Synthesis of Cembranoid Analogues through Ring-Closing Metathesis of Terpenoid Precursors: A Challenge Regarding Ring-Size Selectivity

Article abstract of DOI:10.1002/chem.201502051

A systematic study on ring-closing metathesis with Grubbs II catalyst to cembranoid macrocycles is described. Acyclic terpenoids with a functional group X in the homoallylic position relative to an RCM active terminus and substituents R, R¹ directly attached to the other terminal double bond were prepared from geraniol derived trienes and fragments that are based on bromoalkenes and dimethyl malonate. Such terpenoids were suitable precursors, despite the presence of competing double bonds in their framework. The size of R and R¹ is crucial for successful macrocyclization. Whereas small alkyl substituents at the double bond directed the RCM towards six-membered ring formation, cross metathesis leading to dimers dominated for bulkier alkyl groups. A similar result was obtained for precursors without functional group X. In the case of unsymmetrically substituted terpenoid precursor (R=Et, R¹=Me) with homoallylic OTBS or OMe group, the RCM could be controlled towards formation of macrocyclic cembranoids, which were isolated with excellent E-selectivity. The role of the substituents was further studied by quantum chemical calculations of simplified model substrates. Based on these results a mechanistic rationale is proposed.

Full text of DOI:10.1002/chem.201502051

DOI: 10.1002/chem.201502051
Full Paper
&
Ring-Closing Metathesis
Synthesis of Cembranoid Analogues through Ring-Closing
Metathesis of Terpenoid Precursors: A Challenge Regarding Ring-
Size Selectivity
Tanja Heidt,[a] Angelika Baro,[a] Andreas Kçhn,[b] and Sabine Laschat*[a]
Dedicated to Professor Hans-Joachim Werner on the occasion of his 65th birthday
Abstract: A systematic study on ring-closing metathesis
with Grubbs II catalyst to cembranoid macrocycles is de-
scribed. Acyclic terpenoids with a functional group X in the
homoallylic position relative to an RCM active terminus and
substituents R, R1 directly attached to the other terminal
double bond were prepared from geraniol derived trienes
and fragments that are based on bromoalkenes and dimeth-
yl malonate. Such terpenoids were suitable precursors, de-
spite the presence of competing double bonds in their
framework. The size of R and R1 is crucial for successful mac-
rocyclization. Whereas small alkyl substituents at the double
bond directed the RCM towards six-membered ring forma-
tion, cross metathesis leading to dimers dominated for bulki-
er alkyl groups. A similar result was obtained for precursors
without functional group X. In the case of unsymmetrically
substituted terpenoid precursor (R=Et, R1 =Me) with homo-
allylic OTBS or OMe group, the RCM could be controlled to-
wards formation of macrocyclic cembranoids, which were
isolated with excellent E-selectivity. The role of the substitu-
ents was further studied by quantum chemical calculations
of simplified model substrates. Based on these results
a mechanistic rationale is proposed.
Introduction
Ring-closing metathesis (RCM) became an important tool in or-
ganic synthesis after the early discoveries by Schrock[1] and
Grubbs,[2] and the reaction is certainly one of the most valu-
able catalytic CÀC bond-forming reactions available, giving
access to a large variety of cycloalkenes. Many researchers
worldwide have contributed to the field, and their efforts have
led to improvements with respect to the scope of substrates,
catalyst design, stereoselectivity, optimization of preparative
procedures and mechanistic understanding.[3] In particular, nat-
ural product synthesis has profited immensely from the advan-
ces of RCM, which can now tolerate even sensitive precursors[4]
with complex substitution patterns. Macrocyclization by RCM
appears to be an attractive way to construct cembranoids,
a class of naturally occurring diterpenes produced by marine
organisms and plants. These compounds have a 14-membered
carbon ring system in common, which is decorated with a vari-
ety of oxygen functions such as alcohols, esters, epoxides, or
furans (Figure 1).[5,6] Their wide range of biological activities
Figure 1. Basic cembranoid skeleton 1[5e] and two typical members of this
class: flexilarin B (2) from Sinularia flexibilis[6a,7] and sarcophytonolide C (3)
from soft corals of the genus Sarcophyton.[5a,8]
makes cembranoids interesting as potential pharmacological
and agrochemical lead structures.
The use of RCM reactions to construct these macrocycles
provides a particular synthetic challenge because acyclic 1,w-
diene precursors carrying additional C=C double bonds might
lead to RCM products with different ring sizes or dimers[9] be-
cause of cross metathesis. Furthermore, C=C double bond iso-
merization gave varying ring sizes, as recently reported by
De Bo and Markó for medium-sized rings.[10] Although dimer
formation can be suppressed under high dilution conditions,[11]
ring-size selectivity remains an issue. In the case of polyenes,
different reactivities of the double bonds in the backbone de-
termined the preferred formation of the desired macrocycle.[12]
Furthermore, the presence of a polar functional group (e.g., an
ester) at a suitable distance from the reactive double bonds
was reported to be important for effective RCM towards mac-
[a] T. Heidt, Dr. A. Baro, Prof. S. Laschat
Institut für Organische Chemie, Universität Stuttgart
Pfaffenwaldring 55, 70569 Stuttgart (Germany)
E-mail: sabine.laschat@oc.uni-stuttgart.de
[b] Prof. A. Kçhn
Institut für Theoretische Chemie, Universität Stuttgart
Pfaffenwaldring 55, 70569 Stuttgart (Germany)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201502051.
Chem. Eur. J. 2015, 21, 12396 – 12404
12396
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
rocycles.[13] However, successful ring-closing metathesis of sub-
strates containing competing double bonds of similar reactivity
requires alternative RCM concepts. For example, the macrocyc-
lization step in the total synthesis of oximidine III was realized
by relay ring-closing metathesis (RRCM),[14] which was devel-
oped as a strategy to direct metal movement throughout
metathesis sequences.[15] However, RRCM does not always pro-
vide a solution. In the syntheses of archazolid[16] and (À)-dic-
tyostatin[17] relay-loss and backbiting was observed. Further
routes to favor the construction of macrocyclic cores over
competing six-membered ring formation utilized well-defined
substrates with additional substituents at the alkene termi-
nus[18] or the combination of different reactions for macrocycli-
zation, for example, lactonization followed by transannular
RCM, as recently applied in the total synthesis of (+)-sarcophy-
tonolide C.[19]
Scheme 1. Retrosynthetic pathway to cembranoid analogues 5. Numbering
in 5 and 4 for NMR assignment.
During our ongoing work on cyclic terpenes, we envisaged
the use of ring-closing metathesis to create cembranoid ana-
logues in a more direct fashion. Taking previously reported re-
sults into account, we studied RCM of precursors 4 with re-
spect to an additional substituent X in the vicinity of the termi-
nal C=C double bond and substituents R and R1 directly at-
tached to the double bond, as illustrated schematically in
Figure 2. We anticipated that chelation as well as steric effects
should guide the RCM to efficient macrocycle formation. The
results are reported below.
Scheme 2. Preparation of steering fragments 8. E/Z ratios were determined
1
by H NMR spectroscopic analysis.
Figure 2. Schematic illustration of RCM of alkenes 4 with competing double
bonds as well as additional substituent X and different substituents R, R1.
Results and Discussion
Synthesis of precursors 4
Scheme 3. Preparation of substituted 11-bromo-5,9-dimethylundeca-1,5,9-tri-
enes 7a,b. Reagents and conditions: a) Ac2O, DMAP, CH2Cl2, RT, 8 h, 99%;
b) Grubbs II catalyst (5 mol%), methacrolein, CH2Cl2, 408C, 7 h, 82%;
c) H2C=CHCH2MgCl, THF, À788C, 19 h, 94%; d) TBSCl, NEt3, DMAP, DMF, RT,
11 h, 94%; e) p-TsOH, MeOH, À208C, 2 h, 91%; f) 1. NEt3, MsCl, CH2Cl2,
À408C, 1.5 h; 2. LiBr, THF, À208C, 22 h, 97%.
The retrosynthetic concept that was used to plan the construc-
tion of the desired cembranoid macrocycles 5 is shown in
Scheme 1. The envisioned acyclic tetraene precursors 4 with
different substitution patterns in the homoallylic position to
the RCM entering terminus (X) and the terminal C=C bond (R,
R1) were traced back to building blocks 7 and 8. The latter was
derived from bromoalkenes 10 and dimethyl malonate 11. Ter-
penoid bromides 7 were accessible starting from geraniol (9).
Homoallylic bromoalkenes 10a–g, which were accessible by
using reported procedures,[20–23] were reacted with dimethyl
malonate (11) to afford the steering fragments 8a–g in 55–
95% yield (Scheme 2).
The synthesis of terpenoid building blocks 7 started from
commercially available geraniol (9) (Scheme 3). Geraniol (9)
was acetylated to 12,[24] which was submitted to a Ru-catalyzed
cross metathesis with methacrolein in the presence of second
generation Grubbs catalyst (Grubbs II). The resulting com-
Chem. Eur. J. 2015, 21, 12396 – 12404
www.chemeurj.org
12397
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
pound 13 was treated with allylmagnesium chloride, providing
diol 14. The latter was either brominated under Appel condi-
tions[24,25] to give derivative 7b or submitted to OH protection
with TBSCl to give 15. Selective deprotection of the primary
silyl ether with TsOH to give 16 followed by a two-step, one-
pot bromination yielded bromoalkene 7a (Scheme 3).
Alkylation of fragments 8a–g with bromo terpenoids 7a,b
and unsubstituted 11-bromo-5,9-dimethylundeca-1,5,9-triene
(7c; X=H)[11a,26] provided the acyclic tetraene precursors 4a–
l in 65–98% yield. The unsymmetrically substituted tetraenes
4m–p were finally accessible by deprotection and derivatiza-
tion of the respective precursor tetraenes 4c and 4j,m,l, re-
spectively (Scheme 4).
Ring-closing metathesis of terpenoid substrates 4
Tetraenes 4 were then submitted to RCM in the presence of
Grubbs II catalyst (5 mol%) in CH2Cl2 heated to reflux under
high dilution conditions (2 mm of 4). The crude product mix-
tures were purified by chromatography and then analyzed by
NMR spectroscopy. The results are summarized in Table 1. It
should be noted that further byproducts were formed by
cross-metathesis reaction; these byproducts are described in
detail in the Supporting Information.
Scheme 4. Preparation of RCM precursors 4. E/Z ratios were determined by
1H NMR spectroscopic analysis.
We first studied the role of terminal substituents R and R1 in
the series of OTBS-substituted RCM precursors 4a–g. We
found that treatment of tetraene 4a, with one terminal (E)-
methyl group, with 5 mol% Grubbs II catalyst afforded the de-
sired macrocycle 5a in 25% yield (E/Z ratio 87:13), albeit with
cyclohexene 6 as the major product (62%) (Table 1, entry 1). To
shift the olefin metathesis in favor of macrocyclization, the
steric bulk at the alkenyl terminus was increased by the inclu-
sion of a second methyl group. The RCM of 4b afforded mac-
rocycle 5a in slightly improved yield of 29% (E/Z ratio 99:1)
but was still accompanied by formation of 6 (entry 2). When
diethyl- and diisopropyl-substituted substrates 4 f and 4g, re-
Table 1. Ring-closing metathesis (RCM) of acyclic tetraenes 4 to cembranoid macrocycles 5.
Entry
4
R
R1
X
5
Yield [%][a]
E/Z ratio[b]
Yield of 6 [%][a]
Dimer 17
Yield [%][a]
E/Z ratio[b]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
4a
4b
4c
4d
4e
4f
4g
4h
4i
4k
4m
4n
4o
4p
Me
Me
Et
iPr
tBu
Et
iPr
Me
Et
Me
Et
H
OTBS
OTBS
OTBS
OTBS
OTBS
OTBS
OTBS
H
H
OH
OH
OAc
OMe
OMe
5a
5a
5a
5a
5a
5a
5a
5b
5b
5c
5c
5d
5e
5e
25
29
61
30
–
87:13
99:1
99:1
99:1
–
62
36
12
–
–
–
17a
17b
17c
17d
17e
17f
17g
17h
17i
17k
17m
17n
17o
17p
–
–
7
64
99
51
60
–
24
–
56
–
21
61
–
–
–
Me
Me
Me
Me
Et
iPr
Me
Me
Me
Me
H
57:43
60:40
64:36
66:34
–
12
–
99:1
–
–
18
29
24
23
–
86:14
89:11
99:1
99:1
–
37
27
–
–
79
7
–
–
51:49
–
66:34
70:30
Me
Et
iPr
Me
Me
63
32
99:1
99:1
–
[a] Isolated yields based on conversion/recovered tetraene 4. [b] E/Z ratios were determined by 13C NMR spectroscopic analysis.
Chem. Eur. J. 2015, 21, 12396 – 12404
www.chemeurj.org
12398
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
spectively, were submitted to olefin metathesis reaction, the
situation grew even less favorable. Macrocyclization to triene
core 5a was suppressed and cross metathesis of 4 f and 4g
dominated, resulting in the formation of dimers 17 f and 17g
in 51 and 60% yield, respectively (entries 6 and 7). Given that
dimethyl-substituted tetraene 4b tended to favor six-mem-
bered ring formation and diethyl substituted 4 f gave mainly
cross metathesis dimer 17 f, we anticipated that a tetraene
precursor with different alkyl substituents R and R1 might force
the macrocycle formation. Indeed, the substitution pattern in
acyclic tetraene 4c steers the metathesis pathway in the direc-
tion of macrocyclization, affording cembranoid analogue 5a in
61% yield (E/Z ratio 99:1). Cross metathesis and competing six-
membered ring formation was clearly decreased (entry 3).
However, further increasing the size of substituent R in 4d and
4e directed the reaction to cross metathesis (entries 4 and 5).
Whereas 4d still gave macrocycle 5a as a minor product, 4e
reacted exclusively to give dimer 17e. Clearly, there is a delicate
steric balance between macrocycle RCM, six-membered ring
RCM and cross metathesis.
the steric hindrance at the alkene terminus drives the reaction
to cross metathesis.
The results presented in Table 1 demonstrate that the ring-
size selectivity depends on a subtle balance of steric hindrance
at one alkenyl chain guiding the ruthenium carbene to the
other alkene terminus. In contrast, we did not find a general
trend of homoallylic substituent X. The steric factor in the RCM
of tetraenes 4 agreed well with previous results that showed
that a methyl group on the more reactive diene of a bis-diene
substrate directed RCM to macrocyclization.[18] However, to our
knowledge, homoallylic substituents have not been reported
so far, and the influence of allylic substituents is less clearcut
in the literature. For example, allylic hydroxy groups were
found to exert an activating effect on the RCM rate constants
for cyclopentene derivatives, whereas allylic methoxy groups
retarded the RCM.[27] In the total synthesis of terpestacin, the
free allylic alcohol function is essential to produce the 15-
membered carbocycle,[28] the corresponding TBS ether did not
undergo RCM to the terpestacin intermediate.[29] It should be
mentioned that neither competing five- nor six-membered
ring formation occurred. In the total synthesis of pochonin E
and F, allylic OTBS substitution surprisingly gave lower yields
of the metathesis than in the case of substituent-free sub-
strates.[30] As previously reported, the stereochemistry of the al-
lylic methoxy group in complex migrastatin precursors is criti-
cal for the success of the ring-closing metathesis, which can
either be promoted or completely blocked.[31] The involvement
of chelate complexes formed by substrates carrying functional
groups such as ester, ketone, ether, or urethane in the catalytic
cycle have been reported by Fürstner and Langemann.[13d]
However, a prerequisite for cyclization is a suitable distance be-
tween these groups from the reacting Ru carbene terminus.
Mechanistic studies on the Grubbs II catalyst by using low-tem-
perature NMR spectroscopy revealed for a,a’-bisallylmalonates
that the intermediate ruthenacyclobutanes are much lower in
energy than the corresponding ruthenium carbene com-
plexes,[32] and a detailed DFT study[33] rationalized the beneficial
effect of allylic alcohol substrates on ring opening/cross meta-
thesis (ROCM) reaction with Grubbs II catalyst by hydrogen
bonding and minimization of destabilizing donor–donor inter-
action between the two chloride ligands. It should be noted
that Cossy previously discovered a beneficial effect of TBDPS-
protected homoallylic alcohols on the regioselectivity of cross-
metathesis reaction catalyzed by second generation Hoveyda–
Grubbs catalyst.[34]
To check whether the 14- versus 6-membered ring selectivity
was caused by alkyl substitution in the steering subunit or by
a combined influence of both alkenyl substituents and homo-
allylic substituent X, tetraene precursors 4h and 4i (X=H)
were treated with Grubbs II catalyst under identical conditions.
In the case of 4h, six-membered ring formation was dominant
and macrocycle 5b was isolated in only 18% yield. For sub-
strate 4i, macrocyclization and cyclohexene formation compet-
ed, giving the respective products 5b and 6 in comparable
yield (Table 1, entries 8 and 9). Additionally, cross metathesis to
give dimer 17i occurred. For both precursors 4h and 4i, de-
creased E/Z selectivity for the RCM product was observed.
The type of O-protecting group turned out to play a role, as
shown for substrates 4a and 4n. In contrast to 4a, with an
OTBS group, the homoallylic acetyloxy functionalized substrate
4n did not undergo macrocyclization to 5d but afforded the
competing cyclohexene 6 in 79% yield (Table 1, entries 1 and
12). To study the influence of size of substituent X, metathesis
precursors 4o and 4p, with the small OMe group, were treated
with Grubbs II catalyst under the standard conditions. The
RCM of 4o gave macrocycle 5e in 63% yield and E/Z ratio of
99:1, albeit with competing cyclohexene 6 and dimer 17o (7
and 21%, respectively) (entry 13). Metathesis of 4p yielded
macrocycle 5e in only 32%, but 61% of dimer 17p (entry 14).
These results agree well with those obtained with OTBS-substi-
tuted counterparts 4c and 4d (entries 3 and 4). Although
OTBS and OMe clearly differ in steric bulk, both groups be-
haved similarly in the metathesis.
Quantum chemical calculations and mechanistic proposal
As compared to 4c and 4o, with OTBS and OMe group, re-
spectively, olefin metathesis of the corresponding hydroxy
functionalized tetraene 4m resulted in a reversed RCM/cross
metathesis ratio, giving dimer 17m as the major product
(Table 1, entries 3, 11, and 13). Under Grubbs II catalysis, macro-
cyclization of OH-substituted 4k to 5c occurred (entry 10), but
the RCM was accompanied by various side reactions (see the
Supporting Information). Thus, the hydroxy group seems to
disfavor RCM to the macrocycle and six-membered ring, and
To shed light on the role of the different residues R, R1, and X
in steering the RCM towards the desired macrocycle, we car-
ried out a number of quantum chemical model calculations
using density functional theory. We used a simplified model for
the substrate, either 19 to model RCM starting on one end of
the substrate (in the vicinity of the X residue), or 19’ to model
an RCM starting on the opposite end, where the R and R1 resi-
dues are located (Scheme 5).
Chem. Eur. J. 2015, 21, 12396 – 12404
www.chemeurj.org
12399
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
catalytic cycle, we believe that the present computational data
is sufficient to augment the experimental findings and to put
forward—in concert with experiment—a mechanistic proposal.
The computed energy levels are presented in Figure 3. Ac-
cording to the calculations, the initial p-complexes that can be
formed with either terminal C=C double bond of the substrate
clearly differ in energy. In particular, sterically demanding sub-
stitution (R=iPr; 20’d) strongly decreases the binding energy
and very likely suppresses this pathway. With small substitu-
ents in this position (R=Me, R1 =H; 20’a), the energetic differ-
ence to attack at the other vinyl terminus becomes small, the
other considered substitution patterns have intermediate ener-
gies for the p-complex. Hence, the computations clearly sup-
port the experimental finding that substitution in the R and R1
position leads to a subtle triggering of different pathways,
leading to different products (see below).
Scheme 5. Simplified model substrates 19, 19’, and carbene 18 as well as
calculated structures 20–22 and 20’ thereof: p-complexes 20 and 20’, metal-
lacyclobutane 21 as well as carbene 22.
Figure 3. Energy level diagram (in kJmolÀ1) based on quantum chemical cal-
culations (PBE+D/def2-TZVPP, solvent effects treated by COSMO approach,
see also Tables S2, S3). The molecular structures are defined in Scheme 5.
The full Grubbs II catalyst was used in the model calcula-
tions, for which we placed a dimethyl-carbene group coordi-
nated to Ru, which models the resting state of the catalyst for
the case R=R1 =Me (the phenyl-carbene group of the original
catalyst will only be present in the initial catalysis cycle). Coor-
dination of the free site of the resting state 18 and the car-
bene intermediate 22 by one explicit solvent molecule (CH2Cl2)
was found to be energetically favorable. However, this finding
is not of particular importance for the following, except for de-
fining the energetic zero point. We optimized the Ru complex
with p-coordinated substrate 20 (attack on one side of the full
substrate 4) and 20’ (attack on the opposite side), the metalla-
cyclobutane intermediate 21 and the carbene intermediate 22
after cycloreversion. In addition, searches were made for O-co-
ordinated complexes (for X=OMe and OTBS) and chelate com-
plexes (both s- and p-coordination). Care was taken to find
comparable conformations for each of the cases X=H, OMe,
and OTBS and the different cases considered for R/R1. These
were, in each case, the most stable among the considered con-
formations. However, it must be stated that a full search of the
conformational space was not attempted; nevertheless, we be-
lieve that the conformations presented here are sufficient for
our conclusions. Furthermore, we did not attempt to calculate
transition structures, but we assume that the changes in rela-
tive energies of the intermediates are sufficient to estimate the
effect of the different substituents. Whereas a more rigorous
computational treatment will be necessary to fully model the
Further calculations were carried out to elucidate the effect
of X substitution. For p-complexes 20, metallacyclobutanes 21,
and post-cycloreversion intermediates 22, the energetic esti-
mates for all three residues X are energetically very similar; the
results obtained at this computational level therefore pre-
cludes any reliable interpretation of the small differences that
are apparent in Figure 3. Only for carbene 22c, with the OTBS
ligand, is a significant energy decrease apparent, even at the
highest computational level considered here. We also consid-
ered alternative interaction patterns with the catalyst. Coordi-
nation through oxygen is only favorable for X=OMe (binding
energy À4 kJmolÀ1), but even then it is not competitive with
p-complexation. Likewise, our calculations do not suggest that
a chelate with both s- and p-coordination is competitive with
the simple p-complex.
The calculations show that the presence of the large OTBS
ligand has no adverse effect on the efficiency of the catalysis
cycle. As can be seen in Figure 4, the OTBS ligand remains in
the periphery. However, the calculated energies do not explain
the observed gain of regioselectivity for substrates with the
OTBS ligand (which enhances the yield of the macrocycle in
the experiments). Nevertheless, a structural (and entropic) ar-
gument may be deduced from the calculations. The structure
of the carbene intermediate after cycloreversion 22c (Figure 4)
suggests that the bulky OTBS ligand induces and stabilizes the
wrapped structure of the substrate, which both directs the ter-
Chem. Eur. J. 2015, 21, 12396 – 12404
www.chemeurj.org
12400
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 4. Modeled p-complexes 20a–c (left column) and carbene intermedi-
ates 22a–c after cycloreversion (right column) based on DFT calculations
(PBE-D3/def2-TZVPP). A solvent molecule CH2Cl2 is coordinated to the car-
bene intermediates.
minal C=C double bond (which is not present in our simplified
model) towards the center of the catalyst and inhibits the
attack by a second substrate molecule.
Based on these studies, we propose the mechanistic ration-
ale that is outlined in Scheme 6. As supported by the above
quantum chemical results, the steric effects can be attributed
to two different intermediate Ru carbene species 23 and 24
upon treatment of tetraene 4 with Grubbs II catalyst. For small
substituents at the double bond, that is, R=Me and R1 =H,
Me, complex 23 undergoes immediate formation of the six-
membered ring 6. With increasing steric bulk of R and R1, the
ratio 23/24 is shifted towards complex 24. The computations
clearly show that large substituents such as R=iPr clearly dis-
favor this pathway. For smaller substituents, computations still
predict 23 (as modeled by 20’) to be less stable than 24 (mod-
eled by 20) for all cases. This, however, may be a shortcoming
of the model, which uses only fragments of the real substrate;
on the other hand, even in the case of a slight energetic favor-
ing of 24, a fast formation of the six-membered ring 6 from 23
may drive the observed product distribution in this direction.
Experiments show that substituents R=Et, R1 =Me have suffi-
cient bulk to favor the pathway via complex 24 over six-mem-
bered ring formation via 23, leading, via 25, 26, and carbene
intermediate 27, to the desired 14-membered ring 5. However,
due to the equilibrium between carbene complex 27 and car-
bene intermediate 28, in which a second substrate molecule 4
is p-bonded, cross metathesis can occur to provide dimer 17.
For bulkier substituents R=Et, iPr, tBu and R1 =Me, Et, iPr, car-
bene intermediate 28 and thus cross metathesis even domi-
nates, giving mainly dimer 17. It should be noted that complex
Scheme 6. Proposed mechanistic rationale considering the quantum chemi-
cal model calculations. Postulated intermediates 24–26 are tracked back to
the calculated structures 20–22.
24 may also lead to further byproducts through cross-metathe-
sis reactions (see the Supporting Information, Scheme S1). Ac-
cording to our mechanistic proposal, homoallylic substitution,
however, seems to have less effect on olefin metathesis, and
this is supported by the DFT calculations.
Conclusions
We could demonstrate that even terpenoid substrates 4, with
four competing double bonds, may be successfully submitted
to ring-closing metathesis to afford cembranoid macrocycles 5
when steric effects are well balanced. In this case, macrocycli-
zation was clearly favored and competing six-membered ring
formation of cyclohexene 6 and cross metathesis to dimers 17
were decreased. Optimum yields of macrocyclic products 5
were realized for a combination R=Et, R1 =Me at one alkenyl
terminus, regardless of the homoallylic OTBS or OMe group at
the opposite alkenyl terminus. Quantum chemical model calcu-
lations gave some insight into the various substituent effects.
Future work must demonstrate whether this methodology can
be used to control the ring-size selectivity in RCM reactions of
Chem. Eur. J. 2015, 21, 12396 – 12404
www.chemeurj.org
12401
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
non-terpenoid substrates with a more polar substitution pat-
tern.
Dimethyl 9-{[tert-butyl(dimethyl)silyl]oxy}-4,8-dimethylcy-
clotetradeca-3,7,11-triene-1,1-dicarboxylate (5a): 1H NMR
(500 MHz, CDCl3): d=À0.02 (s, 3H; Si-CH3), 0.03 (s, 3H; Si-CH3),
0.87 (s, 9H; Si-C(CH3)3), 1.52 (s, 3H; 16-H), 1.55 (s, 3H; 15-H),
1.61–1.73 (m, 3H; 14-Ha or 14-Hb, 13-H), 2.04–2.15 (m, 5H; 5-Ha
or 5-Hb, 6-Ha or 6-Hb, 10-H, 14-Ha or 14-Hb), 2.20–2.28 (m, 1H;
5-Ha or 5-Hb), 2.31–2.41 (m, 1H; 6-Ha or 6-Hb), 2.52–2.58 (m,
1H; 2-Ha or 2-Hb), 2.83–2.90 (m, 1H; 2-Ha or 2-Hb), 3.71 (s, 3H;
CO2CH3), 3.73 (s, 3H; CO2CH3), 3.82–3.86 (m, 1H; 9-H), 4.77–
4.82 (m, 1H; 3-H), 4.91–4.95 (m, 1H; 7-H), 4.96–5.02 (m, 1H;
11-H), 5.34–5.41 ppm (m, 1H; 12-H); 13C NMR (125 MHz, CDCl3):
d=À4.90 (Si-CH3), À4.7 (Si-CH3), 10.5 (C-16), 14.9 (C-15), 18.2
(Si-C(CH3)3), 24.4 (C-6), 25.9 (Si-C(CH3)3), 26.4 (C-13), 29.8 (C-2),
32.0 (C-14), 37.4 (C-10), 38.8 (C-5), 52.5 (CO2CH3), 52.6 (CO2CH3),
56.5 (C-1), 79.9 (C-9), 118.9 (C-3), 126.2 (C-7), 127.6 (C-11), 130.7
(C-12), 135.7 (C-8), 137.8 (C-4), 171.8 (CO2CH3), 172.3 ppm
(CO2CH3); FTIR (ATR): n˜ =2953 (m), 2927 (m), 2885 (w), 2855
(m), 2360 (w), 1736 (vs), 1667 (w), 1472 (w), 1453 (m), 1433 (m),
1388 (w), 1361 (w), 1316 (w), 1295 (m), 1271 (m), 1252 (m),
1235 (m), 1218 (m), 1200 (m), 1172 (m), 1108 (w), 1065 (s), 1060
(s), 1005 (w), 960 (w), 890 (w), 863 (m), 835 (s), 811 (w), 776
(m), 670 (w), 620 cmÀ1 (w); MS (ESI): m/z (%): 487.3 (100)
[M+Na]+, 393.3 (8), 350.2 (1), 333.2 (8) [MÀOSi(CH3)2C(CH3)3],
301.2 (3), 273.2 (2), 251.1 (2); HRMS (ESI): m/z: calcd for
C26H44O5SiNa: 487.2850 [M+Na]+; found: 487.2842.
Experimental Section
Materials and methods: NMR spectra were recorded with
a Bruker Avance 300 or Avance 500 spectrometer with TMS as
internal standard. In the case of E/Z isomers, data of the major
isomer are given. IR spectra were recorded with a Bruker
Vektor 22 spectrometer equipped with an MKII golden gate
single reflection diamond ATR system. Mass spectra were re-
corded with a Varian MAT 711 spectrometer (EI, 70 eV) and
a Bruker Daltonics micrOTOF Q (ESI) with nitrogen as carrier
gas. Control of reactions and purity was performed with a Hew-
lett–Packard HP 6890 equipped with a HP-5 column (30 m”
0.32 mm). Chromatography was performed on silica gel (grain
size 40–63 mm, Fluka). All reactions were performed under ni-
trogen in oven-dried glassware. All reagents were used as pur-
chased unless otherwise noted. Solvents for chromatography
were distilled prior to use. Tetrahydrofuran (THF) was distilled
from potassium/benzophenone, Et2O from sodium/benzophe-
none, CH2Cl2, Et3N and N,N-dimethylformamide (DMF) from
CaH2, and MeOH from magnesium. The reactions were moni-
tored by TLC (Macherey–Nagel silica gel 60 F254 plates) and vi-
sualized with an ethanolic solution of p-anisaldehyde and sul-
furic acid or an aqueous solution of potassium permanganate,
potassium carbonate and NaOH.
Dimethyl 4,8-dimethylcyclotetradeca-3,7,11-triene-1,1-di-
carboxylate (5b): 1H NMR (500 MHz, CDCl3): d=1.53 (s, 3H;
16-H), 1.56 (s, 3H; 15-H), 1.64–1.71 (m, 2H; 13-H), 1.90–1.96 (m,
2H; 14-H), 2.00–2.07 (m, 4H; 9-H, 10-H), 2.13–2.22 (m, 4H; 5-H,
6-H), 2.70 (d, J=7.2 Hz, 2H; 2-H), 3.72 (s, 6H; 2”CO2CH3), 4.81
(dd, J=7.2, 7.2 Hz, 1H; 3-H), 4.84 (dd, J=5.0, 5.0 Hz, 1H; 7-H),
5.19 (dt, J=15.3, 6.8 Hz, 1H; 11-H), 5.37 ppm (dd, J=15.3,
6.3 Hz, 1H; 12-H); 13C NMR (125 MHz, CDCl3): d=15.03 (C-16),
15.04 (C-15), 24.9 (C-6), 25.9 (C-13), 28.2 (C-9), 29.9 (C-2), 31.5
(C-14), 38.9 (C-5), 40.8 (C-10), 52.5 (2”CO2CH3), 56.6 (C-1), 118.8
(C-3), 126.1 (C-7), 129.0 (C-12), 130.8 (C-11), 132.8 (C-8), 138.0
(C-4), 172.1 ppm (2”CO2CH3). FTIR (ATR): n˜ =2923 (m), 2852
(m), 1733 (vs), 1434 (m), 1382 (w), 1269 (s), 1198 (s), 1169 (s),
1076 (m), 959 (w), 913 (w), 887 (w), 813 (m), 734 (m), 578 (w),
521 cmÀ1 (w); MS (ESI): m/z (%): 357.2 (100) [M+Na]+; HRMS
(ESI): m/z: calcd for C20H30O4Na: 357.2036 [M+Na]+; found:
357.2036.
Computational details: Density functional theory with the
PBE and PBE0 functionals[35] as implemented in the TURBO-
MOLE program package[36] was used. The multipole accelerated
resolution of the identity (RI) approximation was employed
throughout.[37] The basis sets def2-SVP and def2-TZVPP[38] were
used together with the corresponding auxiliary basis for the RI
approximation.[39] Empirical dispersion corrections with damp-
ing as defined previously[40] were added (PBE+D3). Full struc-
ture optimizations were carried out at each level of theory. Sol-
vation effects were tested by single-point calculations at the
PBE/def2-TZVPP level, using the COSMO approach[41] with e=9
(which approximately equals the dielectric constant of CH2Cl2).
Starting from PBE+D3/def2-SV(P) calculations, we checked the
influence of including exact exchange (hybrid functional PBE0),
larger basis set (def2-TZVPP) and solvation (continuum solva-
tion model COSMO). Details of the results are given in the Sup-
porting Information; in the text, we base our discussion on the
PBE+D3/def2-TZVPP results including the COSMO correction.
General procedure for the RCM with Grubbs II catalyst:
Grubbs II catalyst [(1,3-bis(2,4,6-trimethylphenyl)-2-imidazolidi-
nylidene)dichloro(phenylmethylene)(tricyclohexylphosphine)ru-
thenium] (2.54 mg, 2.99 mmol) was dissolved in CH2Cl2 (12 mL)
in a Schlenk flask under N2 atmosphere. A solution of the re-
spective diene 4 (2 mm in CH2Cl2, 3 mL, 29.9 mmol) was added
dropwise and the reaction mixture was heated for 72 h at
408C. The catalyst was filtered off through SiO2 and the filtrate
was concentrated under reduced pressure. The residue was
purified by chromatography on SiO2 to give macrocyclic prod-
ucts 5 as light-yellow oils.
Dimethyl
9-hydroxy-4,8-dimethylcyclotetradeca-3,7,11-
triene-1,1-dicarboxylate (5c): 1H NMR (500 MHz, CDCl3): d=
1.56 (s, 3H; 15-H), 1.58 (s, 3H; 16-H), 1.65–1.76 (m, 2H; 13-H),
1.79–1.98 (m, 2H; 14-H), 2.10–2.21 (m, 3H; 5-Ha or 5-Hb, 6-Ha or
6-Hb, 10-Ha or 10-Hb), 2.21–2.27 (m, 1H; 5-Ha or 5-Hb), 2.28–2.36
(m, 2H; 6-Ha or 6-Hb, 10-Ha or 10-Hb), 2.64 (dd, J=15.8, 6.0 Hz,
1H; 2-Ha or 2-Hb), 2.78 (dd, J=15.8, 8.4 Hz, 1H; 2-Ha or 2-Hb),
3.72 (s, 3H; CO2CH3), 3.73 (s, 3H; CO2CH3), 3.94–4.01 (m, 1H; 9-
H), 4.80 (dd, J=8.4, 6.0 Hz, 1H; 3-H), 5.02 (dt, J=15.3, 7.8 Hz,
1H; 11-H), 5.04–5.08 (m, 1H; 7-H), 5.47 ppm (dt, J=15.3,
6.7 Hz, 1H; 12-H); 13C NMR (125 MHz, CDCl3): d=11.2 (C-16),
14.9 (C-15), 24.4 (C-6), 26.5 (C-13), 29.9 (C-2), 32.1 (C-14), 35.8
(C-10), 38.8 (C-5), 52.5 (2”CO2CH3), 52.6 (CO2CH3), 56.5 (C-1),
77.8 (C-9), 119.1 (C-3), 126.5 (C-11), 126.8 (C-7), 131.9 (C-12),
135.1 (C-8), 137.6 (C-4), 171.8 (CO2CH3), 172.0 ppm (2”CO2CH3);
Chem. Eur. J. 2015, 21, 12396 – 12404
www.chemeurj.org
12402
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[5] Reviews on cembranoids: a) L.-F. Liang, Y.-W. Guo, Chem. Biodiversity
2013, 10, 2161–2196; b) J. R. Hanson, Nat. Prod. Rep. 2012, 29, 890–
898; c) B. Yang, X.-F. Zhou, X.-P. Lin, J. Liu, Y. Peng, X.-W. Yang, Y. Liu,
Curr. Org. Chem. 2012, 16, 1512–1539; d) Y. Li, G. Pattenden, Nat. Prod.
Rep. 2011, 28, 429–440; e) P. A. Ferchmin, O. R. Pagµn, H. Ulrich, A. C.
FTIR (ATR): n˜ =2992 (w), 2953 (m), 2918 (m), 2853 (w), 1752 (s),
1732 (vs), 1667 (w), 1453 (m), 1433 (m), 1387 (w), 1362 (w),
1315 (w), 1295 (m), 1271 (m), 1238 (m), 1219 (m), 1200 (s),
1172 (m), 1102 (w), 1075 (w), 1038 (m), 1019 (m), 1001 (w), 974
(w), 960 (w), 888 (w), 844 (w), 731 cmÀ1 (w); MS (ESI): m/z (%):
373.2 (100) [M+Na]+, 333.2 (2), 301.2 (2); HRMS (ESI): m/z:
calcd for C20H30O5Na: 373.1985 [M+Na]+; found: 373.1992.
´
Szeto, R. M. Hann, V. A. Eterovic, Toxicon 2009, 54, 1174–1182; f) H.
Gross, G. M. Kçnig, Phytochem. Rev. 2006, 5, 115–141.
[6] Reviews on furanocembranoids: a) Y. Li, G. Pattenden, Nat. Prod. Rep.
2011, 28, 1269–1310; b) P. A. Roethle, D. Trauner, Nat. Prod. Rep. 2008,
25, 298–317.
Dimethyl
9-methoxy-4,8-dimethylcyclotetradeca-3,7,11-
triene-1,1-dicarboxylate (5e): 1H NMR (500 MHz, CDCl3): d=
1.51 (s, 3H; 16-H), 1.57 (s, 3H; 15-H), 1.61–1.72 (m, 3H; 13-Ha
or 13-Hb, 14-H), 2.04–2.17 (m, 4H; 5-Ha or 5-Hb, 6-Ha or 6-Hb,
10-Ha or 10-Hb, 13-Ha or 13-Hb), 2.21–2.31 (m, 2H; 5-Ha or 5-Hb,
10-Ha or 10-Hb), 2.40–2.50 (m, 1H; 6-Ha or 6-Hb), 2.50–2.58 (m,
1H; 2-Ha or 2-Hb), 2.88 (dd, J=15.7, 10.5 Hz, 1H; 2-Ha or 2-Hb),
3.16 (s, 3H; OCH3), 3.36 (dd, J=10.7, 5.1 Hz, 1H; 9-H), 3.71 (s,
3H; CO2CH3), 3.73 (s, 3H; CO2CH3), 4.77–4.84 (m, 1H; 3-H),
4.96–5.04 (m, 1H; 11-H), 5.05–5.10 (m, 1H; 7-H), 5.36–5.47 ppm
(m, 1H; 12-H); 13C NMR (125 MHz, CDCl3): d=9.9 (C-16), 15.0
(C-15), 24.5 (C-6), 26.3 (C-14), 29.8 (C-2), 31.8 (C-13), 34.4 (C-10),
38.7 (C-5), 52.5 (CO2CH3), 52.6 (CO2CH3), 55.5 (OCH3), 56.5 (C-1),
88.7 (C-9), 119.0 (C-3), 127.1 (C-11), 129.8 (C-7), 131.1 (C-12),
132.8 (C-8), 137.6 (C-4), 171.7 (CO2CH3), 172.2 ppm (CO2CH3);
FTIR (ATR): n˜ =2924 (m), 2854 (m), 1732 (vs), 1435 (m), 1387
(w), 1364 (w), 1294 (s), 1269 (s), 1235 (s), 1200 (s), 1171 (s),
1097 (s), 887 (s), 835 (w), 696 cmÀ1 (w); MS (ESI): m/z (%): 387.2
(100) [M+Na]+, 325.3 (1), 297.2 (1), 251.2 (41), 213.2 (1), 183.1
(22), 157.1 (1), 109.1 (4); HRMS (ESI): m/z: calcd for C21H32O5Na:
387.2142 [M+Na]+; found: 387.2146.
[7] Y.-S. Lin, C.-H. Chen, C.-C. Liaw, Y.-C. Chen, Y.-H. Kuo, Y.-C. Shen, Tetrahe-
dron 2009, 65, 9157–9164.
[8] R. Jia, Y.-W. Guo, E. Mollo, G. Cimino, Helv. Chim. Acta 2005, 88, 1028–
1033.
[9] a) Z. Cai, N. Yongpruksa, M. Harmata, Org. Lett. 2012, 14, 1661–1663;
b) R. Kadyrov, Chem. Eur. J. 2013, 19, 1002–1012.
[10] G. De Bo, I. E. Markó, Eur. J. Org. Chem. 2011, 1859–1869.
[11] a) S. Yildizhan, J. van Loon, A. Sramkova, M. Ayasse, C. Arsene, C. ten
Broeke, S. Schulz, ChemBioChem 2009, 10, 1666–1677; b) L. F. Tietze,
C. C. Brazel, S. Hçlsken, J. Magull, A. Ringe, Angew. Chem. Int. Ed. 2008,
47, 5246–5249; Angew. Chem. 2008, 120, 5324–5327.
´
[12] A. Fürstner, C. Nevado, M. Waser, M. Tremblay, C. Chevrier, F. Teply, C.
A¨ıssa, E. Moulin, O. Müller, J. Am. Chem. Soc. 2007, 129, 9150–9161.
[13] a) A. Fürstner, O. R. Thiel, C. W. Lehmann, Organometallics 2002, 21,
331–335; b) A. Fürstner, F. Jeanjean, P. Razon, C. Wirtz, R. Mynott, Chem.
Eur. J. 2003, 9, 320–326; c) A. Fürstner, F. Jeanjean, P. Razon, Angew.
Chem. Int. Ed. 2002, 41, 2097–2101; Angew. Chem. 2002, 114, 2203–
2206; d) A. Fürstner, K. Langemann, J. Am. Chem. Soc. 1997, 119, 9130–
9136; e) A. Fürstner, K. Langemann, J. Org. Chem. 1996, 61, 3942–3943.
[14] a) X. Wang, E. J. Bowman, B. J. Bowman, J. A. Porco, Angew. Chem. Int.
Ed. 2004, 43, 3601–3605; Angew. Chem. 2004, 116, 3685–3689; b) F.
Grisi, C. Costabile, A. Grimaldi, C. Viscardi, C. Saturnino, P. Longo, Eur. J.
Org. Chem. 2012, 5928–5934.
[15] T. R. Hoye, C. S. Jeffrey, M. A. Tennakoon, J. Wang, H. Zhao, J. Am. Chem.
Soc. 2004, 126, 10210–10211.
[16] B. R. King, S. M. Swick, S. L. Schaefer, J. R. Welch, E. F. Hunter, G. W.
O’Neil, Synthesis 2014, 46, 2927–2936.
Acknowledgements
[17] W. Zhu, M. JimØnez, W.-H. Jung, D. P. Camarco, R. Balachandran, A. Vogt,
B. W. Day, D. P. Curran, J. Am. Chem. Soc. 2010, 132, 9175–9187.
[18] X. Wang, J. A. Porco, J. Am. Chem. Soc. 2003, 125, 6040–6041.
[19] H. Takamura, K. Iwamoto, E. Nakao, I. Kadota, Org. Lett. 2013, 15, 1108–
1111.
Generous financial support by the Ministerium für Wissen-
schaft, Forschung und Kunst des Landes Baden-Württemberg
(Landesgraduierten fellowship for T. H.), the Fonds der Chemi-
schen Industrie, and the Gesellschaft Deutscher Chemiker
(travel grant for T. H.) is gratefully acknowledged.
[20] A. Ortiz, A. Quesada, A. Sanchez, J. Chem. Ecol. 2004, 30, 991–1000.
[21] W. Biernacki, A. Gdula, Synthesis 1979, 37–38.
[22] R. J. Anderson, V. L. Corbin, G. Cotterrell, G. R. Cox, C. A. Henrick, F.
Schaub, J. B. Siddall, J. Am. Chem. Soc. 1975, 97, 1197–1204.
[23] P. S. Koranne, T. Tsujihara, M. A. Arai, C. B. Bajracharya, T. Suzuki, K. Onit-
suka, H. Sasai, Tetrahedron: Asymmetry 2007, 18, 919–923.
[24] X.-J. Yu, H. Zhang, F.-J. Xiong, X.-X. Chen, F.-E. Chen, Helv. Chim. Acta
2008, 91, 1967–1974.
Keywords: macrocycles
substituent effects · terpenoids
·
metathesis
·
regioselectivity
·
[25] R. Appel, Angew. Chem. Int. Ed. Engl. 1975, 14, 801–811; Angew. Chem.
1975, 87, 863–874.
[26] M. J. Geier, M. R. GagnØ, J. Am. Chem. Soc. 2014, 136, 3032–3035.
[27] T. R. Hoye, H. Zhao, Org. Lett. 1999, 1, 1123–1125.
[28] B. M. Trost, G. Dong, J. A. Vance, J. Am. Chem. Soc. 2007, 129, 4540–
4541.
[1] R. R. Schrock, J. S. Murdzek, G. C. Bazan, J. Robbins, M. DiMare, M.
O’Regan, J. Am. Chem. Soc. 1990, 112, 3875–3886.
[2] a) S. T. Nguyen, L. K. Johnson, R. H. Grubbs, J. W. Ziller, J. Am. Chem. Soc.
1992, 114, 3974–3975; b) P. Schwab, R. H. Grubbs, J. W. Ziller, J. Am.
Chem. Soc. 1996, 118, 100–110.
[3] Reviews: a) R. H. Grubbs, Angew. Chem. Int. Ed. 2006, 45, 3760–3765;
Angew. Chem. 2006, 118, 3845–3850; b) R. R. Schrock, Angew. Chem. Int.
Ed. 2006, 45, 3748–3759; Angew. Chem. 2006, 118, 3832–3844; c) R. H.
Grubbs, Handbook of Metathesis, Vol. 1–3, Wiley-VCH, Weinheim, 2003;
d) A. Gradillas, J. PØrez-Castells, Angew. Chem. Int. Ed. 2006, 45, 6086–
6101; Angew. Chem. 2006, 118, 6232–6247.
[4] a) A. Fürstner, Angew. Chem. Int. Ed. 2000, 39, 3012–3043; Angew.
Chem. 2000, 112, 3140–3172; reviews on RCM in total synthesis:
b) K. C. Nicolaou, P. G. Bulger, D. Sarlah, Angew. Chem. Int. Ed. 2005, 44,
4490–4527; Angew. Chem. 2005, 117, 4564–4601; c) A. H. Hoveyda,
A. R. Zhugralin, Nature 2007, 450, 243–251; d) A. Deiters, S. F. Martin,
Chem. Rev. 2004, 104, 2199–2238; e) A. Fürstner, Chem. Commun. 2011,
47, 6505–6511; f) J. Cossy, S. Arseniyadis, C. Meyer, Metathesis in Natural
Product Synthesis Wiley-VCH, Weinheim, 2010.
[29] B. M. Trost, G. Dong, J. A. Vance, Chem. Eur. J. 2010, 16, 6265–6277.
[30] a) G. Karthikeyan, C. Zambaldo, S. Barluenga, V. Zoete, M. Karplus, N.
Winssinger, Chem. Eur. J. 2012, 18, 8978–8986; b) E. Moulin, S. Barluen-
ga, F. Totzke, N. Winssinger, Chem. Eur. J. 2006, 12, 8819–8834.
[31] L. C. Dias, F. G. Finelli, L. S. Conegero, R. Krogh, A. D. Andricopulo, Eur. J.
Org. Chem. 2010, 6748–6759.
[32] E. F. van der Eide, W. E. Piers, Nat. Chem. 2010, 2, 571–576.
[33] S. Torker, R. K. M. Khan, A. H. Hoveyda, J. Am. Chem. Soc. 2014, 136,
3439–3455.
[34] S. BouzBouz, R. Simmons, J. Cossy, Org. Lett. 2004, 6, 3465–3467.
[35] a) J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865–
3868; b) J. P. Perdew, M. Ernzerhof, K. Burke, J. Chem. Phys. 1996, 105,
9982–9985.
Chem. Eur. J. 2015, 21, 12396 – 12404
www.chemeurj.org
12403
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[36] Turbomole V6.6 2014, a development of University of Karlsruhe and
Forschungszentrum Karlsruhe GmbH, 1989–2007, TURBOMOLE GmbH,
since 2007; available from http://www.turbomole.com.
[40] a) S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys. 2010, 132,
154104; b) S. Grimme, S. Ehrlich, L. Goerigk, J. Comput. Chem. 2011, 32,
1456–1465.
[41] A. Klamt, G. Schüürmann, J. Chem. Soc. Perkin Trans. 2 1993, 799–805.
[37] M. Sierka, A. Hogekamp, R. Ahlrichs, J. Chem. Phys. 2003, 118, 9136–
9148.
[38] F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005, 7, 3297–3305.
[39] F. Weigend, Phys. Chem. Chem. Phys. 2006, 8, 1057–1065.
Received: June 26, 2015
Published online on July 28, 2015
Chem. Eur. J. 2015, 21, 12396 – 12404
www.chemeurj.org
12404
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim