Interkingdom Complementation Reveals Structural Conservation and Functional Divergence of 14-3-3 Proteins

The 14-3-3s are small acidic cytosolic proteins that interact with multiple clients and participate in essential cellular functions in all eukaryotes. Available structural and functional information about 14-3-3s is largely derived from higher eukaryotes, which contain multiple members of this protein family suggesting functional specialization. The exceptional sequence conservation among 14-3-3 family members from diverse species suggests a common ancestor for 14-3-3s, proposed to have been similar to modern 14-3-3ε isoforms. Structural features of the sole family member from the protozoan Giardia duodenalis (g14-3-3), are consistent with this hypothesis, but whether g14-3-3 is functionally homologous to the epsilon isoforms is unknown. We use inter-kingdom reciprocal functional complementation and biochemical methods to determine whether g14-3-3 is structurally and functionally homologous with members of the two 14-3-3 conservation groups of the metazoan Drosophila melanogaster. Our results indicate that although g14-3-3 is structurally homologous to D14-3-3ε, functionally it diverges presenting characteristics of other 14-3-3s. Given the basal position of Giardia in eukaryotic evolution, this finding is consistent with the hypothesis that 14-3-3ε isoforms are ancestral to other family members.


Introduction
The 14-3-3s comprise a family of small acidic polypeptides, which form dimers and are ubiquitous in all eukaryotic kingdoms [1,2,3]. The key feature of 14-3-3s is their ability to bind hundreds of client proteins, generally phosphorylated on a serine or threonine within short consensus binding motifs [4]. The broad 14-3-3 interaction network [5,6,7,8,9] and the essential role of these proteins may be reflected in the multiple distinct isoforms present in higher eukaryotes. The structure of both animal and plant 14-3-3s is conserved, with the monomer comprised of a bundle of nine alpha-helices (A to I) organized in a cup-like shape. All 14-3-3s contain three distinct blocks of highly invariant aminoacids [2], proposed essential for their distinct topology and functionality [10]. The most invariant block in the central region of all isoforms suggests evolution from a common ancestor before animal and plant separation [2]. Dimerization, is the elemental functional property of 14-3-3s and is mediated by interaction of largely conserved amino-acids in the amino-terminal helices A and B of one monomer with C and D of the other [11]. The identity of conserved amino-acids within these helices determines dimer stability and limits homo and hetero-dimerization options among isoforms. Dimer composition is functionally relevant because it determines the range of client proteins it interacts with.
Another remarkable characteristic of 14-3-3s is the high aminoacid sequence similarity among orthologs from different species, rather than with other isoforms of the same organism [2,3]. This is particularly prominent among animal epsilon isoforms [3], suggesting that 14-3-3e may be close to the common ancestral isoform. However, a central question in studying the 14-3-3 family is whether sequence conservation among orthologs also reflects functional conservation and differentiates them from other family members. We addressed this question based on the similarity of the single 14-3-3 protein of Giardia duonenalis (syn. G. lamblia, G. intestinalis), a worldwide parasite of the upper part of the small intestine of mammals including human [12], with the fruit fly Drosophila melanogaster 14-3-3s, asking for interspecies functional complementarity.
The protist genus Giardia belongs to the diplomonads, a group of anaerobic/microaerophilic binucleated flagellates. The highly divergent Giardia spp. are considered candidate early-branching eukaryotes and provide unique opportunities for gaining insights into key events of eukaryotic evolution [13,14,15]. In this emerging eukaryotic model, a single 14-3-3 (g14-3-3) engages an interaction network comparable to that of the yeast isoforms [5] and is critically important for parasite differentiation [16]. In contrast, the metazoan D. melanogaster has only two 14-3-3-encoding genes [17], but alternative splicing of the 14-3-3fencoding gene leonardo results in three spatially restricted and apparently functionally distinct isoforms [17,18].
In addition to alternative splicing, posttranslational modifications are utilized to increase the number of distinct isoforms temporally or cell-type-specifically, to cope with functional diversity when limited by the number of distinct 14-3-3 encoding genes in a genome [19]. Unlike the Drosophila isoforms [20], g14-3-3 is constitutively phosphorylated and stage-dependent polyglycylated, modifications essential for parasite differentiation from trophozoites into infectious cysts [21,22]. Whether modified and unmodified g14-3-3 have distinct functions and whether they can functionally complement one or more isoforms of a higher eukaryote is currently unknown.
Because of g14-3-3s similarity to Drosophila and human 14-3-3e isoforms, we investigated whether it is functionally orthologous to D14-3-3e and conversely whether any of the fly 14-3-3 proteins function in G. duodenalis (henceforth referred to as Giardia). The two systems are ideally suited for this approach as they are highly distant evolutionarily, have well developed genetic arsenals and flies have only two 14-3-3 genes to the single protozoan counterpart. Of the three fly 14-3-3f (LEONARDO) isoforms, we chose LEOII because similar to D14-3-3e it is broadly expressed spatially and temporally [18] and parallels in principle the broad distribution of g14-3-3 in Giardia [16,22].

Cultures and Transfection
Trophozoites of G. duodenalis strain WB-C6 were axenically grown and encystation induced as previously described [22]. Transgenic Giardia lines were generated by electroporation, selected and maintained under 100 mM puromycin (Invivogen, Toulouse, France).

Preparation of Soluble Protein Fractions
Soluble Giardia proteins were prepared from 2610 9 trophozoites or encysting parasites as previously described [22].
For affinity purifications and mass spectrometry analysis, soluble proteins were obtained from 30-40 flies chilled at 4uC for 1 h, then frozen at -20uC followed by homogenization in 1 ml EB (EB, 30 mM Tris-HCl, 1 mM DTT, and 1 mM EDTA, pH 7.4), supplemented with protease and phosphatase inhibitor cocktails (Sigma), then sonicated and the lysate treated as described above.

Expression and Purification of Recombinant Proteins
GST and GST-Difopein were purified from E. coli by affinity chromatography on glutathione-sepharose 4B (GE Healthcare, Little Chalfont, England) and eluted with 10 mM glutathione (pH 8.0) as described by the manufacturer.
To purify 14-3-3 proteins from Giardia or Drosophila on GSTdifopein, 3 mg of soluble proteins were incubated with 15 mg of glutathione-sepharose immobilized GST or GST-difopein in HT buffer at 4uC for 2 h. After extensive washes with HT buffer, the 14-3-3 proteins were eluted with 100 ml of 2 mM A8Ap synthetic phosphopeptide (ARAApSAPA, where pS is a phosphoserine) reproducing a 14-3-3 binding motif [26], in HT buffer. An aliquot of eluted material was run in SDS-PAGE and subjected to immunoblot as described above.

Affinity Purifications, Pull-down and Immunoprecipitations
FLAG-tagged proteins were purified using anti-FLAG M2 mAb bound to agarose beads (Sigma-Aldrich) per manufacturer's instructions. An equal amount of Giardia soluble proteins from control WB-C6 strain and transgenic lines were incubated with anti-FLAG beads at 4uC for 3 h and washed with 100 bed volumes of K-HT buffer. To selectively remove proteins bound to 14-3-3, the resin was incubated with 2 mM A8Ap synthetic phosphopeptide, at 4uC for 1 h. Finally, FLAG-fusion proteins were eluted from the resin by incubation with 200 mM synthetic FLAG-peptide at 4uC for 1 h and stored at 270uC until use.
For purification of His-tagged protein, 1 mg of Drosophila soluble proteins was incubated with 120 ml pre-cleared Ni beads (Talon) according to manufacturer, for 4 h at 4uC, followed by multiple washes in lysis buffer with 10 mM imidazole. Bead-bound proteins were recovered by addition of one volume of Laemmli buffer and heating at 95uC for 5 min.
For cross-linking experiments, 6xHis-g14-3-3 was purified from 1 mg of Drosophila protein lysate on Ni-beads as described above. Cross-linking was attained by incubating lysates in 3 mM BS 3 (Pierce) as described previously [20] and then processing the samples for Western blots, or Ni affinity purification.
For cell counting, parasites cultured in encysting medium for 12 h were stained with Cy3-conjugated anti-FLAG mAb, FITCconjugated anti-CWP mAb and DAPI. For each transgenic parasite line, the total amount of cells spotted on the glass-slide and positive for the anti-FLAG staining were counted. The percentage of trophozoites, encysting parasites and cysts (distinguished from trophozoites by co-staining with anti-CWP mAb) was quantified and the mean and standard error of three independent experiments calculated.

Mass Spectrometry
Aliquots of affinity purified 14-3-3s (with either anti-FLAG or GST-difopein) were separated on a 1D-gel NuPAGE 4-12% (Novex, Invitrogen) run in morpholinepropanesulfoninic acid (MOPS) buffer and stained with the Colloidal Blue Staining kit (Invitrogen). Slices were excised and digested with modified sequencing-grade trypsin (Promega) according to published protocols [27]. Mass spectrometry analyses of peptide mixtures were performed with a Voyager DE-STR (Applied Biosystems) in positive reflector mode. To enhance the efficiency of phosphopeptide ionization in MALDI, 30 mg/ml of 2,5-dihydroxybenzoic acid (DHB, Sigma Aldrich) at the concentration of 30 mg/ml in 50% CH 3 CN, 1% orthophosphoric acid was used as matrix. Peptides were measured in the mass range of 750 to 4000 Da and all spectra were internally calibrated and processed via the Data Explorer software using ions coming from trypsin autodigestion of FLAG-g14-3-3.

Sequence Analysis
Multiple alignments were performed using the Clustal W program at http://www.ebi.ac.uk/clustalw/.
To determine whether g14-3-3 is a functional ortholog of D14-3-3e we took a complementary approach expressing the Giardia protein in Drosophila and the fly proteins in the protozoan. Drosophila 14-3-3s were introduced in Giardia by cloning D14-3-3e and leoII cDNAs [18] under the constitutive Giardia glutamate dehydrogenase promoter ensuring stable mid range expression in all stages, except for a small decrease during encystation [28]. A transgenic line over-expressing a FLAG-g14-3-3 was used as control. The recombinant proteins expressed in trophozoites and during encystation appeared as single species in all transfectants and absent from the parental WB-C6 strain ( Figure 2A). The apparent molecular weight of g14-3-3 was larger than predicted from the primary sequence due to the previously described posttranslational modifications [16]. In encysting parasites, FLAG-g14-3-3 displayed the expected decrease in apparent molecular weight consequent to the previously described shortening of the polyglycine tail occurring normally during cyst formation [22]. Intriguingly, a smaller molecular weight shift was also evident for FLAG-D14-3-3e, suggesting modification by polyglycylation or phosphorylation. In contrast, FLAG-LeoII appeared to migrate in its predicted molecular weight irrespective of the stage of the parasites expressing it, suggesting lack of such post-translational modifications ( Figure 2A).
Ubiquitous expression of transgenic 6XHis-g14-3-3 in Drosophila was achieved under the b-Tubulin-Gal4 (TubG4) enhancer [29] ( Figure 2B). The Giardia protein was expressed at relatively high levels and exhibited the expected molecular weight. Importantly, the polyclonal antibodies against D14-3-3e and Leo did not crossreact with the Giardia protein, which is consistent with the differences in primary sequence and facilitates differentiation between the three proteins when co-expressed. The 3 Leo species resolved into two bands in this particular gel system as reported before [20].
Because g14-3-3 polyglycylation has been linked to its nuclear localization [22], we investigated the distribution of the transgenic Drosophila proteins in Giardia trophozoites and parasites undergoing encystation. Encystation was monitored with an antibody against Giardia cyst wall protein (CWP) [16,21], which decorates the encystation-specific vesicles (ESV) and in the early stage of encystation, the endoplasmic reticulum also. As previously shown [16], both anti-FLAG and anti-g14-3-3 antibodies exclusively decorated the cytoplasm in control trophozoites expressing FLAG-g14-3-3. The transgenic protein was absent from the ESVs of the encysting parasites, demonstrated by lack of co-localization with the CWP protein ( Figure 2C). Similarly, FLAG-D14-3-3e was present in the cytoplasm of trophozoites, but was not in the nuclei of encysting parasites. By contrast, FLAG-LeoII localized both in the nuclei and cytoplasm of trophozoites and encysting parasites. These immunolocalization results support the notion borne from the western analysis that FLAG-D14-3-3e could be polyglycylated and excluded from nuclei, whereas FLAG-LeoII is not.
The subcellular localization of 6xHis-g14-3-3 under the neuronal-specific ElavG4 driver was examined in fly embryonic and adult neurons selected for their accessibility. In comparison to the nuclear Elav protein [30], 6XHis-g14-3-3 was clearly localized in the perinuclear space of the cell bodies and in axons (arrow) of adult mushroom body (MB) neurons ( Figure 1D). The transgenic protein was similarly perinuclear in adult optic lobe and embryonic ventral nerve cord neurons. Similar localization was observed for non-neuronal tissues expressing 6XHis-g14-3-3 under TubG4 (not shown). Therefore, 6xHis-g14-3-3 is cytoplasmic in Drosophila cells as in Giardia trophozoites.

Selective Dimerization of g14-3-3 with Leo
The default state of 14-3-3s is dimeric [10,11,31] and in transfected Giardia and transgenic Drosophila, 14-3-3s from different species co-exist. Because g14-3-3 is the sole isoform in Giardia, it normally forms homodimers. Heterodimer formation with either fly 14-3-3 would provide significant insight into their structural properties. Therefore, FLAG-tagged proteins were purified from Giardia trophozoites by affinity chromatography as detailed in experimental procedures. Tagged 14-3-3s were present only in the transfectant FLAG-eluted fractions, but not in controls ( Figure 3A). As demonstrated already [16], endogenous g14-3-3 co-purified with FLAG-g14-3-3 ( Figure 3B), but intriguingly it also co-purified with FLAG-LeoII to a comparable extent, demonstrating that the two isoforms heterodimerize. Importantly, no signal could be detected with anti-g14-3-3 in the immunopurified FLAG-D14-3-3e fraction ( Figure 3B). Although a band with molecular weight compatible with g14-3-3 was visible in the coomassie stained gel ( Figure 3A), this was identified by MS analysis as D14-3-3e (data not shown) and likely correspond to an N-terminal truncated FLAG-D14-3-3e, since it was not detectable by the anti-FLAG antibody.
Similarly, nickel affinity chromatography of 6XHis-g14-3-3 from transgenic Drosophila lysates demonstrated exclusive copurification with the endogenous Leo, but not with D14-3-3e ( Figure 3C), irrespective of whether the transgenic Giardia protein was expressed, only in neurons (ElavG4) or throughout the adult fly (TubG4). Interestingly, His-g14-3-3 appeared to interact with one of the Leo species resolved (Fig. 3C, lane 6), but currently its identity is under investigation. To further demonstrate the differential dimerization of 6xHis-g14-3-3 with the endogenous Leo, we chemically cross-linked the proteins with BS3, resolved the complexes by SDS-PAGE and used antibodies to probe the constitution of the dimers. The results revealed strong dimerization of Leo isoforms [20], homodimerizarion of 6xHis-g14-3-3, but also Leo/6xHis-g14-3-3 heterodimers ( Figure 3D). To ascertain Leo/6xHis-g14-3-3 heterodimerization in Drosophila, we collected the complexes containing cross-linked 6xHis-g14-3-3 with nickel beads and asked whether such dimers contain Leo. Indeed, as shown in Figure 3D, His-containing dimers also co-purified with endogenous Leo, demonstrating heterodimerization of the Giardia and Drosophila proteins. These dimers did not contain detectable D14-3-3e (data not shown). Collectively, these results suggest that g14-3-3 heterodimerizes preferentially with Leo either in Giardia or Drosophila, a property consistent with a 14-3-3e-like protein.

D14-3-3e is Post-translationally Modified in Giardia
In Giardia g14-3-3 is constitutively phosphorylated at Thr 214 and polyglycylated at Glu 246 [21,22]. Residues surrounding Thr 214 of g14-3-3 are well conserved in Drosophila 14-3-3s ( Figure 4A) as in place of Thr 214 of g14-3-3 there is the phosphorylatable Ser 210 in D14-3-3e. In contrast, it is replaced with the polar uncharged Asn 210 in LeoII. Moreover, a hypothetical polyglycylation signature sequence, where X is a polar or a negatively charged amino acid and gE is a polyglycylated glutamic acid, has been previously suggested for the g14-3-3 [22] at the C-terminus. Although as commonly observed for 14-3-3s [2], the C-termini of the three proteins diverge ( Figure 4A), we searched for a similar sequence in the fly isoforms. As shown ( Figure 4B), multiple alignment of the Cterminal amino acids of the D14-3-3e and LeoII respectively, with those of g14-3-3 surrounding the polyglycylated Glu 246 identified a polyglycylation signature sequence at the D14-3-3e C-terminus which was clearly absent from LeoII. As reference and controls, we included in this comparison the verified polyglycylation sites of aand b-tubulin from G. duodenalis, Tetrahymena thermophila and Paramecium tetraurelia [21,22].
To ascertain that D14-3-3e was modified in Giardia as convergently suggested by sequence comparison, immunofluorescence and immunoblot analysis, affinity purified FLAG-tagged proteins from trophozoites were analyzed by MALDI-MS. The mass spectrum of FLAG-g14-3-3 revealed the presence of peaks    MH + = 2104.92) [21], which is consistent with the addition of up to 24 glycines ( Figure 4C, right panel). Conversely, the spectrum of FLAG-D14-3-3e ( Figure 4D, left panel) revealed two peaks, corresponding to peptide 197 AAFD-DAIAELDTLS 210 EESYK 216 in the un-phosphorylated (MH + = 2087.9) and phosphorylated version (MH + = 2167.9), suggesting that a fraction of FLAG-D14-3-3e is phosphorylated likely on Ser 210 corresponding to Thr 214 in g14-3-3 ( Figure 4A). However, we cannot exclude the possibility that the observed shift results from phosphorylation of either Thr 208 or Ser 213 . Further scrutiny of the FLAG-D14-3-3e spectra revealed the peak of unmodified C-terminal peptide 249 EQIQDVEDQDVS 260 (MH + = 1404.6), but in addition a pattern of peaks signifying addition of 7-8 glycines ( Figure 4D, right panel). Sequence homology suggests polyglycylation of Glu 249 ( Figure 4A). In addition, a peak corresponding to the phosphorylated peptide 249-260 was detected (MH + = 1484.6), with the only possible phosphorylation being obligatorily on Ser 260 ( Figure 4D, right panel). It is evident then, that in Giardia, D14-3-3e is phosphorylated and a limited number of glycines are added at the carboxyterminus, strongly suggesting that this Drosophila protein is at least in part, processed by the parasites similarly to the endogenous g14-3-3.
In contrast, tryptic digestion of FLAG-LeoII generated only the peaks corresponding to unmodified peptide 197 QAFD-DAIAELDTLN 210 EDSYK 215 with MH + = 2157.98 and 226 DNLTLWTSDTQGDEAEPQEGGDN 248 (2492.03 MH + ) ( Figure 4E, right panel), indicating that as expected, substitution of Ser 210 with Asn 210 abolished phosphorylation. Moreover, this result indicates that FLAG-LeoII is not modified detectably on Thr 208 or Ser 213 or any other residues in the vicinity, a conclusion also extended to FLAG-D14-3-3e. Importantly, this result provides functional support for the essential role of the identified polyglycylation consensus whose absence in LeoII predicts the observed lack of modification. Therefore, D14-3-3e is posttranslationally modified in Giardia similarly with the endogenous g14-3-3.
The peptide mass fingerprint of proteins purified from wild type flies identified D14-3-3e and all Leo isoforms (data not shown). The same result was obtained with proteins purified from transgenic flies, where in addition to the endogenous 14-3-3s, 6xHis-g14-3-3 was also successfully identified (data not shown). The mass spectrum of 6xHis-g14-3-3 only revealed peaks corresponding to the unmodified peptides 202 AFDAAITDLDKL-TEESYK 219 at MH + = 2029.9 ( Figure 5B Figure 5D). Further inspection of all spectra could not detect any other modified peptides. These results demonstrate that endogenous Drosophila 14-3-3s are not subjected to any major posttranslational modification, at least under our experimental conditions and in the context of whole fly lysates. This is also congruent with the lack of phosphorylation or polyglycylation on the relevant peptides of the transgenic His-g14-3-3 expressed in Drosophila, suggesting that flies may lack the requisite enzymes.

Negative Functional Consequences of Exogenous 14-3-3 Expression in Giardia and Drosophila
Over-expression of wild type D14-3-3e, or Leo isoforms in Drosophila does not appear detrimental [17]. However, overexpression of g14-3-3 phosphorylation and polyglycylation mutants or deglycylation enzymes in Giardia has been reported to affect encystation [16]. Because LeoII is not post-translationally modified and D14-3-3e presents limited polyglycylation in the parasite, we tested whether their expression may affect cyst development negatively. For each transgenic line the percentage of trophozoites, encysting parasites and cysts present in cultures after 12 hrs of growth in the encysting medium was estimated. Encysting parasites and cysts were distinguished based on expression of the CWP2 protein.
Compared to Giardia expressing FLAG-g14-3-3 ( Figure 6A), responsiveness to encystation stimuli was significantly reduced resulting in a 40% and 25% cyst reduction in transformants expressing FLAG-D14-3-3e and FLAG-LeoII respectively. A concomitant increase in the number of encysting parasites accompanied cyst reduction and was nearly double of that observed in the FLAG-g14-3-3-expressing line. Therefore, both Experimentally defined polyglycylated glutamic acid residues are black boxed and highlighted in bold white letters. Putative polyglycylated glutamic acid residues are only black boxed. The underlined glutamic acid of Tetrahymena a-tubulin is predicted to be polyglycylated. C-E) MALDI-MS analysis of affinity purified FLAG-tagged transfected proteins from Giardia trophozoites. MALDI-MS spectra of FLAG-g14-3-3 (C), FLAG-D14-3-3e (D) and FLAG-LEOII (E), encompassing the MH + range of 1400-3600. Mono-isotopic masses of relevant peaks are shown. Peptides are indicated by the positions of their N-and C-termini and numbered as in the protein sequence. Peaks shifted from the theoretical MH + are indicated by arrows as for the 80 Da shift due to phosphorylation. For each protein, analysis of phosphorylation is reported on the left and the analysis of C-terminal polyglycylation is reported on the right panels. C) For FLAG-g14-3-3, the calculated MH + peak for peptide 202-219 at 2029.9 is clearly shifted to 2109.93 indicating phosphorylation on Thr 214 ( 202 AFDAAITDLDKLTEESYK 219 ). On the right, the peaks corresponding to polyglycylation of the peptide ( 230 DNLNLWVTDSAGDDNAEEK 248 ) with predicted MH + = 2047.92 was clearly shifted to 2105.9 indicating multiple glycines on Glu246 as revealed by their number in the lateral chain. The insert shows the lack of the peak corresponding to the unmodified 230-248 peptide. D) For D14-3-3e-FLAG, the predicted peak at MH + = 2087.9 for the peptide ( 197 AAFDDAIAELDTLSEESYK 21 ) was shifted by 80 kDa to MH + = 2167.9, indicating phosphorylation at Ser 210 . On the right, the peaks corresponding to the unmodified ( 249 EQIQDVEDQDVS 260 = 1404.6 MH + ) peptide and the shifts to higher MH + consistent with addition of 7 or 8 glycines (right) and phosphorylation at Ser 260 (+80kDa, arrow on the left) are shown. C) For LEOII-FLAG only the peaks corresponding to unmodified ( 197 QAFDDAIAELDTLNEDSYK 215 = 2157.98 MH + ) and ( 226 DNLTLWTSDTQGDEAEPQEGGDN 248 = 2492.03 MH + ) peptides were evident. doi:10.1371/journal.pone.0078090.g004

14-3-3 Functional Conservation and Divergence
PLOS ONE | www.plosone.org Figure 5. MALDI-MS analysis of affinity purified endogenous and transgenic g14-3-3 proteins from Drosophila. A) Western blot analyses of GST-difopein purified 14-3-3s (1:10) from wild type flies and His-g14-3-3 -expressing transgenic flies and the relevant controls (WB-C6 for Giardia and w 1118 for flies) were separated on 12% SDS-PAGE, blotted and the membranes probed with the indicated antibodies. The AXO49 mAb detecting the Giardia polyglycylated protein did not cross react with the fly isoforms, which are detected with the anti-pan14-3-3 antibody, while the transgenic g14-3-3 protein is detectable with the anti-His in fly lysates. TR denotes trophozoite lysates, while ENC indicates lysates from 12hr encysting parasites. MALDI-MS spectra of the transgenic His-g14-3-3 (B) and the endogenous Drosophila D14-3-3e (C) and Leo (D) encompassing the MH + range of 1400-3600. Mono-isotopic masses of relevant peaks are shown. Peptides are indicated by the positions of their NH 2 -and C-termini and Drosophila 14-3-3s interfere with the cyst differentiation process, although probably for different reasons. Given g14-3-3 heterodimerization with LeoII and the lack of phosphorylation and polyglycylation on the latter, it is not surprising that such heterodimers may be functionally impaired for cyst formation. This is also consistent with the dominant negative effects of the g14-3-3 phosphorylation mutant T214A when co-expressed with the endogenous protein [16]. Surprisingly however, the effect seems greater upon overexpression of D14-3-3e, which is at least in part phosphorylated and polyglycylated as is the endogenous g14-3-3. Provided that D14-3-3e does not heterodimerize with the endogenous protein, it is likely that it competes with the endogenous g14-3-3 for relevant clients and/or modifying enzymes (such as deglycylases) impairing the encystation process.
Flies expressing 6xHis-g14-3-3 pan-neuronally under ElavG4, or ubiquitously with TubG4 seemed to develop normally and yield adults of the expected numbers without apparent morphological aberrations. Furthermore, they exhibited grossly normal behaviours and were not deficient in olfactory associative learning (data not shown). This lack of deficits prompted use of 6xHis-g14-3-3expressing flies to complement loss of function mutations in Leo or D14-3-3e. Loss of Leo is lethal in Drosophila and homozygotes for the null transposon mutation leo P1188 are not recovered [25,32]. Although similarly with Leo 6xHis-g14-3-3 is not post-translationally modified, its pan-neuronal or ubiquitous expression did not reverse the lethality of these and other null leo homozygotes (not shown). Thus, as indicated by their sequence divergence, g14-3-3 is not functionally homologous to Leo.
Because 6xHis-g14-3-3 expression does not affect wild type fly development, it cannot be argued that the enhanced lethality results from interference with D14-3-3e or Leo-interacting proteins, essential for embryonic development. An alternative explanation is that g14-3-3 interferes with, or negates the compensation by Leo over-expression in the null embryos [17]. To address this hypothesis, we quantified the levels of D14-3-3e and Leo in late D14-3-3e ex4 homozygous embryos. Although very low levels of maternal D14-3-3e (Acevedo and Skoulakis, unpublished) remained, Leo levels were significantly elevated in D14-3-3e ex4 null homozygotes as expected [17]. Surprisingly however, Leo levels in null homozygotes also expressing 6xHis-g14-3-3 were significantly lower than in animals not expressing the transgene ( Figure 6C). These results indicate that in Drosophila g14-3-3 may be recognized as an endogenous 14-3-3 thereby negating the normal homeostatic response to D14-3-3e loss [17,20]. Because g14-3-3 does not functionally substitute for either of the two fly isoforms this results in the observed enhanced lethality of D14-3-3e ex4 nulls. This in turn confirms that the post-translational modifications of g14-3-3 are essential for the functionality of the Giardia protein [16,[21][22].

g14-3-3 and D14-3-3e are Closely Related
Our data clearly support the proposal [22] that g14-3-3 is homologous to the epsilon group of animal 14-3-3s, as it is structurally and functionally closest to D14-3-3e of the two classes (e and f) of fly 14-3-3s [33]. Because Giardia is suggested to represent an early branching eukaryote [15], our results are congruent with the notion that 14-3-3e represents the member of numbered as in the protein sequence. For each protein, the analysis of phosphorylation is reported on the left panels and the analysis of C-terminal polyglycylation is reported on the right panels. Only peaks corresponding to the unmodified peptides were visible and reported here. doi:10.1371/journal.pone.0078090.g005 Figure 6. Functional consequences of exogenous 14-3-3 expression in Giardia and Drosophila. A) The number of trophozoites, encysting parasites and cysts expressing the FLAG-tagged Drosophila 14-3-3s were estimated by co-staining with anti-CWP and anti-FLAG antibodies after 12 h of growth in encysting medium and counting. Their numbers were compared to those of similarly cultured parasites expressing g14-3-3-FLAG. Results from three independent experiments are reported and the bars represent the total number of cells observed (approximately 1000 parasites per experiment/transfected line) and the error is represented as the standard deviation. B) The number of adult flies homozygous for the null mutation D14-3-3e ex4 alone or expressing g14-3-3 ubiquitously is reported relative to controls arbitrarily set to 100. Error bars are standard errors of the mean. C) Semi-quantitative Western blots from late embryonic extracts of the levels of endogenous Drosophila 14-3-3s in homozygotes for the D14-3-3e ex4 mutation alone or expressing g14-3-3 ubiquitously. Both anti-Leo and anti-D14-3-3e antibodies were applied simultaneously. The level of Tubulin in the lysates was used to control for loading. The ratio of each endogenous 14-3-3 to Tub in control animals was arbitrarily set to 1 and similarly calculated ratios in the experimental genotypes are reported relative to it. The blot is representative of three in total and error bars in the graphs represent the standard errors of the mean. doi:10.1371/journal.pone.0078090.g006 the 14-3-3 family more similar to the ancestral protein which may have given rise to other isoforms by duplication and divergence.
As for the endogenous protein, polyglycylation appears sufficient to exclude D14-3-3e from the nuclei of Giardia trophozoites, in agreement with the suggested role of the polyglycine chain in preventing g14-3-3 nucleo-cytoplasmic shuttling. In contrast, D14-3-3e was not found in the nuclei of encysting parasites despite the presence of a short polyglycine chain. This suggests that polyglycine chain length may be critically important for nuclear import factor interaction and its reduced length on D14-3-3e prohibits translocation. Alternatively, putative import factors are required in addition to length-independent polyglycine chain recognition and they fail to recognize features on D14-3-3e conducive to nuclear localization. In fact, only a fraction of D14-3-3e is phosphorylated in the peptide Ala 197 -Lys 216 and phosphorylation of this region has been proven critical for g14-3-3 function, at least during encystation [16]. The inefficient phosphorylation and polyglycylation of D14-3-3e in Giardia, functionally confirms the importance of its sequence divergence surrounding the posttranslational modification sites on g14-3-3. In agreement with this notion, LeoII (and LeoI and LeoIII) being more divergent in these sequences is not modified at all and as the polyglycylation defective E246A g14-3-3 mutant [16], appears in Giardia nuclei stageindependently. Moreover, the post-translational modifications on g14-3-3 and the consequent changes in subcellular localization may reflect an adaptation similar to that used by Drosophila to increase the functional 14-3-3f isoforms by tissue specific alternative splicing.

g14-3-3 is not Functionally Homologous to D14-3-3e
The kinase responsible for Thr 214 phosphorylation of g14-3-3 in the parasite is currently unknown. Hence we cannot determine whether failure to phosphorylate g14-3-3 in flies is due to absence of a functional homolog. However, g14-3-3 polyglycylation in Giardia is performed by the bifunctional enzyme gTTLL3, a member of the tubulin tyrosine ligase-like family [21]. Interestingly, Drosophila possesses polyglycylated microtubules [34], and 7 distinct genes encoding putative TTLL proteins (Flybase), two of which have been annotated as DmTTLL3A and DmTTLL3B [35]. In Drosophila, DmTTLL3A mono-and polyglycylates the aand the b-tubulin, whereas DmTTLL3B mono-and polyglycylates non-tubulin proteins [35].
It is unlikely that lack of D14-3-3e and g14-3-3 polyglycylation in Drosophila results from distinct spatial distributions with DmTTLL3s. D14-3-3e is present in all Drosophila developmental stages and tissues tested [17,25] and at least for the gonads, it is coexpressed with DmTTLL3A [35,36]. At least DmTTLL3A should be co-expressed with 6xHis-g14-3-3 under the ubiquitous TubGal4 driver [36]. It is likely then that the sequence recognized on g14-3-3 and D14-3-3e by gTTLL3 is not a polyglycylation site for the Drosophila enzymes. Because polyglycylation is not evident in the fly 14-3-3s, this mechanism of g14-3-3 compartmentalization and stage-specific functions in Giardia does not seem to be utilized in Drosophila. Therefore, a major functional modification of g14-3-3 in the parasite is absent from its Drosophila homolog.
Our date indicate that despite the similarity of the two proteins, g14-3-3 is not a functional ortholog of D14-3-3e, at least for processes essential for embryonic viability. The dominant negative effect of g14-3-3 expression in D14-3-3e nulls is probably a consequence of its recognition as an endogenous 14-3-3 protein by the cellular mechanisms responsible for 14-3-3 homeostasis [17,20]. Because g14-3-3 heterodimerizes with Leo, it is likely that the presence of Leo heterodimers, which are absent in D14-3-3e nulls, negate the homeostatic response. The lack of posttranslational modifications and/or aberrant folding potentially rendering g14-3-3 non-functional in Drosophila, suggest that g14-3-3/Leo heterodimers may also contribute to the dominant negative effects of the Giardia protein in the fly embryos.
In conclusion, our data indicate that despite the high level of sequence conservation and structural homology among eukaryotic 14-3-3 proteins, functionality and specificity likely reside in posttranslational modifications and particular amino-acids as previously suggested [17,18]. We propose that the single g14-3-3 from the early branching but highly specialized protozoan Giardia, may represent an ancestral 14-3-3 containing features common to different subgroups of 14-3-3 and that gene duplication and divergence events have subdivided specific 14-3-3 functions, residing in single of a few amino-acid changes, giving rise to the multimember 14-3-3 families in vertebrates and plants.