Chromosome-centric Human Proteome Project ... - ACS Publications

Subscriber access provided by UNIV WAIKATO

Article

Chromosome-Centric Human Proteome Project: Deciphering Proteins Associated with Glioma and Neurodegenerative Disorders on Chromosome 12 Manoj Kumar Gupta, Savita Jayaram, Anil K Madugundu, Sandip Chavan, Jayshree Advani, Akhilesh Pandey, Visith Thongboonkerd, and Ravi Sirdeshmukh J. Proteome Res., Just Accepted Manuscript • DOI: 10.1021/pr500023p • Publication Date (Web): 07 May 2014 Downloaded from http://pubs.acs.org on June 2, 2014

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Proteome Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Chromosome-Centric

Human

Proteome

Project:

Deciphering

Proteins

Associated with Glioma and Neurodegenerative Disorders on Chromosome 12 Manoj Kumar Gupta1,2*, Savita Jayaram1,2*, Anil K. Madugundu1, Sandip Chavan1,2, Jayshree Advani1, Akhilesh Pandey1,3, Visith Thongboonkerd4,5, Ravi Sirdeshmukh1,6# *Both the authors contributed equally to this work 1

Institute of Bioinformatics, International Tech Park, Bangalore, India

2

Manipal University, Madhav Nagar, Manipal, India

3

McKusick-Nathans Institute of Genetic Medicine, Johns Hopkins University School of Medicine,

Baltimore, USA 4

Medical Proteomics Unit, Office for Research and Development, Faculty of Medicine Siriraj Hospital,

Mahidol University, Bangkok, Thailand 5

Center for Research in Complex Systems Science, Mahidol University, Bangkok, Thailand

6

Mazumdar Shaw Centre for Translational Research, Narayana Health, Bangalore, India

# Corresponding author Dr. Ravi Sirdeshmukh, Institute of Bioinformatics, International Technology Park, Bangalore- 560066, India Ph: 0091-9885090963; FAX: 0091-80-28416132 E-mail: [email protected], [email protected]

Abstract In line with the aims of the Chromosome-Centric Human Proteome Project (C-HPP) to completely annotate proteins of each chomosome and biology/disease driven HPP (B/DHPP) to decipher their relation to diseases, we have generated a non-redundant catalogue of protein-coding genes for chromosome 12 and further annotated proteins associated with major neurological disorders. Integrating high level proteomic evidences from four major databases (neXtProt, Global Proteome Machine (GPMdb), PeptideAtlas and Human Protein Atlas (HPA)) along with Ensembl data resource resulted in the identification of 1066 protein coding genes, of which 171 were defined as ‘missing proteins’ based on the weak or complete absence of experimental evidence. With functional annotations using DAVID and GAD about 40% of the proteins could be grouped as brain related with implications in cancer or neurological disorders. We used published and unpublished high confidence mass

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

spectrometry data from our group and other literature consisting of more than 5000 proteins derived from clinical specimens from patients with human gliomas, Alzheimer’s and Parkinson’s disease and mapped it on to Chr. 12. We observed a total of 202 proteins mapping to human Chr. 12, 136 of which were differentially expressed in these disease conditions as compared to the normal. Functional grouping indicated their association with cell cycle, cell-to-cell signaling and other important processes and networks, whereas their disease association analysis confirmed neurological diseases and cancer as the major group along with psycological disorders. Using multiple strategies and bioinformatics tools, we identified 103 differentially expressed proteins to have secretory potential, 17 of which have already been reported in direct analysis of the plasma or cerebrospinal fluid (CSF) from the patients and 21 of them mapped to cancer associated protein (CAPs) database that are amenable to Selective Reaction Monitoring (SRM) assays for targeted proteomic analysis. Our analysis also reveales, for the first time, mass spectrometric evidence for two ‘missing proteins’ from Chr. 12, namely, synaptic vesicle 2-related protein (SVOP) and IQ motif containing D (IQCD). The analysis provides a snapshot of chromosome 12 encoded proteins associated with gliomas and major neurological conditions and their secretability which can be used to drive efforts for clinical applications.

Keywords CHPP, B/D-HPP, Chromosome 12, Glioma, Alzheimer’s disease, Parkinson’s disease, secretory proteins, ‘missing proteins’

Introduction The Human Proteome Project (HPP) initiative has been launched by the Human Proteome Organization (HUPO) to identify and characterize the full protein complement of the genome contained in each human chromosome in terms of biological functions and disease relationships. The HPP has been delineated into two broad areas: 1. Gene or chromosomecentric approach (C-HPP) to identify and catalogue all known and well characterized proteins as well as less defined proteins lacking sufficient protein evidence1, 2 and 2. Biology and disease centric HPP (B/D-HPP) to decipher, using experimental or bioinformatics approach, a list of proteins that are directly linked to a specific disease and understand them in depth. The concerted global effort is expected to provide a comprehensive map of human biology and its dynamics in years to come. The main target of B/D-HPP is to develop a wellannotated knowledgebase of disease-associated genes/proteins supported with mass spectrometric evidence or protein capture using antibodies.3 A consortium formed among the Asian research teams (Hong Kong, India, Singapore, Taiwan and Thailand), aims to


Page 2 of 31

Page 3 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


systematically survey Chromosome 12 (Chr. 12) proteins, their disease context as well as explore experimental evidence for the ‘missing proteins’. As a member of the consortium, our group is focused on the analysis of available mass spectrometry data on gliomas and major neurological disorders and we report here our first extensive analysis of the available experimental and bioinformatics/database information on Chr. 12.

Human Chr. 12 spans 132 megabases (Mb) and represents ~4.5% of the total DNA in the cells with an estimate to possess nearly 1400 protein coding genes4. It is rich in disease associated loci and a total of 762 loci on Chr. 12 have been directly linked to human diseases (http://www.omim.org). Gene amplification mechanisms are often exploited by cancer cells to increase copy number and expression of dominant cancer genes. Several chromosomal amplicons found to be localized on Chr. 12, particularly in the 12q13-q15 region, are implicated in human osteosarcomas5, 6 and human malignant gliomas.7, 8, 9 The major genes amplified in this region are MDM2 - a key downregulator of p53, CDK4 involved in cell cycle progression, SAS - regulates cell growth and motility and GLI1 oncogenes.7, 8, 10 Earlier studies have also provided some evidence of the presence of susceptibility loci on Chr. 12 for two common types of adult-onset neurodegerative disorders, Alzheimer’s Disease (AD), a type of dementia and Parkinson’s Disease (PD), a progressive movement disorder; that are characterized by loss of specific neuronal populations11-13. Alpha-2macroglobulin (A2M), a protease inhibitor that mediates the clearence and degradation of Abeta, a major component of beta-amyloid deposits, is located close to AD susceptibility locus and glyceraldehyde-3-phosphate dehydrogenase (GAPDH), a glycolytic enzyme that promotes apoptosis, has recently been implicated in AD. Variations in Leucine-rich repeat serine/threonine-protein kinase 2 (LRRK2, also known as PARK8) gene, induces neuronal degeneration and has been linked to PD14. In addition to these, many important neuronal disease related genes like, protein tyrosine phosphatase (PTPN11), low density lipoprotein receptor-related protein 1 (LRP1), Nitric-oxide synthase (NOS1),15 Contactin 1 (CNTN1), are found on Chr. 12. Mutations in PTPN11 have been linked to Noonan syndrome and acute myeloid leukemia. Excess nitric oxide (NO) production by neuronal NOS1 has been implicated in neuronal tissue damage seen in AD and PD. Genetic predisposition to psychiatric conditions such as, Schizophrenia, Bipolar Disorder and Major Depression Disorder (MDD), also confirms the presence of neuro related genes in 12q22-23 region.16 The main aim of our study was to revisit and generate a non-redundant catalogue of proteincoding entries on Chr. 12 using updated data resources and assess their functional and disease association in gliomas - major tumors of the central nervous system (CNS) and AD



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

and PD - major neurodegenerative conditions, using published literature as well as unpublished data from our lab. We confirm that nearly 40% of the Chr. 12 proteins are brain related with many proteins implicated in the tumor condition or neurological disorders. We also provide for the first time, protein-level, mass spectrometric evidence for two of the ‘missing proteins’ on Chr. 12. With this, we provide a snap-shot of Chr.12 proteins associated with these diseases, which can be further explored in depth by us and others in the context of their respective biology and disease involvement.

Materials and methods Data resources and pipeline for Chr. 12 parts list To come up with the parts list for the protein-coding genes on Chr. 12, four major knowledgebases were used as primary resources, namely: neXtProt database, Global Proteome Machine (GPMdb), PeptideAtlas and the Human Protein Atlas (HPA) in addition to Ensembl v74 database resource for gene related information. neXtProt (Release: 2014-0225) has been a reference knowledgebase and integration platform for HPP, which integrates highest quality proteomics ‘Gold’ data (≤1% error) or ‘Silver’ data (≤5% error) mapping to Ensembl and Swiss-Prot/Uniprot entries17. PeptideAtlas (Release: 2014-03-14)18 and the GPMdb (Release: 2014-03-15) databases were built from raw spectra with increasingly stringent criteria of 1% FDR at the protein level and 0.2% FDR at the peptide level. The HPA (version 12.0; Release: 2014-03-13) data resource categorize the protein evidence for genes as high, moderate or low, based on manual curation of immunohistochemistry, immunoflurosence and western blot data. We have used the following set of thresholds and criteria as per the HUPO defined guidelines19 for mapping proteins with high confidence evidence20, 21- neXtProt (evidence: Protein level) or PeptideAltas (presence level: canonical) or GPMdb (evidence: high quality). HPA was used as an additional supportive evidence for all the proteins. For ‘missing proteins’ as those with low or no supportive protein level information from any of these resources, the criteria used was: neXtProt (evidence: transcript level, homology, predicted or uncertain) or PeptideAltas (non-canonical) or GPMdb (low quality or no credible evidence). All the proteins that mapped to Chr. 12 were also checked for subcellular localization and functional annotations using Human Protein Reference Database (HPRD)22 and chromosomal locations were obtained from Ensembl database. A systematic pipeline for this is represented in Figure 1. Bioinformatics analysis and functional annotation of Chr. 12 proteins implicated in gliomas, Alzheimer’s and Parkinson’s disease


Page 4 of 31

Page 5 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


A thorough scan and survey of the literature published in peer reviewed journals during the last 10 years on high throughput proteomic studies in human gliomas, Alzheimer’s disease and Parkinson’s disease was done to map the proteins implicated in these conditions to Chr. 12. The datasets used were generated using 2DE-MS or gel-free LC-MS/MS based mass spectrometry approaches and quantifications based on image analysis (2DE-MS) or iTRAQ technology (gel free LC-MS/MS). These published studies and datasets used for the analysis are listed in Table 1. The analytical rigor used for these data was found to be as per standard parameters followed for high confidence identifications and quantifications with accepted false discovery rate (≤1%).23 Apart from published data, we used unpublished data in case of gliomas that was also generated using iTRAQ-based LC-MS/MS analysis with comparable

analysis

parameters.

These

datasets

have

been

deposited

to

the

ProteomeXchange with identifier number PXD000848. To substantiate proteins and their roles in these disease conditions, we further added supportive evidence from transcriptomic studies wherever available. Ingenuity Pathway Analysis (IPA) was done to identify networks, molecular and cellular functions, and diseases and disorders. We further referred to an integrated cancer-associated protein database (CAP) to augment these data with functional annotations and amenability to SRM assays.24 The pipeline used for mapping and annotation of these proteins is explained under Results section. Further we assesed the secretory potential of Chr. 12 proteins by using several bioinformatics tools. We used HPRD to find membrane and extracellular localization; proteins with signal peptide and transmembrane domain were predicted using SignalP 4.0 and TMHMM 2.0 software tools, respectively and we used SecretomeP to identify proteins which may follow a non-classical secretory pathway. In addition, ExoCarta database51 (containing 2576 exosome proteins) was used to map the exosomal proteins. The experimental detectability of these proteins in the body fluids was veryfied by comparing with the protein profiles from normal cerebrospinal fluid (CSF)52 and plasma53. Aplying the same startegies, we also mapped subset of Chr.12 proteins implicated in Gliomas, AD and PD, for their secretory potential, along with the differentially altered proteins identified from direct analysis of CSF/plasma from patients with these neurogical disorders.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Results and discussion In order to fully annotate the current proteomic information on Chr. 12, we examined four major proteomic data resources: (neXtProt database, Global Proteome Machine (GPMdb), PeptideAtlas and the Human Protein Atlas (HPA)) apart from Ensembl v74 database for gene specific information and enlisted all the non-redundant protein entries mapping to Chr. 12. We identified the total number of protein coding gene entries to be 1031 as per neXtProt and 1035 as per Ensembl, after removing 'uncertain' and 'dubious' entries as recently suggested by Lane et al as per HUPO guidelines19. Majority of the genes (n=1000) were common between the two resources whereas 31 were unique to neXtProt and 35 were unique to Ensembl. Combining the two databases resulted in a total of 1066 non-redundant entries of protein coding genes on Chr. 12 (Table 2 and Supplementary Table 1A and 1B).

The ‘missing proteins’ estimates could be somewhat variable based on the criteria/metrices used for their determination. As per HUPO guidelines, Lane et al19 considered protein level evidence only from neXtProt to group 'missing proteins' but have suggested that one could use additional evidence from GPMdb. Although, neXtProt includes evidence from PeptideAtlas, it does not take into account GPMdb evidence at any level. We observed that 41 of the entries that do not have protein level evidence in neXtProt, have high quality evidence in GPMdb. We believe this may be considered and incorporate this information as described in the pipeline (refer Figure 1 under method section). Thus, of the total 1066 protein coding genes entries on Chr. 12, 895 had high level protein coding evidence and 171 were catagorised as ‘missing proteins’ on account of weak or no evidence at protein level in any of the high quality proteomic databases - neXtProt, PeptideAtlas and GPMdb (Supplementary Table 1A and 1B). These strategies may evolve and revise the number of ‘missing proteins’ as databases get curated with further information. Chromosomal location of all these 1066 protein-coding genes was obtained using Ensembl database and their Gene ontology information, molecular function and sub-cellular localization were extracted using HPRD (Supplementary Table 1A; Supplementary Figure 1A and 1B).

One of the key objectives of B/D-HPP is to map aberrantly expressed genes and their protein products to chromosomes and explore the underlying molecular networks that regulate disease processes. Two functional annotation resources were used to map the disease association of Chr. 12 proteins, namely DAVID and Genetic Associtation Database (GAD). DAVID (http://david.abcc.ncifcrf.gov/home.jsp) integrates functional genomic level annotations from GenBank, UniGene, RefSeq, Locuslink, OMIM etc with GO terms, protein domains and KEGG and Biocarta pathways to identify enriched biological themes, discover functionally-related gene groups, and cluster genes based on redundant annotation terms.


Page 6 of 31

Page 7 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


GAD (http://geneticassociationdb.nih.gov) is focused on archiving published gene level information from candidate gene studies and genome wide association studies on genetic variations and polymorphisms positively associated with disease MeSH terms. Using these resources, we observed about 40% of the 1066 proteins to be brain localized and associated with cancer and neurological disorders as shown in the Figure 2A and 2B. Apart from this, a major category of proteins also map to cardiovascular and metabolic disorders. Further, many genetic defects and diseases with their associated susceptibility loci have shown linkage to Chr. 12 using Online Mendelian Inheritance of Man (OMIM) database - a collection of information on genetic disorders and traits (Supplementary Table 2). Several chromosomal loci such as, 12p11.2-q13.1, 12q24.31, 12p13, 12q22-q23.3 and others are found to be associated with neurological disorders.16, 54-58 Based on the above observations, we planned to focus on genes/proteins implicated in gliomas and neurological disorders and understand their role in the disease process. Gliomas are the most common brain tumors of the adult CNS, originating from the neuroepithelial tissue of which astrocytomas are the most predominant variety. Among these, Glioblastoma multiforme (GBM; WHO grade IV) is a highly aggressive tumor, characterized

by

neurodegeneration.

59-62

uncontrolled

proliferation,

diffused

tissue

invasion

and

Increased levels of glutamate causes neuronal cell death and is

considered to play a vital role in glioma-induced neurodegeneration which promotes malignant glioma progression.63 Thus, there may be several proceses that may be common between glioma and neurodegenerative disorders. AD and PD are two common neurodegenerative disorders occuring in older adults where brain cells (neurons) are damaged, ultimately leading to dementia. We used published and unpublished, high confidence proteomic datasets generated with clinical specimens of gliomas (high and low grades) and from the published literature on clinical specimens from neurological diseases (AD and PD), to make additional integrated readouts for Chr. 12 with regard to the protein functions and disease association. The workflow that was followed to map the disease association of Chr. 12 proteins is depicted in Figure 3. The analysis resulted in identification of 202 non-redundant proteins, of which, 136 proteins (125 in glioma and 35 in AD or PD) showed differential expression of ≥1.5 fold, some of them with support from the transcript level (Supplementary Table 3). Given the possibility that some of the neurological tumor related processes may share commonality with neuro degenerative conditions, we looked for common proteins between the two sets of conditions. A total of 24 proteins were common between glioma and AD/PD as shown in the Figure 4 and listed in Supplementary Table 3.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The proteins and genes that represent the above mentioned neurological conditions (n=136) were further overlapped with the positional gene sets corresponding to each cytogenetic band using Gene Set Enrichment Analysis tool that uses the Molecular Signatures Database (GSEA/MSigDB; version 4.01 release Aug 9, 2013). Further, genes implicated in different cancers, listed chromosome wise in Atlas of Genetics and Cytogenetics in Oncology and Haematology (http://atlasgeneticsoncology.org) was also used to identify gene clusters on Chr. 12, as shown in Figure 5. Differentially expressed Chr. 12 protein genes from our analysis are highlighted in red in the background of other cancer genes present in the region. These gene clusters due to their proximity to each other and to other known cancer related genes, are helpful in identifying positional effects related to chromosomal deletions or amplifications that are often observed as somatic variations in the tumors. They may also be useful in identifying dosage compensation, epigenetic silencing and other regional effects. Genetic variations (SNPs) showed significant association (P< 0.001) with 12p13 and adjacent regions (12p13.3, 12p13.31), showing linkage to increased glioma risk in patients with a family history of primary brain tumors.64 Strong association for genetic variants was seen for three genes in this region, namely serine/threonine/tyrosine kinase 1 (STYK1), protein arginine methyl transferase 8 (PRMT8) and SRY-box 5 (SOX5) that play diverse roles in cellular processes and development and regulation of cell fate, cell proliferation, differentation and survival. Highly amplified regions of Chr. 12 have been found to cluster genes implicated in glial tumors.7, 8, 10 The 12q13-15 region represents one of the best mapped amplicons in gliomas. Amplifications of multiple genes, either together or independently, in 12q13-15 segment, accompanied by strong overexpression of CDK4, SAS and MDM2 is seen in malignant gliomas7 as discussed earlier.65 Genes such as GADD153, GLI1, RAP1B, A2MR, and IFNG, that are located close to these amplification clusters may be coamplified in these tumors, suggesting possible positional effects.8 This region also carries several genes that are found to be overexpressed: KUB3, WNT1, CTDSP2, CDK4, OS9, DCTN2, RAB3IP, FRS2, GAS41 and RAP1B.66, 67 From our datasets of differentially expressed proteins in glioma, AD and PD, we observe 37 overexpressed genes maping to the 12q13-15 region (30 from the gliomas and 11 from AD/PD; Supplementary Table 3). Although it is not clear whether all represent amplified genes, it is interesting that they include LRP1, OS9 and RAP1B, already known to be amplified in this amplicon.66 The protein differentials (n=136) from the present neuro disease dataset were further mapped to known networks and processes using IPA. One of the top networks identified was cell death and survival, cancer, cell-to-cell signaling and interaction. Neurological,


Page 8 of 31

Page 9 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


psychological disease and cancer came up as top diseases and disorders in this analysis confirming the results. Cell cycle and cell to cell signaling were top molecular and cellular functions (Supplementary Table 4A, 4B and 4C). Identification of candidate biomarkers in body fluids is critical in the context of clinical applications and disease management as mentioned above; exploring secretabilty of proteins in CSF or plasma in the case of cancers and neurological disorders, would thus be important.68, 69 Towards this goal, we examined the total Chr. 12 proteins dataset (n=1066) listed in Supplementery Table 1A for their reported detection in the normal CSF or plasma and further their secretory potential using 1. SignalP 4.0 (i.e. by the endoplasmic reticulam or golgi-dependent pathway (SignalP probability ≥0.90) based on presence of signal peptide), 2. SecretomeP 2.0 (i.e. by nonclassical secretory pathway (SignalP probability

Chromosome-centric Human Proteome Project ... - ACS Publications

Recommend Documents