Session Poster. There are 154 abstracts in this session.

Session: Poster, poster number: 001-ML

Multiplexed Quantification Strategy for Candidate Biomarker Discovery and Verification in Alzheimer’s Disease

Xiaofang Zhong; Qinying Yu; Fengfei Ma; Dustin Frost; Lingjun Li
University of Wisconsin-Madison, Madison, WI

Absolute quantification in targeted proteomics is challenging due to a variety of factors, including low specificity in complex backgrounds, limited analytical throughput, and wide dynamic range. To address these problems, we developed a hybrid offset-triggered multiplex absolute quantification (HOTMAQ) strategy that combines cost-effective mass difference and isobaric DiLeu (N,N-dimethyl leucine) tags to enable simultaneous construction of an internal standard curve in the MS1 precursor scan, real-time identification of peptides at the MS2 level, and mass offset-triggered accurate quantification of target proteins in synchronous precursor selection (SPS)-MS3 spectra. This approach increases the analytical throughput of targeted quantitative proteomics by up to 12-fold. The HOTMAQ strategy was employed to verify candidate protein biomarkers in preclinical Alzheimer’s disease (AD) with high accuracy. The greatly enhanced throughput and quantitative performance, paired with sample flexibility, makes HOTMAQ broadly applicable to targeted peptidomics, proteomics, and post-translational modifications.

Besides of protein biomarkers, the glycosylation-based biomarker is another promising direction, as glycosylation is crucial for many brain functions, including learning and memory. In our quantitative N-glycoproteome study, cerebrospinal fluid (CSF) samples from 48 healthy controls, patients at mild cognitive impairment (MCI) and AD dementia were compared by 12-plex isobaric DiLeu labeling strategy. After enrichment, above 2000 N-glycopeptides corresponding to 112 N-glycoproteins were quantified in Fusion Lumos. The majority of proteins in CSF are either fucosylated or sialylated. Through the integrated proteomic and glycoproteomic study, we demonstrated N-glycoproteins could be candidate biomarkers in Alzheimer’s disease.

Tips and Tricks (if submitted):

Session: Poster, poster number: 002

A Multiple Reaction Monitoring Assay for Simultaneous Quantification of Mitochondrial Proteins in Mouse

Alexsandr Stotland; Weston Spivia; Amanda Orosco; Allen Andres; Jennifer Van Eyk; Roberta Gottlieb; Sarah Parker
Cedars Sinai Medical Center, Los Angeles, CA

Mitochondria are the major source of cellular energy (ATP), as well as critical mediators of widespread functions such as cellular redox balance, apoptosis, and metabolic flux. Mitochondrial dysregulation occurs in a variety of cardiovascular pathologies, including atherosclerosis, heart failure, and vascular aneurysms. Mitochondrial protein abundance is a key indicator of mitochondrial quantity as well as composition. Methods to quantify mitochondrial proteins are limited to low throughput immunoassays or untargeted, relative quantification by shotgun data-dependent or data-independent mass spectrometry. We developed a targeted, scheduled multiple reaction monitoring assay, termed ‘Mitoplex’, to enable the precise and focused quantification of a set of 32 mitochondrial proteins critical to central carbon chain metabolism and overall mitochondrial function. Mitoplex quantification of proteins was reproducible and sensitive in both mitochondrial enrichments and total lysates from cultured cell lines (mouse C2C12 skeletal muscle myoblasts) and tissue (mouse soleus muscle). Further, the Mitoplex replicated and extended previous observations of a large increase in mitochondrial protein expression during the C2C12 myoblast to myotube differentiation process. We then quantified a novel reduction in mitochondrial protein expression following exposure of C2C12 cells to the HMG-CoA reductase inhibitor simvastatin, which was further associated with a failure of myoblasts to fully differentiate and a reduction in overall mitochondrial ATP production. Finally, the Mitoplex quantified widespread reduction of skeletal muscle protein abundance in vivo following treatment of mice with simvastatin.  We conclude that Mitoplex is a focused, sensitive, and high-throughput quantitative tool to monitor mitochondrial protein content. We’ve demonstrated its the biological relevance by characterization of mitochondrial protein depletion as a potential mechanism driving the skeletal muscle myopathy following statin treatment, which may ultimately aid in identification of ways to prevent this debilitating drug side effect in humans.  

Tips and Tricks (if submitted):

Session: Poster, poster number: 003

A novel UHPLC-MRM-MS methodology for accurate, reproducible, and fast quantification of histone PTMs

Joseph Cesare; Zuofei Yuan; Steven Zhao; Peder Lund; Yekaterina Kori; Simone Sidoli; Josue Baeza; Hee Jong Kim; Kathryn E. Wellen; Benjamin A. Garcia
University of Pennsylvania, Philadelphia, PA

Histones are structural proteins that modulate chromatin structure and gene expression. Their post translational modifications (PTMs) recruit transcription factors and directly affect chromatin state through chemical interactions.  Aberrant levels of these modifications are found in many diseases such as over methylation of Histone 3 Lysine 4 (H3K4) and Histone 3 Lysine 27 (H3K27) in acute myeloid leukemia (AML) and multiple lineage leukemia (MLL) (Chi et al, Nature Reviews Cancer 2010). The ability to quantify these changes reliably, at low costs, and in high throughput experiments will allow for large scale studies and has the potential to revolutionize clinical diagnostics. Mass spectrometry (MS) is the method of choice for this analysis; however, it requires expensive instrumentation, each sample requires at least one hour of run time, and the use of nano-liquid chromatography is susceptible to batch effects. To address these limitations, we coupled analytical flow ultra-high-pressure liquid chromatography (UHPLC) to a triple quadrupole mass spectrometer using multiple reaction monitoring (MRM) acquisition mode to quantify 93 peptides and 80 known PTMs. Our method obtained highly reproducible results with a median coefficient of variation 3.9% in 20 minutes using a heavy labeled peptide library. This method was applied to human derived cell lines and mouse liver tissue and demonstrated more accurate quantification than nano-liquid chromatography coupled to data independent acquisition (DIA). For quality control and correction for batch effects, we included a peptide to indicate efficient digestion and chemical derivatization in each analysis. Furthermore, we corrected for ionization biases. The data analysis was integrated into our EpiProfile 2.0 software and found to be comparable to data analysis completed through manual inspection and Skyline software. Our new UHPLC-MRM-MS method has reduced the analysis time and increased reproducibility and accuracy in quantification opening new opportunities for translational medicine and clinical diagnostics.

Tips and Tricks (if submitted):

Session: Poster, poster number: 004-TL

Circulating hydroxylated bradykinin level corresponds with pancreatic tumor hypoxia and predicts treatment outcome

Yang Liu; Christopher Lyon; Jia Fan; Tony Hu
Arizona State University, Tempe, AZ

Tumor hypoxia is linked to poor prognosis due to its role in promoting cancer progression and metastasis. Evaluating tumor hypoxia to predict patient outcomes is challenging, however, due to the lack of rapid, accurate and non-invasive methods. Hypoxia positively regulates the α1 subunit of prolyl-4-hydroxylase (P4H) in several tumors, and P4Hα1 expression is involved in hypoxia-inducible modification of the plasma protein bradykinin (BK) to hydroxyprolyl BK (Hyp-BK), leading us to hypothesize tumor hypoxia could be detected by plasma Hyp-BK analysis. Here we show that P4Hα1 expression is selectively increased in malignant vs. non-malignant pancreatic cancer tissue, corresponding to tumor expression of Carbonic Anhydrase 9, a reported hypoxia marker. Hyp-BK/BK ratios in pre-treatment plasma samples of pancreatic cancer patients are also inversely correlated with their response to treatment, suggesting that plasma Hyp-BK/BK may have value as a non-invasive, surrogate marker for tumor hypoxia and disease prognosis in response to anticancer therapy.

Tips and Tricks (if submitted):

Session: Poster, poster number: 005-ML

Utilizing parallel reaction monitoring for a high throughput diagnostic pipeline to establish cardiac Troponin I phosphorylation as a biomarker.

Daniel Soetkamp; Weston Spivia; Qin Fu; Jennifer E. Van Eyk
Cedars-Sinai Medical Center, Beverly Hills, California

Introduction and Objectives

Cardiac Troponin I (cTnI) is the preferred diagnostic biomarker for various myocardial diseases. Phosphorylation of specific Ser/Thr residues on cTnI are relevant for several pathological conditions. In heart failure (HF) phosphorylation of residues S22 and S23 are reduced, whereas residues T143 and S199 are increased and are associated with decreased force generation and hypertrophic cardiomyopathy, respectively. This study focuses on introducing a highly sensitive, high throughput quantitative screening for cTnI, and its disease-associated posttranslational modifications in blood samples by mass spectrometric analysis.


An automated robotic high throughput intact protein immunoprecipitation assay was developed and optimized for identification and quantification of cTnI phosphorylation in patient plasma. Targeted approaches on different MS instruments (Qtrap 6500, Obitrap Elite and Orbitrap Fusion Lumos) using alternate flow velocities (high, low and nano) were tested for optimal sensitivity.   

Results and Discussion

Multiple factors in cTnI immunoprecipitation workflow were tested and optimized manually using recombinant cTn-complex in a cTnI depleted plasma matrix, e.g., optimal antibody concentrations, pull down duration, bead washing and the elution process. An established parallel reaction monitoring method on the Orbitrap Fusion Lumos was proven to provide the desired sensitivity to measure cTnI phosphorylation in plasma from patients with myocardial infarcts. Further optimization of automated antibody based cTnI enrichment from actual undiluted patient plasma led to alterations in the workflow of the automated IP assay, e.g., increased antibody-antigen incubation period, reduced washing and increased elution periods.


An automated high-throughput assay was successfully established that can quantify total cTnI and phosphorylation on residues S22, S23, T143 and S199 using a targeted MS approach in patient plasma samples with approximately 90% accuracy for total cTnI when compared to an ELISA based method.

Tips and Tricks (if submitted):

Session: Poster, poster number: 006-TL

Top-Down Proteogenomics Analysis of Serum Autoantibody Repertoire for the Discovery of Biomarker of Systemic Lupus Erythematosus

Zhe Wang1; Xiaowen Liu2; Kenneth Smith3; Si Wu1
1University of Oklahoma, Norman, OK; 2School of Informatics and Computing, IUPUI, Indianapolis, IN; 3Oklahoma Medical Research Foundation, Oklahoma City, OK

Systemic Lupus Erythematosus (SLE) is a multi-organ, systemic autoimmune disorder, affecting over 1.5 million Americans. The hallmark of SLE is the production of serum autoantibodies that are directly pathogenic, eventually causing organ damage and early mortality. Autoantibodies accrue years before clinical symptoms arise which may allow early diagnosis and treatment. While current diagnostic platforms can screen total autoantibodies, finding specific autoantibody biomarkers is impossible.

Top-down proteomics analyzes intact proteoforms with sequence variations and post-translational modifications, holding great potential to analyze serum autoantibodies of high similarity and diversity. However, there are two major challenges: (1) The serum autoantibody repertoire is extremely complexed (e.g., a wide range of highly homologous autoantibodies), requiring advanced separation approaches; (2) Personalized autoantibody database must be generated due to high diversity of autoantibody repertoire.

Herein, we developed a top-down proteogenomics platform integrating (1) a long-column ultra-high-pressure RPLC to efficiently separate highly homologous autoantibodies based on the hydrophobicity differences from sequence variations; (2) a personalized protein database from the B-cell next-generation sequencing. 86 Fab-related mass features were identified in longitudinal SLE serum samples from a patient, representing first top-down demonstration of serum autoantibody repertoire analysis. Two monoclonal autoantibodies with same CDR3 sequence were identified, one of which was functional characterized to have high affinity to a known antigen SmD through the fully human recombinant antibody. These two autoantibodies may result from the somatic gene conversion.  However, interestingly, they have different expression patterns over the patient’s disease progress, which may relate to the affinity maturation of individual antibodies.  Only CDR3 sequences (i.e., possibly from bottom-up) cannot pinpoint mature autoantibodies which are potential disease biomarkers. 

Our results demonstrate the potential application of top-down proteogenomics analysis of serum autoantibody repertoire. It holds great promise for discovering novel serum autoantibody biomarkers, as well as promoting our understanding of pathogenic autoimmune processes.

Tips and Tricks (if submitted):

Session: Poster, poster number: 007

Development of an MRM Assay to Distinguish Active and LAP TGF-β in Urine

Chelsea Boo; Raghothama Chaerkady; Sonja Hess
MedImmune, Gaithersburg,

Transforming growth factors β (TGF-β) are important cytokines that regulate cell proliferation, differentiation, and other biological processes. After cleavage of the signal peptide, they form latent TGF-β complexes in which TGF-β is noncovalently bound to latency-associated peptide (LAP). TGF-β mediates immune and anti-inflammatory responses. A method to distinguish between active TGF-β and LAP would provide critical insight into biological/disease processes; however, to the best of our knowledge, there are currently no mass spectrometry (MS)-based assays that differentiate these forms. Herein we have developed a multiple reaction monitoring (MRM) assay to distinguish active TGF-β and LAP.

A critical evaluation of sample solubility, reduction, and alkylation was performed in TGF-β-spiked urine and sample preparation conditions were selected to prevent protein precipitation and encourage complete carbamidomethylation of nine cysteines. Urine was obtained from normal and chronic kidney disease (CKD) donors and following optimized reduction and alkylation, the proteins were fractionated using a C3 column (acidic HPLC conditions). Fractions containing total active TGF-β and total LAP were spiked with iRT peptides and digested with Lys-C followed by trypsin. Relative quantitation was performed using isotopically-labeled (13C15N) synthetic standards to peptide targets from both active TGF-β and LAP.

The LC-MS/MS method is a robust method for the optimized detection of total active TGF-β and total LAP in normal and CKD donor urine. Active TGF-β and LAP levels were markedly elevated in the urine of CKD donors versus normal donors. The detection and quantitation of LAP was particularly critical to this project since previously, an ELISA assay was used to detect and quantitate active TGF-β and the presence of LAP in urine was unconfirmed. Current investigations now involve measuring fibrosis levels in urine donor samples and correlating to active TGF-β levels.


Tips and Tricks (if submitted):

Session: Poster, poster number: 008-TL

Signatures of ambient exposure to benzene and other air pollutants in the human serum albumin Cys34 adductome

Joshua Smith1; Robert O'Meally1; Derek Ng1; Thomas Kensler1, 2; Robert Cole1; John Groopman1
1Johns Hopkins University, Baltimore, MD; 2Fred Hutchinson Cancer Research Center, Seattle, WA

Benzene and other components of outdoor air pollution are human carcinogens. However, airborne concentrations of pollutants may not accurately predict exposure or internal dose. Covalent adducts to the Cys34 thiol in human serum albumin (HSA) can serve as internal dosimeters, while the t1/2 of HSA facilitates biomonitoring of the previous ~3 months. Analyzing the HSA adductome can simultaneously reveal unique biomarkers for many pollutants in a complex mixture. We have used Cys34 adductomics to characterize longitudinal exposures in Qidong, China, which experiences substantial airborne pollution and a rising incidence of lung cancer. Non-smoking women (n=8) from the placebo arm of a published clinical chemoprevention trial were selected for this pilot study. HSA was precipitated from serum collected at baseline (day 0) and day 56, digested with lysyl endopeptidase (LysC), and Cys34-adducted peptides were separated and quantified by online nanoflow LC interfaced with a Thermo Scientific Orbitrap Fusion Tribrid mass spectrometer. We detected four Cys34 adducts arising from three benzene metabolites: benzoquinone (BQ; BQ1, BQ2), benzene oxide (BO), and benzene diolepoxide (BDE). Incubation of control serum with 1,4-BQ in vitro demonstrated dose-dependence of BQ adducts and that they may reflect keto-enol tautomerism of a 1,4-BQ-derived adduct, but do not arise from 1,2-BQ. Along with benzene-derived adducts, we identified other putative adducts (ethylene oxide, nitriles, aldehydes, oxidation, disulfides) and revealed shifts in the Cys34 adductome during the 8-week follow-up: decreased BQ2 (p=0.08), BDE (p=0.06), acrylonitrile (p=0.08), homocysteine (p=0.05), and cysteinylglycine (p=0.08), but increased unassigned adducts (p=0.02). To our knowledge, this is the first report of benzene-derived adducts in humans with ambient exposures, detected with an adductomics approach. Work is underway to expand sample number, exposure timeline, and adduct identification. Cys34 adductomics represents a promising approach for the discovery of air pollutant biomarkers and quantification of an individual’s exposure to air pollutants.

Tips and Tricks (if submitted):

Session: Poster, poster number: 009

Global proteome and phosphoproteome alterations reveal novel drug targets to circumvent 3rd-generation EGFR TKI resistance in human lung adenocarcinoma

Xu Zhang1; Tapan Maity1; Karen Ross2; Shaojian Gao1; Khoa Dang Nguyen1; Fatos Kirkali1; Cathy Wu3; Udayan Guha1
1Thoracic and GI Malignancies Branch, CCR, NCI, NIH, Bethesda, MD; 2Georgetown University Medical center, Washington DC, DC; 3University of Delaware, Newark, DE

Osimertinib, a 3rd generation epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor (TKI) is approved for the treatment of lung adenocarcinoma patients harboring TKI-sensitizing or resistant (EGFRT790M) EGFR mutants. Here, we sought to identify novel resistance mechanisms to two 3rd-generation EGFR TKIs, osimertinib and rociletinib using SILAC-based quantitative mass spectrometry. The parental lung adenocarcinoma cell line, H1975 (EGFRL858R/T790M), and the isogenic osimertinib (AZR3,4) and rociletinib (COR1,10) resistant cells were employed in 3-5 biological replicates of global proteome and phosphoproteome estimation with/without TKIs. TiO2-enriched phosphopeptides underwent tandem mass spectrometry in QE-HF.  Overall, we identified 6752 proteins and 17,646 phosphosites. The abundance of hundreds of proteins and phosphopeptides was significantly altered between the TKI-sensitive and resistant cells. Surprisingly, there was a greater degree of inhibition of phosphorylation in resistant cells, including of EGFR and the downstream MAPK pathway. Several phosphatases, including PTPN11, PTPN12, and PTPN1 were predicted to be activated in the resistant cells. Proteins involved in epithelial mesenchymal transition (EMT) were expressed more in TKI-resistant cells, suggesting EMT as a mechanism of resistance. IPA analysis of proteins with significantly altered phosphorylation identified several signaling pathways activated (such as p53, PTEN, RhoGDI, RhoA and PPAR) or inhibited (including ERBB, AMPK, ERK/MAPK, RAC) in the resistant cells. We used iPTMnet, a bioinformatic resource that integrates data from text mining of scientific literature and other PTM databases to identify upstream kinases of phosphorylation targets. AKT and PKA kinase substrates were hyper-phosphorylated in TKI-resistant cells. We also used kinase-inhibitor target data (KinomeScan) and phosphoproteomic data (P100) generated by the NIH LINCS Program to identify drugs that might overcome resistance. The TKI resistant cells were indeed sensitive (EC50:18-130nM) to one of the predicted drugs, dactolisib, a PI3K/MTOR inhibitor.  In vivo studies are underway in mouse models to demonstrate the efficacy of dactolisib in circumventing TKI resistance.

Tips and Tricks (if submitted):

Session: Poster, poster number: 011

Identifying breast cancer vulnerabilities by mapping interactome dysregulations in primary tumor samples

Robert Morris1, 2; Johannes Kreuzer1, 2; Ridwan Ahmad1, 2; Cyril H. Benes1, 2; Dennis C. Sgroi1, 2; Wilhelm Haas1, 2
1Massachusetts General Hospital, Charlestown, MA; 2Harvard Medical School, Boston, MA

The promise of cancer precision medicine is to treat patients based on the specific molecular properties of their tumors.  However, identifying patient-specific tumor vulnerabilities still remains an enormous challenge.  We have developed a novel technology that allows mapping dysregulations of protein-protein interactions (PPI) in an interactome-wide manner.  The method is based on (i) high-throughput quantitative proteome mapping of tumor samples using TMT-11 reagents, and MS2/SPS-MS3 method on Orbitrap Fusion and Lumos instruments, (ii) identification of PPIs through protein co-regulation analysis, and (iii) mapping of PPI dysregulations through monitoring sample specific deviations of protein-protein co-regulations to identify tumor vulnerabilities.  Our previous work with cancer cell lines has shown the high potential of this approach, and we have now applied the method in 320 samples of primary breast cancer tumors (254) and normal breast tissue (66) received from the MGH breast cancer tumor bank.  The samples represent all subtypes according to the Perou classification – luminal A, luminal B, HER2+, and triple-negative (TNBC).  To our knowledge this is currently the largest proteomics dataset of primary tumor samples.  Across all 300 samples, we have quantified more than 11,000 proteins with an average of 7,995 proteins quantified for each sample.  Protein concentration co-regulation analysis resulted in identifying more than 20,000 protein-protein association.  PPI dysregulation analysis revealed an average of more than 200 dysregulated interactions tumor.  We will present tumor and subtype-specific vulnerabilities and therefore potential targets in subtype-specific treatment strategies.  We found that the identified interactome dysregulation include pathways known to be perturbed in the specific pathways based on the known subtype-specific driver mutations.  We will also present a strategy to identify novel treatment target candidates by overlaying interactome dysregulation data from tumors and cancer cell lines, which vulnerabilities have been mapped by genome-wide genetic drop-out screens.

Tips and Tricks (if submitted):

Session: Poster, poster number: 012

A novel role α (1,6) fucosyltransferase, a glycosylated related enzyme in the biology of castration resistant prostate cancer.

Naser Uddin
Johns Hopkins School Of Medicine, Baltimore, Maryland

Glycosylation is recognized as one of the most common modifications on proteins. Recent studies have shown that aberrant expression of α (1,6) fucosyltransferase (FUT8), which catalyzes the transfer of fucose from GDP-fucose to core-GlcNAc of the N-linked glycoproteins, modulates cellular behavior that could lead to the development of aggressive prostate cancer. While the relationship between the abnormal expression of FUT8 and glycoprotein fucosylation in different prostate cancer cells has been demonstrated, there is no evidence that shows dysregulated fucosylation might be involved in prostate cancer progression from androgen-dependent to castration-resistant disease. In this study, using a proteomics approach, we analyzed androgen-dependent and androgen-resistant LAPC4 cells and identified FUT8 to be significantly overexpressed in the androgen-resistant LAPC4 cells. These findings were independently confirmed in LAPC4 cells that were treated with non-steroidal anti-androgen (bicalutamide) and in the in vivo castrated tumor xenograft models. We further evaluated whether FUT8 expression alone might be responsible for the androgen resistant phenotype, by exogenously overexpressing FUT8 in prostate cancer cells, we found a substantial decreased in cellular PSA expression that was consistent with FUT8 expression and PSA production in prostate cancer specimens. To our knowledge, this is the first study reporting the functional role of fucosylated enzyme in the development of castration-resistant prostate cancer.

Tips and Tricks (if submitted):

Session: Poster, poster number: 014-WTT

Integrating kinetic and quantitative proteomics to investigate autophagy substrates in tumors

Monique Speirs; John Price
Brigham Young University, Provo, UT

The activation of the catabolic pathway autophagy is strongly connected to cancer stress tolerance and chemoresistance. Pharmacological inhibition of autophagy is a potential strategy to sensitize tumors to traditional chemotherapeutics. Current clinical trials are attempting to control autophagy flux using broad-spectrum inhibitors with limited knowledge of the side effects of these drugs on protein homeostasis in healthy surrounding tissues. We used D2O-based metabolic labeling and mass spectrometry combined with standard quantitative proteomics techniques to measure changes in protein degradation in a xenograft mouse model upon autophagy inhibition. We found that autophagy substrates differ in the tumor versus liver tissue. Our results suggest that autophagy is used to selectively degrade mitochondria (mitophagy) in tumor cells, perhaps to regulate mitochondrial function and biogenesis. In the liver, autophagy is used broadly as a bulk degradation system to maintain global proteostasis. These experiments support a model of modified mitochondrial metabolism, enforced in part by rapid autophagic degradation, as a pro-survival mechanism in cancer. We also observed off-target effects in the liver consistent with a toxicity response to treatment. These methods are directly applicable to humans and could test for subject-specific differences in autophagy targeting and guide the development and translation of substrate-specific autophagy inhibitors.

Tips and Tricks (if submitted):

Utilizing protein turnover and concentration to monitor in vivo protein degradation rates

Session: Poster, poster number: 015

A chemical biology screen identifies a metabolic vulnerability of neuroendocrine cancer cells to SQLE inhibition

Sebastian Hayes
Agios, Cambridge, MA

Aberrant metabolism of cancer cells is well appreciated, but the identification of cancer subsets with specific metabolic vulnerabilities remains challenging.  We conducted a chemical biology screen and identified a subset of neuroendocrine tumours displaying a striking pattern of sensitivity to inhibition of the cholesterol biosynthetic pathway enzyme squalene epoxidase (SQLE).  Using parallel approaches, including deep proteomic profiling of 42 small cell lung cancer lines with varying degrees of sensitivity to SQLE inhibition, we were able to identify a subset of cell lines with distinct proteomic and transcriptomic signatures as being the most susceptible to SQLE inhibition.  We employed mass spectrometry-based metabolomics analyses to demonstrate that sensitivity to SQLE inhibition results not from cholesterol biosynthesis pathway inhibition but from the specific and toxic accumulation of the SQLE substrate, squalene. These findings highlight SQLE as a potential therapeutic target in a subset of neuroendocrine tumors, particularly small cell lung cancers.

Tips and Tricks (if submitted):

Session: Poster, poster number: 016

TRIM28 as a candidate mutant p53 interacting partner in cancer cells

Mariel Mendoza; Katherine Alexander; Enrique Lin Shiao; Charly Ryan Good; Benjamin A. Garcia; Shelley L. Berger
University of Pennsylvania, Philadelphia, PA

p53 is a transcription factor that is mutated in over 50% of cancers. Missense mutations in the DNA binding domain of p53 can result in a gain-of-function (GOF) phenotype, leading to increased cell proliferation and tumor formation. Our lab previously showed that prevalent mutant p53 (mtp53) forms modify chromatin through their interaction with ETS2 and activation of non-canonical transcriptional targets (MOZ, MLL1, and MLL2). Aside from ETS2, other mtp53 partners that have been identified, including Sp1, NF-Y, and PML. However, whether specific proteins are critical for the stability and the GOF effect of mtp53 remains to be seen. To this end, we developed a quantitative mass spectrometry-based strategy, combined with molecular and genomic approaches, to identify and validate novel mtp53 binding partners from cancer cell lines with varying GOF p53 mutations. Our preliminary data identified the transcriptional corepressor TRIM28 as a candidate mtp53 interacting partner, as it was identified in all 4 GOF cell lines tested (VU1365, HUPT3, MDA468, and PANC1). Knockdown of TRIM28 in MDA468 cells (contains R273H mutation in p53) caused a decrease in cell viability. TRIM28 has been shown to interact with MDM2 to promote wild type p53 ubiquitylation and degradation; however, its role in regulating mtp53 has not been determined. Ultimately, our studies will identify and validate further novel proteins critical for the GOF activity of mtp53. Characterizing these novel interacting partners of mtp53 will shed light into the molecular mechanisms underlying cancer and thus will provide new therapeutic targets to destabilize mutant p53 interactions in cancer cells.

Tips and Tricks (if submitted):

Session: Poster, poster number: 017-ML

Detection of malignancy-associated proteome and phosphoproteome alterations in human colorectal cancer induced by cell surface binding of growth-inhibitory galectin-4

Malwina Michalak1; Uwe Warnken2; Hans-Joachim Gabius3; Martina Schnölzer2; Jürgen Kopitz1
1Heidelberg University Hospital, Heidelberg, Germany; 2German Cancer Research Center (DKFZ), Heidelberg, Germany; 3Ludwig-Maximilians-University Munich, Heidelberg, Germany

Carbohydrate-binding proteins, called lectins, translate glycan-encoded information on the cell surface into effects on fundamental biological processes, including cell growth, differentiation and adhesion. One of them, galectin-4 (gal-4), exhibits tumor suppressing actions and has been shown to influence colorectal cancer tumorigenesis and progression. Accordingly, we have found that gal-4 cell surface binding induced cell differentiation and reduced proliferation in five human colorectal cancer cell lines (LS 180, Vaco 432, Colo 205, CX 1, HCT 116). In order to explore changes underlying observed shifts in growth behavior and morphology on a molecular level, SILAC-based proteomics and phosphoproteomics were applied in LS 180 cells treated with gal-4. Proteome investigation disclosed down-regulation of DNA replication-associated processes, while protein presence for secretory and transport functions appeared increased. Profound down-regulation was observed for malignancy-associated proteins, such as EPCAM (epithelial cell adhesion molecule; ~6-fold), LMNB1 (lamin-B1; ~5-fold) and common cell proliferation marker Ki-67 (~4-fold), accompanied by strongly up-regulated CALB2 (calretinin; ~24-fold) and transglutaminase TGM2 (~11-fold), among others. As a part of interactive galectin network, gal-4 caused an intracellular increase of adhesion/growth regulatory galectin-1 (~4-fold). Further phosphoproteomic analysis at two time points revealed early and sustained effects induced by gal-4. Early reaction indicated modulation of membrane organization, vesicle-mediated transport as well as microtubule-kinetochore attachment by BET1 hypophosphorylation (~3-fold) and CENPF hyperphosphorylation (centromere protein F; ~2-fold), among others. Sustained phosphorylation changes suggested the role of cytoskeleton organization and glutamine metabolism by altered phosphorylation of CFL1 (cofilin S3; ~5-fold) and glutamine transporter SLC1A5 (S503; ~3-fold), proven by significant effect of gal-4 treatment on glutamine uptake in the analyzed cell lines. Altogether, detected proteome and phosphoproteome changes are potentially relevant for the observed phenotypic shift, giving new insights into gal-4 functions and revealing the capacity of galectins to affect human colon cancer cells at multiple sites.

Tips and Tricks (if submitted):

Session: Poster, poster number: 018

Proteomic analysis of uveal melanoma derived exosomes

Blake Ebert; Alex J. Rai
Columbia University, New York, NY

Uveal melanoma (UM) is a rare form of non-cutaneous melanoma that develops from melanocytes in the uveal tract, forming a malignant tumor of the eye. Metastatic disease frequently involves the liver and is associated with poor outcomes – resulting almost always in death within 15 months. Recent evidence suggests that organotropic metastasis can be mediated through exosomes, which are membrane bound nanovesicles released from all cells and contain proteins, nucleic acid, and other small molecules. We have characterized the exosomal proteome in an effort to gain insight into the molecular mechanisms of metastasis. We first demonstrated successful exosomal isolation through western blotting of core exosomal proteins and visualization of these vesicles using transmission electron microscopy. We treated exosomes with crizotinib, a chemotherapeutic agent currently in clinical trials for UM, and performed comprehensive, label-free, proteomics on exosomes from the treated and untreated cells. We identified >600 proteins in total and ~150 proteins unique to each group. Based on gene ontology and literature review, we identified 16 high-value biomarker candidates and assessed their levels in exosomes derived from 92.1 (UM) cell culture media and clinical samples from UM patients. Finally, to determine the effect of exosomes and drug treatment on metastasis, we treated liver cells with exosomes isolated from cell culture media. Exosomes from untreated cells induced multiple changes in target hepatocytes, including alterations in signaling pathways affecting angiogenesis, extracellular matrix invasion, and cell-cell adhesion.  These changes were abrogated when hepatocytes were treated with UM exosomes isolated after incubation with crizotinib. We demonstrate the effect of crizotinib on the selective packaging and distribution of proteins within exosomes, and develop a list of protein biomarkers potentially useful for clinical management of UM patients. Further, our studies suggest that exosomes can induce changes in hepatocytes and that crizotinib treatment dampens this response. 

Tips and Tricks (if submitted):

Session: Poster, poster number: 019-ML

Bioinformatics approach for understanding the role of intrinsic disordered regions in cancer-related proteins 

Rita Hayford; Cathy Wu; Cecilia Arighi
University of Delaware, Newark, DE

Intrinsically disordered regions (IDRs) are protein regions that do not fold into stable secondary or tertiary structure under physiological conditions. They are usually biologically active and associated with important cellular activities such as regulation, signaling, and control. IDRs have been implicated in many diseases, such as cancer, cardiovascular and neurodegenerative diseases. Given that IDRs provide structural plasticity and functional diversity, we hypothesize that proteins associated with cancer contain disordered regions which play a critical role in regulating their function. Here we collected a list of human proteins involved in cancer from the UniProt Knowledgebase and the curated cancer gene list from the Catalogue of Somatic Mutations in Cancer (COSMIC). The cancer-related proteins were screened for the presence of disordered regions using MobiDB, a database for disordered proteins. Then, we mapped potential regulatory features, such as post-translational modifications (PTMs), motifs, regions, and repeats from the individual proteins to the IDRs, and we also looked for variants or mutagenesis information on these regions. Our data suggest that a large percentage of the cancer-related proteins contain IDRs spanning at least nine amino acid residues long (~70% of 278 total proteins collected). We found that IDRs of proteins associated with cancer were rich in PTMs (mainly phosphorylated sites); their IDRs encompassed regions involved in protein interactions, and approximately 11% of the non-synonymous SNPs mapped to IDR regions. Additionally, the functional classification of the IDR protein set under study was analyzed and the terms enriched in this set included transcriptional regulation and transcription. The pathways enriched in the IDRs set include the Wnt signaling pathway. Our data integrates the current knowledge on IDRs and functional information in cancer proteins and may provide new light into cancer disease mechanism and development of therapeutic approaches. 

Tips and Tricks (if submitted):

Session: Poster, poster number: 020

Early urine proteome changes in the Walker-256 tail-vein injection rat model.

Jing Wei1; Na Ni2; Wenshu Meng1; Youhe Gao1
1Beijing Normal University, Beijing, China; 2Chongqing Medical University, Chongqing, China

Detection of cancer at its early stage is important for treatment. Urine, which is not regulated by homeostatic mechanisms, reflects early systemic changes throughout the whole body and can be used for the early detection of cancer. In this study, the Walker-256 tail-vein injection rat model was established to find whether the urine proteome could reflect early changes if tumor grown in lung. Urine samples from the control group (n=7) and Walker-256 tail-vein injection group (n=7) on days 2, 4, 6 and 9 were analyzed by label-free proteomic quantitative methods. On day 2, when lung tumor nodules did not appear, 62 differential proteins were identified. They were associated with epithelial cell differentiation, regulation of immune system processes and the classical complement activation pathway. On day 4, when lung tumor nodules appeared, 72 differential proteins were identified. They were associated with the innate immune response and positive regulation of phagocytosis. On day 6, when body weight began to decrease, 117 differential proteins were identified. On day 9, the identified 125 differential proteins were associated with the B cell receptor signaling pathway and the positive regulation of B cell activation. Our results indicate that (1) the urine proteome changed even on the second day after tail-vein injection of Walker-256 cells and that (2) compared to previous studies, the urine proteomes were different when the same cancer cells were grown in different organs.

Keywords: cancer biomarkers, urine proteome, lung tumor, early detection

Tips and Tricks (if submitted):

Session: Poster, poster number: 021

Identification of Smyd1’s Chromatin Binding Partners via ChIP-MS

Anna Bakhtina; Aman Makaju; Sarah Franklin
University of Utah, Salt Lake City, UT

As the heart experiences stress, it compensates by undergoing hypertrophic growth which, left unchecked, eventually progresses to heart failure. Pathological cardiac hypertrophy is preceded by the re-expression of genes normally only activate during development which can only be accomplished by significant remodeling of the chromatin in order for key regions of DNA to be silenced or reactivated. Although recent studies have demonstrated that altering key epigenetic mechanisms can inhibit these gene-expression changes and prevent disease-induced growth, our knowledge regarding the chromatin modifiers driving cardiac pathology is quite limited. We have previously shown that loss of the histone methyltransferase, Smyd1, can be deleterious for the heart, however, the mechanisms by which the two Smyd1 variants (Smyd1a and Smyd1b) regulate gene expression changes and how this maintains basal cardiac function and influences heart failure onset is not known. In this study we used chromatin-immunoprecipitation coupled with tandem mass spectrometry (ChIP-MS) to identify the Smyd1 binding partners on chromatin and the post-translational modifications found on nucleosomes they bind. Specifically, we utilized adenoviral-mediated expression of Flag-tagged Smyd1a or Smyd1b in C2C12 cells and primary rat ventricular myocytes (under normal and hypertrophic conditions). Our analyses identified a number of novel interactors of Smyd1a and Smyd1b on chromatin (31 and 12, respectively) which included both shared and variant-specific proteins. Interestingly, we identified BTF3 and Ybx1 as novel transcription factors which were found to exclusively bind Smyd1a and Smyd1b, respectively. Additionally, we showed that nucleosomes bound by Smyd1 are enriched in histone H3 di- or tri-methylated on K37, mono-methylated on R40/K80, and void of methylation or acetylation marks on K9/K14. In addition, these nucleosomes are enriched in histone H2A type-1. Overall, these results begin to unravel the mechanistic basis by which these histone methyltransferases regulate gene expression changes in the genome to influence myocyte morphology and physiology.

Tips and Tricks (if submitted):

Session: Poster, poster number: 022

Conservation and Divergence of Protein Pathways in the Vertebrate Heart

Vcrdxtm4iq93+Oy3zoeuximmp8ll0uo5zndtcdo3jpi= Conlon
University of North Carolina, Chapel Hill, NC

Congenital malformations, or structural birth defects, are now the leading cause of infant mortality in the United States and Europe. Of the congenital malformations, congenital heart disease (CHD) is the most common.  Studies of human heart disease and treatments have relied on vertebrate model systems with a dual circulatory system, including Xenopus, mouse, and pig. However, little is known about which proteins and protein pathways are conserved and which have diverged between these model systems. To address these questions we examined heart tissue proteomes of Xenopus tropicalis, Xenopus laevis, Mus musculus, and Sus scrofa and assessed protein abundance changes in the context of pathways, protein complexes, and enrichment of corresponding genes in human heart diseases.  This was achieved by dissecting heart tissue and subjecting the tissue to differential solubilization followed by gel fractionation, digestion, and analysis on a Q Exactive HF (ThermoScientific).  Differential heart proteome assessment was performed by label-free quantitation using Proteome Discoverer 2.1 and Scaffold.  Over 9,000 proteins were identified across all four species and were mapped to orthologous human proteins to assist with bioinformatic analysis.  Quantitative differences revealed species-specific enrichments in selected pathways.  One example is a pronounced enrichment of cell cycle associated proteins in Xenopus laevis compared to the other species, which we confirmed via targeted mass spectrometry. Thus, the dataset provided here will facilitate selection of appropriate models for studies of specific genes.  To demonstrate the utility of the dataset, we tested the role of the Kcp protein, which is present in the hearts of humans and frogs, but not present in mice or pigs.  Strikingly, germ-line null mutations in Kcp in Xenopus lead to cardiac failure and death.  Thus, our data provide a new and better road-map for establishing animal models for the genetic basis human cardiac development and disease. 



Tips and Tricks (if submitted):

Session: Poster, poster number: 023-ML

Broad time-dependent proteomic and metabolomic effects of atorvastatin on hepatocytes

Akos Vertes1; Albert-Baskar Arul1; Andrew R. Korte1; Peter Avar1; Lida Parvin1; Ziad J. Sahab1; Deborah I. Bunin2; Merrill Knapp2; Andrew Poggio2; Carolyn L. Talcott2; Brian M. Davis3; Christine A. Morton3; Christopher J. Sevinsky3; Maria I. Zavodszky3
1Dept. of Chemistry, The George Washington Univ., Washington, DC; 2SRI International, Menlo Park, CA; 3GE Global Research, Niskayuna, NY

Cholesterol plaques built up in arteries can slowly reduce blood flow leading to ruptures and heart disease. Atorvastatin, and other drugs from the statin class, provide cardiovascular protection by blocking hydroxymethylglutaryl coenzyme A reductase (HMGCR), an essential enzyme in the cholesterol biosynthesis pathway. Over and above its primary action, atorvastatin has a range of well-established adverse effects and cholesterol-independent beneficial or “pleiotropic” effects, including reduction of inflammatory mediators, improvement of endothelial function and antioxidant mechanism.
While hypothesis driven investigations on the cellular level have unveiled many of these effects, systematic mass spectrometry-based experiments have the potential to provide further insight into related molecular mechanisms. Here we show that cell culture based untargeted profiling of time-dependent proteomic and metabolomic effects of atorvastatin can capture changes related to the known mechanism of action, and also provides a comprehensive view of adverse and pleiotropic effects. Our metabolome analysis captured the changes in many lipid features, for example the decreases of lanosterol and 4-methylzymosterol levels, in terpenoid biosynthesis and related pathways, whereas comprehensive analysis of the protein abundances showed significant changes for 86 proteins, including increased abundances of HMGCR, squalene synthase and apolipoprotein-A. As most features affected by atorvastatin are already well identified in the literature, we were able to pinpoint changes that either support existing hypotheses or represent undiscovered details of the diverse effects of atorvastatin. Among others we observed decreased abundances for DNA topoisomerase 2-alpha and increased abundances for Alpha-1-anti-trypsin. Our time-dependent analysis (based on 10 timepoints) revealed that approximately one fifth of the significant changes were missed by traditional single timepoint analyses, which demonstrates the importance of temporal profiling and systemic approaches.

Tips and Tricks (if submitted):

Session: Poster, poster number: 024

Site-specific proteomic profiling using a novel chemical probe identifies new members in the deubiquitinase family

Taylur Ma1; David Hewings5; Johanna Heideker6; Andrew Ah Young1; Farid El Oualid2; Alessia Amore2; Gregory Costakes2; Daniel Kirchhofer1; Bradley Brasher3; Thomas Pillow1; Nataliya Popovych1; Till Maurer1; Carsten Schwerdtfeger3; William Forrest1; John Flygare3; Matthew Bogyo4; Ingrid Wertz1; Kebing Yu1
1Genentech, Inc, South San Francisco, ; 2UbiQ Bio BV, Amsterdam, Netherlands; 3Boston Biochem Inc, Cambridge, MA; 4Stanford University, Stanford, CA; 5Roche Inc, Basel, Switzerland; 6Center for Aids Research, Thermo Fisher, San Diego, CA

Activity-based protein profiling (ABPP) may be a powerful tool to explore in situdeubiquitinase (DUB) activity to aid the discovery of ubiquitin-based therapeutics. A common workflow consists of DUB enrichment by an activity-based probe and subsequently identified with mass spectrometry. In principle, inferring DUB activity from probe reactivity requires that the probe reacts with the enzyme at its active site; however, this is rarely verified since the labeling sites are usually not determined. In some cases, there is no indication that a protein enriched by the probe is truly an active DUB, or merely an ubiquitin binder. Here we present an enhanced chemoproteomic approach to discover deubiquitinating enzymes and study their activity through reaction sites.

Tips and Tricks (if submitted):

Session: Poster, poster number: 025

Use of peptide biosensor and PRM for measuring kinase activity

Tzu-Yi Yang1; Monica Johnson2; Laurie L. Parker3
1University of Minnesota, Minneapolis, MN; 2UMN, Minneapolis, MN; 3University of Minnesota Twin Cities, Minneapolis, MN

Our group uses peptide biosensors for measuring kinase activity in cells. Peptide sequences that can be selectively phosphorylated by the Abl kinase are synthesized. Stable isotope labeled Alanine is incorporated to create four isocratic biosensors (mass differences +0, +4, +7, +10 Da). Each isocratic biosensor was used for measuring Abl kinase activity in K562 cells under different treatment conditions. Equal amounts of lysates from different treatments were mix and the biosensors’ Tyrosine phosphorylation was quantified using PRM. PRM results show about 2% and 20% biosensor phosphorylation under vehicle and pervanadate treatments, respectively. Phosphorylation was not detected when cells were treated with Gleevec, an Abl kinase inhibitor. Our next goal is to quantify both biosensor and endogenous kinase phosphorylation using SWATH-MS.

Tips and Tricks (if submitted):

Session: Poster, poster number: 026

Changes to Human Chromatin Induced by Cytomegalovirus Immediate Early Proteins

Abigail A. Lemmon1; Tyler T. Miller2; Jen Liddle2; Katarzyna Kulej2; Simone Sidoli1; Daphne C. Avgousti2; Benjamin A. Garcia1; Matthew D. Weitzman1, 2
1University of Pennsylvania School of Medicine, Philadelphia, PA; 2Children's Hospital of Philadelphia, Philadelphia, PA

The herpesvirus Human Cytomegalovirus (HCMV) latently infects over half the adult population and can cause serious disease in immunocompromised patients. The double-stranded DNA herpesvirus expresses immediate early proteins, most abundantly immediate early 1 (IE1) and immediate early 2 (IE2), to promote viral infection through largely unknown mechanisms. IE1 interacts with host chromatin by binding to host histones through its chromatin-tethering domain (CTD), which binds the acidic patch between histones H2A and H2B. We investigate the effects of IE1 binding on chromatin. Using mass spectrometry, we assessed the differences in histone post-translational modifications (PTMs) in cells expressing IE1, IE1∆ (deleted CTD), or IE2. Our findings reveal significant changes in several histone post-translational modifications with over-expression of IE1. Modifications to the core histones correspond with cell cycle progression and changes in gene transcription, and the functional significance of a double modified linker histone H1 peptide is still under investigation.

Tips and Tricks (if submitted):

Session: Poster, poster number: 028

Adapting EasyPepTM MS Sample Preparation for 96-well Automated Liquid Handling Systems

Sergei Snovida1; Ryan Bomgarden1; Amarjeet Flora1; Xinyu Zhang2; Emily I. Chen2; John C. Rogers1
1Thermo Fisher Scientific, Rockford, IL; 2Thermo Fisher Scientific, Cambridge, MA


Advances in mass spectrometry (MS) instrumentation has enabled routine analysis of complex protein samples. However, sample preparation methods are not standardized with many protocols taking 8-24hrs in addition to suffering from low peptide yields, poor digestion efficiency and low reproducibility. Recently, we developed a new, simplified sample prep kit containing pre-formulated reagents and a standardized protocol for processing 10µg to 100µg protein samples in less than 2 hours. In this study, we adapted our chemistry to use a 96-well filter plate for peptide clean up and assessed this format using an automated liquid handling system.


Cellular protein extracts and human plasma samples were diluted in lysis buffer. A universal nuclease was added to cellular extracts to reduce sample viscosity. Protein samples were heated at 50°C for 10 minutes in the presence of a combined reduction/alkylation solution before digestion using a trypsin/LysC protease mixture.  A mixed mode peptide clean-up procedure using a 96-well filter plate format was evaluated using a vacuum manifold or automated positive-pressure system. Peptides were quantified and normalized using the PierceTM Quantitative Colorimetric Peptide Assay prior to LC-MS analysis using a Thermo Scientific™ Q Exactive™ HF hybrid quadrupole-Orbitrap™ mass spectrometer.


Preliminary Data

Although our optimized protocol using in our EasyPep kit significantly reduces both hands-on and total sample processing time, peptide clean up using microcentrifuge spin columns is still time consuming with larger sample numbers.  We developed a new 96-well filter plate format to support higher sample processing throughput.  This format showed nearly identical performance in terms of peptide yield, identification rates, alkylation efficiency and digestion efficiency compared to the manual spin column protocol with better reproducibility among replicates. 


Overall, we demonstrate that our chemistry is readily adaptable to automated liquid handling system which provides excellent reproducibility and greatly simplifies proteomic sample preparation.

Tips and Tricks (if submitted):

Session: Poster, poster number: 029

timsTOF Pro: Maximum throughput, robustness and analytical depth for shotgun proteomics

Scarlet Koch1; Tharan Srikumar2; Christopher Swift2; Christopher Adams3; Heiner Koch1; Thomas Kosinski1; Gary Kruppa2
1Bruker Daltonic, Bremen, Germany; 2Bruker Daltonic, Billerica, MA; 3Bruker Daltonic, San Jose, CA

Sample throughput (>40 runs in 24 hr) in proteomics with reasonable proteome depth is highly desirable. The timsTOF Pro with trapped ion mobility spectrometry (TIMS) offers additional separation power and increased peak capacity. TIMS enables the powerful Parallel Accumulation Serial Fragmentation (PASEF) method [1,2] for very high sequencing speed. These features are perfectly suited for proteome analysis on short gradients. Here, we demonstrate the performance of the timsTOF Pro mass spectrometer using short gradients, small sample loads and hundreds of LC MSMS runs in a 24 hr time period.

Hela digest (< 250 ng) was LC separated either using the nanoElute HPLC (Bruker Daltonics) or Evosep One system (Evosep) on-line coupled to a high-resolution TIMS-QTOF instrument (timsTOF Pro, Bruker Daltonics). For all analyses we applied the PASEF method. Data analysis was performed using PEAKS (Bioinformatics solutions Inc.) and MaxQuant (Jürgen Cox, Max Planck Institute of Biochemistry).

When analyzing 50 ng of HeLa digest using the Evosep One system in combination with a TIMS-QTOF mass spectrometer, 1400 proteins could be reproducibly (R2 > 0.97) identified using short gradients of 5.6 min. This very short gradient gives the possibility to analyze up to 200 samples/day. Increasing the measurement time per sample (measurement of up to 50 samples/day) and using a nanoElute HPLC system allows the identification of more than 4000 protein groups from 250 ng of HeLa digest, enabling the analysis of large sample cohorts with a reasonable proteome depth of coverage and short measurement time. Moreover, measurements of 24 concatenated fractions of HeLa digest allowed the identification of more than 100,000 unique peptides and 9,052 protein groups in less than 12 h measurement time. Our results show that an ultimate analytical depth can be achieved using short gradients.

Tips and Tricks (if submitted):

Session: Poster, poster number: 030

Identification of Inconsistent Peptide Recovery and Aberrant Peptide Termini as Sources of Sample Variability in Patient-derived Tumor Samples

Meghan Burke; Zheng Zhang; Yuri Mirokhin; Dmitrii Tchekhovskoi; Stephen Stein
NIST, Gaithersburg, MD

NIST performance metrics (Rudnick et al, MCP, 2009) have been extended to individual samples in multiplexed experiments, such as iTRAQ- or TMT-labeled peptides, as well as results obtained from the recently developed hybrid mass spectral library search.  In addition, we have compared normalized reporter ion abundance, relative to a reference channel, across retention time between digests which has identified tumor samples with inconsistent peptide recovery thereby resolving the previously unexplained bimodal reporter ion intensity distribution observed in data obtained from Clinical Proteomic Tumor Analysis Consortium (CPTAC). 

Variation in reporter ion relative abundance across over 300 distinct DeltaMass values, obtained from hybrid search results, was also analyzed.  The hybrid search can identify peptides that contain unanticipated modifications by matching both ion m/z and mass losses (Burke et al, JPR, 2016).  The DeltaMass, or mass difference between the query and reference spectral library peptide, is expected to correspond to a modification.  Outlier modifications identified include amino acid losses and additions. Further analysis revealed amino acid additions were found to be localized to the C-terminus and losses localized to the N-terminus.  Moreover, select amino acids are preferred, meaning that not all amino acids are equally likely to be truncated or added to a peptide sequence, and this preference also varies across laboratory and experiment.  The presence and variability observed in peptides containing aberrant termini, which may be due to non-tryptic proteolytic activity or sample degradation, is found to contribute to sample variation. 

Tips and Tricks (if submitted):

Session: Poster, poster number: 031

MaxQuant software for trapped ion mobility enhanced shotgun proteomics

Christoph Wichmann1; Nikita Prianichnikov1; Heiner Koch2; Marcus Lubeck2; Chris Adams3; Scarlet Koch2; Gary Kruppa4; Juergen Cox1
1Max Planck Institute of Biochemistry, Martinsreid, Germany; 2Bruker Daltonic, Bremen, Germany; 3Bruker Daltonic, San Jose, CA; 4Bruker Daltonic, Billerica , MA

Trapped ion mobility spectrometry (TIMS) provides an additional dimension of separation to LC-MS which boosts proteome coverage, quantification accuracy and dynamic range in shotgun proteomics experiments.  Required for this is suitable software that extracts the information contained in the 4D data space spanned by m/z, retention time, trapped ion mobility and signal intensity. Here, we describe the trapped ion mobility enhanced MaxQuant software, which utilizes the added data dimension.  

We performed benchmark measurements on a Bruker timsTOF Pro instrument with ‘Parallel Accumulation Serial Fragmentation’ (PASEF) functionality for the acquisition of MS/MS spectra. We generated benchmark datasets in which HeLa proteins were spiked with E. coli and yeast proteomes at 1:2 and 1:5 ratios and compared the replicates with the MaxLFQ algorithm for label-free quantification.

We adapted the complete MaxQuant shotgun proteomics workflow to process data with the added trapped ion mobility dimension. Most adaptations were done in the feature detection workflow which now produces 4D peaks. De-isotoping and assembling of MS1 labeling multiplets utilize intensity profile correlations also over the ion mobility direction. Matching MS1 features between runs, the transfer of identifications without MS/MS identifications, is making use of the ion mobility coordinates to become more specific. Peptide identification, FDR, protein assembly, label-free quantification and isobaric labeling quantification algorithms are included. Application to benchmark dataset showed unprecedented identification depth in single shot experiments and precise ratio quantification. MaxQuant offers an end to end computational workflow for the identification and quantification in LC-TIMS-MS/MS shotgun proteomics data including CCS value determination to reduce proteomics FDR and enhance the LFQ performance and the match between runs functionality using the 3D (m/z, RT and CCS).

Tips and Tricks (if submitted):

Session: Poster, poster number: 032

Reproducible quantitative mass spectrometry-based research: the MSstats perspective

Meena Choi1; Ting Huang1; Tsung-Heng Tsai1; Eralp Dogu2; Sara Mohammad Taheri1; Olga Vitek1
1Northeastern University, Boston, MA; 2Mugla Sitki Koçman University, Mugla, Turkey

We present a statistical perspective on reproducible quantitative mass spectrometry-based proteomics. Statistical components of reproducibility include experimental design, from both biological perspective (which proteins and samples, and how many, do we need to quantify?) and technological perspective (are the assays appropriate for the task? Do the experimental steps run properly?). Statistical components of reproducibility also include data processing (which features should we use to quantify a protein?) and downstream statistical analysis (how to detect changes in protein abundance? Are our conclusions consistent with prior results?). Answering these questions requires the availability of statistical methods, and but also of publicly available data that help understand the advantages and the limitations of the methods. We developed a family of R/Bioconductor packages, called MSstats, for protein significance analysis in mass spectrometry-based proteomic experiments. MSstats supports DDA, SRM, and DIA acquisition, label-free experiments or experiments with tandem mass tag (TMT) labeling. MSstats supports assay characterization and longitudinal profiling and statistical quality control. MSstats includes new functionalities for selecting informative features, and works seamlessly with numerous data processing tools. We will highlight the new aspects and extensions of these packages and the contributions of our lab to these components of reproducible research.

Tips and Tricks (if submitted):

Session: Poster, poster number: 033

2018-2019 Metrics from the HUPO Human Proteome Project:  Progress on Identifying and Characterizing the Human Proteome

Gilbert Omenn1; Lydie Lane2; Christopher Overall3; Fernando Corrales4; Jochen Schwenk5; Young-Ki Paik6; Jennifer Van Eyk7; Liu Siqi8; Michael Snyder9; Mark Baker10; Eric Deutsch11
1University of Michigan, Ann Arbor, MI; 2CALIPHO Group, SIB Swiss Institute of Bioinformati, Geneva, Switzerland; 3University of British Columbia, Vancouver, BC; 4Centro Nacional de Biotecnologia, Madrid, Spain; 5Science for Life Laboratory, Solna, N/A; 6Yonsei Proteome Research Center, Seodaemoon-ku, Seoul, Korea; 7Cedar Sinai Medical Center, Los Angeles, CA; 8BGI Human Genome Center · Department of Bioinforma, Beijing, China; 9Stanford University, Stanford, CA; 10Macquarie University, Sydney, Australia; 11Institute for Systems Biology, Seattle, WA


     The Human Proteome Project (HPP) annually reports on progress throughout the field in credibly identifying and characterizing the human protein parts list and making proteomics an integral part of multiomics studies in medicine and the life sciences.  NeXtProt release 2018−01−17 contains 17 470 PE1 proteins, 89% of all neXtProt predicted PE1−4 proteins, up from 17 008 in release 2017−01.  Conversely, the number of neXtProt PE2,3,4 missing proteins (MPs) was reduced from 2579 to 2186.  Of the PE1 proteins, 16 092 are based on mass spectrometry results, and 1378 on other protein studies, notably protein−protein interaction.  PeptideAtlas 2018-01 has 15 798 canonical proteins, up 625 over the previous year, including 269 from SUMOylation studies.  The largest reason for MPs is low abundance.  The 6th annual HPP special issue of JPR presents evidence for 104 MPs that meet the HPP guidelines plus 107 identified in MassIVE (JPR 2018; 17:4023).  The 2019 releases from PeptideAtlas and neXtProt will be available for updates in time for the US HUPO meeting.

     Meanwhile, the Human Protein Atlas has released its Cell Atlas, Pathology Atlas, updated Tissue Atlas, and recommendations on antibody validation.  Finally, multiplex organ-specific popular proteins targeted proteomics has been applied to various diseases.

Tips and Tricks (if submitted):

Session: Poster, poster number: 034

Integrated Proteogenomic Data Analysis Pipeline and Its Applications to Post-translational Modification Investigation

Yingwei Hu; Minghui Ao; Jianbo Pan; David J. Clark; Weiming Yang; Punit Shah; Michael Schnaubelt; Lijun Chen; Jiang Qian; Zhen Zhang; Daniel W. Chan; Hui Zhang
Johns Hopkins University, Baltimore, Maryland

Proteogenomic analysis provides comprehensive and integrated characterization of specimens using multiple genomic and proteomic data generation pipelines, which is critical to understand the regulations in the alterations of genes, proteins, and post-translational modifications (PTMs). There are needs for a data analysis pipeline to integrate genomic and proteomic data as well as PTM information. Here, we describe an proteogenomic data analysis pipeline with the purpose of investigation of PTMs (e.g. Phosphorylation and Glycosylation). 

With the standardized analysis templates implemented in the pipeline, it provides a series of routine data analysis outputs for further investigation by integration of existing tools and in-house scripts, which include quality control assessment, differential expression analysis, pathway analysis, phenotype associations analysis, PTMs identification and investigation, and etc. The pipeline has been applied in the latest studies of Clinical Proteomic Tumor Analysis Consortium (CPTAC) and the HIV latency study from an amfAR project. In practice, the pipeline not only significantly shortened the period from data quality assessment to preliminary conclusion generation, but also showed the potential of the utility of using integrated data analysis on multi-omics data. For example, the integrated analysis of the expression of glycosylated peptides and glycosylation enzymes revealed the differential glycosylation mechanisms in ovarian tumor and non-tumor tissues in the ovarian cancer study.

To conclude, our results showed that the pipeline provided standard quality control assessment as well as efficient integrated data analysis methods to accurate knowledge discovery from multi-omics studies.

Tips and Tricks (if submitted):

Session: Poster, poster number: 035

Integrated Machine Learning Pipeline Reveals Fingerprints of the Oxidative Stress-Sensitive Post-translational Modification during Cardiac Remodeling

Howard Choi1; Bilal Mirza1; Jie Wang1; Jessica M Lee1; Dominic CM Ng1; Neo Christopher Chung1, 2; Ding Wang1; David A. Liem1; J. Harry Caufield1; Henning Hermjakob1, 3; Wei Wang1; Yibin Wang4; John R. Yates III1, 5; Peipei Ping1
1NIH BD2K Center of Excellence at UCLA, Los Angele, CA; 2Institute of Informatics, University of Warsaw, Warsaw, Poland; 3European Bioinformatics Institute (EMBL-EBI), Hinxton, United Kingdom; 4Department of Anesthesiology, UCLA, Los Angeles, CA; 5Department of Chemical Physiology, TSRI, La Jolla, CA

Oxidative stress is a common pathological stimulus contributing to cardiac disease, yet our understanding of oxidative stress-induced protein targets and their dynamics remain elusive. This is largely due to the complex nature of oxidative stress-induced post-translational modifications (O-PTMs) of proteins. Accordingly, machine learning (ML) approaches can unveil complex patterns in large biological datasets. However, the successful application of ML to elucidate O-PTM signatures in cardiac pathology present challenges in sparsity, temporal nature, and high-dimensionality of data that require further innovations.

Longitudinal datasets over a 14-day duration were generated from 6 genetically distinct murine strains with variable disease susceptibility. Mice were infused with isoproterenol, a model of cardiac disease and elevated oxidative stress. Functional phenotyping ensued via quantifying 7 chemically-distinct O-PTMs and developing a novel ML-based platform to link O-PTM fingerprints with cardiac pathological progression. Molecular signatures of cardiac phenotypes and O-PTM dynamic patterns were identified using feature selection algorithms following random forest-based missing data imputation and cubic spline-based clustering, respectively. Unique molecular signatures were validated using three independent mechanisms: computational, biological, and translational.

Cardiac remodeling potentiated O-PTMs in 6,478 of 8,227 murine cardiac proteins, including 25,148 modified sites in 549 cellular pathways. ML analysis produced 8 O-PTM sites representing a cardiac pathological phenotype; known hypertrophic pathways such as muscle contraction, TCA, beta-oxidation, branched chain amino acid catabolism (BCAA), and Ca2+ regulation showed enrichment. Importantly, O-PTM combinations in these pathways exhibited distinct temporal profiles, discovering novel molecular signatures that define different phases of cardiac disease progression.

Our findings provide the first O-PTM molecular map that serves as an avenue for molecular signature discovery and design of therapeutics. Our novel ML-based integrative analytics approaches discerned the relationships of O-PTM alterations and functional remodeling. This pipeline has been demonstrated to be effective for conducting deep molecular phenotyping in disease progression of cardiovascular medicine.

Tips and Tricks (if submitted):

Session: Poster, poster number: 036-TL

Improved Protein Inference from Multiple Protease Bottom-Up Mass Spectrometry Data with MetaMorpheus

Rachel Miller1; Rob Millikin1; Connor Hoffman1; Stefan Solntsev1; Gloria Sheynkman2; Michael Shortreed1; Lloyd Smith1
1University Wisconsin-Madison, Madison, WI; 2Dana-Farber Cancer Institute, Boston, MA

Protein inference is a process in bottom-up proteomics to infer the presence of a set of proteins from the identified peptides. Peptides with only one protein of origin ("unique peptides") provide enough evidence for confident identification of that protein. Identified peptides with multiple possible proteins of origin yield ambiguous protein identifications (“protein groups”). One strategy employed to increase the number of identified unique peptides is to use multiple proteases. Each protease produces a different set of peptides from the same set of proteins, which increases the chances of identifying a unique peptide and subsequently identifying each parent protein. Protein inference, when performed separately for each protease’s identified peptides, necessitates reconciling the protein inference results, which is time consuming and can lead to errors. We recently updated the search software MetaMorpheus with the ability to perform coincident protein parsimony using bottom-up data from multiple different proteases. Combined protein inference disambiguates many protein groups into single protein identifications and reduces the number of erroneous protein identifications. In addition, PTM identification and localization are markedly improved.

Tips and Tricks (if submitted):


Session: Poster, poster number: 037-ML

EpiProfile 2.0: A Computational Platform for Processing Epi-Proteomics Mass Spectrometry Data

Zuofei Yuan; Simone Sidoli; Dylan M. Marchione; Johayra Simithy; Kevin A. Janssen; Mary R. Szurgot; Benjamin A. Garcia
University of Pennsylvania, Philadelphia, PA

Histone post-translational modifications (PTMs) contribute to chromatin function through their chemical properties which influence chromatin structure, and their ability to recruit chromatin interacting proteins. Nanoflow liquid chromatography coupled with high resolution tandem mass spectrometry (nanoLC-MS/MS) has emerged as the most suitable technology for global histone modification analysis due to the high sensitivity and the high mass accuracy that provide confident identification. However, the histone nanoLC-MS/MS data analysis is even more challenging due to the large number and variety of isobaric histone peptides, and the high dynamic range of histone peptide abundances. Therefore, quantification of histone PTMs has become a high priority to investigate cell regulation and epigenetics.


Here, we present EpiProfile 2.0, an extended version of our 2015 software (v1.0) for accurate quantification of histone peptides based on liquid chromatography – tandem mass spectrometry analysis. EpiProfile 2.0 is now optimized for data-independent acquisition through the use of precursor and fragment extracted ion chromatography to accurately determine the chromatographic profile and to discriminate isobaric forms of peptides. The software uses fragment ions and an intelligent retention time prediction trained on the analyzed samples to enable accurate peak detection. EpiProfile 2.0 supports label-free and isotopic labeling, different organisms, known sequence mutations in diseases, different derivatization strategies, and unusual PTMs (such as acyl-derived modifications). In summary, EpiProfile 2.0 is a universal and accurate platform for the quantification of histone marks via LC-MS/MS. Being the first software of its kind we anticipate that EpiProfile 2.0 will play a fundamental role in epigenetic studies relevant to biology and translational medicine. EpiProfile is freely available at

Tips and Tricks (if submitted):

Session: Poster, poster number: 038

UniRef clusters as a resource for protein annotation propagation and prediction

Yuqi Wang1, 3; Hongzhan Huang1, 3; Peter McGarvey2, 3; Cecilia Arighi1, 3; Cathy Wu1, 2; UniProt Consortium3
1CBCB, University of Delaware, Newark, DE; 2BMCB, Georgetown University Medical Center, Washington, DC; 3EBI-EMBL, UK; SIB, CH and PIR, USA, Washington, DC

UniRef databases provide full-scale clustering of UniProt protein sequences to hide the redundant sequences and obtain complete coverage of the sequence space at three resolutions: 100, 90 and 50 % identity, respectively. They have been utilized in a broad range of applications. UniRef100 has become a standard sequence database for Mass Spectrometry search while UniRef90/50 are used in similarity-based functional annotation. We have assessed the intra-cluster molecular function consistency with Gene Ontology (GO), and further found that such results were highly coherent in all three GO domains.

Recently, we have implemented GO annotations to UniRef90/50 clusters when all members with GO annotation contained common GO terms, or had a meaningful common ancestor. Such GO terms should be safely propagated to other UniProtKB members in the cluster lacking any GO annotation. In other words, we could predict the GO terms using the UniRef clusters.

To evaluate this, we took all the UniRef90/50 clusters that provided any GO annotation in a previous release and examine if an un-annotated member from such cluster has received GO annotations from UniProtKB in a later release. A new GO term would be counted as false positive if it does not match previous prediction, but has a non-root common ancestor with it. We compared the GO annotations between Releases through 2017 and 2018 for UniRef90/50 and the results indicate that more than 97% of the new GO annotations match the previous predictions.

In conclusion, the members of a UniRef90/50 cluster are highly coherent and provide a good potential method for annotation prediction/propagation. In fact UniRef clusters have been widely used for the purpose, such as GO enrichment, gene prediction, metagenomics functional profiling and transcriptome assembly. Here, we provided a formal evaluation of the UniRef-based annotation method.

Web access at and 

Tips and Tricks (if submitted):

Session: Poster, poster number: 039

Proteomic Data Commons (PDC): A Node in the NCI Cancer Research Data Commons

Ratna Thangudu1; Michael Holck1; Deepak Singhal1; Paul Rudnick2; Nathan Edwards3; Karen Ketchum1; Christopher Kinsinger4; Izumi Hinkson5; Anand Basu1; Michael Maccoss6
1ESAC, Inc., Rockville, Maryland; 2Spectragen Informatics LLC, Bainbridge Island, WA; 3Georgetown University, Washington, DC; 4Natl Cancer Institute, NIH, Bethesda, MD; 5NCI CBIIT, Rockville, MD; 6University of Washington, Seattle, WA

The creation of a National Cancer Data Ecosystem to establish data science infrastructure necessary to connect repositories, analytical tools, and knowledge bases was a key recommendation of the 2016 Beau Biden Cancer MoonshotSM initiative.  These resources serve to support data aggregation, query, analysis, and visualization in unique and powerful ways within and across data types.

To this end, the National Cancer Institute (NCI) has begun the development of the NCI Cancer Research Data Commons (CRDC) that included a Genomic Data Commons, Proteomic Data Commons and Cloud Resources. In the near future, the CRDC will launch an Imaging Data Commons and a Cancer Data Aggregator.

As a node in the CRDC, the overarching goal of the NCI Proteomic Data Commons (PDC) project is to democratize access to cancer-related proteomic datasets as well as to provide sustainable computational support to the cancer research community.  The PDC seeks to enable and empower the cancer-research community, including scientists working in both intramural and extramural laboratories, with the necessary informatics capabilities to carry out large-scale, multi-omic data analysis.

We recently launched a minimal viable product (MVP) for the PDC with the goal of getting a prototype product quickly to the user community in order to collect feedback to drive the future development of features and capabilities of the final product. The prototype consists of a data portal and workspace components, deployed in Amazon Web Services. In this presentation we will describe the current status of the PDC, including: available datasets, application programming interface, analysis pipelines, data model, underlying architecture, integration with the CRDC, and the overall philosophy of the PDC. The prototype can be accessed from

Tips and Tricks (if submitted):

Session: Poster, poster number: 040-TL

Optimized cross-linking mass spectrometry for in situ interaction proteomics

Zheng Ser1, 2; Paolo Cifani1; Alex Kentsis1, 2
1Sloan Kettering Institute, New York, NY; 2Tri-Institutional PhD Program in Chemical Biology, New York, NY

Recent advancements in mass spectrometer cleavable protein cross-linkers and algorithms for cross-linked peptide identification now enable proteome-scale cross-linking mass spectrometry (XL-MS). Here, we optimized the use of mass spectrometer cleavable cross-linker, disuccinimidyl sulfoxide (DSSO), for labeling protein complexes in live human cells. To control the sensitivity and specificity of identified cross-linked peptides, we applied a generalized linear mixture model to calibrate cross-link peptide-spectra matching (CSM) scores. Using specific CSM score thresholds to control the false discovery rate, we found that higher-energy collisional dissociation (HCD) and electron transfer dissociation (ETD) can both be effective for proteome-scale XL-MS protein interaction mapping. We demonstrate that the density and coverage of protein-protein interaction maps can be significantly improved through the use of multiple proteases. In addition, the use of sample-specific search databases can be used to improve the specificity of cross-linked peptide spectral matching. Application of this approach to cross-linked human chromatin in live cells recapitulated known and revealed new protein interactions of nucleosomes and other chromatin-associated complexes in situ. In particular, we found unanticipated interactions of HMGN proteins with the H1 linker histone, providing the sought-after mechanism to explain its structural effects on chromatin directly. Live cell cross-linking was performed on human cancer cells to define the architecture of the mutated chromatin remodeling BAF complex. These studies identified aberrant protein-protein interactions retained by the mutant BAF complex in rhabdoid tumor cells, thereby identifying neomorphic BAF complexes induced by oncogenic mutations.  The optimized approach for mapping native protein interactions can be applied to study a wide range of unanswered biological problems.

Tips and Tricks (if submitted):

Session: Poster, poster number: 041

Accelerating DIA Studies to Extend Workflow Utility, Using Fast Microflow LC Gradients

Christie Hunter1; Zuzana Demianova3; Nick Morrice2
1SCIEX, Redwood City, CA; 2SCIEX, Warrington, UK; 3SCIEX, Darmstaedt, Germany

Microflow LC has been used increasing in quantitative proteomics in combination with SWATH® Acquisition, to provide better robustness and higher throughput when measuring larger sample cohorts. With higher flow rates, sample loading, trap/column washing, equilibration and gradient formation are all accelerated, allowing much faster run times to be achieved. Here, the impact of shortening gradient length on protein quantitation results with DIA was explored. Using microflow LC on the TripleTOF® 6600 System, a range of gradient lengths from 5 – 45 mins was explored, and key acquisition parameters for SWATH® Acquisition were varied to optimize for the much faster run times. 

Using complex digested cell lysates, SWATH protein quantitation results were assessed as a function of gradient length. Fast MS/MS acquisition rates were found to be critical because this enabled more smaller variable Q1 windows to improve S/N for quantitation.  Even with the fastest gradients, methods with 60-100 windows with very fast accumulation times of 15 msec improved results. As expected, total # of protein quantified decreased when shortening the gradient from 45 to 5mins. However, with the 10min gradient, the peptide and protein quantified was ~60 and 70% respectively of results from the 45min gradient for the three matrices tested (on 2 instruments). With the Pan Human Library on a 1ug load of HEK lysate, ~2100 and ~3400 proteins were quantified with the 5 and 10 min gradients respectively. The full optimization results will be presented.

Tips and Tricks (if submitted):

Session: Poster, poster number: 042

Simplifying the Use of Ion Libraries During Data Processing of Data Independent Acquisition Proteomics Data

Christie Hunter1; Nick Morrice2
1SCIEX, Redwood City, CA; 2SCIEX, Warrington, UK

As the use of data independent acquisition (DIA) grows in proteomics research, the need for improved data processing workflows increases. The most common data processing workflow is to use spectral ion libraries to drive targeted extraction of peptide / fragment areas from the data, using the m/z and retention time information contained in the library. Increasing the size and quality of the ion library has been shown to increase the proteins reliably quantified from a dataset. Retention time (RT) correlation between ion library and the dataset is another key factor that determines quality of data extraction. Two algorithms were explored to simplify the data extraction workflow for SWATH® Acquisition data within the OneOmics™ Project cloud processing pipeline, the automatic merging of related ion libraries followed by auto RT calibration.

The quality of library merging on SWATH results was explored with a series of small datasets. Non-linear retention time alignment was shown to be effective for merging of libraries with identical or differing gradient lengths.  By merging libraries from replicates of a 1D dataset, a 20% gain in peptides was observed. The gain in peptides identified and quantified by SWATH is primarily attributed to an increase in peptide coverage rather than improved ion selection. Full pipeline was then used on some biological datasets (PBMC and Mouse cell lines) to measure performance, where a series of libraries were available to merge. Results will be discussed.

Tips and Tricks (if submitted):

Session: Poster, poster number: 043-WTT

Rapid Qualitative and Absolute Quantification of Plasma based proteins using a Novel Scanning Quadrupole DIA Acquisition Method

Roy Martin1; Lee Gethings2
1Waters, Beverly, MA; 2Waters, Wilmslow, UK

Quantitative proteomics often incorporates the use of stable isotope labels (SILs) to provide absolute quantification. Recent advancements have seen the introduction of peptide panels allowing the quantification of over 500 proteins in plasma sample sets. However, this is technically challenging when attempting to acquire the data using more traditional MS acquisition modes such as multiple reaction monitoring (MRM), since the duty cycle of the instrument is compromised and therefore results in under sampling. Data independent analysis (DIA) methodology provides high throughput while also ensuring high rates of data acquisition and specificity. Here we describe the use of a kit consisting of SIL’s capable of quantifying >500 plasma-based proteins in conjunction with a novel scanning quadrupole DIA acquisition schema in order to quantify proteins of interest for patient cohorts diagnosed with respiratory disorders.

Undepleted human plasma originating from controls and patients diagnosed with chronic obstructive pulmonary disorder (COPD) and asthma were reduced, alkylated and tryptically digested overnight. Prior to LC-MS analysis, samples for their individual groups were pooled and spiked with PQ500 SIL peptides (Biognosys).  In all cases, samples were separated using various LC gradients (15 and 45 minutes) whilst employing 1mm scale chromatography for high throughput analyses. MS data were acquired using SONAR, whereby the quadrupole (MS1) was continuously scanned between m/z 400-900 using a quadrupole transmission width of approximately 20 Da, whilst the TOF scanned between m/z 50-2000. In all cases, precursor and product ion information were collected. The LC-MS data were processed with Spectronaut Pulsar X (Biognosys AG and Progenesis QI for Proteomics (Non-linear Dynamics). Multivariate statistical analysis showed distinct differences between all three cohorts and proteins corresponding to 86% of the spectral library were quantified. CV’s for each group were all found to be <5% in all cases and was maintained across all gradient lengths.

Tips and Tricks (if submitted):

Scaling up high throughput Quantitive proteomics to 1mm scale provides some benefits in terms of robustness and throughput, but requires different sample loads and faster data cycle times. The presentation will focus on what needs to be considered and some guidelines for good results.

Session: Poster, poster number: 044

Introducing the “DataCrusher SuperComputing Service”: A DIA pipeline that’s ~10x Easier, ~1000x Faster, and ~10x More Robust

Gautam Saxena; Japheth Odonya; Boutaskiouine Mustapha
DeepDIA, Bethesda, MD

Data Independent Acquisition (DIA) has gained tremendous popularity in the last five years, but its tremendous power is rarely fully harnessed due to three main factors:

1) the complexity of the experimental setup (eg the tremendous informatics and MS time needed to create DDA-based spectral libraries)

2) the computing resource requirements, which results in days/weeks of computation time for even small projects

3) the lack of a reasonably simple, single informatics workflow that can be executed with minimal training

Herein, we present for the first time the “DataCrusher SuperComputing Service,” a beginning-to-end informatics pipeline which runs on a supercomputer, can process both MS files created from a traditional DIA protocol as well as those created using a novel DIA protocol called “Global Staggered DIA” (gsDIA), and which can be accessed by the scientific MS community using a modern web browser.

The DataCrusher SuperComputing Service

1) Does not require a spectral library to be generated

2) Runs in approximately 15 minutes, irrespective of the number of samples submitted

3) Does not require retention time standards to be injected

4) Demonstrates substantively superior number of low-abundant analytes quantified reproducibly (CV < 5%) than traditional DIA protocols

5) Is dead-easy to use (most of the relevant information can be derived from the MS data itself without querying user)

6) Includes an automated, near-real-time, and extraordinarily fast MS data importer (~30s/file)

7) Is accessed through a user’s browser, and so does not involve any installation by the average user

8) Exists simultaneously in multiple parts of the world (US, Europe, Asia) and therefore provides low latency, high fault tolerance, and adheres to various legal requirements for data storage

9) Includes an interactive graphical system for all processed data, from raw MS data to MS2 features to pseudo mgfs to lists of peptides and proteins.

Tips and Tricks (if submitted):

Session: Poster, poster number: 045-ML

The nature of phosphatidylinositol mannosidases (PIMs) interaction with the PPE68 protein – revealing novel insights in its immunogenicity and virulence

Nagender Rameshwaram; Sangita Mukhopadhyay
CDFD, India, Hyderabad, India

Background: Mycobacterial protein glycosylation is understudied and for only very few glycoproteins glycosylation sites have been described. Mycobacteria have been suggested to possess O-linked glycosylation pathways that display many commonalities with their eukaryotic and archaeal counterparts as well as some unexpected variations. Studies have demonstrated that PPE68 is a glycoprotein and plays an important role in M. tuberculosis pathogenicity and represent as potential vaccine candidate and diagnostic tool. Our preliminary results and shreds of evidence have shown that M. tuberculosis PPE68 protein stained positive for glycosylation in vitro (PAS staining) and glycosylated PPE68 is highly immunogenic in contrast to the deglycosylated PPE68. However, it is still unclear why does PPE68 wears an attire of such a modification? If PPE68 has the potential of being a therapeutic target it seems interesting to understand the protein glycosylation pattern to identify attributes connected to its immunogenicity using advanced strategies of glycomics and glycoproteomics.

Methods: Dot-blotting/electroblotting, Release of N-linked glycans, Release/reduction of O-linked glycans, Desalting of reduced N- and O-linked glycans, PGC-LC-ESI MS/MS of glycans

Results: Using glycomics and glycoproteomics, we demonstrated that the glycosylation of the PPE68 protein is due to its attachment to PIMs. The preliminary finding suggests that this is a covalent bond, rather than just non-covalent interaction. The bond is labile against acidic and alkaline hydrolysis and the glycosylation can partly be removed from the protein core by nucleophiles. Results also suggested the participation of aspartic and glutamic acid carboxylic side chain in the binding of the PIMs to PPE68.

Conclusions: Our study hints to a novel role of PPE68 as a PIM anchored protein that implicates its role in immune signaling in M. tuberculosis survival inside the host. This would also improve our understanding of the intricacies of mycobacterial protein glycosylation systems.

Keywords: PGC-LC-ESI MS/MS of glycans, glycoproteomics, glycomics

Tips and Tricks (if submitted):

Session: Poster, poster number: 046-TL

Integrated Glycoproteomic Characterization of Human High-Grade Serous Ovarian Cancer

Yingwei Hu; Jianbo Pan; Punit Shah; Minghui Ao; Stefani Thomas; Yang Liu; Lijun Chen; Michael Schnaubelt; David Clark; Qing Li; Jiang Qian; Zhen Zhang; Daniel Chan; Hui Zhang
Johns Hopkins University, Baltimore,

Many gene products exhibit great structural heterogeneity due to an array of modifications. These modifications are not directly encoded in the genomic template but often affect the functionality of proteins. Protein glycosylation plays a vital role in proper protein functions. However, the analysis of glycoproteins has been limited compared to other protein modifications, such as phosphorylation. Here, we performed an integrated proteomic and glycoproteomic analysis of 83 prospectively collected high-grade serous ovarian carcinoma (HGSC) and 23 non-tumor tissues. Integration of the expression data from global proteomics and glycoproteomics revealed tumor-specific glycosylation, uncovered different glycosylation associated with three subtypes of HGSC, and identified glycosylation enzymes that were correlated with glycosylation. In addition to providing a valuable resource, these results provide new insights into the potential roles of glycosylation in the pathogenesis of HGSC with the possibility of distinguishing pathological outcomes of ovarian tumors from non-tumors, as well as tumor subtypes.

Tips and Tricks (if submitted):

Session: Poster, poster number: 048

Organelle glycoproteomics by using lectin chromatography and HCD/ETD mass spectrometry

Junfeng Ma1; Jason Maynard2; Alma Burlingame2; Gerald Hart3
1Georgetown University, Washington, DC; 2UCSF, Sausalito, CA; 3University of Georgia, Athens, GA

Glycosylation, a large and structurally diverse category of post-translational modifications, regulates a wide range of protein functions and cellular processes. However, it has been a major challenge to study glycosylation due to technical difficulties for site mapping and structural characterization. To enrich sub-stoichiometric glycosylation, a number of methods have been developed recently. Meanwhile multiple mass spectrometric approaches have been adopted to facilitate site assignment and structural elucidation. In this study, we compiled a map for glycoproteins (including N-glyco-, O-GalNAc, and O-GlcNAc proteins) in different organelles by analyzing glycan-containing peptides. Specifically, we prepared organelle-enriched samples from human cancer cells. Glycopeptides were enriched by multiple rounds of wheat germ agglutinin (WGA) affinity chromatography. Enriched samples were then dried and fractioned by high pH RPLC. To improve HCD/ETD detection, each fraction was injected twice, once allowing for fragmentation of precursor ions 3+ and above, and one injection for only 2+ precursor ions. Indeed, we found a wide distribution of glycoproteins identified: O-GlcNAcylation occurs mainly on proteins in nuclear and plasma membrane while N-glycosylation and O-GalNAcylation occur on proteins localized in ER, Golgi, lysosome, and plasma membrane. Specifically, we identified 106 unique O-GlcNAc peptides with unambiguous site localization, 293 O-GlcNAc peptides with ambiguous site localization, 918 unique peptides with other glycosylation forms (unambiguous sites assigned), and 287 unique peptides with ambiguous localization of N or O-GalNAc modifications. More importantly, besides the identity of glycosylation sites, our results reveal the nature of the oligosaccharide structures on these proteins. We also report an iterative approach that allows us to explore the diverse nature and distribution of oligosaccharides present in glycopeptide mixtures.  

Tips and Tricks (if submitted):

Session: Poster, poster number: 049-ML

Glycoproteomics-based Signatures for Tumor Subtyping and Clinical Outcome in Human High-Grade Serous Ovarian Cancer

Jianbo Pan; Yingwei Hu; Shisheng Sun; Lijun Chen; Jian-Ying Zhou; Michael Schnaubelt; Minghui Ao; Jiang Qian; Zhen Zhang; Daniel Chan; Hui Zhang
Johns Hopkins School of Medicine, Baltimore, MD

In different individuals, tumor heterogeneity could exhibit significant differences at the molecular level due to an array of co-occurring genomic, transcriptional, translational, and post-translational regulations. As one of the most abundant and complex protein-translational modifications, protein glycosylation is known to be associated with tumor progression, metastasis and survival. To investigate the role of protein glycosylation in tumor heterogeneity of high-grade serous ovarian carcinoma (HGSC), we performed mass-spectrometry (MS)-based glycoproteomic quantification on 119 TCGA HGSC tissues using 4-plex iTRAQ reagents. Bi-cluster analysis of intact glycoproteomics identified 3 major tumor subtypes and 5 groups of intact glycopeptides, and showed a strong relationship between N-glycan structures and molecular subtypes, for example, fucosylation and mesenchymal subtype. Further survival analysis displayed that intact glycopeptide signature of mesenchymal subtype indicated a poor clinical outcome in HGSC. In addition, glycoprotein synthesis involving a series of reactions and enzymes is regulated by several factors that are related to gene expression, protein expression, and protein functions. Through integration of genomics, transcriptomics and proteomics in previous studies with glycoproteomics, we studied the expression correlation between different types of glycoprotein product (i.e. mRNA, protein, deglycosite peptide and intact glycopeptide). Although glycoprotein expression was major regulated by substrates in this study, we investigated and found some glycosylation enzymes could relate to the production of different types of glycan peptide, and further coordinate the tumor heterogeneity. Deep understanding of the glycosylation process and production in different types of HGSC may provide clues for precision medicine and targeted therapy on glycosylation genes.

Tips and Tricks (if submitted):

Session: Poster, poster number: 050

Development of hybrid glycoproteomic workflows for site-specific characterization of intact glycopeptides

Matthew Glover; Kristen Lekstrom; Raghothama Chaerkady; Sonja Hess
MedImmune, Gaithersburg, MD

Protein glycosylation is an abundant post-translational modification that regulates a variety of important molecular and cellular processes. The analysis of intact glycopeptides remains an enormous analytical challenge. Because of this, the majority of glycoproteomic studies have relied on enzymatic removal of the glycan side chain. Although this is an effective strategy for identifying glycosylation sites, valuable information about glycan microheterogeneity and the biological function of specific glycans is lost. Here, we describe the development of a hybrid analytical workflow that combines hydrophilic interaction liquid chromatography (HILIC) and strong anion exchange cartridges operated in electrostatic repulsion-hydrophilic interaction chromatography mode (SAX-ERLIC) for glycopeptide enrichment with electron-transfer/higher-energy collision dissociation (EThcD) fragmentation for site-specific characterization of intact glycopeptides.

Preliminary analysis of CHO digest (1 mg starting material) enriched with HILIC and analyzed in a 90-minute LC-MS/MS method with EThcD fragmentation yielded > 1000 N-linked glycopeptide identifications. In agreement with previous results, preliminary results suggest that SAX-ERLIC is superior for enrichment of O-linked glycopeptides compared to HILIC. Therefore, ongoing experiments are being performed to analyze the utility of a serial enrichment strategy in which HILIC enrichment is performed prior to SAX-ERLIC. The aim of this method is to use HILIC to remove the majority of N-linked glycopeptides prior to SAX-ERLIC to increase the O-glycoproteome coverage. This serial enrichment strategy combined with EThcD-enabled glycopeptide characterization has the potential to provide valuable site-specific information about the role of N- and O-linked glycans in a variety of biological systems.

Tips and Tricks (if submitted):

Session: Poster, poster number: 051-TL

Developing and Characterizing FLIM Probes to Detect Subcellular Tyrosine Kinase Activity

Sampreeti Jena; Oscar Bastidas; Erica Pratt; Scout Allendorf; Blanche Cizubu; Laurie L. Parker
University of Minnesota Twin Cities, Minneapolis, MN

Tyrosine kinases are regulators of the signal transduction process and play a key role in cellular processes such as proliferation, migration, metabolism, differentiation and survival. The goal of this project is to develop biosensors (probes) to measure and spatially map kinase activity in live, intact cells with subcellular resolution. The fluorophore labeled peptide probes comprise a motif that facilitates cell penetration and a motif recognized by a specific kinase family. Fluorescence lifetime imaging microscopy (FLIM) measures the decay rate of fluorescence excitation at a picosecond scale resolution and enables single cell mapping of fluorophore lifetime. In this study, fluorophore labeled probes were incubated with cultured MDA-MB-231 cells. Fluorescence lifetime distributions of the probes were mapped within cells, before and after kinase stimulation to detect changes induced due to phosphorylation. The effect of the respective inhibitor drugs on kinase activity were also gauged from shifts in fluorophore lifetime. Identical probes with a phosphorylated tyrosine and a Y to F mutant served as the positive and negative controls for these experiments. FLIM is capable of superior multiplexed analysis and intensity based read out techniques since it enables fluorophores to be distinguished by their lifetime behaviors in addition to their excitation and emission properties.

The shifts in fluorescence lifetime upon phosphorylation arise due to Src homology 2 (SH2) domains within the kinase serving as ‘reader’ proteins that recognize and interact with phosphorylated probes (protein-protein interaction). In vitro, high-throughput-compatible fluorescence polarization assays and Isothermal Titration Calorimetry (ITC) were employed to characterize the binding affinity between recombinantly expressed SH2 domains and their corresponding peptide probes (phospho-, unphospho-  and Y to F mutants). This work should lead to a suite of novel, complementary probes for quantification and localization of PTM enzymatic function in cells as well as serve a high throughput pipeline for drug screening.

Tips and Tricks (if submitted):

Session: Poster, poster number: 052

Enrichment and Identification of the class I MHC phosphopeptides from the resected head and neck cancer tumors

Mohammad Ovais Aziz-Zanjani1; Feng Shi3; Sean Sepulveda3; Jeffrey Shabanowitz1; Dina L. Bai1; Donald F. Hunt1, 2; Mark Cobbold3
1Department of Chemistry, University of Virginia, Charlottesville, VA; 2Department of Pathology, University of Virginia, Charlottesville, VA; 3MGH Cancer Center, Harvard Medical School, Charlestown, MA

Patients with recurrent or metastatic head and neck cancers (HNCs) have a median overall survival of less than 1 year.1 The recurrence rate is ~50% in patients with locally advanced HNCs within 3 years.1 The effectiveness of cancer immunotherapies for the treatment of HNCs patients1 are based on the fact that they have T lymphocyte capable of recognizing tumor antigens.2 Tumor antigens are presented to the immune system through the major histocompatibility complex (MHC)-I pathway.2 Previous studies revealed that deregulated protein phosphorylation in cancer results in the presentation of class I MHC phosphopeptides on the cell surface3 that are potential targets for T cell recognition.3,4 In this study, class I MHC phosphopeptides were identified from five resected HNC tumors. After the cell lysis and the immunopurification of class I MHC molecules, MHC-bound peptides were eluted by the acid. Then, samples were desalted via the STop and Go Extraction (STAGE) tip technique.5 Finally, the immobilized metal affinity chromatography (IMAC) protocol was used for the enrichment of the class I MHC phosphopeptides before the analysis with high-performance liquid chromatography coupled to electrospray ionization–tandem mass spectrometry (HPLC-ESI-MS/MS).6 Complementary data obtained from the collisional activated dissociation (CAD) and electron transfer dissociation (ETD) fragmentation techniques were used for the manual sequencing and determination of the phosphosite. This technology enabled the identification and sequencing class I MHC phosphopeptides at the attomole level from resected tumor samples as low as 0.1g. About two-thirds of 122 identified class I MHC phosphopeptides from HNCs tumor samples were already identified in other cancer types like melanoma, colorectal, liver, and ovarian cancers.5 The overlapping of the identified class I MHC phosphopeptides among various cancer types and between different cancer patients shows their immunotherapeutic merits.

Tips and Tricks (if submitted):

Session: Poster, poster number: 053

Modulating immunopeptidomes through ex vivo manipulation

Marlene Heberling; Niclas Olsson; Joshua Elias
Stanford University, Stanford,

Tumor neoantigens hold immense potential for cancer immunotherapies but are limited by challenges with in vivo detection and validation. This is due to a dearth of neoantigens presented on cell surfaces via Human Leukocyte Antigen–Class I and –Class II, otherwise known as the immunopeptidome. I hypothesize that HLA–Class I and –Class II immunopeptidomes can be modulated to increase neoantigen presentation, expanding the therapeutic opportunities for adoptive T-cell therapy or personalized cancer vaccines. Initial experiments demonstrate that rapamycin treatment causes protein abundance changes that are apparent in the proteome. Normally less abundant HLA molecules such as HLA-C and HLA-DQ are consistently unpregulated. NetMHCpan-4.0 binding affinity estimates of eluted HLA–I peptides show that mean binding affinity is significantly lower for alleles HLA–B*15:10 and HLA–C*04:01. There is also a reproducible increase in high-affinity pHLA–I for all alleles following 48 hours of rapamycin treatment. For both HLA–C*03:04 and HLA–C*04:01, the increase in unique pHLA–I is 2-fold. Thus, rapamycin-induced changes in HLA expression consistently alter the immunopeptidomic landscape.These data indicate a potential role for rapamycin in increasing neoantigen presentation and expanding immunotherapy opportunities.

Tips and Tricks (if submitted):

Session: Poster, poster number: 054

Unbiased solution for MS-based immunopeptidomics with high sensitivity and accuracy

Katherine Tran1; Hieu Tran2; Baozhen Shan1
1Bioinformatics Solutions Inc, Waterloo, Canada; 2University of Waterloo, Waterloo, Canada

Identification of tumor-specific antigens (neoantigens) is needed for development of effective cancer immunotherapy and a good source for such antigens are the pools of HLA-bound peptides presented exclusively by the tumor cells. Mass spectrometry (MS) has evolved as the method of choice for the exploration of the human immunopeptidome (HLA class-I and class-II peptides). Workflows for immunopeptidomics are different from ones for more established shotgun proteomics which is biased to tryptic peptides, yet inherent differences between these two fields create significant drawbacks of current data analysis algorithms for the former. We provide de novo sequencing for peptide identification to address the barriers for data analysis, e.g., diverse C-termini of HLA-peptides, lack of sequence library for spliced peptides. With deep learning technology, we first time proposed peptide de novo sequencing algorithms for data independent acquisition (DIA) method. Further, de novo sequencing was integrated with motif-constrained database search for HLA-peptide identification to increase peptide sequencing coverage, depth, and confidence. Our approach was tested with several public data sets for the identification of HLA-peptides or neopeptides, including DDA (PXD007596, PXD006939, PXD004964, and PXD002431) and DIA (PXD001094). On average, 60% more HLA-peptides were identified by validation with the public immunopeptidome databases, IEDB and NeuroPep. Preliminary results showed that the de novo sequencing-based data analysis would provide a novel solution for immunopeptidomics with high sensitivity and accuracy.

Tips and Tricks (if submitted):

Session: Poster, poster number: 055-ML

A New Strategy for the Global Identification and Validation of Post-Translationally Spliced Peptides with Neo-Fusion

Zach Rolfs; Stefan Solntsev; Michael Shortreed; Brian Frey; Lloyd Smith
University of Wisconsin - Madison, Madison, WI

A controversy has emerged regarding the existence and prevalence of post-translationally spliced peptides in biology. Such peptides have been reported as both autoantigens in pancreatic islets and as HLA-I associated antigens in tumors, and thus have major potential implications for autoimmunity and cancer immunotherapy. The investigation of spliced peptides by mass spectrometry is currently limited by the nontrivial nature of spliced peptide identification and is typically accompanied by a high false-positive rate. We developed Neo-Fusion, a software program to assist in the identification and validation of post-translationally spliced peptides in complex samples. Neo-Fusion uses separate b- and y-ion databases to identify the N-terminal and C-terminal halves of each spliced peptide and determine the best matching sequence. We present a new and comprehensive strategy to increase the sensitivity and confidence of spliced peptide identification. This strategy uses Neo-Fusion and the auto de novo program PEAKS to first identify putative spliced peptide sequences.  These candidate sequences are then manually validated through comparisons of theoretical and experimental retention times and peptide fragmentation patterns. Through this work, we have identified common pitfalls in the field that can lead to misidentification of post-translationally spliced peptides and massively underestimated false discovery rates. Thorough reanalysis of published data from manuscripts reporting large scale presentation of post-translationally spliced peptides indicates that most of the reported spliced peptides were assigned incorrect sequences. Freely available tools exist for the rapid evaluation of reported spliced peptide identifications, and these tools should be employed whenever spliced peptide discovery is attempted. Using our strategy, we estimate that, at most, 1-4% of all HLA-I associated peptides can be explained through post-translational splicing.

Tips and Tricks (if submitted):

Session: Poster, poster number: 056-TL

SILAC-based quantitative proteogenomics unveil altered MHC-associated peptidome in osimertinib resistant human lung adenocarcinoma

Yue Qi; Tapan Maity; Meriam Bahta; Khoa Dang Nguyen; Constance Cultraro; Xu Zhang; Udanya Guha
Thoracic and GI Malignancies Branch, CCR, NCI, NIH, Bethesda, MD

The 3rd generation epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor (TKI), osimertinib is very effective against EGFR mutant lung adenocarcinoma. Unfortunately, all patients develop resistance. EGFR mutant patients respond poorly to immune checkpoint inhibitors. Adoptive T-cell therapy (ACT) holds great potential for treating metastatic cancers. Since cytotoxic T cells largely recognize the major histocompatibility complex (MHC) class I-associated peptides, tumor-specific neo-peptides are attractive targets for ACT. Here, we employed a proteogenomic approach using SILAC-based quantitative proteomics to quantify the class I-associated neo-peptides from parental PC9 (EGFRDel) lung adenocarcinoma cells, and isogenic osimertinib resistant PC9 (PC9-OsiR-NCI-1) following pan-MHC class I pull-down from cellular lysates and detection of class I-associated peptides by tandem mass spectrometry in a QE-HF mass spectrometer. Cell line-specific databases were built using whole exome sequencing (WES) data. PEAKS and Maxquant were used for identification and quantification of peptides. We identified 10711 class I-associated peptides which were 6-15 amino acid residues in length and 11 variant peptides, many of which were not predicted as neoantigens using epitope prediction algorithms. Interestingly, 90% of the peptides contained either a lysine, arginine or both; therefore, SILAC could be used for quantitation. Among the peptides identified in both cell lines, 17 and 440 had increased and decreased abundance, respectively in the resistant cells. 178 peptides were identified only in PC9 parental cells and 10 peptides only in resistant cells. The SILAC data of class I-associated peptidome demonstrate that there may be significant reduction of class I presentation by osimertinib resistant cells. In summary, we have established an integrated proteogenomic platform to provide direct evidence of neo-peptide presentation and to estimate the abundance of the MHC peptidome. Further experiments are underway to identify such variant neo-peptides in tumor tissue from osimertinib resistant patients enrolled in our clinical protocol at the NIH Clinical Center.

Tips and Tricks (if submitted):

The separation of MHC binding peptides from MHC pulled down complex is still challenge. We optimized the peptides separation from MHC molecules and W6/32 antibody with serial buffer elution from C18 desalting column. We found 40% acetonitrile buffer yield the maximal peptides but minimal macromolecule contamination.   

Session: Poster, poster number: 057-ML

Uncovering the Prognostic and Therapeutic Potential of N-Acetyl-Aspartyl-Glutamate Metabolism in Cancer

Sunag Udupa1; Tu Nguyen1; Brian Kirsch1, 2; Ryoichi Asaka1; Karim Nabi1; Addison Quinones1; Jessica Tan1; Marjorie Antonio1; Felipe Camelo1; Ting Li1; Stephanie Nguyen1; Giang Hoang1; Kiet Nguyen1; Christos Sazeides3; Yao-An Shen1; Amira Elgogary1; Juvenal Reyes1; Liang Zhao4; Andre Kleensang4; Kaisorn Chaichana1; Thomas Hartung4; Michael Betenbaugh2; Suely Marie5; Jin Jung1; Tian-Li Wang1; Edward Gabrielson1; Anne Le1
1Johns Hopkins Medicine, Baltimore, MD; 2Johns Hopkins Whiting School of Engineering, Baltimore, MD; 3University of Pennsylvania Perelman SOM, Philadelphia, PA; 4Johns Hopkins Bloomberg School of Public Health, Baltimore, MD; 5University of São Paulo, São Paulo, Brazil

N-acetyl-aspartyl-glutamate (NAAG) is a peptide-based neurotransmitter that has been extensively studied in many neurologic diseases. In this study, we unveiled NAAG's role in the promotion of cancer growth and its elevated presence in more aggressive cancers. In cancers expressing glutamate carboxypeptidase II (GCPII), the enzyme that catalyzes the hydrolysis of NAAG to glutamate and N-acetyl-aspartate (NAA), we demonstrated that NAAG acted as a glutamate provider, restoring cell proliferation upon glutamate deprivation. GCPII was thus identified as an alternative target enzyme for cancer therapy, especially in cancers showing resistance to glutaminase inhibition therapy in current clinical trials. In GCPII deficient-cancers, NAAG also enhanced cancer cell growth but through a metabotropic glutamate receptor 3 (mGlu3)-dependent pathway. Moreover, NAAG was found to be a potential non-invasive metabolite monitor for cancer progression. This newly discovered role of NAAG has greatly reshaped our understanding of its multi-faceted prognostic and therapeutic significance.

Tips and Tricks (if submitted):

Session: Poster, poster number: 058

Uncovering the Role of N-Acetyl-Aspartyl-Glutamate as a Glutamate Reservoir in Cancer

Giang Hoang1; Tu Nguyen1; Brian Kirsch1, 2; Ryoichi Asaka1; Karim Nabi1; Addison Quinones1; Jessica Tan1; Marjorie Antonio1; Felipe Camelo1; Ting Li1; Stephanie Nguyen1; Kiet Nguyen1; Sunag Udupa1; Christos Sazeides3; Yao-An Shen1; Amira Elgogary1; Juvenal Reyes1; Liang Zhao4; Andre Kleensang4; Kaisorn Chaichana1; Thomas Hartung4; Michael Betenbaugh2; Suely Marie5; Jin Jung1; Tian-Li Wang1; Edward Gabrielson1; Anne Le1
1Johns Hopkins Medicine, Baltimore, MD; 2Johns Hopkins Whiting School of Engineering, Baltimore, MD; 3University of Pennsylvania Perelman SOM, Philadelphia, PA; 4Johns Hopkins Bloomberg School of Public Health, Baltimore, MD; 5University of São Paulo, São Paulo , Brazil

N-acetyl-aspartyl-glutamate (NAAG) is a peptide-based neurotransmitter that has been extensively studied in many neurological diseases. In this study, a specific role of NAAG is identified for the first time in cancer.  We found that NAAG is more abundant in higher-grade cancers and is a source of glutamate in cancers expressing glutamate carboxypeptidase II (GCPII), the enzyme that hydrolyzes NAAG to glutamate and N-acetyl-aspartate.   Knocking down GCPII expression through genetic alteration or pharmacological inhibition of GCPII results in decreased glutamate concentrations and reduced cancer cell proliferation in vitro, as well as tumor weights in vivo. Moreover, targeting GCPII in combination with glutaminase inhibition accentuates the reduction of glutamate concentrations, cell proliferation and tumor weights. These findings suggest that NAAG serves as an important reservoir to provide glutamate to cancer cells through GCPII when glutamate production from other sources is limited. Thus, GCPII is a viable target for cancer therapy, either alone or in combination with glutaminase inhibition.

Tips and Tricks (if submitted):

Session: Poster, poster number: 059

The Intestinal Microbiome And Its Metabolites Are Unaltered by Pathogen-Specific Monoclonal Antibodies

Omari Jones-Nelson1; Matthew Glover1; Andrey Tovchigrechko1; Taylor S. Cohen1; Fiona Fernandes2; Udaya Rangaswamy2; Liu Hui2; David E. Tabor2; Paul Warrener1; Jose Martinez1; Jamese Hilliard1; C. Ken Stover1; Wen Yu1; Gina Dangelo1; Sonja Hess1; Bret R. Sellman1
1MedImmune, Gaithersburg, ; 2Medimmune, South San Francisco, California

The microbiome plays a key role in human health and its perturbation is increasingly recognized as contributing to many human diseases. Although antibiotic therapy for bacterial infections has revolutionized medicine, it is now clear that broad-spectrum antibiotics alter the composition and function of the host microbiome. This adverse effect on the microbiome coupled with the emergence of multi-drug resistant pathogens has spurred development of pathogen-specific strategies, such as monoclonal antibodies (mAbs), to combat bacterial infection. We hypothesized that pathogen-specific mAbs against prevalent antibiotic resistant pathogens would not significantly disrupt the intestinal microbiota as compared to conventional antibiotics. We treated 7-week-old, female, C57BL/6 mice with either a single systemic mAb dose or (saline as a control) or human equivalent doses of the antibiotics vancomycin, linezolid, or levofloxacin for 5 consecutive days. Independent experiments were conducted for each indication. We performed microbiome and metabolomic studies using fecal pellets collected from individual animals prior to treatment and for up to 14 days.

The taxonomic abundance and diversity of the bacterial genera and key metabolites (short-chain fatty acids, bile acids and tryptophan metabolites) in the fecal pellets from mice treated with pathogen-specific mAbs were minimally affected and were similar to that in the control groups. In contrast, as expected, metabolites were drastically changed in antibiotic-treated mice (12 hr, and day 7). This correlated with dramatic changes in the relative abundance, as well as alpha- and beta-diversity in the feces of all antibiotic treated groups. While the metabolites showed some restoration at day 14, changes in the microbiome persisted until the end of the study. Taken together, these results indicate that pathogen-specific mAbs do not alter the fecal microbiome like broad-spectrum antibiotic therapies do and may therefore represent an approach to antibacterial therapy that leaves a microbiome intact.

Tips and Tricks (if submitted):

Session: Poster, poster number: 060

Temporal changes in protein abundance and subcellular location reveal HL60 cellular response to methotrexate treatment.

Danielle B. Gutierrez1; Melissa A. Farrow2; Carrie E. Romer1; Jamie L. Allen1; Yuan-Wei Nei3; Zachary Jenkins1; Matthew Hensen5; KayCei Moton-Melancon2; Tina Tsui1; James C. Pino1; Michael Ripperger1; Nicole D. Muszynski1; Salisha Hill1; Kristie L. Rose1; Randi L. Gant-Branum4; Stacy D. Sherrod1; Carlos F. Lopez1; John A. McLean1; John P. Wikswo1; D. Borden Lacy2; Eric P. Skaar2; Jeremy L. Norris1; Richard M. Caprioli1
1Vanderbilt University, Nashville, TN; 2Vanderbilt University Medical Center, Nashville, TN; 3Quest Diagnostics, Chantilly, VA; 4Lackland Airforce Drug Testing Laboratory, Lackland AFB, TX; 5Tempus, Chicago, IL

Elucidation of protein responses to biological stimuli (e.g., therapeutics) is primarily based on the measurement of abundance changes. Recently, large-scale measurement of protein translocation events across subcellular compartments in an untargeted fashion has been demonstrated in a comparison of control and treated cells. However, the comprehensive understanding of cellular response to perturbation requires knowledge of both abundance and localization changes over time. The spatial proteome is dynamic, complicating the validation of a protein’s assignment to a particular subcellular compartment and its translocation between two compartments. We present a strategy to investigate both protein abundance and localization changes in a temporally resolved fashion with specific application to elucidating, in HL60 cells, a comprehensive mechanism of action for methotrexate. HL60 cells were exposed to methotrexate for durations ranging from 20 seconds to 48 hours. Vehicle-treated control and methotrexate-treated samples were processed for whole cell lysate (4 replicates, 12 time points), phosphoenrichment (SILAC labeled, 2 replicates, 10 time points) or fractioned via centrifugation to obtain nuclear, membrane, and cytosolic compartments (3 replicates, 5 time points). Samples were processed by automated digestion (abundance changes) or filter-aided sample preparation (subcellular fractions), desalted, and analyzed by mass spectrometry. In total, 2,102 proteins and 3,253 phosphoproteins changed significantly in abundance and 4,265 proteins from subcellular fractions changed significantly. Of these, 3,228 proteins changed significantly in abundance within a compartment and 2,795 proteins translocated from one compartment to another, with 1,758 proteins overlapping. Temporal abundance and location changes were incorporated into a custom data analysis pipeline for efficient integration and extraction of biological relevance. DNA damage, cell cycle arrest, and apoptosis were explored. This approach provides efficient workflows to gain comprehensive insight into the dynamics of temporal and spatial proteomics during cellular response to perturbation.

Tips and Tricks (if submitted):

Session: Poster, poster number: 061

Multi-omic profiling of TKI resistant K562 cells suggests metabolic reprogramming to promote cell survival

Laura Marholz
University of Minnesota, Minneapolis, MN

Resistance to chemotherapy can occur through a wide variety of mechanisms. Typically, resistance to tyrosine kinase inhibitors (TKIs) is thought to arise from kinase mutations or signaling pathway reprogramming—however, “off-target” adaptations enabling survival in the presence of TKIs without resistant mutations are poorly understood. Previously, we established cell line resistance models for the three most commonly used TKIs in chronic myeloid leukemia treatment, and found that their resistance to cell death was not attributed entirely to failure of kinase inhibition. In the present study, we performed global, integrated proteomic and transcriptomic profiling of these cell lines to describe the mechanisms of resistance at the protein and gene expression level. The proteomic and transcriptional data were correlated to generate an integrated understanding of the gene expression and protein alterations associated with TKI resistance. We identified mechanisms of resistance that were unique to each TKI. Additionally, we defined mechanisms of resistance that were common to all TKIs tested. Resistance to all of the TKIs was associated with the oxidative stress responses, hypoxia signatures, and apparent metabolic reprogramming of the cells. Metabolite profiling and glucose dependence experiments showed that the resistant cells relied on glycolysis (particularly through the pentose phosphate pathway) more heavily than the sensitive cells, which supported the idea that metabolism alterations were associated with resistant cell survival. These experiments are the first to report a global, integrated proteomic and transcriptomic analysis of TKI resistance. These data suggest that targeting metabolic pathways along with TKI treatment may overcome pan-TKI resistance.

Tips and Tricks (if submitted):

Session: Poster, poster number: 062

Integrative proteomics for non-canonical protein and proteome discovery using the ProteomeGenerator

Paolo Cifani; Zining Chen; Avantika Dhabaria; Akihide Yoshimi; Abdel-Wahab Omar; John T. Poirier; Alex Kentsis
Sloan-Kettering Institute, New York, NY

Modern mass spectrometry now permits genome-scale measurements of biological proteomes. However, analysis of specific specimens is hindered by the incomplete representation of protein sequence variability, and by the technical demands for the construction a sample-specific sequence databases. Here, we describe ProteomeGenerator, a framework for de novo and reference-assisted proteogenomic analysis based on sample-specific genome and transcriptome sequencing, and high-accuracy mass spectrometry proteomics. This enables the assembly of sample-specific proteomes encoded by mutant and non-canonical genes, including protein isoforms resulting from alternative mRNA transcription, splicing, or editing. To improve the accuracy of protein identification in non-canonical proteomes, this method relies on statistical target-decoy database matching calibrated using sample-specific controls. The current version of ProteomeGenerator integrates automated peptide-spectral matching using MaxQuant, and is implemented as a Snakemake workflow within a Singularity container for one-step installation in diverse computing environments. We applied ProteomeGenerator for the proteogenomic analysis of splicing factor SRSF2 mutant leukemia cells, demonstrating high-confidence identification of non-canonical protein isoforms arising from alternative transcriptional start sites, intron retention, and cryptic exon splicing as well as improved accuracy of genome-scale proteome discovery. Additionally, we report proteogenomic performance metrics for SEQUEST HT, MaxQuant, Byonic, and PEAKS mass spectral analysis algorithms, obtained using a generalized target-decoy strategy based on inclusion of both decoy spectra and decoy sequences. Thus, ProteomeGenerator should facilitate the discovery of non-canonical and neomorphic biological proteomes, as required for metaproteomic, immune, and cancer proteomic studies.


Tips and Tricks (if submitted):

Session: Poster, poster number: 063-ML

Coordination between TGF-β cellular signaling and epigenetic regulation during epithelial to mesenchymal transition

Congcong Lu1; Simone Sidoli1; Katarzyna Kulej1; Karen Ross2; Cathy H Wu2; Benjamin A Garcia1
1University of Pennsylvania, Philadelphia, PA; 2University of Delaware, Newark, DE

Epithelial to mesenchymal transition (EMT) is a reversible and plastic process in which epithelial cells lose their junctions and polarity to gain a motile mesenchymal phenotype. It plays a crucial role in cancer propagation. EMT can be orchestrated by the activation of multiple signaling pathways, where have been found highly coordinated with many epigenetic regulators. Although the mechanism of EMT has been studied over decades, crosstalk between signaling and epigenetic regulation is not fully understood.

Here we present a time-resolved multi-omics strategy, which featured the identification of the correlation between protein changes (proteome), signaling pathways (phosphoproteome) and chromatin modulation (histone modifications) dynamics during TGF-β-induced EMT. By taking the advantages of SILAC (stable isotope labeling by amino acids in cell culture) based quantification, proteins concentrated at areas of cell-cell adhesion and cell migration were grouped after 1-day TGF-β stimulation, overall suggesting that structural proteins rearrangements evoked in the later transition state. Significant regulations of phosphorylations were detected after only 5 min treatment, leading to the prediction that Erk signaling was activated upon immediate stimulation. The comprehensive profiling of histone post-translational modifications identified H3K27me3 (histone H3 lysine 27 trimethylation) as the most significantly up-regulated mark. This PTM is catalyzed by Ezh2, which is itself regulated by phosphorylations. We thus speculated and confirmed that a combined inhibition of Erk signaling and Ezh2 was more effective in blocking EMT progress than individual inhibitions.

In summary, our data provided a more detailed map of cross-talk between signaling pathway and chromatin regulation comparing to previous EMT studies. Our findings point to a promising therapeutic strategy for EMT-related diseases by combining Erk inhibitor (singling pathway) and Ezh2 inhibitor (epigenetic regulation).

Tips and Tricks (if submitted):

Session: Poster, poster number: 064

Hdac4 interactions in Huntington’s Disease viewed through the prism of multiomics

Joel Federspiel; Todd Greco; Ileana Cristea
Princeton University, Princeton, NJ

Huntington’s disease (HD) is a monogenic disorder, driven by the expansion of a trinucleotide (CAG) repeat within the huntingtin (Htt) gene and culminating in extensive neuronal degeneration in the brain, predominantly in the striatum and cortex. The histone deacetylase 4 (Hdac4) was found to contribute to the disease progression, providing a potential therapeutic target. Hdac4 knockdown reduced the accumulation of misfolded Htt protein and improved HD phenotypes. However, the underlying mechanism remains unclear, given its independence on deacetylase activity and the predominant cytoplasmic Hdac4 localization in the brain. Here, we undertook a multiomics approach to study the Hdac4 contribution to HD pathogenesis. We characterized the interactome of endogenous Hdac4 in the brains of HD mouse models. Alterations in interactions were investigated in response to Htt polyQ length, comparing mice with normal (Q20) and disease (Q140) Htt, at both pre- and post-symptomatic ages (2 and 10 months, respectively). Parallel interaction analyses for Hdac5, a related class IIa Hdac enzyme, highlighted the unique interaction network established by Hdac4. To validate interactions and distinguish those specifically enhanced in an HD-vulnerable brain region, we next characterized endogenous Hdac4 interactions in dissected striata from this HD mouse series. We discovered that Hdac4 associations were largely polyQ-dependent in the striatum, but not in the whole brain, particularly in symptomatic mice. In contrast, Hdac5 interactions did not exhibit polyQ dependence. To identify which functions and proteins within Hdac4 interactions could participate in HD pathogenesis, we integrated our interactome with proteome and transcriptome datasets generated from striata of the same HD mouse models. We found an overlap in enriched functional classes with the Hdac4 interactome, particularly in vesicular trafficking and synaptic functions. This study expands the knowledge of Hdac4 regulation and functions in the context of HD, adding to the understanding of the molecular underpinning of HD phenotypes.

Tips and Tricks (if submitted):

Session: Poster, poster number: 065

A multi-omic approach identifies DLK1 as a candidate oncoprotein and immunotherapeutic target in high-risk neuroblastoma​

Amber K. Weiner1, 2; Alexander B. Radaoui1; Simone Sidoli2; Karina L. Conkrite1; Zalman Vaksman1; Komal S. Rothi1; Pichai Raman1; Jo Lynne Rokita1; Tina Glisovic-Aplenc1; Dan Martinez1; Tricia Bhatti1; Matthew Tsang1; Bruce Pawel1, 2; Benjamin A. Garcia2; John M. Maris1, 2; Sharon J. Diskin1, 2
1Children's Hospital of Philadelphia, Philadelphia, PA; 2University of Pennsylvania, Philadelphia, PA

Immunotherapeutic strategies have produced remarkable results in some malignancies; however, optimal cell surface targets in many cancers remain elusive. We have developed a multi-omic approach to identify high-confidence cell surface oncoproteins for immunotherapy development and applied it to neuroblastoma (NB), an often fatal childhood cancer. We first utilized an optimized sucrose density gradient methodology followed by nano-liquid chromatography coupled to mass spectrometry (nLC-MS/MS) to identify proteins on the surface of NB cells, including 12 human-derived NB cell lines and 10 patient-derived xenograft (PDX) models. This yielded on average 66% (range:60-68%) membrane protein enrichment with high reproducibility between biological replicates (80%; range:78-84%) and identified 1,010 unique proteins containing an extracellular domain. Proteomic data was integrated with extensive NB and normal tissue expression data to discover NB-specific (or enriched)  proteins for validation and assessment of biological relevance. This approach confirmed known cell surface proteins under development as immunotherapeutic targets in NB (ALK, GPC2, DLL3, NCAM1, CD276), and prioritized Delta Like Non-Canonical Notch Ligand 1 (DLK1) for further study. High DLK1 expression in patient tumors obtained at diagnosis predicted poor outcome. Integration of H3K27ac chromatin immunoprecipitation (ChIP) sequencing and RNA-sequencing from NB cells revealed that DLK1 expression was uniquely driven by the presence of a super-enhancer at the locus. DLK1 protein expression was validated by Western blotting, immunofluorescence, flow cytometry, and immunohistochemistry on both NB and normal tissue microarrays (TMAs). Silencing of DLK1 in NB cell models using short hairpin RNAs (shRNAs) resulted in significant growth inhibition and induction of neuronal differentiation. Taken together, we have defined the first MS-based surfaceome of NB and identified DLK1 as a candidate oncoprotein and immunotherapeutic target. Ongoing in vivo studies in PDX models of NB will assess the efficacy of DLK1-based immunotherapy and efforts are underway to extend this multi-omic approach to additional high-risk childhood cancers.

Tips and Tricks (if submitted):

Session: Poster, poster number: 066

Network Integration of Omics Data for Fast-track Identification of the Mechanisms of Action for Drug Candidates

Akos Vertes1; Albert Arul1; Peter Avar1; Andrew Korte1; Lida Parvin1; Ziad Sahab1; Deborah Bunin2; Merrill Knapp2; Andrew Poggio2; Carolyn Talcott2; Brian Davis3; Christine Morton3; Christopher Sevinsky3; Maria Zavodsky3
1George Washington University, Washington, DC; 2SRI International, Menlo Park, CA; 3GE Global Research, Niskayuna, NY

Identification of the mechanism of action (MoA) for drug candidates through targeted assays typically takes years and contributes to the lengthy drug approval process. This challenge is best addressed through high-throughput omics technologies to monitor the cellular response to drug exposure, followed by network mapping of the resulting data for the reconstruction of their MoAs. The response of hepatocytes (HepG2/C3A) exposed to Nexturastat A—an HDAC-6 inhibitor and leukemia drug candidate—was followed at 10 different time-points ranging from 10 s to 48 h. Transcripts and proteins were then analyzed by microarray-based transcriptomics and multiplex shotgun proteomics, respectively. Non-targeted metabolomics experiments were also conducted using UPLC-MS and laser desorption ionization MS from silicon nano-post arrays (NAPA). A network was constructed based on correlation analysis of the molecular time profiles, and the MoA was extracted based on Markov clustering and modularity class analysis. Conventional assays were used to validate the MoA. Omics analyses resulted in the identification and quantitation of >67,000 transcripts, >3,000 proteins, and >400 metabolites over 10 time-points. The statistically-significant regulation of 2,089 transcripts (|FC| >2.0), 235 protein groups (|FC| >1.5), and 203 metabolites (|FC| >1.3) was observed. Spearman correlation-based network mapping of these biomolecules (rs > 0.96) resulted in the formation of 20,286 edges and 19 modularity classes. Gene set enrichment analysis (GSEA) was then performed to identify the biological processes overrepresented in each class. This linked the MoA to cell cycle arrest, DNA damage/apoptosis, and chromatin modification, and to lipid accumulation in the cells. Flow cytometry analysis confirmed G1/S checkpoint inhibition. BODIPY stain for lipid droplets and immunofluorescence against pan-acetyl-lysine confirmed the accumulation of lipid droplets and the upregulation in lysine acetylation, respectively. Our omics approach followed by network mapping rapidly identified the MoA of a drug candidate.

Tips and Tricks (if submitted):

Session: Poster, poster number: 067

UniProt Genomic Mapping for Deciphering Functional Effects of Missense Variants

Peter McGarvey1; Andrew Nightingale2; Hongzhan Huang3; Maria Martin2; Cathy Wu3; UniProt Consortium2
1Georgetown University Med Center, Washington, DC; 2European Bioinformatics Institute (EMBL-EBI), Hinxton, Cambridge, UK; 3University of Delaware, Newark, DE

Understanding the association of genetic variation with its functional consequences in proteins is essential for the interpretation of genomic data and identifying causal variants in diseases.  Integration of protein functional knowledge with genome annotation can assist in comprehending genetic variation within complex biological processes. Here, we describe mapping UniProtKB human sequences and positional annotations such as active sites, binding sites, and variants to the human genome (GRCh38) and the release of a public genome track hub for genome browsers. 

To demonstrate the power of combining protein and genome annotations, we present specific biological examples in disease-related genes and illustrate how missense variants in key functional features of a protein can alter a protein’s structure and function and if severe enough are classified as harmful. Comparisons of UniProtKB annotations and protein variants with ClinVar clinically annotated SNP data show that 32% of UniProtKB variants co-locate with 8% of ClinVar SNPs.  The majority of co-located UniProtKB disease-associated variants (86%) map to 'pathogenic' ClinVar SNPs. To get an overview of variants in different functional features we examined SNPs from ClinVar that overlap selected protein features and plot the percentage of ClinVar SNPs, in each ACMG annotation category, that overlap each protein feature. Six features have more pathogenic variants than either benign or uncertain (Disulfide Bonds, Initiator Methionine, Intramembrane Region, Natural Variant, DNA Binding Domain, Active Site).

In summary, linking annotated data with assertions, publications and other evidence from UniProtKB, ClinVar or other datasets via co-location on the genome, can help integrate protein and genomic analyses and improve interoperability between the genomic and proteomic communities to determine the functional effects of genome variation on proteins.  The location of a variant within functional features may correlate with pathogenicity and would be a useful attribute for use in variant prediction algorithms, including machine-learning approaches.

Tips and Tricks (if submitted):

Session: Poster, poster number: 068

Developing an integrated global, glyco-, and phosphoproteomic workflow for large scale analysis of tissue samples

Yangying Zhou; Tung-Shing Mamie Lih; Ganglong Yang; Shao-Yung Chen; Lijun Chen; Hui Zhang; Qing Kay Li
Johns Hopkins Medical Institutions, Baltimore, MD

The rapid technology development and applications of mass spectrometry-based technologies have enabled the comprehensive proteomic analysis of global proteins and protein post-translational modifications (PTMs). The conventional workflows for characterization of global proteome and proteins with PTMs are often carried out independently and separately. Thus, there is a crucial need for developing an effective and integrated proteomic workflow for global proteins and their PTMs. In this study, we first investigated the impact of protein glycosylation and phosphorylation to global proteomics and found that the use of glycosidase (PNGase F) and phosphatase (CIP) to remove the modifications of global proteins could improve the number of peptide and protein identification of global proteome of breast cancer xenograft tissue samples. We then developed a new integrated workflow for global, glyco-, and phospho-proteomics analysis where samples (after trypsin digestion) went through IMAC enrichment for phosphopeptides followed by MAX enrichment for glycopeptides. The peptides after depletion of phophopeptides and glycopeptides were analyzed for global proteomics. Similar to what we observed by enzymatic de-glycosyaltion and de-phosphorylation, the proteomic analysis of peptides with depleted glycopeptides and phosphopeptides showed 1.18-, 1.19- and 1.30-fold increase in the number of identified proteins, peptides, and PSMs comparing to those without depletion (i.e., global proteome obtained before going through IMAC and MAX enrichment). There was high reproducibility of our integrated workflow for global proteome and suggesting that using PTM-depleted peptides improved the identification of global proteome adequately. Moreover, for the enriched phosphopeptides, we identified an average 5,798 and 9,317 unique phosphopeptides and phospho-PSMs from three replicates with specificity values over 98%. For glycopeptides, we also confirmed the phenomenon of co-capture and elution of glycopeptides from phosphopeptide enrichment and observed different glycan distribution profiles of glycopeptides enriched by MAX. In summary, our integrated workflow provided a new dimension of multi-proteomics analysis of tissue samples. 

Tips and Tricks (if submitted):

Session: Poster, poster number: 069

Mathematic, Machine Learning Methods on DESI Mass Spectrometry of Brain Tissue

Austin Ahlstrom
Brigham Young Univ., Provo, UT, Provo,

Imaging mass spectrometry generates information rich patterns of chemical spectra.  Application of imaging mass spectrometry to spatial regulation of metabolism is an exciting new opportunity.  Independent of this research, methods in mathematics, programming, and machine learning have continued to progress, providing new tools to solve address complex problems. Effective application of such methods could allow significant headway to be made into analysis of mass spectrometric data, uncovering more about the molecular processes that accompany neurodegeneration. We utilized mass spectrometry to collect images of mouse brain tissue. Here, I demonstrate methods for automatic brain region detection and analysis by applying mathematical and machine learning algorithms to mass spectrometry based images of brain tissue.

Tips and Tricks (if submitted):

Session: Poster, poster number: 070-TL

An integrated multi-omic analysis in iPSC-derived motor neurons from C9ORF72 ALS patients

Victoria Dardov1; Ryan Lim2; Vidya Venkatraman1; Jie Wu2; NeuroLINCS Consortium3; Leslie Thompson2; Clive Svendsen1; Jennifer Van Eyk1
1Cedars Sinai Medical Center, Los Angeles, CA; 2University of California, Irvine, Irvine, CA; 3NIH, Bethesda, MD

The NeuroLINCS consortium produced a detailed molecular characterization of motor neurons from induced pluripotent stem cells (iPSCs) derived from patients with amyotrophic lateral sclerosis who carried hexanucleotide expansions in C9orf72 - the most common known cause of ALS. Searching for early molecular differences, we characterized cellular states through ATAC-seq, RNA-seq, and data-independent acquisition mass-spectrometry (DIA-MS) proteomics. Several pathways, including biological adhesion and extracellular matrix organization were altered across epigenomic, transcriptomic, and proteomic data. In addition, protein isoforms predicted from the transcriptomics data were analyzed in the matching proteomics data set to determine if isoforms were translated and if there were quantity changes between C9ORF72 ALS and control. A similar workflow was employed to investigate the presence of novel splice variants identified in the transcriptomic data within the DIA-MS data and if there were quantity changes between C9ORF72 ALS and control. We were able to detect 11 potentially novel splice variants within the DIA data, with five of these being differential between ALS and Control.

Tips and Tricks (if submitted):

Session: Poster, poster number: 071-ML

Host-centric stool proteomics reveals latent-phase-expressed host protease inhibitors modulate EAE severity.

Carlos Gonzalez1; Stephanie Tankou2; Laura Cox2; Howard Weiner2; Josh Elias1
1Stanford University, Stanford, CA; 2BWH, Harvard School of Medicine, Boston, MA

Diseases are often diagnosed once classically recognized symptoms arise, ignoring the latent period prior to disease when biological pathways contributing to disease are altered but disease state is not outwardly detectable. For example, experimental autoimmune encephalomyelitis (EAE), a model for multiple sclerosis, exhibits a latent period of approximately 10 days before overt symptom onset, but changes prior to these symptoms are not well characterized. Recent evidence suggests microbial community structures and specific microbial constituents influence disease outcome of EAE, yet its impact on the latent phase remains unknown. Here, we show the phenotypically latent period between immunization and EAE’s overt symptom onset is characterized by clear host responses, as measured by host-centric proteomics. In particular, we found a sharp, transient increase in protease inhibitor abundance inversely correlated with EAE severity. Antibiotic administration largely attenuated the increase in protease inhibitors. These results strengthen previously discovered ties between proteases and their inhibitors play critical roles in the severity of EAE. More broadly, our findings highlight the utility host-derived stool-omics can have for revealing complex, dynamic biology.  

Tips and Tricks (if submitted):

Session: Poster, poster number: 072

Investigating Modified Lipid Metabolism in Brain Tissue Associated with Varying Isoforms of Apolipoprotein in Relation to Alzheimer’s Disease

John Holman; Peter Jones; Monique Speirs; Russell Denton; John Price
Brigham Young University, Provo, NV

Lipid transport and metabolism are connected to multiple predictors of Alzheimer’s Disease (AD) risk.  Apolipoprotein (APoE), a protein involved in lipid transport is the strongest genetic risk factor. APoE2 demonstrates a low risk for AD development, APoE3 contains neutral risk for AD development, and APoE4 is associated as a risk factor for AD development. Bioactive lipid signaling, has been associated across a variety of cell-signaling pathways and plays a key role in determining cell fate. While it has been shown that possessing the APoE4 isoform increases likeliness of AD development, the connection between APoE4 and modified lipid metabolism has not been fully explored. Using mouse brain tissue collected from the 3 APoE isoforms, I will present our recent work investigating changes in lipid transport and metabolism. This will help determine how lipid transport and metabolism impacts biochemical pathways within the brain to modify AD risk.

Tips and Tricks (if submitted):

Session: Poster, poster number: 073-ML

Proteomic analysis of the developing inner ear in Xenopus laevis

Aparna B. Baxi1, 2; Sally A. Moody1; Peter Nemes1, 2
1George Washington University, Washington, DC; 2University of Maryland, College Park, MD

Congenital hearing loss is one of the most prevalent birth defects in which 1.4 in 1000 infants screened in the US have hearing loss. To understand congenital hearing loss, we need to expand our current knowledge of the proteins that play a role in forming the inner ear during development.

The South African clawed frog (Xenopus laevis) is a well-characterized animal model in developmental biology whose inner ears are evolved for land-based hearing. Recent studies in Xenopus have explored the transcriptomic profiles of developed inner ears to reveal human hearing loss gene orthologs in this animal model. However, the information on the numerous proteins and their dynamics during the formation of the inner ear is still lacking. To address this knowledge gap, we here developed a methodology using liquid chromatography and high-resolution mass spectrometry to quantify protein production in the developing inner ear (otic tissue) of Xenopus laevis. Briefly, otic tissues were manually dissected from five larval stages that represent key transitions in otic patterning starting from the otic vesicle up to the formation of all primary inner ear components. Proteins were extracted from the dissected tissues and digested with trypsin. The resulting peptides from each larval stage were barcoded with a unique tandem mass tag (TMT) and analyzed together to quantify relative proteomic changes between stages of inner ear formation.  Our preliminary analysis led to the identification of over 3,000 protein groups, of which ~74 proteins are associated with hearing loss in humans. We are now exploring this information to identify unique proteomic trends during inner ear patterning. Our approach to study the temporal dynamics of proteins during otic development holds a potential to identify novel proteins guiding the formation of the vertebrate ear.

Tips and Tricks (if submitted):

Session: Poster, poster number: 074

Identification and Validation Protein Phosphorylations Regulating Synapse Loss in Schizophrenia

Matthew Macdonald1; Megan Garver1; Ying Ding1; David Lewis1; Nathan Yates2; Robert Sweet1
1University of Pittsburgh, Pittsburgh, Pennsylvania; 2University of Pittsburgh BioMS Center, Pittsburgh, Pennsylvania

Reductions in dendritic spines, the postsynaptic component of excitatory synapses, have been reproducibly observed in multiple brain areas in schizophrenia (Sz) and are believed to underlie cortical processing deficits. Synaptic plasticity is regulated by synaptic protein network features such as protein trafficking and activity, both of which are mediated by posttranslational modifications (e.g. phosphorylation). Additionally, a significant number of Sz risk genes code for synaptic proteins. Here, in brain tissue from 50 Sz and 50 matched control subjects, we utilized parallel microscopy, proteomic, and phosphoproteomic approaches to identify protein phosphorylations highly correlated with both dendritic spine loss and synaptic protein level alterations in Sz. The effects of one candidate phosphorylation (MAP2 S426) were evaluated in HEK293 cells and mice.

We observed robust changes to synaptic and phosphorylation levels of canonical postsynaptic proteins in Sz (q < 0.05) that were not explained by postmortem interval or protein expression. WGCNA and cross-network analyses observed significant correlations (q < 0.05) between synaptosome, phosphorylation, and dendritic spine alterations in Sz. Nine phosphorylations on eight proteins were highly correlated with both synaptic protein alterations and spine loss.

Our findings suggest that Sz genetic risk and synaptic protein network pathology manifests in processes beyond gene and protein expression, such as protein trafficking and activity. Of the eight proteins with spine and synaptic protein level correlated phospho-alterations, all but one have well-documented roles in vesicular trafficking of postsynaptic glutamate receptors and/or regulating dendritic spines, indicating that they could be upstream of Sz spine loss. Preliminary analysis of one candidate phosphorylation (MAP2 S426E) found that this single modification was capable of impairing MAP2-microtuble binding and resulted in decreased cortical volume, a well replicated finding in Sz, further supporting a role for these phosphorylations in Sz pathology.

Tips and Tricks (if submitted):

Session: Poster, poster number: 075-ML

Understanding epigenome and proteome remodeling caused by novel germline histone H3.3 mutations during neurodevelopment 

Khadija Wilson1; Geoffrey Dann1; Elizabeth J. Bhoj2; Hakon H. Hakonarson2; Benjamin A. Garcia1
1University of Pennsylvania School of Medicine, Philadelphia, PA; 2Children's Hospital of Philadelphia, Philadelphia , PA

Histone H3.3 (H3.3) is a histone variant often found at actively transcribed loci. H3.3 also plays a role in cellular inheritance as ablation of H3.3 expression leads to loss of active gene states and dysfunction of heterochromatin telomeric structures.H3F3Aand H3F3B, the two genes known to encode H3.3, are ubiquitously expressed in all human cells with higher expression in the ovaries, testes, uterus and brain. Somatic mutations in the H3F3Agenes have been reported as oncogenic drivers of pediatric glioblastomas. More recently, here at the Center for Applied Genomics at the Children’s Hospital of Philadelphia novel germline mutations in both H3F3genes have been discovered in a small cohort of patients who suffer from a common pattern of neurodevelopmental disorders, seizures and craniofacial abnormalities. Here we hypothesize that mutant H3.3 (mH3.3) change the regulatory capacity of mH3.3 containing chromatin and therefore modulate gene expression and ultimately the cellular proteome. To this end, we developed stable murine astrocyte cell lines expressing either wild type H3.3 or mH3.3 (G90R, T45I) to study by quantitative mass spectrometry the histone post translational modification changes to the mH3.3 containing nucleosomes and subsequent proteome alterations. Our preliminary data has identified downregulation of chromatin remodeler proteins, as well as upregulation of mitochondrial proteins following reprogramming of the reprogrammed mH3.3 encrusted epigenome. Ultimately, our studies aim to identify and validate potential epigenetic and proteome-wide factors involved in these novel neurodevelopmental disorders. Understanding the basic mechanisms of how these new histone mutations function in neurodevelopment may allow re-classification of epigenome reprogramming in neurological contexts. 

Tips and Tricks (if submitted):

Session: Poster, poster number: 076-TL

Measuring Parkison’s Disease Mitochondrial Protein Turnover Rates in Human IPSC-Derived Organoids by Mass Spectrometry

Anthony Duchesne; Nguyen-Vi Mohamed; Wei Yi; Jean Francois Trempe
McGill University, Montréal, Canada

Parkinson’s Disease (PD) is a currently incurable neurodegenerative disorder that manifests in the elderly through motor symptoms of bradykinesia, rigidity and tremor. PD causes a dopamine deficit, which leads to ineffective neural motor function. For unknown reasons, certain neuronal populations involved in the disease will die whilst others nearby that are very similar the same will survive. One of the prevalent theories explaining this selective death is the mitochondria stress hypothesis, where the neuron populations associated with PD are under more stress than others, resulting in their death from mitochondrial quality control mechanisms. Therefore, understanding the mechanisms of mitochondrial quality control in these PD-associated neural populations is critical. There are two PD-associated proteins: PINK1, a mitochondrial-targeted kinase, and Parkin, a ubiquitin ligase. Previous studies have found that the regulation of mitochondrial proteins in fruit flies is impaired by mutations in Parkin and PINK1, characterized by a deficit in mitochondrial protein turnover. Whether mitochondrial proteins are similarly regulated in mammals has yet to be confirmed. In the current study, our aim was to validate the human induced pluripotent stem cell organoids model in measuring protein turnover. We used mass spectrometry proteomics to examine the effect of a Parkin mutation in a human induced pluripotent stem cell (IPSC) organoid model. We used stable isotope labelling (SILAC) in the amino acid leucine to measure protein half-life and turnover in time-course experiments. Our preliminary results have shown the incorporation of heavy isotope labels into the organoids is associated with a decrease in protein turnover in Parkin mutated organoids. Additional experiments are underway to optimize mass spectrometry methods and proteomic software to maximize the detection and identification of proteins affected by the mutation. In the future, we will expand to other PD-related organoid models to test drugs that stimulate mitochondrial quality control.

Tips and Tricks (if submitted):

Session: Poster, poster number: 077-ML

Early candidate urine biomarkers for detecting Alzheimer’s disease before amyloid-β plaque deposition in an APP (swe)/PSEN1dE9 transgenic mouse model.

Fanshuang Zhang1; Jing Wei2; Xundou Li1; Chao Ma3; Youhe Gao2
1Basic Medicine Peking Union Medical College, Beijing, China; 2Beijing Normal University, Beijing, China; 3Chinese Academy of Medical Sciences, Beijing, China

Alzheimer’s disease (AD) is an incurable age-associated neurodegenerative disorder that is characterized by irreversible progressive cognitive deficits and extensive brain damage. The identification of candidate biomarkers before amyloid-β plaque deposition occurs is therefore of great importance for the early intervention of AD. Urine, which is not regulated by homeostatic mechanisms, theoretically accumulates changes associated with AD earlier than cerebrospinal fluid and blood. In this study, an APP (swe)/PSEN1dE9 transgenic mouse model was used to identify candidate biomarkers for early AD. Urine samples were collected from 4-, 6-, and 8-month-old transgenic mouse models, and the urinary proteomes were profiled using liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS). The levels of 29 proteins differed significantly between wild type and 4-month-old mice, which had not started to deposit amyloid-β plaque. Among these proteins, 13 have been associated with the mechanisms of AD, while 9 have been suggested as AD biomarkers. Our results indicated that urine proteins enable detecting AD before amyloid-β plaque deposition, which may present an opportunity for intervention.

Key words: Alzheimer’s disease (AD), urine proteome, early diagnosis, APP (swe)/PSEN1dE9

Tips and Tricks (if submitted):

Session: Poster, poster number: 078

Single Step Protein Extraction from Trace-amount Human Hair for Genetically Variant Peptide Detection

Zheng Zhang; Meghan Burke; William Wallace; Yuxue Liang; Sergey Sheetlin; Yuri Mirokhin; Dmitrii Tchekhovskoi; Stephen Stein
NIST, Gaithersburg, MD

Recent reports have demonstrated that genetically variant peptides (GVPs) derived from human hair shaft proteins can be used to differentiate individuals of different biogeographic origin. Human hair has the potential to be a suitable alternative to DNA for human identification if GVPs existing in hair cuticular keratins and keratin related proteins can be reliably and accurately identified. To detect them, we first need a sensitive method to extract proteins out of trace-amount human hair shaft. However, hair protein extraction remains challenging due to the extensive cross-linking and poor solubility. We developed a new method called single-step extraction that can extract hair proteins, mainly hair cuticular keratins, from as little as 1 cm-long human hair shaft. We also compared this newly developed method to two previously published methods: 1) ProteaseMax based method; 2) NaOH with SDS repeated extraction method. We found that our newly developed single step protein extraction method is the most sensitive method among the three methods. Furthermore, the construction of a human hair specific peptide mass spectral library, including previously reported GVPs, enables the evaluation of human hair proteome coverage and the effect of each protein extraction method on hair specific peptide identification and GVP detection.

Tips and Tricks (if submitted):

Session: Poster, poster number: 079-WTT

Spray-Capillary: An Electrohydrodynamic Spray Assisted Device for Quantitative Ultra-Low Volume Extraction

Lushuang Huang; Zhe Wang; Si Wu
University of Oklahoma, Norman, OK

Analysis of ultra-low-volume samples provides invaluable insights in studying complex biological systems.  In genomics, single-cell sequencing (SCS) has been advanced to study rare cell population, heterogeneity, etc. However, the proteomics or metabolomics analysis of ultra-low-volume samples remains underachieved mainly because of lacking microsample handling and analysis techniques.


Efforts have been made to improve the sensitivity and throughput of ultra-low-volume sample analysis in MS-based omics, including advanced MS instrumentation, high-resolution separation, and efficient spray-MS interfaces. Although the sensitivity of such approaches has been dramatically improved, considerable sample losses during sample preparation largely limit the application of MS-based omics on ultra-low-volume samples, requiring the development of novel microsample handling techniques.


Herein, we proposed an electrohydrodynamic spray-assisted device for quantitative ultra-low-volume sample extraction, named Spray-Capillary. We utilized the ESI process for quantitative sample extraction through an etched long capillary (bare capillaries for proof-of-principle and coated CE capillaries for on-line separation and detection).  Specifically, ESI voltage is applied to initiate the sample injection. The spray of droplets results in a small vacuum force in the capillary which drives the solution in the capillary move from the sample vial to the ESI emitter. We evaluated the reproducibility and accuracy of sample injection amounts using our Spray-Capillary by varying the ESI voltage, capillary ID, and length. Our results demonstrate that the Spray-Capillary allows reproducible, accurate, and quantitative sample injection ranging from 1 nL to 100 nL or higher.  We further etched a LPA coated capillary for online CE-MS/MS of both peptides and protein mixtures. 


The Spray-Capillary is a very simple quantitative sample handling device for ultra-low-volume samples. It can be directly integrated to a CE capillary for sensitive sample injection and separation without any additional devices, and therefore holds great potential for the high-throughput omics analysis of ultra-low-volume samples, such as single0cell mass spectrometry.


Tips and Tricks (if submitted):

Session: Poster, poster number: 080-WTT

Total solubilization of FFPE samples for high throughput clinical proteomics

John P. Wilson1; Ilyana Ilieva2; Darryl J. Pappin1, 3; John B. Wojcik2
1ProtiFi, LLC, Farmingdale, NY; 2University of Pennsylvania, Philadelphia, PA; 3Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

Formalin fixed paraffin embedding (FFPE) is a decades-old sample preparation technique common in experimental research and medicine. FFPE embedded samples can be stored indefinitely at room temperature, resulting in an exceptionally large and rich worldwide collection. Despite its potential, proteomic analysis of FFPE samples has lagged. Traditionally, samples are first laboriously deparaffinized with often-toxic organic solvents. Subsequent protein extraction is extremely critical but no consensus has been reached as to an optimal protocol. Standardization is fully lacking.

Here, we present a one-pot solution which eliminates deparaffinization and which employs 5% SDS and S-Trap sample processing to exhaustively solubilize entire FFPE samples evaluated on human liver. Samples were paired and split, half flash-frozen and half fixed in formalin and paraffin embedded according to standard histopathology procedures. All samples were extracted with SDS using standard techniques (pulverization, syringe needles), probe sonication or Covaris AFA ultrasonication. SDS was removed by standard precipitation or S-Traps. Protein identification rates and reproducibility were evaluated after analysis on a Thermo QE HF-X or Fusion mass spectrometer.

Compared to standard procedures, the use of S-Traps resulted in significant increases in peptide (≥30%) and protein identification rate (≥20% increase) with greater reproducibility. The use of AFA decreased hands-on time, increased ID rates an additional 6% – 8% and significantly increased protein yield from FFPE samples (80% – ≥200%). The combination of S-Traps and AFA yielded ID rates comparable to those obtained from fresh frozen tissue (101%/ 97% ID rates for peptides/proteins) while eliminating toxic xylene and saving approximately 6 hrs in sample processing by avoiding deparaffinization. 

Our system solves the problem of extraction bias and achieves the goal of reproducible standardized protein recovery from FFPE samples in a workflow suited to automated, high-throughput analyses. We anticipate this workflow will assist to usher in a new era of clinical proteomics.

Tips and Tricks (if submitted):

Huge FFPE sample collections exist; their analysis has been hindered by inefficient processing. We present a highly efficient, one-pot system that uses 5% SDS to dissolve everything, paraffin included, with exceptionally high yield, superior reproducibility and ID rates like fresh-frozen. It saves many hours and eliminates toxic organic deparaffinization steps.

Session: Poster, poster number: 081

Evaluation of a novel LC system that embeds analytes in pre-formed gradients for rapid, ultra-robust proteomics

Nicolai Bache1; Philipp Geyer2; Dorte Bekker-Jensen3; Ole Hoerning1; Lasse Falkenby1; Peter Treit2; Sophia Doll2; Igor Paron2; Florian Meier2; Jesper Olsen3; Ole Vorm1; Matthias Mann2
1Evosep, Odense, Denmark; 2Max Planck Institute of Biochemistry, Martinsried, Germany; 3University of Copenhagen, Copenhagen, Denmark

Mass spectrometry-based proteomics and metabolomics are fast growing and powerful technologies, with the potential to revolutionize health care and precision medicine. However, available separation systems have so far limited throughput and robustness and thereby prevented omic technologies from being fully integrated and routinely applied in clinical settings. Here, we evaluate a conceptually novel liquid chromatography (LC) system that significantly increases robustness and sample throughput for cutting edge proteomic workflows.

The new system, called Evosep One, uses four low-pressure pumps in parallel to elute samples from a disposable and single use trap column while also forming a chromatographic gradient. The sample and gradient are moved into a storage loop that subsequently is switched in-line with a single high-pressure pump and an analytical separation column for separation. 

We evaluated and benchmarked how to use the Evosep One in cutting edge LC-MS workflows to significantly increase overall performance and throughput and demonstrated the value of being able to increase the sample size in studies. The cross contamination (<0.07%), retention time stability and peak properties were monitored over 1500 HeLa runs and the short overhead time of approximately 2 min allows us to efficiently measure 300, 200, 100, 60 or 30 samples per day with corresponding gradient lengths of 3.2, 5.6, 11.5, 21 and 44 minutes, respectively. From fractionated HeLa cell lysates, deep proteomes covering more than 130,000 sequence unique peptides and around 10,000 proteins were rapidly acquired (18 h total instrument time). Using this data as a library for data independent acquisition, we demonstrate the quantitation of up to 5200 proteins in only 21 minutes.

We also demonstrate how this can be applied to clinical research workflows that require uninterrupted analysis of thousands of crude biological samples as well as routine applications such as doping and drug screening. 

Tips and Tricks (if submitted):

Session: Poster, poster number: 082

PASEF on a TIMS-QTOF instrument is reproducible, sensitive and high-throughput for shotgun proteomic workflows

Heiner Koch3; Tharan Srikumar4; Marcus Lubeck3; Thomas Kosinski3; Romano Hebeler3; Florian Meier1; Christopher Adams2; Scarlet Koch3; Gary Kruppa4; Andreas Brunner1; Matthias Mann1
1Max-Planck Inst. for Biochemistry, Martinsried, N/A; 2Bruker Daltonic, San Jose , CA; 3Bruker Daltonic, Bremen, Germany; 4Bruker Daltonic, Billerica, MA

In data dependent acquisition (DDA) experiments 20% of eluting peptide features are targeted for sequencing by MS/MS currently, due to limitations in sequencing speed, sensitivity and resolution. The previously introduced “Parallel Accumulation Serial Fragmentation” method (PASEF, Meier et al., JPR 2015) on a trapped ion mobility spectrometry quadrupole time of flight mass spectrometer has been shown to increase the sequencing speed and sensitivity of MS/MS scans at high resolution. We demonstrate instrument performance on low sample amounts (< 200 ng), offering unprecedented possibilities to investigate samples at high sensitivity, high throughput and high reproducibility.

We observed a sequencing speed > 100 Hz with a gain in sensitivity and no loss in resolution. In each 90 min gradient we could identify more than 5300 proteins (from 200 ng HeLa digest injected on column) and a total of 5672 proteins covering a dynamic range of 4-5 orders of magnitudes. Comparing protein identities between the triplicate analyses, we observed that more than 90% of the identified proteins were identified in each of them. This indicates high reproducibility with the PASEF method. Also, the reproducibility of the label-free intensities determined by MaxQuant is excellent with a R2 = 0.98. To evaluate the accuracy of label free quantification we spiked in E.coli and yeast in two different concentrations into the HeLa proteome. All species could be nicely separated in the expected ratios of 1:2 and 1:5. Using very low sample amounts (<12 ng) of Hela digest and a 60 min gradient resulted in 20,000 unique peptides identified, corresponding to 2000 protein groups. Also using short gradient of 20 min and a sample amount of only 100 ng resulted in the identification of around 25,000 unique peptides and 3000 proteins groups.

Tips and Tricks (if submitted):

Session: Poster, poster number: 083-WTT

30 second analysis of histone post-translational modifications by direct infusion mass spectrometry

Yekaterina Kori1; Simone Sidoli1; Mariana Lopes2; Zuo-Fei Yuan1; Hee Jong Kim1; Katarzyna Kulej1; Kevin Janssen1; Laura Agosto1; Julia Pinheiro Chagas da Cunha2; Benjamin A. Garcia1
1University of Pennsylvania, Philadelphia, PA; 2Instituto Butantan, Sao Paulo, Brazil

Proteomics has become a fundamental and widely recognized discipline for basic science. The next frontier is its application in translational medicine. While much progress has been achieved, proteomics integration in diagnostics still has issues of throughput and repeatability (batch effects). These drawbacks are due to nano-liquid chromatography, as chromatographic columns have highly variable performance and require frequent replacement. We present a novel workflow for the analysis of histone post-translational modifications (PTMs) via direct injection in mass spectrometry (DI-MS) without the usage of liquid chromatography. Histone PTMs play fundamental roles in modulating DNA readout, including gene expression and DNA repair, but additionally, their disregulated abundance serve as drivers and potential diagnostic markers of many types of cancers.

We reduced the two-day sample preparation required to purify and digest histones into a 7 hour protocol (from tissue to ready-to-inject sample) which can be easily parallelized for ~750 samples per batch. The protocol includes a spike-in of a synthetic peptide that benchmarks the quality of sample preparation. Samples are eluted from desalting columns into 96-well plates and are ready to spray using dedicated equipment such as the NanoMate (Advion). Each injection replaces the tip used for pickup, solving the issue of carry-over and more importantly batch effects. Spectra are acquired using a multiplexed targeted selected ion monitoring (tSIM-MSX) combined with MS/MS events to discriminate isobaric forms, collecting the signal of >200 different histone PTMs in less than 30 seconds. Additionally, we have developed a new freely available computational software, EpiProfileLite, for raw file processing and peptide quantification. In conclusion, our workflow is capable of extracting and analyzing >1,000 histone samples per day, compared to the 15-20 we could with our previous state-of-the-art system. We believe this represents an important first step towards more robust analysis of epigenomes for clinical applications.

Tips and Tricks (if submitted):

Session: Poster, poster number: 084-WTT

Kinetics of acetone precipitation: Optimizing conditions to efficiently concentrate and purify protein samples in minutes with the ProTrap XG

Jessica Nickerson; Alan A. Doucette
Dalhousie University, Halifax, Canada


Acetone precipitation is a widely used method of concentrating and purifying complex protein samples in top-down proteomic workflows [3]. Our group previously established that the addition of 1-30 mM salt facilitates protein recoveries >95% [2]. We later developed a filtration cartridge (ProTrap XG), which automates precipitation and SDS depletion [1]. The current bottleneck with acetone precipitation relates to the lengthy (overnight) incubation. In the interest of maximizing protein recovery, purity, and throughput, we investigated the kinetics of protein precipitation, optimizing conditions to achieve the most efficient protein recovery.


Aqueous protein samples (BSA, cytochrome c, and a yeast proteome) were combined with 4 volumes of acetone at a defined temperature (-20 to +37degC), either in conventional vials, or within the ProTrap XG, which contains a PTFE membrane to trap aggregated protein. The samples were incubated for times ranging from 2 minutes to 24 hours followed by centrifugation. The isolated supernatant was retained and residual protein was quantified by a Pierce BCA assay.


It was found that increased temperature, protein and salt concentration provided the most efficient protein recovery. Under optimal conditions, >99% recovery was achieved after just 2 minutes of incubation. Kinetically slower conditions, such as low protein concentration, showed improved rates when additional salt was present or incubation temperature was increased. 

Precipitated yeast proteins will be visualized by SDS PAGE (BioRad) and identified through a standard bottom-up LC-MS/MS approach.

Facilitated by the ProTrap XG and an improved understanding of precipitation kinetics, sample preparation ahead of top-down proteomic analysis is incredibly reliable and efficient. 


1. Crowell et al. (2015). Journal of Proteomics118, 140-150.

2. Crowell et al. (2013). Analytica Chimica Acta796, 48-54.

3. Zhang et al. (2015). Analytical Biochemistry484, 40-50.


Tips and Tricks (if submitted):

Session: Poster, poster number: 085-WTT

Size Dependence for Protein Precipitation: Optimized Conditions for Efficient Recovery of Low-Mass Proteins in the ProTrap XG

Venus Baghalabadi; Alan A. Doucette
Dalhousie University, Halifax, Canada

Introduction: An essential consideration of any protein sample cleanup approach is maintaining high recovery of all sample components during purification. With organic solvent precipitation, low molecular weight proteins are generally less susceptible to aggregation in organic solvent. A question remains as to what conditions enable maximum precipitation of low-mass proteins. Our group previously demonstrated the importance of salt in maximizing recovery of proteins during solvent precipitation [1]. We also demonstrated that aggregated proteins could be rapidly isolated in a two-stage filtration cartridge called the ProTrap XG [2]. We herein extend our previous work with an objective to establish precipitation protocols for the purification and recovery of low molecular weight proteins and peptides, suitable for both top-down and/or middle-down proteomic applications.

Methods: Pepsin-digested BSA was used to generate a mixture of low molecular weight protein fragments (<5 kDa). The sample was precipitated in the ProTrap XG by addition of organic solvent (acetone, acetonitrile, etc.) with inclusion of various concentrations of salt (NaCl, Na2SO4, ZnSO4, etc.). The pellet was isolated from the supernatant and subject to analysis by HPLC with UV quantitation, as well as SDS PAGE and LC-MS/MS to identify proteins/ peptides recovered in the resulting fractions.

Preliminary Results: Our findings indicate a strong dependence on the type of salt in the sample. The highest recoveries were determined using divalent ions (ZnSO4 > Na2SO4 > NaCl). SDS PAGE confirms the recovery of low molecular weight peptides (<5 kDa) in the pellet fraction with intensity similar to that of the control. Subsequent work will evaluate the size trends and properties of peptides recovered in the pellet vs. supernatant following optimized precipitation.



1. Crowell et al., Analytica Chimica Acta, 796 (2013) 48–54.

2. Crowell et al., Journal of Proteomics, 118 (2015) 140-150.

Tips and Tricks (if submitted):

Session: Poster, poster number: 086

Hands-off: fully automated & TMT-compatible sample preparation in less than 4 hours on the PreOn platform employing the iST-NHS technology

Fabian Hosp; Doris Jansen; Nils Kulak
PreOmics, Martinsried, Germany

Sample preparation is a very important component of the overall mass spectrometry-based proteomics workflow and remains to be a limiting factor for high-throughput analyses. One method to increase the throughput is sample multiplexing using chemical labeling approaches such as iTRAQ or TMT. Here, we present a fully automated end-to-end solution for standardized sample processing, including cell lysis, digestion, TMT labeling and peptide cleanup within 4 hours.

In order to minimize sample loss and improve reproducibility, we sought to completely automate TMT sample processing from cell lysis to ready-to-measure labeled peptides. To facilitate this, we aimed to combine the NHS adaption of the iST technology (Kulak et al., 2014) with a newly developed automation platform called PreOn (HSE AG).

The workflow described here integrates robotic handling with chemical labeling of peptides in the very same reaction device used for cell lysis, protein denaturation, reduction, alkylation, digestion as well as the peptide cleanup. This workflow minimizes sample loss, overall hands-on time and the amount of required chemical labels by 50%. The PreOn platform features a built-in centrifuge, ultrasound liquid detection, deck load check and a drag-and-drop method selection for easy, fast and convenient user guidance.

We demonstrate successful labeling of several cancer cell lines using 11-plex TMT achieving >99% labeling efficiency and a reproducibility of R2 = 0.98 for biological replicates (CVs <10%). Furthermore, we present data on more complex samples such as yeast or plasma and combination with peptide fractionation to increase proteome depth.

The described hands-off workflow enables sample processing and TMT-labeling of up to 11 samples in parallel in a fully automated fashion and in less than 4 hours, scaling from 1-100 µg of protein input material. We envision processing of tissue samples combined with chemical labeling to be implemented soon.

Tips and Tricks (if submitted):

Session: Poster, poster number: 087

High-throughput and robust plasma analysis with capillary-flow LC and high-resolution accurate-mass (HRAM) mass-spectrometry

Oleksandr Boychenko1; Jenny Ho2; Christopher Pynn1; Angelito Nepomuceno3
1Thermo Fisher Scientific, Germering, Germany; 2Thermo Fisher Scientific, Hemel Hempstead, UK; 3Thermo Fisher Scientific, West Palm Beach, FL

Analysis of blood and biofluids is a key step in clinical research. The different blood components (drugs, metabolites, electrolytes, proteins) are routinely measured in laboratories. The most widespread methods for quantification of individual proteins rely on antibodies. However, immunoaffinity methods have multiple limitations related to limited specificity and problems with multiplexing. In contrast, LCMS become a reliable technology that combines significant depth into the proteome and impeccable specificity including differentiation of post-translational modifications.

Limiting factors of LCMS adoption are high costs, low throughput, and insufficient robustness of typical nanoLCMS proteomics methods that are used in research. In this work, we developed a new capillary-flow LCMS method with a throughput of 100 samples per 24 hours. The UltiMate™ 3000 RSLCnano system was coupled with Q Exactive™ HF-X HRAM mass-spectrometer. Crude and top 12 depleted enzymatically digested plasma samples were separated on EASY-Spray™ (ES800, 75 µm x 150 mm, 3 µm) column with flow rate 1 µL/min. Low gradient delay volume, fast sample loading (9 sec) and parallel washing of trap cartridge and analytical column allowed to achieve > 85% MS utilization. CapLCMS method was tested in two laboratories by analyzing of > 500 depleted and crude serum samples. The robust retention times (SD < 0.1 min), peak areas (RSD, % < 20), highly symmetric and sharp peaks (PWHM ca. 5 sec) confirmed that method is suitable for the analysis of large sample cohorts. More than 1500 peptide groups from plasma proteins which abundance cover 5 orders of magnitude can be identified in less than 15 min of analysis time.

Thus, capLCMS has a strong potential for providing a reliable, specific and multiplexed solution for identifying plasma protein fingerprints on a routine basis with low-costs per sample.

Tips and Tricks (if submitted):

Session: Poster, poster number: 088

Quantifying Ubiquitination Signaling with a Chemical Proteomics Strategy

Yunan Li; Ang Luo; Luke Erber; Yue Chen
University of Minnesota, Minneapolis,

Stoichiometric analysis of posttranslational modifications is an emerging strategy for absolute quantification of the modification’s fractional abundance. In this study, we present a quantitative chemical proteomic workflow for stoichiometric analysis of ubiquitination, named Isotope-BAlanced Quantification of Ubiquitination (IBAQ-Ub). The strategy utilizes a new amine-reactive chemical tag (AcGG-NHS) for highly efficient labeling of amine residues. The new chemical tag is structurally homologue to the GG remnant of ubiquitin on modified lysine following trypsin cleavage and therefore enables the generation of structurally identical peptides from ubiquitinated and unmodified lysine residues following trypsin digestion and the secondary stable isotopic labeling. Quantitative analysis of labeled peptides with HPLC and high resolution mass spectrometry allows unbiased measurements of absolute site-specific ubiquitination stoichiometry. Using recombinant protein standards and isotopic serial dilution, we showed that the strategy is highly robust, sensitive and accurate. We further designed a SILAC-based experiment to demonstrate the efficient labeling of lysine with the new AcGG tag in complex cell lysates. Application of the IBAQ-Ub workflow for the analysis of enriched histone fractions allowed us to measure the stoichiometry of histone H2B C-terminal ubiquitination. Comparative analysis of stoichiometric dynamics upon proteasome inhibitor treatments showed a dramatic decrease of histone H2B ubiquitination levels and a significant increase of polyubiquitination K48 linkage. Thus, this work provides an efficient chemical proteomics tool for quantitative stoichiometric analysis of ubiquitination signaling pathways in cells and diseases.

Tips and Tricks (if submitted):

Session: Poster, poster number: 089-ML

Building an antiviral platform: Nuclear protein oligomerization as a key contributor to innate immune response

Tim Howard; Krystal Lum; Catherina Pan; Ileana Cristea
Princeton University, Princeton, NJ

    Innate immune responses to invading pathogens rely on the ability of specific host proteins to identify pathogen molecules and induce signals. The interferon-inducible protein IFI16 was shown to recognize pathogenic DNA after a variety of viral infections, including herpesviruses and HIV, as well as bacterial infections. Given the expression of IFI16 in numerous tissues relevant to host-pathogen interactions, including skin, nasal, and oral mucosa, this protein is thought to play a critical role in maintaining balanced immune responses for a healthy system. IFI16 has two functions in host defense against the dsDNA viruses herpes simplex virus type 1 (HSV-1) and human cytomegalovirus (HCMV). It induces antiviral cytokine expression and suppresses viral gene expression. At the core of these antiviral responses are its functional domains: a PYRIN domain (PYD) that mediates its oligomerization and two HIN-200 domains that bind pathogenic DNA. Our lab has previously demonstrated that PYD is required for IFI16 localization to incoming viral DNA at the nuclear periphery. However, how the PYD induces IFI16 oligomerization and the function of this nuclear aggregation in antiviral response remained unknown.

    Here, we used an integrative approach of molecular biology, virology, microscopy, and quantitative proteomics to explore the role of IFI16 oligomerization during herpesvirus infection. We first discover solvent-exposed charged residues that regulate IFI16 oligomerization. This allowed us to generate IFI16 mutants that lack the ability to oligomerize. We find that cells expressing oligomerization-incompetent IFI16 permit more productive infections and result in increased virus protein production. Using immunoaffinity purifications of IFI16 mutants and parallel reaction monitoring (PRM) MS/MS, we determine that IFI16 oligomerization mediates interactions with other critical antiviral proteins. We further establish the necessity of IFI16 oligomerization for immune signaling. Altogether, we provide mechanistic insights into the contributions of PYD oligomerization in viral DNA sensing for innate immune response.

Tips and Tricks (if submitted):

Session: Poster, poster number: 090-TL

Location, location, location:

Using spatial proteomics to uncover functional protein translocations during viral infection

Michelle Kennedy; Ileana Cristea
Princeton University, Princeton, NEW JERSEY

Protein movement between organelles lies at the core of numerous essential cellular processes, such as gene expression, immune signaling, and apoptosis. As obligate intracellular parasites, viruses must co-opt these pathways, and consequently, viral infections induce diverse changes in proteome organization. Our lab has previously characterized the broad changes that occur in the protein composition of organelles during human cytomegalovirus (HCMV) infection. This led us to identify temporal alterations in protein localizations during infection, suggesting the existence of numerous, yet undiscovered, protein movement events. However, the knowledge of the global regulation of protein translocations during HCMV infection remains limited, and the analysis of proteins localized to multiple compartments remains challenging. Here, we designed a computational approach that is robust for multiply-localized proteins and can characterize protein movements between diverse cellular organelles during the progression of HCMV infection. Indeed, we uncovered translocations of both cellular and viral proteins, which we further confirmed and investigated using live cell microscopy. Additionally, this analysis allowed us to discover proteins that co-translocate with one another, pointing to putative functionally-related complexes. Consequently, these movements may contribute to biological processes that underlie either virus replication or host defense. For example, we found that the unconventional myosin, MYO18A, co-translocates with myosin motors and clathrin-associated adaptor proteins to sites of virus assembly. Follow-up functional analyses, including microscopy, virus genome and titer measurements, and knockdown assays, showed that MYO18A is required for virus production. Specifically, MYO18A is hijacked by the virus late in infection to aid the trafficking of viral particles for cellular egress. Overall, this project integrates spatial proteomics, computational biology, and molecular virology methods to gain new functional insights into the contribution of proteome organization to the intricate interplay between viruses and their hosts. Furthermore, this computational method can be applied to characterize translocating proteins during diverse biological processes.

Tips and Tricks (if submitted):

Session: Poster, poster number: 091

Identifying the molecular mechanisms of sex-specific severity of the Middle East Respiratory Syndrome Coronavirus (MERS-CoV) using proteomics

Natarajan Bhanu1; Simone Sidoli1; Ranran Wu1; Neeltje van Doremalen2; Vincent Munster2; Angela Rasmussen3; Benjamin A Garcia1
1University of Pennsylvania School of Medicine, Philadelphia, PA; 2National Institutes of Health, Hamilton, MO; 3Columbia University, New York, NY

The novel Middle East Respiratory Syndrome Coronavirus (MERS-CoV) is a poorly understood pathogen causing high mortality rate (~35%) and surprising higher lethality in male vs female patients. The transgenic human DPP4 mouse is currently the most valuable model to study the poorly understood pathogenesis of MERS-CoV infection.  We applied state-of-the-art mass spectrometry and unbiased non-parametric statistics in a transcriptomics/proteomics project to identify proteins modulated in response to MERS-CoV infection.

We analyzed 90 Trizol homogenates of lung tissue from male and female hDPP4 mice inoculated with high (105 TCID) and low (102 TCID) doses of MERS-CoV obtained on day 1, 2, 3, 4 and 5, along with mock controls. The quantification of the viral proteome showed that high dose inoculation led to similar viral load in the first 3 days, but then this dose decreased in females while it kept growing monotonically in males (Mann-Kendall test). Kaplan-Meier analysis of our mice models confirmed a longer survival rate for females than males. When comparing the proteomics response of high vs low dose of viral inoculation, we observed a remarkably different enrichment in protein functions between males and females; females translated mostly proteins related to defense response, while males “delayed” the response by activating first proteins related to generic translation. We identified remarkable differences in the abundance trends (Kolmogorov-Smirnov goodness of fit test) of proteins involved in the B- cell receptor signaling between sexes (confirmed by transcriptomics analysis), showing that this pathway is activated at an earlier stage in females. Together, we have identified a critical difference in immune response in male vs female mice infected by MERS-CoV using non-parametric statistics on large proteomics data.

Tips and Tricks (if submitted):

Session: Poster, poster number: 092-TL

Antiviral function of mitochondrial sirtuin 4 during human cytomegalovirus infection

Cora Betsinger; Elizabeth Rowland; Ileana Cristea
Princeton University, Princeton, NJ

Sirtuins (SIRTs) are evolutionary-conserved enzymes, known to be ubiquitously expressed in mammalian cells and critical regulators of core cellular pathways. We previously discovered that SIRTs are broad-spectrum viral restriction factors, protecting host cells against DNA and RNA viruses. Among SIRTs, knockdown of mitochondrial SIRT4 led to the largest increase in replication of tested DNA viruses, including the β-herpesvirus human cytomegalovirus (HCMV). HCMV is a prevalent human pathogen, estimated to infect over half of the world’s population. Although HCMV infection is typically asymptomatic, HCMV is a major concern for immunocompromised individuals and the leading infectious cause of congenital birth defects. Here we aimed to define mechanisms underlying SIRT4-mediated antiviral functions and their temporal regulation during HCMV infection. Although initially thought to be a deacetylase and ADP-ribosyltransferase, we discovered that SIRT4 is a more potent lipoamidase. SIRT4 removes the rare but essential posttranslational modification lipoylation from the pyruvate dehydrogenase complex (PDH), thereby inhibiting PDH activity. Therefore, SIRT4 functions as a critical regulator of cellular metabolism by modulating carbon entry from glycolysis into the TCA cycle. To determine the mechanism of SIRT4 antiviral function, we next used a multidisciplinary approach, integrating proteomics, microscopy, enzyme activity assays, and molecular virology. We monitored SIRT4 lipoamidase activity and the regulation of PDH lipoylation and PDH activity throughout the HCMV replication cycle. These findings were placed in the context of the requirement for upregulation of the TCA cycle and fatty acid biosynthesis during HCMV replication. Furthermore, characterization of SIRT4 protein interactions using mitochondrial enrichment and immunoaffinity purification-mass spectrometry led to the identification of a previously uncharacterized viral protein as a novel SIRT4 interaction. Follow-up functional assays via generation of virus strain mutants and enzymatic activity assays led us to demonstrate that this interaction represents a mechanism acquired by HCMV to inhibit the antiviral activity of SIRT4.

Tips and Tricks (if submitted):

Session: Poster, poster number: 093

Developing an analysis pipeline for PfEMP1s in parasites isolated from children presenting with malaria.

Patricia Gonzales Hurtado1; Robert Morrison1; Jose M. C. Ribeiro2; Hussein Magale1; Oumar Attaher3; Bacary Diarra3; Almahamoudou Mahamar3; Amadou Barry3; Alassane Dicko3; Patrick Duffy1; Michal Fried3
1NIAID/LMIV, Rockville, MD; 2NIAID/LMVR, Rockville, MD; 3MRTC-Univ of Sciences Techniques and Technologies, Bamako, Mali

In 2017, an estimated 210 million cases of malaria occurred in the world and 435,000 died globally. Plasmodium falciparum is the deadliest of the four parasites that cause malaria and infects humans.  The parasite proteins on the surface of infected erythrocytes are an important target for vaccines to prevent severe malaria.  PfEMP1 which are surface proteins are important targets for developing a protective immunity to malaria caused by P. falciparum.  The major challenge in P. falciparum proteomics studies is identifying these PfEMP1s at the protein level due to extensive antigenic variation.  To identify these PfEMP1s using shotgun proteomics, we developed a pipeline that searches high resolution mass spectrometry spectra against a custom protein sequence database.  We analyzed parasites from thirty-one P.falciparum infected Malian children. These parasites were grown in culture and then membrane proteins were extracted and electrophoresed followed by in-gel digestion with trypsin.  The pipeline was first validated in the analysis of a laboratory strain with a known PfEMP1, then it was implemented on the analysis of parasite isolates from malaria-infected pregnant women and finally on the analysis of parasite isolates from malaria infected children.

Tips and Tricks (if submitted):

Session: Poster, poster number: 094-TL

Dynamic regulation of mitochondria morphology, composition, acetylation, and function during viral infection

Xinlei Sheng; Laura Murray; Ileana Cristea
Princeton University, Princeton, NJ

Mitochondria govern a wide range of cellular processes, including metabolism, innate immunity, cell death, and signaling pathways. As viruses rely on the regulation of these cellular processes for their replication and spread, mitochondria shape and functions are known to be altered during viral infections. For example, the prominent herpesvirus, human cytomegalovirus (HCMV), induces fragmentation and increased motility of mitochondria as a means to regulate mitochondrial metabolism and inhibit apoptosis and immune response. Further highlighting the critical roles of mitochondria during infection, we have reported that the mitochondrial deacetylase sirtuin 3 (SIRT3) restricts HCMV, suggesting that protein acetylation status is involved in antagonizing viral infection. Here, we aimed to uncover the mechanisms underlying mitochondrial regulation and the role of mitochondrial acetylation during HCMV infection. We first defined the temporal acetylome during HCMV infection, which we placed in the context of protein abundance by using parallel acetyl-peptide enrichment and proteome analyses. Organelle-specific acetylomes were derived from integration of protein localization information based on density fractionation. We found the mitochondrial acetylome to be the most elevated by infection, when compared to other organelles. Mitochondrial protein acetylations were confirmed by mitochondria enrichment and parallel reaction monitoring (PRM). We next interrogated whether this is controlled enzymatically or non-enzymatically. We observed a decrease in mitochondrial pH during infection, indicating that non-enzymatic regulation is unlikely to be the driving force. We hypothesized that SIRT3 interactions contribute to these acetylation changes, and performed immunoaffinity purification to investigate SIRT3 associations with substrates during infection. Proteins involved in oxidative phosphorylation and mitochondrial dynamics displayed changes in both acetylation and SIRT3 association. Indeed, we determine that mitochondrial membrane potential decreased upon infection, and that SIRT3 normally functions to maintain it. Altogether, we provide insights into the regulation of mitochondrial composition and acetylation status during infection, and mechanisms underlying SIRT3 antiviral functions.

Tips and Tricks (if submitted):

Session: Poster, poster number: 095

Quantitative membrane proteomics analysis of low density and high density neutrophils from Staphylococcus aureus infected diabetic mice

Raghothama Chaerkady; Virginia Takahashi; Wen Yu; Taylor S. Cohen; Sonja Hess
MedImmune, Gaithersburg,

Staphylococcus aureus infection is one of the major infections that are difficult to control diabetes pathogenesis. Neutrophils play an important role in defense against S. aureus infection by phagocytosis, oxidative burst and neutrophil extracellular traps (NETs) formation. Low density neutrophils (LDN), a subpopulation of neutrophils has been implicated in many pathological conditions. Understanding distinct features of low (LDN) and high density neutrophils (HDN) in antibacterial defense is important to monitor the effect of therapeutic drugs in treatment of diabetic infections. We carried out quantitative membrane proteomic analysis to unravel the differentially expressed protein machinery in these cell types.

High density neutrophils (HDN) and low density neutrophils (LDN) were isolated from red blood cells and enriched for membrane proteins. Digested membrane proteins were labeled using 6-plex TMT reagents. Labeled peptides were further fractionated into 10 fractions using basic reversed-phase fractionation on Oasis HLB material. LC-MS/MS analysis of TMT labeled peptides was carried out on an Orbitrap Fusion TribridTM (Thermo Fisher Scientific). Each fraction was also analyzed in multinotch MS/MS/MS mode. Mass spectrometry data was analyzed using Proteome Discoverer 2.2 (Thermo Fisher Scientific) software with search engines Mascot (version 2.6.0).

We have identified ~2,330 membrane associated proteins from neutrophils subtypes at 1% false discovery cutoff. Principal Component (PCA) analysis revealed distinct differentially expressed proteins in three LDN and HDN groups. Proteins encoded by Akt2 Ppp2r5d, Ptpa and Tfrc were significantly down-regulated in LDNs. In contrast, proteins encoded Tlr2, Cxcr2, Rab2a, Itm2b, Cd9, Cd177, C5ar1, Fpr2 were up-regulated in LDNs. Pathway analysis revealed protein changes associated with NET formation, AKT activation and central role of GSK3β in neutrophil function. The data from this study helped to understand the contribution of LDNs in S. aureus in diabetic mice model.

Tips and Tricks (if submitted):

Session: Poster, poster number: 096-TL

Quantitative proteomics reveals host protein SLFN5 as a target of HSV ICP0-mediated ubiquitination and degradation.

Joseph M. Dybas1; Eui Tae Kim1; Emigdio D. Reyes1, 2; Katarzyna Kulej1; Jennifer C. Liddle1, 2; Benjamin A. Garcia2; Matthew D. Weitzman1, 2
1Children's Hospital of Philadelphia, Philadelphia, PA; 2Perelman School of Med, University of Pennsylvania, Philadelphia, PA

Viruses have evolved mechanisms to hijack the host ubiquitin system in order to control ubiquitination of host proteins and thereby modify the cellular proteome and subvert host immune responses. Herpes simplex virus 1 (HSV) encodes “infected cell protein 0” (ICP0), a viral E3 ubiquitin ligase that is necessary for efficient infection. ICP0 is known to ubiquitinate and degrade host anti-viral factors, but the extent to which HSV utilizes ubiquitination during infection is not well understood. In order to identify host proteins that associate with the viral genome during infection, we used a quantitative proteomics approach that combines the “isolation of proteins on nascent DNA” (iPOND) technique with mass spectrometry (MS). We hypothesized that iPOND-MS data generated during wildtype or ΔICP0 HSV infection would identify potential ICP0 substrates by quantifying ICP0-dependent changes in the proteome associated with the viral genome. Using a PCA-based clustering strategy, we identify SLFN5 as a host protein that exhibits abundance patterns similar to known ICP0 substrates during infection, suggesting that SLFN5 is a potential ICP0 substrate. Additionally, di-glycine remnant profiling, combined with whole cell proteomics data, show that SLFN5 is ubiquitinated by ICP0 and decreased during infection. Subsequent experiments confirm the proteasomal degradation of SLFN5 during HSV infection and show that transcription of viral genes is enhanced by SLFN5 knockdown. These data suggest that SLFN5 is a restriction factor of HSV that inhibits viral gene transcription. Our study demonstrates that identification of the proteome associated with viral genomes during infection with mutant viruses can reveal host proteins that are degraded or mislocalized as the virus counteracts intrinsic host defenses. We used quantitative proteomics to identify the host protein SLFN5 as a novel substrate of HSV ICP0.Our data provide the first indication that SLFN5 is a restriction factor for DNA viruses and is degraded by a viral activator of transcription.

Tips and Tricks (if submitted):

Session: Poster, poster number: 097

Quantitative crotonylome analysis expands the roles of p300 in the regulation of the lysine crotonylation pathway.

Mathew Perez-Neut; Huang He; Yejing Wang; Yingming Zhao
University of Chicago, Chicago, IL


We recently discovered nine chemically distinct lysine modifications that can occur on histone proteins, including lysine crotonylation (Kcr)1,2. We demonstrated that lysine crotonylation can be enzymatically installed by acetyl-transferase, p300. p300-catalyzed histone Kcr is associated with active gene expression3,4. Nevertheless, the substrates regulated by p300 remain largely unknown, hindering efforts to establish other potential mechanistic links between Kcr and p300-madiated phenotypes. Here we used quantitative proteomics to characterize the p300-regulated lysine crotonylome in wildtype (WT) and in p300 knockout (KO) cells. A total of 816 unique endogenous crotonylation sites were identified across 392 protein, with 88 sites from 69 protein dynamically decreasing by 0.7-fold (log2 <-0.5) in response to p300 KO and 31 sites from 17 proteins increasing by 1.4-fold (log2>0.5) in response to p300 KO. The most significantly altered crotonylome pathways in the p300 KO cells are centered around key components of nonsense-mediated decay, infectious disease pathways, and viral/eukaryotic translation pathways. Further network analysis highlights the complexity and interconnectivity of the p300-targted Kcr substrates. Moreover, we reveal that some p300-targeted Kcr substrates are potentially linked to diseases such as cancer. Taken together, our results reveal the p300-regulated lysine crotonylome, shedding light on the role for lysine crotonylation in the regulation of diverse cellular processes, and providing the foundation for the discovery of new mechanistic roles for p300. References: 1. Tan, M., Wysocka, J., Ye, Y., Khochbin, S., Ren, B.,  Zhao, Y. (2011). Cell146(6), 1016-28. 2. Bhanu, N., Yuan, Z., Garcia, B. (2018) H2AV Lysine Crotonylation: An Epigenetic Switch During Human Myogenic Differentiation. World HUPO. 3. Sabari, B. R., Huang, H., Zhao, Y., Roeder, R. G., Allis, C. D. (2015). Molecular cell, 58(2), 203-15. 4. Li, Y., Sabari, B. R., Huang, H., Tang, Z., Zhao, Y., Roeder, R. G., Shi, X., Allis, C. D., Li, H. (2016). Molecular cell62(2), 181-193.

Tips and Tricks (if submitted):

Session: Poster, poster number: 098

Development of Mass Spectrometry-Compatible Peptide Biosensors to Detect Kinase Activity

Nicole Wolter; Tzu-Yi Yang; Naomi Widstrom; Laurie Parker
University of Minnesota-Twin Cities, Minneapolis, MN

Protein kinases are responsible for attaching phosphate groups to other molecules as part of a signaling pathway. Kinases are strictly regulated, and can cause cancer when mutated; this is exemplified by the Philadelphia chromosome, a translocation mutation which renders Abl kinase constitutively active and causes chronic myeloid leukemia (CML). Mutations in other kinases, such as Janus kinase 2 (Jak2), have also been implicated in CML. Treatment of CML was progressed following development of kinase inhibitor imatinib (Gleevec®); unfortunately, a significant portion of patients are resistant to the inhibitor, rendering this therapy futile. Because of the importance of kinase response to inhibitor treatment, it is advantageous to quickly measure kinase activity in response to treatment. This can be accomplished through the use of a biosensor-a small peptide phosphorylated by the kinase and ideally analyzed by mass spectrometry (MS), due to the sensitivity of this method to analyze clinically-relevant concentrations of cell samples. The goal of this project is to use KINATEST-ID, an in silico kinase substrate generator, to create a biosensor compatible with MS analysis methods to detect kinase activity. Biosensors were generated by KINATEST-ID, and then run through machine-learning algorithms in CONSeQuence to screen for predicted MS ionization efficiency, an indicator of MS detection compatibility. These substrates were then synthesized and purified, and used in an in vitro kinase assay with enzyme-linked immunosorbent assay (ELISA) analysis to determine biosensor specificity; these experiments were then repeated with other kinases to test for biosensor selectivity. We plan to continue analyzing ionization efficiency with synthesized Jak2 substrates, and will move forward by developing and analyzing substrates for Src family and Syk kinases.  

Tips and Tricks (if submitted):

Session: Poster, poster number: 099-ML

cGAS is Regulated by Phosphorylation and Acetylation during Infection with Herpes Simplex Virus 1

Bokai Song; Krystal Lum; Ileana Cristea
Princeton University, Princeton,

Cyclic GMP-AMP synthase (cGAS) is a DNA sensor critical for initiating innate immunity in response to diverse types of pathogen infections. cGAS detects pathogenic DNA and induces the host expression of Type I interferons and other antiviral cytokines upon infection with both cytoplasmic- and nuclear-replicating DNA viruses, as well as upon bacterial infections. cGAS binds double stranded DNA and synthesizes cyclic GMP-AMP (cGAMP) to activate the STING-TBK1-IRF3 pathway and induce the cytokine expressions. With this central immune signaling axis defined, focus now starts to be placed on understanding the regulation of cGAS. What triggers cGAS activation and localization to pathogenic DNA? Although posttranslational modification (PTMs) has emerged as an important facet of cGAS regulation, only few cGAS modifications are known to date. Here, we aimed to acquire a better understanding of PTMs that can regulate cGAS during either cellular homeostasis or active immune signaling. We enriched cGAS via immunoaffinity purification and used quantitative mass spectrometry to characterize the cGAS modification status in uninfected primary human fibroblasts and upon sensing viral DNA. We compared cGAS PTMs upon infection with herpes simplex virus type 1 (HSV-1) and detection of cytoplasmic vaccinia virus DNA. These analyses led us to discover novel cGAS phosphorylation and acetylation sites during active immune signaling. We further validated and quantified infection-induced changes in these PTMs using parallel reaction monitoring (PRM). To characterize the function of these modifications, we generated a series of cGAS mutants that either mimic the acetylation or phosphorylation or an unmodified state. As a positive control, we compared these mutations to the mutation of the previously characterized S305 phosphorylation site. This allowed us to identify cGAS PTMs that impact cytokine induction, providing new insights into the dynamic regulation of cGAS during active immune signaling.

Tips and Tricks (if submitted):

Session: Poster, poster number: 100

Mapping Disulfide Linkages Without Having to Detect Disulfide-linked Peptides

Tommy K Cheung; Twyla Lombana; Marissa Matsumoto; David Arnott
Genentech Inc, South San Francisco, CA

Disulfide bonds are important structural elements of proteins; biological activity depends on their correct formation, and their assignment and verification are essential to the development and production of protein therapeutics such as monoclonal antibodies. Assigning disulfide connectivity is, however, a frequently difficult task, primarily because of the need to detect and perform tandem mass spectrometry on the crosslinked peptides. These can be large, sometimes multiply branched, and generally different in their chromatographic and mass spectrometric behaviors from the reduced and alkylated peptides routinely analyzed in the proteomics field.  We therefore took an alternative approach that enables disulfide mapping on fully reduced and alkylated peptides that are chemically tagged in such a way that information about their original linkages is preserved.

Proteins were treated with dithiothreitol (DTT), varying time or temperature to yield partial reduction of the disulfides.  Treatment with an alkylating reagent (e.g. iodoacetamide) marked all cysteines that were in their reduced forms.  Proteins were then fully reduced and all remaining cysteines labeled with a second alkylating reagent (e.g. n-isopropyl iodoacetamide or acrylamide). Alternatively, isotopically labeled alkylating reagents such as deuterated versions or Iodo-TMT can be employed. Enzymatic digestion (typically tryptic) was performed and conventional reverse phase chromatography and tandem mass spectrometry were used to detect the now-fully reduced and alkylated peptides.  The fraction of each cysteine alkylated with each reagent was used to calculate the amount of each cysteine that was in its reduced form after partial reduction.  Symmetry requires that cleavage of a disulfide bond releases the involved cysteines to identical degree, so pairs of cysteines with the same extent of reduction are inferred to have been disulfide-linked.  Examples of therapeutically important proteins including human growth hormone and monoclonal antibodies demonstrate that disulfide linkages can be assigned without directly detecting the disulfide-linked peptides. 

Tips and Tricks (if submitted):

Session: Poster, poster number: 101-ML

Ethionine, produced by commensal Lactobacillus reuteri, is immunomodulatory, proteogenic, and leads to ethylation of human proteins

Daniel Röth1; Abby Chiang1; Gabriel Gugiu1; Christina Morra2; James Versalovic2; Markus Kalkum1
1City of Hope, Duarte, CA; 2Baylor College of Medicine, Houston, TX

Colonization of the human gut by immunomodulatory Lactobacillus reuteri strains reduces the risks for inflammatory diseases and colorectal cancer. The underlying molecular mechanisms, however, are not fully understood. Our previous research showed that the immunomodulation is dependent on L. reuteri’s folate metabolism. Mass spectrometric (MS) analysis of L. reuteri strain 6475 folyl-polyglutamates revealed a novel folate derivative, corresponding to 5,10-methenyl tetrahydrofolate that carried an additional methyl group. It was identified by MS fragmentation and NMR to be 5,10-ethenyl tetrahydrofolate (Et-THF). Isotopic labeling of folic acid precursors and substrates of the folate cycle demonstrated that the two ethenyl carbons originated from acetate, thereby identifying acetate as a novel carbon source for the folate cycle in this commensal bacterium. Moreover, Et-THF not a dead-end product, but a substrate of the folate cycle. It enables the transfer of ethyl groups instead of methyl groups onto homocysteine, leading to the production of ethionine. The unconventional amino acid ethionine has immunomodulatory functions: Treatment of human monocytic THP-1 cells with ethionine, resulted in reduced proliferation and LPS-induced immune activation. Mass spectrometric analysis of histones of ethionine-treated THP-1 cells revealed a reduction of lysine methylation. Furthermore, we detected partial ethylation of histone tail lysine residues as a novel posttranslational protein modification. Ethionine was also incorporated into proteins in place of methionine, demonstrating the proteogenic potential of this uncommon amino acid. Our study implicates human microbiome as capable in modifying key human host proteins involved in gene regulation.

Tips and Tricks (if submitted):

Session: Poster, poster number: 102-TL

A Novel Method of Quantifying Protein Methylation Utilizing DIA-MS

Aaron Robinson1; Ronald Holewinski1; Vidya Venkatraman1; Jose Mato2; Shelly Lu1; Jennifer Van Eyk1
1Cedars Sinai Medical Center, Los Angeles, CA; 2CIC bioGUNE, Bizkaia, Spain

Protein methylation of Arginine and Lysine is an important post-translational modification in disease but little is known about its effect in-vivo. Global protein methylation analysis has recently been enabled by the development of peptide immunoprecipitations against methylated Lysine and Arginine and subsequent mass spectrometry.  However, methyl-enrichments require large amounts of sample and multiple antibodies to cover all methyl forms, the antibodies contain a sequence bias masking a portion of the methylome, and total protein quantification needs to be done separately from a methyl enrichment inhibiting the ability to study site occupancy. We have developed a method to differentiate an unmodified peptide from a peptide containing mono-, di- or tri-methylated Lysine or Arginine through a data independent acquisition approach utilizing small precursor mass windows.  We then applied our method to complex cellular lysate from differentially methylated in-vivo non-alcoholic steatohepatitis mouse models. By first creating a biologically hyper-methylated peptide assay library to which each experimental sample is compared, we were able to identify and quantify protein methylation without the need to enrich each sample.  In conclusion, our method provides the framework needed to study global protein methylation, as well as total protein quantification from the same data independent acquisition.

Tips and Tricks (if submitted):

Session: Poster, poster number: 103

Method development for phosphorylation and glycosylation detection using Orbitrap Fusion Lumos

Susanne B. Breitkopf; Jeffrey A. Culver; Michelle F. Clasquin; Bei Betty Zhang; Mara Monetti
Pfizer, Inc., Cambridge,

Nearly all proteins undergo post-translational modifications (PTMs) which are crucial for function, structure, activity, expression of the proteins and protein interactions. Particularly glycosylation and phosphorylation are involved in pathways critical for signaling, and are often altered in disease states. Mass spectrometry (MS) is a powerful tool for the identification of protein glycosylation and phosphorylation due to its sensitivity of detection and its ability to analyze complex mixtures. Combining multiple MS fragmentation techniques (i.e. HCD, ETD, etc) allows for a comprehensive structural characterization of modified proteins.

Phosphorylation occurs on serine, threonine, and tyrosine residues. Typical methods for enrichment include immobilized metal affinity chromatography (IMAC) or titanium dioxide (TiO2) beads, which have selective affinity and interact with phosphopeptides. Comparing both methods we identified more unique phosphosites with IMAC compared to TiO2 enrichment. MS fragmentation techniques are also crucial for successful identification of PTMs and the most phosphosites were identified with IT-(ion trap)-HCD fragmentation using a top speed method (scan cycle with a maximum of 3 seconds).

Glycosylation can occur on asparagines (N-linked) or on serine or threonine residues (O-linked). Enrichment strategies such as TiO2, which is selective for sialic acid-containing glycopeptides, and lectins, with their ability to bind glycoconjugates, are often used. In our comparison the lectin enrichment exceeded the TiO2 strategy. In addition lectin enrichment on Filter Aided Sample Prep (FASP) filters is preferable for O-linked glycopeptides, lectin enrichment on agarose beads is superior for N-linked glycosylation. Similarly to phosphorylation, we compared different MS fragmentation techniques and identified the IT-ETD_OT (orbitrap)-HCD as the method that provides the most complete protein glycosylation coverage.

Tips and Tricks (if submitted):

Session: Poster, poster number: 104-TL

ELTA-MS: Labeling, enrichment and identification of ADP-ribosylated peptides by mass spectrometry

Robert Lyle Mcpherson; Anthony Kar Lun Leung
Johns Hopkins University, Baltimore, MD

The post-translational modification of polypeptides is a ubiquitous mechanism by which cells control the function, localization, and stability of cellular proteins. The modification of proteins with monomers or polymers of the small molecule ADP-ribose is known as ADP-ribosylation and is implicated in a wide variety of cellular processes including DNA repair and antiviral responses. In many cases the identities of ADP-ribosylated substrates and specific sites of conjugation are, however, unknown. In recent years, liquid chromatography-tandem mass spectrometry (LC-MS/MS) has proven to be an unparalleled tool for the global identification of PTMs on substrate proteins. Because ADP-ribosylation is not an abundant modification, the modified substrates must be enriched before LC-MS/MS analyses. Existing approaches utilize immobilized chemical groups and protein domains with an affinity for the ADP-ribose moiety to enrich for ADP-ribosylated peptides. These protocols are capable of identifying hundreds of sites of ADP-ribosylation. However, there is incomplete overlap between the different data sets, suggesting that the complete ADP-ribosylome has yet to be uncovered and highlighting the need for an approach with higher sensitivity. Recently our group has developed a novel method called ELTA for labeling ADP-ribosylated substrates with chemical groups, including tags such as biotin and clickable functional groups. Here we have applied this approach to optimize a workflow (ELTA-MS) for the specific labeling and functionalization of ADP-ribosylated peptides with biotin followed by selective enrichment and identification by mass spectrometry.

Tips and Tricks (if submitted):

Session: Poster, poster number: 105

Detection of BTK activity using a terbium-chelating peptide biosensor

Naomi Widstrom1; Minervo Perez1, 2; John Blankenhorn1; Lindsay Breidenbach1; Hannah Peterson1; Laurie L. Parker1
1University of Minnesota, Minneapolis, MN; 2Purdue University, West Lafayette, IN

Protein tyrosine kinases are involved in many crucial cell signaling pathways. Due to this, tyrosine kinases frequently exhibit altered activity in many types of cancer, making them excellent targets for cancer therapeutics. Efficient and effective assays are needed to test candidate tyrosine kinase inhibitors. To address this need, we develop peptide biosensors that are both good substrates and selective for the kinase of interest. In addition, several biosensors have been designed to selectively chelate to terbium. Terbium binds with stronger affinity to the phosphorylated tyrosine and has a longer luminescence lifetime, allowing a clear distinction between the phosphorylated and unphosphorylated forms of the peptide. We have illustrated this concept using a biosensor for Bruton’s tyrosine kinase (BTK), a nonreceptor tyrosine kinase implicated in B-cell leukemias. Preliminary work identified sequences preferentially phosphorylated by BTK using phospho-proteomics. These sequences were run through the in-silico pipeline KINATEST-ID to generate a library of putative BTK substrates. Of these substrates, BTK artificial substrate D (BAStide-D) was designed to selectively chelate to terbium. We demonstrated that the percentage of phosphorylated peptide present can be determined using time-resolved luminescent detection. Kinase assays demonstrated that BAStide-D is rapidly phosphorylated by BTK, indicating it is an efficient substrate. We aim to further characterize BAStide-D by characterizing the kinetics of the biosensor and kinase interaction. Additionally, we aim to use this assay to evaluate the response of BTK to inhibitors. This efficient antibody-free assay can be used to screen potential BTK inhibitors.

Tips and Tricks (if submitted):

Session: Poster, poster number: 106

Human Breast Cancer Cell Line Phosphoproteome Revealed by an Automated and Highly Selective Enrichment Workflow

Shuai Wu; Linfeng Wu
Agilent Technologies, Santa Clara, CA


Immobilized metal affinity chromatography (IMAC) using a nitrilotriacetic acid (NTA) chelating ligand functionalized with Fe(III) is one of the most widely adopted phosphopeptide enrichment techniques for LC/MS applications. Agilent AssayMAP Bravo platform is able to provide a fully automated and highly selective phosphopeptide enrichment workflow using high capacity Fe(III)-NTA cartridges. In this study, we evaluated how the ratio of total peptide sample amount to affinity resin affects the performance of the enrichment results.



The Agilent AssayMAP was used for automated phosphopeptide enrichment with Fe(III)-NTA cartridge. Human MCF7 breast cancer cell line tryptic digest was loaded onto the cartridges with four different sample to resin ratios. With a 90-min nanoLC gradient, 6550 iFunnel Q-TOF was used for phosphopeptide discovery. MRM analysis of 20 light and heavy pre-spiked phosphopeptide standards was performed on 6495B QQQ to measure their overall recovery from the enrichment procedure.



Prior to enrichment, about 1.5% of peptides were identified as phosphopeptides. After enrichment, phosphopeptide selectivity was routinely above 90% across all 12 samples. Injecting 1/5 of the final enriched sample, Spectrum Mill was able to identify about 1,200 to 1,500 total number of phosphopeptides from the 4 ratios of sample. About 55% of all phosphorylation sites were assigned. The recovery rates were consistent for all the standards across the 4 ratios of sample.



AssayMAP Bravo provides fully automated, highly selective and reproducible enrichment for phophopeptides. Nanodapter effectively converts UHPLC to nanoflow UHPLC which allows users to have both standard flow and nanoflow LC in one system. Together with nanoESI source, 6550 Q-TOF offers the ultimate sensitivity for maximum number of phosphopeptide identification. 6495B Triple Quadrupole offers fast, accurate and robust MRM based peptide quantitation.


Keywords: AssayMAP, Fe(III)-NTA, phosphopeptide enrichment, peptide quantitation

Tips and Tricks (if submitted):

Session: Poster, poster number: 107

Ischemic Stress to Kidneys from SIRT5 Mice is mitigated by Succinylation Response

Kevin Peasley1; Anja Holtz2; Nathan Basisty2; Takuto Chiba1; Birgit Schilling2; Sunder Sims-Lucas1; Eric Goetzman1
1University of Pittsburgh, Pittsburgh, PA; 2Buck Institute for Research on Aging, Novato, CA

Oxidative stress and kidney dysfunction are relevant pathological consequences due to ischemia/reperfusion-induced acute kidney injury (AKI). In previous studies we observed protection against AKI in SIRT5 knockout (KO) mice. Here, we investigated the effect of ischemic stress on the kidney succinylome in SIRT5KO versus wild-type (WT) mice.


In WT and SIRT5KO mice (N=3), one kidney was subjected to 22 minutes of ischemia while the other kidney served as uninjured control. The protein succinylome was determined in whole kidney. Following tissue lysis and tryptic digestion, succinylated peptides were enriched with the PTMScan succinylation affinity enrichment kit (CST). Samples were subjected to data-independent acquisitions (DIA) on a TripleTOF 6600 system. Succinylation sites were quantified using Skyline.


Data-independent acquisitions for succinyl-lysine from mouse kidney yielded overall over 1000 succinylation sites.  We identified more unique succinylated sites (731) in injured WT than uninjured WT (528); however, we detected fewer unique sites in injured Sirt5KO (891) versus uninjured SIRT5KO (989).  Quantitative comparisons between the different conditions revealed mostly robust up-regulation of succinylation levels in KO vs. WT, both from uninjured and injured kidneys.  More than 95% of the succinylation sites were on mitochondrial proteins. Pathway analysis indicated that the most significantly affected pathway was mitochondrial fatty acid oxidation (FAO). Interestingly, for many of the Sirt5-regulated sites on FAO enzymes and respiratory chain targets, the fold changes between KO and WT were lower in the injured kidneys than in the uninjured kidneys. We also quantified 50 sites that were significantly less succinylated in injured KO kidneys compared to injured WT kidneys, while in uninjured KO kidneys only hypersuccinylation was observed. We believe that the total succinylation load increases during ischemic injury in WT but not KO kidneys, possibly due to higher mitochondrial metabolism in the WT driving succinyl-CoA formation thus leading to higher protein succinylation levels.

Tips and Tricks (if submitted):

Session: Poster, poster number: 108

Adenoviral proteins E1B55K and E4orf6 use non-degradative ubiquitination to regulate viral late protein expression

Christin Herrmann1, 2; Jen Liddle1, 2; Joseph Dybas1, 2; Benjamin A. Garcia2; Matthew Weitzman1, 2
1Children's Hospital of Philadelphia, Philadelphia, PA; 2University of Pennsylvania, Philadelphia, PA

During the early phase of infection, viruses establish favorable conditions for late phase infection by co-opting host signaling processes and countering host antiviral defenses.  Adenovirus is a DNA virus important both for its role in human diseases and for its archetypal insights in understanding virus-host interactions. Two early adenoviral proteins, E1B55K and E4orf6, associate with the host Cullin-5 ubiquitin ligase complex and redirect substrate recognition in order to promote an environment conducive for viral replication.  Mutating these viral genes impairs production of late viral proteins and reduces viral progeny.  Previous work has identified a limited number of substrates of the E1B55K/E4orf6 complex which are ubiquitinated and degraded upon infection.  However, these known targets do not account for the decrease in late viral protein production.  Here we identify new targets of the viral ubiquitin ligase that may account for the late phase defects in the E1B55K-deficient virus.  We transduced HeLa cells with E1B55K and E4orf6, enriched for ubiquitinated proteins over a timecourse of E1B55K/E4orf6 expression, and applied mass spectrometry to identify potential substrates.  We combined these results with proteomic analysis to normalize ubiquitin levels against protein abundance.  In total, we identified >6000 proteins and >4000 modified peptides.  Our integrative analysis reveals that ubiquitination occurs both on proteins that decrease over time, characteristic of degradative signaling, and on proteins whose abundances are unchanged, characteristic of non-degradative signaling.  Our non-degraded ubiquitination targets are enriched in RNA-binding proteins, with hnRNP-C and RALY among the most abundantly modified.  We show that these homologous proteins are specifically ubiquitinated by the E1B55K/E4orf6 complex and that they play a functional role in viral late protein production in an E1B55K-dependent context.  Overall, we provide the first known case of viral-mediated non-degradative ubiquitin signaling affecting protein production.

Tips and Tricks (if submitted):

Session: Poster, poster number: 109-ML

Lysine benzoylation is a histone mark regulated by SIRT2.

Mathew Perez-Neut; He Huang; Di Zhang; Yingming Zhao
University of Chicago, Chicago, IL


Metabolic regulation of histone marks is associated with diverse biological processes through dynamic modulation of chromatin structure and function1. Here, we report the identification and characterization of a histone mark, lysine benzoylation (Kbz)2. Our study identifies 22 Kbz sites on histones from HepG2 and RAW cells. This histone mark can be stimulated by sodium benzoate (SB), an FDA-approved drug and a widely used chemical food preservative, via generation of benzoyl-CoA. ChIP-seq and RNA-seq analysis demonstrate that histone Kbz marks are associated with gene expression and have physiological relevance distinct from histone acetylation. In addition, we demonstrate that SIRT2, an NAD+-dependent protein deacetylase, removes histone Kbz both in vitro and in vivo. This study reveals a new type of histone mark with potential physiological relevance and identifies possible non-canonical functions of a widely used chemical food preservative. References: 1. Sabari, B. R., Zhang, D., Allis, C. D., & Zhao, Y. (2016). Metabolic regulation of gene expression through histone acylations. Nature reviews. Molecular cell biology18(2), 90-101. 2. Huang, H., Zhang, D., Wang, Y., Perez-Neut, M., Han, Z., Zheng, Y. G., Hao, Q., … Zhao, Y. (2018). Lysine benzoylation is a histone mark regulated by SIRT2. Nature communications9(1), 3374. doi:10.1038/s41467-018-05567-w

Tips and Tricks (if submitted):

Session: Poster, poster number: 110

Response to variables for mice on caloric restriction experiments

Elizabeth Wuerch; Lavender H Lin; Richard Carson; Nathan Zuniga; Eston Dunn; John Price
Brigham Young University, Provo, UT

Caloric restriction experiments show that dieting leads to more stable and functionally folded proteins. The purpose of this study is to determine the effects other variables have on caloric restricted mice. Mice are divided into two groups. Group 1 are put on a diet, limiting their food intake. Each day they are fed at a specific time with a specific amount of food. Group 2, the control group, have continuous access to food. Both groups are evaluated for: weight change, movement/activity level, O2consumption, and quantitative proteomics. Movement/activity levels are measured through the use of a laser beam system. The laser system attaches to the outside of the cage and sends beams through the cage. Movement is recorded each time the light beam is broken inside the cage. The mice’s weight is recorded routinely every seven days. O2consumption is measured through blood gas samples. A mass spectrometer gathered quantitative proteomics data. The determining variables that are analyzed include: mice strain, company purchased from, and the length of time on the caloric restriction diet prior to proteomic analyzation. Data gathered would demonstrate the effects different variables have on the weight change, movement/activity levels, O2consumption, and quantitative proteomics. 


Tips and Tricks (if submitted):

Session: Poster, poster number: 112-TL

Acetylation of the lamina promotes the integrity of the nuclear periphery and inhibits virus production

Laura Murray; Xinlei Sheng; Ileana Cristea
Princeton University, Princeton, NJ

The regulation of the nuclear shape, composition, and function is at the core of numerous types of viral infections. For example, HIV induces perforations in the nuclear lamina, while herpesviruses induce nuclear swelling. One of the most striking alterations of nuclear shape occurs during infection with human cytomegalovirus (HCMV), a widely-spread pathogen that infects 50-90% of the global population and for which there are no effective treatments or vaccine. As HCMV is a nuclear-replicating virus, its capsids are assembled within the nucleus. However, as the viral capsid is too large to exit via nuclear pores, the lamina at the nuclear periphery must be disrupted, forming infoldings that allow capsid budding through the nuclear membrane. Additionally, the nucleus acquires a kidney-like shape to facilitate the formation of an assembly complex juxtaposed to the nucleus where the viral capsid undergoes maturation before cellular egress. Here, we report our discovery that acetylation is a critical regulator of viral capsid egress. Using acetyl-peptide enrichment and quantitative mass spectrometry, we find acetylations on lamina proteins that are temporally regulated during the progression of infection. Among these, K134 acetylation on lamin B1 (LMNB1) displayed one of the largest increases late in infection. To determine the function of acetylation, we generated acetyl-mimic (K-to-Q) and charge-mimic (K-to-R) LMNB1 mutants, and found that acetylation induces reductions in both infectious extracellular and intracellular virus. Live cell microscopy showed that LMNB1 acetylation induced the accumulation of viral capsids within the nucleus. Moreover, cells expressing LMNB1 acetyl mimic had reduced nuclear infoldings and curvature. Altogether, we demonstrate that LMNB1 K134 acetylation acts in host defense by stabilizing the nuclear periphery and inhibiting viral capsid nuclear egress. As K134 is conserved among mammalian species, we predict this acetylation can serve as a general regulator of nuclear morphology under different biological conditions.

Tips and Tricks (if submitted):

Session: Poster, poster number: 113

Global PTM analysis to study cross-talk between lysine acetylome and tyrosine phosphoproteome in EGFR TKI resistant human lung adenocarcinoma

Yue Qi; Tapan Maity; Xu Zhang; Udanya Guha
Thoracic and GI Malignancies Branch, CCR, NCI, NIH, Bethesda, MD

Osimertinib, a 3rd generation epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor (TKI) has been approved to treat lung cancer patients harboring EGFRT790M; unfortunately, all patients eventually develop acquired resistance. Here, we employed a mass spectrometry (MS)-based proteomic approach to identify and quantify post-translational modifications (PTMs) to understand the crosstalk between tyrosine phosphorylation (pY) and lysine acetylation (lysAc) in osimertinib-sensitive lung adenocarcinoma cell line, PC-9, that harbor the EGFRDel746-750 mutant, and its isogenic resistant counterpart, PC-9-OsiR-NCI-1. We used stable isotope labeling in cell culture (SILAC) and dual PTM immunoprecipitation to quantify the pY and lysAc sites. The enriched PTM peptides were analyzed with data dependent acquisition on a tandem MS (Q-Exactive-HF). The raw file library search was performed by MaxQuant and PEAKS.  In total, we quantified 1064 pY and 475 lysAc sites, of which 191 pY sites were up-regulated and 53 sites were down-regulated in the resistant cells; 43 lysAc sites were upregulated and 7 sites were down-regulated in the resistant cells compared to the sensitive cells. Interestingly, we discovered that EGFR phosphorylation is down-regulated at Y1197, Y1172, and Y772 residues in the resistant cells indicating that EGFR signaling may be dampened in osimertinib-resistant cells. LysAc of the tumor suppressor protein p53 (TP53) significantly decreased at K120 postition in the resistant cells. Interestingly, 54 proteins were modified both at pY and lysAc sites. For example, lysine acetylation and phosphorylation at cyclin-dependent kinase 1 (CDK1)-K33 lysAc and pY15 sites were significantly reduced but pY19 phosphorylation was marginally increased in the resistant cells. We hypothesize that the lysAc and pY of CDK1 may be co-regulated to modulate CDK1 function in osimertinib acquired resistance. CDK1 may be a potential drug target to overcome resistance. Further experiments are underway to validate these findings and assess the efficacy of CDK1 inhibition in overcoming osimertinib resistance.

Tips and Tricks (if submitted):

Session: Poster, poster number: 113-ML

Tissue-Specific Protein Sulfhydrome Analysis in Mice as a Function of Age, Sex, and Diet

Nazmin Bithi; Belinda Willard; Christopher Hine
Cleveland Clinic Lerner Research Institute, Cleveland, OH

Endogenously produced hydrogen sulfide (H2S) is a redox reactive metabolite that serves as a vasorelaxant and signaling molecule. One potential mechanism of action for endogenous H2S is through posttranslational modifications of protein thiols, termed sulfhydration. Sulfhydration, similar to other posttranslational modifications, alters a protein’s structure, function, localization, stability, and/or interactions with other macromolecules. Despite the importance of sulfhydrated proteins in various physiological processes, global sulfhydrome profiles in multiple mammalian tissues as a function of age, sex, and diet, which are known modulators of endogenous H2S levels and production, are yet to be determined. We first hypothesized that caloric restriction (CR), one of the best studied interventions to extend lifespan, improve stress resistance, and boost endogenous H2S production in a number of tissues, increases and/or diversifies tissue-specific sulfhydrome profiles. Male C57BL/6J mice were fed ad libitum or 50% CR for one week prior to tissue harvest followed by a combinatorial sulfhydration proteomics approach utilizing an optimized biotin-thiol pulldown assay with LC-MS/MS quantitative proteomics and bioinformatics. We identified 1006, 1107, 474, 450, 905, and 165 individual sulfhydrated proteins in the liver, kidney, heart, muscle, brain and serum, respectively. 50% CR enriched for sulfhydrated proteins in the liver, kidney, muscle, brain and serum, but not in the heart. Pathway enrichment analysis indicated CR promoted sulfhydration of proteins involved in redox homeostasis, metabolism, and biosynthesis pathways. In addition to diet, we determined age and sex act as important biological variables impacting the global sulfhydrome. Aged mice had increased numbers, but altered pathway enrichment, of sulhydrated proteins in the brain while female mice had similar numbers, but different pathway enrichments, of sulhydrated proteins compared to male mice.  In summary, we identified tissue-specific sulfhydrated proteins and the impact of diet, age, and sex on their individual abundance with implications in multiple biological pathways.

Tips and Tricks (if submitted):

Session: Poster, poster number: 114-TL

Systems-Level Identification of PKA-Independent Vasopressin Signaling in Renal Epithelial Cells

Arnab Datta; Chin-rang Yang; Raghuram Viswanathan; Mark A. Knepper
NHLBI, NIH, Bethesda,

Vasopressin signaling in the renal collecting duct is believed to be mediated predominantly by the activation of protein kinase A (PKA). The recent generation of mouse cell lines (“PKA-null” cells), derived from mpkCCD cells, in which both PKA catalytic subunit genes have been deleted via CRISPR-Cas9 affords us an opportunity to identify PKA-independent signaling. Here, we used SILAC coupled protein mass spectrometry to profile proteome-wide phosphorylation changes in response to vasopressin (dDAVP, 0.1 nM, 30 min) in PKA-null cells and in control mpkCCD (PKA-intact) cells. Experiments were done in three different PKA-null lines and three different PKA-intact cell lines.

Overall, >14,800 unique phosphopeptides were quantified in both PKA-null and -intact cells. In PKA-intact cells, 452 distinct phosphorylation sites were altered in abundance in response to dDAVP. In contrast, in PKA-null cells, only 61 phosphorylation sites were altered in abundance, indicating that a large component of the vasopressin response is PKA-dependent. Notably, phosphorylation of Ser256 of AQP2 (thought to be a PKA site) increased in response to dDAVP in PKA-null cells, indicating that other kinases can phosphorylate this site. The upregulated phosphorylation sites in PKA-null cells mapped to an R-X-N-(S/T)*-X motif, consistent with activation of other basophilic kinases. An expanded list of vasopressin-regulated phosphorylation sites were obtained as a byproduct of this approach. Interestingly, cAMP measurements showed that baseline cAMP levels in PKA-null cells were ~10-fold higher than in PKA-intact cells, although both PKA-null and -intact cells showed significant increases in cAMP levels in response to dDAVP. This indicates that there is likely to be a PKA-dependent feedback on some component of the Avpr2-Gαs-Adcy6 signaling pathway responsible for cAMP generation.

Tips and Tricks (if submitted):

Session: Poster, poster number: 115

Quantitative Protein Expression Biomarker Feasibility in Advanced Ovarian Cancer

Punit Kaur1; Alexzander Asea1; Luisa Manning2; Ernest Bognar2; Heidi Zupanc2; Gladice Wallraven2; Khalil Choucair1; Lance Dworkin1; John Nemunaitis1
1University of Toledo, Toledo, OH; 2Gradalis, Inc., Carrollton, TX

This study quantitates the differentially expressed proteins from tumor tissue derived from ovarian cancer patients treated with Vigil, an expressive bi-shRNA-furin and GM-CSF DNA transfected autologous tumor vaccine. Briefly, tumor tissue was removed and transfected with Vigil. Pre- and post-Vigil transfected samples were lysed and equal concentrations of protein were trypsin digested, and run on a Synapt G2-Si high definition mass spectrometer. Protein quantitation statistics was run using Progenesis-QIP software and IPA to provide more detailed information regarding the pathways in which the target genes were involved. Proteomic analysis by mass spectrometry of two blinded samples split from the same patient (#046) pre- and post-Vigil transfection revealed that furin and downstream signal pathway protein TGFβ1 and TGFβ2 were significantly downregulated post-bi-shRNA-furin knockdown DNA transfection. These results are consistent with reduced TGFβ1 (32 pg/10^6 cells to below detection) from prior ELISA of same sample. Additional samples were assayed from patient #046 who had a poor response to Vigil (overall survival (OS) < 1 year) and compared to patient #1017 who had a good response to Vigil (OS > 4 years). Proteins (109) were identified as significantly downregulated and 161 as significantly upregulated between the two patients. Two relevant protein signals, endoglin (167-fold) and retinoid x receptor (RXR) gamma (15600-fold) were identified as dramatically upregulated in the good response patient. Interestingly, patient #1017 also had much more activity by ELISA of knockdown related to TGFβ1 (TGFβ1 pre-transfection 709 pg/10^6 cells to post 0 pg/10^6 cells). These studies demonstrate the feasibility of quantifying differentially expressed proteins from tumor tissue derived from ovarian cancer patients treated with Vigil, and support the applicability of proteomics towards new biomarker discovery.

Tips and Tricks (if submitted):

Session: Poster, poster number: 116-TL

Comparing personalized profiles of host-expressed proteins and microbes in human stool reveals complementary inter-subject distinction.

Ellen Casavant; Les Dethlefsen; Kris Sankaran; Daniel Sprockett; Susan Holmes; David Relman; Joshua Elias
Stanford University, Stanford,

Measuring host stool proteins through noninvasive stool-based assays opens new avenues for characterizing states of gastrointestinal health. However, the extent to which these proteins vary over time and between healthy subjects is poorly characterized. Here, we investigate how host stool protein measurements quantitatively and qualitatively depend on technical and biological sources of variability when measured by mass spectrometry.  We further compare these measurements to parallel gut microbiota surveys. Examining the proteomes and microbiomes measured from five self-reported healthy individuals –each sampled eight times over four weeks, let us investigate four sources of measurement variance.  We find that mass spectrometry-based identification and label-free quantification of stool proteins was consistent over technical and preparative replicates, supporting this assay’s utility for biomarker measurement. Although several human stool proteins tended to vary significantly over time within individuals, we observed far more protein variation between subjects. Both proteome and microbiome data sets consistently distinguish different subjects. Host-proteins measured from stool unambiguously point to clear innate and adaptive immune response mechanisms which varied over time and between subjects. Inter-subject differences were further observed for canonical IBD biomarkers such as fecal calprotectin (subunits S100A8 and S100A9). We conclude that host-centric protein stool measurements indicate a surprisingly wide range of “healthy” protein variation, adding a valuable functional complement to high throughput microbiota surveys. For example, one subject’s proteins support a chronic sub-clinical inflammation state – an observation which would have been difficult to make solely with microbe-derived data.

Tips and Tricks (if submitted):

Proteomics data is heavily enriched in zero values (also known as a zero-inflated mixture). I have developed a statistical approach (trick) to assess a subset of proteins for whether they are significantly variable by identification and abundance in multiple conditions (ex: technical replicates, preparative replicates, biological replicates).

Session: Poster, poster number: 117

Development of mitoTurboID to Unveil Components of the Mitochondrial Signaling Hub Following Innate Immune Activation

Clinton Bradfield

Macrophage innate‑immune signaling utilizes receptor‑ligand interactions to incite antimicrobial defenses, rewire transcription, and promote activation of the adaptive immune system. In the case of the classical toll-like receptor 4 (TLR4) response to lipopolysaccharide (LPS), distinct spatial utilization of signaling adaptors Myd88 and Trif leads to discrete transcriptional responses. This differential utilization of plasma-membrane and endosomal TLR signaling highlights the need to understand organelles as dynamic signaling hubs. While it is well appreciated that many transcriptional responses following TLR stimulation rely on alterations of mitochondrial physiology, namely through the glycolytic shunt, mechanistic understanding of how these changes occur and how they promote altered transcriptional responses remain understudied. Herein we utilized a new proximity‑labeling system (mito‑TurboID) expressed on the outer membrane of the mitochondria to dynamically assess proteins associated with the mitochondrial signaling scaffold following immune stimulation. 

Tips and Tricks (if submitted):

Session: Poster, poster number: 118

Protein ADP-ribosylation and its impact on protein complex formation and release in the LPS activated human macrophage

Casey Daniels1; Pauline Kaplan1; Clinton Bradfield2; Iain Fraser2; Aleksandra Nita-Lazar1
1CNP, NIAID, NIH, Bethesda, N/A; 2SSS, LISB, NIAID, NIH, Rockville, N/A

Protein complex formation and release regulates many cellular processes, including the immune response to lipopolysaccharide (LPS). Protein ADP-ribosylation, a polymeric post-translational modification (PTM), is known to serve as a scaffold for complexes and the quick formation and release of this polymer by Poly(ADP-ribose) Polymerases (PARPs) and Glycohydrolases (PARGs) is an understudied mechanism for protein complex regulation. Here we have separated protein complexes from their monomeric subunits by Size Exclusion Chromatography (SEC), using mass spectrometry to detect proteins as they co-elute with smaller and larger macromolecules. Time course application of this method to human macrophages stimulated with LPS has revealed several PARP proteins moving between fractions as macrophages are activated, as well as proteins whose movement into or out of complexes upon LPS stimulation is reversed by PARP inhibition. Use of our powerful, in-house ADP-ribose site identification method has allowed for detection of ADP-ribosylation events during macrophage LPS activation, and detection of PTMs which vary between monomeric and complexed proteoforms. This work sheds light on an understudied aspect of both LPS signaling and PAR proteomics. This research was supported by the Intramural Research Program of NIAID, NIH.

Tips and Tricks (if submitted):

Session: Poster, poster number: 119-ML

Composition of the myddosome during the innate immune response

Joseph Gillen1; Aleksandra Nita-Lazar2
1NIH-NIAID, Bethesda, MD; 2NIAID, NIH, Bethesda, MD

Critical for the innate immune response to PAMPs, including viral RNA, the myddosome is a complex of proteins characterized by the presence of myeloid differentiation factor 88 (Myd88).   The myddosome acts to transfer signals from the toll-like receptor (TLR) proteins to TNF receptor-associated factor 6 (TRAF6).  In addition to Myd88, the myddosome contains multiple copies of the interleukin-1 receptor associated kinases (IRAK) 2 and 4 and is theorized to form following stimulation of the TLR proteins.  Using affinity purification – mass spectrometry analysis (AP-MS), we identified the Myd88-associated proteins in mouse immortalized bone marrow-derived macrophages before and after lipopolysaccharide (LPS) treatment.  Prior to LPS treatment, we found the stable association of Myd88 to Tripartite motif-containing protein 21 (TRIM21) and other inhibitory factors.  Following LPS treatment, the inhibitory factors were lost and Myd88 associated to the downstream effector proteins, including IRAK1, IRAK4, and TRAF6.   Using stable isotope labeling by amino acids in cell culture (SILAC), we found that the association of Myd88 to the IRAK proteins changes as the LPS response progresses.  Because of this, we are examining the mechanism(s) that regulate the activity of the myddosome.  We are currently using AP-MS followed by phosphopeptide enrichment with MS/MS analysis to identify possible phosphorylation sites.  Using these data, we will evaluate the role of phosphorylation on myddosome activity.  This research was supported by the Intramural Research Program of the NIH, National Institute of Allergy and Infectious Diseases.

Tips and Tricks (if submitted):

Session: Poster, poster number: 120

Identification of Dynamic Nuclear Receptor Co-Regulatory Protein Complexes by Reverse Phase Protein Array and Mass Spectrometry

Shixia Huang1; Kalpana D Acharya2; Sandy L. Grimm1; Sung Yun Jung1; Celetta G Callaway1; Kimal Pajapakshe1; Charles E Foulds1; Cristian Coarfa1; Marc J Tetel2; Dean P Edwards1
1Baylor College of Medicine, Houston, TX; 2Wellesley College, Wellesley, MA

Nuclear hormone receptors (NHR) are a family of ligand-dependent transcription factors regulating expression of important gene networks.  NHRs bind specific DNA sequences and recruit transcriptional co-regulators (CoRs) that are a dynamic complex of proteins .

In this study we combined mass spectrometry (MS) and reverse phase protein array (RPPA) proteomic platforms to identify the changes in composition of CoRs complexes associated with progesterone receptor (PR) in response to physiological stimuli.  Nuclear proteins from mouse brain tissues or human cancer cells were incubated respectively with mouse or human PR (A or B isoforms) bound to bead-immobilized DNA harboring hormone response elements.  Assembled DNA-bound PR-CoR complexes were treated with progesterone (P4) or ATP, eluted and analyzed by MS profiling and RPPA, the later with > 200 validated antibodies to various protein signaling pathways including a group of known CoRs.

In mouse brain tissues, a total of 50 proteins were identified by MS and RPPA in isolated PR-CoR complexes in ligand dependent manner. Among them, 33 were identified by MS and 17 by RPPA. The majority of proteins detected by RPPA were not detected by MS including the well known steroid receptor coactivators 1,2 (SRC-1, -2), regulatory proteins involved in signal transduction and metabolism and phosphorylated forms of MAPK1 and src tyrosine kinase.

In isolated human PR-CoR complexes from cancer cells, a total of 45 proteins were detected by MS and RPPA.  Since the RPPA platform contains > 70 antibodies to phosphorylated forms of proteins, it further identified recruitment of 19 phosphorylated proteins to the PR-CoR complex by ATP treatment that was not possible by MS profiling. 

Conclusion:  Combining MS profiling and targeted RPPA enabled a more comprehensive identification of known and novel protein interactions indicating that these are complementary proteomic platforms for analysis of the dynamics of cellular protein complexes and interactomes.

Tips and Tricks (if submitted):

Session: Poster, poster number: 121

Determination of Host Cell Proteins In Antibody Preparations Using PASEF on the timsTOF PRO

Stuart Pengelley1; Scarlet Koch1; Christopher Swift2; Guillaume Tremintin3; Christopher Adams3; Detlev Suckau1; Jochen Boosfeld1; Gary Kruppa2; Michael Greig3
1Bruker Daltonic, Bremen , Germany; 2Bruker Daltonic, Billerica, MA; 3Bruker Daltonic, San Jose, CA

The analysis of host cell proteins (HCP) at the ppm level is critical. ELISA is currently the gold standard for QC applications, but the advantages of mass spectrometry are abundant. We show how PASEF (parallel accumulation and serial fragmentation), as implemented on the timsTOF PRO QTOF, can be applied to HCP analysis for highly sensitive detection with enhanced data quality.

NISTmAb Reference Material 8671 and the Universal Proteomics Standard (UPS1, Sigma) were reduced, and digested. Peptides were separated on an Intensity Solo 2 1.8µm C18 column using an Elute UHPLC coupled to a timsTOF Pro ion mobility QTOF mass spectrometer (all Bruker Daltonics). A 150 minute gradient was used. For nanospray, a nanoElute UHPLC was fitted with an IonOpticks C18 column using a 210 minute gradient.

The UPS1 standard was used to make a 5-step 1:3 dilution series in a constant background of NIST mAb over a concentration range from 0.3 to 934 ppm. PASEF enabled the detection of UPS1 proteins down to low single digit ppm concentrations in the presence of 25 µg NIST mAb. Linear response for the UPS1 protein beta-2-microglobulin in the concentration range from 132 ppm to 1.6 ppm was observed. The dilution series indicated that PASEF enabled detection of HCPs in the range of 1 to 100 ppm of the therapeutic protein. Nano UHPLC was also evaluated for its suitability for HCP identification. This setup has already been established as the new benchmark for bottom-up proteomics applications [1, 2]. More than 200 HCPs were identified in 1.5 µg NIST mAb, including expected and previously unreported HCP proteins. The depth of HCP coverage and sequence coverage was further extended to >280 HCP identifications by employing an alternate digestion method in which only the HCPs are digested allowing the mAb to be removed prior to analysis.

Tips and Tricks (if submitted):

Session: Poster, poster number: 122

Multiplexed interactomics reveals coordination of proteostasis network remodeling and mechanisms of protein quality control

Madison T. Wright; Lars Plate
Vanderbilt University, Nashville, TN

Proteins are folded and transported to the appropriate subcellular locations with the guidance of various protein folding and trafficking factors, which together comprise the proteostasis network. Coordinated activity of these factors governs protein quality decisions ensuring correct folding and localization of properly matured client proteins while degrading unstable, misfolded proteins. Imbalances in proteostasis activity resulting from genetic defects, stress or aging are associated with etiologically-diverse gain- and loss-of-function protein misfolding diseases, due to decreased protein quality control. Consequently, remodeling of the proteostasis network has emerged as a promising therapeutic strategy to correct the underlying quality control defects for diverse proteins associated with such disorders. However, it often remains difficult to define the underlying molecular mechanisms responsible for the protein quality control defects that lead to improper partitioning of destabilized, disease-associated proteins between folding, trafficking and degradation. Protein quality control is dictated by transient interactions between client proteins and individual proteostasis network components and mass spectrometry-based interactomics can globally profile changes among these interactions. Here, we discuss integration of tandem mass tag (TMT)-based multiplexed quantitative interactomics as a method to further elucidate dynamic proteostasis network processing in order to define altered protein recognition associated with destabilized protein variants responsible for aggregation diseases. Furthermore, we determine how remodeling of the endoplasmic reticulum (ER) proteostasis network can alter client protein engagement to influence trafficking and degradation decisions. Our results provide mechanistic insight how remodeling of the ER proteostasis network can improve quality control decisions for aggregation-prone protein variants, specifically clients associated with amyloid or loss-of-function protein misfolding diseases. The work highlights the utility of multiplexed quantitative interactomics to define mechanisms of aberrant protein quality control that can be exploited for therapeutic intervention in protein misfolding diseases.

Tips and Tricks (if submitted):

Session: Poster, poster number: 123

Influence of Post-Translational Modifications on Protein Stability


Nathan Zuniga; Lavender Lin; John Price
Brigham Young University, Provo, UT

Recent studies demonstrate that imbalances in the quality and concentration of most proteins in each cell (proteostasis) are at the core of many of today’s devastating diseases. The causes in these imbalances are poorly understood. Thus, it is important to create diagnostic methods that can detect changes in the quality and concentration of many proteins within a patient’s proteome. We recently identified changes in structure and post translational modifications of abundant proteins in the blood serum samples of rheumatoid arthritis (RA) patients relative to healthy controls. These results suggest that either protein modification or changes in secondary structure could influence disease etiology.

Our long-term goal involves understanding the role of disease specific post-translational modifications on protein turnover and protein stability. We observed differences in protein denaturation curves between serum samples of RA and otherwise “healthy” patients. Subsequent proteomics studies confirmed a change in secondary structure of abundant blood proteins.  Based on this preliminary data, we hypothesize that modifications of protein folding are related to disease.  Our first objective focuses on the use of chemical probes to monitor protein unfolding energies across the proteome by modifying surface exposed AA sites during denaturation.  We are optimizing use of tyrosine, tryptophan, and methionine. These modifications are introduced at different degrees of protein unfolding, and measured with mass spectrometry.   Use of these three probes should result in the ability to monitor 98% of the proteome.  This will allow us to test how protein homeostasis changes with RA development. This understanding will also lead to greater knowledge of the control of protein homeostasis during development and aging.


Tips and Tricks (if submitted):

Session: Poster, poster number: 124-TL

Quantitative Interactomics to Determine Protein Quality Control Mechanisms Dictating Thyroglobulin Secretion

Madison Wright; Lars Plate
Vanderbilt University, Nashville, TN

The proteostasis network (PN) is composed of folding, trafficking, and degradation factors that facilitate protein quality control (PQC). Perturbations to the PN caused by environmental, age-related, or genetic alterations are linked to a variety of disease states, including protein aggregation and amyloid diseases that are often difficult to treat. These detrimental states are a result of diminished PQC capacity, leading to aberrant processing of client proteins during biogenesis and downstream handling. Modulation of PN components or entire pathways has shown promise as a potential therapeutic strategy to ameliorate such protein misfolding diseases. In the current study, we utilize thyroglobulin (Tg), the 330 kDa thyroid prohormone, as a model to elucidate mechanisms of PQC directed by PN pathways. Tg maturation relies heavily on PN components for proper domain assembly, trafficking, and ultimately secretion due to its large size and extensive requirements for post-translational modification (PTM). Once folded, Tg is secreted and further processed to create T3 and T4 hormones that are critical for fetal development and utilized in adulthood to regulate intermediary metabolism. Many patient-derived Tg mutations associated with congenital hypothyroidism lead to aberrant processing, resulting in decreased or complete loss of Tg secretion and subsequent hormone production. We have implemented a multiplexed quantitative affinity purification – mass spectrometry (AP-MS) method to delineate the wild-type Tg interactome from pathological variants in order to identify PN dynamics facilitating aberrant processing of these clinical mutations. An improved understanding of the PQC defects for Tg variants provides new therapeutic strategies directed at recovering Tg secretion and hormone production, and may be applicable to other protein misfolding diseases.

Tips and Tricks (if submitted):

Session: Poster, poster number: 125

Reference Materials for Proteomic Investigations

Ashley Beasley-Green; Lisa Kilpatrick; Mark Lowenthal; Karen Phinney; David Bunk
NIST, Gaithersburg, MD

High-throughput, bottom-up proteomic experiments typically involve complex experimental designs, both for sample preparation and for instrumental analysis.  The complex workflows frequently used in proteomics are difficult to optimize and trouble-shoot, impacting the quality of the proteomic investigative results.  NIST has developed two reference materials which were designed to support the quality of proteomic results:  RM 8323 Yeast Protein Extract and RM 8321 Peptide Mixture for Proteomics. 

RM 8323 (Yeast Protein Extract) is a frozen aqueous solution of extracted Saccharomyces cerevisiae yeast proteins. The S. cerevisiae yeast proteome is an attractive material for proteomic reference materials for several reasons including its proteome complexity and availability.  RM 8323 can be utilized in the design and optimization of proteomics-based methodologies from sample preparation to data analysis.

In contrast to RM 8323, RM 8321 (Peptide Mixture for Proteomics) is composed from approximately 440 synthetic peptides in a frozen aqueous solution.  Selection of the constituent peptides was based on several factors including, their identification as “proteotypic” peptides in several archived MS/MS databases and chromatographic retention.  The reference material was formulated with several tiered peptide concentrations in order to evaluate a LC-MS/MS platform’s performance for a range of peptide concentrations in a single analysis.  Both discovery-based and targeted LC-MS/MS methods were used to identify the peptides present and a confidence-ranking system was developed to combine results of peptide identities from these methods.

We provide examples of how NIST RM 8323 and RM 8321 can be used as a quality control material for method optimization in common proteomics-based workflows to improve measurement quality.  These reference materials were used to develop digestion metrics, to optimize LC-MS/MS parameters, and to evaluate the search engines used in discovery-based proteomics.

Tips and Tricks (if submitted):

Session: Poster, poster number: 126

SASP Atlas: A Comprehensive and Unbiased Proteomic Database of the Senescence Associated Secretory Phenotype

Nathan Basisty1; Abhijit Kale1; Okhee Jeon1; Chisaka Kuehnemann1; Therese Payne1; Chirag Rao1; Anja Holtz1; Samah Shah1; Luigi Ferrucci3; Judith Campisi1, 2; Birgit Schilling1
1The Buck Institute, Novato, CA; 2Lawrence Berkeley Laboratory, Berkeley, CA; 3National Institute on Aging, Baltimore, MD

Introduction: The senescence-associated secretory phenotype (SASP) has recently emerged as both a driver of -- and promising therapeutic target for -- a multitude of chronic age-related conditions, ranging from neurodegeneration to cancer.  The complexity of the SASP has been greatly underappreciated and a small set of factors cannot explain the diverse phenotypes it produces in vivo.  Here, we present ‘SASP Atlas’, a comprehensive proteomic database of SASPs, including a novel exosome SASP phenotype, driven by multiple inducers of senescence in different human cell types. We also propose that SASP proteins are promising biomarkers to assess senescent cell burden in aging and disease.

Methods: Secretomes of senescent cells were characterized by comparing the secreted soluble (sSASP) and exosome (eSASP) proteins of irradiation-, oncogenic RAS-, or HIV drug atazanavir-induced senescent human fibroblasts to non-senescent controls. Secreted proteins and exosomes were obtained from the medium of cells cultured for 24 hours in serum-free conditions. Data-independent acquisitions (DIA) were performed on a TripleTOF 6600.

Results: While 172 proteins increased in all inducers and 53 proteins increased in all inducers and cell types, the SASP of each inducer and cell type were largely distinct, totaling over 1000 unique proteins. CXCL1, MMP1, and STC1 were consistently among the top increased proteins in response to all inducers. Senescent cells, on average, secreted larger exosomes containing a highly distinct set of proteins compared with the soluble secretome, including proteins involved in G-protein and RAS signaling, prostaglandin regulation and the complement system.

Discussion: Together, the data demonstrate that the SASP is a complex and highly diverse set of secretory phenotypes, including a unique exosome SASP. This resource will aid in identifying the proteins that drive senescence-associated phenotypes and provide comprehensive catalogs of potential biomarkers that can be used to assess the burden and origin of senescent cells in vivo. 

Tips and Tricks (if submitted):

Session: Poster, poster number: 127

Physical Activity Associated Proteomics of Skeletal Muscle: Being Physically Active in Daily Life Protect Skeletal Muscle from Aging

Ceereena Ubaida-Mohien; Marta Gonzalez-Freire; Alexey Lyashkov; Ruin Moaddel; Chee Chia; Eleanor Simonsick; Ranjan Sen; Luigi Ferrucci
NIA/NIH, Baltimore, Maryland


Muscle strength declines with aging and exercise is the only intervention known to attenuate this decline. In order to adequately investigate both preventive and therapeutic interventions against sarcopenia, a better understanding of the biological changes that are induced by exercise in aging skeletal muscle is required.


In an effort to determine the effect of physical activity on the skeletal muscle proteome, we utilized liquid-chromatography mass spectrometry to collect quantitative tandem mass tag proteomics data on human skeletal muscle biopsies obtained from 60 well-characterized healthy individuals (20-87 years) who reported heterogeneous levels of physical activity (not active, active, moderately active, and highly active).


Over 4000 proteins were analyzed, and higher self-reported physical activity was associated with substantial overrepresentation of proteins associated with mitochondrial, TCA cycle, structural and contractile muscle, and genome maintenance. Conversely, proteins related to the spliceosome, transcription regulation, immune function, apoptosis, DNA damage, and senescence were underrepresented with higher self-reported activity. These differences in observed protein expression were related to different levels of physical activity in daily life and not intense competitive exercise. In most instances, differences in protein levels were directly opposite to those observed with aging. These data suggest that being physically active in daily life has strong beneficial effects on muscle.

Tips and Tricks (if submitted):

Session: Poster, poster number: 128-TL

Multifaceted proteomic evaluation of lysosome dysfunction in neurodegeneration via human iPSC-derived neurons

Ling Hao1; Ryan Prestil1, 2; Michael Fernandopulle1, 2; Stewart Humble1; Daniel Lee3; Saadia Hasan1; Maia Parsadanian1; Richard Youle1; Michael Ward1
1National Inst of Neurological Disorders and Stroke, Bethesda, MD; 2University of Cambridge, Cambridge, UK; 3Cornell University, Ithaca, NY

Lysosome dysfunctions have been associated with neurodegeneration, but the molecular mechanisms remain unclear. Here, we developed novel proteomic approaches in iPSC-derived neurons to study the molecular correlations between lysosomal biology and neurodegeneration. 1) To characterize protein turnover in neurons, I developed a dynamic SILAC labeling method to calculate protein half-lives in iPSC-neurons. 2) To characterize lysosome interactions with other organelle and cytosolic proteins, I developed a proximity labeling method by stably expressing Lamp1-APEX tag to biotinylate proteins within ~20 nm of lysosomes in living neurons. 3) We then applied these two methods to evaluate the changes of lysosome functions when knocking out genes that cause neurodegenerative diseases. The first gene we studied is GRN, whose mutations lead to the deficiency of progranulin, a lysosomal glycoproteins, and cause frontotemporal dementia. Loss of PGRN impairs lysosomal activity, but the precise function of PGRN remains unknown. 

By dynamic SILAC proteomics,  I found that the neuronal protein degradation curve follows the 1st order kinetics. I developed a multi-time point and a single time point methods to calculate protein half-lives, both yielded reproducible and consistent results. Neuronal proteins showed longer half-lives compared to other cell types. Histones and nucleoporins have super long half-lives (<300 hrs) in neurons, and proteins related to neuron projection and ubiquitin proteasome pathways have much shorter half-lives. Lamp-Apex proteomics captured the transient lysosomal interacting proteins in living neurons, and 1548 out of 4413 proteins were truly interacted with lysosome after comparing to the cytosolic-Apex as a spatial control. Knocking out GRN resulted dramatic changes in protein half-lives and lysosomal interacting proteins in neurons, which are related to the regulation of hydrolase activity and vesicle-mediated transport (crucial lysosomal functions). PGRN deficiency also reduced recruitment of multiple annexins to the lysosome, in particular annexin A11 whose mutations cause autosomal dominant neurodegeneration. 

Tips and Tricks (if submitted):

Session: Poster, poster number: 129

Effects of Oxidized Lipoprotein (oxLDL) on the Proteome in Retinal Pigment Epithelial Cells

Sarka Beranova-Giorgianni; Francesco Giorgianni
University of Tennessee Health Science Center, Memphis, TN

Retinal pigment epithelium (RPE) are specialized support cells essential for vision. Loss of these cells contributes to development of age-related macular degeneration (AMD). Oxidative stress and ensuing molecular and cellular damage are major contributors to the pathological mechanisms of AMD. Oxidatively modified low-density lipoprotein (oxLDL), formed from LDL in a pro-oxidant milieu, induces oxidative stress and is cytotoxic to the RPE. In the sub-RPE space, oxLDL is found in extracellular deposits (drusen), which are a prominent clinical feature of AMD. Information on how the RPE respond to exposure to oxLDL is therefore a critical element in the mechanistic understanding of RPE dysfunction and death in the context of AMD. The purpose of this study was to determine protein expression changes in the RPE proteome induced by exposure of the cells to sub-lethal levels of oxLDL.  ARPE-19 cells were treated with oxLDL (100 µg/mL) for 24 h. Proteins extracted from control and oxLDL-treated cells were digested and analyzed with high-resolution LC-MS/MS on a Synapt G2-Si system in the data-independent acquisition mode (MSE) and using ion mobility. Protein identification and quantification was performed with Progenesis QIP. With this workflow, close to 3,000 RPE proteins were identified.  Treatment with oxLDL resulted in expression changes of 300 proteins (fold change ≥ 1.5, q value < 0.05). Functional analysis of the differentially expressed proteins with DAVID and STRING tools showed that upon oxLDL uptake the RPE cells mount a complex molecular response affecting multiple biological processes. This response involves upregulation of proteins related to autophagy, and of anti-oxidative stress proteins whose expression is controlled by the transcription factor Nrf2. Among the proteins down-regulated by oxLDL treatment are ribosome components and proteins involved in translation initiation, which suggests that the mechanisms through which RPE cells respond to oxLDL exposure include inhibition of protein synthesis.

Tips and Tricks (if submitted):

Session: Poster, poster number: 130

Calorie Restriction Conditions May Modulate Aging Rates by Altering Ribosomal Maintenance and Quality

Richard Carson1; Bradley Naylor2; John Price1
1Brigham Young University, Provo, UT; 2University of Utah, Salt Lake City, UT

One of the hallmarks of aging is decreased proteome quality, leading to increased risk of age-related diseases.  Calorie restriction is the gold standard intervention for increasing lifespan and healthspan in laboratory animals; decades of study suggest that these benefits result from an overall improvement in protein quality and maintenance.  However, the mechanisms by which the proteome control machinery is adjusted to do so remain incompletely understood.  Using cohorts of mice placed on calorie restriction with varying dietary perturbations, we performed an analysis of kinetic and quantitative proteomic data combined with RNA-sequencing data to investigate nodes of proteome control.  Our data show that a general slowdown of protein turnover results from calorie restriction; additionally, the metabolic flux of ribosomal proteins shifts, suggesting significant changes in translational quality and ribosomal maintenance.  Further, the data suggest that this slowdown in protein translation may be accomplished by alterations to the peripheral cellular protein translational machinery.  Improvements in ribosome quality leading to increased translational fidelity, along with secondary effects such as increased time for co-translational protein folding, are the expected positive effects on overall proteome quality due to slower translation rates.  These data provide evidence for a model in which the phenotypic benefits of calorie restriction arise from the cell actively perturbing translational controls to slow down proteome turnover.

Tips and Tricks (if submitted):

Session: Poster, poster number: 132

Characterization of Changes in the Insolublome with Aging and Age-related Diseases

Xueshu Xie; Dipa Bhaumik; Kathleen Dumas; Manish Chamoli; Anja Holtz; Suzanne Angeli; Renuka Sivapatham; Julie Andersen; Gordon Lithgow; Birgit Schilling
Buck Institute for Research on Aging, Novato, CA

We and others have shown that increased protein insolubility is a feature of normal aging. The proteins that enter the insolublome are enriched for lifespan determining functions, but it is not clear how normal aging processes and Alzheimer’s disease (AD) processes are related.  To investigate SDS insoluble protein aggregates, we are using data-dependent acquisitions (DDA) for protein discovery and identification, and data-independent acquisitions (DIA or SWATH) for comprehensive sampling of protein aggregates or the ‘insolublome’. Taking advantage of the powerful genetic tools available in C. elegans models, we have assessed and compared conditions, such as young and old (for N2 and Ju775 strains), but also ‘human A-beta’ expressing worms vs. control. We routinely identify and quantify 500-1000 proteins in the insoluble protein fractions.  Significantly regulated proteins observed and increased in aggregates from old vs young worms were compared to insoluble proteins significantly regulated in ‘human A-beta’ expression models vs control.  Additional time courses throughout aging were assessed monitoring dynamic changes over time in the protein insolublome.

Tips and Tricks (if submitted):

Session: Poster, poster number: 133

Changes in Skeletal Muscle Proteins: A Story of Splicing, Mitochondria, and Immune Function

Ceereena Ubaida-Mohien1; Alexey Lyashkov1; Marta Gonzalez-Freire1; Ravi Tharakan1; , Michelle Shardell1; Ruin Moaddel1; Richard Semba2; Chee Chia1; Ranjan Sen1; Luigi Ferrucci1
1National Institute on Aging, Baltimore, MD; 2Johns Hopkins Medical Institute, Baltimore, MD


Muscle degeneration is a substantial cause of weakness and frailty in older persons, but the cause of such degeneration remains unknown and no previous study assessed the human skeletal muscle proteome over the course of aging. We have examined the proteome of human skeletal muscle utilizing a Tandem Mass Tag (TMT) based protein quantification approach in very healthy persons dispersed over a wide age-range.


Skeletal muscle biopsies were collected from 60 healthy human donors ranging in age from 22 to 87 years. TMT 6plex was used for relative protein quantification. Skeletal muscle proteins were extracted, trypsin digested, reduced and alkylated. MS/MS peaks from the samples were searched and analyzed. For final representation, spectra were normalized by median polish and median sweep, protein identifications were quantified and annotated. Relationships of protein levels with aging were examined by the linear regression model.


Several functional classes of proteins were underrepresented or overrepresented with aging. Proteins related to energetic metabolism were underrepresented in older participants, including those related to the TCA cycle, respiratory and electron transportation chain, mitochondria respiratory complex assembly and glycolysis. Several proteins that play important roles in innate and adaptive immunity were overrepresented with older age. Similarly, proteins of the major spliceosome complexes were also overrepresented with older age and this was paralleled by a global change in alternative splicing. Our analysis indicates that the skeletal muscle proteome undergoes substantial changes with healthy aging, particularly evident in the areas of energy metabolism, inflammation, proteostasis, and splicing.


Our preliminary analysis shows that the skeletal muscle proteome undergoes substantial changes with healthy aging, indicating profound changes in energy metabolism and spliceosome complex.

Tips and Tricks (if submitted):

Session: Poster, poster number: 134-TL

Using protein stability as a matric of protein quality in protein homeostasis

Lavender Hsien-Jung Lin; Nathan Zuniga; Joseph Creery; Marcus Hadfield; John Price
Brigham Young University, Provo, UT

The lifespan of US population has been increased from 45 in 1913 to 77 in 2015. People lives longer, in the same time age-related disease, such as cancer, heart diseases, Parkinson’s disease, and Alzheimer’s disease, also becomes more common. Slowing down the onset of these diseases may be accomplished by understanding the biochemical mechanisms of aging. It has been thought that aging is related to the decline of protein homeostasis (proteostasis), which is the decline of the functional balance in proteome. However, what contributes to the proteostasis decline remains unclear. We are creating a system scale model of proteostasis which includes synthesis, folding, and degradation, together to describe the phenomenon of proteostasis. The model we propose utilizes protein concentration, protein turnover rate and protein stability to understand how protein population and quality has changed overtime and between different condition. Using mouse models we have measured changes in proteostasis during periods where the aging rate is modified.  The literature and our preliminary data suggests that during slow aging, the protein turnover slows down and the protein stability decreases. Here we investigate how this change in proteome scale turnover affects protein quality by comparing changes of protein turnover verses the protein fold stability across multiple protein groups and mouse models.

Tips and Tricks (if submitted):

Session: Poster, poster number: 135

Identifying ApoE isoform-dependent changes for protein turnover in the brain

Joseph Creery; Russell Denton; John Price
Brigham Young University, Provo, UT

It is known that a significant contributor to proteopathy in late-onset Alzheimer’s disease (AD) is the genetic variation of Apolipoprotein E (ApoE). There is evidence to support that relative to isoform 3, isoform type 4 is implicated in increased AD risk, and isoform 2 protects against AD. It is also known that many specific protein-protein interactions contribute, even accelerate disease progression, including the Tau effect and amyloid-creating-peptidase proteins. Although this is known, there is no clear picture of total protein homeostasis and turnover in the presence of the different isoforms of ApoE. Comparing transgenic mice (C57BL/6), expressing one of the three isoforms of human ApoE, we can observe differences in relative concentrations of proteins that do not directly interact with ApoE on a protein-protein level. The differences in protein concentrations of these seemingly unrelated proteins have been observed by others as well as ourselves. We hypothesize that there are systematic responses to the ApoE isoform differences to compensate for a lack of control in the ApoE dominion or subsection of the system. From the combination of protein turnover and concentration data, we can determine whether a change in synthesis or degradation is controlling for the change of protein concentration for each of the proteins in the proteome.

Tips and Tricks (if submitted):

Session: Poster, poster number: 136-ML

Proteomic Approaches to Study Cell Death Mechanisms in Human Stem Cell-Derived Retinal Ganglion Cells

Joseph Mertz; Xitiz Chamling; Ah Young Lee; Xiaoli Chang; Byoung-Kyu Cho; David Clark; Cynthia Berlinicke; Hui Zhang; Donald Zack
Johns Hopkins Medical School, Baltimore, MD

Glaucoma is characterized by progressive loss of retinal ganglion cells (RGCs) and while reasonable therapies exist – primarily to lower intraocular pressure – they are not always effective. To date, there are no FDA-approved neuroprotective agents that directly halt vision loss. Our goal is to increase our understanding of RGC neuroprotective signaling networks during axonal injury using global and phosphoproteomics in human stem cell-derived RGCs (hscRGCs), a highly disease-relevant model. Our primary focus is on the dual leucine zipper kinase (DLK)/leucine zipper kinase (LZK)/c-Jun N-terminal kinase (JNK)/stress activated protein kinase (SAPK) pathway and the phosphorylation events that bring about cell death upon axonal injury.

Human ES cells expressing the murine cell-surface protein THY1.2 under BRN3B promoter control were differentiated into RGCs for 35 days, immunopurified through THY1.2, and cultured as a purified population for 7 days. Axonal injury was induced by the microtubule destabilizer colchicine for 5 minutes to 48 hours with a subset simultaneously subjected to pharmacological DLK inhibition by VX-680. Viability, morphology, and SAPK pathway activation under these conditions were assessed, samples prepared for LC-MS/MS analysis including phosphopeptide enrichment by Fe3+ immobilized metal affinity chromatography (IMAC), and LC-MS/MS performed on Orbitrap instruments.

At 48 hours, 1µM colchicine significantly decreased cell viability (13.1±0.9%vs ctl) and neurite length (6.9±2.1% vs ctl), and engaged SAPK signaling (increased phospho-cJun) after only 30 minutes. DLK inhibition by VX-680 strongly suppressed all of these effects (87.1±2.3% survival, 47.5±6.4% neurite length, and no detectable phospho-cJun). Global and phosphoproteome samples have been prepared and LC-MS/MS analysis is underway. We expect to quantify over 9,000 proteins and 10,000 phosphorylation sites.

Multi-tiered proteomic analysis of signaling networks involved in axon injury-induced cell death in conjunction with pharmacological modulation of these networks will provide insights into key signaling events in RGC death, and hopefully lead to improved therapeutic approaches.

Tips and Tricks (if submitted):

Session: Poster, poster number: 137

Body-wide dynamics of organ proteomes for health and diseases

Bingyun Sun
Simon Fraser University, Burnaby, Canada

Body-wide organs and tissues are closely related and function as an integrated network. Pathological response of one organ often has effects to other organ and tissue types. Some of these relationships are well established such as the dysregulated insulin secretion in pancreas of diabetic patients causes complications in cardiovascular, urinary, and nervous systems. Nevertheless, in many other pathological conditions, multiorgan responses have been overlooked, such as in drug induced toxicity. For instance, in acetaminophen overdose, whole-body histology has been performed in animal models, from which several extraheptic organs besides the liver shown toxic response. Using organ specific proteins, multiorgan responses had been detected in the blood of over dosed mice. We here further demonstrate that by studying the proteome of different organs during the course of the acetaminophen overdose, it is possible to infer the trend of muliorgan response observed at histology and blood proteome level. When superimpose the organ proteome dynamics caused by drug perturbation to the proteome of different in-bred strains, we can also start to reveal the pathological phenotypes of different mouse strains. Our results were summarized from proteomes of upto 13 major organs in four in-bred strains of mice, and some of the organ proteomes had been further characterized for their changes in response to half lethal dose of acetaminophen.

Tips and Tricks (if submitted):

Session: Poster, poster number: 138-WTT

Burkholderia rewires its proteome to lower antibiotics sensitivities and to support biofilm formation

Mohd M. Khan1, 5; Supaksorn Chattagul2; Bao Q. Tran3, 6; Jeffrey A. Freiberg4; Aleksandra Nita-Lazar5; Mark E. Shirtliff4; Rasana W. Sermswan2; Robert K. Ernst4; David R. Goodlett3
1University of Maryland School of Medicine, Baltimore, MD; 2Melioidosis Research Center, Khon Kaen University, Khon Kaen, Thailand; 3University of Maryland School of Pharmacy, Baltimore, MD; 4University of Maryland School of Dentistry, Baltimore, MD; 5LISB, NIAID, National Institutes of Health (NIH), Bethesda, MD; 6U.S. Army Edgewood Chemical Biological Center, Gunpowder, MD

The melioidosis causing opportunistic pathogen Burkholderia pseudomallei, a tier 1 overlap select agent by the Centers for Disease Control and Prevention (CDC), imparts a significant public health burden in Southeast Asia and northern Australia. Unfortunately, Burkholderia antibiotics sensitivities are lowered in biofilm forms while treating drug-resistant strains is still a challenge. Further, lack