Three-dimensional correlation of steric and electronic free energy relationships guides asymmetric propargylation

Article abstract of DOI:10.1126/science.1206997

Chemical reaction outcomes are often rationalized on the basis of independent analyses of steric and electronic effects. We applied three-dimensional free energy relationships correlating steric and electronic effects to design and optimize a ligand class

Full text of DOI:10.1126/science.1206997

Three-Dimensional Correlation of Steric and Electronic Free Energy
Relationships Guides Asymmetric Propargylation
Kaid C. Harper and Matthew S. Sigman
Science 333, 1875 (2011);
DOI: 10.1126/science.1206997
This copy is for your personal, non-commercial use only.
If you wish to distribute this article to others, you can order high-quality copies for your
colleagues, clients, or customers by clicking here.
Permission to republish or repurpose articles or portions of articles can be obtained by
following the guidelines here.
The following resources related to this article are available online at
www.sciencemag.org (this information is current as of November 11, 2014 ):
Updated information and services, including high-resolution figures, can be found in the online
version of this article at:
http://www.sciencemag.org/content/333/6051/1875.full.html
Supporting Online Material can be found at:
http://www.sciencemag.org/content/suppl/2011/09/28/333.6051.1875.DC1.html
A list of selected additional articles on the Science Web sites related to this article can be
found at:
http://www.sciencemag.org/content/333/6051/1875.full.html#related
This article cites 40 articles, 4 of which can be accessed free:
http://www.sciencemag.org/content/333/6051/1875.full.html#ref-list-1
This article has been cited by 4 articles hosted by HighWire Press; see:
http://www.sciencemag.org/content/333/6051/1875.full.html#related-urls
This article appears in the following subject collections:
Chemistry
http://www.sciencemag.org/cgi/collection/chemistry
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the
American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright
2011 by the American Association for the Advancement of Science; all rights reserved. The title Science is a
registered trademark of AAAS.

REPORTS
tion state energies (DDG‡) (18, 35–38). Although
Three-Dimensional Correlation of Steric
and Electronic Free Energy Relationships
these efforts led to qualitative insight into impor-
tant structural features responsible for asymmetric
catalysis, extrapolation to (or identification of )
new catalysts was challenging. Therefore, we de-
veloped a method to simultaneously examine and
correlate the steric effects of two key ligand sub-
stituents through the use of mathematical surface
models (38). This approach combines experimen-
tal design principles with the systematic synthesis
of ligand libraries. Manipulation of the resultant
surface models allowed for successful prediction
of previously untested catalysts through extra-
polation and conclusive identification of opti-
mal catalysts for a given Nozaki-Hiyama-Kishi
(NHK) carbonyl allylation, wherein a carbonyl-
Guides Asymmetric Propargylation
Kaid C. Harper and Matthew S. Sigman*
Chemical reaction outcomes are often rationalized on the basis of independent analyses of steric and
electronic effects. We applied three-dimensional free energy relationships correlating steric and
electronic effects to design and optimize a ligand class for the enantioselective Nozaki-Hiyama-Kishi
propargylation of ketones. The resultant mathematical model describing the steric and electronic
parameter relationship is highly reliant on the synergistic interactions of these two effects.
omplex chemical reactions can proceed asymmetric catalyst design principle (13, 14). bearing substrate reacts with a Cr(III)-allyl spe-
through numerous pathways to many dif- However, electronic effects, both inductive and cies generated from the reaction of Cr(II) and
ferent products, although synthetic utility conjugative, can have an equally important im- an allyl halide (39). We were initially interested
C
generally requires enrichment of a single product, pact on reaction enantioselectivity (15–18). Al- in expanding the use of this ligand-based de-
presumably through a preferred low-energy path- though computational techniques accounting for sign of experiments approach to other NHK-
way. This selectivity is commonly ascribed to two steric and electronic effects on asymmetric cat- type reactions.
specific properties: electronic and steric effects. alyst performance have made great strides over
Of several possible reaction types, the NHK
Electronic effects refer to variations in orbital elec- the past decade to aid catalyst design (19–30), it propargylation of ketones drew our interest
tron density through bonding or spatial interac- is difficult to implement these approaches for because the homopropargylic tertiary alcohol
tions (1). An electronic effect is a general term mechanistically and structurally ill-defined reac- products constitute a synthon for use in target-
that can include inductive effects, conjugative tions (31, 32). This is especially true for a priori oriented synthesis (40). Shibasaki and colleagues
(resonance) effects, or molecular orbital sym- understanding of catalyst electronic effects and (41) and Fandrick et al. (42) have recently and
metry considerations. Steric effects arise from their potential synergistic relationship to steric effects. independently reported catalytic, enantioselective
electrostatic repulsion as atoms or molecules ap- Therefore, a practitioner of asymmetric catalyst de- ketone propargylations via a Cu-catalyzed addi-
proach each other, as commonly modeled by the velopment often relies on an independent qualita- tion of boronic ester derivatives to ketones. Cata-
Lennard-Jones potential (1–4).
tive evaluation of catalyst steric and electronic lytic enantioselective propargylation of ketones
Quantitative parameterization of electronic properties, although quantitative structure-selectivity using the commercially available and attractive
effects has been extensively used in physical relationships, analogous to the structure-activity propargyl halide reagents in the NHK protocol
organic chemistry as a means of probing re- approaches used in pharmaceutical optimization, has not been reported. Therefore, we submitted
action mechanisms, often by application of the have generally been used for ex post facto rationale the oxazoline-proline ligand library 3 to the enan-
Hammett s-value framework (5, 6). These pa- of catalyst performance (19–22, 33, 34). Here, tioselective ketone propargylation using propargyl
rameters reflect the relative acidity of benzoic we describe a facile method to quantitatively and bromide (Fig. 2). A mathematical surface mod-
acid derivatives (Fig. 1) and are generally consid- simultaneously probe steric and electronic ef- el (Eq. 1) was fit to the results:
ered robust. In contrast, quantifying steric effects fects, using basic principles of physical organic
DDG‡ ¼ −1:87 þ 1:34X þ
has been historically controversial, with several chemistry in the context of a chiral ligand class
different sets of parameters developed over the for enantioselective ketone propargylation.
3:03Y − 0:99Y2 − 1:19XY þ
years, each based on different measurable
Previous efforts in our group have focused on
0:62XY2 − 0:73YX 2
ð1Þ
Qualitative evaluation reveals a relatively shal-
quantities or computations (7). Steric parameters independently correlating electronic and steric
introduced by Taft in the 1950s were derived effects as a function of catalyst substituent with
from the relative rate of acid-catalyzed hydrolysis the observed enantiomeric ratio (e.r.), which is
of methyl esters (Fig. 1), which were later cor- directly proportional to the difference in transi- low surface, with a predicted optimal catalyst
related to the van der Waals radii of the substit-
uent by Charton (7–12). The Taft-Charton steric
Fig. 1. Classic steric and electronic
parameters have seen the most extensive use, al-
parameters in organic chemistry.
though debate on their validity continues (7).
The importance of both steric and electron-
ic effects is magnified in asymmetric catalysis,
where the diastereomeric transition states leading
to enantiomerically enriched products can differ
by as little as 1 to 3 kcal/mol. Of the two, steric
effects, generally through the introduction of a
large group to enhance repulsive interactions be-
tween a substituent(s) on the catalyst and the
substrate, are most commonly implicated as an
Department of Chemistry, University of Utah, Salt Lake City, UT
84112, USA.
*To whom correspondence should be addressed. E-mail:
sigman@chem.utah.edu
www.sciencemag.org SCIENCE VOL 333 30 SEPTEMBER 2011
1875

REPORTS
performance of only 75:25 e.r. for this ligand for the systematic evaluation of electronic param- and 6a, respectively) were synthesized and eval-
scaffold. This modeling suggests that satisfactory eters; we also incorporated a steric element into uated in the enantioselective NHK propargylation
enantioselectivity could not be achieved using its substitute, motivated by the observed enhance- of acetophenone. Substituting pyridine for the
this ligand class, and thus a different ligand scaf- ment of enantioselectivity with larger oxazoline oxazoline did not lead to improvement in enantio-
fold was targeted.
substituents, as seen for ligand 4 (Fig. 3A). Spe- selectivity, but instead suggested that an aromatic
To design a distinct ligand template, we opted cifically, we selected aromatic heterocycles as a heterocycle was a functional oxazoline analog
to conserve the proline component of 3 because replacement allowing access to similar chelation of our previous catalyst system. Use of the larger
of its proven effectiveness in asymmetric cataly- environments about the Cr catalyst, while also heterocycle (6a), quinoline, yielded a higher e.r.
sis and the relative ease with which its steric allowing for systematic electronic perturbations. for this difficult reaction class.
profile can be synthetically modulated (43, 44). To test the viability of this ligand design, both the
The quinoline-proline (QuinPro)–based lig-
We replaced the oxazoline component to allow pyridine-proline and quinoline-proline ligands (5 and structure is perfectly suited to simultane-
A
BB
CC
X = Me; Y = tBu
er = 75:25
Fig. 2. (A) Ligand library based
evaluation of ketone propargylation. (B) 3D surface derived
from ligand library evaluation, which relates the Charton
value (u) of groups X and Y to the measured enantioselectivity. The surface is described by Eq. 1 and the
predicted maximum is depicted. (C) Contour plot of the same surface (1).
A
B
C
D
Fig. 3. (A) Ligand evolution for the Nozaki-Hiyama-Kishi propargylation of modeled by Eq. 4 (raw data in table S2). (C) Contour plot of Eq. 4. (D) Com-
ketones, which led to the development of the QuinPro ligand scaffold. (B) Surface parison of the DDG‡ values predicted by Eq. 4 and the averaged measured value
derived from the evaluation of the QuinPro library with replicate runs and experimentally observed for each of the nine ligands in the QuinPro library.
1876
30 SEPTEMBER 2011 VOL 333 SCIENCE www.sciencemag.org

REPORTS
ously optimize the effects of electronic and steric verified by a p test. The design and response ketones, where the same catalyst was found to
parameters. The ligand substituents that con- matrices were used to solve for the matrix of be optimal.
stitute the library were selected on the basis of parameters (C) using the linear algebra definition
experimental design principles, where the incor- of regression (Eq. 3) (46, 47):
poration of substituents evenly spans a wide pa-
The standard approach to asymmetric cat-
alyst development is to evaluate catalyst proper-
ties independently and individually. Inspection
of the data in Fig. 3B reveals that this approach
would be unlikely to expose the synergy between
C ¼ ðXT XÞ−1ðXT YÞ
ð3Þ
rameter range of both Hammett and Charton
values (5, 6, 9–12). A nine-membered (3 × 3)
ligand library was targeted to balance the syn-
Equation 4 defines the model with the most the steric and electronic effects in this catalyst
thetic effort with the statistical requirements statistical significance:
for developing a three-dimensional (3D) mod-
system. For example, steric effects are substan-
tially mitigated for electron-poor ligands (E =
CF3), whereas a considerable enhancement of the
steric effect is observed for electron-rich ligands
(E = OMe). Furthermore, the electronic effect for
smaller substituents (S = Me) is diminished as
compared to the effect observed for the series
DDG‡ ¼ −1:20 þ 1:22E þ
el (38).
The desired ligands were synthesized and
evaluated with replicates in the propargylation of
acetophenone under the conditions specified in
Fig. 3A (45). The data set required that a 3D
surface function be developed to accurately mod-
2:84S − 0:85S2
−
3:79ES þ 1:25ES2
ð4Þ
The surface in Fig. 3B describes this equa- containing S = tBu. Simply put, a poor choice
el the data. From inspection of the raw data, we tion, and the resultant contour plot is depicted of one variable could not be superseded by an
chose as the base function a full third-order poly- in Fig. 3C. Evaluation of the goodness of fit ideal choice in the other.
nomial (Eq. 2) that appeared to adequately describe (Fig. 3D) shows good correlation of the orig-
This empirical inspection is strongly rein-
the observed variation:
inal experimentally measured enantioselect- forced by the derived mathematical model de-
ivities with those output by the final model scribed in Eq. 4. Inclusion of crossterms between
(predicted value). The optimal ligand as de- the putatively independent variables not only con-
scribed by Eq. 4 lies nearest E = O-methyl siderably improves the model, but also mathemat-
(OMe), S = tert-butyl (tBu). The relative ease of ically defines a significant synergistic relationship
synthesis as well as the general availability of between the electronic and steric interactions in
the ligand precursors makes this particular lig- the catalyst. Specifically, the terms ES and ES2
DDG‡ ¼ z0 þ aE þ bS þ cE2 þ
dS2 þ fES þ gE3 þ hS3 þ
iES2 þ jSE2
ð2Þ
The data were arrayed into matrix format with and desirable. Although this ligand is incor- suggest that the electronic nature of the catalyst
the design matrix X (18 data points, nine individ- porated in the original data set, the combination and the steric environment in the transition state
ual ligands with their replicates, to allow for sta- of experimental design (evenly spreading the are closely correlated. Because the electronic
tistically meaningful modeling of the 10 parameters: parameters over a large range) (47) and math- variable itself contains no chiral information, the
1, E, S, E2, S2, ES, E3, S3, ES2, and SE2) and the ematical modeling of the data provides a high electronic effect on enantioselectivity presum-
corresponding response matrix Y (18 × 1) of level of confidence that 6b is the optimal ligand. ably is through mitigation of the Lewis acidity of
observed enantioselectivity. The polynomial base If the optimal ligand were not incorporated into the Cr center.
model was refined using backward elimination the original library, the modeling should allow
Noting that both aromatic and aliphatic ke-
regression techniques, wherein terms were added for effective new catalyst selection, as previously tones remain difficult substrate classes for these
to or omitted from the base polynomial function reported (38). Finally, the QuinPro library was carbonyl addition reactions, we first explored
iteratively according to the magnitude of the evaluated in the enantioselective propargylation the substrate scope by evaluating various aryl ke-
covariance (error) associated with each term as of 2-hexanone, a model substrate for aliphatic tone substrates using ligand 6b (Fig. 4). Generally,
Fig. 4. Substrate scope of the enantioselective ketone propargylation reaction.
www.sciencemag.org SCIENCE VOL 333 30 SEPTEMBER 2011
1877

REPORTS
27. P. H.-Y. Cheong, C. Y. Legault, J. M. Um, N. Çelebi-Ölçüm,
K. N. Houk, Chem. Rev. 111, 5042 (2011).
28. K. B. Lipkowitz, M. Pradhan, J. Org. Chem. 68, 4648 (2003).
29. P. J. Donoghue, P. Helquist, P.-O. Norrby, O. Wiest,
J. Chem. Theory Comput. 4, 1313 (2008).
30. P. J. Donoghue, P. Helquist, P.-O. Norrby, O. Wiest,
J. Am. Chem. Soc. 131, 410 (2009).
31. C. Allemann, R. Gordillo, F. R. Clemente, P. H.-Y. Cheong,
K. N. Houk, Acc. Chem. Res. 37, 558 (2004).
32. T. Dudding, K. N. Houk, Proc. Natl. Acad. Sci. U.S.A. 101,
5770 (2004).
33. S. E. Denmark, N. D. Gould, L. M. Wolf, J. Org. Chem. 76,
4337 (2011).
34. J. D. Oslob, B. Åkermark, P. Helquist, P.-O. Norrby,
Organometallics 16, 3015 (1997).
35. M. S. Sigman, J. J. Miller, J. Org. Chem. 74, 7633
(2009).
high enantioselectivity and good yield were ob- ing the interplay of two (or more) effects on a
served for aryl methyl ketones, with a modest reaction outcome.
drop in enantioselectivity for electron-poor sub-
strates. Other aryl ketones gave rise to good to
References and Notes
excellent e.r.’s, as highlighted by substrate 2h.
Heteroaromatics were also well tolerated un-
der the reaction conditions, with thiophene- and
furan-derived ketones undergoing highly enantio-
selective propargylation reactions (2i and 2j).
Additionally, an a,b-unsaturated ketone was an
excellent substrate for enantioselective propar-
gylation, leading to a 95:5 e.r. (2p).
In the case of aliphatic ketones, the catalyst
selected on the basis of steric differentiation, giv-
ing higher e.r.’s with increased steric bulk on a
single side of the ketone (2l to 2o). An e.r. of
85:15 observed for 2-hexanone (2l) is relatively
impressive, considering that the catalyst is differ-
entiating a methyl from an n-butyl group. Groups
with substitution at the a-position of the ketone
substantially enhanced the e.r.’s, as highlighted
by ketones with a cyclohexyl (96:4 e.r., 2n) and
a t-butyl group (98:2 e.r., 2o). Furthermore, a
g-butyrolactone (2q) could be synthesized with a
good e.r. from propargylation of an aliphatic ke-
tone with a pendant ester.
Our data suggest that steric-electronic corre-
lations provide a means for efficient optimization
of a catalytic system and are evidence for a syn-
ergistic relationship between these two classically
independent variables in reactions. This is es-
pecially attractive for optimizing reactions with
limited detailed mechanistic and structural
understanding and, considering that the model-
ing is tied to basic physical organic precepts, a
greater understanding of the underlying fea-
tures of asymmetric catalysis should result. The
application of this method is not limited to asym-
metric catalysis but can potentially be applied
to broad areas of chemistry dependent on evaluat-
1. E. V. Anslyn, D. A. Dougherty, in Modern Physical
Organic Chemistry, J. Murdzek, Ed. (University Science,
Sausalito, CA, 2006), pp. 441–482.
2. M. S. Newman, Ed., Steric Effects in Organic Chemistry
(Wiley, New York, 1956).
3. J. E. Jones, Proc. R. Soc. London Ser. A 106, 441 (1924).
4. J. E. Jones, Proc. R. Soc. London Ser. A 106, 463 (1924).
5. L. P. Hammett, J. Am. Chem. Soc. 59, 96 (1937).
6. C. Hansch, A. Leo, R. W. Taft, Chem. Rev. 91, 165 (1991).
7. C. Hansch, A. Leo, in Exploring QSAR: Fundamentals and
Applications in Chemistry and Biology, S. R. Heller,
Ed. (American Chemical Society, Washington, DC, 1995),
pp. 69–96.
8. R. W. Taft Jr., J. Am. Chem. Soc. 74, 3120 (1952).
9. M. Charton, J. Am. Chem. Soc. 97, 1552 (1975).
10. M. Charton, J. Am. Chem. Soc. 97, 3691 (1975).
11. M. Charton, J. Am. Chem. Soc. 97, 3694 (1975).
12. M. Charton, J. Org. Chem. 41, 2217 (1976).
13. E. N. Jacobsen, A. Pfaltz, H. Yamamoto, Eds.,
Comprehensive Asymmetric Catalysis I-III (Springer,
New York, 1999).
14. P. J. Walsh, M. C. Kozlowski, in Fundamentals of
Asymmetric Catalysis, J. Murdzek, Ed. (University Science,
Sausalito, CA, 2009), pp. 1–240.
15. A. L. Casalnuovo, T. V. RajanBabu, T. A. Ayers,
T. H. Warren, J. Am. Chem. Soc. 116, 9869 (1994).
16. D. W. Nelson et al., J. Am. Chem. Soc. 119, 1840 (1997).
17. M. Palucki et al., J. Am. Chem. Soc. 120, 948 (1998).
18. K. H. Jensen, M. S. Sigman, J. Org. Chem. 75, 7194 (2010).
19. M. C. Kozlowski, S. L. Dixon, M. Panda, G. Lauri,
J. Am. Chem. Soc. 125, 6614 (2003).
20. M. C. Kozlowski, M. Panda, J. Org. Chem. 68, 2061 (2003).
21. J. C. Ianni, V. Annamalai, P.-W. Phuan, M. Panda,
M. C. Kozlowski, Angew. Chem. 118, 5628 (2006).
22. K. B. Lipkowitz, C. A. D’Hue, T. Sakamoto, J. N. Stack,
J. Am. Chem. Soc. 124, 14255 (2002).
23. S. J. Zuend, E. N. Jacobsen, J. Am. Chem. Soc. 131,
15358 (2009).
24. C. Uyeda, E. N. Jacobsen, J. Am. Chem. Soc. 133, 5062
(2011).
25. S. J. Zuend, E. N. Jacobsen, J. Am. Chem. Soc. 129,
15872 (2007).
26. P.-O. Norrby, T. Rasmussen, J. Haller, T. Strassner,
K. N. Houk, J. Am. Chem. Soc. 121, 10186 (1999).
36. J. L. Gustafson, M. S. Sigman, S. J. Miller, Org. Lett. 12,
2794 (2010).
37. J. J. Miller, M. S. Sigman, Angew. Chem. Int. Ed. 47,
771 (2008).
38. K. C. Harper, M. S. Sigman, Proc. Natl. Acad. Sci. U.S.A.
108, 2179 (2011).
39. The resultant Cr(III)-alkoxide releases product through
reaction with TMSCl and Mn subsequently reduces the
catalyst to Cr(II).
40. C.-H. Ding, X.-L. Hou, Chem. Rev. 111, 1914 (2011).
41. S.-L. Shi, L.-W. Xu, K. Oisaki, M. Kanai, M. Shibasaki,
J. Am. Chem. Soc. 132, 6638 (2010).
42. K. R. Fandrick et al., J. Am. Chem. Soc. 133, 10332
(2011).
43. B. List, Tetrahedron 58, 5573 (2002).
44. J. J. Miller, M. S. Sigman, J. Am. Chem. Soc. 129, 2752
(2007).
45. Table S2 contains the raw data used to initiate the
modeling process.
46. See supporting material on Science Online.
47. S. N. Deming, S. L. Morgan, Experimental Design: A
Chemometric Approach (Elsevier, New York, ed. 2, 1993).
Acknowledgments: Supported by NSF grant CHE-0749506.
Supporting Online Material
www.sciencemag.org/cgi/content/full/333/6051/1875/DC1
Materials and Methods
Tables S1 to S26
References (48, 49)
14 April 2011; accepted 13 July 2011
10.1126/science.1206997
Experimental psychologists have repeatedly
demonstrated that positive and negative affect are
independent dimensions. Positive affect (PA) in-
cludes enthusiasm, delight, activeness, and alert-
ness, whereas negative affect (NA) includes distress,
fear, anger, guilt, and disgust (6). Thus, low PA
indicates the absence of positive feelings, not the
presence of negative feelings.
Diurnal and Seasonal Mood Vary
with Work, Sleep, and Daylength
Across Diverse Cultures
Scott A. Golder* and Michael W. Macy
Laboratory studies have shown that diurnal
mood swings reflect endogenous circadian rhythms
interacting with the duration of prior wakefulness
We identified individual-level diurnal and seasonal mood rhythms in cultures across the globe,
using data from millions of public Twitter messages. We found that individuals awaken in a good or sleep. The circadian component corresponds
mood that deteriorates as the day progresses—which is consistent with the effects of sleep and
circadian rhythm—and that seasonal change in baseline positive affect varies with change in
to change in core body temperature, which is lowest
at the end of the night and peaks during late
daylength. People are happier on weekends, but the morning peak in positive affect is delayed by 2 afternoon. The sleep-dependent component is
hours, which suggests that people awaken later on weekends.
elevated at waking and declines throughout the
day (7). Other studies have variously observed a
ndividual mood is an affective state that is hormones (e.g., cortisol) (3). Mood is also ex- single PA peak 8 to 10 hours after waking (8), a
important for physical and emotional well- ternally modified by social activity, such as daily
I
being, working memory, creativity, decision- routines of work, commuting, and eating (4, 5).
Department of Sociology, Cornell University, Ithaca, NY
14853, USA.
making (1), and immune response (2). Mood is Because of this complexity, accurate measure-
influenced by levels of dopamine, serotonin, and ment of affective rhythms at the individual level
other neurochemicals (1), as well as by levels of has proven elusive.
*To whom correspondence should be addressed. E-mail:
sag262@cornell.edu
1878
30 SEPTEMBER 2011 VOL 333 SCIENCE www.sciencemag.org