Intramolecular nucleophilic catalysis and the exceptional reactivity of N-benzyloxycarbonyl α-aminophosphonochloridates

Article abstract of DOI:10.1039/b204935j

The intramolecular nucleophilic catalysis and the exceptional reactivity of N-benzyloxycarbonyl α-aminophosphonochloridates were discussed. It was found that N-benzyloxycarbonyl derivatives of α-aminophosphonochloridates were highly reactive because of in

Full text of DOI:10.1039/b204935j

View Article Online / Journal Homepage / Table of Contents for this issue
Intramolecular nucleophilic catalysis and the exceptional reactivity
of N-benzyloxycarbonyl ꢀ-aminophosphonochloridates
Paul M. Cullis and Martin J. P. Harger
2
Department of Chemistry, University of Leicester, Leicester, UK LE1 7RH
Received (in Cambridge, UK) 22nd May 2002, Accepted 2nd July 2002
First published as an Advance Article on the web 29th July 2002
The chloridate BnOCONHCH2P(O)(OMe)Cl is 103–104 times more reactive than ClCH2P(O)(OMe)Cl in
substitution with PriOH; the α,α-dimethyl analogue is no less reactive and the N-methyl derivative is more reactive;
nucleophilic participation (catalysis) by the carbamate group is implicated.
Aminophosphonic acids, especially the α-amino compounds,
are important transition-state analogues for mechanistic studies
of enzyme catalysis and they have been shown to inhibit a
number of proteases, notably angiotensin-converting enzyme.1
They have found use as haptens for obtaining catalytic anti-
bodies (abzymes), especially those that catalyse dipeptide
formation, and in some cases they display significant levels of
antimicrobial, herbicidal and neurophysiological activity.1 It is
hardly surprising that aminophosphonic acids and their deriv-
atives have become the focus of much synthetic work.1b In
developing efficient synthetic strategies studies of reactivity
have an important part to play.2
by comparing the reactivity of the chloridate 4 (X = Cl) and its
analogues 5 and 6 (X = Cl) alkylated at Cα or on the N atom.†
Results and discussion
Phosphonochloridates
The aminomethylphosphonate H2NCH2P(O)(OMe)2 was
obtained from commercially-available dimethyl phthal-
imidomethylphosphonate by deprotection with H2NNH2 in
methanol,10 and the 1,1-dimethyl-substituted analogue
H2NCMe2P(O)(OMe)2 was prepared from HP(O)(OMe)2 by
reaction with an acetone–ammonia mixture.11 The amino
phosphonates were then treated with benzyl chloroformate
in a two-phase system (CHCl3–aqueous NaHCO3) to give the
N-benzyloxycarbonyl derivatives 4 (X = OMe) and 5 (X =
OMe). Partial hydrolysis of the former (aqueous NaOH at
room temperature) and demethylation of the latter (NaI in
acetone at 55 ЊC) gave the crystalline monoesters 4 (X = OH)10
and 5 (X = OH). Alkylation of the diester 4 (X = OMe) with
NaH and MeI in DMF prior to hydrolysis afforded the
N-methyl monoester 6 (X = OH) (an oil). In the case of
the N-methyl compound the two rotamers resulting from slow
rotation about the carbamate C–N bond were seen to be pres-
ent (in CDCl3 solution) in a 55 : 45 ratio by NMR spectroscopy
The α-amidophosphonate ester 1 undergoes hydrolysis (acid
or base) much faster than analogous esters lacking the amide
function.3 Here, and also in related systems in which the
CO–NH sequence (relative to the P᎐O group) is reversed,4–6 the
᎐
high reactivity seems to result from intramolecular nucleophilic
catalysis by the carboxamide group, the carbonyl oxygen atom
attacking at phosphorus to form a reactive five-membered
cyclic intermediate (not detected).3–5 A more modest rate
enhancement is seen in the base-catalysed hydrolysis of the
β-amidophosphonate 2 and in this case it has been attributed to
intramolecular electrophilic catalysis, i.e. hydrogen bonding
by the amide NH activating the phosphonyl centre towards
normal intermolecular nucleophilic attack.7 Most recently the
reactivity of the protected α-aminophosphonochloridate 3, in
particular its apparent ability to form phosphonylammonium
salts,2 has also been attributed to intramolecular electro-
philic catalysis by the acidic hydrogen of the NH group.8 The
reactivity of 3 could however be equally well explained in
terms of intramolecular nucleophilic catalysis, especially if
the supposed phosphonylammonium salts are actually cyclic
oxazaphospholine oxides.9
1
[31P: δP 25.9 (major) and 25.5; H: two distinct CH2, NMe and
OMe signals].‡ The corresponding NH compound 4 (X = OH)
also displayed two rotamers in the 31P NMR spectrum [δP 25.2
(major) and 24.4] but in a more unequal 90 : 10 ratio, while for
the 1,1-dimethyl compound 5 (X = OH) one of the rotamers
appeared to be totally dominant (δP 29.7 only).
The monoesters were converted into the chloridates using
SOCl2 in CH2Cl2 or CDCl3. Even at low concentrations (0.1–0.2
mol dmϪ3 SOCl2) and in the absence of DMF or any other
catalyst, the reactions proceeded readily at room temperature.
The ease of conversion (complete in ∼0.5 h) must surely be a
The distinction between electrophilic and nucleophilic
catalysis is important, not only because of their contrasting
geometrical requirements (disposition of amide NH or CO in
† We chose OMe rather than OEt as the ‘spectator’ ligand on the phos-
phorus atom because then the sterically hindered dialkyl phosphonate
precursors 5 and 10 (X = OMe) can the more easily be dealkylated by
nucleophilic attack (at carbon) by iodide or tert-butylamine.
‡ Rotamer interconversion is sufficiently fast at room temperature to
cause pronounced line broadening in the 400 MHz 1H NMR spectra of
the N-methyl compounds 6 (X = OH, Cl, OPri etc.) but at 0 ЊC the lines
were sharp.
relation to P᎐O) and steric implications (relative insensitivity of
᎐
intramolecular nucleophilic attack to steric hindrance) but also
because a fully substituted amide group (no H on N) can act as
a nucleophile but not as an electrophile. We hoped to learn
more about the relative importance of the two types of catalysis
1538
J. Chem. Soc., Perkin Trans. 2, 2002, 1538–1543
This journal is © The Royal Society of Chemistry 2002
DOI: 10.1039/b204935j

View Article Online
result of intramolecular catalysis by the carbamate group
(cf. intermolecular catalysis by DMF); without it ClCH2P(O)-
(OH)OMe, for example, is not only less reactive towards SOCl2
but in the absence of DMF the first-formed anhydride is only
very slowly converted into the chloride. The two NH chlor-
idates, 4 (X = Cl), δP 37.3 (minor rotamer 36.8) and 5 (X = Cl),
δP 47.6, were initially obtained as oils but could be obtained as
crystalline solids by carrying out the SOCl2 reaction at ∼30 ЊC
in a small volume of a solvent in which the product is only
sparingly soluble (ether or ether-light petroleum) and then cool-
ing the reaction mixture to 0 ЊC. They were however extremely
sensitive to moisture, and any manipulation (in the absence of
SOCl2) generally introduced some of the anhydride (pyrophos-
phonate) [31P NMR: 2 peaks (diastereoisomers) at ca. δP 17 or
23]. The NMe chloridate 6 (X = Cl), δP 36.5 and 36.3 (rotamers;
ratio ca. 1 : 1) was not isolated but was used directly in the
reaction mixture in which it was prepared (HCl, SO2, excess
SOCl2 still present).
after 5 months. Relative to a Cl substituent on Cα, therefore, the
BnOCONH group accelerates the reaction with PriOH by a
factor of 103–104 in the case of 4 (X = Cl) and at least 106 in the
case of 5 (X = Cl).
The fact that the two α-chloro phosphonochloridates differ
greatly in reactivity is not surprising, given the general high
sensitivity of nucleophilic substitution at a P᎐O centre to steric
᎐
effects in the substrate.13 In the particular case of ClCMe2-
P(O)(OMe)Cl we have previously seen that a nucleophile such
as PhNH2 or ButNH2 may actually dealkylate the P-methoxy
group by nucleophilic attack at carbon more readily than it
displaces the chloride leaving group from the phosphorus
atom.12,14 The fact that the carbamate-containing substrate 5
(X = Cl) is not less reactive than its unhindered counterpart
4 (X = Cl) is therefore diagnostically important. If the
BnOCONH group were providing electrophilic catalysis of
substitution (activation of the P᎐O group by hydrogen bond-
᎐
ing) the rate-limiting step would still involve intermolecular
nucleophilic attack at phosphorus; substitution would still be
fully exposed to the effects of steric crowding, and the hindered
compound 5 (X = Cl) could not possibly react so readily. If
there is nucleophilic catalysis however, the leaving group will be
displaced by intramolecular nucleophilic attack forming a
reactive cyclic intermediate (Scheme 1). The same cyclic species
Reactivity of phosphonochloridates
The reactivity of the phosphonochloridates was examined
using PriOH as the nucleophile, a secondary alcohol being
preferred to MeOH or EtOH as a more rigorous test of
phosphonylating ability and of potential synthetic value.
The two NH chloridates 4 and 5 (X = Cl) reacted rapidly and
cleanly with PriOH in CDCl3 at room temperature giving the
expected isopropyl phosphonates 4 (X = OPri), δP 22.8 (minor
1
rotamer 22.2) and 5 (X = OPri), δP 27.8, with characteristic H
NMR signals for the new alkoxy group [δH(CDCl3) 4.7 (1H,
d × septet, JPH ∼ JHH ∼ 6) and 1.3 (6H; two d, JHH 6, diastereo-
topic CH3 groups)]. Even at high dilution (0.08 mol dmϪ3
PriOH) both reactions were у 90% complete in 6 minutes. To
obtain a more precise measure of relative reactivity an equi-
molar mixture of the phosphonochloridates was allowed to
react with 0.5 equiv. PriOH. The 31P NMR spectrum of the
reaction mixture (Fig. 1) showed both substrates to be ca. 50%
Scheme 1
is probably the immediate precursor of the oxazaphospholine
oxides now thought to be formed when the NH chloridates are
treated with Et3N.9 ¶ Such intramolecular attack could benefit
from the gem-dimethyl effect15 resulting in an enhanced rate, as
in hydantoin-forming cyclisation of the ester 7 (1100 times
faster with R = Me than with R = H at pH р 2),16 or at least a
smaller than usual ‘neopentyl’ steric retardation, as in cyclis-
ation of the aminoalkyl bromide 8 and in sulfur-assisted
methanolysis of the tosylate 9 (<10 times slower with R = Me
than with R = H).17,18 For the α,α-dimethyl phosphono-
chloridate 5 (X = Cl) to be as reactive as its unmethylated
counterpart 4 (X = Cl) but not more reactive, the beneficial
gem-dimethyl effect would have to be balanced by the adverse
effect of steric crowding at the P᎐O centre in the rate-limiting
᎐
cyclisation (Scheme 1). Alternatively, if cyclisation is revers-
ible a relatively fast intramolecular formation of the cyclic
intermediate could be balanced by a relatively slow (steric
hindrance) intermolecular reaction of the intermediate with the
external nucleophile.
Fig. 1 31P NMR spectrum of the mixture (1 : 1) of 4 (X = Cl)
(rotamers; ratio ∼15 : 1) (᭹) and 5 (X = Cl) (᭢) before (top) and after
reaction (bottom) with PriOH (0.5 equiv.).
An obvious distinction between the two types of catalysis lies
in their dependence on the acidic hydrogen of the carbamate
group; without it electrophilic catalysis cannot occur whereas
nucleophilic catalysis is possible and could even benefit from
converted into product, implying that their reactivities are
very similar. Thus the presence of two methyl groups on the
α carbon atom of 5 (X = Cl) has no significant effect on its
reactivity towards PriOH in nucleophilic substitution.
To gain some measure of the importance of intramolecular
catalysis by the carbamate group, a comparison was made with
the corresponding α-chloro compounds ClCH2P(O)(OMe)Cl§
and ClCMe2P(O)(OMe)Cl,12 reasoning that the inductive and
steric effects of the carbamate group should be matched better
by a Cl atom at Cα than by a H or alkyl group. Using PriOH in
CDCl3 (∼0.12 mol dmϪ3; slight excess) the α-chloromethyl
compound was only half-consumed in 5 days (T ∼20 ЊC) while
the more alkylated compound was practically unchanged (99%)
§ Our ClCH2P(O)(OMe)Cl contained substantial amounts of ClCH2-
P(O)Cl2 and ClCH2P(O)(OMe)2 [reaction of ClCH2P(O)Cl2 with 1
equiv. MeOH shows rather poor selectivity]; its (slow) reaction with
PriOH was complicated by preferential reaction with traces of moisture
and by degradation of the product (acid-catalysed dealkylation of
P–OPri group).
¶ The oxazaphospholine oxides have not been isolated but the spectro-
scopic evidence of structure is good. They react very readily with nucleo-
philes but should be isolable if moisture can be completely excluded.
J. Chem. Soc., Perkin Trans. 2, 2002, 1538–1543
1539

View Article Online
diester 10 (X = OMe) was carried out by heating with ButNH2
(large excess; 40 h at 55 ЊC). Dilution of the reaction mixture
with ether gave a precipitate of the pure anhydrous tert-butyl-
ammonium salt of the monoester (10; X = OϪ ϩNH3But), and
this reacted readily with SOCl2 in CDCl3. The product, how-
ever, was not the expected phosphonochloridate 10 (X = Cl).
Instead, the benzyl group was removed as benzyl chloride
[δH 4.59 (2H, s); δC 46.7; Mϩ (GC–MS) 126 and 128] leaving
what appears to be the oxooxazaphospholane oxide 13 (δP 31.1;
νmax/cmϪ1 1775; Mϩ 193). In accord with the proposed cyclic
structure 13, addition of MeOH to the reaction mixture pro-
duced the (known) aminophosphonate MeNHCMe2P(O)-
(OMe)2 corresponding to nucleophilic attack at phosphorus
and ring opening followed by loss of CO2. We have found no
other examples of this ring system but it is not difficult to
rationalise its formation. Nucleophilic participation by the
carbamate group will convert the chloridate 10 (X = Cl) or
more probably its precursor 10 (X = OSOCl) into the cyclic
intermediate 12 (Scheme 2) but because of the methyl groups
alkylation of the N atom. The behaviour of the N-methyl
substrate 6 (X = Cl) is therefore of particular importance.
Under similar conditions (0.08 mol dmϪ3 PriOH in CDCl3) it
formed the corresponding isopropyl ester 6 (X = OPri), δP 22.7
and 22.15 (ratio 57 : 43, rotamers) at least as quickly as the NH
substrates 4 and 5 (X = Cl), and when required to compete with
4 (X = Cl) for a limited amount of PriOH (0.43 equiv.) it was
seen to be substantially more reactive (Fig. 2). High reactivity of
Scheme 2
on Cα it is the benzylic carbon atom that is attacked by chlor-
ide ion rather than the sterically hindered P᎐O centre. A
᎐
similar combination of intramolecular nucleophilic attack and
debenzylation by chloride ion is well known for acyl chlorides
derived from N-benzyloxycarbonyl-protected α-amino carb-
oxylic acids.19 The formation of 13 can therefore be seen as
powerful support for nucleophilic participation by the carb-
amate group, although it is unfortunate that the very efficiency
of participation has thwarted our attempts to prepare the
phosphonochloridate we hoped to study.
Fig. 2 31P NMR spectrum of the mixture (1 : 2) of 4 (X = Cl) (᭹) and
6 (X = Cl) (rotamers; ratio ∼1 : 1) (᭢) before (top) and after reaction
(bottom) with PriOH (0.43 equiv.) [minor (highfield) rotamer of 4
(X = Cl) hidden by lowfield rotamer of 6 (X = Cl); product 4 (X = OPri)
(if present) hidden by lowfield rotamer of 6 (X = OPri); change in
substrate ratio (1 : 2.0 to 1 : 0.8) implies product derived largely from 6
(X = Cl)].
Conclusion
N-Benzyloxycarbonyl derivatives of α-aminophosphono-
chloridates are highly reactive because of intramolecular
catalysis by the carbamate group. The high reactivity is not
diminished when the acidic hydrogen of the carbamate NH
moiety is replaced by an alkyl group, as would be the case for
the chloridates under these non-basic conditions (no scavenging
of liberated HCl) therefore does not depend on the presence of
an NH group or by implication on intramolecular electrophilic
catalysis.
intramolecular electrophilic catalysis (activation of the P᎐O
᎐
group by hydrogen bonding), or when alkylation of the α
It had been our intention to include the fully methylated
chloridate 10 (X = Cl) in our study of reactivity but unfortu-
nately we were unable to prepare it. The diester 5 (X = OMe)
was successfully methylated on nitrogen (NaH then MeI in
DMF) and demethylated at oxygen (NaI in acetone at 55 ЊC)
giving an aqueous solution of the sodium salt of the monoester
10 (X = OH). However, acidification and extraction with CHCl3
did not give the free monoester but rather the phosphonic acid
11 corresponding to loss of the remaining O-methyl group.
Apparently the free monoester has only limited stability and is
rapidly degraded by hydrolytic displacement of the OMe group
from phosphorus (or less plausibly dealkylation of the OMe
group by nucleophilic attack at carbon). When the acidification
and extraction were carried out in an NMR tube and the
spectrum of the CDCl3 layer recorded as quickly as possible
some degradation of the monoester was already apparent
carbon atom impedes intermolecular nucleophilic attack at the
P᎐O centre (steric hindrance). It thus seems that intramolecular
᎐
nucleophilic catalysis is responsible for the high reactivity, and
that the external nucleophile reacts rapidly with the resulting
cyclic species rather than with the phosphonochloridate itself.
An understanding of the reactivity of these α-aminophos-
phonic acid derivatives will, we think, enhance their value in the
synthesis of biologically significant phosphonates.
Experimental
Mps were determined using a Kofler hot-stage apparatus and
1
are uncorrected. H NMR spectra were recorded on Bruker
ARX 250, DPX 300 or DRX 400 spectrometers (Me4Si internal
standard; coupling constants J given in Hz) at room temper-
ature or (where indicated) 0 ЊC (to reduce rate of interconver-
sion of rotamers); 31P NMR spectra (1H decoupled) were also
recorded on these instruments at 101, 122 or 162 MHz (positive
chemical shifts downfield from 85% H3PO4).
[δP 29.0
30.2; δH 3.79 (d, JPH 12)
3.50 (s) (MeOH)].
To obtain a salt that is soluble in aprotic solvents and avoid
the need to isolate the free monoester the demethylation of the
Mass spectra were obtained in EI (70 eV) or FAB (NBA
matrix) mode using a Kratos Concept spectrometer or in ES
mode using a Micromass Quattro spectrometer. DMF and
isopropanol were repeatedly dried over 3 Å molecular sieves
1540
J. Chem. Soc., Perkin Trans. 2, 2002, 1538–1543

View Article Online
(pellets and powder respectively). SOCl2 was distilled. Ether
refers to diethyl ether.
301.1080. C13H20NO5P requires: C, 51.8; H, 6.7; N, 4.65%; M,
301.1079.
A solution of the diester 5 (X = OMe) (602 mg, 2.0 mmol)
in acetone (4 ml) containing NaI (1.50 g, 10.0 mmol) was
maintained at 55 ЊC for 18 h. Volatile material was evaporated
in vacuo and the residue was dissolved in water (10 ml).
The solution was washed with CHCl3 (2 ml) and acidified with
12 mol dmϪ3 HCl (0.33 ml) and the product was extracted into
CHCl3 (30 ml; 4 × 10 ml). The extract was dried (Na2SO4) and
concentrated; crystallisation from CHCl3 diluted with ether
and light petroleum (bp 60–80 ЊC) afforded methyl hydrogen
N-(Benzyloxycarbonyl)aminomethylphosphonates
Following the published procedure10 dimethyl phthalimido-
methylphosphonate (4.25 g, 15.9 mmol) was treated with
hydrazine hydrate (0.88 g, 17.5 mmol) in MeOH (25 ml)
at ∼26 ЊC. After 49 h reaction was still incomplete (31P
NMR spectroscopy) but the yield of dimethyl aminomethyl-
phosphonate (∼70%) was not improved by extending the reac-
tion time because of degradation of the product (δP 31.8
15.4 ppm; probably demethylation). After isolation the crude
amino compound was immediately dissolved in CHCl3 (45 ml),
stirred vigorously with NaHCO3 (1.8 g, 21.5 mmol) in water
(18 ml) and cooled; benzyl chloroformate (3.5 g, 20.8 mmol)
was added and stirring was continued for 25 min at room
temperature. The organic layer was separated, washed with
water (2 × 10 ml), dried (Na2SO4) and concentrated to give
N-(benzyloxycarbonyl)-1-amino-1-methylethylphosphonate
5
(X = OH) (528 mg, 91%); mp 112–113.5 ЊC; (EI) m/z 287 (Mϩ,
3%), 192 [Mϩ Ϫ P(O)(OH)OMe, 30], 148 (25) and 91 (100);
δP (CDCl3) 29.7; δH (CDCl3, 300 MHz) 12.01 (1H, s, OH), 7.33
(5H), 5.42 (1H, s, NH), 5.05 (2H, s), 3.73 (3H, d, JPH 11, OMe)
and 1.575 (6H, d, JPH 15.5, Me2C); νmax (Nujol)/cmϪ1 3335
(NH), 3000–2000 (OH) and 1730 (C᎐O). Found: C, 50.3; H, 6.4;
᎐
N, 4.9%; Mϩ, 287.0923. C12H18NO5P requires: C, 50.2; H, 6.3;
N, 4.9%; M, 287.0923.
dimethyl N-(benzyloxycarbonyl)aminomethylphosphonate
4
(X = OMe) (an oil; >95% by 31P NMR spectroscopy) (see
below). Purification was not attempted at this point.
N-(Benzyloxycarbonyl)-N-methylaminomethylphosphonates
The crude dimethyl ester was hydrolysed by stirring with
2.0 mol dmϪ3 aqueous NaOH (20 ml) at 20–25 ЊC for 40 min.
The resulting solution was washed with CHCl3, acidified by
careful addition of H2SO4, and extracted repeatedly with
CHCl3 (6 × 10–15 ml). The combined extracts were dried
(Na2SO4) and concentrated. Crystallisation from toluene
afforded methyl hydrogen N-(benzyloxycarbonyl)aminomethyl-
phosphonate 4 (X = OH) (2.50 g, 61% overall); mp 103–
104 ЊC (lit.,10 106–106.5 ЊC); δP (CDCl3) 25.2 (rotamer 24.4);
δH (CDCl3, 250 MHz) 9.9 (br s, OH), 7.33 (5H), 5.35 (br, NH),
5.11 (2H, s), 3.71 (3H, d, JPH 11, OMe) and 3.615 (2H, d, JPH 12,
CH2P); νmax(Nujol)/cmϪ1 3340 (NH), 3000–2000 (OH) and 1720
A solution of dimethyl N-(benzyloxycarbonyl)aminomethyl-
phosphonate 4 (X = OMe) (519 mg, 1.9 mmol) in DMF (4 ml)
was stirred and cooled in ice while NaH (64 mg, 2.65 mmol) was
added. The mixture was allowed to warm to room temperature
and after 15 min was again cooled while MeI (480 mg, 3.4
mmol) was added. After a further 15 min at room temperature
MeOH (100 µl) was added (to destroy any remaining NaH) and
volatile material was removed in vacuo at ∼ 40 ЊC giving a mix-
ture (70 : 30 by 31P NMR in MeOH) of the required dimethyl
phosphonate 6 (X = OMe) and the corresponding monomethyl
compound (Na salt) contaminated with much DMF. The
dimethyl phosphonate was hydrolysed by stirring the mixture
with 1.5 mol dmϪ3 NaOH (4 ml) for 2 h. More water (3 ml) was
added and the solution was washed repeatedly with CHCl3
to remove all DMF and give an aqueous solution of pure
sodium methyl N-(benzyloxycarbonyl)-N-methylaminomethyl-
phosphonate 6 (X = ONa), δP (H2O) 19.85 and 19.5 (ratio ∼1 : 1;
rotamers). Acidification with 2 mol dmϪ3 HCl and repeated
extraction with CH2Cl2 afforded methyl hydrogen N-(benzyloxy-
carbonyl)-N-methylaminomethylphosphonate 6 (X = OH) as an
oil, (EI) m/z 273 (Mϩ, 10%), 256 (20) and 91 (100); δP (CDCl3)
25.9 and 25.5 (ratio 55 : 45; rotamers); δH (CDCl3, 400 MHz at
0 ЊC) 8.0 (br s, OH ϩ H2O), 7.4–7.25 (5H), 5.14 (2H, s), 3.77
(major rotamer) and 3.73 (total 2H; both d, JPH 10.5, CH2P),
3.75 (major) and 3.645 (total 3H; both d, JPH 11, OMe) and
3.06 and 3.05 (major) (total 3H; both s, NMe); νmax (CDCl3)/
cmϪ1 1700 (C᎐O), 1220 (P᎐O) and 1065. Found: Mϩ, 273.0766;
(C᎐O).
᎐
Pure dimethyl N-(benzyloxycarbonyl)aminomethylphosphon-
ate 4 (X = OMe) was obtained by treating 4 (X = OH) with
diazomethane: (EI) m/z 273 (Mϩ, 15%), 166 (Mϩ Ϫ PhCH2O,
25) and 91 (100); δP (CDCl3) 25.2; δH (CDCl3, 250 MHz) 7.4–
7.25 (5H), 5.90 (1H, br t, NH), 5.10 (2H, s), 3.72 (6H, d, JPH 11,
OMe) and 3.62 (2H, dd, JPH 11, JHH 6, CH2P); νmax (film)/cmϪ1
1725 (C᎐O).
᎐
N-(Benzyloxycarbonyl)-1-amino-1-methylethylphosphonates
Ammonia was passed through stirred and ice-cooled acetone
(11.6 g, 0.20 mol) and dimethyl phosphite (16.5 g, 0.15 mol)
was added dropwise over 45 min.11 While continuing to pass
ammonia the mixture was allowed to warm to room temper-
ature; it was then heated at 50 ЊC (bath temp.) for 10 min.
Volatile material was evaporated in vacuo and the residue was
extracted with ether to give a mixture (18.65 g) of dimethyl
1-amino-1-methylethylphosphonate, δP (CDCl3) 34.1, δH (CDCl3,
250 MHz) 3.80 (d, JPH 12, OMe) and 1.30 (d, JPH 15, Me2C) and
the corresponding 1-hydroxy compound, δP 29.7, δH 3.81 (d, JPH
12) and 1.43 (d, JPH 14), in a ratio ∼ 2 : 1. A portion of the
mixture (10.3 g) was dissolved in CHCl3 (80 ml) and was stirred
vigorously with NaHCO3 (5.05 g, 60 mmol) in water (45 ml);
benzyl chloroformate (8.2 g, 48 mmol) was added and stirring
was continued for 1.3 h. The organic layer was collected,
washed with water (2 × 50 ml) and dried (Na2SO4). The solvent
was evaporated and the resulting solid was washed with ether
and crystallised from a small volume of CHCl3 diluted with
ether and light petroleum (bp 60–80 ЊC) to give dimethyl
᎐
᎐
C11H16NO5P requires M, 273.0766. This compound is much
more stable than 10 (X = OH) but was used with a minimum of
delay.
A pure sample of dimethyl N-(benzyloxycarbonyl)-N-methyl-
aminomethylphosphonate 6 (X = OMe) was obtained by treating
6 (X = OH) with diazomethane: (EI) m/z 287 (Mϩ, 20%), 152
(Mϩ Ϫ BnOCO, 30) and 91 (100); δP (CDCl3) 25.1 and 24.5
(ratio ∼3 : 2; rotamers); δH (CDCl3, 300 MHz) 7.45–7.3 (5H),
5.15 (2H, s), 3.76 (6H, d, JPH 10.5, OMe), 3.73 and 3.67 (major
rotamer) (total 2H; both d, JPH 11, CH2P) and 3.05 (3H, s,
NMe); νmax (CDCl3)/cmϪ1 1705 (C᎐O), 1220 (P᎐O), 1065 and
᎐
᎐
1045. Found: Mϩ, 287.0923; C12H18NO5P requires M, 287.0923.
N-(Benzyloxycarbonyl)-N-methyl-1-amino-1-methylethyl-
phosphonates
N-(benzyloxycarbonyl)-1-amino-1-methylethylphosphonate
5
(X = OMe) (9.7 g, 39% overall); mp 113.5–114 ЊC; (EI) m/z 301
(Mϩ, 2%), 192 [Mϩ Ϫ P(O)(OMe)2, 30], 148 (25) and 91 (100);
δP (CDCl3) 29.9; δH (CDCl3, 300 MHz) 7.45–7.35 (5H), 5.06
(2H, s), 5.05 (1H, s, NH), 3.79 (6H, d, JPH 10.5, OMe) and 1.605
(6H, d, JPH 16.5, Me2C); νmax (CDCl3)/cmϪ1 3420 (NH), 1735
(C᎐O) and 1260 (P᎐O). Found: C, 51.8; H, 6.6; N, 4.7%; Mϩ,
NaH (96 mg, 4 mmol) was added to a stirred solution of
dimethyl N-(benzyloxycarbonyl)-1-amino-1-methylethylphos-
phonate 5 (X = OMe) (602 mg, 2.0 mmol) in DMF. Much solid
precipitated. After 20 min the mixture was cooled in ice and
MeI (850 mg, 6.0 mmol) was added; stirring was continued for
᎐
᎐
J. Chem. Soc., Perkin Trans. 2, 2002, 1538–1543
1541

View Article Online
30 min at room temperature, then NH4Cl (160 mg, 3 mmol)
was added to destroy any remaining NaH. The mixture was
diluted with CH2Cl2 (40 ml) and filtered through Celite,
and volatile material was removed as completely as possible
in vacuo (some DMF remained). Without further purification
the dimethyl phosphonate was demethylated by heating with
NaI (1.45 g, 9.8 mmol) in acetone (4 ml) at 55 ЊC for 15 h (31P
The same method, but with a mixture of ether and
light petroleum (bp 40–60 ЊC) as solvent, was used to convert
5 (X = OH) into methyl N-(benzyloxycarbonyl)-1-amino-
1-methylethylphosphonochloridate 5 (X = Cl) (88%); mp 71–
72 ЊC; δP (CDCl3) 47.6 (trace 23.2 and 22.8, pyrophosphonate
diastereoisomers); δH (CDCl3, 300 MHz) 7.4–7.25 (5H), 5.39
(1H, s, NH), 5.06 (2H, s), 3.85 (3H, d, JPH 12.5, OMe) and 1.69
(6H, d, JPH 19, Me2C); νmax (CDCl3)/cmϪ1 3420 (NH), 1735
NMR: δP 30.7
25.0), giving a solution of sodium methyl
N-(benzyloxycarbonyl)-N-methyl-1-amino-1-methylethylphos-
phonate 10 (X = ONa), (ϪES) m/z 300.
(C᎐O), 1260 (P᎐O) and 1050.
᎐ ᎐
The N-methyl compound 6 (X = OH) was treated with SOCl2
(1.5 equiv.) in CDCl3 to give methyl N-(benzyloxycarbonyl)-N-
methylaminomethylphosphonochloridate 6 (X = Cl) which was
used without isolation: δP (CDCl3) 36.5 and 36.3 (ratio ∼1 : 1,
rotamers); δH (CDCl3, 400 MHz at 0 ЊC) 7.45–7.3 (5H), 5.18
and 5.165 (total 2H; both s), 4.13 (ABP; δA 4.25, δB 4.01, JAB 16,
JPH 6.5) and 4.08 (ABP; δA 4.19, δB 3.97, JAB 16, JPH 6.5) (total
2H; CH2P), 3.93 and 3.85 (total 3H; both d, JPH 13, OMe) and
3.13 and 3.11 (total 3H; both s, NMe); νmax (CDCl3)/cmϪ1 1710
In an unsuccessful attempt to obtain the hydrogen methyl
phosphonate, acetone was evaporated off and the residue was
dissolved in water. Acidification (HCl) and extraction with
CHCl3 gave an oil but this contained no P–OMe group (1H
NMR); crystallisation from ether afforded N-(benzyloxy-
carbonyl)-N-methyl-1-amino-1-methylethylphosphonic acid 11
(420 mg, 74%); δP (CDCl3) 30.4; δH (CDCl3, 250 MHz) 10.83
(2H, s), 7.4–7.3 (5H), 5.14 (2H, s), 3.06 (3H, s, NMe) and 1.66
(6H, d, JPH 15). [A similar experiment was conducted in an
NMR tube using a D2O solution of 10 (X = ONa), CF3CO2H
for acidification, and CDCl3 for extraction; the NMR spectrum
of the CDCl3 extract showed demethylation (hydrolysis) to be
(C᎐O), 1235 (P᎐O) and 1050.
᎐
᎐
Reaction of salt 10 (X ؍ O؊ ؉NH3But) with thionyl chloride
Thionyl chloride (excess) was added to a solution of the salt 10
(X = OϪ ϩNH3But) in CDCl3 giving a product δP 31.1; νmax/cmϪ1
appreciable within just a few minutes: δP 29.0
(d, JPH 12) 3.50 (s) (MeOH)].
30.2; δH 3.79
1775 (C᎐O), 1280, 1235 and 1055; (EI) m/z 193 (Mϩ, 45%), 178
᎐
The phosphonic acid 11 was treated with diazomethane to
give pure dimethyl N-(benzyloxycarbonyl)-N-methyl-1-amino-
1-methylethylphosphonate 10 (X = OMe) (an oil), (EI) m/z 315
(Mϩ, 0.2%), 206 [Mϩ Ϫ P(O)(OMe)2, 40], 162 (35) and 91 (100);
(FAB) m/z 316 (M ϩ Hϩ, 95%), 206 (100) and 162 (55);
δP (CDCl3) 29.9; δH (CDCl3, 300 MHz) 7.5–7.3 (5H), 5.12 (2H,
s), 3.73 (6H, d, JPH 10, OMe), 3.075 (3H, s, NMe) and 1.69 (6H,
d, JPH 15.5, Me2C); νmax (film)/cmϪ1 1715 (C᎐O), 1255 (P᎐O),
(Mϩ Ϫ Me, 100) and 136 (20) (Found: Mϩ, 193.0504. C6H12-
NO4P requires M, 193.0504); the H and 13C NMR spectra of
1
the reaction mixture were consistent with the oxooxazaphos-
pholane oxide 13, δH (300 MHz) 3.96 (3H, d, JPH 11, OMe), 2.85
(3H, s, NMe), 1.51 (3H, d, JPH 16.5, CMe) and 1.45 (3H, d, JPH
15.5, CMe), δC (75 MHz) 149.1 (d, JPC 11, CO), 56.1 (d, JPC 130,
Cα), 55.1 (d, JPC 7, OMe), 26.5 (d, JPC 11, NMe), 21.8 (s) and
21.55 (d, JPC 3.5), accompanied by PhCH2Cl, δH 7.45–7.3 (5H)
and 4.59 (2H, s); δC 137.9 (s), 129.1 (s), 129.0 (s), 128.8 (s) and
46.7 (s) and ButNH3ϩClϪ (partially insoluble), δH 8.3 (br s) and
1.47 (s), δC 53.4 (s) and 28.1 (s). The presence of PhCH2Cl was
confirmed by GC–MS, (EI) m/z 128 and 126 (Mϩ, 80%) and
91 (100).
Addition of MeOH to the reaction mixture converted the
product into MeNHCMe2P(O)(OMe)2 (initially as the hydro-
chloride) which was isolated and characterised by NMR
spectroscopy, δP (CDCl3) 33.3; δH (CDCl3, 300 MHz) 3.80 (6H,
d, JPH 10), 2.46 (3H, s), 1.88 br (s, NH) and 1.32 (6H, d, JPH
15.5), and by conversion into the picrate, mp 146–148 ЊC (lit.,14
147–148 ЊC).
᎐
᎐
1060 and 1040. Found: Mϩ, 315.1236. C14H22NO5P requires M,
315.1236.
A solution of 10 (X = OMe) (490 mg, 1.55 mmol) in tert-
butylamine (3 ml) was heated in a sealed vessel at 55 ЊC for 40 h
(δP 30.3
20.2). Most of the tert-butylamine was evaporated
off and ether (10 ml) was added; on refrigeration tert-butyl-
ammonium methyl N-(benzyloxycarbonyl)-N-methyl-1-amino-
1-methylethylphosphonate 10 (X = OϪ ϩNH3But) (448 mg, 78%)
precipitated: mp 110–112 ЊC (sealed tube; softens at variable
lower T ); (FAB) m/z 375 (M ϩ Hϩ, 100%) and 302 (75); (ϪES)
m/z 300; δP (CDCl3) 21.1; δH (CDCl3, 300 MHz) 8.4 (3H, br s,
ϩH3NBut), 7.45–7.25 (5H), 5.06 (2H, s), 3.55 (3H, d, JPH 9.5,
OMe), 3.14 (3H, s, NMe), 1.62 (6H, d, JPH 13.5, Me2C) and 1.33
(9H, s, But); νmax (Nujol)/cmϪ1 3500–2500 (NH), 1705 (C᎐O),
᎐
Phosphonochloridate reactions with isopropanol
1465 and 1170. Found: C, 54.6; H, 8.4; N, 7.4%; M ϩ Hϩ
(FAB), 375.2048. C17H31N2O5P requires: C, 54.5; H, 8.35; N,
7.5%; M ϩ H, 375.2049.
Isopropanol (30 mg, 0.5 mmol) was added to a solution of the
phosphonochloridate 4 (X = Cl) (69 mg, 0.25 mmol) in CDCl3
(0.6 ml). Reaction was complete (δP 37.4
23.5) inside 10 min.
Volatile material was removed in vacuo to give isopropyl methyl
N-(benzyloxycarbonyl)aminomethylphosphonate 4 (X = OPri);
(EI) m/z 301 (Mϩ, 15%), 259 (Mϩ Ϫ C3H6, 30), 152 (50), 108
(35) and 91 (100); (FAB) m/z 302 (M ϩ Hϩ, 100%) and 260
(M ϩ Hϩ Ϫ C3H6, 30); δP (CDCl3) 22.8 (rotamer 22.2);
δH (CDCl3, 300 MHz) 7.4–7.25 (5H), 5.64 br (1H, NH), 5.11
(2H, s), 4.715 (1H, d × septet, JPH ∼ JHH ∼ 6, OCHMe2), 3.71
(3H, d, JPH 11, OMe), 3.61 (2H, dd, JPH 11.5, JHH 6, NHCH2P)
and 1.31 and 1.275 (both 3H, d, JHH 6; OCHMe2); νmax (film)/
cmϪ1 3235 (NH), 1725 (C᎐O), 1275, 1235 (P᎐O) and 1010.
Phosphonochloridates
Thionyl chloride (45 mg, 0.38 mmol) was added to a suspension
of methyl hydrogen N-(benzyloxycarbonyl)aminomethylphos-
phonate 4 (X = OH) (69 mg, 0.27 mmol) in ether (0.5 ml). The
mixture was maintained at 25–30 ЊC with frequent agitation for
40 min during which time the initial solid passed into solution
and a new solid began to precipitate. After cooling at 0–5 ЊC for
30 min the supernatant was removed using a fine-tipped pipette
(minimum exposure to moisture) and the solid was pumped at
0.2 mmHg to give methyl N-(benzyloxycarbonyl)aminomethyl-
phosphonochloridate 4 (X = Cl) (69 mg, 93%); δP (CDCl3) 37.3
(rotamer 36.8); δH (CDCl3, 300 MHz) 7.34 (5H), 5.72 (1H, br t,
᎐
᎐
Found: M ϩ Hϩ (FAB) 302.1157. C13H20NO5P requires: M ϩ
H, 302.1157.
The corresponding reaction of 5 (X = Cl) with isopropanol
JHH ∼ 6, NH), 5.13 (2H, s) (rotamer 5.08), 3.96 (2H, dd, JPH
∼
(δP 47.5
27.8; complete inside 10 min) gave isopropyl methyl
JHH ∼ 6, NHCH2P) and 3.86 (3H, d, JPH 13, OMe); νmax
N-(benzyloxycarbonyl)-1-amino-1-methylethylphosphonate
5
(CDCl3)/cmϪ1 3430 and 3280 br (NH), 1735 (C᎐O), 1245 (P᎐O)
(X = OPri); (FAB) m/z 330 (M ϩ Hϩ, 100%) and 288 (M ϩ Hϩ
Ϫ C3H6, 20); δP (CDCl3) 27.8; δH (CDCl3, 300 MHz) 7.4–
7.25 (5H), 5.21 (1H, d, JPH 5.5, NH), 5.06 (2H, s), 4.73 (1H,
d × septet, JPH ∼ JHH ∼ 6, OCHMe2), 3.75 (3H, d, JPH 10.5,
OMe), 1.60 and 1.59 (both 3H, d, JPH 16; Me2C) and 1.33 and
᎐
᎐
and 1055. This material is hydrolysed very readily; the mp (73–
75 ЊC) is probably unreliable and a MS could not be obtained;
the 31P NMR spectrum contained evidence of some hydrolysis
(δP 17.3 and 17.2; pyrophosphonate diastereoisomers).
1542
J. Chem. Soc., Perkin Trans. 2, 2002, 1538–1543

View Article Online
1.315 (both 3H, d, JHH 6; OCHMe2); νmax (film)/cmϪ1
3 J. Rahil and R. F. Pratt, J. Chem. Soc., Perkin Trans. 2, 1991, 947.
4 R. Kluger and J. L. W. Chan, J. Am. Chem. Soc., 1976, 98, 4913.
5 R. Kluger, J. F. Chow and J. J. Croke, J. Am. Chem. Soc., 1984, 106,
4017.
6 L. A. Reiter and B. P. Jones, J. Org. Chem., 1997, 62, 2808.
7 R. A. Naylor and A. Williams, J. Chem. Soc., Perkin Trans. 2, 1976,
1908.
3240 (NH), 1735 (C᎐O), 1260, 1240 (P᎐O) and 1005. Found:
᎐
᎐
M ϩ Hϩ (FAB) 330.1471. C15H24NO5P requires: M ϩ H,
330.1470.
The corresponding reaction of 6 (X = Cl) with isopropanol
[δP 36.45 and 36.25 (rotamers)
23.05 and 22.45; complete
8 A. B. Smith, III, L. Ducry, R. M. Corbett and R. Hirschmann,
Org. Lett., 2000, 2, 3887.
inside 10 min] gave isopropyl methyl N-(benzyloxycarbonyl)-N-
methylaminomethylphosphonate 6 (X = OPri); (FAB) m/z 316
(M ϩ Hϩ, 100%); δP (CDCl3) 22.7 and 22.15 (ratio 57 : 43,
rotamers); δH (CDCl3, 400 MHz at 0 ЊC) 7.45–7.3 (5H), 5.15
(2H, s), 4.82–4.68 [1H, m; with 31P decoupling, 4.765 (major
rotamer) and 4.73 (both septet, JHH 7, OCHMe2)], 3.85–3.61
(2H, m, CH2P), 3.75 (major) and 3.63 (total 3H; both d, JPH
10.5, OMe), 3.07 (major) and 3.065 (total 3H, both s; NMe)
and 1.34 and 1.305 (major), 1.315 and 1.27 (total 6H; all d, JHH
6.5; OCHMe2); νmax (film)/cmϪ1 1710 (C᎐O), 1250, 1220 (P᎐O)
9 P. M. Cullis and M. J. P. Harger, Chem. Commun., 2002, 458.
10 N. A. Jacobsen and P. A. Bartlett, J. Am. Chem. Soc., 1981, 103, 654.
11 K. A. Petrov, V. A. Chauzov, I. V. Pastukhova and N. N. Bogdanov,
J. Gen. Chem. USSR (Engl. Transl.), 1976, 46, 1226.
12 M. J. P. Harger and A. Williams, J. Chem. Soc., Perkin Trans. 1,
1986, 1681.
13 R. S. Edmundson, in Chemistry of Organophosphorus Compounds,
ed. F. R. Hartley, Wiley, Chichester, 1996, vol. 4, ch. 6. For
some particularly relevant examples (P–Cl bond cleavage) see:
A. A. Neimysheva and I. L. Knunyants, J. Gen. Chem. USSR
(Engl. Transl.), 1968, 38, 575; I. L. Knunyants and A. A.
Neimysheva, J. Gen. Chem. USSR (Engl. Transl.), 1972, 42, 2415.
14 M. J. P. Harger and A. Williams, J. Chem. Soc., Perkin Trans. 1,
1989, 563.
᎐
᎐
and 1005. Found: M ϩ Hϩ (FAB) 316.1314. C14H22NO5P
requires: M ϩ H, 316.1314.
15 B. Capon and S. P. McManus, Neighbouring Group Participation,
Plenum, New York, 1976, vol. 1, ch. 2; E. L. Eliel and S. H. Wilen,
Stereochemistry of Carbon Compounds, Wiley, New York, 1994,
pp. 682–684.
16 I. V. Blagoeva, D. T. Tashev and A. J. Kirby, J. Chem. Soc., Perkin
Trans. 1, 1989, 1157.
17 R. F. Brown and N. M. van Gulick, J. Org. Chem., 1956, 21, 1046.
18 E. L. Eliel and D. E. Knox, J. Am. Chem. Soc., 1985, 107, 2946.
19 M. Bodanszky, Principles of Peptide Synthesis, Springer-Verlag,
Berlin, 1984, pp. 10–11.
References
1 (a) P. Kafarski and B. Lejczak, Phosphorus, Sulfur Silicon Relat.
Elem., 1991, 63, 193; J. Huang and R. Chen, Heteroat. Chem., 2000,
11, 480; (b) V. P. Kukhar and H. R. Hudson, Aminophosphonic and
Aminophosphinic Acids, Wiley, Chichester, 2000; . See also ref. 2 for a
summary and leading references.
2 R. Hirschmann, K. M. Yager, C. M. Taylor, J. Witherington,
P. A. Sprengeler, B. W. Phillips, W. Moore and A. B. Smith, III,
J. Am. Chem. Soc., 1997, 119, 8177.
J. Chem. Soc., Perkin Trans. 2, 2002, 1538–1543
1543