Mechanisms of decomposition of aryl N-(methoxycarbonyl) sulfmates in aqueous media

Article abstract of DOI:10.1021/ja981014w

Rate constant and products are reported for the decomposition of aryl N- (methoxycarbonyl)sulfamates 1a-h (range) in aqueous solutions (pH 0-14) at 50 °C. The pK(a) values at 25 °C of 1a-h are 0.46-2.40. The pH-rate profiles of 1 indicate a rate law that includes three terms: two pH independent terms, k(a) in acid and k(p) around neutral pH, with k(a) > k(p), and a hydroxide ion dependent term, k(OH). In acid, product analysis reveals that both S-O (k(S02) and C - O (k(CO) bond cleavage reactions are involved. From an analysis of β(1g), solvent isotope effects, and solvent isotopic labeling of the products, it is concluded (a) that the C-O cleavage reaction involves protonation of the leaving group methanol and its expulsion from the dipolar intermediate ArOS0₂N-CO-O⁺HCH₃ and (b) that the S-O cleavage reaction may involve either an intra- or an intermolecular general acid-catalyzed decomosition of 1 or 1^-, respectively. In contrast to k(a), the spontaneous hydrolysis reaction of 1^-, k(p), takes place exclusively via S-O bond fission. The k(OH) reaction of 1^- is rationalized by OH^- attack either at the carbonyl center (1b-h) or at the aromatic ring (1a). It is concluded that the -SO₂NHCO-group may be viewed as an attractive phosphate analogue.

Full text of DOI:10.1021/ja981014w

9574
J. Am. Chem. Soc. 1998, 120, 9574-9583
Mechanisms of Decomposition of Aryl
N-(Methoxycarbonyl)sulfamates in Aqueous Media
Patrick Blans and Alain Vigroux*
Contribution from the Laboratoire de Synthe`se et physico-chimie organique associe´ au CNRS,
UniVersite´ Paul Sabatier, 118 route de Narbonne, 31062 Toulouse Cedex 4, France
ReceiVed March 25, 1998
Abstract: Rate constants and products are reported for the decomposition of aryl N-(methoxycarbonyl)sulfamates
1a-h (range) in aqueous solutions (pH 0-14) at 50 °C. The pKa values at 25 °C of 1a-h are 0.46-2.40.
The pH-rate profiles of 1 indicate a rate law that includes three terms: two pH independent terms, ka in acid
and kp around neutral pH, with ka > kp, and a hydroxide ion dependent term, kOH. In acid, product analysis
reveals that both S-O (kSO2) and C-O (kCO) bond cleavage reactions are involved. From an analysis of âlg,
solvent isotope effects, and solvent isotopic labeling of the products, it is concluded (a) that the C-O cleavage
reaction involves protonation of the leaving group methanol and its expulsion from the dipolar intermediate
ArOSO2N-CO-O+HCH3 and (b) that the S-O cleavage reaction may involve either an intra- or an
intermolecular general acid-catalyzed decomposition of 1 or 1-, respectively. In contrast to ka, the spontaneous
hydrolysis reaction of 1-, kp, takes place exclusively via S-O bond fission. The kOH reaction of 1- is rationalized
by OH- attack either at the carbonyl center (1b-h) or at the aromatic ring (1a). It is concluded that the
-SO2NHCO- group may be viewed as an attractive phosphate analogue.
Introduction
strategies, the extended use of the -SO2NHCO- groupsas a
replacement for the diphosphate or sulfate moiety of a variety
of critical biomoleculessproved to be successful as well.2-5
Also, the ionized form of this spacer was used to design an
unusual water-soluble ACAT inhibitor which is well absorbed
and thus exhibits improved bioavailability.6
Scrutiny of the above-mentioned work clearly indicates that
the [(carbonyl)amino]sulfonyl spacer moiety plays an important
role in both antiviral and inhibitory activities. Despite the
exciting prospects presented by these bioisosterism-based
therapeutics,7 we are unaware of any previously reported study
intended to investigate the aqueous reaction chemistry of this
attractive phosphate isostere.8 In addition to successfully
mimicking the diphosphate bridge of a variety of biomolecules,
including coenzymes and secondary messengers, we wondered
whether the -SO2NHCO- linkage would also be qualified to
carry out the biological tasks which naturally lie with the
phosphate group in nucleic acids and lipids.10 For instance,
replacement of the phosphodiester backbone by appropriate
Phosphate esters and anhydrides are of fundamental impor-
tance in biological systems. The genetic materials DNA and
RNA are phosphodiesters, and most of the coenzymes are esters
of phosphoric or pyrophosphoric acid. Recently, a new type
of analogue of uridine 5′-diphosphate glucose (UDP-Glc), in
which the naturally occurring diphosphate group has been
replaced by an isosteric O-SO2NHCO-O group, were found
to interfere with protein glycosylation, inhibiting the glycosy-
lation of viral proteins to a greater extent than glycosylation of
cellular proteins.1 Although these analogues were initially
(2) Pe´rez-Pe´rez, M. J.; Balzarini, J.; de Clercq, E.; Camarasa, M. J.
Bioorg. Med. Chem. 1993, 1, 279.
(3) Paul, P.; Lutz, T. M.; Osborn, C.; Kyosseva, S.; Elbein, A. D.;
Towbin, H.; Radominska, A.; Drake, R. R. J. Biol. Chem. 1993, 268, 12933.
(4) Zhu, B. C. R.; Drake, R. R.; Schweingruber, H.; Laine, R. A. Arch.
Biochem. Biophys. 1995, 319, 355.
(5) Woo, L. W. L.; Purohit, A.; Reed, M. J.; Potter, V. L. Bioorg. Med.
Chem. Lett. 1997, 7, 3075.
(6) Sliskovic, D. R.; Krause, B. R.; Picard, J. A.; Anderson, M.; Bousley,
R. F.; Hamelehle, K. L.; Homan, R.; Julian, T. N.; Rashidbaigi, Z. A.;
Stanfield, R. L. J. Med. Chem. 1994, 37, 560.
(7) Patani, G. A.; LaVoie, E. J. Chem. ReV. 1996, 96, 3147.
(8) Previous investigations into the mechanisms of this group only
focused on synthetic applications in aprotic solvents, to obtain olefins in
high yields from sodium N-(alkoxycarbonyl)sulfamate derivatives.9
(9) Atkins, G. M., Jr.; Burgess, E. M. J. Am. Chem. Soc. 1972, 94, 6135
and references therein.
designed to interfere with the glycosylation process and, do in
fact, inhibit it, their mode of action proved to be more complex
since they also inhibit DNA synthesis.1b In recent therapeutic
* Author to whom correspondence should be addressed at the follow-
ing: phone 33 5 61 55 62 97; fax 33 5 61 55 60 11; e-mail vigroux@iris.ups-
tlse.fr.
(1) (a) Camarasa, M. J.; Fernandez-Resa, P.; Garcia-Lopez, M. T.; de
las Heras, F. G.; Mendez-Castrillon, P. P.; Alarcon, B.; Carrasco, L. J. Med.
Chem. 1985, 28, 40. (b) Alarcon, B.; Gonzalez, M. E.; Carrasco, L.; Mendez-
Castrillon, P. P.; Garcia-Lopez, M. T.; de las Heras, F. G. Antimicrob. Agents
Chemother. 1988, 32, 1257.
(10) Westheimer, F. H. Science 1987, 235, 1173.
S0002-7863(98)01014-2 CCC: $15.00 © 1998 American Chemical Society
Published on Web 09/05/1998

Aryl N-(Methoxycarbonyl)sulfamates in H2O
J. Am. Chem. Soc., Vol. 120, No. 37, 1998 9575
Table 1. Effect of the Aryl Group on Product Distributions, Rate Parameters and Solvent Deuterium Isotope Effects for Hydrolysis of Aryl
N-(Methoxycarbonyl)Sulfamates 1 at 50 °C in 1.0 M DCl Solution Containing 16% CD3CN, unless Otherwise Noted, µ ) 1.0 M
a
compd
phenyl
% MeOHa
% H2NCO2CH3
kobsd, s-1 b,c
kCO, s-1 b,d
kSO , s-1 b,e
kH O/kD O
b,f
ArOSO2NHCO2CH3
substituent
pKArOH (C-O cleavage) (S-O cleavage)
2
2
2
1a
1b
1c
1d
1e
1f
2,4-(NO2)2
2-Cl-4-NO2
2-NO2-4-Cl
4-NO2
4-CN
4-CH3CO
4-Cl
4.11
5.45
6.48
7.15
7.95
8.05
9.38
9.92
0.0
28.9
36.6
51.5
63.1
78.1
93.6
96.9
100
71.1
63.4
48.5
36.9
21.9
6.4
7.38 × 10-4
7.38 × 10-4
2.00
1.55
1.49
1.40
1.38
1.08
1.03
1.03
6.75 × 10-5 1.95 × 10-5 4.80 × 10-5
7.42 × 10-5 2.71 × 10-5 4.70 × 10-5
2.56 × 10-5 1.32 × 10-5 1.24 × 10-5
1.97 × 10-5 1.24 × 10-5 7.25 × 10-6
1.58 × 10-5 1.24 × 10-5 3.46 × 10-6
1.30 × 10-5 1.22 × 10-5 8.33 × 10-7
1.13 × 10-5 1.10 × 10-5 3.50 × 10-7
1g
1h
4-H
3.1
a Percentages of methanol and methyl carbamate were determined by H NMR spectroscopy with an estimated error (1% based on an average
of three measurements. b Standard errors are (5%. c Overall first-order rate constant measured spectrophotometrically. d Rate constant of formation
of methanol determined from kobsd and product ratio given in the previous column. e Rate constant of formation of methyl carbamate determined as
1
f
indicated in the previous column. Overall first-order rate constants ratio measured spectrophotometrically at 50 °C in 1.0 M LCl solutions containing
0.7% dioxane (µ ) 1.0 M).
Scheme 1a,b
moieties would be particularly valuable for modified oligo-
nucleotides to be used in antisense and antigene therapies.11 To
address this question, and then help the medicinal chemist to
rationally extend the use of the [(carbonyl)amino]sulfonyl
moiety in chemotherapy, we have been engaged in a detailed
quantitative study of the aqueous chemistry of a wide range of
[(carbonyl)amino]sulfonyl-linked derivatives, G1-SO2NHCO-
G2.12 In the present work, we report our investigation of the
mechanisms of the aqueous chemistry of aryl N-(methoxycar-
bonyl)sulfamates 1 (Table 1).
a Not detected in strong acidic media due to concomitant hydrolysis
Due to the low pKa of compounds 1 (pKa ) 0.5-2.4), our
affording methyl carbamate. b Not detected after completion of the
work is relevant to the aqueous chemistry of unusually weak
kinetic runs of 1 due to concomitant hydrolysis affording the corre-
nitrogen anions in comparison with the aqueous chemistry of
sponding substituted phenols.
the oxyanions encountered in phosphate and sulfate esters. From
a mechanistic perspective, we wondered whether such anions
relative to the residual HDO peak at δ 4.68). GC/MS and EIMS spectra
would (i) decompose through the sulfonyl or the carbonyl group,
or through both of them,13 (ii) behave like oxyanions in the
hydrolysis of phosphate and sulfate esters, (iii) thus involving
the formation ofsdiffusible or notspotent electrophiles (Scheme
1) capable of modifying critical macromolecules.
were recorded on a Nermag R 10-10 mass spectrometer.
All aryl N-(methoxycarbonyl)sulfamates 1 were obtained in 42-
61% yield by reacting [N-(methoxycarbonyl)sulfamoyl] triethylammo-
nium inner salt with corresponding substituted phenols as described
by Burgess et al.9 N-Methylation of compound 1a by the Mitsunobu
reaction,15 as described by Criton et al.,16 gave 2,4-dinitrophenyl
N-methyl-N-(methoxycarbonyl)sulfamate 2 in 51% yield. The spec-
troscopic and analytical data for these compounds are given in the
Supporting Information.
Experimental Section
In general, chemicals were purchased as the best available com-
mercial grade. Organic chemicals were purified by distillation or
recrystallization prior to use. Water was distilled and deionized on a
Water-Purification System Milli-Q (Millipore Apparatus). Melting
points were taken on a Kofler hot-stage apparatus and are uncorrected.
Elemental analyses were performed by Analytic Service, University
Paul Sabatier. IR spectra were recorded using a Perkin-Elmer 883
spectrometer. 1H and 13C NMR spectra were obtained with a Bruker
Methods. a. Kinetics. Decomposition reactions of substrates 1
and 2 in aqueous media were monitored either by UV-vis spectro-
photometry or 1H NMR spectroscopy or by use of the HPLC technique
depending on specific experimental requirements. All of these reactions
gave, except where otherwise stated, satisfactory first-order rate
constants. The solutions employed were HCl (pH 0-2.0),17 trichlo-
roacetate (pH 0.5-1.1), difluoroacetate (pH 0.9-1.5), dichloroacetate
(pH 1-1.5), cyanoacetate (pH 2.1-3.2), acetate (pH 4-5.3), phosphate
(pH 5.9-7.2), borate (pH 8.3-9.6), generally at 0.5 M, and NaOH
(pH 11-13). The pH of the reaction mixtures was measured upon
initiation and after completion of the runs using a Tacussel pH-meter
1
AC-80 or AC-250 spectrometer (250 MHz for H and 50.32 MHz for
1
13C). Chemical shifts for H NMR spectra of chloroform-d solutions
(containing 1 to 2 drops of DMSO-d6 to improve solubility) are relative
to internal TMS. For deuterium oxide solutions, all spectra were
recorded with the Bruker AC-250 apparatus (1H chemical shifts are
(14) The rate constants ratio, kCO/kSO2, for the E1cB decompositions of
the corresponding conjugate bases of CH3NHCO-OAr (kCO) and CH3-
NHSO2-OAr (kSO2), was estimated from the literature data to ca. 109 when
the leaving group is p-nitrophenol. See: (a) Al-Rawi, H.; Williams, A. J.
Am. Chem. Soc. 1977, 99, 2671. (b) Williams, A.; Douglas, K. T. J. Chem.
Soc., Perkin Trans. 2 1974, 1727.
(15) Mitsunobu, O. Synthesis 1981, 1.
(16) Criton, M.; Dewynter, G.; Aouf, N.; Montero, J. L.; Imbach, J. L.
Nucleosides Nucleotides 1995, 14, 1795.
(11) For recent reviews, see: Uhlmann, E.; Peyman, A. Chem. ReV. 1990,
90, 543. de Mesmaeker, A.; Ha¨ner, R.; Martin, P.; Moser, H. E. Acc. Chem.
Res. 1995, 28, 366.
(12) Vigroux, A.; Bergon, M.; Bergonzi, C.; Tisne`s, P. J. Am. Chem.
Soc. 1994, 116, 11787.
(13) Given that nitrogen anions decompose ca. 109 times faster through
carbonyl than sulfonyl group,14 compounds 1 might prove to be good
candidates to exhibit ambivalent reactivity due to the difference in pKa that
may lie between methanol and ArOH (e.g., when ArOH is p-nitrophenol,
∆pKa ≈ 8.5).
(17) No correction was applied for the small error incurred in reading
pH values between pH 0 and 1.

9576 J. Am. Chem. Soc., Vol. 120, No. 37, 1998
Blans and Vigroux
Table 2. Values of Equilibrium and Rate Constants Determined as Kinetically Apparent Constants (50 °C and µ ) 1.0 M with KCl) and via
Spectrophotometric Titration (25 °C and µ ) 1.0 M with KCl)
a
b
compd
pKapp
pKa
ka, s-1 a,c
kp, s-1 a,d
kOH, M-1 s-1 a
kOH, M-1 s-1 e
1a
1b
1c
1d
1e
1f
0.79 ( 0.30
1.21 ( 0.06
1.06 ( 0.02
1.58 ( 0.03
1.56 ( 0.01
1.80 ( 0.08
2.13 ( 0.17
2.16 ( 0.12
0.46 ( 0.04
1.16 ( 0.02
0.97 ( 0.03
1.55 ( 0.02
1.71 ( 0.01
1.85 ( 0.02
2.10 ( 0.02
2.40 ( 0.04
(1.54 ( 0.09) × 10-3
(1.18 ( 0.06) × 10-4
(1.28 ( 0.03) × 10-4
(3.17 ( 0.1) × 10-5
(2.19 ( 0.02) × 10-5
(1.57 ( 0.1) × 10-5
(9.25 ( 2) × 10-6
(1.07 ( 0.01) × 10-3
(2.09 ( 0.09) × 10-6
(4.37 ( 0.08) × 10-7
(1.40 ( 0.06) × 10-8 f
(3.05 ( 0.03) × 10-9 f
(2.48 ( 0.3) × 10-9 f
(1.44 ( 0.4) × 10-9 f
(3.06 ( 0.5) × 10-9 f
(1.05 ( 0.03) × 10-2
(9.51 ( 0.4) × 10-5
(9.74 ( 0.2) × 10-5
(6.72 ( 0.3) × 10-5
g
(1.32 ( 0.09) × 10-2
(1.07 ( 0.02) × 10-4
(9.23 ( 0.3) × 10-5
(7.50 ( 0.2) × 10-5
g
(3.74 ( 0.3) × 10-5
(2.18 ( 0.4) × 10-5
(1.87 ( 0.2) × 10-5
(4.54 ( 0.2) × 10-5
(3.35 ( 0.1) × 10-5
(2.29 ( 0.07) × 10-5
1g
1h
(8.22 ( 0.8) × 10-6
a Kinetically apparent constant obtained at 50 °C by fitting of eq 1 to experimental data points. b Ionization constant determined at 25 °C by
spectrophotometric titration. The effect of pKArOH (given in Table 1) on the pKa of the corresponding aryl N-(methoxycarbonyl)sulfamates 1 measured
at 25 °C obeys the following equation: pKa ) (0.32 ( 0.03)pKArOH - (0.8 ( 0.2). c Overall first-order rate constant corresponding to the acidic
pH-independent region of the pH-rate profiles shown in Figure 1. d Spontaneous first-order rate constant corresponding to the plateau region of the
pH-rate profiles (Figure 1) around neutral pH. e Second-order rate constant for hydroxide ion determined at 50 °C from the linear plots of kobsd vs
[OH-] in the alkaline pH-dependent region (pH 11-13). f Determined from extrapolation of an Eyring plot obtained between 100 and 90 °C by the
initial rates method using UV-vis spectrophotometry. g Interference from reaction of hydroxide ion with the cyano group prevented measurement.
Table 3. Activation Parametersa and Solvent Deuterium Isotope Effects for the Hydrolysis Reactions of Compounds 1 and 2
reactionb
∆Hq, kcal/mol
∆Sq, cal/(deg mol)
kH O/kD O
2 2
c
compd
1a
kp (ka)
kOH
20.4 ( 0.3d (21.9 ( 0.2)d
-9.2 ( 0.9d (-4.0 ( 0.7)d
1.16e (2.00)e,f
16.6 ( 0.2g
-17.0 ( 0.8g
1b
1c
1d
1e
1f
1g
1h
2
kp (kobsd
kp (kobsd
kp (kobsd
kp (kobsd
kp (kobsd
kp (kobsd
kp (kobsd
)
)
)
)
)
)
)
23.8 ( 0.4h
-11.0 ( 1.3h
(1.25)
(1.09)
(0.91)
(1.05)
(0.89)
(0.91)
(0.84)
(1.93)
24.9 ( 0.6h
-10.9 ( 1.8h
24.2 ( 1.3i
-19.6 ( 3.6i
21.3 ( 0.9i
-31.8 ( 2.4i
21.8 ( 0.9i
-30.7 ( 2.5i
23.2 ( 0.2i
-27.2 ( 0.5i
21.3 ( 0.3i
-31.7 ( 0.8i
kOH (ka)
21.0 ( 0.2g (19.9 ( 0.4)h
-4.5 ( 0.8g (-26.7 ( 1.4)h
a The temperature dependence of rate constants is given in Table S3 in the Supporting Information. b ka refers to the pH-independent reaction in
strong acidic media, kp to the spontaneous pH-independent hydrolysis reaction around neutral pH, kobsd to the pH-dependent reaction observed at
pH ) 3.60 in formate buffer (0.5 fraction base) and kOH to the hydroxide ion-catalyzed hydrolysis reaction at pH > 12 for 1a and at pH > 6 for
2. c At 75 °C for 3 half-lives unless otherwise stated. d Measurements at six temperatures ranging between 35 and 60 °C in phosphate buffer pH
f
5.95 (kp) and in 1.0 M HCl solution (ka). e At 50 °C. See Table 1 for the solvent deuterium isotope effects for the ka reaction of 1b-h. g Measurements
at five temperatures ranging between 30 and 50 °C. h Measurements at five temperatures ranging between 50 and 70 °C (initial rates method used
by UV-vis spectrophotometry). i Determined at three temperatures (90, 95, and 100 °C at pH ) 7.25 in phosphate buffer 0.8 fraction base) by
means of the initial rates method using UV-vis spectrophotometry, see Experimental Section.
(TT processor 2) using XC 111 or U402-M3-S7/60 Ingold combination
electrodes with calibrations carried out with commercially available
standards at 50 °C. Kinetic runs exhibiting pH drift greater than (0.02
unit were discarded. The kinetic runs in D2O were carried out under
the same conditions described for H2O. The value of pD was obtained
by adding 0.29 to the observed pH of solutions in D2O.18
kinetic measurements carried out either by UV-vis spectrophotometry,
1H NMR spectroscopy or by use of the HPLC technique are given in
the Supporting Information.
b. Product Distributions. The product distributions for the
hydrolysis of aryl N-(methoxycarbonyl)sulfamates 1 in DCl and
1
buffered deuterium oxide solutions were determined using H NMR
UV-vis spectrophotometry was used for all reactions except those
of substrates 1f-h in the pH range 0-2 (see below). Reactions were
monitored at the appropriate wavelengths using either a Perkin-Elmer
Lambda 7 or a Hewlett-Packard 8453 UV-vis spectrophotometer
attached to thermostated water bath. Runs were carried out in 3 mL
of reaction solution contained in a 1.0 cm quartz cuvette equilibrated
at 50.0 ( 0.2 °C. Suitable wavelengths for the kinetic studies were
selected by repetitive spectral scanning of the reactions. The very slow
hydrolysis of compounds 1 at 50 °C in the pH range 3-10 was
measured by following initial rates of release of the phenols. This
method was generally used for reactions with half-lives longer than
ca. 1 day at 50 °C (see Table S4 in the Supporting Information for the
specific method employed for each compound studied in this work).
For compounds 1d-h, kobsd values at neutral pH were extrapolated at
50 °C from measurements carried out at 100, 95, and 90 °C (initial
rates method used) using activation parameters given in Table 3. In
the pH range 0-2, decomposition reactions of substrates 1f-h were
monitored by means of the HPLC technique due to the concomitant
hydrolysis of the aryl sulfamate intermediates formed during the reaction
(see Scheme 1) which prevents reliable measurements of first-order
rate constants by UV-vis spectrophotometry. 1H NMR Spectroscopy
was used to complement the kinetic study of compounds 1 carried out
by UV-vis spectrophotometry. The experimental procedures for
spectroscopy. In the pH range 0-2, reactions were initiated in a 5
mL vial by mixing acetonitrile solutions containing substrates into acidic
buffers preincubated at 50 °C (ionic strength maintained at 1.0 M with
KCl) affording initial substrate concentrations of (3.3-4.3) × 10-2 M.
The final concentrations of acetonitrile were 16 vol % for reasons of
solubility. When the reactions were complete (at 4-6 half-lives,
determined from the corresponding kobsd values obtained by UV
spectrophotometry in D2O, using the same experimental conditions and
the same amounts of acetonitrile as cosolvent), the solutions were
transferred to NMR tubes. A spectrum was recorded at 4 and 6 half-
lives of the stated reaction time in order to check products’ stability.
Methanol, methyl carbamate, and substituted phenols were identified
at completion by a comparison of NMR peaks with those of authentic
samples of pure material obtained in the same experimental conditions
1
(the H NMR signal for the methyl group of methanol obtained from
the hydrolysis of substrates 1 appears at δ 3.1-3.5 and the signal for
the methyl group of methyl carbamate at δ 3.5-3.8, depending on
acidity). Yields of product methanol, methyl carbamate, and substituted
phenols were checked at 6 half-lives and were within 3% of those
determined at 4 half-lives in the case of all substrates studied, attesting
to the stability of these products under the reaction conditions.
The pH-independent acidic hydrolysis reactions (ka) of aryl N-
(methoxycarbonyl)sulfamates 1 produce two sets of products (Scheme
1). The first set is methanol and substituted phenylsulfamates which
(18) Fife, T. H.; Bruice, T. C. J. Phys. Chem. 1961, 65, 1079.

Aryl N-(Methoxycarbonyl)sulfamates in H2O
J. Am. Chem. Soc., Vol. 120, No. 37, 1998 9577
substrates 1 at the same temperature. The values of Ka
determined spectrophotometrically at 25 °C and µ ) 1.0 M with
KCl, are also summarized in Table 2. These values are in
reasonable agreement with the apparent Kapp values obtained at
50 °C by fitting of eq 1 to experimental data points.
Products. The pH-independent acidic hydrolysis of com-
pounds 1, ka, yields four products in two sets of two. One set
is methanol and substituted phenylsulfamates (the C-O cleavage
products, kCO in Scheme 1). Detection of the aryl sulfamate
esters in acidic media was complicated (except for compounds
1f-h) by concomitant hydrolysis affording the corresponding
substituted phenols (Scheme 1). The second set is methyl
carbamate and substituted phenols (the S-O cleavage products,
kSO2 in Scheme 1). In 1.0 M DCl solution, the relative amounts
of the two sets of products vary linearly from 0 to 100% when
the pKa of the leaving group ArOH lies between ca. 4.2 and
9.5 (see Figure S1 in the Supporting Information). For pKaArOH
values less than 4.2, the hydrolysis reaction occurs only via kSO2
pathway whereas for those pKaArOH values greater than 9.5 kCO
is the unique pathway (Table 1).
Figure 1. Plots of log kobsd, the overall first-order rate constants for
hydrolysis of aryl N-(methoxycarbonyl)sulfamates 1, as a function of
pH at 50 °C and 1.0 M ionic strength (KCl). The curves (solid lines)
were computer-generated by iteratively fitting the experimental points
to eq 1 given in the text (see Results). The values of constants that
provided the optimal fits are summarized in Table 2. Compounds 1
are represented by symbols as follows: 1a, (9); 1b, (4); 1c, (]); 1d,
([); 1f, (2); 1g, (O); 1h, (0). The numbering system for substrate
identification is in Table 1. Data points for which log kobsd is lower
than -7.5 were extrapolated from Eyring plots obtained between 100
and 90 °C by the initial rates method using UV-vis spectrophotometry.
Fit for 1e is omitted for clarity (due to the competitive reaction of
hydroxide ion with the cyano group of 1e, it has not been possible to
study this compound in solutions of pH > 8).
The pH-independent rate constant kp for hydrolysis of
compounds 1a-c in the pH region around 7 yields the
corresponding substituted phenols ArOH and N-(methoxycar-
bonyl)sulfamate -OSO2NHCOOMe as the final products,
indicating the kSO2 route in Scheme 1 is the unique pathway
for spontaneous hydrolysis of these compounds: kp ) kSO2. The
1H NMR signal due to the methyl group of product N-
(methoxycarbonyl)sulfamate formed during the kinetic runs of
1a-c was found to be identical (δ ) 3.65 ( 0.05) to that of an
authentic sample of N-(methoxycarbonyl)sulfamoyl chloride
ClSO2NHCOOMe recorded in the same reaction conditions.
Since rapid hydrolysis of S-Cl bond is expected to occur from
the latter compound to yield N-(methoxycarbonyl)sulfamate, we
actually consider this similarity of NMR peaks as good evidence
for the formation of -OSO2NHCOOMe during spontaneous
hydrolysis of 1a-c. To ensure complete comparability with
the S-O cleavage products formed in acidic media (i.e.,
substituted phenols coupled with methyl carbamate), we checked
and confirmed that N-(methoxycarbonyl)sulfamate hydrolyzes
to methyl carbamate during the kinetic runs of compounds 1
performed at 50 °C in 1.0 M DCl solution.
decompose during the kinetic runs into the corresponding substituted
phenols. The second set is methyl carbamate and substituted phenols.
At the pH-independent portion ka of the pH-rate profiles shown in Figure
1, the relative amounts of the two sets of products depend on the aryl
group of substrates. The relative amount of methanol was determined
in 1.0 M DCl solutions by measuring the ratio of the intensity of the
signal due to this product (δ ) 3.1) to the combined signal intensities
due to methyl carbamate (δ ) 3.35) and methanol. For each set of
product analysis, particular precautions were taken to check the stability
of methyl carbamate with respect to its possible convertion into
methanol. Partial rates for the two modes of breakdown (kCO and kSO2
)
were obtained from the product ratios and the first-order rate constants
observed (kobsd) at 50 °C in DCl solutions containing the appropriate
percentage of acetonitrile (16 vol %).
The experimental procedures for H218O experiments and pKa
For compounds 1d-h, the spontaneous hydrolysis reaction
kp was found to yield, in phosphate buffers, substituted phenols,
and methanol as the final products. Although methanol is
obtained from the latter compounds, we still interpret kp as the
determinations are given in the Supporting Information.
Results
k
SO2 route of Scheme 1. In effect, while N-(methoxycarbonyl)-
Compounds 1a-h. pH-Rate Profiles. Figure 1 shows the
plots of the log of the overall first-order rate constants kobsd vs
pH at 50 °C. The pH-rate constant profiles for hydrolysis of
compounds 1a-h shown in Figure 1 are characterized by four
distinct regions: (i) the appearance of a plateau below pH 1
(ka) followed by (ii) a decrease of log kobsd with increasing pH,
then (iii) a pH-independent region (kp) preceding (iv) a
hydroxide ion-catalyzed hydrolysis reaction (kOH) at high pH.
The latter reaction is characterized by plots of log kobsd vs pH
which are linear with slopes of +1.0. The experimental points
were fit to eq 1, where aH is the hydrogen ion activity measured
at 50 °C. The values of the constants ka, kp, kOH, and Kapp
required to fit the experimental rate constants to eq 1 are
recorded in Table 2. KW is the autoprotolysis constant of water
sulfamate (δ ) 3.44) hydrolyzes to methyl carbamate (δ ) 3.35)
in 1.0 M DCl solution, we found that its neutral hydrolysis
investigated in phosphate buffer (pD 7.84) at 100 and 95 °C
led to the formation of methanol (δ ) 3.28 ( 0.05) with the
observed first-order rate constants k′ ) 1.26 × 10-4 s-1 at 100
°C and k′ ) 8.40 × 10-5 s-1 at 95 °C (kinetic runs carried out
by 1H NMR for 3 half-lives) giving rise to k′/kp values ranging
between ca. 50 (1d) and 580 (1g). Since the kp terms for
compounds 1d-h were extrapolated at 50 °C from measure-
ments carried out at 100, 95, and 90 °C (see the Experimental
Section), we explain the formation of methanol during hydrolysis
of 1d-h at those temperatures by the fact that N-(methoxycar-
bonyl)sulfamate is a steady-state intermediate in the pH(D)-
independent decay of 1d-h ([1d-h] ) (5-8) × 10-2 M,
[buffer] ) 0.5 M). In every case, i.e., from 1d to 1h, the
putative N-(methoxycarbonyl)sulfamate intermediate is sub-
stantially more reactiveswith regard to its conversion into
methanolsthan its substrate precursors 1d-h (k′ > kp in eq 2
kobsd ) (kaaH2 + kpKappaH + kOHKWKapp)/(aH2 + KappaH) (1)
at 50 °C, and Kapp the apparent acid dissociation constant of

9578 J. Am. Chem. Soc., Vol. 120, No. 37, 1998
Blans and Vigroux
1
below) so that the rate of methanol formation measured by H
NMR by the initial rates method (see Supporting Information)
is obtained in approximately quantitative yields. In Tables S1
and S2 in the Supporting Information are listed rate data and
the second-order rate constants knuc for the reactions of sodium
azide, pyridine, and substituted pyridines with the SO2 group
of 1a measured at 39 °C in 0.5 M Tris buffer (pH 8.2, ionic
strength 1.0 M with KCl). Measurements were carried out by
UV-vis spectrophotometry at 360 nm. Attack of pyridines at
the aromatic ring would give a stable cationic product19 with
no free electron pair on nitrogen; the latter product does not
absorb at 360 nm so that the fraction of optical density at 360
nm due to 2,4-dinitrophenolate absorption was considered as a
measure of the fraction of S-O bond cleavage.20 Since the
release of 2,4-dinitrophenolate anion in most of the pyridine-
catalyzed reactions is nearly quantitative, one may conclude that
pyridines must attack predominantly at sulfur.21 In Figure S2
(Supporting Information) is shown the linear plot of log knuc
for S-O bond cleavage versus pKa of the attacking nucleophile
for the reaction of sodium azide and substituted pyridines with
1a. The Bro¨nsted-type coefficient is ânuc ) 0.14 ( 0.02.
H218O Experiments. In an attempt to determine the positions
of bond cleavage during the acidic and basis hydrolyses of
compounds 1a and 1d, experiments were conducted to examine
if 18O oxygen atoms originally present in an isotopically mixed
16O/18O reaction mixture would be incorporated into products
methanol and 4-nitrophenol (1d) and 2,4-dinitrophenol (1a). At
the end of the acidic (1.0 M HCl) and basic (1.0 M NaOH)
hydrolysis reactions of 1d, product methanol was analyzed by
use of the GC/EIMS technique, and the analysis in every case
indicated no detectable signal corresponding to Me18OH. The
isotopic abundance of 18O in 4-nitrophenol formed in acidic
and basic 42.1% 18O-enriched reaction mixtures was found to
be 0.71 and 0.77%, respectively, for a theoretical value of 0.84%
based on “natural abundance.” On the contrary, the 18O content
in product 2,4-dinitrophenol formed after basic hydrolysis of
1a in the same isotopically mixed 16O/18O reaction mixture was
found to be 43.7%, for a theoretical value of 1.28% based on
“natural abundance”.22
can be considered a measure of kSO2:kobsd ) kp ) kSO2
.
Additional evidence that N-(methoxycarbonyl)sulfamate is a
1
steady-state intermediate was found in the H NMR spectra
which were recorded versus time during the first 15% of the
hydrolysis reaction of 1d in phosphate buffer ([1d] ) 6.5 ×
10-2 M, pD ) 7.84) at 100, 95, and 90 °C. In each case the
methyl group of N-(methoxycarbonyl)sulfamate was identified
at δ ) 3.65 ( 0.05 at the steady concentration of ca. 1.0 ×
10-3 M whereas methanol concentration increased in the same
time from ca. 1.0 × 10-3 M to 9.8 × 10-3 M (i.e., more than
830% increase). While the observed rate constant kp markedly
decreases on going from substrate 1a to 1d, it remains virtually
constant on going from 1e to 1h suggesting a change in
mechanism as the leaving group ArOH becomes poorer than
ca. 4-nitrophenol. In the hydrolysis process of 1e-h, the point
at which k′ . kp is reached and the sulfamate intermediate is
no longer detected.
Product analysis performed in the pH portion where 1a-h
hydrolysis is hydroxide ion catalyzed suggests that compound
1a differs from the others compounds in that it produces 2,4-
dinitrophenol and N-(methoxycarbonyl)sulfamate as the ultimate
products whereas the final products for 1b-h hydrolysis are
substituted phenols and methanol. In contrast to the spontaneous
reaction kp, the formation of methanol cannot be explained from
the basic hydrolysis of the putative intermediate N-(methoxy-
carbonyl)sulfamate, as in eq 2, since the latter compound was
found to be quite stable during the hydrolysis reaction time of
1b-h at high pD. Indeed, hydrolysis of N-(methoxycarbonyl)-
sulfamate was examined by 1H NMR spectroscopy. Only 5.6%
hydrolysis into methanol could be detected after 5 days at 50
°C in 0.5 M NaOD solution while all substrates 1b-h, during
the same period of time and in the same reaction conditions,
were found to give methanol in quantitative yields. Since
N-(methoxycarbonyl)sulfamate was never detected in the basic
hydrolysis of 1b-h, it may be deduced that it never forms.
Buffer Catalysis. The overall first-order rate constants kobsd
were found to be quite insensitive (typically, the values of kobsd
changed by less than 2%) to changes in buffer concentrations
that spanned from 0.05 to 0.5 M, indicating that the hydrolysis
of compounds 1 is not buffer-catalyzed. The buffer solutions
employed for detecting catalysis were trichloroacetate (pH 0.5-
1.1), difluoroacetate (pH 0.9-1.5), dichloroacetate (pH 1-1.5),
cyanoacetate (pH 2.1-3.2), acetate (pH 4-5.3), phosphate (pH
5.9-7.2), and borate (pH 8.3-9.6).
Further experiments were carried out to investigate the
possibility of exchange of the carbonyl oxygen of 1d with that
of water enriched in H218O. Unreacted 1d from acidic (1.0 M
HCl) and basic (1.0 M NaOH) hydrolysis in enriched water
(42.1% H218O) gave abundance ratios (M + 2)+/M+ equal to
5.63 and 5.25%, respectively, compared to a theoretical value
of 27.46% assuming there is total incorporation of 18O only at
the carbonyl moiety of 1d. A control value for (M + 2)+/M+
ratio of unreacted 1d hydrolyzed in H216O was shown to be
5.47 and 5.07% in 1.0 M HCl and NaOH solutions, respectively,
for a theoretical value of 6.41% based on “natural abundance”.
Compound 2. The hydrolysis of the nonionizable N-methyl
analogue of 1a, namely, 2,4-dinitrophenyl N-methyl-N-(meth-
oxycarbonyl)sulfamate 2, gave under the same reaction condi-
tions used in the hydrolysis of 1a (at 50 °C and µ ) 1.0 M),
(19) The second-order rate constant for attack of pyridine on 1-chloro-
2,4-dinitrobenzene is several hundred times smaller than that measured for
attack on 1a and the product of attack at carbone, N-2,4-dinitrophenyl-
pyridinium chloride is hydrolyzed 44 times more slowly (kobsd ) 8.0 ×
10-6 s-1) than 1a (3.5 × 10-4 s-1) under the same reaction conditions.
(20) Kirby, A. J.; Jencks, W. P. J. Am. Chem. Soc. 1965, 87, 3209.
(21) Some C-O fission occurs with 4-aminopyridine (38% at 0.1 M
free amine concentration) and sodium azide (17% at 0.25 M total azide
concentration).
Effect of Added Nucleophiles on 1a. The kinetics of the
reaction of amine nucleophiles with 1a are satisfied by the
expression given in eq 3. The nucleophilic reaction knuc may
occur via attack on the aromatic ring leading to C-O bond
cleavage and anilide formation or attack on sulfur leading to
S-O bond fission and N-(methoxycarbonyl)sulfamoylamine.
(22) Control runs indicated that the (M + 2)+/M+ ratio of signals in
authentic methanol, 4-nitrophenol, and 2,4-dinitrophenol that were subjected
to the same hydrolysis conditions than 1d and 1a (i.e., at 50 °C in exactly
the same acidic or basic isotopically mixed H216O/H218O reaction mixtures)
is not appreciably different than predicted on the basis of “natural
abundance”.
kobsd ) kp + knuc[nuc]
(3)
Our primary interest is focused on the amine reactions that lead
to S-O bond cleavage, i.e., those for which 2,4-dinitrophenol

Aryl N-(Methoxycarbonyl)sulfamates in H2O
J. Am. Chem. Soc., Vol. 120, No. 37, 1998 9579
of kCO/kSO2 reflects the greater ease for the C-O bond to be
cleaved during anion breakdown. Therefore, the most likely
E1cB reaction that might take place in the hydrolysis process
of these compounds is anion decomposition through carbonyl
group as in eq 5. However, if one considers that the sensitivities
Figure 2. Values of log k0 and log kobsd, the buffer-independent rate
constants for hydrolysis of 2,4-dinitrophenyl N-methyl-N-(methoxy-
carbonyl)sulfamate 2 (9) and 2,4-dinitrophenyl N-(methoxycarbonyl)-
sulfamate 1a (0), respectively, against pH in aqueous solutions, 50
°C, µ ) 1.0 M (with KCl).
of log kSO2 and log kCO toward the basicities of the leaving
groups R1O- and R2O-, respectively, are close to unity (i.e.,
âlg ≈ -1),23 one may then expect an ambivalent reactivity to
show up, i.e., kCO ≈ kSO2, if the acidity of the leaving group
R1OH attached to the sulfonyl moiety becomes ca. 9 orders of
quantitative amounts of 2,4-dinitrophenol. Decomposition
magnitude greater than that of R2OH, i.e., ∆pKa ) pKR OH
-
2
1
pKR OH ≈ 9 (eq 6). That is the case of compounds 1a-h whose
reaction of compound 2 in aqueous media was monitored by
UV-vis spectrophotometry at appropriate wavelengths which
allowed us to follow either the disappearance of the starting
compound 2 (λ ) 250 nm) or the formation of 2,4-dinitrophenol
(λ ) 360 nm at pH > 4 and λ ) 340 nm at pH < 4). The
kinetics obtained at either wavelength given above were identical
and cleanly first-order for between 4 and 5 half-lives of reaction.
In contrast to compounds 1, the hydrolysis reaction for 2 was
found to be buffer-catalyzed in the whole pH range investigated
(pH 1-10). In the typical buffer concentration range of 0.05-
0.5 M, increases in kobsd due to catalysis of up to approximately
4-fold were observed. Values of the rate constants, k0, for the
buffer-independent rate constant for hydrolysis were determined
as the intercepts of plots of kobsd against buffer concentration.
Plot of log k0 as a function of pH is presented in Figure 2. The
data in Figure 2 (for 2) are well fit, at zero buffer concentration,
∆pKa values, as just defined, range from 11.4 to 5.6 on going
from 1a to 1h, respectively. The product distributions obtained
in 1 M DCl solution (Table 1) show that the relative amounts
of methanol and methyl carbamate vary linearly within this
∆pKa range from ca. 0 to 100% (see Figure S1 in the Supporting
Information). According to the linear equation of Figure S1,
an equal partitioning between methanol and methyl carbamate
should set up at ∆pKa ) 8.6 ( 1, a value that agrees well with
that estimated above.
However, any reaction pathway involving 1- as the single
reactive species is considered unlikely. Such a possibility might
be suggested on the basis of the above consideration and the
fact that compounds 1 have pKa values (Table 2) that are far
below the pKa values usually observed for >N-H bond
ionization. This is due to the combined electron-withdrawing
conjugative effects of sulfonyl and carbonyl groups which
strongly contribute to stabilizing the negative charge on the
nitrogen atom. There should therefore be a significant amount
of anion present in solution even in the more acidic media (e.g.,
at pH ) 1, more than 77% of 1a is present in the anionic form
at 25 °C), so that it does not seem unreasonable to envisage
that 1- could be the reactive species over the entire pH range
as suggested by Scheme 1. However, to be consistent with the
observed rate law at pH < 7, either reaction of Scheme 1 should
be general acid-catalyzed. In that case, the catalytic role of
hydronium ion would be to assistsby proton transfersdeparture
of the leaving groups ArO- and CH3O- during the course of
the two rate-limiting S-O and C-O bond fissions, respectively.
Reactions involving general acid/base catalysis and rate-limiting
proton transfer are known to exhibit a significant solvent
deuterium isotope effect kH2O/kD2O. While the reaction with
S-O bond cleavage shows relatively large isotopic dependence
(kH2O/kD2O ) 2 for 1a), that with C-O bond cleavage exhibits
only low values of kH2O/kD2O (see compounds 1f-h in Table
1). The absence of isotopic effect on kCO is then inconsistent
-
by the rate law of eq 4 for which k0 ) kH2O + kOHaOH . The
k
H2O and kOH values are (3.8 ( 0.5) × 10-7 s-1 and 11.6 ( 0.9
M-1 s-1, respectively.
kobsd ) kH2O + kOHaOH- + kB[B]
(4)
Deuterium Solvent Kinetic Isotope Effects and Activation
Parameters. These are summarized for compounds 1 and 2
in Table 3. For compounds 1, solvent deuterium isotope effects
measured in 1.0 M LCl solutions are presented in Table 1.
Discussion
Acid-Catalyzed Hydrolysis. Reactive Species. The sharp
difference obtained in the pH-rate profiles (Figure 2) of 1a and
its structurally related N-methylated analogue 2, for which a
simple bimolecular displacement by a molecule of water
probably occurs at pH < 5, is suggestive of different mechanistic
paths for the two ester types. Such a difference, very likely,
bears directly on the ability of substrates 1 to ionize at low pH.
Identity and distribution of products formed during the acidic
hydrolysis of compounds 1a-h are consistent with the mecha-
nistic pathway shown in Scheme 1 in which two competitive
E1cB reactions (kCO and kSO2) are involved. From the literature
data,14 one may estimate the rate constant ratio kCO/kSO2 at ca.
109 when the oxy leaving groups attached to either side of the
[(carbonyl)amino]sulfonyl moiety are identical. This large value
(23) The Bro¨nsted exponent, âlg, for the E1cB reaction of aryl N-
methylsulfamates in 50% aqueous ethanol is -1.8.14b In pure aqueous media,
the âlg values for the E1cB reactions of aryl N-methylcarbamates,14a aryl
carbamates14a and aryl sulfamates24 are -1.1, -1.15, and -1.2, respectively.

9580 J. Am. Chem. Soc., Vol. 120, No. 37, 1998
Blans and Vigroux
Scheme 2a
(2) The Bro¨nsted coefficient for the leaving group ArOH,
âlg, for the partial rate of release of methanol and arylsulfamates
(kCO) is 0, consistent with the absence of S-OAr bond cleavage
at the transition state I.
(3) The absence of detectable signal of Me18OH at the end
of the acidic hydrolysis reaction of 1d conducted in 42% 18O-
enriched water indicates that product methanol is not formed
from a bimolecular attack of water on the methyl group of
compounds 1 (an SN2 reaction with methyl-oxygen fission).
The results given in points 1-3 definitely locate the reaction
site at the -NHCOO- moiety.
(4) The possibility of reversible attack of water at the carbonyl
group of the neutral species 1 can be ruled out because of the
absence of 18O exchange from H218O into unreacted starting
material after 1 half-time of hydrolysis in 42% 18O-enriched
1.0 M HCl solution.
a In acid kobsd ) ka with ka ) kHKa + kz (paths a and c) or ka ) ki +
kz (paths b and c).
with a mechanism in which the expulsion of methanol from 1-
would be general acid-catalyzed.
The reaction pathway shown in Scheme 2 is consistent with
the rate law for decomposition of substrates 1 in acid (where
kobsd ) ka, see eq 1) and is not inconsistent with any other
experimental results obtained in the present work.
According to Scheme 2, the reaction with C-O bond cleavage
(kCO) involves protonation of the leaving group methanol and
its expulsion from the dipolar intermediate 1( with kz (path c)
and a transition state (I). This provides both an internal
nucleophile in the form of a nitrogen anion and an activated
methanol leaving group. As is shown in Scheme 2, our
(5) Also, the possibility that the nucleophilic addition of water
at the carbonyl group of the neutral form of 1 occurs at the rate
determining step is ruled out based on the following facts: (i)
The observed rate constants kCO (Table 1) are ca. 50-fold larger
than observed for hydrolysis of 2 in acid (kH2O ) 3.8 × 10-7
s-1, Figure 2). If the reactions of 1 that lead to methanol
proceed by addition of water to the carbonyl group, then the
rate constants kCO for compounds 1 should be similar to that
observed for 2 in acid, i.e. kH2O, which, very likely, corresponds
to nucleophilic addition of water to the carbonyl group. (ii)
Buffer bases are known to increase the nucleophilic reactivity
of water toward addition to the carbonyl group, and such
catalysis is generally associated with large values of kH2O/kD2O
and large negative values of ∆Sq as observed for 2 (see Results
and Table 3). In contrast, the kCO pathway for 1 is characterized
by the absence of both buffer acid-base catalysis and solvent
isotope effect (see compounds 1f-h in Table 1).
(6) Extrapolation of the Bro¨nsted plot log kp vs pKlg obtained
for the uncatalyzed decomposition of the anions of aryl
N-(phenylsulfonyl)carbamates,25 as indicated in footnote 26, and
taking into account the stabilizing effect of the phenolic oxygen
of 1 (estimated to ca. 2.3 kcal/mol27) suggests that the
zwitterions 1( involved in Scheme 2 may exist as discrete
experimental results do not distinguish between two possible
alternatives, paths a and b, for the reaction with S-O bond
cleavage, kSO2. Specifically, kH[H+] (path a) represents hydro-
nium ion-catalyzed decomposition of 1- with a transition state
(II), while ki (path b) refers to intramolecularly acid-catalyzed
expulsion of ArO-: the proton transfer to the leaving group
may either be direct (III) or via a bridging water molecule (IV);
the latter would avoid ring strain at the transition state and is
more likely. The kinetic ambiguity between paths a and b bears
directly on the failure to observe buffer catalysis for the reactions
of 1 (Results).
species in aqueous solutions28 with kz < 1 × 1010 s-1
.
(24) Thea, S.; Cevasco, G.; Guanti, G.; Williams, A. J. Chem. Soc., Chem.
Commun. 1986, 1582.
(25) The rate constant kp for the decomposition of the anions
PhSO2N-CO-OAr obeys the following equation: log(kp, s-1) ) -1.29
pKArOH + 8.4 (50 °C), see ref 12.
(26) (a) From the linear Bro¨nsted relationship given in footnote 25 one
may calculate pKROH2 values for which kp is greater than 1013 s-1. An
+
important assumption involved in this calculation is that the Bro¨nsted
relationship of footnote 25 can be applied to the zwitterionic species
PhSO2N-CO-HO+R with kp ) kz and neutral leaving groups ROH.
+
Otherwise this calculation gives pKROH2 < -3.6 (at 50 °C). Assuming
+
∆pKa ) pKROH - pKROH2 ≈ 17 (based on the fact that for methanol and
phenol ∆pKa is 15.5 - (-2.5) ) 18 and 10 - (-6.7) ) 16.7, respectively)
+
the trend obtained in terms of pKROH2 would then correspond to pKROH
<
ca. 13.4. (b) A similar calculation can be made to estimate the pKROH range
for which 1010 < kp ) kz < 1013, which corresponds to the kinetic case
where the diffusion-controlled formation of the zwitterion, arbitrary taken
to 1010 M-1 s-1, is partially or fully rate determining (in this case the
zwitterion is assumed to exist as an identifiable species but is too unstable
to diffuse through the solvent; also, the assumption that the formation of
the zwitterion may be limited by diffusion implicitly suggests that the proton
is transferred by a water mediated proton switch mechanism). This
calculation gives ca. 13.4 < pKROH < ca. 15.8 (at 50 °C).
(27) The C-O bond cleavage reaction leading to phenolate expulsion
from phenyl N-(phenoxycarbonyl)sulfamate anion (kp ) 9.36 × 10-7 s-1
at 50 °C, unpublished results) occurs 35 times slower than from phenyl
N-(phenylsulfonyl)carbamate anion (kp ) 3.27 × 10-5 s-1 at 50 °C, see ref
12). Assuming that the same stabilizing effect holds for the zwitterionic
species Ph(O)SO2N-CO-HO+R, one may re-estimate the pKROH range
obtained in footnote 26b for which 1010 < kz (Scheme 2) < 1013. This
gives 12.2 < pKROH < 14.6.
Methanol Route. The following accumulated evidence led
us to propose transition state I (path c) as the most likely
mechanism for the reaction with C-O bond cleavage.
(1) Aryl sulfamate esters were detected as intermediates
during the kinetic runs of compounds 1 while, in the same
reaction conditions, N-(methoxycarbonyl)sulfamate anion was
shown to hydrolyze to methyl carbamate, not methanol (see
Scheme 1). This indicates that the reaction site for methanol
formation is neither at the sulfur atom nor at the aromatic ring
but instead at the N-(methoxycarbonyl) moiety.

Aryl N-(Methoxycarbonyl)sulfamates in H2O
J. Am. Chem. Soc., Vol. 120, No. 37, 1998 9581
(7) The small values of kH2O/kD2O observed for kCO in 1.0 M
LCl are not incompatible with the elimination pathway c (eq
7). The solvent deuterium isotope effect on the mechanism of
eq 7 is due solely to secondary effects at exchangeable sites.
The expected effect can then be deduced on the basis of
fractionation factor theory.29 The expression for the solvent
deuterium isotope effect is presented in eq 8, where φNH, φN-,
φOH+, and kzH/kzD represent the fractionation factors for the NH
group, its anion in 1(, the oxonium ion in 1(, and the isotope
effect on kz (eq 7), respectively. A suitable value for φOH+ is
Figure 3. Values of log kSO2, the partial rate constant for the formation
of substituted phenols and methyl carbamate (the S-O cleavage
products) determined from product ratio, for hydrolysis of aryl
N-(methoxycarbonyl)sulfamates 1 at 50 °C in 1.0 M DCl solution
containing 16% CD3CN (µ ) 1.0 M), against pKa of the corresponding
leaving group ArOH. The linear regression equation is log kSO2
(-0.54 ( 0.04)pKArOH - (1.0 ( 0.3).
kCOH/kCOD ) φNH(1/φN-)(1/φOH+)kzH/kzD
(8)
)
the fractionation factor for the lyonium ion (φ ) 0.69);29 the
value of φNH/φN- ) 0.77 for 1f can be estimated from the
measured solvent deuterium isotope effect on equilibrium
is formed in association with methyl carbamate (corresponding
to 48.5% of total 4-nitrophenol formed in 1.0 M DCl solu-
tion).
(3) The latter result coupled with the fact that quite a good
correlation exists for all compounds 1, including 1a, between
the partial rate kSO2 and the pKa of the leaving group ArOH
(Figure 3) strongly suggests that (i) a single mechanism is
involved and (ii) that this mechanism is not SNAr.
(4) Although the transition state structure V may account for
the observed solvent isotope effect (Table 1) and âlg parameter
(see below), it is not likely in view of (i) the near-zero value
for the entropy of activation of 1a given in Table 3 (mechanisms
involving water-catalyzed nucleophilic attack of water as V are
commonly characterized by strongly negative entropies of
activation typically in the range -30-60 cal deg-1 mol-1), (ii)
the lack of buffer catalysis and (iii) the sharp difference obtained
between the pH-rate profiles of 1a and 2.
H
D
ionization of this compound, Ka /Ka , using φOH+ ) 0.69 for
the hydronium ion.30 Thus, the isotope effect on Ks (eq 7) is
found to be slightly greater than 1: KsH/KsD ) (φNH/φN-)(1/
φOH+) ) (0.77)(1.45) ) 1.12. The value of kzH/kzD is unknown,
but some consideration allows kz /kz ≈ 1 to be deduced.31
Thus, it is conceivable that the small values of kH2O/kD2O
observed for kCO (Table 1) could result from the solvent
deuterium isotope effect on Ks in eq 8.
H
D
Methyl Carbamate Route. The transition state structures
shown in II, III, and IV are supported by the following
accumulated evidence.
(1) Methyl carbamate was shown to be the hydrolysis product
of N-(methoxycarbonyl)sulfamate anion in acidic conditions (see
Scheme 1). The latter compound may result from the nucleo-
philic attack of water on the neutral species of 1 either at the
aromatic ring leading to C-O bond cleavage or at the sulfur
atom leading to S-O bond fission. However, the sharp
difference obtained in the pH-rate profiles (Figure 2) of 1a and
its N-methylated analogue 2, for which a simple bimolecular
displacement by a molecule of water probably occurs at pH <
5, is suggestive of different mechanistic paths for the two ester
types.32
(2) Partial formation of 4-nitrophenol via a SNAr mechanism
can be ruled out for 1d because of the absence of 18O
incorporation from H218O into the fraction of 4-nitrophenol that
(5) The Bro¨nsted coefficient for the leaving group, âlg,
obtained for kSO2 is -0.54 (Figure 3), consistent with some
degree of S-OAr bond cleavage at the transition state. Figure
S3 under Supporting Information shows the Hammett relation-
ship between the hydrolysis rate kSO2 and σ parameter for
electron-withdrawing para substituents. The correlation (r )
0.999) exhibits a F of 2.03 ( 0.06. This F value is probably an
overestimate of the expected F value in pure water because of
(28) An entity is not considered as an intermediate if its lifetime is less
than h/kT (see below), i.e., ca. 1.6 × 10-13 s at 25 °C, that is, if the rate
constant for its decomposition is on the order of a bond vibration frequency,
i.e., ca. 6.2 × 1012 s-1 at 25 °C. h/kT is the universal frequency factor
comprising the Boltzmann and Planck constants, and absolute temperature.
See: Frost, A. A.; Pearson, R. G. Kinetics and Mechanism, 2nd ed.;
Wiley: 1961; p 91.
(29) Kresge, A. J.; More O’Ferral, R. A.; Powell, M. F. In Isotopes in
Organic Chemistry; Buncel, E., Ed.; Elsevier: Amsterdam, The Netherlands,
1987; Vol 7.
(32) The apparent second-order rate constant for a hypothetical bimo-
lecular attack of water on the neutral 2,4-dinitrophenyl ester 1a, kH2O )
(30) The value of φNH/φN- for 1( can be estimated from the following
equation: KaH/KaD ) (φNH/φN-)(φOH+)-3. For compound 1f, ∆pKa ) pKaD
- pKaH ) 0.37 ( 0.04 (25 °C) gives φNH/φN- ) 0.77 ( 0.07. An important
assumption involved in this estimate is that the fractionation factor for the
nitrogen ion in 1- is the same as in 1(.
ka/55.5 ) 2.77 × 10-5 M-1 s-1 (50 °C), would be more than 3 orders of
magnitude larger than that for its N-methyl analogue 2 for which the real
second-order rate constant for the water-induced process is kH2O ) (3.8 ×
10-7)/55.5 ) 6.85 × 10-9 M-1 s-1 (50 °C). Such a difference is considered
as good evidence especially against a nucleophilic aromatic substitution
for 1a since, under the assumption of such a mechanism, similar rates would
have been expected in acidic media for both compounds (regardless of which
step of the SNAr mechanism is rate determining) due to the presumably
comparable electronic effects and leaving group abilities of the N-
(methoxycarbonyl)sulfamate moieties of 1a and 2 in the pH region where
the N-H bond remains undissociated.
(31) The driving force for methanol expulsion from 1( is likely to be
very large. As a result, the transition state I should be quite early, that is,
it should closely resemble 1(. Therefore, the value of kzH/kz ) (φN-
D
)(φOH+)(1/φN q)(1/φOH q) is expected to be maximal, that is, scarcely below
δ-
δ+
the upper limit of 1 since (φN q)(φOH q) < (φN-)(φOH+) (the fractionation
factor increases in going from 1( {(φN-)(φOH+) < 1} to the immediate
products of the reaction {φ ≈ 1} (eq 7)).
δ-
δ+

9582 J. Am. Chem. Soc., Vol. 120, No. 37, 1998
Blans and Vigroux
the presence of 16% of acetonitrile33 (see Experimental Section).
However, it means that there is a significant charge buildup on
the phenolic oxygen in going from the initial to the transition
state consistent with S-O bond cleavage.34 The “effective
charge”35 on the phenolic oxygen of the reactant ester 1- is
unknown, but its value is expected to be close to that observed
on the phenolic oxygen of related esters such as, e.g., monoan-
ions of aryl phosphate (+0.7436) and aryl sulfate (+0.737)
monoesters.38 Assuming that the “effective charge” is +0.7 in
1- and, therefore, > +0.7 in 1, then the âlg value of -0.54
implies that there must be +0.16 or >+0.16 unit of “effective
charge” on the aryl oxygen at the transition state depending on
whether the reactant species is 1- or 1, respectively. Such
positive values are consistent with the fact that the rate constants
kSO2 are well correlated with σ constants.
(6) There seems to be good evidence that the leaving group
expulsion from either 1- or 1 is assisted by proton transfer as
shown in II, III, or IV. The overall change of the “effective
charge” (ec) on the aryl oxygen in going from the reactant
species 1-/1 (ec g ca. +0.7 see above) to the final species ArOH
(ec ) 0) or ArO- (ec ) -1) is ca. -0.7 or -1.7, respectively.
A âlg of -0.54 associated to an overall charge change of -1.7
would indicate that the “effective charge” on the aryl oxygen
at the transition state is some 32% of the difference between
that in ground (compound 1) and product (ArO-) states. In
comparison, the âlg value for ArO- expulsion from 1- is -1.39
(vide infra) indicating that the S-O bond fission is well
advanced (82%) in the transition state. The fact that the
transition state for the expulsion of ArO- from 1 would be some
61% earlier than that from 1- is not consistent with expecta-
tion: 1- has more driving force than 1 to expel ArO- so that
the transition state for ArO- expulsion should be earlier for
1-. This “contradiction” with expectation can be resolved if
there is protonation of the leaving group as in mechanisms II,
III, or IV. In that case, the âlg value of -0.54 is associated to
an overall “effective charge” change of ca. -0.7 consistent with
a late transition state.
Figure 4. Plot of the log of kp, the spontaneous hydrolysis reaction
for aryl N-(methoxycarbonyl) sulfamates 1a-h at 50 °C (µ ) 1.0 M
with KCl), against pKa of the corresponding leaving group ArOH. The
linear regression equation obtained with compounds 1a-f is log kp )
(-1.39 ( 0.1)pKArOH + (2.43 ( 0.7), r ) 0.987. For compounds 1d-
h, values of kp were extrapolated at 50 °C from Eyring plots obtained
between 100 and 90 °C.
s-1 for leaving phenols of pKa < ca. 9.4 (at 50 °C),28 that is,
only a concerted mechanism is possible for compounds 1. Paths
a and b represent the two types of concerted general acid
catalysis that may be considered to “bypass” the zwitterionic
species in eq 9. As mentioned in an earlier section, unambigu-
ous distinction between inter- and intramolecular general acid
catalysis (i.e., between paths a and b, respectively) will depend
on the observation or not of buffer acid catalysis. While the
observation of buffer catalysis should readily resolve the kinetic
ambiguity in favor of path a, the failure to observe any buffer
effect, as in the case of compounds 1, should maintain the kinetic
ambiguity between the two paths.
“Neutral” Hydrolysis. In contrast to the acidic hydrolysis
reaction ka, the pH-independent “neutral” hydrolysis kp (corre-
sponding to the plateau regions around pH 7 in Figure 1) takes
place exclusively via S-O bond fission (see Results). The
hydrolysis rates for the anions of N-(methoxycarbonyl)sulfamate
esters 1a-f depend very strongly on the basicity of the leaving
group, the slope of log kp against pKlg being -1.39 (Figure 4).
This figure is consistent with a transition state in which the
Intermolecular (Path a) or Intramolecular (Path b)
General Acid Catalysis? The concerted general acid catalysis
of path a or b (Scheme 2) for the reaction with S-O bond
cleavage appears to be enforced39,40 by the disappearance of
the barrier for leaving group expulsion from the zwitterionic
species in eq 9. By extrapolating the Bro¨nsted line shown in
breaking of the bond to the leaving group is well advanced.
2- 41
This is typical of many phosphoryl (-PO3
)
and sulfuryl
- 42
(-SO3
)
group transfer reactions including hydrolysis. The
low Bro¨nsted exponent for attack of sodium azide and substi-
tuted pyridines on the sulfur center of 1a (ânuc ) 0.14, Figure
S2 in the Supporting Information) indicates weak interaction
with nucleophile. Thus, although the second-order kinetics
demands that the nucleophiles take part in the rate-limiting step,
rate constants are virtually independent of the basicity of the
nucleophiles, from pyridines with pKa of 1.45 to pyridines with
pKa of 9.20 (Tables S1 and S2 in the Supporting Information).
Again, this behavior is typical of phosphoryl and sulfuryl transfer
reactions.41,42 The large value of âlg associated with the small
Figure 4 (vide infra), in the same way as indicated in footnote
26, we find that the rate constant kz (eq 9) is greater than 1013
(33) For the dissociation of benzoic acids, a maximum increase of 141%
is observed for F at 25 °C in going from 100% water to 100% acetonitrile.
See: Kolthoff, I. M.; Chantooni, M. K. J. Am. Chem. Soc. 1971, 93, 3843.
(34) The exact charge on the aryl oxygen is not known and not directly
comparable to pure water, but the differences in conditions are small and
do not invalidate the conclusion that there is significant S-O bond cleavage
at the transition state.
ânuc coefficient is consistent with an “exploded” transition state
(35) Williams, A. AdV. Phys. Org. Chem. 1991, 27, 1.
for substitution that may occur by a concerted or stepwise
preassociation mechanism.43 Although there is evidence for free
intermediates of type O2SdNR when R ) H and Me,14b,24 we
failed in the present case to demonstrate that the putative
(36) Bourne, N.; Williams, A. J. Org. Chem. 1984, 49, 1200.
(37) Hopkins, A. R.; Day, R. A.; Williams, A. J. Am. Chem. Soc. 1983,
105, 6062.
(38) The ArO- species is defined as possessing 1 unit of negative charge
on its oxygen, in comparison with zero in the neutral phenol (ArOH). Thus,
the “effective charges” of +0.74 and +0.7 on the phenolic oxygen of
phosphate and sulfate monoesters, respectively, mean that -PO3H- and
-SO3- groups are effectively more electron withdrawing than the hydrogen
(proton) when covalently linked to an aryl oxygen.
(41) Cox, J. R.; Ramsay, O. B. Chem. ReV. 1964, 64, 317 and references
therein. Benkovic, S. J.; Schray, K. J. In The Enzymes; 3rd ed.; Boyer, P.
D., Ed.; Academic Press: New York, 1973; Vol. 8, pp 201-238 and
references therein.
(39) Jencks, W. P. Acc. Chem. Res. 1976, 9, 425.
(40) Jencks, W. P. Acc. Chem. Res. 1980, 13, 161.
(42) Williams, A.; Douglas, K. T. Chem. ReV. 1975, 75, 627 and
references therein.

Aryl N-(Methoxycarbonyl)sulfamates in H2O
J. Am. Chem. Soc., Vol. 120, No. 37, 1998 9583
N-(methoxycarbonyl)sulfimide species (below) could be a
liberated intermediate in aqueous solution.
attack at the aromatic ring followed by C-O bond cleavage
(see Results). For all compounds 1 (except for 1a and 1e), the
plot of log kOH against the pKa of the corresponding phenols is
linear with a slope âlg ) -0.15 (Figure S4, Supporting
Information). The latter value probably reflects the sensitivity
of the negative charge of 1- to be more extensively delocalized
to the sulfonyl moiety through the electron-withdrawing effects
of the aryl substituents. Consequently, compounds whose
negative charge will provide the lowest electrostatic barrier
toward OH- attack are those with powerful electron-withdraw-
ing aryl substituents as illustrated by the kOH values given in
Table 2.
It is of interest to note that the hydroxide ion catalyzed
reaction of 1a is characterized by a significantly more negative
entropy of activation (∆Sq ) -17 eu) than that observed for its
N-methylated analogue 2 (∆Sq ) -4.5 eu). Such a relatively
large difference may be explained in part by the fact that the
attack of OH- at the aromatic ring of 1a corresponds presumably
to the rate-determining step of the addition-elimination process
whereas for 2 the attack (either at the aromatic ring or at the
sulfur or carbonyl centers46) is likely to be faster than the
subsequent breakdown of the corresponding addition intermedi-
ate. Under this assumption, the expected ∆Sq value for 2 should
reflect both the associative and dissociative nature of the
stepwise process, that is, it should not be far from zero, as
observed.
With an increase in leaving group basicity there is a change
to a coupled concerted mechanism that is characterized by a
decrease in âlg and a large negative value of ∆Sq. There is a
sharp break in âlg plot as âlg changes from -1.39 to 0 and ∆Sq
changes from -10 eu for compounds 1a-c to -30 eu for
compounds 1e-h. This change is presumably due to the larger
nucleophilic assistance required for the cleavage of the S-O
bond with more basic leaving groups. The fact that the reaction
of 1a with pyridines is a nucleophilic displacement that occurs
at the same rate regardless of the nucleophilicity of the incoming
reagent suggests that water is also involved in the transition
state with weak binding to the sulfur center. However, as in
the case of phosphate and sulfate monoesters, the exact role
played by water is not clear. In effect, water might provide
part of the driving force for fragmentation either by differential
solvation of ground and transition states and/or as a nucleophile.
Under the latter hypothesis, the question as to whether water
behaves rather as a nucleophilic acceptor (to quench the forming
N-(methoxycarbonyl)sulfimide species) or as a true nucleophile
(which weakly binds to the sulfur center to assist leaving group
expulsion) remains uncertain for 1a-c. In contrast the role
played by water becomes clearer as the leaving group of
compounds 1 becomes poorer. In effect, the Bro¨nsted exponent
âlg for aryl N-(methoxycarbonyl)sulfamates 1 changes sharply
(from -1.39 to ca. 0) when the pKa of the leaving group exceeds
ca. 7.8 (Figure 4), which strongly suggests a changeover in
mechanism to stronger participation of water (as a nucleophile)
in the transition state.45 This is further confirmed by the
entropies of activation given in Table 3. The difference in the
observed ∆Sq values for compounds 1e-h (range) and 1a-c
(range) is such as to suggest considerably more involvement of
water in the transition state for the hydrolysis of 1e-h. In fact,
entropies of activation of the order of -30 eu are generally
associated with bimolecular mechanisms so that it seems
reasonable to envisage that the neutral hydrolysis of compounds
1e-h proceeds through an addition-elimination mechanism
with a transition state as VI.
Conclusions
The present study suggests that the -SO2NHCO- group may
be viewed as an attractive phosphate analogue in which the
naturally occurring oxyanions of phosphates have been replaced
by a nitrogen anion of comparable basicity. Interestingly,
according to the criteria developed by Westheimer10 the -SO2-
NHCO- linkage may be envisaged as a possible alternative to
phosphate esters in nucleic acids DNA and RNA. Indeed, as
phosphodiesters (1) it is conveniently made by ester bonds, (2)
it can link two groups (e.g. nucleosides) and still ionize at
physiological pH, and (3) it is hydrolytically stable. Moreover,
our detailed kinetic study of aryl N-(methoxycarbonyl)sulfamates
1 shows that the reaction of 1 that occurs with sulfur-oxygen
bond cleavage bears very strong similarities to the hydrolytic
mechanisms of both phosphate and sulfate monoesters. Other
aspects of the chemistry of the [(carbonyl)amino]sulfonyl group
are currently under investigation.
Acknowledgment. We thank professors James Fishbein
(Wake Forest University) and John Richard (University at
Buffalo, SUNY) for their critical reading and helpful comments.
We also thank Susy Richelme, Marie-Pierre Ferte´ and Eric
Leroy from the Service Commun de Spectrome´trie de Masse,
Universite´ Paul Sabatier, for their assistance with the GC/EIMS
technique. Special thanks go to Dr. Marcel Perry for his
invaluable help in developing original data analysis programs.
Hydroxide Ion Catalyzed Hydrolysis. The OH- reaction
observed with compounds 1b-h (except for 1e, see legend of
Figure 1) undoubtedly reflects attack of OH- at the carbonyl
moiety whereas with 1a the experimental evidence supports
Supporting Information Available: Spectroscopic and
analytical data for compounds 1 and 2; text giving details of
procedures for kinetic measurements, H218O experiments, and
pKa determinations; and figures S1-4 and Tables S1-4 (see
text) (13 pages, print/PDF). See any current masthead page
for ordering information and Web access instructions.
(43) Jencks44 has pointed out that a reaction is necessarily preassociative
if either the reaction intermediate, e.g., O2SdNCO2Me, does not exist (in
this case the mechanism is preassociative concerted) or if it is so unstable
that it collapses back to starting materials faster than the nucleophilic
acceptor can diffuse away from the complex (in this case the mechanism is
preassociative stepwise).
(44) Jencks, W. P. Chem. Soc. ReV. 1981, 10, 345.
JA981014W
(45) A somewhat similar change in âlg was observed for the E1cB
decomposition of N-methylsulfamate esters with poor leaving groups, see
ref 14b.
(46) Clearly our data on the hydrolysis of 2 are insufficient to elucidate
the mechanism in greater detail.