Selective Protein Staining On Native Gel
- 1. Subscriber access provided by UIC Library Article Selective Staining of Proteins with Hydrophobic Surface Sites on a Native Electrophoretic Gel Martina Bertsch, and Richard J. Kassner Journal of Proteome Research, 2003, 2 (5), 469-475• DOI: 10.1021/pr025579+ • Publication Date (Web): 16 August 2003 Downloaded from http://pubs.acs.org on April 2, 2009 More About This Article Additional resources and features associated with this article are available within the HTML version: • Supporting Information • Links to the 1 articles that cite this article, as of the time of this article download • Access to high resolution figures • Links to articles and content related to this article • Copyright permission to reproduce figures and/or text from this article Journal of Proteome Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036
- 2. Selective Staining of Proteins with Hydrophobic Surface Sites on a Native Electrophoretic Gel Martina Bertsch† and Richard J. Kassner* Department of Chemistry, University of Illinois at Chicago, 845 West Taylor Street, Suite 4500, Chicago, Illinois 60607 Received October 14, 2002 Chemical proteomics aims to characterize all of the proteins in the proteome with respect to their function, which is associated with their interaction with other molecules. We propose the identification of a subproteomic library of expressed proteins whose native structures are typified by the presence of hydrophobic surface sites, which are often involved in interactions with small molecules, membrane lipids, and other proteins, pertaining to their functions. We demonstrate that soluble globular proteins with hydrophobic surface sites can be detected selectively by staining on an electrophoretic gel run under nondenaturing conditions. The application of these staining techniques may help elucidate new catalytic, transport, and regulatory functionalities in complex proteomic screenings. Keywords: chemical proteomics • subproteomic library • hydrophobic surface sites • polarity sensitive staining • nondenaturing gel electrophoresis Introdution proteomic sample. Developed by O’Farrell in 1975, this tech- nology separates proteins based on differences in molecular One of the goals of proteomics is to characterize all of the charge and mass.2 The protein mixture of interest, which may proteins in the proteome with respect to their structure and be a full tissue or body fluid extract or a fraction of the same, function. Thousands of new proteins have been predicted from is applied to an isoelectric focusing (IEF) gel or an IPG strip the 30 000-40 000 genes identified in the human genome. A with little or no prior sample preparation. In an electric field, still much larger number of proteins are thought to result from a protein will migrate toward the pH region on the IEF gel or post-transcriptional and post-translational modifications, e.g., the IPG strip in which the pH equals its isoelectric point. After glycosylation, phosphorylation, hydroxylation, methylation, the first dimension separation based on differences in isoelec- acetylation, ubiquitinylation, and so forth. Thus, although the tric points, the IEF gel or the IPG strip is laid atop the second number of unmodified proteins coded by the genome may be between 80 000 and 100 000, the complete human proteome dimension polyacrylamide sodium dodecyl sulfate gel (PAGE- is thought to contain close to 1 000 000 proteins.1 To date, only SDS), and the proteins are run in an orthogonal direction, to a fraction of these proteins has been fully characterized. achieve separation by size of fully unfolded, SDS coated protein molecules. The gels are stained with chromophores, fluoro- The major technology platforms that have been developed for proteomics include 2-D gel electrophoresis, gel image phores, or silver that bind to all proteins nonspecifically. The analysis software, affinity and multidimensional high perfor- stained gels are then photographed utilizing digital imaging mance liquid chromatography (HPLC), mass spectrometry systems, with a broad range of image enhancing features. Appli- (MS), protein shotgun sequencing, isotope-coded affinity tag cation of fluorescent staining dyes, which bind noncovalently (ICAT) reagent comparative quantification method, chip based to proteins, allows for the extraction of selected spots after microassays, two-hybrid assays, phage display assays, biomo- excision and further investigation via proteolytic fragmentation lecular interaction analysis through surface plasmon resonance and MS identification of signature peptides. This technology (SPR) coupled with matrix-assisted laser desorption ionization has provided important information about the expression, time-of-flight (MALDI-TOF) MS, bioluminescence and fluores- molecular weight, isoelectric point, and sequence of new cence resonance energy transfer (BRET and FRET), X-ray and proteins. NMR analysis of protein three-dimensional structure. By examining images of 2-D gels of protein extracts from Two-dimensional (2-D) electrophoresis has been the method tissues or cells at different stages of development, under of choice for the separation of the proteins in a complex different growth conditions, in the presence of drug molecules or upon infection with a pathogen, researchers have looked * To whom the correspondence should be addressed. Telephone: (312) for changes in the spotting pattern: (1) movements of indi- 996-5202. Fax: (312) 996-0431. E-mail: [email protected]. † Current address: Department of Molecular Pharmacology and Biological vidual spots, which may be caused by alterations in the post- Chemistry (S215), The Feinberg School of Medicine, Northwestern University. translational modification, e.g., phosphorylation, or (2) the 10.1021/pr025579+ CCC: $25.00 © 2003 American Chemical Society Journal of Proteome Research 2003, 2, 469-475 469 Published on Web 08/16/2003
- 3. research articles Bertsch and Kassner disappearance of a spot or the creation of a new one, signaling cellular Zn-endopeptidase with its peptide substrate is an the shutting off or turning on of the gene expression.3 illustration of a hydrophobic enzyme-protein substrate as- With significant technical improvements, 2-D gel electro- sociation and is an important subject of metastasis research.15 phoresis continues to be the primary high-resolution separation The most abundant bee toxin peptide, melittin, is shown to method for complex protein mixtures. Although this platform bind bovine brain calmodulin through hydrophobic interac- can lead to the establishment of a more complete library of tions.16 The surface hydrophobicities of eleven staphylococcal expressed proteins and thus the identification of new proteins enterotoxins (SE) have been estimated and compared with and their primary structural characterization, other approaches those of standard proteins on an octyl agarose column by high are necessary to elucidate the native structure and function of performance hydrophobic interaction chromatography.17 these new proteins. Fibronectin is a structural protein that promotes cell attach- Nondenaturing or native electrophoresis offers a different ment. Fibronectin activation is ascribed to the nonspecific perspective to the protein-mixture composition. Native three- binding of hydrophobic substrates.18 The glycine rich structural dimensional structure or conformation is maintained through protein (GRP) of French bean is part of a repair mechanism to support of stabilizing noncovalent interactions. Disulfide bridges strengthen the cell walls during elongation growth of the plant. and noncovalent linkages between subunits are preserved in a The complete extraction of GRP is only possible by using a biologically functional multimeric protein. The metal ions and detergent, indicating the hydrophobic character of interprotein prosthetic groups are still attached to polypeptide chains in a contacts.19 holoprotein unit. The pattern of separation resulting from this Among the examples of membrane protein-protein hydro- type of electrophoresis emphasizes the native charge-to-mass phobic interaction is an oncoprotein binding to an integral ratio. The separation may be followed by a selective staining membrane protein. A study of a highly hydrophobic bovine technique, which highlights the presence of proteins that bind papillomavirus protein involved in cell transformation reveals certain ligand types, e.g., glycoprotein staining or staining of its specific binding to β-type platelet derived growth factor proteins with hydrophobic surfaces developed in this work. receptor.20,21 The receptor activation seems to be a direct Excision and subsequent structure and function assays of consequence of the complex formation. selected spots containing proteins of interest, may also be Integral membrane protein-lipid hydrophobic interactions implemented, if the staining is reversible and the dyes may be are exemplified in studies of cytochrome c oxidase solubiliza- dialyzed after excision. tion in phosphatidylcholate and membrane reconstitution,22,23 We propose the establishment of a subproteomic library of the antimicrobial activity of certain natural peptides due to expressed soluble globular proteins whose native structures are membrane permeabilization and extensive pore formation24 typified by the presence of surface hydrophobic sites. Although and a molecular dynamics simulation of an interaction of the the majority of hydrophobic amino acid groups are buried in enzyme phospholipase A2 with a phospholipid monolayer.25 a nonpolar protein interior, in some proteins nonpolar residues The hydrophobic sites on protein surfaces may also become are found on the surface, which is expected to be polar due to exposed as a consequence of conformational changes in re- its aqueous exposure. These nonpolar surface residues may sponse to environmental stress or during protein denatur- form hydrophobic sites of biologically relevant interaction with ation.26 small effector molecules, e.g., receptor-ligand, enzyme- The presence of hydrophobic groups on the surface of native substrate, enzyme-inhibitor or enzyme-activator binding, as protein molecules may thus provide important information well as interaction with lipids, peptides and other proteins. about the function of the proteins. Hence, the identification Numerous instances of hydrophobic interaction between of a protein fraction of complex tissue and body fluid protein small molecules and protein targets have been observed, extracts with hydrophobic surface sites gains importance as a including the binding of two methylepoxyoctadecane enanti- tool in chemical and structural proteomics.27 It could poten- omers to two pheromone binding protein (PBP) variants;4 the tially aid in the discovery of new enzymes, receptors, and binding of odorant molecules to the hydrophobic interior of regulatory and transporter proteins. the β-barrel of the porcine odorant binding protein (pOBP);5 In this manuscript, we describe the separation of protein an interaction of R1-acid glycoprotein (AAG) with dipy- mixtures by nondenaturing gel electrophoresis and their selec- ridamole;6 a binding interaction of the protein kinase A catalytic tive staining with polarity sensitive fluorophoric and chro- unit and balanol;7 ligand binding to γ-amino butyric acid mophoric dyes that appear to bind primarily by hydrophobic (GABA)/benzodiazepine receptors;8 the δ-opioid receptor in- interactions. teraction with 4-(N,N-diarylamino) piperidines;9 histamine H3 receptor-antagonist interaction.10 Experimental Section Illustrations of hydrophobic protein-protein interactions Acrylamide monomers, TRIS base, ammonium persulfate, include pathological fibrillar aggregations, oligomerizations, TMED, chicken egg albumin, carbonic anhydrase from bovine enzyme-peptide substrate and toxin binding and structural erythrocytes, bovine serum albumin, β-lactoglobulin A from support. Hydrophobic associations implicated in various pa- bovine milk, L-lactic dehydrogenase and phosphorylase b from thologies include neurofibrillary tangles of the hyperphospho- rabbit muscle and soybean trypsin inhibitor, type I-S were rylated τ protein (P-τ) that impair synaptic transmission in obtained from Sigma. The sodium salt of bromophenol blue Alzheimer disease (AD)11 and a strongly hydrophobic aggrega- (BPB) was purchased from Aldrich; 8-anilino-1-naphthalene- tion of the eye lens crystallins, associated with the loss of the sulfonic acid (ANSA) was obtained from Molecular Probes. lens transparency.12 Ligand induced oligomerizations are ex- Premixed 30% (29:1) acrylamide/bisacrylamide solution, native emplified by a homooligomerization of an E. coli transcription TRIS buffer (10X) and BioSafe Coomassie stain were obtained factor, MalT, induced by maltotriose binding to a ligand site from Bio-Rad. Gradient 4-20% native polyacrylamide gels were inside a novel superhelix fold13 and the biotin-induced tet- purchased from Bio-Rad. Nondenaturing one-dimensional ramerization of chicken avidin.14 An interaction of an extra- PAGE was carried out with a Bio-Rad Mini-Protean II cell 470 Journal of Proteome Research • Vol. 2, No. 5, 2003
- 4. Staining Proteins with Hydrophobic Surface Sites research articles following the protocol provided by Bio-Rad. For the 2-D native 100 mL of this solution and shaken at 100 rpm for 30 min on electrophoresis, native 7 cm Immobiline Dry Strip, pH range the Lab Line Orbit Environ-Shaker. The gel was destained by 3 3-10, and the IPG Native Buffer were obtained from Pharmacia quick rinses with d.d. H2O. Better results were achieved if a Biotech. The proteins were separated by isoelectric point on small volume of 1 M phosphate buffer, pH 7.0 was used for the Pharmacia Biotech Multiphor II system. For the second destaining. Selection of a phosphate buffer as a destaining dimension, gradient 4-20% TRIS-glycine gels from Novex and agent was based on the antichaotropic quality of H2PO4- ion, Invitrogen native TRIS-glycine buffer were used. which strengthens hydrophobic interactions of BSA and BPB In one-dimensional electrophoresis experiments, selectivity in water. The dark blue bands were visualized by a dual UV- of the staining dyes was probed on gradient 4-20% native visible light Alpha Innotech Corp. IS 1000 transilluminator. polyacrylamide gels. Aliquots of 20 µL of protein solutions in (2) A 4.0 × 10-5 M ANSA solution was also prepared in 0.100 TRIS-glycine native sample buffer were loaded onto the gels. M TRIS solution, pH 8.0. The gel was immersed in ap- The running buffer was TRIS-glycine at pH 8.3. Electrophoreses proximately 100 mL of this solution and shaken at 100 rpm for were run for 90 min at a constant voltage of 125 V over a 9 cm 20 min. Destaining was achieved by quick rinsing of the gels minigel path. in d.d. H2O. For BPB selectivity experiments, seven gel lanes were loaded For the competitive displacement gel experiments (CDGE), with 20 µL aliquots of different proteins, so that the protein continuous 15% native gels were used. Two lanes on each gel concentration in each lane was 12 µM, equivalent to 0.8 µg/µL were loaded with 20 µL of 15.2 µM BSA (or 20 µg of BSA) and BSA. For ANSA selectivity experiments, gel lanes were loaded the 1-D native electrophoreses were performed as described with 20 µL aliquots of 6 µM protein solutions. A duplicate gel, above. After the runs, the gels were cut in half and treated as containing all proteins at 0.8 µg/µL, was run simultaneously follows: (1) One-half of the first gel was stained with 50 mL of and stained with Coomassie general stain, to visualize all the 23 µM BPB. The other half was immersed in 50 mL of 23 µM proteins. BPB, also containing 1.67 mM ibuprofen. (2) The first half of For sensitivity assays, continuous 15% polyacrylamide gels the second gel was stained in 50 mL of 25 µM ANSA. The were prepared from the premixed 30% acrylamide/bisacryla- second half was immersed in 50 mL of 25 µM ANSA, also mide solution. The gel lanes were loaded with 20 µL aliquots containing 50 µM BPB. All the solutions were prepared in 0.100 of BSA of decreasing concentration. The running buffer was M TRIS buffer, pH 8.0. After being shaken at 100 rpm for 20 TRIS-glycine at pH 8.3. Electrophoreses were run for 90 min min, the gels were destained by 3 quick rinses with d.d. H2O. at a constant voltage of 125 V over a 9 cm minigel path. For ANSA responsive protein bands were visualized by a UV the BPB sensitivity test, BSA concentrations ranged from 50 to transilluminator, as the excitation wavelength is 370 nm. The 0.5 µg per 20 µL well. For ANSA, 20 µL wells contained 66.7 to BSA-ANSA complex emission maximum at about 465 nm 0.067 µg BSA. produces an aqua blue signal. The gels were photographed In the two-dimensional native electrophoresis experiments, using a UV/visible Alpha Innotech Corp. IS 1000 digital imaging the first dimension was carried out on a native 7 cm Immo- system. biline Dry Strip with a pH range 3-10. The strip was rehydrated Duplicate control gels were stained with BioSafe Coomassie overnight with 130 µL of the protein mixture in Pharmacia stain following the protocol provided by Bio-Rad. The protein Biotech IPG native buffer, pH 3-10 using the sample cups on bands were visualized by a visible light transilluminator. the sample bar via anodic application, to ensure maximal BPB and Coomassie stained gels were scanned using Hewlett- adsorption of the large proteins onto the strip. For the selectiv- Packard 5100C Scanjet, to obtain color images. All images were ity assay, the protein mixture contained: 6 µg of ovalbumin, formatted in Paint Shop Pro 4.0. PSP parameters are as 112 µg glycogen phosphorylase b from rabbit muscle, 4 µg follows: for BPB stained gels, hue 175, brightness 25%, contrast carbonic anhydrase from bovine erythrocytes, 2.5 µg soybean 15%. trypsin inhibitor, 8 µg BSA, and 78 µg L-lactate dehydrogenase from rabbit muscle. For the urine protein analysis, about 1 mg Results and Discussion of lyophilized urine proteins was dissolved in 130 µL of the native buffer. A sample of urine proteins for 2-D native The current search for staining dyes was initiated by the electrophoresis was prepared as earlier described.28 Briefly, a observation of an unusual complex between a common elec- 40-mL portion of urine was centrifuged at 10 000 rpm for 10 trophoresis tracer bromophenolblue (BPB) and the cytochrome min and 20 mL of the supernatant was transferred to a Pierce c′ from Chromatium vinosum. The complex formation resulted Slide-A-Lyzer 10 K Dialysis Cassette and dialyzed at 4 °C against in a mobility shift on a native gel and could also be observed two changes of 2 L of double distilled water over 2 d. The spectroscopically as a shift and a change in the intensity of dialyzed urine sample was then lyophilized. After the separation the visible absorption band of BPB in solution.29 It was of proteins according to their isoelectric points, the strip was concluded that the mode of binding in this complex was equilibrated in the Novex TRIS-glycine native sample buffer hydrophobic, based on a similar change in absorption maxi- (2×) containing 0.01% SDS for 15 min and dried. SDS was mum of BPB in the presence of nonionic TRITON-100 micelles. added to facilitate a more complete transfer of the high MW Additionally, the binding of BPB to the cytochrome was not proteins from the strip onto the second dimension gel. Second affected by the ionic strength of the solution, excluding the dimension electrophoreses were run on a pre-cast polyacry- possibility that the binding occurs through an ionic interac- lamide mini-gel system, under a constant voltage of 125 V for tion.29 Similarly, the visible spectra of BPB solutions exhibit a 90 min. peak shift and absorption coefficient change upon addition of Visualization of the spots containing proteins with hydro- propanol.30 It has also been demonstrated30 that BPB binds to phobic surfaces was achieved by staining with the following BSA, a protein that has been shown to have hydrophobic BPB and ANSA solutions: binding sites.31 The absorption maximum of the BSA‚BPB (1) An 8.67 × 10-5 M BPB solution was prepared in 0.100 M complex exhibits a red shift with respect to that of unbound TRIS buffer, pH 8.0. The gel was immersed in approximately BPB, characteristic of a change to a more nonpolar environ- Journal of Proteome Research • Vol. 2, No. 5, 2003 471
- 5. research articles Bertsch and Kassner ment.30 Thus, a stain on a gel immersed in a BPB solution indicates a hydrophobic type interaction between a protein from the mixture and the dye. A literature search26,32-34 for other suitable dyes yielded 1-anilinonaphthalene-8-sulfonic acid (1,8-ANSA), a lipophilic fluorophore. ANSA is virtually nonfluorescent in an aqueous medium, but becomes strongly fluorescent as the polarity of a medium decreases. ANSA has also been shown to bind to hydrophobic sites on HSA.35 The fluorescent response on an ANSA stained gel may then be attributed to a hydrophobic interaction between a protein from the mixture and the dye. To demonstrate the feasibility of these dyes as staining agents, selectivity and sensitivity tests were performed. A selectivity test was designed to show that under the staining conditions, only a select subset of components of a complex protein mixture would yield a detectable signal, indicative of binding of the staining dye to the molecules of the subset. The images of one-dimensional gel selectivity assays stained with BPB and ANSA, including the Coomassie stained control gel, are presented in Figures 1A-C. As shown in Figure 1A, BSA monomer and dimer, phospho- rylase b and L-lactic dehydrogenase produced a visible spot upon staining with BPB. As shown in Figure 1B, the same proteins produced a fluorescent signal upon staining with ANSA, indicating the binding of both polarity sensitive dyes to these proteins. The results shown in Figure 1 suggest the presence of solvent accessible hydrophobic sites of interaction of these proteins with the polarity sensitive dyes. The binding of ANSA to BSA is consistent with the binding of ANSA to HSA, which has been shown to have at least two binding sites to which a variety of hydrophobic drug molecules may bind: site I in subdomain Figure 1. Selectivity assays. Lanes 1 through 7 contain 20 µL IIA and site II in subdomain IIIA. Crystal structures of HSA aliquots of chicken egg albumin (lane 1; 45 000 Da), phospho- complexes with warfarin,36 an anticoagulant, and the local rylase b from rabbit muscle (lane 2; 97 600 Da) carbonic anhy- anesthetics halothane37 and propofol37 have been reported. drase from bovine erythrocytes (lane 3; 31 000 Da), soybean Warfarin binds to site I, which is predominantly hydrophobic. trypsin inhibitor S (lane 4; 21 500 Da), bovine serum albumin or The site is formed by the packing of all six helices of subdomain BSA (lane 5; 66 000 Da), β-lactoglobulin A from bovine milk (lane IIA and consists of two chambers in to which a warfarin 6, 17 600 Da), and L-lactic dehydrogenase from rabbit muscle molecule fits quite snugly.36 The benzyl moiety interacts with (lane 7, 209 000 Da). Lane 8 contained 20 µL of the mixture of all the above proteins, at a concentration equivalent to those of the residues F211, W214, L219, L238, R218, and H242, whereas the separate protein solutions. (A) One-dimensional native gel coumarin moiety binds to R222, F223, L238, V241, R257, I260, stained with 8.67 × 10-5 M BPB. Concentrations of proteins in I264, I290, and A291. We propose that ANSA, a molecule with µg/µL are: lane 1, 0.54; lane 2, 1.17; lane 3, 0.37; lane 4, 0.26; a similar topology (a naphthalene bicyclic structure and an lane 5, 0.79; lane 6, 0.21 and lane 7, 2.50, which are equivalent anilino moiety with a one atom linker) could also bind to site to 12 µM; (B) One-dimensional native gel stained with 4.0 × 10-5 I. Furthermore, it was recently observed that ANSA com- M ANSA. Concentrations of proteins in µg/µL are: lane 1, 0.27; petitively displaces BPB from BSA, which was monitored by lane 2, 0.58; lane 3, 0.18; lane 4, 0.13; lane 5, 0.40; lane 6, 0.10 absorption difference spectroscopy.30 Upon titration of the BSA‚ and lane 7, 1.25, which are equivalent to 6 µM; (C) Control gel BPB complex by increasing concentrations of ANSA, the stained with BioSafe Coomassie. Concentrations of all proteins absorption difference maximum corresponding to the BSA‚BPB are 0.8 µg/µL. complex decreases, suggesting competitive displacement of bovine carbonic anhydrase, trypsin inhibitor, and β-lactoglo- BPB by ANSA. The competitive binding of BPB and ANSA bulin on the gels do not stain in the presence of BPB or ANSA, suggests that BPB and ANSA interaction occur at a common but are stained in the presence of a nonspecific dye, e.g., binding site on BSA. Thus, it can be concluded that ANSA and Coomassie. It was not unexpected that the protein from chicken BPB bind to a common hydrophobic locus on BSA, analogous egg did not stain with BPB and ANSA. Although earlier named to the warfarin binding site on HSA. albumin, it is actually a serpin (serine protease inhibitor) with The staining of phosphorylase b was not unexpected, a completely different structure and function than that of BSA because it has earlier been shown to bind several amphiphilic or HSA. The results show that BPB and ANSA are selective with ligands, including pyridoxal-5′-phosphate and isopropoxy- respect to the proteins to which they bind and thus detect those carbonyl-6-methyl-pyridinium (crystal structure with the which have solvent accessible hydrophobic sites. Brookhaven PDB code 3amv),38 heptulopyranosonamide (1fs4),39 A sensitivity test was carried out to show a detectability limit oxoethyl amide (1c50),40 and caffeine (1c8l),41 Likewise, LDH of each staining technique. The sensitivity assay results are has earlier been reported to bind BPB, although the mode of presented in Figure 2A-B. The detection limit for our system binding was not proposed,42 By contrast, chicken egg albumin, is determined to be at 0.50 µg protein per 20 µL for a 8.67 × 472 Journal of Proteome Research • Vol. 2, No. 5, 2003
- 6. Staining Proteins with Hydrophobic Surface Sites research articles Table 1. Molecular Weights and PI Values for Proteins Used in Gel Assays 1-D gel 2-D gel ANSA/BPB ANSA/BPB protein MW (kDa) pI stain visiblea stain visiblea BSA 66 5.4, 5.5, 5.6 + + carbonic 31 5.9, 6.0 - - anhydrase LDH 209 8.5 + + ovalbumin 45 4.5 - - phosphorylase b 97.6 6.7 + + trypsin inhibitor 21.5 4.7 - - a A “+” sign indicates that a protein stains with a polarity sensitive probe, whereas a “-” sign indicates a negative result. The l-lactic dehydrogenase and phosphorylase b dissociate into subunits in the presence of 0.01% SDS during the separation in the second dimension on the 2-D gel. 10-5 M BPB staining solution, which corresponds to a 380 nM BSA solution. The detection limit is at 67 ng BSA per 20 µL sample for a 4.0 × 10-5 M ANSA solution, corresponding to a BSA concentration of 51 nM. If gels with narrower lanes were used, then the detection limit would presumably be even lower. Advanced digital imaging may further enhance the detectability Figure 2. Sensitivity assays. (A) One-dimensional native gel stained with 8.67 × 10-5 M BPB. Lanes 1 through 7 contain 50, of a protein spot stained with these polarity sensitive reagents. 20, 10, 5.0, 2.0, 1.0, and 0.50 µg BSA per 20 µL, corresponding to The sensitivity of both the BPB and ANSA staining solutions BSA concentrations from 38 to 0.38 µM; (B) One-dimensional is sufficient to enable selective staining of more abundant native gel stained with 4.0 × 10-5 M ANSA. Lanes 1 through 7 proteins with hydrophobic surfaces in whole cell lysates, body contain 66.7, 33.3, 6.7, 3.3, 0.67, 0.33, and 0.067 µg BSA per 20 fluids, and other complex protein mixtures. Even with the µL, corresponding to BSA concentrations from 51 µM to 51 nM. current gel bandwidths and without any digital image enhance- ment, the sensitivity of the ANSA stain (67 ng per band) is SDS-PAGE, e.g., MALDI-TOF mass spectrometry and sequenc- comparable to that of the Coomassie general protein stain ing, may establish a library of proteins with hydrophobic which averages at 40 ng, or the BioSafe Coomassie by Bio-Rad, surface sites, containing information on their molecular weight whose detection limit is at 28 ng. The detection of hydrophobic and sequence to complement the binding assays and structure- surface sites on proteins of lower abundance is also possible function correlation, which are unique to nondenaturing upon prefractionation and/or enrichment of a sample. approaches. The photographs of two-dimensional gels stained with ANSA An application of the above approach to the characterization and Coomassie are presented in Figure 3A,B. The molecular of proteins found in urine is shown in Figure 4. A nondena- weights and pI values of the proteins used in this study are turing 2-D gel of urine proteins stained with ANSA is shown in listed in Table 1. Native 2-D electrophoretic patterns reveal a Figure 4.A. Stained bands indicate the presence of proteins with greater complexity of the mixture. The isozyme triplets of the a solvent accessible hydrophobic site. Figure 4B shows a BSA monomeric and dimeric structures stain with both ANSA nondenaturing 2D gel of the same sample of urine proteins and Coomassie. Further, at least three bands of LDH and at stained with Coomassie, the general stain for all proteins. It is least two phosphorylase b traces stained with ANSA, possibly apparent that only a fraction of the total proteins stain with corresponding to the light and heavy subunit of dimeric LDH ANSA. and the phosphorylase b monomeric and dimeric structures, The identification of soluble globular proteins with surface respectively. The bovine carbonic anhydrase, chicken ovalbu- hydrophobic sites by staining with polarity sensitive dyes may min and soybean trypsin inhibitor stain only with Coomassie also be used to screen pharmacologically active molecules as stain. drug candidates. A drug candidate will be effective in vivo only The identification of a subproteomic library that exhibits if it is able to achieve and maintain therapeutic concentration particular structural features related to protein functions could at the site of action.43 Pharmacological characteristics, e.g., facilitate functional and structural characterization of new solubility, partition coefficient, permeability and protein bind- proteins. As noted in the Introduction, a number of proteins ing contribute to its in vivo disposition. Many drug candidates have been identified that have hydrophobic surface sites are lipophilic, such that their efficacy may depend on their involved in binding of small molecules, lipids, peptides and transport by serum proteins, e.g., HSA, to their target, e.g., other proteins. Native 2-D electrophoresis of complex protein another protein. Additionally, their approval may be limited samples, followed by staining with a polarity sensitive dye by their interactions with other proteins that lead to undesirable facilitates the discovery of soluble globular proteins whose side effects. The advances of combinatorial chemistry and drug hydrophobic surface sites may be implicated in their transport, discovery have increased the interest in rapid and cost-effective catalytic or regulatory functions. The staining is reversible, such solutions for prescreening of novel pharmaceuticals and early that the dyes may be dialyzed after excision, enabling subse- elimination of possible problem candidates. quent structure and function assays of selected spots containing Complex protein mixtures, e.g., serum, urine, or cerebrospi- proteins of interest, including proteins unique to certain disease nal fluid samples or soluble cell extracts may be subjected to states, or development stages. Furthermore, implementation the native 2-D electrophoresis protocol, followed by Coomassie, of modern bioanalytical techniques already in use with 2-D silver, or Ruby staining for total protein and the novel proce- Journal of Proteome Research • Vol. 2, No. 5, 2003 473
- 7. research articles Bertsch and Kassner Figure 5. Competitive displacement gel experiment. Two lanes were loaded with equivalent amounts of BSA, e.g., 20 µL of 1.0 µg/µL solution (15 µM). After native electrophoresis, the gel was cut in half and treated as described in the Material and methods section. The left half was stained with 23 µM BPB. The right half of the gel was stained with a solution containing 23 µM BPB and 1.67 mM ibuprofen. the drug would bind the dye. Ideally, only the expected binding of the drug to a target carrier protein would be observed. Loss of additional signals upon binding of the drug with respect to the selectively stained gel in the absence of the drug may indicate possible side effects due to nonspecific binding of the drug. Figure 3. Two-dimensional selectivity assays. The protein mix- A simple demonstration of the competitive displacement gel ture contained the following proteins: chicken egg albumin (6 µg), phosphorylase b (112 µg), bovine carbonic anhydrase (4 µg), experiment (CDGE) was carried out using the drug ibuprofen, soybean trypsin inhibitor (2.5 µg), BSA (8 µg) and lactic dehy- for which earlier studies have shown affinity toward HSA and drogenase (78 µg) in 130 µL. (A) Two-dimensional native gel BSA binding.44,45 Figure 5 shows that BSA exposed to the stained with 4.0 × 10-5 M ANSA; (B) Control gel stained with polarity sensitive chromophore BPB only, produces a normal BioSafe Coomassie. signal due to the interaction of the dye with a hydrophobic region on the surface of the protein. When the drug ibuprofen is introduced into the staining solution, the signal is greatly diminished. The loss of the stained spot in the presence of ibuprofen can be explained by the competitive displacement of the dye by the drug. The displacement efficacy at a certain pH, ionic strength, and temperature is dependent upon two factors: (1) the strength of binding, characterized by the free energy of binding or the binding constant; and (2) the drug vs dye concentration ratio. The higher the binding constant of the drug, the lower the concentration limit at which the displacement is detectable as a loss of the signal. The particular concentration ratio for this assay was selected based on the binding constant deter- mined previously by a spectrophotometric competitive titra- tion.30 For a novel protein-drug interaction, for which an initial assessment of protein-drug binding affinity is not available, a millimolar to micromolar drug concentration range seems most appropriate. Alternatively, the drug concentration may be chosen based on blood concentrations during treatment with analogous agents. A similar experiment served as qualitative evidence that BPB and ANSA bind to the same site on the protein surface and Figure 4. Two-dimensional native gel of urine proteins. The first that the site is indeed hydrophobic. Again, two gel lanes dimension Immobiline Dry Strip was rehydrated with 130 µL of contained the same protein amount. BSA exposed to the native buffer, pH 3-10, containing about 1 mg of lyophilized urine environment polarity sensitive fluorophore ANSA only, pro- proteins. (A) Two-dimensional native gel stained with 4.0 × 10-5 duces a fluorescent signal due to the interaction of the dye with M ANSA; (B) Control gel stained with BioSafe Coomassie. a hydrophobic region on the surface of the protein. Figure 6 shows that when a competing dye, BPB, is intro- dure for staining of soluble globular proteins with hydrophobic duced into the ANSA staining solution, the fluorescent signal surface sites. On the Coomassie stained gels, all of the detect- is reduced to less than one-half of the initial signal. This is able proteins will be visualized. The selectively stained samples explained by the competitive displacement of ANSA by BPB should have distinct patterns, since a selective marker, e.g., BPB from a common binding locus on the protein. On the basis of or ANSA, would stain only the proteins with hydrophobic the concentration ratio of BPB to ANSA and the extent of surface sites. On a gel treated with a mixture of a polarity displacement, the binding constants of the two dyes are similar. sensitive dye and a lipophilic (hydrophobic) drug molecule, In summary, a rapid and sensitive method for selective only the proteins whose hydrophobic surface sites do not bind staining of soluble globular proteins with hydrophobic surface 474 Journal of Proteome Research • Vol. 2, No. 5, 2003
- 8. Staining Proteins with Hydrophobic Surface Sites research articles (12) Liang, J. N.; Li, X. Y. Exp. Eye Res. 1991, 53 (1), 61-66. (13) Steegborn, C.; Danot, O.; Huber, R.; Clausen, T. Structure (Camb.) 2001, 9 (11), 1051-1060. (14) Laitinen, O. H.; Marttila, A. T.; Airenne, K. J.; Kulik, T.; Livnah, O.; Bayer, E. A.; Wilchek, M.; Kulomaa, M. S. J. Biol. Chem. 2001, 276 (11), 8219-8224. (15) Makinen, P. L.; Clewell, D. B.; An, F.; Makinen, K. K. J. Biol. Chem. 1989, 264 (6), 3325-3334. (16) Moorthy, A. K.; Gopal, B.; Satish, P. R.; Bhattacharya, S.; Bhatta- charya, A.; Moorthy, M. R.; Surolia, A. FEBS Lett. 1999, 461 (1- 2), 19-24. (17) Eriksson, K. O.; Naidu, A. S.; Kilar, F.; Wadstrom, T.; Hjerten, S. Figure 6. Competitive displacement gel experiment. Two lanes APMIS 1989, 97 (12), 1081-1087. were loaded with equivalent amounts of BSA, e.g., 20 µL of 1.0 (18) Klebe, R. J.; Bentley, K. L.; Schoen, R. C. J. Cell Physiol. 1981, 109 µg/µL solution (15 µM). After native electrophoresis, the gel was (3), 481-488. (19) Ringli, C.; Hauf, G.; Keller, B. Plant Physiol. 2001, 125 (2), 673- cut in half and treated as described in the Material and methods 682. section. The left half was stained with 25 µM ANSA. The right (20) Goldstein, D. J.; Andresson, T.; Sparkowski, J. J.; Schlegel, R. EMBO half of the gel was stained with a solution containing 25 µM J. 1992, 11 (13), 4851-4859. ANSA and 50 µM BPB. (21) Goldstein, D. J.; Li, W.; Wang, L. M.; Heidaran, M. A.; Aaronson, S.; Shinn, R.; Schlegel, R.; Pierce, J. H. J. Virol. 1994, 68 (7), 4432- sites has been developed. A fluorescent and visible spectro- 4441. photometric approach is offered through ANSA and BPB (22) Seelig, A.; Seelig J. Hoppe Seylers Z. Physiol. Chem. 1978, 359 (12), staining, respectively, for additional versatility of the technique. 1747-1756. (23) Volwerk, J. J.; Mrsny, R. J.; Patapoff, T. W.; Jost, P. C.; Griffith, O. Staining with both reagents is reversible, enabling excision of H. Biochemistry 1987, 26 (2), 466-475. stained spots, removal of dyes by dialysis or chromatography (24) Kolusheva, S.; Boyer, L.; Jelinek, R. Nat. Biotechnol. 2000, 18 (2), and subsequent structure and function assays. 225-227. (25) Zhou, F.; Schulten, K. Proteins 1996, 25 (1), 12-27. As elaborated earlier, the CDGE offers the possibility of (26) Sackett, D.; Wolff, J. Anal. Biochem. 1987, 167, 228-234. evaluating the specificity of drug-target protein interaction for (27) Edwards, A. M.; Arrowsmith, C. H.; Des Pallieres, B. Modern Drug many lipophilic drugs. For a highly specific drug, the staining Discovery 2000, 9, 35-44. pattern in the presence of the drug will lack only the band (28) Rasmussen, H. H.; Orntoft, T. F.; Wolf, H.; Celis, J. E. J. Urol. 1996, 155 (6), 2113-2119. corresponding to the target protein. An analogous assay can (29) Mayburd, A. L.; Tan, Y.; Kassner, R. J. Arch. Biochem. Biophys. aid environmental investigations of protein-toxin interactions. 2000, 378 (1), 40-44. The technique also enables qualitative prescreening of pro- (30) Bertsch, M.; Mayburd, A. L.; Kassner, R. J. Anal. Biochem. 2002, 313, 187-195. teomic samples against drugs and drug candidates, whose (31) Kragh-Hansen, U. Dan. Med. Bull. 1990, 37, 57-84. pharmacological potency is indicated by cell-based assays, but (32) Fitzgerald, R. J.; Swaisgood, H. E. Arch. Biochem. Biophys. 1989, their protein target is unknown. 268 (1), 239-248. (33) Musci, G.; Berliner, L. J. Biochemistry 1985, 24, 3852-3856. (34) Slavik, J. Biochim. Biophys. Acta 1982, 694, 1-25. References (35) Weber, G.; Young, L. B. J. Biol. Chem. 1964, 239, 1415-1423. (36) Petitpas, I.; Bhattacharya, A. A.; Twine, S.; East, M.; Curry, S. J. (1) Jung, E.; Veuthey, A. L.; Gasteiger, E.; Bairoch, A. Proteomics 2001, Biol. Chem. 2001, 276 (25), 22 804-22 809. 1 (2), 262-268. (37) Bhattacharya, A. A.; Curry, S.; Franks, N. P. J. Biol. Chem. 2000, (2) O′Farrell, P. J. Biol. Chem. 1975, 250 (10), 4007-4021. 275 (49), 38 731-38 738. (3) Willis, R. Modern Drug Discovery 2001, 10, 63-66. (38) Oikonomakos, N. G.; Tsitsanou, K. E.; Zographos, S. E.; Skamnaki, (4) Plettner, E.; Lazar, J.; Prestwich, E. G.; Prestwich, G. D. Biochem- V. T.; Goldmann, S.; Bischoff, H. Protein Sci. 1999, 8 (10), 1930- istry 2000, 39 (30), 8953-8962. 1945. (5) Vincent, F.; Spinelli, S.; Ramoni, R.; Grolli, S.; Pelosi, P.; Cambillau, (39) Oikonomakos, N. G.; Kontou, M.; Zographos, S. E.; Tsitoura, H. C.; Tegoni, M. J. Mol. Biol. 2000, 300 (1), 127-139. S.; Johnson, L. N.; Watson, K. A.; Mitchell, E. P.; Fleet, G. W.; Son, (6) El-Gamel, S.; Wollert, U.; Muller, W. E. J. Pharm. Pharmacol. 1982, J. C.; Bichard, C. J.; et al. Eur. J. Drug Metab. Pharmacokinet. 1994, 34 (3), 152-157. 19 (3), 185-192. (7) Wong, C. F.; Hunenberger, P. H.; Akamine, P.; Narayana, N.; (40) Oikonomakos, N. G.; Skamnaki, V. T.; Tsitsanou, K. E.; Gavalas, Diller, T.; McCammon, J. A.; Taylor, S.; Xuong, N. H. J. Med. Chem. N. G.; Johnson, L. N. Structure Fold Des. 2000, 8 (6), 575-584. 2001, 44 (10), 1530-1539. (41) Tsitsanou, K. E.; Skamnaki, V. T.; Oikonomakos, N. G. Arch. (8) Harris, D. L.; DeLorey, T. M.; He, X.; Cook, J. M.; Loew, G. H. Biochem. Biophys. 2000, 384 (2), 245-254. Eur. J. Pharmacol. 2000, 401 (3), 271-287. (42) Towell, J. F., III; Woody, R. W. Biochemistry 1980, 19, 4231-4237. (9) Podlogar, B. L.; Poda, G. I.; Demetar, D. A.; Zhang, S. P.; Carson, (43) Parikh, H. H.; McElwain, K.; Balasubramanian, V.; Leung, W.; J. R.; Neilson, L. A.; Reitz, A. B.; Ferguson, D. M. Drug Des. Discov. Wong, D.; Morris, M. E.; Ramanathan, M. Pharm. Res. 2000, 17 2000, 17 (1), 34-50. (5), 632-637. (10) De Esch, I. J.; Mills, J. E.; Perkins, T. D.; Romeo, G.; Hoffmann, (44) Whitlam, J. B.; Crooks, M. J.; Brown, K. F.; Pedersen, P. V. Biochem. M.; Wieland, K.; Leurs, R.; Menge, W. M.; Nederkoorn, P. H.; Pharmacol. 1979, 28, 675-678. Dean, P. M.; Timmerman, H. J. Med. Chem. 2001, 44 (11), 1666- (45) Hage, D. S.; Noctor, T. A.; Wainer, I. W. J. Chromatogr. Anal. 1995, 1674. 93 (1), 23-32. (11) Ruben, G. C.; Ciardelli, T. L.; Grundke-Iqbal, I.; Iqbal, K. Synapse 1997, 27 (3), 208-229. PR025579+ Journal of Proteome Research • Vol. 2, No. 5, 2003 475