Proteomic Analysis of Phytophthora infestans Reveals the Importance of Cell Wall Proteins in Pathogenicity
2017; Elsevier BV; Volume: 16; Issue: 11 Linguagem: Inglês
10.1074/mcp.m116.065656
ISSN1535-9484
AutoresSvante Resjö, Maja Brus-Szkalej, Ashfaq Ali, H.J.G. Meijer, Marianne Sandin, Francine Govers, Fredrik Levander, Laura J. Grenville‐Briggs, Erik Andréasson,
Tópico(s)Potato Plant Research
ResumoThe oomycete Phytophthora infestans is the most harmful pathogen of potato. It causes the disease late blight, which generates increased yearly costs of up to one billion euro in the EU alone and is difficult to control. We have performed a large-scale quantitative proteomics study of six P. infestans life stages with the aim to identify proteins that change in abundance during development, with a focus on preinfectious life stages. Over 10 000 peptides from 2061 proteins were analyzed. We identified several abundance profiles of proteins that were up- or downregulated in different combinations of life stages. One of these profiles contained 59 proteins that were more abundant in germinated cysts and appressoria. A large majority of these proteins were not previously recognized as being appressorial proteins or involved in the infection process. Among those are proteins with putative roles in transport, amino acid metabolism, pathogenicity (including one RXLR effector) and cell wall structure modification. We analyzed the expression of the genes encoding nine of these proteins using RT-qPCR and found an increase in transcript levels during disease progression, in agreement with the hypothesis that these proteins are important in early infection. Among the nine proteins was a group involved in cell wall structure modification and adhesion, including three closely related, uncharacterized proteins encoded by PITG_01131, PITG_01132, and PITG_16135, here denoted Piacwp1–3. Transient silencing of these genes resulted in reduced severity of infection, indicating that these proteins are important for pathogenicity. Our results contribute to further insight into P. infestans biology, and indicate processes that might be relevant for the pathogen while preparing for host cell penetration and during infection. The mass spectrometry data have been deposited to ProteomeXchange via the PRIDE partner repository with the data set identifier PXD002446. The oomycete Phytophthora infestans is the most harmful pathogen of potato. It causes the disease late blight, which generates increased yearly costs of up to one billion euro in the EU alone and is difficult to control. We have performed a large-scale quantitative proteomics study of six P. infestans life stages with the aim to identify proteins that change in abundance during development, with a focus on preinfectious life stages. Over 10 000 peptides from 2061 proteins were analyzed. We identified several abundance profiles of proteins that were up- or downregulated in different combinations of life stages. One of these profiles contained 59 proteins that were more abundant in germinated cysts and appressoria. A large majority of these proteins were not previously recognized as being appressorial proteins or involved in the infection process. Among those are proteins with putative roles in transport, amino acid metabolism, pathogenicity (including one RXLR effector) and cell wall structure modification. We analyzed the expression of the genes encoding nine of these proteins using RT-qPCR and found an increase in transcript levels during disease progression, in agreement with the hypothesis that these proteins are important in early infection. Among the nine proteins was a group involved in cell wall structure modification and adhesion, including three closely related, uncharacterized proteins encoded by PITG_01131, PITG_01132, and PITG_16135, here denoted Piacwp1–3. Transient silencing of these genes resulted in reduced severity of infection, indicating that these proteins are important for pathogenicity. Our results contribute to further insight into P. infestans biology, and indicate processes that might be relevant for the pathogen while preparing for host cell penetration and during infection. The mass spectrometry data have been deposited to ProteomeXchange via the PRIDE partner repository with the data set identifier PXD002446. Phytophthora infestans is a devastating oomycete plant pathogen that causes the disease late blight on potato and several related plants. It originated in South or Central America and reached Europe during the middle of the nineteenth century. The first described major outbreak culminated in the Great Irish Famine of 1845–1852 (1.Agrios G.N. Plant pathology. Elsevier Academic Press, Amsterdam2004Google Scholar). Since then, it has been the most damaging pathogen for potato growers in Europe and North America. The costs of reduction in yield due to late blight and of measures to control the disease have been estimated to more than one billion euro in the EU alone (2.Haverkort A. Boonekamp P. Hutten R. Jacobsen E. Lotz L. Kessel G. Visser R. van der Vossen E. Societal costs of late blight in potato and prospects of durable resistance through cisgenic modification.Potato Res. 2008; 51: 47-57Crossref Scopus (306) Google Scholar). Current strategies for late blight control largely depend on regular application of a combination of fungicides. Developing high quality potato varieties with durable resistance has not been very successful. P. infestans has the capacity to rapidly change its repertoire of effectors and thereby escape recognition by resistant potato varieties. Phytophthora infestans belongs to the oomycetes, a class in the stramenopile lineage that together with the alveolate and Rhizaria lineages forms SAR, a supergroup that is distinct from animals, plants and fungi (3.Burki F. The eukaryotic tree of life from a global phylogenomic perspective.Cold Spring Harb. Perspect. Biol. 2014; 6: a016147Crossref PubMed Scopus (237) Google Scholar). The asexual life cycle of P. infestans comprises six life stages (Fig. 1A) (4.Judelson H.S. Blanco F.A. The spores of Phytophthora: weapons of the plant destroyer.Nat. Rev. Microbiol. 2005; 3: 47-58Crossref PubMed Scopus (295) Google Scholar). Hyphae grown on agar medium or colonizing an infected leaf produce sporangiophores with air-borne sporangia, which in turn release motile zoospores. When the zoospores touch a barrier, they encyst. The cysts germinate, and when the tip of the germ tube encounters a hydrophobic surface like a leaf surface, an appressorium is formed. The appressorium adheres to the surface and is triggered to form a penetration peg that pierces the cuticle and penetrates the epidermal cell where an infection vesicle is formed. Thereafter, hyphae grow into the mesophyll layer, where they form an intercellular hyphal network and produce feeding structures called haustoria that penetrate the mesophyll cells to maximize nutrient uptake from the host. As the lesion expands P. infestans switches to a more destructive necrotrophic mode of growth in the center of lesion, while continuing to grow as a biotroph at the edges. Depending on the conditions, it takes about 4–7 days before sporangiophores emerge from the stomata and produce sporangia, which will spread and initiate new infections. The fact that oomycetes and fungi, two major groups of filamentous plant pathogens, are not closely related implies that comparative analyses of oomycetes and fungi is limited in drawing conclusions about the molecular mechanisms underlying infection. This makes it important to acquire new information about the infection process at a molecular level. An improved understanding of the infection process is useful for the development of cultivars with durable resistance as well as new fungicides. A characterization of mRNA or protein abundance in different life stages can contribute important information about what differentiates the life stages from each other, and thus indicate which properties are critical during infection. Judelson et al. (5.Judelson H.S. Ah-Fong A.M. Aux G. Avrova A.O. Bruce C. Cakir C. da Cunha L. Grenville-Briggs L. Latijnhouwers M. Ligterink W. Meijer H.J. Roberts S. Thurber C.S. Whisson S.C. Birch P.R. Govers F. Kamoun S. van West P. Windass J. Gene expression profiling during asexual development of the late blight pathogen Phytophthora infestans reveals a highly dynamic transcriptome.Mol. Plant Microbe Interact. 2008; 21: 433-447Crossref PubMed Scopus (86) Google Scholar) performed a microarray analysis of several life stages using a set of expressed sequence tags (ESTs) 1The abbreviations used are: EST, Expressed Sequence Tag; MGF, Mascot Generic Format; PSM, Peptide-Spectrum Match; dpi, days post inoculation; RNAi, RNA interference; ROS, reactive oxygen species; CD36, Cluster of Differentiation 36. 1The abbreviations used are: EST, Expressed Sequence Tag; MGF, Mascot Generic Format; PSM, Peptide-Spectrum Match; dpi, days post inoculation; RNAi, RNA interference; ROS, reactive oxygen species; CD36, Cluster of Differentiation 36. generated by Randall et al. (6.Randall T.A. Dwyer R.A. Huitema E. Beyer K. Cvitanich C. Kelkar H. Fong A.M. Gates K. Roberts S. Yatzkan E. Gaffney T. Law M. Testa A. Torto-Alalibo T. Zhang M. Zheng L. Mueller E. Windass J. Binder A. Birch P.R. Gisi U. Govers F. Gow N.A. Mauch F. van West P. Waugh M.E. Yu J. Boller T. Kamoun S. Lam S.T. Judelson H.S. Large-scale gene discovery in the oomycete Phytophthora infestans reveals likely components of phytopathogenicity shared with true fungi.Mol. Plant-Microbe Interact. 2005; 18: 229-243Crossref PubMed Scopus (122) Google Scholar). They found differences in expression of genes with several different functions, such as metabolism, regulation of DNA synthesis, cellular structure, pathogenicity, as well as several genes encoding known effector proteins. Previous proteomics studies focused on P. infestans have been rather small-scale 2D gel based experiments. In two studies, screening of ∼200 spots allowed identification of a small number of specific proteins that were more abundant during appressorium formation and were involved in amino acid biosynthesis or cellulose synthesis or that exhibited changes in abundance during preinfection stages (7.Grenville-Briggs L.J. Anderson V.L. Fugelstad J. Avrova A.O. Bouzenzana J. Williams A. Wawra S. Whisson S.C. Birch P.R.J. Bulone V. van West P. Cellulose synthesis in Phytophthora infestans is required for normal appressorium formation and successful infection of potato.Plant Cell. 2008; 20: 720-738Crossref PubMed Scopus (109) Google Scholar, 8.Grenville-Briggs L.J. Avrova A.O. Bruce C.R. Williams A. Whisson S.C. Birch P.R. van West P. Elevated amino acid biosynthesis in Phytophthora infestans during appressorium formation and potato infection.Fungal Genet. Biol. 2005; 42: 244-256Crossref PubMed Scopus (88) Google Scholar). In another study Grenville-Briggs et al. (9.Grenville-Briggs L.J. Avrova A.O. Hay R.J. Bruce C.R. Whisson S.C. van West P. Identification of appressorial and mycelial cell wall proteins and a survey of the membrane proteome of Phytophthora infestans.Fungal Biol. 2010; 114: 702-723Crossref PubMed Scopus (30) Google Scholar) specifically analyzed proteins from the cell wall of sporulating mycelium, nonsporulating mycelium and appressoria (9.Grenville-Briggs L.J. Avrova A.O. Hay R.J. Bruce C.R. Whisson S.C. van West P. Identification of appressorial and mycelial cell wall proteins and a survey of the membrane proteome of Phytophthora infestans.Fungal Biol. 2010; 114: 702-723Crossref PubMed Scopus (30) Google Scholar) using LC-MS/MS and could identify four proteins as unique to the P. infestans appressorium cell wall. Proteomics studies on other Phytophthora spp. include the study of Savidor et al. who used a label-free approach to compare mycelium and germinating cysts in P. sojae and P. ramorum and could identify several candidate proteins distinguishing the preinfection stage from the mycelial stage (51). In addition, Meijer et al. performed an analysis of proteins from isolated P. ramorum cell walls. They identified 17 proteins, mostly mucins or mucin-like proteins, and glycoside hydrolases, but also two transglutaminases and one elicitin (10.Meijer H.J.G. van de Vondervoort P.J.I. Yin Q.Y. de Koster C.G. Klis F.M. Govers F. de Groot P.W.J. Identification of cell wall-associated proteins from Phytophthora ramorum.Mol. Plant Microbe In. 2006; 19: 1348-1358Crossref PubMed Scopus (68) Google Scholar). These studies on P. sojae and P. ramorum did not consider appressoria. To acquire a larger data set of proteins spanning the major stages of development in P. infestans, we have performed a large-scale quantitative analysis of P. infestans proteins. In this study, we compare the abundance of more than 2000 proteins in six life stages. By combining proteomics with techniques for validation using RT-qPCR and RNA interference, we further confirm the abundances and functional role of selected proteins. P. infestans strain 88069 was cultivated on rye sucrose medium (11.Caten C.E. Jinks J.L. Spontaneous Variability of Single Isolates of Phytophthora Infestans .I. Cultural Variation.Can. J. Botany. 1968; 46: 329-348Crossref Google Scholar). Hyphae, sporangia, zoospores, cysts, germinated cysts and appressoria were isolated essentially as described by Resjö et al. (12.Resjo S. Ali A. Meijer H.J. Seidl M.F. Snel B. Sandin M. Levander F. Govers F. Andreasson E. Quantitative label-free phosphoproteomics of six different life stages of the late blight pathogen Phytophthora infestans reveals abundant phosphorylation of members of the CRN effector family.J. Proteome Res. 2014; 13: 1848-1859Crossref PubMed Scopus (21) Google Scholar) and Grenville-Briggs et al. (8.Grenville-Briggs L.J. Avrova A.O. Bruce C.R. Williams A. Whisson S.C. Birch P.R. van West P. Elevated amino acid biosynthesis in Phytophthora infestans during appressorium formation and potato infection.Fungal Genet. Biol. 2005; 42: 244-256Crossref PubMed Scopus (88) Google Scholar). Hyphae were grown for 5 days in liquid rye sucrose medium, harvested and washed briefly with water before freezing in liquid nitrogen. To produce sporangia, zoospores, cysts, germinated cysts, and appressoria, P. infestans was grown on rye sucrose agar plates. Sporangia were harvested after 14 days by flooding the plates with cold water and gentle rubbing. The sporangial suspension was removed and filtered through a 40-μm nylon mesh to remove hyphal remnants. Zoospore release was induced by incubation of sporangia at 10 °C for 2–3 h. Encystment was induced by vortexing for 2 min. Sporangia, zoospores and cysts were harvested by centrifugation (13,000 × g, 1 min). Cysts were incubated at 10 °C in Petri dishes to induce germination and formation of appressoria. Germinated cysts were harvested by centrifugation after 3–4 h of incubation. Appressoria were harvested after 24 h by removing the water, adding 1 ml of extraction buffer and scraping with a cell scraper. All samples were frozen after isolation. Hyphal samples were ground in a mortar with liquid nitrogen. The other life stages were ground with a plastic pestle in an Eppendorf tube. Proteins from all P. infestans preparations were extracted in a buffer consisting of 50 mm Tris-HCl, pH 7.4, 2 mm EDTA, 0,25% SDS, 50 mm sodium pyrophosphate, 1 mm sodium fluoride, 50 μm sodium orthovanadate, 5 nm Calyculin A, 1 mm phenylmethanesulfonyl fluoride (PMSF) and 20 μm leupeptin, and separated on a short SDS-PAGE. The entire lane was excised, washed and the proteins digested with trypsin (Promega Trypsin Gold, Mass Spectrometry Grade). The tryptic digest was desalted using 200 μl C18 StageTips (Proxeon Biosystems) according to the manufacturer's instructions. The tryptic digest, corresponding to 10 μg protein loaded on the preparative SDS-PAGE, was subjected to HPLC-MS/MS analysis using an Eksigent nanoLC2D HPLC system coupled to an LTQ Orbitrap XL ETD. The peptides were loaded onto a precolumn (Agilent Zorbax 300SB-C18, 0.3 mm ID, 5 mm, 5 μm particle size) connected to an in-house packed picofrit column (Agilent Zorbax 300SB C18, 75 μm ID, 150 mm, 3.5 μm particle size). The analytical column was preequilibrated for 10 min using buffer consisting of 0.1% formic acid (FA), 5% ACN at a flow rate of 10 μl/min and the peptides were separated in an 0.1% FA buffer using a 55 min linear gradient from 5% to 40% ACN followed by a 5 min linear gradient from 40% to 80% ACN, at a flow rate of 350 nl/min. The eluted peptides were analyzed online using an LTQ Orbitrap XL ETD. The Orbitrap was operated in data dependent mode to automatically perform Orbitrap-MS and LTQ-MS/MS analysis. Survey scan spectra (400–2000 Da) were acquired using the Orbitrap mass analyzer with the resolution r = 60,000. Automatic gain control was enabled (Target value for LTQ MSn was 1 × 104 and the target value for FT MS was 5 × 105). The seven most intense ions were selected for fragmentation in the LTQ, using a mass window of 2 Da for precursor ion selection. The precursor ions were fragmented with a normalized collision energy of 35 (with activation Q set to 0.25 and an activation time of 30 ms). Dynamic exclusion was enabled with a repeat count of 2, a repeat duration of 20 s, an exclusion duration of 120 s, an exclusion list size of 499 and a 10 ppm exclusion mass width relative to both low and high. Biological replicates were from separate cultures, totaling six hyphal, four sporangial, four zoospore, four cyst, three germinated cyst and six appressorial biological replicates. Each biological replicate was analyzed once, except for one of the hyphal samples, which was reanalyzed because of suspected poor HPLC performance, and all available replicates were used for peptide identification and feature matching. For the quantitative analysis, three samples were excluded because of poor HPLC performance, low protein content and contamination, respectively. Further details about the reason for excluding these files can be found in supplemental Table S6. Given this design, and the label-free workflow with MS1 precursor matching that was used, there was no need for repeated technical replicates. The statistical analysis of the proteomics data was performed using the software programs Qlucore Omics Explorer 3.0 (Qlucore, Lund, Sweden) and DanteR (13.Taverner T. Karpievitch Y.V. Polpitiya A.D. Brown J.N. Dabney A.R. Anderson G.A. Smith R.D. DanteR: an extensible R-based tool for quantitative analysis of -omics data.Bioinformatics. 2012; 28: 2404-2406Crossref PubMed Scopus (114) Google Scholar). Qlucore was used as a tool to identify abundance profiles among the MS peptide features. Feature data was imported into Qlucore, log transformed and a heat map was generated in which samples and MS features were ordered by hierarchal clustering. Peptides from different life stages were analyzed using Qlucore's Multi Group comparison, (ANOVA with an F-test), followed by the adjustment of the p values for multiple comparisons using the Benjamini-Hochberg procedure for calculating the q-value (14.Benjamini Y. Hochberg Y. Controlling the false discovery rate - a practical and powerful approach to multiple testing.J. Roy Stat. Soc. B. Met. 1995; 57: 289-300Google Scholar, 15.Reiner A. Yekutieli D. Benjamini Y. Identifying differentially expressed genes using false discovery rate controlling procedures.Bioinformatics. 2003; 19: 368-375Crossref PubMed Scopus (1335) Google Scholar). Only peptides with q < 0.01 were used for the heat map. Protein-level statistical analysis was performed in DanteR. The method called "Model Based Filter/Impute/Anova" in Dante R is a statistical model for filtering and imputation of missing values, followed by ANOVA (16.Karpievitch Y. Stanley J. Taverner T. Huang J. Adkins J.N. Ansong C. Heffron F. Metz T.O. Qian W.J. Yoon H. Smith R.D. Dabney A.R. A statistical framework for protein quantitation in bottom-up MS-based proteomics.Bioinformatics. 2009; 25: 2028-2034Crossref PubMed Scopus (119) Google Scholar). The method named "Protein-level Anova" in DanteR is an ANOVA-based method that combines peptide data to generate p values at the protein level and is described in detail in (17.Oberg A.L. Mahoney D.W. Eckel-Passow J.E. Malone C.J. Wolfinger R.D. Hill E.G. Cooper L.T. Onuma O.K. Spiro C. Therneau T.M. Bergen 3rd., H.R. Statistical analysis of relative labeled mass spectrometry data from complex samples using ANOVA.J. Proteome Res. 2008; 7: 225-233Crossref PubMed Scopus (142) Google Scholar). ANOVA-based statistics were used because log-transformed peptide intensities in general follow a normal distribution. The raw data from the Orbitrap were converted to Mascot Generic Format (MGF) and mzML (18.Martens L. Chambers M. Sturm M. Kessner D. Levander F. Shofstahl J. Tang W.H. Rompp A. Neumann S. Pizarro A.D. Montecchi-Palazzi L. Tasman N. Coleman M. Reisinger F. Souda P. Hermjakob H. Binz P.A. Deutsch E.W. mzML–a community standard for mass spectrometry data.Mol. Cell. Proteomics. 2011; 10Abstract Full Text Full Text PDF PubMed Scopus (451) Google Scholar) using ProteoWizard (version 2.1.2430) (19.Kessner D. Chambers M. Burke R. Agus D. Mallick P. ProteoWizard: open source software for rapid proteomics tools development.Bioinformatics. 2008; 24: 2534-2536Crossref PubMed Scopus (1217) Google Scholar). The Proteios software environment (2.20 dev build 4523) (20.Hakkinen J. Vincic G. Mansson O. Warell K. Levander F. The proteios software environment: an extensible multiuser platform for management and analysis of proteomics data.J. Proteome Res. 2009; 8: 3037-3043Crossref PubMed Scopus (83) Google Scholar) was used to search the MGF files with Mascot (version 2.3.01) and X!Tandem ("Jackhammer" 2013.06.15) against a database consisting of all P. infestans proteins in UniProt as of 2010–04-22, concatenated with an equal size decoy database (random protein sequences with conserved protein length and amino acid distribution, in total 36,512 target and decoy protein entries) (21.Elias J.E. Gygi S.P. Target-decoy search strategy for increased confidence in large-scale protein identifications by mass spectrometry.Nat. Methods. 2007; 4: 207-214Crossref PubMed Scopus (2826) Google Scholar). Search tolerances were set to 7 ppm for MS and 0.5 Da for MS/MS. The enzyme used to generate the peptides was trypsin and one missed cleavage was allowed. Carbamidomethylation of cysteine residues was selected as a fixed modification and oxidation of methionine residues was selected as a variable modification. Search results were exported from Mascot as XML, and results, including the top ranked peptide for each spectrum, were imported to Proteios, where the peptide-spectrum level search results from X-tandem and Mascot were combined, and q values were calculated using the target-decoy method as described by Käll et al. (22.Kall L. Storey J.D. MacCoss M.J. Noble W.S. Assigning significance to peptides identified by tandem mass spectrometry using decoy databases.J. Proteome Res. 2008; 7: 29-34Crossref PubMed Scopus (441) Google Scholar). The search results were then filtered at a q-value of 0.01, to obtain a peptide-spectrum-match (PSM) false discovery rate of 1% in the filtered list. The mass, charge and peptide identification scores for all identified peptides can be found in supplemental Table S5. The MS proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) (23.Vizcaino J.A. Cote R.G. Csordas A. Dianes J.A. Fabregat A. Foster J.M. Griss J. Alpi E. Birim M. Contell J. O'Kelly G. Schoenegger A. Ovelleiro D. Perez-Riverol Y. Reisinger F. Rios D. Wang R. Hermjakob H. The PRoteomics IDEntifications (PRIDE) database and associated tools: status in 2013.Nucleic Acids Res. 2013; 41: D1063-D1069Crossref PubMed Scopus (1594) Google Scholar) via the PRIDE partner repository with the data set identifier PXD002446. For quantitative analysis, a label free approach based on precursor intensities was used (24.Sandin M. Krogh M. Hansson K. Levander F. Generic workflow for quality assessment of quantitative label-free LC-MS analysis.Proteomics. 2011; 11: 1114-1124Crossref PubMed Scopus (28) Google Scholar) with all data processing steps performed within Proteios. The feature detection step was performed on mzML files using msInspect (25.Bellew M. Coram M. Fitzgibbon M. Igra M. Randolph T. Wang P. May D. Eng J. Fang R. Lin C. Chen J. Goodlett D. Whiteaker J. Paulovich A. McIntosh M. A suite of algorithms for the comprehensive analysis of complex protein mixtures using high-resolution LC-MS.Bioinformatics. 2006; 22: 1902-1909Crossref PubMed Scopus (225) Google Scholar) and subsequent feature matching and alignment between LC-MS/MS runs with a previously described workflow (26.Sandin M. Ali A. Hansson K. Mansson O. Andreasson E. Resjo S. Levander F. An adaptive alignment algorithm for quality-controlled label-free LC-MS.Mol. Cell. Proteomics. 2013; 12: 1407-1420Abstract Full Text Full Text PDF PubMed Scopus (29) Google Scholar). The quantitative, un-normalized data for all peptides used for the quantitative analysis is shown in supplemental Table S1. The resulting peptide data was linearly normalized against the median sum of all features in all conditions. The normalized data was analyzed using the software programs Qlucore Omics Explorer 3.0 (Qlucore, Lund, Sweden) and DanteR (13.Taverner T. Karpievitch Y.V. Polpitiya A.D. Brown J.N. Dabney A.R. Anderson G.A. Smith R.D. DanteR: an extensible R-based tool for quantitative analysis of -omics data.Bioinformatics. 2012; 28: 2404-2406Crossref PubMed Scopus (114) Google Scholar). Qlucore was used as a tool to identify abundance profiles among the MS features. Feature data was imported into Qlucore, log transformed and a heat map was generated in which samples and MS features were ordered by hierarchal clustering. Peptides from different life stages were analyzed using Qlucore's Multi Group comparison, (an Anova with an F-test), followed by the adjustment of the p values for multiple comparisons using the Benjamini-Hochberg procedure for calculating the q-value (14.Benjamini Y. Hochberg Y. Controlling the false discovery rate - a practical and powerful approach to multiple testing.J. Roy Stat. Soc. B. Met. 1995; 57: 289-300Google Scholar, 15.Reiner A. Yekutieli D. Benjamini Y. Identifying differentially expressed genes using false discovery rate controlling procedures.Bioinformatics. 2003; 19: 368-375Crossref PubMed Scopus (1335) Google Scholar). Only peptides with q < 0.01 were used for the heat map. The resulting heat map (shown in Fig. 1B) was then manually inspected to select abundance profiles of interest. Eight profiles that were considered of biological interest, or that were highly abundant were selected for protein level analysis. The abundance profiles and the rules for including a protein in them are described in detail in Table I. The DanteR software was used for protein-level analysis. In DanteR, the data was log transformed and then subjected to two different analyses. First, the "Model Based Filter/Impute/Anova" feature of DanteR was used. Because this procedure has been shown to filter out proteins that are unique to one condition (27.Webb-Robertson B.J. McCue L.A. Waters K.M. Matzke M.M. Jacobs J.M. Metz T.O. Varnum S.M. Pounds J.G. Combined statistical analyses of peptide intensities and peptide occurrences improves identification of significant peptides from MS-based proteomics data.J. Proteome Res. 2010; 9: 5748-5756Crossref PubMed Scopus (69) Google Scholar) a second analysis was carried out in which all missing values were imputed to 1. A "protein-level Anova" was then performed (order peptides using median, minimum number of peptides = 2, maximum number of peptides = 10, no weighting). For both procedures, p values were adjusted using the Benjamini-Hochberg procedure for calculating the q-value (14.Benjamini Y. Hochberg Y. Controlling the false discovery rate - a practical and powerful approach to multiple testing.J. Roy Stat. Soc. B. Met. 1995; 57: 289-300Google Scholar, 15.Reiner A. Yekutieli D. Benjamini Y. Identifying differentially expressed genes using false discovery rate controlling procedures.Bioinformatics. 2003; 19: 368-375Crossref PubMed Scopus (1335) Google Scholar). In both analyses, peptides were grouped into proteins based on the first listed peptide ID if the peptide could be assigned to more than one protein. The lists of proteins identified as belonging to a given abundance profile by Model Based Filter/Impute/Anova or protein-level Anova were then combined. The profiles are listed in supplemental Table S2.Table IProtein abundance profiles, selection criteria and number of proteins identifiedProfile numberProfile nameCriteria for selectionNumber of proteins1Up in hyphaeHigher levels of protein in hyphae than in all other life stages, with q-values for the comparisons < 0.051342Down in hyphaeLower levels of protein in hyphae than in all other life stages, with q-values for the comparisons < 0.052983Down in sporangiaLower levels of protein in sporangia than in all other life stages, with q-values for the comparisons < 0.0554Up in cyst (compared to hyphae, sporangia and zoospores)Higher levels of protein in cysts than in hyphae, sporangia and zoospores, with q-values for the comparisons < 0.05. The average protein levels in cysts were required to be higher than in germinated cysts and in appressoria, but no q-value cutoff was applied.205Down in hyphae and sporangiaLower levels of protein in hyphae and sporangia combined (i.e. sporangia and hyphae were treated as one group in the comparison) than in all other life stages, with q-values for the comparisons < 0.05646Up in sporangia and zoosporesHigher levels of protein in hyphae and sporangia combined (i.e. sporangia and hyphae were treated as one group in the comparison) than in all other life stages, with q-values for the comparisons < 0.051077Down in sporangia and zoosporesLower levels of protein in sporangia and zoospores combined (i.e. sporangia and zoospores were treated as one group in the comparison) than in all other life stages, with q-values for the comparisons < 0.05218
Referência(s)