FTMS for Eukaryotic Proteomics, Enzymology, and Structural Biology

General

Our laboratory has thrusts in three main areas: custom instrumentation for Fourier Transform Mass Spectrometry (FTMS), Nuclear Signaling and Natural Products. More specifically, our main interests lie in the enzymology of natural product biosynthesis, mass spectrometric-based studies of the "Histone Code," and development of Fourier Transform Mass Spectrometry (FTMS) for Top Down Proteomics (i.e. analyzing intact proteins directly; no proteases).

A Schematic Overview of the Kelleher Laboratory

A core activity is measuring chemical modifications to proteins in both hypothesis-driven and discovery modes. Our pioneering efforts in "Top Down" proteomics involve fragmenting intact protein ions in the gas phase and developing custom bioinformatics to characterize unexpected post-translational modifications (PTMs) in methane-producing microbes, yeast, and human cancer cells. In both human cell biology and antibiotic biosynthesis, key proteins harbor over 20 PTMs that present a "code" of biological logic written in the language of protein modifications. We construct, automate, and apply custom mass spectrometry and algorithms to detect and decode this logic.

We currently have 13 graduate students, 2 senior associates, 4 postdoctoral associates, and 9 undergraduate students.

Instrumentation for Proteomics using Intact Protein Ions

We are developing a general measurement platform dedicated to the efficient detection of covalent modifications to proteins. To this end, our primary enginer is Electrospray Ionization combined with a 9.4 Tesla Quadrupole-FTMS Hybrid and the database retrieval software contained in a Web-accessible environment [ProSight PTM]. This "Q-FTMS" instrument enables the interrogation of intact proteins directly by tandem MS (MS/MS) and the mechanistic studies of 100-700 kDa enzymes with complex arrays of covalently-bound intermediates. We are targeting and discovering systems where protein mass information can be translated into enzyme functional or mechanistic insight.

Our short-term goal is to apply this custom platform to characterize the protein inventory of Mycoplasma pneumoniae, Methanococcuss jannaschii, and S. cerevisiae with unprecedented molecular detail to elucidate the structure and biogenesis of unexpected (even novel) covalent modifications. In the long run, we aim to collect data for a more complete understanding of post-translational modifications (PTMs) to human proteins [HUPO] and their role in cellular dynamics. As we emerge from a "tooling up" phase, we will begin pursuing biochemical and genetic follow up work to contextualize the most interesting protein features we uncover (e.g., novel PTMs).

The interdisciplinary nature of the program requires a lab outfitted with state-of-the-art analytical instrumentation, the tools of biochemistry/molecular biology, and advanced use of algorithms and databases. To date, we have established the infrastructure and conceptual framework to achieve the high throughput analysis of proteins in a unique fashion. Ours will be the first lab to realize a robust implementation of the "Top Down" approach (See above figure for example) and will therefore be positioned competitively for discoveries of novel processing events to yeast and human proteins over the next few years. We seek to increasingly automate and refine a unique measurement platform which includes three major aspects (described below click here for a schematic overview):

1. The "Front End" (from cells to mass spectrometer, Click here for paper): protein solubilization and 2-Dimensional (2-D) fractionation using continuous elution Gel Electrophoresis (GE), an Acid-Labile Soap (ALS), and Reversed-Phase Liquid Chromatography (RPLC). This new ALS-PAGE/RPLC approach produces fractions each containing proteins within a known molecular weight range (~5 kDa)

1. The "Instrument" : automated acquisition of protein ion fragmentation data using a nanospray robot coupled to a Quadrupole-Fourier Transform Hybrid Mass Spectrometer (Q-FTMS). (See Q-FTMS Schematic Here)

1. The "Back End" : data processing/protein identification programs and protein characterization environments customized to facilitate whole protein analysis (web-based, accessible to a broad community).

FTMS-Based Enzymology on polyketide and non-ribosomal peptide synthetases

The biosynthesis of several clinically-used antibiotics and immunosuppressants occurs on large enzymes (100-1600 kDa) known to have complex arrays of covalent intermediates, ideal for study by FTMS. With 10-20 biosynthetic intermediates bound to these enzymes via acid-stable thioester linkages, natural product assembly is thought to occur in a "conveyor belt" style. By targeting polyketide and non-ribosomal peptide synthetases (NRPS) involved in erythromycin and gramicidin S production, respectively, we will gain detailed insight into this "thiotemplate mechanism". By measuring the MW values and ratios of the peptides carrying the various biosynthetic intermediates vs. time, we are poised to generate unique data in this field. The high performance of FTMS (e.g., ultrahigh resolving power and mass accuracy) are crucial for extracting the kinetic data from complex proteolytic mixtures (>100 peptides per spectrum) of such large enzymes. Beyond the basic science, detailed mechanistic models would be invaluable in the design of nonnatural enzymes for (combinatorial) generation of new bioactive compounds by polyketide and peptide syntheses.

For the first time, the percent occupancies of intermediates at multiple carrier sites are being correlated with one another to decipher between overall "crawling" vs. "fast translocation" kinetic models of how these enzymes function. Our technology allows us to semi-quantitatively determine the occupancy and heterogeneity at multiple, specific carrier sites with enough efficiency to perform kinetics. Focusing on the biosynthesis of Yersiniabactin (See Figure Below) and Epothilone (an antitumor compound in Phase III trials), we are illustrating the feasibility and generality of using MS as the primary assay.

"Assembly line" biosynthesis of yersiniabactin

We will extend these studies to focus on fully reconstituted systems in vitro, and directly observe the enzymes as they are acylated in vivo. Further, we will use the MS-based approach to access systems where the substrates are unknown and mechanistic issues like "skipping" and "stuttering" of intermediates is hypothesized. In the coming years, our lab will continue to illuminate fundamental mechanisms, elucidate the structure of putative intermediates, and facilitate the design of non-natural enzymes of (combinatorial) generation of new bioactive compounds by NRPS and PKS systems.

The Histone Code

Eukaryotic cells have evolved a complex language to facilitate cell-cell communications, perform signal transduction, and regulate other diverse cellular functions. To the extent this language is conveyed in the post-translational modifications (PTMs) of proteins, we are uniquely positioned to capitalize on its development of a new mass spectrometric methodology to read out this part of cell biology at the molecular level.

The prototypical systems that embody this line of thinking are human histones. Through certain combinations of modifications such as methylation, phosphorylation, and acetylation central processes such as genome packaging/duplication, gene transcription, and epigenetic information transfer are somehow encoded. Our group's most recent breakthroughs in technological development have come from combining histone biology, ultra-high performance mass spectrometry, and computer science.

Visualization of the multiple histone forms by Top Down Mass Spectrometry. Chemical modifications are assigned to specific amino acids by high resolution MS/MS.

Histone proteins serve as a structural scaffold for packaging DNA into the nucleus. Combinations of PTMs on histones H2A, H2B, H1, H3, and H4 create a putative "Code." Essential for complete understanding of this code is an efficient methodology for detection, exact localization, and quantitation of modifications at specific sites. We combine gas-phase concentration of protein ions inside a Quadrupole-FTMS hybrid with Top Down fragmentation using Electron Capture Dissociation (ECD) and large scale PTM prediction. This prediction uses a new type of protein database that has been "shotgun annotated" by assigning site-specific PTMs (and all their combinations) prior to searching for best matches with ECD data. This approach considers PTMs during a database search and now enables automated characterization of human histones harboring specific combinations of modifications from human cells.

We now seek to aggresively extend the findings from our initial work, using only one histone, to all human and yeast histones through the cell cycle and move toward "biological endpoints" through use of diverse experimental methods. For example, we have found that di-methylation of Lys20 on histone H4 is the most abundant PTM and that mono- and di-methylation on newly synthesized histone H4 proceeds through M phase (or very early in G1) in the cell cycle. With Prof. Craig Mizzen (UIUC), we are using RNA interference to knock down a known methyl transferase. There is a cell cycle arrest in these cells with is currently under further study.

Top Down Proteomics

Within the arena of mass spectrometry-based protein analysis, a new type of strategy has been in development in the Kelleher laboratory. What it represents is a major step in the evolution of tandem mass spectrometry to high mass ions.

Overview of Top Down Proteomics

Our measurement platform dedicated to Top Down MS/MS is based on a home-built 9.4 Tesla Q-FTMS and custom software Prosight PTM. We have established the infrastructure to acheive the analysis of intact proteins in a unique fashion. Ours is the first lab to realize this approach on a reasonable scale and will therefore be positioned competitively for discoveries of novel processing events to microbial and human proteins over the next several years.

In our laboratory, both discovery and hypothesis-driven modelsof research are utilized. In hypothesis-driven mode, we either find interesting cases by sequence gazing or target the analysis of specific protein classes known to harbor extensive arrays of PTMs. For discovery, we are focusing on methanogenic Archaea and yeast. We have detected exotic PTMs, such as dimethylated Proline that places a permanent positive charge at the N-terminus of a ribosomal yeast protein. The methyltransferase is not known and is currently being targeted for identification, in vitro reconsitution (using a MS-based assay), and genetic disruption. Such projects are beginning to provide the kind of opportunitites for internal synergy within the enzymology and proteomics subgroups envisioned years ago.

In either mode of operation, a key question has been the extensibility of Top Down MS to higher eukaryotes. Very recently, we have extended the concept of "Shotgun Annotation" to enable the identification and characterization of wild-type human proteins. This includes precise and automated identification of proteins (specific to one member of a gene family), robust genotyping of known polymorphisms, characterization of alternatively spliced transcripts, and detection of diverse modifications. The extension of the Shotgun approach to the Human Proteome is a significant bioinformatic feat, but more importantly begins a Top Down proteome project for synchronized human cells focusing initially on nuclear extracts and transcription complexes.

The ability of top-down mass spectrometry to dissect slightly different molecular forms of proteins will have a significant impact on the future of systems biology and various human proteome projects. Associating specific protein forms with particular phenomena and phenotypes will be a major part of cell and chemical biology for decades to come. A glimpse of this future can be gleaned from our recent work with histones.

As we accelerate past an extensive "tooling up" period, the application of our technology (at much higher magnetic fields) to timely biological drivers is a clear way forward. Throughout the next several years, our laboratory will continue leading the charge for Top Down Mass Spectrometry toward the dissection of histone modifications and enumeration of the "Histone Code." We will continually refine our mass spectrometry, but also embed ever more biological knowledge and skill sets within the group. Specifically, we will extend our Top Down analyses to other human histones and connect these to studies of post-translational dynamics of specific transcriptional complexes through the cell cycle. The main limitations of Top Down MS - low applicability to very large and hydrophobic proteins - will also be addressed on a continuing bases.

Through our new ability to detect allele-specific splicing and modification patterns, we will be able to dissect very subtle molecular distinctions between different protein forms. This takes proteomics to a new level by better capturing the combinatorial dynamics that have evolved to modulate protein activity, in vivo half-life, sub-cellular localization, and non-covalent associations. Such advances are critical to provide a molecular understanding of cell biology and the progression of complex human disease.

Integrated Teaching Initiative

I have developed a course entitled, "Genomics, Proteomics, and Bioinformation: A Survey of Modern Bioanalytical Methods." (See recent syllabus here) (Note : The course number has changed from CHEM/BIOCH 473 to CHEM 574 /MCB 554). Course Perspective.