Retention and loss of RNA interference pathways in trypanosomatid protozoans

Abstract

RNA interference (RNAi) pathways are widespread in metaozoans but the genes required show variable occurrence or activity in eukaryotic microbes, including many pathogens. While some Leishmania lack RNAi activity and Argonaute or Dicer genes, we show that Leishmania braziliensis and other species within the Leishmania subgenus Viannia elaborate active RNAi machinery. Strong attenuation of expression from a variety of reporter and endogenous genes was seen. As expected, RNAi knockdowns of the sole Argonaute gene implicated this protein in RNAi. The potential for functional genetics was established by testing RNAi knockdown lines lacking the paraflagellar rod, a key component of the parasite flagellum. This sets the stage for the systematic manipulation of gene expression through RNAi in these predominantly diploid asexual organisms, and may also allow selective RNAi-based chemotherapy. Functional evolutionary surveys of RNAi genes established that RNAi activity was lost after the separation of the Leishmania subgenus Viannia from the remaining Leishmania species, a divergence associated with profound changes in the parasite infectious cycle and virulence. The genus Leishmania therefore offers an accessible system for testing hypothesis about forces that may select for the loss of RNAi during evolution, such as invasion by viruses, changes in genome plasticity mediated by transposable elements and gene amplification (including those mediating drug resistance), and/or alterations in parasite virulence.

Figure 1. Tests of RNAi pathway activity in L. braziliensis using GFP reporters.

Panel A. Schematic map of the SSU rRNA locus in Leishmania and an example of targeting using the SwaI GFP65-StL fragment derived from the GFP65-StL construct pIR1SAT-GFP65-StL. Regions are the SSU rRNA (gray box), GFP65 ORF (striped arrow), nourseothricin resistance gene ORF (SAT), stem-loop stuffer fragment (black box), and the rRNA promoter (P_rRNA). Panel B. siRNA analysis of WT L. braziliensis M2903 and GFP65-StL transfectants of T. brucei and L. braziliensis M2903 SSU:GFP65-StL, hybridized with a radiolabeled GFP65 probe. The star marks the mobility of a 26 nt standard; CHB is a cross hybridizing band that serves as a loading control. Panel C. GFP flow cytometry of L. braziliensis M2903 transfectants expressing either the AT-rich GFP65* or the GC-rich GFP+ reporters, alone or in combination with a GFP65-StL. Profiles are labeled and color-coded as follows: Black, WT M2903; green, GFP65+GFP65-StL (SSU:GFP65-StL + SSU:GFP65*, clone 8); blue, GFP65 (SSU:GFP65*, clone 10); blue-green, GFP+(SSU:GFP+, clone 38); and purple, GFP65-StL + GFP+ (SSU:GFP++SSU:GFP65-StL, clone 60). Panel D. Northern blot analysis of L. braziliensis M2903 derived lines; WT, SSU:GFP65-StL, SSU:GFP65, SSU:GFP65 + SSU:GFP65-StL, SSU:GFP+, and SSU:GFP++SSU:GFP65-StL. The hybridization probe was radiolabeled GFP65. Hybridization with a α-tubulin probe was used as a loading control and the migration of rRNAs (1.5, 1.8 and 2.2×103 nt; see GenBank AC005806) are indicated by dots. Panel E. siRNA analysis of lines described in panel C, probed with radiolabeled GFP65. The star marks the mobility of a 26 nt standard and CHB is a cross hybridizing band that serves as a loading control. Panel F. Northern blot of analysis of lines described in Panel C, hybridized with the GC-rich GFP+ probe. Hybridization with a α-tubulin probe was used as a loading control and the migration of rRNAs are indicated by dots. Panel G. Western blot of lines described in panel C probed with anti-GFP antisera. The filled arrowhead indicates a cross-reactive band (CRB) that serves as a loading control.

Figure 2. RNAi of endogeneous Leishmania braziliensis genes: effects on mRNA, protein or LPG expression.

Panel A. Northern blot analysis of L. braziliensis transiently transfected with α-tubulin dsRNA. Cells were electroporated with 20 µg dsRNA from the T. brucei paraflagellar rod protein gene (lane 0) or with 20 µg dsRNA derived from L. braziliensis α-tubulin (lane 20). Panel B. mRNA levels of LPG1-StL, LPG2-StL, LPG3-StL or HGPRT-StL L. braziliensis M2903 stable transfectants, as determined by quantitative RT-PCR, relative to WT and/or control transfectants. The average and standard deviation from 4–6 transfectants for each construct are shown. Panel C. Northern blot analysis of LPG2-StL transfectants. Total RNAs were hybridized with a radiolabeled L. braziliensis LPG2 probe, located outside of the LPG2 ‘stem’. Lane 1, LPG2-StL-F; lane 2, LPG2.-StL-R transfectant; lane 3, WT M2903; lanes 4 and 5, empty vectors (StL-F and StL-R, respectively). Panel D. Western blot of HGPRT protein; Lbr, M2903; Lbr+HGPRT-StL, independent Lbr SSU:HGPRT-StL transfectants. Panel E. LPG in LPG2-StL transfectants. LPG was isolated from WT and independent SSU:LPG2-StL transfectants and quantitated, and expressed as percent WT levels.

Figure 3. Ablation of paraflagellar rod synthesis following RNAi of PFR1 or PFR2.

L. braziliensis M2903 was transfected with constructs expressing PFR1-StL or PFR2-StL via integration into the SSU rRNA locus yielding clonal lines with typical transfection frequencies. Several of these, along with WT, were fixed, stained and subjected to transmission electron microscopy as described in the methods. The location of the paraflagellar rod adjacent to the flagellar axonemes is shown (PFR); its presence or absence was scored in 200 cells as indicated below the figure.

Figure 4. RNAi of AGO1.

LUC assays of L. braziliensis M2903 lines bearing the indicated constructs. WT, L. braziliensis M2903; LUC control, SSU::IR2HYG-LUC(b); LUC SR, SSU:IR2SAT-LUC-StL(a)-LUC(b); LUC SR + AGO1 StL, SSU:IR2HYG-LUC-StL(a)-LUC-(b) + SSU:IR1SAT-AGO1-StL(b). Standard deviations are shown; measurements were made in triplicate of the control lines, while the LUC SR+ AGO1 StL represents the average of 12 independent clones, each measured in duplicate.

Figure 5. GFP siRNAs in Leishmania species.

The indicated species were electroporated with the targeting fragment from pIR1SAT-GFP(65)-StL, yielding SSU:SAT-GFP(65)-StL transfectants. These were confirmed by PCR tests for the marker and presence of the inverted GFP65 repeats, and RNA was isolated and subjected to Northern blotting for siRNAs using a GFP65 probe. CHB indicates a cross hybridizing band that serves as a loading control, and the arrow head indicates the position of a 26 nt DNA marker. Panel A and B samples were run on one gel, Panel C and D samples on another one.

Figure 6. Evolutionary tree of trypanosomatid housekeeping genes, AGO1s and Dicers.

Panel A. Protein-based phylogeny of trypanosomatid species considered in this work. We identified the predicted protein sequences for PTR1, (pteridine reductase 1), GSH1 (γ-glutamylcysteine synthetase) and APRT (adenine phosphoribosyl transferase) in public databases (www.genedb.org) or preliminary genome sequence assemblies from Crithidia fasciculata. For each species the three protein sequences were concatenated, aligned using the ClustalW algorithm, and a neighbor joining tree was generated using the MEGA4 software . The scale corresponds to inferred number of amino acid substitutions. The tree shown agrees well with consensus evolutionary trees presented elsewhere . Panel B. Argonautes. A molecular tree was created as described in the legend to Panel A using representative metazoan Argonaute sequences as well as T. brucei AGO1, L. braziliensis AGO1, Crithidia fasciculata AGO1 (this work), and predicted AGO1s for T. congolense and T. vivax (www.genedb.org). Panel C. Trypanosomatid Argonautes. A molecular tree was generated as described in panel B, including only the eight trypanosomatid AGO1s. Panel D. Trypanosomatid DCL1s. A molecular tree was generated as described in panel B, including only the five sequenced trypanosomatid DCL1s.

Figure 7. Retention and loss of RNAi machinery and activity during trypanosomatid evolution.

A consensus evolutionary tree is shown; the scale corresponds to the degree of evolutionary divergence amongst these organisms (Fig. 6). Lineages lacking RNAi activity and/or genes are indicated at the termini, with RNAi-deficient lineages colored red and RNAi-proficient lineages colored blue. Bicoloring along the T. cruzi lineage signifies that the point at which RNAi was lost is unknown. The + or − symbols indicate the presumptive presence and/or loss of RNAi during evolution.

Figure 8. Efficiency of RNAi is reduced in L. guyanensis M4147 independent of LRV status.

Panel A. LUC RNAi reporter assays. pIR2SAT constructs expressing LUC alone (black boxes) or the LUC RNAi self reporter (LUC SR; white boxes) were introduced separately into L. braziliensis M2903, L. guyanensis M4147 (LRV1-4 virus-containing), or L. guyanensis M4147/pX63HYG (virus-free). SSU-integrated clonal lines were obtained and assayed for luciferase activity (n = 4 for M2903; n = 6 for L. guyanensis; the average and standard deviations are shown). The ratio of luciferase activities between the LUC SR and LUC expressing clones of each of the three lines are shown below the graph. Panel B. PCR confirmation of LRV1-4 virus status in parental and transfectant L. guyanensis M4147. PCR primers were LRV1-4 set 1 (lanes 3,5,7,9,11) or set 2 (lanes 2,4,6,8,10,12) (Table S1). RT-PCR reactions were performed with RNAs isolated from L. braziliensis M2903 (virus-free control; lanes 1,2), M4147 (obtained from two sources; lanes 3,4 and 5,6), M4147+LUC SR (lanes 7,8), M4147/pX63HYG (lanes 9,10), or M4147/pX63HYG + LUC SR (lanes 11,12). M, molecular size marker.