Gas-phase enantioselectivity of chiral amido[4]resorcinarene receptors

Article abstract of DOI:10.1002/chem.200600102

Diastereomeric proton-bound [1_LHA]⁺ complexes between selected amino acids (A = phenylglycine (Phg), tryptophan (Trp), tyrosine methyl ester (TyrOMe), threonine (Thr), and allothreonine (AThr)) and a chiral amido[4]resorcinarene receptor (1_L) display a significant enantioselectivity when undergoing loss of the amino acid guest A by way of the enantiomers of 2-aminobutanes (B) in the gas phase. The enantioselectivity of the B-to-A displacement is ascribed to a combination of thermodynamic and kinetic factors related to the structure and the stability of the diastereomeric [1_LHA]⁺ complexes and of the reaction transition states. The results of the present and previous studies allow classification of the [1_LHA]⁺ complexes in three main categories wherein: i) guest A does not present any additional functionalities besides the amino acid one (alanine (Ala), Phg, and phenylalanine (Phe)) ; ii) guest A presents an additional alcohol function (serine (Ser), Thr, and AThr); and iii) guest A contains several additional functionalities on its aromatic ring (tyrosine (Tyr), TyrOMe, Trp, and 3,4-dihydroxyphenylalanine (DOPA)). Each category exhibits a specific enantioselectivity depending upon the predominant [1 _LHA]⁺ structures and the orientation of the 2-aminobutane reactant in the relevant adducts observed. The results may contribute to the understanding of the exceptional selectivity and catalytic properties of enzyme mimics towards unsolvated biomolecules.

Full text of DOI:10.1002/chem.200600102

DOI: 10.1002/chem.200600102
Gas-Phase Enantioselectivity of Chiral Amido[4]resorcinarene Receptors
Bruno Botta,[a] Fabiana Caporuscio,[a] Ilaria DꢀAcquarica,[a] Giuliano Delle Monache,[a]
Deborah Subissati,[a] Andrea Tafi,*[b] Maurizio Botta,[b] Antonello Filippi,[a] and
Maurizio Speranza*[a]
Abstract:
Diastereomeric
proton-
[1LHA]+ complexes and of the reaction
transition states. The results of the
present and previous studies allow clas-
sification of the [1LHA]+ complexes in
three main categories wherein: i) guest
A does not present any additional
functionalities besides the amino acid
one (alanine (Ala), Phg, and phenylala-
nine (Phe)); ii) guest A presents an ad-
ditional alcohol function (serine (Ser),
Thr, and AThr); and iii) guest A con-
tains several additional functionalities
on its aromatic ring (tyrosine (Tyr),
TyrOMe, Trp, and 3,4-dihydroxyphenyl-
alanine (DOPA)). Each category exhib-
its a specific enantioselectivity depend-
ing upon the predominant [1LHA]+
structures and the orientation of the 2-
aminobutane reactant in the relevant
adducts observed. The results may con-
tribute to the understanding of the ex-
ceptional selectivity and catalytic prop-
erties of enzyme mimics towards unsol-
vated biomolecules.
bound [1LHA]+ complexes between se-
lected amino acids (A=phenylglycine
(Phg), tryptophan (Trp), tyrosine
methyl ester (TyrOMe), threonine
(Thr), and allothreonine (AThr)) and a
chiral amido[4]resorcinarene receptor
(1L) display a significant enantioselec-
tivity when undergoing loss of the
amino acid guest A by way of the
enantiomers of 2-aminobutanes (B) in
the gas phase. The enantioselectivity of
the B-to-A displacement is ascribed to
a combination of thermodynamic and
kinetic factors related to the structure
and the stability of the diastereomeric
Keywords: enantioselectivity · gas-
phase reactions · host–guest sys-
tems · mass spectrometry · molecu-
lar dynamics
Introduction
Biomolecules are often chiral and many biochemical pro-
cesses show preference of one enantiomer over the other.
The exceptional selectivity of biochemical systems is ascri-
bed to size- and shape-specific interactions between large
“host” molecules and small “guest” species that bind on or
within the host. Usually, the interactions are non-covalent,
typically involving van der Waals, electrostatic, or hydrogen-
bonding attractions hampered by steric repulsions.
[a] Prof. B. Botta, Dr. F. Caporuscio, Dr. I. DꢀAcquarica,
Dr. G. Delle Monache, Dr. D. Subissati, Dr. A. Filippi,
Prof. M. Speranza
Dipartimento degli Studi di Chimica e Tecnologia delle
Sostanze Biologicamente Attive
Università “La Sapienza”
00185 Roma (Italy)
Fax : (+39)06-4991-3602
E-mail: maurizio.speranza@uniroma1.it
Enzymes exhibit amazing enantioselectivities. Their asym-
metric structure is often due to complicated supramolecular
assemblies of biopolymers and, therefore, not easily amena-
ble to detailed studies. For this reason, insight into their be-
havior can be gained by investigating tailor-made simplified
host–guest models under conditions, that is, the gas phase,
mimicking the extensive desolvation of the guest molecule
inside the enzyme cavity.
A number of excellent reviews in the area of gas-phase
chiral recognition have recently appeared, which confirm
the importance and the broad applicability of this growing
field.[1–13] Gas-phase chiral recognition experiments are usu-
[b] Prof. A. Tafi, Prof. M. Botta
Dipartimento Farmaco Chimico Tecnologico
Università di Siena
53100 Siena (Italy)
Fax : (+39)0577-234333
E-mail: tafi@unisi.it
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author: Ion abundances
and kinetic plots of the gas-phase reaction between [1LHA]+
(A=Phg, TyrOMe, Trp, Thr, and AThr) and B (BR: black; BS: red;
AD: open; AL: full). Docking and Molecular Dynamics Simulations:
partial atomic charges, MCMM/MOLS docking and MD graphics.
8096
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 8096 – 8105

FULL PAPER
ally carried out by mass spectrometry through the measure-
ment of: i) the relative abundance of noncovalent diastereo-
meric adducts between a chiral host and the two enantiom-
ers (one isotopically labeled) of a guest;[1,2,14] ii) the relative
stability of diastereomeric adducts by equilibrium measure-
ments[3,6,15,16] or by collision-induced dissociation (CID) ex-
periments (Cooksꢀ kinetic method),[17–25] and iii) the rates of
ion/molecule reactions between diastereomeric adducts and
suitable chiral or achiral reactants.[5,26–35]
Figure 1. Formula and side view of local minimum geometry of flattened-
Our group has focused on the last two types of experi-
ments. The emphasis was mainly put on the understanding
of the fundamental interactions responsible of chiral dis-
crimination by calixarenes, which have been long recognized
as potential enzyme mimics.[36–46] In particular, we used an
electrospray-ionization Fourier-transform ion cyclotron reso-
nance mass spectrometer (ESI-FT-ICR) to carry out a kinet-
ic study on the base-induced displacement reaction (1),
wherein [MHA]+ are proton-bonded diastereomeric com-
plexes between some representative amino acids (A= Ala,
Ser, Leu, Pro, picolinic acid (Pip), Phe, Tyr, DOPA) and
suitable calixarene hosts (M), that is, the amido[4]resorcinar-
ene 1L (Figure 1)[31,32] and most of its isomers,[33] and the
enantiomers of 2-aminobutane (B).
cone 2,8,14,20-tetrakis(l-valinamido)[4]resorcinarene (1L).
the four pendants containing the chiral l-valine group which
may be spatially oriented to generate chiral cavities of dif-
ferent size and shape.[33] In the mentioned ESI-FT-ICR stud-
ies,[31–33] we pointed out that the efficiency of the gas-phase
reaction (1) depends on: i) the configuration of both A and
B; ii) the size and the shape of the chiral host M; and iii) the
basicity of the amino acid guest A and the nature of its func-
tional groups. For instance, the heterochiral [1LHAlaD]+
complex is more reactive than the homochiral [1LHAlaL]+
one, whereas the reverse is true for [1LHSerD]+ and
[1LHSerL]+. These findings have been explained in terms of
the effects of the host asymmetric frame on the structure,
stability, and rearrangement dynamics of the diastereomeric
[MHA]+ complexes and the orientation of the amine reac-
tant B in encounters with them.
þ
þ
½MHAŠ þ B ! A þ ½MHBŠ
ð1Þ
The molecular asymmetry of the selected hosts M is due to
With the aim of elucidating the intrinsic factors, which de-
termine the enantioselectivity in noncovalent host/guest sys-
tems, we now extended the investigation to other amino
acid guests with different structure and functionalities (i.e.,
Phg, Trp, and tyrosine methyl ester (TyrOMe)) and with a
second chiral center (i.e., Thr and allothreonine (AThr)).
The relevant kinetic results will be discussed in the light of
molecular mechanics (MM) calculations and molecular dy-
namics (MD) simulations and compared with those obtained
in previous related studies.[31,32]
Abstract in Italian: I complessi diasteromerici non covalenti
del tipo [1LHA]+,in cui 1L › un ammido[4]resorcinarene
chirale con configurazione definita e A sono gli enantiomeri
di alcuni amminoacidi (A=fenilglicina (Phg),triptofano
(Trp),tirosina metilestere (TyrMOe),treonina (Thr),and al-
lotreonina (AThr)),mostrano unꢀelevata enantioselettività
quando reagiscono con gli enantiomeri del 2-amminobutano
in fase gassosa. Lꢀorigine della misurata enantioselettività ›
attribuita alla combinazione di fattori cinetici e termodinami-
ci ed › principalmente determinata dagli effetti della struttura
asimmetrica dellꢀammido[4]resorcinarene sulla struttura e la
stabilità dei complessi diasteromerici [1LHA]+ e delle relative
strutture di transizione nel processo di sostituzione dellꢀam-
minoacido da parte del 2-amminobutano. Lo studio permette
di classificare i complessi [1LHA]+ in tre categorie principali
dove: i) lꢀamminoacido non presenta nessun gruppo funzio-
nale ulteriore (alanina (Ala),Phg,and fenilalanina (Phe)); ii)
lꢀamminoacido presenta anche una funzione alcoolica (serina
(Ser),Thr,and AThr); e iii) lꢀamminoacido presenta anche
gruppi funzionali nellꢀanello aromatico (tirosina (Tyr),
TyrOMe,Trp,and 34,-diidrossifenilalanina (DOPA)). Cia-
scuna categoria mostra una enantioselettività specifica deter-
minata da una struttura predominante per relativo complesso
[1LHA]+ e dallꢀorientamento del reagente amminico. I risul-
tati ottenuti contribuiscono allo sviluppo di un modello dina-
mico per il riconoscimento chirale di biomolecole da parte di
enzimi artificiali allo stato non solvatato.
Materials andMethods
Materials: Enantiomerically pure 1L, in the flattened-cone conformation,
was synthesized and purified according to established procedures.[47] The
d- and l enantiomers of the amino acids A (A=Phg, Trp, Tyr, Thr, and
AThr) were purchased from a commercial source (Aldrich Co.) and used
without further purification. The same company provided (R)-(À)- (BR)
and (S)-(+)-2-butylamine (BS), which were purified in the vacuum with
several freeze–thaw cycles. A simple and safe procedure was used to pre-
pare the pure enantiomers of TyrOMe: acetyl chloride (5 mL) was added
dropwise to cooled dry methanol (50 mL; T=08C), in which the amino
acid (1.7 mmol) has been dissolved. The mixture was heated under reflux
for about 2 h, and then evaporated to dryness.
FT-ICR Experiments: The experiments were carried out as described
elsewhere.[31,32] In particular, they were performed at room temperature
in an APEX 47e FT-ICR mass spectrometer equipped with an ESI
source (Bruker Spectrospin) and a resonance cell (“infinity cell”) situated
between the poles of a superconducting magnet (4.7 T). Stock solutions
of 1L (1”10À5 m) in H2O/CH3OH 1:3, containing a fivefold excess of the
appropriate amino acid A, were electrosprayed through a heated capilla-
ry (1308C) into the external source of the FT-ICR mass spectrometer.
Chem. Eur. J. 2006, 12, 8096 – 8105
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8097

A. Tafi, M. Speranza et al.
All the formed ions were transferred into the resonance cell by a system
of potentials and lenses and quenched by collisions with methane pulsed
into the cell through a magnetic valve. Abundant signals, corresponding
to the natural isotopomers of the proton-bound complex [1LHA]+, were
monitored and isolated by broad-band ejection of all the accompanying
ionic species. The [1LHA]+ family was then allowed to react with the
chiral amine B present in the cell at a fixed pressure whose value ranges
from 2.3”10À8 to 1.3”10À7 mbar depending upon its reactivity.
Computational details: All computational calculations were carried out
on a SGI Origin 300 server and visualized on a SGI Octane workstation.
Molecular mechanics (MM) calculations (docking) and molecular dynam-
ics simulations (MD) were performed using the AMBER* force field as
implemented in MacroModel 5.5.[48] No cutoff was applied for the non-
bonded interactions and the calculations were performed in the gas
phase selecting the constant dielectric treatment (dielectric constant e=
1.0). Partial atomic charges to be used in the docking and MD simula-
tions were obtained through quantum-mechanics calculations performed
with the semiempirical program MOPAC distributed by Accelrys Inc.,
using the AM1-Mulliken method. The atomic partial charges used for
Trp, Tyr and TyrOMe are reported in the Supporting Information and
are specified in the last column of the three Cartesian coordinate struc-
tures (.pdb format generated by InsightII) describing the lowest energy
conformation of their l enantiomer. The partial charges used for [1L·H]+
have already been reported elsewhere.[32]
Figure 2. a) Structural descriptor (SD) defined to classify the output
MCMM/MOLS docking geometries and to graph MD simulations for
[1LHA]+; b) Structural descriptor (SD) defined to graph MD simulations
Some insights in the structure and the dynamics of the proton-bound
[1LHA]+ complexes were obtained by using two computational method-
ologies, that is, i) the statistical Monte Carlo Multiple Minimum
(MCMM) conformational search of A and [1LH]+, coupled with random
rototranslations of the A guest (MOLS command) relative to the [1LH]+
host standing still in the 3D space (MCMM/MOLS docking); and ii) con-
stant temperature MD runs. Although MCMM/MOLS is indeed a well
described and validated docking protocol,[32] nevertheless we are aware
that the number of rotatable bonds moved in each MCMM/MOLS run
(more than 20) is well over the maximum allowed to guarantee exhaus-
tive searches. This problem can be solved by combining docking studies
with MD simulations so as to ensure a complete and reasonably homoge-
neous sampling of the whole potential energy hypersurface of our sys-
tems. Each MCMM/MOLS docking run was made of 20000 steps. With
A=Tyr and TyrOMe, a total number of 25 torsional degrees of freedom
was analyzed, while with A=Trp the number of rotatable bonds was 24.
In the docking calculations, the rototranslations of A with respect to 1L
were limited by the maximum values of 1808 for the rotational angle and
of 3 Š for the translational movement. At the same time, a randomly var-
iable number of rotatable bonds of the side chains of 1L (the flexibility of
the resorcarene skeleton was not directly sampled)[47] and A, ranging
from 2 to NÀ1 (N represents the overall number of variable torsion
angles defined in the command file), was subjected to random step varia-
tions in the range 60–1808. Energy minimizations were performed using
the Truncated Newton conjugate gradient (TNCG) procedure and were
terminated when the energy gradient root mean square (rms) fell below
À
for the rotation of the Cring Ca bond of the high-energy up complex be-
tween A and a simplified version of 1L.
cation into the achiral upper rim cavity (henceforth denoted as up). The
SD shown in Figure 2b with A=TyrL as guest describes the rotation of
the Cring Ca bond of the high-energy up complex between guest A and a
À
simplified version of 1L.
Constant temperature MD simulations with generation of the canonical
ensemble were performed at 300 K. Coupling between the temperature
bath and the molecules was updated every 0.2 ps. The equilibration
period was 50 ps for every run, while the total simulation time was 20 ns.
During each trajectory, 5000 structures (frames) were sampled at regular
intervals throughout the time course, which were graphed relating SD to
the frame number. Each recognition simulation (both docking and mo-
lecular dynamics) was repeated few times starting from different arbitra-
ry geometries to produce a complete sampling of the whole potential
energy hypersurface of the selected [1LHA]+ systems. The convergence
of the results guarantees the completeness of the study.
Results andDiscussion
FT-ICR experiments: The pseudo-first-order rate constant
k’ of Reaction (1) was obtained from the slopes of the rele-
vant lnACHTRE(UNG I/I0) versus t plots, where I is the intensity of com-
10 JmolÀ1 À1. To eliminate duplicate conformations, a comparison was
Š
performed on the heavy atoms, selecting 1.0 Š as the maximum allowa-
ble separation between couples of corresponding atoms after superimpo-
sition. All the conformers were saved that differed from the global mini-
mum-energy conformation by less than 20 kJmolÀ1. The overall confor-
mation of each output docking geometry was classified by the values
taken by the ad hoc defined structural descriptor (SD) shown in Figur-
e 2a.
plex [1LHA]+ at the delay time t and I0 is the sum of the in-
tensities of [1LHA]+ and [1LHB]+ (see, for instance,
Figure 3). The corresponding second-order rate constants
k=k’/[B] are denoted according to the configuration of the
leaving amino acid guest A (kD or kL) or to that of the
amine reactant B (kR or kS). Enantioselectivity is defined by
the 1=kD/kL ratio, when referred to the configuration of the
amino acid A, or by the x=kR/kS one, when referred to the
configuration of the amine B. A value of 1>1 indicates that
amine B displaces the d enantiomer of A (AD) faster that
the l enantiomer (AL) from the relevant diastereomeric
[1LHAD]+ and [1LHAL]+ complexes. The opposite is true
when 1<1. A value 1 = 1 corresponds to equal displace-
The SD shown in Figure 2a with A=TyrOMeD as guest describes the in-
termolecular out-of-plane bending defined by picking four atoms as fol-
lows: assuming C2 as the carbon bearing the protonated side chain, the
four designated atoms were, in sequential order, C28, C25, C27 for the
host (Figure 1) and the chiral carbon atom for the guest. The conse-
quence of this choice is that SD values centered around 08 correspond to
an external lower rim location of guest A in proximity of the protonated
pendant of the host (henceforth denoted as ext). Values around À908 sug-
gest a lower rim location of the guest, among the pendants of the hosting
resorcin[4]arene (henceforth denoted as down), and values of +908 its lo-
8098
www.chemeurj.org
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 8096 – 8105

Gas-Phase Enantioselectivity
FULL PAPER
Table 1. Percent distribution of isomeric [1LHA]+ structures.
Guest (A)
A
[1LHA]+
ACHTREUNG
slow
fast
d-TyrOMe
l-TyrOMe
d-Trp
l-Trp
d-DOPA
l-DOPA
20Æ3
23Æ2
32Æ5
35Æ3
20Æ4
19Æ3
80Æ3
77Æ2
68Æ5
65Æ3
80Æ4
81Æ3
listed in Table 2. Their values, compared with the relevant
collision rate constant (kcoll),[49] provide a measure of the ef-
ficiency of the reaction (eff=k/kcoll). Analysis of Table 2
confirms the view that the efficiency of the gas-phase Reac-
tion (1) depends on the configuration of both A and B and
that the corresponding enantioselectivity factors 1 and x are
dramatically affected by the functional groups present in the
amino acid guests. Indeed, those containing an OH function-
ality, such as Tyr, TyrOMe, DOPA, Ser, Thr, and Athr,
confer to the corresponding [1LHA]+ complexes a 1ꢀ1
factor, while 1ꢁ1 factors are invariably measured for the
complexes with A=Phg, Phe and Ala. The diastereomeric
[1LHA]+ (A=Trp) complexes do not follow an univocal
trend, since they display 1ꢁ1 factors in the reaction with
BR and 1<1 factors in that with BS. This opposite enantiose-
lectivity is reflected in the corresponding x terms which ex-
hibit the largest variation ever measured for this kind of re-
actions (x=0.37–0.95). Comparison of the kinetic results of
[1LHA]+ (A=Tyr, TyrOMe) indicates that the simple meth-
ylation of the carboxyl function of the guest induces a dra-
matic change in the kinetics and the selectivity of Reaction
(1). In fact, while [1LHTyr]+ follow a mono-exponential
decay curve (1ꢀ1; x<1), [1LHTyrOMe]+ exhibit bi-expo-
nential kinetics (1<1; x>1) (see Supporting Information
and Table 2). Significant enantioselectivity differences have
also been noticed for [1LHA]+ (A=Ser, Thr, and AThr).
Indeed, the complexes with A=AThr appear approximately
as selective as those with A=Ser, but appreciably more se-
lective than those with A=Thr, thus indicating that the
presence and the configuration of a further chiral center
may play a significant role on the reaction (1) efficiency.
The kinetic results of Table 2, in particular the opposite 1
values measured with [1LHTrp]+, confirms previous conclu-
sions that the observed enantioselectivities are determined
by a combination of kinetic and thermodynamic factors, that
is, by the relative stability of the diastereomeric [1LHA]+
complexes and by the effects of the resorcin[4]arene frame
upon the transition structures involved in their displacement
reactions (1). In this view, any attempt of rationalizing the
kinetic results of Table 2, in particular the bi-exponential ki-
netics observed with [1LHA]+ (A = TyrOMe, Trp, and
DOPA) and the relative enantioselectivities, requires a de-
tailed structural and energetic analysis of the relevant
proton-bonded diastereomeric [1LHA]+ reactants. This task
has been undertaken by using MM calculations and MD
simulations.
Figure 3. a) Kinetic plot for the gas-phase reaction between (S)-(+)-2-bu-
tylamine (pB =2.9”10À8 mbar) and [1LHPhgD]+; b) kinetic plot for the
gas-phase reaction between (R)-(À)-2-butylamine (pB =1.1”10À7 mbar)
and [1LHTrpD]+.
ment rates. Analogously, a value of x > 1 indicates that the
displacement of the A guest from a given [1LHA]+ diaster-
eomer is faster with the (R)-amine (BR) than with the (S)-
one (BS). Again, the opposite is true when x<1. A value x
= 1 corresponds to equal displacement rates.
Irrespective of the configuration of B, linear rate plots are
observed with A=Phg, Thr, and AThr (corr. coeff. 0.985 <
r2 < 0.998; see, for instance, Figure 3a), whereas reaction
(1) follows bi-exponential kinetics with A=TyrOMe and
Trp (see, for instance, Figure 3b). As pointed out in related
studies,[5,28,31–33] the bi-exponential kinetics is consistent with
the occurrence of two stable isomeric [MHA]+ structures,
one less reactive ([MHA]+slow) and the other more reactive
([MHA]+fast). The time dependence of [MHA]+ (e.g. solid
fast
circles in Figure 3b) can be inferred from the overall
[MHA]+ decay (e.g. open squares in Figure 3b) after sub-
tracting the first-order decay of [MHA]+
(e.g. upper line
slow
in Figure 3b). The two isomeric structures react with the
amine B at rates differing by a factor ranging from about 6
to over 16. The Y intercepts of the first-order decay of
[MHA]+
and [MHA]+ provide an estimate of their rel-
slow
fast
ative distribution. The relevant data are reported in Table 1
together with those concerning the isomeric [MHA]+ (A=
DOPA) structures.[33]
The second-order rate constants k for all displacement re-
actions of Equation (1), which have been investigated, are
Chem. Eur. J. 2006, 12, 8096 – 8105
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8099

A. Tafi, M. Speranza et al.
Table 2. Exchange rate constants (k”10À10 cm3 moleculeÀ1 sÀ1).
(docking) give steric energies
corresponding to enthalpies at
0 K. Therefore, to get average
Guest (A)
(R)-(À)-C4H9NH2
(S)-(+)-C4H9NH2
k
eff
1
k
eff
1
x
d-Phg
l-Phg
d-Phe
l-Phe
d-Tyr
l-Tyr
4.98Æ0.12
4.86Æ0.10
3.60Æ0.03
2.20Æ0.03
0.86Æ0.02
1.07Æ0.02
1.63Æ0.19
2.61Æ0.30
0.12Æ0.01
0.18Æ0.01
2.28Æ0.08
3.00Æ0.09
0.07Æ0.01
0.10Æ0.01
1.82Æ0.04
1.14Æ0.04
0.15Æ0.01
0.14Æ0.01
7.69Æ0.25
6.87Æ0.15
4.59Æ0.06
5.05Æ0.05
2.89Æ0.04
3.56Æ0.08
1.75Æ0.07
4.35Æ0.09
0.43
0.42
0.32
0.20
0.08
0.10
0.15
0.23
0.01
0.02
0.20
0.27
0.01
0.02
0.16
0.10
0.01
0.01
0.69
0.45
0.41
0.62
0.25
0.30
0.15
0.37
1.02Æ0.05
5.19Æ0.06
4.24Æ0.06
3.56Æ0.04
2.28Æ0.04
1.42Æ0.02
1.36Æ0.02
1.30Æ0.04
1.80Æ0.09
0.08Æ0.01
0.16Æ0.01
1.26Æ0.05
1.82Æ0.20
0.08Æ0.01
0.16Æ0.01
1.92Æ0.04
2.33Æ0.20
0.23Æ0.01
0.37Æ0.07
7.06Æ0.08
5.89Æ0.08
3.70Æ0.06
7.56Æ0.06
2.79Æ0.05
3.65Æ0.13
2.49Æ0.01
4.38Æ0.08
0.45
0.37
0.32
0.21
0.13
0.13
0.12
0.16
0.01
0.01
0.11
0.16
0.01
0.01
0.12
0.16
0.02
0.03
0.63
0.53
0.34
0.68
0.24
0.31
0.21
0.37
1.22Æ0.04
0.96Æ0.03 enthalpies at room tempera-
1.14Æ0.05
1.01Æ0.02
0.96Æ0.03
0.61Æ0.02
0.79Æ0.02
1.25Æ0.19
1.45Æ0.25
1.43Æ0.15
1.17Æ0.11
1.81Æ0.12
ture, more appropriate for large
and flexible systems, a constant
temperature MD simulation is
needed.
Previous experimental and
MCMM simulations revealed
that the energetically most fav-
oured [1LHA]+ structures are
1.64Æ0.04
0.80Æ0.04
0.62Æ0.17
0.63Æ0.06
0.76Æ0.04
0.73Æ0.13
1.60Æ0.11
1.07Æ0.13
1.52Æ0.29
0.67Æ0.02
0.81Æ0.03
0.40Æ0.03
1.56Æ0.04
1.04Æ0.03
0.72Æ0.10
0.51Æ0.07
0.69Æ0.11
0.81Æ0.19
0.82Æ0.10
0.62Æ0.18
1.20Æ0.11
0.49Æ0.01
0.78Æ0.05
0.57Æ0.01
ACHTREUNG
ACHTREUNG
ACHTREUNG
ACHTREUNG
ACHTREUNG
T
1.65Æ0.24 invariably obtained by locating
N
1.14Æ0.28
1.27Æ0.25
0.95Æ0.04
0.49Æ0.06
0.65Æ0.04
0.37Æ0.09
1.09Æ0.26
0.86Æ0.17
1.24Æ0.05
0.91Æ0.02
the proton between the amido
groups signalled in Figure 1 by
the broken circles. Indeed, loca-
tion of the proton on one of the
other amido groups is calculat-
ed to cost over 17 kJmolÀ1
more at 300 K.[32]
ACHTREUNG
ACHTREUNG
ACHTREUNG
ACHTREUNG
ACHTREUNG
d-Ala
l-Ala
d-Ser
l-Ser
d-Thr
l-Thr
d-Athr
l-Athr
The results of docking calcu-
1.03Æ0.04 lations for the homochiral
[1LHTyrL]+
and
0.97Æ0.06
0.70Æ0.03
0.99Æ0.04
[1LHTyrOMeL]+ complexes are
illustrated in Figure 4a and b,
respectively. Similar patterns
have been obtained with the
Kinetics andenantioselectivity of [1 LHA]+ (A=Tyr,
TyrOMe) complexes: As pointed out in a previous paper,[32]
a quantitative estimate of the [1LHA]+ energetics by MM
calculations and MD simulations is prevented because of
the high flexibility of the 1L host which makes the number
of its conceivable adducts with the selected amino acids ex-
ceedingly large. Besides, any computational attempt to re-
produce quantitatively the small activation free energy dif-
ferences derived from the measured enantioselectivity
values (<2.5 kJmolÀ1 at 300 K; Table 2) is thwarted by the
relatively large uncertainty associated with the computation-
al approaches. As a consequence, we can just provide a
qualitative description of the structural features of the
[1LHA]+ adducts which, however, may reveal very helpful
in rationalizing the FT-ICR experimental results. This is the
case of the diastereomeric [1LHTyr]+ and [1LHTyrOMe]+
complexes, the first ones exhibiting mono-exponential kinet-
ics with 1 and x below unit and the latter ones characterized
by bi-exponential kinetics with 1<1 and x>1.
corresponding heterochiral adducts. The distribution of the
SD values shown in Figure 4a indicates that, at 0 K, TyrL
As reported in the Computational Details section, a com-
plete and reasonably homogeneous sampling of the whole
potential energy hypersurface of the selected [1L·H·A]+ sys-
tems requires the combination of docking studies with MD
simulations. The time evolution of the molecular motions of
the host and guest moieties, in fact, is expected to let the
system move among many conformations populated at room
temperature, by crossing over low energy barriers and by fa-
vouring large flat minima with respect to narrow ones, so as
to provide a dynamic picture of the recognition process. Fur-
thermore, energy minimizations in molecular mechanics
Figure 4. Docking of a) [1LHTyrL]+ and b) [1LHTyrOMeL]+.
8100
www.chemeurj.org
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 8096 – 8105

Gas-Phase Enantioselectivity
FULL PAPER
preferentially occupies two specific regions of [1LH]+. In
fact, the lowest-energy MCMM/MOLS host–guest com-
plexes, collected within 20 kJmolÀ1, exhibit a clear prefer-
ence of TyrL for regions of three-dimensional space at SD
values ranging around +908 (the up region) or À408. This
latter value corresponds to a structure with the TyrL guest
placed halfway from both the down (SD=À908) and ext
(SD=08) regions of [1LH]+ complex. This position is hence-
forth denoted as the half region. The distribution of the SD
values shown in Figure 4b suggests that both the up and the
half regions of [1LH]+ are mostly suited for hosting TyrO-
MeL as well. In the latter case, TyrOMeL appears to be lo-
cated at the most frequent SD value of about À458, instead
of the À408 value derived for the TyrL guest (Figure 4b).
This small difference indicates that, relative to TyrL, the
TyrOMeL guest is located slightly more inwards the host
cavity and more surrounded by its chiral pendants (Figure
S13a and b, Supporting Information).
Molecular dynamics simulations on [1LHTyrL]+ suggest
that, at 300 K, the amino acid guests can be permanently
trapped nearby the chiral lower rim of [1LH]+ (the half
region; Figure 5b) as well as on its apparently achiral upper
rim (the up region; Figure 5a). Similar patterns have been
observed with the up and half structures of the diastereo-
meric [1LHTyrOMe]+ complexes as well (Figure S14 of Sup-
porting Information). The half and up regioisomers of either
[1LHTyrOMeL]+ and [1LHTyrOMeD]+ exhibit almost the
same average potential energy (within 1 kJmolÀ1), when
scaled to 300 K. In contrast, the average potential energy of
the up regioisomer of [1LHTyrL]+ exceeds that of the half
regioisomer by 9 kJmolÀ1, in agreement with the largely pre-
dominant occurrence of the latter structure under the ESI-
FT-ICR conditions.
The results of docking calculations and MD simulations
on the half regioisomers of [1LHTyrL]+ (Figure 5b) and
[1LHTyrOMe]+ can be compared to those carried out in pre-
vious studies on the diastereomeric [1LHDOPA]+ com-
plexes.[32] In this case, several recognition geometries were
found, characterized by SD values between À90 and 08.
Moreover, it was observed that, during MD simulations
starting from ext geometries (SDꢂ08), DOPA moved to-
wards the lower rim of [1L·H]+ (the down region). By this
movement, the two phenolic OH of DOPA could form hy-
drogen bonds with the polar amido groups of the host at the
lower rim. This final structure somewhat resembles the
global minimum geometry of the complex [1LHTyrL]+ and
[1LHTyrOMeL]+ complexes, shown in Figure S13 of the Sup-
porting Information. Differently from the case of DOPA,
however, the “canyon-shaped” architecture adopted by the
host appears to be less sharp when interacting with TyrL and
the TyrOMe enantiomers, probably due to the lack of the
meta OH group in these guests which may allow the location
of the guest slightly outside the host chiral cavity (SD-
Figure 5. Molecular dynamics of a) half and b) up regioisomers of
[1LHTyrL]+.
MD simulations have been also applied to the rotatable
+
À
Cring Ca bond of the high-energy up [1LHTyrL] complex; its
structural descriptor (SD) is illustrated in Figure 2b. The re-
sults of the MD runs are reported in Figure 6a together with
the up geometry of the [1LHDOPAL]+ complex, shown in
the inset. The SD values point to the fast interconversion of
two guest geometries at the upper rim region of the host.
Analysis of Figure 6a helps clarifying the reasons why up
TyrL complexes are less stable than to the corresponding
half structures. The main reason can be found in the fact
that the aromatic ring of TyrL lacks of the meta-OH group,
which is present in DOPA. This causes a sort of “head-
down” orientation adopted by TyrL to integrate the p–p aro-
matic stacking at the upper rim of the host, with a fruitful
C
aromOH····OMe interaction (atoms belonging to the guest
are henceforth denoted in italic). As a consequence, the
polar interactions between the amino acid group of the
guest and the upper rim of the host are severely weakened.
Differently from TyrL, the up region of [1LH]+ is almost
as suited as the half one for hosting the TyrOMe enantiom-
ers. Indeed, as pointed out above, at 300 K, both up
[1LHTyrOMe]+ complexes have an average enthalpy very
close to that of the corresponding half complexes (within
1 kJmolÀ1). The results of MD simulations, applied to the
+
À
rotatable Cring Ca bond of the up [1LHTyrOMeL] complex,
are reported in Figure 6b. MD simulations on the up
[1LHTyrOMeD]+ complex lead to the same picture. As for
[1LHTyrL]+ (Figure 6a), the up structures of [1LHTyrOMe]+
are characterized by the p–p aromatic stacking accompanied
by HN-H···OMe and CaromOH···OMe interactions. The evi-
A
(TyrOMe)=À458 vs SD
ACHTRE(UNG DOPA)=
À908) and, therefore, the establishment of more interactions
among the host pendants.
Chem. Eur. J. 2006, 12, 8096 – 8105
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8101

A. Tafi, M. Speranza et al.
The comparable 1<1 values, measured for the reaction
with [1LHTyr]+ and [1LHTyrOMe]+, can be explained by
the remarkable similarity of the corresponding half struc-
À
tures, both characterized by intense C=O···H N and
C
aromOH···O=C intermolecular interactions (Figure S13a,b
of Supporting Information). The significant difference in the
measured x terms (Table 2) is due to the fact that, as point-
ed out before, TyrOMe is located more inwards the host
cavity and surrounded by its chiral pendants than Tyr. As a
consequence, relative to Tyr, displacement of TyrOMe re-
quires that amine B enters more in depth into the host
cavity which, therefore, may exert more the effects of its
asymmetry towards it.
Kinetics andenantioselectivity of [1 LHA]+
ACHTRE(UGN A=Trp) com-
plexes: The computational protocol used for the [1LHA]+
(A=Tyr, TyrOMe) systems has been extended to [1LHTrp]+
as well. The results of docking calculations on the homochi-
ral [1LHTrpL]+ and the heterochiral [1LHTrpD]+complexes
are shown in Figure 7a and b, respectively. The distribution
of the relevant SD values as a function of the energy of
output complexes (<20 kJmolÀ1 over the global minimum)
points to both Trp enantiomers as residing preferentially at
the up and down regions of the [1LH]+ host.
À
Figure 6. Molecular dynamics of the rotatable Cring Ca bond of the up re-
gioisomers of a) [1LHTyrL]+ and b) [1LHTyrOMeL]+.
dent difference in the dynamics of [1LHTyrOMe]+, by one
side, and of [1LHTyrL]+, by the other, is to be attributed to
the fact that, despite the co-operation of HN-H···OMe and
C
aromOH···OMe interactions in all these structures, the
[1LHTyrOMe]+ complexes lack of a free acid group which
allows the establishment of a COOH···OMe interaction to
the expenses of the CaromOH···OMe one. Besides, the stabili-
ty of the up [1LHTyrOMeL]+ complex is enhanced by the
presence of the freely rotating estereal group in TyrOMe
which may establish additional dispersive interactions with
the closest aromatic ring of the host.
The concomitant occurrence of stable half and up re-
gioisomers of the [1LHTyrOMe]+ complexes is consistent
with their bi-exponential reaction kinetics (see Supporting
Information). In analogy with the [1LHDOPA]+ systems[32]
and in qualitative agreement with the computed enthalpy
gap between the up and half [1LHTyrOMe]+ structures, the
most stable half regioisomer is identified as the less reactive
([1LHTyrOMe]+)slow component and amounts to about 80%
of the total ion abundance. The less stable up regioisomer is
obviously identified as the most reactive ([1LHTyrOMe]+)fast
component and accounts for the residual ca. 20% of the
total ion abundance (Table 1). In contrast, the large stability
gap between the half and up regioisomers of the [1LHTyr]+
complexes (9 kJmolÀ1) prevents the formation of the latter
one and, therefore, only the stable half regioisomer is
formed under ESI-FT-ICR conditions. This accounts for the
mono-exponential kinetics followed by the [1LHTyr]+ com-
plexes (see Supporting Information and Table 2).
Figure 7. Docking of a) [1LHTrpL]+ and b) [1LHTrpD]+.
Molecular dynamics simulations starting from the global
minima of [1LHTrpD]+ suggest that the guest can be perma-
nently trapped in the chiral lower rim of the [1LH]+ (the
down region) as well as in its apparently achiral upper rim
(the up region) (Figure S15 of Supporting Information).
8102
www.chemeurj.org
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 8096 – 8105

Gas-Phase Enantioselectivity
FULL PAPER
Similar patterns have been observed with [1LHTrpL]+ as
well (Figure S16 of Supporting Information). The occur-
rence of stable down and up structures for the diastereomer-
ic [1LHTrp]+ complexes accounts for their bi-exponential re-
action kinetics ((Figure S3 and S4 of Supporting Informa-
tion). The average potential energy gaps, scaled to 300 K,
between the down and up regioisomers of [1LHTrpD]+ and
[1LHTrpL]+ amount to 3 and 8 kJmolÀ1, respectively, in
favour of the down structures. The more stable down struc-
tures are associated with the most abundant ([1LHTrp]+)slow
fractions (65–68%; Table 1), while the less stable up one
with the less abundant ([1LHTrp]+)fast components (32–35%;
Table 1).
In spite of the evident differences in the complexation
modes, chiral recognition of DOPA and Trp by [1LH]+ ex-
hibits close similarities. The steric energy difference between
the relevant up and down structures is indeed rather close
and unaffected by the configuration of the amino acid guest
(the global minima of the diastereomeric complexes are
quasi-degenerate). In analogy with DOPA, the [1LH]+ host
adopts a pre-organized “canyon-shaped” architecture by hy-
drogen bondings between adjacent pendants to embody Trp.
Such hydrogen bonds are maintained during the MD runs
and, therefore, the indole moiety of Trp is firmly accommo-
dated in the “canyon-shaped” cleft, where it finds sterically
complementary and electronically appropriate surfaces of
the host. However, according to Figure 8a, the amino acid
groups of TrpD interact with two different chiral pendants of
the host (four hydrogen bonds), while in the case of TrpL,
the indolic NH and the amino acid groups form three hydro-
gen bonds with two different host pendants (Figure 8b).
These different interactions are responsible of the quasi-or-
thogonal orientation of the Trp enantiomers in the down
cavity of [1LH]+. The consequence is a different host/guest
packing for the relevant down
The average enthalpies of complexation of the diastereo-
meric [1LHTrp]+ structures [DDHav
=
(DHav)homo
À
(DHav)hetero], calculated from the difference between the
average combined enthalpy of the two isolated components
and that of the complexes, are estimated to amount to À7
and À2 kJmolÀ1 for the down and up structures, respective-
ly.[50] Let us now consider these stability gaps in the light of
the structures of the relevant global minima (Figure 8).
[1LHTrp]+ structures which
may account for the opposite
1(>1 with BR; <1 with BS) and
the largely different x enantio-
selectivity factors, measured for
([1LHTrp]+)slow (Table 2).
The reasons of the up-trap-
ping of Trp can be certainly as-
cribed to both the low polarity
of the guest and the structure
of its side chain, which allow
the guest to establish a very ef-
fective p–p stacking interaction
with the electron-rich cavity at
the upper rim of the host,
strengthened by a dipolar inter-
action of the NH group, proper-
ly located between the methoxy
groups of two facing aromatic
rings (Figure 8c, d). Such a fa-
Figure 8. Front view of the fully minimized structures of: 1) the down regioisomers of a) [1L·H·TrpD]+ and b) vourable spatial orientation
[1L·H·TrpL]+ complexes; 2) the up regioisomers of c) [1L·H·TrpL]+ and d) [1L·H·TrpD]+ complexes. Hydrogen-
bonding interactions are depicted as dotted lines.
shows close analogies with that
of the complex with DOPA,
wherein the dipolar interaction
of the indole NH functionality
The distribution of SD values shown in Figure 7 suggests
that, in contrast to DOPA, the ext region of [1LH]+ is not
suitable for hosting Trp. This different behavior is probably
caused by the absence of appropriate OH “hooks” on the
Trp aromatic rings. Indeed, their presence in DOPA allows
the formation of high-energy ext structures. The observation
that, like Trp, both Tyr and TyrOMe never reside at the ext
region of [1LH]+ suggests that the presence of only one OH
“hook” on their aromatic ring is not enough for the forma-
tion of the ext structure.
is replaced by that of the meta-OH group (inset of Figure
6a). In this connection, the similar 1 and x values, mea-
sured for the up and down [1LHTrp]+ structures, are rather
surprising since, differently from down regioisomers, the up
ones display a lower-rim cavity which is not appreciably per-
turbed by the presence of the guest (Figure 8c,d). Thus, no
appreciable enantioselectivity (1ꢂ1) would be expected in
the reaction of the [1LHTrp]+ up regioisomers with both
amine enantiomers. In contrast, 1=1.60Æ0.11 (with BR) and
1=0.82Æ0.10 (with BS) (Table 2). However, a closer inspec-
Chem. Eur. J. 2006, 12, 8096 – 8105
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8103

A. Tafi, M. Speranza et al.
tion of Figure 8c,d reveals that the [1LHTrp]+ up structures
do have a molecular chirality due to the bent position of the
guest in the upper-rim cavity of the host. Therefore, despite
the chiral lower-rim cavity of the diastereomeric [1LHTrp]+
up complexes is unaffected by the guest, nevertheless its
presence at the up region of the host generates two supra-
molecular diastereomeric forms which may be responsible
of the observed enantioselectivity. In other words, the enan-
tioselective behaviour of the diastereomeric [1LHTrp]+ up
complexes towards the amine enantiomers leads one to con-
jecture on the occurrence of a heterotropic allosteric effect
in these supramolecular systems.[51]
high-energy transition structure. The relatively small effi-
ciencies, measured for the heterochiral complexes, further
support this view. In this frame, the [1LHSerD]+
>
[1LHThrD]+ > [1LHAThrD]+ efficiency trend (Table 2) may
be regarded as due to increasingly congested transition
structures in the down region of the [1LH]+ host.
Conclusion
The gas-phase base-induced displacement reaction between
the 2-aminobutane enantiomers (B) and the diastereomeric
[1LHA]+ complexes (A=amino acid) exhibits a significant
enantioselectivity which has been discussed in the light of
molecular mechanics calculations and molecular dynamics
simulations. A comprehensive analysis of the experimental
and computational results extends and reinforces previous
conclusions that the enantioselectivity of these reaction are
mainly determined by the effects of the host asymmetric
frame upon the structure and the stability of the diastereo-
meric [1LHA]+ complexes and the orientation of the amine
reactant B in its encounters with them.
Kinetics andenantioselectivity of [1 LHA]+ (A=Thr, AThr)
complexes: As shown in Table 2, the [1LHA]+ (A=Thr,
AThr) systems follow mono-exponential kinetics much like
the [1LHSer]+ one. As pointed out in previous papers,[31,32]
this kinetic behaviour is due to: i) the exclusive formation of
a single stable regioisomer, for example, the down structure
for [1LHSerD]+; or ii) several unstable regioisomers with the
guest moving fast around the host on the reaction time
scale, for example, all [1LHSerL]+ regioisomers rapidly
moving over both the ext and the down regions of the host.
Enantioselectivity of [1LHSer]+ is mainly controlled by the
relative stability of the diastereomeric complexes (1<1) and
by the specific path followed by B for removing serine (x>1
with the down [1LHSerD]+ structure and xꢀ1 with the rap-
idly interconverting [1LHSerL]+ ones). The reaction efficien-
cies and the enantioselectivity parameters, measured with
the [1LHA]+ (A=Thr, AThr) systems, are qualitatively simi-
lar to those obtained with [1LHSer]+, thus suggesting a dis-
tinct prevalence of the stable down regioisomer for the het-
erochiral complexes and of rapidly interconverting struc-
tures for the homochiral ones.
In particular, the [1LHA]+ complexes can be classified in
three broad categories: i) those in which the amino acid
guest does not present any additional functionality besides
the amino acid one (Ala, Phg, and Phe); ii) those in which
the amino acid guest presents an additional alcoholic func-
tion (Ser, Thr, and AThr); and iii) those in which the amino
acid guest contains several additional functionalities on its
aromatic ring (Tyr, TyrOMe, Trp, and DOPA). Reaction (1)
with class i) complexes is characterized by an appreciable
sensitivity to the configuration of A (1>1), but it is essen-
tially unaffected by that of B (xꢂ1). Reaction (1) with class
ii) complexes exhibits an appreciable sensitivity to both the
Comparative analysis of the kinetic results for [1LHA]+
(A=Ser, Thr, AThr) points to an appreciable decrease of
the reaction efficiency in going from [1LHSerD]+ to
[1LHThrD]+ to [1LHAThrD]+ (Table 2), while their homochi-
ral analogues follow a different efficiency order, that is,
[1LHSerL]+ > [1LHAThrL]+ > [1LHThrL]+. These different
trends have to be ascribed to the presence and the configu-
ration of the chiral C(3) center of the amino acid and its ef-
fects on the relevant adduct structures. The limited effect of
the B configuration on the reaction efficiencies of the homo-
chiral [1LHAL]+ (A=SerL, ThrL, AThrL) complexes (0.91ꢀ
xꢀ0.99; Table 2) confirms that the guest may move around
all the host positions, including the ext one. In this case,
amine B can remove the amino acid guest without entering
A and B configuration (1<1; x¼
6
1). Reaction (1) with class
iii) complexes displays a significant sensitivity both to the A
and B configuration which markedly depends on the func-
tionalities of guest A. The origin of these pronounced selec-
tivity differences may be mainly attributed to the effects of
the 1L asymmetric frame upon the structure and the stability
of the diastereomeric transition structures involved in Reac-
tion (1). Thus, class i) enantioselectivities are attributed to
the attack of the amine B on the amino acid guest placed
outside the host in proximity of two adjacent pendants (the
ext region). Class ii) enantioselectivities are determined by
the greater stability of the heterochiral [1LHAD]+ down
complexes, relative to that of their homochiral [1LHAL]+ ext
analogues. In class iii complexes, the amino acid guest may
permanently occupy both the up and down (or half) regions
of the host (A=TyrOMe, Trp, and DOPA) or just the last
one (A=Tyr). Their occurrence accounts for the observa-
tion of bi-exponential reaction kinetics in the first case and
of mono-exponential ones in the latter. The large variations
in class iii) enantioselectivities are attributed to the specific
effects of the A functionalities on its location in the up and
down (or half) regions of the host. The results of this com-
prehensive investigation may be regarded as contributing to
the chiral-lower rim cavity of the host. The [1LHSerL]+
>
[1LHAThrL]+ > [1LHThrL]+ efficiency trend probably re-
flects a reverse [1LHA]+ stability order. On the contrary, the
largely variable effect of the B configuration on the reaction
efficiency of the heterochiral [1LHAD]+ (A=SerD, ThrD,
AThrD) complexes (0.70ꢀxꢀ1.24; Table 1) complies with a
stable down structure for these complexes. Here, amine B
must partially enter the host chiral cavity to oust the amino
acid guest through the involvement of a more congested,
8104
www.chemeurj.org
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 8096 – 8105

Gas-Phase Enantioselectivity
FULL PAPER
[27] J. Ramirez, S. Ahn, G. Grigorean, C. B. Lebrilla, J. Am. Chem. Soc.
2000, 122, 6884–6890.
[28] J. F. Gal, M. Stone, C. B. Lebrilla, Int. J. Mass Spectrom. 2003, 222,
259–267.
the formulation of a model for chiral recognition of biomo-
lecules by enzyme mimics in the unsolvated state.
[29] S. Ahn, J. Ramirez, G. Grigorean, C. B. Lebrilla, J. Am. Soc. Mass
Spectrom. 2001, 12, 278–287.
[30] G. Grigorean, C. B. Lebrilla, Anal. Chem. 2001, 73, 1684–1691.
[31] B. Botta, M. Botta, A. Filippi, A. Tafi, G. Delle Monache, M. Sper-
anza, J. Am. Chem. Soc. 2002, 124, 7658–7659.
[32] A. Tafi, B. Botta, M. Botta, G. Delle Monache, A. Filippi, M. Sper-
anza, Chem. Eur. J. 2004, 10, 4126–4135.
[33] B. Botta, D. Subissati, A. Tafi, G. Delle Monache, A. Filippi, M.
Speranza, Angew. Chem. 2004, 116, 4871–4874; Angew. Chem. Int.
Ed. 2004, 43, 4767–4770.
Acknowledgement
Work supported by the Ministero dellꢀIstruzione dellꢀUniversità e della
Ricerca (MIUR, COFIN), the Consiglio Nazionale delle Ricerche
(CNR) and Istituto Pasteur Fondazione Cenci Bolognetti. B.B. and A.T.
gratefully acknowledge Project “FIRB 2003” from MIUR (Ministero
della Università e della Ricerca). The authors express their gratitude to
F. Angelelli for technical assistance.
[34] A. Filippi, M. Speranza, Int. J. Mass Spectrom. 2000, 199, 211–219.
[35] H. Bagheri, H. Chen, R. G. Cooks, Chem. Commun. 2004, 2740–
2741.
[36] C. D. Gutsche, Calixarenes: Monographs in Supramolecular Chemis-
try (Ed.: J. F. Stoddart), Royal Society of Chemistry, Cambridge,
1989.
[37] C. D. Gutsche, Calixarenes Revisited: Monographs in Supramolecu-
lar Chemistry (Ed.: J. F. Stoddart), Royal Society of Chemistry,
Cambridge, 1998.
[38] “Calixarenes: A Versatile Class of Macrocyclic Compounds”: Topics
in Inclusion Science,Vol. III (Eds.: J. Vicens, V. Bçhmer), Kluwer
Academic Publishers, Dordrecht, The Netherlands, 1991.
[39] V. Bçhmer, Angew. Chem. 1995, 107, 785–818; Angew. Chem. Int.
Ed. Engl. 1995, 34, 713–745.
[1] M. Sawada, Biological Mass Spectrometry: Present and Future (Eds.:
T. Matsuo, R. M. Caprioli, M. L. Gross, T. Seyama), Wiley, New
York, 1994, p. 639–646.
[2] M. Sawada, Mass Spectrom. Rev. 1997, 16, 73–90.
[3] C. A. Schalley, Int. J. Mass Spectrom. 2000, 194, 11–39.
[4] A. Filippi, A. Giardini, S. Piccirillo, M. Speranza, Int. J. Mass Spec-
trom. 2000, 198, 137–163.
[5] C. B. Lebrilla, Acc. Chem. Res. 2001, 34, 653–661.
[6] D. V. Dearden, Y. Liang, J. B. Nicoll, K. A. Kellersberger, J. Mass
Spectrom. 2001, 36, 989–997.
[7] C. A. Schalley, Mass Spectrom. Rev. 2001, 20, 254–309.
[8] M. Speranza, Adv. Phys. Org. Chem. 2004, 39, 147–281.
[9] M. Speranza, Int. J. Mass Spectrom. 2004, 232, 277–317.
[10] M. Speranza, M. Satta, S. Piccirillo, F. Rondino, A. Paladini, A.
Giardini, A. Filippi, D. Catone, Mass Spectrom. Rev. 2005, 24, 588–
610.
[40] A. M. A. van Wageningen, W. Verboom, D. N. Reinhoudt, Pure
Appl. Chem. 1996, 68, 1273–1277.
[41] A. Ikeda, S. Shinkai, Chem. Rev. 1997, 97, 1713–1734.
[42] Calixarenes in Action (Eds.: L. Mandolini, R. Ungaro), Imperial
College Press, London, 2000.
[11] K. A. Schug, W. Lindner, J. Sep. Sci. 2005, 288, 1932–1955.
[12] M. G. Finn, Chirality 2002, 14, 534–540.
[43] Z. Asfari, V. Bçhmer, J. Harrowfield, J. Vicens, Calixarenes 2001,
Kluwer Academic Publishers, Dordrecht, The Netherlands, 2001.
[44] J. M. Harrowfield, W. R. Richmond, A. N. Sobolev, A. H. White, J.
Chem. Soc. Perkin Trans. 2 1994, 5–9.
[13] J. S. Splitter, F. Turecek, Applications of Mass Spectrometry to Or-
ganic Stereoselectivity, VCH, New York, 1994.
[14] E. N. Nikolaev, E. V. Denisov, V. S. Rakov, J. H. Futrell, Int. J. Mass
Spectrom. 1999, 183, 357–368.
[45] R. Ludwig, Fresenius,J. Anal. Chem. 2000, 367, 103–128.
[46] G. Arena, A. Contino, A. Magri, D. Sciotto, G. Spoto, A. Torrisi,
Ind. Eng. Chem. Res. 2000, 39, 3605–3610.
[47] B. Botta, G. Delle Monache, P. Salvatore, F. Gasparrini, C. Villani,
M. Botta, F. Corelli, A. Tafi, E. Gacs-Baitz, A. Santini, C. F. Carval-
ho, D. Misiti, J. Org. Chem. 1997, 62, 932–938.
[15] D. V. Dearden, Y. Liang, J. B. Nicoll, K. A. Kellersberger, J. Mass
Spectrom. 2001, 36, 989–997.
[16] Y. Liang, J. S. Bradshaw, D. V. Dearden, J. Phys. Chem. A 2002, 106,
9665–9671.
[17] K. VØkey, G. Czira, Rapid Commun. Mass Spectrom. 1995, 9, 783–
787.
[48] F. Mohamadi, N. G. J. Richards, W. C. Guida, R. Liskamp, M.
Lipton, C. Caufield, G. Chang, T. Hendrickson, W. C. Still, J.
Comput. Chem. 1990, 11, 440–467.
[49] T. Su, J. Phys. Chem. 1988, 88, 4102–4103; T. Su, J. Phys. Chem.
1988, 88, 5355.
[50] S. Chimin, K. N. Houk, J. Am. Chem. Soc. 1996, 118, 8056–8070.
[51] a) B. Botta, F. Caporuscio, D. Subissati, A. Tafi, M. Botta, A. Filippi,
M. Speranza, Angew. Chem. 2006, 118, 2783–2786; Angew. Chem.
Int. Ed. 2006, 45, 2717-2720.
[18] W. A. Tao, L. Wu, R. G. Cooks, Chem. Commun. 2000, 2023–2024.
[19] W. A. Tao, R. G. Cooks, Anal. Chem. 2003, 75, 25A–31A.
[20] W. A. Tao, D. Zhang, E. N. Nikolaev, R. G. Cooks, J. Am. Chem.
Soc. 2000, 122, 10598–10609.
[21] W. A. Tao, R. G. Cooks, Eur. J. Mass Spectrom. 2002, 8, 107–115.
[22] W. A. Tao, R. L. Clark, R. G. Cooks, Anal. Chem. 2002, 74, 3783–
3789.
[23] W. A. Tao, R. G. Cooks, Angew. Chem. 2001, 113, 779–782; Angew.
Chem. Int. Ed. 2001, 40, 757–760.
[24] W. A. Tao, F. G. Gozzo, R. G. Cooks, Anal. Chem. 2001, 73, 1692–
1698.
[25] L. Wu, R. G. Cooks, Anal. Chem. 2003, 75, 678–684.
[26] J. Ramirez, F. He, C. B. Lebrilla, J. Am. Chem. Soc. 1998, 120, 7387–
7388.
Received: January 23, 2006
Revised: March 28, 2006
Published online: July 31, 2006
Chem. Eur. J. 2006, 12, 8096 – 8105
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8105