An RNA Interference (RNAi) Toolkit and Its Utility for Functional Genetic Analysis of Leishmania (Viannia)

RNA interference (RNAi) is a powerful tool whose efficacy against a broad range of targets enables functional genetic tests individually or systematically. However, the RNAi pathway has been lost in evolution by a variety of eukaryotes including most Leishmania sp. RNAi was retained in species of the Leishmania subgenus Viannia, and here we describe the development, optimization, and application of RNAi tools to the study of L. (Viannia) braziliensis (Lbr). We developed vectors facilitating generation of long-hairpin or “stem-loop” (StL) RNAi knockdown constructs, using GatewayTM site-specific recombinase technology. A survey of applications of RNAi in L. braziliensis included genes interspersed within multigene tandem arrays such as quinonoid dihydropteridine reductase (QDPR), a potential target or modulator of antifolate sensitivity. Other tests include genes involved in cell differentiation and amastigote proliferation (A600), and essential genes of the intraflagellar transport (IFT) pathway. We tested a range of stem lengths targeting the L. braziliensis hypoxanthine-guanine phosphoribosyl transferase (HGPRT) and reporter firefly luciferase (LUC) genes and found that the efficacy of RNAi increased with stem length, and fell off greatly below about 128 nt. We used the StL length dependency to establish a useful ‘hypomorphic’ approach not possible with other gene ablation strategies, with shorter IFT140 stems yielding viable cells with compromised flagellar morphology. We showed that co-selection for RNAi against adenine phosphoryl transferase (APRT1) using 4-aminopyrazolpyrimidine (APP) could increase the efficacy of RNAi against reporter constructs, a finding that may facilitate improvements in future work. Thus, for many genes, RNAi provides a useful tool for studying Leishmania gene function with some unique advantages.


Introduction
More than 1.7 billion people are at risk for the 'neglected tropical disease' leishmaniasis, with nearly 12 million exhibiting symptomatic disease and more than 50,000 deaths annually, and upwards of 100 million harboring asymptomatic infections [1][2][3][4][5]. Leishmania sp. have two distinct growth stages, the promastigote in the sand fly vector, and the intracellular amastigote residing within cellular endocytic pathways in the mammalian host. While the promastigote stage is readily cultured in the laboratory and amenable to molecular techniques, amastigotes require the use of macrophage infection systems or infections of animal models replicating key aspects of human disease.

Western-Blot Analysis
Leishmania promastigotes were collected and resuspended in phosphate-buffered saline (PBS) at 1 × 10 8 cells/mL. Cell extracts were prepared and Western blots performed after separation by SDS-polyacrylamide gel electrophoresis as previously described [26]. For HGPRT, primary antibodies were anti-L. donovani HGPRT and APRT antiserum [27] was used at a titer of 1:5000 and 1:1000, respectively. For normalization anti-L. major H2A [28] was used at a titer of 1:100,000, with goat anti-rabbit IgG as the secondary antibody (1:20,000, Licor Inc., Lincoln, NE USA.). Similar procedures were used for QDPR Western blot analysis with extracts from 1.6 × 10 7 cells. Gels were transferred to nitrocellulose membranes, which were blocked with a 5% skim milk solution and incubated with 1:500 dilution of rat anti-QDPR [29], a 1:1000 dilution of rabbit anti-PTR1 [30] or anti-L. major H2A as described above. IRDye™ anti-rat or rabbit goat immune globulin G were used as the secondary antisera at 1:10,000 dilution. Antibody binding to blots was detected and quantified using an Odyssey infra-red imaging system (Li-Cor).

Quininoid Dihydropteridine Reductase Assay
Parasites were harvested at log phase (4-6 × 10 6 cells/mL) and collected by centrifugation at 1250× g for 10 min at 26 • C, washed twice with PBS, and resuspended at 2 × 10 9 cells ml −1 in 10 mL of Tris-Cl, pH 7.0, with 1 mM EDTA and a mixture of protease inhibitors as described [31]. Cells were lysed by three rounds of freeze thawing and sonication, and the extracts clarified by centrifugation at 15,000× g for 30 min at 4 • C. Protein concentrations were determined using Qubit Fluorometric Quantification (Invitrogen). Quininoid dihydropteridine reductase activity was measured at 25 • C as described [32] using quinonoid dihydrobiopterin generated continuously by horseradish peroxidase mediated oxidation of H 4 -biopterin. The standard reaction mixture contained 50 mM Tris-HCl, pH 7.2, 20 µg of horseradish peroxidase, 0.1 mM H 2 O 2 , 20 µM of H 4 -biopterin, 100 µM NADH, and purified QDPR or parasite lysates; all components were incubated for 3 min prior to initiation of reaction by addition of H 4 B. The activity was measured by monitoring NADH consumption at 340 nm (ε 340 for NADH is 6200 M −1 cm −1 ) in a Beckman DU-640 spectrophotometer.

Transmission Electron Microscopy
Leishmania promastigotes were harvested in logarithmic growth phase and fixed in 2% paraformaldehyde/2.5% glutaraldehyde (Polysciences Inc., Warrington, PA, USA) in 100 mM phosphate buffer, pH 7.2, for 1 h at room temperature. Samples were washed in phosphate buffer and postfixed in 1% osmium tetroxide (Polysciences Inc., Warrington, PA, USA) for 1 h, rinsed extensively in water, and block stained with 1% aqueous uranyl acetate (Ted Pella Inc., Redding, CA, USA) for 1 h. Following several rinses in water, samples were dehydrated in a graded series of ethanol solutions and embedded in Eponate 12 resin (Ted Pella Inc.). Sections of 95 nm were cut with a Leica Ultracut UCT ultramicrotome (Leica Microsystems Inc., Bannockburn, IL, USA), stained with uranyl acetate and lead citrate, and viewed on a JEOL 1200 EX transmission electron microscope (JEOL USA Inc., Peabody, MA, USA).

Construction of the Integrating pIR-GW Destination Vectors Facilitating Generation of StL Constructs Using Gateway Site Specific Recombinase
We first assembled a construct bearing a PEX11-MYC 'loop' fragment flanked by inverted Gateway ® cassettes; these contain ccdB, a lethal gene that targets DNA gyrase, and CmR encoding chloramphenicol-resistance, flanked by two attR sequences necessary for site-specific recombination (attR1-ccdB-CmR-attR2; www.Invitrogen.com (accessed on 26 December 2022). For propagation, all constructs bearing the Gateway cassettes were propagated in Escherichia coli (E. coli) DB3.1 which contains a gyrA462 mutation conferring resistance to ccdB toxicity. First, the PEX11-MYC loop was excised from pIR1SAT-GFP65-StL (B4733) [7,9,10] and inserted into pGEMT yielding pGEMT-stuffer (B5974). The Gateway cassette (attR1-ccdB-CmR-attR2) was amplified from the pDONR221 (Invitrogen) with primers S1 and S2, and inserted by blunt end ligation into the NheI site of pGEMT-stuffer (B5974), yielding pGEMT-stuffer -one GW (B6158). The second GW cassette was inserted by blunt end ligation into the pGEMT-stuffer -one GW AvrII site, in inverted orientation to the first cassette (with both in divergent orientation relative to the ccdB/CmR ORFs), yielding (B6218).

Generation of Target Stem-Loop (StL) Constructs for RNAi
The molecular constructs used in this work and their synthesis are summarized in Table S2. Briefly, for StL constructs the 'stem' was obtained by PCR, inserted into the pCR8/GW/TOPO vector by TA cloning, (Invitrogen # K250020), and their orientation (same direction as attL2) confirmed by sequencing and restriction digestion. The stems were transferred from the pCR8/GW/TOPO donor vector to the pIR1-GW or pIR2-GW destination vectors described above; these contain sequences from the parasite small subunit rRNA locus to enable integration into the genome, and inverted attR1-ccdB-CmR-attR2 cassettes. The gene of interest in pCR8/GW/TOPO was transferred to the pIR-GW vector with LR Clonase II (Thermo Fisher, Waltham MA USA) in an overnight reaction at room temperature. Reactions were terminated by incubating with proteinase K for 1 h at 37 • C (Gateway Technology with Clonase TM II manufacturer' protocol, Version A: 24 June 2004). Reactions were transformed into E.coli TOP10 TM or DH5α (which select against both ccdB cassettes). All final stem-loop (StL) constructs named as StL expressers were confirmed by restriction enzyme digestion and DNA sequencing ( Figure 1). Prior to transfection, constructs were digested with SwaI to expose the SSU rRNA segments mediating homologous integration.

Generation of Target Stem-Loop (StL) Constructs for RNAi
The molecular constructs used in this work and their synthesis are summarized in Table S2. Briefly, for StL constructs the 'stem' was obtained by PCR, inserted into the pCR8/GW/TOPO vector by TA cloning, (Invitrogen # K250020), and their orientation (same direction as attL2) confirmed by sequencing and restriction digestion. The stems were transferred from the pCR8/GW/TOPO donor vector to the pIR1-GW or pIR2-GW destination vectors described above; these contain sequences from the parasite small subunit rRNA locus to enable integration into the genome, and inverted attR1-ccdB-CmR-attR2 cassettes. The gene of interest in pCR8/GW/TOPO was transferred to the pIR-GW vector with LR Clonase II (Thermo Fisher, Waltham MA USA) in an overnight reaction at room temperature. Reactions were terminated by incubating with proteinase K for 1 h at 37 °C (Gateway Technology with Clonase TM II manufacturer' protocol, Version A: 24 June 2004). Reactions were transformed into E.coli TOP10 TM or DH5α (which select against both ccdB cassettes). All final stem-loop (StL) constructs named as StL expressers were confirmed by restriction enzyme digestion and DNA sequencing ( Figure 1). Prior to transfection, constructs were digested with SwaI to expose the SSU rRNA segments mediating homologous integration. Figure 1. Flowchart for generation of entry and StL constructs using Gateway site-specific recombinase technology followed by introduction into Leishmania to express dsRNA. A stem for each GOI (gene of interest) was amplified by PCR and then cloned into pCR8/GW/TOPO vector (Invitrogen) to generate an "entry" clone bearing a flanking attL1 and attL2 sites for recombination-based transfer. The pIR-GW destination vectors bear two opposing ccdB/CmR cassettes, each flanked by attR1 and attR2 sites. Following site-specific recombination by the Gateway LR reaction, the desired StL for each GOI was obtained and confirmed. For biological tests, each StL DNA was linearized with SwaI, electroporated into Leishmania, where it integrated into the small subunit ribosomal RNA locus Figure 1. Flowchart for generation of entry and StL constructs using Gateway site-specific recombinase technology followed by introduction into Leishmania to express dsRNA. A stem for each GOI (gene of interest) was amplified by PCR and then cloned into pCR8/GW/TOPO vector (Invitrogen) to generate an "entry" clone bearing a flanking attL1 and attL2 sites for recombination-based transfer. The pIR-GW destination vectors bear two opposing ccdB/CmR cassettes, each flanked by attR1 and attR2 sites. Following site-specific recombination by the Gateway LR reaction, the desired StL for each GOI was obtained and confirmed. For biological tests, each StL DNA was linearized with SwaI, electroporated into Leishmania, where it integrated into the small subunit ribosomal RNA locus (SSU). There it is transcribed by the strong rRNA pol I promoter to generate RNAs whose dsRNA regions are processed by the RNAi machinery.

Rapid Generation of 'Stem-Loop' Constructs as RNAi Triggers
To trigger the RNAi response, dsRNA generating 'stem-loop' constructs have been used in previous studies, often assembled in three steps with two 'stem' segments cloned in opposite orientations, separated by a short spacer/loop [10]. To accelerate this process, we utilized Gateway TM (Invitrogen) technology, incorporating precise, site-specific recombina-tion [22,33]. First we engineered a 'destination' expression vector, based on the Leishmania pIR vector series which achieves high levels of RNA expression following integration into the ribosomal small subunit RNA locus [10,34]. The final destination vector (pIR-GW generically) bears two Gateway recipient cassettes, arranged in inverted orientation and separated by a short segment destined to become the 'loop' inserted into one of the strong expression sites of pIR vectors ( Figure 1). Each Gateway cassette contained a positive CmR and negative ccdB marker, flanked by attR1 and attR2 sites. Then, each stem to be tested was inserted into an 'entry' vector (pCR8/GW/TOPO), where it was flanked by donor attL1 and attL2 sites ( Figure 1). Gateway LR reactions between the destination vector and entry DNAs were performed in vitro, and transformed into E.coli TOP10 or DH5α TM which selected against the presence of the target ccdB/CmR cassettes, which could be rapidly confirmed by loss of chloramphenicol resistance. Typically we obtained numerous recombinants of which more than 70% bore the expected configuration of the desired StL configuration, and one correct representative was selected for introduction into Leishmania. Prior to transfection, constructs were digested with SwaI to expose the SSU rRNA termini to direct integration into the SSU rRNA locus, where strong expression is driven from the rRNA promoter ( Figure 1).

Testing the Effect of Stem Length on RNAi Activity Using L. braziliensis HGPRT-StL and Luciferase-StL Series Constructs
In T. brucei, stem lengths ranging in length between 100 and 1500 bp are effective in knocking down target genes [35,36], and stem lengths as short as 29-100 nt are active in other metazoan species [37,38]. To explore this in Leishmania, we varied the stem length in RNAi constructs targeting an integrated firefly luciferase reporter gene (LUC) as well as an endogenous cellular gene (HGPRT), assayed following transfection into WT Lbr. For HGPRT increasing the stem from 494 nt to 1005 nt reduced expression from 55% to 92% ( Figure 2A). No strong differences were seen between similarly sized stems targeting the HGPRT ORF or 3 untranslated region ( Figure 2A).
For luciferase, we tested LUC StL constructs following introduction into a LbrM2903 transfectant stably expressing high levels of luciferase activity ( Figure 2B). As with HGPRT, the longer luciferase stems resulted in greater reductions in LUC activity ( Figure 2B). While stems of 158 bp or greater showed strong reductions (>93%), stems of 128 or 54 nt showed much smaller effects (35-32%; Figure 2B). Preliminary analysis did not suggest a strong association of the activity of the stems tested with siRNA abundance 'hot spots', small regions where siRNA levels greatly exceed the average across a gene, as seen with LUC or Leishmania Virus 1 (LRV1) in our prior study [19]. We did not observe synergy when two weak constructs were transfected simultaneously (not shown).
These data together with those shown below for IFT140 (Section 3.6) suggested that for the strongest effect stem lengths of >500 nt were preferable, with a significant drop-off below 128 nt. We did not see strong 'positional' effects of the location of the stems within the target gene, however the constructs tested tended to be progressively truncated from one side so it is possible these may have been overlooked. Importantly, our findings suggest differences in the dsRNA trigger length dependency in Leishmania relative to trypanosomes or other organisms. Further studies will be required to explore the basis for this. Genes 2023, 14, x FOR PEER REVIEW 7 of 18

StL-Mediated Specific RNAi of the Metabolic Target Quinonoid Dihydropteridine Reductase (QDPR) Interspersed within a Tandem Repeated Gene Array
We examined the efficacy of StL RNAi against the L.braziliensis quinonoid dihydropteridine reductase (QDPR) gene, encoding an enzymatic step of the reduced pteridine pathways involved in recycling oxidized qH2-biopterin or qH2-folates back to the active tetrahydro state in Leishmania [29]. In all Leishmania, QDPR genes are interspersed in a tandem array with two other genes, ORFq (q: a hypothetical protein) and β7-proteasome (β7: 20S proteasome β7 subunit), the latter gene being essential where tested in other species including trypanosomes ( Figure 3A; [29]). This rendered selective deletion of the interspersed QDPR targets considerably more challenging, but provided a suitable opportunity for testing the utility of RNAi in this context.
We introduced a StL construct bearing a 588 nt QDPR stem into WT parasites successfully. Relative to WT, QDPR protein was reduced more than 87.4% ( Figure 3B) while QDPR activity was reduced by more than 88% individually in StL knockdowns ( Figure  3C). The Lbr QDPR StL knockdowns grew normally in culture, suggesting that RNAi specifically targeted QDPR without significantly impacting the flanking essential proteasome subunit, although a partial reduction in expression cannot be ruled out. Thus RNAi can be used to successful probe the consequences of metabolic gene expression depletion in this most challenging context.

StL-Mediated Specific RNAi of the Metabolic Target Quinonoid Dihydropteridine Reductase (QDPR) Interspersed within a Tandem Repeated Gene Array
We examined the efficacy of StL RNAi against the L.braziliensis quinonoid dihydropteridine reductase (QDPR) gene, encoding an enzymatic step of the reduced pteridine pathways involved in recycling oxidized qH 2 -biopterin or qH 2 -folates back to the active tetrahydro state in Leishmania [29]. In all Leishmania, QDPR genes are interspersed in a tandem array with two other genes, ORFq (q: a hypothetical protein) and β7-proteasome (β7: 20S proteasome β7 subunit), the latter gene being essential where tested in other species including trypanosomes ( Figure 3A; [29]). This rendered selective deletion of the interspersed QDPR targets considerably more challenging, but provided a suitable opportunity for testing the utility of RNAi in this context.
We introduced a StL construct bearing a 588 nt QDPR stem into WT parasites successfully. Relative to WT, QDPR protein was reduced more than 87.4% ( Figure 3B) while QDPR activity was reduced by more than 88% individually in StL knockdowns ( Figure 3C). The Lbr QDPR StL knockdowns grew normally in culture, suggesting that RNAi specifically targeted QDPR without significantly impacting the flanking essential proteasome subunit, although a partial reduction in expression cannot be ruled out. Thus RNAi can be used to successful probe the consequences of metabolic gene expression depletion in this most challenging context.

RNAi of an L. braziliensis Gene Important for Amastigote Replication
An important area of Leishmania biology is the study of genes that impact survival of the amastigote stage in the vertebrate host. Many (not all) Leishmania species can differentiate to amastigote-like forms in culture (axenic amastigotes), when grown at conditions resembling those within the parasitophorous vacuole, e.g. elevated temperature and low pH [40,41]. Here, we made use of a clonal derivative of Lbr M2903 (LbrM2903SA2) which had been adapted for growth as axenic amastigotes [24].
As a test we focused on the A600 locus, which in L. mexicana comprises four genes whose deletion had little impact on promastigotes, but precluded axenic amastigote replication in vitro [42]. The A600 copy number varies amongst species [42], with only two found in Lbr (A600-1 and A600-4) which show 88% nucleotide identity. We targeted these sequences simultaneously by a single StL construct using the A600-1 sequence, which shows long stretches of identity relative to A600-4 (118 nt, 51 nt and 45 nt).
We transfected the LbrA600-StL construct into Lbr SA2 promastigotes, obtaining many clonal transfectants all of which grew normally. Several were then inoculated into axenic amastigote growth medium, where they showed a severe growth defect, with 10fold fewer amastigotes than WT (Figure 4), similar to the results obtained with L. mexciana A600 knockouts. These results establish the utility of RNAi knockdowns for the study L. braziliensis genes involved in amastigote differentiation and proliferation. QDPR genes are tandemly repeated with ORFq (Q) and β7 proteasome (β7) up to 9 times in L. braziliensis genomes; only the ORFs are depicted, and the entire QDPR ORF was used as the tested stem. Proteasome subunits are known to be essential, while the requirement for ORFq has not been tested. (B) QDPR protein expression in QDPR-StL knockdowns. QDPR was detected with anti-QDPR antiserum and expression was calculated relative to that of H2A, from Western blots as described in the methods. (C) QDPR enzymatic activity in QDPR-StL knockdowns. QDPR enzymatic activity was assayed and normalized to total cellular protein. Data were from three independent experiments each performed in triplicate. * indicates p < 0.0235. (D) Growth inhibition by inhibitor TQD. WT and QDPR-StL transfected Lbr were inoculated into media containing 0-50 µM 5, 6, 7, 8-Tetrahydro-2, 4-quinazolinediamine (TQD) at 2 × 10 5 cells/mL and allowed to grow until WT had reached late log phase, at which time parasite numbers were determined. The experiment was repeated three times, with results similar to that shown; the EC50s for the QDPR-StL (red) and WT (blue) lines were 0.15 ± 0.015 vs. 3.49 ± 0.59 µM (p < 0.0006).

RNAi of an L. braziliensis Gene Important for Amastigote Replication
An important area of Leishmania biology is the study of genes that impact survival of the amastigote stage in the vertebrate host. Many (not all) Leishmania species can differentiate to amastigote-like forms in culture (axenic amastigotes), when grown at conditions resembling those within the parasitophorous vacuole, e.g. elevated temperature and low pH [40,41]. Here, we made use of a clonal derivative of Lbr M2903 (LbrM2903SA2) which had been adapted for growth as axenic amastigotes [24].
As a test we focused on the A600 locus, which in L. mexicana comprises four genes whose deletion had little impact on promastigotes, but precluded axenic amastigote replication in vitro [42]. The A600 copy number varies amongst species [42], with only two found in Lbr (A600-1 and A600-4) which show 88% nucleotide identity. We targeted these sequences simultaneously by a single StL construct using the A600-1 sequence, which shows long stretches of identity relative to A600-4 (118 nt, 51 nt and 45 nt).
We transfected the LbrA600-StL construct into Lbr SA2 promastigotes, obtaining many clonal transfectants all of which grew normally. Several were then inoculated into axenic amastigote growth medium, where they showed a severe growth defect, with 10-fold fewer amastigotes than WT (Figure 4), similar to the results obtained with L. mexciana A600 knockouts. These results establish the utility of RNAi knockdowns for the study L. braziliensis genes involved in amastigote differentiation and proliferation.

Targeting Essential Genes of the L. braziliensis Intraflagellar Transport (IFT) Pathway
The Leishmania flagellum plays key roles in the promastigote and amastigote stages, and previously we reported successful knockdowns of the paraflagellar rod proteins PFR1 and PFR2 which showed phenotypes comparable to complete deletions [10]. We extended these studies to a set of genes associated with intraflagellar transport (IFT), which mediates transport of cargo in both anterograde and retrograde directions and is required for proper flagellar assembly [43].
StL constructs targeting Lbr IFT122, IFT140, IFT172, representing both the anterograde and retrograde IFT, were transfected into WT Lbr (Table 1). However, we were unable to recover transfectant colonies, despite multiple attempts and success targeting a nonessential gene, LbrPFR1-StL (Table 1). To establish that this result was dependent on the RNAi pathway, we introduced these same constructs into an RNAi-deficient mutant obtained by homozygous deletion of the AGO1 gene (Δago1 − ), an Argonaute family protein encoding the key 'slicer' activity required for RNAi activity in Lbr [10,44]. Now, all LbrIFT StL constructs successfully yielded transfectants ( Table 1), establishing that RNAi of the StL-derived dsRNA trigger was responsible for the lack of transfectants.
These data are consistent with other studies in trypanosomes and Leishmania showing that IFT gene ablation displays an array of phenotypes, including essentiality [45][46][47][48]. Negative control 0 0

Targeting Essential Genes of the L. braziliensis Intraflagellar Transport (IFT) Pathway
The Leishmania flagellum plays key roles in the promastigote and amastigote stages, and previously we reported successful knockdowns of the paraflagellar rod proteins PFR1 and PFR2 which showed phenotypes comparable to complete deletions [10]. We extended these studies to a set of genes associated with intraflagellar transport (IFT), which mediates transport of cargo in both anterograde and retrograde directions and is required for proper flagellar assembly [43].
StL constructs targeting Lbr IFT122, IFT140, IFT172, representing both the anterograde and retrograde IFT, were transfected into WT Lbr (Table 1). However, we were unable to recover transfectant colonies, despite multiple attempts and success targeting a nonessential gene, LbrPFR1-StL (Table 1). To establish that this result was dependent on the RNAi pathway, we introduced these same constructs into an RNAi-deficient mutant obtained by homozygous deletion of the AGO1 gene (∆ago1 − ), an Argonaute family protein encoding the key 'slicer' activity required for RNAi activity in Lbr [10,44]. Now, all LbrIFT StL constructs successfully yielded transfectants (Table 1), establishing that RNAi of the StLderived dsRNA trigger was responsible for the lack of transfectants. These data are consistent with other studies in trypanosomes and Leishmania showing that IFT gene ablation displays an array of phenotypes, including essentiality [45][46][47][48].
The table shows the number of colonies (per plate) obtained after transfection of various StL constructs into WT or RNAi-deficient ∆ago1 − Lbr. PFR1-StL was used as a positive control [10].

Systematic Generation of Hypomorphic Mutants by Exploiting the Stem Length-Dependency of RNAi in L. braziliensis
While plus/minus tests of gene essentiality are useful they provide little information about the role of the encoded protein within the cell. When inducible systems are available, observations following shutoff and prior to cell death can be informative, however presently this can only be achieved using conditional degradation domains in Leishmania [49]. The studies of the stem length-dependency of the efficacy of RNAi described in Section 3.2 above suggested that it could be possible to engineer partial loss of function mutants through successively reducing the stem length, until a point was reached where viable cells could be obtained, which ideally might show informative defects. We tested this with the IFT140 gene, which is essential in trypanosomes and probably Lbr (Table 1). Curiously, viable deletion mutants with different phenotypes were possible in L. major and L. donovani [47,48] suggesting there is considerable variation amongst Leishmania species for the consequences of IFT ablation.
Transfection of a 941 nt IFT140 StL construct yielded no transfectants in WT Lbr (Table 1), however as the stem length was reduced to 562 nt and further to 131 nt, increasing numbers of transfectants could be recovered ( Figure 5A). Notably, the 562 nt stem yielded about 10% as many transfectants, and the cells recovered from these colonies grew slowly in culture ( Figure 5A). In contrast, transfectants recovered with constructs with stems shorter than 562 nt were recovered at normal frequencies, grew like WT, and showed no obvious differences in flagellar length or motility.
The table shows the number of colonies (per plate) obtained after transfection of various StL constructs into WT or RNAi-deficient Δago1 − Lbr. PFR1-StL was used as a positive control [10].

Systematic Generation of Hypomorphic Mutants by Exploiting the Stem Length-Dependency of RNAi in L. braziliensis
While plus/minus tests of gene essentiality are useful they provide little information about the role of the encoded protein within the cell. When inducible systems are available, observations following shutoff and prior to cell death can be informative, however presently this can only be achieved using conditional degradation domains in Leishmania [49]. The studies of the stem length-dependency of the efficacy of RNAi described in Section 3.2 above suggested that it could be possible to engineer partial loss of function mutants through successively reducing the stem length, until a point was reached where viable cells could be obtained, which ideally might show informative defects. We tested this with the IFT140 gene, which is essential in trypanosomes and probably Lbr (Table 1). Curiously, viable deletion mutants with different phenotypes were possible in L. major and L. donovani [47,48] suggesting there is considerable variation amongst Leishmania species for the consequences of IFT ablation.
Transfection of a 941 nt IFT140 StL construct yielded no transfectants in WT Lbr (Table 1), however as the stem length was reduced to 562 nt and further to 131 nt, increasing numbers of transfectants could be recovered ( Figure 5A). Notably, the 562 nt stem yielded about 10% as many transfectants, and the cells recovered from these colonies grew slowly in culture ( Figure 5A). In contrast, transfectants recovered with constructs with stems shorter than 562 nt were recovered at normal frequencies, grew like WT, and showed no obvious differences in flagellar length or motility. This suggested that the 562 StL IFT140 transfectant was a candidate hypomorph, as it displayed a shortened flagellum and weak motility (not shown). Transmission EM of one clone showed that these knockdown cells exhibited modified shape and accumulated vesicles in the flagellar pocket similar to the phenotypes in other trypanosomes, such as T. brucei where ablation of IFT140 was initiated by conditional RNAi prior to cell death [45], or in L.mexicana IFT140 knockouts [48] ( Figure 5C). These data suggest that the stem-length dependency of RNAi in L. braziliensis will prove a useful tool in generating informative, hypomorphic mutants of otherwise essential genes, facilitating inquiries into their cellular targets and/or mechanism of action.

A Negative Selection System for Enhancement of RNAi Activity in L. braziliensis
It is often helpful to have screens or selections to monitor the efficacy of RNAi under various circumstances; for example, in screening for genes acting within the RNAi pathway, or when RNAi exhibits considerable clonal variability, as in some fungi [50][51][52]. GFP and luciferase based screens are commonly used, which both perform well in Leishmania [10], but here we sought a selectable system which could facilitate other applications.
We tested the 4-aminopyrazolpyrimidine/adenine phosphoryl transferase 1 (APP/APRT1) system in L. braziliensis. APRT converts APP into a toxic metabolite inhibiting Leishmania growth, and thus cells with decreasing APRT levels show increasing resistance to APP. Importantly, APRT1 is not essential in most media due to the alternative salvage route through adenosine aminohydrolase [53].
First, we introduced an APRT1-StL construct into WT Lbr, where transfectants were readily obtained and grew normally. Western blot analysis confirmed significant reduction to undetectable level in APRT1 expression ( Figure 6A). Unlike WT parasites whose growth was completely inhibited by 500 µM APP, LbrAPRT1-StL transfectants were able to grow at the highest concentration tested, albeit at reduced growth rates ( Figure 6B,C). Other studies showed that the APP EC50s were 50-100 µM for WT and 500-1000 µM for APRT1-StL (not shown). Thus, RNAi knockdown of APRT1 leads to APP resistance as expected.
We developed a dual stem-loop reporter parasite that simultaneously expresses an APRT1-StL along with a LUC-StL with a 508 nt stem. This was obtained by transfection an Lbr line expressing luciferase with a second construct bearing both APRT1-StL and LUC-StL cassettes (Section 2.2). This reporter line allows manipulations of overall cellular RNAi activity to be tracked through simple luciferase assays. We chose a mid-sized stem as we wanted to provide some range for detection of elevated RNAi activity before saturation.
To illustrate its utility, APRT1-StL+LUC-StL/LUC parasites were plated on increasing concentrations of APP, which would select for low APRT1 expression which could potentially arise through alterations in overall cellular RNAi activity. Colonies were picked and parasites were grown thereafter in 50% of the selective concentration of APP used in plating.
In the absence of APP selection, luciferase expression was reduced 122 fold in the APRT1-StL+LUC-StL/LUC parasites (Figure 7), slightly less than the 200-300 fold reduction seen previously [10] but greater than the 30-50 fold seen in Figure 2B with similarly sized stems. These experimental series were performed months or even years apart, and we suspect small differences in vector design or clonal variation may be a contributing factor. Since each series of experiments was internally controlled, we consider these differences of neglible significance in this context.
In the APRT1-StL+LUC-StL/LUC parasites subjected to APP selection, luciferase activity decreased an additional 2.9 to 6.8 fold, as the APP concentrations were increased from 125 to 1000 µM (Figure 7). Overall relative to WT parasites, the efficacy of LUC silencing increased from 122-to nearly 830-fold. Thus, in one step we were readily able to identify parasite showing elevated levels of RNAi activity against a luciferase reporter. This proof of principle experiment establishes a useful tool for both monitoring and manipulating RNAi activity amongst clones or mutants that may prove of great utility in future studies. We developed a dual stem-loop reporter parasite that simultaneously expresses an APRT1-StL along with a LUC-StL with a 508 nt stem. This was obtained by transfection an Lbr line expressing luciferase with a second construct bearing both APRT1-StL and LUC-StL cassettes (Section 2.2). This reporter line allows manipulations of overall cellular RNAi activity to be tracked through simple luciferase assays. We chose a mid-sized stem as we wanted to provide some range for detection of elevated RNAi activity before saturation. from 125 to 1000 μM (Figure 7). Overall relative to WT parasites, the efficacy of LUC silencing increased from 122-to nearly 830-fold. Thus, in one step we were readily able to identify parasite showing elevated levels of RNAi activity against a luciferase reporter. This proof of principle experiment establishes a useful tool for both monitoring and manipulating RNAi activity amongst clones or mutants that may prove of great utility in future studies. Figure 7. A negative selectable system to identify enhanced RNAi activity in L. braziliensis. Double StLs (LUC and APRT1) expressed in a single construct were transfected in the Lbr LUC reporter line and selected on M199 plates containing various concentrations of APP. Colonies were readily obtained and grown in APP, and luciferase activity was measured from 10 6 parasites as described in the methods. The fold reduction of colonies selected in varying concentrations of APP relative to no drug are shown in brackets above the plotted luciferase activity. The average and standard deviations calculated from at least 10 transfectant colonies is shown. *** indicates a p < 0.001 for comparisons with no APP selection. Figure 7. A negative selectable system to identify enhanced RNAi activity in L. braziliensis. Double StLs (LUC and APRT1) expressed in a single construct were transfected in the Lbr LUC reporter line and selected on M199 plates containing various concentrations of APP. Colonies were readily obtained and grown in APP, and luciferase activity was measured from 10 6 parasites as described in the methods. The fold reduction of colonies selected in varying concentrations of APP relative to no drug are shown in brackets above the plotted luciferase activity. The average and standard deviations calculated from at least 10 transfectant colonies is shown. *** indicates a p < 0.001, ** indicates a p < 0.01 for comparisons with no APP selection.

Discussion
RNA interference has proven a valuable tool for the study of gene regulation in many eukaryotes including African trypanosomes. While lost in many Leishmania sp., those of the subgenus Viannia retained a functional pathway, opening up its use as a tool for genetic analysis. In this work we describe a useful Gateway™ site specific recombination based system for rapidly and efficiently generating stem-loop constructs suitable for RNAi tests, and applied it towards a spectrum of Leishmania genes of interest to illustrate its range and potential. For perspective, Table 2 summarizes the genes and the outcomes obtained in this or previous studies.  Of the 22 targets listed, 6 (27%) showed changes in survival, growth and/or morphology when knocked down (PFR1, PFR2, IFT140, IFT122, IFT172, VTC4). When other criteria were evaluated phenotypes were detected in a further 9 (41%; APRT1, A600, QDPR, AGO1, δ-amastin, and 4 LRV1s). In contrast, no phenotypes were detected with the methods applied with 7 target genes (HGPRT1, LPG1, LPG2, LPG3, BBS1, BBS3 or BBS4). Our findings can be compared with the more extensive studies from African trypanosomes, where about 1/3 of genes targeted in chromosomal surveys or genome wide studies show changes in survival, growth or morphology [16,55]. As in trypanosomes, genes involved in the synthesis of abundant surface glycoconjugates such as Leishmania lipophosphoglycan LPG1-3 showed little phenotype, while other genes such as those implicated in the parasite flagellum (PFR or IFT) yielded readily detected phenotypes. Importantly phenotypes were obtained in several repetitive gene families, including Lbr A600, PFRs, δ-amastins and QDPR, the latter having additional complexity as an 'interspersed' repetitive gene family. Similarly we have been able to target cytoplasmic RNA viruses effectively such as the LRV1 totiviruses of Lbr and Lgy. These data suggest that RNAi offers another strong option for functional genomics in Leishmania.
The specificity of RNAi knockdowns can be assessed in several ways. First, the dependency of RNAi on a functional Argonaute (AGO1) can be used to support conclusions about the essentiality of a given RNAi target, as illustrated by studies with several essential IFT genes (Table 1). Secondly, the highly expressed integrated SSU:IR-StL constructs used here typically yield very high levels of siRNAs and occasionally dsRNA, both of which can have off target effects [19]. The specificity for the target gene may be assessed by reintroduction of a 'recoded' RNAi-resistant target gene [21], or by selection for rare spontaneous excision of the integrated SSU:IR-StL construct from the rRNA gene array [56], both of which should restore the WT phenotype.
While most studies summarized in Table 2 have been carried out with L. braziliensis, RNAi was also effective in L. guyanensis (VTC4, LRV1). However, in previous work we showed in both LUC and LRV1 studies that the efficacy of RNAi was significantly less than in Lbr [10,19]. The factors responsible have not been studied, and in our previous studies we noted that even with strong RNAi induction target RNA levels sometimes did not decline [10]. Thus, the 'penetrance' of RNAi efficacy can vary widely amongst specific genes, and/or species or strains, something that should be considered in future studies. Fortunately, the ease by which RNAi constructs may be generated using site-specific recombinase technology (Figure 1) allows this to be explored with minimal effort.
Interestingly, with the selectable APRT1/APP system we were able to increase the strength of RNAi nearly 7 fold in Lbr (Figure 7), suggesting it may be possible in the future to develop lines where the efficacy of RNAi is enhanced. The feasibility of this was shown in mammalian cells where increased expression of Argonaute-2 enhanced RNAi activity [27].
While the introduction of CRISPR/Cas9 technology provides an attractive alternative to RNAi for gene ablation, there are some useful applications of RNAi as well. Inducible RNAi systems are readily reversible, and the stem length dependency shown here offers the possibility of developing 'graded' RNAi responses to yield stable hypomorphic lines. This was illustrated with the Lbr IFT140 gene, where shorter stems led to viable cells, with a key one exhibiting reduced transfection efficiency, slower growth and flagellar defects ( Figure 5), allowing further exploration of the impact of RNAi on flagellar biology or cell physiology. While the data provided here serve as a guide, it is likely that the optimal stem length (and possibly position) will have to be evaluated empirically for any given gene and/or phenotype. Nonetheless, the ability to systematically generate stable hypomorphic mutants adds another dimension to the utility of RNAi in Leishmania.