A FYVE-containing unusual cyclic nucleotide
phosphodiesterase from Trypanosoma cruzi
Stefan Kunz, Michael Oberholzer and Thomas Seebeck
Institute of Cell Biology, University of Bern, Bern, Switzerland
The cell biology of Trypanosoma cruzi, the causative
agent of South American Chagas’ disease, has been
extensively studied. Surprisingly, still very little is
known about the role of cyclic nucleotide signaling in
this organism [1]. A number of earlier studies have
indicated a role of cAMP in differentiation [2,3], and
the existence of a nitric oxide regulated guanylyl
cyclase has been suggested [4]. In T. cruzi epimasti-
gotes, a cAMP-regulated transcript has been identified
that can be induced by elevated cAMP levels [5]. The
corresponding gene, TC26, was later found to code for
an RNaseH and to be localized on a large family of
repetitive genetic elements. More recently, the T. cruzi
genome was shown to code for several adenylyl cyclas-
es, all predicted to be similarly organized, consisting of
a large N-terminal, presumably extracellular region,
which is followed by a single transmembrane helix and
a C-terminal catalytic domain [6]. The structure of
these cyclases is entirely different from that of their
mammalian counterparts, but closely similar to that of
the cyclases characterized in Leishmania donovani [7]
and in African trypanosomes [8–10]. One of these
adenylyl cyclases, TczAC, was found to interact with a
paraflagellar rod protein, and is most likely located in
the flagellum [11].
A cAMP-specific phosphodiesterase (PDE) activity
has been demonstrated in T. cruzi by various laborat-
ories [12,13]. Recently, the first cyclic-nucleotide-
specific PDE from T. cruzi has been identified and
characterized at the molecular level [14]. This enzyme,
Keywords:
Chagas’ disease; cyclic nucleotides; FYVE
domain; kinetoplastids; phosphodiesterase
Correspondence
T. Seebeck, Institute of Cell Biology,
Baltzerstrasse 4, CH-3012 BERN,
Switzerland
Fax: +41 31 6314684
Tel: +41 31 6314649
E-mail: thomas.seebeck@izb.unibe.ch
Website: http://www.izb.unibe.ch
Nucleotide sequence data have been sub-
mitted to the DDBJ ⁄EMBL ⁄GenBank data-
bases under the accession numbers
AJ889575 and AJ889576 for TcrPDEC alle-
les 1 and 2, respectively.
(Received 26 August 2005, revised 20 Octo-
ber 2005, accepted 27 October 2005)
doi:10.1111/j.1742-4658.2005.05039.x
Cyclic-nucleotide-specific phosphodiesterases (PDEs) are key players in the
intracellular signaling pathways of the important human pathogen Trypano-
soma cruzi. We report herein the identification of an unusual PDE from
this protozoal organism. This enzyme, TcrPDEC, is a member of the class
I PDEs, as determined from the presence of a characteristic signature
sequence and from the conservation of a number of functionally important
amino acid residues within its catalytic domain. Class I PDEs include a
large number of PDEs from eukaryotes, among them all 11 human PDE
families. Unusually for an enzyme of this class, TcrPDEC contains a
FYVE-type domain in its N-terminal region, followed by two closely
spaced coiled-coil domains. Its catalytic domain is located in the middle of
the polypeptide chain, whereas all other class I enzymes contain their cata-
lytic domains in their C-terminal parts. TcrPDEC can complement a PDE-
deficient yeast strain. Unexpectedly for a kinetoplastid PDE, TcrPDEC is a
dual-specificity PDE that accepts both cAMP and cGMP as its substrates.
Abbreviations
DMSO, dimethyl sulfoxide; EHNA, erythro-9-(2-hydroxy-3-nonyl)adenosin; FYVE, domain containing Fab1p, YOTB, Vac1p and EEA1PDE; GST,
glutathione-S-transferase; IBMX, isobutyl methyl xanthine; LmPDEC, phosphodiesterase from Leishmania major; PtdIns(3)P, phosphatidyl
inositol-3-phosphate; TbPDEC, phosphodiesterase from Trypanosoma brucei; TcrPDEC, phosphodiesterase from Trypanosoma cruzi.
6412 FEBS Journal 272 (2005) 6412–6422 ª2005 The Authors Journal compilation ª2005 FEBS
TcPDE1 is entirely cAMP-specific, and it is located
along the flagellum. In terms of its amino acid
sequence, TcPDE1 is a close homologue of the Trypano-
soma brucei PDEs TbPDE2B [15] and TbPDE2C [16],
as well as of LmPDEB1 and LmPDEB2 of Leishmania
major (Johner et al., unpublished results). All of these
enzymes belong to the class I PDEs [17]. TcPDE1 con-
tains two GAF domains [18,19] in its N-terminal moi-
ety, and a C-terminal catalytic domain. TcPDE1, as all
its other kinetoplastid homologues, is highly cAMP-
selective.
This study reports the identification and characteri-
zation of a novel and rather unusual PDE from T. cruzi.
According to the recently proposed unifying nomen-
clature for kinetoplastid PDEs [20], this enzyme was
designated as TcrPDEC. Based on the amino acid
sequence of its catalytic domain, TcrPDEC unambigu-
ously belongs to the class I PDEs. However, it is a
rather unusual PDE in several respects: (a) unlike all
other class I PDEs, its catalytic domain is localized in
the middle of the polypeptide chain, and not at its
C-terminus; (b) the N-terminal region of TcrPDEC
contains a FYVE-type domain [21,22], a functional
domain that has not been found in any PDE so far;
and (c) TcrPDEC is the first dual-substrate PDE, with
similar K
m
values for cAMP and cGMP, that has been
identified in kinetoplastids.
Results
Identification of TcrPDEC
When the T. cruzi database (http://www.genedb.org/
genedb/tcruzi) was screened for putative PDEs, a gene
was identified that codes for a rather unusual PDE,
TcrPDEC (temporary gene identification number
Tc00.1047053506697.20). The open reading frame of
TcrPDEC was amplified from genomic DNA, and sev-
eral PCR products were sequenced. This analysis
revealed the presence of two distinct alleles which dif-
fer by 62 bp (out of the 2775 bp of the entire open
reading frame; 2.2% sequence divergence). These single
nucleotide polymorphisms translate into 38 amino acid
changes (4.1% amino acid substitutions; 21 conserved,
17 nonconserved). Only six of these substitutions occur
in the catalytic domain of the enzyme, and none of
them affects a residue that is crucial for function (see
below). Southern blot analysis of T. cruzi genomic
DNA by hybridization with the complete open reading
frame of TcrPDEC results in restriction enzyme pat-
terns that are compatible with the nucleotide sequence
of TcrPDEC, demonstrating that it represents a single
copy gene (Fig. 1). EcoRI, PstI and EcoRV cut once
within the open reading frame, BamHI does not cut,
and HindIII cuts three times, resulting in two frag-
ments too small to be detected by hybridization and
one fragment of two kilobases. The hybridization of
SalI-digested DNA confirmed the polymorphism of
one SalI site detected by the sequence analysis of the
two alleles. The establishment of TcrPDEC as a single-
copy gene is not entirely trivial, as large parts of the
T. cruzi genome have undergone a duplication [23].
Functional domains of TcrPDEC
The open reading frame of TcrPDEC codes for a pro-
tein of 924 amino acids (calculated relative molecular
mass of 103 169, calculated pI ¼5.91) with several
functional domains (Fig. 2A). The N-terminus (P
10
–
G
73
) contains a FYVE-type domain, an acronym
composed of the designations of the first four repre-
sentatives Fab1p, YOTB, Vac1p and EEA1 [21], that is
followed by two closely spaced coiled-coil regions
(D
144
–D
179
and K
207
–E
264
). FYVE-type domains are
zinc-finger-like structures, which are currently divided
10
8
6
5
4
3
2
1.5
1
0.5
EcoRI SalI EcoRV
BamHI PstI HindIII
kb
Fig. 1. TcrPDEC is a single copy gene. Genomic DNA of T. cruzi
hybridized with a probe representing the entire open reading frame
of TcrPDEC.
S. Kunz et al. Novel phosphodiesterase from Trypanosoma cruzi
FEBS Journal 272 (2005) 6412–6422 ª2005 The Authors Journal compilation ª2005 FEBS 6413
into two classes: The classical FYVE domains (exam-
ples: (HsEEA1, accession number Q15075; DmHRS,
accession number Q960 ·8; and ScVps27p, accession
number P40343) share three consensus motifs (WXXD,
R + HHCR and RVC), and they bind specifically
to membrane-embedded-phosphatidyl-inositol-3-phos-
phate [PtdIns(3)P]. FYVE-variant domains such as
HsDFCP1 and AtPRAF1 lack some of the conserved
residues (Fig. 2B), but still bind PtdIns(3)P, albeit with
lower affinity. The FYVE-related domains (e.g. rabphi-
lin 3 A (P47709) or human Rim1 (Q86UR5) exhibit
a still higher sequence divergence in the consensus
region. Their function is still undetermined. The align-
ment of the FYVE-type domain of TcrPDEC places it
close to the FYVE-variant domains. All eight cysteine
residues predicted to be involved in Zn
2+
-binding are
fully conserved (Fig. 2C), as are the two predicted heli-
cal regions. The two hydrophobic amino acids that are
inserted in the membrane upon PtdIns(3)P binding
(L
185
and L
186
in ScVps27p [22]); are represented by
L
30
and F
31
of TcrPDEC. When matched with the
WxxD...R + HHCR...RVC motif of FYVE domains,
the sequence of TcrPDEC exhibits several alterations.
A glutamate residue at position four of the first block
is substituted by aspartate, arginine at the beginning of
the second block is replaced by an alanine, the two
adjacent histidine residues are replaced by serine and
glutamine, the subsequent arginine is replaced by pro-
line, and finally the arginine of the third block is
replaced by a lysine. The consequence of these replace-
ments is a decrease of the overall net charge of the
motive from +4 to +1. The effect of these changes
on a putative membrane binding of the FYVE-type
domain of TcrPDEC remains to be explored, but they
render the TcrPDEC domain unlikely to bind to
PtdIns(3)P. This prediction is confirmed by the obser-
vation that the recombinant FYVE domain of
TcrPDEC does not bind to PtdIns(3)P, nor to
PtdIns(3,4)P
2
, PtdIns(4,5)P
2
, PtdIns(3,4,5)P
3
, phos-
phatidic acid, phosphatidyl choline, phosphatidyl ser-
ine or phosphatidyl inositol in a dot-spot assay [24,25]
(data not shown).
12
34
AB
FYVE
DmHrs
PHD
HsKAP-1
FYVE-related
HsRIM1
FYVE-related
RnRPH3A
TcPDEC
FYVE-variant
AtPRAF1
FYVE-variant
HsDCFP1
FYVE
ScVps27p
FYVE
HsEEA1
C
Fig. 2. The FYVE-type domain of TcrPDEC. (A) Functional organization of TcrPDEC: 1, FYVE-type domain; 2 and 3, coiled-coil regions; 4, cata-
lytic domain. (B) Dendrogram of FYVE domains: TcrPDEC; HsEEA1, human early endosome antigen 1 (accession number Q15075); DmHrs,
Drosophila Hrs (Q960 ·8); ScVps27p, S. cerevisiae vacuolar sorting protein (P40343); HsDFCP1, human double FYVE-containing protein 1
(Q9HBF4); AtPRAF1, Arabidopsis PRAF1 (Q947D2); RnRPH3A, rat rabphilin-3 A (P47709); HsRIM1, human Rim1 (Q86UR5). (C) Alignment of
FYVE and FYVE-related domains. Grey boxes: the conserved Zn
2+
coordinating cysteine residues. The FYVE domain signature motifs WxxD,
R + HHCRxCG and RVC are given in bold, underlined letters. Box 1: turret loop [41]; box 2: a-helix found in the structures of HsEEA1,
DmHrs, ScVps27p and RnRabphilin-3 A. Horizontal box: putative dimer interface of EEA1 [42].
Novel phosphodiesterase from Trypanosoma cruzi S. Kunz et al.
6414 FEBS Journal 272 (2005) 6412–6422 ª2005 The Authors Journal compilation ª2005 FEBS
Downstream of the FYVE domain, TcrPDEC is pre-
dicted to contain two closely spaced coiled-coil regions
(S
150
–L
174
and K
207
–D
264
). These might serve to dimer-
ize the FYVE domains in a way similar to the struc-
ture that was determined for EEA1 [26].
To explore if these regions are indeed essential for
stabilizing the FYVE domain in the dimeric state, the
FYVE domain was expressed either alone (amino acids
1–74 of TcrPDEC), or in conjunction with the coiled-
coil region (amino acids 1–272). Gel filtration analysis
demonstrated that already the FYVE domain alone
migrates as a stable dimer (calculated molecular mass
8.2 kDa; apparent molecular mass upon gel filtration:
16.2 kDa) (Fig. 3A). The construct containing the two
coiled-coil regions in addition to the FYVE domain
(calculated molecular mass 30.5 kDa) eluted with an
apparent mass of 199.7 kDa, indicating the formation
of a higher order complex (Fig. 3B).
In TcrPDEC, the catalytic domain is located in the
middle of the polypeptide chain of 924 amino acids
(T
291
–S
657
; Fig. 2A). This is very unusual for a class I
PDE, as all other members of this PDE class contain
the catalytic domain in their C-terminal portions. Nev-
ertheless, the catalytic domain of TcrPDEC unambigu-
ously identifies it as a class I PDE (Fig. 4). This PDE
class includes all 11 human PDE families, and all
of its members contain the signature motif HD(LIV-
MFY)xHx(AG)xxNx(LIVMFY). Their catalytic
domains share 30–40% amino acid sequence identity
between families [17]. The overall sequence of the
TcPDE-FYVE catalytic domain conforms well with
that of other class I PDEs. It shares between 24 and
33% of amino acid sequence identity with the 11
human PDE families, and 45 and 57% sequence iden-
tity with its putative orthologs in L. major (lmjPDEC)
and T. brucei (TbrPDEC; unpublished data). All resi-
dues that have been identified as important for sub-
strate recognition, selectivity and catalysis in the
human PDEs [27,28] are conserved in TcrPDEC with
respect to HsPDE4B2 (Fig. 4). The two metal binding
sites are predicted to be formed by residues H
368
,H
372
,
H
409
,E
410
, His413, N
418
,L
438
,D
439
,E
481
,M
482
and
E
521
[28]. The hydrophobic pocket that accommodates
the purine moiety of the substrate is also conserved
and predicted to consist of Y
367
,I
522
,A
524
,S
525
,A
532
,
W
535
,L
536
,I
538
,L
539
,G
559
,S
564
,V
566
,S
569
,Q
570
, and
F
573
. Interestingly, of the two residues in this pocket
that contribute to selectivity for cAMP over cGMP
in HsPDE4, only Q
570
(corresponding to Q
443
in
HsPDE4B2) is conserved, while the residue corres-
ponding to N
395
in HsPDE4B2 is substituted by A
524
.
The side chain of this N
395
in the structure of
HsPDE4B forms two hydrogen bonds with adenine,
with the 6-NH
2
atom and with the 7-N ring atom,
while it might only form a single interaction with
either the 7-N atom or the 2-NH
2
atom of guanine.
While the presence of an N residue at this position
favors cAMP binding, it does not preclude cGMP
binding. However, in several of the human PDEs that
accept cGMP as a substrate, the position of N
395
is
substituted by an alanine (HsPDE5 and HsPDE6) or
glycine (HsPDE3 [27]). The presence of an alanine resi-
due at this position implies that TcrPDEC might be
capable of using cGMP as a substrate. This could be
experimentally confirmed (see below). Compared with
the set of 19 residues that are completely conserved
in all 11 human PDEs [27], four substitutions in
Fig. 3. Dimeric structure of the FYVE-type domain. (A) Gel filtration
(Superdex 75 PC 3.2 ⁄30) of an N-terminal fragment of TcrPDEC
containing the FYVE-variant domain through the coiled-coil region
(amino acids 1–272). (B) Gel filtration of FYVE-type domain alone
(amino acids 1–74 of TcrPDEC). Elution positions of the FYVE-con-
taining polypeptides are indicated by asterisks. Vertical black arrow:
elution position of TEF protease (27 kDa). Elution position of
molecular weight markers (open arrows): 1, aldolase (158 kDa); 2,
bovine serum albumin (67 kDa); 3, chymotrypsingen A (25.0 kDa);
4, aprotinin (6.5 kDa); 5, vitamin B12 (1.35 kDa).
S. Kunz et al. Novel phosphodiesterase from Trypanosoma cruzi
FEBS Journal 272 (2005) 6412–6422 ª2005 The Authors Journal compilation ª2005 FEBS 6415
TcrPDEC are notable. In the predicted helix 9, a con-
served alanine is substituted by S
429
in TcrPDEC. In
predicted helix 10, the first of the two vicinal histidines
is replaced by L
441
, and in predicted helix 11, a con-
served alanine is substituted by H
479
. Finally, in pre-
dicted helix 14, a conserved aspartate is replaced by
E
547
. The three substitutions represented by L
441
,H
479
and E
547
are specific for TcrPDEC. Another class1
PDE of T. cruzi, TcPDE1, that has a low K
m
and is
cAMP-selective [14], conforms to the mammalian PDE
pattern in all three positions. These three substitutions
in TcrPDEC may contribute to its dual-substrate spe-
cificity, and ⁄or to its relatively high K
m
for both sub-
strates (see below).
Functional complementation of a PDE-deficient
S. cerevisiae strain
Deletion of the two PDE genes ScPDE1 and ScPDE2
from the S. cerevisiae genome leads to an accumula-
tion of cAMP in the cells, leading to marked heat-
shock sensitivity [29]. Heterologous complementation
of the heat-shock sensitivity phenotype of PDE-defici-
ent yeast strains has proven to be a highly sensitive
functional validation for suspected PDE genes [20–32].
The full-size open reading frame of TcrPDEC, as well
as the predicted catalytic domain (T
291
–S
657
) were
expressed in the PDE-deficient S. cerevisiae strain PP5
[31]. In addition, the same two constructs carrying
Fig. 4. Sequence alignment of catalytic domains. Grey vertical boxes indicate residues that are identical in all human PDEs as well as in
TcPDE1 and TcrPDEC; open vertical boxes indicate residues that are conserved in all 11 human PDEs and in TcPDE1, but are substituted in
TcrPDEC; black dots represent residues that are necessary for positioning the catalytically important histidine residue (arrow); boxes indicate
the consensus helices; asterisk and bold, asparagine residue which confers cAMP preference (N
395
in HcPDE4B). m, metal binding pocket;
q, Q-domain [28].
Novel phosphodiesterase from Trypanosoma cruzi S. Kunz et al.
6416 FEBS Journal 272 (2005) 6412–6422 ª2005 The Authors Journal compilation ª2005 FEBS