Decoding Arabidopsis thaliana CPK/SnRK Superfamily Kinase Client Signaling Networks Using Peptide Library and Mass Spectrometry.

Members of the calcium-dependent protein kinase (CDPK/CPK) and SNF-related protein kinase (SnRK) superfamilies are commonly found in plants and some protists. Our knowledge of client specificity of the members of this superfamily is fragmentary. As this family is represented by over 30 members in Arabidopsis thaliana, the identification of kinase-specific and overlapping client relationships is crucial to our understanding the nuances of this large family of kinases as directed towards signal transduction pathways. Herein, we used the kinase client (KiC) assay-a relative, quantitative, high-throughput mass spectrometry-based in vitro phosphorylation assay-to identify and characterize potential CPK/SnRK targets of Arabidopsis. Eight CPKs (1, 3, 6, 8, 17, 24, 28, and 32), four SnRKs (subclass 1 and 2), and PPCK1 and PPCK2 were screened against a synthetic peptide library that contains 2095 peptides and 2661 known phosphorylation sites. A total of 625 in vitro phosphorylation sites corresponding to 203 non-redundant proteins were identified. The most promiscuous kinase, CPK17, had 105 candidate target proteins, many of which had already been discovered. Sequence analysis of the identified phosphopeptides revealed four motifs: LxRxxS, RxxSxxR, RxxS, and LxxxxS, that were significantly enriched among CPK/SnRK clients. The results provide insight into both CPK- and SnRK-specific and overlapping signaling network architectures and recapitulate many known in vivo relationships validating this large-scale approach towards discovering kinase targets.

Biocontrol of Rice Seed-Associated Fungal Pathogens Using Green Synthesis Approaches.

Rice (Oryza sativa) is a major cereal crop that balances the food demand of the worldwide population. The crop quality drops daily due to their exposure to biotic and abiotic stresses, especially pathogens. It needs to be improved to maintain the consumption level to cope with increasing population demands for food. The current study was designed to analyze the comparison of the effects of green synthesis approaches on pathogens associated with rice seeds. In this study, essential oils were extracted from Cymbopogon citratus, Thymus vulgaris, and Origanum vulgaris medicinal plants and used as fungicides on fungal strains of Aspergillus spp. T. vulgaris effectively controlled the growth of Aspergillus niger, Aspergillus flavus, and Aspergillus terreus as compared with O. vulgaris and Cymbopogon. Further, silica nanoparticles (SiNPs) were synthesized from rice husk to evaluate their antifungal activities. SiNPs were characterized by ultraviolet-visible spectroscopy with a broad peak at 281.62 nm. Fourier-transform infrared spectroscopy spectrum confirms the presence of Si-H, Si-OH, and Si-O-Si bonds functional groups, and SiO4 tetrahedral coordination unit. X-ray diffraction pattern describes the crystalline structure with a sharp peak at 2θ = 22°. Scanning electron microscopy and energy-dispersive spectroscopy confirmed the spherical shape, size 70-115 nm, and elemental composition with pure silica contents. SiNPs showed no significant antifungal activity against Aspergillus strains. Moreover, Trichoderma was isolated from the rhizosphere of rice fields and showed a surprising antifungal effect against A. terreus, A. niger, and A. flavus. The current study successfully revealed environment-friendly and cost-effective green synthesizing approaches for analyzing biocontrol potential against rice seed-related Aspergillus spp. They will also help to improve pathogen control strategies in other cereals.

Dose-dependent anti-hyperglycemic & anti-dyslipidemic potential of aqueous leaves extract of Typha elephantina in-vivo and in-vitro.

Diabetes mellitus is among the fundamental causes of illness and millions of deaths around the globe are directly attributed to it each year. Current antidiabetic medications often lack sustained glycemic control and carry significant risks of side effects. As a result, the use of plant-based treatments has gained popularity. In this experimental study, we evaluated the aqueous extracts (LQE) of Typha elephantina (also known as Elephant grass) leaves collected from freshwater marshes, for their potential anti-hyperglycemic and anti-hyperlipidemic antioxidant effects in healthy streptozotocin caused diabetic-mice. We employed glucose adsorption tests at different glucose levels and glucose diffusion tests to assess the in-vitro antidiabetic action of plant extract. For the in-vivo trail, we measured fasting blood glucose (FBG), glucose tolerance (GTT), as well as long-term anti-diabetic, anti-hyperlipidemic, and antioxidant activities. Our results from the glucose diffusion test indicated that the extract was highly effective at both low glucose concentrations (5 mmol L) and high glucose concentrations (100 mmol L). However, the glucose-diffusion ability reached its peaked at an excessively high dosage of the aqueous extract, suggesting a dose-related effect. Similarly, we observed that high doses of TEL.AQ extracts (400 mg/kg body weight) significantly reduced blood glucose levels in healthy mice during the glucose tolerance test (GTT) at 3 h and fasting blood glucose studies (FBG) at 6 h. Furthermore, the high-dose TEL.AQ extract effectively reduced liver-related serum markers and blood-glucose concentration (BGC) in severely chronic diabetic rats. The extract dosage also influenced lipid profile, conjugate and unconjugated bilirubin levels, cholesterol, triglycerides, HDL, and total bilirubin levels. Additionally, after administering a high extract dose, we observed considerable improvement in the liver homogenate markers CAT, POD, and SOD. In contrast, the extract at a low dosage (100 mg/kg), showed minimal, while a moderate dose (200 mg/kg), yielded promising results.

Protein-Protein Interactions and Quantitative Phosphoproteomic Analysis Reveal Potential Mitochondrial Substrates of Protein Phosphatase 2A-B'ζ Holoenzyme.

Protein phosphatase 2A (PP2A) is a heterotrimeric conserved serine/threonine phosphatase complex that includes catalytic, scaffolding, and regulatory subunits. The 3 A subunits, 17 B subunits, and 5 C subunits that are encoded by the Arabidopsis genome allow 255 possible PP2A holoenzyme combinations. The regulatory subunits are crucial for substrate specificity and PP2A complex localization and are classified into the B, B', and B" non-related families in land plants. In Arabidopsis, the close homologs B'η, B'θ, B'γ, and B'ζ are further classified into a subfamily of B' called B'η. Previous studies have suggested that mitochondrial targeted PP2A subunits (B'ζ) play a role in energy metabolism and plant innate immunity. Potentially, the PP2A-B'ζ holoenzyme is involved in the regulation of the mitochondrial succinate/fumarate translocator, and it may affect the enzymes involved in energy metabolism. To investigate this hypothesis, the interactions between PP2A-B'ζ and the enzymes involved in the mitochondrial energy flow were investigated using bimolecular fluorescence complementation in tobacco and onion cells. Interactions were confirmed between the B'ζ subunit and the Krebs cycle proteins succinate/fumarate translocator (mSFC1), malate dehydrogenase (mMDH2), and aconitase (ACO3). Additional putative interacting candidates were deduced by comparing the enriched phosphoproteomes of wild type and B'ζ mutants: the mitochondrial regulator Arabidopsis pentatricopeptide repeat 6 (PPR6) and the two metabolic enzymes phosphoenolpyruvate carboxylase (PPC3) and phosphoenolpyruvate carboxykinase (PCK1). Overall, this study identifies potential PP2A substrates and highlights the role of PP2A in regulating energy metabolism in mitochondria.

Peroxisomal protein phosphatase PP2A-B' theta interacts with and piggybacks SINA-like 10 E3 ligase into peroxisomes.

Protein phosphatase 2A (PP2A) is targeted to the plant peroxisome via a C-terminal SSL sequence on its regulatory B' theta (θ) subunit. To date the substrates of peroxisomal PP2A are unknown but are thought to be recruited by the regulatory B'θ subunit. Employing yeast two hybrid screening, we have identified Arabidopsis E3 ligase SINA-like 10 as a B'θ binding partner. The E3 ligase SINA-like 10 was found to harbor the PP2A B'-binding Short Linear interaction Motif or SLiM, LxxIxE. This interaction was further verified both in vitro and in vivo using direct pulldown assays and bimolecular fluorescence complementation. Utilizing peroxisomal targeted and a cytosolic version of B'θ (lacking its C-terminal peroxisomal targeting sequence SSL>) bimolecular fluorescence complementation suggests an interaction to occur in the cytosol followed by piggybacking E3 ligase SINA-like 10 into peroxisomes. These results identify a first peroxisomal PP2A interactor, which also obtains a PP2A B'-binding SLiM.

Identification of Arabidopsis Protein Kinases That Harbor Functional Type 1 Peroxisomal Targeting Signals.

Peroxisomes are eukaryotic specific organelles that perform diverse metabolic functions including fatty acid ß-oxidation, reactive species metabolism, photorespiration, and responses to stress. However, the potential regulation of these functions by post-translational modifications, including protein phosphorylation, has had limited study. Recently, we identified and catalogued a large number of peroxisomal phosphorylated proteins, implicating the presence of protein kinases in this organelle. Here, we employed available prediction models coupled with sequence conservation analysis to identify 31 protein kinases from the Arabidopsis kinome (all protein kinases) that contain a putative, non-canonical peroxisomal targeting signal type 1 (PTS1). From this, twelve C-terminal domain-PTS1s were demonstrated to be functional in vivo, targeting enhanced yellow fluorescent protein to peroxisomes, increasing the list of presumptive peroxisomal protein kinases to nineteen. Of the twelve protein kinases with functional PTS1s, we obtained full length clones for eight and demonstrated that seven target to peroxisomes in vivo. Screening homozygous mutants of the presumptive nineteen protein kinases revealed one candidate (GPK1) that harbors a sugar-dependence phenotype, suggesting it is involved in regulating peroxisomal fatty acid ß-oxidation. These results present new opportunities for investigating the regulation of peroxisome functions.

Multi-targeted trehalose-6-phosphate phosphatase I harbors a novel peroxisomal targeting signal 1 and is essential for flowering and development.

MAIN CONCLUSION: This work reveals information about new peroxisomal targeting signals type 1 and identifies trehalose-6-phosphate phosphatase I as multitargeted and is implicated in plant development, reproduction, and stress response. A putative, non-canonical peroxisomal targeting signal type 1 (PTS1) Pro-Arg-Met > was identified in the extreme C-terminus of trehalose-6-phosphate phosphatase (TPP)I. TPP catalyzes the final step of trehalose synthesis, and the enzyme was previously characterized to be nuclear only (Krasensky et al. in Antioxid Redox Signal 21(9):1289-1304, 2014). Here we show that the TPPI C-terminal decapeptide ending with Pro-Arg-Met > or Pro-Lys-Met > can indeed function as a PTS1. Upon transient expression in two plant expression systems, the free C- or N-terminal end led to the full-length TPPI targeting to peroxisomes and plastids, respectively. The nucleus and nucleolus targeting of the full-length TPPI was observed in both cases. The homozygous T-DNA insertion line of TPPI showed a pleiotropic phenotype including smaller leaves, shorter roots, delayed flowering, hypersensitivity to salt, and a sucrose dependent seedling development. Our results identify novel PTS1s, and TPPI as a protein multi-targeted to peroxisomes, plastids, nucleus, and nucleolus. Altogether our findings implicate an essential role for TPPI in development, reproduction, and cell signaling.

A Framework to Investigate Peroxisomal Protein Phosphorylation in Arabidopsis.

Peroxisomes perform essential roles in a range of cellular processes, highlighted by lipid metabolism, reactive species detoxification, and response to a variety of stimuli. The ability of peroxisomes to grow, divide, respond to changing cellular needs, interact with other organelles, and adjust their proteome as required, suggest that, like other organelles, their specialized roles are highly regulated. Similar to most other cellular processes, there is an emerging role for protein phosphorylation to regulate these events. In this review, we establish a knowledge framework of key players that control protein phosphorylation events in the plant peroxisome (i.e., the protein kinases and phosphatases), and highlight a vastly expanded set of (phospho)substrates.

Protein Kinases and Phosphatases of the Plastid and Their Potential Role in Starch Metabolism.

Phospho-proteomic studies have confirmed that phosphorylation is a common mechanism to regulate protein function in the chloroplast, including the enzymes of starch metabolism. In addition to the photosynthetic machinery protein kinases (STN7 and STN8) and their cognate protein phosphatases PPH1 (TAP38) and PBCP, multiple other protein kinases and phosphatases have now been localized to the chloroplast. Here, we build a framework for understanding protein kinases and phosphatases, their regulation, and potential roles in starch metabolism. We also catalog mapped phosphorylation sites on proteins of chloroplast starch metabolism to illustrate the potential and mostly unknown roles of protein phosphorylation in the regulation of starch biology.

Methylation of protein phosphatase 2A-Influence of regulators and environmental stress factors.

Protein phosphatase 2A catalytic subunit (PP2A-C) has a terminal leucine subjected to methylation, a regulatory mechanism conserved from yeast to mammals and plants. Two enzymes, LCMT1 and PME1, methylate and demethylate PP2A-C, respectively. The physiological importance of these posttranslational modifications is still enigmatic. We investigated these processes in Arabidopsis thaliana by mutant phenotyping, by global expression analysis, and by monitoring methylation status of PP2A-C under different environmental conditions. The lcmt1 mutant, possessing essentially only unmethylated PP2A-C, had less dense rosettes, and earlier flowering than wild type (WT). The pme1 mutant, with 30% reduction in unmethylated PP2A-C, was phenotypically comparable with WT. Approximately 200 overlapping genes were twofold upregulated, and 200 overlapping genes were twofold downregulated in both lcmt1 and pme1 relative to WT. Differences between the 2 mutants were also striking; 97 genes were twofold upregulated in pme1 compared with lcmt1, indicating that PME1 acts as a negative regulator for these genes. Analysis of enriched GO terms revealed categories of both abiotic and biotic stress genes. Furthermore, methylation status of PP2A-C was influenced by environmental stress, especially by hypoxia and salt stress, which led to increased levels of unmethylated PP2A-C, and highlights the importance of PP2A-C methylation/demethylation in environmental responses.

PLATINUM SENSITIVE 2 LIKE impacts growth, root morphology, seed set, and stress responses.

Eukaryotic protein phosphatase 4 (PP4) is a PP2A-type protein phosphatase that is part of a conserved complex with regulatory factors PSY2 and PP4R2. Various lines of Arabidopsis thaliana with mutated PP4 subunit genes were constructed to study the so far completely unknown functions of PP4 in plants. Mutants with knocked out putative functional homolog of the PSY2 LIKE (PSY2L) gene were dwarf and bushy, while plants with knocked out PP4R2 LIKE (PP4R2L) looked very similar to WT. The psy2l seedlings had short roots with disorganized morphology and impaired meristem. Seedling growth was sensitive to the genotoxin cisplatin. Global transcript analysis (RNA-seq) of seedlings and rosette leaves revealed several groups of genes, shared between both types of tissues, strongly influenced by knocked out PSY2L. Receptor kinases, CRINKLY3 and WAG1, important for growth and development, were down-regulated 3-7 times. EUKARYOTIC ELONGATION FACTOR5A1 was down-regulated 4-6 fold. Analysis of hormone sensitive genes indicated that abscisic acid levels were high, while auxin, cytokinin and gibberellic acid levels were low in psy2l. Expression of specific transcription factors involved in regulation of anthocyanin synthesis were strongly elevated, e.g. the master regulator PAP1, and intriguingly TT8, which is otherwise mainly expressed in seeds. The psy2l mutants accumulated anthocyanins under conditions where WT did not, pointing to PSY2L as a possible upstream negative regulator of PAP1 and TT8. Expression of the sugar-phosphate transporter GPT2, important for cellular sugar and phosphate homeostasis, was enhanced 7-8 times. Several DNA damage response genes, including the cell cycle inhibitor gene WEE1, were up-regulated in psy2l. The activation of DNA repair signaling genes, in combination with phenotypic traits showing aberrant root meristem and sensitivity to the genotoxic cisplatin, substantiate the involvement of Arabidopsis PSY2L in maintenance of genome integrity.

Light regulation of nitrate reductase by catalytic subunits of protein phosphatase 2A.

MAIN CONCLUSION: PP2A catalytic subunit C2 is of special importance for light/dark regulation of nitrate reductase activity. The level of unmethylated PP2A catalytic subunits decreases in darkness. Protein phosphatase 2A (PP2A) dephosphorylates and activates nitrate reductase (NR) in photosynthetically active tissue when plants are transferred from darkness to light. In the present work, investigation of Arabidopsis thaliana PP2A mutant lines revealed that one of the five PP2A catalytic subunit genes, e.g., C2, was of special importance for NR activation. Impairment of NR activation was, especially pronounced in the c2c4 double mutant. Though weaker, NR activation was also impaired in the c2 single mutant, and c1c2 and c2c5 double mutants. On the other hand, NR activation in the c4c5 double mutant was as efficient as in WT. The c4 single mutant had low PP2A activity, whereas the c2 single mutant possessed WT levels of extractable PP2A activity. PP2A activity was low in both c2c4 and c4c5. Differences in extracted PP2A activity among mutants did not strictly correlate with differences in NR activation, but underpinned that C2 has a special function in NR activation in vivo. The terminal leucine in PP2A catalytic subunits is generally methylated to a high degree, but regulation and impact of methylation/demethylation is barely studied. In WT and PP2A mutants, the level of unmethylated PP2A catalytic subunits decreased during 45 min of darkness, but did not change much when light was switched on. In leucine carboxyl methyl transferase1 (LCMT1) knockout plants, which possess mainly unmethylated PP2A, NR was still activated, although not fully as efficient as in WT.

Towards understanding peroxisomal phosphoregulation in Arabidopsis thaliana.

MAIN CONCLUSION: This work identifies new protein phosphatases and phosphatase-related proteins targeting peroxisomes, and raises the question of a novel protein import pathway from ER to peroxisomes involving peroxisomal targeting signal type 1 (PTS1) Plant peroxisomes are essential for several processes, for example lipid metabolism, free radical detoxification, development, and stress-related functions. Although research on peroxisomes has been intensified, reversible phosphorylation as a control mechanism in peroxisomes is barely studied. Therefore, it is crucial to identify all peroxisomal proteins involved in phosphoregulation. We here started with protein phosphatases, and searched the Arabidopsis thaliana genome for phosphatase-related proteins with putative peroxisomal targeting signals (PTS). Five potential peroxisomal candidates were detected, from which four were confirmed to target peroxisomes or have a functional PTS. The highly conserved Ser-Ser-Met> was validated for two protein phosphatase 2C (PP2C) family members (POL like phosphatases, PLL2 and PLL3) as a functional peroxisomal targeting signal type 1 (PTS1). Full-length PLL2 and PLL3 fused with a reporter protein targeted peroxisomes in two plant expression systems. A putative protein phosphatase, purple acid phosphatase 7 (PAP7), was found to be dually targeted to ER and peroxisomes and experiments indicated a possible trafficking to peroxisomes via the ER depending on peroxisomal PTS1. In addition, a protein phosphatase 2A regulator (TIP41) was validated to harbor a functional PTS1 (Ser-Lys-Val>), but the full-length protein targeted cytosol and nucleus. Reverse genetics indicated a role for TIP41 in senescence signaling. Mass spectrometry of whole seedlings and isolated peroxisomes was employed, and identified new putative phosphorylated peroxisomal proteins. Previously, only one protein phosphatase, belonging to the phospho-protein phosphatase (PPP) family, was identified as a peroxisomal protein. The present work implies that members of two other main protein phosphatase families, i.e. PP2C and PAP, are also targeting peroxisomes.

Protein phosphatase 2A regulatory subunits affecting plant innate immunity, energy metabolism, and flowering time--joint functions among B'η subfamily members.

Protein phosphatase 2A (PP2A) is a heterotrimeric complex comprising a catalytic, scaffolding, and regulatory subunit. The regulatory subunits are essential for substrate specificity and localization of the complex and are classified into B/B55, B', and B" non-related families in higher plants. In Arabidopsis thaliana, the close paralogs B'η, B'θ, B'γ, and B'ζ were further classified into a subfamily of B' called B'η. Here we present results that consolidate the evidence for a role of the B'η subfamily in regulation of innate immunity, energy metabolism and flowering time. Proliferation of the virulent Pseudomonas syringae in B'θ knockout mutant decreased in comparison with wild type plants. Additionally, B'θ knockout plants were delayed in flowering, and this phenotype was supported by high expression of FLC (FLOWERING LOCUS C). B'ζ knockout seedlings showed growth retardation on sucrose-free medium, indicating a role for B'ζ in energy metabolism. This work provides insight into functions of the B'η subfamily members, highlighting their regulation of shared physiological traits while localizing to distinct cellular compartments.

MAP kinase phosphatase 1 harbors a novel PTS1 and is targeted to peroxisomes following stress treatments.

In Arabidopsis thaliana, twenty mitogen-activated protein kinases (MAPKs/MPKs) are regulated by five MAP kinase phosphatases (MKPs). Arabidopsis MKP1 has an important role in biotic, abiotic and genotoxic stresses and has been shown to interact with and negatively regulate specifically MPK3 and MPK6. MKP1 has been reported to have a role in negative regulation of reactive oxygen species (ROS) and salicylic acid (SA) production. As essential organelles involved in production of ROS and SA, peroxisomes could possibly be an important compartment for MKP1 activity, however MKP1 was previously reported to be cytosolic. By screening Arabidopsis protein phosphatases for peroxisomal targeting signal 1 (PTS1), we identified MKP1 as a putative peroxisomal protein. Arabidopsis MKP1 was found to harbor a non-canonical PTS1-like tripeptide (Ser-Ala-Leu>) that is conserved in MKP1 orthologs. We show experimentally that the C-terminal Ser-Ala-Leu> can function as a novel PTS1, and alanine in position -2, adds more relaxation to the plant PTS1 motif. The full-length MKP1 remained in the cytosol when transiently expressed in Arabidopsis mesophyll protoplasts under standard conditions. When different biotic and abiotic stresses were applied to mesophyll protoplasts, the full length protein changed its targeting to unidentified organelle-like structures that subsequently fused with peroxisomes. Our results identify MKP1 as a protein dually targeted to cytosol and peroxisomes. The finding that MKP1 targets peroxisomes by a non-canonical PTS1 under stressful conditions highlights the complexity of peroxisomal targeting mechanism.

Protein phosphatase 2A holoenzyme is targeted to peroxisomes by piggybacking and positively affects peroxisomal ß-oxidation.

The eukaryotic, highly conserved serine (Ser)/threonine-specific protein phosphatase 2A (PP2A) functions as a heterotrimeric complex composed of a catalytic (C), scaffolding (A), and regulatory (B) subunit. In Arabidopsis (Arabidopsis thaliana), five, three, and 17 genes encode different C, A, and B subunits, respectively. We previously found that a B subunit, B'θ, localized to peroxisomes due to its C-terminal targeting signal Ser-Ser-leucine. This work shows that PP2A C2, C5, andA2 subunits interact and colocalize with B'θ in peroxisomes. C and A subunits lack peroxisomal targeting signals, and their peroxisomal import depends on B'θ and appears to occur by piggybacking transport. B'θ knockout mutants were impaired in peroxisomal ß-oxidation as shown by developmental arrest of seedlings germinated without sucrose, accumulation of eicosenoic acid, and resistance to protoauxins indole-butyric acid and 2,4-dichlorophenoxybutyric acid. All of these observations strongly substantiate that a full PP2A complex is present in peroxisomes and positively affects ß-oxidation of fatty acids and protoauxins.

Protein phosphatases PP2A, PP4 and PP6: mediators and regulators in development and responses to environmental cues.

The three closely related groups of serine/threonine protein phosphatases PP2A, PP4 and PP6 are conserved throughout eukaryotes. The catalytic subunits are present in trimeric and dimeric complexes with scaffolding and regulatory subunits that control activity and confer substrate specificity to the protein phosphatases. In Arabidopsis, three scaffolding (A subunits) and 17 regulatory (B subunits) proteins form complexes with five PP2A catalytic subunits giving up to 255 possible combinations. Three SAP-domain proteins act as regulatory subunits of PP6. Based on sequence similarities with proteins in yeast and mammals, two putative PP4 regulatory subunits are recognized in Arabidopsis. Recent breakthroughs have been made concerning the functions of some of the PP2A and PP6 regulatory subunits, for example the FASS/TON2 in regulation of the cellular skeleton, B' subunits in brassinosteroid signalling and SAL proteins in regulation of auxin transport. Reverse genetics is starting to reveal also many more physiological functions of other subunits. A system with key regulatory proteins (TAP46, TIP41, PTPA, LCMT1, PME-1) is present in all eukaryotes to stabilize, activate and inactivate the catalytic subunits. In this review, we present the status of knowledge concerning physiological functions of PP2A, PP4 and PP6 in Arabidopsis, and relate these to yeast and mammals.

Non-canonical peroxisome targeting signals: identification of novel PTS1 tripeptides and characterization of enhancer elements by computational permutation analysis.

BACKGROUND: High-accuracy prediction tools are essential in the post-genomic era to define organellar proteomes in their full complexity. We recently applied a discriminative machine learning approach to predict plant proteins carrying peroxisome targeting signals (PTS) type 1 from genome sequences. For Arabidopsis thaliana 392 gene models were predicted to be peroxisome-targeted. The predictions were extensively tested in vivo, resulting in a high experimental verification rate of Arabidopsis proteins previously not known to be peroxisomal. RESULTS: In this study, we experimentally validated the predictions in greater depth by focusing on the most challenging Arabidopsis proteins with unknown non-canonical PTS1 tripeptides and prediction scores close to the threshold. By in vivo subcellular targeting analysis, three novel PTS1 tripeptides (QRL>, SQM>, and SDL>) and two novel tripeptide residues (Q at position -3 and D at pos. -2) were identified. To understand why, among many Arabidopsis proteins carrying the same C-terminal tripeptides, these proteins were specifically predicted as peroxisomal, the residues upstream of the PTS1 tripeptide were computationally permuted and the changes in prediction scores were analyzed. The newly identified Arabidopsis proteins were found to contain four to five amino acid residues of high predicted targeting enhancing properties at position -4 to -12 in front of the non-canonical PTS1 tripeptide. The identity of the predicted targeting enhancing residues was unexpectedly diverse, comprising besides basic residues also proline, hydroxylated (Ser, Thr), hydrophobic (Ala, Val), and even acidic residues. CONCLUSIONS: Our computational and experimental analyses demonstrate that the plant PTS1 tripeptide motif is more diverse than previously thought, including an increasing number of non-canonical sequences and allowed residues. Specific targeting enhancing elements can be predicted for particular sequences of interest and are far more diverse in amino acid composition and positioning than previously assumed. Machine learning methods become indispensable to predict which specific proteins, among numerous candidate proteins carrying the same non-canonical PTS1 tripeptide, contain sufficient enhancer elements in terms of number, positioning and total strength to cause peroxisome targeting.

Experimental and statistical post-validation of positive example EST sequences carrying peroxisome targeting signals type 1 (PTS1).

We recently developed the first algorithms specifically for plants to predict proteins carrying peroxisome targeting signals type 1 (PTS1) from genome sequences. As validated experimentally, the prediction methods are able to correctly predict unknown peroxisomal Arabidopsis proteins and to infer novel PTS1 tripeptides. The high prediction performance is primarily determined by the large number and sequence diversity of the underlying positive example sequences, which mainly derived from EST databases. However, a few constructs remained cytosolic in experimental validation studies, indicating sequencing errors in some ESTs. To identify erroneous sequences, we validated subcellular targeting of additional positive example sequences in the present study. Moreover, we analyzed the distribution of prediction scores separately for each orthologous group of PTS1 proteins, which generally resembled normal distributions with group-specific mean values. The cytosolic sequences commonly represented outliers of low prediction scores and were located at the very tail of a fitted normal distribution. Three statistical methods for identifying outliers were compared in terms of sensitivity and specificity." Their combined application allows elimination of erroneous ESTs from positive example data sets. This new post-validation method will further improve the prediction accuracy of both PTS1 and PTS2 protein prediction models for plants, fungi, and mammals.