The regulatory subunit of a cGMP-regulated protein kinase A of Trypanosoma brucei Tarek Shalaby, Matthias Liniger and Thomas Seebeck{ Institute of Cell Biology, University of Bern, Switzerland This study reports the identification and characterization of the regulatory subunit, TbRSU, of protein kinase A of the parasitic protozoon Trypanosoma brucei. TbRSU is coded for by a single copy gene. The protein contains an unusually long N-terminal domain, the pseudosubstrate site involved in binding and inactivation of the catalytic subunit, and two C-terminally located, closely spaced cyclic nucleotide binding domains. Immunoprecipitation of TbRSU copre- cipitates a protein kinase activity with the characteristics of protein kinase A: it phosphorylates a protein kinase specific substrate, and it is strongly inhibited by a synthetic protein kinase inhibitor peptide. Unexpectedly, this kinase activity could not be stimulated by cAMP, but by cGMP only. Binding studies with recombinant cyclic nucleotide binding domains of TbRSU confirmed that both domains bind cGMP with K d values in the lower micromolar range, and that up to a 100-fold excess of cAMP does not compete with cGMP binding. Keywords: sleeping sickness; protein kinase A; African trypanosomes; cyclic nucleotide signalling. The concept of cellular signaling by cyclic AMP (cAMP) has been maintained throughout evolution, from bacteria to mammals. However, the only component of this signalling pathway that has been strictly conserved is the second messenger molecule itself, cAMP, while the enzymatic machinery that generates and transduces the signal exhibits great variety. This is exemplified by the adenylyl cyclases, which have developed into many different molecular structures [1–3], although their function is invariably to convert ATP to cAMP. A similarly wide range of structure and sequence diversity of functionally similar enzymes is found within the cAMP-specific phosphodiesterases (PDEs). On the basis of sequence comparison as well as of pharmacological criteria, two distinct classes of eukaryotic PDEs are currently distinguished, class I and class II [4,5], with no significant sequence similarities between them. Besides these, many PDEs have been identified in bacteria that share no significant sequence homology with either the class I or the class II of the eukaryotic PDEs [6]. An even greater variety is encountered with the down- stream effectors of cAMP signalling. cAMP can bind directly to and regulate a number of different ion channels, such as cyclic nucleotide gated ion channels [7,8] or hyperpolarization-activated cyclic nucleotide gated chan- nels [9]. On the other hand, cAMP can bind to and stimulate drug efflux pumps, e.g. in the human erythrocyte [10]. Furthermore, recent data have demonstrated that the guanine nucleotide exchange factor Epac is a cAMP-binding protein [11], and that binding of cAMP modulates its activity. This interaction potentially allows a crosstalk between cAMP pathways and ras-mediated pathways in cell cycle control. In addition to its many roles as an intracellular messenger, cAMP also can act as an extracellular signalling molecule, either directly, as in the aggregation of the slime mold Dictyostelium discoideum [12], or indirectly via extracellu- lar conversion into adenosine and the subsequent activation of adenosine receptors in the brain [13]. In mammalian systems, the most extensively studied downstream effector of cAMP is the cAMP-regulated protein kinase A (PKA) [14 –18]. According to the current paradigm, PKA is an R 2 C 2 heterotetramer consisting of two catalytic and two regulatory subunits. The regulatory subunits contain a dimerization domain in their N-terminal regions, followed by an autoinhibitor sequence that resembles a PKA substrate. This region binds to the active site of the catalytic subunit, inactivating it while it is in the R 2 C 2 complex. The C-terminus of the regulatory subunit contains two adjacent cAMP-binding domains. Domain A is not accessible for cAMP in the R 2 C 2 complex. cAMP first binds to domain B, triggering a conformational change that renders domain A more accessible. The two cAMP-binding domains are biochemically distinct, both in terms of binding kinetics and in their preference for substituted cAMP analogs. The three-dimensional structure of the cAMP- binding domain of a bovine type I regulatory subunit has been determined [19]. Binding of cAMP to the regulatory subunits releases the active catalytic subunits from the complex. These proceed to phosphorylate a plethora of proteins, among them transcription factors such as CREB [20,21]. The current view is that most of the downstream effects of cAMP in eukaryotic cells are mediated through Note: a web site is available at http://www.izb.unibe.ch/res/seebeck/sehome.html Correspondence to T. Seebeck, Institute of Cell Biology, University of Bern, Baltzerstrasse 4, CH-3012 Bern, Switzerland. Fax: 1 41 31 631 46 84, Tel.: 1 41 31 631 46 49, E-mail: thomas.seebeck@izb.unibe.ch (Received 27 June 2001, revised 20 September 2001, accepted 1 October 2001) Abbreviations: PKA, protein kinase A; cGMP, cyclic guanosine monophosphate; cAMP, cyclic adenosine monophosphate; cNMP, cyclic nucleoside monophosphate; TbRSU, regulatory subunit of trypanosomal PKA. Eur. J. Biochem. 268, 6197–6206 (2001) q FEBS 2001 the alteration of transcription via PKA-mediated phos- phorylation of transcription factors. Interestingly, in at least in some instances, the activity of mammalian PKA appears to be stimulated by cGMP rather than by cAMP [22]. In the unicellular eukaryote Trypanosoma brucei, the causative agent of human sleeping sickness in Africa, cAMP signalling and its role in parasite proliferation and host/ parasite interaction are still poorly understood [23]. A large number of genes coding for different adenylyl cyclases have been identified [3,24], and one of these enzymes, GRESAG 4.4B, has been further characterized [25]. Also, several cAMP-specific phosphodiesterases have recently been iden- tified and characterized [26] (S. Kunz, P. Bern, A. Rascon, S. H. Soderling and J. Beavo, personal communication; A. Rascon and J. Beavo, personal communication, University of Seattle, WA, USA). Little is currently known about the biological role of cAMP signalling in these organisms. A role for cAMP in the differentiation of long, slender to short, stumpy forms in the bloodstream of the mammalian host has been proposed [27]. PKA activity has also been implicated in a mechanism by which T. brucei can remove bound host antibody from its cell surface [28]. The enzyme itself has not yet been characterized in any of the kinetoplastids, although previous work demonstrated the presence of a PKA-like kinase activity in T. cruzi [29]. The current study describes the identification and characterization of the regulatory subunit of trypanosomal PKA (TbRSU). Many of the structural features are well conserved between TbRSU and its mammalian counterparts. Despite this overall similarity between mammalian and trypanosomal regulatory subunits, the trypanosomal homo- log binds cGMP rather than cAMP, and the trypanosomal PKA is activated by cGMP, but not by cAMP. TbRSU thus represents yet another facet in the amazing kaleidoscope of cyclic nucleotide signalling. MATERIALS AND METHODS Materials Enzymes were obtained from Roche Diagnostics (Rotkreuz, Switzerland), and culture media were purchased from Difco. Radiochemicals were from Dupont-NEN (Regensdorf, Switzerland), while chemicals were obtained from SIGMA or Fluka (Buchs, Switzerland). Talonw immobilized-cobalt resin was from Clontech (Basel, Switzerland). DNA sequencing was outsourced to Microsynth GmbH, Balgach, Switzerland where the reactions were run with BigDye terminators (PE-Biosystems) and were analyzed on an ABI Prism 377 instrument. Cell culture T. brucei strain 427 (derived form MiTat 15a), was grown as procyclic forms at 27 8C in SDM medium [30]. Mono- morphic bloodstream forms of strain 221 (MiTat 1.2) were cultivated as described by Hesse et al. [31]. Drosophila Schneider 2 (S2) cells and expression vectors were obtained from Invitrogen (Carlsbad, CA, USA). Cells were passaged at cell densities between 6 and 20 Â 10 6 mL 21 by splitting at a 1 : 2 to 1 : 5 dilution in complete DES TM medium (Invitrogen) containing 10% heat-inactivated fetal bovine serum. S2 cells are density- sensitive and do not proliferate when seeded at less than 5 Â 10 5 mL 21 . Cells were cultured in a 22– 24 8C incubator with no extra CO 2 supplied. Cell viability was checked using the Trypan Blue exclusion test and was routinely found to be between 95 and 99%. Transfection of S2 cells S2 cells were prepared for transfection by seeding 3 Â 10 6 cells in 3 mL DES TM medium into a 35-mm Petri dish. The culture was incubated at 24 8C until a cell density of 2–4 Â 10 6 mL 21 was reached (6– 16 h). Immediately before transfection, the following two solutions were prepared separately (per 35-mm dish). Tube A: 36 mL2 M CaCl 2 and 19 mg vector DNA, in a final volume of 300 mL H 2 O. Tube B: 300 mL50mM Hepes, pH 7.1, 1.5 mM NaH 2 PO 4 , 280 mM NaCl. The contents of tube A were added slowly (over 1–2 min) to tube B under continued mixing. The final mixture was incubated at room temperature for 30– 40 min to allow the precipitate to form. The suspension was then well resuspended and added dropwise to the medium of the cell culture. After incubation for 16–24 h, cells were washed twice with medium to remove the calcium-phosphate precipitate, suspended in fresh growth medium, and further incubated. Expression of the recombinant protein was induced by the addition of 15 mL 100 m M CuSO 4 per 3 mL culture medium (final concentration 500 m M), and protein expression was assayed 12– 48 h after induction. When stable cell lines were desired, the cells were cotransfected with plasmid pCoHYGRO (Invitrogen) and were selected for growth in 300 mg : mL 21 hygromycin B. Preparation of the PKA-specific substrate An expression plasmid coding for a 28-kDa His 6 -tagged green fluorescent protein with a protein kinase A specific phosphorylation sequence (GFP227-RRRRSII) at its C-terminus was provided by K. Shokat, Princeton University, NJ, USA [32]. The plasmid was transfected into BL21DE, and positive colonies were identified by their fluorescence under UV light. Liquid cultures were grown to a D 595 of < 0.4 and were then induced for 4 h with 0.5 mM isopropyl thio-b-D-galactoside. Cells were suspended in 1–2% of the original culture volume of ice-cold 50 m M sodium phosphate, pH 7.0, 300 mM NaCl, and were lysed by sonication. The lysate was cleared by centrifugation for 20 min at 7000 g, and the recombinant protein was adsorbed batchwise to Talonw immobilized-cobalt resin (Invitrogen) and purified according to the manufacturer’s protocol. Immunoprecipitation For the preparation of antibody-coated beads, protein G–Sepharose beads (Amersham-Pharmacia) were washed twice in NaCl/P i and then suspended as a 50% slurry in 100 m M phosphate buffer, pH 8.2. Fifty-microliter aliquots of this slurry were incubated for 1 –3 h at 4 8C in 500 mL phosphate buffer containing the antibody to be coupled (rat polyclonal antibody against TbRSU1 or a control polyclonal rat antibody directed against an irrelevant protein). Beads were then washed twice in 100 m M phosphate buffer and once in HB buffer (25 m M Tris/HCl, pH 8.0, 50 mM NaCl). 6198 T. Shalaby et al. (Eur. J. Biochem. 268) q FEBS 2001 For immunoprecipitation, 2 Â 10 8 trypanosomes were sedimented at 1300 g for 10 min and were washed twice in ice-cold NaCl/P i (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na 2 HPO 4 , 1.4 mM KH 2 PO 4 , pH 7.3). The final pellet was suspended in 240 mL HB buffer and 30 mL Completew protease inhibitor mix (Roche Molecular Biochemicals). Thirty microliters lysis buffer (10% deoxycholate, 10% NP40 in HB buffer) were added, and the mixture was extensively vortexed. After 30 min incubation on ice, the lysate was centrifuged for 5 min at 10 000 g at 4 8C. 200 mL of the supernatant was transferred to a fresh tube containing 25 mL of the antibody-coated protein G– Sepharose beads. The slurry was gently rocked for 1 h to overnight at 4 8C. Beads were then washed on ice three times with cold WBI buffer (0.5% NP 40, 0.05% deoxycholate and 0.05% SDS in NaCl/P i , pH 7.5), twice with cold WBII buffer (125 m M Tris/HCl, pH 8.2, 500 mM NaCl, 1 mM EDTA, 0.5% NP40), and finally once with 500 mL kinase buffer (15 m M NaCl, 5 mM MgCl 2 ,10mM Hepes, pH 7.5). Protein kinase assay For assaying protein kinase activity in the immunoprecipi- tates, the following reaction mix was prepared: 1 mL [ 32 P]gATP ( 5 mCi : mL 21 , 150 mM ATP), 4 mLof5Â kinase buffer (75 m M NaCl, 25 mM MgCl 2 ,50mM Hepes, pH 7.5), 1 mL kinase substrate (0.5–1 mg), further additions as required, and H 2 O to a final volume of 20 mL. These 20 mL were added to 10 mLwashed immunoprecipitation beads (corresponding to 2 Â 10 6 trypanosomes), and the suspension was incubated at 30 8C for 30 min. The reaction was stopped by the addition of 5 mL5Â SDS sample buffer and boiling for 3 min. Expression of a recombinant GST–RSU fusion protein A sequence fragment of the TbRSU gene recovered from the T. brucei genome project at The Institute for Genetic Research (TIGR) was used to design two PCR primers (RSU-1, 5 0 -GAGAGTCGACGCTCAAGGTAGAAGGTA CGG-3 0 , and RSU-2, 5 0 -AGACTCGAGCTACTTCCTCCC CTCTGCCC-3 0 ; added SalI and XhoI restriction sites underlined, respectively). The expected 600-bp fragment was amplified from genomic DNA of T. brucei, confirmed by DNA sequencing and introduced into the multicloning site of the expression vector pGEX-4T2 (Amersham- Pharmacia). The vector was transformed into Escherichia coli BL21DE and the recombinant protein was expressed in high amounts in an insoluble form. The protein was solubilized from the inclusion bodies in 100 m M Tris/HCl, pH 7.5, 5 m M EDTA, 6 M urea. For renaturation, several procedures were tried, all of which lead to soluble fusion protein unable to bind to glutathione–Sepharose. Thus, the fusion protein was purified by gel filtration on a Superdex 200 column, followed by gel electrophoresis. After blotting the protein to nitrocellulose (Schleicher & Schuell BA 85), the 50-kDa fusion protein band was excised, dissolved in dimethylsulfoxide and used for immunization. Expression of cNMP-binding domains in Drosophila S2 cells For the expression of the cNMP-binding domains of TbRSU1 in S2 cells, the respective gene fragments were amplified and cloned into the pMT/V5-His B vector (Invitrogen). In this vector, expression is regulated by a metallothioneine promotor, and it allows induction of expression by the addition of Cu 21 to the growth medium. The recombinant proteins carry a V5 immunological tag and a His 6 -tag at their C-termini, which allow for easy detection and purification. The cNMP-binding domain A (amino acids 231–367) was amplified using primers Adom-F [5 0 - TAT ACTAGTATGG(2531)CACTCATCTTGAAGTTGT-3 0 , added Spe I site and start codon, bold underlined] and Adom-R [5 0 -TATCTCGAGA(2938rev)AGGCCACTGAG GAAC-3 0 , added Xho I site underlined]. Domain B (amino acids 352– 499) was amplified using primers Bdom-F [5 0 -TATACTAGTATGC(2921)CGTTCCTCAGTGG-3 0 , added SpeI site and start codon, bold underlined] and Bdom-R [5 0 -TATCTCGAG(3334rev)CTTCCTCCCCT CTG-3 0 , added XhoI site underlined]. For amplification of the joint domains (amino acids 231–499), primers Adom-F and Bdom-R were used. The PCR products were cloned into the pGEM T-Easy vector, verified by DNA sequencing and were finally subcloned into the expression vector pMT/ V5-His B. Purification of recombinant cNMP-binding domains from S2 Drosophila cell s Cells were collected by centrifugation at 500 g for 5 min at 4 8C. The cell pellet was suspended in ice-cold lysis buffer (50 mL per mL cell culture; 50 m M Tris/HCl, pH 7.8, 150 m M NaCl, 1% Nonidet P-40; Completew protease inhibitor cocktail was added immediately before use). The lysate was incubated on ice for 20–30 min, briefly homogenized in a glass/Teflon homogenizer and finally centrifuged at 7000 g for 20 min at 2 8C. To the cleared supernatant, 1/10 volume of a 50% (v/v) suspension of Talonw beads in NaCl/P i was added, and the suspension was incubated on a rocking platform for 3 h at 4 8C. After incubation, the suspension was poured into a small column and was washed extensively with 50 m M sodium phosphate buffer, pH 7.0, 300 m M NaCl. Recombinant protein was finally eluted with four aliquots of 100 mL elution buffer (50 sodium phosphate, pH 7.0, 300 m M NaCl, 150 mM imidazole). The protein containing fractions were pooled, aliquoted, snap-frozen and stored at 270 8C. Cyclic nucleotide binding assays Binding assays were performed in 5 m M sodium phosphate, pH 6.8, 1 m M EDTA, 25 mM 2-mercaptoethanol, 0.2 mM isobutyl-methyl-xanthine, 1.5 mg purified protein and increasing concentrations of [ 3 H]cGMP (NEN, catalogue no. NET-337) adjusted to a specific activity of 1 mCi : nmol 21 . cAMP competition and kinetic experiments were carried out in the presence of 0.4 m M [ 3 H]cGMP. Initial experiments were carried out in the presence of 500 mg : mL 21 histone VIII-S, which increased the binding efficiency by about 50%. Histone was omitted in later experiments. The binding reactions were incubated on ice q FEBS 2001 PKA regulatory subunit from T. brucei (Eur. J. Biochem. 268) 6199 overnight. Reactions were stopped by the addition of 1 mL ice cold 10 m M sodium phosphate, pH 6.8, 1 mM EDTA and were filtered immediately through prewetted Millipore HAWP filters (0.45 m M). Filters were rinsed three times with 1 mL ice-cold buffer each, thoroughly dried and counted in a toluene-based scintillator. Dissociation rate constants were determined by overnight equilibration on ice of the binding reaction containing 0.4 m M [ 3 H]cGMP. After the addition of a 100-fold excess of unlabelled cGMP, aliquots were withdrawn and processed for filtration at time points between 0 and 30 min. All reactions were done in triplicate. Binding parameters were determined by curve fitting using the PRISM software package of GraphPad Inc., San Diego, CA, USA. RESULTS Identification of TbRSU1 The DNA database of the T. brucei genome project was searched for predicted proteins containing putative cAMP- binding domains. This search resulted in a 600-bp DNA sequence which was predicted to code for the C-terminal fragment of a protein with high similarity to the regulatory subunits of eukaryotic PKAs. From the retrieved sequence, PCR primers were designed (see Materials and methods) and were used to amplify the corresponding fragment from genomic DNA of T. brucei. The resulting PCR fragment of 600 bp was cloned and verified by sequencing. It was then used to hybridize genomic blots of T. brucei DNA in order to establish the number of corresponding genes present in the genome. When genomic DNA was digested with enzymes that did not cut within the DNA sequence of the hybridization probe (Xho I, Stu I, Spe I, Pst I, Nhe I, Kpn I and HindIII), all digests resulted in a single hybridizing band (Fig. 1A), strongly indicating that the new gene, TbRSU, is coded for by a single-copy gene. The 600-bp PCR fragment was then used to screen a genomic library of T. brucei in a lambda phage vector [33]. This screening resulted in several independent phages containing the same Fig. 1. TbRSU is a single-copy gene. (A) Digests of genomic DNA of T. brucei were hybridized with a 600-bp PCR fragment representing the conserved cNMP-binding domain of TbRSU. (B) Map of the TbRSU locus. Nucleotides 1 – 483 code for the C-terminus of a protein of unknown function (TbTAS ). Nucleotides 1838–3334 represent the open reading frame of TbRSU. Nucleotides 3335– 3566 represent a part of the 3 0 untranslated region of TbRSU. The grey boxes designated A and B represent the predicted cyclic nucleotide binding domains of the TbRSU protein. The sequence has been deposited at GenBank under the accession number AF326975. Fig. 2. Gene and amino-acid sequence of TbRSU. The pseudosub- strate sequence is indicated by the grey box. The two cyclic-nucleotide binding domains A and B are boxed. Shaded boxes in domain A: Glu311 (is Ala in all homologs, see Fig. 3); Thr318 (is Arg in all homologs); Val319 (is Ala in cAMP and Thr in cGMP binding domains). Shaded boxes in domain B: Glu435 (is Ala in all homologs); Asn442 (is Arg in all homologs). 6200 T. Shalaby et al. (Eur. J. Biochem. 268) q FEBS 2001 locus. A 3-kb Eco RI fragment was subcloned into pBlueskriptSK1 and both strands were completely sequenced. The sequence analysis demonstrated that this fragment contained the entire open reading frame of the TbRSU gene (Fig. 1B). In parallel, a cDNA library of procyclic T. brucei was also screened with the same PCR fragment, resulting in three independent phages that all contained a 1500-bp cDNA fragment. All three were sequenced and were shown to contain a short 5 0 untranslated region, a complete open reading frame of 1497 bp, and a 3 0 untranslated region terminated by a polyA tract. Although all three cDNA clones were terminated with this sequence, this polyA tract probably does not represent the polyA tail of the mRNA because a sequence of 12 adenosine residues following T3566 was is present in a genomic clone of the T. brucei genome project (accession number AQ 644384) that extends beyond this region. The sequences of the open reading frames of all three cDNAs were identical to that obtained from the genomic fragment. Upstream of the TbRSU gene, the 3 0 end of an open reading frame was identified (nucleotides 1–486 of the genomic fragment), which coded for an unidentified protein termed TbTAS. The stop codon of this open reading frame is separated from the start codon of TbRSU by 1352 bp, including a pyrimidine-rich region. Predicted amino-acid sequence of TbRSU The open reading frame of TbRSU predicts a protein of 499 amino acids, with a calculated M r of 56 725 (Fig. 2). Overall, the protein shares extensive sequence homology with mammalian PKA regulatory subunits type I. The N-terminal domain of TbRSU (amino acids 1–242) is longer than the N-termini of its mammalian homologs, and it bears no identifiable functional domains. In analogy to mammalian type I regulatory subunits, the cysteine residues Cys15 and Cys67 may be involved in dimer formation, although such dimers could not be detected in cell lysates analysed by gel filtration chromatography (data not shown). In these experiments, TbRSU always migrated as a monomer. Residues 202–206 (-ArgArgThrThrVal-) rep- resent the pseudo-inhibitor site which is involved in the interaction with the catalytic domain [34]. Amino acids 243– 360 and 363–483 form the cyclic nucleotide binding domains A and B, respectively. Based on the structural model of the bovine regulatory subunit RIa [17], Glu309 and Glu433 form a hydrogen bond with the 2 0 hydroxyl of the ribose of the bound cNMP, while Leu310 and Leu434 interact with a nitrogen of the pyrimidine ring of the base. Tyr370 and Tyr482 are probably the functional homologs of Trp260 and Tyr371 in bovine RIa, allowing base-stacking with the purine residue. Unexpectedly, a strongly conserved arginine residue, which forms a hydrogen bond to the phosphate group, is replaced by threonine (Thr318) and asparagine (Asn442) in domains A and B, respectively (Fig. 3). Sequencing errors or allelic variation at these sites are unlikely as identical sequences have been obtained by independent sequencing of TbRSU from different trypano- some strains (accession nos AQ638897 and AF182823). A further difference between TbRSU and the PKA regulatory subunits from other eukaryotes is seen in Val319. All cAMP-binding domains of the regulatory subunits carry an alanine residue at this position, while the closely related cGMP-binding domains of protein kinase G always contain either threonine or serine residues. TbRSU mRNA is more abundant in bloodstream forms To explore if TbRSU is differentially expressed in the different life stages of T. brucei, total RNA was extracted both from bloodstream and from procyclic forms and was analyzed by Northern blotting and hybridization. RNA loading was quantitated by ethidium bromide staining to visualize the ribosomal RNA before blotting the gel, and by hybridization of the filter with a DNA probe specific for b-tubulin [35]. The extent of hybridization of both probes was quantitated using a PhosphorImager. TbRSU mRNA is clearly detectable in both life cycle stages (Fig. 4A). Fig. 3. Sequence comparison of cNMP-binding domains of PKA regulatory subunits and of protein kinase G. cNMP-binding domains A and B are indicated by grey boxes. Amino-acid numbering of the respective proteins is given. A: Rattus norvegicus type I (accession number P09456); B: D. melanogaster (P16905); C: Caenorhabditis elegans (P30625); D: D. discoideum (P05987); E: S. cerevisiae (P07278); F: Schizosaccharomyces pombe (P36600); G: TbRSU (AF326975); H: Homo sapiens protein kinase G (O13237); I: D. melanogaster protein kinase G (O03043). Filled circles denote amino acids conserved in all sequences. Open squares denote amino acids which are conserved in all sequences, but differ in TbRSU. q FEBS 2001 PKA regulatory subunit from T. brucei (Eur. J. Biochem. 268) 6201 However, the steady-state level of TbRSU mRNA in blodstream forms is about five times higher than it is in procyclic forms. The TbRSU protein is present both in bloodstream and in procyclic forms To follow up the results of the Northern blotting experiments on the protein level, whole cell lysates were analyzed by immunoblotting, using an affinity-purified polyclonal antibody raised against recombinant TbRSU (see Materials and methods). This polyclonal antibody not only recognizes TbRSU in trypanosomes, but it also detects PKA regulatory subunits in other organisms such as Saccharomyces cerevisiae and mammalian cells (Fig. 4B). The TbRSU protein is readily detectable both in bloodstream and in procyclic forms, and it migrates as a single band of a M r of 55 000, in agreement with its calculated M r of 56 726. Similarly to what was observed with TbRSU mRNA, the TbRSU protein is much more abundant in bloodstream than in procyclic forms. Co-immunoprecipitation of PKA with TbRSU Sequence analysis clearly established TbRSU as a homolog of the type I regulatory subunits of mammalian PKA. In order to functionally verify if TbRSU is associated with a kinase in vivo, TbRSU was immunoprecipitated from whole cells lysates using the polyclonal rat antibody. Immunopre- cipitates were first analyzed by immunoblotting with a polyclonal rabbit antibody against the catalytic subunit of bovine PKA. In these experiments, the antibody detected a protein with a M r of about 40 000, suggesting that the catalytic subunit of trypanosomal PKA does in fact coprecipitate with TbRSU. Inspection of the T. brucei databases identified several DNA sequences that code for a homolog of a PKA catalytic subunit. The catalytic activity of the immunoprecipitates was then analysed by incubation in kinase reaction buffer in the presence or absence of a recombinant PKA-specific substrate [32] and 20 m M cAMP. Analysis of the reaction products by gel electrophoresis and autoradiography (Fig. 5) demonstrated that the coimmuno- precipitates did indeed contain a kinase activity which phosphorylated the PKA-specific substrate. No phosphoryl- ation of the substrate was observed when either no antibody, or an irrelevant antibody, was used for immunoprecipitation, or when the TbRSU antibody was used in the absence of cell lysate. Unexpectedly, the addition of 20 m M cAMP to the reactions did not stimulate the kinase activity, but had either Fig. 4. TbRSU is more abundant in bloodstream than in procyclic forms. (A) Northern blot analysis. Ten-microgram aliquots of total RNA of procyclic (PC) or bloodstream form (BSF) trypanosomes were loaded per slot. After transfer, the filter was successively hybridized with a TbRSU probe (a) and a probe for b-tubulin (b). After electrophoresis, the gel was stained with ethidium bromide to control for equal loading (c). (B) Hybridization was quantified using a PhosphorImager. (a) Hybridization with TbRSU; (b) hybridization with a b-tubulin probe. Grey bars, procyclics; black bars, bloodstream forms. (C) Immunoblot analysis. (a) The polyclonal antibody raised against recombinant TbRSU recognizes homologs in a wide spectrum of species. 1, Whole cell lysate from E. coli expressing the GST– TbRSU fusion protein used for raising the antibody; 2, whole cell lysate of T. brucei; 3, whole cell lysate of S. cerevisiae; 4, whole cell lysate of COS (monkey) cells. (b) Immunoblot of equivalent amounts of whole cell lysates of bloodstream (B) and procyclic (P) trypanosomes. Molecular mass markers are indicated for each panel. Fig. 5. The TbRSU antibody coimmunoprecipitates a protein kinase activity which phosphorylates a PKA-specific substrate. Protein kinase activity assays of immunoprecipitates. (Top) Coomassie- stained gels, molecular mass markers are IgG heavy chain (50 kDa) and the PKA substrate (30 kDa); (bottom) corresponding autoradiographs, arrowheads indicate the position of the PKA substrate. (A) Immunoprecipitation with no antibody; (B) immunoprecipitation with TbRSU antibody; (C) immunoprecipitation with irrelevant antibody (against the phosphodiesterase TbPDE1; S. Kunz, personnal communi- cation); (D) immunoprecipitation with TbRSU antibody, but without cell lysate. Beads were incubated for activity assays as follows: lanes 1: kinase buffer; lanes 2: kinase buffer plus 20 m M cAMP; lanes 3: kinase buffer plus PKA-substrate; lanes 4: kinase buffer plus PKA-substrate plus 20 m M cAMP 6202 T. Shalaby et al. (Eur. J. Biochem. 268) q FEBS 2001 no effect or inhibited it. While the absence of stimulation by cAMP was consistent in all of the many independent experiments carried out (see also below), the inhibitory effect of cAMP was observed in some, but not in all experiments. Phosphorylation of the PKA-specific substrate by the immunoprecipitates was time-dependent, Mg 21 -dependent and was quenched by an excess of unlabelled ATP (data not shown). These results demonstrated that a protein kinase activity was coimmunoprecipitated with TbRSU under our conditions. Phosphorylation of the PKA-specific substrate by this activity suggested that it represented PKA. This was further corroborated by the observation that the co-immunoprecipitating kinase activity was inhibited by the highly PKA-specific peptide inhibitor PKI [36] (Fig. 6). PKA activity is stimulated by cGMP, but not by cAMP When kinase activity of TbRSU immunoprecipitates was assayed in the presence or absence of 20 m M cAMP, no stimulation of phosphorylation of the PKA-specific substrate could be detected. In contrast, control reactions using mammalian COS cell lysates precipitated by the same antibody, exhibited the expected stimulation of kinase activity by cAMP (Fig. 7A). This unexpected absence of stimulation of trypanosomal PKA activity by cAMP was consistently observed over many independent experiments (see above). However, when similar experiments were performed with cGMP instead of cAMP, a marked stimulation of kinase activity was consistently observed (Fig. 7B–D). The phosphorylation reactions followed a similar time course in the presence and in the absence of cGMP (Fig. 7B), but the overall kinase activity was stimulated threefold to fourfold by cGMP. The kinase reaction was stimulated to a similar extent in immunoprecipitates from procyclic and bloodstream form trypanosomes (Fig. 7C,D), with maxi- mum stimulation reached around 20 m M cGMP. These unexpected findings suggested that the trypanosomal TbRSU, in contrast to its homologs in all other eukaryotes analysed so far, is activated by cGMP rather than by cAMP. cGMP binding to the cyclic nucleotide binding domains A and B sites of TbRSU In order to directly confirm if TbRSU does in fact bind cGMP, expression of the recombinant domains was attempted in E. coli. Expression of domain B alone produced ample recombinant protein, but all in insoluble form. Expression of the combined A and B domains resulted in much less protein (all insoluble). Expression of domain A alone proved impossible, despite much effort, in agreement with earlier observations that this domain is highly toxic for E. coli [37]. Thus, domains A and B were expressed individually in the Drosophila cell line S2, under the control of a Cu 21 -inducible metallothionein promoter. Similarly to what was observed in E. coli, domain B was well expressed, while domain A again resulted in very poor cell growth and in low amounts of recombinant protein. The individual domains A and B were purified by cobalt-affinity chromatography, and were assayed for cGMP binding Fig. 6. PKI inhibits the activity of coimmunoprecipitating kinase. Immunoprecipitates were incubated for 10 min under phosphorylation conditions with PKA substrate in the presence or absence of PKI inhibitor peptide (10 mg per 30 mL reaction mix). (A) autoradiogram of PKA substrate; (B) Coomassie-stained PKA substrate; (C) Phosphor- Imager analysis of the gel shown in (A). Fig. 7. The kinase activity which coimmunoprecipitates with TbRSU is stimulated by cGMP, but not by cAMP. (A) Whole cell lysates from mammalian COS cells and from T. brucei were immunoprecipitated with antibody against TbRSU, and the immuno- precipitates were assayed for PKA activity in the presence or absence of 20 m M cAMP. (B) Time course of kinase activity of immunoprecipitates from T. brucei in the presence (grey boxes) or absence (white boxes) of cGMP. (C and D) Effect of increasing cGMP concentrations on the kinase activity of immunoprecipitates (autoradiographs). (C 0 and D 0 ) Coomassie stained PKA substrate. bl, blank reaction incubated in the presence of 20 m M cGMP, but without protein substrate. (C and C 0 ) procyclics; (D and D 0 ) bloodstream forms. q FEBS 2001 PKA regulatory subunit from T. brucei (Eur. J. Biochem. 268) 6203 (Fig. 8). Both domains exhibited very similar K d values for cGMP (domain A: 7.51 ^ 1.97 m M, n ¼ 3; domain B: 11.43 ^ 2.24 m M, n ¼ 3). For both domains, cAMP did not measurably compete with cGMP binding up to a 100-fold excess of cAMP over cGMP. Dissociation rate constants for cGMP were also very similar between the two domains (domain A 0.24 min 21 , n ¼ 3, and domain B 0.36 ^ 0.18 min 21 , n ¼ 3). DISCUSSION The current study reports the identification of the regulatory subunit of PKA from the parasitic protozoon T. brucei, TbRSU. Several previous attempts to purify the PKA holoenzyme from this organism had failed, although an activity resembling the catalytic subunit could be identified [29]. Similarly, attempts in several laboratories, including our own, to demonstrate cAMP-specific protein phosphoryl- ation in T. brucei were unsuccessful. TbRSU was originally identified by searching of the T. brucei sequence databases for putative cAMP-binding proteins. The full gene was then isolated by screening genomic and cDNA libraries. Sequence analysis demonstrated that TbRSU is closely related to the mammalian type I PKA regulatory subunits, with the only major difference being the significantly longer N-terminus of the trypanoso- mal protein. The two cyclic nucleotide binding domains exhibit sequence similarities with both the cAMP-binding domains of the PKA regulatory subunits from yeast to mammals, as well as with the cGMP-binding domains of protein kinase G. Unexpectedly, one absolutely conserved arginine residue in each of the two domains is replaced by Thr318 and Asn442 in TbRSU. In the bovine regulatory subunit, and by inference also in all its homologs, these arginine residues form a hydrogen bond to the phosphate group of the bound nucleotide [19]. Sequencing errors can be ruled out as a simple reason for this variation, as this region was independently sequenced by three different laboratories using different trypanosome strains. The functional implication of this amino-acid substitution remains to be explored. Eight amino acids before Thr318 and Asn442, another amino-acid substitution peculiar to TbRSU has occurred: Glu311 and Glu435 replace otherwise invariant alanine residues. Thirdly, Val319 represents another substitution that sets TbRSU apart from its homologs. At the equivalent position, the other PKA regulatory domains contain an alanine residue while protein kinases G contain serine or threonine. The hydroxyl side chain of either one of these residues interacts with the C2 amino group of cGMP and is essential for full activation of cGMP dependent protein kinases [38]. The gene encoding TbRSU is expressed both in the bloodstream and in the procyclic forms of the parasite, but at much higher levels in the bloodstream form. The TbRSU protein levels in both life cycle stages closely correspond to the mRNA levels. Immunoprecipitation of TbRSU consistently coprecipi- tated a protein kinase activity exhibiting many character- istics of the catalytic subunit of trypanosomal PKA. The coprecipitated kinase is recognized by an antibody against the bovine PKA catalytic subunit, the enzyme phosphorylates a PKA-specific substrate [32], and its activity is strongly inhibited by the rabbit PKI inhibitor peptide [36]. While the protein kinase activity recovered in the immunoprecipitates exhibited all the characteristics of a canonical PKA catalytic subunit, no stimulation by cAMP could be detected. On the contrary, cAMP appeared to inhibit the protein kinase activity in some, but not all, experiments. Surprisingly, a marked stimulation of protein kinase activity was consistently found with cGMP. This stimulation was concentration-dependent, reaching its maximum at < 20 m M cGMP. The interaction of TbRSU with cyclic nucleotides was further investigated using the recombinant cNMP-binding domains A and B. Both domains did bind cGMP with K d values in the low micromolar range (7.5 and 11.4 m M, respectively). This value is unexpectedly high when compared to the K d values determined for cAMP of mammalian PKA regulatory subunits (1.2 and 1.7 n M for domains A and B, respectively [39]). However, the results are in good agreement with the PKA activation experiments presented in this study, which exhibited a maximal activation of the kinase at about 20 m M cGMP. This value is almost 200-fold higher than the apparent activation constant of mammalian PKA (120 n M; [17]). Binding of cGMP was not affected by cAMP, up to an excess of at least 100-fold. Again, the results agree well with the observations that the kinase activity was not stimulated by cAMP at concentrations of up to 20 m M. In marked contrast to the mammalian regulatory subunit where the two domains differ considerably in their dissociation rate constants (0.15 min 21 vs. 0.04 min 21 [17]), both domains of TbRSU behave very similarly (0.24 min 21 for domain A and 0.36 min 21 for domain B). The observation that protein kinase A in T. brucei (and probably also in other kinetoplastids) is regulated by cGMP rather than by cAMP implies that cGMP has an important signalling role in this group of organisms. Earlier work had demonstrated the presence of cGMP in T. cruzi [40], and several members of a family of recently identified cAMP- specific phosphodiesterases of T. brucei [26] (A. Rascon, S. H. Soderling & J. Beavo, personal communication) contain one or two GAF-domains [41] that may be involved in cGMP binding. While these phosphodiesterases may represent an interconnection between the cAMP- and the cGMP-signalling pathways in T. brucei, the cGMP-regu- lated TbRSU/PKA kinase may well represent the major effector of cGMP signalling in these organisms. Fig. 8. The cNMP-binding domains of TbRSU bind cGMP, but not cAMP. Saturation binding of of cGMP to recombinant domains A and B of TbRSU. Data represent one of three very similar experiments. 6204 T. Shalaby et al. (Eur. J. Biochem. 268) q FEBS 2001 ACKNOWLEDGEMENTS We are grateful to Kevin Shokat (Princeton University, Princeton, NJ, USA) for providing his plasmid for the expression of recombinant PKA substrate, to Brian Hemmings (Friedrich Miescher Institute, Basel) for his generous supply of antibody against bovine heart PKA catalytic subunit, and to Ursula Kurath and Erwin Studer for producing the trypanosomes. Special thanks go to Min Ku for her careful reading of the manuscript, and to Michael Boshart (Free University, Berlin) for communicating unpublished results and for stimulating discussions. This work was supported by grants 31-046760.96 and 31-058927.99 of the Swiss National Science Foundation, grant C98.0060 of COST program B9 of the European Union, and by the UNDP/World Bank/ WHO Special Programme for Research and Training in Tropical Diseases REFERENCES 1. Hurley, J.H. (1998) The adenylyl and guanylyl cyclase superfamily. Curr. Opin. Struct. Biol. 8, 770–777. 2. Chen, Y., Cann, M.J., Litvin, T.N., Iourgenko, V., Sinclair, M.J., Livin, L.R. & Buck, J. (2000) Soluble adenylyl cyclase as an evolutionarily conserved bicarbonate sensor. Science 289, 625–628. 3. Naula, C. & Seebeck, T. (2000) Cyclic AMP signaling in trypanosomatids. Parasitol. Today 16, 35–38. 4. Nikawa, J., Sass, P. & Wigler, M. (1987) Cloning and characterization of the low-affinity cyclic AMP phosphodiesterase gene of Saccharomyces cerevisiae. Mol. Cell. Biol. 7, 3629– 3636. 5. Sass, P., Field, J., Nikawa, J., Toda, T. & Wigler, M. (1986) Cloning and characterization of the high-affinity cAMP phosphodiesterase of Saccharomyces cerevisiae. Proc. Natl Acad. Sci. USA 83, 9303–9307. 6. Imamura, R., Yamanaka. K., Ogura, T., Hiraga, S., Fujita, N., Ishihama, A. & Niki, H. (1996) Identification of the cpdA gene encoding cyclic 3 0 5 0 -adenosine monophosphate phosphodiesterase in Escherichia coli. J. Biol. Chem. 271, 25423–25429. 7. Kraus-Friedman, N. (2000) Cyclic nucleotide gated channels in non-sensory organs. Cell Calcium 27, 127–138. 8. Broillet, M.C. & Firestein, S. (1999) Cyclic nucleotide-gated channels. Molecular mechanisms of activation. Ann. NY Acad. Sci. 868, 730–740. 9. Clapham, D. (1998) Not so funny anymore: pacing channels are cloned. Neuron 21, 5–7. 10. Schultz, C., Vaskinn, S., Kidalsen, H. & Sager, G. (1998) Cyclic AMP stimulates the cyclic GMP egression pump in human erythrocytes: effects of probenecid, verapamil, progesterone, theophylline, IBMX, forskolin and cyclic AMP on cyclic GMP uptake and association to inside-out vesicles. Biochemistry 37, 1161–1166. 11. De Rooji, J., Zwartkruis, F.J.T., Verhekgen, M.H.G., Cool, R.H., Nijman, S.M.B., Witinghofer, A. & Bos, J.L. (1998) Epac is a Rap1 guanine nucleotide exchange factor directly activated by cyclic AMP. Nature 396, 474–477. 12. Weeks, G. (2000) Signalling molecules involved in cellular differentiation during Dictyostelium morphogenesis. Curr. Opin. Microbiol. 3, 625–639. 13. Rosenberg, P.A. & Li, Y. (1995) Adenylyl cyclase activation underlies intracellular cyclic AMP accumulation, cyclic AMP transport, and extracellular adenosine accumulation evoked by beta-adrenergic receptor stimulation in mixed cultures of neurons and astrocytes derived from rat cerebral cortex. Brain Res. 692, 227–232. 14. Taylor, S.S., Radzio-Andzelm, E., Madhusudan, A., Cheng, X., Ten Eyck, L. & Narayana, N. (1999) Catalytic subunit of cyclic AMP dependent protein kinase: structure and dynamics of the active site cleft. Pharmacol. Ther. 82, 133 –141. 15. Francis, S.H. & Corbin, J.D. (1999) Cyclic nucleotide-dependent protein kinase: intracellular receptors for cAMP and cGMP action. Crit. Rev. Lab. Sci. 36, 275–328. 16. Thevelein, J.M. & de Winde, J.H. (1999) Novel sensing mechanisms and targets for the cAMP-protein kinase A pathway in the yeast Saccharomyces cerevisiae. Mol. Microbiol. 33, 904–918. 17. Francis, S.H. & Corbin, J.D. (1994) Structure and function of cyclic nucleotide-dependent protein kinases. Annu. Rev. Physiol. 56, 237–272. 18. Taylor, S.S., Buechler, J.A. & Yonemoto, W. (1990) cAMP- dependent protein kinase: framework for a diverse family of regulatory enzymes. Annu. Rev. Biochem. 59, 971–1005. 19. Su, Y., Dostmann, W.R.G., Herberg, F.W., Durick, K., Xuong, N.H., Ten Eyck, L., Taylor, S.S. & Varughese, K.I. (1995) Regulatory subunit of protein kinase A: structure of deletion mutant with cAMP binding domains. Science 269, 807–813. 20. De Cesare, D. & Sassone-Corsi, P. (2000) Transcriptional regulation by cyclic AMP-responsive factors. Prog. Nucleic Acid Res. Mol. Biol. 64, 343–369. 21. Shaywitz, A.J. & Greenberg, M.E. (1999) CREB: a stimulus- induced transcription factor activated by a diverse array of extracellular signals. Annu. Rev. Biochem. 68, 821 – 861. 22. Forte, L.R., Thorne, P.K., Eber, S.L., Krause, W.J., Freeman. R.H., Francis, S.H. & Corbin, J.D. (1992) Stimulation of intestinal Cl-transport by heat-stable enterotoxin: activation of cAMP dependent protein kinase by cGMP. Am. J. Physiol. 263, C607–C615. 23. Seebeck, T., Gong, K.W., Kunz, S., Schaub, R., Shalaby, T. & Zoraghi, R. (2001) cAMP signalling in T. brucei. Int. J. Parasitol. 31, 491–498. 24. Alexandre, S., Paindovoine, P., Hanocq-Quertier, J., Paturiaux- Hanocq, F., Tebabi, P. & Pays, E. (1996) Families of adenylate cyclase genes in Trypanosoma brucei. Mol. Biochem. Parasitol. 77, 173–182. 25. Naula, C., Schaub, R., Leech, V., Melville, S. & Seebeck, T. (2001) Spontaneous dimerization and leucine-zipper induced activation of the recombinant catalytic domain of a new adenylyl cyclase, GRESAG4.4B. Mol. Biochem. Parasitol. 112, 19 – 28. 26. Zoraghi, R., Kunz, S., Gong, K.W. & Seebeck, T. (2001) Characterization of TbPDE2A, a novel cyclic nucleotide specific phosphodiesterase from the protozoan parasite Trypanosoma brucei. J. Biol. Chem. 276, 11559–11566. 27. Vassella, E., Reuner, B., Yutzy, B. & Boshart, M. (1997) Differentiation of African trypanosomes is controlled by a density sensing mechanism which signals cell cycle arrest via the cAMP pathway. J. Cell Sci. 110, 2661 –2671. 28. O’Beirne, C., Lowry, C.M. & Voorheis, H.P. (1998) Both IgM and IgG anti-VSG antibodies initiate a cycle of aggregation- disaggregation of bloodstream forms of Trypanosoma brucei without damage to the parasite. Mol. Biochem. Parasitol. 91, 165–193. 29. Ochatt, C.M., Ulloa, R.M., Torres, H.N. & Tellez. Inon, M.T. (1993) Characterization of the catalytic subunit of Trypanosoma cruzi cyclic AMP-dependent protein kinase. Mol. Biochem. Parasitol. 57, 73 –81. 30. Brun, R. & Scho ¨ nenberger, M. (1979) Cultivation and in vitro cloning of procyclic culture forms of Trypanosoma brucei in a semi-defined medium. Acta Tropica 36, 289– 292. 31. Hesse, F., Selzer, P.M., Mu ¨ hlstadt, K. & Duszenko, M. (1995) A novel cultivation technique for long-term maintenance of blood- stream form trypanosomes in vitro. Mol. Biochem. Parasitol. 70, 157–166. 32. Yang, F., Liu, Y., Bixby, S.D., Friedman, J.D. & Shokat, K. (1999) Highly efficient green fluorescent protein-based kinase substrates. Anal. Biochem. 266, 167–173. 33. Schla ¨ ppi, K., Deflorin, J. & Seebeck, T. (1989) The major q FEBS 2001 PKA regulatory subunit from T. brucei (Eur. J. Biochem. 268) 6205 component of the paraflagellar rod of Trypanosoma brucei is a helical protein that is encoded by two identical, tandemly linked genes. J.Cell Biol. 109, 1695 –1709. 34. Zetterqvist, O. & Ragnarsson, U. (1982) The structural require- ments of substrates of cyclic AMP-dependent protein kinase. FEBS Lett. 139, 287–290. 35. Seebeck, T., Whittaker, P.A., Imboden, M.A., Hardman, N. & Braun, R. (1983) Tubulin genes of Trypanosoma brucei: a tightly clustered family of alternating genes. Proc. Natl Acad. Sci. USA 80, 4634–4638. 36. Cheng, H.C., Kemp, B.E., Pearson, R.B., Smith, A.J., Misconi, L., Van Patten, S.M. & Walsh, D.A. (1986) A potent synthetic peptide inhibitor of the cAMP-dependent protein kinase. J. Biol. Chem. 261, 989–992. 37. Gosse, M.E., Padmanabhan, A., Fleischmann, R.D. & Gottesmann, M.M. (1993) Expression of Chinese hamster cAMP-dependent protein kinase in Escherichia coli results in growth inhibition of bacterial cells: a model system for the rapid screening of mutant type I regulatory subunits. Proc. Natl Acad. Sci. USA 90, 8159–8163. 38. Taylor, M.K. & Uhler, M.D. (2000) The amino-terminal cyclic nucleotide binding site of the type II cGMP-dependent protein kinase is essential for full cyclic nucleotide-dependent activation. J. Biol. Chem. 275, 28053–28062. 39. Døskeland, S.O., Øgreid, D., Ekanger, R.E., Sturm, P.A., Miller, J.P. & Suva, R.H. (1983) Mapping of the two intrachain cyclic nucleotide binding sites of adenosine cyclic 3 0 ,5 0 -dependent protein kinase I. Biochemistry 22, 1094–1101. 40. Paveto, C., Pereira, C., Espinosa, J., Montagna, A.E., Farber, M., Esteva, M., Flawia, M.M. & Torres, H.N. (1995) The nitric oxide transduction pathway in Trypanosoma cruzi. J. Biol. Chem. 270, 16576–16579. 41. Ho, Y.S., Burden, L.M. & Hurley, J.H. (2000) Structure of the GAF domain, a ubiquitous signaling motif and a new class of cyclic GMP receptor. EMBO J. 19, 5288–5299. 6206 T. Shalaby et al. (Eur. J. Biochem. 268) q FEBS 2001 . least in some instances, the activity of mammalian PKA appears to be stimulated by cGMP rather than by cAMP [22]. In the unicellular eukaryote Trypanosoma brucei, the causative agent of human sleeping. The regulatory subunit of a cGMP-regulated protein kinase A of Trypanosoma brucei Tarek Shalaby, Matthias Liniger and Thomas Seebeck{ Institute of Cell Biology, University of Bern, Switzerland This. regulatory subunits, the trypanosomal homo- log binds cGMP rather than cAMP, and the trypanosomal PKA is activated by cGMP, but not by cAMP. TbRSU thus represents yet another facet in the amazing