Sequence alignment

L. braziliensis RPL13a and RPS16 protein and gene sequences were obtained from TriTryp data base (https://tritrypdb.org/tritrypdb/app/). The respective sequences were used as inputs to perform multiple global alignment queries in Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/) to a sequence comparison. The RP nomenclature used herein is based on the annotation of the Leishmania genome database (https://tritrypdb.org/tritrypdb/app/). If using the universal nomenclature system proposed by Ban et al (2014), in which homologous ribosomal proteins are assigned the same name, regardless of the species, L13a would be named uL13 and S16, uS9. For structural alignments, the predicted protein structures for both variants of L13 were obtained from AlphaFold database (https://alphafold.ebi.ac.uk/entry/Q4QFG2 for LMJF_15_0200 and https://alphafold.ebi.ac.uk/entry/Q4Q3B9 for LMJF_34_0860). The URLs corrresponding to mmCIF files were submitted to the online Pairwise Structure Alignment tool (https://www.rcsb.org/docs/tools/pairwise-structure-alignment) (https://alphafold.ebi.ac.uk/files/AF-Q4QFG2-F1-model_v4.cif for LMJF_15_0200 and https://alphafold.ebi.ac.uk/files/AF-Q4Q3B9-F1-model_v4.cif for LMJF_34_0860) and the default alignment method (jFATCAT-rigid) was applied to generate the superposition. Alignments were visualized using an integrated tool within the Pairwise Structure Alignment tool.

Parasite culture, differentiation, and transfection

Procyclic promastigotes of Leishmania major strain LV39 (MRHO/SU/59/P) were cultivated in M199 medium (Sigma Aldrich) supplemented with 10% heat-inactivated fetal bovine serum. To obtain metacyclic promastigotes, procyclic forms were cultivated for 5 days in M199 and metacyclics were enriched from stationary phase cultures by Ficoll^® gradient [22]. Transfections were performed as described previously [23]. Briefly, 1x10⁷ promastigotes were resuspended in 100 μL Tb-BSF buffer and added to 60 μL of DNA generated by PCR (for Δ or tagging). All this volume was transferred to an electrolytic cuvette and transfection was performed using the X-001 program of Amaxa Nucleofector instrument (LONZA). Cultures were maintained in M199 at 26°C before selection with the appropriate drugs (16 μg/mL hygromycin B; 20 μg/mL blasticidin; 16 μg/ mL puromycin) on solid M199 media. After 15–20 days, individual colonies were collected and transferred to liquid M199 medium containing the respective drug of selection and homozygosity was confirmed by DNA extraction and conventional PCR and sequencing. Using specific primers (ST.3), the tagged regions were amplified by conventional PCR, cloned into a pCR4-TOPO plasmid and sequenced by the Sanger method with M13 primers, confirming the correct insertion of 3 copies of the myc tag (S6 Fig)

Immunofluorescence

RP subcellular localization was analyzed by immunofluorescence: a total of 1.5x10⁶ cells were centrifuged at 1,400 x g for 5 minutes at RT, followed by washing with 500 μl of PBS. Cells were fixed for 10 minutes at RT in 500 μL of 3% paraformaldehyde in PBS, pelleted and washed once with 500 μL PBS. The pellet was resuspended in 100 μL of 0.1% glycine in PBS and 30 μL of the total cells were added to poly-lysine slides and left to adhere for 30 minutes. Then, fixed cells were permeabilized with 30 μl of 0.2% Triton X-100 in PBS solution for 5 minutes at RT, then washed 5x in 1x PBS. Blocking was performed for 30 min with 5% skimmed milk powder dissolved in TBS-T. α-myc (Sigma C3956) primary antibody was diluted in blocking solution (1:4000) ratio and incubated for two hours. After incubation, five washes were performed with 1x PBS then secondary antibodies conjugated with Alexa Fluor 488 (Invitrogen A11001) were 1:500 diluted in PBS and incubated for 30 minutes at RT protected from the light. Nuclei and kinetoplast staining were performed using HOESCHT (Invitrogen H3570) at 2 μg·mL^-1 for 15 min. Images were acquired on a Leica DMI6000B fluorescence microscope at 60x magnification and processed using the Fiji Image software J (https://imagej.net/Fiji/Downloads) [24].

Scanning Electron Microscopy (SEM)

10⁷ parasites were fixed for 2 hours in 3% paraformaldehyde and 2% glutaraldehyde in PBS supplemented with 0.9 mM CaCl₂ and 0.5 mM MgCl₂ at RT. The parasites were post-fixed in 2% OsO₄ for 2 h and incubated with a thiocarbohydrazide (TCH) solution for 10 minutes, followed by ethanol and acetone dehydration. Then, parasite cells were mounted on a support and subjected to gold-coated metal plating. Cells were analyzed using an electron microscope scan (JEOL-JSM-5200). Images were captured in the Electron Microscopy Multiuser Laboratory (Department of Cellular and Molecular Biology, Ribeirão Preto Medical School, USP).

Western blotting

Parasites were pelleted (1,400 x g for 5 minutes at RT), washed once with 500 μL of cold-PBS supplemented with protease inhibitors (Roche), resuspended in 10 μL of extraction buffer⁵⁴ and boiled for 10 min. Total protein was quantified in Nanodrop One spectrophotometer (Thermo Scientific) by measuring the absorbance at 280 nm. For every protein sample, sample buffer was added [23] and boiled for 3 minutes. 30 μg of total protein for each sample was fractionated in a 12% polyacrylamide gel. Proteins were transferred to nitrocellulose membranes (GE Healthcare Life Sciences: 10600003) and blocked for 1 h with TBS-T buffer (Tris-saline-Tween buffer: 10 mM Tris, 100 mM NaCl, pH 7.6, Tween20 0.1%) containing 5% milk powder. Immune detection was performed with the appropriate primary and secondary antibodies, following the manufacturer’s recommendations: α-myc (1:4000; Sigma C3956) and α -EF1α: (1:80000; Merck 05–235). Both primary antibodies were diluted in solution of TBS-T with 2.5% powdered milk and incubated for 1 h at RT, followed by incubation with secondary antibody under the same conditions. Membrane visualization was performed via chemiluminescence (ECL kit–GE Healthcare: RPN2232) and images were obtained on the ImageQuant LAS 4000 equipment (GE Healthcare). Band quantification was performed using ROI manager tool of Fiji ImageJ software to determine the band skew of each sample. Band skew is a parameter used to quantify the pixels of one image, and has no specific unit [25]. Increase in the band skew means higher pixel’s intensity, here directly correlated to the protein amount. All band skews were compared between the samples.

RNA extraction and transcripts quantification

Cells were pelleted (1,400 x g for 5 minutes at RT), lysed with TRIzol reagent (Invitrogen) and RNA extraction was performed using DirectZol RNA Miniprep kit (Zyme Research). Total RNA was treated with DNase Turbo (ThermoFisher Scientific) and RT-qPCR performed and analyzed as described by Freitas Castro and cols [26], using G6PDH as housekeeping genes for normalization.

Pull-down assay

The regions corresponding to the 3’UTR of the RPS16 and RPL13a genes were retrieved from TriTryp data base (https://tritrypdb.org/tritrypdb/app/). These sequences were cloned into pUC-56 plasmid between a T7 promoter and 4xS1m aptamer sequences, adapted from Leppek’s strategy [27]. Then, RNA was in vitro transcribed (MEGAscript T7 transcription kit–ThermoFisher AM1334). Thirty micrograms of purified RNA were immobilized on Streptavidin magnetic beads (NEB) at 4°C for 8h under orbital rotation. 10⁸ parasites were lysed on ice by physical pressure using a 19G needle with 1mL of SA-RNP-Lyse buffer (20mM Tris-HCl pH7.5, 150mM NaCl, 1.5mM MgCl₂, 2mM DTT, 2mM RNAse inhibitor, 1 protease inhibitor cocktail tablet, 1% Triton X-100). Biotinylated proteins were previously removed from the extract by incubating the lysed extract for 8h at 4°C with the streptavidin beads. Then, the supernatant was incubated with the bead-immobilized RNA sequences for 8h at 4°C. An empty plasmid with no sequence between T7 promoter and 4xS1m aptamer was used as RNA control to identify unspecific proteins. After that, the beads were washed three times with wash buffer (20mM Tris-HCl pH7.5, 300mM NaCl, 5mM MgCl₂, 2mM DTT, 2mM RNAse inhibitor, 1 protease inhibitor cocktail tablet), resuspended in 35 μL of Laemmli buffer [28] and boiled for 10 mins before application into 12% polyacrylamide gel. Samples were run at 110V until the samples reached the separation gel.

Proteomic analysis, mass spectrometry, DataBase searching and criteria for protein identification

Gel bands containing the samples were sent to protein identification by mass spectrometry analysis (Proteomics Platform of the CHU de Québec Research Centre, Quebec, Canada). Three biological replicates were evaluated for each protein and for the control. Results were obtained and analyzed using the software Scaffold Protein. The list of identified proteins was filtered using a protein threshold of 99%, a peptide threshold of 95% and a minimum of 1 peptide identified for all the samples. Proteins interacting with the control RNA sequence were unconsidered and the results were based on the proteins specifically interacting with the 3’UTR sequences in triplicate. Detailed information on Mass Spectrometry, Database searching and criteria for protein identification have been described elsewhere [23].

Starvation resistance assay

Nutritional stress was evaluated by incubating 10⁶ parasites per well in a 96 well plate in PBS for 4 h. After that, plate was centrifuged 1,400 x g for 5 minutes at RT, cells were resuspended in complete M199 with MTT (3-(4,5-DIMETHYL-2-THIAZOLYL)-2,5-DIPHENYL-2H-TETRAZOLIUM·BROMIDE) 1 mg·mL^-1 and incubated for 24 h at 26°C. After that, parasites were pelleted and 200 μL of DMSO was added to solubilize the tetrazolium crystals. Absorbance was measured at 560 nm considering the substrate conversion for cell non-starved as 100% of nutritional response. Experiments represent three biological replicates performed as quintuplicate technical replicates.

Sucrose density gradient

Promastigotes extracts from Leishmania major under normal growth condition were fractionated on sucrose gradient [29]. Briefly, 5 x10⁸ cells were previously incubated in cycloheximide 100 μg·mL^-1 for 10 min. Cells were kept on ice and washed once with TKM buffer (10 mM Tris, 10 mM KCl, 1 mM MgCl₂ and pH 7.5) supplemented with 100 μg·mL^-1 cycloheximide, 10 μg·mL^-1 heparin, 10 μg·mL^-1 E-64 (cysteine protease irreversible inhibitor–Sigma-Aldrich) and protease inhibitor cocktail (Roche). For polysome dissociation, cells were treated with puromycin 2mM. Cells were pelleted and 100 μL of lysis buffer (TKM supplemented with 10% surfactant NP-40 and 2 M sucrose) were lysed by 10 x up and down agitation with a P200 pipette, followed by centrifugation at 18,000 x g at 4°C for 10 min. The supernatant was added on the top of linear 10–50% sucrose density gradient prepared in the same buffer. The system was centrifuged at 39,000 rpm at 4°C for 2 h in a Beckman SW41 rotor. After centrifugation, the gradient fractions were collected using the ISCO gradient fractionation system with the same sensitivity in the instrument (1.0) and starting the gradient profile at the baseline of 30 cm on the ISCO paper for all gradient replicates, using 30 μL of each sample for blotting assays. For a semi-quantitative analysis, the scale with arbitrary values of A254 nm was inserted into the polysomal profile graph, since the ISCO instrument does not provide absolute absorbance values. To assign these values, we relied on the centimeter scale of the ISCO paper on which the gradients are plotted, considering the maximum absorbance as 1. The intensity values of the 80S and 40S peaks were determined considering the y axis values. Peaks height were estimated considering the values from their base to the top. In this way, the intensity of the 80S peak relative to the 40S peak can be determined.

Statistical analysis

Statistical t-test (Mann–Whitney) and one-way ANOVA followed by Tukey’s multiple comparison tests were performed using GraphPad Prism 8, considering as significant a p-value <0.05.

RPs from paralog genes have different expression levels and undistinguishable subcellular distribution

We investigated the expression of two pairs of duplicated paralogous genes encoding for RPs in L. major LV39 parasites: RPS16 (Fig 1A) and RPL13a (Fig 1B). RPS16 was studied as a model for paralogs with identical coding sequences and divergent UTRs (Fig 1C), whilst RPL13a we selected as a representative for paralogs with non-identical coding sequences and divergent UTRs (Fig 1D). Seven nucleotide substitutions were found when comparing the two copies of RPL13a genes; two of them led to a codon modification and to non-conserved amino acid substitutions (Fig 1D) however, such alterations lead to no detectable changes in their 3D structures (S1D Fig), suggesting the RPL13a paralogs may retain conserved functions. Next, we focused on the roles of the divergent 3’UTRs, which are known to be involved in the interaction with RNA binding proteins [30] and, therefore, in the control of gene expression by regulating or modulating the mRNAs transcriptional and translational rates, as well as their stability.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Paralog genes and proteins alignments.

(A) 40S ribosomal protein S16 genes are found in tandem on the minus strand of chromosome 26 and (B) 60S ribosomal protein L13a genes are found on distinct chromosomes (15 and 34) and in opposite directions. (C) RPS16 and (D) RPL13a gene alignments, highlight the divergences in the UTR sequences. (E) RPL13a amino acid sequence alignment showing the two amino acid substitutions (*).

https://doi.org/10.1371/journal.pone.0292152.g001

To investigate the fate of each of these RPs in L. major, we used CRISPR/Cas9 to fuse an endogenous epitope tag to the N-terminus of each protein as previously described [31]. Briefly, donor DNA containing 3xmyc epitopes and a blasticidin resistance gene was used to repair the Cas9-induced double strand break at 5’-end of each one of the paralogs (Fig 2A). The insertion of the tag in each copy of the RPS16 paralogs was confirmed by PCR and sequencing, showing that each individual paralog, RPS16_80 or RPS16_90, was efficiently tagged in both alleles (S1 and S6 Figs). For the RPL13a isoforms, due to the similarity of their amplicon size, the PCR was carried out for both copies, RPL13a_15 and RPL13a_34, allowing the confirmation of tag insertion in only one paralogue with the maintenance of unmodified paralogous copy (S1 Fig). These confirmed transfectants were subsequently used to check the individual expression level and subcellular localization of all four RP isoforms.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Protein levels and subcellular localization of duplicated S16 and L13a ribosomal proteins.

(A) Strategy for 3xmyc tag insertion in only one of the copies of the RP genes using CRISPR/Cas9. (B) Procyclic promastigotes were cultivated until the log phase from which metacyclic forms were purified by Ficoll gradient (see methods). Their morphological differences were confirmed by scanning electron microscopy. (C) Western blots probed with α-myc show that, for both RPs, the levels of protein for one paralog are consistently reduced in comparison to the other, and moreover, abundance is further reduced in metacyclic compared to procyclic promastigotes. (D) Immunofluorescence analysis confirms the cytoplasmic localization for all RP (green) proteins in procyclic promastigotes (nucleus [N] and kinetoplast [K] are indicated and stained in blue using Hoechst).

https://doi.org/10.1371/journal.pone.0292152.g002

Log-phase procyclic promastigotes and metacyclic promastigotes, purified from cultures in stationary phase by a Ficoll gradient fractionation [22], were isolated from axenic cultures and their morphologies examined by scanning electron microscopy to confirm their lifecycle stages; metacyclic promastigotes have elongated flagella and reduced cell body size when compared to the procyclic form (Fig 2B).

The relative protein abundance derived from each paralogous gene was evaluated by western blotting (Fig 2C) in both procyclic and metacyclic promastigotes. RPS16_80 and RPL13a_15 are consistently present in higher levels than their respective paralog in both lifecycle stages. Such differences in protein abundance were more prominent between the RPS16_80 and RPS16_90 isoforms in metacyclic promastigotes; RPS16_80 levels remain unchanged between procyclic and metacyclic stages but RPS16_90 was found to be lower in abundance when compared to RPS16_80; such difference between RPS16_80 e RPS16_90 levels is more striking in metacyclics.

The subcellular localization of RPS16 and RPL13a and their paralogs were investigated by indirect immunofluorescence. We detected myc signal for all RPs localizing (Fig 2D) exclusively to the cytoplasm, with no detectable co-localization with the nuclear or kinetoplastid Hoechst signal. Additionally, direct immunofluorescence performed for RPS16 using an α-RPS16 antibody raised in L. major LV39 confirmed the cytoplasmic localization in the myc-tagged RPS16 cell lines (S2 Fig). Importantly, the intensity of the immunofluorescence signal corroborated the western blotting results and confirmed the differences in the abundance of each RP isoform in procyclic promastigotes.

Profile of proteins interacting with the UTRs of the studied RP transcripts

Since the 3’UTRs of both duplicated genes are not conserved and the expression levels of each paralog differed, we performed pull-down assays using the S1m in vitro system [27] to identify proteins potentially involved in such regulatory mechanisms. Briefly, the 3’UTR sequences were determined based on TriTrypDB annotations (Fig 3A), fused to the S1m aptamer and transcribed in vitro. After RNA immobilization on streptavidin-coated magnetic beads, the 3’UTR sequences were incubated with log-phase L. major LV39 protein extract. The bound proteins were isolated and identified by mass spectrometry (MS). Gene ontology (GO) analysis revealed that the proteins interacting with RPL13a and RPS16 3’UTRs are mostly involved in peptide and ribosome biogenesis, respectively (Fig 3B). Among these proteins, six were identified as binding to all four 3‘UTR sequences (Fig 3C), and four of them are directly related to protein folding and ribosomal biogenesis and processing. We speculate that these four proteins common to all the examined UTRs are possibly core proteins that bind to mRNA of RP genes and might be involved in intra nuclear trafficking and nucleolar activities (Fig 3D).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Identification of proteins interacting with the 3’UTRs of the RPS16 and RPL13a paralogs mRNA.

(A) For the pulldown assays, fragments from the 3’UTRs of each gene were amplified and cloned into PUC-54-4xS1m; the position of the primers and lengths of the fragments are indicated. (B) Gene ontology analysis of proteins identified in vitro interacting with individual 3’UTR of each RPS16 and RPL13a paralog genes (shades of blue), and proteins interacting with both paralogs (red). (C) Specific and shared proteins for each duplicated gene were identified, with total of 24 and 18 for RPS16_80 and 90, respectively. For RPL13a paralogues, similar numbers of specific proteins were obtained: 30 and 31 proteins for RPL13a_15 and 34, respectively. (D) All 3’UTRs of ribosomal protein duplicated genes bound to proteins involved in ribosome pathways, when considered the five most relevant p-values (D). All the analyses were based on the results from three independent assays.

https://doi.org/10.1371/journal.pone.0292152.g003

Increased response to starvation and RP levels of mutant parasites

Considering that ribosomal activity is affected under nutritional stress potentiating or uncovering non-canonical functions of ribosomal proteins, as shown for other organisms [5–8], we analyzed the resistance to starvation by comparing parental cells (pT007) to RP transfectants knocked out for each one of the paralogs. To this end, cells carrying a myc tag in one of the paralogous gene was knocked out for its corresponding counterpart (Fig 4A) with all knockouts (Δ) confirmed by PCR (S3 Fig). The ability to rapidly adapt and respond to starvation is critical for Leishmania differentiation and survival. Procyclic promastigotes were incubated for 4h in PBS, with recovery of 24h in fresh medium, and viability measured via colorimetric assays using MTT (Fig 4B). Curiously, an increment of the parasite resistance to the nutritional stress was observed for all the Δ parasites compared to parental cells, as indicated by the higher mitochondrial activity measured by MTT conversion (Fig 4C).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Nutritional stress response and expression levels of RPs under starvation.

(A) After tagging one of the paralogues, the other one was deleted using CRISPR/Cas9. (B) Procyclic promastigotes in exponential growth were exposed to total starvation in PBS for 4h at 27°C, followed by viability assays, blotting and immunofluorescence analysis. (C) Cell viability was quantified by MTT assay at 24h post-starvation, to examine the recovery capacity of the mutant cells compared to parental line (pT007). (D) Expression levels of RPs were lower under starvation (St), particularly for RPS16_90 and RPL13a_34, with general recovery of parental levels after 24h of recovery in fresh medium (Rv) in comparison to the non-starved cells (control–Ct)–detection by western blotting with α-myc. (E) Immunofluorescence revealed no significant difference between the RPs distribution (green) under stress, but a subtle accumulation can be observed by increased signal in some regions of the cytoplasm (arrows) after 4h of total starvation (D).

https://doi.org/10.1371/journal.pone.0292152.g004

Additionally, the RP levels were evaluated by western blotting prior to and after 4 hours of starvation in PBS, as well as at 24hs post recovery in fresh medium. As shown (Fig 4D), upon starvation, a subtle decrease in the levels of both L13 and S16 proteins was observed, which were completely recovered to the unstarved levels after 24h in fresh medium. Interestingly, this trend was also observed for RPS16_90 and RLP13_34 which are consistently expressed at lower levels under physiological conditions (Fig 4D). Additionally, under stress conditions no marked changes in the subcellular distribution of these RPs were observed (Fig 4E).

Knockout of RPL13a paralogs reduced the amount of 60S and 80S ribosomal subunits

Given that CDSs encoding the RPL13a proteins are non-identical with two non-conserved amino acids substitutions at the N-terminus (Fig 1E), we investigated the ribosome organization of these Δ cells under physiological condition. In that direction, we conducted polysomic profile analyses to investigate the status of the assembly of the ribosomal subunits, translation activity and biomolecule complexes [29,32]. Cell extracts of parental and Δ cells bearing only one of the RPL13a paralogs were fractionated by a sucrose gradient after treatment with cycloheximide or puromycin. For comparison purposes, we first analyzed the profile of polysomic distribution in parental and Δ cells in the presence of cycloheximide and observed a decreased peak for the 60S ribosome, as well as a subtle decrease in the 80S complex with a slight increase in the 40S subunit for the ΔRPL13a_15 and ΔRPL13a_34. As shown in Fig 5A, the intensity of the 80S peak for the ΔRPL13a_15 and ΔRPL13a_34 mutants relative to the 40S peak (80S/40S) is 2.2-fold and 1.6-fold, respectively. For the parental cell, 80S peak intensity relative to 40S peak (80S/40S) is 1.5-fold. When this 80S:40S ratio is compared between the three profiles, a 1.5- fold (2.2 or 2.3/1.5) reduction in the intensity of the peaks corresponding to the monosomal fraction of the both RPL13amutants is observed with respect to the parental strain. Additionally, polysomal peaks were significantly reduced compared to the parental cell line. These findings suggest that knockout of both RPL13a_15 and RPL13a_34 affects protein synthesis, even in the presence of the corresponding paralog. Next, we analyzed the distribution of each RPL13a paralog throughout the non-polysomic and polysomic fractions. No clear distinction in the distribution of a specific paralog in the absence of the other was observed, with both paralogous proteins being detected at similar levels at sucrose densities corresponding to monosomes and light or heavy polysomes (Fig 5B). Additionally, both RPL13a paralogs were detected mainly at the density of light polysome fractions when treated with puromycin, as if large and small subunits were assembled but not organized for translation (Fig 5C).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Knockout of RPL13a paralogs decreases 60S, 80S and polysomes.

(A) RPL13a Δ mutants and the parental cell lines were subjected to sucrose gradient fractionation and polysomal profile determination, where FP: free polysomes fraction, PP: pre-polysomes, LP: light polysomes and HP: heavy polysomes. 254nm absorbance values were set arbitrarily and the relative intensity of the 80S and 40S peaks are shown in the dashed lines. Accumulation of both RPL13a isoforms in light polysomes fraction was determined by western blotting after cycloheximide treatment. (B) Additionally, both RPL13a isoforms were found in light polysome fraction (LP), even after ribosome dissociation by puromycin treatment. Experiments were performed in biological duplicate and the polysome profiles correspond to the average of each sample.

https://doi.org/10.1371/journal.pone.0292152.g005

Duplicated RPs present a paralog compensation that maintains protein levels

Next, we sought to check if different paralogs could compensate for the absence of each other in a compensatory mechanism to maintain the protein at parental levels. Protein levels were evaluated in both procyclic and metacyclic parental and knockout promastigotes by western blotting using a α-myc antibody. For all the studied proteins, when one of the genes was deleted, the tagged paralog protein levels increased compensating the lack of its counterpart (Fig 6A). This compensatory effect is more evident when the gene that is expressed at lower levels (RPS16_90 and RPL13a_34) is kept and their highly expressed paralog deleted (Fig 2C). Additionally, to confirm that the presence of at least one of the paralogs was necessary for parasite viability, we proceeded to knock out both RPS16 paralogs, but no double Δ clones could be recovered. Recovered clones had inserted two drug resistance genes from two rounds of transfections using CRISPR/Cas9 but all transfectants had retained at least one intact copy of the parental gene (S3 Fig).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Compensatory expression of RP paralogs in knockout cells.

(A) Expression levels in procyclic and metacyclic promastigotes were analyzed by western blotting using α-myc. The presence and absence (Δ) of the correspond RPS16 and RPL13a paralogues are indicated by + and -, respectively. Band intensity was quantified in ImageJ software and the results were plotted in the corresponding graph with the increment of band color indicated. (B) RT-qPCR was used to quantify the relative expression of RPL13a transcripts in the parental (WT) and Δ promastigotes. Expression of the gene of interest was normalized to the G6PDH housekeeping gene.

https://doi.org/10.1371/journal.pone.0292152.g006

In addition to protein abundance, gene expression was also evaluated at the transcript levels to evaluate at what level(s) the compensatory mechanism may occur. For that, total RNA from both ΔRPL13a procyclic promastigotes was extracted and transcript levels quantified by RT-qPCR utilizing specific primers for the divergent UTR regions of each L13a gene copy. Transcript levels in the Δ and parental cell lines were compared, and relative expression was normalized to the expression of the housekeeping gene Glucose-6-phosphate dehydrogenase (G6PDH). No significant differences in transcript levels were observed for either RPL13a paralogs (Fig 6B). This result indicates that RPL13a compensatory mechanism involves either alterations to the translation rate or protein stability. However, it is unlikely that it is related to increases in transcript levels and/or stability.