Theoretical and X-ray crystallographic evidence of a fluorine-imine gauche effect: An addendum to dunathan's stereoelectronic hypothesis

Article abstract of DOI:10.1002/chem.201100644

The preference of β-fluoroimines to adopt a gauche conformation has been studied by single-crystal X-ray diffraction analysis and DFT methods. Empirical and theoretical evidence for a preferential gauche arrangement around the NCCF torsion angle (φ) is pr

Full text of DOI:10.1002/chem.201100644

DOI: 10.1002/chem.201100644
Theoretical and X-ray Crystallographic Evidence of a Fluorine-Imine Gauche
Effect: An Addendum to Dunathanꢀs Stereoelectronic Hypothesis
Christof Sparr,[a] Evdokiya Salamanova,[b] W. Bernd Schweizer,[a]
Hans Martin Senn,*[b] and Ryan Gilmour*[a]
Abstract: The preference of b-fluoro-
imines to adopt a gauche conformation
has been studied by single-crystal X-
ray diffraction analysis and DFT meth-
ods. Empirical and theoretical evidence
for a preferential gauche arrangement
around the NCCF torsion angle (f) is
presented ((E)-2-fluoro-N-(4-nitroben-
zylidene)ethanamine: fNCCF =70.08). In
the context of this study, the analysis of
was performed, a species that is impli-
cated in the inhibition of pyridoxal
phosphate (PLP)-dependent enzymes
by b-fluoroamine derivatives. The
gauche preference of the internal aldi-
mine (=NCH2CH2F) that can be ra-
tionalized by stereoelectronic argu-
ments does not hold for the corre-
sponding
external
system
(N=
CHCH2F) (Emin when
fNCCF =1208).
ꢀ
Moreover, the C F bond is lengthened
by more than 0.02 ꢁ at fNCCF =ꢁ908,
when it is exactly antiperiplanar to the
conjugated imine. This activation of
Keywords: aldimines
· conforma-
ꢀ
the C F s bond by an adjacent p
tion analysis · density functional
theory · fluorine · gauche effect
system constitutes an addendum to Du-
nathanꢀs stereoelectronic hypothesis.
a
pyridoxal-derived b-fluoroaldimine
Introduction
Dunathanꢀs stereoelectronic model to correlate the confor-
mation and stereospecificity of pyridoxal phosphate (PLP
1)-dependent enzymatic processes remains a milestone dis-
covery in rationalizing the mechanistic intricacies of vitamin
B6 enzymes.[1] Of principal importance is the notion that s
bonds may be activated by an adjacent
p system
(Scheme 1). Consequently, the cofactor-derived Schiff bases
2 that are central to a plethora of enzyme mechanisms, in-
cluding those of decarboxylases and racemases, share a
common topological transition-state dependence on the s
bond being aligned with the p-system of the pyridoxal
imine: in essence, maximum s–p overlap regulates reaction
specificity.[2] Unsurprisingly, this mechanistic understanding
has led to the rational design of numerous small molecule
inhibitors for the treatment of neurodegenerative diseases,
as well as PET imaging agents.[3,4] Many of these pharma-
ceuticals are fluorinated at the b position such that elimina-
tion from the transient quinoid intermediate 3 generates an
Scheme 1. The Dunathan hypothesis to correlate reaction specificity and
conformation in PLP-dependent enzymes.[1]
electron sink 4, which can covalently modify the enzyme
active site, thus rendering it inactive. Pertinent examples of
pharmaceuticals that exploit this design approach include a-
fluoromethyl-DOPA (5), a-difluoromethyl-DOPA and relat-
ed medicinally important compounds.[5,6]
Despite the wealth of literature pertaining to the reactivi-
ty and conformational dynamics of pyridoxal-derived Schiff
bases,[7] including b-fluoroimines,[8] an important conforma-
tional issue remains unaddressed: do b-fluoroimines prefer-
entially adopt a gauche conformation in a manner consistent
with other fluorinated compounds containing a vicinal elec-
tron-withdrawing substituent?[9,10] Herein, we describe the
synthesis and solid state analysis of three representative b-
fluoroimines (6–8), including a vitamin B6-derived aldimine
8, and quantify the empirical findings by theoretical meth-
ods. In view of the prominent role that b-fluoroamines have
[a] C. Sparr, Dr. W. B. Schweizer, Prof. Dr. R. Gilmour
Swiss Federal Institute of Technology (ETH) Zꢂrich
Laboratory for Organic Chemistry
Department of Chemistry and Applied Biosciences
Wolfgang-Pauli-Str. 10, 8093 Zꢂrich (Switzerland)
E-mail: ryan.gilmour@org.chem.ethz.ch
[b] E. Salamanova, Dr. H. M. Senn
WestCHEM and School of Chemistry
University of Glasgow
Glasgow G12 8QQ, Scotland (UK)
Fax : (+44)141-330-4888
E-mail: senn@chem.gla.ac.uk
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201100644.
8850
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 8850 – 8857

FULL PAPER
historically played in the study and inhibition of PLP-depen-
dent enzymes, this study is placed in the context of a wider
theoretical conformational analysis of the transient inter-
mediates involved. The conformational preferences of repre-
sentative aldimine and quinoid intermediates are described
b-fluoroimine 8 (Figure 1, right); a species that is structural-
ly analogous to the transient intermediates implicated in the
disruption of many PLP-dependent enzymatic processes (2,
Scheme 1). Analysis of the solid-state structure of this aldi-
mine 8 reveals a racemic and an enantiomerically pure
form.[13] Both forms have two symmetry-independent mole-
cules in the asymmetric unit within the centrosymmetric tri-
ꢀ
together with a study of C F activation as a function of
fNCCF
.
¯
clinic space group P1 and the polar triclinic space group P1,
respectively. Importantly, a clear gauche effect is observed
(fNCCF =ꢁ678). The orthogonal orientation of the (normally
phosphorylated) primary hydroxyl group relative to the pyr-
idine ring is also noteworthy, and all chemically intuitive hy-
drogen-bonding patterns are satisfied. It is important to
note that in all cases the (E)-configured imines were isolat-
ed exclusively, thus minimizing steric stress, and that the
non-sterically congested systems (6 and 8) adopt conforma-
tions in which the nitrogen lone pair points away from the
fluorine atom to minimize elec-
Results and Discussion
Experimental structures in the solid state
As a start point for this study, b-fluoroimines 6–8 were pre-
pared under dehydrative conditions from the corresponding
aldehyde and 2-fluoroamine, and then evaluated by single-
crystal X-ray diffraction analysis (Figure 1). Imine 6, derived
tronic repulsion. Whereas the
solid state conformations of 6
ꢀ
and 8 clearly show opposing C
ꢀ
F and C N dipoles, this obser-
vation does not hold true for b-
fluoroimine 7, in which the ni-
ꢀ
trogen lone pair and the C F
bond are parallel (Scheme 2).
Conformational analysis
Conformational preferences of
b-fluoroimines: In an attempt to
quantify the gauche conforma-
tional preference of the b-fluo-
roethylimines
described
in
Figure 1, we embarked upon a
Figure 1. X-ray structures of b-fluoroimines 6, 7 and 8.[11–13]
from 2-fluoroethylamine and p-nitrobenzaldehyde (Figure 1,
left), in which there are no dominant steric interactions,
crystallizes in the orthorhombic space group Pna21.[11] While
the interatomic distances and angles are within the expected
range, a clear gauche conformation is observed between the
fluorine and nitrogen centres (fNCCF =ꢁ70.08). Interestingly,
the planar section of the b-fluoroimine 6 forms a p–p inter-
action at a distance of 3.4 ꢁ with a neighboring molecule
translated along the c axis.
To compare this structural analysis with that of a more
sterically demanding system, the valine-derived b-fluoro-
imine 7 was prepared (Figure 1, centre). This b-fluoroimine
crystallizes in the polar space group C2221 with the fluorine
adopting a gauche orientation relative to the imine nitrogen
(fNCCF =ꢀ74.28). The nitrobenzyl group is rotated out of the
imine plane (fNCCC =ꢀ25.58),[12] likely induced by crystal
packing forces where p–p stacking with an adjacent nitro-
benzyl moiety (distance of the two ring centroids=3.75 ꢁ)
is the most dominant interaction. Finally, our attention was
focussed on preparing and analyzing the pyridoxal-derived
Scheme 2. Newman projections of b-fluoroimines 6, 7 and 8.
theoretical study of related systems at the DFT level. Initial-
ly, the NCCF bond rotational profile of the imine derived
from b-fluoroethylamine and acetaldehyde was calculated in
the vacuum and compared with the corresponding de-halo-
genated system (Figure 2; Table 1, entries 1–6). The parent
system shows the usual symmetric rotation profile with
three equivalent minima separated by barriers of about
13 kJmolꢀ1.
In the b-fluoro system (Table 1, entries 4–6), the symme-
try of the bond rotational profile is broken, with the
ꢀgauche conformer being the most stable, followed by anti
(+4 kJmolꢀ1 relative to ꢀgauche) and
+gauche
Chem. Eur. J. 2011, 17, 8850 – 8857
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8851

R. Gilmour, H. M. Senn et al.
17 kJmolꢀ1 (ꢀgauche!anti); the anti!+gauche barrier is
about 13 kJmolꢀ1. These energetic data suggest that all
three conformers are accessible in equilibrium at 298 K,
with relative populations of 1:0.2:0.03 (ꢀgauche/anti/+
gauche).
Based on the electronic origin of the preference for
ꢀgauche over anti, a dependence on the electronic nature of
the imine is expected: the anti conformer should be the
more disfavored the more electron-withdrawing the imine.
To test this hypothesis, we investigated related b-fluoro-
imines derived from selected aldehydes RCHO (R=H, 4-
NO2Ph (6), 4-MeOPh); see Table 1, entries 7–15. The com-
parison revealed only a small effect of the R group on the
conformational preference of the b-fluoroimines. The stabili-
zation of ꢀgauche over anti is slightly more pronounced (6
vs. 3 kJmolꢀ1) for the electron-deficient aromatic imine
(R=4-NO2Ph, Table 1, entries 10–12) as compared with the
electron-rich species (R=4-MeOPh, Table 1, entries 13–15).
Albeit small, this effect is consistent with the stereoelec-
tronic explanation of the gauche effect, whereby the anti ar-
rangement of acceptor bonds is disfavored. In all cases, the
order of stability remains constant: ꢀgauche<anti<+
gauche.
Figure 2. Bond-rotation profiles of acetaldehyde-derived imines (F vs. H),
calculated in vacuum at the M06-2X/6-311G(2df,p) level.
AHCTUNGTRENNUNG
Table 1. Calculated relative conformational energies and torsion angles
of ethylimines derived from selected aldehydes RCOH.[a]
Entry
X
R
Conformer DE [kJmolꢀ1
]
fACTHNGUTERN(UNG NCCX) [8]
1
2
3
4
5
6
7
8
H
H
H
F
F
F
F
F
F
F
F
F
F
F
F
Me
Me
Me
Me
Me
Me
H
H
ꢀgauche
+gauche
anti
ꢀgauche
+gauche
anti
ꢀgauche
+gauche
anti
ꢀgauche
+gauche
anti
0
0
0
0
9
4
0
8
4
0
8
6
0
8
3
ꢀ64
55
176
ꢀ68
69
175
ꢀ68
69
175
ꢀ67
68
Natural bond orbital analyses: The extent of the stereoelec-
tronic gauche effect can be quantified by considering the
stabilization gained from the hyperconjugative donor–ac-
ceptor interaction. Within the framework of the natural
bond orbital (NBO) approach,[15] the interaction energy due
to the delocalization of electron density from a donor NBO
to an acceptor NBO can be calculated from a second-order
perturbation theory expression. NBOs are localized orbitals
constructed such as to provide the most accurate Lewis-like
bonding picture, based on a given N-electron density, which
can be obtained from any electronic-structure method. Spe-
cifically, for the case of the NCCF torsion, the NBO method
9
H
10
11
12
13
14
15
4-NO2Ph (6)
4-NO2Ph (6)
4-NO2Ph (6)
4-MeOPh
4-MeOPh
4-MeOPh
176
ꢀ68
68
ꢀgauche
+gauche
anti
176
[a] Results obtained with M06-2X/6-311GACHTNUTRGNEUNG(2df,p) in vacuum.
(+9 kJmolꢀ1). An explanation for the preference of
ꢀ
ꢀgauche over anti is a stabilizing hyperconjugative sCꢀ
!
yields localized bonding and anti-bonding NBOs for the C
H
ꢀ
ꢀ
s*Cꢀ interaction. To maximize the stabilizing overlap be-
F, C H, and C N bonds, and it quantifies the energy due to
any donor–acceptor interactions between them.
F
tween the strongly electron-accepting s*Cꢀ orbital and the
F
donating vicinal sCꢀ orbital (which is a better donor than
In agreement with previous studies,[9f,h] it was found that
the total donor–acceptor stabilization within the NCH2CH2F
fragment is dominated by interactions between anti-oriented
vicinal bonds. The vicinal gauche interactions are minor and
change little upon rotation; the geminal interactions, al-
though significant, are unaffected by rotation. We therefore
used the sum of the six vicinal anti donor–acceptor interac-
tion energies as a measure for the hyperconjugative stabili-
zation, Edeloc. For instance, for the three conformers of the 4-
nitrobenzaldimine 6, Edeloc amounts to ꢀ66, ꢀ64 and
ꢀ57 kJmolꢀ1 for the ꢀgauche, +gauche and anti conform-
ers, respectively. In the gauche conformers, the single largest
H
ꢀ
ꢀ
sCꢀN), the C F bond is oriented anti to either vicinal C H
bond. Consequently, there is a stereoelectronic preference
ꢀ
ꢀ
for the C F bond being gauche to C N. This deviation from
simple steric and electrostatic arguments is known as the
stereoelectronic “gauche effect”, which is epitomized by 1,2-
difluoroethane.[14] However, the gauche effect does not dif-
ferentiate between the two possible gauche conformers (ꢁ
g). The observed destabilization of the +gauche conformer
in the b-fluoroimines investigated here is a possible conse-
quence of the overriding electrostatic repulsion between the
imine and fluorine lone pairs, as schematically indicated in
Figure 2. Qualitatively and quantitatively similar behavior
has previously been reported for other systems containing
fluorine in the b position with respect to a lone-pair-bearing
heteroatom, such as b-fluoroamines or b-fluoroalcohols.[9c,f,h]
The preferred ꢀgauche minimum is protected by barriers of
around 33 kJmolꢀ1 (ꢀgauche!+gauche) and around
contributor is indeed the sCꢀH! s*Cꢀ interaction, which is
F
worth about ꢀ20 kJmolꢀ1. In 6, the stereoelectronic prefer-
ence for ꢀgauche over anti is therefore DE
ꢀ8.5 kJmolꢀ1.
N
8852
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 8850 – 8857

Evidence of a Fluorine-Imine Gauche Effect
FULL PAPER
gauche preference, due to the less electron-withdrawing
nature of the imine. However, DE
(pKa =8.2). We therefore restricted ourselves to studying
overall neutral forms of 8 as they are prevailing in physio-
logical solutions, and considered a systematic study of the
cations, which may well be relevant inside enzyme active
sites, to be outside the scope of the present work.
(ꢀg, a)=ꢀ8.2 kJmolꢀ1
for the 4-methoxybenzaldimine, insignificantly different
from 6, which hence cannot account for the difference in
gauche preference. We therefore inspected the atomic par-
tial charges derived from the NBO procedure. They are es-
sentially independent of the torsional conformation.
However, in the gauche conformers, the two electronega-
tive atoms F and N, which also bear negative NBO charges,
are relatively close to one another (ꢂ2.85 ꢁ). On electro-
static grounds, the gauche conformers are thus destabilized
relative to the anti conformer. Whereas the NBO charge on
fluorine is the same in both benzaldimines (qF =ꢀ0.40e), the
imine nitrogen is slightly more negatively charged in the
more electron-rich 4-methoxy derivative (qN =ꢀ0.48e) than
in the 4-nitro derivative 6 (qN =ꢀ0.45e). This amounts to a
difference in electrostatic repulsion of about 5 kJmolꢀ1 (also
In both vacuum and water, the enol-imine form 8a is pre-
ferred. This is in agreement with the X-ray structure of 8
(Figure 3), which was unambiguously identified with the
enol-imine form 8a by comparing computed with experi-
mental structural parameters. The keto–enamine tautomer
8b is energetically competitive in water, where its equilibri-
um concentration amounts to about 0.5% at 298 K. The qui-
noid tautomer 8c and the zwitterion 8d, however, are only
present in vanishingly small amounts under equilibrium con-
ditions in either medium. As expected, the zwitterion is
strongly stabilized in the polar solvent as compared with
vacuum. For the low-energy tautomers 8a,b as well as for
the mechanistically important quinoid 8c, we calculated
NCCF bond rotation profiles in water (Figure 3 and
Figure 4) and optimized the torsional conformers (in both
vacuum and water; Table 3).
ꢀ
considering the small change in the F N distance, which is
0.02 ꢁ shorter in 6). The larger gauche preference in 6 com-
pared with the 4-methoxybenzaldimine can thus be ex-
plained by a smaller electrostatic destabilization of the
gauche conformer. It may therefore be concluded from this
study that the electronic nature of the imine (electron rich
vs. electron deficient) does have an effect, albeit small, on
the NCCF torsional preference of b-fluoroimines. The
gauche effect is slightly more pronounced in electron-poor
imines. However, the reason for this is not the larger hyper-
conjugative stabilization, but the reduced electrostatic repul-
sion between the two electronegative atoms F and N.
Conformational preferences of b-fluoropyridoximine: For the
pyridoximine 8, we first considered the relative stabilities of
the possible tautomeric forms with overall neutral charge,
8a–d (Table 2). The focus on neutral forms was motivated
two-fold: Firstly, our structural studies reported above have
been limited to neutral b-fluoroimines. Secondly, recent de-
tailed NMR spectroscopic investigations by Limbach and
co-workers[16] have shown that the pyridine nitrogen of PLP
aldimines is not protonated in aqueous solution near neutral
pH (pKa =5.8). This contrasts with the situation for PLP
itself, which is N-protonated under the same conditions
Figure 3. Bond-rotation profiles of b-fluoropyridoximine tautomers 8a
and 8b, calculated in water at the M06-2X/6-31+GCAHTUNGTRENNUNG
energies of stationary points using the larger 6-311G
ACHTUNGTRENNUNG
differ by 1 kJmolꢀ1 (data not shown).
Table 2. Calculated relative stabilities of tautomeric and protonation forms of neutral
pyridoximine 8 in vacuum and water.[a]
Focussing on 8a,b (Figure 3; Table 3, entries 1–6),
we first note that their torsional preferences are es-
sentially identical. Comparing 8a,b in vacuum to
the b-fluoroacetaldimine (Figure 2; Table 1, en-
tries 4–6) and the b-fluorobenzaldimines (Table 1,
entries 10–15) discussed above, the most obvious
difference is the stabilization of the +gauche con-
former, making the anti conformer now the least
stable one. The order of stability for b-fluoropyri-
doximine 8 is therefore: ꢀgauche<+gauche<anti.
The energy of the anti relative to ꢀgauche in 8a,b
is, however, the same as for the electron-deficient
imine 6.
(ꢀg)-8a
0
(ꢀg)-8b
27
(E)-8c
(Z)-8c
(ꢀg)-8d
115
DE [kJmolꢀ1
]
vacuum
water
76 (ꢀac)
67 (syn)
79 (anti)
71 (anti)
0
13
65
[a] Relative energies of the most stable conformer of each form at the M06-2X/6-
311+G(2df,p) level in vacuum and continuum water, respectively.
ACHTUNGTRENNUNG
Chem. Eur. J. 2011, 17, 8850 – 8857
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8853

R. Gilmour, H. M. Senn et al.
The stabilization of the +gauche conformer in 8a results
from a stronger hydrogen bond in solution; the OH···N dis-
tance reduces by 0.03–0.05 ꢁ upon solvation, which further
mitigates the lone-pair repulsion.
The quinoid tautomer 8c can exist as two stereoisomers
with respect to the configuration about the external (N=Ca)
imine double bond. Depending on which of the two Ca pro-
tons in 8a is abstracted, either (E)- or (Z)-8c is formed. The
most stable conformer of the E isomer is always preferred
(by about 4 kJmolꢀ1) to the most stable conformer of the Z
isomer (Table 3, entries 7–12; Figure 4). It is important to
note that the central bond of the NCCF torsion in 8c is sp2–
sp3, rather than sp3–sp3 as in all the previous cases. This cre-
ates a different steric and stereoelectronic environment,
thus we can expect torsional preferences to differ from the
previously discussed systems.
The rotational profile for the less favored (Z)-8c
(Figure 4) is still similar to 8a,b in that the minima are at
fNCCF ꢂꢁ608 and 1808. However, the most stable conformer
is now anti; and especially in water, the energy differences
between the conformers are small (<3 kJmolꢀ1). This is
consistent with an electrostatic control (repulsion between
N and F), which favors anti; in water, the electrostatic inter-
action is screened, thus the differences between conformers
are reduced. No stereoelectronic gauche preference is op-
erational in this system. The barriers separating the minima
are significantly lower in (Z)-8c than in 8a,b, especially the
central ꢀgauche!+gauche barrier, which is almost halved.
The rotational barriers in (Z)-8c are not caused by syn-vici-
nal interactions, but by relatively weak 1,3-allylic (A1,3)
strain between the fluoromethyl group and the vinylic CH
(Scheme 3).
Figure 4. Bond-rotation profiles of the quinoid b-fluoropyridoximine tau-
tomers (E)- and (Z)-8c, calculated in water at the M06-2X/6-31+G(d, p)
level. Data for both isomers are plotted relative to the lowest stationary
point at this level of theory, (ꢀac)-(E)-8c. Full black symbols show con-
former energies obtained with the larger 6-311G (2df, p) basis set, rela-
tive to the lowest stationary point at that level, which is (syn)-(E)-8c.
AHCTUNGTRENNUNG
Table 3. Calculated relative conformational energies of different forms
of b-fluoropyridoximine 8.[a]
Entry
Form
Conformer
DE [kJmolꢀ1
]
fACTHNGUTERN(UNG NCCF) [8]
1
2
3
4
5
6
7
8
8a
8a
8a
8b
8b
8b
(E)-8c
(E)-8c
(E)-8c
(Z)-8c
(Z)-8c
(Z)-8c
ꢀgauche
+gauche
anti
ꢀgauche
+gauche
anti
ꢀanticlinal
+anticlinal
syn
ꢀgauche
+gauche
anti
0 (0)
6 (3)
6 (7)
0 (0)
4 (2)
6 (8)
0 (2)
0 (2)
2 (0)
9 (6)
9 (8)
3 (5)
ꢀ66 (ꢀ65)
64 (62)
176 (176)
ꢀ63 (ꢀ61)
62 (60)
176 (176)
ꢀ133 (ꢀ123)
135 (124)
0 (0)
9
10
11
12
ꢀ62 (ꢀ62)
61 (61)
179 (178)
[a] Results obtained with M06-2X/6-311GACTHNUTRGNEU(GN 2df,p) in vacuum and in con-
tinuum water (values in parentheses). Energies are relative to the most
stable conformer of the respective form in vacuum or water, respectively.
Scheme 3. Unfavorable interactions responsible for the highest NCCF ro-
tational barrier in (E)- and (Z)-8c, respectively.
The stabilization of the +gauche conformer in 8a,b is
readily understood when one considers that the nitrogen
lone pair is now accepting a hydrogen bond (8a) or is part
of the conjugated system (8b). Repulsion with the fluorine
lone pairs, responsible for the +gauche destabilization, is
therefore significantly reduced. Otherwise, 8a,b show quali-
tatively and quantitatively very similar conformational pref-
erences to the previous b-fluoroimines. The NBO-derived
Finally, in (E)-8c the NCCF rotation is almost unhin-
dered, and the conformers are essentially energetically de-
generate in both vacuum and water. Most notable is the in-
terchange between minima and maxima in the torsional pro-
file for (E)-8c as compared with all previous cases. The con-
formers of (E)-8c are at fNCCF ꢂꢁ1208 and 08, whereas
fNCCF ꢂꢁ608 and 1808 correspond to (low) barriers. The bar-
riers are caused by unfavorable syn-vicinal interactions
(Scheme 3). These are, however, much weaker than in, for
example, 8a, due to the wider sp2 bond angle at Ca (N-Ca-
Cb is 1318 in 8c, but 1118 in 8a), which positions the termi-
nal atoms of the NCCF torsion further apart (2.93 ꢁ in 8c
vs. 2.57 ꢁ in 8a at fNCCF =08). The flatness of the potential
energy surface for NCCF rotation in (E)-8c is also reflected
stereoelectronic preference in favor of ꢀgauche is DEdeloc
(ꢀg, a)=ꢀ8.7 kJmolꢀ1 for 8a, compared with ꢀ8.5 kJmolꢀ1
for 6.
-
ACHTUNGTRENNUNG
The net effect of solvation in 8a,b is to stabilize the +
gauche conformer and destabilize the anti relative to
ꢀgauche, however without changing the order of stability.
The anti destabilization is in agreement with an enhanced
stereoelectronic preference; DE
(ꢀg, a)=ꢀ10.1 kJmolꢀ1
for 8a in solution, compared with ꢀ8.7 kJmolꢀ1 in vacuum.
8854
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 8850 – 8857

Evidence of a Fluorine-Imine Gauche Effect
FULL PAPER
in the relatively large influence of basis set and solvation on
relative energies (several kJmolꢀ1) and exact positions of
the minima (ꢂ108).
Therefore in the E-quinoid the stereoelectronic situation
ꢀ
is such that the C F bond (or any other acceptor bond) is
preferentially activated when it is aligned with the extended
p system. Moreover, this conformation is accessible at very
little energetic cost as the NCCF rotation is practically un-
hindered, accounting for the facile elimination of fluoride
from the transient quinoid species to generate the Michael
acceptor that is necessary for inhibition of PLP-dependent
enzymes by b-fluoroamine derivatives.
ꢀ
C F activation in b-fluoropyridoximines: Returning to Du-
nathanꢀs stereoelectronic hypothesis, we finally analyzed the
effect of the NCCF torsion angle on the activation of the
ꢀ
ꢀ
C F bond in the quinoid (E)-8c. As a simple measure of C
ꢀ
F activation, the C F bond length as a function of fNCCF was
ꢀ
determined (Figure 5). The C F bond is lengthened by
Conclusion
We have presented theoretical and crystallographic evidence
that b-fluoroimines preferentially adopt one of two possible
gauche conformations, and that the overall order of stability
is ꢀgauche<anti<+gauche. A stereoelectronic rational for
this conformational preference is a stabilizing hyperconjuga-
tive interaction from the sCꢀ bond into the low lying s* or-
bital of the C F bond (sC HH!s*C F).[14] In the context of
ꢀ
ꢀ
ꢀ
this study, a theoretical approach has enabled us to compare
and contrast the conformational preferences of both the “in-
ꢀ ꢀ ꢀ
ꢀ
ternal” (F C C N=C) aldimine and “external” quinoid (F
ꢀ
ꢀ
C C=N C) intermediates that are implicated in the mecha-
nism of PLP-dependent enzyme inhibition. The gauche pref-
erence of the “internal” aldimine (=NCH2CH2F), which can
be rationalized by stereoelectronic arguments, does not hold
for the corresponding “external” system (N=CHCH2F)
ꢀ
Figure 5. Variation of the C F bond length with NCCF torsion in (E)-8c,
calculated using M06-2X/6-31+G
A
ꢀ
more than 0.02 ꢁ at fNCCF =ꢁ908, when it is exactly copla-
nar with the imine p system, compared with the syn confor-
mer, when it is perpendicular to the p system. In the two an-
(Emin when fNCCF =08). Moreover, the C F bond is length-
ened by more than 0.02 ꢁ at fNCCF =ꢁ908, when it is exactly
orthogonal to the plane of the conjugated imine. This activa-
ꢀ
ꢀ
ticlinal conformers (fNCCF ꢂꢁ1208), the C F bond is almost
tion of the C F s bond by an adjacent p system is consistent
with Dunathanꢀs stereoelectronic model.
as elongated as at fNCCF =ꢁ908. The NBO analysis revealed
that there is indeed substantial pN C!s*Cꢀ donation when
=
F
ꢀ
the C F bond is co-planar with the p-system. The antibond-
ing s*Cꢀ orbital has an occupancy of 0.05e, which weakens
the C F bond. In that conformation, the overlap between
the donor and acceptor NBOs is optimal (Figure 6). The
energy gained from this donor–acceptor interaction
(ꢀ38 kJmolꢀ1) is second only to the contributions from de-
localization within the p system itself. Contrastingly, in the
F
Experimental Section
ꢀ
Synthesis and structure determination:
See the Supporting Information for full experimental details.
Data for 6: Isolated as
(400 MHz, CDCl3): d=8.41 (s, 1H; CHN), 8.27 (ddd, J=8.9, 1.9, 1.9 Hz,
a
yellow solid (m.p. 102–1038C); 1H NMR
3
3
ꢀ
ꢀ
H
syn conformer, the C F bond is antiperiplanar to the Ca
2H; CHCNO2), 7.92 (ddd, J=9.1, 2.0, 2.0 Hz, 2H; CHCCNO2), 4.76 (dt,
2JHF =47.2, 4.9, 2H; CH2F), 3.96 ppm (dt, 3JHF =28.8, 4.7, 2H; CH2N);
13C NMR (100 MHz, CDCl3): d=161.3 (CHN), 149.2 (CNO2), 141.3
bond, which results in a much weaker sCꢀH!s*Cꢀ interac-
F
tion. The s*CꢀF orbital is occupied by only 0.01e, and the de-
localization energy amounts to ꢀ15 kJmolꢀ1.
1
(CCHN), 129.0 (CCNO2), 123.9 (CCCNO2), 82.4 (d, JCF =169.8 Hz; CF),
61.3 ppm (d, 2JCF =19.8 Hz, CCF); 19F NMR (376 MHz, CDCl3): d=
2
ꢀ222.5 ppm (tt, JFH =47.2, 28.6 Hz); IR (neat): n˜max =3104 (w), 2969 (w),
2903 (w), 1644 (m), 1603 (m), 1514 (s), 1433 (w), 1415 (w), 1392 (w),
1329 (s), 1290 (m), 1225 (m), 1104 (m), 1056 (m), 1028 (s), 1009 (m), 976
(w), 913 (m), 859 (s), 828 (s), 749 (m), 690 (m), 768 (w), 662 cmꢀ1 (w);
HRMS (ESI): m/z: calcd for: C9H10FN2O2+: 197.0721 [MH+]; found:
197.0714.
Data for 7: Isolated as an off white solid; m.p. 147–1488C; [a]2D0 +118.2
(c=1.05 in CHCl3); 1H NMR (400 MHz, CDCl3): d=8.19 (ddd, 3J=9.0,
2.0, 2.0 Hz, 2H; CHCNO2), 7.84 (s, 1H; CHN), 7.71 (ddd, 3J=9.0, 2.0,
2.0 Hz, 2H; CHCCNO2), 7.61–7.67 (m, 2H; Ph), 7.35–7.43 (m, 2H;
Ph),7.14–7.31 (m, 6H; Ph), 3.76 (dd, 3JHF =22.0, 5.0 Hz, 1H; CHCF),
2.21–2.38 (m, 1H; CHMe2), 0.85–0.96 ppm (m, 6H; Me); 13C NMR
2
ꢀ
Figure 6. Illustration of the stereoelectronic C F activation in (E)-8c at
(100 MHz, CDCl3): d=160.1 (CHN), 149.0 (CNO2), 142.6 (d, JCF
=
f
NCCF =908. Shown are the occupied pN and the vacant s*CꢀF NBOs.
23.2 Hz; Ph1), 141.4 (CCHN), 141.3 (d, 2JCF =22.7 Hz; Ph1’), 128.7
=
C
Chem. Eur. J. 2011, 17, 8850 – 8857
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8855

R. Gilmour, H. M. Senn et al.
(CCCNO2), 128.1 (Ph3), 127.9 (d, 4JCF =1.6 Hz; Ph3’), 127.6 (Ph4), 127.4
(Ph4’), 126.1 (d, 3JCF =10.3 Hz; Ph2), 125.8 (d, 3JCF =9.3 Hz; Ph2’), 123.8
(CCNO2), 100.8 (d, 1JCF =185.4 Hz; CF), 83.6 (d, 2JCF =23.9 Hz; CCF),
[4] For examples, see a) B. Lippert, B. W. Metcalf, R. J. Resvick, Bio-
chem. Biophys. Res. Commun. 1982, 108, 146–152; b) W. S. Faraci,
C. T. Walsh, Biochemistry 1989, 28, 431–437; c) D. B. Berkowitz,
K. R. Karukurichi, R. De La Salud-Bea, D. L. Nelson, C. D.
McCune, J. Fluorine Chem. 2008, 129, 731–742; d) D. B. Berkowitz,
K. R. Karukurichi, R. De La Salud-Bea, G. Maiti, J. M. McFadden,
M. L. Morris, ACS Symp. Ser. 2009, 1009, 288–303; e) D. B. Berko-
witz, R. de La Salud-Bea, W.-J. Jahng, Org. Lett. 2004, 6, 1821–
1824.
[5] D. Murali, L. G. Flores, A. D. Roberts, R. J. Nickles, O. T. DeJesus,
Appl. Radiat. Isot. 2003, 59, 237–243.
[6] M. G. Palfreyman, C. Danzin, P. Bey, M. J. Jung, G. Ribereau-
Gayon, M. Aubry, J. P. Vevert, A. Sjoerdsma, J. Neurochem. 1978,
31, 927–932.
[7] For selected examples, see a) M.-D. Tsai, H. J. R. Weintraub, S. R.
Byrn, C. Chang, H. G. Floss, Biochemistry 1978, 17, 3183–3188; L.
Casella, M. Gullotti, J. Am. Chem. Soc. 1983, 105, 803–809; b) M.
Nagata, M. Doi, M. Inoue, T. Ishida, M. Kamigauchi, M. Sugiura, A.
Wakahara, J. Chem. Soc. Perkin Trans. 2 1994, 983–988.
[8] a) C. Bꢄguin, S. Hamman, Org. Magn. Reson. 1981, 16, 129–132;
b) M. C. Salon, S. Hamman, C. G. Bꢄguin, Org. Magn. Reson. 1983,
21, 265–270; c) M. C. Salon, S. Hamman, C. G. Bꢄguin, J. Fluorine
Chem. 1985, 27, 361–370; d) H. ꢅnver, E. Kendi, K. Gꢂven, T. N.
Durlu, Z. Naturforsch. B 2002, 57, 685–690.
4
29.9 (CMe2), 21.6 (d, JCF =3.3 Hz; Me), 19.3 ppm (d, 4JCF =3.4 Hz; Me’);
19F NMR (376 MHz, CDCl3): d=ꢀ155.3 ppm (d, 3JHF =20.3 Hz); IR
(neat): n˜max =2963 (w), 1645 (w), 1598 (m), 1517 (s), 1450 (m), 1338 (s),
1201 (w), 1145 (w), 1102 (w), 1061 (m), 993 (m), 941 (w), 882 (w), 856
(m), 789 (m), 757 (s), 747 (s), 696 (s), 668 (m), 638 cmꢀ1 (w); HRMS
(ESI): m/z: calcd for C24H24FN2O2+: 391.1816 [MH+]; found: 391.1824.
Data for 8: Isolated as
a
yellow solid; m.p. 140–1418C; 1H NMR
(400 MHz, CDCl3): d=13.71 (brs, 1H; OH), 8.86 (s, 1H; CHN), 7.83 (s,
1H; CHpy), 4.73 (s, 2H; CH2OH), 4.67 (dt, 2JHF =47.3, 4.8 Hz, 2H;
CH2F), 3.91 (dtd, 3JHF =27.9, 4.7, 1.2 Hz, 2H; CH2N), 2.46 (s, 3H; CH3),
2.02 ppm (brs, 1H; OH); 13C NMR (100 MHz, CDCl3): d=165.0 (CHN),
154.6 (Cpy), 151.0 (Cpy), 138.1 (Cpy(6)), 131.0 (Cpy), 119.8 (Cpy), 82.0 (d,
1JCF =171.2 Hz; CF), 60.7 (CH2OH), 59.7 (d, 2JCF =19.8 Hz; CCF),
2
19.0 ppm (CH3); 19F NMR (376 MHz, CDCl3): d=ꢀ222.9 ppm (tt, JFH
=
47.1, 27.9 Hz); IR (neat) n˜max =3126 m, 2836 (w), 1631 (s), 1403 (s), 1340
(m), 1294 (m), 1260 (m), 1211 (m), 1119 (w), 1085 (w), 1023 (s), 988 (w),
964 (w), 906 (w), 856 (s), 786 (w), 756 (w), 714 (m), 640 cmꢀ1w; HRMS
+
(ESI): m/z: calcd for C10H14FN2O2 213.1034 [MH+]; found: 213.1034.
Computational Details
[9] For selected examples, see a) D. OꢀHagan, C. Bilton, J. A. K.
Howard, L. Knight, D. J. Tozer, J. Chem. Soc. Perkin Trans. 2 2000,
605–607; b) C. R. S. Briggs, D. OꢀHagan, J. A. K. Howard, D. S.
Yufit, J. Fluorine Chem. 2003, 119, 9–13; c) C. R. S. Briggs, M. J.
Allen, D. OꢀHagan, D. J. Tozer, A. M. Z. Slawin, A. E. Goeta,
J. A. K. Howard, Org. Biomol. Chem. 2004, 2, 732–740; d) N. E. J.
Gooseman, D. OꢀHagan, A. M. Z. Slawin, A. M. Teale, D. J. Tozer,
R. J. Young, Chem. Commun. 2006, 3190–3192; e) N. E. J. Goose-
man, D. OꢀHagan, M. J. G. Peach, A. M. Z. Slawin, D. J. Tozer, R. J.
Young, Angew. Chem. 2007, 119, 6008–6012; Angew. Chem. Int. Ed.
2007, 46, 5904–5908; f) D. Y. Buissonneaud, T. van Mourik, D.
OꢀHagan, Tetrahedron 2010, 66, 2196–2202; g) for an excellent
review see D. OꢀHagan, Chem. Soc. Rev. 2008, 37, 308–319; h) I.
Hyla-Kryspin, S. Grimme, S, Hruschka, G. Haufe, Org. Biomol.
Chem. 2008, 6, 4167–4175.
[10] For examples from this laboratory see a) C. Sparr, W. B. Schweizer,
H. M. Senn, R. Gilmour, Angew. Chem. 2009, 121, 3111–3114;
Angew. Chem. Int. Ed. 2009, 48, 3065–3068; b) C. Bucher, C. Sparr,
W. B. Schweizer, R. Gilmour, Chem. Eur. J. 2009, 15, 7637–7647;
c) C. Mondelli, C. Bucher, A. Baiker, R. Gilmour, J. Mol. Catal. A:
Chem. 2010, 327, 87–91; d) C. Sparr, R. Gilmour, Angew. Chem.
2010, 122, 6670–6673; Angew. Chem. Int. Ed. 2010, 49, 6520–6523;
e) C. Bucher, R. Gilmour, Angew. Chem. 2010, 122, 8906–8910;
Angew. Chem. Int. Ed. 2010, 49, 8724–8728; f) S. Paul, W. B. Schwe-
izer, M.-O. Ebert, R. Gilmour, Organometallics 2010, 29, 4424–
4427.
All calculations were done with Gaussian 09[17] using DFT. The M06–2X
hybrid meta-GGA exchange–correlation functional was used throughout,
which has been shown to yield superior accuracy not only for main-group
thermochemistry, but in particular for non-covalent interactions, includ-
ing dispersion and hydrogen bonding.[18] Even though such interactions
are not expected to be dominant in the systems studied here, they may
well influence conformational preference. We used standard Pople-style
basis sets, starting with 6-31+GACTHNUTRGNE(NUG d,p), in which the addition of diffuse
functions on non-hydrogen atoms enormously improves energetics when
using DFT methods, comparable to what is achieved with triple-z basis
sets.[19] Minima were re-optimized using 6-311G
ACHTUNGTRENNUNG
with 6-311+GACHTUNGTRENNUNG
tive energies was <1 kJmolꢀ1. Calculations in water solvent used the
IEF-PCM polarizable continuum model with UFF atomic radii for the
cavity construction. Non-electrostatic contributions to solvation were not
included (which corresponds to the default solvation treatment in Gaussi-
an 09). Default convergence criteria were applied for the SCF and in ge-
ometry optimizations. All stationary points were confirmed as minima by
a positive-definite Hessean. Values for rotational barriers were estimated
from the maxima of the torsion profiles.
Acknowledgements
[11] Selected crystallographic data for 6: Mr =196.181; orthorhombic
Pna21; a=10.3049(5), b=20.0536(12), c=4.4424(2) ꢁ; a=90.00, b=
90.00, g=90.008; fNCCF =ꢁ70.08.
We gratefully acknowledge generous financial support from the Alfred
Werner Foundation (assistant professorship to R.G.), the Roche Re-
search Foundation, Novartis AG (doctoral fellowships to C.S.), and the
ETH Zꢂrich.
[12] Selected crystallographic data for 7: Mr =390.458; orthorhombic
C2221; a=8.0748(2), b=15.9339(4), c=31.8928(8) ꢁ; a=90.00, b=
90.00, g=90.008; fNCCF =ꢀ74.28; polar space group; configuration of
the starting material known.
¯
[13] Selected crystallographic data for 8: Mr =212.224; triclinic P1; a=
[1] H. C. Dunathan, Proc. Natl. Acad. Sci. USA 1966, 55, 712–716.
[2] a) C. T. Walsh, Enzymatic Reaction Mechanisms, 2nd ed., W. H.
Freeman, San Fransisco, 1979; b) J. C. Vederas, H. G. Floss, Acc.
Chem. Res. 1980, 13, 455–463; c) H. Hayashi, H. Wada, T. Yoshi-
mura, N. Esaki, K. Soda, Annu. Rev. Biochem. 1990, 59, 87–110;
d) H. Hayashi, J. Biochem. 1995, 118, 463–473; e) A. C. Eliot, J. F.
Kirsch, Annu. Rev. Biochem. 2004, 73, 383–415; f) T. D. H. Bugg in
Introduction to Enzyme and Coenzyme Chemistry, 2nd ed., Wiley-
Blackwell, New York, 2004, pp. 211–226.
7.166(2), b=11.489(3), c=12.575(3); a=96.917(14), b=89.939(14),
g=101.869(13)8; fNCCF =ꢁ67.0/ꢁ66.9;[a] triclinic P1; a=4.4723(1),
b=10.1486(1), c=11.9847(2) ꢁ; a=105.698(1), b=92.052(1), g=
96.745(1)8; fNCCF =67.08/ꢀ66.38.[b] [a] Centrosymmetric space group
with 2 symmetry independent molecules in the asymmetric unit.
[b] Polar space group with two symmetry independent molecules in
the unit cell, both gauche conformers observed.
[14] L. Goodman, H. B. Gu, V. Pophristic, J. Phys. Chem. A 2005, 109,
1223–1229.
[3] C. T. Walsh, Annu. Rev. Biochem. 1984, 53, 493–535.
8856
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 8850 – 8857

Evidence of a Fluorine-Imine Gauche Effect
FULL PAPER
[15] F. Weinhold, C. R. Landis, Valency and Bonding: A Natural Bond
Orbital Donor–Acceptor Perspective, Cambridge University Press,
Cambridge, 2005.
[16] a) S. Sharif, M. C. Hout, P. M. Tolstoy, M. D. Toney, K. H. M. Jons-
son, H.-H. Limbach, J. Phys. Chem. B 2007, 111, 3869–3876; b) M.
Chan-Huot, C. Niether, S. Sharif, P. M. Tolstoy, M. D. Toney, H.-H.
Limbach, J. Mol. Struct. 2010, 976, 282–289; c) M. Chan-Huot, S.
Sharif, P. M. Tolstoy, M. D. Toney, H.-H. Limbach, Biochemistry
2010, 49, 10818–10830.
Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J.
Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar,
J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A.
Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, ꢆ.
Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian,
Inc., Wallingford CT, 2009.
[17] Gaussian 09, Revision A.02, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V.
Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X.
Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Son-
nenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,
M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven,
J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J.
[18] a) Y. Zhao, D. G. Truhlar, Theor. Chem. Acc. 2008, 120, 215–241;
b) Y. Zhao, D. G. Truhlar, Acc. Chem. Res. 2008, 41, 157–167.
[19] B. J. Lynch, Y. Zhao, D. G. Truhlar, J. Phys. Chem. A 2003, 107,
1384–1388.
Received: February 28, 2011
Published online: July 5, 2011
Chem. Eur. J. 2011, 17, 8850 – 8857
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8857