ABSTRACT
The diastereomers of adenosine and uridine 2
'
,3
'
- cyclic phosphorothioates were tested as substrates for 2
'
,3
'
-cyclic nucleotide 3
'
-phosphodiesterase from bovine brain. The enzyme cleaves the Sp (or exo)
diastereomers efficiently, whereas the Rp (or endo) diastereomers are resistant
to hydrolysis, even after long incubation. As the enzyme exhibits strong
substrate inhibition the precise determination of kinetic parameters posed
problems, particularly with phosphorothioates. The stereoselectivity of this
enzyme is opposite to that of RNase T
1
and RNase A and thus could be a useful complement in determination of the
configuration of nucleoside 2
'
,3
'
-cyclic phosphorothioates resulting from hydrolysis reactions of unknown
stereochemical course.
Nucleoside phosphate analogues, where a phosphate oxygen has been exchanged for
sulphur generating a chiral phosphorothioate, have found widespread application
in biochemistry and molecular biology (
1
). For example, comparison of the configuration of reactants and products where
the phosphate reaction centre has been replaced by a phosphorothioate gives
important information on the stereochemical course of reactions at phosphate
centres. The configuration of the reactants and products can be determined by various methods (
1
), some of which are based on enzymatic degradation, where an enzymes preference for a particular phosphorothioate stereoisomer is exploited. For example, pancreatic ribonuclease (RNase A) exhibits a preference for the Rp isomer of uridine 2',3'-cyclic phosphorothioate (2',3'-cUMPS) and cytidine 2',3'-cyclic
phosphorothioate and has been used to investigate the stereochemistry of
hammerhead ribozyme cleavage (
2
). However, the stereoselectivity of RNase A is not complete and the Sp isomer
is also hydrolysed, albeit at a much slower rate (
1
,
3
). Similarly, ribonuclease T
1
(RNase T
1
), which converts guanine 2',3'-cyclicphosphate (2',3'-cGMP) to guanine 3'-phosphate, also hydrolyses
the Rp isomer of guanine 2',3'-cyclic phosphorothioate, but in this case the Sp isomer
resists hydrolysis and acts as a competitive inhibitor (
4
) (Fig.
1
).
2',3'-Cyclic nucleotide-3'-phosphodiesterase (CNPase, EC 3.1.4.37) has been shown to hydrolyse nucleoside 2',3'-cyclic monophosphates
to their corresponding 2'-phosphates (
5
) with a preference for the purine nucleotides (
5
-
7
). Other phosphodiesters, such as 3',5'-cyclic nucleotides, dinucleotides and nucleoside 2'- or 3'-propyl or benzyl phosphate esters
are resistant to hydrolysis (
5
). The enzymes structure, properties and biological relevance have been reviewed
(
8
). Here we report the stereochemical preference of CNPase for hydrolysis of the
Sp isomer of nucleoside 2',3'-cyclic phosphorothioates to yield the 2'-phosphorothioates. The products from hydrolysis
of 2',3'-cUMPS have been confirmed by co-injection with authentic material. Under the
conditions used the Rp isomer is resistant to hydrolysis and thus CNPase
demonstrates a stereochemical preference which is opposite to that of RNase A
and T
1
.
Adenosine 2',3'-cyclic phosphate (2',3'-cAMP) and uridine 2',3'-cyclic phosphate (2',3'-cUMP) were obtained from Sigma Chemical Co. and used as supplied. The diastereomers of adenosine 2',3'-cyclic phosphorothioate (2',3'-cAMPS) and 2',3'-
cUMPS were prepared as described by Ludwig and Eckstein (
9
). 2',3'-Dideoxythymidine (ddT) and thymidine (dT) were supplied by
United States Biochemicals and Biomol Feinchemikalien GmbH (Ilvesheim, Germany), respectively. CNPase from bovine brain was supplied by
Sigma Chemical Co. as a 60% glycerol solution containing 50 mM MES and 100 mM
NaCl, pH 6.5. The glycerol solution contained 2 mg protein/ml with a specific activity of 31 U/mg protein (units based on hydrolysis of 2',3'-cNADP;
10
).
The CNPase catalysed hydrolysis of 2',3'-cAMP was initially investigated, in order to standardize the
enzyme solution and to give a comparison with the rate of hydrolysis of the
phosphorothioate analogue. The observed rate constants (
k
obs
) were determined for hydrolysis of 2',3'-cAMP and the dependence of the observed rate on the initial
substrate concentration, [
S
]
0
, is shown in Figure
2
. The plot shows that CNPase catalysed hydrolysis deviates from Michaelis-Menten kinetics at high substrate concentration, which is characteristic
of substrate inhibition. It was assumed that the observed rate constants determined at low substrate concentrations (0.5-2.0 mM) were relatively unaffected by inhibition and they were plotted on a Eadie-Hofstee plot giving a
K
m
of 0.9 mM and a
V
max
of 0.69 [mu]mol/min/U (pH 6.0, 37oC). The enzyme units used were those based on hydrolysis of 2',3'-cNADP (
10
). The values for the
K
m
and
V
max
were consistent with the values reported by other investigators (
7
,
10
-
13
) using similar conditions, for example Sogin (
10
) determined a
K
m
of 0.2 mM and a
V
max
of 1.1 [mu]mol/min/U (
V
max
has been converted to units based on the hydrolysis of 2',3'-cNADP).
CNPase is a unique RNase in that it only cleaves nucleoside 2',3'-cyclic phosphates and not the RNA internucleotide linkage,
like other RNases such as RNase A and RNase T
1
. We became interested in this enzyme as it offered the opportunity of analysing
the stereochemical course of either protein or ribozyme catalysed reactions
resulting in nucleoside 2',3'-cyclic phosphates or oligonucleotides terminating in such a
nucleotide. A precondition for such an application would be that the enzyme accept 2',3'-cyclic phosphorothioates as substrates and cleaves them
stereospecifically.
Various methods have been used to determine the kinetic parameters
K
m
and
V
max
of CNPase; these include a colorometric assay (
7
), spectroscopic and fluorometric assays (
10
,
12
) and an assay involving microHPLC (
11
). Only some of these methods could be adapted to study CNPase-mediated hydrolysis of 2',3'-cAMPS. For example, the colorometric assay mentioned above is based on detection of inorganic phosphate
produced by alkaline phosphatase catalysed hydrolysis of the 2'-monophosphate reaction product. 2'-Nucleoside monophosphorothioates, produced on
hydrolysis of 2',3'-cAMPS, are resistant to further hydrolysis by alkaline
phosphatase (
1
) and this renders the method unsuitable for our purposes. A HPLC-based assay was found to be most suitable, since 2',3'-cAMP and 2',3'-cAMPS (
9
) were readily available and are easily separated from their respective
hydrolysis products by reverse phase chromatography. CNPase-mediated hydrolysis of 2',3'-cAMP and the diastereomers of 2',3'-cAMPS was monitored by periodic
removal of aliquots of the reaction mixture and quantitating the amount of
substrate remaining by HPLC. The ratio of the remaining substrate to the
internal standard was determined and, in conjunction with the initial ratio of
substrate to standard, the amount of product produced at a given time could be
quantified. It was established that dT and ddT do not inhibit CNPase and were
suitable for use as internal standards (data not shown).
Determination of kinetic parameters for this enzyme proved to be difficult because of deviation from Michaelis-Menten behaviour. This was attributed to substrate inhibition. In mammals CNPase consists
of two protein components with slightly different shifts on SDS-PAGE (
14
-
17
), from which molecular weights of 46 and 48 kDa have been estimated. Both
components have been characterized and shown to have identical amino acid
sequences, except that the heavier component has an additional 20 amino acids
attached to the N-terminus (
14
). This dimeric compostion might well be the source of the substrate inhibition
phenomena observed (
18
,
19
). General kinetic schemes have been documented (
18
,
19
) which display variations of
k
obs
with substrate concentration similar to those observed with hydrolysis of 2',3'-cAMP and Sp-2',3'- cAMPS. As a result of substrate
inhibition, comparison of the efficiency of cleavage of the Sp diastereomer of
2',3'-cAMPS with that of cAMP was not possible in a precise
manner; comparison of the rates at 3 mM substrate concentration, where turnover
is highest, indicates that the efficiency is reduced ~5-6 fold. It is also interesting to note that no substate inhibition
was observed with 2',3'-cUMP. This is attributed to the high
K
m
value, since it is unlikely that the substrate concentration was high enough to
observe any inhibition phenomena.
It has been demonstrated that CNPase is able to specifically hydrolyse Sp-2',3'-cyclic phosphorothioates of both purine and pyrimidine
nucleosides, which is in sharp contrast to RNases A and T
1
, where the Rp diastereomer is preferred. It has also been established that
CNPase hydrolyses the terminal cyclic phosphate of an RNA without affecting the
internucleotide linkages(
5
). These two unique properties of CNPase thus offer an alternative method to
total digestion when determining the stereochemistry of intramolecular
displacement reactions at oligoribonucleotide phosphate centres. For example,
the products of hammerhead ribozyme catalysed reactions on substrates with a
phosphorothioate linkage at the cleavage site (
2
,
20
) could be characterized by further hydrolysis with CNPase. Oligoribonucleotide
products terminating in a 2'-phosphorothioate, which only result from hydrolysis of the Sp isomer of a 2',3'-cyclic phosphorothioate cleavage product, can
be differentiated by conventional or mercury-derivatized PAGE (
21
) and would thus indicate the configuration of the cyclic phosphorothioate
cleavage product from the ribozyme. Previously this characterization required
total digestion of the substrate and laborious co-injection of the 2',3'-cyclic phosphorothioate products with authentic
samples, which are not available commercially.
We are very grateful to S.Th.Sigurdsson and J.B.Thomson for critical reading of
the manucript.
REFERENCES
Return