Use of an 18O-labelled phosphonamidic-sulfonic anhydride to learn more about the mechanism by which O-sulfonyl-N-phosphinoylhydroxylamines rearrange

Article abstract of DOI:10.1039/b603051c

The ¹⁸O-labelled phosphonamidic-sulfonic anhydride was used to learn more about the mechanism by which O-sulfonyl-N-phosphinoylhydroxylamines rearrange. It was observed that the anhydride intermediate could be attacked at sulfur instead of phos

Full text of DOI:10.1039/b603051c

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/obc | Organic & Biomolecular Chemistry
Use of an 18O-labelled phosphonamidic-sulfonic anhydride to learn more about
the mechanism by which O-sulfonyl-N-phosphinoylhydroxylamines rearrange
Martin J. P. Harger*
Received 28th February 2006, Accepted 28th March 2006
First published as an Advance Article on the web 12th April 2006
DOI: 10.1039/b603051c
With ButNH2 the anhydride R(PhNH)P(O)OS18O2Bn 5 (R =
But), like the hydroxylamine derivative R(Ph)P(O)NH-
OS18O2Bn 4 (R = PhMeCH), gives ButNHS18O2Bn containing
only a part (76–78%) of the available 18O label; this is a result
with any straightforward transposition mechanism. Our concern
now is with the possibility that the distribution of the label in the
anhydride intermediate might not be the same when it reacts as
when it is formed.
of partial scrambling of the label in 5 (OS18O2
ꢀ
18OSO2)
Ideally we would have used the anhydride 5 (R = PhMeCH)
labelled with 18O specifically in the bridging (SOP) or the
non-bridging (SO2) position but it could not be obtained.
Phosphonylation4 of labelled sulfonate (BnS18O3−) cannot possibly
give a product with 18O in just one position, and sulfonylation of
phosphonate generally fails because the phosphonate displaces
sulfonate from the anhydride as it is formed, giving largely
pyrophosphonate.5 This pyrophosphonate formation involves
nucleophilic attack at the P atom of the anhydride; we thought it
might usefully be suppressed (steric hindrance) if the alkyl group
on phosphorus (R = PhMeCH) were changed to one even more
bulky (R = But).
while it is reacting; there is no need to postulate scrambling
in the rearrangement of 4 to 5 (R = PhMeCH) or to exclude
a concerted mechanism.
The O-sulfonyl derivatives 1 (R = alkyl or phenyl) of N-
phosphinoylhydroxylamines react with alkoxides or aliphatic
amines to give products 3 (X = OMe, NHBut, etc.) in which
a phenyl group has migrated from phosphorus to nitrogen.1 It
now seems that this migration is just one half of a transposition
reaction, the other half being migration of the sulfonate group
from nitrogen to phosphorus.2 The result is a phosphonamidic-
sulfonic anhydride 2 which reacts rapidly with the nucleophile
(HX) to give the product 3.
The tert-butylphosphonamidic chloride 3 (R = But, X = Cl) (dP
53.5; M+ 231,233) was prepared from ButP(O)Cl2 and LiNHPh
in THF at −40 ◦C and was hydrolysed using aqueous NaOH
or H218O–Et3N.† The resulting phosphonamidate 7 (R = But)
could be sulfonylated without much pyrophosphonate formation
by addition of BnSO2Cl [unlabelled or labelled (57.5 mol% one
18O atom)3] to a suspension of the Et3NH+ salt in diethyl ether.‡
Samples of the anhydride 5 (R = But) were thus obtained
In principle, the anhydride intermediate could be attacked at
sulfur instead of phosphorus but in only one case (Scheme 1,
R = PhMeCH) has appreciable competition been seen.3 That
observation opened the way to an investigation of the trans-
position mechanism using the hydroxylamine derivative 4 (R =
PhMeCH) labelled specifically with 18O in the sulfonyl (SO2)
group. By measuring the distribution of the label between the
products 6 and 7 resulting from attack at sulfur, the distribution
of 18O (SO2 vs. SOP) in the anhydride intermediate 5 could, we
supposed, be deduced.3 The results, however, were not consistent
◦
containing no label [mp 142–143 C; m/z (FAB) 367 (M+) and
368 (MH+)] and with 18O in the SO2 group (sample A) [57.5 mol%
one 18O (FAB MS); dP 36.4 only (no high field P–18O peak)] or in
the phosphonate group (sample B) [95.5 mol% one 18O; dP 36.3
(ca. 5%) with much larger peaks at higher field, DdP 0.034 (P–18O)
and 0.047 ppm (P 18O), ratio ca. 1 : 1].
=
Reaction of the labelled anhydride 5 (R = But) (sample A or B)
◦
with ButNH2 (50 equiv.; 2.0 mol dm−3 in CH2Cl2) at 30 C was
complete inside 0.5 h (31P NMR). The dominant products were
the phosphonic amide 9 (R = But) (dP 30.7) and the sulfonate 8
(dH 4.05) resulting from attack at phosphorus (ca. 90%), but minor
amounts of the sulfonamide 6 (dH 4.25) and the phosphonamidate
7 (R = But) (dP 26.5) were also formed as a result of attack at
sulfur. The salts (7 + 8) were removed by aqueous extraction and
esterified (CH2N2) and both the esters and the amides (6 + 9) were
analysed by GC–MS to determine their 18O content (Table 1).§
The results for the sulfonate and phosphonic amide (attack at P)
are as expected, reflecting (within experimental error) the content
and distribution of 18O in the anhydride, but the results for the
sulfonamide and phosphonamidate (attack at S) are not. A faithful
reflection of the labelling in the anhydride would give sulfonamide
6 with 57.5 mol% one 18O from sample A, not 45 mol%, and with
no 18O from sample B, not 20.5 mol%. Also, the phosphonamidate
from sample A would not contain any 18O. There must, it seems,
Scheme 1
Department of Chemistry, University of Leicester, Leicester, UK LE1
7RH. E-mail: mjph2@le.ac.uk; Fax: +44(0)116 2523789; Tel: +44(0)116
2522127
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 1863–1864 | 1863
©

View Article Online
Table 1 18O Content (mol% one 18O atom) of products from reaction of
18O-labelled phosphonamidic-sulfonic anhydride 5 (R = But) (sample A
the anhydride) and 7 (R = But) in much the same way even though
the label is all in the SO2 group of the anhydride to begin with. It
therefore seems likely that our earlier inference was wrong and that
the anhydride intermediate 5 (R = PhMeCH) is actually formed
with all the label in the SO2 group. If that is so, the transposition
reaction 4 → 5 (conjugate bases) must surely be concerted with a
transition state resembling 10. The alternative transition state 11
can be discounted, as can a non-concerted mechanism in which
the three sulfonate O atoms become equivalent; in neither case
could the resulting anhydride produce sulfonamide containing
more than two-thirds of the available label.
a
or B) with ButNH2 (2 mol dm−3) in CH2Cl2
Sample A
Sample B
Sulfonamide 6
45
20.5
70b
49
Phosphonamidate 7 (R = But)
Sulfonate 8
9.5b
57.5
0
Phosphonic amide 9 (R = But)
47
a Sample A, label (57.5 mol% one 18O) in SO2; sample B, label (95.5 mol%
one 18O) shared equally between P–O and P O; products 7 and 8 analysed
=
as methyl esters. b Lower-than-expected 18O content of 7 is attributable to
traces of moisture [7 is only a minor product from 5 + ButNH2 (attack at
S) but a major product from 5 + H2O (attack at P; cleavage of the P–18O
bond)].
be some scrambling of the label between the bridging (SOP)
and non-bridging (SO2) positions of the anhydride while it is
reacting.
Scrambling could easily be explained if the sulfonate anion
released during the reaction were to displace the sulfonate leaving
group from the anhydride yet to react. Such simple exchange
seems not to occur, however, since labelled anhydride (sample B)
suffers no change when dissolved in CDCl3 containing unlabelled
The assistance of Jaswinder Gill with preliminary experiments
is gratefully acknowledged.
sulfonate (ButNH3 salt; 1 equiv.) (31P NMR: size of P–18O peak
Notes and references
+
unchanged after 25 h at 30 ◦C). Any scrambling clearly occurs only
when the amine is present and must, we think, be coupled with
the product-forming reactions of the anhydride. For substitution
at phosphorus, the structure of the anhydride (bulky alkyl group,
acidic NH group) and the nature of the nucleophile (ButNH2) will
hinder SN2(P) but will assist an elimination–addition mechanism
(Scheme 2). In this, a reactive three-coordinate metaphosphonimi-
date is generated by the amine acting as a base and is trapped by the
amine acting as a nucleophile.6,7 If sulfonate anion competes with
amine for the metaphosphonimidate so that some of it returns, any
18O in the sulfonate anion will be shared between the bridging and
non-bridging positions of the resulting anhydride. The sulfonate
anion (1 equiv. at most) is unlikely to compete effectively in the
bulk solution where the amine (50 equiv.) is in large excess; more
likely is some direct recombination of the metaphosphonimidate
and the sulfonate leaving group before they diffuse apart.
† Two equivalents of LiNHPh are required because the phosphonamidic
chloride is acidic (NH) and cooling must be continued until the reaction has
been quenched (CF3CO2H) because the conjugate base (NLi) is of limited
stability (elimination of LiCl). The phosphonamidic chloride (0.4 mmol)
can be hydrolysed in CHCl3 (1.5 ml) by addition of H218O (2 equiv.) and
then Et3N (0.8 ml), with vigorous stirring for 4 h at 30 ◦C; in this way
pyrophosphonate formation is largely avoided without the need of a large
excess of H218O.
‡ A small excess of BnSO2Cl (1.2 equiv.) was used and a little Et3N
(0.4 equiv.) was added after 3–4 min. Because chloride ion (from BnSO2Cl)
tends to displace sulfonate from the anhydride (attack at P), diethyl ether
was used as the solvent (Et3NHCl precipitates out) and the reaction was
quenched (slightly acidic ice-cold water) after just 5 min. The product
(a foam, initially) was purified by washing with warm light petroleum
and diethyl ether and crystallisation from CH2Cl2–diethyl ether (1 : 8);
dH(CDCl3) 7.5–7.0 (10 H), 5.40 (d, JPH = 10 Hz, NH), 4.67 (AB quartet,
dA 4.71, dB 4.63, JAB = 14 Hz, CH2Ph) and 1.23 (d, JPH = 19 Hz, But);
dC(CDCl3) 140–120, 60.1 (s), 35.0 (d, JPC = 125 Hz) and 24.3 (s).
§ Mass spectra were recorded in EI mode. The proportion of molecules
containing the 18O label was determined from the abundance of (M + 2)+
ions (relative to M+) corrected for the contribution of ions containing 16
O
and natural abundance 18O, 34S or 13C (2 atoms). The molecular ion was
of very low abundance in the case of sulfonamide 6 and (M+ − Me) was
used.
¶ A similar study using 4 (R = But) is not possible because the requisite
N-phosphinoylhydroxylamine [ButPhP(O)NHOH] cannot be prepared
(steric hindrance).
Scheme 2
1 M. J. P. Harger, J. Chem. Soc., Perkin Trans. 1, 1983, 2699; M. J. P.
Harger and A. Smith, J. Chem. Soc., Perkin Trans. 1, 1987, 683.
2 M. J. P. Harger and A. Smith, J. Chem. Soc., Perkin Trans. 1, 1990, 2507;
M. J. P. Harger and R. Sreedharan-Menon, J. Chem. Soc., Perkin Trans.
1, 1994, 3261.
In our earlier study the 18O label was confined to the SO2 group
of the hydroxylamine derivative 4 (R = PhMeCH) but was shared
in the product between the sulfonamide 6 (76% of the available
18O) and the phosphonamidate 7 (R = PhMeCH).3 ¶ We inferred
that the label in the anhydride intermediate 5 (R = PhMeCH) was
shared in the same way between the SO2 group and the bridging
O atom but then struggled to relate the labelling pattern to a
reasonable mechanism for the rearrangement of 4 to 5. In the
present study the similar anhydride 5 (R = But) (sample A) gives
products in which the label is shared between 6 (78% of the 18O in
3 M. J. P. Harger, Org. Biomol. Chem., 2003, 1, 3390.
4 J. Wasiak and J. Michalski, Tetrahedron Lett., 1994, 35, 9473.
5 W. Dabkowski, J. Michalski, C. Radziejewski and Z. Skrzypczynski,
Chem. Ber., 1982, 115, 1636; J. Michalski, W. Dabkowski and J. Wasiak,
Pol. J. Chem., 1992, 66, 879.
6 P. S. Traylor and F. H. Westheimer, J. Am. Chem. Soc., 1965, 87, 553.
7 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.
1864 | Org. Biomol. Chem., 2006, 4, 1863–1864
This journal is
The Royal Society of Chemistry 2006
©