Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

Prediction of protein–protein interactions between Theileria parva and Bos taurus based on sequence homolo ...

Prediction of protein–protein interactions between Theileria parva and Bos taurus based on... Abstract Theileria parva induces pathogenesis, characteristic of cancer cell transformation and associated with invasion, proliferation and altered gene expression of infected bovine host leucocytes. Interactions among proteins are an important basis for biological functions and underlie processes essential to pathogenesis during infection. Knowledge or prediction of host–pathogen molecular interactions may suggest mechanisms of pathogen interference and can be useful for selecting potential therapeutic targets. Using information on conserved protein interactions in other organisms (interologs), protein interactions and orthologous relationships were predicted between T. parva parasite and Bos taurus (the bovine mammalian host). Among the predicted interactions were Theileria HSP90 and glutaredoxin-like protein and bovine c-JUN, AKT1, Rac1, STAT3 and HIF-1-α proteins, observed to act as hubs connecting the predicted interactions to protein interactions within host. Bovine proteins were enriched in pathways that reflect known phenotype of Theileria infection such as induction or inhibition of apoptosis signalling, metastasis and tissue invasion, IL-10 signalling, NF-κB/IKK activation, PI-3K pathway, TGF-β signalling, modulation of immune and inflammatory responses. Support vector machine classifiers trained with the predicted interactions identified known protein interactions with 86.22% accuracy, 84.72% precision, 89.88% sensitivity and 84.39% specificity measures. Predicted interactions provide insight into Theileria- and bovine-encoded interactomes that contribute to infection, providing a candidate set for subsequent experimental studies with the possible use for defining functional annotation to uncharacterized parasite proteins. Theileria parva, protein–protein interaction, East Coast fever, host–pathogen, support vector machines Introduction Theileria parva, a protozoan parasite (phylum Apicomplexa) transmitted by an ixodid tick is the aetiological agent for East Coast fever (ECF). ECF is a lymphoproliferative disorder that culminates in disorganization and destruction of the host lymphoid system (Dobbelaere and Heussler, 1999, Bishop et al., 2004) and is prevalent across eastern, central and southern parts of Africa, where it inflicts heavy economic losses to mainly small-scale resource-poor households (Gachohi et al., 2012). An infected tick inoculates Theileria sporozoites, which rapidly invade cells of the lymphoid lineage and differentiate into multinucleate schizonts. After entry, the enveloping host cell membrane is eliminated and the schizont resides freely in the host cell cytoplasm, a perfect position to interfere with host cell signalling pathways that regulate host cell proliferation, survival and tissue invasion in a tumour-like manner (Shaw, 2003, Dobbelaere and Kuenzi, 2004). Parasite multiplication is semi-synchronized with the infected host cell cycle, and the tight association of the schizont with the host cell mitotic apparatus ensures distribution of the schizont over the daughter cells (Ahmed, Schnittger and Mehlhorn, 1999). Although host cell signal transduction pathways that are perturbed during transformation have been described, parasite molecules responsible for initiation and regulation of the transformation events are yet to be fully validated. Existing data indicate activation of signalling pathways that regulate apoptosis, proliferation, immune and inflammatory responses, metastasis and cellular adhesion (Dobbelaere et al., 1999, Heussler et al., 1999, and reviewed in Dobbelaere and Rottenberg, 2003). A number of bovine (cattle) cell kinases with key roles in regulation of T-cell activation, proliferation and transformation have been found activated in a parasite-dependent manner in transformed T cells (Galley et al., 1997, Baumgartner, Chaussepied and Moreau, 2000). In this work, inter-species protein–protein interactions were predicted between the intracellular T. parva schizont and bovine proteins based on sequence homology, and information from experimental protein–protein interactions, with the assumption that interactions in other organisms are conserved between the target species (interolog mapping). Features such as co-localization of possible interacting pairs in the cytoplasm, gene ontology (GO) annotation, enriched pathways and gene expression data were considered. In addition, sequence and structural features of the predicted interacting proteins were used in a supervised learning computational approach to check if the predicted interactions can correctly identify an independent data set of known (experimentally verified) interacting proteins. Functionally relevant interactions between T. parva and bovine host proteins were inferred with support from reviewed literature, indicating possible mechanisms of parasite manipulation of bovine cellular processes in ECF pathogenesis. Materials and methods Data sets A list of 211 candidate T. parva manipulators of host cell phenotype was selected on the basis of expression during the parasite's schizont stage, presence of transmembrane domains and a glycosyl-phosphatidylinositol anchor for localization on the cell surface, localization to the nucleus, a signal peptide or ‘PEST’ motif for secretion into host cell cytoplasm (Swan et al., 2001, Swan et al., 2003, Shiels et al., 2004, Bishop et al., 2005, Shiels et al., 2006, Schneider et al., 2007). Putative interactions between parasite proteins and bovine leucocyte cytoplasmic proteins were predicted based on query-target–template interactions of their orthologs using BIANA Interolog Prediction Server (Garcia-Garcia et al., 2010). Homology conditions were filtered as 1e-10 e-value, 50% identity, 50% query sequence coverage, 90% template sequence coverage. Sequence similarity was measured as a function of the percentage of identical residues and the percentage of their sequence being aligned (query and template coverage). An interaction between bovine protein (Bp) and T. parva protein (Tp) was eliminated if Tp was orthologous to a known interacting partner of Bp in the bovine (intra-species) protein interaction network and vice-versa. Data on intra-species interactions for both Bos taurus and T. parva were retrieved from STRING database v9.1. The predicted interaction formed a positive training data set required for supervised machine learning. Since there are no known non-interacting proteins, an equal number of protein pairs were generated from artificial protein sequences and randomly paired to yield a negative training data set. An independent test set also required for supervised learning was constructed with 600 experimentally known, randomly selected protein interactions between human and Plasmodium falciparum (Vignali et al., 2008), and 600 artificial protein pairs, forming positive and negative independent testing data sets, respectively. Python and R code were used to generate the artificial protein sequences and for pairing the sequences. Sequence redundancy was removed using the CD-HIT programme with a sequence identity threshold of 25%. Protein features and vector encoding for supervised learning Training and testing data sets were encoded using the PROFEAT online server (Rao et al., 2011) and the feature space defined by calculating amino acid and dipeptide compositions, distribution and transition of amino acid properties such as hydrophobicity, polarizability, normalized van der Waals volume, polarity, side chain volume, solvent accessibility and net charge based on AAindex (Kawashima and Kanehisa, 2000). Feature encoding captures information of protein sequence content in real values and reflects some of the structural properties that dominate protein interactions (Bradford and Westhead, 2005, Shen et al., 2007). Feature vectors were labelled ‘+1’ and ‘−1’ for positive (interacting) and negative (non-interacting) instances respectively, and values linearly scaled to the range of [−1,+1]. Vector spaces for every interacting pair were concatenated to represent their combined interaction features. Parameter optimization To select optimal parameters for C (regularization parameter) and γ (kernel type parameter) for Support Vector Machine‘s (SVM) Gaussian Radial Basis Function (RBF) kernel, a grid (factorial) search approach was adopted within a limited range (−3:12 and −3:15 for logC and logγ, respectively) using 5-fold cross validation. Parameter optimization identifies a combination of the two parameters with minimal test error, good overall accuracy and best performance of the training data. Optimal values of C and γ were determined at 8.0 and 0.0048, respectively, with prediction accuracy of 99%, corresponding to maximum accuracy at which sensitivity and specificity are nearly equal. SVM learning and evaluation SVM supervised learning was implemented using the SVM_light programme and applied to train predictive binary classifiers from feature vectors of the predicted interactions. For a set of labelled input vectors, SVM classification function maps vectors into a higher dimensional feature space and seeks a separation hyperplane with a maximum margin to divide positive and negative data instances. SVM models based on RBF kernel functions were trained and tested on an independent labelled data set to validate the use of sequence composition features. To assess performance of the trained SVM models, various threshold dependent measures were computed: sensitivity, specificity, accuracy, precision and Matthews correlation coefficient (MCC). A receiver operating characteristic (ROC) curve, a threshold independent performance measure was plotted for true-positive rate ‘tpr’ (sensitivity, y-axis) against false-positive rate ‘fpr’ (1-specificity, x-axis) for a ‘0’ cut-off point for all predictions of testing data. ‘ROCR’ CRAN package in R was used to plot ROC curves, compute area under curve and other quantitative performance measures. Functional annotation of predicted interactions GO attributes significantly enriched in the predicted interacting proteins were computed using the GO::TermFinder tool. Significance p-value (≤0.05) was considered for the enriched GO term(s) for molecular function, cellular component and biological function. The following attributes were excluded particularly for B. taurus proteins to cater for spatio-temporal constraints: ‘mitochondria’, ‘nucleotide binding’, ‘ribosome’, ‘intracellular membrane-bounded organelle’, ‘nuclear chromatin’, ‘helicase activity’, ‘nuclear binding’ and ‘proteolysis‘. Comparison to gene expression data Predicted interactions were also correlated to available gene expression data sets obtained from in vitro and in vivo studies of Theileria infection. In particular, differentially expressed bovine leucocyte genes contained in microarray data were extracted from the Gene Expression Omnibus (GEO) database (GSE44414). Results and discussion Predicted T. parva–B. taurus protein interactions In this work, 5294 interactions were inferred between 24 T. parva proteins and 2997 unique B. taurus proteins based on sequence similarity between host and parasite proteins, and known interacting proteins, using BIANA tool for information integration and network management. BIANA inserts IntAct, DIP (Database of Interacting Protein), HPRD (Human Protein Reference Database), BioGRID (Biological General Repository for Interaction Datasets), STRING (Search Tool for Retrieval of Interacting Genes/Proteins) and MINT (Molecular INTeraction) protein–protein interaction databases, which are manually curated, and hence their information is of higher quality. T. parva proteins in the predicted interactions are listed in Table 1. Functional relevance of host proteins in the predicted interactions was evaluated. Bovine proteins were enriched in pathways that reflect known pathology of Theileria infection such as induction or inhibition of apoptosis signalling, metastasis and tissue invasion, interleukin 10 (IL-10) signalling, nuclear factor-kappa light-chain enhancer of B cells/IκB kinase complex (NF-κB/IKK) activation, phosphoinositide 3-kinase (PI-3K) pathway, transforming growth factor-beta (TGF-β) signalling, modulation of immune and inflammatory responses as discussed below. Table 1. Theileria parva proteins in the predicted host–parasite protein–protein interactions Parasite protein ID . Description . Q4N8A9, Q4N0T3, Q4N841, Q4N621, Q4N8J0, Q4N6T4, Q4N6C8, Q4N4C3, Q4N3M3, Q4N2S1, Q4MZ58, Q4MYF6, Q4N7V1, Q4N2R7, Q4MYG8, Q4MYM9, Q4N527 Uncharacterized protein; ‘putative’; Q4N068, Q4N069 Cysteine proteases (peptidase C1) Q4N5N0 Falcilysin (peptidase M16) Q4MYN7 Signal peptidase Q9BH70 Glutaredoxin-like protein* P24724 HSP90 Parasite protein ID . Description . Q4N8A9, Q4N0T3, Q4N841, Q4N621, Q4N8J0, Q4N6T4, Q4N6C8, Q4N4C3, Q4N3M3, Q4N2S1, Q4MZ58, Q4MYF6, Q4N7V1, Q4N2R7, Q4MYG8, Q4MYM9, Q4N527 Uncharacterized protein; ‘putative’; Q4N068, Q4N069 Cysteine proteases (peptidase C1) Q4N5N0 Falcilysin (peptidase M16) Q4MYN7 Signal peptidase Q9BH70 Glutaredoxin-like protein* P24724 HSP90 *Ebel et al., 1997. Open in new tab Table 1. Theileria parva proteins in the predicted host–parasite protein–protein interactions Parasite protein ID . Description . Q4N8A9, Q4N0T3, Q4N841, Q4N621, Q4N8J0, Q4N6T4, Q4N6C8, Q4N4C3, Q4N3M3, Q4N2S1, Q4MZ58, Q4MYF6, Q4N7V1, Q4N2R7, Q4MYG8, Q4MYM9, Q4N527 Uncharacterized protein; ‘putative’; Q4N068, Q4N069 Cysteine proteases (peptidase C1) Q4N5N0 Falcilysin (peptidase M16) Q4MYN7 Signal peptidase Q9BH70 Glutaredoxin-like protein* P24724 HSP90 Parasite protein ID . Description . Q4N8A9, Q4N0T3, Q4N841, Q4N621, Q4N8J0, Q4N6T4, Q4N6C8, Q4N4C3, Q4N3M3, Q4N2S1, Q4MZ58, Q4MYF6, Q4N7V1, Q4N2R7, Q4MYG8, Q4MYM9, Q4N527 Uncharacterized protein; ‘putative’; Q4N068, Q4N069 Cysteine proteases (peptidase C1) Q4N5N0 Falcilysin (peptidase M16) Q4MYN7 Signal peptidase Q9BH70 Glutaredoxin-like protein* P24724 HSP90 *Ebel et al., 1997. Open in new tab Several orthologous relationships were identified in the predicted interactions to highlight conservation of gene families and biological function. One orthologous group constituted ATP-dependent RNA helicases (A1A4H6, Q29S22, Q4N797). A small number of RNA helicases are potentially secreted from the parasite into host cell cytoplasm and may contribute to transformation by enhancing expression of proto-oncogenes (Shiels et al., 2006). A second group was identified for chaperone proteins or heat shock protein 90 (HSP90) homologues (Q76LV1, Q76LV2, Q4N786), which potentially influence activity of key signalling proteins that are components of oncogenic transformation of infected cells. Theileria-encoded HSP90 released by infective schizont may participate in IKK constitutive activation in transformed cells (Heussler et al., 1999). Functions of predicted interactions Pro-apoptotic and pro-survival signalling Direct interactions were predicted between serine/threonine-specific kinase AKT1 and five T. parva proteins (Q4N802, Q4N8J0, Q4N621, P24724 and Q4N6T4). Q4N841 T. parva protein was predicted to interact with JUN proteins (JUN, JUNB and JUND). Based on GO annotations, these T. parva proteins are annotated as hypothetical polypeptides and AKT1 and c-Jun factors are involved in apoptosis signalling pathway (Dobbelaere et al., 1999, Heussler et al., 1999). JUN, JUND and JUNB transcription factors also show elevated expression in Theileria-infected B lymphocytes (Lizundia et al., 2006, Kinnaird et al., 2013). c-Jun N-terminal kinase (JNK) and its cellular substrate c-Jun play an active role in Fas-mediated cell death by regulating Fas ligand (FasL) expression, by repressing anti-apoptotic Bcl-2 (B-cell lymphoma 2)-family proteins, and by inducing the pro-apoptotic Bcl-2-family protein Bim (reviewed in Dobbelaere and Kuenzi, 2004 and Luder et al., 2009). AKT1 in PI-3K/AKT pathway promotes resistance to apoptosis and acquisition of survival mechanisms (Heussler et al., 2001). In the predicted interaction network, AKT1 and JUN proteins act as hubs to the bovine intra-PPI network, linking the putative host–parasite interactions to other signalling effectors such as mitogen-activated protein kinase 1 (MAPK1), tumour necrosis factor receptor superfamily member 1A (TNFRSF1A) transmembrane receptor and protein kinase C, Alpha (PRKCA). Tumour necrosis factor alpha (TNF-α) cytokines signal through TNF membrane receptors to induce either pro-apoptotic or pro-survival signalling (reviewed in Tretina et al., 2015). Immune and inflammatory responses Fifteen T. parva proteins were predicted to interact with various bovine proteins associated with inflammatory responses. An interaction was predicted between bovine STAT3 (signal transducer and activator of transcription 3) and Q4N841 T. parva protein, and consequently STAT3 was seen to link this host–parasite interaction to other bovine proteins. STAT3 transcription factor (Janus Kinase 2/JAK2-STAT3 pathway) is activated and phosphorylated in a parasite-dependent GM-CSF (granulocyte-macrophage colony stimulating factor) autocrine loop (Dessauge et al., 2005a) in Theileria-transformed leucocytes. In cancer cells, STAT3 has dual roles: by suppressing anti-tumour immunity, it increases tumour cell proliferation, invasion and survival, while at the same time promotes pro-oncogenic inflammatory pathways (Yu, Pardoll and Jove, 2009). JUN/JUND proteins were also annotated as participating in inflammatory responses. The activated protein 1 (AP1) complex (partly composed of c-Jun factors) is heavily dependent on JNK activation in Theileria-infected bovine leucocytes, and in turn, regulates expression of genes in response to inflammatory stimuli (Lizundia et al., 2006). Other bovine proteins associated with inflammatory responses (Heussler et al., 2001) and predicted to interact with T. parva include PLCG1 (phospholipase C gamma 1), AKT1, MYH7/2 (myosin heavy chain), Toll-like receptors (TLR2/4/6/10), IRAK1/2 (interleukin-1 receptor associated kinase) and serine/threonine-specific kinase RAF1. Association with TLR-mediated signalling as predicted, may indicate parasite's (negative) control of host immune responses rendering the infected cell refractory to further immune stimulation and facilitating pathogen survival (reviewed in Tretina et al., 2015). Bovine TNF-α, which plays a role in systemic inflammation and immune development, was predicted to interact with Q4N2S1 T. parva protein. TNF-α was linked to the maintenance of infected cell proliferation and induction of inflammatory responses in T. parva-infected B cells through c-Jun N-terminal kinase/MAPK pathways and NF-kB activation (Guergnon et al., 2003). Modulation of metabolic energetics T. parva Q4N841 protein was predicted to interact with hypoxia inducible factor (HIF-1α), a mediator of the anti-oxidant response. HIF-1α induction is parasite-dependent and leads to increased expression of hif-1-α target genes, generating a host cell hypoxic response and promoting Warburg-like glycolysis (Metheni et al., 2014, 2015). AKT activation augments HIF-1-α expression by increasing its translation in hypoxic conditions, and c-Jun was previously reported to cooperate with HIF-1-α subunit to modulate HIF-1-mediated induction of target genes (Medjkane et al., 2014). Modulation of cellular morphology and cytoskeletal organization T. parva protein Q4N3V0 was predicted to interact with bovine profilin proteins (PF2, PF3), Destrin (actin depolymerizing factor) and Villin-1 associated with actin organization in the cytoskeleton, while bovine F-actin capping proteins (CAPZB, CAZA1, CAZA2) were predicted to interact with a parasite protein Q4N0T3. Bovine molecular motors (myosin proteins MYO1A, MYOC, MYOD1, MYO10 and MYOG) were also predicted to interact with three other T. parva proteins. These motors simultaneously bind to actin and cell membrane (Nambiar, McConnell and Tyska, 2010), a potential indicator of parasite‘s manipulation of host cell cytoskeleton dynamics. Bovine tubulin beta chain proteins that function as structural cytoskeletal components in cytokinesis were also predicted to interact with six parasite proteins. Other bovine serine/threonine kinase proteins (PLK1 and P21 protein (Cdc42/Rac1)-activated kinase 1) categorized as involved in the regulation of cell morphology and actin cytoskeleton were predicted to interact with T. parva proteins Q4N841, Q4N621, Q4N8J0 and Q4MYF6. PLK1 (Polo-like kinase) is a cell mitotic kinase recruited by the schizont and required for association with host cell central spindles during parasite segregation between the daughter cells in cytokinesis (Von-Schubert et al., 2010). Similarly, bovine calmodulin (CALM) was predicted to interact with T. parva proteins. CALM (calcium binding protein) stimulates a large number of enzymes like protein kinases and phosphatases, regulating events such as cytoskeletal remodelling through cytokinesis. CALM localizes to spindle pores and spindle microtubules during mitosis and has been shown to functionally interact with actin, myosin, microtubule-associated and ras-GTPase proteins. Association with host cell cytoskeletal network, as seen with predicted interactions, ensures successful dissemination on to each daughter cell with great efficiency to maintain infection rate. Several bovine proteins in the predicted interactions were found to be involved in TGF-β signalling. TGF-β regulates cytoskeleton dynamics via transcription-dependent and transcription-independent processes (Haidar et al., 2015). TGF-β was found to trigger a parasite-dependent invasive motility programme in the host cell through cytoskeleton remodelling via activation of Rho-associated kinase (ROCK) (Chaussepied et al., 2010). SVM modules of predicted host–parasite interactions Predicted interactions were assessed by means of supervised learning algorithm on the basis of sequence composition and structural features. The RBF kernel-based SVM model achieved recall of 81.2%, precision of 84.72%, 89.88% sensitivity and 84.39% specificity. The model correctly classified a good number of interacting and non-interacting test protein pairs with an accuracy of 86.22%. The model had MCC value of 0.51. An area under receiver operation characteristic curve obtained with RBF kernel for all predictions on the test data was computed and an optimal value of 0.79 obtained (Fig. 1). Figure 1. Open in new tabDownload slide ROC curve. Performance of the SVM models based on RBF kernel function is shown in a ROC plot, sensitivity (tpr) against specificity (false-positive rate), which equals 1-specificity. A ROC plot displays the trade-off between sensitivity and specificity, and area under the curve (AUC) measures accuracy. Comparison of predicted interactions to microarray data Predicted protein interactions were correlated with schizont and bovine genes altered at RNA or protein level during infection. Gene expression profiling data obtained from microarray analysis was extracted from the GEO database. Bovine matrix metallo-protease 9 (MMP9) predicted to interact with a T. parva protein (Q4N621), and FYN, (a Src family tyrosine kinase also predicted to interact three T. parva proteins (Q4N802, Q4N841 and Q4N0T3)) have been shown to be differentially expressed and are linked to metastasis, parasite dissemination and proliferation phenotypes of Theileria-infected cells. T. parva‘s Q4N621 is a transmembrane protein, and its interaction with MMP9 possibly contributes to matrix degradation reported in Theileria-infected cells (Baylis, Megson and Hall, 1995, Adamson et al., 2000, Baumgartner, 2011). NF-kappa-B essential modulator (NEMO) and IKK-β catalytic subunit, central regulators of NF-kB signalling pathway, were identified as putative candidates for interaction with two T. parva proteins (Q4N0T3 and Q4N841). The loci encoding NEMO was previously reported as transcriptionally silent in non-infected cells, and its increased expression during infection may represent its importance in Theileria-mediated infection mechanisms (Durrani et al., 2012, Kinnaird et al., 2013). Similarly, transcriptional regulators JUN, JUNB, JUND, c-Fos and c-MYC proto-oncogenes predicted here to interact with various T. parva proteins were previously reported to have increased expression in Theileria-infected cells (Chaussepied et al., 1998). These transcriptional regulators are reversibly expressed (or down-regulated) upon drug-induced parasite death (Galley et al., 1997, Dessauge et al., 2005b, Chaussepied et al., 2010). A close correlation between high levels of transcription and protein abundance in the schizont stage was observed for P24724 T. parva proteins (Gardner et al., 2005). P24724 protein, a molecular chaperone, was predicted to interact with TLR proteins (TLR2/4/6/10), which show decreased expression in Theileria-infected cells (Kinnaird et al., 2013). Repression of genes encoding TLRs in the presence of viable parasite, and the possible interaction with parasite‘s molecular chaperones may indicate modification of perception of inflammatory stimulation, in turn enhancing survival of the infected cell. T. parva proteins P24724 and Q4N6T4 were predicted to interact with bovine Fas receptors. Expression of FasL and Fas death receptors were previously reported in T. parva-infected T cells where the normally pro-apoptotic influence of FasL ligation is suppressed (Kuenzi, Schneider and Dobbelaere, 2003). An interaction between parasite factors and Fas receptors as predicted in this study may indirectly or directly contribute to the decreased sensitivity to Fas/FasL-induced apoptosis in parasitized cells, a similar mechanism observed in bacterial secretory effectors (Caulfield and Lathem, 2014). The gene PAK1 encoding p21(Cdc42-Rac)-activated kinase 1 is significantly up-regulated in (22-fold) in Theileria-infected leucocytes (Durrani et al., 2012). PAK1 regulates actin and microtubule cytoskeleton can phosphorylate cortactin to regulate dynamics of branched actin filaments and enhances cell migration and proliferation via AKT (Huynh et al., 2010). PAK1 has been reported to operate in the pathogen-dependent activation of NF-κB (Neumann et al., 2006), and it is possible that T. parva–PAK1 interaction alters cell morphology and actin cytoskeleton organization to direct parasite-dependent NF-κB activation. TNF-α is significantly induced in Theileria-infected cells compared to uninfected cells (Durrani et al., 2012). This chronic increase in TNF-α was recently revealed to influence cell morphology, mobilization behaviour and matrix invasion of infected bovine cell through induction of MAP4K4, an evolutionary conserved kinase that controls actin cytoskeleton dynamics and cell motility (Ma and Baumgartner, 2014). Theileria infection is also associated with significant repression of bovine TLR4 at both mRNA and protein level (Durrani et al., 2012). Cellular activation via TLR4 receptor operates as an extrinsic route for NF-kB signalling that can induce apoptosis, a mechanism bypassed in the presence of parasites in infected leucocytes (Heussler et al., 2002). T. parva interaction with the receptor as predicted here, in addition to the experimental evidence of TLR4 down-regulation, may reduce the influence of TLR4-mediated pathways to enhance pro-survival and promote parasite establishment. Conclusions The predicted protein interaction network provides a glimpse into the relationship between T. parva and infected host cell, where parasite proteins impacted on a variety of pathways and targeted signalling hubs in the host, indicating that the parasite uses its protein repertoire to influence signalling and regulation host processes. For instance, JUN proteins were seen as highly connected proteins appearing in numerous signalling pathways, suggesting that the parasite takes advantage of the host network at the pathway level. Interaction between T. parva proteins and bovine CALM was predicted, and these interactions may regulate functioning of host cell CALM-centric network. T. parva (schizont) proteins potentially released into host cell cytoplasm and predicted to interact with host cell proteins, and for which no orthologs were identifiable in the bovine genome, may represent potential parasite-specific therapeutic targets. This study provides new testable hypotheses that should be explored for experimental efforts to identify T. parva–B. taurus protein–protein interactions. Author biography Everlyn Kamau is carrying out her postgraduate studies in Bioinformatics and Molecular Biology. Her current interests are in computational approaches of systems biology to study biological mechanisms that govern inflammation, infection and immunity. Future plans include doctoral studies and research on immunity in cancer. Acknowledgements The authors acknowledge Joyce Njuguna (International Livestock Research Institute) and Alan Orth (International Livestock Research Institute) for their immense technical assistance during the study. The authors also acknowledge the African Bioscience Challenge Fund (ABCF) fellowship awarded to EK, which supported part of this study. References Adamson , R. , Logan , M., Kinnaird , J. et al. . ( 2000 ) Loss of matrix metalloproteinase 9 activity in Theileria annulata—attenuated cells is at the transcriptional level and is associated with differentially expressed AP-1 species , Molecular and biochemical parasitology , 106 , 51 – 61 . Google Scholar Crossref Search ADS PubMed WorldCat Ahmed , J. S. , Schnittger , L. and Mehlhorn , H. ( 1999 ) Theileria schizonts induce fundamental alterations in their host cells , Parasitology research , 85 , 527 – 538 . Google Scholar Crossref Search ADS PubMed WorldCat Baumgartner , M. , Chaussepied , M. and Moreau , M. F. ( 2000 ) Constitutive PI3-K activity is essential for proliferation, but not survival, of Theileria parva-transformed B cells , Cellular microbiology , 2 , 329 – 339 . Google Scholar Crossref Search ADS PubMed WorldCat Baumgartner , M. ( 2011 ) Theileria annulata promotes Src kinase-dependent host cell polarization by manipulating actin dynamics in podosomes and lamellipodia , Cellular microbiology , 13 , 538 – 553 . Google Scholar Crossref Search ADS PubMed WorldCat Baylis , H. A. , Megson , A. and Hall , R. ( 1995 ) Infection with Theileria annulata induces expression of matrix metalloproteinase 9 and transcription factor AP-1 in bovine leucocytes , Molecular and biochemical parasitology , 69 , 211 – 222 . Google Scholar Crossref Search ADS PubMed WorldCat Bishop , R. , Musoke , A., Mozaria , S. et al. . ( 2004 ) Theileria: intracellular protozoan parasites of wild and domestic ruminants transmitted by ixodid ticks , Parasitology , 129 , S271 – S283 . Google Scholar Crossref Search ADS PubMed WorldCat Bishop , R. , Shah , T., Pelle , R. et al. . ( 2005 ) Analysis of the transcriptome of the protozoan Theileria parva using MPSS reveals that the majority of genes are transcriptionally active in the schizont stage , Nucleic acids research , 33 , 5503 – 5511 . Google Scholar Crossref Search ADS PubMed WorldCat Bradford , J. R. and Westhead , D. R. ( 2005 ) Improved prediction of protein–protein binding sites using a support vector machines approach , Bioinformatics (Oxford, England) , 21 , 1487 – 1494 . Google Scholar Crossref Search ADS PubMed WorldCat Caulfield , A. J. and Lathem , W. W. ( 2014 ) Disruption of Fas-Fas ligand signaling, apoptosis, and innate immunity by bacterial pathogens , PLoS pathogens , 10 ( 8 ), e1004252 . Google Scholar Crossref Search ADS PubMed WorldCat Chaussepied , M. , Lallemand , D., Moreau , M. F. et al. . ( 1998 ) Upregulation of Jun and Fos family members and permanent JNK activity lead to constitutive AP-1 activation in Theileria-transformed leukocytes , Molecular and biochemical parasitology , 94 , 215 – 226 . Google Scholar Crossref Search ADS PubMed WorldCat Chaussepied , M. , Janski , N., Baumgartner , M. et al. . ( 2010 ) TGF-b2 induction regulates invasiveness of Theileria-transformed leukocytes and disease susceptibility , PLoS pathogens , 6 , e1001197 . Google Scholar Crossref Search ADS PubMed WorldCat Dessauge , F. , Lizundia , R., Baumgartner , M. et al. . ( 2005 a) Taking the Myc is bad for Theileria , Trends in Parasitology , 21 , 377 – 385 . Google Scholar Crossref Search ADS PubMed WorldCat Dessauge , F. , Hilaly , S., Baumgartner , M. et al. . ( 2005 b) c-Myc activation by Theileria parasites promotes survival of infected B-lymphocytes , Oncogene , 24 , 1075 – 1083 . Google Scholar Crossref Search ADS PubMed WorldCat Dobbelaere , D. , Fernandez , P., Machado , J. et al. . ( 1999 ) Interference by the intracellular parasite Theileria parva with T-cell signal transduction pathways induces transformation and protection against apoptosis , Veterinary immunology and immunopathology , 72 , 95 – 100 . Google Scholar Crossref Search ADS PubMed WorldCat Dobbelaere , D. A. and Kuenzi , P. ( 2004 ) The strategies of the Theileria parasite: a new twist in host-pathogen interactions , Current opinion in immunology , 16 , 524 – 530 . Google Scholar Crossref Search ADS PubMed WorldCat Dobbelaere , D. A. and Rottenberg , S. ( 2003 ) Theileria-induced leukocyte transformation , Current opinion in microbiology , 6 , 377 – 382 . Google Scholar Crossref Search ADS PubMed WorldCat Dobbelaere , D. and Heussler , V. ( 1999 ) Transformation of leukocytes by Theileria parva and Theileria annulata , Annual review of microbiology , 53 , 1 – 42 . Google Scholar Crossref Search ADS PubMed WorldCat Durrani , Z. , Weir , W., Pillai , S. et al. . ( 2012 ) Modulation of activation-associated host cell gene expression by the apicomplexan parasite Theileria annulata , Cellular microbiology , 14 , 1434 – 1454 . Google Scholar Crossref Search ADS PubMed WorldCat Ebel , T. , Middleton , J. F. S., Frisch , A. et al. . ( 1997 ) Characterization of secretory type Theileria parva glutaredoxin homologue identified by novel screening procedure , The Journal of biological chemistry , 272 , 3042 – 3048 . Google Scholar Crossref Search ADS PubMed WorldCat Gachohi , J. , Skilton , R., Hansen , F. et al. . ( 2012 ) Epidemiology of East Coast fever (Theileria parva infection) in Kenya: past, present and the future , Parasites & Vectors , 5 , 194 . Google Scholar Crossref Search ADS PubMed WorldCat Galley , Y. , Hagens , G., Glaser , I. et al. . ( 1997 ) Jun NH2-terminal kinase is constitutively activated in T cells transformed by the intracellular parasite Theileria parva , Proc Natl Acad Sci U S A , 94 , 5119 – 5124 . Google Scholar Crossref Search ADS PubMed WorldCat Garcia-Garcia , J. , Guney , E., Aragues , R. et al. . ( 2010 ) BiANA: a software framework for compiling biological interactions and analyzing networks , BMC Bioinformatics , 11 , 56 . Google Scholar Crossref Search ADS PubMed WorldCat Gardner , M. J. , Richard , B., Trushar , S. et al. . ( 2005 ) Genome sequence of Theileria parva, a bovine pathogen that transforms lymphocytes , Science (New York, N.Y.) , 309 ( 5731 ), 134 – 137 . Google Scholar Crossref Search ADS PubMed WorldCat Guergnon , J. , Chaussepied , M., Sopp , P. et al. . ( 2003 ) A tumour necrosis factor alpha autocrine loop contributes to proliferation and nuclear factor-kappaB activation of Theileria parva transformed B cells , Cellular microbiology , 5 , 709 – 716 . Google Scholar Crossref Search ADS PubMed WorldCat Haidar , M. , Echebli , N., Ding , Y. et al. . ( 2015 ) TGF-b2 promotes transcription of COX2 and EP4 leading to a PGE2-driven auto-stimulatory loop that enhances virulence of Theileria-transformed macrophages , Infection and immunity , 83 , 1869 – 1880 . Google Scholar Crossref Search ADS PubMed WorldCat Heussler , V. T. , Machado , J. Jr, Fernandez , P. C. et al. . ( 1999 ) The intracellular parasite Theileria parva protects infected T cells from apoptosis , Proc Natl Acad Sci U S A , 96 , 7312 – 7317 . Google Scholar Crossref Search ADS PubMed WorldCat Heussler , V. T. , Kuenzi , P., Fraga , F. et al. . ( 2001 ) The Akt/PKB pathway is constitutively activated in Theileria-transformed leucocytes, but does not directly control constitutive NF-kB activation , Cellular Microbiology , 3 , 537 – 550 . Google Scholar Crossref Search ADS PubMed WorldCat Heussler , V. T. , Rottenberg , S., Schwab , R. et al. . ( 2002 ) Hijacking of host cell IKK signalosomes by the transforming parasite Theileria , Science (New York, N.Y.) , 298 , 1033 – 1036 . Google Scholar Crossref Search ADS PubMed WorldCat Huynh , N. , Liu , K. H., Baldwin , G. S. et al. . ( 2010 ) P21- activated kinase 1 stimulates colon cancer cell growth and migration/invasion via ERK- and AKT-dependent pathways , Biochimica et biophysica acta , 1803 , 1106 – 1113 . Google Scholar Crossref Search ADS PubMed WorldCat Kawashima , S. and Kanehisa , M. ( 2000 ) AAindex: amino acid index database , Nucleic acids research , 28 , 374 . Google Scholar Crossref Search ADS PubMed WorldCat Kinnaird , J. H. , Weir , W., Durrani , Z. et al. . ( 2013 ) A bovine lymphosarcoma cell line infected with Theileria annulata exhibits an irreversible reconfiguration of host cell gene expression , PLoS ONE , 8 ( 6 ), e66833 . Google Scholar Crossref Search ADS PubMed WorldCat Kuenzi , P. , Schneider , P. and Dobbelaere , D. A. ( 2003 ) Theileria parva-transformed T cells show enhanced resistance to Fas/Fas ligand-induced apoptosis , Journal of immunology , 171 , 1224 – 1231 . Google Scholar Crossref Search ADS WorldCat Lizundia , R. , Chaussepied , M., Huerre , M. et al. . ( 2006 ) c-Jun NH2-terminal kinase/c-Jun signaling promotes survival and metastasis of B lymphocytes transformed by Theileria , Cancer research , 66 , 6105 – 6110 . Google Scholar Crossref Search ADS PubMed WorldCat Luder , C. G. , Stanway , R. R., Chaussepied , M. et al. . ( 2009 ) Intracellular survival of apicomplexan parasites and host cell modification , International journal for parasitology , 39 , 163 – 173 . Google Scholar Crossref Search ADS PubMed WorldCat Medjkane , S. , Perichon , M., Marsolier , J. et al. . ( 2014 ) Theileria induces oxidative stress and HIF1-α activation that are essential for host leukocyte transformation , Oncogene , 33 , 1809 – 1817 . Google Scholar Crossref Search ADS PubMed WorldCat Metheni , M. , Echebli , N., Chaussepied , M. et al. . ( 2014 ) The level of H(2)O(2) type oxidative stress regulates virulence of Theileria transformed leukocytes , Cellular microbiology , 16 , 269 – 279 . Google Scholar Crossref Search ADS PubMed WorldCat Metheni , M. , Lombes , A., Bouillaud , F. et al. . ( 2015 ) Microreview: HIF-1a induction, proliferation and glycolysis of Theileria-infected leukocytes , Cellular microbiology , 17 , 467 – 472 . Google Scholar Crossref Search ADS PubMed WorldCat Ma , M. and Baumgartner , M. ( 2014 ) Intracellular Theileria annulata promote invasive cell motility through kinase regulation of the host actin cytoskeleton , PLoS pathogens , 10 , e1004003 . Google Scholar Crossref Search ADS PubMed WorldCat Nambiar , R. , McConnell , R. E. and Tyska , M. J. ( 2010 ) Myosin motor function. The ins and outs of actin-based membrane protrusions , Cellular and molecular life sciences : CMLS , 67 , 1239 – 1254 . Google Scholar Crossref Search ADS PubMed WorldCat Neumann , M. , Foryst-Ludwig , A., Klar , S. et al. . ( 2006 ) The PAK1 autoregulatory domain is required for interaction with NIK in Helicobacter pylori induced NF-kappaB activation , Biological chemistry , 387 , 79 – 86 . Google Scholar Crossref Search ADS PubMed WorldCat Rao , H. B. , Zhu , F., Yang , G. B. et al. . ( 2011 ) Update of PROFEAT: a web server for computing structural and physico-chemical features of proteins and peptides from amino acid sequence , Nucleic acids research , 39 , W385 – W390 . Google Scholar Crossref Search ADS PubMed WorldCat Schneider , I. , Haller , D., Kullmann , B. et al. . ( 2007 ) Identification, molecular characterization and subcellular localization of a Theileria annulata parasite protein secreted into the host cell cytoplasm , Parasitology research , 101 , 1471 – 1482 . Google Scholar Crossref Search ADS PubMed WorldCat Shaw , M. K. ( 2003 ) Cell invasion by Theileria sporozoites , Trends in parasitology , 19 , 2 – 6 . Google Scholar Crossref Search ADS PubMed WorldCat Shiels , B. R. , McKellar , S., Katzer , F. et al. . ( 2004 ) A Theileria annulata DNA binding protein located to the host cell nucleus alters the phenotype of a bovine macrophage cell line , Eukaryotic Cell , 3 , 495 – 505 . Google Scholar Crossref Search ADS PubMed WorldCat Shiels , B. , Langsley , G., Weir , W. et al. . ( 2006 ) Alteration of host cell phenotype by Theileria annulata and Theileria parva: mining for manipulators in the parasite genomes , International journal for parasitology , 36 , 9 – 21 . Google Scholar Crossref Search ADS PubMed WorldCat Shen , J. , Zhang , J., Luo , X. et al. . ( 2007 ) Predicting protein-protein interactions based only on sequences information , PNAS Biophysics , 104 ( 11 ), 4337 – 4341 . Google Scholar OpenURL Placeholder Text WorldCat Swan , D. G. , Stern , R., McKellar , S. et al. . ( 2001 ) Characterization of a cluster of genes encoding Theileria annulata AT hook DNA-binding proteins and evidence for localization to the host cell nucleus , Journal of cell science , 114 , 2747 – 2754 . Google Scholar PubMed OpenURL Placeholder Text WorldCat Swan , D. G. , Stadler , L., Okan , E. et al. . ( 2003 ) TashHN, a Theileria annulata encoded protein transported to the host nucleus displays an association with attenuation of parasite differentiation , Cellular microbiology , 5 , 947 – 956 . Google Scholar Crossref Search ADS PubMed WorldCat Tretina , K. , Gotia , H. T., Mann , D. J. et al. . ( 2015 ) Theileria-transformed bovine leukocytes have cancer hallmarks , Trends in parasitology , 31 ( 7 ), 306 – 314 . Google Scholar Crossref Search ADS PubMed WorldCat Vignali , M. , McKinlay , A., LaCount , D. J. et al. . ( 2008 ) Interaction of an atypical Plasmodium falciparum ETRAMP with human apolipoproteins , Malaria journal , 7 , 211 . Google Scholar Crossref Search ADS PubMed WorldCat Von-Schubert , C. , Xue , G., Schmuckli-Maurer , J. et al. . ( 2010 ) The transforming parasite Theileria co-opts host cell mitotic and central spindles to persist in continuously dividing cells , PLoS Biology , 8 , 43 . Google Scholar Crossref Search ADS WorldCat Yu , H. , Pardoll , D. and Jove , R. ( 2009 ) STATs in cancer inflammation and immunity: a leading role for STAT3 , Nature reviews. Cancer , 9 ( 11 ), 798 – 908 . Google Scholar Crossref Search ADS PubMed WorldCat © The Author 2016. Published by Oxford University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com © The Author 2016. Published by Oxford University Press. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png BioScience Horizons Oxford University Press

Prediction of protein–protein interactions between Theileria parva and Bos taurus based on sequence homolo ...

8 pages

Loading next page...
 
/lp/oxford-university-press/prediction-of-protein-protein-interactions-between-theileria-parva-and-U0cVnS4SUH

References (56)

Publisher
Oxford University Press
Copyright
Copyright © 2022 Oxford University Press
eISSN
1754-7431
DOI
10.1093/biohorizons/hzw006
Publisher site
See Article on Publisher Site

Abstract

Abstract Theileria parva induces pathogenesis, characteristic of cancer cell transformation and associated with invasion, proliferation and altered gene expression of infected bovine host leucocytes. Interactions among proteins are an important basis for biological functions and underlie processes essential to pathogenesis during infection. Knowledge or prediction of host–pathogen molecular interactions may suggest mechanisms of pathogen interference and can be useful for selecting potential therapeutic targets. Using information on conserved protein interactions in other organisms (interologs), protein interactions and orthologous relationships were predicted between T. parva parasite and Bos taurus (the bovine mammalian host). Among the predicted interactions were Theileria HSP90 and glutaredoxin-like protein and bovine c-JUN, AKT1, Rac1, STAT3 and HIF-1-α proteins, observed to act as hubs connecting the predicted interactions to protein interactions within host. Bovine proteins were enriched in pathways that reflect known phenotype of Theileria infection such as induction or inhibition of apoptosis signalling, metastasis and tissue invasion, IL-10 signalling, NF-κB/IKK activation, PI-3K pathway, TGF-β signalling, modulation of immune and inflammatory responses. Support vector machine classifiers trained with the predicted interactions identified known protein interactions with 86.22% accuracy, 84.72% precision, 89.88% sensitivity and 84.39% specificity measures. Predicted interactions provide insight into Theileria- and bovine-encoded interactomes that contribute to infection, providing a candidate set for subsequent experimental studies with the possible use for defining functional annotation to uncharacterized parasite proteins. Theileria parva, protein–protein interaction, East Coast fever, host–pathogen, support vector machines Introduction Theileria parva, a protozoan parasite (phylum Apicomplexa) transmitted by an ixodid tick is the aetiological agent for East Coast fever (ECF). ECF is a lymphoproliferative disorder that culminates in disorganization and destruction of the host lymphoid system (Dobbelaere and Heussler, 1999, Bishop et al., 2004) and is prevalent across eastern, central and southern parts of Africa, where it inflicts heavy economic losses to mainly small-scale resource-poor households (Gachohi et al., 2012). An infected tick inoculates Theileria sporozoites, which rapidly invade cells of the lymphoid lineage and differentiate into multinucleate schizonts. After entry, the enveloping host cell membrane is eliminated and the schizont resides freely in the host cell cytoplasm, a perfect position to interfere with host cell signalling pathways that regulate host cell proliferation, survival and tissue invasion in a tumour-like manner (Shaw, 2003, Dobbelaere and Kuenzi, 2004). Parasite multiplication is semi-synchronized with the infected host cell cycle, and the tight association of the schizont with the host cell mitotic apparatus ensures distribution of the schizont over the daughter cells (Ahmed, Schnittger and Mehlhorn, 1999). Although host cell signal transduction pathways that are perturbed during transformation have been described, parasite molecules responsible for initiation and regulation of the transformation events are yet to be fully validated. Existing data indicate activation of signalling pathways that regulate apoptosis, proliferation, immune and inflammatory responses, metastasis and cellular adhesion (Dobbelaere et al., 1999, Heussler et al., 1999, and reviewed in Dobbelaere and Rottenberg, 2003). A number of bovine (cattle) cell kinases with key roles in regulation of T-cell activation, proliferation and transformation have been found activated in a parasite-dependent manner in transformed T cells (Galley et al., 1997, Baumgartner, Chaussepied and Moreau, 2000). In this work, inter-species protein–protein interactions were predicted between the intracellular T. parva schizont and bovine proteins based on sequence homology, and information from experimental protein–protein interactions, with the assumption that interactions in other organisms are conserved between the target species (interolog mapping). Features such as co-localization of possible interacting pairs in the cytoplasm, gene ontology (GO) annotation, enriched pathways and gene expression data were considered. In addition, sequence and structural features of the predicted interacting proteins were used in a supervised learning computational approach to check if the predicted interactions can correctly identify an independent data set of known (experimentally verified) interacting proteins. Functionally relevant interactions between T. parva and bovine host proteins were inferred with support from reviewed literature, indicating possible mechanisms of parasite manipulation of bovine cellular processes in ECF pathogenesis. Materials and methods Data sets A list of 211 candidate T. parva manipulators of host cell phenotype was selected on the basis of expression during the parasite's schizont stage, presence of transmembrane domains and a glycosyl-phosphatidylinositol anchor for localization on the cell surface, localization to the nucleus, a signal peptide or ‘PEST’ motif for secretion into host cell cytoplasm (Swan et al., 2001, Swan et al., 2003, Shiels et al., 2004, Bishop et al., 2005, Shiels et al., 2006, Schneider et al., 2007). Putative interactions between parasite proteins and bovine leucocyte cytoplasmic proteins were predicted based on query-target–template interactions of their orthologs using BIANA Interolog Prediction Server (Garcia-Garcia et al., 2010). Homology conditions were filtered as 1e-10 e-value, 50% identity, 50% query sequence coverage, 90% template sequence coverage. Sequence similarity was measured as a function of the percentage of identical residues and the percentage of their sequence being aligned (query and template coverage). An interaction between bovine protein (Bp) and T. parva protein (Tp) was eliminated if Tp was orthologous to a known interacting partner of Bp in the bovine (intra-species) protein interaction network and vice-versa. Data on intra-species interactions for both Bos taurus and T. parva were retrieved from STRING database v9.1. The predicted interaction formed a positive training data set required for supervised machine learning. Since there are no known non-interacting proteins, an equal number of protein pairs were generated from artificial protein sequences and randomly paired to yield a negative training data set. An independent test set also required for supervised learning was constructed with 600 experimentally known, randomly selected protein interactions between human and Plasmodium falciparum (Vignali et al., 2008), and 600 artificial protein pairs, forming positive and negative independent testing data sets, respectively. Python and R code were used to generate the artificial protein sequences and for pairing the sequences. Sequence redundancy was removed using the CD-HIT programme with a sequence identity threshold of 25%. Protein features and vector encoding for supervised learning Training and testing data sets were encoded using the PROFEAT online server (Rao et al., 2011) and the feature space defined by calculating amino acid and dipeptide compositions, distribution and transition of amino acid properties such as hydrophobicity, polarizability, normalized van der Waals volume, polarity, side chain volume, solvent accessibility and net charge based on AAindex (Kawashima and Kanehisa, 2000). Feature encoding captures information of protein sequence content in real values and reflects some of the structural properties that dominate protein interactions (Bradford and Westhead, 2005, Shen et al., 2007). Feature vectors were labelled ‘+1’ and ‘−1’ for positive (interacting) and negative (non-interacting) instances respectively, and values linearly scaled to the range of [−1,+1]. Vector spaces for every interacting pair were concatenated to represent their combined interaction features. Parameter optimization To select optimal parameters for C (regularization parameter) and γ (kernel type parameter) for Support Vector Machine‘s (SVM) Gaussian Radial Basis Function (RBF) kernel, a grid (factorial) search approach was adopted within a limited range (−3:12 and −3:15 for logC and logγ, respectively) using 5-fold cross validation. Parameter optimization identifies a combination of the two parameters with minimal test error, good overall accuracy and best performance of the training data. Optimal values of C and γ were determined at 8.0 and 0.0048, respectively, with prediction accuracy of 99%, corresponding to maximum accuracy at which sensitivity and specificity are nearly equal. SVM learning and evaluation SVM supervised learning was implemented using the SVM_light programme and applied to train predictive binary classifiers from feature vectors of the predicted interactions. For a set of labelled input vectors, SVM classification function maps vectors into a higher dimensional feature space and seeks a separation hyperplane with a maximum margin to divide positive and negative data instances. SVM models based on RBF kernel functions were trained and tested on an independent labelled data set to validate the use of sequence composition features. To assess performance of the trained SVM models, various threshold dependent measures were computed: sensitivity, specificity, accuracy, precision and Matthews correlation coefficient (MCC). A receiver operating characteristic (ROC) curve, a threshold independent performance measure was plotted for true-positive rate ‘tpr’ (sensitivity, y-axis) against false-positive rate ‘fpr’ (1-specificity, x-axis) for a ‘0’ cut-off point for all predictions of testing data. ‘ROCR’ CRAN package in R was used to plot ROC curves, compute area under curve and other quantitative performance measures. Functional annotation of predicted interactions GO attributes significantly enriched in the predicted interacting proteins were computed using the GO::TermFinder tool. Significance p-value (≤0.05) was considered for the enriched GO term(s) for molecular function, cellular component and biological function. The following attributes were excluded particularly for B. taurus proteins to cater for spatio-temporal constraints: ‘mitochondria’, ‘nucleotide binding’, ‘ribosome’, ‘intracellular membrane-bounded organelle’, ‘nuclear chromatin’, ‘helicase activity’, ‘nuclear binding’ and ‘proteolysis‘. Comparison to gene expression data Predicted interactions were also correlated to available gene expression data sets obtained from in vitro and in vivo studies of Theileria infection. In particular, differentially expressed bovine leucocyte genes contained in microarray data were extracted from the Gene Expression Omnibus (GEO) database (GSE44414). Results and discussion Predicted T. parva–B. taurus protein interactions In this work, 5294 interactions were inferred between 24 T. parva proteins and 2997 unique B. taurus proteins based on sequence similarity between host and parasite proteins, and known interacting proteins, using BIANA tool for information integration and network management. BIANA inserts IntAct, DIP (Database of Interacting Protein), HPRD (Human Protein Reference Database), BioGRID (Biological General Repository for Interaction Datasets), STRING (Search Tool for Retrieval of Interacting Genes/Proteins) and MINT (Molecular INTeraction) protein–protein interaction databases, which are manually curated, and hence their information is of higher quality. T. parva proteins in the predicted interactions are listed in Table 1. Functional relevance of host proteins in the predicted interactions was evaluated. Bovine proteins were enriched in pathways that reflect known pathology of Theileria infection such as induction or inhibition of apoptosis signalling, metastasis and tissue invasion, interleukin 10 (IL-10) signalling, nuclear factor-kappa light-chain enhancer of B cells/IκB kinase complex (NF-κB/IKK) activation, phosphoinositide 3-kinase (PI-3K) pathway, transforming growth factor-beta (TGF-β) signalling, modulation of immune and inflammatory responses as discussed below. Table 1. Theileria parva proteins in the predicted host–parasite protein–protein interactions Parasite protein ID . Description . Q4N8A9, Q4N0T3, Q4N841, Q4N621, Q4N8J0, Q4N6T4, Q4N6C8, Q4N4C3, Q4N3M3, Q4N2S1, Q4MZ58, Q4MYF6, Q4N7V1, Q4N2R7, Q4MYG8, Q4MYM9, Q4N527 Uncharacterized protein; ‘putative’; Q4N068, Q4N069 Cysteine proteases (peptidase C1) Q4N5N0 Falcilysin (peptidase M16) Q4MYN7 Signal peptidase Q9BH70 Glutaredoxin-like protein* P24724 HSP90 Parasite protein ID . Description . Q4N8A9, Q4N0T3, Q4N841, Q4N621, Q4N8J0, Q4N6T4, Q4N6C8, Q4N4C3, Q4N3M3, Q4N2S1, Q4MZ58, Q4MYF6, Q4N7V1, Q4N2R7, Q4MYG8, Q4MYM9, Q4N527 Uncharacterized protein; ‘putative’; Q4N068, Q4N069 Cysteine proteases (peptidase C1) Q4N5N0 Falcilysin (peptidase M16) Q4MYN7 Signal peptidase Q9BH70 Glutaredoxin-like protein* P24724 HSP90 *Ebel et al., 1997. Open in new tab Table 1. Theileria parva proteins in the predicted host–parasite protein–protein interactions Parasite protein ID . Description . Q4N8A9, Q4N0T3, Q4N841, Q4N621, Q4N8J0, Q4N6T4, Q4N6C8, Q4N4C3, Q4N3M3, Q4N2S1, Q4MZ58, Q4MYF6, Q4N7V1, Q4N2R7, Q4MYG8, Q4MYM9, Q4N527 Uncharacterized protein; ‘putative’; Q4N068, Q4N069 Cysteine proteases (peptidase C1) Q4N5N0 Falcilysin (peptidase M16) Q4MYN7 Signal peptidase Q9BH70 Glutaredoxin-like protein* P24724 HSP90 Parasite protein ID . Description . Q4N8A9, Q4N0T3, Q4N841, Q4N621, Q4N8J0, Q4N6T4, Q4N6C8, Q4N4C3, Q4N3M3, Q4N2S1, Q4MZ58, Q4MYF6, Q4N7V1, Q4N2R7, Q4MYG8, Q4MYM9, Q4N527 Uncharacterized protein; ‘putative’; Q4N068, Q4N069 Cysteine proteases (peptidase C1) Q4N5N0 Falcilysin (peptidase M16) Q4MYN7 Signal peptidase Q9BH70 Glutaredoxin-like protein* P24724 HSP90 *Ebel et al., 1997. Open in new tab Several orthologous relationships were identified in the predicted interactions to highlight conservation of gene families and biological function. One orthologous group constituted ATP-dependent RNA helicases (A1A4H6, Q29S22, Q4N797). A small number of RNA helicases are potentially secreted from the parasite into host cell cytoplasm and may contribute to transformation by enhancing expression of proto-oncogenes (Shiels et al., 2006). A second group was identified for chaperone proteins or heat shock protein 90 (HSP90) homologues (Q76LV1, Q76LV2, Q4N786), which potentially influence activity of key signalling proteins that are components of oncogenic transformation of infected cells. Theileria-encoded HSP90 released by infective schizont may participate in IKK constitutive activation in transformed cells (Heussler et al., 1999). Functions of predicted interactions Pro-apoptotic and pro-survival signalling Direct interactions were predicted between serine/threonine-specific kinase AKT1 and five T. parva proteins (Q4N802, Q4N8J0, Q4N621, P24724 and Q4N6T4). Q4N841 T. parva protein was predicted to interact with JUN proteins (JUN, JUNB and JUND). Based on GO annotations, these T. parva proteins are annotated as hypothetical polypeptides and AKT1 and c-Jun factors are involved in apoptosis signalling pathway (Dobbelaere et al., 1999, Heussler et al., 1999). JUN, JUND and JUNB transcription factors also show elevated expression in Theileria-infected B lymphocytes (Lizundia et al., 2006, Kinnaird et al., 2013). c-Jun N-terminal kinase (JNK) and its cellular substrate c-Jun play an active role in Fas-mediated cell death by regulating Fas ligand (FasL) expression, by repressing anti-apoptotic Bcl-2 (B-cell lymphoma 2)-family proteins, and by inducing the pro-apoptotic Bcl-2-family protein Bim (reviewed in Dobbelaere and Kuenzi, 2004 and Luder et al., 2009). AKT1 in PI-3K/AKT pathway promotes resistance to apoptosis and acquisition of survival mechanisms (Heussler et al., 2001). In the predicted interaction network, AKT1 and JUN proteins act as hubs to the bovine intra-PPI network, linking the putative host–parasite interactions to other signalling effectors such as mitogen-activated protein kinase 1 (MAPK1), tumour necrosis factor receptor superfamily member 1A (TNFRSF1A) transmembrane receptor and protein kinase C, Alpha (PRKCA). Tumour necrosis factor alpha (TNF-α) cytokines signal through TNF membrane receptors to induce either pro-apoptotic or pro-survival signalling (reviewed in Tretina et al., 2015). Immune and inflammatory responses Fifteen T. parva proteins were predicted to interact with various bovine proteins associated with inflammatory responses. An interaction was predicted between bovine STAT3 (signal transducer and activator of transcription 3) and Q4N841 T. parva protein, and consequently STAT3 was seen to link this host–parasite interaction to other bovine proteins. STAT3 transcription factor (Janus Kinase 2/JAK2-STAT3 pathway) is activated and phosphorylated in a parasite-dependent GM-CSF (granulocyte-macrophage colony stimulating factor) autocrine loop (Dessauge et al., 2005a) in Theileria-transformed leucocytes. In cancer cells, STAT3 has dual roles: by suppressing anti-tumour immunity, it increases tumour cell proliferation, invasion and survival, while at the same time promotes pro-oncogenic inflammatory pathways (Yu, Pardoll and Jove, 2009). JUN/JUND proteins were also annotated as participating in inflammatory responses. The activated protein 1 (AP1) complex (partly composed of c-Jun factors) is heavily dependent on JNK activation in Theileria-infected bovine leucocytes, and in turn, regulates expression of genes in response to inflammatory stimuli (Lizundia et al., 2006). Other bovine proteins associated with inflammatory responses (Heussler et al., 2001) and predicted to interact with T. parva include PLCG1 (phospholipase C gamma 1), AKT1, MYH7/2 (myosin heavy chain), Toll-like receptors (TLR2/4/6/10), IRAK1/2 (interleukin-1 receptor associated kinase) and serine/threonine-specific kinase RAF1. Association with TLR-mediated signalling as predicted, may indicate parasite's (negative) control of host immune responses rendering the infected cell refractory to further immune stimulation and facilitating pathogen survival (reviewed in Tretina et al., 2015). Bovine TNF-α, which plays a role in systemic inflammation and immune development, was predicted to interact with Q4N2S1 T. parva protein. TNF-α was linked to the maintenance of infected cell proliferation and induction of inflammatory responses in T. parva-infected B cells through c-Jun N-terminal kinase/MAPK pathways and NF-kB activation (Guergnon et al., 2003). Modulation of metabolic energetics T. parva Q4N841 protein was predicted to interact with hypoxia inducible factor (HIF-1α), a mediator of the anti-oxidant response. HIF-1α induction is parasite-dependent and leads to increased expression of hif-1-α target genes, generating a host cell hypoxic response and promoting Warburg-like glycolysis (Metheni et al., 2014, 2015). AKT activation augments HIF-1-α expression by increasing its translation in hypoxic conditions, and c-Jun was previously reported to cooperate with HIF-1-α subunit to modulate HIF-1-mediated induction of target genes (Medjkane et al., 2014). Modulation of cellular morphology and cytoskeletal organization T. parva protein Q4N3V0 was predicted to interact with bovine profilin proteins (PF2, PF3), Destrin (actin depolymerizing factor) and Villin-1 associated with actin organization in the cytoskeleton, while bovine F-actin capping proteins (CAPZB, CAZA1, CAZA2) were predicted to interact with a parasite protein Q4N0T3. Bovine molecular motors (myosin proteins MYO1A, MYOC, MYOD1, MYO10 and MYOG) were also predicted to interact with three other T. parva proteins. These motors simultaneously bind to actin and cell membrane (Nambiar, McConnell and Tyska, 2010), a potential indicator of parasite‘s manipulation of host cell cytoskeleton dynamics. Bovine tubulin beta chain proteins that function as structural cytoskeletal components in cytokinesis were also predicted to interact with six parasite proteins. Other bovine serine/threonine kinase proteins (PLK1 and P21 protein (Cdc42/Rac1)-activated kinase 1) categorized as involved in the regulation of cell morphology and actin cytoskeleton were predicted to interact with T. parva proteins Q4N841, Q4N621, Q4N8J0 and Q4MYF6. PLK1 (Polo-like kinase) is a cell mitotic kinase recruited by the schizont and required for association with host cell central spindles during parasite segregation between the daughter cells in cytokinesis (Von-Schubert et al., 2010). Similarly, bovine calmodulin (CALM) was predicted to interact with T. parva proteins. CALM (calcium binding protein) stimulates a large number of enzymes like protein kinases and phosphatases, regulating events such as cytoskeletal remodelling through cytokinesis. CALM localizes to spindle pores and spindle microtubules during mitosis and has been shown to functionally interact with actin, myosin, microtubule-associated and ras-GTPase proteins. Association with host cell cytoskeletal network, as seen with predicted interactions, ensures successful dissemination on to each daughter cell with great efficiency to maintain infection rate. Several bovine proteins in the predicted interactions were found to be involved in TGF-β signalling. TGF-β regulates cytoskeleton dynamics via transcription-dependent and transcription-independent processes (Haidar et al., 2015). TGF-β was found to trigger a parasite-dependent invasive motility programme in the host cell through cytoskeleton remodelling via activation of Rho-associated kinase (ROCK) (Chaussepied et al., 2010). SVM modules of predicted host–parasite interactions Predicted interactions were assessed by means of supervised learning algorithm on the basis of sequence composition and structural features. The RBF kernel-based SVM model achieved recall of 81.2%, precision of 84.72%, 89.88% sensitivity and 84.39% specificity. The model correctly classified a good number of interacting and non-interacting test protein pairs with an accuracy of 86.22%. The model had MCC value of 0.51. An area under receiver operation characteristic curve obtained with RBF kernel for all predictions on the test data was computed and an optimal value of 0.79 obtained (Fig. 1). Figure 1. Open in new tabDownload slide ROC curve. Performance of the SVM models based on RBF kernel function is shown in a ROC plot, sensitivity (tpr) against specificity (false-positive rate), which equals 1-specificity. A ROC plot displays the trade-off between sensitivity and specificity, and area under the curve (AUC) measures accuracy. Comparison of predicted interactions to microarray data Predicted protein interactions were correlated with schizont and bovine genes altered at RNA or protein level during infection. Gene expression profiling data obtained from microarray analysis was extracted from the GEO database. Bovine matrix metallo-protease 9 (MMP9) predicted to interact with a T. parva protein (Q4N621), and FYN, (a Src family tyrosine kinase also predicted to interact three T. parva proteins (Q4N802, Q4N841 and Q4N0T3)) have been shown to be differentially expressed and are linked to metastasis, parasite dissemination and proliferation phenotypes of Theileria-infected cells. T. parva‘s Q4N621 is a transmembrane protein, and its interaction with MMP9 possibly contributes to matrix degradation reported in Theileria-infected cells (Baylis, Megson and Hall, 1995, Adamson et al., 2000, Baumgartner, 2011). NF-kappa-B essential modulator (NEMO) and IKK-β catalytic subunit, central regulators of NF-kB signalling pathway, were identified as putative candidates for interaction with two T. parva proteins (Q4N0T3 and Q4N841). The loci encoding NEMO was previously reported as transcriptionally silent in non-infected cells, and its increased expression during infection may represent its importance in Theileria-mediated infection mechanisms (Durrani et al., 2012, Kinnaird et al., 2013). Similarly, transcriptional regulators JUN, JUNB, JUND, c-Fos and c-MYC proto-oncogenes predicted here to interact with various T. parva proteins were previously reported to have increased expression in Theileria-infected cells (Chaussepied et al., 1998). These transcriptional regulators are reversibly expressed (or down-regulated) upon drug-induced parasite death (Galley et al., 1997, Dessauge et al., 2005b, Chaussepied et al., 2010). A close correlation between high levels of transcription and protein abundance in the schizont stage was observed for P24724 T. parva proteins (Gardner et al., 2005). P24724 protein, a molecular chaperone, was predicted to interact with TLR proteins (TLR2/4/6/10), which show decreased expression in Theileria-infected cells (Kinnaird et al., 2013). Repression of genes encoding TLRs in the presence of viable parasite, and the possible interaction with parasite‘s molecular chaperones may indicate modification of perception of inflammatory stimulation, in turn enhancing survival of the infected cell. T. parva proteins P24724 and Q4N6T4 were predicted to interact with bovine Fas receptors. Expression of FasL and Fas death receptors were previously reported in T. parva-infected T cells where the normally pro-apoptotic influence of FasL ligation is suppressed (Kuenzi, Schneider and Dobbelaere, 2003). An interaction between parasite factors and Fas receptors as predicted in this study may indirectly or directly contribute to the decreased sensitivity to Fas/FasL-induced apoptosis in parasitized cells, a similar mechanism observed in bacterial secretory effectors (Caulfield and Lathem, 2014). The gene PAK1 encoding p21(Cdc42-Rac)-activated kinase 1 is significantly up-regulated in (22-fold) in Theileria-infected leucocytes (Durrani et al., 2012). PAK1 regulates actin and microtubule cytoskeleton can phosphorylate cortactin to regulate dynamics of branched actin filaments and enhances cell migration and proliferation via AKT (Huynh et al., 2010). PAK1 has been reported to operate in the pathogen-dependent activation of NF-κB (Neumann et al., 2006), and it is possible that T. parva–PAK1 interaction alters cell morphology and actin cytoskeleton organization to direct parasite-dependent NF-κB activation. TNF-α is significantly induced in Theileria-infected cells compared to uninfected cells (Durrani et al., 2012). This chronic increase in TNF-α was recently revealed to influence cell morphology, mobilization behaviour and matrix invasion of infected bovine cell through induction of MAP4K4, an evolutionary conserved kinase that controls actin cytoskeleton dynamics and cell motility (Ma and Baumgartner, 2014). Theileria infection is also associated with significant repression of bovine TLR4 at both mRNA and protein level (Durrani et al., 2012). Cellular activation via TLR4 receptor operates as an extrinsic route for NF-kB signalling that can induce apoptosis, a mechanism bypassed in the presence of parasites in infected leucocytes (Heussler et al., 2002). T. parva interaction with the receptor as predicted here, in addition to the experimental evidence of TLR4 down-regulation, may reduce the influence of TLR4-mediated pathways to enhance pro-survival and promote parasite establishment. Conclusions The predicted protein interaction network provides a glimpse into the relationship between T. parva and infected host cell, where parasite proteins impacted on a variety of pathways and targeted signalling hubs in the host, indicating that the parasite uses its protein repertoire to influence signalling and regulation host processes. For instance, JUN proteins were seen as highly connected proteins appearing in numerous signalling pathways, suggesting that the parasite takes advantage of the host network at the pathway level. Interaction between T. parva proteins and bovine CALM was predicted, and these interactions may regulate functioning of host cell CALM-centric network. T. parva (schizont) proteins potentially released into host cell cytoplasm and predicted to interact with host cell proteins, and for which no orthologs were identifiable in the bovine genome, may represent potential parasite-specific therapeutic targets. This study provides new testable hypotheses that should be explored for experimental efforts to identify T. parva–B. taurus protein–protein interactions. Author biography Everlyn Kamau is carrying out her postgraduate studies in Bioinformatics and Molecular Biology. Her current interests are in computational approaches of systems biology to study biological mechanisms that govern inflammation, infection and immunity. Future plans include doctoral studies and research on immunity in cancer. Acknowledgements The authors acknowledge Joyce Njuguna (International Livestock Research Institute) and Alan Orth (International Livestock Research Institute) for their immense technical assistance during the study. The authors also acknowledge the African Bioscience Challenge Fund (ABCF) fellowship awarded to EK, which supported part of this study. References Adamson , R. , Logan , M., Kinnaird , J. et al. . ( 2000 ) Loss of matrix metalloproteinase 9 activity in Theileria annulata—attenuated cells is at the transcriptional level and is associated with differentially expressed AP-1 species , Molecular and biochemical parasitology , 106 , 51 – 61 . Google Scholar Crossref Search ADS PubMed WorldCat Ahmed , J. S. , Schnittger , L. and Mehlhorn , H. ( 1999 ) Theileria schizonts induce fundamental alterations in their host cells , Parasitology research , 85 , 527 – 538 . Google Scholar Crossref Search ADS PubMed WorldCat Baumgartner , M. , Chaussepied , M. and Moreau , M. F. ( 2000 ) Constitutive PI3-K activity is essential for proliferation, but not survival, of Theileria parva-transformed B cells , Cellular microbiology , 2 , 329 – 339 . Google Scholar Crossref Search ADS PubMed WorldCat Baumgartner , M. ( 2011 ) Theileria annulata promotes Src kinase-dependent host cell polarization by manipulating actin dynamics in podosomes and lamellipodia , Cellular microbiology , 13 , 538 – 553 . Google Scholar Crossref Search ADS PubMed WorldCat Baylis , H. A. , Megson , A. and Hall , R. ( 1995 ) Infection with Theileria annulata induces expression of matrix metalloproteinase 9 and transcription factor AP-1 in bovine leucocytes , Molecular and biochemical parasitology , 69 , 211 – 222 . Google Scholar Crossref Search ADS PubMed WorldCat Bishop , R. , Musoke , A., Mozaria , S. et al. . ( 2004 ) Theileria: intracellular protozoan parasites of wild and domestic ruminants transmitted by ixodid ticks , Parasitology , 129 , S271 – S283 . Google Scholar Crossref Search ADS PubMed WorldCat Bishop , R. , Shah , T., Pelle , R. et al. . ( 2005 ) Analysis of the transcriptome of the protozoan Theileria parva using MPSS reveals that the majority of genes are transcriptionally active in the schizont stage , Nucleic acids research , 33 , 5503 – 5511 . Google Scholar Crossref Search ADS PubMed WorldCat Bradford , J. R. and Westhead , D. R. ( 2005 ) Improved prediction of protein–protein binding sites using a support vector machines approach , Bioinformatics (Oxford, England) , 21 , 1487 – 1494 . Google Scholar Crossref Search ADS PubMed WorldCat Caulfield , A. J. and Lathem , W. W. ( 2014 ) Disruption of Fas-Fas ligand signaling, apoptosis, and innate immunity by bacterial pathogens , PLoS pathogens , 10 ( 8 ), e1004252 . Google Scholar Crossref Search ADS PubMed WorldCat Chaussepied , M. , Lallemand , D., Moreau , M. F. et al. . ( 1998 ) Upregulation of Jun and Fos family members and permanent JNK activity lead to constitutive AP-1 activation in Theileria-transformed leukocytes , Molecular and biochemical parasitology , 94 , 215 – 226 . Google Scholar Crossref Search ADS PubMed WorldCat Chaussepied , M. , Janski , N., Baumgartner , M. et al. . ( 2010 ) TGF-b2 induction regulates invasiveness of Theileria-transformed leukocytes and disease susceptibility , PLoS pathogens , 6 , e1001197 . Google Scholar Crossref Search ADS PubMed WorldCat Dessauge , F. , Lizundia , R., Baumgartner , M. et al. . ( 2005 a) Taking the Myc is bad for Theileria , Trends in Parasitology , 21 , 377 – 385 . Google Scholar Crossref Search ADS PubMed WorldCat Dessauge , F. , Hilaly , S., Baumgartner , M. et al. . ( 2005 b) c-Myc activation by Theileria parasites promotes survival of infected B-lymphocytes , Oncogene , 24 , 1075 – 1083 . Google Scholar Crossref Search ADS PubMed WorldCat Dobbelaere , D. , Fernandez , P., Machado , J. et al. . ( 1999 ) Interference by the intracellular parasite Theileria parva with T-cell signal transduction pathways induces transformation and protection against apoptosis , Veterinary immunology and immunopathology , 72 , 95 – 100 . Google Scholar Crossref Search ADS PubMed WorldCat Dobbelaere , D. A. and Kuenzi , P. ( 2004 ) The strategies of the Theileria parasite: a new twist in host-pathogen interactions , Current opinion in immunology , 16 , 524 – 530 . Google Scholar Crossref Search ADS PubMed WorldCat Dobbelaere , D. A. and Rottenberg , S. ( 2003 ) Theileria-induced leukocyte transformation , Current opinion in microbiology , 6 , 377 – 382 . Google Scholar Crossref Search ADS PubMed WorldCat Dobbelaere , D. and Heussler , V. ( 1999 ) Transformation of leukocytes by Theileria parva and Theileria annulata , Annual review of microbiology , 53 , 1 – 42 . Google Scholar Crossref Search ADS PubMed WorldCat Durrani , Z. , Weir , W., Pillai , S. et al. . ( 2012 ) Modulation of activation-associated host cell gene expression by the apicomplexan parasite Theileria annulata , Cellular microbiology , 14 , 1434 – 1454 . Google Scholar Crossref Search ADS PubMed WorldCat Ebel , T. , Middleton , J. F. S., Frisch , A. et al. . ( 1997 ) Characterization of secretory type Theileria parva glutaredoxin homologue identified by novel screening procedure , The Journal of biological chemistry , 272 , 3042 – 3048 . Google Scholar Crossref Search ADS PubMed WorldCat Gachohi , J. , Skilton , R., Hansen , F. et al. . ( 2012 ) Epidemiology of East Coast fever (Theileria parva infection) in Kenya: past, present and the future , Parasites & Vectors , 5 , 194 . Google Scholar Crossref Search ADS PubMed WorldCat Galley , Y. , Hagens , G., Glaser , I. et al. . ( 1997 ) Jun NH2-terminal kinase is constitutively activated in T cells transformed by the intracellular parasite Theileria parva , Proc Natl Acad Sci U S A , 94 , 5119 – 5124 . Google Scholar Crossref Search ADS PubMed WorldCat Garcia-Garcia , J. , Guney , E., Aragues , R. et al. . ( 2010 ) BiANA: a software framework for compiling biological interactions and analyzing networks , BMC Bioinformatics , 11 , 56 . Google Scholar Crossref Search ADS PubMed WorldCat Gardner , M. J. , Richard , B., Trushar , S. et al. . ( 2005 ) Genome sequence of Theileria parva, a bovine pathogen that transforms lymphocytes , Science (New York, N.Y.) , 309 ( 5731 ), 134 – 137 . Google Scholar Crossref Search ADS PubMed WorldCat Guergnon , J. , Chaussepied , M., Sopp , P. et al. . ( 2003 ) A tumour necrosis factor alpha autocrine loop contributes to proliferation and nuclear factor-kappaB activation of Theileria parva transformed B cells , Cellular microbiology , 5 , 709 – 716 . Google Scholar Crossref Search ADS PubMed WorldCat Haidar , M. , Echebli , N., Ding , Y. et al. . ( 2015 ) TGF-b2 promotes transcription of COX2 and EP4 leading to a PGE2-driven auto-stimulatory loop that enhances virulence of Theileria-transformed macrophages , Infection and immunity , 83 , 1869 – 1880 . Google Scholar Crossref Search ADS PubMed WorldCat Heussler , V. T. , Machado , J. Jr, Fernandez , P. C. et al. . ( 1999 ) The intracellular parasite Theileria parva protects infected T cells from apoptosis , Proc Natl Acad Sci U S A , 96 , 7312 – 7317 . Google Scholar Crossref Search ADS PubMed WorldCat Heussler , V. T. , Kuenzi , P., Fraga , F. et al. . ( 2001 ) The Akt/PKB pathway is constitutively activated in Theileria-transformed leucocytes, but does not directly control constitutive NF-kB activation , Cellular Microbiology , 3 , 537 – 550 . Google Scholar Crossref Search ADS PubMed WorldCat Heussler , V. T. , Rottenberg , S., Schwab , R. et al. . ( 2002 ) Hijacking of host cell IKK signalosomes by the transforming parasite Theileria , Science (New York, N.Y.) , 298 , 1033 – 1036 . Google Scholar Crossref Search ADS PubMed WorldCat Huynh , N. , Liu , K. H., Baldwin , G. S. et al. . ( 2010 ) P21- activated kinase 1 stimulates colon cancer cell growth and migration/invasion via ERK- and AKT-dependent pathways , Biochimica et biophysica acta , 1803 , 1106 – 1113 . Google Scholar Crossref Search ADS PubMed WorldCat Kawashima , S. and Kanehisa , M. ( 2000 ) AAindex: amino acid index database , Nucleic acids research , 28 , 374 . Google Scholar Crossref Search ADS PubMed WorldCat Kinnaird , J. H. , Weir , W., Durrani , Z. et al. . ( 2013 ) A bovine lymphosarcoma cell line infected with Theileria annulata exhibits an irreversible reconfiguration of host cell gene expression , PLoS ONE , 8 ( 6 ), e66833 . Google Scholar Crossref Search ADS PubMed WorldCat Kuenzi , P. , Schneider , P. and Dobbelaere , D. A. ( 2003 ) Theileria parva-transformed T cells show enhanced resistance to Fas/Fas ligand-induced apoptosis , Journal of immunology , 171 , 1224 – 1231 . Google Scholar Crossref Search ADS WorldCat Lizundia , R. , Chaussepied , M., Huerre , M. et al. . ( 2006 ) c-Jun NH2-terminal kinase/c-Jun signaling promotes survival and metastasis of B lymphocytes transformed by Theileria , Cancer research , 66 , 6105 – 6110 . Google Scholar Crossref Search ADS PubMed WorldCat Luder , C. G. , Stanway , R. R., Chaussepied , M. et al. . ( 2009 ) Intracellular survival of apicomplexan parasites and host cell modification , International journal for parasitology , 39 , 163 – 173 . Google Scholar Crossref Search ADS PubMed WorldCat Medjkane , S. , Perichon , M., Marsolier , J. et al. . ( 2014 ) Theileria induces oxidative stress and HIF1-α activation that are essential for host leukocyte transformation , Oncogene , 33 , 1809 – 1817 . Google Scholar Crossref Search ADS PubMed WorldCat Metheni , M. , Echebli , N., Chaussepied , M. et al. . ( 2014 ) The level of H(2)O(2) type oxidative stress regulates virulence of Theileria transformed leukocytes , Cellular microbiology , 16 , 269 – 279 . Google Scholar Crossref Search ADS PubMed WorldCat Metheni , M. , Lombes , A., Bouillaud , F. et al. . ( 2015 ) Microreview: HIF-1a induction, proliferation and glycolysis of Theileria-infected leukocytes , Cellular microbiology , 17 , 467 – 472 . Google Scholar Crossref Search ADS PubMed WorldCat Ma , M. and Baumgartner , M. ( 2014 ) Intracellular Theileria annulata promote invasive cell motility through kinase regulation of the host actin cytoskeleton , PLoS pathogens , 10 , e1004003 . Google Scholar Crossref Search ADS PubMed WorldCat Nambiar , R. , McConnell , R. E. and Tyska , M. J. ( 2010 ) Myosin motor function. The ins and outs of actin-based membrane protrusions , Cellular and molecular life sciences : CMLS , 67 , 1239 – 1254 . Google Scholar Crossref Search ADS PubMed WorldCat Neumann , M. , Foryst-Ludwig , A., Klar , S. et al. . ( 2006 ) The PAK1 autoregulatory domain is required for interaction with NIK in Helicobacter pylori induced NF-kappaB activation , Biological chemistry , 387 , 79 – 86 . Google Scholar Crossref Search ADS PubMed WorldCat Rao , H. B. , Zhu , F., Yang , G. B. et al. . ( 2011 ) Update of PROFEAT: a web server for computing structural and physico-chemical features of proteins and peptides from amino acid sequence , Nucleic acids research , 39 , W385 – W390 . Google Scholar Crossref Search ADS PubMed WorldCat Schneider , I. , Haller , D., Kullmann , B. et al. . ( 2007 ) Identification, molecular characterization and subcellular localization of a Theileria annulata parasite protein secreted into the host cell cytoplasm , Parasitology research , 101 , 1471 – 1482 . Google Scholar Crossref Search ADS PubMed WorldCat Shaw , M. K. ( 2003 ) Cell invasion by Theileria sporozoites , Trends in parasitology , 19 , 2 – 6 . Google Scholar Crossref Search ADS PubMed WorldCat Shiels , B. R. , McKellar , S., Katzer , F. et al. . ( 2004 ) A Theileria annulata DNA binding protein located to the host cell nucleus alters the phenotype of a bovine macrophage cell line , Eukaryotic Cell , 3 , 495 – 505 . Google Scholar Crossref Search ADS PubMed WorldCat Shiels , B. , Langsley , G., Weir , W. et al. . ( 2006 ) Alteration of host cell phenotype by Theileria annulata and Theileria parva: mining for manipulators in the parasite genomes , International journal for parasitology , 36 , 9 – 21 . Google Scholar Crossref Search ADS PubMed WorldCat Shen , J. , Zhang , J., Luo , X. et al. . ( 2007 ) Predicting protein-protein interactions based only on sequences information , PNAS Biophysics , 104 ( 11 ), 4337 – 4341 . Google Scholar OpenURL Placeholder Text WorldCat Swan , D. G. , Stern , R., McKellar , S. et al. . ( 2001 ) Characterization of a cluster of genes encoding Theileria annulata AT hook DNA-binding proteins and evidence for localization to the host cell nucleus , Journal of cell science , 114 , 2747 – 2754 . Google Scholar PubMed OpenURL Placeholder Text WorldCat Swan , D. G. , Stadler , L., Okan , E. et al. . ( 2003 ) TashHN, a Theileria annulata encoded protein transported to the host nucleus displays an association with attenuation of parasite differentiation , Cellular microbiology , 5 , 947 – 956 . Google Scholar Crossref Search ADS PubMed WorldCat Tretina , K. , Gotia , H. T., Mann , D. J. et al. . ( 2015 ) Theileria-transformed bovine leukocytes have cancer hallmarks , Trends in parasitology , 31 ( 7 ), 306 – 314 . Google Scholar Crossref Search ADS PubMed WorldCat Vignali , M. , McKinlay , A., LaCount , D. J. et al. . ( 2008 ) Interaction of an atypical Plasmodium falciparum ETRAMP with human apolipoproteins , Malaria journal , 7 , 211 . Google Scholar Crossref Search ADS PubMed WorldCat Von-Schubert , C. , Xue , G., Schmuckli-Maurer , J. et al. . ( 2010 ) The transforming parasite Theileria co-opts host cell mitotic and central spindles to persist in continuously dividing cells , PLoS Biology , 8 , 43 . Google Scholar Crossref Search ADS WorldCat Yu , H. , Pardoll , D. and Jove , R. ( 2009 ) STATs in cancer inflammation and immunity: a leading role for STAT3 , Nature reviews. Cancer , 9 ( 11 ), 798 – 908 . Google Scholar Crossref Search ADS PubMed WorldCat © The Author 2016. Published by Oxford University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com © The Author 2016. Published by Oxford University Press.

Journal

BioScience HorizonsOxford University Press

Published: Jan 1, 2016

There are no references for this article.