HHMI-NIH Research Scholars
Learn about the HHMI-NIH Research Scholars Program, also known as the Cloister Program. More

Janelia Farm Research Campus
Learn about the new HHMI research campus located in Virginia. More

Summary: Natalie Ahn integrates technologies of proteomics and mass spectrometry with biochemical and molecular biological methods to study the control of cellular responses by MAP kinases and Rho GTPase pathways.

Our goal is to discover new mechanisms underlying the regulation and function of signaling pathways. To investigate cell responses to signal transduction pathways, we are using mass spectrometry to develop and apply new methods for global profiling. A second focus of our lab is the examination of internal protein motions in protein kinases, demonstrating coupling between enzyme dynamics and catalytic function.

To investigate cancer progression, we are using melanoma as a model, with stages of dysplastic or atypical nevi, early radial growth phase and advanced vertical growth phase primary melanoma, and metastatic melanoma. We are investigating four signaling pathways that are activated in melanoma cells and influence cancer cell behavior (Figure 1). B-Raf, oncogenic in two-thirds of melanoma tissues and cell lines, promotes melanoma cell proliferation and survival, as well as cell migration and invasion. Rho GTPases Rac and RhoA promote cell migration and invasion in melanoma cell lines, and their activation is correlated with stage transitions. The noncanonical Wnt ligand Wnt5A is elevated in advanced melanomas and enhances invasion when overexpressed in melanoma cell lines.

	Figure 1: Signaling pathways dysregulated in melanoma Figure 2: Individual proteins in cell extracts

To analyze molecular responses to signal transduction pathways, we are using various methods for surveying changes in the "proteome"—the total number of expressed proteins in a cell. To do this, we take advantage of the capability of mass spectrometry (MS) to detect and sequence peptides at low femtomole quantities. In one approach, we extract proteins from cells treated to elicit different signal transduction responses, separate them by two-dimensional polyacrylamide gel electrophoresis (2D-PAGE), and visualize them by silver staining (Figure 2). Changes in protein patterns between different cellular states are noted, and proteins altered in expression or gel mobility are identified by in-gel elution, proteolytic digestion, and MS sequencing.

Using this strategy, we analyzed protein changes responsive to activation of RhoA in early nonmetastatic melanoma cell models and then examined changes that occur in cells during progression from primary to metastatic cancer. One target, protein-tyrosine phosphatase 1B (PTP1B), was phosphorylated and inhibited in response to RhoA, resulting in increased tyrosine phosphorylation of the adapter, p130^Cas. Another protein, which has not been characterized before in human systems, is induced in response to RhoA in metastatic cells and mediates RhoA-dependent cell invasion. The behavior of this protein suggests a role in focal adhesion disassembly and turnover, consistent with its ability to promote cell invasion.

In other studies, we are implementing shotgun proteomics technologies by multidimensional LC/MS/MS as a powerful alternative to 2D-PAGE. Proteins in complex mixtures are proteolyzed in solution, and peptides are separated by ion-exchange and reversed-phase HPLC (high-performance liquid chromatography) and sequenced by high-throughput MS/MS. Surveying signaling responses requires comprehensive profiling. However, most studies in mammalian cells report ~2,000 proteins in any one cell type, far lower than the proteome size of ~12,000 proteins. Limitations remain in (1) validating peptide sequences, (2) inferring protein identity from peptide information, (3) quantifying protein abundances, and (4) phosphorylation site mapping. We are developing new strategies to address these problems.

We developed three strategies to improve accuracy and sensitivity of peptide assignments. The manual analysis emulator (MAE) program scores chemical plausibility of observed MS/MS spectra, evaluating both fragment ion intensity and mass by using a recent kinetic model for gas phase fragmentation to simulate spectra. The MSPlus algorithm improves filtering of incorrect assignments by evaluating peptide physicochemical properties. These algorithms improve accuracy of matching MS/MS spectra to database peptide sequences. The IsoformResolver algorithm uses the novel approach of "peptide-centric" databases: each peptide sequence is represented once and then linked to all protein entries containing that sequence. This accurately reports ambiguities in protein identification due to database sequence redundancies. Together, MAE, MSPlus, and IsoformResolver capture data reliably and have enabled identification of >6,000 proteins from one human cell type.

Current methods for protein quantitation often use stable isotope labeling, where samples are differentially labeled and mixed and changes in peptide abundance are determined from intensity ratios of isotope-labeled peaks. Many human samples are not, however, easily analyzed by isotopic labeling. We investigated label-free methods, first counting MS/MS spectra for each protein and showing that statistical methods for serial analysis of gene expression (SAGE) could quantify relative protein changes. We also developed algorithms to quantify proteins by comparing ion intensities of peptides. We validated both methods with samples separated by multidimensional chromatography, allowing ~2-fold changes in protein abundances to be quantified with 95 percent confidence.

We are also investigating new approaches to identify and profile phosphopeptides, which are often difficult to analyze because of weak signal-to-noise ratio and competition with unphosphorylated peptides during MS/MS. To selectively report phosphopeptides from signature fragmentation products of –79 Da (PO₃^–), we are exploring the use of a negative-ion mode MS approach described by Steven Carr (Broad Institute, MIT) and Roland Annan (GlaxoSmithKline). We are developing strategies to use this method to analyze complex protein mixtures extracted from melanoma cells. Proteins phosphorylated in response to B-Raf/MAPK signaling in melanoma cells can be identified using precursor ion scanning and phosphopeptide detection from complex protein mixtures.

We are applying these technologies to profile responses to pathways dysregulated in melanoma. Our goals are to develop a robust, routine protocol for comprehensive profiling and, by doing so, gain insight into how integration of signaling from these pathways controls melanoma progression.

A third project uses hydrogen-exchange mass spectrometry (HX-MS) to investigate effects of protein kinase activation on internal protein motions. One line of study uses HX-MS measurements to reveal changes in conformational mobility between inactive and active forms of several protein kinases. Our findings have established that protein flexibility and enzyme activity are coupled in MAP kinases and MAP kinase kinase, a hypothesis that was difficult to address by other means. Using site-directed spin-labeling EPR (electron paramagnetic resonance), we have shown that kinase activation controls protein dynamics. A second study monitors changes in solvent protection to map regions of protein-protein interactions important for binding substrates or scaffold proteins. We have observed that enzyme activation or substrate binding leads to localized changes in protein flexibility, demonstrating regulated dynamic motions in protein kinases. These occur in regions far from sites of conformational changes, suggesting that internal motions are modulated over long distances.

We are now combining HX-MS with other biophysical approaches to confirm and extend these conclusions. Using site-directed mutagenesis and NMR (nuclear magnetic resonance) relaxation measurements, we are exploring how these protein motions control enzyme function, and how motions are regulated intramolecularly over long distances. Our findings provide insight into conformational dynamics of protein kinases and their control by enzyme activation and substrate binding. Such effects could be exploited to design more-specific inhibitors as potential cancer therapeutics.

This research has been supported in part by the National Institutes of Health.

Last updated: July 5, 2007