Catalytic regioselective sulfonylation of α-chelatable alcohols: Scope and mechanistic insight

Article abstract of DOI:10.1021/ja016031r

This paper describes a convenient protocol for the regioselective sulfonylation of α-chelatable alcohols. Typically, the reaction of α-heterosubstituted alcohols with 1 equiv of p-TsCl and 1 equiv of Et₃N in the presence of 2 mol % of Bu₂SnO leads to rapid, regioselective, and exclusive monotosylation. The pK_a of the amine was correlated to the reaction rate. A plausible mechanism for this reaction has been proposed on the basis of ¹¹⁹Sn NMR studies.

Full text of DOI:10.1021/ja016031r

Published on Web 03/15/2002
Catalytic Regioselective Sulfonylation of r-Chelatable
Alcohols: Scope and Mechanistic Insight
Michael J. Martinelli,* Rajappa Vaidyanathan,*,† Joseph M. Pawlak,
Naresh K. Nayyar, Ulhas P. Dhokte, Christopher W. Doecke, Lisa M. H. Zollars,
Eric D. Moher, Vien Van Khau, and Berta Kosˇmrlj
Contribution from Chemical Process R&D, Lilly Research Laboratories, Lilly Corporate Center,
Eli Lilly and Company, Indianapolis, Indiana 46285-4813
Received April 17, 2001. Revised Manuscript Received January 12, 2002
Abstract: This paper describes a convenient protocol for the regioselective sulfonylation of R-chelatable
alcohols. Typically, the reaction of R-heterosubstituted alcohols with 1 equiv of p-TsCl and 1 equiv of Et3N
in the presence of 2 mol % of Bu2SnO leads to rapid, regioselective, and exclusive monotosylation. The
pKa of the amine was correlated to the reaction rate. A plausible mechanism for this reaction has been
proposed on the basis of 119Sn NMR studies.
Introduction
observed. However, the unavoidable coproduction of a stoi-
chiometric amount of lipophilic tin waste, usually separable only
The regioselective monofunctionalization of hydroxy groups
in polyol substrates has been of considerable interest to organic
chemists.1 Several reports describing the use of stoichiometric
reagents to effect sulfonylation,2 alkylation,3 and acylation4 have
appeared in the literature. Since Shanzer and co-workers first
reported it,5 the monoderivatization of diols via their stannylene
acetals has been explored and reviewed thoroughly.6 Typically,
the reaction of a 1,2-diol 1 with Bu2SnX, where X ) O2 or
(OMe)2,7 is driven to completion by removal (azeotropic or
desiccant) of water or methanol to afford the required stannylene
acetal 2 (eq 1). After solvent exchange, these acetals undergo
selective alkylation, acylation, sulfonylation, and phosphoryl-
ation, usually at the primary position, or silylation with variable
regioselectivity.8,9 This derivatization step is generally done in
the presence of a catalytic amount of base (typically a tertiary
amine such as triethylamine). The stannylene acetal protocol
accomplishes primary hydroxyl activation and temporary sec-
ondary hydroxyl protection in a single operation.
by chromatography, poses a significant problem in large-scale
applications. Dibutyltin oxide has been employed in a catalytic
fashion to effect macrolactonization under neutral conditions,
presumably as a template for ionic interactions with the
carboxylate and alcohol termini.10 Herrado`n and co-workers
reported the use of catalytic Bu2SnO under microwave-assisted
conditions to accelerate benzoylation of polyols.11 We recently
disclosed a convenient procedure for the primary selective
sulfonylation of glycols using catalytic dibutyltin oxide in the
presence of stoichiometric base.12 Subsequently, dimethyltin
dichloride was reported as a catalyst for the selective mono-
benzoylation of diols, with powdered K2CO3 as adjuvant.13 An
asymmetric version of this reaction using a chiral dialkyltin
dibromide has also been reported.14 More recently, Curran and
co-workers have described the use of a recoverable fluorous
tin oxide to effect regioselective monosulfonylation of diols.15
(5) Shanzer, A. Tetrahedron Lett. 1980, 21, 221.
(6) (a) David, S.; Hanessian, S. Tetrahedron 1985, 41, 643. (b) Grindley, T.
B. AdV. Carbohydr. Chem. Biochem. 1998, 53, 17.
(7) Boons, G.-J.; Castle, G. H.; Clase, J. A.; Grice, P.; Ley, S. V.; Pinel, C.
Synlett 1993, 913.
(8) Leigh, D. A.; Martin, R. P.; Smart, J. P.; Truscello, A. M. J. Chem. Soc.,
Chem. Commun. 1994, 1373.
(9) Reginato, G.; Ricci, A.; Roelens, S.; Scapecchi, S. J. Org. Chem. 1990,
55, 5132.
(10) Steliou, K.; Nowosielska, A. S.; Favre, A.; Poupart, M. A.; Hanessian, S.
J. Am. Chem. Soc. 1980, 102, 7579.
(11) (a) Morcuende, A.; Valverde, S.; Herrado´n, B. Synlett 1994, 89. (b)
Herrado´n, B.; Morcuende, A.; Valverde, S. Synlett 1995, 455. (c) Mor-
cuende, A.; Ors, M.; Valverde, S.; Herrado´n, B. J. Org. Chem. 1996, 61,
5264.
This stannylene-based approach has been widely employed
primarily because of the high yields and regioselectivities
* To whom correspondence should be addressed. E-mail: mjm@lilly.com.
† Present address: Chemical Process R&D Division, Pharmacia Corp.,
1500-91-201, 7000 Portage Road, Kalamazoo, MI 49001.
(1) Kolb, H. C.; Van Nieuwenhze, M. S.; Sharpless, K. B. Chem. ReV. 1994,
94, 2483.
(2) O’Donnell, C. J.; Burke, S. D. J. Org. Chem. 1998, 63, 8614.
(3) Bouzide, A.; Sauve´, G. Tetrahedron Lett. 1997, 38, 5945.
(4) Sekine, M.; Kume, A.; Hata, T. 3617; Ishihara, K.; Kurihara, H.; Yamamoto,
H. J. Org. Chem. 1993, 58, 3791.
(12) (a) Martinelli, M. J.; Moher, E. D. U.S. Patent 9809942, 1998; Chem. Abstr.
1998, 175904. (b) Martinelli, M. J.; Nayyar, N. K.; Moher, E. D.; Dhokte,
U. P.; Pawlak, J. P.; Vaidyanathan, R. Org. Lett. 1999, 1, 447. (c) Martinelli,
M. J.; Vaidyanathan, R.; Khau, V. V. Tetrahedron Lett. 2000, 41, 3773.
(13) Maki, T.; Iwasaki, F.; Matsumura, Y. Tetrahedron Lett. 1998, 39, 5601.
(14) (a) Iwasaki, F.; Maki, T.; Nakashima, W.; Onomura, O.; Matsumura, Y.
Org. Lett. 1999, 1, 969. (b) Iwasaki, F.; Maki, T.; Onomura, O.; Nakashima,
W.; Matsumura, Y. J. Org. Chem. 2000, 65, 996.
(15) Bucher, B.; Curran, D. P. Tetrahedron Lett. 2000, 41, 9617.
9
3578
J. AM. CHEM. SOC. 2002, 124, 3578-3585
10.1021/ja016031r CCC: $22.00 © 2002 American Chemical Society

Sulfonylation of R-Chelatable Alcohols
A R T I C L E S
Table 1. Tosylation of Diol 3 under Various Conditions
tin free
standard
catalytic
diol 3 (equiv)
TsCl (equiv)
Et3N (equiv)
Bu2SnO (equiv)
time (min)
% 5
1.0
1.0
1.0
0.1
1.0
1.0
1.0
1.0
1.0
0
1.0
0.02
15
<1%
1080
>10%
1080
<1%
We report herein the full details of our investigation into the
regioselective tosylation of R-chelatable alcohols with catalytic
Bu2SnO.
Results and Discussion
During the course of our efforts to synthesize analogues of
cryptophycin,16 we discovered the ability of Bu2SnO to function
as a catalyst in the selective monosulfonylation of diols. Mono-
(tosylate) 4 (eq 2) was a key intermediate in our synthetic route.
The results of our preliminary experiments are listed in Table
1. Attempts to access 4 by treating diol 3 with p-TsCl and Et3N
(“tin-free” conditions; Table 1) led to the undesired production
of the bis(tosylate) 5 (∼10%) in addition to the desired mono-
(tosylate) 4. Diol 3 was converted to the corresponding
stannylene acetal by treatment with Bu2SnO in toluene with
azeotropic removal of water. After solvent exchange into
CH2Cl2, the stannylene was treated with p-TsCl (1 equiv) and
Et3N (0.1 equiv) to furnish mono(tosylate) 4 as the exclusive
product (“standard” conditions; Table 1).17 The “standard”
protocol employs 0-10 mol % Et3N, presumably since the weak
product-tin complex remains intact until workup. While the
stannylene is a tight covalent complex and quite stable, its
stability should be significantly diminished upon primary alcohol
functionalization. We therefore speculated that excess Et3N
might enhance turnover through competitive tin binding to
displace the product with concomitant neutralization of the
newly formed HCl. Thus, treatment of diol 3 with p-TsCl and
Et3N (1 equiv each) and catalytic Bu2SnO (2 mol %) led to the
results under the “catalytic” column. It is noteworthy that
excellent regioselectivity (comparable to the “standard” protocol)
is achieved in this case. More significant is the observed rate
acceleration under these conditions, compared with the “tin-
free” protocol.
Figure 1. Rates of monotosylation of 6.
Figure 2. Variation of rate of reaction with tin species used: (1)Bu2SnO;
(2) Bu2SnCl2; (3) Bu2Sn(OMe)2; (4) Bu2Sn(OAc)2; (5) Bu3SnCl.
SnO or without added Bu2SnO. The conversion to mono-
(tosylate) 7 was followed by 1H NMR spectroscopy as a function
of time (Figure 1). Analysis of the data demonstrates that the
Bu2SnO-catalyzed reaction goes to completion substantially
faster than the uncatalyzed version. Apparently, in the uncata-
lyzed version, as the reaction time is extended, competitive
secondary sulfonylation becomes significant.
Encouraged by these initial results, we sought to explore the
scope and mechanistic aspects of this reaction. We examined
the rate of the reaction and product distribution with respect to
variation of several parameters, such as the Sn species used as
catalyst, solvent, stoichiometry and nature of base added, and
substrate structure.
Nature of Tin Species. Different tin reagents were evaluated
as catalysts in the tosylation reaction. In a typical procedure, a
0.5 M solution of diol 6 in CD2Cl2 was treated with 1 equiv of
p-TsCl, 0.02 equiv of the tin reagent, and 1.0 equiv of Et3N,
In an effort to clarify the catalytic effect of dibutyltin oxide
on the tosylation reaction, a rate study was conducted. Diol 6
was now chosen as the test substrate. In separate experiments,
1-phenyl-1,2-ethanediol (6) was treated with p-TsCl (1.05 equiv)
and Et3N (1 equiv) in CD2Cl2, in the presence of catalytic Bu2-
1
and the progress of the reaction was monitored by H NMR
spectroscopy. The results are summarized in Figure 2. Interest-
ingly, Bu2SnO, Bu2Sn(OMe)2, Bu2SnCl2, and Bu2Sn(OAc)2
exhibited essentially similar catalytic activity. Bu3SnCl, how-
ever, proved to be a much less efficient catalyst. We believe
that this trend reflects the ability of the tin reagent to readily
form a strong complex with the diol (stannylene acetal). Bu2-
(16) Trimurtulu, G.; Ohtani, I.; Patterson, G. M.; Moore, R. E.; Corbett, T. H.;
Valeriote, F. A.; Demchik, L. J. Am. Chem. Soc. 1994, 116, 4729.
(17) Treatment of stannylene acetal 8 with 1 equiv of p-TsCl in the absence of
Et3N led to clean conversion to tosylate 7 within 15 min. The time (1080
min) under the “standard” conditions (Table 1) refers to the time required
for stannylene acetal formation and tosylation.
9
J. AM. CHEM. SOC. VOL. 124, NO. 14, 2002 3579

A R T I C L E S
Martinelli et al.
Figure 3. Variation of rate of reaction with solvent used: (1) acetonitrile;
(2) methylene chloride; (3) tetrahydrofuran; (4) toluene; (5) methanol.
Figure 4. Variation of rate of reaction with base used. Numbers in
parentheses refer to the pKa of the conjugate acid of the base used: (1)
Et3N (10.72); (2) i-Pr2NEt; (3) PMP (11.25); (4) NMM (7.13); (5) K2CO3;
(6) pyridine (5.17).
SnO, Bu2Sn(OMe)2, Bu2SnCl2, and Bu2Sn(OAc)2 are all capable
of forming stannylene acetals upon treatment with diols. Once
the stannylene acetal is formed, the nature of the “nonalkyl”
ligand on tin (in the catalyst) becomes irrelevant. Thus, not
surprisingly, the tosylation reactions conducted in the presence
of these catalysts were comparable in rate, going to completion
within ca. 15 min. Bu3SnCl, on the other hand, is incapable of
forming stannylene acetals with diols. This could account for
the fact that the tosylation reaction conducted in the presence
of Bu3SnCl was much slower. We chose to use Bu2SnO for
our future reactions primarily because it was a well-behaved
solid, which would facilitate handling.
Figure 5.
morpholine (pKa 7.13) and proton sponge (pKa 12.0) were
considerably slower, while pyridine (pKa 5.17) did not progress
beyond 40% conversion. The reactions using bidentate bases
such as DABCO (1,4-diazabicyclo[2.2.2]octane) and TMEDA
(N,N,N′,N′-tetramethylethylenediamine) went to completion
within 15 min, similar to Et3N, but 1,1,3,3-tetramethylguanidine
(pKa 13.6) predominantly afforded N-tosylation. The reaction
with proton sponge intermediary in effectiveness could be due
to the insoluble, heterogeneous reaction mixture.
Solvent Study. The catalytic tosylation reaction was carried
out in a variety of solvents in our quest for optimal conditions.
In a typical procedure, a 0.5 M solution of diol 6 in the
appropriate solvent was treated with 1 equiv of p-TsCl, 0.02
equiv of Bu2SnO, and 1.0 equiv of Et3N, and the progress of
The stoichiometry of the base was then considered. The
stannylene acetal of 1-phenyl-1,2-ethanediol (8) (Figure 5) was
synthesized and subsequently treated with p-TsCl in CH2Cl2
and varying amounts of Et3N (0.1-1.0 equiv). In all cases, clean
and efficient primary regioselective tosylation was observed,
irrespective of the stoichiometry of Et3N. Treatment of stan-
nylene acetal 8 with 1 equiv of p-TsCl in the absence of Et3N
led to clean conversion to tosylate 7 within 15 min. This suggests
that, in the stoichiometric tin protocol, exclusive primary
monotosylation can be obtained even in the absence of any
nucleophile or base; however, for the catalytic tin protocol, a
full 1 equiv of base is required to regenerate the tin catalyst.
The same net reaction in the catalytic protocol was then
investigated. Diol 6 was dissolved in CD2Cl2 and treated with
p-TsCl (1.05 equiv), Bu2SnO (0.02 equiv), and varying amounts
of Et3N. We discovered that, under these conditions, the percent
conversion to the tosylate was equal to the added equivalents
of base. In other words, with 0.5 equiv of base, the reaction
stopped at 50% conversion. It could be driven to completion
by simply adding the balance of Et3N. Thus, it is evident that,
in the catalytic reaction, a full 1 equiv of the base is required,
presumably to remove HCl from the cycle. From these
preliminary studies, we surmised that the optimal conditions
for the tosylation reaction would be catalytic Bu2SnO and
stoichiometric triethylamine in CH2Cl2 as the solvent.
1
the reaction was monitored by H NMR spectroscopy. The
results are summarized in Figure 3. The reactions in acetonitrile
and methylene chloride were essentially complete within ca.
20 min, while in THF the reactions were slower; the reaction
in toluene was significantly slower, presumably due to the
limited solubility of the diol in toluene. The use of MeOH as
the solvent proved deleterious, probably as a result of either
competitive binding of MeOH at the tin center and/or competi-
tive tosylation of MeOH. However, it is noteworthy that
tosylation of 6 went to 20% conversion even in such a large
excess of methanol.
Nature and Stoichiometry of Base. To better understand
the role of the amine base in the tosylation reaction, we
undertook a systematic investigation. In these experiments, a
solution of diol 6 in CD2Cl2 was treated with 1 equiv of p-TsCl,
0.02 equiv of Bu2SnO, and 1.1 equiv of the appropriate base,
and the reaction progress was monitored by 1H NMR spectros-
copy. The results are depicted in Figure 4. The reaction rate
appears to depend on the pKa of the conjugate acid of the base
used. There was virtually no difference in rate between Et3N,
i-Pr2NEt,18 suspended K2CO3, and 1,2,2,6,6-pentamethylpip-
eridine, bases with pKa above 10. Reactions with N-methyl-
(18) We had earlier incorrectly observed that addition of i-Pr2NEt proved
deleterious to the reaction rate and efficiency.12b
9
3580 J. AM. CHEM. SOC. VOL. 124, NO. 14, 2002

Sulfonylation of R-Chelatable Alcohols
A R T I C L E S
Table 2. Substrate Effect in the Monotosylation Reaction
Catalyzed by Bu2SnO
a For a typical procedure, see Experimental Section. b Using 0.02 equiv
of Bu2SnO. c Refers to the isolated yield of the primary mono(tosylate).
Substrate Study. Our next endeavor was to examine the
range of substrates to which the catalytic reaction may be
applicable. Table 2 shows a series of substrates subjected to
the Bu2SnO-catalyzed tosylation reaction. The yields refer to
the amount of monotosylated product isolated, with the balance
being a mixture of the starting material and the bis(tosylate). It
is evident that substrates in entries 1-3 are capable of forming
five-membered chelates with Bu2SnO. In each of these cases,
dramatic rate acceleration was observed in the presence of
catalytic Bu2SnO, concomitant with excellent regioselectivity.
The rate differences between the diol (entry 1) and alkoxy
alcohol (entry 3) substrates may be a reflection of haptophilicity.
The diol in entry 4 is predisposed to form a six-membered
stannylene acetal with Bu2SnO. The catalyzed reaction in this
case was virtually indistinguishable from the uncatalyzed version
in terms of reaction rate.19 The same was true in the case of
phenethyl alcohol, where bidentate chelation onto tin was not
possible (entry 5). From these observations, is clear that an
R-chelating moiety is crucial for efficient catalysis of the
tosylation reaction by Bu2SnO.
Bu2SnO within 45 min to afford the desired mono(tosylate) in
97% yield. The corresponding uncatalyzed reaction went to
approximately 13% conversion in 4 h (entry 1). A similar rate
enhancement was observed in the case of cis-cyclohexane-1,2-
diol (entry 3). The reactions of the corresponding trans-diols
were much slower under the same conditions, and less signifi-
cant rate acceleration was observed with added Bu2SnO (entries
2 and 4). These results are consistent with the intermediacy of
a five-membered ring stannylene acetal, in accordance with the
stoichiometric precedent.20 The formation of such a species
would be geometrically more facile in the case of cis-1,2-diols
than trans-1,2-diols.
We also examined more diverse primary alcohols with
R-chelatable substituents. Alcohol 13 was sulfonylated more
rapidly in the presence of catalytic Bu2SnO to afford tosylate
14 (98% in 6 h; eq 3). The corresponding uncatalyzed reaction
was much slower (95% in 23 h). Glucofuranose 15 also
furnished the expected primary tosylate 16 (Eq 4) under these
conditions in 74% yield within 2 h (18% yield for the
uncatalyzed reaction under identical conditions). Hexane-1,2,6-
triol (17) underwent tosylation predominantly at the 1-position
to afford a 9:1 mixture of mono- and bis-primary tosylates 18
and 19, respectively, within 2 h in 73% yield (eq 5). This
example again serves to highlight the importance of the
R-chelatable substituent in the catalytic tosylation reaction.
Cyclic 1,2-Diols. In the case of cyclic 1,2-diols, significant
rate differences were observed between the catalyzed and
uncatalyzed versions (Table 3). For example, cis-cyclopentane-
1,2-diol was efficiently tosylated in the presence of catalytic
Interestingly, the tosylation of hydrobenzoin 24 led to the
formation of trans-stilbene oxide (26) (ca. 5%) in addition to
Table 3. Tosylation of Cyclic 1,2-Diols Catalyzed by Bu2SnOa
(19) However, the regioselectivity was slightly better in the catalyzed version.
This suggests that the incipient stannylene acetal may be competing with
the non-Sn-mediated background process.
a All reactions were performed in CH2Cl2. See Experimental Section for
details.
9
J. AM. CHEM. SOC. VOL. 124, NO. 14, 2002 3581

A R T I C L E S
Martinelli et al.
Scheme 1
Scheme 2
the desired mono(tosylate) 25, which was isolated in 80% yield
in 40 min (eq 6).21 The corresponding uncatalyzed reaction was
much slower in this case also (<5% conversion in 2 h).
Sulfonylation of diethyl tartrate (27) under the Bu2SnO-mediated
conditions led to a mixture of mono(tosylate) 28 (65%) and
bis(tosylate) 29 (8%), the latter arising presumably due to the
coordination of tin to the ester oxygen (eq 7).
(32 and 31, respectively) were obtained in 70% overall yield
(77:23). This is in contrast to the results of Tsuda and co-workers
who observed a preponderance of the 4-O-tosylate 31 when they
treated the stannylene acetal derived from R-methyl-D-xylose
with p-TsCl and DMAP (cat) in dioxane (Scheme 1).23
This trend may be explained as follows: We have already
shown that an R-chelatable moiety should be present for efficient
catalysis of tosylations by Bu2SnO.12a For cyclic compounds,
we have demonstrated that a cis-R-chelatable moiety is slightly
favored over the trans congener (Table 3). The trans-1,3-
cyclohexanediol reacts about 3× slower than the cis-1,2-diol,
with significantly lower yield. In the case of the R-anomer 30,
one may envision a five-membered intermediate arising out of
tin coordination to the methoxy and 2-hydroxy groups (38,
Scheme 2), thereby leading to tosylation at the 2-position to
afford 32. The 4-O-tosylate 31 may be formed either via the
six-membered stannylene acetal intermediate 39 or via an
uncatalyzed reaction. Since five-membered rings are kinetically
favored over six-membered rings, one would expect 38 to be
formed faster than 39, leading to the observed product distribu-
tion. On the other hand, the product distribution observed by
Tsuda and co-workers may be attributed to the greater contribu-
tion of intermediate 39 (preformed stannylene acetal), where
the 4-O-Sn bond is the most reactive to a bulky electrophile
for steric reasons. The fact that cis-chelation is more effective
than trans-chelation is well established.20 The most spectacular
example is the efficient reaction at O-3′ of methyl â-D-lactoside
stannylene acetal with allyl bromide in the presence of six other
hydroxyl moieties. The evidence in Table 2 suggests that
Carbohydrate Examples. Stannylene acetals have been
extensively employed as intermediates in the regioselective
derivatization of carbohydrates.6b We sought to expand the scope
of our catalytic protocol to carbohydrate examples (R- and
â-methyl-D-xylose, Scheme 1). If the hypothesis of a five-
membered intermediate were correct, one would expect the
R-anomer 30 to undergo preferential tosylation at the 2-position
to afford 32. However, when the Bu2SnO-catalyzed tosylation
of R-methyl-D-xylose 30 was run in CH2Cl2, the conversion
was low (25%), and the predominant product was the 2,4-bis-
(O-tosylate) (33). Similar results were obtained in the absence
of Bu2SnO. We reasoned that this was due to the higher
solubility of the mono(tosylate)s 31 and 32 (compared to the
starting material) in CH2Cl2. However, when the reaction was
carried out in dioxane,22 the 2-O- and 4-O-tosylated products
(20) Alais, J.; Maranduba, A.; Veyrie`res, A. Tetrahedron Lett. 1983, 24, 2383.
(21) Mono(tosylate) 25 is prone to decomposition and has to be isolated
immediately after the reaction. The compound does not survive mass
spectral conditions.
(23) Tsuda, Y.; Nishimura, M.; Kobayashi, T.; Sato, Y.; Kanemitsu, K. Chem.
Pharm. Bull. 1991, 39, 2883.
(22) The Bu2SnO-catalyzed tosylations are slower in dioxane than in CH2Cl2.
9
3582 J. AM. CHEM. SOC. VOL. 124, NO. 14, 2002

Sulfonylation of R-Chelatable Alcohols
A R T I C L E S
chelation to a cis-R-methoxy moiety is possible (entry 3) and
could be in competition with that to a trans-OH (Table 3). When
the stannylene acetal is formed in advance, the equilibrium is
pushed to trans-diol complexation by removal of water. It should
be recognized that alternative explanations for formation of
2-substituted products are possible, but the complexity of the
system makes it difficult to rule out alternatives.
The â-anomer 34, on the other hand, underwent tosylation
preferentially at the 4-position to furnish 35, consonant with
Tsuda’s results (Scheme 1); however, the conversion was low.
In this case, the trans complexation between the less hindered
hydroxy groups would be more favorable, and hence the 4-O-
tosylate (35) would be formed. Moreover, the formation of the
six-membered stannylene acetal 40 via tin coordination to the
4- and 2-hydroxy groups may not be favored (1,3-diaxial
interaction), leading to the observed lower conversion.
Figure 6. Dimeric structures of stannylene acetals.
monomeric subunits in these dimers, there is one apical
dicoordinate oxygen and one equatorial tricoordinate oxygen.
The apical oxygen is reportedly more reactive than the tri-
coordinate equatorial oxygen. Although steric factors may dictate
which dimer is more populated, the 1,2-dimer 42 or the 1,1-
dimer 43 reacts to give the observed primary tosylate. Therefore,
to simplify our discussion, we will consider only dimer 43
henceforth.
It was next important to consider whether Et3N is a ligand
for tin or simply a base to neutralize the formed HCl. This would
provide critical insight to whether Et3N becomes relevant pre-
tosylation or post-tosylation, to design catalytic systems that
could employ chiral bases. Substitution of Et3N with Bu4NCl
in the catalytic protocol did not lead to any tosylation, suggesting
that the primary role of Et3N is as a base. The utilization of
suspended K2CO3 performed similarly to Et3N although slower,
and “proton sponge” (1,8-bis(dimethylamino)naphthalene) was
only slightly less effective. These data further support the notion
that the base is likely not a ligand for tin.
However, in the stoichiometric version, it has also been
suggested that the amine (or nucleophile) helps break up the
dimer.6b Grindley et al have observed significant secondary
tosylation in the absence of added nucleophiles (in the stoichio-
metric tin protocol), especially when the group on the secondary
carbon was bulky and the two alkyl groups on tin were large
or rigid.24b This reversal in regioselectivity in the absence of
added nucleophiles prompted the authors to implicate a mon-
omeric stannylene acetal species complexed with the nucleophile
(in the presence of added nucleophiles) as the reactive species.
The addition of excess Bu4NI with stoichiometric pyridine,
under the standard conditions, did not provide any beneficial
effect relative to the reaction without added iodide.
Mechanistic Studies. Having developed the catalytic tosy-
lation protocol, we deemed it appropriate to investigate the
mechanism of this reaction. It has to be mentioned that the
products of these reactions do not rearrange, unlike acylations
and silylations.24 One may also safely assume that the origin of
selectivity is the same for both the stoichiometric and catalytic
tosylation reactions. The selectivity observed in the sulfonylation
of stannylene acetals (stoichiometric tin protocol) has been
ascribed to the existence of these compounds as dimers (41-
43, Figure 6).6 However, our first approach was to determine
which species, the monomer or dimer, dominated in solution.
We chose 1-phenyl-1,2-ethanediol (6) as our test substrate.
Our attempts to identify reaction intermediates by NMR
spectroscopy (1H and 119Sn) were hampered by the limited
solubility of Bu2SnO in CH2Cl2. If indeed the catalytic tosylation
proceeds via the stannylene acetal intermediate (vide supra),
we reasoned that the reaction could be catalyzed by the
stannylene acetal 8. Indeed, the sulfonylation of 6 using 1.05
equiv of p-TsCl, 1.1 equiv of Et3N, and 2 mol % of 8 was
essentially complete within ca. 8 min. Thus, acetal 8 was used
instead of Bu2SnO for our NMR studies. The 119Sn NMR
spectrum of acetal 8 in CD2Cl2 (0.2 M) at room temperature
shows a broad peak around -130 ppm, similar to what was
observed by Roelens,25 indicating the presence of a penta-
coordinate tin species. In an effort to mimic the reaction
conditions, we acquired the 119Sn NMR spectrum of a mixture
of diol 6 containing 5 mol % of acetal 8.26 The presence of a
broad peak centered around -125 ppm suggests that the
stannylene acetal still exists as a dimer under the reaction
conditions in the catalytic tin protocol. Three rapidly equilibrat-
ing dimeric structures are possible, namely, the 2,2-dimer (41),
the 1,2-dimer (42), and the 1,1-dimer (43).27 In each of the
We also considered the fact that the catalytic reaction might
proceed through such an intermediate arising out of precom-
plexation of the stannylene acetal dimer with Et3N to afford a
monomeric species, which would then undergo tosylation.
Addition of stoichiometric Et3N to stannylene acetal 8 did not
1
alter either the H or the 119Sn NMR spectrum. Significant
changes have been reported at lower temperatures,6b and this
only shows that if an amine intermediate is formed, it is less
stable than the dimer by >2 kcal/mol, which is significantly
less than the activation barrier. These data do not support
precomplexation of the stannylene acetal with Et3N.28
Moreover, when the catalytic tosylation of 6 was conducted
using 0.02 equiv of Bu2SnO and 0.5 equiv of p-TsCl in the
presence of a chiral, optically pure base such as cinchonidine
or sparteine, no kinetic resolution was observed. However,
modest selectivities could be detected in the presence of a chiral
tin dibromide.29 Therefore, it may be inferred that the amine is
involved in the post-tosylation step. Even if the base-complexed
monomer were the intermediate, the two most electronegative
(24) Roelens, S. J. Org. Chem. 1996, 61, 5257. (b) Kong, X.; Grindley, T. B.
Can. J. Chem. 1994, 72, 2396.
(25) Luchinat, C.; Roelens, S. J. Org. Chem. 1987, 52, 4449.
(26) The spectrum was acquired using 52 mg of diol 6 (0.38 mmol) and 7 mg
of 8 (0.02 mmol) in 0.75 mL of CD2Cl2.
(27) The dimers are named by using the number of the two tricoordinate oxygen
atoms. See: Grindley, T. B.; Thangarasa, R. Can. J. Chem. 1990, 68, 1007.
9
J. AM. CHEM. SOC. VOL. 124, NO. 14, 2002 3583

A R T I C L E S
Martinelli et al.
Figure 7. Proposed mechanism of the Bu2SnO-catalyzed tosylation reaction.
oxygen atoms could not both be apical if they are in a ring.
Therefore, one oxygen would be apical and the other would be
equatorial, as the two butyl groups would be equatorial in the
trigonal bipyramid at tin. The amine could adopt the other apical
position, but the attack would occur on the more reactive apical
oxygen atom, which is remote from the nitrogen substituents
and unable to influence stereoselectivity. Thus, it is difficult to
fully rationalize the chiral amine ligand experimental outcome
without additional studies.
outlined in Figure 7. The reaction of the diol with Bu2SnO
produces the stannylene acetal 8 in situ, which exists as a
dimeric species 43. Sulfonylation by p-TsCl leads to intermedi-
ates 44 and 45. Et3N could break up the dimer to give a species
such as 46. Displacement of the tosylate by another molecule
of the diol would furnish intermediate 47. Triethylamine can
then scavenge HCl to regenerate the stannylene acetal 8 and
complete the catalytic cycle.
This mechanism accounts for the following observations: (1)
The stannylene acetal catalyzes the tosylation just as Bu2SnO
does. (2) The stannylene acetal exists as a dimer under the
reaction conditions. (3) 1 equiv amount of base is required to
ensure complete reaction. (4) The extent and rate of reaction
depend on the pKa of the conjugate acid of the base used. (5)
The acetoxy, alkoxy, oxo, or halogen ligands on tin have little
effect on the reaction rate, although other alkyl ligands have
been shown to play a significant role.24b (6) This catalytic
version may contrast the stoichiometric, in which added
nucleophiles can alter the regioselectivity.24b
Conclusion. We have described a convenient protocol for
the regioselective sulfonylation of R-chelatable alcohols using
catalytic amounts (2 mol %) of Bu2SnO. The reactions are
extremely rapid and highly selective. We varied several reaction
parameters and described the optimal reaction conditions. The
role of the amine has been elucidated, and a plausible mecha-
nism for the Bu2SnO-catalyzed tosylation reaction has been
proposed. The proposed mechanism is based on our observations
and literature precedent. These new insights will provide a good
starting point for the design of novel chiral and resin-bound tin
reagents. This catalytic procedure obviates the need for extensive
chromatographic removal of lipophilic tin oxides. The high
regioselectivity and the dramatic rate enhancement observed
render this reaction particularly attractive for large-scale sul-
fonylations.
Next, the 119Sn NMR spectrum of a mixture of diol 6 in the
presence of 8 (5 mol %) and p-TsCl (1 equiv) was acquired.30
The 119Sn NMR spectrum of this sample contains two equally
intense peaks at -140.97 and -91.97 ppm. By analogy to
literature data,31 these peaks may be attributed to a [Bu2Sn-
•
•
(OR)2 ] and a [Bu2SnCl2 ] fragment, respectively. Thus, one can
postulate structure 44 for this complex (Figure 7).31 The
dibutyltin dichloride may be loosely bound to the dialkoxytin
fragment allowing for rapid intra- and intermolecular exchange
of Bu2SnCl2. The 119Sn NMR spectrum also shows three other
significant peaks at -120.48, -374.91, and 108.37 ppm. The
peak at -120.48 ppm may arise from the stannylene acetal
dimer 43. The peak at -374.91 ppm could be due to the
presence of higher oligomeric species. The upfield peak at
108.37 may be attributed to the presence of monomeric Bu2-
SnCl2.32
On the basis of our observations and literature precedent, a
plausible mechanistic postulate for the catalytic protocol is
(28) Matsumura and co-workers have obtained high primary selectivity in the
presence of a non-nucleophilic base (K2CO3 or Na2CO3) while using both
a sterically “unencumbured” tin species (Me2SnCl2)13 as well as a sterically
demanding binaphthyl tin species14 as the catalyst. In the former case, the
reactions were faster with Et3N than K2CO3, but no differences in
regioselectivity were observed. In the latter case, there was no difference
in regioselectivity in the presence of a coordinating base (Et3N) or a
noncoordinating base (Na2CO3), although there were differences in yields
and enantioselectivities. In other words, there was no reversal in regiose-
lectivity due to either the steric bulk of the alkyl group on tin or the nature
of the base used; however, it must be reiterated that the catalytic reaction
does not proceed in the absence of the base, and hence, the product
distribution in the absence of base cannot be determined! These results,
coupled with the lack of spectroscopic evidence for monomeric tin species
in the presence of Et3N, lead us to believe that the catalytic reaction proceeds
via the dimeric species.
Experimental Section
Materials. All chemicals were purchased from Aldrich Chemical
Co. and were used as received, unless otherwise stated. (R)-1-Phenyl-
1,3-propanediol was purchased from Fluka. The carbohydrate samples,
R- and â-methyl-D-xylosides, were purchased from Pfaltz and Bauer,
Inc., and were used as received.
General Procedure for Kinetics Experiments. To a solution of
alcohol 6 (5 mmol) in the appropriate deuterated solvent (10 mL) were
added the appropriate tin reagent (0.1 mmol), p-TsCl (5.25 mmol), and
the appropriate base (5 mmol). The reaction mixture was stirred at room
(29) Martinelli, M. J.; Vaidyanathan, R. Unpublished results. See also ref 14.
(30) The spectrum was acquired using 52 mg of diol 6 (0.38 mmol), 7 mg of 8
(0.02 mmol), and 72 mg of p-TsCl (0.38 mmol) in 0.75 mL of CD2Cl2.
(31) Bredenkamp, M. W.; Spies, H. S. C.; van der Merwe, M. J. Tetrahedron
Lett. 2000, 41, 547.
(32) The complexity of the 119Sn and 1H NMR spectra of a mixture of acetal 8
and p-TsCl and the fact that there is no obvious general relationship between
δ(119Sn) and other empirical parameters preclude a more rigorous inter-
pretation of the data.
9
3584 J. AM. CHEM. SOC. VOL. 124, NO. 14, 2002

Sulfonylation of R-Chelatable Alcohols
A R T I C L E S
temperature. Aliquots (0.3 mL) were withdrawn at different time
intervals, and the percent conversion was determined by 1H NMR
spectroscopy.
Hz, 2 H), 5.84 (d, J ) 3.5 Hz, 1 H), 4.54 (d, J ) 3.0 Hz, 1 H), 4.27-
4.25 (m, 1 H), 4.16-4.15 (m, 1 H), 4.09-4.03 (m, 2 H), 3.88 (d, J )
3.0 Hz, 1 H), 3.43 (s, 3 H), 2.91 (d, J ) 6.5 Hz, 1 H), 2.43 (s, 3 H),
1.72-1.38 (m, 10 H); 13C NMR (125 MHz, CDCl3) δ 144.9, 132.4,
129.9, 128.0, 112.6, 104.6, 84.2, 80.9, 78.9, 72.3, 67.3, 57.9, 36.3, 35.6,
24.8, 23.8, 23.5, 21.6; HRMS found m/z 446.1838 (M + NH4)+, calcd
for C20H32NO8S m/z 446.1848.
General Procedure for the Sulfonylation of r-Chelatable Alco-
hols. To a solution of the alcohol (10 mmol) in CH2Cl2 (20 mL) were
added Bu2SnO (0.2 mmol), p-TsCl (10 mmol), and Et3N (10 mmol).
The reaction mixture was stirred until TLC indicated complete
consumption of the starting material. The reaction was quenched with
water, and the layers were separated. The aqueous layer was extracted
with CH2Cl2 (2 × 10 mL), the combined organic layers were washed
sequentially with water and brine, dried (Na2SO4), and filtered, and
the filtrate was concentrated in vacuo. The residue was crystallized or
chromatographed to afford the desired mono(tosylate).
cis-Cyclopentane-1,2-diol 1-tosylate (20): IR (CHCl3) 3022, 1743,
1
1728, 1600, 1362, 1176 cm-1; H NMR (500 MHz, CDCl3) δ 7.79
(dd, J ) 7.0 Hz, 2.0 Hz, 2 H), 7.33 (d, J ) 8.0 Hz, 2 H), 4.64 (ddd,
J ) 6.0 Hz, 6.0 Hz, 4.0 Hz, 1 H), 4.10 (ddd, J ) 6.0 Hz, 6.0 Hz, 4.0
Hz, 1 H), 3.05 (br s, 1 H), 2.42 (s, 3 H), 1.84-1.75 (m, 4 H), 1.69-
1.64 (m, 1 H), 1.49-1.45 (m, 1 H); 13C NMR (125 MHz, CDCl3) δ
144.9, 133.4, 129.8, 127.7, 84.1, 72.7, 29.8, 27.7, 21.6, 18.7; HRMS
found m/z 257.0828 (M + H)+, calcd for C12H17O4S m/z 257.0848.
trans-Cyclopentane-1,2-diol 1-tosylate (21): IR (CHCl3) 2969,
1600, 1358, 1189, 1176 cm-1; 1H NMR (500 MHz, CDCl3) δ 7.77 (d,
J ) 8.5 Hz, 2 H), 7.33 (d, J ) 8.5 Hz, 2 H), 4.58-4.55 (m, 1 H),
4.26-4.23 (m, 1 H), 2.89 (br s, 1 H), 2.42 (s, 3 H), 2.01-1.90 (m, 2
General Procedure for the Sulfonylation of r- and â-Methyl-D-
xylosides. To a solution of the xyloside (2.5 mmol) in dioxane (5 mL)
were added Bu2SnO (0.05 mmol), p-TsCl (2.63 mmol), and Et3N (2.75
mmol). The reaction mixture was stirred at room temperature overnight
and concentrated in vacuo. The residue was chromatographed (SiO2,
60% EtOAc/hexanes) to afford the mono(tosylates).
H), 1.70-1.63 (m, 1 H), 1.49-1.45 (m, 3 H), 1.56-1.50 (m, 1 H); 13
C
Sulfonylation Products. Tosylates 7,33 9,34 10,35 11,36 12,37 28,38
31,23 32,23 33,23 35,23 36,23 and 3723 were characterized by 1H and 13
C
NMR (125 MHz, CDCl3) δ 145.2, 133.7, 130.2, 128.0, 89.2, 77.3, 31.7,
30.1, 21.9, 20.7; HRMS found m/z 257.0831 (M + H)+, calcd for
C12H17O4S m/z 257.0848.
NMR and were compared to literature values or commercial samples,
as appropriate.
119Sn NMR Experiments. 119Sn NMR spectra were collected on a
Varian Inova 500 MHz spectrometer at a frequency of 186.52 MHz.
Unless stated, all spectra were accumulated at a probe temperature of
25 °C. Data were collected using the INEPT pulse sequence with a
sweep width of 119.9 kHz using a 64K data set. The pulse width was
approximately 70° with an acquisition time of 0.192 s coupled with a
2 s recycle delay. Data were accumulated using H decoupling and
transformed using a 3 Hz line broadening. Tetramethyltin was used as
the standard.
(R)-1-Phenyl-1,3-propanediol 3-tosylate (11):36 1H NMR (300
MHz, CDCl3) δ 7.79 (d, J ) 8.2 Hz, 2 H), 7.36-7.26 (m, 7 H), 4.81
(m, 1 H), 4.28 (m, 1 H), 4.08 (m, 1 H), 2.45 (s, 3 H), 2.03 (m, 3 H);
13C NMR (75 MHz, CDCl3) δ 145.1, 142.5, 142.0, 132.0, 129.8, 127.8,
121.6, 121.4, 117.1, 117.0, 70.1, 67.1, 64.0, 21.4
2-((Tosyloxy)methyl)-1,4-benzodioxan (14): mp ) 79-81 °C; IR
(CHCl3) 3035, 1597, 1494, 1369, 1269, 1177 cm-1; 1H NMR (300 MHz,
CDCl3) δ 7.79 (d, J ) 8.2 Hz, 2 H), 7.34 (d, J ) 8.0 Hz, 2 H), 6.75
(m, 4 H), 4.41-4.00 (m, 5 H), 2.45 (s, 3 H); 13C NMR (75 MHz,
CDCl3) δ 145.1, 142.5, 142.0, 132.0, 129.8, 127.8, 121.6, 121.4, 117.1,
117.0, 70.1, 67.1, 64.0, 21.4; HRMS found m/z 320.0713 (M+), calcd
for C16H16O5S m/z 320.0718.
cis-Cyclohexane-1,2-diol 1-tosylate (22): IR (CHCl3) 3028, 2946,
2869, 1704, 1600, 1449, 1361, 1175 cm-1; 1H NMR (500 MHz, CDCl3)
δ 7.79 (d, J ) 8.5 Hz, 2 H), 7.32 (d, J ) 8.5 Hz, 2 H), 4.59 (ddd, J )
8.0 Hz, 3.0 Hz, 3.0 Hz, 1 H), 3.80-3.79 (m, 1 H), 2.42 (s, 3 H), 2.20
(br s, 1 H), 1.92-1.85 (m, 1 H), 1.75-1.69 (m, 1 H), 1.62-1.45 (m,
4 H), 1.31-1.23 (m, 2 H); 13C NMR (125 MHz, CDCl3) δ 144.7, 134.0,
129.8, 127.6, 83.1, 68.9, 30.1, 27.6, 21.6, 21.5, 20.7; HRMS found
m/z 271.0987 (M + H)+, calcd for C13H19O4S m/z 271.1004.
trans-Cyclohexane-1,2-diol 1-tosylate (23): IR (CHCl3) 2947, 2868,
1
1
1600, 1454, 1360, 1175 cm-1; H NMR (500 MHz, CDCl3) δ 7.82
(dd, J ) 2.5 Hz, 8.0 Hz, 2 H), 7.34 (dd, J ) 2.5 Hz, 8.0 Hz, 2 H),
4.31-4.25 (m, 1 H), 3.59-3.54 (m, 1 H), 2.44 (s, 3 H), 2.38 (br s, 1
H), 2.04-1.95 (m, 2 H), 1.67-1.66 (m, 2 H), 1.47-1.39 (m, 1 H),
1.32-1.19 (m, 3 H); 13C NMR (125 MHz, CDCl3) δ 144.8, 133.9,
129.9, 127.7, 86.7, 72.0, 32.3, 30.8, 23.8, 23.2, 21.6; HRMS found
m/z 271.1004 (M + H)+, calcd for C13H19O4S m/z 271.1004.
Hydrobenzoin tosylate (25):21 1H NMR (500 MHz, CDCl3) δ 7.45-
7.07 (m, 14 H), 5.45 (d, J ) 3.5 Hz, 1 H), 5.10 (d, J ) 3.5 Hz, 1 H),
2.35 (s, 3 H), 1.98 (br s, 1 H); 13C NMR (125 MHz, CDCl3) δ 144.3,
138.3, 134.3, 133.5, 129.4, 128.7, 128.1, 128.0, 127.8, 127.7, 127.1,
126.9, 86.2, 76.2, 21.5.
1,2-O-Cyclohexylidene-3-O-methyl-r-D-glucofuranose 6-tosylate
(16): IR (CHCl3) 3029, 2941, 2865, 1600, 1450, 1368, 1177 cm-1; 1H
NMR (500 MHz, CDCl3) δ 7.79 (d, J ) 8.0 Hz, 2 H), 7.34 (d, J ) 8.0
Acknowledgment. The authors thank Professors Leo Paquette,
Peter Wipf, William Roush, and Marvin Miller for helpful
discussions and Professor T. Bruce Grindley for suggestions
during the preparation of this manuscript.
(33) Uccello-Barretta, G.; Cuzzola, A.; Balzano, F.; Menicagli, R.; Iuliano, A.;
Salvadori, P. J. Org. Chem. 1997, 62, 827.
(34) Eliel, E. L.; Bai, X.; Ohwa, M. J. Chin. Chem. Soc. 2000, 47, 63.
(35) Valentine, D.; Johnson, K. K.; Priester, W.; Sun, R. C.; Toth, K.; Saucy,
G. J. Org. Chem. 1980, 45, 3698.
(36) Mukaiyama, T.; Tominori, K.; Oriyama, T. Chem. Lett. 1985, 1359.
(37) Commercially available from SALOR. CAS Registry Number: 4455-09-
8.
(38) Crout, D. H. G.; Gaudet, V. H. B.; Hallinan, K. O. J. Chem. Soc., Perkin
Trans. 1 1993, 805.
Supporting Information Available: 119Sn NMR spectra of
stannylene acetal 8 under both reaction conditions (with and
without TsCl) (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
JA016031R
9
J. AM. CHEM. SOC. VOL. 124, NO. 14, 2002 3585