Chemical shift assignment of geminal protons in 3,7-diazabicyclo[3.3.1]nonanes: An unexpected deviation from the axial/equatorial chemical shift order

Article abstract of DOI:10.1002/(SICI)1097-458X(199701)35:1<13::AID-OMR27>3.0.CO;2-M

The chemical shift order of axial and equatorial methylene protons in 1,5-disubstituted 3,7-diazabicyclo[3.3.1]nonan-9-ones may be altered by substituents in the 1,5-positions, but the corresponding alcohols behave differently. Unambiguous signal assignme

Full text of DOI:10.1002/(SICI)1097-458X(199701)35:1<13::AID-OMR27>3.0.CO;2-M

MAGNETIC RESONANCE IN CHEMISTRY, VOL. 35, 13È20 (1997)
Chemical Shift Assignment of Geminal Protons in
3,7-Diazabicyclo[3.3.1]nonanes: An Unexpected
Deviation from the Axial/Equatorial Chemical
Shift Order
Adolf Gogoll,* Helena Grennberg and Andreas Axen
Department of Organic Chemistry, University of Uppsala, Box 531, 751 21 Uppsala, Sweden
The chemical shift order of axial and equatorial methylene protons in 1,5-disubstituted 3,7-
diazabicyclo[3.3.1]nonan-9-ones may be altered by substituents in the 1,5-positions, but the corresponding alcohols
behave di†erently. Unambiguous signal assignments for a series of the title compounds are provided, based on 3J
CH
coupling constants and on {1H} 13C heteronuclear Overhauser e†ects. Substituent anisotropy e†ects as a source of
the chemical shift changes are discussed.
Magn. Reson. Chem. 35, 13È20 (1997) No. of Figures: 8 No. of Tables: 6 No. of References: 23
Keywords: NMR; 1H NMR; 13C NMR; 3,7-diazabicyclo[3.3.1]nonane; heteronuclear Overhauser e†ect; HSBC; 2D hetero-
nuclear correlation
Received 19 February 1996; revised 15 August 1996; accepted 19 August 1996
no vicinal coupling constants are available to dis-
tinguish axial and equatorial methylene protons. In
view of the considerable interest in 3,7-
INTRODUCTION
diazabicyclo[3.3.1]nonane derivatives and their element
analogues, it is surprising that a sound assignment of
these protons has rarely been attempted. With few
exceptions, reported assignments of proton spectra are
either based on empirical rules for cyclohexane deriv-
atives,5b,c or they are done without any explanation at
all, or signals are unassigned or not listed.4b,7 However,
an unequivocal assignment is possible based on (i) the
One of the most applicable rules for the assignment of
proton NMR spectra concerns the relative chemical
shift order of axial and equatorial protons in cyclo-
hexane derivatives and their element analogues.1 It is
usually observed that the chemical shift di†erence for
protons in the cyclohexane chair conformation,
*d
\ d [ d , is positive, owing to the anisotropy
eq, ax
eq
ax
of the CÈC single bonds. Exceptions can be observed
in the presence of magnetically anisotropic substit-
uents.2 We report here the complete signal assignments
for a series of 3,7-diazabicyclo[3.3.1]nonane (bispidine)
derivatives, including a deviation from this rule.
heteronuclear vicinal coupling constant (3J ) between
CH
the methylene protons, i.e. H or H and the keto or
ax
eq
the ester carbonyl carbons and (ii) the heteronuclear
Overhauser e†ect from these protons to the same
carbons.
Bispidine derivatives can be involved in conforma-
tional equilibria (Fig. 1), which in simple cases are
revealed by the number of 13C NMR signals.8b If the
nitrogens are unsubstituted (R \ H) or have electron-
RESULTS AND DISCUSSION
2
withdrawing substituents, the twin chair (CC) confor-
mation, possibly Ñattened, is preferred.8,9b The
Bispidine derivatives are of interest as pharmacolo-
gically active compounds3 and as rigid cyclohexane
models,4 and have recently gained new interest as metal
ion scavengers.5 Current work in our laboratory is
aimed at their use as ligands in palladium complexes.6
In the course of this work, we have prepared a series of
1,5-dicarboxymethyl bispidinones (1) (Scheme 1). The
methylene proton signals of these compounds appear as
pairs of AB doublets. Owing to the substitution pattern,
boatÈchair (BC) conformation has been reported for
compounds with substituent R in an endo position,9 or
2
with large substituents on the ring carbons C-2, C-4,
C-6 and C-8.8b,c,9 The degenerate equilibrium CB H BC
has a *Gº between 35 and 40.6 kJ mol~1, depending
on the nitrogen substituents.8b,10 BB conformers have
not been reported so far.8a,c Although we did not Ðnd
any indication for conformers other than CC, even at
low temperatures, initially we cannot completely rule
out fast equilibria with low energy barriers. Therefore,
we consider the two possibilities for the methylene
proton arrangement shown in Fig. 2.
* Correspondence to: A. Gogoll.
Contract grant sponsor: Swedish Natural Science Research Council.
Contract grant sponsor: Magn. Bergvalls Stiftelse.
In the chair (of CC or BC conformer), H and H
would have di†erent angles with the carbonyl carbon
ax
eq
CCC 0749È1581/97/010013È08
( 1997 by John Wiley & Sons, Ltd.

14
A. GOGOLL, H. GRENNBERG AND A. AXEŠ N
Scheme 1.
not perfect, but a distinction of H and H is possible.
ax
eq
On the other hand, in the ring with a boat conforma-
tion of the BC conformer, the coupling constants for
H
and H with the carbonyl carbon (C-9) would not
ax
eq
di†er very much from each other (Table 1). Dihedral
angles between the methylene protons and the carbon
attached to C-1 or C-5 (e.g. the carboxy carbon) are
expected to be similar for H and H in the CC con-
Figure 1. Conformational equilibria of bispidinone derivatives.
ax
eq
former, whereas in the BC conformer, H should have a
ax
larger coupling to these carbons than H . Coupling
eq
constants were measured from cross peaks in HSBC
(Heteronuclear Single quantum multiple Bond
Correlation) spectra13 (Figs 3 and 4 and Table 2). These
cross peaks contain the active, long-range heteronuclear
Figure 2. Newman projection along the C-1—C-2 bond for chair
and boat conformers of the bispidinones.
coupling constant as an antiphase splitting along f .
2
The coupling constant was determined from f traces
2
with J-doubling.14
(C-9). Molecular modelling yields dihedral angles
between 170¡ and 180¡ for the equatorial proton and
between 61¡ and 71¡ for the axial proton, in agreement
with a reported x-ray crystallographic investigation of a
related bispidine derivative11 (Table 1). These angles
correspond to 3J of 9 Hz (H ) and 1È2 Hz (H ), as
Inspection of the HSBC spectrum of 3a shows a pro-
nounced cross peak between the methylene proton at
lower chemical shift, previously assigned as H by its
ax
chemical shift and the ipso carbon of the phenyl substit-
uent. The methylene proton at higher chemical shift has
instead a cross peak with the keto carbonyl carbon (Fig.
3), conÐrming the previous assignments based on
CH
eq
ax
estimated from a Karplus equation:12
*d
.5b However, in the carboxymethyl compounds
3J \ 4.50 [ 0.87 cos h ] 4.03 cos h
(1)
eq, ax
CH
1aÈc the picture is reversed. The cross peaks now indi-
cate that the equatorial proton is the one at lower
chemical shift (Fig. 4). For 1c, which has a Ðxed confor-
Since we have no substituent corrections, the match
between calculated and observed coupling constants is
Table 1. Molecular geometry and calculated coupling constants in bispidinone derivatives
Dihedral angles (¡) and J
(Hz)a
Distances (Ó)
-R -C-9
CH
-C-9
eq
R
R
H
-C-9
H
-R
ax
b
H
H
-R
b
H
-C-9
H
b
H
H
-R
eq 1
b
1
2
ax
1
eq
1
ax
ax
1
eq
N,N¾-Diphenylbispidinec
H
Ph
Ph
Ph
69.0¡
1.2
49.5¡
(J ¼4.1)
174.0¡
9.3
67.5¡
1.9
2.82
2.77
3.29
2.37
2.64
2.56
3.45
3.43
3.23
2.48
2.78
3.29
J
J
J
¼
¼
¼
CH
CH
CH
HH
1b, CCd
1b, BCd
CO Me
71.0¡
1.0
51.0¡
171.4¡
9.2
66.6¡
1.4
2
3.1
5.6¡
7.6
CO Me
124.9¡
3.6
119.2¡
2.8
121.6¡
3.1
2
a Calculated coupling constants 3J ÍEqn (1)Ë for the indicated dihedral angles.
CH
b Angle or distance to the first atom of the substituent R .
1
c From x-ray crystallography.11
d Conformation: CC ¼chair–chair, BC ¼boat–chair; values are for the respective ring conformation. Conformations were optimized with
molecular modelling (MM ½ parameter set16).

PROTON CHEMICAL SHIFTS IN 3,7-DIAZABICYCLO[3.3.1]NONANES
15
Table 2. Vicinal C–H coupling constantsa 3J
and hetero-
CH
nuclear Overhauser e†ect data for various bispidine
derivatives
d
3J
CH
(Hz)
Compound
H
C
NOEc
Proton assignment
1a
3.28
203.6
170.6
203.6
170.6
201.7
170.6
201.7
170.6
200.3
72.4
3.1
4.5
4.7
3.2
4.6
4.7
5.5
4.6
3.7
6.3
6.1
2
½
½
É
½
½
½
É
½
ax
3.15
4.13
3.91
3.70
3.60
3.04
2.38
3.93
3.18
3.65
3.22
eq
ax
eq
ax
eq
eq
ax
eq
ax
eq
ax
1b
1c
2a
2b
3a
Figure 3. Expansion of the HSBC spectrum of 3a showing corre-
200.3
72.4
lation through 3J
from H
2, 4vax
to the carbonyl and the ipso-phenyl carbons
.
CH
and H
2, 4veq
215.0
20.0
7.5
n.d.b
n.d.b
4.8
5.8
n.d.b
n.d.b
5.1
7.2
6.6
4.6
7.3
215.0
20.0
mation, a pronounced cross peak is observed between
H
and NÈCH ÈN, whereas H only has a weak
214.0
18.3
ax
2
eq
cross peak with this carbon, in agreement with the dihe-
dral angles between the atoms involved.
214.0
18.3
The inaccuracy of the calculated 3J values prevents
CH
a quantitative estimation of BC conformers, but the
210.9
142.6
210.9
142.6
É
É
½
½
similarity of 1c with both 1a and 1b indicates that their
contribution must be small.
Additional and independent conÐrmation for the
proton assignments is provided by heteronuclear Over-
hauser e†ects from the protons to keto carbons (C-9)
and the carbons attached to C-1 and C-5 (i.e. carboxyl,
a Measured from HSBC traces with J-doubling (CDCl solutions,
3
25 ¡C).
b Not detected.
c ½, NOE detected; É, no NOE detected; no entry, not measured.
methyl or ipso-Ph). Based on the geometry of the CC
conformers (Fig. 2), H is expected to produce an
ax
equally large NOE for both C-9 and the carboxyl or
ipso-Ph carbons, whereas H should have an e†ect
eq
almost exclusively to the latter. This is conÐrmed by the
experiment (Fig. 5 and Table 2). For the BC conformers,
we would predict a strong e†ect from H to the carbox-
ax
yl (or ipso-Ph) carbon and no (or very weak) e†ects
from any methylene proton to C-9. Since this was not
the case, we conclude that the compounds occur pre-
dominantly if not exclusively in their CC conformation.
In the methyl derivatives 2, it is more convenient to
observe the homonuclear NOE from the methyl
protons to H , which is larger than the e†ect to H
e.g. 6% vs. 2.5% for 2b, which Ðts well to the chairÈchair
geometry.
,
Figure 4. Expansion of the HSBC spectrum of 1a showing corre-
to the carbonyl and the carboxyl carbons from
ax
eq
lation through 3J
H
CH
2, 4veq
and H
.
2, 4vax
Figure 5. Heteronuclear NOE (Ê1HÌ ] 13C) difference spectra for 1a. (a) 13C NMR reference spectrum; (b) saturation of proton signal at
3.28 ppm (c) saturation of proton signal at 3.15 ppm.

Table 3. 1H and 13C NMR data for the bispidine derivatives (CDCl solution, 25 ÄC, 100–200 mg)
3
1a
1b
1c
2a
2b
2ca
d, 13C
d, 1H
d, 13C
d, 1H
d, 13C
d, 1H
d, 13C
d, 1H
d, 13C
d, 1H
d, 13C
d, 1H
CO
203.4
170.2
137.1
128.7
128.1
127.1
60.7
201.7
169.6
149.6
129.0
120.9
117.4
200.3
168.8
215.0
214.0
215.4
CO CH
2
ipso
3
138.4
128.7
128.2
127.0
61.3
149.5
115.4
129.0
118.9
ortho
meta
para
7.31
7.01
6.81
7.31
7.31
3.96
3.28
3.15
7.27
6.94
7.32
7.18
6.77
PhCH
72.4b
4.05
3.53
3.04
2.38
2
CH
59.7
57.8
4.13
3.91
60.7
3.70
3.60
65.5
62.3
3.93
3.18
61.8
3.36
2
ax
ax
ax
eq
ax
eq
ax
eq
ax
2.95
eq
eq
eq
J ¼11.1
3.86
J ¼12.0
3.73
J ¼12.8 Hz
3.76
J ¼11.0 Hz
J ¼12.2 Hz
J ¼12.2 Hz
CH O C
52.1
58.9
52.6
59.0
52.3
55.9
3
C-1, C-5
2
46.7
20.0
47.0
18.3
49.4
17.2
C-CH
0.96
1.16
0.88
3
3a
4a
4b
4c
4d
d, 13C
d, 1H
d, 13C
d, 1H
d, 13C
d, 1H
d, 1H
d, 1H
CO
210.9
COO
173.8
173.5
Ph-ipso
Bz-ipso
Bz-ortho
Bz-meta
Ph-meta
Bz-para
Ph-ortho
Ph-para
142.6
137.9
129.0
128.3
127.7
127.4
126.8
126.5
ipso-7
137.7
Ph-ipso
150.2, 149.6
129.0, 128.9
119.4, 119.1
115.8
ipso-3
meta
136.7
Ph-meta
Ph-para
Ph-ortho
7.20
6.79
6.92
7.36
meta : 7.19
para : 6.72
7.51
128.3
7.19–7.38
4.24
7.46
7.36
7.38
7.29
7.27
ortho
128.9, 128.7
127.32, 127.30
74.8
ortho : 6.80
para
CH(OH)
CH(OH)
70.9
4.62
2.86
3.33
CH O
3
C-6, C-8
52.4
59.6
3.71
CH O
3
C-6, C-8
52.7
55.7
3.83
3.79
CH : 1.08
1.14
3.55
2.75
3.30
3.12
3
2.61
eq
2.50
eq
eq
ax
eq
ax
eq
ax
eq
ax
2.47
ax
3.33
3.63
3.75
1.84
2.64
2.58
ax
eq
ax
CH
64.8
3.65
3.22
C-2, C-4
53.4
2.63
eq
C-2, C-4
2
eq
ax
3.60
ax
J ¼10.5
3.82
PhCH
61.8
54.4
PhCH -7
61.24
61.20
49.2
3.50
3.66
PhCH -7: 3.50
2
2
PhCH -3
2
2
PhCH -3: 3.56
2
C-1, C-5
48.1
OH: 5.13
OH: 3.46
OH: 4.57
OH: 2.02
a NH d 2.67.
b N—CH —N.
2

PROTON CHEMICAL SHIFTS IN 3,7-DIAZABICYCLO[3.3.1]NONANES
17
H-2 and H-4 are revealed by their HSBC cross peak
with C-6/C-8. In addition, all equatorial protons are
ax
ax
Table 4. Comparison of methylene proton chemical shifts in
structurally related bispidine derivatives (R = CH
having cross peaks to the other CH carbon in the same
1
3
vs. R = COOCH )
2
ring, e.g. H-2 with C-4. The signals of C-2/4 are also
eq
1
3
Compound
R
d H
d H
DdH
a
DdH
a
Dd
eq, ax
0.66
identiÐed by a cross peak with CH(OH).
1
ax
eq
ax
eq
Finally, it should be noted that the methylene proton
signals in alcohols 4aÈd are not doublets as in the
parent ketones. Instead, the protons of each piperidine
ring give rise to an AA@BB@ pattern. Additional coup-
lings into the other piperidine ring result in
AA@BB@CC@DD@ spin systems (Fig. 6). Coupling con-
stants were determined by spin simulation using full-
lineshape iteration (simulations were performed with
gNMR v. 3.615). Resulting parameter sets are shown in
Table 5.
2a
CH
2.38 3.04
3.28 3.15
3.18 3.93
4.13 3.91
2.58 2.64
3.60 2.63
1.84 2.50
2.47 2.61
3.12 3.30
3.75 3.63
2.75 3.55
3.33 3.79
3
COOCH
1a
2b
1b
0.90
0.95
1.02
0.63
0.63
0.58
0.11 É0.13
0.75
3
3
3
3
3
3
CH
3
COOCH
É0.02 É0.22
0.06
4c, C /C
CH
2
4
4
8
8
3
4a, C /C
COOCH
É0.01 É0.97
0.66
2
4c, C /C
CH
6
4a, C /C
6
3
COOCH
0.11
0.14
0.18
4d, C /C
CH
2
4
4
8
8
3
4b, C /C
COOCH
0.33 É0.12
0.80
2
4d, C /C
CH
6
4b, C /C
6
3
Compounds 1 are b-keto esters, and an interesting
COOCH
0.24
0.46
question is whether the observed chemical shift order,
a d
Éd
.
R1¼COOCH3
R1¼CH3
i.e. *d
eq, ax
\ 0, is related to the conformationally
restricted bispidinone skeleton. Unexpectedly, relevant
data on 2-carboalkoxycyclohexanones are scarce in the
literature at best. We therefore investigated the keto-
carboxylic acid esters 5 and 6. A complete proton signal
assignment for their keto isomers (Table 6) shows that
Alcohols 4aÈd show the normal shift order, i.e.
*d
[ 0, except for the protons on the OH-side, i.e.
eq, ax
C-2/C-4 of 4a and 4b (Tables 3 and 4). Here, the assign-
indeed *d
\ 0 for the methylene protons on C-3,
eq, ax
ment of the axial protons H-6 and H-8 is based on
but not for those on C-5. The similarity of 5 and 6
proves that the nitrogen atom is not important. Based
on the appearance of the spectra, it is reasonable to
assume the main conformer of 5 and 6 to be a Ñattened
chair with H-2 in an axial position (Table 6). This is
supported by a molecular modelling and conforma-
ax
ax
their NOE with CH(OH). Because of the CC geometry
of these molecules, the axial methylene protons also
have a strong cross peak with C-2/C-4 in the HSBC
spectrum, which is not observed for the equatorial
protons H-6 and H-8 . Likewise, the axial protons
eq
eq
Figure 6. (a) Methylene proton signals of 4b (CDCl solution, 400 MHz, 25 ¡C). The FID was processed with Gaussian apodization prior
to Fourier transformation. (b) Simulation of the signals with the parameters from Table 5, using a Gaussian lineshape.
3
Table 5. Spin simulation results for the methylene protons of alcohols 4a–da
Compound
R
R
dH-2
ax
dH-2
eq
dH-4
ax
dH-4
eq
dH-6
ax
dH-6
eq
dH-8
ax
dH-8
1
2
eq
4a
4b
4c
4d
CO Me
Bz
Ph
Bz
Ph
3.597
3.747
2.581
3.123
2.627
3.625
2.636
3.303
3.597
3.747
2.581
3.123
2.627
3.625
2.636
3.303
2.467
3.325
1.840
2.750
2.608
3.792
2.499
3.554
2.467
3.325
1.840
2.750
2.608
3.792
2.499
3.554
2
CO Me
2
Me
Me
J
J
J
J
J
J
J
J
J
2eq, 2ax
2eq, 4eq
1.00
2ax, 4eq
6eq, 6ax
6eq, 8eq
1.60
6eq, 8ax
2ax, 8ax
2eq, 8eq
—
2eq, 8ax
4a
4b
4c
4d
CO Me
Bz
Ph
Bz
Ph
É11.28
É12.10
É10.59
É11.67
É0.57
É0.46
É0.63
É0.21
É10.76
É12.03
É10.89
É11.84
É0.64
É0.77
É0.52
É0.89
0.27
1.43
0.03
1.97
0.40
0.40
0.54
0.41
2
CO Me
2
0.43
0.80
0.54
1.16
1.86
1.25
0.06
—
Me
Me
—
a Simulated for 400 MHz observation frequency with full-lineshape iteration of experimental spectra.15 J in Hz. Further couplings:
J
J
¼J
.
, J
¼J
, J
¼J
, J
¼J
, J
¼J
, J
¼J
, J
¼
4eq, 4ax
2eq, 8ax
2eq, 2ax
2eq, 4ax
2ax, 4eq
8eq, 8ax
6ax, 6eq
6eq, 8ax
6ax, 8eq
4ax, 6ax
2ax, 8ax
4eq, 6eq
2eq, 8eq
4eq, 6ax

18
A. GOGOLL, H. GRENNBERG AND A. AXEŠ N
Table 6. 1H NMR data for the keto isomers of 5a (benzene-d solution) and 6b (CDCl solution)c
6
3
5
H-2
H-3
ax
H-3
eq
H-4
ax
H-4
eq
d
3.05
J
1.95
J
1.66
J
1.00
J
1.33
J
J
(Hz)
¼10.3
¼5.6
¼13.6
¼10.3
¼3.4
¼5.6
¼2.0
¼2.0
¼12.8
¼1.3
HH
2, 3ax
3ax, eq
3eq, 4eq
3eq, 4ax
3eq, 5eq
4ax, eq
4eq, 6eq
J
J
J
J
J
J
2, 3eq
2, 6ax
3ax, 4ax
3ax, 4eq
¼1.3
6
d
3.45
J
3.08
J
2.93
J
J
(Hz)
¼7.7
¼5.0
¼1.4
¼11.6
¼1.3
¼1.6
HH
2, 3ax
3ax, eq
3eq, 5eq
J
J
J
2, 3eq
2, 6ax
3ax, 5ax
5
H-5
H-5
H-6
H-6
eq
CH CH
2
CH CH
2
ax
eq
ax
3
3
d
1.21
J
1.31
J
1.81
J
2.19
4.01
0.98
J
(Hz)
¼13.2
¼10.7
¼5.0
¼5.6
¼5.0
¼13.6
¼13.6
J ¼7.2
J ¼7.2
HH
5ax, eq
5eq, 6ax
5eq, 6eq
6ax, eq
J
J
J
5ax, 6ax
5ax, 6eq
6
d
2.76
J
2.82
J
2.54
J
2.64
4.22
1.26
J
(Hz)
¼11.3
¼8.0
¼4.8
¼6.5
¼5.5
J ¼7.2
Bz: 3.64
J ¼7.2
HH
5ax, eq
5ax, 6ax
5ax, 6eq
5eq, 6eq
J
6ax, eq
J
J
5eq, 6ax
Ph: 7.19–7.31
a Ketone : enol ¼20 : 80 (benzene-d solution), 40 : 60 (CDCl solution).
b Ketone : enol ¼38 : 62 (CDCl solution).
c For better comparability, corresponding positions in 5 and 6 are numbered equally (Scheme 1).
6
3
3
tional search, which show that conformers with an
equatorial carboethoxy substituent account for ca. 80%
of the population (using the MM] force Ðeld in Hyper-
chem 4.016). Small four-bound couplings between
protons assigned to occupy axial positions indicate the
presence of further conformers in solution.
The source of the reversed chemical shift order of
methylene protons in compounds 1aÈc, 5 and 6 and in
part alcohols 4a and b, is not immediately obvious. It
has previously been speculated whether variations of
since the anisotropic shielding by these bonds is pri-
marily responsible for the chemical shift di†erence of
axial and equatorial methylene protons.1 However, cal-
culations did not provide any support, such as a pattern
of bond orders, electron densities or molecular geome-
tries, for this hypothesis (the semi-empirical methods
AM1 and PM3 as implemented in MOPAC 6.0 were
used).19
The anisotropy of the carboxyl ester substituent
might be another source of the observed chemical shift
changes. We therefore consider the methylene proton
chemical shifts in the cyclohexane derivatives of Scheme
*d
in piperidone derivatives could be a result of the
eq, ax
spatial arrangement of the nitrogen lone pair, with a
shift to higher frequencies for protons syn to it.8a,17 This
possibility has been rejected by others,18 and can also
be discarded in the present case.
2, and in the bispidine derivatives with either CH or
3
COOCH in positions 1 and 5 (*dH and *dH
,
3
ax
eq
Table 4). Obviously, the carboxyl ester group increases
the chemical shifts of the axial protons more than that
of the equatorial protons, especially with a keto group
in the b-position, which might inÑuence its orientation.
It is tempting to assume that a change in electron
density of the bonds between C-1/C-5 and C-9 causes
the di†erent chemical shift order for compounds 1aÈc,
Scheme 2.

PROTON CHEMICAL SHIFTS IN 3,7-DIAZABICYCLO[3.3.1]NONANES
19
dinones (0.5 mmol) with NaBH in THFÈwater (1 : 1) (5 ml). The reac-
If there were free rotation, equally large chemical shift
increases for both axial and equatorial protons would
be expected, which is not observed.
4
tion mixture was stirred for 1 h at room temperature, then cooled on
an ice-bath and the pH was adjusted to ca. 1 with 50% aqueous HCl.
Stirring was continued for an additional 30 min at room temperature.
The reaction mixture was transferred to a separation funnel and the
pH was adjusted to ca. 10 with 20% aqueous NaOH. This solution
was extracted with chloroform (2 ] 20 ml) and the combined organic
phases were washed with water (50 ml). After evaporation, the remain-
ing material was recrystallized from diethyl ether at [ 20 ¡C. The
colorless crystals were collected and dried in vacuo. 4a: yield 34%;
m.p. 100È104 ¡C; IR (KBr), 3488, 1734, 1721, 1260 cm~1; MS, m/z 410
EXPERIMENTAL
(M`, 100%), 258, 105, 77; analysis, calculated for C
H
N O , C
Spectra
23 26
2 5
67.34, H 6.34; found, C 66.45, H 6.31%. 4b: yield 80%; m.p. 122 ¡C;
IR (KBr), 3278, 2833, 1332, 1124 cm~1; MS, m/z 438 (M`, 10.4%),
Spectra were measured on a Varian Unity instrument at
347, 302, 134, 91; analysis, calculated for C
H
N O , C 68.51, H
25 30
2 5
25 ¡C on solutions in CDCl at 400 MHz for 1H and
6.84; found, C 68.39, H 6.86%. 4c: yield 72%; m.p. 112È113 ¡C; 13C
3
NMR (CDCl ), d 138.9, 137.9, 128.6, 128.5, 128.2, 128.0, 126.9, 126.7,
100 MHz for 13C, respectively. Chemical shifts are indi-
3
81.0, 65.0, 61.8, 61.7, 59.4, 36.6, 23.7 ppm; IR (KBr), 3354, 2948, 2929,
rectly referenced to TMS via the solvent signals (1H,
residual CHCl at 7.26 ppm; 13C, CDCl at 77.0 ppm).
2868, 2801, 1494. 1453 cm~1; MS, m/z 350 (M`, 10%), 259, 198, 91;
analysis, calculated for C
H
N O, C 78.82, H 8.63; found, C 78.61,
3
3
23 30
2
Signal assignments were derived from HSQC,13
HSBC,20 P.E.COSY21 and NOE di†erence22 spectra.
Heteronuclear Overhauser e†ect di†erence spectra23
were obtained on non-degassed samples of 100È200 mg
by selectively saturating the signal of the respective
12C-bound proton for 30 s, followed by acquisition
of the 13C NMR spectrum with Waltz-16 proton
decoupling, acquisition time 1.2 s, 384 transients and
experiment time 3 h 20 min. A reference spectrum was
obtained with the decoupler set o†-resonance.
H 8.72%. 4d: yield, 22%; m.p. 97 ¡C; 13C NMR (CDCl ), d 150.3 (2C),
3
128.9 (2C), 118.3, 117.7, 114.9, 114.3, 77.9, 60.7, 53.2, 36.4, 22.4 ppm;
IR (CDCl solution), 3064, 2253, 1598, 1503, 1464 cm~1; MS, m/z 322
3
(M`, 10%), 184, 120, 106; analysis, calculated for C
H
N O, C
21 26
2
78.22, H 8.13; found, C 78.39, H 8.83%.
Methylcyclohexane 7: 1H NMR (CDCl ), d 1.68 (m, 2H, H-3 ),
eq
1.65 (m, 2H, H-2 ), 1.62 (m, 1H, H-4 ), 1.34 (m, 1H, H-1 ), 1.23 (m,
eq eq ax
2H, H-3 ), 1.14 (m, 1H, H-4 ), 0.88 (m, 2H, H2 ), 0.86 (d, J \ 6.5 Hz,
ax ax ax
CH ). Cyclohexane carboxylic acid 8: 1H NMR (CDCl ), d 2.34 (tt,
3
3
3
J \ 3.7, 11.2 Hz, 1H, H-1), 1.94 (m, 2H, H-2 ), 1.77 (m, 2H, H-3 ),
ax eq
1.65 (m, 1H, H-4 ), 1.46 (m, 2H, H-2 ), 1.30 (m, 2H, H-3 ), 1.25 (m,
eq
ax
ax
1H, H-4 ).
ax
Materials
Acknowledgements
Bispidinone derivatives (Scheme 1) were prepared by a Mannich reac-
tion from the corresponding acyclic ketones and benzylamine or
aniline;6b 3a, 3b and 6 were prepared as described in the literature.5a,c
Complete NMR data are given in Table 3. Compounds 5, 7 and 8
were commercially available and used without further puriÐcation.
Alcohols 4aÈd were prepared by reduction of the corresponding bispi-
We thank Professor R. Freeman for supplying us with the source code
for his J-doubling routine. The Swedish Natural Science Research
Council and Magn. Bergvalls Stiftelse are acknowledged for Ðnancial
support.
REFERENCES
1. (a) F. A. Bovey, Nuclear Magnetic Resonance Spectroscopy,
2nd ed. Academic Press, San Diego (1988); (b) J. W. Emsley,
J. Feeney and L. N. Sutcliffe, High Resolution NMR Spectros-
copy, p 696. Pergamon Press, Oxford (1966); (c) H. Booth,
Prog. Nucl. Magn. Reson. Spectrosc. 5, 149 (1969); (d) R. A.
Johnson, J. Org. Chem. 33, 3627 (1968); (e) H. P. Hamlow,
S. Okuda and N. Nakagawa, Tetrahedron Lett. 2553, (1964);
(f) F. Bohlmann, D. Schumann and H. Schulz, Tetrahedron
Lett. 173, (1965).
2. (a) K. M. Wellman and F. G. Bordwell, Tetrahedron Lett. 25,
1703 (1963); (b) K. L. Williams and W. S. Johnson, J. Org.
Chem. 20, 4623 (1961).
3. (a) U. Holzgrabe and E. Erciyas, Arch. Pharm. (Weinheim)
325, 657 (1992); (b) R. Jeyaraman and S. Avila, Chem. Rev.
81, 149 (1981).
4. (a) B. W. Gung and M. A. Wolf, J. Org. Chem. 61, 232
(1996); (b) E. M. Gonikberg and W. J. le Noble, J. Org.
Chem. 60, 7751 (1995); (c) A. S. Cieplak, J. Am. Chem. Soc.
103, 4540 (1981).
5. (a) G. D. Hosken and R. D. Hancock, J. Chem. Soc., Chem.
Commun., 1363 (1994); (b) D. St. C. Black, G. B. Deacon
and M. Rose, Tetrahedron 51, 2055 (1995); (c) D. St. C.
Black, M. A. Horsham and M. Rose, Tetrahedron 51, 4819
(1995).
6. (a) A. Gogoll, J. Gomes, M. Bergkvist and H. Grennberg,
Organometallics 14, 1354 (1995); (b) A. Gogoll, H. Grenn-
berg and A. Axen, Organometallics (in press).
7. (a) R. Caujolle, P. Castera and A. Lattes, Bull Soc. Chim. Fr.
414 (1984); (b) T. Sasaki, S. Eguchi, T. Kiriyama and Y.
Sakito, J. Org. Chem. 38, 1648 (1973).
8. (a) J. E. Douglass and T. B. Ratliff, J. Org. Chem. 33, 355
(1968); (b) P. H. McCabe, N. J. Milne and G. A. Sim, J.
Chem. Soc., Chem. Commun. 625 (1985); (c) N. S. Zefirov,
Russ. Chem. Rev. 44, 196 (1975); (d) P. J. Cox, P. H.
McCabe, N. J. Milne and G. A. Sim, J. Chem. Soc., Chem.
Commun. 626 (1985).
9. (a) H. Quast, B. Muller, E.-M. Peters, K. Peters and H. G. von
Schnering, Chem. Ber. 115, 3631 (1982); (b) L. M. Jackman,
T. S. Dunne, B. Muller and H. Quast, Chem. Ber. 115, 2872
(1982).
10. Y. Takeuchi, P. Scheiber and K. Takada, J. Chem. Soc., Chem.
Commun., 403 (1980).
11. O. I. Levina, K. A. Potekhin, E. N. Kurkutova, Y. T. Struchkov,
V. A. Palyulin and N. S. Zefirov, Dokl. Akad. Nauk SSSR 277,
367 (1984).
12. R. Aydin and H. Gunther, Magn. Reson. Chem. 28, 448
(1990).
13. K. E. Kover, O. Prakash and V. J. Hruby, Magn. Reson. Chem.
31, 231 (1993).
14. (a) L. McIntyre and R. Freeman, J. Magn. Reson. 96, 425
(1992); (b) F. del Rio-Portilla, V. Blechta and R. Freeman, J.
Magn. Reson. A111, 132 (1994).
15. gNMR v. 3.6. Cherwell Scientific, Oxford (1995).
16. Hyperchem 4.0. Hypercube, Waterloo, Ont (1994).
17. (a) J. B. Lambert, R. G. Keske, R. E. Cathcart, A. P. Jovano-
vich, J. Am. Chem. Soc. 89, 3761 (1967); (b) S. F. Nelsen, P.
J. Hintz and R. T. Landis, J. Am. Chem. Soc. 94, 7105
(1972).

20
A. GOGOLL, H. GRENNBERG AND A. AXEŠ N
18. M. J. T. Robinson, Tetrahedron Lett. 1153 (1968).
19. D. R. Armstrong, P. G. Perkins and J. J. P. Stewart, J. Chem.
Soc., Dalton Trans., 838 (1973).
21. L. Mueller, J. Magn. Reson. 72, 191 (1987).
22. J. K. M. Sanders and J. D. Mersh, Prog. Nucl. Magn. Reson.
Spectrosc. 15, 353 (1982).
20. (a) G. Bodenhausen and D. J. Ruben, Chem. Phys. Lett. 69,
185 (1980); (b) A. M. Torres, T. T. Nakashima and R. E. D.
McClung, J. Magn. Reson. A102, 219 (1993).
23. K. E. Kover and G. Batta, Prog. Nucl. Magn. Reson. Spectrosc.
19, 223 (1987).