Diastereoselective electrophilic substitution of α-amino-substituted benzylic organometallics

Article abstract of DOI:10.1039/b109633h

Reductive metallation of a diastereoisomeric bicyclic 2-phenyloxazolidine derived from 2-hydroxymethylpiperidine occurs with racemization at the benzylic carbon atom. Reaction of intermediate organometallics with alkyl halides affords substituted amino alcohols in a highly syn-selective fashion. Observed diastereoselectivities are rationalized in terms of rapidly equilibrating epimeric intermediate organometallics, one of which reacts preferentially under appropriate reaction conditions. Deuteration of the same intermediates usually leads to deuterated amino alcohols with low diasteroselectivities, unless lithium is employed as the reducing agent and the resulting mixture is allowed to equilibrate before deuteration.

Full text of DOI:10.1039/b109633h

View Article Online / Journal Homepage / Table of Contents for this issue
Diastereoselective electrophilic substitution of ꢀ-amino-substituted
benzylic organometallics
Ugo Azzena
1
Dipartimento di Chimica, Università di Sassari, via Vienna 2, 07100 Sassari, Italy.
E-mail: ugo@ssmain.uniss.it
Received (in Cambridge, UK) 22nd October 2001, Accepted 27th November 2001
First published as an Advance Article on the web 9th January 2002
Reductive metallation of a diastereoisomeric bicyclic 2-phenyloxazolidine derived from 2-hydroxymethylpiperidine
occurs with racemization at the benzylic carbon atom. Reaction of intermediate organometallics with alkyl halides
affords substituted amino alcohols in a highly syn-selective fashion. Observed diastereoselectivities are rationalized
in terms of rapidly equilibrating epimeric intermediate organometallics, one of which reacts preferentially under
appropriate reaction conditions. Deuteration of the same intermediates usually leads to deuterated amino alcohols
with low diasteroselectivities, unless lithium is employed as the reducing agent and the resulting mixture is allowed
to equilibrate before deuteration.
Introduction
Reductive cleavage of arylmethyl alkyl ethers by electron trans-
fer from alkali metals in ethereal solvents results in the
regioselective cleavage of the arylmethyl carbon–oxygen bond,
with generation of arylmethyl organometallics.1 Application
Scheme 1 Synthesis of 3-phenylhexahydrooxazolo[3,4-a]pyridine 1.
Reagents, conditions and yield: (i) CH3COOH, C6H6, reflux, 4 h (83%,
of this procedure to the reductive metallation of aromatic
acetals or α-N,N-dialkylamino-substituted benzyl alkyl ethers
allows the generation of, respectively, α-oxygen-2 and α-amino-
substituted3 arylmethyl organometallics. Furthermore, this
procedure is well suited to generate oxy- or amino-function-
alized arylmethyl organometallics via reductive metallation
of suitable α-aryl-substituted heterocycles.3–6 Interestingly, a
dilithium derivative generated by reductive metallation of an
oxazolidine, i.e., an organolithium bearing an oxyanionic
group, was more stable under the reaction conditions than a
similar monolithium derivative generated from an open-chain
amino ether.3
92 : 8 mixture of diastereoisomers).
assigned according to 1H NMR NOE difference spectra;
in particular, presaturation of the H3 resonance resulted in
enhancement (4%) of the signal for the H8a proton and
vice-versa (Fig. 1).
The configurational stability of chiral α-substituted aryl-
methyl organometallic reagents and their ability to react with
electrophiles in a stereodefined manner is a topic of current
interest in organic chemistry.7–12 Investigations on the carbo-
lithiation of cinnamyl alcohol and derivatives have shown that
appropriately positioned substituents are sometimes able to
coordinate the organometallic centre, fixing its configuration
and allowing a diastereoselective alkylation.13–16
I wish now to report that reductive metallation of the
diastereoisomeric bicyclic 2-phenyloxazolidine 1 followed by
reaction with alkyl halides allows the highly diastereoselective
synthesis of N-(α-alkylbenzyl)-2-(hydroxymethyl)piperidines,
via intermediate formation of α-amino-substituted organo-
metallic derivatives bearing a potentially coordinative oxy-
anionic group. A preliminary report has already appeared.17
Fig. 1 Stereochemistry of the major stereoisomer of the oxazolidine 1,
1
as determined by H NMR NOE difference spectra. Besides geminal
correlations, H1 shows correlations with H8a; H3 shows NOE
correlations with H5 and H8a; H8a shows NOE correlations with H1,
H3 and H5.
Reductive metallation reactions of 1
The reduction of 1 was carried out under Ar with Li or K metal
in the presence of a catalytic amount of naphthalene in
tetrahydrofuran (THF) (Scheme 2, Table 1). The results of D2O;
t-BuOD- and 2,6-[(CH3)3C]2C6H3OD-quenching experiments,
carried out to check the formation of carbanionic inter-
mediates, as well as their behaviour as a function of reaction
time, temperature and concentration, are reported in Table 1.
Reaction of the oxazolidine 1 with Li metal (5 equiv.) in the
presence of a catalytic amount of naphthalene (10 mol%) in
THF at Ϫ20 ЊC for 4 h furnished a deep red heterogeneous
reaction mixture which, upon aqueous work-up, afforded
amino alcohol 3 in quantitative yield (Table 1, entry 1).
Results
Starting materials
The oxazolidine 1 was synthesized in 83% isolated yield as a
92 : 8 mixture of diastereoisomers by the acid-catalyzed reac-
tion of benzaldehyde with rac-2-(hydroxymethyl)piperidine,
according to a known procedure (Scheme 1).18 Selective 1H–1H
decoupling experiments allowed assignment of the 1H
resonances and a syn-relationship between H3 and H8a was
360
J. Chem. Soc., Perkin Trans. 1, 2002, 360–365
This journal is © The Royal Society of Chemistry 2002
DOI: 10.1039/b109633h

View Article Online
Table 1 Reductive metallation of oxazolidines 1
Product
E
Diastereoisomeric
Entry
Metal
Time (t/h)
EX
Temp. (θ/ЊC)a
Yield (%)b
ratio 4a : 4bb
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Li
Li
Li
Li
Li
Li
Li
Li
Li
Li
Li
Li
K
4
4
4
4
4
4
6
8
22
22
8
15
1
1
4
H2O
D2O
t-BuOD
ArODc
D2O
t-BuOD
D2O
D2O
D2O
ArOD
D2O
D2O
Ϫ20
Ϫ20
Ϫ20
Ϫ20
Ϫ80d
Ϫ80d
Ϫ20
Ϫ20
Ϫ20
Ϫ20
Ϫ20
Ϫ20
Ϫ20
Ϫ20
Ϫ20
Ϫ20
3, H
>95
>95
>95
>95
>95
>95
>95
>95
>95
>95
>95
>95
>95
83
4a/4b, D
4a/4b, D
4a/4b, D
4a/4b, D
4a/4b, D
4a/4b, D
4a/4b, D
4a/4b, D
4a/4b, D
4a/4b, D
4a/4b, D
3, H
52 : 48
55 : 45
56 : 44
55 : 45
56 : 44
76 : 24
94 : 6
94 : 6
84 : 16
61 : 39e
61 : 39e
H2O
D2O
D2O
D2O
K
K
K
4a/4b, D
4a/4b, D
4a/4b, D
74 : 26
75 : 25
75 : 25
81
51
24
a All reactions run in the presence of 5 equiv. of the metal and 10 mol% of naphthalene; starting concentration of 1 is 0.05 M, unless otherwise
indicated. b As determined by 1H NMR spectroscopy. c ArOD = 2,6-[(CH3)3C]2C6H3OD. d Reductive metallation performed at Ϫ20 ЊC, D2O added at
Ϫ80 ЊC. e Starting concentration of 1 is 0.01 M.
amount of naphthalene (10 mol%) in THF (Table 1, entry 13).
Quenching of the reaction mixture with D2O showed inter-
mediate formation of organometallic reagent(s) and afforded
amino alcohols 4a and 4b in a 74 : 26 diastereoisomeric ratio;
although the intermediate organopotassium reagent slowly
decays, the diastereoisomeric ratio of recovered amino alcohols
is practically time independent (Table 1, entries 14–16).
Reductive metallation and alkylation of 1
The reactivity of the intermediate organometallic reagent(s)
obtained as described in the above paragraph towards alkyl
halides was investigated (Scheme 2, Table 2). Reductive lithi-
ation followed by reaction with PhCH2Cl, n-C6H13Br, n-C4H9I
or n-C4H9Br afforded amino alcohols 5a and 5b, 6a and 6b and
7a and 7b, respectively, with yields ranging from 75 to 45% and
good diastereoselectivities (Table 2, entries 1–6). Furthermore,
the diastereoisomeric ratio of recovered amino alcohols 5a and
Scheme 2 Reductive metallation and reaction with electrophiles of
5b and 7a and 7b is time independent (Table 2, entry 1 vs. entry
the oxazolidine 1. Reagents and conditions: (i) Li or K, naphthalene
2 and entry 4 vs. entry 5).
A comparable result was obtained in the K-mediated
(10 mol%), THF; (ii) EX, then water.
reductive cleavage of 1 followed by quenching with n-C4H9Br
Quantitative formation of intermediate organometallic(s)
(Table 2, entry 7).
was evidenced after quenching of the reductive lithiation
In contrast to the above reported results, low yields and
mixture with D2O. Under these conditions, amino alcohols 4a
low diastereoselectivites were observed upon quenching of
and 4b were obtained in a 52 : 48 ratio (Table 1, entry 2);
the mixture of reductive lithiation of the oxazolidine 1 with
quenching the reaction mixture with t-BuOD or with 2,6-
n-C4H9Cl (Table 2, entries 8 and 9).
[(CH3)3C]2C6H3OD, or lowering the temperature to Ϫ80 ЊC
prior to quenching, did not affect this result (Table 1, entries
Relative stereochemistry of amino alcohol 4a
3–6).
However, when the reaction mixture was stirred for 6–22 h
at Ϫ20 ЊC before quenching with D2O, the ratio between the
diastereoisomers rose to 94 : 6 (Table 1, entries 7–9); a similar
result was obtained upon quenching of the reaction mixture
with 2,6-[(CH3)3C]2C6H3OD after 22 h at Ϫ20 ЊC (Table 1,
entry 10).
The above-described dramatic effect of reaction time on
the diastereoisomeric ratio of recovered amino alcohols 4a
and 4b was observed in reactions run with a starting 0.05 M
concentration of the oxazolidine 1. Lowering the starting
concentration of the oxazolidine 1 to 0.01 M led to a com-
pletely different result. Indeed, under the new conditions,
quenching of the reaction mixture with D2O afforded amino
alcohols 4a and 4b in a 61 : 39 diastereoisomeric ratio,
independent of the reaction time (Table 1, entries 11 and 12).
Reductive cleavage of the oxazolidine 1 can be obtained also
by the action of K metal (5 equiv.) in the presence of a catalytic
To determine the stereochemistry of the reductive metallation
reaction followed by deuteration of intermediate organo-
metallic(s), amino alcohols 3 and 4a were cyclized to the
corresponding 1,2,3,4-tetrahydroisoquinolines, in order to
assess the relative orientation of protons at stereocentres by
means of NOE correlations.
According to a procedure developed by Mendelson et al.,19,20
the tetrahydroisoquinoline 8 was obtained in 21% isolated yield
by reaction of amino alcohol 3 in fused AlCl3 and NH4Cl.
Application of the same procedure to a 94 : 6 mixture of deu-
terated amino alcohols 4a and 4b (%D > 95) led to isolation
of the corresponding deuterated tetrahydroisoqunolines 9a and
9b (%D > 95) as an 89 : 11 mixture of diastereoisomers in 22%
isolated yield (Scheme 3).
Although a minor variation of the diastereoisomeric ratio
was evidenced, it was confidently assumed that a comparison
between the relative stereochemistry of starting material and
J. Chem. Soc., Perkin Trans. 1, 2002, 360–365
361

View Article Online
Table 2 Reductive alkylation of oxazolidines 1
Product
E
Diastereoisomeric
Entry
Metal
Time (t/h)
EX
Temp. (θ/ЊC)a
Yield (%)b
ratio a : bc
1
2
3
4
5
6
7
8
9
Li
Li
Li
Li
Li
Li
K
4
8
4
4
8
4
4
4
4
PhCH2Cl
PhCH2Cl
n-C6H13Br
n-C4H9I
Ϫ20
Ϫ20
Ϫ20
Ϫ20
Ϫ20
Ϫ20
Ϫ20
Ϫ20
0e
5a/5b, PhCH2
5a/5b, PhCH2
6a/6b, n-C6H13
7a/7b, n-C4H9
7a/7b, n-C4H9
7a/7b, n-C4H9
7a/7b, n-C4H9
7a/7b, n-C4H9
7a/7b, n-C4H9
74d
75
93 : 7
92 : 8
>95 : <5
>95 : <5
94 : 6
>95 : <5
>95 : <5
63 : 27
50d
57d
45
n-C4H9I
n-C4H9Br
n-C4H9Br
n-C4H9Cl
n-C4H9Cl
56
63
<5
23
Li
Li
67 : 23
a All reactions run in the presence of 5 equiv. of the metal; starting concentration of 1 is 0.05 M. b As determined by 1H NMR spectroscopy, unless
otherwise indicated. c As determined by GC–MS; compounds 5–7a showed higher retention times than diastereoisomers 5–7b. d Isolated yield.
e Reductive metallation performed at Ϫ20 ЊC, n-C4H9Cl added at 0 ЊC.
Scheme 3 Synthesis of isoquinolines 8 and 9: Reagents and conditions:
(i) AlCl3–NH4Cl, 180 ЊC, 15 h.
reaction product is reliable: indeed, no loss of deuterium at the
benzylic position was observed, and a comparable result was
reported in the cyclization of a similar benzylaminoethanol
selectively deuterated at the carbon atom in the α-position to
the hydroxy group.20
Scheme 4 Cyclization of amino alcohols 3a (E = H) and syn-4a (E =
A comparison of the 1H NMR spectra of isoquinolines 8 and
D) and stereochemistry of the isoquinoline 8 (E = H6) and 9a (E = D),
9 allowed us to assign the configuration of the major stereo-
isomer of the deuterated product and, therefore, to assign the
relative stereochemistry to the major diastereoisomer of amino
as determined by 1H NMR NOE difference spectra: besides geminal
correlations, H6 shows NOE correlations with H4 and H11a; H6Ј
shows NOE correlations with H4Ј; H4 shows correlatons with H6 and
H11a.
alcohols 4a and 4b.
1
1
Selective H–1H decoupling experiments as well as H–13C
COSY experiments allowed assignment of the 1H resonances of
the isoquinoline 8, and a syn-relationship between H6 and
H11a was assigned according to 1H NMR NOE difference
spectra. In particular, presaturation of the H6 resonance in the
oxazolidine 8 resulted in enhancement (2%) of the signal for the
H11a proton.
1
The H NMR spectrum of the deuterated isoquinolines 9a
Scheme
5
Syntheses of diastereoisomeric alcohols 7a and 7b.
and 9b showed two α-aminobenzylic protons in a 89 : 11 ratio
corresponding, respectively, to the H6Ј and H6 resonances of
the isoquinoline 8 (see Experimental section). This finding is
indicative of a syn relationship between the α-aminobenzylic
deuterium and H11a in the major stereoisomer, i.e., the
isoquinoline 9a. Accordingly, the anti configuration can be
ascribed to the deuterated isoquinoline 9a, and the syn
configuration to the deuterated amino alcohol 4a (Scheme 4).
Reagents and conditions: (i) Li or K, THF, Ϫ20 ЊC; then BuBr: 7a : 7b =
>95 : <5; (iii) BuMgBr, THF, 0 ЊC: 7a : 7b = 30 : 70.
oxazolidines should occur with an overall prevailing inversion
of configuration at the electrophilic carbon21–25 thus leading, in
the present example, to the preferential formation of the anti
amino alcohol. Accordingly, the anti configuration can be
ascribed to amino alcohol 7b, and the syn configuration to
amino alcohol 7a, as well as to amino alcohols 5a and 6a.
Relative stereochemistry of amino alcohols 5–7
To determine the relative stereochemistry of the reductive
alkylation products, amino alcohols 7a and 7b were synthesized
by reaction of the oxazolidine 1 with n-C4H9MgBr (Scheme 5).
Under these conditions, amino alcohols 7a and 7b were
obtained as a 30 : 70 diastereoisomeric mixture (as determined
by GC–MS) in 32% isolated yield. According to the literature,
reaction of Grignard reagents with N-substituted-2-aryl-
Discussion
This paper describes the generation of organometallic deriv-
atives which behave in a markedly different way with different
electrophiles. Indeed, reaction of both lithium and potassium
intermediates with deuterating reagents usually led, with
one notable exception discussed below, to the formation of
362
J. Chem. Soc., Perkin Trans. 1, 2002, 360–365

View Article Online
deuterated amino alcohols 4 with low diastereoselectivities
(Table 1, entries 2–7, 11, 12 and 14–16) whilst, in contrast,
reaction of the same intermediates with alkyl halides usually
led to amino alcohols 5–7 with good to very good diastereo-
selectivities (Table 2, entries 1–7).
It is accordingly proposed that reductive metallation of the
oxazolidine 1 occurs with racemization, giving rise to a pair
of rapidly interconverting diastereoisomeric intermediates,
syn-2 and anti-2. Racemization at the benzylic carbon atom is
supported not only by the above-mentioned deuteration
experiments, but also by the reaction mechanism usually
considered for the reductive metallation of benzyl alkyl ethers,
which occurs via intermediate formation of configurationally
labile benzylic radicals.1, †
Observed diastereoselectivities can be rationalized by
assuming that the configuration at the benzylic carbon of
organometallic intermediates is fixed by coordination of
the carbanionic mojety to the alkoxy group.13–16 One of the
intramolecularly coordinated organometals reacts preferen-
tially and stereoselectively with carbon electrophiles (either
with retention or inversion of configuration),‡ i.e., the
behaviour of the intermediate organometallics with alkyl
halides can be rationalized by assuming that the activation
energy of the electrophilic reaction is higher than the activation
energy of their epimerization.9,13 It is particularly interesting to
note that potassium as a counter-ion is, in the present example,
as effective as lithium for such coordination. On the other hand,
deuteration of the equilibrating reaction mixture requires an
activation energy lower than (or comparable to) the activation
energy of epimerization.9,13 This hypothesis is represented in
Scheme 6, where a specific coordination between the alkoxy
diastereoselectivity (Table 1, entries 7–10). Such behaviour
could be, in principle, an indication of thermodynamic
resolution of the intermediates. Alternatively, owing to the
heterogeneity of the reaction mixture, I suggest dynamic
resolution through preferential crystallization of one diastereo-
isomeric intermediate.28 Indeed, under the above-reported
reaction conditions (initial concentration of the oxazolidine 1 =
0.05 M), resolution could result from a selective dissolution/
epimerization/selective crystallization process.
The completely different result obtained upon performing
the reductive cleavage reaction at higher dilution (initial con-
centration of the oxazolidine 1 = 0.01 M, Table 1, entries 11 and
12), i.e., no variation of the diastereoisomeric ratio of amino
alcohols 4a and 4b with time, supports this hypothesis.
Although always in the presence of an heterogeneous reaction
mixture, dilution could well affect the solubility of inter-
mediates and the selectivity of a dissolution/crystallization
process, whilst it should not affect thermodynamic resolution
to such an extent.
The finding that resolution of the intermediates did not affect
the result of alkylation reactions (Table 2, entry 1 vs. entry 2,
and entry 4 vs. entry 5) can be accordingly rationalized. As
mentioned above, I suggest that the activation energy of alkyl-
ation reactions is higher than the activation energy of inter-
conversion of the epimeric intermediate organometallics.
According to the Curtin–Hammett principle, the product
composition of a similar reaction is solely dependent on the
difference in energies of the transition states; as a consequence,
whatever the relative amount of the most reactive intermediate
in the reaction mixture after equilibration, it should drive the
reaction to an unchanging diastereoisomeric composition of
reaction products.
Whatever the reason of the resolution process, it strongly
supports the hypothesis concerning the formation of inter-
converting diastereoisomeric organometallic intermediates
and rules out a possible alternative explanation of the above-
reported results, i.e., formation of essentially planar organo-
metallic intermediates which react stereoselectively with alkyl
halides, and non-stereoselectively with deuterating agents.
The last hypothesis appears unlikely also in view of the
results obtained in the synthesis of amino alcohols 7a and 7b
by alkylation of intermediates 2 (M = Li), where n-C4H9I and
n-C4H9Br afforded higher diastereoselectivities than did the less
reactive n-C4H9Cl (Table 2, entries 4–6, 8 and 9). Apparently,
the low reactivity of n-C4H9Cl rendered the alkylation step less
stereospecific, i.e., one (or both) intermediates 2 (M = Li)
reacted comparatively well with this electrophile under reten-
tion or inversion of configuration at the organometallic centre.
This hypothesis is in agreement with the statement that
activation barriers for retentive and inverse electrophilic attack
at organometallic derivatives should not differ very much.13,29
In summary, the results reported in the present investigation
show that reductive metallation of the diastereoisomeric oxazol-
idine 1 is a practical approach to a highly syn-selective synthesis
of N-(α-substituted benzyl)-2-(hydroxymethyl)piperidines,
illustrate an interesting example of resolution of diastereo-
isomeric organometallics, and add additional evidence to the
hypothesis that coordination with an appropriately positioned
substituent enhances the configurational stability of organo-
metallic centres.
Scheme 6 Epimerization and reaction with electrophiles of syn-2 and
anti-2; M = Li or K; AlkylX = PhCH2Cl, n-C6H13Br, n-C4H9I or n-
C4H9Br; R = D, (CH3)3C or 2,6-[(CH3)3C]2C6H3.
and the carbanionic mojety is not shown. Indeed, fixation of
configuration at the benzylic carbon could involve chelation of
benzylic metal by oxygen or interaction of alkoxide metal with
the benzylic carbon,16 and it has been recently pointed out that
the actual nature of such a coordination depends on several
parameters like, inter alia, solvation and formation of
aggregates.14
Experimental
Interestingly, equilibration of intermediates 2 (M = Li) before
deuteration afforded amino alcohols 4a and 4b with high
General remarks
Boiling points are uncorrected; the air-bath temperature on
bulb-to-bulb distillations is given as the boiling point. Starting
materials were of the highest commercial quality and were used
without further purification. D2O was 99.8% isotopic purity.
THF was distilled from Na–K alloy under N2 immediately prior
† Examples of configurationally stable Cr(CO)3-complexed benzylic
radicals were recently reported.26
‡ Reaction of benzyllithium derivatives with different electrophiles can
occur either with retention or with inversion of configuration.27
J. Chem. Soc., Perkin Trans. 1, 2002, 360–365
363

View Article Online
1
to use. H NMR spectra were recorded at 300 MHz, and 13C
known. Other products were characterized as follows. Amino
alcohols 5b and 6b were not separated from the corresponding
major stereoisomer; their relative amount was determined
by GC–MS, and they were characterized by means of their
repective MS fragmentation pattern, closely related to the
fragmentation pattern of the corresponding major stereo-
isomer. Yields are reported in the Tables.
NMR spectra were recorded at 75 MHz, in CDCl3 (unless
otherwise indicated) with SiMe4 as internal standard. Deuter-
ium incorporation was calculated by monitoring the H NMR
1
spectra of the crude mixtures and comparing the integration of
the signal corresponding to the proton in the arylmethyl pos-
ition with that of known signals. Mass spectra were recorded on
a quadrupole mass spectrometer operating at 70 eV, interfaced
with a gas chromatograph equipped with a DBS 30-m capillary
column (i.d. 0.25 mm). IR spectra were recorded on thin films.
Elemental analyses were performed by the Microanalytical
Laboratory of the Dipartimento di Chimica, Università di
Sassari.
Grignard reactions were performed according to a general
procedure described in ref. 25. Petroleum ether refers to the
fraction with distillation range 35–60 ЊC. TLC analyses were
performed on aluminium plates precoated with silica gel 60
F254, developed with the solvent systems reported for flash
chromatography purifications, and visualized by UV light.
N-(ꢀ-Deuterio benzyl)-2-(hydroxymethyl)piperidine (syn-4a :
anti-4b ؍ 94 : 6). Oil; purified by flash chromatography
(AcOEt–petroleum ether–Et3N = 3 : 7 : 1), Rf = 0.34; bp 120–
125 ЊC/1 mmHg; νmax 3410 cmϪ1; δH 1.35 (2H, m, CH2), 1.63
(4H, m, 2 × CH2), 2.14 (1H, m, CHN), 2.46 (1H, m, CHN), 2.58
(1H, br s, OH), 2.86 (1H, m, CHN), 3.29 (0.94H, br s, CHD,
4a), 3.51 (1H, dd, J = 10.8, 3.9 Hz, CHO), 3.87 (1H, dd, J =
10.8, 4.2 Hz, CHO), 4.03 (0.06H, br s, CHD, 4b), 7.28 (5H, m,
ArH); δC 23.4, 24.1, 27.3, 50.7, 57.3 (t, J = 21.0 Hz), 60.8, 62.2,
127.0, 128.3, 128.8, 138.9 (Found: C, 75.9; H ϩ D, 10.0; N, 6.5.
C13H18DNO requires C, 75.7; H ϩ D, 9.8; N, 6.8%).
syn-N-(1Ј,2Ј-Diphenylethyl)-2-(hydroxymethyl)piperidine 5a.
Oil; purified by flash chromatography (AcOEt–petroleum
ether–Et3N = 4 : 6 : 1), Rf = 0.43; bp 200–205 ЊC/1 mmHg; νmax
3400 cmϪ1; δH 1.49 (6H, m, 3 × CH2), 2.50 (1H, br s, OH), 2.82
(3H, m, 3 × CHN), 2.95 (1H, dd, J = 13.5, 10.2 Hz, PhCH ),
3.31 (1H, dd, J = 13.5, 4.2 Hz, PhCH ), 3.60 (1H, dd, J = 10.5,
5.1 Hz, CHO), 3.69 (1H, dd, J = 10.5, 6.6 Hz, CHO), 4.33 (1H,
dd, J = 10.2, 4.2 Hz, PhCHN), 6.96 (2H, m, ArH), 7.15 (8H, m,
ArH); δC 22.0, 22.7, 24.8, 36.5, 43.2, 55.8, 60.8, 64.4, 125.8,
Preparation of starting materials
Oxazolidine derivative 1 was obtained in 83% isolated yield
according to a procedure described in ref. 18 and purified by
vacuum distillation; diastereoisomeric ratio was calculated by
monitoring 1H NMR spectra and comparing integrations
of clearly separated singlet resonances of H3 protons. The
oxazolidine 1 was characterized as follows:
127.0, 128.0, 128.0, 128.8, 129.2, 139.4, 141.7; m/z 264 (Mϩ
Ϫ
3-Phenylhexahydrooxazolo[3,4-a]pyridine 1. Oil; bp 117–120
ЊC/1 mmHg (lit.,18 134–136/8 Torr); δH (syn stereoisomer,
CD3OD) (Hs at C6, C7 and C8 superimpose between δ 1.28 and
1.90), 1.49 (4H, m), 1.84 (2H, m), 2.04 (1H, td, J = 11.0, 3.3 Hz,
H5), 2.51 (1H, m, H8a), 2.71 (1H, m, H5Ј), 3.67 (1H, dd,
J = 9.9, 6.6 Hz, H1Ј), 4.07 (1H, t, J = 6.6 Hz, H1), 4.57 (1H, s,
H3), 7.34 (3H, m, ArH), 7.44 (2H, m, ArH); δC (syn stereo-
isomer, CD3OD) 24.5, 25.7, 27.7, 48.3, 64.2, 72.1, 98.1, 129.2,
129.2, 130.2, 139.9; (anti stereoisomer, CD3OD) 23.2, 24.2,
26.1, 47.5, 57.5, 70.9, 95.7, 128.1, 129.1, 129.5 139.9.
31), 204 (Mϩ Ϫ 91, 100%), 181 (Mϩ Ϫ 114) (Found: C, 81.2; H,
8.8; N, 4.6. C20H25NO requires C, 81.3; H, 8.6; N, 4.7%).
syn-N-(1Ј-Phenylheptyl)-2-(hydroxymethyl)piperidine 6a. Oil
purified by flash chromatography (AcOEt–petroleum ether–
Et3N = 2 : 8 : 1), Rf = 0.44; bp 146–148 ЊC/1 mmHg; νmax 3378
cmϪ1; δH 0.84 (3H, t, J = 7.0 Hz, CH3), 1.02 (2H, m, CH2), 1.24
(8H, m, 4 × CH2), 1.65 (6H, m, 3 × CH2), 2.47 (1H, br s, OH),
2.73 (3H, m, 3 × CHN), 3.48 (1H, dd, J = 10.5, 5.1 Hz, CHO),
3.65 (1H, dd, J = 10.5, 7.1 Hz, CHO), 3.96 (1H, dd, J = 10.4, 3.7
Hz, PhCHN), 7.28 (5H, m, ArH); δC 14.0, 21.9, 22.4, 22.6, 24.3,
26.7, 29.5, 29.9, 31.7, 42.6, 55.3, 60.3, 62.7, 126.9, 128.2, 128.5,
142.7; m/z 258 (Mϩ Ϫ 31), 204 (Mϩ Ϫ 85), 91 (Mϩ Ϫ 198, 100%)
(Found: C, 78.5; H, 11.0; N, 4.6. C19H31NO requires C, 78.8; H,
10.8; N, 4.8%).
Reductive cleavage of the oxazolidine 1 and reaction with
electrophiles—general procedure
Li metal (5–10 equiv. of a 30% wt. dispersion in mineral oil) was
placed under Ar in a 100 ml two-necked flask equipped with
reflux condenser and magnetic stirrer, washed with THF (3 × 10
ml), and suspended in THF (30 ml). Alternatively, K dispersion
was prepared in a similar apparatus by vigorous stirring of the
freshly cut metal (5 equiv.) in THF (30 ml) at reflux temperature
for 10 min; the unstirred metal suspension was then allowed to
cool to rt. A catalytic amount of naphthalene (10 mol%) was
added to the suspension of the metal, and the mixture was
stirred until a dark green colour appeared. The mixture was
chilled to the reported temperature (Tables) and a solution of
the substrate (5 mmol) in THF (2 ml) was added dropwise.
After stirring of this mixture for the reported time (Tables),
a solution of the appropriate electrophile (1.1 equiv.) in THF
(5 ml) was slowly added to the dark red heterogeneous reaction
mixture. After stirring for 15–60 min, the mixture was quenched
by slow dropwise addition of water (10 ml, CAUTION:), the
cold bath was removed, and the resulting mixture was extracted
with Et2O (3 × 20 ml). The organic phase was washed with
brine (10 ml), dried (K2CO3), and evaporated.
syn-N-(1-Phenylpentyl)-2-(hydroxymethyl)piperidine 7a. Oil;
purified by flash chromatography (AcOEt–petroleum ether–
Et3N = 3 : 7 : 1), Rf = 0.49; bp 125–130 ЊC/1 mmHg; νmax 3420
cmϪ1; δH 0.83 (3H, t, J = 7.2 Hz, CH3), 1.04 (2H, m, CH2), 1.29
(4H, m, 2 × CH2), 1.67 (6H, m, 3 × CH2), 2.74 (3H, m, 3 ×
CHN), 2.95 (1H, br s), 3.49 (1H, dd, J = 10.2, 5.1 Hz, CHO),
3.65 (1H, dd, J = 10.2, 7.2 Hz, CHO), 3.96 (1H, dd, J = 10.2, 3.9
Hz, PhCHN), 7.28 (5H, m); δC 14.0, 21.9, 22.4, 22.9, 24.3, 29.0,
29.7, 42.6, 55.3, 60.4, 62.7, 126.9, 128.2, 128.6, 142.7; m/z 230
(Mϩ Ϫ 31), 204 (Mϩ Ϫ 57), 91 (Mϩ Ϫ 170, 100%) (Found: C,
77.9; H, 10.6; N, 5.2. C17H27NO requires C, 78.1; H, 10.4; N,
5.4%).
anti-N-(1-Phenylpentyl)-2-(hydroxymethyl)piperidine 7b. Oil;
purified by flash chromatography (AcOEt–petroleum ether–
Et3N = 3 : 7 : 1), Rf = 0.51; bp 125–130 ЊC/1 mmHg; νmax 3420
cmϪ1; δH 0.85 (3H, t, J = 6.9 Hz, CH3), 1.21 (6H, m, 3 × CH2),
1.51 (4H, m, 2 × CH2), 1.74 (3H, m), 2.11 (1H, m, CHN), 2.57
(1H, m, CHN), 2.93 (1H, m, CHN), 3.50 (1H, dd, J = 10.5, 3.9
Hz, CHO), 3.99 (1H, t, J = 6.6 Hz, PhCHN), 4.01 (1H, dd, J =
10.5, 4.5 Hz, CHO), 7.28 (5H, m, ArH); δC 13.9, 22.8, 22.9,
24.2, 27.4, 29.0, 32.9, 43.7, 56.0, 61.6, 61.7, 127.0, 127.9, 128.9,
139.3; m/z 230 (Mϩ Ϫ 31), 204 (Mϩ Ϫ 57), 91 (Mϩ Ϫ 170, 100%)
(Found: C, 77.9; H, 10.6; N, 5.2. C17H27NO requires C, 78.1; H,
10.4; N, 5.4%).
D2O quenching was performed by dropwise addition of 1 ml
(0.055 mol) of the electrophile dissolved in THF (5 ml) during
2 min. After stirring of this mixture for 15 min, water (10 ml,
CAUTION:) was slowly added dropwise to quench unchanged
Li metal, the cold bath was removed, and the mixture worked
up as described above.
Crude products were purified by flash chromatography
(AcOEt–petroleum ether–Et3N); compound 330 is already
364
J. Chem. Soc., Perkin Trans. 1, 2002, 360–365

View Article Online
Synthesis of 1,2,3,4-tetrahydroisoquinolines 8 and 9
Acknowledgements
Tetrahydroisoquinolines 8 and 9 were synthesized according to
a modification of a procedure described in ref. 19. The starting
amino alcohol (2.5 mmol) was dissolved in decalin (1 ml) in a 25
ml flask equipped with reflux condenser and magnetic stirrer
under dry Ar. To this solution were added NH4Cl (0.10 g, 1.83
mmol) and AlCl3 (0.64 g, 4.8 mmol), and the resulting mixture
was vigorously stirred at 185 ЊC. Two portions of AlCl3 (0.32 g,
2.4 mmol, and 0.64 g, 4.8 mmol) were added after 40 and 70
min, respectively, and the mixture was stirred at 185 ЊC for
another 15 h. After chilling of the mixture to 100 ЊC, PhCl was
added (4 ml) and the mixture was stirred at 185 ЊC for a few
minutes before chilling to 0 ЊC and quenching with ice-cold 1 M
HCl (10 ml). The organic phase was washed with ice-cold 1 M
HCl (2 × 20 ml). The collected aqueous phases were extracted
with Et2O (3 × 10 ml), then basified with solid NaOH. The
basic mixture was extracted with Et2O (3 × 20 ml), and these
extracts were dried K2CO3, filtered, and evaporated. Reaction
products were characterized as follows.
Financial support from Università di Sassari (ex 60% funds) and
Regione Autonoma della Sardegna is gratefully acknowledged.
References
1 U. Azzena, Trends Org. Chem., 1997, 6, 55.
2 U. Azzena, G. Melloni, L. Pisano and B. Sechi, Tetrahedron Lett.,
1994, 35, 6759.
3 U. Azzena, L. Pilo and E. Piras, Tetrahedron, 2000, 56, 3775.
4 U. Azzena, S. Demartis and G. Melloni, J. Org. Chem., 1996, 61,
4913.
5 U. Azzena and L. Pilo, Synthesis, 1999, 664.
6 U. Azzena, S. Demartis, L. Pilo and E. Piras, Tetrahedron, 2000, 56,
8375.
7 O. Stratmann, B. Kaiser, R. Fröhlich, O. Meyer and D. Hoppe,
Chem. Eur. J., 2001, 7, 423.
8 D. Hoppe and T. Hense, Angew. Chem., Int. Ed. Engl., 1997, 36, 2282.
9 P. Beak, A. Basu, D. J. Gallagher, Y. S. Park and S. Thayumanavan,
Acc. Chem. Res., 1996, 29, 552.
10 H. Ahlbrecht, J. Harbach, R. W. Hoffmann and T. Ruhland, Liebigs
Ann., 1995, 211.
11 R. W. Hoffmann, T. Rühl, F. Chemla and T. Zahneisen, Liebigs Ann.
Chem., 1992, 719.
12 R. W. Hoffmann, T. Rühl and J. Harbach, Liebigs Ann. Chem., 1992,
725.
13 C. Mück-Lichtenfeld and H. Ahlbrecht, Tetrahedron, 1999, 55, 2609.
14 C. Mück-Lichtenfeld and H. Ahlbrecht, Tetrahedron, 1996, 52,
10025.
15 S. Klein, I. Marek and J.-F. Normant, J. Org. Chem., 1994, 59, 2925.
16 T. Kato, S. Marumoto, T. Sato and I. Kuwajima, Synlett, 1990, 671.
17 U. Azzena, L. Pilo and E. Piras, Tetrahedron Lett., 2001, 42, 129.
18 S. Kasuga and T. Taguchi, Chem. Pharm. Bull., 1965, 13, 233.
19 W. L. Mendelson, A. J. Villani, L. A. Petka and C. B. Spainhour, Jr.,
J. Labelled Compd. Radiopharm., 1984, 21, 961.
20 W. L. Mendelson, C. B. Spainhour, Jr., S. S. Jones, B. L. Lam and
K. L. Wert, Tetrahedron Lett., 1980, 21, 1393.
1,3,4,6,11,11a-Hexahydro-2H-pyrido[1,2-b]isoquinoline
8.
From amino alcohol 3 (0.51 g, 2.5 mmol). The product was
obtained as an oil (99 mg, 0.53 mmol, 21%), which solidified
upon storage and was purified by flash chromatography
(AcOEt–petroleum ether–Et3N = 8 : 2 : 1), Rf = 0.53; bp 125 ЊC/
1 mmHg (lit.,31 160 ЊC/15 Torr); δH (Hs at C1, C2 and C3 super-
impose between δ 1.28 and 1.96) 1.37 (2H, m), 1.77 (4H, m),
2.12 (1H, m, H4), 2.27 (1H, m, H11a), 2.75 (2H, m, 2H at C11),
3.08 (1H, m, H4Ј), 3.39 (1H, d, J = 15.3 Hz, H6), 3.83 (1H, d,
J = 15.3 Hz, H6Ј), 7.12 (4H, m, ArH); δC (C1, C2 and C3
resonate between δC 24.2 and 33.6) 24.2 (t), 25.8 (t), 33.6 (t),
36.7 (t, C11), 56.1 (t, C4), 58.3 (d, C11a), 58.3 (t, C6), 125.5 (d,
aromatic), 125.9 (d, aromatic), 126.1 (d, aromatic), 128.0 (d,
aromatic), 133.9 (s, aromatic), 134.2 (s, aromatic).
21 H. Takahashi, Y. Chida, K. Higashiyama and H. Onishi, Chem.
Pharm. Bull., 1985, 33, 4662.
22 H. Takahashi, B. C. Hsieh and K. Higashiyama, Chem. Pharm.
Bull., 1990, 38, 2429.
23 M. Yamato, K. Hashigaki, S. Ishikawa and N. Qais, Tetrahedron
Lett., 1988, 29, 6949.
24 L. Neelakantan, J. Org. Chem., 1971, 36, 2256 see also G. Just,
P. Potvin, P. Uggowitzer and P. Bird, J. Org. Chem., 1983, 48, 2923.
25 M.-J. Wu and L. N. Pridgen, J. Org. Chem., 1991, 56, 1340 and
references therein.
26 H.-G. Schmalz, C. B. de Koning, D. Bernicke, S. Siegel and
A. Pfletschinger, Angew. Chem., Int. Ed., 1999, 38, 1620.
27 D. Hoppe, A. Carstens and T. Krämer, Angew. Chem., Int. Ed. Engl.,
1990, 29, 1424 see also refs. 8, 9 and 13.
28 For related examples, see M. Marsch, K. Harms, O. Zschange,
D. Hoppe and G. Boche, Angew. Chem., Int. Ed. Engl., 1991, 30, 321;
M. Schlosser and D. Limat, J. Am. Chem. Soc., 1995, 117, 12342.
29 E. D. Jemmins, J. Chandrasekhar and P. v. R. Schleyer, J. Am. Chem.
Soc., 1979, 101, 527.
30 B. Weber, J. Schwerdtfeger, R. Fröhlich, A. Göhrt and D. Hoppe,
Synthesis, 1999, 1915.
1,3,4,6,11,11a-Hexahydro-6-deuterio-2H-pyrido[1,2-b]-
isoquinoline (anti-9a : syn-9b ؍ 89 : 11). From a mixture of
amino alcohols 4a and 4b (94 : 6 mixture; 0.52 g, 2.5 mmol) The
product was obtained as an oil (0.1 g, 0.55 mmol, 22%), which
solidified upon storage and was purified by flash chrom-
atography (AcOEt–petroleum ether–Et3N = 8 : 2 : 1), Rf = 0.53;
bp 125 ЊC/1 mmHg; δH (Hs at C1, C2 and C3 superimpose
between δ 1.28 and 1.96) 1.37 (2H, m), 1.77 (4H, m), 2.12 (1H,
m, H4), 2.27 (1H, m, H11a), 2.75 (2H, m, 2H at C11), 3.08 (1H,
m, H4Ј), 3.35 (0.11H, br s, 9b, H6), 3.80 (0.89H, br s, 9a, H6Ј),
7.12 (4H, m, ArH); δC (C1, C2 and C3 resonate between δC 24.2
and 33.6) 24.2 (t), 25.9 (t), 33.6 (t), 36.7 (t, C11), 56.1 (t, C4),
57.9 (t, J = 19.5 Hz, C6), 58.3 (d, C11a), 125.5 (d, aromatic),
125.9 (d, aromatic), 126.1 (d, aromatic), 128.0 (s, aromatic),
134.0 (s, aromatic), 134.2 (s, aromatic); m/z 188 (M ϩ), 187 (Mϩ
ؒ
Ϫ 1, 100%), 105 (Mϩ Ϫ 82) (Found: C, 82.8; H ϩ D, 9.9; N, 7.3.
C13H16DN requires C, 82.9; H ϩ D, 9.7; N, 7.4%).
31 v. J. Braun and W. Pinkernelle, Ber. Dtsch. Chem. Ges., 1931, 64,
1871.
J. Chem. Soc., Perkin Trans. 1, 2002, 360–365
365