Leishmania-Specific Surface Antigens Show Sub-Genus Sequence Variation and Immune Recognition

Background A family of hydrophilic acylated surface (HASP) proteins, containing extensive and variant amino acid repeats, is expressed at the plasma membrane in infective extracellular (metacyclic) and intracellular (amastigote) stages of Old World Leishmania species. While HASPs are antigenic in the host and can induce protective immune responses, the biological functions of these Leishmania-specific proteins remain unresolved. Previous genome analysis has suggested that parasites of the sub-genus Leishmania (Viannia) have lost HASP genes from their genomes. Methods/Principal Findings We have used molecular and cellular methods to analyse HASP expression in New World Leishmania mexicana complex species and show that, unlike in L. major, these proteins are expressed predominantly following differentiation into amastigotes within macrophages. Further genome analysis has revealed that the L. (Viannia) species, L. (V.) braziliensis, does express HASP-like proteins of low amino acid similarity but with similar biochemical characteristics, from genes present on a region of chromosome 23 that is syntenic with the HASP/SHERP locus in Old World Leishmania species and the L. (L.) mexicana complex. A related gene is also present in Leptomonas seymouri and this may represent the ancestral copy of these Leishmania-genus specific sequences. The L. braziliensis HASP-like proteins (named the orthologous (o) HASPs) are predominantly expressed on the plasma membrane in amastigotes and are recognised by immune sera taken from 4 out of 6 leishmaniasis patients tested in an endemic region of Brazil. Analysis of the repetitive domains of the oHASPs has shown considerable genetic variation in parasite isolates taken from the same patients, suggesting that antigenic change may play a role in immune recognition of this protein family. Conclusions/Significance These findings confirm that antigenic hydrophilic acylated proteins are expressed from genes in the same chromosomal region in species across the genus Leishmania. These proteins are surface-exposed on amastigotes (although L. (L.) major parasites also express HASPB on the metacyclic plasma membrane). The central repetitive domains of the HASPs are highly variant in their amino acid sequences, both within and between species, consistent with a role in immune recognition in the host.


Introduction
Kinetoplastid parasites of the genus Leishmania cause a diverse spectrum of infectious diseases, the leishmaniases, in tropical and subtropical regions of the world (reviewed in [1]). Mammalianinfective Leishmania species are divided into two subgenera, Leishmania (Leishmania) and Leishmania (Viannia), that differ in their developmental cycles within the female sandfly vector. Transmission of species of both subgenera from vector to mammalian host requires parasite differentiation into non-replicative flagellated metacyclic promastigotes. These forms are inoculated when a female sandfly takes a blood meal; the parasites enter resident dermal macrophages and transform into replicative amastigotes that can be disseminated to other tissues, often inducing immunoinflammatory responses and persistent infection. The fate of Leishmania amastigotes in the host determines disease type, which can range from cutaneous or mucocutaneous infection to diffuse cutaneous or the potentially fatal visceral leishmaniasis [1].
Comparative sequencing of three Leishmania genomes, L. (L.) major and L. (L.) infantum from the L. (Leishmania) sub-genus and L. (V.) braziliensis from the L. (Viannia) sub-genus, has revealed high conservation of gene content and synteny across the genus [2,3,4]. A number of loci show significant variation in size and gene complement between species, however. One example is the GP63 locus, containing tandemly arrayed genes coding for surface glycoproteins that are critical for macrophage invasion and virulence [5,6]. This locus is present in all three sequenced Leishmania species but varies considerably in size and number of genes present. Another example is the LmcDNA16 locus, originally identified on chromosome 23 of L. (L.) major [7,8,9,10] but since also found in L. (L.) donovani [11], L. (L.) infantum [3] and other L. (Leishmania) species. This locus is characterised by the presence of two Leishmania-specific gene families encoding hydrophilic acylated surface proteins (HASPs; [7,9,10,12,13]) and small hydrophilic endoplasmic reticulum associated proteins (SHERPs; [14]). The HASPs have conserved N-and C-termini but a sub-set, the HASPBs, possess divergent central domains containing hydrophilic amino acid repeats that exhibit both interand intra specific variation in their size and composition [10,11,15]. Acylation of the HASPBs involves N-terminal myristoylation and palmitoylation, modifications that are required for protein targeting to the parasite plasma membrane [13]. In L. major, HASPB expression is confined to mammalian-infective stages of the parasite life cycle, the metacyclics and amastigotes. While HASPBs are antigenic in the host [16,17] and can induce protective immune responses [18,19,20], the biological functions of both the HASP and SHERP proteins remain unresolved [7].
To date, while the LmcDNA16 locus has been identified in all L. (Leishmania) species analysed, expression and localization of the encoded proteins has not been studied in New World L. (Leishmania) species. Here, we present analysis of HASPB expression in two representative sub-species, L. m. mexicana and L. m. amazonensis. Furthermore, the LmcDNA16 locus has been reported as absent from the published L. (V.) braziliensis genome, one of the few chromosomal regions showing strong divergence between sub-genera [3]. Instead, an apparently unrelated region containing several putative genes of unknown coding capacity is found in this position on chromosome 23 [3].
In this paper, we investigate this region further and identify at least two novel but closely-related L. (V.) braziliensis genes coding for putatively acylated repeat-containing proteins. These, like the HASPB proteins in L. (L.) mexicana but unlike those in L.major, are predominantly expressed on the plasma membrane of amastigotes. We name these proteins orthologous HASPs (oHASPs) and refer to the locus as the orthologous HASP locus (OHL). Sequencing one of these new L. (V.) braziliensis genes in clinical isolates taken from Brazilian leishmaniasis patients has identified extensive sequence variation in the amino acid repeat regions, while some but not all sera samples taken from the same patients recognise recombinant protein expressed from the same open reading frame expressed in E. coli. These data identify a new molecular marker for L. (V.) braziliensis infection and suggest the potential for antigenic change within this class of amastigote proteins.
N-terminal myristoylation and palmitoylation sites in target sequences were predicted using NMT -The MYR Predictor [26,27] and CSS-Palm 2.0 [28] with default settings. CLUSTAL alignments were generated for inter-and intra-species analysis of the oHASP protein repetitive regions using the CLUSTALW2 program (default settings) hosted by EBI.

Leishmania species and strains
The Leishmania species and strains used in this study are described in Table 1 and include 11 L. (V.) braziliensis clinical isolates, provided as genomic DNA by the Leishmaniasis Immunobiology Laboratory, Institute of Tropical Pathology and Public Health, Goiás Federal University (Leishbank -IPTSP/ UFG/GO). The identities of species and strains were confirmed using restriction fragment length polymorphism (RFLP) analysis [29]. The clinical isolates were identified as L. (V.) braziliensis by PCR-typing with ribosomal DNA and glucose-6-phosphate dehydrogenase/META2 genes as described [30,31,32].
L. (L.) major, L. (V.) braziliensis and L. (L.) infantum parasites were maintained in culture as described [33]. L. (L.) mexicana and L. (L.) amazonensis parasites were maintained in culture and differentiated according to the method of Bates [34]. L. (V.) braziliensis promastigotes and intramacrophage amastigotes were generated and purified as described [33]. In brief, macrophages were incubated with stationary-phase L. (V.) braziliensis at a ratio of 1:10 for 2 hr at 34uC, prior to washing twice with DMEM, replacement with fresh complete DMEM and further incubation for 48 hr at 34uC before amastigote harvesting, using 0.05% saponin and a single density isotonic Percoll gradient.

DNA extraction and analysis
Genomic DNA from each species and strain was extracted as follows: 5610 8 -5610 9 parasites were pelleted by centrifugation (2000 g, 10 min, 4uC) and washed twice with sterile PBS. Pellets were resuspended in 9 ml NET Buffer (0.01 M Tris pH 8.0, 0.05 M EDTA, 0.1 M NaCl) and 1 ml 10% SDS, ribonuclease A (Sigma Aldrich) added to a final concentration of 100 mg/ml and the mixture incubated at 37uC for 30 min. 200 ml proteinase K (20 mg/ml) was added and the mixture incubated at 55uC overnight. Parasite genomic DNA was extracted with phenolchloroform, washed twice in 70% ethanol, resuspended in TE buffer and stored at 4uC.

Author Summary
Single-celled Leishmania parasites, transmitted by sand flies, infect humans and other mammals in many tropical and sub-tropical regions, giving rise to a spectrum of diseases called the leishmaniases. Species of parasite within the Leishmania genus can be divided into two groups (referred to as sub-genera) that are separated by up to 100 million years of evolution yet are highly related at the genome level. Our research is focused on identifying gene differences between these sub-genera that may identify proteins that impact on the transmission and pathogenicity of different Leishmania species. Here we report the presence of a highly-variant genomic locus (OHL) that was previously described as absent in parasites of the L. (Viannia) subgenus (on the basis of lack of key genes) but is present and well-characterised (as the LmcDNA16 locus) in all members of the alternative subgenus, L. (Leishmania). We demonstrate that the proteins encoded within the LmcDNA16 and OHL loci are similar in their structure and surface localisation in mammalian-infective amastigotes, despite significant differences in their DNA sequences. Most importantly, we demonstrate that the OHL locus proteins, like the HASP proteins from the LmcDNA16 locus, contain highly variable amino acid repeats that are antigenic in man and may therefore contribute to future vaccine development.
Surface Proteins on Leishmania Amastigotes www.plosntds.org PCR primers were designed using the Primer3 web utility [35] with default settings and synthesised by Eurogentec. All primer sequences used are shown in Table S1. PCR amplifications for sequencing and cloning were carried out in either a Peltier PTC-200 Thermocycler (MJ Research) or a TechGene Thermocycler (Techne) using the Kod polymerase (Novagen) in 3-step reactions, according to the manufacturer's instructions. Briefly, the initial denaturing step required a 2 min incubation at 94uC and was followed by 35 reaction cycles (1 cycle = 95uC, 30 sec; 55uC, 10 sec; 72uC, 40 sec) and a final extension step of 40 sec at 72uC.
Southern blotting was carried out as described [7] with DIGlabeled probes and hybridization reagents (Roche) using the manufacturer's protocols. Primers for probe amplification were targeted against the intergenic region within the OHL locus (Table  S1). The membrane was exposed to autoradiography film (Amersham Hyperfilm HP) and processed using a XoGraph Compact x4 (XoGraph Imaging Systems).
DNA sequencing of the repeat domains of oHASP genes utilised cloned PCR products amplified with suitable flanking primers (Table S1) and cloned into pGEM-T easy vector. All sequencing was carried out on an Applied Biosystems 3130 sequencer, using T7 forward and Sp6 reverse primers, in the University of York Technology Facility; all data were analysed using Applied Biosystems Sequence Scanner v1.0.

RNA isolation and analysis
Total RNAs (15 mg per track) from procyclic, metacyclic and amastigotes of L. (L.) mexicana and L. (L.) amazonensis, generated by axenic culture [34], were extracted and analysed by formaldehyde denaturing electrophoresis in the presence of commercial RNA markers, prior to blotting and hybridization as described [14]. The radioactive probe used for hybridization, NREP, was an oligolabelled PCR product generated from the repetitive central domain of the HASPB gene (GenBank: AJ251974.1) using primers NREP1 and NREP2 (Table S1).
L. (V.) braziliensis amastigote pellets were resuspended using TRIzol Reagent (Invitrogen) and total RNA extracted according to the manufacturer's instructions. Further purification and quantitative real-time PCR (RT-qPCR) analysis was carried out as described [33]. The data generated were normalised using the constitutively expressed c-glutamyl cysteine synthetase (LbrM18_V2.1700) [36].

Protein expression, antibody generation and immunodetection
The L. mexicana HASPB open reading frame (ORF) was amplified using the primers LEXP5 and LEXP32 (Table S1) prior to cloning into the Nde1site of pET15b and expression in E. coli BL21 (DE3) pLysS [14]. His-tagged recombinant protein was purified by affinity chromatography, checked for purity by SDS-PAGE, and used to raise polyclonal antibodies (anti-Lmex HASPB) in rabbits, as described in [14].
The Lbr1110 ORF was PCR-amplified from L. (V.) braziliensis genomic DNA (wild-type strain), using the Lb1110 primers (Table  S1) and subject to ligation-independent cloning within the University of York HiTel facility (http://www.york.ac.uk/depts/ biol/tf/hitel/index.htm). The resulting recombinant plasmid was introduced into E.coli Rosetta 2 and expression achieved in autoinduction medium [37] with overnight growth at 30uC.
For protein purification, bacterial cells were resuspended in 70 ml buffer containing 300 mM NaCl, 20 mM sodium phosphate pH 7.4, 20 mM imidazole, protease inhibitors and DNAse I. Lysis was performed by one pass through a continuous flow French Press at 20 kPSi and 4uC. The crude lysate was cleared by centrifugation at 50,000 g for 40 min at 4uC followed by filtration of the supernatant through a 0.8 mm membrane. All purification steps were carried out on an AKTA100 (GE) fitted with a direct loading pump. The lysate was loaded directly onto an equilibrated 1 ml HisTrap column (GE) at a flow rate of 1 ml/min. Following a 10 column volume (CV) wash with buffer A (300 mM NaCl, 20 mM sodium phosphate pH 7.4, 20 mM imidazole), bound proteins were eluted with buffer B (300 mM NaCl, 20 mM sodium phosphate pH 7.4, 0.5 M imidazole) using a gradient of 0-100% B over 10 CV. Fractions of 1 ml were collected and analysed by SDS-PAGE; peak fractions were pooled and concentrated to ,2 ml. Gel filtration was then performed using a Superdex 75 16/ 60 column (GE) and PBS buffer at a flow rate of 1 ml/min, collecting 1 ml fractions for SDS-PAGE analysis. Purified protein (final yield, ,4 mg/L cells) was concentrated, stored at 220uC in PBS containing 25% glycerol and used for polyclonal antibody production in rabbits (Eurogentech). Antibodies were purified using a 1 ml NHS-activated HP column (GE) coupled with 1 mg recombinant Lbr1110 protein.
Following column equilibration with 10 ml binding buffer (20 mM sodium phosphate pH7, 150 mM NaCl), 15 ml rabbit serum was loaded onto the column at 0.3 ml/min. Unbound sample was removed with 5 ml binding buffer and antibody eluted at low pH (in 0.1 M glycine pH2.7, 0.5 M NaCl) in 0.5 ml fractions directly into tubes containing 50 ml 1 M Tris-HCl pH9 for neutralisation and storage.
For detection by confocal microscopy, antibody-labelling was performed on live parasites, to detect surface Lb1110, and on permeabilised cells, to detect total Lb1110 localisation. 2610 7 parasites were collected by centrifugation at 800 g for 10 min, washed and resuspended in 100 ml of 1% fatty acid-free BSA blocking solution (BB International) for 20 min. Live parasites were labelled with rabbit anti-Lb1110 (1:100) for 30 min at 20uC, then fixed in 4% paraformaldehyde (PFA) before secondary detection with AlexaFluor-488-conjugated goat anti-rabbit IgG (1:250 in blocking solution; Invitrogen). Labelling was also carried out on permeabilised cells which were first fixed in 4% PFA, washed, then incubated with 0.1% Triton-X100 (Sigma) for 10 min, washed and then incubated in 1% BSA blocking solution for 20 min at 20uC before labelling as above. Parasites were allowed to adhere to polylysine slides (Sigma) for 20 min and coverslips mounted with Vectashield containing DAPI (Vector Laboratories), prior to imaging using a Ziess LSM 510 meta with a Plan-Apochromat 63X/1.4 oil DIC I objective lens. Images were acquired using LSM510 version 3.5 software.
For detection by epifluorescence microscopy ( Figure 1C, lower panel), axenic amastigotes of L. mexicana were fixed and permeabilised as described above before labelling with anti-LmexHASPB (1:100) and detection with goat-anti-rabbit-FITC secondary antibody (Sigma). Fluorescent parasites were viewed using a Nikon Microphot FX epifuorescent microscope, images captured with a Photometrics CH350 CCD camera and data analysed via IPLab Spectrum software (Scanalytics). Intramacrophage L. mexicana amastigote infections were carried out as described above for L. braziliensis, except that macrophages were grown on glass coverslips. Infected macrophages were fixed and permeabilised as described above. HASPB localisation ( Figure 1C, upper panel) was determined using anti-Lmex HASPB, with detection by AlexaFluor-488-conjugated goat anti-rabbit IgG (1:250; Invitrogen). For analysis by flow cytometry, parasites were labelled live as described above. Samples were analysed on a Dako CyAn ADP and data evaluated by Summit 4.3 Software.

Clinical samples
Sera samples taken from 6 patients, from whom L. (V.) braziliensis parasites were also isolated, were kindly provided by the Leishmaniasis Immunobiology Laboratory, Institute of Tropical Pathology and Public Health, Goiás Federal University (Leishbank -IPTSP/UFG/GO; see Table 1). The blood samples were collected as part of the initial diagnostic procedure, at the time of first clinical evaluation and prior to treatment.

Results
The L. (L.) mexicana HASPB genes are expressed predominantly in intracellular amastigotes

Replacement of the LmcDNA16 locus in L. (Viannia) species
Although conserved in L. (Leishmania) species, the LmcDNA16 locus was reported as missing in L. (V.) braziliensis, one of the few chromosomal regions showing significant divergence between Leishmania species sequenced to date [3]. Instead, a non-syntenic region of ,7Kb (named here the OHL locus) is positioned at the same chromosomal location in L. (V.) braziliensis, as determined by examination of the LmcDNA16 locus flanking regions that contain genes that are conserved in all sequenced L. (Leishmania) species (Figure 2A). In addition, partial genome sequencing of Leptomonas seymouri, a related insect parasite, has identified a similar variable region between the same conserved flanking genes which is of reduced size and contains two ORFs (Figure 2A).
To confirm the content of the Leishmania loci, PCR amplification was used to probe L. (Viannia) and L. (Leishmania) species for HASPB and SHERP sequences, as well as for the two new ORFs identified in the OHL region of L. (V.) braziliensis (LbrM23V2.1110 and LbrM23V2.1120). This analysis confirmed the presence of conserved HASPB and SHERP genes in all analysed L. (Leishmania) species and their absence in L. (Viannia) species (data not shown). Similarly, the newly identified ORFs were only detected in the L. (Viannia) species although notably, the sizes of the bands observed were variable in both number and size (data not shown). Previous studies have shown that the genes within the LmcDNA16 locus exhibit both inter-and intra-species variation in size and content [11,15].  (Figure 2A).
Additional bioinformatics analysis revealed a sequence misassembly derived from a ,3.2 kb collapsed repeat sequence within the OHL locus. Collapsed repeats of this type frequently arise during automated genome assembly when sequence reads originating from distinct repeat copies are incorrectly joined to generate a single unit. They are identified as genomic regions with significantly increased read depth. The collapsed repeat identified here contains conserved ,1.2 kb sequences (A) flanked by ,0.8 kb sequences (B) forming an ABAB motif, as shown in Figure 2B. Each A sequence contains a putative ORF containing multiple iterations of conserved 30 nt sequences that code for a large amino acid repeat domain (see Figure 3). The number of repetitive 30 nt sequences varies between the two ORFs identified in GeneDB (http://www.genedb.org) as LbrM23V2.1110 and LbrM23V2.1120, with 4 and 14 iterations respectively. It is important to note however that these two ORFs differ only in the number of repeat units present.
Examination of the individual sequence reads that map to the collapsed repeat region reveal the presence of another ORF variant (containing 12 iterations of the repeat motif). While only three variant ORFs were detected in this analysis, the increased read depth within the OHL region suggests that multiple copies of each motif could be present and that the structure of this repeat region consists of a tandemly repeated ABAB pattern, with sequence diversity within the iterated sequences, spanning more than 12 Kb of genomic DNA. Further analysis to more precisely define the size and composition of the OHL locus is in progress.
Characterisation of the putative ORFs within the L. braziliensis OHL region  Figure 3Ac). This information suggested that the 39-UTRs of these genes are extensive, in common with other Leishmania genes. While these predicted RNA processing sites require experimental verification, their positions confirm that the OHL genes span the A and B sequences in Figure 2B, with the repeats arranged in an AB, AB reiterating pattern for RNA expression.
Comparative analysis of the putative proteins encoded by the OHL ORFs revealed significant conservation although, as described above, variation was observed in the composition and number of iterations of the 30 nt repeats that code for hydrophilic 10 amino acid repeats (Figures 3B). Of particular interest is the presence of conserved N-terminal residues, including a 2 nd position glycine and a 5 th position cysteine, confirmed as potential sites for N-myristoylation and palmitoylation using the NMT-The MYR Predictor and CSS-PALM predictive tools [27][28]. By contrast, screening for potential prenylation sites (by PrePS), GPImodification sites (by big-PI Predictor) or GPI-anchor signal sequences (by GPI-SOM ) returned no positive predictions. Overall, these data indicate that the AB sequence repeats embedded within the OHL locus have the necessary sequence components for identification as functional genes coding for proteins that contain large internal hydrophilic repeat domains and may be modified both co-and post-translationally by Nmyristoylation and palmitoylation. The OHL ORFs, therefore, have very similar characteristics to the L. (Leishmania) HASPB proteins, features evident in the comparisons and alignments presented in Figure 3 and Figure S2.
A similar analysis of the two L. seymouri ORFs reveals that both contain large hydrophilic amino acid repeat domains that are larger than those observed in the HASPs and oHASPs and also  Figure 3B). Predicted sites for N-myristoylation and palmitoylation are also found in these deduced protein sequences ( Figure 3B).

Expression and localisation of the OHL gene products
To investigate RNA expression from the two characterised OHL genes (LbrM23_V2.1110 and LbrM23_V2.1120), RT-qPCR was used for quantitative analysis of transcript levels in macrophage-derived amastigotes and axenic procyclic and metacyclics of L. (V.) braziliensis. The results were normalised using the experimentally-characterised c-glutamyl cysteine synthetase (LbrM18_V2.1700) as a constitutive control and Meta1 (LbrM17_V2.0980) as a marker for metacyclic expression [36]. Data were analysed using the Pfaffl method [33] and showed that the transcript abundances from both genes are increased 5 -10 fold in the infective metacyclic and amastigote stages relative to the procyclic stages of the parasite ( Figure 4). As expected, expression of the Meta1 gene was highly up-regulated in metacyclics versus amastigotes, as previously reported [36].  Figure 5A). In view of the RNA expression analysis above, this observation suggests that expression of the Lb1110 gene may be regulated translationally, unlike the L. (L.) mexicana homologue described in Figure 1. A second protein, migrating at ,2x the observed molecular mass of recombinant Lb1110, was also detected by anti-Lb1110 in amastigote lysates although this molecule was also detectable at low levels in metacyclics and procyclics (when compared to the differential loading indicated by detection of the constitutive marker, EF1a). It is likely that this protein is an additional oHASP containing a more extended repeat domain that is recognised by anti-Lb1110. Other OHL genes may also be expressed as proteins that are of lower abundance and therefore not readily detectable by immunoblotting.
Flow cytometry was used as a second method to quantify Lb1110 protein surface expression in procyclic, metacyclic and amastigote L. (V.) braziliensis. In these experiments, surfaceexposed Lb1110 was detected by live primary antibody labelling prior to fixation and detection with AlexaFluor 488-conjugated goat anti-rabbit IgG (see Materials and Methods). Prior to live cell staining, the amine-reactive fluorophore sulfo-succinimidyl-7amino-4-methylcoumarin-3-acetic acid (Sulfo-NHS-AMCA) was used to confirm cell viability; dead cells staining with this reagent emit a strong blue fluorescence and could be omitted from further analyses. As shown in Figure 5B, live cell staining with anti-Lb1110 was not detectable in control (no primary antibody used) or procyclic parasite populations. Conversely, 13% of parasites in a metacyclic population stained with anti-Lb1110 while 90% of amastigotes were positive with this antibody. These results confirm the stage-specificity of Lb1110 expression and demonstrate its surface exposure on the majority of amastigotes. The 59-UTR is defined by a trans-splicing acceptor site (a) and translation initiation site (b), identified using consensus sequences derived by [42]. The putative polyadenylation site (c) was predicted by PREDATERM [43]. Within the oHASP ORF (grey bar), the central black domain represents the amino acid repeats, which comprise .60% of the ORF. Primers designed to amplify the repeat region are indicated by the red arrows (d) and are shown in Table S1 As a third approach to determining the localisation of the Lb1110 protein, expression was visualised in either live or permeabilised and fixed L. (V.) braziliensis by indirect immunofluorescence and confocal microscopy ( Figure 5C). Antibody labelling was carried out either pre-or post-fixation at 20uC, in order to compare antigen localisation at the surface membrane with that detected both externally and internally within the parasite. DAPI staining of the parasite nucleus and kinetoplast was used as a counter-stain in these experiments. As shown in the upper panel of Figure 5C, anti-Lb1110 staining is specific to L. (V.) braziliensis amastigotes and, in live antibody labelled cells, Lb1110 localises to a site close to the protrusion of the rudimentary flagellum, which could be indicative of antibody capping of the surface exposed protein. In permeabilised cells (labelled Total Lb1110), by comparison, staining is evident in a punctate pattern indicative of plasma membrane and flagellar localisation on both faces of the membrane bi-layer. In the lower panel of Figure 5C, a single L. (V.) braziliensis amastigote is shown at higher magnification, clearly demonstrating the plasma membrane localisation following permeabilisation but surface localisation to the rudimentary flagellum in the non-permeabilised L. (V.) braziliensis cell. In contrast, the live labelling pattern on the L. (L.) mexicana amastigote in the same figure ('Surface HASPB') is very similar to the total labelling pattern on the fixed L. (V.) braziliensis amastigote (and to the fixed labelling seen in Figure 1C), suggesting that antibody capping is minimal on live L. (L.) mexicana under the labelling conditions used. Overall, these data suggest that the surface distribution of Lb1110 to the amastigote flagellum is not an artefact of antibody capping in these live cells.
Given the amastigote-dominant expression of Lb1110 and its surface exposure on live parasites, in a pattern similar to that observed for HASPB expression in L. (L.) major metacyclic parasites [40], we next investigated whether this protein is recognised by human immune serum collected from patients infected with L. (V.) braziliensis. Six serum samples derived from infections with the L. (V.) braziliensis clinical isolates listed in Table 1 were used to probe blots of separated parasite proteins from different stages, together with recombinant protein (as used in Figure 5A). Two examples of these immunoblots, representative of the patterns observed, are shown in Figure 6, probed with serum taken from HPV-06 and TMB-06 infections (using the same serum dilution and length of chemical exposure for all blots). Recognition of a broad size range of proteins in total parasite extracts was evident in all stages with each antiserum, while normal human serum detected few proteins above background levels and did not recognise recombinant Lb1110. Interestingly, the recombinant protein was strongly detected by HPV-06 but not by TMB-06. Overall, these data confirm the antigenicity of Lb1110 and, as with the central repetitive domain of L. (L.) major HASPB, it can be predicted that the Lb1110 repeats may provide dominant epitopes for antibody recognition. It is also evident that not all antisera taken from infected patients recognise Lb1110, suggesting that this antigen could be unstable or variant in vivo. To investigate this further, oHASP gene repeats were analysed in a number of L. (V.) ) braziliensis was carried out by RT-qPCR as described [33] with results displayed as ratios of fold increase in expression in metacyclics (black bars) and amastigotes (hatched bars) relative to procyclic parasites. Error bars represent standard errors of the mean. In this analysis, c-glutamyl cysteine synthetase (LbrM18_V2.1700) was used as a constitutive control and Meta1 (LbrM17_v2.0980) as a stage-specific control for metacyclic parasites [36].  Table 1).

Sequencing of the variable repeat domains within the oHASPs
Variation in the number of repeat iterations present in each of the 2 OHL ORFs described above (LbrM23V2.1110 and LbrM23V2.1120), coupled with the large size of the non-syntenic region, raised the possibility that further ORFs with distinct repeat regions might be present in this region, as discussed earlier. To verify this prediction, genomic DNA from the L. (V.) braziliensis genome strain (MHOM/BR/75/M2904) was subjected to PCR with primers designed to amplify the repetitive domain in the oHASPs (Figure 3Ad, Table S1). The PCR products were subcloned into the pGEM-T-easy vector, 10 clones of each selected and their insertions sequenced. The repeat domain structure was then determined for each clone and each unique sequence translated and Anti-EF1a was used as a constitutive control for protein loading on the re-probed blot below. B. Analysis of Lb1110 expression by flow cytometry in live parasites using the antibody described in A. Surface-exposed Lb1110 was detected by live primary antibody labelling prior to fixation and detection with AlexaFluor 488-conjugated goat anti-rabbit IgG. Total Lb1110 was detected by antibody-labelling post-fixation. Prior to live cell staining, the amine-reactive fluorophore sulfo-succinimidyl-7-amino-4-methylcoumarin-3-acetic acid (Sulfo-NHS-AMCA) was used to confirm cell viability; dead cells stained with this reagent emit a strong blue fluorescence and can be omitted from further analyses. The experiment shown was one of two conducted, both of which showed similar % cell counts. Control, no primary antibody. C. Use of confocal microscopy to detect either total or surface-exposed   (Figure 7). While the composition of the repeat domains varied both between and within strains, it is interesting to note that the most prevalent motif within the repeat region across all strains (excepting LTB300 and RPL-05) is GGDHGHEHMD ( Figure 7B, sequence G). Also of note, the repeat domains in the LTB300 strain are very similar to those observed in L. (V.) peruviana strain LpLCA08-90. Similarly, the RPL-05 strain shares greater conservation with the L. (V.) guyanensis strain LgM4147-75 repeat domains than with the L. (V.) braziliensis genome strain (M2904-75). In these cases, the most prevalent repeat unit appears to be GGDHVPEKAN ( Figure  7B, sequence X) and GGDHGHGNMD ( Figure 7B, sequence F) respectively. Intriguingly, the overall structure of the repeat domains is well conserved between the representative sequences from each strain with the individual motifs occupying very specific positions (Figure 7). The functional significance of this conservation has not yet been investigated further. In

Discussion
The LmcDNA16 locus, identified on chromosome 23 in all L. (Leishmania) species examined to date, contains two unusual and  Table 1). Each unique OHL sequence identified per strain (i -vi) was analysed and the position (within the repeat domain, shown in A) and composition of each individual 10-amino acid repeat unit (lettered A-Z, numbered 1-9 plus #) recorded in B. Sequences marked as single occurrence only (*) contain several low quality reads; a reliable sequence could not be definitively determined. doi:10.1371/journal.pntd.0000829.g007 Surface Proteins on Leishmania Amastigotes www.plosntds.org apparently unrelated gene families (encoding the HASPs and SHERPs), both of which are preferentially expressed during infective stages of the parasite life cycle. Ongoing functional characterisation using transgenic parasite lines lacking this locus has revealed an essential role for members of these gene families in facilitating differentiation of L. (L.) major parasites in the sandfly vector, Phlebotomus papatasi [41]. These observations suggest that the HASP and/or SHERP proteins are also likely to be essential for parasite transmission from vector to host in L. (L.) major. The absence of the HASP/SHERP locus from the L. (V.) braziliensis genome assembly, and the identification in this study of the distinct, if related, OHL region encoding proteins that also contain amino acid repeats and localise predominantly to the amastigote (but not metacyclic) plasma membrane, raises questions regarding the role of these parasite proteins in transmission from vector to host in L. (Viannia) species.
The data generated in this study demonstrate that the OHL and LmcDNA16 loci are subgenus specific (found in L. (Viannia) and L. (Leishmania) respectively), yet probably arose from a common ancestor, as suggested by analysis of the syntenic region in the monogenetic L. seymouri. Interestingly, in all Leishmania species examined so far, this region of chromosome 23 encodes gene families with similar features. These include (a) the presence of large hydrophilic amino acid repeat domains within proteins that are potentially N-terminally acylated; and (b) localisation and exposure of at least one of the encoded proteins at the plasma membrane during infective stages of the parasite life cycle. The similarity in expression patterns and localisation of the HASP and oHASP proteins supports the proposal that the encoding genes are orthologous.
The HASPBs have been previously shown to be recognition targets for host immune responses [16,17,18,19,20], possibly due to their high charge and the presence of extended hydrophilic amino acid repeat domains. Intriguingly, the variations observed in the size and composition of the oHASP repeats, both between L. (Viannia) species and within L. (V.) braziliensis strains, are similar to those observed in the HASPBs. These data support the proposal that the oHASP and HASPB proteins may have conserved functions, although the role of the repeat domains in both proteins is still unclear. While amino acid repeats are frequently involved in protein-protein contacts and could facilitate key interactions during parasite differentiation in the sand fly, the repeat domains of HASPB (and oHASP) are also expressed and diversified as surface antigens in the host, as reported in L. major [15,16] and in this paper. The detection of Lmex HASPB and Lb1110 predominantly in amastigotes of L. (L.) mexicana and L. (V.) braziliensis respectively suggests a dominant role for these proteins in the host rather than the vector for these species. Perhaps the significant sequence variation observed between the repeat domains of the oHASP proteins in the clinical isolates of L. (V.) braziliensis used here could be a consequence of variable host immune pressure.

Evolution of the LmcDNA16 loci and LmcDNA16 replacement regions
In addition to the complete genomes of L. (L.) major, L. (L.) infantum and L .braziliensis [2,3], sequence data are also currently available for L. seymouri, a monogenetic protozoan that parasitizes insects, nematodes and ciliates and is the closest sequenced relative to Leishmania. The presence of a syntenically-positioned locus containing ORFs that code for putative N-acylated proteins containing large hydrophilic amino acid repeat domains suggests the presence of this hypermutable locus in the pre-Leishmania state. Whether this locus is present in Crithidia species remains unknown. Given the comparative simplicity of the locus in L. seymouri, the expansion seen in Leishmania spp. could be representative of the shift from the monogenetic life cycle of ancestral Leishmania to the digenetic life cycle of parasites from the Leishmania sensu strictu genus. A key step in this process is the evolution of the parasiteparasitized insect relationship allowing Leishmania to use sand flies as their vector. Our recent observation that the LmcDNA16 locus is essential for L. (L.) major differentiation in Phlebotomus papatasi [41] may be of relevance in this respect.

Concluding remarks
Recent studies have demonstrated the importance of HASP proteins for L. major differentiation in the sand fly vector, while the antigenic properties of these molecules suggest their suitability as targets for vaccine development. Previous comparative genomic analyses of L. (V.) braziliensis, L. (L.) major and L. (L.) infantum, however, reported the absence of the HASP/SHERP (or LmcDNA16) locus on chromosome 23 in L. (V.) braziliensiswith a smaller non-syntenic locus (the OHL locus) found at that location.
In this paper, we show that the oHASP proteins coded within the OHL locus are orthologues of HASPB, possessing similar expression, localisation and antigenic properties. Of particular interest is the inter-and intra-species variation in the size and composition of the oHASP repeat domains (also observed in HASPBs) which could indicate that host (and/or vector) immune pressure is driving sequence diversification within this locus. Further study is now required to investigate the antigenic properties of the oHASPs, explore their interaction with the host immune system and investigate their utility as diagnostic agents for L. (L.) Viannia clinical infections.  Table 1 (Lb from strain M290475) were digested with XhoI and HinDIII, size separated through 0.6% agarose and hybridized with a digoxigenin probe targeting a repetitive intergenic region (vertical black bars in Figure 2B