INTRODUCTION
Visceral leishmaniasis (VL) or kala-azar (KA), fatal if not treated, is caused by protozoan parasites of the Leishmania donovani complex, comprising of L. d. donovani, L. d. infantum and L. d. chagasi. More than 90% of the VL cases in the world are reported from Bangladesh, Brazil, India and Sudan (Desjeux, Reference Desjeux2001). In India, L. d. donovani is the primary causative agent of VL and states of Bihar, Uttar Pradesh and West Bengal are highly endemic foci of KA where periodic epidemics are common (Sundar and Rai, Reference Sundar and Rai2002).
Dimorphic Leishmania parasites pass through markedly different environments with a rise in ambient temperature and acidification encountered by promastigotes upon entry into the macrophage cell of the mammalian host, leading to transformation to the amastigote stage. The parasites adapt themselves to alterations in the environment, in order to survive and proliferate inside their hosts. The knowledge of the molecular basis of the promastigote-to-amastigote transformation is limited, although it is well established that several environmental factors such as pH and temperature trigger this cytodifferentiation process and have been utilized to develop in vitro culture conditions that mimic the differentiation seen in vivo (Zilberstein and Shapira, Reference Zilberstein and Shapira1994; Debrabant et al. Reference Debrabant, Joshi, Pimenta and Dwyer2004). Apart from the temperature difference in the two environments; the insect gut and the macrophage phagolysosome also offer different qualitative and quantitative composition of carbon sources. The different parasite stages therefore, must accomplish an adaptation of the metabolic pathways. The in vitro-cultured (axenic) amastigotes provide sufficient numbers of parasites to permit the molecular, biological and biochemical approaches that would be difficult or impossible with intracellular amastigotes. Although the axenic parasites may not be identical in every way to the amastigotes in an infected mammal, previous, detailed studies have documented their similarity to animal-derived Leishmania cells (Duncan et al. Reference Duncan, Alvarez, Jaffe, Wiese, Klutch, Shakarian, Dwyer and Nakhasi2001; Debrabant et al. Reference Debrabant, Joshi, Pimenta and Dwyer2004).
Microarray techniques using genomic DNA, cDNA or oligonucleotide microarrays are high throughput approaches for large-scale gene expression analysis and enable the investigation of the mechanism of fundamental processes and the molecular basis of disease on a genomic scale. In parasitological studies, microarray technology has been used successfully to identify critical genes expressed during development of Plasmodium (Young et al. Reference Young, Fivelman, Blair, de la Vega, Le Roch, Zhou, Carucci, Baker and Winzeler2005), Trypanosoma (El-Sayed et al. Reference El-Sayed, Hegde, Quackenbush, Melville and Donelson2000; Imai et al. Reference Imai, Mimori, Kawai and Koga2005) and Leishmania (Saxena et al. Reference Saxena, Worthey, Yan, Leland, Stuart and Myler2003; Akopyants et al. Reference Akopyants, Matlibs, Bukanova, Smeds, Brownstein, Stormo and Beverley2004; Almeida et al. Reference Almeida, Gilmartin, McCann, Norrish, Ivens, Lawson, Levick, Smith, Dyall, Vetrie, Freeman, Coulson, Sampaio, Schneider and Blackwell2004; Goyal et al. Reference Goyal, Duncan, Selvapandiyan, Debrabant, Baig and Nakhasi2006; Holzer et al. Reference Holzer, McMaster and Forney2006).
We developed a prototype chip for L. donovani and used it to compare gene expression in Leishmania parasites isolated from KA and post-kala-azar dermal leishmaniasis patients (Duncan et al. Reference Duncan, Salotra, Goyal, Akopyants, Beverley and Nakhasi2004; Salotra et al. Reference Salotra, Duncan, Singh, Subba Raju, Sreenivas and Nakhasi2006). The genomic microarray was expanded to 4224 clones and used in the present study to look for the gene expression changes that occur during the promastigote-to-amastigote differentiation. Whereas there are several studies on stage-specific expression of genes in promastigotes and terminally differentiated amastigotes (Coulson and Smith, Reference Coulson and Smith1990; Joshi et al. Reference Joshi, Dwyer and Nakhasi1993; Zhang and Matlashewski, Reference Zhang and Matlashewski1997; Krobitsch et al. Reference Krobitsch, Brandau, Hoyer, Schmetz, Hubel and Clos1998; Wu et al. Reference Wu, El-Fakhry, Sereno, Tamar and Papadopoulou2000), the knowledge of genes that are expressed during the differentiation process is limited (Duncan et al. Reference Duncan, Alvarez, Jaffe, Wiese, Klutch, Shakarian, Dwyer and Nakhasi2001). We therefore chose to investigate the changes in parasite gene expression during a time-course of differentiation of Leishmania promastigotes into amastigotes, to search for genes that may contribute to this process and hence may play a role in Leishmania virulence.
MATERIALS AND METHODS
Parasite isolation and culture
The L. donovani parasite strain used to prepare genomic DNA fragments printed on the microarray, induced to differentiate in culture and for preparation of RNA for microarray hybridization is an isolate from an Indian KA patient (K59) originating from Bihar and reporting to Safdarjung Hospital (SJH), New Delhi. The parasite strain was prepared from bone-marrow aspirates as before (Salotra et al. Reference Salotra, Sreenivas, Pogue, Lee, Nakhasi, Ramesh and Negi2001; Sreenivas et al. Reference Sreenivas, Singh, Selvapandian, Negi, Nakhasi and Salotra2004). Diagnosis of KA was confirmed by demonstrating the presence of LD bodies in the bone marrow. Informed consent was obtained from patients before collecting the bone-marrow samples, according to the guidelines of the Ethical Committee, SJH. The K59 parasite strain was characterized by isoenzyme analysis, immunofluorescence assay with monoclonal antibodies and genomic fingerprinting by AP-PCR as in an earlier study (Sreenivas et al. Reference Sreenivas, Singh, Selvapandian, Negi, Nakhasi and Salotra2004). The promastigotes (Pro) were cultured in Medium 199 with 10% fetal calf serum and subjected to in vitro transformation into amastigotes (Am) by serially adapting them to grow at elevated temperatures and reduced pH conditions as described (Debrabant et al. Reference Debrabant, Joshi, Pimenta and Dwyer2004). Briefly, the promastigotes grown at 26°C in Medium 199 supplemented with 10% FCS were adapted to grow in the same medium at pH 6·8. After several serial passages, the parasites were adapted to grow at pH 5·5 in RPMI medium supplemented with 20% fetal calf serum and 20 mm MES buffer at 26°C. The temperature was increased stepwise to 30°C, 33°C and finally 37°C to adapt the parasites to increased temperature conditions. Subsequent to several passages under these conditions, the parasites transformed and grew as intermediate forms. When these adapted intermediate form parasites were transferred to 37°C in an atmosphere of 5% CO2 in the same medium, at pH 5·5, the parasites were fully transformed and grew as axenic amastigotes. Once adapted, the axenic amastigotes were maintained by shuttling between promastigote and amastigote culture conditions at each passage. The cell differentiation was confirmed by measuring the expression of an amastigote-specific gene A2, by RT-PCR.
For generating PA24 parasites, the adapted amastigote parasites were differentiated back to the promastigote form. The fully differentiated flagellated promastigotes were incubated in amastigote culture conditions (pH 5·5, 37°C in an atmosphere of 5% CO2) for 24 h.
Isolation of RNA
Total RNA was isolated from 3 Leishmania life-stages: promastigotes (Pro), an intermediate differentiation stage i.e. promastigotes after incubating in amastigote culture conditions for 24 h (PA24) and terminally differentiated axenic amastigotes (Am) using Trizol reagent (Invitrogen) according to the manufacturer's instructions. To reduce variations in the quality of RNA, samples were collected from a comparable number of parasites at identical growth points (late log phase) and grown under identical conditions.
Leishmania donovani genomic microarray
The construction of the L. donovani genomic microarray is described elsewhere (Duncan et al. Reference Duncan, Salotra, Goyal, Akopyants, Beverley and Nakhasi2004). Briefly, a library of 1–1·5 kb randomly sheared fragments of genomic DNA from a fresh isolate of L. donovani prepared from an Indian KA patient was ligated into the pZeRO vector (Invitrogen). A total of 4224 spots, representing ∼37% of the expressed genes (Akopyants et al. Reference Akopyants, Clifton, Martin, Pape, Wylie, Li, Kissinger, Roos and Beverley2001), comprised of 4188 PCR amplified inserts from library clones along with alien external DNA controls, 24 known Leishmania genes and 12 negative controls, were printed in triplicate arrays on poly-L-lysine coated slides. A sampling of the Leishmania genomic library clones printed on the array was sequenced and the sequence searched against the Leishmania major genome database (www.genedb.org/genedb/leish/index.jsp). From among the random samples, all of the inserts aligned with the L. major genome with nucleotide sequence homology of 85% or better, most were close enough to an annotated open reading frame (less than 500 bp on the 5′ side, overlapping or less than 1500 bp on the 3′ side) to assign the clone to a gene. A small proportion of the sequences aligned at a greater distance from an annotated open reading frame; so these have been labelled ‘intergenic sequence’ or ‘telomeric sequence’ as appropriate, although in some cases a strong signal on microarray hybridization suggests the sequence is transcribed.
Microarray hybridization
Production of differential probes
cDNA was prepared using 20 μg of total RNA spiked with alien RNAs (Stratagene), 1 μg oligo (dT)20 primer, 10 mm each of dATP, dCTP, dGTP, 6 mm dTTP, 4 mm amino-ally-dUTP, 9 mm DTT and 400 units Superscript II reverse transcriptase (Invitrogen) in the provided reaction buffer. The RNA and primer were incubated at 70°C for 5 min and snap-chilled on ice before other components of the reaction were assembled. The reaction was incubated at 42°C for 1 h. The residual RNA in the reaction tube was degraded and the reaction neutralized. The cDNA products were purified using MinElute PCR purification kit (Qiagen) and concentrated using a vacuum concentrator. The dried cDNA was resuspended in 5 μl of 0·2 m NaHCO3 coupling buffer, and 5 μl of Cy3 or Cy5 monofunctional dye solution (the dry contents of 1 tube were resuspended in 62 μl of DMSO) was added, mixed and incubated in a dark box for 1 h at room temperature. Subsequently, the fluorescent cDNA was purified using the MinElute PCR purification kit. The cDNA from the reference sample, Pro, was labelled with one fluorochrome and mixed with cDNA from either PA24 or Am labelled with a contrasting fluorochrome.
Hybridization and scanning of the microarray
The 2 fluorescently labelled probes were mixed and hybridized with the microarray in a 40 μl solution containing 3·5× SSC, 0·3% SDS, 10 μg Cot-1 DNA, 4 μg yeast tRNA, 10 μg polyA in a hybridization chamber for 16 h at 65°C. The hybridized microarrays were washed at room temperature for 2 min in 2× SSC/0·1% SDS, and for 10 min each in 1× SSC, 0·2× SSC and 0·05× SSC. The arrays were spun-dried, scanned in a laser scanner (Axon 4100A) and visualized using GenePixPro5.0 software. Replicate experiments with 3 biological preparations comparing either PA24 or Am with Pro samples were performed. A dye-swap experiment was performed to eliminate dye bias in differential expression.
Analysis of microarray data
The individual features on the microarray were manually examined to assess their quality and those exhibiting a sum of Cy3 and Cy5 intensity value lower than the control features included on the array (<1500) or those having saturation intensity or poor quality were flagged and discarded from further analysis. Local background was subtracted from the intensity value of each feature on the array. The relative intensity of the two signals identified clones expressed differentially at either PA24 or Am stage in comparison with Pro stages. Analysis was carried out using statistical analysis software package Acuity 3.1 and MS Excel. The microarray data were normalized based on medians of ratios as well as by Z-score transformation as described by Cheadle et al. (Reference Cheadle, Vawter, Freed and Becker2003). In order to compare expression levels between RNAs, the medians of Cy5/Cy3 ratios were calculated from the normalized values of replicate arrays. The ratios from reverse-labelled experiments were reciprocated before analysis. Clones showing consistent ratios in at least 8/9 spots in replicate experiments were selected for further analysis. These fold changes were relative to the amount of RNA in the cell. For calculation of Z-ratios of dye-flip experiments, the ratios were inversed before analysis. Z-scores were calculated by subtracting the average gene intensity within 1 array replicate from the trimmed mean scaled intensity and then dividing the result by the standard deviation of all intensities. This step corrects the data internally in each hybridization and expresses the intensity values as units of standard deviation away from the mean of the array, which is set to zero, while still reflecting the quantitative integrity of the data. Z-ratios were calculated by taking the difference of the average of the replicate Z-scores between 2 samples and then dividing by the standard deviation of all the averaged Z-score differences for that particular comparison. Average Z-scores were also subjected to sample comparisons by grouping the replicates by sample and then applying the t-test comparison measure. Only 2 samples were compared in any one analysis (Pro versus PA24 or Pro versus Am). Each comparison produced a single P-value for each gene, which were corrected for multiple testing by Bonferroni's correction with an error rate of 0·05.
The microarray work complied with MIAME standards and all details of microarray chip construction, hybridization, scanning and analysis protocols have been submitted to Arrayexpress with the Accession number E-MEXP-866 (http://www.ebi.ac.uk/aerep/login. Reviewer account: Login: Reviewer_E-MEXP-866, Password: 1161778431481).
Nucleotide sequence analysis of selected clones
Single-pass automated sequencing of the plasmid DNA that contained the Leishmania clones identified on the microarray was performed. The sequence generated was sufficient to find the homologous regions of the L. major and L. infantum genomes by BLAST searches of the databases (www.geneDB.org). All the single-pass sequences of L. donovani clones were submitted to the NCBI Genbank-GSS database and Accession numbers are listed in Tables 1 and 2.
Table 1. Clones upregulated at PA24 stage
a The sequences of clones were BLAST analysed in L. major and L. infantum gene databases. All clones aligned with the respective L. major sequences with an E-value of 10−40 or smaller.
§ Clones showing transiently higher expression at PA24 stage.
* Clones showing different alignment in L. infantum database.
15B2 – calpain-like protein, 63C5 – Zn finger domain-like protein.
# Classified as intergenic region as the distance between two ORFS is large and therefore no ORF could be assigned to these DNA clones.
Table 2. Clones down regulated at PA24 stage
a The sequences of clones were BLAST analysed in L. major and L. infantum gene databases. All clones aligned with the respective L. major sequences with an E-value of 10−40 or smaller.
* Clone 23B8 aligns to peptide chain release like protein in L. infantum database.
# Clones showing transient down regulation at PA24 stage with >1·5-fold higher expression at Am stage.
Northern blot analysis
In selected clones, the differential expression in each of the Leishmania life-stages was validated using Northern hybridizations. Total RNA (10 μg) was size fractionated on agarose gel and blotted onto nylon membrane. Digestion of pZeRO vector with XbaI and HindIII released an insert of 1–1·5 kb that was labelled with 32P and used as probe. The hybridization was carried out at 42°C. The membranes were washed in a solution containing 0·2× SSC/0·1% SDS at 65°C for 15 min, exposed at −70°C to an X-ray film for 5 days and the signal was quantitated using AlphaImager 5.5.
RT-PCR
First strand cDNA was prepared for reverse transcriptase-PCR using 5 μg total RNA. The RNA and 1 μg oligo (dT)20 primer were incubated at 70°C for 5 min and snap-chilled on ice and 10 mm each of dATP, dCTP, dTTP, dGTP, 9 mm DTT and 200 units Superscript II reverse transcriptase (Invitrogen) were added in the provided reaction buffer. The reaction was incubated at 42°C for 1 h. Amplification reactions were conducted with gene-specific primers or with primers for the constitutively expressed housekeeping gene α-tubulin for normalization of amounts of RNA. Control experiments were performed with amplification using total RNA but no reverse transcriptase to check for DNA contamination.
RESULTS
Changes in mRNA abundance during differentiation of promastigotes into amastigotes were examined by genome-wide expression profiling using genomic DNA microarrays. Initially, the expresson of amastigote specific gene A2 was evaluated by RT-PCR in all 3 stages of Leishmania. The expression of A2 was evident in Am stage whereas in Pro stage it was not detectable (data not shown). To incur an estimate of accuracy and precision of the system, microarray hybridization was carried out comparing Cy3-labelled promastigote RNA to itself labelled with Cy5, which showed gene expression changes with ratio >2·0 fold in 15/4224 (∼0·35%) clones indicating little or no dye bias or high ratios due to random variation in the technique (data not shown).
Replicate experiments with 3 biological preparations were performed comparing PA24 or Am with promastigotes. A scatter-plot of a representative hybridization comparing the expression in PA24 and Pro following LOWESS normalization is shown in Fig. 1. Normalization was carried out based on the premise that most genes on the array are not differentially expressed; therefore, the arithmetic mean of the ratios from every feature on the array is equal to 1. From the scatter-plot, it is seen that the majority of the log ratio values are clustered close to zero reflecting no change in expression as expected, with outliers representing the differentially expressed genes. Similar plots were obtained in all the hybridizations comparing promastigote with PA24 or amastigote RNA expression (data not shown).
Fig. 1. Scatter-plot comparing the log ratios against the average of the Cy3 and Cy5 intensities for each feature following LOWESS normalization using Acuity software. PA24 total RNA was used to synthesize Cy3 labelled cDNA and mixed with promastigote cDNA labelled with Cy5. The average intensity (A) of each spot is plotted along the x-axis against the LOWESS normalized log ratio (M) along the y-axis. The vertical line defines the cut off used and only spots to the right of this line were considered for analysis.
Analysis of microarray experiments revealed a number of DNA clones showing differential expression in PA24/Pro and Am/Pro microarrays. Of the 4224 genomic DNA clones, those showing ⩾2-fold differential expressions in either of Leishmania life-cycle stages were considered for further analysis. At this cut off value, there were 1·82% (77/4224) clones in PA24 vs Pro and 5·04% (213/4224) clones in Am vs Pro microarrays showing differential expression. Of these, the clones showing significant and consistently higher expression with ratio ⩾2·0 in at least 8/9 spots (s.d. <1), Z-ratio >1·9 and P-value <0·05 in 3 microarray hybridizations and reproducibility in dye-flip microarray experiments were chosen for further analysis. Z-ratios are a direct measure of the likelihood that an observed change is an outlier in an otherwise normal distribution and are independent from their underlying intensity values. All the raw data have been submitted to ArrayExpress (Accession no. E-MEXP-866).
Fifty-five clones ranking top in the fluorescence intensities with ratio ⩾2·0 and reproducibility in replicate experiments were selected and sequenced for further analysis. Of the 55 clones, 28 were upregulated at the intermediate PA24 stage (Table 1) while 27 were found to be down-regulated at the PA24 stage (Table 2). The identities of these genomic clones were assigned by homology to regions in the ORF, 5′UTR (up to ∼500 bp) or in 3′UTR (within 1·5 kb) of known Leishmania genes. The identities of these selected clones, their NCBI-GSS Accession numbers, mean expression ratios from 3 replicate arrays along with standard deviation, Z-ratios and the gene ontology category to which they belong are given in Tables 1 and 2. In addition to identifying homologues in L. major, the clone sequences were also searched by BLAST in the L. infantum genome. Except for 3 clones indicated in the tables, all other clones matched homologous genes in the 2 databases. Based on their identities, putative functions were assigned by homology to proteins of known function. The genes coding for proteins involved in basic cell metabolism for maintenance of life were grouped as maintenance genes whereas those coding for surface proteins were classified as surface genes. Kinases/phosphatases form the category of genes involved in the signalling cascasde and HSP/Chaperonins represent the chaperone category. The category, intergenic region, consists of the clones that are homologous to a region in L. major genome which does not code for any protein as they are located outside the coding or untranslated region (5′ or 3′UTR) of mRNA.
It is evident from Tables 1 and 2 that among the 28 clones overexpressed at PA24 stage, 15 clones showed further increase in their expression during the differentiation into Am. The expressions of 13 clones showed a transient increase at PA24 while their expression declined in Am although the expressions in Am stage remained significantly higher compared to Pro. Protein kinases and HSP10 were found to follow this pattern.
Among the 27 clones showing 2-fold down regulation in PA24 stage compared to Pro stage, 10 clones were under-expressed at PA24 and Am stages. In 13 clones, although the expression was low in PA24, their expression was regained as the parasite fully differentiated into Am. Of particular interest were 4 other clones which showed transient decrease at PA24 stage while their expression level increased by >1·5-fold in Am in comparison to Pro. The identity of these clones revealed them to be surface molecules such as antigen proteins and amino acid transporters.
Analysis of the differentially expressed clones in PA24 vs Pro and Am vs Pro microarray revealed 5 different patterns of gene expression (Fig. 2A–E). Figure 2A shows the expression profiles of 13 clones that increased transiently at PA24 stage when compared to Pro and Am. Of the clones with increased expression at the PA24 stage, 4 maintained the same level of expression as they further differentiated to Am (Fig. 2B). Eleven clones showed continuous increase in their expression levels as the parasites fully transformed into Am (Fig. 2C). The expression pattern of 17 clones showing transient down regulation at PA24 stage is shown in Fig. 2D. Of these, 3 clones under expressed at PA24 by −4-fold were significantly overexpressed in Am compared to Pro. The fifth pattern of gene expression consisted of 10 clones showing down regulation at both PA24 and Am stages (Fig. 2E). The clones are grouped by gene ontology to facilitate interpretations about the correspondence between the pattern of gene expression changes with differentiation and the gene functions which follow that pattern.
Fig. 2. Graphs showing various patterns of gene expression observed among clones upregulated at the PA24 stage. The log2 expression ratio in PA24/Am vs Pro is plotted at each time-point. (A) Expression pattern of clones showing transient higher expression at PA24 stage. (B) Expression pattern of clones showing uniform expression at PA24 and amastigotes compared to promastigotes. (C) Expression patterns of clones showing continuous increase in their expression during differentiation of promastigotes into amastigotes. (D) Expression pattern of clones showing transient down regulation at PA24 stage. (E) Expression patterns of clones showing down regulation at both PA24 and Am stages.
The expression changes in representative clones from various categories were verified by RT-PCR and Northern hybridizations (Fig. 3). The gene expression changes in 5 clones representing different categories were validated by RT-PCR in 3 different patient isolates of L. donovani (Fig. 3A) and 5 other clones were tested in Northern blots in 2 patient isolates (Fig. 3B). Northern blots and RT-PCR with different parasite lines gave similar expression levels. Although the fold changes observed by Northern blots and RT-PCR were different from those seen in microarray results, the expression patterns were found to be similar to microarray results. PCR reactions were also carried out with parasite RNA without a reverse transcriptase step for all of the above-mentioned primer sets, but no amplification was seen with RNA indicating the absence of DNA contamination in the RNA samples.
Fig. 3. Validation of microarray results by RT-PCR and Northern blots. Expression pattern was verified in 3 different field isolates of Leishmania donovani and representative picture with the RNA of the parasite used in the microarray is shown. The RNAs from the 3 stages were reverse transcribed and amplified by gene-specific primers using α-tubulin for normalization (A). Ethidium bromide-stained rRNA bands were used as loading controls in Northern blots (B). Fold changes at the 3 stages are indicated.
DISCUSSION
The present study is focused on differences in gene expression that can be seen as the visceral parasite, L. donovani, differentiates from the life-cycle stage found in the insect vector (Pro) into the human infectious stage (Am) in culture. Identification of genes whose expression changes during differentiation could help reveal the mechanism of Leishmania gene regulation and identify targets for controlling the diseases caused by the human pathogen. Such studies are facilitated by the availability of a host-free promastigote to amastigote differentiating system for L. donovani (Debrabant et al. Reference Debrabant, Joshi, Pimenta and Dwyer2004). The differential expression of amastigote specific gene A2 gene in axenic amastigotes confirmed that the in vitro-generated axenic amastigotes mimic the true amastigotes. Gene expression in Pro, PA24 (an intermediate stage of differentiation) and axenic Am stages were compared using L. donovani genomic microarrays. DNA microarrays either in the form of random genomic fragments or cDNA have proven their utility in identifying differentially expressed genes in Leishmania (Saxena et al. Reference Saxena, Worthey, Yan, Leland, Stuart and Myler2003; Holzer et al. Reference Holzer, McMaster and Forney2006). Leishmania microarray studies carried out previously have been largely confined to L. major, comparing the procyclic stage with metacyclic (Saxena et al. Reference Saxena, Worthey, Yan, Leland, Stuart and Myler2003; Akopyants et al. Reference Akopyants, Matlibs, Bukanova, Smeds, Brownstein, Stormo and Beverley2004) or with amastigote stage (Akopyants et al. Reference Akopyants, Matlibs, Bukanova, Smeds, Brownstein, Stormo and Beverley2004; Holzer et al. Reference Holzer, McMaster and Forney2006). Genomic arrays have the advantage over cDNA arrays that the genes represented on the array are not biased by the expression levels in the RNAs used for microarray construction. Therefore to screen for expressed genes in an unbiased way we utilized a microarray constructed of random genomic fragments. Furthermore, to insure representation of important virulence-associated genes, we prepared the genomic clones from parasite DNA obtained from a strain derived from a kala-azar patient, after a minimum number of in vitro passages. Earlier studies on transcriptome analysis have revealed only a low level of gene expression changes during the Leishmania life-cycle, accordingly clones showing ⩾2-fold differential expression at either of the stages were chosen for further analysis.
Previous observations (Duncan et al. Reference Duncan, Alvarez, Jaffe, Wiese, Klutch, Shakarian, Dwyer and Nakhasi2001) have suggested 24 h to be a relevant time-point at which to study the events during the differentiation process in Leishmania. DNA microarrays identified several differentially expressed genes. In comparison with the promastigote stage, the number of transcripts showing up or down regulation was 3-fold higher in Am than in the PA24 stage, suggesting gene alteration to be more prominent in the two extreme life-cycle stages where cell maintenance and survival processes are more active than at the intermediate PA24 stage where the cell's primary activity is differentiation.
Our study highlights the importance of studying an intermediate stage during Leishmania differentiation where some of the genes are transiently up or down regulated. The expression of a variety of genes was found to be modulated, the prominent ones being the surface molecules, kinases and maintenance genes. Earlier studies have revealed the upregulation of amastin, protein kinases and phosphomannomutase in the amastigotes of L. major (Akopyants et al. Reference Akopyants, Matlibs, Bukanova, Smeds, Brownstein, Stormo and Beverley2004) and L. mexicana (Holzer et al. Reference Holzer, McMaster and Forney2006) as compared to the promastigotes.
Of the over-expressed genes, the putative signal-transducing molecules such as protein kinases (clones 28F11, 31C2, 45F10, 63C8) showed higher expression during the first 24 h than that observed in fully differentiated amastigotes. This pattern is consistent with the PA24 stage as a transitional stage where signalling molecules (kinases) are most active in regulating the process of differentiation. This inference is also consistent with lower expression of kinases in the fully differentiated amastigotes. In eukaryotes, the signal transducers mediate phosphorylation and dephosphorylation of serine, threonine and tyrosine residues in proteins and function as control switches in cellular networks. MAP kinase in Leishmania has been shown to be required for the maintenance of a full-length flagellum, promastigote shape as well as cell growth (Wiese, Reference Wiese1998; Weise et al. Reference Wiese, Kuhn and Grünfelder2003; Erdmann et al. Reference Erdmann, Scholz, Melzer, Schmetz and Wiese2006). The transient increase in the expression of these protein phosphorylation molecules, which would otherwise be undetected in studies with the two extreme life-cycle stages, suggests an important role of these enzymes in the differentiation process that may contribute to virulence.
Apart from protein phosphorylation, another important phenomenon that occurs during the temperature increase as the parasite travels from the insect to the mammalian host is the production of various heat shock proteins (HSPs) that may protect the parasite against the adverse effects of elevated temperature by chaperoning proteins. Earlier studies have reported L. donovani CPN10/HSP10 to be expressed preferentially under heat stress and in axenic amastigotes unlike Hsp70 and Hsp90 that show high constitutive abundance and only a marginal increase in the amastigote stage (Brandau et al. 1995; Zamora-Veyl et al. Reference Zamora-Veyl, Kroemer, Zander and Clos2005). The recently published protein profiling study with L. panamensis promastigotes and axenic amastigotes revealed over expression (1·8-fold) of heat shock protein 83 in the amastigote stage, which is similar to our findings at transcriptome level (Walker et al. Reference Walker, Vasquez, Gomez, Drummelsmith, Burchmore, Girard and Ouellette2006). In our array data, CPN10 (clone 28F12) and HSP83 (clone 16C11) were upregulated in PA24 as well as Am stages with CPN10 transiently higher at PA24 whereas HSP83 expression was elevated at PA24 and Am to similar levels. This indicates a more significant role of CPN10 during the early stages of differentiation.
The expression of surface genes amastin (clones 29B8 and 14D7) and proteophosphoglycan (PPG, clone 18D8) increased progressively during the differentiation process of the promastigotes into amastigotes, suggesting a more important role in fully differentiated amastigotes. These genes are known to be expressed specifically at the amastigote stage in Leishmania (Piani et al. Reference Piani, Ilg, Elefanty, Curtis and Handman1999; Boucher et al. Reference Boucher, Wu, Dumas, Dube, Sereno, Breton and Papadopoulou2002) and are thought to play a role in the intracellular survival and pathogenesis in the mammalian host.
The expression of PSA (clone 42A11), surface antigen protein 2 (49A11), amino acid permeases (37G1) and a hypothetical protein (clone 61C9) were unique compared to other patterns of expression of genes observed in this study. The 4 genes were found to be highly expressed in Pro as well as Am stages but their expression at PA24 stage was low. The 3 cell surface protein genes that are preferentially expressed in amastigotes although significantly down regulated at the intermediate stage (amino acid permease, PSA2 and surface antigen protein 2 which showed 1·8–3·2-fold higher expression at amastigotes and ∼4-fold down regulation at intermediate PA24 stage in comparison with promastigotes) suggest an increase in complexity of the cell surface that is at the interface with the harsh environment of the phagolysosome. The increased expression of amino acid permeases in Leishmania promastigotes and amastigotes may be due to the nutritional requirement of the parasite for its survival as the metabolism of Leishmania parasites is strongly based on amino acid consumption and the amino acid permeases are required for the uptake of amino acids (Geraldo et al. Reference Geraldo, Silber, Pereira and Uliana2005). However, since at the PA24 stage there is limited metabolism taking place inside the cell, the expression levels of amino acid permeases fall to basal levels. Promastigote surface antigen (PSA) of Leishmania has previously been shown to increase about 30-fold as in vitro-cultured parasites progress from logarithmic to stationary phase, growth phases that are, respectively associated with parasites having low and high infectivity to mammals (Lincoln et al. Reference Lincoln, Ozaki, Donelson and Beetham2004; Beetham et al. Reference Beetham, Donelson and Dahlin2003). These antigenic proteins might have a role in immune evasion during insect to mammal transition, but this function is not required in the initial infection process that coincides with the first 24 h of differentiation. However, the exact reason for this transient low expression at PA24 remains unclear and further studies are required in this direction.
The membrane-bound kinesin (clone 23G11) and V-type ATPases (clone 44D6) were found to be over expressed in Pro in our microarray analysis. Kinesins perform a variety of cellular functions that include cell division, signal transduction, microtubule dynamics and trafficking of macromolecular complexes and organelles including mitochondria, lysosomes and synaptic vesicles (Rodrigues et al. Reference Rodrigues, Scott and Docampo1999). V-type ATPases are present in acidocalcisomes of Leishmania and play a role in pH homeostasis. This molecule is believed to protect the parasites from osmotic shock inside the sandfly. The expression of these genes correlated with the known functions of these genes at the two stages (Dutoya et al. Reference Dutoya, Gibert, Lemercier, Santarelli, Baltz, Baltz and Bakalara2001).
From our observation, among the 55 identified genes, 12/27 (∼47%) of the PA24 down-regulated genes were maintenance genes. The maintenance category included some of the enzymes required for the cell's metabolism and maintenance and none of the maintenance genes were found to be transiently upregulated suggesting some metabolic functions are put ‘on hold’ while the cell undergoes the transition. In fully differentiated amastigotes, a substantial number of maintenance genes were down regulated while other genes were upregulated. As the parasite differentiates from Pro to Am, there is a shift in the metabolic pathway due to differences in the environment and consequent change in nutrient availability at different stages. Thus, a large turnover in the expression of these enzymes during the process of differentiation is not unexpected. However, the increased expression of the enzyme NAD/FAD dependent dehydrogenase (clone 29C8) in PA24 and Am stages and relatively very low expression in Pro suggests this molecule to be an extremely important molecule involved in the parasite's survival in the mammalian host. There are indications that this enzyme might be involved in the activation of ubiquitin-activating enzyme E1 involved in iNOS degradation in macrophages (Kolodziejski et al. Reference Kolodziejski, Musial, Koo and Eissa2002; Yoshida and Xia, Reference Yoshida and Xia2003; Mitani et al. Reference Mitani, Terashima, Yoshimura, Nariai and Tanigawa2005). Nevertheless, more studies with this molecule are needed to establish this fact.
From our array work, 16/55 identified genes were identified as hypothetical genes that showed differential expression at either of the Leishmania life-cycle stages. This can be explained by the fact that a large proportion of the differentially expressed genes comprised of hypothetical proteins indicating the extent of genetic difference between this parasite and the other well-studied organisms. In the recently completed L. major genome 4673 (>50%) out of 8370 coding sequences are conserved hypothetical proteins (www.geneDB.org; Ivens et al. Reference Ivens, Peacock, Worthey, Murphy, Aggarwal, Berriman, Sisk, Rajandream, Adlem, Aert, Anupama, Apostolou, Attipoe, Bason, Bauser, Beck, Beverley, Bianchettin, Borzym, Bothe, Bruschi, Collins, Cadag, Ciarloni, Clayton, Coulson, Cronin, Cruz, Davies, De Gaudenzi, Dobson, Duesterhoeft, Fazelina, Fosker, Frasch, Fraser, Fuchs, Gabel, Goble, Goffeau, Harris, Hertz-Fowler, Hilbert, Horn, Huang, Klages, Knights, Kube, Larke, Litvin, Lord, Louie, Marra, Masuy, Matthews, Michaeli, Mottram, Muller-Auer, Munden, Nelson, Norbertczak, Oliver, O'neil, Pentony, Pohl, Price, Purnelle, Quail, Rabbinowitsch, Reinhardt, Rieger, Rinta, Robben, Robertson, Ruiz, Rutter, Saunders, Schafer, Schein, Schwartz, Seeger, Seyler, Sharp, Shin, Sivam, Squares, Squares, Tosato, Vogt, Volckaert, Wambutt, Warren, Wedler, Woodward, Zhou, Zimmermann, Smith, Blackwell, Stuart, Barrell and Myler2005) meaning no homologue that is functionally characterized has been found in other organisms. Characterization of some of these genes may help in studying the processes involved in Leishmania pathogenesis.
While this manuscript was in the final stages of preparation, another group published a similar study of gene expression in in vitro-differentiating L. donovani, using a microarray comprising of L. major genomic DNA. Of the 55 clones identified in our study, only 4 clones were common with the differentially expressed genes reported in their study (Saxena et al. Reference Saxena, Lahav, Holland, Aggarwal, Anupama, Huang, Volpin, Myler and Zilberstein2007). Our study using L. donovani genomic microarray with a virulent, well-characterized field isolate of L. donovani adds 51 genes to the list of differentially expressed genes reported by Saxena et al. (Reference Saxena, Lahav, Holland, Aggarwal, Anupama, Huang, Volpin, Myler and Zilberstein2007) in which RNA from a laboratory strain of L. donovani was hybridized to L. major genomic array. In common with their study, the 4 genes, amastin-like protein, PPG4, PSA2 and HSP83 showed similar trends in the ordered progression of expression changes with differentiation. As per the nature of microarray comprising of random genomic clones, different sets of genes were identified in the 2 studies, although the gene ontology categories showed similar trends in the ordered progression of expression changes with differentiation.
In summary, we have demonstrated the utility of L. donovani genomic microarrays for the study of early events occurring during Leishmania differentiation. The work presented here represents the genome-wide analysis of changes in gene expression of a virulent strain of L. donovani during the transition from promastigotes to amastigotes, a process which is impossible to follow until host-free differentiation systems have been developed. We report the significance of studying early stages of differentiation in which we have found a few genes to be transiently modulated that would not have been observed while studying the two extreme stages. Our array identified several differentially expressed molecules that might enable the parasite to survive within the phagolysosome of vertebrate macrophages and be transmitted by sandfly vectors. We intend to pursue the functional study of some of the identified genes that might provide useful insights in Leishmania survival and pathogenesis.
Financial support for the study, provided by the Indo-US VAP funded by DBT, India is gratefully acknowledged. S. G. is grateful to the Council of Scientific and Industrial Research, India, for providing a fellowship.
