Cell lines and culture

Leishmania donovani infantum MHOM/MA/67/ITMAP-263 promastigotes were cultured at 25°C in SDM-79 medium (pH 7.0) supplemented with 10% heat-inactivated fetal calf serum (Multicell, Wisent Inc.) and 5 mg/ml hemin. Axenic amastigotes were grown in MAA/20 medium (pH 5.8) at 37°C with 5% CO₂ as described [27] and used after two to three passages, when fully differentiated. Differentiating amastigotes were grown for 8 to 24 hours in MAA/20 medium. For transfectant selection, 30 μg/ml G418 and 160 μg/ml hygromycin B were used. For growth phenotype analysis, 10⁶ stationary-phase promastigotes were counted on a Malassez (hemocytometer) and then inoculated in 5 ml of fresh SDM media to measure the O.D._600nm every 24 hours. For axenic amastigotes, a similar approach was used but with 10⁷ parasites as a starter. Experiments were done in triplicate, using in average passage 4–5 for promastigotes and passage 3 for amastigotes.

DNA manipulations and plasmid preparation

LiAlba3^-/- (LinJ.34.2410; previously named LiAlba20^-/-) and LiAlba1^-/- (LinJ.13.0270; previously named LiAlba13^-/-) knockout mutants and add-back strains obtained by episomal transfection of plasmids pSP-alphaIRNEOalphaIR-HA-LiAlba3 and pSP-alphaIRNEOalphaIR-LiAlba1-HA into the corresponding null mutant background were described previously [18]. The strategy used for episomal expression of Leishmania proteins of interest fused to fluorescent proteins or for genomic integration is presented in S1A Fig. The fluorescent coding sequence (e.g. mCherry, and yellow fluorescent protein (eYFP)) was fused at the N-terminal part of LiAlba1, LiAlba3 or LiNop10 to keep the natural context of 3'UTR, hence allowing efficient regulation of the transcripts. Briefly, a first cassette containing a selectable marker gene (NEO or HYG), the alpha-tubulin intergenic region (alphaIR) for optimal processing and expression of the selectable marker, and the fluorescent protein without the stop codon was prepared by Phusion PCR using Phusion enzyme (Thermo Scientific) according to the manufacturer instructions. Primers (from Integrated DNA Technologies) used for PCR amplifications are listed in S1 Table. The NEO-alphaIR-eYFP and HYG-alphaIR-mCh cassettes were cloned into pGEM-T Easy (Promega) for sequencing and further sub-cloning experiments. The second step consisted in the amplification of a 500 bp fragment corresponding to the 5'-intergenic region of the target gene, and the ORF (cloned in frame with the mCherry and eYFP fluorescent proteins) together with the corresponding 3'-intergenic region (~500 bp). These two fragments were fused by PCR and an EcoRV site was added at the junction. Cloning in pGEM-T Easy and digestion using EcoRV allowed us to perform a blunt insertion of the cassette to finally obtain the pGEM-NEOalphaIR-eYFP-LiAlba1, pGEM-HYGalphaIR-mCh-LiAlba3 and pGEM-HYGalphaIR-mCh-LiNop10 vectors. The mCh-LiAlba3 cassette was amplified from pGEM-HYGalphaIR-mCh-LiAlba3 using primers containing XbaI (5'end) and NdeI (3'end) restriction sites. PFR2C-HA was produced by PCR amplification with specific primers harboring the HA sequence and XbaI and NdeI sites. Both constructs (S1B Fig) were sub-cloned into pSP-alphaIRNEOalphaIR XbaI and NdeI sites to generate pSPalphaIRNEOalphaIR-mCh-LiAlba3 and pSP-alphaIRNEOalphaIR-PFR2C-HA vectors, respectively as previously described for the Alba-HA constructs [18]. Transfections were performed by electroporation as previously described [28].

Sequence alignments

All the Alba protein sequences from trypanosomatids were downloaded from TritrypDB.org (v4.0). For both LiAlba sequences, protein secondary structure prediction was performed using the Phyre webserver [29]. Alba-domain protein alignments in Trypanosoma and Leishmania species were carried out with MEGA5 software using ClustalW (Gonnet matrix and 10 gaps allowed). Neighbor-joining tree was build using MEGA6 bootstrap method and Poisson model for substitutions. Phyre results and ClustalW alignment were merged using the BioEdit software.

Immunofluorescence studies

For indirect immunofluorescence, we used parasites episomally expressing pSP-alphaIRNEOalphaIR-LiAlba1-HA and pSP-alphaIRNEOalphaIR-HA-LiAlba3 in the L. infantum LiAlba1^-/- and LiAlba3^-/- null mutant background, respectively. Approximately 10⁶ parasites were centrifuged, washed 3 times with PBS and kept on ice before attachment on microscopy slides for 10 min at room temperature. Cold 100% methanol was used for cell fixation for 10 min at -20°C and 3 x 5 min washes with PBS were performed prior to a 30 min permeabilization at room temperature in the following solution (1X PBS, 0.2% Triton X-100, 5% Fetal Calf Serum). Blocking buffer (1X PBS; 2% BSA; 0.02% Tween-20; 0.1% Triton X-100) was added for overnight incubation at 4°C. An anti-HA antibody (Applied Biological Materials Inc.) diluted 1:400 in blocking buffer was used as primary antibody for 1 hour incubation at room temperature, and unbound material was removed by 5 x 5 min washes in PBS. 2 μg of secondary antibody (Alexa Fluor 488 goat anti-mouse, LiveTech) were used under similar conditions in the dark and washed as described for the anti-HA antibody. For nuclear and kinetoplast DNA staining, DAPI solution (Sigma) diluted to 0.5 μg/ml in 1X PBS was added for 5 min at room temperature and then washed 2 x 5 min with 1X PBS. After drying the slides, a drop of mounting solution (Mobi Glow) and a cover slide were added and sealed with nail polish. For fluorescent protein imaging, we used approximately 10⁶ L. infantum episomally transfected or co-transfected with pGEM-HYGalpha-mCh-LiAlba3, pGEM-NEOalphaIR-eYFP-LiAlba1 and/or pGEM-HYGalphaIR-mCh-LiNop10. Parasite cultures were centrifuged, washed in PBS twice, and incubated in 2% paraformaldehyde (Sigma) in PBS for 10 min at room temperature. Cells were dropped on slide and dried prior DAPI staining as described above. After slide drying, Mobi Glow and cover slide were added and sealed with nail polish. All slides were analyzed by epifluorescence microscope using a Nikon Eclipse TE300 with a 100X objective and oil immersion. Images acquisition was performed with ImagePro Plus 5 software, and ImageJ or Photoshop were used for picture merging and montage.

Co-immunoprecipitation studies

For identifying interacting partners of the Alba-domain proteins, we carried out co-immunoprecipitation studies from recombinant L. infantum expressing pSP-alphaIRNEOalphaIR-LiAlba1-HA or pSP-alphaIRNEOalphaIR-HA-LiAlba3 vectors. For amastigotes and promastigotes, 5x10⁹ cells in exponential phase were pelleted and resuspended in 1 ml of cytoplasmic extraction buffer (CEB: 10 mM Hepes, 3 mM MgCl2, 14 mM KCl, pH 7.5, 5% glycerol and 0.2% NP-40), supplemented with 1 mM PMSF and 2X cOmplete ULTRA Tablets Mini, EDTA-free EASYpack (Roche). Cells were lysed with 15 to 20 strokes using a Douncer homogenizer and lysis was verified under the microscope before centrifugation (13 000 rpm) for 10 min at 4°C to remove insoluble material. To avoid identification of proteins bound to the beads, protein extracts were washed on a rotor with 60 μl of Dynabeads Protein G (Invitrogen) for 1 hour at 4°C, and resuspended in 200 μl PSB-0.1% Tween. In parallel, another 60 μl of beads were resuspended in 200 μl PSB-0.1% Tween, incubated with 10 μg of an anti-HA antibody (Applied Biological Materials Inc.) during 1 hour at 4°C on a rotor. Unbound antibody was removed by three washes of 5 min in 500 μl PSB-0.1% Tween and the “washed”proteins were added to the beads and antibody mix. Incubation was carried out for 90 min at 4°C on a rotor and unbound material was removed by 5 washes using PSB-0.1% Tween supplemented with 1X cOmplete ULTRA Tablets Mini, EDTA-free EASYpack. The mix was transferred to a new Eppendorff tube and elution was performed by direct addition of 60 μl of 2x Laemmli Sample Buffer and incubation for 5 min at 100°C. Elutions were loaded on 12% SDS-PAGE and migration was stopped as soon as the bromophenol blue marker entered to the resolving part of the gel. After Coomassie staining and destaining (to remove traces of SDS), bands were excised and conserved in 1% acetic acid for MS/MS peptide identification (see below). The results presented in Table 1 are only peptides corresponding to proteins co-immunoprecipitated with the HA-Alba proteins and are absent from the wild type negative control (data not shown). Immunoprecipitations from L. infantum wild type, HA-LiAlba3 and LiAlba1-HA were all run in triplicates.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. L. infantum Alba-domain protein putative partners identified by co-immunoprecipitation and LC-MS/MS analysis.

https://doi.org/10.1371/journal.pone.0137243.t001

Flagellum purification and flagellar proteome analysis

To purify intact flagella from Leishmania parasites, we used the method recently described in [30]. 3x10⁹ parasites were resuspended in 2 mL of Buffer A (25 mM tricine; 0.2 mM EDTA; 5 mM MgCl₂; 12 mM beta-mercaptoethanol, pH 7.0) supplemented with 0.32 M sucrose, 1% BSA, 0.1 mM CaCl₂ and protease inhibitor tablet (Roche). To detach the flagellum, cells were vortexed for 10 min at 4°C and centrifuged at 420 g for 10 min. The supernatant was decanted and recentrifuged twice, and the flagella were pelleted on a table top centrifuge at 13,000 rpm for 5 min at 4°C. Sediment was resuspended in 500 μl of Buffer A + 0.32 M Sucrose, layered on top of a discontinuous sucrose gradient of 1.49/1.66/1.84/2.02 M prepared in Buffer A, and centrifuged (28,000 rpm) overnight at 4°C in a Beckman SW40 Ti rotor, and the flagellum were enriched in the 1.66 M sucrose fraction. This fraction was recovered and centrifuged in table top centrifuge at 13,000 rpm for 30 min at 4°C. The pellet was resuspended in Buffer A and centrifuged at maximum speed for 15 min, and finally resuspended in 25 μl of 2X Laemmli buffer and loaded on a 12% SDS-PAGE. Then, the gel was used for Western blot analysis or MS/MS peptide identification as described below.

Mass spectrometry and database searching

Peptide samples were separated by online reversed-phase (RP) nanoscale capillary liquid chromatography (nanoLC) and analyzed by electrospray mass spectrometry (ES MS/MS). The experiments were performed with Agilent 1200 nanopump connected to a 5600 TripleTOF + mass spectrometer (AB Sciex, Framingham, MA, USA) equipped with a nanoelectrospray ion source. Peptide separation took place on a self-packed PicoFrit column (New Objective, Woburn, MA) packed with Jupiter (Phenomenex) 5u, 300A C18, 15 cm x 0.075 mm internal diameter. Peptides were eluted with a linear gradient from 2–50% solvent B (acetonitrile, 0.1% formic acid) in 30 min, at 300 nL/min. Mass spectra were acquired using a data dependent acquisition mode using Analyst software version 1.6. Each full scan mass spectrum (400 to 1250 m/z) was followed by collision-induced dissociation of the twenty most intense ions. Dynamic exclusion was set for a period of 12 sec and a tolerance of 100 ppm. All MS/MS peak lists (MGF files) were generated using ProteinPilot (AB Sciex, Framingham, MA, USA, Version 4.5) with the paragon algorithm. MGF sample files were then analyzed using Mascot (Matrix Science, London, UK; version 2.4.0) to search the TriTrypDB database (8 May 2014 TriTrypDB 8.0 released). Mascot was searched with a fragment ion mass tolerance of 0.10 Da and a parent ion tolerance of 0.10 Da. Scaffold (version 3.4.5), Proteome Software Inc., Portland, OR) was used to validate MS/MS based peptide and protein identifications. Protein probabilities were assigned by the Protein Prophet algorithm [31]. Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony.

Polysome fractionation

Polysome fractionation was analyzed as previously described [32]. Briefly, 3×10⁹ wild type parasites were treated with 100 μg/ml cycloheximide (Sigma) for 10 min and washed with cycloheximide-containing PBS buffer. Cell lysis was performed with a Dounce homogenizer in lysis buffer [10 mM Tris-HCl pH 7.4, 10 mM MgCl2, 150 mM NaCl, 0.5% IGEPAL, 100 μg/ml cycloheximide, 1 mM PMSF, 100 U/ml RNAseOUT (Amersham), 2X final concentration of protease inhibitor tablet (Roche)]. Lysates were divided into two tubes, and one of them was treated 10 min at room temperature with 50 mM EDTA as control. All the lysates were pelleted by centrifugation and the supernatant (40 OD_{260 nm} units) was layered on top of a 15% to 45% linear sucrose gradient (10 ml) prepared in gradient buffer (50 mM Tris-HCl pH 7.4, 10 mM MgCl2, 50 mM KCl, 100 U/ml RNaseOUT). Ultracentrifugation (35000 rpm for 2.15 hours at 4°C) was performed in a Beckman SW40 Ti rotor, and fractions were collected with an ISCO Density Gradient Fractionation System under constant monitoring of the absorbance at 254 nm. Proteins from the various fractions were concentrated using the TCA method, and finally resuspended in 50 μl of 2X Laemmli buffer. 10 μl of each fraction were loaded on 12% SDS-PAGE and transferred for Western blot analysis. Anti-HA (100 ng/ml, Applied Biological Materials Inc.) or anti-TbAlba340/2 (GeneID: Tb927.4.2040 [26],) and HRP-linked anti-mouse (NEB) antibodies were used for 1 h at room temperature.

Alba-domain proteins of Leishmania

In contrast to the T. brucei genome that encodes four Alba-domain proteins (Alba1-4) corresponding to the Rpp20- and Rpp25-like eukaryotic subgroups, the related Leishmania species sequenced to date [7, 33, 34] possess only two Alba-domain proteins (LinJ.13.0270 (LiAlba1) and LinJ.34.2410 (LiAlba3) in L. infantum), as illustrated in the Neighbor-joining tree presented in Fig 1A. LiAlba1 is part of the Rpp20-like subgroup whereas LiAlba3 belongs to the Rpp25-like subgroup. This suggests a duplication of Alba genes after Leishmania and Trypanosoma speciation. In T. cruzi, similarly to T. brucei, there are four Alba-domain proteins based on sequence identity (from 65–73%) that are annotated as hypothetical proteins (Fig 1 and not shown).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Sequence alignment and phylogeny of Alba-domain proteins in Leishmania, Trypanosoma brucei and T. cruzi species.

(A) Neighbor-joining tree showing the phylogenetic relationship between the Alba-domain proteins of TriTryps. Evolutional distances (scale) were estimated as the number of amino acid substitutions per site, considering Poisson correction. The two subgroups Rpp20-like and Rpp25-like are marked. (B) ClustalW alignment of Rpp20-like Alba-domain proteins merged with the in silico structure prediction of LiAlba1 (LinJ.13.0270) using the Phyre algorithm. The best score was obtained with the Alba protein from Sulfolobus solfataricus (Ss)(NCBI WP_010923153.1). ss: secondary structure in silico prediction; C: coil; H: helix; E: Sheet. Red squares indicate amino acids known to be phosphorylated on these specific genes (T. cruzi and T. brucei). The black star shows the expected position for Sir2 acetylation and the red stars underline the signature motif of the subgroup. Sequence variations in coiled regions between Trypanosoma spp. and Leishmania spp. are underlined with a black bar. LiAlba1 was used for Phyre structure prediction. (C) As in B for the Rpp25-like Alba-domain proteins. Structure prediction of LiAlba3 using Phyre outputs Alba2 from Aeropyrum pernix (Ap) K1 (NCBI WP_010866616.1) as the best match. LiAlba3 (LinJ.34.2410) was used for Phyre structure prediction. LinJ: L. infantum; LmjF: L. major; LtaP: L. tarentolae; LbrM: L. braziliensis; Tb: T. brucei; Tc: T. cruzi.

https://doi.org/10.1371/journal.pone.0137243.g001

Members of the Rpp20-like subgroup have a molecular weight of 12–14 kDa. Surprisingly, although they exhibit a basic pI in Trypanosoma spp., the Leishmania orthologs have an acidic pI (pI 4.8 in the case of L. infantum compared to 9.8 in T. brucei). This suggests that Leishmania Alba-domain proteins of the Rpp20-like subgroup have different charged amino acids, underlying differences in their abilities to bind their substrates. The charge differences are mainly observed in the C- and N-terminal regions (Fig 1B). Within Leishmania species, Rpp20-like proteins are well conserved sharing amino acid sequence identities up to 88–99%. However, identity of LiAlba1 protein with T, brucei Rpp20-like proteins TbAlba2 (Tb11.02.2030) (Tb927.11.4450) and TbAlba1 (Tb11.02.2040) (Tb927.11.4460) drops to 57% and 56%, respectively.

Given the lower primary sequence similarity between the Leishmania and Trypanosoma Alba proteins and particularly LiAlba1 and TbAlba2, we carried out in silico structural comparative analysis using the Phyre algorithm. The best score was obtained with the Alba protein from Sulfolobus solfataricus, Sso10b1 [35], with an E-value of 8.2e-09. The βαβαββ topology is conserved in LiAlba1 with an additional α helix at the C-terminal region (Fig 1B), corresponding to the previously described synapomorphic C-terminal extension containing the FDxh signature motif, which is specific of the Rpp20-like subgroup [20]. The coil regions are the most divergent in trypanosomatids but are well conserved between Leishmania spp. (positions 3–13, 17–22, 80–90 and 116–131 of the alignment in Fig 1B). The lysine residue usually deacetylated by Sir2 (reviewed in [36, 37]) is not conserved in trypanosomes (position 19 of the alignment) but we cannot rule out the possible acetylation on neighbor lysine residues. While the acetylation of Alba-domain proteins by Sir2 has been characterized in Archaea [38] and Plasmodium falciparum [22], there are no published reports in trypanosomatids to date. Another possible well-known posttranslational modification (PTM) is protein phosphorylation. It has been reported that TbAlba2 in T. brucei bloodstream forms is phosphorylated at Serine 24 [39], an amino acid that is not conserved in Leishmania spp. (Fig 1B). However, in T. cruzi metacyclics, Serine 2 of Tc00.1047053504089.70 (TcCLB.504089.70) is phosphorylated and this amino acid is well conserved in Leishmania [40] (Fig 1B). Although the Alba-domain proteins from L. infantum have not been identified in phosphoproteome studies so far [41–43], the conservation of Ser 2 suggests that these proteins might be phosphorylated in Leishmania. Phosphorylation of Alba-domain proteins on Ser/Thr residues has also been reported in Toxoplasma gondii and P. falciparum [44–46]. In P. falciparum, phosphorylation by protein kinase CK2 [47] has been characterized for PfAlba1 and 2, belonging to the Rpp20-like subgroup.

In trypanosomatids, Rpp25-like proteins have a molecular weight ranging from 20 to 25 kDa and a pI ranging from 9.9 to 10.3. Within Leishmania spp., Rpp25-like proteins share 83–96% identity, with L. tarentolae being the most divergent (83%). Sequence comparison of LiAlba3 with Trypanosoma spp. Rpp25-like proteins revealed 53–60% identity. The sequence of LiAlba3 shares 56% identity with the TbAlba3 (Tb927.4.2040) and 53% with TbAlba4 (Tb927.4.2030). Structure prediction of LiAlba3 using Phyre outputs Alba2 from Aeropyrum pernix K1 [48] as the best match (E-value 1.2e-07). The βαβαββ secondary structure seems conserved in LiAlba3 with an additional α helix in the N-terminal region and a longer C-terminal extension containing the RGG-box domain, which is specific of Rpp25-like proteins [20] together with the signature sequence GYQxP (position 146–150 of the alignment) (Fig 1C). As for LiAlba1, the coil regions are the ones with the highest divergence between Leishmania spp. and Trypanosoma spp. (positions 15–23 and 107–118 in the alignment of Fig 1C). Here, the lysine residue deacetylated by Sir2 is replaced by a threonine in Leishmania (position 38 of the alignment). In P. falciparum, deacetylation of PfAlba3 (Rpp25-like subgroup) has been reported [21]. In T. brucei procyclic form, TbAlba3 is phosphorylated at tyrosine 4 [49]. Tyrosine phosphorylation is rare in trypanosomatids and this conserved amino acid in Trypanosoma spp. is replaced by an arginine in all Leishmania species. In T. cruzi metacyclics, Ser residues 144 and 157 are phosphorylated [40], and only Ser144 is conserved in trypanosomatids, suggesting a potential phosphorylation of LiAlba3 on this residue.

Taken together, these sequence predictions indicate that Alba-domain proteins have a conserved structure in trypanosomatids, but Leishmania harbors specific mutations in key regulatory sequences shown to be phosphorylated in T. brucei or T. cruzi. The RGG-box remains also a good candidate for arginine methylation. Recent work in T. brucei and Leishmania showed that orthologs of LiAlba3 are methylated [50, 51]. In L. major and L. infantum, the RGG-box demonstrates a good conservation of the direct repeat RGGRGGS, resulting in three SR sites, a structure known to be phosphorylated in yeast and to play a role in nuclear import [52].

To look for putative PTMs in LiAlba3 protein, we loaded protein samples from L. infantum promastigotes and differentiating amastigotes on 2D SDS-PAGE and carried out a Western blot using the T. brucei TbAlba2040/2 antibody produced in mice [26] that specifically recognizes the LiAlba3 protein [18]. Interestingly, we detected 4 to 5 different forms of LiAlba3 in promastigotes with pI values ranging from 10 to 11 (S2A Fig), which suggests different levels of PTMs. During amastigote differentiation, only one prominent form was seen at pI 10.25 (S2B Fig).

Alba-domain protein associated partners

To get further insight into the functional role of LiAlba1 and LiAlba3 proteins, we carried out co-immunoprecipitation studies coupled with MS/MS peptide identification to characterize putative Alba-domain interacting partners. For these studies, we used L. infantum add-back strains episomally expressing the LiAlba1-HA or HA-LiAlba3 proteins into the LiAlba1^-/- and LiAlba3^-/- background, respectively (previously named LiAlba13^-/- and LiAlba20^-/-, respectively in [18]) to more closely mimic endogenous levels of Alba proteins. We have shown previously [18] that episomal expression of Alba proteins in the knockout strain background resulted in similar expression levels to those of the endogenous Alba proteins. The MS/MS results from both L. infantum promastigote and axenic amastigote forms, representative of three independent experiments for each strain, are shown in Table 1. Alba1 and Alba3 are known to complex together in trypanosomes [21, 25, 53] and this interaction was also revealed in Leishmania both by co-immunoprecipitation (Table 1) and co-localization studies (Fig 2). This explains why most of the co-immunoprecipitated proteins were found to interact with both HA-LiAlba3 and LiAlba1-HA. LiAlba1 and LiAlba3 interact with the poly(A)-binding proteins (PABP1-3) in both parasite lifestages (Table 1 and S3A Fig) as also observed in T. brucei [25], P. berghei [23] and T. gondii [54]. Nevertheless, interaction of LiAlba1 and LiAlba3 with PABPs may be not direct as RNase A treatment prior to immunoprecipitation and Western blot analysis did not detect PABP1 (S3B Fig). LiAlba1 and LiAlba3 interact also with the DEAD-box ATP-dependent RNA helicase LinJ.32.0410, a homolog of the yeast Ded1p implicated in ribosome scanning and the distribution of mRNAs between P-bodies and polysomes [55]. We have previously shown that this RNA helicase protects rRNA from degradation under conditions of stress and cell death [32]. A 25 kDa RNA-binding protein (LinJ.32.0790) with a nucleolar RNA-binding domain specific to trypanosomatids and a ZC3H41 helicase homolog (LinJ.27.1220 [56]) were also found to interact with LiAlba1 and LiAlba3 in both developmental stages. To our knowledge, these proteins have not been studied yet. An NTF2-like protein (for Nuclear Transport Factor 2), LinJ.21.0490, interacts also with both Alba-domain proteins. In other organisms, NTF2 is involved in RanGTP transport and cargo shuttling between the cytoplasm and nucleoplasm (reviewed in [57]), and some of these factors have been studied in T. brucei for their implication in cell death [58]. Ribosomal proteins of the 40S subunit (SA, S3A, S6, S14, S24e) or 60S (L1a, L7a) were also found associated with LiAlba1 and LiAlba3 both in promastigotes and amastigotes, although a higher number of ribosomal proteins was seen in co-immunoprecipitation studies using amastigote extracts (Table 1). Association of LiAlba1 and LiAlba3 with ribosomal subunits (40S and 60S) was also suggested by sucrose gradient fractionation (Fig 3A). In the closely related species T. brucei, Alba proteins were found to co-sediment with polysomal fractions [25], but in the case of Leishmania they were associated predominantly with ribosomal subunits, although a small fraction was also seen in the polysomes (Fig 3A). Interestingly, under conditions of decreased global translation following O/N exposure to heat stress (Fig 3B upper panel and [59]), the enrichment of LiAlba3 in ribosomal subunits was significantly increased (Fig 3B, lower panel), suggesting a role of this protein in translational repression, as also reported in Plasmodium [23]. Higher co-fractionation of HA-LiAlba3 in ribosomal subunits upon heat-stress was not due to changes in protein expression, as recombinant HA-LiAlba3 remained stable between unstressed and heat-stressed parasites (S4C Fig). Also, no changes were observed in HA-LiAlba3 and LiAlba1-HA proteins between promastigotes and differentiating amastigotes (S4A and S4B Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Alba-domain proteins co-localize to the cytoplasm of promastigote and amastigote Leishmania life stages.

Direct fluorescence images of recombinant L. infantum promastigotes (Pro) and axenic amastigotes (Ama) at passage 4 co-expressing eYPF-LiAlba1 (green) and mCh-LiAlba3 (red) proteins. Green and red pixels overlapped in the digital images yielding yellow/orange signals. The nucleus (N) and kinetoplastid DNA (K) were stained with DAPI (blue).

https://doi.org/10.1371/journal.pone.0137243.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Alba-domain proteins are associated with ribosomal subunits.

Polysome fractionation of L. infantum expressing HA-tagged Alba1 and Alba3 proteins by 15–45% sucrose gradient was carried out using logarithmic phase promastigotes (26°C) (A) or heat-stressed parasites grown O/N at 37°C (B). Graphical representations present the RNA content of each collected fraction after ultracentrifugation on 15–45% sucrose gradient. F: Free RNA; 40S, 60S and 80S: ribosomal subunits and monosomes, respectively. Each fraction was loaded on 12% SDS-PAGE and transferred on a nylon membrane for Western blot analysis to detect HA-LiAlba3 and LiAlba1-HA proteins using an anti-HA antibody. As a control, half of the protein extracts were incubated with EDTA before ultracentrifugation to disrupt association of the polyribosomes with mRNAs.

https://doi.org/10.1371/journal.pone.0137243.g003

Proteins found to interact with LiAlba1 and LiAlba3 specifically in amastigotes include DRBD2 (LinJ.35.2240, a possible homolog of yeast Gbp2p involved in mRNA export), a homolog of Rap55 (LinJ.25.0550), a putative vacuolar protein (LinJ.19.0020 [60]), and a second NTF-2 like protein specific of trypanosomatids (LinJ.18.0300). On the other hand, a 44 kDa RNA-binding protein specific to Leishmania (LinJ.18.0590), a seryl-tRNA synthetase, and a dipeptidyl-peptidase III (LinJ.05.0960), mostly found in Leishmania spp. and metazoans, seem to interact with LiAlba1 or LiAlba3, specifically in promastigotes.

Alba-domain proteins are localized to distinct subcellular compartments throughout Leishmania development

Fluorescent protein imaging studies using L. infantum promastigotes and axenic amastigotes expressing either eYFP-LiAlba1 or mCh-LiAlba3 demonstrated that LiAlba1 and LiAlba3 proteins co-localize to the cytoplasm of both major parasite lifestages (Fig 2), as also described in T. brucei [25, 26] and consistent with a role of these proteins in RNA metabolism. In other organisms as plants, Alba-domain proteins are described as stress responsive proteins [61, 62] and in T. brucei they accumulate in starvation granules [25]. Therefore, we investigated the subcellular localization of Alba-domain proteins upon amastigote differentiation triggered in vitro by exposure of promastigotes to combined heat stress (25°C to 37°C) and acidic environment (from neutral to acidic pH). We used the add-back strains previously described [18] where HA-tagged versions of LiAlba1 and LiAlba3 were episomally expressed into the LiAlba1^-/- and LiAlba3^-/- background, respectively. Expression of recombinant HA-LiAlba3 and LiAlba1-HA proteins remained stable between promastigotes and differentiating amastigotes (S4A and S4B Fig). Immunolocalization studies confirmed the cytoplasmic co-localization of LiAlba1-HA and HA-LiAlba3 proteins in promastigotes (Fig 4A), in agreement with data obtained in Fig 2 using fluorescent Alba proteins. The cytoplasmic localization of Alba proteins in promastigotes is not homogenous (Figs 2 and 4A), suggesting that they may associate with subcellular compartments. Digitonin fractionation of L. infantum promastigotes showed that native LiAlba3 (S5A Fig) as well as HA-tagged LiAlba3 (S5B Fig) and LiAlba1 proteins (S5C Fig) were not only enriched in fraction 2 (mainly cytosolic proteins) but also in fraction 3 (enriched for organelle-located proteins) [63]. To rule out the possibility that Alba proteins associate with the endoplasmic reticulum or the mitochondrion, we carried out a co-immunolocalization studies with BiP, an ER chaperone, and MitoTracker staining, respectively. Our results indicate that LiAlba proteins do not associate with the endoplasmic reticulum (S6A Fig) or the mitochondrion (S6B Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Alba-domain proteins translocate from the cytoplasm to the flagellum and the nucleolus upon Leishmania amastigote differentiation.

Subcellular localization of LiAlba1-HA and HA-LiAlba3 proteins in promastigotes (A) and upon amastigote differentiation (8 h in MAA medium pH 5.8 at 37°C) (B) was assessed by indirect immunofluorescence studies using an anti-HA antibody as described in Materials and Methods. DAPI staining (red) allows detection of the nucleus (N) and kinetoplastid DNA (K). C) Immunofluorescence images of wild type L. infantum episomally co-expressing pSP-NEOalphaIR-eYPF-LiAlba1 and pSP-HYGalphaIR-mCh-LiNOP10 grown as promastigotes (Pro), differentiating amastigotes (Diff) and amastigotes (Ama). LiNOP10 was used as a nucleolar (Nu) control.

https://doi.org/10.1371/journal.pone.0137243.g004

Interestingly, upon amastigote differentiation (from 8 h up to overnight exposure to the differentiation signals), the vast majority of Alba1 and Alba3 proteins were not cytoplasmic but were enriched inside the nucleus and all along the flagellum (Fig 4B and data not shown). During differentiation, Alba proteins were also enriched in fraction 5 corresponding to insoluble or membrane-associated material as shown by digitonin fractionation and Western blot experiments (S5 Fig, right panels). Within the nucleus, LiAlba proteins accumulate in a region with a weaker DAPI staining (data not shown), usually corresponding to the nucleolus as described previously [64, 65]. We therefore investigated whether Alba proteins were localized to the nucleolus. As a control, we used the Nop10 protein (S7 Fig), previously characterized as a nucleolar component in T. brucei [66]. To confirm nucleolar localization of Alba proteins upon differentiation, we generated a recombinant strain episomally co-expressing eYFP-LiAlba1 and mCh-LiNop10 proteins in the wild type background. A co-localization signal with LiNop10 confirmed the presence of LiAlba proteins in the nucleolus (Fig 4C). It is noticeable that during differentiation the LiNop10 signal is stronger and this could correlate with nucleolar rearrangements during stress in eukaryotes (reviewed in [67, 68]). Even if some parasites show only nucleolar or only flagellar localization of Alba proteins, it is experimentally difficult to obtain a clear chronological order of these events. The flagellar and nucleolar localization of Alba proteins is detected only upon amastigote differentiation and the proteins are relocalized back to the cytoplasm in fully differentiated amastigotes (Fig 2). Differential localization of Alba proteins is a regulated process as when differentiating amastigotes were switched back to promastigote growth conditions, the nucleolar and flagellar signals were lost, and Alba proteins were seen again predominantly in the cytoplasm (S8 Fig).

To determine if LiAlba proteins form different complexes upon amastigote differentiation, hence explaining their differential localization, we carried out co-immunoprecipitation experiments with parasites grown O/N under conditions of amastigote differentiation in vitro. These experiments did not reveal any new Alba protein associations, which could explain translocation of Alba proteins to the nucleolus and the flagellum (data not shown).

Heat-shock triggers translocation of Alba-domain proteins from the cytoplasm to the nucleolus and the flagellum

To investigate further the differentiation signal(s) necessary for Alba protein translocation into the nucleolus and the flagellum, we exposed L. infantum promastigotes to either temperature stress (from 25° to 37°C in SDM medium pH 7.0 O/N) or acidic pH (pH 5.8 in HCl acidified SDM medium at 25°C O/N). Temperature stress and acidic pH represent the most important triggering factors of Leishmania differentiation into amastigote forms inside macrophages [4]. Temperature stress was sufficient for LiAlba1 protein translocation into the nucleolus and the flagellum (Fig 5, upper panel). Similar results were obtained with LiAlba3 (data not shown). In contrast, pH stress did not have any effect on Alba protein differential localization (Fig 5, lower panel).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Heat stress triggers differential localization of Alba-domain proteins in Leishmania.

Immunofluorescence images for eYPF-LiAlba1 (green) and mCh-LiNOP10 (red) in L. infantum promastigotes co-expressing pSP-NEOalphaIR-eYPF-LiAlba1 and pSP-HYGalphaIR-mCh-LiNOP10 submitted to heat stress (from 25°C to 37°C) or to acidic pH (pH 5.8) O/N. Nu: nucleolus.

https://doi.org/10.1371/journal.pone.0137243.g005

In a time course experiment, we evaluated the kinetics of Alba protein flagellar localization upon exposure to temperature stress using microscopic examination. The LiAlba3^-/- knockout strain episomally expressing mCh-Alba3 was used for these experiments. Parasites with Alba3 localized to the flagellum were observed 2 hours following temperature stress and after 8 hours more than 26% of the total population showed this pattern (S9 Fig). Exposure of parasites to temperature stress can induce, under certain conditions, parasite killing [69]. Therefore, we investigated whether Alba protein localization in the nucleolus or the flagellum is associated with dying cells. Using flow cytometry, we compared the percentage of cells having a flagellar signal with the percentage of cells in SubG1 population, representative of dead cells with nuclei containing degraded DNA. During the 8 h period of temperature stress, the proportion of SubG1 population remained stable ranging from 2% to 2.5% (S9 Fig), a value clearly inferior to the 26% of parasites showing Alba3 flagellar localization (S9 Fig). Thus, these data exclude any correlation between flagellar localization of LiAlba3 and cell death.

Inhibition of transcription using Actinomycin D is another type of stress previously reported in Leishmania and also in T. cruzi to induce nucleolar accumulation of cytoplasmic RNA-binding proteins [65]. We therefore tested whether transcriptional stress induces Alba protein accumulation in the nucleolus but we did not observe any of this upon Actinomycin D treatment (data not shown).

Partial characterization of the flagellar proteome of Leishmania promastigotes and differentiating amastigotes

The presence of RNA-binding proteins such as Alba proteins in the Leishmania flagellum is an intriguing finding. We therefore carried out additional experiments to confirm by other means the presence of Alba proteins in the flagellum and to examine if other RNA-binding proteins were also present in this organelle. For this purpose, we adapted a recently published protocol in T. brucei [30] to purify intact flagella from wild type L. infantum promastigotes and parasites undergoing amastigote differentiation (8 h exposure to differentiation signals) (Fig 6A and 6B) and analyzed this material by mass spectrometry. S3 Table lists all the components of the Leishmania flagellum proteome of promastigotes and also of differentiating amastigotes identified in this study and S4 Table provides the detailed peptide and spectrum reports of these experiments. Ninety-eight genes coding for various proteins previously described to be associated with the flagellum were identified here, including components of the paraflagellar rod (15 proteins) and the axoneme such as radial spoke proteins (8 proteins), components of the dynein arm (7 proteins) and other proteins (17 proteins) conserved in species with motile flagella (CMF, components of motile flagella) [70]. In addition, proteins normally found in the membrane or matrix fraction of flagella were detected such as intraflagellar transport proteins [71], SMP-1 [72], FLAM7 and 14-3-3 protein [30] or flabarin [73] (Fig 6C, Tables 2 and S3). Seventy-five proteins involved in RNA metabolism and translation, 108 proteins participating in mitochondrial function and metabolic processes, proteins involved in protein folding (19 proteins), post-translational modifications (11 proteins), protein degradation (32 proteins), and intracellular transport (13 proteins) were also identified as part of the Leishmania flagellum proteome (Fig 6C, Tables 2 and S3). Table 2 summarizes the best-characterized proteins belonging to flagellar structures and RNA metabolism found in our experiments. Leishmania shares 57% of the identified proteins in common with the flagellum proteome of T. brucei procyclics characterized using a similar strategy [30]. The most important differences between these two related parasitic protozoa include components of the proteasome machinery seen only our Leishmania flagellar fraction (17 in total) and 18 proteins that are specific to Leishmania.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. List of the most representative components of the flagellar structure and RNA metabolism enriched in the L. infantum flagellum proteome.

https://doi.org/10.1371/journal.pone.0137243.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Flagellum purification demonstrates an enrichment of LiAlba3 in the Leishmania flagellum during heat stress.

Phase contrast images of intact L. infantum promastigotes (A) and of purified flagella after sucrose gradient isolation (B) as described in Materials and Methods. (C) Summary of MS/MS identified proteins from four independent experiments of flagellum purification (two from promastigote cell extracts and two from 8 h-differentiating amastigotes). Identified genes were classified according to their gene ontology and to characterized orthologs in Trypanosoma spp. based on GeneDB and TriTrypDB gene annotations. Only genes identified at least twice with a minimum of 2 peptides are shown here (see Tables 2, S3 and S4 for the complete list of the identified L. infantum flagellum proteins). (D) Confirmation of flagellar localization of PFR2C-HA in recombinant L. infantum expressing pSP-alphaIRNEOalphaIR-PFR2C-HA. (E) Western blot analysis and quantification of endogenous LiAlba3 in purified flagellum fractions upon promastigote conditions of growth (Pro) or following 8 h of temperature stress (37°C). After flagellum purification, flagella from promastigotes and heat-stressed parasites were counted on a Malassey (hemocytometer) to load an equivalent number of flagella on the gel. Total proteins of promastigote cells were loaded as a control. Relative quantification was performed using ImageJ blot.

https://doi.org/10.1371/journal.pone.0137243.g006

Several RNA-binding proteins were found in the Leishmania flagellar fraction, including several ATP-dependent RNA helicases, translation initiation factors, and zinc-finger domain proteins (Tables 2 and S3). Only a single peptide for LiAlba1 was identified under differentiation in two immunoprecipitation experiments (data not shown). This result could be explained if the amount of Alba proteins in the flagellum was too low relative to the major skeletal components such as those in the PFR or the axoneme. However, two amastin surface proteins were found in the flagellum of differentiating amastigotes (S3 Table). We have shown recently that Alba-domain proteins regulate the stability of amastin transcripts [18].

To further confirm the presence of Alba-domain proteins in the flagellar fraction, the experiment was repeated with extracts from promastigotes and parasites exposed for 8 hours to heat stress. As a control for flagellar localization, we used recombinant parasites episomally expressing the paraflagellar rod gene PFR2C tagged with an HA epitope at the C-terminus (PFR2C-HA). Specific localization of this protein in the flagellum was shown by immunofluorescence (Fig 6D). After flagellum purification, the samples were used for Western blot analysis against PFR2C-HA (control) and endogenous Alba3 antibodies in parallel (Fig 6E). This purification strategy indicated that LiAlba3 is enriched by 2.2-fold in the flagellar fraction of L. infantum parasites exposed to heat stress (Fig 6E), hence confirming the immunofluorescence data in Figs 4B and 5.

Interesting differences were observed in the composition of the flagellar fraction proteome between L. infantum promastigotes and differentiating amastigotes. For example, the flagellum of parasites undergoing differentiation lacks 2 paraflagellar rod components, 5 dyneins, 1 radial spoke protein (RSP11), 2 flagellar transport proteins, and few (10) other proteins associated with flagellum functions. There is also a large number of hypothetical proteins and most of the hypothetical Leishmania-specific proteins (48 in total) that were not detected in the flagellum of differentiating amastigotes (Fig 6C and S3 Table). On the other hand, several proteins involved in protein degradation, protein folding, intracellular protein transport, GTP/ATP binding, cell-redox homeostasis, and oxidation-reduction processes were mostly detected in the flagellum of differentiating parasites (S3 Table). In total, 79 proteins were enriched in the promastigote flagellum and 122 in that of differentiating amastigotes (Fig 6C). Although the length of the flagellum between promastigotes and differentiating parasites was relatively similar (data not shown) and co-immunoprecipitation studies were carried out under the same conditions, additional quantitative proteomic studies are needed to conclude about differences in flagellum composition between promastigotes and differentiating amastigotes.