Diastereo- and enantioselective addition of anilide-functionalized allenoates to N-acylimines catalyzed by a pyridylalanine-based peptide

Article abstract of DOI:10.1021/ja412996f

A selective peptide-catalyzed addition of allenic esters to N-acylimines is reported. Tetrasubstituted allenes were achieved with up to 42:1 diastereomeric ratio and 94:6 enantiomeric ratio (up to 99:1 er after recrystallization of the major diastereomer). An exploration of the role of individual amino acids within the peptide was undertaken. The scope of the reaction was explored and revealed heightened reactivity with thioester-containing allenes. A mechanistic framework that may account for the observed reactivity is also described.

Full text of DOI:10.1021/ja412996f

Article
pubs.acs.org/JACS
Diastereo- and Enantioselective Addition of Anilide-Functionalized
Allenoates to N‑Acylimines Catalyzed by a Pyridylalanine-Based
Peptide
Curren T. Mbofana and Scott J. Miller*
Department of Chemistry, Yale University, P.O. Box 208107, New Haven, Connecticut 06520-8107, United States
S
* Supporting Information
ABSTRACT: A selective peptide-catalyzed addition of allenic
esters to N-acylimines is reported. Tetrasubstituted allenes were
achieved with up to 42:1 diastereomeric ratio and 94:6
enantiomeric ratio (up to 99:1 er after recrystallization of the
major diastereomer). An exploration of the role of individual
amino acids within the peptide was undertaken. The scope of
the reaction was explored and revealed heightened reactivity
with thioester-containing allenes. A mechanistic framework that
may account for the observed reactivity is also described.
INTRODUCTION
■
The allene moiety is an important functional group in organic
synthesis and medicine. It is found in a number of natural
products and pharmaceutical compounds.1,2 Moreover, the
unique reactivity of allenes renders them versatile intermediates
in organic synthesis.3−10 Numerous methods for stereoselective
synthesis of di- and trisubstituted allenes have been
developed.11,12 However, access to chiral tetrasubstituted
allenes remains a challenge, and limited examples exist.13−16
Our laboratory has been interested in the use of nitrogen-
based catalysts for additions of nucleophiles,17−19 including
allenoate esters20,21 to various electrophiles such as α,β-
unsaturated carbonyl compounds and N-acylimines. In 2009,
our laboratory disclosed an enantioselective coupling reaction
of allenic esters to imines using a pyridylalanine (Pal)-based
peptide catalyst (eq 1).16 As a nontrivial extension to this work,
we envisioned that the coupling of a racemic allene (1) to an N-
acylimine (2) would deliver a tetrasubstituted allenoate (3).
Such products (3) possess two elements of stereochemistry, a
stereogenic center and an axis of chirality (eq 2). Thus, catalytic
reaction (eq 4).23 Notably, the reaction proceeded with
reactions of this type would allow an examination of the
interplay of the stereogenic center with the axis of chirality
during a complex reaction mechanism. We hoped to control
both diastereo- and enantioselectivity using peptide-based
catalysts. To our knowledge, the only example of a similar
transformation is a report published recently by Maruoka and
co-workers of a diastereo- and enantioselective addition of di-
tert-butyl allenic esters to sulfonyl imines, using phase-transfer
catalysis (eq 3).22
acceptable conversion, despite the increased steric demands
RESULTS AND DISCUSSION
■
Preliminary studies in our laboratory focused on the γ,γ-
substituted allene substrate 1a, employing conditions that had
been previously established for the unsubstituted addition
Received: December 20, 2013
Published: February 14, 2014
© 2014 American Chemical Society
3285
dx.doi.org/10.1021/ja412996f | J. Am. Chem. Soc. 2014, 136, 3285−3292

Journal of the American Chemical Society
Article
a
of the γ,γ-disubstitution in the substrate. With 30 mol% of
catalyst 4a, the reaction proceeded smoothly to deliver 3a in
67% isolated yield as a combined mixture of stereoisomers
within 23 h. However the diastereomeric ratio (dr) was poor
(1.0:1.2), and the dominant diastereomer exhibited modest, yet
appreciable, enantiomeric ratio (er) of 38.5:61.5. Changing
reaction variables such as solvent, temperature, and incorpo-
ration of Lewis acidic additives did not improve dr.24
We hypothesized that employing a substrate that contained
an additional hydrogen bond donor or acceptor might perturb
diastereoselectivity. Our analysis was superficial but hinged on
the idea that peptide catalyst−substrate interactions involving
hydrogen bonds are supported by extensive experimental
evidence.25,26 Switching to the anilide allenic ester 1b provided
an observation consistent with this hypothesis (eq 5). With 20
Table 1. Diastereomeric Peptide Catalyst Screen
b
yield
er (min,
d
c
entry catalyst i + 1
i + 3
X
(%)
dr
maj)
1
4a
D-Pro
Phe
NMe2
85
1.0:2.7 45:55,
39:61
2
4b
D-Pro
D-Phe
OMe
80
1.0:3.0 46:54,
22:78
3
4
5
4c
4d
4e
Pro
Phe
OMe
OMe
NMe2
85
76
71
1.0:1.1 ND
1.0:1.1 ND
Pro
D-Phe
D-Phe
D-Pro
1.0:3.4 39:61,
22:78
a
b
Reactions were run with 1.2 equiv of allenic ester. Total yield of
both diastereomers determined by 1H NMR using an internal
standard. Determined by H NMR. Determined using HPLC with
c
d
1
chiral stationary phase. ND: Not determined.
reaction time of 3 h. For aliphatic groups (Table 2, entries 3−
5) the bulkier isopropyl group gave the highest dr of 1.0:6.2.
Unfortunately, the tert-butyl allenic ester could not be
successfully synthesized. In addition, more electron-rich ester
groups required longer reaction times. For example, with
isopropyl allenic ester the reaction required 8 h to reach
completion (Table 2, entry 5). It is worth noting that although
significant perturbations were observed in terms of the dr and
reaction times, the identity of the ester moiety had only a small
effect on the observed er of the major diastereomer.
The effect of the reaction solvent was also investigated.
Generally, aromatic solvents performed similarly to toluene in
both yield and selectivity (Table 2, entries 6−8), with the
exception of electron-deficient hexafluorobenzene, in which a
decreased dr of 1.0:5.0 and only 48% yield were observed
(Table 2, entry 8). CHCl3 provided 3g in 1.0:4.4 dr, while
DCM and THF revealed lower dr’s of 1.0:2.5 and 1.0:2.2 dr,
respectively (Table 2, entries 9−11). In addition, lower yields
were observed in these solvents, ranging from 55% to 73% in 8
h. Hence, toluene was retained as the solvent of choice.
Next, we set out to investigate the impact of the imine-
protecting group in Table 3. N-Diphenylphosphinoylimine gave
no reaction after 24 h (Table 3, entry 1). Yet, N-carboxyethyl-
imine gave an improved dr of 1.0:9.3 but with a slightly
diminished er of 25:75 (Table 3, entry 2). Switching to the
bulkier Boc group resulted in a further increase in dr to 1.0:12,
although again a lower er was observed (27:73; Table 3, entry
3). Cyclohexylcarbonyl imine 2e reactivity was similar to the
Boc group (Table 3, entry 4). We chose to proceed with Boc
imine due to ease of synthesis.
mol% of catalyst 4a the reaction delivered product 3b in 85%
isolated yield in just 1 h and an improved dr of 1.0:2.7 while
maintaining an er of 39:61 for the major diastereomer. Use of
pyridine as the catalyst showed no diastereoselectivity, and the
reaction was considerably slower, requiring 6 h to reach a
comparable level of conversion. Moreover, replacing the anilide
group with an ester group eroded dr to 1.0:1.2 (eq 6). These
observations support the importance of the anilide group,
possibly as a hydrogen bond participant.
Encouraged by these results, we performed a catalyst screen
to investigate the optimum diastereomer for the peptide
catalyst. In an arbitrary manner, we initially elected to maintain
the identity of the amino acids to allow a direct assessment of
stereochemical issues. The results are shown in Table 1. The
most notable change resulted from switching the i + 3 residue
from Phe to D-Phe (Table 1, entry 2, peptide 4b) leading to a
slight increase in dr to 1.0:3.0 and an improved er of 22:78 for
the major diastereomer. The minor diastereomer er was
marginally decreased. However switching the i + 1 residue
from D-Pro to Pro led to diminished dr (1.0:1.1) regardless of
which Phe enantiomer was at the i + 3 position (Table 1,
entries 3 and 4). Further optimization of peptide 4b by
switching the C-terminal functional group to N,N-dimethyla-
mide slightly improved the dr to 1.0:3.4. The er of the major
diastereomer remained unchanged, while the er of the minor
diastereomer slightly increased to 39:61 (Table 1, entry 5,
peptide 4e).
Encouraged by these results, we examined additional peptide
sequences, and the results are shown in Table 4. For this
peptide series, the relative stereochemistry in the sequence was
kept constant, while the identities of several amino acid side
chains were altered. For the i + 3 position, switching Aib (2-
aminoisobutyric acid) with Acp (1-aminocyclopropane-1-
carboxylic acid) gave a minute change in selectivity (Table 4,
entries 1 and 2). However, Aic (2-aminoindane-2-carboxylic
acid) replacement in peptide 4h resulted in significant er
increase to 20:80 (Table 4, entry 3). Owing to the ready
availability of trans-4-hydroxy-L-proline, a pseudoenantiomeric
Having identified a more robust peptide catalyst 4e, we set
out to investigate the effect of the allenic ester group.
Replacement of the phenyl ester (Table 2, entry 1) with a
benzyl ester resulted in an increase in dr to 1.0:4.3 (Table 2,
entry 2). In addition, the reaction required an extended
3286
dx.doi.org/10.1021/ja412996f | J. Am. Chem. Soc. 2014, 136, 3285−3292

Journal of the American Chemical Society
Article
a
Table 2. Variation to the Allenic Ester Group and Solvent Screen
b
c
d
entry
R1
solvent
toluene
rxn time (h)
Prd
yield (%)
dr
er (maj)
1
OPh, 1b
OBn, 1d
OMe, 1e
OEt, 1f
OiPr, 1g
OiPr, 1g
OiPr, 1g
OiPr, 1g
OiPr, 1g
OiPr, 1g
OiPr, 1g
1
3
8
8
8
8
8
8
8
8
8
3b
3d
3e
3f
71
>95
>95
>95
93
1.0:3.4
1.0:4.3
1.0:5.0
1.0:5.0
1.0:6.2
1.0:6.1
1.0:6.2
1.0:5.0
1.0:4.4
1.0:2.5
1.0:2.2
22:78
22:78
21:79
21:79
22:78
21:79
26:74
ND
2
toluene
3
toluene
4
toluene
5
toluene
3g
3g
3g
3g
3g
3g
3g
6
benzene
92
7
mesitylene
hexafluorobenzene
chloroform
dichloromethane
94
8
48
9
55
34:66
ND
10
11
73
tetrahydrofuran
60
ND
a
b
c
Reactions were run with 1.2 equiv of imine. Total yield of both diastereomers determined by 1H NMR using an internal standard. Determined by
d
1H NMR. Determined using HPLC with chiral stationary phase. ND: Not determined.
a
peptide of 4h, peptide 4i, was synthesized. The result was a
Table 3. Variation to the Imine Protecting Group
remarkable improvement in both dr (1:17) and er (89:11)
(Table 4, entry 4). It is worth noting that although the
preferred enantiomer was of course reversed, the major
diastereomer remained the same. Further changes to 4i in the
i + 3 position by replacing phenylalanine with phenylglycine or
O-benzyltyrosine were not beneficial (Table 4, entries 5 and 6).
Elongating the catalyst sequence to a pentamer did not improve
selectivity either (Table 4, entries 7 and 8). Hence, 4i was
identified as the lead peptide catalyst.
A screen of catalyst loading revealed that conversion
decreased with lower amounts of catalyst. The addition
reaction was carried out with 10, 5, and 2 mol % of peptide
4i, resulting in 79, 62, and 29% of product, respectively, within
24 h (Table 4, entries 9−11). Nonetheless, higher product yield
was achieved with longer reaction times; 88% yield was
achieved with 10 mol % of catalyst after 36 h (Table 4, entry 9),
while 5 mol % of catalyst gave 79% yield in 48 h (Table 4, entry
10). Peptide catalyst 4i was used to examine the substrate scope
b
rxn time
(h)
yield
(%)
er
c
c
entry
PG
prd
dr
(maj)
1
2
3
4
POPh2, 2b
CO2Et, 2c
24
10
8
3h
3i
NR
72
−:−
1.0:9.3
−:−
25:75
27:73
28:72
t
CO2 Bu, 2d
3j
78
1.0:12
d
C(O)cyclohexyl,
24
3k
>95
1.0:12
2e
a
b
Reactions were run with 1.2 equiv of imine. Total yield of both
diastereomers determined by 1H NMR using an internal standard.
c
d
Determined using HPLC with chiral stationary phase. Determined
1
by H NMR. NR: no reaction.
a
Table 4. Peptide Catalyst Reoptimization and Catalyst Loading Screen
b
c
c
entry
catalyst
catalyst (mol %)
yield (%)
dr
er (maj)
1
Boc-Pal-D-Pro-Aib-D-Phe-NMe2, 4f
20
20
20
20
20
20
20
20
10
5
78
75
1.0:12
1.0:11
1.0:13
1.0:17
1.0:12
1.0:15
1.0:11
1.0:12
1.0:16
1.0:18
1.0:15
27:73
26:74
20:80
89:11
85:15
87:13
86:14
83:17
88:12
88:12
89:11
2
Boc-Pal-D-Pro-Acp-D-Phe-NMe2, 4g
3
Boc-Pal-D-Pro-Aic-D-Phe-NMe2, 4h
81
4
Boc-D-Pal-Hyp(OBn)-Aic-Phe-NMe2, 4i
Boc-D-Pal-Hyp(OBn)-Aic-Phg-NMe2, 4j
Boc-D-Pal-Hyp(OBn)-Aic-Tyr(OBn)-NMe2, 4k
Boc-D-Pal-Hyp(OBn)-Aic-Phe-D-Phe-NMe2, 4l
Boc-D-Pal-Hyp(OBn)-Aic-Phe-Phe-NMe2, 4m
Boc-D-Pal-Hyp(OBn)-Aic-Phe-NMe2, 4i
Boc-D-Pal-Hyp(OBn)-Aic-Phe-NMe2, 4i
Boc-D-Pal-Hyp(OBn)-Aic-Phe-NMe2, 4i
92
5
87
6
>95
81
7
8
79
d
9
79 (88)
e
10
11
62 (79)
29
2
a
b
c
1
Reactions were run with 1.2 equiv of imine. Total yield of both diastereomers determined by H NMR using an internal standard. Determined
d
e
using HPLC with chiral stationary phase. Yield after 36 h. Yield after 48 h.
3287
dx.doi.org/10.1021/ja412996f | J. Am. Chem. Soc. 2014, 136, 3285−3292

Journal of the American Chemical Society
Article
a
Table 5. Substrate Scope for Allenic Esters 1 and N-Acylimines 2 To Give Addition Products 3 Using Peptide Catalyst 4i
a
b
c
All data is the average of two experiments run using 1.2 equiv of imine. Total isolated yield of both diastereomers. Total yield of both
d
e
diastereomers determined by 1H NMR using an internal standard. Determined using HPLC with chiral stationary phase. After recrystallization of
major diastereomer. ND: not determined.
while maintaining a 20 mol % catalyst loading due to shorter
reaction times.
Further substrate and reaction condition optimization
revealed that allenic thioester substrates at low temperature
have superior reactivity. Allenic thioester 1i and Cbz-imine 2j at
−20 °C reacted to give product 3r in excellent selectivity
(Table 5, entry 8; 1:25 dr and 94:6 er). As observed with the
allenic oxo-ester reactions, the electron-rich p-methyl sub-
stitution in 2k had little effect in selectivity (Table 5, entry 9).
o-Methyl product 3t (Table 5, entry 10) was obtained in
excellent dr of 1.0:42 while maintaining a 94:6 er. The electron-
deficient p-CF3 product 3u gave slightly diminished selectivity
(Table 5, entry 11). Also in line with earlier observations,
sterically bulky substituents significantly slowed the reaction
and resulted in decreased selectivity. Product 3v was obtained
after 72 h in a diminished 1.0:13 dr and 87.5:12.5 er (Table 5,
entry 12). This effect was even more pronounced for the
bulkier isobutyl allenic ester 1k, which required 96 h at 23 °C
to deliver 58% isolated yield of product 3w in a poor 1.0:5.8 dr
and 75.5:24.5 er (Table 5, entry 13). Nonetheless, it is quite
interesting that even with significantly extended reaction time
the reaction still proceeded cleanly to the addition product.
A catalyst comparison study was undertaken to better
understand the reactivity of allenic oxo- and thioesters under
pyridine and peptide catalysis (eq 7). With peptide catalyst 4i,
As outlined in Table 5, the optimized catalytic system was
successfully applied to a broad substrate scope in good yields
and selectivity. Addition product 3j was obtained in 91%
isolated yield, 1.0:16 dr, and 88:12 er after 24 h (Table 5, entry
1). As evidenced by product 3l (Table 5, entry 2), electron-rich
substitution on the imine aryl ring had a negligible effect on dr
and er. However, electron-deficient substitution diminished
selectivity. p-Bromo product 3m had a dr of 1.0:10 and
85.5:14.5 er (Table 5, entry 3). Furthermore, the pure major
diastereomer was enantioenriched by recrystallization of the
racemate from toluene, leaving enantioenriched 3m (98:2 er) in
the mother liquor with 44% recovery (based on the sample
subjected to recrystallization). The electronic effect was even
more pronounced for the p-nitro product 3n, with a poor dr
(1.0:4.3) and er (20.5:79.5). In addition, there was significant
rate acceleration, and the reaction was complete in 6 h (Table
5, entry 4). Similarly, the 3n major diastereomer was
enantioenriched to 99:1 er by recrystallization from ethyl
acetate. Once again, the racemate crystallized while the mother
liquor was enriched. The addition reaction tolerated the 2-
thienyl heterocycle giving a decent dr of 1.0:8.7 and a
comparable er of 88:12 (Table 5, entry 5). Replacement of
the Boc group with the Cbz group gave a marginally improved
er of 89:11, though the dr was greatly diminished to 1.0:10
(Table 5, entry 6). In addition, this reaction was slightly faster,
reaching completion in 18 h. It emerged that the reaction was
sensitive to steric bulk on the allenic ester substrate. The
reaction with ethyl substrate 1h required 120 h to deliver
product 3q in 70% isolated yield with a moderate selectivity of
1.0:11 dr and 71.5:28.5 er (Table 5, entry 7).
3288
dx.doi.org/10.1021/ja412996f | J. Am. Chem. Soc. 2014, 136, 3285−3292

Journal of the American Chemical Society
Article
the reaction proceeded to allow isolation of the product in 91%
yield and 1.0:16 dr in 24 h at 23 °C with 66:34 and 88:12 er for
the minor and major diastereomers, respectively. On the other
hand, with pyridine as catalyst, only 38% yield was observed
within the same time frame, with a dr of 1.0:3.4 and racemic
products. To reach a comparable yield (79%), the pyridine-
catalyzed reaction required 108 h of reaction time.
Scheme 1. Possible Mechanistic Pathways to Addition
Product 3 under Basic Conditions
Likewise for the thioester reaction (eq 8), at −20 °C the
peptide-catalyzed reaction gave a higher yield of 81% within 15
give product 3 (Scheme 1, path A).22 Under the strongly basic
conditions of the Maruoka study (5 equiv of K2CO3 or 50%
KOHaq), this pathway seems plausible. On the other hand, for
nitrogen-containing bases an alternative pathway involving
nucleophilic catalysis in analogy to the Morita−Baylis−
Hillman-type mechanism could be considered.27−33 Nucleo-
philic addition to the allenic β-carbon would result in
zwitterionic intermediate int-2. Addition of this species to the
imine would ultimately lead to product 3 (Scheme 1, path B). It
seems plausible that this alternative mechanism might operate
in the presence of weaker bases that are strongly nucleophilic,
as we have previously considered.34
h and a 1.0:25 dr with 69:31 and 94:6 er for the minor and
major diastereomers, respectively. Yet, the pyridine-catalyzed
reaction gave only 14% yield in 15 h; perhaps surprisingly, in
this case a high dr of 1.0:11 was noted. The high dr could be
the result of an apparent hydrogen bond present in the crystal
structure of the major diastereomer. Figure 1 shows the X-ray
To further probe the deprotonation mechanism (path A)
versus the addition mechanism (path B), we conducted several
experiments, yielding results consistent with our assertions for
the peptide-catalyzed pathway. Deuterium-exchange studies
were carried out with allenic diester 1c under various basic
conditions with D2O or MeOD-d4 as the source of the
deuteron electrophile, using 1H NMR. Notably, under
Maruoka’s phase-transfer conditions,22 H/D exchange occurred
rapidly revealing 100% deuterium incorporation with K2CO3
(Figure 2a, blue line) and 85% with KOH (Figure 2b, red line)
Figure 1. ORTEP representation of the X-ray crystal structure of the
major diastereomer racemate of 3m showing the relative stereo-
chemistry of an arbitrary enantiomer. Ellipsoids drawn at 30%
probability level.
structure of addition product 3m. In Figure 1, the carbamate
carbonyl oxygen and anilide nitrogen are 3.19 Å apart, possibly
indicative of a hydrogen bond that may develop in the
transition state. If the major diastereomer exhibiting this
favorable interaction is preferred even with an unfunctionalized,
nonpeptidic catalyst, it may be that the peptides amplify the dr
in an as yet unknown manner. We note that the X-ray structure
shown in Figure 1 was derived from crystals obtained from
racemic material of the major diastereomer. Thus, the absolute
configuration of our products remains unassigned and is only
drawn in analogy to that observed in our earlier study.23
We then turned our attention to mechanistic issues,
ultimately in pursuit of an understanding of both bond-forming
steps and the mode of stereoinduction by the chiral catalyst.
We consider two possible mechanistic pathways for the
formation of addition products under basic conditions as
shown in Scheme 1. In one case, proposed by Maruoka and co-
workers, a cumulenolate intermediate of type int-1, resulting
from α-deprotonation of the allenic ester undergoes addition to
Figure 2. Plot of % deuterium incorporation of 1c to D-1c against
time for the following reaction conditions in toluene-d8: (a) K2CO3 (5
equiv), Bu4NBr (1 equiv), MeOD-d4; (b) 50% KOH/D2O, Bu4NBr (1
equiv); (c) peptide 4i (1 equiv), MeOD-d4; (d) pyridine (1 equiv),
MeOD-d4; (e) quinuclidine (1 equiv), MeOD-d4; (f) Et3N (1 equiv),
MeOD-d4; (g) 2,6-lutidine (1 equiv), MeOD-d4.
within 5 min. These results are consistent with the role of the
proposed cumulenolate intermediate int-1a en route to D-1c as
shown in Scheme 2a. However, H/D exchange was much
slower for peptide catalyst 4i (Figure 2c, purple line) and
pyridine (Figure 2d, green line), both of which are relatively
weaker bases, reaching 90% and 86%, respectively, over 23 h.
Yet, other nitrogen bases that are of various orders of
3289
dx.doi.org/10.1021/ja412996f | J. Am. Chem. Soc. 2014, 136, 3285−3292

Journal of the American Chemical Society
Article
Scheme 2. Pathways to Account for the Observed Deuterium
Incorporation Pattern
Table 6. Kinetic Order of Allenoate and Peptide for the
a
Transformation in eq 7
kinetic
order
linear fit correlation power fit exponent (correlation
compd
constant
constant)
1g
4i
1
1
0.999
0.993
1.00 (0.999)
0.966 (0.987)
a
See Supporting Information for reaction details and rate plots at the
various concentrations. Linear and power fits and correlation constants
were determined using the trendline feature in Microsoft Excel.
Table 7. Kinetic Isotope Effect for Allenoate α-Proton for
a
the Transformation in eq 7
b
parameter
average
kH (min‑1)
kD (min‑1)
kH/kD
0.424
0.415
1.02
a
b
See the Supporting Information for reaction rate plots. Average of
three runs.
magnitude stronger as bases than pyridine induced slower H/D
exchange than that of the pyridyl catalysts. Quinuclidine
(Figure 2e, pink line) gave 75%, and Et3N (Figure 2f, aqua line)
gave only 21%, while no deuterium incorporation was observed
for 2,6-lutidine (Figure 2g, orange line) within 23 h.
Correlation of kinetic rate data to thermodynamic basicity
data must be done cautiously.35−37 Nonetheless, the lack of a
rigorous pKa correlation in the deuteration data of Figure 2
supports a possible difference in mechanism. Notably, in scales
of nucleophilicity that take into account both electronic and
steric effects, pyridine emerges a better nucleophile than does
2,6-lutidine38 or Et3N.39,40 Other scales reveal a different
hierarchy.41 Nevertheless, it is plausible that the deuteration
rates are consistent with a possible change in mechanism, from
general base catalysis to nucleophilic catalysis,42 as one switches
from phase transfer to pyridine-induced catalysis.
It is perhaps not surprising that the two mechanisms might
induce H/D exchange at very different rates. Scheme 2b shows
a pathway that could account for H/D exchange under
nucleophilic catalysis. Zwitterionic enolate int-2a may react
with a deuteron to result in intermediate int-2b and, upon
subsequent deprotonation and catalyst elimination, yield D-1c.
Given the highly substituted nature of the allenic substrates, it is
conceivable that addition of the relatively bulky Et3N or
quinuclidine or sterically hindered 2,6-lutidine would lead to
undesirable steric interactions, ultimately slowing down the
reaction significantly when compared to pyridine-based
catalysts.
reaction. When the rates exhibited by 1g (α-H) and D-1g (α-
D) were compared, the kinetic isotope effect, kH/kD, was found
to be 1.02 (Table 7). If deprotonation to form int-1 were the
rate-determining step, a significant primary KIE would be
observed. Taken together, our kinetic studies are consistent
with an addition mechanism (path B) for the peptide-catalyzed
reaction.
Scheme 3 shows a speculative, more detailed mechanistic
hypothesis for the pyridyl-catalyzed addition transformation
Scheme 3. Mechanistic Hypothesis for Peptide-Catalyzed
a
Addition Reaction
To further interrogate the peptide-catalyzed transformation,
we carried out kinetic studies employing allenoate 1g, imine 2d,
and peptide catalyst 4i (i.e., eq 7; see the Supporting
Information for experimental details and rate plots). The
reaction was found to be first order in allenoate and peptide
(Table 6), but the rate dependence on imine was very small at
synthetically relevant concentrations and could be interpreted
as close to zero order and consistent with imine saturation (see
Supporting Information).43 These results parallel the exper-
imental observations with the simple allenoates studied
previously34 and imply that the rate-determining step could
involve allenoate and peptide only. Hence, the formation of
intermediate int-1 or int-2 (Scheme 1) is plausibly the rate-
limiting step.
Perhaps more revealing, however, are kinetic isotope effects
(KIE) studies that support path B for the peptide-catalyzed
a
R2 = methyl, R3 = anilide.
3290
dx.doi.org/10.1021/ja412996f | J. Am. Chem. Soc. 2014, 136, 3285−3292

Journal of the American Chemical Society
Article
involving nucleophilic catalysis. In this mechanism, peptide
catalyst 4i engages in nucleophilic attack on the allenic β-
carbon to form a zwitterionic intermediate I, which may exist as
a variety of isomers, (e.g., Ia and Ib). However, the delocalized
nature of the zwitterion suggests that interconversion among
these isomers is possible and possibly subject to catalyst
control. Notably, at this stage in the reaction coordinate, any
conformational bias resulting from the axial chirality of the
racemic allenic starting material enantiomers is likely lost. Next,
intermediate I attacks the N-acylimine to form the C−C bond
in intermediate II, an event that also sets the stereogenic N-
bearing carbon center. At least four isomers involving the newly
formed stereocenter and the double bond configuration may be
considered (IIa−d), and again, the catalyst would still exert
conformational bias on these isomers. Subsequent α-proton
transfer and catalyst elimination would give the addition
products, and this final event would define the axis of chirality.
Thus, it is possible that the catalyst addition step is rate-
determining, but not the stereochemistry-determining step.34
Another critical aspect of our study concerning stereo-
induction concerns the actual three-dimensional architecture of
the catalyst. Data obtained from 1H NMR experiments (1H−1H
NOESY) (Figure 3a) are consistent with a β-turn secondary
structure as shown in Figure 3b.44,45 (Details may be found in
the Supporting Information.)
1
Figure 4. Catalyst and allenoate H NMR spectra in toluene-d8: (a)
peptide 4i; (b) allenoate 1g and peptide 4i; (c) allenoate 1g; (d)
⧫
allenoate 1g and pyridine; (e) pyridine. ( : allenoate 1g NH peak, ★:
peptide 4i NHPhe peak).
invoked in the intramolecular H-bonding array of the β-hairpin.
Based on this result, it is conceivable that the catalyst β-turn
structure and catalyst−substrate H-bonding interactions may be
maintained simultaneously in the transition state leading to
significant levels of selectivity.
With these experimental observations in hand, we have
considered a working model that is consistent with the
significant selectivity we have observed. Figure 5 shows an
Figure 3. (a) NOE contacts for peptide catalyst 4i. (b) β-Turn
secondary structure consistent with the observed NOE data.
Even with knowledge about the three-dimensional structure
of the catalyst, the exact origin of stereoinduction is not fully
understood. A detailed confirmation of a catalyst−substrate
adduct (e.g., of type Ia or Ib, Scheme 3) was not possible.
However, when probed by 1H NMR (see the Supporting
Information for full details), a comixture of catalyst and
substrate reveals some differences (the stacked spectra are
shown in Figure 4). Notably, when Figure 4c (allenoate 1g) is
compared to Figure 4b (4i + allenoate 1g), a significant
downfield shift of the allenic amide NH resonance (Figure 4⧫),
by approximately 0.2 ppm is observed. However, for Figure 4c
(allenoate 1g) and 4d (1g + pyridine) no shifting of 1g peaks is
detected under these conditions. This observation, which is
consistent with a hydrogen bond between peptide catalyst and
substrate, may portend a multifunctional interaction between
catalyst and substrate in the transition state. As noted above (eq
6), in the absence of the NH in the allenic ester, the observed
selectivity in the couplings is greatly eroded. A putative
interaction of this type could serve to activate the allene,
consistent with the heightened reactivity observed with pyridyl
peptide catalyst compared to pyridine (vide supra, eqs 7 and 8).
It may also be of significance that, as shown in Figure 4c,
only a slight perturbation (0.03 ppm) is observed for the
diagnostic44,45 downfield NHPhe peak (Figure 4★), which is
Figure 5. Speculative model for the addition reaction leading to 3j.
allenoate−catalyst adduct I (Scheme 3) with the peptide
maintaining its β-turn secondary structure (Figure 3b) and the
allenic NH contacting the catalyst Phe-carbonyl oxygen via H-
bonding. (We reiterate that in the absence of this NH in the
allenic ester, we observe low selectivity; eq 6.) Carbon−carbon
bond formation with the imine might then occur with
simultaneous association of the imine electrophile through an
additional H-bond, with an available NH of the catalyst (e.g.,
Aic-NH). This highly ordered, multivalent assembly46 would
place the imine in close proximity to the α-carbon of the
zwitterion, facilitating bond formation. These notions are
consistent with the heightened reactivity observed with pyridyl
peptide catalyst compared to pyridine as catalyst (eqs 7 and 8).
This model also accounts for the observation that peptide
catalyst 4a (Table 1), with the epimeric Phe residue at the i + 3
position, exhibits lower selectivity (Table 1, entries 1 and 2),
since the allenic anilide and the Phe side chain would compete
for the same space in this scenario. Yet it is appropriate to state
that these models remain speculative in the absence of more
definitive studies.
3291
dx.doi.org/10.1021/ja412996f | J. Am. Chem. Soc. 2014, 136, 3285−3292

Journal of the American Chemical Society
Article
(16) Pu, X. T.; Qi, X. B.; Ready, J. M. J. Am. Chem. Soc. 2009, 131,
10364.
(17) Imbriglio, J. E.; Vasbinder, M. M.; Miller, S. J. Org. Lett. 2003, 5,
CONCLUSION
■
In conclusion, we have developed a highly diastereo- and
enantioselective addition of allenic esters to N-acylimines. Our
results include the study of allenoates containing unsymmetrical
carboxylic acid functional groups (ester versus amide) at the
allene termini in these additions for the first time, revealing
mechanistic consequences. These chiral tetrasubstituted allenes
could serve as chiral building blocks for applications in
synthetic chemistry. In addition, the differential substitution
pattern in these products allows for orthogonal functionaliza-
tion. Our data supports nucleophilic catalysis as the mode of
reaction, and while our mechanistic data is still preliminary we
have proposed a hypothetical mechanistic model that is
consistent with the observed reactivity and selectivity.
3741.
(18) Aroyan, C. E.; Vasbinder, M. M.; Miller, S. J. Org. Lett. 2005, 7,
3849.
(19) Vasbinder, M. M.; Imbriglio, J. E.; Miller, S. J. Tetrahedron 2006,
62, 11450.
(20) Evans, C. A.; Miller, S. J. J. Am. Chem. Soc. 2003, 125, 12394.
(21) Evans, C. A.; Cowen, B. J.; Miller, S. J. Tetrahedron 2005, 61,
6309.
(22) Hashimoto, T.; Sakata, K.; Tamakuni, F.; Dutton, M. J.;
Maruoka, K. Nat. Chem. 2013, 5, 240.
(23) Cowen, B. J.; Saunders, L. B.; Miller, S. J. J. Am. Chem. Soc.
2009, 131, 6105.
(24) Saunders, L. B. Doctoral Thesis, Yale University, 2012.
(25) Jakobsche, C. E.; Peris, G.; Miller, S. J. Angew. Chem., Int. Ed.
2008, 47, 6707.
(26) Vasbinder, M. M.; Jarvo, E. R.; Miller, S. J. Angew. Chem., Int. Ed.
2001, 40, 2824.
(27) Price, K. E.; Broadwater, S. J.; Walker, B. J.; McQuade, D. T. J.
Org. Chem. 2005, 70, 3980.
(28) Price, K. E.; Broadwater, S. J.; Jung, H. M.; McQuade, D. T. Org.
Lett. 2005, 7, 147.
ASSOCIATED CONTENT
■
S
* Supporting Information
Full experimental details, characterization, and X-ray crystallo-
graphic data. This material is available free of charge via the
Internet at http://pubs.acs.org.
(29) Robiette, R.; Aggarwal, V. K.; Harvey, J. N. J. Am. Chem. Soc.
2007, 129, 15513.
(30) Aggarwal, V. K.; Fulford, S. Y.; Lloyd-Jones, G. C. Angew. Chem.,
Int. Ed. 2005, 44, 1706.
(31) Wenzel, A. G.; Jacobsen, E. N. J. Am. Chem. Soc. 2002, 124,
12964.
(32) Raheem, I. T.; Jacobsen, E. N. Adv. Synth. Catal. 2005, 347,
1701.
(33) Buskens, P.; Klankermayer, J.; Leitner, W. J. Am. Chem. Soc.
2005, 127, 16762.
(34) Saunders, L. B.; Cowen, B. J.; Miller, S. J. Org. Lett. 2010, 12,
4800.
(35) Hall, H. K.; Bates, R. B. Tetrahedron Lett. 2012, 53, 1830.
(36) Brotzel, F.; Kempf, B.; Singer, T.; Zipse, H.; Mayr, H. Chem.
Eur. J. 2007, 13, 336.
AUTHOR INFORMATION
■
Corresponding Author
scott.miller@yale.edu
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We are grateful to the National Science Foundation (CHE-
0848224) for financial support and to Louise Guard for X-ray
crystallography. Partial support from the Israel−US Binational
Science Foundation is also acknowledged. All experiments
involving allenic substrate 1a were conducted by Megan
Brennan and Lindsey B. Saunders.
(37) Ammer, J.; Baidya, M.; Kobayashi, S.; Mayr, H. J. Phys. Org.
Chem. 2010, 23, 1029.
(38) Odiaka, T. I.; Kane-Maguire, L. A. P. J. Chem. Soc., Dalton Trans.
1981, 1162.
REFERENCES
■
(1) Hoffmann-Roder, A.; Krause, N. Angew. Chem., Int. Ed. 2004, 43,
1196.
(2) Krause, N.; Hoffmann-Roder, A. In Modern Allene Chemistry;
Krause, N., Hoffmann-Roder, A., Eds.; Wiley-VCH: Weinheim, 2004;
(39) Pearson, R. G.; Sobel, H. R.; Songstad, J. J. Am. Chem. Soc. 1968,
90, 319.
(40) Bruckner, R. Organic Mechanisms Reactions, Stereochemistry and
Synthesis; Springer Verlag: Berlin, 2010.
(41) Wei, Y.; Sastry, G. N.; Zipse, H. J. Am. Chem. Soc. 2008, 130,
3473.
(42) Denmark, S. E.; Beutner, G. L. Angew. Chem. Int. Ed 2008, 47,
1560.
(43) Anslyn, E. V.; Dougherty, D. A. Modern Physical Organic
Chemistry; University Science: Sausalito, CA, 2006.
(44) Haque, T. S.; Little, J. C.; Gellman, S. H. J. Am. Chem. Soc. 1996,
118, 6975.
̈
̈
p 997.
(3) Yu, S. C.; Ma, S. M. Angew. Chem., Int. Ed. 2012, 51, 3074.
(4) Brummond, K. M.; Chen, H. In Modern Allene Chemistry; Krause,
N., Hoffmann-Roder, A., Eds.; Wiley-VCH: Weinheim, 2004; p 1041.
̈
(5) Tejedor, D.; Mendez-Abt, G.; Cotos, L.; Garcia-Tellado, F. Chem.
Soc. Rev. 2013, 42, 458.
(6) Brasholz, M.; Reissig, H. U.; Zimmer, R. Acc. Chem. Res. 2009, 42,
45.
(7) Aubert, C.; Fensterbank, L.; Garcia, P.; Malacria, M.; Simonneau,
A. Chem. Rev. 2011, 111, 1954.
(45) Rose, G. D.; Gierasch, L. M.; Smith, J. A. Adv. Protein Chem.
1985, 37, 1.
(46) Knowles, R. R.; Jacobsen, E. N. Proc. Natl. Acad. Sci. U.S.A.
2010, 107, 20678.
(8) Bates, R. W.; Satcharoen, V. Chem. Soc. Rev. 2002, 31, 12.
(9) Cowen, B. J.; Miller, S. J. Chem. Soc. Rev. 2009, 38, 3102.
(10) Na, R. S.; Jing, C. F.; Xu, Q. H.; Jiang, H.; Wu, X.; Shi, J. Y.;
Zhong, J. C.; Wang, M.; Benitez, D.; Tkatchouk, E.; Goddard, W. A.;
Guo, H. C.; Kwon, O. Y. J. Am. Chem. Soc. 2011, 133, 13337.
(11) Ohno, H.; Nagaoka, Y.; Tomioka, K. In Modern Allene
Chemistry; Krause, N., Hoffmann-Roder, A., Eds.; Wiley-VCH:
̈
Weinheim, 2004; p 141.
(12) Ogasawara, M. Tetrahedron: Asymmetry 2009, 20, 259.
(13) Li, Z. J.; Boyarskikh, V.; Hansen, J. H.; Autschbach, J.; Musaev,
D. G.; Davies, H. M. L. J. Am. Chem. Soc. 2012, 134, 15497.
(14) Hayashi, T.; Tokunaga, N.; Inoue, K. Org. Lett. 2004, 6, 305.
(15) Hammel, M.; Deska, J. Synthesis 2012, 44, 3789.
3292
dx.doi.org/10.1021/ja412996f | J. Am. Chem. Soc. 2014, 136, 3285−3292