Comparative Study of [Three] LC-MALDI Workflows for the Analysis of Complex Proteomic Samples Stephen J. Hattan,* ,† Jason Marchese, † Nikita Khainovski, † Steve Martin, ‡ and Peter Juhasz ‡ Applied Biosystems, 500 Old Connecticut Path, Framingham, Massachusetts 01701, and BG Medicine, 40 Bear Hill Rd., Waltham, Massachusetts 02451 Received April 11, 2005 Large-scale proteomic analyses frequently rely on high-resolution peptide separation of digested protein mixtures in multiple dimensions to achieve accuracy in sample detection and sensitivity in dynamic range of coverage. This study was undertaken to demonstrate the feasibility of MALDI MS/MS with off-line coupling to HPLC for the analysis of whole cell lysates of wild-type yeast by three different workflows: SCX-RPHPLC-MS/MS, high-pH SAX-RPHPLC-MS/MS and RP (protein)-SCX-RPHPLC-MS/ MS. The purpose of these experiments was to demonstrate the effect of a workflow on the end results in terms of the number of proteins detected, the average peptide coverage of proteins, and the number of redundant peptide sequencing attempts. Using 60 μg of yeast lysate, minor differences were seen in the number of proteins detected by each method (800-1200). The most significant differences were observed in redundancy of MS/MS acquisitions. Keywords: yeast • two dimensional chromatography • LC-MALDI • MALDI • workflow • proteomics • shotgun proteomics Introduction Exhaustive proteomic analyses of complex samples use either 2D-poly(acrylamide) gel electrophoresis (PAGE) for separating proteins 1-3 or multidimensional peptide chromatography of the proteolytic digest of the entire sample 4-6 or other combinations of protein and peptide fractionation steps. 7-9 Protein identifica- tion based on the peptide mass fingerprinting of in-gel digested gel spots relies most commonly on MALDI TOF mass spec- trometry, 10-12 but the method of choice for peptide identifica- tions using MS/MS experiments is on-line LC-ESI MS/MS. 13-15 More recently, with the introduction of MS/MS capabilities on MALDI TOF instruments 16-17 or with the implementation of MALDI ionization sources on MS/MS capable mass analy- zers, 18-20 the use of an LC-MALDI analytical platform for proteomic studies is becoming more widespread. 21,22 While preliminary comparisons of data collected on proteomic samples run in parallel by LC-ESI MS/MS and LC-MALDI show the two techniques to be somewhat complementary, 23-25 in regards to experimental design and result-dependent data analysis the off-line coupling of LC-MALDI platform has distinct advan- tages. 26-33 First, the archiving of the LC separation on a MALDI plate allows independent and repetitive MS and/or MS/MS analyses of the same separation without having to worry about the reproducibility of capillary HPLC separations. For instance, the selection of precursor species for MS/MS can be chosen to coincide with the peak of the HPLC elution; it can be based on relative abundances of peptide pairs in a differential isotope labeling experiment 34-37 (expression-dependent MS/MS). Ac- commodation of different acquisition conditions for different precursors 38 can also be easily accomplished whereas it would be very difficult in an on-line LC-ESI MS/MS configuration. Soon, a wide range of result-dependent MS/MS strategies 39 are expected to emerge that utilize off-line LC-MALDI coupling. Another area where LC-MALDI offers a distinct advantage has to do with the wider range of chromatography variables that are compatible with this mode of ionization. 40-44 The use of ion pairing agents such as TFA, in the HPLC buffers does not affect MALDI sensitivity, and even surfactants and phos- phate additives such as ammonium-phosphate or sodium- phosphate can be added to improve chromatographic resolu- tion. 45 The range of flow rates and column sizes has also substantially less effect on mass spectrometric sensitivity if MALDI is used, allowing for decoupled optimization of chro- matography and mass spectrometry. The present research attempts to characterize different peptide-based workflows in proteomics where LC-MALDI MS/ MS is utilized for peptide identification. For such a study the reproducible selection of precursors for MS/MS is critical, and LC-MALDI was found to be a better choice than LC-ESI. Replicate analyses of complex peptide mixtures have been reported to yield only partially overlapping lists of identified peptides with LC-ESI MS/MS, the overlap at times being as low as 25%. 46 Similar experiments with LC-MALDI routinely result in 60% or higher overlap (data were generated from digest mixtures of complex yeast fractions; not shown here). There- fore, results derived from LC-MALDI MS/MS experiments are * To whom correspondence should be addressed. E-mail: hattansn@ appliedbiosystems.com. † Applied Biosystems. ‡ BG Medicine. 10.1021/pr050099e CCC: $30.25  2005 American Chemical Society Journal of Proteome Research 2005, 4, 1931-1941 1931 Published on Web 10/21/2005