Prins Cyclization Catalyzed by a FeIII/Trimethylsilyl Halide System: The Oxocarbenium Ion Pathway versus the [2+2] Cycloaddition

Article abstract of DOI:10.1002/chem.201502488

The different factors that control the alkene Prins cyclization catalyzed by iron(III) salts have been explored by means of a joint experimental-computational study. The iron(III) salt/trimethylsilyl halide system has proved to be an excellent promoter in

Full text of DOI:10.1002/chem.201502488

DOI: 10.1002/chem.201502488
Full Paper
&
Reaction Mechanisms |Hot Paper|
Prins Cyclization Catalyzed by a FeIII/Trimethylsilyl Halide System:
The Oxocarbenium Ion Pathway versus the [2+2] Cycloaddition
Sixto J. PØrez,[a] Martín Purino,[a] Pedro O. Miranda,[c] Víctor S. Martín,*[a] Israel Fernµndez,*[b]
and Juan I. Padrón*[a, c]
Abstract: The different factors that control the alkene Prins
cyclization catalyzed by iron(III) salts have been explored by
means of a joint experimental–computational study. The
iron(III) salt/trimethylsilyl halide system has proved to be an
excellent promoter in the synthesis of crossed all-cis disub-
stituted tetrahydropyrans, minimizing the formation of prod-
ucts derived from side-chain exchange. In this iron(III)-cata-
lyzed Prins cyclization reaction between homoallylic alcohols
and non-activated alkenes, two mechanistic pathways can
be envisaged, namely the classical oxocarbenium route and
the alternative [2+2] cycloaddition-based pathway. It is
found that the [2+2] pathway is disfavored for those alco-
hols having non-activated and non-substituted alkenes. In
these cases, the classical pathway, via the key oxocarbenium
ion, is preferred. In addition, the final product distribution
strongly depends upon the nature of the substituent adja-
cent to the hydroxy group in the homoallylic alcohol, which
can favor or hamper a side 2-oxonia-Cope rearrangement.
Introduction
Tetrahydropyrans (THPs) are common structural units present
in many natural products,[1] from marine natural products such
as brevetoxin B or gambierol to biologically active macrolides
and macrolactones, such as dactylolide and bryostatin 1, re-
spectively (Figure 1).[2,3] Therefore, THP derivatives represent
very attractive targets for synthetic organic chemists. Among
the different methods and strategies to synthetize THP rings,[4]
the so-called Prins cyclization reaction has emerged as a power-
ful tool allowing the efficient access to these oxacycles. For
this reason, it is not surprising that this process has been
widely applied to the synthesis of natural products.[5] This cycli-
zation is based on the reaction between a homoallylic alcohol
or derivative with aldehydes promoted or catalyzed by Brønst-
ed as well as Lewis acids (Scheme 1).[6] The keystone in the
widely accepted mechanism is the generation of an oxocarbe-
[a] S. J. PØrez, Dr. M. Purino, Prof. V. S. Martín, Dr. J. I. Padrón
Instituto Universitario de Bio-Orgµnica “Antonio Gonzµlez”
Universidad de La Laguna, C/Francisco Sµnchez 2
38206 La Laguna, Tenerife (Spain)
Figure 1. Selected examples of natural products containing tetrahydropyran
subunits.
Fax: (+34)922318571
E-mail: vmartin@ull.es
[b] Dr. I. Fernµndez
Departamento de Química Orgµnica I, Facultad de Ciencias Químicas
Universidad Complutense de Madrid, 28040 Madrid (Spain)
E-mail: israel@quim.ucm.es
nium ion, which drives the cyclization. However, this species
can also undergo an oxonia-Cope rearrangement as a competi-
tive pathway, leading to two deleterious problems, namely
a mixture of products by side-chain exchange and racemiza-
tion. These processes have been thoroughly studied by the
groups of Rychnovsky and Willis.[7]
[c] Dr. P. O. Miranda, Dr. J. I. Padrón
Instituto de Productos Naturales y Agrobiología
Consejo Superior de Investigaciones Científicas (CSIC)
C/Francisco Sµnchez 3, 38206 La Laguna, Tenerife (Spain)
Fax: (+34)922260135
From the very beginning, this cyclization was promoted by
using strongly acidic conditions in stoichiometric amounts and
mixed acetals as starting materials. This strategy ensured the
E-mail: jipadron@ipna.csic.es
Supporting information and ORCID(s) from the author(s) for this article are
available on the WWW under http://dx.doi.org/10.1002/chem.201502488.
Chem. Eur. J. 2015, 21, 15211 – 15217
15211
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 1. Oxocarbenium ion-driven Prins cyclization.
reaction proceeded through an oxocarbenium intermediate as
the unique mechanistic pathway.[7,8] Recently, many different
Lewis acids have been used in the synthesis of substituted tet-
rahydropyrans with different substituents at the C4 position,
that is, halide[9] and hydroxy moieties.[6f,10] Strikingly, the com-
bination of substoichiometric amounts of Lewis acid with tri-
methylsilyl halide (TMSX) led to a significant improvement in
this cyclization reaction. Loh and co-workers applied a combi-
nation of indium salts and chlorotrimethylsilane (TMSCl) as cat-
alyst to the synthesis of 4-halotetrahydropyran rings.[11] More
recently, the combination of TMSX with iron(III) salts (FeX3 and
Fe(acac)3) permitted our group to catalyze the Prins cyclization
with non-activated alkenes and alkynes in a sustainable metal
catalysis context (Scheme 1).[12]
Scheme 2. Possible mixture of tetrahydropyrans in the Prins cyclization by
participation of oxonia-Cope rearrangement.
also undergo an oxonia-Cope rearrangement to generate
a new homoallylic alcohol 1* and an aldehyde that can partici-
pate in a new Prins cyclization process, thus generating the
undesired species 3 and 5, besides the tetrahydropyran 4
(Scheme 2). Furthermore, due to the involvement of an oxocar-
benium ion, a mixture of diastereoisomers would be expected.
Therefore, it was necessary to carry out a study on the influ-
ence of the oxonia-Cope rearrangement on the tetrahydropyr-
an distribution, as well as the possible consequences when
using non-activated alkenes.
This catalytic system is widely applicable and promotes the
construction of substituted six-membered oxa- and azacycles,
leading to the corresponding chloro, bromo and iodo hetero-
cycles by a suitable combination of an iron(III) source, a trime-
thylsilyl halide, and the solvent. The devised catalytic cycle
relies on a ligand exchange between the iron complex and
a halosilane, which regenerates the iron(III) halide due to the
more oxophilic character of the silicon atom. This system pres-
ents several advantages, such as the use of reduced amounts
of metal, tolerance of functional groups in the aldehyde, and
cleaner reactions. By using this methodology, the synthesis of
enantiomerically pure 4-hydroxytetrahydropyran derivatives
was recently accomplished by Feng and co-workers in an ex-
cellent study, through a sequential ene/Prins cyclization with
FeCl3/TBSCl (TBS= tert-butyldimethylsilyl) as a catalyst.[6f ]
Herein, we describe the different factors that control the
alkene Prins cyclization catalyzed by iron(III) salts in the pres-
ence of SiMe3X (X=Cl, Br) towards the synthesis of crossed dis-
ubstituted tetrahydropyrans. Density functional theory (DFT)
calculations support the preference of the classical oxocarbeni-
um ion pathway over the possible [2+2] cycloaddition path-
way and explain the origin of the observed product distribu-
tion, which depends on the nature of the substituent at the
homoallylic alcohol and the iron(III) source.
In a preliminary work using stoichiometric amounts of iron(-
III) chloride, we observed that the distribution of tetrahydro-
pyrans mainly depends on two factors, namely the bulkiness
of the substituent R1 and the electronic effect of the substitu-
ents directly attached to R1 if R1 is an aryl group. An increase
in the bulkiness of R1 (methyl, ethyl and cyclohexyl) increased
the formation of the desired THP 4 with a subsequent de-
crease in that of THP 5. This steric control therefore favored
the Prins cyclization to 4 over the 2-oxonia-Cope rearrange-
ment. Moreover, when R1 was an aromatic ring, the presence
of electron-deficient substituents favored the Prins cyclization,
yielding THP 4 as the major product.[13]
To better understand the effect of the catalytic system on
the competitive processes (Prins cyclization vs. oxonia-Cope re-
arrangement), different sources of FeIII salts, as well as TMSX,
were evaluated (Table 1). Moreover, these results can be com-
pared to our previous studies using stoichiometric amounts of
FeCl3.[13]
When R1 is a phenyl ring and R2 an iso-butyl group, the cycli-
zation under stoichiometric conditions leads to a 50:50 mixture
of tetrahydropyrans 4 and 5 in 70% yield.[13] We focused on
this particular case as a starting point of our comparative
study. In the case of allyl phenyl carbinol 1 (R1 =Ph) the per-
centage of the desired THP 4 was improved to 87% under cat-
alytic conditions, and electron-deficient substituents were not
necessary to achieve this result (Table 1, entry 1).[7a] This is a re-
markable result, because Willis and co-workers had to intro-
duce an electron-deficient substituent on the aromatic ring
and use stoichiometric amounts of Lewis acid to obtain the
Results and Discussion
Influence of the oxonia-Cope rearrangement in the alkene
Prins cyclization using the catalytic system FeIII/TMSX
As discussed above, the Prins cyclization mechanism is based
on the generation of an oxocarbenium ion intermediate 2,
which drives the formation of the corresponding tetrahydro-
pyrans (Scheme 2). However, this cationic intermediate may
Chem. Eur. J. 2015, 21, 15211 – 15217
www.chemeurj.org
15212
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
We have therefore solved some of the side reactions, avoiding
the exchange of aldehyde and alcohol side chains.
Table 1. Effect of catalytic system (FeIII/TMSX) on the product distribution
of direct Prins cyclization versus 2-oxonia-[3,3]-rearrangement.[a]
Theoretical calculations on iron(III)-catalyzed alkene Prins
cyclization
DFT calculations were carried out (PCM(CH2Cl2)B3LYP/def2-
TZVP//B3LYP/def2-SVP level)[16,17] to gain more insight into the
reaction mechanism of the above iron(III)-catalyzed Prins cycli-
zation reactions. To this end, we selected the reaction between
the model reactants 1 (R1 =Me) and MeCHO in the presence of
FeCl3 as catalyst.
Entry R1
R2
X
FeIII
Selectivity [%][b,c]
Yield [%]
4
5
1
Ph
iBu
Cl FeCl3
Fe(acac)3
c-Hex Cl FeCl3
Fe(acac)3
Ph(CH2)2 c-Hex Cl FeCl3
Fe(acac)3
Cl FeCl3
Fe(acac)3
c-Hex Cl FeCl3
87
87
89
85
94
94
94
93
93
13
13
11
15
6
6
6
7
7
86
80
90
90
90
88
80
75
85
75
95
96
95
95
87
90
95
90
74
84
90
87
95
99
2
Ph
3
At this point, it is important to consider the recent work of
Feng and co-workers, based on labeling as well as DFT stud-
ies.[6f] These authors suggest a new mechanism for the FeCl3-
catalyzed Prins cyclization between homoallylic alcohols and
aldehydes. It was proposed that the entire catalytic cycle pro-
ceeds via the following three consecutive steps: 1) initial [2+2]
cycloaddition reaction leading to the oxetane intermediate A;
2) nucleophilic attack of the OH group at the electrophilic CÀ
OFeCl3 center with concomitant ring opening to form inter-
mediate B; 3) intramolecular proton transfer to produce com-
pound C (Scheme 3). Step (2) occurs with an activation barrier
of approximately 27 kcalmolÀ1 and constitutes the rate-deter-
mining step of the entire transformation.[6f]
4
Ph(CH2)2 Ph
5
n-Hex
iBu
Fe(acac)3 100
0
6
iBu
Cl FeCl3
100
0
Fe(acac)3 100
0
7
iBu
c-Hex Cl FeCl3
100
0
Fe(acac)3 100
0
8
Ph
iBu
Br FeBr3
Fe(acac)3
c-Hex Br FeBr3
Fe(acac)3
c-Hex Br FeBr3
Fe(acac)3
Br FeBr3
80
75
79
89
95
20
25
21
11
5
9
Ph
10
11
12
n-Hex
iBu
95
5
iBu
100
0
Fe(acac)3 100
0
iBu
c-Hex Br FeBr3
Fe(acac)3
95
96
5
4
[a] Reactions conditions:
1
(1.0 mmol), R2CHO (1.2 mmol) and FeIII
(0.1 mmol)/TMSX (1.1 mmol) in dry CH2Cl2 (10 mL) at rt for 15 h. [b] Yields
of products after purification by silica gel column chromatography. [c] We
did not detect any THP 3 product.
same product distribution.[7a] The cyclization works well with
both iron(III) salts, FeCl3 and Fe(acac)3, leading to similar reac-
tion yields. The combination of these conditions with a bulkier
group at R2 (c-Hex) results in a slight increase of the percent-
age of THP-4 up to 89% (Table 1, entry 2). With homobenzyl as
the substituent at R1, the crossed 4-chloro-2,6-trisubstituted
THP 4 was obtained almost exclusively, minimizing the com-
petitive 2-oxonia-Cope rearrangement and side-chain ex-
change (Table 1, entry 3).[14] The best results were obtained
when combining n-hexyl and iso-butyl as R1 with c-hexyl as R2
(Table 1, entries 5 and 7). In these cases, the crossed THP 4 was
obtained exclusively with excellent yields. Similar results were
obtained with the bromo version of the catalytic system
(Table 1, entries 8–12). The best ratios of crossed 4-bromo-2,6-
trisubstituted THP 4 were obtained with c-hexyl as R2 and
Fe(acac)3 as iron(III) source (Table 1, entries 10 and 12). The
nature of the silyl additive did not influence either the ratio or
the yield of the Prins cyclization catalyzed by iron(III).[15]
Scheme 3. Feng’s mechanistic model for the FeCl3-catalyzed Prins cyclization
This alternative reaction mechanism can be only considered
as a special class of the Prins cyclization because of the partic-
ular nature of the substrates used by Feng and co-workers.
These authors use a homoallylic alcohol with an activated
alkene (typically R1 =aromatic) having an ester group as R2,
which favors the initial stepwise [2+2] cycloaddition step
(Scheme 3). Feng and co-workers also showed that homoallylic
alcohols with 4-unsubstituted alkenes but keeping the ester at
R2 did not undergo the Prins cyclization.[6f] In spite of this, we
were curious to calculate an analogous reaction pathway for
our substrates, where R1 =H and R2 ¼ ester. From the barrier
energies given in Figure 2, it seems that the pathway leading
to the axial-hydroxy derivative INT4 (C in Scheme 3) is feasible
from the initially formed oxetane intermediate INT2 (A in
Scheme 3). However, the formation of the latter intermediate is
clearly more difficult for systems lacking an aromatic substitu-
At this moment, the combination of catalytic amounts of
iron(III) salts and TMSCl as additive is effective and nearly elimi-
nates the 2-oxonia-Cope rearrangement, favoring the direct
Prins cyclization under the specified experimental conditions.
Chem. Eur. J. 2015, 21, 15211 – 15217
www.chemeurj.org
15213
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 2. Computed reaction profile of the reaction between 1 and MeCHO in the presence of FeCl3: [2+2] cycloaddition pathway. Relative energies (ZPVE in-
cluded) are given in kcalmolÀ1. All data were computed at the PCM(CH2Cl2)B3LYP/def2-TZVP//B3LYP/def2-SVP level
ent at the 4-position relative to the oxygen. Not surprisingly,
the computed activation barrier for the second step of the
[2+2] cycloaddition (which produces INT2) is much higher
(29.4 kcalmolÀ1, corresponding free energy of 38.7 kcalmolÀ1
)
than the analogous process reported by Feng and co-workers
(4.7 kcalmolÀ1, R1 =Ph, R2 =CO2Et, R3 =Et, at the similar
PCM(CH2Cl2)//B3LYP/6-31G** level).[6f] In addition, all of our at-
tempts to locate the corresponding carbocationic intermediate
INT1 on the potential energy surface (even including the sol-
vent during the geometry optimizations) met with no success,
that is, only dissociation into the initial reactants was found.
These results suggest that the [2+2] cycloaddition is disfa-
vored when using homoallylic alcohols with non-activated and
non-substituted alkenes. In the absence of substituents at the
double bond, as in the substrates considered in the present
work, this pathway can be ruled out and therefore, the so-
called “classical” pathway seems to be operating. Further ex-
perimental support for this finding is given by the absence of
the corresponding 4-hydroxy derivatives from INT4 (no traces
of this type of product were detected in the reaction crude
product mixtures). Moreover, if the [2+2] mechanism were op-
erating, adding 4-hydroxy or 4-OTMS THP (compounds 6 and
7, respectively) would produce the observed 4-equatorial chlo-
ride product 4 under the same reaction conditions (Scheme 4).
However, the latter reaction product 4 was only observed in
very low yield by treatment of the 4-OTMS THP 7 under cata-
lytic conditions [Scheme 4, Eq. (2)].
Scheme 4. Control experiments to test for plausible hydroxy intermediates
of the [2+2] cycloaddition pathway.
an endothermic process (DE=14.8 kcalmolÀ1). This intermedi-
ate is transformed into the metalla-1,3-dioxetane INT8 through
the transition state TS5, a saddle point associated with the for-
mation of the new CÀO bond, with an activation barrier of
12.5 kcalmolÀ1 (the total activation energy from INT6 to TS5 is
27.3 kcalmolÀ1). INT8 then evolves into INT9 via TS6, a transi-
tion state associated with the protonolysis of FeÀO promoted
by HCl. The ease of the latter reaction step is confirmed by the
low computed activation barrier (DE° =6.2 kcalmolÀ1) and
high exothermicity (DE=À16.0 kcalmolÀ1). It can be suggested
that INT9 is then converted into INT10 in a slightly endother-
mic reaction (DE=3.2 kcalmolÀ1). This reaction very likely pro-
ceeds via decoordination of the FeCl3 catalyst followed by a nu-
cleophilic substitution reaction between the readily formed al-
Therefore, it can be concluded that the [2+2] pathway does
not occur within the substrates and reaction conditions consid-
ered in this work. Once the classical pathway was confirmed to
be preferred, we then computationally studied the formation
of the key oxocarbenium ion 2 (Scheme 2). The computed re-
action profile for the reaction between 1 (R1 =Me) and the
iron(III)-coordinated aldehyde MeCHO is depicted in Figure 3.
Our calculations suggest that the process begins with the co-
ordination of the initial reactants to the FeCl3 to form INT6.
This complex is stabilized by an intramolecular Cl···HO hydro-
gen bond which weakens the corresponding FeÀCl bond. As
a result, a molecule of HCl can be released to produce INT7 in
À
cohol and SiMe3Cl to produce the cation INT10 and FeCl4 as
counteranion. We were not able to locate a transition state as-
sociated with the intramolecular nucleophilic addition of the
C=C double bond with concomitant release of SiMe3OH in
INT10. Instead, we found a SN1-type mechanism, which in-
volves the initial release of the SiMe3OH fragment to produce
the cationic intermediate INT11. This step is highly exothermic
Chem. Eur. J. 2015, 21, 15211 – 15217
www.chemeurj.org
15214
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 3. Computed reaction profile of the reaction between 1 and MeCHO in the presence of FeCl3 and SiMe3Cl: oxocarbenium ion pathway. Relative ener-
gies (ZPVE included) are given in kcalmolÀ1 and bond distances in angstroms. All data were computed at the PCM(CH2Cl2)B3LYP/def2-TZVP//B3LYP/def2-SVP
level.
(DE=À14.9 kcalmolÀ1) as a result of the stabilization of the
carbocation by the adjacent oxygen atom. Then, the intramo-
lecular cyclization reaction occurs via TS7, which is associated
with the formation of the new CÀC bond with an activation
barrier of only 6.8 kcalmolÀ1. This process leads to the forma-
tion of INT12, in which both methyl substituents are placed in
equatorial positions. The geometry of this carbocation directs
À
the subsequent nucleophilic attack, either by FeCl4 or simply
by ClÀ, from the equatorial position, as the axial delivery is
sterically hampered (see inset in Figure 3).
Scheme 5. Contribution of the acetal to the oxocarbenium ion pathway.
From the data in Figure 3, the initial CÀO bond formation
and subsequent FeÀO protonolysis constitute the rate-deter-
mining steps of the entire transformation. In addition, the
overall relative reaction profile for this classical oxocarbenium
pathway is energetically favored over the alternative [2+2] cy-
cloaddition pathway (see Figure 2), which is in good agree-
ment with the experimental observations.
instance, use of Fe(OTf)3/TMSCl as the catalytic system leads to
a decrease in the amount of THP 4 and an increase in the
amounts of THPs 5 and 3, the latter being observed for the
first time (Table 2, entry 3).
Similar results were obtained when using allyl phenyl carbi-
nol (1, R1 =Ph) and a reaction temperature of 08C. In this case,
we obtained a mixture of THPs 3, 4 and 5 regardless of the
iron(III) salt used (Table 2, entries 4–6).[18] When using Fe(OTf)3,
the major product was THP 5 (Table 2, entry 6). This behavior
was also observed when the temperature of the reaction was
changed to 408C (Table 2, entries 10–12). However, when the
homoallylic alcohol 1 bore an aliphatic substituent, the Prins
cyclization led exclusively to the THP 4, regardless of the reac-
tion temperature or iron ligands (Table 2, entries 1, 2, 7–9, and
13–15).
The occurrence of a mixture of products (3, 4 and 5,
Scheme 2) by means of competitive 2-oxonia-Cope rearrange-
ment and side-chain exchange reaction is fully compatible
with the above described classical pathway via the key oxocar-
benium ion. However, the final product distribution strongly
depends on the experimental conditions, type of alkene, and
iron(III) source. For instance, we observed significant differen-
ces when using stoichiometric instead of catalytic conditions;
the relative amount of 5 is in general higher with stoichiomet-
ric amounts of iron(III) salts than under catalytic conditions.
Moreover, acetal 8, which contributes to the formation of oxo-
carbenium ion 9, is only formed under stoichiometric condi-
tions (Scheme 5).
From all this data in Table 2, it is clear that the oxonia-Cope
rearrangement is favored when using allyl phenyl carbinols, tri-
flate as the iron salt, and when the temperature of the cycliza-
tion is modified from room temperature.
At this point, we tried to rationalize the finding of the for-
mation of the mixture of THPs when R1 =Ph. To this end, we
carried out a theoretical study, comparing the oxonia-Cope re-
arrangement with phenyl and iso-butyl as R1 substituents
The iron ligand and the reaction temperature both also have
a direct effect on the tetrahydropyran distribution when the
homoallylic alcohol 1 bears at R1 a phenyl group (Table 2). For
Chem. Eur. J. 2015, 21, 15211 – 15217
www.chemeurj.org
15215
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
to 2-iBu, which suggests an equilibrium between both spe-
cies). Therefore, our computational data indicate that the
oxonia-Cope rearrangement is favored when R1 =Ph, which is
in agreement with the experimental results gathered in
Table 2.
Table 2. Iron ligand and temperature effect on the ratio of direct Prins
Cyclization versus product derived from the 2-oxonia-[3,3]-rearrange-
ment.[a]
Finally, racemization in the Prins cyclization has also been
suggested as an indicator for participation of the oxonia-Cope
rearrangement in the process.[7] Thus, the next step was to
check the level of racemization in the process under our cata-
lytic conditions that, in principle, should be low based on the
data in Table 1. For this purpose, we carried out the Prins cycli-
zation between alcohol (R)-1 and benzaldehyde using our cata-
lytic system (Scheme 6). The use of 10 mol% of FeCl3 yielded
the desired product 8 with 96% ee, which nicely indicates that
no competitive 2-oxonia-Cope rearrangement or side-chain ex-
change occurred.
1
Entry
R
R2
T [8C] FeL3
Selectivity [%]
Yield [%]
4
5
3[b]
1
2
3
4
5
6
7
8
iBu
iBu
RT
Fe(OTf)3 100
Fe(OTf)3 100
0
0
0
0
90
>99
74
75
87
75
80
84
73
80
75
64
>99
95
iBu c-Hex RT
Ph
Ph
Ph
Ph
iBu
iBu
iBu
iBu
RT
0
Fe(OTf)3
FeCl3
30 45 25
22 59 19
23 62 15
0
0
Fe(acac)3
Fe(OTf)3
15 85
0
0
0
0
0
0
iBu c-Hex
iBu c-Hex
iBu c-Hex
Ph
Ph
Ph
iBu c-Hex 40
iBu c-Hex 40
iBu c-Hex 40
0
FeCl3
100
0
0
0
0
0
Fe(acac)3 100
Fe(OTf)3 100
9
10
11
12
13
14
15
iBu
iBu
iBu
40
40
40
FeCl3
Fe(acac)3
Fe(OTf)3
FeCl3
10 90
25 75
28 57 14
100
0
0
0
0
0
0
Fe(acac)3 100
Fe(OTf)3 100
>99
[a] Reactions conditions:
1
(1.0 mmol), R2CHO (1.2 mmol) and FeIII
(0.1 mmol)/TMSCl (1.1 mmol) in dry CH2Cl2 (10 mL) at RT for 15 h or 08C
at 24 h; 4 h at 408C. [b] Yields of products after purification by silica gel
column chromatography.
Scheme 6. Iron(III) salt/TMSCl system catalyzes Prins cyclization preventing
racemization.
Conclusion
We have established a method to obtain almost exclusively
crossed 2,4,6-trisubstituted tetrahydropyrans, avoiding side-
chain exchange and racemization processes, by using a catalyt-
ic system of iron(III) salts and TMSX. Two mechanistic pathways
can be envisaged for this iron(III)-catalyzed Prins cyclization re-
action between homoallylic alcohols and non-activated al-
kenes, namely the classical oxocarbenium route and the alter-
native [2+2] cycloaddition-based pathway. Our joint experi-
mental and computational study clearly indicates that the
[2+2] pathway is disfavored for alcohols having non-activated
and non-substituted alkenes. In these cases, the classical path-
way, via the key oxocarbenium ion 2, is preferred. Furthermore,
the nature of the substituent adjacent to the hydroxy group in
the homoallylic alcohol is decisive in controlling the possible
oxonia-Cope rearrangement of the corresponding oxocarbeni-
um ion. Thus, whereas alkyl substituents do not favor this side
reaction and therefore lead almost exclusively to crossed tetra-
hydropyran derivatives, the 2-oxonia-Cope rearrangement is
strongly thermodynamically favored in the presence of
a phenyl group. As a result, a mixture of tetrahydropyrans is
observed in these cases.
Figure 4. Comparative computed reaction profiles for the oxonia-Cope rear-
rangement with phenyl and alkyl as R1 substituents. Relative energies (ZPVE
included) are given in kcalmolÀ1. All data were computed at the
PCM(CH2Cl2)B3LYP/def2-TZVP//B3LYP/def2-SVP level.
(Figure 4). Although the computed activation barrier is quite
similar in both cases (DDE° of only 0.2 kcalmolÀ1), the process
is greatly thermodynamically favored (DDE=9.3 kcalmolÀ1) for
the oxocarbenium ion with
a phenyl substituent (2-Ph,
Experimental Section
Figure 4). This can of course be ascribed to the stabilization by
p-conjugation exerted by the phenyl group in 2’-Ph, which
constitutes the driving force for the rearrangement. In contrast,
the stabilization by hyperconjugation exerted by the isobutyl
group is comparatively much weaker, making the rearrange-
ment much more difficult (i.e., 2’-iBu is practically isoenergetic
General methods and computational details are given in the Sup-
porting Information.
General procedure for iron(III) salt/TMSX catalyzed Prins cycliza-
tion: To a solution of homoallylic alcohol (1.0 equiv) in anhydrous
CH2Cl2 (0.1m) was added the corresponding aldehyde (1.2 equiv)
Chem. Eur. J. 2015, 21, 15211 – 15217
www.chemeurj.org
15216
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
ani, S. Gastaldi, M. P. Bertrand, Tetrahedron 2010, 66, 413–445; d) E. A.
Crane, K. A. Scheidt, Angew. Chem. Int. Ed. 2010, 49, 8316–8326; Angew.
Chem. 2010, 122, 8494–8505; e) I. M. Pastor, M. Yus, Curr. Org. Chem.
2012, 16, 1277–1312; f) K. Zheng, X. Liu, S. Qin, M. Xie, L. Lin, C. Hu, X.
Feng, J. Am. Chem. Soc. 2012, 134, 17564–17573; g) S. J. Greco, G.
Fiorot, V. Lacerda, R. B. dos Santos, Aldrichimica Acta 2013, 46, 59–67.
[7] a) S. R. Crosby, J. R. Harding, C. D. King, G. D. Parker, C. L. Willis, Org. Lett.
2002, 4, 577–580; b) C. S. Barry, N. Bushby, J. R. Harding, R. A. Hughes,
G. D. Parker, R. Roe, C. L. Willis, Chem. Commun. 2005, 3727–3729; c) S.
Marumoto, J. J. Jaber, J. P. Vitale, S. D. Rychnovsky, Org. Lett. 2002, 4,
3919–3922; d) R. Jasti, S. D. Rychnovsky, J. Am. Chem. Soc. 2006, 128,
13640–13648.
and then the iron(III) salt (0.1 equiv) and TMSX (1.1 equiv), in this
order. The reaction mixture was stirred at room temperature for
15 h. The reaction was quenched by addition of water with stirring
for 90 min, and the mixture extracted with CH2Cl2. The combined
organic layers were dried over magnesium sulfate and filtered and
the solvent was removed under reduced pressure. This crude reac-
tion mixture was purified by flash column chromatography on
silica gel (n-hexane/EtOAc solvent systems).
Acknowledgements
[8] a) R. C. Winstead, T. H. Simpson, G. A. Lock, M. D. Schiavelli, D. W.
Thompson, J. Org. Chem. 1986, 51, 275–277; b) L. E. Overman, A. Casta-
Çeda, T. A. Blumenkopf, J. Am. Chem. Soc. 1986, 108, 1303–11304.
[9] a) Y. Kishi, H. Nagura, S. Inagi, T. Fuchigami, Chem. Commun. 2008,
3876–3878; b) Y. Kishi, H. Nagura, S. Inagi, T. Fuchigami, Eur. J. Org.
Chem. 2009, 103–109; c) U. Biermann, A. Lützen, J. O. Metzger, Eur. J.
Org. Chem. 2006, 2631–2637; d) A. P. Dobbs, L. Pivnevi, J. J. Penny, S.
Martinovic, J. N. Iley, P. T. Stephenson, Chem. Commun. 2006, 3134–
3136; e) J. Vitale, S. D. Rychnovsky, J. Am. Chem. Soc. 2004, 126, 9904–
9905; f) R. Jasti, C. D. Anderson, S. D. Rychnovsky, J. Am. Chem. Soc.
2005, 127, 9939–9945; g) F. Liu, T. P. Loh, Org. Lett. 2007, 9, 2063–2066;
h) G. Sabitha, K. B. Reddy, M, Bhikshapathi, J. S. Yadav, Tetrahedron Lett.
2006, 47, 2807–2810; i) J. S. Yadav, B. V. S. Reddy, G. G. K. S. N. Kumar,
G. M. Reddy, Chem. Lett. 2007, 36, 426–427.
[10] a) I. T. Kay, D. Bartholomew, Tetrahedron Lett. 1984, 25, 2035–2038;
b) W. C. Zhang, G. S. Viswanathan, C. J. Li, Chem. Commun. 1999, 291–
292; c) J. S. Yadav, B. V. S. Reddy, G. M. Kumar, C. V. S. R. Murthy, Tetrahe-
dron Lett. 2001, 42, 89–91; d) C. C. K. Keh, V. V. Mamboodiri, R. S. Varma,
C. J. Li, Tetrahedron Lett. 2002, 43, 4993–4996; e) D. J. Hart, C. E. Ben-
nett, Org. Lett. 2003, 5, 1499–1502; f) C. S. J. Barry, S. R. Crosby, J. R.
Hardy, R. A. Hughes, C. D. King, G. D. Parker, C. L. Willis, Org. Lett. 2003,
5, 2429–2532; g) V. K. Yadav, N. V. Kumar, J. Am. Chem. Soc. 2004, 126,
8652–8653; h) J. S. Yadav, B. V. S. Reddy, G. G. K. S. N. Kumar, S. Aravind,
Synthesis 2008, 395–400; i) K. Tadpetch, S. D. Rychnovsky, Org. Lett.
2008, 10, 4839–4842; j) B. V. S. Reddy, A. Venkateswarlu, G. G. K. S. N.
Kumar, A. Vinu, Tetrahedron Lett. 2010, 51, 6511; k) T. Voigt, C. Gerding-
Reimers, T. T. N. Tran, S. Bergmann, H. Lachance, B. Schçlermann, A.
Brockmeyer, P. Janning, S. Ziegler, H. Waldmann, Angew. Chem. Int. Ed.
2013, 52, 410–414; Angew. Chem. 2013, 125, 428–432.
We first thank the referees of this work for several suggestions
which strengthened the manuscript. This research was sup-
ported by the Spanish MINECO, co-financed by the European
Regional Development Fund (ERDF) and FEDER (CTQ2011-
28417-C02–01/BQU, CTQ2014-56362-C2–1-P, CTQ2011-22653,
CTQ2013-44303-P) and IMBRAIN project (FP7-REGPOT-2012-
CT2012-31637-IMBRAIN), funded under the 7th Framework Pro-
gramme (CAPACITIES). S.J.P. thanks the Spanish MINECO for
a F.P.U fellowship. The research leading to these results re-
ceived funding from the People Programme (Marie Curie Ac-
tions) of the European Union’s Seventh Framework Programme
(FP7/2007-2013 under REA grant agreement no. 623155 to
P.O.M). The ORFEO-CINQA network is also acknowledged.
Keywords: cyclization
· density functional calculations ·
heterocycles · iron · Lewis acids
[1] Selected reviews: a) N. Li, Z. Shi, Y. Tang, J. Chen, X. Li, Beilstein J. Org.
Chem. 2008, 4, 48; b) B. G. Wang, J. B. Gloer, N. Y. Ji, J. C. Zhao, Chem.
Rev. 2013, 113, 3632–3685.
[2] a) J. C. Morris, A. J. Phillips, Nat. Prod. Rep. 2011, 28, 269–289, and previ-
ous issues in the series; b) D. Baden, K. Rein, L. Fieber, S. Parker, Patent
US20050124685A1, 2005.
[3] a) K. J. Hale, M. G. Hummersone, S. Manviazar, M. Frigerio, Nat. Prod.
Rep. 2002, 19, 413–453; b) P. A. Wender, J. L. Baryza, M. K. Hilinski, J. C.
Horan, C. Kan, V. A. Verma in Drug Discovery Research: New Frontiers in
the Post-Genomic Era, (Ed.: Z. Huang), John Wiley & Sons, Hoboken, NJ,
2007, pp. 127–162; c) A. E. Wright, J. C. Botelho, E. GuzmHn, D. Harmo-
dy, P. Linley, P. J. McCarthy, T. P. Pitts, S. A. Pomponi, J. K. Reed, J. Nat.
Prod. 2007, 70, 412–416; d) A. Cutignano, I. Bruno, G. Bifulco, A. Casa-
pullo, C. Debitus, L. Gómez-Paloma, R. Riccio, Eur. J. Org. Chem. 2001,
775–778; e) G. R. Pettit, C. L. Herald, D. L. Doubek, D. L. Herald, E.
Arnold, J. Clardy, J. Am. Chem. Soc. 1982, 104, 6846–6848.
[4] a) T. L. B. Boivin, Tetrahedron 1987, 43, 3309–3362; b) P. A. Clarke, S.
Santos, Eur. J. Org. Chem. 2006, 2045–2053; c) I. Larrosa, P. Romea, F.
Urpí, Tetrahedron 2008, 64, 2683–2723; d) J. Muzart, J. Mol. Catal. A
2010, 319, 1–29; e) T. Martín, J. I. Padrón, V. S. Martín, Synlett 2014, 12–
32.
[11] a) K. P. Chan, T. P. Loh, Org. Lett. 2005, 7, 4491–4494; b) K. P. Chan, Y. H.
Liu, T. P. Loh, Chem. Commun. 2007, 939–941.
[12] a) P. O. Miranda, R. M. Carballo, V. S. Martín, J. I. Padrón, Org. Lett. 2009,
11, 357–360; b) R. M. Carballo, G. Valdomir, M. Purino, V. S. Martín, J. I.
Padrón, Eur. J. Org. Chem. 2010, 2304–2313; c) M. Purino, M. A. Ramírez,
A. H. Daranas, V. S. Martín, J. I. Padrón, Org. Lett. 2012, 14, 5904–5907.
[13] P. O. Miranda, L. G. León, V. S. Martín, J. I. Padrón, J. M. Padrón, Bioorg.
Med. Chem. Lett. 2006, 16, 3135–3138.
[14] This data is in agreement with the results of Chan and Loh except in
the meso-type compound obtained. In our case, 5% yield of the meso
product 5 was obtained. However, 5% of the meso product 3 was ob-
tained by Chan and Loh using InIII salt instead of FeIII salts. See
ref. [11a].
[15] We obtained the same results using TMSCl and TBSCl as silyl additive
with no difference in the 4/5 ratio or yields.
[5] a) D. L. Aubele, S. Wan, P. E. Floreancig, Angew. Chem. Int. Ed. 2005, 44,
3485–3488; Angew. Chem. 2005, 117, 3551–3554; b) S. K. Woo, M. S.
Kwon, E. Lee, Angew. Chem. Int. Ed. 2008, 47, 3242–3244; Angew. Chem.
2008, 120, 3286–3288; c) P. A. Wender, B. A. DeChristopher, A. J. Schrier,
J. Am. Chem. Soc. 2008, 130, 6658–6659; d) D. K. Mohapatra, P. P. Das,
M. R. Pattanayak, J. S. Yadav, Chem. Eur. J. 2010, 16, 2072–2078; e) B.
Lin, Y.-C. Lai, Y. Zhao, Y. H. Wong, Z.-L. Shen, T. P. Loh, Angew. Chem. Int.
Ed. 2012, 51, 10619–10623; Angew. Chem. 2012, 124, 10771–10775;
f) X. Han, G.-R. Peh, P. E. Floreancig, Eur. J. Org. Chem. 2013, 1193–1208.
[6] a) E. Arundale, L. A. Mikeska, Chem. Rev. 1952, 51, 505–555; b) I. M.
Pastor, M. Yus, Curr. Org. Chem. 2007, 11, 925–927; c) C. Olier, M. Kaafar-
[16] This level was selected because a similar level (PCM(CH2Cl2)/B3LYP/6–
31G**) was used by Feng and co-workers in a related study. See ref. [6f]
[17] For computational details, see the Supporting Information.
[18] We did not detect benzyl chloride resulting from benzyl carbocation
formation.
Received: June 26, 2015
Published online on September 10, 2015
Chem. Eur. J. 2015, 21, 15211 – 15217
www.chemeurj.org
15217
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim