Using data mining via the OmniBLAST server we searched the whole L. major genome database http://www.ebi.ac.uk/parasites/leish.html for a putative sequence that shared similarities with the vertebrate TRF1 and TRF2 proteins. For this search, we used the most conserved part of both human proteins, the C-terminal fragment containing the Myb-like DNA binding domain. The search returned a single sequence (GenBank acc. no. XP_001682531.1) that encoded a hypothetical protein (GenBank acc. no. Q4QDR7, GeneDB_Lmajor LmjF18.1250), the C-terminus of which shared ~30% identity and 50-55% similarity with the vertebrate TRF Myb-like domain, according to the blast2 sequence analysis (Table 1).

Based on the L. major sequence, primers were designed for PCR amplification of the entire homologous sequence from L. amazonensis with genomic DNA as the template. PCR products of 2,931 bp were cloned into the vector pCR2.1 and both insert strands were sequenced (data not shown). The deduced polypeptide sequence of 796 amino acid residues contained a putative C-terminal Myb-like DNA binding domain between residues 684-733, according to psi-blast (Fig 1 - top). The LaTRF gene (GenBank acc. no. EF559263) shared high sequence identity and similarity to the putative L. major TRF, and to hypothetical L. infantum and L. braziliensis TRFs (Table 1). The sequence conservation between LaTRF and the trypanosome TbTRF and the putative TcTRF homologues decreased to 35-45% identity (Table 1), consistent with the known evolutionary relationships among these organisms. The Leishmania TRF homologues encode the largest TRF protein (~82.5 kDa) described so far. The fact that the Leishmania proteins showed much greater homology with each other than with other protozoan proteins and that they are the largest TRF described so far resembles the situation for Leishmania telomerase protein [25].

In addition, like TbTRF, LaTRF shared sequence similarities with the canonical Myb-like domain and with the TRFH dimerization domain of human TRF1 and TRF2 (Fig 1-bottom and Table 1), but no sequence similarities were found with any other telobox protein (data not shown). Together, these results indicate that although LaTRF shares high sequence similarity with TbTRF, probably because the two species are phylogenetically related [26], further studies are required to confer any functions to the Leishmania TRF homologue identified here.

In exponentially growing L. amazonensis promastigotes, LaTRF was detected only in nuclear protein extracts. A single ~82.5 kDa protein band was detected using anti-LaTRF serum (Fig 2 - top panel: lane 1). No protein was detected in cytoplasmic and total protein extracts (Fig 2 - top panel: lanes 2 and 3), indicating that LaTRF is a nuclear protein with very low intracellular abundance. As a control, Western blots were revealed with anti-LaRPA-1 serum, which recognizes a ~51.2 kDa telomeric protein band [23] (Fig 2 - bottom panel: lane 1) and also its phosphorylated forms (Fig 2 - bottom panel: lane 2; da Silveira & Cano, unpublished data).

We also developed an immunofluorescence assay combined with FISH, using anti-LaTRF serum and a PNA-telomere probe specific for TTAGGG repeats. As shown in Fig 3 (panels p1-4, merged images a and b), LaTRF is a nuclear protein that partially co-localizes with parasites telomeres, since some of the LaTRF signal coincided with telomeric foci and some did not (Fig 3, panels p1-4). In most cells, LaTRF appears as a diffuse signal spread all over the nucleoplasm and only in some cases it forms large punctuated foci, which seems to co-localize with the telomeric DNA (yellow dots in Fig 3, panels p2 and p4). Similarly, in humans, the hTRF2 protein also appears in the form of punctuate foci that does not completely associate with telomeres [18], which is in agreement with other cellular functions played by this protein [27,28]. In contrast, the T. brucei TRF protein (TbTRF) appears to co-localize with most telomeres at all stages of the cell cycle in both bloodstream and procyclic forms [24]. Whether LaTRF also has other cellular roles or if its association with telomeres occurs in a cell cycle dependent manner is not clear at this stage.

EMSA assays were done with renatured protein extracts containing full length LaTRF, the Myb-like DNA binding domain (LaTRF^Myb) (Figs 4 and 5, see additional file 1) and with L. amazonensis nuclear extracts (Fig 6), to investigate whether LaTRF, like its vertebrate and trypanosome counterparts [18,24], was able to bind double-stranded telomeric DNA in vitro.

The full-length recombinant protein and its deletion mutant were expressed in very low amounts and in non-soluble form in the E. coli system (data not shown) making their purification by conventional chromatography very difficult. Therefore, protein expression was checked by Western blot using anti-LaTRF serum and anti-His tag monoclonal antibody (data not shown).

As shown in Fig 4, recombinant full length LaTRF and the mutant bearing only the C-terminal Myb-domain were able to bind specifically the double-stranded telomeric DNA (LaTEL). Competition assays showed that the complexes formed by both recombinant proteins were completely abolished in the presence of excess unlabeled LaTEL and that there was no competition for binding when excess of non-specific poly [dI-dC] [dI-dC] double-stranded DNA was used (Fig 4, lanes 4, 5, 8 and 9). Supershift assay with anti-LaTRF serum, which recognizes a N-terminal epitope in the protein, confirmed that full length LaTRF forms a robust complex with labeled LaTEL (Fig 4, lane 6), possibly because the binding of anti-LaTRF stabilized the LaTRF-LaTEL complex, blocking the action of other non-specific binding activity in the extract. When competitors were added to the supershift reactions with anti-LaTRF serum, the binding specificity of recombinant LaTRF for LaTEL was confirmed (Fig 5, lanes 2-4). The complex was almost totally abolished in the presence of excess unlabeled LaTEL (Fig 5, lane 3) and no competition was detected in the presence of non-specific DNA (Fig 5, lane 4). The results presented above suggest that recombinant LaTRF binds LaTEL potentially via the putative Myb-like DNA binding domain indicating a role for the C-terminal region of LaTRF in mediating sequence-specific binding to telomeric DNA.

Nuclear extracts were obtained from log phase L. amazonensis promastigotes in order to check if native LaTRF was also able to bind double-stranded telomeric DNA (LaTEL) in vitro, (Fig 6). The results showed the presence of LaTRF activity in these extracts, as part of the complex formed with the nuclear proteins and LaTEL (Fig 6, lane 2) was supershifted by the anti-LaTRF serum (Fig 6, lane 6). In addition, competition assays showed that this complex was unaffected by excess of poly [dI-dC] [dI-dC] (Fig 6, lane 3), used as the non-specific competitor, but it was almost completely abolished in the presence of excess unlabeled LaTEL (Fig 6, lane 4). Supershift experiments using anti-LaTRF serum were done in the presence of competitor to confirm that LaTRF was actually involved in the formation of the retarded band (Fig 6, lane 5). Note that the retarded shifted band disappeared due to the competition by non-labeled LaTEL. Thus, these results indicate that LaTRF is in part responsible for the binding activity shown in these extracts and is probably a component of the Leishmania telomeric complex.

Chromatin immunoprecipitation experiments also suggested that LaTRF is a telomeric protein. The anti-LaTRF serum immunoprecipitated L. amazonensis telomeric DNA (Tel1) in vivo (Fig 7 - left) but did not immunoprecipitate the GT-rich kinetoplast DNA (kDNA) (Fig 7 - right). The kDNA control represented by the UMS (universal mini-circle sequence) albeit GT-rich, is very representative of the general base composition of Leishmania genomic DNA. In addition, it is a good control, since we were able to show that it was co-immunoprecipitated by two other Leishmania telomeric protein [17,23]. In a previous study, we described LaTBP1, a protein that specifically binds telomeric and GT-rich DNA in Leishmania. LaTBP1 has a centrally positioned Myb-like DNA binding domain and is most likely a non-telobox protein that is apparently related to the multifunctional yeast RAP1 telomeric protein and TFIIIB B" transcription factor [17]. Together with the putative LaTRF described here, these are the only descriptions of proteins bearing a Myb-like DNA binding domain that interact with double-stranded telomeric DNA in Leishmania.

As mentioned here and elsewhere [26], the huge evolutionary distance between this protozoan and higher eukaryotes presents a barrier when searching for protein homologues in the genomes of these parasites. For example, no TRF1 homologues were found in trypanosomatid genomes but the expression of hTRF1 in procyclic forms of T. brucei caused telomere shortening and cell cycle arrest, probably by displacing an unknown endogenous telomeric factor [29]. RNAi knockdown of TbTRF arrested bloodstream cells in G2/M and most of the procyclic cells in the S phase, and caused shortening of G-rich single-stranded telomeric DNA. These findings suggest that TbTRF is probably the unknown endogenous telomeric factor, which resembles the function of mammalian TRF2 at parasite telomeres [24]. The functions of LaTRF at Leishmania telomeres remain to be determined.

L. amazonensis promastigotes (MHOM/BR/73/M2269) were grown in M199 medium (Cultilab) supplemented with 10% fetal calf serum (Cultilab), 25 mM HEPES and 1 × antibiotic/antimycotic solution (Cultilab) at 28°C.

Total genomic DNA of L. amazonensis was prepared as previously described [31]. LaTRF was cloned using a PCR-based strategy. Primers were designed based on the putative sequence LM16.2.Contig67 from L. major (GeneDB_Lmajor LmjF18.1250) for amplification of the complete LaTRF open reading frame (ORF) (See additional file 2: Table S1). The PCR product spanning the entire L. amazonensis TRF ORF (2,391 bp) was obtained by using the primers F1 and R1 and 1U of Platinum Taq (Invitrogen) followed by cloning into the pCR^® 2.1 cloning vector (Invitrogen). The PCR product was sequenced using specific primers and primers from the vector (See additional file 2: Table S1). The primers F1 and R1 contained restriction sites for NdeI and XhoI (See additional file 2: Table S1) to allow further cloning of the gene in-frame with a N-terminal 6× His-tag into plasmid pET-28a+ (Novagen).

Amino acid sequence alignments were done with blastp, bl2seq, cds http://blast.ncbi.nlm.nih.gov/Blast.cgi and ClustalW http://www.ebi.ac.uk/clustalw/ using default parameters. The sequences used for these analyses were: hTRF2 (GenBank acc. no. Q15554), hTRF1 (GenBank acc. no. P54274.2), TbTRF (GenBank acc. no. AY910010), putative LmTRF (TrEMBL acc. no. Q4QDR7, GeneDB_Lmajor LmjF18.1250), TcTRF (GenBank acc. no. XP_819954.1), LiTRF (GenBank acc. no. XP_001464939.1) and LbTRF (GenBank acc. no. XP_001564056.1). The L. amazonensis LaTRF gene sequence was submitted to GenBank and is available under the accession number EF559263.

To verify the existence of a Myb-like DNA-binding domain at the C-terminus of the protein, a deletion mutant was constructed. The primers F3 and R1 (See additional file 2: Table S1) were used to amplify the deletion mutant LaTRF^Myb from genomic DNA, which contained the putative C-terminal Myb-like DNA binding domain. This mutant has approximately 665 bp that span nt 1726-2391. As with full length LaTRF, the LaTRF^Myb mutant was cloned into the pCR^® 2.1 cloning vector (Invitrogen), sequenced and subcloned into a pET28a+ expression vector.

Full length LaTRF and the deletion mutant LaTRF^Mybcloned into a pET 28a+ vector, were transformed in E. coli strain BL21 DE3 RP codon plus cells for expression in the presence of 1 mM IPTG. Both proteins were expressed in low amounts and in non-soluble form, preventing them from being purified by affinity chromatography based on the 6× His-tag. To overcome this problem, the non-soluble bacterial pellets containing both proteins were solubilized in 7 M urea, sonicated in the presence of 10 U of DNAse I (Sigma) and renatured by dialysis in 50 mM glycine, pH 8.0. The presence of each protein in the extracts was checked by electrophoresis in 10% SDS-PAGE followed by Western blot probed with anti-LaTRF serum and with an anti-His tag monoclonal antibody (Novagen).

Promastigotes in mid-exponential growth were used to obtain both extracts. Nuclear and cytoplasmic extracts were prepared with a Nuclear Extract Kit (Active Motif) adapted for L. amazonensis promastigotes in the presence of phosphatase and protease inhibitors.

Total protein extracts were obtained using RIPA buffer (150 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% Triton X-100 and 0.1% SDS) in the presence of 10 U of DNase I and 1X protease inhibitor cocktail (Calbiochem) and incubated for 15 min at 4°C. Cell lysates were homogenized by vortexing at maximum speed (5 bursts of 10 s each). Extracts were cleared by centrifugation at 9,300 ×g for 8 min at 4°C, to separate the total protein (supernatant) from the cellular debris (pellet).

Both extracts were stored at -80°C and their protein concentrations were measured by the Bradford dye-binding assay, using bovine serum albumin as standard.

Different protein extracts obtained from 10⁷ parasites were separated by SDS-PAGE on 10% polyacrylamide gels and transferred to nitrocellulose membranes (BIO-RAD) in Tris-glycine-methanol at 16°C. The membranes were probed with rabbit anti-TRF2 serum raised against the synthetic peptide Nt-APAVTTRKRPRSSDSP-Ct (Sigma). The extracts were also probed with anti-LaRPA-1 serum as a control [23,32]. In both cases, immunoreactive bands were revealed by using an Amplified Alkaline Phosphatase Immun-Blot Assay Kit, according to the manufacturer's instructions (BIO-RAD).

This assay was performed using previously described protocols [32,33] with minor modifications. For IF, 10⁶ promastigotes cells were washed in 1X PBS and fixed in 1% formaldehyde for 5 min at 4°C. After permeabilization with 0.1% Triton X-100 (in 1X PBS) for 10 min at room temperature, cells were incubated with 0.1 M Glycine (in 1X PBS) and attached to glass coverslips coated with 0.1% poly-L-Lysine (Sigma). Anti-LaTRF serum was used to detect LaTRF with Alexa Fluor 555-labeled goat anti-rabbit IgG (Invitrogen) as the secondary antibody followed by telomere detection using a Telomere PNA FISH Kit/FITC (DakoCytomation). VECTASHIELD^® Mounting Medium with DAPI (Vector Labs) was used as the anti-fade mounting solution and to stain nuclear and kinetoplast DNA. The images were analyzed with a Nikon 80i fluorescence microscope and captured with a digital camera (Nikon). When necessary, images were superimposed using NIS elements software (v. Br 2.30).

All of the conditions for binding reactions and EMSA, including binding temperature, protein concentrations in the extracts and the double-stranded DNA probe (LaTEL), were standardized in preliminary experiments. LaTEL was constructed by using the γ [³²P]ATP 5'-end-labeled oligonucleotides ssTel78G and ssTel78C, as described by Lira et al. [17]. Assays were done by mixing 10 μg of renatured bacterial extracts containing full length LaTRF or LaTRF^Myb with approximately 2 pmol of labeled probe (LaTEL) in 30 μl of EMSA buffer (20 mM HEPES, 2.5 mM MgCl_2, 0.1 mM EDTA, 0.1 M KCl, 10% glycerol, 0.5 mM DTT, pH 8.0) containing 10 ng of poly [dI-dC] [dI-dC] and 10 ng of poly [dA-dT] [dA-dT]. Total protein extracts of non-transformed E. coli were used as controls. The reactions were incubated for 30 min at room temperature and loaded onto a non-denaturing 4% polyacrylamide gel (acrylamide:bis-acrylamide, 19:1, w/w) in 1X TBE. After electrophoresis, the gels were exposed to X-ray film.

Binding reactions were also done with crude nuclear extracts obtained from 10⁸ parasites (~2.3 μg of total proteins) and γ [³²P]ATP labeled LaTEL (2 pmol) in EMSA buffer containing a mixture of 10 ng of poly [dI-dC] [dI-dC] and 10 ng of poly [dA-dT] [dA-dT].

Competition assays to test the binding specificity of proteins in both recombinant and nuclear extracts, were done using 20 fold excess of unlabeled LaTEL (in relation to the labeled probe) as the specific competitor and a 100 fold excess (in relation to the labeled probe) of unlabeled double-stranded DNA poly [dI-dC] [dI-dC] as the non-specific competitor. Supershift assays were done using full-length recombinant LaTRF (10 μg) or native nuclear extracts from 10⁸ parasites in the presence of ~30 μg of anti-LaTRF serum in EMSA buffer containing labeled LaTEL as probe and both poly [dI-dC] [dI-dC] and poly [dA-dT] [dA-dT] as above described. These assays were also performed in the presence of 20 fold excess of non-labeled LaTEL and 100 fold excess of poly [dI-dC] [dI-dC] as described above.

Formaldehyde cross-linked chromatin was obtained from promastigote forms of L. amazonensis parasites (0.8 × 10⁸ cells/experiment) as described by Lira et al. [17]. Chromatin was immunoprecipitated with anti-LaTRF serum and DNA was extracted after cross-link reversal. DNA samples were slot-blotted and hybridized with Tel1 and kDNA probes by using a previously established protocol. Aliquots of 1% and 10% of total DNA used in each experiment (input) were tested separately. Control assays included immunoprecipitation of chromatin with pre-immuneserum (pre-immune) or without serum (mock). The probes used were 5'-end labeled with γATP [³²P]: Tel1 (5'TTAGGG-3')₃ and kDNA (5'-TTTCGGCTCGGGCGGTGAAAACTGGGGGTTGGTGTAAAAT-3'), according to Lira et al. [17].