Protected propargylic acetals. Nicholas-Prins cyclization leading to functionalized 2-alkynyl-tetrahydropyrans. Intramolecular trapping by allenes

Article abstract of DOI:10.1016/j.tetlet.2007.09.021

Dicobalt hexacarbonyl complexes derived from propargylic acetals undergo Lewis acid-mediated Nicholas-Prins cyclization in the presence of homoallylic alcohols. The trapping of the resulting cyclic carbocation enables the synthesis of a series of functionalized tetrahydropyrans. The complexation of the triple bond contributes to the suppression of the side reactions and very significantly increases both the yield and the diastereoselectivity of the reaction. Homoallenic alcohols lead to dienic compounds through proton elimination.

Full text of DOI:10.1016/j.tetlet.2007.09.021

Tetrahedron Letters 48 (2007) 7801–7804
Protected propargylic acetals. Nicholas–Prins cyclization
leading to functionalized 2-alkynyl-tetrahydropyrans.
Intramolecular trapping by allenes
a
a
b
a,
*
`
´
´
Clarisse Olier, Stephane Gastaldi, Gerard Gil and Michele P. Bertrand
a
ˆ
LCMO, UMR 6517, Boite 562, Universite´ Paul Ce´zanne, Faculte´ des Sciences et Techniques, St Je´rome,
13397 Marseille Cedex 20, France
bLaboratoire de Ste´re´ochimie Dynamique et Chiralite´, UMR 6180, Universite´ Paul Ce´zanne, Faculte´ des Sciences et Techniques,
ˆ
St Je´rome, 13397 Marseille Cedex 20, France
Received 11 July 2007; revised 7 August 2007; accepted 4 September 2007
Available online 8 September 2007
Abstract—Dicobalt hexacarbonyl complexes derived from propargylic acetals undergo Lewis acid-mediated Nicholas–Prins
cyclization in the presence of homoallylic alcohols. The trapping of the resulting cyclic carbocation enables the synthesis of a series
of functionalized tetrahydropyrans. The complexation of the triple bond contributes to the suppression of the side reactions and
very significantly increases both the yield and the diastereoselectivity of the reaction. Homoallenic alcohols lead to dienic
compounds through proton elimination.
Ó 2007 Elsevier Ltd. All rights reserved.
1. Introduction
2. Results and discussion
The reactivity of alkyne dicobalt hexacarbonyl-stabi-
lized carbocations, commonly called Nicholas cations,
has been extensively studied. Nicholas reaction has
emerged as a powerful tool for the formation of C–C
and C-heteroatom bonds. Both inter- and intramolecu-
lar trapping by nucleophiles have led to brilliant syn-
thetic applications.1
As exemplified in Scheme 1, for the synthesis of fluori-
nated derivatives, Prins cyclization is preceded by the
intermolecular trapping of the ‘doubly stabilized’ cation
B by the homoallylic alcohol, immediately followed by
the formation of a new oxocarbenium ion D, which is
itself captured intramolecularly by the double bond.
The resulting carbocation E is then trapped by the exter-
nal nucleophile.
We have recently reported, that functionalized alkynyl
cyclohexanes could be prepared via intramolecular trap-
ping of Nicholas cation by terminal double bonds.2 This
reaction was unexpected in view of previous reports on
the trapping of complexed propargylic cations by non-
activated alkenes.3 Following on this study, we have
investigated the formal synthesis of 2-alkynyl tetrahydro-
pyrans4 through Nicholas–Prins cyclization. The
sequence occurs upon the treatment of cobalt complexed
propargylic acetals with homoallylic or homoallenic alco-
hols in the presence of different Lewis acids.5 The protec-
tion of the triple bond was shown to strongly improve
both the yield and the diastereoselectivity of the reaction.
The results obtained with homoallylic alcohols and o-vi-
nyl phenols are given in Table 1. The different propargy-
lic acetals and the unsaturated alcohols used in this
study are given in Figures 1 and 2. The products’ struc-
tures are given in Figure 3. For comparison sake, the re-
HO
R2
BF4
O
R2
OR HBF4
R1
(CO)6Co2
R1
R1
(CO)6Co2
-ROH
OR
OR
B
OR
C
(CO)6Co2
A
-ROH
HBF4
R1
R1
O
R1
(CO)6Co2
O
O
F
(CO)6Co2
(CO)6Co2
R2
R2
BF4
R2
Keywords: Alkyne complexes; Nicholas reaction; Prins cyclization;
Carbocations; Tetrahydropyrans.
E
F
D
*
Scheme 1. Reaction mechanism.
Corresponding author. E-mail: michele.bertrand@univ-cezanne.fr
0040-4039/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2007.09.021

7802
C. Olier et al. / Tetrahedron Letters 48 (2007) 7801–7804
Table 1. Reactions with homoallylic alcohols, o-vinyl phenols, and
homoallenic alcohols
O
O
R1
Me
(CO)6Co2
R2
X
(CO)6Co2
Entry Alcohol
(substrate)
Lewis acid Products isolated Cis/trans
12b
yield (%)
ratio
: R1= Me, R2= H, X= Cl
1
2
3
4
5
6
7
8
9
4a (1)a
4a (1)a
4a (10)
4a (1)a
4a (10)
4a (1)a
4a (10)
4a (2)a
4a (3)a
4b (1)a
4b (10)
4b (1)a
TiCl4
7a (58)
64/36
83/17
41/59
>99/1
63/37
93/7
57/43
42/58
41/59
7/93
O
: R1=Me, R2= H, X= F
Me
BF3ÆOEt2
BF3ÆOEt2
TMSOTf
TMSOTf
HBF4
HBF4
HBF4
HBF4
TiCl4
8a (52)
: R1=Me, R2= H, X= OTf
: R1= Ph, R2= H, X= F
8ab (20)d
9a (66)
(CO)6Co2
14b
: R1= SiMe3, R2= H, X= F
: R1= Me, R2= H, X=NHAc
: R1= Ph, R2= H, X=NHAc
: R1= Me, R2= Me, X= Cl
: R1= Me, R2= Me, X= F
: R1= Me, R2= Me, X=OEt
9ab (40)d
8a (63)
8ab (51)d
10a (85)
11a (80)
7b (66)
O
Me
Co2(CO)6
17c
10
11
12
TiCl4
BF3ÆOEt2
7b (27)
8b/12b
nd
64/36
17a : R3 = Me
18a : R3 = Ph
85/15 (64)
13b/12b
70/30 (53)
8b/12b
O
O
19a : R3 = SiMe3
c
4b (1)a
4b (1)a
BF3ÆOEt2
70/30
6/94
R3
Co2(CO)6
17b
Me
13
14
Co2(CO)6
HBF4
Figure 3. Products’ structures.
96/4 (65)
14b (29)
15a (43)
16a (32)
15
16
17
5b (1)a
4a (1)
4a (2)
TMSOTf
e
HBF4
HBF4
>99/1
>99/1
in yields ranging from 52% to 85% (entries 1, 2, 6, 8, and
9). High diastereoselectivity was observed for the forma-
tion of fluoride 8a (entries 2 and 6). The cis/trans ratio
was far lower for the reaction leading to chloride 7a
(entry 1). The cyclization carried out in the presence of
TMSOTf (entry 4) led cleanly to triflate 9a isolated as
a single cis stereoisomer. The structures of cis and trans
isomers were assigned from the characteristic coupling
patterns of the protons a to the polar substituents.6
When 3-hexen-1-ol (4c) bearing a non-terminal double
bond was used, the reaction led to a complex mixture
of products that could not be isolated as pure samples
and thereby were not identified.
e
a The reactions were carried out at room temperature in dichloro-
methane, on 0.3–0.4 mmol of complex in the presence of 1.1 equiv of
alcohol and 2.0 equiv of Lewis acid, unless otherwise stated.
b The products are intended non-complexed.
c The reaction was performed in nitromethane.
d 1H NMR yield using pentamethylbenzene as internal standard.
e The reaction was performed in freshly distilled dry acetonitrile.
OR
OR
OEt
OEt
R1
Me
(CO)6Co2
1 : R = Et, R1= Me
2 : R = Me, R1= Ph
3 : R = Et, R1= SiMe3
1'
Comparatively, the reactions performed on the non-
complexed acetal 10 led to the expected tetrahydropy-
rans but in far lower yields and selectivity (entries 3, 5,
7, and 11).7 As an example, 3-ethoxybut-2-enal and 4-
ethoxybut-3-en-2-one resulting from the degradation
of 10 were detected in the crude mixture (20% overall
NMR yield) in the presence of HBF4 (entry 7).8
Figure 1. Propargylic acetals.
HO
HO
4b
HO
4a
5a
4c
OH
OH
Prins cyclization is reputedly stereoselective.9 According
to DFT calculations, the secondary 4-tetrahydropyranyl
cation E adopts a chair conformation, where stabiliza-
tion occurs through the overlap of both the oxygen
equatorial lone pair and the empty p orbital at the cat-
ionic center with the C2–C3 and C5–C6 bonds orbi-
tals.10 Optimal delocalization places the hydrogen
atom at C4 in a pseudoaxial position. Thereby, preferen-
tial equatorial attack of the nucleophile leading to cis
isomers should be expected. This is the case for the fluo-
rinated product 8a in entries 2 and 6, and for triflate 9a
in entry 4. In this respect, the bulky complexed triple
bond is likely to lock the conformation of the cyclic
intermediate E and thus it contributes to increase the
diastereomeric ratio compared to non-complexed sub-
strate (entries 2/3; 4/5; and 6/7).
5b
•
OH
•
•
OH
OH
6a
6d
6c
6b
•
•
OH
•
OH
6e
OH
6f
Figure 2. Unsaturated alcohols.
sults of Prins cyclizations performed on non-complexed
acetal 10 in the presence of alcohol 4a are given in entries
3, 5, and 7, and in entry 11 for alcohol 4b.
When conducted in dichloromethane in the presence of
TiCl4, HBF4 or BF3ÆOEt2, the reactions carried out on
complexes 1–3 and alcohol 4a led to 4-halogenated
tetrahydropyrans isolated as a mixture of stereoisomers
The formation of a large amount of trans isomer, as ob-
served in entries, 1, 8, and 9, might result from thermo-
dynamic control. Epimerization at C1 is possible via

C. Olier et al. / Tetrahedron Letters 48 (2007) 7801–7804
Table 2. Intramolecular trapping with allenic alcohols15
7803
reversible Nicholas reaction.4a When a pure isolated
sample of cis-7a was allowed to react with TiCl4 for
30 min, a 60/40 ratio of cis and trans isomers was
formed.
Entry
Alcohol
(substrate)
Lewis acid
Products isolated
yield (%)
1
2
3
4
5
6a (1)a
6a (2)a
6a (3)a
6b (1)a
6c (1)a
HBF4
HBF4
HBF4
HBF4
HBF4
17a (87)
18a (76)
19a (50)
17b (55)
17c (55)
Another rationale, in agreement with Rychnovsky pro-
posal for trans selective Prins reactions, involves the
concept of ‘least motion pathway’ from a tight ion-pair
(D).9a,11 As stated in our previous note,2 the latter are
likely to be involved in a non-dissociative solvent like
CH2Cl2.
a The reactions were carried out at room temperature in dichloro-
methane, on 0.3–0.4 mmol of complex in the presence of 1.1 equiv of
alcohol and 1.0 equiv of Lewis acid.
When the reaction led to a tertiary 4-tetrahydropyranyl
cation, the formation of the tertiary chloride 7b was
highly trans selective (entry 10). The overall yield in ter-
tiary fluoride 8b was lowered by the regioselective for-
mation of alkene 12b through competitive elimination
(entries 12–14). The amount of elimination was slightly
higher when using BF3ÆOEt2 (entries 12/14). Concomi-
tantly, the cis/trans ratio increased as if the olefin was
mainly formed at the expense of the trans fluoride.12
The amount of 12b increased even more when dichloro-
methane was replaced by nitromethane (entry 13), a sol-
vent which has a medium dielectric constant and favors
the formation of loose ion pairs. However, there was no
more trapping of the cation by the fluoride anion. Only
the trapping by ethanol released in the reaction medium
during the reaction was detected. This argues in favor of
the model of tight ion pairs in the low polarity solvent.
ity of the double bond due to the electron-donating
effect of the methyl group cannot be neglected.14
Much to our disappointment, the reactions performed in
acetonitrile led to amides 15a and 16a in rather moder-
ate yields, but with a high selectivity (entries 16–17).
Regarding the use of allenes as internal traps (Table 2),
only homoallenic alcohol led to reactions having some
synthetic value. In this case, the final cation would lose
regioselectively a proton to give conjugated dienes (17–
19). In the case of allene 6c, with no gem dimethyl group
at C5, the proton would be abstracted from C5. To the
best of our knowledge, there are rather few examples of
Prins cyclization involving allenes as internal trap. Prins
cyclizations through TMSOTf-mediated cyclization
onto an activated silylated allene was shown to give con-
jugated dienes.16 Aza-Prins cyclization onto allenes has
also been reported by Hanessian.17 No cyclization prod-
uct was isolated from alcohols 6d and 6e.18 No 5-endo
ring closure was observed using allenic alcohol 6f.
The diastereomeric ratios of the halogenated products in
entries 10 and 14 are also in agreement with previous re-
ports on Prins reaction.13 In contrast to the secondary
cation, the tertiary one is more stable in a planar geom-
etry10 and the reaction would lead to the thermody-
namic product with the halogen atom in the axial
position.
In conclusion, protection of the triple bond as dicobalt
hexacarbonyl complex enabled the synthesis of 2-alky-
nyl tetrahydropyrans functionalized at C4 via Lewis
acid mediated-Prins cyclization in the presence of homo-
allylic alcohols. Both the yield and the diastereoselectiv-
ity were improved compared to the use of non-
complexed acid-sensitive substrates. The outcome of
the reaction depended on the nature of both the Lewis
acid and the solvent. When acetonitrile was used as
the solvent, a Prins–Ritter sequence led to amides in
moderate yields. The use of homoallenic alcohols led
to dienic products.
It can be noted that the substituent at the terminus of
the ‘formal’ triple bond has also a significant influence
upon the diastereoselectivity (entries 6, 8, and 9). Due
to the change in the hybridization state of the triple
bond carbons upon complexation, the bulk of the re-
mote substituent influences the diastereoselectivity sig-
nificantly. The diastereoselectivity should be influenced
by the preferred conformation around the equatorial
C–C bond adjacent to the complex. Bulky substituents
(R1) may hinder one face and destabilize the cis isomer.
The reactions where o-vinylphenols 5a–b were used as
partners for complexed acetal 1 led to rather disappoint-
ing results (Table 1, entry 15). This is probably due to
stereoelectronic factors. In the absence of the methyl
group geminated to the aromatic ring, there is no imped-
iment to the conjugation of the double bond with the
aromatic system. The double bond is pushed outward,
therefore, the preferred conformation of the p system
does not enable overlapping with the empty orbital at
the cationic carbon center in the oxocarbenium ion.
No cyclization was observed. However, due to the steric
effect of the methyl group that prevents conjugation and
makes the overlap possible, phenol 5b led to 14b but in
rather low yield. In addition, the increased nucleophilic-
References and notes
1. (a) Caffyn, A. J.; Nicholas, K. M.. In Comprehensive
Organometallic Chemistry II; Abel, E. W., Stone, F. G.,
Wilkinson, G., Hegedus, L. S., Eds.; Pergamon: Oxford,
1995; Vol. 12, Chapter 7.1 (b) Nicholas, K. M. Acc. Chem.
Res. 1987, 20, 207; (c) Green, J. R. Curr. Org. Chem. 2001,
5, 809; (d) Teobald, B. J. Tetrahedron 2002, 58, 4133; (e)
Alvaro, E.; De la Torre, M. C.; Sierra, M. A. Chem. Eur.
J. 2006, 12, 6403; (f) Fletcher, A. J.; Christie, S. D. R. J.
Chem. Soc., Perkin Trans. 1 2000, 1657.
2. Olier, C.; Gastaldi, S.; Christie, S. D. R.; Bertrand, M. P.
Synlett 2007, 423.
3. For the synthesis of fused ring systems see: (a) Tyrrell, E.;
Claridge, S.; Davis, R.; Lebel, J.; Berge, J. Synlett 1995,

7804
C. Olier et al. / Tetrahedron Letters 48 (2007) 7801–7804
714; (b) Berge, J.; Claridge, S.; Mann, A.; Muller, C.;
12. Stereochemical assignment was based on the chemical
shift difference between the diastereotopic protons at C6.
Dd values are far larger when the halogen atom is
equatorial (cis isomers) as compared to the trans ones.
Furthermore, the axial 19F nucleus is more shielded than
the equatorial one. As an example, d19F (/CFCl3) is:
ꢀ170.4 ppm in cis-8a, and ꢀ185.2 ppm in trans-8a, (very
close chemical shifts were observed for 10a and 11a); d19F
is: ꢀ121.6 ppm in cis-8b, and ꢀ153.0 ppm in trans-8b.
13. Epstein, O. L.; Rovis, T. J. Am. Chem. Soc. 2006, 128,
16480.
Tyrrell, E. Tetrahedron Lett. 1997, 38, 685; (c) Tyrrell, E.;
Tillett, C. Tetrahedron Lett. 1998, 39, 9535; (d) Tyrrell, E.;
Skinner, G. A.; Bashir, T. Synlett 2001, 1929; For
intermolecular reactions, see: (e) Krafft, M. E.; Cheung,
Y. Y.; Wright, C.; Cali, R. J. Org. Chem. 1996, 61, 3912.
4. Tetrahydropyran units are encountered in many natural
products, for selected examples, see: (a) Clarke, P. A.;
Santos, S. Eur. J. Org. Chem. 2006, 2045; (b) Tanaka, S.;
Isobe, M. Tetrahedron 1994, 50, 5633; (c) Hosokawa, S.;
Isobe, M. J. Org. Chem. 1999, 64, 37; For strategies
involving Prins reaction see also: (d) Vasconcellos, M. L.
A. A.; Miranda, L. S. M. Quim. Nova 2006, 29, 834; Chan,
K.-P.; Ling, Y. H.; Loh, T.-P. Chem. Commun. 2007, 939;
14. For a nucleophilicity scale of carbon-centered nucleophiles
in intermolecular reactions see: Mayr, H.; Kempf, B.;
Ofial, A. R. Acc. Chem. Res. 2003, 36, 66, N = 2.35 for a-
methyl styrene; N = 0.78 for styrene.
´
´
´
(e) Leon, L. G.; Miranda, P. O.; Martın, V. S.; Padron, J.
´
I.; Padron, J. M. Bioorg. Med. Chem. Lett. 2007, 17, 3087.
15. Typical experimental procedure: Hexacarbonyl[l-g4-{5,6-
dihydro-5,5-dimethyl-2-(2-phenylethynyl)-3-(prop-1-en-2-
yl)-2H-pyran}] dicobalt (18b). 2,2,5-Trimethylhexa-3,4-
dien-1-ol (43 mg, 0.31 mmol) and HBF4 (38 lL,
0.28 mmol) were successively added, under inert atmo-
sphere, to a solution of 2 (131 mg, 0.28 mmol) in freshly
distilled dichloromethane (0.9 mL). The reaction mixture
was stirred for 30 min at room temperature. The reaction
was quenched by the addition of water (5 mL) and the
aqueous layer was extracted twice with dichloromethane.
The organic layer was washed with brine, dried over
MgSO4, and concentrated under reduced pressure. Liquid
chromatography on silica gel (pentane) afforded 18b
(99 mg, 0.21 mmol, 76%). 1H NMR (300 MHz, CDCl3):
d 0.94 (s, 3H), 1.10 (s, 3H), 1.89 (s, 3H), 3.36 (d,
J = 11.1 Hz, 1H), 3.63 (d, J = 11.3 Hz, 1H), 4.86 (s, 1H),
4.97 (s, 1H), 5.76 (s, 1H), 5.86 (s, 1H), 7.33 (m, 3H), 7.45
(m, 2H). 13C NMR (75 MHz, CDCl3): d 21.7 (CH3), 24.9
(CH3), 27.3 (CH3), 32.5 (C), 71.5 (CH2), 73.7 (CH), 93.9
(C), 98.0 (C), 114.9 (CH2), 127.8 (CH), 128.9 (CH), 129.9
(CH), 133.9 (CH), 137.2 (C), 138.9 (C), 140.6 (C), 199.8
(CO). HRMS (TOF MS ES+); MH+, Calcd [M+1] for
C24H20O7Co2: 538.9945; found: 538.9939.
5. For the formation of 2-aryl-tetrahydropyrans from arene-
chromiumtricarbonyl complexes, through a similar path-
way see: Davies, S. G.; Donohoe, T. J.; Lister, M. A.
Tetrahedron: Asymmetry 1991, 2, 1089.
6. As examples, the axial pseudo-propargylic protons give a
4
ddd (J = 11.3, 1.7 Hz, JHF = 1.7 Hz) in cis-8a, a dd
(J = 11.1, 1.9 Hz) in cis-7a, and a dd (J = 11.3, 1.9 Hz) in
cis-9a. The axial protons at C4 are characterized by dtt
(2JHF = 49.1 Hz, JHH = 11.0, 4.9 Hz) in 8a, tt patterns
(J = 11.4–11.9, 4.3–5.1 Hz) in 7a and 9a. The equatorial
proton at C4 in trans-8a exhibits a t (J = 3.0 Hz).
1
7. Only H NMR yields, based on the characteristic signals
of protons at the stereocenters, were determined using
pentamethylbenzene as internal standard. The isolation of
the pure products was made difficult due to the low
polarities of all products.
8. The characteristic signals of 4-ethoxybut-3-en-2-one are:
1H NMR,
d 5.60 (d, J = 12.8 Hz, 1H), 7.55 (d,
J = 12.8 Hz, 1H); 13C NMR, d 107.6 (CH), 176.7 (CH),
197.8 (C@O). The characteristic signals of 3-ethoxybut-2-
enal are: 1H NMR, d 5.39 (d, J = 7.6 Hz, 1H), 9.75 (d,
J = 7.6 Hz, 1H); 13C NMR, d 105.2 (CH), 162.8 (C), 190.9
(C@O). Their identification was further confirmed by
performing a blank experiment by reacting 10 with HBF4
in dichloromethane.
16. Cho, Y. S.; Karupaiyan, K.; Kang, H. J.; Pae, N. A.; Cha,
J. H.; Koh, H. Y.; Chang, M. H. Chem. Commun. 2003,
2346.
9. (a) Jasti, R.; Rychnovsky, S. D. Org. Lett. 2006, 8, 2175;
(b) Jasti, R.; Anderson, C. D.; Rychnovsky, S. D. J. Am.
Chem. Soc. 2005, 127, 9939.
10. Alder, R. W.; Harvey, J. N.; Oakley, M. T. J. Am. Chem.
Soc. 2002, 124, 4960.
11. (a) Jasti, R.; Vitale, J.; Rychnovsky, S. D. J. Am. Chem.
Soc. 2004, 126, 9904; (b) Zhang, Y. D.; Reynolds, N. T.;
Manju, K.; Rovis, T. J. Am. Chem. Soc. 2002, 124, 9720.
17. Hanessian, S.; Tremblay, M.; Marzi, M.; del Valle, J. R. J.
Org. Chem. 2005, 70, 5070.
18. Tertiary homoallenic alcohol 6e might be too crowded to
trap the very first cationic intermediate. No product could
be isolated from the reaction of complex 1 with this
alcohol.