Osmium-mediated oxidative cyclizations: A study into the range of initiators that facilitate cyclization

Article abstract of DOI:10.1002/asia.200900168

A general route to prepare substituted, saturated five-membered heterocycles has been developed. The application of a wide range of starting materials to the osmium-catalyzed oxidative cyclization reaction is described. Diols, hydroxy-amides, hydroxy-sulf

Full text of DOI:10.1002/asia.200900168

DOI: 10.1002/asia.200900168
Osmium-Mediated Oxidative Cyclizations: A Study into the Range of
Initiators That Facilitate Cyclization
Timothy J. Donohoe,*[a] Katherine M. P. Wheelhouse (nꢀe Gosby),[a]
Peter J. Lindsay-Scott,[a] Gwydion H. Churchill,[a] Matthew J. Connolly,[a]
Sam Butterworth,[a] and Paul A. Glossop[b]
Abstract: A general route to prepare
substituted, saturated five-membered
heterocycles has been developed. The
application of a wide range of starting
materials to the osmium-catalyzed oxi-
dative cyclization reaction is described.
Diols, hydroxy-amides, hydroxy-sulfo-
namides, and carbamates all cyclize in
moderate to excellent yields to give
cis-tetrahydrofurans and pyrrolidines,
depending upon the position of the
heteroatoms in the starting materials.
These cyclizations all proceed with
near total selectivity for the cis-hetero-
cycles, and with stereospecific introduc-
tion of a hydroxy group adjacent to the
ring. Moreover, routes to enantiopure
starting materials are described, which
give enantiopure products upon cycli-
zation. Catalyst loadings of as low as
one mol percent have been successfully
employed for this transformation.
Keywords: catalysis · heterocycles ·
osmium · pyrrolidines · tetrahydro-
furans
Introduction
The oxidative cyclization of 1,5-dienes to give 2,5-disubsti-
tuted cis-tetrahydrofurans (THFs) has received a considera-
ble amount of attention since its discovery by Klein and
Rojahn in 1965, with a number of metal-oxo species now ca-
pable of accomplishing this transformation.[1] Recent work
from our laboratory has shown that catalytic osmium tetrox-
ide in the presence of trimethylamine N-oxide (TMO) and
trifluoroacetic acid (TFA) was a mild and effective system
for the oxidative cyclization of 1,5-dienes (Scheme 1).[2] This
reaction was stereoselective for the formation of cis-hetero-
Scheme 1. Catalytic oxidative cyclization of 1,5-dienes. TMO=trimethyl-
amine N-oxide dihydrate.
cycles and stereospecific with respect to alkene oxidation, as
confirmed by X-ray crystal structure analysis on the cyclized
compound 2.
[a] Prof. T. J. Donohoe, Dr. K. M. P. Wheelhouse (nꢀe Gosby),
P. J. Lindsay-Scott, Dr. G. H. Churchill, M. J. Connolly,
Dr. S. Butterworth
The proposed mechanism for this transformation involved
osmylation of a single alkene unit, followed by an acid-pro-
moted cycloaddition of the osmate ester to give the corre-
sponding THF via transition structure 5 (Figure 1). This was
analogous to the mechanism proposed by Piccialli et al.[1k]
and in accordance with the mechanism of permanganate-
mediated oxidative cyclization originally proposed by Bald-
win et al.[1c]
The bidentate nature of the initially formed osmate ester
enforced the 2,5-cis selectivity of the cyclization, because
models indicated that such a bidentate osmate ester species
could not attain sufficient orbital overlap with the alkene to
Department of Chemistry
University of Oxford
Chemistry Research Laboratory, Mansfield Road
Oxford, OX1 3TA (UK)
Fax : (+44)186-527-5708
E-mail: timothy.donohoe@chem.ox.ac.uk
[b] P. A. Glossop
Pfizer Global R&D
Pfizer Limited
Ramsgate Road, Sandwich, Kent CT13 9NJ (UK)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/asia.200900168.
Chem. Asian J. 2009, 4, 1237 – 1247
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1237

FULL PAPERS
droxylated in situ to prevent unwanted dihydroxylation of
the substrate olefin, a process clearly detrimental to the
yield of THF (see below). Herein, we report a full account
of our investigations into different initiating functionalities
for osmium in the oxidative cyclization reaction for the for-
mation of both oxygen and nitrogen heterocycles. The aim
of these studies was to increase the scope of the oxidative
cyclization to encompass new functionality adjacent to the
THF ring, and also to investigate cyclization to form pyrroli-
dines through a completely new mode of reaction.
Figure 1. Proposed mechanism for catalytic oxidative cyclization (likely
osmium oxidation states shown).
give the trans-product (see 5); furthermore, the related
rheniumACHTUNGTRENNUNG(VIII) oxide mediated cyclizations of 5-hydroxyal-
kenes, proposed to proceed via a monodentate rhenium
ester, afforded 2,5-trans-THFs owing to an “open” transition
state.[3]
However, the diene cyclization reaction offered no scope
for the synthesis of enantiopure THFs. Addition of cinchona
alkaloid ligands to a catalytic oxidative cyclization led to no
absolute stereochemical induction, likely owing to the low
pH leading to protonation of the chiral amine ligand and
possibly also to the initial dihydroxylation step being a
ligand-independent, second-cycle process.[4] A range of pH
values and oxophilic Lewis acids were screened in the oxida-
tive cyclization reaction in a search for acidic conditions
that would activate the metal-oxo species but not deactivate
the amine, but to no avail.
Results and Discussion
Oxidative Cyclization of 1,3-Diol Initiators
Initially, we chose to examine the oxidative cyclization of
1,3-diols to give the corresponding cis-THFs. Success in this
endeavor would expand the range of functionalities that
chelate to osmium and facilitate cyclization, increasing the
number of structural units accessible from this methodology.
Therefore, the starting 1,3-diols were synthesized from the
corresponding aldehydes by aldol reaction and subsequent
reduction or organometallic addition (Scheme 3). Thus, the
known aldehyde 7 was synthesized from diol 6 (derived in
We then uncovered an analogous oxidative cyclization uti-
lizing enantiopure diols, derived from 1,5-dienes by Sharp-
less asymmetric dihydroxylation, to afford enantiopure
THFs with no erosion of enantiomeric purity (Scheme 2).[5]
This transformation greatly improved the scope of the oxi-
dative cyclization, especially with respect to the formation
of single product enantiomers, and was subsequently em-
ployed in the synthesis of the natural products cis-solamin
and cis-sylvaticin.[6]
During these studies, it was found that an excess of a sac-
rificial alkene (isoprene shown in Scheme 2) in the reaction
mixture ensured high yields and satisfactory reaction
times.[5a] We postulated that this second alkene was dihy-
Scheme 3. Synthesis of 1,3-diol substrates for oxidative cyclization.
Scheme 2. Oxidative cyclization of enantiopure diols.
LDA=lithium diisopropylamide.
1238
www.chemasianj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2009, 4, 1237 – 1247

Osmium-Mediated Oxidative Cyclizations
two steps from geraniol)[7] by oxidative cleavage with
sodium periodate in 92% yield. Subsequent aldolizaton with
the lithium anion of ethyl acetate gave the hydroxy-ester 8
in good yield. Lithium borohydride reduction to the 1,3-diol
9 proceeded in 92% yield, while the addition of an excess of
methylmagnesium bromide furnished the tertiary alcohol 10
in 57% yield. In order to probe the stereospecificity of the
cyclization reaction, the corresponding 1,3-diol 14 bearing a
cis-alkene was prepared by the same sequence, beginning
with the known diol 11 derived from nerol.[8] A second 1,3-
diol 17 was readily prepared from commercially available
cis-4-heptenal (15) in two facile steps (Scheme 3).
During optimization studies on the attempted cyclization
of 1,3-diol 9, the reaction was found to be much more slug-
gish than for the corresponding 1,2-diol using isoprene as a
sacrificial alkene, with only 40–50% product being obtained
even after 36–48 h. This difference in reactivity was attribut-
ed to the weaker chelating ability of 1,3-diols over 1,2-diols
for large metals such as osmium (larger metals generally
have a preference for five-membered chelates over six,
owing to more favorable bond angles).[9] A consequence of
this weaker binding would be a lower concentration of the
“loaded” 1,3-diol osmate ester and, therefore, a slower over-
all reaction. In order to try to speed up catalytic turnover of
any osmium that was bound to the 1,3-diol, we proposed the
use of a sacrificial alkene bearing a neighboring group capa-
ble of assisting with hydrolysis and ligand exchange after
cyclization. We reasoned that an increase in the turnover of
the catalyst might compensate for the poorer binding of the
1,3-diol and thereby increase the overall reaction rate.
Sharpless and Fokin had shown previously that addition of
citric acid to otherwise slow dihydroxylation reactions can
lead to a dramatic rate increase, owing to the free carboxyl-
ate group of the citrate ligand assisting with hydrolytic re-
lease of the product.[10] a,b-Unsaturated carboxylic acids
also undergo rapid aminohydroxylation for the same reason,
and the amino alcohol products are excellent ligands for ac-
celerating both aminohydroxylation and dihydroxylation re-
actions.[11] As a result, it was proposed that an a,b-unsaturat-
ed carboxylic acid would be a suitable sacrificial alkene,
since the resulting bound diol (derived from dihydroxyla-
tion) should allow faster exchange with the 1,3-diol sub-
strate and hydrolytic release of the product, effectively
acting as a second cycle ligand. Pleasingly, the use of trans-
cinnamic acid (TCA) resulted in a significant increase in
rate (complete consumption of starting material was ob-
served after 16–24 h) for the cyclization of 1,3-diol 9
(Scheme 4), while control experiments showed that this rate
increase was unlikely to be derived from the simple pH ef-
fects of adding extra acid to a cyclization.[12] Application of
these optimized reaction conditions to the cyclization of 1,3-
diols 9, 10, 14, and 17 delivered the THFs 18–21 in 66-82%
Scheme 4. Oxidative cyclization of 1,3-diols. TCA=trans-cinnamic acid.
the alkene unit. Given these observations, and the substan-
tial precedent that exists for this oxidative cyclization reac-
tion, we make the reasonable assumption that addition
across the alkene is syn and the stereochemistry is therefore
as shown in Scheme 4.
Next, we sought to prepare and test an enantiopure 1,3-
diol in the oxidative cyclization reaction (Scheme 5). Thus,
aldehyde 7 underwent efficient aldolization with Evans aux-
iliary 22[13] in the presence of nBu2BOTf and triethylamine
to give the syn-aldol product 23 in 87% yield. After reduc-
tive removal of the auxiliary, the diol 24 was subjected to
the optimized cyclization conditions and yielded 81% of the
desired hydroxy-THF 25 (Scheme 5).
Scheme 5. Synthesis and cyclization of an enantiopure 1,3-diol substrate.
Limitations to the Methodology
Subsequently, a series of other initiator groups were investi-
gated in the oxidative cyclization reaction (Scheme 6). In
the first instance, a number of compounds were prepared
which did not give the expected heterocycles under the con-
ditions developed. Bis homo-allylic alcohol 26 failed to cy-
clize under a variety of osmium-catalysis conditions, in con-
trast to the procedures documented for rhenium and
cobalt.[14] The formation of a seven-membered osmate ester
derived from 28 was assumed to be unlikely, given the rela-
1
yield as single product diastereomers (by H and 13C NMR).
Nuclear Overhauser effect (nOe) experiments on 18–21
showed that the THFs had a 2,5-cis relationship, and the dif-
ferences in the NMR spectra of 18 and 20 indicated that the
reaction was stereospecific with respect to addition across
Chem. Asian J. 2009, 4, 1237 – 1247
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
1239

T. J. Donohoe et al.
FULL PAPERS
Scheme 6. Limitations of diol cyclizations. In each case, a variety of dif-
ferent cyclization conditions (based around catalytic osmium, acid, reoxi-
dant, and solvent) were screened.
tive unreactivity of 9. Moreover, the osmate ester derived
from 1,3-diol 30 suffered from undue strain, which made
overlap of the olefin with an osmium-oxo ligand difficult
and prohibited cyclization. The lack of reactivity of phenol
32 was perhaps more puzzling, but demonstrated that pla-
narity along the 1,3-diol backbone was not tolerated in this
reaction. Finally, the failure of the 1,2-diol 34 to cyclize to
the corresponding tetrahydropyran 35 under a variety of dif-
ferent conditions (some forcing) was surprising, and attribut-
ed to excessive steric strain in the transition state.[15]
Scheme 7. Synthesis of amino alcohols 39, 40, 44, and 45 (Mosherꢂs ester
formation from alcohol 45 and subsequent separation of diastereomers
furnished enantiopure samples of 45). DHP=dihydropyran; Ts=tolu-
ene-4-sulfonyl; Cbz=benzyloxycarbonyl; PPTS=pyridinium para-tolu-
ene sulfonate; DMAP=4-dimethylaminopyridine; THP=tetrahydropyr-
an.
Pleasingly, submission of the sulfonamide and carbamate
protected amino alcohols 39, 40, 44, and 45 to the optimal
conditions for 1,3-diol cyclization led to excellent yields of
the corresponding amino-THFs 46–49 (Scheme 8). X-ray
crystal structure analysis of 46 and 48 indicated the reaction
to be both stereoselective for formation of 2,5-cis-THFs and
Formation of Amino-THFs by Oxidative Cyclization
Having shown that dienes, 1,2-diols, and 1,3-diols were suita-
ble initiators for the oxidative cyclization reaction, attention
was then turned to protected amino alcohols. Sharplessꢂ
work on second-cycle ligands for dihydroxylation indicated
that such substrates were good ligands for osmium, and the
fact that catalytic aminohydroxylation was possible indicated
that the binding constants should not be so high as to pre-
clude useful catalyst turnover.[16]
Amino alcohols 39 and 40 were both synthesized in five
steps from commercially available cis-4-heptenal (15) in
49% and 29% overall yields, respectively.[12] Thus, Henry re-
action, alcohol protection, nitro reduction, and amine pro-
tection, all followed by alcohol deprotection, provided a
facile route to these compounds (Scheme 7). Similarly,
amino alcohol 44 was synthesized from commercially avail-
able trans-4-decenal (41) in 50% overall yield utilizing the
same five-step sequence. The hydroxy-amide 45 was also
constructed from 41 and resolved by separation of its Mosh-
erꢂs ester derivatives to furnish enantiopure material.[12]
Scheme 8. Oxidative cyclization to form amino-THFs 46–49.
1240
www.chemasianj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2009, 4, 1237 – 1247

Osmium-Mediated Oxidative Cyclizations
stereospecific with respect to addition across the alkene
moiety.[17] The enantiomeric excess of THF 49 was deter-
mined by conversion into the Mosherꢂs ester derivative and
comparing the 19F NMR spectra with those for the deriva-
tive of racemic material.[12] However, it was evident that car-
bamate-protected amino alcohol 40 cyclized much more
slowly, and in slightly lower yield, than its sulfonamide ana-
logue 39. Moreover, a carbamate-protected nitrogen atom
was consisitently shown to be a worse ligand for osmium
than a sulfonamide-protected nitrogen atom throughout
these studies (see below).
Mosherꢂs ester derivatives and by comparing the 19F NMR
spectra with those for the derivatives of racemic material.[12]
Subjection of compounds 53, 55, and 57 to catalytic
osmium tetroxide in acidic dichloromethane furnished the
corresponding pyrrolidines 58, 59, and 60 in high yields
(Scheme 10). The use of camphorsulfonic acid (CSA) in di-
chloromethane gave the pyrrolidine products in higher
yields than the TFA/acetone/water mixture that had proven
optimal for diol cyclization.
Pyrrolidine Formation Using Oxidative Cyclization
With an effective protocol in hand for the formation of
THFs and amino-THFs, attention was turned to the con-
struction of pyrrolidines by osmium-catalyzed oxidative cyc-
lization from structurally-related amino alcohols to those
shown above. The synthesis of the pyrrolidine scaffold in a
concise, stereodefined manner has remained a challenging
task in organic chemistry.[18] However, a novel nitrogen cyc-
lization procedure would allow facile access to this structur-
al moiety commonly found in many natural products.[19] Ac-
cordingly, amino alcohols 53, 55, and 57 were readily pre-
pared from enantiopure glycidol utilizing cross-metathesis
and Wittig methodology (Scheme 9).[12] Thus, allyl Grignard
addition to commercially available trityl glycidol 50 and sub-
sequent Mitsunobu reaction provided protected amino alco-
hol 52, which was elaborated into amino alcohols 53, 55, and
57 as shown. The trans and cis geometries of 55 and 57 were
assigned from the olefinic coupling constants of 15.7 and
9.8 Hz, respectively. The enantiomeric excesses of amino al-
cohols 55 and 57 were determined by conversion into the
Scheme 10. Oxidative cyclization to form pyrrolidines 58–60. CSA=(ꢀ)-
camphorsulfonic acid.
Oxidative cyclization was also attempted with only one
mole percent of osmium catalyst. Pleasingly, pyrrolidines 58
and 59 were formed in 75% and 80% yield, respectively.
Amino alcohol 57 was considerably slower to cyclize than
the others owing to the steric bulk of the trans tert-butyl-
substituted olefin, but it proved possible to achieve a viable
reaction with one mole percent of catalyst by performing
the reaction at 408C. The relative product stereochemistries
were determined by X-ray crystal structure analyses of 58,
59, and 60 and the enantiomeric excess of 59 and 60 by for-
mation of the Mosherꢂs ester derivatives. The X-ray crystal
structures again confirmed that the reaction was stereoselec-
tive for formation of 2,5-cis heterocycles and stereospecific
with respect to syn addition across the alkene. Thus, this
new nitrogen cyclization reaction displayed the same stereo-
chemical preferences as the parent THF-forming reactions,
which was taken as good evidence that a similar mechanism
was in operation. However, pyrrolidine formation was limit-
ed to the use of sulfonamide protecting groups, since the
cyclization of the Cbz analogue of amino alcohol 53 (not
shown) failed under a range of reaction conditions utilizing
catalytic osmium (in agreement with the general observation
that a sulfonamide-protected nitrogen atom was a better
ligand for osmium than a carbamate-protected nitrogen
atom).
Scheme 9. Synthesis of amino alcohols 53, 55, and 57. Tr=trityl; Boc=
tert-butoxycarbonyl; DIAD=diisopropyl azodicarboxylate; KHMDS=
potassium bis(trimethylsilyl)amide.
Next, 1,3-amino alcohols 65 and 66 were synthesized from
the commercially available aldehydes 15 and 41 to examine
Chem. Asian J. 2009, 4, 1237 – 1247
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
1241

T. J. Donohoe et al.
FULL PAPERS
the effects of different initiator functionalities upon the oxi-
dative cyclization reaction. Hence, aldolization, lithium bor-
ohydride ester reduction, selective protection of the primary
hydroxyl group as the corresponding trityl ether, Mitsunobu
reaction, and subsequent deprotection of both acid-labile
protecting groups furnished the amino alcohols 65 and 66
(Scheme 11).
Mechanistic Investigations
Having observed that pyrrolidine formation was optimal
with CSA in dichloromethane, while THF formation re-
quired the use of TFA in aqueous acetone, we were in-
trigued as to the relative rates of formation of the two types
of heterocycle and whether it would be possible to form a
pyrrolidine selectively over a THF (or vice versa) when a
choice existed. To this end, competition substrate 72 was
prepared in five steps from commercially available 4-pente-
nal (69) in 24% overall yield.[12] Thus, Henry reaction, alco-
hol protection, nitro reduction, and amine protection fol-
lowed by alcohol deprotection provided amino alcohol 72 as
a
single diastereoisomer (Scheme 13). Intriguingly, the
Scheme 11. Synthesis of 1,3-amino alcohols 65 and 66.
Compounds 65 and 66 were then subjected to the oxida-
tive reaction conditions that had proven optimal for the cyc-
lization of 1,2-amino alcohols, to form pyrrolidines 67 and
68 as single diastereoisomers by 13C NMR (Scheme 12).
Scheme 13. Preparation of competition substrate amino alcohol 72.
Henry reaction of 69 was essentially unselective, furnishing
a 1.6:1 diastereomeric mixture. However, when this mixture
was reduced with lithium aluminum hydride, a single diaste-
reomer was formed (after tosylation and alcohol deprotec-
tion), suggesting deprotonation a to the nitrogen atom had
occurred during the reaction, with subsequent selective re-
duction.
Oxidative cyclization of 72 utilizing TFA in aqueous ace-
tone and CSA in dichloromethane furnished pyrrolidine 73
in 46% and 61% yields, respectively (Scheme 14). The
structure of pyrrolidine 73 was confirmed by X-ray crystal
structure analysis.
Scheme 12. Oxidative cyclization of 1,3-amino alcohol initiators 65 and
66.
Interestingly, no THF formation was observed when
either set of reaction conditions was applied to 72. This ex-
However, the times taken for the reactions to reach full con-
version were longer and the yields poorer than for the corre-
sponding 1,2-amino alcohols, precluding the use of lower
catalyst loadings. Presumably, the 1,3-amino alcohols were
again poorer ligands for the osmium metal, resulting in
lower concentrations of the “loaded” 1,3-amino alcohol
osmate ester cyclization precursors and therefore slower
overall reaction rates. The relative stereochemistry of 67
was confirmed by X-ray crystal structure analysis;[17] given
this result and the substantial precedent from the osmium-
mediated oxidative cyclization reactions we have studied,
we make the assumption that addition across the alkene
moiety for 66 was also syn.
Scheme 14. Oxidative cyclization of competition amino alcohol 72.
1242
www.chemasianj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2009, 4, 1237 – 1247

Osmium-Mediated Oxidative Cyclizations
periment exemplified the overriding preference for the for-
mation of nitrogen over oxygen heterocycles, providing fur-
Role of Acid in Oxidative Cyclization
ꢁ
ther evidence for the enhanced activity of an Os N unit
Interestingly, when the azaglycolate osmate ester 75 was ex-
posed to basic or neutral conditions, no cyclization was ob-
served. However, treatment with acid furnished the corre-
sponding pyrrolidine 58 in 91% yield. We believe that pro-
tonation of an Os=O unit in the intermediate osmate ester
allows the bound osmium species to participate in a pericy-
clic process with the pendant alkene, facilitating cyclization.
Protonation would be expected to lower the energy of the
orbitals around Os, thus increasing the interaction with the
HOMO of a relatively electron-rich alkene.[22]
ꢁ
over an Os O bond when a choice exists, as previously
found for both aminohydroxylation and diamination.[20] To
investigate why this should be the case for the oxidative cyc-
lization reaction, we developed a stoichiometric model
(Scheme 15). Hence, treatment of amino alcohols 53 and 74
with stoichiometric osmium(VI) and TMEDA at pH 7 gave
the stable azaglycolate osmate esters 75 and 76.[21]
Osmium Oxidation State Considerations
The condensation of osmium(VI) with amino alcohols 53
and 74 (and subsequent cyclization) revealed that osmi-
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
rimental to yields and reaction times, since substrate dihy-
droxylation could occur instead of cyclization (see above).
Hence, the role of the sacrificial alkene in the catalytic pro-
tocol is to be preferentially dihydroxylated by the excess
osmiumACHTNUTRGNE(UNG VIII) to give a greater concentration of osmium-
(VI) in the reaction mixture. Therefore, the mechanism
shown previously in Figure 1 must be treated as an oversim-
plification, since it is likely that a dihydroxylated molecule
of TCA remains bound to the osmium catalyst throughout
the catalytic cycle, exerting an effect similar to the “citrate
effect” noted by Sharpless.[10] This osmium species can dihy-
droxylate other molecules of TCA upon oxidation to
osmiumACHTNUTRGNE(NUG VIII), then suffer ligand exchange with a substrate
molecule before undergoing an acid-mediated oxidative cyc-
lization reaction across the pendant alkene.
Scheme 15. Mechanistic probe for oxidative cyclization. TMEDA=
N,N,N’,N’-tetramethylethylenediamine.
Gratifyingly, when compounds 75 and 76 were exposed to
acidic dichloromethane, efficient cyclization to the corre-
sponding pyrrolidines 58 and 77 occurred. The structure of
azaglycolate osmate ester 76 was confirmed by X-ray crystal
analysis, which revealed that the nitrogen atom bound to
osmium was planar rather than pyramidal. This observation
suggests significant double bond character between the ni-
trogen and osmium atoms. We suggest that the lesser elec-
Stoichiometric Probe of Osmium Oxidation States
To determine how the oxidation state of osmium affected
the propensity of an osmate ester to undergo cyclization, we
performed some stoichiometric studies on geranyl trichlor-
oacetamide 78 (Scheme 16). These experiments were con-
ducted using a system originally developed by Sharpless and
co-workers to examine the effect of the second catalytic
cycle in the asymmetric dihydroxylation reaction.[4]
ꢁ
tronegativity of nitrogen over oxygen allows for better Os
ꢁ
N p overlap than for Os O, leading to greater participation
ꢁ
of the Os N unit in a pericyclic cyclization process. This ra-
tionale would help to explain why there is a greater propen-
ꢁ
ꢁ
sity for cyclization of an Os N unit over an Os O unit.
Moreover, the sp3 nature of a sulfonamide-protected ni-
trogen atom compared to the sp2 nature of a carbamate-pro-
Thus, addition of osmium tetroxide to quinuclidine in di-
chloromethane at ꢁ788C furnished the orange osmium-
ꢁ
tected nitrogen atom suggests that the Os N interaction
(within an azaglycolate osmate ester) will be weaker for a
carbamate-protected nitrogen atom (which will lose delocal-
ization upon binding to osmium) than a sulfonamide-pro-
tected nitrogen atom (which will not). This effect was evi-
dent from the slower cyclization of carbamate 40 over sulfo-
namide 39 in the formation of amino-THFs 47 and 46, re-
spectively (see Scheme 8 above). Also, the failure of carba-
mate-protected amino alcohols to cyclize to the
corresponding pyrrolidines was in stark contrast to the suc-
cessful cyclization of sulfonamide-protected amino alcohols
(see above).
ACHTUNGTREN(NGNU VIII) complex 79. Subsequent addition of diene 78 and
warming to ꢁ508C resulted in the formation of the green
osmium(VI) ester 80.[23] Exposure of this osmate ester to
basic conditions only liberated the corresponding diol. How-
ever, acidification with CSA gave the THF 82 in 61% yield
(whose structure was proven by conversion to the corre-
sponding para-nitrobenzoyl derivative and subsequent X-ray
crystal structure analysis), providing further evidence that
osmium(VI) facilitates oxidative cyclization. Interestingly,
treatment of osmate ester 80 with TMO gave the red
osmate ester 81[21] (not characterized), which also furnished
Chem. Asian J. 2009, 4, 1237 – 1247
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
1243

T. J. Donohoe et al.
FULL PAPERS
In addition to the simplified catalytic cycle shown in
Figure 2, osmiumACTHNUTRGNEUNG(VIII) is present in the reaction mixture.
We have no evidence to suggest that a substrate molecule
(e.g., a diol or amino alcohol) can condense with osmium-
AHCTUNGTREG(NNNU VIII), since dihydroxylation of the starting material ap-
pears to be a more favorable process.[12] Hence, this side-re-
action is tempered by the presence of TCA, which provides
a higher concentration of osmium(VI), promoting oxidative
cyclization over dihydroxylation. In addition, it is quite pos-
sible that the chelated osmium(VI) osmate ester could un-
dergo oxidation to the corresponding chelated osmium-
AHCTUNGTREG(NNNU VIII) species, which could also cyclize to give the product
heterocycle (see Scheme 16 above).[4] This would liberate os-
mium(VI) to the reaction mixture, available for further cata-
lytic reactions. Of course, a bound osmiumACTHUNRTGNEUNG(VIII) species
also has the potential to dihydroxylate other substrates and
TCA molecules instead of undergoing cyclization. Thus, the
proposed catalytic cycle in Figure 2 should be viewed as a
simplified platform upon which several other catalytic pro-
cesses operate, many with the same net result.
Scheme 16. Stoichiometric investigations to determine whether osmium
oxidation state affects the propensity for oxidative cyclization (osmium
oxidation states are shown in parentheses). P=trichloroacetamide.
Conclusions
THF 82 upon acidification with TMO. This experiment re-
veals that in the right setting osmiumACTHNUTRGNEUNG(VIII) is also capable
of promoting acid-mediated oxidative cyclization reactions.
A powerful method for the construction of five-membered
heterocycles has been developed using the catalytic
osmium-mediated oxidative cyclization reaction. Judicious
choice of chelating functionalities in the starting material
gave rise to THF and pyrrolidine products in moderate to
high yields, with near total selectivity for the formation of
2,5-cis heterocycles. 1,2-Diols, 1,2-amino alcohols, 1,2-hy-
droxy amides, and 1,3-diols were shown to cyclize to the cor-
responding THFs, while pyrrolidine formation was observed
upon cyclization of 1,2- and 1,3-amino alcohols. Further-
more, enantiopure starting materials were shown to cyclize
efficiently, allowing access to enantiopure heterocycles. Ad-
ditionally, several mechanistic insights were revealed
through competition experiments and stoichiometric studies.
Thus, a mechanistic rationale was developed and the scope
and limitations of this methodology were explored. The ap-
plication of this synthetic tool to the construction of com-
plex natural products bearing THF and pyrrolidine units is
currently underway in our laboratories and will be reported
in due course.
Proposed Catalytic Cycle
Although the exact nature of the osmium oxidation state
during the catalytic reaction is unclear, the following conclu-
sions can be drawn regarding a simplified osmium(VI)–os-
mium(IV) cycle: 1) osmium(VI) chelates efficiently to the
substrate (see Scheme 15 above); 2) the subsequent os-
mium(VI) ester 83 suffers an acid-mediated oxidative cycli-
zation reaction via transition structure 84 (Figure 2); 3) the
resultant weakly bound osmium(IV) ester is hydrolyzed to
give the product heterocycle and osmium(IV), which is re-
oxidized.
Experimental Section
General: Tetrahydrofuran and dichloromethane were purified prior to
use by filtration through two activated alumina columns (activated basic
aluminum oxide, Brockmann I, standard grade, ca. 150 mesh, 58 ꢃ). Re-
agents obtained from Acros, Aldrich, Avocado, Fluka, and Lancaster fine
chemicals suppliers were used directly. Flash column chromatography
was carried out using silica gel 60 (0.040–0.063 mm; Merck) using head
pressure by means of head bellows. Thin layer chromatography was per-
formed on commercially available precoated aluminum-backed plates
(0.25 mm silica gel with fluorescent indicator UV254). Visualization was
achieved by either the quenching of UV fluorescence, KMnO4, or vanillin
stain. 1H and 13C NMR spectra were recorded on a Bruker AVANCE
Figure 2. Simplified oxidative cyclization catalytic cycle for an osmi-
ACHTUNGTRENNUNGum(VI)–osmium(IV)-mediated process (a diol has been arbitrarily
chosen as the cyclization substrate and the dihydroxylated TCA ligand
that is presumably bound to osmium has been omitted for clarity).
1244
www.chemasianj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2009, 4, 1237 – 1247

Osmium-Mediated Oxidative Cyclizations
AV400 (400 MHz and 100.6 MHz), Bruker DPX400 (400 MHz and
100.6 MHz), or a Bruker AVANCE AV500 (500 MHz and 125.7 MHz)
spectrometer. Signal positions were recorded in d (ppm) with the abbre-
viations s, d, t, q, qu, br, app, and m denoting singlet, doublet, triplet,
quartet, quintet, broad, apparent, and multiplet, respectively. All NMR
chemical shifts were referenced to residual solvent peaks or to SiMe4 as
an internal standard. All coupling constants, J, are quoted in Hz. IR spec-
tra were recorded on a Bruker Tensor 27 FTIR spectrometer. Spectra
were analyzed either as thin films between NaCl plates, KBr disks, or in
a chloroform solution cell. Mass spectra (m/z) and HRMS were recorded
under the conditions of electrospray (ESI), chemical (CI), and field (FI)
ionization. Melting points were obtained using a Leica VMTG heated-
stage microscope and are uncorrected. “Petrol” refers to the fraction of
petroleum ether boiling in the range 40–608C unless otherwise stated and
“ether” refers to diethyl ether.
(ꢀ)-(E)-8-(Benzyloxy)-6-methyloct-6-ene-1,3-diol (9): Lithium borohy-
dride (0.21 g, 9.7 mmol) was added to a stirred solution of hydroxy-ester
8 (2.0 g, 6.4 mmol) in THF (50 mL) at 08C, then warmed to room tem-
perature and stirred for 4 h. The reaction was quenched with water
(50 mL) and extracted with ethyl acetate (3ꢄ30 mL). The combined or-
ganic extracts were washed with brine (50 mL), dried over MgSO4, and
concentrated. Purification by flash column chromatography (SiO2, 1:2
acetone/petrol) gave diol 9 (1.6 g, 6.0 mmol, 92%) as an oil; IR (thin
film): n˜max =3408, 2936, 1719, 1452, 1277, 1068 cmꢁ1 1H NMR (CDCl3,
;
400 MHz): dH =7.36–7.35 (4H, m), 7.31–7.27 (1H, m), 5.44 (1H, t, J 6.7),
4.51 (2H, s), 4.03 (2H, d, J 6.7), 3.89–3.76 (3H, m), 2.73 (2H, brs), 2.22–
2.06 (2H, m), 1.71–1.58 ppm (7H, m); 13C NMR (CDCl3, 100 MHz): dC =
140.3, 138.4, 128.4, 127.8, 127.6, 121.1, 72.2, 71.8, 66.5, 61.6, 38.3, 35.6,
35.5, 16.5 ppm; MS m/z (ESI+) 287 (100%, M+Na+); HRMS (ESI+)
calcd for C16H24O3Na [M+Na+]: 287.1618; found: 287.1618 (+0.1 ppm).
General pocedure A: Aldol reaction: nButyllithium (1.6m in hexanes,
(ꢀ)-(S)-2-(Benzyloxy)-1-((2S,5R)-5-(2-hydroxyethyl)-2-methyltetrahydro-
furan-2-yl)ethanol (18): The diol 9 (80 mg, 0.30 mmol) was subjected to
general procedure B. Purification by flash column chromatography (flori-
sil, eluting with 4:1 petrol/acetone) gave the THF 18 (69 mg, 0.25 mmol,
82%) as an oil; IR (thin film): n˜max =3417, 2930, 1680, 1453, 1260,
3.1 equiv) was added dropwise to
a solution of diisopropylamine
(3.1 equiv) in THF (7.7 mL per mmol substrate) and stirred for 20 min at
08C before cooling to ꢁ788C. Freshly distilled ethyl acetate (3.0 equiv)
was added dropwise and the mixture was stirred for a further 20 min. Al-
dehyde (1.0 equiv) was added dropwise and the mixture stirred at ꢁ788C
for 2 h. The reaction was quenched by dropwise addition of saturated
aqueous NH4Cl (5 mL) and extracted with ether (3ꢄ25 mL). The com-
bined organic extracts were washed with brine (25 mL), dried over
Na2SO4, and concentrated to yield the crude product, which was purified
as indicated.
1073 cmꢁ1 1H NMR (CDCl3, 400 MHz): dH =7.35–7.27 (5H, m), 4.56
;
(2H, dd, J 18.9, 11.8), 4.23–4.16 (1H, m), 3.77 (2H, t, J 5.5), 3.67 (1H,
dd, J 7.4, 3.4), 3.59 (1H, dd, J 9.7, 3.4), 3.49 (1H, dd, J 9.7, 7.4), 2.81 (2H,
brs), 2.17–1.99 (2H, m), 1.78–1.74 (2H, m), 1.69–1.56 (2H, m), 1.17 ppm
(3H, s); 13C NMR (CDCl3, 100 MHz): dC =137.9, 128.4, 127.8, 127.7, 84.4,
78.5, 75.5, 73.5, 71.2, 61.1, 37.5, 34.4, 31.9, 22.7 ppm; MS m/z (ESI+) 303
(100%, M+Na+); HRMS (ESI+) calcd for C16H24O4Na [M+Na+]:
303.1567; found: 303.1567 (+0.0 ppm).
General procedure B: TFA-mediated oxidative cyclization: The cycliza-
tion substrate (1.0 equiv) and trimethylamine N-oxide dihydrate
(5.0 equiv) were dissolved in a 9:1 mixture of acetone/water (20 mL per
mmol substrate) and the solution cooled to 08C. trans-Cinnamic acid
(5.0 equiv) and trifluoroacetic acid (10 mL per mmol substrate) were
added, followed immediately by potassium osmate dihydrate (5 mol%).
The reaction was warmed to room temperature and stirred for 16 h, then
quenched by addition of solid sodium sulfite (0.5 equiv) and stirred for
30 min. The resulting mixture was cooled to 0 C and neutralized by care-
ful addition of aqueous NaOH (40% by weight, saturated with NaCl).
Ethyl acetate (20 mL) was added and the layers separated. The organic
layer was washed with water (20 mL) and the combined aqueous phases
extracted with ethyl acetate (3ꢄ25 mL). The combined organic extracts
were dried over Na2SO4 and concentrated to give the crude product,
which was purified as indicated.
(ꢀ)-(E)-9-(Benzyloxy)-2,7-dimethylnon-7-ene-2,4-diol (10): Methylmag-
nesium chloride (3.0m solution in THF, 1.9 mL, 5.7 mmol) was added
dropwise over 2 min to a solution of hydroxy-ester 8 (350 mg, 1.1 mmol)
in THF at ꢁ788C under an atmosphere of argon. The mixture was
warmed to room temperature and stirred for 16 h. Saturated aqueous
NH4Cl (5 mL) was added followed by ethyl acetate (10 mL). The separat-
ed aqueous layer was extracted with ethyl acetate (2ꢄ20 mL) and the
combined organic layers were washed with brine (20 mL), dried over
MgSO4, and then concentrated to give the crude product. Purification by
flash column chromatography (SiO2, 1:4 acetone/petrol) to give the terti-
ary alcohol 10 (240 mg, 0.83 mmol, 73%) as a viscous oil; IR (thin film):
n˜max =3420, 2971, 1718, 1452, 1377, 1275, 1118 cmꢁ1 1H NMR (CDCl3,
;
400 MHz): dH =7.35–7.28 (5H, m), 5.44 (1H, t, J 6.8), 4.51 (2H, s), 4.01
(2H, d, J 6.8), 4.03–3.97 (1H, m), 3.51 (1H, brs), 3.05 (1H, brs), 2.20–
2.05 (2H, m), 1.66 (3H, s), 1.68–1.49 (4H, m), 1.30 (3H, s), 1.26 ppm
(3H, s); 13C NMR (CDCl3, 100 MHz): dC =140.4, 138.4, 128.4, 127.8,
127.6, 121.0, 72.1, 71.7, 69.5, 66.5, 47.7, 36.1, 35.5, 32.1, 27.7, 16.5 ppm;
MS m/z (ESI+) 315 (100%, M+Na+); HRMS (ESI+) calcd for
C18H28O3Na [M+Na+]: 315.1936; found: 315.1930 (ꢁ0.2 ppm).
(ꢀ)-1-((2R,5S)-5-((S)-2-(Benzyloxy)-1-hydroxyethyl)-5-methyltetrahydro-
furan-2-yl)-2-methylpropan-2-ol (19): The tertiary alcohol 10 (105 mg,
0.36 mmol) was subjected to general procedure B. Purification by flash
column chromatography (florisil, eluting with 3:1 petrol/acetone) gave
General procedure C: CSA-mediated oxidative cyclization: The cycliza-
tion substrate (1.0 equiv), (ꢀ)-camphorsulfonic acid (6.0 equiv), and tri-
methylamine N-oxide dihydrate (4.0 equiv) were dissolved in CH2Cl2
(25 mL per mmol substrate). trans-Cinnamic acid (5.0 equiv) was added,
followed immediately by osmium tetroxide (5 mol%), and the resulting
solution was stirred at room temperature for 16 h. In the case of starting
material remaining after this time, a further portion of (ꢀ)-camphorsul-
fonic acid (6.0 equiv) was added and the reaction monitored by TLC
until complete. The reaction was quenched by addition of solid sodium
sulfite (0.5 equiv) and stirred for 30 min. The reaction mixture was
washed with aqueous NaOH (2m, 25 mL) and extracted with ethyl ace-
tate (3ꢄ25 mL). The combined organic extracts were dried over Na2SO4
and concentrated to give the crude product, which was purified as indi-
cated.
the THF 19 (74 mg, 0.24 mmol, 67%) as an oil; IR (thin film): n˜max
=
3426, 2970, 1719, 1275, 1095 cmꢁ1 1H NMR (CDCl3, 400 MHz): dH =
;
7.37–7.28 (5H, m), 4.55 (2H, s), 4.36–4.29 (1H, m), 3.65 (1H, dd, J 6.9,
3.3), 3.60 (1H, dd, J 9.7, 3.3), 3.49 (1H, dd, J 9.7, 6.9), 2.20–2.13 (1H, m),
2.09–2.02 (1H, m), 1.72 (1H, dd, 13.3, 10.6), 1.62–1.53 (3H, m), 1.28 (3H,
s) 1.22 (3H, s), 1.18 ppm (3H, s); 13C NMR (CDCl3, 100 MHz): dC =
138.0, 128.5, 127.8, 127.5, 85.0, 76.5, 75.2, 73.5, 71.2, 70.5, 47.6, 33.9, 32.6,
30.8, 28.2, 23.2 ppm; MS m/z (ESI+) 331 (100%, M+Na+); HRMS
(ESI+) calcd for C18H28O4Na [M+Na+]: 331.1880; found: 331.1876
(+1.0 ppm).
(ꢀ)-(E)-Ethyl 8-(benzyloxy)-3-hydroxy-6-methyloct-6-enoate (8): (E)-6-
(Benzyloxy)-4-methylhex-4-enal (7; 0.75 g, 3.4 mmol) was subjected to
general procedure A. Purification by flash column chromatography
(SiO2, 1:2 ether:petrol) gave hydroxy-ester 8 (0.92 g, 3.0 mmol, 88%) as
an oil; IR (thin film): n˜max =3452, 2978, 1725, 1276 cmꢁ1; 1H NMR(CDCl3,
400 MHz): dH = 7.35–7.27 (5H, m), 5.44 (1H, t, J 6.8), 4.51 (2H, s), 4.18
(2H, q, J 7.2), 4.03 (2H, d, J 6.8), 4.01–3.97 (1H, m), 2.97 (1H, brs), 2.51
(1H, dd, J 16.4 and 3.4), 2.43 (1H, dd, J 16.4 and 8.7), 2.26–2.19 (1H, m),
2.15–2.05 (1H, m), 1.66 (3H, s), 1.62–1.53 (2H, m), 1.28 ppm (3H, t, J
7.2); 13C NMR (CDCl3, 100 MHz): dC = 172.9, 139.8, 138.5, 128.3, 127.8,
127.5, 121.3, 72.1, 67.7, 66.5, 60.7, 41.3, 35.4, 34.4, 16.5, 14.2 ppm; MS m/z
(ESI+) 329 (100%, M+Na+); HRMS (ESI+) calcd for C18H26O4Na
[M+Na+]: 329.1723; found: 329.1713 (+3.2 ppm).
(ꢀ)-(Z)-Ethyl 8-(benzyloxy)-3-hydroxy-6-methyloct-6-enoate (13): (Z)-6-
(Benzyloxy)-4-methylhex-4-enal 12 (0.730 g, 3.3 mmol) was subjected to
general procedure A. Purification by flash column chromatography
(SiO2, 1:2 ether/petrol) gave hydroxy-ester 13 (0.97 g, 3.2 mmol, 95%) as
an oil; IR (thin film): n˜max =3457, 2980, 1724, 1452, 1375, 1276, 1027 cmꢁ1
;
1H NMR (CDCl3, 400 MHz): dH =7.35–7.34 (4H, m), 7.30–7.27 (1H, m),
5.48 (1H, t, J 7.0), 4.52 (2H, s), 4.16 (2H, q, J 7.1), 4.05 (1H, dd, J 11.0,
Chem. Asian J. 2009, 4, 1237 – 1247
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
1245

T. J. Donohoe et al.
FULL PAPERS
7.3), 3.98–3.94 (2H, m), 3.36 (1H, brs), 2.44–2.42 (2H, m), 2.33–2.26
(1H, m), 2.15–2.09 (1H, m), 1.76 (3H, s), 1.59–1.53 (2H, m), 1.27 ppm
(3H, t, J 7.1); 13C NMR (CDCl3, 100 MHz): dC =172.7, 141.0, 138.2,
128.4, 127.9, 127.6, 122.0, 72.3, 67.0, 65.9, 60.6, 41.5, 34.4, 27.8, 23.2,
14.2 ppm; MS m/z (ESI+) 329, (100%, M+Na+); HRMS (ESI+) calcd
for C18H26O4Na [M+Na+]: 329.1723; found: 329.1726 (ꢁ0.9 ppm).
(CDCl3, 100 MHz): dC =82.7, 79.4, 73.2, 61.6, 37.3, 31.6, 25.8, 23.8,
10.4 ppm; MS m/z (ESI+) 197 (100%, M+Na+); HRMS (ESI+) calcd
for C19H18O3Na [M+Na+]: 197.1148; found: 197.1149 (ꢁ0.4 ppm).
(S)-4-Benzyl-3-((2S,3R,E)-8-(benzyloxy)-3-hydroxy-2,6-dimethyloct-6-
enoyl)oxazolidin-2-one (23): Dibutylboryl trifluoromethanesulfonate
(1.0m solution in ether, 2.6 mL, 2.6 mmol) was added dropwise over
5 min to a solution of (S)-4-benzyl-3-propionyloxazolidin-2-one (0.50 g,
2.1 mmol) in CH2Cl2 (20 mL) at 08C. Triethylamine (0.41 mL. 3.0 mmol)
was then added dropwise over 5 min to give a light yellow solution which
was cooled to ꢁ788C. (E)-6-(Benzyloxy)-4-methylhex-4-enal (7, 0.70 g,
3.2 mmol) was added dropwise over 5 min. After 30 min the solution was
warmed to 08C and stirred at that temperature for 4 h. Phosphate buffer
(pH 7, 3 mL) and methanol (5 mL) were then added followed by a mix-
ture of aqueous hydrogen peroxide (30%, 3 mL) and methanol (5 mL)
dropwise by syringe over 5 min. The mixture was stirred for 1 h and then
concentrated. The mixture was extracted with ether (3ꢄ50 mL) and the
combined organic extracts were washed sequentially with saturated aque-
ous sodium hydrogencarbonate (50 mL) and brine (50 mL), dried over
MgSO4, and then concentrated. The crude residue was purified by flash
column chromatography (SiO2, 1:9!1:4 acetone/petrol) to give the alco-
hol 23 (0.83 g, 1.8 mmol, 86%) as a viscous oil; ½aꢂ2D2 =+40.0 (c=1.0 in
CHCl3); IR (thin film): n˜max =3480, 2937, 1779, 1699, 1453, 1384, 1212,
(ꢀ)-(Z)-8-(Benzyloxy)-6-methyloct-6-ene-1,3-diol (14): Lithium borohy-
dride (0.093 g, 4.3 mmol) was added to a stirred solution of hydroxy-ester
13 (0.87 g, 2.9 mmol) in THF (20 mL) at 08C, then warmed to room tem-
perature and stirred for 4 h. The reaction was quenched with water
(20 mL) and extracted with ethyl acetate (3ꢄ20 mL). The combined or-
ganic extracts were washed with brine (20 mL), dried over MgSO4, and
concentrated. Purification by flash column chromatography (SiO2, 1:2
acetone/petrol) gave diol 14 (0.54 g, 2.1 mmol, 72%) as a viscous oil; IR
(thin film/cmꢁ1): n˜max =3408, 2936, 1719, 1452, 1277, 1068; 1H NMR
(CDCl3, 400 MHz): dH =7.36–7.27 (5H, m), 5.54 (1H, t, J 7.3), 4.53 (2H,
s), 4.09 (1H, dd, J 10.9, 7.7), 3.89 (1H, dd, J 10.9, 7.0), 3.77–3.71 (3H, m),
2.92 (2H, brs), 2.40 (1H, dt, J 13.2, 8.1), 2.05 (1H, dt, J 13.2, 6.0), 1.75
(3H, s), 1.67–1.56 ppm (4H, m); 13C NMR (CDCl3, 100 MHz): dC =142.6,
137.7, 128.4, 128.1, 127.9, 121.4, 72.6, 70.6, 65.6, 61.9, 38.3, 35.0, 27.6,
23.0 ppm; MS m/z (ESI+) 287 (100%, M+Na+); HRMS (ESI+) calcd
for C16H24O3Na [M+Na+]: 287.1618; found: 287.1618 (+0.0 ppm).
1110 cmꢁ1 1H NMR (CDCl3, 400 MHz): dH =7.37–7.28 (8H, m), 7.21
;
(ꢀ)-(R)-2-(Benzyloxy)-1-((2S,5R)-5-(2-hydroxyethyl)-2-methyltetrahy-
drofuran-2-yl)ethanol (20): Diol 14 (160 mg, 0.61 mmol) was subjected to
general procedure B. Purification by flash column chromatography (flori-
sil, eluting with 4:1 petrol/acetone) gave the THF 20 (140 mg, 0.49 mmol,
80%) as an oil; IR (thin film): n˜max =3410, 2935, 1718, 1452, 1374, 1276,
(2H, d, J 7.1), 5.46 (1H, t, J 6.8), 4.72–4.66 (1H, m), 4.51 (2H, s), 4.23–
4.16 (2H, m), 4.04 (2H, d, J 6.8), 3.97–3.93 (1H, m), 3.77 (1H, dq, J 6.9,
2.7), 3.25 (1H, dd, J 13.4, 3.3), 2.94 (1H, brs), 2.79 (1H, dd, J 13.4, 9.5),
2.30–2.22 (1H, m), 2.15–2.06 (1H, m), 1.74–1.65 (1H, m), 1.67 (3H, s),
1.60–1.52 (1H, m), 1.28 ppm (3H, d, J 6.9); 13C NMR (CDCl3, 100 MHz):
dC =177.4, 153.0, 139.9, 138.5, 135.0, 129.4, 129.0, 128.4, 127.8, 127.6,
127.5, 121.3, 72.1, 71.0, 66.6, 66.2, 55.1, 42.2, 37.8, 35.9, 31.7, 16.5,
10.6 ppm; MS m/z (ESI+) 474 (100%, M+Na+); HRMS (ESI+) calcd
for C27H33NO5Na [M+Na+]: 474.2251; found: 474.2254 (ꢁ0.8 ppm).
1072 cmꢁ1 1H NMR (CDCl3, 400 MHz) dH =7.38–7.29 (5H, m), 4.58
;
(2H, s), 4.23–4.06 (3H, m), 3.73 (1H, dd, J 7.7, 3.1), 3.66 (1H, dd, J 9.7,
7.7), 3.49 (1H, dd, J 9.7, 7.7), 2.21–2.08 (2H, m), 1.88–1.76 (2H, m), 1.64–
1.54 (2H, m), 1.28 ppm (3H, s); 13C NMR (CDCl3, 100 MHz): dC =138.0,
128.4, 127.7, 127.7, 84.2, 77.2, 75.4, 73.4, 71.1, 62.0, 34.9, 33.0, 31.6,
22.5 ppm; MS m/z (ESI+) 303 (100%, M+Na+); HRMS (ESI+) calcd
for C16H24O4Na [M+Na+]: 303.1567; found: 303.1568 (ꢁ0.6 ppm).
AHCTUNGTERG(NNUN 2R,3R,E)-8-(Benzyloxy)-2,6-dimethyloct-6-ene-1,3-diol (24): Lithium
borohydride (0.087 g, 4.0 mmol) was added to a stirred solution of alco-
hol 23 (1.2 g, 2.7 mmol) in THF (20 mL) containing water (1 mL) at 08C,
then warmed to room temperature and stirred for 16 h. The reaction was
quenched with water (20 mL) and extracted with ethyl acetate (3ꢄ
30 mL). The combined organic extracts were washed with brine (30 mL),
dried over MgSO4, and concentrated. Purification by flash column chro-
matography (SiO2, 1:6!1:3 acetone/petrol) gave diol 24 (0.52 g,
1.9 mmol, 71%) as a viscous oil; ½aꢂ2D2 =+4.0 (c=1.0 in CHCl3); IR (thin
(ꢀ)-(Z)-Ethyl 3-hydroxynon-6-enoate (16): cis-4-Heptenal 15 (0.580 mL,
4.5 mmol) was subjected to general procedure A. Purification by flash
column chromatography (SiO2, 1:3 ether/petrol) furnished hydroxy-ester
16 (0.89 g, 4.5 mmol, 100%) as an oil; IR (thin film): n˜max =3346, 1730,
1648, 1373, 1301, 1189 cmꢁ1 1H NMR (400 MHz, CDCl3): dH =5.43–5.37
;
(1H, m), 5.35–5.28 (1H, m), 4.17 (2H, q, J 7.2), 4.05–3.97 (1H, m), 3.01
(1H, brs), 2.50 (1H, dd, J 16.4, 3.4), 2.41 (1H, dd, J 16.4, 8.9), 2.21–2.11
(2H, m), 2.05 (2H, qu, J 7.5), 1.63–1.54 (1H, m), 1.51–1.42 (1H, m), 1.27
(3H, t, J 7.2), 0.95 ppm (3H, t, J 7.5); 13C NMR (100 MHz, CDCl3): dC =
173.0, 132.5, 128.1, 67.5, 60.7, 41.3, 36.4, 23.2, 20.5, 14.3, 14.2 ppm; MS m/
z (ESI+) 223 (100%, M+Na+); HRMS (ESI+) calcd for C11H20O3Na
[M+Na+]: 223.1305; found: 223.1307 (ꢁ0.9 ppm).
film): n˜max =3425, 2958, 1032 cmꢁ1 1H NMR (CDCl3, 400 MHz): dH =
;
7.35–7.27 (5H, m), 5.44 (1H, t, J 7.0), 4.51 (2H, s), 4.03 (2H, d, J 7.0),
3.79–3.78 (1H, m), 3.66 (2H, d, J 5.4), 2.82 (2H, m), 2.25–2.18 (1H, m),
2.11–2.05 (1H, m), 1.81–1.75 (1H, m), 1.66 (3H, s), 1.65–1.49 (2H, m),
0.89 ppm (3H, d, J 7.1); 13C NMR (CDCl3, 100 MHz): dC =140.4, 138.3,
128.4, 127.9, 127.6, 121.0, 73.9, 72.2, 66.9, 66.6, 39.2, 36.3, 31.7, 16.5,
10.3 ppm.
(ꢀ)-(Z)-Non-6-ene-1,3-diol (17): Lithium borohydride (0.760 g,
35.0 mmol) was added to a stirred solution of hydroxy-ester 16 (1.7 g,
8.7 mmol) in THF (110 mL) at 08C, warmed to room temperature, and
stirred for 16 h. The reaction was quenched with water (50 mL) and ex-
tracted with ether (3ꢄ30 mL). The combined organic extracts were
washed with brine (30 mL), dried over Na2SO4, and concentrated. Purifi-
cation by flash column chromatography (SiO2, eluting with 1:4 acetone/
petrol) gave diol 17 (1.1 g, 6.9 mmol, 79%) as an oil; IR (thin film):
(R)-2-((2R,5S)-5-((S)-2-(Benzyloxy)-1-hydroxyethyl)-5-methyltetrahydro-
furan-2-yl)propan-1-ol (25): Diol 24 (120 mg, 0.44 mmol) was subjected
to general procedure B. Purification by flash column chromatography
(florisil, eluting with 4:1 petrol/acetone) gave the THF 25 (105 mg,
0.36 mmol, 81%) as an oil; ½aꢂD22 =+9.5 (c=1.0 in CHCl3); IR (thin film):
n˜max =3428, 2967, 1715, 1454, 1372, 1095 cmꢁ1
;
1H NMR (CDCl3,
n˜max =3333, 1654, 1060 cmꢁ1 1H NMR (400 MHz, CDCl3): dH =5.45–5.31
;
400 MHz): dH =7.36–7.27 (5H, m), 4.58 (1H, d, J 11.8), 4.52 (1H, d, J
11.8), 4.11–4.06 (1H, m), 3.70–3.46 (5H, m), 2.79 (2H, brs), 2.11–1.98
(2H, m), 1.90–1.72 (2H, m), 1.65–1.58 (1H, m), 2.91 (3H, s), 0.87 ppm
(3H, d, J 7.1); 13C NMR (CDCl3, 100 MHz): dC =137.9, 128.5, 127.7,
127.7, 83.4, 81.6, 75.5, 73.5, 71.2, 65.5, 38.0, 34.9, 27.6, 22.1, 12.6ppm; MS
m/z (ESI+) 317 (100%, M+Na+), 295 (25%, MH+); HRMS (ESI+)
calcd for C17H26O4Na [M+Na+]: 317.1729; found: 317.1729 (+0.1 ppm).
(2H, m), 3.92–3.80 (3H, m), 2.58 (2H, s), 2.23–2.09 (2H, m), 2.10–2.03
(2H, m), 1.77–1.65 (2H, m), 1.62–1.48 (2H, m), 0.97 ppm (3H, t, J 7.5);
13C NMR (100 MHz, CDCl3): dC =132.5, 128.3, 72.1, 61.8, 38.3, 37.6, 23.4,
20.5, 14.3 ppm; MS m/z (ESI+) 181 (100%, M+Na+); HRMS (ESI+)
calcd for C9H18O2Na [M+Na]: 181.1199; found: 181.1203 (ꢁ2.4 ppm).
(ꢀ)-(R)-1-((2S,5R)-5-(2-Hydroxyethyl)tetrahydrofuran-2-yl)propan-1-ol
(21): Diol 17 (180 mg, 1.1 mmol) was subjected to general procedure B.
Purification by flash column chromatography (florisil, eluting with 3:1
petrol/acetone) gave the THF 21 (130 mg, 0.75 mmol, 66%) as an oil; IR
(ꢀ)-(R)-1-((2S,5R)-5-(2-Hydroxyethyl)-1-tosylpyrrolidin-2-yl)propan-1-ol
(67): Amino alcohol 65 (0.19 g, 0.61 mmol) was subjected to general pro-
cedure C. Flash column chromatography (SiO2, 15:85 acetone/petrol)
gave pyrrolidine 67 (0.091 g, 0.28 mmol, 46%) as needles; m.p. 86–878C;
(thin film): n˜max =3383, 2964, 1657, 1462, 1072 cmꢁ1 1H NMR (CDCl3,
;
400 MHz): dH =4.09 (1H, qu, J 6.7), 3.88 (1H, dt, J 7.4, 3.4), 3.81 (2H, t,
J 5.4), 3.72–3.68 (1H, m), 2.06–1.88 (2H, m), 1.82–1.75 (3H, m), 1.64–
1.55 (1H, m), 1.48–1.40 (2H, m), 1.00 ppm (3H, t, J 7.4); 13C NMR
IR (KBr disc): n˜max =3416, 1598, 1494, 1449, 1340, 1158, 1091 cmꢁ1
;
1H NMR (400 MHz, CDCl3): dH =7.73 (2H, d, J 8.2), 7.35 (2H, d, J 8.2),
4.04–3.84 (4H, m), 3.52 (1H, ddd, J 9.8, 7.8, 2.4), 2.67 (2H, brs), 2.45
1246
www.chemasianj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2009, 4, 1237 – 1247

Osmium-Mediated Oxidative Cyclizations
(3H, s), 1.98 (1H, tdd, J 12.8, 9.5, 6.7), 1.83 (1H, dddd, J 14.1, 10.9, 5.9,
3.0), 1.65–1.55 (2H, m), 1.45–1.36 (3H, m), 1.10 (1H, tt, J 12.6, 7.4),
1.03 ppm (3H, t, J 7.4); 13C NMR (100 MHz, CDCl3): dC =144.0, 133.9,
129.9, 127.7, 73.4, 66.7, 59.6, 59.4, 37.0, 30.1, 25.9, 23.2, 21.6, 10.8 ppm;
MS m/z (ESI+) 350 (60%, M+Na+), 386 (80%, M+NH4++MeCN), 677
(10%, 2m+Na+), 713 (100%, 2m+NH4++MeCN); HRMS (ESI+) calcd
for C16H25NO4SNa [M+Na+]: 350.1397; found: 350.1396 (+0.3 ppm).
[8] P. Grieco, Y. Musaki, D, Boxler, J. Am. Chem. Soc. 1975, 97, 1597–
1599.
[9] R. D. Hancock, A. E. Martell, Chem. Rev. 1989, 89, 1875–1914.
[10] P. Dupau, R. Epple, A. A. Thomas, V. V. Fokin, K. B. Sharpless,
Adv. Synth. Catal. 2002, 344, 421–433.
[11] a) V. V. Fokin, K. B. Sharpless, Angew. Chem. 2001, 113, 3563–3565;
b) M. A. Andersson, R. Epple, V. V. Fokin, K. B. Sharpless, Angew.
Chem. 2002, 114, 490–493; Angew. Chem. Int. Ed. 2002, 41, 472–
475.
[12] T. J. Donohoe, G. H. Churchill, K. M. P. Wheelhouse (nꢀe Gosby),
P. A. Glossop, Angew. Chem. 2006, 118, 8193–8196; Angew. Chem.
Int. Ed. 2006, 45, 8025–8028.
(ꢀ)-(S)-1-((2S,5R)-5-(2-Hydroxyethyl)-1-tosylpyrrolidin-2-yl)hexan-1-ol
(68): Amino alcohol 66 (0.23 g, 0.65 mmol) was subjected to general pro-
cedure C. Flash column chromatography (SiO2, eluting with 15:85 ace-
tone/petrol) gave pyrrolidine 68 (0.13 g, 0.34 mmol, 53%) as an oil; IR
(thin film): n˜max =3426, 1598, 1494, 1454, 1338, 1158, 1092 cmꢁ1 1H NMR
;
[13] J. R. Gage, D. A. Evans, Org. Synth. Coll. Vol. VIII, 1993, 339–443.
[14] Rhenium: a) see reference [3]; cobalt: b) S. Inoki, T. Mukaiyama,
Chem. Lett. 1990, 67–70.
[15] For reports on the oxidative cyclization of 1,6-dienes using rutheni-
um, see: a) V. Piccialli, Tetrahedron Lett. 2000, 41, 3731–3733; b) S.
Roth, B. W. Stark, Angew. Chem. 2006, 118, 6364–6367; Angew.
Chem. Int. Ed. 2006, 45, 6218–6221; manganese: c) R. L. Cecil;
R. C. D. Brown, Tetrahedron Lett. 2004, 45, 7269–7271; R. C. D.
Brown, Tetrahedron Lett. 2004, 45, 7269–7271.
(400 MHz, CDCl3): dH =7.74 (2H, d, J 8.1), 7.35 (2H, d, J 8.1), 4.13–4.07
(1H, m), 3.98 (1H, ddd, J 12.2, 9.8, 3.8), 3.75 (1H, dt, J 12.2, 3.8), 3.49
(1H, app. q, J 7.7), 3.41 (1H, td, J 8.5, 2.1), 2.45 (3H, s), 2.37 (2H, brs),
1.76–1.55 (5H, m), 1.50–1.21 (9H, m,), 0.90 ppm (3H, t, J 7.0); 13C NMR
(100 MHz, CDCl3): dC =144.3, 134.0, 130.0, 127.7, 74.8, 66.6, 59.2, 58.9,
38.2, 34.0, 32.0, 30.1, 27.8, 25.1, 22.6, 21.6, 14.1 ppm; MS m/z (ESI+) 392
+
(10%, M+Na+), 428 (100%, M+NH4++MeCN), 797 (35%, 2m+NH4
+MeCN); HRMS (ESI+) calcd for C19H31NO4SNa [M+Na+]: 392.1866;
found: 392.1866 (+0.1 ppm).
[16] a) See reference [4]; b) K. MuÇiz, I. Almodovar, J. Streuff, M.
Nieger, Adv. Synth. Catal. 2006, 348, 1831–1835.
[17] Data were collected at low temperature using an Enraf–Nonius
KCCD diffractometer [Z. Otwinowski, W. Minor, Processing of X-
ray Diffraction Data Collected in Oscillation Mode, In Methods En-
zymol., Vol. 276 (Eds.: C. W. Carter, R. M. Sweet), Academic Press,
1997]. The crystal structures were solved using SIR92 [A. Altomare,
G. Cascarano, C. Giacovazzo, A. Guagliardi, M. C. Burla, G. Poli-
dori, M. Camalli, J. Appl. Crystallogr. 1994, 27, 435.] and refined
using the CRYSTALS software suite [P. W. Betteridge, J. R. Carruth-
ers, R. I. Cooper, K. Prout, D. J. Watkin, J. Appl. Crystallogr. 2003,
36, 1487], as per the CIF. Crystallographic data (excluding structure
factors) for compound 67 have been deposited with the Cambridge
Crystallographic Data Centre (CCDC 730223) and copies of these
data can be obtained free of charge via www.ccdc.cam.ac.uk/data_
request/cif.
[18] For examples of pyrrolidine synthesis using catalysis, see: a) C. V.
Galliford, M. A. Beenen, S. T. Nguyen, K. A. Scheidt, Org. Lett.
2003, 5, 3487–3490; b) Q. Yang, J. E. Ney, J. P. Wolfe, Org. Lett.
2005, 7, 2575–2578; c) C. Welter, A. Dahnz, B. Brunner, S. Streiff, P.
Dꢆbon, G. Helmchen, Org. Lett. 2005, 7, 1239–1242; d) F. A. Davis,
H. Xu, Y. Wu, J. Zhang, Org. Lett. 2006, 8, 2273–2276.
[19] For examples of pyrrolidines in natural products, see: a) J. W. Daly,
T. F. Spande, H. M. Garraffo, J. Nat. Prod. 2005, 68, 1556–1575;
b) N. Asano, R. J. Nash, R. J. Molyneux, G. W. Fleet, Tetrahedron:
Asymmetry 2000, 11, 1645–1680; J. D. Scott, R. M. Williams, Chem.
Rev. 2002, 102, 1669–1730.
Acknowledgements
We thank the EPSRC, Pfizer and AstraZeneca for funding this project,
Merck for unrestricted support, Oxford Chemical Crystallography Serv-
ice for the use of their instrumentation, and Jess Kershaw for X-ray crys-
tal structure analyses.
[1] For a review see: a) V. Piccialli, Synthesis 2008, 2585–2607; manga-
nese: b) E. Klein, W. Rojahn, Tetrahedron 1965, 21, 2353–2358;
c) J. E. Baldwin, M. J. Crossley, E. M. M. Lehtonen, J. Chem. Soc.
Chem. Commun. 1979, 918–920; d) D. M. Walba, M. D. Wand,
M. C. Wilkes, J. Am. Chem. Soc. 1979, 101, 4396–4397; e) R. C. D.
Brown, J. F. Keily, Angew. Chem. 2001, 113, 4628–4630; Angew.
Chem. Int. Ed. 2001, 40, 4496–4498; f) R. C. D. Brown, C. J. Bataille,
R. M. Hughes, A. Kenney, T. J. Luker, J. Org. Chem. 2002, 67,
8079–8085; ruthenium: g) P. H. J. Carlsen, T. Katsuki, V. S. Martin,
K. B. Sharpless, J. Org. Chem. 1981, 46, 3936–3938; h) L. Albarella,
D. Musumeci, D. Sica, Eur. J. Org. Chem. 2001, 997–1003; i) V. Pic-
cialli, N. Cavallo, Tetrahedron Lett. 2001, 42, 4695–4699; j) S. Roth,
S. Gohler, H. Cheng, C. B. W. Stark, Eur. J. Org. Chem. 2005, 4109–
4118; osmium: k) M. de Champdore, M. Lasalvia, V. Piccialli, Tetra-
hedron Lett. 1998, 39, 9781–9784.
[2] T. J. Donohoe, S. Butterworth, Angew. Chem. 2003, 115, 978–981;
Angew. Chem. Int. Ed. 2003, 42, 948–951.
[20] Aminohydroxylation: a) D. W. Patrick, L. K. Truesdale, S. A. Biller,
K. B. Sharpless, J. Org. Chem. 1978, 43, 2628–2638; diamination:
b) A. O. Chong, K. Oshima, K. B. Sharpless, J. Am. Chem. Soc.
1977, 99, 3420–3426.
[21] T. J. Donohoe, K. M. P. Wheelhouse (nꢀe Gosby), P. J. Lindsay-
Scott, P. A. Glossop, I. A. Nash, J. S. Parker, Angew. Chem. 2008,
120, 2914–2917; Angew. Chem. Int. Ed. 2008, 47, 2872–2875.
[22] For a related inverse electron-demand cycloaddition, see: L. R.
Domingo, Eur. J. Org. Chem. 2000, 2265–2272.
[3] a) R. M. Kennedy, S. Tang, Tetrahedron Lett. 1992, 33, 3729–3732;
b) S. Tang, R. M. Kennedy, Tetrahedron Lett. 1992, 33, 5299–5302.
[4] J. S. M. Wai, I. Markꢅ, J. M. Svendsen, M. G. Finn, E. N. Jacobsen,
K. B. Sharpless, J. Am. Chem. Soc. 1989, 111, 1123–1125.
[5] a) T. J. Donohoe, S. Butterworth, Angew. Chem. 2005, 117, 4844–
4846; Angew. Chem. Int. Ed. 2005, 44, 4766–4768; for a related
chromium-mediated cyclization, see: b) D. M. Walba, G. S. Stoudt,
Tetrahedron Lett. 1982, 23, 727–730.
[6] cis-Solamin: a) see reference [5a]; cis-sylvaticin: b) T. J. Donohoe,
R. M. Harris, J. Burrows, J. Parker, J. Am. Chem. Soc. 2006, 128,
13704–13705.
[7] M. Z. Hoemann, Z. Michael, K. A. Agrios, J. Aube, Tetrahedron
1997, 53, 11087–11098.
[23] M. Schrçder, A. J. Neielson, W. P. Griffith, J. Chem. Soc. Dalton
Trans. 1979, 1607–1611.
Received: May 7, 2009
Published online: July 3, 2009
Chem. Asian J. 2009, 4, 1237 – 1247
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
1247