Abstract
A.1
PPKs mediate direct signal transfer from phytochrome photoreceptors to transcription factor PIF3
Weimin Ni1, Shou-Ling Xu2, Eduardo González-Grandío1, Robert J. Chalkley3, Andreas F. R. Huhmer4, A. L. Burlingame3, Zhi-Yong Wang2, Peter H. Quail1
1University of California, Berkeley, CA, USA; 2Carnegie Institution, Stanford, CA, USA; 3University of California, San Francisco, CA, USA; 4Thermo Scientific, San Jose, CA, USA
Upon light-induced nuclear translocation, phytochrome (phy) sensory photoreceptors interact with, and induce rapid phosphorylation and consequent ubiquitin-mediated degradation of, transcription factors, called PIFs, thereby regulating target gene expression and plant development. Nevertheless, the biochemical mechanism of phy-induced PIF phosphorylation has remained ill-defined. Here we identify a family of nuclear protein kinases, designated Photoregulatory Protein Kinases (PPK1-4; formerly called MUT9-Like Kinases (MLKs)), that interact with PIF3 and phyB in a light-induced manner in vivo. Genetic analyses demonstrate that the PPKs are collectively necessary for the normal light-induced phosphorylation and degradation of PIF3. PPK1 directly phosphorylates PIF3 in vitro, with a phosphosite pattern that strongly mimics the light-induced pattern in vivo. These data establish that the PPKs are directly involved in catalyzing the photoactivated-phy-induced phosphorylation of PIF3 in vivo, and thereby are critical components of a transcriptionally-centered signaling hub that pleiotropically regulates plant growth and development in response to multiple signaling pathways.
This work was supported by NIH (2R01 GM-047475), DOE (DEFG03-87ER13742), and USDA ARS Current Research Information System (5335-21000-032-00D) to P.H.Q.; by NIH (5R01GM066258) and DOE (DEFG02-08ER15973) to Z.Y.W.; and by NIH (8P41GM103481) to A.L.B.
A.2
Mass Spectrometry-based Proteomics of Human Breast Milk to Identify Potential Breast Cancer Biomarkers
Roshanak Aslebagh1, Devika Channaveerappa1, Kathleen F. Arcaro2, Costel C. Darie1
1Clarkson University, Potsdam, NY, USA; 2University of Massachusetts, Amherst, MA, USA
Early detection of breast cancer, as the second common cancer and second leading cause of cancer death in American women, has been a challenging subject of study especially in young women. Mammography is not effective on the dense breast tissue of young women, and they also face a transient increased risk of pregnancy-associated breast cancer. Investigation of biomarkers in different types of bodily fluids has been a potential tool for early detection of cancers. In this study, we applied mass spectrometry- based proteomics on human breast milk as an appropriate cancer microenvironment for identification of potential protein biomarkers of breast cancer. Intact human breast milk samples from breast cancer suffering mothers and matched controls were subjected to one dimensional and two dimensional polyacrylamide gel electrophoresis (2D-PAGE), quantified, and the protein spots were cut and digested with trypsin. The peptides were extracted and zip-tipped followed by nanoliquod chromatography-tandem mass spectrometry (nanoLC-MS/MS) analysis using a nanoAcquity UPLC coupled with a Q-TOF Ultima Mass spectrometer. Data base search was done by Mascot Daemon (version 2.5.1.) software and the results were analyzed by Scaffold Proteome (version 4.2.1) software. Protein alterations (upregulation and downregulation in protein expression) in breast milk samples were identified and are under further investigation and considered as potential breast cancer biomarkers.
Acknowledgement: The authors would like to thank all the participants for generously donating their breast milk and participating in this study.
A.3
Profiling Biochemical Individuality: Human Personal Omics Profiling (hPOP)
Sara Ahadi, Hannes Rost, Liang Liang, Mike Snyder
Stanford University, Stanford, CA, USA
Recent advances in high throughput technologies allow profiling of thousands of analytes within a single experiment. These measurements could potentially be used to diagnose disease early, monitor treatment progression and stratify patient groups to ensure each individual obtains the treatment best suited to their needs. This personalized approach to medicine would include continuous monitoring of thousands of parameters over a whole lifetime. However, in order to be able to interpret such data, we need to have a better understanding of the underlying natural variation of biological molecules in large crowds. If we know the natural ranges of individual analytes, the expected responses to perturbations and the long-term trends in their levels, we can draw meaningful conclusions from comprehensive personalized profiling. Tryptic peptides of plasma samples were separated on a NanoLC™ 425 System (SCIEX). 5ul/Min flow was used with trap-elute setting using a 0.5 x 10 mm ChromXP™ (SCIEX). LC gradient was set to a 43-minute gradient from 4–32% B with 1 hour total run. Mobile phase A was 100% water with 0.1% formic acid. Mobile phase B was 100% acetonitrile with 0.1% formic acid. 8ug load of undepleted plasma on 15cm ChromXP column. MS analysis were performed using SWATH® Acquisition on a TripleTOF® 6600 System equipped with a DuoSpray™ Source and 25μm I.D. electrode (SCIEX). Variable Q1 window SWATH Acquisition methods (100 windows) were built in high sensitivity MS/MS mode with Analyst® TF Software 1.7. We were able to successfully run a pilot study on human Personal Omics Profiling (hPOP) in March 2016 during US HUPO conference in Boston and then launched the study at 2016 HUPO overall about 150 individuals participated in the study so far and their urine, stool and blood samples have been collected. Our preliminary data on plasma proteins quantified suggests large amount of variance in the data and there is variance we currently cannot explain. With little technical variance, individual variance is large in certain proteins. Using our current SWATH-MS, we used 8ug of undepleted plasma for one hour LC/MS run. In total 530 proteins with 1%FDR in peptide level were quantified over 31 subjects. Over 80% of all proteins showed less than 10% CV in SWATH data on hPOP pilot plasma samples whereas when we add technical and biological variance, the 80% of all proteins have over 50% variance. Overall there is 46.5% biological variance which includes the unknown genetic components and unknown environmental components and then there is the 21% of the variance (on average) that we can explain with variables such as age, sex, BMI and ethnicity. We expect these studies to reveal novel, multi-analyte signatures from the blood that are associated with aging and we hope to discover additional markers that are different in ethnic groups. The data to be generated here represent an unprecedented level of detail for blood-based profiling among a diverse panel of individuals and we hope that this study can propel additional multi omics work exploring factors leading to discovery of markers for aging.
A.4
Towards a Human Chemo-proteomic Database: Profiling Structurally Diverse Chemical Space to Map Proteome-wide Interactions
Francisco Garcia, Kyeongmi Cheon, Huijun Wang, Andy Liaw, Vidhi Mishra, Anne Mai Wassermann, Brian Dill, Benjamin Ruprecht, An Chi, Xudong Qiao, and Ivan Cornella-Taracido
Merck Research Laboratories, Boston, MA, USA
Chemoproteomics utilizes affinity chromatography, high resolution mass spectrometry, biostatistics, and informatics for large-scale identification and quantification of proteins perturbed by chemical probes in biological samples. Chemoproteomics is used extensively to assess drug-target engagement, selectivity, biomarker discovery, target identification, and drug repurposing. Chemical biologists routinely apply chemoproteomics to characterize compound-target interactions for discrete project specific biosamples. A comprehensive and systematic evaluation of bioactive molecules against the whole human proteome however, has not been reported. We therefore aim to analyze proteome-wide interactions for hundreds of chemical probes against a selection of physiologically relevant and distinct human cell and tissue samples to generate an unprecedented first draft map of the human chemo-proteome. Herein we provide initial datasets towards this first-in-class experimental and unbiased map of the human “interactome” connecting biological samples with small molecules through the discovery of proteins that bind to them in affinity enrichment proteomic experiments. We compare various mass spectrometry workflows to enable robust protein identification and quantification while minimizing instrument time. We examine several statistical methods to maximize identification of compound-protein interacting partners. Furthermore, we discuss recent advances towards a medium throughput approach to survey a large set of structurally diverse compounds against biological matter to boost the chemical and biological space profiled at a time. These efforts to develop a unique human chemo-proteomic database, together with chemo-genomic and transcriptomic approaches, provide chemical biologists the means to prosecute novel target identification and subsequent validation studies to support relevant disease areas.
A.5
Development of Quantitative Mass Spectrometry Assays for Species-matched Surrogate Antibody Fragments in Ocular Matrices
Hilda Hernandez-Barry, Robert F. Kelley, Devin Tesar, Whitney Shatz, Laetitia Comps-Agrar, Joyce Chan, Keyang Xu, Luna Liu, Yanqiu Liu, Mauricio Maia, Kelly Loyet
Genentech Inc., South San Francisco, CA, USA
Introduction: Species-matched antibody fragments (Fabs), e.g., rabbit and cynomolgus monkey Fab, have been developed for pharmacokinetic (PK) and safety assessments, investigating long-acting ocular delivery technologies without invoking an immune response to a humanized molecule in an animal model. Such anti-therapeutic antibodies confound the interpretation of assay results and other study end-points. Distinguishing the species-matched dosed surrogate Fab from endogenous antibodies to assess pharmacokinetics requires either a labeled Fab or a specific assay, such as quantitative MS or ELISA. Labeling the Fab may alter its properties and a specific ELISA requires the time-consuming process of molecule-specific antibody generation. We developed an LC-MS/MS approach that uses signature peptides from Complementarity Determining Regions (CDRs) of the Fabs, allowing for flexibility and faster assay development. Methods: All recombinant proteins and samples were digested using Trypsin (Promega) and various timepoints and temperatures were investigated for optimal digestion. Detection of most abundant peptides was performed using Orbitrap-Elite (Thermo Fisher Scientific). For all Fabs, determination of most sensitive peptide in region of interest (CDR) and MRM method development was performed by a Sciex - Qtrap 6500 mass spectrometer coupled to Acquity-Waters HPLC. Isotopic labeled peptides were obtained from New England Peptide (Gardner, MA). Different amounts of acetonitrile (Fisher Scientific) were added into samples in order to determine the optimal MS solvent. Preliminary data: In order to quantitate species-matched Fabs in rabbit or cynomolgus monkey vitreous humor, several parameters and conditions were assessed and optimized. First, in-silico digest by several enzymes was necessary to localize peptides in the regions of interest (CDRs) that met the mass spectrometry range criteria. After the possible peptides were identified, digestion was performed, investigating temperature and effects of ocular matrices on the assay sensitivity of specific peptides of interest. Once the most sensitive (and less prone to modification) peptides were identified, their heavy isotopic labeled counterparts were generated in order to perform absolute quantitation by MS. Validation of the methods was performed by analyzing vitreous humor samples spiked with known concentrations of rabbit Fab. Rabbits were dosed with the rabbit Fab via an intravitreal injection, and samples were obtained at defined time points. A sandwich ELISA that utilized anti-idiotype antibodies against the rabbit Fab was used to quantify rabbit Fab in the vitreous humor for comparison to results obtained with the MS assay. Since the cynomolgus monkey Fab consisted of the CDRs from a human Fab that were grafted onto cynomolgus monkey framework, the development and validation of a quantitative MS assay for the cynomolgus monkey Fab could be accomplished using the human Fab. This enabled considerable time savings in that the assay could be developed concurrent with the cynomolgus monkey Fab design, generation, and protein purification. In addition, vitreous humor samples collected following intravitreal injection of the human Fab into rabbit eyes could be employed for analysis with a generic human Fab ELISA or generic affinity capture LC-MS/MS method in order to compare with the current MS assay. In conclusion, a practical, robust, and sensitive method for quantitation of species-matched Fabs in preclinical samples was developed and qualified to support early-stage drug development studies. Novel Aspect: Quantitation of rabbit and cynomolgus monkey species-matched Fabs in ocular matrices using signature peptides from Fab CDR regions.
A.6
The functional landscape of human protein phosphorylation
David Ochoa, Andrew Jarnuczak, Juan A. Vizcaíno, Pedro Beltrao
European Bioinformatics Institute (EMBL/EBI), Hinxton, Cambridgeshire, UK
Cells need to constantly adapt to internal and external conditions in order to maintain their homeostasis. One of the fastest response mechanisms is the targeted protein phosphorylation mediated by kinases. Through these post-translational modifications the cell can quickly alter enzymatic activities, protein interactions or sub-cellular localizations in order to activate or inactivate specific processes. However, our understanding of the more than 200.000 potentially relevant human phosphorylation sites has been limited so far by our ability to distinguish highly relevant events affecting protein function from those with minor or no impact. In this study, we aim to catalogue and functionally characterize recurrent human phosphorylation events in order to prioritize those with relevant biological consequences. By re-analyzing a large compendium of phospho-enriched proteomics studies deposited in PRIDE, we defined a human reference phosphoproteome supported by available MS evidence. Modified residues were then scored using a machine learning approach combining more than 30 different features regarding sequence, structural, regulatory or evolutionary properties. The resulting scores not only separate known from unknown regulatory sites, but also reports significantly higher values for disease-related phosphosites. Additionally, the functional score successfully discriminates those phosphorylation sites regulating certain biological processes from other unspecific phosphorylations that we aim to validate. The resulting reference phosphoproteome, together with the predicted functional relevance and annotation, represents a resource to explore the human signaling under perturbation or disease.
A.7
Direct Proximity Tagging of Small Molecule Protein Targets
Zachary Hill1, Samuel B. Pollock1, Min Zhuang2, James A. Wells1
1University of California, San Francisco, CA, USA; 2ShanghaiTech University, Shanghai, China
Identifying the protein targets of bioactive small molecules remains a major challenge in the development of new small-molecule therapeutics. Knowing a compound's targets and mechanism of action is important from both an efficacy and safety perspective, and is increasingly required by regulatory bodies. While affinity resins and activity-based probes have both had successes in identifying the targets of small molecules, both have limitation. Here we will present a method for direct proximity tagging of proteins that bind small molecules. Recently, our lab, as well as several others, has developed proximity-tagging enzymes capable of covalently tagging the interaction partners of proteins with biotin. Our system, based on the NEDD8-E2-ligase Ubc12, is able to directly transfer a small ubiquitin-like protein, NEDD8, onto lysines of the target proteins. If the NEDD8 used is biotinylated, then the target proteins can be enriched and then identified using liquid-chromatography tandem mass spectrometry (LC-MS/MS). While these proximity-tagging technologies have previously been used to look at protein-protein interactions, here we will present an extended application of our technology to identifying the protein targets of a small molecule. Utilizing the drug dasatinib, we have shown dasatinib-directed NEDDylation of endogenously expressed dasatinib targets in complex cell lysates. In addition, we have been able to improve our engineered NEDDylation enzyme through rational mutagenesis. Finally, we have been able to apply our proximity tagging technology to two other protein-ligand interactions beyond dasatinib and kinases, demonstrating the generality of this method. Proximity tagging using our engineered ligase requires direct binding of the target and, thus, provides a useful and orthogonal approach to small molecule target identification.
A.8
iST: a reproducible sample preparation method for in-depth proteome discovery and interaction proteomics
Garwin Pichler1, Fabian Hosp1, Nils Kulak1, Matthias Mann2
1PreOmics, Martinsried, Germany; 2Max Planck Institute of Biochemistry, Max Planck Institute of Biochemistry, Munich, Germany
Mass spectrometry (MS)-based proteomics typically employs multiple sample processing steps, representing a crucial part of routine MS analyses. Complex workflows, extensive sample fractionation and proteolytic digestion are highly time-consuming and restrict the overall technical reproducibility. The accuracy and robustness of the MS platform is also strongly affected by sample quality reasoning for high quality proteomic samples. Here, we present the straightforward and robust in-StageTip (iST) method for streamlined sample processing of complete proteomes and immunoprecipitations (IPs). The iST method is a 3-step procedure performed in a single, enclosed volume, which thereby circumvents the likelihood of contamination and sample loss. Due to the straightforward nature, the method can readily be performed in a 96-well format on liquid handling robotic systems. The method is highly compatible with established and novel StageTip based pre-fractionation methods and thereby allows in-depth analysis of complex proteomic samples. Moreover, the iST method can also be applied to process immunoprecipitated protein samples or to clean-up samples with high detergent contamination. Applying the iST method to the well-studied cancer cell line HeLa allowed us to estimate protein copy-numbers of 9,667 proteins. The results demonstrated excellent reproducibility (R2 = 0.97) in quadruplicates measurements, reflecting the overall strength of the method. In addition to cell lines, we applied this sample preparation workflow to process yeast, tissue samples or body fluids such as blood, urine or CSF and obtained in-depth proteome coverage. Finally, we demonstrate the compatibility of the iST method with immunoprecipitation approaches utilizing agarose and magnetic beads, enabling streamlined and robust sample processing for IP samples. Our workflow retrieves known complex members and efficient enrichment of DNA-repair factors to a bait protein involved the base excision repair pathway. The iST method opens up opportunities for clinical applications and protein interactomics, while keeping exceptional sample quality at low-cost effort.
A.9
Requirements of O-GlcNAcylation in hepatocytes and liver metabolic function
Krista Kaasik, Robert J. Chalkley, A. L. Burlingame
University of California, San Francisco, CA, USA
O-linked-β-N-acetylglucosamine transferase (OGT) post-translationally modifies serine and threonine residues of proteins and is required for stem cell viability and embryogenesis. OGT knockout is early embryonic lethal but not much is known about the different tissue-specific functions of OGT and O-GlcNAcylation in differentiated adult tissues. We have focused on O-GlcNAcylation in hepatocyte function in conditional liver-specific OGT knockout mice. Albumin-Cre driven Ogt knockout mice develop normally. Remarkably, hepatocytes are quite resistant to the loss of OGT enabling a model system for proteome-wide analysis of reduced O-GlycNAcylation in vivo. O-GlcNAcylation is drastically reduced in mutant liver tissue in vivo during postnatal development analyzed by WGA chromatography. The loss of OGT in hepatocytes increased cell proliferation compared to control littermates. Co-expressed Cre dependent lineage marker indicates unexpected prolonged survival of hepatocytes upon loss of OGT activity. In addition, we have characterized phosphorylation, acetylation, N and O-glycosylation changes in the hepatic proteome in response to loss of OGT activity. We have discovered that Glycogen Synthase (Gys) activity is regulated by the attachment of a single N-acetylglucosamine in O-glycosidic linkage to Gys protein, catalyzed by Ogt. We have identified the single site of Gys protein by ESI-ETD-MS/MS method and showed that O-GlcNAcylated Gys is active while phospho-form is inhibited in glycogen synthesis. This site co-localizes with mutations in patients with GSD 0. Manipulating the levels of nutrients (high glucose vs. low glucose) increases O-GlcNAcylated Gys. Our metabolic experiments using mice with modified OGT activity mimic human GSD 0 displaying a hypoglycemic phenotype with very low levels of glycogen. These data indicate OGT mediated control of glucose immobilization by GS in liver.
A.10
A Tale of Two – Data Independent Acquisition applied to maximize proteome coverage and throughput
Roland M. Bruderer1, Oliver M. Bernhardt1, Tejas Gandhi1, Jan Muntel1, Sebastian Müller1, Polina Mironova2, Ondine Walter2, Jérôme Carayol2, Jörg Hager2, Armand Valsesia2, Loic Dayon2, Arne Astrup3, Wim H. M. Saris4, Florian Marty1, Lukas Reiter1
1Biognosys, 8952 Zurich-Schlieren, Switzerland; 2Nestlé Institute of Health Sciences, Lausanne, Switzerland; 3Department of Nutrition, Exercise and Sports, Faculty of Science, University of Copenhagen, Copenhagen, Denmark; 4NUTRIM, School for Nutrition, Toxicology and Metabolism, Department of Human Biology, Maastricht University Medical Centre, Maastricht, The Netherlands
Data independent acquisition (DIA) is a powerful approach for proteome quantification providing high sensitivity, quantitative precision and reproducibility. People are often not applying DIA because they think they have to sacrifice either proteome coverage or throughput. In this presentation we show that both high throughput and high proteome coverage can be achieved by applying either new data analysis strategies or modifying the experimental set up. In the first part we will focus our presentation on the newly implemented directDIA workflow enabled in Spectronaut Pulsar on data generated on the Thermo Fisher Q Exactive HF-X. Best performance in DIA is achieved by a spectral library typically generated with data dependent acquisition (DDA). Since spectral library generation requires significant overhead, particularly in small-scale experiments, spectral library-free DIA analysis techniques (Silva et al. 2006, Tsou et al. 2015) offer the promise of significant savings in instrument time and thus increasing throughput. Here, we present an algorithm for direct analysis of DIA data which is integrated into Spectronaut Pulsar. We applied this algorithm to data generated on several instruments, among them the new Q Exactive HF-X. Among the three data sets analyzed and presented the first one comprises samples derived from cancerous lung tissue of twelve individuals and corresponding adjacent healthy tissue. These 24 samples were acquired in DIA mode on the Q Exactive HF-X. On a 120min gradient we identified over 4,300 protein groups (FDR 1% on peptide and protein level) in all 24h runs. Data completeness achieved on the protein level was > 98%. In the second part we are focusing on the capillary flow system set up to measure 1546 human plasma samples. The proteins from the blood circulatory system are indicative for the status of an individual. Comprehensive, robust high throughput analysis of the proteome will enable holistic analysis of the health state. Recently, three studies were published advancing this task based on nano-flow LC-MS DDA combined with TMT or label free quantification and MS1 alignment (Cominetti et al. 2015 JPR, Geyer et al. 2016 MSB, Geyer et al. 2016 Cell Systems). Nano-flow setups are delicate and require long gradient overheads cutting down throughput for large scale studies. To significantly reduce this limitation, we developed a simple and robust workflow for label-free proteome quantification with high quantitative precision and high reproducibility using a robust capillary-flow LC-MS-DIA. This workflow is well suited for cohorts of over a thousand samples. After optimizing the method, we ran 1564 samples from the DiOGenes consortium with a 40min gradient (6min overhead). All samples were measured using a single column and analyzed at once with Spectronaut. On average we quantified 500 protein groups per run (1% FDR at peptide and protein level). The data presented here illustrate the current proteome coverage and throughput achievable with current DIA workflows.
A.11
Quantifying early events on mitochondria in the PINK1/PARKIN ubiquitin ligase signaling cascade
Alban Ordureau, Joao A. Paulo, Jiuchun Zhang, Wade Harper
Harvard Medical School, Boston, MA, USA
Recent work indicates that a major role for PINK1 and PARKIN, two proteins mutated in Parkinson's disease, is in mitochondrial outer membrane (MOM) ubiquitylation, which promotes mitophagy. We have previously found that PARKIN can build K6, K11, K48, and K63 ubiquitin (Ub) chains in response to mitochondrial depolarization (1). The ability of PARKIN to build these chains is greatly stimulated by PINK1-dependent phosphorylation of not only PARKIN, but also Ub itself on S65. A critical feature of the system is that phosphorylation of Ub chains by PINK1 as they are assembled by PARKIN leads to further recruitment of PARKIN through its phospho-Ub binding site (1,2). To measure chain assembly activity and kinetics, we previously developed a quantitative proteomic-based assay that measures total Ub, each individual linkage type as well as Ub phosphorylation of Serine 65 (2). We are now building assays that include over 100 peptides representative of 19 proteins to examine primary ubiquitylation and phosphorylation sites on MOM proteins. The quantitative proteomic-based assay we have developed robustly works in HeLa cells, but we've now applied successfully this system in different types of neurons in culture or in vivo. We have performed experiment using mitochondria from WT or gene-edited ES cells differentiated to neuronal precursor cells (NPCs) and have detected UB chain assembly, UB phosphorylation, and subset of MOM proteins ubiquitylation upon activation of the PINK1/PARKIN pathway. The ability to accurately monitor PARKIN activity at the primary substrate level is critical for multiple purposes, including: a) measuring PINK1/PARKIN pathway activity precisely in cells, b) measuring the effect of small molecule activators or inhibitors of the pathway on mitochondrial ubiquitylation, c) examining if there is selectivity in PARKIN substrate targeting in various experimental settings. I will show we can now quantitatively monitor ubiquitylation of primary ubiquitylation sites in a cohort of substrates simultaneously using this novel assay and advance more generally our understanding of Ub signaling cascades.
This work was supported by the Michael J. Fox Foundation and the Lefler Foundation.
1. Ordureau A, et al (2014) Mol Cell2. Ordureau A, et al (2015) PNAS.
A.12
Pyrraline, Argpyrimidine and carboxymethyllysine adducts of advanced glycated apoA-IV do not change the apolipoprotein ability in removing cell cholesterol and inhibiting inflammation in macrophages
Ligia S. Okuda1,3, Russell Pickford2, Mili Patel3, Tom Woods4, Margaret Brimble3, Kerry-Anne Rye3, Marisa Passarelli1
1Faculdade de Medicina, Universidade de Sao Paulo, Brazil; 2Bioanalytical Mass Spectrometry Facility and 3Lipid Research Group, University of New South Wales, Sydney, Australia; 4School of Chemical Sciences, University of Auckland, Aukland, New Zealand
Background: Advanced glycation end products (AGE) are related to atherogenesis by impairing cholesterol efflux and anti-inflammatory role of apolipoprotein A-I. We aimed to characterize the glycation profile of apoA-IV in comparison to apoA-I and its effects in apoA-IV anti-inflammatory properties and ability to mediate cholesterol efflux.
Methods: Recombinant apoA-IV (E coli) or apoA-I isolated from healthy human plasma (1 mg) were submitted to in vitro glycation by incubation with 1 mM glycolaldehyde (GAD) or 1 mM methylglyoxal (MGO) at 37°C for 24h. Pyrraline (PYR), carboxymethyllysine (CML) and argpyrimidine (ARG) were determined by LC-MS/MS (nmol of PYR or CML or ARG/mmol of Lysine or Arginine) (TSQ Quantum Access triple quadrupole mass spectrometry - Thermo Fisher Scientific). Cholesterol-loaded J774 or bone marrow-derived macrophages were pre-incubated with C or AGE-apoA-IV (50 μg/mL, 24 h). TNF secretion was determined by ELISA after exposure to lipopolyssacharide (LPS; 1 μg /mL; 24 h). Cholesterol efflux was measured by incubating 14C-cholesterol and acetylated LDL-loaded macrophages with C or AGE-apoA-IV (50 μg /mL) for 8h. Results: In comparison to apoA-I, incubation of apoA-IV with GAD generated less CML (563 ± 85 vs 379 ± 76, p = 0.001, n = 3) and more PYR (73 ± 14 vs 313 ± 60, p = 0.0003, n = 3); while incubation with MGO produced more ARG (6161 ± 3116 vs 11922 ± 3636; p = 0.009, n = 3). The amount of CML, PYR and ARG was 468, 1305 and 332 times higher in AGE-apoA-IV in comparison to C-apoA-IV. The LPS-mediated TNF secretion was inhibited by C-apoA-IV (625 ± 37 pg/mL) as well as by GAD-apoA-IV (413 ± 127 pg/mL) and MGO-apoA-IV (357 ± 12 pg/mL) compared to incubation with LPS alone (1451 ± 674 pg/mL; p < 0.05, n = 6). The % of 14C-cholesterol efflux was similar among C (25 ± 4.5, n = 6), GAD (25 ± 3.2, n = 6) and MGO-apoA-IV (22 ± 6, n = 6). Conclusion: AGE-apo A-IV has more PYR and ARG and less CML compared to AGE-apoA-I and differently from what is described for apoA-I does not have its anti-inflammatory properties or the cholesterol efflux capacity affected by advanced glycation.
Acknowledgement: This work was supported by FAPESP, Brazil (Fundação de Amparo à Pesquisa do Estado de São Paulo) (2013/02854-7; 2013/23392-1).
A.13
High-Throughput Screening of Proteoform-Specific Binding of a Chemical Library
Daojing Wang, Geuncheol Gil, Pan Mao
Newomics Inc., Emeryville, CA, USA
Identification of the specific binding between small molecules and proteins is a critical step in drug and its target discovery. Current quantitative high-throughput screening (qHTS) are predominantly based on phenotypic assays using fluorescence or chemiluminescence readouts. Here we describe a mass spectrometry-based method for high-throughput screening of interactions between proteins and natural products in a chemical library. Our method identifies and measures the stoichiometry of the specific binding between small molecules and their protein targets at the proteoform level in a proteoform-specific manner. Covalent from noncovalent binding can be further differentiated. A workflow for automatic operation of 24-plex multinozzle emitter array (MEA) chip was developed to perform high-throughput screening. Custom-built LabVIEW software realized full automation of the workflow. The chip consisted of 24 identical units, each of which contained an inlet hole for sample injection, a multinozzle emitter for nanoelectrospray, and a connecting channel bridging the hole and emitter. Manifolds were built for chip assembly that provided sample delivery from external HPLC source. Using this system, we demonstrated a high-throughput assay to screen the proteoform-specific binding of the over 100 natural products to HSA. For each sample, the acquisition time was 1 min and total time was 5 min, and the screening of the whole library took about 12 hrs. The throughput can be improved by reducing sample injection time as well as system swept volume in the future. Under our direct infusion native nanoESI-MS conditions, we looked at both the non-native (denatured and partially denatured) and native forms of HSA to determine whether the binding was covalent, non-covalent, or no binding. In addition, the binding stoichiometry (the number of molecules per HSA) can be determined by quantifying the relative abundance of their free and bound proteoforms. Out of the over 100 compounds, we identified different numbers of covalent hits, non-covalent hits, and non-hits. Interestingly, several compounds were found to have both covalent and non-covalent binding with HSA. We also studied how binding mode was affected by changing ionic strength or hydrophobicity of the reaction buffer. In summary, our silicon microfluidic chip platform enables high-sensitivity and high-throughput mass spec-based screening of proteoform-specific binding between small molecules and target proteins (not limited to HSA).
A.14
Accurate measurement of acetylation stoichiometry by quantitative MS
Brian Weinert, Chunaram Choudhary, Bogi Karbech Hansen
University of Copenhagen, Copenhagen, Denmark
In recent years, mass spectrometry (MS) enabled the identification of lysine N-ε-acetylation sites on thousands of proteins throughout the cell, suggesting that acetylation regulates nearly all aspects of cellular function. However, the initial excitement that followed these discoveries has given way to a more nuanced view of acetylation. The technical bias to detect acetylation sites on abundant proteins led to a widespread misconception that acetylation is overrepresented on metabolic proteins; it is not (Nature Methods 2015, 12:1003–4). Results from a number of groups indicate that most of the acetylation sites identified in cells occur by nonenzymatic reaction with acetyl-CoA. Other acyl-CoA metabolites similarly modify lysine residues, resulting in lysine succinylation, malonylation, and so on. This provides an attractive model whereby metabolite levels regulate metabolism through nonenzymatic modification. Mutational analysis of acetylation sites on metabolic proteins mostly shows that acetylation reduces their catalytic activity. Therefore, determining the stoichiometry (degree) of acetylation is important for understanding how acetylation impacts metabolic protein function. To study acetylation stoichiometry, we developed a method based on comparing chemically acetylated peptides to native acetylated peptides. We applied this method to provide the first global analysis of acetylation stoichiometry in yeast cells (Mol Syst Biol 2014, 10: 716). We subsequently measured stoichiometry in mouse liver tissue to demonstrate that the mitochondrial sirtuin deacetylase SIRT3 suppresses acetylation to very low stoichiometry (EMBO J 2015, 34:2620–32), supporting a model in which SIRT3 acts as a protein repair factor targeting acetyl lesions. Recently we developed a method to control for quantitative accuracy that was applied to measuring stoichiometry in E. coli (Mol Cell Proteomics 2017, 16:759–769). These data revealed striking similarities in how sirtuin deacetylases interact with low stoichiometry acetylation. In unpublished work we are characterizing acetylation stoichiometry in human (HeLa) cells. These data are being analyzed to reveal functional categories modified by high stoichiometry acetylation, to study the relationship between stoichiometry and regulation by deacetylases and acetyltransferases, and as a means to distinguish regulatory acetylation events from the vast background of nonenzymatic acetylation.
A.15
Extraction and Characterization of Prenylated Proteins
James Wilkins, A. L. Burlingame
University of California, San Francisco, CA, USA
A number of proteins are known to be posttranslationally modified at their C-terminus with either C15 (farnesyl) or C20 (geranylgeranyl) isoprenoid species in a process known as prenylation. These proteins, many of which are GTPase switches, are involved in critical signaling and cellular transport activities. The enzymatically-catalyzed addition of an isoprenoid moiety to cysteine residues located at the C-terminus of these proteins facilitates their interaction with membranes and is absolutely required for their normal cellular function and subcellular localization. However direct characterization of these modifications has been limited by the lack of broadly applicable analytical approaches, although various chemical labeling and tagging strategies have been employed. We have approached this problem using high resolution LC coupled with MS technologies taking advantage of tissue extraction methods known to be selective for disruption of selected areas of protein-lipid interaction. Increasing numbers of known prenylated proteins were extracted by titration using a known lipid raft disruptor: methyl β cyclodextrin. We identified a number of these proteins using a “shotgun” proteomics approach coupled with targeted informatics searches to find and identify peptides derived from the proteins of interest. In addition, chemically modified synthetic peptides were used as aids in development of optimal liquid chromatographic approaches, particularly to enhance chromatographic efficiency prior to high resolution mass spectrometry. The ultimate goal of this work is a generalized method for detailed profiling of a wide range of prenylated proteins including their posttranslational modifications.
Acknowledgment: Supported by NIH grant 8P41GM103481 and Sheldon G. Adelson Medical Research Foundation.
A.16
Profiling the Distribution of N-Glycosylation in Therapeutic Antibodies using LC or CE separation in combination with a QTRAP® 6500 System
Jenny Albanese1, Christie L. Hunter1, Ningombam Sanjib Meitei2, Andras Guttman1, Marton Szigeti3
1Sciex, Redwood City, CA, USA; 2PREMIER Biosoft, Palo Alto, CA, USA; 3University of Debrecen, Debrecen, Hungary
Immunoglobulin G molecules have become attractive as targeted therapeutic proteins, due to their high specificity and long circulation time. Glycosylation patterns determine the stability and bio-disposition of these recombinant protein drugs in vivo, as well as the efficacy, folding, binding affinity, specificity and pharmacokinetic properties. Therefore, a complete characterization of the biotherapeutic IgG glycosylation is desirable. In this study, we demonstrated how a comprehensive MS/MS analysis of the glycopeptides can be achieved by targeting the known nature of the glycosylation structures. We used Trastuzumab, a humanized mAb, and SilumAb in this proof of concept study. Monoclonal IgGs have a known glycosylation site and set of glycan isotypes (complex, high mannose and hybrid). We applied a robust MRM triggered MS/MS workflow to profile and confirm glycopeptides and their glycan isoforms. Overlaying several specific MRMs for each glycopeptide group and searching each full scan MS/MS spectrum with SimGlycan® provides added confidence to the identification of each glycopeptide isoform. In addition, to a front end LC-MS approach we started to explore a different CESI-MS separation using the same MS instrument. Integration of CE and electrospray ionization (ESI) into a single dynamic process (CESI) provides the capability of performing CE separation and MS ionization with ultra-low flow rates, resulting in reduced ion suppression and improved sensitivity. We present a set of proof-of-principle experiments as an alternative workflow to RPLC.
A.17
Identifying the N-terminal acetyltransferases involved in N-terminal Acetylation of RNA polymerase II subunits
Anita Saraf1, Amber Mosley2, Yan Hao1, Michael P. Washburn1, Laurence Florens1
1Stowers Institute for Medical Research, Kansas City, MO, USA; 2Indiana University School of Medicine, Bloomington, IN, USA
N-terminal acetylation (Nt-acetylation) of proteins is one of the most common modifications that occurs both in prokaryotes and eukaryotes and can occur co- or post- translationally. N-terminal acetylation in eukaryotes occurs co-translationally on nascent polypeptide chains as they leave the ribosomes. The main target of this modification is the initiator methionine residue or the second residue if methionine is cleaved. Protein N-terminal methionine excision is performed by essential methionine aminopeptidase enzymes if the second residue has a side chain with a small radius of gyration. Nt-acetylation in eukaryotes is catalyzed by N-terminal acetyltransferases which transfer acetyl group from acetyl-CoA to the termini of α-amino group of the first amino acid residue of a protein. Nt- acetylation of proteins in eukaryotes serves as a multifunctional regulator where they can be involved in protein stability, protein degradation, subcellular location of proteins and formation of protein complexes. Our goal in this study is to identify the N-acetyl transferases involved in the Nt-acetylation of RNAPol II sub-units in S. cerevisiae. RNA Pol II complex was affinity purified from the wild type and three yeast strains deleted for subunits of different N-acetyl transferases, namely Ard1p, Mak3p, Nat3p. Affinity purified RNAPol II complex was subjected to proteolytic digestion using Trypsin or Endoproteinase Lys-C. To make the detection and identification of the Nt-acetylated peptides more robust, we used an enrichment approach that takes advantage of the fact that Nt-acetylated peptides lack a positive charge at their N-terminal α-amino groups and can be separated efficiently from the internal peptides based on charge differences. The enrichment approach we used involves the chemical modification (demethylation) of the peptide mixture followed by off-line Strong Cation Exchange (SCX) chromatography. Dimethylation increases the charge differences between the Nt-acetylated peptides and the internal peptides. Nt-acetylated peptides eluted in the flow through and some early SCX fractions. The resulting SCX-fractions were analyzed using reverse-phase nano-liquid chromatography (RP-nLC) coupled to high resolution tandem mass spectrometry. We reproducibly detected most of the well-known Nt-acetylation of RNAPol II sub-units along with the unmodified N-terminus, if present. These acetylations were detected at high levels in agreement with the literature, illustrating how comprehensive our approach is. The characterization of significant differences between Nt-acetylation pattern between the wild-type and the deletion strains of S. cerevisiae will be helpful in identifying the N-acetyl transferases involved in the Nt-acetylation of RNAPol II subunits.
A.18
Deciphering the Mechanism of Interferon Regulation in pDCs Using Phospho and Ubiquitin Remnant Motif Proteomics
Dirk M. Walther, Alex Pellerin, Kejie Li, Dania Rabah, Peter Juhasz
Biogen, Cambridge, MA, USA
Plasmacytoid dendritic cells (pDCs) are responsible for type I interferon secretion in response to virus infection after sensing of viral DNA by Toll-like receptor(TLR) 9. In patients with the autoimmune disorder systemic lupus erythematosus (SLE), TLR signaling in pDCs is activated by endogenous nucleic acids, resulting in dysregulated type I interferon production. Type I interferon production in pDCs is tightly regulated by several non-redundant inhibitory receptors. The signaling events downstream of these receptors, however, are poorly understood. To fill this gap, we interrogated TLR and inhibitory pathways in a pDC cell line. Cells were stimulated with the potent TLR9 agonist CpG, crosslinking of an ITAM-mediated signaling receptor, or both stimuli at the same time for durations ranging from 5 to 120 min. Cells were lysed and digested, and samples were subjected to sequential enrichment for phosphorylated or ubiquitin remnant motif-containing peptides using immobilized metal ion affinity chromatography or antibody-based enrichment, respectively. Peptides were analyzed on a QE HF mass spectrometer and quantified with TMT isobaric labeling. Our dataset comprises more than 24,000 distinct phosphorylation and 6,500 ubiquitin remnant motif sites with high localization probablilities, making it the most comprehensive PTM study on TLR signaling to date. 15 min of CpG stimulation significantly changed over 1% of quantified sites at the phospho proteome level which include previously reported modifications on proteins such as p38-alpha/MAPK14, MAPKAPK2 and 3 as well as hundreds of novel sites. A ruleset was established to identify modifications involved in pathway crosstalk, yielding a small group of phosphorylation sites with differential regulation by co-stimulation with CpG and receptor crosslinking compared to CpG alone. Affected proteins had functions in diverse biological processes, including vesicle trafficking and cellular signaling. In summary, our data not only expands our knowledge of TLR signaling in general, but also sheds light on the mechanism governing the negative regulation of interferon production in pDCs.
A.19
Ancient regulatory role of lysine acetylation
Samuel Payne, Ernesto S. Nakayasu
Pacific Northwest National Laboratory, Richland, WA, USA
Acetylation of lysine residues is a common protein post-translational modification in bacteria and eukaryotes. Unlike phosphorylation, whose functional role in signaling has been established, it is unclear what regulatory role acetylation plays and whether it is conserved across evolution. By performing a phyloproteomics analysis of 48 phylogenetically distant bacteria, we discovered conserved acetylation sites on catalytically essential lysine residues that are invariant throughout evolution. Lysine acetylation removes the residue's charge and changes the shape of the pocket required for substrate/cofactor binding, which blocks enzyme activity. Two-thirds of glycolytic and TCA cycle enzymes are acetylated in these critical sites. Our data suggest that, unlike phosphorylation whose biological function is usually in signal transduction, acetylation may play a role in metabolic regulation by switching off enzyme activity. We propose that protein acetylation is an ancient and widespread mechanism of protein activity regulation.
Acknowledgment: This work was supported by the U.S. Department of Energy, Office of Science, Office of Biological and Environmental Research, Early Career Research Program.
A.20
Mechanistic characterization of key factors that drive antibody drug conjugate in vitro efficacy: intracellular payload deconjugation and accumulation relative to free drug potency
Josefa Chuh
Genentech Inc, South San Francisco, CA, USA
Antibody-drug conjugates (ADCs) combine the selectivity of a monoclonal antibody with the cytotoxic effects of payload drugs to deliver tumor-targeted killing. For an ADC to be successful, it needs to bind to its target antigen on tumor cells, internalize via receptor-mediated endocytosis and be transported to the lysosome where subsequent intracellular processing of the ADC will release the biologically active drug. While a direct measurement of ADC cytotoxic potency allows efficient screening and confirmation that new drugs conjugated to antibodies result in proper processing in cells, additional mechanistic characterization is often needed to provide information-rich data to guide further optimization of the ADC design. For example, a quantitative understanding of how ADCs are processed intracellularly can help identify which processing step is impacting payload delivery. Here we describe an in vitro cell based LC/MS drug release assay in a 96-well format to measure the cellular accumulation of cytotoxic payload. The data generated by this assay provided correlation between cellular potency and the amount of cytotoxic payload accumulated inside the cell. In parallel, the biochemical characterization of ADC processing performed using cathepsin B and purified human liver enzyme extract can inform how catalytic cleavage efficiency of linker drugs from ADC impact the intracellular payload accumulation. Taken together, we demonstrated the utility of quantitative LC-MS assays in characterizing the cytotoxic payload release mechanism of ADCs in guiding the design of more effective ADCs.
A.21
Label-free Quantification of Nascent Proteins in Primary Astrocyte Cells by Incorporation of O-Propargyl-puromycin, Affinity Purification, and LC-MS/MS Analysis
Indrek Koppel1, Nancy J. Phillips2, Qian Zhao2,3, Robert J. Chalkley2, Juan A. Oses-Prieto2, Michael Fainzilber1, A. L. Burlingame1
1Weizmann Institute of Science, Rehovot, Israel; 2University of California, San Francisco, CA, USA; 3Hong Kong Baptist University, Kowloon, Hong Kong
The control of gene expression at the level of mRNA translation is central to how cells function and react rapidly to environmental cues but is challenging to assess experimentally. To directly measure nascent protein production in native cells by mass spectrometry, we have employed O-propargyl-puromycin (OPP), an aminonucleoside tRNA mimic that is incorporated into polypeptide chains undergoing elongation. This tag enables subsequent alkyne-azide cycloaddition (click chemistry) for the capture and analysis of OPP-labeled proteins. In this study, we investigated nascent protein synthesis in cultured murine cortical astrocytes in the presence and absence of forskolin, an adenylate cyclase activator and differentiating agent. A rapidly growing body of evidence supports their central roles in central nervous system development and homeostasis. However, compared to neurons these cells remain relatively poorly investigated, including proteomic studies. To investigate the immediate effect of adenylate cyclase activation on protein synthesis in astrocytes, we pre-treated with forskolin or vehicle for 2 h in a serum-free medium and simultaneously labeled with either 10 μM OPP or 0.1% DMSO as a control. Whole cell lysates were reacted with biotin-azide and OPP-tagged proteins were captured on streptavidin magnetic beads and digested with trypsin. Digests were analyzed by LC-MS/MS using a Q Exactive Plus hybrid quadrupole-orbitrap mass spectrometer employing HCD. The relative abundance of identified proteins was calculated as the sum of the individual peptide intensities obtained from extracted ion chromatograms (XICs). Overall, we identified and quantified 2301 unique proteins from 3 biological replicates. To differentiate OPP-enriched proteins from background proteins that bind nonspecifically to streptavidin beads, we evaluated +OPP/-OPP intensity ratios (with and without forskolin) for all proteins by plotting log2 mean values and corresponding p-values derived from the 3 replicate measurements. This graphical approach distinguished 1424 OPP-enriched proteins, including a cluster of 274 strongly enriched proteins with log2 fold-changes >10 and p-values <0.01. The full list of OPP-enriched proteins was further narrowed down to 836 proteins by applying a 1.0 log2 fold-change cutoff for +/- forskolin treatment, of which 139 were in the top-ranked cluster. In the filtered list we observed forskolin-mediated induction of known cAMP-activated immediate early proteins c-fos and jun-B. In addition, forskolin induced rapid synthesis of aquaporin-4 and glutamate transporter GLT-1, known astrocyte markers associated with differentiation. Interestingly, synthesis of Arnt2, a transcription factor associated with neuronal differentiation and identity in the CNS was detected in control cells, but not in forskolin-treated cells, suggesting its role as a negative regulator of astrocyte differentiation. Pathway analysis revealed forskolin-mediated regulation of several with cAMP signaling-associated nodes including G-protein and PKA subunits, as well as nodes associated with cytoskeletal remodeling such as the Arp2/3 complex and Rho GEFs. This work was supported by the Dr. Miriam and Sheldon G. Adelson Medical Research Foundation and NIH grants P41GM103481 and 1S10OD016229.
A.22
Multiple QconCAT biosynthesis in a cell-free protein synthesis system
Nobuaki Takemori1, Ayako Takemori1, Yuki Tanaka1, Yaeta Endo1, Jane L. Hurst2, Guadalupe Gómez-Baena2, Victoria M. Harman2, Robert J. Beynon2
1Proteo-Science Center, Ehime University, Eihme, Japan; 2Institute of Integrative Biology, University of Liverpool, England, UK
[Objective] We propose a novel experimental workflow to create a large-scale standard peptide library for absolute proteome quantification through high-throughput synthesis of QconCATs in a cell-free protein synthesis system. [Background] QconCATs are artificial proteins that are concatenations of standard peptides from multiple natural proteins, expressed heterologously in the presence of stable isotope precursors. When mixed with their natural counterparts and proteolysed, the outcome is a mixture of labelled (standard) and unlabelled (analyte) peptide pairs that can be analysed by LC-MS to yield absolute quantification of the parent proteins. QconCATs are established as powerful tools for absolute proteome quantification, and are widely accepted as an established methodology. One of the barriers to uptake, however, is the complexity associated with bacterial expression of QconCATs. If a workflow for mass production of QconCATs could be established, quantitative proteomic information on human and various model organisms can be acquired comprehensively, which can be a 'big bang' in the field of systems biology. [Preliminary Results] Towards the goal of global proteome quantification by QconCAT approach, we have introduced a wheat germ cell-free protein synthesis system (WGCFS) into the workflow of QconCAT synthesis. WGCFS can provide the highest translation efficiency among eukaryotic cell-free systems and has a very high success rate for the production of difficult-to-express proteins. First, we have demonstrated that a high-throughput parallel production of over 70 QconCATs targeting 1,400 proteins at 2 peptides/protein was achieved within two days. This would take many months of work by traditional bacterial expression methods. Secondly, the total synthesis was of the order of tens of micrograms of each QconCAT, and the yield was sufficiently high to support hundreds of selected reaction monitoring assays using nanoLC-MS. Finally, an extremely high level of incorporation (99.5%) of stable isotope labelled amino acids was achieved. [Conclusion] The ability to high-throughput QconCAT synthesis in WGCFS introduces the potential of readily deployed large-scale standards or of kits that can be used by proteomics laboratories to create instantaneous standards in quantities far in excess of that required for modern mass spectrometers. This enhancement, readily deployed in most laboratories, could catalyze a substantial increase in quality and throughput of quantitative proteomics research.
A.23
Defining the Cell Surface Proteome of MLL Rearranged Leukemia
Matthew Nix, Arun P. Wiita
University of California, San Francisco, CA, USA
B-acute lymphoblastic leukemia (B-ALL) patients that harbor rearrangements of the Mixed-lineage leukemia gene (MLLr) have a particularly dismal clinical outcome. Although it is known that this oncoprotein radically alters epigenetic regulation, the precise molecular mechanisms that define the cellular phenotype of these cancers remain poorly understood. Furthermore, although flow cytometric immunophenotyping has been performed for decades on B-ALL blasts, relatively little is known about the cell surface proteome of these cancers outside of canonical CD markers. The cell surface therefore represents an attractive target for unbiased proteomic profiling due to its role in modulating cellular behavior as well as accessibility to immunotherapy. Here, we aim to resolve the cell surface proteomic landscapes of MLLr and non-MLLr B-ALL cellular models. Our goal is to generate new mechanistic insights into the molecular re-wiring that occurs in these forms of cancer and identify novel candidates for targeted immunotherapy. We examined 10 B-ALL cell lines that harbor different oncogenic fusions including the MLL-AF4, BCR-ABL, and ETV6-RUNX1 translocations, respectively, using both “classical” lines as well as low-passage patient-derived lines, in addition to non-malignant B-cells as a control. We performed unbiased labeling of cell surface proteins by biotinylation of glycosylated amino acids (3 biological replicates each) followed by affinity pull-down. After on-bead digestion, peptides from bound glycoproteins were analyzed on a Thermo Q-Exactive Plus Mass spectrometer using 4-hour LC runs, data-dependent acquisition, and label-free quantification. We quantified on average >900 high-confidence membrane proteins (FDR=0.05) per cell type, comprising ∼40% of all proteins identified in each experiments. Principal component analysis identified unique cell surfaceome signatures that distinguish several of the leukemia subtypes and normal B-cells from each other, implying different cell-surface phenotypes associated with specific B-ALL genetic alterations. As positive controls, we found several proteins differentially expressed in MLLr that were also previously identified based on transcriptomic profiling of MLLr patient samples, such as FLT3 and PROM1. In our data, the MLLr surfaceome was surprisingly enriched for Ephrin receptors, matrix metalloproteases, and other proteins that play a role in cell motility, invasion of tissues, and modulation of the tumor microenvironment. We are currently validating several of these receptors by flow cytometry and have thus far found excellent agreement with proteomic data on cell lines. We will next move to patient-derived xenografts and tissues for further validation. These findings suggest these novel markers may be candidates for downstream immunotherapy development, potentially specific to MLLr B-ALL, where there is particularly high clinical need. Overall, our work demonstrates the power of performing un-biased interrogation of the tumor surfaceome to identify new therapeutic targets for immunotherapy, as well as provide a resource to advance our phenotypic understanding of these aggressive cancer subtypes.
A.24
Reliable protein quantification in the OpenMS software framework
Lars Nilse, Oliver Schilling, Martin Biniossek
Institute of Molecular Medicine and Cell Research, Freiburg University, Freiburg, Germany
Introduction: Quantitative LC-MS based proteomics has become a standard technique in many biological and clinical studies. The reliable detection of MS1-based peptide features is one key algorithmic step in the analysis of such data. We developed and optimized such a feature detection algorithm within the OpenMS software framework. The detection becomes increasingly difficult as fold changes increase and peptide abundances decrease. Often it is far from clear which of the reported protein fold changes can be trusted and which ones should be taken with a grain of salt. Objectives: In this study we test the reliability of protein quantification within the OpenMS framework in a systematic way. We demonstrate that independent of the mass spectrometry hardware used, OpenMS workflows allow for the correct characterization of proteomes over a wide dynamic range of abundances and fold changes. But our study also investigates the limits of modern mass spectrometry hardware and software analysis. Materials and Methods: Our study focuses on two very different proteome samples: a titration series of dimethyl-labelled E. coli samples with fixed fold changes, and a SILAC-labelled human SH-SY5Y cell line in undifferentiated and differentiated form. Both samples are measured on three different machines: Bruker Impact II, Sciex TripleTOF 6500 and Thermo Q Exactive. While the fold changes in the E. coli sample are a priori know, we cross-check selected protein quantifications in the SH-SY5Y dataset with SRM and Western blotting. Results: Shotgun proteomics is stochastic in nature. Consequently, the three lists of reported protein fold changes do differ. But proteins common in all three lists show a good agreement over a wide dynamic range. Our results are corroborated by SRM measurements and traditional Western blotting. Conclusion:OpenMS is a major, versatile software platform. Experimental data from different mass spectrometry machines can be reliably analyzed within the same software framework. That makes it especially suited for large projects and core facilities.
A.25
Engineering peptide ligase specificity to enable unbiased global sequencing of proteolytic cleavage sites
Amy Weeks, James A. Wells
University of California, San Francisco, CA, USA
Enzyme-catalyzed peptide ligation is a powerful tool for site-specific protein bioconjugation and N-terminal capture, but stringent enzyme-substrate specificity limits its utility. We developed a proteomic approach for comprehensive characterization of ligase specificity. This method utilizes database-searchable, proteome-derived peptide libraries as ligase substrates, enabling selective isolation of ligated peptides, sequencing by tandem mass spectrometry, and rapid determination of positional enrichment or de-enrichment at each site. We used this strategy to characterize the ligation efficiency for >20,000 enzyme-substrate pairs in the context of the engineered peptide ligase subtiligase, leading to the identification of variants with activity toward sequences that were previously recalcitrant to modification. We have applied these mutants individually for site-specific bioconjugation of purified proteins and in algorithmically selected combinations for global sequencing of proteolytic cleavage sites with reduced bias in captured sequences. These studies establish a platform for defining and engineering the specificity of peptide ligases, and greatly expand the toolbox of enzymes available for site-specific modification of protein N termini.
A.26
Mitochondrial ATP synthase disorders investigated by quantitative proteomics of CRISPR-Cas9 knockout cell lines
Marek Vrbacky1, Jana Kovalcikova1, Karel Harant2, Alena Pecinova2, Josef Houstek1, Tomas Mracek1
1Institute of Physiology, Czech Academy of Sciences, Prague, Czech Republic; 2BIOCEV & Faculty of Science, Charles University, Prague, Czech Republic
Mitochondria are cellular organelles producing the vast majority of the key energetic molecule ATP and the mitoproteome is composed of approx. 1500 proteins. ATP is synthesized by the mitochondrial multi-subunit ATP synthase that together with the respiratory chain constitutes the oxidative phosphorylation (OXPHOS) system. Mutations in proteins affecting OXPHOS function comprise broad group of mitochondrial diseases and among them, disorders of ATP synthase belong to the most severe. The majority of ATP synthase disorders are caused by the mutations in the nuclear TMEM70 gene encoding mitochondrial TMEM70 transmembrane protein localized in the inner mitochondrial membrane and involved in the biogenesis of the eukaryotic ATP synthase. TMEM70 is expressed at very low levels as compared with structural subunits of ATP synthase and its molecular function is not yet known. To better understand the role of this factor, we generated TMEM70 knockout HEK293 cell lines by CRISPR-Cas9 technology. Several biochemical methods, including blue native electrophoresis, ATPase enzyme activity assay, and mitochondrial oxygen consumption revealed a defect of ATP synthase, which was fully restored upon expression of wt TMEM70. Label-free (LFQ) as well as SILAC quantitative MS analysis of knockout cell lines showed an isolated decrease of all subunits of mitochondrial ATP synthase. Co-immunoprecipitation performed as an affinity enrichment LFQ MS analysis of cellular lysates identified a few mitochondrial OXPHOS complex I proteins as interaction partners of TMEM70. Collectively, TMEM70 knockout cell lines recapitulate the pathology observed in fibroblasts from human patients and demonstrate that CRISPR-Cas9 knockouts and quantitative MS are promising tools in studying human pathologies. This project is supported by the GACR (14–36804G) and the AZV (16–33018A).
A.27
Modification site-specific pathway analysis applied to phosphoproteomes detects unique perturbation signatures that are not accessible to gene-centric pathway analysis
Karsten Krug1, Philipp Mertins1, Bin Zhang2, Peter Hornbeck2, Rajesh Raju3, Rushdy Ahmad1, Matt Szucs1, Pablo Tamayo4, Jill P. Mesirov4, Jacob D. Jaffe1, Steven A. Carr1, D. R. Mani1
1The Broad Institute of MIT and Harvard, Cambridge, MA, USA; 2Cell Signaling Technology, Beverly, MA, USA; 3Rajiv Gandhi Centre for Biotechnology (RGCB), Kerala, India; 4University of California, San Diego, CA, USA
Post translation modifications (PTMs) of proteins play a crucial role in practically every cellular process by regulating activity, localization and interaction of proteins. Mass spectrometry (MS)-based proteomics facilitates profiling of tens of thousands of PTM sites in a single experiment, most importantly phosphorylation, acetylation and ubiquitylation. Databases annotating biological pathways such as KEGG, Reactome or MSigDB are gene centric and do not capture information for individual PTM sites. Therefore, a first step in pathway analysis of PTM datasets combines measurements of PTM sites detected on the same protein into a single measurement representing the corresponding gene, thereby, neglecting the additional information provided by individual PTM sites. The lack of appropriately curated databases has impeded development of bioinformatics analysis tools that leverage site-centric PTM signatures. We have now developed PTMsigDB, a curated database of modification site-specific signatures. A unique feature of PTMsigDB over other databases is the annotation of each PTM site with quantitative information, i.e. increased or decreased expression upon a specific perturbation or in a pathway. We further present a modified version of the widely used single sample Gene Set Enrichment Analysis (ssGSEA) approach that enables PTM Signature Enrichment Analysis (PTM-SEA). The current version of PTMsigDB (v1.3) comprises 357 signatures of which 114 are perturbation signatures, 207 kinase-substrate signatures and 36 signatures of molecular pathways. To assess the utility of our approach, we applied PTM-SEA to phosphoproteomes of various cancer cell lines treated with different perturbagens. The application of PTM-SEA to the phosphoproteome of human breast cancer cells (T47D) treated with DMSO and a specific PI3K inhibitor for 6h and 24h, respectively, detected signatures of PI3K, MEK and EGFR inhibition in increased significance in the 24h timepoint compared to 6h treatment. We also detected anti-signatures of certain signaling pathways such as ghrelin, GLP1 and leptin pathway pointing to an inhibition of these pathways upon PI3K inhibition. Importantly, no significant signatures were detected in the DMSO treated control experiments demonstrating the specificity of signatures in PTMsigDB. We further applied PTM-SEA to a large-scale perturbation phosphoproteome dataset comprising 26 perturbagens applied to three different human cell lines (PC3, HL60, MCF7). Focusing on four compounds that inhibit cell cycle, PTM-SEA detected CDK signatures in anisomycin, GW8510 and irinotecan whereas anti-CDK signatures were detected in cells treated with paclitaxel which is in agreement with the different mechanisms of action of the corresponding compounds. Taken together, PTMSigDB and PTM-SEA present promising new resources to perform PTM-specific signature analysis that can be directly applied to phosphoproteomes to detect novel signatures that are not accessible to gene-centric pathway analysis. Although the current version of PTMsigDB solely consists of phospho signatures, the database can be easily extended to other types of PTMs like ubiquitylation or lysine acetylation. This project was funded by National Cancer Institute (NCI) CPTAC awards U24CA210986, U24CA160034 and U24CA210979.
A.28
Identification of Novel Autoantigens in Autoimmune Encephalitis
Giselle Knudsen1, Caleigh Mandel-Brehm1, Chris Liverman2, Yvette Schollmeier3, Brian D. O'Donovan1, Trung Huynh1, Sam J. Pleasure1, Michael R. Wilson1, A. L. Burlingame1, Joseph L. DeRisi1
1University of California, San Francisco, CA, USA; 2University of New Mexico Medical School, Albuquerque, USA; 3Roche, in Mannheim, Baden-Wurttemberg, Germany
Significance: Over half of all encephalitis cases worldwide have no known etiology (Pillai SC Pediatrics 2015; Singh Neurol. 2015; Glaser CID 2006; Olsen EID 2015 and references therein). Treatment depends upon a key decision point between infection (commonly viral or bacterial) vs. sterile inflammation, mainly due to autoimmune disease. Current methods for diagnosis are largely candidate driven, both to detect the presence of individual pathogens (for which hundreds of pathogens are known to cause encephalitis) and for detecting the presence of known autoantibodies to targets such as the NMDA receptor (Lee, J. Epilepsy Res 2006). Our research efforts in autoimmune disease diagnosis are focused on developing novel detection methods for identifying the targets of autoantibodies from patient serum and cerebrospinal fluid (CSF) samples. The outcomes of this study include novel markers and improved methods for diagnostics, as well as insights into the molecular mechanisms for how autoantibodies participate in disease. Our diagnostic method prioritizes patient samples with specific immunohistochemical staining of mouse brain tissue. These sera and CSF samples are then further probed by immunoprecipitation-mass spectrometry (IPMS) using protein A/G capture and mouse brain lysates. Captured proteins are digested with trypsin then subjected to LC-MS/MS peptide sequencing for protein identification. To identify specific interacting proteins, the captured proteins are compared to samples obtained from patients with unrelated neuroinflammatory disorders as well as healthy donor controls. Identified proteins are ranked by Z-score as well as SAINT scoring. To date, several clinically validated autoimmune markers have been recapitulated with our method, including anti-VGKC (voltage gated potassium channel, one marker for limbic encephalitis) and the antigenic target for anti-Hu encephalitis antibodies. Future directions: Work in progress is aimed at sequencing the major autoimmune IgG antibody species in one of our patients to enable cloning and recombinant expression of a specific target IgG for validation assays.
Acknowledgements: Mass spectrometry analysis was provided by the UCSF Mass Spectrometry Facility directed by A. L. Burlingame, supported by 8P41GM103481 and the Adelson Medical Research Foundation. This work was also supported by the Sandler Family Foundation and the William K. Bowes, Jr. Foundation.
A.29
Combined Top-down and Bottom-up Proteomics using Capillary Electrophoresis–Mass Spectrometry
Chien-Hyun Chen, Aaron Gajadhar, Ioanna Ntai, Andreas F. R. Huhmer
Thermo Fisher Scientific, San Jose, CA, USA
Bottom-up proteomics has been widely used for resolving the sequence of a single protein or components of protein complexes. Bottom-up approaches for single protein characterization enable the localization of PTMs via digested peptide; however, the stoichiometry of PTMs are not able to be resolved easily. Top-down approaches have proved to be a better solution for understanding combinatorial PTM profiles. A combined top-down and bottom-up proteomics workflow is a promising solution for comprehensive protein characterization. Current analytical strategies mainly rely on liquid chromatography–mass spectrometry (LC-MS) for top-down and bottom-up analysis. However, protein separation used for top-down and peptide separation used for bottom-up requires different LC platforms; therefore, the combined workflow using LC-MS is usually time consuming. In this study, we introduce a new platform of capillary electrophoresis–mass spectrometry (CE-MS) for proteomics research. On this platform, intact proteins and peptides can be monitored on the same microfluidic CE chip, and top-down and bottom-up analysis can be combined on one CE-MS platform. The whole workflow only takes 18 min including 3 min for intact analysis, 3 min for top-down analysis and 8 min for bottom-up analysis. Two model proteins cytochrome c (12.4 kDa) and carbonic anhydrase (29.0 kDa) have been investigated on this platform. For cytochrome c, 56% residue cleavage from top-down and 97.1% sequence coverage from bottom-up was observed with high-energy collisional dissociation (HCD). In the case of carbonic anhydrase, 19% residue coverage from top-down and 96.5% sequence coverage from bottom–up was detected. These model proteins demonstrated the capability for comprehensive top-down and bottom-up analysis on this CE-MS platform within 18 min. To demonstrate the utility of this combined platform, we utilized human recombinant KRAS (21.2 kDa) with a known G12D oncogenic mutation and PD-L1 (33.3 kDa), both expressed in HEK293 cells to investigate disease relevant PTMs. KRAS plays an important role in the regulation of cell proliferation and promoting oncogenic progression. It has abundant PTMs including phosphorylation, acetylation, and ubiquitylation. PD-L1 is involved in the suppression of the immune system and is frequently upregulated in numerous human tumors. It is a transmembrane protein known to contain glycosylation, acetylation, and phosphorylation. The localization and stoichiometry of multiple PTMs are expected to be acquired on this CE-MS platform with combined top-down and bottom-up workflow in a timely manner. Overall, this novel platform represents a universal approach for comprehensive analysis of any protein of health and life science interest.
A.30
AltitudeOmics: Red Blood Cell Metabolic Adaptation to High Altitude Hypoxia
Travis Nemkov1, Angelo D'Alessandro1, Andrew A. Monte1, Andrew W. Subhudi2, Andrew T. Lovering3, Kirk C. Hansen1, Robert C. Roach1
1University of Colorado Denver, Aurora, CO, USA; 2University of Colorado, Colorado Springs, CO, USA; 3University of Oregon, Eugene, OR, USA
Adjustment to hypoxia is a key challenge for millions of people either living at or traveling to high altitudes. The physiological process of acclimatization enables adaptation to characteristically low oxygen levels. Exposure to acute hypoxia for unadjusted individuals results in a marked drop in exercise capacity, impaired cognitive function, and possible loss of consciousness, all of which can be reversed upon acclimatization. The AltitudeOmics research program was designed to provide insights into acclimatization and retention of these adaptations upon reascent to high altitudes. Mechanisms of acclimatization can be studied through the correlation of physiological and neurocognitive readouts with data obtained from omics technologies including metabolomics, proteomics, epigenomics, and transcriptomics on key biological matrices. Here, we assessed the metabolic response of red blood cells (RBCs) to high altitude hypoxia. Insights gained from understanding RBC response to hypoxia not only provide a mechanistic understanding of RBC biology, but also can be applied to practices of transfusion medicine and treatment of trauma/hemorrhagic shock-induced hypoxemia. Exposure to high-altitude hypoxia triggers dramatic metabolic modulation of RBCs in humans. These metabolic adaptations could contribute to molecular mechanisms of acclimatization that could be retained upon later reascent, thus expanding the central role of erythrocyte metabolism in mitigating problematic effects of prolonged exposure to acute hypoxia.
A.31
Sharing Annotated Spectra when Standard Formats are not an Option
Robert J. Chalkley, Peter Baker
University of California San Francisco, CA, USA
The requirement by journals for authors to make available annotated proteomic results has been driven by the recognition that authors cannot look at all of their reported results, so by sharing data associated with a manuscript, readers can make an independent assessment and reanalysis of results of particular interest. This process has been heavily facilitated by the development of standard file formats. However, for some search engines or results types, outputs in these formats are not an option, presenting a challenge for authors to comply with journal guidelines. MS-Viewer, part of the free Protein Prospector package, can produce annotated spectra from many search engine outputs, including simple tab-delimited text files. This allows it to annotate data from search engines that do not support standard formats. It can also annotate cross-linked peptide identifications. A website allows submission of results to a public repository, and there is a searchable browser that allows one to see all submissions associated with publications. There are links to associated proteomeXchange submissions, and PRIDE provides links from datasets in their repository to MS-Viewer. MS-Viewer has been used to publish results associated with about fifty publications in ten different journals, and is also used for sharing results between laboratories. It is particularly popular with MaxQuant users, as it allows production of annotated spectra from much smaller file uploads than through using the MaxQuant Viewer and can handle large datasets that the MaxQuant Viewer struggles to display. Working with authors, we have successfully produced annotated spectra from all submissions.
Acknowledgment: Supported by NIH grant 8P41GM103481 and Sheldon G. Adelson Medical Research Foundation.
A.32
Characterizing covalent protein-DNA adducts in the human cell
Emily Myers1, Kostantin Kiianitsa2, Nancy Maizels2,3,4, Shao-En Ong1
Departments of 1Pharmacology, 2Immunology, 3Biochemistry and 4Pathology, University of Washington, Seattle, WA, USA
A subset of very potent cancer drugs target proteins that form covalent bonds with DNA as obligatory reaction intermediates, trapping those intermediates on DNA and forming cytotoxic protein-DNA adducts. Screening for new bioactive drugs that work through this mechanism is challenging as isolation and quantification of protein-DNA adducts is labor intensive and slow. To address this, we developed RADAR (Rapid Approach to DNA Adduct Recovery), an ELISA based assay to rapidly and sensitively detect covalent DNA adducts. Here, we describe RADAR-SILAC, a method combining unbiased quantitative mass spectrometry with RADAR to identify endogenous and drug-induced protein-DNA adducts. As proof of concept, we tested two cancer drugs previously shown to trap protein-DNA adducts, topotecan and etoposide, against a lymphoblast leukemia cell line (CCRF-CEM). These chemotherapeutics target topoisomerase 1 and topoisomerase 2A, respectively. Briefly, CCRF-CEM cells are SILAC labeled with heavy arginine and lysine or normal isotope abundance amino acids for five doublings. One cell population is treated with either 10 μM topotecan or 50 μM etoposide for 15 min. Cells are lysed in buffer containing 5 M guanidinium isothiocyanate, 2% Sarkosyl, 0.1 M sodium acetate, 20 mM EDTA, 20 mM Tris-HCl (pH 6.5), and 10 mg/ml DTT. Lysates are brought to 2 M LiCl and nucleic acids are precipitated with isopropanol. Pellets are solubilized in 8 mM NaOH, treated with RNAse A, mixed with lysis buffer and re-precipitated with isopropanol. The isolates from control and drug-treated cells are mixed 1:1 before treatment with Benzonase nuclease. Samples are adjusted to 8 M urea, reduced with 1 mM TCEP and alkylated with 2 mM chloroacetamide. Proteins are digested with trypsin and desalted using C18 StageTips. Peptides are analyzed with 90 min gradients of 10–30% acetonitrile at 200 nL/min on an Orbitrap Elite. MaxQuant v.1.5.0.30 and the associated Andromeda search engine was used to search a Uniprot human database. Data were analyzed with the Perseus and R environments. In a series of two-label swap biological replicate experiments, we found TOP1 consistently enriched on DNA from cells treated with topotecan compared to DMSO. Interestingly, two topoisomerases, TOP2A and TOP2B, are enriched with etoposide treatment over control. The SILAC ratio was 4-fold higher for TOP2A than TOP2B, possibly suggesting higher selectivity of etoposide for TOP2A in vivo. We also identified peptides from SUMO enriched in the etoposide treated samples, consistent with previous reports that TOP2A and TOP2B-DNA adducts may be SUMOylated. These results indicate our approach is promising for sensitive and specific identification of targets of drugs that trap protein-DNA adducts. We believe our RADAR-SILAC method can be used to identify new drug-target pairs that act through this mechanism and provide insights into the mechanism of action and side effects of these drugs.
A.33
Unbiased phosphoproteomics reveals a therapeutically-relevant connection between pre-mRNA splicing and proteasome inhibition in multiple myeloma
Hector H. Huang1, Christine Lam1, Alexis Thornton2, Margarette Mariano1, Ian Ferguson1, Byron C. Hann1, Angela N. Brooks2, Arun P. Wiita1
1University of California, San Francisco, CA, USA; 2University of California, Santa Cruz, CA, USA
Proteasome inhibitors (PIs) are first-line therapeutic agents for the blood cancer multiple myeloma. While these small molecules have revolutionized myeloma therapy, nearly all myeloma patients treated with PIs eventually experience drug resistance and relapse. We have previously examined the global effects of PIs on the transcriptional and translational landscape of multiple myeloma plasma cells (Wiita et al, eLife (2013) 2:01236). Here, we hypothesized that investigation into the intracellular signaling response to PI-induced stress would help elucidate additional PI mechanisms of action in this cancer. We aim to reveal novel therapeutic strategies to potentiate PI therapy or overcome resistance.
We initially applied an unbiased phosphoproteomic screen to MM.1S myeloma cells treated with the PI carfilzomib across a 24-hour time course. We used IMAC enrichment chromatography and a single-shot 4-hour LC method at each time point on a Thermo Q-Exactive Plus, with phosphopeptide identification and label-free quantification performed in MaxQuant. Among the more than 5,000 phosphopeptides quantified, the most profoundly upregulated phosphosites mapped to numerous serine-and-arginine rich (“SR”) core components of the pre-mRNA spliceosome. Simultaneous mRNA-seq confirmed no change in splicing factor abundance in response to PI. Targeted quantification using parallel reaction monitoring (PRM) on SILAC-labeled, carfilzomib treated MM.1S cells, as well as another multiple myeloma cell line, AMO1, at 24 hr of treatment confirmed the novel connection between proteasome inhibition and splicing factor modulation. We further used paired-end mRNA-seq analyzed with JuncBASE software to correlate changes in splicing factor phosphorylation with alternative splicing events. Intriguingly, intron retention appeared to be a dominant signature in pre-mRNA splicing following proteasome inhibition.
This connection led us to investigate whether induction of intron retention may be sufficient to lead to myeloma tumor cell death. We therefore obtained the clinical candidate E7107, a specific inhibitor of the core spliceosome factor SF3B1, known to strongly induce intron retention across the transcriptome. We found that a panel of myeloma cell lines were highly sensitive to E7107 at low nanomolar concentrations, which are easily achievable in patients based on prior Phase I studies. We further found evidence of synergy in vitro between E7107 and PI treatment, and E7107 was equally efficacious in naïve and PI-resistant AMO1 cells. Furthermore, pre-clinical experiments with disseminated murine xenograft models of myeloma showed a brief two-week treatment with E7107 significantly decreased tumor burden and significantly increased lifespan.
Our approaches here reveal a novel connection between PI-induced stress and modulation of the core spliceosome machinery not previously observed in any other system. By capitalizing on this vulnerability, we further propose that targeting the spliceosome machinery may be a novel therapeutic option for multiple myeloma patients, either in combination with PIs or in the PI-refractory setting, addressing a major clinical need.
A.34
Unraveling White Lupin's Signal Transduction in Response to Phosphorus Deficiency using iTRAQ Labeling, Phosphopeptide Enrichment, and Tandem Mass Spectrometry
Michael Amadi1, Claudia Uhde-Stone1, Kathy Li2, Robert J. Chalkley1, A. L. Burlingame1
1California State University, East Bay, Fremont, CA; 2University of California, San Francisco, CA, USA
Phosphorus (P) availability is one of the most limiting factors for crop production worldwide. To breed or engineer crops that do well in poor soils, researchers are investigating plants that are well adapted to P deficiency, such as the legume white lupin (Lupinus albus L.). While great progress has been made in understanding white lupin's strategies to acquire P, not much is known about the signal transduction pathways that link sensing of P deficiency with the appropriate responses. Phosphorylation of proteins plays a central role in signaling processes. Proteins that change phosphorylation status when exposed to P deficiency are potential members of such signal transduction pathways. However, quantifying the dynamics of protein phosphorylation is challenging, because of their transient nature and low abundance. A sensitive method is needed to identify proteins that become phosphorylated after exposure to P deficiency. To this end, white lupin was grown in hydroponics with sufficient phosphorus (+P). After 3 weeks, half of the lupin plants were transferred into P-deficient nutrient solution (-P). After 24 hours, total protein was isolated from roots of two +P and two –P24h plants, for a total of two independent biological replications. We used iTRAQ 4-plex tags to label the +P and –P peptides from the two replicates. From these, we enriched phosphorylated peptides using titanium dioxide chromatography and performed UPLC-MS/MS to quantify the iTRAQ-labeled phosphorylated peptides. Using a translation of cDNA as a database reference, from phosphopeptide enriched and flow-through samples we identified 111 phosphorylation sites and around 350 proteins. These results included a phosphopeptide that was 1.78-fold more abundant in –P, compared to +P. Bioinformatic analysis of the corresponding cDNA and the translated protein identified this protein as a likely leucine-rich repeat receptor kinase. The protein contains a signal peptide, a transmembrane-spanning domain, an extracellular leucine-rich receptor domain, and an intracellular kinase domain. We are currently performing follow-up experiments to confirm thatthe phosphorylation status of this receptor kinase increases in response to P deficiency, rather than a change of protein expression. If confirmed, the next steps will include identification of the ligand binding to the receptor domain as well as the substrate that becomes phosphorylated by this kinase. The results presented here offer a potential starting point to further unravel the signal transduction cascade and regulatory network that link sensing of P deficiency to the appropriate plant responses.
Acknowledgment: Supported by NIH grant 8P41GM103481 and Sheldon G. Adelson Medical Research Foundation.
A.35
Identification of Lysyl Oxidase Generated Protein Crosslinks in Collagen
Kirk Hansen1, Alex Barrett1, Peter Baker2, Robert J. Chalkley2
1University of Colorado Denver, Aurora, CO, USA; 2University of California, San Francisco, CA, USA
The collagen family of proteins, the most abundant proteins in the human body, represent a basic building block within nearly every tissue and organ and are crosslinked by the lysyl oxidase (LOX) family of enzymes. The LOX family of ECM crosslinking enzymes (LOX & LOXL1–4) catalyze the conversion of lysine (Lys) residues to a semialdehydes (allylysine). This reactive side-chain further reacts with Lys or hydroxylsine (Hyl) to form a di-valent crosslink, or another allylysine to form an aldol-product that goes on to react with a Lys or Hyl to form tri-valent crosslinks or another aldol product to make tetra-valent desmosines. LOX enzymes have been shown to play a role in the generation of a pathologic tissue microenvironment in the progression of several diverse diseases and have been shown in animal models to limit reversal of fibrotic disease, promote tumor progression and cause preconditioning of distant organs prior to metastatic dissemination. In addition, a strong relationship between crosslinked products and tissue stiffness has been shown. However, very little is known about the substrate specificity of the LOX family members or resulting architectures of modified collagens. An analytical and data-analysis workflow was developed to identify in vivo enzymatic crosslinks from tissue samples. We utilized a decellularization procedure to isolate insoluble, crosslinked, extracellular matrices (iECM) from bovine and human tissues. These iECM fractions were subjected to chemical and enzymatic digestion and the resulting peptides were fractionated by cation exchange or size exclusion chromatography to enrich for crosslinked peptides. Protein Prospector was used to interrogate LC-MS/MS data and the results were used to refine the analytical workflow. We focused our analysis efforts on the N-terminal crosslinked peptide of collagen I (NTX) as this peptide is being followed in the clinic to monitor disease of the bone and connective tissues. A wide range of crosslinked NTX species were identified that differed in crosslinked amino acids, proline hydroxylation and other PTMs. It is now possible to search multiple LOX generated crosslinks in one search using Protein Prospector. The results generated here are being used to develop a spectral-library search routine to facilitate crosslinked peptide identification so that the relationship between crosslink species and resulting ECM architectures and biomechanics can be explored at the molecular level. Our goal is to study the specificity of the individual LOX enzymes to better understand their role in disease progression.
Acknowledgment: Supported by NIH grant 8P41GM103481 and Sheldon G. Adelson Medical Research Foundation.
A.36
Top-down strategies to better localize post-translational modifications of integral membrane proteins
Julian Whitelegge
University of California Los Angeles, CA, USA
Top-down mass spectrometry has great potential to precisely locate post-translational modifications while retaining the context provided by the whole protein. In this way we may start to address complex logic enabled by multiple modifications. For top-down mass spectrometry samples from liquid chromatography are analyzed off line to yield the full benefits of modern Fourier-transform instruments. Membrane protein preparations are typically delipidated using organic solvent precipitation and subjected to reverse-phase chromatography with a split eluent line for fraction collection and low-resolution electrospray-ionization mass spectrometry (LC-MS+). Data-directed top-down FT-ICR MS is performed on selected fractions from the primary separation using LC-MS+ derived information. Both MS3 and middle-down approaches maybe necessary for precise localization of post-translational modifications. While sample preparation and chromatography conditions do not favor native structure, it is likely that transmembrane alpha-helices remain folded and once ionized remain quite stable within the mass spectrometer. Thus, it was previously shown that ion activation was necessary for effective ECD analysis of a two-transmembrane helix membrane protein (Zabrouskov and Whitelegge, 2007, J Proteome Res. 6(6): 2205–10). The rational is that while extensive ECD may be occurring, hydrogen bonding within alpha helices prevents release of product ions. Use of an IR laser irradiation within the ICR cell allowed thermal activation of ions, breakage of hydrogen bonds and appearance of ECD product ions within the mass spectrum. Another way to improve efficacy of ECD is to work with more highly charged ions. Consequently, we have investigated the use of a supercharging reagents, such as propylene carbonate, in ECD analysis of a proteins with one or more transmembrane helices. While little supercharging was observed there was a clear improvement in yield of product ions during ECD, presumably due to denaturation of the protein prior to ionization. Strategies to maximize sequence coverage in top-down membrane protein analysis will be compared.