ANALYTICAL BIOCHEMISTRY Analytical Biochemistry 329 (2004) 289–292 www.elsevier.com/locate/yabio
A proteomic approach to identify phosphoproteins encoded by cDNA libraries Xudong Shi, Robert J. Belton Jr., Heather R. Burkin, Ana P. Vieira, and David J. Miller¤ Department of Animal Sciences, University of Illinois, 1207 West Gregory Drive, Urbana, IL 61801, USA Received 19 December 2003 Available online 6 May 2004
Abstract We report a method for large-scale rapid analysis of phosphoproteins in tissues or cells by combining immobilized metal aYnity chromatography (IMAC) with phage display cDNA library screening. We expressed a testis cDNA library as fusion proteins on phage and, using IMAC, enriched for sequences encoding phosphoproteins. Selected clones were polymerase chain reaction ampliWed and sequenced. The majority of the clones sequenced (80%) encoded known proteins previously identiWed as phosphoproteins. Immunoblotting with phosphotyrosine antibodies conWrmed that some of the selected sequences encoded tyrosine phosphorylated proteins when expressed on phage. An advantage of this method is the rapid identiWcation of phosphoproteins encoded by a cDNA library, which can identify proteins that are potentially phosphorylated in vivo. When this method is combined with limited enzymatic digestion and tandem mass spectrometric techniques, the speciWc phosphorylation site in a protein can be identiWed. This technique can be used in proteomics studies to eVectively detect phosphorylated proteins and avoid time-consuming and expensive peptide sequencing. 2004 Elsevier Inc. All rights reserved.
As the number of genomes being sequenced increases, emphasis is shifting toward translating DNA sequence information to develop an understanding of protein function. One approach to the study of protein function is through the use of recombinant protein expression. However, the expression and characterization of individual proteins is time consuming. Herein we demonstrate a method for the isolation and characterization of numerous cDNA sequences that encode phosphoproteins. Current estimates are that at least 30% of all proteins are phosphorylated [1]. Protein phosphorylation can alter the function of the protein by altering its enzyme activity (allosteric regulation), its aYnity for other proteins, its stability, and/or its localization [2]. Because phosphorylation is a dynamic process, to understand phosphoprotein regulation in a mammalian cell, it is necessary to acquire an inventory of phosphoproteins and their phosphorylation sites under diVerent conditions, recently termed phosphoproteomics [3]. Because many proteins are phosphorylated on multiple sites and the site at ¤
Corresponding author. Fax: 1-217-333-8286. E-mail address:
[email protected] (D.J. Miller).
0003-2697/$ - see front matter 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2004.03.033
which they are phosphorylated dictates the speciWc functional changes, it is of value to identify the speciWc phosphorylated amino acids in the protein and study the dynamic changes of the phosphoproteins under diVerent conditions. However, current methods have pitfalls in isolation, identiWcation, and characterization of phosphoproteins. Due to the high number and low abundance of many phosphoproteins, particularly those involved in signal transduction, one of the major challenges is to enrich the abundance of phosphoproteins so that they can be sequenced by conventional methods such as Edman sequencing or by more sensitive mass spectrometric techniques. Traditionally, tyrosine phosphorylated proteins have been separated by immunoaYnity chromatography using phosphotyrosine antibodies and identiWed by Edman degradation or mass spectrometry [1,4]. This labor-intensive technique is especially diYcult when studying proteins in low-abundance, a characteristic of most tyrosine phosphorylated proteins. Detecting phosphorylated proteins and other low-abundance proteins and identifying phosphorylation sites of the proteins are diYcult tasks using current technology, even though
290
X. Shi et al. / Analytical Biochemistry 329 (2004) 289–292
mass spectrometric techniques are much more sensitive than conventional Edman degradation for protein sequencing. Although immobilized metal aYnity chromatography (IMAC)1 was Wrst used to isolate metal-ion-binding proteins containing many imidazole and thiol groups [5], in recent years, IMAC has been used successfully to isolate phosphoproteins from cell homogenates [6,7]. IMAC using Fe3+ appears particularly useful for this purpose [8]. However, due to the complexity of the proteins in eukaryotes and the low abundance of signaling proteins, it is not likely to yield good coverage of all the phosphorylated proteins without further ampliWcation and enrichment of less-abundant phosphoproteins [9]. We combined IMAC with expression screening on phage to address this problem.
Materials and methods Preparation of the aYnity column Nickel nitrilotriacetic acid agarose resin (Ni-NTA; Qiagen, Valencia, CA) was stripped of Ni2+ and recharged with Fe3+ as follows: 200 l Ni-NTA agarose was washed three times with phosphate-buVered saline and three times with distilled H2O and then resuspended in 1000 l of 50 mM EDTA for 30 min at room temperature. The EDTA treatment was repeated twice and the resin was washed with 1 ml distilled H2O four times. Immediately before use, the resin was charged with FeCl3; 300 l of 20 mM FeCl3 was incubated with NTA resin for 30 min at room temperature. After the procedure was repeated once, excess FeCl3 was removed and the resin was washed four times with 10 ml H2O. Phage library production and selection A pig testis cDNA library in T7 phage (10-3b vector, Novagen, Madison, WI) was constructed. The vector displays 5–15 fusion peptides per phage as fusions of library-encoded proteins to the C terminus of a capsid protein. The library was screened by aYnity of the phage for Fe-NTA agarose beads. BLT5615 cells (Novagen, Madison, WI) were incubated in 20 ml of LB overnight. In a 50-ml conical tube, 1 ml of overnight cultured BLT5615 cells was incubated with 20 ml of LB media for 1 h. Isopropyl -D-thiogalactoside (1 mM Wnal concentration) was added to induce the expression of capsid protein along with 20 l of cDNA library (108 phage); 10 mM ATP and phosphatase inhibitors (5 mM NaF and 2 mM sodium vanadate, Sigma, St. Louis, MO) were added. The BLT5615 cells were lysed by library phage in 1
Abbreviations used: IMAC, immobilized metal aYnity chromatography; NTA, nitrilotriacetic acid.
2–3 h; 100 l of chloroform was added to kill cells and the samples were centrifuged at 10,000g for 20 min. The phage lysate in the supernatant was pooled and the pH adjusted to 3–3.5 with acetic acid. Phage were added to the Fe-NTA column equilibrated in the same solution. After a 1-h incubation, unbound phage were washed from the column with 6 ml of 40 mM acetic acid. Retained phage were eluted with 200 mM sodium phosphate at pH 8.4 and added to BLT5615 cells for multiplication and subsequent rounds of selection, as described above. After Wve rounds of selection, the collected phage speciWcally bound to the aYnity column were diluted to 108 to 106 pfu/ml for plaque assay. Plaque assay After Wve rounds of aYnity selection, the selected phage was serially diluted using LB medium. Top agar was melted in a 55 °C water bath. In a 15-ml conical tube, 1 ml of the prepared BLT5615 cells was added. Then 100 l of the diluted phage and 10 ml of the melted top agar were mixed together and poured on top of a LB agar plate with 100 g/ml of ampicillin. The plate was incubated at room temperature for 30 min and then transferred to a 37 °C incubator for 2 h. Plaques were counted on the plate. Individual clones were PCR ampliWed and sequenced. PCR ampliWcation and DNA sequencing Phage clones were picked up from the plaque assay using a glass pipette and incubated with 50 l of LB media for 15 min. The LB solution containing phage was heated at 65 °C for 10 min followed by centrifugation at 15,000g for 5 min. The phage lysate was used as a template for PCR using T7 primers (Novagen, Madison, WI). The following components were added to a thinwalled PCR tube for a 50-l PCR: 2 l of phage lysate as a template, 5 l of 10£ Taq PCR buVer (100 mM KCl, 100 mM Tris–HCl, pH 8.3), 3 l of 5 mM MgCl2, 1 l of T7 select UP primer (5 pmol/l), 1 l of T7 select DOWN primer (5 pmol/l), 1 l of dNTP mixture (10 mM each dATP, dGTP, dCTP, and dTTP), 2.5 l of Taq DNA polymerase, and 37.5 l of water. The PCR condition used were 95 °C for 1 min and then repeated for 35 cycles as 94 °C for 1 min 60 °C for 1 min, and 72 °C for 1 min. Last cycle extension was for 10 min at 72 °C. Eight tubes of 50 l PCR were combined and products puriWed with a Qiagen PCR puriWcation kit (Qiagen, Valencia, CA) for DNA sequencing. Western blot After the fourth and Wfth rounds of selection, the phage selected were ampliWed and protein from 300 l of a phage solution containing 4 £ 108 pfu/l was separated
X. Shi et al. / Analytical Biochemistry 329 (2004) 289–292
on an SDS–PAGE gel and transferred to nitrocellulose membrane. The membrane was blocked with 5% BSA for 1 h at room temperature. Tyrosine-phosphorylated proteins were detected by immunoblotting with an antiphosphotyrosine monoclonal antibody (1:1000 dilution, Upstate Biotechnology, Charlottesville, VA).
Results and discussion To identify DNA sequences encoding phosphoproteins, we fused sequences from a pig testis cDNA library to a gene encoding a phage coat protein and the fusion protein was expressed on the phage surface. Because genes encoding kinases, phosphatases, and other enzymes from the library are also expressed as fusions on phage, posttranslational modiWcations of the proteins that occur in the tissue might also be found on the phage surface, if conditions are appropriate and substrates and cofactors are available. Using this approach, phage expressing testis cDNAs were aYnity-selected using Wve successive rounds of IMAC. Fifteen clones with aYnity for the matrix were randomly selected for DNA sequence analysis. All 15 porcine sequences were greater than 90% homologous to human or mouse genes. Among the 15 IMAC-selected phage clones, 12 clones (80%) had library inserts encoding putative phosphoproteins, according to previous studies of these proteins. Of the 12 cDNAs encoding phosphoproteins, 9 clones (60%) encoded serine/threonine phosphorylated proteins and 3 clones (20%) encoded tyrosine phosphoproteins (Table 1). This list includes kinases that can phosphorylate themselves and proteins that lack kinase activity. Three selected DNA sequences encoded proteins that do not have phosphorylation sites. Their selection can probably be explained by the aYnity of their acidic D/E-rich domains (enriched in
291
aspartate and glutamate) for the IMAC resin, as previous reports have found [7,10,11]. As a control, IMAC selection was performed in the absence of ATP and phosphatase inhibitors. As assessed by plaque assays and phosphotyrosine antibody analysis, no increase in phosphotyrosine-containing proteins was observed (data not shown). To conWrm that fusion proteins on the selected phage were phosphorylated, we performed Western blots using phosphotyrosine antibodies. Results showed that some mammalian proteins expressed on the phage surface were phosphorylated on tyrosine residues (Fig. 1). Because tyrosine phosphorylation is typically less abundant than serine and threonine phosphorylation, serine and threonine phosphorylation are also expected to occur. Additionally, the observation that proteins were phosphorylated demonstrates that kinases from the testis library were active, because bacteria and phage have a limited repertoire of protein kinases [12–15]. With each round of selection, we would expect that fewer nonspeciWcally bound proteins would adhere to the Fe3+ resin during IMAC, as supported by Fig. 1. Because multiple rounds of selection would also separate phosphoproteins from kinases that were not phosphoproteins, with several rounds of IMAC selection one would expect to eventually Wnd fewer phosphoproteins, as the kinases were removed. Therefore, although they may increase the speciWcity, additional rounds of selec-
Table 1 Proteins encoded by Fe-NTA-selected phage clones Protein selected
Residue phosphorylateda
Paxillin Proto-oncogene tyrosine-protein kinase Lck Shc transforming protein cAMP-dependent protein kinase Glutamate (NMDA) receptor subunit zeta Glycogen phosphorylase Phenylalanine-4-hydroxylase (PAH) Phosphorylase b kinase, alpha regulatory chain Sodium channel protein, brain II alpha Pleckstrin (platelet p47 protein) Protamine (clupeine) YI Nucleolin Major centromere auto antigen B Protein HSPC020 Testican -3 precursor
Y Y Y S S S S S S S/T S/T S/T
a
Y, tyrosine; S, serine; and T, threonine.
Fig. 1. Western blot with phosphotyrosine antibody detected phosphorylated proteins produced by selected phage. A pig testis cDNA phage display library was screened with IMAC for Wve rounds. After the fourth and Wfth rounds, the phage selected were ampliWed and protein from 300 l of a phage solution (4 £ 108 pfu/l) was separated on an SDS–PAGE gel and transferred to nitrocellulose. The signal was detected with phosphotyrosine antibodies and horseradish-peroxidase-conjugated secondary antibody. Lanes 1 and 2, testis phage display cDNA library; lane 3, testis cDNA library after four rounds of IMAC selection; lane 4, testis cDNA library after Wve rounds of IMAC selection. No bands were observed in the absence of phosphotyrosine antibody.
292
X. Shi et al. / Analytical Biochemistry 329 (2004) 289–292
tion are expected to result in a loss of phosphoproteins. However, additional rounds of selection are expected to enrich those phosphoproteins that are also kinases and that phosphorylate themselves. Phage display with IMAC selection is a powerful method to identify signaling proteins and determine phosphorylation sites of the proteins. The high percentage of phosphorylated proteins in the sequenced phage clones conWrms the eYciency and reliability of the method used. Furthermore, the high sequence coverage of phage display allows the identiWcation of a large number of the phosphoproteins expressed by the tissue from which the library was constructed. Several proteins were identiWed that, at least in vivo, have multiple phosphorylation sites, suggesting that this method can help identify speciWc peptides that are phosphorylated. A drawback of this method is that the majority of proteins expressed on the phage surface are not fulllength proteins. Furthermore, expression of library cDNAs as fusion proteins with a phage coat protein may aVect phosphorylation status. It is not known how closely the protein phosphorylation pattern in the selected phage represents the phosphorylation pattern in the tissue from which the library was constructed. Finally, due to removal of kinases during IMAC selection, the repertoire of selected phosphoproteins may change with additional selection rounds. However, this method can yield sequence information for a large number of phosphoproteins very rapidly and has the potential to be very important in identifying phosphoproteomes.
[2]
[3] [4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
Acknowledgments This work was supported by the University of Illinois Agricultural Experiment Station as part of Hatch Project ILLU-35-0335 and the National Institutes of Health (HD38311).
[12]
[13]
[14]
References [15] [1] S.B. Ficarro, M.L. McCleland, P.T. Stukenberg, D.J. Burke, M.M. Ross, J. Shabanowitz, D.F. Hunt, F.M. White, Phosphoproteome
analysis by mass spectrometry and its application to Saccaromyces cerevisiae, Nat. Biotechnol. 20 (2002) 301–305. T.P. Conrads, H.J. Issaq, T.D. Veenstra, New tools for quantitative phosphoproteome analysis, Biochem. Biophys. Res. Commun. 290 (2002) 885–890. D.F. Stern, Phosphoproteomics, Exp. Mol. Pathol. 70 (2001) 327– 331. A. Stensballe, S. Andersen, O.N. Jensen, Characterization of phosphoproteins from electrophoretic gels by nanoscale Fe(III) aYnity chromatography with oV-line mass spectrometry analysis, Proteomics 1 (2001) 207–222. J. Porath, J. Carlsson, I. Olsson, G. Belfrage, Metal chelate aYnity chromatography, a new approach to protein fractionation, Nature 258 (1975) 598–599. K.A. Powell, V.A. Valova, C.S. Malladi, O.N. Jensen, M.R. Larsen, P.J. Robinson, Phosphorylation of dynamin I on Ser-795 by protein kinase C blocks its association with phospholipids, J. Biol. Chem. 275 (2000) 11610–11617. D.C.A. Neville, C.R. Rozanas, E.M. Price, D.B. Gruis, A.S. Verkman, R.R. Townsend, Evidence for phosphorylation of serine 753 in CFTR using a novel metal–ion aYnity resin and matrixassisted laser desorption mass spectrometry, Protein Sci. 6 (1997) 2436–2445. T.S. Nuhse, A. Stensballe, O.N. Jensen, S.C. Peck, Large-scale analysis of in vivo phosphorylated membrane proteins by immobilized metal ion aYnity chromatography and mass spectrometry, Mol. Cell. Proteomics (2003). S. Ficarro, O. Chertihin, V.A. Westbrook, F. White, F. Jayes, P. Kalab, J.A. Marto, J. Shabanowitz, J.C. Herr, D.F. Hunt, P.E. Visconti, Phosphoproteome analysis of capacitated human sperm. Evidence of tyrosine phorphorylation of A Kinase-Anchoring Protein 3 and valosin-containing protein/p97 during capacitation, J. Biol. Chem. 278 (2003) 11579–11589. G. Muszynska, G. Dobrowolska, A. Medin, P. Ekman, J.O. Porath, Model studies on iron(III) ion aYnity chromatography. II. Interaction of immobilized iron(III) ions with phosphorylated amino acids, peptides and proteins, J. Chromatogr. 604 (1992) 19–28. M.C. Posewitz, P. Tempst, Immobilized gallium(III) aYnity chromatography of phosphopeptides, Anal. Chem. 71 (1999) 2883– 2892. A.J. Cozzone, Regulation of acetate metabolism by protein phosphorylation in enteric bacteria, Annu. Rev. Microbiol. 52 (1998) 127–164. E.S. Robertson, A.W. Nicholson, Protein kinase of bacteriophage T7 induces the phosphorylation of only a small number of proteins in the infected cell, Virology 175 (1990) 525–534. I. Mijakovic, S. Poncet, G. Boel, A. Maze, S. Gillet, E. Jamet, P. Decottignies, C. Grangeasse, P. Doublet, P. Le Marechal, J. Deutscher, Transmembrane modulator-dependent bacterial tyrosine kinase activates UDP-glucose dehydrogenases, EMBO J. 22 (2003) 4709–4718. P.J. Kennelly, Archael protein kinases and protein phosphatases: insights from genomics and biochemistry, Biochem. J. 370 (2003) 373–389.
