Protein Staining Influences the Quality of Mass ... - ACS Publications

Aug 6, 2004 - Brilliant Blue (CBB) and silver staining, respectively. We have compared the quality of mass spectra for proteins obtained from gels sta...
0 downloads 0 Views 127KB Size
Protein Staining Influences the Quality of Mass Spectra Obtained by Peptide Mass Fingerprinting after Separation on 2-D Gels. A Comparison of Staining with Coomassie Brilliant Blue and Sypro Ruby Boel Lanne* and Oleg Panfilov AstraZeneca R&D, Research Area CV & GI, S-431 83 Mo¨lndal, Sweden Received August 6, 2004

Abstract: When separating protein mixtures on 2-D gels for proteomics purposes, fluorescent staining is superior in sensitivity and linear response as compared to Coomassie Brilliant Blue (CBB) and silver staining, respectively. We have compared the quality of mass spectra for proteins obtained from gels stained with CBB and SYPRO Ruby (SR) and found significant differences. These differences can be seen both in inferior signal/noise ratios and number of peptides detected with the fluorescent stain. Keywords: protein identification • protein staining • gel electrophoresis • mass spectrometry • MALDI • peptide mass fingerprinting • significance testing • proteomics • cystein residues

Introduction A new protein stain, SYPRO Ruby, a fluorescent organic rhutenium complex that binds noncovalently to proteins, was introduced a couple of years ago.1 This stain is as sensitive as silver stains but much simpler to use. Furthermore, SR staining shows a linear response over 3 orders of magnitude.2 Extended linear response is a property of the outmost importance for protein expression comparisons in analyses using 2-D gels and peptide-mass fingerprinting (PMF). As different fluorescent protein stains become more and more frequently used for proteomics purposes, it is crucial to test whether the stain itself or the staining procedure influence the mass spectrometric (MS) analysis of the protein spots in a negative way. It has been noted that SR staining can give lower score rates and lower base-peak heights than MS analyses performed with CBB stained gels.3 We therefore investigated if this lowering of MS response was observed also in our laboratory. We have shown that, using Duracryl 2-D gels (250 × 200 × 1 mm), the MS signals of tryptic peptides from CBB stained gels were better in terms of S/N ratios, sequence coverage and scores than those obtained from SR stained gels. The inferior spectra obtained after SR staining partly depended on the frequent absence of signals corresponding to Cys-containing peptides.

analyses, see ref 4. Livers of Zucker fatty rats (fa/fa) were homogenized4 in 8 vol. of 7 M urea, 2 M thiourea, 20 mM dithiothreitol (DTT), 0.5% (v/v) IPG-buffer pH 3-10 NL, 4% (w/v) CHAPS and protease inhibitors (Complete Mini, 1 tablet per 10 mL, Roche Diagnostics, complete with EDTA, #1836153). Protein standards, broad range, 161-0317, and peptide standards, 161-0326, were obtained from Bio-Rad, Hercules, CA, USA; bovine serum albumin (BSA), Fraction V, A-2153 and bakers yeast enolase were obtained from Sigma-Aldrich; and recombinantly expressed Helicobacter pylori-antigen A (HpaA, Swiss-Prot id HPAA_HELPJ, 31 kDa), was prepared and purified by Susanne Nystro¨m (present affiliation Medivir AB, Huddinge, Sweden). The protein standards were dissolved in an “SDS buffer” (0.1 M Tris-HCl, pH 6.8, 4% SDS, 20% (v/v) glycerol, 10 mM DTT, 0.01% (w/v) bromophenol blue (BPB)). The protein standards were in some cases subjected to alkylation under reducing conditions (10 mM DTT) by addition of iodo acetamide (IAA), 60 mM, 5-10 min, and room temperature. The mixture of protein standards was applied to strips of filter paper, 200 × 4 mm, attached using 0.6% agarose in running buffer to a Duracryl SDS-PAGE gel (250 × 200 × 1 mm). For 2-D gel analysis of rat-liver proteins, the isoelectric focusing (IEF) strips were rehydrated for 12 h (30 V) in 350 µL “Urea buffer” (7 M urea, 2 M thiourea, 10% glycerol, 0.02% bromophenol blue, 20 mM DTT, 0.5% (v/v) IPG-buffer pH 3-10NL or 4-7L, 4% (w/v) CHAPS with protease inhibitors) with rat-liver proteins dissolved. The strips were focused under a layer of Drystrip cover fluid, at 20 °C, in an IPG-Phor unit. The focusing was carried out in three steps: 500 V for 1 h, 1 kV for 1 h, and 8 kV until focusing had been obtained (65 kVh, all steps 1) as compared to the SR stain. In one case, the CBB staining was almost three times better than the SR staining. Although in all analyses CBB gave better S/N, scores, number of identified peptides and sequence coverage, the number of analyses (four) were too few to reach statistical significance, Table 1. Furthermore, none of the 17 theoretically possible peptides with cysteine residues could be

technical notes

Lanne and Panfilov

Table 2. Quality of MALDI-TOF MS Obtained for Standard Proteins after Staining the Gels with Either CBB or SR, Experiment 2, Gel Plugs 2 mma

protein, conditions

enolase, nonalkylated, 280 fmol/plug enolase, alkylated, 280 fmol/plug HpaA, nonalkylated, 750 fmol/plug HpaA, alkylated, 750 fmol/plug BSA, alkylated, 650 fmol/plug

N

ratio S/N CBB/SR

no. of theor. Cys peptides

cys detected in CBB (range)

Cys detected in SR

average score CBB

ratio CBB/SR average score SR

score

matching peptides

sequence coverage

8

1.7b

1

0

0

169

98.3

1.7

1.7

1.9

8

1.1

1

0

0

100

78.8

1.3

1.3

1.3

8

1.7c

0

0

0

135

98.3

1.4

1.3

1.2

8

1.5

0

0

0

67.6

1.5

1.4

1.5

8

1.2d

17

3 (2-4)

0

84.4

1.3

1.3

1.4

p e 0.031

p e 0.031

p e 0.031

95.8 108

pe 0.031

a For calculation of p-values Wilcoxon sign-test, one-sided, was used. b S/N values were calculated for the following enolase peptides: 1412.85, 1627.95, 1821.92, 1912.03, 2441.12. c S/N values were calculated for the following HpA peptides: 1465.78, 1862.05, 1887.88, 1975.05, 2088.07. d BSA, see Table 1 for peptides used for S/N calculations.

Table 3. Quality of MALDI-TOF MS Obtained for Standard Proteins after Staining the Gels with Either CBB or SR, Experiment 3, Gel Plugs 2 mma ratio CBB/SR Cys detected in CBB (range)

Cys detected in SR (range)

average score CBB

protein, conditions

N

no. of theor. Cysa peptides

phosphorylase B, nonalkylated, 1.5 pmol/plug BSA, nonalkylated, 2.2 pmol/plug carbonic anhydrase, nonalkylated, 4.9 pmol/plug ovalbumin, nonalkylated, 13.4 pmol/plug lysozyme, nonalkylated, 10 pmol/plug triosephosphate isomerase, nonalkylated, 12 pmol/plug phosphorylase B, alkylated, 1.5 pmol/plug BSA, alkylated, 2.2 pmol/plug carbonic anhydrase, alkylated, 4.9 pmol/plug ovalbumin, alkylated, 3.4 pmol/plug lysozyme, alkylated, 10 pmol/plug triosephosphate isomerase, alkylated, 12 pmol/plug

8

8

1

1

224

8

17

3 (2-4)

0

143

8

0

0

0

8

6

0

0

141

8

6

4

3

113

8

5

5

0

171

8

8

2 (2-3)

0

197

8 8

17 0

2 0

0 0

8

6

1

1

137

8

6

6 (5-7)

4 (3-5)

113

8

5

4 (4-5)

1 (0-2)

194

98.1

92.3 85.9

average score SR

193

score

sequence coverage

1.2

1.1

1.1

83.0

1.7

1.6

1.6

80.3

1.2

1.4

1.3

0.9

1.0

1.0

83.9

1.4

1.3

1.2

67.6

2.5

2.4

2.0

1.8

1.6

1.3

1.3 1.4

1.3 1.6

1.5 1.4

1.0

0.9

0.9

91.6

1.2

1.4

1.2

67.9

2.9

3.6

2.6

157

111 68.9 63.2 143

pe 0.019 a

matching peptides

pe 0.019

pe 0.019

For calculation of p-values Wilcoxon sign-test, one-sided, was used.

detected in SR stained samples whereas a number of such peptides were found in samples stained with CBB. In experiment 2, two proteins with fewer Cys than BSA were tested. These were enolase with one Cys and HpaA without free Cys (the N-terminal Cys is lipidated). Both were analyzed in the nonaklylated and the alkylated form. In addition, BSA in alkylated form was included to investigate if the detection of the Cys-CH2CONH2 modified peptides also differed in the CBB and the SR stained gels. In Figure 1 the MS of two enolase peptides are shown from gels stained with either CBB or SR. For both peptides the S/N and signal intensity were better for the CBB stained sample. The values for S/N, Table 2, were compared as described in the Experimental Section using the peptides listed in Table 2.

The Cys-containing residue of enolase was not detected, neither in nonalkylated nor alkylated forms (theoretical mass of m/z ) 1316.61 (z ) +1) or 1373. 61 Da, respectively). To assess the influence of the chemicals used in the alkylation procedure on the quality of MS data, HpaA which lacks free Cys was subjected to mock reduction by DTT and alkylation by IAA. On the basis of S/N, number of detected Cys peptides, scores, matching peptides and sequence coverage all samples from CBB stained gels gave better MS than those from SR stained gels (p e 0.032, Table 2). The S/N values and scores obtained for enolase and HpaA treated with IAA under reducing conditions were lower than for the nonalkylated samples. One possible explanation could be that alkylation and reduction introduce more contaminants into the samples. Journal of Proteome Research • Vol. 4, No. 1, 2005 177

technical notes

Staining with Coomassie Blue and Sypro Ruby

Figure 1. Comparison of MALDI-TOF MS of two enolase peptides (105-109, 1412.85 and 312-329, 1912.03) obtained by tryptic in-gel digestion of enolase, separated by 2-D SDS-PAGE and then stained with either CBB (top) or SR (bottom). Relative and absolute intensities are given on the Y-axis to the left and right, respectively. Table 4. Quality of MALDI-TOF MS Obtained from Liver Proteins Separated on 2-D Gels and Stained with Either CBB or SR ratio CBB/SR protein, conditions

N

no. of theor. Cys peptides

Cys detected in CBB

Cys detected in SR

score

matching peptides

sequence coverage

ATP synthase β chain glycerol-3-phosphate dehydrogenase L-lactate dehydrogenase actin superoxide dismutase cytochrome b5 apolipoprotein A-I 3-hydroxyanthranilate 3,4-dioxygenase

2 2

0 10

0 3

0 0

0.8 5.3

1.1 3.1

1.3 3.8

2 2 2 2 2 2

4 4 3 0 0 4

2 2 0 0 0 0

0 0 0 0 0 0

1.1 1.0 1.1 0.9 3.7 1.3

1.0 1.0 1.0 0.8 4.8 1.3

1.5 1.1 1.0 0.7 4.8 1.2

In experiment 3, mixtures of purified protein standards were analyzed either in a nonalkylated form or pre-alkylated with IAA. All proteins chosen, except for carbonic anhydrase, contain peptides with Cys in the mass region 850-3500 Da. When alkylated, the Cys-containg peptides were detected as either carboxyamidomethylated or propionamide derivatized. Again it was evident that more Cys-containing peptides were found when the gels had been stained with CBB compared to those stained with SR. Except for ovalbumin, all analyses showed better MS result for CBB-stained gels (p e 0.019, Table 3). 2-D Gel Electrophoresis of Rat-Liver Proteins. To perform a comparison of the two staining methods using a natural protein mixture, rat liver proteins were separated on 2-D gels. The quality of MS data obtained was assessed using the same criteria as previously for experiments 1-3. Minor differences were observed for the majority of the analyzed protein spots with the exception of glycerol-3-phosphate dehydrogenase and apolipoprotein A-I. Glycerol-3-phosphate dehydrogenase is rich in Cys and the observed superiority of CBB is due to a number of Cys-containing peptides detected in these samples.

Discussion We wanted to investigate if the fluorescent protein stain SuproRuby was compatible with PMF by MALDI-TOF MS, and we designed our analyses so that they allowed repetitive and 178

Journal of Proteome Research • Vol. 4, No. 1, 2005

well controlled comparisons. Instead of running ordinary 2-D gel separations of reference proteins, we applied them to either filter papers or to IEF-strips that were rehydrated in the protein solution but not focused (not shown). In this way, the same gel could be divided and stained in different ways and many gel excissions could be done from the same original gel ensuring equal amount of protein in each MS sample. One advantage of using the large gels with well defined homogeneous protein bands is that gel pieces can be repetatively excised until satisfactory reproducibility in the excision has been achieved. Excision accuracy is especially hard to attain with SR stained gels since the bands are invisible to the naked eye and the gels have to be aligned with a printed scan image while CBB-stained plugs are easy to excise with high accuracy since they are visible directly on the gel. The most reliable way of measuring the quality of mass spectra is to compare S/N of specified peaks between different acquisitions. This eliminates influences of e.g., the laser energy, crystallization or ionization, which measurements of base-peak height does not. Since different staining methods could give rise to different contaminants or possibly to different modifications of the proteins, modifications that are not included in the mass-score algoritms, it is also important to report comparisons of score rates.

technical notes The scope of the literature on proteomics has increased substantially over the past decade and many papers deal with various aspects of optimization of the proteomics technique. Due to the laborious 2-D gel procedure, conclusions presented in many papers are based on very few replicates why statistical methods are seldom applied. As the 2-D gel technique is inherently prone to large gel-to-gel and day-to-day variations, statistics is a prerequisite to support many of the claims in these publications. In our comparisons of the quality of the MS data obtained after staining with CBB and SR, we used an experimental protocol that allowed statistical analyses to be performed. The fluorescent protein stain SR is claimed to be fully compatible with PMF using trypsin cleavage and MALDI-TOF MS analyses.1,2 However, of the four proteins analyzed by Berggren et. al,1 two gave equal number of peptides with CBB and SR staining while the other two gave more peptides when stained with CBB. Although three of the proteins contained Cys in the mass region analyzed, 920-2500 Da; (BSA 17, soy-bean trypsin inhibitor 8, chicken ovalbumin 5) none was detected in the SR-stained gels (Table 2 in ref 1). Some of the Cyscontaining peptides, however, were detected in the spectra from CBB stained gels. Unfortunately neither of the manuscripts from this group1,2 describe how the Cys-containing peptides are searched for, i.e., as un-derivitized [m+H]+ or as acrylamide-derivitized [m+71]+. On the basis of our analyses, it can be concluded that SR lowers the chances of obtaining protein identity by PMF. However, due to the superior linear response curve of this stain, its sensitivity as well as ease of handling SR still is our first choice for 2-D gel stainings aimed at protein-expression profiling. But one must be aware of that it may be nescesary

Lanne and Panfilov

to stain gels with CBB in parallell to improve the possibility of obtaining protein identification. To elucidate the exact mechanism of the loss of Cyscontaining peptides after SR staining further investigations would be needed, e.g. by using isotope-labeled dye. It should be noted that Cys-containing peptides were in our analyses more seldom observed in MS irrespective if they are unmodified; alkylated or acrylamide modified. If the dye simply modifies some residues in such a way that they still are detectable by MS at an off-set weight, then the modification could actually improve the quality of the PMF data.

Acknowledgment. We wish to thank Dr Susann Nystro¨m for preparing the protein HpaA, Magnus A° strand for the statical analyses and Dr Tasso Miliotis and Dr Bjo¨rn Dahllo¨f for critical reading of the report. References (1) Berggren, K.; Chernokalskaya, E.; Steinberg, T. H.; Kemper, C.; Lopez, M. F.; Diwu, Z.; Haugland, R. P.; Patton, W. F. Electrophoresis 2000, 21, 2509-2521. (2) Berggren, K. N.; Schulenberg, B.; Lopez, M. F.; Steinberg, T. H.; Bogdanova, A.; Smejkal, G.; Wang, A.; Patton, W. F. Proteomics 2002, 2, 486-498. (3) Voshols, J.; Novartis Pharma, F. G. A. P. S. U., Basel, Switzerland 2003, Personal communication. (4) Lanne, B.; Potthast, F.; Ho¨glund, A° .; Brockenhuus von Lo¨wenhielm, H.; Nystro¨m, A.-C.; Nilsson, F.; Dahllo¨f, B. Proteomics 2001, 1, 819-828. (5) Go¨rg, A.; Postel, W.; Weser, J.; Gu ¨ nter, S.; Strahler, J. R.; Hanash, S. M.; Somerlot, L. Electrophoresis 1987, 8, 122-124. (6) Doherty, N.; Littman, B.; Reilly, K.; Swindell, A.; Buss, J.; Anderson, N. L. Electrophoresis 1998, 19, 355-363.

PR040005L

Journal of Proteome Research • Vol. 4, No. 1, 2005 179