Influence of the Microenvironment in the Transcriptome of Leishmania infantum Promastigotes: Sand Fly versus Culture

Abstract

Zoonotic visceral leishmaniasis is a vector-borne disease caused by Leishmania infantum in the Mediterranean Basin, where domestic dogs and wild canids are the main reservoirs. The promastigote stage replicates and develops within the gut of blood-sucking phlebotomine sand flies. Mature promastigotes are injected in the dermis of the mammalian host and differentiate into the amastigote stage within parasitophorous vacuoles of phagocytic cells. The major vector of L. infantum in Spain is Phlebotomus perniciosus. Promastigotes are routinely axenized and cultured to mimic in vitro the conditions inside the insect gut, which allows for most molecular, cellular, immunological and therapeutical studies otherwise inviable. Culture passages are known to decrease infectivity, which is restored by passage through laboratory animals. The most appropriate source of promastigotes is the gut of the vector host but isolation of the parasite is technically challenging. In fact, this option is not viable unless small samples are sufficient for downstream applications like promastigote cultures and nucleic acid amplification. In this study, in vitro infectivity and differential gene expression have been studied in cultured promastigotes at the stationary phase and in promastigotes isolated from the stomodeal valve of the sand fly P. perniciosus. About 20 ng RNA per sample could be isolated. Each sample contained L. infantum promastigotes from 20 sand flies. RNA was successfully amplified and processed for shotgun genome microarray hybridization analysis. Most differentially regulated genes are involved in regulation of gene expression, intracellular signaling, amino acid metabolism and biosynthesis of surface molecules. Interestingly, meta-analysis by hierarchical clustering supports that up-regulation of 22.4% of the differentially regulated genes is specifically enhanced by the microenvironment (i.e. sand fly gut or culture). The correlation between cultured and naturally developed promastigotes is strong but not very high (Pearson coefficient R2 = 0.727). Therefore, the influence of promastigote culturing should be evaluated case-by-case in experimentation.

Fig 1. Sand fly gut dissection, in vitro infectivity and mRNA amplification.

(A) Detail of Pro-Pper (40X) within the anterior thoracic midgut (10X). SV: stomodeal valve. Sand flies of an established colony were dissected to extract the whole guts. After that, the anterior part of the thoracic midgut, behind the SV, was separated in a PBS drop and slightly squeezed with a coverslip. Then, the PBS drop containing Pro-Pper in suspension was recovered, thus avoiding carryover of gut tissue as much as possible. (B) The U937 cell line was differentiated with phorbol esters on 8-well chamber slides and in vitro infected with L. infantum Pro-Pper and Pro-Stat (approximately 5 x 10⁴ promastigotes at a phagocyte:promastigote ratio 1:5 were added). The preparations were fixed and stained with modified Giemsa and 100 cells were randomly counted per replicate. The average number of amastigotes per infected cell was measured at 48 h post-infection. Mean ± SD: 2.7 ± 0.4 in the case of Pro-Stat and 4.8 ± 0.9 in the case of Pro-Pper. (C) Agarose gel electrophoresis of aaRNA samples used for the microarray analysis after synthesis of labeled cDNA. Total RNA was purified from isolated Pro-Pper immediately after dissection and doubly amplified (aaRNA) with MessageAmpII aRNA Amplification Kit (Life Technologies), due to sample amount requirements. Pro-Stat RNA was isolated and processed following the same procedure as for Pro-Pper. Five μl aliquots of the aaRNA samples were run at 5 V/cm in a 1.5% agarose gel prepared with RNase-free water after treatment of the electrophoresis cell, tray and comb with hydrogen peroxide. Three biological replicates of the microarray hybridization experiment were performed.

Fig 2. M/A scatter plot of the Pro-Pper/Pro-Stat microarray hybridization experiment.

M = (log₂Ri–log₂Gi) and A = [(log₂Ri + log₂Gi)/2], where R and G are, respectively, red (Cy5) and green (Cy3) fluorescence intensity values previously normalized by the LOWESS per pin algorithm. Red spots correspond to selected clones containing at least a 2-fold up-regulated gene and green spots a 2-fold down-regulated one. Only statistically significant differences were selected (Student’s t-test, p<0.05). A fluorescence intensity filter was applied to select these data as well. The dot cloud is not dispersed and it is symmetric about the M = 0 line (lack of differential expression), as expected in a microarray analysis. The Pearson correlation coefficient between normalized fluorescence intensity values of Pro-Pper and Pro-Stat is R² = 0.727. Consequently, both samples are strongly correlated but important differences are also observed.

Fig 3. Multi-level bar graph of GO biological function terms annotated in the differentially regulated genes between Pro-Pper and Pro-Stat.

The GO terms were assigned to the differentially expressed genes between Pro-Pper and Pro-Stat following the pipeline of the BLAST2GO software. First, the BLAST step was run. Then, the GO term assignment step was performed and terms from the KEGG and the InterPro databases were subsequently assigned, as well as the IUBMB EC identifiers. Then, direct acyclic graphs (DAGs) were generated and multi-level sector graphs generated on the basis of those DAGs were retrieved. Finally, the format of the multi-level sector graphs was changed to bar graphs. Accession numbers of GO terms from top to bottom: GO:0015986; GO:0005975; GO:0007049; GO:0008652; GO:0044265; GO:0051276; GO:0051649; GO:0006631; GO:0009064; GO:0018130; GO:0006886; GO:0008610; GO:0034660; GO:0006259; GO:0055114; GO:0043687; GO:0064681; GO:0006508; GO:0009207; GO:0050794; GO:0042221; GO:0050896; GO:0042254; GO:0006396; GO:0006412; GO:0055085; GO:0016192.

Fig 4. Differential expression profiles of Pro-Pper and Pro-Stat.

Protein products in red correspond to genes up-regulated in Pro-Pper and those in green to Pro-Stat. The most relevant changes in transcript relative abundance are discussed in the text. More detailed discussion of all differentially regulated genes is provided in S1 Text. The asterisk indicates that N-acetyl hexosamines are building blocks for PI-bound molecules. Abbreviations not found in the main body of the text: ABC, ATP-binding cassette; AGXT, alanine-glyoxylate transaminase activity; AP3δ1, adaptor complex protein 3 δ subunit 1; coxV, cytochrome oxidase V subunit; CUL, cullin; DECS, sphingolipid δ4-desaturase; DHAK, dihydroxyacetone kinase; EHHADH, 3-hydroxyacyl-ACP dehydratase; erv25, COP-coated vesicle membrane protein erv25; Fe/ZnT, iron/zinc transporter; His-tRNA^S, histidyl-tRNA synthase; mACDH, mitochondrial acyl-CoA dehydrogenase; MAPK, mitogen-activated protein kinase; MTP, mitochondrial carrier protein; mvaK, mevalonate kinase; NSDHL, NAD(P)-dependent steroid dehydrogenase-like protein; NsT1, nucleoside transporter 1; NUP155, nuclear pore complex protein 155; PNPP, p-nitrophenylphosphatase; PT, pteridine transporter; SbGRP, sodium stibogluconate-resistance protein; TGL, triacylglycerol lipase; TPN2, transportin 2; UbqA-E1, ubiquitin activating enzyme E1; UbqC, ubiquitin conjugating enzyme-like protein; UbqC-E2, ubiquitin conjugating enzyme E2; vamp, vesicle-associated membrane protein; vATPSc, subunit c of the vacuolar ATP synthetase; VPSL, vacuolar protein sorting-like protein.

Fig 5. HCL-ST comparison of Pro-Pper with cultured promastigotes and amastigotes of L. infantum.

(A) Pro-Pper vs. Pro-Stat (this study) and Amas [24]. (B) Pro-Stat vs. Pro-Pper (this study), Pro-Log and Amas [28]. A description of clones not found in S4–S8 Tables can be found in [28]. All the microarray hybridization experiments were performed by the same procedure and clone nomenclature is equivalent (see Availability of the Supporting Data). For obvious reasons, when more than one gene is represented by a given clone, independent qRT-PCR analysis was performed. This approach allows determining the actual gene that is differentially expressed in each different biological condition (Tables 2 and 3, S4 and S5 Tables [24,28]).