Extracellular-signal regulated kinase 8 of Trypanosoma brucei uniquely phosphorylates its proliferating cell nuclear antigen homolog and reveals exploitable properties

ABSTRACT

The Trypanosoma brucei subspecies T. brucei gambiense and T. brucei rhodesiense are vector-borne pathogens that cause sleeping sickness also known as Human African Trypanosomiasis (HAT), which is fatal if left untreated. The drugs that treat HAT are ineffective and cause toxic side effects. One strategy for identifying safer and more effective HAT drugs is to therapeutically exploit essential gene targets in T. brucei. Genes that make up a basic mitogen-activated protein kinase (MAPK) network are present in T. brucei. Tb927.10.5140 encodes an essential MAPK that is homologous to the human extracellular-signal regulated kinase 8 (HsERK8) which forms a tight complex with the replication factor proliferating cell nuclear antigen (PCNA) to stabilize intracellular PCNA levels. Here we demonstrate that (TbPCNA) is uniquely phos-phorylated on serine (S) and threonine (T) residues in T. brucei and that TbERK8 phosphorylates TbPCNA at each of these residues. The ability of an ERK8 homolog to phosphorylate a PCNA homolog is a novel biochemical property that is first demonstrated here in T. brucei and may be unique to this pathogen. We demonstrate that the potent HsERK8 inhibitor Ro318220, has an IC₅₀ for TbERK8 that is several hundred times higher than its reported IC₅₀ for HsERK8. This indicated that the active sites of TbERK8 and HsERK8 can be selectively inhibited, which provides a rational basis for discovering inhibitors that specifically target this essential parasite MAPK to kill the parasite.

Introduction

Two Trypanosoma brucei subspecies, T. brucei gambiense and T. brucei rhodesiense, are the causative agents of Human African Trypanosomiasis (HAT) or sleeping sickness, a meningoencephalitic disease. This parasite is endemic in the sub-Saharan region of Africa where it is transmitted by the bite of infected tsetse fly from the Glossina species and is most often fatal if left untreated.¹ The available HAT treatments suffer from toxic side effects, difficulty in administration, and a lack of targets that can be selectively inhibited in the parasite. The T. brucei genome possesses the basic components of a mitogen-activated protein kinase (MAPK) signaling network^2,3 found in eukaryotic cells. Such networks are generally regulated by sequential phosphorylation of a 3-tiered superfamily of serine/threonine protein kinases consisting of MAPK kinase kinases, MAPK kinases, and MAPKs.^4-6 Once activated, MAPKs phosphorylate downstream effectors that include transcription factors, protein kinases, and phosphatases.⁷ In T. brucei, MAPKs have been shown to regulate differentiation, stress response, and transmission in both the insect and mammalian stages of this heteroxenous parasite.^8-11 The important role that MAPKs have in regulating many critical processes in T. brucei makes them attractive for therapeutic exploitation. We previously demonstrated that depleting Tb927.10.5140 expression in bloodstream form T. brucei arrests proliferation and leads to its death, discovering that this MAPK is essential for the pathogenic form of the parasite.¹² We demonstrate in this study that the Tb927.10.5140 gene, which encodes an extracellular signal-regulated kinase 8 (TbERK8) homolog uniquely phosphorylates the essential replication factor proliferating cell nuclear antigen of T. brucei (TbPCNA) in vitro and in the parasite. We discuss the distinct biochemical properties of TbERK8, which opens opportunities for discovering new small molecules that can kill T. brucei by selectively inhibiting TbERK8.

Results

Phylogenetic analysis of TbERK8

The amino acid sequence of Tb927.10.5140 was analyzed with BLAST on the Orthologous group tool¹³ to retrieve a non-biased group of orthologous sequences. Redundant sequences were filtered out of the group while TbMAPK1 (Tb927.10.7780) and TbMAPK3 (Tb927.8.3550) were added in to optimize the ClustalW alignment among the orthologous groups. Table 1 lists the genes identified from the search that were used to build this phylogenetic tree (Fig. 1A). Human sarcoma (SRC) tyrosine kinase was used as the out-group because tyrosine kinases are not encoded in the T. brucei genome.³ Based on the phylogenetic tree, we found that Tb927.10.5140 was located in the clade with HsERK8, whereas TbMAPK1 and TbMAPK3 were more distantly related (Fig. 1A). ClustalW alignment of Tb927.10.5140 with the open reading frame of HsERK8 demonstrated that these 2 kinases were about 36% identical at the amino acid level. The kinase domains of these 2 proteins from divergent species are about 51% identical at the amino acid level based on ClustalW alignment. Tb927.10.5140 also has a long C-terminus that extends 100 amino acids beyond the kinase domain that is not conserved, which is a defining characteristic of the ERK8 family members¹⁴ (Fig. 1B). We adopted the nomenclature of TbERK8 for Tb927.10.5140 because its coding region encodes a protein that was most closely related to HsERK8.^12,14

Table 1.

ERK8 orthologs. The Tb927.10.5140 ORF was searched using BLASTO on the Orthologous group tool to retrieve an unbiased group of related eukaryotic MAPKs listed below.

Kinase	Accession Number
Tb ERK8*	Tb927.10.5140
Ce ERK7	Q11179
Dm ERK7	AAF46481
Hs ERK8/MAPK15	NP_620590
Sp MAPK Spk1	NP_594009
Hs MAPK3	NP_002737
Sc MAPK1/3 fus3	NP_009537
Hs MAPK7	NP_002740
Ce MAPK-1	CE24971
Hs MAPK2	Hs20986529
Hs MAPK1	NP_620407
Hs MAPK13	NP_002745
Dm p38b	NP_477361
Dm MAPK2	NP_477163
Sp Pmk1	NP_595289
Hs MAPK14	AAP36304
Hs MAPK11	NP_002742
Hs MAPK12/p38gamma	AAB40118
Sc KSS1p	NP_011554
Sc Kdx1p	NP_012761
Hs MAPK4	CAA42411
Hs MAPK6	NP_002739
Sc Smk1p	NP_015379
Ce Protein SMA-5, isoform a	NP_741909
Ce NLK Isoform c	NP_001022807
Ce lit-1	NP_001022806
Ce lit-1_b	NP_001022805
Tb MAPK3	Tb927.8.3550
Tb MAPK1	Tb927.10.7780
SRC_tyrosine_kinase	NP_033297

Note. Tb: Trypanosoma brucei, Ce: Caenorhabditis elegans, Dm: Drosophila melanogaster, Sp: Schizosaccharomyces pombe, Sc: Saccharomyces cerevisiae, Hs: Homo sapiens.

Figure 1.

Phylogenetic tree and amino acid sequence alignment of ERK8-related kinases. (A) Phylogenetic relationship of the sequences and the credibility weight of the groupings were inferred using MrBayes after running 100,000 generations. Relationships with confidence values greater than 0.5 are shown at the branch points of the phylogenetic tree. The tree shown is derived from alignment with the ends trimmed to remove regions that were represented by only a few sequences. A separate analysis was done using the entire alignment and the tree topology was nearly identical to the tree shown here. (B) ClustalW alignment of ERK8 homologs. Polypeptide sequences obtained from TbERK8 BLAST on Orthologous groups were aligned using ClustalW with a gap opening penalty of 5 and a gap extension penalty of 0.05. The kinase domain is shown inside the box and the triangles (▴) highlight the PIP-box sequences. Tb: Trypanosoma brucei, Hs: Homo sapiens, Ce: Caenorhabditis elegans, Dm: Drosophila melanogaster, Sp: Schizosaccharomyces pombe, Sc: Saccharomyces cerevisiae, Rat: Rattus rattus.

Examining the kinase activity of TbERK8

We tested TbERK8 purified from baculovirus (TbERK8_BV) for its ability to autophosphorylate by incubating it in buffer K that contained ³²P-γ-ATP. Autophosphorylation was determined in TbERK8 by using autoradiography to detect its incorporation of ³²P, indicating that recombinant TbERK8 purified as an active protein (Fig. 2A). We further designed a C-terminal truncated form of TbERK8 that expressed only the kinase domain residues (1-343 of TbERK8) fused to GST (GST-TbERK8_KD) and re-tested its ability to autophosphorylate. The GST-TbERK8_KD also incorporated ³²P when incubated in kinase buffer with ³²P-γ-ATP substrate, indicating that it was able to autophosphorylate (Fig. 2B). Typically, ERKs become fully activated by upstream dual-specific MAPK-kinases that phosphorylate a conserved T-X-Y motif in the activation loop to pT-X-pY.^7,15 ERK8 family members are atypical because full activation is catalyzed by autophosphorylation of the T-X-Y residues that make up the activation loop instead of by a dual-specific upstream MAPK-kinase.^7,14-16 We demonstrated that TbERK8 could autophosphorylate tyrosine residues by using the PY20 antibody (Fig. 2C). The tryptic peptides of phosphorylated TbERK8 were analyzed by LC-MS/MS to further investigate if the tyrosine residue in the activation loop was phosphorylated. This analysis revealed the phosphorylated residues pT₁₇₄ and pY₁₇₆, which correspond to the activation loop motif of TbERK8 (Fig. 2D). Based on these observations, we conclude that TbERK8 has dual-specific activity that allows it to autophosphorylate the T and Y residues that make up its activation loop.

Figure 2.

Autophosphorylation of TbERK8. Top panels: (A) Autoradiograph of purified recombinant TbERK8_BV incubated in kinase reaction buffer without (−) or with (+) 10 µCi ³²P-γ-ATP; (B) Autoradiograph showing the autophosphorylation of the GST-TbERK8 kinase domain (GST-TbERK8_KD) that lacks the C-terminus. GST-TbERK8_KD was tested for autophosphorylation in a kinase reaction without (−) or with (+) 10 μCi ³²P-γ-ATP; (C) Immunoblot using PY20 antibodies to detect phosphotyrosine residues associated with autophosphorylation of TbERK8_BV. Bottom panels are StainFree® gels showing the loading controls for TbERK8 samples. (D). LC-MS/MS mass spectrum of ₁₆₆EQVARPVLT(phospho)DY(phospho)IMTR₁₈₀ peptide from TbERK8 containing the phosphorylated T₁₇₄ and Y₁₇₆ residues, with m/z = 651.30 (3+), 0.52 ppm error.

Examination of the putative TbERK8 PIP-box

Many proteins form stable interaction with PCNA through the conserved PIP-box motif. TbERK8 contains a putative PIP-box motif that is 75% identical to the functional PIP-box of HsERK8 (Fig. 1B arrows, and 3A). One of the proposed functions for the PIP-box motif of HsERK8 is to mediate its interaction with HsPCNA and form a stable complex that prevents PCNA from being degraded by the proteasome.¹⁷ We generated the GST-TbERK8-PIP-box fusion protein (GSTtbE8-PIP) depicted in Figure 3B to test the ability of this motif to interact with TbPCNA. Figure 3C shows a representative immunoblot of lysates from tetracycline-induced T. brucei expressing TbPCNA_HA. Pull-down assays were done by incubating T. brucei lysates from tetracycline-induced parasites with GST control (GST-Ctrl) or with GSTtbE8-PIP beads. The GSTtbE8-PIP beads pulled down about 5 times more TbPCNA than the GST-Ctrl beads (Fig. 3D and E). To assess if the putative TbERK8 PIP-box motif was fully accessible when fused to the end of GST, we did orthogonal pull down experiments using 2 canonical PIP-box motifs from unrelated proteins: (1) T. brucei aurora kinase 1 (TbAUK1) and (2) paraflagellar complex (TbPFC19) (Tb927.11.8220 and Tb927.10.10140, respectively). The PIP-box of either TbAUK1 or TbPFC19 was similarly fused with GST and immobilized on GST beads. We found that these 2 PIP-box motifs were respectively 5-fold and 3-fold more efficient at pulling down TbPCNA_6H than the TbERK8 PIP-box motif as represented by the experiments in Figure 3F. In comparison to the TbERK8 peptide sequence, the TbAUK1 and TbPFC19 PIP-boxes pulled down about 5x and 3x more TbPCNA_6H with p values of 0.045 and 0.047 respectively as determined by Student's t-test (Fig. 3G and H). This indicates that each of the motifs that we tested was exposed in our GST-fusion proteins; however, the TbERK8 motif was less efficient at pulling down TbPCNA than the TbAUK1 or TbPFC19 motifs were. We also did in vivo pulldown assay with hemagglutinin-tagged TbERK8 (TbERK8_HA) expressed in T. brucei followed by mass spectrometry analysis. This pull-down strategy was not successful at detecting stable interactions between TbERK8 and TbPCNA (data not shown). Based on these observations, we conclude that the putative PIP-box of TbERK8 is very poor at forming a stable complex with TbPCNA in comparison with functional canonical PIP-box motifs from other T. brucei proteins.

Figure 3.

Testing putative TbERK8 PIP-box in TbPCNA pull-down assays_. (A) Alignment of putative PIP-box motifs from ERK8 homologs of T. brucei (Tb), T. congolense (Tco), T. cruzi (Tcr), and H. sapiens (Hs). The putative PIP-box motifs are 100% identical in related trypanosome species and are 75% identical to the functional PIP-box of HsERK8. (B) Schematic presentation of GST-T. brucei PIP-box fusion proteins. The PIP_box in the schematic represents the putative TbERK8 PIP-box or canonical PIP-box sequences from TbAUK1 or TbPFC19. (C) Immunoblot analysis of wt-TbPCNA_HA in control (Tet −) or tetracycline-induced (Tet +) bloodstream form T. brucei with antisera to hemagglutinin (α-HA). Bottom panel shows Coomassie-stained gel image of variant surface antigen (VSG) loading control. (D) Examination of wt-TbPCNA_HA pulled down by GST-control (GST-Ctrl) or GST-TbERK8-PIP-box (GSTtbE8PIP-box) glutathione agarose beads. Top panel shows the Coomassie-stained gel image of GST fusion proteins, middle panel shows the immunoblot analysis using α-HA antibodies to detect wt-TbPCNA_HA, and lower panel shows Coomassie-stained gel image of VSG loading control. (E) Quantitation of wt-TbPCNA_HA pull down. Bar graph showing results from densitometry trace comparing the amount of TbPCNA pulled down by GST-Ctrl beads to that pulled down by GST-tbERK8-PIP beads. Values are mean arbitrary units (AU) with standard deviation for 3 independent experiments. (F) Pull down of recombinant TbPCNA_6H by GST-bound TbERK8 PIP-box (GSTtbE8-PIP) and GST-bound canonical PIP-boxes from TbAUK1 (GSTtbAUK1-PIP) and TbPFC19 (GSTtbPFC-PIP). The top 2 panels show StainFree® gels of TbPCNA_6H bound to the beads in comparison to the loading controls. The third StainFree® gels of the GST-PIP-box proteins used in pull-downs. The bottom 2 panels show the immunoblot analysis of TbPCNA_6H bound to the beads in comparison to the loading controls. (G) Mean quantitation values from 3 independent experiments of TbPCNA_6H pulled down by various GST-PIP-boxes. Error bars show standard deviation (*** indicate p-value < 0.05). (H) TbPCNA calibration curve from StainFree® gels showing average RFU intensity values with standard deviations from replicate experiments. (Relative Fluorescent Units, RFU), (Arbitrary Units, AU).

Examination of TbPCNA from in vivo pull-down assays

The reciprocal in vivo pulldowns were also done with hemagglutinin-tagged TbPCNA (TbPCNA_HA). We immunoprecipitated TbPCNA from stably transfected T. brucei strains that express wt-TbPCNA_HA at endogenous or overexpressed levels and subjected it to LC-MS/MS analysis. This pulldown strategy was also ineffective at detecting interactions between TbERK8 and TbPCNA (data not shown). In addition we examined the phosphorylation status of TbPCNA_HA pulled down from T. brucei by LC-MS/MS. This analysis showed that TbPCNA_HA pulled down from T. brucei was phosphorylated at residues T₂₀₂ (Fig. 4A) and S₂₁₆ (Fig. 4B) whether it was isolated from T. brucei that expressed it at endogenous or overexpressed levels. These phosphorylated residues are situated in the unique insertion region of TbPCNA (Fig. 4C). Furthermore, this observation indicates that an unknown kinase activity within T. brucei is able to phosphorylate TbPCNA at these residues.

Figure 4.

In vivo phosphorylation of TbPCNA_HA immunoprecipitated from T. brucei. (A) LC-MS/MS mass spectrum of ₁₉₄SKGESDVKT(phospho)EDEEADACSVR₂₁₃ peptide from TbPCNA containing a phosphorylated S₂₁₁ residue, with m/z = 764.65 (3+), 2.7 ppm error. (B) LC-MS/MS mass spectrum of ₂₁₄THS(phospho)AKGKDGPLGIGVDVR₂₃₁ peptide from TbPCNA containing a phosphorylated S₂₁₆ residue, with m/z = 629.65 (3+), 8.9 ppm error. (C) ClustalW alignment comparing TbPCNA with HsPCNA showing unique insertion between residues 191 and 228. Arrows show T202 and S216 residues that are phosphorylated.

TbERK8 phosphorylates TbPCNA in its unique insertion

Both wt-TbERK8_HA and K42A-TbERK8_HA mutant was overexpressed in stable T. brucei transfectants by tetracycline induction (Fig. 5A). We demonstrated that the K42A-TbERK8_HA mutant was defective in ³²P-γ-ATP kinase autophosphorylation assays (Fig. S1). These overexpressed TbERK8 proteins were adsorbed to anti-HA affinity beads for use in kinase assays with buffer K and ³²P-γ-ATP to verify their activity. The generic kinase substrate myelin basic protein (MBP) and recombinant TbPCNA_6H purified from E. coli were tested as substrates in the kinase assay. The autoradiographs from the kinase assays demonstrated that TbPCNA was phosphorylated with a greater efficiently than MBP by TbERK8 immunoprecipitated from T. brucei (Fig. 5B). As a mock control, the affinity beads were incubated in lysates from non-transgenic T. brucei 221 and only background labeling of both MBP and TbPCNA was detected (Fig. 5C). When anti-HA affinity beads adsorbed with the K42A-TbERK8_HA mutant were used in kinase reactions, only background levels of phosphorylation similar to those of the mock control were observed (Fig. 5D). The wt-TbERK8_HA phosphorylated TbPCNA_6H about 20-fold above the mock control and about 7-fold above the K42A-TbERK8_HA (Fig. 5E and F). The ability to label TbPCNA_6H in this kinase assay clearly depended on the K42 residue of TbERK8 being intact. We conclude from these results that the kinase activity of TbERK8 was responsible for phosphorylating TbPCNA in the kinase assay. We repeated the kinase assays with affinity beads adsorbed with wt-TbERK8_HA in buffer K that had cold ATP to analyze the phosphorylated TbPCNA_6H substrate by mass spectrometry. LC-MS/MS analysis of tryptic TbPCNA_6H peptides from the kinase assay identified peptides that were phosphorylated at residues T₂₀₂ and S₂₁₁ (Fig. 5G). The pT₂₀₂ residue matched the position phosphorylated in TbPCNA_HA immunoprecipitated from T. brucei and the pS₂₁₁ residues is also within the unique insertion region of TbPCNA.

Figure 5.

Activity assay with immunoprecipitated TbERK8. (A) Immunoblot verifying the expression of wt-TbERK8_HA and K42A-TbERK8_HA in the parasite after tetracycline induction (+). Arrow points to the 52-kDa HA-tagged TbERK8 band recognized by α-HA antibodies. T. brucei variant surface antigen (VSG arrow) was used for the loading control. ³²P-Kinase assay with wt-TbERK8_HA immunoprecipitated from T. brucei. Top panel shows the autoradiograph of kinase immunoprecipitation assays done with immuno-affinity beads pre-adsorbed with (B) wt-TbERK8_HA, (C) T. brucei 221 lysate and (D) K42A-TbERK8_HA. Each of the 3 sets of beads were tested with either no substrate (first lane of each experiment in the top row), 100 μmol of MBP (second lane of each top panel), or 100 μmol of TbPCNA (third lane of each panel). The second panel shows the StainFree® loading controls for MPB and TbPCNA_6H. The third panel shows the α-HA immunoblot detecting the HA-tagged TbERK8 bound to the immuno-affinity beads (absent in beads pre-incubated with T. brucei 221 lysates). The bottom panel is a α-IgG immuno-blot as a loading control for the affinity beads. (E) LC-MS/MS mass spectrum of ₁₉₄SKGESDVKT(phospho)EEDEEADAC(cam)S(phosphor)VR₂₁₃ peptide from TbPCNA containing phosphorylated T₂₀₂ and S₂₁₁ residues, with m/z = 791.31.65 (3+), 1.9 ppm error.

We also tested TbERK8_BV for activity against TbPCNA using kinase buffer K with ³²P-γ-ATP. The autoradiographs from kinase reactions revealed that both TbERK8_BV and the 36-kDa TbPCNA_6H were phosphorylated (Fig. 6A). We repeated the kinase reactions in buffer containing cold ATP and resolved the phosphorylated TbPCNA by SDS-PAGE for examination by LC-MS/MS. More than 96% of the TbPCNA sequence was covered by the peptides detected by LC-MS/MS. Phosphorylated residues corresponding to S₂₁₁ (Fig. 6B) and S₂₁₆ (Fig. 6C) were identified on 2 of the peptides. The phosphorylated S₂₁₆ residue matched the phosphorylated serine residue identified from TbPCNA that was immunoprecipitated from T. brucei. The phosphorylated S₂₁₁ identified by this analysis is also located within the unique TbPCNA insertion.

Figure 6.

Recombinant TbERK8_BV phosphorylates TbPCNA_6H in vitro. (A) LC-MS/MS mass spectrum of ₁₉₄SKGESDVKTEDEEADACS(phospho)VR₂₁₃ peptide from TbPCNA containing a phosphorylated S₂₁₁ residue, with m/z = 1146.47 (2+), 5.1 ppm error. (B) LC-MS/MS mass spectrum of ₂₁₄THS(phospho)AKGKDGPLGIGVDVR₂₃₁ peptide from TbPCNA containing a phosphorylated S₂₁₆ residue, with m/z = 629.65 (2+), 1.9 ppm error.

In vivo phosphorylation of TbPCNA reduced upon TbERK8 depletion–We examined the effect of TbERK8 depletion on TbPCNA phosphorylation in T. brucei using the plasmid, pHAR (Fig. 7A). T. brucei stably transfected with this plasmid simultaneously expressed HA-tagged TbPCNA under the control of the endogenous TbPCNA allele and tetracycline-inducible RNAi expression against TbERK8 mRNA. Tetracycline induction of T. brucei stably expressing pHAR resulted in a decrease of TbERK8 mRNA levels as well as decreased proliferation (Fig 7B). This result was consistent with our previous study of depleting TbERK8 in bloodstream form T. brucei.¹² We also observed that the in vivo levels of endogenously-tagged TbPCNA decrease after 48 h of TbERK8 depletion (Fig. 7C). Neither T₂₀₂, S₂₁₁ nor S₂₁₆ residues were phosphorylated in TbPCNA_HA immunoprecipitated from T. brucei after 48 h of TbERK8 depletion (data not shown). The mass spectrometry analysis of the TbPCNA_TSS triple mutant control expressed and immunoprecipitated in T. brucei did not detect any phosphorylated residues. Table 2 summarizes the phosphorylated TbPCNA residues that we have identified in vivo and in vitro. We conclude that TbERK8 phosphorylates TbPCNA in vivo.

Figure 7.

TbERK8 levels have an effect on TbPCNA phosphorylation in vivo. (A) Map of the plasmid pHAR. The plasmid has an HA-tag for endogenous expression of TbPCNA and an RNAi construct for TbERK8 silencing. (B) Decreased proliferation in pHAR stable parasites with tetracycline induction to reduce TbERK8 mRNA levels. Sybersafe gels at the top show the reduction of TbERK8 mRNA levels as detected by RT-PCR after 48 h of induction (upper panel). α-tubulin was used as a control (bottom panel). The bottom graph shows the growth curve for pHAR stable parasites with (+Tet) ( ) or without (−Tet) ( ) tetracycline induction for TbERK8 silencing. The growth curve represents parasite proliferation monitored during 6 d in 3 replicates. (C) Reduction in TbPCNA levels in pHAR stable parasites after tetracycline induction to reduce TbERK8 mRNA levels. Western blot shows the levels of expression of TbPCNA after 48 h of tetracycline induction (upper). The StainFree® gel shows VSG as a loading control (bottom).

Table 2.

Summary of phosphorylated residues identified on TbPCNA immunoprecipitated from T. bruei.

	pS202	pS211	pS216
O/E-TbPCNAHA	+	−	+
EL-TbPCNAHA	+	−	+
pHAR-TbPCNAHA	−	−	−

Note. Overexpressed (O/E), Endogenous Levels (EL), after TbERK8 depletion (pHAR)

Small molecule inhibitors distinguish between active sites of HsERK8 and TbERK8

We probed the active site of TbERK8 with the inhibitor SCYX5070 to test for selectivity. This 2,4-diaminopyrimidine was a logical first choice because it kills T. brucei at submicromolar concentrations and TbERK8 (formerly annotated as Tb10.70.2070) was one of the many kinases that its immobilized version pulled down from parasite lysates.¹⁸ We tested SCYX5070 by HTRF and determined its IC₅₀ for TbERK8 to be about 1.6 mM. At a concentration of 1 mM, SCYX5070 inhibited HsERK8 activity to 37% ± 7% versus 59% ± 11% inhibition for TbERK8 which is significantly different with a p-value = 0.02 (Fig. S2). Ro318220 is an HsERK8 inhibitor with reported IC₅₀ values that ranges from 5-10 nM.¹⁹ We tested this inhibitor at 1 μM and 10 μM against TbERK8 as shown in the representative kinase assay shown in Figure 8A. This inhibitor was able to reduce the activity of TbERK8 to 49% and 27% at 1 μM and at 10 μM respectively (Fig. 8B). Eight-point dose response curves done with Ro318220 against TbERK8 were used to calculate an IC₅₀ of 1.4 μM (Fig. 8C). This IC₅₀ of Ro318220 for TbERK8 ranges between 100 -to- 300 times higher than the IC₅₀ values of this compound reported for HsERK8.

Figure 8.

Testing the HsERK8 inhibitor Ro318220 against TbERK8. (A) Kinase assay using MBP as the substrate to tests Ro318220 against TbERK8 at 1 μM and 10 μM. (B) Bar graph showing how much the activity of TbERK8 is inhibited by Ro318220 at 1 μM and 10 μM. (C) Eight-point dose curve of Ro318220 against TbERK8 in kinase assay using MBP to calculate IC₅₀. Graphs show the mean values and standard deviations for triplicate experiments.

Discussion

TbERK8 is a member of the ERK8 family of MAPKs first identified in human cells. These family members have characteristically long C-terminal extensions that are not conserved.⁷ The ERK8 MAPKs are atypical because they become fully activated by autophosphorylation instead of by an upstream MAPK kinase.¹⁴ This study reveals that TbERK8 has some biochemical properties that distinguish it from its HsERK8 homolog.

We conclude that TbERK8 phosphorylates TbPCNA in T. brucei. By using this parasite system, we also present the first example of a PCNA homolog being phosphorylated on serine and threonine residues. All of the TbPCNA residues phosphorylated by TbERK8 identified in this study were located in an insert that is predicted to form an extended backside loop (Fig. 9A). The presence of such a loop is a feature that has only been identified in the PCNA homologs of kinetoplastid parasites. The in vitro kinase assays clearly demonstrated that TbERK8 was able to phosphorylate TbPCNA_6H at the same residues that we identified from TbPCNA directly immunoprecipitated from T. brucei. Interestingly, the phosphorylation patterns of TbPCNA that was immunoprecipitated from the parasite did not exactly match those obtained from recombinant TbPCNA that was phosphorylated by recombinant TbERK8 in vitro. One possible explanation for such discrepancies is that an unidentified factor co-purifies with immunoprecipitated TbERK8 and regulates the way that this kinase phosphorylates TbPCNA. Our inability to phosphorylate TbPCNA with HsERK8 (Fig. S3) or to phosphorylate HsPCNA with TbERK8 (data not shown) and results from a previous study¹⁷ strengthen the hypothesis that this mechanism of PCNA serine/threonine phosphorylation is unique to T. brucei. However, this mechanism of PCNA phosphorylation might also be possible in other kinetoplastid parasites.

Figure 9.

Model for phosphorylated TbPCNA. (A) TbPCNA model showing that the phospohorylated residues *T₂₀₂, *S₂₁₁ and *S₂₁₆ are situated in its extended backside loop formed by its unique insertions. (B) ClustalW alignment of TbPCNA and HsPCNA with black arrow showing tyrosine 211, which is phosphorylated by receptor tyrosine kinases. Note that TbPCNA has a phenylalanine residue at the corresponding position. *denotes phospho-serine/threonine residues.

One perceived weakness in our discovery is the observation that none of the TbPCNA residues phosphorylated by TbERK8 had a proline at the +1 position. ERKs are proline-directed serine/threonine ((S/T)P) kinases in general and such activity has been demonstrated in HsERK8.¹⁹ However, the presence of a Pro at the +1 position is not an absolute substrate requirement for ERKs and other MAPKs.^20,21 We have identified only one potential site with a ((S/T) P) motif within TbPCNA (₂₆₃AEKSP₂₆₇), but no phosphorylated Ser₂₆₆ residue was detected by our mass spectrometry analysis of phosphorylated TbPCNA from either in vitro or in vivo preparations. We speculate that the residues that we identified by mass spectrometry represent TbPCNA sites that are the most accessible to TbERK8 and make up the majority of phosphorylation events. The other sites might represent ones that are phosphorylated at low frequencies below the detection limit of mass spectrometry analysis. The notion that TbERK8 can phosphorylate TbPCNA at other residues is strengthened by in vitro kinase assays using the TbPCNA_TSS triple mutant (Fig. S3).

All reported cases of HsPCNA phosphorylation occur on tyrosine residues.^22-24 The phosphorylation of HsPCNA is mediated by tyrosine kinases such as epidermal growth factor receptor (EGFR)^23,24 and the Recepteur d'Origine Nantais (RON)/c-ABL pathway.^25,26 Receptor tyrosine kinase-mediated phosphorylation of HsPCNA occurs at Y₂₁₁ which correlates with aggressive cancer either by increased DNA synthesis activity²⁴ or inhibition of mismatch repair.²⁷ Such tyrosine kinase-mediated mechanisms of phosphorylation are not possible in T. brucei because the parasite lacks genes for any tyrosine kinases^2,3 and because TbPCNA has a phenylalanine residue at the position that corresponds to Y₂₁₁ in HsPCNA (Fig. 9B).

The strengths of this study are that it reveals that the replication factor TbPCNA is the first known biological substrate for TbERK8, which is an essential MAPK in T. brucei. Further studies are needed to elucidate the biological role of serine/threonine phosphorylation in TbPCNA. Overall, very few substrates have been identified for ERK8 homologs. The only known cellular substrate of HsERK8 to date is the RNA binding protein HuR,²⁸ which is encoded by the ELAVL1 gene. HuR binds to the 3′UTR of the PDC4 tumor suppressor to protect it from miRNA 21 mediated gene silencing.^29-31 Ro318220 is a potent inhibitor of human ERK8 and several other human protein kinases.^19,32,33 The ability of Ro318220 to inhibit HsERK8 with such a low IC₅₀ in comparison to TbERK8 presents evidence that the active sites of HsERK8 and TbERK8 can be selectively inhibited. This creates a rational basis for discovering and developing small molecules that selectively inhibit TbERK8 to kill T. brucei and have fewer off-target effects in the host. TbERK8 selective inhibitors can also be used to chemically map out its biological function in T. brucei.

Materials and methods

Culturing of T. brucei

Bloodstream form parasites (Lister 427 VSG 221 strain) were incubated in 5% CO₂ at 37 °C in complete HMI-9 medium purchased from Axenia Biologix (http://www.axeniabio.com/) using the formulation of Hirumi and Hirumi.³⁴

Primers for TbERK8 and TbPCNA constructs

The coding region for wild-type TbERK8 (wt-TbERK8) was amplified with the forward primer 5′-CGCGCCAAGCTTATGT-CATCAGAAATAGAGCC-3′ and the reverse primer 5′-CTTAAGT-TTGTGCAACACACGAGAGGC-3′. The K42A point mutation (K42A-TbERK8) was made in TbERK8 cDNA using the K42A forward primer 5′-GGTTGTAGCGTTAGCGAAGATATACGAC-GC3 and reverse 5′-GCGTCGTATATCTT-CGCTAACGCTACAACC-3′ by the annealing overlapping PCR method.³⁵

The wild-type TbPCNA coding region (wt-TbPCNA) was amplified by PCR from T. brucei genomic DNA using the forward primer 5′-AAGCTTATGCTTGAGGCTCAG-GTTCTG-3′ and the reverse primer 5′-CTTAAGCTCGGCGTCGTCCACCTT-TGG-3′. Primers to generate the PIP-box from TbERK8 were 5′-CGATACTTAAGGGATCCATGTC-CCCTATACTAGGT-3′ and 5′-GCGAGA-GTCGACTCAGTGGAAAGCCG-CCACATACG-G3′. Primers to generate the PIP-box from TbAUK1 were forward primer 5′-GATCCCAGTTCCTTCTCAAGTATTATTATG-3′ and reverse primer 5′-TCGACATAATAATACTTGAGAAGGAACTGG-3′. Primers to generate the PIP-box from TbPFC19 were forward primer 5′-GATCCCAGGGCCACTTGTTTA-CCTACTACG-3′ and reverse primer 5′ -TACGGTGGCCCTGG-3′. These primers were annealed and subcloned into BamHI/SalI restriction sites of pGSTag.³⁶

Expression and purification of recombinant proteins in E. coli

wt-TbPCNA was subcloned into the BamHI/HindIII sites of pTrcHis A (Invitrogen, https://www.thermofisher.com/order/catalog/product/V36520?ICID=search-product) and expressed with a 6×His tag in BL21-DE3. The recombinant protein, wt-TbPCNA_6H, was purified by nickel agarose affinity chromatography (Cat# 88222, Lot# PL208040, Thermo Scientific, https://www.thermofisher.com/us/en/home.html) as recommended by the manufacturer.

To express the TbERK8 kinase domain without the C-terminal extension, the nucleotides from 1-1019 of its coding region were subcloned into the BamHI/EcoRI sites of pGSTag to express it as a GST fusion protein (GST-TbERK8_KD) in E. coli. The full-length TbERK8 (nucleotides 1-1329) was subcloned into BamHI/HindIII site of the pGSTag to express it as a GST fusion protein (GST-TbERK8). The pGEX-2T-HsERK8 expression plasmid was kindly provided by the laboratory of Deborah Lannigan (Vanderbilt University). E. coli expressing either GST-TbERK8, GST-TbERK8_KD, or GST-HsERK8 fusion protein were grown at 37 °C until they reached an A₆₀₀ of 1.0, then chilled on ice before inducing with 0.2 mM isopropyl β-D-1-thiogalactopyranoside at 16 °C for 12 h. The E. coli were pelleted by centrifugation at 7,000 rpm in a Beckman JLA 10.5 rotor at 10 °C for 15 minutes. Pellets were lysed by sonication and clarified by centrifugation at 15,000 rpm using a Beckman JA 25.5 rotor at 4 °C. These GST fusion proteins were purified by GSTPrep FF 16/10 affinity chromatography (Cat# 28-9365-50, Lot# 10233091, GE Healthcare, http://www.gelifesciences.com/webapp/wcs/stores/servlet/ProductDisplay?categoryId=114-38&catalogId=10101&productId=17628&storeId=11787&langI-d=-1) as recommended by the manufacturer.

Purification of TbERK8_BV from SF9 insect cells

A recombinant baculovirus encoding the full coding region of TbERK8 was constructed using the pFastBac1 baculovirus expression system (ThermoFisher Scientific, https://www.thermofisher.com/order/catalog/product/10712024). Sf9 insect cells (4 × 250 ml) were grown to a density of 2.0 × 10⁶ cells/ml and then infected with the recombinant virus at a multiplicity of infection of 3-5. After rotation at 150 rpm for 72 h at 28 °C, the cells were harvested by centrifugation, flash-frozen in liquid nitrogen, and stored at −80 °C. The frozen pellet was thawed on ice and then resuspended in 200 ml of 50 mM Tris-HCl (pH 7.5), 75 mM NaCl, 0.1% Nonidet P-40, 1 mM EDTA, 0.5 mM dithiothreitol (DTT), 1 mM phenylmethanesulfonyl fluoride, 1 mM benzamidine-HCl, 1 μg/ml leupeptin, 2 μg/ml aprotinin, and 1 μg/ml pepstatin. After 30 min on ice, the lysate was cleared by centrifugation at 35,000 rpm for 30 min in a Beckman 45 Ti rotor at 4 °C. The clarified lysate was loaded on a carboxymethyl (CM) chromatography column (Cat# 17-0719-01, GE Healthcare, http://www.gelifesciences.com/webapp/wcs/stores/servlet/Home/en/GELifeSciences-us/). Protein was eluted from the CM column with buffer A (150 mM NaCl in 20 mM Tris, pH7.4, 10% Glycerol and 1 mM EDTA). These fractions were pooled and loaded onto hydroxyapatite (HAP) chromatography resin (Cat# 130-0151, BioRad, http://www.bio-rad.com/en-us/category/cht-ceramic-hydroxyapatite-crystalline-hydroxyapatite-resins). TbERK8_BV was eluted from the HAP column in buffer B (HEPES, pH 7.4 10% glycerol, 75 mM NaCl, 75 mM phosphate ions). Fractions from the HAP column were pooled and loaded onto an S-Fast Flow column (Cat# 17-0511-01, GE Healthcare Life Sciences, http://www.gelifesciences.com/webapp/wcs/stores/servlet/Home/en/GELifeSciences-us/) and eluted with buffer A containing 250 mM NaCl.

Autophosphorylation of recombinant TbERK8_BV and GST-TbERK8_KD

We incubated 0.1 μg of purified TbERK8_BV or GST-TbERK8_KD protein in 30 μl of buffer K (30 mM Tris pH 7.4, 10 mM MgCl₂, 1 mM DTT, 5% glycerol, [1mM cold ATP for mass spectrometry assays or 10-100 μM cold ATP for ³²P incorporations assays] and 0.1 mg/ml bovine serum albumin). Ten μCi of ³²P-γ-ATP were added for detections by autoradiography in labeling experiments. The kinase reactions were incubated at 30 °C for 30 min. SDS loading buffer was added to the reaction mixtures and boiled for 1 min to stop the reaction. TbERK8 autophosphorylation by this method was detected by SDS-PAGE and autoradiography. Alternatively, TbERK8 autophosphorylation was done using 1 mM of cold ATP. In this case, TbERK8 autophosphorylation was detected by immunoblot analysis using the anti-phosphotyrosine antibody PY20 (Cat# GTX10321, Lot# 33072, GeneTex Inc. http://www.genetex.com/) at 1/500. PY20 was recognized by a horseradish peroxidase-conjugated secondary antibody to the mouse IgG2b (Cat# GTX77294, Lot# 821600258, GeneTex Inc. http://www.genetex.com/).

Kinase activity assays

Recombinant full-length TbERK8_BV purified from Sf9 insect cells (100 nmol) was incubated with 100 μmol TbPCNA_6H. For activity assays, 10 μCi of ³²P-γ-ATP (Cat# BLU502H250UC, Perkin Elmer, http://www.perkinelmer.com/product/atp-g-32p-blu002250uc?searchTerm=32p&pushBackUrl=?searchName=32p&No=0&Ns=Best%20Match) were added to 30 μL reactions that used buffer K. The samples were incubated for 5 min at 30 °C and stopped as described above. Reactions were resolved in an SDS-PAGE gel, and bands were visualized by autoradiography. TbPCNA phosphorylation was quantified with a Typhoon FLA 7000 PhosphorImager (GE Healthcare). To determine the K_m of TbERK8_BV for TbPCNA_6H, kinase assays were performed with ³²P-γ-ATP and varying concentrations of purified recombinant TbPCNA_6H ranging from 10 nM to 5 μM. Standard curves for autoradiography quantitation by PhosphorImager were made from 2-fold serial dilutions of 10 μl Ci of ³²P-γ-ATP in 100 μM cold ATP. Steady-state kinetics of TbERK8_BV were calculated by fitting data to the Michaelis-Menten curve using Prism Graphpad 5.0 (Fig. S4).

Expression of hemagglutinin tagged (HA-tagged) recombinant proteins in T. brucei

pLEW11-3HA was generated by modifying the plasmid pLE-W111,³⁷ which allows for tetracycline inducible protein expression, to include a C-terminal hemagglutinin tag (HA-tag). The coding region for wt-TbERK8, K42A-TbERK8, and wt-TbPCNA were subcloned into the HindIII and AflII restriction sites of pLEW11-3HA. Nucleofections were carried out with 1 μg to 10 μg DNA using the Lonza Nucleofector II™ program X-001 and the Nucleofector™ kit for Human T cell transfection (Cat# VPA-1002, Lonza Pharma&Biotech, http://www.lonza.com/products-services/bio-research/transfection.aspx). Stable transfectants were selected based on their resistance to phleomycin. HA-tagged recombinant proteins wt-TbERK8_HA, K42A-TbERK8_HA, and wt-TbPCNA_HA were induced with tetracycline and expressed in the T. brucei 90-13 strain as described by Wirtz et al.³⁸ The expression of recombinant proteins was confirmed by immunodetection with rabbit anti-HA (Cat# H6908, Sigma-Aldrich, http://www.sigmaaldrich.com/catalog/product/sigma/h6908?lang=en&region=US) antibody and by mass spectrometry.

pHAR construct

The TbERK8 RNAi construct from pZJM³⁹ was digested with BamHI and KpnI and subcloned into the modified pLEW11-3HA plasmid containing wt-TbPCNA. We named the resulting plasmid pHAR because of its ability to express HA tagged TbPCNA and TbERK8 RNAi cassette simultaneously. The plasmid pHAR was linearized with the restriction enzyme BsaBI, which has a unique site in the TbPCNA coding region, and 10 μg of the linearized plasmid was nucleofected into the Tb90-13 to produce a stably transfected T. brucei strain that expressed endogenous levels of HA-tagged TbPCNA and the tetracycline inducible TbERK8 RNAi cassette. TbPCNA protein levels in these parasites were monitored by immunoblot as described below.

Immunoblot analysis of HA-tagged proteins

Parasites from cultured bloodstream form T. brucei were harvested at 2 × 10⁶ cells/ml, lysed with Tb lysis buffer (1.0% Triton X-100, 10 mM Tris pH 7.5, 25 mM KCl, 150 mM NaCl, 1 mM MgCl₂, 0.2 mM EDTA, 1 mM dithiothreitol, 20% glycerol), and incubated on ice for 30 min, followed by microcentrifugation at 15,000 rpm for 5 min to produce a clarified lysate. For immunoblotting, 25 μg of clarified lysate was resolved by SDS-PAGE and transferred to a polyvinylidene difluoride membrane (PVDF). After transferring and blocking, the membrane was incubated with rabbit anti-HA (1: 2,000 dilution) for 1 h and washed 3 times for 5 min with TBST (10 mM Tris, pH 7.4, 150 mM NaCl, 0.4% Tween 20). After the third wash, horseradish peroxidase-conjugated donkey anti-rabbit IgG (1: 1,000 dilution) (Cat# 32460, Thermo-Scientific, https://www.thermofisher.com/antibody/product/Goat-anti-Rabbit-IgG-H-L-Secondary-Antibody-Polyclonal/32460) was added to the blots for 1 h. The blots were then washed again 3 times for 5 min each and examined by enhanced chemiluminescence (ECL) (Cat# RPN2232 GE Healthcare Life Sciences, http://www.gelifesciences.com/webapp/wcs/stores/servlet/productById/en/GELifeSciences-us/28980926).

TbPCNA Pull Down with GST-PIP-box Fusion Constructs–The 3 GST-PIP-box fusion proteins: TbERK8 PIP-box (GSTtbE8-PIP), TbAUK1 PIP-box (GSTtbAUK1-PIP), and TbPFC19 PIP-box (GSTtbPFC-PIP) were overexpressed in E. coli and captured on glutathione agarose beads (Cat# G-250-10) (Gold Bio Inc., https://www.goldbio.com/product/1252/glutathione-agarose-resin) following the protocol suggested by the manufacturer. Briefly, the pellets from 250 ml cultures were lysed using 10 mL of buffer 4 (50 mM Tris, 1% DOC, 0.3 M urea, pH 8, 1 mM phenylmethanesulfonyl fluoride, 1 mM benzamidine-HCL) as described by Danilevich et al.⁴⁰ and clarified by centrifugation (4 °C at 15,000 rpm) using a Beckman JA 25.5 rotor. GST beads (200 μl) were added to the clarified lysate and incubated overnight at 4 °C rotating to saturate the beads. The saturated beads were washed 5 times with PBS. The washed GST agarose beads were incubated with lysates from T. brucei that overexpressed wt-TbPCNA_HA for 2 hours at 4 °C. Beads were washed 5 times with PBS, boiled in 5x SDS loading buffer, and resolved by polyacrylamide gel electrophoresis. The loading controls were visualized by Coommassie stain or by StainFree®, (http://www.bio-rad.com/en-us/product/mini-protean-precast-gels/mini-protean-tgx-stain-free-precast-gels), which contains a trihalo compound that produces a fluorescent product when covalently crosslinked to protein tryptophan residues. These gels reduced the need to stain gels and allowed bands to be visualized with a BioRad ChemDoc MP. TbPCNA bands were detected by immunoblots using anti-HA (Cat# MAB3834, Thermo Scientific, http://www.emdmillipore.com/US/en/product/Anti-Polyhistidine-Tag-Antibody,MM_NF-MAB3834) or anti-His antibodies (Cat# H6908, Sigma-Aldrich, http://www.sigmaaldrich.com/catalog/product/sigma/h6908?lang=en&region=US). Standard curves for proteins quantified with StainFee gels and ChemDoc MP were made from 2-fold serial dilutions of TbPCNA.

Immunoprecipitation of HA-tagged proteins

Stable T. brucei transfectants that expressed either HA-tagged TbERK8 or TbPCNA were induced with tetracycline to a final concentration of 1 μg/ml for 24 hours. Extracts from induced parasites were pelleted by centrifugation and lysed with 1 ml of Immunoprecipitation (IP) lysis buffer on ice for 30 min (1% Triton X-100, 10 mM Tris pH 7.9, 25 mM KCl, 150 mM NaCl, 1 mM MgCl₂, 0.2 mM EDTA, 1 mM dithiothreitol (DTT), 10% glycerol, and 1× complete protease inhibitor mix (Cat# 11873580001, Roche, http://www.sigmaaldrich.com/catalog/product/roche/COEDTAFRO?lang=en&region=US). Lysates were incubated with 30µl anti-HA immuno-affinity beads from (Cat# 26181, Thermo Scientific, https://www.thermofisher.com/order/catalog/product/26181) and rotated overnight at 4 °C. The beads were washed 5 times with 1 ml of IP lysis buffer.

TbPCNA_6H pull down with full length TbERK8_HA

wt-TbERK8_HA prepared from T. brucei lysates was adsorbed onto 30 μl of immuno-affinity beads and washed 5 times in 1 ml of immunoprecipitation lysis buffer. Ten microliters of the anti-HA immunoaffinity beads with wt-TbERK8_HA were incubated with a final concentration of 15 ng/μl of TbPCNA_H6 in 300 μl reactions with rotating at 4 °C for 2 hours. Beads were washed 5 times with 1 ml of PBS, resolved by SDS-PAGE, followed by anti-His tag immunoblotting for detection of bound TbPCNA_6H.

Autophosphorylation and kinase activity analysis with in vivo expressed TbERK8

Ten microliters of the HA-immuno-affinity beads with either wt-TbERK8_HA or the K42A-TbERK8_HA mutant bound were tested for autophosphorylation as described above. To test the activity of the in vivo expressed proteins, we preincubated 10 microliters of the immuno-affinity beads with 10 μM ATP, 10 μCi of ³²P-γ-ATP in 30 μl of buffer K. Subsequently, 100 µmol of either TbPCNA_6H or the generic kinase substrate myelin basic protein (MBP) (Cat# 13-110, Millipore, https://www.emdmillipore.com/US/en/product/MBP,-Dephosphorylated,MM_NF-13-110) were added, and the reactions incubated at 30 °C for 30 min. Reactions were stopped by adding 5x SDS-PAGE loading buffer and boiling for 1 minute. Ten microliters of the kinase IP reaction were resolved by SDS-PAGE and examined by autoradiography. We used a mock reaction (no substrate) for each set of experiments as well as affinity beads incubated in lysates from the non-transfected T. brucei 221 wild type strains as negative controls to detect background labeling.

Peptide sequencing by mass spectrometry

Peptide sequencing and protein identification was performed using liquid chromatography-tandem mass spectrometry (LC-MS/MS). For protein identification, we excised and prepared bands from SDS-PAGE gels and subjected them to in-gel trypsin digestion for peptide identification.⁴¹

Sequencing was performed using either an LTQ-Orbitrap XL or an LTQ-Orbitrap Velos mass spectrometers (Thermo Scientific, Rockford, IL), each equipped with a 10,000 psi system nanoACUITY (Waters, City State) UPLC instrument and EZ-Spray source (Thermo Scientific). Reversed phase liquid chromatography was performed using an EZ-Spray C18 column (Thermo, ES800, PepMap, 3 μm bead size, 75 μm × 15 cm). The LC was operated at 600 nL/min flow rate for loading and 300 nL/min for peptide separation over a linear gradient for 60 min from 2% to 30% acetonitrile in 0.1% formic acid. Tandem mass spectrometry experiments using the LTQ OrbitrapXL and the LTQ Orbitrap Velos instruments both used survey scans recorded over a 350-1500 m/z range. The LTQ Orbitrap XL was operated in collision induced dissociation (CID) mode, and MS/MS CID scans were performed on the 6 most intense precursor ions, with the following parameters: a minimum of 1,000 counts precursor, isolation width 2.0 amu, and 35% normalized collision energy. The LTQ Orbitrap Velos was operated in higher energy collision dissociation (HCD) mode, and MS/MS HCD scans were performed on the 6 most intense precursor ions with the following parameters: 3,000 count threshold, isolation width 2.0 amu, and 30% normalized collision energy. Internal recalibration to a polydimethylcyclosiloxane (PCM) ion with m/z = 445.120025 was used for both MS and MS/MS scans on both instruments.²²

Mass spectrometry centroid peak lists were generated using in-house software called PAVA, and data were searched using Protein Prospector software v. 5.10.17.⁴² For protein identification, searches were performed against the curated T. brucei 972 genome in the TriTrypDB database v 8.0 (http://tritrypdb.org/tritrypdb/, June, 2014) containing 30,295 entries. This database was concatenated with a fully randomized set of 30,295 entries for estimation of false discovery rate.⁴³ Data were also searched against the SwissProt database (downloaded January 11, 2012) for identification of common contaminant proteins. Peptide sequences were matched as tryptic peptides with up to 2 missed cleavages, and carbamidomethylated cysteines as a fixed modification. Variable modifications included: oxidation of methionine, N-terminal pyroglutamate from glutamine, start methionine processing, protein N-terminal acetylation, and S/T phosphorylation. Mass accuracy tolerance was set to 20 ppm for parent and 0.6 Da fragment for CID experiments, and 20 ppm for parent and 30 ppm for fragment mass error for HCD experiments.

For reporting of protein identifications from this database search, score thresholds were selected that resulted in a protein false discovery rate of less than 1%. The specific Protein Prospector parameters were: minimum protein score of 22, minimum peptide score of 15, and maximum expectation values of 0.02 for protein and 0.05 for peptide matches. Site Localization In Peptide (SLIP) scoring⁴⁴ was used to score the site assignment for phosphorylation sites, with manual validation of all reported phosphorylation spectra.

Inhibitors

SCYX5070 was a kind gift from William Janzen and Melissa A. Porter, The University of North Carolina at Chapel Hill. Ro318220 (Cat# SC-200619) was purchased from Santa Cruz Biotechnology (http://www.scbt.com/datasheet-200619-ro-31-8220.html). Each inhibitor was pre-incubated with 100 ng of GST-TbERK8 or GST-HsERK8 to a final concentration of 1 μM at 30 °C for 30 minutes before performing a kinase assay using ³²P-γ-ATP. The kinase reaction was stopped by boiling in 5x SDS loading buffer then resolved by SDS-PAGE. Standard curves for autoradiography quantitation by PhosphorImager were made from 2-fold serial dilutions of 10 μl Ci of ³²P-γ-ATP in 100 μM cold ATP.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgments

We would like to thank Zhijian Jake Tu¹ in the Virginia Tech Department of Biochemistry for technical expertise with phylogenetic analyses. Janet Webster and Ling Chen in the Fralin Life Science Institute for critical reading.

Notes on contributors

AVL and ZBM performed and analyzed the experiments shown in Figures 1–8 except for mass spectrometry. GMK performed and analyzed mass spectrometry experiments shown in the manuscript. All authors reviewed the results and approved the final version of the manuscript. ZBM conceived and coordinated the study, wrote the paper, and analyzed the experiments.

Funding

This work was supported in part by Start-up fund NIFA139696, VT Drug Discovery Center VTCDD ##119286. MS analysis was performed in the Bio-Organic Biomedical Mass Spectrometry Resource at UCSF (A.L. Burlingame, Director) supported by GM103481.