PDE7A1 hydrolyzes cCMP

Article abstract of DOI:10.1016/j.febslet.2014.08.005

The degradation and biological role of the cyclic pyrimidine nucleotide cCMP is largely elusive. We investigated nucleoside 3′,5′-cyclic monophosphate (cNMP) specificity of six different recombinant phosphodiesterases (PDEs) by using a highly-sensitive HPLC-MS/MS detection method. PDE7A1 was the only enzyme that hydrolyzed significant amounts of cCMP. Enzyme kinetic studies using purified GST-tagged truncated PDE7A1 revealed a cCMP K_M value of 135 ± 19 μM. The V_max for cCMP hydrolysis reached 745 ± 27 nmol/(min mg), which is about 6-fold higher than the corresponding velocity for adenosine 3′,5′-cyclic monophosphate (cAMP) degradation. In summary, PDE7A is a high-speed and low-affinity PDE for cCMP.

Full text of DOI:10.1016/j.febslet.2014.08.005

FEBS Letters 588 (2014) 3469–3474
journal homepage: www.FEBSLetters.org
PDE7A1 hydrolyzes cCMP
Maike Monzel a, Maike Kuhn a,1, Heike Bähre a,b, Roland Seifert a, Erich H. Schneider a,
⇑
a Institute of Pharmacology, Hannover Medical School, Carl-Neuberg-Str. 1, 30625 Hannover, Germany
b Core Facility Mass Spectrometry and Metabolomics, Institute of Pharmacology, Hannover Medical School, Carl-Neuberg-Str. 1, 30625 Hannover, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
The degradation and biological role of the cyclic pyrimidine nucleotide cCMP is largely elusive. We
investigated nucleoside 30,50-cyclic monophosphate (cNMP) specificity of six different recombinant
phosphodiesterases (PDEs) by using a highly-sensitive HPLC–MS/MS detection method. PDE7A1 was
the only enzyme that hydrolyzed significant amounts of cCMP. Enzyme kinetic studies using purified
Received 13 June 2014
Revised 2 August 2014
Accepted 4 August 2014
Available online 13 August 2014
GST-tagged truncated PDE7A1 revealed a cCMP KM value of 135 19 lM. The Vmax for cCMP hydro-
lysis reached 745 27 nmol/(min mg), which is about 6-fold higher than the corresponding velocity
for adenosine 30,50-cyclic monophosphate (cAMP) degradation. In summary, PDE7A is a high-speed
and low-affinity PDE for cCMP.
Edited by Peter Brzezinski
Keywords:
Phosphodiesterase
HPLC–MS
Ó 2014 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
Cyclic nucleotides
Cyclic CMP
Second messenger
Enzyme kinetics
1. Introduction
to detect even very low concentrations of cyclic pyrimidine nucle-
otides in cells, tissues and enzymatically digested samples [4,5].
Physiological role and metabolism of the pyrimidine nucleotide
30,50-cCMP are still largely elusive. The existence of a cCMP-form-
ing cytidylyl cyclase was proposed [1], but technical problems
resulted in a false-positive detection of cCMP [2]. The lack of sen-
sitive detection methods hampered research activity in this field
for many years. With the advent of highly sensitive and selective
mass spectrometry-based detection systems [3], it became possible
cCMP is synthesized by soluble adenylyl cyclase [4], by soluble gua-
nylyl cyclase [6] and by the Pseudomonas aeruginosa exotoxin ExoY
[7]. It occurs in numerous cell types like human embryonic kidney
cell line (HEK-293) cells and rat B103 neuroblastoma cells and pri-
mary cells from the neuronal, mesenchymal or epithelial lineage
[8]. The activity of hyperpolarization-activated cyclic nucleotide-
gated cation channel (HCN) ion channels is modulated by cCMP
[9]. Furthermore, cCMP-agarose binds protein kinase A (PKA) [10]
and cCMP regulates both PKA and protein kinase G (PKG) in vitro
[11]. cCMP causes guanosine 30,50-cyclic monophosphate (cGMP)
kinase I activation and subsequent relaxation of murine aorta
smooth muscle cells [12]. If cCMP is a second messenger like the
purine nucleotides adenosine 30,50-cyclic monophosphate (cAMP)
and cCMP, it needs inactivation mechanisms to switch off the intra-
cellular signal. A so-called ‘‘cCMP-specific‘‘ phosphodiesterase
(PDE) was claimed with KM values in the millimolar range [13–
16]. Moreover, a ‘‘multifunctional cCMP-PDE’’ was proposed with
Abbreviations: A549, human lung carcinoma cell line; AMP, adenosine 50-
monophosphate; cAMP, adenosine 30,50-cyclic monophosphate; B103, rat neuro-
blastoma cells; CMP, cytidine 50-monophosphate; cCMP, cytidine 30,50-cyclic
monophosphate; CyaA, Bordetella pertussis adenylyl cyclase toxin; ExoY, P. aeru-
ginosa exotoxin Y; GMP, guanosine 50-monophosphate; cGMP, guanosine 30,50-
cyclic monophosphate; GST, glutathione-S-transferase protein tag; HCN, hyperpo-
larization-activated cyclic nucleotide-gated cation channel; HEK-293, human
embryonic kidney cell line; HIS, hexahistidine tag; HPLC–MS/MS, high performance
liquid chromatography tandem mass spectrometry; HuT 78, human lymphoblast
cell line; IMP, inosine 50-monophosphate; cIMP, inosine 30,50-cyclic monophos-
phate; MRP5, multidrug resistance-associated protein 5; cNMP, nucleoside 30,50-
cyclic monophosphate; PDE, phosphodiesterase; PKA, protein kinase A; PKG,
protein kinase G; Sf9, Spodoptera frugiperda insect cell line; TMP, thymidine 50-
monophosphate; cTMP, thymidine 30,50-cyclic monophosphate; UMP, uridine 50-
monophosphate; cUMP, uridine 30,50-cyclic monophosphate; U-937, human prom-
onocytic cell line; XMP, xanthosine 50-monophosphate; cXMP, xanthosine 30,50-
cyclic monophosphate
KM values <200 l
M and activity for both 30,50- and 2030-cNMPs
[16–21]. However, the identity of these proteins is still elusive.
We studied nucleoside 30,50-cyclic monophosphate (cNMP)
specificity of eight recombinant PDEs (PDE1B, 2A, 3A, 3B, 4B, 5A,
8A and 9A), but, despite their broad substrate specificity, none of
these enzymes accepted cCMP as a substrate [5]. In this publication
we identified PDE7A as a cCMP-degrading PDE.
⇑
Corresponding author.
1
Current affiliation: TWINCORE Center for Experimental and Clinical Infection
Research, Feodor-Lynen-Str. 7, 30625 Hannover, Germany.
http://dx.doi.org/10.1016/j.febslet.2014.08.005
0014-5793/Ó 2014 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

3470
M. Monzel et al. / FEBS Letters 588 (2014) 3469–3474
to short centrifugations (at least every 30 min) to avoid concentra-
tion changes by evaporation- and condensation processes.
2. Materials and methods
2.1. Enzymes and reagents
2.4. Determination of cCMP KM and Vmax for the GST-tagged purified
PDE7A1/2
Recombinant truncated PDE7A1/2 (shared sequence of PDE7A1
and PDE7A2, starting at amino acid 121 of Gene bank
#NM_002603) was obtained from BPS Bioscience (San Diego, CA,
USA). The enzyme was N-terminally GST-tagged, had a purity of
Michaelis Menten kinetics was studied in PDE buffer + 0.05%
(m/v) BSA at an enzyme concentration of 0.95 lg/ml, correspond-
ing to 210 pmol/(min ml) in the presence of increasing concentra-
tions of cyclic nucleotide. Enzyme-free control samples were run in
parallel for each individual cyclic nucleotide concentration. After
an incubation of 1 h at 30 °C, the samples were stopped, processed
and analyzed by HPLC–MS/MS as described in Section 2.2 and in
the Supplementary Methods. Only samples with <40% substrate
hydrolysis were included into data analysis.
A rough estimation of the kinetics of PDE7A1/2-mediated cAMP
hydrolysis was performed by incubating 50 nM, 75 nM, 100 nM
and 250 nM of cAMP for 6 min at 30 °C in the presence of 0.06–
>44% and an activity of 221 pmol/min/lg. Full length N-terminally
HIS-tagged PDE1A3, 6AB, 7A1, 10A1 and 11A1 (lysates from baculo-
virus-infected Spodoptera frugiperda insect cell line (Sf9) cells,
0.5 pmol/min/ll) and a control lysate from uninfected Sf9 cells were
purchased from sb drug discovery (Glasgow, UK). Detailed informa-
tion on the purity and quality of the different PDE preparations is
provided in the Supplementary Methods (1.3). Calmodulin was
obtained from Calbiochem (Merck, Darmstadt, Germany). The inter-
nal standard tenofovir was obtained from the NIH AIDS Research and
Reference Reagent Program (Germantown, MD, USA). The selective
and competitive PDE7A inhibitor BRL-50481 (5-Nitro-2,N,N-trim-
ethylbenzenesulfonamide) was obtained from TOCRIS Bioscience
(Bristol, UK) and dissolved in DMSO to yield a 5 mM stock solution.
Adenosine 50-monophosphate (AMP), cytidine 50-monophosphate
(CMP), guanosine 50-monophosphate (GMP), thymidine 50-
monophosphate (TMP), uridine 50-monophosphate (UMP), inosine
50-monophosphate (IMP) and xanthosine 50-monophosphate
(XMP) were from Sigma–Aldrich (Steinheim, Germany). The cyclic
30,50-nucleotides cAMP, cCMP, cGMP, inosine 30,50-cyclic monophos-
phate (cIMP), thymidine 30,50-cyclic monophosphate (cTMP),
uridine 30,50-cyclic monophosphate (cUMP) and xanthosine
30,50-cyclic monophosphate (cXMP) were purchased from Biolog Life
Science Institute (Bremen, Germany). All other reagents were
analytical grade and from standard suppliers.
0.24 lg/ml of enzyme. Only samples with <60% substrate degrada-
tion in at least one of the duplicates were included into analysis.
2.5. Determination of the effect of the competitive inhibitor BRL-50481
on PDE7A activity
The effect of the competitive inhibitor BRL-50481 on PDE7A1/2
activity was studied at concentrations between 10 nM and 100 lM
in PDE buffer + 0.05% (m/v) BSA. Incubations were performed for
15 min (cAMP) or 4 h (cCMP) at 30 °C in the presence of 500 nM
(cAMP) or 10
PDE7A1/2 was used at 0.48
for cAMP hydrolysis. For cCMP experiments 0.95
l
M (cCMP) of cyclic nucleotide. The GST-tagged
g/ml, corresponding to 105 pmol/(min ml)
g/ml enzyme
l
l
were used, corresponding to 210 pmol/(min ml). DMSO content
of all samples was adjusted to 2%.
2.2. Screening of enzymes with various cNMPs
2.6. Quantitation of cyclic nucleotides (cNMPs) and nucleotide
monophosphates (NMPs) by HPLC–MS/MS and data analysis
Hydrolytic activity of various PDEs was determined in 1x PDE
buffer (50 mM Tris–HCl, 1.7 mM EDTA, 8.3 mM MgCl2). For analysis
The concentrations of the cNMPs and NMPs were determined
by HPLC–MS/MS, consisting of an UFLC HPLC system (Shimadzu,
Kyoto, Japan) and the QTRAP5500™ triple quadrupole mass spec-
trometer (ABSciex, Framingham, MA, USA). A detailed description
of HPLC configuration, eluent composition, m/z values, retention
times and data analysis is provided in the Supplementary Methods
(1.1 and 1.2).
of PDE1A3 activity, 100 lM CaCl2, 100 mM EGTA and 100 nM cal-
modulin were added. PDE1A3, PDE6AB, PDE7A1 and PDE11A1 were
analyzed at a volume activity of 5 pmol/(min ml). PDE10A1 was
added at a volume activity of 12.5 pmol/(min ml). GST-tagged and
truncated PDE7A was used at a concentration of 0.95
sponding to 210 pmol/(min ml). Cyclic nucleotides were added to
yield a final concentration of 3 M. Incubation was performed for
lg/ml, corre-
l
1–24 h at 30 °C. Enzyme was inactivated by 15 min incubation at
95 °C. Protein was precipitated by freezing and removed by centri-
fugation. The supernatant was diluted 1:5 with purified water and
then additionally mixed with an equal volume of tenofovir solution
(internal standard), yielding a final tenofovir concentration of
50 ng/mL. The samples were analyzed by high performance liquid
chromatography tandem mass spectrometry (HPLC–MS/MS) as
described in the Supplementary Methods.
2.7. Preparation of whole cell homogenates and Western blots
The preparation of whole cell homogenates and Western blot-
ting for Suppl. Fig. 1 are described in detail in the Supplementary
information (Suppl. Methods 1.4 and 1.5).
3. Results
2.3. Time course of PDE7A activity
3.1. Analysis of substrate specificity and identification of PDE7A1 as
cCMP-hydrolyzing PDE
Time course experiments were performed in PDE buffer + 0.05%
(m/v) BSA at a cAMP- or cCMP-concentration of 3
GST-tagged and truncated PDE7A1/2 was added at a concentration
of 0.95 g/ml, corresponding to 210 pmol/(min ml). Crude PDE7A1-
lM. The purified
The substrate specificity of PDE1A3, 6AB, 7A1, 10A1 and 11A1
(HIS-tagged full-length enzymes) was characterized in crude Sf9 cell
lysates. The PDEs were incubated in the presence of 30,50-cAMP,
l
containing Sf9 cell lysate was added to yield a final activity of
5 U/ml. The samples were incubated at 30 °C under constant
shaking. Aliquots were drawn at appropriate times, processed as
described for the enzyme screening experiments (Section 2.2) and
analyzed by HPLC–MS/MS as described in the Supplementary
Methods. In all experiments, the samples were repeatedly subjected
-cGMP, -cCMP, -cUMP, -cXMP, -cIMP and -cTMP (3 lM each) for
24 h. In case of <100 % hydrolysis, the velocity of the enzymatic reac-
tion was calculated. When 100 % of the specific cNMP was hydro-
lyzed within 24 h, incubation was shortened until hydrolysis was
<100%. Sincetheenzymeswerenotpurified, a lysatefromuninfected
Sf9 cells was used as a negative control. The PDE-overexpressing Sf9

M. Monzel et al. / FEBS Letters 588 (2014) 3469–3474
3471
cell lysates were diluted several-fold by the manufacturer to yield
the specified activity. Thus, the control lysate should show the max-
imum possible background activity, and we only interpreted PDE
activities higher than the corresponding activity of the control
sample.
CMP increased linearly for up to 8 h (Fig. 1A). The purified GST-
tagged and truncated enzyme, PDE7A1/2, also converted cCMP to
CMP (Fig. 1B). However, while it took >4 h to degrade 50% of cCMP
(Fig. 1B), the corresponding amount of cAMP was converted to
AMP in <30 min (Fig. 1C). A comparison of the linear curve parts
(Fig. 1D; first 30 min for cAMP and first 60 min for cCMP) shows
that the activity of PDE7A1/2 for cCMP was ꢀ13% of its cAMP
activity under these experimental conditions. Since the KM value
of PDE7A for cAMP is 100–200 nM [22], the 3 lM of cAMP in our
time course experiments should saturate the enzyme, yielding the
Vmax of cAMP hydrolysis. The very slow cCMP hydrolysis, however,
suggests that the corresponding KM value is much higher than 3 lM.
We confirmed the broad substrate specificity of PDEs (Suppl.
Table 1) [5]. PDE1A3 hydrolyzed cXMP and cIMP with similar activ-
ity as cAMP and cGMP. PDE6AB hydrolyzed cIMP with similar effi-
ciency as cAMP and cGMP, and PDE7A1 showed some activity for
cTMP. Most interestingly PDE6AB and PDE7A1 exhibited activity
for cCMP. Hydrolysis of cCMP was undetectable in the control lysate
and can therefore be attributed to the over-expressed PDEs. Neither
PDE1A3, 10A1, 11A1, nor the eight PDEs analyzed in our previous
report [5] degraded cCMP. Thus, PDE6AB and PDE7A1 represent
the first PDEs that hydrolyze cCMP. PDE7A1 hydrolyzed cCMP with
considerably higher activity than PDE6AB, comparable to its activ-
ity for the ‘‘standard’’ second messengers cAMP and cGMP.
Therefore, we focused on PDE7A1 and characterized its cCMP-
hydrolyzing activity. To minimize background PDE activity we
used a GST-tagged affinity-purified (>44%) and truncated enzyme,
representing the consensus sequence of the splice variants PDE7A1
and PDE7A2 (=‘‘PDE7A1/2’’).
3.3. Determination of KM and Vmax of PDE7A1/2 for cCMP and of Vmax
for cAMP
Michaelis–Menten analysis was performed with cCMP concen-
trations of up to 2 mM. Fig. 2A shows that detection of CMP accu-
mulation allowed a highly sensitive analysis of enzyme activity,
yielding KM and Vmax values of 135 19
respectively (Fig. 2A, Table 1). Measurement of cCMP degradation
up to a maximum substrate concentration of 250 M yielded
lM and 745 27 nmol/(min mg),
l
comparable results (Fig. 2B, Table 1). For comparison, the Vmax of
3.2. Time course of PDE7A1/2 activity
PDE7A1/2 for cAMP hydrolysis was calculated from Fig. 1D
(saturation of PDE7A1/2 with
3 lM of cAMP), yielding
The time course of PDE7A1-mediated cCMP hydrolysis was
113.8 8.7 nmol/(min mg). We determined the speed of cAMP
degradation in the presence of 50, 75, 100 and 250 nM of substrate
(2–3 experiments in duplicates, mean SEM, data not shown) and
determined in the presence of 3
lM of substrate. PDE7A1 Sf9 cell
lysate hydrolyzed cCMP, and the concentration of the product
Fig. 1. Time course of PDE7A-mediated hydrolysis of 30,50-cAMP and 30,50-cCMP. (A) formation of 50-CMP by a PDE7A1-containing Sf9 cell lysate in the presence of 3
l
M of
30,50-cCMP; (B) degradation of 30,50-cCMP (3 M) and formation of 50-CMP by purified GST-tagged PDE7A1/2; (C) degradation of 30,50-cAMP (3 M) and formation of 50-AMP
l
l
by GST-tagged purified PDE7A1/2; (D) comparison of the initial linear parts of the curves for 30,50-cAMP and 30,50-cCMP hydrolysis by purified GST-tagged PDE7A. All
experiments were performed as indicated in Section 2. Data are means SEM. (A) shows merged data from two independent experiments in duplicates with partially
different sampling times; (B–D) are from 2 independent experiments in duplicates with the same sampling.

3472
M. Monzel et al. / FEBS Letters 588 (2014) 3469–3474
Fig. 2. Michaelis Menten kinetics of PDE7A-mediated hydrolysis of 30,50-cCMP. (A) Velocity of 50-cCMP formation and (B) 30,50-cCMP degradation during incubation of GST-
tagged PDE7A1/2 with increasing concentrations of 30,50-cCMP. All experiments were performed as indicated in Section 2. Data are means SEM from 2 to 4 independent
experiments in duplicates.
Table 1
Characterization of cCMP hydrolysis by GST-tagged PDE7A1/2.
Parameter
CMP formation
cCMP hydrolysis
AMP formation
cAMP hydrolysis
Vmax
KM
745 27 nmol/(min mg)
560 110 nmol/(min mg)
n.d.
n.d.
5.37 0.15
n.d.
113.8 8.7 nmol/(min mg)
<100 nM (estimation)
5.31 0.11
135 19
lM
61 29
l
M
pIC50 (BRL-50481)
Ki (BRL-50481)
6.98 0.10
98 nM
7.01 0.14
91 nM
n.d.
n.d. = not determined.
Fig. 3. Inhibition of PDE7A-mediated 30,50-cAMP and 30,50-cCMP hydrolysis by BRL-50481. (A) 30,50-cAMP (500 nM) degradation and 50-AMP formation in the presence of
increasing concentrations of BRL-50481. (B) 30,50-cCMP (10 M) degradation and 50-CMP formation in the presence of increasing concentrations of BRL-50481. All
experiments were performed as indicated in Section 2. Curves were normalized to 100% cyclic nucleotide concentration in the presence of 100 M of inhibitor (=100%
l
l
inhibition) and to 100% product formation in the absence of inhibitor (=100% enzyme activity and no inhibition). In both graphs, the data point at logꢁ9 represents the
inhibitor-free control. Data are means SEM from 3 independent experiments in duplicates.
found velocities of 69 4, 89 2, 117 10 and 132 29 nmol/(min mg),
respectively. Thus, under our experimental conditions, the KM
value of PDE7A1/2 for cAMP is probably around 50 nM.
assumption of steady-state conditions, the Ki value of BRL-50481
(calculated from IC50, cCMP concentration and KM) was 98 nM
and 91 nM for CMP formation and cCMP hydrolysis, respectively
(Table 1). Since cAMP hydrolysis was almost 100%, even in the
presence of 100 nM of inhibitor, steady-state conditions were not
guaranteed any more, and we did not calculate the BRL-50481 Ki
value for cAMP hydrolysis.
3.4. Inhibition of PDE7A1/2-mediated hydrolysis of cAMP and cCMP by
BRL-50481
BRL-50481 is a specific competitive inhibitor of PDE7A-medi-
ated cAMP hydrolysis [23]. PDE7A1/2 activity for cCMP (10
l
M)
3.5. Validation of integrity and identity of the PDE7A used in the
experiments
was reduced by 50% in the presence of ꢀ100 nM of BRL-50481
(Fig. 3B, Table 1). By contrast, the enzymatic activity in the pres-
ence of 500 nM of cAMP was reduced with an IC50 of 4–5
l
M for
In order to confirm the presence and integrity of PDE7A1 in the
Sf9 cell lysate and of PDE7A1/2 in the recombinant preparation, we
performed a Western blot with the enzyme preparations. PDE7A1
was detected in the corresponding Sf9 cell lysate (Suppl. Fig. 1, lane
both cAMP hydrolysis and AMP formation (Fig. 3A, Table 1).
Despite the long incubation time of 4 h, only 50–60% of the cCMP
were hydrolyzed in the absence of inhibitor (Fig. 3B). Under the

M. Monzel et al. / FEBS Letters 588 (2014) 3469–3474
3473
1), while no band was visible in the equally concentrated PDE7A1-
free control lysate (Suppl. Fig. 1, lane 2). PDE7A1 from Sf9 cells
migrated as two bands and slightly faster than the expected molec-
ular weight of 57 kDa. This may be caused by a different glycosyl-
ation pattern in insect cells. Analysis of the recombinant purified
GST-tagged PDE7A1/2 (Suppl. Fig. 1, lane 3) showed the predicted
molecular mass of 60–70 kDa.
Next we aimed at identifying a cell line with high PDE7A1
expression, which may be suited for future functional studies.
Since high PDE7A1 expression was reported for human lympho-
blast cell line (HuT 78) lymphoma cells, in contrast to rather low
expression levels in human promonocytic cell line (U-937) cells
[23], we chose these two cell lines for Western blotting. In fact,
U937 cells (Suppl. Fig. 1, lane 4) exhibited only a very weak signal,
while HuT 78 cells (Suppl. Fig. 1, lane 5) showed an intense band at
the molecular weight corresponding to PDE7A1.
of PDE7A1 are quite different from the features reported for
‘‘cCMP-specific PDE’’ and ‘‘multifunctional PDE’’ (Suppl. Table 2).
While ‘‘cCMP-specific PDE’’ hydrolyzes only cCMP with a KM value
in the millimolar range [13–15], the PDE7A1/2 in our study shows
a >10-fold lower KM-value and hydrolyzes also 30,50-cAMP. The
‘‘multifunctional PDE’’ [17–20] has a KM value of ꢀ180
lM for
M of PDE7A1/2. However,
30,50-cCMP, which is close to the 135
l
‘‘multifunctional cCMP-PDE‘‘ also accepts 20,30-cNMPs as sub-
strates [21], but we did not detect such 20,30-cNMP-hydrolytic
activity for PDE7A1/2 (data not shown). Moreover, PDE7A1 and
PDE7A2 have molecular masses of 57 and 50 kDa, respectively
[29], while the ‘‘multifunctional PDE’’ was purified as a protein
with 33 kDa [21] and the molecular weight of ‘‘cCMP-specific
PDE’’ is only 28 kDa. Therefore, we conclude that we have discov-
ered a third type of cCMP-hydrolyzing PDE, which is distinct from
the cCMP-hydrolyzing activities reported to date (Suppl. Table 2).
The PDE7-specific competitive inhibitor BRL-50481 [23] inhib-
ited PDE7A1/2-mediated cCMP hdrolysis with a Ki value of 98 nM
or 91 nM for CMP formation or cCMP hydrolysis, respectively. This
corresponds well to the Ki value of 180 nM that was previously
reported for inhibition of cAMP hydrolysis by PDE7A1 expressed
in Sf9 cells [23]. This suggests that cCMP may bind to the same
binding site of PDE7A1 as cAMP. From a different point of view,
cCMP could also be regarded as an inhibitor of PDE7A1-mediated
cAMP hydrolysis or vice versa.
4. Discussion
Out of 13 enzymes (five enzymes analyzed in this paper and
eight in Ref. [5]), we identified PDE7A1 as the first PDE that hydro-
lyzes cCMP. Except for a very low activity of PDE6AB, the other
enzymes described in this paper did not hydrolyze cCMP. However,
they partially showed activity for ‘‘exotic’’ cyclic nucleotides like
cXMP, cIMP and cTMP.
Only PDE7A1 and – to a lesser extent – PDE6AB hydrolyze cCMP.
Due to its low Vmax for cAMP, it has been suggested that PDE7A may
need additional factors to be activated or may simply play a role in
regulating ‘‘basal’’ cAMP concentrations [22]. Our findings suggest
that PDE7A may be additionally involved in the degradation of
intracellular cCMP. Analysis of HEK-293 and B103 cells has demon-
strated that cCMP is generated by soluble adenylyl cyclase in mam-
malian cells, reaching concentrations comparable to the established
second messenger cGMP [4]. Since cCMP is biologically active
[9,11,12], it is conceivable that its intracellular concentration has
to be regulated, e.g. by the activity of PDEs. PDE7A1/2 is a low-affin-
PDE7A1/2 has a Vmax value for cCMP, which is 6–7-fold higher
than for cAMP. This may result in some cCMP ‘‘specificity’’ of
PDE7A1/2. Similarly, although the ‘‘cCMP-specific PDE‘‘ has KM val-
ues in the millimolar range that are very similar for both cAMP and
cCMP (Suppl. Table 2), its ‘‘specificity’’ for cCMP is caused by the
ꢀ200-fold higher Vmax value for cCMP hydrolysis, as compared to
cAMP degradation [16].
It may be argued that the observed cCMP-hydrolyzing activity is
not caused by PDE7A1, but due to a contamination. However, the
following arguments support the notion that PDE7A1 hydrolyzes
cCMP: First, the activity was observed in two PDE7A1-containing
preparations that were very different (Sf9 cell lysate and purified
protein) and from two independent commercial suppliers. Second,
as shown in this paper and in our preceding publication [5], the
cCMP-degrading activity was absent in all other enzyme prepara-
tions from these suppliers (except for the low activity observed
with PDE6AB). Finally, the cCMP-hydrolyzing activity was
eliminated by the PDE7A1-selective inhibitor BRL-50481 with a
Ki value that corresponds to the literature [23].
Our results also show distinct substrate profiles for the investi-
gated PDEs (Suppl. Table 1), suggesting PDE-specific ‘‘cNMP signa-
tures’’. Besides the unique cCMP-hydrolyzing activity of PDE7A1,
we also found that PDE1A3 hydrolyzes cXMP and cIMP much faster
than any of the other tested enzymes. This finding confirms our
previous results with purified recombinant PDE1B [5], indicating
that this substrate profile may be a feature shared by all members
of the PDE1 family.
In conclusion, PDEs should be re-visited and studied with the
complete spectrum of cNMPs and with substrate concentrations
up to the micromolar range. Low-affinity- and high-Vmax-PDE
activity as well as PDE activity for ‘‘uncommon’’ cyclic nucleotide
substrates may represent a novel and yet unappreciated compo-
nent of physiological PDE function.
ity PDE for cCMP with a KM value around 135 lM. Thus, the ques-
tion arises, whether this activity is relevant under physiological
conditions. In fact, cCMP occurs in mammalian cells, and HEK-293
cells contain ꢀ30 pmol of cCMP per 106 cells under basal conditions
[8]. Since the volume of a HEK-293 cell is ꢀ1000 femtoliters [24],
this corresponds to a cCMP concentration around 30 lM. This cal-
culation, however, disregards the fact that in many cell types a large
part of the volume is occupied by the nucleus. Thus, if cCMP is con-
fined to the cytoplasm, its actual concentration may be even higher.
In summary, it is likely that cCMP concentrations in human cells
can reach levels, where PDE7A1/2 activity becomes relevant.
The cCMP degrading activity of PDE7A may even become more
important during bacterial infections, since bacterial toxins like
edema factor (EF) [25,26] or adenylyl cyclase toxin CyaA [26], both
from Bacillus anthracis, produce cCMP. In addition, the P. aeruginosa
type III secretion protein ExoY, which shows highly toxic effects in
the rat model [27], produces a significant amount of cCMP in B103
and human lung carcinoma cell line (A549) cells [7]. If a cell is
flooded by cCMP during a bacterial infection, detoxifying processes
are required. PDE7A1-mediated cCMP hydrolysis may be one of
them, in addition to the very recently described export of cCMP
by the multi-drug resistance protein MRP5 [28]. Both PDE7A1
and MRP5 represent low-affinity and high-capacity mechanisms
for the removal of cCMP, which additionally supports a potential
function in the detoxification of exceedingly high cCMP concentra-
tions. cCMP-degrading activity has been described in several types
of cells and tissues [13–21] and was attributed to a ‘‘cCMP-specific
PDE’’ and a ‘‘multifunctional PDE’’ (Suppl. Table 2). However, the
molecular identity of these activities is still elusive. We have
identified PDE7A1 as cCMP-hydrolyzing PDE, but the properties
Acknowledgments
We thank Prof. Harald Genth for excellent scientific advice and
we highly appreciate the helpful comments of the reviewers. More-
over, we thank Ms. Annette Garbe for excellent technical support
and Mr. Justin Kewney from sb drug discovery for his efforts to pro-
vide us with all requested information about the Sf9 cell lysates.

3474
M. Monzel et al. / FEBS Letters 588 (2014) 3469–3474
[15] Kuo, J.F., Brackett, N.L., Shoji, M. and Tse, J. (1978) Cytidine 30:50-
Appendix A. Supplementary data
monophosphate phosphodiesterase in mammalian tissues. Occurrence and
biological involvement. J. Biol. Chem. 253, 2518–2521.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.febslet.2014.08.
005.
[16] Newton, R.P., Bayliss, M.A., Khan, J.A., Bastani, A., Wilkins, A.C., Games, D.E.,
Walton, T.J., Brenton, A.G. and Harris, F.M. (1999) Kinetic analysis of cyclic
CMP-specific and multifunctional phosphodiesterases by quantitative
positive-ion fast-atom bombardment mass spectrometry. Rapid Commun.
Mass Spectrom. 13, 574–584.
[17] Helfman, D.M., Shoji, M. and Kuo, J.F. (1981) Purification to homogeneity and
general properties of a novel phosphodiesterase hydrolyzing cyclic CMP and
cyclic AMP. J. Biol. Chem. 256, 6327–6334.
References
[1] Cech, S.Y. and Ignarro, L.J. (1977) Cytidine 30,50-monophosphate (cyclic CMP)
formation in mammalian tissues. Science 198, 1063–1065.
[2] Gaion, R.M. and Krishna, G. (1979) Cytidylate cyclase: the product isolated by
the method of Cech and Ignarro is not cytidine 30,50-monophosphate. Biochem.
Biophys. Res. Commun. 86, 105–111.
[18] Lavan, B.E., Lakey, T. and Houslay, M.D. (1989) Resolution of soluble cyclic
nucleotide phosphodiesterase isoenzymes, from liver and hepatocytes,
identifies a novel IBMX-insensitive form. Biochem. Pharmacol. 38, 4123–4136.
[19] Worby, A., Mensah, L.M. and Murray, K.J. (1991) Identification of the novel rat
liver IBMX-insensitive phosphodiesterase as a non-specific phosphodiesterase
capable of hydrolysing cCMP. Biochem. Pharmacol. 42, 1318–1321.
[20] Mendel, D.B., Cihlar, T., Moon, K. and Chen, M.S. (1997) Conversion of 1-[((S)-
2-hydroxy-2-oxo-1,4,2-dioxaphosphorinan-5-yl)methyl]cytosine to cidofovir
by an intracellular cyclic CMP phosphodiesterase. Antimicrob. Agents
Chemother. 41, 641–646.
[3] Beste, K.Y., Burhenne, H., Kaever, V., Stasch, J.P. and Seifert, R. (2012)
Nucleotidyl cyclase activity of soluble guanylyl cyclase
51, 194–204.
[4] Hasan, A., Danker, K.Y., Wolter, S., Bähre, H., Kaever, V. and Seifert, R. (2014)
Soluble adenylyl cyclase accounts for high basal cCMP and cUMP
concentrations in HEK293 and B103 cells. Biochem. Biophys. Res. Commun.
448, 236–240.
a1b1. Biochemistry
[21] Helfman, D.M. and Kuo, J.F. (1982)
A
homogeneous cyclic CMP
phosphodiesterase hydrolyzes both pyrimidine and purine cyclic 20:30- and
30:50-nucleotides. J. Biol. Chem. 257, 1044–1047.
[5] Reinecke, D., Burhenne, H., Sandner, P., Kaever, V. and Seifert, R. (2011) Human
cyclic nucleotide phosphodiesterases possess
a much broader substrate-
[22] Bender, A.T. and Beavo, J.A. (2006) Cyclic nucleotide phosphodiesterases:
molecular regulation to clinical use. Pharmacol. Rev. 58, 488–520.
[23] Smith, S.J., Cieslinski, L.B., Newton, R., Donnelly, L.E., Fenwick, P.S., Nicholson,
A.G., Barnes, P.J., Barnette, M.S. and Giembycz, M.A. (2004) Discovery of BRL
specificity than previously appreciated. FEBS Lett. 585, 3259–3262.
[6] Beste, K.Y. and Seifert, R. (2013) CCMP, cUMP, cTMP, cIMP and cXMP as
possible second messengers: development of a hypothesis based on studies
with soluble guanylyl cyclase a1b1. Biol. Chem. 394, 261–270.
50481 [3-(N,N-dimethylsulfonamido)-4-methyl-nitrobenzene],
a selective
[7] Beckert, U., Wolter, S., Hartwig, C., Bähre, H., Kaever, V., Ladant, D., Frank, D.W.
and Seifert, R. (2014) ExoY from Pseudomonas aeruginosa is a nucleotidyl
cyclase with preference for cGMP and cUMP formation. Biochem. Biophys. Res.
Commun. 450, 870–874.
[8] Hartwig, C., Bähre, H., Wolter, S., Beckert, U., Kaever, V., Seifert, R. (2014) cAMP,
cGMP, cCMP and cUMP concentrations across the tree of life: high cCMP and
cUMP levels in astrocytes. Neurosci. Lett. (in press). http://dx.doi/org/10.1016/
j.neulet.2014.07.019.
[9] Zong, X., Krause, S., Chen, C.C., Krüger, J., Gruner, C., Cao-Ehlker, X., Fenske, S.,
Wahl-Schott, C. and Biel, M. (2012) Regulation of hyperpolarization-activated
cyclic nucleotide-gated (HCN) channel activity by cCMP. J. Biol. Chem. 287,
26506–26512.
[10] Hammerschmidt, A., Chatterji, B., Zeiser, J., Schröder, A., Genieser, H.G., Pich,
A., Kaever, V., Schwede, F., Wolter, S. and Seifert, R. (2012) Binding of
regulatory subunits of cyclic AMP-dependent protein kinase to cyclic CMP
agarose. PLoS One 7, e39848.
[11] Wolter, S., Golombek, M. and Seifert, R. (2011) Differential activation of cAMP-
and cGMP-dependent protein kinases by cyclic purine and pyrimidine
nucleotides. Biochem. Biophys. Res. Commun. 415, 563–566.
[12] Desch, M., Schinner, E., Kees, F., Hofmann, F., Seifert, R. and Schlossmann, J.
(2010) Cyclic cytidine 30,50-monophosphate (cCMP) signals via cGMP kinase I.
FEBS Lett. 584, 3979–3984.
[13] Cheng, Y.C. and Bloch, A. (1978) Demonstration, in leukemia L-1210 cells, of a
phosphodiesterase acting on 30:50-cyclic CMP but not on 30:50-cyclic AMP or
30:50-cyclic GMP. J. Biol. Chem. 253, 2522–2524.
inhibitor of phosphodiesterase 7: in vitro studies in human monocytes, lung
macrophages, and CD8+ T-lymphocytes. Mol. Pharmacol. 66, 1679–1689.
[24] Vecsler, M., Lazar, I. and Tzur, A. (2013) Using standard optical flow cytometry
for synchronizing proliferating cells in the G1 phase. PLoS One 8, e83935.
[25] Göttle, M., Dove, S., Kees, F., Schlossmann, J., Geduhn, J., König, B., Shen, Y.,
Tang, W.J., Kaever, V. and Seifert, R. (2010) Cytidylyl and uridylyl cyclase
activity of bacillus anthracis edema factor and Bordetella pertussis CyaA.
Biochemistry 49, 5494–5503.
[26] Beste, K.Y., Spangler, C.M., Burhenne, H., Koch, K.W., Shen, Y., Tang, W.J.,
Kaever, V. and Seifert, R. (2013) Nucleotidyl cyclase activity of particulate
guanylyl cyclase A: comparison with particulate guanylyl cyclases E and F,
soluble guanylyl cyclase and bacterial adenylyl cyclases CyaA and edema
factor. PLoS One 8, e70223.
[27] Stevens, T.C., Ochoa, C.D., Morrow, K.A., Robson, M.J., Prasain, N., Zhou, C.,
Alvarez, D.F., Frank, D.W., Balczon, R. and Stevens, T. (2014) The Pseudomonas
aeruginosa exoenzyme Y impairs endothelial cell proliferation and vascular
repair following lung injury. Am. J. Physiol. Lung Cell. Mol. Physiol. 306, L915–
24.
[28] Laue, S., Winterhoff, M., Kaever, V., van den Heuvel, J.J., Russel, F.G. and Seifert,
R. (2014) cCMP is
a substrate for MRP5. Naunyn Schmiedebergs Arch.
Pharmacol. 387, 893–895.
[29] Han, P., Zhu, X. and Michaeli, T. (1997) Alternative splicing of the high affinity
cAMP-specific phosphodiesterase (PDE7A) mRNA in human skeletal muscle
and heart. J. Biol. Chem. 272, 16152–16157.
[14] Newton, R.P. and Salih, S.G. (1986) Cyclic CMP phosphodiesterase: isolation,
specificity and kinetic properties. Int. J. Biochem. 18, 743–752.