Metal-cation-mediated hydrolysis of phosphonoformate diesters: Chemoselectivity and catalysis

Article abstract of DOI:10.1021/ja030437h

Hydrolyses in D₂O (pD 1.7-3.1) of dimethyl (7), methyl phenyl (8), phenyl methyl (9), and diphenyl (10) phosphonoformate diesters are substantially accelerated by Ce(IV), Th(IV), Zr(IV), and Hf(IV) cations. Chemoselectivity is observed, whereby

Full text of DOI:10.1021/ja030437h

Published on Web 08/14/2004
Metal-Cation-Mediated Hydrolysis of Phosphonoformate
1
Diesters: Chemoselectivity and Catalysis
†
Robert A. Moss,* Hugo Morales-Rojas, Saketh Vijayaraghavan, and Jingzhi Tian
Contribution from the Department of Chemistry and Chemical Biology, Rutgers, The State
UniVersity of New Jersey, New Brunswick, New Jersey 08903
Received July 18, 2003; E-mail: moss@rutchem.rutgers.edu
Abstract: Hydrolyses in D
10) phosphonoformate diesters are substantially accelerated by Ce(IV), Th(IV), Zr(IV), and Hf(IV) cations.
Chemoselectivity is observed, whereby Zr(IV) and Hf(IV) principally direct P-OR hydrolysis, whereas Th-
IV) and Ce(IV) mainly direct C-OR hydrolysis. Leaving group efficiency (OMe vs OPh) modulates the
2
O (pD 1.7-3.1) of dimethyl (7), methyl phenyl (8), phenyl methyl (9), and diphenyl
(
(
chemoselectivity. The metal cations also mediate further hydrolytic reactions of the initially produced
phosphonoformate monoesters 14, 15, 18, and 20. The origins of the P-OR/C-OR selectivity are discussed
in terms of the metal cation hydroxo species likely to be present in solution and the kinetics of the reactions.
The phosphonoformate trianion (“foscarnet”), 1 (PFA), is
We, and others,5a,7 have been interested in PFA diesters 3 as
PFA prodrugs. The diesters are relatively stable toward hy-
drolysis in neutral and mildly acidic aqueous solutions, whereas
triesters 2 are more labile and can also suffer P-C backbone
cleavage. Representative of the diesters, diethyl phosphonofor-
mate (3, R1 ) R2 ) Et) cleaves in pH 7.0 Tris buffer with kobs
active against herpes simplex virus and AIDS-related human
2
cytomegalovirus. However, PFA is largely trianionic at physi-
3
ological pH and displays poor membrane permeability and low
bioavailability. Interest therefore focuses on PFA mono-, di-,
4-6
and triesters as PFA prodrugs.
For example, P-monoesters
-
8
-1
5b
are more easily biotransformed to PFA than C-monoesters, and
) 5 × 10
s
(t1/2 ≈ 160 d), whereas base catalyzed
4
aryl P-monoesters are more active than their alkyl analogues.
hydrolysis occurs more rapidly (by C-O cleavage) to yield
Additionally, the activities of PFA triesters 2, bearing an
aromatic P-ester residue, and the corresponding monoanionic
PFA diesters 3 were similar, suggesting that the aryl residue is
P-monoesters 4, which can be further hydrolyzed to PFA; eq
1.
4
selectively hydrolyzed to the corresponding diester. Mechanistic
In a seminal and stimulating study, Thatcher and Ferguson
reported that aminocyclodextrins accelerated the aminolysis,
esterolysis, and hydrolysis of C-phenyl PFA diesters (3, R1 )
studies of PFA solvolysis have appeared.5c,d,6 Of special interest
is the observation that PFA triesters are orders of magnitude
more reactive at P than simple phosphonate diesters due to the
electronic influence of the R-carbonyl group.5
5
b
Ph). Hydrophobic bonding of the phenyl moiety inside the (R
or â) cyclodextrin cavity and electrostatic interactions of the
phosphate group with the positively charged ammonium groups
c,e
†
Current address: Centro de Investigaciones Quimicas, Universidad
5b
on the rim of the torus contributed to the observed reactivity.
Autonoma del Estado de Morelos, Cuernavaca, Morelos, Mexico.
Our interest in PFA diester hydrolysis derives from previous
studies of metal-cation-mediated scission of phosphate diesters
and phosphonate monoesters. Ce(IV), Th(IV), Zr(IV), and Hf-
(
(
(
1) Taken in part from the Ph.D. dissertations of H.M.-R., S. V. Rutgers
University, New Brunswick, NJ, 2001 and 2002, respectively.
2) Physicians Desk Reference; Medical Economics Co.: Montvale, NJ, 2000;
pp 600-603. See also: Obert, B. Pharmacol. Ther. 1989, 40, 213.
a
3) The pK values of PFA at µ ) 0.1 are 0.78, 3.61, and 7.57: Song, B.;
(
IV) salts in aqueous solution manifest high reactivity in the
Chen, D.; Bastian, M.; Martin, R. B.; Sigel, H. HelV. Chim. Acta 1994, 77,
738.
8
1
hydrolysis of both activated and unactivated phosphate diesters.
(
4) Noren, J. O.; Helgstrand, E.; Johansson, N. G.; Misiorny, A.; Stening, G.
J. Med. Chem. 1983, 26, 264.
4+
4+
Soluble forms of Ce and Th afford accelerations of ∼2.4-
2.8 billion in the hydrolysis of bis-p-nitrophenyl phosphate.8
More remarkably, an acidic aqueous solution of uncomplexed
,9
(
5) (a) Ferguson, C. G.; Gorin, B. I.; Thatcher, G. R. J. J. Org. Chem. 2000,
6
5, 1218. (b) Ferguson, C. G.; Thatcher, G. R. J. Org. Lett. 1999, 1, 829.
(
c) Krol, E. S.; Thatcher, G. R. J. J. Chem. Soc., Perkin Trans. 2 1993,
793. (d) Krol, E. S.; Davis, J. M.; Thatcher, G. R. J. Chem. Commun. 1991,
18. (e) Thatcher, G. R. J.; Krol, E. S.; Cameron, D. R. J. Chem. Soc.,
1
(7) Gorin, B. I.; Ferguson, C. G.; Thatcher, G. R. J. Tetrahedron Lett. 1997,
Perkin Trans. 2 1994, 683.
38, 2791.
(
6) (a) Mitchell, A. G.; Nicholls, D.; Irwin, W. J.; Freeman, S. J. Chem. Soc.,
Perkin Trans. 2 1992, 1145. (b) Mitchell, A. G.; Nicholls, D.; Walker, I.;
Irwin, W. J. Freeman, S. J. Chem. Soc., Perkin Trans. 2 1991, 1297.
(8) Bracken, K.; Moss, R. A.; Ragunathan, K. G. J. Am. Chem. Soc. 1997,
119, 9323.
(9) Moss, R. A.; Zhang, J.; Bracken, K. Chem. Commun. 1997, 1639.
10.1021/ja030437h CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 10923-10936
9
10923

A R T I C L E S
Moss et al.
Ce4 hydrolyzes dimethyl phosphate (5, DMP) with an ac-
+
Scheme 1
8
celeration of ∼2 × 10 , reducing its half-life from >8000 years
at pH 7 to only 22 min.10 Subsequent H2 O labeling studies
18
4+
revealed that Ce -mediated hydrolysis of DMP occurs with
1
1
9
1% P-O cleavage and about 9% O-Me cleavage. The
resemblance between DMP and dimethyl phosphonoformate (3,
R1dR2)Me, DMPF) is both clear and suggestive. Zr4+ and Hf4+
survive with substrate 9, where the “dice are loaded” in favor
of C-O cleavage? We also revisit the question of the aggrega-
tion states of the chemoselective M catalysts.
1
2
are also active in the mediated hydrolysis of phosphate diesters,
4
+
4
+
4+
4+
while Ce , Th , and Zr display huge kinetic advantages in
the cleavage of p-nitrophenyl methylphosphonate, 6.
Our studies of metal-mediated phosphorolysis appeared within
1
3
Results
14
a matrix of many related contributions. With respect to
Synthesis. The sodium salt of DMPF (7) was prepared by
1
0,11
15
16
DMP,
for example, Co(III)-cyclen, and Cp2MoCl2 have
4
sodium iodide demethylation of commercial trimethyl phospho-
also been shown to educe very large hydrolytic rate accelera-
tions. And the hydrolyses of more reactive phosphodiesters are
noformate (2, R1 ) R2 ) R3 ) Me) in refluxing acetone. It
1
31
was characterized by H and P NMR spectroscopy, as well as
1
7
speeded by a variety of highly charged metal cations.
elemental analysis.
Substrate 8 was prepared by the more complicated sequence
of Scheme 1. Trimethyl phosphonoformate was converted in
Given the strong catalysis exhibited by metal cations in
phosphodiester hydrolysis, we examined the reactivity of Ce-
(
IV), Th(IV), Zr(IV), and Hf(IV) toward DMPF, 7 (i.e., 3 with
9
1% yield to P-bis-trimethylsilyl C-methyl phosphonoformate
18
R1 ) R2 ) Me). Our principal finding was an unprecedented
(11) with trimethylsilyl bromide. Treatment of the latter with
PCl5 gave 69% of the phosphonodichloridate 12, which, upon
reaction with phenol and DBU in CH2Cl2, afforded C-methyl
diphenyl phosphonoformate 13. Hydrolysis of 13 with dilute
base gave the desired C-methyl P-phenyl phosphonoformate 8,
in 60% yield for the last two steps. Substrate 8 was characterized
1
31
by H and P NMR and by elemental analysis. It has been
previously prepared by Noren using a less accessible procedure.4
Substrate 9 was made by reacting phenyl chloroformate with
trimethyl phosphate to yield C-phenyl dimethyl phosphonofor-
mate (2, R1 ) Ph, R2 ) R3 ) Me) in 97% yield, followed by
sodium iodide demethylation of the latter in refluxing acetone
4+
chemoselectivity in which acidic aqueous solutions of Ce and
4+
4+
Th directed the C-OMe hydrolysis of DMPF, whereas Zr
4
+
18
and Hf afforded preferential P-OMe hydrolysis. All the
hydrolyses were accelerated, relative to simple acidic catalysis.
The origins of the accelerations and chemoselectivity were
4+
ascribed to the nature of the M -DMPF interactions and the
(
80%). Phenyl methyl phosphonoformate was fully character-
ized.
Finally, diphenyl phosphonoformate (10) was prepared by
aggregation states of M4 in aqueous solutions.
+
18
In this full paper, we extend our studies of metal-cation-
mediated hydrolysis to phosphonoformates 8-10, in which the
Me esters of 7 are systematically replaced by the more
hydrolytically sensitive Ph esters. An underlying question is how
robust is the P-OR > C-OR chemoselectivity exhibited by
conversion of diphenyl H-phosphite [(PhO)2PH] to the P-anion
with NaH in THF, followed by reaction of the latter with phenyl
chloroformate to give triphenyl phosphonoformate (2, R1 ) R2
) R3 ) Ph). This was directly hydrolyzed to 10 by sodium
bicarbonate in aqueous acetonitrile. Substrate 10 was purified
by column chromatography and fully characterized (as the
hemihydrate).
4+
4+
Zr and Hf toward phosphonoformate diesters? For instance,
does the kinetically controlled P-OR > C-OR discrimination
(
(
(
(
(
10) Moss, R. A.; Ragunathan, K. G. Chem. Commun. 1998, 1871.
11) Moss, R. A.; Morales-Rojas, H. Org. Lett. 1999, 1, 1791.
Reactions and Kinetics; DMPF. We first examined the
12) Moss, R. A.; Zhang, J.; Ragunathan, K. Tetrahedron Lett. 1998, 39, 1529.
13) Moss, R. A.; Ragunathan, K. G. Langmuir 1999, 15, 107.
4
+
4+
4+
hydrolysis of DMPF (7), mediated by Ce , Th , Zr , and
14) For “early” studies, see: (a) Breslow, R.; Huang, D. L. Proc. Natl. Acad.
Sci. U.S.A. 1991, 88, 4080. (b) Morrow, J. R.; Buttrey, L. A.; Shelton, V.
M.; Berback, K. A. J. Am. Chem. Soc. 1992, 114, 1903. (c) Morrow, J. R.;
Buttrey, L. A.; Berback, K. A. Inorg. Chem. 1992, 31, 16. (d) Komiyama,
M.; Matsumura, K.; Matsumoto, Y. Chem. Commun. 1992, 640. (e)
Schneider, H.-J.; Rammo, J.; Hettich, R. Angew. Chem., Int. Ed. Engl. 1993,
4
+ 18
Hf . Reactions were initiated by mixing 10 mM of DMPF
with 25 mM ZrCl4, HfCl4, Ce(NH4)2(NO3)6, or Th(NO3)4‚4H2O
at 25 °C in D2O containing 0.5 M NaClO4. Pyrazine was added
as an internal H NMR integration standard; O-methyl meth-
ylphosphonate served a similar purpose for P NMR spectros-
copy. The pD’s of the unbuffered reaction solutions were
established by hydrolysis of the metal cations: Zr , 1.7; Hf ,
.2; Ce , 1.9; Th , 3.1. The reactions were followed by H
and P NMR, and products were verified by NMR spiking
experiments with authentic materials (see Experimental Section).
For P NMR experiments, it was necessary to remove the metal
1
3
2, 1716. (f) Rammo, J.; Schneider, H.-J. Liebigs Ann. 1996, 1757. (g)
31
Takasaki, B. K.; Chin, J. J. Am. Chem. Soc. 1993, 115, 9337. (h) Breslow,
R.; Zhang, B. J. Am. Chem. Soc. 1994, 116, 7893.
(
(
(
15) Kim, J. H.; Chin, J. J. Am. Chem. Soc. 1992, 114, 9792.
16) Kuo, L. Y.; Barnes, L. A. Inorg. Chem. 1999, 38, 814.
17) (a) Williams, N. H.; Takasaki, B.; Wall, M.; Chin, J. Acc. Chem. Res. 1999,
4
+
4+
4+
4+
1
2
3
2, 485. (b) Blasko, A.; Bruice, T. C. Acc. Chem. Res. 1999, 32, 475. (c)
3
1
Kr a¨ mer, R. Coord. Chem. ReV. 1999, 182, 243. (d) Roigk, A.; Hettich, R.;
Schneider, H.-J. Inorg. Chem. 1998, 37, 751. (e) Molenveld, P.; Engbersen,
J. F. J.; Reinhoudt, D. N. Chem. Soc. ReV. 2000, 29, 75.
31
(
18) Moss, R. A.; Morales-Rojas, H. J. Am. Chem. Soc. 2001, 123, 7457.
10924 J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004

Hydrolysis of Phosphonoformate Diesters
A R T I C L E S
Scheme 2a
a
Species are depicted in anticipated ionization states in D2O at experimental pD’s.
Table 1. Kinetic Data for Dimethyl Phosphonoformate (7) and 14a
cations by chelation with EDTA or tartrate in order to suppress
line broadening. Reaction aliquots, so treated, were then
examined at appropriate time intervals.
Under the foregoing conditions, Zr - or Hf -mediated
hydrolyses of DMPF occur mainly at P-OMe. Thus, the proton
NMR spectrum of DMPF in D2O shows Me signals at δ 3.71
product distributions for
04 kobs (s-1
b
cleavage of 7 (%)c
1
)
M(IV)
DMPF (7)
14
k
M(IV)/kD+d
14
15
4+
4+
Zr
4.4
4.0
1.3
5.2
0.022
0.026
0.42
3300
3100
980
79
90
21
10
>95
90
Hf
Th
Ce
e
0.93e
(
s, C-OMe) and 3.58 (d, JP-H ) 11.1 Hz, P-OMe). Reaction
3900
10
4
+
with, e.g., Hf , mainly removes the P-OMe doublet. In its
place, singlets appear at δ 3.33 (MeOD) and 3.79 for the product
C-monoester 14. Similar changes attend the reaction of DMPF
a
Conditions [M(IV)] ) 25 mM, [substrate] ) 10 mM, 0.5 M NaClO4
in D2O, 25 °C, pD 1.7 (Zr), 2.2 (Hf), 3.1 (Th), 1.9 (Ce). b Monitored by 1H
NMR integration of released MeOD relative to an internal pyrazine standard.
c
31
Errors of duplicate runs ( 8% (maximum). From P NMR spectra of
d
aliquots quenched at 15, 30, or 60 min with EDTA; errors ( 2%. kD+ )
-
7
-1
1
.3 × 10
s
for the acid-catalyzed cleavage of 7 at pD 1.7-2.2; see
e
text. See text.
observed over 5 half-lives with kobs ) 4.0 × 10 s-1 (Hf )
-
4
4+
4+
-4 -1
4+
with Zr . A minor competitive pathway, in each case, involves
C-OMe cleavage to P-monoester 15. Here, the C-OMe singlet
of DMPF is reduced, and the P-OMe doublet of 15 appears at
δ 3.74 (JP-H ) 11.7 Hz); see Scheme 2.
and 4.4 × 10
s
(Zr ). These values appear in Table 1.
4
+
4+
4+
Contrasting behavior is found for Ce and Th . Th
cleaves DMPF only at C-OMe, affording P-monester 15, which
is stable to Th for at least 5 days. In the H NMR, MeOD (δ
3.33) and the doublet of 15 at δ 3.74 grow in as DMPF decays;
4
+
1
The 14/15 distributions of the Hf and Zr4+ cleavages of
DMPF were best ascertained by 31P NMR spectroscopy of
EDTA-quenched reaction aliquots taken at 15, 30, or 60 min.
4+
31
P-OMe scission to 14 is not competitive. In the P NMR,
formation of 15 is signaled by growth of a broad signal at δ
6.3; spiking with authentic 15 supported this assignment.
Similarly, Th(IV) produces only product 15 when the cleavage
1
31
19
The H-decoupled P NMR resonances of DMPF, 14, and 15
1
9
in D2O appeared at δ -2.5, -5.2, and 6.9, respectively.
Integration gave the 14/15 distributions ((2%) as 79:21 from
of DMPF is carried out in H2O instead of D2O.
4
+
4+
20
4+
Zr and 90:10 from Hf ; cf. Table 1. The 14/15 distribution
With Ce a similar pattern ensues: DMPF decays in favor
4+
1
19
for the Zr cleavage of DMPF was similarly determined by
of 15 ( H NMR doublet at δ 3.61 ) and MeOH. However,
precipitation occurs after ∼1000 s, so that the rate constant,
3
1
P NMR in H2O (locked to D2O in a capillary tube). After
integration of the appropriate signals, the distribution was 80:
0, experimentally identical to the cleavage distribution found
in D2O.
The reaction kinetics were followed by monitoring MeOD
formation in the proton NMR. Pseudo-first-order behavior was
-
4
-1
k
obs
) 5.2 × 10 s , is obtained from the initial rate. About
4+
2
10% of C-monoester 14 also forms by Ce P-OMe cleavage
of DMPF.
4+
We also observed M -mediated reactions of the monoesters
4+
4+
1
4 and 15. Zr and Hf slowly cleaved C-monoester 14 to
3
1
-6 -1
PFA (1) [ P at δ 4.2], with rate constants 2.2 × 10
s
and
(
19) The 31P NMR chemical shifts vary somewhat in the presence of the different
metal cations. They are also sensitive to pD or pH.
-6 -1
4+
4+
2
.6 × 10 s , respectively. Zr and Hf induced decar-
1
8
21
(
20) A labeling experiment was conducted in 47% O-H
2
O with DMPF (P-
boxylation of P-monoester 15 to D-phosphonate 16. Further
13
31
1
O CH
of 14, with stoichiometric O incorporation, and released CH
Therefore, P-O-Me scission of DMPF by Hf(IV) and (presumably) Zr-
IV) occurred only by P-O cleavage.
3
) and Hf(IV). P and H NMR monitoring showed the formation
18
13
16
3
OH (only).
(21) 31P NMR: δ 9.03 (t, JP-D ) 98 Hz). This signal becomes a triplet of quartets
1
(
in the H-coupled spectrum.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004 10925

A R T I C L E S
Moss et al.
Table 2. Kinetic Data for Phosphonoformates 8 and 18a
product distributions for
hydrolysis of 16 then led to MeOH release and the formation
-
22
of DPO3D . The rate constants for this latter reaction,
10 kobs (s-1
2
)
cleavage of 8 (%)b
-
5
-1
measured by MeOH release, were 2.4 × 10
10 (Hf). About 10% of PFA also formed in these
reactions.
s
(Zr) and 2.1
-
5
-1
M(IV)
8
18
1.54
14
18
×
s
Zrc
2.33
>95
>95
c
Th4+ slowly cleaves C-monoester 14 with kobs ) 4.2 × 10-5
Hf
0.65
1.05
Th
0.016d
0.0026c
>95
s- . PFA appears to be the final product of this reaction as
1
3
1
a
identified by P NMR with tartrate added to chelate Th(IV).
Conditions: as those in Table 1 (a), except [substrate ] ) 0.2 mM for
b
4+
UV kinetics (10 mM for NMR kinetics). Product distributions determined
Ce induces net hydrolysis and decarboxylation of 14 to
3
1
c
by P NMR of solutions after quenching with EDTA or tartrate. Kinetics
3
1
-5 -1
phosphate anion ( P δ 2.1) with kobs ) 9.3 × 10 s . PFA is
d
by UV monitoring of PhOH release at 260, 270, and 280 nm. Kinetics by
H NMR in D2O solution following disappearance of substrate C-OMe.
4+
4+
1
stable to Ce . With P-monoester 15, Ce induces decarboxy-
lation to methyl phosphate (17) ( P δ 3.07, q, JP-H ) 10.8
Hz).
31
in the UV experiment. Qualitatively, however, P-OPh cleavage
In Scheme 2, we collect the set of reactions described above.
Here, branch A pertains to initial P-OMe scission, whereas
branch B comprises reactions that involve initial C-OMe
cleavage. Rate constants for reactions of DMPF and C-
monoester 14, as well as 14/15 distributions, appear in Table
1
of 8 is faster than P-OMe cleavage of DMPF. H NMR
confirms the production of phenol in the cleavages of 8; only
traces of MeOD appear after full release of the phenol. Product
31
31
1
4 is exclusively formed ( P δ -5.2); no 18 ( P δ -1.18),
the product of C-OMe cleavage, forms in these reactions (see
Scheme 3).
1
.
Control experiments with DCl/D2O furnished rate constants
for the acid-catalyzed hydrolysis of DMPF in the absence of
4
+
M . At pD 1.7, 1.9, and 2.2, kobs ) 1.35, 1.42, and 1.21 ×
-
7
-1
-7 -1
1
0
s , respectively (average kobs ) 1.32 × 10 s ). The
products at pD 1.7 were D-phosphonate 16 from C-OMe
cleavage followed by decarboxylation; subsequent hydrolysis
produced DPO3D anion.
Although 18 does not form from 8 by Zr(IV)- or Hf(IV)-
mediated cleavages, these cations do rapidly cleave 18: kobs-
With kD+ ) 1.3 × 10 s-1 for the acid-catalyzed hydrolysis
-
7
-2 -1
-2 -1
(Zr) ) 1.5 × 10
s
and kobs(Hf) ) 1.05 × 10 s , as
4+
of DMPF, we find that the M -mediated hydrolyses occur
determined by released phenol (UV). D-phosphonate 19 and
DPO D are the products, according to P NMR, along with a
2
3
-
31
∼
1000-4000 times more rapidly, as shown in Table 1. Note,
however, that acid catalysis induces only C-OMe scission and
is more akin to the Th - and Ce -mediated cleavages than
the Zr and Hf reactions in which P-OMe scission is the
major pathway (Scheme 2).
3
trace of PFA (analogously to the Zr(IV) or Hf(IV) cleavage of
P-monoester 15; cf. Scheme 2).
4+
4+
4
+
4+
The intriguing question with substrate 8 is whether Th(IV),
which manifests C-OMe specificity toward DMPF, will
maintain that selectivity despite the availability of a labile
In summary, Zr(IV) and Hf(IV) are chemoselective for the
P-OMe cleavage of DMPF, whereas Th and Ce direct
C-OMe scission. This pattern recurs in cleavages of monoesters
4 and 15: P-monoester 15 is not appreciably cleaved by Th-
IV) or Ce(IV), but C-monoester 14 is cleaved 16-42 times
more rapidly by Th(IV) or Ce(IV) than by Zr(IV) or Hf(IV).
Methoxycarbonyl Phenyl Phosphonate (8). Given the
P-OMe chemoselectivity toward DMPF expressed by Zr(IV)
and Hf(IV), it is reasonable to expect analogous selectivity in
the phosphorolysis of substrate 8, where the P-ester bears the
more reactive phenoxide leaving group. This expectation is
fulfilled.
4+
4+
25
P-OPh linkage in 8. In the event, C-OMe cleavage prevails.
1
4+
Followed by H NMR, Th mediates hydrolysis of 8 (δ 3.83),
-
4 -1
1
(
yielding MeOD with kobs ) 1.6 × 10 s , similar to k ) 1.3
× 10 for the Th -catalyzed C-OMe scission of DMPF
(Tables 2 and 1). After 5 h, quenching of the reaction solution
with tartrate (to remove Th ) reveals P-monoester 18 (and PFA)
as the only products ( P NMR).
-
4
-1
4+
s
4+
3
1
P-OPh cleavage of 18 by Th(IV) is slower than C-OMe
cleavage but does occur. UV-monitored phenol release gives
-5 -1
kobs ) 2.6 × 10 s . PFA is the product from P-OPh cleavage
of 18.
Table 2 collects rate constants pertaining to substrate 8, while
Scheme 3 summarizes the reactions. Note that the kinetic data
for the M cleavages of C-monoester 14 appear in Table 1
Zr(IV)- and Hf(IV)-mediated hydrolyses of 8 in D2O at pD
-
2
-1
1
.7 or 2.2 occur with kZr ) 2.33 × 10
s
and kHf ) 0.65 ×
4
+
-2
-1
24
10
s , respectively; see Table 2. We note greater reactivity
and are discussed in the section on DMPF hydrolysis.
of 8 compared to the P-OMe scission of DMPF. Thus, the
Zr(IV) and Hf(IV) P-OPh cleavages of 8 were too fast to be
followed by NMR and required UV monitoring. The UV rate
1
The acid-catalyzed hydrolysis of 8 was followed by H NMR
spectroscopy, monitoring the decrease of substrate at pD 1.7
4
+
-7 -1
1
(DCl) in the absence of M . We obtained k ) 1.1 × 10 s ,
constants are not directly comparable to those obtained by H
+
-7
similar to the D -catalyzed hydrolysis of DMPF (1.3 × 10
NMR because the ratio of metal cation/substrate is much higher
-
1
-
s , see above), with DPO3D as the product. Rate enhance-
3
4+
5
4+
(
(
22) 31P NMR: δ 3.2 (t, JP-D ) 97 Hz).
ments of 10 (Th ) to 10 (Zr ) are obtained for the M(IV)
5
7
23) DMPF is ∼10 -10 times more reactive than phosphodiesters (e.g., 5),
+
cleavages of 8, relative to D . Again, we note that the acid-
probably because of ground-state destabilization due to unfavorable
5
b,d
carbonyl/phophoryl electronic interactions.
Therefore, the catalytic
catalyzed process (C-OMe cleavage, decarboxylation) differs
from the P-OPh cleavage induced by Zr(IV) or Hf(IV). Most
4+
3
enhancement of DMPF hydrolysis at P-OMe by M is compressed (∼10 )
1
0
relative to the much larger enhancements observed with 5.
(
24) Kinetics were monitored by UV spectroscopy, following the release of
phenol at 260, 270, or 280 nm. Reactions were pseudo-first-order over 5
half-lives.
(25) Due to its proclivities for precipitation and substrate oxidation, Ce(IV) was
not studied with this substrate.
10926 J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004

Hydrolysis of Phosphonoformate Diesters
A R T I C L E S
Scheme 3
Table 3. Kinetic Data for Phosphonoformates 9 and 20a
does not persist with (C-OPh) substrate 9. Nevertheless, it is
remarkable that the P-OMe ester of 9 is at all cleaved by Zr-
product distributions for
1
02 kobs (s-1
)
cleavage of 9 (%)b
(IV) or Hf(IV) in competition with the C-OPh ester, consider-
M(IV)
9
20
15
90
79
>95
20
ing the differences in leaving group and carbonyl vs phosphoryl
site reactivities. Both differences favor C-esterolysis, so that
the P-selective character of Zr(IV)- and Hf(IV)-mediated
cleavages persists, even if P-esterolysis does not dominate with
substrate 9.
Zrc
1.79
0.61
0.18
0.020
0.011
0.24
10
21
c
Hf
c
Th
a
Conditions: as those in Table 1 (a), except [substrate ] ) 0.2 mM for
b
UV kinetics (10 mM for NMR kinetics). Product distributions determined
Reactions of C-monoester 20 with Zr(IV) or Hf(IV) proceed
3
1
c
by P NMR of solutions after quenching with EDTA. Kinetics by UV
monitoring of phenol release at 260, 270, and 280 nm.
-
4
-1
-4 -1
with kZr ) 2.0 × 10
s
and kHf ) 1.1 × 10 s , with the
release of phenol followed by UV spectroscopy. Only PFA is
produced after 8 h of reaction at 25 °C with either metal cation.
As expected, these C-OPh scissions of 20 to PFA are faster
than the corresponding conversions of C-OMe monoester 14
importantly, the chemoselectivity found with DMPF (P-OR
cleavage by Zr(IV) or Hf(IV) but C-OR cleavage by Th(IV))
is maintained with substrate 8 and echoed by P-monoester 18
-
6
-1
-6 -1
(
kZr ) 2.2 × 10 s , kHf ) 2.6 × 10 s ; see Table 1).
Cleavage of 20 with Th affords only PFA with kTh ) 0.24 ×
s .
(Table 2), where P-OPh cleavage by Zr(IV) or Hf(IV) is 400-
4
+
6
00 times faster than Th(IV) cleavage.
-
2
-1
1
0
Phenoxycarbonyl Methyl Phosphonate (9). An important
Reaction of diester 9 with Th4 (0.18 × 10 s ) gives only
+
-2 -1
question to be answered here is whether the P-OR chemose-
lectivity of Zr(IV) and Hf(IV) is sufficiently robust to endure
with substrate 9, where the C-OPh site is highly labile.
Followed by UV spectroscopic monitoring of released phenol
in D2O (conditions as above), Zr(IV)-, Hf(IV)-, and Th(IV)-
mediated hydrolyses of 9 exhibited good pseudo-first-order
1
31
C-OPh scission to phenol and C-monoester 15 ( H and
P
NMR described above). Given the C-esterolytic chemoselectivity
4
+
of Th and the good C-OPh leaving group of 9, this result is
unsurprising. Scheme 4 outlines the reactions described in this
section. The reactions of Zr(IV) and Hf(IV) with 15 are
discussed above in connection with DMPF.
-
2
-1
-2 -1
behavior with kZr ) 1.79 × 10 s , kHf ) 0.61 × 10 s ,
-
-
2
-1
Acidic hydrolysis of 9 (to 16 and DPO D ) at pD 1.7 (HCl)
and kTh ) 0.18 × 10 s ; see Table 3.
3
-
7
-1
1
occurs with kD+ ) 6.2 × 10 s , as followed by NMR
H NMR spectroscopy revealed that both MeOD (from
4
+
spectroscopy of the substrate. Therefore, M catalysis of the
P-OMe scission) and PhOD (from C-OPh scission) were
liberated by Zr(IV) or Hf(IV). Correspondingly, C-monoester
4
4+
cleavage of 9 occurs with enhancements of ∼2.9 × 10 (Zr )
3
4+
to 2.9 × 10 (Th ).
2
0 and P-monoester 15 were produced in these reactions, as
well as D-phosphonate 16. These products were observed by
Diphenyl Phosponoformate (10). Finally, we examined the
4
+
4+
Zr - and Th -mediated hydrolyses of diphenyl phosphono-
formate (10). Unfortunately, 10 mM 10 immediately precipitated
with 25 mM Zr(IV) or Th(IV), our standard concentrations. It
was necessary to use a 1:1 D2O/CD3CN mixture as the solvent
for reactions of 10.
3
1
31
4+
1
P NMR spectra obtained after 10 min of reaction (EDTA
The cleavage of 10 ( P δ -8.85) by Zr , followed by H
21
31
quench): δ 6.9 (15), -5.75 (20), and 9.0 (16). After 50 min
of reaction, only DPO3D anion (from 16) and PFA (from 20)
were present; see Scheme 4.
NMR and P NMR, revealed the formation of phenol (δ 7.25-
3
1
19
7.34) and C-monoester 20 ( P δ -5.25 ), the products of
P-OPh scission; see Scheme 5, branch A. No other signals were
observed, so that the Zr(IV)-mediated hydrolysis of 10 was
chemospecific at P-OPh. UV monitoring of released phenol
The distribution of 20/15 [P-OMe/C-OPh cleavage] was
0:90 for Zr and 21:79 for Hf , as determined from the P
4
+
4+
31
1
-
2
-1
NMR spectra at short reaction times. Thus, the P-OR chemose-
lectivity of Zr(IV) observed with (C-OMe) substrates 7 and 8
gave kZr ) 1.29 × 10
s
for the P-OPh cleavage of 10 by
Zr with [10] ) 0.2 mM and [Zr ] ) 25 mM; see Table 4.
4
+
4+
J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004 10927

A R T I C L E S
Moss et al.
Scheme 4
Scheme 5
Table 4. Kinetic Data for Phosphonoformate 10a
P-OPh vs C-OPh chemoselectivities, respectively, in parallel
to their behavior with substrate 7 (DMPF); see Table 1.
A control experiment in which DMPF was cleaved by Zr
product distributions for
cleavage of 10 (%)b
4+
2
M(IV)
10 kobs (s-1
)
18
20
in 1:1 CD3CN/D2O gave P-O scission to 14 and C-O scission
Zrc-e
Th
1.29
0.47
>95
10
31
to 15 (see Scheme 2) in a ratio of 37:63 as determined by
P
c-e
90
NMR. The P-OMe/C-OMe chemoselectivity of 80:20 ob-
4
+
served for Zr in 100% D2O (Table 1) is thus “reversed” in
a
Conditions: as those in Table 1 (a), except [substrate ] ) 0.2 mM for
b
CD3CN/D2O. Substrate cleavage is also ∼4.5 times slower (0.96
UV kinetics (10 mM for NMR studies). Product distributions determined
3
1
c
-4 -1 1
by P NMR of solutions after quenching with EDTA. Kinetics by UV
×
10 s , H NMR) in the binary solvent, compared to D2O
d
monitoring of phenol release at 270 nm. For kinetic data of 18 and 20,
see text and Tables 2 and 3. Solvent is 1:1 MeCN/D2O for product studies
-
4
-1
e
(4.4 × 10 s ). This influence of solvent on the chemose-
but D2O for UV kinetic studies.
lectivity of the Zr(IV)-mediated hydrolysis may reflect solvent
4
+
effects on the stabilities of the Zr
complexes that are
This rate constant is similar to that for the analogous P-OPh
responsible for the P-O selectivity (see Discussion).
Th cleavage of DMPF in 1:1 CD3CN/D2O gave only
4
+
4+
cleavage of 8 (see Table 2). Continued incubation with Zr
3
1
led to slow cleavage of 20 to PFA, as verified by P NMR,
C-OMe scission to 15, as in D2O alone, and with an almost
-
4
-1
-4 -1
1
with k ) 7.54 × 10
s
(UV). This value is somewhat higher
unchanged rate constant (1.16 × 10
s
by H NMR vs 1.3
4
+
-4 -1
4+
than the one observed in the direct reaction of 20 with Zr
× 10
s
in D2O, Table 1). Th -mediated hydrolysis is thus
-
4
-1
(
2.0 × 10 s , Table 3).
unaffected by the solvent change.
4+
In contrast, reaction of 10 with Th gave mainly P-monoester
8 via C-OPh scission, verified by 31P NMR spiking experi-
We also carried out rate constant vs M(IV) concentration
studies, from which a correlation of log k vs log[M(IV)] might
reveal the reaction order in metal cation. We performed these
experiments for substrate 8 with Zr(IV) and substrate 9 with
Th(IV), but we could not conveniently do so for DMPF (7),
where the dual necessities of a minimum substrate concentration
for NMR analysis and a high [M(IV)]/substrate ratio to ensure
pseudo-first-order behavior would require impractically high
metal ion concentrations.
With 8-Zr(IV) and 9-Th(IV), full P-OPh and C-OPh
cleavage selectivity (respectively) was observed (see above).
The kinetics of cleavage were readily followed by UV monitor-
ing of the released phenol in both D2O and H2O, which
1
ments; see Scheme 5, branch B. About 10% of 20 was also
observed in this cleavage. UV monitoring of released phenol
-
3
-1
4+
gave kTh ) 4.72 × 10 s . Continued incubation with Th
31
converted 18 to D-phosphonate 19 ( P δ 3.3, t, JP-D ) 99 Hz)
by decarboxylation and then to DPO3D anion by the hydrolytic
loss of phenol. The UV-determined rate constant for the latter
reaction was 1.05 × 10 s . It is of interest that the Th(IV)
cleavage of 18 in D2O gives PFA (Scheme 3), whereas
decarboxylation to 19 prevails in CD3CN/D2O.
-
4
-1
The main result, as portrayed in Table 4 and Scheme 5, is
that the Zr(IV) or Th(IV) hydrolyses of 10 manifest opposite
10928 J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004

Hydrolysis of Phosphonoformate Diesters
A R T I C L E S
permitted the evaluation of an apparent solvent isotope effect.
In these experiments, the maximum [M(IV)] ) 25 mM, [8 or
The well-known ability of metal ions to activate coordinated
water molecules (i.e., to decrease the pKa of water) leads to the
2
6,27
9
] ) 0.2 mM, [NaClO4] ) 0.5 M, and the solution pH or pD
formation of polynuclear hydroxo-metal species.
At suf-
-
5
was established by the natural hydrolysis of M(IV) at 25 mM,
with the pH or pD adjusted at lower metal ion concentrations
by the addition of DCl or HCl as required.
ficiently low concentration (<1 × 10 M), cations generally
appear to give mononuclear hydroxo species, but, at ordinary
-
3
concentrations (>1 × 10 M), they often produce several
polymeric species. Polynuclear hydroxo-complex formation is
established or suspected for the great majority of 3+ and 4+
ions, as well as for many 2+ ions. Thus, at concentrations
from 1 mM to 25 mM in weakly acidic solutions, metal ions
such as Ce(IV) and Zr(IV) are most likely present as discrete
polynuclear hydroxo species rather than as simple aquocations
The dependence of kobs on the cleavage of 8 by Zr(IV) in
H2O at pH 1.0 showed linear behavior within the range of metal
ion concentrations employed (∼2.5-25 mM). The same reaction
in D2O at pD 1.7 displayed linear behavior at lower concentra-
tions of metal ion, with some saturation beginning at higher
concentrations (see Figure S1 in the Supporting Information).
2
7
4
+
-
1
2 x
[M(H O) ] or the related monohydroxo species [M(OH)-
The second-order rate constants were k2 ) 0.86 ( 0.03 M
(H O)x-1](4-1)+.
-1
-1 -1
s
in H2O and 1.29 ( 0.06 M
s
in D2O. From these results
2
Despite much effort, identification of these polymeric species
has been difficult, in part due to the necessity of operating within
narrow concentration and pH ranges. There are only a few
detailed studies of phosphate and phosphonate ester cleavage
that characterize the active species of tri- or tetravalent metal
the apparent solvent isotope effect, k2(H2O)/k2(D2O) ) 0.66.
Although the difference between pH (1.0) and pD (1.7)
complicates a mechanistic interpretation, it seems clear that the
H2O to D2O change does not cause a major alteration in the
mechanism. Figure S2 shows the dependence of log kobs vs log-
2
8-30
cations in solution.
To rationalize their chemoselectivity,
[Zr(IV)] for the cleavage of 8 in both H2O and D2O at 25 °C
we need to understand the aqueous solution chemistry and
polynuclear hydroxo-metal species of Ce(IV), Th(IV), Zr(IV),
and Hf(IV). We make use of the available literature data,2
noting that a particularly detailed account of metal ion polym-
and 0.5 M NaClO4; good linear behavior was observed with
slopes of 0.84 ( 0.01 and 0.79 ( 0.03, respectively.
6,27,31
In contrast to the 8-Zr results, the dependence of kobs for
the cleavage of 9 by Th(IV) in H2O at pH 2.3 showed saturation
with increasing [Th(IV)]. Saturation was also observed in D2O
at pD 2.9. (See Figure S3 in the Supporting Information.)
Apparent second-order rate constants could be estimated from
the linear portions of the curves, affording k2 ) 0.44 ( 0.03
2
7
erization is offered by Baes and Mesmer.
A. Ce(IV). Based on several studies of Ce(IV) salts, with
both perchlorate and nitrate counterions, Baes and Mesmer
describe the formation of several Ce(IV) species as represented
2
7
-1
-1
-1 -1
by the set of equilibria in Chart 1. The log Kn,m data of Chart
can be processed by the program SPE given by Martell and
M
s
in H2O and 0.66 ( 0.01 M
s
in D2O. For this
1
system too, an apparent solvent isotope effect of 0.66 is obtained.
The apparent reaction order can be extracted from the linear
portions of the curves. Figure S4 (Supporting Information)
shows the dependence of log kobs on log[Th(IV)] for the
cleavages of 9 in H2O (pH 2.3) and D2O (pD 2.9) at 25 °C and
32
Motekaitis to afford the distribution of equilibrium concentra-
tions of the various Ce(IV) species at selected values of [Ce-
(IV)]total and pH. The SPE program approximates the degree of
formation of each metal ion species depending on a given set
of stability constants, total metal ion concentration, and solution
pH. Program output includes a “complete listing of each species,
its stoichiometry, concentration, and relative percentages all as
0
.5 M NaClO4. The slopes calculated from the linear fit were
0
.82 ( 0.04 and 0.84 ( 0.06, respectively. We will return to
the log kobs vs log[M(IV)] results in the Discussion section.
3
2
a function of pH.” A detailed discussion, complete program
code, and appropriate floppy disks are contained in ref 32. We
thus obtain [Ce(IV)] species distributions as a function of pH
which can be converted to the format of Figure 1 using the
program Origin 4.0 (Microcal).
Finally, we note that the saturation kinetics observed for the
9/Th(IV) system, in comparison to the linear behavior exhibited
by 8/Zr(IV), reflects differences in the degree of protonation of
the substrates under the experimental conditions where their
kinetics are measured. PFA diesters such as 8 and 9 should have
pKa’s similar to the first pKa of the parent PFA (∼0.78).
Therefore, at pH 1.0, substrate 8 should be ∼30% in the
protonated form, decreasing its apparent binding to the metal
cation. In contrast, at pH 2.3, substrate 9 should be fully
deprotonated and display saturation at a lower concentration of
metal ion.
At pH ∼2, dimeric and hexameric Ce(IV) species are
predicted to dominate. Structural information is scarce, although
1
2+
hexanuclear Ce6(OH)12
has been related to the crystal
4
+
structure of Ce6O4(OH)4(SO4)6, in which 6 Ce ions are in an
octahedral arrangement, and each of the four O and OH ions
are centered on the octahedral faces. Interestingly, in recent
2-
-
2
7
studies on the cleavage of cyclic adenosine 3′,5′-monophosphate
(
cAMP) by Ce(IV) acidic solutions, Komiyama and co-workers
Discussion
(
26) (a) Burgess, J. Metal Ions in Solution; Halstead Press: New York, 1978.
b) Burgess, J. Ions in Solution: Basic Principles of Chemical Interactions;
Ellis Horwood: New York, 1988. (c) Choppin, G. R. Radiochim. Acta
983, 32, 43.
(
Dominant Forms of the Hydroxo-Metal Species. Ce(IV),
Th(IV), Zr(IV), and Hf(IV) catalyze cleavages of phosphono-
formate substrates 7-10 by ∼3-4 orders of magnitude.
Moreover, the cations manifest an unprecedented chemoselec-
tivity in which Ce(IV) and Th(IV) favor C-OR scission,
whereas Zr(IV) and Hf(IV) prefer P-OR cleavage (given
identical leaving groups at both C-OR and P-OR, e.g., R )
Me or Ph). In this discussion we will consider these issues of
chemoselectivity and catalysis.
1
(
27) Baes, C. F., Jr.; Mesmer, R. E. The Hydrolysis of Cations; Wiley-
Interscience: New York, 1976.
(28) Komiyama, M.; Takeda, N.; Shigekawa, H. Chem. Commun. 1999, 1443.
29) Blewett, F. McC.; Watts, P. J. Chem. Soc. B 1971, 881.
(
(30) Hurst, P.; Takasaki, B. K.; Chin, J. J. Am. Chem. Soc. 1996, 118, 9982.
(
31) Smith, R. M.; Martell, A. E. Critical Stability Constants, Plenum Press:
New York, 1976; Vol. 4-6.
(32) Martell, A. E.; Motekaitis, R. J. Determination and Use of Stability
Constants, 2nd ed.; VCH: New York, 1992; p 173. The “SPE” program
was supplied as a floppy disk that accompanied the book and was
implemented on our PC.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004 10929

A R T I C L E S
Moss et al.
Chart 1. Equilibria for Ce(IV)a
a
b
-
Data and equations are taken from ref 27, Tables 7.3 and 7.4, and the discussion on pp 139-146. Values at I ) 3 M (ClO4 ). In parentheses, values
-
27
at I ) 3 M (NO3 ). Data unavailable for recalculation at I ) 0.5.
3
5 17
38
1
36
angle X-ray scattering (SAXS),
O and H NMR, Raman
3
7
39
spectroscopy, light scattering, and ultracentrifugation.
Despite intense investigation, there is still debate about the exact
nature of the species. Hafnium(IV) is much less studied; its
chemistry is usually considered to be very similar to that of
Zr(IV). Baes and Mesmer reviewed pre-1975 data3
8-40
and
proposed the set of equilibria shown in Chart 3 for Zr(IV)
species.
This set of equilibria only indicates the overall metal-to-
hydroxide stoichiometry of the species; no structural information
is available in most of the cases except for Zr4(OH)88 , which
+
is believed to exist as hydroxyl-bridged tetramers, [Zr4(OH)8-
(
H2O)16(Clx](
8-x)+
(21), similar to the unit in the crystal structure
of ZrOCl2‚8H2O.
Based on the foregoing, it has been generally accepted that
the predominant Zr(IV) species in acidic homogeneous aqueous
3
6,37
solutions is the hydroxyl-bridged tetramer (21).
However,
3
5c
in a recent SAXS study, Singhal et al. showed that the
tetramer was not the only species present. At 0.05 M ZrOCl2
in aqueous solutions in the pH range 0.1-1.5, an equilibrium
Figure 1. Speciation of Ce(IV) as a function of pH, with [Ce(IV)] ) 0.025
-
M, I ) 3 M (ClO4 ). The equilibrium concentrations of each Ce(IV) species
at each pH were calculated from the equilibrium constants in Chart 1 (taken
from ref 27), using the iterative program SPE (see ref 32). A description of
this program appears in the text.
2
+
between tetramers [Zr4(OH)8(H2O)16 Cl6] and octamers [Zr8-
(OH)20(H2O)24Cl12] was found (Chart 4, a) with a quotient, Q
)
0.2 ( 0.05 M (Chart 4, b). The octamer was proposed to
exist in a stacked form comprised of two tetramers (cf., 22,
water molecules are omitted for clarity).
directly correlate the observed reactivity with the dimeric Ce-
35c
Therefore, an
IV
4+ 28,33
(
IV)-hydroxo species [Ce 2(OH)4] .
additional equilibrium can be added to Chart 3, with a formation
constant (K8,20) calculated from K4,8 and the quotient Q; cf.,
Chart 4, c and d.
Evaluation of the Zr(IV) species at equilibrium can now be
made32 using the stability constants (Kn,m) given in Charts 3
B. Th(IV). From very careful potentiometric studies with
-
-
-
ClO4 , Cl , and NO3 counterions, as well as studies with other
techniques including ultracentrifugation, light scattering, and
X-ray scattering, Baes and Mesmer represent the formation of
27
various Th(IV) species by the equilibria in Chart 2. As with
Ce(IV), speciation diagrams can be calculated32 from these
equilibrium constants for different concentrations of Th(IV).
Figure 2 depicts the results with [Th(IV)] ) 0.025 M. At pH 3,
both monomeric and dimeric Th(IV) species exist. The dimer
Th2(OH)26 appears in studies with all 3 counterions27 and is
strongly supported by X-ray scattering results in which two
bridging hydroxides link the metal centers with a Th-Th
distance of 3.95 Å.2
and 4 (d). The speciation for Zr(IV) at 0.025 M as a function
of pH appears in Figure 3. Additionally, similar diagrams for
different [Zr(IV)]o concentrations could be evaluated so that
equilibrium concentrations for Zr tetramers and octamers as a
function of the total [Zr] at a given pH could also be obtained;
see Figure S5 in the Supporting Information.
+
(
34) Veyland, A.; Dupont, L.; Pierrand, J.-C.; Rimbault, J.; Aplincourt, M. Eur.
J. Inorg. Chem. 1998, 1765.
6a
(35) (a) Hu, M. Z.-C.; Zielke, J. T.; Lin, J.-S.; Byers, C. H. J. Mater. Res. 1999,
14, 103. (b) Singhal, A.; Toth, L. M.; Beaucage, G.; Lin, J.-S.; Peterson,
C. Zr(IV) and Hf(IV). The chemistry of zirconium (IV) salts
has been studied extensively in acidic aqueous solutions using
J. J. Colloid Interface Sci. 1997, 194, 470. (c) Singhal, A.; Toth, L. M.;
Lin, J.-S.; Affholter, K. J. Am. Chem. Soc. 1996, 118, 11529.
(36) A° berg, M.; Glaser, J. Inorg. Chim. Acta 1993, 206, 53.
34
various instrumental techniques, including protometry, small-
(
37) Hannane, S.; Bertin, F.; Bouix, J. Bull. Soc. Chim. Fr. 1990, 127, 43.
38) Angstadt, R. L.; Tyree, S. Y. J. Inorg. Nucl. Chem. 1962, 24, 913.
(
(
33) Sumaoka, J.; Takeda, N.; Okada, Y.; Takahashi, H.; Shigekawa, H.;
(39) Johnson, J. A.; Kraus, K. A. J. Am. Chem. Soc. 1956, 78, 3937.
(40) Zielen, A. J.; Connick, R. E. J. Am. Chem. Soc. 1956, 78, 5785.
Komiyama, M. Nucleic Acid Symp. Ser. 1998, 39, 137.
10930 J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004

Hydrolysis of Phosphonoformate Diesters
A R T I C L E S
Chart 2. Equilibria for Th(IV)a
a
b
Data and equations are taken from ref 27, Tables 8.5 and 8.8, and the discussion on pp 158-168. Calculated at I ) 0.5 M from equations in ref
2
7.
Th(IV) species, as opposed to tetrameric or octameric Zr(IV)
and Hf(IV) species. The hydrolytic models follow.
The chemoselectivity of cleavage for substrates 7-10 must
depend on which M-OH species are present under the reaction
conditions and the particular mode of substrate binding in each
case. For the PFA diesters, which at pD 1.7-3.1 are present in
-
3
aqueous solutions as the P-O monoanions, binding to the
metal species (i.e., Lewis activation) and the simultaneous
supply of a metal-bound hydroxide nucleophile to attack either
the phosphoryl or the carbonyl groups must account for the
observed hydrolytic accelerations in comparison with the acid-
3
0,42
catalyzed reactions (see Tables 1-4).
However, it should
+
be noted that D induces only C-O ester scission and is more
comparable to Th(IV) and Ce(IV), rather than to Zr(IV) and
Hf(IV), which mainly display reactivity at the phosphoryl ester.
1
1,18,28
We represented
Ce(IV) cleavage of (e.g., DMPF) by
23, where intramolecular hydroxide attack at the carbonyl center
Figure 2. Speciation of Th(IV) as a function of pH, with [Th(IV)] ) 0.025
(a) occurs via a five-membered transition state, whereas attack
at phosphorus (b) requires a more strained four-centered
transition state.
-
M, I ) 0.5 M (ClO4 ). The equilibrium concentrations of each Th(IV)
species at each pH were calculated from the equilibrium constants in Chart
2
(taken from ref 27), using the iterative program SPE (see ref 32). A
description of this program appears in the text.
From Figure 3, we reach a conclusion similar to that of
3
5c
Singhal: the tetramer dominates at high acidity (pH 0-0.6),
whereas the octamer is more prevalent in the 1.0-1.5 pH range.
Under the latter conditions, other species are practically
negligible. We should note that Zr speciation may be quite
different at low Zr concentrations. For example, Aplincourt et
-
5
-3
al. found that with 8 × 10 < [Zr] < 8 × 10 M, species
such as Zr(OH)3 , Zr2(OH)7 , and Zr(OH)4 may dominate.
Recent SAXS data strongly support polymeric species at
ZrOCl2] > 0.05 M.35 Unfortunately, there are no experimental
+
+
34
This analysis suggests that C-O scission is kinetically
preferred to P-O scission (given comparable leaving groups
at C and P). By implication, other dimeric or monomeric metal
complexes (cf., Th) should also be C-O selective. This is also
the case for trivalent Co, La, and Eu; see below.
[
-
3
-2
data covering the concentration interval 8 × 10 -5 × 10
M.
Chemoselectivity. The preceding discussion indicates that,
under the acidic conditions employed, dimeric or monomeric
Ce(IV) and Th(IV) species and tetrameric or octameric Zr(IV)
and (presumably) Hf(IV) species are the prevalent forms of the
metal cations present during phosphonoformate diester hydroly-
sis. We have seen that Ce(IV) and Th(IV) foster C-ester scission,
It is also possible to envisage cleavage of the bound
phosphonoformate by a bridging hydroxyl group, cf. 24.1
7a,42,43
(
41) This includes acceleration of C-OMe scission. Ce(IV)-mediated cleavage
of DMPF is 98 times faster than the analogous process for methyl acetate,
-6
-1
where kobs ) 5.3 × 10
s .
18
(42) Takasaki, B. K.; Chin, J. J. Am. Chem. Soc. 1995, 117, 8582.
(
whereas Zr(IV) and Hf(IV) favor P-ester cleavage. Previously,
we associated this chemoselectivity with dimeric Ce(IV) and
43) Moss, R. A.; McKernan, B. A.; Sauers, R. R. Tetrahedron Lett. 2002, 43,
5925.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004 10931

A R T I C L E S
Moss et al.
Chart 3. Equilibria for Zr(IV)a
a
b
Data and equations are taken from ref 27, Tables 8.3 and 8.4, and the discussion on pp 155-158. Calculated at I ) 0.5 M from equations in ref
2
7.
Chart 4. Additional Equilibria for Zr(IV)
occur at phosphorus via a six-membered transition state. The
tripodal phosphonate product (26) reproduces the binding
4
5
scheme of the lamellar zirconium phosphonates. Cleavage of
DMPF by Hf(IV), which also affords P-O selective cleavage,
2
0
must follow an analogous mechanism.
A. What Evidence Supports These Models? The mecha-
nism represented by 25 f 26 implies that alteration of the Zr-
(IV) octamers should impact the expressed P-O chemoselec-
tivity. Tris(2-hydroxymethyl)aminomethane (27) is reported to
form 1:1 complexes with Zr(IV) that are able to cleave bis-p-
nitrophenyl phosphate in acidic aqueous solution.46 Indeed,
addition of 1 or 2 equiv of 27 to the Zr reaction mixture of
Table 1 changes the 14/15 product distribution from ∼80:20 to
4
0:60 or 15:85, respectively; the P-O chemoselectivity shifts
to C-O selectivity. With Hf(IV), similar additions of 1 or 2
equiv of 27 change the 14/15 distribution from 90:10 (Table 1)
to 66:34 or 19:81, respectively.
Note that the interaction of 27 with Zr only occurs via the
tris-alcohol functionality; the amino group must be deuterated
under the reaction conditions. The addition of 27 raised the
solution pD by 0.2-0.6, so that we also examined the effect of
adding 1 equivalent of NaOD to the Zr(IV) or Hf(IV) solutions
Figure 3. Speciation of Zr(IV) as a function of pH, with [Zr(IV)] ) 0.025
-
M, I ) 0.5 (ClO4 ). The equilibrium concentrations of each Zr(IV) species
4
7
at each pH were calculated from the equilibrium constants in Charts 3 and
4
(taken from ref 27), using the iterative program SPE (see ref 32). A
description of this program appears in the text.
Again, C-attack corresponds to a (nominally) five-membered
ring transition state (or intermediate), while P-attack transits a
four-membered cycle. Examples of 24 and related constructs
with M ) La were examined computationally, and preferred
(2 equivalents of NaOD induced turbidity). Again, the 14/15
distributions changed, shifting to 50:50 (Zr) or 58:42 (Hf). The
addition of small quantities of NaOH to aqueous Zr(IV)
solutions is known to promote the formation of higher oligo-
43
C-ester attack was indicated.
For Zr(IV) we expect octamers 22 to be the dominant species
at pD 1.5-1.7;2
7,32,35c
cf. Figure 3. The octamers are most readily
(44) Mark, T. C. Can. J. Chem. 1968, 46, 3493.
(45) (a) Amicangelo, J. C.; Leestra, W. R. J. Am. Chem. Soc. 1998, 120, 6181.
represented as stacked hydroxyl-bridged tetramers,3 which also
occur as the unit cell in the crystal structure of ZrOCl2‚8H2O.
5c
(
b) Mitchell, M. C.; Kee, T. P. Coord. Chem. ReV. 1997, 158, 359 and
4
4
especially 362, 363. (c) Clearfield, A. Prog. Inorg. Chem. 1998, 47, 371.
46) Ott, R.; Kr a¨ mer, R. Angew. Chem., Int. Ed. 1998, 37, 1957.
47) Control experiments show that, in the absence of M(IV), added Tris does
not cleave DMPF at pD 2.2 over 72 h at 25 °C.
(
(
Binding of DMPF to the Zr(IV) octamer can be represented as
-
that in 25, where intracomplex OH cleavage can now readily
10932 J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004

Hydrolysis of Phosphonoformate Diesters
A R T I C L E S
not in straightforward accord with the kinetic data as expressed
in the correlations of log kobs vs log[M(IV)]. These determina-
tions were carried out in both H2O and D2O for the cleavages
of substrate 8 by Zr(IV) and substrate 9 by Th(IV). The slopes
of the linear correlations for 8 with Zr(IV) were 0.84 (H2O)
and 0.79 (D2O), and for 9 with Th(IV) they were 0.82 (H2O)
and 0.84 (D2O).
In principle, the dependence of log kobs on log[M(IV)] should
give the reaction order in the metal ion if the kinetics followed
the simple relation: kobs ) k2[M]total. However, the experimental
dependences give slopes that are <1, indicating a more complex
relation between kobs and [M(IV)]. At face value, these results
indicate that higher order complexes (dimers, tetramers) are not
the major catalytic species involved in these reactions. If they
were, we would expect an upward curvature in the kobs vs
meric Zr species.48 Thus, the P-O selectivity of Zr(IV) and
Hf(IV) toward DMPF appears to be a property of M(IV)
tetramers or octamers and can be shifted toward C-O selectivity
by modifying these complexes. We can also imagine that the
use of 1:1 acetonitrile/water as the solvent, necessary for
substrate 10, would perturb the speciation of Zr(IV) and alter
the P-O/C-O regiochemistry of DMPF cleavage (see above).
These speciation and binding models help rationalize the
opposing hydrolytic chemoselectivities expressed by Ce(IV) and
Th(IV) or Zr(IV) and Hf(IV) toward substrates 7-10. By
implication, other dimeric or monomeric metal complexes (or
simple nucelophilic reagents) should, like Ce(IV) or Th(IV),
be C-O selective. We have found that Eu(III) and La(III)
cations, and their bis-tris propane complexes, do indeed mediate
the hydrolysis of DMPF with C-OMe chemospecificity and
substantial rate enhancement.4 These lanthanide metal com-
plexes are most likely dimers.2
[
M(IV)] plot, and the log kobs vs [M(IV)] plot should give a
5
3
slope > 1.
3,49
C. Can One Model the Observed Kinetics with Higher
Order M(IV) Species? In an effort to link the possible
contributions of different metal species to the dependence of
kobs on [M(IV)], we evaluated the equilibrium concentrations
of Zr(IV) species within the concentration range studied (2.5-
25 mM), in water at pH 1.0 (see Table S1 in the Supporting
Information and Figure 3). According to the calculated equi-
librium concentrations, Zr tetramers and octamers are dominant
under these conditions and together account for 92-98% of
the total Zr(IV). If these species are the major contributors to
the catalysis, the observed rate constant could follow a
dependence like kobs ) k4[tetramer] + k8[octamer]. The equi-
librium concentrations of Zr(IV) tetramers and octamers as a
function of the total concentration of Zr(IV) (see Figure S5 in
the Supporting Information) do not curve upward, and the log-
log correlations of kobs for the cleavage of 8 with the concentra-
tions of Zr tetramers or octamers are linear, with slopes of 1.38
and 0.69, respectively (see Figure S6 in the Supporting
Information).
7,30,42,43
Similarly, 10 mM DMPF was hydrolyzed at pD 6.8 by 25
1
5,50
mM of the monomeric Co(III)cyclen complex 28
with kobs
-
4
-1
18,51
)
9.5 × 10
s
to yield only 15 by C-OMe cleavage.
-
7
Additionally, the slow cleavage of DMPF (kobs ) 7.6 × 10
-1
s
at pH 8.4) by cetyltrimethylammonium iodosobenzoate also
52
occurs with only C-OMe cleavage. However, the scission of
8
, as might be expected, occurs with 90:10 P-OPh/C-OMe
5
2
cleavage.
B. What Evidence Opposes These Models? Although the
foregoing models rationalize the M(IV)-specific chemoselec-
tivity and are consistent with the speciation studies, they are
Multiple regression analysis of the observed rate constants,
fit to the equation kobs ) k4[tetramer] + k8[octamer], afforded
second-order rate constants k4 ) 6.92 ( 1.4 and k8 ) 5.47 (
(
48) Clearfield, A. J. Mater. Res. 1990, 5, 161.
(
49) With substrate 8, where P-O scission is advantaged by a phenoxide leaving
4
3
group, these complexes afford P-O/C-O cleavage ratios of 1.1-1.3.
(
(
(
50) Chin, J.; Banaszcyzk, M.; Jubian, V.; Zou, X. J. Am. Chem. Soc. 1989,
-1 -1
2
0
.59 M s , with R ) 0.996. The theoretical fit of the
1
11, 186.
51) Analogous results with other Co complexes are presented in Morales-Rojas,
H. Ph.D. dissertation, Rutgers University, 2001.
experimental kobs using these parameters and the equilibrium
concentrations of tetramers and octamers is remarkably good,
as shown in Figure S7 (Supporting Information). The calculated
52) Moss, R. A.; Vijayaraghavan, S.; Kanamathareddy, S. Langmuir 2002, 18,
2
468.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004 10933

A R T I C L E S
Moss et al.
rate constants for the Zr(IV) tetramers and octamers are nearly
equal, indicating similar reactivity of each species toward
substrate 8, in agreement with the model proposed in 25 and
solutions. For deuterated aqueous solutions, pD was calculated accord-
ing to the following equation: pD ) pH meter reading + 0.40.
Melting points are uncorrected. Microanalyses were performed by
Quantitative Technologies, Inc., Whitehouse, NJ.
5
5
2
6.
The foregoing results indicate that the observed rate constant
Materials. Zirconium tetrachloride anhydrous (ZrCl
oxychloride hydrate (ZrOCl ‚8H O), hafnium tetrachloride anhydrous
HfCl ), ceric ammonium nitrate (Ce(NH (NO ), and thorium nitrate
hydrate (Th(NO ‚H O) were purchased from Strem Chemicals.
4
), zirconium
2
2
does not correlate with the concentration of a single Zr(IV)
species but follows a more complex dependence that can be
modeled in a linear fashion by the contributions of two higher
order Zr complexes as shown in Figure S7. This does not
exclude the intervention of still other Zr species, but it is of
interest that the model predicts the observed linear behavior of
kobs and [Zr(IV)]total. Linear correlations of kobs vs total metal
concentration have also been found in the cleavage of phosphate
diesters by dinuclear lanthanide species.54
(
4
)
4 2
3 6
)
3
)
4
2
Trimethyl phosphonoformate, phenyl chloroformate, trimethyl phos-
phite, trimethylsilyl bromide, phosphorus pentachloride, 1,8-diazabicyclo-
[5.4.0]undecene (DBU), and phenol were purchased from Aldrich
Chemical Co.. Phosphonoformate trisodium salt was purchased from
ICN Chemicals. 1,4,7-Triazacyclononane(3HCl) was purchased from
TCI America.
Deutereated solvents (D O, CDCl ) and MeOH (99% 13C) were
2
3
obtained from Cambridge Isotope Laboratories. H
5% of enrichment was obtained from ICON Laboratories. Anhydrous
18O with 43.85% or
2
8
Conclusions
THF and DMF were used as supplied from Aldrich. All other solvents
and salts were analytical grade unless otherwise stated.
Substrates. A. Dimethyl Phosphonoformate (7). To a solution of
trimethyl phosphonoformate (2.5 g, 14.9 mmol) in acetone (30 mL)
Ce(IV), Th(IV), Zr(IV), and Hf(IV) cations markedly ac-
celerate the hydrolyses of phosphonoformate diester substrates
7-10. Most importantly, our studies on the metal-mediated
was added a solution of sodium iodide (2.23 g, 14.9 mmol) in acetone
cleavages of these bifunctional esters illuminate a feature not
previously associated with “bare” tetravalent metal salts in
aqueous solutions: chemoselectiVe ester hydrolysis. Unprec-
edented chemoselectivity is displayed in these reactions, whereby
Zr and Hf direct scission at the P-O ester site, while Ce and
Th activate the C-O ester site. Leaving group efficacy
modulates the chemoselectivity: with substrate 9, C-OPh
cleavage dominates P-OMe cleavage even with Zr(IV) and Hf-
,56
(
30 mL).4 The reaction mixture was stirred and heated under reflux
for 2 h. Diester 7 was precipitated as a white solid. After filtration, the
solid was washed with cold acetone and dried under vacuum to give
57
2
.1 g of the sodium salt (80% yield); mp 179-181 °C (lit. 183 °C).
1
O): 3.71 (s, 3H), 3.58 (d, 3H, JP-H ) 11.1 Hz). 31P
O): -2.6 (s, H-decoupled; q H-coupled). Anal. Calcd for
H NMR (D
2
19
NMR (D
2
3 6 5
C H O
PNa: C, 20.5; H, 3.4. Found: C, 20.5; H, 3.2.
B. Methoxycarbonyl Phenylphosphonate (8). Following the method
of McKenna,58 trimethylsilylbromide (22.5 g, 14.7 mmol) was added
with stirring to trimethyl phosphonoformate (12.5 g, 7.14 mmol), under
(IV) catalysts, although the innate P-O selectivity of these
cations is sufficiently robust to impose some competitive
P-OMe cleavage, even in this biased case.
2
N , previously cooled in an ice water bath; see Scheme 1. After addition
was complete, the reaction mixture was allowed to warm to room
temperature. After removal of methyl bromide under vacuum, the
sample was purified by distillation under reduced pressure (54 °C, P
We consider a model in which the P-O selectivity of Zr-
(
IV) and Hf(IV) toward phosphonoformate diesters is linked to
the presence of octameric/tetrameric complexes of the cations
and can be shifted toward C-O selectivity by modification of
the complexes, whereas C-O esterolytic selectivity is associated
with dinuclear or mononuclear species of, e.g., Ce(IV), Th(IV),
Co(III), Eu(III), or La(III). Binding of the substrate (Lewis acid
activation) and reaction with a metal-coordinated water or
hydroxide nucleophile account for both the observed reactivity
and chemoselectivity. However, the kinetic dependences of the
metal-ion-mediated cleavages on [M(IV)]total are not in full
agreement with higher order metal complexes as the major
catalytic entities. Therefore, the proffered models should not
be regarded as established, but there is no gainsaying the
unprecedented, clearly expressed, chemoselective proclivities
of the several metal cations toward the phosphonoformate diester
substrates.
<
(
0.01 mmHg) to give 18.5 g of colorless liquid bis-trimethylsilyl-
methoxycarbonyl) phosphonate 11 (91%). 31P NMR (CDCl
): δ -25.1.
Next, 18.5 g (6.5 mmol) of this liquid was dissolved in 50 mL of
CH Cl and added dropwise to a solution of PCl (29 g, 13.1 mmol)
suspended in 50 mL of CH Cl under N , previously cooled to 0 °C.
3
2
2
5
2
2
2
The reaction mixture was stirred at room temperature for an additional
4 h. The volatiles were evaporated, and the final liquid was distilled
under reduced pressure (35 °C, P < 0.01 mmHg) to give 8 g of colorless
3
1
liquid methoxycarbonyl phosphonodichloridate 12 (69%); P NMR
(
CDCl
To a solution of methoxycarbonyl phosphonodichloridate (300 mg,
.75 mmol) in 15 mL of dry CH Cl under N , previously cooled to 0
3
): δ 9.8.
1
2
2
2
°
C, was added dropwise a solution of phenol (330 mg, 3.5 mmol) and
DBU (0.53 g, 3.5 mmol) in 15 mL of dry CH Cl . The mixture was
allowed to warm and stir for an additional 2 h. The organic phase was
extracted twice with aqueous NaHCO solution (0.06 g/50 mL) and
then washed with 50 mL of water. The organic phase was dried over
Na SO and evaporated under vacuum to give a liquid containing
methoxycarbonyl diphenylphosphonate 13 as the major component
68%, 31P NMR CDCl
, δ -15.3, s). The liquid was redissolved in 15
mL of CH CN and 20 mL of a NaHCO solution (0.088 g) was added.
2
2
3
Experimental Section
2
4
1
General Methods. Routine H NMR spectra were determined at
(
3
3
1
3
00 or 400 MHz. P NMR spectra were obtained at 121.4 or 161.9
3
3
1
3
MHz. C NMR spectra were obtained at 100.6 MHz. Chemical shifts
are reported as ppm (δ) relative to TMS or DSS for H, external H -
3
PO (85%) for P, or dioxane (67.4 ppm) for C. UV-vis kinetics
The mixture was stirred at room temperature for 3 h. Acetonitrile was
removed in a vacuum, and the aqueous solution was extracted with (2
1
3
1
13
4
×
30 mL) of CH
2 2
Cl . The aqueous phase was lyophilized to give solid
studies were obtained with a Hewlett-Packard 8453A diode array
spectrophotometer with thermostated cells. pH measurements were
made with an Orion 7110SC electrode, referenced with standard buffer
4
8
, 0.25 g (60% yield); mp 80-82 °C (lit. 79-81).
(
55) Glasoe, P. K.; Long, F. A. J. Phys. Chem. 1960, 64, 188.
56) Zervas, L.; Dilaris, I. J. Am. Chem. Soc. 1955, 77, 5354.
(
(
53) (a) Wahnon, D.; Hynes, R. C.; Chin, J. Chem. Commun. 1994, 1441. (b)
Jurek, P. E.; Jurek, A. M.; Martell, A. E. Inorg. Chem. 2000, 39, 1016.
54) Gomez-Tagle, P.; Yatsimirsky, A. K. Inorg. Chem. 2001, 40, 3786.
(57) U.S. Patent 4018854; Chem. Abstr. CAN 87: 135910.
(58) McKenna, C. E.; Higa, M. T.; Cheung, N. H.; McKenna, M.-C. Tetrahedron
Lett. 1977, 155.
(
10934 J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004

Hydrolysis of Phosphonoformate Diesters
1H NMR (D O): 7.09-7.33 (m, 5H), 3.74 (s, 3H). 31P NMR (D
A R T I C L E S
O): 3.49 (d, JP-H ) 10.8 Hz). 31P NMR (D
O): 3.3 (s,
1H NMR (D
H-decoupled).
2
2
O)
2
2
-
7.4 (s, H-decoupled). Anal. Calcd for C PNa: C, 40.3; H, 3.38.
8
H
8
O
5
Found: C, 40.2; H, 3.41.
C. Phenoxycarbonyl Methylphosphonate (9). Trimethyl phosphite
C. O-Phenyl disodium oxycarbonyl phosphonate (18) was pre-
pared by the hydrolysis of 8 with 1 equiv of NaOH solution.
Lyophilization of the aqueous solution gave a white solid; mp > 300
°C.
(10.5 g, 0.08 mol) was added dropwise over a period of 30 min with
2
stirring to phenylchloroformate (12.6 g, 0.08 mol), under N , and
previously cooled in an ice-water bath. After addition was complete
the mixture was allowed to warm and stir for an additional 3 h. Volatiles
were removed under vacuum to give 18 g (97%) of colorless liquid
phenoxycarbonyl dimethylphosphonate. No further purification was
1H NMR (D
2
O): 7.0-7.3 (m). 31P NMR (D O) 4.2.19
2
D. Disodium phenoxycarbonyl phosphonate (20) was prepared
from bis-trimethylsilyl phenoxycarbonyl phosphonate in a fashion
similar to that for 14; mp > 300 °C.
3
1
1
31
19
performed on this product. P NMR (CDCl
3
): δ -4.72 (s proton
decoupled; septet proton coupled). H NMR (CDCl ): δ 7.37 (t, 2H),
.24 (t, 1H), 7.12 (d, 2H), 3.98 (d, 3H, JP-H ) 11.4 Hz).
To a solution of phenoxycarbonyl dimethylphosphonate (3.45 g, 15
mmol) in 30 mL of acetone under N was added a solution of sodium
H NMR (D O): 7.09-7.41 (m). P NMR (D O): -3.2.
2
2
1
3
Kinetics. Data for substrates 7, 8, 9, 14, and 15 were obtained by
NMR monitoring of the corresponding released alcohol fragment,
MeOD. Kinetics data were also obtained from the disappearance of
the substrate if its signal was resolved enough to give accurate integrals.
For rapid kinetics, 40 µL of a 0.5 M substrate stock solution was added
7
2
4,56
iodide (2.25 g, 15 mmol) in 30 mL of acetone. The reaction mixture
was stirred and heated under reflux for 3 h. The solvent was removed
to give a sticky solid, which was dissolved in acetone. Upon addition
of ethyl ether, and after standing for 2 h, a white solid precipitated.
The solid was filtered and dried under vacuum to give the sodium salt
-
3
to 2 mL of a 0.025 M M(IV) solution containing 2.5 × 10 M of
pyrazine (integration standard) and 0.5 M NaClO . The pD of the
4
M(IV)-pyrazine-NaClO4 solution was established by the natural
hydrolysis of the metal salt under these conditions. No DCl or NaOD
was added unless otherwise noted. The vial was shaken to ensure
homogeneity, and 700 µL were quickly transferred to a 5 mm NMR
tube. After loading, equilibrating and tuning the NMR spectrometer at
25 °C (∼3 min), we acquired the spectra at preset time intervals. In a
typical experiment, the parameters were 1.6 s acquisition time, 1.0 s
delay, 5000 Hz sweep width, 16 K data block, and 32 accumulated
transients. For control experiments, the reaction mixtures were prepared
as described above. The pH value was adjusted with 1 N DCl. The
NMR tube was incubated in a water bath at 25 °C. At suitable time
intervals, the tube was removed and NMR spectra were collected.
of diester 9 (2.85 g, 80% yield); mp 220-222 °C.
1
H NMR (D
2
O): 7:12-7.43 (m, 5H), 3.72 (d, 3H, JP-H ) 10.8 Hz).
O): -3.26 (s, H-decoupled). Anal. Calcd for C
PNa: C, 40.3; H, 3.38. Found: C, 40.0; H, 3.30.
31
P NMR (D
2
8 8 5
H O
D. Diphenyl Phosphonoformate (10). Diphenyl phosphite (1.5 g,
5
.68 mmol) was dissolved in 75 mL of dry, distilled THF in a 250-mL
round-bottom flask, cooled in an ice-salt bath. Then, 0.32 g of a 60%
NaH dispersion in mineral oil (7.85 mmol) was added. The contents
were stirred at 0 °C for about 3 h. Then, 2.2 g (14 mmol) of
phenylchloroformate in 5 mL of THF was added dropwise to the flask.
After the addition, the contents were allowed to warm to room
temperature and stirred for about 24 h. THF was removed on a rotary
Kinetics were followed up to 5 half-lives, 25 to 40 points were taken,
and the endpoint (I ) was obtained after 10 half-times. The raw integrals
∞
evaporator. About 15 mL of cold H
decompose excess NaH, and the aqueous layer was extracted with 2 ×
0 mL of methylene chloride. The organic layer and the extract were
combined and dried over NaSO and then evaporated to give crude
triphenyl phosphonoformate triester (3.0 g) as a colorless liquid. The
crude ester (0.75 g) was not purified and was directly hydrolyzed with
2
O was added to the flask to
obtained in each spectrum were normalized by dividing by the integral
of the pyrazine standard. The pseudo-first-order rate constants were
4
calculated by plotting ln(I
ln(I ) vs time for the disappearance of substrates. Rate constants are
the averages of duplicate runs.
∞
- I
t
) vs time for the released species or
4
t
Kinetics for the cleavage of substrates 8-10, 18, and 20 were
monitored by UV spectroscopy following the appearance of the released
phenol at 260, 270, and 280 nm. In a typical experiment, 50 µL of a
12.5 mM stock solution of the substrate was added to 3 mL of 25 mM
1
50 mg of NaHCO
stirring at room temperature for about 4 h. (Longer times or an excess
of NaHCO led to further decomposition.) After the reaction, the
3 2 3
in a 5 mL of H O/5 mL of CH CN mixture with
3
acetonitrile and most of the water were removed by rotary evaporation,
and the crude residue was washed with ether. At this point, a white
solid product precipitated. It was filtered, dried, and column-chromato-
graphed after adsorption onto silica gel and eluted with 10% MeOH/
4 2 2
metal(IV) solution containing 0.5 M NaClO in D O (or H O). Clean
pseudo-first-order profiles were obtained over 5 half-lives. Rate
constants are the averages of duplicate runs.
Product Studies by 31P NMR. For these studies, 80 µL of a 0.5 M
stock solution of substrates 7-9, 14, 15, 18, or 20 were added to 4 mL
of a solution of 0.025 M M(IV) and 0.5 M NaClO in D O. After
9
0% EtOAc to give ∼100 mg of the purified sodium diphenyl
phosphonoformate (10) as a white solid, mp 366-368 °C.
4
2
1
31
O) - 8.23. 13C NMR
PNa‚
H NMR (D
O): 121, 122, 125, 128, 130. Anal. Calcd for C13
.5H O: C, 50.5; H, 3.26; P, 10.0. Found: C, 50.6, H, 3.61; P, 10.0.
Intermediates.59 A. Disodium Methoxycarbonyl Phosphonate
14). Bis-trimethylsilyl methoxycarbonyl phosphonate [11, see above
and Scheme 1] (2 g, 7 mmol) in 30 mL of ether was extracted with 2
30 mL of a solution of 1.14 g of NaHCO in 60 mL of water. Strong
CO evolution occurred and left a neutral aqueous solution from which
lyophilization gave white solid 14; mp > 300 °C.
2
O): 7.00-7.40 (m). P NMR (D
2
shaking to ensure homogeneity, each vial was incubated at 25 °C.
Aliquots (1 mL) were taken from the reaction mixture at different time
intervals, and 35 mg of EDTA disodium salt was added. A solid
precipitated right after the addition. After standing for 1-2 h at room
temperature, the mixture became homogeneous and colorless. For
substrates 8, 18, and 20, in the presence of Th(IV), the solutions were
quenched with 50 mg of tartrate disodium salt instead of EDTA.
Next, 700 µL of each of the above solutions (pH 3.5) were transferred
to 5-mm NMR tubes. After careful tuning and shimming of the NMR
spectrometer, the proton coupled and decoupled 31P NMR spectra were
3 4
acquired at 25 °C, referenced to external 85% H PO . In a typical case
the following parameters were used: 1.6 s acquisition time, 8000 Hz
sweep width centered at -1100 Hz, 32 K data block, 1.0 s recycle
delay, and 800 accumulated transients. Product identification was
performed by the additions of small aliquots of authentic materials.
For substrate 10, a 5-mL solution containing 25 mM metal ion [Zr-
(D
2
10 5
H O
0
2
(
×
3
2
1
31
2 2
H NMR (D O): 3.6 (s). P NMR (D O): -5.2.
B. O-Methyl disodium oxycarbonyl phosphonate (15) was pre-
pared by the hydrolysis of DMPF (7) with 1 equiv of NaOH solution.
Lyophilization of the aqueous solution gave a white solid; mp > 300
°
C.
(
59) Reaction intermediates are characterized spectroscopically; elemental
analyses were not attempted. Note that the chemical shifts are somewhat
(IV) or Th(IV)] and 0.5 M NaClO
4
, in 50% D
2 3
O-50% CD CN, was
4
+
different in the absence of M . However, the various “synthetic”
intermediates were identical to the corresponding “reaction” intermediates
in NMR spiking experiments.
prepared and thermostated at 25 °C. The solution was allowed to reach
its pH by natural hydrolysis. To this was then added 100 µL of a
J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004 10935

A R T I C L E S
Moss et al.
solution of 0.5 M substrate 10 (or 7) in D
2
O. Reactions were quenched
Research Office for financial support. H.M-R. acknowledges a
Fullbright-CONACyT Fellowship.
at preset time intervals by removing 700-µL aliquots from the reaction
mixture and adding ∼20 mg of sodium tartrate. The solution was then
transferred to a 5-mm NMR tube, and the 31P NMR spectrum was
Supporting Information Available: Figures of kobs vs [M(IV)]
and log kobs vs log[M(IV)] for substrate 8 with Zr(IV) and
substrate 9 with Th(IV); calculated equilibrium concentrations
and rate constants for Zr(IV) tetramers and octamers in the
cleavage of 8. This material is available free of charge via the
Internet at http://pubs.acs.org.
determined at 25 °C in the presence of 2mM O-methyl methyl
phosphonate (MMP) as an internal integration standard. In a typical
NMR experiment, about 1500 transients (scans) were performed, and
3
2 000 data points were collected, with an acquisition time of 1.6 s.
Acknowledgment. We thank a reviewer for helpful discus-
sions concerning the kinetics. We are grateful to the U.S. Army
JA030437H
10936 J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004