The META1 sequence is highly conserved across the Leishmania genus 16 . Based on this high degree of conservation and the convenient accessibility of L. major genome sequence and other related resources over those for other Leishmania species, all of our bioinformatics analyses have used L. major META1 as the representative species. We began our study by using BLASTP to search for homologs of META1 protein against the NCBI NRDB. Homologs of META1 were found in all other Leishmania species and Trypanosomes (Additional file 1).

Since BLASTP search failed to detect any functionally characterized homolog within or outside the Trypanosomatidae family, a PSI-BLAST search was performed against NCBI NRDB. This search yielded hits beyond the trypanosomatids as early as the second iteration. Interestingly, these significant hits were all homologs of the heat shock inducible protein HslJ from bacteria. Table 1 lists representative hits after 3 iterations. All sequence homologs of META1 (trypanosomatid as well as bacterial) are characterized by the presence of at least one META domain. Although there is a low direct pairwise sequence identity of 25% between META1 and HslJ (Figure 1 ), bacterial HslJ is currently the closest relative of Leishmania META1 outside the family Trypanosomatidae. The most likely explanation for this unusual phyletic relationship is that META1 in trypanosomatids originated as a result of a lateral gene transfer (LGT) event from bacteria.

Most genomes have distinct overall GC content and the genes in their genome are fairly similar in base composition and patterns of codon usage. A laterally transferred gene continues to carry imprints of its origin and ancestry whilst residing in its current recipient genome. Such 'foreign genes' can be identified by their unusual sequence characteristics vis a vis the recipient genome 19 20 21 . GC content at first and third codon positions and synonymous codon usage are generally used for such analysis. ORFs are identified as atypical if their GC contents at first and third codon positions (GC₁ & GC₃ respectively) are two or more standard deviations (SDs) higher or lower than the respective means for all genes in that genome. Our analysis revealed that the total GC content of L. major META1 is 54%, which is 6.3 SD lower than the average of the total GC content of all genes in L. major genome. In addition, LmjMETA1 has a GC₁ content of 48.67% (15.8 SD lower than GC₁ of all genes in L. major genome) and a GC₃ content of 77% (within the range of SD for GC₃ of L. major genome), as summarized in Figure 2A. GC content anomaly results establish that META1 nucleotide composition is distinct from the overall Leishmania genome and therefore, is possibly of foreign origin.

Codon adaptation index (CAI), on the other hand, gives an account of similarity of synonymous codon usage to that of a standard set of highly expressed genes for that organism. Thus, for a native gene CAI is high and for a gene of exogenous origin CAI value is low. As part of CAI determination, we analyzed a set of other genes for reference along with META1 (Additional file 2 ). The genes that were included: Amastin & GP63 (specific to trypanosomatids), Hexokinase & Ribosomal Protein S8 (present in all taxa), α-Tubulin (a pan-eukaryotic structural protein) and PTR-1 & Coproporphyrinogen, previously predicted to be laterally transferred in L. major from bacteria 22 . META1, PTR-1 and Coproporphyrinogen all have low CAI values when compared to the other reference genes (Figure 2B ), suggesting again, that META1 is of exogenous origin. We also carried out a similar analysis on META1 homologs in T. cruzi (Additional file 3 ); where once again the CAI of the META homologs is lower than that of other Trypanosoma genes (RP S8 and α-Tubulin). These analyses, and the striking frequency of bacterial HslJ hits, point to bacterial hslJ being the exogenous source of META1 into trypanosomatids.

In order to study the phylogenetic linkage between META1 and HslJ, we generated a phylogenetic tree based on the protein sequences of META1 homologs in trypanosomatids and HslJ homologs in bacteria. A multiple sequence alignment containing 30 sequences (16 trypanosomatid and 14 bacterial) was used to create a phylogenetic tree using PhyML. Topology of the phylogenetic tree and the bootstrap values of the branch points (Figure 3 ), again indicate the occurrence of a LGT event of META/hslJ from bacteria to a trypanosomatid.

Estimation of synonymous (K_s) and non-synonymous substitution (K_a) rates is an important tool for understanding molecular sequence evolution and the ratio of non-synonymous to synonymous rate of substitution (K_a/K_s) reflects the selection pressure on a gene 23 . We observed that META1 and the META domain in bacterial hslJ homologs are under strong purifying selection as indicated by the K_a/K_s values < 1 in a pairwise comparison of LmjMETA1 and META domain of 14 bacterial hslJ homologs (Figure 4). Therefore, lower substitution rates for META1 emphasizes the selection pressure on it, a constraint on the Leishmania genome not to change it rapidly and thus, the importance of conservation of this gene.

hslJ transcript is up-regulated in more pathogenic strains of E. coli 24 and the protein is heat inducible 25 . To delineate similarities between META1 and hslJ, we investigated the expression pattern of META1. META1 transcript levels were determined by QRT-PCR at different developmental stages of virulent and attenuated lines of L. donovani. META1 transcript is up-regulated from log to stationary phase promastigotes as well as from promastigote to amastigote stage (Figure 5A ). META1 protein in L. donovani is maximally expressed in amastigotes (Figure 5B ) even though META1 transcript levels in amastigotes are lower than stationary phase promastigotes. META1 protein is induced within 24 hours of shift to amastigote culture conditions, increases further upto 48 hours and its expression decreases within 24 hours of reversal back to promastigote conditions (Figure 5C ). The most striking observation is that attenuated Leishmania have reduced META1 expression under all conditions, as compared to virulent cells (Figure 5).

Acidic pH and increased temperature (shift from 26°C to 37°C) are very important cues for the in vivo as well as in vitro conversion of promastigotes to amastigotes 26 27 28 . Since we found that META1 is up-regulated in amastigotes stage, we investigated the role of these two parameters individually. We found that while both acidic pH and increased temperature induce META1 expression, increased temperature has a greater effect (Figure 5D ). Levels of META1 on western blots have been quantified as shown in Additional file 4.

In the absence of a solved structure of Leishmania META1 protein, we used 3D-JURY 29 structure prediction to look for structural homologs of META1. Consistent with our earlier searches, we found the top scoring model for META1 was based on E. coli HslJ. HslJ structure has 10 β-strands, which resembles a β-barrel structure except of the discontinuity caused due to the presence of the α-helix. We also generated a structural model by homology modeling using MODELLER 9v2 30 for Leishmania META1 using E. coli HslJ (PDB code 2kts) as the template. The final model (Figure 6A) comprises of complete META1 sequence (1-112) threaded on known NMR structure of E. coli HslJ (PDB code 2kts) for residues 25-140 (Figure 6B). The two proteins share a similar structural fold given that 107 C_α atoms of total 116 C_α atoms across the length of the proteins could be superposed with a root mean square deviation (RMSD) value of 1.2 Å and a sequence identity of ~ 28%. The striking structural similarity underscores that despite low sequence identity, META1 and HslJ have been under evolutionary constraints to retain a similar fold.

Since little is known about E.coli HslJ's function with respect to its structure, we decided to look for additional proteins with similar fold. In DALI server 31 , we found a structural homolog of HslJ: the Shigella flexneri protein MxiM (Figure 6A and Additional file 5 ). Despite a low sequence identity between HslJ and MxiM of only ~13%, 94 C_α atoms of the total 116 C_α atoms could be superposed with an RMSD value of 2.5 Å, as determined by DALI pairwise comparison 32 . Similarly, META1 and MxiM share a sequence identity of ~9%; yet, 94 C_α atoms out of the total 115 C_α atoms could be superposed with an RMSD of 2.6 Å. MxiM is a type III secretion apparatus (TTSS) pilot protein, which is a 142-residue lipoprotein, essential for the assembly and membrane association of the Shigella secretin, MxiD 33 . A deficiency in MxiM leads to complete loss of TTSS function and virulence 34 . The structure of MxiM contains a 'cracked barrel' motif that creates a striking cleft in the centre of the protein. The cavity is entirely hydrophobic in nature and binds to lipid moieties of outer membrane (OM) of the bacteria 35 . Likewise, we observed that META1 and HslJ also have a similar hydrophobic cavity upon their superposition to MxiM (Figure 6C ). However, unlike in MxiM, the putative hydrophobic cavity in META1 has at least one charged residue in its core (Figure 6B ). Overall, structural comparisons between META1, HslJ and MxiM further reinforce their fold similarity. This pattern of structural and not sequence conservation signifies selection pressure on the structure of these proteins, probably for maintaining similar functions.

In L. major, META1 is reported to localize around the flagellar pocket 16 , the site for exo- and endocytosis in Leishmania 36 . MxiM, in addition to its known participation in Shigella TTSS, is also known to be structurally similar to the bacterial lipocalin 35 . Lipocalins are known extracellular transporters that bind small hydrophobic molecules 37 . Against this background of information, the striking structural similarity between META1 and MxiM prompted us to examine whether Leishmania META1 had a role in Leishmania secretory processes. Secreted acid phosphatase (SAP) is the most abundant secreted glycoprotein of L. donovani 38 39 40 and estimation of extracellular SAP activity is routinely used as an indicator of the status of secretory processes in the parasite 41 42 43 . Therefore, SAP activity was assayed in culture supernatants in the context of different levels of META1. Extracellular SAP activity was found to be lower at both log and stationary phase in L. donovani virulent promastigotes that have more META1 protein, when compared to attenuated promastigotes (Figure 7). In order to determine whether extracellular SAP activity is directly related with levels of META1 expression, we ectopically expressed META1 in both virulent and attenuated L. donovani promastigotes. The ectopic copy of L. donovani META1 protein was tagged with GFP at the C-terminus to distinguish from the endogenous protein. On META1 overexpression, extracellular SAP activity further reduced in both virulent and attenuated overexpression lines in comparison to their respective wild-types (Figure 7 ). Overexpression of META1 at protein level and its effect on growth kinetics is shown in Additional file 6. This suggests that cells having higher amounts of META1 protein have lower extracellular SAP activity, indicating a role for META1 in secretory processes in Leishmania.

To examine whether the predicted hydrophobic cavity has any role in META1's function, we decided to mutate smaller hydrophobic residues in the cavity to bulkier ones, wherein we replaced leucines at 58^th and 80^th position with a phenylalanine individually and both simultaneously (L58F or/and L80F). L58 is located at the entry of the cavity and is conserved across trypanosomatids and bacterial HslJ. On the other hand, L80 is inside the hydrophobic cavity and is conserved in trypanosomatids, with the exception of L. braziliensis and T. cruzi as shown in Additional file 7. Mutant proteins were C-terminally tagged with GFP, in order to distinguish from endogenous protein and were ectopically expressed in L. donovani promastigotes.

Extracellular SAP activity increased by 40% in L. donovani overexpressing L58F mutant META1 protein, as compared to cells overexpressing the wild-type (WT) META1 (Figure 8A ). However, L80F mutation in META1 decreased the extracellular SAP activity with respect to WT overexpression. Moreover, extracellular SAP activity of the double mutant L58,80F was found to be similar to that of L80F mutant in L. donovani (Figure 8A ). We also determined the intracellular SAP activity in all of these Leishmania lines. Consistent with the changes in extracellular activity, the percentage of intracellular v/s total SAP activity is higher in cells overexpressing WT, L80F and L58,80F META1, but lower in the L58F mutant, as compared to GFP control (Figure 8B ). These results suggest that the 58^th and 80^th position are critical for the function of META1 in L. donovani. Similar results were obtained when L58F and L80F mutations were introduced into L. major META1, either individually or simultaneously (Additional file 8 ), reiterating that the L. donovani and L. major META1 proteins have a conserved function. We also observed a growth defect only in the double mutant in both L. donovani and L. major, which have higher multiplication rate at late logarithmic phase and start declining sooner than WT overexpression cells (Figure 8C and Additional file 8 respectively). The difference in SAP activity between cells overexpressing WT versus META1 mutant forms was not due to different amounts of META1 protein being expressed, as shown by western blots (Additional files 6 and 8 ). In order to eliminate the possibility of misreporting extracellular SAP activity due to cell lysis, a western blot on culture supernatants was done with GP63 and BiP antibodies, along with whole cell lysates as control. The culture supernatants that we used for the experiment did have extracellular GP63 but lacked ER chaperone BiP, as shown in Additional file 9. Overall, alteration in extracellular SAP activity in cells expressing META1 mutants points to a role for META1 in Leishmania secretory processes and highlights the importance of the putative hydrophobic pocket to META1's function.

BLASTP and PSI-BLAST searches 61 were performed against the NRDB using the NCBI server http://www.ncbi.nlm.nih.gov/BLAST. Pfam search 62 was performed using HMMer via the Pfam server at http://pfam.sanger.ac.uk/. All the sequences (DNA/protein) used in the study for codon bias, phylogenetic analysis and synonymous & non-synonymous rate estimation were obtained from NCBI. For E. coli HslJ, sequence information was used from accession number NP_4158971.1. Sequence alignment was done with CLUSTALW available at http://www.ebi.ac.uk/clustalw/ or EMBOSS http://www.ebi.ac.uk/Tools/emboss/align/index.html. Sequence to structure alignment was done with PROMALS3D software available at http://prodata.swmed.edu/promals3d/promals3d.php 63 .

Total GC content and GC content at codon positions for all L. major ORFs and META1 was obtained using

I

G

I

P

T

For calculating Codon Adaptation Index (CAI), the Relative Synonymous Codon Usage (RSCU) values of the highly expressed genes of L. major were obtained from 65 . CAI values of reference genes from Additional file 2 and META1 were calculated manually as described earlier 66 .

The sequences were aligned using MUSCLE 67 on default settings. Gaps were removed from the alignment manually. Phylogenetic tree was generated by PhyML 68 using the evolutionary model WAG, the gamma-distribution model, four rate categories and invariant position. The WAG model was found to be the best-fitting model using PROTTEST 69 . The gamma parameter and the fraction of invariant positions were estimated from the data. The trees were optimized for topology, branch length and rate parameters. Each tree was subjected to 100 bootstrap replicates. iTOL 70 was used for viewing the phylogenetic tree and later on modified in Adobe Photoshop CS4 for publication. Re-rooting of the tree was also done using iTOL. The accession numbers of the protein sequences used are given in Additional file 11.

Synonymous (K_s) and non-synonymous (K_a) substitution rates were estimated using the methods as described earlier 23 as implemented in yn00 in the PAML software 71 .

Consensus structure prediction for META1 was performed using 3D-Jury http://bioinfo.pl/meta; 29 ). MODELLER 9v2 30 software was used to obtain the final model of META1 based on the NMR structure of an E. coli HslJ (PDB code 2kts). The model was viewed using the program O 72 and was compared with that of the template and related structures. The program O was also used for superposition of META1 and HslJ structures on that of S. flexneri secretin pilot protein, MxiM (PDB code 1y9t) for comparison. The figures were made using PyMol 73 .

L. donovani MHOM/IN/1983/AG83 and L. major MHOM/Su73/5ASKH were used for all the experiments. The conditions for promastigotes Leishmania cultures are as described earlier 74 . For all experiments, cell numbers were estimated by direct counting in a hemocytometer under a light microscope. The promastigotes were routinely inoculated at a starting density of 10⁶cells/ml; typically, logarithmic growth was observed between day 2-4 and stationary phase between day 5 and 7. Hence, in all experiments, 3 days and 7 days old cultures were used as representative of log phase and stationary phase respectively. For axenic amastigotes, log phase promastigotes were pelleted down by centrifugation at 3000 rpm for 10 min, washed with PBS and resuspended in amastigote media (Complete HOMEM, 20% FCS, pH 5.5) and incubated at 37°C with 5% CO₂ for 48 hrs_. Promastigotes convert into amastigotes within 48 hrs. For generating attenuated parasite line, virulent promastigotes were passaged in HOMEM medium once a week until 30 passages and thereafter, infectivity of the parasites were checked in BALB/c mice after 4 weeks of inoculation.

Total RNA was isolated using the acid guanidinium isothiocyanate method 75 . 1 μg of total RNA from L. donovani was treated with DNAse I (Ambion Inc.). Synthesis of cDNA was performed by using First Strand Synthesis kit and the Superscript III Reverse Transcriptase (Invitrogen), according to the manufacturer's instructions. All real-time PCR experiments were performed in ABI prism 7900 HT sequence detection system (ABI) as described earlier 76 . The PCR conditions were as follows: 95°C for 10 min, 95°C for 15 sec, 58°C for 30 sec and 72°C for 30 sec for 40 cycles. Following primers were used in QRT-PCR analysis; for LdMETA1: TV359 & TV360, for LmjMETA1: TV672 & TV758, for neomycin: TV 397 & TV398, for GFP: TV276 & TV396 and for GAPDH: TV366 & TV367. Sequences of all the primers are mentioned in Additional file 12. All the reactions were analyzed using the software (SDS 2.3) provided with the instrument. The relative expression of the genes was calculated by using 2^-ΔΔCt formula using GAPDH as a normalizer. The values reported are the mean of at least three biological replicates. The standard deviation from the mean is shown as error bars in each group.

The LdMETA1 coding region was PCR amplified from L. donovani genomic DNA, using primers TV213 & TV215 and cloned in EcoRI-HindIII site of pET28a+ vector (Novagen). Primer sequences are listed in Additional file 12. The recombinant constructs were transformed in E. coli BL21 (DE3) strain and protein expression was induced at A₆₀₀ of 0.6 with 1 mM of isopropyl-1-β-D-galactopyranoside at 37 °C for 3 hours. Recombinant 6X-His-META1 protein was purified from bacterial cells in native conditions using the Ni²⁺-nitrilo triacetic acid agarose resin (Qiagen), according to the manufacturer's instructions.

For raising polyclonal antibodies against recombinant 6X-His-META1 protein in the mouse, 50 μg/animal of protein was mixed with an equal volume of Freund's complete adjuvant. The mixture was made into an emulsion by passing through a 2 ml syringe (1.5 inch 19G needle) and intermittently keeping it at 4°C. This emulsion was injected into the mouse subcutaneously. After 14 days, the first booster dose was given in a similar manner, except that Freund's complete adjuvant was replaced with Freund's incomplete adjuvant. Subsequent booster doses were given after a 14-day interval and the antibody titer was checked using dot blot. For determination of antibody titer 20 ng of purified protein was spotted in a row, air-dried and incubated with different concentrations of immune serum (1:1,000 to 1:5,000). Western blot experiments were carried out when the appropriate titer (1:5,000) was obtained following booster doses. Mice were sacrificed and blood collected from the inferior vena cava using a 23G needle and a 2 ml syringe. The blood was allowed to stand at room temperature for a couple of hours and then kept at 4°C for overnight to allow formation of a firm clot. Red blood cells were removed by centrifugation, the sera were collected in a fresh tube, 0.2% NaN₃ was added and stored at 4°C. All experiments involving animals were performed in compliance with recommendations and permission of the CCMB Institutional Animal Ethics Committee.

The META1 gene of both L. donovani and L. major were cloned in frame with GFP at the C-terminus and further this META1-GFP cassette was ligated into BamHI-XbaI site of pX63-NEO vector. The L. donovani and L. major META1 genes were amplified from genomic DNA by using the primer combinations TV494 & TV757 and TV494 & TV758 respectively and GFP was amplified using TV276 & TV277 from pX63-EGFP vector. PCR amplicons of META1 gene of both L. donovani & L. major and GFP were first cloned using pMOS blunt end cloning vector kit (GE Healthcare) as per manufacturer's instructions. Inserts were released from pMOS by using restriction enzymes inserted in the primer sequences. Extra sequences were added in the primers to maintain the reading frame and/or to add restriction sites for cloning: for META1, TV494 with a BamHI and TV757 & TV758 with NcoI site respectively and for GFP, TV276 & TV277 with an NcoI and XbaI site respectively. Primer sequences are listed in Additional file 12.

Leishmania promastigotes were grown to late log phase (~2 × 10⁷cells/ml). Transfection protocol is the same as described earlier 77 . Briefly, cells were washed and resuspended in ice cold electroporation buffer (137 mM NaCl, 5 mM KCl, 0.7 mM Na₂HPO₄, 6 mM glucose, 21 mM HEPES pH 7.5) at a density of 10⁸ cells/ml. 400 μl of cells and 100 μg of plasmid DNA were mixed well into 0.2 mm cuvettes (Biorad) then electroporated in Genepulser II apparatus (Bio-Rad Laboratories) with the pulse of 450 volts (2.25 V/cm) and capacitance 500 μF. Cuvettes were immediately placed on ice for 10 minutes. Cells were transferred to 5 ml of complete HOMEM media. After 48 hours, transformants were selected in HOMEM containing G418 until the control, untransfected cells died completely in the presence of the antibiotic. Concentrations of G418 used were: 40 μg/ml for virulent L. donovani & L. major promastigotes and 200 μg/ml for attenuated L. donovani promastigotes. For further experiments, pool populations were maintained in liquid media containing G418 in above mentioned concentrations for the respective cell types.

For generating constructs with point mutations in META1, site-directed mutagenesis (SDM) was carried out using QuickChange-XL kit from Stratagene as per manufacturer's instructions. pX63-LdMETA1::GFP (Ld/Ld-WT) and pX63-LmjMETA1::GFP (Lmj/Lmj-WT) were used as templates for SDM of L. donovani and L. major META1 mutagenesis. Details of the primers used are given in Additional file 12.

Leishmania promastigotes or amastigotes at different stages of growth were harvested, washed in PBS and then resuspended in 1X-PBS with 1X-protease inhibitor cocktail (Roche) and then sonicated using VibraCell (Sonics & Materials Inc.) with 50% amplitude and three pulses for 30 sec each. 30 μg of protein lysate from each sample were separated on 15% SDS-PAGE and transferred onto nitrocellulose Hybond ECL membranes (Amersham Biosciences). Membranes were blocked with 5% non-fat dried milk in 1X-TBS-T (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% Tween-20). The following antibodies were used: mouse anti-LdMETA1 (1:5,000) was raised against recombinant N-terminally 6X-His tagged LdMETA1 protein, rabbit polyclonal BiP antisera (1:15,000), rabbit polyclonal GFP antibody (Abcam; 1:4000), sheep polyclonal GP63 antisera (1:10,000), anti-mouse or anti rabbit biotinylated antibodies (1:5,000, Amersham), avidin conjugated with horseradish peroxidase (1:10,000, Amersham Biosciences) and anti-sheep conjugated with horseradish peroxidase (1:5,000, Santa Cruz Biotechnology). After incubation, membranes were washed three times with 1X-TBS-T. Immunodetection was carried out using the ECL western blotting detection system (GE Healthcare) according to the manufacturer's instructions and images were obtained using BIOMAX™-XBT X-ray film (Kodak). Images of western blots were quantified by ImageJ software available at http://rsbweb.nih.gov/ij/ 78 .

For western blots on the culture supernatants, media from cultures were passed through a 0.22 μm membrane and concentrated 50-fold using Millipore Amicon 5 kDa cut-off and loaded onto SDS-PAGE gel along with whole cell lysates as control.

SAP activity was determined in culture supernatants as previously described 41 . After seeding cells at a density of 10⁶cells/ml, aliquots of culture medium were removed at log and stationary phase, filtered through 0.22 μm pores membrane to remove cells and debris. 138 μl supernatant was then incubated in a final volume of 200 μl with 50 mM para-nitrophenyl phosphate (pNPP), 50 mM sodium acetate pH 4 and 0.1% β-mercaptoethanol (v/v) for 30 minutes at 37 °C. The reaction was stopped with 800 μl of 2M sodium hydroxide and the absorbance of the released para-nitrophenol was determined at 410 nm. For intracellular SAP activity, cell pellets were washed with PBS and were lysed on ice for 30 min at 10⁸ cells/ml in 50 mM acetate buffer and 1% Triton X-100. 100 μl of lysate was then used for SAP assay with 50 mM pNPP.