Phosphorylation-dependent protein interaction with Trypanosoma brucei 14-3-3 proteins that display atypical target recognition.

BACKGROUND
The 14-3-3 proteins are structurally conserved throughout eukaryotes and participate in protein kinase signaling. All 14-3-3 proteins are known to bind to evolutionally conserved phosphoserine-containing motifs (modes 1 and/or 2) with high affinity. In Trypanosoma brucei, 14-3-3I and II play pivotal roles in motility, cytokinesis and the cell cycle. However, none of the T. brucei 14-3-3 binding proteins have previously been documented.


METHODOLOGY/PRINCIPAL FINDINGS
Initially we showed that T. brucei 14-3-3 proteins exhibit far lower affinity to those peptides containing RSxpSxP (mode 1) and RxY/FxpSxP (mode 2) (where x is any amino acid residue and pS is phosphoserine) than human 14-3-3 proteins, demonstrating the atypical target recognition by T. brucei 14-3-3 proteins. We found that the putative T. brucei protein phosphatase 2C (PP2c) binds to T. brucei 14-3-3 proteins utilizing its mode 3 motif (-pS/pTx(1-2)-COOH, where x is not Pro). We constructed eight chimeric PP2c proteins replacing its authentic mode 3 motif with potential mode 3 sequences found in Trypanosoma brucei genome database, and tested their binding. As a result, T. brucei 14-3-3 proteins interacted with three out of eight chimeric proteins including two with high affinity. Importantly, T. brucei 14-3-3 proteins co-immunoprecipitated with an uncharacterized full-length protein containing identified high-affinity mode 3 motif, suggesting that both proteins form a complex in vivo. In addition, a synthetic peptide derived from this mode 3 motif binds to T. brucei 14-3-3 proteins with high affinity.


CONCLUSION/SIGNIFICANCE
Because of the atypical target recognition of T. brucei 14-3-3 proteins, no 14-3-3-binding proteins have been successfully identified in T. brucei until now whereas over 200 human 14-3-3-binding proteins have been identified. This report describes the first discovery of the T. brucei 14-3-3-binding proteins and their binding motifs. The high-affinity phosphopeptide will be a powerful tool to identify novel T. brucei 14-3-3-binding proteins.


Introduction
Trypanosoma brucei is the causative agent of sleeping sickness in man and nagana disease in cattle and one of the most divergent eukaryotes from mammals. The disease is spread by the tsetse fly, in which the procyclic forms (PCF) proliferate and differentiate into bloodstream forms (BSF), the life stage that then proliferates in the mammalian host. The disease is fatal if left untreated and no effective drug is currently available for treatment of the late stage of the disease (i.e., involvement of the central nervous system). The 14-3-3 proteins are highly conserved dimeric acidic proteins acting as phosphoserine/phosphothreonine-dependent chaperones [1,2]. Homologues of 14-3-3 proteins have been found in all eukaryotes [3,4]. Every organism expresses at least one 14-3-3 protein that binds to phosphopeptides containing consensus motifs (mode 1 and/or mode 2) with high affinity (nanomolar levels). The motifs include both RSxpSxP (mode 1) and RxY/ FxpSxP (mode 2) where pS is phosphoserine [5], and the recently identified -pS/pTx 1-2 -COOH (mode 3) where x is not Pro [6]. Only limited number of proteins are known to have the mode 3 motif [7]. 14-3-3 proteins also have the ability to bind other than the modes 1-3 motifs [8,9,10,11]. The latest bioinformatic and experimental survey of 14-3-3-binding sites reveal that alternative mode 1 Rxx(pS/pT)xP motifs dominate, although the last Pro occurs less than half [12]. When 14-3-3 proteins bind to their partners, the interacting partners may change their intracellular localization, preference of interacting partners, or enzymatic functions through conformational changes or masking of the functional amino acid residues [8,9,10,11]. In mammalian cells, the characterization of signal transduction pathways involving kinase/phosphatase has progressed extensively through the discovery of more than 200 14-3-3-interacting proteins, mainly mediated by phosphorylated serine/threonine residue(s) of the target proteins [13].
There is still a gap in our understanding of signal transduction pathways in protozoan parasites including T. brucei. Although we have previously reported that both T. brucei 14-3-3I and II proteins play important roles in cell motility, cytokinesis and the cell cycle [14], phosphoserine-dependent T. brucei 14-3-3-interacting proteins have not been found until now in spite of extensive efforts. Therefore, we examined the differences between human 14-3-3 isoforms and T. brucei 14-3-3 isoforms with respect to affinities to various ligands. Here we provide several lines of evidence that the 14-3-3I, and especially the II, isoforms bind far less efficiently to the conventional consensus motifs (modes 1 and 2). In addition, heterodimerized 14-3-3I and II, the major existing forms in vivo ( [14] and unpublished data), showed detectable affinities to the chimeric proteins containing the mode 3 motif, leading us to identify the T. brucei 14-3-3 binding proteins. The overall data highlight the scarcity of 14-3-3 target proteins with high affinity in the T. brucei cells and may indicate the divergent roles of T. brucei 14-3-3 proteins. The newly identified phosphopeptide that binds to T. brucei 14-3-3 proteins may be utilized in isolating a novel class of T. brucei 14-3-3 binding proteins, since over 200 human 14-3-3binding proteins can be purified from HeLa cell extracts by a competitive elution from 14-3-3 affinity columns with alternative mode 1 phosphopeptide or high-affinity peptide antagonist of 14-3-3 proteins [13,15].

The target proteins for T. brucei 14-3-3 interaction
Mammalian or yeast 14-3-3 proteins have been successfully used as probes in far-Western blot (Far-WB) to identify direct interactions with numerous target partners [13,15,19], although the Far-WB assay has certain limitations related to the conformational state of the protein. Therefore, we used far-Western blot (Far-WB) analysis to search for binding proteins of T. brucei 14-3-3. Lysates were prepared from HeLa and T. brucei PCF cells treated with or without calyculin A (CalA), a serine/threonine phosphatase inhibitor, to increase the number and amount of phosphorylated proteins [13,20]. None of the proteins, except 14-3-3 isoforms including putative dominant negative forms (I K77E and II K56E) that correspond to human 14-3-3f K49E and 14-3-3t K49E, respectively, interacted with T. brucei 14-3-3I and II as shown the bands of approximately 28-30 kDa ( Figure S1) [8]. The 14-3-3t proteins interacted with various proteins including isoforms of human 14-3-3 proteins ( Figure S2A). The number and amounts of the 14-3-3t-interacting protein in both PCF and HeLa cells increased upon CalA treatment ( Figure S2A, GST-t), as reported previously [13]. However, they were much less in PCF cells than those in HeLa cells when human 14-3-3t protein was used as a probe ( Figure S2A, GST-t), suggesting that molecular recognition of 14-3-3 proteins may co-evolve with their ligands. In Arabidopsis, some isoforms of 14-3-3 increase the phosphoserinedependent target binding in the presence of divalent cations [21]. Therefore, we carried out Far-WB analysis in the presence of 1 mM CaCl 2 . However, no major difference in the interacting protein was observed between CalA (2) and (+) in the presence or absence of calcium ( Figure S2B), indicating that the binding of T. brucei 14-3-3 to phosphopeptide-containing proteins is Ca 2+independent. The slight differences between Figure S2A (GST-II) and Figure S2B (GST-II) are due to differences in the concentration of the probes (1 mg/ml in Figure S2A, 2 mg/ml in Figure S2B). The two minor bands (,66 and 70 KD) in the blots incubated with GST alone were the non-specific bands from anti-GST antibodies ( Figure S2). Similar results were obtained when digoxigenin-labeled non-fusion recombinant 14-3-3 proteins were used as probes (data not shown), suggesting that the failure of bindings of 14-3-3I and II is not due to the steric hindrance of GST moiety. These results are consistent with the results of the GST pull-down assay and the surface plasmon resonance analysis. We have also used the heterodimerized form of GST-14-3-3II +I (described in the next section) as a probe in Far-WB analysis and obtained similar results (data not shown). Taken together, our results strongly support the notion that T. brucei 14-3-3 proteins do not interact with the majority of human 14-3-3-binding proteins as other 14-3-3 proteins from other eukaryotic organisms do.

Identification of T. brucei 14-3-3 binding protein with high affinity
Since T. brucei 14-3-3 did not show high affinity to the mode 1 or 2 peptide, we searched the Trypanosoma brucei genome database (GeneDB: http://www.genedb.org/genedb/tryp/index.jsp) for the potential binding partners of T. brucei 14-3-3 proteins containing a motif of -pS/pTx 1-2 -COOH (mode 3: where x is not Pro). We first used Motif Search to extract the sequences containing the mode 3 motif and then selected for the sequences that could be phosphorylated by AGC kinase (PKA, PKG, PKC or related kinases). We selected the putative AGC substrate sequences, since AGC kinase are known to mediate diverse and important cellular functions in mammalian cells. The selected sequences were listed in Table 1.
We first examined whether a putative protein phosphatase 2c (PP2c) in the list interacts with T. brucei 14-3-3 proteins, since the phenotype of knockdown of 14-3-3I and/or II resembled that of okadaic acid, a potent serine/threonine phosphatase inhibitor, -treated T. brucei cells and PP2c has a putative mode 3 motif [22]. The N-terminally V5-tagged PP2c protein was expressed in human HEK293T cells and purified by immunoprecipitation using anti-V5 monoclonal antibody (Ab). Far-WB probing with GST-14-3-3 proteins was used to detect the direct interaction. We have recently identified the heterodimeric form as the major form of T. brucei 14-3-3 proteins (unpublished data) and thus, recombinant heterodimeric GST-II +I was used as a probe. The results showed that V5-tagged PP2c (but not S744A mutant protein which has Ala instead of Ser in the mode 3 motif), directly bound to GST-14-3-3f with high affinity and to T. brucei GST-II +I (heterodimerized recombinant proteins) with lower affinity (Figure 2A upper panel). We further confirmed that the binding was mediated by intact mode 3 motif ( Figure 2B). Of note, PP2c mutant without W at the C-terminal end (-W) showed much higher affinity to dimeric T. brucei 14-3-3 proteins than wild type PP2c ( Figure 2B, lane 2). We next examined whether T. brucei 14-3-3 proteins recognize mode 1, 2 and 3 synthetic peptides spotted on a nitrocellulose filter. Heterodimeric T. brucei 14-3-3 proteins failed to show the interactions with these phosphopeptides (Table 1) including PP2c and PP2c (-W), whereas human 14-3-3f showed strong interactions ( Figure 2C). The binding of T. brucei 14-3-3 proteins with PP2c might require an additional sequence(s) to establish a stable interaction. The other putative mode 3 synthetic phosphopeptides derived from ACS (Table 1) would not interact either T. brucei 14-3-3 or human 14-3-3f ( Figure 2C). We next tested the association of ectopically expressed V5-tagged PP2c with endogenous 14-3-3I and II proteins in a Tet-inducible V5-tagged PP2c-expressing T. brucei PCF clone. Immunoprecipitation with V5 monoclonal Ab followed by Western blotting with a mixture of a-14-3-3I and II Ab failed to show the interaction of 14-3-3 with PP2c ( Figure 2D lower panel), suggesting that the interaction detected by Far-WB is not stable enough to detect by immunoprecipitation. Furthermore, knockdown of PP2c gene did not affect the morphology and the growth of the T. brucei PCF cells (data not shown), suggesting that PP2c may not be a physiological target for T. brucei 14-3-3 proteins in vivo.
To identify 14-3-3 interacting proteins containing a mode 3 motif with higher affinity than PP2c, we constructed chimeric molecules by replacing the C-terminal end of PP2c with various sequences of mode 3 motif found in the database (Table 1, except for ACS sequence). Those chimeric proteins were expressed in HEK293T cells and purified by immunoprecipitation, and the affinity to GST-II +I was compared with wild type PP2c by Far-WB analyses. Chimeras containing -RRRNSV (Tb09.211.0210) and -KRRRSV (Tb10.70.2780: predicted SAP domain protein termed p31-SAP in this manuscript) associated with GST-II +I more tightly than wild type PP2c ( Figure 3A, upper panel). In order to examine whether p31-SAP binds to T. brucei 14-3-3 proteins through the mode 3 motif in vivo, Tet-inducible V5-tagged p31-SAP or p31-SAP S286A (mode 3 motif mutant)-expressing T. brucei PCF clones were established. Tet-induced or uninduced cell lysates were subjected to immunoprecipitation with a-V5 monoclonal Ab followed by Western blotting with a mixture of a-I and a-II-specific Ab ( Figure 3B, upper panel). Western blotting with a-V5 monoclonal Ab ( Figure 3B, lower panel) serves as the immunoprecipitation control. Importantly, p31-SAP but not p31-SAP S286A, a mode 3 mutant, binds to 14-3-3I and II in vivo (PCF cells) suggesting that the interaction is mediated by the mode 3 motif ( Figure 3B). We then examined whether the p31-SAPderived mode 3 phosphopeptide interact with T. brucei 14-3-3 proteins. The synthetic phosphopeptide MGGGHVSGLKRR-RpSV derived from p31-SAP was clearly detected by heterodimeric T. brucei 14-3-3I+II proteins whereas MGGGLTRSRpSL dereived from PP2c-W was not detected as demonstrated previously ( Figure 2C, Figure 4). The human 14-3-3f interacts with both phosphopeptides but with the preferential binding to the mode 3 peptide derived from p31-SAP (Figure 4). The human 14-3-3f shows slightly higher affinity to the p31-SAP derived phosphopeptide than T. brucei 14-3-3 ( Figure 4). Overall data suggest that human 14-3-3 proteins have higher affinity to all the binding motifs used in our experiments than T. brucei 14-3-3 proteins. In addition, mutations in the putative critical amino acid residues in both 14-3-3I (K77E) and II (K56E) ( Figure S1) prevent the binding to the p31-SAP derived peptide (Figure 4, GST-I K77E +II K56E), suggesting that the structure of an amphipathic groove that mediate the association of 14-3-3 proteins with phosphopeptides are evolutionally conserved between far distant organisms. Subtle difference(s) in the structure of the amphipathic groove or the distinctive difference(s) in the structure of the N-and/or C-termini, may affect the binding to the motifs. Since no phosphopeptide was known to interact with T. brucei 14-3-3 proteins until now, the affinity purification method eluting with a specific phosphopeptide or a high-affinity peptide antagonist of 14-3-3 proteins, which is successfully employed to isolate a great number of 14-3-3-interacting proteins in other organisms [13,15], has not been possible. The newly identified high-affinity phosphopeptide (HVSGLKRRRpSV) is the first available phosphopeptide that can be utilized for the affinity purification of 14-3-3-binding proteins in T. brucei and for the subsequent identification of novel binding motifs (unpublished data).
T. brucei 14-3-3 proteins exhibit far lower affinity to the evolutionally conserved consensus binding motifs (modes1 and 2) and slightly lower affinity to the newly identified mode 3 sequence when compared to those in human 14-3-3 proteins, representing the atypical nature of T. brucei 14-3-3 proteins. Considering the affinity to the phosphopeptides, T. brucei 14-3-3 proteins might not only act as phosphoserine-dependent binding proteins but also act as binding proteins utilizing hitherto unknown consensus motifs. Thus, the functions of 14-3-3 proteins in protozoan organisms such as trypanosomatids, the most divergent eukaryote from mammals, may need to be reconsidered. Further investigation of the atypical nature of T. brucei 14-3-3 proteins and the unidentified binding proteins may help identify novel drug targets since 14-3-3 proteins are essential to the survival of the parasite [14].

Construction of pLew82T7bsr N-V5 PP2c, pCR3 N-V5 PP2c
The PP2c-coding sequence (Tb927.7.4020: protein phosphatase 2C, putative) was amplified by PCR using primers 59-atgtataccagtgttagaaagcct-39 and 59-gggccgcaacctgtctcctcataacat-39. The  Table 1 have far less affinity to T. brucei 14-3-3 than human 14-3-3f proteins. Indicated biotinylated phosphopeptides were mixed with streptavidine and spotted to nitrocellulose filters and dried. The filters were then incubated with indicated GST-14-3-3 proteins and detected with a-GST antibodies. (D) Tet-inducible T. brucei expression vector, pLew82 T7bsr-N-V5-PP2c was transfected and clones were isolated with blasticidin selection. Tet-uninduced (Tet2) or -induced (Tet +) cell lysates of the clone were subjected to immunoprecipitation followed by Western blotting using a mixture of a-I and II Ab (lower panel), and horseradish peroxidase (HRP)-labeled a-V5 Ab (upper panel). doi:10.1371/journal.pone.0015566.g002 amplified PP2c sequence was inserted into pLew82T7bsr N-V5 (pLew82 is a kind gift from Dr. George Cross) and PCR3 N-V5 (PCR3 is purchased from Invitrogen) at the HpaI site. (The pLew82 vector was modified by the insertion of annealed oligonucleotides of weak T7 promoter at the BamHI site. The sequences of oligonucleotides are as follows: 59-GATCCTTA-ATACGTCTCACTATAGGGA-39, and 59-GATCTCCCTA-TAGTGAGACGTATTAAG-39. Permanent transfectants with pLew82 T7 vector do not require 1 ng/ml Tet for maintenance of the clones, while pLew82 transfectants do. The N-terminal V-5 tag sequence and HpaI site were introduced in the pLew82T7 vector to make pLew82T7 N-V5. Drug selection maker was replaced with a blasticidin-resistance gene, which was designated as pLew82T7bsrN-V5.)

Construction of pLew82T7bsrN-V5 p31-SAP
The P31-SAP-coding (Tb10.70.2780: predicted SAP domain protein) sequence was amplified by PCR using primers 59atgaggaaacccgggcggaaaatt-39 and 59-acgggtgctgataatgtaaccaa-39. The amplified sequence was inserted in the vector. The mutants were created using PCR and the sequences were confirmed.

Construction of C-terminal PP2C mutants
The C-terminal HpaI site in the PP2c gene and the XbaI site in the vector were used to insert the annealed oligonucleotides encoding the indicated amino acid sequences.

Raf259 peptide pull-down assay
Ten ml of streptavidin agarose beads (Sigma Chemical Co., St. Louis, MO) were incubated with 10 ml of 100 mM of pRaf 259 (biotin-MAGGGRQRSTSTPN) and/or pSRaf 259 at room temperature for 30 min and the beads were washed three times with lysis buffer (150 mM NaCl, 10 mM HEPES, pH 7.5, 0.2%  NP-40, 50 mM NaF, 1 mM Na 3 VO 4 and Roche protease inhibitor tablets). Next, 1610 9 cells of T. brucei PCF were lysed with 2 ml of the lysis buffer on ice. Insoluble proteins were pelleted and 1 ml of the supernatant was used for each peptide pull-down assay. The lysates were incubated with each bead, rotated at 4uC for 2 hrs and washed with the lysis buffer three times. Twenty ml of 2xSDS Gel loading buffer with 2-mercaptoethanol were added to the washed beads, and 5 ml of eluates were run on 10-20% SDS-PAGE. The blots were detected with a-I and/or a-II Abs as described previously [14].

Transfection and establishment of clones
Transfection and cloning of T. brucei 29-13 procyclic cells (a kind gift from George Cross) were performed as previously described [23]. Finally, 1 mg/ml of Tet was added upon induction of the genes. Transfection of SV40 large T-antigen-transformed human embryonic kidney (HEK 293T) cells (Gene Hunter Corporation) was performed using Fugene 6 transfection reagent (Roche).

Interaction of GST-14-3-3I, -II and t with human c-Raf
HeLa cell lysates were prepared using NP-40 lysis buffer (150 mM NaCl, 10 mM HEPES, pH 7.5, 0.2% NP-40, 50 mM NaF, 1 mM Na 3 VO 4 and Roche protease inhibitor tablets) and subjected to GST pull-down assay. Bound proteins were separated with SDS-PAGE, transferred to a PVDF membrane and detected with anti-human c-Raf-1 antibodies (Santa Cruz Biotechnology, Santa Cruz, CA).

Far-Western blot analysis
Far-Western blot analysis of Ser/Thr phosphatase inhibitor Calyculin A (CalA) treated + and/or untreated -of HeLa and/or 29-13 PCF cell lysates was performed. HeLa cells and/or 29-13 PCF cells were lysed with SDS-gel loading buffer supplemented with 2-mercaptoethanol (Bio-Rad, Hercules, CA) and sonicated. The cell lysates equivalent to 5610 5 HeLa cells and 1610 7 29-13 PCF cells were applied on 4-20% SDS-PAGE (Daiichi, Japan) and transferred to nitrocellulose filters (Millipore, Bedford, MA). The filters were denatured with denaturation buffer (6M guanidine-HCl, 50 mM Tris-HCl, pH 8.0, and 1 mM dithiothreitol [DTT]). The denaturation buffer was then diluted with an equal volume of 50 mM Tris-HCl, pH 8.0 with 1 mM DTT. After 15 min of denaturation, the filters were then treated with 2x diluted denaturation buffer. The same step was repeatedly carried out eight times and then renatured filters were washed twice with TTBS (50 mM Tris-HCl, pH 7.4, 0.1% Tween 20, and 150 mM NaCl). The resultant filters were blocked with TTBS containing 4% skim milk plus 1 mM DTT. Then the filters were incubated overnight with 1 or 2 mg/ml of GST-14-3-3 probes in 4% skim milk containing the blocking solution at 4uC. The filters were washed three times with TTBS and incubated with anti-GST polyclonal antibodies (Sigma) followed by horseradish peroxidase (HRP)-labeled anti-rabbit goat IgG (Jackson Immunoresearch Laboratories, West Grove, PA).

Immunoprecipitation
Immunoprecipitation was carried out using NP-40 lysis buffer. Anti-V5 monoclonal Ab and sepharose suspension protein G (protein G beads) were purchased from Nacalai and Invitrogen, respectively. In brief, cells were lysed on ice for 30 min and spun at 15,000 rpm for 10 min. The resultant supernatants were used for immunoprecipitation using 5 ml of protein G beads and 1 mg of a-V5 monoclonal Ab.

NC-filter binding assay
Five ml of 1 mM biotinylated peptides, purchased from Scrum (Japan) were mixed with 15 ml of streptavidine (1 mg/ml) in the presence of 0.1% Tween20, spotted on NC-filters and dried. Filters were washed with TTBS, blocked with Protein-Free Blocking Buffer (Pierce), and then incubated with GST-14-3-3 proteins in TTBS containing 20% Protein-Free Blocking Buffer and 1 mM DTT. The filters were extensively washed four times with TTBS, and detected with anti-GST antibodies. Figure S1 Amino acid sequence alignment of T. brucei 14-3-3I, II, human 14-3-3t, and f. The amphipathic groove structures are comprised of a-helices 3, 5, 7 and 9 as shown in green lines. Amino acid residues directly engaged in the conserved phosphopeptide bindings are boxed in red. Identical amino acid residues are colored in magenta. (TIF)