Oral and Craniofacial Sciences Faculty






















The proteome can be viewed as the total content and state of protein in a living system. The proteome is dynamic, responding to changes in the cellular environment caused by normal and pathogenic physiological states associated with any aspect of cellular function or disease process. The sequencing of the human genome was a significant benchmark in scientific endeavor. From this genetic archive, the regulation of transcription, translation, processing, and post-translational modifications offer many levels of control that ultimately result in the state of the cellular proteome. Work in the Mass Spectrometry facility centers on applying the tools of proteomics to reveal the dynamic qualitative and quantitative state of protein associated with cellular functions. The maintenance and operation of mass spectrometers is central to this work, and as Director of the Biological Mass Spectrometry and Proteomics Facility, my goal is to make this technology accessible, understandable, and above all, useful to the UMKC educational and research community.

My own research is currently focused on two proteomics challenges. First, the problem of identifying proteins in very divergent organisms: In proteomics facilities worldwide, protein identification is routinely accomplished by comparing mass spectrometry data to predicted, virtual spectra, based on the known protein sequences held in protein databases. In fact, this strategy has actually been called, “looking up the answer in the back of the book.” Peptides have fairly predictable fragmentation behavior in mass spectrometers, and this simple comparison efficiently finds matches in the MS data, but only if the sequence is represented in the protein database. It is an ideal bioinformatics problem for computer automated searching. Obviously, success based on this paradigm requires that the MS data of peptides from the unknown protein is represented in the database, which is true for commonly studied species. In fact, even moderate sequence divergence, like polymorphism that can be found in some peptides of a protein is typically compensated for by the presence of other conserved peptides in the same protein, allowing the protein identification to be made. More extreme differences, like those encountered in very divergent species not present in the database, however, can severely diminish confidence in protein identifications by reducing the number of matching peptides. Interestingly, we can correctly predict the sequence of unknown, naturally occurring proteins based on well documented substitution patterns exhibited by the known proteins from sequenced organisms. Using this strategy, we have designed a web-based virtual sequence database engineering program that successfully predicts naturally occurring protein sequences that are not represented in the protein databases, but are present and identifiable in mass spectrometry data (http://ms-virtmorph.umkc.edu/). We are developing this technology to improve protein identification in divergent proteins that are not present in the existing database.

The second area of interest in this laboratory is differential expression analysis using non-labeling strategies. Various methods of isotopic or affinity labeling methods have been applied towards the problem of detecting and quantifying differential expression between samples (control versus experimental). Stable isotopes are quite useful for this purpose since chemical characteristics are not changed in isotopically labeled peptides; peptides from labeled or unlabeled samples behave identically in chromatographic separations, differing only in measurable mass, which is the basis of the separate quantitative measurements. Affinity tagging and isobaric labeling (iTrac) rely on chemistry which can be quite effective in many circumstances, but also quite expensive. The use of endogenous peptides for quantitative loading control (non-labeled quantitation) has two major advantages: minimal sample handling, and major cost savings. isotope labeling (patient samples). Furthermore, sometimes it is not possible to apply uniform isotope labeling (patient samples, for instance), and no internal quantitative control can be more accurate, on average, than the unmodified peptides present in many proteins in a cell that do not change expression level (internal quantitative control). This laboratory has a significant initiative to develop standard methods for preparing and analyzing proteomes for differential expression analysis to facilitate biomarker discovery and basic research in various cellular processes.

Andrew KeightleyAdjunct Faculty

Contact Info