Secretome weaponries of Cochliobolus lunatus interacting with potato leaf at different temperature regimes reveal a CL[xxxx]LHM - motif

Background Plant and animal pathogenic fungus Cochliobolus lunatus cause great economic damages worldwide every year. C. lunatus displays an increased temperature dependent-virulence to a wide range of hosts. Nonetheless, this phenomenon is poorly understood due to lack of insights on the coordinated secretome weaponries produced by C. lunatus under heat-stress conditions on putative hosts. To understand the mechanism better, we dissected the secretome of C. lunatus interacting with potato (Solanum tuberosum L.) leaf at different temperature regimes. Results C. lunatus produced melanized colonizing hyphae in and on potato leaf, finely modulated the ambient pH as a function of temperature and secreted diverse set of proteins. Using two dimensional gel electrophoresis (2-D) and mass spectrometry (MS) technology, we observed discrete secretomes at 20°C, 28°C and 38°C. A total of 21 differentially expressed peptide spots and 10 unique peptide spots (that did not align on the gels) matched with 28 unique protein models predicted from C. lunatus m118 v.2 genome peptides. Furthermore, C. lunatus secreted peptides via classical and non-classical pathways related to virulence, proteolysis, nucleic acid metabolism, carbohydrate metabolism, heat stress, signal trafficking and some with unidentified catalytic domains. Conclusions We have identified a set of 5 soluble candidate effectors of unknown function from C. lunatus secretome weaponries against potato crop at different temperature regimes. Our findings demonstrate that C. lunatus has a repertoire of signature secretome which mediates thermo-pathogenicity and share a leucine rich “CL[xxxx]LHM”-motif. Considering the rapidly evolving temperature dependent-virulence and host diversity of C. lunatus, this data will be useful for designing new protection strategies.

'Secretome' refers to a set of secreted proteins at a given physiologic condition; which plays a key role in cell signaling, intracellular trafficking and migration of invasive weaponries (i.e. candidate effectors) in pathogenic interactions. C. lunatus has attracted the interest of many workers on various aspects viz., induce-virulence variation, virulence differentiation and heat-dependent aggressiveness [7][8][9][10][11][12]. Experimentally, extracellular weaponries secreted by pathogens are crucial for increased virulence and disease development in the context of plant-pathogen interaction sensu stricto. Candidate effector molecules are believed to manipulate host cell structure and function, thereby facilitating infection and suppression of the host immune responses [13,14]. Once candidate effectors are deployed, they act either in the exhaustorial matrix, the extracellular space or within the host cell cytoplasm to promote invasion and pathogenicity [13][14][15]. In conditions where candidate effectors are recognized by the host disease resistance (R) proteins, hallmark resistance occurs via programmed cell death. In this case candidate effectors are considered to have an avirulence activity. Often, fungi discharge their candidate effectors into their surroundings via a non-classical pathway which does not require an Nterminal signal peptide [15]. On the contrary in classical secretory pathway, candidate effectors are directed by the N-terminal peptide signal via the endoplasmic reticulum and Golgi systems to their extracellular locations [13][14][15][16].
Frequently, pathogens differentially produce enzymes based on the environmental conditions [16]. Fluctuations of temperature in most cases play a decisive role in the development of disease, since the physiology of either the host or pathogen can change and significantly modulate the interaction dynamics. Interestingly, C. lunatus virulence increases with ambient temperature upto 38°C [9][10][11][12]. Nevertheless, whether C. lunatus discharges secretome weaponries under heat stress conditions on putative host is not known. Thus, examining C. lunatus temperature-dependent secretome on a putative host is important and can permit the discovery of candidate effectors that govern its virulence and thermo-pathogenicity. To date, the secretome architecture of C. lunatus is not explored and could be of value for designing a suitable control measure in the context of the current rise in global temperature. In this study, microscopic analysis was performed first to decipher the nature of potato leaf invasion in the liquid phase. Subsequently, we used 2-D, MS-technology and in silico tools to analyze C. lunatus secretome discharged during interaction with potato leaf in the liquid phase. Our work provided first analysis of C. lunatus temperature-dependent secretome weaponries deployed during invasion in potato crop.

Plant growth and microorganism culturing conditions
Potato cv. Kufri Jyoti was grown in a plant growth chamber (U-CON250, Danihan Labtech Co., Ltd) at 20°C. The average light intensity was 180 μmolm −2 s −1 with photoperiod of 16 h light and 8 h darkness. Potato cv. Kufri Jyoti is widely cultivated in India and shows salient resistance features to Phytophthora infestans (http://nhb.gov.in/ vegetable/potato/pot013.pdf), is moderately susceptible to C. lunatus [4] and thermotolerant at 35°C [17]. C. lunatus strain btl5 (GenBanK® accession JX907828) was grown on V8 agar medium (Himedia®). The Czapek Dox Broth (CDB) medium composed of 30 g sucrose, 3 g NaNO 3 , 1 g K 2 HPO 4 , 0.5 g MgSO 4 , 0.5 g KCl, 0.01 g FeSO 4 and 500 mg chloramphenicol in 1 L water was used. The CDB medium was buffered with 100 mM of citric acid-sodium citrate buffer at pH 7.3. Only 10 mm diameter mycelia plug was inoculated in 100 ml CDB medium in a 250 ml conical flask. Five groups of treatments were established as follows. The first control flask contained only C. lunatus. Another control flask contained 3 g of disease-free potato leaf fragments devoid of C. lunatus. Treatment flasks contained C. lunatus and 3 g of diseasefree potato leaf fragments. These treatment flasks were incubated independently at 20°C, 28°C and 38°C for 2 weeks under the same photoperiodic conditions that plants were grown and shaken daily at 180 rpm for 10 min.

Biomass, pH variations and harvest of secreted proteins
After incubation, mycelia and conidia were removed by centrifugation at 13,000 g for 40 min at 4°C. The pH of the supernatant was determined for all the independent replicates using a pH meter (Eutech pH700, Thermo-Scientific®, Germany). Fresh weight of the interacting complex was measured on a sensitive balance (MicroBal-ance® C-35, ThermoScientific®, Germany) and 3 g was deducted. The 3 g is assumed to be the equivalent fresh weight of potato leaf added prior to interaction. Next, the complex matter was lyophilized and dry mass was measured in independent replicates. Subsequently, supernatant was chilled at -20°C for 2 h and secreted proteins were isolated by treating the supernatant with 1% sodium deoxycholate (w/v, Sigma®, Saint Louis, USA) and mixed by inversion. The protein complex was precipitated with 15% v/v solution of precipitating agent mixture (100% TCA: 100% acetone, 1:1% v/v) overnight at -20°C. After centrifugation (13,000 g, 30 min, 4°C), protein pellet was washed 5 times with pre-chilled extrapure acetone. Additional cleaning and depigmentation of protein was achieved using clean-up kit (Bio-Rad® laboratories, USA). The precipitates were air-dried for 30 min and dissolved in isoelectric focusing rehydration buffer containing 8 M urea, 2% CHAPS, 50 mM dithiothreitol (DTT), 0.2% (w/v) Bio-Lyte® ampholytes, and bromophenol blue trace (Bio-Rad® laboratories, USA). The protein concentration was determined by the dye-binding method [18]. We used bovine serum albumin for establishing the standard curve and protein aliquots were stored at -80°C till further use.

Test for leaf invasion in the liquid phase
In order to investigate whether C. lunatus established an intimate relationship with potato leaf in CDB medium during interaction, we aseptically removed intact leaf pieces from the reaction flask after 2 weeks of inoculation. Leaves were cleared in glacial acetic acid-ethanol (1:1% v/v) solution at 40°C overnight and rinsed in sterile water with four changes. Here, the chlorazole-black E-KOH staining technique [19] was used for studying the colonization of the abaxial leaf surface. In a randomized block design, we counted necrotic zones every 200 μm 2 for 10 leaf pieces. Intact leaf pieces were scarce and often difficult to handle for treatment at 38°C. The observation was performed with a microscope coupled with DP7M5.0.0.5 software and an Olympus DP70 camera (Olympus BX61®, USA).

Data analysis
One-way Anova associated with Tukey's HSD Post Hoc test were performed to determine the mean significant differences between treatments at P < 0.05. Data were computed in SPSS software v.20.

Two dimensional gel electrophoresis (2-D)
Aliquot of 140 μg of protein sample was used for rehydrating immobilized pH gradient strips (IPG; 7 cm) of pH gradient 4 to 7 (Bio-Rad® laboratories, USA) for 16 h in a passive mode. The pH 4-7 range was predetermined after trials with other focusing range for best resolution. Isoelectric focusing (IEF) was performed at 20°C for a total of 20 KVh using a default rapid ramp protocol on Protean®i12 IEF CELL (Bio-Rad® laboratories, USA). IPG strips were equilibrated twice for 40 min in equilibration buffer I [50 mM Tris-HCl pH 8.8, 6.5 M urea, 30% glycerol (v/v), traces of bromophenol blue and 2% DTT (w/v)] and in equilibration buffer II (50 mM Tris-HCl pH 8.8 and 2.5% iodoacetamide), respectively. The second dimension electrophoresis was performed at 16°C in a Mini-Protean® Tetra Cell (Bio-Rad® laboratories, USA) on a 15% resolving gel. The run was terminated when the dye front reached the lower end of the gel. Gels were calibrated with Precision-PlusProtein™ WesternC™ Standards (Bio-Rad® laboratories, USA). The gels were stained with Coomassie brilliant blue R250 (CBR) in a solution containing 50% methanol (v/v), 7% glacial acetic acid (v/v) and 0.3% CBR (w/v) overnight at 38°C. Subsequently, gels were destained adequately in a solution containing 30% methanol (v/v) and 7% glacial acetic acid (v/v) until visible spots appeared. Imaging was performed in Molecular Imager Versa DOCMp (Bio-Rad® laboratories, USA).

Image processing and data analysis
Quality control for gel images and statistical analyses were performed in Progenesis SameSpots v.4.1 suite (TotalLab®, USA). Spots with pixel intensity less than 120, spots in damaged areas and at the edge of the gel were excluded prior to nonlinear dynamics alignment. Spot volume (pixel-by-pixel intensity) were normalized as parts per million (ppm) of the total spot volume to determine the fold expression. Importantly, results were validated by performing pixel-to-pixel correlation analysis with an Anova P-value ≤0.05 at a fold expression cut-off value (F) ≥1.0. Here, differentially expressed spots and unique spots that did not align, judged not to be false positive based on eight gel runs were manually excised for downstream analysis.

Protein digestion and mass spectrometry
Protein digestion was performed as previously described [20]. Briefly, 0.45 μl of digested protein solution was mixed in 0.45 μl of α-cyano-4-hydroxycinnamic acid solution on matrix-assisted laser desorption/ionization timeof-flight/time-of-flight mass spectrometry (MALDI-TOF/ TOF MS) 4800 proteomics analyzer targeted plate. Peak lists were processed and exported through 4000 Series Explorer Software (Applied Biosystem, MA, USA) at default settings. Homology search was performed using MAS-COT v.2.3 (MatrixScience, London, UK) through Proteome Discoverer v.1.3.0.339 (ThermoScientific, Germany) against filtered predicted protein models (originated from expression sequence tag) from C. lunatus 20120521 m118 v2.0 genome peptides available at http://genome.jgi.doe. gov. The search parameters were: Enzyme, trypsin; Fixed modifications, carbamidomethyl (C); Variable modification, oxidation (M); Peptide mass tolerance, 40-100 ppm; Maximum missed cleavages, 2. The accepted MOWSE score threshold was inferred at P < 0.05. A few peptide peak lists that failed to match C. lunatus m118 v2.0 model genome peptides were queried against all updated entries from the NCBInr and Fungi MSDB sequence databases via in-house MASCOT server (v.2.3 MatrixScience, London, UK) using identical search parameters. In case of homologous proteins having similar MOWSE scores, we gave preference to proteins with best matched theoretical and experimental pI. False-discovery rate (FDR) [21] for the peptide search match was calculated using a decoy database (http://www.matrixscience.com/help/decoy_help. html). Here, we set FDR of 1% as a cut-off to export results from the analysis. Among the 60 spots excised from the gels that were analysed, only 39 were validated at FDR ≤1% and reported. Each step in the identification process was verified manually.
In silico characterization of secretome and de novo motif searches Signal peptide was predicted in SignalP 4.1 server [22]. A cut-off discriminatory score (D) was used to discern signal peptide with or without transmembrane (TM) network as follows. D = 0.45 for signal peptide without TM network and D = 0.50 for signal peptide with TM network. Subcellular localization of candidate effectors target was predicted at default settings in TargetP v.1 server [23]. Theoretical pI was predicted as earlier described [24] and theoretical molecular weight and protein net charge was predicted at http://www.encorbio. com/protocols/Prot-MW.htm. Glycosylphosphatidylinositol (GPI) anchored was predicted in big-pi web server [25]. Adhesin was predicted in Faapred [26] and Fun-galRV [27] web servers. Motif search was performed in MEME v.4.9 suite [28]. Here, we used sequences from C. lunatus model proteins only (elaborated in Additional file 1: Fasta) with significant hits (E-value <10 −5 ). The search was set to account for a maximum of 3 different motifs with 4 to 10 amino acids width with shuffling option. The motifs dispersion was set at zero or one per sequence. Identified motifs at P-value <10 −5 were retained and queried against GenBank® 23 fungal genomes peptides to determine the best likelihood matches using MAST v.4.9 suite [29]. The fidelity of the best hit motif was confirmed by scanning for the occurrences in 23 fungal genomes peptides using FIMO v.4.9 [30].

RNA isolation and gene expression analysis of candidate effectors
After a quick separation of mycelia mat from the leaf fragments at the same time points, total RNA was isolated from mycelium using TRIzol® reagent (Invitrogen, USA) following the manufacturer's protocol. Total RNA (1 μg) was treated with 1U RNAse-free DNase I (Amplification grade; Invitrogen, USA) and used in subsequent steps. The concentration of total RNA was determined by spectrophotometer (BioSpec-nano, Shimadzu®, Japan) and the integrity was assessed by performing denaturing formaldehyde-agarose gel electrophoresis. Gene expression analysis was performed by semi-quantitative reverse transcriptase-polymerase chain reaction (RT-PCR). The first-strand complementary DNA (cDNA) was synthesized using SuperScript® VILA™ cDNA synthesis kit (Invitrogen, USA) following the manufacturer's protocol. PCR was carried out using gene specific primers and expression was evaluated at the exponential phase of amplification. In brief, PCR was performed using 1 μl of cDNA template (diluted 10 times), 0.2 mM of dNTPs, 0.2 μM each of forward and reverse primers, 1 U of GoTaq® DNA Polymerase (Promega, USA) and 1X PCR buffer (Promega, USA) in a final volume of 25 μl. The thermal cycling conditions were 94°C for 4 min, 25-30 cycles of 94°C for 30 s, 55-56°C for 30 s, 72°C for 30 s and a final extension at 72°C for 7 min. Expression of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) gene was used as internal control to equalize cDNA quantities in all the reactions. All the primer sequences and annealing temperatures are elaborated (Additional file 2: Table S1).

Results
C. lunatus produced large biomass and modified the ambient pH C. lunatus produced floccose mycelia mat in CDB medium. Importantly, interaction of C. lunatus with potato leaf at 38°C produced the maximum significant biomass (6.27 g, P < 0.003, F = 18.08). Lowest biomass (1.82 g, P < 0.003, F = 18.08) was scored in the absence of potato leaf at 20°C ( Figure 1A). Curiously, significant changes in pH i.e. from 7.30 to 5.91 (P = 0.00, F = 102. 19) was observed between C. lunatus-leaf interaction at 20°C. In this study we selected 20°C because potato is a winter crop. Other temperatures such as 25°C, 28°C, 35°C and 38°C were tested in standardizing the experiment. When we compared the secretome maps of 25°C, 28°C, 35°C and 38°C with that of 20°C treatment in the presence of potato leaf, only 38°C produced a unique pattern of peptide spots. Therefore, we focused on 38°C as optimal heat-stress condition for downstream secretome analysis. The heat stress was aimed at providing an intuitive insight on the thermo-pathogenic nature of C. lunatus vis-à-vis potato host. Interestingly, in flask with potato leaf at 38°C, only a slight change in medium pH i.e. from 7.30 to 7.04 was scored against a significant biomass produced (6.27 g, P < 0.003, F = 18.08: Figure 1A, B). The result shows that C. lunatus modified the ambient pH during interaction with potato leaf as a function of temperature. That is, C. lunatus proliferated by changing the neutral medium (pH 7.30) to: 1) acidic medium (pH 5.91) at 20°C with potato leaf and 2) maintained a neutral medium (pH 7.04) at 38°C with potato leaf. This ability to modulate the pH of the interaction medium enables C. lunatus to secrete different set of candidate effectors. This was phenotypically manifested in the variable colors of conidia ( Figure 1C) and variable colors of the culture medium ( Figure 1D-G). No change in pH was observed in the control flask containing potato leaf without C. lunatus (data not shown), and we also failed to harvest proteins from the liquid medium.
C. lunatus invaded potato leaf in the liquid phase C. lunatus penetrated the leaf tissues during interaction. No appressoria-like structure was observed. Remarkably, the infectious hyphae penetrated the leaf tissues subcutaneously, forming necrotic epidermal zones ( Figure 2), dead epidermal cells ( Figure 2A) and pigmented hyphae on the leaf surface ( Figure 2B,C). Worth mentioning, the infectious melanized hyphae of C. lunatus at 20°C often penetrated via the epidermal anticlinal cell wall killing the surrounding cells as manifested by brownto-black tissue ( Figure 2D). Germinating conidia on the leaf-CDB interface ( Figure 2E) and germinating conidia in CDB medium ( Figure 2F-H) were often observed. As shown (Figure 2), two weeks after inoculation, colonization was an active ongoing process and C. lunatus intimately interacted with the potato leaf. A mean of 32.33 necrotic zones per 200 μm 2 of leaf tissue (P < 0.05; F = 8.57) was scored at 20°C. At the other extreme temperature, 57.67 necrotic zones per 200 μm 2 of leaf tissue (P < 0.05; F = 3.10) was scored at 38°C ( Figure 3).

C. lunatus produced diverse repertoire of secretomes
In the course of standardizing this experiment, we found that addition of 1% sodium deoxycholate (w/v) to the filtered CDB interaction medium without chilling at −20°C resulted in: 1) difficulties in eliminating pigments, 2) low number of distinct peptide spots, and 3) vertical striking on the gel. Hence, the interaction medium was chilled at -20°C for 2 h before treatment with 1% sodium deoxycholate (w/v). Importantly, when proteins were fully dried by lyophilization, we faced difficulties in re-solubilizing the samples. Therefore, in this protocol we do not recommend lyophilization of the protein sample prior to re-suspension in 2-D rehydration buffer (Bio-Rad® laboratories, USA). In order to check the quality of our extraction procedure we performed a one dimensional gel electrophoresis (1-D). The 1-D profile of C. lunatus total secretome interacting at 20°C with potato leaf, 28°C with potato leaf and 38°C with potato leaf indicated a differential expression pattern ( Figure 4). This result suggested that C. lunatus adapts to different temperature regimes during interaction with potato crop by differentially expressing diverse set of secretome ( Figure 4).
Following normalization of 2-D gels in Progenesis SameSpotv.4.1, the following reproducible peptide spot counts were present in at least 3 replicates per treatment: 1) 4 peptides spots without potato leaf (Additional file 3: Figure S1), 2) 92 peptide spots at 20°C in the presence of potato leaf ( Figure 5), 3) 46 peptide spots at 28°C in the presence of potato leaf (Additional file 4: Figure S2) and 4) 39 peptide spots at 38°C in the presence of potato leaf (Figures 6 and 7A). The plot showing positional normalized spot volumes on the gels are depicted (Additional file 5: Figure S3). As shown ( Figure 7A), C. lunatus differentially expressed less peptide spots at 38°C. A comparative analysis of the secretome maps for the treatment with potato leaf at 20°C and 38°C revealed that peptide spots# 11, 12, 28, 29, 30, 31, 32 and 39 were unique ( Figures 5 and 6). Thus, these peptide spots did not align in all gels (8 gels, 4 biological replications for each treatment, data not shown). Peptide spots# 28, 29, 30, 31, 32 and 39 were only observed at 38°C ( Figure 6). Overall, we observed that abundant peptide spots were of low molecular mass (Mr) ranging from 9.33-46 KDa; and having an experimental pI ranging from 4.26 to 6.62 on a 4-7 pI scale. Key features like homology, putative function etc. for the identified peptide spots are elaborated (Table 1:  Additional file 6: Table S2A, B).
Next, we performed correlation analysis with peptide spots that aligned to determine the changes in their expression using Progenesis SameSpot v.4.1. This analysis was  important since strong correlation indicated how these peptides are co-regulated and therefore, their involvement in virulence during invasion. Significantly expressed peptide spot# 13 (P = 2.e -0.6 , 2.3 folds, 21.48 KDa) was identified as pleckstrin-like related protein (  Figure 7B: Table 1: Additional file 6: Table S2A).
In the process of search performed against unique protein models predicted from C. lunatus m118 v.2   model genome peptides, 31 out of the 39 identified peptide spots produced 21 unique significant hits (Evalue <10 −5 ) and the profile for the 10 best hits is shown (Additional file 7: Figure S4). 5 out of the 31 peptide spots were without known functions. 8 out of 39 peptide spots failed to match with unique protein models from C. lunatus m118 v.2 model genome peptides. Of the 8 peptide spots, we performed combined searches in fungi MSDB and NCBInr data bases. Once more, 5 out of 8 peptide spots were without known putative functions (Additional file 6: Table S2B). Based on the putative functions of the identified peptides, 10 were unknown (8 up-regulated, 1 induced, 1 unique spot), 6 carbohydrate active-enzymes (3 up-regulated, 3 unique spots), 2 heat shock related proteins (1 up-regulated, 1 unique spot), 6 cellular traffickers (6 up-regulated), 4 protease/proteolysis related-proteins (3 up-regulated, 1 unique spot) and 11 nucleic acid metabolism related (5 upregulated, 6 unique spots) and their relative abundance is represented ( Figure 8A,B). Importantly, the peptides spot# 12, 24 and 26 that were significantly expressed had putative proteolytic activity. HSP70-like related proteins were peptide spot# 7 (P = 0.568; 1.2 folds; 45.23 KDa) and peptide spot# 9 (P = 0.445; 1.9 folds; 35.10 KDa).

C. lunatus used both classical and non-classical secretory pathways
Using SignalP4.1server [22], we found peptide spot# 38 (Cytochrome C oxidase related protein) was secreted through conventional pathway having an N-terminal signal peptide and forming TM-networks only (Additional file 6: Table S2A: Additional file 8: Figure S5). To address the potential destination of the proteins, Target P1.1 [23] algorithm was used. We found that 12 peptide spots could target the mitochondrion, 16 peptide spots could target any subcellular organelle and 11 peptide spots had no predictable targets (Additional file 6: Table  S2A, B). Throughout this study, no protein was predicted to have GPI-anchor motif (data not shown). This was a confirmation that the proteins observed in the secretome were not an outcome of fortuitous discharge from fungal cell surface or leaf tissue. Only 10 out of 31 proteins that matched unique protein models predicted from C. lunatus m118 v.2 genome peptides were found to have adhesin motifs (Additional file 6: Table S2A, B). Overall, only 10 out of 31 unique protein models from C. lunatus m118 v.2 model genome peptides identified in this study had a net negative charge (Table 1). This data revealed that proteins secreted during colonization of potato by C. lunatus are mostly positively charged.

C. lunatus secretomes on potato leaf contained a CLxxxxLHM-motif
Inspired by the lack of putative catalytic domains in some candidate effectors, we searched for conserved motifs in MEME v.4.9 [28].

We found that motifs [FY][MR][HY] V[AE]Y[PR]CM, CL[AK][TW]LHM and [WI][HG]N
[WE] were distributed on peptides sequences with nonoverlapping sites (P-value < 10 −5 ). Conserved amino acid residues in motif 01 were V, Y, C and M and organized in an xxxxxxVxxYxxCM pattern dispersed over the N-terminal ( Figure 9A). 'X' represents any variable amino acid residue in the pattern. Motif 02 contained C, L, L, H and M conserved amino acid residues organized in a CLxxxxLHM pattern and distributed within the range of 450 to 830 amino acids residues after the Nterminal cleavage site. Motif 03 contained a major conserve N (or asparagine) residue, W, H and W with four possible sites. This motif could be organized as xWxHNxW. Amino acids organization is elaborated in the motif logos ( Figure 9B). We tested the strength of motifs against GenBank® 23 fungal genomes peptides, essentially, cross kingdom pathogens such as Candida dubliniensis CD36 uid38659 (CD), Candida glabrata (CG), Cryptococcus gattii WM276 (CG*), Cryptococcus    [29] and occurrence rate using FIMO v.4.9 [30]. Motif 02 (CLxxxxLHM) produced the best fidelity match and occurrence hits (Additional file 9: Table S3). The median occurrence score of motif 02 was 260 per fungal genome peptides. Furthermore, motif 02 maximum and minimum hits of 470 and 56 was scored in Thielavia terrestris NRRL8126 and Encephalitozoon intestinalis ATCC 50506uid5160 genome peptides, respectively ( Figure 10).

Expression profiling of genes encoding secreted candidate effector proteins
The identification of differentially expressed peptide spots with unknown catalytic domains and other virulencerelated peptides prompted us to analyze their gene expression using semi-quantitative RT-PCR. Consistent and reproducible differential expressions were obtained for 16 candidates genes ( Figure 11A). The 5 unique putative C. lunatus effector candidates (ClEfc1, ClEfc2, ClEfc3, ClEfc4 and ClEfc5) with unknown catalytic domains were expressed at both 20°C and 38°C in the presence of potato leaf. Virulence related genes namely heat shock protein 70 (HSP70), scytalone dehydratase (SDR),  Figure 11A). Total RNA profile is shown ( Figure 11B).

Discussion
A great challenge in secretomics is how to trigger a pathogen to secrete proteins as starting material for downstream analysis. Studies based on genome scanning for secretome through in silico approach is less cumbersome and generates large data [15,[25][26][27][28]. Nevertheless, the likelihood that predicted proteins expressed are during host colonization is very low. Moreover, the physiological conditions for secretion cannot be predicted in silico.
During matrix-assisted laser desorption/ionization timeof-flight/time-of-flight mass spectrometry (MALDI-TOF/ TOF MS) we found that several peptide spots showed incongruities between their experimental and predicted molecular mass (Mr) and pI. These inconsistencies in experimental and theoretical values of pI and molecular masses have been observed in several proteomics studies [31][32][33][34]. While predicted values are based on whole protein sequence, experimental values are a function of protein subunits only detected on gels. Ecological niche can cause dramatic changes in experimental pI of proteins from acidic pI to basic pI [32]. Interestingly, it is reported that in vivo or in vitro protein degradation can result in experimental Mr significantly lower than predicted Mr [34]. This might hold true for this study given C. lunatus secreted proteolytic proteins (Table 1). Furthermore, posttranslational modifications such as ubiquitination, sumoylation, glycosylation, alternative splicing and endoproteolytic cleavage can cause discrepancies in predicted and expected pI and Mr values [33,35]. We found that C. lunatus modified the ambient pH. Some fungi have developed complicated regulatory mechanism to sense and respond to ambient pH dependent signals during prolonged interaction with the environment [36]. In addition, a few fungi secrete different set of proteins at different ambient pH [37]. For a successful invasion of a host, pathogens require coordinated secretion of chemical weaponries at the onset of interaction. For instance, an important hemibiotroph such as Botrytis cinerea differentially secretes enzymes in various plant tissues during invasion [16]. Cooperation of various extracellular proteins, though a few may be 'real' virulence factors [37] is decisive for the successful suppression of the host defense response. The changes in biomass and pH observed in this study at different temperature regimes demonstrated that C. lunatus successfully thrived on potato leaf by finely adjusting its microenvironment ( Figure 1A,B).
The secretome studies of B. cinerea revealed a dark colored culture medium at pH 4 and pH 6 [37]. This dark color was due to secondary metabolites produced during the fungus growth [37]. In the present secretome analysis, the intense yellowish-and milky-colored culture medium ( Figure 1D-G) and melanized colonizing hyphae ( Figure 2)  indicated that C. lunatus secretome comprises of colored secondary metabolites along with their candidate effectors. Previous study revealed that C. lunatus strains which produced high level of melanin are more virulent than those producing low level melanin [38]. Melanin and related secondary metabolites are virulence factors in C. lunatus pathogenicity [7,8,38]. A closer inspection of the colonizing hyphae ( Figure 2D) suggested that the darkbrown pigments might be an adaptive feature required to sustain C. lunatus virulence at higher temperature. This assumption is based on previous findings that proteins related to melanin synthesis and heat-stress responses are up-regulated in C. lunatus [7,8] during host colonization.
In this study, we found that C. lunatus invaded potato leaf in CDB medium. This suggested that CDB may not have negatively affected the interaction dynamics. Higher temperature-virulence phenomenon has been observed in Cryptococcus neoformans and Aspergillus fumigatus [39,40]. In these cross-kingdom fungi, loss of genes required for higher temperature growth resulted to attenuated virulence and at times death [39,40]. In our study, we found that conidia as well as hyphae at 28°C and 38°C were highly pigmented ( Figure 2). Additionally, more necrotic zones were formed around the melaninated penetrating hyphae at 38°C (Figure 3). On the other hand, conidia were present and actively germinating, an important prerequisite for persistent virulence [9,12]. Put together, we speculated that C. lunatus concomitantly sporulates, discharges toxic-colored metabolites ( Figure 1G) and produces diverse peptides during colonization of potato leaf (Figures 4, 5 and 6). Elsewhere, melaninated appressorial was reported in secretome analysis of Colletotrichum higginisianum, an important plant-pathogenic hemibiotroph [41]. Secreted secondary metabolites and candidate effectors can change the phenotypic characteristics of a host during interaction. Thus, this possibly explains why C. lunatus produces variable symptoms as earlier observed on different hosts [3,4,[8][9][10][11][12].
In 2-D analysis, some of the spots were unique at 20°C with potato leaf and did not match with spots detected at 38°C with potato leaf where negligible change in pH was observed ( Figures 5, 6 and 1B). Based on the changes in the secretome set (Figure 7), it is tempting to suggest that acidic microenvironment triggers C. lunatus to release more proteins. Also, we could not abandon the possibility that leaf protein underwent proteolysis during interaction. However, we failed to precipitate protein in treatments containing leaf without C. lunatus and also failed to find GPI-anchors on identified proteins. Based on these observations, we assumed that identified peptide spots were products of interaction secreted uniquely by C. lunatus. Nonetheless, the highest homology ranked peptide was CCO (spot# 38) secreted via classical pathway (Additional file 7: Figure S4 and Additional file 8: Figure S5). Adhesins are primary virulence factors used by pathogens to adhere to host prior to invasion [26,27] and only 10 peptide spots were found with adhesins motifs (Additional file 6: Table S2A, B). We concluded that majority of C. lunatus secreted proteins are highly soluble, positively charged (Table 1), discharged via the non-classical pathway and lack anchors to the cell surface.
Most phytopathogenic fungi progressively adapt to the genetic background of their host plants, thus creating a new type of virulent or physiological race [42]. This progressive process largely depends on the intensity of the pathogen-host specific interaction [43]. Nevertheless, during interaction, pathogenic fungi survived host counter attacks through coordinated secretion of their secretome weaponries primarily to harness nutrients and proliferate. Notably, the large number of up-regulated nucleic acid metabolic-related proteins, such as protein kinases ( Figure 8: Table 1) indicated that under heat-stress or biological stress (i.e. with leaf), C. lunatus prioritized its proliferation. This is in agreement with the concept that protein profile of microorganisms change under biological or abiotic stress condition [44], which is also observed in this study ( Figure 4). Intuitively, the abundant secretion of nucleic acid metabolic-related proteins enhances virulence as proliferation directly leads to large fungal biomass production. The secretome changes demonstrated that C. lunatus secretes varied repertoire of proteins depending on the ambient temperature (Figures 4 and 8). This is supported by unique unaligned peptide (spots# 27, 32, 34, 35, 36, 37 and 38) expressed only at 38°C (Figure 6: Additional file 6: Table  S2A, B). This possibly could explain the survival and successful lifestyle of C. lunatus and related species at higher temperature [9][10][11][12]. Cautiously, the unique peptide spots could represent signature proteins required to sustain C. lunatus at elevated temperature only, not necessarily involved in pathogenicity sensu stricto.
Besides the peptide spots with known putative functions, abundantly expressed peptide spots with no catalytic domain and no predictable subcellular target got our notice. C. lunatus is a successful cross-kingdom pathogen and would cogently not over-secrete peptide spots# 6, 19 and 21 at 20°C with potato leaf ( Figure 5: Table 1) and spot# 27 ( Figure 6) at 38°C with potato leaf without a possible role in pathogenicity. As the normalization plot depicts (Additional file 5: Figure S3), C. lunatus secreted discrete extracellular weaponries at different temperature. Assuming that secreted proteins operates in a cooperative manner [37], it is reasonable to suggest that C. lunatus candidate effectors with no assignable functions exert an unforeseen synergistic role in temperature-dependent virulence. Secreted proteins with unknown functions is a common phenomenon in fungal-plant interactions [14,31,37,41]. In recent years, progress has been made in effectoromics [36,[45][46][47][48][49][50][51]. Nonetheless, effectors are still ill-defined, but widely accepted that they manipulate host cell processes for the pathogen benefit [13][14][15][16].
Candidate effectors are understood to perform key functions prior to host colonization such as the provision of nutrients, mask recognition by the host defense system, cell-to-cell communication, detoxification of the environment, suppression of host-cell programmed cell death and killing of potential competitors [13,[45][46][47][48]. Therefore, fungal secretome analysis is important. For instance, mining protein effectors that matches specific resistance-proteins can assist the deployment of engineered resistance-genes in crops and can also permit scientist to survey crops when a particular pathotype overcome the resistance-genes [47,48]. Importantly, effectors can be used as molecular probes to dissect focal immune responses, unravel the diversity of secretory vesicles and the cargo they transport [46]. We found that the secretome of C. lunatus on potato leaf shared motif-CLxxxxLHM (Figure 9). Cysteine is important and maintains effector stability in the extracellular space [36]. In addition, we observed a significant occurrence rate of CLxxxxLHM motif in other GenBank® fungal genome peptides ( Figure 10). This shows that cross-kingdom pathogens share some common pathogenicity signatures in their secretome. For instance, in powdery mildew and rust fungi, motif Y/F/WxC [49] and Oomycetes effector motif RxLR-dEER [45,50] shared no positional conserved amino acid residues. Noteworthy, these pathogens colonize only plant hosts. Intriguingly, C. lunatus colonizes plant hosts [4][5][6][7][8] and animal hosts [1,2]. Critically, motif-CLxxxxLHM share features to motif-Y/F/WxC and motif-RxLR with conserved residue of cysteine (C) and leucine (L), respectively. In addition, histidine (H)-residue like arginine (R) is a basic amino acid residue and is rare in effector motif, but reported elsewhere [51]. Motifs with positional constraint amino acids such as M, L, H, N, Y, F, W and C have been reported in other fungi candidate effectors [49][50][51][52]. A closer inspection suggests histidine in motif-CLxxxxLHM is a unique signature in C. lunatus secretome on potato crop during invasion.
Pioneering effectoromics study indicated that fungal candidate effectors are most likely to be soluble and secreted into the extracellular matrix but do not form cross-link with fungal cell wall and show no homology to proteins with known functions [41]. Based on these criteria, we concluded that spots# 3 (C. lunatus genome peptide m118 v.2 accession no: 118437|e_gw1.13.184.1/ ClEfc1), spot# 6 (19257|fgenesh1_pg.7_#_229/ClEfc2), spot# 19 (46252|estExt_Genemark1.C_210087/ClEfc3), spot# 21 (64448|estExt_fgenesh1_pg.C_160063/ClEfc4) and spot# 27 (132047|estExt_Genewise1.C_4_t40098/ ClEfc5) (Table 1: Figure 11) with no homology to known proteins are candidate effectors of C. lunatus on potato crop. We found candidate effectors ClEfc1, ClEfc2, ClEfc3, ClEfc4 and ClEfc5 were secreted via non-classical pathway. As such, the candidate effectors did not have Nterminal signal peptide in agreement with conditions for non-conventional secretions [15,53]. Sophistication in C. lunatus invasion strategies is further enhanced by its ability to use the classical and non-classical secretory pathways to explore its microenvironment. Since only spot# 38 (Additional file 8: Figure S5) was identified with a TMnetworking, this agrees with previous finding that many fungal effectors are soluble and do not often cross-linked with the fungal cell wall after secretion [41]. Our findings demonstrate that C. lunatus secretes majority of its effectors via non-classical pathway.
At 20°C with potato leaf, known virulence related proteins such as HSP70-related protein [7,41], tetratrico-peptide (TPR)-related protein [54] and pleckstrin-like proteins were up-regulated. Based on the observation that the expression fold of HSP70-related proteins was greater than the cut-off limit, we concluded that it plays a decisive role in C. lunatus virulence on potato as previously reported in maize [7]. TPR-like proteins have a versatile helix associated with protein-protein interaction and intercellular trafficking [54] and a virulence domain [55]. Pleckstrin-like proteins controls diverse cellular trafficking interrelated with cell growth, survival, proliferation and metabolism [56]. In Candida albicans, pleckstrin-like proteins are understood to be involved in virulence and morphological switching [57]. Other putative proteins including protein kinases (spot# 1, 5 and 28), oxido-reductase (OXR, spot# 4) and cytochrome c oxidase (CCO, spot#38) were up-regulated. Although similar up-regulation was observed at the transcript level, their role in C. lunatus-increased temperature dependent-virulence is not clear ( Figure 11A). Also, C. lunatus discharged an array of extracellular proteins that are often related with carbohydrate metabolism and proteolysis. Proteases play an important role in pathogenicity and maintenance of life cycle of fungi [37,46,58] and in this study, subtilisin protease (Pte) was constitutively expressed ( Figure 11A). Toxin production in C. lunatus is believed to be positively correlated with up-regulation of NTR [7,8]. Furthermore, SCD, SDR and NTR are important genes for melanin biosynthetic pathway [7,8,41]. It was showed that up-regulation of SCD and NTR increased melanin pigments [7,8] and also improved the penetration strength of melanized appressorial [41]. Therefore, up-regulation of SCD, SDR and NTR might play important role in C. lunatus temperature-dependent virulence ( Figure 11A). However, an evident up-regulation of intertrafficking transcripts such as WD40, Grlp and ALP was observed ( Figure 11A) which requires detailed investigation to decipher their possible role. Meanwhile, we suggest that WD40, Grlp and ALP may synergistically cooperate with candidate effectors ClEfc1, ClEfc2, ClEfc3, ClEfc4 and ClEfc5 to cause disease on potato.

Conclusion
Classical 2-D is a powerful approach for direct analysis of secretome [16,37] and for the discovery of virulence factors in plant-pathogen interactions [7,8,[58][59][60][61]. In this study, we have successfully used 2-D, MS-technology, RT-PCR and in silico tools to identify a set of 5 soluble candidate effectors of unknown function from C. lunatus secreted against potato crop at different temperature regimes. The findings throw light on the increased temperaturedependent virulence of C. lunatus in accordance with recent findings related to the release of pigments on maize and overexpression of stress related proteins [7]. Importantly, the discrete set of secretome provided new insight that can explain why C. lunatus can successfully thrives on a broad-spectrum of hosts under heat-stress conditions. The results showed that the sophisticated infection process of potato crop by C. lunatus comprises of: 1) intimate penetration of the leaf, 2) modulation of the ambient microenvironment pH, 3) release of colored secondary metabolites and 4) secretion of effectors candidate with diverse functions. In future, it will be interesting to integrate the analysis of secondary metabolites and the current secretome data to confirm the hypothesis generated in this study on C. lunatus thermo-pathogenicity.