Selectivity in the addition reactions of organometallic reagents to aziridine-2-carboxaldehydes: The effects of protecting groups and substitution patterns

Article abstract of DOI:10.1002/chem.201101168

Good to excellent stereoselectivity has been found in the addition reactions of Grignard and organozinc reagents to N-protected aziridine-2-carboxaldehydes. Specifically, high syn selectivity was obtained with benzyl-protected cis, tert-butyloxycarbonyl-protected trans, and tosyl-protected 2,3-disubstituted aziridine-2-carboxaldehydes. Furthermore, rate and selectivity effects of ring substituents, temperature, solvent, and Lewis acid and base modifiers were studied. The diastereomeric preference of addition is dominated by the substrate aziridines' substitution pattern and especially the electronic character and conformational preferences of the nitrogen protecting groups. To help rationalize the observed stereochemical outcomes, conformational and electronic structural analyses of a series of model systems representing the various substitution patterns have been explored by density functional calculations at the B3LYP/6-31G* level of theory with the SM8 solvation model to account for solvent effects. Grignard and bear it: Good to excellent stereoselectivity is reported for the addition reactions of Grignard and organozinc reagents to N-protected aziridine-2-carboxaldehydes (see scheme). High syn selectivity was obtained with benzyl-protected cis, tert-butyloxycarbonyl-protected trans, and tosyl-protected 2,3-disubstituted aziridine-2-carboxaldehydes. Copyright

Full text of DOI:10.1002/chem.201101168

DOI: 10.1002/chem.201101168
Selectivity in the Addition Reactions of Organometallic Reagents to
Aziridine-2-carboxaldehydes: The Effects of Protecting Groups and
Substitution Patterns
Aman Kulshrestha,[a] Jennifer M. Schomaker,[b] Daniel Holmes,[a] Richard J. Staples,[a]
James E. Jackson,*[a] and Babak Borhan*[a]
Abstract: Good to excellent stereo-
selectivity has been found in the addi-
and base modifiers were studied. The
diastereomeric preference of addition
is dominated by the substrate aziri-
dinesꢀ substitution pattern and espe-
cially the electronic character and con-
formational preferences of the nitrogen
protecting groups. To help rationalize
the observed stereochemical outcomes,
conformational and electronic structur-
al analyses of a series of model systems
representing the various substitution
patterns have been explored by density
functional calculations at the B3LYP/6-
31G* level of theory with the SM8 sol-
vation model to account for solvent ef-
fects.
ACHTUNGTRENNUNG
tion reactions of Grignard and organo-
zinc reagents to N-protected aziridine-
2-carboxaldehydes. Specifically, high
syn selectivity was obtained with
benzyl-protected cis, tert-butyloxycar-
bonyl-protected trans, and tosyl-pro-
tected 2,3-disubstituted aziridine-2-car-
boxaldehydes. Furthermore, rate and
selectivity effects of ring substituents,
temperature, solvent, and Lewis acid
Keywords: aziridines
· conforma-
tion analysis · density functional
calculations · Grignard reaction ·
invertomers
·
stereochemistry
·
substituent effects
Introduction
Aziridines are important subunits in chiral auxiliaries, phar-
maceutical intermediates, and biologically active natural
products.[1] Available through several recently developed
stereo- and enantioselective syntheses,[1c,2] these compact
and robust heterocycles are endowed with controllable re-
Scheme 1. Ts-protected aziridine carboxaldehydes serve as precursors for
the tandem aza-Payne/hydroamination methodology.
ACHTUNGTRENNUNGactivity due to their three-membered-ring strain and their
amine siteꢀs acid- and substituent-governed activation. Thus,
through modern regio- and stereoselective ring-opening
methods, aziridine building blocks offer efficient, versatile
access to structurally complex targets. For example, we have
recently reported that N-tosyl-2,3-disubstituted aziridine-2-
carboxaldehydes (1) undergo addition reactions with
Grignard reagents to form syn adducts 2 (Ri =CH3, Rt =
alkyl, Rc =H) with excellent diastereoselectivity (>99:1;
Scheme 1). The use of these products as starting materials
enabled our discovery of a one-pot tandem aza-Payne/hy-
droamination reaction, a process that yields highly function-
alized, densely substituted pyrrolidines 3 (Scheme 1).[3] No-
tably, it is the newly generated carbinol stereocenter that is
the key to the success of these reactions.
This report explores the elaboration of aziridine-2-carbox-
aldehydes through nucleophilic carbonyl addition reactions
that install a stereocenter adjacent to and guided by the
rigid, stereochemically defined aziridine ring. With three
consecutive stereocenters, the resulting aziridinyl carbinols
are poised for application in a wide range of synthetic trans-
formations.[3–4] Such chemistry has proven useful with epox-
ides, the oxygen-containing analogues of aziridines; the ad-
dition reactions of organometallic reagents to a,b-epoxy al-
dehydes have been extensively studied and exploited in the
syntheses of natural products.[5] These additions (Scheme 2)
form a new stereogenic center, giving rise to either syn or
anti adducts depending on the choice of metal. Those that
are strong coordinators favor syn selectivity, which can be
rationalized by a chelation-based transition-state model.[5b]
[a] Dr. A. Kulshrestha, Dr. D. Holmes, Dr. R. J. Staples,
Prof. J. E. Jackson, Prof. B. Borhan
Department of Chemistry, Michigan State University
East Lansing, MI 48824 (USA)
Fax : (+1)517-353-1793
E-mail: jackson@chemistry.msu.edu
babak@chemistry.msu.edu
[b] J. M. Schomaker
Department of Chemistry, University of Wisconsin
1101 University Avenue, Madison, WI 53706 (USA)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201101168.
12326
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12326 – 12339

FULL PAPER
reported excellent syn selectivity with trans substrates, a
result they attributed to N-centered chelation control (Sche-
me 3b).[4b] This selectivity was completely lost in the case of
cis substrates, the cyclic chelate intermediates of which the
authors judged to be unfavorably strained. Contrarily, in a
related study of benzyl (Bn)-protected aziridine carboxalde-
hydes,[4a] Andres and co-workers observed high syn selectivi-
ty specifically in the case of cis substrates and invoked che-
lation control to justify the selectivity. These two reports
appear contradictory; more precisely, taken together, they
suggest that chelation control and the resulting product se-
lectivities are a function of both N protecting group and
substitution pattern.
Noting the limited number of studies on stereoselective
additions of carbon nucleophiles to aziridine-2-carboxalde-
hydes, and intrigued by the widely varying diastereoselectiv-
ities in our own studies of N-tosyl (Ts)-protected systems
(Scheme 1), we recognized the need for a set of principles
to rationalize observations and guide synthetic decisions. We
therefore undertook a combined theoretical and experimen-
tal study of nucleophilic reagent additions to cis- and trans-
substituted aziridine-2-carboxaldehydes with various sub-
stituents on nitrogen.
Scheme 2. The addition of organometallic reagents to epoxy aldehydes.
Rc, Rt, and Ri represent substituents that are cis, trans, or ipso to the al-
dehyde, respectively. Ra denotes the nucleophile added to the carbonyl.
On the other hand, metals that coordinate poorly or condi-
tions that suppress chelation favor the anti adduct predicted
by the Felkin–Anh model.[6] The substitution pattern in the
starting material also influences the syn/anti selectivity, the
effects of which are understood for a,b-epoxy aldehydes.[5b]
Aziridine-2-carboxaldehydes, however, gain additional elec-
tronic and steric degrees of freedom from the nitrogenꢀs pyr-
amidal geometry and substituent variability.
A limited number of other reports have explored the re-
actions of organometallic reagents with aziridine carbonyl
species. Addition of boron enolates to N-tert-butyloxycar-
bonyl (N-Boc) trans-substituted aziridine-2-carboxaldehydes
proceeds with high anti selectivity.[4c] Reductions of aziridine
ketones with various reagents have led to similar results,
that is, the nucleophile attacks by the pro-anti approach.[7]
In both cases, the stereochemical outcomes were rational-
ized in terms of Felkin–Anh transition states for attack on
the carbonyl site (Scheme 3a). It should be noted, however,
that the latter precedent is not universally applicable; the
pro-syn approach can dominate under some conditions.[8]
Investigating the addition of Grignard reagents to N-Boc-
protected aziridine-2-carboxaldehydes, Righi and co-workers
To establish a common reference framework, an initial set
of detailed quantum chemical studies were carried out on
truncated model systems. By exploring the ground-state con-
formational preferences of the various substitution patterns,
our aim was to develop not only useful insights, but also a
simple procedure to routinely predict reactionsꢀ stereochem-
ical outcomes.
To map the effects of electronic variations at nitrogen on
reaction selectivity, we chose the common nitrogen protect-
ing groups: alkyl, alkoxycarbonyl, and alkylsulfonyl. In the
abbreviated models used to simplify the theoretical analysis,
these were methyl (Me), methoxycarbonyl (Moc), and meth-
anesulfonyl (Ms) groups; the corresponding experimental
systems employed Bn, Boc and Ts groups. A simple alkyl or
Bn group leaves the nitrogen comparatively electron rich
and amine-like; Bocꢀs moderate ability as a p-acceptor and
strong conformational preferences offer intermediate activa-
tion; Ts is the most electron-withdrawing group but is less
conformationally defined than Boc. The results of our study
provide intuitive insight and a working model for controlling
the selective addition of organometallic reagents to aziri-
dine-2-carboxaldehydes. The unique combinations of nitro-
gen protecting groups and aziridine substitution patterns
that yield the highest stereoselectivity are identified; inter-
estingly, these are different for each of the protecting
groups.
Results and Discussion
To develop an understanding of the relevant conformational
preferences and to set the stage for analysis of experimental
results, fifteen model substituted aziridine-2-carboxalde-
hydes were studied by quantum chemical modeling. Specifi-
Scheme 3. a) The Felkin–Anh model. b) The Chelation-control model.
The N-bound substituent is denoted RN; see Scheme 2 for details of the
other labels.
Chem. Eur. J. 2011, 17, 12326 – 12339
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12327

B. Borhan, J. E. Jackson et al.
cally, 1) gas-phase DFT simulations of substratesꢀ ground-
state conformations were run in the hope that the prefer-
ACHTUNGTRENNUNGences uncovered would offer simple explanations for the ex-
perimental stereochemical outcomes; 2) solvation effects of
the CH2Cl2 reaction medium were simulated through single
point SM8 calculations on the minima computed; 3) the val-
idity of the gas-phase DFT energetic analyses was verified
through comparison of a subset of the systems studied with
next lowest energy conformation of a different category.
Two binary choices define the four categories of conforma-
tions listed; these are 1) the relationship (anti or syn) be-
tween the N substituent (RN) and the aldehyde moiety; and
2) the rotamer orientation of the aldehyde carbonyl oxygen,
that is, exo, with the hydrogen over the ring, or endo, with
the oxygen over the ring. This choice deserves comment
here; the conjugation of carbonyl groups with adjacent cy-
clopropyl rings in the bisected endo (s-cis) and exo (s-trans)
conformations has long been known (Figure 1a).[14] Like cy-
clopropane, the aziridine ring favors bisected conformations
for attached carbonyl groups, leading to this simple exo/
endo choice. This conjugative preference is surprisingly in-
sensitive to the replacement of carbon with a heteroatom, in
terms of preferred rotameric angle.[15] The tHiCCHCHO (for
which Hi is analogous to Ri in Table 1) dihedral angle values
deviate only by ꢁ158 relative to the 08 and 1808 seen in cy-
clopropane carboxaldehyde. Nucleophilic addition reactions
to cyclopropyl carbonyl compounds have also been system-
atically investigated[16] and often show selectivity governed
by the more stable bisected conformation.[16a,17]
The aforementioned conformational choices are coded A
versus S (anti vs. syn), and X versus N (exo vs. endo), re-
spectively, leading to the four categories AX, AN, SX, and
SN. If more than four conformers exist due to RN rotation,
the energies listed in the table represent the lowest energy
conformer in each category.
G3ACHTUNGTRENNUNG(MP2) composite thermochemistry calculations; and
4) selected potential-energy functions for aldehyde rotation
and nitrogen inversion were also explored in several model
aziridine-2-carboxaldehydes.
All calculations were performed with the Spartan 08[9]
software package, by using the B3LYP functional,[10] the 6-
31G* basis set,[11] and the SM8 solvation modeling
scheme[12] to account for the effects of the CH2Cl2 solvent
environment. To check the performance of this DFT-based
method, the G3ACHTUNGTRENNUNG(MP2) composite scheme was used; this
method offers near-experimental accuracy for the heats of
formation and reaction energetics of small gas-phase organic
molecules.[13]
Conformational analysis of ground state aziridine-2-carbox-
aldehydes: The model systems considered are shown in
Table 1, together with energetic information on their pre-
ferred conformations and the energy gap (in kcalmolꢀ1) that
separates each compoundꢀs lowest energy minimum from its
As noted above, the model
RN groups CH3, COOCH3, and
SO2CH3 were chosen to simu-
Table 1. Conformer energies of modeled aziridine-2-carboxaldehydes.
late Bn, Boc, and Ts substitu-
ents, respectively. Even without
inspecting the full conformer
list, including RN rotamers, it is
immediately evident that rela-
tively few of the model sub-
strates considered here have a
strong structural bias toward a
particular conformation. How-
ever, it is noteworthy that the
structures calculated to be most
conformationally defined corre-
spond to analogues with strong
experimental stereoselectivities.
Also, in cases for which AN is
low enough in relative energy,
this conformation allows for
chelation to exert further di-
recting effects with Grignard
and organozinc nucleophiles.
Before delving into the sig-
nificance of the conformational
findings in Table 1, several
checks on the appropriateness
of this level of calculation
should be discussed. The
1.42 kcalmolꢀ1 AX preference
Name
RN
Ri
Rt
Rc Lowest E[a]
DE[a] to
2nd lowest
AX[b] SX[b] AN[b] SN[b]
conformer 2nd lowest E[a] conformer
conformer
Me-H3
CH3
CH3
CH3
CH3
CH3
H
CH3
H
H
H
CH3
H
H
H
H
AX
AX
SN
1.42
0.44
0.30
0.48
0.19
0.53
0.01
1.20
0.21
0.23
1.30
0.18
0.50
0.42
3.14
AN
SX
AX
AN
SN
AX
SN
SX
0.00 2.60 1.42 2.53
0.00 0.44 2.59 0.95
0.30 0.31 1.66 0.00
0.00 5.47 0.48 4.03
1.80 0.00 4.22 0.19
0.53 0.93 1.76 0.00
0.28 0.00 2.37 0.01
1.32 1.20 2.32 0.00
0.00 1.77 0.21 0.26
1.97 0.23 3.69 0.00
0.00 2.06 1.30 4.72
0.18 0.00 2.23 3.35
0.50 0.00 1.37 2.45
0.00 5.84 0.42 6.66
4.03 0.00 5.95 3.14
Me-ipso
Me-trans
Me-cis
Me-2,3Di
Moc-H3
Moc-ipso CO2CH3 CH3
Moc-trans CO2CH3
Moc-cis CO2CH3
Moc-2,3Di CO2CH3 CH3 CH3
Ms-H3
Ms-ipso
Ms-trans
Ms-cis
H
CH3
H
AX
SX
CH3 CH3
H
CO2CH3
H
H
CH3
H
H
SN
H
H
CH3
H
SX
H
H
SN[c]
AX
SN
AN
SX
SO2CH3
SO2CH3 CH3
SO2CH3
SO2CH3
SO2CH3 CH3 CH3
H
H
H
CH3
H
H
H
AX
SX
AN
AX
AX
AN
SN
H
H
H
SX
CH3
H
AX
SX[d]
Ms-2,3Di
[a] All energies shown in kcalmolꢀ1; [b] AX, SX, AN, and SN label the conformers as anti or syn (relationship
of RN to CHO) and exo or endo (orientation of the CHO group relative to the aziridine ring); [c] An SN con-
formation like that here in Moc-trans is seen in the X-ray crystal structure of the related trans N-Boc analogue
8 (see below); [d] An SX conformation like that in Ms-2,3Di is seen in the X-ray crystal structure of related
trans N-Ts-2,3-disubstituted 2-carboxaldehyde analogue 15.
12328
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12326 – 12339

Selectivity in the Addition Reactions of Organometallic Reagents
FULL PAPER
conformations must first be made. Then directing effects
due to the roles of chelation and sterically interfering sub-
stituents can be explored.
Carbonyl conformation: endo versus exo: Focusing on the
steric demands of a carbon nucleophile alone, it might be
expected that an endo-oriented carboxaldehyde conformer
should be more reactive than an exo one, as attack along
the Bꢂrgi–Dunitz angle would be likely to encounter more
interference approaching the exo conformer. However, an
organometallic compound, such as a Grignard or organozinc
reagent, likely follows a less open trajectory as its metal
component coordinates to the increasingly negative carbonyl
oxygen. Nonetheless, the assumption is supported by
B3LYP/6-31+G* calculations on the transition states for di-
methylzinc attacking cyclopropane carboxaldehydes (Fig-
ure 1b). Although the substrateꢀs ground state conformation
is found to prefer the exo geometry by 0.1–0.2 kcalmolꢀ1
(0.27 experimental),[20] the endo attack transition state (TS)
is favored over the exo by 1.7 kcalmolꢀ1 (1.5 with vibration-
al, thermal, and solvation corrections), which is enough to
create a >10:1 preference for the endo path at relevant re-
action temperatures. Although further skewed by the aziri-
dine nitrogenꢀs own chelation and conformational dynamics,
this steric bias towards attack on endo aldehyde conforma-
tions remains true for aziridine-2-carboxaldehydes in which
the substituent ipso to the aldehyde is hydrogen.
As is evident from Table 1, replacement of the ipso hydro-
gen with an alkyl group in the Ri site has two notable effects
on ground-state conformations: 1) it adds steric bulk on the
trans face which, in turn, decreases the anti–syn energy gap
by shifting the anti form up in energy relative to the syn in-
vertomers at nitrogen regardless of the nature of RN; and
2) it adds a stereoelectronic preference for the exo over the
endo aldehyde rotamer. Although the origin of the latter
preference is not analyzed here, the added invertomer bias
can be as large as 2–3 kcalmolꢀ1, whereas the aldehyde rota-
mer preference is in the 0.5–1 kcalmolꢀ1 range. To illustrate,
comparison of the Me-H3 and Me-ipso models, both of
which have AX lowest energy conformers, is instructive:
1) The energy changes for N-CH3 inversion from syn to anti
are shifted to favor the syn isomer upon adding the methyl
group, that is, on going from Me-H3 (DESX!AX =ꢀ2.60 and
DESN!AN =ꢀ1.11) to Me-ipso (DESX!AX =ꢀ0.44 and DESN!
AN = +1.64). Thus, the ipso CH3 group adds 2.16 and
2.75 kcalmolꢀ1, respectively, to the steric cost for the N-CH3
to be on the trans face, balancing (but not completely over-
coming) the preference for the anti forms. Analogous num-
bers for the Ts model (Ms series) are 2.24 and 2.30 kcal
molꢀ1, which are quite similar to those seen for the Me set;
in this case the anti preference is overcome, moving SX to
the lowest energy spot. Interestingly, the Moc series is only
shifted by 0.68 and 0.60 kcalmolꢀ1, which is consistent with
Mocꢀs strong preference (due to N conjugation) for orient-
ing its thinnest dimension towards vicinal substituents.
2) Comparisons analogous to those above, but comparing al-
dehyde rotamers are shifted as follows: DEAN!AX =ꢀ1.42
Figure 1. a) Bisected exo (s-trans) and endo (s-cis) conformations of cy-
clopropyl carbonyl systems. b) B3LYP/6-31+G*-optimized TS structures
(relative energies with vibrational, thermal, and solvent corrections) for
attack of (CH3)2Zn on endo and exo conformations of cyclopropane car-
boxaldehyde.
computed for Me-H3 is consistent with the conformation as-
signed to N-tert-butylaziridine-2-carboxaldehyde by Pierre
et al. on the basis of variations in the H-C-C(=O)-H
1H NMR coupling constant with temperature and solvent
polarity.[18] Likewise, the small energy differences calculated
for Me-trans conformations are consistent both with the in-
vertomer distribution of 2-acetyl-N,3-dimethyl aziridine ob-
served previously, and with our own findings for the closely
related trans-N-benzyl-3-ethylaziridine-2-carboxaldehyde 7
discussed below.[19] As noted in Table 1, X-ray crystal struc-
tures of N-Boc-trans-3-cyclohexylACTHNUGTRNEUNGaziridine-2-carboxaldehyde
8 and N-Ts-trans-3-benzyloxymethyl-2-methylaziridin-2-car-
boxaldehyde 15, the experimental analogues of Moc-trans
and Ms-2,3Di, confirm the calculated minimum energy con-
formational assignments. Broader support is found in the
G3ACHTUNGTRENNUNG(MP2) calculations on the N-methyl series of model sub-
strates. As noted above, this level of calculation is designed
to provide energetics comparable to experiment in quality.
Table 2 compares their “CH2Cl2 solvated” enthalpies as cal-
culated at the DFT and G3
fect match, the energy variations seen in the DFT results
generally mirror the G3(MP2) data.
ACHTUNGTRNE(NUNG MP2) levels. Although not a per-
AHCTUNGTRENNUNG
To translate the conformational preferences computed
above into predictions regarding the stereochemistry of nu-
cleophilic attack, some estimate of the intrinsic reactivities
and stereoselectivities of the exo versus endo CHO rotamer
Table 2. Comparison of the relative energies of B3LYP/6-31G*/SM8
CH2Cl2 (DFT)and G3ACHTUNGTRENNUNG
AX
DFT
AN
DFT
SN
DFT
G3
DFT
G3
G3
G3
Me-H3
0.00
0.00
0.30
0.00
1.80
0.00
0.00
0.00
0.00
1.32
2.60
0.44
0.31
5.47
0.00
2.98
0.82
0.26
5.56
0.00
1.42
2.59
1.66
0.48
4.22
1.79
2.87
1.80
0.98
4.08
2.53
0.95
0.00
4.03
0.19
3.54
1.89
0.75
5.00
1.01
Me-ipso
Me-trans
Me-cis
Me-2,3Di
Chem. Eur. J. 2011, 17, 12326 – 12339
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12329

B. Borhan, J. E. Jackson et al.
and DESN!SX = +0.07 kcalmolꢀ1 for Me-H3 versus DEAN!
AX =ꢀ2.59 and DESN!SX =ꢀ0.51 kcalmolꢀ1 for Me-ipso.
Thus, the ipso CH3 group favors the exo rotamer by 1.17
and 0.58 kcalmolꢀ1, respectively; analogous pairs of values
for the Moc and Ms systems are 0.86/0.94 and 0.75/0.69 kcal
molꢀ1.
But what of the transition states? How does addition of
an ipso alkyl group affect the relative energetics of endo
versus exo attack? As illustrated in Figure 1b, attack on the
endo rotamer is favored over the exo by 1.5 kcalmolꢀ1 if the
ipso substituent is a hydrogen atom. But just as methyl sub-
stitution adds a bias favoring exo aldehyde rotamers in the
ground-state structures, it also stabilizes the exo relative to
the endo attack TS, leaving the two paths essentially equal
in energy. Thus, ipso substitution is expected to both en-
hance the proportion of exo conformations and lower the
barrier to attack on such forms, favoring product formation
through the AX and SX forms. Herein, we will refer to the
latter phenomenon as the ꢃipso effectꢀ during our discus-
sions.
conformations represents a substantial extra energy cost for
pro-anti attack. This finding concurs with the previously re-
ported experiments and analysis by Pierre et al., in which
CH3Li and PhLi addition to N-tert-butyl-aziridine-2-carbox-
aldehyde showed a strong preference for pro-syn attack, de-
spite the absence of any perturbing cis substituents.[8] As the
experiments below illustrate, specific chelation effects are
easily revealed by simple competition with multidentate li-
gands, such as tetramethylethylenediamine (TMEDA).
Steric effects of cis substituents: It seems clear that a cis
substituent (either the group on C3 or the nitrogen inverto-
mer with the protecting group syn to the aldehyde) should
hinder the approach of nucleophiles to the blocked face of
the aldehyde carbonyl. However, such steric reasoning does
not include the effects of polarity in the N protecting
groups, which may interact with the carbonyl groupꢀs dipole,
orienting it and potentially modifying its reactivity. Such ef-
fects have been clearly observed in infrared studies focusing
on the carbonyl groupꢀs resonance as a probe of conforma-
tional preferences.[21] Similarly, in the case of the Moc sub-
stituent in Moc-trans, the favorable electrostatic interaction
between the aldehydeꢀs carbonyl oxygen and the Moc
groupꢀs electrophilic center plays a role akin to metal chela-
tion of the aziridine nitrogen, inducing a substantial prefer-
ence for the strongly pro-syn-selective SN conformation.
The ꢃelectronically chelatedꢀ Moc/aldehyde group (see
Figure 6 below for an illustration of the aforementioned in-
teraction) places the Moc substituent syn to the aldehyde,
which is in contrast to the benzyl protected aziridines that
necessarily have the protecting group anti to the aldehyde as
a result of the chelating arrangement of the nitrogen lone
pair and the carbonyl group. This orientation effect is also
seen in a related X-ray crystal structure of Boc-protected
aziridine 8 described below. In the analogous framework,
but with the negatively polarized Ts group in Ts-trans, the
aldehyde carbonyl orients away, weakly favoring the SX
conformation. If that conformation is more strongly pre-
ferred, as in Ts-2,3Di, for which the ipso substituent exerts
extra exo preference, the SX conformation dominates, the
Ts group effectively blocks approach to the pro-anti face,
and the resulting pro-syn addition yields exclusively syn
products.
Chelation effects: Chelation of the organometallicꢀs metal
center may drastically alter the conformational energy land-
scape by selectively favoring the AN conformation (the only
conformation of the four that is able to support chelation
between the nitrogen and the aldehyde) provided that AN is
not too much higher in energy than the other rotamers.
Meanwhile, the carbonyl polarization due to Lewis acid
complexation further activates the carbonyl in this already
reactive endo rotamer for pro-syn attack (Figure 2). Because
of the intimate involvement of solvent and reagent aggrega-
tion, it is difficult for theory to directly compare energies of
chelating and non-chelating paths on an equal footing. How-
ever, between the competing pro-syn and pro-anti chelating
paths calculated for dimethyl zinc attack on Me-H3 in
CH2Cl2 (Figure 2), the pro-syn route is substantially favored,
as its TS is closer to the ground state AN conformation with
the CHO moiety bisecting the aziridine ring. In contrast, the
CHO group in the pro-anti TS appears nearly perpendicular
to either of the ground state bisecting conformers of the al-
ꢀ
dehyde. Noting that the aziridine CHO rotational barrier in
isolation is calculated to be 6 kcalmolꢀ1, this difference in
Bn-protected aziridine-2-carboxaldehydes: Addition of eth-
ylmagnesium bromide to N-benzyl-cis-3-ethylaziridin-2-car-
boxaldehyde 6 (Table 3, entry 1) provided the syn[22] adduct
6a exclusively. The corresponding zinc reagent afforded a
cleaner reaction with improved yield and similarly high syn/
anti ratio (Table 3, entry 2). However, addition of TMEDA,
a strong chelator of magnesium ions, lowered the syn selec-
tivity (Table 3, entry 3). Thus, chelation of the organometal-
lic compoundꢀs metal center by the electron-rich nitrogen
and the carbonyl oxygen appears to be the key to the selec-
tivity of the reaction (Figure 3a). To avoid eclipsing steric
interactions, the N-Bn substituent in the cis-substituted sub-
strate is expected to prefer the anti conformation, with the
Figure 2. B3LYP/6-31+G*-optimized TS structures (relative energies)
for chelation-controlled attack on the AN conformer of the model com-
pound Me-H3. Notably, the pro-anti path (b) is calculated to be 1.9 kcal
molꢀ1 higher in energy than the pro-syn path (a).
12330
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12326 – 12339

Selectivity in the Addition Reactions of Organometallic Reagents
FULL PAPER
Table 3. Addition of organometallic reagents to Bn-protected cis-aziri-
dine-2-carboxaldehydes.
reaction pathway with CH3Li and PhLi, yielding products in
syn/anti ratios of 4:1 and 2.5:1, respectively.[8] In related
work, Noh et al. reported a similar chelation-controlled
strong preference for syn products in organometallic addi-
tions to an N-alkylated aziridine-2-carboxaldimine lacking
any additional substituents on the ring.[23] Similarly, Hou
et al. have reported additions of phosphite ions to N-benzyl-
2-aziridinesulfinimines; as expected for a chelation con-
trolled process, more Lewis acidic metal ions favored syn se-
lectivity, which plummeted when less coordinating ions, such
as Na+, were used.[24]
Entry
Reagent
Additive
6a/6b[a]
Yield
1
2
3
EtMgBr
Et2Zn
EtMgBr
–
–
>99:1[b]
>99:1[b]
62:38
75
85
65
TMEDA[c]
Added TMEDA destroys the reactionꢀs selectivity, pre-
sumably by blocking chelation in the Grignard reaction of
compound 6 (Table 3, entry 3). Thus, an uncomplexed neu-
tral aldehyde apparently has little intrinsic selectivity. Com-
putational modeling on Me-cis, the analogue of 6 used for
computational analysis (see Table 1), shows a small energy
difference between endo- and exo-CHO rotamers for the
anti series (the syn series are not considered here since they
are energetically inaccessible). Thus, an incoming nucleo-
phile could react with either the pro-syn or pro-anti face of
the aldehyde with nearly equal probability, even as it avoids
the face shielded by the cis oriented ethyl group. Interest-
ingly, despite the (presumably anti-selective) AX conforma-
tion being lower in energy (see Table 1) and thus prevalent,
a 62:38 preference for syn addition is still seen; evidently
the smaller population of AN outcompetes AX, supporting
the notion that endo conformations are more reactive.
With chelationꢀs observably important role in stereochem-
ical control, it was natural to attempt a direct study of the
chelating properties of the N-benzyl aziridine-2-carboxalde-
hydes by NMR spectroscopy. Unfortunately, the cis substi-
[a] Diastereomeric ratios were determined by NMR spectroscopy.
[b] Only the syn compound was observed by NMR spectroscopy.
[c] TMEDA (10 equiv) was used.
tuted N-benzyl aziridine-2-carboxaldehyde
6 uniformly
formed precipitates when combined with MgBr2, MgCl2, or
ZnCl2 in CDCl3 or [D8]toluene. Evidently, association does
occur with these Lewis acidic MX2 species. However, in
Grignard reactions of 6, no precipitates were noted. Perhaps
when the X (Cl or Br) on the metal is small and polar, the
resulting complex is also polar enough to aggregate and pre-
cipitate, whereas with an organic group in place of a simple
halide the complexes remain soluble.
In contrast to the cis case, trans aziridine carboxaldehyde
7 (Table 4) was found to exist as a mixture of two inverto-
mers in a 1:1.2 ratio. Structural assignments based on
1H NMR spectroscopy (NOE analysis) suggest that the fa-
vored invertomer is syn with the benzyl and aldehyde sub-
stituents on the same side of the aziridine ring (e.g., struc-
tures SX or SN in Figure 4). This finding is in accord with
conventional notions of steric demand that suggest the alde-
hyde should be smaller than the ethyl group (their respec-
tive A values are 0.56–0.8 and 1.79).[25] Furthermore, for
trans-1,3-dimethyl-2-acetyl aziridine, Pierre et al. found a 3:1
preference for the syn conformation.[21] Also, the results
stand in agreement with our DFT results for the model aziri-
dine Me-trans in simulated CH2Cl2, in which, for the respec-
tive syn and anti analogues of 7, the SN conformer lies
0.3 kcalmolꢀ1 lower than AN. Perhaps even more important
Figure 3. a) Proposed syn-selective attack on N-benzyl-cis-3-ethylaziri-
dine-2-carboxaldehyde 6 under chelation control. b) RHF/3-21G-opti-
mized structure of 6 chelating ZnACHTNUGTRENUNG(CH3)2. The structure rationalizes the
observed high syn selectivity, as the anti approach is blocked by the cis-
ethyl substituent.
nitrogen atom lone pair on the same face as the aldehyde
(the AN conformation). This arrangement ideally sets the
stage for chelation to control the aldehyde carbonyl groupꢀs
orientation. In the resulting chelate, the anti approach
should be blocked by the cis ethyl substituent leaving syn
addition as the dominant pathway, consistent with the
strong syn preference seen in these reactions (Figure 3b). It
should be noted that chelation offsets the intrinsic prefer-
ence for the AX conformer, which is favored by only
ꢂ0.5 kcalmolꢀ1 in the model Me-cis system. This line of rea-
soning is exactly how Pierre et al. rationalized the strong
stereoselectivity they observed for LiAlH4 reductions of
aziridinyl ketones in diethyl ether.[8] Although the cis sub-
stituent clearly augments the observed selectivity,[4a] it is not
the only directing factor; even N-tert-butylaziridine-2-car-
boxaldehyde, with no additional substituents and the tert-
butyl group oriented trans to the aldehyde, prefers the syn
Chem. Eur. J. 2011, 17, 12326 – 12339
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12331

B. Borhan, J. E. Jackson et al.
Table 4. Addition of organometallic reagents to Bn-protected trans-aziri-
appears that complexation is unimportant in reactions with
this substrate. This finding makes sense in light of the high
relative energy calculated for the AN conformer, a signifi-
cant point of contrast with the chelation-capable cis sub-
strate 6. For 6, chelation locks the carbonyl rotamer in the
AN geometry, and the ethyl substituent offers strong stereo-
differentiation by blocking one face of the aldehyde. We
were thus surprised at first to find that, as with 6, addition
of Lewis acidic MX2 to 7 quantitatively precipitated com-
plexed aldehyde, removing it from solution in both toluene
and chloroform; evidently, these systems are capable of
strong monodentate complexation, a conclusion supported
by the structural studies of Bartnik et al.[26] Notably, in the
X-ray crystal structure of a ZnBr2 complex of 2-benzoyl azir-
idine, the heterocycle only serves as a monodentate ligand
through nitrogen, despite the carbonyl group being in an
endo conformation. This finding suggests that chelation in
the aziridine-2-carbonyl framework is at most capable of
modest energy lowering, on the same energy scale as confor-
mational variations.
dine-2-carboxaldehydes.[a]
Entry
Reagent
Additive
7a/7b[b]
Yield
1
2
EtMgBr
EtMgBr
–
55:45[c]
55:45[c]
75
85
TMEDA[d]
[a] Exists as 55:45 mixture of invertomers at N observable by NMR spec-
troscopy at 258C. [b] Diastereomeric ratios were determined by NMR
spectroscopy. [c] The syn and anti compounds could not be separated.
NMR analysis of the product mixtures showed a 55:45 mixture of unas-
signed stereoisomers. [d] TMEDA (10 equiv) was used.
For substrate 7, our analysis points to the SN conforma-
tion (Figure 4) as the lowest energy form. Here, a favorable
ꢀ
dipole–dipole interaction between the Bn N and aldehyde
C=O moieties favors the endo orientation of the aldehyde
carbonyl (Figure 5). This conformation itself would be ex-
ꢀ
Figure 5. Dipole interactions between the Bn N and CHO moieties in 7,
proposed to explain the relative order of stability of the conformers. Rel-
ative energies shown are those calculated for the model system Me-trans.
Based on the simple steric size order CHO<Et, the syn systems SN and
SX should be favored. However, the added dipole repulsion in AN and
attractions in AX and SN ultimately disfavor chelation in this system by
pushing AN to a relatively high energy. Presumably, the sterically un-
AHCTUNGERTGfNNUN avorable AX isomer lies close in energy to SX as a result of AXꢀs favor-
able dipole interaction counterbalancing SXꢀs less than ideal dipole ar-
rangement.
Figure 4. Energy-minimized conformers of Bn-protected trans substrate
7. Among these forms, both pro-syn and pro-anti faces are equally avail-
able. Of the Me-trans model structures, the SX and AX conformers lie
only 0.3 kcalmolꢀ1 above the lowest energy SN form; B3LYP/6-31G*/
SM8 energies of the actual structures space them more widely but still
compress their overall energy range to ꢂ1 kcalmolꢀ1. In contrast, the po-
tentially chelating AN conformer lies an additional 1.3 kcalmolꢀ1 higher,
which explains the apparent lack of chelation in additions to 7.
pected to have a modest syn preference, but the two low-
lying exo forms (SX and AX) that are only a fraction of a
kcalmolꢀ1 higher in energy offer reactivity favoring the anti
product. This equilibrating mix of conformations explains
the unselective addition reactions found both in the pres-
ence and absence of the TMEDA additive (Table 4,
entry 2). Thus, for N-benzyl-protected trans substrates,
almost no de is seen in the mixture of syn and anti adducts
as neither chelation nor a single strong candidate among the
free conformers is available to control the stereochemistry
of the Grignard addition reaction.
for the reactionꢀs stereochemical outcome is the calculated
energy difference between the model endo and exo aldehyde
rotamers SN and SX (see Table 1), which for Me-trans is
also quite small, suggesting little differentiation between
access to the two faces of the aldehyde carbonyl. One might
expect chelate formation to offer stereocontrol through the
AN conformation of 7. However, upon addition of ethyl
magnesium bromide to compound 7, an almost equimolar
mixture of syn and anti adducts 7a and b was obtained
(Table 4, entry 1). Notably, this diastereomeric ratio was un-
affected by addition of TMEDA (Table 4, entry 2); thus, it
Boc-protected aziridine-2-carboxaldehydes: In contrast to
the N-benzyl cases discussed above, for which the cis-substi-
12332
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12326 – 12339

Selectivity in the Addition Reactions of Organometallic Reagents
FULL PAPER
tuted system was most selective, in the N-Boc series, it is the
trans substrates that show the greatest selectivity. Addition
of methyl magnesium bromide to Boc-protected trans-3-cy-
clohexylaziridine-2-carboxaldehyde 8 led exclusively to the
formation of syn[27] alcohol 8a (Table 5, entry 1). To our sur-
Table 5. Addition of organometallic reagents to Boc-protected trans-azir-
idine-2-carboxaldehydes.
Figure 6. Comparison of the crystal structure of 8 (a) with its B3LYP
energy minimized model (b) illustrates excellent similarity. Both struc-
tures indicate the SN isomer is the most stable conformation. The dotted
line highlights the favorable electrostatic arrangement of the aldehyde
with the Boc carbonyl (electrostatically chelated).
Entry
Additive
8a/8b[a]
Yield
1
2
3
4
–
–
>99:1[b]
>99:1[b,c]
>99:1[b]
>99:1[b]
85
83
80
78
TMEDA[d]
lattice, so its internal structural preferences would likely
dominate the conformation that crystallizes. Also, for both
the actual substrate 8 and, more rigorously, on our model
system Moc-trans, calculations find a strong preference for
the same conformation that is found in the X-ray crystal
structure, as shown in Figure 6b. The contact distance (2.8
and 2.9 ꢄ for X-ray and calculated structures, respectively)
between the aldehyde oxygen and the Boc carbonyl carbon
is substantially below that of a van der Waals contact
(3.2 ꢄ), suggesting a favorable interaction. This notion is re-
inforced by the observation of pyramidalization at the Boc
carbon; the N-C=O-O torsion angle is compressed (1728 ex-
perimental; 1738 theory) from the expected 1808, suggesting
the beginning of a Bꢂrgi–Dunitz-like attack trajectory and
the corresponding electrophilic activation of the aldehyde
carbonyl.[29] The effect of this aldehyde–Boc dipole–dipole
attraction is to favor both the endo orientation of the alde-
hyde and the nitrogen invertomer with the Boc group cis to
the aldehyde moiety, that is, the SN conformation. In a simi-
lar manner to the chelate structure invoked for addition to
cis-Bn system 6, this SN form favors attack on the more ac-
cessible and reactive pro-syn face of the aldehyde, yielding
the observed high selectivity (compare Figures 3b and 6b to
note similarities in conformation and the resulting face se-
lectivity).
[e]
MgBr2·OEt2
[a] Diastereomeric ratios were determined by NMR spectroscopy.
[b] Only the syn compound was observed by NMR spectroscopy. [c] Re-
action was carried out at ꢀ788C. [d] TMEDA (10 equiv) was used.
[e] Freshly prepared MgBr2·OEt2 (2.5 equiv) was used.
prise, despite the N activation expected from the Boc group,
ring-opened products were not observed during these addi-
tions even near ambient temperatures; comparable yields
and selectivities of product 8a were obtained at 08C and at
ꢀ788C (Table 5, entry 2). Addition of strong chelating re-
agents, such as TMEDA, or Lewis acids, such as
MgBr2·OEt2, had no effect on the isomer distribution of the
product; in fact, the anti addition compound 8b was never
seen to form under any of the reaction conditions studied,
as confirmed through independent Mitsunobu conversion of
8a into 8b (see the Supporting Information) and spectro-
scopic analysis. Thus, in contrast to the N-benzyl series, for
the N-Boc aziridine-2-carboxaldehydes, it is the trans sub-
strates that show excellent syn selectivity in a non-chelation-
controlled Grignard addition (Table 5, entries 3 and 4). This
finding is at odds with a previous report[4b] in which a che-
late is suggested between the nitrogen lone pair, the alde-
hyde carbonyl oxygen and the magnesium atom. However,
the observed failure of additives to alter the selectivity
ratios argues against this scenario. Although the Boc
groupꢀs carbonyl oxygen could offer a coordination site to
participate in chelation, the unfavorable 7-membered-ring
size and the Boc groupꢀs strong preference for conforma-
tions that maintain conjugation between the nitrogen lone
pair and the Boc carbonyl group make chelation appear un-
likely.
The case of Boc-protected cis aziridine-2-carboxaldehyde
is less clear-cut than that of the trans compound in terms of
selectivity; addition of Grignard reagents resulted in moder-
ate yields with variable low to high syn selectivity (Table 6).
Ring-opening products were frequently observed if the re-
AHCTUNGTREGaNNUN ction was carried out at room temperature. As a result,
Grignard additions were run at 08C and then slowly
warmed to room temperature. Notably, syn selectivity im-
proved as the size of the nucleophile increased (compare
Table 6, entries 1 vs. 5). Chelation appears to exert weak ef-
fects on this addition; formation of the Lewis acidic dialkyl
zinc[5a,30] reagent (Table 6, entry 3) led to a modest increase
in syn selectivity, whereas addition of TMEDA decreased it
(Table 6, entries 2 and 6). Simple steric arguments suggest
that the Boc protecting group should be situated trans to the
aziridine substituents. Our modeling studies agree, but find
only small energy differences (0.1–0.5 kcalmolꢀ1) among the
This issue is further illuminated by the X-ray crystal struc-
ture of Boc-protected trans substrate 8 (Figure 6a), in which
the Boc and carboxaldehyde moieties are in the SN confor-
mation. Interpretation of crystal structures as models for rel-
evant solution state conformations of reactive species must
be approached skeptically,[28] due to crystal packing interac-
tions and the absence of solvation effects. However, this rel-
atively rigid and hydrophobic compound is only capable of
modest dipole–dipole and van der Waals interactions in the
Chem. Eur. J. 2011, 17, 12326 – 12339
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12333

B. Borhan, J. E. Jackson et al.
Table 6. Addition of organometallic reagents to Boc-protected cis-aziri-
dine-2-carboxaldehydes.
ence for the syn addition product. However, although it re-
mains pyramidalized due to being in a three-membered ring,
the somewhat delocalized carbamate nitrogen would only
be capable of weak binding at best, unlike in N-Bn cis sub-
strate 6, for which chelation is clearly the controlling factor.
As a result, transition states based on the AX conformation
of the aldehyde may compete; with the cis substituent
blocking the approach from the pro-syn face, these confor-
mations should lead to predominantly anti addition (Fig-
ure 7b). Thus the opposing effects of weakly chelation-fa-
vored endo, aldehyde–Boc attraction, and energetically
close exo aldehyde orientations may explain the limited se-
lectivity found in these additions.
Entry
R
T [8C]
Additive
9a/9b[a]
Yield
1
2
3
4
5
6
7
C2H
C2H
C2H
C2H
Ph
0
0
0
–
59:41[b]
50:50[b]
70:30[b]
52:48[b]
90:10
53
51
60
50
55
52
49
TMEDA[c]
[d]
ZnCl2
ꢀ78
–
–
0
Ph
Ph
0
ꢀ78
TMEDA[c]
–
61:39
88:12
In the case of Boc-protected 2,3-disubstituted aziridine-2-
carboxaldehyde 10, the reaction occurred with moderate to
good syn selectivity. Both reactivity and conversion rates
were found to diminish at lower temperatures, as seen with
the cis substrates (Table 7, entries 3 and 4). Addition of
[a] Diastereomeric ratios were determined by NMR spectroscopy. [b] Ste-
reochemistry of the anti alcohol was confirmed by X-ray crystallography.
[c] TMEDA (10 equiv) was used. [d] Anhydrous ZnCl2 (2.5 equiv) was
used.
Table 7. Addition of organometallic reagents to Boc-protected 2,3-disub-
stituted aziridine-2-carboxaldehydes.
AX, AN, and SN conformers for the Moc-cis model system.
Although the AX conformer is lowest in energy, AN and SN
are only 0.2–0.3 kcalmolꢀ1 higher in energy. Thus, two fac-
tors seem to be relevant in this system: 1) Weak chelation
between the CHO and N groups in the AN conformer shifts
the proportions of nearly equienergetic conformers towards
those that favor the syn products. 2) Larger nucleophiles dis-
criminate more strongly against exo conformers (AX in this
case) than smaller nucleophiles, presumably owing to the
exo rotamer being harder to attack, as discussed above (Fig-
ure 1b). Thus, the greater selectivity seen with the larger nu-
cleophile suggests that the endo conformers AN and SN
dominate those additions. Interestingly, the favorable
dipole–dipole interaction between the aldehyde carbonyl
and the Boc group appears to almost completely offset the
steric effects of the cis hexyl side chain, placing the SN in-
vertomer only 0.3 kcalmolꢀ1 above AX in the Moc-cis
model, and 0.4 kcalmolꢀ1 higher in compound 9 itself. This
last finding highlights the favorable aldehyde–Boc dipole–
dipole interaction seen in the trans case to control both
ground-state conformation and addition stereochemistry.
Formation of a cyclic chelate (Figure 7a) with the metal
of methylmagnesium chloride coordinated between nitrogen
and the aldehyde oxygen might explain the modest prefer-
Entry
T [8C]
Additive
10a/10b[a]
Yield
1
2
3
4
0
0
–
85:15[b]
85:15[b]
60:40[b]
71:29[b]
85
72
TMEDA[d]
–
ꢀ78
51[c]
48
ꢀ78
TMEDA[d]
[a] Diastereomeric ratios were determined by NMR spectroscopy. [b] Ste-
reochemistry was determined as discussed in the Supporting Information;
[c] Conversion=79%; for all other entries, conversion=100%.
[d] TMEDA (10 equiv) was used.
TMEDA also showed temperature-dependent selectivity ef-
fects, enhancing syn adduct formation at ꢀ788C (Table 7,
entry 4) but not at 08C (Table 7, entry 2). The calculated
lowest energy conformation of 10 has both the Boc and the
aldehyde on the same side of the ring, but between the com-
peting effects of the alpha substituent, which favors the exo
rotamer (Figure 8a), and the dipole–dipole interaction with
the Boc group, which favors the endo rotamer (Figure 8b),
there is little ground-state conformational preference.
The temperature effect disclosed in Table 7 seems coun-
terintuitive. We speculate, however, that the higher syn/anti
ratio at higher temperatures might be due to an increase in
population of the highly syn-selective SX conformer. Based
on the energetics of the Moc-2,3Di model system, the slight-
ly lower energy SN isomer should predominate (Figure 8).
However, if the SX isomer had both higher reactivity and
stronger syn selectivity (as expected due to the shielding of
the pro-anti approach by the Boc group), its contribution to
product formation would grow quickly with increasing tem-
perature. As noted earlier, ipso alkyl substitution enhances
the reactivity of exo versus endo aldehyde rotamers; thus
SX reactivity would be enhanced by the methyl substitution
Figure 7. a) Chelated AN conformation of a model simulating Boc-pro-
tected cis-aziridine 9 would favor syn-selective addition (pictured with
MeMgCl). b) The exo isomer (AX) is not capable of chelation and would
favor a pro-anti approach of the nucleophile.
12334
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12326 – 12339

Selectivity in the Addition Reactions of Organometallic Reagents
FULL PAPER
dehydes, only one invertomer
needs to be considered, since
the protecting group is situated
trans to the substituent. Al-
though the calculated lowest
energy conformation is exo,
multiple exo and endo confor-
mations lie within 0.5 kcalmolꢀ1
of the lowest form, differing
only in the rotameric orienta-
tions of the ethyl and tosyl
groups. Thus both faces of the
Figure 8. Low energy conformers of 10 showing near isoenergetic a) SX (0.23 kcalmolꢀ1) and b) SN (0 kcal
molꢀ1) CHO rotamers [note: these energies are for the Moc-2,3Di modeled SX and SN isomers].
at the ipso position (see discussion of the ipso effect above).
At ꢀ788C, the less syn selective SN isomer would dominate
due to both the decreased temperature and potentially due
to Lewis acid amplification of the Boc groupꢀs pseudochelat-
ing interaction with the aldehyde.[31] The latter notion is fur-
ther supported by the results for TMEDA addition at
ꢀ788C, which increases the syn/anti product ratio, presuma-
bly by disrupting the proposed SN-supporting complexation.
carbonyl are expected to be equally prone to nucleophilic
attack, which explains the lack of diastereoselectivity.
Addition of Grignard reagents to trans-substituted N-Ts
aziridine-2-carboxaldehydes proceeded with moderate to ex-
cellent syn selectivity and good yields (Table 9, entries 1–7).
In this case, nucleophilic ring-opening reactions were com-
pletely suppressed at ꢀ788C. Chelation appears to play no
role in these reactions as addition of MgBr2·OEt2 (Table 9,
entry 3) or TMEDA (Table 9, entry 2) had no remarkable
effect either on selectivity or yield.
Ts-protected aziridine-2-carboxaldehydes: Grignard addition
to Ts-protected cis-aziridine-2-carboxaldehydes was accom-
panied by undesired nucleophilic aziridine ring opening. To
suppress these reactions, it was crucial to carry out the addi-
tion at ꢀ788C. The cis substrate 11 exhibited poor reactivity
towards nucleophilic addition at the aldehyde carbonyl at
this temperature. Conversion was low even after adding an
excess (10 equiv) of the Grignard reagent. In all cases, the
desired alcohol was either difficult to isolate, or if isolable,
was recovered in low yield (Table 8, entry 2).
Table 9. Addition of organometallic reagents to Ts-protected trans-aziri-
dine-2-carboxaldehydes.
Entry
Substrate
R
Additive
a/b[a]
70:30
70:30
69:31
70:30
72:28
67:33
>99:1
Yield
Table 8. Addition of organometallic reagents to Ts-protected cis-aziri-
dine-2-carboxaldehydes.
1
2
3
4
5
6
7
12
12
12
13
13
14
14
C2H
C2H
C2H
C2H
C2H
C2H
Ph
–
76
73
75
81
69
71
85
TMEDA[b]
[c]
MgBr2·OEt2
–
[c]
MgBr2·OEt2
–
–
[a] Diastereomeric ratios were determined by NMR spectroscopy.
[b] TMEDA (10 equiv) was used. [c] Freshly prepared MgBr2·OEt2
(2.5 equiv) was used.
Entry
R
Additive
11a/11b[a]
Yield
1
2
3
4
Et
–
–
50:50
56:43
50:50
50:50
n.d.[b]
20
C2H
C2H
C2H
TMEDA[c]
MgBr2·OEt2
n.d.[b]
85
[d]
In the calculated minimum energy conformation of 13
(Figure 9), the aldehyde moiety has a strong preference for
the exo orientation. The Ms-trans model indicates that the
SX isomer is 0.5 kcalmolꢀ1 more stable than the AX isomer.
As such, the bulky tosyl group in the SX isomer blocks one
face of the carbonyl and directs the nucleophile to approach
from the opposite direction, yielding the syn addition prod-
uct. On the other hand, the AX isomer is less syn selective,
owing to the fact that the Ts group is on the opposite face
of the aziridine ring. This is presumably the cause of the
lowered overall syn/anti product ratios. With this in mind,
we hypothesized that increasing the size of the incoming nu-
cleophile would increase the stereodifferentiation since the
pro-anti approach of the AX isomer would be sterically re-
[a] Diastereomeric ratios were determined by NMR spectroscopy.
[b] Yields could not be determined due to undesired side reaction prod-
ucts. [c] TMEDA (10 equiv) was used. [d] Freshly prepared MgBr2·OEt2
(2.5 equiv) was used.
In an attempt to promote the reaction by Lewis acid acti-
vation, MgBr2·OEt2 was added. In this case, higher yields of
the alcohol products were isolated (Table 8, entry 4), but no
higher selectivity between the syn and anti adducts was ob-
served than in the other cases. The product ratios were not
altered under chelating or non-chelating conditions, which
argues against the involvement of a chelated transition state.
As observed with the Bn-protected cis-aziridine-2-carboxal-
Chem. Eur. J. 2011, 17, 12326 – 12339
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12335

B. Borhan, J. E. Jackson et al.
Figure 9. The lower energy SX conformation of 13 is syn selective due to
steric blocking afforded by the Ts protecting group. The AX conformer
exhibits less steric bias and thus would reduce the overall stereoselectiv-
ity. The modest selectivity in Ts-protected trans-aziridine-2-carboxalde-
hyde could arise from the energetically close SX and AX conformers.
Figure 10. Comparison of the molecular structure of 15, as determined by
X-ray diffraction (b), with the lowest energy B3LYP/6-31G*-optimized
aziridine conformer (a) of 15. The Ts-2,3Di model structure shows an
analogous SX-type lowest conformation.
tarded. Indeed, addition of phenyl magnesium bromide to
substrate 14 afforded exclusively the syn adduct 14a-Ph in
good yield (Table 9, entry 7).
Both high yield and exceptional syn selectivity were ach-
ieved with 2,3-disubstituted aziridine-2-carboxaldehyde 15
(Table 10). Similar to the trans system, selectivity was not at
strong conformational preference and reaction stereospeci-
ficity.
The structural similarity between the most stable ground-
state conformations calculated for trans aziridines 12–14 and
2,3-disubstituted aziridine-2-carboxaldehyde 15 does not en-
tirely justify the exceptional syn selectivity observed in the
case of the 2,3-disubstituted aziridine-2-carboxaldehyde. The
extra methyl substituent at the alpha position clearly makes
a considerable difference in yield and level of stereoselectiv-
ity. A careful analysis of energy-minimized conformations
for 13 and 15 suggests that the difference in stereoselectivity
is a result of two factors, both of which favor a more defined
ground-state conformation in 15 as compared to 13. These
factors are that 1) the tosyl group is sterically limited to a
syn relationship with the aldehyde by the two alkyl substitu-
ents on the opposite ring face; and 2) the methyl group
alpha to the aldehyde favors the exo aldehyde rotamer.
Scheme 4 illustrates the expected invertomer equilibria for
13 and 15, based on the energies calculated for the syn and
anti (specifically, SX and AX) conformations of model com-
pounds Ms-trans and Ms-2,3Di, respectively. As expected,
the syn invertomers are favored in both cases, but the AX–
SX DG energy difference is significantly larger (4.03 vs.
0.50 kcalmolꢀ1; see Table 1) for Ms-2,3Di than for Ms-trans,
and hence by analogy for 15 than for 13.
Table 10. Addition of organometallic reagents to Ts-protected 2,3-disub-
stituted aziridine-2-carboxaldehyde.
Entry
R
Additive
a:b[a]
Yield
1
2
3
4
C2H
C2H
C2H
Ph
–
>99:1[b,c]
>99:1[b,c]
>99:1[b,c]
>99:1[b,c]
92
81
79
85
[d]
MgBr2·OEt2
TMEDA[e]
–
[a] Diastereomeric ratios were determined by NMR spectroscopy.
[b] Only the syn isomer was seen by NMR spectroscopy. [c] Stereochem-
istry was confirmed by X-ray crystallographic analysis. [d] Freshly pre-
pared MgBr2·OEt2 (2.5 equiv) was used. [e] TMEDA (10 equiv) was
used.
all affected by addition of TMEDA (Table 10, entry 3) or
MgBr2·OEt2 (Table 10, entry 2). As in the trans case, the
energy-minimized conformation (Figure 10a) orients the al-
dehyde carbonyl exo, such that the tosyl group blocks the
pro-anti face of the carbonyl, thus favoring syn-selective ap-
proach of the nucleophile. Based on the Ms-2,3Di model,
the SX isomer is by far the most stable conformation. The
high exo tendency is not only due to the Ts protecting
group, but also, as described above, the result of the ipso
methyl substitution. The highly favored syn orientation, as
expected, originates from the nitrogen atom adopting the
necessary invertomer to avoid a steric clash with the two
substituents on the anti face of the aziridine ring. The crystal
structure obtained for aziridine-2-carboxaldehyde 15 (Fig-
ure 10b) is in close agreement with our optimized structures,
supporting the above rationalization of this substrateꢀs
Scheme 4. a) The N invertomers of Ts-protected 2,3-disubstituted aziri-
dine-2-carboxaldehyde. b) The N invertomers of Ts-protected trans aziri-
dine aldehyde.
12336
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12326 – 12339

Selectivity in the Addition Reactions of Organometallic Reagents
FULL PAPER
Although the steric bias toward syn (Scheme 4a) in 15 is
straightforward, that in 13 results from a delicate balance
between n-alkyl and CHO groups as substituents cis to the
Ts group. The 0.50 kcalmolꢀ1 calculated anti/syn energy dif-
ference for Ms-trans predicts an invertomer ratio of 2.3:1 at
238C. Because the barrier to nitrogen inversion is higher in
aziridines than in other pyramidal nitrogen heterocycles,
their invertomer spectra can readily be resolved in low-tem-
perature NMR studies.[32] We therefore resorted to dynamic
1H NMR measurements to explore the conformational be-
havior of the trans and 2,3-disubstituted aziridine-2-carbox-
aldehydes 13 and 15 at different temperatures (Figure 11).
In the case of 15, resonances displayed essentially identical
line shapes at temperatures from 25 to ꢀ708C, confirming
the exclusive presence of one conformation (presumably
15b in Scheme 4a), which is expected (see Figure 11a vs. b
for NMR spectra). In contrast, compound 13 showed broad-
ening of most peaks as a function of lowering temperature
(Figure 11c vs. d). Of particular interest is the aldehyde res-
onance, which showed considerable broadening upon cool-
ing and eventually exhibited decoalescence below ꢀ648C.
At the lowest temperature observed, ꢀ708C, two distinct
broad peaks were visible in a ratio of 88:12 (integrated
ratio); this translates into a DDG8 of 0.80 kcalmolꢀ1. Con-
firming the Ms-trans model systemꢀs calculated 0.58 kcal
molꢀ1 free-energy difference, a detailed conformational
search and optimization of 13 at the B3LYP/6-31G* level
with single-point SM8 “CH2Cl2” solvation calculations yield-
ed a calculated DDE of 0.63 kcalmolꢀ1 between the inverto-
mers 13-SX and 13-AX; the resulting prediction of a 4.8:1
invertomer ratio is in reasonably good agreement with the
ꢂ7:1 value observed. Full line-shape analysis of the alde-
hyde resonance (Figure 11e and f),[33] used to determine the
activation parameters, leads to the calculation of the energy
barrier for nitrogen inversion of 13 (DG° =10.7ꢁ1.4 kcal
molꢀ1, DH° =7.99ꢁ0.3 kcalmolꢀ1, and DS° =ꢀ8.94ꢁ
0.3 calmolꢀ1 Kꢀ1, data from an Eyring plot, see the Support-
ing Information). The corresponding B3LYP/6-31G*/SM8-
CH2Cl2 calculation for Ms-trans finds an inversion DG° of
10.2 kcalmolꢀ1, quite similar to the experimental value for
13.[34] With this modest inversion barrier, both invertomers
would be available throughout the reaction with the organo-
metallic nucleophile. Presumably, this mix of substrate con-
formers also leads to the observed moderate selectivity in
the addition of organometallic reagents to Ts-protected
trans-aziridine-2-carboxaldehyde 13, in contrast to the exclu-
sive syn selectivity seen with SX-dominated 2,3-disubstituted
aziridine-2-carboxaldehyde 15.
Conclusion
The selectivity of organometallic addition to N-protected
cis, trans, and 2,3-disubstituted aziridine-2-carboxaldehydes
is governed by multiple intimately coupled factors. Most im-
portant are the rotameric orientation of the aldehyde func-
tionality and invertomeric preference of the nitrogen pro-
tecting group. As with cyclopropane carboxaldehydes, the
CHO moiety strongly prefers bifurcated geometries relative
to the ring, with the oxygen oriented either endo or exo; ro-
tation to other orientations can cost as much as 6 kcalmolꢀ1.
Also, although conjugation might have been expected to
oppose it, the aziridine nitrogen remains strongly pyramidal-
ized regardless of which substituent it carries; invertomeric
alternative conformations therefore must be considered.
These aspects are further modulated by the N substituentsꢀ
electronic nature, and by the steric environment set by the
other substituents on the aziridine ring (see Table 11 for a
summary). Additional variation arises from the use of differ-
ent metal and alkyl components of the organometallic nu-
cleophiles.
Figure 11. Aldehyde resonance of 15 at a) ꢀ128C and b) ꢀ708C. Alde-
hyde resonance of 13 at c) ꢀ128C (expansion shown in the inset) and
d) ꢀ708C. e) Simulated aldehyde resonance of 13 at ꢀ128C (expansion
shown in the inset). f) Simulated aldehyde resonance of 13 at ꢀ708C.
For Bn protected aziridine carboxaldehydes, selectivity
originates from chelation, consistent with findings in the lit-
Chem. Eur. J. 2011, 17, 12326 – 12339
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12337

B. Borhan, J. E. Jackson et al.
Table 11. Summarized results of the selectivity (syn/anti) observed in the
addition reactions of organometallic reagents to aziridine carboxalde-
hydes.
Acknowledgements
Generous support was provided by the NIH (No. R01-GM082961). The
authors are thankful to Dr. Daniel C. Whitehead for fruitful and illumi-
nating discussions.
N protecting group (R)
Substitution pattern
Bn
Boc
Ts
cis
trans
2,3Di
excellent
poor
–
poor
excellent
moderate
poor
moderate
excellent
[1] a) G. Cardillo, L. Gentilucci, A. Tolomelli, Aldrichimica Acta 2003,
36, 39–50; b) V. H. Dahanukar, I. A. Zavialov, Curr. Opin. Drug
Discovery Dev. 2002, 5, 918–927; c) D. Tanner, Angew. Chem. 1994,
106, 625–646; Angew. Chem. Int. Ed. Engl. 1994, 33, 599–619;
d) A. H. Li, L. X. Dai, V. K. Aggarwal, Chem. Rev. 1997, 97, 2341–
2372; e) W. McCoull, F. A. Davis, Synthesis 2000, 1347–1365; f) J. B.
Sweeney, Chem. Soc. Rev. 2002, 31, 247–258.
[2] a) J. C. Antilla, W. D. Wulff, J. Am. Chem. Soc. 1999, 121, 5099–
5100; b) J. C. Antilla, W. D. Wulff, Angew. Chem. 2000, 112, 4692–
4695; Angew. Chem. Int. Ed. 2000, 39, 4518–4521; c) T. B. Bisol,
M. M. Sa, Quim. Nova 2007, 30, 106–115; d) C. A. Olsen, H. Fran-
zyk, J. W. Jaroszewski, Eur. J. Org. Chem. 2007, 1717–1724;
e) H. M. I. Osborn, J. Sweeney, Tetrahedron: Asymmetry 1997, 8,
1693–1715; f) J. Vesely, I. Ibrahem, G. L. Zhao, R. Rios, A. Cꢅrdo-
va, Angew. Chem. 2007, 119, 792–795; Angew. Chem. Int. Ed. 2007,
46, 778–781; g) A. A. Desai, W. D. Wulff, J. Am. Chem. Soc. 2010,
132, 13100–13103.
[3] J. M. Schomaker, A. R. Geiser, R. Huang, B. Borhan, J. Am. Chem.
Soc. 2007, 129, 3794–3795.
[4] a) J. Andres, N. de Elena, R. Pedrosa, A. Perez-Encabo, Tetrahedron
1999, 55, 14137–14144; b) G. Righi, S. Pietrantonio, C. Bonini, Tet-
rahedron 2001, 57, 10039–10046; c) G. Righi, S. Ciambrone, Tetrahe-
dron Lett. 2004, 45, 2103–2106.
[5] a) R. Noyori, M. Kitamura, Angew. Chem. 1991, 103, 34–55;
Angew. Chem. Int. Ed. Engl. 1991, 30, 49–69; b) H. Urabe, O. O.
Evin, F. Sato, J. Org. Chem. 1995, 60, 2660–2661; c) M. T. Reetz,
Angew. Chem. 1984, 96, 542–555; Angew. Chem. Int. Ed. Engl.
1984, 23, 556–569.
[6] a) N. Anh, Top. Curr. Chem. 1980, 88, 145–162; b) G. Righi, S.
Ronconi, C. Bonini, Eur. J. Org. Chem. 2002, 1573–1577; c) H.
Urabe, F. Sato, J. Org. Chem. Jpn. 1993, 51, 14.
erature. Excellent syn selectivity (>99:1 syn/anti) is achieved
for N-Bn cis aziridine carboxaldehydes, whereas the trans
substrates undergo almost completely unselective (55:45
syn/anti) addition. For this case, as well as the N-Boc-pro-
tected cis and 2,3-disubstituted systems, the low selectivity
seems to be due to the substratesꢀ conformational ambiguity,
both between the aldehyde endo/exo rotamer and N inverto-
mer structures. However, for the conformationally well-de-
fined N-Boc trans substrates, extremely high syn selectivity
(>99:1 syn/anti) is observed that is unaffected by chelation
modifiers. This strong stereopreference appears to be dictat-
ed by the favorable chelation-like interaction between the
aldehyde carbonyl and the syn-Boc groupꢀs electrophilic
carbon. Because of amide resonance, the Boc substituent
adopts a tangential rotameric orientation relative to the azir-
idine three-membered ring; this is the opposite of the bifur-
cated geometry preferred by the C-bonded aldehyde, and
leads to their meshing in this complementary manner, caus-
ing mild Lewis acid activation of the aldehyde specifically in
the SN conformation. For the N-Ts protected series, diaste-
reoselectivity is again largely understandable in terms of the
systemsꢀ conformational preferences. Although cis substrates
do not enjoy any kind of selectivity, their trans analogues
afford the addition product with moderate to high syn
(70:30 syn/anti) selectivity, and exceptional stereospecificity
is achieved in the case of N-Ts-2,3-disubstituted aziridine-2-
carboxaldehyde 15 (>99:1 syn/anti). In the trans case, the
SX (i.e., syn N invertomer and exo aldehyde rotamer) con-
former dominates, but only weakly; in the 2,3-disubstituted
case, the additional methyl substituent alpha to the carbonyl
more definitively locks the SX conformer, enforcing syn-se-
lective addition by attack on the aldehyde face away from
the Ts group.
[7] a) B. C. Kim, W. K. Lee, Tetrahedron 1996, 52, 12117–12124; b) C. S.
Park, H. G. Choi, H. Lee, W. K. Lee, H.-J. Ha, Tetrahedron: Asym-
metry 2000, 11, 3283–3292.
[8] J. L. Pierre, H. Handel, P. Baret, Tetrahedron 1974, 30, 3213–3223.
[9] a) Spartan ꢃ08, Wavefunction, Irvine; b) Y. Shao, L. F. Molnar, Y.
Jung, J. Kussmann, C. Ochsenfeld, S. T. Brown, A. T. B. Gilbert,
L. V. Slipchenko, S. V. Levchenko, D. P. OꢀNeill, R. A. DiStasio,
R. C. Lochan, T. Wang, G. J. O. Beran, N. A. Besley, J. M. Herbert,
C. Y. Lin, T. Van Voorhis, S. H. Chien, A. Sodt, R. P. Steele, V. A.
Rassolov, P. E. Maslen, P. P. Korambath, R. D. Adamson, B. Austin,
J. Baker, E. F. C. Byrd, H. Dachsel, R. J. Doerksen, A. Dreuw, B. D.
Dunietz, A. D. Dutoi, T. R. Furlani, S. R. Gwaltney, A. Heyden, S.
Hirata, C. P. Hsu, G. Kedziora, R. Z. Khalliulin, P. Klunzinger,
A. M. Lee, M. S. Lee, W. Liang, I. Lotan, N. Nair, B. Peters, E. I.
Proynov, P. A. Pieniazek, Y. M. Rhee, J. Ritchie, E. Rosta, C. D.
Sherrill, A. C. Simmonett, J. E. Subotnik, H. L. Woodcock, W.
Zhang, A. T. Bell, A. K. Chakraborty, D. M. Chipman, F. J. Keil, A.
Warshel, W. J. Hehre, H. F. Schaefer, J. Kong, A. I. Krylov, P. M. W.
Gill, M. Head-Gordon, Phys. Chem. Chem. Phys. 2006, 8, 3172–
3191.
Thus, extremely high syn selectivity can be achieved with
cis substrates utilizing Bn, trans substrates utilizing Boc, and
2,3-disubstituted aziridine-2-carboxaldehyde substrates bear-
ing Ts as the nitrogen protective groups. The corresponding
anti adducts can easily be accessed by Mitsunobu inversion,
or by complementary hydride reduction of the related ke-
tones, guided by the early conformational and synthetic
analyses of Pierre et al., augmented by the more detailed
conformational insights developed here with the aid of
quantum chemical modeling.
[10] A. D. Becke, J. Chem. Phys. 1993, 98, 1372–1377.
[11] P. C. Hariharan, J. A. Pople, Theor. Chim. Acta 1973, 28, 213–222.
[12] A. V. Marenich, R. M. Olson, C. P. Kelly, C. J. Cramer, D. G. Truh-
lar, J. Chem. Theory Comput. 2007, 3, 2011–2033.
[13] L. A. Curtiss, P. C. Redfern, K. Raghavachari, V. Rassolov, J. A.
Pople, J. Chem. Phys. 1999, 110, 4703–4709.
[14] a) M. J. Aroney, K. E. Calderbank, H. J. Stootman, J. Chem. Soc.
Perkin Trans. 2 1973, 1365–1368; b) M. Pelissier, A. Serafini, J. De-
vanneaux, J. F. Labarre, J. F. Tocanne, Tetrahedron 1971, 27, 3271–
3284; c) A. D. Walsh, Nature 1947, 159, 712–713.
12338
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12326 – 12339

Selectivity in the Addition Reactions of Organometallic Reagents
FULL PAPER
[15] These structural and electronic issues will be explored in detail in a
future manuscript.
[28] H. Kessler, G. Zimmermann, H. Forster, J. Engel, G. Oepen, W. S.
Sheldrick, Angew. Chem. 1981, 93, 1085–1086; Angew. Chem. Int.
Ed. Engl. 1981, 20, 1053–1055.
[16] a) P. H. M. Delanghe, M. Lautens, Tetrahedron Lett. 1994, 35, 9513–
9516; b) Y. Kazuta, H. Abe, T. Yamamoto, A. Matsuda, S. Shuto,
Tetrahedron 2004, 60, 6689–6703; c) C. Mioskowski, S. Manna, J. R.
Falck, Tetrahedron Lett. 1983, 24, 5521–5524; d) S. Ono, S. Shuto,
A. Matsuda, Tetrahedron Lett. 1996, 37, 221–224.
[17] Y. Kazuta, H. Abe, A. Matsuda, S. Shuto, J. Org. Chem. 2004, 69,
9143–9150.
[18] J. L. Pierre, H. Handel, P. Baret, Org. Mag. Res. 1972, 4, 703–707.
[19] H. Handel, P. Baret, J. L. Pierre, C. R. Acad. Sci. C Chim. 1973, 276,
511–514.
[20] J. R. Durig, S. Y. Shen, Spectrochim. Acta A Mol. Biomol. Spectrosc.
2000, 56, 2545–2561.
[21] R. Barlet, P. Baret, H. Handel, J. L. Pierre, Spectrochim. Acta Part
A Mol. Spectrosc. 1974, 30, 1471–1485.
[22] All of the spectroscopic data of the isolated compound 6a matches
with the syn addition product, as reported previously in the litera-
ture, see: Ref. [4a].
[23] H.-Y. Noh, S.-W. Kim, S. I. Paek, H.-J. Ha, H. Yun, W. K. Lee, Tetra-
hedron 2005, 61, 9281–9290.
[29] H. B. Buergi, J. D. Dunitz, Acc. Chem. Res. 1983, 16, 153–161.
[30] M. Bhupathy, T. Cohen, Tetrahedron Lett. 1985, 26, 2619–2622.
[31] Chelation of the Boc carbonyl with divalent magnesium has been in-
voked previously in a number of publications; for an example, see:
P. Merino, A. Lanaspa, F. L. Merchan, M. Tejero, Tetrahedron Lett.
1997, 38, 1813–1816. To our knowledge these authors have not re-
ported confirmatory experiments by using, for example, TMEDA or
other complexants to address the chelation possibility.
[32] a) F. A. L. Anet, R. D. Trepka, D. J. Cram, J. Am. Chem. Soc. 1967,
89, 357–362; b) H. Kessler, Angew. Chem. 1970, 82, 237–253;
Angew. Chem. Int. Ed. Engl. 1970, 9, 219–235; c) E. R. Johnston,
Mag. Res. Chem. 1995, 33, 664–668; d) H. Nakanishi, O. Yamamoto,
Tetrahedron 1974, 30, 2115–2121; e) J. D. Andose, J. M. Lehn, K.
Mislow, J. Wagner, J. Am. Chem. Soc. 1970, 92, 4050–4056; f) S. J.
Brois, J. Am. Chem. Soc. 1967, 89, 4242–4243; g) J. B. Lambert, B. S.
Packard, W. L. Oliver, J. Org. Chem. 1971, 36, 1309–1310; h) H. A.
Dabbagh, A. R. Modarresi-Alam, J. Chem. Res. 2000, 2000, 190–
192.
[33] P. H. M. Budzelaar, gNMR 5.0.6.0, NMR Simulation Program, Ivor-
ySoft, 2006.
[34] The calculated Gibbs energy of activation is a measure of all pro-
cesses that contribute to the observed line broadening of the alde-
hyde resonances and, thus, is not solely a measure of the nitrogen
[24] B.-F. Li, M.-J. Zhang, X.-L. Hou, L.-X. Dai, J. Org. Chem. 2002, 67,
2902–2906.
[25] E. L. Eliel, S. H. Wilen, M. P. Doyle, Basic Organic Stereochemistry,
Wiley-Interscience, New York, 2001.
[26] R. Bartnix, S. Lesniak, A. Laurent, Tetrahedron Lett. 1981, 22,
4811–4812.
ꢀ
inversion. In fact, it is anticipated that rotation about the C3 CO
bond in 13 contributes to the calculated barrier.
[27] The spectroscopic data of the alcohol matches that of the syn com-
pound reported by Righi and co-workers in their previous study,
see: Ref. [4b].
Received: April 16, 2011
Published online: September 16, 2011
Chem. Eur. J. 2011, 17, 12326 – 12339
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12339