pubs.acs.

org/joc

Unexpected Cleavage of 2-Azido-2-(hydroxymethyl)oxetanes:

Conformation Determines Reaction Pathway?
Elisa Farber, Jackson Herget, Jos
e A. Gasc

on, and Amy R. Howell*
Department of Chemistry, University of Connecticut, Storrs, Connecticut 06269-3060, United States

*Corresponding author. amy.howell@uconn.edu Phone: 860-486-3460. Fax: 860-486-2981.

Received July 6, 2010

An unanticipated cleavage of 2-azido-2-(hydroxymethyl)oxetanes is reported. In attempts to oxidize

the title oxetanyl alcohols to the corresponding carboxylic acids with RuO4, cleaved nitriles were
formed as the sole isolable products, while a closely related tetrahydrofuran gave solely the expected
carboxylic acid. Quantum chemical calculations suggest that the divergent outcomes are governed by
conformational differences in the azidoalcohols.

Introduction et al., demonstrated that Pd(II) could promote a ring expan-

sion involving an oxidative cleavage of cyclic 2-azidoalcohols
We have been interested in the synthesis of oxetane con-
to azaheterocycles.5
taining natural product derivatives. As part of this focus we Ruthenium tetroxide is well-known for oxidizing primary
targeted oxetane analogs of hydantocidin, a natural product alcohols to carboxylic acids, as well as for cleaving double bonds
which has shown potent herbicidal activity1 (Figure 1). In an to carbonyl products.6 A limited number of other C-C cleavage
attempt to oxidize a model system to the corresponding reactions mediated by RuO4 have been described. In 1994,
carboxylic acid, 2-azido-2-(hydroxymethyl)oxetane 1a was Ranganathan and co-workers reported the C-C cleavage of a
treated with RuO4, generated in situ from RuCl3 3 3H2O and β-hydroxyamide from a protein backbone using RuO4.7 Also,
NaIO4; only nitrile 2a was isolated. This was a surprising cleavage of β-hydroxyethers under similar oxidative conditions
outcome since Sano and co-workers previously reported that was noted by Ferraz and co-workers.8 However, C-C scission
azidoalcohol 3 was oxidized to carboxylic acid 4 in good of β-hydroxyazides with RuO4 has not been previously reported.
yield.2 No mention was made of nitrile formation (Figure 1). Due to the unexpected result when 1a was treated with RuO4, we
To our knowledge, there are only three examples of decided to investigate the reactivity of other 2-azido-2-(hydroxy-
β-hydroxyazide oxidative cleavage described in the literature methyl)oxetanes under these conditions.
(Figure 2). In 2004, Suarez and co-workers reported a radical
fragmentation with the use of a hypervalent iodine reagent.3 Results and Discussion
Then, in 2006, Ye and co-workers noted that oxidative
cleavage to nitriles resulted when β-azidoalcohols were treated 2-Azido-2-(hydroxymethyl)oxetanes 1 were synthesized from
with PCC.4 This outcome was attributed to the lability of the corresponding β-lactones in four steps (Table 1). β-Lactones
β-azidoaldehydes under the conditions. The third, by Chiba 6a-6c and 6e were prepared following the Mukaiyama aldol-
lactonization protocol developed by Yang and Romo.9 Com-
pound 6d was synthesized under Mitsunobu conditions as
(1) (a) Cseke, C.; Gerwick, B. C.; Crouse, G. D.; Murdoch, M. G.; Green,
S. B.; Heim, D. R. Pest. Biochem. Physiol. 1996, 55, 210. (b) Siehl, D. L.;
Subramanian, M. V.; Walter, E. W.; Lee, S.-F.; Anderson, R. J.; Toschi, (5) Chiba, S.; Xu, Y.-Z.; Wang, Y.-F. J. Am. Chem. Soc. 2009, 131, 12886.
A. G. Plant Physiol. 1996, 110, 753. (6) (a) Plietker, B. Synthesis 2005, 2453. (b) Arends, I. W. C. E.; Kodama,
(2) Sano, H.; Mio, S.; Kitagawa, J.; Shindou, M.; Honma, T.; Sugai, S. T.; Sheldon, R. A. Top. Organomet. Chem. 2004, 11, 277.
Tetrahedron 1995, 51, 12563. (7) Ranganathan, D.; Vaish, N. K.; Shah, K. J. Am. Chem. Soc. 1994, 116,
(3) Hernandez, R.; Leon, E. I.; Moreno., P.; Concepcion, R. F.; Suarez, 6545.
E. J. Org. Chem. 2004, 69, 8437. (8) Ferraz, H. M. C.; Longo, L. S. Org. Lett. 2003, 5, 1337.
(4) Fan, Q.; Ni, N.; Li, Q.; Zhang, L.; Ye, X. Org. Lett. 2006, 8, 1007. (9) Yang, H. W.; Romo, D. J. Org. Chem. 1997, 62, 4.

DOI: 10.1021/jo101328c Published on Web 10/18/2010 J. Org. Chem. 2010, 75, 7565–7572 7565
r 2010 American Chemical Society
JOC Article Farber et al.

FIGURE 1. Outcome of reaction of oxetane 1a and psico-furanose 3 with RuO4.

TABLE 1. Synthesis of 2-Azido-2-(hydroxymethyl)oxetanes

entry reactant yield (%) of 7 yield (%) of 1a

1
1 (a) R = CH2OTBDPS; R = CH3 55 67
2 (b) R = (CH2)6CH3; R1 = CH3 67 30
3 (c) R = (CH2)2Ph; R1 = CH3 7411 34
4 (d) R = H; R1 = Ph 7611 5614
5 (e) R = c-Hexyl; R1 = H 33b 32
a
Percent yield over 3 steps. bThis compound is somewhat volatile.

deduced from NOESY experiments.15 Compound 1d was

known.14
Initially, the conditions to oxidize oxetane 1a were based on
Kumaraswamy and co-workers’ report for the oxidation of a
2-hydroxymethyloxetane.16 First, RuO4 was generated from
RuCl3 3 3H2O (0.06 equiv) and NaIO4 (4 equiv) in a biphasic
solvent system (CCl4/CH3CN/H2O). Then, the supernatant was
added to a solution of 2-azido-2-(hydroxymethyl)oxetane 1a in
CH3CN, followed by additional NaIO4 (2 equiv). The resulting
mixture was stirred for 3 h at rt. NMR spectra (1H and 13C) of
the crude product showed only nitrile 2a and no carboxylic acid.
FIGURE 2. β-Hydroxyazide oxidative cleavages described in the However, the mass balance was low, and the yield of isolated
literature. nitrile was only 16%. To ascertain if other products were form-
ing and being degraded, the reaction was repeated and mon-
itored by NMR. After 15 min, starting material was still present;
previously described.10 β-Lactones 6 were then subjected to
however, after 30 min, no starting material remained. The NMR
methylenation with the Petasis reagent.11 Subsequent epox-
spectra (1H and 13C) of the crude product (after workup)
idation with acetone-free dimethyldioxirane,12 produced
showed the nitrile as the major component. In addition, there
1,5-dioxaspiro[3.2]hexanes 8 in quantitative yields.13 Diox-
were minor peaks which could not be assigned to any specific,
aspirohexanes 8 were then treated with azidotrimethylsilane,14
isolable byproduct. The NMR spectra (1H and 13C) of the crude
followed by deprotection of the primary alcohol with tetra-n-
product after 3 h of reaction was cleaner than the corresponding
butylammonium fluoride or potassium carbonate. The latter
30 min reaction. To better understand this result, two reactions,
deprotection method was utilized only for t-butyldiphenylsilyl-
one for 1 h and one for 2 h, were run. Again, the NMR spectra
containing oxetane 1a. The diastereomers of 1 were separable.
after workup were not as clean as the 3 h reaction, and in both
The relative stereochemistries of the diastereomers of 1a-c were
reactions the nitrile was the only isolated product. This suggested
that byproducts formed were further degraded at longer reac-
(10) Hmamouchi, M.; Prud’homme, R. E. J. Polym. Sci. Part A: Polym.
Chemistry 1991, 29, 1281. tion times and became either water-soluble or volatile. Then, to
(11) Dollinger, L. M.; Ndakala, A. J.; Hashemzadeh, M.; Wang, G.; clarify if the nitrile was stable under these conditions, nitrile 2b
Wang, Y.; Martinez, I.; Arcari, J. T.; Galluzzo, D. J.; Howell, A. R.; was subjected to the same conditions. After stirring the reaction
Rheingold, A. L.; Figuero, J. S. J. Org. Chem. 1999, 64, 7074.
(12) Ferrer, M.; Gibert, M.; Sanchez-Baeza, F.; Messeguer, A. Tetrahe-
dron Lett. 1996, 37, 3585. (15) See Supporting Information.
(13) Ndakala, A. J.; Howell, A. R. J. Org. Chem. 1998, 63, 6098. (16) Kumaraswamy, G.; Padmaja, M.; Markondaiah, B.; Jena, N.;
(14) Howell, A. R.; Ndakala, A. J. Org. Lett. 1999, 1, 825. Sridhar, B.; Kiran, M. U. J. Org. Chem. 2006, 71, 337.

7566 J. Org. Chem. Vol. 75, No. 22, 2010

Farber et al.
JOC Article
TABLE 2. Synthesis of Nitriles

entry reactant diastereomer yield (%) of 2 (%) of recovered 1

1 (a) R = CH2OTBDPS; R1 = CH3 antia 20 33
2 (b) R = (CH2)6CH3; R1 = CH3 syn/antia 26 47
3 (b) R = (CH2)6CH3; R1 = CH3 syna 23 26
4 (b) R = (CH2)6CH3; R1 = CH3 antia 13 25
5 (c) R = (CH2)2Ph; R1 = CH3 antia 15 24
6 (d) R = H; R1 = Ph antib 22 0c
7 (e) R = c-Hexyl; R1 = H d
16 0c
a
Syn/anti defined by relative stereochemistry of N3 and R. Syn/anti defined by relative stereochemistry of N3 and R1. cSee text for discussion.
b
d
Reaction done on minor diastereomer; relative stereochemistry not determined.

mixture at rt for 2 h, only nitrile (>90%) was recovered. Thus, would result. Further oxidative cleavage could lead to water-
any loss of material balance could not be associated with soluble and/or volatile byproducts.
degradation of the nitrile under the reaction conditions. To ascertain if the cleavage was being effected by RuO4 or
Next, in an attempt to minimize side reactions, no NaIO4 simply by NaIO4, β-hydroxyazide 1b was treated with 4 equiv of
was added after the RuO4 supernatant was transferred to the NaIO4 in CCl4/CH3CN/H2O, with no addition of RuO4. After
azidoalcohol solution. After 25 min at rt, the NMR spectra 25 min nitrile 2b was isolated in 36% yield; no starting material
(1H and 13C) of the crude product showed mostly starting was recovered. The fact that NaIO4 effects nitrile formation is
material and nitrile, and it was considerably cleaner than the interesting and will be discussed later. However, these condi-
previous reactions. On the basis of this result, these condi- tions were different from the conditions utilized in Table 2,
tions were utilized to study the behavior of other 2-azido- where only the supernatant was transferred. To mimic these
2-(hydroxymethyl)oxetanes 1. conditions the supernatant of a mixture of 4 equiv of NaIO4 in
The reactivities of oxetanes 1a-1e were examined. Besides CCl4/CH3CN/H2O was added to a solution of β-hydroxyazide
recovered starting material, nitriles were the only isolable 1b in acetonitrile. After 25 min, only 6% of nitrile 2b was
compounds (Table 2). This was true for both diastereomers, isolated. This implies that RuO4 was the main promoter for the
although the two isomers did not react with equal efficiency reactions shown in Table 2.
(entries 3 and 4). A slower rate of reaction for the isomers The unexpected nitrile formation resulting from 2-azido-
where the 2-azido and 3-methyl groups were on the same face 2-(hydroxymethyl)oxetane cleavage with RuO4 led to three
of the oxetane was demonstrated by reaction of a mixture of key questions for us: (1) Is this pathway unique to oxetane
the diastereomers of 1b (entry 2). The anti-diastereomer systems? (2) Which oxidation state--the primary alcohol, the
(with the methyl group on the same side of the ring as the aldehyde or the carboxylic acid--of the R-hydroxymethyl
azide) was consumed more slowly, as evidenced by a change group is the immediate precursor of oxidative cleavage? (3)
in ratio of the diastereomers in the reactants, compared to What is the role of the metal in the reaction?
the recovered starting materials. A more rapid consumption To address the first question, other β-azidoalcohol sys-
of the syn-diastereomer was also observed when a reaction tems were treated with RuO4. First, psico-furanose 919 was
was done with a mixture of 1c. This outcome will be discussed subjected to our optimized RuO4 conditions. After 30 min
further in the quantum calculations section. In all cases, even carboxylic acid 10 was observed, and substantial starting
with shorter reaction times and recovery of some of the material remained (based on 1H and 13C NMR). When more
starting material, the material balance was not accounted RuO4 (0.5 equiv) was added, following the conditions uti-
for. Since nitrile 2b was shown to be stable under the reaction lized by Sano and co-workers,2 only amide 11, derived from
conditions, we postulated that the azidooxetanes were not carboxylic acid 10, was ultimately isolated. The oxidation
themselves stable and that their byproducts may have been with the increased equivalents of RuO4 took ∼1.5 h. Next,
subjected to further oxidative degradation. For β-azidoalco- azidophytosphingosine 1220 was treated with 0.06 equiv of
hols 1d and 1e the expected nitriles were the major products RuO4, and only carboxylic acid 13 was isolated after 3 h
isolated. However, no starting material was recovered, and (Scheme 1). These results suggested that there is something
based on the NMR spectra prior to purification, many unique about the β-azidoalcoholoxetanes 1.
additional products were observed. For substrate 1d this A second issue considered was the oxidation state of the
outcome was attributed to aromatic ring degradation, as cleaved carbon. Ignoring for now the role of the metal,
previously reported by Sharpless and co-workers.17 In addi- plausible cleavage pathways can be drawn from each of the
tion, considering 1e, tertiary carbons can be oxidized by possible oxidation states (Figure 3). Each reaction was care-
RuO4.18 If this occurs at the cyclohexyl tertiary carbon, fully monitored (13C NMR) for aldehyde and/or carboxylic
along with the cleavage giving the nitrile, a vicinal diol acid formation; at no point was either observed. Never-
theless, oxidation to either of these states, followed by rapid
(17) Carlsen, P. H. J.; Katsuki, T.; Martin, V. S.; Sharpless, K. B. J. Org.
Chem. 1981, 46, 3936.
(18) (a) Bakke, J. M.; Bethell, D. Acta Chem. Scand. 1992, 46, 644. (19) Mio, S.; Kumagawa, Y.; Sugai, S. Tetrahedron 1991, 47, 2133.
(b) Tenaglia, A.; Terranova, E.; Waegell, B. J. Org. Chem. 1992, 57, 5523. (20) Dere, R. T.; Zhu, X. Org. Lett. 2008, 10, 4641.

J. Org. Chem. Vol. 75, No. 22, 2010 7567

JOC Article Farber et al.

SCHEME 1. Oxidation of psico-Furanose 9 and Azidophytosphingosine 12 with RuO4

FIGURE 4. Oxidative cleavage mechanism proposed by Ye and

co-workers.

SCHEME 2. Oxidation of Protected Azidophytosphingosine

FIGURE 3. Possible states involved in the oxidative cleavage.

cleavage, could not be ruled out. Consequently, one goal was

to access an aldehyde and acid by alternative pathways to
investigate their behavior under the reaction conditions.
All attempts to synthesize R-azidoaldehydes from azido-
hydroxyoxetanes were unsuccessful. 2-Azido-2-(hydroxy-
methyl)oxetanes 1a and 1b were treated with a variety of species,21 it is logical to assume that, if the proposed mechanism
oxidants, including IBX, Dess-Martin periodinane, Swern con- is correct, the equilibrium for the intramolecular addition favors
ditions, TEMPO/NMO and PCC/NaOAc. In all cases, the the aldehyde with ultimate irreversible conversion to the nitrile
starting material was consumed, but no isolable product driving the reaction to completion over time. We examined one
resulted. Similarly, when 1c was treated with PDC in DMF of the Ye examples, azidophytosphingosine 12, in more detail
no carboxylic acid was observed. Thus, the ability to more (Scheme 2). Compound 12, as well as its corresponding
directly probe the nature of the intermediate that undergoes R-azidoaldehyde 14 (prepared in 99% yield by oxidation of
the cleavage has to this point been precluded because 12 with IBX), required more than two days in the presence of
of the sensitivity of the 2-azido-2-(hydroxymethyl)oxetanes. PCC for complete conversion to nitrile 15. The conversion of 12
Nevertheless, there are a number of observations that suggest to 15 using PCC was slower than that of 14 to 15. Both reactions
that it is the hydroxymethyloxetanes that are cleaved. were checked intermittently by NMR. Over the course of
Our conviction that the azidoalcohol is cleaved is based both reaction of alcohol 12, starting material, aldehyde and the nitrile
on a comparison of our results to those reported for the PCC could be seen for almost the entire time of monitoring. As
mediated cleavage of β-azidoalcohols and on the fact that both expected, the reaction of aldehyde 14 showed starting material
RuO4 and IO4- effect the cleavage. The mechanism of oxidative and nitrile. In neither case was any carboxylic acid observed.
cleavage proposed by Ye and co-workers4 invoked an intramo- These results are consistent with the mechanism proposed by
lecular reaction between the distal nitrogen of the azide and the Ye. However, for 2-azido-2-(hydroxymethyl)oxetane cleavage
carbonyl of the presumed aldehyde intermediate, followed by there are several key differences. The reaction is much more
proton transfer, oxidation, and loss of CO and N2 (Figure 4). rapid, with complete consumption of the starting material
Considering that R-azidoaldehydes are readily isolated, stable occurring in less than an hour. Also, with the oxetanes there is
no hydrogen on the carbon attached to the azide, precluding the
(21) (a) Liu, K. K. C.; Kajimoto, T.; Chen, L.; Zhong, Z.; Ichikawa, Y.; type of rearrangement shown in going from A to B (Figure 4).
Wong, C.-H. J. Org. Chem. 1991, 56, 6280. (b) Sugiyama, M.; Hong, Z.; Thus, the cleavage observed with the 2-azido-2-(hydroxy-
Liang, P.; Dean, S. M.; Whalen, L. J.; Greenberg, W. A.; Wong, C.-H. J. Am.
Chem. Soc. 2007, 129, 14811. (c) Kandula, S. R. V.; Kumar, P. Tetrahedron methyl)oxetanes could not proceed by the same pathway as
Asymmetry 2005, 16, 326. the PCC mediated oxidative cleavage.
7568 J. Org. Chem. Vol. 75, No. 22, 2010
Farber et al.
JOC Article

FIGURE 7. Computed minimum energy configurations of model

FIGURE 5. Proposed mechanism of 2-azido-2-(hydroxymethyl)- complexes of oxetane (16a) and tetrahydrofuran (17a) obtained at
oxetane oxidative cleavage. DFT level.

energy conformers will be essentially unpopulated at room

temperature (kBT ≈ 0.5 kcal/mol). The structural differences
between 16a and 17a therefore suggest that the most stable open
conformations of the azidooxetanyl alcohols may allow them to
accommodate the RuO4 (or IO4-) in a manner that leads to
FIGURE 6. Model oxetanes and tetrahydrofurans for quantum cleavage. In contrast, the closed conformation of the tetrahy-
chemical analysis. drofuran precludes this, and standard oxidation of the alcohol
to the carboxylic acid occurs. Conformational influences on
On the basis of observations thus far, we think that the oxidation reaction pathways are known for proteins,5 but reports
RuO4 (or NaIO4) plays an integral role and that the cleavage for simple organic molecules are hard to find. Nevertheless, the
is most likely occurring from the alcohol oxidation state. We complexity of many oxidation pathways and the difficulty of
propose that the azido nitrogen directly bound to C-2 and the reliably oxidizing alcohols with R-oxidation suggest that con-
oxygen from the primary alcohol react with RuO4 forming a formation may play a greater role than has been recognized.
five-membered intermediate (Figure 5). Subsequent loss of We further explored whether these structural differences
N2 occurs with the regeneration of the double bond between persist in the aldehyde and acid oxidation states. Figure 8 shows
oxygen and ruthenium and the loss of formaldehyde. Finally, the computed minimum energy structures for the aldehydes
nitrile formation and oxetane ring-opening result from the (denoted as 16b and 17b) and the acids (denoted as 16c and 17c).
regeneration of RuO4. It is important to note that we have The lowest energy structure for both the oxetanyl alde-
observed no aldehyde or carboxylic acid in any of our hyde 16b and acid 16c is an open conformation. Compound
reactions. Further support for the alcohol being the species 16b also presents a stable closed conformation with energies
cleaved comes from the outcome with NaIO4 (vide supra). 0.8 kcal/mol higher than the open one. In the acid state (16c),
Nitrile formation was observed from the reaction of NaIO4 however, the open conformation is isoenergetic with the
and oxetane 1b. Periodate is not used to oxidize primary closed one. For the aldehyde (17b) oxidation state in the
alcohols without an appropriate catalytic oxidant, while it is tetrahydrofuran model, the open conformation is unstable.
widely used for oxidative cleavages. However, even if the In the acid state, it becomes stable but 4.6 kcal/mol higher in
alcohol is the precursor and RuO4 or NaIO4 is integral to the energy than the closed conformation. Thus, we conclude that
cleavage process, a question remains: why do the 2-azido- this fundamental structural difference between the model
2-(hydroxymethyl)oxetanes, but not the corresponding tet- oxetane and tetrahydrofuran (open versus closed) persists up
rahydrofurans, undergo cleavage to nitriles? to the aldehyde oxidation state.
Quantum Chemical Analysis. To answer the question pre- Analysis of the molecular electrostatic potential (MEP)
sented above, a series of density functional theory (DFT) surface reveals a potential explanation for why the oxetane
calculations (see Experimental section for details on the level prefers the open conformation, while the tetrahydrofuran
of theory) for model oxetanes 16 and tetrahydrofurans 17 favors the closed. Figure 9 shows a crucial difference in
(Figure 6) was undertaken. the electrostatic interaction between the azide and the rest
DFT calculations reveal an important structural difference of the molecule in 16a open and 17a open. More precisely,
between 16a and 17a. Oxetane 16a exhibits what we refer to as the unfavorable electrostatic repulsion between the distal
an “open” conformation, characterized by a dihedral angle azide nitrogen and one of the oxygens in the formal group is
between atoms C1-C2-N1-N2 of φ=171°. On the other hand, not present in the lowest energy conformation of 17a.
the minimum energy structure of 17a has a “closed” conforma- To test whether the formal group is entirely responsible for
tion with φ = 70° (Figure 7). The possibility for 16a to present a this conformational difference a tetrahydrofuran 18 in which
stable closed conformation was explored, as was the potential this group was removed from 17a was evaluated. This change
for 17a to present an open conformation. In the case of 16a, gave rise to an open conformation 1.5 kcal/mol higher than the
a closed conformation is stable, but it is 0.7 kcal/mol higher closed one, thus reducing the energy difference by 3.5 kcal/mol.
in energy. For 17a, the open conformation is also stable, but Although this is consistent with the electrostatic argument
5.0 kcal/mol in energy above the closed one. Thus, these higher above, it is apparent that the formal is not entirely responsible
J. Org. Chem. Vol. 75, No. 22, 2010 7569
JOC Article Farber et al.

FIGURE 8. DFT computed minimum energy configurations of model oxetane and tetrahydrofuran in their aldehyde (16b and 17b,
respectively) and acid (16c and 17c, respectively) oxidation states.

In conclusion, an unexpected cleavage of 2-azido-

2-(hydroxymethyl)oxetanes appears to be a result of a con-
formational preference that allows RuO4 or IO4- to interact
with both the alcohol and azido moieties, leading to cleav-
age. The lowest energy conformation for a closely related
tetrahydrofuran blocks similar access of the oxidant, and this
is consistent with the observed conversion of the alcohol to
the corresponding carboxylic acid.

Experimental Section
Typical Procedure for the Synthesis of Methyleneoxetanes
FIGURE 9. Molecular electrostatic potential obtained from the (7). trans-4-(tert-Butyldiphenylsilanyloxymethyl)-3-methyl-2-methy-
DFT electron density. Negative and positive potentials are repre- leneoxetane (7a). Dimethyltitanocene (2.54 mmol, 5.00 mL, 0.50 M
sented by red and blue, respectively. The double arrows mark the
in toluene)11 and trans-4-(tert-butyldiphenylsilanyloxymethyl)-3-
crucial interaction that makes the open conformation in 17a less
stable relative to the same conformation in 16a.
methyloxetan-2-one (6a) (0.36 g, 1.05 mmol) were stirred at 80 °C
under N2 in the dark. After 2 h, TLC (petroleum ether/EtOAc, 98:2)
for the differential preference of the closed versus the open showed the presence of starting material; so more dimethyltitano-
conformation. cene (1.0 mL, 0.50 mmol) was added. After 40 min, TLC indicated
Another question that might be addressed by a quantum reaction completion. The solution was then cooled to rt, and
petroleum ether (10 mL) was added, at which point a yellow
mechanical analysis is the observed difference in reactivity
precipitate formed. The resulting mixture was stirred for 18 h at rt.
between the syn and anti stereoisomers of azidooxetanes The solid residue was filtered through a pad of Celite, rinsing with
1a-c (see Table 2 for how syn and anti are defined). The petroleum ether. The solvent was removed under reduced pressure
anti-isomer (which has the 2-azido and 3-methyl groups on to 5 mL (total volume), and the residue was purified by flash
the same face) gave a lower yield of nitrile (see entries 3 and 4, chromatography on silica gel, packing the column with petroleum
Table 2) and reacted at a slower rate (see discussion of ether/triethylamine (96:4) and eluting with petroleum ether/EtOAc/
Table 2) than the syn. The corresponding energies of the triethylamine (97.5:2.0:0.5) to afford methyleneoxetane 7a as white
open and closed conformations for the anti isomer of 16a, as crystals (196 mg, 55%): mp 52-54 °C; IR (neat): 3072, 2930, 1691
well as the transition state between the two, were computed. cm-1; 1H NMR (300 MHz, CDCl3) δ 7.66 (m, 4H), 7.39 (m, 6H),
For the anti-isomer the open and closed conformations have the 4.42 (ddd, J = 4.3, 4.3, 4.3 Hz, 1H), 4.08 (m, 1H), 3.82 (m, 2H), 3.73
(m, 1H), 3.30 (m, 1H), 1.26 (d, J = 7.1 Hz, 3H), 1.05 (s, 9H); 13C
same energy, and they are separated by a 0.5 kcal/mol barrier.
NMR (75 MHz, CDCl3) δ 168.7, 135.9, 135.7, 129.9, 127.9, 127.9,
Considering that only cleavage was observed, the result suggests 86.6, 78.0, 65.2, 38.2, 27.0, 19.5, 16.6; HRMS (ESI) calcd for
that, for the anti-isomer case, RuO4 promoted cleavage com- C22H28NaO2Si (Mþ þ Na) m/z 375.1751, found 375.1739.
petes with the interconversion between the open and closed Typical Procedure for the Synthesis of 1,5-Dioxaspiro-
forms. Furthermore, the oxidation to the carboxylic acid [3.2]hexanes (8). (2S*,3R*,4S*/R*)-2-(tert-Butyldiphenylsilany-
of azidotetrahydrofuran 9 is somewhat slower than the loxymethyl)-3-methyl-1,5-dioxaspiro[3.2]hexanes (8a). A flask was
cleavage/degradation of the azidooxetanes 1. By transitivity, charged with trans-4-(tert-butyldiphenylsilanyloxymethyl)-3-meth-
this last observation suggests that the back and forth equilibra- yl-2-methyleneoxetane (7a) (0.72 g, 2.0 mmol) in dry CH2Cl2 (6 mL),
tion between the open and closed forms in the oxetane anti- and the resulting solution was cooled to -78 °C (dry ice/acetone
isomers will also compete with the oxidation to the carboxylic bath) under N2. A solution of dimethyldioxirane13 (9.50 mL, 4.08
mmol, 0.43 M in CH2Cl2) was added dropwise. The reaction
acid. That is: although we have argued that oxidation to the
solution was stirred for 1 h at -78 °C under N2. The solvent was
carboxylic acid occurs from the closed form, it is possible that removed under reduced pressure, and the resulting clear oil 8a (3:1
the fact that no carboxylic acid was observed for the anti mixture of diastereomers) was used in the next reaction without
isomers of 1a-c could be due to rapid cleavage depleting the purification: IR (neat): 3071, 3050, 3014, 2998, 2931, 2857, 1589,
open azidoalcohols and conversion between the forms being 1472, 1462, 1428 cm-1; 1H NMR (300 MHz, CDCl3) minor
faster than oxidation to the carboxylic acid. diastereomer: δ 7.75 (m, 4H), 7.43 (m, 6H), 4.24 (ddd, J = 4.3,

7570 J. Org. Chem. Vol. 75, No. 22, 2010

Farber et al.
JOC Article
4.3, 4.3 Hz, 1H), 3.93 (m, 2H), 3.26 (qd, J = 6.8, 6.4 Hz, 1H), 2.89 (d, 13
C NMR (100 MHz, CDCl3) δ 135.7, 135.7, 132.8, 132.7, 130.3,
J = 3.2 Hz, 1H), 2.79 (d, J = 3.2 Hz, 1H), 1.27 (d, J = 7.2 Hz, 3H), 128.2, 120.8, 72.5, 65.4, 29.6, 27.0, 19.5, 14.8; HRMS (ESI)
1.14 (s, 9H); major diastereomer: δ 7.75 (m, 4H), 7.43 (m, 6H), 4.39 calcd for C21H27NNaO2Si (Mþ þ Na) m/z 376.1703, found
(ddd, J = 3.7, 3.7, 3.7 Hz, 1H), 3.93 (m, 2H), 3.45 (qd, J = 7.0, 6.4 376.1728.
Hz, 1H), 3.00 (d, J = 3.1 Hz, 1H), 2.70 (d, J = 3.2 Hz, 1H), 1.27 (d, 1-β-Azido-1-r-carbamoyl-1-dehydro-1-deoxy-5-O-benzoyl-
J = 7.2 Hz, 3H), 1.13 (s, 9H); 13C NMR (75 MHz, CDCl3) minor 2,3-O-isopropylidene-D-ribofuranose (11). 2-Azido-2-deoxy-6-O-
diastereomer: δ 135.8, 135.7, 133.5, 133.3, 129.9, 129.9, 127.9, 127.9, benzyl-2,3-O-isopropylidene-β-D-psicofuranose (9)19 (99 mg,
92.1, 80.6, 65.4, 49.1, 38.3, 26.9, 19.4, 14.1; major diastereomer: δ 0.30 mmol) and sodium periodate (0.33 g, 1.5 mmol) in CH3CN
135.8, 135.7, 133.5, 133.3, 129.9, 129.9, 127.8, 127.9, 91.5, 83.2, 65.1, (1.6 mL), CCl4 (1.6 mL) and H2O (2.4 mL) were stirred vigorously
51.1, 36.5, 26.9, 19.4, 12.7; HRMS (ESI) calcd for C22H28O3SiNa in the presence of RuCl3 3 3H2O (32 mg, 0.15 mmol) at rt for 90 min.
(Mþ þ Na) m/z 391.1705, found 391.1743. Then, the mixture was diluted with CH2Cl2 (10 mL) and H2O
Typical Procedure for the Synthesis of 2-Azido-2-(hydroxy- (5 mL), and the layers were separated. The aqueous layer was
methyl)oxetanes (1). (2R*/S*,3R*,4R*)-2-Azido-4-heptyl-2-(hy- extracted with CH2Cl2 (3  5 mL), the combined organic layers
droxymethyl)-3-methyloxetanes (1b). (2S*,3S*,4S*/R*)-2-Heptyl- were dried (MgSO4), and the solvents were removed in vacuo to
3-methyl-1,5-dioxaspiro[3.2]hexanes (8b) (1.83 mmol) were dissolved give a mixture of carboxylic acids (diastereomers). The mixture
in Et2O (3 mL) at rt under N2. Trimethylsilyl azide (0.36 mL, was then dissolved in dry THF (3 mL) and treated with triethyla-
2.74 mmol) was added dropwise, and the resulting solution was mine (0.09 mL, 0.62 mmol) and ethyl chloroformate (0.08 mL,
stirred overnight at rt. Then, the solvent was removed in vacuo to 0.86 mmol) at 0 °C. After 5 min, NH3 (gas) was bubbled through
afford a yellow oil which was dissolved in THF (9.2 mL), and the the solution for 10 min. Then, H2O (5 mL) was added, and the two
resulting solution was cooled to 0 °C (ice bath). Tetra-n-butylam- layers were separated. The aqueous layer was washed with MTBE
monium fluoride (2.74 mmol, 2.74 mL, 1 M in THF) was added (5  10 mL), and the combined organic layers were dried (MgSO4)
dropwise, and the solution was stirred for 2 h at 0 °C. Then, the and concentrated. The crude product was purified by flash chro-
solvent was removed under reduced pressure, and CH2Cl2 (10 mL) matography on silica gel (petroleum ether/EtOAc, 50:50) to afford
was added. The resulting solution was washed with H2O (10 mL) amide 11 (50 mg, 50%) as a white solid: [R]23D -47 (c 0.10,
and brine (10 mL), dried (MgSO4) and the solvent removed in vacuo. CH2Cl2); mp 170-171 °C; IR (mineral oil) 3446, 3162, 2850, 2118,
Purification by flash chromatography on silica gel (petroleum 1721, 1657, 1459 cm-1; 1H NMR (500 MHz, CDCl3) δ 8.07 (d, J =
ether/EtOAc, 90:10) provided β-azidoalcohols 1b as a clear oil 7.2 Hz, 2H), 7.57 (t, J = 7.4 Hz, 1H), 7.45 (t, J = 7.8 Hz, 2H), 6.61
(diastereomeric ratio 4:1) (132 mg, 30% over 3 steps). The two (bs, 1H), 5.81 (bs, 1H), 4.86 (dd, J = 5.7, 1.2 Hz, 1H), 4.76 (d, J =
diastereomers were separated by careful chromatography: Charac- 5.7 Hz, 1H), 4.72 (ddd, J = 6.5, 6.5, 1.0 Hz, 1H), 4.53 (dd, J =
terization of (2S*,3R*,4R*)-2-azido-4-heptyl-2-(hydroxymethyl)-3- 11.8, 6.5 Hz, 1H), 4.47 (dd, J = 11.8, 6.7 Hz, 1H), 1.49 (s, 3H), 1.32
methyloxetane (minor diastereomer): IR (neat) 3447, 2930, 2114, (s, 3H); 13C NMR (125 MHz, CDCl3) δ 167.2, 166.3, 133.7, 130.0,
1458, 1379, 1251 cm-1; 1H NMR (300 MHz, CDCl3) δ 4.06 (ddd, 129.6, 128.8, 114.2, 100.9, 86.3, 86.0, 82.1, 64.1, 26.6, 24.9; HRMS
J = 6.7, 6.7, 6.7 Hz, 1H), 3.62 (m, 2H), 2.73 (dq, J = 7.0, 7.0 Hz, (ESI) calcd for C16H19N4O6 (Mþ þ H) m/z 363.1299, found
1H), 2.17 (m, 1H), 1.79-1.63 (m, 2H), 1.25-1.21 (m, 13H), 0.86 (t, 363.1296.
J = 6.9 Hz, 3H); 13C NMR (75 MHz, CDCl3) δ 97.1, 82.2, 64.0, (2R,3S,4R)-2-Azido-3,4-O-isopropylidene-1-octadecanoic acid
44.4, 36.8, 31.9, 29.5, 29.3, 24.4, 22.8, 14.2, 12.1; HRMS (ESI) calcd (13). Sodium periodate (0.11 g, 0.52 mmol) was added to a flask
for C12H23N3NaO2 (Mþ þ Na) m/z 264.1682, found 264.1671. charged with CCl4/CH3CN/H2O (1:1:1, 0.8 mL), and the resulting
Characterization of (2R*,3R*,4R*)-2-azido-4-heptyl-2-(hydroxy- mixture was stirred at 0 °C (ice bath). RuCl3 3 3H2O (1.60 mg, 0.0078
methyl)-3-methyloxetane (major diastereomer): IR (neat) 3432, mmol) was added, and the biphasic orange mixture was vigorously
2924, 2857, 2115, 1456.9, 1260 cm-1; 1H NMR (300 MHz, CDCl3) stirred for 1 h. Then, the supernatant was added to (2S,3S,4R)-2-
δ 4.35 (ddd, J = 6.8, 6.8, 6.8 Hz, 1H), 3.52 (dd, J = 12.4, 4.8 Hz, azido-3,4-O-isopropylidene-1-octadecanol (12) (50 mg, 0.13 mmol)
1H), 3.40 (dd, J = 12.4, 8.5 Hz, 1H), 2.82 (dq, J = 7.1, 7.1 Hz, 1H), in CH3CN (0.3 mL) at rt, followed by the addition of more sodium
2.25 (dd, J = 8.4, 5.2 Hz, 1H), 1.72-1.62 (m, 2H), 1.25 (m, 10H), periodate (56 mg, 0.26 mmol). The resulting mixture was vigorously
1.14 (d, J = 7.1 Hz, 3H), 0.85 (t, J = 6.9 Hz, 3H); 13C NMR (75 stirred for 3 h at rt and then diluted with CH2Cl2 (5 mL). The two
MHz, CDCl3) δ 99.0, 85.8, 66.1, 41.1, 36.3, 31.9, 29.6, 29.4, 24.6, layers were separated, and the aqueous layer was extracted with
22.8, 14.2, 13.0; HRMS (ESI) calcd for C12H23N3NaO2 (Mþ þ Na) CH2Cl2 (3  5 mL). The combined organic layers were dried
m/z 264.1682, found 264.1696. (MgSO4) and concentrated. Purification by flash chromatography
Typical Procedure for the Synthesis of Nitriles (2). (2S*,3S*)- on silica gel (petroleum ether/EtOAc, 80:20) yielded carboxylic acid
4-(tert-Butyldiphenylsilanyloxy)-3-hydroxy-2-methylbutanenitrile 13 as a white solid (30 mg, 60%): [R]23D -8.0 (c 0.12, CH2Cl2); mp
(2a). A flask was charged with CCl4:CH3CN:H2O (1:1:1, 0.9 mL) 69-71 °C; IR (neat) 3162, 2919, 2850, 2103, 1702; 1H NMR (400
and NaIO4 (133 mg, 0.62 mmol). The resulting mixture was cooled MHz, CDCl3) δ 4.23 (m, 2H), 3.84 (d, J = 8.4 Hz, 1H), 1.62 (m,
to 0 °C (ice bath), and RuCl3 3 3H2O (2.0 mg, 0.009 mmol) was 3H), 1.45 (s, 3H), 1.34 (s, 3H), 1.24 (m, 24 H), 0.86 (t, J = 7.0 Hz,
added at once. The mixture was stirred for 1 h at 0 °C. Then, the 3H); 13C NMR (100 MHz, CDCl3) δ 173.6, 109.3, 77.9, 76.3, 61.1,
supernatant was added to (2R*,3R*,4S*)-2-azido-4-(tert-butyldi- 32.2, 29.9, 29.9, 29.9, 29.8, 29.7, 29.6, 28.9, 28.0, 27.0, 25.7, 22.9, 14.3;
phenylsilanyloxymethyl)-2-(hydroxymethyl)-3-methyloxetane (1a) HRMS (ESI) calcd for C21H39N3NaO4 (Mþ þ Na) m/z 420.2833,
(64 mg, 0.16 mmol) in CH3CN (0.30 mL), and the mixture was found 420.2815.
stirred at rt for 25 min. It was then diluted with CH2Cl2 (5 mL) and (2R,3S,4R)-2-Azido-3,4-O-isopropylidene-1-octadecanal (14).
H2O (3 mL), and the layers were separated. The aqueous layer was 2-Iodoxybenzoic acid (0.95 g, 3.4 mmol) was added to
extracted with CH2Cl2 (3  5 mL), the combined organic extracts (2S,3S,4R)-2-azido-3,4-O-isopropylidene-1-octadecanol (12)
were dried (MgSO4), and the solvents removed under reduced (0.43 g, 1.1 mmol) in EtOAc (17 mL) at rt. The reaction mixture
pressure. Purification by flash chromatography on silica gel was refluxed at 90 °C for 3.5 h. Then, it was cooled to rt and
(petroleum ether/EtOAc, 85:15) provided recovered starting material, filtered through a pad of Celite, rinsing with EtOAc. The filtrate
(2R*,3R*,4S*)-2-azido-4-(tert-butyldiphenylsilanyloxymethyl)- was concentrated in vacuo to afford aldehyde 14 (0.43 g, 99%)
2-(hydroxymethyl)-3-methyloxetane (1a) (20 mg, 33%), and nitrile as a pale yellow oil which solidified overnight: [R]23D -8.3 (c
2a as a clear oil (10 mg, 20%): IR (neat) 3462 (br), 3072, 2930, 2857, 0.10, CH2Cl2); mp 52-54 °C; IR (neat) 2987, 2925, 2114, 1735,
2117, 1472, 1428 cm-1; 1H NMR (400 MHz, CDCl3) δ 7.63 (m, 1468, 1372 cm-1; 1H NMR (400 MHz, CDCl3) δ 9.71 (s, 1H),
4H), 7.41 (m, 6H), 3.70 (m, 3H), 2.84 (dq, J = 7.2, 4.2 Hz, 1H), 4.19 (m, 2H), 3.90 (d, J = 7.6 Hz, 1H), 1.58 (m, 3H), 1.44 (s, 3H),
2.49 (d, J = 4.6 Hz, 1H), 1.30 (d, J = 7.1 Hz, 3H), 1.06 (s, 9H); 1.32 (s, 3H), 1.24 (m, 23H), 0.85 (t, J = 6.9 Hz, 3H); 13C NMR

J. Org. Chem. Vol. 75, No. 22, 2010 7571

JOC Article Farber et al.

(100 MHz, CDCl3) δ 197.1, 109.4, 77.9, 76.8, 66.4, 32.1, 29.9, be unaltered by such correction, we reported energies without
29.9, 29.8, 29.8, 29.7, 29.6, 29.4, 27.8, 26.9, 25.4, 22.9, 14.3, 1.21; the inclusion of zero point energy and thermal effects. Transi-
HRMS (ESI) calcd for C21H39N3O3 (Mþ) m/z 381.2986, found tion state calculations were carried out with the program Jaguar
381.2956. using the quadratic synchronous transit (QST) method. Anal-
Computational Method and Theory Level. Full geometry ysis of the Molecular Electrostatic Potential was performed with
optimization at the DFT level were carried out using the hybrid the quantum chemical software Jaguar,23 at the same theory
functional B3LYP with basis set 6-31 g(d,p) using the program level specified above. Image rendering in Figures 7, 8, and 9 was
Gaussian 09.22 All energies reported were obtained in vacuum. performed with the program Maestro.24
On benchmark calculations of 16a and 17a, in which we
computed the relative energy between the open and closed Acknowledgment. Dr. Martha Morton is acknowledged
conformation, we found that inclusion of zero point energy for assistance with NMR experiments. Professors Nicholas
effects and thermal corrections only changes the energy differ- Leadbeater and Mark Peczuh are acknowledged for helpful
ences by a tenth of a kcal/mol. Since all of our conclusions would
discussion. This manuscript is based upon work partially
(22) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
supported by the National Science Foundation (NSF) under
M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, Grant No. CHE-0111522. J.A.G. acknowledges financial
G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; support from the Camille and Henry Dreyfus foundation
Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.;
Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; and an NSF Career award (CHE-0847340).
Nakai, H.; Vreven, T.; Montgomery, Jr., J. A.; Peralta, J. E.; Ogliaro, F.;
Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.;
Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Supporting Information Available: General experimental
Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, N. J.; Klene, M.; methods and procedures, spectroscopy data and 1H and 13C
Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; NMR spectra for all new compounds. This material is available
Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.;
Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth,
free of charge via the Internet at http://pubs.acs.org.
G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas,
€ Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,
O.; (23) Jaguar version 7.5; Schrodinger LLC: New York, NY, 2008.
Revision A.1; Gaussian, Inc.: Wallingford, CT, 2009. (24) Maestro, version 9.0; Schrodinger LLC: New York, NY, 2010.

7572 J. Org. Chem. Vol. 75, No. 22, 2010