High Sensitivity Detection and Quantitation of Phosphoproteins in 1D and 2D Gels using the ProXPRESS 2D Proteomic Imager and ProFINDER 2D Image Analysis Software GEL IMAGING A P P L I C A T I O N N O T E The detection and quantitation of specific subsets of posttranslationally modified proteins has become a major area of activity in the proteomic analysis of biological systems. Phosphoproteins are a focus for much of this research owing to their importance in cell signaling and metabolic control. The determination of the phosphorylation state of proteins is important with respect to defining protein kinase substrates, as well as revealing the activation state of signal transduction pathways. These in turn have important implications with respect to the understanding of pathophysiological processes, such as cancer. This application note describes how phosphoproteins can be selectively imaged using the ProXPRESS 2D proteomic imager. The ProXPRESS 2D proteomic imaging system is a highly flexible, high performance CCD camera-based instrument suitable for a wide range of gel and non-gel based applications. Its various imaging modes cover fluorescence, transillumination (absorbance) and chemiluminescence applications and the flexible illumination modes include bottom, top and edge illumination. The sensitivity and linearity of detection of phosphoproteins separated by 1D-PAGE and fluorescently stained with the phosphatespecific dye Pro-Q Diamond Authors Elaine Scrivener1, Peter Jackson1, Wayne F Patton1, Alaine Levine2, Cedric Absalon2, Francoise Vannier2 and Simone J Seror2 1. PerkinElmer Life and Analytical Sciences 2. Institut de Genetique et Microbiologie, Université Paris-Sud
phosphoprotein gel stain (Molecular Probes, Inc., Eugene, OR) is demonstrated. This imaging technology is applied to the analysis of phosphoproteins present in complex mixtures of proteins extracted from Bacillus subtilis, a model Gram-positive bacterium. Model systems such as this are of increasing interest to researchers because of the role phosphorylation plays in pathway signal transduction. For example, the PrkC signaling pathway is implicated in sporulation and by biofilm formation and may have a role in cell wall biosynthesis. For example, the extra cellular domain of PrkC may bind penicillin. So these model systems are extremely useful for the study of antibiotics. An experiment is described showing how a phosphoproteinselective fluorescent image of a 2D gel of the bacterial extract is matched to a total protein image of the same gel to enable the excision of protein spots for subsequent proteolysis and analysis by peptide mass profiling. This work demonstrates the capability of the ProXPRESS 2D imager and ProFINDER 2D software for studying the phosphoproteins of bacteria. The steps described in this application note can easily be applied to the study of phosphorylation in other organisms. High sensitivity phosphoprotein detection In proteomics studies it is often the case that proteins of functional interest are present in low abundance in a biological medium. Phosphoproteins are frequently present at low levels and hence sensitive detection is important. The sensitivity of detection of phosphoproteins, separated by 1D PAGE and stained with the phosphate-selective Pro-Q Diamond dye, was measured using the ProXPRESS 2D imager. Salient features of the dye have been described previously (Steinberg et al., 2003). A dilution series of a mixture of five standard proteins, including two phosphoproteins, ovalbumin and beta-casein, was used to determine the limits of detection. Samples were applied to four replicate 1D gels covering the range from 1000 ng down to 0.4 ng for each individual protein (a separate dilution series for each gel). Following electrophoresis the gels were stained with Pro-Q Diamond according to the suppliers instructions as follows: The gels were fixed by gentle shaking for 30 min in 200 ml of an aqueous solution of 40% (v/v) methanol, 10% (v/v) acetic acid. This fixing step was repeated and the gels left in the fixing solution overnight. The gels were washed 3 times for 15 min in 250 ml water before being stained once for 2h with 50 ml per gel Pro-Q Diamond stain solution. The gels were destained by washing 3 times for 15 min in 100 ml Pro-Q Diamond destain solution (as supplied by Invitrogen) and subsequently, twice for 15 min in 250 ml water. Imaging was performed on the ProXPRESS 2D imager using the top illumination mode with the appropriate filter set (Excitation 540/25 nm, Emission 590/25 nm). Top illumination uses a high intensity xenon arc lamp to produce a broad spectrum illumination source encompassing the visible light range. Broadband xenon arc lamp excitation is especially advantageous for gel imaging compared to laser excitation because of the flexibility of being able to choose an optimum wavelength for excitation using filter sets optimized for the particular dye being imaged. With laserbased imagers, excitation efficiency is limited to how far the laser line is displaced from the excitation maximum of the dye. Because laser lines are fixed, it is rarely possible to match laser excitation wavelength with the dye s excitation maximum as well as can be done with broadband excitation and optimized filters. This is particularly important for imaging Pro-Q Diamond phosphoprotein gel stain, which absorbs maximally at 555 nm and emits maximally at 580 nm. An optimized filter set is also used in the light emission path to minimize the transmission of excitation light to the camera. For imaging the gels in this type of low-light application, the instrument was fitted with an absorbent glass base-plate which minimizes stray light, resulting in very low background levels. The gels were placed directly onto this base (supplied with instrument) and a sheet of low fluorescence glass was placed on the top of the gel cassette. Flat fielded images were acquired (i.e. a correction was made for spatial variation in illumination) and the 16-bit TIFF files obtained were analyzed using Phoretix 1D software (Nonlinear Dynamics, Newcastle upon Tyne, England). A representative image is shown in Figure 1. 2
Figure 1. A 1D gel image of a Pro-Q Diamond stained dilution series of five standard proteins including two phosphoproteins. The inset on the right shows the profile of the most heavily loaded lane and shows that the phosphoproteins give the most significant signals and that saturation is not observed. The profiles on the left show that the 0.4 ng protein bands are visible as distinct peaks. Images were obtained using an exposure time of 20 seconds, which was chosen to provide a strong signal from the highestloaded protein lane without showing saturation, as can be seen in the top inset. The two phosphoproteins are detected down to a level of 0.4 ng, demonstrating excellent sensitivity. Some nonspecific staining was observed with this dye, but as can be seen from the profile of the most heavily loaded lane, this was small compared to the staining of the phosphoproteins. The fluorescent intensities of the phosphoproteins were roughly 200-fold higher than the nonspecific staining of the other standards. The insets on the left in Figure 1 show distinct peaks at 0.4 ng phosphoprotein. The signal-to-noise ratios of the 0.4 ng bands from each image were all greater than 3.0 (Table 1), demonstrating detection limits as low as 0.4 ng. This compares favorably with imaging of the dye using laserbased gel scanners, with reported detection limits of 1-8 ng for the same model phosphoproteins (Steinberg et al., 2003).
Sensitive fluorescent detection of Bacillus subtilis phosphoproteins A protein extract of Bacillus subtilis was separated by 2D PAGE and subsequently stained with the Pro-Q Diamond dye. The image of Pro-Q Diamond stained phosphoproteins was acquired on the ProXPRESS 2D imager using the appropriate filter set (Excitation 540/25 nm, Emission 590/25 nm). A portion of this image* is shown in Figure 3a. Some of the fainter spots in this image are thought to be owing to non-specific staining of non-phosphorylated proteins as some spots have the characteristic of heavily loaded protein spots, a non-symmetrical shape in the vertical plane with the strongest fluorescence in the lower part of the spot. Many of the distinct, small, symmetrical spots have been identified as phosphoproteins by mass spectrometry. Examples of phosphorylated and possible non-phosphorylated spots are highlighted in Figure 3a. Full details of this work are submitted for publication and hence only a part of this work can be shown at this time. Possbile non-phoshorylated protein spot (not identified yet) Phoshorylated protein spot Figure 3a: A ProXPRESS 2D image of a Pro-Q Diamond stained 2D gel of Bacillus subtilis. Phosphorylated and possible non-phosphorylated spots are indicated. Figure 3b: A ProXPRESS 2D image of a SYPRO Ruby stained 2D gel of Bacillus subtilis. The region of the gel is the same as shown in 3a. For protein identification by mass spectrometry it was necessary to excise protein spots from the gel for subsequent proteolysis and peptide mass profiling. To detect spots for automated spot picking using the ProXCISION gel picker, it was necessary to post-stain with SYPRO Ruby total protein stain according to the supplier s instructions and re-image for this stain using the ProXPRESS 2D imager (the filters used were; excitation, 460/80 nm and emission, 650/150 nm). Figure 3b shows a region of the SYPRO Ruby image which is equivalent to that of the Pro-Q Diamond image in Figure 3a. 5
Table 1: Signal to noise ratio of the 0.4 ng phosphoprotein bands from each image Signal to noise ratio of the 0.4 ng phosphoprotein bands Ovalbumin Gel 1 6.11 Gel 2 5.11 Gel 3 5.19 Gel 4 6.82 Beta Casein Gel 1 5.69 Gel 2 4.43 Gel 3 4.83 Gel 4 4.24 The figures used to calculate signal to noise ratio were taken from the ProSCAN software (ProXPRESS Imager acquisition software) using the statistics function. Signal = the upper quartile from a region two pixels deep across the protein band minus the mean of the background (three areas were selected and a mean value was taken from each, the average of these was calculated). Noise = the mean of three standard deviation values (of background) from the previously selected regions. A wide linear dynamic range demonstrated The 16-bit TIFF images obtained in the sensitivity experiment were analyzed using Phoretix 1D analysis software. Background subtracted band volumes from the phosphorylated proteins were measured from the four images (each showing a different gel, each with a separate protein dilution series). The mean of these values was plotted against sample loading and is shown in Figure 2a. Excellent linearity is obtained over this range of greater than 3 orders of magnitude. Figure 2b shows a plot from a single image of a 1D gel with a sample loading covering 4 orders of magnitude (4,000 ng down to 0.4 ng). Linearity is demonstrated over the full four orders of magnitude for the phosphoprotein bands measured, which compares very favorably with laser-based gel scanners, demonstrating a 2-3 order of magnitude linear dynamic range (Steinberg et al., 2003). Figure 2a: Mean band volume plotted against sample loading of the two phosphoproteins from the 1D gel sensitivity experiment. A double logarithmic scale has been used to allow easy visualization of response at low sample loadings. The error bar values were taken from the standard deviation of the four replicate band volumes. Figure 2b: Band volume plotted against sample loading of the two phosphoproteins from a single 1D-gel which was loaded with a four order of magnitude dilution series. A double logarithmic scale has been used to allow easy visualization of response at low sample loadings. 4
In order to define a list of protein spots for excision, spots of interest from the Pro-Q Diamond image were matched to corresponding spots in the SYPRO Ruby image using the ProFINDER 2D image analysis software. Figures 4a and 4b show corresponding spots in the two images. Triangulation (to map spot coordinates from the SYPRO Ruby image acquired using the ProXPRESS 2D imager to the ProXCISION SYPRO Ruby picking image) could then be carried out to enable spot picking. Summary This application note demonstrates that the ProXPRESS 2D Proteomic Imaging System can image Pro-Q Diamond stained phosphoproteins in 1D and 2D gels with high performance. Excellent detection sensitivity and linearity is demonstrated. The correspondence of 2D gel spot patterns from Bacillus subtilis, identified using both the Pro-Q Diamond phosphoprotein gel stain and the SYPRO Ruby total protein gel stain are shown. These images were used for subsequent 2D spot pattern matching using ProFINDER 2D software to enable phosphoprotein excision for subsequent spot identification by mass spectrometry. Figure 4a. ProFINDER 2D analysis software view: Selected spots of interest in the Pro-Q Diamond image. Reference Steinberg TH, Agnew BJ, Gee KR, Leung WY, Goodman T, Schulenberg B, Hendrickson J, Beechem JM, Haugland RP, Patton WF. Global quantitative phosphoprotein analysis using Multiplexed Proteomics technology. Proteomics. 2003 Jul;3(7):1128-44. Figure 4b. ProFINDER 2D analysis software view: Protein spots in the SYPRO Ruby image corresponding to those selected from the Pro-Q Diamond image. PerkinElmer Life and Analytical Sciences 710 Bridgeport Avenue Shelton, CT 06484-4794 USA Phone: (800) 762-4000 or (+1) 203-925-4602 For a complete listing of our global offices, visit /lasoffices 2005 PerkinElmer, Inc. All rights reserved. The PerkinElmer logo and design are registered trademarks of PerkinElmer, Inc. ProXPRESS and ProFINDER are trademarks of PerkinElmer, Inc. or its subsidiaries, in the United States and other countries. Pro-Q and SYPRO are registered trademarks of Molecular Probes, Inc. Phoretix is a trademark of Nonlinear Dynamics, Ltd. ProXCISION is a trademark of Genomic Solutions, Inc. All other trademarks not owned by PerkinElmer, Inc. or its subsidiaries that are depicted herein are the property of their respective owners. PerkinElmer reserves the right to change this document at any time without notice and disclaims liability for editorial, pictorial or typographical errors. 007281_01 Printed in USA