Hydrolysis of 8-quinolyl phosphate monoester: Kinetic and theoretical studies of the effect of lanthanide ions

Article abstract of DOI:10.1021/jo801870v

8-Quinolyl phosphate (8QP) in the presence of the trivalent lanthanide ions (Ln = La, Sm, Eu, Tb, and Er) forms a [Ln ? 8QP]⁺ complex where the lanthanide ion catalyzes hydrolysis of 8QP. In reactions with Tb ³⁺ or Er³⁺, t

Full text of DOI:10.1021/jo801870v

Hydrolysis of 8-Quinolyl Phosphate Monoester: Kinetic and
Theoretical Studies of the Effect of Lanthanide Ions
Bruno S. Souza,† Tiago A. S. Branda˜o,† Elisa S. Orth,† Ana C. Roma,‡ Ricardo L. Longo,‡
Clifford A. Bunton,§ and Faruk Nome*,†
Departamento de Qu´ımica, UniVersidade Federal de Santa Catarina, Floriano´polis,
Santa Catarina, 88040-900, Brazil, Departamento de Qu´ımica Fundamental, UniVersidade Federal de
Recife, Pernambuco, 50740-540, Brazil, and Department of Chemistry and Biochemistry, UniVersity of
California, Santa Barbara, California, 93106-9510
faruk@qmc.ufsc.br
ReceiVed August 21, 2008
8-Quinolyl phosphate (8QP) in the presence of the trivalent lanthanide ions (Ln ) La, Sm, Eu, Tb, and
Er) forms a [Ln · 8QP]+ complex where the lanthanide ion catalyzes hydrolysis of 8QP. In reactions with
Tb3+ or Er3+, there is evidence of limited intervention by a second lanthanide ion. Rate constants are
increased by more than 107-fold, and kinetic data and B3LYP/ECP calculations indicate that the effects
are largely driven by leaving group and metaphosphate ion stabilization. The lanthanides favor a single-
step DNAN mechanism with a dissociative transition state, with limited nucleophilic assistance, consistent
with the low hydroxide ion dependence and the small kinetic effect of Ln3+ radii.
SCHEME 1
Introduction
Phosphoryl transfer is very important in the control of
biological activity,1,2 notably with very slow phosphate mo-
noester spontaneous hydrolysis, that occur in seconds when
catalyzed by kinases or phosphatases.3 This remarkable ac-
celeration of reactions with poor leaving groups, e.g., alcohol
and sugar derivatives, involves precisely located metal ions.
Mechanistic differences are striking as compared with reactions
in aqueous solution, where in many instances it has been
suggested that unlike strictly “dissociative” reactions, enzyme
catalysis involves metaphosphate ion-like transition states, with
large amounts of bond breaking, and metal ions and cationic
side chains should favor the increased negative charge in the
transition state.3-5
Reactions of dianionic phosphate esters, whose leaving groups
contain strongly electron-withdrawing substituents, are often
written as involving “dissociation”, to the very labile meta-
phosphate ion, PO3-, which should be rapidly trapped by a
nucleophilic solvent, e.g., H2O, present in the substrate’s solvent
shell. In some reactions, an intermediate can be trapped without
affecting the reaction rate, but absent such evidence, the
* Corresponding Author. Tel: +55-48-37216849. Fax: +55-48-37216850.
† Universidade Federal de Santa Catarina.
‡ Universidade Federal de Pernambuco.
§ University of California.
(1) Barford, D.; Das, A. K.; Egloff, M. P. Annu. ReV. Biophys. Biolmol. Struct.
1998, 27, 133–164.
(4) Lahiri, S. D.; Zhang, G.; Dunaway-Mariano, D.; Allen, K. N. Science
2003, 299, 2067–2071.
(5) Cleland, W. W.; Hengge, A. C. Chem. ReV. 2006, 106, 3252–3278.
(2) Jackson, M. D.; Denu, J. M. Chem. ReV. 2001, 101, 2313–2340.
(3) Williams, N. H. Biochim. Biophys. Acta 2004, 1697, 279–287.
1042 J. Org. Chem. 2009, 74, 1042–1053
10.1021/jo801870v CCC: $40.75 2009 American Chemical Society
Published on Web 12/24/2008

Lanthanide-Catalyzed Hydrolysis of a Phosphate Monoester
distinction between “associative” and “dissociative” depends on
indirect evidence, for example, by drawing a distinction between
substrate solvation and partial formation of a new covalency,
relaying on theoretical modeling to make this distinction, or by
estimating life times of putative intermediates.
We use the term dissociative to indicate that there is
significant P-O1 bond breaking in the transition state. Insofar
as water is omnipresent it is difficult to draw a clear distinction
between solvation and formation of a new covalent bond. There
are many examples of attempts to make this distinction by using
ab initio/DFT calculations, and a variety of mechanistic routes
are available for the different phosphate derivatives. Triesters
such as O,O-diethyl p-nitrophenyl phosphate (paraoxon) and
related compounds react with hydroxide ion and other nucleo-
philes via rate-limiting pentacoordinate intermediate formation.6-8
Similarly, the reaction of dineopentyl phosphate anion diester
with water proceeds through an associative mechanism which
results in a phosphorane intermediate.9 Simple models of
phosphate ester monoanions show that both associative and
dissociative mechanisms are represented on the potential energy
surface, and the associative mechanism seems to be preferred
for metallo-enzymes, while enzymes that function without a
metal may react via a dissociative path.10 Phosphate monoester
dianions show a flat potential energy surface for the associative
and dissociative mechanisms, and the transition state changes
from associative to dissociative upon a decrease in the pKa of
the leaving group.11,12
FIGURE 1. Absorbances at 257 nm (A257nm) as a function of time in
the hydrolysis of 8QP (33.3 µM) in the presence of 1.28 mM of La3+
(0) or 0.98 mM of Eu3+ (O). Reactions at pH 7.00, 0.01 M BTP at
25.0 °C. Absorbances were set to zero at time ) 0.
ases.24-26 They are potentially useful in detoxification of
pesticides and chemical warfare nerve agents.15,27,28
Catalysis by Ln3+ of phosphate monoester reactions are
models for phosphoryl transfer and molecular recognition in
biological systems, and in this respect, 8-quinolyl phosphate,
8QP, is a useful substrate in investigating metal ion effects on
dephosphorylation of phosphomonoesters. For example, there
is a 106-fold rate enhancement in the Cu2+-catalyzed hydrolysis
of 8QP due to a favorable interaction with oxygen in the leaving
group, and in some of these reactions there is high nucleophi-
licity of metallo-bound H2O.29,30 Here, we show effects of
trivalent lanthanide ions, Ln3+ ) La, Sm, Tb, Eu, and Er, on
the hydrolysis of 8QP (Scheme 1) and we compare the
conclusions with those from B3LYP/ECP calculations.
Models sharing enzymic characteristics have shown how
catalysis may involve multiple interactions, as in the metallo-
phosphatase associative mechanisms.13-15 Recently, lanthanide
ions have shown very strong catalytic efficiencies due to their
higher charge densities and coordination numbers than other
metal ions, including some with biological relevance,15,16 and
lanthanide ion catalysis is important in synthetic organic
chemistry, especially for slow or regioselective reactions.17,18
Results and Discussion
Lanthanide ion complexes are medicinally important, e.g.,
as magnetic resonance imaging contrast agents,19,20 luminescent
probes,21-23 and artificial phosphatases, nucleases, and ribonucle-
Kinetic and Product Studies. Addition of an excess of any
of the trivalent lanthanide ions (Ln ) La, Sm, Eu, Tb, and Er)
results in fast complexation with 8QP, and the complexed
organic phosphate subsequently hydrolyzes. The complexation
step for the lanthanide ions was much faster than subsequent
hydrolysis, and typical first-order plots were observed for
absorbances as a function of time (Figure 1). Reactions of 8QP
in the presence of the trivalent lanthanide ions were followed
by measuring the absorbance at 257 nm and reactions are first
order, except for La3+, where with [La3+]/[8QP] < 40 the initial
complex formation affects the initial part of the kinetics, but
they can be conveniently treated as consecutive first-order
reactions (Figure 1). The reaction gives 8-quinolinol (8QOH),
or its anion, and inorganic phosphate, identified by absorption
and NMR spectroscopy, with comparisons with the spectra of
standard samples.31
(6) Zheng, F.; Zhang, C. G.; Ornstein, R. L. J. Chem. Soc., Perkin Trans. 2
2001, 12, 2355–2363.
(7) Bhattacharya, S.; Vemula, P. K. J. Org. Chem. 2005, 70, 9677–9685.
(8) Kumar, V. P.; Ganguly, B.; Bhattacharya, S. J. Org. Chem. 2004, 69,
8634–8642.
(9) Kamerlin, S. C. L.; Williams, N. H.; Warshel, A. J. Org. Chem. 2008,
73, 6960–6969.
(10) Kamerlin, S. C. L.; Wilkie, J. Org. Biomol. Chem. 2007, 5, 2098–2108.
(11) Kla¨hn, M.; Rosta, E.; Warshel, A. J. Am. Chem. Soc. 2006, 128, 15310–
15323.
(12) Iche-Tarrat, N.; Ruiz-Lopez, M.; Barthelat, J.-C.; Vigroux, A.
Chem.sEur. J. 2007, 13, 3617–3629.
(13) Holmes, R. R. Acc. Chem. Res. 2004, 37, 746–753.
(14) Mitic´, N.; Smitgh, S. J.; Neves, A.; Guddat, L. W.; Gahan, L. R.; Schenk,
G. Chem. ReV. 2006, 106, 3338–3363.
(15) Schneider, H.-J.; Yatsimirsky, A. K. The Lanthanides and Their
Interrelations with Biosystems; New York, 2003; Vol. 40, pp 369-462.
(16) Williams, N. H.; Takasaki, B.; Wall, M.; Chin, J. Acc. Chem. Res. 1999,
32, 485–493.
(17) Marks, T. J.; Hong, S. Acc. Chem. Res. 2004, 37, 673–686.
(18) Tosaki, S.; Tsuji, R.; Ohshima, T.; Shibasaki, M. J. Am. Chem. Soc.
2005, 127, 2147–2155.
(24) Zhu, B.; Xue, D.; Wang, K. Biometals 2004, 17, 423–433.
(25) Liu, C.; Wang, M.; Zhang, T.; Sun, H. Coord. Chem. ReV. 2004, 248,
147–168.
(26) Chang, C. A.; Wu, B. H.; Duan, B. Y. Inorg. Chem. 2005, 44, 6646–
6654.
(19) Zhou, J.; van Zijl, P. C. M. Prog. Nucl. Magn. Reson. Spectrosc. 2006,
48, 109–136.
(20) Aime, S.; Crich, S. G.; Gianolio, E.; Giovenzana, G. B.; Tei, L.; Terreno,
E. Coord. Chem. ReV. 2006, 250, 1562–1579.
(21) Pandya, S.; Yu, J.; Parker, D. J. Chem. Soc., Dalton Trans. 2006, 2757–
2766.
(27) Yang, Y.-C. Acc. Chem. Res. 1999, 32, 109–115.
(28) Yang, Y.-C.; Baker, J. A. Chem. ReV. 1992, 92, 1729–1743.
(29) Murakami, Y.; Sunamoto, J. Bull. Chem. Soc. Jpn. 1971, 44, 1827–
1834.
(30) Hay, R. W.; Basak, A. K.; Pujari, M. P.; A., P. J. Coord. Chem. 1991,
23, 43–55.
(22) Hemmila¨, I.; Laitala, V. J. J. Fluoresc. 2005, 15, 529–542.
(23) Song, J.-L.; Lei, C.; Mao, J.-G. Inorg. Chem. 2004, 43, 5630–5634.
(31) Sapelli, E.; Branda˜o, T. A. S.; Fiedler, H. D.; Nome, F. J. Colloid
Interface Sci. 2007, 314, 212–222.
J. Org. Chem. Vol. 74, No. 3, 2009 1043

Souza et al.
FIGURE 2. (a) UV-vis spectra of 8QO- in the presence of 0.32 to 3.2 mM [LaCl3], at pH 7.0 and 25 °C. (b) Absorbances of [La8QO]2+ complex
at 257 nm (A257nm) as a function of [LaCl3], at pH 7.00, 0.01 M BTP and 25 °C. The solid line is a fit by using eq 1.
TABLE 1. K◦assoc and Absorbance Values for the [Ln ·8QO]2+
complex at 257 nm, pH 7.00, 0.01 M BTP and 25.0 °C
SCHEME 2
◦
a
Ln3+
log (Kassoc, M-1
)
ALn8QO
La
Sm
Eu
Tb
Er
6.24 ( 0.01
6.55 ( 0.01
6.93 ( 0.02
6.92 ( 0.02
7.25 ( 0.04
0.879 ( 0.009
1.019 ( 0.006
0.963 ( 0.017
0.867 ( 0.008
0.791 ( 0.014
The reaction products are complexes of Ln3+ and 8QO- as
shown by their spectra, which were identical to those obtained
from 8QOH titration with Ln3+. Figure 2a shows the UV-vis
spectra taken after at least 8 half-lives of the hydrolysis of 8QP
in the presence of different La3+ concentrations, and there is a
pronounced spectral change as a function of [La3+], consistent
with a 1:1 equilibrium complexation of 8QOH with La3+ (Figure
2b). Complex formation decreases the pKa of the hydroxyl group
in 8QOH and allows formation of a stronger chromophore
([La8QO]2+) and increasing absorbance. The spectral changes
reflect the equilibrium depicted in Scheme 2 with derivation of
eq 1 and describe observed absorbance changes.
a K◦assoc values were corrected according to the free Ln3+ fraction.33
A
Ln8QOK◦assocKa[Ln3+]0
K◦assocKa[Ln3+]0 + Ka + [H+]
A257nm
)
(1)
For the other lanthanides, the observed behavior was similar
to that of La3+. The thermodynamic association constants
between 8QOH and Ln3+ (Ka◦ssoc) were obtained by nonlinear
fitting of the absorbance at 257 nm (A257nm) versus [Ln3+] data
by using eq 1, where ALn8QO is the maximum absorbance of the
FIGURE 3. Absorbance at 250 nm of 8QP in the presence of (9) LaCl3,
(O) SmCl3, (2) EuCl3, (0) TbCl3, and ([) ErCl3, at pH 7.0 and 25 °C.
Solid lines calculated using eq 2.
complex and Ka is the 8QOH dissociation constant (2.40 ×
32
10-7
)
as shown in Table 1.
results32 and indicative of strong interactions of Ln3+ with the
aryl oxide ion.
The observed spectra indicate that ligand-Ln3+ bonds have
significant ionic character rather than the usual charge-transfer
coordination of these ligands with transition-metal ions, and as
a consequence, the spectra in the presence of Ln3+ are similar
to that of 8-quinolinate ion (8QO-).31 Stability constants for
complexes of 8QO- with a variety of lanthanides and metal
ions have been estimated and are fully consistent with our
In order to determine the association constant between 8QP
and the lanthanides ions, the UV-vis spectrum of the mixture
was registered immediately after fast mixing in order to
minimize spectral changes due to hydrolysis. Increasing the
metal ion concentration (for all the lanthanides) decreased the
absorbance at 250 nm (Figure 3).
The observed absorbance changes (Figure 3) are consistent
with a 1:1 complex between 8QP and the Ln3+ (Scheme 3),
which allows the derivation of eq 2. Increasing the metal ion
concentration decreases the absorbance at 250 nm, allowing
(32) Martell, A. E.; Smith, Z. M.; Motekaitis, R. J. NIST Critical Stability
Constants of Metal Complexes Database: NIST Standard Reference Database
46; NIST: Gaithersburg, 1993.
(33) Go´mez-Tagle, P.; Yatsimirsky, A. K. Inorg. Chem. 2001, 40, 3786–
3796.
1044 J. Org. Chem. Vol. 74, No. 3, 2009

Lanthanide-Catalyzed Hydrolysis of a Phosphate Monoester
SCHEME 3
sis of 8QP in their presence, and insolubility of the 8QP-
lanthanide complexes prevents titration at low temperature (10
°C).
ATR-FTIR Studies. As a consequence of the ionic nature
of ligand-Ln3+ bonds, UV-vis spectra of 8QP-Ln3+ complexes
at pH 7 were indistinguishable from those without Ln3+
.
Therefore, attenuated total reflectance infrared (ATR-FTIR)
spectroscopy is a useful method for product identification but
because of the form of the equipment not for kinetic work. It
TABLE 2. KQP, K◦ , ALn8QP, and A8QP Values for the [Ln ·8QP]+
QP
Complex at 250 nm, pH 7.00, 0.01 M BTP and 25.0 °C
was employed to give evidence on Ln3+ complexation with 8QP.
log
log
◦
a
a
-
Ln3+ (KQP, M-1
)
(KQP, M-1
)
A8QP
ALn8QP
Figure 5a shows the spectra for aqueous 8QP and H2PO4
/
HPO4 (Pi) in the absence of La3+, and Figure 5b shows the
2-
La
Sm
Eu
Tb
Er
3.96
4.14
4.25
4.30
4.36
4.07
4.25
4.36
4.41
4.47
0.313 ( 0.004 0.070 ( 0.003
0.313 ( 0.002 0.030 (0.003
0.313 ( 0.009 0.018 ( 0.011
0.313 ( 0.002 0.016 ( 0.002
0.313 ( 0.003 0.020 ( 0.003
spectra at the beginning and end of hydrolysis in the presence
of La3+
.
Without La3+, the Pi vibrational spectrum shown in Figure
2-
5a is formally that of H2PO4-/HPO4 at pH 7.0, due to the
a KQP values were corrected according to the free Ln3+ fraction and
are accurate to ( 0.05 units.33
asymmetric stretching in the PO3 and PO3H moieties, as in the
literature.35 Comparison of the 8QP spectrum at pH 9.0
(phosphate moiety as a dianion) with the vibrational frequencies
for a series of aromatic phosphate monoester dianions,36 allows
assignment of the bands at 979 and 1107 cm-1 to symmetric
and asymmetric stretchings in the PO3 moiety, respectively.
Similarly, the peaks at 1142 and 1202 cm-1 correspond to C-O
and P-O-C stretchings.37
determination of the apparent association constants (KQP) by
the nonlinear fit of the absorbance at 250 nm (A250nm) versus
[Ln3+] data with eq 2, and fitted curves are shown in Figure 3.
The thermodynamic association constant for the complex
formation (K◦QP) was calculated by considering the fraction of
the phosphate monoester that is deprotonated at the nitrogen
atom (eq 3).
Comparatively, the symmetric and asymmetric stretching
vibrations of the PO3 moiety in the La ·8QP spectrum are
systematically shifted toward lower frequencies from 979 and
1107 cm-1 to 914 and 1001 cm-1, respectively, while the C-O
band is displaced from 1142 to 1108 cm-1. In addition, there is
a broad band at 1151 cm-1 that can also be assigned to the
asymmetric stretching vibrations in the PO3 moiety, probably
due to symmetry lowering caused by coordination of the
phosphate group with Ln3+. The spectrum at the end of the
hydrolysis reaction is basically identical with the La ·Pi vibra-
tional spectrum at pH 7.0, but the asymmetric vibrations of the
PO3H moiety are missing, due to displacement of hydrogen by
La3+, and there is a broadband with a maximum at 1054 cm-1
due to asymmetric stretching of the PO3 moiety. The band at
1104 cm-1 due to the 8-quinolinolate C-O stretching is similar
to the C-O band of 8QP in the complex with La3+ (Figure
5b). The band positions of 8QP and of the products in the
presence of the other Ln3+ are similar (not shown), with similar
vibrational frequencies through the series and are essentially
identical to those shown in Figure 5b, because interactions are
similar.
The ATR-FTIR spectra were taken as a function of time with
approximately equimolar concentrations of 8QP and lanthanide
ion. The results show disappearance of the band at 1150 cm-1
(asymmetric stretching in the PO3 moiety of the [Ln ·8QP]+
complex) and appearance of a band at 1050 cm-1, assigned as
the asymmetric stretching vibrations of the PO3 moiety in the
[Ln·PO4 ·8QO] complexes (Figure 6). Examples of these
spectral changes for La3+, Tb3+, and Er3+ are given in Figure
S2 of the Supporting Information.
1
A250nm ) (ALn8QPKQP[Ln3+]0 + A8QP
)
3+
(
)
1 + KQP[Ln ]0
(2)
(3)
K◦QP ) KQP ⁄ ꢀQP
In eqs 2 and 3, ALn8QP and A8QP are absorbances of the
complex and that of free 8QP, respectively; [Ln3+]0 is the total
metal ion concentration, and ꢀQP is the mole fraction of 8QP
deprotonated at the nitrogen atom. The equilibrium constants
KQP and K◦QP correspond to the apparent and thermodynamic
association constants, respectively, for the equilibrium depicted
in Scheme 3. The parameters used in the calculation and the
calculated equilibrium constants are in Table 2.
Titration Studies. Potentiometric titrations in the pH range
3-11 gave pKa values for 8QP of 4.30 ( 0.01 and 6.62 ( 0.02.
Spectrophotometric titration was used to assign these pKa values
to hydroxyl or quinolinium groups. Spectral scans from pH 2.0
to 7.8 at 242 and 247 nm (isosbestic point) were used to
calculate pKa values of 4.40 ( 0.03 and 6.48 ( 0.03 (Figure
4a). We observed a smaller absorbance change for the first
relative to the second pKa, indicating that it corresponds to
deprotonation of the quinolinium moiety as confirmed by 31P
1
and H NMR titrations. As shown in Figure 4b, deprotonation
of the phosphate hydroxyl group induced large upfield shifts
on the 31P signal, while deprotonation at nitrogen did not. The
1
dependence of H chemical shifts on pH shows upfield shifts
on H-2 and H-4 signals with phosphate group deprotonation,
consistent with interaction of the phosphate dianion with the
acidic hydrogen quinolinium atom, in the so-called Proton
Sponge effect.34 This effect shifts the pKa of the NH+ group
from 4.94 in the 8-hydroxyquinolinium ion32 to 6.48 in 8QP.
Potentiometric titrations at 25 °C of 8QP/Ln3+/BTP mixtures
were not possible with lanthanide ions, because of fast hydroly-
Figure 7 presents the kinetic results for the lanthanide
catalyzed hydrolysis of 8QP with respect to [Ln3+] at different
pHs in all cases with a significant acceleration. All kinetic
(35) Tejedor-Tejedor, M. I.; Anderson, M. A. Langmuir 1990, 6, 602–611.
(36) Cheng, H.; Nikolic-Hughes, I.; Wang, J. H.; Deng, H.; O’Brien, P. J.;
Wu, L.; Zhang, Z.-Y.; Herschlag, D.; Callender, R. J. Am. Chem. Soc. 2002,
124, 11295–11306.
(37) Silverstein, R. M.; Webster, F. X. Spectrometric Identification of Organic
Compounds; John Wiley & Sons: New York, 2005; p 126.
(34) Staab, H. A.; Saupe, T. Angew. Chem., Int. Ed. 1988, 27, 865–879.
J. Org. Chem. Vol. 74, No. 3, 2009 1045

Souza et al.
FIGURE 4. (a) Spectrophotometric scans of aqueous 33.3 µM 8QP at pH 2.0 (a), 4.9 (b), and 7.8 (c). The inset corresponds to absorbances at 242
1
(9) and 247 nm (0) as functions of pH, at 25.0 °C. (b) 31P (2) and H NMR chemical shifts for H-2 (b) and H-4 (O) as a function of pD, 0.04
M 8QP at 25 °C.
FIGURE 6. Kinetic profile for the hydrolysis of 8QP in the presence
of Ln3+; aqueous 1.0 mM 8QP (initial concentration) and 1.0 mM Eu3+
and 2.0 mM Sm3+ 0.01 M of BTP at 25 °C. The absorbances were
normalized in relation to the highest values.
SCHEME 4
In the lanthanide-mediated hydrolysis of bis(2,4-dinitrophe-
nyl)phosphate (BDNPP), increasing [BTP] from 0.01 to 0.1 M
inhibits the reaction, probably because coordination of lanthanide
ions with BDNPP is weaker than that with BTP (Kassoc ) 200
M-1).39 In the present reaction, an increase in [BTP] in the same
FIGURE 5. ATR-FTIR spectra (in transmittance) of aqueous 1.0 mM
8QP and Na2HPO4 (Pi) in the absence of La3 (a) and at beginning (solid)
range, at pH 8.0, and with 1.0 mM La3+ did not affect the
and end (dotted) of hydrolysis of 1.0 mM 8QP in the presence of 2.0
hydrolysis of 8QP. This behavior is general40 and is fully
mM La3+ at pH 7.0 (b). The spectra in the presence of La3+ were at
consistent with the fact that the association constant of the
pH 7.0 and in its absence at pH 9.0 (8QP) and 7.0 (Pi).
lanthanides with this organic substrate (KQP) is at least 2 orders
experiments were performed in the presence of BTP which,
of magnitude larger relative to the BTP-La3+ complex (Table
besides acting as a buffer, complexes with Ln3+ and, therefore,
2). Kinetics of the lanthanide catalyzed hydrolyses show two
prevents aggregation and precipitation of the free lanthanide
ions.38
(39) Longhinotti, E.; Domingos, J. B.; Silva, P. L. F.; Spoganicz, B.; Nome,
F. J. Phys. Org. Chem. 2005, 18, 167–172.
(40) Bruice, T. C.; Tsubouchi, A.; Dempcy, R. O.; Olson, L. P. J. Am. Chem.
Soc. 1996, 118, 9867–9875.
(38) Longhinotti, E.; Domingos, J. B.; Szpoganicz, B.; Neves, A.; Nome, F.
Inorg. Chem. 2005, 358, 2089–2092.
1046 J. Org. Chem. Vol. 74, No. 3, 2009

Lanthanide-Catalyzed Hydrolysis of a Phosphate Monoester
FIGURE 7. (a-e) Variation of kobs for the hydrolysis of 8QP with respect to concentrations of various lanthanides as a function of pH, 0.01 M
BTP and 25.0 °C. In (f) are the data for all lanthanides at pH 7.00. The solid lines are fits modeled according to eq 4, with equilibrium and kinetic
parameters in Table 3.
J. Org. Chem. Vol. 74, No. 3, 2009 1047

Souza et al.
TABLE 3. Association and Rate Constants for Hydrolysis of 8QP in the Presence of Ln3+ at Different pH, with 0.01 M BTP at 25.0 °Ca
b
pH
log(K , M-1
)
(ꢀLn
)
104 × k1, s-1
k2, M-1s-1 (%)c
kin
QP
La3+
7.00
7.25
7.50
8.00
8.50
7.00
7.50
7.00
7.00
7.50
8.00
7.00
7.50
3.79 ( 0.17
0.986
0.970
0.945
0.832
0.582
0.964
0.362
0.873
0.924
0.328
0.055
0.678
0.011
5.50 ( 0.04
6.57 ( 0.06
9.46 ( 0.06
13.4 ( 0.3
23.8 ( 0.2
16.5 ( 0.3
50.9 ( 0.7
16.5 ( 0.8
9.1 ( 0.3
10.9 ( 0.1
11.5 ( 0.2
6.80 ( 0.40
9.50 ( 1.5
Sm3+
3.95 ( 0.19
Eu3+
Tb3+
4.36 ( 0.05
4.42 ( 0.13
0.26 ( 0.04 (3.0 - 23.5)
0.25 ( 0.01 (2.3 - 18.9)
0.59 ( 0.06 (4.9 - 34.0)
0.55 ( 0.05 (4.0 - 35.0)
0.30 ( 0.02 (1.5 - 13.2)
Er3+
4.43 ( 0.05
a
Values of log K were corrected considering ꢀLn.
b Mole fractions of free Ln3+ (ꢀLn), obtained from potentiometric titrations of the Ln3+/BTP
kin
QP
systems.33,39 c Contribution of k2 to the overall reaction in the range of 0.1-1 mM [Ln3+].
behaviors, both independent of the presence of BTP: (i) with
La3+, Sm3+ and Eu3+ the reaction follows a classic saturation
profile of stoichiometry 1:1 (Ln/8QP); whereas (ii) with Er3+
and Tb3+, saturation is followed by a modest linear increase in
kobs (Figure 7). The following model, depicted in Scheme 4, is
consistent with these observations, which can be expressed as
kobs ) (k1 + k2[Ln3+]0)ꢀc
(4)
where ꢀc is the mole fraction of the complex given by
KkinKa[Ln3+]0
KkinKa[Ln3+]0 + Ka + [H+]
QP
ꢀc )
(5)
QP
where Ka is the acid dissociation constant of the quinolinium
kin
group in 8QP, K is the apparent association constant for
QP
complexation of 8QP with Ln3+, and [Ln3+]0 represents the total
lanthanide ion concentration. The kinetic parameters obtained
by least-squares fits are given in Table 3, and in all cases, the
FIGURE 8. Logarithm of thermodynamic association constants for
[Ln·8QP]+ at pH 7.00 as a function of Ln3+ radii, 0.01 M BTP at 25.0
°C. Ionic radii for coordination number eight are from Shannon.41
kin
K
QP
values agree with those obtained from the complexation
studies (Table 2).
kin
The K values for the formation of the [Ln·8QP]+ complex
QP
thanide complex seems to be the dominant reactive form, and
responsible for about 95-99% of the reaction, we focus our
studies on it.
(Table 3) are similar to those determined spectroscopically
(Table 2) and are at least 2 orders of magnitude greater than
those for the interaction of lanthanides and BTP (vide supra), a
fact fully consistent with the lack of an effect of BTP in this
particular reaction. In the hydrolysis reactions in the presence
of La3+, Sm3+, and Eu3+, the kinetic data can be adequately
explained with only the k1 term. Conversely, in the presence of
Tb3+ and Er3+ there is evidence of a small contribution of the
second order term k2. The percentages of reaction going through
the k2 pathway are given in Table 3, and the contribution of the
k2 term to the overall reaction is significantly smaller than that
of k1. With 0.1 mM Tb3+ (or Er3+), the reaction proceeds with
a contribution of a second lanthanide ion, which is less than
5% (Table 3). It is important to note that, although in Scheme
4 Ln3+ is shown as promoting reaction of the first complex, the
reaction may actually correspond to the complexation of 8QP
to a dinuclear lanthanide complex, such as those participating
in the hydrolysis of BDNPP or BNPP.33,39 The existence of
binuclear polyhydroxo species has been postulated, but there is
the difficulty in identifying them as reactive species, and
potentiometric titrations by Go´mez-Tagle and Yatsimirsky33
identified [Eu2(BTP)2(OH)2]4+ as the main species for Eu3+ at
pH 7.8 and Longhinotti et al.39 concluded that for Sm3+, the
main species is [Sm2(BTP)2(OH)5]+. However, because under
relatively low concentrations of Ln3+ (0.1 mM) the monolan-
While the k1 values for reaction with the largest lanthanides
(La3+ and Sm3+) increase monotonically over the pH range,
these increases are less significant for the smaller lanthanides,
probably because La3+ and Sm3+ can expand its coordination
number with an additional hydroxo ligand as the pH increases,
whereas the smaller ions would not have this ability. Figure 8
shows the dependence of the thermodynamic association
constant for the complex formation (K◦assoc) with respect to the
lanthanide radii at pH 7.0. Effects on K◦assoc are small, so that
log K◦assoc increases modestly from La3+ to Er3+
.
This overall effect is quite general over a number of ligands
but it is subject to uncertainties. Basically, the increase of
K◦assoc with decreasing Ln3+ diameter has been ascribed to the
higher charge density of the smaller lanthanide ions.33,42,43
Regarding the steric requirements for complexation of the
lanthanides and 8QP, it is known that all lanthanides can form
bonds as long as 2.5 Å, which would uniformly allow com-
plexation with the quinolinic nitrogen and the phosphate oxygen
(41) Shannon, R. D. Acta Crystallogr. 1976, A32, 751–767.
(42) Moss, R. A.; Jiang, W. Langmuir 2000, 16, 49–51.
(43) Roigk, A.; Hettich, R.; Schneider, H.-J. Inorg. Chem. 1998, 37, 751–
756.
1048 J. Org. Chem. Vol. 74, No. 3, 2009

Lanthanide-Catalyzed Hydrolysis of a Phosphate Monoester
atoms (vide infra). A probable model for the small effect of the
diameter of the lanthanide on the K◦assoc values involves
solvation and coordination numbers of the Ln3+. Insofar as
hydration energies of the lanthanides decrease for the higher
coordination numbers, which are ca. 9 for La3+ to Eu3+ and 8
for those with smaller radii,44,45 the increasing stability with a
larger charge to radius (Z/r) ratio is accompanied by a hydration
barrier hindering coordination with 8QP.
The influence of Ln3+ radii on k1 values is small in the pH
range studied. Below pH 7.0, the reaction is much slower and
the kinetic behavior is complex, most probably because the
complexation becomes rate determining and thermodynamically
unfavorable (Supporting Information). The observed increase
in rate constants follow complexation in solution, and the
observed increase in k1 may be related to formation of
monohydroxo [La ·8QP(OH)]2+ complexes, 1, as for the hydroxo
complexes in the Ln3+/BTP systems.33,39
FIGURE 9. Dissociative reaction coordinates for reaction of 8QP in
the absence of Ln3+. Structures were optimized at the B3LYP level
with basis sets O, N, and P (6-31+G*); C (6-31G*) and H (6-31G).
whose numbering schemes are as in 2 and 3, i.e., Gibbs free
energies of reaction and activation are compared for
[8QP·(H2O)3]2- and [Ln·8QP·(H2O)6]+. Figures 9 and 10
represent optimized reaction coordinates for the dissociative
mechanism of 8QP in the absence and presence of lanthanide
ions, respectively. Tables 4 and 5 include some selected
geometric parameters for reactants and the transition states
without and with Ln3+, respectively.
The extent of catalysis promoted by the lanthanide ions
formally involves the reactivity of 8QP2-, which is very
unreactive without the lanthanides. Although the rate constant
for this spontaneous reaction is unknown, it can be predicted,
from hydrolyses of 8QP at higher temperatures, to be of the
order of 5 × 10-10 s-1 at pH 7.0 and 25 °C.46 Thus, the
lanthanide ion causes a greater than 107-fold rate enhancement
at pH 7.0, which may increase modestly with other Ln3+ due
to contribution of the k2 reaction.
A probable factor in the lanthanide acceleration of the
hydrolysis of 8QP2- is the decrease of the leaving group pKa
(pKlg), which, by assuming k1 values for La3+, and comparing
them to those of phosphate monoesters assigns a leaving group
pKa of about 5.5. Provided that the mechanism promoted by
lanthanide ions resembles those in their absence, the catalyzed
hydrolysis should have a typical DNAN dissociative character,47
with a very loose transition state, and limited nucleophilic
assistance, as shown by the low hydroxide ion dependence and
the small effect of Ln3+ radii on rate constants, k1. This
mechanism differs from that proposed for hydrolysis of phos-
phate monoesters catalyzed by transition metals where associa-
tion with an aquo complex is important and there is significant
water attack on phosphorus. There is probably little charge
transfer between the Ln3+ and 8QP, although it is important in
transition-metal complexes. Quantum mechanical calculations
were performed on the catalytic mechanism for hydrolysis of
8QP to test these conclusions.
The reaction coordinate for the reaction of 8QP2- was
modeled by the [8QP ·(H2O)3]2- structure because gas-phase
8QP2- has a bond length P-O1 of 2.015 Å, which is ∼0.3 Å
longer than for phosphate monoester dianions in the solid state
(∼1.7 Å).48 This difference is general for calculated structures
of gas-phase monoester phosphate dianions, which relates their
dissociation to the stability of the gas-phase metaphosphate
anion.49 Thus, inclusion of solvent effects is essential, by a
polarizable continuum method (PCM-type approach), or by
adding a few water molecules (explicit solvation) for hydration
of polar and ionic groups,50,51 i.e., the phosphoryl oxygen atoms.
Both approaches are limited, because PCM neglects specific
interactions of water molecules with the phosphoryl group52 and
explicit solvation should consider fluctuations with averaging
over large numbers of water molecules and of accessible
configurations,53 which is very demanding and involves the
interaction potential between solute and solvent.
Computational Studies. Probable reaction paths for the 8QP
hydrolysis in the absence and presence of Ln3+ were modeled
with the [8QP ·(H2O)3]2- and [Ln ·8QP·(H2O)6]+ structures,
(48) Jones, P. G.; Kirby, A. J. J. Am. Chem. Soc. 1984, 106, 6207–6212.
(49) Henchman, M.; Viggiano, A. A.; Paulson, J. F.; Freedam, A.; Worm-
houdt, J. J. Am. Chem. Soc. 1985, 107, 1453–1455.
(50) Wang, Y.-N.; Topol, I. A.; Collins, J. R.; Burt, S. K. J. Am. Chem. Soc.
2003, 125, 13265–13273.
(51) Kirby, A. J.; Dubba-Roy, N.; Silva, D.; Goodman, J. M.; Lima, M. F.;
Roussev, C. D.; Nome, F. J. Am. Chem. Soc. 2005, 127, 7033–7040.
(52) Cramer, C. J.; Truhlar, D. G. Chem. ReV. 1999, 99, 2161–2200.
(53) Åqvist, J.; Kolmodin, K.; Florian, J.; Warshel, A. Chem. Biol. 1999, 6,
R71–R80.
(44) Choppin, G. R.; Bu¨nzli, J.-C. G. Lanthanide probes in life, chemical
and earth sciences - Theory and practice; Elsevier: Amsterdan, 1989; pp
1-41.
(45) Kodama, M.; Koike, T.; Mahatma, A. B.; Kimura, E. Inorg. Chem. 1991,
30, 1270–1273.
(46) Murakami, Y.; Sunamoto, J.; Sadamori, H. J. Chem. Soc., Chem.
Commun. 1969, 983–984.
(47) Kirby, A. J.; Varvoglis, A. G. J. Am. Chem. Soc. 1967, 89, 415–423.
J. Org. Chem. Vol. 74, No. 3, 2009 1049

Souza et al.
FIGURE 10. Dissociative reaction coordinates for reaction of 8QP in the presence of Ln3+. Structures of the lanthanide complexes are similar and
only those with La3+ are shown. Structures were optimized at the B3LYP level by using the following basis sets: La3+ (ECP46MWB); Sm3+
(ECP51MWB); Tb3+ (ECP54MWB); Er3+ (ECP57MWB); O, N, and P (6-31+G*); C (6-31G*), and H (6-31G).
TABLE 4. Selected Distances (Å) and Angles (deg) for Optimized
Structures of 8QP2-, [8QP(H2O)3]2-, and TS1a
The equilibria in Scheme 3 and Figure 10 are markedly
displaced toward formation of the [Ln ·8QP(H2O)5]+ complexes,
∆G° ≈ -2000 kJ mol-1, due to the strong bond with Ln3+, the
charge decrease, and the small Ln-OH2 bond energies. We note
that in this model, a few water molecules provide the solvation,
so that major differences relative to the aqueous phase are
mainly due to the last two factors, which are reduced by the
solvent. Thus, for comparison, effects on the charge, and
partially on the Ln-OH2 bond energies, are reduced by scaling
to La3+. As a result, for Sm3+, ∆∆G° ) (∆G°Sm - ∆G°La) )
-28.4 kJ mol-1, which decreases toward Er3+, namely, ∆∆G°
) -45.5 kJ mol-1. This scaling yields a trend that to some
extent follows the experimental data (Table 6).
8QP2-
[8QP(H2O)3]2-
TS1
C-O1
1.287
2.015
1.522
1.530
1.529
118.2
118.2
116.4
1.317
1.796
1.531
1.540
1.541
116.1
116.1
114.5
1.262
3.455
1.509
1.510
1.510
119.8
119.8
119.5
P-O1
P-O2
P-O3
P-O4
O2-P-O3
O2-P-O4
O3-P-O4
a B3LYP calculations with basis sets: O, N, and P (6-31+G*), C
(6-31G*), and H (6-31G).
The reaction coordinate from ground state, [8QP(H2O)3]2-
,
We consider explicit solvation with a few water molecules,
mainly because PCM-type approaches have not been fully tested
and validated for lanthanide complexes. Also, in order to avoid
many local minima on the potential energy surface (PES), only
three water molecules were explicitly added in hydration of the
phosphoryl oxygens of 8QP2-, establishing the [8QP ·(H2O)3]2-
complex with a Cs symmetry (Figure 9). The calculated P-O1
bond length of 1.796 Å is consistent with that in solution, being
less than 0.1 Å longer than in the solid, where packing forces
and the absence of explicit hydration shorten bonds.
With lanthanide ions, reaction path calculations were per-
formed on the [Ln·8QP(H2O)6]+ complexes (Figure 10), which
were selected from optimizations of the [8QP ·(H2O)3]2- and
[Ln(H2O)8]3+ systems leading to [Ln ·8QP(H2O)5]+ + 2[H2O]3
(clusters). Initial structures of these complexes were created by
coordination at the probable sites of 8QP, namely, of the
quinolinic nitrogen and phosphoryl oxygens to Ln3+ and by
placing water molecules at ca. 2.5 Å from Ln3+ to complete
coordination numbers of 8 or 9, typical for lanthanide ions.
As shown in Table 5, the main effects of Ln3+ complexation
are on the structure of the phosphate group. Bond lengths change
slightly from La3+ to Er3+, with 0.009 Å shortening of the P-O1
bond, and there are evident changes in the bond angles, mainly
O2-P-O3, which in the complexes are about 12° smaller than
in [8QP(H2O)3]2-. There is also a gradual decrease of 1.3° in
the O2-P-O3 angles with decreasing lanthanide radii. Changes
in angles are energetically much less demanding than in bond
length.
to TS1 is indicative of a very loose transition state with O-P-O
angles near to 120° and a P-O1 bond longer by 1.665 Å and
of C-O1 shorter by 0.111 Å. The calculated activation Gibbs
energy, ∆G‡ ) 57.3 kJ mol-1, is typical of a gas phase reaction
and is lower than in solution with no desolvation, solvent
reorganization and differential solvation between ground and
transition states. For example, the hydrolysis of p-nitrophenyl
phosphate dianion has ∆G‡ ) 124 kJ mol-1 at 25 °C in water,
but in DMSO/H2O 95:5, it decreases to 93.6 kJ mol-1, mainly
due to partial desolvation of the phosphoryl group.54 A similar
argument can be made for the calculated values of ∆G‡ for
reactions in the presence of Ln3+, which are systematically
smaller than those calculated from values of k1 at pH ) 7.00.
The behavior of TS2 differs from that of TS1 relative to their
respective ground states, for example, the P-O1 bond length
is longer in TS1 by about 1.7 Å, but 0.8 Å in TS2 (0.809 Å to
La3+ and 0.774 Å to Sm3+). This observation indicates a stronger
P-O1 bond, and the putative metaphosphate ions in TS2 are
practically planar and loose, but with distorted O-P-O bond
angles, namely, ∼108° and ∼126°, which are ∼6° and ∼3°
larger than in the ground state. In structure 1, HO- coordinated
to La is shown as attacking phosphorus, as in dephosphorylations
by H2O/HO- complexed to a transition metal ion. In this respect,
we note that TS2 is consistent with a single step DNAN
mechanism with a rather loose transition state, where the planar
(54) Grzyska, P. K.; Czyryca, P. G.; Golightly, J.; Small, K.; Larsen, P.;
Hoff, R. H.; Hengge, A. C. J. Org. Chem. 2002, 67, 1214–1220.
1050 J. Org. Chem. Vol. 74, No. 3, 2009

Lanthanide-Catalyzed Hydrolysis of a Phosphate Monoester
TABLE 5. Selected Atomic Distances (Å), Angles (deg), and Dihedral Angles (deg) for Optimized Structures of [Ln ·8QP(H2O)5]+ and TS2a
[Ln·8QP(H2O)5]+
TS2
Sm
La
Sm
Tb
Er
La
Tb
C-O1
1.374
1.784
1.579
1.569
1.475
2.615
2.563
2.461
2.681
1.374
1.780
1.579
1.570
1.475
2.555
2.478
2.383
2.594
1.373
1.781
1.578
1.571
1.474
2.526
2.418
2.356
2.552
1.373
1.777
1.580
1.573
1.473
2.517
2.364
2.325
2.513
1.353
2.574
1.534
1.531
1.470
2.392
2.592
2.593
2.642
1.353
2.545
1.533
1.534
1.470
2.318
2.521
2.490
2.550
1.354
2.512
1.534
1.535
1.470
2.283
2.480
2.443
2.506
P-O1
P-O2
P-O3
P-O4
Ln-O1
Ln-O2
Ln-O3
Ln-N
O2-P-O3
O2-P-O4
O3-P-O4
C-O1-Ln-N
102.5
102.0
101.6
101.2
108.2
107.7
107.4
123.8
123.4
22.2
124.2
123.5
22.8
124.5
123.7
23.4
124.6
123.8
23.8
125.6
125.9
10.5
126.0
126.0
9.5
126.2
126.1
8.3
a B3LYP calculations with basis sets: La3+ (ECP46MWB), Sm3+ (ECP51MWB), Tb3+ (ECP54MWB), Er3+ (ECP57MWB), O, N, and P (6-31+G*),
C (6-31G*), and H (6-31G).
TABLE 6. B3LYP/ECP Gibbs Energy of Reaction and of
Activation, Enthalpy and Entropy of Activation, Experimental
Gibbs Energies, And Relative Gibbs Energies Associated with
entropy difference between transition states and reactants.
However, despite this less favorable entropic contribution,
calculated Gibbs energies indicate the strong catalytic effects
of Ln3+. There is a correlation between calculated values of
∆G‡ and the ionic radii of Ln3+, namely, ∆G‡ decreases linearly
as a function of decrease in the Ln3+ radius (Supporting
Information, Figure S4), and when this linear relationship is
extrapolated to the Er3+ radius a ∆G‡ ) 5.5 kJ mol-1 is
predicted, hence the difficulties in describing TS2 for this ion.
Although B3LYP-calculated values of ∆G‡ are much smaller
than those estimated from k1, it is clear from Table 6 that
differences are mainly due to hydration effects which are not
included in the B3LYP calculations. The semiquantitative trends
observed for the hydrolysis in absence and in the presence of
Ln3+ are quite well reproduced by the calculations. In addition,
the estimated hydration effects from the difference between the
experimental and calculated values of ∆G‡ follow a systematic
increase with the decrease of the Ln3+ radii, consistent with
the increase of charge densities of the ions. These comparisons
between the calculated and experimental values of ∆G‡ support
the qualitative results and conclusions obtained from the B3LYP
calculated properties.
Structures in Figures 9 (Absence of Ln3+) and 10 (Presence of
a
Ln3+
)
TS1
La
Sm
Tb
Er
∆∆G° (B3LYP)b
∆H‡ (B3LYP)
T∆S‡ (B3LYP)
∆G‡ (B3LYP)
∆∆G‡ (B3LYP)d
∆G‡ (exp)e
0
28.3
6.7
21.6
35.7
91.6
34.6
0
-28.4
19.9
6.3
13.6
43.7
88.9
37.3
-1.03
75.3
-39.0
14.3
5.9
-45.5
79.2
21.9
57.3
0
126.2
0
8.3
5.5c
51.8
91.1
49.0
90.3
35.9
-1.94
82.0
∆∆G‡ (exp)d
35.1
∆∆G° (exp)b,f
-2.28
g
∆
hyd∆G‡
68.9
70.0
85.6
a All quantities calculated at 298.15 K and expressed in kJ mol-1
.
B3LYP calculations with the following ECPs and basis sets: La3+
(ECP46MWB), Sm3+ (ECP51MWB), Tb3+ (ECP54MWB), Er3+
(ECP57MWB), O, N, and P (6-31+G*), C (6-31G*), and H (6-31G)).
b ∆∆G° ) ∆G°Ln - ∆G°La. c Estimated from the linear relationship
d
between ∆G‡ (B3LYP) and Ln3+ radii. ∆∆G‡
) -
∆G‡(TS1)
e
∆G‡(TS2). Estimated from ∆G‡ ) RT ln[(kBT)/(hk1)], with kB and h
being the Boltzmann and Planck constants, respectively, and k1 was
taken from Table 1 at pH ) 7.00, and T is the temperature. f Calculated
from ∆G° ) -RT ln K◦QP g Calculated as ∆G‡(exp) - ∆G‡(B3LYP).
.
Analyzing the effects of Ln3+ in the ground states shows that
there is an increase in the C-O1-Ln-N dihedral angle from
La3+ to Er3+, which is opposite to the trends in the transition
states. Such behavior indicates that lanthanide ions with smaller
radii should interact favorably with N and O1 atoms during
leaving group departure. Because catalysis proceeds without an
increase in the P-O1 bond length, it seems that electronic, rather
than structural effects, are responsible for acceleration of 8QP
hydrolysis, and atomic partial charges were estimated with the
ChelpG method for the optimized structures (Table 7). Forma-
tion of a complex with Ln demonstrates the decrease of initial
state free energy.
PO3- would allow limited “association” with a water molecule
not included in the reaction scheme. Here the distinction between
“associative” and “dissociative” is essentially one of timing.
Computation indicates extensive formation of PO3 in the
transition state, but it is not evident from experiment and a single
step DNAN is the most probable mechanism.
-
The B3LYP calculated values of ∆G‡ for the gas phase
indicate that these reactions are strongly favored by lanthanide
ions with decreasing activation Gibbs energies from 56.9 kJ
mol-1 in their absence to 21.6, 14.6, and 8.3 kJ mol-1 in the
presence of La3+, Sm3+, and Tb3+, respectively. These calculated
values are largely independent of the basis sets used for the P
atom. The following changes of ∆G‡ are calculated: 56.9, 57.3,
and 59.9 kJ mol-1 with the 6-31G*, 6-31+G*, and 6-311+G(3df)
basis sets, respectively, in the absence of Ln3+ and 21.7, 21.6,
and 26.7 kJ mol-1 in the presence of La3+ (Supporting
Information, Table S16). We note that the lanthanide ions
energetically favor the reaction with significant decreases of the
activation enthalpy, whereas the entropic contributions become
less favorable with Ln3+. This conclusion is reasonable probably
because the PO3- moiety can coordinate to the lanthanide ion,
thus making TS2 more compact than TS1, which decreases the
Differences in atomic charges in the ground ([8QP(H2O)3]2-)
and transition states (TS1) of the uncatalyzed reaction indicate
a largely dissociative mechanism, with decrease of the partial
positive charge on P and the O1 atom becoming more negative
with metaphosphate and aryloxide ion formation.
Despite postulated electronic effects of Ln3+, partial atomic
charges in ground and transition states, TS2 and TS1 of
catalyzed and uncatalyzed reactions, are unexpectedly similar
and vary on P and O1 atoms by only ∼0.04e and ∼0.17e,
respectively, even though, P-O1 bond lengths change by ∼0.8
and ∼1.7 Å, respectively. In addition, while P-O1 bond lengths
J. Org. Chem. Vol. 74, No. 3, 2009 1051

Souza et al.
TABLE 7. ChelpG Atomic Partial Charges at the B3LYP/ECP Level for [8QP(H2O)3]2-, [Ln ·8QP(H2O)5]+, and Its Respective Transition
States, TS1 and TS2a
[8QP(H2O)3]2-
[Ln·8QP(H2O)5]+
GS
TS
La
Sm
Tb
Er
TS2 (La)
TS2 (Sm)
P
C
N
O1
O2
O3
O4
Ln
1.23
1.17
0.38
1.40
0.09
1.40
0.03
1.40
0.01
1.39
0.03
1.38
0.18
1.37
0.20
0.32
-0.63
-0.42
-0.84
-0.84
-0.84
-
-0.69
-0.62
-0.67
-0.68
-0.68
-
-1.04
-0.60
-0.86
-0.91
-0.65
2.57
-1.03
-0.58
-0.86
-0.93
-0.64
2.53
-1.01
-0.54
-0.86
-0.93
-0.63
2.45
-1.01
-0.54
-0.87
-0.95
-0.63
2.45
-0.88
-0.74
-0.83
-0.86
-0.56
2.58
-0.90
-0.75
-0.82
-0.85
-0.56
2.55
a ECPs and basis sets: La3+ (ECP46MWB), Sm3+ (ECP51MWB), Tb3+ (ECP54MWB), Er3+ (ECP57MWB), O and N (6-31+G*), C and P (6-31G*),
and H (6-31G).
in [8QP(H2O)3]2- and [8QP·Ln(H2O)5]+ complexes are similar,
1.80 and 1.78 Å, respectively, ChelpG charges are P+1.23-O-0.42
and P+1.40-O-0.60, respectively, indicating increased polarization.
This difference could indicate a higher stability of structure 4
relative to 5, indicating that lanthanide ions favor leaving group
departure in a loose transition state, which is consistent with a
single step DNAN mechanism. However, we note that partial
atomic charges are only as reliable as the energies calculated
with DFT. Thus, caution should be exercised when using atomic
charges to interpret reaction mechanisms, and other electronic
quantities should also be used.
reaction coordinates for these associative paths are illustrated
in Figures S5 and S6 in the Supporting Information.
It is clear that the reaction paths involving the direct attack
of the water molecule in the absence and presence of La3+ are
highly energetic and, are thus, unfavored compared to the
dissociative paths presented in Figures 9 and 10. In addition,
the reaction profiles illustrated in Figures S5 and S6 (Supporting
Information) are very steep, which make determination of the
transition-state structures very difficult, and even after more than
20 attempts such determinations were unsuccessful. However,
despite the lack of these transition-state structures, given the
large energetic differences between these reaction paths and the
dissociative ones (Figures 9 and 10), the former would be
disallowed. In order to avoid comparison between different
model systems (Figure 9 and Figure S4, Supporting Informa-
tion), the dissociative reaction profile P-O1 has been calculated
with the structure in Figure 9 with one additional water
molecule. The energy maximum in this P-O1 reaction coor-
dinate is ∼90 kJ mol-1 (Supporting Information, Figure S7)
and is significantly lower that that found in Figure 9 (∼270 kJ
mol-1) and very similar to ∆H‡ (79.2 kJ mol-1) in Table 6,
thus validating the previous statement.
Lanthanide ions increase the P-O1 bond polarization bringing
the electronic structure of the reactant (ground state) closer to
that of the transition state for P-O1 dissociation (in hydrolysis)
and, in addition to this polarization, coordination of Ln3+ to
O1 decreases the C-O1 bond order. Wiberg bond indexes on
the natural atomic orbital (NAO) basis are 1.17 and 0.97,
consistent with the overlap-weighted NAO bond orders: 0.94
and 0.80 for [8QP(H2O)3]2- and [8QP·La(H2O)5]+ complexes.
These results favor structure 5, where the metaphosphate ion is
stabilized, which is close to a transition state in a DNAN type
dissociative mechanism. The similarities of structure 5 and the
transition state, induced by Ln3+, is also shown by the decreasing
C-O1 bond distance which decreases by only ∼0.02 Å (1.37
Å × 1.35 Å) from the ground state to TS2 in the
[8QP·Ln(H2O)5]+ complex.
Conclusions
Hydrolysis of 8-quinolyl phosphate (8QP) in the presence of
trivalent lanthanides (Ln ) La, Sm, Eu, Tb, and Er) proceed
through formation of an [Ln·8QP]+ complex that decomposes
into products. The computational studies indicate that there is
significant P-O1 bond breaking in the transition state, a result
consistent with the fact that phosphate monoesters dianions show
a flat potential energy surface and that the transition state
changes from associative to dissociative upon decrease in the
pKa of the leaving group.11 The theoretical results were
fundamental in facilitating understanding of the mechanistic
results, especially because kinetics, activation parameters and
isotope effects did not allow a conclusive distinction between
associative and dissociative mechanisms. Apparently, formation
of the [Ln ·8QP]+ complex alters electron distribution and
effectively lowers the pKa of the leaving group, favoring the
dissociative pathway. The reactions exhibit more than 107-fold
rate enhancements, which B3LYP calculations indicate are
driven by leaving group and metaphosphate stabilization in a
single step DNAN-type dissociative mechanism, consistent with
limited nucleophilic assistance, as shown by the low hydroxide
ion dependence and the small effects of Ln3+ radii on rate
constants, k1.
The coordination of PO3- to the Ln3+ ion changes the features
of the ground and transition states, because without lanthanide
-
ion the interaction of the PO3 with water is significant (see
Figures 9 and 10), whereas upon coordination, these interactions
lose their importance (see Figure 10), which considerably alters
the reaction paths.
The reaction paths involving a direct water attack on the P
atom, without and with the presence of the lanthanide ion, have
also been explored by the same computational approach. The
model systems used differ slightly from those presented in
Figures 9 and 10 by the presence of one additional hydration
water molecule, which will be responsible for the attack. The
1052 J. Org. Chem. Vol. 74, No. 3, 2009

Lanthanide-Catalyzed Hydrolysis of a Phosphate Monoester
Spectrophotometric pH Titration. Absorbances were monitored
on a diode-array spectrophotometer with a thermostatted cell holder
at 25.0 °C with 33.3 µM of 8QP with 0.01 M buffer: HCOOH (pH
3-4.5); CH3COOH (pH 4-5.5); NaH2PO4 (pH 5.5-7.8); H3BO3
(pH 7.8-9.0).
Experimental Section
Materials. Inorganic salts, buffers, and 8-hydroxyquinoline
(8QOH), analytical grade, were used without further purification.
Bis-tris-propane (BTP) and hydrated lanthanide chlorides
(LnCl3 ·xH2O), >99% purity, were used as received. Solvents were
dried over drying agents and distilled before use. Distilled, deionized
water was used in all studies, and CO2 was removed by boiling.
1H and 31P NMR spectra were recorded at 200 and 81 MHz,
respectively, in D2O with sodium 3-(trimethylsilyl)propionate (TSP)
NMR titration was at 25.0 °C. The 8QP solution (10 mg/mL)
was titrated with NaOD with pD ) pHread + 0.4.55
Calculations were performed with Gaussian 9856 at the DFT
level, with the B3LYP functional.57,58 Defaults for convergence
and optimization were used without any symmetry constraints. The
46 + 4fN electrons of Ln3+ were treated as core electrons (MWB46
La3+, MWB51 Sm3+, MWB54 Tb3+, and MWB57 Er3+) described
by the effective core potential (ECP) of Dolg et al.59 which
describes the valence electrons by the contracted basis sets (7s6p5d)/
[5s4p3d]. The N, O, and P atoms were described by the 6-31+G*
basis sets and C and H by 6-31G* and 6-31G basis sets,
respectively.60,61
1
as internal reference for H NMR and 85% phosphoric acid as
external reference for 31P NMR spectra.
Synthesis of Quinolinium-8-yl Hydrogen Phosphate (8-Quinolyl
Phosphate, Zwitterion Form).29,30 A solution of PCl5 (718 mg, 3.45
mmols) in CHCl3 (15 mL) was added dropwise to 8QOH (500 mg,
3.45 mmol in CHCl3, 15 mL) in an ice-water bath. The mixture
was stirred at room temperature for 60 min. Water was then added
(∼3 equiv) and the mixture left to react overnight. The solvent
was removed under reduced pressure, and acetone (10 mL) and
water (3 mL) were added to the crude oil. Pale fine crystals formed
slowly and were collected by filtration and acetone washed to
The initial conformation of 8QP2- without lanthanide was
obtained conventionally. With a lanthanide, initial structures of the
complexes were based on coordination numbers of 8 and 9, as
probable binding sites of 8QP, the quinolinic nitrogen and phosphate
oxygen atoms. Water molecules and other ligands were positioned
at a distance of ca. 2.5 Å from the metal. The critical points in the
Potential Energy Surface (PES), namely, reactants, products,
intermediates and transition states, were properly characterized by
their force constants, which were all positive, except for the
transition state with its imaginary frequency. The transition states
were found through a continuous structural search on the PES and
then optimized by using an eigenvalue-following algorithm. The
charges were obtained with fully optimized structures and the
ChelpG procedure,62 with lanthanide radii of La3+ 1.216, Sm3+
1.132, Tb3+ 1.095 and Er3+ 1.062 Å.41 The natural orbital analysis
for the La3+ complexes was performed with the WMB28 ECP basis
set.59
give313 mg (40%, mp 200-202 °C dec). Results of H and 31P
1
NMR spectroscopy are consistent with a sample with purity >99%.
At pD ) 2.15 in D2O, the following signals were observed: 31P
1
NMR δ 3.72 ppm; H NMR δ 7.86 (dd, 1H, J56 ) 6.7 Hz and J67
) 7.8 Hz), δ 7.96 (d, 2H), δ 8.08 (dd, 1H, J23 ) 7.1 Hz and J34
)
5.6 Hz), δ 9.09 (d, 1H, J34 ) 5.6 Hz) and δ 9.12 ppm (d, 1H, J23
) 7.1 Hz).
Kinetics and Products. ATR-FTIR Studies. Reactions were
monitored at 25 ( 2 °C on a spectrometer with a MCT detector,
and a CIRCLE cell (Spectra Tech) mounted with a ZnSe crystal,
10 cm path length. Parameters were set to 10.0 Å aperture,
automatic gain and 0.20 cm s-1 scan rate. Each spectrum was
recorded in the 800-4000 cm-1 range with spectral resolution, 1.0
cm-1, and is the result of 64 interferograms. Reactions were started
by addition of 200 µL of 0.05 M stock aqueous solutions of 8QP
or K2HPO4 into 10.0 mL of reaction mixture.
Acknowledgment. We dedicate this paper to Prof. Francisco
Carlos Nart (in memoriam), and we acknowledge the Brazilian
agencies FUNCITEC (PRONEX) and CNPq for their financial
support and the Office of International Programs, NSF.
UV-vis Spectrophotometry. Buffered solutions were prepared
by addition of aqueous standard HCl (0.1M; Merck) to aqueous
BTP (0.01 M), and the pH of each reaction mixture was measured
at the beginning and end of each run. Hydrolysis of 8QP in the
presence of the Ln3+ were followed by monitoring the lanthanide
8-quinolinolate complex, [Ln·8QO]2+, at 257 nm and 25.0 °C in
quartz cuvettes controlled with a thermostated water-jacketed cell
holder. Reactions were started by injection of 10 µL of 10 mM
stock solutions of 8QP in water (pH ∼ 10 and stored in a refrigerator
to minimize hydrolysis) into 3 mL of aqueous 0.01 M BTP giving
33.3 µM 8QP. Absorbance versus time data (at least 90% reaction)
were stored directly on a microcomputer. First-order rate constants,
kobs, were estimated from linear plots of ln(A∞ - At) against time,
except when [La3+]/[8QP] < 40, where reactions were consecutive
and rate constants were estimated from eq 6
Supporting Information Available: Potentiometric titration
of 8QP, Cartesian coordinates of the B3LYP-optimized struc-
tures, and the energy profile for the associative and dissociative
path containing four water molecules. This material is available
free of charge via the Internet at http://pubs.acs.org.
JO801870V
(55) Schowen, K. B. J. Transition States of Biochemical Processes; Springer:
New York, 1978; pp 225-284.
(56) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.;
Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.;
Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson,
G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.;
Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Stefanov, B. B.;
Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez,
C.; Challacombe, M.; Gill, P. M. W.; Johnson, B. G.; Chen, W.; Wong, M. W.;
Andres, J. L.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98,
revision A.6; Gaussian, Inc.: Pittsburgh, PA, 1998.
kae-k t - k1e-k t
k1 - ka
1
a
At ) 1 +
A∞
(6)
{
}
where ka is the rate constant for formation of the [8QO ·La]
complex, with ka * k1 . rate constant for reversion of [8QO ·La]
to reagents. All correlation coefficients were >0.996, as estimated
by iterative least-squares fits.
Potentiometric Titration. The pKa values of 8QP were deter-
mined with a digital pH meter and a combined glass electrode.
Titrations were in a 150-mL thermostatted cell, under N2 at 25.0
°C, ionic strength 0.1 M, KCl, and 1.0 mM initial 8QP. The solution
was titrated with small increments of 0.1008 M KOH, CO2-free,
and precautions were taken to eliminate carbonate and CO2 during
the titration. The program BEST732 was used to calculate the
dissociation constants.
(57) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(58) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785–789.
(59) Dolg, M.; Stoll, H.; Savin, A.; Preuss, H. Theor. Chim. Acta 1989, 75,
173–194.
(60) Ditchfield, R.; Hehre, W. J.; Pople, J. A. J. Chem. Phys. 1971, 54, 724–
728.
(61) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972, 56, 2257–
2261.
(62) Breneman, C. M.; Wiberg, K. B. J. Comput. Chem. 1990, 11, 361–373.
J. Org. Chem. Vol. 74, No. 3, 2009 1053