Cell culture

Trypanosoma brucei Lister 427 procyclic cells in logarithmic growth were used for all experiments. Cells were grown in SDM-79 supplemented with 5% FCS and 75 μg/ml hemin at 27^oC and 5% CO₂. [56] Transgenic trypanosomes were kept under the respective selection (G418 disulfate (15 μg/ml), blasticidin S (10 μg/ml), puromycin dihydrochloride (1 μg/ml) and hygromycin B (25 µg/ml); for selection after a transfection, antibiotic concentrations were doubled). Growth was monitored by sub-culturing cells repeatedly to 10⁶ cells/ml and measuring densities 24 hours later using a Coulter Counter Z2 particle counter (Beckman Coulter). Transgenic trypanosomes were generated by standard procedures, relying on the transfection of linearized DNA by electroporation [57]; correct integration of plasmids/PCR products and/or correct protein expression was confirmed via PCR from genomic DNA and/or western blot, respectively.

Auxin inducible protein degradation

The auxin-inducible degron system (AID2) [58] was adapted to trypanosomes and kindly provided by Prof. Mark Carrington, University of Cambridge (manuscript in preparation). Briefly, the mother cell line was made by integrating five proteins from Oryza sativa, each fused to a Ty1 epitope tag for western blot detection, into the tubulin locus of procyclic trypanosomes for endogenous expression, via two plasmid transfections. The first plasmid transferred the genes for OsSKP1, OsCUL1, OsRBX1; the second the genes for OsTIR1(F74G) and OsARF. In this cell line, both alleles of the target protein were modified to express 3HA-AID2 fusion variants (either N or C terminal tagged), by two subsequent transfections using PCR tagging with the (modified) pPOTv7 system [59] with two different selection markers (one for each allele). Protein degradation was induced by adding 50 µ M of 5-Ph-IAA (MedChem Express, HY-134653) directly to the culture, from a 50 mM 5-Ph-IAA stock (in DMSO, stored at -20^oC for short-term use or -80^oC for long-term).

Plasmids and PCR products

Detailed information on plasmids, PCR products and primers used in this study is available in S1 Table. To generate the knockouts, PCR products were created by amplification of the resistance marker with 80 nucleotides overhangs of the 5’ and 3’ untranslated region (UTR) of the gene of interest [60]. The expression of fusion proteins from the endogenous locus was done using the pPOTv7 system [59]. A cell line used as a negative control of the TurboID experiment expressed a codon-optimized eYFP-TurboID-HA [61] with the La protein NLS [62] integrated into the tubulin locus.

Western blotting

Western blotting was performed using standard methods, usually loading about 5x10⁶ cells per lane. Primary antibodies were rat anti-HA (3F10, Roche) (1:1000) and anti-BiP (1:100,000) (kind gift from Jay Bangs) and secondary antibodies were IRDye^® 680 RD and 800 CW rat and rabbit anti-goat (LI-COR) (1:30,000). Detection and quantification was done using the Odyssey Infrared Imaging System (LI-COR Biosciences, Lincoln, NE). For quantification, the average of a 3-pixel width line at the top and bottom of each band was subtracted from each pixel for background correction.

Immunofluorescence

Imaging of TurboID-HA tagged proteins was done with anti-HA and fluorescent streptavidin, essentially as described [55]. Briefly, about 1 ×  10⁷ procyclic-form T. brucei cells, harvested at a density of 5 ×  10⁶ cells/ml, were washed once in 1 ml hemin- and serum-free SDM79, then fixed in 1 ml of 4% paraformaldehyde. After addition of 7 ml PBS supplemented with 20 mM glycine, cells were pelleted, resuspended in 150 μl PBS, and spread on poly-lysine-coated slides. Cells were permeabilized with 0.5% Triton X-100 in PBS, then rinsed in PBS, and blocked in 3% bovine serum albumin (BSA) in PBS for 30 min, followed by 60 min incubation with rabbit mAb-anti-HA C29F4 (1:500 dilution; Cell Signaling Technology) and Streptavidin-Cy3 (1:200 dilution; Jackson Laboratories) in PBS/3% BSA. Slides were washed in PBS (three times for 5 min), then incubated with anti-rabbit Alexa Fluor Plus 488 (1:500 dilution, A32731 Invitrogen) in PBS/3% BSA for 50 min and a further 10 min upon addition of 4’,6-diamidino- 2-phenylindole (DAPI) (0.1 μg/ml). Slides were washed 3 ×  5 min in PBS and embedded into ProLong Diamond Antifade Mountant (Thermo Fisher Scientific).

Affinity purification of biotinylated proteins, mass spectrometry and analysis

Affinity purification of biotinylated proteins and tryptic digest and peptide preparation were performed as described [61], except that 1 mM biotin was added to the on-bead tryptic digests, to improve the elution of biotinylated peptides. Eluted peptides were analyzed by liquid chromatography coupled to tandem mass spectrometry (LC-MSMS) on an Ultimate3000 nano rapid separation LC system (Dionex) coupled to an Orbitrap Fusion mass spectrometer (Thermo Fisher Scientific).

Spectra were processed using the intensity-based label-free quantification (LFQ) in MaxQuant version 2.1.3.0 [63,64], searching the T. brucei brucei 927 annotated protein database (release 64) from TriTrypDB [65]. Analysis was done using Perseus [66] essentially as described in [67]. Briefly, known contaminants, reverse and hits only identified by site were filtered out. LFQ intensities were log₂-transformed and missing values imputed from a normal distribution of intensities around the detection limit of the mass spectrometer. A Student’s t-test was used to compare the LFQ intensity values between the duplicate samples of the bait with an untagged control (WT parental cells) triplicate samples. The -log₁₀ p-values were plotted versus the t-test difference to generate multiple volcano plots (Hawaii plots). Potential interactors were classified according to their position in the Hawaii plot, applying cut-off curves for significant class A (SigA; FDR =  0.01, s0 =  0.1) and significant class B (SigB; FDR =  0.05, s0 =  0.1). The cut-off is based on the false discovery rate (FDR) and the artificial factor s0, which controls the relative importance of the t-test p-value and difference between means (at s0 =  0 only the p-value matters, while at non-zero s0 the difference of means contributes). Perseus was also used for principal component analysis (PCA), the profile plots and to determine proteins with similar distribution in the plot profile using Pearson’s correlation. Gene Ontology analysis was done on TriTrypDB [65], with the option to use only slim terms and the p-value was adjusted with a Bonferroni correction. All proteomics data have been deposited at the ProteomeXchange Consortium via the PRIDE partner repository [1] with the dataset identifier PXD050946.

Single molecule fluorescence in situ hybridization (smFISH)

smFISH was done as previously described [68]. The sequences used to design the Affymetrix FISH probes are in S1 Table.

Microscopy and quantification of microscopy data

For Fig 2B, S2B Fig and S6 Fig, images were acquired using a DMI8 widefield microscope (Thunder Imager, Leica Microsystems, Germany) with a HCX PL APO CS objective (100x, NA =  1.4, Leica Microsystems) and Type F Immersion Oil (refractive index =  1.518, Leica Microsystems). The microscope was controlled using LAS-X software (Leica Microsystems). Samples were illuminated with an LED8 light source (Leica). Excitation light was selected by using the following filter sets: DAPI - Ex 391/32 nm; DC 415nm; Em 435/30 nm (used for DAPI). FITC - Ex 436/28 nm; DC 500 nm; Em 519/25 nm. mCherry – Ex 578/24 nm; DC 598 nm; Em 641/78 nm. Cy5 – Ex 638/31 nm; DC 660 nm; Em 695/58 nm. Images were captured using a K5 sCMOS camera (Leica, 6.5 µm pixel size). Z-stacks (75 slices, 140 nm step size) were recorded. For Fig 2A and S5 Fig, a custom-built TILL Photonics iMIC microscope equipped with a 100x, 1.4 numerical aperture objective (Olympus, Tokyo, Japan) and a sensicam qe CCD camera (PCO, Kehlheim, Germany) was used. eYFP was monitored with the FRET-CFP/YFP-B-000 filter, mCherry with the ET-mCherry-Texas-Red filter and DNA with the DAPI filter (Chroma Technology CORP, Bellows Falls, VT). Z-stacks (75 slices, 140 nm step size) were recorded. All images were deconvolved using Huygens Essential software (SVI, Hilversum, The Netherlands). All data were analysed with Fiji [69].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Growth of Y14/Magoh knockout strains and cells upon auxin-induced degron depletion of eIF4AIII.

A) Growth of wild type cells (WT), heterozygote (+/-) and homozygote (-/-) Y14 and Magoh knockout cells and Magoh/Y14 double knockout cells was monitored over 96 hours. The population doubling time (PDT) is indicated on the right (mean of PDT of all clones ±  standard deviation) B and C) Both eIF4AIII alleles were replaced by eIF4AIII-AID2-3HA in a cell line expressing all proteins required for the AID2 degron system. eIF4AIII degradation was induced by the addition of 50 µ M 5-Ph-IAA. The loss in eIF4AIII-AID2-3HA protein was monitored on a western blot, loaded with protein samples taken after 2 hours of induction, and protein samples of three different clonal cell lines taken after 96 hours of induction (B). For two clonal cell lines, the growth was monitored in the presence and absence of 5-Ph-IAA (C). Note that the degron mother cell line has a slightly reduced growth rate.

https://doi.org/10.1371/journal.pone.0315659.g001

BLASTp screen for presence of UPF sequences

The UPF1 and UPF2 sequences from Trypanosomatids and other kinetoplastids were obtained by using T. brucei (Tb927.5.2140 = UPF1; Tb927.11.6410 = UPF2) protein sequence as query on BLASTp [70] searches on TritrypDB [65], GenBank [71] and ENA [72] databases. The absence of UPF3 in Kinetoplastida was confirmed by blasting with human UPF3. The UPF sequences from other eukaryotes were obtained searching the same databases with human UPF sequences. Sequences from organisms that were not available on the online searches were obtained by downloading full transcriptome data and using it on local BLASTp searches on BioEdit 7.2.5 [73]. Sequence results with e-value less than 1x10^-3 were discarded. All sequences are available in S3 Table. Alignments were visualized using Jalview [74].

Flow cytometry and Northern blot analysis of PTC reporters

Northern blots were done as described [75]. For flow cytometry, data were acquired using FACScan (Becton Dickinson, Franklin Lakes, NJ) and analysed using Cell Quest V3.3 software. Flow Check Fluorospheres (Beckman Coulter, Pasadena, CA) were used for normalisation.

RT-qPCR analysis of PTC reporters

eGFP sequences, either with a native stop codon or premature stop codons at positions 60 and 120, were integrated into tubulin locus for constitutive expression. All cell lines were confirmed by Sanger sequencing. UPF1 degradation was induced for 2 hours with 50 µ M 5-Ph-IAA and RNA was extracted (RNeasy Mini Kit, Qiagen, according to manufacturer’s protocol and including the optional in column DNAse digestion step). RNAs were diluted to 50 ng/µl and used for RT-qPCR with iTaq Universal One-Step RT-qPCR Kit (BIORAD) in a StepOne™ Real-Time PCR System (ThermoFisher). The reaction was done in 10 µl total volume using 100 ng RNA, 1 µ M of each forward and reverse oligo and the cycler program of 50 °C for 20 min, 95 °C for 1 min and 40 cycles of 95 °C for 15 sec and of 60 °C for 10 sec. Each sample was analysed in triplicates. GAPDH oligos were taken from [76]. eGFP oligos were designed using the IDT PrimerQuest™ tool (sequences in S1 Table). The CT values for eGFP were controlled for loading with GAPDH (ΔCT) and the average value of the native stop codon without 5-Ph-IAA treatment was set as parental for the calculation of the 2-ΔΔCT.

eIF4AIII is essential, Magoh and Y14 are not

We attempted to generate genomic knock-out cell lines for Y14, Magoh and eIF4AIII. We transfected PCR products consisting of a resistance marker flanked by 80 nucleotides with homology to the respective 5’- and 3’ regions to replace both alleles of each gene, in two subsequent transfections [59]. For Magoh and Y14, multiple double-resistant cell lines were readily obtained, and the absence of the wild type gene was confirmed by PCR analyses (S2A–S2C Fig). Furthermore, we could delete both Magoh alleles in a Y14 knockout strain, creating a double knockout. All knockout strains displayed only a minor growth phenotype (10% reduced growth rate for the Y14 knockout, 20% for the Magoh knockout and Y14/Magoh double knockout) (Fig 1A) which could not be assigned to a block in a specific cell cycle phase (S2D Fig). Silencing of Magoh in Paramecium or knockout of Y14 in Y. lipolytica caused changes in intron retention, indicating a function in splicing [77,78]. We therefore analysed the T. brucei Y14/Magoh double knockout for splicing defects, by probing for exon and intron of the intron-containing mRNA DBP2B of T. brucei, with dual colour single molecule smFISH [68]. We observed no differences in numbers or cytoplasmic/nuclear RNA distribution between wild type cells and Y14/Magoh double knockout strains, indicating that Y14 and Magoh are not required for processing, export or stability of cis-spliced mRNA (S3 Fig). Surprisingly, we were able to detect both intron-containing mRNAs and intron alone in the cytoplasm, even in wild type cells, representing additional evidence for the absence of a tight quality control system of mRNA nuclear export in trypanosomes.

For eIF4AIII, the first allele could readily be deleted, but multiple trials to knock out the second allele failed, indicating that eIF4AIII is essential. We therefore employed the auxin-inducible degron system (AID2) [58] which was recently adapted to trypanosomes (Mark Carrington, manuscript in preparation). Briefly, both eIF4AIII alleles were replaced by eIF4AIII fused to OsAID2-3HA in a cell line constitutively expressing the five Oryza sativa proteins required for the degron system (details in material and methods). The absence of a wild type eIF4AIII allele was confirmed by PCR (S4A–S4C Fig). Induction of the system with the auxin analogue 5-Ph-IAA resulted in almost complete degradation of eIF4AIII within 2 hours which was not reversed (Fig 1B, Fig 1 in S1 Raw Images). Cell growth was significantly reduced (Fig 1C). No specific cell cycle phase was affected, albeit there was a small increase in the number of “zoids”, anucleate cells that have a kinetoplast only (S5 Fig).

In conclusion, eIF4AIII is essential in cultured trypanosomes, but Magoh and Y14 are not, arguing against an exclusive role of these proteins within the same complex. This is in agreement with data from previous RNAi experiments [42,43].

eIF4AIII, Magoh and Y14 are nuclear proteins, with eIF4AIII being preferentially nucleolar

To determine the subcellular localization of Magoh, Y14 and eIF4AIII, we expressed these proteins as eYFP fusions, by modifying one endogenous allele [59]. Magoh and eIF4AIII were expressed as C-terminal fusions, Y14 was tagged at the N-terminus. The RNA helicase mChFP-DHH1, expressed from the endogenous locus, served as a cytoplasmic control. eYFP fusions of all three proteins were highly enriched in the nucleus, while DHH1 was entirely cytoplasmic (Fig 2A and S6 Fig). Magoh, Y14 and eIF4AIII eYFP fusion proteins were present in both the nucleolus (defined by the absence of DAPI stain) and the nucleoplasm. However, we noticed two differences between Y14/Magoh and eIF4AIII localisations: (i) eIF4AIII was more enriched in the nucleolus, while Y14 and Magoh were more evenly spread throughout the entire nucleus and (ii) a small fraction of eIF4AIII, but not of Magoh or Y14, was cytoplasmic (see fluorescence-profiles across the nuclei in Fig 2A), in agreement with eIF4AIII shuttling between nucleus and cytoplasm, as previously reported [79]. The data agree with localizations obtained by the genome-wide localization project TrypTag [80] and for Y14 also with immunofluorescence data using fluorescent anti-IgG against Y14 fused to the TAP tag [42].

We observed no major changes in protein localization when we manipulated mRNA metabolism with actinomycin D (inhibition of transcription), sinefungin (inhibition of trans-splicing) or PBS-starvation (global inhibition of translation) (S6 Fig). All three proteins remained highly enriched in the nucleus. We observed a minor increase in cytoplasmic localization for eIF4AIII at starvation stress; as this protein appears to shuttle the inhibition of global mRNA metabolism may slightly shift the distribution of eIF4AIII between the two compartments (S6C Fig). Moreover, Y14 fluorescence appeared shifted from the nucleoplasm to the nucleolus at sinefungin treatment (S6A Fig).

For eIF4AIII (alias Hel46), our data contradict a range of immunofluorescence experiments done in T. cruzi and T. brucei, with antibodies against the native protein or anti-Protein A to a PTP-tagged version [79]. The previous authors reported a dominant cytoplasmic localization of eIF4AIII, with a nuclear localization being enforced only when nuclear export was blocked or a putative NES removed. It is possible, that the nuclear localization we and others observe with fusion proteins to GFP derivatives is an artefact of the tag, but, we consider this unlikely, because eYFP tagging at both the C- and the N-terminus results in the same nuclear localization [80] and, we do know that a C-terminal fusion to AID2-3xHA (which is similar in size to eYFP) is fully functional in the absence of wild type eIF4AIII (Fig 1B and 1C, Fig 1 in S1 Raw Images). An alternative explanation for this discrepancy is that antibodies fail to label eIF4AIII inside the nucleolus, because they cannot penetrate phase-separated areas. We have previously observed this for the T. brucei nucleolar protein NOG1 and for a range of proteins localized to other phase-separated compartments [55]. In fact, an older study shows marked differences in eIF4AIII localization between antibody detection (nuclear) or detection of an eYFP-fusion (nucleolus) [43]. We have recently proposed the use of biotin proximity labeling in combination with streptavidin to overcome antibody-accessibility problems: the target protein is fused to the biotin ligase TurboID and the resulting (auto-)biotinylation of this bait and adjacent proteins is detected by fluorescent streptavidin, which readily enters phase-separated areas [55]. Hence, we endogenously expressed TbeIF4AIII C-terminally fused to a tandem tag of the biotin-ligase TurboID [81] and 3HA epitope tags. We imaged eIF4AIII localization in parallel with both fluorescent streptavidin and anti-HA (Fig 2B and S7 Fig). The streptavidin signal was in the nucleolus and partially in the nucleoplasm, resembling the eYFP signal. In contrast, anti-HA weakly stained the nucleoplasm and partially the cytoplasm but was excluded from the nucleolus, proving that eIF4AIII can indeed not be correctly localized with immunofluorescence detection and providing an explanation for the discrepancies between the data. We have repeated the experiment with Magoh and Y14 (Fig 2B and S7 Fig). With streptavidin, we obtained both nucleolar and nucleoplasmic localisation, in agreement with the eYFP fluorescence data (Fig 2A), while the anti-HA signal was restricted to the nucleoplasm and absent from the nucleolus (Fig 2B). Interestingly, the cytoplasmic HA signal was significantly stronger for eIF4AIII than for Magoh and Y14 (images and profiles in Fig 2B), providing further evidence for eIF4AIII being a shuttling protein. This was not observed for streptavidin; a possible reason is that streptavidin detects all eIF4AIII-proximal proteins and the majority of eIF4AIII interactors may not be shuttling.

To summarise, imaging with either eYFP or streptavidin localized Magoh, Y14 and eIF4AIII to the nucleoplasm and the nucleolus, with eIF4AIII being mostly nucleolar and Y14 and Magoh being more equally spread between the nucleolus and the nucleoplasm. eIF4AIII, but not Y14 and Magoh, was also partially cytoplasmic. As the exon junction complex functions in RNA splicing and processing, the expected localization is nucleoplasmic. The observed nucleolar localization of eIF4AIII, and to a lesser extent of Magoh and Y14, therefore contradicts such canonical roles. Moreover, the differences in localizations argue against the exclusive presence of the three proteins in one complex.

eIF4AIII and Magoh/Y14 associate with different cohorts of proteins

To explore the interactome of eIF4AIII, Y14 and Magoh, we employed the available cell lines that expressed the TurboID-HA fusion proteins (Fig 2B). For each bait protein, we carried out two independent streptavidin-affinity purifications and subjected peptides eluted by on-bead tryptic digest to liquid chromatography coupled to tandem mass spectrometry (LC-MSMS) analyses. Intensities obtained from label-free quantification formed distinct replicate clusters in a principal component analysis for all bait and control groups (Fig 3A). For Magoh and Y14, 97 and 367 protein groups (Fig 3B and 3C; S2 Table), respectively, were found enriched (SigB; see methods section) indicating a considerable degree of bystander-labelling. However, Magoh was enriched in the Y14 TurboID and vice versa, strongly suggesting interactions between these two factors. eIF4AIII was absent from the Y14 and Magoh interactomes and, consistently, neither Magoh nor Y14 were among the 131 protein groups enriched (SigB) in the eIF4AIII TurboID experiment (Fig 3B and 3C; S2 Table). Instead, two nucleolar proteins, CWC22 and NOM1 (Sgd1p in yeast), were specifically enriched with eIF4AIII, but not with Y14 or Magoh. Both are homologs of known eIF4AIII interactors in opisthokonts [24,82]. Interestingly, CWC22 is a spliceosomal factor, required for EJC formation in mammalian cells [24] and NOM1 was proposed to function, EJC independently, in pre-ribosomal RNA processing [82]. The potential BTZ ortholog, RBP25, was enriched for all three bait proteins. Further candidate interactors are listed in S2 Table but remain at a low confidence level due to apparent high-level non-specific background, which may be partly attributed to the high abundance of the three bait proteins [83]. GO-term enrichment analyses for the three bait proteins (Fig 3D; S2 Table) reveals differences between eIF4AIII and Y14/Magoh. Whilst all baits show similar enrichment of the GO component ‘nucleus’, components ‘nucleoplasm’ and ‘nucleosome’ are enriched exclusively for Y14 and Magoh, while ‘nucleolus’ is unique for eIF4AIII. However, it is unclear whether these differences reflect different interactions or are merely reflecting bystander labeling profiles evoked by the more dominant eIF4AIII nucleolar localization.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Localisation of the EJC complex components.

A) Cells expressing eYFP fusion proteins of Y14, Magoh and eIF4AIII in a cell line co-expressing mChFP-DHH1 were fixed and imaged; DNA was stained with DAPI. Z-stack projections (75 slices a 140 nm, sum slices) of one representative cell are shown; the corresponding nucleus is shown enlarged. For each cell line, fluorescence profiles through three nuclei are shown. More images, including images of cells treated with actinomycin D, sinefungin or starvation are in S5 Fig B) Cells expressing eIF4AIII-HA-TurboID were labelled with streptavidin-Alexa 488 (green) and anti-HA (secondary antibody Alexa 594 labeled, shown in pink). DNA was stained with DAPI (blue). Deconvolved Z-stack projections (100 slices a 100 nm) of one representative cell are shown; the corresponding nucleus is shown enlarged. For each cell line, fluorescence profiles through three nuclei are shown.

https://doi.org/10.1371/journal.pone.0315659.g002

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Magoh, Y14, and eIF4AIII TurboID interactomes.

A) Principal component analysis (PCA) of the LFQ values obtained from TurboID duplicate experiments with Magoh (red), Y14 (blue), eIF4AIII (green) and three replicate parental (WT) controls (white). B) Corresponding Hawaii-plot (multiple comparative volcano plots) for statistical analysis of TurboID experiments. To generate the volcano plots, the −  log₁₀P-value was plotted versus the t-test difference, comparing each respective bait experiment to the parental control. Potential interactors were classified according to their position in the plot, applying two statistical cut-off curves, termed SigA and SigB (see Methods section). Selected protein groups are coloured (green for eIF4AIII and previously described interactors; red for Magoh; blue for Y14; white for the BTZ homolog RBP25). C) Diagram showing log₁₀-transformed normalized LFQ intensities of baits and high-confidence interactors for each experiment replicate (color coded as in (A); blank cells indicate no detection) including a nuclear eYFP-TurboID control (eYFP_NLS; see methods section). D) Diagram showing selected GO terms enriched within the SigB cut-off from any of the three bait experiments (color coded as in (A)) with p-values <  0.001. Numbers in each cell indicate the enrichment factor and the intersection size (number of protein groups with the respective GO-term within SigB) in brackets. The selection size (total number of protein groups in SigB) is indicated at the bottom; the category size (total number of GO-term occurrence among all 927 detected protein groups) is indicated on the left next to GO IDs; white cells indicate no or insignificant GO enrichment. For all proteomics data and complete GO enrichment analyses see S2 Table.

https://doi.org/10.1371/journal.pone.0315659.g003

Of interest was the absence of the NTF2L protein from the TurboID interactomes of all three EJC proteins, even though the protein was previously coprecipitated with T. brucei Y14 [42] and T. cruzi eIF4AIII [44]. Given that all three EJC complex proteins have exclusive (Magoh and Y14) or dominant (eIF4AIII) nuclear localization, while NTF2L is in the cytoplasm [80], it is possible that the interactions observed in the immunoprecipitations are an in vitro artefact of the cell lysis that allows interactions of proteins that are normally separated. However, this does not fully rule out an artefactual NTF2L absence in the TurboID. RNAi depletion of NTF2L caused only a minor growth effect and no nuclear mRNA accumulation, arguing against any essential function [44]. We show that even a fast reduction of NTF2L protein to 10% using the AID2 system caused only a minor reduction in growth (S4D Fig and S8 Fig, Fig 3 in S1 Raw Images). Our data provide no evidence for an essential function of NTF2L, within the EJC or elsewhere, in cultured procyclic trypanosomes.

UPF1 dependent NMD is functional but not essential

The absence of a canonical EJC does not exclude the presence of an NMD pathway: NMD is EJC-independent in the majority of eukaryotes. T. brucei has homologues to UPF1 and UPF2 that interact, but RNAi depletion in either procyclic or bloodstream form trypanosomes had no effect on growth or the transcriptome [47]. Consistently, we observed only a minor growth defect upon reducing UPF1 levels to approximately 2% via the AID2 degron (Fig 4A and 4B, Fig 2 in S1 Raw Images and S4E Fig). We specifically looked for changes in intron retention in the UPF1 depleted cells with smFISH, by probing for DBP2B, as intron retained DBP2B contains a stop codon downstream of the intron but detected no differences (S3D Fig). Next, we expressed an eGFP reporter gene with either an early PTC (position 60), a late PTC (position 120), or the complete ORF (native, stop at position 240) constitutively from the tubulin locus (Fig 4C). While both premature stop codons resulted in the expected loss of GFP protein (Fig 4D), northern blotting revealed that the presence of a PTC significantly reduced the reporter mRNA levels in a position specific manner (Fig 4E, Fig 4 in S1 Raw Images): Early and late PTC evoked approximate mRNA decreases of 94% and 37%, respectively, when compared to the mRNA encoding native eGFP. This is consistent with earlier studies reporting instability of PTC-containing reporters [47,48,86]. Employing our AID2 degron system we could now test whether UPF1 depletion effects the stability of these PTC reporters. For technical reasons we had to switch from Northern blots to qPCR, but we observed a similar PTC-related decrease in mRNA levels as with northern blot, albeit slightly less severe for the early PTC reporter (by only 74% instead of 94%). To our surprise, depletion of UPF1 almost doubled the levels of the early PTC reporter, reaching 59% of the levels of the native GFP mRNA. This strongly suggests that UPF1-dependent NMD is functional (Fig 4F). Unexpectedly, the AID2 degron depletion of UPF1 also led to a mild but significant decrease of native eGFP mRNA levels by approximately 14% which could be a rather unspecific consequence triggered by the observed loss of fitness. This effect may have prevented that mRNA levels of the early-PTC reporter were fully restored to the levels of the native GFP. Moreover, it could also explain why late PTC mRNA levels remained unaffected by UPF1 depletion (Fig 4F). Another potential factor obscuring the rescue is the residual abundance of transcripts formed before complete UPF1 depletion during the 2h induction time, chosen to restrict secondary effects. Nevertheless, our data show that UPF1 contributes to NMD, but they do not fully rule out the possibility that additional factors contribute. Notably, UPF1 deletion in other organisms with an EJC-independent NMD, such as S. pombe, also does not completely restore the steady state levels of early PTC reporters [87]. We thus conclude that there is an EJC-independent NMD-like pathway in T. brucei which significantly relies on UPF1.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. NMD in Kinetoplastea.

A and B) Both UPF1 alleles were replaced by UPF1-AID2-3HA in a cell line expressing all proteins required for the AID2 degron system. UPF1 degradation was induced by the addition of 50 µ M 5-Ph-IAA. The loss in UPF1-AID2-3HA protein was monitored on a western blot and quantified (A). For two clonal cell lines, growth was monitored in the presence and absence of 5-Ph-IAA (B). Note that the degron mother cell line has a slightly reduced growth rate. C) A schematic of the reporters that express eGFP with native stop codon and premature stop codons (PTC) at the positions 60 and 120 (in scale). D) Flow cytometry analysis shows loss of eGFP fluorescence in the presence of a PTC. E) Northern blot of RNA samples of three clones expressing the different eGFP reporters. The membrane was probed for eGFP mRNAs and rRNA (loading control). The bar graph shows average quantified GFP mRNA levels relative to rRNA from the three independent clones. F) Relative RNA levels of the reporters upon UPF1 depletion using the auxin inducible degron (2 hours induction). The bar graph shows average values from triplicate experiments and corresponding error bars. eGFP RNA levels were quantified using RT-qPCR. Levels were normalised with GAPDH levels and calculated with the 2-ΔΔCT method. The average ΔCT values of the reporter with the native stop codon, without 5-Ph-IAA treatment, was used as the parental ΔCT values to calculate the 2-ΔΔCT values of all samples. G) The presence (coloured) or absence (not-coloured) of UPF1, 2 and 3 is shown schematically for Euglenozoa. The tree is based on 18S rDNA phylogeny data taken from [84]; the Trypanosomatida branch was manually added [85] and is not to scale.

https://doi.org/10.1371/journal.pone.0315659.g004

Homologues to UPF1, UPF2 and UPF3 are present across the eukaryotic kingdom (S9 Fig) and must have been present in the last eukaryotic common ancestor. Losses have since occurred in all branches of the eukaryotes with rather patchy distribution, mainly of UPF2 and UPF3, and very occasionally of the more conserved UPF1. Most of the few UPF1 losses are isolated, e.g., in single organisms within UPF1-encoding clades; these may be very recent losses or could be attributed to incomplete genome sequence information. The one notable exception is the group of Kinetoplastea: UPF3 is absent in all representative organisms, some have additionally lost UPF2 and others have lost all three UPF variants. Notably, the losses of UPF1 and UPF2 must have occurred independently on several occasions, for example in T. congolense and in the ancestor of the Leishmaniinae or in Paratrypanosoma (Fig 4G). The loss of UPF isoforms may be the direct consequence of the almost complete loss of introns that is characteristic for Trypanosomatids. Consistent with this hypothesis, all organisms of the two intron-rich sister groups of the Kinetoplastids, Diplonemids and Euglenids, possess all three UPF isoforms. Notably, unlike cis-splicing, trans-splicing does not create PTCs. In the (near) absence of the former, NMD may have become redundant, hence lifting the selective pressure to maintain the UPF1/2/3 genes.

In conclusion, we provide evidence for the presence of an EJC independent NMD-like pathway in T. brucei that is fully or partially dependent on UPF1. Based on the missing essentiality of UPF1 together with the occurrence of UPF1 losses within Kinetoplastida and the failure of UPF1 depletion to fully restore the stability of early PTC reporter RNAs (Fig 4) it remains possible that additional, UPF1-independent mechanisms exist, but this possibility requires further exploration. Ribosome processivity, for example, influences mRNA stability in trypanosomes [88] and yeast [89]. Given that abnormally long 3’-UTRs contribute to NMD triggering [90] a further open question how NMD is mediated in the presence of long 3’-UTRs present in many trypanosome genes [47].