The Leishmaniases are widespread parasitic diseases, with about 360 million people at risk (Desjeux 1996). Unfortunately, anti-leishmanial drugs, mainly based on antimonial therapy, are toxic, and vaccines, recently developed and tried, showed a relatively low protection under field conditions (Kubar and Fragaki 2005).

The Leishmania are digenetic life cycle parasites that alternate between two stages: flagellated promastigotes, which develop in the midgut of the insect vector, and amastigotes, which multiply in the host macrophage (Alexander and Russell 1992; Handman 1999). At this latter stage, Leishmania parasites are sequestered within the phagolysosomal vacuoles, which result from the fusion of phagosomes with lysosomes (Amer and Swanson 2002; Cunningham 2002; Sacks 2001; Duclos and Desjardins 2000; Handman 1999). The two developmental stages display distinct morphologic and metabolic characteristics consistent with a tightly regulated differential expression of parasitic proteins. The identification of stage-specific genes would constitute an important step for understanding the biology of the parasite and for the development of new drugs or vaccines targeting the product of these genes (Stober 2004).

The recent publication of the whole genome sequence of Leishmania major has led to several studies aiming at the identification of stage-regulated genes. Several methods were used that target either the messenger RNA (mRNA) or the protein levels. Studies, mainly using the DNA microarray technique (Holzer et al. 2006; Akopyants et al. 2004; Almeida et al. 2004; Saxena et al. 2003; Almeida et al. 2002), indicate that hundreds of genes are likely to be developmentally regulated. However, the confirmation of their stage-specific expression requires a more precise technique. Relative quantification of mRNA using real-time reverse transcription–polymerase chain reaction (RT-PCR) is one of the most accurate techniques available to date for the analysis of gene expression (Ding and Cantor 2004). Usually, expression of a target gene is related to the stable expression of a housekeeping gene simultaneously determined in the same sample (Huggett et al. 2005). As Leishmania alternate between different developmental stages, a stably expressed reference gene is needed for the identification of stage-regulated genes. The aim of this study is to select from a panel of L. major potential reference genes, those that are the most stably expressed during the various parasite developmental stages, to use them as reference in real time RT-PCR studies.

Six sequences were selected (Table 1). One of them, N-myristoyltransferase (NMT), was reported as constitutively expressed at all stages of L. major parasites (Price et al. 2003). The five other sequences [pteridine transporter (PT), rRNA45, rRNA46, NH39, and NH78] were selected because they were shown to be stably expressed using microarray analysis (Akopyants et al. 2004). Small hydrophilic endoplasmic reticulum associated protein (SHERP), a gene previously reported as metacyclic specific, was used as a control stage-specific sequence (Knuepfer et al. 2001).

Table 1 Primers used in real time RT-PCR experiments
Full size table

Three L. major strains were used: two isolates (Zymodeme MON25, MHOM/TN/94/GLC94 and MHOM/TN/94/GLC32), obtained from human zoonotic cutaneous leishmaniasis lesions (Louzir et al. 1998; Kebaier et al. 2001) and a clone derived from GLC94 (cGLC94). For each strain, three parasite stages were compared for gene expression analysis: procyclics, metacyclics, and amastigotes. Procyclic promastigotes were collected at the logarithmic-growth phase culture in complete medium consisting on RPMI-1640 (Sigma, St. Louis, MO, USA) containing 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, and 10% heat-inactivated fetal calf serum. For differentiation into metacyclics, a fresh sub-culture was grown, and a growth curve was plotted to ascertain when the cells entered stationary phase (between days 6 and 7). Metacyclics were then purified from stationary phase cultures using Ficoll density gradient centrifugation as previously described (Spath and Beverley 2001). Amastigotes were isolated from L. major-harboring skin lesions of BALB/c mice as described previously (Barral et al. 1983), then purified over a discontinuous gradient centrifugation (Jaffe and Rachamim 1989). For each L. major strain and parasite stage (procyclic and metacyclic promastigotes and amastigote), at least three RNA extractions were performed from different parasite culture. Total RNA was extracted by homogenization in 1 ml TRIzol reagent (Gibco-BRL Life Technologies, Cergy Pontoise, France) according to the manufacturer’s instructions. Any possible genomic DNA contamination was eliminated by treatment with deoxyribonuclease I (Dnase I, Boehring Mannheim, Mannheim, Germany). RNA was quantitated using a spectrophotometer. Examination of purified total RNA by gel electrophoresis revealed prominent 18S, 24Sα, and 24Sβ ribosomal bands, indicating that the RNA was not degraded. To minimize the eventual RNA variability, RNA preparations from the same strain and parasite stage were pooled, and 2 μg of the pooled RNA were reverse transcribed as previously described (Chenik et al. 2005) and used for real-time PCR experiments.

Primers, two forward and two reverse, were designed for each sequence using Primer express software (version 1.5, PE Applied Biosystems, Foster City, CA, USA). The optimal upstream and downstream primer sequences for the different genes are shown in Table 1. The experiments were performed using SYBR green I Universal PCR MasterMix (PE Applied Biosystems), and PCR reactions were performed using the ABI Prism 7700 sequence detection system (with version 1.9.1 software, PE Applied Biosystems) in a standard program of 40 cycles (Chenik et al. 2005). For each PCR reaction, the efficiency value (Eff) was calculated from the standard curve based on tenfold serial dilutions of parasite genomic DNA (Table 1). The PCR efficiency ranges from 86 to 98%. All experiments were performed twice using two different pools of parasite RNA preparations.

For all selected reference genes and for each determination, experiments were performed in triplicate. The mean Ct-values were then used for the analysis of gene expression stability using two programs: GeNorm version 3.4 (Vandesompele et al. 2002) and NormFinder (Andersen et al. 2004).

GeNorm relies on the principle that the expression ratio of two perfect reference genes should be identical regardless of the experimental condition or cell type (Vandesompele et al. 2002).

The Ct values for each strain are first transformed into relative quantities using PCR efficiencies as follows: Q = E (min Ct−sample Ct) in which min Ct corresponds to the lowest Ct value between the three developmental stages and sample Ct corresponds to the Ct in the considered stage of the parasite. For each reference gene, the stability measure (M) is estimated by calculating the average pair-wise variation for one gene against all the other tested genes. Those with the lowest M values have the most stable expression among the three strains and the three developmental stages of the parasite. Hence, by a stepwise exclusion of the gene with the highest M value and recalculation of the M values, the combination of the most stable expressed genes is selected. Finally, a normalization factor (NF) is generated over the most stable genes by calculating the geometric mean of their relative quantities.

In addition, pair-wise variations (Vn/n + 1) are calculated for every series of NFn and NFn + 1 to determine the effect of adding a (n + 1) gene in normalization. A large pair-wise variation means that the added gene has a significant effect on normalization and, thus, should preferably be included for calculation of a reliable normalization factor (Vandesompele et al. 2002).

Hence, to study the possible inter-strain variations not considered in GeNorm, we performed additional analyses using Normfinder. In this program, the stability value is based on the combined estimate of intra- and inter-group expression variations of the studied genes. A low gene stability value indicating a low combined intra- and inter-group variation proves high expression stability of this gene (Andersen et al. 2004).

The ranking of the six tested genes according to their M value (GenNorm) was (from the least stable to the most stable): NMT, rRNA46, PT, NH78, NH39, and rRNA45 (Fig. 1a). It is interesting to note that a similar order, in terms of expression stability, was obtained when using Normfinder program (Fig. 1a). Indeed, using the two programs, rRNA45 emerged as the most stable gene in the three developmental stages of the parasite and, thus, can be used as a reference gene in gene expression studies in L. major.

Fig. 1
figure 1

Selection of reference gene in different Leishmania developmental stages. a Stability ranking of control genes tested using GeNorm (line) and NormFinder (bars). For GeNorm, the expression stability value (M) was calculated for each gene, and the least stable gene with the highest M value was automatically excluded for the next calculation round. Genes are ranked from left to right in order of increasing expression stability, as indicated by lower M values. For NormFinder, high expression stability is indicated by a low stability value as an estimate of the combined intra- and inter-group variation of the individual gene. b Pair-wise variation (Vn/n + 1) analysis between normalization factors NFn and NFn + 1 to determine the optimal number of internal control genes required for reliable normalization. The genes under consideration are rRNA45, NH39 NMT, PT, rRNA46, and NMT ranked according to their expression stability (from the most to the least stable). V2/3 Pair-wise variation of normalization factor obtained when using the two best reference genes compared to normalization factor obtained when using the three best reference genes; V3/4 pair-wise variation of normalization factor obtained when using the three best reference genes compared to normalization factor obtained when using the four best reference genes; V4/5 pair-wise variation of normalization factor obtained when using the four best reference genes compared to normalization factor obtained when using the five best reference genes; V5/6 pair-wise variation of normalization factor obtained when using the five best reference genes compared to normalization factor obtained when using the six best reference genes

Full size image

Although most studies used only one single gene as an internal control for normalization, it has been suggested that the use of more than one gene might generate more reliable results (Vandesompele et al. 2002). We therefore evaluated the optimal number of genes required for accurate normalization using the GeNorm program.

To estimate the benefit of using additional (n + 1) control gene, pair-wise variation (Vn/n + 1) was calculated between consecutively ranked normalization factors. As a significant increase between V2/3 (0.108) and V3/4 (0.120) is obtained (Fig. 1b), we consider that use of the four most stable genes (rRNA45, NH39, NH78, and PT) for normalization should allow a more accurate normalization.

To validate the reference genes identified in this study, we compared the gene expression profiles of a stage-specific gene (SHERP) in the three L. major strains using either rRNA45 alone as reference gene or the combination of the four most stable reference genes (Fig. 2).

Fig. 2
figure 2

Expression profile of SHERP in the three developmental stages of L. major. Gene expression experiments were conducted on cDNAs prepared by reverse transcription of total RNAs purified from the three L. major strains S1 (GLC94), S2 (GLC32), and S3 (cGLC94) in the three developmental stages of the parasite (procyclic, metacyclic, and amastigote). Normalizations were made either by the most stable reference gene (rRNA45) or by a combination of the four most stable reference genes (NH39, rRNA45, NH78, and PT)

Full size image

Relative expression of SHERP in the different developmental stages of the parasite shows the same general trend in the three strains using either rRNA 45 or the four reference genes for normalization. Indeed, SHERP is overexpressed in the metacyclic stage of the parasite compared to the procyclic and the amastigote stages. The most important differences in terms of expression were found between the metacyclic and amastigote stage, in agreement with previous works (Almeida et al. 2004; Knuepfer et al. 2001). However, the use of four reference genes seems to slightly reduce the ratio of relative expression.

In conclusion, the proper selection of a reference gene in gene expression studies targeting the different developmental stages of L. major is crucial to generate data with biological significance. In this study, the analysis of a panel of different potential candidate reference genes allowed us to identify a suitable reference gene (rRNA45), characterized by a stable expression through the various developmental stages of L. major. In addition, we demonstrate that normalization with a single reference gene could generate results comparable to those normalized against a combination of the four most stable genes (rRNA45, NH39, NH78, and PT). These results may provide a guideline for future works on gene expression in the different Leishmania stages using real-time PCR.