Fluoride ion-promoted α-ketol rearrangement during unmasking of silyl-protected medium-ring dihydroxy ketones

Article abstract of DOI:10.1002/adsc.200404002

A new strategy for implementing α-ketol rearrangements under mild conditions is presented. The reactants are mono- and disilylated stereoisomers of α,α′- and α,β-dihydroxycycloheptanones and -cyclooctanones. Compounds of this class experience ready deprotection upon treatment with tetrabutyl-ammonium fluoride. In certain examples, this process is accompanied by structural isomerization. Since the product diols are stable to the reaction conditions, these chemical changes have been attributed to kinetically controlled events following transient alkoxide generation resulting from cleavage of the O-Si bond proximal to the carbonyl group. The experimental findings are evaluated against a backdrop of calculated (MM3) steric energies of the silicon-free products and their response to equilibration under strongly basic conditions.

Full text of DOI:10.1002/adsc.200404002

COMMUNICATIONS
Fluoride Ion-Promoted a-Ketol Rearrangement during
Unmasking of Silyl-Protected Medium-Ring Dihydroxy Ketones
Ryan E. Hartung, David G. Hilmey, Leo A. Paquette*
[a]
Evans Chemical Laboratories, The Ohio State University, Columbus, Ohio 43210, USA
Fax: (þ1)-614-292-1685, e-mail: paquette.1@osu.edu
Received: January 5, 2004; Accepted: April 2, 2004
Dedicated to Joe P. Richmond on the occasion of his 60th birthday.
Abstract: A new strategy for implementing a-ketol
rearrangements under mild conditions is presented.
The reactants are mono- and disilylated stereoisom-
ers of a,a’- and a,b-dihydroxycycloheptanones and
-cyclooctanones. Compounds of this class experience
ready deprotection upon treatment with tetrabutyl-
ammonium fluoride. In certain examples, this proc-
ess is accompanied by structural isomerization. Since
the product diols are stable to the reaction condi-
tions, these chemical changes have been attributed
to kinetically controlled events following transient
alkoxide generation resulting from cleavage of the
trast, heteroatomic substituent effects in the vicinity of
the carbonyl group appear not to have been given prior
attention. The recent availability of monosilylated me-
dium-ring a,a’-dihydroxycycloalkanones in our labora-
tory[11] has prompted a probe of the manner in which
the individual cis- and trans-isomers react when desily-
lated. The generation of tautomeric a-ketol products
has been found to occur under conditions sufficiently
mild[1d,12] that the resulting dihydroxy ketones are not
concurrently equilibrated. These observations consti-
tute a previously unrecognized means for implementing
a-ketol rearrangements.
À
O Si bond proximal to the carbonyl group. The ex-
perimental findings are evaluated against a backdrop
of calculated (MM3) steric energies of the silicon-
free products and their response to equilibration un-
der strongly basicconditions.
Keywords: desilylation; fluoride ion; a-ketols; me-
dium-sized rings; MM3 calculations; rearrangement
We first addressed the direct desilylation of 4 and its
diprotected congener 5 with tetra-n-butylammonium
fluoride (TBAF) in tetrahydrofuran at room tempera-
Many a-hydroxy ketones have the capacity for conver- ture for 30 min. In the event, the dihydroxy ketones 6
sion to an isomeric a-ketol. Catalysis by a base,[1] inor- and 7 were formed in an identical ratio of 2:1 in these ex-
ganicsalt, [2] Br˘nsted acid,[3] or Lewis acid[4] is generally amples. The structural assignments to the products are
required. Two mechanistic pathways are possible de- based upon direct spectral comparisons with authentic
pending on structural features. When the hydroxy group samples. The trans relationship of the hydroxy substitu-
forms part of a tertiary carbinol, a 1,2-alkyl or aryl shift ents in 7 presumably stems from operation of a suprafa-
asdepicted for
operates.[5] Secondary a-keto car- cial 1,2-hydride shift in the alkoxide intermediate(s).
binols can also isomerize via an enediol intermediate
typified by 3.[6 10] In both cases, the rearrangement is re-
versible, such that thermodynamic control is operative
and the more stable isomer results. If the enthalpically
advantaged member of the pair is already in hand, a
chemical transformation will not likely operate under
the predescribed conditions provided that the energy
difference is sufficiently great.
Extensive attention has been accorded to acyclic, cy-
cis-Isomer 8 exhibits less of a driving force to isomer-
clic, bridged, and steroidal a-ketols, such that ring-size ize. Under entirely comparable reaction conditions, 9
effects and the impressively high levels of attainable se- and 10 are formed with a distribution of 10:1. Product
lective bond migration are well appreciated.[5] In con- formation in the case of both cycloheptanone stereo-
Adv. Synth. Catal. 2004, 346, 713 716
DOI:10.1002/adsc.200404002
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
713

Ryan E. Hartung et al.
COMMUNICATIONS
isomers appears to be under kineticcontrol, as 6, 9, and
In line with expectations based on our earlier ration-
10 undergo no further chemical change when reexposed ale, neither 19 nor 21 experienced rearrangement
to TBAF for the same time periods.
when desilylated. Also, resubmission of 9, 10, 15, 17,
18, and 20 to the standard protocol resulted in no further
chemical change. When control experiments were con-
ducted instead with methanolic potassium tert-butoxide
in an effort to simulate more conventional a-ketol rear-
rangement conditions, both 6 and 7 were independently
converted into a 2:1 mixture of the trans tautomers with
the predominant component being 6.
Beyond this, the 2,3-dihydroxy derivatives 11 and 12
were found to undergo desilylation smoothly without
rearrangement to furnish exclusively 7. When the reac-
tivity of 13 was likewise scrutinized, high-yield conver-
sion to 10 (95% yield) was noted. Since the desilylations
of 11 and 13 do not proceed via an a-ketol alkoxide, the
absence of isomerization is not unexpected. This mech-
anisticscenario does not apply to 12, thereby signaling
the lack of a driving force for the isomerization of
trans-2,3-dihydroxycycloheptanone to its trans-2,7-di-
hydroxy tautomer.
Noteworthily, 9 was equilibrated with 10 (ratio 8:1),
while 17 was equilibrated with 18 (1:1 ratio). Compara-
ble treatment of 15 and 20 in independent experiments
gave rise to a 2:1 mixture of 15 and 20 in both cases.
In light of the conformational and reactivity differen-
ces known to distinguish medium rings of different size,
several structurally related cyclooctanone derivatives
were also studied. Ketone 14 was the lead-off eight-
membered candidate to be treated with fluoride ion.
Unlike 4, this a-ketol provided only 15 and exhibited
no detectable tendency to isomerize. This adherence
to a non-isomerative pathway was not shared by the cor-
responding cis-isomer 16. In this instance, activation
with TBAF led reproducibly to a 2:1 mixture of 17
and 18. As before, systematic spectroscopic compari-
sons were made to achieve unequivocal definition of
stereochemistry and structure.
To gain a different vantage point regarding this mat-
ter, the global minimum energy conformations and the
associated steric energies were calculated for each dihy-
droxy ketone by the MM3 method using 1500 Monte
Carlo simulations in MacroModel version 5.0. For con-
venient reference, the 7- and 8-membered ring systems
are grouped separately in Figures 1 and 2, respectively.
As concerns the dihydroxycycloheptanones, two impor-
tant relationships can be readily derived from these
data: a) the cis- and trans-isomers in both the 2,3- and
714
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2004, 346, 713 716

Fluoride Ion-Promoted a-Ketol Rearrangement
COMMUNICATIONS
Experimental Section
For Desilylation of Mono-TBS Protected Substrates;
General Procedure A
The reactant (1 mmol) was dissolved in dry THF (25 mL),
treated with a 1 M solution of TBAF in THF (1.05 mmol),
and stirred for 30 min prior to solvent evaporation. The residue
was directly submitted to chromatography on silica gel (elution
with 1:1 hexane/ethyl acetate).
For Desilylation of Di-TBS Protected Substrates;
General Procedure B
Figure 1. Global minimum energy conformations of the 2,3-
and 2,7-dihydroxycycloheptanones.
As above, except that 2.5 equivs. of TBAF were added.
Control Experiments Involving TBAF
Identical to procedure A.
Control Experiments Involving Potassium tert-
Butoxide
The reactant (1 mmol) was dissolved in methanol (25 mL),
treated with potassium tert-butoxide (1.1 mmol), and stirred
for 30 min prior to solvent evaporation. The residue was direct-
ly submitted to chromatography on silica gel (elution with 1:1
hexane/ethyl acetate).
Figure 2. Global minimum energy conformations of the 2,3-
and 2,8-dihydroxycyclooctanones.
Spectral Data of Products
2,7-disubstituted series are essentially isoenergetic; and
b) the 2,7-dihydroxy arrangement is thermodynamically
favored in both substrates irrespective of diol stereo-
chemistry. Intramolecular hydrogen bonding is availa-
For 6: yellowish oil; IR (neat): n¼3414, 1710, 1452, 1270 cmÀ1
;
1HNMR (300 MHz, CDCl3): d¼4.25 (dd, J¼9.8, 2.9 Hz, 2H),
3.10 (s, 2H), 2.12 2.04 (m, 2H), 1.98 1.90 (m, 2H), 1.79
ble to both structural classes, with the 7-membered 1.25 (series of m, 4H); 13C NMR (75 MHz, CDCl3): d¼214.2,
77.4 (2C), 32.2 (2C), 26.9 (2C); ES HRMS: m/z (MþNa)þ
ring possessing sufficient flexibility to accommodate
the necessary topological arrangements. All hydroxy
groups are projected pseudoequatorially.
The energeticinterrelationships determined in the cy-
clooctyl series are entirely parallel (Figure 2). There is
little difference between configurational diastereomers,
and the 2,8-dihydroxycyloooctanones are seen to be less
sterically disadvantaged than the 2,3 isomers.
calcd.: 167.0678; obsd.: 167.0674.
For 7: colorless oil;IR(neat): n¼3418, 1703, 1647, 1451 cmÀ1
;
1H NMR (300 MHz, CDCl3): d¼4.45 (d, J¼1.5 Hz, 1H), 4.26
4.22 (m, 1H), 2.75 2.69 (m, 1H), 2.53 2.46 (m, 1H), 2.23 2.18
(m, 1H), 1.99 1.93 (m, 1H), 1.88 1.79 (m, 2H), 1.74 1.61 (m,
2H); 13C NMR (75 MHz, CDCl3): d¼212.2, 80.8, 73.1, 41.0,
34.2, 23.0, 22.9; ES HRMS: m/z (MþNa)þ calcd.: 167.0678;
obsd.: 167.0669.
Our global experimental results reflect a tendency on
For 9: white solid, mp 878C; IR (neat): n¼3406, 1711,
1
the part of 4 and 8 to undergo structural isomerization to 1451 cmÀ1; H NMR (300 MHz, CDCl3): d¼4.41 (dd, J¼7.6,
4.5 Hz, 2H), 3.39 (br s, 2H), 2.06 2.00 (m, 2H), 1.83 1.76 (m,
2H), 1.69 1.59 (m, 4H); 13C NMR (75 MHz, CDCl3): d¼
214.9, 74.9 (2C), 33.7 (2C), 24.8 (2C); ES HRMS: m/z (Mþ
Na)þ calcd.: 167.0678; obsd.: 167.0684.
some degree upon treatment with fluoride. The process
operates more extensively in the trans series than the cis.
The converse is true for the related cyclooctanones.
While trans-ketone 14 exhibits no proclivity for conver-
sion to the 2,3-isomer, cis-derivative 16 leads only to
modest amounts of 18. While the rearrangement is ster-
eospecific in all cases, the calculations do not corrobo-
rate perfectly with the experimental results.
For 15: white solid, mp 78 798C; IR (neat): n¼3419, 1700,
1
1653 cmÀ1; H NMR (500 MHz, CDCl3): d¼4.62 (dd, J¼6.0,
2.9 Hz, 2H), 2.47 2.40 (m, 2H), 2.09 2.01 (m, 2H), 1.79
1.71 (m, 2H), 1.60 1.54 (m, 2H), 1.39 1.31 (m, 2H) (OH not
seen); 13C NMR (125 MHz, CDCl3): d¼218.7, 75.7 (2C), 33.5
(2C), 24.9 (2C), 21.0; ES HRMS: m/z (MþNa)þ calcd.:
181.0835; obsd.: 181.0837.
Adv. Synth. Catal. 2004, 346, 713 716
asc.wiley-vch.de
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
715

Ryan E. Hartung et al.
COMMUNICATIONS
For 17: white solid, mp 120 1218C; IR (neat): n¼3693,
1 1977, 893; c) I. Elphimoff-Felkin, B. Tschoubar, C. R.
Hebd. Seance Acad. Sci. 1954, 237, 1425; d) H. Ziehe,
R. Wartchow, H. B¸tenschˆn, Eur. J. Org. Chem. 1999,
64, 823.
1
1705, 1246 cmÀ1; H NMR (300 MHz, CDCl3): d¼4.41 (dd,
J¼7.7, 3.7 Hz, 2H), 3.26 (s, 2H), 2.33 2.27 (m, 2H), 1.97
1.90 (m, 2H), 1.84 1.66 (m, 4H), 1.48 1.40 (m, 2H);
13C NMR (75 MHz, CDCl3): d¼218.2, 74.1 (2C), 33.4 (2C),
24.6, 21.9 (2C); ES HRMS: m/z (MþNa)þ calcd.: 181.0835;
obsd.: 181.0829.
[2] a) I. Elphimoff-Felkin, G. LeNy, B. Tschoubar, Bull. Soc.
Chim. Fr. 1958, 522; b) E. Keinan, Y. Mazur, J. Org.
Chem. 1978, 43, 1020; c) T. Ishiguro, Y. Kondo, T. Take-
moto, Tetrahedron Lett. 1975, 315; d) G. Appendino, J.
Jakupovic, G. Cravotto, M. Varese, Tetrahedron Lett.
1994, 35, 6547.
[3] H. Brunner, F. Stˆhr, Eur. J. Org. Chem. 2000, 2777.
[4] a) X. Creary, P. A. Inocencio, T. L. Underiner, R. Kostri-
min, J. Org. Chem. 1985, 50, 1932.
[5] L. A. Paquette, J. E. Hofferberth, Org. React. 2003, 62,
477.
[6] L. A. Paquette, P. C. Lobben, J. Org. Chem. 1998, 63,
5604 and relevant references cited therein.
For 18: white solid, mp 798C; IR (neat): n¼3614, 1477,
1236 cmÀ1
;
1H NMR (300 MHz, CDCl3): d¼4.03 3.99 (m,
2H), 3.96 (d, J¼2.5 Hz, 1H), 2.37 (s, OH), 2.10 2.04 (m,
2H), 1.92 1.80 (m, 4H), 1.71 1.34 (m, 4H); 13C NMR
(75 MHz, CDCl3): d¼77.6, 76.1 (2C), 32.6 (2C), 29.1, 22.9
(2C); ES HRMS: m/z (MþNa)þ calcd.: 183.0991; obsd.:
183.0993.
For 20: colorless oil; IR (neat): n¼3394, 1703, 1478 cmÀ1
;
1H NMR (300 MHz, CDCl3): d¼4.28 (t, J¼8.2 Hz, 1H), 3.38
(d, J¼8.7 Hz, 1H), 3.31 3.26 (m, 1H), 3.18 (s, 2H), 2.72 2.66
(m, 1H), 2.51 2.44 (m, 1H), 2.11 2.04 (m, 1H), 1.93 1.80
(m, 2H), 1.67 1.57 (m, 2H), 1.54 1.47 (m, 1H), 1.02 0.94
(m, 1H); 13C NMR (75 MHz, CDCl3): d¼213.7, 79.7, 77.6,
41.6, 29.2, 28.2, 21.0, 20.7; ES HRMS: m/z (MþNa)þ calcd.:
181.0835; obsd.: 181.0841.
[7] M. J. Dominguez, E. Mˆssner, M. C. de la Torre, B. Ro-
driguez, Tetrahedron 1998, 54, 14377.
[8] L. J. Benjamin, G. Adamson, L. N. Mander, Heterocycles
1999, 50, 365.
[9] J. D. White, N. S. Cutshall, T.-S. Kim, H. Shin, J. Am.
Chem. Soc. 1995, 117, 9780.
[10] A. J. Hall, D. Ferreira, D. G. Roux, J. Chem. Soc. Perkin
Trans. 1 1980, 1205.
Acknowledgements
[11] L. A. Paquette, R. E. Hartung, J. E. Hofferberth, I. Vilo-
tijevic, J. Yang, J. Org. Chem. 2004, 69, in press.
[12] It is well recognized that anionicaly accelerated rear-
rangements are capable of proceeding at significantly
lower temperatures: a) S. R. Wilson, Org. React. 1993,
43, 93; b) L. A. Paquette, Tetrahedron 1997, 53, 7403.
This investigation was made possible by the receipt ofunrestrict-
ed funds from the S. T. Li Foundation.
References and Notes
[1] a) S. Gelin, R. Gelin, J. Org. Chem. 1979, 44, 808;
b) D. N. Kirk, C. R. McHugh, J. Chem. Soc. Perkin Trans.
716
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2004, 346, 713 716