2',3'-cAMP hydrolysis by metal-dependent phosphodiesterases containing DHH, EAL, and HD domains is non-specific: Implications for PDE screening

Article abstract of DOI:10.1016/j.bbrc.2010.06.107

The recent report of 2',3'-cAMP isolated from rat kidney is the first proof of its biological existence, which revived interest in this mysterious molecule. 2',3'-cAMP serves as an extracellular adenosine source, but how it is degraded remains unclear. He

Full text of DOI:10.1016/j.bbrc.2010.06.107

Biochemical and Biophysical Research Communications 398 (2010) 500–505
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications
journal homepage: www.elsevier.com/locate/ybbrc
0
0
2
,3 -cAMP hydrolysis bymetal-dependent phosphodiesterases containing DHH, EAL,
and HD domains is non-specific: Implications for PDE screening
*
Feng Rao , Yaning Qi, Elavazhagan Murugan, Swathi Pasunooti, Qiang Ji
School of Biological Sciences, Nanyang Technological University, Singapore 637551, Singapore
a r t i c l e i n f o
a b s t r a c t
0
0
Article history:
Received 11 June 2010
Available online 1 July 2010
The recent report of 2 ,3 -cAMP isolated from rat kidney is the first proof of its biological existence, which
0
0
revived interest in this mysterious molecule. 2 ,3 -cAMP serves as an extracellular adenosine source, but
how it is degraded remains unclear. Here, we report that 2 ,3 -cAMP can be hydrolyzed by six phospho-
diesterases containing three different families of hydrolytic domains, generating invariably 3 -AMP but
not 2 -AMP. The catalytic efficiency (kcat/K
against the widely used non-specific substrate bis(p-nitrophenyl)phosphate (bis-pNPP), indicating that
,3 -cAMP is a previously unknown non-specific substrate for PDEs. Furthermore, we show that the
exclusive formation of 3 -AMP is due to the P–O2 bond having lower activation energy and is not the
result of steric exclusion at enzyme active site. Our analysis provides mechanistic basis to dissect protein
function when 2 ,3 -cAMP hydrolysis is observed.
0
0
0
Keywords:
0
0
0
m
) of each enzyme against 2 ,3 -cAMP correlates with that
0
0
2
,3 -cAMP
Phosphodiesterase
Hydrolysis
Non-specific
0
0
2
0
0
0
0
Ó 2010 Elsevier Inc. All rights reserved.
0
0
0
1
. Introduction
families have been reported to have 2 ,3 -cyclic-nucleotide 2 -phos-
0
0
0
phodiesterase activity [6–9], 2 ,3 -cyclic-nucleotide 3 -phosphodi-
esterase activity has been confined to the 2H phosphoesterase
family that includes CNPase [10]. What governs this differential
specificity remains a mystery for decades.
Earlier reports suggested that the 2 ,3 -cyclic ribose structure is
unstable due to the strain of the five-membered ring [11]. It was
0
0
It is well-known that RNA with 2 ,3 -cyclic phosphodiester end
exists as an common intermediate after cleavage by RNases, like
RNase A [1]. Meanwhile, the existence of small 2 ,3 -cyclic mono-
nucleotides in biological system has remained mysterious for a
long time till the recent discovery of 2 ,3 -cAMP being released
from rat kidney [2]. The released 2 ,3 -cAMP is further broken
down into 2 -AMP and 3 -AMP, which can serve as sources of extra-
cellular adenosine [3]. How 2 ,3 -cAMP is synthesized or degraded
remains unknown. However, reports of phosphodiesterases (PDEs)
that can hydrolyze 2 ,3 -cAMP are abundant. A family of metal-
independent 2 ,3 -cyclic-nucleotide 3 -phosphodiesterase (CNPase,
EC 3.1.4.37) that hydrolyzes 2 ,3 -cAMP into 2 -AMP was discov-
ered in brain since 1960s [4]. Though the physiological substrate
of CNPase remains unclear, recent evidence suggests that it can
function as a tRNA splicing enzyme in vivo [5]. On the other hand,
there are numerous reports of 2 ,3 -cyclic-nucleotide 2 -phospho-
diesterases that generate 3 -AMP (EC 3.1.4.16, see http://www.
brenda-enzymes.org/php/result_flat.php4?ecno=3.1.4.16). These
reports were based on in vitro enzymatic assays using synthetic
substrates; their physiological relevance therefore remains
obscure. In general, while PDEs belonging to diverse domain
0
0
0
0
0
0
0
0
II
also shown that a small Cu complex, through an activated hydrox-
0
0
0
0
ide group bound by the metal, could efficiently hydrolyze 2 ,3 -
cAMP [12]. To date, at the enzyme level, it remains unknown
whether PDEs and other metallo-hydrolases, whose active sites
are generally endowed with specificity towards cognate substrate,
can also non-specifically hydrolyze 2 ,3 -cyclic nucleotides like
2 ,3 -cAMP. This would have broad implications regarding the
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
physiological relevance of the known 2 ,3 -cyclic nucleotide phos-
0
0
phodiesterases, as well as the recent approach of using 2 ,3 -cyclic
nucleotides as part of a natural substrate collection to screen cog-
nate PDEs out of proteins with unknown functions but known
structures [13]. More and more such proteins are made available
by efforts of the Structural Genomics Initiatives [6,7,14], which
calls for an urgent and critical assessment of the mechanism of
0
0
0
0
0
0
2 ,3 -cyclic nucleotide hydrolysis.
0
0
Here, we compare the hydrolytic pattern of 2 ,3 -cAMP and bis-
pNPP, a widely used non-specific PDE substrate, by six phosphohy-
drolases (Fig. 1A), five of which have been previously characterized
as metal-dependant PDEs with known substrates (Table 1). Our re-
sults demonstrate that hydrolysis of 2 ,3 -cAMP into 3 -AMP by
these enzymes is non-specific and provide mechanistic insight.
Our results and analysis also suggest that, depending on products
Abbreviations: bis-pNPP, bis(p-nitrophenyl)phosphate; AP, alkaline phospha-
0
0
0
tase; CNPase, 2 ,3 -cyclic-nucleotide 3 -phosphodiesterase; PDE, phosphodiester-
ase; PPT, phosphopantetheinyl; HPLC, high performance liquid chromatography.
0
0
0
*
Corresponding author. Fax: +65 67913856.
E-mail addresses: RAOF0001@ntu.edu.sg, raofeng@gmail.com (F. Rao).
0
006-291X/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.bbrc.2010.06.107

F. Rao et al. / Biochemical and Biophysical Research Communications 398 (2010) 500–505
501
(
DE3). Subsequent expression and purification procedures were
the same as used for YybT [15], except a new lysis buffer was used
50 mM Tris (pH 8.0), 150 mM NaCl, 5% Glycerol, 1 mM phenyl-
[
methylsulfonyl fluoride, and 0.5 mM ethylenediaminetetraacetic
acid].
0
0
2
.3. PDE activity assay against 2 ,3 -cyclic nucleotides
For RocR and DGC2, enzymatic reaction condition used was:
1
2
00 mM Tris–HCl (pH 8.0), 20 mM KCl and 25 mM MgCl . For YybT,
YtqI, 3DMA, and PaAcpH, reaction condition was the same as de-
scribed for assaying the c-di-AMP hydrolysis by YybT [15]:
1
00 mM Tris–HCl (pH 8.3), 20 mM KCl, 0.5 mM MnCl
2
. Reactions
were stopped by adding 1/10 volume of 0.5 M EDTA and the pro-
0
0
gress of 2 ,3 -cAMP hydrolysis was monitored using an HPLC
system with reverse phase XDB-C18 column (4.6 ꢀ 150 mm) [mo-
bile phase: 20 mM triethylammonium bicarbonate (pH 7.0, pH ad-
justed with acetic acid), 9% methanol, 0.9 mL/min], as was adapted
from Ref. [16]. Initial velocity at a certain substrate concentration
was obtained from a series of reactions with varying incubation
m
time. The kinetic parameters kcat and K were obtained by fitting
the initial velocities at various substrate concentrations to the
Michaelis–Menten equation using the software Prism (GraphPad).
0
0
HPLC condition used was the same for 2 ,3 -cGMP, but slightly
0
0
modified (methanol concentration changed to 6%) for 2 ,3 -cCMP
0
0
and 2 ,3 -cUMP.
2
.4. Activity assay against bis-pNPP and pNPP
Fig. 1. (A) Domain composition of the six enzymes. (B) Characterization of YtqI and
DMA by size exclusion chromatography. Inset: SDS–PAGE gel of purified proteins.
3
Enzymatic reaction conditions were the same as those used for
,3 -cAMP. Hydrolysis was followed using a Shimadzu UV-1700
0
0
2
0
0
generated, 2 ,3 -cAMP can be used as a probe to screen for enzymes
spetrophotomer to monitor p-nitrophenol formation at 410 nm.
The extinction coefficient used to calculate p-nitrophenol concen-
tration was 18,300 M cm . A control with the enzyme storage
buffer only was used for baseline correction. Due to the solubility
issues of bis-pNPP, high substrate concentration could not be
0
0
that work on 2 ,3 -cyclic phosphodiester structure under physio-
logical context or merely PDEs in general.
ꢁ1
ꢁ1
2
. Materials and methods
reached to accurately determine the kcat and K
m
for all enzymes.
Catalytic efficiency was calculated as 1/slope(double-reciprocal plot)
.
2.1. Materials
0
0
2.5. Computational modeling
The 2 ,3 -cyclic nucleotide substrates, their respective linear
nucleotide standards, bis-pNPP and pNPP are from sigma. Calf
intestine aphosphatase (AP) was from Roche and was used
according to manufacturer’s instruction. Custom synthesized YtqI
and 3DMA genes were obtained from GenScript Corporation (NJ,
USA).
0
0
The modeling of 2 ,3 -cyclic nucleotide (taken from PDE id 1JH7)
into of the EAL domain PDE BlrP1 (PDB id 3GG0) and the CNPase
(PDB id 1WOJ) was done in Pymol (DeLano Scientific). The scissile
phosphate groups were aligned by pair-fitting 1: the phosphorus
atom under attack; 2: the leaving oxygen atom.
2.2. Protein cloning, expression and purification
2.6. Statistical analysis
The genes encoding 3DMA and YtqI proteins were cloned into
PET28(a+) (Novagen) between the NdeI and XhoI restriction sites,
resulting in two N-terminal His -tagged recombinant constructs.
Correlation analysis was performed using the software Prism
(GraphPad). Nonparametric correlation (spearman) with one-
6
The plasmids were transformed into Escherichia coli strain BL21
tailed p value was determined.
Table 1
0
0
Steady-state kinetic parameters for the six proteins against 2 ,3 -cAMP and bis-pNPP.
0
0
ꢁ1
ꢁ1
0
0
Enzyme
PDE fold
Cognate substrate
Ref.
2 ,3 -cAMP
cat (minꢁ1)
bis-pNPP kcat/K
m
(min mM
)
3 ,5 -cAMP hydrolysis
ꢁ1
ꢁ1
k
K
m
(mM)
m
kcat/K (min mM )
YybT
YtqI
DHH
DHH
DHH
EAL
EAL
HD
c-di-AMP
nano-RNA
–
c-di-GMP
c-di-GMP
PPT arm
[17]
[32]
–
[18]
[15]
[20]
250 ± 31
610 ± 30
15 000 ± 300
2.6 ± 0.5
0.13 ± 0.01
0.21 ± 0.02
15.8 ± 6.2
3.5 ± 0.58
0.85 ± 0.07
0.05 ± 0.02
8.7 ± 1.0
15.8 ± 6.5
174 ± 30
0.55 ± 6.5
110 ± 10
670 ± 20
0.51 ± 0.03
0.004 ± 0.0005
0.03 ± 0.01
No
No
No
No
No
Residual
a
4
3
DMA
(1.7 ± 0.1) ꢀ 10
DGC2
RocR
PaAcpH
52 ± 14
0.015 ± 0.002
0.21 ± 0.06
1.0 ± 0.3
a
The 3DMA protein has not been characterized but its structure was solved by Northeast Structural Genomics Consortium (PDB id: 3DMA). We speculate it is an YtqI
homolog.

5
02
F. Rao et al. / Biochemical and Biophysical Research Communications 398 (2010) 500–505
0
0
0
3
. Results and discussion
3.1. Kinetic evidence that hydrolysis of 2 ,3 -cAMP into 3 -AMP is non-
specific
Among the six proteins we characterized (Fig. 1A and Table 1),
0
0
YybT, YtqI and 3DMA contain the DHH domain [17]. YybT and YtqI
are PDEs for cyclic-di-AMP and nano-RNA, respectively [15,18].
The 3DMA protein was uncharacterized, but its structure was
solved by Northeast Structural Genomics Consortium (PDB id:
We initially observed hydrolysis of 2 ,3 -cAMP in our screening
for potential substrate of YybT (Fig. 2A, B, and E). The end product
0
0
was exclusively 3 -AMP. No hydrolysis was observed when 3 -AMP
or 2 -AMP was used as substrate, in agreement with the lack of
0
3
DMA). We speculate that it is an YtqI homolog. RocR and DGC2
reactivity against phosphatase substrate described previously
0
0
contain the EAL domains that specifically hydrolyze cyclic-di-
GMP [19,20]. PaAcpH is a HD domain PDE [21] that hydrolyzes
the phosphodiester bond between acyl carrier protein and the
phosphopantetheinyl (PPT) arm [22]. Structural data showed that
DHH, EAL, and HD domain proteins are all metalloenzymes that
require divalent metal ion to assist hydrolysis, at least in part to
activate the attacking water nucleophile [23–25].
For YybT, RocR, DGC2, and PaAcpH, purified recombinant pro-
teins were obtained according to published procedures described
before [15,19,20,22]. For YtqI and 3DMA, dimeric proteins were ob-
tained as described in material and methods and used for enzy-
matic assays (Fig. 1B). All six proteins displayed bis-pNPP
hydrolysis activity with the assistance of divalent metal ions, in
agreement with their PDE function, but their catalytic efficiency
[15]. Since the cognate substrate of YybT is c-di-AMP and 2 ,3 -cyc-
lic nucleotides are prone to hydrolysis [11,12], we tested the
0
0
hypothesis that 2 ,3 -cAMP is a non-specific substrate for PDEs
0
0
and found that all six proteins actively hydrolyzed 2 ,3 -AMP into
0
2+
2+
3 -AMP in the presence of Mg or Mn ions. We believe that the
0
0
PDE active sites are responsible for 2 ,3 -cAMP hydrolysis, as we
found that, for multi-domain proteins RocR and YybT, mutations
of catalytic residues (E352A for RocR and D420A for YybT) that
were crucial for their PDE activity [15,20] diminished the hydroly-
sis of 2 ,3 -cAMP to below 0.1%. We also observed significant
hydrolysis of 2 ,3 -cGMP, 2 ,3 -cCMP, and 2 ,3 -cUMP by YybT and
0
0
0
0
0
0
0
0
0
the other enzymes, generating nucleotide 3 -phosphates as prod-
ucts. The less than 10-fold difference in rate (Fig. 2F) suggested
0
0
that hydrolysis was due to the instability of 2 ,3 -cyclic ribose
structure and was not base-specific. Indeed, the natural substrate
5
varied up to 10 fold (Table 1).
0
0
0
0
0
0
Fig. 2. (A) HPLC trace of 2 ,3 -cAMP hydrolysis by AP, YybT and a combination of AP and YybT. The standard retention times for 3 -AMP, 2 -AMP, Adenosine, and 2 ,3 -cAMP
0
0
are: 6.0 min, 9.4 min, 10.1 min, and 11.2 min. (D) Cross-comparison of the catalytic efficiency of the six proteins against 2 ,3 -cAMP (x-axis) and bis-pNPP (y-axis). (E) Scheme
0
0
of 2 ,3 -cAMP hydrolysis.

F. Rao et al. / Biochemical and Biophysical Research Communications 398 (2010) 500–505
503
0
0
for PaAcpH bares no structural resemblance to 2 ,3 -cAMP other
than the phosphodiester moiety [22], suggesting that interaction
PDEs belonging to three different folds all sterically hindered the
binding of 2 ,3 -cAMP in a conformation poised to break P–O3
0
0
0
0
0
with other parts of 2 ,3 -cAMP is not necessary for catalysis by
bond. We addressed this through computational docking. The
EAL domain PDEs have highly conserved active sites, known struc-
tures and well-characterized catalytic mechanism [24], they are
therefore selected for docking analysis. Steric hindrance is unlikely
0
0
PaAcpH. Meanwhile, AP did not hydrolyze 2 ,3 -cAMP after
prolonged incubation (Fig. 2C), but generated adenosine when
0
0
used in combination with YybT (Fig. 2D), suggesting that 2 ,3 -
cAMP is susceptible to PDEs only and not any metal-assisted
phosphohydrolase.
0
0
the reason here because docking of 2 ,3 -cyclic nucleotide into an
EAL domain revealed no steric clash between the protein and
nucleotide, regardless of whether the nucleotide was modeled to
0
0
Kinetic measurements using 2 ,3 -cAMP as a substrate revealed
values of sub-to-high mM range (0.045–31 mM) (Table 1). Such
0
0
K
m
generate 3 -AMP or 2 -AMP (Fig. 3A and B). As a comparison, when
0
0
high K
nucleotides (3 ,5 -c(A/G)MP, c-di-AMP, or c-di-GMP), where their
values are in the range of sub-to-low M [15,19,26], suggesting
m
values are not observed for cognate PDEs of signaling
2 ,3 -cyclic nucleotide was docked into CNPase in a conformation
for attacking the P–O2 bond, we observed significant steric clash
0
0
0
K
m
l
between the ribose, base and the enzyme (Fig. 3D). No clash was
observed when the substrate was docked in a conformation for
0
0
non-specific binding of 2 ,3 -cAMP. Importantly, we observed a
correlation (spearman correlation coefficient = 0.95, p = 0.008) be-
tween the catalytic efficiency of individual protein against bis-
pNPP and that against 2 ,3 -cAMP (Fig. 2G). This correlation
strongly suggests that, similar to the case of bis-pNPP, hydrolysis
of 2 ,3 -cAMP here is a non-specific task and its rate primarily
depends on a combination of catalytic competence and active site
accessibility of the PDE used. To our knowledge, no such correla-
tion has been reported before, nor has the non-specific nature of
0
attacking the P–O3 bond (Fig. 3C). Thus, CNPase enforces selective
0
0
0
cleavage of the P–O3 bond of 2 ,3 -cAMP by steric exclusion, a
property not shared by metal-dependant PDEs that cleaves the
0
0
0
P–O2 bond.
0
0
0
The second potential explanation is that the P–O2 bond is ener-
getically less stable and hence more prone to hydrolysis. This is
0
consistent with the products generated (>95% 3 -AMP) by the small
metal–hydroxide complex, which does not show steric preference
and cleaves bond according to strength [12]. This would also ex-
plain why CNPase requires a steric exclusion property, since it spe-
0
0
2
,3 -cAMP hydrolysis by enzymes.
0
0
3
.2. The weaker P–O2 bond, but not steric exclusion, was responsible
cifically breaks the more energy-demanding P–O3 bond. To verify,
0
for the exclusive formation of 3 -AMP
we attempted to determine the activation energy of the respective
P–O bonds by monitoring spontaneous hydrolysis at a series of ele-
vated temperatures. Unfortunately, the temperature-dependent
cleavage rate cannot be accurately measured due to complications
0
We were intrigued by the selective formation of 3 -AMP but not
-AMP and examined two possible explanations. First, the six
0
2
0
0
0
Fig. 3. (A,B) Modeling of 2 ,3 -cyclic nucleotide into the active site of a EAL domain PDE. The nucleotide was modeled for in-line water attack (red arrow) on the P–O3 bond
0
(
A) or the P–O2 bond (B). The catalytic mechanism of EAL domain is detailed in Ref. [24]: two divalent metal ions (cyan spheres) coordinated by side-chains of active site
residues (gray lines) activate the attacking the water nucleophile (green sphere). (C,D) Modeling of 2 ,3 -cyclic nucleotide into the active site of human brain CNPase The
nucleotide was modeled for in-line water attack (red arrow) on the P–O3 bond (C) or the P–O2 bond (D). The catalytic mechanism of CNPase is detailed in Ref. [36]: a
conserved histidine residue activated by another main-chain carbonyl group in turn activates the attacking water (green sphere). (For interpretation of the references to
colour in this figure legend, the reader is referred to the web version of this article.)
0
0
0
0

5
04
F. Rao et al. / Biochemical and Biophysical Research Communications 398 (2010) 500–505
0
0
0
less stable than 3 ,5 -cAMP, with the P–O2 bond having lower acti-
0
0
vation energy than the P–O3 bond. To understand why the P–O2
bond is weaker, we surveyed crystal structures of various 2 ,3 -cyc-
lic nucleotides and found that its bond length is longer by 0.011–
0
0
0
.07 Å (Table 2) [28–31]. Here, it is conceivable that, due to the
0
0
0
low activation energy of the P–O2 bond in 2 ,3 -cAMP, the energy
barrier for reaching transition state is low, and a metal activated
water/hydroxide group in a given PDE active site can launch a
nucleophilic attack on the phosphorus atom as long as the 2 ,3 -
cAMP molecule can access the active site with proximity.
The mechanistic insights of 2 ,3 -cAMP hydrolysis we gained
here led us to analyze existing reports of proteins hydrolyzing
2
0
0
0
0
0
0
,3 -cAMP, other than the undisputed CNPase family. We found
that our conclusion can guide proper interpretation of results of
0
other proteins (Table 3). For the five enzymes that generate 3 -
AMP as the sole or predominant product, only in vitro enzymatic
characterization data was available, four of them contain the
metallo-beta-lactamas PDE fold [32]. The activity of the PhnP
protein, for instance, was discovered by large-scale substrate
screening [7]. Although the authors discussed how the in vitro
0
0
2
,3 -cAMP hydrolysis might be related to signaling pathways
in vivo, the reactivity could be a non-specific property of PDEs in
general and biologically irrelevant. In comparison, for the two en-
zymes where involvement in RNA modification is obvious, both
0
can generate 2 -AMP. The CthPnkP protein, a homolog of the
0
0
0
0
Fig. 4. HPLC trace of 2 ,3 -cAMP (A) and 3 ,5 -cAMP (B) incubated at 80° without any
enzyme. Buffer conditions: 100 mM Tris, pH 8.3, 20 mM KCl, 0.5 mM Mn2 . The
arrows on top mark the retention times of the respective nucleotides based on
standard. Peaks marked by asterisks are candidate products resulting from ligation
of cyclic nucleotides.
0
k-phage T4 polynucleotide kinase 3 -phosphatase, hydrolyzed
,3 -cAMP into adenosine with 2 -AMP being an intermediate
33], according to the results of its phosphatase inactive mutant
CthPnkP-H198D [34]. The tRNA-NT protein, an E. coli tRNA nucleo-
tidyltransferase that adds a CCA tri-nucleotide to the carboxy-ter-
minus of a tRNA for maturation, has a remarkable metal
independent activity that generates 2 -AMP [35], which is the same
as CNPase and very likely to be its physiological function.
Taken together, our results and analysis suggest that 2 ,3 -cAMP
can be used as a probe to screen for enzymes that work on 2 ,3 -
cyclic phosphodiester structure under physiological context when
2 -AMP is the product, or merely PDEs in general when 3 -AMP is
the product (Fig. 2H). This implies that the numerous reports on
+
0
0
0
2
[
0
arising from formation of ligation product(s), as reported recently
elsewhere [27]. Nevertheless, it was clearly noted that, after incu-
bation at 80 °C, most 2 ,3 -cAMP was broken down into 3 -AMP but
not 2 -AMP (Fig. 4A), while no breakdown products of 3 ,5 -cAMP
3 -AMP or 5 -AMP) were observed (Fig. 4B). We also detected little
0
0
0
0
0
0
0
0
0
0
0
0
(
0
0
0
0
or no hydrolysis of 3 ,5 -cAMP by the six PDEs (Table 1). These re-
sults are consistent with the notion that 2 ,3 -cAMP is energetically
0
0
Table 2
0
0
0
0
The P–O2 and P–O3 bond distance of various 2 ,3 -cyclic nucleotides.
0
0
Nucleotide
CSD refcodea or PDB idb
–c
P–O2 (Å)
P–O3 (Å)
D
(P–O) (Å)
Ref.
0
0
0
0
0
0
0
0
2
2
2
2
,3 -cAMP
1.623
1.617
1.622
1.70
1.612
1.604
1.597
1.63
0.011
0.013
0.025
0.07
[24]
[24]
[25]
[26]
[27]
c
,3 -cGMP
–
a
,3 -cCMP
CYCYPH10
1GSP
b
,3 -cGPS
0
0
d
b
Uridine 2 ,3 -Vanadate
1JH7
1.75
1.71
0.04
a
b
c
Cambridge Structure Database refcode id.
PDB code id.
Id not found.
d
0
0
0
0
0
0
Uridine 2 ,3 -vanadate is an analog of 2 ,3 -UMP where the phosphorus atom is replaced by vanadium. The corresponding bonds should be V–O2 and V–O3 , respectively.
Table 3
0
0
0
0
Steady-state kinetic parameters for proteins that does not hydrolyze 3 ,5 -cAMP but hydrolyze 2 ,3 -cAMP.
cat (min 1
ꢁ
)
K
(mM)
k
cat/K
m
(min
ꢁ1
lM
ꢁ1
)
Ref.
Enzyme
PDE fold
Product(s)
k
m
0
0
CvfA
DR1281
PhnP
YfcE-C74H
Rv0805
CthPnkP
tRNA-NT
HD
3 -AMP
55
17
3.21
[7]
[5]
[6]
[8]
[8]
[29]
[30]
a
Metallo-beta-lactamse
Same as above
Same as above
Same as above
Same as above
HD
3 -AMP
22.8 ± 0.6
103 ± 4
2.6 ± 0.5
150 ± 20
18 ± 1.4
460 ± 30
0.55 ± 0.05
0.11 ± 0.01
0.045 ± 0.015
1.6 ± 0.4
536 ± 24
0.49 ± 0.04
41.5 ± 3.9
936 ± 92
57 ± 22
94 ± 27
30 ± 3
0
3 -AMP
0
3 -AMP
0
0
b
3 -AMP > 2 -AMP
Adenosine
0
c
2 -AMP or mixture
940 ± 100
a
b
c
These metallo-beta-lactamse fold proteins are also referred to as longing to calcineurin-type phosphatase superfamily.
0
0
0
Rv0805 generates a mixture of 3 -AMP and 2 -AMP, with 3 -AMP being the predominant species.
0
0
0
tRNA-NT generates 2 -AMP in the absence of divalent metal ions, but a mixture of 2 -AMP and 3 -AMP in the presence.

F. Rao et al. / Biochemical and Biophysical Research Communications 398 (2010) 500–505
505
0
0
0
2
,3 -cyclic-nucleotide 2 -phosphodiesterase activity (EC 3.1.4.16,
see http://www.brenda-enzymes.org/php/result_flat.php4?ecno=
.1.4.16) could be attributed to non-specific activity of some
[13] M. Proudfoot, E. Kuznetsova, S.A. Sanders, C.F. Gonzalez, G. Brown, A.M.
Edwards, C.H. Arrowsmith, A.F. Yakunin, High throughput screening of purified
proteins for enzymatic activity, Methods Mol. Biol. 426 (2008) 331–341.
3
[
14] D.J. Miller, L. Shuvalova, E. Evdokimova, A. Savchenko, A.F. Yakunin, W.F.
Anderson, Structural and biochemical characterization of a novel Mn2+-
dependent phosphodiesterase encoded by the yfcE gene, Protein Sci. 16 (2007)
PDE(s). The real functions of these proteins remain to be eluci-
dated. Another implication of our results is that the scarcity of re-
ports of 2 ,3 -cyclic mononucleotides isolation from biological
systems may be due to their susceptibility to hydrolysis by non-
specific PDEs. Treatment with divalent metal chelating agents in-
stead of a specific class of PDE inhibitor will facilitate the isolation
process. Finally, we suggest the use of 2 ,3 -AMP as a tool to char-
acterize PDE and other phosphohydrolyase, the advantages of
which, in comparison to the widely used bis-pNPP, is discussed
in Supplementary material.
1338–1348.
0
0
[
[
[
15] F. Rao, R.Y. See, D. Zhang, D.C. Toh, Q. Ji, Z.X. Liang, YybT is a signaling protein
that contains a cyclic dinucleotide phosphodiesterase domain and a GGDEF
domain with ATPase activity, J. Biol. Chem. 285 (2010) 473–482.
16] F. Rao, S. Pasunooti, Y. Ng, W. Zhuo, L. Lim, W. Liu, Z.-X. Liang, Enzymatic
synthesis of c-di-GMP using
Biochem. 389 (2009) 138–142.
a thermophilic diguanylate cyclase, Anal.
0
0
17] L. Aravind, E.V. Koonin, A novel family of predicted phosphoesterases includes
Drosophila prune protein and bacterial RecJ exonuclease, Trends Biochem. Sci.
23 (1998) 17–19.
[18] U. Mechold, G. Fang, S. Ngo, V. Ogryzko, A. Danchin, YtqI from Bacillus subtilis
has both oligoribonuclease and pAp-phosphatase activity, Nucleic Acids Res.
3
5 (2007) 4552–4561.
Acknowledgments
[
[
19] F. Rao, Y. Qi, H.S. Chong, M. Kotada, B. Li, J. Lescar, K. Tang, Z.-X. Liang, The
functional role of a conserved loop in EAL domain-based c-di-GMP specific
phosphodiesterase, J. Bacteriol. 191 (2009) 4722–4731.
This work is supported by Singapore Biomedical Research
Council (Grant 06/1/22/19/464). We thank Prof. Zhao-Xun Liang
for his encouragement, discussion, and support; and I. Soehano
for her experimental assistance.
20] F. Rao, Y. Yang, Y. Qi, Z.X. Liang, Catalytic mechanism of c-di-GMP specific
phosphodiesterase:
Pseudomonas aeruginosa, J. Bacteriol. 190 (2008) 3622–3631.
[21] L. Aravind, E.V. Koonin, The HD domain defines a new superfamily of metal-
dependent phosphohydrolases, Trends Biochem. Sci. 23 (1998) 469–472.
a study of the EAL domain-containing RocR from
[
22] E. Murugan, R. Kong, H. Sun, F. Rao, Z.X. Liang, Expression, purification and
characterization of the acyl carrier protein phosphodiesterase from
Pseudomonas aeruginosa, Protein Expr. Purif. 71 (2010) 132–138.
Appendix A. Supplementary data
[
23] I.P. Fabrichniy, L. Lehtio, M. Tammenkoski, A.B. Zyryanov, E. Oksanen, A.A.
Baykov, R. Lahti, A. Goldman, A trimetal site and substrate distortion in a
family II inorganic pyrophosphatase, J. Biol. Chem. 282 (2007) 1422–1431.
24] T.R.M. Barends et al., Structure and mechanism of a bacterial light-regulated
cyclic nucleotide phosphodiesterase, Nature 459 (2009) 1015–1018.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bbrc.2010.06.107.
[
[
References
25] T. Hogg, U. Mechold, H. Malke, M. Cashel, R. Hilgenfeld, Conformational
antagonism between opposing active sites in
homolog modulates (p)ppGpp metabolism during the stringent response
corrected], Cell 117 (2004) 57–68.
a bifunctional RelA/SpoT
[
1] C.M. Cuchillo, X. Pares, A. Guasch, T. Barman, F. Travers, M.V. Nogues, The role
0
0
of 2 ,3 -cyclic phosphodiesters in the bovine pancreatic ribonuclease
catalysed cleavage of RNA: intermediates or products?, FEBS Lett 333 (1993)
07–210.
2] J. Ren, Z. Mi, N.A. Stewart, E.K. Jackson, Identification and quantification of
A
[
[
[
[
26] A.T. Bender, J.A. Beavo, Cyclic nucleotide phosphodiesterases: molecular
regulation to clinical use, Pharmacol. Rev. 58 (2006) 488–520.
27] G. Costanzo, S. Pino, F. Ciciriello, E. Di Mauro, Generation of long RNA chains in
water, J. Biol. Chem. 284 (2009) 33206–33216.
2
[
[
[
[
0
0
2
,3 -cAMP release by the kidney, J. Pharmacol. Exp. Ther. 328 (2009) 855–865.
0
0
3] E.K. Jackson, J. Ren, Z. Mi, Extracellular 2 ,3 -cAMP is a source of adenosine, J.
Biol. Chem. 284 (2009) 33097–33106.
28] K. Sieroslawski, K. Slepokura, T. Lis, M. Bogucka, J. Lutomska, A. Kraszewski,
0
0
Different nucleobase orientations in two cyclic 2 ,3 -phosphates of purine
0
0
4] G.I. Drummond, N.T. Iyer, J. Keith, Hydrolysis of ribonucleoside 2 ,3 -cyclic
phosphates by a diesterase from brain, J. Biol. Chem. 237 (1962) 3535–3539.
0
0
0
0
ribonucleosides: Et3NH(2 ,3 -cAMP) and Et3NH(2 ,3 -cGMP)ꢂH2O, Acta
Crystallogr. C 62 (2006) o405–o409.
0
0
5] B. Schwer, A. Aronova, A. Ramirez, P. Braun, S. Shuman, Mammalian 2 ,3 cyclic
nucleotide phosphodiesterase (CNP) can function as a tRNA splicing enzyme
in vivo, RNA 14 (2008) 204–210.
[
[
29] B.S. Reddy, W. Saenger, Molecular and crystal structure of the free acid of
0
0
cytidine 2 ,3 -cyclophosphate, Acta Crystallogr. B 34 (1978) 1520–1524.
30] I. Zegers, R. Loris, G. Dehollander, A. Fattah Haikal, F. Poortmans, J. Steyaert, L.
Wyns, Hydrolysis of a slow cyclic thiophosphate substrate of RNase T1
analyzed by time-resolved crystallography, Nat. Struct. Biol. 5 (1998) 280–283.
31] A. Hofmann, M. Grella, I. Botos, W. Filipowicz, A. Wlodawer, Crystal structures
of the semireduced and inhibitor-bound forms of cyclic nucleotide
phosphodiesterase from Arabidopsis thaliana, J. Biol. Chem. 277 (2002) 1419–
[
6] D.H. Shin, M. Proudfoot, H.J. Lim, I.K. Choi, H. Yokota, A.F. Yakunin, R. Kim, S.H.
Kim, Structural and enzymatic characterization of DR1281: a calcineurin-like
phosphoesterase from Deinococcus radiodurans, Proteins 70 (2008) 1000–1009.
7] K. Podzelinska, S.M. He, M. Wathier, A. Yakunin, M. Proudfoot, B. Hove-Jensen,
D.L. Zechel, Z. Jia, Structure of PhnP, a phosphodiesterase of the carbon–
phosphorus lyase pathway for phosphonate degradation, J. Biol. Chem. 284
[
[
1425.
(
2009) 17216–17226.
[
[
32] L. Aravind, An evolutionary classification of the metallo-beta-lactamase fold
proteins, In Silico Biol. 1 (1999) 69–91.
33] N. Keppetipola, J. Nandakumar, S. Shuman, Reprogramming the tRNA-splicing
activity of a bacterial RNA repair enzyme, Nucleic Acids Res. 35 (2007) 3624–
[
8] M. Nagata, C. Kaito, K. Sekimizu, Phosphodiesterase activity of CvfA is required
for virulence in Staphylococcus aureus, J. Biol. Chem. 283 (2008) 2176–2184.
[
9] N. Keppetipola, S. Shuman,
metallophosphodiesterase enzymes is
nucleotide phosphodiesterase activity, J. Biol. Chem. 283 (2008) 30942–30949.
A phosphate-binding histidine of binuclear
0 0
determinant of 2 ,3 -cyclic
a
3
630.
0
0
[
[
[
34] N. Keppetipola, S. Shuman, Characterization of the 2 ,3
cyclic
[
[
10] R. Mazumder, L.M. Iyer, S. Vasudevan, L. Aravind, Detection of novel members,
structure–function analysis and evolutionary classification of the 2H
phosphoesterase superfamily, Nucleic Acids Res. 30 (2002) 5229–5243.
11] F.J. Marsh, P. Weiner, J.E. Douglas, P.A. Kollman, G.L. Kenyon, J.A. Gerlt,
Theoretical calculations of hydrolysis energies of ‘‘high-energy” molecules. 3.
phosphodiesterase activities of Clostridium thermocellum polynucleotide
kinase-phosphatase and bacteriophage lambda phosphatase, Nucleic Acids
Res. 35 (2007) 7721–7732.
35] A.F. Yakunin, M. Proudfoot, E. Kuznetsova, A. Savchenko, G. Brown, C.H.
Arrowsmith, A.M. Edwards, The HD domain of the Escherichia coli tRNA
0
0
0
0
Theoretical calculations on the geometric destabilization of 3 ,5 - and 2 ,3 -
cyclic nucleotides, J. Am. Chem. Soc. 102 (1980) 1660–1665.
0
0
0
nucleotidyltransferase has 2 ,3 -cyclic phosphodiesterase, 2 -nucleotidase, and
phosphatase activities, J. Biol. Chem. 279 (2004) 36819–36827.
0
0
[
12] M. Wall, B. Linkletter, D. Williams, R.C. Hynes, J. Chin, Rapid hydrolysis of 2 ,3 -
cAMP with a Cu(II) complex: effect of intramolecular hydrogen bonding on the
basicity and reactivity of a metal-bound hydroxide, J. Am. Chem. Soc. 121
36] Y. Sakamoto, N. Tanaka, T. Ichimiya, T. Kurihara, K.T. Nakamura, Crystal
0
0
0
structure of the catalytic fragment of human brain 2 ,3 -cyclic-nucleotide 3 -
phosphodiesterase, J. Mol. Biol. 346 (2005) 789–800.
(
1999) 4710–4711.