Secreted parasite Pin1 isomerase stabilizes host PKM2 to reprogram host cell metabolism

Metabolic reprogramming is an important feature of host–pathogen interactions and a hallmark of tumorigenesis. The intracellular apicomplexa parasite Theileria induces a Warburg-like effect in host leukocytes by hijacking signaling machineries, epigenetic regulators and transcriptional programs to create a transformed cell state. The molecular mechanisms underlying host cell transformation are unclear. Here we show that a parasite-encoded prolyl-isomerase, TaPin1, stabilizes host pyruvate kinase isoform M2 (PKM2) leading to HIF-1α-dependent regulation of metabolic enzymes, glucose uptake and transformed phenotypes in parasite-infected cells. Our results provide a direct molecular link between the secreted parasite TaPin1 protein and host gene expression programs. This study demonstrates the importance of prolyl isomerization in the parasite manipulation of host metabolism.

T he metabolic switch to aerobic glycolysis is an important characteristic of tumorigenesis and cellular reprogramming 1,2 . Metabolic exchange is also a key factor in parasite-host interactions and the manipulation of host cell phenotypes. Several parasites enter into intricate metabolic exchange with their host cells 3 . Theileria parasites are remarkable for their ability to interfere with host signaling pathways, activate nuclear transcription factors (e.g., c-Myc, HIF1α, and AP-1) and transform host leukocytes [4][5][6][7] . We previously described a Warburg-like phenotype in infected leukocytes associated with stabilization of hypoxia induced factor 1α (HIF1α) and induction of aerobic glycolytic genes 4,8,9 . We also discovered that Theileria parasites secrete a Peptidyl-prolyl isomerase (TaPin1) into the host cell, which induces proliferation via the host transcription factor c-Jun 10 . We found that TaPin1 is targeted by the theilericidal drug Burparvaquone, though there may be additional pathways targeted by this drug. In this study, we set out to identify molecular mechanisms that could link the secreted parasite TaPin1 protein to host signaling pathways. We show that TaPin1 interacts with the host Pyruvate Kinase Isoform M2 (PKM2), leading to its stabilization and subsequent HIF1αdependent induction of glycolytic enzymes that contribute to host transformed phenotypes.

Results
Parasite TaPin1 stabilizes host PKM2 protein. To search for Pin1 interactors, we expressed ectopic, tagged Pin1 in fibroblasts and performed immunoprecipitation followed by mass spectrometry analysis ( Supplementary Fig. 1a). We identified several potential interacting proteins in the cytoplasm. This list of interacting proteins is unlikely to be exhaustive, as the previously identified FBW7 protein was not found in this screen 11 . One of the most abundant Pin1-interactors was PKM2 (Supplementary Data 1). We investigated whether GST-TaPin1 could also interact with host PKM2 in extracts from bovine leukocyte cell lines infected with either T. annulata or T. parva. (Fig. 1a). To confirm this interaction, we transfected Flag-tagged PKM2 into TBL3 infected cells and showed that it interacts with the endogenous parasite TaPin1 protein (Fig. 1b). To examine the consequences of the TaPin1-PKM2 interaction, we monitored the levels of the endogenous PKM2 protein in parasite-infected TBL3 cells, compared to non-infected BL3 cells (Fig. 1c) and observed elevated levels in parasitized cells. Furthermore, treatment with two pharmacological PPIase inhibitors, Buparvaquone or Juglone 10 , led to a reduction in the PKM2 protein levels in TBL3 parasitized cells ( Fig. 1d and quantification in Supplementary Fig. 1b). Treatment with Buparvaquone or Juglone had no effect on the levels of PKM2 mRNA in parasitized TBL3 cells ( Supplementary  Fig. 1c). Inhibition of TaPin1 with Buparvaquone or Juglone or ectopic expression of TaPin1 did not change basal PKM2 protein levels in control BL3 cells ( Supplementary Fig. 1b, d). It could be that BL3 cells lack effectors required for the TaPin1 effects. To test whether parasite TaPin1 could regulate bovine PKM2 protein stability, we investigated PKM2 ubiquitination and half-life. We found that Buparvaquone/Juglone treatment induced the ubiquitination of PKM2 (Fig. 1e) and reduced the half-life of PKM2 in parasitized TBL3 cells ( Fig. 1f and Supplementary Fig. 1e) as measured by cycloheximide pulse-chase experiments. Together these results showed that the Theileria parasite TaPin1 prolyl isomerase interacts (directly or indirectly) with host bovine PKM2 and leads to its stabilization.
Parasite TaPin1 regulates host cell metabolism. In addition to its role in phosphoenolpyruvate phosphorylation, PKM2 acts as a cofactor for HIF1α, a transcription factor critical for the Warburg effect and the transcription of glycolytic enzymes in cancer cells [12][13][14] . We tested whether stabilization of host PKM2 by the parasite TaPin1 protein could affect HIF1α functions. We observed that TaPin1 inhibition (via Buparvaquone or Juglone treatment) reduced transcriptional activity of HIF1α (40-50%), measured by a hypoxia-responsive element (5×HRE) Luciferase reporter (Fig. 2a). The reduced HIF1α activity correlated with reduced expression of HIF1α-target genes linked to host cell metabolism, namely genes encoding the glycolytic enzymes Hexokinase 2 [HKII], the Glucose transporter 1 [GLUT1], Pyruvate dehydrogenase kinase [PDK1] and Lactate dehydrogenase [LDHA]. The expression of HIF1α-target genes was reduced at the mRNA level (40-70%) (Fig. 2b) and protein levels (Fig. 2d). The PPIase inhibitors had no effect on the expression of HIF1α transcripts (Fig. 2b). Experiments in control BL3 cells indicated that Buparvaquone or Juglone treatment did not affect the expression of glycolytic enzymes in unparasitized cells (Supplementary Fig. 2a). To show that the regulation of metabolic enzymes could be via parasite TaPin1-dependent stabilization of host PKM2 protein, we transfected exogenous PKM2 into TBL3 cells prior to treatment with TaPin1 inhibitors. The forced expression of PKM2 rescued expression of the metabolic enzymes (GLUT1, LDHA, and PDK1) in the presence of Juglone inhibitor (Fig. 2c). Conversely, siRNA silencing of endogenous, bovine PKM2 reduced HIF1α-activity using the 5xHRE luciferase reporter assay in TBL3 cells (Fig. 2e, f). This correlated with a marked reduction in the expression of host glycolytic enzymes without affecting the levels of HIF1α transcripts in TBL3 cells (Fig. 2g). Notably, siRNA directed against PKM2 did not decrease the expression of the glycolytic enzymes in control BL3 cells ( Supplementary Fig. 2b, c).
Parasite TaPin1 can partially rescue PKM2 regulation. To provide further support for a role of TaPin1-PKM2 in the regulation of host metabolic enzymes, we tested TaPin1's ability to rescue the Buparvaquone effects. Overexpression of TaPin1 could partially rescue the expression of PKM2 proteins in TBL3 cells treated with Buparvaquone ( Fig. 3a) leading to a partial rescue of the expression of glycolytic enzymes (Fig. 3b). To demonstrate the importance of the prolyl isomerase activity, we tested TaPin1-K38A and TaPin1-S42E catalytic mutants 10,15 . These failed to rescue PKM2 stabilization or glycolytic enzyme expression ( Supplementary Fig. 3a, b). We previously reported a mutation in the parasite TaPin1 gene that resulted in a Buparvaquone-resistant protein 10 . Interestingly, transfection experiments in TBL3 cells showed that the TaPin1-A53P mutant was not sensitive to Buparvaquone effects on PKM2 protein levels ( Fig. 3c) and the expression of the glycolytic enzymes in TBL3 cells (Fig. 3d). The TaPin1-A53P mutant remains sensitive to the Juglone drug 10 and failed to maintain PKM2 protein levels in Juglone-treated TBL3 cells (Fig. 3c) or glycolytic enzyme expression (Fig. 3d). Finally, we performed knockdown experiments to exclude a role for the endogenous bovine Pin1 protein in regulating PKM2 and metabolic enzymes in TBL3 cells. We tested siRNA knockdown of endogenous bovine Pin1 and showed that siBtPin1 did not affect PKM2 protein expression ( Supplementary  Fig. 4a) or HIF1α-target gene expression ( Supplementary Fig. 4b).
These combined data support a role for the secreted TaPin1 protein in regulating host gene expression through prolyl isomerase stabilization of host PKM2 and the HIF1α pathway.
The TaPin1-PKM2 axis is important for glucose uptake and host cell transformation. To test the functional significance of TaPin1-PKM2 signaling on host transformation, we measured the effects of TaPin1 inhibitors on glucose uptake in parasitized cells. Parasitized TBL3 cells exhibited a 9-fold increase in glucose uptake compared to control BL3 cells (Fig. 4a). This parasiteinduced effect was reduced by treatment with Juglone or Buparvaquone ( Fig. 4a) with no effect in BL3 cells. As demonstrated above for PKM2 stabilization and glycolytic enzyme expression, the Buparvaquone and Juglone effects on glucose uptake could be partially rescued by overexpression of ectopic PKM2 in TBL3 cells (Fig. 4b, c). Furthermore, knockdown of endogenous PKM2 in parasitized cells also led to a reduction in glucose uptake (Fig. 4d). Notably, the overexpression of ectopic PKM2 or siRNA-mediated knockdown PKM2 did not affect the glucose uptake of control BL3 cells (Supplementary Fig. 2d and Fig. 4d). Finally, we tested the contribution of TaPin1 and PKM2 to the transformed phenotype of parasitized cells. Treatment of parasite-infected TBL3 cells with TaPin1 inhibitors, Juglone or Buparvaquone, or knockdown of endogenous PKM2, each led to a marked decrease in cell proliferation ( Supplementary Fig. 5a, b) and the growth of colonies in soft agar, an effective in vitro assay for transformation (Fig. 4e, f). Once again, siPin1 knockdown of bovine Pin1 did not affect colony growth of TBL3 cells ( Supplementary Fig. 4c). These combined experiments suggest that the stabilization of host bovine PKM2 protein by the secreted parasite TaPin1 PPIase leads to activation of HIF1α-dependent metabolic genes that are essential for glucose metabolism and host cell transformation.

Discussion
Metabolic reprogramming is a hallmark of cancer cells and is critical for tumor cell survival and proliferation 2,16 . Previous  reports of Warburg-like "aerobic glycolysis" in bovine leukocyte cell lines transformed by Theileria parasites 17,18 lacked a direct molecular link between intracellular parasites and host metabolic gene expression. Here, we identified the TaPin1-PKM2-HIF1α axis as an integrator of parasite-host interaction. We showed that TaPin1 and bovine PKM2 interact (either directly or indirectly) and that this interaction leads to PKM2 stabilization and HIF1αdependent glycolytic enzyme expression. PKM2 plays a critical role in the metabolic rewiring that underlies tumorigenesis 19 . Non-canonical PKM2 functions include nuclear transcriptional regulation [12][13][14]20 and enhanced HIF1α binding to HRE in metabolic genes 13,14 . The PKM2 protein is finely tuned by cancer cells, involving mRNA splicing, ERK-dependent phosphorylation, and subcellular relocalization. We propose prolyl isomerization and protein stabilization as another level of regulation that parasites exploit to maintain a Warburg-like phenotype. These data add to the descriptions of the complex metabolic relationships between parasites and their host cells 3 . Increased glucose flux and energy metabolism could support the biosynthetic requirements of both the parasite and the hyperproliferating host cell. The secreted parasite TaPin1 protein could activate multiple pathways that are important for host cell proliferation and metabolism. For example, TaPin1 stabilizes the host transcription factor c-Jun by regulating a ubiquitin ligase, FBW7 10 . Additional secreted proteins may contribute to subverting the host cell and perhaps TaPin1 has additional targets. Here, we show that TaPin1 also activates HIF1α-regulated target genes via stabilization of PKM2. These results provide a molecular link between prolyl isomerization and metabolic manipulation by intracellular parasites.

Methods
Cell lines and culture conditions. All infected bovine cell lines were previously described: the TBL3 cell line was derived from in vitro infection of the spontaneous bovine-B lymphosarcoma cell line, BL3, with Hissar stock of T.annulata. The TpMD409 lymphocyte cell line is infected with T. parva. The culture conditions of these cell lines were described previously 10 . Parasite-infected cell lines were provided by the Langsley laboratory. Cells were cultured in a humidified 5% CO 2 atmosphere at 37°C in RPMI 1640 (Gibco-BRL), supplemented with 4 mM L-Glutamine, 25 mM HEPES, 10 µM Beta-mercaptoethanol, 10% heat-inactivated Fetal calf serum and 100 µg/ml penicillin/streptomycin. All cell lines were mycoplasma negatives. The anti-parasite drug Buparvaquone (BW720c) was used for 72 h at 200 ng/ml (Chemos GmbH, Ref: 88426-33-9) (Supplementary Fig. 5a, b). Cells were treated 72 h with Juglone at 5 µM resuspended in Ethanol (Sigma, Ref: H47003).
Plasmids and transfection. Bovine PKM2 (NM_001205727/NP_001192656.1) was cloned between restriction sites HindIII and EcoRI in p3xFlag-myc-CMV-24 using the following oligonucleotides: Fwd-cccaagctttcgaagcaccacagcgacg and Rev-CGGAATTCGAtggcacaggaactacacg. Parasite gene TaPin1 (TA18945) was cloned between restriction sites XhoI and NotI in pREV-HA-Flag-RIL2 using oligonucleotides: Fwd-CCGCTCGAGGCCCACTTGCTACTAAAG and Rev-ATAAGAATGCGGCCGCTTATGCGATTCTATATATAAGATG. Point Original blot images are shown in Supplementary Fig. 7e. e siPKM2 decreased HIF1α activity on 5xHRE Luciferase reporter. f Efficiency of siPKM2. Bovine Beta-actin was used as a loading control. Original blot images are shown in Supplementary Fig. 7f. g PKM2 knockdown decreased the mRNA levels of the HIF1α targets (siPKM2 or siControl). Bovine Beta-actin and H2A mRNAs were used for normalization. Data represent three independent experiments (average ± sd). The p-values were calculated using the Dunnett test for multiple comparisons (Fig. 2a, b). The p-values were calculated using the Mann-Whitney test (Fig. 2c-f). *p < 0.05, **p < 0.01, ***p < 0.001 ARTICLE COMMUNICATIONS BIOLOGY | https://doi.org/10.1038/s42003-019-0386-6 mutations TaPin1 K38A, A53P 10 and S47E were created from pRev-HA-Flag-TaPin1 WT-RIL2 using a set of primers following a 3-step PCR using the following oligos: Immunoblot analysis and immunostaining. Total proteins were extracted with Laemmli lysis buffer. Samples were sonicated: 30 s ON/30 s OFF for 5 min, and then resolved on 10.5% acrylamide/bis-acrylamide SDS-PAGE gels and transferred to nitrocellulose membranes (Thermo Fisher Scientific, MA, USA) in transfer buffer. Protein transfer was assessed by performing a Ponceau-red staining. Membranes were then blocked for 1 h at room temperature in Trisbuffered saline pH 7.4 containing 0.1% Tween-20 and 5% milk. Incubations with primary antibodies were performed at 4°C overnight using antibody dilutions as manufacturer recommendations in Tris-buffered saline pH 7.4, 0.05% Tween-20 and 5% milk. Proteins were detected by chemiluminescence (Thermo Fisher Scientific) following the manufacturer's instructions following 1 h incubation at room temperature with an anti-rabbit or anti-mouse peroxidase- Cycloheximide chase assay. This was performed as previously reported 10 Supplementary Fig. 7h. Data represent three or four (Fig. 3b) independent experiments (average ± sd). The p-values were calculated using the Bonferroni test for multiple comparisons (Fig. 2b) and Dunnett test for multiple comparisons to the control (Fig. 2d). Con, control. **p < 0.01, ***p < 0.001, ****p < 0.0001 lysed using Laemmli sample buffer, resolved by SDS-PAGE and analyzed by western blot using the indicated antibodies. Relative quantification indicates the PKM2/ Tubulin ratios calculated with Image J software (NIH) and PKM2 levels at time 0 was set as 1. Cycloheximide chase experiments were repeated for four independent biological replicates. GST pull-down. This was performed as previously reported 10 . Briefly, TaPin1 and hPin1 were cloned between restriction sites BamHI and EcoRI in pGEX-2T plasmid which was kindly provided by G. Del Sal (LNCIB-Laboratorio Nazionale CIB, Trieste, Italy). Plasmid constructs were expressed in E. coli strain BL21 and then purified using glutathione-sepharose beads. Concentration of purified protein was estimated using Coomassie blue staining. One microgram of GST fusion proteins coated beads were incubated with 250 µl of cell lysate in 50 mM Tris pH7.6, 150 mM NaCl, 0.1% Triton, for 2 h at 4°C. Beads were washed five times with 50 mM Tris pH7.6, 300 mM NaCl, 0.5% Triton. Proteins were than revealed by Western Blot analysis using indicated antibodies.
Complex immunopurification and mass spectrometry analysis. We used retroviral transduction strategy to establish NIH/3T3 cell lines expressing doubletagged proteins. Polyclonal NIH/3T3 cell lines stably expressing Flag-HA-tagged hPin1 or TaPin1 were established. All the proteins were tagged with double-HA (Haemagglutinin) and double-Flag epitopes at the N-terminus. A control cell line transduced with the empty pREV vector was established. We carried out doubleaffinity purification of Flag-HA-hPin1 from NIH/3T3, using either nuclear soluble or cytoplasmic fractions. Both fractions were then subjected to a two-step immunopurification with Flag and HA antibodies as described previously 21 . Mass spectrometry identification of proteins was carried out in the Taplin Biological Mass Spectrometry Facility (Harvard Medical School, Boston, USA) and the results are shown in Supplementary Data 1.
Data and statistical analysis. The GraphPad PRISM 7 was used for statistics. In all the figures the results represent the mean ± sd of at least three independent experiments. Statistical analysis was performed using the Dunnet for multiple comparisons with the control condition or Bonferroni test for multiple comparisons between samples or Mann-Whitney test for the comparison between two conditions. p values of <0.05 were considered statistically significant and are indicated with asterisks *p < 0.05; **p < 0.01; ***p < 0.001, ****p < 0.0001.
Reporting summary. Further information on experimental design is available in the Nature Research Reporting Summary linked to this article.

Data availability
All data generated or analysed during this study are included in this published article, Supplementary