Bisphosphonate derivatives of nucleoside antimetabolites: Hydrolytic stability and hydroxyapatite adsorption of 5′-β,γ-methylene and 5′-β,γ-(1-hydroxyethylidene) triphosphates of 5-fluorouridine and ara-cytidine

Article abstract of DOI:10.1021/jo800317e

(Chemical Equation Presented) Kinetics of the hydrolytic reactions of four bisphosphonate derivatives of nucleoside antimetabolites, viz., 5-fluorouridine 5′-β,γ-(1-hydroxyethylidene) triphosphate (4), 5-fluorouridine 5′-β,γ-methylene triphosphate (5), ara-cytidine 5′-β,γ-(1-hydroxyethylidene) triphosphate (6), and ara-cytidine 5′-β,γ-methylene triphosphate (7), have been studied over a wide pH range (pH 1.0-8.5) at 90°C. With each compound, the disappearance of the starting material was accompanied by formation of the corresponding nucleoside 5′-monophosphate, the reaction being up to 2 orders of magnitude faster with the β,γ-(1-hydroxyethylidene) derivatives (4, 6) than with their β,γ-methylene counterparts (5, 7). With compound 7, deamination of the cytosine base competed with the phosphate hydrolysis at pH 3-6. The measurements at 37°C (pH 7.4) in the absence and presence of divalent alkaline earth metal ions (Mg²⁺ and Ca²⁺) showed no sign of metal ion catalysis. Under these conditions, the initial product, nucleoside 5′-monophosphate, underwent rapid dephosphorylation to the corresponding nucleoside. Hydrolysis of the β,γ-methylene derivatives (5, 7) to the corresponding nucleoside 5′-monophosphates was markedly faster in mouse serum than in aqueous buffer (pH 7.4), the rate-acceleration being 5600- and 3150-fold with 5 and 7, respectively. In human serum, the accelerations were 800- and 450-fold compared to buffer. In striking contrast, the β,γ-(1-hydroxyethylidene) derivatives did not experience a similar decrease in hydrolytic stability. The stability in human serum was comparable to that in aqueous buffer (τ_1/2 = 17 and 33 h with 4 and 6, respectively), and on going to mouse serum, a 2- to 4-fold acceleration was observed. To elucidate the mineral-binding properties of 4-7, their retention on a hydroxyapatite column was studied and compared to that of zoledronate (1a) and nucleoside mono-, di-, and triphosphates.

Full text of DOI:10.1021/jo800317e

Bisphosphonate Derivatives of Nucleoside Antimetabolites:
Hydrolytic Stability and Hydroxyapatite Adsorption of
5′-ꢀ,γ-Methylene and 5′-ꢀ,γ-(1-Hydroxyethylidene) Triphosphates of
5-Fluorouridine and ara-Cytidine
Mikko Ora,† Tuomas Lo¨nnberg,† Diana Florea-Wang,† Shawn Zinnen,‡
Alexander Karpeisky,‡ and Harri Lo¨nnberg*,†
Department of Chemistry, UniVersity of Turku, FIN-20014 Turku, Finland, and MBC Pharma Inc.,
Aurora, Colorado 80047
harlon@utu.fi
ReceiVed February 7, 2008
Kinetics of the hydrolytic reactions of four bisphosphonate derivatives of nucleoside antimetabolites,
viz., 5-fluorouridine 5′-ꢀ,γ-(1-hydroxyethylidene) triphosphate (4), 5-fluorouridine 5′-ꢀ,γ-methylene
triphosphate (5), ara-cytidine 5′-ꢀ,γ-(1-hydroxyethylidene) triphosphate (6), and ara-cytidine 5′-ꢀ,γ-
methylene triphosphate (7), have been studied over a wide pH range (pH 1.0-8.5) at 90 °C. With each
compound, the disappearance of the starting material was accompanied by formation of the corresponding
nucleoside 5′-monophosphate, the reaction being up to 2 orders of magnitude faster with the ꢀ,γ-(1-
hydroxyethylidene) derivatives (4, 6) than with their ꢀ,γ-methylene counterparts (5, 7). With compound
7, deamination of the cytosine base competed with the phosphate hydrolysis at pH 3-6. The measurements
at 37 °C (pH 7.4) in the absence and presence of divalent alkaline earth metal ions (Mg2+ and Ca2+)
showed no sign of metal ion catalysis. Under these conditions, the initial product, nucleoside
5′-monophosphate, underwent rapid dephosphorylation to the corresponding nucleoside. Hydrolysis of
the ꢀ,γ-methylene derivatives (5, 7) to the corresponding nucleoside 5′-monophosphates was markedly
faster in mouse serum than in aqueous buffer (pH 7.4), the rate-acceleration being 5600- and 3150-fold
with 5 and 7, respectively. In human serum, the accelerations were 800- and 450-fold compared to buffer.
In striking contrast, the ꢀ,γ-(1-hydroxyethylidene) derivatives did not experience a similar decrease in
hydrolytic stability. The stability in human serum was comparable to that in aqueous buffer (τ1/2 ) 17
and 33 h with 4 and 6, respectively), and on going to mouse serum, a 2- to 4-fold acceleration was
observed. To elucidate the mineral-binding properties of 4-7, their retention on a hydroxyapatite column
was studied and compared to that of zoledronate (1a) and nucleoside mono-, di-, and triphosphates.
Introduction
osteonecrosis of the jaw.2b In addition to marked bone-seeking
and antiresorptive properties, experimental data suggest that
bisphosphonates have direct and indirect effects on tumor cells.
However, as compared to inhibition of bone resorption, cytotoxic
effects require much higher concentration (not practically
achievable clinically) with potentially greater systemic toxicity.3
The specific delivery of antineoplastic agents to bone tumor
sites, exploiting bone-seeking properties of bisphosphonates, has
been a target for many research projects over several decades.4
Previously, a number of research groups had designed, synthe-
sized, and tested bisphosphonate-antineoplastic drug conju-
Bisphosphonates, particularly zoledronate (1a) or pamidronate
(1b), are effective bone-specific palliative treatments that reduce
tumor-induced skeletal complications.1 However, there are
several limitations on the use of these drugs, which include poor
bioavailability (<1-10%), potential for renal toxicity,2a and
* To whom correspondence should be addressed. Phone: +358-2-333 6770.
Fax: +358-2-333 6776.
† University of Turku.
‡ MBC Pharma Inc.
10.1021/jo800317e CCC: $40.75
Published on Web 05/13/2008
2008 American Chemical Society
J. Org. Chem. 2008, 73, 4123–4130 4123

Ora et al.
gates.5 The cytotoxic drugs, methotrexate, melphalan, doxoru-
bicin, or cis-platinin, were directly covalently attached to the
terminal amino group of bisphosphonate, pamidronate (1b).
However, the anticancer properties of these conjugates tested
in certain animal models proved to be encouraging but far from
perfect and, in the case of doxorubicin, even negative.5b
Unfortunately, no data were presented on the stability of these
conjugates and the kinetics of accumulation in the bone. Thus,
it is possible that the covalent bond between the components
of the conjugates was either too stable or too unstable in vivo
to provide the desired concentrations of a cytotoxic compound
in bone. Most likely, intact, negatively charged conjugates with
a stable covalent bond would not be able to interact effectively
with intracellular molecular targets. On the other hand, an
unstable bond would cleave in the blood stream and could not
provide a sufficient increase of the active compound at bone
tissues. Recently it has been suggested that bisphosphonates
covalently linked via a phosphate group to cytotoxic agents,
such as structurally modified nucleosides provide an optimal
stability. This linkage allows the intact conjugate to survive in
the circulation, bind the mineral matrix of bone, and subse-
quently release both drugs in the bone microenvironment.6
Compounds designed in such a way will localize at the site of
tumor cell induced bone destruction and combine antiresorptive
and antitumor activities. These conjugates may become promis-
ing drugs that would increase the local concentration of the
cytotoxic agent, greatly improving efficacy without increasing
systemic toxicity of chemotherapeutic.
To learn more about the physicochemical properties and
serum stability of nucleoside bisphosphonate conjugates, a
quantitative kinetic investigation of the hydrolytic reactions of
nucleoside 5′-ꢀ,γ-methylene and 5′-ꢀ,γ-(1-hydroxyethylidene)
triphosphates has been carried out over a wide pH range at
elevate temperature and under physiological conditions in the
presence and absence of Mg2+ and Ca2+ ions. In addition, the
hydrolytic stability in human and mouse serum and the affinity
toward hydroxyapatite have been investigated. The compounds
studied include two conjugates of etidronate (2), viz. 5′-ꢀ,γ-
(1-hydroxyethylidene) triphosphates of 5-fluorouridine (4) and
ara-cytidine (6), and two conjugates of medronate (3), viz. 5′-
CHART 1
ꢀ,γ-methylene triphosphates of 5-fluorouridine (5) and ara-
cytidine (7) (Chart1).
Results
Product Distribution of the Hydrolysis of 4-7. First-order
rate constants for the hydrolysis of the bisphosphonate analogues
of 5-fluorouridine and ara-cytidine 5′-triphosphates (4-7) were
determined over a wide pH range (from pH 1.0 to 8.5, I ) 1.0
mol L-1 with NaCl) at 90 °C by analyzing the compositions of
the aliquots withdrawn at appropriate time intervals from the
reaction mixture by RP HPLC. Each reaction mixture contained
EDTA (2 mmol L-1) to suppress possible metal ion catalysis.
The products were characterized either by spiking with authentic
samples or by mass spectrometric analysis (HPLC/ESI-MS).
Over the pH range studied, disappearance of the starting material
was accompanied by formation of the corresponding nucleoside
5′-monophosphate, viz. 5-fluorouridine 5′-monophosphate (8)
from 4 and 5 and ara-cytidine 5′-monophosphate (9) from 6
and 7 (Scheme 1). With the ꢀ,γ-(1-hydroxyethylidene) ana-
logues (4, 6), the starting material was entirely converted to
the corresponding monophosphate (4 to 8 and 6 to 9) before
any other products appeared. For example, the first-order rate
constant for the hydrolysis of 6 to 9 was (1.14 ( 0.07) × 10-2
s-1 at pH 7.5, while the first-order rate constant for the
disappearance of 9 was (3.4 ( 0.2) × 10-7 s-1. The latter
reaction proceeded by two parallel routes, viz. dephosphorylation
to ara-cytidine (10, 15%) and deamination to ara-uridine 5′-
monophosphate (11, 85%). The methylene analogues (5 and 7)
did not, in turn, demonstrate quantitative accumulation of
monophosphate prior to the formation of secondary products.
Dephosphorylation to nucleoside, either 5-fluorouridine (12) or
ara-cytidine (10), was clearly visible at the late stage of the
hydrolysis, but the maximum time-dependent concentration of
the monophosphate still was more than 90%. The situation was
more complicated only under mildly acidic conditions (pH 3-6)
and only with 7. In addition to ara-cytidine 5′-monophosphate
(9), two other products accumulated, as shown by the time-
dependent product distribution in Figure 1. In parallel to
formation of 5′-monophosphate 9, the starting material was
deaminated to ara-uridine 5′-ꢀ,γ-methylenetriphosphate (13),
as indicated by the m/z [M - H]- 481.2 and the UV absorption
(1) (a) Coleman, R. E. Clin. Cancer Res. 2006, 12, 6243s–6249s. (b) Body,
J.-J. SupportiVe Care Cancer 2006, 14, 408–418. (c) Lipton, A.; Berenson, J. R.;
Body, J. J.; Boyce, B. F.; Bruland, O. S.; Carducci, M. A.; Cleeland, C. S.;
Clohisy, D. R.; Coleman, R. E.; Cook, R. J.; Guise, T. A.; Pearse, R. N.; Powles,
T. J.; Rogers, M. J.; Roodman, G. D.; Smith, M. R.; Suva, L. J.; Vessella, R. L.;
Weilbaecher, K. N.; King, L. Clin. Cancer Res. 2006, 12, 6209s–6212s.
(2) (a) Body, J. J.; Diel, I.; Bell, R. Semin. Oncol. 2004, 31, 73–78. (b)
Cameron, D. Semin. Oncol. 2004, 31, 79–82. (c) Picket, F. A. J. Dent. Hyg.
2006, 80, 10. (d) Sanna, G.; Preda, L.; Bruschini, R.; Rocca, M. C.; Ferretti, S.;
Adamoli, L.; Verri, E.; Franceschelli, L.; Goldhirsch, A.; Nole, F. Ann. Oncol.
2006, 17, 1512–16.
(3) (a) Daubine, F.; Le Gall, C.; Gasser, J.; Green, J.; Clezardin, P. J. Natl.
Cancer Inst. 2007, 99, 322–330. (b) Clezardin, P.; Ebetino, F. H.; Fournier,
P. G. Cancer Res. 2005, 65, 4971–4974. (c) Green, J. R. Oncologist 2004, 9,
3–13. (Suppl. 4)
(4) For reviews see: (a) Uludag, H. Curr. Pharm. Des. 2002, 8, 1929–1944.
(b) Hirabayashi, H.; Fujisaki, J. Clin. Pharmacokinet. 2003, 42, 1319–30.
(5) (a) Hosain, F.; Spencer, R. P.; Couthon, H. M.; Sturtz, G. L. J. Nucl.
Med. 1996, 37, 105–107. (b) Sturtz, G. A.; Appere, G. A.; Breistol, K. A.;
Fodstad, O. A.; Schwartsmann, G.; Hendriks, H. R. Eur. J. Med. Chem. 1992,
27, 825–833. (c) Sturtz, G. L.; Couthon, H. M.; Fabulet, O.; Mian, M. A.; Rosini,
S. Eur. J. Med. Chem. 1993, 28, 899–903. (d) Fabulet, O.; Sturtz, G. A.
Phosphorus Sulfur Silicon Relat. Elem. 1995, 101, 225–234.
(6) (a) Karpeisky, M. Y.; Padyukova, N. S.; Mikhailov, S. N.; Dixon, H. B. F.;
Tzeitline, G. US patent 6,214,812 B1. (MBC Research, Inc., Boulder, CO, 2001).
(b) Karpeisky, M. Y.; Padyukova, N. S.; Mikhailov, S. N.; Dixon, H. B. F.;
Tzeitline, G. US patent 6,750,340 B2(MBC Research, Inc., Boulder, CO, 2004).
(c) Reinholz, G. G.; Getz, B.; Sanders, E. S.; Karpeisky, M. Y.; Padyukova,
N. S.; Mikhailov, S. N.; Ingle, J. N.; Spelsberg, T. C. Breast Cancer Res. Treat.
2002, 71, 257–268.
4124 J. Org. Chem. Vol. 73, No. 11, 2008

Bisphosphonate DeriVatiVes of Nucleoside Antimetabolites
SCHEME 1
maximum 262 nm of the HPLC signal of this product at tR )
4.0 min (eluent 0.1 mol L-1 aq KH2PO4, flow rate 1.0 mL
min-1). This product, as well as the ara-cytidine 5′-monophos-
phate (9), was then slowly converted to ara-uridine 5′-
monophosphate (11), exhibiting tR ) 6.5 min, m/z [M + H]+
325.4, and UV absorption maximum 262 nm.
rate constants obtained at pH 4 at the buffer concentrations of
0.05 and 0.2 mol L-1 were identical within the limits of
experimental errors.
As seen, the ꢀ,γ-(1-hydroxyethylidene) derivatives, 4 and 6,
are hydrolyzed over the whole pH range studied, i.e., from pH
1.0 to 8.5, approximately as rapidly. Evidently the base moiety
does not participate in the hydrolysis of these compounds. The
reaction is first order in hydroxide ion concentration at pH >6
and the rate levels off to a constant value upon approaching
pH 8, which suggests that the starting materials undergo a
kinetically significant deprotonation around pH 7.5. The situation
pH-Rate Profiles for the Hydrolysis of 4-7. The pH-rate
profiles for the hydrolysis of 4-7 at 90 °C (I ) 1.0 mol L-1
)
are depicted in Figure 2 and the numerical values of the rate
constants are listed in Table 1. The rate constants were usually
obtained at only one buffer concentration, but the contribution
of buffer catalysis may in all likelihood be neglected, since the
FIGURE 2. pH-rate profiles for the hydrolysis of the bisphosphonate
analogues of nucleoside 5′-triphosphates (4-7) to nucleoside 5′-
monophosphates (8, 9). Notation: 5-fluorouridine 5′-ꢀ,γ-(1-hydroxy-
ethylidene) triphosphate (4, b), ara-cytidine 5′-ꢀ,γ-(1-hydroxyeth-
ylidene) triphosphate (6, O), 5-fluorouridine 5′-ꢀ,γ-methylene triphosphate
(5, ∆), ara-cytidine 5′-ꢀ,γ-methylene triphosphate (7, 2). Open squares
(0) refer to the deamination of 7. (T ) 90 °C, I ) 1.0 mol L-1 with
NaCl, [EDTA] ) 2 mmol L-1.)
FIGURE 1. Time-dependent product distribution for the reaction of
ara-cytidine 5′-ꢀ,γ-methylene triphosphate (7) at pH 4 and 90 °C.
Notation: (0) 7, (9) ara-cytidine 5′-monophosphate (9), (b) ara-uridine
5′-monophosphate (11), and (O) ara-uridine 5′-ꢀ,γ-methylene triph-
osphate (13). The pH was adjusted with HEPES buffer (0.05 mol L-1
)
and the ionic was maintained at 1.0 mol L-1 with sodium chloride and
the concentration of EDTA at 2 mmol L-1
.
J. Org. Chem. Vol. 73, No. 11, 2008 4125

Ora et al.
TABLE 1. First-Order Rate Constants for the Hydrolysis of 5-Fluorouridine 5′-ꢀ,γ-(1-Hydroxyethylidene) Triphosphate (4) and
5′-ꢀ,γ-Methylene Triphosphate (5) to 5-Fluorouridine 5′-Monophosphate (8) and the Hydrolysis of ara-Cytidine 5′-ꢀ,γ-(1-Hydroxyethylidene)
Triphosphate (6) and 5′-ꢀ,γ-Methylene Triphosphate (7) to ara-Cytidine 5′-Monophosphate (9) at 90 °C (I ) 1.0 mol L-1 with NaCl; [EDTA] )
2 mmol L-1
)
k/10-4 s-1
pH
solution
4
6
5
7
1.0
2.0
3.0
4.0
5.0
6.0
6.5
7.0
7.5
8.0
8.5
0.1 mol L-1 HCl
58.5 ( 1.3
33.5 ( 0.9
14.5 ( 0.6
11.40 ( 0.7
7.68 ( 0.06
11.3 ( 0.2
33.5 ( 1.1
67.2 ( 1.2
82.0 ( 3.0
128 ( 3
41.1 ( 1.8
31.2 ( 0.7
16.4 ( 3.0
9.32 ( 0.09
7.34 ( 0.08
11.6 ( 0.4
31.3 ( 1.1
70.9 ( 1.1
114 ( 7
4.65 ( 0.23
1.66 ( 0.03
0.279 ( 0.008
0.074 ( 0.002
0.043 ( 0.001
0.068 ( 0.002
3.45 ( 0.06
1.31 ( 0.02
0.01 mol L-1 HCl
0.05 mol L-1 (HCOOH/HCOONa)
0.05 mol L-1 (HOAc/NaOAc)
0.05 mol L-1 (HOAc/NaOAc)
0.05 mol L-1 (HEPES)
0.05 mol L-1 (HEPES)
0.05 mol L-1 (HEPES)
0.05 mol L-1 (Glycine)
0.05 mol L-1 (Glycine)
0.05 mol L-1 (Glycine)
1.35 ( 0.08a
1.09 ( 0.03a
0.248 ( 0.001a
0.115 ( 0.001a
0.330 ( 0.002
0.664 ( 0.032
0.755 ( 0.057
0.740 ( 0.029
0.400 ( 0.009
115 ( 4
0.763 ( 0.049
0.883 ( 0.043
111 ( 4
a At pH 3-6, the predominant reaction of 7 is deamination of the starting material to ara-uridine 5′-(ꢀ,γ-methylene) triphosphate (13). This reaction
is at pH 5 3.6 times as fast as the phosphate hydrolysis.
TABLE 2. First-Order Rate Constants for the Hydrolysis of
is less clear on going from pH 6 to more acidic solutions. The
5-Fluorouridine 5′-ꢀ,γ-(1-Hydroxyethylidene) Triphosphate (4)
hydrolysis becomes oxonium ion catalyzed at pH <5, but the
reaction order in oxonium ion concentration is clearly less than
unity. Even around pH 1, where the rate reaches a plateau value,
the hydrolysis still is slower than at pH 8. Evidently several
kinetically significant protolytic equilibria overlap in the pH
range 1-5.
and 5′-ꢀ,γ-Methylene Triphosphate (5) to 5-Fluorouridine
5′-Monophosphate (8) and the Hydrolysis of ara-Cytidine
5′-ꢀ,γ-(1-Hydroxyethylidene) Triphosphate (6) to ara-Cytidine
5′-Monophosphate (9) at Different Temperatures at pH 5.0 (0.05
mol L-1 AcOH/AcONa buffer; I ) 1.0 mol L-1 with NaCl)a
k/10-6 s-1
T/°C
4
6
5
7
The ꢀ,γ-methylene triphosphate analogues, 5 and 7, are
hydrolytically considerably more stable than their ꢀ,γ-(1-
hydroxyethylidene) counterparts discussed above. The stability
difference ranges from 1 order of magnitude to more than 2
orders of magnitude. As with the ꢀ,γ-(1-hydroxyethylidene)
analogues, the cleavage of 5 and 7 is clearly first order in
hydroxide ion concentration at pH 6-7 and reaches a plateau
value upon approaching pH 8. At pH <4, the hydrolysis of 5 is
approximately first order in oxonium ion concentration. With
7, the situation is more complicated. As discussed above, this
compound undergoes hydrolytic deamination of the cytosine
base. For this reason, the overall hydrolytic stability is at pH
3-6, much lower than that of 5, having a 5-fluorouracil base.
The phosphate hydrolysis, which occurs as a minor side reaction,
is first order in oxonium ion concentration (and slightly faster
than that of 5) at pH 3-5, turns pH independent on passing pH
3, and is again acid-catalyzed at pH <2.
50.0
60.0
70.0
80.0
90.0
13.9 ( 0.6
42.4 ( 0.9
129 ( 6
320 ( 10
768 ( 6
15.1 ( 0.7
46.1 ( 4.5
135 ( 7
327 ( 11
734 ( 8
0.188 ( 0.003
0.763 ( 0.018
2.32 ( 0.04
3.63 ( 0.02
7.19 ( 0.08
15.9 ( 0.1
24.8 ( 0.1
4.43 ( 0.15
a The rate constants given for ara-cytidine 5′-ꢀ,γ-methylene tri-
phosphate (7) refer to the disappearance of the starting material,
including contributions of deamination and phosphate hydrolysis.
reliable extrapolation to physiological temperature is not pos-
sible. At pH 7.4, the half-lives are 2400 and 1100 h for the
hydrolysis of 5 and 7, respectively.
The enthalpies and entropies of activation for the hydrolysis
of the ꢀ,γ-(1-hydroxyethylidene) phosphonates (4 and 6) are at
pH 5.0 quite similar: for 4 ∆H‡ ) (95 ( 2) kJ mol-1 and ∆S‡
) -(44 ( 5) J K-1 mol-1, and for 6 ∆H‡ ) (92 ( 2) kJ mol-1
and ∆S‡ ) -(53 ( 6) J K-1 mol-1. The markedly slower
hydrolysis of the ꢀ,γ-methylene phosphonates is of enthalpic
origin, the activation parameters obtained with 5 being: ∆H‡ )
(116 ( 4) kJ mol-1 and ∆S‡ ) -(25 ( 10) J K-1 mol-1. With
7, deamination predominates at pH 5 and, hence, activation
parameters for the phosphate hydrolysis cannot be reliably
determined.
18O Incorporation and Temperature Dependence. To
clarify whether the hydrolysis to nucleoside 5′-monophosphates
proceeds by the rupture of the P(R)-O or P(ꢀ)-O bond, 6 and
7 were hydrolyzed in 18O-enriched water at pH 8.0 (90 °C) and
the nucleoside monophosphate product, 9, was analyzed by
HPLC-ESI-MS. No incorporation of 18O into 9 was observed
in either case. Accordingly, the hydrolysis proceeds by cleavage
of the P(ꢀ)-O bond.
The dependence of the hydrolysis rate of 4-7 on temperature
was studied at pH 5, a pH where the reaction is almost pH
independent. Table 2 records the rate constants obtained at 10
°C intervals in the range from 50 to 90 °C with 4 and 6, and
from 60 to 90 °C with 5 and 7. Extrapolation to physiological
temperature by the Arrhenius equation indicates that the half-
lives of the hydrolysis of the ꢀ,γ-(1-hydroxyethylidene) phos-
phonates, 4 and 6, at 37 °C are 62 ( 5 and 50 ( 20 h,
respectively. For comparison, the half-lives for the same
compounds at pH 7.4, i.e. at the upper plateau, are 16 and 12 h
at 37 °C (I ) 0.14 mol L-1 with NaCl). The ꢀ,γ-methylene
phosphonates, 5 and 7, are hydrolyzed at pH 5 so slowly that
The Effect of Mg2+ and Ca2+on the Hydrolysis of 4-7.
The effect of Mg2+ and Ca2+ on the hydrolysis of 4-7 was
studied under modeled physiological conditions (pH 7.4; T )
37 °C; I ) 0.14 mol L-1 with NaCl). The time-dependent
product distribution under these conditions differs from that at
90 °C in the sense that the accumulation of the nucleoside 5′-
monophosphate (8 or 9) is not quantitative, but hydrolysis of
the starting material to nucleoside 5′-phosphate is followed by
dephosphorylation to nucleoside. While the monophosphate
accumulation still is almost quantitative with the ꢀ,γ-(1-
hydroxyethylidene) phosphonate derivatives (4, 6) (Figure 3),
the appearance of 5′-monophosphates during the hydrolysis of
the more stable ꢀ,γ-methylene analogues (5, 7) is barely
4126 J. Org. Chem. Vol. 73, No. 11, 2008

Bisphosphonate DeriVatiVes of Nucleoside Antimetabolites
TABLE 3. First-Order Rate Constants for the Cleavage of 4-7 in
the Presence of Mg2+ and Ca2+ Ions at 37 °C and pH 7.4 (I ) 0.14
mol L-1 with NaCl)
[Mg2+]/
[Ca2+]/
compd
mmol L-1
k/10-6 s-1
mmol L-1
k/10-6 s-1
4
none
0.1
1
12.4 ( 0.17
9.31 ( 0.088
5.08 ( 0.096
5.47 ( 0.043
15.7 ( 0.286
11.8 ( 0.170
7.40 ( 0.049
6.23 ( 0.066
0.09 ( 0.005
2.0
10.0
50.0
6.48 ( 0.061
3.68 ( 0.023
3.05 ( 0.026
2
6
5
7
none
0.1
1.0
2.0
none
0.1
1.0
2.0
none
0.1
1.0
2.0
2.0
10.0
50.0
7.60 ( 0.051
4.30 ( 0.038
3.57 ( 0.043
2.0
10.0
50.0
0.09 ( 0.004
0.12 ( 0.005
0.19 ( 0.005
0.08 ( 0.003
0.06 ( 0.001
0.20 ( 0.004
0.16 ( 0.004
0.11 ( 0.003
0.09 ( 0.002
2.0
10.0
50.0
0.11 ( 0.004
0.18 ( 0.007
0.14 ( 0.095
FIGURE 3. Time-dependent product distribution for the reaction of
5-fluorouridine 5′-ꢀ,γ-(1-hydroxyethylidene) triphosphate (4) at pH 7.4
and 37 °C in the presence of Mg2+ (0.1 mmol L-1). The pH was
adjusted with a HEPES buffer and the ionic was maintained at 0.14
mol L-1 with NaCl. Notation: 4 (b), 5-fluorouridine 5′-monophosphate
(8, O), and 5-fluorouridine (12, 0).
TABLE 4. Half-Lives for the Hydrolysis of 5-Fluorouridine
5′-ꢀ,γ-(1-Hydroxyethylidene) Triphosphate (4) and 5′-ꢀ,γ-Methylene
Triphosphate (5) to 5-Fluorouridine 5′-Monophosphate (8) and the
Hydrolysis of ara-Cytidine 5′-ꢀ,γ-(1-Hydroxyethylidene)
Triphosphate (6) and 5′-ꢀ,γ-Methylene Triphosphate (7) to
ara-Cytidine 5′-Monophosphate (9) in Human and Mouse Serum at
37 °C
t1/2/ h
compd
human serum
mouse serum
4
5
6
7
3.5
2.9
5.7
2.4
17.4
0.4
33
0.3
phorylation. Hydrolysis of the ꢀ,γ-methylene triphosphate
analogues (5, 7) is slightly retarded by Mg2+, but accelerated
by Ca2+. Acceleration of the cleavage by Ca2+ was accompanied
by diminution of the proportion of deamination to ara-uridine.
Kinetic Measurements in Human and Mouse Serum. The
hydrolytic stability of 4-7 was additionally determined in
human and mouse serum. In each case, the disappearance of
the starting material obeyed first-order kinetics. Table 4 records
the half-lives observed. The product distribution was quite
similar to that in aqueous buffers. In other words, the predomi-
FIGURE 4. Time-dependent product distribution for the reaction of
ara-cytidine 5′-ꢀ,γ-methylene triphosphate (7) at pH 7.4 and 37 °C in
the presence of Mg2+ (2.0 mmol L-1). The pH was adjusted with
HEPES buffer and the ionic was maintained at 0.14 mol L-1 with NaCl.
Notation: 7 (0), ara-cytidine 5′-monophosphate (9, 9), ara-uridine (14,
b), and ara-cytidine (10, O).
noticeable. Among the hydrolysis products of 5, 5-fluorouracil
was observed in addition to 5-fluorouridine (12) and its
monophosphate (8). In other words, 5-fluorouridine and possibly
also 5 itself underwent hydrolysis to 5- fluorouracil. Hydrolysis
of 7 gave ara-uridine as the main product (Figure 4). In addition,
ara-cytidine (10) and ara-cytidine 5′-monophosphate (9) were
formed. Deaminated starting material, the 5′-ꢀ,γ-methylene
triphosphate of ara-uridine (13), appeared as an intermediate,
but at a very low level (1-2%), and only traces of ara-uridine
5′-monophosphate (11) could be detected at any stage of the
reaction.
Table 3 summarizes the rate constants obtained for the
disappearance of 4-7 in the presence of Mg2+ (e2 mmol L-1
)
and Ca2+ (e50 mmol L-1) ions. With the ꢀ,γ-(1-hydroxyeth-
ylidene) triphosphate analogues (4, 6), hydrolysis to the 5′-
monophosphates (8 or 9) is somewhat retarded by both Mg2+
and Ca2+, and, hence, intermediary accumulation of these
compounds is diminished. Previously5 published results show
that Mg2+ and Ca2+ do not noticeably accelerate the dephos-
FIGURE 5. Time-dependent product distribution for the reaction of
ara-cytidine 5′-ꢀ,γ-(1-hydroxyethylidene) triphosphate (6) in human
serum at 37 °C. Notation: 6 (b), ara-cytidine 5′-monophosphate (9,
O), and ara-cytidine (10, 0).
J. Org. Chem. Vol. 73, No. 11, 2008 4127

Ora et al.
TABLE 5. Elution Times of Various Phosphate and Phosphonate
Compounds on a Hydroxyapatite Column (Bio-Scale CHT10-I, 12
mm × 88 mm) with Elution with a 1.5 mol L-1 and 0.5 mol L-1
Sodium Phosphate Buffer at a Flow Rate of 2 mL min-1
SCHEME 2
tR/min
1.5 mol L-1 buffer
0.5 mol L-1 buffer
compd
pH 5.8 pH 6.8 pH 7.8 pH 6.8 RP-columna
4
6
5
7
5.5
5.5
3.9
3.8
3.3
3.4
3.7
2.6
24.6
6.0
5.9
3.8
3.7
3.4
3.5
3.7
2.5
17.4
6.6
6.8
3.9
3.8
3.5
3.4
3.6
2.6
14.9
17.7
20.3
6.1
6.1
4.3
4.1
3.4
3.9
3.3
5.2
5′-UMP
5′-UDP
5′-UTP
oligodeoxynucleotide
zoledronate
2.8
3.6
12.7b
2.9
a Atlantis C18 column (4.6 × 250 mm, 5 µm), isocratic elution (1 mL
min-1
) , pH 6.0),
with potassium phosphate buffer (0.1 mol L-1
containing EDTA (2 mmol L-1). b Refers to gradient elution from aq
NH4OAc (50 mmol L-1) to a 1:1 mixture of aq NH4OAc (50 mmol
L-1) and MeCN in 30 min. The sequence is 5′-TAA GTA GCG AAC
ACC AAA GAT GAT AT-3′.
ion (reaction a in Scheme 2), (ii) attack of water on the
triphosphate tetraanion (reaction b), or (iii) attack of hydroxide
ion on the triphosphate trianion (reaction c).
A dissociative mechanism is usually utilized only by phos-
phomonoesters, i.e., when formation of a metaphosphate-like
intermediate, though preassociated with a molecule of water,
is possible.8 This is not possible upon cleavage of the P(ꢀ)-O
bond and, hence, mechanistic alternative a in Scheme 2 appears
less probable than alternative b or c, which may be either
associative reactions giving a phosphorane intermediate or
synchronous SN2-type displacements. Extensive studies with
3′,5′-dinucleoside monophosphates and their phosphotriester
counterparts have led to the conclusion that among the two
kinetically equivalent mechanisms, viz. attack of either 2′-OH
on deprotonated (monoanionic) phosphate or 2′-O- on proto-
nated (neutral) phosphate, the latter reaction is faster.9 Although
an analogous mechanism, viz. an attack of hydroxide ion on
protonated (neutral) ꢀ-phosphate (reaction c in Scheme 2),
cannot be strictly excluded, attack of water on anionic ꢀ-phos-
phate (reaction b in Scheme 2) appears, however, more attractive
in this particular case. The vicinal hydroxyethylidene group may
effectively hydrogen bond the negatively charged phosphate,
facilitating the nucleophilic attack. The marked rate-accelerating
effect of the carbon-bound hydroxy group may be attributed to
stabilization of the phosphorane intermediate by hydrogen
bonding to one of the oxygen atoms of the ꢀ-phosphate, as
depicted in Scheme 3. The breakdown of this intermediate then
proceeds by P(ꢀ)-O bond cleavage concerted with a water-
mediated proton transfer from the phosporane hydroxyl ligand
to the departing R-phosphate.
Two previous observations lend support to the significance
of the intramolecular hydrogen bond stabilization described in
Scheme 3. First, diribonucleoside-3′,3′-monophosphates undergo
transesterification to nucleoside 2′,3′-monophosphate by attack
of one of the 2′-OH groups on the phosphorus atom, followed
by departure of the other 3′-linked nucleoside.10 Methylation
of one of the 2′-OH groups retards the reaction by a factor of
23 (when the statistic correction by a factor of 2 has been taken
into account). Comparative measurements with tri(ribonucleo-
side)-3′,3′,5′-triphosphates has shown that the rate-accelerating
effect of the 2′-OH of the departing nucleoside may be attributed
to the stabilization of the intermediate rather than the leaving
group by hydrogen bonding (as depicted by structure a in
nant reaction was hydrolysis to nucleoside 5′-monophosphate.
The time-dependent product distribution obtained with ara-
cytidine 5′-ꢀ,γ-(1-hydroxyethylidene) triphosphate (6) in human
serum is shown in Figure 5 as an example. In mouse serum, a
small amount of an unidentified product was formed in addition
to the nucleoside and nucleoside 5′-monophosphate. This
product is formed mainly from nucleoside monophosphate and/
or nucleoside. The amount remained below 5%, except with 4,
which gave it up to 10%. Otherwise the product distribution
remained constant over the pH range studied. No depyrimidi-
nation was observed to take place.
Hydroxyapatite-Binding Affinities. The tendency of 4-7
to undergo adsorption to bone material was elucidated by
determination of their elution times on a hydroxyapatite column
(Bio-Scale CHT10-I, 12 mm × 88 mm). Compounds 4-7 were
eluted with a 1.5 mol L-1 sodium phosphate buffer at three
different pH values (pH 5.8, 6.8, and 7.8) at a flow rate of 2
mL min-1. To verify that no structural changes took place during
the elution, the eluted compounds were subjected to RP HPLC
by coinjection with the original sample. The results are listed
in Table 5.
Discussion
The phosphonate analogues, 4-7, of nucleoside 5′-triphos-
phates studied are decomposed both in aqueous buffers and in
human and mouse serum by departure of the pyrophosphonate
moiety, i.e., by cleavage of the bond between R- and ꢀ-phos-
phoric acid residues. Over the whole pH range studied, the ꢀ,γ-
(1-hydroxyethylidene) triphosphate analogues, 4 and 6, react
from 1 to more than 2 orders of magnitude faster than the ꢀ,γ-
methylene triphosphate analogues, 5 and 7. With both types of
compounds, the hydrolysis turns from a hydroxide ion-catalyzed
to a pH-independent reaction on passing pH 7. At pH >7, the
predominant ionic form of the triphosphate moiety is a tetraan-
ion. Since 18O is not incorporated into the nucleoside mono-
phosphate product when the hydrolysis is carried out in 18O-
enriched water, the solvent-derived nucleophile, either water
or the hydroxide ion, must attack the ꢀ-phosphorus atom.
Accordingly, three alternative mechanisms appear worth con-
sidering: (i) dissociative breakdown of the triphosphate tetraan-
(7) Kuusela, S.; Lo¨nnberg, H. J. Phys. Org. Chem. 1993, 6, 347–356.
4128 J. Org. Chem. Vol. 73, No. 11, 2008

Bisphosphonate DeriVatiVes of Nucleoside Antimetabolites
SCHEME 3
SCHEME 4
for the terminal monohydrogen phosphonate group in all
likelihood is around 7.13 Consistent with the preceding discus-
sion, this reaction most likely proceeds by the nucleophlic attack
of water (instead of hydroxide ion) on the ꢀ-phosphorus atom.
The values for the entropy of activation of the hydrolysis of
4-7 are negative and, hence, consistent with the proposed
associative nature of the reaction.
The decomposition of 4-7 becomes hydronium ion catalyzed
at pH <5, but the apparent reaction order in oxonium ion
concentration remains considerably below unity. Consistent with
the preceding discussion, the reaction may be assumed to
proceed, depending on pH, by an attack of water on the
ꢀ-phosphorus atom of the dianionic, monoanionic, and possibly
neutral phosphate/phosphonate moiety. The pKa values of the
terminal dihydrogen phosphonate moiety probably fall in the
range 2-3, while conversion of the monoanion and the neutral
forms takes place at pH <2.13,14 Owing to the existence of the
dianionic and monoanionic species in several tautomeric forms,
the observed rate constant for the hydrolysis cannot be attributed
to any single ionic species, but it refers to several parallel
reactions, the contributions of which vary with pH. This also
explains why the reaction order deviates markedly from unity.
Scheme 4 for the diester): the departure of both the 3′-and 5′-
linked nucleosides is accelerated as much.11 Second, the
internucleosidic 3′-O-P-CH(OH)-5′ linkage of a hydroxyphos-
phonate analogue of adenylyl-3′,5′-adenosine is isomerized to
the corresponding 2′,5′-linkage considerably faster than its deoxy
counterpart, 3′-O-P-CH2-5′.12 This acceleration has been at-
tributed to stabilization of the mono- and dianionic phosphorane
intermediates by intramolecular hydrogen bonding of the
hydroxyl group to a phorphorus-bound oxyanion (structure b
in Scheme 4). It is worth noting that this stabilization allows
hydroxide ion-catalyzed isomerization via a dianionic phospho-
rane, the reaction which is otherwise not encountered, owing
to the fact that a dianionic phosphorane is too unstable to
pseudorotate. Evidently, kinetically invisible proton transfer
between the carbon- and phosphorus-bound oxygens stabilizes
the dianionic structure to such an extent that pseudoration
becomes possible.
As mentioned above, ara-cytidine 5′-ꢀ,γ-methylene triphos-
phate (7) undergoes at pH 3-5 deamination of the cytosine base
in addition to departure of the bisphosphonate moiety. At pH
5, the deamination to uridine 5′-ꢀ,γ-methylene triphosphate (13)
is considerably faster than hydrolysis to ara-cytidine 5′-
monophosphate (9), the first order rate constants for the
deamination and the hydrolytic departure of the bisphosphonate
moiety being 2.5 × 10-5 and 6.9 × 10-6 s-1, respectively, at
90 °C. Under these conditions, the deamination of 7 most likely
proceeds by an attack of a molecule of water on C6 of the N3-
protonated cytosine moiety resulting in saturation of the 5,6-
double bond, which is followed by rapid hydrolysis of the
amidine moiety and dehydration to uracil base (Scheme 5).15
Participation of the 2′-OH by an intramolecular nucleophilic
attack on the C6 atom, suggested for the alkaline cleavage of
ara-cytidine,16 appears less since the 2′-O-methylation of ara-
cytidine has been shown to retard the deamination of ara-
cytidine only by a factor of 3.17 At least this participation is
less marked than with ara-uridine, with which the 2′-O-
methylation retards the breakdown by more than 2 orders of
magnitude.
At pH 7.4, Mg2+ and Ca2+ ions at physiological concentra-
tions retard the hydrolysis of the ꢀ,γ-(1-hydroxyethylidene) 5′-
triphosphate analogues (4, 6) by a factor of 2-4. Mg2+ also
retards the hydrolysis of the 5′-ꢀ,γ-methylene triphosphates (5,
7), while Ca2+ slightly accelerates the cleavage of these
compounds. Since Mg2+ and Ca2+ ions do not noticeably
accelerate the dephosphorylation,7 the intermediary accumula-
tion of the nucleoside 5′-monophosphate turns less quantitative,
except with 5 and 7 in the presence of Ca2+
.
As shown in Figure 1, the hydrolysis of 4-7 is pH
independent over a narrow pH range around pH 5, i.e., under
the conditions where the predominant ionic form is trianion,
the γ-phosphonate group being singly protonated. The pKa value
(8) (a) Grzyska, P. K.; Czyryca, P. G.; Purcell, J.; Hengge, A. C. J. Am.
Chem. Soc. 2003, 125, 13106–13111. (b) Wilkie, J.; Gani, D. J. Chem. Soc.,
Perkin Trans. 2 1996, 783–787. (c) Hengge, A. C.; Edens, W. A.; Elsing, H.
J. Am. Chem. Soc. 1994, 116, 5045–5049. (d) Thatcher, G. R.; Kluger, R. AdV.
Phys. Org. Chem. 1989, 25, 99–265. (e) Herschlag, D.; Jencks, W. P. J. Am.
Chem. Soc. 1989, 111, 7579–7586. (f) Friedman, J. M.; Freeman, S.; Knowles,
J. R. J. Am. Chem. Soc. 1988, 110, 1268–1275. (g) Buchwald, S. L.; Friedman,
J. M.; Knowles, J. R. J. Am. Chem. Soc. 1984, 106, 4911–4916. (h) Guthrie,
J. P. Acc. Chem. Res. 1983, 16, 122–129. (i) Westheimer, F. H. Chem. ReV.
1981, 81, 313–326. (j) Ramirez, F.; Marecek, J. F. Pure Appl. Chem. 1980, 52,
1021–1045. (k) Kirby, A. J.; Varvoglis, A. G. J. Am. Chem. Soc. 1967, 89, 415–
423.
(13) (a) Freedman, L. D.; Doak, G. O. Chem. ReV. 1957, 57, 479–523. (b)
Solinova, V.; Kasicka, V.; Koval, D.; Cesnek, M.; Holy, A. Electrophoresis
2006, 27, 1006–1019.
(14) (a) Sajadi, S. A. A.; Song, B.; Gregan, F.; Siegel, H. Inorg. Chem. 1999,
38, 439–448. (b) Siegel, H.; Hofstetter, F.; Martin, R. B.; Milburn, R. M.;
Scheller-Krattiger, V.; Scheller, K. H. J. Am. Chem. Soc. 1984, 106, 7935–
7946.
(9) (a) Mikkola, S.; Kosonen, M.; Lo¨nnberg, H. Curr. Org. Chem. 2002, 6,
523–538. (b) Kosonen, M.; Lo¨nnberg, H. J. Chem. Soc., Perkin Trans. 2 1995,
1203–1209. (c) Kosonen, M.; Oivanen, M.; Lo¨nnberg, H. J. Org. Chem. 1994,
59, 3704–3708.
(15) Kusmierek, J.; Ka¨ppi, R.; Neuvonen, K.; Shugar, D.; Lo¨nnberg, H. Acta
Chem. Scand. 1989, 43, 196–202.
(10) Kiviniemi, A.; Lo¨nnberg, T.; Ora, M. J. Am. Chem. Soc. 2004, 126,
11040–11045.
(16) Notari, R. E.; Chin, M. L.; Wittebort, R. J. Pharm. Sci. 1972, 61, 1189–
1195.
(11) Lo¨nnberg, T.; Korhonen, J. J. Am. Chem. Soc. 2005, 127, 7752–7758.
(12) Lo¨nnberg, T.; Kralikova, S.; Rosenberg, I.; Lo¨nnberg, H. Collect. Czech.
Chem. Commun. 2006, 71, 859–870.
(17) Lo¨nnberg, H.; Suokas, P.; Ka¨ppi, R.; Darzynkiewicz, E. Acta Chem.
Scand. 1986, B40, 798–805.
J. Org. Chem. Vol. 73, No. 11, 2008 4129

Ora et al.
SCHEME 5
known pKa values of the buffer acids under the experimental
conditions. Low buffer concentration was used (50 mmol L-1). The
initial substrate concentration was ca. 0.15 mmol L-1. The
composition of the samples withdrawn at appropriate intervals was
analyzed on an Aquasil C18 column (4 × 150 mm, 5 µm), using
a mixture of KH2PO4 (0.050 mol L-1) and 0.1% MeCN. The
observed retention times (tR/min) for starting materials were 1.7
(4), 1.7 (5), 1.5 (6), and 1.8 (7) and retention times (tR/min) for the
hydrolysis products were 3.1 (8), 2.6 (9), 3.1 (11), and 2.0 (13). In
the presence of metal ions and in serums, the decomposition of
starting material was followed by an Atlantis C18 column (4.6 ×
250 mm, 5 µm), using 0.1 mol L-1 KH2PO4 with 2 mM EDTA as
an eluent. Signals were recorded on a UV-detector at a wavelength
of 270 nm. The observed retention times (tR/min-1) for the products
on RP HPLC (flow rate was 1 mL min) were 4.0 (13), 4.7 (9), 6.4
(11), 6.5 (8), 9.2 (5-F-Ura), 17.5 (10), 29.0 (14), and 30.0 (12).
The observed retention times (tR/min) for starting materials were
6.3 (4), 3.8 (5), 3.3 (6), and 3.4 (7). The reaction products were
identified by the mass spectra (LC/MS). Aqueous ammonium
acetate (5 mmol L-1) was used as an eluent. The contribution of
the buffer catalysis to the observed rate constants was insignificant
even at the higher buffer concentration employed (0.2 mol L-1).
Calculation of the Rate Constants. The pseudo-first-order rate
constants for the decomposition of starting materials were obtained
by applying the integrated first-order rate equation to the time-
dependent diminution of the HPLC peak area of the starting
material.
Under the modeled physiological conditions (pH 7.4, T )
37 °C, [Mg2+] ) 2 mmol L-1, I ) 0.14 mol L-1 with NaCl),
ara-uridine (14) is obtain as the main product, while its potential
deaminated precursors, ara-uridine 5′-(ꢀ,γ-methylene) triphos-
phate (13) and 5′-monophosphate (11), appear only at low steady
state level (<2%) during a kinetic run. ara-Cytidine (10) is
formed as a minor product (25% of 14). The fact that 10 and
14 seem to be formed in parallel rather than consecutively
(Figure 4) suggests that 14 is formed via 13 and 11, and possibly
via 9 and 11, not via 9 and 10 (Scheme 1).
Compounds 4-7 are hydrolyzed in human and mouse serum
to nucleoside 5′-phosphates. The striking difference with hy-
drolysis in buffers is that the 5′-ꢀ,γ-methylene triphosphate
analogues are in mouse serum exceptionally labile: 5 and 7 are
hydrolyzed 5000 and 3000 times as fast as in buffer, respec-
tively, suggesting their ability to act as substrate for degradative
enzymes. In human serum, the acceleration is more modest:
700- and 400-fold with 5 and 7, respectively. In striking contrast
to the situation in buffers mimicking physiological conditions,
deamination does not detectably compete with the phosphate
hydrolysis. 5′-ꢀ,γ-(1-Hydroxyethylidene) triphosphates (4, 6),
which in physiological buffers are much more labile that their
ꢀ,γ-methylene triphosphate counterparts, do not exhibit a similar
increase in hydrolysis rate on going from buffers to serum. Their
stability in mouse serum is comparable to the stability in buffer
(pH 7.4) and a 2- to 4-fold increase in the hydrolysis rate takes
place on going to human serum. Accordingly, the 1-hydroxy-
ethylidene derivatives are in human serum slightly more stable
than their methylene counterparts and in mouse serum the
stability difference is 40- to 100-fold.
Mineral Binding. The mineral-binding properties of various
phosphonate and phosphate compounds were followed by HPLC
on a hydroxyapatite column (Bio-Scale CHT10-1, 12 mm × 88
mm); 1.5 mol L-1 potassium phosphate buffers (pH 5.8, 6.8, and
7.8) and 0.5 mol L-1 sodium phosphate (pH 6.8) were used as
eluents. The flow rate was 2 mL min-1. Compounds 4-7 were
detected by UV absorbance at 270 nm. Nucleoside mono-, di-, and
triphosphates and oligodeoxynucleotide were recorded at a wave-
length of 260 nm and zoledronate at a wavelength of 218 nm. The
Experimental Section
concentration of the eluted compounds was 0.3 mmol L-1
.
Materials. Preparation of compounds 4-7 used in kinetic
measurements has been described previously.6,18
Kinetic Measurements. The reactions were carried out in sealed
tubes immersed in a thermostated water bath (90.0 ( 0.1 or 37.0
( 0.1 °C). The oxonium ion concentration of the reaction solutions
was adjusted with hydrogen chloride, sodium hydroxide and
formate, acetate, N-[2-hydroxyethyl]piperazine-N-[2-ethanesulfonic
acid] (HEPES), and glycine buffers. The oxonium ion concentra-
tions of the buffer solutions were calculated with the aid of the
Acknowledgment. The authors thank Evelina Laitinen,
M.Sc., and Heidi Isotalo for their skillful help in performing
the kinetic measurements and Dr. Sergei Mikhailov and Dr.
Nelly Padyukowa (Engelhardt Institute of Molecular Biology,
Russian Academy of Sciences, Moscow, Russia) for the
preparation of compounds 4 and 5 and Dr. Andrei Guzaev (AM
Chemicals, Oceanside, CA, USA) for the preparation of
compounds 6 and 7.
(18) Andreeva, O. I.; Efimtseva, E. V.; Padyukova, N. S.; Kochetkov, S. N.;
Mikhailov, S. N.; Dixon, H. B. F.; Karpeisky, M. Y. Mol. Biol. 2001, 35, 717–
729.
JO800317E
4130 J. Org. Chem. Vol. 73, No. 11, 2008