Formation and reactivity of new Nicholas-Ferrier pyranosidic cations: Novel access to oxepanes via a 1,6-hydride shift/cyclization sequence

Article abstract of DOI:10.1039/c0cc00586j

Pyranosidic allylic (Ferrier) cations that share dicobalt hexacarbonyl propargyl (Nicholas) stabilization at C-1 display a remarkable reactivity leading to either substituted oxepanes or 3-C-branched pyranosides, depending on the substituent at O-6.

Full text of DOI:10.1039/c0cc00586j

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Formation and reactivity of new Nicholas–Ferrier pyranosidic cations:
novel access to oxepanes via a 1,6-hydride shift/cyclization sequencew
´
Ana M. Gomez,* Fernando Lobo, Daniel Perez de las Vacas, Serafın Valverde and
J. Cristobal Lopez*
´
´
´
´
Received 25th March 2010, Accepted 25th June 2010
DOI: 10.1039/c0cc00586j
Pyranosidic allylic (Ferrier) cations that share dicobalt hexa-
carbonyl propargyl (Nicholas) stabilization at C-1 display a
remarkable reactivity leading to either substituted oxepanes or
3-C-branched pyranosides, depending on the substituent at O-6.
oxepanes through a reaction sequence that involves: (i) 1,6
hydride transfer; (ii) two-step electrophilic addition (AdE);9
and (iii) pyranosidic ring opening.
The precursors for the Nicholas–Ferrier cations 5 were
derivatives 9, prepared from enynes 8.10 The latter were readily
obtained from D-glucals 6,11,12 via lactones 713 (Scheme 2).
The reaction of tri-O-benzyl glucal 9a with allyltrimethyl-
silane (3 equiv.) in CH2Cl2 at ꢀ20 1C, in the presence of BF3ꢁEt2O
(1.2 equiv.), yielded 3-C-allyl derivative 10 (49%) accompanied
by oxepane 11 (which had not incorporated the allyl moiety,
8%) (Scheme 3a). When the reaction was carried out in the
absence of allyltrimethylsilane, oxepane 11 was obtained in
34% yield, although the major reaction product was 1,6-anhydro
derivative 12 (63%) (Scheme 3b). This (11 : 12) proportion
could be reversed if molecular sieves were added to the
reaction media, then oxepane 11 could be isolated in 61%
yield (Scheme 3c).z
The Nicholas reaction1 and the Ferrier (I) rearrangement2 are
two well-known,3 useful, synthetic transformations,4,5 which
have in common the intermediacy of cationic species such as 1
and 3, respectively (Scheme 1). The reaction of these cations
with nucleophiles (NuH) leads to propargylated derivatives
(e.g. 2) or allylic glycosides (e.g. 4), respectively. Recently,
some examples of novel Nicholas cations have appeared in the
literature, such as a Nicholas dehydrotropylium ion,6 or
Nicholas dipoles derived from alkyne–cycloalkane dicobalt
complexes.7 In this context, we have studied the behavior of
Nicholas pyranosidic oxocarbenium ions, which displayed an
unexpected synthetic behavior.8 Along this line, we became
interested in the study of ‘‘Nicholas–Ferrier’’ cations, e.g. 5,
since they could display new reactivity regarding unsaturated
carbohydrate derivatives and/or Nicholas cations.
We now report that the reactivity of Nicholas–Ferrier
cations, e.g. 5, can be modulated by changes in the protecting
group at O-6, thus leading to 3-C-branched pyranosides
(by reaction with an external nucleophile), or to functionalized
Scheme 2 Synthesis of ‘‘Nicholas–Ferrier’’ cation precursors, 9.
Scheme
1 Dicobalt hexacarbonyl coordinated propargyl cation
(Nicholas cation) (1), Ferrier allylic cation (3), and Nicholas–Ferrier
cation (5).
´
´
Instituto de Quımica Organica General (CSIC), Juan de la Cierva 3,
28006 Madrid, Spain. E-mail: anagomez@iqog.csic.es,
clopez@iqog.csic.es
w Electronic supplementary information (ESI) available: Experimental
details and NMR spectra. See DOI: 10.1039/c0cc00586j
Scheme 3 Reaction of 9a with BF3ꢁEt2O, in the presence or absence
of allyltrimethylsilane.
ꢂc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 6159–6161 | 6159

View Article Online
Table 1 Synthesis of oxepanes 20, from differently substituted
glycals, 19a–f
Starting
Entry material Ar
Products (Isolated yields %)
i
19a
12 (79)
20a (traces)
Scheme 4 Proposed reaction pathways for the formation of 11 and 12.
The structural assignment of compound 11 was based on the
inspection of its NMR spectra (see Supplementary Informationw).
The stereochemistries at C-2 and C-3 (oxepane numbering)
were assigned on the basis of a large J2,3 = 10.2 Hz coupling
constant, and observed NOEs between H-2, H-7ax and H-4ax
(see Scheme 3).
ii
19b
12 (27)
20b (38)b
iii
iv
19c
19d
12 (30)
12 (43)
20c (62)
20d (49)
From these results, we were able to advance a mechanistic
rationale that accounted for the formation of compounds 11
and 12 (Scheme 4). The first step would be a 1,6-hydride
transfer14,15 (13 - 14, Scheme 4) to generate a highly reactive
oxocarbenium ion 14. The latter could then follow two sets of
transformations: (a) it could experience hydrolysis to a 6-OH
derivative 15, which upon protonation of the activated enyne
moiety will form a pyranosidic Nicholas oxocarbenium ion8
16, that will cyclize to 12; or (b) it could experience a two-step
AdE reaction9,16 to give 17, and thence 18, whose ring opening
would finally lead to oxepane 11.
v
19e
19f
12 (21)
12 (15)
20e (54)
20f (69)
According to our proposed reaction pathway, the formation
of oxepane 11, over 1,6-anhydro derivative 12, could be
favored if hydrolysis of the oxocarbenium ion in intermediate
14 is decelerated, as was indeed observed when the reaction
was carried out in the presence of freshly activated, crushedy
4 A molecular sieves (which excluded water needed for
the hydrolysis of 14 from the reaction media, compare in
Scheme 3, entries b and c).
vi
a
Carbohydrate numbering is shown, to highlight the atom correlation
b
between 19 and 20. Compound 15 was also isolated (19% yield).
In order to evaluate the scope of this transformation, aiming
at the preparation of functionalized oxepanes, we prepared
compounds 19a–f,z differing in the nature of the aromatic
group at O-6, and submitted them to treatment with BF3ꢁEt2O
(CH2Cl2, ꢀ20 1C, 4 A molecular sieves, Table 1). This study
was also of interest from a mechanistic standpoint, since
electron donating groups at the aromatic residue that will
favor hydride transfer (13 - 14), will render the ensuing
oxocarbenium ions less electrophilic, whereas electron withdrawing
groups that would favor the electrophilic addition step (14 - 17)
will make the hydride transfer step more difficult.
yields of oxepanes 20d and 20e (Table 1, entries iv and v,
respectively).
A naphthyl derivative (19c) displayed a behavior similar to
that of the parent phenyl derivative, 9b, (compare Table 1,
entry iii, with Scheme 3c). According to these results, electron
withdrawing substituents in the aromatic ring, which render
oxocarbenium intermediates, e.g. 14, more electrophilic (therefore
facilitating the AdE step) seem to favor oxepane formation.
Two new stereogenic centers are created in this process.
The stereogenic center at C-2 (carbohydrate numbering, see
Table 1) in oxepanes 11 and 20 is dictated by geometric
restrictions in the approach of the C-6 substituent towards
C-2 (Scheme 4). The stereogenic center at C-60 (carbohydrate
numbering, see Table 1) is the result of a preferred rotamer of
the oxocarbenium ion (14), which locates the aryl residue away
from the bulky dicobalt hexacarbonyl moiety in the transition
state leading to cyclization (14 - 17, Scheme 4).
From the results in Table 1, it became apparent that substrates
with electron donating substituents in the aromatic residue did
not favor formation of oxepanes 20 (Table 1, entries (i, ii)).
Conversely, p-fluoro derivative 19f furnished a significant
yield of oxepane 20f (69%, Table 1, entry (vi)). p-Iodo and
p-bromo derivatives, 19d and 19e, respectively, gave moderate
ꢂc
This journal is The Royal Society of Chemistry 2010
6160 | Chem. Commun., 2010, 46, 6159–6161

View Article Online
Table 2 Reaction of 9b with carbon nucleophiles
Notes and references
z The crucial role played by the dicobalt hexacarbonyl complex in
these reactions (Scheme 3) was proved when compound 8a (devoid of
dicobalt hexacarbonyl) was treated with BF3ꢁEt2O to yield a complex
mixture of compounds.
y When molecular sieves, 4 A beads, were used, instead of powder
activated 4 A molecular sieves, the isolated yields of isolated 11 and 12
were 45% and 34%, respectively.
z Compounds 19a–f were readily obtained from 8b, by a reaction
sequence that involved: desilylation, alkylation, and cobaltation.
Entry
Nucleophile
t/min
Product
Yield (%)
70
i
30
21a
1 Reviews: J. R. Green, Eur. J. Org. Chem., 2008, 6053;
B. J. Teobald, Tetrahedron, 2002, 58, 4133; J. R. Green, Curr.
Org. Chem., 2001, 5, 809; K. M. Nicholas, Acc. Chem. Res., 1987,
20, 207.
ii
50
21b
59
2 Reviews: R. J. Ferrier and H. O. Hoberg, Adv. Carbohydr. Chem.
Biochem., 2003, 58, 55; R. J. Ferrier and O. A. Zubkov, Org.
React., 2003, 62, 569; R. J. Ferrier, Top. Curr. Chem., 2001, 215,
1535; R. J. Ferrier, Adv. Carbohydr. Chem. Biochem., 1969, 58, 199.
3 See for example: J. J. Li, in Name Reactions, Springer, Berlin,
iii
iv
20
80
21c
21d
57
93
3rd edn, 2006; L. Kurti and B. Czabo, in Strategic Applications of
¨
Named Reactions in Organic Synthesis, Elsevier Academic Press,
Burlington, 2005.
4 Nicholas reaction: D. D. Dı
Synlett, 2007, 343.
5 Ferrier (I) rearrangement: B. Fraser-Reid and J. C. Lo
Org. Chem., 2009, 13, 532; J. C. Lopez, A. M. Gomez, S. Valverde
az, J. M. Betancort and V. S. Martın,
´ ´
´
pez, Curr.
´
´
Since the formation of oxepanes seemed to be triggered by
the presence of benzyl-type substituents at O-6, it could be
eliminated with the presence of a different substituent at O-6.
Accordingly, the reaction of 6-O-triisopropylsilyl (TIPS) derivative
9b, with carbon nucleophiles (3 equiv.) in CH2Cl2 at ꢀ20 1C,
in the presence of BF3ꢁEt2O (1.2 equiv.) led, in a stereo-
controlled manner, to 3-deoxy-3-C-hex-1-enitol derivatives 21,
in moderate to good yields (Table 2).
and B. Fraser-Reid, J. Org. Chem., 1995, 60, 3851.
6 S. Amiralaei and J. R. Green, Chem. Commun., 2008, 4971.
7 E. A. Allart, S. D. R. Christie, G. J. Pritchard and M. R. J.
Elsegood, Chem. Commun., 2009, 7339; S. D. R. Christie,
R. J. Dacoile, M. R. J. Elsegood, R. Fryatt, R. C. F. Jones and
G. J. Pritchard, Chem. Commun., 2004, 2474.
´ ´
8 A. M. Gomez, C. Uriel, S. Valverde and J. C. Lopez, Org. Lett.,
2006, 8, 3187.
9 W. A. Smit, R. Caple and I. P. Smoliakova, Chem. Rev., 1994, 94,
2359.
10 F. E. McDonald, H. Y. H. Zhu and C. R. Holmquist, J. Am.
Chem. Soc., 1995, 117, 6605.
11 H. Imai, T. Oishi, T. Kikuchi and M. Hirama, Tetrahedron, 2000,
56, 8451.
12 N. V. Bovin, S. Zurabyan and A. Khorlin, Carbohydr. Res., 1981,
98, 25.
The reaction of 9b with allyltrimethylsilane furnished
3-C-branched derivative 21a (Table 2, entry (i)). Likewise,
the reaction of 9b with N-methylindole, N-methylpyrrole, and
furan yielded compounds 21b, 21c, and 21d, respectively
(Table 2, entries ii, iii and iv, respectively).
These transformations were completely regioselective,
following a pattern commonly observed for allylic Nicholas
cations, in which the nucleophile reacts at the terminus remote
from the organometallic substituent,17 rather than a Ferrier-
type behavior, where the nucleophile normally enters at C-1.2
The configuration at C-3 in compounds 21a–d could be
established based on the observed large J3,4 coupling constants
in 1H NMR. The stereochemistry of the C-3 branch is the
result of a preferred anti approach of the incoming nucleophile
with respect to the substituent at C-4.18
13 P. Rollin and P. Sinay, Carbohydr. Res., 1981, 98, 139.
¨
14 For related 1,6-hydride transfers on glycals, see: M. A. Brimble
and T. J. Brenstrum, J. Chem. Soc., Perkin Trans. 1, 2001, 1612;
M. A. Brimble and T. J. Brenstrum, Tetrahedron Lett., 2000, 41,
1107; A. L. J. Byerley, A. M. Kenwright, C. W. Lehmann,
J. A. Hugh MacBride and P. G. Steel, J. Org. Chem., 1998, 93, 163.
15 For
D. D. Dı
2006, 12, 2593, and references cited therein.
a
related 1,5-hydride transfer to Nicholas cations, see:
´
az, M. A. Ramırez and V. S. Martın, Chem.–Eur. J.,
´
´
16 G. S. Mikaelian, A. S. Gybin, W. A. Smit and R. Caple,
Tetrahedron Lett., 1985, 26, 1269.
17 (a) E. Alvaro, M. C. de la Torre and M. A. Sierra, Chem.–Eur. J.,
2006, 12, 6403; (b) E. Alvaro, M. C. de la Torre and M. A. Sierra,
Chem. Commun., 2006, 985; (c) S. Padmanabhan and
K. M. Nicholas, Tetrahedron Lett., 1982, 23, 2555.
In summary, novel Nicholas–Ferrier pyranosidic cations,
with benzyl-type substituents at O-6, undergo a synthetically
useful, stereocontrolled, transformation19 to substituted
oxepanes,20 which involves 1,6 hydride shift/cyclization/ring
opening. Work is currently underway with derivatives of
type 9 to further explore its reactivity and potential synthetic
applications.
18 (a) W. Damm, B. Giese, J. Hartung, T. Hasskerl, K. Houk,
O. Huter and H. Zipse, J. Am. Chem. Soc., 1992, 114, 4067;
mez, S. Mantecon, S. Valverde and J. C. Lopez,
´ ´ ´
¨
(b) A. M. Go
J. Org. Chem., 1997, 62, 6612.
19 A related transformation involving a 1,5-hydride shift/cyclization
process has recently appeared: K. M. McQuaid and D. Sames,
J. Am. Chem. Soc., 2009, 131, 402.
20 Reviews: (a) V. Piccialli, Synthesis, 2007, 2585; (b) M. C. Elliott,
J. Chem. Soc., Perkin Trans. 1, 2002, 2301. See also: N. V. Ganesh,
S. Raghothama, R. Sonti and N. Jayaraman, J. Org. Chem., 2010,
75, 215, and references cited therein.
We acknowledge the support of Ministerio de Ciencia e
´
Innovacion (Grants CTQ-2006-C03, and CTQ2009-10343),
and Comunidad de Madrid (Grant S2009/PPQ-1752). F.L. is
´
grateful to Consejo Superior de Investigaciones Cientıficas
(CSIC) for a predoctoral scholarship (JAE program).
ꢂc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 6159–6161 | 6161