Stereochemistry of the methoxide induced rearrangement of an α-bromophosphonamidate: Cleavage of the P-N and P-C bonds in the azaphosphiridine oxide intermediate

Article abstract of DOI:10.1039/a604093d

Menthyl P-(bromomethyl)-N-tert-butylphosphonamidate 5 has been prepared from (1R,2S,5R)-(-)-menthol and the S_P diastereoisomer has been isolated. This rearranges with methoxide [Me₃(PhCH₂)N⁺ MeO^- in THF-MeOH] to give only the S_P diastereoisomer of menthyl methyl (tert-butylamino)-methylphosphonate 6 and very largely (95%) the S_P diastereoisomer of menthyl methyl N-tert-butyl-N-methylphosphoramidate 7. The configurations of these products show that when the (postulated) azaphosphiridine oxide intermediate 16 suffers ring opening by methoxide, the P-N bond is cleaved (to give 6) with inversion of configuration but the P-C bond is cleaved (to give 7) with predominant retention. These contrasting stereochemistries suggest that nucleophilic attack on the P=O group of the azaphosphiridine oxide generates a five-coordinate intermediate (not merely a transition state) that exists long enough to undergo pseudorotation.

Full text of DOI:10.1039/a604093d

View Article Online / Journal Homepage / Table of Contents for this issue
Stereochemistry of the methoxide induced rearrangement of an
á-bromophosphonamidate: cleavage of the P᎐N and P᎐C bonds in
the azaphosphiridine oxide intermediate1
Martin J. P. Harger * and Ramesh Sreedharan-Menon
Department of Chemistry, The University, Leicester LE1 7RH, UK
M enthyl P -(bromomethyl)-N-tert-butylphosphonamidate 5 has been prepared from (1R,2S,5R)-(Ϫ)-
menthol and the SP diastereoisomer has been isolated. This rearranges with methoxide [M e3(PhCH 2)N +
M eOϪ in TH F –M eOH ] to give only the SP diastereoisomer of menthyl methyl (tert-butylamino)-
methylphosphonate 6 and very largely (95%) the SP diastereoisomer of menthyl methyl N-tert-butyl-N-
methylphosphoramidate 7. The configurations of these products show that when the (postulated)
azaphosphiridine oxide intermediate 16 suffers ring opening by methoxide, the P᎐N bond is cleaved (to
give 6) with inversion of configuration but the P᎐C bond is cleaved (to give 7) with predominant retention.
These contrasting stereochemistries suggest that nucleophilic attack on the P᎐O group of the
᎐
azaphosphiridine oxide generates a five-coordinate intermediate (not merely a transition state) that exists
long enough to undergo pseudorotation.
Two types of rearrangement product are formed when an α-
chlorophosphonamidate 1 reacts with methoxide, the α-amino-
phosphonate 3 and the phosphoramidate 4.2,3 In both types the
α-chlorine atom has been replaced, but not directly by meth-
oxide. Rather, it is the nitrogen atom of the amide group that has
displaced chlorine, and the phosphorus atom that has acquired
an extra methoxy group at the expense of the P᎐N or P᎐C
bond. Some substrates give predominantly just one type of
product but others (e.g. 1, R = H, RЈ = But) give substantial
amounts of both.2,3 The mechanism may involve initial base
induced cyclisation (elimination of HC1) to give the azaphos-
phiridine oxide 2. This could then break down, as a result of
nucleophilic attack at phosphorus, with cleavage of the P᎐N
bond to give 3 or the P᎐C bond to give 4. There is no direct
especially with regard to five-coordinate intermediates and
pseudorotation.7 A stereochemical study of the rearrangement
of an α-halogenophosphonamidate might therefore afford
valuable information, not only on that reaction in particular
but on nucleophilic substitution at phosphorus in general.
Results and discussion
Substrate
The menthyl α-bromomethylphosphonamidate 5 seemed a par-
ticularly appropriate substrate for a number of reasons. First,
it should give substantial amounts of both types of rearrange-
ment product, so that the stereochemistry of both P᎐N and
P᎐C cleavage can be examined in a single reaction. Second, it
and its rearrangement products (6 and 7) are chiral at phos-
O
O
R
Cl
R
P
OMe
P
OMe
N
NHR′
R′
O
P
O
1
2
Br
O
P
O
ButNH
O
O
P
NHBut
OMe
R
R
P
OMe
N
R′
OMe
5
6
R′NH
OMe
OMe
3
4
O
P
Me
But
evidence for an azaphosphiridine oxide intermediate, but some-
thing other than nucleophilic attack at phosphorus is clearly
rate-determining, since the reactivity of the substrate is notably
insensitive to steric hindrance.2,3 Also, the composition of the
product seems to depend on the relative abilities of the α-
carbon and nitrogen atoms to be displaced from phosphorus,
with anionic character, by an attacking nucleophile.3
N
O
OMe
7
phorus, as is necessary, but not at the α-C atom, which would
introduce needless complexity. Third, it contains a chiral ligand
(menthoxy) that, while playing no direct part in the chemistry,
opens the way to (i) obtaining stereoisomers of the substrate
that differ only in the configuration at phosphorus, by the sep-
aration of diastereoisomers rather than enantiomers; (ii) pro-
ducing stereoisomers of each type of product that differ only in
the configuration at phosphorus but which, being diastereo-
isomers, can be quantified relatively easily; (iii) deducing con-
figurations at phosphorus, relative to the known configuration
of the chiral ligand, by conventional X-ray crystallography.
Little is known about azaphosphiridine oxides, the phos-
phorus analogues of α-lactams. Other types of three-membered
ring containing a P᎐O group have been encountered as short-
᎐
lived intermediates,4 and some have actually been isolated.5
Even for these, however, there is no information on the stereo-
chemistry of nucleophilic substitution at the phosphorus atom.
By contrast, stereochemical studies with four- and five-
membered cyclic compounds6 have contributed much to our
understanding of nucleophilic substitution at phosphorus,
J. Chem. Soc., Perkin Trans. 1, 1997
527

View Article Online
Finally, it is a bromide rather than a chloride so it should react
under relatively mild conditions. This is especially important
because the P-methoxy groups in the products render them
liable to degradation, either structurally by demethylation [
O
O
BrCH2
ButNH
P
P(O)OMe + ϪOMe
P(O)OϪ + Me2O] or configuration-
P(O)OMeЈ +
5
ally by exchange [ P(O)OMe + ϪOMeЈ
ϪOMe]. The latter process would obviously compromise an
investigation of the stereochemistry, but so would the former if
one diastereoisomer were degraded more rapidly than the other.
Bromomethylphosphonic dibromide 8 was first prepared, by
Fig. 1
ried out using a small excess of Me3(PhCH2)N+ ϪOMe (1.5 mol
equiv.) as a dilute solution (0.2 mol dmϪ3 initial concentration)
in THF–methanol (9:1).
+
Ϫ
controlled hydrolysis of the complex (BrCH2PBr3 AlBr4
)
formed by reaction of CH2Br2 with PBr3–AlBr3.8 Our intention
had been to convert this into the phosphonamidic bromide 9,
but we were unable to selectively replace just one of the Br
atoms with ButNH2: even before the dibromide 8 (δP 1.0) had all
been consumed, some of the phosphonamidic bromide (δP 19.2)
underwent further reaction to the phosphonic diamide 11 (δP
15.8). This would not be expected if substitution proceeds by
the normal associative SN 2(P) mechanism; more likely, the high
reactivity of 9 is a consequence of its ability to react by a dis-
For the single diastereoisomer of 5, reaction was complete
inside 4.5 h at room temperature and gave two products in a 5:1
ratio (31P NMR). The major product was isolated by extraction
(from diethyl ether solution) into aqueous acid and was identi-
fied as the α-aminophosphonate 6, δp 28.4. The 1H NMR spec-
trum clearly showed a methylene group with phosphorus-
coupled diastereotopic protons, δH (CDCl3) 2.922 (ABP, δA
2.941, δB 2.903, JAB 13.9, JAP 14.9, JBP 15.7; CH2P), in addition
to the expected signals for P–O-menthyl (including δH 4.273, 1
H, dddd, JPH 6.9, JH H ~11, 10.7 and 4.5), P–OMe (δH 3.765, 3
H, d, JPH 10.8) and N–But (δH 1.076, 9 H, s) groups. Although
predominantly one diastereoisomer, minor peaks in both the
31P (δP 28.1) and 1H (δP 3.790, d, JPH 10.7; POMe) NMR spectra
indicated 4–5% of the other diastereoisomer.
The minor product, which did not extract into aqueous acid,
was identified as the phosphoramidate 7, δP 11.5; the 1H NMR
spectrum showed a phosphorus-coupled N–Me group at δH
2.663 (d, JPH 9.6) in addition to the P–O-menthyl (δH includes
4.137, 1 H, dddd, JPH 7.5, JH H ~11, 10.6 and 4.5), P–OMe (δH
3.614, 3 H, d, JPH 11.4) and N–But (δH 1.314, 9 H, s) signals.
This also contained ca. 5% of a second diastereoisomer; δP 10.5;
δH 3.628 (d, JPH 11.3, POMe).
sociative elimination–addition mechanism with
a three-
coordinate PV intermediate 10 (Scheme 1).9 Previously, selective
O
P
O
P
Br
O
ButNH2
Br
ButNH2
(–HBr)
Br
Br
Br
P
NBut
NHBut
Br
8
9
10
ButNH2
R*OH–Et3N
O
O
O
P
Br
Br
Br
ButNH2
P
OR*
P
OR*
NHBut
NHBut
NHBut
Using the 60:40 mixture of the diastereoisomers of 5, the
same two products were again formed in a 5:1 ratio, showing
that the two diastereoisomers of the substrate react in essen-
tially the same way. However, the diastereoisomer ratios of the
products—43:57 for the α-aminophosphonate 6 and 41:59 for
the phosphoramidate 7—differed greatly from those seen before.
This difference is important: it shows that stereospecificity (a
consequence of the mechanism of reaction), not stereoselectiv-
ity (a consequence of the relative stabilities of stereoisomeric
products), is chiefly responsible for the stereochemistry seen
using the single diastereoisomer of the substrate.
Br
12
5
11
R* = menthyl
Scheme 1
replacement of just one of the Cl atoms in ClCH2P(O)Cl2 with
ButNH2 did not present a problem.3 However, a better leaving
group (Br instead of Cl) may accelerate the dissociative second
substitution more than the associative first one, so that it does
become competitive.
In these experiments, the diastereoisomer ratios of the pro-
ducts are close to that of the substrate, but not identical. The
significance of the discrepancies was investigated by repeating
the reaction of the single diastereoisomer using a longer reac-
tion time (6 h) and a much shorter one (quenched after 16 min;
ca. 93% completion). The α-aminophosphonate product 6 from
these reactions contained 12% and 1% respectively of the minor
diastereoisomer, showing that it is prone to epimerisation (P–
OMe exchange) under the conditions of the reaction, and sug-
gesting that it is actually formed as just one diastereoisomer
(100% stereospecificity). The phosphoramidate product 7, on
the other hand, contained 5% of the minor diastereoisomer
regardless of the reaction time, implying that it is configura-
tionally stable. Since epimerisation requires attack at phos-
phorus by methoxide, this is not surprising; on both steric and
electronic grounds, the phosphoramidate 7 will be less suscep-
tible to nucleophilic attack than the α-aminophosphonate 6.
Given its configurational stability, it must be that the phos-
phoramidate product is not formed entirely as one diastereo-
isomer (<100% stereospecificity).
The alternative approach requires that the menthoxy group
be introduced first (Scheme 1). The phosphonic dibromide was
therefore treated with (Ϫ)-menthol and Et3N (8
12), then
with ButNH2. This gave the phosphonamidate 5 as a 54:46
mixture of diastereoisomers (31P NMR and capillary GLC),
together with some by-products. Crystallisation removed the
impurities and afforded a small sample of a single diastereo-
isomer, δP 18.0 (diastereoisomer ratio 1:99 at least) as well as a
60:40 mixture of both, δp 18.3 (major) and 18.0. The EI mass
spectrum of 5 did not show the molecular ion but the CI spec-
trum contained (M + H)+ peaks at 368 and 370 (5%; ratio ca.
1:1) as expected. Single-crystal X-ray analysis of the single
diastereoisomer of 5 revealed that it had the S configuration at
phosphorus (Fig. 1)1 relative to the known configuration of the
menthyl group (1R,2S,5R).10
Reaction with methoxide
Preliminary experiments indicated that the α-bromomethyl-
phosphonamidate 5, in spite of its good leaving group, reacted
with sodium methoxide in methanol only under quite forcing
conditions (2 mol dmϪ3 NaOMe; 50 ЊC) and that product de-
gradation (demethylation) was then quite extensive. Reaction
proceeded much more readily with benzyltrimethylammonium
methoxide in THF–methanol and demethylation of the pro-
ducts was not now a problem. All reactions were therefore car-
Product stereostructures
As anticipated, the products from the rearrangement of 5 (sin-
gle diastereoisomer) were non-crystalline, so it was not possible
to determine directly their configurations. For the α-aminophos-
phonate 6 the problem was readily solved, by treatment with
528
J. Chem. Soc., Perkin Trans. 1, 1997

View Article Online
NO2
O
O
MeO
(a)
P
+
O
O
MeO
–S
ButNH2CH2
+
NH2
P
O2N
NO2
2
O–
Fig. 2
(b)
O
O
MeO
P
picric acid. Crystallisation of the resulting salt afforded a pure
sample of the dominant diastereoisomer and the configuration
was determined by X-ray crystallography (Fig. 2).1
ButN
R
The phosphoramidate product 7 is a tertiary amide and there
is no simple way of converting it into a crystalline derivative.
That being so, the related secondary amide 13 was prepared, by
treating POCl3 first with (Ϫ)-menthol–Et3N,11 then with
ButNH2 and finally with NaOMe in methanol. Repeated chro-
matography of the resulting mixture of the diastereoisomers of
13 (δP 7.2 and 6.7) eventually yielded a small sample of the
lowfield diastereoisomer (δP 7.2; >96%). This was crystalline
and on N-methylation (NaH in DMF; MeI) it gave the same
diastereosiomer of the tertiary phosphoramidate 7 (lowfield δP)
as was dominant in the product from the rearrangement of 5
(single diastereoisomer). Since N-methylation does not involve
Fig. 3
O–
OR
OR
OR
O
–
O
OR
OR
P
P
ButN
P
S
ButN
+
–S
+
OR
C
S
ButN
C
S
S
C
S
Scheme 3
Reaction stereochemistry
Comparing the stereostructures of the α-bromophosphon-
amidate (Fig. 1) and its rearrangement products [Figs.
2 and 3(b)] it can be seen that methoxide cleaves the P᎐N
bond, to form the α-aminophosphonate, with inversion of
configuration at phosphorus, but the P᎐C bond, to form the
phosphoramidate, with retention. These contrasting stereo-
chemistries can be rationalised in terms of nucleophilic attack
on the azaphosphiridine oxide intermediate 16 (Scheme 4)† if
the P᎐O centre, the tertiary amide 7 and the secondary amide 13
᎐
must have the same configuration at phosphorus. Determin-
ation of the configuration of 13 will therefore reveal the con-
figuration of the rearrangement product 7. Unfortunately, des-
pite its crystallinity, we were unable to obtain crystals of 13 that
were adequate for X-ray analysis. Derivatisation of 13 thus
became a necessity. To avoid any risk of compromising the con-
figuration at phosphorus, modification of the NHBut group
was to be preferred. The derivatives 14a–c were prepared (NaH
OR*
O
OMe
OR*
O
Br
CH2
P
CH2
ButNH
P
MeO
P
OR*
O
NBut
NHBut
CH3
O
O
(S )–5
p
(S )–6
p
(S )–7
p
H
i, ii
R
N
P
O
N
P
O
But
But
–HBr
P
N
P
C
OMe
OMe
13
OMe
OR*
O–
OR*
O
MeO
H2C
MeO
NC
CH2
CH2
14 a, R =
b, R =
–OMe
ψ
H2C
P
H2C
P
P
OR*
NBut
17b
O–
NBut
NBut
16
i, iii
17a
O
O
O
N
CH2
c, R =
HS
P
O
O–
OMe
R*O
O
R*O
H2C
P
C
OMe
P
P
OMe
NBut
CH3 NBut
15
R* = menthyl
18
(R )–7
p
Scheme 2 Reagents: i, NaH in DMF; ii, RCl or RBr; iii, CS2 then
H3O+
Scheme 4
a five-coordinate intermediate is involved.7 Attack opposite the
more apicophilic ring-nitrogen atom would give the phospho-
rane 17a. This could collapse directly to product, by cleavage of
the apical P᎐N bond, giving the α-aminophosphonate (SP)-6
with inversion of configuration. Alternatively, it could pseudo-
rotate to the new phosphorane 17b. This, on cleavage of the
newly-apical P᎐C bond, would give the phosphoramidate (SP)-7
with retention of configuration. Granted that nitrogen is not
strongly apicophilic, the possibility of some competing attack
opposite carbon must also be considered. Direct breakdown of
the resulting phosphorane 18, by cleavage of the P᎐C bond,
in DMF; RCl or RBr) but unfortunately none was crystalline.
Replacement of one of the groups attached to the P atom thus
seemed unavoidable. The secondary amide 13 (>96% δP 7.2)
was again converted into the anion (NaH in DMF), but
now this was treated with CS2 to give, after acidification, the
phosphorothioic acid 15 (>96% one diastereoisomer). With
dicyclohexylamine the acid formed a crystalline salt and X-ray
crystallography was used to determine its stereostructure [Fig.
3(a)].1 Given the mechanism of the reaction (Scheme 3), no
change in the configuration at phosphorus would be expected
when the ButNH group is replaced by HS. More important,
several examples of this type of transformation have been
shown to proceed with retention of configuration at phos-
phorus.12 We can therefore deduce the configuration of the
amide 13 with confidence [Fig. 3(b), R = H], and hence the con-
figuration of the phosphoramidate 7 (dominant diastereo-
isomer) formed in the rearrangement of the α-bromophos-
phonamidate [Fig. 3(b), R = Me].
† Scheme 4 shows the behaviour of the pure SP diastereoisomer of 5;
the RP diastereoisomer was not obtained pure, but the results for the
diastereoisomer mixture (60% RP) suggest that it reacts with compar-
able stereospecificity, i.e. (RP)-5 gives predominantly (RP)-6 (inversion
of configuration) and (RP)-7 (retention) (Scheme 4 with OR* and O
interchanged).
J. Chem. Soc., Perkin Trans. 1, 1997
529

View Article Online
would in this case give the phosphoramidate (RP)-7 with inver-
sion of configuration at phosphorus. The fact that the phos-
phoramidate was actually formed as a 95:5 mixture of diaster-
eoisomers is therefore easily accommodated and might even
have been anticipated. Of course, substitution with inversion of
configuration does not actually require a phosphorane inter-
mediate—a five-coordinate transition state will suffice—but
substitution with retention of configuration surely does. And if
some of the nucleophilic attack on the azaphosphiridine oxide
proceeds via a phosphorane intermediate, the likelihood must
be that all of it does. In one sense this may seem improbable: the
leaving group (nitrogen or carbon) is part of a three-membered
ring and its departure will afford much relief of strain. Even so,
it seems a concerted pathway like that for nucleophilic substitu-
tion (SN 2) at a tetrahedral carbon is still not as favourable as the
normal non-concerted addition–elimination mechanism of
PBr3 and AlBr3 was dissolved in CH2Cl2 and carefully hydro-
lysed at ca. Ϫ25 ЊC to give bromomethylphosphonic dibromide
8; δP(CH2Cl2) 0.5; δH (CDCl3) 4.13 (d, JPH 4).
A mixture of dried (Ϫ)-menthol (2.55 g, 16.3 mmol) and
Et3N (1.58 g, 15.6 mmol) in CH2Cl2 (15 ml) was added over 10
min to a stirred (powerful magnet) solution of bromomethyl-
phosphonic dibromide 8 (4.68 g, 15.6 mmol) in CH2Cl2 (15 ml).
After 20 min more (Ϫ)-menthol–Et3N (0.8 mmol) was added to
complete the formation of 12 (two diastereoisomers, δP 19.0
and 17.0). Stirring was continued for a further 30 min, then tert-
butylamine (4.66 g, 63.9 mmol) was added dropwise (exo-
thermic) [to give product 5, two diastereoisomers, δP 18.4
(major) and 18.1, ratio 54:46; several by-products]. After 20
min, all volatile material was evaporated and the residue was
partitioned between toluene and water. The organic portion
was dried (MgSO4) and concentrated; crystallisation from
EtOH–H2O (9:1; 20 ml) gave 5 (1.53 g, 27%) as a 46:54 mixture
of diastereoisomers (highfield δP now in excess). Recrystal-
lisation from EtOH–H2O (12:1; 10 ml; overnight refriger-
ation) gave some material (374 mg) having a diastereoisomer
ratio 3:97 and this, on crystallisation from light petroleum
(20 ml), gave the pure highfield (31P NMR) diastereosiomer
of menthyl P-(bromomethyl )-N-tert-butylphosphonamidate 5
(sample A), mp 155–155.5 ЊC; δP(CH2Cl2) 18.0; δH (CDCl3,
300 MHz) 4.225 (1 H, ddt, JPH 8.0, JH H 4.5 and 10.6), 3.278
(2 H, ABP, δA 3.341, δB 3.211, JAB 13.0, JAP 10.3, JBP 7.1,
PCH2Br), 2.646 (1 H, d, JPH 6.9, NH; exchanged with D2O),
2.287 (1 H, m), 2.077 (1 H, d septet, JH H 2.5 and 7.0), 1.663
(2 H, m), 1.55–0.75 (14 H) and 1.355 (9 H, d, JPH 0.6,
PNBut); m/z (no M+) 354, 352 (M+ Ϫ Me, 7%; ratio 1:1),
232, 230 (33; ratio 1:1) and 216, 214 (M+ Ϫ Me Ϫ C10H18,
100; ratio 1:1); m/z (CI) 370, 368 (M + H+, 5%; ratio 1:1),
290 (30) and 74 (100); νmax(Nujol)/cmϪ1 3220 (NH) and 1245
nucleophilic substitution at phosphorus. In fact, when the P᎐O
᎐
group is part of a small ring there can in any case be consider-
able relief of strain: the ideal bond angle at phosphorus in the
tetrahedral substrate is 109.5Њ but in a trigonal-bipyramidal
phosphorane intermediate it will be only 90Њ, between apical
and equatorial positions. Provided nucleophilic attack on the
P᎐O group of 16 occurs opposite one of the ring atoms (N
᎐
or C), the phosphorane will be formed with the ring apical,
equatorial and angle strain will be much reduced.
Angle strain may also be important in determining the struc-
tures of the products, as well as their configurations. The men-
thoxy group (OR*) is inherently the best leaving group in the
azaphosphiridine oxide 16, yet when methoxide attacks, the
P᎐O bond remains intact while the P᎐N or P᎐C bond breaks.
Direct displacement of the menthoxy group is unlikely, since
nucleophilic attack opposite the OR* group of 16 would place
the three-membered ring diequatorial in the resulting inter-
mediate (or transition state), but the menthoxy group occupies
an apical position in the phosphorane 17b and might be
expected to depart. If it were to do so, however, the angle strain
(P᎐O); t 14.7 min (BP 5, 210 ЊC) (Found: C, 48.85; H, 8.1;
᎐
R
N, 3.8. C15H31BrNO2P requires C, 48.9; H, 8.5; N, 3.8%). A
portion of this diastereoisomer (100% by GLC) was recrystal-
lised from CH2Cl2–light petroleum and then slowly from
aqueous ethanol to produce a sample for single crystal X-ray
analysis.1
relieved in the formation of the phosphorane (16
17) would
be reimposed in the formation of the product (16 with OMe in
place of OR*). By contrast, there will be a further easing of
angle strain when 17b forms the phosphoramidate product by
cleavage of the P᎐C bond.
Material from the mother liquors (see above) was crystallised
from light petroleum to give a sample of the phosphonamidate 5
as a 60:40 mixture of diastereoisomers (sample B), mp 113–
115 ЊC; δP(CH2Cl2) 18.3 (major) and 18.0; δH (CDCl3, 300 MHz)
4.34–4.17 (1 H, m), 3.39–3.17 (2 H, m), 2.731 (major) and 2.650
(total 1 H; both d, JPH 7.2 or 7.5, NH), 2.34–2.03 (2 H, m),
1.72–1.62 (2 H, m), 1.55–0.75 (14 H) and 1.355 and 1.351
(major) (total 9 H; both d, JPH 0.7); νmax(Nujol)/cmϪ1 3300,
3260 (NH) and 1235 (P᎐O); t 14.3 (major) and 14.7 min
Experimental
Mps were determined using a Kofler hot-stage apparatus and
are uncorrected. 1H NMR spectra were recorded at 90 MHz on
a Varian EM 390 spectrometer or (where indicated) at 300 MHz
on a Bruker AM-300 (Me4Si internal standard; coupling con-
stants, J, given in Hz) and 31P NMR spectra (1H decoupled)
were recorded at 36.2 MHz on a JEOL JNM-FX90Q spec-
trometer (positive chemical shifts downfield from external 85%
H3PO4). Routine mass spectra were obtained in EI or (where
indicated) CI mode on a VG 16-B or Kratos Concept spec-
trometer and high resolution spectra were recorded by the
SERC Mass Spectrometry Service at Swansea. GLC analyses
were performed using a Philips capillary chromatograph
(helium carrier gas; flame-ionisation detector) fitted with a 25
m × 0.22 mm column containing a 0.25 µm film of BP 5 (equiva-
lent to SE 54) and TLC analyses were performed on silica gel 60
F254 (0.2 mm layer on aluminium foil). Amines were dried over
KOH, CH2Cl2 was distilled from CaH2, THF was distilled from
sodium–benzophenone and DMF was distilled under reduced
pressure from P2O5. Solutions of benzyltrimethylammonium
methoxide in MeOH (40% w/w) were used as supplied. Light
petroleum refers to the fraction with bp 60–80 ЊC unless other-
wise indicated, and ether to diethyl ether.
᎐
R
(BP 5, 210 ЊC).
Reaction of menthyl P-(bromomethyl)-N-tert-butylphosphon-
amidate 5 with methoxide
A solution of the diastereoisomerically enriched bromomethyl-
phosphonamidate 5 (sample A or B) (48 mg, 0.13 mmol) in
THF (0.45 ml) was added to a mixture of methanolic benzyl-
trimethylammonium methoxide (40% w/w) (0.1 ml, 0.2 mmol)
and THF (0.45 ml) (reaction medium: 0.2 mol dmϪ3 methoxide
in 9:1 THF–MeOH). After 4.5 h (or 16 min or 6 h) the reaction
was quenched (NH4Cl) and the reaction mixture was examined
by 31P NMR spectroscopy to determine the ratio of the two
types of rearrangement product and the diastereoisomer com-
position of each. Volatile material was removed, the residue was
partitioned between ether and water and the ether portion was
washed with aqueous NaOH and water. To separate the pro-
ducts the ether portion was then extracted with hydrochloric
acid (1 mol dmϪ3) and water. Concentration of the ether
portion afforded menthyl methyl N-tert-butyl-N-methylphos-
phoramidate 7 (see below) and basification of the combined
aqueous extracts liberated menthyl methyl (tert-butylamino)-
methylphosphonate 6 (see below).
Menthyl P-(bromomethyl)-N-tert-butylphosphonamidate 5
Based on the method of Cade,8 the complex (δP 30.2) formed by
addition of CH2Br2 (7 mol equiv.) to an equimolar mixture of
530
J. Chem. Soc., Perkin Trans. 1, 1997

View Article Online
Menthyl methyl N-tert-butyl-N-methylphosphoramidate 7
5), 3.692 (3 H, d, JPH 11.3, OMe) 3.360 (2 H, ABP, δA 3.391, δB
The reaction of the bromomethylphosphonamidate 5 (sample
B) with methoxide gave the phosphoramidate 7 as a 41:59 mix-
ture of diastereoisomers, bp 74 ЊC (oven temp.) at 0.2 mmHg;
δP(CDCl3) 11.5 and 10.5 (major); δH (CDCl3, 300 MHz) 4.20–
4.08 (1 H, m), 3.628 (major) and 3.612 (total 3 H; both d, JPH
11.3 or 11.4, OMe), 2.665 (major) and 2.661 (total 3 H; both d,
JPH 9.7 or 9.5, NMe), 2.34–2.10 (2 H, m), 1.71–1.58 (2 H, m),
1.54–0.74 (14 H) and 1.315 (9 H, s); m/z (CI) 320 (M + H+,
3.329, JAB ~JAP ~JBP ~15), 2.13–2.02 (1 H, m), 1.916 (1 H, m),
1.72–1.58 (2 H, m), 1.30–0.70 (14 H) and 1.470 (9 H, s). A
portion of this diastereoisomerically pure picrate was allowed
to crystallise slowly from ether–light petroleum to give a sample
suitable for single crystal X-ray analysis.1
Menthyl methyl N-tert-butylphosphoramidate 13
(a) A mixture of (Ϫ)-menthol (3.13 g, 20 mmol) and Et3N (2.02
g, 20 mmol) in light petroleum (25 ml) was added over 20 min to
a cooled (ice bath) and stirred solution of phosphorus oxychlo-
ride (3.07 g, 20 mmol) in light petroleum (15 ml). The mixture
was stirred at room temperature for 3.5 h and the precipitate
(Et3NHCl) was removed by filtration (moisture excluded). The
filtrate was concentrated to give menthyl phosphorodichlori-
date (5.06 g, 93%);11 δP(CH2Cl2) 5.5; νmax(film)/cmϪ1 1290
48%), 264 (M + H+ Ϫ H C᎐CMe , 13), 182 (M + H+ Ϫ C H ,
᎐
2
2
10 18
100), 166 (M+ Ϫ Me Ϫ C10H18, 59) and 126 (M + H+ Ϫ C10H18
Ϫ H C᎐CMe , 68); νmax(film)/cmϪ1 1250 (P᎐O) (Found: C, 59.4;
᎐
᎐
2
2
H, 10.4; N, 4.4; M + H+, 320.2355. C16H34NO3P requires C,
60.2; H, 10.7; N, 4.4%; M + H, 320.2355).
The reaction of 5 (sample A) with methoxide gave the phos-
phoramidate
7 as a 95:5 mixture of diastereoisomers;
δP(CDCl3) 11.5 (major) and 10.5; δH (CDCl3, 300 MHz) for
major component: 4.137 (1 H, dddd, JPH 7.5, JH H ~11, 10.6,
4.5), 3.614 (3 H, d, JPH 11.4, OMe), 2.663 (3 H, d, JPH 9.6,
NMe), 2.31–2.11 (2 H, m), 1.70–1.60 (2 H, m), 1.53–0.79 (14 H,
(P᎐O), 990 and 970. (b) tert-Butylamine (2.66 g, 36.4 mmol) in
᎐
benzene (5 ml) was added dropwise over 5 min to a cooled (ice
bath) and stirred solution of the crude phosphorodichloridate
(4.97 g, 18.2 mmol) in benzene (20 ml). After 10 min the mix-
ture was warmed to room temperature and stirred for a further
2 h. The precipitate (Et3NHCl) was removed by filtration and
the filtrate was concentrated to give menthyl N-tert-butylphos-
phorochloramidate (5.35 g, 95%) as an oil that slowly solidified;
δP(CDCl3) 10.5 (major) and 9.8 (diastereoisomers; ratio 65:35)
(impurity ~10%); δH (CDCl3) 4.63–4.13 (1 H, m), 3.23 (1 H, br,
NH), 2.43–0.70 (18 H) and 1.33 (9 H, s). This material was used
without purification. (c) A solution of the impure phosphoro-
chloramidate (4.3 g, 14 mmol) in light petroleum (bp 40–60 ЊC)
(8 ml) was stirred and cooled (ice bath) while a 2 mol dmϪ3
solution of NaOMe in MeOH (14 ml; 2 mol equiv.) was added.
After 10 min the remaining methoxide was quenched (NH4Cl)
and volatile material was removed. The residue was partitioned
between light petroleum and water and the organic portion was
dried (MgSO4) and concentrated to give the crude product (3.6
g, 85%); δP(CH2Cl2) 7.2 (major) and 6.7 (diastereoisomers; ratio
60:40). Crystallisation from light petroleum (bp 40–60 ЊC) at
Ϫ40 ЊC gave menthyl methyl N-tert-butylphosphoramidate 13,
mp 73–76 ЊC (diastereoisomer ratio now 53:47, lowfield δP still
in excess); δH (CDCl3, 300 MHz) 4.22–4.09 (1 H, m), 3.683 and
3.670 (major) (total 3 H; both d, JPH 11.3 or 11.4, OMe), 2.483
and 2.436 (major) (total 1 H; both d, JH H 7.3 or 7.1, NH;
exchanges with D2O), 2.39–2.08 (2 H, m), 1.72–1.59 (2 H, m),
1.54–0.77 (14 H) and 1.275 (major) and 1.271 (total 9 H; both
d, JPH 0.7 or 0.8); m/z 305 (M+, 1%), 290 (M+ Ϫ Me, 8), 168
(M+ Ϫ C10H17, 50), 152 (M+ Ϫ Me Ϫ C10H18, 100) and 112
m) and 1.314 (9 H, s); νmax(film)/cmϪ1 1250 (P᎐O).
᎐
Authentic samples of the phosphoramidate 7 were also pre-
pared, by methylation of menthyl methyl N-tert-butylphos-
phoramidate 13. Thus 13 (mixture of diastereoisomers) (98 mg,
0.32 mmol) was stirred with NaH (18 mg, 0.75 mmol) in DMF
(0.75 ml) for 2.7 h. Methyl iodide (136 mg, 0.96 mmol) was then
added. After 1 h, volatile material was removed and the residue
was partitioned between light petroleum and water. The organic
portion was dried (MgSO4) and concentrated to give the phos-
phoramidate 7 as a 58:42 mixture of diastereoisomers with the
lowfield (31P NMR) diastereoisomer in excess. In like manner,
the sample of 13 (>96% one diastereoisomer) recovered
unchanged from its reaction with NaH–CS2 (see below) gave
the phosphoramidate 7 as a 98:2 mixture of diastereoisomers
with the lowfield (31P NMR) diastereoisomer dominant;
δH (CDCl3) 3.628 and 3.612 (major) (3 H, d, JPH 11.3 or 11.4,
OMe).
Menthyl methyl (tert-butylamino)methylphosphonate 6
The reaction of the bromomethylphosphonamidate 5 (sample
B) with methoxide gave the α-aminophosphonate 6 as a 43:57
mixture of diastereoisomers; δP(CDCl3) 28.4 and 28.1 (major);
δH (CDCl3, 300 MHz) 4.33–4.21 (1 H, m), 3.791 (major) and
3.765 (total 3 H; both d, JPH 10.6 or 10.8, OMe), 2.922 and
2.920 (major) (total 2 H; ABP pattern or d, JPH 15.2, PCH2N),
2.22–2.07 (2 H, m), 1.72–1.61 (2 H, m), 1.51–1.28 (2 H, m),
1.26–0.68 (13 H; includes NH) and 1.085 (major) and 1.078
(total 9 H; both s); m/z 319 (M+, 13%), 304 (M+ Ϫ Me, 39), 182
(M+ Ϫ C10H17 Ϫ H C᎐CMe , 60); νmax(Nujol)/cmϪ1 3220 (NH)
᎐
2
2
and 1240 (P᎐O) (Found: C, 58.9; H, 10.3; N, 4.4; M+, 305.2127.
᎐
(M+ Ϫ C10H17, 18) and 166 (M+ Ϫ Me Ϫ C10H18, 100); νmax
-
C15H32NO3P requires C, 59.0; H, 10.6; N, 4.6%; M, 305.2120).
Chromatography of a 60:40 mixture of the diastereoisomers
of 13 [rotating silica layer (chromatatron), eluent 1:1 ethyl
acetate–light petroleum (bp 40–60 ЊC)] followed by the repeated
chromatography of enriched fractions afforded a small sample
(3% yield) that was >96% the lowfield (31P NMR) diastereo-
isomer.
(film)/cmϪ1 3280 (NH) and 1245 (P᎐O). Treatment with picric
᎐
acid in benzene and crystallisation from CH2Cl2–ether–light
petroleum afforded the picrate, mp 158–163 ЊC; δP(CDCl3) 17.0
(major) and 16.2 (diastereoisomers); δH (CDCl3, 300 MHz)
8.886 (2 H, s), 4.37–4.16 (1 H, m), 3.676 (major) and 3.649
(total 3 H; both d, JPH 11.3 or 11.4, OMe), 3.50–3.30 (2 H, m),
2.12–1.84 (2 H, m), 1.70–0.69 (16 H) and 1.475 (9 H, s) (Found:
C, 48.3; H, 6.7; N, 10.1. C22H37N4O10P requires C, 48.2; H, 6.8;
N, 10.2%).
N-Alkyl derivatives of menthyl methyl N-tert-butylphosphor-
amidate 13
The reaction of 5 (sample A) with methoxide gave the α-
aminophosphonate 6 as a 95:5 mixture of diastereoisomers;
δP(CDCl3) 28.4 (major) and 28.1; δH (CDCl3, 300 MHz) major
diastereoisomer: 4.273 (1 H, dddd, JPH 6.9, JH H ~11, 10.7 and
4.5), 3.765 (3 H, d, JPH 10.8, OMe), 2.922 (2 H, ABP, δA 2.941,
δB 2.903, JAB 13.9, JAP 14.9, JBP 15.7, PCH2N), 2.26–2.06 (2 H,
m), 1.73–1.60 (2 H, m), 1.58–0.75 (15 H; includes NH) and
Although the phosphoramidate 13 was crystalline, all attempts
to grow crystals suitable for X-ray studies were unsuccessful.
Several derivatives were prepared as summarised below, but
none of them was obtained in a crystalline form.
(a) The phosphoramidate 13 (mixture of diastereoisomers)
(153 mg, 0.5 mmol) was stirred with NaH (31 mg, 1.3 mmol) in
DMF (1 ml). After 1.3 h, 4-methoxybenzyl chloride (235 mg,
1.5 mmol) was added and the mixture was stirred for a further
30 min. The excess NaH was quenched with methanol (60 µl) and
solvent was removed. The residue was partitioned between
ether and water and the organic portion was dried (Na2SO4)
and concentrated. Chromatography [silica layer; 1:3 EtOAc–
1.076 (9 H, s); νmax(film)/cmϪ1 3290 (NH) and 1245 (P᎐O).
᎐
Treatment with picric acid in benzene afforded, after crystal-
lisation from ether–light petroleum, a single diastereoisomer of
the picrate, mp 118.5–120.5 ЊC; δP(CDCl3) 16.4; δH (CDCl3, 300
MHz) 8.889 (2 H, s), 4.259 (1 H, dddd, JPH 6.8, JH H ~11, 11 and
J. Chem. Soc., Perkin Trans. 1, 1997
531

View Article Online
light petroleum (bp 40–60 ЊC); Rf 0.21] afforded menthyl methyl-
N-(4-methoxybenzyl )-N-tert-butylphosphoramidate 14a (171
mg, 79%) as a semi-solid (completely molten above 50 ЊC);
δP(CH2Cl2) 11.9 (major) and 10.8 (diastereoisomers; ratio
60:40); δH (CDCl3) 7.07 (4 H, AAЈBBЈ, δA 7.32, δB 6.82, JAB 9),
4.5–3.9 (1 H, m), 4.20 (2 H, br d, JPH 13, PNCH2Ar), 3.76 (3 H,
s, OMe), 3.63 (3 H, d, JPH 11, POMe), 2.5–0.8 (18 H) and 1.30
(major) and 1.28 (total 9 H; both s, NBut); m/z (CI) 426
(M + H+, 11%), 368 (10), 288 (M + H+ Ϫ C10H18, 17) and 230
(100). (CAUTION: An attempt to dry 4-methoxybenzyl chlor-
ide over molecular sieves for 24 h resulted in decomposition
with a build-up of pressure in the container; it was therefore
used without drying).
Dicyclohexylammonium O-menthyl O-methyl phosphorothioate
The phosphorothioic acid 15 (mixture of diasteroisomers) was
treated with dicyclohexylamine in light petroleum (bp 40–
60 ЊC). Crystallisation of the product from light petroleum (bp
40–60 ЊC) gave dicyclohexylammonium O-menthyl O-methyl
phosphorothioate as an equal mixture of diastereoisomers, mp
151–154 ЊC; δP(CDCl3) 56.7 and 56.0; δH (CDCl3, 300 MHz)
8.85 (2 H, br s, +NH2), 4.25–4.05 (1 H, m), 3.612 and 3.600
(total 3 H; both d, JPH 13.2 or 13.0, OMe), 3.12–2.93 (2 H, m)
and 2.47–0.77 (38 H); m/z [negative ion FAB (NOBA matrix)]
265 (100%); νmax(Nujol)/cmϪ1 3200–2140 (NH) (Found: C, 61.6;
H, 10.1; N, 3.0. C23H46NO3PS requires C, 61.7; H, 10.4; N,
3.1%). Similarly, the phosphorothioic acid 15 [>96% one dia-
stereoisomer, δP(CDCl3) 61.2] was converted into the dicyclo-
hexylammonium salt; crystallisation from light petroleum (bp
40–60 ЊC) afforded a sample that was у99.5% one diastereo-
isomer, δP(CDCl3) 55.9; δH (CDCl3, 300 MHz) 8.82 (v br, +NH2),
4.117 (1 H, dddd, JPH ~10, JH H ~10, 10 and 4.3), 3.612 (3 H, d,
JPH 13.3, OMe), 3.11–2.95 (2 H, m), 2.49–2.38 (1 H, m), 2.256 (1
H, m), 2.17–2.04 (4 H, m), 1.89–1.76 (3 H, m) and 1.75–0.79 (29
H). This sample was allowed to crystallise slowly from a mix-
ture of ether, CH2Cl2 and light petroleum to give a crystal suit-
able for X-ray analysis.1
(b) Similar alkylation of the phosphoramidate 13 with 4-
cyanobenzyl bromide gave, after chromatography, menthyl
methyl N-(4-cyanobenzyl )-N-tert-butylphosphoramidate 14b
(17%) as a glass; δP(CH2Cl2) 11.4 and 10.4 (diastereoisomers;
ratio 50:50); δH (CDCl3) 7.51 (4 H, br s), 4.42 and 4.40 (total 2
H; both d, JPH 12 or 13, PNCH2Ar), 4.12 (1 H, m), 3.66 (3 H, d,
JPH 11, POMe), 2.45–0.65 (18 H) and 1.29 and 1.27 (total 9 H;
both s); m/z (CI) 421 (M + H+, 80%), 405 (M+ Ϫ Me, 17), 365
(M + H+ Ϫ H C᎐CMe , 10), 306 (36), 283 (M + H+ Ϫ C H ,
᎐
2
2
10 18
100) and 267 (M+ Ϫ Me Ϫ C10H18, 75); νmax(film)/cmϪ1 2220
(C᎐N).
᎐
᎐
Acknowledgements
We thank the SERC for a research studentship (to R. S.-M.)
and for access to the Mass Spectrometry Service at Swansea.
(c) Similar alkylation of the phosphoramidate 13 with N-
(bromomethyl)phthalimide (reaction allowed to proceed over-
night) gave, after chromatography (silica layer; ethyl acetate; Rf
0.52), menthyl methyl N-tert-butyl-N-(phthalimidomethyl )-
phosphoramidate 14c (39%) as an oil; δP(CDCl3) 10.8 (major)
and 9.9 (diastereoisomers; ratio 63:37); δH (CDCl3) 7.78 (4 H,
m), 5.34–5.10 (2 H, m, NCH2N), 4.22 (1 H, m), 3.74 (3 H, d, JPH
12, POMe), 2.47–0.55 (18 H) and 1.42 (major) and 1.40 (total 9
H; both s); m/z (CI) 465 (M + H+, 65%), 449 (M+ Ϫ Me, 20),
References
1 Preliminary communication: J. Fawcett, M. J. P. Harger, D. R.
Russell and R. Sreedharan-Menon, J. Chem. Soc., Chem. Commun.,
1993, 1826 (Contains crystallographic structures and details of
X-ray measurements. Atomic coordinates, bond lengths and angles,
and thermal parameters have been deposited at the Cambridge
Crystallographic Data Centre).
2 M. J. P. Harger and A. Williams, J. Chem. Soc., Perkin Trans. 1,
1986, 1681. See also K. A. Petrov, V. A. Chauzov, T. S. Erokhina
and I. V. Pastukhova, J. Gen. Chem. USSR (Engl. Transl.), 1977, 47,
2501.
409 (M + H+ Ϫ H C᎐CMe , 8), 327 (M + H+ Ϫ C H , 69),
᎐
2
2
10 18
311 (M+ Ϫ Me Ϫ C10H18, 100) and 271 (M + H+ Ϫ H C᎐
᎐
2
CMe2 Ϫ C10H18, 55); νmax(film)/cmϪ1 1770 and 1710 (C᎐O).
᎐
O-Menthyl O-methyl phosphorothioate 15
The phosphoramidate 13 (mixture of diastereoisomers) (765
mg, 2.5 mmol) was stirred with NaH (103 mg, 4.3 mmol) in
DMF (4.6 ml). After 1.7 h, CS2 (0.76 g, 10 mmol) was added
(deep red colour) and the mixture was left overnight. The excess
NaH was quenched with MeOH (100 µl). Solvent was removed
and the residue was partitioned between water and ether. The
aqueous portion was acidified (HCl) to pH р 1 and the liber-
ated free phosphorothioic acid 15 was extracted into light
petroleum. Purification of the acid by crystallisation of its
ammonium salt from CH2Cl2–light petroleum (bp 40–60 ЊC)
failed to remove a by-product (15%; believed to be 15 with O
in place of S, resulting from reaction of the phosphoramidate
anion with CO2 instead of CS2). The triethylammonium salt of
15, δP(CH2Cl2) 58.3 and 58.0 (diastereoisomers) (by-product δP
0.3) was therefore prepared and partitioned between CH2Cl2
and water; the by-product passed into the aqueous phase allow-
ing the pure acid 15 (545 mg, 82%) to be liberated from the
organic phase. Crystallisation from light petroleum (bp 40–
60 ЊC) at Ϫ20 ЊC afforded O-menthyl O-methyl phosphorothio-
ate 15 as an equal mixture of diastereoisomers, mp 65–75 ЊC;
δP(CDCl3) 61.0 and 60.6; δH (CDCl3) 7.57 (1 H, br s, OH), 4.53–
4.05 (1 H, m), 3.73 (3 H, d, JPH 13, OMe) and 2.40–0.67 (18 H);
νmax(Nujol)/cmϪ1 3600–1800 (several maxima; OH) and 820
3 M. J. P. Harger and A. Williams, J. Chem. Soc., Perkin Trans. 1,
1989, 563.
4 See, for example, P. Burns, G. Capozzi and P. Haake, Tetrahedron
Lett., 1972, 925; A. J. Fry and L.-L. Chung, Tetrahedron Lett.,
1976, 645; K. A. Petrov, V. A. Chauzov, T. S. Erokhina and I. V.
Pastukhova, J. Gen. Chem. USSR (Engl. Transl.), 1976, 46, 2387.
5 H. Quast, Nachr. Chem. Tech. Lab., 1979, 27, 120; (Chem. Abstr.,
1979, 90, 187 010) (review); H. Quast, M. Heuschmann and
M. O. Abdel-Rahman, Leibigs Ann. Chem., 1981, 943; H. Quast and
M. Heuschmann, Leibigs Ann. Chem., 1981, 967; 1981, 977;
T. Oshikawa and M. Yamashita, Bull. Chem. Soc. Jpn., 1986, 59,
3293.
6 R. F. Hudson and C. Brown, Acc. Chem. Res., 1972, 5, 204;
C. R. Hall and T. D. Inch, Tetrahedron, 1980, 36, 2059.
7 G. R. J. Thatcher and R. Kluger, Adv. Phys. Org. Chem., 1989, 25,
99; A. Yliniemala, T. Uchimaru, K. Tanabe and K. Taira, J. Am.
Chem. Soc., 1993, 115, 3032; G. R. J. Thatcher and A. S. Campbell,
J. Org. Chem., 1993, 58, 2272 and references cited in these.
8 J. A. Cade, J. Chem. Soc., 1959, 2266.
9 cf. M. J. P. Harger, J. Chem. Soc., Perkin Trans. 1, 1983, 2127;
S. Freeman and M. J. P. Harger, J. Chem. Soc., Perkin Trans. 2, 1988,
81.
10 Dictionary of Organic Compounds, ed. J. Buckingham, Chapman
and Hall, London, 5th edn., 1982, I-01370.
11 R. J. W. Cremlyn, R. M. Ellam and N. Akhtar, Phosphorus Sulfur
Relat. Elem., 1978, 5, 1.
12 W. J. Stec, Acc. Chem. Res., 1983, 16, 411; B. Krzyzanowska and
W. J. Stec, Heteroat. Chem., 1991, 2, 123; D. Bouchu, F. Tardy,
M. Moreau, J. Dreux, A. Skowronska and J. Michalski, Tetrahedron
Lett., 1985, 26, 443 and references cited in these.
(P᎐S).
᎐
A portion of the phosphoramidate 13 (52 mg, 0.17 mmol)
having >96% the lowfield diastereoisomer (31P NMR) was
treated as above to give the phosphorothioic acid 15 [>96% one
diastereoismer, δP(CDCl3) 61.2] and some unreacted substrate
(12 mg) which was converted into 7 by N-methylation (see
above).
Paper 6/04093D
Received 11th June 1996
Accepted 11th September 1996
532
J. Chem. Soc., Perkin Trans. 1, 1997