Lewis base catalysis of bromo- and iodolactonization, and cycloetherification

Article abstract of DOI:10.1073/pnas.1005296107

Lewis base catalyzed bromo- and iodolactonization reactions have been developed and the effects of catalyst structure on rate and cyclization selectivity have been systematically explored. The effects of substrate structure on halolactonization reactions

Full text of DOI:10.1073/pnas.1005296107

Lewis base catalysis of bromo- and iodolactonization,
and cycloetherification
1
Scott E. Denmark and Matthew T. Burk
Department of Chemistry, Roger Adams Laboratory, University of Illinois, Urbana, IL 61801
Edited by Eric N. Jacobsen, Harvard University, Cambridge, MA, and approved July 13, 2010 (received for review April 17, 2010)
Lewis base catalyzed bromo- and iodolactonization reactions
have been developed and the effects of catalyst structure on rate
and cyclization selectivity have been systematically explored. The
effects of substrate structure on halolactonization reactions and
the interaction of those effects with the effects of catalyst
structure have been investigated, leading to synthetically useful
improvements in cyclization selectivity. The knowledge acquired
Fig. 1. General scheme of halocyclization reactions.
was applied to the development of Lewis base catalyzed bromo-
and iodocycloetherification reactions. The ability of some of the
surveyed catalysts to influence the cyclization selectivity of
halolactonization reactions demonstrates their presence in the
transition structure of the product-determining cyclization step.
This observation implies that chiral derivatives of these catalysts
have the potential to provide enantioenriched products regardless
of the rates or mechanisms of halonium ion racemization.
properties of chiral Lewis base catalysts from those of readily
available achiral analogs is highly desirable.
Halolactonization and cycloetherification reactions can pro-
duce either of two constitutional isomers, arising from cyclization
in an exo or endo fashion (24). The ratio of the two isomers is
influenced largely by the substrate, the identity of the halogen,
and the choice of reaction conditions. Under conditions wherein
the halonium ion is undergoing rapid exchange, the product-
determining step must also be the stereochemistry-determining
step. Therefore, the ratio of constitutional isomers produced
in the presence of an achiral Lewis base catalyst serves as an in-
dicator of the presence of the catalyst in the stereo-determining
transition structure. This knowledge would allow the search for
an enantioselective catalyst to focus on classes of compound
whose presence in and capacity to influence the relevant transi-
tion state structure has been demonstrated.
In this paper we report a systematic investigation into the
influence of Lewis base catalysts on the rate and constitutional
site selectivity in bromo- and iodo- lactonization reactions.
The changes in the ratio of 6-endo to 5-exo cyclization products
induced by the presence of a wide range of achiral Lewis base
catalysts provides evidence for the presence of the Lewis base
in the site-selectivity-determining step, whereas in situ IR mon-
itoring provides comparative rate data. These findings are then
applied to the development of Lewis base catalysis of bromo-
and iodo-etherification reactions.
halocyclofunctionalization ∣ halogenation
lectrophilic halocyclizations of olefins, in which electrophilic
E
halonium ions are generated from olefins and opened intra-
molecularly by nucleophilic functional groups (Fig. 1), are versa-
tile synthetic transformations with proven applications to the
synthesis of biologically relevant molecules (1–6). The develop-
ment of catalytic enantioselective halocyclization methods is a
topic of increasing interest in synthetic organic chemistry, one
that presents unique challenges and opportunities for various
paradigms of catalysis in addition to the obvious importance
and utility of this transformation. To date, only a few notable suc-
cesses have been reported; these include the use of chiral Ti-salen
complexes in iodoetherification (7) and cinchonidinium phase-
transfer catalysts in iodolactonization (8). Recently, modified
Cinchona alkaloids have been successfully employed in catalytic
enantioselective chlorolactonization (9), as well as a catalytic
enantioselective bromolactonization of 1,3-eneynes via conjugate
opening of achiral bromonium ions (10–15*).
Results
The program to evaluate the feasibility of Lewis base catalysis of
halofunctionalization of isolated double bonds would involve
electrophilic bromine and iodine sources (halosuccinimides) in
conjunction with unsaturated carboxylic acids and alcohols. For
the initial survey of Lewis bases, 5-phenyl-4-pentenoic acid (1a)
was chosen as the test substrate and a standard experimental pro-
cedure was adopted. All reactions were run in dichloromethane
at 0.15 M in substrate with 1.2 equivalent (equiv) of electrophile
and 0.05 equiv of Lewis base. The temperature was adjusted
such that the background (uncatalyzed) reaction was negligible.
A broad selection of Lewis bases was evaluated for kinetic com-
Careful mechanistic studies by Brown and coworkers and from
these laboratories identified a serious obstacle to the develop-
ment of catalytic enantioselective iodination and bromination
methods, namely the propensity of iodonium and bromonium
ions (but not chloronium ions) to undergo degenerate halogen
exchange with olefins (Fig. 2) (16–19). This process racemizes
the halonium ions at rates that can compete with nucleophilic
capture. Chiral Lewis base catalysts have the potential to prevent
this or other racemization processes, if they remain bound to the
halonium ion until the newly created stereocenters are irreversi-
bly set, thus maintaining a chiral environment regardless of
exchange. Lewis base catalysis of halogenation has been reported
in several different contexts (20–23), as have enantioselective
halocyclization reactions promoted by stoichiometric amounts
of chiral Lewis bases (13–15), but the paucity of catalytic, enantio-
selective Lewis base catalyzed halogenations suggests that some
of the aforementioned systems do not preserve the stereochemi-
cal integrity of intermediates as described here. Given the signif-
icant effort required for the preparation of new chiral Lewis
bases, and the structural and functional diversity of the reported
achiral Lewis base catalysts, a systematic means of inferring the
Author contributions: S.E.D. designed research; M.T.B. performed research; and S.E.D.
and M.T.B. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
*
Wacker-type catalytic enantioselective bromination and chlorination have been
reported (11, 12), as have enantioselective iodination reactions using stoichiometric
amounts of chiral ligands (13–15).
1
To whom correspondence should be addressed. E-mail: sdenmark@illinois.edu.
This article contains supporting information online at www.pnas.org/lookup/suppl/
doi:10.1073/pnas.1005296107/-/DCSupplemental.
www.pnas.org/cgi/doi/10.1073/pnas.1005296107
PNAS ∣ November 30, 2010 ∣ vol. 107 ∣ no. 48 ∣ 20655–20660

Table 1. Catalyst survey for Lewis base catalyzed
bromolactonization
t1∕2
(min)*
Reaction
time (min)
Yield
(%)
†
‡
§
Entry
Catalyst
2aa∶3aa
1
2
3
4
5
6
7
8
9
None
>180
>180
>180
>180
25
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
6
>180
<0.5
<0.5
0.5–1
<0.5
<0.5
<0.5
180
180
180
180
60
8
8
8
8
8
13
36
47
15
93
71
82
89
78
87
89
94
8
78
88
84
75
86
64
25∶1
23∶1
51∶1
50∶1
23∶1
7.3∶1
91∶1
75∶1
25∶1
3.4∶1
27∶1
19∶1
N/D
85∶1
8.1∶1
94∶1
38∶1
6.7∶1
400∶1
ðMe NÞ C¼O
2
2
n-Bu P¼O
3
ðMe NÞ P¼O
2
3
Me2S¼O
ðMe2NÞ C¼S
2
Ph3P¼S
n-Bu3P¼S
Cy P¼S
3
10
11
12
13
14
15
16
17
18
19
ðMe NÞ P¼S
2
3
ðCH2Þ S
8
4
Me2S
20
180
8
8
8
5
8
8
ðPhSÞ
2
n-Bu3P¼Se
ðMe NÞ P¼Se
2
3
Fig. 2. The racemization of halonium ions by degenerate exchange; nonde-
generacy induced by the presence of a chiral Lewis base.
ðPhSeÞ
2
n-Bu3P
ðMe2NÞ P
3
petence. The optimal catalyst was then employed in a brief survey
of olefinic substrates with varying double bond geometries and
substituents.
Br2
*
Determined by React-IR monitoring, reactions performed on 0.2 mmol of
a, in 1.5 mL of CH2Cl2
1
†
Time elapsed before quenching
Bromolactonization. The influence of Lewis bases on the rate
and cyclization selectivity in the bromolactonization of 1a with
N-bromosuccinimide (NBS) was evaluated first. Reaction pro-
‡
Determined by integration of 1H NMR signals for H-6 against
1
;2;4;5-C6H2Cl4 internal standard
§
Determined by integration of 1H NMR signals for H-6
gress was measured by the disappearance of the IR band at
67 cm− . The yields and ratios of cyclization products 2aa and
1
9
3
1
aa were assayed by H NMR integration against an internal
3.4∶1–94∶1. Highly reactive catalysts that also lead to increases in
standard.
the already high endo/exo selectivity included Ph P¼S (entry 7),
3
The results from the Lewis base survey are compiled in Table 1
n-Bu P¼S (entry 8), n-Bu P¼Se (entry 14), ðPhSeÞ (entry 16),
3
3
2
and are organized by donor heteroatom. Under the standard re-
action protocol, the mixture was homogenous and the rate of the
background reaction in the absence of catalyst was insignificant
and n-Bu P (entry 17). Of the Lewis base catalysts surveyed, the
3
highest endo selectivity was observed in the presence of Ph P¼S
3
and ðPhSeÞ , but the highest overall endo selectivity was observed
2
(
entry 1). The oxygen donors ðMe NÞ C¼O, n-Bu P¼O, and
in the control experiment with a catalytic quantity of Br (entry
2
2
3
2
ðMe NÞ P¼O (entries 2–4) were only marginally faster than
19). Substantial reductions in the ratio of 2aa to 3aa were
observed in the presence of ðMe NÞ C¼S, ðMe NÞ P¼S,
2
3
the background rate and were still incomplete after 3 h. On
the other hand DMSO (entry 5), which could function as either
an oxygen or sulfur donor, led to complete reaction after only 1 h.
Gratifyingly, very high rate acceleration was observed with many
Lewis bases bearing sulfur, selenium, and phosphorus donor
atoms. The thiono containing bases (thiourea, phosphine sul-
fides, or a thiophosphoramide, entries 6–10) were all indistin-
2
2
2
3
ðMe NÞ P¼Se, and ðMe NÞ P, and (entries 6, 10, 15, 18, resp.).
2
3
2
3
Among the thiono donors an apparent correlation of increasing
steric bulk with decreasing endo/exo selectivity was noted (entries
6–10).
To evaluate the preparative utility of this bromolactonization,
the reaction of test substrate 1a was run on a 1.0 mmol scale
and the cyclization products were isolated in 81% yield and a
90∶1 ratio of isomers favoring 2aa (Fig. 3). Next, unsaturated
acids 1b and 1c, representing the effects of olefin geometry
and conjugation, were examined. Under the optimized reaction
conditions, both 1b and 1c afforded mixtures of constitutional
isomers now weakly favoring exo-cyclization product 3. Interest-
guishably rapid (t1 < 35 sec) and provided good yields of the
∕2
cyclization products. Interestingly, other divalent sulfur donors
behaved quite differently. Whereas tetrahydrothiophene (entry
1
1) was equipotent with the thiono bases, reaction using dimethyl
sulfide (entry 12) was somewhat slower and diphenyl disulfide
entry 13) was completely ineffective. All of the selenium- (en-
(
tries 14–16) and phosphorus- (entries 17 and 18) based donors
were powerful catalysts and within the resolution of this measure-
ment, indistinguishable. In a final control experiment, bromine, a
common impurity in NBS, was shown to be an effective catalyst
ingly, the use of ðMeNÞ P¼S, which previously displayed the low-
3
est endo selectivity (Table 1, entry 6), produced 3ba with high exo
selectivity. In the cyclization of 1c, the exo selectivity increased
with the use of ðMe NÞ P¼S, but only marginally.
2
3
(
entry 19), most likely due to formation of HBr by rapid uncata-
lyzed bromolactonization.
Bromocycloetherification. The preparative utility of the catalytic
bromofunctionalization was next extended to include Lewis base
catalyzed bromocycloetherification. Unsaturated alcohols 4a–c
were chosen to demonstrate the effects of conjugation and alkene
geometry. In situ FTIR monitoring of the cyclization of 4a
under the previously optimized reaction conditions (5 mol% of
In addition to dramatic differences in the rates of the catalyzed
reactions, the steric course of the reaction was also significantly
influenced by the action of the different catalysts. Although sub-
strate 1a is intrinsically biased toward endocyclic closure (entry
1
), the endo/exo selectivity in the catalyzed reactions ranged from
2
0656
∣
www.pnas.org/cgi/doi/10.1073/pnas.1005296107
Denmark and Burk

Table 2. Catalyst survey for Lewis base catalyzed
iodolactonization
T1∕2
(min)*
Reaction
Yield
(%)
‡
Entry
Catalyst
None
time (h)
†
2ab∶3ab
§
1
2
3
4
5
6
7
8
9
>180
>180
>180
>180
>180
4
35
7
8
4
2
3
3
3
3
1
1
1
1
1
1
3
3
1
1
1
2
2
3
3
4
10
14
33
4
91
92
72
93
97
97
45
2
77
95
82
94
53
10
0
9.5∶1
20∶1
16∶1
20∶1
N/D
4.7∶1
5.5∶1
6.1∶1
1.7∶1
2.5∶1
5.8∶1
5.8∶1
N/D
8.5∶1
4.6∶1
10∶1
5.1∶1
1.4∶1
24∶1
N/D
ðMe2NÞ C¼O
2
n-Bu3P¼O
ðMe2NÞ P¼O
3
Me2SO
ðMe NÞ C¼S
2
2
Ph P¼S
3
n-Bu3P¼S
Fig. 3. Scope of Lewis base catalyzed bromolactonization.
Cy P¼S
3
10
11
12
13
ðMe2NÞ P¼S
3
ðCH2Þ S
10
4
Ph P¼S), revealed that only ca. 50% of 4a was consumed before
3
Me S
>180
>180
9
4
15
52
40
>180
>180
2
the reaction stalled. A previous report from these laboratories on
selenocyclization revealed a similar trend for unsaturated acids
and alcohols which could be ameliorated by the addition of a
weak Brønsted acid. Accordingly, bromocycloetherification of
ðPhSÞ
2
14
15
n-Bu P¼Se
3
ðMe2NÞ3P¼Se
16
17
18
19
20
ðPhSeÞ
2
n-Bu3P
4
a was carried out in the presence of 1.0 equiv of AcOH with
ðMe NÞ P
2
3
the intention of producing a level of acidity similar to that present
during bromolactonization. Gratifyingly, this simple remedy
rescued the reactivity of 4a and allowed the reaction to proceed
rapidly to completion in good yield and with good endo selectivity
I
2
TFA
*
Determined by React-IR monitoring, all reactions run on 0.2 mmol of 1a
†
Time elapsed before quenching
Determined by integration of 1H NMR signals for H-6 against PhMe6
(
Fig. 4).
‡
Bromocycloetherification of 4b in the presence of Ph P¼S and
internal standard
Determined by integration of 1H NMR signals for H-6
3
§
AcOH afforded the product of exo cyclization, 6ba in high yield
and with good selectivity. In the hope of obtaining even higher exo
selectivity, this reaction was attempted in the presence of
ðMe NÞ P¼S, however no alteration in the isomer ratio was ob-
progress was monitored by in situ FTIR analysis and the yield and
product ratios were obtained from H NMR spectra.
1
2
3
served. Similarly, the (Z)-alkenol 4c produced 6ca in good yield
and with high exo selectivity.
Unsurprisingly, iodolactonization is much faster than bromo-
lactonization such that suppressing the uncatalyzed reaction of 1a
with NIS in dichloromethane required executing the experiments
at −40 °C (Table 2, entry 1). Consequently, the solubility of NIS
in dichloromethane, which is only modest at room temperature,
was quite low.
Iodolactonization. Iodolactonization is an extensively studied
cyclofunctionalization process and often provides different,
sometimes complementary selectivity compared to bromolactoni-
zation (25), typically producing a greater proportion of 5-exo cy-
clization. Given the successful application of Lewis base catalysis
to bromolactonization and the range of selectivities observed,
catalysis of iodolactonization presented an attractive target. As
before, 1a was selected as a representative substrate and was sub-
jected to the action of N-iodosuccinimide (NIS) in the presence
of a wide variety of Lewis bases (Table 2). As above, the reaction
The results from the survey of Lewis bases are compiled in
Table 2 in the same order as was presented in Table 1. Here
as well, the oxygen-based donors ðMe NÞ C¼O, n-Bu P¼O,
2
2
3
and ðMe NÞ P¼O, (entries 2–4)) provided at best only modest
2
3
rate acceleration. However, the complete lack of reactivity of
DMSO (entry 5) diverged markedly from its behavior in bromo-
lactonization. As was the case in bromolactonization, the highest
rates were observed in the presence of thiono- (ðMe NÞ C¼S,
2
2
n-Bu P¼S, Cy P¼S, ðMe NÞ P¼S, entries 6, 8–10) and seleno-
3
3
2
3
based donors (n-Bu P¼S and ðMe NÞ P¼Se (entries 14, 15),
3
2
3
though Ph P¼S was less effective (entry 7). Similarly, the divalent
3
chalcogen donors tetrahydrothiophene (entry 11) and ðPhSeÞ
2
(
entry 16) also gave rise to high reaction rates. On the other hand,
dimethyl sulfide (entry 12) was ca. 20 times less reactive than tet-
rahydrothiophene, the pnictogens ðMe NÞ P and n-Bu P (entries
2
3
3
1
7, 18) were substantially less reactive, and ðPh SÞ (entries 13)
2
2
was completely ineffective. In contrast to bromolactonization,
the addition of 5 mol% of I2 had a negligible impact on the
rate of iodolactonization, although the endo/exo ratio increased
(
entry 19).
The intrinsic preference for endocyclization of 1a persisted,
but with attenuated magnitude ranging from 1.4∶1–20∶1 ratios
for 5aa∕6aa. In fact, only the oxygen-based donors (entries 2–4)
afforded improved selectivities compared to background reac-
tion. Moreover, substantial reductions in the endo/exo selectivity
Fig. 4. Scope of Lewis base catalyzed bromocycloetherification.
Denmark and Burk
PNAS
∣
November 30, 2010
∣
vol. 107
∣
no. 48
∣
20657

were observed in the presence of Cy P¼S, ðMe NÞ P¼S, and
3
2
3
ðMe NÞ P (entries 9, 10, 18). Diphenyl diselenide had no effect
2
3
on the product ratio (entry 16), and the other Lewis bases tested
reduced the exo selectivity modestly.
These results were both disappointing and also perplexing. In
preliminary experiments, n-Bu P¼S displayed a greatly increased
3
endo selectivity compared to the result shown in entry 8. After
several unsuccessful attempts to reproduce the original, more
promising result, it was hypothesized that some component of
the original reaction mixture had been contaminated in some
way. Among several potential contaminants examined, a trace
amount of TFA, in combination with n-Bu P¼S resulted in the
3
Fig. 6. Stability of isomeric iodolactones to iodolactonization conditions.
almost exclusive formation of 2ab (Fig. 5). To exclude the possi-
bility that this outcome was the result of thermodynamic control
of the product ratio, 2ab and 3ab, the latter produced under es-
tablished conditions of thermodynamic control (26, 27), were
submitted to the reaction conditions as single purified isomers
Discussion
The development of Lewis base catalyzed bromo- and iodolacto-
nization described above has led to insights into the effects of
catalyst structure on the rate and selectivity of cyclization, high-
lighted the effects of substrate structure and the identity of the
halogen on cyclization selectivity, and produced synthetically use-
ful improvements in cyclization selectivity which will be discussed
along with their implications for the development of enantiose-
lective catalysts.
(
Fig. 6). Almost no isomerization was observed, thus demonstrat-
ing that the product ratio was the result of kinetically controlled
selectivity.
Substrates 1b and 1c were again chosen to explore the scope of
this iodolactonization and to demonstrate the effects of olefin
geometry and conjugation. Iodolactonization of 1b in the pre-
sence of ðMe NÞ P¼S afforded 3bb in good yield and exo selec-
2
3
Rates of Iodolactonization. The first readily apparent trend in
the rates of catalyzed iodolactonization is that the donor atom
has a significant influence as follows: R C¼S ≈ R P¼S ≈ R P¼
tivity (Fig. 5). The iodolactonization of 1c was attempted in
the presence of Ph P¼S and Cy P¼S separately. As was seen
3
3
2
3
3
previously, one of the least endo selective catalysts, Cy P¼S, pro-
3
Se > R P > R P¼O, (Table 2). The ordering of the chalcogen
3
3
vided slightly higher exo selectivity, allowing 3cb to be isolated in
derivatives parallels the differences in softness between the iso-
lated atoms (ionization potentials (IP): Se, 9.75 eV; S, 10.36 eV;
P, 10.49 eV; O, 13.61 eV) as well as the differences in ionization
potentials for the (ðMe NÞ P¼X chalcogenides which Bruno
good yield.
4
Iodocycloetherification. To further explore the scope of Lewis
2
3
base catalysis of iodocyclization, the insights gained from the de-
velopment of iodolactonization were applied to the development
of iodocycloetherification. As was done for bromocycloetherifica-
tion, the conjugated (E)-alkenol 4a and the unconjugated (E)-
and (Z)-alkenols 4b and 4c were chosen to represent the effects
of conjugation and olefin geometry. The conditions that were
previously optimized for the iodolactonization of 1a were applied
et. al. have estimated from the charge transfer band of their
I2 complexes (IP ðMe NÞ P¼O − ðMe NÞ P¼S ¼ 1.5 eV;
2
3
2
3
ðMe NÞ P¼S − ðMe NÞ P¼Se ¼ 0.18 eV) (28).
2
3
2
3
The trend in rate with variation in the substituents on the P
(III) and P(V) donors also follows the order ðMe NÞ P¼
2
3
Y > alkyl P¼Y > Ph P¼Y. This order is the same as that of their
3
3
enthalpies of I complexation (29).
2
to the iodocycloetherification of 4a (5 mol% of n-Bu P ¼ S,
3
5
mol% TFA, −45 °C), and afforded the product of endo cycliza-
Rates of Bromolactonization. The most readily apparent trend
in the rates of catalyzed bromolactonization parallels that
of iodolactonization, namely: ðMe NÞ C¼S ≈ R P¼Se ≈ R P¼
tion, 5ab, in a good yield and high selectivity (Fig. 7). Aside from
extending the reaction time, no further changes to the reaction
conditions were required in this case. The aliphatic alkenols
2
2
3
3
S ≈ R P ≫ R P¼O, with any rate difference between the sulfur
3
3
and selenium homologs and their parent phosphines obscured by
the rates of the data acquisition (Table 1). Although this
ordering is similar to the trend that was observed in iodolactoni-
zation, there are discernable differences in the relative reactiv-
ities of some Lewis bases. Diphenyldiselenide is slightly slower
than the phosphine sulfide and selenide donors, where as in
4
b and 4c were both cyclized in good yield and high 5-exo selec-
tivity. No measurable effect on the endo/exo ratio was observed
for the more exo selective catalyst ðMe NÞ P¼S.
2
3
Fig. 5. Scope of Lewis base catalyzed iodolactonization.
Fig. 7. Scope of Lewis base catalyzed iodocycloetherification.
2
0658
∣
www.pnas.org/cgi/doi/10.1073/pnas.1005296107
Denmark and Burk

iodolactonization it was faster than Ph PS. In iodolactonization
kinetic control, the same consideration applies to TFA as applies
to Lewis bases, namely that to affect the endo/exo ratio the agent
must be present in the product-determining transition state struc-
ture. The acidity of TFA naturally leads to the hypothesis that
this is due to proton transfer from TFA, perhaps by causing a
change in the counter ion, protonation state, or the timing of
proton transfer. This hypothesis is reinforced by the extremely
high endo selectivity observed in the bromolactonization of
3
the reactivity of ðMe NÞ P¼O is greater than n-Bu P¼O and
2
3
3
ðMe NÞ C¼O while in bromolactonization it is the least active
2
2
of the three (Table 1, entries 16–18; Table 2, entries 15–17).
Substrate Effects on Selectivity. The preference for 5-exo cyclization
over 6-endo cyclization observed with substrates 1c and 4b–c,
which lack a conjugating substituent on the alkene is typical
for a variety of halocyclization reactions (2–4). Substrates 1a
and 4a favor 6-endo cyclization as a result of the greater stability
of positive charge localized on benzylic carbons. The greater pro-
portion of 6-endo cyclization observed in bromolactonization and
bromocycloetherification compared to their iodine counterparts
has ample precedent and appears to be a general phenomenon,
although a satisfactory explanation has yet to be proposed (25).
The superposition of these factors frequently leads to modest
selectivity in bromolactonizations, such as was observed in the
case of 1c.
1a in the presence of Br , which likely produces HBr through
2
uncatalyzed bromolactonization. Two plausible mechanisms of
halolactonization can be envisioned that differ by the rate and
timing of proton transfer (Fig. 8), one in which the nucleophile
is deprotonated prior to (or during) irreversible cyclization (Fig. 8,
path A), and one in which the nucleophile is deprotonated after
cyclization (path B). The former path should be favored by the
presence of base, such as the succinimidate anion generated
during the reaction, while the latter should be favored under
conditions with added TFA. Proton transfer from any of the
intermediates depicted in Fig. 8 to succinimidate anion (pKa
of succinimide 9.6 (31); pKa of TFA 0.26; pKa of acetic acid
Alkene geometry can have a strong influence on site selectivity
as well; the preference of 1b for 5-exo cyclization despite the
directing effect of the phenyl group has been observed previously
4
.76 (32); pKBH MeOAc-3.9 (33) is thermodynamically favor-
þ
(
30), and a less dramatic effect was seen in bromocycloetherifi-
able, whereas transfer to trifluoroacetate is only favorable from
the protonated lactones.
cation of 4c.
The most critical conclusion from all of these experiments is
that those catalysts that influenced the endo/exo ratio in halolac-
tonization must have been present in the product-determining
transition structure. This conclusion implies that chiral analogs
of these catalysts have the potential to catalyze enantioselective
halocyclization reactions. This conclusion does not mean that
catalysts that did not measurably influence the endo/exo ratio
must not have been present in the product-determining transition
structure, simply that the experiments described in this paper
do not provide evidence that they were. Similarly, the absence
of an observable catalyst effect on the endo/exo ratios of halocy-
cloetherification does not prove that the catalysts are not present
in this product-determining transition structure either. It is pos-
sible that the conditions or substrates chosen were simply not
conducive to observing such an effect.
Catalyst Effects on Selectivity. The least endo selective catalysts
for 1a and most exo selective catalysts for 1b and 1c share two
important structural features: (i) they possess two symmetry
unique points of branching, either dimethylamino groups or
cyclohexyl groups, and (ii) donate a nonbonding pair from an ele-
ment softer than oxygen. Tetramethylthiourea and ðMe NÞ P, as
2
3
well as the latter's sulfide and selenide, are exo selective in both
iodo- and bromolactonization. In contrast, Cy P¼S, which lacks
3
electron donating dimethylamino groups, is less electron rich but
still very sterically demanding, promotes exo iodolactonization
but not exo bromolactonization of 1b and 1c. Such exceptional
behavior of Cy P¼S may be a result of the differences in reaction
3
conditions rather than differences in the halogen. The iodolacto-
nizations were conducted at lower temperatures, which should
entropically favor ligand association and under conditions of
low NIS solubility, which should limit competitive binding of
the catalyst to free NIS. These factors should allow somewhat
weaker binding ligands to have noticeable effects on the outcome
of the reaction.
Conclusions
A systematic examination of the affect of a wide range of Lewis
bases on the rate and constitutional site selectivity in halofunc-
tionalization reactions has been conducted. Cyclization reactions
of unsaturated acids and alcohols with NBS and NIS are drama-
tically accelerated by Lewis bases that contain sulfur, selenium,
and phosphorus donor atoms. The mode of cyclization (exo vs.
endo) is primarily controlled by the structure of the substrate
such that conjugated (E)-alkenes undergo highly endo selective
cyclization whereas conjugated (Z)-alkene and alkenes of either
geometry bearing aliphatic substituents undergo exo selective cy-
clization. In both cases the cyclization diastereoselectivities are
perfect. Lewis bases significantly influence the constitutional site
selectivity in all halolactonizations indicating that the Lewis base
must be present in the stereochemistry-determining transition
structure. This observation implies that chiral derivatives of these
catalysts have the potential to provide enantioenriched products
Comparing the catalyst effects on bromo- and iodo- lactoniza-
tion of 1c, which lacks a conjugating substituent, illuminates the
roles played by the steric and electronic properties of the Lewis
bases. In the bromolactonization of 1c the magnitude of the cat-
alyst effect on cyclization selectivity is small (Fig. 3) although the
reduction in endo/exo ratio using ðMe NÞ P¼S is consistent with
2
3
the direction of the effect observed in the bromolactonization of
1
a (Table 1). Similarly, the endo/exo ratio observed in the iodo-
lactonization of 1c (Fig. 5) was only modestly reduced by the use
of Cy P¼S as the catalyst, although the greater preference for
3
exo cyclization is in the same direction as that observed in the
iodolactonization of 1a. In contrast the endo/exo ratios in the bro-
molactonizations of 1a–b, which possess conjugating substituents,
were both greatly reduced by the use of ðMe NÞ P¼S compared
2
3
to Ph P¼S (Table 1, Fig. 3) despite the inversion of overall selec-
3
tivity. This selectivity suggests that the effect of ðMe NÞ P¼S on
2
3
the bromolactonizations of 1a–b predominately arises from mod-
ulation of the directing ability of the phenyl group, possibly by
reducing the positive charge localized on the electrophilic
carbons.
The stability of the isomers 2ab and 3ab to the reaction con-
ditions under which 2ab was formed (n-Bu P¼S∕TFA, 1∶1; 5 mol
3
%
, NIS, −40 °C, Fig. 7) demonstrates that the inclusion of a
cocatalytic amount of TFA does not result in thermodynamic con-
Fig. 8. Proposed role of timing of proton transfer on the mechanism of
halolactonization.
trol of the endo/exo ratio. Therefore, because the ratio is under
Denmark and Burk
PNAS
∣
November 30, 2010
∣
vol. 107
∣
no. 48
∣
20659

regardless of the rates or mechanisms of halonium ion race-
mization.
in CH2Cl2 (2.0 mL) by cannula, followed by a solution of Ph3P¼S in CH2Cl2
(
(
0.5 mL, 0.05 equiv, 0.05 mmol). The solution was stirred for 2 h, quenched
butyl vinyl ether in EtOH (2.5 mL, 1.2 M)), sat. aq. Na2S2O3 solution (5 mL),
Materials and Methods
Bromolactonization of 1a. Preparation of rel-(5R,6S)-5-Bromotetrahydro-6-
phenyl-2H-pyran-2-one (2aa) (34). To a solution of NBS (213 mg, 1.2 mmol,
sat. aq. NaHCO3 solution (5 mL)), diluted with H2O, and extracted with EtOAc.
The combined organic extracts were dried over MgSO4, filtered, concen-
trated in vacuo, and purified by column chromatography (silica gel, hex-
ane/EtOAc, 80∶20) to provide 183.4 mg (61%) of 2ab as a white solid.
Melting point 68–76 °C (decomposes).
1
.2 equiv) in CH2Cl2 (5.0 mL) under Ar in the dark was added a solution
of 1a (176.1 mg, 1.0 mmol, 1.0 equiv) in CH2Cl2 (2.0 mL) by cannula, followed
by a solution of Ph3P¼S in CH2Cl2 (0.5 mL, 0.05 equiv, 0.05 mmol). The solu-
tion was stirred at 23 C for 5 min, quenched (sat. (saturated) aq. Na2S2O3
solution (5 mL)), diluted with H2O, and extracted with CH2Cl2. The combined
organic extracts were dried over MgSO4, filtered, concentrated in vacuo, and
purified by column chromatography (silica gel, CH2Cl2) to provide 207 mg
Iodocycloetherification of 4a. Preparation of rel-(2R,3S)-3-Iodotetrahydro-2-
phenyl-2H-pyran (5ab). To a −45 °C suspension of NIS (270 mg, 1.2 mmol,
1.2 equiv) in CH2Cl2 (5.0 mL) under Ar in the dark was added a solution
of 4a (162.2 mg, 1.0 mmol, 1.0 equiv) and TFA (3.8 μL, 0.05 mmol, 0.05 equiv)
in CH2Cl2 (2.0 mL) by cannula, followed by a solution of n-Bu3P¼S in CH2Cl2
(
81%) of 2aa as a white solid. Melting point (mp) 104–106 ° C.
(
0.5 mL, 0.05 equiv, 0.05 mmol). The solution was stirred for 2 h, quenched
Bromocycloetherification of 4a. Preparation of rel-(2R,3S)-3-Bromotetrahydro-
-phenyl-2H-pyran (5aa) (35). To a solution of NBS (213 mg, 1.2 mmol, 1.2 equiv)
in CH2Cl2 (5.0 mL) under Ar in the dark was added a solution of 4a (162 mg,
.0 mmol, 1.0 equiv) and AcOH (57 μL, 1.0 mmol, 1.0 equiv) in CH2Cl2 (2.0 mL)
by cannula, followed by a solution of Ph3P¼S in CH3Cl2 (0.5 mL, 0.05 equiv,
.05 mmol). The solution was stirred at 23 C for 5 min, quenched (sat. aq.
Na2S2O3 solution (5 mL), sat. aq. NaHCO3 solution (5 mL)), diluted with
H2O, and extracted with EtOAc. The combined organic extracts were dried
over MgSO4, filtered, concentrated in vacuo, and purified by column chroma-
tography (silica gel, hexane/EtOAc, 95∶5) to provide 165.5 mg (69%) of 5aa as
colorless needles. Melting point 41–42 °C.
(butyl vinyl ether in EtOH (2.5 mL, 1.2 M), sat. aq. Na2S2O3 solution (5 mL), sat.
aq. NaHCO3 solution (5 mL)), diluted with H2O, and extracted with EtOAc.
The combined organic extracts were washed with sat. NaHCO3 solution, dried
over MgSO4, filtered, concentrated in vacuo, and purified twice by column
chromatography (silica gel, hexane/EtOAc, 95∶5) to provide 215.7 mg
(75%) of 5ab as a colorless oil. TLC∶Rf 0.19 (hexanes/EtOAc, 19∶1) [UV].
2
1
0
Supporting Information Available. Procedures for the preparation, character-
ization, and cyclofunctionalization of all substrates along with full character-
ization of halofunctionalization products see SI Appendix.
ACKNOWLEDGMENTS. Peter Yao is thanked for preliminary experiments
Iodolactonization of 1a. Preparation of rel-(5R,6S)-5-iodotetrahydro-6-phenyl-
and helpful discussions. We are grateful to the National Science Foundation
2
H-pyran-2-one (2ab) (8). To a −45 °C suspension of NIS (270 mg, 1.2 mmol,
.2 equiv) in CH2Cl2 (5.0 mL) under Ar in the dark was added a solution
of 1a (176.1 mg, 1.0 mmol, 1.0 equiv) and TFA (3.8 μL, 0.05 mmol, 0.05 equiv)
(
NSF CHE-0717989) and National Institutes of Health (GM R01-085235) for
generous financial support. M.T.B. thanks Abbott Laboratories for a Gradu-
ate Fellowship in Synthetic Organic Chemistry.
1
1
. Ranganathan S, Muraleedharan KM, Vaish NK, Jayaraman N (2004) Halo- and seleno-
lactonization: the two major strategies for cyclofunctionalization. Tetrahedron
18. Brown RS (1997) Investigation of the early steps in electrophilic bromination
through the study of the reaction with sterically encumbered olefins. Acc Chem
Res 30:131–137.
6
0:5273–5308.
2
3
4
5
6
. Laya MS, Banerjee AK, Cabrera EV (2009) Iodolactonization: past and present
examples. Curr Org Chem 13:720–730.
. Cardillo G, Orena M (1990) Sterocontrolled cyclofunctionalizations of double bonds
through heterocyclic intermediates. Tetrahedron 46:3321–3408.
. Dowle MD, Davies DI (1979) Synthesis and synthetic utility of halolactones. Chem Soc
Rev 8:171–197.
. French AN, Bissmire S, Wirth T (2004) Iodine electrophiles in stereoselective reactions:
recent developments and synthetic applications. Chem Soc Rev 33:354–362.
. Montana AM, Batalla C, Barcia JA (2009) Intramolecular haloetherification and trans-
annular hydroxycyclization of alkenes. A synthetic methodology to obtain polycyclic
ethers and amines. Curr Org Chem 13:919–938.
. Kwon HY, Park CM, Lee SB, Youn J-H, Kang SH (2008) Asymmetric iodocyclization
catalyzed by salen-CrIIICl: its synthetic application to swainsonine. Chemistry—A Eur-
opean Journal 14:1023–1028.
. Wang M, et al. (2004) Enantioselective iodolactonization catalyzed by chiral quatern-
ary ammonium salts derived from cinchonidine. J Org Chem 69:2874–2876.
. Whitehead DC, Yousefi R, Jaganathan A, Borhan B (2010) An organocatalytic asym-
metric chlorolactonization. J Am Chem Soc 132:3298–3300.
19. Denmark SE, Burk MT, Hoover AJ (2010) On the absolute configurational stability of
bromonium and chloronium ions. J Am Chem Soc 132:1232–1233.
20. Mellegaard-Waetzig SR, Wang C, Tunge JA (2006) Selenium-catalyzed oxidative
halogenation. Tetrahedron 62:7191–7198.
21. Ahmad SM, Braddock DC, Cansell G, Hermitage SA (2007) Dimethylformamide,
dimethylacetamide, and tetramethylguanidine as nucleophilic organocatalysts for
the transfer of electrophilic bromine from N-bromosuccinimide to alkenes. Tetrahe-
dron Lett 48:915–918.
22. Castellote I, et al. (2007) Reaction of imines with N-iodosuccinimide (NIS): unexpected
formation of stable 1∶1 complexes. Chem Commun 43:1281–1283.
23. Zhang W, Xu H, Xu H, Tang W (2009) DABCO-catalyzed 1,4-bromolactonization of
conjugated enynes: highly stereoselective formation of a stereogenic center and
an axially chiral allene. J Am Chem Soc 131:3832–3833.
24. Bartlett PA (1984) Olefin cyclization processes that form carbon-heteroatom bonds.
Asymmetric Synthesis, ed JD Morrison (Academic, FL), 3, pp 411–454.
25. Snider BB, Johnston MI (1985) Regioselectivity of the halolactonization of γ,δ-unsatu-
rated acids. Tetrahedron 26:5497–5500.
7
8
9
26. Bartlett PA, Myerson J (1978) Stereoselective epoxidation of acyclic olefinic carboxylic
acids via iodolactonization. J Am Chem Soc 100:3950–3952.
1
1
1
1
1
0. Zhang W, et al. (2010) Enantioselective bromolactonization of conjugated (Z)-enynes.
J Am Chem Soc 132:3664–3665.
1. El-Qisairi A, Hamed O, Henry PM (1998) A new palladium(II)-catalyzed asymmetric
chlorohydrin synthesis. J Org Chem 63:2790–2791.
2. El-Qisairi AK, et al. (2003) New palladium(II)-catalyzed asymmetric 1,2-dibromo
synthesis. Org Lett 5:439–441.
3. Garnier JM, Robin S, Rousseau G (2007) An approach to enantioselective 5-endo
halo-lactonization reactions. Eur J Org Chem 2007:3281–3291.
4. Haas J, Bissmire S, Wirth T (2005) Iodine monochloride-amine complexes: an experi-
mental and computational approach to new chiral electrophiles. Chemistry—A Eur-
opean Journal 11:5777–5785.
5. Sakakura A, Ukai A, Ishihara K (2007) Enantioselective halocyclization of polyprenoids
induced by nucleophilic phosphoramidites. Nature 445:900–903.
6. Brown RS, et al. (1994) Stable bromonium and iodonium ions of the hindered olefins
adamantylideneadamantane and Bicyclo[3.3.1]nonylidenebicyclo[3.3.1]nonane. X-ray
structure, transfer of positive halogens to acceptor olefins, and ab initio studies. J Am
Chem Soc 116:2448–2456.
27. Brown HC, Brewster JH, Shechter H (1954) An interpretation of the chemical behavior
of five- and six-membered ring compounds. J Am Chem Soc 76:467–474.
28. Bruno P, Caselli M, Fragale C, Magrino S (1977) Charge transfer complexes of hexam-
ethylphosphoramide chalcogenides. J Inorg Nucl Chem 39:1757–1759.
29. Glera J, Sobczyk L, Paetroldt R (1980) Correlation between the Dipole moments
and thermodynamical data of the iodine complexes with organic oxides, sulfides,
and selenides. J Phys Chem 84:2602–2605.
30. Julia MM, Guy-Roualt A (1964) Iodolactonization and dihydroxylation of 5-phenyl
pentenoic acids (in French). C R Hebd Seances Acad Sci 258:3728–3730.
31. Walton H, Schilt A (1952) The ionization constant and rate of hydrolysis of succinimide.
J Am Chem Soc 74:4995–4996.
32. Goldberg RN (1999) Thermodynamic quantities for the ionization reactions of buffers.
J Phys Chem Ref Data 31:231–370.
33. Bagno A, Scorrano G (1988) Acid-base properties of organic solvents. J Am Chem Soc
110:4577–4582.
34. Crich D, Beckwith ALJ, Filzen GF, Longmore RW (1996) Free radical chemistry of
lactones: ring contractions and expansions. J Am Chem Soc 118:7422–7423.
35. Galpin DR, Bobbink SR, Ary TE (1972) cis- and trans-3-Methylamino-2-phenyltetrahy-
dropyran. J Pharm Sci 61:963–964.
1
1
þ
þ
1
7. Neverov AA, Brown RS (1996) Br and I transfer from the halonium ions of adaman-
tylideneadamantane to acceptor olefins. halocyclization of 1,ω-alkenols and alkenoic
acids proceeds via reversibly formed intermediates. J Org Chem 61:962–968.
2
0660
∣
www.pnas.org/cgi/doi/10.1073/pnas.1005296107
Denmark and Burk