Session MOB. There are 6 abstracts in this session.

Session: Cascadia Proteomics Symposium Session, time: 09:50 AM - 10:05 AM

Sample Preparation of OCT-embedded Brain Tissue for Quantitative Proteomics

Julia Robbins1; Gennifer Merrihew1; Tom E. Montine2; Michael MacCoss1
1University of Washington, Seattle, WA; 2Stanford University, Palo Alto, CA

Alzheimer’s Disease is a growing public health burden with limited interventions. Further
therapeutic development rests on a more detailed understanding of molecular pathogenesis.
The vast majority of data for AD pathogenesis in humans is from standard neuropathologic
assessments or biochemical measurement of a limited number of analytes directly related to
neuropathologic features (amyloid β, hyperphosphorylated tau). Although the current standard
neuropathologic evaluation is highly valuable because it remains the only tool that provides
comprehensive assessment of diseases that afflict an individual brain, it also is severely limited
as a way to investigate molecular pathogenesis.

A major challenge of performing proteomics on human brain is the complexity of the mixture and
the abundance of lipids in the tissue. Furthermore, Optimal Cutting Temperature (OCT)
compound is often used to embed sample tissue prior to frozen sectioning on a
microtome-cryostat. Unfortunately, the analysis of OCT-embedded tissues has been difficult
because of the interference of OCT in the mass spectrometry analysis. OCT binds to the C18
reversed phase chromatography column, elutes during the peptide separation, and creates ion
suppression during electrospray ionization. Consequently we systematically evaluated different
homogenization, solubilization, digestion, and sample cleanup methods with the goal of
performing quantitative proteomics measurements. We wanted a methodology that not only
improved the number of total peptides that could be detected from digested brain tissue but also
improved the quantitative precision. We will present the different strategies that were evaluated
using data-independent acquisition mass spectrometry. The result is a protocol that reproducibly
minimizes OCT and lipid interference and maximizes peptide digestion and recovery.

Tips and Tricks:

Session: Cascadia Proteomics Symposium Session, time: 10:05 AM - 10:20 AM

The proteome landscape of childhood cancers  Implications on xenograft models, drug sensitivity and optimal diagnostic timepoint

Amanda Lorentzian1, 2; Anuli Uzozie1, 2; James Lim1, 2; Christopher Maxwell1; Gregor Reid1; Philipp Lange1, 2
1University of British Columbia, Vancouver, Canada; 2BC Children's Hospital, Vancouver, Canada

Personalized treatments have significant potential to improve therapy options for high-risk cancers and reduce late effects in general. The approach is still in its infancy and much work is needed to improve its utility and efficacy. This is particularly true for childhood cancers for which positive outcome rates remain in the low percent range. To better understand the molecular mechanisms contributing to this we investigate the validity of three key concepts and tools upon which precision oncology is built.

We employ a proteogenomic workflow combined with in vitro drug sensitivity assays and murine patient derived xenografts. Deep sequence coverage of relevant genome variants is enabled by a childhood cancer specific targeted sequencing panel, which we first validated against whole genome sequencing. Data independent acquisition, a newly developed N termini enrichment approach (HUNTER) and sensitive phosphopeptide enrichment allows us to derive functional information beyond protein abundance from only 20µg of limited patient specimen per analysis. We present unpublished data from >100 biopsies from patients diagnosed with diverse pediatric malignancies spanning sarcomas, neuroblastomas and acute lymphoblastic leukemias.

Based on integration of the genomic, proteomic and PTM data we, for the first time, show that the proteome landscape of primary pediatric acute leukemias is largely recapitulated in murine patient derived xenografts. We also present data showing that select pathways differ between primary and xenograft cells and that the host environment has greater impact on post translational modifications than protein abundance. We then explore if protein and PTM level variation can explain observed disconnects between genomic biomarkers and in vitro or in vivotreatment response. Lastly, we explore the stability of genome and proteome variants between initial diagnosis and later relapses. Together these findings support xenografts as excellent model system for most studies and point out areas where addition of proteome analysis would likely improve treatment stratification.

Tips and Tricks:

Session: Cascadia Proteomics Symposium Session, time: 10:20 AM - 10:35 AM

Miro: a method for the proteome-wide identification of amino acid substitutions that are deleterious for protein function

Ricard Rodriguez-Mias; Kyle Hess; Bianca Ruiz; Stephanie Zimmerman; Ian Smith; Anthony Valente; Yang Lu; William S. Noble; Stanley Fileds; Judit Villen
University of Washinngton, Seattle, WA

DNA sequencing methods have led to the discovery of millions of mutations that change the encoded protein sequences, but the impact of nearly all of these mutations on protein function is unknown. We addressed this scarcity of functional data by developing Miro, a technology that leverages mistranslation of non-canonical amino acids to produce protein variants en masse followed by their biochemical characterization using high throughput proteomics assays.

I will present early in vitro and in vivo applications of the Miro technology,  in conjunction with a diverse suit of biochemical assays, to reveal deleterious proline to azetidine-2-carboxylic acid substitutions that impact protein structure, ligand-binding, protein-protein interactions, post-translational modifications and protein thermal stability. 

Additionally, we have screened a panel of non-canonical amino acids and demonstrated that Miro is cost effective, scalable to most types of amino acids, and flexible with respect to the model system. Miro is also versatile in terms of protein functions that can be probed; beyond the assays presented here, Miro can readily be extended to study the effects of amino acid substitutions on protein aggregation, enzymatic activity, subcellular localization, as well as other activities. 

Further development of Miro will enable the generation of proteome-wide mutational sensitivity maps for human proteins as an essential companion to the human genome These maps will allow us to understand human variation and to interpret the clinical significance of millions of mutations in the human genome.

Tips and Tricks:

Session: Cascadia Proteomics Symposium Session, time: 10:35 AM - 11:50 AM

Absolute quantification of transcription factors reveals principles of gene regulation during human erythropoiesis

Mark A. Gillespie*1; Carmen G. Palii*2; Daniel Sanchez-Taltavull*3; Theodore J. Perkins2; Marjorie Brand2; Jeffrey A. Ranish1
1Institute for Systems Biology, Seattle, WA; 2Ottawa Hospital Research Institute, Ottawa, ON; 3Visceral Surgery, University of Bern, Bern, Switzerland

Dynamic changes in transcription factor (TF) abundance and stoichiometry drive cell state changes. However, our understanding of these complex relationships is limited by the paucity of nuclear protein concentration information. To address this, we developed and employed SRM-based targeted mass spectrometry assays, together with stable isotope labeled standard peptides, to quantify the absolute abundances of 103 TF and cofactors across 13 sequential time points during human erythropoiesis. We observe a range of correlations between protein and mRNA abundances, suggesting a role for post-transcriptional regulatory mechanisms. In addition, we define the protein concentration (copies/nucleus) for master regulators of hematopoiesis/erythropoiesis, as well as coregulators of transcription, thereby providing a quantitative scale for TFs in the nucleus. Notably, we observe that corepressors are significantly more abundant than coactivators at the protein level, but not at the RNA level. Finally, we integrate these absolute protein abundances with mRNA measurements to generate a dynamic gene regulatory network of erythroid commitment. These data provide unique and essential information for understanding the transcriptional regulatory programs controlling erythropoiesis, as well as general mechanisms that may regulate other cell fate decisions.

Tips and Tricks:

Session: Cascadia Proteomics Symposium Session, time: 10:50 AM - 11:00 AM

Comparative proteomics of serum from 31 mammalian species

Benjamin Neely1; Magnus Palmblad2; Phillip Wilmarth3; Alison Bland4, 5; Michael Janech4, 5
1National Institute of Standards and Technology, Charleston, SC; 2Leiden University Medical Center, Leiden, The Netherlands; 3Oregon Health & Science University, Portland, OR; 4Hollings Marine Laboratory, Charleston, SC; 5College of Charleston, Charleston, SC

Systematically characterizing the diversity of mammalian proteomes will improve our understanding of biologically interesting phenotypes and recent adaptation. Our first step to build the molecular cartography of mammalian blood proteomes and provide a foundational understanding for future biomimicry studies has focused on 31 species with annotated genomes spanning three placental superordinal clades: Afrotheria, Laurasiatheria, and Euarchontoglires. Undepleted sera from up to four individuals from each species were digested with trypsin and analyzed by data-dependent acquisition. The NIST SRM 909c human serum was used throughout the experiment to confirm digestion consistency and provide a human reference. Data were analyzed in both an ID-free and ID-based manner. For the former, the compareMS2 molecular phylogenetic approach was employed that used the fraction of shared spectra to calculate pairwise distances between all samples. The calculated relationships are based not only on the sequence-dependent fragmentation spectra, which includes some of the genetic variation, but also the relative abundance and post-translational modifications of the proteins. The resulting phylogenetic tree correctly recapitulated mammalia, with the major exception that cetacean blood proteomes were more similar to carnivora than even-toed ungulates. The data was also searched using standard database approaches and the resulting identified proteins were matched to human ortholog sequences using BLAST. On average, 342 proteins were identified (240 to 502) with 57 being shared across all species, though within each clade there were more shared proteins (i.e., pinnipeds shared 205 proteins). Combining an ID-free method with ID-based methods is particularly useful when covering a large number of species whose genomes are not necessarily assembled and annotated to the same standard, as it provides an independent overview of the quality of and similarity between all datasets in a study. Ongoing analysis will identify differences in the blood proteome that may be linked to phenotypes within or between clades.

Tips and Tricks:

Session: Cascadia Proteomics Symposium, time: 11:00 AM - 11:10 AM

“Borrelia PeptideAtlas” a resource for Lyme disease research

Jaipalreddy Panga1; Zhi Sun1; Michael R. Hoopmann1; David S Campbell1; David Baxter1; Strle Klemen2; Leroy Hood1; Kai Wang1; Piesman J4; Allen Steere3; Robert L. Moritz1
1Institute for Systems Biology, Seattle, WA; 2Massachusetts General Hospital, MA & Wadsworth Cen, New York, MA; 3Massachusetts General Hospital, New York, MA; 4CDC, Salt Lake City, UT

Despite discovering the causative agent for Lyme disease and the availability of the B. burgdorferi B31 reference genome sequence, there is no truly sensitive and specific test for detecting Lyme disease at early stages of infection. Currently, the approved two-step diagnostic test only detects post-infection immune response to the Lyme disease pathogen, with limited sensitivity and specificity. A comprehensive genomic and proteomic characterization of infective B. burgdorferi isolate is essential to correlate Lyme disease severity with these isolates. Comprehensive plasmid analysis of over thirty infective isolates of B. burgdorferi obtained from around the US provides evidence that the genome structure is quite different in plasmid number. Further, analysis of the B. burgdorferi B31 reference isolate (ATCC) has identified new plasmids previously unseen. To enable the development of new proteomic based signatures of B. burgdorferi infection, we have created the “Borrelia PeptideAtlas”, which is derived from multiple clinical isolates and consolidates all these data into a comprehensive and publicly accessible web based, searchable resource for the Lyme research community. Furthermore, we are using this information to develop pathogen specific biomarker panels to detect B. burgdorferi.

Tips and Tricks: