Knockout of protein phosphatase 1 in Leishmania major reveals its role during RNA polymerase II transcription termination

Abstract

The genomes of kinetoplastids are organized into polycistronic transcription units that are flanked by a modified DNA base (base J, beta-D-glucosyl-hydroxymethyluracil). Previous work established a role of base J in promoting RNA polymerase II (Pol II) termination in Leishmania major and Trypanosoma brucei. We recently identified a PJW/PP1 complex in Leishmania containing a J-binding protein (JBP3), PP1 phosphatase 1, PP1 interactive-regulatory protein (PNUTS) and Wdr82. Analyses suggested the complex regulates transcription termination by recruitment to termination sites via JBP3-base J interactions and dephosphorylation of proteins, including Pol II, by PP1. However, we never addressed the role of PP1, the sole catalytic component, in Pol II transcription termination. We now demonstrate that deletion of the PP1 component of the PJW/PP1 complex in L. major, PP1-8e, leads to readthrough transcription at the 3’-end of polycistronic gene arrays. We show PP1-8e has in vitro phosphatase activity that is lost upon mutation of a key catalytic residue and associates with PNUTS via the conserved RVxF motif. Additionally, purified PJW complex with associated PP1-8e, but not complex lacking PP1-8e, led to dephosphorylation of Pol II, suggesting a direct role of PNUTS/PP1 holoenzymes in regulating transcription termination via dephosphorylating Pol II in the nucleus.

Graphical Abstract

Graphical Abstract.

INTRODUCTION

Kinetoplastids are flagellated protozoans that are among the earliest diverging eukaryotes with a mitochondrion, whose members include pathogens responsible for multiple human diseases including human African trypanosomiasis (African sleeping sickness) and leishmaniasis. Kinetoplastid parasites, which include Trypanosoma brucei and Leishmania major, progress through life stages by cycling between an insect vector and a mammalian host. Unlike most other eukaryotes, the entire genome of kinetoplastids is arranged into polycistronic transcription units (PTUs), consisting of tens to hundreds of genes co-transcribed from an initiation site at the 5’-end to the termination site at the 3’-end of the PTU (1,2). PTUs can be arranged in opposing directions in the genome where transcription terminates at convergent strand switch regions (cSSRs) (3). PTUs are also adjacently arranged on the same DNA strand in what is called a head-to-tail (H-T) or unidirectional arrangement, such that transcription of an upstream cluster terminates and initiation of a downstream gene cluster occurs (4–5). Pre-mRNAs are processed through trans-splicing with the addition of a 39 nucleotide spliced leader (SL) sequence to the 5’ end of mRNAs (6–11) (reviewed in (12)), which is coupled to the 3’ polyadenylation of the upstream transcript (13). This unique genomic organization has led to a model that gene expression is primarily controlled via post-transcriptional mechanisms such as mRNA processing, mRNA stability, and translation efficiency.

Regulation of RNA polymerase II (Pol II) transcription termination within PTUs provides a novel way to regulate protein coding gene expression in these parasites. Multiple chromatin modifications are enriched at sites of Pol II termination, including histone variants and the DNA modification base J that could be utilized by the parasites to regulate transcription (4,1,14,15). Base J consists of a glucosylated thymidine (16) and has been found only in the nuclear DNA of kinetoplastids, Diplonema, and Euglena (17,18). In kinetoplastids base J is found at almost all Pol II transcription termination sites (15) (19). J biosynthesis occurs through the hydroxylation of thymidine by either J-binding protein 1 (JBP1) or JBP2 (20), forming hydroxymethyluridine, followed by the transfer of a glucose by the glucosyltransferase enzyme JGT (21,22) (reviewed in (23,24)). Both JBPs have an N-terminal dioxygenase domain analogous to the TET proteins in mammals (25,26), and utilize 2-oxoglutarate, oxygen, and Fe²⁺ in the hydroxylation reaction (20). JBP1 also contains a J-binding domain (27–29), while JBP2 contains a SWI/SNF2 helicase C-terminal domain presumably involved in chromatin binding/remodeling (30,31). Addition of the 2-oxoglutarate structural analog dimethyloxalylglycine (DMOG) to the growth medium inhibits hydroxylase activity of JBP1/2 and thus enables J reduction in cells without genetic modification (20). Reduction of J in Leishmania and T. brucei using genetic KO of JBPs or inhibition of hydroxylase activity via DMOG led to readthrough transcription at termination sites, suggesting a critical role for J in transcription termination (19,32–35). For several PTUs in T. brucei and L. major, we found J promoting Pol II termination prior to the end of the gene cluster, leading to silencing of the downstream genes (32,36). Loss of J from these premature termination sites results in readthrough transcription and derepression of the downstream genes. H3.V co-localizes with J at Pol II termination sites in T. brucei (15) and L. major (36) and has been shown to play a similar role in regulating transcription termination (33,35). J and H3.V independently function to promote termination, such that the combined loss of J and H3.V results in a synergistic increase in read through transcription, including at PTU internal termination sites (33,36). H4.V, which co-localizes with H3.V and J at termination sites in T. brucei, has recently been shown to play a similar synergistic role in transcription termination (34). Genes regulated by this ‘premature termination’ process include many that are developmentally regulated and, in the case for T. brucei, code for proteins involved in optimal growth and immune evasion during infection of the mammalian host (35,36).

During our studies of base J synthesis and function, we identified a PJW/PP1 complex in kinetoplastids composed of a J-binding protein (JBP3), protein phosphatase 1 (PP1), PP1 interactive-regulatory protein (PNUTS) and Wdr82 (37). Ablation of JBP3, Wdr82 or PNUTS in T. brucei causes read-through transcription at termination sites, indicating the role of the complex in kinetoplastid transcription termination (37). JBP3 has also recently been shown to play a role in transcription termination in L. tarentolae (38). A similar PTW/PP1 complex (containing PP1, PNUTS, Wdr82 and the DNA binding protein Tox4) has been shown to play a role in regulating transcription in humans and yeast (39–46). Critical to this process is the regulation of protein phosphorylation by the major eukaryotic protein serine/threonine protein phosphatase, PP1. PP1-dependent dephosphorylation of Spt5 and Pol II leading to decreased Pol II elongation, enhances the capture by the torpedo exonuclease allowing Pol II dissociation and termination (39,41,46,47). PP1 function is modulated by its association with the PP1 regulatory factor PNUTS (PP1 nuclear targeting subunit) via the canonical PP1 RVxF interaction motif. Based on genetic and biochemical analysis of the structural components (PNUTS, JBP3 and Wdr82), we proposed the kinetoplastid PJW/PP1 complex regulates transcription termination by recruitment to termination sites via JBP3-base J interactions and dephosphorylation of specific proteins by PP1, similar to the control of termination in higher eukaryotes by PTW/PP1 (see Supplementary Figure S1A). Central to this model is the sole catalytic subunit PP1. PP1, however, does not seem to be consistently included in the PJW complex across kinetoplastids (37). Purification of PNUTS from T. brucei extracts identified the associated proteins JBP3 and Wdr82, but not PP1. Furthermore, an obvious ortholog for the PP1 identified in the Leishmania PJW/PP1 complex is not present in the T. brucei genome. There are eight isoforms of PP1 in the Leishmania and Trypanosome genome distributed among five different clades (Supplementary Figure S1B). The eight PP1 paralogs in T. brucei have been numbered 1–8 (48). In accordance with the numbering nomenclature introduced by Li and coworkers and labeling the clades A-E (38) in the phylogenetic analysis, we now refer to the PP1 component of the Leishmania PJW/PP1 complex as PP1-8e (Supplementary Figures S1B and S2). Interestingly, PP1-8e belongs to a clade that lacks an ortholog in T. brucei, although present in other kinetoplastids including T. cruzi and more distantly related kinetoplastids such as Bodo saltans (37,38). TbPNUTS contains a ‘RVxF’ docking motif and presumably associates with at least one of the PP1 isotypes in vivo in a manner that is unstable during our purification methods and transcription termination proceeds by a similar mechanism among the kinetoplastids (37). However, due to the lack of an obvious PP1 component of the T. brucei complex, we sought to investigate whether the PJW/PP1 complex regulates transcription termination in a PP1-dependent manner by exploring the role of PP1-8e in Leishmania.

As stated above, PP1 mediated de-phosphorylation of both Pol II and Spt5 regulates the transition from transcription elongation to termination and Pol II release in humans and yeast (39,41,46,47). Interestingly, the only established direct substrate for PNUTS-PP1 is the C-terminal domain (CTD) of Pol II (39,49). During the transcription cycle, the largest subunit of Pol II, RPB1, becomes post-translationally modified in its CTD, which is an unstructured domain consisting of 52 repeats in humans (50). Coordinated reversible phosphorylation of the CTD regulates its association with factors involved in initiation, elongation and termination as well as co-transcriptional RNA processing (51–53). While the CTD of RPB1 in kinetoplastids is unique in that it does not contain the heptad or other repetitive motifs, 17 phospho-sites have been identified in this C-terminal serine-rich region in T. brucei (54,55). While this non-canonical CTD was shown to be essential for Pol II function (56,57), the functional significance of its phosphorylation remains unclear. Furthermore, a Pol II-CTD specific phosphatase has not been identified in kinetoplastids. While a CTD phosphatase RPAP2/Rtr1 homolog has been shown to associate with the T. brucei Pol II complex in vivo (58), phosphatase activity has not been studied.

Although PP1 phosphatases have been known to perform many essential functions in the life cycle of trypanosomatid parasites, including kinetoplastid segregation, cytoskeletal integrity, cytokinesis and nuclear positioning, little information is available regarding their role in transcription. No information is available regarding the function of the PP1-8e isotype in any kinetoplastid. In this work we sought to resolve the question regarding the essential role of PP1-8e in Leishmania and to test for its role in Pol II transcription termination by using a gene knockout (KO) strategy. Our data show that the PP1 isotype present in the PJW/PP1 complex has in vitro phosphatase activity and plays a central role in Pol II termination, where its catalytic activity is required for proper termination of Pol II transcription and repression of specific genes at the end of polycistronic units. Transcriptional and gene expression defects are similar to those seen upon the reduction in base J, directly linking DNA modification and PP1-8e protein phosphatase activity in the termination mechanism. Additionally, we show that Pol II is a direct substrate for PNUTS-PP1-8e in vitro. Together, these findings suggest a direct role of PNUTS/PP1 holoenzymes in regulating transcription termination via dephosphorylating Pol II in the Leishmania nucleus.

MATERIALS AND METHODS

Parasite culture

Promastigote Leishmania major strains were grown in M199 medium, supplemented with 10% FCS at 26°C. Transfections were performed with exponentially growing cells using the Amaxa electroporation system (Human T Cell Nucleofactor Kit, program U-033). After transfection, cells were split in two and allowed to recover for 24 hrs. before plating into 96 well plates to obtain clonal cell lines. Where appropriate, the following drug concentrations were used: 15 μg/ml Blasticidin, 50 μg/ml Hygromycin, 10 μg/ml Puromycin and 50 μg/ml Nourseothricin. DMOG treatment of cells was performed by supplementing media with 2.5- or 5 mM DMOG for 5 days. Control cells were treated with an equal amount of DMSO. Promastigote form L. tarentolae were cultured in SDM79 medium and transfections performed as previously described (37). Where appropriate, the following drug concentrations were used: 50 μg/ml G418 and 10 μg/ml Puromycin.

DNA constructs and cell line generation

Endogenous HA-tagging in L. tarentolae. A background L. tarentolae cell line was established in which Cas9 and T7 polymerase are expressed from the tubulin array. WT cells were transfected with plasmid pTB007 (59), digested with PacI, to generate the Cas9/T7-expressing cell line. To tag the endogenous PP1-8e, RBP1 and PABP1 locus with 6xHA tag, this cell line was then used in a single round of transfection to generate the PP1-HA, RBP1-HA and PABP1-HA cell lines with gRNAs and donor fragments, as previously described (59). gRNAs were designed with LeishGEdit and generated in vivo upon transfection with the appropriate DNA fragment generated by PCR. The donor fragments were PCR-amplified from pGL2314 plasmid with 30 nt homology flanks specific to the target loci, as previously described (59).

For generation of C-terminal multi (Streptavidin Binding Protein, Protein A and FLAG) tagged constructs in L. tarentolae, the coding region of LtPNUTS without a stop codon was amplified and cloned into the BamH1 and XbaI sites of pSNSAP1. The resulting construct is referred to as PNUTS-Pd. The PNUTS-Pd plasmid was transfected into the PP1-8e-HA cell line and WT L. tarentolae. PP1-PNUTS fusions were generated, with a flexible linker between PP1 and PNUTS, as described previously for PP1-NIPP1 (60). The DNA was synthesized by Genewiz to include N-terminal Protein A, Streptavidin-Binding Protein (PS) tag. The synthesized fusion protein DNA was sub-cloned into the pSNSAP1 plasmid by Gibson cloning to exclude the triple tag on the plasmid. The resulting construct is referred to as PP1-PNUTS fusion-PS. The sequences of all final constructs were confirmed by sequencing prior to transfection.

L. major PP1-8e KO. A background L. major cell line was established in which DiCre is expressed from the ribosomal locus and both Cas9 and T7 polymerase are expressed from the tubulin array. WT cells were transfected with plasmid pGL2399 (61), digested with PacI and PmeI, to generate Di-Cre-expressing cells. These cells were then transfected with plasmid pTB007 (59), digested with PacI, to generate the DiCre/Cas9/T7-expressing cell line. This cell line was then used to flank both copies of PP1-8e with LoxP sites, in a single round of transfection to generate the PP1-8e^Flox cell line. Donor fragment for Cas9-mediated replacement of endogenous PP1-8e was generated by PCR. WT PP1-8e was cloned into the NdeI and SpeI sites of the HA tagging/LoxP containing plasmid (pGL2341) (62). The resulting construct was used in a PCR reaction to generate the donor fragment flanked by 30 nucleotide sequence homology to the targeting integration sites, and transfected into the DiCre/Cas9/T7-expressing cell line. gRNAs were designed with LeishGEdit and generated in vivo upon transfection with the appropriate DNA fragment generated by PCR. PP1-8e KO cell lines were generated by adding 300 nM Rapamycin to the culture medium, reconstituting active Cre recombinase. After 14 days Rapamycin growth, clonal cell lines were obtained by limiting dilution in 96 well plates with the addition of 4.10⁵ WT cells/ml. PP1-8e rescue constructs (FLAG-tagged and untagged) were generated by subcloning a L. major PP1 ORF-Sat^R fragment from the pXNG4 plasmid into the SpeI and BsiW I sites from the pTB007 plasmid, digested with PacI, resulting in constitutive expression from the tubulin locus. Untagged PP1-1 and PP1-7 over-expression constructs were made the same way. All final constructs were sequenced prior to electroporation. Primers sequences used in the analysis are indicated in Supplementary Table S3.

Co-immunoprecipitation and peptide competition assay

5 × 10⁸ of L. tarentolae cells were lysed and PNUTS-Pd was affinity purified using 50 μl IgG Sepharose beads as previously described (37). After incubation with cell extract, the beads were washed 3 times in 10 ml PA-150 buffer and boiled for 5 min in 1x SDS-PAGE sample buffer for western blot analysis with anti-protein A and anti-HA antibodies. For peptide competition assay, beads were resuspended in 0.3 ml PA-150 buffer and incubated with the indicated concentrations of WT (KPAEAPSRKRVCWADEGHTDVSRGL) or mutant (KPAEAPSRKRACAADEGHTDVSRGL) RVXF peptide while rotating at room temperature for 30 min. The IPs were then washed 3 times in 1 ml PA-150 buffer and proteins eluted by boiling for 5 min in 1× SDS-PAGE sample buffer and analyzed by western blot with anti-protein A and anti-HA antibodies.

pNPP phosphatase assay

Phosphatase activity was assayed by using the generic phosphatase substrate p-Nitrophenyl Phosphate (pNPP). PNUTS-Pd and PP1-PNUTS-PS fusion were affinity purified from 8 × 10⁹ of L. tarentolae cells using the IgG Sepharose beads as above. The beads were washed 3 times in 10 ml PA-150 buffer. Following the final wash, the beads were resuspended in 10 ml PA-150 buffer. 0.1 ml bead slurry was taken, and beads collected by centrifugation and boiled in 1× sample buffer for western blot analysis with anti-protein A and anti-EF1α antibodies. The remaining 9.9 ml bead slurry was centrifuged to collect the beads, which were then resuspended in 0.2 ml assay buffer (50 mM Tris–HCl, pH 7.7, 1 mM DTT, 0.3 μM MnCl₂). The phosphatase reactions were started by addition of 50 mM pNPP and were monitored by continuously following production of p-nitrophenol (absorbance at 405 nm).

Rpb1 phosphatase assay

Phosphorylation status of RBP1-HA was determined by western blot, following phosphatase treatment of cell lysates as previously described (63). Briefly, 4.5 × 10⁸ cells were resuspended in 80 ul 0.1 M Tris–HCl, pH 6.8. Cells were lysed by addition of 20 ul 10% SDS and boiling for 5 min at 95°C. Cells lysates were then diluted 40-fold with dilution buffer (100 mM NaCl, 3 mM MgCl₂, 20 mM Tris–HCl, pH 7.5) to reduce the SDS concentration. Diluted cell lysates were then concentrated with Amicon spin column to 100 ul and treated with or without NEB Quick CIP and 1x PhosStop phosphatase inhibitor (Roche) for 30 min at 37°C.

In vitro Rpb1 phosphatase assays were done by the addition of purified PNUTS-PP1 complex to purified Pol II complex from L. tarentolae. PNUTS was purified from L. tarentolae cells expressing PNUTS-Pd (PNUTS^WT and PNUTS^RACA) using anti-FLAG magnetic beads (Invitrogen). To avoid interference of the cas9-Flag tagged protein in the PNUTS purification, we generated WT cells expressing PNUTS-Pd without HA-tagged PP1 through cas9 mediated editing. 8 × 10⁹ cells were suspended in 20 ml lysis buffer (150 mM NaCl, 50 mM Tris–HCl pH 7.7, 1 mM EDTA, 0.5% NP-40, 10% glycerol) with protease inhibitors (8 μg/ml Aprotinin, 10 μg/ml Leupeptin, 1 μg/ml Pepstatin, 1 mM PMSF and 1x cOmplete Mini, EDTA free protease inhibitor cocktail; Roche). Cells were lysed by sonication for 5 times (15′ on/45′ off, 50% amplitude, large tip) on ice. The cell lysates were cleared by centrifugation at 21 000 × g at 4°C for 10 min and incubated with 200 μl anti-FLAG magnetic beads for 4 h at 4°C on rotor. Beads were then washed 3 times with 20 ml wash buffer (50 mM HEPES/KOH pH 7.5, 300 mM KCl, 2 mM MgCl₂, 0.5% NP-40) and PNUTS (and associated proteins) eluted off with 0.5 ml (400 ug/ml) 3× FLAG peptide twice at room temperature for 30 min while rotating. The two eluted fractions were pooled. The HA-tagged Rpb1 was purified from 3 × 10⁹L. tarentolae cells using 150 μl anti-HA magnetic beads as described for IgG Sepharose purification of PNUTS-Pd but with buffers supplemented with 1x PhosSTOP, 100 mM beta-glycerophosphate and 25 mM NaF to inhibit endogenous phosphatase activity. After incubation with cell extract, the beads were washed 3 times with PA-150 buffer. The bead-immobilized Rpb1 was incubated with 1 ml PNUTS eluents for indicated time periods at 30°C while rotating. After incubation, Rpb1 was sedimented and eluted by boiling for 5 min in 1× sample buffer for western blot analysis with anti-HA antibody.

Phosphorylation status of PABP1-HA was determined by western blot, following lysis of cells in 1.0% NP-40. 2.5 × 10⁸ cells were resuspended in 0.2 ml lysis buffer (150 mM NaCl, 50 mM Tris–HCl pH 7.7, 1 mM EDTA, 1.0% NP-40, 10% glycerol) with protease inhibitors (8 μg/ml Aprotinin, 10 μg/ml Leupeptin, 1 μg/ml Pepstatin, 1 mM PMSF and 1× cOmplete Mini, EDTA free protease inhibitor cocktail; Roche) and incubated at 30°C with or without 1× PhosStop phosphatase inhibitor. Samples were taken at 0, 5, 15 and 30 min after incubation and reactions stopped by adding SDS-PAGE sample buffer and boiling for 5 min. In vitro PABP1 phosphatase assays were done by the addition of purified PNUTS^WT-PP1 complex to purified PABP1 from L. tarentolae as described above for the Rpb1 assay. Incubation of PABP1 in reaction buffer without addition of the PNUTS-PP1 complex was done as a control.

Western blotting

Proteins from 7.5 × 10⁶ cell equivalents were separated by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE 10% gel), transferred to nitrocellulose and probed with anti-HA antibodies (Sigma, 3F10, 1:3000) and anti-Elongation Factor 1A (Sigma, 05-235, 1:20 000) was used as a loading control. Bound antibodies were detected by an Alexa Fluor 800 labelled secondary goat anti-rat antibody and an Alexa Fluor 680 labelled secondary goat anti-mouse (LiCor) and analyzed with Image Studio software (LiCor)

RT-PCR analysis

Total RNA was isolated with Tripure Isolation Reagent (Roche). cDNA was generated from 0.5 to 1 μg Turbo™ DNase (ThermoFisher) treated total RNA with Superscript™ III (ThermoFisher) according to the manufacturer's instructions with either oligo dT primers or strand specific oligonucleotides. Strand specific RT reactions were performed with the strand specific oligonucleotide and an antisense blasticidin oligonucleotide. Equal amounts of cDNA were used in PCR reactions with Ready Go Taq Polymerase (Promega). A minus-RT control was used to ensure no contaminating genomic DNA was amplified.

Quantitative RT-PCR analysis

Total RNA was isolated and Turbo™ DNase treated as described above. Quantification of Superscript™ III generated cDNA was performed using an iCycler with an iQ5 multicolor real-time PCR detection system (Bio-Rad). Triplicate cDNA’s were analyzed and normalized to enolase cDNA. qPCR oligonucleotide primers combos were designed using Integrated DNA Technologies software. cDNA reactions were diluted 20-fold and 5 μl was analyzed. A 15 μl reaction mixture contained 4.5 pmol sense and antisense primer, 7.5 μl 2× iQ SYBR green super mix (Bio-Rad Laboratories). Standard curves were prepared for each gene using 5-fold dilutions of a known quantity (100 ng/μl) of WT gDNA. The quantities were calculated using iQ5 optical detection system software. Primer sequences used in the analysis are indicated in Supplementary Table S3.

Phylogenetic analysis

Phylogenetic analysis was performed with Maximum Likelihood method and JTT matrix-based model, using MEGA11 software.

Determination of the genomic level of base J

To quantify the genomic J levels, DNA was isolated and utilized in the anti-J DNA immunoblot assay as described previously. Briefly, serially diluted genomic DNA was blotted onto a nitrocellulose membrane, followed by incubation with anti-J antisera. Bound antibodies were detected by an Alexa Fluor 800 labelled secondary goat anti-rabbit antibody. The membrane was stripped and incubated with Methylene Blue stain for DNA visualization.

Strand-specific RNA-seq library construction

For mRNA-seq, total RNA was isolated from wild-type and PP1-8e KO L. major cultures using TriPure. Six mRNA-seq libraries were constructed (triplicate samples for WT and PP1-8e KO) using Illumina TruSeq Stranded RNA LT Kit following the manufacturer's instructions with limited modifications. The starting quantity of total RNA was adjusted to 1.3 μg, and all volumes were reduced to a third of the described quantity. High throughput sequencing was performed on an Illumina Novaseq6000 instrument.

RNA-seq analysis

Raw reads from mRNA-seq were first trimmed using fastp with default settings (v0.23.2; (64)). Clean reads were locally aligned to the L. major Friedlin strain genome version 9.0 using bowtie2 tool (65) with ‘very high sensitivity’ parameter and further processed with samtools version 1.7 (66). To ensure proper read placement, alignments with multiple low-quality hits and mapping quality (MQ) scores less than 30 were removed. Strand-specific read coverage was calculated from BAM files obtained from Bowtie2 using customized pysam scripts (https://github.com/pysam-developers/pysam). To compare read coverage for different sites the mean read coverage was calculated and normalized by total number of read number of sequencing library for each category. To compare read coverage of dSSRs, we analyzed the SSRs and the 5-kb flanking regions with DeepTools (3.5.0) using 100 bp bins flanking the SSRs and dividing each SSR into 50 equally sized bins (67). For each sample, FeatureCounts (v2.0.1) (68) was used to count reads for each reference transcript annotation, followed by normalization/variance stabilization using DESeq2 (v1.26.0) (69). Differentially expressed genes (DEGs) were identified using the DESeq2 by comparing WT and PP1 KO samples in triplicate by setting log2 fold change >1 and FDR-adjusted P-value <0.001 (Supplementary Table S1). The average mapping rate for mRNA-seq replicates of WT and PP1 KO was 0.92 (Supplementary Table S2). DMOG RNA-seq data shown here are from previously published work (36).

RESULTS

PP1-8e has in vitro phosphatase activity and associates with PNUTS via the RVxF motif

Leishmania PP1-8e contains all the invariant structural motifs (-GDXHG-, -GDXVXRG- and -GNH-) described in the members of the PPP family as well as all the kinetoplastid serine/threonine protein phosphatases characterized to date (70). As such, it possesses all six conserved residues in the catalytic site to coordinate two manganese ions (Supplementary Figure S2). Since all attempts to purify soluble recombinant PP1-8e from E. coli failed, we utilized affinity purified enzyme expressed in L. tarentolae in order to examine in vitro protein phosphatase activity. Purification of PP1 from Leishmania extracts would also co-purify PNUTS as well as other PP1 regulatory proteins. To help ensure purified PP1-8e is associated specifically with PNUTS, we expressed PS-tagged PP1-PNUTS fusions joined via a flexible linker (Figure 1A and B). In addition to the WT fusion (PP1-PNUTS), we used a fusion with an inactive catalytic moiety (PP1HK-PNUTS). PP1HK has a mutated metal-coordinating residue in the active site (H92K). Similar H66K mutation with the human PP1 catalytic site has been shown to eliminate its activity (71). Purified PP1-PNUTS fusion showed phosphatase activity with pNPP as a substrate (Figure 1C). As expected, the PP1HK-PNUTS fusion was largely inactive. The results indicate PP1-8e-PNUTS complex has PP1-dependent in vitro phosphatase activity.

Figure 1.

The PJW/PP1 complex exhibits phosphatase activity in vitro. (A–C) Phosphatase assay for PP1-PNUTS fusions. (A) Schematic of Ps-tagged PP1-PNUTS fusions. The black box indicates the N-terminal Protein A, Streptavidin Binding Protein tag. (B) PP1-PNUTS fusions were expressed in WT L. tarentolae and purified with anti-protein A resin. Control, WT cells; WT, cells expressing WT PP1-PNUTS fusion; HK, expressing H92K PP1 mutant fusion protein. (C) PP1-PNUTS fusion immunoprecipitates were assayed for phosphatase activity with pNPP as a substrate. Anti-protein A immunoprecipitates from WT cell extracts were included as a negative control. Absorbances at 405 nm were measured. The changes in absorbance were plotted against time course. Equal level of input fusion protein was used in the phosphatase assay as shown by the western blot (B). The bar graph represents the relative changes in absorbance after 2-h incubation above the background control, with the value for WT fusion immunoprecipitates set as 1. (D, E) The PJW/PP1 complex exhibits in vitro phosphatase activity in a PP1-dependent manner. (D) Equal levels of pd-tagged WT PNUTS (PNUTS^WT) or PP1-8e binding deficient PNUTS^RACA proteins were purified from cell extracts by anti-protein A affinity, as shown by the western blot with anti-protein A. Anti-EF1a serves as a loading control. Right, the conserved RVXF motif and mutations in PNUTS^RACA. (E) The immunoprecipitates were assayed for phosphatase activity against pNPP as described above. Error bars indicate standard error of the mean (SEM) from three experiments. The blots are representative of three independent experiments. In, Input; IP, bound fraction; FT, unbound fraction.

To understand the function of the PP1-8e phosphatase in this complex, we initially sought to biochemically characterize the Leishmania PNUTS/PP1 complex. We verified the binary interaction between PNUTS and PP1-8e by co-immunoprecipitation experiments (Co-IP) in vivo. To do this we HA-tagged the endogenous locus of PP1-8e using Cas9 and expressed a Pd-tagged version of PNUTS from a plasmid in L. tarentolae. As expected, immunoprecipitation of PNUTS-Pd resulted in co-precipitation of PP1-8e-HA (Figure 2A). There is no precipitation of PP1-8e-HA in cells lacking PNUTS-Pd expression. To demonstrate the involvement of the canonical primary PP1-binding sequence, we created a cell line expressing PNUTS-Pd with V97A and W99A substitutions in the predicted RVXF motif (37). This PNUTS^RACA mutant has minimal interaction with PP1 by Co-IP (Figure 2A). Furthermore, a short peptide from PNUTS that contains the RVXF motif is able to disrupt the PP1-PNUTS association, while the identical peptide with V97A and W99A substitutions is not (Figure 2B). We then purified the PNUTS complex from L. tarentolae cells expressing PNUTS-Pd (WT and RACA mutant) (Figure 1D) to ascertain the effect that loss of PP1-8e association had on in vitro phosphatase activity. The PNUTS-Pd^WT immunoprecipitate has phosphatase activity, while the purified PNUTS-Pd^RACA immunoprecipitate lacking PP1-8e does not (Figure 1D and E). Altogether, the data shows that Leishmania PP1-8e is a functional protein phosphatase and associates with the complex via the PNUTS RVXF PP1 binding motif.

Figure 2.

PP1-8e binds PNUTS via the RVXF motif. Analysis of the interactions between PNUTS and PP1-8e in vivo. (A) Co-immunoprecipitation analysis showing that V97 and W99 of the LtPNUTS ‘RVXF’ interaction motif is essential for interaction between PP1-8e and PNUTS. PP1-8e was endogenously tagged with HA tag, and Pd-tagged wild type or mutant PNUTS (with alanine substitutions for V and W in the RVXF motif) was over-expressed from the pSNSAP1 vector. Cell extracts from the indicated cell lines were purified by anti-protein A affinity resin and analyzed by western blot with anti-protein A and anti-HA. In; input (equivalent to the amount of protein added to the IP reaction mixture), IP; precipitated immunocomplexes, FT; flow through or non-bound supernatant. EF1α provides a loading control and negative control for the IP. (B) A peptide based on the PP1 interaction motif RVXF displaces PP1 from PNUTS-Pb bound affinity matrix. Pd-tagged PNUTS was purified from cell extract with anti-protein A affinity resin, and IPs were incubated with indicated concentrations of synthetic 25-mer peptides with either WT (RVCW) or mutated (RACA) RVXF motif for 30 min. Dissociation of PP1 from the PNUTS IP was examined by western blotting of sedimented fractions. The blots shown are representative of three independent experiments. PP1 protein levels were quantified by densitometric analysis and normalized to PNUTS protein level in the IPs. The bar graph represents the mean ± SD for three independent experiments.

PJW/PP1 dephosphorylates pol II in vitro

Dephosphorylation of the largest subunit of the Pol II complex, RPB1, has been shown to underlie the mechanism by which the PTW/PP1 complex regulates transcription termination in human cells. To investigate if the PJW/PP1 complex similarly regulates transcription termination by dephosphorylation of RPB1, we characterized the phosphatase activity of the PJW/PP1 complex from L. tarentolea expressing PNUTS-FLAG in an in vitro assay with Pol II that has been separately purified via HA-RBP1 expression. The phosphorylation level of Pol II RBP1 was first visualized by western blot analysis using anti-HA. In SDS-PAGE, trypanosome RBP1 migrates as a doublet where the upper band is phosphorylated RBP1 and the lower band is the unphosphorylated RBP1 (57,63,72). We found the Leishmania Pol II RBP1 appears as a doublet in SDS-PAGE gels, suggesting it is also phosphorylated (Figure 3A). To confirm the presence of these two forms was due to differences in the level of protein phosphorylation, extracts were treated with alkaline phosphatase (CIP). Phosphatase treatment led to the disappearance of the upper band and accumulation of the lower band (Figure 3A). This shift was completely blocked with phosphatase inhibitor. Therefore, we conclude that the upper and lower bands in our anti-HA western blots represented phosphorylated RBP1 (pRPB1) and dephosphorylated RBP1, respectively.

Figure 3.

Pol II is a substrate for PP1-8e. PJW/PP1 subunit PP1-8e dephosphorylates Pol II in vitro. (A) LtRNA Pol II is phosphorylated. Western blot of parasite lysate expressing RPB1 endogenously tagged with HA-tag treated without (-) or with (+) calf intestinal phosphatase (CIP) in the absence or presence (+Inhibitor) of a phosphatase inhibitor cocktail (PhosSTOP). Blots were probed with anti-HA. Phosphorylated (pRPB1) and dephosphorylated (RPB1) forms of Rpb1 are indicated. Ratio of pRPB1/RPB1 from densitometric quantification of the blot is indicated below; with the ratio in initial extracts arbitrarily set to 1. (B) Purification of the PJW/PP1 complex. Pd-tagged WT PNUTS or RVXF mutant PNUTS (RACA) were purified with anti-FLAG affinity resin and eluted via 3× FLAG peptide and processed for Western blotting using anti-HA and anti-EF1α antibodies. In, Input; Elu, FLAG peptide elutant; FT, flow through unbound fraction. (C) Pol II in vitro phosphatase assay. Equal inputs of immunoprecipitates from WT and mutant PNUTS, as shown in (B), were incubated with HA-immobilized Rpb1 for the indicated times. The last lane represents in vitro incubation of WT PNUTS-PP1-8e complex with Rpb1 for 4 hours with the addition of phosphatase inhibitor. HA-Immobilized Rpb1 proteins were processed for Western blotting using anti-HA. The blots shown are representative of at least three independent experiments. Densitometric quantification of pRPB1 and RPB1 forms from three independent in vitro phosphatase assays is shown below. The pRPB1/RPB1 ratio at T0 was arbitrarily set to 1. The bar graph represents the mean ± SD for three independent experiments. (D) LtPABP1 is phosphorylated. Western blot of parasite lysate expressing PABP1 endogenously tagged with HA-tag lysed in buffer II with 1% NP-40 and incubated at 30°C with or without phosphatase inhibitor for the indicated times. Ratio of pPABP1/PABP1 from densitometric quantification of the blot is indicated below; with the ratio in initial extracts (T0) arbitrarily set to 1. (E) PABP1 in vitro phosphatase assay. Experiments were performed and analyzed as in (C). HA-Immobilized PABP1 was incubated with and without immunoprecipitates from WT PNUTS for the indicated times.

Purified HA-RBP1 was then tested as a substrate for purified PNUTS-PP1-8e complex. Analysis of the complex was necessary because isolated phosphatases generally have little specificity, and target-specific dephosphorylation often relies on cofactors/regulatory proteins. As described above, PNUTS purification allows the isolation of the PJW/PP1 complex from L. tarentolae cells (37,38). Isolation of the PNUTS-PP1-8e component of the complex via PNUTS-Pd purification is confirmed in Figure 2A. We therefore utilized the PNUTS-Pd immunoprecipitate, via affinity purification of FLAG-PNUTS (Figure 3B), in an in vitro phosphatase assay with purified HA-RBP1. To test if PP1 is necessary for the phosphatase activity, PNUTS with a mutant RVxF motif (RACA) that lacks PP1-8e association (Figure 2A) was similarly purified (Figure 3B), and tested for Rpb1 in vitro phosphatase activity. Anti-HA immunoblotting clearly showed that the pRBP1 signal became diminished upon incubation with WT PNUTS-PP1-8e immunoprecipitate, whereas dephosphorylated RBP1 increased (Figure 3C). The dephosphorylation of Pol II by PNUTS-PP1-8e immunoprecipitate was completely blocked by the addition of phosphatase inhibitor and little to no dephosphorylation was seen with addition of PNUTS (RACA) mutant immunoprecipitate, lacking associated PP1-8e. These observations provide evidence that dephosphorylation of Pol II by PP1-8e was not mediated by a contaminating phosphatase and that the shift to the high mobility band was not an artifact resulting from proteolysis of Pol II.

To test if the in vitro phosphatase activity is specific, we characterized PP1-PNUTS activity on another phosphorylated protein, Poly-A binding protein 1 (LtPABP1). PABP1 orthologues from both T. brucei and Leishmania species have been shown to be targeted by serine/threonine phosphorylation events within its linker domain leading to multiple isoforms visible on the SDS-PAGE gel (73–76). LtPABP1 was represented by at least two distinctly migrating bands with apparent molecular masses of approximately 69 and 75 kDA, with the larger band representing the phosphorylated form of the protein, as previously described in Leishmania (73–75) (Figure 3D). To confirm the migration pattern of PABP1 on SDS gel was indeed due to hyperphosphorylation, detergent lysis of parasites allows native phosphatases to convert the phosphorylated state to the dephosphorylated state, an activity that is inhibited by phosphatase inhibitors (Figure 3D). In contrast to the high dephosphorylation activity toward Pol II (Figure 3C), no dephosphorylation of purified HA-PABP1 was seen following incubation with WT PNUTS immunoprecipitate (Figure 3E). These results show the PP1-8e subunit of PJW/PP1 exhibits specific phosphatase activity toward Rpb1 in vitro and suggest that PP1-8e could be involved in the regulation of Pol II-CTD phosphorylation in the Leishmania nucleus.

DiCre approach for assessment of PP1 function in L. major

To investigate the phenotypes resulting from loss of PP1-8e in L. major promastigotes, an inducible knockout strain was generated using the DiCre system (Supplementary Figure S3). An L. major strain expressing dimerizable split Cre recombinase was modified to carry 6xHA epitope-tagged alleles of PP1 flanked by LoxP sites, referred to as PP1::6xHA^−/-flox. Growth curves showed that the addition of loxP sites and the HA tag did not lead to any significant growth impairment (see below). KO induction of PP1-8e was attempted by rapamycin-mediated DiCre activation in logarithmically growing cultures (Figure 4). PCR analysis of these populations at 72 h after rapamycin addition revealed that the PP1::6xHA^flox alleles had been excised (Figure 4A and B). Controls without addition of rapamycin showed no gene excision (Figure 4B). The levels of PP1::6xHA protein were assessed by western blot, revealing a >90% reduction in the rapamycin-treated sample compared to the control sample (Figure 4C). After 15 days of growth in rapamycin a clone was obtained, referred to PP1 KO cB5. The PP1 KO cB5 grew normally as per the wild-type and parental DiCre strain (Figure 4D and data not shown). This phenotype was reproducible and observed in an independent clonal cell line (Supplementary Figure S4A–C). The loss of PP1-8e gene products in both KO clones was confirmed by RT-PCR (Supplementary Figure S4D and E). We conclude that PP1-8e is not essential for Leishmania promastigote viability.

Figure 4.

PP1-8e is non-essential in L. major. (A) Illustration of PP1-HA^Flox excision catalyzed by DiCre, as induced by rapamycin. Primers used in PCR analysis of genomic DNA, are shown in red. Please see material and methods (and Supplementary Figure S3) for details on generation of the PP1-HA^Flox cell line. (B) PCR analysis of genomic DNA extracted from the PP1-HA^Flox without the addition (0) and 3–10 days after addition (+RAP) of rapamycin, leading to DiCre induction. Approximated annealing positions for the primers are shown in (A). DNA from wildtype cells (WT) is included as a control. KO clone obtained after 15 days growth in rapamycin. (Intact) and (Excised), PP1^Flox and PP1^Flox after excision, respectively. (C) Western blotting analysis of whole cell extracts of the PP1-HA^Flox without the addition (0) and 3–6 days after addition (+RAP) of rapamycin, leading to DiCre induction; extracts were probed with anti-HA antiserum and anti-EF1α was used as loading control (protein sizes are indicated, kDa). (D) Growth curves of wild-type (WT, black), PP1-HA^Flox (red) and the PP1-8e KO clone 5 (blue); cells were seeded at ∼10⁵ cells/ml at day 0 and diluted back to that density every 3 days; cell density was assessed every 24 h, and error bars depict standard deviation from three replicate experiments.

Loss of PP1-8e impairs pol II transcription in L. major

To assess the role of PP1-8e in transcription termination and mRNA gene expression, RNAs were isolated from the PP1 KO and WT cell lines and used to generate strand-specific RNA-seq libraries. For each condition (WT and PP1-8e KO) three independent mRNA-seq libraries were sequenced. Illumina sequencing reads were mapped to the L. major reference genome, and normalized read counts were calculated for every gene. First, we analyzed the read coverage for 5 kb on either side of all 146 transcription termination sites (TTSs) in the L. major genome. This includes convergent strand-switch regions (cSSRs) where the 3’ termini of two PTUs converge and TTSs between head-to-tail (unidirectionally) oriented PTUs (Figure 5A–C). As expected, in WT cells the mean-normalized coverage on the top (coding) strand decreased sharply at the TTS of cSSRs (Figure 5A). However, when PP1-8e is deleted, the read coverage downstream of the TTS (top strand) was significantly higher (P value < 0.001), suggesting that loss of PP1-8e resulted in significant transcriptional readthrough. At these convergent sites, analysis of bottom (noncoding) strand also reveals significant differences between WT and the PP1-8e KO, where readthrough from the adjacent convergent PTU results in increased antisense transcripts past the TTS into the adjacent PTU (Supplementary Figure S5A). For example, at a cSSR on chromosome 22 (cSSR 22.3) we see that the loss of PP1-8e leads to readthrough transcription on the top and bottom strands into the adjacent PTU (Figure 5B). RNA-seq read mapping data for all three replicates are shown in Supplementary Figure S6. Quantitation of readthrough transcription at cSSRs genome-wide, by measuring the change in the ratio of antisense:sense reads in a 10kB window, indicates a 5.68-fold increase upon deletion of PP1-8e (Supplementary Figure S7). Readthrough transcription was also seen at TTS between head-to-tail (H-T) oriented PTUs (Figure 5C). While the gap between adjacent PTUs is small, termination defects at H-T sites are indicated by the increase in the top (coding) strand transcriptome reads from the 3’ end of the last gene in the array into the downstream gene array. In many cases, the actual TTS is located prior to the final annotated gene in the PTU (32,36) (as discussed below). This would explain the increased strand coverage prior to the end of the upstream PTU at head-to-tail TTSs (Figure 5C). However, this increase in readthrough transcription did not occur to the same extent at different types of TTSs (Supplementary Figure S5). One variable is the presence of noncoding RNA genes transcribed by Pol III (i.e. tRNA and 5S RNAs) (Supplementary Figure S5B and C and Figure 5D). Readthrough was modest at cSSRs containing Pol III transcribed RNA genes compared with cSSRs that lack these RNA genes. Furthermore, there is little to no readthrough at centromeric cSSRs (Supplementary Figure S5D) and telomeric localized TTSs (Supplementary Figure S5I). Interestingly, while the presence of RNA gene at H-T sites had a similar (although smaller) effect on readthrough, centromeric localization had little to no effect (Supplementary Figure S5F–H and Figure 5D). A similar analysis of transcript abundance surrounding transcription start sites (TSSs) revealed no significant changes in sense RNA downstream of TSSs at dSSRs due to the loss of PP1-8e (Supplementary Figure S8), except for the small increase in top (coding) strand coverage at head-to-tail oriented PTUs that is presumably due to readthrough from the upstream PTU (Figure 5B). Furthermore, there was little to no increase in RNA upstream of TSSs at dSSRs that would correspond to bi-directional activity of Pol II promoters (Supplementary Figure S8) (37).

Figure 5.

Depletion of PP1 results in readthrough at transcription termination sites. (A) Mean top strand coverage at each nucleotide position (bp) in the 10 kbp surrounding the transcription termination site (TTS) at all 39 cSSRs for WT cells (blue line) or PP1 KO (orange line). The schematic represents the protein-coding genes associated with each strand at an ‘average’ convergent TTS (cTTS). Plots are orientated that transcription proceeds from the left and terminates at ‘0’, with the top strand being the coding strand on the left side of the TTS. Dashed arrows represent transcription direction. (B) A region on chromosome 22 from 503–513kb where J (36) and PP1 regulate transcription termination at a cSSR (cSSR 22.3) is shown. Top; map of the cSSR. ORFs are shown with the top strand in blue and the bottom strand in red. The red arrow indicates read through transcription following the loss of J (36) and loss of PP1. Bottom; Strand-specific mRNA-seq reads from the indicated cell lines are mapped. Reads that mapped to the top strand are shown in blue and reads that mapped to the bottom strand in red. (C) Mean top strand coverage at each nucleotide position in the 10 kb surrounding the 52 TTSs between head-to-tail (unidirectional) PTUs. The schematic represents the protein-coding genes associated with each strand at an ‘average’ head-to-tail (HT) TTS. HT regions that are transcribed in the opposite direction of the diagram are reoriented so that the transcribed genes are represented on the top strand. The flag indicates the transcription start site for the downstream gene cluster as indicated by H3 acetylation localization (1). The region spans 5 kb flanking the transcription start site downstream of the TTS. (D) Box-and-whisker plots showing the median top strand coverage in the 5-kb region downstream of all 146 TTSs. Separate plots are shown for all the TTSs at cSSRs (Conv All), 24 TTSs at cSSRs that lack an RNA gene (Conv –), 8 TTSs at cSSRs that contain one or more RNA genes (Conv +), 7 TTSs at cSSRs adjacent to a centromere (Cent), TTSs between all head-to-tail PTUs (H-T All), 27 TTSs between head-to-tail PTUs that lack an RNA gene (H-T –),12 TTSs between head-to-tail PTUs that contain one or more RNA genes (H-T +), 13 TTSs between head-to-tail PTUs adjacent to a centromere (H-T Cen), and 55 TTSs at telomeres (Tel). Multiple comparisons were conducted by wilcoxon test. P values were presented on top of each compared group.

To further characterize these transcription termination defects, normalized read counts mapped to the genome were calculated for every gene. Differential expression analysis (DESeq2 module) revealed that 34 genes had significantly higher mRNA abundances (>2-fold; P < 0.001) in the PP1-8e KO compared to the WT cell line (Supplementary Table S1). Interestingly, 24 of the 34 upregulated genes are located adjacent to transcription termination sites (TTSs). Therefore, the majority of the upregulated genes in the L. major PP1-8e KO are located at the end of a gene cluster immediately downstream or within a J peak, where J-mediated transcription termination attenuates transcription of downstream genes, similar to findings we previously described in L. major and T. brucei (32,36). This epigenetic regulated termination of Pol II transcription prior to the last ORF of the PTU we have referred to as premature termination. For example, on chromosome 9 only one gene is upregulated in the PP1 KO, and represents the last gene of a gene cluster at cSSR 9.1 (Figure 6A). Identical specific upregulation of this gene was seen following the loss of J at the TTS following DMOG treatment of WT cell (36). Interestingly, of the 21 genes that were upregulated upon the loss of base J (WT + DMOG) (36), 18 are also up in the PP1-8e KO (Supplementary Table S1). At the premature termination site at cSSR 9.1, loss of PP1-8e leads to readthrough transcription on the top strand that results in the complete transcription of the final annotated gene, which is processed to mature mRNA (Figure 6B). No expression change is detected for the genes immediately upstream or the final few genes of the adjacent convergent gene cluster in the PP1-8e KO. Gene expression changes were confirmed by RT-qPCR (Figure 6C and D). As previously demonstrated upon the loss of base J, readthrough transcription then extends down to a Pol III-transcribed gene in the cSSR (a 5S rRNA gene on the bottom strand and tRNA genes on the top strand), which are known to terminate Pol II transcription independent of J in Leishmania (19,32), as discussed above. Thus, termination defects at this specific TTS in the PP1 KO are limited to changes in the top strand within the SSR and little to no change in antisense transcription into the adjacent PTU on either strand. Therefore, the presence of RNAP III-transcribed genes at this site seems to terminate transcription on both strands independent of PP1 (Figure 6B). This is in contrast to TTSs at cSSRs that lack Pol III genes (for example cSSR 22.3) where readthrough in the PP1 KO leads to antisense transcription into the adjacent PTU on both strands (Figure 5B). Additional example of PP1 regulation of gene expression via termination includes H-T region 26.5 (Figure 7) where J is found upstream of the last gene within the gene cluster. The expression of the downstream (and final) gene is increased upon the loss of J following DMOG treatment (36) or deletion of PP1-8e (Figure 7). Adjacent genes, upstream of the TTS or within the neighboring gene cluster, are not affected.

Figure 6.

Decreased efficiency of RNAP II termination and increased gene expression following the loss of PP1. (A) Gene map of chromosome 9 is shown. mRNA coding genes on the top strand are indicated by black lines in the top half of the panel, bottom strand by a line in the bottom half. Genes on the top strand are transcribed from left to right and those on the bottom strand are transcribed from right to left, indicated by blue arrows. Panel below (WT + DMOG and PP1) indicates the location of the single mRNA (LmjF09.0690) found upregulated by at least two-fold in WT cells treated with DMOG relative to WT (36) and PP1 KO relative to WT. No other expression changes (up or downregulated) were detected. (B) A region on chromosome 9 from 263–285 kb where J (36) and PP1 regulates transcription termination and gene expression at a cSSR is shown (cSSR 9.1). Top; map of the cSSR. The vertical arrow indicates the proposed TTS. ORFs are shown with the top strand in blue and the bottom strand in red. The red arrow indicates read through transcription following the loss of J (36) and loss of PP1. Green boxes indicate RNAP III transcribed genes (tRNA and 5S rRNA). The numbered genes (1) and (2) that flank the TTS refer to LmjF09.0680 and LmjF09.0690, respectively. Bottom; Strand-specific mRNA-seq reads from the indicated cell lines are mapped. Reads that mapped to the top strand are shown in blue and reads that mapped to the bottom strand in red. (C and D) Gene expression changes for the genes flanking the TTS and indicated (numbered) in the ORF map in B were confirmed by qRT-PCR in the indicated cell lines. mRNA levels in WT set to one. Error bars indicate the standard deviation between two biological replicates analyzed in triplicate. WT; wild-type, PP1 HA; PP1-HA^Flox, PP1 KO; PP1-HA^Flox excised, +PP1 Flag; PP1-KO + PP1-Flag tagged, +PP1,; PP1-KO + PP1 untagged. P values were calculated using Tukey's multiple comparisons test, ****P values < 0.0001; NS, not significant.

Figure 7.

PP1 regulates RNAP II termination and gene expression at head-tail regions within gene clusters. (A) Gene map of chromosome 26 is shown where loss of J leads to upregulation of a single mRNA at the end of a gene cluster at a head-tail region. Labeling is as described in Figure 6. (B) Top; ORFs are plotted for the head-tail region on chromosome 26 from 912–922 kb, as described in Figure 3. Base J localizes at the transcription termination site (TTS). The vertical arrow indicates the proposed TTS (36). The black dashed arrow above the map indicates the direction of transcription and the dashed red arrow indicates read through transcription past the TTS. The upregulated gene, downstream of the TTS, is LmjF26.2280 (2280). The flag indicates the transcription start site for the downstream gene cluster as indicated by H3 acetylation localization (1). Bottom: Plot of the mRNA-seq data for the region above, as described in Figure 6. (C and D) Gene expression changes for the genes flanking the TTS and indicated in the ORF map in B were confirmed by RT-qPCR in the indicated cell lines, as described in Figure 6 (C and D). Error bars indicate the standard deviation between two biological replicates analyzed in triplicate. P values were calculated using Tukey's multiple comparisons test, ****P values < 0.0001; NS, not significant.

Additional evidence for readthrough transcription following PP1 loss is provided by strand-specific RT-PCR analysis (Figure 8A). Following the loss of PP1-8e we detect an increase in a nascent transcript that extends through the TTS, and further downstream, at a cSSR (Figure 8B). As we previously shown (32,36), the level of this RNA species increases along with decreasing levels of J in WT parasites treated with DMOG (Figure 8C and D). This increase in readthrough of nascent RNA leads to increased expression of downstream genes (Supplementary Figure S9). Interestingly, reduction of base J levels in the PP1-8e KO upon treatment with DMOG results in further increase in readthrough transcription and appearance of a significant growth defect (Figure 8C–E and Supplementary Figure S9).

Figure 8.

PP1 regulates nascent readthrough RNA synthesis. (A) Schematic representation (not to scale) of cSSR 22.3 illustrating the nascent RNA species assayed by RT-PCR. The dashed red arrow indicates read through transcription past the transcription termination site (TTS). Arrows indicate the location of primers utilized for strand-specific RT-PCR analysis. (B) Strand-specific RT-PCR analysis of readthrough transcription. Read through transcription on the top strand for the indicated cell lines was quantitated by performing site-specific cDNA synthesis using primer RT illustrated in the diagram above, followed by PCR using primers A and B. Abundance was normalized using blasticidin marker (a gene specific primer against blasticidin was added to the same cDNA synthesis reaction with primer RT, followed by PCR using blasticidin primers). Fold increase in nascent readthrough RNA species relative to background levels in WT is based on A + B qPCR analysis, normalized to blasticidin qPCR. (C) Strand-specific RT-PCR analysis of readthrough transcription as in (B) for the indicated cell lines grown in the presence of DMOG. (D) Serially diluted genomic DNA from wild-type and PP1-8e KO cells, grown with 0, 2.5mM and 5mM DMOG for 5 days, were incubated with anti J antisera. DNA loading was verified by Methylene Blue staining. (E) Growth curves of wild-type and PP1-8e KO cells in the presence of 0, 2.5 mM and 5 mM DMOG. Cumulative cell numbers were calculated after passaging 10⁵ cells/ml after 3 days of initial growth in medium with fresh DMOG.

To ensure the phenotype was specific to PP1 deletion and not due to off-target effects, an allele of PP1::Flag was reintroduced to the PP1 KO. Unfortunately, the use of Flag tag was not ideal with expression of Cas9-Flag in the same cell, and background not allowing detection of PP1-Flag expression by western blot (data not shown). Thus, expression of PP1-Flag was confirmed at the RNA level (Supplementary Figure S4D), with expression ∼3-fold higher than WT endogenous levels. Termination of the PP1-Flag-complementation strain, measured by upregulation of the gene downstream of TTS at cSSR 9.1, was restored to levels seen in the PP1::6xHA (Figure 6D). While not 100% complementation (compared to WT), it potentially reflects the negative effect of C-terminal tag on PP1 function, demonstrating the requirement for PP1 for transcription termination. It appears that the replacement of WT PP1-8e locus with PP1-8e-6xHA tagged version already led to defects in termination at the level of nascent RNA and gene expression changes (Figures 6D and 8B). Rapamycin-mediated PP1-HA allele excision had additional effect on termination defects measured by gene expression and nascent RNA. Therefore, the presence of either C-terminal Flag or HA tag may significantly affect PP1-8e function. To explore the effect of C-terminal tags, we expressed an untagged version of PP1 in these cells. Expression of the untagged version of PP1-8e, with similar 3-fold higher levels of mRNA than WT (Supplementary Figure S4D), resulted in an almost complete rescue in readthrough transcription measured by both gene expression and nascent RNA changes (Figures 6D and 8B). Taken together, these findings suggest PP1 phosphatase functions as part of the PP1/PJW complex in the promotion of Pol II termination in Leishmania.

To explore possible redundancy of PP1 isotype function, we over-expressed an untagged version of PP1-1a and PP1-7d isotypes in the PP1-8e KO. Among the 8 PP1 isotypes in T. brucei, PP1-1 and PP1-7 been shown to primarily localize within the nucleus (Tryptag.org; (77)). Interestingly, over-expression of LmPP1-1a or LmPP1-7d were able to partially rescue the PP1-8e KO (p-value of 0.35 and 0.035, respectively) (Supplementary Figure S10). To determine whether PP1-8e catalytic activity is required for Pol II termination in L. major, we expressed an untagged PP1-8e H92K mutant in the PP1 KO. H92K mutation of PP1-8e completely eliminated phosphatase activity toward pNPP in vitro (Figure 1C). Importantly, L. major cells that exclusively express PP1-8e H92K, at identical levels of mRNA as WT from the native locus, were unable to properly terminate Pol II transcription (Supplementary Figure S10). Hence, termination of Pol II transcription requires both PP1-8e expression and activity.

DISCUSSION

These findings significantly expand our understanding of the mechanism of Pol II transcription termination in highly divergent organisms that utilize polycistronic transcription and therefore, need to decouple termination from 3’-end formation of individual genes. Pol II-mediated transcription termination of most protein-encoding genes in eukaryotes is directly linked to 3’-end formation where, according to the ‘torpedo’ model, cleavage of the nascent transcript at the poly(A) site provides access for the 5’-3’ RNA exonuclease Xrn2/Rat1 (human/yeast) (46,78–81) that eventually leads to dissociation of the polymerase from the DNA template. Thus, termination is enhanced by mechanisms that decelerate Pol II and biases the kinetic competition between Pol II and the exonuclease torpedo chasing it down (78). Pol II speed is reduced by the PP1-mediated dephosphorylation (as a component of the mammalian PTW/PP1 complex) of the transcription elongation factor Spt5, resulting in deceleration of transcription downstream of poly(A) sites enhancing torpedo dislodgment of Pol II (46). PP1 mediated dephosphorylation of Pol II CTD coordinates the recruitment of factors involved in 3’ end formation and termination (42). In yeast, PP1 dephosphorylation of both Pol II CTD and Spt5 is thought to orchestrate the recruitment of the termination factor Seb1 and the transition from elongation to termination and Pol II release (82). Identification of an analogous PJW/PP1 complex in Leishmania suggested a similar Pol II termination mechanism in kinetoplastids (37). Substitution of the mammalian Tox4 DNA binding protein with the base J binding protein (JBP3), presumably allows the complex to terminate Pol II transcription at specific sites at the end of PTUs marked by base J. According to this model, PP1-mediated Pol II deceleration would be stimulated upon reaching base J rather than poly(A) sites within the polycistronic gene array. The coupling of trans splicing and polyadenylation, preventing the generation of the 5’ phosphate substrate for the 5’-exonuclease torpedo, may explain the ability to bypass the link between Pol II termination and 3’ end formation within the gene array. Whether termination then proceeds via the ‘torpedo’ model upon recruitment of the PJW/PP1 complex remains to be tested. The analysis of the structural components of the PJW/PP1 complex (PNUTS, Wdr82 and JBP3) in T. brucei and JBP3 mutant in L. tarentolae indicated the critical role of the complex in transcription termination (37,38). While PP1 is the only catalytic component of the mammalian PTW/PP1 complex and dephosporylation by PP1-PNUTS is directly involved in regulating Pol II termination in human and yeast, analysis of PP1 function in kinetoplastid Pol II transcription has not been addressed until now (2,3,60,61). We now show that the PP1 component of the PJW/PP1 complex, PP1-8e, plays a key role in controlling the termination of Pol II transcription in kinetoplastids since the deletion of PP1-8e leads to defects in transcription termination at the 3’ ends of PTUs in L. major, similar to phenotypes seen following the knockdown of PNUTS, JBP3 or Wdr82 in T. brucei (37) and JBP3 in L. tarentolae (38). We also show that PP1-8e is a protein phosphatase and is able to directly dephosphorylate Pol II in vitro. This is the first demonstration that the dephosphorylation of Pol II in kinetoplastids is mediated by PP1 (PJW/PP1 complex) in vitro and supports our overall model of PJW/PP1 complex function in Pol II termination (Supplementary Figure S1).

Interestingly, termination defects are not seen to the same extent at all TTSs in the PP1-8e KO. The presence of Pol III transcribed RNA genes downstream of the TTS in cSSRs appears to effectively block Pol II readthrough transcription, as seen previously in Leishmania with reduced levels of base J (19,32), and there is essentially no readthrough at cSSRs at centromeric or telomeric locations. Similar differential defects in termination were recently described in the Leishmania JBP3 mutant (38). This suggests that other factors other than the PJW/PP1 complex and base J can play a role in reducing transcriptional readthrough at these loci. For example, apparent reduced read-through defects at H-T regions may reflect altered chromatin structures at termination sites adjacent to Pol II initiation complex assemblies. The impact of chromatin in Kinetoplastid transcription termination is illustrated by the role of H3V and H4V enriched at termination sites (32–34,36). Chromatin components associated with Pol II initiation may help prevent interference from upstream transcription elongation independent of PJW/PP1 complex function. The lack of defects at centromeric or telomeric locations may similarly reflect unique chromatin structures as well as compartmentalization within the nucleoplasm. Telomeres tend to be close to the nuclear periphery in trypanosomes. Compartmentalization of centromeric heterochromatin has been demonstrated in mammalian cells, often localized near nuclear lamina where chromatin is largely silenced. Furthermore, several histone markers have recently been characterized as uniquely associated with centromeres in Leishmania as well as being enriched for base J (83). Additional work is needed to identify other factors involved in termination. Transcriptional defects in the T. brucei PJW complex mutants included de-repression of genes located upstream of transcription start sites resulting from transcription between diverging PTUs (37). Apparently, similar to mammalian and yeast, Pol II transcription initiation sites are bi-directional in T. brucei giving rise to transcription in both sense and divergent antisense directions where unidirectional transcription is ensured by early termination of antisense RNA by the PJW complex. In contrast to T. brucei, we find little evidence for antisense transcription between diverging PTUs in L. major, even after the loss of PJW/PP1 complex function in the PP1-8e KO. Similar lack of antisense transcription at the 5’ ends of PTUs was also characterized in the LtJBP3 mutants (38). It is unclear why bi-directional activity and regulated antisense transcription in these regions would be restricted to Trypanosomes. However, in contrast to T. brucei, dSSRs of Leishmania lack the presence of transcriptionally regulated genes as well as being smaller in size and mostly lacking base J.

Importantly, similar to the transcription termination defects seen in base J and H3.V mutants in T. brucei and L. major (32,33,36), and PJW complex mutants in T. brucei (37), the loss of PP1-8e results in the upregulation of mRNA levels for protein-coding genes downstream of base J and H3.V marked TTS at the 3’-end of PTUs. In fact, many of upregulated genes are shared between the base J mutant and the PP1 KO in L. major. Growth of WT L. major in DMOG resulted in 10-fold reduction of the modified DNA base and termination defects. Further reduction of base J levels at termination sites by DMOG treatment of the H3.V KO revealed greater termination defects, more significant gene expression changes, and greatly reduced cell growth (36). Interestingly, reducing J levels in the PP1-8e KO by DMOG treatment led to a similar additive increase in readthrough transcription and appearance of growth defects (Figure 8C-E and Supplementary Figure S9). While readthrough transcription in L. major include the extension of Pol II onto the adjacent opposing gene cluster and dual strand transcription, we saw no evidence of transcription interference resulting in significant downregulation of mRNAs on the opposing gene cluster in cells with reduced base J (36). Similarly, we see no evidence of transcription interference in the PP1-8e KO here. Taken together these results indicate a conserved role for J and the PJW/PP1 complex in regulating Pol II transcription termination and expression of genes within polycistronic gene clusters in kinetoplastids, and suggest that the essential nature of J and the PJW/PP1 complex (32,36,38,84) in Leishmania is related to their role in repressing specific genes at the end of gene clusters rather than the prevention of dual strand transcription. Overall, these data suggest that PP1-8e works with the PJW complex to terminate Pol II transcription at the end of PTUs, sometimes leading to premature termination thereby shaping the transcriptome.

However, while we now provide a mechanistic link between J modification of DNA at termination sites, Pol II dephosphorylation and transcription termination in kinetoplastids, how do we explain the lack of PP1 in the T. brucei PJW complex and the non-essential nature of LmPP1-8e? Especially when the other structural components (PNUTS, Wdr82 and JBP3) appear to be essential for termination as well as parasite growth (37,38). PP1-8e is the only PP1 isotype associated with the Leishmania PJW/PP1 complex in vivo (37,38). Sequence analysis suggests this is due to unique insertions within the PP1 catalytic subunit and C-terminal tail important for PNUTS association (Supplementary Figure S2). While this hypothesis remains to be tested, this is reminiscent of the diversity of PP1 tails in Drosophila and mammals presumably involved in regulator protein interactions (85). In fact, the Drosophila PP1 regulatory protein ASPP (apoptosis-stimulating protein of p53) can discriminate between different PP1 isoforms based on the PP1 C-tail (85,86). The absence of these LmPP1-8e unique sequences in all 8 PP1 isoforms of T. brucei may explain why none is stably associated with the complex in T. brucei under our purification conditions. The finding that TbPNUTS contains the RVXF PP1-binding motif and is involved in termination, along with JBP3 and Wdr82, (37,38) combined with the data here showing LmPP1 function in termination, strongly suggests PP1 is involved in termination via the PJW complex in T. brucei. TbPNUTS presumably associates with at least one of the PP1 isotypes in vivo in a manner that is unstable during our complex purification methods. Interestingly, PP1-1 is closely related to the PP1 orthologue (Glc7) involved in transcription termination in yeast (Supplementary Figure S1) and has been shown to localize within the T. brucei nucleus (Tryptag.org; (77)) and associate with the PIP5Pase/RAP1 complex bound to telomeric VSG expression sites (87). The ability of LmPP1-1, or another isoform, to function in the PJW complex to a certain extent may explain why the LmPP1-8e KO is viable and the reduction of base J via DMOG results in additional defects in termination and defects in cell growth. The ability of LmPP1-1 and LmPP1-7 over-expression to partially rescue the PP1-8e KO, supports this idea. This redundancy is understandable since protein phosphatases use structurally related catalytic domains that are remarkably well conserved and shown to be relatively promiscuous in vitro (88). Furthermore, the ability of distinct phosphatases to functionally substitute has been demonstrated in vivo. For example, PP1 and PP2A-B56 phosphatases are recruited via their respective motif (RVxF and LxxIxE, respectively) containing regulatory proteins to allow control over different substrates and different mitotic processes (89–91). Removing either phosphatase produces markedly distinct phenotypic effects. However, modifying the PP1-binding motif of the regulatory protein to allow the alternative phosphatase be recruited in its place has shown they can functionally substitute for each other at kinetochores (92). Therefore, we believe some low level PNUTS complex association among the remaining PP1 isotypes allows sufficient termination control in the PP1-8e KO to remain viable.

Overall, this work provides new insight into the molecular mechanism utilized to control transcription termination at the end of PTUs in these divergent parasites. The data support conserved function of proteins involved in transcription termination among eukaryotes, despite the need of kinetoplastids to bypass termination at the 3’ end of every gene and terminate in base J specific manner within or at the end of polycistronic gene arrays. Further work is needed to understand details of this mechanism, including identifying other potential substrates of PP1-PNUTS.

DATA AVAILABILITY

All sequencing data discussed in this publication have been deposited in NCBI’s Gene Expression Omnibus and are accessible through GEO Series accession numbers GSE200788.

SUPPLEMENTARY DATA

Supplementary Data are available at NAR Online.

FUNDING

National Institutes of Health [R01AI109108 to R.S.]; Office of the Vice President for Research (to R.J.S.). Funding for open access charge: Foundation for the National Institutes of Health [R01AI109108].

Conflict of interest statement. None declared.