Schedule

Lorne Taylor,  MDS Sciex

Information not available for posting

"Mapping Human Protein-interaction Networks Using Targeted Peptide Arrays and Bioinformatics"

Shawn Shun-Cheng Li, University of Western
Ontario
sli@uwo.ca

A growing body of research has concentrated on the identification and definition of conserved sequence motifs. It is widely recognized that these conserved sequences and structural units mediate protein functions and interactions. Hence domains that facilitate similar or related functions may appear in interacting proteins such as those that constitute a protein complex. In this study we present a new method of identifying functionally correlated domains in protein complexes from Saccharomyces cerevisiae using curated and high-throughput datasets. Two random models were used to generate probability values that represent the degree of correlation between domains. The statistical associations between domains, generated from the two random models, were visualized in the form of domain correlations networks. Our results show that high-quality curated datasets produce domain networks that map to known biological assemblies such as Ribosome, RNA polymerase, Proteasome regulators, transcription initiation and Histones. In contrast, the high throughput datasets are comprised of one large network of domain associations. High connectivity of RNA processing and binding domains in the high-throughput datasets reflects the abundance of RNA binding proteins in yeast and is in agreement with a previous report that identified a nucleolar protein cluster from these complexes. We compared these results with binary protein interactions that were collected from DIP, BIND and literature-based survey. A conservative calculation shows that about 20% of the functional associations are directly supported by protein interactions where each protein contains one of the correlated domains.

We describe a new de novo sequencing software package, PEAKS, to extract amino acid sequence information without the use of databases.  PEAKS uses a new mathematical model and a new algorithm to efficiently compute the best peptide sequence whose fragment ions can best interpret the peaks in the MS/MS spectrum.  The output of the software gives amino acid sequences with confidence scores for the entire sequence as well as an additional novel positional scoring scheme for portions of the sequence. The performance of PEAKS is compared with Lutefisk, a well known de novo sequencing software, using quadrupole-time-of-flight (Q-TOF) data obtained for several tryptic peptides from standard proteins.  PEAKS showed a clear advantage in the comparison.

Recently, we further added a database search step to PEAKS after de novo sequencing.  The peptide sequences or sequence tags obtained by de novo sequencing are searched for in a protein database to identify proteins.  The proteins identified by the sequence search are in turn used to interpret the MS/MS spectra.  Our experiments showed that this approach works significantly better than the “de novo sequencing” or the “matching uninterpreted spectra” approaches.

We have developed a new database search algorithm, called SIMS (Simple Interval Motif Search), which is optimized for identifying post-translational modifications and sequence substitutions in peptide tandem mass spectra in an unbiased and comprehensive manner. SIMS is effective at detecting alterations at specific amino acid residues without prior knowledge or expectation of the chemical delta mass or modification site. It also predicts peptide sequence substitutions wherein one amino acid is replaced with a different amino acid due to polymorphism, mutation, or database error. To date, we have shown that SIMS is effective at correctly identifying:
 - Sequences with common modifications such as phosphorylation, acetylation and carboxymethylation using MS/MS spectra generated by ion trap MS.
 - Peptides with database sequence errors and unknown substitutions.
 - Sequences with unknown chemical modifications.
 - Unmodified peptides.
The results were validated using the database search algorithm SEQUEST [Eng et al. J Am Soc Mass Spectrom 1994; 5: 976-989] and StatQuest [Kislinger et al. Mol Cell Proteomics 2003; 2: 96 - 106].  SIMS has been optimized for the Linux operating system and can be run in a parallel, distributed computing environment.

The ciliated protozoan Tetrahymena thermophila is a unicellular eukaryote with a degree of cellular structural and functional complexity comparable to that of human and other metazoan cells. It has proven extremely valuable as a model organism for many genetic and molecular biological studies. It contains an estimated 20-40,000 proteins, a rich resource for proteome analysis.  An ongoing EST sequencing project has identified many interesting proteins which have database matches to proteins found in mammalian cells but not in yeast. An impressive array of novel molecular genetic technologies places Tetrahymena at the forefront of experimental, in vivo functional genomics research. Sequencing of the genome of T. thermophila has recently reached 7X coverage. This as well as ongoing EST sequencing, complement functional analysis such as gene knockout and antisense ribosome technologies and provide a basis for exploiting functional analysis. To initiate functional analysis, we have undertaken an analysis of aspects of the proteome of Tetrahymena. Our initial focus has been on cilia, an important organelle with a tractable protein complement, an organelle not present in another unicellular model organism, the yeast Saccharomyces cerevisiae. We will describe the isolation of ciliary membranes and axonemes, separation of their protein complements by 1D and 2D gel electrophoresis, and protein identification by mass spectrometry including peptide sequence analysis by MS/MS. Methods used and developed for this work will be presented. We will describe the identification of a number of novel and interesting proteins from cilia and will present data about possible post-translational modifications observed.

The leading cause of death from gynecologic malignancies in the United States is ovarian cancer.  Cisplatin is the mainstay of treatment for many solid tumors. The response rate of advanced stage ovarian cancer to platinum-based primary therapy is 70-80%, with most patients relapsing and dying of disease having developed clinical resistance to multiple chemotherapeutic agents. Enhanced Transcription-coupled Nucleotide Excision
Repair (TC-NER) is implicated as the cause for Cisplatin resistance in ovarian cancer cells. Recent studies show that Et-743, a potent antitumor agent, counteracts this resistance by specifically targeting and killing the cancer cells containing TC-NER factors. We have developed proteomic tools with which to evaluate the expression of the protein components of the NER
pathway and to correlate those results to platinum and Et-743 sensitivity in culture and in vivo. This tool will be used in a forthcoming phase II clinical trial of Et-743 in epithelial ovarian cancer patients.  This will hopefully allow us to determine which drug treatment, either Cisplatin or Et-743, would benefit patients the most.  The aim of our research is to contribute to the improvement of cancer chemotherapy.

Recent findings have highlighted the fact that the regulation of poly(ADP-ribose) (pADPr) metabolism is not only important for DNA repair but is also involved in apoptosis signaling and in the regulation of circadian rhythms. To gain insight into the cellular pathways involving poly(ADP-ribose) polymerases (PARPs), poly(ADP-ribose) glycohydrolase (PARG) and pADPr, we have initiated a study to identify pADPr-binding proteins as well as protein interactors of PARG and three nuclear PARPs, PARP-1, PARP-2 and PARP-3, using a proteomic approach.To identify pADPr-binding proteins, Rotofor fractions of a HeLa cellular extract were screened with ³²P-labeled pADPr using a protein blot assay. The most abundant pADPr-binding proteins were identified by peptide mass finger printing as hnRNPs, a family of proteins that bind pre-mRNA into functional complexes involved in mRNA maturation and transport to the cytoplasm.PARP interactors were isolated by immunoprecipitations. Endogenous PARP-1 was immunoprecipitated from HeLa S3 nuclei using PARP-1 specific antibodies. PARP-2 and PARP-3, transiently expressed in COS-7 cells as N-terminal fusions with the FLAG epitope, were immunoprecipitated with the M2 anti-FLAG antibody. Immunoprecipitated proteins digested with trypsin were identified by mass spectrometry and Western analysis. This approach identified previously reported PARP-1 interactors such as topoisomerase I, hnRNPs and nucleolin, as well as novel interactors such as EBNA1-binding protein. The identification of PARP-3 interactors suggests that it participates in the cellular response to DNA damage. The identification of PARP-2 interactors isolated by immunoprecipitation and of PARG interactors isolated by a yeast two-hybrid screen is ongoing and results of this analysis will be presented. This study will contribute to our global understanding of the role of cellular poly(ADP-ribosyl)ation.

Endometrial carcinoma (EmCa) is a cancer of the uterine epithelium.  It is the fourth most common malignancy for women in Canada, representing 3,500 cases in 2001.  The pathologic diagnosis of EmCa (in fact, almost all human cancers) is currently based upon the microscopic recognition of cellular and tissue phenotypes.  Advances in proteomic analysis techniques, especially those that are based on high-resolution mass spectrometry, now permit the recognition of both normal and diseased tissues and cellular phenotypes by means of biomarkers in protein/peptide profiles.  These biomarkers may then be used in the screening or diagnosis of disease through serum, cytologic, or tissue testing.  We are currently engaging in the proteomics of endometrium and EmCa, and in developing methodologies and technologies in mass spectrometry (MS).  At present, over 300 consented tissue and tumor samples have been banked and histologically classified.  Tens of potential cancer markers have been recognized; a number of them have been positively identified and two have high probability for clinical utility.  Marker discovery is carried out by (1) comparing protein expression profiles after fast separation using solid-phase extraction or selective adsorption on functionalized surfaces (e.g. protein chips of Ciphergen); and (2) comparing relative abundances of tryptic peptides labeled with isotope-coded affinity tags (ICAT) after nanobore liquid chromatography (nanoLC) separation.  The former methodology is typically implemented on matrix-assisted laser desorption/ionization (MALDI) hybrid tandem quadrupole time-of-flight (QqTOF) mass spectrometers and SELDI (surface-enhanced equivalent of MALDI) on protein chips coupled to QqTOF mass spectrometers.  The latter methodology is employed with nanoLC coupled to nanospray QqTOF MS.  Methodology No. 1 is relatively easy to implement, but tends to give fewer and is more amenable to smaller proteins; in addition, protein identification is typically difficult and slow.  Methodology No. 2 is technically more demanding, with two-dimensional LC separation (first dimension of strong cation exchange fractionation followed by second dimension of reverse-phase nanoLC) and ICAT-labeling, large-scale protein identification and quantification by isotope-dilution mass spectrometry is achieved, although it is labor-intensive and throughput is medium as protein identification is repeated in every run.