Moving from unsequenced to sequenced genome: Reanalysis of the proteome of Leishmania donovani

Abstract

The kinetoplastid protozoan parasite, Leishmania donovani, is the causative agent of kala azar or visceral leishmaniasis. Kala azar is a severe form of leishmaniasis that is fatal in the majority of untreated cases. Studies on proteomic analysis of L. donovani thus far have been carried out using homology-based identification based on related Leishmania species (L. infantum, L. major and L. braziliensis) whose genomes have been sequenced. Recently, the genome of L. donovani was fully sequenced and the data became publicly available. We took advantage of the availability of its genomic sequence to carry out a more accurate proteogenomic analysis of L. donovani proteome using our previously generated dataset. This resulted in identification of 17,504 unique peptides upon database-dependent search against the annotated proteins in L. donovani. These peptides were assigned to 3999 unique proteins in L. donovani. 2296 proteins were identified in both the life stages of L. donovani, while 613 and 1090 proteins were identified only from amastigote and promastigote stages, respectively. The proteomic data was also searched against six-frame translated L. donovani genome, which led to 255 genome search-specific peptides (GSSPs) resulting in identification of 20 novel genes and correction of 40 existing gene models in L. donovani.

Biological significance

Leishmania donovani genome sequencing was recently completed, which permitted us to use a proteogenomic approach to map its proteome and to carry out annotation of it genome. This resulted in mapping of 50% (3999 proteins) of L. donovani proteome. Our study identified 20 novel genes previously not predicted from the L. donovani genome in addition to correcting annotations of 40 existing gene models. The identified proteins may help in better understanding of stage-specific protein expression profiles in L. donovani and to identify novel stage-specific drug targets in L. donovani which could be used in the treatment of leishmaniasis.

1. Introduction

Leishmaniasis refers to a broad range of diseases caused by protozoan parasite of the genus Leishmania and is the second most common cause of mortality amongst various tropical infections [1]. Leishmania is a kinetoplastid protozoan and digenic parasite of blood feeding insects and vertebrates including humans [2]. The promastigote motile form of the parasite resides in the gut of the phlebotomine sand fly. It infects the vertebrate host when the female sand fly takes a blood meal and, once inside the host, the promastigotes are engulfed by circulatory macrophages by phagocytosis. The parasite is transformed into a non-motile amastigote within the macrophage phagolysosome. The spectrum of clinical manifestations caused by the Leishmania species depends on several factors including species of infecting parasite and host factors such as immunological response resulting in cutaneous, mucocutaneous, or visceral leishmaniasis [3].

Visceral leishmaniasis, also known as Kala azar, is one of the most challenging infectious ailments and is fatal if untreated. It is caused by a complex of species comprising L. donovani, L. infantum and L. chagasi. L. donovani is predominantly found in the Indian subcontinent, Asia, and Africa while L. infantum or L. chagasi exist in the Mediterranean region, Southwest and Central Asia and South America [4]. In subclinical visceral leishmaniasis, which is newly acquired, there are no indications of disease and, as it progresses, it becomes oligosymptomatic. If not treated, it becomes fully established with hepatosplenomegaly and associated anemia, leukopenia or thrombocytopenia and hypergammaglobulinemia leading to vulnerability to secondary infections and finally death. Post Kala azar dermal leishmaniasis is a complication characterized by a macular, maculopapular and nodular rash in patients who have recovered from visceral leishmaniasis. Increasing HIV-visceral leishmaniasis co-infection worldwide is also a cause for concern [5].

The L. donovani genome sequencing was recently completed and provided insights into the organization of genes in this protozoan parasite. The L. donovani genome has 36 chromosomes comprising 32.4 million base pairs and 8195 annotated genes [6]. The other old world Leishmania species – L. major and L. infantum – have 36 chromosomes each [7,8]. However, the new world Leishmania species – L. brazilensis and L. mexicana – have 35 and 34 chromosomes [9], respectively which can be explained by fusion of chromosomes [10]. The details of the sequenced Leishmania species till date are provided in Table 1.

Table 1.

Summary of Leishmania species whose genomes have been sequenced.

	Organism	Year	Genome size	Chromosomes	Protein coding genes
1.	Leishmania major	2004	32.8 Mbp	36	8311
2.	Leishmania infantum	2007	32.1 Mbp	36	8154
3.	Leishmania braziliensis	2007	30 Mbp	35	8153
4.	Leishmania mexicana	2011	32 Mbp	34	8250
5.	Leishmania donovani	2011	32.4 Mbp	36	8195

Searching tandem mass spectrometry (MS/MS) data against six-frame translated nucleotide sequences from a genome provides the most direct evidence for protein coding genes. Thus, the use of proteomic data to annotate sequenced genomes is a complementary strategy to conventional genome annotation [11–13]. One of the first approaches to search mass spectrometry data against a six-frame translated nucleotide database was described by Yates et. al. [14] and Choudhary et. al. suggested the use of EST databases along with genome database [15]. This proteogenomic approach has been used to identify and correct annotated genes in Homo sapiens [16–19], Anopheles gambiae [20,21], Mycobacterium tuberculosis [22], Candida glabrata [23], Plasmodium falciparum [24], Toxoplasma gondii [25], Schmidtea mediterranea [26], Yersinia pestis [27], Pristionchus pacificus [28], Deinococcus deserti [29] and many other organisms. This approach provides useful information regarding the existence of novel genes, translational start sites from N-terminally acetylated peptides, frame shifts, splice variants, truncation and extension of existing proteins. This approach is mostly useful in scenarios where the genome sequence is available for the organism under study.

Non-availability of genome sequence can impede identification of proteins owing to lack of the most appropriate database against which the MS/MS spectra must be searched. However, this limitation can be overcome by using protein and genome sequence data from taxonomically related species, if they are available. Using this comparative proteogenomics approach, we had previously mapped the proteome of L. donovani [30] using protein data from three related Leishmania species (L. infantum, L. major and L. braziliensis). This proteomic analysis resulted in identification of 22,322 unique peptides which mapped to 9431 proteins from the three related Leishmania species. Out of these, a majority of peptides (12,141) mapped to 3711 proteins in L. infantum, which is taxonomically the closest species to L. donovani. We identified 1387 proteins to be expressed in both amastigote and promastigote life stages of L. donovani. However, 1423 proteins were identified in amastigotes and 901 proteins were identified in promastigotes.

In the present study, we used the mass spectrometry data from our previous protemic analysis of L. donovani and searched it against the newly annotated protein and six-frame translation of the recently sequenced L. donovani genome. This proteogenomic analysis resulted in identification of 3999 proteins in L. donovani, accounting for 50% of the annotated L. donovani proteome. In addition, this mass spectrometry data was searched against a six-frame translated genome of L. donovani resulting in the identification of 255 GSSPs which led to identification of 20 novel genes as well as 14 N-terminal and 26 C-terminal extensions of existing gene models in L. donovani.

2. Materials and methods

2.1. Leishmania donovani sample preparation, fractionation and LC–MS/MS analysis

We used axenic promastigote and amastigote stages of Dd8 strain of L. donovani (MHOM/IND/80/Dd8) in our previous comparative proteogenomic analysis. We fractionated cell lysates from both the life stages by SDS-PAGE followed by in-gel trypsin digestion. Similarly, we carried out in-solution digestion of cell lysates followed by strong cationic exchange (SCX) chromatography. A total of 112 fractions obtained from 2 different fractionation techniques were analyzed by LC–MS/MS using LTQ-Orbitrap Velos ETD mass spectrometer. MS was acquired with FT analyzer at the resolving power of 60,000 at 400 m/z. MS/MS was carried out in HCD mode with resolving power of 15,000 at 400 m/z. The lock mass option was enabled for accurate mass measurements.

2.2. Mass spectrometry data analysis

A non-redundant (nr) protein database of BPK282A1 strain of L. donovani (n = 8032) database available from NCBI (http://www.ncbi.nlm.nih.gov) as of February 20, 2012 was used. Sequest and Mascot (version 2.2) search engines were used to search L. donovani mass spectrometry data and searches were submitted through Proteome Discoverer software (version 1.3). The search parameters used were as follows: a) trypsin as a proteolytic enzyme (with up to one missed cleavage); b) peptide mass error tolerance of 20 ppm; c) fragment mass error tolerance of 0.1 Da; d) carbamidomethylation of cysteine as fixed modification and oxidation of methionine, deamidation of asparagine, glutamine and acetylation of protein N-termini were included as variable modifications. Sequest and Mascot peptide data were extracted with 1% false discovery rate (FDR) threshold. Unique peptide data obtained from both Sequest and Mascot search algorithms after protein database dependent searches were used for further analysis.

2.3. Proteogenomic data analysis

The genome sequence of BPK282A1 strain of L. donovani was downloaded from NCBI and a six-frame translated database (n = 1,096,898) was created using in-house python scripts. The six-frame translated genome database was created by masking the gaps in the genome. The translation was truncated at a codon penultimate to the gap in the genome and the first codon after the gap was treated as a start codon for a new translated protein entry. This strategy was used since it is unbiased and does not rely on pre-computed gene models. Most of the commonly occurring contaminants such as trypsin, keratins and albumin were added to both the protein and six-frame translated genome databases that were used for MS/MS ion search. Searches were submitted through Proteome Discoverer software (Thermo Scientific, version 1.3) to Sequest and Mascot (version 2.2) search engines as described above.

The L. donovani six-frame translated genome database searches resulted in identification of unique peptide data which includes GSSPs. This unique list of GSSPs was generated by comparing the unique peptide data with the protein database search results. Peptides that do not map to the protein database but map uniquely to the six-frame translated genome database were considered as GSSPs. The genomic regions to which these GSSPs map were further analyzed to identify novel genes or corrections to existing annotations (N- and C-terminal extensions). GSSPs were categorized either as 1) mapping intergenic regions or 2) overlapping annotated genes in BPK282A1 strain of L. donovani. Additionally, we used two different gene prediction programs – FgeneSH and GeneMark – to identify alternative gene models. Hence, novel genes or extensions to the existing L. donovani gene models obtained using GSSP evidence and gene prediction tools were further checked for their conservation across members of the genus Leishmania.

2.4. Bioinformatics analysis

Proteins identified in our current analysis were categorized into groups based on their primary subcellular localization e.g. cytoplasm, nucleus, extracellular. In addition to this, the proteins were also classified based on biological process (e.g., cell cycle proteins). These analyses were performed in compliance with Gene Ontology (GO) standards.

2.5. Proteomic data availability

To make our observations publicly available and accessible to other researchers, the complete set of raw mass spectrometry data (.raw files) generated from this study has been made available through the Tranche server (http://proteomecommons.org/tranche). The Tranche Hash is:

$$\begin{array}{l}\mathrm{t}+0\mathrm{N}3\mathrm{q}+\text{IYbj}+6\text{iSsAkkVwPow}0\text{FVKpJV}1\text{YIsHERg}9\text{DDwPc}\\ \text{EtPz}2\mathrm{T}7\text{LE}2\text{mBpK}8\text{yxH}0\text{ux}3\text{cpz}2\text{zN}/\text{zRTA}2\text{xTJy}8\text{TobE}22\text{cAAA}\\ \text{AAAAA}4\text{fw}==.\end{array}$$

3. Results and discussion

3.1. Summary of proteomic data

The aim of this study was to carry out a detailed genome annotation of L. donovani using the recently sequenced genome sequence data. The strategy used for proteomic profiling and genome annotation of L. donovani is outlined in Fig. 1. In our previous study, we had used 2 different life stages of L. donovani i.e. promastigote and amastigote stages to map the proteome of L. donovani using a comparative proteogenomic approach. The 112 LC–MS/MS data files that were acquired in the previous study were used for the current analysis. This dataset included ~1.5 million MS/MS spectra that were searched using Sequest and Mascot search algorithms using the recently available L. donovani protein and six-frame translated genome database. In all, 203,342 peptide spectrum matches (PSMs) that led to 17,759 unique peptides at a 1% FDR cutoff were identified from both protein and six-frame translated genome database searches. Supplementary Tables 1 and 2 provide a complete list of proteins and peptides identified in protein database searches.

Fig. 1.

Proteogenomic workflow for analysis of the Leishmania donovani proteome. Mass spectrometry data from L. donovani promastigote and amastigote was searched against L. donovani protein database. In addition, a six-frame translated L. donovani genome database was searched to identify unique genome search-specific peptides (GSSPs) which should not map to protein database. These GSSPs were further analyzed resulting in identification of novel genes and corrections to the existing gene models in L. donovani.

3.2. Confirmation of annotated proteins in L. donovani from protein database searches

The mass spectrometry data was searched against non-redundant (nr) protein database of BPK282A1 strain of L. donovani. This resulted in identification of 17,504 unique peptides which mapped to 3999 proteins thus confirming the existence of corresponding protein annotations in L. donovani which accounted for 50% of L. donovani proteome. The peptide data from our current (17,504 peptides) and previous (22,322 peptides) study was compared for identification of orthologs and paralogs in different Leishmania species. We identified many peptides which are shared between 2 or more Leishmania species this shows the homology between orthologs from different Leishmania species. L. infantum is taxonomically the closest species to L. donovani and majority of our peptides (12,177) from L. donovani mapped to L. infantum thus validating their phylogenetic relationship.

3.3. Confirming start sites of known proteins

Identifying correct translational start sites (TSS) is a challenging task for predicted transcripts. Most of the methods used for TSS determination depend on annotating the longest open reading frame in a given nucleotide sequence [31]. Thus, additional methods such as homology-based bioinformatics approach and mass spectrometry-based approaches have been used to determine N-terminal acetylation of proteins [17,32,33]. N-terminal acetylation occurs in a majority of eukaryotic proteins and it is catalyzed by N-acetyltransferases (NAT) which transfer acetyl groups from acetyl-CoA to the intact N-termini or to one of the internal amino acids after the initiator methionine alone or additional residues are cleaved. There are 5 groups of N-terminal acetyl transferases (A to E) which carry out N-terminal acetylation of majority of eukaryotic proteins [34,35]. Confirmation of exact TSS based on N-terminal acetylation can be determined by using mass spectrometry-based approaches [36,37]. N-terminally acetylated peptides can be determined by searching mass spectrometry data using N-terminal acetylation as a modification during database searches. In our present analysis we identified 113 N-terminally acetylated peptides and among these the majority had methionine (43%) residue acetylated. In addition to this, alanine (17%) and serine (30%) residues were also found to be modified. The complete list of N-terminally acetylated peptides and the corresponding proteins is shown in Supplementary Table 2.

3.4. Proteogenomic analysis of L. donovani

3.4.1. Identification of novel genes in L. donovani

GSSPs identified in six-frame translated genome database searches were categorized as intergenic if they mapped to a particular region in the genome with no annotated protein coding genes. Such matches suggest the presence of novel protein coding genes which are not annotated in L. donovani genome. In the present study, we identified 113 peptides mapping to the intergenic region which in turn resulted in identification of 20 novel protein coding genes in L. donovani. A complete list of GSSPs with the genome co-ordinates which resulted in identification of novel genes in L. donovani is provided in Supplementary Table 3.

An illustrative example of a novel gene identified in L. donovani is where a novel gene coded for a flagellar radial spoke protein. This protein forms a part of the flagellar apparatus where it plays an important role in controlling flagellar movements. We identified 4 unique GSSPs mapping to intergenic region on chromosome 29 on the negative strand. Upon gene prediction analysis, we identified a 702 amino acid long protein coding gene in L. donovani (Fig. 2). An ortholog for flagellar radial spoke protein exists in the closely related species L. infantum which is coded by gene LinJ29_0690 and it codes for 702 amino acids. Additionally, there is a truncated flagellar radial spoke protein coding gene (LDBPK_290690) on chromosome 29 on the negative strand which codes for 36 amino acid long truncated protein. This indicated a wrong annotation of flagellar radial spoke protein coding gene and the full length gene was annotated based on our analysis. A pair wise alignment of novel flagellar radial spoke protein identified in L. donovani and its ortholog, LINJ29_0690, in L. infantum (Supplementary Fig. 1) shows that the full length novel flagellar radial spoke protein (FgeneSH) identified in our study is highly conserved when compared to its L. infantum ortholog (LINJ29_0690).

Fig. 2.

Identification of a novel protein coding gene based on peptides mapping to an intergenic region in L. donovani. (A) Chromosomal position of the truncated flagellar radial spoke protein coding gene LDBPK_290690 (blue box) which is a part of the current genome annotation in L. donovani. Four intergenic peptides (marked as red bars) mapped to a region where no known gene was predicted in L. donovani. Gene prediction analysis of this region resulted in identification of full length flagellar radial protein coding gene (white box) on chromosome 29 in L. donovani. Upon further bioinformatics analysis a corresponding conserved ortholog was identified in L. infantum LINJ_29_0690. (B) A pair wise sequence alignment of full length ortholog from L. infantum LINJ_29_0690 and novel predicted protein identified in FgeneSH gene prediction analysis of L. donovani. Red color sequences denote the GSSPs identified in L. donovani which support the existence of this novel protein coding gene in L. donovani which codes for flagellar radial spoke protein. (C) A representative MS/MS spectrum for identification of genome search specific peptide LPDVSPHHITVAR is shown.

3.4.2. Identification of N-terminal extensions in L. donovani

We also refined existing gene models in L. donovani. In the present study, we identified 26 GSSPs mapping to the 5′ boundary of existing genes which resulted in N-terminal extension of 14 proteins in L. donovani. As an example, we observed an N-terminal extension of an existing hypothetical protein-coding gene (LDBPK_261010), which was predicted to encode a 357 amino acid long protein. Based on three GSSPs mapping to the 5′ region upstream of this gene and additional gene prediction analysis, we were able to annotate a longer protein product of 779 amino acids which contained the 3 GSSPs. This longer protein had a conserved ortholog in L. major, which was encoded by the gene LmjF_26_1030. This ortholog codes for a longer protein product of 779 amino acids (Fig. 3). A pair wise alignment of truncated L. donovani protein LDBPK_261010, L. major ortholog LMJF_26_1030 and the full length N-terminally extended L. donovani protein (FgeneSH) (Supplementary Fig. 2) shows that the N-terminally extended L. donovani protein identified in our study is highly conserved.

Fig. 3.

Identification of N-terminal extended protein in L. donovani. (A) Three peptides (marked as red bars) mapped to the upstream region of L. donovani gene LDBPK_261010 (red box) and this codes for short protein product of 367 amino acids (blue box). However, upon gene prediction analysis the presence of a much longer protein coding gene extending N terminal of LDBPK_261010 gene was identified (white box). Ortholog LMJF_26_1030 codes for 779 amino acid long protein (blue box) in L. major and this supports the N terminal extension. (B) A pair wise sequence alignment of full length ortholog (LMJ_26_1030) from L. major and the truncated L. donovani protein (LDBPK_261010) as well as full length predicted protein identified in FgeneSH gene prediction analysis. Red color sequences denote the GSSPs identified in L. donovani which support N-terminal extension and correction of L. donovani gene LDBPK_261010 and its conservation across the other ortholog in L. major. (C) A representative MS/MS spectrum for identification of genome search specific peptide GIYEALQAFELDR is shown.

3.4.3. Identification of C-terminal extension in L. donovani

In the present study, we identified 116 peptides mapping to the 3′ boundaries of existing genes which resulted in C-terminal extension of 26 proteins in L. donovani. As an illustration for C-terminal extension, we identified a heat shock protein 70 (Hsp70) coded by the gene LDBPK_302540 in L. donovani. It codes for a small truncated protein product (86 amino acids) which is a result of gap due to series of nucleotides (724 Ns) downstream of gene LDBPK_302540 where no nucleotide could be assigned. The full length Hsp70 protein (502 amino acids) is encoded by the gene LDBPK_302480 and is localized to the mitochondria. However, we have identified 11 unique GSSPs which map downstream of the truncated LDBPK_302540 gene. Upon further bioinformatics analysis, we determined that these peptides are well conserved in HSP orthologs from L. infantum (LINJ_30_2530) and L. major (LMJF_30_2550) which are of 661 amino acids. This resulted in C-terminal extension of Hsp70 (LDBPK_302540) by 575 amino acids (Fig. 4). A pair wise alignment of truncated L. donovani heat shock proteins (LDBPK_302540 and LDBPK_302480) and full length orthologs from L. infantum (LINJ30_2870) and L. major (LMJF_30_2550) shows that the C-terminally extended regions identified in our study is highly conserved (Supplementary Fig. 3).

Fig. 4.

Identification of C-terminal extended protein in L. donovani. (A) Eleven peptides (marked as red bars) mapped to the downstream region of the truncated L. donovani Hsp70 gene LDBPK_302540 which codes small truncated protein of 86 amino acids (blue box) which is a part of the current genome annotation in L. donovani. However, upon gene prediction the presence of a much longer gene extending C terminal of the current gene model (white box) was identified in our analysis. Orthologs LINJ_30_2530 and LMJF_30_2550 genes in L. infantum and L. major respectively support (blue box) the C terminal extension. (B) Full length heat shock protein 70 (Hsp70) protein sequence (LINJ_30_2530) and the blue colored sequence represents conserved truncated Hsp70 protein LDBPK_302540 in L. donovani showing complete homology with L. infantum ortholog. The remaining black colored extended sequence in the L. infantum ortholog contains the eleven GSSPs (red colored) which are conserved in L. donovani. (C) A representative MS/MS spectrum for identification of genome search specific peptide NNAETQLTTAER is shown.

In our previous comparative proteogenomic study we had identified 13 N-terminal extensions in L. donovani. We have reanalyzed these previously identified extensions in L. donovani with the availability of genome sequence data and we found that only 2 out of the 13 N-terminal extensions are incorporated in the current genome annotation of L. donovani. This shows the utility of proteogenomic approaches in annotating genomes. Our current proteogenomic analysis of L. donovani using sequenced genome data identified many genes that were erroneously annotated as truncated forms. These truncated forms were a result of gaps arising from the non-sequenced flanking regions with no nucleotide sequence. These truncated genes are frequently referred to in the current genome annotation of L. donovani as “partial” and there are many such truncated partial protein coding genes in L. donovani. However, our proteogenomic analysis of L. donovani identified peptides mapping upstream or downstream of the gap which resulted in N- or C-terminal extension of these protein coding genes. A complete list of GSSPs along with the genome coordinates which resulted in identification of N- and C-terminal extensions to the existing gene models in L. donovani is provided in Supplementary Table 4.

3.5. Expression of proteins in promastigote and amastigote stages of L. donovani

The promastigote stage represents the insect stage and is well suited for survival in the gut of sandflies. Amastigotes on the other hand are adapted to survive in the highly acidic environment inside the phagosome of infected macrophage. A different repertoire of proteomes thus needs to be expressed in order for the parasite to survive in different environments (i.e. gut of sandflies or inside the infected macrophages). Previous transcriptomic [38–42] and proteomic analyses from other groups has provided a set of differentially expressed genes/proteins in these 2 different life stages of Leishmania [43–49]. Many of these studies including our proteogenomics made use of axenic amastigotes which are generated by differentiating promastigotes by culturing them at 37 °C and pH 6 in 5% CO₂ which mimics the intracellular environment inside the macrophages. In a study by Lahav et al. the differentiation from promastigote to amastigote was studied in the context of mRNA and protein levels using a transcriptomic and differential proteomic approach. Various time points during the differentiation of promastigote to amastigote were sampled and it was found that there is a substantial change at mRNA level during early differentiation. However, during the later stages of the differentiation process, translation and post-translational regulation of protein abundance was shown to play an important role [50].

It is debated that the true proteome of intracellular amastigote is different from the axenic amastigotes which are frequently used in proteomic analysis due to the ease of culturing and scaling up of axenic amastigote cultures and frequently these axenic forms are referred to as amastigote-like to differentiate them from true intracellular amastigotes [51,52]. In a recent comparative proteomic study by Biyani et al., intracellular amastigotes and axenic promastigotes from three different clones were used for analysis. This study resulted in identification of differential set of proteins between the two life stages and many of the differentially expressed proteins were conserved across the different clones. In this study, they identified 205 proteins in total and out of these 29 proteins were upregulated and 36 were downregulated in amastigotes compared to promastigotes [53].

In our present proteogenomic analysis, 613 proteins were identified in the amastigote stage and 1090 proteins were identified in the promastigote stage of L. donovani. In contrast 2296 proteins were identified in both the life stages of L. donovani. The distribution of identified proteins between the two life stages of L. donovani is shown in Fig. 5. We identified various classes of proteins expressed in both life stages including proteins involved in cell motility, cytoskeleton, energy metabolism and cellular signaling. In addition to this, many of the factors involved in survival and virulence were also identified in this study.

Fig. 5.

Distribution of proteins identified in the 2 life stages of L. donovani. Distribution of proteins identified in L. donovani between the 2 different life stages i.e. promastigote (insect stage) and amastigote (human stage).

3.6. Identification of virulence factors in L. donovani

We identified various virulence factors known to be associated with infection process. This included previously known virulence factors such as Leishmania homolog of activated C kinase receptor (LACK) which is encoded by the gene LDBPK_282970 and is known to induce suppression of CD4+ T cells upon infection of the host and it may play a crucial role in parasite infection [54,55]. Prohibitins are GPI anchored proteins which are shown to be localized in flagellar and aflagellar pole of the L. donovani promastigote. Prohibitins encoded by the gene LDBPK_161710 in L. donovani. The aflagellar pole is involved in invasion process of macrophage and it involves binding of prohibitin to its corresponding surface HSP70 receptor on macrophage. Hence, it seems to play an important role in host–parasite interaction [56]. Magnesium transporter 2 (MGT2) is a cation transporter encoded by the gene LDBPK_251130 and it is involved in the transport of magnesium in intracellular amastigotes engulfed in phogolysosome of macrophages. MGT2 single allele knockout study showed drastic loss of virulence and it also reduced ability of Leishmania to survive inside acidified phagolysosomes. This was attributed to the inability of promastigote to transform completely to amastigote which is better suited to survive inside the harsh environment of phagolysosome [57].

Sphingolipids are components of lipid rafts and are involved in cellular signaling. The de novo synthesis of sphingolipids involves serine palmitoyltransferase which is a crucial enzyme in the biosynthetic pathway and is coded by the gene LDBPK_350320 in L. donovani. Mutation of spt gene has been shown to result in inability of parasites to differentiate into infective metacyclic forms. Additionally, the undifferentiated parasites die due to accumulation of vesicles [58] and there is defect in membrane trafficking in extracellular Leishmania. It was also seen that the parasites are defective in infecting host macrophages and this was attributed to defect in the localization of GP63 (leishmanolysin) and GPI mediated association GP63 to the lipid rafts [59]. In our study, we identified many other virulence factors such as phosphoglycan beta 1,3 galactosyltransferase, lipophosphoglycan biosynthetic enzyme and surface antigen-like protein which were also identified in our previous comparative proteogenomic analysis of L. donovani. These virulence factors are expressed at a basal level in axenic L. donovani.

3.7. Identification of survival factors in L. donovani

We identified many ATP binding cassette (ABC) class of proteins in our current analysis of L. donovani. ABC proteins have been shown to be associated with drug resistance in Leishmania. ABC proteins characteristically have transmembrane domains that are hydrophobic and form a channel for passage of molecules. In addition to this, they have nucleotide binding domain which bind to ATP and bring about hydrolysis of ATP to drive the process of pumping drugs or other metabolites outside the cell [60,61]. We identified a classic ABC transporter P-glycoprotein (LDBPK_341060) in L. donovani. P-glycoprotein or Pgp was identified initially as an ABC gene in the genus Leishmania and it was shown to be a part of the H-circle DNA which is frequently found to be amplified in drug resistant Leishmania [62,63]. Initially P-glycoprotein was shown to induce methotrexate resistance in Leishmania tarentolae [62] and L. major [64]. It was shown later that P-glycoprotein could confer arsenate resistance in different Leishmania species upon transfection [65] but the amount of resistance to the drugs varied with type of species [66]. In addition to this, the gene amplification of pgpA gene was shown not to be directly related to the level of resistance that it confers to Leishmania [65]. However, null mutants of pgpA gene can result in increased sensitivity to oxyanions such as arsenate and antimony in L. tarentolae, but the PgpA was shown to confer only low level of protection against oxyanions upon pgpA gene transfection in mutant cells and it is not the major transporter of oxyanions [67].

We identified another member of the ABC family multidrug resistance protein in L. donovani which is encoded by the gene LDBPK_282660. This MRP like protein is shown to be involved in active transport of multiple drugs including amphotericin B and sodium stibogluconate [68]. This ABC protein was shown to be 2-fold upregulated in antimony resistant L. donovani clinical isolate [69]. In addition to these, we identified many other ABC proteins in our current analysis of L. donovani and the partial list of the identified ABC proteins is shown in Table 2. We also identified additional proteins which play a crucial role in conjunction with ABC proteins such as PgpA in efflux of various metabolites including drugs. We identified γ-glutamylcysteine synthase (γ-GCS) and ornithine decarboxylase (ODC) which are encoded by genes LDBPK_181660 and LDBPK_120100 in L. donovani, respectively. These proteins are crucial enzymes involved in synthesis of glutathione and polyamine spermidine. γ-glutamylcysteine synthase (γ-GCS) was shown to be upregulated in drug resistant Leishmania [70]. Both these molecules are precursors for synthesis of low molecular weight thiols such as trypanothione [71] which play important role in preventing oxidative damage and maintaining intracellular reducing environment [72]. Trypanothione is synthesized by trypanothione synthetase [73] which we have identified in L. donovani and is coded by the gene LDBPK_271770. This trypanothione is conjugated by a transferase to heavy metals/drugs and is recognized by PgpA and it assists in the active extrusion of trypanothione conjugated molecules outside the cells [74].

Table 2.

List of ATP binding cassette proteins identified in L. donovani.

	Gene	Protein	Unique peptides	Life stage
1	LDBPK_290640	ATP-binding cassette protein subfamily A, member 10	15	Amastigote and promastigote
2	LDBPK_210770	ATP-binding cassette protein subfamily E, member 1	14	Amastigote and promastigote
3	LDBPK_250540	ATP-binding cassette protein subfamily B, member 1	13	Amastigote and promastigote
4	LDBPK_341060	P-glycoprotein	11	Amastigote and promastigote
5	LDBPK_190800	ATP-binding cassette protein subfamily F, member 2	9	Amastigote and promastigote
6	LDBPK_330340	ATP-binding cassette protein subfamily F, member 3	8	Amastigote and promastigote
7	LDBPK_150950	ATP-binding cassette protein subfamily G, member 4	6	Promastigote
8	LDBPK_282660	MRP protein-like protein	4	Amastigote and promastigote

3.8. Analysis of secreted and membrane proteins from L. donovani

In our current study, we carried out bioinformatics analysis to identify membrane and secreted proteins in L. donovani. Of the 3999 proteins identified in L. donovani 202 proteins contained only signal peptides (SP) and 282 proteins contained a transmembrane domain. In addition, 71 proteins contained both a transmembrane domain and a signal peptide (Fig. 6A). Identifying secreted and transmembrane proteins in L. donovani can provide valuable insights into the function of these proteins in relation to virulence and pathogenesis. Amastin which is encoded by gene LDBPK_281510 is known to adjust the intracytoplasmic pH of macrophage engulfed amastigotes in the phagolysosome by pumping ions across the membrane [75]. It is predominantly expressed in the amastigote life stage upon exposure to acidic pH [76]. The amastin gene family is highly diversified in the genus Leishmania [77]. Tuzins are a class of surface proteins identified to be part of amastin family and it has been studied in context of related kinetoplastid Trypanosoma cruzi, where its expression was shown to be post-transcriptionally suppressed resulting in low levels of tuzin [78]. Cysteine peptidase B was shown to inhibit IL-12 production from macrophages stimulated with bacterial lipopolysaccharide (LPS). It was shown to bring about degradation of IKappaB alpha and beta as well as NF-kappaB. Hence, Leishmania is able to inhibit NF-KappaB signaling and in turn IL-12 production [79]. The partial list of secreted or transmembrane molecules identified in this study is shown in Table 3.

Fig. 6.

Feature of proteins identified in L. donovani. (A) Distribution of membrane and secreted proteins identified in L. donovani. (B) Distribution of L. donovani proteins based on biological function. (C) Distribution of L. donovani proteins based on subcellular localization. A great percentage of proteins remain unclassified.

Table 3.

List of transmembrane and secreted proteins identified in L. donovani.

	Gene	Protein	Biological role	Life stage
1	LDBPK_362720	Membrane-bound acid phosphatase	It is a membrane bound acid phosphate involved in protection against oxidative stress to intracellular Leishmania	Amastigote and promastigote
2	LDBPK_150370	Ecotin	It is an inhibitor of host serine peptidases. It modulates Leishmania infectivity and phagocytic activity by inhibiting macrophage serine peptidases [79]. It is also involved in normal development of Leishmania in infected macrophages.	Amastigote and promastigote
3	LDBPK_010660	HSP70-like protein	Stress response protein that has been shown to play crucial role not only in intracellular survival of parasite but also during infection process of host macrophages	Promastigote
4	LDBPK_367280	Protein disulfide isomerase	It is thioredoxin domain containing protein involved in formation of disulfide bonds in newly synthesized proteins [80,81].	Amastigote and promastigote
5	LDBPK_080740	Tuzin	It is a surface protein	Promastigote
6	LDBPK_070600	Cysteine peptidase B	Cysteine peptidase B was shown to inhibit IL-12 production from macrophages stimulated with bacterial lipopolysaccharide (LPS) by degrading IKappaB alpha and beta as well as NF-kappaB.	Amastigote and promastigote
7	LDBPK_281310	Glucose-regulated protein 78	It is a member of HSP 70 family and it is localized to the endoplasmic reticulum.	Amastigote and promastigote
8	LDBPK_291230	Tryparedoxin-like	It is a oxidoreductase involved in maintaining reducing environment within Leishmania	Amastigote and promastigote

3.9. Gene ontology and bioinformatics analysis

Gene ontology analysis was carried out to categorize protein based on biological process and localization. The proteins identified in the current proteomic analysis of L. donovani were classified based on biological process (e.g. cell communication and signaling). This GO analysis resulted in 1761 proteins (44%) being grouped into one of the biological functions (Fig. 6B). In addition to this, proteins were categorized based on primary subcellular localization e.g. nucleus, cytoplasm and cytoskeleton; 1020 proteins (26%) were assigned one of the subcellular locations (Fig. 6C). In our current study, we also analyzed biological domains and motifs for the proteins identified in L. donovani and the details can be found in Supplementary Table 1. In addition to this, L. donovani has many genes that are functionally uncharacterized and do not have sequence similarity with other known genes in protozoa. We identified 1578 that are annotated as hypothetical proteins in recently sequenced L. donovani. Many of these proteins could be interesting targets for drug and vaccine development.

4. Conclusions

In our proteogenomic analysis reported here, we have cataloged the largest set of proteins from the two life stages of L. donovani. This illustrates the importance of high-resolution mass spectrometry derived data in proteomic profiling of L. donovani. Many of the proteins identified in our current analysis using sequenced L. donovani genome and protein data have not been shown to exist previously. However, many proteomic studies in the past, including our previous comparative proteogenomic study, had identified many proteins based on homology with the closest taxonomically related Leishmania species i.e. L. infantum. Significantly, we identified 255 GSSPs from six-frame translated genome searches of L. donovani which led to correction of 40 gene models (14 N-and 26 C-terminal extensions). We also identified 20 novel genes in the recently sequenced L. donovani genome based on genome search-specific peptide evidence. This shows that annotating exact protein-coding genes is a challenging task and proteogenomic approach can help not only to identify novel genes but also refine gene models of existing genes.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.jprot.2013.04.021.

Abbreviations

GSSPs

Genome search specific peptides

FDR

False discovery rate

PSM

Peptide spectrum matches

HCD

High energy collision induced dissociation