Mass Spectrometric Identification of RNA Binding Proteins from Dried EMSA Gels Dominique Stenger, Wilhelm Gruissem, and Sacha Baginsky* Institute of Plant Science, Swiss Federal Institute of Technology, Zu ¨ rich, Switzerland Received January 21, 2004
Abstract: Electrophoretic mobility shift assays (EMSA) are commonly employed for the analysis of nucleic acid/ protein interactions with a native gel system. Here, we report a method to identify RNA binding proteins from a dried EMSA gel by mass spectrometry following autoradiography. Compared to wet gel exposure, our approach resulted in an improved protein identification sensitivity and RNA/protein complex isolation accuracy. The method described here is useful for the large scale characterization of RNA- or DNA-protein complexes. Keywords: mass spectrometry • protein/nucleic acid interaction • EMSA
The analysis of protein/nucleic acid interactions is an important analytical step for the study of regulatory processes in gene expression. Prominent examples are promoter studies conducted with RNA polymerase complexes (reviewed in Martinez, 2002; Schramm and Hernandez, 2002). In principle, these analyses allow for the identification of protein factors involved in processes such as transcription initiation, RNA stability regulation and translation. One method that has been successfully employed to study the interaction of DNA or RNA with a protein complex is the “electrophoretic mobility shift assay (EMSA)”. EMSAs are especially useful to analyze the dynamic subunit composition of nucleic acid binding protein complexes under a variety of conditions especially when they are combined with protein identification techniques. In a typical EMSA experiment, the nucleic acid is radiolabeled (probe) and its interaction with proteins monitored by its mobility shift in a native gel system compared to a control (see Figure 1, A and B). Protein/nucleic acid complexes are visualized by autoradiography. For routine analyses, the gel is dried prior to autoradiography to avoid diffusion of the binary complex during the exposition time. Here, we report the identification of RNA binding proteins from a dried EMSA gel by mass spectrometry. The identification of proteins from dried EMSA gels harbors the possibility to directly identify proteins from a binary protein/nucleic acid complex and to identify RNA or DNA binding proteins on a large scale. The advantage of isolating a protein/nucleic acid complex from a dried gel compared to a wet gel exposure is * To whom correspondence should be addressed. Institute of Plant Sciences, Swiss Federal Institute of Technology, ETH Zentrum, LFW E18, Universita¨tstrasse 2, CH-8092 Zu ¨ rich, Switzerland. Phone: +41 1 632 3866. Fax: +41 1 632 10 79. E-mail:
[email protected].
662
Journal of Proteome Research 2004, 3, 662-664
Published on Web 03/26/2004
Figure 1. (A) Analysis of RNA binding proteins from a complex chloroplast protein extract using the radiolabeled petD-3′-UTR as a probe in a native gel system. Presented is the autoradiograph of an EMSA gel with (dry) or without (wet) previous gel drying on Whatman paper. (B) Analysis of RNA binding proteins from a partially purified chloroplast extract and a commercially available preparation of E. coli polynucleotide phosphorylase (PNPase) (Sigma). The chloroplast protein extract described above (A) was further fractionated by gel filtration chromatography (GF). Fractions corresponding to a molecular mass of 2045 kDa were tested for RNA binding activities as described above. Preparations of E. coli PNPase were assayed for RNA binding activities as described above. Control lanes contained the radiolabeled RNA probe but no protein. RNA/protein complexes 1 to 5 were cut out and subjected to mass spectrometric protein identification.
the higher isolation accuracy of the protein/nucleic acid complex since diffusion effects are minimized (compare Figure 1 A, dry and wet). Additionally, exposition times with dried gels are reduced since the quenching effect of water is diminished and an enhancer screen can be used at -80°. For the experiment reported here, a soluble protein extract was prepared from spinach chloroplasts as described previously (Baginsky and Gruissem, 2001). An RNA probe was synthesized by in vitro transcription with T7 RNA polymerase in the presence of R-32P-UTP following standard protocols. The probe contained the 3′-UTR of the chloroplast gene for subunit IV of the cytochrome b6/f complex (Baginsky and Gruissem, 2001). To analyze 3′-UTR binding protein complexes in the soluble extract we mixed 5 µg of soluble protein with 2 µL of RNA 10.1021/pr049966q CCC: $27.50
2004 American Chemical Society
technical notes substrate (50 000 cpm/µl), 3.75 mM MgCl2, 2 mM DTT, 10 mM KCl, 1 mM sodium phosphate (pH 7) and 20 U RNAsin in a total volume of 10 µL in an Eppendorf tube. The mixture was incubated for 10 min at room temperature to allow protein/ RNA complexes to form. Following the incubation time, the mixture was run on a cooled native gel (5% acrylamide, 0.5 M Tris/HCl pH 8.0) at 200 V for about 40 min. One gel was subsequently dried on Whatman paper and subjected to autoradiography overnight at -80 °C. The other gel was not dried (wet) and subjected to autoradiography at room temperature. Figure 1 shows a comparison between a dry gel exposure and a wet gel exposure. Two major complexes were clearly visible after the dry gel exposure, whereas a diffuse smear was visible after the wet gel exposure (Figure 1 A). The protein/RNA complex with the higher electrophoretic mobility was cut out from both gels (Figure 1 A indicated with 1 (wet gel) or 2 (dry gel) respectively) and the gel slice originating from the dried gel was allowed to rehydrate in water for 10 min. Following rehydration, the gel piece and the Whatman paper were separated by using clean forceps. Both gel pieces were subsequently subjected to in gel tryptic digest essentially as described (Shevchenko et al., 1996). The gel piece from the dried gel still contained paper fibers from the drying procedure which were partially removed by centrifugation of the collected supernatants from the tryptic digest procedure (Eppendorf centrifuge, 10 min, 13 000 rpm). The supernatant was carefully transferred to another tube and the collected supernatants were subsequently evaporated in a speed vac. Following evaporation, the dried peptide mixture (still some paper fibers were visible from the dried gel slice) was resuspended in “buffer A” (10 µl of 94.5% water, 5% acetonitrile and 0.5% formic acid). To separate paper fibers from the peptides and to remove salts originating from the in gel digest the solution was pipeted through C18 material [ZipTips (Millipore, Bedford)] that had been conditioned with “buffer A” prior to use. Under these conditions peptides bind to the C18 material through hydrophobic interactions while salts and paper fibers can be washed away efficiently. Altogether, peptides from both gel slices were bound to three ZipTips. Bound peptides were eluted from the C18 material with “buffer B” (95% acetonitrile, 4.5% water and 0.5% formic acid) and the peptide solution was again evaporated in a speed vac. Following evaporation the dried pellet was resuspended in 10 µL “buffer A” (94.5% water, 5% acetonitrile and 0.5% formic acid) and reversed phase chromatography (C18 material) for peptide separation was performed prior to mass spectrometry (Shabanowitz et al., 2000). MS-scans were defined as “data dependent acquisition” using one full scan and four MS/MS scans with the four most intense ions. Protein identification was performed using the non redundant NCBI protein database (ftp://ftp.ncbi.nih.gov./blast/db) and the SEQUEST search program (ThermoFinnigan, San Jose). For verification suggested peptide identifications from SEQUEST (Xcorr > 2.5, dCN > 0.1, Ion Ratio > 50%) were visually examined. Aldolase and glutamate synthase (FD-GOGAT) were identified from both complexes (Table 1). Alanine aminotransferase, elongation factor TU and porphobilinogen synthase were exclusively identified from complex 2 whereas the large subunit of RubisC/O (LSU) and glutamine synthetase were exclusively identified from complex 1 (Table 1). A complementary experiment using single strand (ss) DNA affinity chromatography was performed to verify the RNA binding activity of the identified proteins (data not shown). RubisC/O (LSU), glutamine synthetase, RubisC/O activase and
Stenger et al. Table 1. Proteins Identified from Complexes 1-5 by LC-ESI-MS/MS complex
peptides
acc. no.
protein
1
3 2 1 1 9 7 3 1 1 1 2 1 1 6 1 1 1 11 7 7 7 5 4 4
gi 2149531 gi 4884970 gi 2494793 gi 68200 gi 68200 gi 2494793 gi 132168 gi 6175148 gi 2494261 gi 99556 gi 68200 gi 336390 gi 2494793 gi 78768 gi 15831867 145341 119201 gi 121416 gi 145341 gi 78768 gi 121882 gi 434010 gi 42837 gi 42818
RubisC/O LSU glutamine synthetase FD-GOGAT fructose-bisphosphate aldolase fructose-bisphosphate aldolase FD-GOGAT RubisC/O activase alanine aminotransferase elongation factor TU porphobilinogen synthase fructose-bisphosphate aldolase GAP-dehydrogenase FD-GOGAT PNPase ferritin argininosuccinate synthase elongation factor TU glycerate kinase argininosuccinate synthase PNPase DNA gyrase pyruvate dehydrogenase ribosomal protein S1 RNA polymerase beta
2
3 4
5
a Proteins identified from the RNA binding complexes are listed (right) including the number of peptide identifications and the NCBI reference entry (gi number, http://www.ncbi.nlm.nih.gov/).
aldolase bound to the ssDNA matrix suggesting that they can either bind nucleic acids directly or interact with a nucleic acid binding protein. We have shown that aldolase can be crosslinked to the petD 3′-UTR which confirms previous reports about the 3′-UTR binding activity of aldolase (Kiri and Goldspink, 2002). This activity, however, is unspecific as suggested by competition experiments with E. coli t-RNA as unspecific competitor (data not shown). It is conceivable that the other identified proteins comigrated with the RNA/protein complex since we used a complex protein mixture for the experiment reported. To test this possibility we separated aldolase from potential contaminants by gel filtration chromatography on Superose 6 (Baginsky and Gruissem, 2001). Fractions containing aldolase (GF 20-45 kDa) were tested for RNA binding activity as described above. One major RNA/protein complex formed during the native gel electrophoresis (Figure 1 B, complex 3) that comigrated with aldolase, glyceraldehyde-3-phosphate dehydrogenase (GAP-DH) and Fd-glutamate synthase (FdGOGAT). Since neither GAP-DH nor Fd-GOGAT bind the ssDNA chromatography matrix efficiently (see above, data not shown), these data strongly suggest that aldolase is the major unspecific RNA binding constituent in the protein fractions analyzed. To further investigate the potential of the described method, we analyzed the RNA binding activities in a commercially available preparation of E. coli polynucleotide phosphorylase (Sigma) (Figure 1B). PNPase is a 3′ to 5′ exonuclease that has a reported RNA binding activity. Two major RNA/protein complexes formed with proteins from this protein preparation (Figure 1B, complex 4 and 5) both of which contained PNPase (Table 1). This suggests that PNPase can potentially form different complexes with the petD 3′-UTR probe. Some of the other proteins identified from complex 5 play a role in RNA or DNA metabolism such as the ribosomal protein S1, the RNA polymerase subunit beta and DNA gyrase. Most likely, these proteins interact with the probe nonspecifically. Journal of Proteome Research • Vol. 3, No. 3, 2004 663
technical notes
Protein/Nucleic Acid Interactions
Together these data emphasize the suitability to combine electrophoretic mobility shift assays with mass spectrometric protein identification for a first analysis of nucleic acid binding proteins from protein mixtures. Further investigations are necessary to confirm and further characterize these activities. Additionally, pre-fractionation of protein extracts prior to EMSA analyses will help to minimize the number of contaminating proteins that fortuitously comigrate with the RNA binding proteins in the native gel system. Taken together protein identification from a dried EMSA gel is feasible and does not result in a considerable loss of sensitivity compared to protein identification from a wet gel (Figure 1, Table 1). In contrast, protein isolation from a dried gel improved protein identification sensitivity as suggested by the number of peptides identified from each protein (Figure 1A, Table 1). Even after years of
664
Journal of Proteome Research • Vol. 3, No. 3, 2004
storage, efficient protein identification from a dried EMSA gel can be achieved with the described method (data not shown).
References (1) Baginsky, S.; Gruissem, W. Methods Enzymol. 2001, 342, 408419. (2) Kiri, A.; Goldspink, G. J. Muscle Res. Cell Motil. 2002, 23, 119129. (3) Martinez, E. Plant Mol. Biol. 2002, 50, 925-947. (4) Schramm, L.; Hernandez, N. Genes Dev. 2002, 16, 2593-2620. (5) Shabanowitz, J.; Settlage, R. E.; Marto, J. A.; Christian, R. E.; White, F. M.; Russo, P. S.; Martin, S. E.; Hunt, D. F. In. Mass Spectrometry in Biology and Medicine; Burlingame, A. L., Carr, S. A., Baldwin, M. A., Eds.; Humana Press: Totowa, N. J., 2000, 163-177. (6) Shevchenko, A.; Wilm, M.; Vorm, O.; Mann, M. Anal. Chem. 1996, 68, 850-858.
PR049966Q
