Intramolecular catalysis of phosphodiester hydrolysis by two imidazoles

Article abstract of DOI:10.1021/ja1034733

Two imidazole groups act together to catalyze the hydrolysis of the phosphodiester bis(2-(1-methyl-1H-imidazolyl)phenyl) phosphate (BMIPP). A full investigation involving searching computational and electrospray ionization (ESI-MS-/MS) and ultra mass spec

Full text of DOI:10.1021/ja1034733

Published on Web 05/28/2010
Intramolecular Catalysis of Phosphodiester Hydrolysis by
Two Imidazoles
Elisa S. Orth,† Tiago A. S. Branda˜o,† Bruno S. Souza,† Josefredo R. Pliego,‡
Boniek G. Vaz,§ Marcos N. Eberlin,§ Anthony J. Kirby,*,| and Faruk Nome*,†
INCT-Cata´lise, Departamento de Qu´ımica, UniVersidade Federal de Santa Catarina,
Floriano´polis, Santa Catarina 88040-900, Brazil, UniVersidade Federal de Sa˜o Joa˜o Del-Rei,
Sao Joao Del Rei, MG, 36301-160, Brazil, ThoMSon Mass Spectrometry Laboratory, Institute of
Chemistry, UniVersity of Campinas-UNICAMP, Campinas, SP, Brazil 88040-900, and UniVersity
Chemical Laboratory, Cambridge CB2 1EW, U.K.
Received April 23, 2010; E-mail: faruk@qmc.ufsc.br
Abstract: Two imidazole groups act together to catalyze the hydrolysis of the phosphodiester bis(2-
(1-methyl-1H-imidazolyl)phenyl) phosphate (BMIPP). A full investigation involving searching compu-
tational and electrospray ionization (ESI-MS-/MS) and ultra mass spectrometry (LTQ-FT) experiments
made possible a choice between two kinetically equivalent mechanisms. The preferred pathway,
involving intramolecular nucleophilic catalysis by imidazole, assisted by intramolecular general acid
catalysis by the imidazolium group, offers the first simple model for the mechanism used by the extensive
phospholipase D superfamily.
Scheme 1a
Introduction
Relatively few enzymes are capable of cleaving a P-OR bond
of an unactivated phosphodiester without the aid of a metal. So
those known to do sosand their number is growingsare of
particular interest. The key to a successful substitution reaction
at an unreactive phosphorus center is always the delivery of
the primary nucleophile. In the familiar ribonuclease mechanism
this is the neighboring 2′-OH group of the substrate, but
otherwise, in the absence of a metal cation, this is typically a
nucleophilic group from an active-site amino acid. In enzymes
belonging to the extensive phospholipase D superfamily, the
nucleophile is a histidine imidazole, thought to be assisted by
a second imidazole properly positioned by the extensive
common sequence motif, HisXLys(X)4Asp(X)6GlySerXAsn,
which occurs twice in the (typically dual-domain or dimeric)
holoenzyme.1-3 In the topoisomerases the nucleophile is the
hydroxyl oxygen of a tyrosine, but the hydrolysis of the
intermediate protein-tyrosine phosphate diester thus formed, at
the 3′-terminus of the DNA chain, is itself catalyzed by a
phospholipase D superfamily enzyme.4
a (A) Enzymes of the phospholipase D superfamily use two active site
histidines to catalyze the cleavage of unreactive phosphodiesters by a
nucleophilic mechanism involving a phosphoryl-enzyme intermediate 1. (B)
The intermediate is rapidly hydrolyzed by the reverse of the cleavage
mechanism, with water replacing the departing nucleotide ROH.
SN2(P) reaction in which nucleophilic attack is assisted by what
is effectively general acid catalysis of the departure of the
leaving group.
No simple model is available for this mechanism. We have
characterized intramolecular general acid catalysis (IGAC) by
the imidazolium group in the hydrolysis of the phosphate
monoester 2 as moderately efficient, by virtue of the intramo-
lecular hydrogen bond stabilizing the product 3 and the transition
state leading to it (Scheme 2).6
We also identified efficient IGAC of the attack of nitrogen
nucleophiles on phosphodiester phosphorus in system 4, which
was designed specifically to generate a strong intramolecular
H-bond in the product 5, and thus also in the transition state
leading to it. The pKa of the general acid in 4 is 7.4, close to
that of imidazole.
The mechanism suggested for this group of enzymes (specif-
ically for the “ubiquitous acid endonuclease” DNase II5) is
outlined in Scheme 1. It involves a (presumably concerted)
† Universidade Federal de Santa Catarina.
‡ Universidade Federal de Sa˜o Joa˜o Del-Rei.
§ University of Campinas-UNICAMP.
| University Chemical Laboratory.
(1) Gottlin, E. B.; Rudolph, A. E.; Zhao, Y.; Matthews, H. R.; Dixon,
J. E. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 9202–9207.
(2) Stuckey, J. A.; Dixon, J. E. Nat. Struct. Biol. 1999, 6, 278–284.
(3) Exton, J. H. ReV. Physiol., Biochem. Pharmacol. 2002, 144, 1–94.
(4) Interthal, H.; Pouliot, J. J.; Champoux, J. J. Proc. Natl. Acad. Sci.
U.S.A. 2001, 98, 12009–12014.
(5) Cheng, Y.-C.; Hsueh, C.-C.; Lu, S.-C.; Liao, T.-H. Biochem. J. 2006,
398, 177–185.
(6) Branda˜o, T. A. S.; Orth, E. S.; Rocha, W. R.; Bortoluzzi, A. J.; Bunton,
C. A.; Nome, F. J. Org. Chem. 2007, 72, 3800–3807.
9
10.1021/ja1034733  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 8513–8523 8513

A R T I C L E S
Orth et al.
Scheme 2. Mechanism of Intramolecular General Acid Catalysis of
the Hydrolysis of Phosphate Monoester 26
systems involving water as the primary nucleophile, and the
mechanisms by which enzymes activate water so effectively in
the absence of a metal remain an important challenge for model
design.10 RNA, and successful systems designed to model the
ribonuclease mechanism, has a nucleophilic OH group built in
(with EM ≈ 107 M in the reactions of ribose derivatives9).
We report new experimental and computational evidence from
a detailed investigation of the hydrolysis of the diester BMIPP,
which includes the identification of the imidazolium derivative
INT as a short-lived intermediate, thus defining our preferred
mechanism as intramolecular nucleophilic catalysis, concerted
with IGAC of the departure of the leaving group by the second,
imidazolium group (path B in Scheme 3).
Because of the extended, linear geometry of the nucleo-
phile· · ·P· · ·OR system in the transition state (Scheme 1), the
only way to introduce a neutral imidazole as a second catalytic
group into a simple diester based on 4 is to attach into the second
(“spectator”) ester group (e.g., the OMe in 4). Introducing a
neighboring imidazole in this way into 2 gave the system
BMIPP (bis(2-(1-methyl-1H-imidazolyl)phenyl) phosphate),
Results and Discussion
Kinetics. pH-rate profiles for the hydrolysis of the diester
BMIPP and the corresponding monoester 2-(1-methyl-2′-
imidazolium)phenyl phosphate (Me-2) are compared with
previous results for the nonmethylated monoester 2-(2′-imida-
zolium)phenyl phosphate (2)6 in Figure 1. The bell-shaped
profile for BMIPP shows a rate maximum at pH 6.6-6.8,
indicating bifunctional catalysis, involving both protonated and
neutral imidazole groups of the zwitterionic species (BMIPP().
which, like enzyme active sites deploying two histidines, has
two catalytic groups in potentially productive proximity of the
unreactive phosphodiester center. We found7 that BMIPP is
hydrolyzed approximately 107 times faster than expected for a
simple diaryl phosphate with leaving groups of similar pKa (∼
7.858), although this value is considerably higher when com-
pared with diphenyl phosphate (vide infra). To account for this
high reactivity we suggested the mechanism shown as path A
in Scheme 3, involving IGAC of intramolecular general base
catalysis (IGBC), and making the reaction formally a model
for the mechanism used by RNase A.7
However, the substantial combined effect of the two catalytic
groups could also be explained in terms of (the kinetically
equivalent) intramolecular nucleophilic catalysis (NUC) by the
neighboring imidazole free base, concerted with the expected
efficient IGAC (Scheme 3, path B). In the absence of a steric
or stereoelectronic barrier, intramolecular nucleophilic catalysis
is reliably more efficient than IGBC, with effective molarities
of up to 109 in conformationally flexible systems, compared
with a limit of about 60 M for IGBC, which has still not been
surpassed.9 This limit applies in particular to all known model
Figure 1. pH-rate profiles for the hydrolysis of BMIPP (2) and Me-2
(9), at 60 °C, I ) 1.0 (KCl). The solid lines for BMIPP and Me-2 represent
fits with eq 1. The bold curve shows the corresponding fit for hydrolysis of
2-(2′-imidazolium)phenyl phosphate (2) at 60 °C, I ) 1.0 (KCl).6
Direct titration of BMIPP gave consistent pKa values for the
imidazole groups of pKa2 ) 6.2 and pKa3 ) 7.8. (Rate and
dissociation constants are defined in Scheme 4.)
The data for BMIPP and Me-2 in Figure 1 were fitted to eq
1, based in Scheme 4.
Scheme 3. Alternative Mechanisms for the Rapid Hydrolysis of the
Phosphodiester BMIPP, Involving Intramolecular General Acid
Catalysis (IGAC) by the Imidazolium Group of General Base (Path
A) or Nucleophilic Catalysis (Path B) by the Imidazole Free Base
kobs ) k1dꢀ1d + k2d + k3dꢀ3d + k4dꢀ4d
(1)
The rate constants k1d, k2d, k3d, and k4d refer to the hydrolysis
of individual ionic species of BMIPP. The molar fractions ꢀ1d,
ꢀ2d, ꢀ3d, and ꢀ4d represent the dicationic, monocationic, zwitte-
rionic, and anionic BMIPP species, respectively. A similar
equation was used to analyze the results for monoesters 2 and
Me-2, and the subscript for molar fractions and rate constants
is m (Scheme 4). Ka1 is the dissociation constant for the
phosphoric acid group and is estimated from kinetic data in the
acidic (h0) range as approximately -1.5, consistent with values
assigned by other authors to similar phosphate esters.6,11 Ka2
and Ka3 (Scheme 4) are the dissociation constants of the
phosphoryl and imidazole groups of Me-2 and BMIPP,
9
8514 J. AM. CHEM. SOC. VOL. 132, NO. 24, 2010

Catalysis of Phosphodiester Cleavage by Imidazoles
A R T I C L E S
Scheme 4. Rate and Equilibrium Constants Accounting for the Hydrolysis of BMIPP over the pH Rangea
a The dashed arrows represent reactions too slow to be observed in this work.
Table 1. Kinetic Parametersa b Compared for the Hydrolysis of
,
respectively. The rate constants k2d and k4d in the diester
hydrolysis and k3m and k4m in the monoester reactions refer to
extremely slow reactions, which were not detected experimen-
tally, so were neglected in the fitting.
The kinetic parameters obtained by fitting the data in Figure
1 are presented in Table 1, and compared with previous results
reported for 2, all according to eq 1.6 The dissociation constants
pKa2 and pKa3 are from titration measurements for BMIPP and
2, while pKa2 for Me-2 was estimated from the kinetic results.
The result is consistent with the value obtained for 2.
Hydrolysis in Acid. Acid catalysis in the H0 region (Ham-
mett’s acidity function, H0 ) -log h0)13 corresponds to the
reaction of the dication BMIPP2+. The developing rate maxi-
mum at H0 ≈ -2, and the similar behavior of 2 in the same
region, is observed for the reactions of many aryl and naphthyl
phosphate esters which involve the attack of a water molecule
Phosphodiester BMIPP and Monoesters 2 and Me-2, at 60 °C
BMIPP
26
Me-2
k1, s-1 (1.52 ( 0.11) × 10-5 (8.67 ( 0.57) × 10-5
s
k2, s-1
s
(5.83 ( 0.18) × 10-6 (3.65 ( 0.06) × 10-6
k3, s-1 (1.98 ( 0.13) × 10-3
s
s
s
c
c
pKa1
pKa2
pKa3
-1.42 ( 0.05
-1.17 ( 0.16
d
d
c
6.10 ( 0.04
7.80 ( 0.05
4.67 ( 0.02
7.43 ( 0.04
4.52 ( 0.03
s
d
d
a The equations used for fitting BMIPP (eq S3) and Me-2 (eq S4),
b
are given in full in the Supporting Information. kobs values in the H0
region have been corrected for water activity (kobs/aw).12 c Estimated
kinetically at 60 °C. d Determined by potentiometric titrations at 25 °C.
The corresponding values calculated from the kinetic fit are pKa2 ) 6.12
and pKa3 ) 6.98 for BMIPP and pKa2 ) 4.90 and pKa3 ) 7.47 for 2.6
in the rate-limiting step. Values of kobs in the H0 region of Figure
1 have been corrected for water activity (kobs/aw).12 Further
evaluation of the acid hydrolysis of BMIPP2+ gave Bunnett
parameters w and Φ (calculated from eq 2 and 3: for data see
Supporting Information) of w ) 9.6 (n ) 4, R ) 0.979) and Φ
) 0.98 (n ) 4, R ) 0.979). These are similar to the values
estimated for the acid hydrolysis of 2 (w ) 11.6, ꢁ )1.31, n )
7, R ) 0.998),6 8-(N,N-dimethylamino)-1-naphthyl phosphates
(w ) 10.9, ꢁ )1.18, n ) 12, R ) 0.997),14 and other aryl
phosphate esters (w ≈ 7, ꢁ ≈ 1.2)15 where the suggested
(7) Orth, E. S.; Brandao, T. A. S.; Milagre, H. M. S.; Eberlin, M. N.;
Nome, F. J. Am. Chem. Soc. 2008, 130, 2436–2437.
(8) Rogers, G. A.; Bruice, T. C. J. Am. Chem. Soc. 1974, 96, 2463–2472.
(9) Kirby, A. J. AdV. Phys. Org. Chem. 1980, 17, 183–278.
(10) Kirby, A. J.; Medeiros, M.; Oliveira, P. S. M.; Brandao, T. A. S.;
Nome, F. Chem.sEur. J. 2009, 15, 8475–8479.
(11) Kirby, A. J.; Lima, M. F.; Silva, D.; Roussex, C. D.; Nome, F. J. Am.
Chem. Soc. 2006, 128, 16944–16952.
(12) Åkerlof, G.; Teare, J. W. J. Am. Chem. Soc. 1937, 59, 1855–1869.
(13) Anslyn, E. V.; Dougherty, D. A. Modern Physical Organic Chemistry;
University Science Books: Sausalito, CA, 2006.
9
J. AM. CHEM. SOC. VOL. 132, NO. 24, 2010 8515

A R T I C L E S
Orth et al.
Table 2. Efficiency of Intramolecular Catalysis of Phosphate Ester Hydrolysisa
Reaction
kcat/knoncat
kH/kD
Bifunctional intramolecular catalysis in the hydrolysis of BMIPP(
Intramolecular general acid catalysis in the hydrolysis of BMIPP2+
Intramolecular nucleophilic-general acid catalysis in the hydrolysis of 720
Intramolecular nucleophilic and general acid catalysis in the hydrolysis of 821
1011
107
1010
108
109
108
109
120
1.42
-
-
1.9
1.0
1.46
1.72
2.12
6
Intramolecular general acid catalysis in the hydrolysis of 2
Intramolecular general acid catalysis in the hydrolysis of 614
Intramolecular general acid catalysis in the hydrolysis of 411
Bifunctional general acid-base catalysis by ꢂ-cyclodextrin bis-imidazole in the
hydrolysis of 4-tert-butylcatechol cyclic phosphate19
a Efficiency (kcat/knoncat) was calculated by comparing with the hydrolysis of diphenyl phosphate (kH O ) 1.54 × 10-14 s-1 at 60 °C; applying
2
Arrhenius equation22).23 In the case of ꢂ-cyclodextrin bis-imidazole, efficiency was calculated by comparing with the spontaneous hydrolysis reaction of
4-tert-butylcatechol cyclic phosphate.
hydrolysis mechanism involves slow proton transfer concerted
with water attack.16
BMIPP2+. This suggests an important catalytic role for the
neutral imidazole, acting as a general base or nucleophile
(Scheme 3). Comparing hydrolysis rate constants (Table 1: k2
for 2 and Me-2; k3 for BMIPP() for the zwitterionic species of
the esters shown in Figure 1 shows that Me-2 is a little less
reactive than 2: presumably the N-methyl group decreases the
planarity of the arylimidazole aromatic system and, therefore,
reduces the catalytic efficiency of the imidazolium general acid.
On the other hand, the zwitterionic species BMIPP( is
hydrolyzed appreciably faster, by a factor of up to 540, than
the corresponding monoester Me-2. Phosphate monoesters,
including systems set up for IGAC, are generally more reactive
than the corresponding diesters: thus the monoester (6) derived
from 8-dimethylamino-1-naphthyl phosphate is hydrolyzed 18
times faster at 60 °C than the corresponding diester (4).11 The
enhanced reactivity of BMIPP( compared to Me-2 substantiates
the importance of the additional imidazole free base in what
must therefore be bifunctional catalysis.
log kobs + H0 ) w log aw + constant
(2)
(3)
log kobs + H0 ) φ(H0 + log[HCl]) + constant
The hydrolysis mechanism suggested for BMIPP2+ in the
H0 region is shown in Scheme 5, together with that proposed
for the reaction of 2 (from Scheme 2),6 where the departure of
the leaving group is assisted by general acid catalysis by the
protonated imidazole group as a water molecule attacks the
phosphorus atom. The same mechanism should apply also to
the hydrolysis of Me-2, although experimental data in the acid
region were not obtained for this ester, which was studied
specifically for comparison with the reaction of the reactive
species BMIPP( at pH > 6. Acid catalysis is also very efficient
in the hydrolysis of BMIPP2+, with an estimated rate enhance-
ment of >105 compared to phosphate esters without the
neighboring general acid.17 Upon comparison of the acid
hydrolysis (k1) constants in Table 1, the diester BMIPP2+ is
observed to be approximately 8 times less reactive than
monoester 2+, which shows one of the most efficient general
acid catalyzed reactions reported in the literature.6 This reduced
reactivity may result from a steric effect of the methyl group in
BMIPP2+, impeding the optimal conformation necessary for
the acid catalysis possible for 2+ (Scheme 5). There is no
evidence for the formation of a pentacovalent addition inter-
mediate, proposed for acid hydrolyses of some phosphate
esters,18 perhaps for steric reasons. In fact computational results
for 2 strongly disfavor this mechanism,6 a conclusion which
should apply to BMIPP2+ also.
This bifunctional intramolecular catalysis could involve
general acid assistance to either general base or nucleophilic
catalysis, since the two pathways (A and B in Scheme 3) are
kinetically equivalent. To compare the catalytic efficiency of
other intramolecular catalytic systems with that shown by
BMIPP, we adopted as a comparison standard the hydrolysis
reaction of diphenyl phosphate in the absence of catalysis, which
is one of the most common phosphate esters used for comparison
purposes in the literature, Table 2. Bifunctional general acid-base
catalysis by ꢂ-cyclodextrin bis-imidazole in the hydrolysis of
4-tert-butylcatechol cyclic phosphate is by far less efficient than
the intramolecular systems and shows a higher isotope effect.19
Intramolecular catalysis in the case of BMIPP( stands among
the most efficient: it is hydrolyzed at approximately the same
rate (half-life of 5.83 min at 60 °C) as the two efficient systems
7 and 8, based on two carboxy groups, which are hydrolyzed
with half-lives of 6.42 and 10.2 min at 50 and 39 °C,
respectively. All three phosphodiesters show pH-rate maxima,
but for the dicarboxy compounds 7 and 8 these are near pH
4.20,21 The balance between the contributions of the two catalytic
groups is expected to change, since the nucleophile is weaker
Scheme 5. Suggested Mechanism for the Hydrolysis of BMIPP2+
in Strong Acid, Compared with the Similar Reaction of 26 (See the
Text)
Bifunctional Catalysis near pH 7. The most reactive diester
species, BMIPP(, is hydrolyzed over 100 times faster than
(14) Kirby, A. J.; Dutta-Roy, N.; Silva, D.; Goodman, J. M.; Lima, M. F.;
Roussev, C. D.; Nome, F. J. Am. Chem. Soc. 2005, 127, 7033–7040.
(15) Bunton, C. A.; Farber, S. J. J. Org. Chem. 1969, 34, 767–772.
(16) Bunnett, J. F. J. Am. Chem. Soc. 1961, 83, 4956–4967.
(18) Bunton, C. A. Acc. Chem. Res. 1970, 3, 257–265.
(17) Kirby, A. J.; Younas, M. J. Chem. Soc. B 1970, 1165–1172.
(19) Breslow, R.; Schmuck, C. J. Am. Chem. Soc. 1996, 118, 6601–6605.
9
8516 J. AM. CHEM. SOC. VOL. 132, NO. 24, 2010

Catalysis of Phosphodiester Cleavage by Imidazoles
A R T I C L E S
unimolecular reactions,25 and values close to -25 e.u. are
commonly observed for bimolecular reactions:15 but these are
guidelines rather than rules. Thus ∆S‡ values of -3.9 e.u. and
-14.4 e.u. have been observed in relevant unimolecular
intramolecular nucleophilic reactions;21,25 and the formally
bimolecular reactions of methyl 8-dimethylamino-1-naphthyl
phosphate with water (∆S‡ ) -12.0 e.u) and three other
nucleophiles (∆S‡ ) 18.8-24.8 e.u.) differ significantly.11 Thus
∆S‡ ) -9.05 ( 0.5 e.u. also is consistent, but not uniquely
consistent, with the bifunctional nucleophilicsgeneral acid
catalysis mechanism for the hydrolysis of BMIPP(. A triester-
like mechanism involving protonation of a nonbridging oxygen26
can be ruled out in the hydrolysis of BMIPP(, since such
catalysis is known to cause small rate enhancements.26
Buffer Catalysis. The much higher EMs typically observed
for intramolecular nucleophilic catalysis9 mean that reactions
with the limited concentrations of external nucleophiles available
in solution cannot compete. So a simple but reliable test of
mechanism which is the presence or absence of such catalysis
was carried out. The hydrolysis reactions of phosphate esters
are often strongly accelerated in the presence of external
nucleophiles, especially the exceptionally reactive R-nucleo-
philes. For example, cleavage of bis(2,4-dinitrophenyl) phos-
phate is 104-fold faster,27 and the reaction of the triester diethyl
8-(N,N-dimethylamino)-1-naphthyl phosphate is 105-fold faster
in the presence of 0.5 M hydroxylamine, compared with the
reactions with water.28 We find that the rate of disappearance
of BMIPP is not affected by added 1 M acetate, 1 M imidazole,
or even the R-nucleophile hydroxylamine (also 1M), over the
pH range 4-8 at 60 °C. Fitting the data to eq 1 gives values
for k3 identical within experimental error (<1.3 × 10-4) to that
obtained previously (Table 1). (For details see Figure S8 and
Tables S9-S10 of the Supporting Information.) This is prima
facie evidence against the IGBC-IGAC mechanism (path A of
Scheme 3), unless the intramolecular general base catalysis
shows unprecedented efficiency.
Product Characterization by NMR. 1H and 31P NMR spectra
taken at regular time intervals during the course of BMIPP
hydrolysis in D2O are shown in Figures 10 and 11 in the
Experimental Section. Three species were observed: BMIPP,
which gives the stable phenol product Me-3 plus the monoester
Me-2 as an intermediate, which is itself hydrolyzed more slowly
to a second molecule of Me-3 and inorganic phosphate.
Figure 2 shows relative concentrations of the three compounds
measured by 31P NMR as they appear or disappear during the
course of the reaction. The solid curves in the figure represent
fits to a consecutive-reaction model. The derived rate constants
(kobs) for the hydrolysis of BMIPP at pD 7.2 and of Me-2 at
pD 1.5 are 3.0 × 10-4 and 2.7 × 10-6 s-1, respectively.
Computational Calculations. To better understand the struc-
tures and likely behavior of reactants, transition states, and
possible intermediates in the reaction path leading from BMI-
PP( to Me-2 and Me-3, we performed computational calcula-
tions at the B3LYP level to compare the two mechanisms
described in Scheme 3: intramolecular general base catalysis
of the attack of a water molecule on phosphorus (path A) and
intramolecular nucleophilic attack of the imidazole group on
Table 3. Kinetic Parameters for the Hydrolysis of BMIPP, Derived
from Data for kobs Measured between pH 5.5-8.0 (See Supporting
Information)
a
a
Conditions
k3 10-3 s-1
pKa2
pKa3
50 °C, H2O
60 °C, H2O
70 °C, H2O
80 °C, H2O
60 °C, D2O
0.77
1.98
3.52
7.15
1.39
6.24
6.12
5.80
5.58
6.91
6.99
6.98
6.83
6.79
7.51
a Values of pKa2 and pKa3 were estimated from the fit of the kinetic
data at each temperature.
in the case of the less basic carboxylate, but the COOH a
correspondingly stronger general acid. The acyl phosphates
produced by the initial nucleophilic attack have been identified
as intermediates in the reactions of 7 and 8,20,21 so the
involvement of the stronger, imidazole, nucleophile in a system
with almost identical geometry and closely similar leaving group
would not be surprising.
Kinetic Isotope Effect and Thermodynamic Parameters. The
hydrolysis of the zwitterionic species BMIPP( is characterized
by a low deuterium isotope effect, kH O/kD O ) 1.42 (data from
2
Table 3), consistent with a single proton tra2nsfer in the transition
state for the bifunctional intramolecular catalysis.6,24 Low values
close to unity are expected for the nucleophilic part of the
mechanism (e.g., kH/kD ) 1.10 for the hydrolysis of 2-carbox-
yphenyl-3-nitrophenyl phosphate, which involves nucleophilic
intramolecular catalysis by carboxylate25) but are observed also
for various reactions involving intramolecular general acid
catalysis of the hydrolysis of phosphate diesters (Table 2). Thus
kH/kD ≈ 1.70 for the hydrolysis of 4,11 where the dimethylam-
monium center promotes efficient intramolecular acid catalysis,
and kH/kD ) 1.9 for the hydrolysis of the anion of 8, which
involves bifunctional nucleophilic-general acid catalysis by the
two carboxy groups.21 The kinetic isotope effect, kH O/kD O
)
2
2
1.42, observed for the hydrolysis of BMIPP(, is consistent,
though not uniquely consistent, with the bifunctional nucleo-
philic-general acid catalysis mechanism and so is not of itself
sufficient evidence to rule out the mechanism (path A in Scheme
3) involving two rather than one proton transfers in the rate-
determining transition state.
The activation parameters calculated for the hydrolysis of
BMIPP(, with the rate constants given in Table 3, are ∆H‡ )
19.0 ( 1.2 kcal/mol and ∆S‡ ) -9.05 ( 0.5 e.u., giving a
Gibbs free energy, calculated at 25 °C, of 21.7 ( 1.4 kcal mol-1.
In principle, entropy values close to zero are expected for
(20) Kirby, A. J.; Abell, K. W. Y. J. Chem. Soc., Perkin Trans. 2 1983, 8,
1171–1174.
(21) Bruice, T. C.; Blasko´, A.; Arasasingham, R. D.; Kim, J. S. J. Am.
Chem. Soc. 1995, 117, 12070–12077.
(22) Jencks, P. J. Catalysis in Chemistry and Enzymology; Dover Publica-
tions, Inc.: New York, 1987.
(26) Anslyn, E. V.; Piatek, A. M.; Gray, M. J. Am. Chem. Soc. 2004, 126,
9878–9879.
(23) Kirby, A. J.; Younas, M. J. Chem. Soc. B 1970, 510–513.
(24) Khan, S. A.; Kirby, A. J. J. Chem. Soc. B 1970, 1172–1182.
(25) Khan, S. A.; Kirby, A. J.; Wakselman, M. J. Chem. Soc. B 1970,
1182–1187.
(27) Domingos, J. B.; Longhinotti, E.; Bunton, C. A.; Nome, F. J. Org.
Chem. 2003, 68, 7051–7058.
(28) Kirby, A. J.; Tondo, D. W.; Medeiros, M.; Souza, B. S.; Priebe, J. P.;
Lima, M. F.; Nome, F. J. Am. Chem. Soc. 2009, 131, 2023–2028.
9
J. AM. CHEM. SOC. VOL. 132, NO. 24, 2010 8517

A R T I C L E S
Orth et al.
Table 4. Selected Distances (Å) and Angles (deg) for Optimized
Structures of BMIPP(.w, TS.GB, TS.NUC, and TS.NUC2a b
,
BMIPP(.w
TS.GB
TS.NUC
TS.NUC2
O3-H1
O4-P
O4-H1
O5-P
N1-H2
N1-P
N2-H1
O1-P-O3-O2
1.50
1.70
2.50
4.02
2.19
4.46
1.09
131.40
2.56
2.32
1.66
2.20
1.73
1.68
2.32
2.26
3.78
2.76
1.67
2.32
2.27
-
-
3.94
1.05
1.90
1.05
1.94
1.05
-165.21
-176.67
-162.67
a B3LYP calculations with basis sets: O, N, and P (6-31+G*), C
(6-31G*), and H (6-31G). b Names are defined in Figures 4, 5, and 7.
Starting from BIMPP(.w the two pathways of Scheme 3 were
evaluated: activation of the coordinated water, to act as the
nucleophile, and direct nucleophilic attack of N1 on the P atom.
The transition states involved in the two mechanisms are labeled
TS.GB and TS.NUC, respectively, and are shown in Figure 4.
Relevant structural details are given in Table 4.
In TS.GB, the imaginary frequency observed corresponds to
O5-P attack, with simultaneous O4-P bond breaking; N1
modestly activates the water molecule, and H1 strongly stabilizes
the leaving group departure through general acid catalysis, with
an O4-H1 bond length of 1.66 Å and a quasi planar arylimi-
dazole aromatic system (O1-P-O3-O2 ) -176.67°). Fur-
thermore, the strong hydrogen bonding between O3-H1 in
BMIPP(.w is not observed in TS.GB due to the strong O4-H1
hydrogen bond.
For nucleophilic attack, structure TS.NUC shows an advanced
transition state with a higher degree of bond formation between
N1-P (1.90 Å) compared with O5-P in TS.GB (2.20 Å) and
an O1-P-O3-O2 dihedral angle of -162.67°. The stabilization
of the leaving phenoxide group by O4-H1 hydrogen bonding
observed in TS.GB is absent in TS.NUC, with a preferred
stabilization of the negative charge at O3, which is hydrogen
bonded to H1 (O3-H1 ) 1.68 Å) as in BMIPP(.w. Aside from
the bond making and breaking at the phosphorus center, the
Figure 2. Relative species concentrations, obtained from repeated 31P NMR
scans over time, for hydrolysis of BMIPP at pD 7.2 and Me-2 at pD 1.5,
in D2O at 60 °C: (0) BMIPP, (b) Me-2, (4) Pi.
phosphorus, also with a water molecule present (path B). Tables
of Cartesian coordinates and thermochemistry data appear in
full in section 3 of the Supporting Information.
In both cases, the first step should be the solvation of the
phosphate zwitterion. Although there is a strong interaction
between O3 and the neighboring imidazolium group, which
partially neutralizes the phosphate negative charge, BMIPP(
interacts favorably with the water molecule through the N1 atom
oftheimidazolegroupandO2,toformtheorganicphosphate-water
complex BMIPP(.w (Figure 3). This interaction has a ∆Geq of
-2.78 kcal ·mol-1, thus BMIPP(.w is taken to be the reacting
species. The use of the gas phase to simplify the computational
calculations is reinforced by the fact that the addition of the
∆Gsolv term, computed by single-point calculations using PCM,
gives very similar results and the free energies differences are
not affected to a large extent (see sections 3.1 and 3.2 of the
Supporting Information). Therefore, all structures and thermo-
dynamic parameters presented here are computed in the gas
phase, without the inclusion of the ∆Gsolv term.
Figure 3. B3LYP gas phase optimized structures of BMIPP( and BMIPP(.w showing the ∆Geq involved in the organic phosphate-water complex formation.
The basis set used is given in the Experimental Section.
Figure 4. B3LYP gas phase optimized structures of TS.GB and TS.NUC. The basis set used is given in the Experimental Section.
9
8518 J. AM. CHEM. SOC. VOL. 132, NO. 24, 2010

Catalysis of Phosphodiester Cleavage by Imidazoles
A R T I C L E S
Table 5. Experimental and B3LYP Gibbs Energy, Enthalpy, and
Entropy of Activation for Structures TS.GB and TS.NUCa Energies
Relative to BMIPP(.w
∆G‡/kcal·mol-1
∆H‡/kcal·mol-1
∆S‡/cal·mol-1 ·K-1
Experimental
TS.GB
TS.NUC
21.7 ( 1.4b
28.4
18.9
19.0 ( 1.2
29.2
18.6
-9.0 ( 0.5
2.7
-1.2
a All quantities calculated at 298.15 K. B3LYP calculations with the
following basis set: O, N, and P (6-31+G*), C (6-31G*), and H
(6-31G). b Calculated at 298.15 K.
Figure 6. B3LYP gas phase optimized of TS.NUC2. The basis set used is
given in the Experimental Section.
Figure 5. Free energy variation along the reaction coordinates involving
TS.GB and TS.NUC. Energies values relative to BMIPP(.w. Structures
calculated at the B3LYP level with the following basis sets: O, N, and P
(6-31+G*), C (6-31G*), and H (6-31G).
major structural difference between TS.NUC and BMIPP(.w
is the twisting of the imidazole ring needed to put N2 in a
reasonable position to attack P.
Figure 7. Variations of bond lengths and energies (∆) along the IRC for
the intramolecular nucleophilic attack reaction in BMIPP( via TS.NUC2
(Note that H1 is within strong H-bonding distance of O4).
Although paths A and B (Scheme 3) are kinetically equiva-
lent, the calculated ∆G‡ is some 9.5 kcal ·mol-1 lower for
TS.NUC than for TS.GB as shown in Table 5, a result which
clearly indicates the nucleophilic pathway as the favored route
for the reaction. For reaction via TS.NUC, the calculated ∆G‡
is only 2.79 kcal ·mol-1 lower than the experimental value and
the agreement of the calculated ∆H‡ is even closer, differing
by only 0.44 kcal ·mol-1 from the experimental value. Although
there is a considerable difference between the experimental and
calculated ∆S‡, the theoretical value is in accord with that
expected for an intramolecular reaction and the difference is
most likely associated with reorganization of solvent molecules,
which is not of course addressed by our gas phase calculations.
As shown in Figure 5, the decomposition of TS.GB leads to
arylimidazole (Me-3) + Me-2, while TS.NUC decomposes to
Me-3 + INT. Calculated structures of the reaction products are
given in the Supporting Information. The reaction path involving
TS.NUC is more efficient and results in the formation of Me-2
+ INT, which are 7.16 kcal ·mol-1 higher in energy than the
actual reaction products ((Me-3) + Me-2). Since the phospho-
rane intermediate was not detected in any of the earlier NMR
experiments, the calculations indicate that INT decomposes
rapidly under mild reaction conditions.
As shown in Table 4, the main bond lengths and angles of
TS.NUC and TS.NUC2 are very similar, as is the central bond
making-breaking process. Thus (to minimize computational
expense) we performed an IRC analysis on TS.NUC2: Figure
7 shows the changes in bond lengths involving this structure.
The approach of N1 toward P can be seen to be concerted with
O4 departure, while H1 is transferred to N2, stabilizing the
leaving group.
Mass Spectrometric Measurements. Building on our experi-
ence in reaction mechanism studies using MS techniques,29,30
we applied ESI-MS-(/MS) and LTQ-FT to monitor the course
of the hydrolysis of BMIPP( in aqueous solution. In the ESI-
MS process used here, solvated ions are “fished” directly from
solution and transferred to the gas phase. The data provide
snapshots of the ions present in the reaction solution, which
have been shown in numerous cases to reflect accurately the
actual ionic composition (see the cited references and a recent
Wiley book ReactiVe Intermediates: MS InVestigations in
Solution, edited by Leonardo S. Santos). After 25 min (4 half-
lives) of reaction at pH 6.5 and 60 °C, samples of reaction
solution, containing all reagents, intermediates, and products,
were transferred directly to the gas phase and detected by
In TS.NUC, the water molecule is not directly involved in
the bond making-breaking process. Thus, we also modeled the
nucleophilic attack of the imidazole group toward P without
the assisting H2O: the results are very similar. Starting from
the BMIPP( zwitterion, ∆G‡ for the N1 attack is 19.57
kcal·mol-1, which is also in excellent agreement with the
experimental value (Table 5): the corresponding transition state
structure TS.NUC2 is shown in Figure 6.
(29) Domingos, J. B.; Longhinotti, E.; Branda˜o, T. A. S.; Santos, L. S.;
Eberlin, M. N.; Bunton, C. A.; Nome, F. J. Org. Chem. 2004, 69,
7898–7905.
(30) Orth, E. S.; da Silva, P. L. F.; Mello, R. S.; Bunton, C. A.; Milagre,
H. M. S.; Eberlin, M. N.; Fiedler, H. D.; Nome, F. J. Org. Chem.
2009, 74, 5011–5016.
9
J. AM. CHEM. SOC. VOL. 132, NO. 24, 2010 8519

A R T I C L E S
Orth et al.
Figure 8. ESI(+) FT-ICR MS after 10 min of hydrolysis of BMIPP in aqueous solution at pH 6.5 and 60 °C.
ESI(()-MS and then the gaseous ionic species transferred to
the gas phase were characterized by ESI(()-MS/MS. Initially,7
ESI-MS-(/MS) showed a series of the major anions from
Scheme 4, which were identified as: (i), the deprotonated form
of the reactant BMIPP of m/z 409; (ii) the monoester Me-2 of
m/z 253; (iii) the phenolate Me-3 of m/z 17; and (iv) inorganic
other as the conjugate acid (paths A and B of Scheme 3). The
imidazolium group clearly acts as a general acid to assist the
departure of the leaving group, but the neutral imidazole could
be involved either as a general base, catalyzing the attack of
water, or directly as a nucleophile, in which case the initial
product must be the cyclic phosphoimidazole INT. There is no
kinetic evidence for the presence of INT during the course of
the reaction. There are good reasons based on the overall
efficiency of catalysis to interpret the kinetic evidence in terms
of the nucleophilic mechanism, but these are not sufficient to
rule out the general base mechanism pathway (A in Scheme 3)
completely. The general base catalysis mechanism was preferred
initially because it is consistent with the kinetic parameters and
because no cyclic intermediate was detected by ESI-MS (in
either positive or negative mode) or by NMR (1H and 31P).
However, the phosphoryl-imidazolium (P-N+) bond is un-
stable,31 so that the nondetection of the intermediate could be
the result of its short lifetime, before breaking down to products,
which are the same from either mechanism. New calculations
strongly favor the nucleophilic mechanism and encouraged us
to investigate the possible presence in the reaction mixture of
low concentrations of INT, using more sensitive ES-MS
techniques. The positive result obtained using LTQ-FT (+)
(Figure 8), showing that INT is indeed present during the course
of the reaction, provides unambiguous support for the nucleo-
philic-IGBC mechanism (path B of Scheme 3). Encouraged by
the positive identification of INT and to obtain positive evidence
in the aqueous phase, we are currently studying the stability of
a series of phosphoryl-imidazolium (P-N+) derivatives, to
obtain better estimates of the expected reactivities of N-
phosphoryl imidazolium systems. Though we cannot rule out
the possibility of a small contribution from the IGBC pathway,
the application of Occam’s razor32 is well justified by the general
arguments from catalytic efficiency and the unambiguous
indications from our calculations.
-
P in the form of PO3 of m/z 79 and H2PO4 of m/z 97. No
cyclic intermediate (INT) could be detected by ESI-MS in either
negative or positive ion modes, which does not rule out its
formation, since this cyclic species could be relatively unstable.
However, some important species could be detected and
characterized via collision-induced dissociation in ESI(-)-MS/
MS experiments, with the resulting spectra supporting the
structural assignments. The ESI(-)-MS/MS of the Me-2 anion
of m/z 253 m/z shows that it dissociates almost exclusively to
-
PO3 of m/z 79. Deprotonated Me-3, of m/z 173, dissociates
mainly by two routes that lead either to the fragment ion of
m/z 158 by the loss of a methyl radical or the fragment of m/z
118, most likely by losing 1-methyl-1H-azirine (C3H5N) to form
the 2-cyanophenoxy anion.
To address further the possibility of the intramolecular
nucleophilic pathway, ultrahigh resolution and accuracy LTQ
FT-ICR MS data, which combine the most advanced Ion Trap
and Fourier Transform Ion Cyclotron Resonance technologies,
were acquired in both the negative and positive mode. In the
negative ion mode, the same anions identified by Q-TOF were
detected. In the positive mode (Figure 8), an ion of m/z 237
was intercepted and assigned to the cyclic intermediate (INT).
Indeed, the program Xcalibur 2.0 (Thermo Scientific), which
calculates molecular composition to within 0.4 ppm, identified
C10H24O3N1P1 as the most probable composition for the ion of
m/z 237.14. Thus, the hydrolysis of BMIPP( is seen to involve
intramolecular nucleophilic attack by an imidazole group.
Conclusions
We have now studied the hydrolysis of BMIPP( in consider-
able depth and have gathered enough evidence to assign the
mechanism with some confidence. The kinetic evidence indi-
cates bifunctional catalysis of the hydrolysis of BMIPP(
involving the two imidazole groups, one as the free base, the
(31) Attwood, P. V.; Piggott, M. J.; Zu, X. L.; Besant, P. G. Amino Acids
2007, 32, 145–156.
(32) Bergson, G.; Linderberg, J. J. Phys. Chem. A 2008, 112, 4235–4240.
9
8520 J. AM. CHEM. SOC. VOL. 132, NO. 24, 2010

Catalysis of Phosphodiester Cleavage by Imidazoles
A R T I C L E S
Figure 9. pH-rate profile for the hydrolysis of BMIPP in (A) H2O (-9-) and D2O (-0-); (B) at 50 °C (-0-), 60 °C (-2-), 70 °C (-b-), and 80 °C (-9-).
The curves show the fits using eq 1, and the parameters obtained are presented in Table 3. The points in H2O at 60 °C, shown for comparison, are from
Figure 1.
1
Figure 10. Progressive H NMR spectra of the hydrolysis reaction of BMIPP. (a) Hydrolysis of BMIPP at pD 7.2 and (b) hydrolysis of the intermediate
monoester Me-2 at pD 1.5, in D2O, at 60 °C. Signals are identified in Table 6.
Experimental Section
hydroxyphenyl)imidazole (500 mg, 2.87 mmol in 15 mL) in an
ice-water bath, under argon. The mixture was stirred at room
temperature for 120 min. Water was then added (0.25 mL), and
the mixture was left to react overnight. The solvent was removed
under reduced pressure, and methanol (5 mL) and water (3 mL)
were added to the resulting crude oil. Fine white crystals (370 mg,
51%) of BMIPP were obtained immediately. 13C NMR (100 MHz,
D2O, internal reference TSP) δ 149.91, 141.57, 134.49, 131.74,
Materials. Inorganic salts were of analytical grade and were used
without further purification. Liquid reagents were purified by
distillation. 1-Methyl-2-(2′-hydroxyphenyl)imidazole (Me-3) was
prepared by the method of Rogers and Bruice.8
Synthesis of Bis(2-(1-methyl-1H-imidazolyl)phenyl)phosphate
(BMIPP). A solution of PCl5 (600 mg, 2,87 mmol) in CHCl3 (15
mL) was added dropwise to a CHCl3 solution of 1-methyl-2-(2′-
9
J. AM. CHEM. SOC. VOL. 132, NO. 24, 2010 8521

A R T I C L E S
Orth et al.
Table 6. NMR Spectra of the Species in the Hydrolysis Reaction
of BMIPP, in D2O at 60 °C
stored directly on a microcomputer, and observed first-order rate
constants, kobs, were calculated from linear plots of ln(A∞ - At)
against time for at least 90% of the reaction using an iterative least-
squares program; correlation coefficients were better than 0.999
for all kinetic runs. The pH was maintained with 0.01 M buffers
of CH2ClCOOH (pH 2-3), HCOOH (pH 3-4.5), CH3COOH (pH
4-5.5), and NaH2PO4 (pH 5.5-6.0). In solutions where [HCl] >
0.5 M, the reactions were self-buffered with no ionic strength
correction. Reactions at pD 7.2 and pD 1.5 were also followed by
NMR to follow the appearance of products and disappearance of
reagents.
1H NMR
δ (ppm)
31P NMR
δ (ppm)
Compound
BMIPP
(pD ) 7.2)
Me-2
(pD ) 7.2)
Me-2
(pD ) 1.5)
Me-3
(pD ) 7.2)
Me-3
δ 3.79 (s, 6H, Me), 7.5-7.6 (m, 8H, Ar),
7.70 (d, 2H, Ar), 7.81 (t, 2H, Ar)
δ 4.03 (s, 3H, Me), 7.5-7.8 (m, 5H, Ar),
7.88 (t, 1H, Ar)
δ 3.62 (s, 3H, Me), 7.5-7.8 (m, 5H, Ar),
7.56 (t, 1H, Ar)
δ 3.97 (s, 3H, Me), 7.3-7.4 (m, 2H, Ar),
7.5-7.8 (m, 4H, Ar)
δ 3.62 (s, 3H, Me), 6.9-7.0 (m, 2H, Ar),
7.30 (d, 1H, Ar), 7.34 (s, 2H, Ar), 7.41
(t, 1H, Ar)
-10.38
1.22
-4.30
Since the pH-rate profile for the hydrolysis of BMIPP( is bell
shaped (at all temperatures investigated and in both H2O and D2O),
the solvent deuterium kinetic isotope effect and activation param-
eters were calculated from the rate constants k3d obtained by fitting
experimental data in the range of pH 5.5-8.0, as shown in Figure
9A and B. The values obtained (including the acid dissociation
constants pKa2 and pKa3) are shown in Table 3. The kinetic isotope
effect is defined by k3H O/k3D O, and activation parameters were
(pD ) 1.5)
Pi
0.57
125.28, 124.02, 120.56, 119.22, 114.41, 34.99; 1H NMR (400 MHz,
D2O, internal reference TSP) δ 3.48 (s, 6H, CH3), δ 7.13 (d, 2H,
JAB ) 8.4 Hz), δ 7.23 (dt, 2H, JCB ) 7.6 and JCD ) 7.6 Hz), δ
2
2
7.30 (dd, 4H, JFG ) 5.2 and JGF ) 5.2 Hz), δ 7.37 (d, 2H, JDC
)
calculated by plotting ln(k3d/T) vs 1/T, according to Eyrings’
7.6 Hz), δ 7.48 (dt, 2H, JBC ) 7.6 and JBA ) 8.4 Hz). 31P NMR
(81 MHz, D2O, external reference H3PO4 85%) δ -10.38 ppm, s;
m/z (ESI-MS) 409,1720 C20H18N4O4P requires 409.1066
equation.13
Mass Spectrometric Measurements. To identify intermediates
and reaction products of BMIPP hydrolysis, direct infusion
electrospray ionization mass spectrometry was performed using a
hydrid triple quadrupole linear ion-trap mass spectrometer.30 The
typical electrospray ionization (ESI-MS) experiment involved 1 mL
of aqueous 1 × 10-6 M BMIPP solution at pH 6.5. A microsyringe
pump delivered the reagent solution into the ESI source at a flow
rate of 10 µL/min. ESI and the QqQ (linear trap) mass spectrometer
were operated in the negative-ion mode. Main conditions: curtain
gas nitrogen flow of 5 mL min-1; capillary voltage of -3500 eV;
cone voltage of -30 eV. Some of the main anionic species detected
by ESI-MS were subjected to ESI-MS/MS by using collision-
induced dissociation (CID) with nitrogen, with collision energies
ranging from 5 to 45 eV for the best dissociation yields. For accurate
mass measurements of possible intermediates, direct infusion
automated chip-based nano-ESI-MS were performed on a Triversa
NanoMate 100 system in both the positive and negative ion modes.
Samples were loaded into 96-well plates (total volume of 100 µL
in each well) and analyzed by a 7.2T LTQ FT Ultra mass
Potentiometric Titrations. The pKa values of BMIPP were
determined using a digital pH meter and a combined glass electrode,
equipped with an automatic buret. Titrations were performed in a
150 mL thermostatted cell, under N2 at 25.0 °C, at ionic strength
0.1 M KCl, and [BMIPP] ) 1.0 mM (initial). The solution was
titrated with small increments of 0.1008 M CO2-free KOH. All
precautions were taken to eliminate carbonate and CO2 during the
titration. The program BEST733 was used to calculate the dissocia-
tion constants, using the value -13.78 for pKw.
Kinetics. Reactions of BMIPP and Me-2 were followed spec-
trophotometrically by monitoring the appearance of 1-methyl-2-
(2′-hydroxyphenyl)imidazole (Me-3) at 290 nm. Reactions were
started by adding 20 µL of 10 mM stock solutions of the substrate
in water (pH ) 10: stored in a refrigerator to minimize hydrolysis)
to 3 mL of aqueous solution in quartz cuvettes, to give a final
concentration of the substrate of 66.7 µM. The temperatures of the
reaction solutions in the cuvettes were controlled with a thermo-
statted water-jacketed cell holder, and ionic strengths were kept
constant at 1.0 M with KCl. Absorbance versus time data were
Figure 11. Progressive 31P NMR spectra for hydrolyses of BMIPP (pD 7.2, 5 min to 5.5 h) and Me-2 (pD 1.5, 5.5 h to 7.5 days), in D2O at 60 °C. Signals
are identified in Table 6.
9
8522 J. AM. CHEM. SOC. VOL. 132, NO. 24, 2010

Catalysis of Phosphodiester Cleavage by Imidazoles
A R T I C L E S
spectrometer. ESI general conditions were as follow: gas pressure
of 0.3 psi, capillary voltage of 1.55 kV, and a flow rate of 250 nl
described by the 6-31+G(d) basis set, and C and H by 6-31G(d)
and 6-31G basis sets, respectively. The geometry optimizations were
carried out in the gas phase, and water solvation free energies
(∆Gsolv) were estimated using PCM39,40 on single-point calculations,
with the molecular cavity computed, including explicit hydrogens,
using the UFF radius.41 Tables of Cartesian coordinates and
thermochemistry data are presented in sections 3.1 to 3.9 of the
Supporting Information.
All reactants and transition states were properly characterized
in force constant calculations by the absence or presence of a single
negative eigenvalue, respectively, and all thermodynamic parameters
were converted from a 1 atm standard state into the standard state
of molar concentration.42,43
min-1
.
1
NMR Spectra. H and 31P NMR spectra taken at regular time
intervals were recorded on a spectrometer (300 MHz for 1H) using
D2O as solvent. The pD values of the solutions were corrected
considering pD ) pHread + 0.4 at 25 °C. Spectra were recorded at
regular intervals at 60 °C during BMIPP hydrolysis at pD 7.2, while
the spectroscopic data for the much slower hydrolysis of the Me-2
intermediate at 60 °C and pD 1.5 were collected at 25 °C. 1H and
31P chemical shifts are referred to sodium 3-(trimethylsilyl) pro-
pionate (TSP, internal) and H3PO4 85% (external reference),
respectively. The relaxation delay was 1 s for all acquisitions.
1
Significant changes in the H NMR spectra are observed in the
For the unassisted nucleophilic attack pathway (Path B, Scheme
3) intrinsic reaction coordinate (IRC) computations44 were per-
formed, at the same level of theory described above, starting from
the optimized transition state structure (TS.NUC), using a step
length of 0.10 (a.m.u)1/2 bohr.
N-methyl region (Figure 10). At pD 7.2 the disappearance of the
BMIPP peak at 3.79 ppm is accompanied by the appearance of
two new signals at 4.03 and 3.97 ppm, which are assigned to Me-2
and Me-3, respectively. At pD 1.5 these two peaks are displaced
upfield to approximately 3.62 ppm and the hydrolysis of Me-2
affords a single peak of product Me-3. The assignments for the
aromatic region are less obvious, but the spectra at pD 7.2 clearly
show the appearance of peaks at 7.88 and 7.34 ppm for Me-2 and
Me-3, respectively. As in the aliphatic region, these peaks are also
displaced upfield at pD 1.5, where the hydrolysis of Me-2 is
observed as the disappearance of the peak at 7.56 ppm and the
appearance of the product of Me-3 at 6.99 ppm.
Acknowledgment. We thank INCT-Cata´lise and the Brazilian
foundations CNPq, FAPESC, FAPESP, and Capes for financial
assistance.
Supporting Information Available: Bunnett plot for acid
hydrolysis of BMIPP. Tables of the Bunnett plot, pH rate profile
for BMIPP and Me-2, and kinetic data in D2O and at different
temperatures. pH rate profile for the reaction of BMIPP with
nucleophiles with data table. Potentiometric titration data and
species distribution. Tables of Cartesian coordinates and ther-
mochemistry data from computational calculations. Complete
ref 34. This material is available free of charge via the Internet
at http://pubs.acs.org.
The 31P NMR spectra offer a clearer view of BMIPP hydrolysis
(Figure 11). At pD 7.2, the BMIPP peak at -10.4 ppm disappears
over time, giving the Me-2 monoester at 1.22 ppm. This peak is
displaced at pD 1.5 to -4.30 ppm, and hydrolysis of Me-2 gives
inorganic phosphate (Pi: chemical shift 0.57 ppm under the
conditions).
Computational Methods. All computational calculations were
performed using Gaussian 0334 at the DFT level employing the
hybrid B3LYP35-38 functional. The P, N, and O atoms were
JA1034733
(38) Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J.
Phys. Chem. 1994, 98, 11623–11627.
(33) Martell, A. E. S., Z. M.; Motekaitis, R. J. NIST Critical Stability
Constants of Metal Complexes Database: NIST Standard Reference
Database 46; NIST: Gaithersburg, MD, 1993.
(34) Frisch, M. J. Gaussian 98, revision D.01; Gaussian, Inc.: Pittsburgh,
PA, 2004.
(35) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(36) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. ReV. B: Condens. Matter
1988, 37, 785–789.
(37) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200–
1211.
(39) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. ReV. 2005, 105, 2999–
3093.
(40) Tomasi, J.; Persico, M. Chem. ReV. 1994, 94, 2027–2094.
(41) Rappe, A. K.; Casewit, C. J.; Colwell, K. S.; Goddard, W. A.; Skiff,
W. M. J. Am. Chem. Soc. 1992, 114, 10024–10035.
(42) Benson, S. Thermochemical Kinetics; Wiley: New York, 1968.
(43) Rastelli, A.; Bagatti, M.; Gandolfi, R. J. Am. Chem. Soc. 1995, 117,
4965–4975.
(44) Fukui, K. Acc. Chem. Res. 1981, 14, 363–368.
9
J. AM. CHEM. SOC. VOL. 132, NO. 24, 2010 8523