Reactions of bis(2,4-dinitrophenyl) phosphate with hydroxylamine

Article abstract of DOI:10.1021/jo034348z

For dephosphorylation of bis(2,4-dinitrophenyl) phosphate (BDNPP) by hydroxylamine in water, pH region 4-12, the observed first-order rate constant, k_obs, initially increases as a function of pH, but is pH-independent between pH 7.2 and pH 10. The initial BDNPP cleavage by nonionic NH 2OH (<0.2 M) involves attack by the OH group and follows first-order kinetics, but the overall initial reaction of BDNPP liberates ca. 1.7 mol of 2,4-dinitrophenoxide ion (DNP). This initial reaction generates a short-lived O-phosphorylated hydroxylamine, 2, followed by three possible reactions: (1) reaction of 2 with hydroxylamine, generating 2,4-dinitrophenyl phosphate (DNPP, 3), which subsequently forms DNP; (2) intramolecular displacement of the second DNP group and rapid decomposition of the cyclic intermediate to form phosphonohydroxylamine and eventually inorganic phosphate; (3) a novel rearrangement with intramolecular aromatic nucleophilic substitution involving a cyclic intermediate and migration of the 2,4-dinitrophenyl group from O to N. Values of k_obs increase modestly with pH > 10, the reaction is biphasic, and the yield of DNP increases. An increase in [NH₂OH] also increases the yield of DNP, due largely to accelerated hydrolysis of DNPP.

Full text of DOI:10.1021/jo034348z

Rea ction s of Bis(2,4-d in itr op h en yl) P h osp h a te w ith
Hyd r oxyla m in e
J osiel B. Domingos,† Elisane Longhinotti,† Clifford A. Bunton,‡ and Faruk Nome*,†
Departamento de Qu´ımica, Universidade Federal de Santa Catarina,
Floriano´polis, SC 88040-900, Brazil, and Department of Chemistry and Biochemistry, University of
California, Santa Barbara, California 93106-9510
faruk@qmc.ufsc.br
Received March 17, 2003
For dephosphorylation of bis(2,4-dinitrophenyl) phosphate (BDNPP) by hydroxylamine in water,
pH region 4-12, the observed first-order rate constant, kobs, initially increases as a function of pH,
but is pH-independent between pH 7.2 and pH 10. The initial BDNPP cleavage by nonionic NH2-
OH (<0.2 M) involves attack by the OH group and follows first-order kinetics, but the overall initial
reaction of BDNPP liberates ca. 1.7 mol of 2,4-dinitrophenoxide ion (DNP). This initial reaction
generates a short-lived O-phosphorylated hydroxylamine, 2, followed by three possible reactions:
(1) reaction of 2 with hydroxylamine, generating 2,4-dinitrophenyl phosphate (DNPP, 3), which
subsequently forms DNP; (2) intramolecular displacement of the second DNP group and rapid
decomposition of the cyclic intermediate to form phosphonohydroxylamine and eventually inorganic
phosphate; (3) a novel rearrangement with intramolecular aromatic nucleophilic substitution
involving a cyclic intermediate and migration of the 2,4-dinitrophenyl group from O to N. Values
of kobs increase modestly with pH > 10, the reaction is biphasic, and the yield of DNP increases. An
increase in [NH2OH] also increases the yield of DNP, due largely to accelerated hydrolysis of DNPP.
In tr od u ction
the reaction with 4-nitrophenyl acetate involves attack
of the hydroxyl group of neutral hydroxylamine with
concerted proton transfer to the amino group and inter-
action with putative anionoid oxygen in the transition
state. This interaction probably assists reactions of the
OH group of NH2OH with other activated esters, leading
to the unexpectedly high nucleophilicity of NH2OH in
terms of Brønsted relationships.
Hydroxylamine is a typical R-effect nucleophile with
unshared electron pairs on oxygen and nitrogen. We were
interested in the possibility that the reaction of NH2OH
with an activated bisaryl phosphate might release more
than one aryloxide ion in the course of reaction. Recently
the reaction of bis(4-nitrophenyl) phosphate anion with
alkaline hydrogen peroxide was shown to involve simul-
taneous liberation of more than 1 equiv of 4-nitrophe-
noxide ion (NP)/mol of substrate, with no evidence for
buildup of an intermediate, and it was suggested that
the first-formed transient peroxophosphate ionic inter-
Mechanistic studies of hydrolyses and nucleophilic
substitutions on phosphodiesters are extensive, and such
reactions are involved in metabolic processes that are still
not fully understood at a molecular level. Reactions of
hydroxylamine with activated acyl derivatives and aryl
sulfonic and phosphate esters involve attack by the
hydroxyl group, although the first-formed products may
subsequently decompose.1-4
Nucleophilicities toward these centers typically fit
Brønsted relationships with linear plots of log k against
pKa of the nucleophile, but the nucleophilicity of hydroxy-
lamine in its reaction with 4-nitrophenyl acetate fits the
Brønsted relationship with pKa ) 6.22, i.e., that of the
amino group.2,3 This evidence with a simple nucleophile
could indicate that NH2 is the attacking group, but
instead deprotonation of the ammonium residue gener-
ates the R-nucleophile, therefore increasing the nucleo-
philicity of the OH group by orders of magnitude, and
the fit to the Brønsted relationship appears to be fortu-
itous. Hess et al.5 used kinetic isotope effects to show that
mediate (Int
1) reacted with H2O2, generating monoester,
or intramolecularly giving the second aryloxide ion
(Scheme 1).6
A similar reaction scheme can be envisaged for the
reaction of NH2OH with a transient intermediate, Int2,
formed by initial attack of NH2OH, which could decom-
pose to the ionic monoester, or lose a second aryloxide
ion (Scheme 2).
† Universidade Federal de Santa Catarina.
‡ University of California.
(1) (a) J encks, W. P. J . Am. Chem. Soc. 1958, 80, 4581. (b) J encks,
W. P. J . Am. Chem. Soc. 1958, 80, 4585.
(2) J encks, W. P. Catalysis in Chemistry and Enzymology; McGraw-
Hill: New York, 1969.
(3) J encks, W. P.; Carriulo, J . J . Am. Chem. Soc. 1960, 82, 1778.
(4) Behrman, E. J .; Bialles, M. J .; Braas, H. J .; Edwards, J . O.; Isaks,
M. J . Org. Chem. 1970, 35, 3069.
Either of the reactions of Int2 could involve interven-
tion by NH2OH, and we must also consider attack of NH2-
(5) Hess, R. A.; Hengge, A. C.; Cleland, W. W. J . Am. Chem. Soc.
1997, 119, 6980.
(6) Mej´ıa-Radillo, Y.; Yatsimirsky, A. K.; Foroudian, H. J .; Gillitt,
N. D.; Bunton, C. A. J . Phys. Org. Chem. 2000, 13, 505.
10.1021/jo034348z CCC: $25.00 © 2003 American Chemical Society
Published on Web 08/15/2003
J . Org. Chem. 2003, 68, 7051-7058
7051

Domingos et al.
to aqueous amine hydrochloride, by borate buffer (0.005 M)
from pH 8 to pH 10, and by NaOH at higher pH.
SCHEME 1
All reactions were done at 25 °C, followed by the appearance
of DNP at 400 nm on a diode-array spectrophotometer with a
thermostated cell holder. The pH of each reaction mixture was
measured at the end of each run. Observed first-order rate
constants (kobs) were calculated through a nonlinear least-
squares fitting of the absorbance vs time with the use of UV-
vis ChemStation software. Second-order rate constants were
obtained from the slopes of plots of the observed first-order
rate constants against the concentration of the nucleophiles
with the Origin 5.0 program.
SCHEME 2
All NMR spectra were obtained in D2O at 25 °C, generally
with a delay time of 1 s. Relaxation is slow at some positions
in nitroarenes, and there were increases of 15-20% in the
integrated areas of some of the signals at complete reaction
when the delay time was increased to 20 s, but these long
delays create problems in examining the spectra in the course
of the reaction. Most of the 31P NMR measurements were
carried out with a delay time of 1 s, but integrated areas did
not change when the delay time was increased to 3 s in
measurements done at the end of the reaction. Both pyri-
dinium and sodium salts were used in the NMR work, but 1H
signals of the pyridinium ion complicate the spectra, and the
later experiments discussed here were made with the sodium
salt. However, both samples gave identical signals in the
spectral regions where there is no interference. The 1H and
13C chemical shifts are referred to internal sodium 3-(trim-
ethylsilyl)propionate (TSP), and those of 31P are referred to
external 85% phosphoric acid. The value of pD was obtained
by adding 0.4 to the observed pH of the solutions in D2O at
25 °C.13
OH on an initially formed monoester, either on phospho-
rus or on the aryl group, since even the less reactive
4-nitrophenyl phosphate dianion reacts with NH2OH
with a rate constant of 3.33 × 10-7 M-1 s-1 7a
.
We chose bis(2,4-dinitrophenyl) phosphate (BDNPP) as
substrate, because 2,4-dinitrophenoxide ion (DNP) is a
good leaving group and there is extensive evidence of the
hydrolysis of 2,4-dinitrophenyl phosphate (DNPP), in-
cluding rate enhancements by tertiary amines.7,8 Most
of our work involved NH2OH, but we have also measured
rate constants of initial reactions with NHMeOH, NMe2-
OH, and MeONH2 with the aim of distinguishing between
initial phosphorylation on oxygen or nitrogen, and com-
plete results of this work will be given elsewhere. Most
of our work was in the pH range 4-12 (should this be,
10k increases at pH > 10) where nonionic hydroxylamine
is dominant9 and the spontaneous hydrolysis of BDNPP
is slow.10 At high pH NH2O- (pKa ) 13.74) is a very
strong nucleophile,11 but we did not investigate reaction
products in detail at high pH because of the instability
of the organic products in basic solutions.
P r od u cts of th e Rea ction of Hyd r oxyla m in e. Most of
the products and unreacted BDNPP were identified by their
absorption and NMR spectra, with comparison with those of
authentic material. BDNPP and DNPP (designated 1 and 3,
respectively) were prepared as described above, 2,4-dinitro-
aniline and 2,4-dinitrophenol were purchased and used at
the appropriate pH, and O-phosphorylated hydroxylamine,
2-
H2NOPO3
(designated 5), was prepared by reaction of
potassium phosphoramidate with hydroxylamine.14
N-(2,4-Din itr oph en yl)-O-ph osph on oh ydr oxylam in e. The
monosodium salt of N-(2,4-dinitrophenyl)-O-phosphonohy-
droxylamine (designated 4) was isolated from a reaction
mixture containing BDNPP (sodium salt) (0.5 g) rapidly
dissolved in ∼50 mL of aqueous 0.1 M hydroxylamine, pH 8.5.
The resulting solution was held at 25 °C overnight. The pH
was decreased to ca. 0 by adding approximately 20 mL of 5 M
HCl, the reaction mixture was cooled (ice bath), and 2,4-
dinitrophenol was removed by filtration. The filtrate contain-
ing 4, as the free acid, was concentrated (rotary evaporator,
∼35 °C), NaCl was precipitated by addition of dry ether, and
after filtration, the ether solution was stirred while a dilute
solution of sodium ethoxide in ethanol was added dropwise,
giving precipitation of a pale yellow solid. This solid was
collected and washed with ether. EI-MS (70 eV): m/z (relative
intensity) 197 (15), 182 (25), 181 (40), 168 (10), 167 (7), 91 (25),
75 (100), 74 (90), 63 (70). Anal. Calcd for C6H5N3O8PNa: C,
23.9; H, 1.66; N, 13.9. Found: C, 24.0; H, 1.77; N, 14.0. The
NMR spectrum of the isolated material was identical with that
observed in the reaction mixture.
Exp er im en ta l Section
Ma ter ia ls. BDNPP as the pyridinium salt was prepared
as descibed.10 The pyridinium ion was exchanged for sodium
ion on a cation-exchange resin (Dowex 50W X8) in the Na+
form. DNPP as the pyridinium salt was prepared by the
procedure of Rawji and Milburn.12 The hydroxylamines, as
their hydrochlorides, other nucleophiles, and 2,4-dinitrophenol
were of the highest purity available and were used as
purchased.
Kin etics. Reactions were started by adding a 30 µL stock
solution of the substrate (10-3 M) in water to 3 mL of the
reaction mixture, which contained a large excess of the
nucleophile (0.1 M), ensuring strictly first-order kinetics for
the initial nucleophilic attack upon the substrate. Solutions
were self-buffered by amine/amine hydrochloride at pH 4.0-
7.0, prepared by addition of aqueous standard NaOH (0.1 M)
The isolated sodium salt of 4 is a substrate for calf intestinal
alkaline phosphatase (activity 1 unit/µL). The reactivity of 4
was tested by using 5 µL of the phosphatase, in 50 mM Tris-
HCl, pH 9.3, 1 mM MgCl2, 0.1 mM ZnCl2, and 1 mM
spermidine in a final volume of 0.2 mL. The reaction mixture
was kept at room temperature for 14 h, and inorganic
phosphate was detected following a previously described
procedure.15
(7) (a) Kirby, A. J .; Varvoglis, A. G. J . Chem. Soc. B 1968, 135. (b)
Kirby, A. J .; Varvoglis, A. G. J . Am. Chem. Soc. 1967, 89, 415. (c) Kirby,
A. J .; Varvoglis, A. G. J . Am. Chem. Soc. 1965, 87, 3209.
(8) Bunton, C. A.; Fendler, E. J .; Fendler, J . H. J . Am. Chem. Soc.
1967, 89, 1221.
(9) Hughes, M. N.; Nicklin, H. G.; Schrimankes, K. J . Chem. Soc. A
1971, 3485.
(10) Bunton, C. A.; Farber, S. J . J . Org. Chem. 1969, 34, 767.
(11) Simanenko, Y. S.; Popov, A. F.; Prokop’eva, T. M.; Savyolova,
V. A.; Belousova, I. A.; Zubareva, T. M. Mendeleev Commun. 1994,
210.
(13) Fife, T. H.; Bruice, T. C. J . Phys. Chem. 1961, 65, 1079.
(14) J encks, W. P.; Gilchrist, M. J . Am. Chem. Soc. 1965, 87, 3199.
(12) Rawji, G.; Milburn, R. M. J . Org. Chem. 1981, 46, 1205.
7052 J . Org. Chem., Vol. 68, No. 18, 2003

Reactions of Diaryl Phosphate with Hydroxylamine
F IGURE 1. k1 (9) and k2 (O) as functions of [NH2OH] for
nucleophilic attack on BDNPP, at pH 9 and 25 °C. Inset:
kinetic plots for (a) [NH2OH] ) 0.1 M and (b) [NH2OH] )
0.4 M.
F IGURE 2. Logarithmic plots of kobs for nucleophilic attack
by NH2OH on BDNPP (9) and H2O ([), at 25 °C.
constant (kobs) could be separated into two rate constants
(k1 and k2), by fitting the formation of DNP to eq 1. Both
reactions were linearly dependent on [NH2OH], and with
increasing [NH2OH] the yield of DNP increased up to
approximately 200 mol % (Table 1).
TABLE 1. Yield of DNP in th e Rea ction of NH2OH w ith
BDNP P , a t p H 9 a n d 25 °C
yield of
DNP,
mol %
yield of
DNP,
mol %
yield of
DNP,
mol %
[NH2OH],
M
[NH2OH],
M
[NH2OH],
M
0.10
0.20
173
179
0.40
0.60
191
197
0.80
199
k1
[DNP] ) [BDNPP]0 1 - e-k1t
-
(e-k1t
-
{
k2 - k1
e
-k2t) + [DNPP]0(1 - e-k2t) + [DNP]0 (1)
P r od u cts of th e Rea ction of O-Meth ylh yd r oxyla m in e.
An authentic sample of N-(2,4-dinitrophenyl)-O-methylhy-
droxylamine was prepared from the reaction of 1 mmol of 2,4-
dinitrofluorobenzene with 3 mmol of O-methylhydroxylamine,
in water at pH 7. The solvent was evaporated and a pale yellow
solid collected by filtration, mp 110-112 °C. EI-MS (70 eV):
m/z 213 (M+). Anal. Calcd for C7H7N3O5: C, 39.4; H, 3.3; N,
19.7. Found: C, 39.5; H, 3.5; N, 19.8. The NMR spectrum of
the synthetic sample was identical with that observed in the
reaction mixture. The NMR spectra of the sodium salt of the
O-methyl-N-(2,4-dinitrophenyl)phosphonohydroxylamine iso-
lated from the reaction mixture was shown to be identical with
that of the reaction mixture. Anal. Calcd for C7H7N3O8PNa:
C, 35.0; H, 2.94; N, 17.5. Found: C, 34.9; H, 3.0; N, 17.6.
Geom etr ica l Op tim iza tion . Theoretical calculations were
performed by using Spartan ‘02 for Windows. Optimizations
were carried out at the PM3 level followed by ab initio
optimizations with the HF 6-31+G* basis set.16 Cartesian
coordinates of the structures presented in the text are given
in Table S1 of the Supporting Information. Hydration energies
were calculated by the semiempirical SM 5.4 treatment of
Cramer and Truhlar.17
}
The data shown in Figure 2 correspond to those for
the initial reaction of BDNPP with NH2OH, and in all
cases the reaction is overall second-order (Figure 1). The
data for the spontaneous hydrolysis of BDNPP are
included for comparison purposes and in fitting the
experimental data, although contributions of the spon-
taneous and basic hydrolyses are minor in the pH 4-10
region of interest.10 The subsequent slow reaction, ob-
served at high pH or higher [NH2OH], corresponds to
reactions of DNPP with NH2OH, NH2O-, or H2O/OH-,
which eventually give the observed 200 mol % yield of
DNP.
The pH rate constant profile for the initial reaction
between BDNPP and hydroxylamine, shown in Figure
2, was fitted by using eq 2, where ko is the rate constant
k
NH2OH[NH2OH]
kobs ) ko + kOH[OH-] +
+
[H+]
Ka1
Ka2
[H+]
1 +
+
Resu lts a n d Discu ssion
Kin etics. Reactions of BDNPP with 0.1 M NH2OH,
monitored by the increasing absorbance at 400 nm at pH
4-12, were first-order for over 3 half-lives, up to pH 11,
and in all cases more than 1.5 mol of DNP was liberated.
However, the reaction is clearly biphasic in some condi-
tions. At pH 12, or higher [NH2OH], the kinetics have
the typical profile of two consecutive reactions, as shown
by curve b in the inset of Figure 1, and the observed rate
k
NH2O-[NH2OH]
(2)
[H+]
1 +
Ka2
for the spontaneous reaction, kOH is that for reaction with
OH-, and kNH2OH and kNH2O are rate constants for
-
reaction with hydroxylamine and its anion. The acid
dissociation constants Ka1 and Ka2 correspond to forma-
tion of nonionic hydroxylamine (a ) and its anionic form
(b), respectively (Scheme 3). Values of the individual rate
and acid dissociation constants are given in Table 2.
(15) Heinonen, J . K.; Lahti, R. J . Anal. Biochem. 1981, 113, 313.
(16) Cramer, C. J . Essentials of Computational Chemistry; J ohn
Wiley & Sons: New York, 2002; Chapter 6.
(17) Cramer, C. J .; Truhlar, D. G. Chem. Rev. 1999, 99, 2161.
J . Org. Chem, Vol. 68, No. 18, 2003 7053

Domingos et al.
F IGURE 3. 1H NMR spectrum of the reaction mixture of 0.01 M BDNPP with 0.1 M NH2OH, at pH 9 and 25 °C.
SCHEME 3
the difference in substrate concentration did not affect
the rate constants and formation of DNP. These concen-
1
trations gave good signal-to-noise ratios for H or 31P, but
because products were changing with time, it was dif-
ficult to use extensive accumulations, especially for the
31P spectra, where weak signals could have been lost in
the noise (Experimental Section). Detection of the 1H
signals was less of a problem, but some of the nuclei,
especially that of H-3 of the dinitroarenes, relax very
slowly, and because the reaction was progressing, we
could not use delay times long enough to ensure linearity
between concentrations and all signal areas (Experimen-
tal Section). Figure 3 shows 1H NMR spectra as a
function of time for the reaction mixture of 0.01 M
BDNPP with 0.1 M NH2OH, and Figure 4 shows the 31P
NMR spectra.
TABLE 2. Ra te a n d Dissocia tion Con sta n tsa Used in th e
Kin etic F ittin g
K0, s-1
k
1.90 × 10-7 kNH2OH, M-1 s-1 3.96 × 10-3 pKa1 6.19
OH, M-1 s-1 2.92 × 10-3
k
NH2O-, M-1 s-1 0.407
pKa2 13.74
a
Dissociation constants are from NIST Standard Reference
Database 46, Version 6.0, distributed by NIST Standard Reference
Data, Gaithersburg, MD 20899.
Although, the rate constants show that the anionic form
of hydroxylamine reacts ca. 100 times faster than the
neutral form, we did not investigate this reaction in detail
due to instability of the organic products. Indeed, experi-
ments at high pH showed, as previously reported in the
absence of hydroxylamine, that 2,4-dinitrophenol is
unstable in basic solutions,10 nitro groups are removed
from the arene ring, and nitrite ion is detected in solution.
We have preliminary kinetic data on the reactions of
BDNPP with the different nonionic methyl derivatives
NHMeOH, NMe2OH, and MeONH2. The results in the
pH-independent region are consistent with little rate
effect of methylation on nitrogen (rate constants are 1.0
× 10-3 and 2.4 × 10-4 s-1 for NHMeOH and NMe2OH,
respectively), but methylation on oxygen strongly inhibits
reaction (kobs ) ca. 1.4 × 10-5 s-1). Rate profiles and
products of dephosphorylation by these N- and O-methyl
derivatives will be discussed in detail elsewhere, but the
kinetics fit our assumption that the initial reaction of
NH2OH is on oxygen.
Our NMR spectra allow identification of the products,
and some of the intermediates, in particular compound
2, formed by the initial attack of NH2OH. They show that
a long-lived species is formed, probably from 2, which
1
does not decompose in the course of the reaction. The H
chemical shifts and multiplicities of the signals indicate
that the 2,4-dinitrophenyl group is always present in the
arene derivatives; i.e., nitro groups are not displaced by
nucleophilic attack at pH < 10 (Table 3).
The NMR spectra in Figures 3 and 4 illustrate,
qualitatively, the changing concentrations of products
and intermediates with time, and Figure 5 gives ex-
amples of the results, among many observations. For
convenience we designate the various species as 1-5,
where 1 is the starting material BDNPP. We used 1H
and 31P to distinguish between species that retain dini-
trophenyl and phosphoryl residues and inorganic phos-
phoryl derivatives. As shown in both panels, at [NH2OH]
) 0.1 and 0.6 M, the disappearance of starting material
1 is associated with the appearance of about 170 mol %
aryloxide ion DNP, and we always detected the transient
species 2, which disappears with time. The rate of
disappearance of 2 depends on [NH2OH], and this reac-
tion generates 3 (DNPP), whose 1H and 31P NMR spectra
are identical with those of authentic DNPP, as well as a
stable product, 4. It is interesting to note that as the
P r od u cts. It is important to note that the yield of DNP
in the first step of the reaction of BDNPP with 0.1 M
NH2OH, estimated spectrophotometrically, is approxi-
mately 1.7 equiv (Table 1); i.e., more than one molecule
of DNP is formed in the first kinetically observed step of
the reaction, indicating a complex reaction.
1
Most of the products were also identified by H, 13C,
and 31P NMR spectroscopy. The conditions, of necessity,
differ from those of the kinetic work in the use of D2O as
solvent and higher substrate concentrations. However,
7054 J . Org. Chem., Vol. 68, No. 18, 2003

Reactions of Diaryl Phosphate with Hydroxylamine
F IGURE 4. 31P NMR spectrum of the reaction mixture of 0.01 M BDNPP with 0.1 M NH2OH, at pH 9 and 25 °C.
TABLE 3. NMR Da ta for th e P r od u cts (Sch em e 4) of th e Rea ction of BDNP P (0.01 M) w ith NH2OH (0.1 M), in D2O, a t
p D 9 a n d 25 °C
31P NMR δ (ppm)
compd
DNP
1H NMR δ (ppm)
(relative integrated area, %)a
13C NMR δ (ppm)
6.74 (d, 1H, J ) 9.6 Hz, Ar), 8.11 (dd, 1H, J ab ) 9.6 Hz and
J bx ) 3.0 Hz, Ar), 8.90 (d, 1H, J ) 3.0 Hz, Ar)
4
2
1
3
7.62 (d, 1H, J ) 9.6 Hz, Ar), 8.38 (dd, 1H, J ab ) 9.6 Hz and
J bx ) 2.7 Hz, Ar), 9.10 (d, 1H, J ) 2.7 Hz, Ar)
6.43 (0)
-2.04 (51)
-13.17 (0)
0.04 (27)
7.78 (22)
C1, 151.2; C2, 133.3; C3, 126.5;
C4, 140.0; C5, 139.5; C6, 118.5
7.75 (d, 1H, J ) 9.1 Hz, Ar), 8.56 (dd, 1H, J ab ) 9.1 Hz and
J bx ) 2.7 Hz, Ar), 8.92 (d, 1H, J ) 2.7 Hz, Ar)
7.80 (d, 1H, J ) 9.1 Hz, Ar), 8.58 (dd, 1H, J ab ) 9.1 Hz and
J bx ) 2.9 Hz, Ar), 8.94 (d, 1H, J ) 2.9 Hz, Ar)
7.86 (d, 1H, J ) 9.4 Hz, Ar), 8.48 (dd, 1H, J ab ) 9.4 Hz and
J bx ) 3.0 Hz, Ar), 8.84 (d, 1H, J ) 3.0 Hz, Ar)
5
a
Relative integrated area of the signals of 1 and compounds identified at complete reaction.
initial [NH2OH] is increased the amount of 3 increases
and that of 4 decreases, showing that these decomposi-
tions of 2 have different dependences on [NH2OH]. With
time, in a reaction that occurs after complete disappeance
of starting material 1 (BDNPP), but corresponds to the
slow increase in absorbance (curve b in the inset of Figure
1), and is linearly dependent on [NH2OH], 3 disappears,
generating aryloxide DNP. This relatively slow decom-
position of DNPP is well studied.7,8
parison of its NMR spectrum with that of authentic
material (Experimental Section) and with the litera-
ture.14,18 Inorganic phosphate (Pi) is observed only after
a long time of reaction (small amounts were detected at
1200 min), and it is probably formed in the spontaneous
hydrolysis of DNPP.
We first consider products of the initial reaction of
BDNPP with NH2OH on the assumption that it is
phosphorylation on oxygen giving DNP and the transient
intermediate 2, which breaks down in three ways.
(i) The first competing reaction involves a novel type
of intramolecular rearrangement, which is somewhat
The 31P NMR chemical shifts, and the relative inte-
grated areas (Table 3), are fully consistent with the
reactions shown in Scheme 4, indicating that at the end
of reaction (ca. 253 min) compounds 3-5 are the main
detectable phosphorus-containing products. Compound 5
was identified as O-phosphonohydroxylamine, by com-
(18) Weiss, P. M.; Hermes, J . D.; Dougherty, T. M.; Cleland, W. W.
Biochemistry 1984, 23, 4346.
J . Org. Chem, Vol. 68, No. 18, 2003 7055

Domingos et al.
F IGURE 5. Relative concentrations of products of the reaction of BDNPP (0.01 M) with (a) [NH2OH] ) 0.1 M and (b) [NH2OH]
) 0.6 M, in D2O at pD 9 and 25 °C: (9) 1, (4) DNP , ([) 3, (1) 4, and (O) 2.
SCHEME 4
TABLE 4. 1H NMR Ch em ica l Sh ifts, p p m , of 4 a s a
similar to both the Smiles rearrangement and the forma-
tion and breakdown of cyclic Meisenheimer complexes.
The reaction involves nucleophilic attack on the arene
ring and is favored by the strong electron withdrawal by
the nitro groups, giving the phosphate ester 4 as one of
the products (Scheme 4). However, the Smiles rearrange-
ment has not been observed with a nitroaryl phosphoryl
derivative, and the formation and breakdown of cyclic
Meisenheimer complexes typically involve alkoxy or
aryloxy derivatives.19 The dependence of the yield of 4
on [NH2OH] allowed us to isolate and fully characterize
this product at low [NH2OH] (Experimental Section). The
mass spectrum of 4 did not show the parent M+ peak,
F u n ction of p D, in D2O a t 25 °C
pD 6
pD 12
Ha
Hb
Hx
7.62
8.38
9.12
7.30
7.52
8.85
because electron impact was used, but it had all the
expected fragments of the N-(2,4-dinitrophenyl)hydroxy-
lamine moiety, and elemental analysis is consistent with
the proposed structure. Addition of aqueous alkali solu-
tion to a solution of 4 promotes a reversible color change
1
associated with an upfield shift of the H NMR signals,
indicating deprotonation at NH (Table 4). Heating the
aqueous solution at 55 °C, pH 0, for 24 h gave only the
(19) March, J .; Smith, M. B. March’s Advanced Organic Chemistry,
5th ed.; J ohn Wiley & Sons: New York, 2001; Chapter 13.
7056 J . Org. Chem., Vol. 68, No. 18, 2003

Reactions of Diaryl Phosphate with Hydroxylamine
SCHEME 5
SCHEME 6
SCHEME 7
(Scheme 5), the transient phosphorylated hydroxylamine
2 can break down by intermolecular attack by NH2OH,
giving DNPP, nitrogen, and ammonia (Scheme 6). This
general reaction had been observed by J andorf, and
Harger reported decomposition of the hydroxylamine
moiety of phosphonylhydroxylamine by reaction with
potassium iodide.20,21
31P NMR signal of H3PO4, and as expected, 4 behaves as
a substrate for alkaline phosphatase (Experimental Sec-
1
tion). H NMR signals of the acid hydrolyzate indicated
the presence of several nitroarenes, 2,4-dinitroanilinium
ion of which was identified as the major product, by
comparison with authentic material, and it is probably
formed by acid hydrolysis of phosphate 4 and subsequent
degradation of the hydroxylamine derivative.
This reaction is similar to the well-known decomposi-
tion of peroxy acids by reaction with H2O2.22,23
(iii) In competition with the above reactions 2 may
generate DNP intramolecularly, or by attack of either
NH2OH or water, which is improbable because the initial
yield of DNP decreases with increasing [NH2OH] (Table
1). We therefore conclude that decomposition involves
intramolecular expulsion of DNP, via a hypothetical cyclic
intermediate, to form the phosphorylated hydroxylamine
5, identified by its 31P NMR signal at 7.78 ppm (Scheme
Our hypothesis that compound 4 is produced by the
intramolecular rearangement of the O-phosphorylated
hydroxylamine 2 (Scheme 4), in competition with the
formation of major products, e.g., 3, formed by reaction
of NH2OH with 2, fits all our experimental evidence. The
terminal NH2 group in 2 can readily attack the dinitro-
phenyl moeity to form a transient cyclic Meisenheimer
complex, 2a , higher in energy than the final product, and,
on the basis of the Hammond postulate, it should rapidly
ring open to give 4. In addition, the Meisenheimer
complex, 2a , must be short-lived, because we saw no
buildup of color or NMR signals typical of a Meisenhe-
imer complex in the course of reaction. Scheme 5 shows
the reaction scheme for the rearrangement, and the
structures are those with optimized geometries calculated
with the HF 6-31+G* basis set. The calculations indicate
that the proposed intermediate has an overall stability
consistent with its role in the reaction (26 kcal/mol
difference between 2 and 2a ), and the angles in both rings
of structure 2a are consistent with a species with little
torsional strain (Table S1, Supporting Information).
(ii) In addition to the rearrangement described above
7). This compound is also formed by reaction of NH2OH
2- 14
with phosphoramidate, H2NPO3
.
This reaction is analogous to that postulated for the
more than stoichiometric production of 4-nitrophenoxide
-
ion in the reaction of HO2 with bis(4-nitrophenyl)
phosphate.6 We assume that this reaction occurs with
NH2OH and bis(2,4-dinitrophenyl) phosphate because the
high leaving group ability of the 2,4-dinitrophenoxide
(20) J andorf, B. J . J . Am. Chem. Soc. 1956, 78, 3686.
(21) Harger, M. J . P. J . Chem. Soc., Perkin. Trans. 1 1981, 3284.
(22) Epstein, J .; Demek, M. M.; Rosenblatt, D. H. J . Org. Chem.
1956, 21, 796.
(23) Creaser, I. I.; Edwards, J . O. Top. Phosphorus Chem. 1972, 7,
379.
J . Org. Chem, Vol. 68, No. 18, 2003 7057

Domingos et al.
TABLE 5. NMR Sp ectr a l Da ta for P r od u cts of th e
Rea ction of BDNP P (0.01 M) w ith NH2OMe (0.1 M), in
D2O a t p D 9 a n d 25 °C
step reaction, an equilibrium with the intermediate in
low concentration, or a stepwise reaction with an inter-
mediate in steady state, and our reactions followed by
formation of DNP should not be strictly first-order, even
with dilute NH2OH. However, several factors allow
observation of reasonable first-order plots: (a) the most
important is that decomposition of intermediate 2 is
relatively fast and its concentration in solution is never
more than ca. 10% of that of the initial substrate; (b)
modest deviations from first-order kinetics are not readily
detected from the kinetic form of an individual reaction.
It is therefore understandable that we observed first-
order rate plots with dilute NH2OH, even though an
intermediate was not strictly in steady state. The sub-
sequent deviations from first-order kinetics, after exten-
sive loss of BDNPP, are also understandable because here
decomposition of DNPP, giving DNP, is accelerated by
attack of NH2OH and a typical profile of consecutive
reactions is observed.
31P NMR
δ (ppm)
compd
1H NMR δ (ppm)
6
3.89 (s, 3H, CH3), 7.52 (d, 1H, J ) 9.6 Hz, Ar),
8.39 (dd, 1H, J ab ) 9.6 Hz and J bx) 2.4 Hz, Ar),
9.08 (d, 1H, J ) 2.4 Hz, Ar)
3.63 (s, 3H, CH3), 7.74 (d, 1H, J ) 9.1 Hz, Ar),
8.54 (dd, 1H, J ab ) 9.1 Hz and J bx ) 2.7 Hz, Ar),
8.90 (d, 1H, J ) 2.7 Hz, Ar)
7
0.36
moiety allows its expulsion by the terminal NH2 group
in compound 2.24
For comparison purposes, we are noting the products
of the reaction of BDNPP with NH2OMe, which, besides
being much slower, gives less than 1 equiv of DNP. This
consequence of blocking the OH group confirms that
attack of the other hydroxylamine derivatives on BDNPP
is preferably by oxygen rather than nitrogen. The reac-
tion with NH2OMe is less simple than expected for a
nucleophile with one nucleophilic center, and we detected
Con clu sion s
The initial reaction of NH2OH with bis(2,4-dinitrophe-
nyl) phosphate involves release of 1 mol of 2,4-dinitro-
phenoxide ion and formation of a phosphorylated hy-
droxylamine, 2, which reacts readily with NH2OH, giving
the monoester 3, which reacts relatively slowly, generat-
ing more aryloxide ion DNP. The first-formed intermedi-
ate 2 also breaks down by two other independent
reactions; one involves intramolecular displacement of
the aryloxide ion DNP, and the other involves migration
of the 2,4-dinitrophenyl group from O to N, probably via
a transient cyclic Meisenheimer complex and formation
of phosphorylated 2,4-dinitrophenylhydroxylamine, 4,
which is hydrolyzed in aqueous acid but is otherwise
unreactive. The products and longer-lived intermediates
are identified by absorption and NMR spectroscopy.
1
by H and 31P NMR spectroscopy products 6 and 7 (see
Table 5 and the Experimental Section), which are formed
in two competitive reactions: (i) aromatic nucleophilic
substitution on BDNPP, which yields DNPP and product
6, and (ii) attack of the amino nitrogen at the phosphorus
atom of BDNPP, giving a phosphorylated product (7) and
DNP.
Ack n ow led gm en t. We are indebted to PRONEX,
CAPES, and CNPq, Brazil, CNPq/NSF, and the U.S.
Army Office of Research for financial support of this
work. We are grateful to Professor A. K. Yatsimirsky
for helpful discussions.
Kin etic F or m a n d P r od u ct F or m a tion . Reactions
in dilute NH2OH followed by the appearance of DNP
follow first-order kinetics (Figures 1 and 2), but as shown
in Scheme 4, some DNP is generated by reactions of the
first-formed intermediate 2, whose concentration is suf-
ficient for its detection by NMR spectroscopy (Table 3).
Adherence to a first-order kinetic form requires a single-
Su p p or tin g In for m a tion Ava ila ble: Optimized geom-
etries data (Cartesian coordinates) of structures 2, 2a , and 4.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(24) Kirby, A. J .; Younas, M. J . Chem. Soc. B 1970, 1165.
J O034348Z
7058 J . Org. Chem., Vol. 68, No. 18, 2003