions

Article abstract of DOI:10.1002/asia.200800352

Intramolecular Diels-Alder (IMDA) transition structures (TSs) and energies have been computed at the B3LYP/6-31+G(d) and CBS-QB3 levels of theory for a series of 1,3,8-nonatrienes, H₂C=CH-CH=CH-CH₂X-Z-CH= CH₂ [-X-Z-=-CH₂-CH₂(1); -O-C(=O)- (2); -CH ₂-C(=O)-(3); -O-CH₂- (4); -NH-C(=O)- (5); -S-C(=O)- (6); -O-C(=S)- (7); -NH-C(=S)- (8); -S-C(=S)- (9)]. For each system studied (1-9), cis- and trans-TS isomers, corresponding, respectively, to endo- and exo-positioning of the -C-X-Z- tether with respect to the diene, have been located and their relative energies (E_rel^TS) employed to predict the cis/trans IMDA product ratio. Although the E_rel ^TS values are modest (typically <3kJmol^-1), they follow a clear and systematic trend. Specifically, as the electronegativity of the tether group X is reduced (X=O-→ NH or S), the IMDA cis stereoselectivity diminishes. The predicted stereochemical reaction preferences are explained in terms of two opposing effects operating in the cis-TS, namely (1) unfavorable torsional (eclipsing) strain about the C4-C5 bond, that is caused by the -C-X-Q=Y)- group's strong tendency to maintain local planarity; and (2) attractive electrostatic and secondary orbital interactions between the ertdo(thio)carbonyl group, C=Y, and the diene. The former interaction predominates when X is weakly electronegative (X=N, S), while the latter is dominant when X is more strongly electronegative (X=O), or a methylene group (X=CH₂) which increases tether flexibility. These predictions hold up to experimental scrutiny, with synthetic IMDA reactions of 1, 2, 3, and 4 (published work) and 5, 6, and 8 (this work) delivering ratios close to those calculated. The reactions of thiolacrylate 5 and thioamide 8 represent the first examples of IMDA reactions with tethers of these types. Our results point to strategies for designing tethers, which lead to improved cis/trans-selectivities in IMDAs that are normally only weakly selective. Experimental verification of the validity of this claim comes in the form of fumaramide 14, which undergoes a more trans-selective IMDA reaction than the corresponding ester tethered precursor 13.

Full text of DOI:10.1002/asia.200800352

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
FULL PAPERS
DOI: 10.1002/asia.200800352
On the Effect of Tether Composition on cis/trans Selectivity in
Intramolecular Diels–Alder Reactions
Michael N. Paddon-Row,*[a] Alistair I. Longshaw,[b] Anthony C. Willis,[b] and
Michael S. Sherburn*[b]
Abstract: Intramolecular Diels–Alder
(IMDA) transition structures (TSs) and
energies have been computed at the
the tether group X is reduced (X=O!
NH or S), the IMDA cis stereoselectiv-
ity diminishes. The predicted stereo-
chemical reaction preferences are ex-
plained in terms of two opposing ef-
fects operating in the cis-TS, namely
(1) unfavorable torsional (eclipsing)
latter is dominant when X is more
strongly electronegative (X=O), or a
methylene group (X=CH2) which in-
creases tether flexibility. These predic-
tions hold up to experimental scrutiny,
with synthetic IMDA reactions of 1, 2,
3, and 4 (published work) and 5, 6, and
8 (this work) delivering ratios close to
those calculated. The reactions of thio-
lacrylate 5 and thioamide 8 represent
the first examples of IMDA reactions
with tethers of these types. Our results
point to strategies for designing tethers,
which lead to improved cis/trans-selec-
tivities in IMDAs that are normally
only weakly selective. Experimental ve-
rification of the validity of this claim
comes in the form of fumaramide 14,
which undergoes a more trans-selective
IMDA reaction than the corresponding
ester tethered precursor 13.
B3LYP/6-31+G(d)
and
CBS-QB3
levels of theory for a series of 1,3,8-
ꢀ
ꢀ
ꢀ
nonatrienes, H2C=CH CH=CH CH2
ꢀ ꢀ ꢀ ꢀ ꢀ ꢀ
ꢀ
ꢀ
X Z CH=CH2 [ X Z = CH2 CH2
ꢀ ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
(1); O C(=O) (2); CH2 C(=O)
(3); O CH2 (4); NH C(=O) (5);
strain about the C4 C5 bond, that is
ꢀ ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ ꢀ ꢀ
ꢀ
caused by the C X C(=Y) groupꢀs
strong tendency to maintain local pla-
narity; and (2) attractive electrostatic
and secondary orbital interactions be-
tween the endo-(thio)carbonyl group,
C=Y, and the diene. The former inter-
action predominates when X is weakly
electronegative (X=N, S), while the
ꢀ ꢀ
ꢀ
ꢀ ꢀ
ꢀ
S C(=O)
(6);
O C(=S)
(7);
ꢀ
ꢀ
ꢀ
ꢀ ꢀ
ꢀ
NH C(=S) (8); S C(=S) (9)].
For each system studied (1–9), cis- and
trans-TS isomers, corresponding, re-
spectively, to endo- and exo-positioning
ꢀ ꢀ ꢀ ꢀ
of the C X Z tether with respect to
the diene, have been located and their
relative energies (ErelTS) employed to
predict the cis/trans IMDA product
Keywords: cycloaddition · density
functional calculations · diastereo-
selectivity · tether groups · transi-
tion structures
TS
ratio. Although the Erel values are
modest (typically <3 kJmolꢀ1), they
follow a clear and systematic trend.
Specifically, as the electronegativity of
Introduction
[a] Prof. M. N. Paddon-Row
School of Chemistry
The intramolecular Diels–Alder reaction (IMDA) offers an
efficient, atom economic, and versatile solution to the prob-
lem of constructing multicyclic molecules.[1] The full poten-
tial of this important reaction will be realized only when the
stereochemical outcome of IMDAs can be predicted and,
hence, controlled experimentally. This goal remains elusive,
notwithstanding recent significant progress towards delineat-
ing the various factors controlling IMDA stereochemis-
try.[2–18] In this paper, we explore, using density functional
theory, the origin of cis/trans selectivity in IMDAs of 1,3,8-
nonatrienes, 1–9, possessing tethers that are three atoms
long (Scheme 1). Given the experimental fact that IMDAs
generally give kinetically controlled product distributions,
the key to understanding cis/trans selectivity in these reac-
University of New South Wales
Sydney, NSW 2052 (Australia)
(computational)
E-mail: m.paddonrow@unsw.edu.au
[b] Dr. A. I. Longshaw, Dr. A. C. Willis,+ Prof. M. S. Sherburn
Research School of Chemistry
Australian National University
Canberra, ACT 0200 (Australia)
Fax : (+61) 2-6125-8114
(synthetic)
E-mail: sherburn@rsc.anu.edu.au
[+] Correspondence author for X-ray crystallography
E-mail: willis@rsc.anu.edu.au
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/asia.200800352.
126
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2009, 4, 126 – 134

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
level of theory and CBS-QB3 gave similar results for the
corresponding IMDA reactions of 2 and 7. All calculations
refer to the gas phase. In a previous study,[12] we have found
that there is excellent agreement between gas phase B3LYP/
6-31+G(d) predicted IMDA cis/trans ratios and the experi-
mental ratios obtained using weakly polar solvents—which
are the solvents of choice for carrying out IMDA reactions.
Consequently, we ignore solvent effects in this study. Both
cis and trans IMDA TSs were located for each molecule.
Harmonic vibrational frequencies (at the same level of
theory) were employed to characterize the TSs as first-order
saddle points. The unscaled vibrational frequencies were
also used to calculate the enthalpies and Gibbs free energies
of the TSs at 298.15 K. Also reported are TS electronic en-
ergies at 0 K corrected for zero-point vibration energies
(ZPE). Cis/trans product ratios were calculated from the
Gibbs free energies of the TSs using the rate expression de-
rived from standard transition state theory.[32] The cis/trans
ratios listed in Table 2 refer to 298.15 K whereas those given
in Scheme 2–7 and Scheme 9 were calculated using the ex-
perimental temperature.
Scheme 1. IMDA reactions under scrutiny. Reactant trienes give rise to
cis and trans fused bicyclic products from cis and trans TSs, respectively.
tions is to calculate and examine the IMDA transition struc-
tures (TSs) leading to the products (Table 1).
The origin of the cis/trans stereoselectivity[19] in the
IMDA of the hydrocarbon 1,3,8-nonatriene[20] 1 has been
previously elucidated by Houk et al.[21–23] In this paper, we
focus primarily on IMDAs of the series of trienes 2–9, with
tethers possessing heteroatoms. Significantly, our results sug-
gest that arguments used to explain the cis/trans selectivity
in the IMDA of 1 are not applicable to reactant molecules
with tethers containing either the ester or amide functionali-
ty, for example, 2, 5–9. This finding has enabled us to predict
ways for improving cis/trans selectivity in IMDAs by judi-
cious choice of tether composition.
The Gaussian 98[33] and 03[34] program packages were used
throughout. Details of the structures (in Cartesian coordi-
nate form) and energies of all computed TSs are given in
the Supporting Information.
Results and Discussion
The numbering scheme for the atoms in the TSs is presented
in Figure 1, together with definitions of salient geometrical
parameters. Two types of asynchronicity characterize IMDA
Computational Details
All IMDA TSS associated with 1–9 were calculated using
two methods: (a) the B3LYP functional[24] and the 6-
31+G(d) basis set,[25,26] and (b) the multilevel CBS-QB3
method. Although B3LYP/6-31+G(d) is known to give ac-
ceptable relative energies and geometries of various DA
TSs,[2–18,27,28] we found that it predicted an incorrect cis/trans
selectivity in the IMDA reaction of the thiolacetate 6. Con-
sequently, in addition to the DFT calculations, we also used
the more accurate (but computationally much more expen-
sive) CBS-QB3 method,[29] a member of the complete basis
set methods developed by Petersson et al.[29,30] The CBS-
QB3 method uses a B3LYP/CBSB7 optimized geometry and
frequencies together with CCSD(T), MP4SDQ, and MP2
single-point calculations and a CBS extrapolation to pro-
duce accurate energies. The CBS-QB3 method has recently
been successfully applied in the investigation of DA reac-
tions involving furan.[31] CBS-QB3 calculations were not car-
ried out for 13 and 14 (Scheme 10), owing to their large size.
It transpired that B3LYP/6-31+G(d) calculations on these
two systems gave results in accord with experiment. Only
B3LYP/6-31+G(d) calculations were carried out on the in-
termolecular TSs for 11 and 12 (Scheme 8) because this
Figure 1. Schematic of IMDA transition structures (cis or trans) for 1–9
depicting the numbering schemes and relevant geometrical parameters.
TSs, namely bond-forming asynchronicity, Dras, defined as
the difference between the lengths of the peripheral (rp) and
internal (ri) forming bond lengths, and the twist-mode asyn-
chronicity, qas,[21] which measures the extent of twisting of
the dienophile double bond about the forming internal bond
(Figure 1). Values of salient geometrical parameters for the
various B3LYP/6-31+G(d) optimized TSs are presented in
Table 1 (the B3LYP/6-31+G(d) and B3LYP/CBSB7 TS geo-
metries are very similar to each other). Relative energies,
enthalpies (H), and Gibbs free energies (G) between the cis
Chem. Asian J. 2009, 4, 126 – 134
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
127

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
M. N. Paddon-Row, M. S. Sherburn et al.
FULL PAPERS
Table 1. Salient IMDA TS (1TS–9TS) B3LYP/6-31+G(d) optimized geometrical parameters (Figure 1), in-
cluding forming bond lengths ri and rp (ꢂ), their difference Dras (ꢂ), and key dihedral angles qas, q1, q2, and q3
(8).
mately 178 in the former, com-
pared to approximately 368 in
the latter TS. These two unfav-
orable features of trans-1TS are
evidently more energetically
costly than the close contact be-
tween H4 and H8 in cis-1TS
(2.35 ꢂ). Within the context of
this study, it is important to
point out that both cis-1TS and
trans-1TS are practically free of
torsional strain about the
TS
ri
rp
Dras
qas
q1
q2
q3
cis-1TS
trans-1TS
cis-2TS
trans-2TS
cis-3TS
trans-3TS
cis-4TS
trans-4TS
cis-5TS
trans-5TS
cis-6TS
trans-6TS
cis-7TS
trans-7TS
cis-8TS
2.346
2.194
2.226
2.159
2.478
2.317
2.300
2.152
2.254
2.176
2.310
2.296
2.266
2.212
2.291
2.228
2.400
2.370
2.162
2.280
2.310
2.355
2.083
2.192
2.211
2.347
2.281
2.323
2.232
2.238
2.284
2.310
2.267
2.290
2.175
2.191
ꢀ0.184
0.086
0.084
15.4
ꢀ12.7
5.9
ꢀ8.8
14.2
ꢀ4.8
16.8
ꢀ15.0
7.9
–
–
–
–
ꢀ40.3
ꢀ44.7
3.4
ꢀ31.0
161.5
ꢀ162.3
101.5
ꢀ133.0
–
0.196
31.0
ꢀ63.9
ꢀ40.7
ꢀ66.9
ꢀ34.8
ꢀ46.5
ꢀ4.7
ꢀ0.395
ꢀ0.125
ꢀ0.089
0.195
0.027
0.147
ꢀ0.078
ꢀ0.058
0.018
ꢀ14.8
15.2
–
–
–
ꢀ33.5
158.6
ꢀ160.4
152.9
ꢀ154.6
160.6
ꢀ160.5
162.5
ꢀ161.5
149.4
ꢀ154.6
ꢀ9.7
33.4
ꢀ61.6
ꢀ
2.0
ꢀ23.0
22.8
2.8 C4 C5 bond, with q3 (Figure 1)
ꢀ1.9
4.5
ꢀ68.8
ꢁ408 and 458, respectively.
ꢀ32.3
0.3
ꢀ64.3
ꢀ4.4
ꢀ63.2
–
As in the case of 1, the
IMDA of acrylate 2 displays cis
selectivity, but is weaker, with
cis-2TS lying only 2.6 kJmolꢀ1
below trans-2TS. The CBS-QB3
cis/trans ratio at 1808C is in
accord with the experimental
0.098
ꢀ6.5
31.0
ꢀ0.024
6.9
ꢀ37.0
35.0
trans-8TS
cis-9TS
trans-9TS
0.062
ꢀ7.8
1.3
0.0
ꢀ0.225
ꢀ0.179
ꢀ25.3
25.0
–
and trans TSs for each reaction (ZrelTS =ZcisꢀTSꢀZtransꢀTS; Z=
E+ZPE, H, G) are given in Table 2. Also given in Table 2
are cis/trans product ratios at 298.15 K. Perhaps not unex-
pectedly for structurally similar TSs, there is little difference
ratio[12] (Scheme 3). Unlike trans-1TS, there is no significant
Table 2. ZPE-Corrected CBS-QB3 (B3LYP/6-31+G(d) values in paren-
TS
TS [a,b]
theses) cis/trans-TS relative energies (Erel
)
[a], enthalpies (Hrel
)
and
TS
between the ErelTS, HrelTS, and Grel values for each system
TS
Gibbs free energies (Grel
)
[a,b] kJmolꢀ1, and cis/trans product ratios.[c]
TS
TS
TS
cis/trans[b]
studied. Thus, unless stated otherwise, the ensuing discussion
refers to ErelTS values derived from the CBS-QB3 method.
Experimentally, the IMDA reaction of 1,3,8-nonatriene 1
gives a 75:25 ratio of cis/trans products at 1908C in cyclo-
hexane,[22] which was successfully reproduced computation-
ally using HF/3-21G and hybrid QM-MM models.[23] Our
CBS-QB3 and DFT calculations also correctly predict the
predominant formation of cis adduct from the IMDA of 1,
with cis-1TS favored over the trans-1TS by 5.34 kJmolꢀ1.
The CBS-QB3 cis/trans ratio, calculated using the experi-
mental temperature (1908C), is in good agreement with the
experimentally determined ratio (Scheme 2). Both cis-1TS
and trans-1TS are only moderately asynchronous, with the
forming peripheral bond slightly shorter than the forming
internal bond (Dras =ꢀ0.18 ꢂ) in cis-1TS, but longer than
the internal bond in cis-1TS (Dras =0.086 ꢂ; Scheme 2). As
noted by Houk et al.,[21] both TSs exhibit marked twist-
mode asynchronicity, amounting to 158 (endo twisting) and
ꢀ138 (exo twisting) in cis-1TS and trans-1TS, respectively,
which was attributed to the tendency of the dihedral angle
Structure
Erel
Hrel
Grel
ꢀ5.34
(ꢀ5.85)
ꢀ5.52
ꢀ4.86
88:12
ACHTUNGTRENNUNG(90:10)
A
A
A
ꢀ2.62
ꢀ5.03
ꢀ5.67
0.66
75:25
88:12
78:22
(ꢀ1.88)
(ꢀ1.75)
(ꢀ2.06)
G
ꢀ5.10
ꢀ6.12
0.82
ꢀ4.85
ꢀ4.38
0.39
(ꢀ5.65)
(ꢀ5.83)
(ꢀ5.01)
(88:12)
(ꢀ6.42)
(ꢀ6.81)
(ꢀ5.59)
ACHTUNGTRENNUNG
(0.59)
(0.78)
(0.28)
ACHTUNGTRENNUNG
0.11
(2.17)
0.26
(2.37)
ꢀ0.36
(1.63)
G
G
G
ACHTUNGTRENNUNG(34:66)
ꢀ4.67
ꢀ1.30
ꢀ0.49
ꢀ4.45
ꢀ1.05
ꢀ0.24
ꢀ5.07
ꢀ1.67
ꢀ1.48
89:11
ACHTUNGTRENNUNG(81:19)
ꢀ
ꢀ
ꢀ
C5 C4 C8 C7 in the developing 5-membered ring to be as
close as possible to zero degrees. This theory was confirmed
(ꢀ3.15)
(ꢀ2.90)
(ꢀ3.62)
by optimizing the structure of cyclopentane in which one
[21,35]
66:34
ꢀ
C C bond length was frozen at 2.2 ꢂ.
(ꢀ0.71)
(ꢀ0.41)
(ꢀ1.19)
A
As proposed earlier, the origin of the cis selectivity in this
reaction may be traced to the presence of an adverse steric
interaction between H4 and H7b in the trans TS—they are
2.41 ꢂ apart—which is absent in the corresponding cis
TS.[21–23] In addition, trans-1TS experiences more torsional
64:36
ACHTUNGTRENNUNG(40:60)
(2.10)
(2.40)
(0.96)
[a] XrelTS =XcisꢀTSꢀXtransꢀTS (X=E, H, or G); a positive value indicates
ꢀ
strain about the C5 C6 bond than does cis-1TS; the dihe-
that the trans-TS is most stable. [b] At 1 bar pressure and 298.15 K.
TS
dral angles between the C5H and C6H bonds are approxi-
[c] Ratio (%cis)/ACHTNGUTERN(NUG %trans) calculated from Grel values at 298.15 K.
128
www.chemasianj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2009, 4, 126 – 134

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Cis/trans Selectivity in Intramolecular Diels–Alder Reactions
selectivity because the dihedral angle, q1, between the car-
bonyl group and the dienophile bond is the same (318) in
both TSs (Table 1). Likewise, the torsional angle about the
ꢀ
O6 C7 bond, q2, is comparable (ca. 1628) in both TSs.
Intriguingly, the only obvious strain difference that we
were able to identify is the presence of torsional strain in
ꢀ
cis-2TS; in this TS, the conformation about the C4 C5 bond
[36]
ꢀ
ꢀ
is unfavorably C H/C H eclipsed
with a 3.48 dihedral
ꢀ
ꢀ
angle, q3, between the C5 H5b and C4 H4 bonds. In con-
trast, the C4 C5 bond in cis-2TS adopts the favored C H/
ꢀ
ꢀ
[36]
ꢀ
C H staggered
conformation, with q3 ꢁ648 (Table 1).
Using pentadienyl formate 10 as a model for estimating tor-
sional strain in cis-2TS and trans-2TS (Figure 2), the stag-
Scheme 2. The cis and trans TSs for the IMDA reaction of 1 with the
CBS-QB3 (at 1908C) and experimental product distribution. Arrows
ꢀ
show the direction of twisting of the dienophile about the C4 C8 pivot.
Figure 2. Models used to estimate the difference in the torsional strain
ꢀ
about the C4 C5 bond in the cis and trans TSs of 2.
gered conformation in trans-2TS is estimated to be approxi-
mately 2.5 kJmolꢀ1 more stable than the eclipsed conforma-
tion in cis-2TS (B3LYP/6-31+G(d)). Note that this model
also takes into account repulsive interactions between C4
ꢀ
and C7 which are slightly larger in trans-2TS (C4 C7 sepa-
ꢀ
ration 2.56 ꢂ; C7 H4 separation 2.56 ꢂ), than in cis-2TS
ꢀ
ꢀ
(C4 C7 separation 2.61 ꢂ; C7 H4 separation 3.29 ꢂ).
The reason why the C4 C5 bond has an eclipsed confor-
mation in cis-2TS, whereas it is staggered in trans-1TS
arises from the resistance of an ester group to out-of-plane
twisting about the alkoxy bond. Achievement of a staggered
ꢀ
ꢀ
conformation about the C4 C5 bond in cis-2TS would re-
quire the TS dihedral angle q2 to be reduced from 1628 to
approximately 1008, and this is energetically too costly. In
the case of cis-1TS, the flexible trimethylene tether is able
ꢀ
to realize a staggered conformation about C4 C5 while, at
the same time, maintaining a strain-free conformation
ꢀ
within the tether. As to be expected, the staggered C4 C5
bond conformation is also found in the ketone cis-3TS (q3 =
408), in which O6 in cis-2TS is replaced by a methylene
group, thereby restoring greater tether flexibility (q2 ꢁ1028),
and also in the ether cis-4TS (q3 =358), in which the conju-
gating tether carbonyl group is absent.
Scheme 3. The cis and trans TSs for the IMDA reaction of 2 with the
CBS-QB3 (at 1808C) and experimental product distribution.
transannular strain between the C4 and C7 groups in trans-
2TS; H4 is 2.57 and 3.41 ꢂ from C7 and O10, respectively,
with the former distance being slightly larger than the sum
of the van der Waals radii of a hydrogen and a carbon atom
(Scheme 3).
Combining the overall energetic preference of approxi-
mately 1.9 kJmolꢀ1 for cis-2TS, relative to trans-2TS, with
ꢀ
the formerꢀs unfavorable C4 C5 torsional strain energy of
approximately 2.5 kJmolꢀ1 suggests the presence of an addi-
tional, stabilizing, factor in cis-2TS, amounting to approxi-
mately 4.4 kJmolꢀ1. We postulate that this factor is a stabi-
Possible differences in conjugation in the cis and trans
IMDA TSs for 2 cannot be responsible for the observed cis
Chem. Asian J. 2009, 4, 126 – 134
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
129