Dissecting the stereocontrol elements of a catalytic asymmetric chlorolactonization: Syn addition obviates bridging chloronium

Article abstract of DOI:10.1021/ja4072145

We report absolute and relative stereochemistry of addition in enantioselective chlorolactonizations of 4-phenyl-4-pentenoic acid and its related t-butyl ester, catalyzed by (DHQD)₂PHAL. Predominant syn addition of the chlorenium and the nucleophile across the olefin is observed. As shown by isotopic labeling, NMR spectroscopy, and derivative studies, the two new stereocenters formed by addition across the double bond are set independently and influenced by different factors. These findings suggest a stepwise process via an intermediate capable of lactone closure with either stereochemistry, in contradistinction to the more familiar scenario in which anti addition is dictated by a bridging chloronium ion intermediate.

Full text of DOI:10.1021/ja4072145

Communication
pubs.acs.org/JACS
Dissecting the Stereocontrol Elements of a Catalytic Asymmetric
Chlorolactonization: Syn Addition Obviates Bridging Chloronium
Roozbeh Yousefi,† Kumar Dilip Ashtekar,† Daniel C. Whitehead,‡ James E. Jackson,*,†
and Babak Borhan*,†
†Department of Chemistry, Michigan State University, East Lansing, Michigan 48824, United States
‡Department of Chemistry, Clemson University, Clemson, South Carolina 29634, United States
S
* Supporting Information
across the double bond. Until recently, however, control of
absolute stereochemistry at the newly formed sp3 centers was
ABSTRACT: We report absolute and relative stereo-
chemistry of addition in enantioselective chlorolactoniza-
tions of 4-phenyl-4-pentenoic acid and its related t-butyl
ester, catalyzed by (DHQD)2PHAL. Predominant syn
addition of the chlorenium and the nucleophile across the
olefin is observed. As shown by isotopic labeling, NMR
spectroscopy, and derivative studies, the two new stereo-
centers formed by addition across the double bond are set
independently and influenced by different factors. These
findings suggest a stepwise process via an intermediate
capable of lactone closure with either stereochemistry, in
contradistinction to the more familiar scenario in which
anti addition is dictated by a bridging chloronium ion
intermediate.
lacking, its development hindered by the following challenges:
(a) To be useful, the catalyzed, stereocontrolled process must
outcompete stereorandom noncatalyzed background reactions.
Thus, the halenium donor must react with the olefin on a
practical time scale, but only when these components are
activated by catalysts. (b) The catalyst must survive the
presence of such an electrophilic halogenating agent while
directing reactions to specific olefin faces. (c) Assuming a
stepwise process, high alkene face selectivity in the initial
electrophilic attack is no guarantee of ultimate enantio-
induction; the putative carbocation (whether or not it
equilibrates with a bridged halonium ion) may undergo bond
rotation, erasing configurational memory at the cationic site.
Thus catalyst control of the final ring closure is essential. (d)
Finally, recent sophisticated studies of bromo- and iodocycliza-
tions have uncovered olefin-to-olefin halenium transfer, another
potential mode of stereorandomization prior to ring closure.1
The analogous exchanges in chlorenium ions, however, were
ruled out, confirming earlier gas-phase and computational
studies and consistent with results described herein.
alocyclization of alkenes is a robust, versatile route to a
Hwide range of heterocycles. Typical textbook halocycliza-
tion mechanisms invoke electrophilic halogen attack to form a
three-membered halonium ion intermediate; ring closure then
ensues (Figure 1, path a) via intramolecular SN2 attack on the
halonium ion. This scenario predicts anti relative addition
Despite the above pitfalls, the last three years have witnessed
great progress,2 highlighted by the discovery of several efficient
catalytic asymmetric halocyclizations with excellent enantiose-
lectivity.3 As more examples emerge, tools to probe the origins
of stereocontrol will be essential to provide mechanistic insight
and guidance to the field.
In recently disclosed efforts, we showed that cinchona
alkaloid dimers, such as (DHQD)2PHAL, can catalyze efficient,
stereoselective chlorolactonization of alkenoic acids 1a, as
depicted in Figure 1.2g,4 Furthermore, the same alkaloid has
demonstrated proficiency in several other catalytic asymmetric
halocyclizations.2f,5 To understand this system’s effectiveness
and generality as a path to chiral heterocyclic frameworks, we
have explored its stereochemical details.
Key questions to address are: (a) Is the chlorenium delivered
in a face-selective manner? (b) What stereochemical relation-
ship, if any, exists between the chlorenium delivery and the
nucleophilic attack? (c) For a stepwise process, what would be
the likely nature of the reactive intermediate, i.e., a bridged
chloronium or a carbocation as in Figure 1?
Figure 1. (DHQD)2PHAL-catalyzed chlorolactonization process.
Carbocation or a bridged chloronium species represent potential
intermediates is in question.
Received: July 15, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja4072145 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
The literature offers a start on this last question. For
substrates like 1a, halocyclization may be a polar, stepwise
process, as in the textbook scenario. However, the open
carbocation should be energetically preferred, as seen in Olah’s
seminal work on chloronium intermediates of similarly
substituted olefins.6 Previous studies7 and our own quantum
chemical modeling support this view; even in the “gas phase”
where bridging is the only available mode of charge
delocalization, no cyclic chloronium ion energy minimum
deep enough to lock in stereochemistry is seen from 1a (or any
α-alkylstyrene) or even from a simple 1,1-dialkyl alkene, such as
2-methylpropene.8 Thus, an open halomethyl carbenium ion
(Figure 1, path b) would be the intermediate expected from
halenium attack on a 1-aryl-1-alkylalkene substrate 1a.
Halocyclization of the 1,1-disubstituted olefin 1a forms
lactone 2 with a single stereocenter; the chlorine resides on a
nonstereogenic carbon, with no record of its attack path on the
alkene. In the expected planar carbenium ion intermediate, the
absolute stereochemistry defined by asymmetric delivery of the
chlorenium to the olefin would be lost, so the reaction’s
enantioselectivity would then be determined at the (presumably
catalyst controlled) ring-closing step. Notably, if the reaction
were to involve a bridged chloronium (i.e., a species in which
both ends are stereochemically committed), stereochemical
definition would be preserved as the carboxylate closed the
ring. Enantioselectivity would be controlled in parallel with the
initial asymmetric chlorenium delivery, yielding anti addition.
To probe the scenarios described above, we required a
labeled substrate that could report on the stereochemical fate of
the chlorine atom, without sacrificing the structural uniqueness
of the 1,1-disubstituted carboxylic acid substrate. Accordingly,
the E-deuterated analog 1a-D was synthesized in short order, as
shown in Figure 2a. The chlorolactonization of 1a-D under
standard catalytic asymmetric conditions yielded the corre-
sponding deuterated product 2-D. As described below, the
absolute stereochemistry of the CHDCl group in the major
isolate was straightforwardly established via NOE analysis of
epoxide 3-D obtained from the chemically transformed
chlorolactone product 2-D. Lithium borohydride reduction of
lactone 2-(R), followed by sodium hydroxide mediated
cyclization of the resulting chlorohydrin intermediate, returned
the 1,1-disubstituted epoxy alcohol 3 in good yield (Figure 2b).
NOESY analysis of 3 showed a pronounced correlation
between the epoxy proton at 2.98 ppm and the methylene
protons of the alkyl chain, indicating that the protons at 2.73
and 2.98 ppm are, respectively, cis (Ha) and trans (Hb) to the
phenyl ring. Conversion of deuterated (5R)-lactone 2-D (91:9
er from catalytic asymmetric chlorolactonization of 1a-D with
(DHQD)2PHAL to the corresponding epoxy alcohol 3-D
yielded a 1H NMR spectrum with the major epoxy CH
resonance (90%) at 2.98 ppm (Hb). With the deuteron cis to
the phenyl group, the carbon bearing it was assigned the S-
configuration, implying the R-configuration for the CHDCl
group in the (5R) 2-D from which epoxide 3-D was formed via
intramolecular SN2 chloride displacement. This result, in turn,
allowed assignment of the 1H resonances from the pro-R (3.74
ppm) and pro-S (3.83 ppm) diastereotopic hydrogens of the
CH2Cl group in 2 and CHDCl in 2-D (see SI). The latter
assignment enables identification of all four isomers obtained
via chlorolactonization of 1a-D and 1b-D.
Figure 3 depicts the results obtained from the chlorolacto-
nization of 1a-D under (DHQD)2PHAL catalyzed reaction
Figure 3. Reaction of deuterated substrate 1a-D under standard
catalytic asymmetric reaction conditions with (DHQD)2PHAL
provided four isomeric products. The anticipated 5R product was
the major HPLC isolate (∼91:9 5R:5S). The level of C-6
stereoisomerism (6R:6S = 97:3 for 5R, 96:4 for 5S) was quantified
1
by H NMR analysis of crude and HPLC purified fractions.
conditions. Though 1a-D was a 83:5:12 mixture of E:Z:H2
isotopomers, numbers provided in Figure 3 are corrected to the
pure E value (see SI). Noncatalyzed cyclization of 1a-D in the
presence of DCDMH yielded a racemic mixture of the two
diastereomeric products in a 1:1 dr.9 Inclusion of quinuclidine
as an achiral catalyst (at 20 mol %, to equal the concentration
of amine sites in (DHQD)2PHAL at 10 mol %) led to a 5:1
anti:syn dr (see SI). Formation of both diastereomers excludes
reaction via a single, stereospecific pathway, requiring instead
an overall sequence capable of multiple stereochemical
outcomes. Reaction with quinuclidine does introduce an anti
stereopreference, consistent with the idea of enhanced
carboxylate nucleophilicity in the presence of the quinuclidine;
in the suggested carbocation pathway, the cation would then be
Figure 2. (a) Synthesis of deuterated substrates 1a-D and 1b-D. (b)
Absolute stereochemical assignment of the deuterated product 2-D via
conversion to the epoxide 3-D.
B
dx.doi.org/10.1021/ja4072145 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
trapped fast enough to retain some conformational memory of
the orientation of chlorenium delivery.
The face selectivity of chlorenium transfer is formally
inconsequential for the enantioselectivity of lactonization with
unlabeled 1,1-disubstituted alkenoic acids, such as 1a. But
revealing the new stereocenters’ absolute preferences (or lack
thereof) sheds light on the factors that control selectivity.
Stereocontrol at both new sp3 sites might reflect binding of the
substrate to the protonated catalyst in a conformation “cocked”
for formation of both new bonds, or the two bond-forming
events could be independently controlled by the same catalyst.
A key hint favoring the latter interpretation is that the minor 5S
enantiomer in the above reaction also showed a strong
preference for the R stereochemistry at the CHDCl site (i.e.,
net anti addition).
To better segregate face selectivity in chlorenium transfer
from the intramolecular cyclization that yields the lactone
product, the transformations of t-butyl esters 1b and 1b-D were
investigated. Chlorolactonization of 1b under standard catalytic
asymmetric conditions with (DHQD)2PHAL led to 2 with the
5S isomer weakly predominant, at 20% ee (Figure 4). These
Olefin-to-olefin chlorenium ion transfer, already noted by
Denmark et al. as least probable,1 is a conceivable stereo-
randomizing factor. This possibility is largely eliminated by the
following observations: (a) For the catalyzed process in Figure
3, the 6R stereopreference is 96:4, allowing for a randomization
process of at most 4%. (b) Quantum chemical models without
catalyst find an open (unbridged) chloromethyl carbenium ion
whose low barrier (<3.5 kcal/mol) for CHDCl rotation
indicates that the newly attached Cl atom exerts no
stereochemical control over bond formation at C-5.8 (c)
Even the slowest of the above reactions shows no loss of
stereochemical integrity for the labeled olefinic carbon of
recovered starting material 1a-D. Since chlorenium ion loss
from the conformationally mobile carbocation would scramble
label in the alkenoic acid, this finding implies that cyclization is
faster than any reversible chlorenium transfer. (d) Cyclization is
also irreversible; control experiments were run using pure
enantiomers (isolated by HPLC) of the p-OMe phenyl analogs
of chlorolactone 2. This substrate, chosen for its potential
ability to form a stabilized benzylic carbenium ion, showed no
loss of stereochemical integrity, even in the presence of reacting
components. Thus, there is no reversibility in the ring closure
step (see SI for details).
Considering the diversity of stereochemical outcomes above,
it appears likely that in the presence of the (DHQD)2PHAL,
the conversion of 1a to 2 also proceeds through the
intermediacy of a carbocation intermediate. If so, the high
enantioselectivity of the final lactone product must be the result
of high facial selectivity in the ring-closing step. However, the
observed ee of the product need not reflect the facial selectivity,
if any, for the chlorenium transfer to the olefin. Using the E-
deuterated analog 1a-D and NMR assignments outlined above
to probe both absolute stereochemistry at the CHDCl site and
the relative addition stereochemistry of the chlorine and the
carboxylate nucleophile, this hypothesis was readily tested.
Chlorolactonization of 1a-D was carried out under optimized
reaction conditions (Figure 2). The proportions of the four
stereoisomeric products 2-D, depicted in Figure 3, were
quantified as follows: chiral HPLC separation of C-5 epimers
enabled isolation of the expected major product, the 5R
diastereomeric mixture. The C-6 CHDCl epimer ratios of the
two C-5 antipodes could then be quantified through NMR
analysis of crude products and HPLC isolates of major and
minor fractions (details in SI).
Figure 4. Chlorolactonization of the t-butyl ester analog 1b under the
standard catalytic asymmetric reaction conditions with
(DHQD)2PHAL produces the 5S product in low ee. The deuterated
1
analog 1b-D yields the same C-5 product stereoisomers. H NMR
analysis of the HPLC purified fractions reveals high olefin face
selectivity in chlorination in the first step as compared to poor
carbocation face selectivity in the second step.
In agreement with our previous results, the R stereochemistry
results suggest that the carboxylic acid moiety is important for
achieving high enantioselectivity, presumably as a result of
hydrogen bonding or ionic interactions with the chiral amine
catalyst. The analogous reaction of the deuterated E-t-butyl
ester analog 1b-D under identical reaction conditions yielded a
surprise: NMR analysis of the HPLC purified fractions as
described above indicated a much higher facial selectivity in
chlorenium transfer to the olefin (via integration of the
diastereotopic protons) than for the cyclization to the lactone
product. As with 1a, the R configuration dominates in the C-6
CHDCl group, but this is true in both 5S and 5R enantiomers.
Thus, the dominant 5S product 2 from 1b actually arises from
mostly anti addition.
of the quaternary carbon for lactone 2-D was favored by 91%
1
(HPLC). Interestingly, H NMR analysis of the major fraction
isolated by HPLC clearly showed one major diastereomeric
1
product. Integration of the diastereotopic H peaks revealed a
97:3 preference for one CHDCl epimer; assignment of the
CH2Cl group’s 1H resonances as discussed above and shown in
Figure 2b proved this dominant form to have a CHDCl group
of R configuration. Thus (DHQD)2PHAL strongly controls
face selectivity of the initial chlorenium delivery and favors syn
addition in the major product. This relative stereochemistry is
opposite to that expected from a bridged chloronium
intermediate. This high level of stereoinduction is striking
and suggests that in systems beyond 1,1-disubstituted olefins,
nonbridged chloronium mechanisms10 may also apply when the
Cl−C bonds form enantioselectively.2f,5
Taken together, the above findings suggest that independent
events determine facial selectivity in chlorination of the olefin
and cyclization of the ensuing intermediate. The high R-
C
dx.doi.org/10.1021/ja4072145 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
H.; Jiang, J.; Chen, W. L.; Lin, L. L.; Feng, X. M. J. Am. Chem. Soc.
2011, 133, 5636. (i) Sakakura, A.; Ukai, A.; Ishihara, K. Nature 2007,
445, 900. (j) Zhang, W.; Zheng, S. Q.; Liu, N.; Werness, J. B.; Guzei, I.
A.; Tang, W. P. J. Am. Chem. Soc. 2010, 132, 3664.
(3) (a) Castellanos, A.; Fletcher, S. P. Chem.−Eur. J. 2011, 17, 5766.
(b) Tan, C. K.; Zhou, L.; Yeung, Y.-Y. Synlett 2011, 1335.
(c) Hennecke, U. Chem. Asian J. 2012, 7, 456. (d) Chen, F.; Tan,
C. K.; Yeung, Y. Y. J. Am. Chem. Soc. 2013, 135, 1232. (e) Jiang, X.;
Tan, C. K.; Zhou, L.; Yeung, Y. Y. Angew. Chem. 2012, 51, 7771.
(4) Yousefi, R.; Whitehead, D. C.; Mueller, J. M.; Staples, R. J.;
Borhan, B. Org. Lett. 2011, 13, 608.
(5) (a) Garzan, A.; Jaganathan, A.; Salehi Marzijarani, N.; Yousefi, R.;
Whitehead, D. C.; Jackson, J. E.; Borhan, B. Chem.−Eur. J. 2013, 19,
9015. (b) Ikeuchi, K.; Ido, S.; Yoshimura, S.; Asakawa, T.; Inai, M.;
Hamashima, Y.; Kan, T. Org. Lett. 2012, 14, 6016. (c) Wilking, M.;
Muck-Lichtenfeld, C.; Daniliuc, C. G.; Hennecke, U. J. Am. Chem. Soc.
2013, 135, 8133.
(6) Olah, G. A.; Bollinger, J. M. J. Am. Chem. Soc. 1968, 90, 947.
(7) (a) Haubenstock, H.; Sauers, R. R. Tetrahedron 2004, 60, 1191.
(b) Haubenstock, H.; Sauers, R. R. Tetrahedron 2005, 61, 8358.
(c) Ohta, B. K.; Hough, R. E.; Schubert, J. W. Org. Lett. 2007, 9, 2317.
(d) Solling, T. I.; Radom, L. Int. J. Mass Spectrom. 1999, 185, 263.
(8) For the C9H10Cl+ ion obtained from chlorenium addition to α-
methylstyrene, B3LYP/6-31G* calculations find only an open
carbocation minimum. The optimized structure does find the C−Cl
bond aligned with the carbocation’s empty 2p orbital, but it shows a
∠CCCl angle of 109°, and the face-switching barrier to rotation of the
CH2Cl group is calculated to be only 1.6 kcal/mol in the gas phase,
roughly half the value for methyl group rotation in ethane. Simulated
solvation in CHCl3 further lowers this value to 0.9 kcal/mol. See SI for
details of calculations and structures.
selectivity in the first step may be due to attack by the
chlorinated hydantoin on the styrenic portion common to
substrates 1a and 1b, bound by catalyst so that only the pro-R
face is accessible.11 The selectivity for the second step, which is
so strongly modified by esterification that its stereopreference is
inverted, would then be controlled by catalyst templating of the
cation’s conformational preferences, setting the face selectivity
of the ring-closing step. Given the hydrogen binding moieties of
acid 1a, largely absent in less polar, more sterically bulky ester
1b, it seems sensible that cyclization of the cation from 1b
should be much less strongly directed by its interaction with the
polyamine catalyst.
In summary, to answer the questions raised at the outset: (a)
Chlorenium delivery to alkene sites in 1a and 1b, catalyzed by
(DHQD)2PHAL, is highly face-selective. But this pro-R face
selectivity is not a sufficient condition to ensure enantiose-
lective lactone production. (b) Following attack of chlorenium
on acid 1a, catalyst-templated nucleophilic closure favors the
5R over the 5S lactone by a factor of >10:1, regardless of the
original Cl+ delivery path. Thus, the two sites’ stereo-
preferences, though determined independently, lead to
predominant syn Cl,O addition from acid 1a. The strong pro-
R preference of chlorenium attack, however, results in net anti
addition in the weakly favored 5S product of ester 1b.12 (c)
With the two stereochemical decisions apparently uncoupled,
the reaction is most straightforwardly understood as stepwise
via a carbocation intermediate. A bridged chloronium ion is
both energetically and stereochemically incompatible with this
and many analogous reported reaction systems. Thus, we
believe that independent catalyst stereocontrol of the two new
bond formation steps is ultimately responsible not only for
controlling which olefin face is attacked by Cl+ but also for
guiding the final enantioselective cyclization.
(9) Although dichlorodiphenyl hydantoin yields slightly higher ees,
we opted for the use of DCDMH because of its better solubility in the
chosen solvents for this study.
(10) (a) Fahey, R. C.; McPherson, C. A. J. Am. Chem. Soc. 1969, 91,
3865. (b) Fahey, R. C.; Schubert, C. J. Am. Chem. Soc. 1965, 87, 5172.
(11) X-ray or NOE results could suggest a structural model for
catalyst−substrate−DCDMH interactions; however, assiduous efforts
to obtain crystals in various solvents and temperature conditions
yielded only uncomplexed catalyst crystals. Likewise, detailed NMR
studies failed to yield interpretable data on the relevant spatial
relationships.
(12) We cannot absolutely exclude the possibility of multiple
competing pathways that independently yield the four stereoisomeric
products. For instance, a double inversion process can be envisioned
wherein a nearby nucleophile (e.g., catalyst’s phthalazine nitrogen
atom) forms a covalent catalyst−substrate intermediate, which then
undergoes SN2 nucleophilic reaction to yield the dominant syn
chlorolactonization product. However, the diversity of stereochemical
outcomes in the products from 1a and 1b argues against such
stereochemically defined intermediates, supporting instead the notion
of a common carbocation which then undergoes ring closing under the
stereocontrol of the catalyst.
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental details, characterization data, and stereochemical
assignments. This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Authors
babak@chemistry.msu.edu
jackson@chemistry.msu.edu
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
We acknowledge the NSF (CHE-0957462) for funding.
■
REFERENCES
■
(1) Denmark, S. E.; Burk, M. T.; Hoover, A. J. J. Am. Chem. Soc.
2010, 132, 1232.
(2) (a) Murai, K.; Matsushita, T.; Nakamura, A.; Fukushima, S.;
Shimura, M.; Fujioka, H. Angew. Chem., Int. Ed. 2010, 49, 9174.
(b) Veitch, G. E.; Jacobsen, E. N. Angew. Chem., Int. Ed. 2010, 49,
7332. (c) Dobish, M. C.; Johnston, J. N. J. Am. Chem. Soc. 2012, 134,
6068. (d) Li, G. X.; Fu, Q. Q.; Zhang, X. M.; Jiang, J.; Tang, Z.
Tetrahedron-Asymmetry 2012, 23, 245. (e) Tan, C. K.; Le, C. C.;
Yeung, Y. Y. Chem. Commun. 2012, 48, 5793. (f) Jaganathan, A.;
Garzan, A.; Whitehead, D. C.; Staples, R. J.; Borhan, B. Angew. Chem.,
Int. Ed. 2011, 50, 2593. (g) Whitehead, D. C.; Yousefi, R.; Jaganathan,
A.; Borhan, B. J. Am. Chem. Soc. 2010, 132, 3298. (h) Cai, Y. F.; Liu, X.
D
dx.doi.org/10.1021/ja4072145 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX