RESEARCH INTERESTSPlant Metabolite Imaging Cross-linking Mass Spectrometry Bio-Oil Analysis
MS Imaging on Single Cell Level with High Spatial Resolution
The capabilities of high spatial resolution MS imaging is crucial for studying the detailed distribution of biologically important analytes. We have recently achieved single-cell level spatial resolution (~12 μm) by reducing the laser beam size with optical fibers (Anal. Chem., 2010, 82, 3255-3265). “Photo-quality” chemical images of surface lipid metabolites on the Arabidopsis whole flower allow the observation of single-cell localization of molecules, showing fine structures of the anther (tip of stamen) and single pollen grains at the stigma (tip of carpel) and anthers (Figure 1).
With this high spatial resolution, we could also obtain high quality chemical images of surface metabolites on the root. Figure 2 is the first time MS imaging of secondary roots for molecular species. We found uncommon metabolites, such as caffeate, ferulate, coumarate esters and sterol, directly on the root surface.
We recently upgraded our vMALDI-LTQ to a new MALDI-LTQ-Orbitrap mass spectrometer and now have high mass resolution imaging capability. High mass resolution has two major advantages: 1) differentiation of isobaric ions and 2) direct identification of unknown metabolites. For example, we can distinguish silver ion adducts of C29 alkane and C28 aldehyde in the spectrum and construct separate MS images (Figure 3).
In another experiment, our capability of single-cell level analysis is demonstrated with astrocyte as a model system (Rapid Commun. Mass Spectrom. 2010, 24, 1147-1154). Membrane-bound cholesterol was monitored as a silver ion adduct as an indication of single cells. Figure 4 shows a good correlation in cell counts between mass spectrometric and optical images at different cell concentrations. The most important advantage of this approach is we can easily construct single cell level population of analytes. For example, Figure 5 shows single cell cholesterol distribution obtained from MS images in Figures (d) and (e). An average measurement of cholesterol does not give this information; and a high resolution image, such as TOF SIMS, requires multiple data acquisition because of the limited field of view.
Plant Functional Genomics in High Localization
We are applying our approach to study functional genomics, particularly by monitoring highly localized metabolite distributions in a few Arabidopsis mutants. For example, Figure 6 shows distributions of a few surface lipids on an Arabidopsis flower comparing the cer1 mutant and the wild type, monitored by LDI MS imaging with colloidal silver (unpublished; presented at ASMS 2009). Most abundant surface lipids such as C29 alkane and C29 ketone are almost diminished but C30 fatty acid is abundant in cer1. The same trend is monitored in other plant organs, such as leaves and stems, but no difference was detected in other metabolites, such as flavonoids. This observation is consistent with the hypothesis that cer1 is blocking biosynthetic pathways from alkane to aldehyde. In fact, hyper-accumulation of aldehyde is observed in cer1, compared to wild-type, in high mass resolution imaging using MALDI LTQ-Orbitrap.
Another application is on a flavonoid mutant, tt7, in which flavonoid synthesis to the precursor of quercetin (Q) and isorhamnetin (I) is blocked. In Figure 7, Q and I and their glyceride derivatives disappeared in the tt7 mutant compared to the wild type, while kaempferol (K) and its derivatives are still present (ASMS 2010; manuscript in preparation). In this experiment, we compared three different matrix conditions: colloidal graphite, 9-aminoacridine (9AA), and no matrix. Flavonoids have good absorptions with N2 laser (377nm) and could be desorbed/ionized and detected without any matrix; however, ion signals for their glycerides derivatives are considerably lower than in others. Colloidal graphite and 9-AA show better ion signals, especially for glyceride derivatives; however, colloidal graphite suffers from matrix background.
Alternate Scanning Strategy to Obtain More Information in Less Time
We recently incorporated a Nd:YAG laser capable of producing laser spots of 10 μm into our commercial (Thermo Scientific) hybrid linear ion trap-Orbitrap mass spectrometer enabling MS imaging with high spatial resolutions. However, acquiring high mass spectral resolution MS images at this spatial resolution requires an unrealistic amount of time for data acquisition. For example, changing raster size from 50 μm to 10 μm requires 25x more data acquisition time and from ion trap scan to Orbitrap requires 3-5x more data acquisition time; totaling about 100x more data acquisition time in high spatial AND high mass resolution imaging. To circumvent this problem, we have developed methods to use the linear ion trap and Orbitrap in tandem to reduce acquisition time (ASMS 2010; manuscript in preparation). Incorporation of linear ion trap and Orbitrap scans during the course of MS imaging experiment reduced the data acquisition time by over 50% when compared to the same size area imaged with only Orbitrap scans (Figure 8 ). In this approach, we maintain high mass resolution information in an Orbitrap scan and high spatial resolution is maintained in the low mass resolution IT scan.
Furthermore, interspersing MS/MS (Figure 9) and MSn scans (Figure 10) can be incorporated to provide more analytical information of the sample, such as the distribution of structural isomers, which are not distinguishable even with Orbitrap high mass resolution analysis. Most of all, we can do all this imaging in a single experiment: high mass resolution Orbitrap imaging, MS/MS and MSn imaging.
This project is being performed in close collaboration with Basil J. Nikolau, R. Sam Houk, and Edward S. Yeung.References:
Sangwon Cha, Zhihong Song, Basil J. Nikolau, and Edward S. Yeung, "Direct Profiling and Imaging of Epicuticular Waxes on Arabidopsis thaliana by Laser Desorption/Ionization Mass Spectrometry Using Silver Colloid as a Matrix", Anal. Chem. 2009, 81, 2991–3000. Sangwon Cha, Hui Zhang, Hilal I. Ilarslan, Eve Syrkin Wurtele, Libuse Brachova, Basil J. Nikolau and Edward S. Yeung, “Direct profiling and imaging of plant metabolites in intact tissues by using colloidal graphite-assisted laser desorption ionization mass spectrometry”, The Plant Journal (2008) 55, 348–360. Sangwon Cha and Edward S. Yeung, “Colloidal Graphite-Assisted Laser Desorption/Ionization Mass Spectrometry and MSn of Small Molecules. 1. Imaging of Cerebrosides Directly from Rat Brain Tissue”, Anal. Chem. 2007, 79, 2373-2385. Hui Zhang, Sangwon Cha, and Edward S. Yeung, “Colloidal Graphite-Assisted Laser Desorption/Ionization MS and MSn of Small Molecules. 2. Direct Profiling and MS Imaging of Small Metabolites from Fruits”, Anal. Chem. 2007, 79, 6575-6584.Top of Page
Determination of the 3D structure of all proteins, aimed by Structural Genomics, seems to be an unachievable goal with only the current x-ray crystallography or NMR technologies, as signified by a low success rate of 3-5% to obtain genome-wide 3D protein structures (Yee et al., 2003). Moreover, the structural proteomics of biological complexes is expected to be a far more daunting process (Sall, 2003). Chemical cross-linking of intact proteins or protein complexes, followed by enzymatic digestion and mass spectrometric analysis (Figure 11) has been suggested as a low-resolution alternative (Sinz, 2006; Lee, 2008). This technique has been successfully utilized to provide protein structures and protein-protein interactions. It has also been utilized to construct high resolution protein structure by combining with x-ray crystallography of N- and C-terminal domains (Forwood et al., 2007). However, inherent difficulty in finding low abundance cross-linked peptides out of thousands of other coexisting non-cross-linked peptides has been a bottleneck in its further development (Sinz, 2006).
We developed a high throughput and high sensitivity shotgun approach to efficiently identify cross-linked peptides (Lee et al., 2007). Our new method is arguably the most sensitive approach reported so far, and we further improved this approach by adopting a probability based scoring system to significantly remove false positives (Lee, 2009). In the database search against large protein sequences, we found partial matching could be a significant problem in cross-linking sites search and we suggested a novel approach to remove such inadvertent false positives by adopting E-value filtering in each peptide level ( Figure 12). It is our expectation that this approach has great potential to advance cross-linking mass spectrometry as a unique tool supplementary to x-ray crystallography or NMR. We have recently been working on improving experimental conditions and operation parameters to enhance detection efficiency of cross-linked proteins, all while minimizing structural distortions caused by the cross-linking procedure. Our long term goal is to advance cross-linking mass spectrometry to fill the current technological gap in structural biology. We currently have several collaborations with biologists, computational scientists, and biophysicists to solve the structural proteomics problem together.References
Adelinda Yee, Keith Pardee, Dinesh Christendat, Alexei Savchenko, Aled M. Edwards, and Cheryl H. Arrowsmith "Structural Proteomics: Toward High-Throughput Structural Biology as a Tool in Functional Genomics", Acc. Chem. Res. 36, 183-189, 2003. Andrj Sall, "NIH Workshop on Structural Proteomics of Biological Complexes", Structure, 11, 1043-1047, 2003. A. Sinz, “Chemical cross-linking and mass spectrometry to map three-dimensional protein structures and protein-protein interactions”, Mass Spectrom. Rev. 25, 663-682, 2006. Young Jin Lee, Christopher B. Whitehurst, Erik J. Soderblom, Michael B. Goshe, Dennis T. Brown, Brett S. Phinney, “Cross-linking Site Mapping of Sindbis Virus Structural Proteins”, Proceed. 54th ASMS Conference on Mass Spectrom. Allied Topics, MP546, 2006. Young Jin Lee, Robert H. Rice, and Young Moo Lee, “Proteome Analysis of Human Hair by Two Dimensional Liquid Chromatography Coupled with Tandem Mass Spectrometry (2-D LC-MS/MS): From Protein Identification to Posttranslational Modification”, Mol. Cell. Proteomics, 5, 789-800, 2006b. Young Jin Lee, Laura Lachner, Jodi Nunnari, Brett S. Phinney, “Shotgun Cross-linking Analysis for Studying Quaternary and Tertiary Protein Structures”, J. Proteom Res. 6, 3908-3917, 2007. Young Jin Lee, “Mass Spectrometry based Cross-linking Sites Mapping for Structural Elucidation of Protein and Protein Complex”, Molecular Biosystems, 4, 816, 2008. Young Jin Lee, “Probability based Shotgun Cross-Linking Sites Analysis”, J. Am. Soc. Mass Spectrom., in print, accepted on Jun 30th, 2009. J. K. Forwood, A. S. Thakur, G. Guncar, M. Marfori, D. Mouradov, W. Meng, J. Robinson, T. Huber, S. Kellie, J. L. Martin, D. A. Hume and B. Kobe, “Structural basis for recruitment of tandem hotdog domains in acyl-CoA thioesterase 7 and its role in inflammation”, Proc. Natl. Acad. Sci. U. S. A., 104, 10382–10387, 2007.Top of Page
With so many problems in petroleum based energy production (limited oil reserves, energy independence, global warming, and oil spills in the Gulf of Mexico), biorenewable energy is arguably the best solution, if not the only solution, that can replace or supplement current petroleum based transportation fuels. It is particularly important in the State of Iowa and Iowa State University because of its high impact on agriculture, as exemplified in the initiation of the Bioeconomy Institute ( http://www.biorenew.iastate.edu/). As analytical chemists, we are contributing to this research field by developing and providing analytical tools that help to characterize the bio-oils down to the molecular level. High resolution mass spectrometry combined with modern ionization sources is best suited for this purpose, due to its high peak capacities and accurate mass measurements. We are specifically interested in two emerging bio-renewable research areas that do not agitate the food market.
Fast Pyrolysis of Biomass (in collaboration with Robert Brown)
Biomass can be converted into bio-oils by thermochemical conversion, specifically through fast pyrolysis process. However, the lack of molecular understanding on final products is a significant bottleneck in further improvement. Analysis of bio-oils has been relying on bulk property measurements such as pH, water quantity, combustion efficiency, and limited molecular information such as GC-MS, FT-IR, and GPC. GC-MS or pyGC-MS provide molecular level information, but only for volatile compounds or broken-down components. Understanding on nonvolatile compounds, the major components in fast pyrolysis bio-oils, has been almost absent. We recently developed high resolution MS techniques to characterize the molecular constituents by adapting similar techniques for petroleomics and assigned chemical compositions of over one hundred compounds (Figure 13; ASMS 2009).
Figure 14 shows DBE vs. the number of carbon plot for O4 and O6 compounds, most dominant heteroatom classes in bimodal distributions of m/z range 250-400 and 400-550. It demonstrates this bio-oil sample is dominated by lignin dimer and trimer units with average monomer chemical composition of C9H10O2. We are further developing this approach using APPI (atmospheric pressure photoionization) FT-ICR to characterize a wider range of molecules (ASMS 2010).
Algal Oil (in Collaboration with Brian Trewyn/Victor Lin)
Green algae is a photosynthetic microorganism that produces abundant fatty acids that can be used for bio-diesel production. Due to algae's fast growth cycle and non-influence on the food market, algae biomass has sparked interest in the renewable energy community. We are developing a high-throughput, high mass resolution assay for algae bio-oil analysis with a major focus on monitoring major algae lipids. With the successful development of the protocol, we will be able to assist developing better mesoporous nanoparticles that can separate specific lipid compounds by the Trewyn/Lin group.Top of Page