Stereocontrolled photocyclization of 1,2-diketones: Application of a 1,3-acetyl group transfer methodology to carbohydrates

Article abstract of DOI:10.1021/jo702663w

(Chemical Equation Presented) Photolysis of 1-glycosyl-2,3-butanodione derivatives using visible light is a mild and selective procedure for the synthesis of chiral 1-hydroxy-1-methyl-5-oxaspiro[3.5]nonan-2-one carbohydrate derivatives. The results strongly suggest that stereocontrol of the cyclization is dependent on conformational and stereoelectronic factors. Further oxidative opening of the 1-hydroxy-1-methyl-2-cyclobutanone moiety affords new C-ketoside derivatives either in C- and O-glycoside series. This tandem two-step process could be considered to be a stereocontrol led 1,3-transference of an acetyl group, and it can be applied either to pyranose and furanose models.

Full text of DOI:10.1021/jo702663w

Stereocontrolled Photocyclization of 1,2-Diketones: Application of a
1,3-Acetyl Group Transfer Methodology to Carbohydrates
Antonio J. Herrera,* María Rondón, and Ernesto Suárez*
Instituto de Productos Naturales y Agrobiología del C.S.I.C. Carretera de La Esperanza 3,
38206 La Laguna, Tenerife, Spain
esuarez@ipna.csic.es; ajherrera@ipna.csic.es
ReceiVed December 14, 2007
Photolysis of 1-glycosyl-2,3-butanodione derivatives using visible light is a mild and selective procedure
for the synthesis of chiral 1-hydroxy-1-methyl-5-oxaspiro[3.5]nonan-2-one carbohydrate derivatives. The
results strongly suggest that stereocontrol of the cyclization is dependent on conformational and
stereoelectronic factors. Further oxidative opening of the 1-hydroxy-1-methyl-2-cyclobutanone moiety
affords new C-ketoside derivatives either in C- and O-glycoside series. This tandem two-step process
could be considered to be a stereocontrolled 1,3-transference of an acetyl group, and it can be applied
either to pyranose and furanose models.
Introduction
been well studied since the early 60s,3 only a few reports on
the use of this methodology appears in the literature3a,4 and, as
far as we know, only one case involved a chiral substrate.5
Diastereoselective C-C bond formation, absence of competitive
cleavage reactions6 or additional reagents, high yields, and
possible use of sunlight are the most attractive features of this
methodology for the synthetic chemist and render exploring its
scope and limitations highly interesting.
The development of efficient methods to generate regio- and
stereocontrolled C-C bonds under mild conditions, especially
when constructing quaternary stereocenters, is highly important
in synthetic organic chemistry.1 Norrish-Yang photocyclization
has been used to accomplish this task, aliphatic and aromatic
ketones and R-keto esters being the most widely used reactants.2
On the other hand, although the Norrish-Yang photocycliza-
tion of 1,2-diketones to generate 2-hydroxy-cyclobutanones has
Herein, we expand a previous report7 on the highly efficient
use of a Norrish-Yang photocyclization to obtain new spirocyclic
monosaccharide-derivatives of types III and IV via a hydrogen
atom transfer (HAT) reaction promoted by a 1,2-diketone I, in
its excited state, followed by C-C tetrasubstituted bond
formation in a diastereoselective manner (Scheme 1). Of special
interest is the study of the tendency to inversion at C-5,8
probably triggered by conformational changes that the 1,4-
(1) For recent reviews on the construction of quaternary carbon stereocenters,
see: (a) Steven, A.; Overman, L. E. Angew. Chem., Int. Ed. 2007, 46, 5488. (b)
García, C.; Martín, V. S. Curr. Org. Chem. 2006, 10, 1849. (c) Christoffers, J.;
Baro, A. AdV. Synth. Catal. 2005, 347, 1473. (d) Quaternary Stereocenters:
Challenges and Solutions for Organic Synthesis; Christoffers, J., Baro, B., Eds.;
Wiley-VCH: Weinheim, 2005. (e) Douglas, C. J.; Overman, L. E. Proc. Natl.
Acad. Sci. U.S.A. 2004, 101, 5363. (f) Ramo´n, D. J.; Yus, M. Curr. Org. Chem.
2004, 8, 149. (g) Christoffers, J.; Baro, A. Angew. Chem., Int. Ed. 2003, 42,
1688. (h) Denissova, I.; Barriault, L. Tetrahedron 2003, 59, 10105. (i)
Christoffers, J.; Mann, A. Angew. Chem., Int. Ed. 2001, 40, 4591. (j) Corey,
E. J.; Guzman-Perez, A. Angew. Chem., Int. Ed. 1998, 37, 388. (k) Fuji, K.
Chem. ReV. 1993, 93, 2037.
(3) (a) Urry, W. H.; Trecker, D. J. J. Am. Chem. Soc. 1962, 84, 118. (b)
Urry, W. H.; Trecker, D. J.; Winey, D. A. Tetrahedron Lett. 1962, 3, 609. (c)
Turro, N. J.; Lee, T.-J. J. Am. Chem. Soc. 1969, 91, 5651. (d) Urry, W. H.;
Duggan, J. C.; Pai, M.-S. H. J. Am. Chem. Soc. 1970, 92, 5785.
(4) (a) Hamer, N. K. J. Chem. Soc., Perkin Trans. 1979, 1, 1285. (b) Hamer,
N. K. Tetrahedron Lett. 1982, 23, 473. (c) Hamer, N. K. J. Chem. Soc., Perkin
Trans. 1983, 61. (d) Hamer, N. K. Tetrahedron Lett. 1986, 27, 2167.
(5) Obayashi, M.; Mizuta, E.; Noguchi, S. Chem. Pharm. Bull. 1979, 27,
1679.
(2) For reviews on Norrish-Yang, see: (a) Wagner, P. J. Acc. Chem. Res.
1989, 22, 83. (b) Wagner, P. J.; Park, B. In Organic Photochemistry; Padwa,
A., Ed.; Marcel Dekker: New York, 1991; p 227. (c) Handbook of Organic
Photochemistry and Photobiology, 2nd ed.; Horspool, W. M., Lenci, F., Eds.;
CRC Press: Boca Raton, FL, 2003; Chapters 52, 55, 57, and 58. (d) Wagner,
P. J. In Synthetic Organic Photochemistry (Molecular and Supramolecular
Photochemistry); Griesbeck, A. G., Mattay, J., Eds.; Marcel Dekker: New York,
2005; Vol. 12, p 11. (e) Wessig, P. In Radicals in Organic Synthesis; Renaud,
P., Sibi, M. P., Eds.; Wiley-VCH: Weinheim, 2001; Vol. 2, p 523. (f) Wessig,
P.; Muehling, O. Eur. J. Org. Chem. 2007, 2219.
(6) Yang, N. C.; Yang, D.-D. H. J. Am. Chem. Soc. 1958, 80, 2913.
(7) Herrera, A. J.; Rondo´n, M.; Sua´rez, E. Synlett 2007, 1851.
3384 J. Org. Chem. 2008, 73, 3384–3391
10.1021/jo702663w CCC: $40.75 2008 American Chemical Society
Published on Web 03/28/2008

Stereocontrolled Photocyclization of 1,2-Diketones
SCHEME 1. Mechanism of Photocyclization13
a C5–alkyl; thus, it will be termed a pseudoanomeric radical.
The photochemical generation of a pseudoanomeric radical at
C5 of the ribose moiety of a nucleotide, mimicking the 4′-RNA
and 4′-DNA radicals, has previously been achieved via a Norrish
type I cleavage to study the DNA and RNA damage.15
Additionally, we have recently shown that the relatively
unstable 2-hydroxy-cyclobutanones can be converted into much
more stable methyl-γ-ketoesters fused to the carbohydrate
skeleton by treatment with periodic acid in methanol following
previously reported procedures (as described in Scheme 4).3d,4b
The two-step tandem process (photocyclyzation followed by
oxidative opening and methyl esterification) can be considered
to be a methodology to achieve a rather interesting diastereo-
controlled CfC-[1,3]-acetyl shift. This methodology has been
shown to tolerate various functional and protecting groups used
in carbohydrate chemistry expanding its applications in organic
synthesis.
diradical intermediate II undergoes in its triplet state, within
its lifetime,9 before the intersystem crossing (ISC) occurs.
In this regard, we probe different substituents and stere-
ochemistries, mainly at positions 5, 6, and 9 of the pyranose
core of II-(Z), to explore the role of the stereoelectronic
interactions,9b,10 conformational restrictions,11 and formation of
intramolecular hydrogen bonds12 in the stereocontrol of this
reaction.
It has been reported that 1,2-diketones mainly abstract
hydrogen atoms from their triplet state and with a very small
rate constant related to alkyl and aryl ketones.3c However, in
our cases, the reactions were accomplished within relatively
short times. A new factor has been introduced in these models.
The presence of the geminate endocyclic oxygen may specially
activate the hydrogen atom transference. Furthermore, the 1,4-
diradical intermediate II may be stabilized by a possible
conjugative stabilizing interaction of the SOMO-C5 radical with
the lone pair at the ring oxygen and the σ*-LUMO of the C6-O
bond (if Y ) OR)10 affecting probably the rate of the HAT
step and the lifetime of intermediate II. To the best of our
knowledge, such studies have not so far been carried out,
although this C5-centered radical resembles the in-depth studied
anomeric radical14 where the C5–H bond has been replaced by
Results and Discussion
With the aim of rationalizing the role of the stereoelectronic
and conformational factors in the stereocontrol of this process,
we have synthesized a number of pyranose and furanose models
carrying a four-carbon 1,2-diketone tether and submitted to
various photolysis conditions as shown in Tables 1-2 and
Scheme 6.16
Preparation of the required 1-glycosyl-2,3-butanodione de-
rivatives was accomplished following three different protocols
in which the 1,2-diketone moiety was prepared by oxidation of
the corresponding alkyne derivative. The synthesis of com-
pounds 4, 8, and 11 are depicted in Scheme 2 as representative
examples of each protocol.17
The primary bromide of the glucose derivative 1 was
substituted by nucleophilic attack of allylmagnesiumchloride in
high yield. This alternative methodology was used because a
more direct route by nucleophilic substitution either of this
bromide, tosylate, or triflate derivatives with 1-propynylmag-
nesium bromide, ethynylmagnesium chloride, or propargylmag-
nesium bromide failed in this case. Transformation of the alkene
2 into the alkyne 3 was accomplished by ozonolysis followed
by the Corey-Fuchs methodology.18 Isomerization of the
terminal alkyne into the methylalkyne functionality under basic
conditions followed by oxidation with the RuO2 ·H2O/NaIO4
system provided the 1,2-diketone 4 in moderate yields.19
A second less efficient alternative procedure to introduce the
alkyne functionality modifying the primary (C-6) position of
suitably protected pyranose models was employed to obtain 8.
Swern oxidation of the primary alcohol of 5 and subsequent
addition of 1-propynylmagnesiumbromide afforded an isomeric
mixture of propargyl alcohol 6 in moderate yield. Barton-
McCombie radical deoxygenation20 through the methylxanthate
derivative leaded to formation of 7 with 44% of yield, which
(8) For examples of epimerization of anomeric and pseudoanomeric radicals,
see: (a) Brunckova, J.; Crich, D.; Yao, Q. Tetrahedron Lett. 1994, 35, 6619. (b)
Yamazaki, N.; Ei-chemberger, E.; Curran, D. P. Tetrahedron Lett. 1994, 35,
6626.
(9) For discussions on the lifetime of diradicals in solution, see: (a) Johnston,
L. J.; Caiano, J. C. Chem. ReV 1989, 89, 521. and references cited therein. (b)
Cai, X.; Cygon, P.; Goldfuss, B.; Griesbeck, A. G.; Heckroth, H.; Fujitsuka,
M.; Majima, T. Chem.-Eur. J. 2006, 12, 4662.
(10) (a) Dupuis, J.; Giese, B.; Ru¨egge, D.; Fischer, H.; Korth, H.-G.; Sustman,
R. Angew. Chem., Int. Ed. Engl. 1984, 23, 896. (b) Giese, B. Angew. Chem.,
Int. Ed. Engl. 1989, 28, 969. (c) Brunckova, J.; Crich, D. Tetrahedron 1995, 51,
11945. and references cited therein.
(11) (a) Griesbeck, A. G.; Mauder, H.; Stadtmu¨ller, S. Acc. Chem. Res. 1994,
27, 70. (b) Ihmels, H.; Scheffer, J. R. Tetrahedron 1999, 55, 885. (c) Giese, B.;
Wettstein, P.; Sta¨helin, C.; Barbosa, F.; Neuburger, M.; Zehnder, M.; Wessig,
P. Angew. Chem., Int. Ed. 1999, 38, 2586. and references cited therein. (d) For
anomeric effect, see: Juaristi, E.; Cuevas, G. Tetrahedron 1992, 48, 5019. and
references cited therein. (e) For anomeric effect of proton, see: David, S.
Carbohydr. Res. 2005, 340, 2569.
(15) Strittmatter, H.; Dussy, A.; Schwitter, U.; Giese, B. Angew. Chem., Int.
Ed. 1999, 38, 135.
(16) Three different light sources were used: Sunlight: Direct sunlight: a bright
sunny day, on cloudy days the reaction was considerably slower; UV lamp: 450
W ACE-Hanovia medium-pressure mercury lamp in an immersion well with
4.8 mm Pyrex walls; Daylight lamp: Philips lamp (master PL electronic, 23
W/865).
(12) Walther, K.; Kranz, U.; Henning, H.-G. J. Prakt. Chem. 1987, 329, 859.
(13) For convenience, the atom-numbering system used throughout this
section corresponds to that depicted in structures of Table 1 and the Schemes,
although a IUPAC systematic nomenclature has been used in the Supporting
Information.
(17) Complete details of the synthesis of all the 1,2-diketone precursors are
provided in the Supporting Information.
(14) For recent reviews, see: (a) Giese, B.; Zeitz, H.-G. In PreparatiVe
Carbohydrate Chemistry; Hanessian, S., Ed.; Marcel Dekker: New York, 1997;
p 507. (b) Pearce, A. J.; Mallet, J.-M.; Sinaÿ, P. In Radicals in Organic Synthesis;
Renaud, P., Sibi, M. P., Eds.; Wiley-VCH: Weinheim, 2001; Vol. 2, p 523.
(18) Corey, E. J.; Fuchs, P. L. Tetrahedron Lett. 1972, 36, 3769.
(19) Zibuck, R.; Seebach, D. HelV. Chim. Acta 1988, 71, 237.
(20) Barton, D. H. R.; McCombie, S. W. J. Chem. Soc., Perkin Trans. 1975,
1, 1574.
J. Org. Chem. Vol. 73, No. 9, 2008 3385

Herrera et al.
SCHEME 2. Synthesis of 1,2-Diketone Precursors
SCHEME 3. Photocyclization of 1,2-Diketone 4
SCHEME 5. Oxidative Opening of 1-Hydroxy-1-methyl-
cyclobutan-2-one Moiety
SCHEME 4. Synthesis of C-Ketosides
Corey–Fuchs protocol, and oxidation of the triple bond to obtain
the 1,2-diketone moiety as shown for the synthesis of compound
11.
Alternatively, the alkyne moiety was converted into the 1,2-
diketone functionality by ozonolysis quenched with Ph3P for
the synthesis of various models.23 It is important to mention
that the yields of formation of the 1,2-diketone moiety, either
in pyranose or furanose models, were lower with substrates
upon ozonolysis followed by quenching with Me2S provided
the product 8 in moderate yield.21
All C-glycoside derivatives were efficiently prepared by using
a four-step methodology: Lewis acid catalyzed C-glycosidation
with allyltrimethylsilane22 followed sequentially by ozonolysis,
(21) Favino, T. F.; Fronza, G.; Fuganti, C.; Fuganti, D.; Graselli, P.; Mele,
A. J. Org. Chem. 1996, 61, 8975.
(23) The use of Ph3P instead of Me2S for this selective oxidation does not
have precedents in the literature and resulted in an improved reagent for various
models.
(22) Levy, D. E.; Tang, C. The Chemistry of C-Glycosides; Pergamon Press:
Oxford, UK, 1995.
3386 J. Org. Chem. Vol. 73, No. 9, 2008

Stereocontrolled Photocyclization of 1,2-Diketones
TABLE 1. Photocyclization of 1,2-Diketones Derived from O-Glycosides via HAT
b
a SL: sunlight, UV: UV-lamp, DL: daylight-lamp. The ratios and stereochemistries were determined by H NMR of the reaction crudes and NOESY
experiments, respectively. For clarity, the underlined major diastereomers are depicted. c Yields refer to purified products. d t-BuOH/CHCl3 (3:1). e 0.4
M Naphthalene, 10 mM substrate. f 0.5 M Benzophenone, 38 mM substrate. g 0.5 M Pyrene, 38 mM substrate.
1
bearing benzyl groups, being approximately around 50%. In fact,
the three above-mentioned procedures to accomplish this alkyne
oxidation can be used to oxidize benzyl groups to benzoate
groups.24
product 12-(2R) was isolated in moderate yield after silica gel
chromatography as a sole stereoisomer (Table 1, entry 1). Side
products of photooxidation and intermolecular reactions were
formed, probably due to the presence of oxygen, a long reaction
time, and restricted mobility of the molecules in the crystalline
net. The observed stereocontrol could have been determined
by the preorganized media in the crystal;26 however, a single
product 12-(2R) was obtained in quantitative yield upon
irradiation with sunlight in solution with CHCl3 or C6H6 as
solvent (Table 1, entry 2). A slightly shorter time was required
upon irradiation with the UV lamp (Table 1, entry 3). The
reaction proceeded with total retention at C5. For this model, a
C5 radical with very slow isomeric interconversion is formed
at the 1,4-diradical intermediate II-4 stage (Scheme 3). This
isomerization is controlled by conformational changes at the
The reaction of photocyclization to obtain the desired
spirocompounds were performed by placing the 1,2-diketones
in a Pyrex vessel and irradiating with one of the three different
light sources16 emitting in the violet-blue region of the spectrum,
employing preferentially deoxygenated solvents. The yields of
isolated spirocompounds for this reaction vary between moderate
and excellent for all the examples assayed (Tables 1 and 2).
Model 4 was selected due to its conformationally restricted
4C1 pyranose ring.25 1,2-Diketone 4 was irradiated with outdoor
sunlight in its crystalline form for 20 h until the yellow color
faded, affording a mixture of compounds from which the main
(24) (a) Schuda, P. F.; Cichowicz, M. B. Tetrahedron Lett. 1983, 24, 3829.
(b) Angibeaud, P.; Defaye, J.; Gadelle, A.; Utille, J.-P. Synthesis 1985, 1123.
(25) The mentioned conformation of substrates was determined by 1H NMR.
(26) Kaanumalle, L. S.; Natarajan, A.; Ramamurthy, V. In Synthetic Organic
Photochemistry (Molecular and Supramolecular Photochemistry); Griesbeck,
A. G., Mattay, J., Eds.; Marcel Dekker: New York, 2005; Vol. 12, p 553.
J. Org. Chem. Vol. 73, No. 9, 2008 3387

Herrera et al.
TABLE 2. Photocyclization of 1,2-Diketones Derived fromC-Glycosides via HAT
a SL: sunlight, UV: UV-lamp. b The ratios and stereochemistries were determined by 1H NMR of the reaction crudes and NOESY experiments,
respectively. For clarity, only the underlined diastereomers are depicted. c Yields refer to purified products. d Another two isomers with undefined
stereochemistry were obtained with the ratio [25-(2S)/two isomers, 4:1.5:1]. e 0.4 M Naphthalene, 10 mM substrate.
ring moiety influenced by its substituents and stereoelectronic
effects within the lifetime of this intermediate before the ISC
occurs. In this case, four main favorable factors can play a major
role leading to the product with retention at C5: Stabilizing
interaction of the SOMO-C5 radical with the lone pair at the
ring oxygen (pseudoanomeric effect) and the σ*-LUMO of the
C6-O bond (ꢀ-oxygen effect) leading both to an axial C2-C5
bond formation via an early transition state,27 the classical
anomeric effect restricting the relative conformation between
endocyclic oxygen and C9 and the possible existence of an
intramolecular hydrogen bond C2-OH· · ·O(Bn)-C6 favoring
their syn approximation. Moreover, steric hindrance and/or the
possible hydrogen bond formation C2-OH· · ·O(Bn)-C6 prob-
ably enhanced the high stereocontrol at C2 of the product,
because the introduction of a protic solvent led to a mixture of
diastereomers (Table 1, entry 4).26
Compound 8, dissolved in chloroform, was irradiated alter-
natively with sunlight, a UV lamp, or 3 daylight lamps affording
products 14-(2S) and 15-(2R) with equal ratio and slightly
different reaction times and yields (Table 1, entries 5–7). This
scarce influence on yields and isomeric ratio of products,
regarding the light source and the use of benzene instead of
chloroform, was observed for all of the assayed models;
therefore, the irradiation with UV light and chloroform as
solvent were chosen to be the standard conditions to compare
the results for each model. Although the 1H NMR spectrum of
the reaction crude showed a clean conversion into the two
desired products, the reaction yield after purification was
diminished in relation to model 4, from which it differs only in
the absence of the oxygenated substitution at C6 (Table 1, entries
3 and 6). The same feature was observed for models 19 and 24
in relation to 16 and 11, respectively (compare entries 9 and
12, Table 1 and entries 1 and 2, Table 2). The main reason was
the reduced stability of their corresponding cyclic products
during the chromatographic purification on silica gel. The
cyclization of model 8 occurred with 70% of retention at C5
and the stereochemistry at C2 of the spiro-product 14-(2S) is
the opposite of the observed for product 12-(1R), indicating an
important contribution of the oxygenated substituent at C6 in
the stereocontrol of the process. Interestingly, analysis of the
spectroscopic data reveals that the pyranose ring of products
14-(2S) and 15-(2R) adopts the 1C4 and 4C1 conformation
respectively, placing the C2 in equatorial position and directing
the hydroxyl group toward the endocyclic oxygen, leading us
to believed that a hypothetic hydrogen bond and/or the
diminished steric shield at the 1,4-diradical intermediate II-6
may play an important role in the stereocontrol at this asym-
metric center (Figure 1). In fact, in an experiment not shown in
Table 1, a mixture of the four possible stereoisomers of cyclic
products is obtained if the reaction is carried out in the presence
of a protic solvent (tert-butanol) in a ratio that is difficult to
estimate due to the complexity of the 1H NMR spectrum of the
reaction crude.
(27) (a) Beckwith, A. L. J.; Duggan, P. J. Tetrahedron 1998, 54, 4623. (b)
Beckwith, A. L. J.; Duggan, P. J. Tetrahedron 1998, 54, 6919.
3388 J. Org. Chem. Vol. 73, No. 9, 2008

Stereocontrolled Photocyclization of 1,2-Diketones
SCHEME 6. Cyclization–Oxidative Opening Tandem
Reactiona
FIGURE 1. Molecular models of 14-(2S) and 15-(2R) obtained using
CS Chem3D Pro v10.0 software. Benzyl and methyl hydrogens have
been omitted for clarity.
It is important to mention that benzophenone, as triplet
sensitizer, and pyrene and naphthalene, as triplet quenchers, have
been assayed with all the substrates of Tables 1-2 with the
aim of modulating the stereoselectivity. Indeed, despite model
16, pyrene strongly inhibits the process, extending reaction times
beyond 48 h and leading to the formation of contaminating side
products, but scant influence on the stereoselectivity was
observed for all other models. On the other hand, performing
the photochemical reactions in the presence of naphthalene
extended reaction times by 3.5–5 longer times but did not affect
notably either the diastereomeric ratios or reaction yields as
exemplified in entries 8 (Table 1) and 5 (Table 2). The presence
of benzophenone significantly reduced reaction times to ap-
proximately 30–50%, but no considerable changes were ob-
served in the stereoisomeric ratios. These experiments suggest
an excited triplet state, but the high efficiency of pyrene in the
stereocontrol of model 16 remains unclear at the present.
The possible stereoelectronic and conformational stability
established by the presence of the oxygenated substitution at
C6 and the axial position of the methoxyl group at the anomeric
position is not present in model 19. As may be expected,
compound 19 cyclized with a lower percentage of retention at
C5, the inversion product 21-(2R) being the major one (Table
1, entry 12). The discussion about the stereocontrol observed
at C2 for products 14-(2S) and 15-(2R) can be extended to
products 20-(2S) and 21-(2R).
a 1,3-Acetyl transference in pyranose and furanose models.
Irradiation of model 16 with a UV-lamp gave the expected
products in excellent yield, as a mixture of three diastereomers
(Table 1, entry 9). The reaction proceeded predominantly with
retention at C5 (95%) [17-(2R) and 17-(2S)], although the
inversion product 18-(2R) was also isolated. In this case,
substrate 16 is in a 4C1 conformation, but the pyranosyl ring of
the intermediate II-11 (Scheme 1) is conformationally more
flexible and distortions are more likely, and even more so when
placing the OMe group in the axial position, because of the
anomeric effect at C9. The favorable contribution of the classical
anomeric effect at C9 to the retention products at C5 seems to
be less than the ꢀ-oxygen effect (compare entries 3, 6 and 9,
Table 1).
The photoreaction of 16 in the presence of the triplet sensitizer
benzophenone led almost to the same ratio of diastereomers
(compare entries 9 and 10, Table 1). The reaction was complete
in half-time, though the yield slightly decreased. In contrast, in
the presence of pyrene, not only as a triplet quencher but also
as a singlet sensitizer, the retention product 17-(2R) was obtained
with absolute diastereoselectivity and in high yield, although a
longer reaction time was required (Table 1, entry 11). The
cyclization of the singlet diradical II-16 (S1-spinisomer) inter-
mediate is obviously so fast that rotations around the C4–C5
single bond do not take place. The configuration of the
asymmetric center C2 in this last experiment is probably
controlled by the existence of an intramolecular hydrogen bond,
steric factors, and restricted rotation of C2–C3 in the short-
lived singlet diradical II-16 intermediate.
Photocyclization of the C-glycoside 11 afforded the four
possible stereoisomers in good yield (67% retention, Table 2,
entry 1). A 1H NMR experiment indicated that starting
4
compound 11 adopts a C1 conformation with the transferring
hydrogen atom in the equatorial position. For this case,
pseudoanomeric and ꢀ-oxygen effects play a disfavored con-
tribution to obtain the retention products triggering conforma-
tional distortions to favor the axial reactivity of the gene-
rated transient radical at C5, explaining the low retention
percentage. This reaction can be performed in a very efficient and
environmentally friendly way, as shown in entry 2 of Table 2.
Upon irradiation under standard conditions, model 24 yielded
three stereoisomers of cyclic products in a ratio of 4:1.5:1 (Table
2, entry 3). The major product was the isomer 25-(2S), but the
stereochemistry of the other two minor stereoisomers could not
be elucidated. Oxidative opening of the cyclobutanol moiety
J. Org. Chem. Vol. 73, No. 9, 2008 3389

Herrera et al.
of the mixture of the two nondetermined diastereomers (1.4:1)
afforded product 35 (Scheme 5) and its epimer at C3 in a ratio
of 1.1:1 respectively in a 56% yield, indicating that one of the
undetermined isomers corresponded to 25-(2R). The absence
of the oxygenated substituent at C6 in 11, as in the case of
1,2-diketone 24, enhanced the percentage of retention product
from 67 to over 77% (compare entries 1 and 3, Table 2),
showing the disfavored contribution of this substituent to the
retention.
The experiments described in entries 4 and 6 demonstrate
that the election of a different protecting group may be essential
for the outcome of the reaction. Models 27 and 30 adopt a 4C1
and 1C4 conformation, respectively.28 The photoreaction of
model 27 involved transfer of a hydrogen atom in an axial
position, affording exclusively the retention products. However,
model 30, which involved the transfer of an equatorial hydrogen
atom, afforded the four possible isomers with 74% of retention.
These examples indicate that conformational and/or steric
shielding factors can play an important role for the stereocontrol
of the process especially related to the asymmetric center C5
of the cyclic products. Model 30 achieved a clean cyclization
and not spin center shift (SCS), resulting in a possible migration
of the vicinal acetyl group to the pseudoanomeric position, was
observed.2f
The 1-hydroxy-1-methyl-cyclobutan-2-one moiety could be
opened with oxidative conditions using periodic acid in methanol
in a preparative manner in excellent yield as depicted in Scheme
4.3d A diastereomeric mixture of compound 28-(2S) and 28-
(2R) led to a single product 33, which after deprotection afforded
a new C-ketoside 34. This procedure introduces a new entry
into C-ketosides that can only be accessed through chemical
synthesis.29 Alternatively, this oxidative opening reaction can
be achieved by treatment with m-chloroperbenzoic acid in
dichloromethane followed by esterification with diazomethane
in diethyl ether but is less efficient than the previously described
procedure for all the assayed R-diketone models 8, 19, 27, 39,
and 41.30
this tandem cyclization-oxidative opening reaction should be
preferentially addressed without purification of the cyclic
intermediates especially those for which the spiro intermediates
prove to be unstable during purification on silica gel. This was
the case for all the furanose models assessed (39, 41, and 43).
Any attempt of isolation of their spiro intermediate resulted in
a considerably lower yield than the direct isolation of the final
γ-ketoester. The Norrish-Yang photocyclization products derived
from models 39 and 43 were isolated in 56 and 67% yield,
respectively, whereas γ-ketoesters 40 and 44 were obtained in
80 and 86% yield with the tandem two-step procedure. The spiro
compound derived from model 41 proved to be remarkably
unstable at room temperature, decomposing over 50% within
24 h of isolation, despite being stored under nitrogen, whereas
the isolated mixture of γ-ketoesters 42-(3R) and 42-(3S) (2.1:
1) remained stable for over two weeks under air.
The results disclosed in Scheme 6 indicate that the Norrish-
Yang reaction for all the furanose models, either C- or
O-glycosides, proceeded predominantly with retention. The
overall retention percentages, measured in the reaction crude
1
by H NMR, were 72, 68, and 73%, respectively, for models
39, 41, and 43, which do not vary considerably with the ratios
of the isolated products shown in Scheme 6. Analysis of the
1H NMR of the crude of the photocyclization reaction with
models 39 and 43 showed the formation of three stereoisomers
for each one, with ratios (4.9:1.4:1) and (1.8:1.3:1), respectively.
However, NOESY experiments were not conclusive for the
assignment of the absolute configuration for any of the spiro
products.
Scarce variations in yields, reaction times, and stereoisomeric
ratios of products were observed upon irradiation with sunlight
or daylight lamps instead of an UV lamp in this cyclization-
oxidative opening tandem reaction, either with pyranose or
furanose models. Glyconate derivatives 35, 36, 37, 38, 40, 42,
and 44 are sugar-fused γ-ketoesters susceptible to further
modifications to engineer versatile scaffolds and building blocks.
Moreover, these oxidative opening conditions have been
applied with good yields to the greater part of the studied
models, whereas a lower efficiency was observed when benzyl
protecting groups were present in the molecule. This is the case
for all the examples shown in Scheme 5, where products 35,
36, and 37 were obtained in moderate yields (63, 61, and 50%,
respectively), because partial oxidation and subsequent hydroly-
sis of the benzyl moiety, to afford more polar products, also
occurs. This reaction was very useful to determine or verify
the configuration of C5 in the cyclic products, especially
valuable with isomeric irresolvable mixtures as shown in eqs 2
and 3 of Scheme 5.
Conclusion
The results described herein demonstrate that the Norrish-
Yang photocyclization of 1,2-diketones is well suited for the
stereocontrolled synthesis of cyclobutanols fused to carbohy-
drate. The stereocontrol of the photocyclization is dependent
on stereochemistries, functional groups, and ring-size of the
precursors, and the results engage with the assumption of similar
features between anomeric and pseudoanomeric radicals despite
the respective lifetime. A tandem photocyclization-oxidative
opening protocol proved to be very useful to prepare sugar-
fused γ-ketoesters derived either from O- and C-glycosides or
from pyranose and furanose models. This tandem protocol
results in an efficient stereocontrolled 1,3-acetyl group transfer
methodology.
These previous results appear promising in expanding the
scope of this Norrish-Yang reaction with 1,2-diketones in
asymmetric synthesis in the same way as arylketones or R-keto
esters have already been used.2 Further studies of this reaction,
applied to noncarbohydrate derivatives as well as noncyclic
chiral compounds, are currently underway in our laboratory.
The two-step conversion of 1,2-diketone derivatives into
γ-ketoesters, as shown in Scheme 6, can be considered to be a
stereocontrolled 1,3-transfer of the acetyl group. This methodol-
ogy can be also extended to furanosyl derivatives. In most cases,
(28) TBS protecting groups of pyranoses causes a flip of their conformation
leading to 1C4-form. Due to this conformational mobility, the L-rhamnose TBS
1
derivatives prepared in this work exhibit broadened signals in their H and 13C
NMR spectra at room temperature. (a) Hosoya, T.; Ohashi, Y.; Matsumoto, T.;
Suzuki, K. Tetrahedron Lett. 1996, 37, 663. (b) Yahiro, Y.; Ichikawa, S.; Shuto,
S.; Matsuda, A. Tetrahedron Lett. 1999, 40, 5527. (c) Shuto, S.; Terauchi, T.;
Yahiro, Y.; Abe, H.; Ichikawa, S.; Matsuda, A. Tetrahedron Lett. 2000, 41, 4151.
(29) See a list of procedures to generate C-ketosides: Roberts, S. W.; Rainier,
J. D. Org. Lett. 2005, 7, 1141.
Experimental Section
General Procedure for the Photocyclization of 1,2-Diketones.
The corresponding 1,2-diketone was placed in a Pyrex glass vessel
(30) (a) Kedar, T. E.; Miller, M. W.; Hegedus, L. S. J. Org. Chem. 1996,
61, 6121. (b) Bueno, A. B.; Hegedus, L. S. J. Org. Chem. 1998, 63, 684.
3390 J. Org. Chem. Vol. 73, No. 9, 2008

Stereocontrolled Photocyclization of 1,2-Diketones
with or without the indicated deoxygenated solvent (approximately
0.055 M). The vessel was purged with argon before being sealed
and irradiated with sunlight, an UV lamp at 10 cm, or three daylight
lamps at 7 cm until the reaction turned colorless. Concentration
under vacuum, if required, afforded a residue that was purified by
column or Chromatotron chromatography (hexanes–EtOAc mix-
tures) to give the spiro compounds. Specific conditions and yields
are shown in Tables 1-2 and Scheme 6.
General Procedure for the Photocyclization of 1,2-Dike-
tones with a Photosensitizers or Quencher. The corresponding
1,2-diketone (0.038 mmol) and photosensitizer (benzophenone) or
quencher (pyrene, naphthalene) (0.5 mmol) was dissolved in CDCl3
(1 mL), placed in a 5-mm NMR tube, which was purged with argon
before being closed. The reaction mixture was irradiated with a
UV–lamp at 10 cm until complete conversion and the solvent was
removed under vacuum. The reaction was monitored by 1H NMR.
Chromatotron chromatography of the residue with (hexanes) to
remove the photosensitizer followed by (hexanes-EtOAc mixtures)
afforded the cyclic compounds.
compound usually as a mixture of various stereoisomers. After the
solvent was removed under vacuum, HIO4 (3.5 mmol) was added
to the solution of this residue in MeOH (30 mL) at 0 °C and stirred
for 1 h at this temperature followed by another hour at rt. The
reaction mixture was quenched with saturated solution of Na2CO3
and extracted 3 times with EtOAc. The organic extracts were dried
1
over Na2SO4 and concentrated in vacuum. The H NMR of the
residue showed the ratio of the two possible stereoisomers.
Chromatotron chromatography (hexanes-EtOAc mixtures) of the
residue afforded the sugar-fused γ-ketoesters.
Acknowledgment. This work was supported by the research
programs Nos. CTQ2004-06381/BQU and CTQ2004-02367/
BQU of the Ministerio de Educación y Ciencia, Spain, cofi-
nanced by the Fondo Europeo de Desarrollo Regional (FEDER).
A.J.H. and M.R. thank the Program I3P-CSIC and ULA-
Venezuela, respectively, for a research contract.
Supporting Information Available: General and chemical
information plus experimental procedures and copies of 1H and
13C NMR spectra for all new compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
General Procedure for the Tandem Photocyclization–Oxi-
dative Opening Reaction from 1,2-Diketones. Following the
general procedure of photocyclization with deoxygenated CHCl3
as solvent and irradiation with an UV lamp for approximately 1 h,
the corresponding 1,2-diketones (1 mmol) afforded the cyclic
JO702663W
J. Org. Chem. Vol. 73, No. 9, 2008 3391