Exploring the Function of Dark Proteome for Biomedical Applications
Abstract
One of the important goals of the HUPO Chromosome-centric Human Proteome Project (C-HPP) is to correctly define the function of less annotated proteins encoded by their cognate open reading frames on each chromosome in the human genome (1). This can be achived by collaboration with various research groups and resource pillars within the HUPO community (2). For example, with collaborative efforts by the proteomics community, the number of missing proteins (MPs) has been reduced from ~6000 to 2,129 out of 19,823 human canonical proteins (neXtProt 1-11-2019 release). In 2018, HUPO C-HPP launched a new pilot project over 3 years, neXt-CP50, that aims to systematically explore the biological functions of ~50 uncharacterized proteins. These comprise some MPs and/or annotated proteins (uPE1s), termed Dark Proteins, and performed in a close collaboration by 15-international C-HPP teams (3). The Dark Proteins are rising to form one of the major targets in proteome biology due to the nature of their unknown functionality and low abundance properties in general, which render them as promising drug target candidates, biomarkers, cellular regulators and structural components. We present the status of our work-in-progress on the study of a few selected Dark Proteins that are located in chromosome 13 by using the combined multi-omics approaches, which employed cell-based assays, CRISPR/cas9 genome editing, Protein-Protein Interaction (PPI) tool (e.g., COFACTOR, I-TASSER) (4) and proteogenomic analyses. We will also discuss some phenotypic changes as well as functional and structural aspects observed from a loss of function mutant of YPRC-DP1. Finally, we address some common technical issues and annotation problems associated with functional characterization of Dark Proteins (Supported by HI13C2098, a grant from the Ministry of Health and Welfare to Y.K.P.).
Identification of Missing Proteins in Human Olfactory Epithelial Tissue by Liquid Chromatography-Tandem Mass Spectrometry
Abstract
We performed proteomic analyses of human olfactory epithelial tissue to identify missing proteins using liquid chromatography-tandem mass spectrometry. Using a next-generation proteomic pipeline with a <1.0% false discovery rate at the peptide and protein levels, we identified 3,731 proteins, among which five were missing proteins (P0C7M7, P46721, P59826, Q658L1, and Q8N434). We validated the identified missing proteins using the corresponding synthetic peptides. No olfactory receptor (OR) proteins were detected in olfactory tissue, suggesting that detection of ORs would be very difficult. We also identified 49 and 50 alternative splicing variants mapped at the neXtProt and GENCODE databases, respectively, and 2,000 additional single amino acid variants. This dataset is available at the ProteomeXchange consortium via PRIDE repository (PXD010025).
Cancer biomarker study has been the major research subject in proteomics. In the early 21 century, the paper reporting the biomarker signatures for early detection in ovarian cancer considerbaly influenced the direction of cancer proteomics (Petricon III EF et al, Lancet 2002). At the same time, the detection limits of proteomics modalities of those dsays were clearly pointed out by the meta-analysis of plasma proteins (Anderson NL et al, Mol Cell Proteomics 2002). Since then, the researchers have developed proteomics modalities with an aim of identifying cancer biomarkers. Many biomarker candidates were discovered by the higly developed tools. Then, we realized the power of proteomics, and at the same time, the difficulty of applying the proteome data to the clinical applications. In the National Cancer Center, we have challenged the biomarker development using a larget format 2D-DIGE and clinical materials since 2001. Based on our experience, three interesting view poins will be discussed in my presentation.
Fundamental difficulty of biomarker study is attributable to the fact that the real world is hard to be reproduced in the laboratory. For example, the number of cancer patients is often close to that of controls in the experiments. However, in contrast, the cancer patients are few in the normal population in the real world. As a consequence, the diagnostic biomarkers with high sensitivity and specificity at the laboratory condition will generate huge number of false positive detection in the real world. We may need to find the research subject, in which the priori probability is similar between the real world and the laboratory condition. In this sence, the predictive or prognostic biomarkers would be more realistic than the biomarkers for early detection, because the patients with poor response or prognosis exist in a relatively large number in the population of cancer patients.
Other inherent difficulty of biomarker study is that the number of samples is always too small to obtain conclusive results, and it is hard to convince the people to launch a large scale validation. To compensate for a small number of samples, we need a possible theory of biomarkers and attract the sample providors, such as physicians. The in vitro data that the predictive or prognostic biomarkers significantly influence the malignant phenotypes of cancers in a reasonable way are mandatory to start biomarker validation. We always use the cell lines to verify the functional properties of biomarker proteins. However, the fully established cell lines and their xenografts do not always preserve the featrues of original tumor cells. Indeed, it is well know that the anti-cancer drugs effective on the cancer models do not always cure the real patients. The patient-derived cancer models such as xenografts and cell lines freshly established from the tumor tissues will a remedy, because they may still preserve the original phenotypes. However, unfortunately, theose models are not well available from publc biobanks. With this notion, we started the project to develop the patient-derived cancer models in 2014. We have challenged tumor tissues from 250 sarcoma cases, and established 40 xenografts and 30 cell lines. We are expanding this project to an international scale in collaboration with Charls River Laboratories; the NCC provides tumor tissues to the CRL, and the CRL establishes the xenografts. We share the established models, and allow the academic researchers and pharmaceutical companies to access them world wide.
In the mass spectrometric protein identification, we always use the publc proteome database such as SwissProt. SwissProt does not include the peptides with minor mutations, which may be unique to certain tumor or patients groups, and we cannot observe them even when mass spectrometry detects the signals. Although the peptides with unique sequence are worth considering biomarker candidates, they may be missed as long as we use the common proteome database. Those peptides should be good candidates for biomarkers because they wouldn’t be produced by the normal cells. We need a proteome database for individual sample. We are developing a software to generate proteme database from geneme data of individual samples; Mutated Nucleotide and Amino acid sequence Generator (MuNAGe). MuNAGe translates the genome data to the proteome data, and allows the identification of peptides with the sequence unique to each sample. Presently, we are examining the mass spectrometry of tumor tissues with genome data.
At the dawn of cancer proteomics, the technological limitation seemed to be the only obstacle. However, the biomarker study has realized us many challenging issues in the clinical application of proteomics. What we have learned from the biomarker development should be applicable to the other subjets of cancer research.
Room 404, Nakseongdae R&D Center, 38 Nakseongdae-ro, Gwanak-gu, Seoul, 08790, Korea
Phone: (+82)2–393–8328 E-mail : admin@khupo.org Homepage : http://www.khupo.org
COPYRIGHT ⓒ 2019 KHUPO. ALL RIGHTS RESERVED.
