Phosphoproteome dynamics reveal heat-shock protein complexes specific to the Leishmania donovani infectious stage

Abstract

Leishmania is exposed to a sudden increase in environmental temperature during the infectious cycle that triggers stage differentiation and adapts the parasite phenotype to intracellular survival in the mammalian host. The absence of classical promoter-dependent mechanisms of gene regulation and constitutive expression of most of the heat-shock proteins (HSPs) in these human pathogens raise important unresolved questions as to regulation of the heat-shock response and stage-specific functions of Leishmania HSPs. Here we used a gel-based quantitative approach to assess the Leishmania donovani phosphoproteome and revealed that 38% of the proteins showed significant stage-specific differences, with a strong focus of amastigote-specific phosphoproteins on chaperone function. We identified STI1/HOP-containing chaperone complexes that interact with ribosomal client proteins in an amastigote-specific manner. Genetic analysis of STI1/HOP phosphorylation sites in conditional sti1^−/− null mutant parasites revealed two phosphoserine residues essential for parasite viability. Phosphorylation of the major Leishmania chaperones at the pathogenic stage suggests that these proteins may be promising drug targets via inhibition of their respective protein kinases.

Kinetoplastid parasites of the genus Leishmania generate a variety of pathologies collectively termed leishmaniasis, afflicting millions of people worldwide (1, 2). During the infectious cycle, these insect-borne parasitic trypanosomatids are exposed to a temperature increase following transmission from the invertebrate to the vertebrate host. The temperature change provides a crucial signal for developmental transition of the promastigote insect form to the amastigote form that thrives inside host phagocytes, generating the disease (3). Despite the relevance of heat-induced stage differentiation for pathogenesis, mechanisms underlying the parasite heat-shock response and its role in the development and survival of the amastigote stage remain poorly understood.

Trypanosomatids express highly conserved members of heat-shock and chaperone protein families, suggesting that the cellular response to heat stress is similar between parasite and host (4, 5). However, in contrast to other eukaryotes that regulate heat-induced expression of molecular chaperones and cytoprotective proteins via a family of heat-shock transcription factors (HSFs) (6), trypanosomatid genomes do not encode for classical transacting nuclear factors (7). Gene expression in these organisms relies on highly parasite-specific mechanisms involving poly- cistronic transcription and transsplicing (8, 9). Expression of the major heat-shock proteins (HSPs) is constitutive, even if heat shock may induce a transient increase in synthesis, which has been shown to be regulated exclusively at the posttranscriptional level (10–13). In contrast to its vertebrate host, both constitutive and inducible expression of Leishmania HSPs occurs from the same set of genes, making constitutive and stress-inducible chaperones indistinguishable at the sequence level (11, 14). This important difference in host and parasite biology raises questions concerning the role of Leishmania HSPs at low temperatures in promastigotes and regulation of their chaperone function upon temperature increase in differentiating and proliferating amastigotes. By combining approaches of quantitative phosphoproteomics, systems biology, and mutagenesis, we have uncovered several unique properties of Leishmania donovani HSPs with respect to protein modifications, complex formation, and the importance of chaperone phosphorylation in parasite viability.

Results and Discussion

Two-dimensional differential gel electrophoresis (2D-DIGE) analysis of affinity-enriched phosphoextracts obtained from L. donovani LD1S promastigotes and axenic amastigotes (15, 16) revealed dramatic differences in protein phosphorylation profiles across the major Leishmania infectious stages (Fig. 1A, Fig. S1B, and Dataset S1). A total of 831 protein spots were detected automatically using the DeCyder Differential Analysis Software Package (GE Healthcare), and >700 spots matched between three gels representing independent biological replicates, indicating little experimental variation between samples and highly reproducible two-dimensional gel electrophoresis (2DE) conditions (Dataset S1). A total of 171 proteins were identified by mass spectrometry using the genome database of highly related Leishmania infantum (WWW.GeneDB.org) (Fig. S1A and Dataset S1), including 55 putative phosphoproteins not identified in our previous study using fluorescent multiplex staining (17). Gene ontology (GO) analysis of the Leishmania phosphoprotein dataset via yeast ortholog mapping (Dataset S2) identified six statistically significant GO categories that were overrepresented in our analysis (Fig. 1B and Table S1). Three of these processes—translation initiation, protein folding, and protein catabolism—have been implicated previously in trypanosomatid differential gene expression (9), emphasizing the importance of protein phosphorylation in posttranslational control of this process.

Fig. 1.

Quantitative analysis of L. donovani stage-specific phosphoproteome. (A) 2D-DIGE analysis. Extracts from promastigotes and host-free amastigotes of three independent biological repeat experiments were differentially labeled with the spectrally resolvable CyDye fluors Cy3 and Cy5 and separated by two-dimensional electrophoresis (2DE) on 11-cm (pH 4–7) IPG strips and 12.5% polyacrylamide gels. A merged image of Cy5-labeled amastigotes (red) and Cy3-labeled promastigotes (green) is shown. The molecular weights of marker proteins (kDa) and the pH of the gradient (pI) are indicated. (B) Gene ontology analysis. Overrepresented categories of GO biological processes identified with the BiNGO plugin and visualized with Cytoscape software are shown. Gray levels indicate type I error level (hypergeometric test P value) after false discovery rate correction. Branches of the network selected for statistical significance are represented (P values <5e-02). The node area is proportional to the number of genes that correspond to a given GO category. (C) Analysis of the amastigote phosphoproteome. The readout of the DeCyder Biological Variation Analysis (BVA) module is shown for the most abundant amastigote phosphoprotein cyclophilin 40 (CYP40, spot 572); one isoform of HSP90 (spot 258); stress-induced protein STI1 (spot 361); and hypothetical proteins LinJ15.0040 (spot 767), LinJ32_V3.2410 (spot 683), and LinJ04.0240 (spot 685). Enlarged regions of 2D-DIGE gels for Cy3-labeled promastigotes (pro, green) and Cy5-labeled amastigotes (ama, red), and the corresponding 3D views, are represented. The Bottom shows a graphic representation of differences in abundance of these proteins across three independent experiments. For normalization purposes, a Cy2-labeled internal standard was included, corresponding to a pool of protein from all extracts used in the analysis (st, standard).

We previously analyzed affinity-enriched L. donovani phosphoproteins by qualitative 2DE analysis and demonstrated the specificity of this procedure combining fluorescent phosphoprotein staining and phosphatase treatment (17). In contrast to this analysis, which suggested only little stage-specific phosphorylation, the quantitative 2D-DIGE analysis revealed a statistically significant difference (P value <0.05) in protein abundance for 318 spots corresponding to 38% of the detected phosphoproteins, with 10.2% of the phosphoproteins showing a statistically significant increase in abundance at the amastigote stage of ≥2-fold (Fig. S1B and Dataset S1). Significantly, amastigote phosphoproteins with increased abundance and thus phosphorylation compared to promastigotes were almost exclusively protein chaperones, including several isoforms of HSP90 family member HSP83 (hereafter referred to as HSP90), various HSP70 family members, stress-induced protein STI1/HOP (referred to hereafter as STI1) (18), cyclophilin 40, and the L. donovani ortholog of tetratricopeptide repeat (TPR) domain-containing peptidyl-prolyl-isomerase-like protein LinJ19_V3.1560 (Fig. 1C and Table S2). Previous proteomic studies that quantified changes in protein abundance during axenic amastigote differentiation (Fig. S2) (14), along with quantitative Western blot analysis of promastigote and amastigote total and phosphoextracts (Fig. 2A), demonstrate that the expression of these protein chaperones across the promastigote and amastigote stages is largely constitutive and not induced by elevated temperature. The increased abundance of these HSPs and chaperones in the amastigote phosphoproteome therefore does not simply result from increased expression, but rather reflects a change in phosphorylation stoichiometry with, for example, an 8-fold increase in the phosphorylation ratio of STI1 at the intracellular stage (Fig. 2A).

Fig. 2.

Analysis of phosphoprotein stoichiometry and phosphorylation site determination. (A) Quantitative Western blot. Promastigote (pro) and amastigote (ama) crude (CE) and phosphoprotein (Phospho) extracts were analyzed by Western blotting using polyclonal anti-Leishmania HSP90 and STI1 and monoclonal anti-tubulin (tub) antibodies. The blots were revealed using ZyMax Cy3-conjugated anti-rabbit and ZyMax Cy5-conjugated anti-mouse secondary antibodies (Invitrogen) and signals were detected using a Typhoon variable mode imager. Relative intensities correspond to quantified signals of HSP90 and STI1 after normalization to α-tubulin (ImageQuant software; GE Healthcare). (B) Phosphopeptide analysis. MALDI-TOF/TOF spectrum of one HSP90 peptide isolated from 2D-DIGE gels after tryptic digestion and TiO₂ enrichment is shown. Peptide 519EGVHFEESEEEKQQR533 (m/z 1940.00) of HSP90 contains y₇, y₈, and y₈* ions, enabling identification of Ser526 as the phosphorylated residue. *, fragment ions arising from loss of phosphoric acid (−98 Da); [MH]⁺, precursor ion; [MH-P-18]⁺, precursor ion with loss of one phosphoric acid (−98 Da); pS, phosphorylated serine. (C) Multiple alignments. Sequence elements encompassing HSP90 phosphorylation sites of three trypanosomatids, human and yeast, were analyzed with ClustalXv2. The phosphoresidue is marked by the asterisk.

Evidence for a potential regulatory role of HSP phosphorylation in Leishmania arises from MALDI-TOF/TOF mass spectrometry analysis of the phosphorylation sites of HSP90 and HSP70. Phosphopeptides were isolated after in-gel digestion and peptide extraction from the 2D gels by TiO₂ enrichment (Fig. 2B and Fig. S3A). Manual analysis of the mass spectrometry spectra identified three phosphorylation sites at HSP90 Thr223 and Ser526 and at HSP70 Thr498. Conservation of the threonine residues between Leishmania and human HSP90 and HSP70 identifies this residue as a putative phosphorylation site in higher eukaryotes as well (Fig. 2C and Fig. S3B). Significantly, whereas the Thr223 residue in L. donovani HSP90 corresponds to serine in human HSP90 and thus may be regulated by phosphorylation, HSP90 Ser526 is unique to Leishmania despite the highly conserved sequence to the human homolog (Fig. 2C). The presence of this phosphorylation site opens up the possibility that the function of this major HSP is regulated in a parasite-specific manner by changes in the protein ionic state through phosphorylation, which may affect protein conformation and interaction with other chaperones, thus adding new regulatory features to L. donovani HSP90. Whereas this phosphorylation site is conserved in Trypanosoma HSP90 as judged by multiple alignment (Fig. 2C), this position is occupied in human and mouse HSP90 by aspartic acid. Thus, the HSP90 configuration in higher eukaryotes may be locked into a conformation that mimics constitutive phosphorylation. These findings indicate that regulation of HSP90 functions through posttranslational modifications may substantially differ between parasite and host despite the highly conserved sequence of this protein from Leishmania to man.

Biological network analysis using PathwayArchitect software applied to the identified parasite phosphoprotein datasets revealed the presence of a protein network formed between six amastigote phosphoproteins (Fig. 3A, Table S3). In other eukaryotes, this multimeric chaperone complex has been shown to provide an important signaling function through its interaction with so-called “client proteins,” which include various steroid receptors and protein kinases (19, 20). Cochaperone STI1 plays a crucial role in formation of this complex, acting as a scaffolding protein that mediates the interaction between HSP90 and the HSP70/client protein complexes through specific TPR-rich domains (21, 22). We investigated the presence of the predicted STI1-containing protein complexes in Leishmania by Blue Native (BN) electrophoresis and Western blot analysis. Amastigote-specific phosphorylation of multiple protein chaperones correlated with the presence of numerous STI1-containing complexes ranging from 66 to 480 kDa (Fig. 3B). As judged by coimmunoprecipitation using anti-STI1 antibody, only the amastigote STI1/HSP90 complex interacts with HSP70/client protein complexes (Fig. 3C). Mass spectrometry analysis of the coprecipitated protein bands identified numerous client proteins impli-cated in the assembly of the protein translation machinery and the control of protein translation (Fig. 3D). The specificity of this interaction has been controlled for by using an isotype-specific control antibody (Fig. S3C) and is further supported by the absence of client protein detection in promastigotes, despite their constitutive expression at both stages (23). Our data correlate STI1 phosphorylation with complex formation linked to protein translation, which may affect resistance of ribosomal client pro-teins against proteasome-dependent degradation (24). Although we detected these complexes only in amastigotes, our data do not rule out the presence of similar complexes in promastigotes, which may have escaped our analysis due to their low abundance. These results are reminiscent of the observation that phosphorylation of murine STI1 affects localization of this protein and thus the types of client proteins that interact with STI1/HSP90/HSP70 (25).

Fig. 3.

Chaperone phosphorylation in amastigotes is linked to formation of a stage-specific multiprotein complex. (A) Biological network analysis. Analysis was carried out with PathwayArchitect software (version 3.0.1; www.stratagene.com). An input set of 43 yeast IDs was used to build the biological interaction network from annotations extracted from scientific literature (automatic scanning of abstracts from PubMed database: www.ncbi.nlm.nih.gov/pubmed). Shaded levels indicate the fold increase in amastigote phosphoprotein abundance compared to promastigote phosphoextracts. (B) Blue-Native PAGE. Native extracts from promastigotes (P) or amastigotes (A) were separated on NativePAGE Novex 4–16% Bis-Tris gels (Invitrogen) and complexes were revealed by colloidal Coomassie staining (Left). Replica gels were electroblotted onto PVDF membranes and proteins detected with anti-STI1 antibody (Right). Molecular weight (MW) of native protein marker (M) is shown. (C) Coimmunoprecipitation. Pro- (P) and amastigote (A) crude extracts were incubated with STI1 polyclonal antibody and protein A MicroBeads (Miltenyi Biotec). Eluates were separated by denaturing SDS/PAGE and stained with SyproRuby. MW of marker proteins in kDa is shown. (D) STI1-associated proteins isolated by immunoprecipitation were identified by MS/MS analysis.

We next used a genetic approach to investigate the biological significance of STI1 phosphorylation by mutagenesis. Three STI1 phospho-site mutants were generated on the basis of previously identified STI1 phosphoresidues in Trypanosoma brucei (26), mouse (27), and human (28) (Fig. 4A). As STI1 appears to be essential, we tested these mutants using a conditional knockout system (29). In this approach, to guard against the lethal phenotype, both chromosomal STI1 alleles were inactivated in the presence of an episomal plasmid expressing WT STI1, yielding the mutant sti1^−/−/pXNG-STI1 (Fig. 4B and Fig. S4). The episomal plasmid pXNG4SAT (29) additionally carries both a fluorescent (GFP) and a negative selectable thymidine kinase (TK) marker rendering parasites susceptible to the antiviral drug ganciclovir (GCV). Thus by FACS or drug selection sti1^−/−/pXNG-STI1 parasites could be tested for their requirement to maintain the ectopic STI1 gene copy. The functionality of mutated STI1 genes was then tested in a “plasmid shuffle” (30), by introducing a second plasmid, and asking whether the WT STI1 borne on pXNG could be lost. As expected for an essential gene, it was not possible to segregate away pXNG-STI1 in the chromosomal sti1^−/− null mutant, even in the presence of GCV (Fig. 4C, parental line). The presence of an additional copy of functional STI1, however, allowed for efficient elimination of pXNG-STI1 during negative selection as judged by the substantial reduction in GFP fluorescence intensity (STI1_WT, Fig. 4 C and D Left).

Fig. 4.

STI1 phosphorylation is essential for L. donovani viability. (A) Multiple alignment of STI1 homologs from Leishmania major (Lmajor, CAJ02290.1), L. infantum (Linfantum, CAM65800.1), Trypanosoma brucei (Tbrucei, CBH11274.1), Trypanosoma cruzi (Tcruzi, EAN97552.1), Mus musculus (Mouse, AAC53267.1), Homo sapiens (Human, AAA58682.1), and Saccharomyces cerevisiae (Yeast, CAA60743.1) using ClustalXv2 is shown. The number indicates the L. major STI1 amino acids targeted for analysis by mutagenesis using the QuikChange Site-Directed Mutagenesis Kit (Stratagene). Phosphorylation sites at serine 15 (S15) and serine 481 (S481) have been experimentally identified in human and T. brucei, respectively (26, 28). Threonine 217 (T217) has been selected on the basis of experimentally identified threonine phosphorylation upstream of the yeast STI1 nuclear localization signal (NLS, dotted line) and the presence of a casein kinase II signature sequence S/T-X-X-D/E (27). The shading indicates the level of amino acid conservation. (B) Generation of L. donovani sti1^−/− conditional null mutants for analysis of phosphorylation site-specific mutants. Heterozygous sti1^+/− null mutants (STI1/Δsti1::PAC) were transfected with episomal vector pXNG carrying the STI1 wild-type ORF (STI1/Δsti1::PAC[pXNG-STI1]), before replacement of the second STI1 allele yielding homozygous sti1^−/− null mutants (Δsti1::BLEO/Δsti1::PAC[pXNG-STI1]). Independent clones were transfected with episomal vector pLEXSY expressing either an additional copy of wild-type STI1 or one of the phospho-site mutants described in A. The ability of episomally expressed STI1 phospho-site mutants S15A, T217A, and S481A to complement for the loss of pXNG-STI1 during negative selection with gangciclovir (GCV) provides a genetic test to distinguish mutations that affect STI1 function (case 1, pXNG-STI1 is maintained) from silent mutations (case 2, pXNG-STI1 is lost). Selections were performed using 25 μg/mL puromycin, 150 μg/mL nourseothricin, 5 μg/mL bleomycin, and 50 μg/mL hygromycin B, respectively. BLEO, bleomycin resistance cassette; GCV, ganciclovir; GFP, green fluorescent protein gene; HYG, hygromycin B resistance cassette; PAC, puromycin resistance cassette; SAT, nourseothricin resistance cassette; TK, herpes simplex virus thymidine kinase gene. (C) Analysis of STI1 mutants by negative selection in conditional sti1^−/− lines. Elimination of pXNG-STI1 was followed by FACS analysis monitoring the levels of GFP expressed from the same episome. GFP mean fluorescence of conditional sti1^−/− null mutant promastigotes treated for three culture passages with 50 μg/mL GCV (Left and Center) or axenic amastigotes treated for 72 h with GCV (Right) is shown. Data are means ± SD of a representative experiment analyzed in triplicate. (D) Histogram plots of one representative analysis of axenic amastigotes after 72 h of GCV selection (shaded histograms). Dotted lines represent the background fluorescence of untransfected control parasites. Solid lines represent the GFP fluorescence levels of parasites cultured without GCV.

This binary readout allowed us now to test for the functionality of the STI1 phosphorylation site mutants. The mutants were expressed in independent sti1^−/−/pXNG-STI1 clones, and their effect on negative selection against pXNG-STI1 was analyzed. Expression of the T217A mutant fully compensated for pXNG-STI1, which was efficiently eliminated during negative selection (Fig. 4 C and D). This result suggests that the T217A mutation does not affect the functional properties of STI1 (case 2, Fig. 4B). In contrast, expression of neither S15A nor S481A was able to compensate for pXNG-STI1, which was maintained at levels comparable to the sti1^−/−/pXNG-STI1 parental line in both promastigote and amastigote stages (Fig. 4 C and D). Together these data suggest that the STI1 residues S15 and S481 are essential phosphorylation sites (case 1, Fig. 4B) required for L. donovani viability in culture and emphasize the importance of chaperone phosphorylation in parasite biology. Further analysis of the role of these phosphorylation sites in STI1 complex formation was precluded by the lethal sti1^−/− phenotype and the requirement to maintain WT STI1.

The manner in which Leishmania regulates the response to stress is fundamentally different from that of other eukaryotes, including the mammalian host. In most organisms, stress-induced expression increases stress tolerance through protection of basic cell functions, without a major impact on cellular morphology and phenotype. In Leishmania, however, stress signals such as low pH and nutritional starvation, or high temperature, induce developmental programs that lead to differentiation of metacyclic promastigotes and amastigotes and adapt the parasite for transmission and intracellular survival (3, 31, 32). This reinterpretation of the stress response likely translates into unique regulatory mechanisms and interactions of Leishmania chaperones. Our data provide important insights into these parasite-specific mechanisms, which may depend on (i) stage-specific chaperone phosphorylation during environmentally induced parasite differentiation; (ii) phosphorylation of unique residues in parasites HSP70, HSP90, and STI1; and (iii) formation of chaperone complexes. The emphasis on posttranslational regulation of the stress response in Leishmania through phosphorylation and other protein modifications may represent an evolutionary adaptation of trypanosomatid parasites to constitutive expression and the absence of transcriptional regulation (9). By analogy to stress regulation through HSFs in other eukaryotes (33), our data indicate that the absence of these factors in trypanosomatids may have been compensated for by the evolution of protein kinases that regulate chaperone function. In resting cells, HSF1 is inactivated in other eukaryotes through its association with HSP70 and HSP90, but is released and activated under stress conditions (34). In a similar fashion, Leishmania chaperones may tether protein kinases that are released and activated during environmentally induced stage differentiation. Indeed, binding of Leish-mania MAP kinases to HSP70 (35) and the direct implication of HSP90 in parasite differentiation established in geldanamycin-treated parasites (36) suggest a unique model in which chaperone/kinase interactions would regulate phosphotransferase activities, which in turn might directly feed back on chaperone functions and differentiation.

Materials and Methods

Cell Culture and Differentiation of L. donovani.

The L. donovani strain 1S2D (MHOM/SD/62/1S-CL2D), clone LdB was cultured and axenic amastigotes were differentiated as described (16, 37).

Preparation of L. donovani total and phosphoprotein extracts.

Axenic promastigotes or axenic amastigotes 48 h after induction of differentiation by pH and temperature shift were harvested from logarithmic cultures. For phosphoprotein purification, protein concentration was adjusted to 0.1 mg/mL, and 2.5-mg extracts were applied onto equilibrated affinity columns of the phosphoprotein purification kit (Qiagen) according to manufacturer's instructions.

Western Blot Analysis.

Proteins were revealed using the following antibodies: polyclonal anti-STI-1 (18) and anti-HSP83, mouse monoclonal anti-α-tubulin antibody (Sigma), and anti-rabbit or anti-mouse ZyMax Cy3 or Cy5 conjugated secondary antibodies (Invitrogen). In some experiments, NativePAGE Novex 4–16% Bis-Tris Gels (Invitrogen) were transferred onto PVDF membranes, and proteins were revealed using the antibodies described above and anti-rabbit HRP-conjugated secondary antibodies (Pierce).

Immunoprecipitation.

Cells were lysed and incubated for 45 min at 4 °C with STI-1 polyclonal antibody (10 μL, 200 μg/mL) and protein A MicroBeads (50 μL; Miltenyi Biotec). Mixtures were loaded on μMACS columns (Miltenyi Biotec). Eluates were separated by denaturing SDS/PAGE, gels were stained with SyproRuby, and bands of interest were excised and further analyzed by MS-MS/MS.

Blue Native PAGE.

Native extracts were centrifuged at 20,000 × g for 30 min at 4 °C. Twenty or 40 μg of protein, containing 0.025% Coomassie Blue G-250, were separated on NativePAGE Novex 4–16% Bis-Tris Gels (Invitrogen) at 150 V for 3 h and 250 V for 1 h at 4 °C.

Sample Preparation and DIGE Labeling.

Phosphoprotein pellets were resuspended in DIGE sample buffer (7 M Urea, 2 M Thiourea, 4% CHAPS, 30 mM Tris, pH 8.5) to a final protein concentration of 5.0 mg/mL. Phosphoprotein extracts from promastigotes and amastigotes were differentially labeled with the spectrally resolvable Cy3 and Cy5, and a pool of both extracts was labeled with Cy2 for normalization purposes, following the manufacturer's recommendations (GE Healthcare). Three independent biological replicates of promastigote and amastigote phosphoextracts were prepared and resolved by 2D-DIGE. In addition, a gel containing a pool of either pro- or amastigote extracts from all extractions was included for normalization purposes.

2DE.

A total of 90 μg of protein sample containing 30 μg of Cy3- and Cy5-labeled samples was pooled together with 30 μg of Cy2-labeled control and adjusted with 150 μL Destreak rehydration buffer (GE Healthcare) containing 0.5% IPG buffer 4-7 and 1.0% DTT. Samples were simultaneously separated in the first dimension by iso-electric focusing (IEF) overnight (see SI Materials and Methods for details).

Staining Procedures and Image Analysis.

After electrophoresis, gels were scanned on a Typhoon 9410 Variable Mode Imager (GE Healthcare) and analyzed with DeCyder 6.5 software (GE Healthcare).

Protein Identification by Mass Spectrometry.

Spots of interest were excised from gels using the ProPic Investigator robotic system (Genomic Solutions). MS and MS/MS raw data for protein identification were obtained as previously described (17) (see SI Materials and Methods for details).

Phosphopeptide Identification.

Protein digests, obtained as described above, were diluted in loading buffer (80% ACN, 5% TFA) (38) and loaded on TiO₂ microcolumns as described previously (39). All phosphorylated peptides were first analyzed for the presence of the major fragment ion [MH-H₃PO₄]⁺ = MH − 98 Da corresponding to the loss of the phosphate moiety and identified positively by MASCOT. In addition, all MS/MS spectra were carefully curated manually for assignment of phosphorylation sites (see SI Materials and Methods for details).

Bioinformatics Approaches.

Retrieval of yeast orthologous sequences was carried out with a BLASTP algorithm (40) by querying the Saccharomyces cerevisiae protein database with the L. infantum protein sequences (default parameters; Refseq protein database build 2.1 released May 17, 2005; www.ncbi.nlm.nih.gov/genome/seq/BlastGen/BlastGen.cgi?taxid=4932). From 92 unique L. infantum IDs used in this analysis, 17 IDs were specific to Leishmania and 75 IDs mapped to 71 yeast orthologs with expectation values ranging from E = 0 to E = 0.003 and alignment scores ranging from 974 to 38.9 bits. Functional enrichment analysis was carried out with the 71 yeast IDs as input using the BiNGO plugin (version 2.1) of the Cytoscape software (version 2.5.1). Biological network analysis was carried out with PathwayArchitect software (version 3.0.1; www.stratagene.com). An input set of 43 yeast IDs was used to build the biological interaction network from the annotations extracted from the scientific literature (automatic scanning of abstracts from PubMed database; www.ncbi.nlm.nih.gov/pubmed/).