cUMP hydrolysis by PDE3A

Article abstract of DOI:10.1007/s00210-016-1328-1

As previously reported, the cardiac phosphodiesterase PDE3A hydrolyzes cUMP. Moreover, cUMP-degrading activity was detected in cow and dog hearts several decades ago. Our aim was to characterize the enzyme kinetic parameters of PDE3A-mediated cUMP hydrolysis and to investigate whether cUMP and cUMP-hydrolyzing PDEs are present in cardiomyocytes. PDE3A-mediated cUMP hydrolysis was characterized in time course, inhibitor, and Michaelis-Menten kinetics experiments. Intracellular cyclic nucleotide (cNMP) concentrations and the mRNAs of cUMP-degrading PDEs were quantitated in neonatal rat cardiomyocytes (NRCMs) and murine HL-1 cardiomyogenic cells. Moreover, we investigated cUMP degradation in HL-1 cell homogenates and intact cells. Educts (cNMPs) and products (NMPs) of the PDE reactions were detected by HPLC-coupled tandem mass spectrometry. PDE3A degraded cUMP (measurement of UMP formation) with a K_M value of ~143?μM and a V_max value of ~42?μmol/min/mg. PDE3A hydrolyzed cAMP with a K_M value of ~0.7?μM and a V_max of ~1.2?μmol/min/mg (determination of AMP formation). The PDE3 inhibitor milrinone inhibited cUMP hydrolysis (determination of UMP formation) by PDE3A (K_i?=?57?nM). Significant amounts of cUMP as well as of PDE3A mRNA (in addition to PDE3B and PDE9A transcripts) were detected in HL-1 cells and NRCMs. Although HL-1 cell homogenates contain a milrinone-sensitive cUMP-hydrolyzing activity, intact HL-1 cells may use additional PDE3-independent mechanisms for cUMP disposal. PDE3A is a low-affinity and high-velocity PDE for cUMP. Future studies should investigate biological effects of cUMP in cardiomyocytes and the role of PDE3A in detoxifying high intracellular cUMP concentrations under pathophysiological conditions.

Full text of DOI:10.1007/s00210-016-1328-1

Naunyn-Schmiedeberg's Arch Pharmacol (2017) 390:269–280
DOI 10.1007/s00210-016-1328-1
ORIGINAL ARTICLE
cUMP hydrolysis by PDE3A
1
1
2
1
Stefan Berrisch & Jessica Ostermeyer & Volkhard Kaever & Solveig Kälble
&
3
1
1
Denise Hilfiker-Kleiner & Roland Seifert & Erich H. Schneider
Received: 1 July 2016 /Accepted: 30 November 2016 /Published online: 14 December 2016
Springer-Verlag Berlin Heidelberg 2016
#
Abstract As previously reported, the cardiac phosphodiester-
ase PDE3A hydrolyzes cUMP. Moreover, cUMP-degrading
activity was detected in cow and dog hearts several decades
ago. Our aim was to characterize the enzyme kinetic parame-
ters of PDE3A-mediated cUMP hydrolysis and to investigate
whether cUMP and cUMP-hydrolyzing PDEs are present in
cardiomyocytes. PDE3A-mediated cUMP hydrolysis was
characterized in time course, inhibitor, and Michaelis-
Menten kinetics experiments. Intracellular cyclic nucleotide
PDE3 inhibitor milrinone inhibited cUMP hydrolysis (deter-
mination of UMP formation) by PDE3A (K = 57 nM).
i
Significant amounts of cUMP as well as of PDE3A mRNA
(in addition to PDE3B and PDE9A transcripts) were detected
in HL-1 cells and NRCMs. Although HL-1 cell homogenates
contain a milrinone-sensitive cUMP-hydrolyzing activity, in-
tact HL-1 cells may use additional PDE3-independent mech-
anisms for cUMP disposal. PDE3A is a low-affinity and high-
velocity PDE for cUMP. Future studies should investigate
biological effects of cUMP in cardiomyocytes and the role
of PDE3A in detoxifying high intracellular cUMP concentra-
tions under pathophysiological conditions.
(
cNMP) concentrations and the mRNAs of cUMP-degrading
PDEs were quantitated in neonatal rat cardiomyocytes
NRCMs) and murine HL-1 cardiomyogenic cells.
(
Moreover, we investigated cUMP degradation in HL-1 cell
homogenates and intact cells. Educts (cNMPs) and products
.
.
.
Keywords Cyclic UMP Phosphodiesterase PDE3A
.
(
NMPs) of the PDE reactions were detected by HPLC-
Neonatal rat cardiomyocytes HL-1 cells
coupled tandem mass spectrometry. PDE3A degraded
cUMP (measurement of UMP formation) with a KM value
of ~143 μM and a Vmax value of ~42 μmol/min/mg. PDE3A
hydrolyzed cAMP with a K value of ~0.7 μM and a V of
Introduction
M
max
~
1.2 μmol/min/mg (determination of AMP formation). The
The cyclic purine nucleotides cAMP and cGMP are well-
established intracellular signaling molecules and fulfill all
criteria of classic second messengers (Seifert 2015).
Although there was some initial interest in the cyclic pyrimi-
dine nucleotides cUMP and cCMP, research in this field was
severely hampered by the lack of sensitive and selective de-
tection methods (Seifert et al. 2015). With the advent of highly
sensitive and specific mass-spectrometry-based methods,
however, detection and quantitation of low concentrations of
cUMP and cCMP became possible even in complex cellular
matrices (Seifert et al. 2015). HPLC-coupled tandem mass
spectrometry (HPLC-MS/MS) has revealed that significant
concentrations of cUMP and cCMP are present in mammalian
cells (Hartwig et al. 2014) and are generated by soluble
guanylyl cyclase (sGC) (Bähre et al. 2014) or soluble adenylyl
cyclase (sAC) (Hasan et al. 2014). Bacterial toxins such as
Electronic supplementary material The online version of this article
doi:10.1007/s00210-016-1328-1) contains supplementary material,
which is available to authorized users.
(
*
Erich H. Schneider
schneider.erich@mh-hannover.de
1
2
Institute of Pharmacology, Hannover Medical School,
Carl-Neuberg-Str. 1, D-30625 Hannover, Germany
Research Core Unit Metabolomics, Institute of Pharmacology,
Hannover Medical School, Carl-Neuberg-Str. 1,
D-30625 Hannover, Germany
3
Molecular Cardiology Research Group, Department of Cardiology
and Angiology, Hannover Medical School, Carl-Neuberg-Str. 1,
D-30625 Hannover, Germany

2
70
Naunyn-Schmiedeberg's Arch Pharmacol (2017) 390:269–280
CyaA from Bordetella pertussis (Göttle et al. 2010) and ExoY
from Pseudomonas aeruginosa (Beckert et al. 2014) are ca-
pable of producing cCMP and cUMP. Infection with
P. aeruginosa results in high cUMP production in the mouse
model, suggesting a pathophysiological role of this cyclic nu-
cleotide (Bähre et al. 2015). In fact, the protein kinases PKA
and PKG are stimulated by high concentrations of cCMP or
cUMP in vitro (Wolter et al. 2011), and cCMP activates HCN
channels with low potency and as a partial agonist (Zong et al.
tenofovir was obtained from the National Institute of Health
(NIH) AIDS Research and Reference Reagent Program
(Germantown, MD, USA). Claycomb medium was purchased
from Sigma Aldrich and supplemented with 0.1 mM of norepi-
nephrine by adding an appropriate volume of a 10-mM norepi-
nephrine stock solution containing 30 mM of ascorbic acid.
Additionally, a mixture of penicillin, streptomycin, and L-glu-
tamine as well as 10% fetal bovine serum (FBS) were added as
previously described (Claycomb et al. 1998). All PDE reactions
were performed in 1× PDE buffer containing 50 mM of Tris–
HCl (pH 7.5), 8.3 mM of magnesium chloride, and 1.7 mM of
EDTA. For the experiments with recombinant enzyme, the 1×
PDE buffer was additionally supplemented with 0.05% BSA.
All other reagents were analytical grade and from standard
suppliers.
2
012). Moreover, cCMP induces caspase-dependent apopto-
sis of mouse lymphoma cells (Wolter et al. 2015). These ex-
amples support the notion that cUMP and/or cCMP produc-
tion by bacterial toxins may exert biological effects and dam-
age host cells.
Thus, it is important to elucidate the inactivation mecha-
nisms of cyclic pyrimidine nucleotides. Multidrug resistance
proteins (MRPs) 4 and 5 export cCMP and/or cUMP (Laue
et al. 2014). A screening of 13 recombinant phosphodiester-
ases (PDEs) revealed only one enzyme (PDE7A) that hydro-
lyzes cCMP, but three enzymes (PDE3A, 3B, and 9A) that
degrade cUMP with significant activity (Reinecke et al.
Time course of PDE3A activity for cAMP and cUMP
Recombinant GST-tagged PDE3Awas added to a solution con-
taining 3 μM of cAMP or cUMP in 1× PDE buffer supplement-
ed with 0.05% of BSA, yielding a final volume of 300 μl and a
final enzyme concentration of 0.2 µg/ml. All samples were run
in duplicates. The total incubation time was 90 min at 30 °C
2
011; Monzel et al. 2014). Recently, our research group has
characterized the enzyme kinetics of PDE7A-mediated cCMP
hydrolysis, revealing that PDE7A1/2 is a low-affinity and
high-velocity enzyme for cCMP (Monzel et al. 2014). In this
paper, we report on the detailed characterization of PDE3A-
mediated cUMP degradation. Moreover, we demonstrate the
presence of significant amounts of cUMP in cardiomyocytes
®
under constant shaking at 300 rpm (Eppendorf Thermomixer
comfort). As 0 min value, the negative control without enzyme
was used. During the incubation time, aliquots of 40 μl were
drawn at defined times and instantly heated to 95 °C for 15 min
to inactivate PDE3A and precipitate protein. After freezing the
samples at −20 °C overnight, they were thawed and centrifuged
(15 min, 20,800×g, 4 °C) to remove precipitated protein. Ten
microliters of the supernatant was diluted 1:4 with purified
water. An equal volume of tenofovir solution (100 ng/ml) was
added to yield a final tenofovir concentration of 50 ng/ml and a
final volume of 80 μl. Degradation of the substrate (cAMP or
cUMP) as well as accumulation of the product (AMP or UMP)
was determined by high-performance liquid chromatography
tandem mass spectrometry (HPLC-MS/MS) as previously de-
scribed (Monzel et al. 2014).
(
neonatal rat cardiomyocytes and murine cardiomyogenic
HL-1 cells) and show that cUMP hydrolysis in HL-1 homog-
enates is sensitive to milrinone. Our results strongly suggest
that the cUMP-degrading PDE activity detected in cow and
dog heart tissue 50 years ago (Hardman and Sutherland 1965)
was at least partially caused by cardiac PDE3A.
Materials and methods
Enzymes, reagents, and buffers
Michaelis-Menten analysis of PDE3A-mediated cUMP
and cAMP degradation
Truncated (amino acids 484–1141) recombinant N-terminally
GST-tagged human PDE3A (lot #130312-G1; GenBank ac-
cession no. NM_000921) with a purity of ≥25%
For the determination of KM and Vmax values of PDE3A-
mediated cAMP and cUMP hydrolysis, substrate concentra-
tions between 0.1 and 6 μM (cAMP) and between 1 μM and
1.5 mM (cUMP) were used in 1× PDE buffer supplemented
with 0.05% of BSA. The reaction was started by adding
PDE3A to yield a final concentration of 0.1 μg/ml and a final
sample volume of 50 μl. Negative controls without PDE were
run in parallel for each substrate concentration. All samples
were run in duplicates and incubated for 15 min (cUMP) or
2 min (cAMP) at 30 °C under constant shaking (300 rpm).
After that, 40 μl aliquots were drawn, heat-inactivated, and
(
http://bpsbioscience.com/pde3a-484-end-60030) and a
specific activity of 1100 pmol/min/μg was obtained from
BPS Bioscience (San Diego, CA, USA). 3′,5′-cAMP, 3′,5′-
cCMP, 3′,5′-cGMP, and 3′,5′-cUMP were provided by Sigma
Aldrich (Steinheim, Germany). The acetoxymethyl (AM) ester
of cUMP, cUMP-AM, was obtained from Biolog Life Science
Institute (Bremen, Germany). The PDE3-selective inhibitor
milrinone (lot #23922111; LKT-M3344) was purchased from
Biomol GmbH (Hamburg, Germany) and dissolved in DMSO
to yield a 200-mM stock solution. The internal standard

Naunyn-Schmiedeberg's Arch Pharmacol (2017) 390:269–280
271
frozen overnight as described in the section about the time
course experiments.
section about cell culture conditions. Cells were trypsinized,
pelleted at 300×g, and washed three times in PBS (1×). After
that, they were re-suspended in BSA-free 1× PDE buffer supple-
mented with a protease-inhibitor mixture consisting of
benzamidine (5 mM), phenylmethylsulfonylfluoride (PMSF;
200 μM), and leupeptin (57.4 μM). After an incubation of
10 min, cells were either homogenized by applying ~50 strokes
in a 5-ml Dounce homogenizer or by ~30 s of sonification
(Branson Sonifier Analog 250). All steps were performed on
ice. After one freeze-thaw cycle, protein concentrations were
In the next step, the samples were thawed and centrifuged
(
15 min, 20,800×g, 4 °C) to remove precipitated protein. In
case of samples with less than 10 μM of cNMPs, 10 μl of the
supernatant was diluted 1:4 with purified water. For cNMP
concentrations of more than 10 μM and less than 0.5 mM,
the supernatant was diluted 1:200 (1:50, followed by 1:4) with
purified water. Substrate concentrations higher than 0.5 mM
were diluted 1:400 (1:100, followed by 1:4). Forty microliters
of these dilutions were mixed 1:1 with tenofovir solution
®
determined using the Pierce BCA Protein Assay from
(
100 ng/ml) to yield a final tenofovir concentration of 50 ng/ml
Thermo Scientific (Rockford, IL, USA) according to the manu-
facturer’s instructions. The cell homogenates were stored at
−80 °C.
and a final volume of 80 μl. The HPLC-MS/MS measurement
was performed as previously described (Monzel et al. 2014).
Inhibition of PDE3A-mediated cUMP and cAMP
hydrolysis by milrinone
Basal cUMP content (HL-1 cells or NRCMs)
and PDE3-mediated cUMP hydrolysis in cell homogenates
(HL-1 cells)
The reaction was started by adding PDE3A to 1× PDE buffer
containing appropriate milrinone concentrations (0–10 μM),
The experiments with HL-1 cell homogenates (prepared as
described in the preceding section) were run in duplicate in
BSA-free PDE buffer and at a final protein concentration of
250 μg/ml. The samples contained 3 μM of substrate (cAMP
or cUMP) and either 10 μM of milrinone or 0.005% of DMSO
(negative control). Aliquots were drawn right at the beginning
(0 min) and after 6 h of incubation at 30 °C (constant shaking
at 300 rpm). The tubes were centrifuged approximately every
30 min to minimize the error caused by condensation of liquid
in the lid. Afterwards, samples were stopped, processed, and
analyzed as described in the section about the time course
experiments.
0
.05% DMSO, and cAMP (1 μM) or cUMP (3 μM). The final
PDE3A concentration was 0.2 μg/ml (cUMP samples) or
.1 μg/ml (cAMP samples). The sample volume was 50 μl
and all samples were run in duplicate. After an incubation for
0 min (cUMP) or 15 min (cAMP) at 30 °C, the samples were
0
1
stopped, processed, and analyzed as described in the section
about the time course experiments. The HPLC-MS/MS mea-
surement was performed as previously described (Monzel
et al. 2014).
Cell culture
For determination of intracellular cNMP concentrations,
HL-1 cells were grown to 70–80% confluence in six-well
plates. The medium was removed and 300 μl of ice-cold ex-
traction solution (acetonitrile/methanol/water; 2:2:1; v/v/v)
containing 25 ng/ml tenofovir as internal standard was added.
The cells were rigorously scraped off and the suspension was
transferred to 2-ml reaction tubes. The wells were washed
twice with 400 μl of ice-cold extraction solution and the sus-
pension from both washing steps was also transferred to the 2-
ml reaction tube, yielding a final volume of 1.1 ml. NRCMs
were isolated from neonatal rat brains and seeded in 6-cm
dishes. On the fourth or fifth day of cultivation, the medium
was removed and 500 μl of ice-cold extraction solution with-
out tenofovir plus 300 μl of extraction solution containing
25 ng/ml tenofovir were added. The following steps were
performed as described for HL-1 cells, but with 500 μl of
extraction solution for washing, yielding a final volume of
The mouse cardiomyogenic cell line HL-1 was cultured in
supplemented Claycomb medium as previously described
(
Claycomb et al. 1998). The media was changed every 24 to
8 h. Prior to seeding the cells, cell culture flasks and dishes
were coated with a dilution of 0.005 mg/ml fibronectin in
.02% gelatin for at least 60 min at 37 °C. Neonatal rat
4
0
cardiomyocytes (NRCM) were isolated as described previous-
ly (Hilfiker-Kleiner et al. 2004) and cultured in 60 × 15 mm
cell culture dishes coated with 1% gelatin at a density of 1.2–
6
1
.5 × 10 cells/dish in Dulbecco’s modified Eagle’s medium
(
DMEM) supplemented with 10% Medium 199, 5% fetal calf
serum (FCS), and 10% horse serum. 3T3-L1-MBX
preadipocyte cells were obtained from LGC Standards/
ATCC (Wesel, Germany) and cultured in DMEM supplement-
ed with 10% neonatal calf serum (NCS), penicillin, strepto-
mycin, and L-glutamine.
1
.8 ml. The suspensions from the extraction of HL-1 cells or
NRCMs were heated to 95 °C for 15 min and then stored at
20 °C overnight. After thawing and centrifuging for 10 min
Preparation of cell homogenates
−
Both HL-1 cells and 3T3-L1-MBX cells were cultured until 70–
at 20,800×g, the supernatant was transferred into a new tube
and evaporated completely at 40 °C under a constant nitrogen
2
80% confluence in 175 cm cell culture flasks as described in the

2
72
Naunyn-Schmiedeberg's Arch Pharmacol (2017) 390:269–280
stream. The residue was dissolved in 150 μl of water. The
cNMP measurement was performed as previously described
stacking and 7.5% resolution gel), which was run at 25 mA
for 10 min and at 35 mA for 60 min. The protein was trans-
2
(
Hasan et al., 2014). The cNMP content was normalized to the
ferred to a PVDF membrane by semi-dry blotting (2 mA/cm ,
®
protein concentration of each sample, which was determined
90 min). After that, the membrane was blocked in Roti
®
using the Pierce BCA Protein Assay from Thermo Scientific
Block protein-free blocking solution (Roth, Karlsruhe,
Germany) for 60 min at ambient temperature and then incu-
bated overnight at 4 °C with a rabbit monoclonal antibody
against PDE3A (EPR11601, cat. #ab169534; Abcam,
(
Rockford, IL, USA) according to the manufacturer’s
instructions.
®
Cambridge, MA, USA) at a 1:1000 dilution in Roti Block.
Effect of milrinone on cNMP concentrations in intact
HL-1 cells after pre-incubation with cUMP-AM
On the next day, the membrane was washed 3 × 7 min with 1×
TBST and then incubated for 120 min at ambient temperature
with an HRP-linked goat-anti-rabbit secondary antibody
HL-1 cells were grown to 70–80% confluence in six-well
plates. On the day of the experiment, the cells were pre-
incubated with 10 μM of cUMP-AM or with 0.1% DMSO
(
#7074S; Cell Signaling Technology, NEB Frankfurt,
®
Germany), diluted 1:3000 in Roti Block. Finally, the blot
was developed by applying a 1:1 mixture of Western Sure
Luminol Enhancer Solution and Western Sure Stable
Peroxidase Solution (LI-COR Biotechnology, Bad
Homburg, Germany). Luminescence was detected using a
(
control) for 5 min. Then the medium was aspirated and the
cells were washed with PBS once. After that, supplemented
Claycomb medium containing either 100 μM of milrinone or
0
.5% DMSO (control) was added to the cells. After incubation
®
C-Digit Blot scanner (LI-COR Biotechnology).
for predefined times (0, 30, 60, and 120 min) at 37 °C and 5%
CO , the medium was removed and 300 μl of ice-cold extrac-
2
Data analysis and statistics
tion solution (acetonitrile/methanol/water; 2:2:1; v/v/v) con-
taining 25 ng/ml tenofovir as internal standard were added.
The samples were processed as described in the preceding
section.
HPLC-MS/MS-data were analyzed with the Analyst Software
(
ABSciex; Framingham, USA), and the results were exported
®
to Microsoft Excel (Microsoft; Redmond, USA). The area
of every cNMP and/or NMP chromatography peak was deter-
mined for the samples as well as for appropriate standard
curves containing AMP, cAMP, UMP, and cUMP (0, 0.24,
Quantitative real-time PCR (qPCR)
Preparation of RNA from HL-1 cells, nRCMs, and 3T3-L1-
MBX cells was performed using the NucleoSpin RNA II kit
0
.49, 0.98, 1.95, 3.9, 7.8, 15.6, 31.3, 62.5, and 125 pmol) or
all four cNMPs and their 5′-NMP hydrolysis products
0.0262, 0.0655, 0.164, 0.410, 1.024, 2.56, 6.4, 16, 40, 100,
(Macherey-Nagel, Düren, Germany) according to the manufac-
(
turer’s instructions. The RNA was reverse-transcribed into
®
and 250 pmol). Using GraphPad Prism 6 (San Diego, CA,
USA), the individual analyte amounts were calculated from
the peak areas. In order to minimize matrix effects, the peak
area of the analyte was related to the peak area of the internal
standard tenofovir in studies with cell homogenates. For the
analysis of enzyme kinetics data, tenofovir was not included.
®
cDNA and analyzed using TaqMan probes for the expression
of mouse (mActb; Mm00607939_s1; ThermoFisher Scientific,
Frankfurt/Main, Germany) and rat (rActb; Rn00667869_m1) β-
actin as well as mouse and rat PDE3A (mouse—
Mm00479581_m1; rat—Rn00569192_m1), PDE3B (mouse—
Mm00691635_m1; rat—Rn00568191_m1), and PDE9A
Data plots and calculations of K , Vmax, and IC50 values were
performed with GraphPad Prism 6 (San Diego, USA).
M
(
mouse—Mm00501039_m1; rat—Rn00593577_m1).
®
®
Reactions were performed with TaqMan Gene Expression
Master Mix (Applied Biosystems™, Frankfurt, Germany). The
mRNA expression levels of the genes were compared after cal-
Results
culating the ΔC values:
T
Time course of PDE3A activity
ΔCT ¼ CT ½target geneŠ− CT ½m or rActbŠ
In order to assess stability and activity of the recombinant
PDE3A used in our experiments, we determined the time
course of PDE3A-mediated cAMP and cUMP hydrolysis.
cAMP (3 μM) was fully hydrolyzed within 20 min (Fig. 1a)
at a constant velocity of about 0.86 μmol/min/mg (calculated
from the first 10 min of the curve) (Fig. 1a, c). By contrast,
about 60 min were required for complete hydrolysis of 3 μM
of cUMP. Interestingly, the initial velocity of cUMP
Western blots
After diluting the samples with 5× Lane Marker Reducing
Sample Buffer
Germany), they were heated to 96 °C for 10 min. Then, the
samples were loaded on an SDS polyacrylamide gel (5%
TM
(#39000; Thermo Scientific, Bremen,

Naunyn-Schmiedeberg's Arch Pharmacol (2017) 390:269–280
273
hydrolysis (0.91 μmol/min/mg) was comparable to the initial
cAMP degradation rate (Fig. 1b, c). After 10 min, however,
cUMP hydrolysis slows down significantly (Fig. 1b).
neonatal rat cardiomyocytes (NRCMs) as model cell systems
to investigate cUMP metabolism in these cells. HPLC-MS/
MS analysis revealed the presence of all four analyzed
cNMPs in both cell systems. The most abundant cNMP was
cAMP with a concentration of 36.3 ± 10.7 pmol/mg in HL-1
cells and 200.8 ± 81.8 pmol/mg in NRCMs (n = 4,
means ± SD) (Figs. 4a and 4b). The cUMP concentration in
HL-1 cells was 3.1 ± 2.0 pmol/mg and was comparable to that
of the established second messenger cGMP that amounted to
2.6 ± 1.0 pmol/mg (n = 4, means ± SD). Similarly, in NRCMs,
the cGMP concentration was 10.8 ± 2.8 pmol/mg, while
cUMP occurred at a concentration of 9.0 ± 2.4 pmol/mg
(n = 4, means ± SD). In both cell types, the least abundant
cNMP was cCMP (HL-1 = 1.6 ± 1.0 pmol/mg;
NRCMs = 4.8 ± 1.4 pmol/mg; means ± SD from n = 4). The
original datasets for all experiments are shown in
Supplementary Table 3.
Michaelis-Menten kinetics of PDE3A for cUMP
and cAMP
For the PDE3A/cAMP Michaelis-Menten kinetics, 0.1–6 μM
of substrate was applied (enzyme concentration 0.1 μg/ml).
Calculation of KM and Vmax from the cAMP hydrolysis data
(
Fig. 2a, Table 1) yielded 0.26 μM and 1.30 μmol/min/mg,
respectively (n = 2 in duplicates). The KM value of AMP
formation (Fig. 2b, Table 1) was 0.71 μM with a Vmax of
1
.15 μmol/min/mg (n = 2 in duplicates). The Michaelis-
Menten kinetics of PDE3A-mediated cUMP hydrolysis was
determined at substrate concentrations between 1 μM and
1
.5 mM (enzyme concentration 0.1 μg/ml). Hydrolysis of
cUMP (Fig. 2c) occurred with a KM value of 98.8 μM and a
Vmax of 37.6 μmol/min/mg. The UMP accumulation data
(
Fig. 2d) yielded a KM value of 142.9 μM and a Vmax value
Expression of cUMP-degrading PDEs in HL-1 cells
and neonatal rat cardiomyocytes
of 42.0 μmol/min/mg (n = 4 in duplicates). These results are
listed in Table 1. The complete datasets for cUMP hydrolysis
and UMP accumulation in the Michaelis-Menten kinetics ex-
periments are shown in Supplementary Tables 1 and 2.
Next, we addressed the question which of the previously de-
scribed (Reinecke et al. 2011) cUMP-hydrolyzing PDEs are
expressed in cardiomyocytes. We isolated RNA from HL-1
cells and NRCMs and analyzed transcription of PDE3A,
PDE3B, and PDE9A genes (Fig 5). The mRNA for all three
cUMP-degrading enzymes was present in HL-1 cells (n = 3-4
Inhibition of PDE3A-mediated cUMP and cAMP
hydrolysis by the PDE3-selective inhibitor milrinone
Milrinone, a selective inhibitor of PDE3, was used at concen-
trations between 1 nM and 100 μM to inhibit PDE3A-
mediated hydrolysis of cUMP (3 μM) and cAMP (1 μM).
As expected, milrinone inhibited cAMP degradation by
PDE3A (Fig. 3a), yielding IC50 values of 899 nM (cAMP
hydrolysis) and 1075 nM (AMP formation) (Table 1).
Milrinone also reduced both PDE3A-mediated cUMP hydro-
lysis and UMP accumulation (Fig. 3b) with IC50 values of
independent RNA isolations, means ± SD) with ΔC values
T
of 10.94 ± 1.36 (PDE3A), 7.96 ± 1.17 (PDE3B), and
13.35 ± 1.49 (PDE9A). Moreover, these mRNAs were detect-
ed in NRCMs (n = 2 independent RNA isolations,
means ± SD) with ΔC values of 6.70 ± 0.31 (PDE3A),
T
7.16 ± 0.31 (PDE3B), and 14.38 ± 0.22 (PDE9A).
According to the literature, 3T3-L1 fibroblasts show low
PDE3B expression, which is only increased upon differentia-
tion to adipocytes (Taira et al. 1993; Niiya et al. 2001). Thus,
RNA from 3T3-L1-MBX fibroblasts (n = 2-3 independent
RNA isolations) was used as a negative control for PDE3B,
showing an approximate ΔC value of 16.5. The ΔC values
1
22 ± 99.6 nM and 57.5 ± 37.3 nM, respectively (n = 3,
means ± SEM). The K values were calculated by applying
i
the Cheng-Prusoff equation (Cheng and Prusoff 1973), using
the corresponding K values of the substrates (see preceding
M
T
T
section and Table 1). For inhibition of PDE3A-mediated
for PDE3A (18.3) and PDE9A (18.5) suggest that 3T3-L1
cells are also suited as negative controls for PDE3A and 9A.
cAMP hydrolysis, the K values of milrinone were 186 nM
i
(
cAMP hydrolysis) and 446 nM (AMP formation). Hydrolysis
All ΔC values were calculated using mActb as housekeeping
T
of cUMP was inhibited by milrinone with a K value of 56 nM
gene.
i
(
UMP formation) and 118 nM (cUMP hydrolysis) (Table 1).
We also tried to detect PDE3A on the protein level by
Western blotting with a monoclonal rabbit anti-PDE3A anti-
body. In Hl-1 cells, the antibody stained two bands around
130 kDa (Supplementary Fig. 1A and B, red boxes) that might
correspond to PDE3A isoforms. A 118-kDa and a 136-kDa
PDE3A isoform have been previously reported for human
myocardium (Wechsler et al. 2002). However, the antibody
also stained numerous bands below 100 kDa, indicating a lack
of selectivity. Moreover, a band close to 130 kDa was also
Basal cNMP levels in HL-1 cells and neonatal rat
cardiomyocytes
PDE3A is expressed in heart (Omori and Kotera 2007) and the
presence of a cUMP-hydrolyzing activity has been previously
reported in cow and dog hearts (Hardman and Sutherland
1
965). Thus, we chose HL-1 cardiomyogenic cells and

2
74
Naunyn-Schmiedeberg's Arch Pharmacol (2017) 390:269–280
Fig. 1 Time course of PDE3A-mediated hydrolysis of 3′,5′-cAMP and
′,5′-cUMP. a Degradation of cAMP (3 μM) and formation of AMP by
purified GST-tagged PDE3A; b degradation of cUMP (3 μM) and
formation of UMP by purified GST-tagged PDE3A; c comparison of
the initial linear parts of the curves for PDE3A-mediated cAMP and
cUMP hydrolysis; Data are means ± SD from n = 3 independent
experiments performed in duplicates
3
detected in cell homogenates of HepG2 liver carcinoma cells,
U937 promonocytes, and 3T3-L1 pre-adipocytes.
the antibody. According to the manufacturer’s datasheet, a
Coomassie-stained SDS-PAGE of this preparation yielded nu-
merous bands below 100 kDa and purity was given as ≥25%,
which supports the notion that the preparation contains major
impurities.
A clear staining between 100 and 130 kDa was observed
with the recombinant enzyme preparation used for the PDE3A
enzyme kinetics experiments in this paper (Supplementary
Fig. 1A and B, red boxes). According to the manufacturer’s
specifications, the enzyme is truncated and bears an N-
terminal GST tag, resulting in a net molecular mass of
Inhibition of cUMP hydrolysis in HL-1 cell homogenates
by milrinone
1
03 kDa. In addition to the high molecular mass signal, how-
ever, the antibody also produced strong staining between 70
and 100 kDa in the recombinant protein sample
In the next step, we addressed the question whether the PDE3
isoforms constitutively expressed in HL-1 cardiomyogenic
cells are capable of hydrolyzing cUMP. This question is specif-
ically interesting because more than 50 years ago, a cUMP-
hydrolyzing PDE activity was reported in beef and dog heart
(
Supplementary Fig. 1A, lanes 2–4). It is unclear if this stain-
ing is due to PDE3A degradation products or caused by a
totally different protein that coincidentally cross-reacts with
Fig. 2 Michaelis-Menten
kinetics of PDE3A-mediated
cNMP hydrolysis. Velocity of
PDE3A-mediated a cAMP
degradation and b AMP
formation in the presence of
increasing cAMP concentrations;
velocity of PDE3A-mediated c
cUMP degradation and d UMP
formation in the presence of
increasing concentrations of
cUMP. Data are means ± SD from
n = 4 (c, d) or n = 2 (a, b)
independent experiments in
M
duplicates. The mean K and
Vmax values are listed in Table 1.
Moreover, the complete datasets
for the cUMP Michaelis-Menten
kinetics are shown in
Supplementary Table 1 (UMP
accumulation) and
Supplementary Table 2 (cUMP
hydrolysis)

Naunyn-Schmiedeberg's Arch Pharmacol (2017) 390:269–280
275
Table 1. Characterization of GST-tagged PDE3A in terms of cUMP and cAMP degradation and UMP and AMP formation (data are means ± SEM,
graphs are shown in Figs. 2 and 3)
Parameter
max [μmol/(min mg)]
(μM)
UMP formation
42.0 ± 2.8
cUMP hydrolysis
AMP formation
1.15
cAMP hydrolysis
1.30
V
37.6 ± 5.2
(n = 4)
a
a
(
n = 4)
142.9 ± 36.1
n = 4)
57.5 ± 37.3
n = 3)
56
(1.27, 1.03)
(1.85, 0.76)
K
M
98.8 ± 36.4
0.71
(0.67, 0.75)
0.26
(0.41, 0.11)
a
a
(
(n = 4)
b
b
IC50 (milrinone) (nM)
122 ± 99.6
1075
(925, 1224)
899
(1099, 699)
a
a
(
(n = 3)
c
d
d
K
i
(milrinone) (nM) (calculated from IC50
)
118
446
186
a
For experiments with n = 2, the results from the individual experiments were used to calculate the mean value. The individual results are given below in
brackets
b
The logarithmic error from the GraphPad prism output was converted to the linear error given in the table by multiplying the log error with the non-
logarithmic mean value and with ln10
c
Calculated with the Cheng-Prusoff equation by using the corresponding K
M
values
d
These K
i
values may be overestimated due to substrate depletion (as explained in the discussion of this article)
tissue (Hardman and Sutherland 1965). We incubated HL-1 cell
homogenates with 3 μM of cUMP for a total time of 6 h in the
presence and absence of milrinone and analyzed the cNMP
content by HPLC-MS/MS (n = 4 homogenates, tested in dupli-
cates). HL-1 cell homogenate hydrolyzed a small but signifi-
cant amount of cUMP (~13%) within 6 h, which was signifi-
cantly inhibited by milrinone. Even after 6 h, the initially added
cUMP concentration was only reduced by ~3%, when the
PDE3 inhibitor was present (Fig. 6). By contrast, cAMP was
completely degraded by the HL-1 cell homogenate during the
5 min with 10 μM of cUMP-AM (or DMSO as solvent con-
trol). Then, the cUMP-AM-containing incubation medium
was removed, the cells were washed once and then incubated
for 120 min in the presence of 100 μM of milrinone or solvent
(DMSO, negative control). The time course of intracellular
cUMP concentrations was followed by HPLC-MS/MS, yield-
ing the data depicted in Fig. 7. Since no difference between
milrinone-containing and milrinone-free samples is to be ex-
pected at t = 0 min, both sets of samples were merged to
compare the average initial cUMP concentrations. At
t = 0 min, the cUMP-AM pre-treated samples had a cUMP
content of 41.1 ± 34.1 pmol/mg (n = 8, mean ± SD) (Fig. 7a),
while the DMSO controls contained only 3.7 ± 0.4 pmol/mg
of cUMP (n = 6, mean ± SD) (Fig. 7c). This indicates that
cUMP-AM pre-incubation significantly increased intracellu-
lar cUMP (p < 0.05, unpaired two-tailed t test).
6-h incubation time. This was not prevented by milrinone (data
not shown), probably due to the presence of other cAMP-
hydrolyzing but milrinone-insensitive PDEs.
Inhibition of cUMP hydrolysis by milrinone in intact HL-1
cells
Intracellular cUMP concentrations dramatically decreased
within 30 min in all cUMP-AM-pre-loaded samples (Fig. 7a,
b). Since the initial cUMP concentration after pre-incubation
Next, we investigated whether milrinone inhibits cUMP hy-
drolysis in living HL-1 cells. The cells were incubated for
Fig. 3 Inhibition of PDE3A-mediated cAMP and cUMP hydrolysis by
milrinone. Comparison of a degradation of 1 μM of cAMP (filled circles)
and formation of AMP (open circles) as well as b degradation of 3 μM of
cUMP (filled squares) and formation of UMP (open squares) by purified
GST-tagged PDE3A in the presence of increasing milrinone
concentrations. Data are means ± SD from n = 2 (cAMP) or n = 3
(cUMP) independent experiments performed in duplicates. The mean K
values for inhibition of cUMP or cAMP degradation are listed in Table 1
i

2
76
Naunyn-Schmiedeberg's Arch Pharmacol (2017) 390:269–280
Fig. 4 Basal cNMP concentrations in HL-1 cardiomyogenic cells (a) and
NRCMs (b), related to the protein amounts determined by the BCA
method. Data are means ± SD from n = 4 independent experiments
with 2–4 replicates. The mean values of the individual experiments are
given in the text of the “Results” section. The complete datasets are given
in Supplementary Table 3.
with cUMP-AM shows high inter-experimental variability
Discussion
(
error bar at t = 0 min in Fig. 7a), all values were normalized
to the cUMP concentration at t = 0 min (Fig. 7b). The DMSO
controls were transformed the same way (Fig. 7c vs. d). The
relative cUMP concentrations after 30, 60, and 120 min were
significantly reduced as compared to t = 0 min in both the
milrinone-containing (p < 0.001) and the milrinone-free
A detailed enzyme kinetics analysis of PDE3A-mediated
cUMP hydrolysis was provided. The initial hydrolysis rate in
time course experiments was comparable for 3 μM of cAMP or
cUMP. After 20 min, however, the reaction speed in the cUMP
samples declined, while cAMP hydrolysis continued at a con-
stant rate. This may be due to largely different KM values.
PDE3A is saturated by 3 μM of cAMP and therefore works
at maximum speed throughout the experiment. By contrast, due
to a higher cUMP KM value, the enzyme is not saturated by
3 μM of cUMP and substrate depletion quickly causes a reduc-
tion of cUMP hydrolysis rate. Alternatively, the cUMP degra-
dation could be slowed down by product inhibition or limited
enzyme stability. The Michaelis-Menten kinetics data, howev-
er, support the notion that PDE3A shows a much lower affinity
(
p < 0.01) samples (one-way ANOVA followed by
Dunnett’s multiple comparison test). After 120 min, the
cUMP concentrations had stabilized at 10.5 ± 4.2 pmol/mg
(
with milrinone) and at 9.2 ± 4.5 pmol/mg (without milrinone)
in the cUMP-AM-pre-incubated samples. In the solvent con-
trols, the 120-min values were at 4.3 ± 2.2 pmol/mg (with
milrinone) and at 3.9 ± 2.6 pmol/sample (without milrinone).
In addition to cUMP, we also quantified cAMP, cCMP,
and cGMP in all samples (data not shown). Pre-incubation
of the cells with cUMP-AM had no significant effect on
cAMP, cGMP, and cCMP contents, and milrinone caused
only a slight non-significant increase of intracellular cAMP
for cUMP than for cAMP. An about 200-fold higher K value
M
and a 30-fold higher maximum reaction speed (Vmax) of
PDE3A for cUMP as compared to cAMP (product accumula-
tion data) indicate that PDE3A is a low-affinity and high-
(
data not shown).
Fig. 5 Analysis of PDE3A, PDE3B, and PDE9A gene expression in HL-
murine cardiomyogenic cells (n = 3-4 independent RNA isolations),
NRCM (n = 2 independent RNA isolations), and murine 3T3-L1-MBX
n = 2-3 independent RNA isolations) pre-adipocytes. a PDE3A, b
PDE3B, c PDE9A. Data were analyzed using the ΔC method with
actin as a housekeeping gene. Please note the inverted relationship
T T
between expression level and ΔC value: a lower ΔC signifies a
1
higher amount of transcript. The data points represent the individual
samples from 2–4 independent RNA isolations. The number of qPCR
samples (measured in duplicates) for each cell type and PDE is given in
round parentheses below the x-axis
(
T

Naunyn-Schmiedeberg's Arch Pharmacol (2017) 390:269–280
277
velocity enzyme for cUMP. Similar properties were previously
observed for PDE7A that shows a 1000-fold lower substrate
affinity, but a 6–7-fold higher Vmax for cCMP as compared to
cAMP (Monzel et al. 2014).
The KM values determined for PDE3A-mediated cAMP hy-
drolysis (Table 1) are in good agreement with the K of 0.18 μM
M
reported in the literature (Bender and Beavo 2006). The Vmax
values determined for cAMP degradation by recombinant
PDE3A (Table 1) are close to the literature range of 3–6 μmol/
min/mg (Bender and Beavo 2006), specifically when it is consid-
ered that, according to the manufacturer’s specifications, our en-
zyme preparation contained up to 75% of non-PDE3A protein.
This confirms that the recombinant enzyme used in our experi-
ments is fully functional. It could be argued, however, that the
observed cUMP hydrolysis was not caused by PDE3A, but by a
contamination, because the manufacturer’s specifications state a
purity of only ≥25%, and our Western blots also show a major
band between 70 and 100 kDa (Supplementary Fig. 1). However,
the PDE3-selective inhibitor milrinone (Ito et al. 1988) potently
inhibited cUMP hydrolysis in our samples, which strongly indi-
Fig. 6 Milrinone-sensitive cUMP-degrading activity in HL-1 cell
homogenates. Depicted are cUMP concentrations in HL-1 cell
homogenate samples in the beginning and at the end of a 6-h
incubation period with 3 μM of cUMP. Milrinone (10 μM)-containing
samples were compared with negative controls containing solvent
(
DMSO). Data are means ± SD from n = 4 different HL-1 cell
homogenates tested in duplicates. The homogenates were independently
prepared on three different days. Two of the homogenates were prepared
with a Dounce homogenizer. The other two homogenates were prepared
using a sonifier. **p < 0.05, one-way ANOVAwith Bonferroni’s multiple
comparison test
cates the presence of PDE3A. Moreover, the K values for
i
milrinone with respect to inhibition of PDE3A-mediated cUMP
hydrolysis (Table 1) correspond well to the K value of 150 nM
i
reported in the literature (inhibition of cardiac PDE3-mediated
cAMP hydrolysis) (Ito et al. 1988). This suggests that cUMP
Fig. 7 Milrinone effect on cNMP metabolism in intact HL-1
cardiomyogenic cells. Cells were pre-incubated for 5 min with 10 μM
of cUMP-AM (a, b) or with DMSO (negative controls; c, d). After
removal of the cUMP-AM stimulus, intracellular concentrations of
cUMP were determined by HPLC-MS/MS after 0, 30, 60, and 120 min
in the presence (filled symbols) and in the absence (open symbols) of the
PDE3 inhibitor milrinone. The figure shows absolute cUMP
concentrations (a, c) as well as relative amounts normalized to the
initial concentration at 0 min (b, d). Significance was calculated by
one-way ANOVA followed by Dunnett’s multiple comparison test
(comparison of the 30, 60, and 120 min points to the 0 min (100%)
point). Asterisk (*): statistics with milrinone-containing samples; plus
sign (+): statistics with milrinone-free samples; two symbols: p < 0.01;
three symbols: p < 0.001. Data are means ± SD from n = 3 (c, d) or n = 4
(a, b) independent experiments with single samples.

2
78
Naunyn-Schmiedeberg's Arch Pharmacol (2017) 390:269–280
binds to the same site as cAMP and competes in a similar way
with milrinone. It should be noted, however, that the milrinone Ki
values calculated from inhibition of cAMP hydrolysis should be
interpreted with caution. The high activity of PDE3A for cAMP
caused complete hydrolysis of cAMP at low substrate concen-
MRP proteins. cCMP is exported by MRP5, while cUMP is
transported by both MRP4 and 5 (Laue et al. 2014). MRP4 and
5 are also expressed in cardiac tissues (Dazert et al. 2003; Sassi
et al. 2012). Thus, MRP-mediated transport may account for
the observed PDE3-independent reduction of intracellular
cUMP in HL-1 cells. This mechanism is supported by recently
published data with HEL human erythroleukemia or K-562
human chronic myelogenous leukemia cells, where the unspe-
cific MRP inhibitor probenecid strongly increased the apoptotic
effects of several cNMPs including cUMP (Dittmar et al. 2016).
Alternatively, the distinct effect of milrinone on HL-1 cell ho-
mogenates as compared to living HL-1 cells may be explained
by PDE3 compartmentalization. Exogenously added cUMP-
AM may deliver the cUMP to an intracellular compartment that
does not contain PDE3.
trations, which may have led to an overestimation of K values.
i
PDE3A is a typical cardiac PDE (Meacci et al. 1992). Our
qPCR data show that transcripts of the cUMP-degrading enzymes
PDE3A, PDE3B, and PDE9A (Reinecke et al. 2011) are present
in both HL-1 cells and nRCMs. It was, however, difficult to prove
the presence of PDE3A on the protein level since the monoclonal
rabbit anti-PDE3A antibody used in our experiments produced
several non-specific signals at lower molecular masses
(<100 kDa). The two bands observed around 130 kDa might
belong to PDE3A isoforms. However, since a PDE3A-free cell
line was not available as a negative control, it cannot be excluded
that these bands represent non-specific staining of another protein
that coincidentally has the “correct” molecular weight.
The presence of PDE3B and PDE9A mRNA is in agreement
with previous reports on expression of PDE3B in mouse (Chung
et al. 2015) and rat (Liu and Maurice 1998) cardiovascular tissues
as well as of PDE9A in cardiac tissues (Kuhn 2015; Lee et al.
Analysis of basal intracellular cNMP concentrations revealed
that the cUMP concentration in HL-1 cells and nRCMs is com-
parable to that of the established second messenger cGMP. This
is in good agreement with recent data describing the presence of
cUMP in zebrafish heart (Dittmar et al. 2015). The low cUMP
affinity of PDE3A and the lacking milrinone effect in intact HL-1
cells, however, suggest that PDE3A does not participate in ho-
meostatic cUMP metabolism. Since cAMP has a several
hundred-fold higher affinity to PDE3A than cUMP, the enzyme
is most likely saturated by cAMP under physiological conditions,
specifically because cardiomyocytes contain far more cAMP
than cUMP. A low-affinity and high-velocity PDE for cUMP
like PDE3A, however, may become important under pathophys-
iological conditions when intracellular cUMP concentration is
dramatically increased. This is, for example, the case when the
exotoxin ExoY is introduced into a host cell after infection with
ExoY-positive P. aeruginosa (Beckert et al. 2014; Bähre et al.
2015). It should be noted that P. aeruginosa does not only infect
the lung but is also involved in infective endocarditis that occurs
in persons abusing intravenous drugs (Sousa et al. 2012).
In summary, we provide for the first time the basic enzyme
kinetics parameters for PDE3A-mediated cUMP hydrolysis and
show that HL-1 cardiomyogenic cells contain a milrinone-
sensitive cUMP-degrading activity. Future studies should inves-
tigate interactions between the PDE3A substrates cAMP, cGMP,
and cUMP in mixtures of these cNMPs that mimic intracellular
conditions. Moreover, the experiment shown in Fig. 7 should be
further developed and modified by changing the type and the
concentration of the PDE inhibitor as well as the time when it
is added to the sample. Moreover, the effect of MRP inhibitors
like probenecid on extrusion of cUMP and other cNMPs by
cardiomyocytes should be investigated. Furthermore, in addition
to Hl-1 cells, cUMP metabolism of primary cardiomyocytes
(e.g., neonatal or adult rat cardiomyocytes) should be analyzed.
Finally, it would be interesting to examine if cUMP-degrading
PDEs like PDE3A are upregulated in cells infected with ExoY-
positive P. aeruginosa and if enhanced intracellular cUMP con-
centrations affect cardiomyocyte function.
2015). In 1965, a cUMP-degrading PDE activity was detected in
beef and dog heart tissue (Hardman and Sutherland 1965). Our
results suggest that this activity may have been caused by
PDE3A.This is supported by our data demonstrating cUMP-
hydrolyzing activity in HL-1 cardiomyocyte homogenates that
was completely eliminated by milrinone. Thus, cUMP hydro-
lysis by HL-1 cell homogenates is mainly caused by PDE3
rather than by the milrinone-insensitive PDE9A. Since
milrinone does not differentiate between PDE 3A and 3B, how-
ever, we could not determine the individual contributions of the
two PDE3 isoforms
In cUMP-loaded intact HL-1 cells, however, milrinone did
not affect cUMP disposal. Independently of the presence of
milrinone, intracellular cUMP quickly decreased within
30 min. It should be noted that cUMP pre-incubation occurred
in the absence of inhibitor and milrinone was only added in the
second phase. This was done to reach comparable intracellular
cUMP starting concentrations for the degradation curves in the
DMSO and milrinone samples. Milrinone, however, may re-
quire some time to inhibit PDE3A because it has to penetrate
the cellular membrane first. Thus, despite the addition of
milrinone, there could have been a time window during which
PDE3Awas able to hydrolyze cUMP. Alternatively, HL-1 cells
may use a PDE-independent mechanism of cUMP removal. It
is well established that the multidrug resistance proteins
(MRPs) MRP4 and MRP5 are capable of exporting cNMPs
like cAMP or cGMP (Ritter et al. 2005). In smooth muscle
cells, export of cGMP by MRP4 is as important as hydrolysis
by PDE5 and regulates cGMP-induced vascular smooth muscle
relaxation (Krawutschke et al. 2015; Stangherlin and Zoccarato
2015). Non-canonical pyrimidine cNMPs are also substrates of

Naunyn-Schmiedeberg's Arch Pharmacol (2017) 390:269–280
279
Acknowledgements We thank Prof. Dr. Ralf Gerhard (Institute of
Toxicology, MHH) and Dr. Sabine Wolter (Institute of Pharmacology,
MHH) for excellent scientific discussions as well as Ms. Annette Garbe
activity of Bacillus anthracis edema factor and Bordetella pertussis
CyaA. Biochemistry 49:5494–5503. doi:10.1021/bi100684g
Hardman JG, Sutherland EW (1965) A cyclic 3′,5′-nucleotide phospho-
diesterase from heart with specificity for uridine 3′,5′-phosphate. J
Biol Chem 240:3704–3705
(Research Core Unit Metabolomics, MHH) and Ms. Martina Kasten
Molecular Cardiology Research Group, Department of Cardiology and
(
Angiology, MHH) for outstanding technical support. Moreover, we thank
the reviewers for their constructive comments.
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 579:
1
83–187. doi:10.1016/j.neulet.2014.07.019
Compliance with ethical standards
Hasan A, Danker KY, Wolter S, Bähre H, Kaever V, 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. doi:10.1016/j.bbrc.2014.04.099
Conflict of interest The authors declare that they have no competing
interests.
Hilfiker-Kleiner D, Kaminski K, Kaminska A, Fuchs M, Klein G, Podewski
E, Grote K, Kiian I, Wollert KC, Hilfiker A, Drexler H (2004) Regulation
of proangiogenic factor CCN1 in cardiac muscle: impact of ischemia,
pressure overload, and neurohumoral activation. Circulation 109:2227–
References
2
233. doi:10.1161/01.CIR.0000127952.90508.9D
Beckert U, Wolter S, Hartwig C, Bähre H, Kaever V, Ladant D, Frank
DW, 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. doi:10.1016/j.
bbrc.2014.06.088
Bender AT, Beavo JA (2006) Cyclic nucleotide phosphodiesterases: mo-
lecular regulation to clinical use. Pharmacol Rev 58:488–520.
doi:10.1124/pr.58.3.5
Bähre H, Danker KY, Stasch JP, Kaever V, Seifert R (2014) Nucleotidyl
cyclase activity of soluble guanylyl cyclase in intact cells. Biochem
Biophys Res Commun 443:1195–1199. doi:10.1016/j.
bbrc.2013.12.108
Ito M, Tanaka T, Saitoh M, Masuoka H, Nakano T, Hidaka H (1988)
Selective inhibition of cyclic AMP phosphodiesterase from various
human tissues by milrinone, a potent cardiac bipyridine. Biochem
Pharmacol 37:2041–2044
Krawutschke C, Koesling D, Russwurm M (2015) Cyclic GMP in vas-
cular relaxation: export is of similar importance as degradation.
Arterioscler Thromb Vasc Biol 35:2011–2019. doi:10.1161
/
ATVBAHA.115.306133
Kuhn M (2015) Cardiology: a big-hearted molecule. Nature 519:416–
17. doi:10.1038/nature14373
Laue S, Winterhoff M, Kaever V, van den Heuvel JJ, Russel FG, Seifert R
2014) cCMP is a substrate for MRP5. Naunyn Schmiedeberg’s
4
(
Bähre H, Hartwig C, Munder A, Wolter S, Stelzer T, Schirmer B, Beckert
U, Frank DW, Tümmler B, Kaever V, Seifert R (2015) cCMP and
cUMP occur in vivo. Biochem Biophys Res Commun 460:909–
Arch Pharmacol 387:893–895. doi:10.1007/s00210-014-1018-9
Lee DI, Zhu G, Sasaki T, Cho GS, Hamdani N, Holewinski R, Jo SH,
Danner T, Zhang M, Rainer PP, Bedja D, Kirk JA, Ranek MJ,
Dostmann WR, Kwon C, Margulies KB, Van Eyk JE, Paulus WJ,
Takimoto E, Kass DA (2015) Phosphodiesterase 9A controls nitric-
oxide-independent cGMP and hypertrophic heart disease. Nature
519:472–476. doi:10.1038/nature14332
Liu H, Maurice DH (1998) Expression of cyclic GMP-inhibited phospho-
diesterases 3A and 3B (PDE3A and PDE3B) in rat tissues: differen-
tial subcellular localization and regulated expression by cyclic AMP.
Br J Pharmacol 125:1501–1510. doi:10.1038/sj.bjp.0702227
Meacci E, Taira M, Moos M, Smith CJ, Movsesian MA, Degerman E,
Belfrage P, Manganiello V (1992) Molecular cloning and expression
of human myocardial cGMP-inhibited cAMP phosphodiesterase.
Proc Natl Acad Sci U S A 89:3721–3725
9
14. doi:10.1016/j.bbrc.2015.03.115
Cheng Y, Prusoff WH (1973) Relationship between the inhibition con-
stant (K ) and the concentration of inhibitor which causes 50 per cent
inhibition (I50) of an enzymatic reaction. Biochem Pharmacol 22:
099–3108
I
3
Chung YW, Lagranha C, Chen Y, Sun J, Tong G, Hockman SC, Ahmad
F, Esfahani SG, Bae DH, Polidovitch N, Wu J, Rhee DK, Lee BS,
Gucek M, Daniels MP, Brantner CA, Backx PH, Murphy E,
Manganiello VC (2015) Targeted disruption of PDE3B, but not
PDE3A, protects murine heart from ischemia/reperfusion injury.
Proc Natl Acad Sci U S A 112:E2253–E2262. doi:10.1073
/pnas.1416230112
Claycomb WC, Lanson NA, Stallworth BS, Egeland DB, Delcarpio JB,
Bahinski A, Izzo NJ (1998) HL-1 cells: a cardiac muscle cell line
that contracts and retains phenotypic characteristics of the adult car-
diomyocyte. Proc Natl Acad Sci U S A 95:2979–2984
Dazert P, Meissner K, Vogelgesang S, Heydrich B, Eckel L, Böhm M,
Warzok R, Kerb R, Brinkmann U, Schaeffeler E, Schwab M,
Cascorbi I, Jedlitschky G, Kroemer HK (2003) Expression and lo-
calization of the multidrug resistance protein 5 (MRP5/ABCC5), a
cellular export pump for cyclic nucleotides, in human heart. Am J
Pathol 163:1567–1577. doi:10.1016/S0002-9440(10)63513-4
Dittmar F, Abdelilah-Seyfried S, Tschirner SK, Kaever V, Seifert R
Monzel M, Kuhn M, Bähre H, Seifert R, Schneider EH (2014) PDE7A1
hydrolyzes cCMP. FEBS Lett 588:3469–3474. doi:10.1016/j.
febslet.2014.08.005
Niiya T, Osawa H, Onuma H, Suzuki Y, Taira M, Yamada K, Makino H
(2001) Activation of mouse phosphodiesterase 3B gene promoter by
adipocyte differentiation in 3T3-L1 cells. FEBS Lett 505:136–140
Omori K, Kotera J (2007) Overview of PDEs and their regulation. Circ
Res 100:309–327. doi:10.1161/01.RES.0000256354.95791.f1
Reinecke D, Burhenne H, Sandner P, Kaever V, Seifert R (2011) Human
cyclic nucleotide phosphodiesterases possess a much broader
substrate-specificity than previously appreciated. FEBS Lett 585:
3259–3262. doi:10.1016/j.febslet.2011.09.004
Ritter CA, Jedlitschky G, Meyerzu Schwabedissen H, Grube M, Köck K,
Kroemer HK (2005) Cellular export of drugs and signaling mole-
cules by the ATP-binding cassette transporters MRP4 (ABCC4) and
MRP5 (ABCC5). Drug Metab Rev 37:253–278. doi:10.1081
/DMR-200047984
(
2015) Temporal and organ-specific detection of cNMPs including
cUMP in the zebrafish. Biochem Biophys Res Commun 468:708–
12. doi:10.1016/j.bbrc.2015.11.020
7
Dittmar F, Wolter S, Seifert R (2016) Regulation of apoptosis by cyclic
nucleotides in human erythroleukemia (HEL) cells and human my-
elogenous leukemia (K-562) cells. Biochem Pharmacol 112:13–23.
doi:10.1016/j.bcp.2016.04.018
Göttle M, Dove S, Kees F, Schlossmann J, Geduhn J, König B, Shen Y,
Tang WJ, Kaever V, Seifert R (2010) Cytidylyl and uridylyl cyclase
Sassi Y, Abi-Gerges A, Fauconnier J, Mougenot N, Reiken S, Haghighi
K, Kranias EG, Marks AR, Lacampagne A, Engelhardt S, Hatem
SN, Lompre AM, Hulot JS (2012) Regulation of cAMP homeostasis

2
80
Naunyn-Schmiedeberg's Arch Pharmacol (2017) 390:269–280
by the efflux protein MRP4 in cardiac myocytes. FASEB J 26:1009–
017. doi:10.1096/fj.11-194027
Seifert R (2015) cCMP and cUMP: emerging second messengers. Trends
Biochem Sci 40:8–15. doi:10.1016/j.tibs.2014.10.008
Wechsler J, Choi YH, Krall J, Ahmad F, Manganiello VC, Movsesian
MA (2002) Isoforms of cyclic nucleotide phosphodiesterase
PDE3A in cardiac myocytes. J Biol Chem 277:38072–38078.
doi:10.1074/jbc.M203647200
1
Seifert R, Schneider EH, Bähre H (2015) From canonical to non-canonical
cyclic nucleotides as second messengers: pharmacological implications.
Pharmacol Ther 148:154–184. doi:10.1016/j.pharmthera.2014.12.002
Sousa C, Botelho C, Rodrigues D, Azeredo J, Oliveira R (2012) Infective
endocarditis in intravenous drug abusers: an update. Eur J Clin
Microbiol Infect Dis 31:2905–2910. doi:10.1007/s10096-012-1675-x
Stangherlin A, Zoccarato A (2015) Relax: it’s not all about degradation.
Arterioscler Thromb Vasc Biol 35:1907–1909. doi:10.1161
Wolter S, Golombek M, 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. doi:10.1016/j.bbrc.2011.10.093
Wolter S, Kloth C, Golombek M, Dittmar F, Försterling L, Seifert R
(2015) cCMP causes caspase-dependent apoptosis in mouse
lymphoma cell lines. Biochem Pharmacol 98:119–131.
doi:10.1016/j.bcp.2015.08.096
/
ATVBAHA.115.306217
Zong X, Krause S, Chen CC, Krüger J, Gruner C, Cao-Ehlker X,
Fenske S, Wahl-Schott C, Biel M (2012) Regulation of
hyperpolarization-activated cyclic nucleotide-gated (HCN)
channel activity by cCMP. J Biol Chem 287:26506–26512.
doi:10.1074/jbc.M112.357129
Taira M, Hockman SC, Calvo JC, Belfrage P, Manganiello VC (1993)
Molecular cloning of the rat adipocyte hormone-sensitive cyclic
GMP-inhibited cyclic nucleotide phosphodiesterase. J Biol Chem
2
68:18573–18579