The properties of phosphodiesterase 11A4 GAF domains
are regulated by modifications in its N-terminal domain
Marco Gross-Langenhoff
1
, Arnulf Stenzl
2
, Florian Altenberend
1
, Anita Schultz
1
and
Joachim E. Schultz
1
1 Pharmazeutisches Institut, Universita
¨
tTu
¨
bingen, Germany
2 Urologische Universita
¨
tsklinik, Universita
¨
tTu
¨
bingen, Germany
The secondary messengers cAMP and cGMP regulate
a variety of signalling pathways in essentially all cells
[1]. Their intracellular levels are balanced by the
rates of biosynthesis via adenylyl cyclase (EC 4.6.1.1)
and guanylyl cyclase (EC 4.6.1.2) and of breakdown
via cyclic nucleotide phosphodiesterase (PDEs;

tTu
¨
bingen, Morgenstelle 8,
72076 Tu
¨
bingen, Germany
Fax: +49 7071 295952
Tel: +49 7071 2974676
E-mail: [email protected]
(Received 22 October 2007, revised 4
December 2007, accepted 5 February 2008)
doi:10.1111/j.1742-4658.2008.06319.x
The tandem GAF domain of human phosphodiesterase 11A4 (hPDE11A4)
requires 72 lm cGMP for half-maximal effective concentration (EC
50
)ofa
cyanobacterial adenylyl cyclase used as a reporter enzyme. Here we exam-
ine whether modifications in the N-terminus of PDE11A4 affect cGMP sig-
nalling. The N-terminus has two phosphorylation sites for cyclic nucleotide
monophosphate-dependent protein kinases (Ser117, Ser168). Phosphoryla-
tion of both by cAMP-dependent protein kinase decreased the EC
50
value
for cGMP from 72 to 23 lm. Phosphomimetic point mutations
(S117D ⁄ S167D), which project complete phosphorylation, lowered the
EC
50
value to 16 lm. Structural and sequence data indicate that 196 amino
acids precede the start of the GAF domain in hPDE11A4. Removal of 197
amino acids yielded unregulated cyclase activity, whereas truncation by 196

extended N-terminus in vitro phosphorylation sites
have been identified at Ser117 and Ser162 [13]. In vivo
phosphorylation and a potential physiological role for
these phosphorylation sites have not been reported.
PDE11 is a true dual-substrate PDE, i.e. it hydro-
lyses cAMP and cGMP with similar K
m
and V
max
val-
ues [13–15,17]. The tissue distribution of the PDE11
isozymes has not yet been fully examined. It is known
that robust expression in humans occurs in the pros-
tate [18,19], testis [20–22], spermatozoa [22] and brain
[23]. The physiological role of PDE11 in the regulation
of cyclic nucleotide levels is currently unclear, partly
because of a lack of any clear pattern in its tissue dis-
tribution and partly because a PDE11-specific inhibitor
has not yet been developed. In studies using a PDE11
knockout mouse model, it has been discussed in terms
of its involvement in sperm development and function
[22]. In addition, it is speculated that PDE11 plays a
role in the pathology of major depressive disorder [23]
and the development of endocrinal tumours [24]. Fur-
thermore, hPDE11 presumably plays a role in various
urological diseases such as benign prostatic syndrome,
erectile dysfunction and premature ejaculation [25,26].
In agreement with the expression of PDE5 and PDE11
in the transitional zone of the prostate, these PDE iso-
forms might also have a role in regulating the prolifer-

the longest N-terminus of human GAF-containing
PDE2, )5, )6, )10 and )11. With 27 strongly basic
amino acids, Arg and Lys, it has a calculated isoelec-
tric point of 10.8 and carries 10 positive charges at
pH 7, i.e. this N-terminus may be prone to interact
with other subdomains. For comparison, the N-termini
of PDE2 and PDE5 have isoelectric points of 5.3 and
5.7, respectively. The PDE11A4 N-terminus has two
phosphorylation sites, Ser117 and Ser162 imbedded in
the signature sequences RRA
117
S for cAMP-dependent
protein kinase (cAK; RRXS) and RKA
162
S for
cGMP-dependent protein kinase (cGK; RKXS). It has
previously been shown that both sites are phosphory-
lated in vitro by cAK or cGK, yet the functional con-
sequences have not been reported [13]. We used the
catalytic subunit of cAK for phosphorylation of the
chimeric protein and observed that the EC
50
for cGMP
stimulation of cyaB1 was reduced to 23 ± 1.2 lm,
whereas basal activity (7.4 and 6.8 nmoles cAMPÆ
mg
)1
Æmg
)1
for the unphosphorylated and phosphory-

)1
). cAMP did not stimulate (data not
shown). Similarly, the K
m
values for ATP and the
V
max
values were unaffected by the mutations.
The PDE11A4 N-terminus affects cGMP
regulation
We have shown that a chimera consisting of the
PDE11A4 tandem GAF domain and the cyaB1 N-ter-
minus in front of cyab1 AC, is not stimulated by
cGMP or cAMP [22]. This indicated that the N-termi-
nus which precedes the GAF tandem domain either
affects intramolecular signal transduction and ⁄ or
directly inhibits the C-terminally located AC. Based on
two available GAF tandem structures and correspond-
ingly adapted sequence alignments, we designated
Lys196 as the last residue of the N-terminus and the
start of GAF-A at Lys197, the position at which a
short a helix that is visible in both structures starts
[24,28]. Accordingly, we generated two shortened
chimeras, one starting at Lys197 [PDE11A4(197-
568)cyaB1 AC] and one at Lys198 [PDE11A4(198-
568)cyaB1 AC]. The data corroborated the aforemen-
tioned conclusions because PDE11A4(198-568)cyaB1
AC could not be stimulated by cGMP, whereas
PDE11A4(197-568)cyaB1 AC was stimulated 3.4-fold
by cGMP with an EC

phosphomimetic mutants in human PDE11A4 tandem GAF domain
constructs.
Chimeric construct
EC
50
cGMP
(l
M)
Maximal stimulation
(fold)
PDE11 ⁄ cyaB1 (wild-type) 72.5 ± 10.1 3.8 ± 0.4
S117D 115.5 ± 26.1 2.3 ± 0.1*
S162D 62.9 ± 14.6 2.5 ± 0.5*
S117D ⁄ S162D 16.3 ± 5.6* 1.8 ± 0.1*
S117E 74.2 ± 2.3 3.6 ± 0.4
S162E 36.4 ± 3.2 2.4 ± 0.2*
S117E ⁄ S162E 38.1 ± 1.2 2.6 ± 0.1*
*P < 0.05 compared with the parent chimera using Dunnett’s ana-
lysis (see text); n = 4–8.
0
40
120
Lys197
–log [cGMP] [M]
6543
35
45
66
116
kDa

cGMP affinity. Thus, removal of at least 176 N-termi-
nal amino acids had a dual effect; it released AC from
apparent inhibition by its N-terminus and increased
the cGMP affinity of the GAF tandem domain up to
20-fold (Figs 1 and 2B). Maximal stimulation by
cGMP was significantly affected in three of the nine
truncated constructs (Fig. 2C), although the reduction
appeared to be minor. By and large, we consider this
to be within the experimental variability for this type
of assay. Matters were clearly different with regard to
the EC
50
values for cGMP stimulation, as determined
by dose–response curves for all truncated versions. In
fact, the data did not require statistical analysis
(Fig. 2C). The EC
50
for cGMP stimulation was
3.5 ± 2.2 lm for the construct beginning at Glu177
and 10.2 ± 4.0 lm for the construct beginning at
Pro187 (Fig. 2C). Thus, only the two latter constructs
were comparable with that without an N-terminal
region, which had an EC
50
value of 7.6 lm cGMP
(Fig. 1). cAMP did not stimulate the reporter enzyme
significantly in any of these constructs (data not
shown). cGMP-binding assays were not carried out
because the lm affinities precluded promising experi-
ments using this technique. For all shortened con-

50
-values for cGMP [µM]
Met-1
Ser-43
Gly-110
Lys-119
Leu-149
Leu-164
Ala-169
Glu-177
Pro-187
Lys-197
ND ND ND
0
1
2
3
4
Met-1
Ser-43
Gly-110
Lys-119
Leu-149
Leu-164
Lys-197
Ala-169
*
Glu-177
*
Pro-187

1646 FEBS Journal 275 (2008) 1643–1650 ª 2008 The Authors Journal compilation ª 2008 FEBS
were not due to proteolysis because the western blot
detected the N-terminal affinity tag and C-terminal
degradation at the catalytic domain would have
resulted in a loss of AC activity. We noticed that basal
AC activities correlated reasonably well with maximal
activation which was attainable with cGMP. There-
fore, we concluded that intramolecular signalling most
likely was unimpaired in the shortened constructs.
Discussion
Previously we reported that the tandem GAF domains
of mammalian PDE2, )5, )10 and )1 functionally cou-
ple to the cyanobacterial AC cyaB1 with retention of
their regulatory potency [12,27,29]. Thus, cyaB1 AC is
regulated by cAMP when coupled to the tandem GAF
domain of PDE10, and by cGMP when linked to the
tandem GAF domain of PDE2, )5or)11. The bio-
chemical properties of signalling by human tandem
GAF domains to cyaB1 AC were by and large in agree-
ment with data obtained in studies using bacterially
expressed GAF domains, e.g. the cyclic nucleotide-
binding affinities of PDE2 and )5 GAF domains. How-
ever, we were able to determine additional GAF
domain properties because of the dissociation of the
allosteric regulators, cyclic nucleotides and the substrate
of the reporter enzyme, ATP. As far as AC regulation
by the PDE11A4 tandem GAF domain is concerned,
we have previously reported that cGMP activates with
an EC
50

affinity. With both positions mutated to Asp, the EC
50
value for cGMP was 16.3 lm compared with 72.5 lm
in the non-mutated chimera. Therefore, as in PDE5,
phosphorylation at the N-terminus of the PDE11A4
GAF tandem may constitute a mechanism regulating
the cGMP affinity of the PDE11A4 tandem GAF
domain and thus allosterically affect PDE11 activity.
The role of the N-terminus of the PDE11A4 GAF
tandem domain
The N-termini that precede the GAF domains in
mammalian PDEs are of significant length, 221 amino
acids in PDE2, 148 in PDE5, 82 in PDE10 and 196 in
PDE11A4, and it is conceivable that they have a func-
tion in conjunction with the regulation of PDE activ-
ity in addition to the phosphorylations in PDE5 and
PDE11A4 (see above) [27]. Indeed, we have shown
that shortening of the PDE5 N-terminus significantly
affected intramolecular signalling in chimeras similar
to those used in this study. To date, cNMP binding
assays using PDE11A4 have been negative and regu-
lation by cyclic nucleotides is uncertain [13,15,17],
possibly because of the low cGMP affinity of the
PDE11A4 GAF domains [12]. Here, we explored
whether the PDE11A4 N-terminus is involved in mod-
ulating cGMP affinity. We generated a panel of nine
N-terminally shortened PDE11A4 tandem GAF chi-
meras according to secondary structure predictions.
This included the complete removal of 196 amino
acids. Interestingly, removal of Lys197 which con-

involved in PDE regulation are concerned [27,29].
Therefore, one may postulate that N-terminal modifi-
cations of the PDE11A4 tandem GAF domains are
required to enable cGMP regulation of PDE11 cata-
lytic activity. Actually, the terminal region may have
two separate effects: one that directly affects cGMP
affinity via the GAF domains, and a second compo-
nent that acts directly on the catalytic activity. To
date, no structure is available for the N-termini of
PDE2, )5, )10 or )11 and it will be interesting to see
whether common structural features exist in the N-ter-
mini of GAF-domain-containing PDEs that might
contribute to intramolecular signalling in a similar
manner. Another point merits discussion. Irrespective
of phosphomimetic mutations or N-terminal shorten-
ings, PDE11A4 GAF-tandem-mediated activation of
cyaB1 AC was always modest, mostly two- to three-
fold, when compared with the effects of the PDE2,
PDE5 and PDE10 tandem GAF domains. One may
ask whether this is physiologically significant because:
(a) a two- to threefold change in the V
max
value will
hasten intracellular adjustment of cAMP or cGMP
levels considerably and thus possibly shorten excited
cell states; (b) regulation of PDE4 isozymes is reported
to involve the phosphorylation of a serine located at
the beginning of the tandem of upstream conserved
regions, which precedes the catalytic domain. PDE4
activation by this phosphorylation is about twofold,

1-568
cyaB1
386-859
chimera [12] served as a template to generate
all mutants and the N-terminally shortened constructs. Sin-
gle- and double-point mutations of PDE11A4 Ser117 and
Ser162 were generated by fusion PCR with Pfu DNA poly-
merase (Promega, Madison, WI, USA) using corresponding
sense and antisense primers (MWG Biotech, Ebersberg,
Germany) and nearby restriction sites (KpnI, SacI, StuI and
MfeI). N-Terminally shortened constructs were created with
respective NdeI sense primers and an MfeI antisense primer
in the expression vector pET16b adding a C-terminal
His10-tag. To generate PDE 11A4 GAF-A(181-370)cyaB1(386-
859) corresponding parts were amplified by PCR and
cloned into pET16b ⁄ pQE30 [27] via BamHI, BglII (GAF-
A) and Bgl II, SmaI (cyaB1 AC), respectively. An MRGS-
His6-tag was located N-terminally.
The hPDE11A4(1-568) construct in pQE60 via 5¢-NcoI
and 3¢-BamHI was obtained using hPDE11 as a PCR tem-
plate. This added a C-terminal GSRSHis6 affinity tag. The
fidelity of all constructs was verified by double stranded
sequencing. Primer sequences are available on request.
All pQE plasmids were obtained from Qiagen (Hilden,
Germany).
Expression and purification of recombinant
proteins
hPDE11A4 ⁄ cyaB1 chimeras were expressed and purified as
described earlier [12]. pQE60 constructs were expressed at
16 °C at 400 lm isopropyl b-d-thiogalactoside overnight

secondary anti-(mouse IgG) (Dianova, Hamburg, Germany).
Peroxidase detection was carried out with the ECL Plus kit
(Amersham-Pharmacia, Freiburg, Germany). Preferably,
western blots of affinity-purified proteins are depicted
because for the current studies it was necessary to ensure that
the constructs did not contain degraded products which
would affect the cyclase reaction. N-Terminally degraded
proteins would not have bound to the Ni
2+
⁄ nitrilotetra ⁄
acetic acid affinity material. C-Terminally truncated proteins
are catalytically inactive, i.e. western blots show the extent of
purification of the relevant recombinant protein species.
Miscellaneous methods
Total protein concentrations were determined using the
method described by Bradford [31], with BSA as the stan-
dard. Data are given as means ± SEM of between four
and eight experiments. The statistical evaluation of data
was carried out using the Student’s t-test and multiple com-
parisons by one-way analysis of variance (ANOVA) fol-
lowed by Dunnett’s posterior test using graphpad prism
software, version 4.0 for Windows (GraphPad Software
Inc., San Diego, CA, USA http://www.graphpad.com).
A value of P < 0.05 was considered significant.
Acknowledgements
We are grateful to Prof. M. Wahl, Univeristy of Tu
¨
bin-
gen, for help with the statistical analysis. This study was
supported by the Deutsche Forschungsgemeinschaft.

Hol WG & Beavo JA (2002) The two GAF domains in
phosphodiesterase 2A have distinct roles in dimerization
and in cGMP binding. Proc Natl Acad Sci USA 99,
13260–13265.
10 Rybalkin SD, Rybalkina IG, Shimizu-Albergine M,
Tang XB & Beavo JA (2003) PDE5 is converted to an
activated state upon cGMP binding to the GAF
A domain. EMBO J 22, 469–478.
11 Mou H & Cote RH (2001) The catalytic and GAF
domains of the rod cGMP phosphodiesterase (PDE6)
heterodimer are regulated by distinct regions of its inhi-
bitory gamma subunit. J Biol Chem 276, 27527–27534.
12 Gross-Langenhoff M, Hofbauer K, Weber J, Schultz A
& Schultz JE (2006) cAMP is a ligand for the tandem
GAF domain of human phosphodiesterase 10 and
cGMP for the tandem GAF domain of phosphodiester-
ase 11. J Biol Chem 281, 2841–2846.
13 Yuasa K, Kotera J, Fujishige K, Michibata H, Sasaki
T & Omori K (2000) Isolation and characterization of
two novel phosphodiesterase PDE11A variants showing
unique structure and tissue-specific expression. J Biol
Chem 275, 31469–31479.
14 Hetman JM, Robas N, Baxendale R, Fidock M, Phillips
SC, Soderling SH & Beavo JA (2000) Cloning and char-
acterization of two splice variants of human phosphodi-
esterase 11A. Proc Natl Acad Sci USA 97, 12891–12895.
15 Fawcett L, Baxendale R, Stacey P, McGrouther C,
Harrow I, Soderling S, Hetman J, Beavo JA & Phillips
SC (2000) Molecular cloning and characterization of a
distinct human phosphodiesterase gene family:

22 Wayman C, Phillips S, Lunny C, Webb T, Fawcett L,
Baxendale R & Burgess G (2005) Phosphodiesterase 11
(PDE11) regulation of spermatozoa physiology. Int J
Impot Res 17, 216–223.
23 Wong ML, Whelan F, Deloukas P, Whittaker P,
Delgado M, Cantor RM, McCann SM & Licinio J (2006)
Phosphodiesterase genes are associated with susceptibil-
ity to major depression and antidepressant treatment
response. Proc Natl Acad Sci USA 103, 15124–15129.
24 Horvath A, Boikos S, Giatzakis C, Robinson-White A,
Groussin L, Griffin KJ, Stein E, Levine E, Delimpasi
G, Hsiao HP et al. (2006) A genome-wide scan identi-
fies mutations in the gene encoding phosphodiester-
ase 11A4 (PDE11A) in individuals with adrenocortical
hyperplasia. Nat Genet 38, 794–800.
25 Bazrafhan S, Uckert S, Mayer M, Stief C & Jonas U
(2007) Phosphodiesterase inhibitors attenuate the con-
tractile activity of human seminal vesicles smooth mus-
cle. Eur Urol 105(Suppl.), 330.
26 Uckert S, Hedlund P, Andersson KE, Truss MC, Jonas
U & Stief CG (2006) Update on phosphodiesterase
(PDE) isoenzymes as pharmacologic targets in urology:
present and future. Eur Urol 50, 1194–1207 [discussion
p. 1207].
27 Bruder S, Schultz A & Schultz JE (2006) Character-
ization of the tandem GAF domain of human
phosphodiesterase 5 using a cyanobacterial adenylyl
cyclase as a reporter enzyme. J Biol Chem 281,
19969–19976.
28 Bruder S, Linder JU, Martinez SE, Zheng N, Beavo JA