Genomes to Life Contractor-Grantee Workshop III
February 6-9, 2005, Washington, D.C.
Genomics:GTL Program Projects
Harvard Medical School
2
Quantitative Proteomics of Prochlorococcus marinus
Kyriacos C. Leptos1* (leptos@fas.harvard.edu), Jacob D. Jaffe1, Eric Zinser2, Debbie Lindell2, Sallie W. Chisholm2, and George M. Church1
1Harvard Medical School, Boston, MA and 2Massachusetts Institute of Technology, Cambridge, MA
With the capability of performing whole-cell proteome analysis, a need to extent the above capability to whole-cell protein quantitation has proven to be a necessity. For this purpose we developed MapQuant, a platform-independent open-source software, which given large amounts of mass-spectrometry data, outputs quantitation for any organic species in the sample. We have previously applied MapQuant in the study of standardization samples at different concentrations on both LCQ and LTQ-FT spectrometers and also in the content of protein mixture of medium complexity and have showed linearity of signal with respect to the quantity of protein introduced.
The Prochlorococcus species is an abundant marine cyanobacterium that contributes significantly to the primary production of the ocean and whose life cycle is synchronized to the solar day (the “diel cycle”). In this study we leverage previously obtained protein identification data and the capabilities of MapQuant to quantify the proteins in a time-series dataset which includes 25 time points distributed along a 48-hour period (two diel cycles) of the strain MED4 of Prochlorococcus marinus. Protein samples from the growing culture were collected in duplicate and digested into peptides using trypsin, each time-point sample subjected to liquid chromatography coupled to hybrid linear ion trap-FTICR mass spectrometry, giving rise to a total 150 LC/MS experiments. The data acquisition took place on a Finnigan LTQ-FT mass spectrometer and it involved the acquisition of maximum two MS/MS spectra per MS spectrum. MS/MS spectra were interpreted using the program SEQUEST. The cross-correlation scores assigned to peptides that scored were filtered using thresholds to take into account false-positive results and the peptides were compiled into a summary list. This list of highly scored peptides was used as landmarks for evaluating MapQuant performance. MapQuant algorithms include morphological operations, noise filtering, watershed segmentation, peak finding and fitting, peak clustering and isotopic-cluster deconvolution and fitting using binomially distributed clusters of gaussioid peaks.
MapQuant outputs a list of potential organic species, by reporting four physical attributes for each isotopic cluster that it deconvolves. Those attributes are the m/z and the retention time (RT) of the monoisotopic peak, its charge and its carbon content. We have employed an m/z, RT and charge matching approach to assigning MapQuant Isotopic Clusters (MQIC) to the landmark peptides identified by SEQUEST in the same run with 91% success. However, MQICs that were assigned to a peptide using SEQUEST constitute 3% of the total MQIC found in a 2-D map. We are in the process of developing a matching algorithm that will be able to assign identities to unassigned MQICs. This approach will utilize SEQUEST peptides identified in the same organism Prochlorococcus marinus MED4 in five LC/LC/MS/MS experiments performed in the past, which correspond to five different environmental conditions. The matching algorithm should enable mapping of many of the remaining (97%) of the unidentified MQICs.
Our end goal is to be able to perform quantitation for most peptides found in the 25 time-points of the two diel cycles and hope to understand how carbon fixation, light-response and cell division are coordinated throughout the daily cycle.
Project Web site: http://arep.med.harvard.edu/DOEGTL/
* Presenting author
