Quantitative Proteome Analysis of Barley Seeds Using Ruthenium(II)-tris-(bathophenanthroline-disulphonate) Staining Katja Witzel,† Giridara-Kumar Surabhi,†,§ Gottimukkala Jyothsnakumari,‡ Chinta Sudhakar,‡ Andrea Matros,† and Hans-Peter Mock*,† Leibniz Institute of Plant Genetics and Crop Plant Research, Corrensstraβe 3, 06466 Gatersleben, Germany, and Department of Botany, Sri Krishnadevaraya University, sv puram, Anantapur, Andhra Pradesh 515003, India Received October 9, 2006
This paper describes the application of the recently introduced fluorescence stain Ruthenium(II)-tris(bathophenanthroline-disulphonate) (RuBP) on a comparative proteome analysis of two phenotypically different barley lines. We carried out an analysis of protein patterns from 2-D gels of the parental lines of the Oregon Wolfe Barley mapping population DOM and REC and stained with either the conventional colloidal Coomassie Brilliant Blue (cCBB) or with the novel RuBP solution. We wished to experimentally verify the usefulness of such a stain in evaluating the complex pattern of a seed proteome, in comparison to the previously used cCBB staining technique. To validate the efficiency of visualization by both stains, we first compared the overall number of detected protein spots. On average, 790 spots were visible by cCBB staining and 1200 spots by RuBP staining. Then, the intensity of a set of spots was assessed, and changes in relative abundance were determined using image analysis software. As expected, staining with RuBP performed better in quantitation in terms of sensitivity and dynamic range. Furthermore, spots from a cultivar-specific region in the protein map were chosen for identification to asses the gain of biological information due to the staining procedure. From this particular region, eight spots were visualized exclusively by RuBP and identification was successful for all spots, proving the ability to identify even very low abundant proteins. Performance in MS analysis was comparable for both protein stains. Proteins were identified by MALDI-TOF MS peptide mass fingerprinting. This approach was not successful for all spots, due to the restricted entry number for barley in the database. Therefore, we subsequently used LC-ESI-Q-TOF MS/MS and de novo sequencing for identification. Because only an insufficient number of proteins from barley is annotated, an EST-based identification strategy was chosen for our experiment. We wished to test whether under these limitations the application of a more sensitive stain would lead to a more advanced proteome approach. In summary, we demonstrate here that the application of RuBP as an economical but reliable and sensitive fluorescence stain is highly suitable for quantitative proteome analysis of plant seeds. Keywords: barley • mass spectrometry • protein identification • protein staining • seed proteome • two-dimensional electrophoresis
1. Introduction Quantitative protein profiling has become the major focus in proteomic studies. Due to the high demand on reliable and sensitive methods for protein quantitation, it is a rapidly developing field. An enormous number of gel-based and nongel-based methods have been developed with particular assets and drawbacks. Several mass spectrometry-based approaches like the differential isotope labeling in vitro or in vivo1, 2 or * To whom correspondence should be addressed. Dr. Hans-Peter Mock, Leibniz Institute of Plant Genetics and Crop Plant Research, Corrensstrasse 3, D-06466 Gatersleben, Phone, +49 39482 5506; Fax, +49 39482 5139; E-mail,
[email protected]. † Leibniz Institute of Plant Genetics and Crop Plant Research. § Present address: Division of Biology & Ecological Genomics Institute, Ackert Hall, Kansas State University, Manhattan, KS 66506. ‡ Sri Krishnadevaraya University. 10.1021/pr060528o CCC: $37.00
2007 American Chemical Society
fluorescent affinity tags for the analysis of a particular subset of peptides3 have been established and enhanced in recent years. But still, in these non-gel-based approaches, the complexity of mass spectra generated from LC separated protein digests hampers their interpretation and identification using a single peptide is insufficient in some cases.4 Although MS-based quantitation methods provide the possibility of the detection of very low-abundant proteins and for automation of the workflow, for obtaining precise quantitative data, 2-D electrophoresis is still dominating the field in largescale proteomics. The reason for this is the opportunity to detect protein modifications beyond phosphorylation or glycosylation and to separate complex protein mixtures with a high-resolution capacity using large format gels and prefractionation approaches. But despite this potential to study protein expression and protein modification of several thousand Journal of Proteome Research 2007, 6, 1325-1333
1325
Published on Web 01/20/2007
research articles proteins at once, this method is far from being perfect. Intrinsic limitations resulting from comigrating proteins and unsatisfactory detection of low-abundant proteins are two of the weak points of this approach.5 Nevertheless, the number of enhancements in the 2-DE methodology increased considerably in the past years to overcome the mentioned bias. Several reviews have provided a comprehensive overview.6-10 The improvements included not only sample preparation or protein separation but also protein visualization by introducing new staining methods using fluorescent dyes. The intrinsic advantage of fluorescence staining is that the detection of the primary signal provides a linear response with respect to the amount of protein over a much wider range than is found for the common staining techniques Coomassie Brilliant Blue (CBB) and silver nitrate.11-17 One of the recently introduced fluorescent dyes is the metal chelate Ruthenium(II)-tris-(bathophenanthrolinedisulfonate) (RuBP) as described by Rabilloud et al.14 This stain is able to join the advances of a highly sensitive (detection threshold of 16 ng protein) and quantitative fluorescence dye with the cost efficiency of a lab-made staining solution. Recent technological developments in fluorescence methodologies and instrumentation have allowed the evolution of new assays with improved sensitivity and specificity permitting the detection and quantification of analytes that would otherwise be difficult to measure by any other approach. A number of recent studies aiming in the research of salt stress responses in plants focused on proteomic approaches using 2-DE.18-20 One of the most salt-tolerant cereals is barley (Hordeum vulgare), an important crop grown both for the feed and malting industries.21 Due to the fact that cultivars of this crop plant display different levels of salt tolerance, it is an appropriate tool to study the qualitative and quantitative changes in protein expression regarding to salt stress. In fact, two cultivars of barley displaying contrasting salt tolerance are H. vulgare cv. DOM and cv. REC, the parental lines of the Oregon Wolfe Barley (OWB) mapping population.22 The OWB population represents a set of doubled haploid spring barley lines, developed from the F1 of a cross between dominant (DOM) and recessive (REC) morphological marker stocks that exhibit an exceptional degree of phenotypic variation within a single reference population.23,24 Besides genome mapping and map-based analysis,25-27 these barley accessions are valuable for the functional characterization of molecular networks of stress mechanisms. Both parental lines showed differences in salt tolerance on the basis of germination experiments, and when using a germination assay for testing, results indicate a higher salt tolerance in germinating seeds of the parental line REC toward salt stress compared to DOM.22 In a preliminary comparative seed proteome analysis of the parental lines DOM and REC of the OWB mapping population, distinct differences in protein pattern were revealed. To investigate these quantitative and qualitative differences in more detail, we took advantage of recent developments in improved staining techniques to visualize proteins after two-dimensional gel electrophoresis. In this report, we compare the RuBP with the established cCBB stain on protein extracts of mature barley seeds. We validated the performance of both staining techniques as well as their compatibility with mass spectrometry and demonstrate that the novel RuBP stain is a highly suitable visualization method for quantitative proteome analysis to evaluate plant genetic material with respect to valuable agronomic traits. 1326
Journal of Proteome Research • Vol. 6, No. 4, 2007
Witzel et al.
2. Materials and Methods 2.1. Materials. The 2-D Quant Kit, IPG buffers, Immobiline DryStrips and agarose NA were purchased from GE Healthcare (Freiburg, Germany). For colloidal coomassie staining GelCodeBlue Stain Reagent from Pierce Chemical Company (Rockford, Illinois) and for Ruthenium staining Ruthenium(II)-tris-(bathophenanthroline-disulphonate) sodium salt solution from Fluka (Sigma-Aldrich, Buchs, Switzerland) were used. Tris base, glycerine, ammonium bicarbonate, acetic acid and Rotiphorese Gel 30 for SDS-PAGE were from Roth (Karlsruhe, Germany). CHAPS and DTT were from AppliChem (Darmstadt, Germany). Iodoacetamide was purchased from Sigma (Sigma-Aldrich, Buchs, Switzerland). High quality SDS was from USB Corporation (Cleveland, Ohio), urea from Invitrogen (Carlsbad, California), and TFA from Merck (Darmstadt, Germany). 2.2. Plant Material and Protein Extraction. Mature seeds from Hordeum vulgare cv. DOM and REC were obtained from green-house grown plants. Extraction of the water-soluble protein fraction was done following the protocol of Østergaard et al.28 Approximately 1 g of seeds was ground under liquid nitrogen in a cooled mortar to a homogeneous flour. Aliquots of 250 mg of flour were thawed in 1250 µL buffer (5 mM Tris/ HCl pH 7.5, 1 mM CaCl2) and incubated for 30 min at 4 °C on a shaker. After centrifugation (15 min, 4 °C), the supernatant was mixed with 4 volumes of ice-cold aceton and incubated at -20 °C for 2 h. Proteins were sedimented by centrifugation (5 min, 4 °C) and dried afterward in a vacuum centrifuge. The pellet was dissolved in lysis buffer (8 M urea, 2% CHAPS, 20 mM DTT, 0.5% IPG buffer) by incubating for 1 h at 37 °C on a shaker. Insoluble material was pelleted by centrifugation (15 min, room temperature). The protein concentration was determined using the 2-D Quant Kit (GE Healthcare) following the manufacturer’s instructions. 2.3. Two-Dimensional Gel Electrophoresis and Protein Staining. Protein extracts were subjected to isoelectric focusing and subsequent SDS-PAGE as described in ref 29. For Ruthenium-stained gels 150 µg of protein and for cCBB-stained gels 200 µg of protein were loaded by rehydration on IPG strips of 13 cm in length with a pH gradient of 3-10. For separation on an IPGphor II unit (GE Healthcare), the following parameters were used: 15 h rehydration, 1 h gradient to 250 V, 1 h gradient to 500 V, 1 h gradient to 4000 V, and 5.30 h 4000 V with a total of about 25 kVh. After IEF, strips were equilibrated for 15 min in buffer A (50 mM Tris-HCl pH 8.8, 6 M urea, 30% v/v glycerin, 2% w/v SDS, 20 mM DTT, 0.01% bromphenol blue). Strips destined for Ruthenium staining were equilibrated additionally in buffer B (50 mM Tris-HCl pH 8.8, 6 M urea, 30% v/v glycerin, 2% w/v SDS, 135 mM iodoacetamide, 0.01% bromphenol blue) for 15 min prior to SDS-PAGE. The strips were then placed on top of an 11.25% SDS polyacrylamide gel and covered with 0.5% agarose. Separation in the second dimension was performed using a Hoefer S600 apparatus (GE Healthcare). Afterward, gels were washed for 5 min with water and proteins were visualized with either cCBB staining using GelCodeBlue Stain Reagent (Pierce Chemical Company) following the manufacturer’s instructions or with RuBP staining using Ruthenium(II)-tris(bathophenanthroline-disulphonate) sodium salt solution (Fluka) following the protocol of Lamanda et al.16 2.4. Image Acquisition and Analysis. Image acquisition was performed using a UMAX Power Look III scanner (Umax Systems GmbH, Willich, Germany) with the MagicScan software
research articles
Barley Seed Proteome
(v4.5, Umax) for cCBB-stained gels. For RuBP-stained gels, the Fuji FLA-5100 (Fuji Film, Tokyo, Japan) with the Image Reader FLA-5000 v1.0 software was used. Scanning parameters were: resolution 100 µm, 16 bit picture, excitation wavelength 473 nm, emission filter 580 nm. For the 2-D image analysis, the TT900 S2S software (Nonlinear Dynamics) was used for image alignment and the Progenesis PG220 v2006 (Nonlinear Dynamics) was used for comparative image analysis. 2.5. Mass Spectrometry. Identification of proteins was accomplished using mass spectrometry. Excision of protein spots was performed manually from the gel. After washing with 400 µL of 10 mM ammonium bicarbonate/50% AcN for 30 min, the gel pieces were dried. For the digestion of proteins, 7.5 µL of trypsin solution (Sequencing Grade Modified Trypsin V511, Promega, Madison, USA, 10 ng/µl in 5 mM ammonium bicarbonate/5% AcN) was added to each sample, followed by an incubation step for 5 h at 37 °C. The digestion was stopped by adding 1 µL of 1% TFA. For MALDI-TOF mass spectrometry, 1 µL of the in-gel digest was placed directly onto a prespotted Anchor Chip target (Bruker Daltonics, Bremen, Germany). After drying, spots were washed with 7 µL of 10 mM ammonium phosphate in 0.1% w/v TFA. The acquisition of peptide mass fingerprint data was performed on a REFLEX III MALDI-TOF mass spectrometer (Bruker Daltonics) operating in reflector mode. Spectra were calibrated using external calibration and subsequent internal mass correction under application of the XMASS software v5.1.5 (Bruker Daltonics). Protein identification was performed with the MASCOT search engine (Matrix Science, London, UK)30 searching for Viridiplantae in the NCBI nonredundant protein sequence database as well as for barley EST Gene Index in the TIGR database. Parameters for the search were the following: monoisotopic mass accuracy, 200 ppm, missed cleavages, 1, allowed variable modifications, oxidation (Met), propionamide (Cys), and carbamidomethyl (Cys). For the LC-ESI-Q-TOF MS and de novo sequencing experiments, 3 µL of the digest was subjected to nanoscale RP LC analysis on a nanoAcquity UPLC system (Waters Corporation, Milford, MA). The mobile phase flow from the binary pump was used to preconcentrate and desalt the digest samples on a 20 mm × 180 µm Symmetry 5 µm C18 precolumn (Waters Corporation) for 3 min at 4 µL/min with an aqueous 0.1% formic acid solution. The peptides were subsequently eluted onto a 10 mm × 75 µm analytical Atlantis C18 column (Waters Corporation) and separated at 0.3 µL/min with an increasing AcN gradient from 2 to 40% B in 30 min. Mobile phase A consisted of 0.1% formic acid in water and mobile phase B of 0.1% formic acid in AcN. The nanoscale LC effluent from the analytical column was directed to the NanoLockSpray source of a Q/Tof Premier hybrid orthogonal accelerated Time-ofFlight (oa-TOF) mass spectrometer (Waters Corporation, MS Technologies Centre, Manchester, UK). The mass spectrometer operated in a positive ion mode with a source temperature of 80 °C and a cone gas flow of 30 L/h. A voltage of approximately 2 kV was applied to the nano flow sample tip. The mass spectra were acquired with the TOF mass analyzer in V-mode of operation and spectra were integrated over 1 s intervals. MS and MS/MS data were acquired in a continuum mode using MassLynx 4.0 software (Waters Corporation). The instrument was calibrated with a multi-point calibration using selected fragment ions of the CID of Glu-Fibrinopeptide B (SIGMAAldrich Chemie GmbH, Taufkirchen, Germany). Automatic data directed analysis was employed for MS/MS analysis on doubly
and triply charged precursor ions. The product ion MS/MS spectra were collected from m/z 50 to 1600. Lock mass correction of the precursor and the product ions was conducted with 150 pmol/µL Glu-Fibrinopeptide B in 0.1% formic acid in AcN/water (25:75, v/v), respectively, and introduced via the reference sprayer of the NanoLockSpray interface. ProteinLynx GlobalSERVER v2.1 software was used as a software platform for data processing, deconvolution and de novo sequence annotation of the spectra, and various database search types. The MS/MS spectra searches were conducted with a protein Viridiplantae index of the nonredundant NCBI database. A 10 ppm peptide, 0.1 Da fragment tolerance, one missed cleavage, and variable oxidation (Met) and propionamide (Cys) were used as the search parameters. BLAST homology and similarity searches were conducted with a protein Viridiplantae index of the nonredundant NCBI database as well as with the barley EST Gene Index in the TIGR database.
3. Results and Discussion 3.1. Comparison of Colloidal Coomassie Blue and Ruthenium Staining. The demand for a protein staining technique that is sensitive, quantitative, and compatible with mass spectrometry has been emerging within proteomic research for a number of years. Fluorescence detection of proteins after electrophoresis is gaining popularity, particularly for laboratories investigating differential protein abundance in large scale proteomics experiments on crude protein mixtures.31,32 Therefore, a major objective of this study was to compare the staining performance of two protein gel stains. The widely used stain cCBB and, as a ruthenium-based fluorescent metal chelate, Ruthenium(II)-tris-(bathophenanthroline-disulphonate) were compared with respect to sensitivity and homogeneity to assess their capability on the seed proteome of barley cultivars, with respect to spot detection and identification. The water-soluble protein fraction of barley flour was extracted according to Østergaard et al.28 to suppress the release of water-insoluble storage proteins representing the vast majority of crude seed protein extracts. For two-dimensional gel analysis, 200 µg of protein were separated for cCBB staining, which gave the best results in previous experiments regarding sensitivity and resolution. With consideration to superior sensitivity of fluorescent stains in detection,33,34 for RuBPstained gels, 150 µg of protein were separated to avoid overloading of the gels. The spot pattern was reproducible for independent extractions and reliability was confirmed in triplicate gels. The comparative analysis of the gel images from both stains displayed distinct differences in spot detection values (Figure 1). A first visual inspection of the gels revealed more spots on the protein maps of DOM and REC when the RuBP stain was used (confer Figure 1a-d). For a detailed comparison of the detection sensitivity concerning the number of visualized protein spots, the image analysis software was used. In cCBBstained gels, a total of 860 spots on the barley cultivar DOM protein map and 720 spots on 2-dimensional gels of barley cultivar REC were differentiated according to the software (Figure 2, gray bars). The number of visualized spots was significantly higher in RuBP-stained gels, although less protein was loaded. Here, 1290 and 1120 spots were differentiated on 2-dimensional gels of barley cultivars DOM and REC, indicating the detection of 430 and 400 additional spots, respectively (Figure 2, black bars). As expected, RuBP staining revealed a large number of low-abundance spots not detected by cCBB. Journal of Proteome Research • Vol. 6, No. 4, 2007 1327
research articles
Witzel et al.
Figure 2. Comparison of the total number of detected protein spots on two-dimensional gels from barley cultivars DOM and REC, visualized by either cCBB (gray) or RuBP (black). Data represent means of three technical replications. Gel images were analyzed using image analysis software. Table 1. Comparison of the Detected Differences in Expression Levels Between Barley Cultivars DOM and REC Evaluated from Both cCBB-Stained and RuBP-Stained Two-Dimensional Gelsa spots with increased abundance
Figure 1. Representative images from two-dimensional gel analysis of proteins from mature seeds from barley cultivars DOM (A, B) and REC (C, D). Sample loading and separation was as described in the Materials and Methods section. Proteins were visualized with either cCBB staining (A, C) or with RuBP staining (B, D). The framed areas of the gels (a-d) are reproduced in the lower panel with higher magnification indicating clear variation in the spot pattern between cultivars DOM (a, b) and REC (c, d), as well as differences in detection sensitivity between cCBB staining (a, c) and RuBP staining (b, d). All gels are shown using similar gray scale range.
This increase in spot detection values of about 34% is a consequence of the better sensitivity and higher protein affinity of RuBP dye, which is especially important for low-abundance proteins, representing the majority of proteins in complex biological samples. For analysis of the variation in spot pattern between cultivars DOM and REC, the differences in protein expression were evaluated from both cCBB-stained and RuBP-stained 2-dimensional gels. Image analysis with subsequent careful manual inspection revealed in total 56 spots showing changes of about 2-fold or more in abundance or appearance in only one of the cultivars when the cCBB stain was used. In contrast, 127 spots were detected using RuBP and displaying the mentioned changes in abundance (Table 1). This indicates an increase of 44% in detected expression variances between the two cultivars when using RuBP staining. For a more detailed look at the effect of the staining technique on the detection of differential protein abundance, seven protein spots, which are differential expressed in the cultivars DOM and REC, were chosen for a direct comparison. 1328
Journal of Proteome Research • Vol. 6, No. 4, 2007
stain
DOM
REC
cCBB RuBP
45 87
11 40
a The numbers of spots showing more than 2-fold significant change in normalized spot volume or with appearance in only one line are presented. Spots depicted as increased in one line should be considered as decreased in the other. Gel images were analyzed with an image analysis software.
Table 2. Comparison of the Detected Differences in Expression Levels Between Barley Cultivars DOM and REC for Selected Barley Seed Proteins Stained with Either cCBB or RuBPa
spot
1 2 3 4 5 6 7
a
protein name
accession number
Protein disulfidegi|1709617 isomerase precursor, Hordeum vulgare Protein z-type serpin, gi|1310677 Hordeum vulgare Serpin, gi|1197577 Hordeum vulgare Serpin, gi|1197577 Hordeum vulgare Fructose-bisphosphate gi|50878433 aldolase, Oryza sativa Disulfide isomerase, gi|493591 Hordeum vulgare Triose phosphate gi|2507469 isomerase, Hordeum vulgare
relative abundance in REC compared to DOM cCBB
RuBP
+2.6
+2.6
+2.5
+2.8
+1.6
+2.4
+1.4
+2.0
+1.1
+2.6
-2.6
+1.2
-1.1
-1.0
Protein identification was performed according to Section 2.5.
To evaluate the potential benefit of the fluorescence stain, given its dynamic range over 3 orders of magnitude compared to cCCB,12 we chose highly abundant spots for comparison. The results are shown in Table 2, summarizing differences in relative spot intensities between cCBB-stained and RuBPstained 2-dimensional gels. In our analysis, we found abundance ratios that were similar in both staining techniques as
Barley Seed Proteome
research articles
Figure 3. MALDI-TOF mass spectra of spot 5 excised from a cCBB-stained gel (A) and from a RuBP-stained gel (B) of two-dimensional separated seed protein extracts of barley cultivar Dom (see Figure 4a,b). Manual spot excision from cCBB-stained gels was performed using a standard white light box, whereas a standard UV light box was used for excision from RuBP-stained gels. MS analysis and subsequent database search of tryptic in-gel digests led to the identification of the spot as fructose-bisphosphate aldolase (Oryza sativa) in both cases.
well as ratios that were different after respective staining. The differences in the relative abundance of spots #2, 3, 4, and 5 were higher when proteins were stained with RuBP. This displays the better performance of RuBP when assessing quantitative expression changes based on threshold levels. In Table 2, spots #3, 4, and 5 exceed the threshold level of 2-fold, which is mostly used for distinguishing significant changes in spot abundance, only when they were visualized with RuBP. Similar results were obtained by analysis of two-dimensional gels of serial diluted marker proteins mixed to the water-soluble protein fraction of barley seeds (data not shown) and are in agreement to the literature.14 For spots #1 and 7, no changes in relative abundance were detected indicating that spot concentrations were at the optimum of the linear range of cCBB as well as RuBP. The huge divergence of the protein abundance ratio of spot #6, ranging from -2.6 with cCBB to +1.2 with RuBP, is propably due to local anomalities in the triplicate cCBB-stained gels, and it was observed exclusively for this spot. In most comparisons, a threshold value is used for detecting differential protein expression. Therefore, the dynamic range of a protein stain is crucial for detection of variation between two samples, and reliable measurement of relative abundances has a deep impact on whole data sets. Most questions in protein profiling are addressed to gain information especially on varying spot patterns and intensities rather than providing a survey of spots.35,36 In our study, the RuBP dye proved to be an advantageous alternative for proteome comparison due to the superior results in detection and quantitation. To enhance the cost-effectiveness ratio of the dye, the synthesis protocol of Rabilloud et al.14 and the staining protocol of Lamanda et al.16 can be applied. In a direct comparison of the purchased RuBP with the synthesized RuBP on serial diluted crude protein extracts on 1-D gels, we found no differences in signal-tobackground ratio (data not shown). 3.2. Comparison of Spot Pattern from Barley Cultivars DOM and REC. RuBP staining of 2-D gels of barley cultivar REC revealed 40 spots with increased and 87 spots with decreased protein expression, in comparison to cultivar DOM. Although the vast majority of protein spots were highly similar (Figure 1), the overall protein pattern displayed one particular region containing characteristic cultivar-specific protein spots (Figure 1, lower panel). For protein identification we have chosen 24 spots from this particular region as well as 7 distinctive spots with differences in relative abundance from the entire gel (Figure 4) to gain
more information about cultivar-specific and common proteins in DOM and REC. This approach represents the initial step for further biochemical characterization of both cultivars regarding cultivar-specific expression of seed proteins, rather than providing a complete list of identified proteins. 3.3. Identification of Proteins from Mature Barley Seeds. Protein identification was carried out from cCBB- and RuBPstained gels for spots visualized by both stains or only from RuBP-stained gels for spots visualized exclusively by this dye. Spots were excised manually and samples were processed as described in Section 2.5. In our experiments, spectra recorded from RuBP-stained gels showed in most cases the same quality as spectra acquired from cCBB-stained gels. As a typical result, MS data of spot number 5 are shown in Figure 3. The numbers of matching peptides as well as the sequence coverage were similar between the two staining techniques. After image analysis, 31 spots (see Figure 4) from a cultivarspecific region in the 2-D gel were chosen for identification via mass spectrometry. Spots were identified using MALDI-TOF MS and peptide mass fingerprinting was also used to confirm identity of the annotated spots from independent gels (data not shown). Proteins not identified by peptide mass fingerprinting data, due to limited database entries of barley, were subjected to LC-ESI-Q-TOF MS for de novo sequencing. For annotation, the criteria of Bevan et al.37 were followed in order to group the identified proteins according to their putative function. Three categories resulted: primary metabolism, disease/defense, and protein destination/storage. Table 3 presents the protein name, accession number, biochemical properties, as well as their detection by stain using the normalized spot volumes given by the image analysis software. Belonging to the group of housekeeping enzymes in the primary metabolism the following enzymes were identified: fructose-bisphosphate aldolase (#5), triosephosphate isomerase (#7), aldose reductase (#8), hexokinase (#12), glyceraldehyde3-phosphate dehydrogenase (#11, 14, 20, 21, 23), glucan endo1-3-beta-glucosidase (#28), and isoamylase 1 (#29). It is evident that proteins of the primary metabolism are involved in stress responses. When rice leaves were subjected to osmotic stress (drought and salt stresses), fructose-bisphosphate aldolase and triose phosphate isomerase were highly upregulated upon treatment.38 In a further proteomic analysis of salt stressresponsive proteins, triosephosphate isomerase was found induced by salt stress in roots of three-week-old rice seedlings.39 In our experiment, triosephosphate isomerase (#7) was not Journal of Proteome Research • Vol. 6, No. 4, 2007 1329
research articles
Witzel et al.
Figure 4. Representative images from two-dimensional gel analysis of proteins from mature seeds from barley cultivars DOM (A, a) and REC (B, b) are shown. Sample loading and separation was as described in Section 2.3 and proteins were visualized with RuBP staining. The framed areas of the gels (A and B) are reproduced in the lower panel with higher magnification of cultivars DOM (a) and REC (b). Numbers are circled in the case where the protein could be visualized only by RuBP but not with cCBB. Identified spots are numbered as listed in Table 3.
found differentially expressed in the seed proteome of the barley lines under investigation. The glycolytic enzyme fructosebisphosphate aldolase (#5) showed an increase of 2.6-fold in expression in the salt-tolerant line REC, leading to the assumption that in this line the energy transduction might be enhanced during germination. Hexokinase (#12) was found exclusively in the salt-sensitive line DOM, but possible comigration with a hordein spot at the same position in the 2D-gel (#13) of REC could hamper the identification of hexokinase in REC. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was found in 5 spots. Two spots (#21, 23) showed a higher expression in REC, one spot (#20) did not show a change in intensity, one spot (#11) was upregulated in DOM, and one (#14) could only be identified in DOM. But also for spot #14, it is possible that comigration with spot #29 (isoamylase 1) interferes with the determination of the spot’s identity. Whether the multiple spot abundance of GAPDH in the barley cultivars is regulated by differential gene expression of isoforms or posttranslational modifications, such as phosphorylation, has to be determined. The differential abundance of housekeeping enzymes detected in our experiment will be used for more detailed studies regarding altered patterns of carbon and energy flux in the saltsensitive line DOM and the salt-tolerant line REC during germination. The group of disease/defense proteins include the stress response enzyme peroxidase which was found in 6 different spots (#9, 10, 15, 16, 17, and 18) varying slightly in pI and Mr. Except for spot #10, which is upregulated in REC, all other spots are either higher (#17) or exclusively expressed in DOM (#9, 15, 16, 18). The physiological role of peroxidases in seeds 1330
Journal of Proteome Research • Vol. 6, No. 4, 2007
comprises not only cell elongation processes by cell wall loosening or cross-linking of cell wall components. As an antioxidant enzyme, peroxidases can oxidize substrates in the presence of H2O2 and also produce reactive oxygen species.40 Multiple isoforms of plant peroxidases are known, demonstrating overlapping activities and reflecting the significant role in germinating and signaling mechanisms. Several studies have shown that peroxidases are differentially regulated upon salt stress. In a proteome approach analyzing the leaf apoplast of tobacco, the spot abundance of two peroxidases was decreased after treatment.19 Contrary to this, peroxidase was found to be upregulated in salt stressed roots of rice seedlings.39 In a comprehensive transcriptional profiling approach, barley seedlings were subjected to a gradually imposed salt stress.41 As early as 3 h after reaching the final concentration of 100 mM NaCl, 3 different probes for peroxidase were downregulated in shoots. Apparently, the organ specificity in accumulation and functionality of peroxidase isoforms determines the response to environmental signals. In our analysis, a higher expression of this enzyme was detected for barley cultivar DOM that was found to be less tolerant against salt stress in germination assays.22 Therefore, the higher expression of peroxidase in DOM is rather related to other cultivar-specific developmental processes than to the different salt tolerance. The following enzymes were counted to the group of protein destination and storage proteins: protein disulfide-isomerase (#1, 6), serpins (#2, 3, 4), B-hordeins (#13, 19, 22, 24, 25, 26, 27, 30, 31) and a seed storage protein (#24). Hordeins, which were found in nine different spots in the analysis, are alcohol-soluble storage proteins and account for 30-50% of the total grain
research articles
Barley Seed Proteome
Table 3. Identification and Quantitative Assessment of Common as Well as Cultivar-Defining Barley Seed Proteinsa Ruthenium staining normalized spot volume (CV)
colloidal Coomassie staining
expression ratio
accession number
pI
Mr (kDa)
DOM
REC
DOM
spot
protein name
1
gi|1709617
5.0
56.4
0.542 (73%)
1.423 (11%)
1
gi|1310677
5.6
43.1
gi|1197577
5.4
42.7
gi|1197577
5.4
42.7
gi|50878433
6.9
38.8
0.063 (45%) 0.055 (32%) 0.188 (22%) 0.247 (50%)
0.180 (28%) 0.132 (19%) 0.389 (14%) 0.655 (8%)
gi|493591
4.8
33.2 26.7
1
-1.03
gi|728592
6.6
35.8
1
-1.03
gi|2624498
6.5
33.8
0.079 (27%) 0.940 (8%) 0.229 (11%) -
1.21
5.4
1
-
gi|167081
7.5
38.7
2.26
6.2
33.2
0.690 (7%) -
1
gi|120668
0.065 (37%) 0.969 (1%) 0.238 (73%) 0.621 (7%) 0.305 (2%) 0.023 (87%)
1
gi|2507469
1
-
26
Protein disulfideisomerase precursor, Hordeum vulgare Protein z-type serpin, Hordeum vulgare Serpin, Hordeum vulgare Serpin, Hordeum vulgare Fructosebisphosphate aldolase, Oryza sativa Disulfide isomerase, Hordeum vulgare Triosephosphate isomerase, Hordeum vulgare Aldose reductase, Hordeum vulgare Peroxidase 1, Hordeum vulgare Peroxidase BP 1, Hordeum vulgare Glyceraldehyde-3phosphate dehydrogenase, Hordeum vulgare Hexokinase, Zea mays B1 hordein fragment, Hordeum vulgare Glyceraldehyde-3phosphate dehydrogenase, Hordeum vulgare Peroxidase BP 1, Hordeum vulgare Peroxidase BP 1, Hordeum vulgare Peroxidase BP 1, Hordeum vulgare Peroxidase BP 1, Hordeum vulgare B3 hordein fragment, Hordeum vulgare Glyceraldehyde-3phosphate dehydrogenase, Hordeum vulgare Glyceraldehyde-3phosphate dehydrogenase, Hordeum vulgare B1 hordein fragment, Hordeum vulgare Glyceraldehyde-3phosphate dehydrogenase, Hordeum vulgare B3 hordein fragment, Hordeum vulgare B3 hordein fragment, Hordeum vulgare B hordein, Hordeum vulgare
27 28
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
29 30 31
gi|21954124
6.0
54.8
gi|18906
7.0
20.5
0.175 (23%) -
gi|66010
6.6
36.5
0.450 (21%)
gi|167081
7.5
38.8
gi|167081
7.5
38.8
gi|167081
7.5
38.8
0.739 (6%) 0.635 (11%) 0.147 (34%) 0.105 (38%) 0.539 (11%) 0.120 (7%)
REC
expression ratio
DOM
REC
DOM
2.62
0.320 (16%)
0.838 (14%)
1
2.61
1
2.85
2.53
2.37
1
1.63
1
2.07
1
1.49
1
2.65
0.327 (12%) 0.112 (26%) 0.254 (23%) 0.254 (23%)
1
1
0.129 (40%) 0.069 (27%) 0.171 (23%) 0.427 (15%)
1
1.15
0.127 (19%) 0.897 (9%) 1.139 (25%) 0.251 (5%) 0.633 (5%) -
0.049 (25%) 0.712 (18%) 0.712 (18%) -
1
-2.58
1
-1.18
1
-1.6
1
-
0.544 (4%) -
1
-
1
-
0.142 (28%) -
-
1
-
1
-
-
1
-
0.032 (75%) -
1 1
-
1.391 (17%) 0.110 (47%)
1
2.58
1
-1.08
1
-4.52
-
REC
-1.16 -
-
1
-
-
-
-
0.232 (9%)
-
1
-
0.815 (8%) 0.861 (17%) -
-
1
-
-
1
-
-
-
-
0.079 (54%) -
-
-
-
-
0.163 (28%) -
0.706 (17%) -
1
4.32
-
-
10.79
0.007 (98%)
0.016 (77%)
1
2.3
1
17.96
-
-
-
-
1
3.03
-
-
-
-
-
1
-
-
-
-
-
1
-
-
-
-
-
1
-
-
-
-
-
1
-
-
-
-
-
1
-
0.117 (31%)
-
1
-
1
-
1
3.04
-
0.233 (44%) -
-
1
-
gi|167081
7.5
38.8
gi|82371
7.7
30.1
gi|120680
6.6
36.5
gi|120680
6.6
36.5
0.003 (120%)
0.029 (78%)
1
gi|829269
8.7
27.6
gi|18978
6.6
36.4
0.002 (159%) 0.019 (3%)
0.027 (25%) 0.0577 (5%)
gi|82371
7.7
30.1
-
gi|82371
7.7
30.1
-
gi|73427781
9.0
34.4
-
B hordein, Hordeum vulgare
gi|73427781
9.0
34.4
-
Glucan endo-1,3beta-glucosidase, Hordeum vulgare Isoamylase 1, Hordeum vulgare B3 hordein fragment, Hordeum vulgare B3 hordein fragment, Hordeum vulgare
gi|585076
9.8
35.0
-
0.201 (18%) 0.232 (22%) 0.082 (39%) 0.135 (66%) 0.126 (47%)
gi|5825470
4.8
18.5
-
gi|82371
7.7
30.1
gi|82371
7.7
30.1
0.079 (31%) -
0.482 (5%) 0.240 (31%) 0.554 (6%)
normalized spot volume (CV)
0.277 (35%)
-
1
-
-
-
1
a Identification was done by database search against Viridiplantae protein index of the nonredundant NCBI database as well as the barley EST Gene Index in the TIGR database using peptide mass fingerprint data and de novo sequencing from MALDI-TOF MS and LC-ESI-Q-TOF MS analysis. The accession numbers are listed below. Theoretical values for molecular mass (Mr) and isoelectric point (pI) were calculated using ExPASy tools (http://www.expasy.ch). Relative protein abundance was determined using an image analysis software. The values given for normalized spot volume represent the mean of three technical replicates. Proteins, which could not be visualized in one of the lines, are indicated by hyphen.
nitrogen content. The subgroup of B-hordeins, a highly polymorphic gene family, represents 75-90% of total hordein content. Although a method for the extraction of water soluble
proteins was applied, also these storage proteins were detected on 2-D gels of barley seed proteins, indicative of different biochemical properties of the extracted B-hordeins. All nine Journal of Proteome Research • Vol. 6, No. 4, 2007 1331
research articles spots were either upregulated (#19, 22, 30) or exclusively expressed in REC (#13, 24, 25, 26, 27, 31), suggesting that the different hordein composition in this cultivar causes a better provision of the seedling with nitrogen and so might contribute to the better performance under stress conditions. The last two categories display the major biochemical processes in the endosperm, designated to stress/defense mechanisms and to protein folding/proteolysis-related enzymes in the group of protein destination and storage. Although we detected 127 differentially expressed spots with RuBP staining of 2-D gels, we restricted our analysis to a limited number of spots from a distinct cultivar-specific region to validate the potential benefit of the fluorescent RuBP stain. However, it seems evident that most identified unique proteins in this study, which differ between DOM and REC (refer Table 1), are housekeeping enzymes, indicating that these substantial enzymes are strongly connected to the phenotypic variation and tolerance against biotic or abiotic stress of both cultivars. In our study, we identified a number of candidate proteins, which could be responsible for salt tolerance. Further biochemical verification and functional elucidation based on methods detached from two-dimensional analysis should now provide more insight into cultivar-specific stress tolerance.
4. Concluding Remarks In the presented study, we aimed to establish a fluorescence staining method for mature seed extracts and to test the sensitivity in a comparison between two barley cultivars on 2-D gels. Our results showed that the number of detected spots by the image processing software was significantly higher in RuBPstained gels of both cultivars in comparison to the cCBB stain. Several proteins that were not detectable in cCBB-stained gels could be visualized with the RuBP gel stain and a number of them was chosen for identification to correlate the detection limits of a given protein spot with the ability for identification (see Figure 4). All protein spots, which were visualized exclusively by RuBP, could be identified by mass spectrometry. Overall, identification was successful for all of the selected spots proving the gain of information for protein maps due to RuBP staining. Also, the assessment of changes in relative abundance for given spots in the comparison was more reliable due to the well-known enhanced dynamic range of fluorophores. In our study we have found a number a spots, which were either regulated or cultivar-specific, and we proved the application of the staining technique in proteomic analysis as well as the biological relevance of our findings. The identified proteins form a basis for further proteomic analysis to bring our results into a biological context. For this purpose, introgression lines from the OWB mapping population, resulting from a cross between DOM and REC, are characterized in further studies. Weidner et al.22 used these lines for a germination assay for testing salt tolerance, and the results indicate that some lines display a higher salt tolerance than REC and some lines are more salt sensitive than DOM. Studies to explore the potential benefit of these lines to find and characterized candidate proteins for salt tolerance are underway.
Acknowledgment. Funding by the BMBF to H.-P.M. (GABISEED II; FKZ 0313115) is gratefully acknowledged. We thank Dr. Andreas Bo¨rner for providing barley seeds and Annegret Wolf for excellent technical assistance. G.-K.S. gratefully acknowledges receipt of DAAD/Leibniz post-doc scholarship (A/03/27383). 1332
Journal of Proteome Research • Vol. 6, No. 4, 2007
Witzel et al.
Note Added after ASAP Publication. The paper published ASAP on January 20, 2007 had an author name and current address and the last sentence of the Acknowledgment missing. The correct version was published ASAP on March 19, 2007.
References (1) Schmidt, A.; Kellermann, J.; Lottspeich, F. Proteomics 2005, 5, 4-15. (2) Gruhler, A.; Schulze, W. X.; Matthiesen, R.; Mann, M.; Jensen, O. N. Mol. Cell. Proteomics 2005, 4, 1697-1709. (3) Brittain, S. M.; Ficarro, S. B.; Brock, A.; Peters, E. C. Nat. Biotechnol. 2005, 23, 463-468. (4) Guina, T.; Purvine, S. O.; Yi, E. C.; Eng, J.; Goodlett, D. R.; Aebersold, R.; Miller, S. I. PNAS 2003, 100, 2771-2776. (5) Gygi, S. P.; Corthals, G. L.; Zhang, Y.; Rochon, Y.; Aebersold, R. PNAS 2000, 97, 9390-9395. (6) Harry, J. L.; Wilkins, M. R.; Herbert, B. R.; Packer, N. H.; Gooley, A. A.; Williams, K. L. Electrophoresis 2000, 21, 1071-1081. (7) Rabilloud, T. Proteomics 2002, 2, 3-10. (8) Ca´novas, F.; Dumas-Gaudot, E.; Recorbet, G.; Jorrin, J.; Mock, H.-P.; Rossignol, M. Proteomics 2004, 4, 285-298. (9) Go¨rg, A.; Weiss, W.; Dunn, M. J. Proteomics 2004, 4, 3665-3685. (10) Rossignol, M.; Peltier, J.-B.; Mock, H.-P.; Matros, A.; Maldonado, A.; Jorrı´n, J. Proteomics 2006, DOI. 10.1002/pmic.200600260. (11) Neuhoff, V.; Arold, N.; Taube, D.; Ehrhardt, W. Electrophoresis 1988, 9, 255-262. (12) Patton, W. F. Electrophoresis 2000, 21, 1123-1144. (13) Lopez, M.; Berggren, K.; Chernokalskaya, E.; Lazarev, A.; Robinson, M.; Patton, W. Electrophoresis 2000, 21, 3673-3683. (14) Rabilloud, T.; Strub, J.-M.; Luche, S.; Dorsselaer, A. v.; Lunardi J. Proteomics 2001, 1, 699-704. (15) Berggren, K. N.; Schulenberg, B.; Lopez, M. F.; Steinberg, T. H. et al. Proteomics 2002, 2, 486-498. (16) Lamanda, A.; Zahn, A.; Ro¨der, D., Langen, H. Proteomics 2004, 4, 599-608. (17) Berger, K.; Wissmann, D.; Ihling, C.; Kalkhof, S. et al. Mol. Cell Endocrinol. 2004, 227, 21-30. (18) Salekdeh, G.; Siopongco, J.; Wade, L.; Ghareyazie, B.; Bennett, J. Field Crops Res. 2002, 76, 199-219. (19) Dani, V.; Simon, W.; Duranti, M.; Croy, R. Proteomics 2005, 5, 737-745. (20) Yan, S.; Tang, Z.; Su, W.; Sun, W. Proteomics 2005, 5, 235-244. (21) Munns, R.; Hare, R.; Colmer, T. Proc. 8th Int. Barley Genet. Symp. 2000, 1, 238-245. (22) Weidner, A.; Dadshani, S.; Hakizimana, S.; Buck-Sorlin, G.; Boerner, A. Vortr. Pflanzenzuechtg 2005, 67, 53-55. (23) Wolfe, R. Barley Genet. Newslett. 1972, 2, 170. (24) Wolfe, R.; Franckowiak, J. Barley Genet. Newslett. 1991, 20, 117121. (25) Costa, J. M.; Corey, A.; Hayes, P. M.; Jobet, C. et al. Theor. Appl. Genet. 2001, 103, 415-424. (26) Cooper, L.; Marquez-Cedillo, L.; Singh, J.; Sturbaum, A. et al. Mol. Gen. Genomics 2004, 272, 181-193. (27) Rostoks, N.; Mudie, S.; Cardle, L.; Russell, J. et al. Mol. Gen. Genomics 2005, 274, 515-527. (28) Østergaard, O.; Melchior, S.; Roepstorff, P.; Svensson, B. Proteomics 2002, 2, 733-739. (29) Schlesier, B.; Mock, H.-P. In Arabidopsis Protocols; SanchezSerrano, J., Salinas, J., Eds.; Humana Press: Totowa, New Jersey, 2006; pp 381-391. (30) Perkins, D.; Pappin, D.; Creasy, D.; Cottrell, J. Electrophoresis 1999, 20, 3551-3567. (31) Skylas, D.; Cordwell, S.; Craft, G.; McInerney, B. et al. Aust. J. Agric. Res. 2005, 56, 145-155. (32) Gil-Agusti, M.; Campostrini, N.; Zolla, L.; Ciambella, C.; et al. Proteomics 2005, 5, 710-718. (33) Steinberg, T. H.; Chernokalskaya, E.; Berggren, K.; Lopez, M. F.; et al. Electrophoresis 2000, 21, 486-496. (34) Chevalier, F.; Rofidal, V.; Vanova, P.; Bergoin, A.; Rossignol, M. Phytochemistry 2004, 65, 1499-1506. (35) Finnie, C.; Melchior, S.; Roepstorff, P.; Svensson, B. Plant Physiol. 2002, 129, 1308-1319. (36) Gallardo, K.; Job, C.; Groot, S.; Puype, M. et al. Plant Physiology 2001, 126, 835-848. (37) Bevan, M.; Bancroft, I.; Bent, E.; Love, K. et al. Nature 1998, 391, 485-488. (38) Salekdeh, G.; Siopongco, J.; Wade, L.; Ghareyazie, B.; Bennett, J. Field Crops Res. 2002, 76, 199-219.
research articles
Barley Seed Proteome (39) Yan, S.; Tang, Z.; Su, W.; Sun, W. Proteomics 2005, 5, 235244. (40) Passardi, F.; Penel, C.; Dunand, C. Trends Plant Sci. 2004, 9, 534540.
(41) Walia, H.; Wilson, C.; Wahid, A.; Condamine, P.; Cui, X.; Close, T. Funct. Integr. Genomics 2006, 6, 143-156.
PR060528O
Journal of Proteome Research • Vol. 6, No. 4, 2007 1333
