Assessable consequences of through-bond donor-acceptor interactions in β-aminoketones

Article abstract of DOI:10.1002/chem.200701220

Reported are the syntheses of ester-functionalized (6-8) and alkylsubstituted (9) 1-aza-adamantanones; the easy handling of the compounds provides an opportunity to comprehensively study the fundamental changes in structure and reactivity that can accompany the donor-acceptor arrangement in rigid β-aminoketones. X-ray structural analysis of trione 6 and dione 7 reveals bond length and angle variations consistent with throughbond (hyperconjugative) donor-acceptor interactions. Observed is a shortening of the C-N bond, elongation of the central C-C bond (to ≈1.6 A), and a significant pyramidalization of the car bonyl carbon within the donor-σ-acceptor pathway. UV/Vis spectra of 6-9 show a new absorption maximum (λ_max = 260-275 nm in three solvents), the so-called σ-coupled transition ; the molar absorptivity scales with the number of carbonyl groups (for trione 6, ε ≈ 3000, for dione 7, ε ≈ 2000) and the band reversibly disappears upon addition of acid. IR and ¹³C NMR spectroscopic data show trends consis tent with through-bond donation to the carbonyl acceptor groups and commensurate weakening of the carbonyl π bond. High yielding acid-mediated fragmentations are used to illustrate the effects of the donor-acceptor arrangement on the reactivity of the molecules. Given that donor-σ-acceptor molecules have recently been found to show self-assembly behavior and macromolecular properties linked to their unusual structure, the current analysis encourages further consideration of the systems in advanced materials applications.

Full text of DOI:10.1002/chem.200701220

DOI: 10.1002/chem.200701220
Assessable Consequences of Through-Bond Donor–Acceptor Interactions in
b-Aminoketones
Andrew J. Lampkins, Yan Li, Alexandre Al Abbas, Khalil A. Abboud,
Ion Ghiviriga, and Ronald K. Castellano*[a]
In memory of Professor Dmitry M. Rudkevich
Abstract: Reported are the syntheses
of ester-functionalized (6–8) and alkyl-
substituted (9) 1-aza-adamantanones;
the easy handling of the compounds
provides an opportunity to comprehen-
sively study the fundamental changes
in structure and reactivity that can ac-
company the donor–acceptor arrange-
ment in rigid b-aminoketones. X-ray
bonyl carbon within the donor–s-ac-
ceptor pathway. UV/Vis spectra of 6–9
tent with through-bond donation to the
carbonyl acceptor groups and commen-
surate weakening of the carbonyl p
bond. High yielding acid-mediated
fragmentations are used to illustrate
the effects of the donor–acceptor ar-
rangement on the reactivity of the mol-
ecules. Given that donor–s-acceptor
molecules have recently been found to
show self-assembly behavior and mac-
romolecular properties linked to their
unusual structure,the current analysis
encourages further consideration of the
systems in advanced materials applica-
tions.
show
a new absorption maximum
(lmax =260–275 nm in three solvents),
the so-called “s-coupled transition”;
the molar absorptivity scales with the
number of carbonyl groups (for trione
6, e ꢁ3000,for dione 7, e ꢁ2000) and
the band reversibly disappears upon
addition of acid. IR and 13C NMR
spectroscopic data show trends consis-
structural analysis of trione
6 and
dione 7 reveals bond length and angle
variations consistent with through-
bond (hyperconjugative) donor–accept-
or interactions. Observed is a shorten-
Keywords: donor–acceptor
sys-
tems · hyperconjugation · reactivity ·
strained molecules · through-bond
interactions
ꢀ
ing of the C N bond,elongation of the
ꢀ
central C C bond (to ꢁ1.6 Š),and a
significant pyramidalization of the car-
Introduction
mantanones 3 (and related structures; Figure 1). Arrival at
the desired donor–acceptor geometry (with respect to the
saturated two-carbon bridge) in the donor–s-acceptor mole-
cules[3] is often associated with a unique (split) p–p* absorp-
tion (s-coupled transition:[4] lmax =220–260 nm; e=500–
Cyclic b-aminoketone derivatives are unique molecular
models to determine the effects of through-bond[1] (hyper-
conjugative[2]) interactions on structure,reactivity,stereose-
lectivity,and photophysical properties. The donor (amine)
and acceptor (carbonyl) interaction is optimized when the
ꢀ
nitrogen lone pair,C
Cb bond,and carbonyl p system
a
(Figure 1) are in a zig–zag arrangement; this situation is
transiently maintained for molecules such as N-methylpiper-
idone 1 and tropinone 2,but permanently fixed in aza-ada-
[a] A. J. Lampkins,Y. Li,A. Al Abbas,Dr. K. A. Abboud,
Dr. I. Ghiviriga,Prof. R. K. Castellano
Department of Chemistry,University of Florida
P.O. Box 117200,Gainesville,FL 32611 (USA)
Fax : (+1)352-846-0296
Figure 1. Donor–s-acceptor arrangement in cyclic b-aminoketones; N-
methylpiperidone (1),tropinone ( 2),1-aza-adamantanone ( 3a),and 13,-
diaza-adamantanone (3b). Note that the appropriate nitrogen configura-
tion is fixed only in 3a and 3b.
E-mail: castellano@chem.ufl.edu
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
1452
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2008, 14,1452 – 1463

FULL PAPER
2500) and a red-shifted (lmax =285–315 nm; e=30–50) and
intensified n–p* absorption in the UV spectra.[4,5] Studies
that attribute ground-state structural and related reactivity
changes to donor–s-acceptor interactions in b-aminoketones
have often done so despite the presence of competing (and
often times stronger) inductive,electrostatic,steric,and/or
solvation effects. Even so,the most carefully constructed ex-
amples show that donor–acceptor through-bond interactions
can play at least a supporting role in perturbations of struc-
ture and reactivity in these systems.
ny the donor–s-acceptor arrangement in these 1-aza-ada-
mantanones,and b-aminoketones in general. Included is dis-
cussion of the first X-ray crystal structure of an AAT (6)
and second reported structure of an aza-adamantanedione
(AAD; 7) that offer compelling structural evidence for
through-bond interactions,complemented by spectroscopic
(NMR,IR,and UV) studies and comparisons to the litera-
ture and appropriate model compounds 8 and 9. The conse-
quences of through-bond interactions on reactivity are dis-
cussed in the context of nitrogen basicity/nucleophilicity and
an acid-promoted core fragmentation. The conclusion from
the studies is that the through-bond donor–acceptor interac-
tions in the aza-adamantanones are both assessable and sig-
nificant enough to play a role in their ground-state structure,
reactivity,and self-assembly properties.
Recently we have exploited the unusual 1-aza-adamanta-
netrione[6] (AAT) (Figure 2) framework to explore the con-
sequences of donor–acceptor through-bond interactions at
Results and Discussion
Synthesis: Several ester-functionalized aza-adamantanones
have been prepared for these studies (Schemes 1 and 2). In
each case,isopropyl esters have been selected due to their
efficient preparation under Fischer esterification conditions,
good solubility (in part by discouraging aggregation of the
core),and stability (to Lewis and protic acids,and mild
bases). The first, 6 (Scheme 1),like 5,bears only one meth-
ylene spacer between the core and ester carbonyl. To pre-
pare 6,triacid 10 (made in three steps from commercially
available 1,3,5-trimethoxybenzene[7b]) is esterified to give
triisopropyl ester 11 which is then deprotected with BBr3 to
afford the phloroglucinol cyclization substrate 12.[8] Ester-
functionalized AAT 6 is then formed upon reaction of 12
with hexamethylenetetramine (HMTA),in the appropriate
alcoholic solvent (iPrOH) in 66% yield. This particular cyc-
lization works best at higher concentrations (ꢂ0.1m). As ex-
pected,triisopropyl ester 6 is markedly more soluble than
the previously synthesized amide derivative 5.
The reasonable solubility of triester 6 allowed us to ex-
plore carbonyl hydride reductions at low temperatures with
the goal of isolating and characterizing the corresponding al-
cohols,and analyzing the structural consequences of
through-bond donor–acceptor interactions within a structur-
ally related aza-adamantanone family. Indeed,the hydride
reduction of cyclic b-aminoketones has been used extensive-
ly to explore the influence of through-bond effects on car-
bonyl facial selectivity,[9] and elegant work with the aza-ada-
mantanones[9e,f,10] (e.g. 3a),diaza-adamantanones (e.g.
3b),[11] and related monoketones[12] have provided among
the more compelling illustrations of the Cieplak hyperconju-
gative transition-state model.[9a] Unexpected reactivity
emerges from treatment of 6 (8 and 9) with LiAlH4,Red-
Al,and DIBAL-H under various conditions—in studies that
parallel (in design) those recently performed on a related
phloroglucinol-derived natural product[13]—and these results
will be reported separately. Of relevance to the current dis-
cussion,treatment of 6 with one equivalent of Red-Al pro-
vides a modest (33%) yield of the exo-mono-reduction
Figure 2. 1-Aza-adamantanones considered in the previous and current
studies.
the macromolecular/supramolecular level.[7] Molecules like
4,for example,form gel phases at low concentration in or-
ganic solution,a consequence of their shape (conforma-
tional preferences),ground-state dipole (and dipolar interac-
[7a]
tions),and aromatic substituents.
Installation of amide
functional groups proximal to the AAT core causes remark-
able changes in macromolecular behavior as shown by 5; ag-
gressive solution-phase assembly is observed that also re-
sults in the moleculeꢁs displaying long-range order in the
solid state.[7b] The enhancement is not,based on IR and
NMR studies,exclusively (or even mainly) the result of
more conventional intermolecular hydrogen bonding be-
tween monomers,but rather well-defined interactions be-
tween the polar amide groups and polar AAT core (that in-
clude weak intramolecular hydrogen-bonding contacts) that
modulate conformational preferences,molecular dipole,and
inter-/intramolecular dipolar interactions. More recent high-
level computational work suggests a dramatic role of these
amides in tuning the fundamental electronic properties of
the molecules making them potentially suitable for electron-
ic and optoelectronic applications.[7c,d] Although multiple in-
teractions are at work,experimental and theoretical studies
support the notion that through-bond interactions a) can be
“tuned” in these systems and b) offer a subtle way to ad-
dress the macromolecular properties that emerge upon mo-
lecular self-assembly.
Reported here are the syntheses of ester-functionalized
aza-adamantanones (6–8,Figure 2) the easy handling of
which provides an opportunity to better study the funda-
mental changes in structure and reactivity that can accompa-
Chem. Eur. J. 2008, 14,1452 – 1463
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
1453

R. K. Castellano et al.
Scheme 1. Synthesis of ester-functionalized 1-aza-adamantanetrione
and exo-mono-alcohol 7.
6
Scheme 3. Synthesis of tripentyl 1-aza-adamantanetrione 9. Tripropyl 9a,
synthesized by an alternate route,has been reported previously (see ref.
[7a]).
product 7 (Scheme 1),in addition to more polar di- and
trireduced compounds.
The synthesis of homologated 1-aza-adamantanetrione
triester 8 is outlined in Scheme 2. Tribromo scaffold 13[7a,14]
is converted to tripropionic acid 16 using a three-step se-
quence involving alkylation with diethyl malonate,saponifi-
cation,and thermally-mediated decarboxylation (78% over-
all yield for three steps). Fischer esterification then provides
demethylation and HMTA cyclization reactions follow to
afford the desired tripentyl AAT 9 (that is somewhat more
soluble than previously prepared 9a[7a]). This procedure ap-
pears as our most general one to prepare alkyl- and aryl-
substituted AATs.
17 in good yield,and a carefully monitored BBr -mediated
X-ray crystal structures of 6 and 7—General features and
packing: X-ray crystallography is among the best experi-
mental techniques to reveal the structural consequences
(e.g.,on bond lengths and angles) of hyperconjugative inter-
actions in donor–s-acceptor molecules.[19] No X-ray struc-
tures of the AAT core have been reported in the Cambridge
Structural Database (v. 5.28) as of May 2007; only after nu-
merous attempts were we finally able to grow suitable single
crystals of 6 (from pentane/ethanol). The asymmetric unit
consists of one molecule of 6 and two half ethanol mole-
cules,and the structure is disordered with respect to both
components. For the AAT,the disorder is localized in the
isopropyl side chains,and could be appropriately modeled
in the absence of the solvent. The refined monomer struc-
ture is shown in Figure 3a as a stick view and Figure 3b as
an ORTEP plot; additional crystallographic details (e.g.,pa-
rameters,bond lengths,and bond angles) are summarized in
the next section as well as the Experimental Section.
3
demethylation affords 18.[8] The Mannich-type cyclization
with HMTA provides AAT 8 in 47% yield,when the reac-
tion is again performed in the appropriate alcoholic solvent
(iPrOH).[15] Triester 8 is found to be readily soluble in
common organic solvents.
The packing structure of 6 is worth mentioning given our
interest in the self-assembly of these and related molecules
in solution and the solid state. Recognizable is an antiparal-
lel alignment of monomers (Figure 3c),a consequence pre-
sumably of favorable dipolar interactions (the calculated
ground state dipole of similar AATs is ꢁ4 D);[7c] the core-
to-core distance is ꢁ6.8 Š. Other intermolecular interac-
tions benefit from this arrangement,including multiple C-
H···O contacts[20] (C···O 3.4–3.6 Š) at the interface of the
two monomers (e.g. C(10)···O(3)’,O(4)···C(8) ’). The cores
further organize into layers (Figure 3d,view along the crys-
tallographic a axis shown),the result of gentle interactions
between the isopropyl side chains at the discernably lipo-
philic interface. The layer-to-layer distance in this packing
motif is ꢁ11.5 Š.
Scheme 2. Synthesis of ester-functionalized 1-aza-adamantanetrione 8.
Using methodology outlined previously for the explora-
tion of diastereoselective additions to crowded phlorogluci-
nol derivatives,[16] tripentyl AAT 9 is prepared starting from
versatile trialdehyde 19.[17] Three-fold addition of n-butyl-
lithium produces triol 20 as a mixture of anti,syn and
syn,syn isomers (Scheme 3); the mixture is then subjected to
trimethylsilyl hydride reduction as described for similar
compounds by Biali and co-workers to give 21.[18] Standard
1454
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2008, 14,1452 – 1463

Donor–Acceptor Interactions in b-Aminoketones
FULL PAPER
Figure 3. X-ray crystal structure of 6. a) Monomer structure; solvent mol-
ecules and hydrogen atoms have been removed for clarity. b) ORTEP
plot where thermal ellipsoids are shown at the 50% probability level. c)
Antiparallel alignment of 6 within layers that features multiple C-H···O
interactions at the “dimer” interface. d) A packing view (along the a
axis) showing interaction of the isopropyl side chains at a lipophilic inter-
face. Atom color codes: C=gray,N =blue,O =red.
Figure 4. X-ray crystal structure of 7. a) Monomer structure; solvent mol-
ecules and hydrogen atoms have been removed for clarity. b) ORTEP
plot where thermal ellipsoids are shown at the 50% probability level. c)
Intermolecular hydrogen bonding (indicated by the dashed line) between
the exo alcohol and neighboring ester carbonyl (O(2)···O(13)=2.72 Š).
d) A packing view (along the a axis) showing formation of layers,domi-
nated by side chain interactions. Atom color codes: C=gray,N =blue,
O=red.
Alcohol 7 could be recrystallized from a concentrated
methylene chloride solution and the refined monomer struc-
ture is shown in Figure 4a and b. The asymmetric unit con-
sists of two molecules of 7 (Figure 4c),and three of the six
isopropyl side chains are disordered (disorder not shown).
The exo alcohol participates in a short hydrogen bond with
a neighboring ester carbonyl group (O(2)···O(13)=2.72 Š;
O(2)-H···O(13)=166.78) in this arrangement. Layers ob-
served by looking along the a axis are again dominated by
interactions involving the isopropyl side chains (Figure 4d).
The ester side chain conformations displayed by the mon-
omers 6 and 7 in the solid state (Figures 3a and 4a),while
certainly influenced by packing of the isopropyl groups,are
ics could play a role in stabilizing this conformation. Re-
gardless of the origin,the result for both 6 and 7 is that
three of the ester oxygens are positioned underneath the
core with an average O···O distance of ꢁ4.1 Š; whether this
space can serve as a binding environment for ions (akin to
valinomycin,[22] where O···O ꢁ4.2 Š for the ester carbonyl
oxygens) is currently being evaluated.
Comparative analysis of bond lengths and angles from X-
ray structural data: Relevant bond lengths and angles for
the monomer structures of 6 and 7 are provided in Table 1
(additional crystallographic data is provided in the Experi-
mental Section); for 7,data for only one of the molecules in
the asymmetric unit is provided and discussed since the
values for the two are similar. The comparable monomer
ꢀ
ꢀ
remarkably similar. The ester C(10) C(11),C(15) C(16),
ꢀ
ꢀ
and C(20) C(21) bonds are antiperiplanar to the Ca Cb
ꢀ
bonds (e.g. C(6) C(8)),and although of “normal” length
(average for 6: 1.497(5) Š; 7: 1.496(4) Š),[21] stereoelectron-
Chem. Eur. J. 2008, 14,1452 – 1463
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
1455

R. K. Castellano et al.
Table 1. Selected bond lengths [Š] and angles [8] for 6 and 7.[a]
Bond length
6
7
Angle
6
7
ꢀ
N(1) C(1)
1.461(4) 1.442(4) C(1)-N(1)-C(8) 110.4(3) 110.7(2)
1.433(4) 1.463(4) C(1)-N(1)-C(9) 110.6(3) 110.1(2)
1.451(4) 1.456(4) C(8)-N(1)-C(9) 111.3(3) 109.3(2)
ꢀ
N(1) C(8)
ꢀ
N(1) C(9)
ꢀ
C(1) C(2)
1.591(5) 1.591(4) N(1)-C(1)-C(2) 111.4(3) 112.3(2)
1.593(5) 1.560(4) N(1)-C(8)-C(6) 111.5(2) 111.7(2)
1.598(5) 1.565(3) N(1)-C(9)-C(4) 111.4(3) 112.5(2)
ꢀ
C(4) C(9)
ꢀ
C(6) C(8)
ꢀ
C(2) C(3)
1.506(5) 1.520(4) C(1)-C(2)-C(3) 103.8(3) 104.0(2)
1.516(5) 1.505(4) C(1)-C(2)-C(7) 103.3(2) 102.3(2)
1.515(5) 1.543(4) C(5)-C(6)-C(8) 103.9(3) 107.8(2)
1.514(5) 1.540(4) C(7)-C(6)-C(8) 103.4(2) 103.5(2)
1.509(5) 1.508(4) C(3)-C(4)-C(9) 103.7(3) 105.0(2)
1.513(5) 1.517(4) C(5)-C(4)-C(9) 103.5(3) 108.0(2)
Figure 5. Atom labeling scheme for the bond length data presented in
Table 2. R=CH2CO2iPr,R 1–R4 =various substituents,Ph =phenyl,Ar =
p-nitrophenyl (24) or p-chlorophenyl (25), AD=various adamantanes.
Reference (CSD code): [a] ref. [24] (KOLSIN); [b] ref. [25] (OCAYEW);
[c] ref. [26] (EJIQUJ); [d] ref. [28] (DESVIG); [e] ref. [29] (LIXFIH); [f]
ref. [29] (LIXFUT); [g] ref. [30] (FITPED).
ꢀ
C(3) C(4)
ꢀ
C(4) C(5)
ꢀ
C(5) C(6)
ꢀ
C(6) C(7)
ꢀ
C(2) C(7)
ꢀ
O(1) C(3)
1.215(4) 1.213(3) C(3)-C(4)-C(5) 111.9(3) 109.5(2)
1.214(4) 1.426(3) C(5)-C(6)-C(7) 112.4(2) 111.7(2)
1.224(4) 1.214(3) C(7)-C(2)-C(3) 112.1(3) 112.2(2)
C(4)-C(5)-C(6) 114.1(3) 110.3(2)
ꢀ
O(2) C(5)
Table 2. Comparison of average bond lengths[a] [Š] from the X-ray crys-
ꢀ
O(3) C(7)
tal structures of 6, 7,and related tricyclic molecules from Figure 5.
C(6)-C(7)-C(2) 114.4(3) 113.9(2)
C(2)-C(3)-C(4) 114.8(3) 114.2(2)
Bond[b]
6[c]
7[d]
23[e] 24[f] 25[g] 26[h], 27[i], 28[j] 29[k] AD[l]
a
1.45 1.44 1.46
1.59 1.59 1.59
1.51 1.52 1.53
–
–
–
–
–
–
–
–
–
–
–
1.55
1.51
1.22
1.50
1.54
–
1.54
–
–
–
–
–
–
–
–
–
–
–
–
1.54
–
–
b, b’
c, c’
d, d’
e, e’
f, f’
g
h, h’
i, i’
j, j’
k
C(2)-C(3)-O(1) 122.8(3) 122.1(3)
C(4)-C(3)-O(1) 121.5(3) 123.4(3)
C(4)-C(5)-O(2) 122.9(3) 108.4(2)
C(6)-C(5)-O(2) 122.1(3) 108.5(2)
C(6)-C(7)-O(3) 122.8(3) 123.1(2)
C(2)-C(7)-O(3) 122.0(3) 122.1(2)
1.22 1.21 1.21 1.21
1.21
1.51
1.54
–
1.54
1.53
1.53
–
–
–
–
–
–
–
–
1.51 1.50 1.51
1.56 1.54 1.55
1.46 1.46 1.48
1.54 1.53 1.53 1.53
[a] Standard deviations are shown in parentheses. Additional crystallo-
graphic details are provided in the Experimental Section.
–
–
–
–
–
–
1.54 1.53
1.53 1.53
1.47 1.49
–
–
[a] The bond lengths for chemically equivalent bonds within each mole-
cule have been averaged. [b] As specified in Figure 5. Crystallographic
details (see the Experimental Section for further details of 6 and 7) in-
cluding CSD codes and references: [c] 6: T=173(2) K, R1 =0.0703. [d] 7:
T=173(2) K, R1 =0.0587. [e] 23 (KOLSIN,Ref. [24]): T=295 K, R1 =
conformations of 6 and 7 (Figure 3a and 4a) indeed encour-
ages this side-by-side comparison. Most striking are the long
ꢀ
ꢀ
ꢀ
ꢀ
Ca Cb bonds (i.e.,C(1) C(2),C(4) C(9),and C(6) C(8)) of
6,up to 1.6 Š,that we take a priori as a signature of hyper-
conjugative interactions. For 7,equivalent elongation is ob-
0.0470. [f] 24 (OCAYEW,ref. [25]):
T=293 K, R1 =0.0580. [g] 25
(EJIQUJ,ref. [26]): T=291 K, R1 =0.0456. [h] 26 (DESVIG,ref. [28]):
T=295 K, R1 =0.0490. [i] 27 (LIXFIH,ref. [29]): T=295 K, R1 =0.0290.
[j] 28 (LIXFUT,ref. [29]): T=295 K, R1 =0.0470. [k] 29 (FITPED,ref.
[30]): T=295 K, R1 =0.0300. [l] AD: various adamantanes from the CSD.
ꢀ
served only for the C(1) C(2) bond,the central bond that is
flanked by two carbonyl acceptors,highlighting the unique-
ness of this arrangement of atoms. Correspondingly,the
ꢀ
carbon carbon bonds for 7 that are not within a donor–s-
ꢀ
ꢀ
acceptor framework (e.g.,C(4) C(5),C(5) C(6)) appear to
be of standard length. Further discussion of bond lengths is
best done in the context of similarly strained cyclic mole-
cules (below); indeed,the bond angles within 6 and 7 devi-
ate appreciably from the optimal values (i.e., a C-C-C ¼
109.58).
Again,the comparison of molecules 6 and 7 to simple
structural analogues like 1 or 2 is tempting but would ignore
any structural changes associated with ring strain. We have
chosen instead to discuss and analyze bond lengths (Figure 5
and Table 2) and angles (Figure 6 and Table 3) in the con-
text of related aza- and deaza-adamantane crystal structures
Figure 6. Atom labeling scheme for the bond angle data presented in
Table 3. R=CH2CO2iPr,R 1-R4 =various substituents,Ph =phenyl,Ar =
p-nitrophenyl. Reference (CSD code): [a] ref. [24] (KOLSIN); [b] ref.
[25] (OCAYEW); [c] ref. [28] (DESVIG); [d] ref. [29] (LIXFIH); [e] ref.
[29] (LIXFUT); [f] ref. [30] (FITPED).
(with crystallographic
Table 2 is the average bond length data for 6 and 7,along
R
factors <0.05).[23] Included in
side data for analogous structures available in the literature
(the di- (CSD code KOLSIN[24]) and mono- (OCAYEW[25]
)
ketones 23 and 24,respectively) and a saturated aza-ada-
LIXFIH (27),[29] and LIXFUT (28)[29]),a selected adamanta-
none (FITPED (29)[30]),and variously substituted adaman-
tanes (AD). A simplified bond labeling scheme (Figure 5)
mantane 25 (EJIQUJ[26]).[27] Also shown are average bond
lengths for three adamantanediones (DESVIG (26),[28]
1456
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2008, 14,1452 – 1463

Donor–Acceptor Interactions in b-Aminoketones
FULL PAPER
Table 3. Comparison of average bond angles[a] [8] from the X-ray crystal
structures of 6, 7,and related tricyclic molecules from Figure 6.
is analogous to the “enhanced” through-bond effects noted
for the diaza-adamantanones like 3b (that feature two
donor nitrogens communicating with one acceptor carbon-
yl).[5d]
Comparison of selected bond angles of 6 and 7 to those of
structural analogues is provided in Figure 6 and Table 3; an
arbitrary atom labeling scheme has been adopted for the
analysis (Figure 6),which should not be confused with the
atom labels reported in the crystallographic data files. There
are no significant trends that can be identified for the angles
that describe the cyclohexane ring that includes atoms 2–5
[b]
Angle[c]
6
7
23[d]
24[e]
26[f]
27[g], 28[h]
29[i]
1–2–3
1–2–4
4–5–6
3–2–4
2–4–5
ꢀCNC
122.3
123.4
122.1
122.1
114.2
112.2
330.1
123.5
121.8
121.4
114.5
105.8
330.2
124.1
123.2
–
112.7
108.9
329.4
124.9
121.2
121.4
113.5
105.8
–
124.1
122.7
123.1
112.9
110.5
–
123.7
123.9
–
112.3
109.5
–
114.4
112.1
332.3
[a] The bond angles for chemically equivalent bonds within each mole-
cule have been averaged. [b] Standard deviations are shown in parenthe-
ses. [c] As specified in Figure 6. CSD codes and references: [d] KOLSIN,
ref. [24]. [e] OCAYEW,ref. [25]. [f] DESVIG,ref. [28]. [g] LIXFIH,ref.
[29]. [h] LIXFUT,ref. [29]. [i] FITPED,ref. [30].
ꢀ
across the series. The C C=O angles (e.g. angle 1–2–3) fall
ꢀ
ꢀ
within 122–1258,and angle 3–2–4 (C C(=O) C) is only
slightly larger (ꢁ18) for 6, 7,and 23 than for the deaza-ada-
mantanones (but certainly larger than 25 and AD,where it
is 109–1108). The greatest variation is found at the bridge-
head carbon angle, 2–4–5,although the angle for 6 and 7 is
only 1.78 larger than what is found for deaza-analogues 27
and 28. Based on the sum of the bond angles at nitrogen in
the aza-adamantane derivatives,the nitrogen atom of struc-
tures 6 and 7 is slightly flattened compared with that of
structures 23, 24,and 25 (ꢀC-N-C =3288 for 25); this distor-
has been adopted to facilitate the comparison; the letters
are kept consistent between structures based on the atom
bonding sequence. For example,path a-b-c-d uniquely de-
scribes the donor–s-acceptor pathway where central bond b
is flanked by two carbonyl acceptors.
ꢀ
The analysis reveals somewhat shortened C N bonds a
ꢀ
ꢀ
(1.45 Š) versus k (1.48 Š),and a considerably elongated C
tion further aligns the nitrogen lone pair with the Ca Cb
C bond b (1.59 Š) versus f (1.55 Š), j (1.53 Š), f’ (1.54 Š),
or j’ (1.53 Š). There is therefore some hint of the bond
length alternation that is a classically predicted consequence
of hyperconjugation,[2c,31] visualized,if even crudely,by the
no-bond resonance structure shown below:
central bond.[19] Without comparing every bond angle for
each structure throughout the series,our conclusion is that
while 6 and 7 (and 23) are certainly strained molecules,it is
not obvious that their respective bond length changes (in
ꢀ
particular elongation of the Ca Cb bonds) are solely due to
this effect. High-level computational studies[33] (e.g. NBO
analysis) could shed additional light on the distortions,and
these are currently in progress.[34]
Further identified from close inspection of the crystal
structures of 6 and 7 is pyramidalization of the carbonyl
carbon atoms (labeled 2 in Figure 6).[1c] Similar distortion
has been noted by Verhoeven and co-workers in 1,1-dicya-
The bond alternation trend does not extend to bonds c
(1.52 Š versus e, c’,and e’ that are all 1.51 Š) or d (including
d’,all C =O bonds are 1.21–1.22 Š). That bond d is not sub-
stantially perturbed may speak to compensatory stereoelec-
tronic effects in these systems.[32] Comparison of bonds b’, f’,
h’,and j’ (1.53–1.55 Š) reveals only a modest sensitivity to
inherent structural changes (and strain) within the deaza-
adamantanes; this further implicates donor–acceptor effects
in the elongation of bond b.
The bond elongation found for b (ꢁ0.05 Š) is the largest
reported for b-aminoketones; in work by Verhoeven and co-
workers with (admittedly less strained) piperidone and tro-
panone derivatives (bearing carbonyl groups converted to
1,1-dicyanovinyl functions, better electron acceptors), the
novinyl-modified
piperidone
and tropanone derivatives.[19,33a]
Figure 7 shows one way to
report this effect;[33a] namely,as
the angle (q) between the car-
bonyl vector and the mean
plane defined by the three core
carbons (2, 3,and 4). For 6, q=
8.78 (the average of three car-
bonyl groups),and for 7 q=
7.38 (average of four carbonyl
groups,two from each molecule
in the asymmetric unit). The
Figure 7. Angle q,which quan-
tifies the pyramidalization of
the carbonyl carbon atoms of 6
and 7,calculated as the angle
between the carbonyl bond
and the mean plane defined by
carbon atoms 2,3,and 4.
ꢀ
Ca Cb bond elongations are ꢁ0.02 Š and taken as fairly de-
largest value reported from Verhoevenꢁs work (recalculated
here in the fashion shown in Figure 7) is 6.18. While the pyr-
amidalization serves to reduce ring strain,that it occurs in
such a way (+q rather than ꢀq) to further position the car-
finitive evidence (along with other data) for through-bond
interactions. Even so,it seems unnecessary,and likely inap-
propriate,[31f] to equate the magnitude of the bond length
changes to the “strength” of the through-bond interactions
in the molecules; suffice it to say,the consequences of
donor–acceptor interactions in these aza-adamantanones are
detectable. That pathway a-b-c-d emerges as unique (one
donor nitrogen communicating with two acceptor carbonyls)
ꢀ
bonyl p bond parallel to the Ca Cb bonds presumably also
optimizes interaction with the nitrogen donor. For 26,that
lacks donor–s-acceptor interactions, q=ꢀ6.78 for one of the
two carbonyl groups (the other features q ꢁ08). Finally,
while the sensitivity of the pyramidalization to packing and
Chem. Eur. J. 2008, 14,1452 – 1463
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
1457

R. K. Castellano et al.
temperature is not known,it is not reproduced with the
same magnitude in the crystal structures of the other mole-
cules shown in Figure 6.
Complementary spectroscopic data: Techniques other than
X-ray crystallography have been employed to explore the
consequences of donor–acceptor interactions in the rigid
aza-adamantanones. Highlighted here is data obtained from
UV,IR,and
13C NMR measurements for 6–9,with some
comparative data provided from the literature.
UV spectroscopy is the most routinely used method to di-
agnose through-bond interactions in b-aminoketones where
a new absorption band appears (the lower energy compo-
nent of a split carbonyl p–p* transition,the so-called s-cou-
pled transition) to signify communication between the
donor and acceptor.[4,5] The absorption data for 6–9 for
three solvents is provided in Table 4. A new absorption
Figure 8. Treatment of a 1.2”10ꢀ5 m solution of 9 (in acetonitrile) with tri-
fluoroacetic acid (TFA) results in disappearance of the s-coupled transi-
tion.
Table 4. UV absorption data for 6–9 in different solvents.[a]
Aza-
ada-
mantane
lmax
e
lmax
e
lmax
e
although this is often done. Additional complication enters
when considering IR data for the aza- and deaza-adamanta-
nones bearing two or more ketones in a 1,3-relationship that
display Fermi-type coupling;[38] we have used the average
carbonyl stretch for making a quick comparison (the com-
plete comparative data,including the data for 6–9,is provid-
ed in the Supporting Information). Across representative
adamantanones (two monoketones,[10a,11a,39] two diones,[40,41]
and one trione[42]) identified from the literature,the carbon-
yl stretching frequency is relatively invariant (average value
ꢁ1720 cmꢀ1). A similar consistency is observed for a series
of representative aza-adamantanones (including 6–
9),[10a,11a,43] although the analysis reveals a ꢁ5–20 cmꢀ1 shift
to lower wavenumber for these donor–s-acceptor molecules.
A similar shift has been identified for the 1,3-diaza-adaman-
tanones (e.g. 3b).[5d,11a] In some previously reported cases
the carbonyl stretch has been restored to higher frequency
upon protonation or methylation at nitrogen,[5d,10a] although
these experiments have not generally considered structural
changes associated with the chemical transformations.[9e,44]
The on-average shift to lower energy is consistent with pre-
dictions based on through-bond donor–acceptor interactions
that would weaken the carbonyl p bond (Figure 1).
[nm]
[mꢀ1 cmꢀ1
]
[nm]
[mꢀ1 cmꢀ1
]
[nm]
[mꢀ1 cmꢀ1
]
cyclohexane
acetonitrile
ethanol
6
7
8
9
270
–
–
2600
–
–
263
265
267
273
3230
2040
3270
2860
265
263
267
275
3010
1910
2980
3270
271
3060
[a] Data was collected from 40–120 mm. The absorption intensity was
shown to vary linearly with concentration in all cases.
maximum (lmax) is indeed observed in the 260–275 nm
window,consistent with (although somewhat red-shifted
from) earlier results.[7a,35,36] The frequency deviates little nei-
ther with solvent polarity—not surprising given the nominal
charge-transfer character of these donor–s-acceptor mole-
cules[37]—nor peripheral substitution. The molar extinction
coefficients,calculated from plots of absorbance versus con-
centration,are ꢁ3000mꢀ1 cmꢀ1 for the triketones 6, 8,and 9;
this value decreases commensurately (i.e.,by about one
third) for diketone 7. An identical trend has been reported
in related aza-adamantanones[35] and speaks to the participa-
tion of all three acceptor carbonyls to the donor–acceptor
system. The lmax values do not deviate significantly between
the di- and triketones in this work,although a survey of the
literature shows that simple (even rigid) b-aminoketones
offer values ꢁ235–245 nm; this increases for the diaza-ada-
mantanones.[4a,5d] Interestingly,the s-coupled transition can
be reversibly abolished[4a] upon protonation by trifluoroace-
tic acid in acetonitrile (Figure 8; see the Supporting Infor-
mation for additional details).
A similar treatment of the 13C NMR data for adamanta-
nones and aza-adamantanones reveals,for both series,that
the carbonyl resonance shifts upfield as the number of car-
bonyl groups increases (the complete comparative data,in-
cluding that of 6–9,is provided in the Supporting Informa-
tion). Such a subtle effect does not emerge from the X-ray
crystallography data. For example, dC
shifts from
=
O
IR and 13C NMR spectroscopic data have been used spor-
adically to report on donor–s-acceptor interactions and in
our estimation interpretation of these data warrants some
caution. Notwithstanding the obvious complications of sol-
vation (and experimental conditions in general),differences
identified from single-point comparisons (e.g. 1 versus 3)[5d]
are difficult to attribute to any one particular phenomenon,
ꢁ218 ppm for adamantanones (in CDCl3),[5e,39,45] to
ꢁ208 ppm for adamantanediones,[46] to ꢁ202 ppm for ada-
mantanetriones.[42] The aza-adamantanones (including
diones and triones) are on-average upfield shifted (ꢁ3–
5 ppm) from these values. From monoketone,to dione,to
trione the values are d ꢁ214 ppm,[5e,11a] 202 ppm (for 7),[47]
and 197–200 ppm,respectively. This trend is again consistent
1458
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2008, 14,1452 – 1463

Donor–Acceptor Interactions in b-Aminoketones
FULL PAPER
with the hyperconjugative interactions depicted in Figure 1,
but such an analysis remains admittedly oversimplified.
bined yield (and ꢁ1:1:1 molar ratio); Scheme 4) that do not
arise from a simple reversal of the Mannich-type cyclization
reaction that generates the starting material (Scheme 1). All
of the structures were determined by 1D and 2D (gHMBC
Consequences of through-bond interactions on b-aminoke-
tone reactivity—General considerations: When molecules
that feature well-characterized donor–acceptor interactions
through s bonds show aberrant chemical behavior the chal-
lenge lies in linking the two phenomena. For b-aminoke-
tones,diminished reactivity at nitrogen (basicity/nucleophi-
licity) is generally taken as a hallmark consequence.[5d,e,h,35,48]
Risch and co-workers have shown that the pKa of the pro-
tonated amine of aza-adamantanes decreases by ꢁthree
units upon introduction of each carbonyl acceptor (from
ꢁ10.7 for aza-adamantane to ꢁ0.3 for aza-adamantane-
triones).[35] Likewise,Sasaki and co-workers have noted that
the pKa1 of protonated diaza-adamantanones like 3b (ꢁ4–
5) is significantly lower than that of aza-adamantanone
(8.6[5e]) (or aza-adamantane).[5d] Related studies have shown
a diminished nucleophilicity at nitrogen,and some have lik-
ened its general reactivity to that of an amide. We found
previously[7a] that treatment of 4 with methyl iodide,hydro-
gen peroxide,or m-CPBA gives no reaction at the bridge-
head amine (or elsewhere).[49] Through-bond (hyperconjuga-
1
and gDQCOSY) H and 13C NMR analysis using a 500 MHz
instrument and confirmed by mass spectrometry. The prod-
ucts include N-alkyl-cis-3,5-disubstituted piperidone[56] 30
(as a mixture of epimers),isolated in a ꢁ1:1 molar ratio
(determined by integration) with a,b-unsaturated acid 31
after column chromatography (the mixture appears as the
higher Rf spot with CH2Cl2/MeOH 20:1),and cis-3,5-disub-
stituted piperidone 32 (isolated as the lower Rf spot). To test
whether placement of the esters or their presence matters to
the products formed (and mechanism), 8 and 9 were subject-
ed to similar conditions. N-Alkyl-cis-3,5-disubstituted piperi-
dones 33 (69% yield,again as a mixture of epimers) and 34
(ꢁ75% yield) could be isolated from the reaction of 8 and
9,respectively. In both cases a small amount ( < 5%) of the
alkenes (35 and 36) were also identified by NMR. The reac-
tions appear to share a common mechanism that is specific
to the AAT core.
tive) interactions likely play some role in this sluggish reac-
[50]
tivity,but general inductive effects
in b-aminoketone-
s[48a,51] are also significant; that N-methylpiperidine is a
much stronger base (pKb 3.92) than N-methylpiperidone
(pKb 6.01)[5e,52] despite “undetectable”[4a,5g] through-bond in-
teractions (e.g. no s-coupled transition) is evidentiary.
Discussed earlier,through-bond influences on carbonyl
reactivity in b-aminoketones have historically been probed
through subtle changes in nucleophile addition/hydride re-
duction stereoselectivity. Molecules 6–9 are complicated
substrates for stereochemical analysis given that their car-
bonyl groups are sterically biased and that some (e.g. 6–8)
bear functional groups capable of metal chelation. Likewise,
our own studies[7a] and those of others[5d,43a] note slow or no
reaction with conventional carbonyl nucleophiles (e.g. hy-
drazines,hydroxylamine,diazomethane,alkyl phosphonium
ylides,etc.),although these results are difficult to assess (or
quantify) in terms of through-bond interactions.
Scheme 4. Acid-mediated fragmentation reactions of aza-adamantane-
triones 6, 8,and 9.
Scheme 5 shows a proposed mechanism to rationalize the
fragmentation products. We favor a first productive step of
carbonyl protonation followed by hydration to form A, a
process already discussed for aza-adamantanetriones under
aqueous acidic conditions.[35] The resulting hydrate can then
collapse to the diketo-acid intermediate B through a retro-
Claisen-like process. A similar mechanistic sequence could
then repeat; carbonyl protonation and hydration of B,fol-
lowed by fragmentation of the core and thermally-mediated
decarboxylation to afford the N-alkyl-cis-3,5-disubstituted
piperidones (30, 33,and 34). The increased nitrogen basicity
of B (relative to the starting material)[35] makes protonation
here plausible; this initiates an alternative hydration/frag-
mentation/decarboxylation pathway that could account for
the formation of piperidone 32 and a,b-unsaturated acid 31
(and 35/36 for 8/9). That the alkene 31 does not come di-
rectly from 30 has been proven by subjecting pure 30 to the
Fragmentation reactions: The orbital arrangement that un-
derlies donor–s-acceptor interactions in b-aminoketones
(Figure 1) is also central to the kinetics and stereoselectivity
of Grob fragmentation reactions.[1c,d,53] This connection
makes fragmentation studies of the molecules presented
here,for which spectroscopic evidence of the intimate rela-
tionship between the nitrogen and carbonyl groups has been
presented,particularly fitting. Risch and co-workers have el-
egantly described heterolytic fragmentations of chloro-sub-
stituted aza-adamantanones[54] and aza-adamantanediones[55]
under basic conditions; an initial report of related reactions
of aza-adamantanetriones (i.e. 6, 8,and 9) under acidic con-
ditions is presented below.
Upon treating 6 with dilute acetic acid (HOAc) at 1008C
for 1 h,three distinct products are formed (in ꢁ98% com-
Chem. Eur. J. 2008, 14,1452 – 1463
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
1459

R. K. Castellano et al.
Experimental Section
Materials and general methods: Reagents and solvents were purchased
from commercial sources and used without further purification unless
otherwise specified. THF,ether,CH 2Cl2,and DMF were degassed in 20 L
drums and passed through two sequential purification columns (activated
alumina; molecular sieves for DMF) under a positive argon atmosphere.
Thin layer chromatography (TLC) was performed on SiO2–60 F254 alumi-
num plates with visualization by UV light or staining. Flash column chro-
matography was performed using Purasil SiO2–60,230–400 mesh from
Whatman. Melting points (m.p.) were determined on a Mel-temp electro-
thermal melting point apparatus and are uncorrected. 300 (75) MHz and
1
500 (125) MHz H (13C) NMR spectra were recorded on Varian Mercury
300,Gemini 300,VXR 300,and Varian Inova (500) spectrometers.
Chemical shifts (d) are given in parts per million (ppm) relative to TMS
and referenced to residual protonated solvent (CDCl3: dH 7.27 ppm, dC
77.00 ppm; [D6]DMSO: dH 2.50 ppm, dC 39.50 ppm). Abbreviations used
are s (singlet),d (doublet),t (triplet),q (quartet),quin (quintet),hp
(heptet),b (broad),and m (multiplet). UV/Vis absorption spectra were
obtained using a Cary 100 Bio spectrophotometer and 1 cm quartz cells.
ESI- and ESI-TOF-MS spectra were recorded on a Bruker APEX II
FTICR and Agilent 6210 TOF spectrometer,respectively. EI-,CI-,and
DIP-CI-MS spectra were recorded on a Thermo Trace GC DSQ (single
quadrupole) spectrometer. The syntheses of 8, 9, 13–22,and 30–35 as
Scheme 5. Proposed mechanistic pathway for the acid-mediated fragmen-
tation of 6, 8,and 9.
1
reaction conditions and monitoring its stability by H NMR
(over four hours). Future work could consider this method-
ology as a general way to prepare highly substituted cis-3,5-
disubstituted piperidones,important building blocks in the
synthesis of alkaloids and analgesics.[57]
1
well as H and 13C NMR spectra for all new compounds are presented in
the Supporting Information.
X-ray crystal structure determination and refinement: Data were collect-
ed at 173 K on a Siemens SMART PLATFORM equipped with a CCD
area detector and a graphite monochromator utilizing MoKa radiation
(l=0.71073 Š). Cell parameters were refined using up to 8192 reflec-
tions. A full sphere of data (1850 frames) was collected using the w-scan
method (0.38 frame width). The first 50 frames were re-measured at the
end of data collection to monitor instrument and crystal stability (maxi-
mum correction on I was < 1%). Absorption corrections by integration
were applied based on measured indexed crystal faces. The structure was
solved by the Direct Methods in SHELXTL6,[58] and refined using full-
matrix least squares on F2. The non-H atoms were treated anisotropically,
whereas the hydrogen atoms were calculated in ideal positions and were
riding on their respective carbon atoms. Details are provided in Table 5.
For 6,the asymmetric unit consists of the molecule and two half ethanol
Conclusion
A combination of experimental techniques and comparative
analysis has shown that the through-bond donor–acceptor
interactions in appropriately functionalized b-aminoketones
can be particularly assessable. Uniquely for the 1-aza-ada-
mantanones,molecules in which the donor and acceptor ori-
entation is permanently optimized,many of the hallmark
consequences of through-bond communication emerge.
Table 5. Crystal data and structure refinement for 6 and 7.
ꢀ
These include elongation of the central C C bond in the
6
7
donor–s-acceptor pathway (to ꢁ1.6 Š),the presence of a
new absorption band in the molecules’ UV/Vis spectra,and
IR/13C NMR spectroscopic shift trends (versus similarly
strained molecules that lack either the donor or acceptor
groups) consistent with theoretical expectations. The ar-
rangement of donor and acceptor groups has functional con-
sequences as well in terms of the molecules’ reactivity pro-
files.
We previously showed that molecules like 4 and 5 self-as-
semble in solution and that the emergent macromolecular
properties respond to molecular-level changes that also in-
fluence the donor–s-acceptor interactions. The current mo-
lecular-level analysis that finds pronounced structural,spec-
troscopic,and reactivity consequences for traditionally weak
effects suggests that they have been largely overlooked on
the supramolecular scale where similarly modest interac-
tions are often amplified. Left now is to introduce these
donor–s-acceptor arrangements into macromolecules,where
even the simplest motifs could offer ways to tune bulk prop-
erties not available to more traditional weak interactions.
empirical formula
formula weight
crystal system
space group
a [Š]
b [Š]
c [Š]
a [8]
b [8]
C26H39NO10
525.58
C24H35NO9
481.53
triclinic
triclinic
¯
¯
P1
P1
11.4801(16)
11.6917(16)
13.3142(19)
104.663(3)
97.753(2)
116.825(2)
1477.2(4)
2
12.7729(16)
14.7468(18)
14.8080(18)
75.036(2)
73.265(2)
84.410(2)
2579.7(5)
4
g [8]
V [Š3]
Z
1calcd [gcmꢀ3
]
1.182
0.18”0.17”0.05
3871
1.240
0.26”0.19”0.11
9042
crystal size [mm3]
independent reflections
observed reflections [I>2s(I)]
index ranges
2736
5600
ꢀ12ꢃhꢃ12
ꢀ12ꢃkꢃ10
ꢀ14ꢃlꢃ14
314
564
1.051
ꢀ15ꢃhꢃ9
ꢀ17ꢃkꢃ17
ꢀ17ꢃlꢃ16
628
1032
1.034
parameters
F
A
goodness-of-fit (on F2)
R1 based on F [I>2s(I)]
wR2 based on F [I>2s(I)]
0.0703
0.2079
0.0587
0.1325
1460
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2008, 14,1452 – 1463

Donor–Acceptor Interactions in b-Aminoketones
FULL PAPER
molecules; each disordered around an inversion center. The latter were
disordered and could not be modeled properly,thus the program
SQUEEZE,[59] a part of the PLATON package[60] of crystallographic soft-
ware,was used to calculate the solvent disorder area and remove its con-
tribution to the overall intensity data.
was purified by flash chromatography (CH2Cl2/MeOH 20:1) to afford a
mixture of 30 (0.115 g,47%) and 31 (0.046 g,51%) as a colorless oil,and
32 (0.081 g,52%) as a colorless solid. 1H NMR (500 MHz,CDCl 3): d =
1.16 (m,18H),2.12 (dd, 3J=16.9,6.1 Hz,1H),2.13 (dd, 3J=17.0,6.2 Hz,
1H),2.40 (t, 3J=11.8 Hz,1H),2.44 (dd, 3J=16.8,6.2 Hz,1H),2.48 (t,
3J=11.7 Hz,1H),2.60 (m,1H),2.64 (m,1H),2.70 (dd,
1H),2.74 (dd, 3J=17.3 Hz,7.0 Hz,1H),2.90 (dd, 3J=12.1,10.2 Hz,1H),
3.02 (dq, 3J=9.8 Hz,6.5 Hz,1H),3.15 (dq, 3J=11.2,5.8 Hz,1H),3.18
3J=11.3,6.1 Hz,2.3 Hz,1H),3.43
3J=13.2,6.0 Hz,
CCDC 656392 (6) and 656393 (7) contain the supplementary crystallo-
graphic data for this paper. These data can be obtained free of charge
from The Cambridge Crystallographic Data Centre via www.ccdc.cam.
ac.uk/data_request/cif.
(dq, 3J=11.9,5.8 Hz,1H),3.31 (ddd,
(ddd, 3J=11.6,6.0,2.7 Hz,1H),4.93 (m,3H),10.63 ppm (s,1H);
13C NMR (125 MHz,CDCl 3): d = 22.0,32.0,34.1,38.6,44.8,45.0,57.3,
Isopropyl
2,2’,2’’-(2,4,6-trimethoxybenzene-1,3,5-triyl)triacetate
(11):
pTsOH·H2O (1.13 g,5.92 mmol) was added to a stirred solution of triacid
10[7b] (0.900 g,2.63 mmol) in isopropanol (25 mL) and the reaction was
heated under reflux overnight. All volatiles were then removed under re-
duced pressure and the residue was taken up with EtOAc and washed
with dilute NaOH. The organic solution was then dried over MgSO4 and
concentrated under reduced pressure and the residue was purified using
column chromatography (hexanes/EtOAc 3:1) to afford 11 (1.13 g,92%)
as a colorless oil. 1H NMR (300 MHz,CDCl 3): d=1.23 (d, 3J=6.3 Hz,
57.8,58.8,68.5,170.5,171.5,176.7,206.4 ppm; MS (ESI):
m/z: calcd for
C23H38NO9: 472.2541; found 472.2525 [M+H]+.
4-Isopropoxy-2-methylene-4-oxobutanoic acid (31): 1H NMR (500 MHz,
CDCl3): d = 1.16 (m,6H),3.23 (s,2H),4.93 (m,1H),5.72 (s,1H),6.34
(s,1H),10.63 ppm (s,1H); 13C NMR (125 MHz,CDCl 3): d = 22.0,37.9,
68.5,130.2,134.2,170.5,171.1 ppm; MS (ESI):
m/z: calcd for C8H13O4:
173.0808; found 173.0801 [M+H]+.
syn-Isopropyl 2,2’-(4-oxopiperidine-3,5-diyl)diacetate (32): 1H NMR
18H),3.63 (s,6H),3.72 (s,9H),5.05 ppm (m,
3J=6.3 Hz,3H); 13C NMR
(500 MHz,CDCl 3): d = 1.16 (m,12H),2.08 (dd, 3J=17.1,6.8 Hz,2H),
(75 MHz,CDCl 3): d=21.7,30.8,61.2,68.0,118.6,157.7,171.5 ppm; MS
2.62 (t, 3J=12.2 Hz,2H),2.66 (dd, 3J=16.9,6.1 Hz,2H),3.01 (dq,
3J=
3J=5.9 Hz,
13C NMR (125 MHz,CDCl 3): d = 22.0,31.9,
(EI): m/z: calcd for C24H36O9: 468.2359; found 468.2355 [M]+.
12.3,6.1 Hz,2H),3.48 (dd,
2H),5.88 ppm (s,1H);
3J=12.6,6.0 Hz,2H),4.92 (hp,
Isopropyl 2,2’,2’’-(2,4,6-trihydroxybenzene-1,3,5-triyl)triacetate (12): BBr3
(1.26 mL,13.3 mmol) was added at ꢀ788C to a stirred solution of 11
(1.04 g,2.22 mmol) in dry dichloromethane (60 mL) and the resulting re-
action mixture was stirred for 2 h before warming to 08C for 20 min. The
reaction was then quenched with saturated aqueous NaHCO3,poured
48.2,52.6,68.3,171.5,207.2 ppm; MS (CI):
m/z: calcd for C15H26NO5:
300.1811; found 300.1829 [M+H]+.
into
a separatory funnel,and extracted with dichloromethane (3”
50 mL). The organics were combined,dried over MgSO and concentrat-
4
Acknowledgements
ed under reduced pressure to afford a crude solid. Purification by column
chromatography (hexanes/EtOAc 3:1) gave 12 (0.85 g,89%) as a color-
less solid. M.p. 163–1648C; 1H NMR (300 MHz,CDCl 3): d=1.28 (d, 3J=
This work was financially supported by the National Science Foundation
CAREER program (CHE-0548003) and the University of Florida. A.J.L.
was supported by a University of Florida Alumni Graduate Fellowship
and A.A. by the National Science Foundation REU program (CHE-
0353828). K.A.A. thanks the National Science Foundation and University
of Florida for funding the X-ray crystallography equipment.
6.6 Hz,18H),3.76 (s,6H),5.03 (m,
3J=6.3 Hz,3H),8.50 ppm (s,3H);
13C NMR (75 MHz,CDCl 3): d=21.6,30.6,69.9,103.0,153.8,175.3 ppm;
MS (EI): m/z: calcd for C21H30O9: 426.1890,found 426.1879 [ M]+.
2,5,7-Tris-(2-isopropoxycarbonylmethyl)-1-aza-adamantane-4,6,10-trione
(6): Hexamethylenetetramine (0.270 g,1.93 mmol) was added to a solu-
tion of 12 (0.820 g,1.93 mmol) in isopropanol (15 mL) and the reaction
mixture was heated to reflux for 20 h. After cooling to RT,all volatiles
were removed under reduced pressure. The residue was purified by
column chromatography (hexanes/EtOAc 3:2) to afford 6 (0.61 g,66%)
[1] a) R. Hoffmann,A. Imamura,W. J. Hehre, J. Am. Chem. Soc. 1968,
90,1499–1509; b) R. Hoffmann, Acc. Chem. Res. 1971, 4,1–9; c) R.
Gleiter,W. D. Stohrer,R. Hoffmann,
Helv. Chim. Acta 1972, 55,
of a colorless solid. M.p. 111–1128C; 1H NMR (300 MHz,CDCl 3): d =
893–906; d) R. Gleiter, Angew. Chem. 1974, 86,770–775; Angew.
Chem. Int. Ed. Engl. 1974, 13,696–701; e) M. N. Paddon-Row, Acc.
Chem. Res. 1982, 15,245–251; f) R. Gleiter,W. Schafer, Acc. Chem.
Res. 1990, 23,369–375.
3
1.22 (d, J=6.0 Hz,18H),2.73 (s,6H),3.76 (s,6H),4.99 (hp,
3J=6.0 Hz,
3H); 13C NMR (75 MHz,CDCl 3): d = 21.6,32.0,68.2,70.3,70.8,169.0,
197.2 ppm; IR (KBr): n˜ =2984,2924,2853,1731,1700,1191,1109,
798 cmꢀ1; UV/Vis (ethanol): lmax (e)=265 nm (3010); MS (EI): m/z:
calcd for C24H33NO9: 479.2155; found 479.2159 [M]+.
[2] For excellent leading references on hyperconjugation see: a) J. B.
Lambert,Y. Zhao,R. W. Emblidge,L. A. Salvador,X. Liu,J.-H. So,
E. C. Chelius, Acc. Chem. Res. 1999, 32,183–190; b) I. V. Alabugin,
T. A. Zeidan, J. Am. Chem. Soc. 2002, 124,3175–3185; c) I. V. Ala-
bugin,M. Manoharan, J. Org. Chem. 2004, 69,9011–9024.
2,5,7-Tris-(2-isopropoxycarbonylmethyl)-10-exo-hydroxy-1-aza-adaman-
tane-4,6-dione (7): Ester 6 (0.180 g,0.376 mmol) was dissolved in dry
THF (15 mL) and cooled to ꢀ788C under argon. To this solution was
slowly added Red-Al (0.376 mmol,0.28 mL,65 wt% in toluene). The so-
lution temperature was maintained at ꢀ788C until the starting material
was consumed (30 min),and then the reaction was quenched with dilute
HCl. The resulting mixture was warmed to RT and the solvent was re-
moved in vacuo. The residue was redissolved in EtOAc,washed with
water and brine,dried over MgSO 4,and concentrated in vacuo. Column
chromatography (hexanes/EtOAc 2:1!1:1) afforded 7 (60 mg,33%) as a
white solid. M.p. 139–1408C; 1H NMR (300 MHz,CDCl 3): d =1.22 (m,
[3] This nomenclature designates the amine (lone pair,n orbital) as the
donor and the carbonyl (p* orbital) as the acceptor,although the or-
bital interactions are even more complex (e.g. n ! s*C–C, s ! p*C
,
=
O
etc.).
[4] a) R. C. Cookson,J. Henstock,J. Hudec, J. Am. Chem. Soc. 1966,
88,1060–1062; b) R. C. Cookson, Proc. R. Soc. London Ser. A 1967,
297,27–39.
[5] a) J. Kuthan,J. Palecek, Collect. Czech. Chem. Commun. 1963, 28,
2260–2264; b) W. A. Ayer,B. Altenkir,R. H. Burnell,M. Moinas,
Can. J. Chem. 1969, 47,449–455; c) J. Hudec, J. Chem. Soc. Chem.
Commun. 1970,829–831; d) T. Sasaki,S. Eguchi,T. Kiriyama,Y.
Sakito, J. Org. Chem. 1973, 38,1648–1652; e) A. W. J. Dekkers,
J. W. Verhoeven,W. N. Speckamp, Tetrahedron 1973, 29,1691–1696;
f) P. Pasman,J. W. Verhoeven,T. J. Deboer, Tetrahedron 1976, 32,
2827–2830; g) A. M. Halpern,A. L. Lyons, J. Am. Chem. Soc. 1976,
18H),2.48 (d, 3J=15.9 Hz,2H),2.56 (d, 3J=15.9 Hz,2H),2.62 (s,2H),
3
2.98 (d, J=12.6 Hz,2H),3.54 (s,2H),3.53 (s,1H),3.80 (d,
3J=12.9 Hz,
2H),4.52 (s,1H),4.98 ppm (m,3H);
13C NMR (75 MHz,CDCl 3): d=
21.6,21.7,31.7,34.6,54.8,59.2,68.0,68.3,68.7,69.9,71.1,169.7,170.9,
201.7 ppm; IR (KBr): n˜ =3467,2981,2936,2877,1726,1697,1374,1195,
1109,823 cm ꢀ1; UV/Vis (ethanol): lmax (e)=263 nm (1910); MS (EI):
m/z: calcd for C24H35NO9: 481.2312; found 481.2298 [M]+.
2-(((syn)-3,5-Di(2-isopropoxy-2-oxoethyl)-4-oxopiperidin-1-yl)methyl)-4-
98,3242–3247; h) J. R. Wiseman,H. O. Krabbenhoft,R. E. Lee,
J.
isopropoxy-4-oxobutanoic acid (30):
0.522 mmol) in HOAc/H2O 1:1 (16 mL) was heated under reflux for 1 h.
The solvent was then removed with reduced pressure and the residue
A
solution of
6
(0.250 g,
Org. Chem. 1977, 42,629–632; i) J. W. Verhoeven,in Electron Trans-
fer-from Isolated Molecules to Biomolecules, Pt. 1, Vol. 106 (Eds.: J.
Jortner,M. Bixon), 1999,pp. 603–644.
Chem. Eur. J. 2008, 14,1452 – 1463
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
1461

R. K. Castellano et al.
[6] N. Risch, J. Chem. Soc. Chem. Commun. 1983,532–533.
aza-adamantanes suggests that the substituents should have a mar-
ginal effect on the bond lengths and angles of the core. Reference
[24] states for 23 that “the deformation of the tricyclic system in-
duced by the incorporation of the phenyl ring is quite small.”
[7] a) H. Li,E. A. Homan,A. J. Lampkins,I. Ghiviriga,R. K. Castella-
no, Org. Lett. 2005, 7,443–446; b) A. J. Lampkins,O. Abdul-Rahim,
H. Li,R. K. Castellano, Org. Lett. 2005, 7,4471–4474; c) B. G.
Sumpter,V. Meunier,A. Vµzquez-Mayagoitia,R. K. Castellano, Int.
J. Quantum Chem. 2007, 107,2233–2242; d) B. G. Sumpter,V. Meu-
[28] P. W. Hickmott,M. G. Ahmed,S. A. Ahmed,S. Wood,M. Kapon,
Chem. Soc. Perkin Trans. 1 1985,2559–2571.
J.
nier,E. F. Valeev,A. J. Lampkins,H. Li,R. K. Castellano,
Chem. C in press.
J. Phys.
[29] M. G. Ahmed,S. M. I. Moeiz,S. A. Ahmed,F. Kiuchi,Y. Tsuda,P.
Sampson, J. Chem. Res. Synop. 1999,316–317.
[8] Notably,treatment of triester-functionalized phloroglucinol deriva-
tives with Lewis or protic acids results,under certain conditions,in
significant lactone formation. We are exploring this chemistry sepa-
rately and will report on it in due course.
[9] See a special issue on diastereoselectivity: a) A. S. Cieplak, Chem.
Rev. 1999, 99,1265–1336; b) W. Adcock,N. A. Trout, Chem. Rev.
1999, 99,1415–1435; c) S. Tomoda, Chem. Rev. 1999, 99,1243–
1263; d) G. Mehta,J. Chandrasekhar, Chem. Rev. 1999, 99,1437–
1467; e) B. W. Gung, Chem. Rev. 1999, 99,1377–1386; f) M. Kaselj,
W. S. Chung,W. J. le Noble, Chem. Rev. 1999, 99,1387–1413.
[10] a) J. M. Hahn,W. J. le Noble, J. Am. Chem. Soc. 1992, 114,1916–
1917; b) E. M. Gonikberg,W. J. le Noble, J. Org. Chem. 1995, 60,
7751–7755; c) B. W. Gung, Tetrahedron 1996, 52,5263–5301; d) M.
[30] V. A. Tafeenko,A. E. Prozorovskii,V. V. Kovalev,I. V. Knyazeva, J.
Struct. Chem. 1987, 28,155–158.
[31] a) D. A. Dougherty,W. D. Hounshell,H. B. Schlegel,R. A. Bell,K.
Mislow, Tetrahedron Lett. 1976, 17,3479–3482; b) D. A. Dougherty,
H. B. Schlegel,K. Mislow, Tetrahedron 1978, 34,1441–1447; c) E.
Osawa,P. M. Ivanov,C. Jaime, J. Org. Chem. 1983, 48,3990–3993;
d) D. A. Dougherty,C. S. Choi,G. Kaupp,A. B. Buda,J. M. Rudzin-
ski,E. Osawa, J. Chem. Soc. Perkin Trans. 2 1986,1063–1070;
e) N. L. Allinger,K. S. Chen,J. A. Katzenellenbogen,S. R. Wilson,
G. M. Anstead, J. Comput. Chem. 1996, 17,747–755; f) K. K. Bal-
dridge,T. R. Battersby,R. VernonClark,J. S. Siegel,
Soc. 1997, 119,7048–7054.
J. Am. Chem.
[32] V. Galasso, THEOCHEM 2000, 528,171–176.
Kaselj,E. M. Gonikberg,W. J. le Noble,
J. Org. Chem. 1998, 63,
[33] a) A. M. Brouwer,B. Krijnen, J. Org. Chem. 1995, 60,32–40;
3218–3223; e) S. Tomoda,T. Senju, Tetrahedron 1999, 55,5303–
5318.
b) W. D. Oosterbaan,R. W. A. Havenith,C. A. van Walree,L. W.
Jenneskens,R. Gleiter,H. Kooijman,A. L. Spek,
J. Chem. Soc.
[11] a) Y. Senda,M. Morita,H. Itoh, J. Chem. Soc. Perkin Trans. 2 1996,
221–224; b) Y. Senda, Chirality 2002, 14,110–120.
Perkin Trans. 2 2001,1066–1074.
[34] Worth noting,DFT-LDA/cc-pVDZ calculations already performed
on 4 (and reported in ref. [7c]) show bond lengths and angles for the
AAT core that agree very well with those reported herein (generally
within 0.01 Š and 1.58,respectively). Calculations at the molecular
mechanics level (e.g. Amber* force field) do not reproduce the
bond length changes,a phenomenon noted in ref. [31e] but argued
against in ref. [31f].
[12] a) J. M. Coxon,K. N. Houk,R. T. Luibrand, J. Org. Chem. 1995, 60,
418–427; b) V. Yadav,G. Senthil,D. Jeyaraj, Tetrahedron 1999, 55,
14211–14218; c) F. Jimenez-Cruz,R. Cetina,H. Rios-Olivares,
Chem. Lett. 2000,956–957; d) D. Moraleda,C. Ollivier,M. Santelli,
Tetrahedron Lett. 2006, 47,5471–5474.
[13] L. Verotta,O. Sterner,G. Appendino,E. Bombardelli,T. Pilati,
Eur.
J. Org. Chem. 2006,5479–5484.
[14] B. R. Cho,K. Chajara,H. J. Oh,K. H. Son,S.-J. Jeon,
[35] D. Kardel,W. Knoche,N. Risch, J. Chem. Soc. Perkin Trans. 2 1993,
1455–1459.
Org. Lett.
2002, 4,1703–1706.
[36] The n–p* transition is too weak to be observed in these systems (see
related results in ref. [5d]).
[37] This is not the case for through-bond interactions that show marked
charge-transfer character. See,for example,ref. [19].
[38] Cooperative stretching of the carbonyls in these systems has been
noted before (see below,ref. [43a]),but we could not find any refer-
ences of a comprehensive study or discussion.
[15] When the cyclization is performed in methanol,the triisopropyl
ester (20%) is obtained in addition to two of the three possible
methyl ester derivatives (the mono-,21%,and bis-methylester,
6%).
[16] A. J. Lampkins,O. Abdul-Rahim,R. K. Castellano, J. Org. Chem.
2006, 71,5815–5818.
[17] B. W. Domagalska,L. Syper,K. A. Wilk,
1931–1939.
Tetrahedron 2004, 60,
[39] D. A. Lightner,T. D. Bouman,W. M. Donald Wijekoon,A. E.
Hansen, J. Am. Chem. Soc. 1984, 106,934–944.
[18] S. Simaan,J. S. Siegel,S. E. Biali,
3701.
J. Org. Chem. 2003, 68,3699–
[40] D. Lenoir,P. Mison,E. Hyson,P. v. R. Schleyer,M. Saunders,P.
Vogel,L. A. Telkowski, J. Am. Chem. Soc. 1974, 96,2157–2164.
[19] a) B. Krijnen,H. B. Beverloo,J. W. Verhoeven,C. A. Reiss,K. Gou-
bitz,D. Heijdenrijk, J. Am. Chem. Soc. 1989, 111,4433–4440;
b) D. J. A. De Ridder,K. Goubitz,H. Schenk,B. Krijnen,J. W. Ver-
hoeven, Helv. Chim. Acta 2003, 86,799–811; c) D. J. A. De Ridder,
[41] R. K. Murray,Jr.,D. L. Goff,T. M. Ford,
3870–3874.
[42] G. E. Henry,H. Jacobs,C. M. S. Carrington,S. McLean,W. F. Rey-
nolds, Tetrahedron Lett. 1996, 37,8663–8666.
J. Org. Chem. 1977, 42,
K. Goubitz,H. Schenk,B. Krijnen,J. W. Verhoeven,
Acta 2003, 86,812–826.
Helv. Chim.
[43] a) N. Risch, Chem. Ber. 1985, 118,4849–4856; b) N. Risch,U. Biller-
beck,E. Krieger, Chem. Ber. 1992, 125,459–465.
[20] R. K. Castellano, Curr. Org. Chem. 2004, 8,845–865.
[21] Compared with similar isopropyl ester substituents in the CSD
(HEFKOS10=1.508 Š; HEFKUY10=1.509 Š).
[44] B. W. Gung,M. A. Wolf, J. Org. Chem. 1996, 61,232–236.
[45] J. M. Grill,J. W. Ogle,S. A. Miller, J. Org. Chem. 2006, 71,9291–
9296.
[22] J. P. Degelaen,P. Pham,E. R. Blout, J. Am. Chem. Soc. 1984, 106,
4882–4890.
[46] H. Duddeck,V. Wiskamp,D. Rosenblum, J. Org. Chem. 1981, 46,
5332–5336.
[23] The X-ray structures of 6 and 7 have been collected at 173 K,while
those taken from the literature for comparison have generally been
collected at room temperature (290–295 K). The consequence of
temperature on through-bond interactions in these systems (in the
solid state) is not known.
[47] The average value for 1-aza-adamantanediones from ref. [43a] is d
ꢁ207 ppm. Such an NMR comparison of course depends significant-
ly on how the chemical shifts are referenced (as well as other pa-
rameters).
[48] a) C. Worrell,J. W. Verhoeven,W. N. Speckamp, Tetrahedron 1974,
30,3525–3531; b) L. Toom,A. Kutt,I. Kaljurand,I. Leito,H. Ottos-
son,H. Grennberg,A. Gogoll, J. Org. Chem. 2006, 71,7155–7164.
[49] The latter two reagents oxidize diaza-adamantanones; see ref. [5d].
[50] K. Bowden,E. J. Grubbs, Chem. Soc. Rev. 1996, 25,171–177.
[51] a) L. M. Stock, J. Chem. Educ. 1972, 49,400–404; b) R. D. Topsom,
Prog. Phys. Org. Chem. 1976, 12,1–20; c) R. Gleiter,M. Kobayashi,
J. Kuthan, Tetrahedron 1976, 32,2775–2781; d) G. Spanka,P. Rade-
macher, J. Org. Chem. 1986, 51,592–596.
[24] N. Risch,M. Langhals,W. Mikosch,H. Bçgge,A. Mꢂller,
J. Am.
Chem. Soc. 1991, 113,9411–9412.
[25] F. Jimꢃnez-Cruz,R. Cetina-Rosado,S. Hernµndez-Ortega,R. A. To-
scano,H. Rꢄos-Olivares, Acta Crystallogr. Sect. C 2001, 57,868–869.
[26] F. Jimꢃnez-Cruz,H. Rꢄos-Olivares,S. Hernµndez-Ortega,A. Cer-
vantes-Nevarez, J. Mol. Struct. 2003, 655,23–29.
[27] Side-by-side analysis of some structures is complicated by substitu-
tion a to the nitrogen,although comparison to similarly substituted
1462
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2008, 14,1452 – 1463

Donor–Acceptor Interactions in b-Aminoketones
FULL PAPER
[52] P. Geneste,I. Hugon,C. Reminiac,G. Lamaty,J. P. Roque,
Bull.
[57] T. Scherer,W. Hielkema,B. Krijnen,R. M. Hermant,C. Eijckelhoff,
F. Kerkhof,A. K. F. Ng,R. Verleg,E. B. Vandertol,A. M. Brouwer,
J. W. Verhoeven, Recl. Trav. Chim. Pays-Bas 1993, 112,535–548.
Soc. Chim. Fr. Part 2 1976,845–846.
[53] a) C. A. Grob,P. W. Schiess, Angew. Chem. 1967, 79,1–14; Angew.
Chem. Int. Ed. Engl. 1967, 6,1–15; b) C. A. Grob, Angew. Chem.
1969, 81,543–554; Angew. Chem. Int. Ed. Engl. 1969, 8,535–546.
[58] SHELXTL6,Bruker-AXS,Madison,Wisconsin,USA,
2000.
[59] P. v. d. Sluis,A. L. Spek, Acta Crystallogr. Sect. A 1990, 46,194–201.
[60] A. L. Spek, J. Appl. Crystallogr. 2003, 36,7–13.
[54] N. Risch,U. Billerbeck,B. Meyerroscher,
Chem. Ber. 1993, 126,
1137–1140.
[55] N. Risch,M. Langhals,T. Hohberg,
4465–4468.
Tetrahedron Lett. 1991, 32,
[56] That the 3,5-substituents are cis diequatorial is verified by coupling
constants; the a-hydrogen (to the ketone) couples with the b-hydro-
Received: August 6,2007
Published online: November 22,2007
3
gen with J
A
Chem. Eur. J. 2008, 14,1452 – 1463
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
1463