1-azabicyclo[3.3.1]nonan-2-one: Nitrogen versus oxygen protonation

Article abstract of DOI:10.1021/jo200195a

Protonation of typical unstrained amides and lactams is heavily favored at oxygen. In contrast, protonation of the highly distorted lactam 1-azabicyclo[2.2.2]octan-2-one is heavily favored at nitrogen. What structures occupy "crossover boundaries" where N- and O-protonation are nearly equienergetic Density function theory calculations at the B3LYP/6-31G* level, as well as QCISD(T)/6-31G* calculations, predict that 1-azabicyclo[3.3.1]nonan-2-one favors N-protonation at nitrogen only very slightly (<2.0 kcal/mol; "gas phase") over O-protonation. ¹H and ¹³C NMR as well as ultraviolet (UV) studies of this lactam, in its combination with sulfuric acid, confirm predominant protonation at nitrogen. Although the calculations very slightly favor the N-protonated chair-chair conformation, experimental spectra clearly support the N-protonated boat-chair. Broadened resonances in the ¹³C NMR spectrum suggest an exchange phenomenon. Variable-temperature studies of the ¹³C NMR spectra support dynamic exchange between the major tautomer (N-protonated) and the minor tautomer (O-protonated) in a roughly 4:1 mixture. The findings also support the published prediction that a twisted bridgehead lactam with the nitrogen lone pair (n_N) as HOMO will protonate at nitrogen.

Full text of DOI:10.1021/jo200195a

ARTICLE
pubs.acs.org/joc
1-Azabicyclo[3.3.1]nonan-2-one: Nitrogen Versus Oxygen
Protonation
Brian Sliter, Jessica Morgan, and Arthur Greenberg*
Department of Chemistry, University of New Hampshire, Durham, New Hampshire 03824, United States
S Supporting Information
b
ABSTRACT: Protonation of typical unstrained amides and
lactams is heavily favored at oxygen. In contrast, protonation of
the highly distorted lactam 1-azabicyclo[2.2.2]octan-2-one is
heavily favored at nitrogen. What structures occupy “crossover
boundaries” where N- and O-protonation are nearly equienergetic? Density function theory calculations at the B3LYP/6-31G* level,
as well as QCISD(T)/6-31G* calculations, predict that 1-azabicyclo[3.3.1]nonan-2-one favors N-protonation at nitrogen only very
1
13
slightly (<2.0 kcal/mol; “gas phase”) over O-protonation. H and C NMR as well as ultraviolet (UV) studies of this lactam, in its
combination with sulfuric acid, confirm predominant protonation at nitrogen. Although the calculations very slightly favor the
N-protonated chairꢀchair conformation, experimental spectra clearly support the N-protonated boat-chair. Broadened resonances
1
3
13
in the C NMR spectrum suggest an exchange phenomenon. Variable-temperature studies of the C NMR spectra support
dynamic exchange between the major tautomer (N-protonated) and the minor tautomer (O-protonated) in a roughly 4:1 mixture.
The findings also support the published prediction that a twisted bridgehead lactam with the nitrogen lone pair (n ) as HOMO will
N
protonate at nitrogen.
6
b
’
INTRODUCTION
favored over its N-protonated tautomer by about 7 kcal/mol.
Aqueous solvation, calculated by using the polarizable continuum
Distortion of the amide linkage, through twisting about the
1
model (PCM), only slightly modifies these energy differences in
NꢀCO bond and/or pyramidalization at nitrogen, can signifi-
17
2
the direction favoring the N-protonated tautomer. The calcu-
lated free energy differences [ΔGN-Prot. ꢀ ΔG (B3PW91/6-
cantly modify its chemistry. Examples include facile thermolysis,
O-Prot.
as well as hydrolysis under neutral aqueous conditions, of deriva-
3ꢀ5
31þG*)] (a) favor N-protonated 1-azabicyclo[2.2.2]octan-2-one
2) [ꢀ19.7 kcal/mol (gas phase); ꢀ21.0 kcal/mol (aqueous)],
(b) favor N-protonated 1-azabicyclo[3.2.2]nonan-2-one (3)
tives of “2-quinuclidone” (1-azabicyclo[2.2.2]octan-2-one),
(
and protonation as well as alkylation on nitrogen instead of
3ꢀ7
oxygen.
A delicate shift in the balance of structural factors
[
ꢀ4.8 kcal/mol (gas phase); ꢀ5.9 kcal/mol (aqueous)], and
dictates O-protonation (and methylation) of 1-azabicyclo[3.3.2]
decan-2-one and N-protonation (and methylation) of its lower
(
c) favor O-protonated N-methyl-4-methyl-2-piperidone (5)
1
7
8
[þ16.2 kcal/mol (gas phase); þ14.0 kcal/mol (aqueous)].
homologue 1-azabicyclo[3.2.2]nonan-2-one. The strain and un-
Protonated lactams 3 and 4 are thus calculated, and experimen-
ique structure of 3,5,7-trimethyl-1-azaadamantan-2-one have pro-
vided a fascinating picture of the four-coordinate intermediate in
4a,d
tally inferred, to be unambiguously N-protonated and O-pro-
tonated, respectively. In contrast, N-protonated 1-azabicyclo
[3.3.1]nonan-2-one (6, Scheme 2) has been calculated to be only
1.4 kcal/mol lower in corrected total energy than the O-proto-
9
,10
amide hydrolysis.
The enhanced reactivities and altered
chemistries of distorted amide linkages have consequences for
1
1
12
the activities of β-lactam antibiotics, mechanisms of proteolytic
6b
1
3
nated tautomer 7. This is close to the “noise level” of such
calculations and therefore worthy of experimental investigation.
Although, as noted above, aqueous solvation may add an addi-
tional 1ꢀ2 kcal/mol to the relative stability of 6, N-protonation
enzymes,
cis-trans-peptidylprolyl isomerases and protein
1
4
15
folding, and autocatalytic proteolysis. The discovery of “cross-
over” geometries in which N- and O-protonation are nearly
equienergetic may provide insights into these important issues.
(
6) and O-protonation (7) remain quite close in energy. The
1a
Ab initio molecular orbital calculations (isolated molecule or
6
calculated DunitzꢀWinkler amide distortion parameters in the
preferred boat-chair conformation of the neutral [3.3.1] lactam
(8BC, see Scheme 3) are as follows: τ (19.8°); χ (49.7°); χ
“
gas phase”) indicate that O-protonation is strongly favored
over N-protonation of unstrained amides and lactams and this is
16
N
C
consistent with experiment. For example, O-protonated N-methyl-
(
7.2°). The corresponding parametersin anundistorted lactamare
2
-pyrrolidone (1, Scheme 1) is calculated, at the 6-31G* level, with
correction for zero-point energy (ZPE) and thermal energy, to be
as follows: τ = χN = χ = 0°. An additional comparison is
CO
furnished by the sum of the three calculated CꢀNꢀC angles:
about 15 kcal/mol more stable than the N-protonated tautomer in
6
b
340.5° in the [3.3.1] lactam versus 360° for planar nitrogen and ca.
the gas phase. In contrast, N-protonation of 1-azabicyclo-
[
2
2.2.2]octan-2-one (see 2) is predicted to be favored by about
3 kcal/mol. Closer to the “borderline” (as noted above), 3 is
6
b
Received: January 26, 2011
Published: March 10, 2011
favored over its O-protonated tautomer by 9 kcal/mol while 4 is
r 2011 American Chemical Society
2770
dx.doi.org/10.1021/jo200195a
|J. Org. Chem. 2011, 76, 2770–2781

The Journal of Organic Chemistry
ARTICLE
Scheme 1
Table 1. Relative Uncorrected Total Energies (ΔE ), Rela-
tive Total Energies Including ZPE and Thermal Corrections
T
(
ΔE [Corr]), Relative Free Energies (ΔΔG) for Optimized
T
B3LYP/6-31G* Structures (gas phase), and Relative Total
Energies for Solvent Continuum (aqueous) B3LYP/6-31G*
Structures, As Well As Relative Energies (ΔE ) for Single
T
Point QCISD(T)/6-31G* Calculations (all values are in kcal/
mol)
QCISD(T)/6-
B3LYP/6-31G*
31G*
protonated
structure
ΔE
T
ΔE
T
T
[Corr] ΔΔG ΔE (aq)
ΔE
T
6
CC
0.00
þ0.28
þ0.40
0.00 0.00
0.00
0.00
þ0.13
þ2.98
Scheme 2
6BC
þ0.36
þ0.43
þ0.71 þ0.13
þ1.38 þ5.03
7
BC
movement of the flagpole hydrogen on C3 of the boat-chair toward the
equatorial position of the chair-chair. These molecules were also
optimized at the B3LYP/6-31G* level by using the solvent polarization
continuum model (SM8) in water. The lowest-energy conformer of the
neutral lactam (boat-chair) was optimized at the B3LYP/6-31G* level of
1
13
theory. Calculations of H and C NMR spectra were performed with
Spartan 08 at the B3LYP/6-31G* level of theory for the lactam and its
protonated tautomers.
Scheme 3
The two low-energy N-protonated conformers (both pairs of en-
antiomers) of 1-azabicyclo[3.3.1]nonan-2-one were initiated as hydro-
gen-bonded dimers at the AM1 level. However, due to Coulombic
repulsion in the “gas phase” calculation these dimers separated during
optimization. Reasoning that solvation in water would mitigate this
repulsion very significantly, the dimer was optimized at the HF/3-21G
level of theory by using the continuum model (SM8) in water. Single-
point energy calculations were performed upon these structures at the
B3LYP/6-31G* level of theory. For purposes of comparison, the two
N-protonated and single O-protonated conformers of the monomer
were optimized at the HF/3-21G level of theory and single-point energy
calculations were performed at B3LYP/6-31G* by using SM8 solvation
in water. A similar study was done for the dimer of N-protonated
3
28.5° for an idealized tetrahedral nitrogen. In addition to estab-
lishing the favored protonation site, experimental investigation of
the [3.3.1] system should also provide a challenging test of the
prediction by Werstiuk et al. that, in twisted bridgehead lactams,
those in which the highest occupied molecular orbital (HOMO)
corresponds to the nitrogen lone pair (n ) protonate at nitrogen
8
N
(e.g., 3), while those in which n is the HOMO protonate at
O
1-azabicyclo[2.2.2]octan-2-one (“2-quinuclidonium”), which was fully
oxygen (e.g., 4).
In the present study, 1-azabicyclo[3.3.1]nonan-2-one is dis-
solved in strong acidic media and the protonated lactam is studied
in solution, using H and C NMR as well as UV spectroscopy.
The experimental results are compared with computational pre-
dictions of total energies, free energies, molecular orbital energies,
and spectra in order to assign the site of protonation (N vs O) as
well as the conformation(s) of the protonated tautomer(s).
optimized at B3LYP/6-31G* by using SM8 solvation in water.
1
13
’
RESULTS AND DISCUSSION
Scheme 3 shows the calculated lowest-energy conformations
for 1-azabicyclo[3.3.1]nonan-2-one (8BC), its O-protonated
conjugate acid (7BC), and the two low-energy conformers for
the N-protonated tautomer (6BC and 6CC), where “B” and “C”
designated boat and chair conformations, respectively. Table 1
lists calculated relative total energies, relative total energies
including unscaled ZPE and thermal corrections, relative total
energies in aqueous solution, and relative free energies at the
B3LYP/6-31G* optimized level for the three protonated species
6BC, 6CC, and 7BC. Additionally, a single-point calculation was
made for each of these three structures at the QCISD(T)/6-
’
COMPUTATIONAL METHODS
18
Calculations were performed with the Spartan 08 as well as the
19
Gaussian 03 suite of programs. A Monte Carlo search algorithm was
initially performed on the N-protonated and O-protonated conjugate
acids of 1-azabicyclo[3.3.1]nonan-2-one at the AM1 level of theory. The
geometry of each conformer was optimized at the B3LYP/6-31G* level
of theory and single-point energy determinations were performed for the
two lowest-energy N-protonated conformers (boat-chair and chair-
chair) and the lowest-energy O-protonated conformer (boat-chair) at
the QCISD(T)/6-31G* level of theory with GAUSSIAN 03. The
transition state between the boat-chair and chair-chair was located by
using Spartan 08 at the B3LYP/6-31G* level of theory. It was confirmed
as a transition state by having only one imaginary vibrational mode and
by IRC in Gaussian 03. The imaginary frequency corresponded to
3
1G* level by using the DFT-optimized geometries. These
calculations support the earlier predictions that N-protonation
6
is favored over O-protonation. They also very slightly favor 6
CC
over 6BC. The NꢀCO bond in 8BC is calculated to be lengthened
relative to that in the unstrained lactams N-methyl-2-pyrrolidone
(1.398 Å vs 1.374 Å) while the CdO bond is calculated to show
little variation (1.222 Å vs 1.221 Å). Upon O-protonation, the
2
771
dx.doi.org/10.1021/jo200195a |J. Org. Chem. 2011, 76, 2770–2781

The Journal of Organic Chemistry
ARTICLE
1
1
Figure 1. (A) H NMR spectrum (400 Hz) of 1-azabicyclo[3.3.1]nonan-2-one in CDCl
3
; (B) H NMR spectrum of 1-azabicyclo[3.3.1]nonan-2-one
2 4
approximately equimolar with H SO in deuterated triflouroacetic acid (TFA-d) solution.
NdCO bond in 7BC is predicted to shorten to 1.312 Å and the
CdO bond lengthen to 1.317 Å. Protonation at nitrogen (e.g. for
6BC) is calculated to lengthen NdCO (1.581 Å) and shorten
CdO (1.187 Å). This is consistent with earlier MP2/6-31G*
whereas the chair-chair mimics the much more strained trans-
1
cyclohexene. The H NMR peak assignments for 8 in Table 2
BC
were aided by NOESY spectroscopic studies (see the Supporting
Information, S.2.13 and S.2.14). Of particular note is the NOESY
interaction between 9H and 3H (calculated distance, 2.177 Å),
6
b
studies.
fp
fp
1
13
The H and C NMR spectra of 1-azabicyclo[3.3.1]nonan-2-
verifying the boat-chair conformation.
one (8BC in Scheme 3) are displayed in Figures 1A and 2A.
The initial experiment involved dissolving 8 in pure deuterated
BC
2
0
1
21
Previous analyses of the H NMR of this compound provided
limited chemical shift assignments due to the complexity of some
of the upfield peaks. Computed and experimental assignments for
trifluoroacetic acid (TFA-d, pK = 0.52). While this acid should
a
22
not protonate an unstrained lactam [e.g., pK (1) = ꢀ0.75], it
a
might be capable of protonating this twisted lactam, which has an
appreciably pyramidal nitrogen. A previous computational study
predicted that the gas-phase proton affinity (PA) of 1-azabicyclo-
[3.3.1]nonan-2-one is slightly higher that that of N-methyl-2-
pyrrolidone [difference = 0.9 (6-31G*), 0.5 kcal/mol (6-31G*,
20d
8BC are listed in Table 2. Buchanan analyzed the long-range,
W-type coupling of 9H (termed 9H in the presentpaper; “fp” =
eq
fp
flagpole) with 8Heq as well as 6Heq in 5-phenyl-1-azabicyclo-
3.3.1]nonan-2-one, the absence of such coupling between 9Hax
and a C-4 proton, and assigned the boat-chair conformation. This
[
6b
with unscaled thermal and ZPE corrections)]. The B3LYP/6-
31G* calculations (including unscaled thermal and ZPE
corrections) in the present study predict a gas-phase PA of
218.3 kcal/mol for N-methyl-2-pyrrolidone (experimental: 220.7
20e
was subsequently supported by an X-ray crystallographic study.
This is consistent with the widely accepted amide-alkene
1,2
analogy: The boat-chair in this molecule mimics trans-cyclooctene
2
772
dx.doi.org/10.1021/jo200195a |J. Org. Chem. 2011, 76, 2770–2781

The Journal of Organic Chemistry
ARTICLE
1
3
13
Figure 2. (A) C NMR spectrum (100 MHz) of 1-azabicyclo[3.3.1]nonan-2-one in CDCl
3
; (B) C NMR spectrum of 1-azabicyclo[3.3.1]nonan-2-
1
3
one in concentrated H SO with added TFA-d [see Figure 3A for the corresponding C NMR spectrum (125 MHz) in approximately equimolar lactam
2
4
and H SO in TFA-d solution].
2
4
2
3
kcal/mol ). The corresponding DFT calculated value to produce
of 8BC and D SO (or H SO ) as solutes. These acids were usually
2 4 2 4
6CC is220.2 kcal/mol, and219.8 kcal/mol to produce6 . Thus, it
dry enough not to induce significant hydrolysis of the lactams to its
BC
was reasonable to investigate whether 8BC could be protonated in
protonated ring-opened amino acid.
1
13
1
TFA. There was no significant change in the H and the C NMR
Figure 1B shows the H NMR spectrum of approximately
spectra in comparing CDCl and TFA-d solutions (see the
equimolar 1-azabicyclo[3.3.1]nonan-2-one and H SO dissolved
3
2
4
13
Supporting Information, S.2.4ꢀS.2.6). Hence TFA does not
in TFA-d (the corresponding C NMR spectrum, Figure 3A).
1
13
protonate 1-azabicyclo[3.3.1]nonan-2-one. Initial H NMR ex-
Figure 2B displays the C NMR spectrum in pure H SO with
2
4
1
periments with mineral acid employed D SO and solvent with
added TFA-d. The H resonances are found in five groups: an
upfield group between 3.5 and 4.1 ppm due to CHꢀN protons
(4H); a multiplet between 2.9 and 3.1 ppm (2H) assigned to the
COꢀCH protons (2H); a multiplet between 2.5 and 2.7 ppm (2H);
a multiplet between 1.8 and 2.0 ppm (3H); and a multiplet between
2
4
added TFA-d as the internal standard, due to the insolubility of
1
3
TMS. The two C quartets of TFA (centered at 116.6 and 164.2
2
4
ppm) occur in spectroscopic “windows”, and the difference in
chemical shifts (47.6 ppm) between the two carbons in TFA
remains essentially constant in the solvent system employed (e.g.,
1.7 and 1.8 ppm (2H). Table 2 lists experimental and calculated
1
4
7.1 ppm in trifluoroacetic acid with added sulfuric acid, see
(B3LYP/6-31G*) H chemical shifts for 8 , calculated chemical
BC
below). The chemical shift of the hydrogen in TFA (11.5 ppm)
served as a useful marker for H NMR. Subsequent experiments
employed TFA-d as the solvent with roughly equimolar amounts
shifts for 6 , 6 ,and7 , as well as the experimental chemical shifts
BC CC
BC
1
of the protonated (i.e., deuterated) lactam (Figure 1B). Although it is
13
the C NMR spectrum that definitively establishes dominant
2
773
dx.doi.org/10.1021/jo200195a |J. Org. Chem. 2011, 76, 2770–2781

The Journal of Organic Chemistry
ARTICLE
Table 2. Comparison between Calculated (B3LYP/6-31G*)
N-protonation (see below), this conclusion is also supported by the
H NMR spectrum. O-Protonation is predicted to be characterized
1
1
and Experimental H NMR Chemical Shifts (CDCl ) for 8BC,
3
Experimental Chemical Shifts Reported for Approximately
by four protons in the range 2.70ꢀ3.13 ppm, one proton at 2.17
Equimolar Amounts of 8BC and H SO in TFA-d, and Com-
ppm, and four protons in the range 1.85ꢀ1.98 ppm. Both 6 and
BC
2
4
parison with Calculated Chemical Shifts for the N-Protonated
Conformers 6BC and 6CC and the O-protonated 7BC
6
CC
are calculated to have their 3H protons well-separated from
the seven remaining higher-field protons, as observed experimen-
tally. The striking similarity (Figure 1) between the four downfield
8
BC
protons (8H , 9H , 9H , 8H ) in the parent lactam and those in
protonated
6
BC
6
CC
7
BC
eq
fp
ax
ax
b
c
a
the protonated lactam is fully consistent with structure 6BC. If 6CC
proton expt calcd
8
BC expt
calcd
calcd
calcd
weredominant, onewould predictadditional long-range(W-type)
20d
3
3
4
4
5
6
6
7
7
8
8
9
9
H
H
H
H
H
H
H
H
H
H
H
H
H
fp
2.47 2.34 2.90ꢀ3.10
2.34 2.04 2.9ꢀ3.1
1.40 1.58 1.7ꢀ2.0
2.26 2.05 2.5ꢀ2.7
2.30 1.95 2.5ꢀ2.7
∼1.82 1.74 1.7ꢀ2.0
∼1.51 1.46 1.7ꢀ2.0
∼1.76 1.90 1.7ꢀ2.0
1.36 0.96 1.7ꢀ2.0
2.82 2.49 3.51
2.97
3.15
2.07
2.69
2.57
2.03
2.04
2.07
2.03
3.35
4.09
3.73
3.49
3.33 (3Heq
)
2.70
3.13
1.85
2.84
2.95
2.17
1.83
1.98
1.94
3.70
3.83
3.42
3.49
coupling between 9H and 4H . The NOESY spectrum (see
ax eq
theSupporting Information, S.2.13andS.2.14) indicatesproximity
eq
ax
eq
3.16 (3Hax
)
of 9H and 3H (calculated distance2.368 Å in 6BC) as well as 5H
fp
fp
2.50
and 4H (calculated distance 2.302 Å in 6BC).
eq
2.35
13
Comparison of the C NMR spectrum of protonated 8BC
2.30
(
(
Figure 2B or Figure 3A) with that of the neutral lactam
ax
eq
ax
eq
ax
eq
fp
2.24
Figure 2A) is very instructive. Table 3 lists experimental and
13
2.28
calculated C chemicalshifts for8BC, calculated chemical shifts for
6BC, 6CC, and 7 , as well as the experimental chemical shifts of
2.24
BC
the protonated lactam (6BC). Calculations with B3LYP/6-31G*
2.25
13
adequately predict the C chemical shifts of the carbons in 8BC
with the exception of the carbonyl carbon (C2), for which the
calculated value is about 14 ppm upfield of the experimental value.
This same single discrepancy (ca. 14 ppm upfield) is also seen in
comparison of B3LYP/6-31G* and experimental data for N-
methyl-2-pyrrolidone (9, see Table 4). N-Methyl-2-pyrrolidone,
as noted earlier, is calculated to favor O-protonation over N-pro-
tonation by 15 kcal/mol. It therefore protonates exclusively on
3.56
4.12 4.15 4.05
3.73
0
3.32 3.23 3.80
3.79 (9Hax
3.61
)
ax
3.05 2.86 3.61
a
6
BC is demonstrated in this paper to be the major component, with 7BC
comprising roughly 20%. For numbering of protons see the Supporting
Information, S.1.1. Chemical shifts assigned with the aid of COSY and
b
c
6,16
oxygen just as do other unstrained amides. Table 4 also shows
the comparison between calculation and experiment for this salt.
The calculated chemical shift of the carbonyl carbon of
NOESY. These may be compared with the chemical shifts reported by
2
0h
Brehm et al.
some cases.
representing unassigned chemical shift ranges in
13
Figure 3. Variable-temperature (25 to ꢀ15 °C) C NMR (125 MHz) spectra (A to D, respectively) for 1-azabicyclo[3.3.1]nonan-2-one and roughly
equimolar sulfuric acid in TFA-d. There was some hydrolysis of the lactam producing protonated 3-(3-piperidinyl)propionic acid, which supplemented
TFA as an internal standard for chemical shift as well as line width.
2
774
dx.doi.org/10.1021/jo200195a |J. Org. Chem. 2011, 76, 2770–2781

The Journal of Organic Chemistry
ARTICLE
Table 3. Comparison between Calculated (B3LYP/6-31G*)
Scheme 4
13
and Experimental C NMR Chemical Shifts (CDCl ) for 8BC,
3
Experimental Chemical Shifts Reported for Approximately
Equimolar Amounts of 8BC and H SO in TFA-d, and Com-
2
4
parison with Calculated Chemical Shifts for the N-Protonated
Conformers 6BC and 6CC and the O-Protonated 7BC
8
BC
protonated
b
a
c
carbon expt
calcd
8
BC expt
6
BC calcd
6
CC calcd
7
BC calcd
carbonyl carbon shift of 8BC moves from 185.0 ppm to higher
field (ca. 182.5 ppm). The carbonyl peak in Figure 2B is very
broad as is the peak at 27.5 ppm, with those at 20.5 and 25.5
ppm also showing some degree of broadening. The upfield shift is
more consistent with the calculated values for 6BC and 6CC after
adding ca. 14 ppm (as for 8BC) to the calculated values (see the
Discussion below). Similarly, correction of the calculated carbo-
nyl chemical shift for 7BC by adding 10ꢀ11 ppm (as for 1, see
Table 4) provides a value of about 192ꢀ193 ppm, some 7ꢀ8
ppm to lower field from the neutral lactams. The prediction of an
upfield shift of the carbonyl carbon of a lactam upon protonation
at nitrogen appears at first to be counterintuitive. However, there
is solid precedent. The annelated [4.3.1] system 10 (Scheme 4),
which also has significantly reduced resonance stabilization, has a
d
d
d
e
C2
C3
C4
C5
C6
C7
C8
C9
185.0 170.9 182.5 (v br)
166.8
163.9
181.6
33.4
25.0
29.8
30.9
20.7
51.6
52.9
33.2
27.9
32.2
32.3
21.7
51.0
52.4
34.5
30.4
22.8
25.4
29.5
18.0
55.7
52.3
34.4
28.4
26.6
26.8
20.6
53.0
56.3
31.6
27.9
37.3
29.6
28.5
57.0
57.7
25.5 (v sl br)
27.5 (v br)
30.5
b
20.5 (sl br)
56.5
56.0
a
6BC is demonstrated in this paper to be the major component, with 7
comprising roughly 20%. These may be compared with the chemical
shifts assigned by Brehm et al. That study definitively assigned the
9.7 ppm chemical shift to the methine carbon C5. With the exception
of C2, the assignments of C chemical shifts to specific carbons are
approximate, based upon comparison with calculated values, and are
discussed further in the text (v br = very broad; sl br = slightly broad; v sl
br = very slightly broad). Calculated and experimental results for 8
BC
b
20h
c
2
13
7c
carbonyl chemical shift of 188.7 ppm. The N-protonated salt
11) is more stable than its O-protonated tautomer: Aube
demonstrated crystallographic proof of N-protonation using
(
d
BC
7c
and for 2 and 9 (Table 4) suggest that calculated CdO chemical shifts
for related species should be corrected by adding 12ꢀ14 ppm. This also
the tetrafluoroborate salt. The carbonyl shift in 11 (177.3
7c
ppm) is some 11ꢀ12 ppm upfield of that in the neutral lactam.
e
makes the results for 6BC and 6CC self-consistent. Calculated and
In the present study, calculations of 1-azabicyclo[4.3.1]decan-10-
one, the core lactam in 10, as well as its N-protonated and
O-protonated derivatives, favor N-protonation over O-proton-
ation by 7.9 kcal/mol (comparison between the most stable
experimental results for 1 (Table 4) suggest that calculated CdO
chemical shifts for related species should be corrected by adding ca.
1
0ꢀ11 ppm to ca. 192ꢀ193 ppm. From a different standpoint,
comparison of the calculated chemical shift difference between 6BC
13
conformers of each). The calculated C chemical shift for the
carbonyl carbon in the N-protonated derivative (166.3 ppm) is
predicted to be almost 8 ppm upfield of that in the neutral lactam
1-azabicyclo[4.3.1]decan-2-one (174.2 ppm). As already noted,
the B3LYP/6-31G* values are about 14 ppm upfield of the
experimental values for lactam carbonyl carbons. The calculated
difference (8 ppm) is quite comparable to the experimental
difference between 10 and 11 (11ꢀ12 ppm). An upfield
chemical shift of the carbonyl carbon upon N-protonation of a
distorted lactam is also seen when one compares the chemical
and 7BC and the experimental value for 6 (179 ppm) obtained from
the variable-temperature C NMR study (see later discussion) suggests
ca. 15 ppm correction to 194 ppm. This will be the chemical shift
BC
13
employed in the present study for 7BC
.
1
3
Table 4. Calculated (B3LYP/6-31G*) and Experimental
Chemical Shifts for N-Methyl-2-pyrrolidone (9) and Its
O-Protonated Salt (1), As Well As Calculated Values for
C
1
-Azabicyclo[2.2.2.]octan-2-one (“ABO”) and Calculated and
5
5a
Experimental Values for Its N-Protonated Salt (2)
shift reported by Tani and Stoltz for 2 (175.9 ppm) with that
4a
reported by Somayaji and Brown for 5,6-benzo-1-azabicyclo-
[
2.2.2]octan-2-one (192.3 ppm), where both are reported for
3
a
9
1
ABO
2
CD CN solutions. Comparison of the calculated value for this
last compound (182.0 ppm vs 192.3 ppm experimental) with the
5
expt
calcd
expt
calcd
calcd
expt
calcd
5a
b
calculated value for the presently unknown neutral parent
compound (185.4 ppm, thus predicting an experimental value
of 195.7 ppm for the N-protonated tautomer) suggests that the
upfield shift in this system (ca. 20 ppm: 195.7 ppm vs 175.9 ppm)
should be even larger than that (11ꢀ12 ppm) recorded for 10
and 11. Table 4 includes calculated and experimental chemical
shifts for the protonated lactam 2 as well as calculated values for its
neutral conjugate base 1-azabicyclo[2.2.2]octan-2-one (“ABO” in
Table 4). Further analysis of the carbonyl C chemical shift in
6BC versus 8BC will be presented below. Of particular note at
present are relative C chemical shift values calculated for N-
methyl-2-pyrrolidone (9): 160.7 ppm (174.9 ppm, experimental);
its O-protonated derivative (1): 169.7 ppm (180.5
ppm experimental); and the N-protonated derivative: 166.2
ppm. For this undistorted lactam, the carbonyl carbon of the
C2: 174.9 C2: 160.7 C2: 180.5 C2: 169.7 C2: 185.8
C3: 30.8 C3: 30.2 C3: 34.7 C3: 32.3 C3: 40.9
C4: 17.8 C4: 20.6 C4: 19.5 C4: 19.4 C4: 28.8
C2: 175.9 C2: 164.0
C3: 40.1 C3: 39.1
C4: 25.7 C4: 26.4
C5: 49.1 C5: 47.9 C5: 57.0 C5: 56.0 C5/8: 27.3 C5/8: 22.7 C5/8: 24.3
C6: 29.7 C6: 28.5 C6: 33.1 C3: 33.1 C6/7: 46.1 C6/7: 48.1 C6/7: 48.6
a
b
ABO = 1-azabicyclo[2.2.2]octan-2-one. Note that the experimental
value for 5,6-benzo-1-azabicyclo[2.2.2]octan-2-one is 192.3 ppm,
whereas the calculated value (B3LYP/6-31G*) is 182.0 ppm. Use of
this difference might suggest that the experimental value for ABO should
be ca. 196 ppm.
4
a
13
13
25
O-protonated N-methyl-2-pyrrolidone (1) is ca. 10ꢀ11
ppm upfield of the experimental value.
Comparison of Figures 2A and 2B (or Figure 3A) and
comparison with Table 3 indicates that upon protonation, the
2
775
dx.doi.org/10.1021/jo200195a |J. Org. Chem. 2011, 76, 2770–2781

The Journal of Organic Chemistry
ARTICLE
N-protonated derivative is 5ꢀ6 ppm downfield of the neutral
lactam. Similarly, the calculated value for N-protonated N,N-
dimethylacetamide (162.9 ppm) is 6ꢀ7 ppm downfield of that in
the planar neutral amide (156.0 ppm). Of course, there are no
experimental NMR data for the N-protonated salts of these
undistorted lactam and amide compounds since they are
O-protonated.
On the basis of the C chemical shift of the very broad
carbonyl carbon peak in protonated 1-azabicyclo[3.3.1]nonan-2-
one and comparisons with experimental values for 2 and 1, as
well as computational chemical shifts (corrected), nitrogen is the
preferred site of protonation and 6BC and/or 6CC is the major
species present in the presence of sulfuric acid. This is certainly
consistent with the calculated values in Table 1.
run, there has been some hydrolysis of the lactam to the
protonated amino acid starting material. This turned out to
be fortuitous since the amino acid augmented TFA as a
chemical shift internal standard as well as a peak-width standard
for nonexchanging resonances. The same 99.98% H SO had
been used for an extended period by the time this phase of the
study was undertaken and had undoubtedly absorbed some
moisture from the atmosphere. Since the melting point of TFA
is ꢀ15 °C and the goal of this study was to investigate the origin
of the broad peaks rather than a detailed kinetic study, the
temperature range investigated was limited (25 to ꢀ15 °C) and
instrument temperature settings rather than precise calibrations
were employed.
2
4
13
At 0 °C (Figure 3), some very significant changes occurred in
13
The next question: Is the dominant conformation the boat-
chair (6BC) or chair-chair (6CC)? There are clear similarities in
appearance between the four low-field CHꢀN protons in parts A
and B of Figure 1. This suggests the same order in chemical shifts
among the CHꢀN protons just as the calculations predict for
the C NMR spectrum. Most obvious is that the broad carbonyl
resonance (182.5 ppm at 25 °C, corresponding to 182.5 ppm in
Table 3) has disappeared into the baseline. Also striking is the
very marked broadening of the peak at ca. 56 ppm, the higher
3
field of the two sp carbons attached to nitrogen. The other very
6
(Table 2). The calculations predict the same order of
broad peak (27.5 ppm) has also disappeared into the baseline
while the remaining two broad peaks (25.5 and 20.5 ppm) have
broadened further. At ꢀ10 °C, a peak at 179 ppm starts to appear
and is more in evidence at ꢀ15 °C. At ꢀ15 °C, the resonance at
25.5 ppm appears to be exhibiting coalescence. Our assignments
of experimental chemical shifts with calculated chemical shifts are
consistent with the broadened peaks corresponding to those with
the largest calculated chemical shift differences between 6BC and
7BC (Table 3).
BC
chemical shifts for the four downfield (CHꢀN) protons as found
experimentally. Furthermore, the calculations for 6 also re-
BC
produce the pattern of two protons in the 2.50ꢀ2.70 ppm range
and five higher-field protons with very similar chemical shifts.
13
The comparison between calculated versus experimental
C
chemical shifts (Table 3) does not allow a clear choice between
6BC and 6CC. Although the B3LYP/6-31G* and QCISD(T)
calculations both predict that 6CC is slightly more stable, the
proton NMR data clearly support 6BC as the more stable
conformer.
The changes observed for the carbonyl resonance in Figure 3
are of particular interest. Parts AꢀD of Figure 4 focus on this
peak. The peak at 179 ppm that starts to emerge at ꢀ10 °C
(Figure 4C) and is more apparent at ꢀ15 °C (Figure 4D)
13
The four broadened C resonances shown in Figure 2B
suggest the possibility of a dynamic exchange process accessible
13
due to the inherently larger range of C chemical shifts. Perhaps
there is exchange between protonated and unprotonated lactam?
Another possibility is exchange between monoprotonated and
diprotonated lactam. However, while diprotonated (but not
triprotonated) urea has been experimentally observed in magic
corresponds to the N-protonated tautomer 6 . For comparison,
BC
the corresponding resonances in 2 (∼176 ppm) and 11 (∼177
ppm) are very similar to this value (although both 2 and 11 are in
acetonitrile solution rather than TFA). The calculated difference
between 8BC and 6BC (4.1 ppm) is close to the experimental
difference (6.0 ppm). This raises the question of the location of
the lower-field carbonyl resonance corresponding to the O-pro-
tonated tautomer 7BC. If one simply employs the calculated
chemical shift difference between 6BC and 7BC (see Table 3), the
peak for 7BC should appear around 194 ppm. Using a weighted
average of chemical shifts, one calculates an equilibrium mixture
of roughly 4:1 (6BC/7BC). This might be a lower limit for this
ratio since there is simply no discernible peak in the vicinity of
194 ppm at ꢀ15 °C. The simulations (Figure 4, parts EꢀM)
indicate that only at very low exchange rate should the minor
peak emerge from the background.
26
acid (FSO H-SbF ), there are no published observations of an
3
5
amide or lactam protonated on both heteroatoms of the same
linkage. To test these possibilities, 1-azabicyclo[3.3.1]nonan-2-
one was dissolved in neat deuterated trifluorosulfonic (triflic)
13
acid (TFSA-d). The resulting C NMR spectrum (see the
Supporting Information, S.2.10) was essentially the same as that
obtained with D SO and thus only monoprotonated species are
2
4
present. When “25% Magic Acid” (4:1 FSO H/SbF ) was em-
3
5
1
13
ployed as solvent and the H and C NMR observed at ambient
temperature, it was clear that extensive decomposition of the
lactam had occurred (see the Supporting Information, S.2.11 and
13
27
S.2.12). Another explanation for the broad C peaks is exchange
Simulation, using Reich’s WinDNMR program, of the
among two or all three species (6BC, 6CC, and a lesser amount
exchange of two singlets in a 4:1 ratio and separated by 15
ppm (179 and 194 ppm) nicely reproduces key features of the
experimental data (see Figure 4, parts EꢀN; the 20% approx-
imation places the time-averaged peak at 182 rather than 182.5
ppm for simplicity’s sake). These include the total disappearance
of the peak due to its very broad coalescence at 0 °C (Figure 4H)
and the emergence of the N-protonated lactam peak at 179
ppm with no visible downfield O-protonated peak at 194 ppm at
this temperature or as low as ꢀ15 °C (e.g., Figure 4, parts IꢀL).
The very approximate calculated exchange rates (k þ k ) are as
1
3
of 7 ). The C chemical shift differences between potentially
BC
exchanging carbons in 6BC and 6CC are relatively small (Table 3).
The calculatedenergy of activation (B3LYP/6-31G*) for inversion
of the slightly flattened chair in 6BC to the slightly flattened boat in
6CC is only 1.5 kcal/mol (1.7 kcal/mol aqueous). This exceedingly
rapid conformational process cannot be responsible for the
broadened peaks in Figure 2B.
To further investigate the source of peak broadening in the
C NMR spectrum of protonated 1-azabicyclo[3.3.1]nonan-2-
1
3
ab
ba
ꢀ1
ꢀ1
one (Figure 2B), a limited variable-temperature study of the
carbonyl resonance was conducted at 125 MHz (Figure 3). In
this experiment, at ambient temperature it is clear that, for this
follows: 200 000 s (25 °C) (Figure 4G); 12 000 s (0 °C)
(Figure 4H; this is most uncertain since there is a very wide range
ꢀ
1
of rates (10 000ꢀ50 000 s corresponding to the disappearance
2
776
dx.doi.org/10.1021/jo200195a |J. Org. Chem. 2011, 76, 2770–2781

The Journal of Organic Chemistry
ARTICLE
1
3
Figure 4. (AꢀD) Expansion of low-field sections of variable-temperature C NMR spectra from Figure 3 from 25 (A) to ꢀ15 °C (D) respectively.
27
Parts EꢀN depict simulated WinDNMR spectra that correspond to a 4:1 ratio of 6BC and 7BC, having carbonyl carbon chemical shifts of 179 and 194
ꢀ
1
6
5
5
4
ppm (estimated), respectively, and calculated exchange rates (kAB þ kBA, s ) as follows: (E) 1 ꢁ 10 ; (F) 5 ꢁ 10 ; (G) 2 ꢁ 10 ; (H) 1.2 ꢁ 10 ; (I) 5 ꢁ
3
3
3
2
2
1
0 ; (J) 2 ꢁ 10 ; (K) 1 ꢁ 10 ; (L) 5 ꢁ 10 ; (M) 1 ꢁ 10 ; N) 0.
ꢀ
1
of peaks at this very broad coalescence); 5000 s (ꢀ10 °C)
absorption band(ca. 220ꢀ280nm) and a pronounced blueshift of
the πꢀπ* band. The UV spectrum of the hydrochloride salt of N,
ꢀ1
(
Figure 4I); and 2000 s (ꢀ15 °C) (Figure 4J). The simulated
30
spectra predict that the O-protonated peak at 194 ppm only starts to
N-dimethylacetamide (DMA), established by X-ray crystallo-
graphy to be protonated on oxygen, is of considerable interest. In
solution, protonation on the DMA oxygen produces a blue shift of
ꢀ
1
emerge from baseline at very slow exchange rates (e.g., <500 s , see
Figure 4L). Figure 4N depicts hypothetical 0 s exchange.
The relatively high observed coalescence temperature (ca.
°C) is a consequence of the huge chemical shift difference (ca.
5 ppm) estimated for carbonyl carbons in two very different
environments. As the simulations (Figure 4EꢀN) indicate, this
chemical shift difference is predicted to yield sharp peaks for the
separated carbonyl resonances only at very low exchange rates.
Ultraviolet (UV) spectroscopy is a useful probe for detecting
the site of protonation on an amide or lactam. Protonation on
nitrogen converts an allylic-type system to an NꢀH-substituted
ketone with the expectation that the nfπ* absorption will move
to longer wavelength, and the πfπ* transition will move to much
ꢀ
1
3c
ca. 8 nm (200 to 192 nm) in the πꢀπ* transition. This blue shift
0
1
might in fact be up to 10 nm greater if correction for solvent effects
31
were included. The DMA hydrochloride is also characterized by
weak bands at 315 and 360 nm (CH Cl solution), strongly
2
2
3c
resembling the spectrum observed in aqueous HCl. A claim to
have observed ca. 0.1% N-protonated DMA appears to have
28
31
been an artifact of solvent effects.
UV spectroscopy isconsistent with protonation of1-azabicyclo-
[3.3.1]nonan-2-one at nitrogen rather than oxygen. Figure 5A
displays the UV spectrum of N-methyl-2-pyrrolidone in water. It is
dominated by an intense absorbance just below 200 nm due to the
2
8
20h
shorter wavelength. Interestingly, UV spectra for unprotonated
and protonated derivatives of 2-quinuclidone were reported nearly
0 years ago. Since both the unprotonated and N-protonated
πꢀπ* transition. In good agreement with an earlier study, λmax
for 1-azabicyclo[3.3.1]nonan-2-one in water occurs at 204 nm (ε
7900, Figure 5B) and is associated with the πꢀπ* transition with a
slight shoulder that is likely due to the nꢀπ* transition. The UV
spectrum of N-methyl-2-pyrrolidone in acid solution (Figure 5C)
strongly resembles that of protonated DMA, reflecting O-proton-
ation. It includes weak absorptions at ∼260 and ∼320 nm and a
very slight blue shift of the πꢀπ* transition in pure sulfuric acid.
The UV spectrum of 1-azabicyclo[3.3.1]nonan-2-one in sulfuric
acid (Figure 5D) is markedly different from that of protonated
3
c
5
derivatives of 2-quinuclidone are effectively ketones, that study has
onlypartialrelevance for thepresentstudy. The nꢀπ* transition in
29
various methylated 2-quinuclidones is, as predicted, found at
higher wavelengths (236ꢀ247 nm in cyclohexane) than in
3c
unstrained lactams or amides (ca. 220 nm). The UV spectrum
of the N-protonated hydrochloride of the 6,6,7-trimethyl deriva-
tive (in water) is reported to include a very broad, indistinct
2
777
dx.doi.org/10.1021/jo200195a |J. Org. Chem. 2011, 76, 2770–2781

The Journal of Organic Chemistry
ARTICLE
Figure 5. (A) Ultraviolet spectrum of 0.00010 M N-methyl-2-pyrrolidone in water; (B) ultraviolet spectrum of 0.00010 M 1-azabicyclo[3.3.1]nonan-2-
one in water; (C) ultraviolet spectrum of 0.00010 M N-methyl-2-pyrrolidone in pure sulfuric acid; and (D) ultraviolet spectrum of 0.00010 M
1-azabicyclo[3.3.1]nonan-2-one in pure sulfuric acid. The quartz UV cell path length was 1.00 cm.
N-methyl-2-pyrrolidone (Figure 5C). There now appear to be two
substantial πꢀπ* absorbances: the larger one very appreciably
blue-shifted and the other very slightly red-shifted to 210 nm.
There appears to be a very small trace of an absorbance at ca.
Scheme 5
3
20 nm in Figure 5D.
The significant differences in the UV spectra depicted in parts
C and D of Figure 5 are consistent with a dominant change in the
site of protonation from oxygen to nitrogen. The features in
Figure 5D suggest the possibility of a lesser amount of 7BC in
equilibrium with 6BC (although a minor quantity of 6CC would
be difficult to distinguish by using UV spectroscopy).
an aqueous continuum) led to 6CC/6CC structures as most
favorable. Combinations of R,R (equivalent to S,S) and R,S
indicated that structure 12 (R,S-6CC, see Scheme 5) has the
lowest energy although the energy differences are very small. In
aqueous solution the total energy of 12 is 2.1 kcal/mol lower than
the separated monomers in aqueous continuum. However, due
to losses in translational and rotational entropy as well as the tight
organizational order due to two hydrogen bonds, the aqueous
monomers are favored by about10 kcal/mol over the dimer. The
In unstrained amides and lactams, the p-orbital of the planar
nitrogen is the major contributor to the HOMO, which resem-
3
2
bles ψ of the allylic system, sometimes termed π . Twisting of
2
N
1b
the amide is accompanied by pyramidalization of nitrogen and
transformation of the nitrogen contribution (in π ) to a lone pair
N
3
2
(
n ). Comparison of the experimental UV photoelectron
N
spectrum with computational results rationalizes the first two
sharp) bands of the spectrum of 1-azabicyclo[3.3.1]nonan-2-
(
32
13
one in terms of n as the HOMO and n as HOMO-1. Our
calculated C chemical shifts for 12 (gas phase for comparison
N
O
finding that this lactam protonates on nitrogen rather than
purposes) predict a value for C2 of 180.6 ppm in seeming
agreement with experiment. However, the caveats in Tables 3
and 4 suggest that this calculated value actually should corre-
spond to an experimental chemical shift of 190ꢀ195 ppm.
Therefore, there is no evidence supporting the dimer. In
principle, 2-quinuclidonium (2) can also form an excellent
8b
oxygen is in accord with the predictions of Werstiuk et al.
13
During the investigation of the broadened C spectra of 6BC
the possibility of hydrogen-bonded dimers playing a role in
exchange was investigated. Molecular models quickly make it
apparent that dimerization of N-protonated chair-chair (6CC) is
far better than 6BC. Optimizations (B3LYP/6-31G*//3-21G) of
the dimer employing 6CB/6CB, 6CB/6CC, and 6CC/6CC (each in
doubly hydrogen-bonded dimer (13) as readily as 6 might.
CC
5
However, this does not appear to be the case and in the crystal
2
778
dx.doi.org/10.1021/jo200195a |J. Org. Chem. 2011, 76, 2770–2781

The Journal of Organic Chemistry
ARTICLE
structure the hydrogen bonding apparent in the crystalline
solution. 2-Quinuclidonium can, in principle, form an analogous
dimer and the energy considerations are almost identical with
those of the aforementioned [3.3.1] system. However, the Tani
and Stoltz study showed no evidence of the dimer in solution. In
þ
5a
tetrafluoroborate involves NꢀH---F. Calculations indicate
that dimerization of the monomer in aqueous solution also
corresponds to a virtually identical favorable change in total
energy (ꢀ2.1 kcal/mol), but loss of translational and rotational
freedom along with the highly ordered structure are factors
heavily favoring the monomer. Although a rate of exchange has
been estimated for transfer of the proton between nitrogen and
oxygen (see above), the kinetics of this process have not been
investigated. While it would seem that exchange rates are likely to
be pseudo-first order (i.e., proton exchange between the mono-
mer and solvent), the potential formation of dimers (12) as
intermediates for exchange would be consistent with a process
that is second-order in monomer, characterized by a substantial
negative entropy of activation.
the solid state, there is hydrogen bonding between 2-quinucli-
ꢀ
donium and its counterion BF
. Although proton exchange is
4
likely to occur between the monomer and solvent, it is con-
ceivable that dimerization may furnish a viable, but entropically
demanding, pathway for simultaneous exchange of two N-pro-
tons with two O-protons. Detailed investigation of the mechan-
ism of the tautomerism between 6BC and 7BC is a project for
future study.
Since pyramidalization at the carbonyl carbon (χ ) in dis-
C
torted lactams and amides is usually quite small, this allows the
distortion energy surface to be three-dimensional. The present
case provides one experimental point on such a surface (τ = 20°;
χ = 50°) where N-protonation and O-protonation are almost
N
’
CONCLUSIONS
-Azabicyclo[3.3.1]nonan-2-one protonates on nitrogen in
equienergetic. Complexing species other than protons should
1
exhibit markedly different distortion energysurfaces. “Hard” Lewis
þ
agreement with calculations for the gas phase and for aqueous
solution. The most obvious support for this conclusion is the
upfield chemical shift, relative to the free lactam, of the carbonyl
acids such as Na should intrinsically favor oxygen complexation
and require greater pyramidalization and twisting to complex
2þ
nitrogen; “softer” Lewis acids such as Cu should require little
1
3
2c
C resonance of the protonated species. Calculations predict
distortion in order to complex to the amide nitrogen. This is
2þ
that the chair-chair conformation of the N-protonated lactam is
consistent with the function of Cu discovered by Eakin et al. in
1
very slightly more stable than the boat-chair, while the H and
reducing the activation barrier to cis-trans peptidyl-prolyl isomer-
1
3
14g
C NMR spectra favor the latter. This is most apparent by
ization of β-2-microglobulin. This is also an area for future study.
comparison of the chemical shifts and coupling patterns of the
four downfield protons in 6BC with those in 8BC. Conformational
exchange between the two conformers is calculated to be
exceedingly rapid (energy of activation <2 kcal/mol) and the
’
EXPERIMENTAL SECTION
1
13
NMR Spectra. H and C NMR spectra were run at 400 and 100.5
NMR spectra arise from the dominant conformer (6 ) with
MHz, respectively, on a Varian Mercury spectrometer. The variable-
temperature C NMR spectra were run at 125.7 MHz on a Varian Inova
BC
1
13
13
little if any broadening due to the similarities in the H and
C
NMR chemical shifts of the two conformers. There is, however,
500 MHz spectrometer. Spectra were referenced (ppm) against TMS
except in the experiments involving protonation with sulfuric acid, where
13
an exchange process apparent through variable-temperature
C
1
3
1
NMR spectroscopy, even at ambient temperature. It corresponds
to dynamic exchange between the dominant N-protonated
tautomer and the minor (ca. 20%) O-protonated tautomer.
The large chemical shift difference between the carbonyl carbons
of 6BC and 7BC (ca. 15 ppm calculated) contributes to making
this exchange observable at moderately low temperatures.
The UV spectrum of N-methyl-2-pyrrolidone in sulfuric acid
bears strong resemblance to the UV spectrum of the hydro-
chloride salt of N,N-dimethylacetamide, established by X-ray
crystallography to be O-protonated. The UV spectrum of
trifluoroacetic acid-d (TFA-d) ( C and residual H) served as internal
1
3
standard as well as the C spectrum in triflic acid (CF
solvent also served as internal standard.
SO H) where the
3
3
UV Spectra. Ultraviolet spectra were obtained with a Cary 50 Bio
(Varian) spectrometer with 1.00 cm thickness quartz cells and the
resulting spectra were analyzed with Cary WinUV version 3.00(182)
software. For 1-azabicyclo[3.3.1]nonan-2-one, 0.0014 g was dissolved in
10.0 mL of solvent (water or 99.98% sulfuric acid) and diluted 1:10 in
the same solvent to produce a 0.00010 M solution. Similarly, 0.0010 g of
N-methyl-2-pyrrolidone was dissolved in 10.0 mL of solvent and diluted
1:10 to produce a 0.00010 M solution.
1
-azabicyclo[3.3.1]nonan-2-one in sulfuric acid shows marked
Synthetic Procedures
differences compared to the corresponding spectrum for N-
methyl-2-pyrrolidone. Its UV spectrum is compatible with the
simultaneous presence of N-protonated and O-protonated tau-
tomers with the former being the major species. Taken together
3
-(3-Piperidyl)propionic Acid Hydrochloride: In a slight
2
0a,b
modification of published procedure,
11.2 g, 0.080 mol) in 100 mL of glacial acetic acid was hydrogenated
over 0.75 g of 10% Pd on charcoal in a stainless steel bomb with stirring
Pressure Products Industries, Inc.) over the course of 6 days maintain-
trans-3-(3-pyridyl)acrylic acid
(
1
3
with the simulated C dynamic NMR study and the predictions
of the calculations, this appears to be the first observation of an
amide or lactam with measurable quantities of N-protonated and
O-protonated species in equilibrium. The results also support the
(
20a,b
ing a maximum hydrogen pressure of 600 psi during this period.
The
resulting solution was filtered through Celite three times and the filtrate
concentrated to a brown syrup on a steam bath/rotoevaporator. The
resulting oil was dissolved in absolute ethanol (1:15 mol ratio) and 6.0 M
hydrochloric acid was added until precipitation was complete. The white
finding of Werstiuk et al. that the nature of the HOMO (n or
O
nN) of a twisted bridgehead lactam determines the site of
protonation (oxygen or nitrogen).
precipitate was filtered; the yield of hydrochloride was 8.0 g (70.7% yield
1
The doubly hydrogen-bridged dimer of chair-chair N-proto-
nated 1-azabicyclo[3.3.1]nonan-2-one has an attractive structure
and, in aqueous solution, has a calculated total energy over 2
kcal/mol lower than the separated monomeric cations. However,
the free energy of the dimer is considerably higher than that of
the two monomers and there is no evidence for the dimer in
overall): mp 219ꢀ220 °C; H NMR (400 MHz, DMSO-d
6
) δ 1.12
(
m, 1H), 1.45 (m, 2H), 1.67 (m, 4H), 2.24 (t, 2H), 2.50 (broad q, 1H),
2.70 (broad q, 1H), 3.15 (broad d, 2H), 8.89 (broad s, 1H), 9.23 (broad s,
1
3
1H), 12.15 (s, 1H); C NMR (100.5 MHz, DMSO-d ) δ 22.3, 28.5, 28.8,
6
1
3
31.3, 33.0, 43.8, 48.2, 174.9; C NMR (100.5 MHz, TFA-d) δ 24.2, 29.6,
30.1, 33.0, 35.7, 48.6, 52.7, 189.5.
2
779
dx.doi.org/10.1021/jo200195a |J. Org. Chem. 2011, 76, 2770–2781

The Journal of Organic Chemistry
ARTICLE
2
0h
1
-Azabicyclo[3.3.1]nonan-2-one: . A 354 mg (1.85 mmol)
Mucsi, Z.; Tsai, A.; Szori, M.; Chass, G. A.; Viskolcz, B.; Czismadia, I. G.
J. Phys. Chem. A 2007, 111, 13245–13254.
sample of 3-(3-piperidyl)propionic acid hydrochloride and 665 mg
2.67 mmol) of di-n-butyltin(IV) oxide in 1200 mL of toluene is refluxed
(
(3) (a) Pracejus, H. Chem. Ber. 1959, 92, 988. (b) Pracejus, H. Chem.
Ber. 1965, 98, 2897. (c) Pracejus, H.; Kehlen, M.; Kehlen, H.; Matschi-
ner, H. Tetrahedron 1965, 21, 2257–2270. (d) Blackburn, G. M.; Skaife,
C. J.; Kay, I. T. J. Chem. Res., Miniprint 1980, 3650–3668. Greenberg, A.;
Wu, G.; Tsai, J. C.; Chin, Y. Y. Struct. Chem. 1993, 4, 127–129.
with a DeanꢀStark trap for 16 h. The solution is concentrated to a
yellow oil with use of rotoevaporation. Flash chromatography on silica
gel with ethanol:hexanes (4:1) furnished the product (mp 79ꢀ81 °C for
20c
20g
1
racemate; lit. values 77ꢀ79 °C and 77ꢀ81 °C ). H NMR (400
(4) (a) Somayaji, V.; Brown, R. S. J. Org. Chem. 1986,
MHz, CDCl
3
) δ 1.36 (m, 1H), 1.40 (ddd, 1H), 1.51 (broad d, 1H),
51, 2676–2686. (b) Somayaji, V.; Skorey, K. I.; Brown, R. S.; Ball,
1
1
.76ꢀ1.82 (m, 2H), 2.26ꢀ2.34 (m, 3H), 2.47 (ddd, 1H), 2.82 (ddd,
13
R. G. J. Org. Chem. 1986, 51, 4866–4872. (c) Bennet, A. J.; Wang, Q. P.;
Slebocka-Tilk, H.; Somayaji, V.; Brown, R. S. J. Am. Chem. Soc. 1990,
H), 3.05 (d, 1H), 3.32 (dd, 1H), 4.12 (ddd, 1H); C NMR (100.5
) δ 20.7, 25.0, 29.8, 30.9, 33.4, 51.6, 52.9, 185.0. [The NMR
MHz, CDCl
spectra include minor contributions from pump oil: H NMR
3
112, 6383–6385. (d) Wang, Q. P.; Bennet, A. J.; Brown, R. S.;
1
Santarsiero, B. D. J. Am. Chem. Soc. 1991, 113, 5757–5765. (e) Bennet,
A. J.; Somayaji, V.; Brown, R. S.; Santarsiero, B. D. J. Am. Chem. Soc.
13
(
m, 0.83ꢀ0.89 ppm; 1.26 ppm); C NMR (29.8 ppm).]
Protonated 1-Azabicyclo[3.3.1]nonan-2-one:
Azabicyclo[3.3.1]nonan-2-one was initially protonated in 99.98%
SO or D SO with CF CO D (TFA-d) as an internal standard.
Subsequent studies were performed with H SO approximately equi-
1-
1
991, 113, 7563–7571.(f) Brown, R. S. In Greenberg, A.; Breneman,
C. M.; Liebman, J. F. The Amide Linkage: Structural Significance in
Chemistry, Biochemistry and Material Science; John Wiley & Sons: New
York, 2000: pp 85ꢀ114.
H
2
4
2
4
3
2
2
4
molar to the lactam in TFA-d as solvent (see the Supporting In-
(5) (a) Tani, K.; Stoltz, B. M. Nature 2006, 441, 731–734. (b) Ly, T.;
Krout, M.; Pham, D. K.; Tani, K.; Stoltz, B. M.; Julian, R. R. J. Am. Chem.
Soc. 2007, 129, 1864–1865.
1
formation). H NMR (400 Hz, TFA-d) δ 1.70ꢀ2.00 (m, 5H),
2
1
(
.50ꢀ2.70 (m, 2H), 2.90ꢀ3.10 (m, 2H), 3.51 (ddd, 1H), 3.61 (d,
13
(
6) (a) Greenberg, A.; Venanzi, C. A. J. Am. Chem. Soc. 1993,
15, 6951–6957. (b) Greenberg, A.; Moore, D. T.; DuBois, T. D. J.
Am. Chem. Soc. 1996, 118, 8658–8668.
7) (a) Szostak, M.; Yao, L.; Aube, J. J. Org. Chem. 2009,
H), 3.80 (broadened d, 1H), 4.05 (broadened d, 1H); C NMR
1
100.5 MHz, TFA-d) δ 20.5, 25.5, 27.5, 30.5, 34.5, 56.0, 56.5, 182.5.
Studies were also performed in triflic acid (CF SO H) as well as in 25%
Magic Acid (i.e., 4:1 fluorosulfuric acid/antimony pentafluoride).
Protonated N-Methyl-2-pyrrolidone: The lactam was dissolved
in 99.98% H SO with TFA-d added as internal standard. C NMR
TFA-d, ppm) δ 19.5, 33.1, 34.7, 57.0, 180.5. This may be compared with
the results for the neutral lactam: C NMR (CDCl
8.5, 173.9 (see the Supporting Information for these spectra).
3
3
(
74, 1869–1875. (b) Szostak, M.; Yao, L.; Aube, J. J. Am. Chem. Soc.
2
010, 132, 2078–2084. (c) Szostak, M.; Yao, L.; Day, V. W.; Powell,
1
3
2
4
D. R.; Aube, J. J. Am. Chem. Soc. 2010, 132, 8836–8837.
8) (a) Werstiuk, N. H.; Brown, R. S.; Wang, Q. Can. J. Chem. 1996,
4, 524–532. (b) Werstiuk, N. H.; Muchall, H. M.; Roy, C. D.; Ma, J.;
Brown, R. S. Can. J. Chem. 1998, 76, 672–677.
9) (a) Kirby, A. J.; Komarov, I. V.; Wothers, P. D.; Feeder, N.
(
(
13
3
) δ 16.8, 28.6, 29.8,
7
4
(
Angew. Chem., Int. Ed. 1998, 37, 785–786. (b) Kirby, A. J.; Komarov,
I. V.; Feeder, N. J. Am. Chem. Soc. 1998, 120, 7101–7102. (c) Kirby, A. J.;
Komarov, I. V.; Kowski, K.; Rademacher, P. J. Chem. Soc., Perkin Trans. 2
’
ASSOCIATED CONTENT
1
S
Supporting Information. Experimental procedures, H
b
1
999, 1313–1316.
10) Morgan, K. M.; Rawlins, M. L.; Montgomery, M. N. J. Phys. Org.
Chem. 2005, 18, 310–314.
11) Lei, Y.; Wrobleski, A. D.; Golden, J. E.; Powell, D. R.; Aube, J. J.
1
3
and C NMR spectra, total energies, full geometries, zero point
energies, and thermal corrections for all structures. This material
is available free of charge via the Internet at http://pubs.acs.org.
(
(
Am. Chem. Soc. 2005, 127, 4552–4553.
’
AUTHOR INFORMATION
(12) Boyd, D. B. In Greenberg, A.; Breneman, C. M.; Liebman, J. F.
The Amide Linkage: Structural Significance in Chemistry, Biochemistry, and
Materials Science; John Wiley & Sons: New York, 2000; pp 337ꢀ375.
(13) Mujika, J. I.; Formoso, E.; Mercero, J. M.; Lopez, X. J. Phys.
Chem. B 2006, 110, 15000–15011.
Corresponding Author
*E-mail: art.greenberg@unh.edu.
(14) (a) Harrison, R. K.; Stein, R. L. Biochemistry 1990,
2
9, 1684–1689. (b) Liu, J.; Albers, M. W.; Chen, C. M.; Schreiber,
’
ACKNOWLEDGMENT
S. L.; Walsh, C. T. Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 2304–2308. (c)
Fischer, G. Chem. Soc. Rev. 2000, 29, 119–127. (d) Cox, C.; Leckta, T.
Acc. Chem. Res. 2000, 33, 849–858. (e) Wedemeyer, W. J.; Welker, E.;
Scheraga, H. A. Biochemistry 2002, 41, 14637–14644. (f) Bhat, R.;
Wedemeyer, W. J.; Scheraga, H. A. Biochemistry 2003, 42, 5722–5728.
The authors wish to thank Professors Richard Johnson, Glen
Miller, and Sterling Tomellini, Dr. Patricia Stone Wilkinson, and
Ilia Terova for helpful discussions and suggestions.
(
2
g) Eakin, C. M.; Berman, A. J.; Miranker, A. D. Nat. Struct. Mol. Biol.
006, 13, 202–208.
’
REFERENCES
(1) (a) Dunitz, J. D.; Winkler, F. K. Acta Crystallogr., Sect. B 1975,
(15) Johansson, D. G. A.; Wallin, G.; Sandberg, A.; Macao, B.;
3
1, 251–263.(b) Wiberg, K. B. In Greenberg, A.; Breneman, C. M.;
Åqvist, J.; H €a rd, T. J. Am. Chem. Soc. 2009, 131, 9475–9477.
(16) (a) Birchall, T.; Gillespie, R. J. Can. J. Chem. 1963,
41, 2642–2649. (b) Olah, G. A.; White, A. M. Chem. Rev. 1970,
70, 561–591. Klumpp, D. A.; Rendy, R.; Zhang, Y.; Gomez, A.; McElrea,
A. Org. Lett. 2004, 6, 1789–1792.
(17) (a) Mujika, J. I.; Mercero, J. M.; Lopez, X. J. Phys. Chem. A 2003,
107, 6099–6107. (b) Mujika, J. I.; Mercero, J. M.; Lopez, X. J. Am. Chem.
Soc. 2005, 127, 4445–4453.
Liebman, J. F. In The Amide Linkage: Structural Significance in Chemistry,
Biochemistry, and Materials Science; John Wiley & Sons: New York, 2000;
pp 33ꢀ45. (c) Yamada, S. J. In Greenberg, A.; Breneman, C. M.;
Liebman, J. F. The Amide Linkage: Structural Significance in Chemistry,
Biochemistry, and Materials Science; John Wiley & Sons: New York, 2000;
pp 215ꢀ246.
(2) (a) Hall, H. K., Jr.; El-Shekeil, A. Chem. Rev. 1983, 83, 549–555.
(
b) Greenberg, A. In Liebman, J. F.; Greenberg, A. Structure and
(18) Spartan 08, 1.2.0, Wavefunction, Inc., Irvine, CA, 2008.
(19) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.;
Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.;
Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson,
Reactivity; VCH Publishers Inc.: New York, 1988; pp 139ꢀ178. (c)
Greenberg, A. In Greenberg, A.; Breneman, C. M.; Liebman, J. F. The
Amide Linkage: Structural Significance in Chemistry, Biochemistry, and
Materials Science; John Wiley & Sons: New York, 2000; pp 47ꢀ83. (d)
2
780
dx.doi.org/10.1021/jo200195a |J. Org. Chem. 2011, 76, 2770–2781

The Journal of Organic Chemistry
ARTICLE
G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.;
Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.;
Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;
Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.;
Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski,
V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick,
D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui,
Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith,
T. M.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.;
Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.;
Pople, J. A. Gaussian 03, Revision E.01; Gaussian Inc., Wallingford, CT,
2004.
(
20) (a) Shaw, R. G., Jr. 1-Azabicyclo[3.3.1]nonan-2-one: A Two
Chair Bicyclic Lactam Containing Bridgehead Nitrogen, Masters Degree
Thesis, The University of Arizona, 1979 [Note: the initial assignment of
the double-chair conformation was due to an error in assigning the
chemical shifts (at 60 MHz) of the two anisochronous 9H protons. The
authors adopted the boat-chair conformation shortly afterward: see ref
17b]; (b) Hall, H. K., Jr. J. Am. Chem. Soc. 1960, 82, 1209–1215. (c) Hall,
H. K., Jr.; Shaw, R. G., Jr.; Deutschmann, A. J. Org. Chem. 1980,
45, 3722–3724. (d) Buchanan, G. L. J. Chem. Soc., Chem. Commun.
1
981, 814–815. (e) Buchanan, G. L.; Kitson, D. H.; Mallinson, P. R.;
Sim, G. A.; White, D. N. J.; Cox, P. J. J. Chem. Soc., Perkin Trans. 2
1983, 1709–1712. (f) Buchanan, G. L. J. Chem. Soc., Perkin Trans. 1
1984, 2669–2670. (g) Steliou, K.; Poupart, M. A. J. Am. Chem. Soc. 1983,
105, 7130–7138. (h) Brehm, R.; Ohnh €a user, D.; Gerlach, H. Helv. Chim.
Acta 1987, 70, 1981–1987.
21) Lide, D. A., Ed.-in-Chief CRC Handbook of Chemistry and
Physics, 90th ed.; CRC Press: Boca Raton, FL, 2009ꢀ2010; pp 8ꢀ42.
22) Cox, R. A.; Druet, L. M.; Klausner, A. E.; Modro, T. A.; Wan, P.;
Yates, P. Can. J. Chem. 1981, 59, 1568–1573.
23) National Institute of Standards and Technology, NIST Chem-
(
(
(
istry WebBook, http://webbook.nist.gov/chemistry/. Accessed on Jan-
uary 1, 2011.
13
(
24) Typical Values of C NMR chemical shifts for trifluoroacetic
acid (Spectral Data Services, NMR Solvent Chemical Shifts Table; www.
sdsnmr.com/cs_table.html; accessed on December 15, 2010).
(
25) An interesting point is that while protonation of N-methylpyr-
rolidone in solution and in the gas phase is widely understood to occur
on oxygen, fragmentation of this ion in electrospray/mass spectrometry
is consistent with tautomerization to the N-protonated ion prior to
fragmentation. Crotti, A. M.; Fonseca, T.; Hong, H.; Staunton, J.;
Galembeck, S. E.; Lopes, N. P.; Gates, P. J. Int. J. Mass Spectrom.
2
004, 232, 271–276. This may make the interpretation of the site of
protonation in the gas phase more difficult. In this regard the collision-
5b
induced dissociation experiments of Ly et al. may also be of interest..
26) Rasul, G.; Prakash, G. K. S.; Olah, G. A. J. Org. Chem. 1994,
9, 2552–2556.
27) Reich, H. J. WinDNMR: Dynamic NMR Spectra for Windows
J. Chem. Educ., Software 3D2.
28) Benderly, H.; Rosenheck, K. J. Chem. Soc., Chem. Commun.
972, 179–180.
(
5
(
(
1
(
29) Bedn ꢀa rov ꢀa , L.; Malo ꢁn , P.; Bou ꢁr , P. Chirality 2007, 19, 775–786.
(30) Benedetti, E.; Di Blasio, B. J. Chem. Soc., Perkin II
1
980, 500–503.
(
(
31) Cox, R. A.; Yates, K. Can. J. Chem. 1981, 59, 1560–1567.
32) (a) Treschanke, L.; Rademacher, P. THEOCHEM 1985,
1
1
22, 35–45. (b) Treschanke, L.; Rademacher, P. THEOCHEM 1985,
22, 47–56.(c) Rademacher, P. In Greenberg, A.; Breneman, C. M.;
Liebman, J. F. The Amide Linkage: Structural Significance in Chemistry,
Biochemistry, and Materials Science; John Wiley & Sons: New York, 2000;
pp 247ꢀ289.
2
781
dx.doi.org/10.1021/jo200195a |J. Org. Chem. 2011, 76, 2770–2781