Specific Covalent Immobilization of Proteins through Dityrosine Cross

Betsy J. Endrizzi, Gang Huang, Patrick F. Kiser, and Russell J. Stewart*. Department of Bioengineering, UniVersity of Utah, 20 South 2030 East, Room 5...
4 downloads 0 Views 225KB Size
Download PDF
Langmuir 2006, 22, 11305-11310

11305

Specific Covalent Immobilization of Proteins through Dityrosine Cross-Links Betsy J. Endrizzi, Gang Huang, Patrick F. Kiser, and Russell J. Stewart* Department of Bioengineering, UniVersity of Utah, 20 South 2030 East, Room 506, Salt Lake City, Utah 84112 ReceiVed June 23, 2006. In Final Form: September 19, 2006 Dityrosine cross-links are widely observed in nature in structural proteins such as elastin and silk. Natural oxidative cross-linking between tyrosine residues is catalyzed by a diverse group of metalloenzymes. Dityrosine formation is also catalyzed in vitro by metal-peptide complexes such as Gly-Gly-His-Ni(II). On the basis of these observations, a system was developed to specifically and covalently surface immobilize proteins through dityrosine cross-links. Methacrylate monomers of the catalytic peptide Gly-Gly-His-Tyr-OH (GGHY) and the Ni(II)-chelating group nitrilotriacetic acid (NTA) were copolymerized with acrylamide into microbeads. Green fluorescent protein (GFP), as a model protein, was genetically tagged with a tyrosine-modified His6 peptide on its carboxy terminus. GFP-YGH6, specifically associated with the NTA-Ni(II) groups, was covalently coupled to the bead surface through dityrosine bond formation catalyzed by the colocalized GGHY-Ni(II) complex. After extensive washing with EDTA to disrupt metal coordination bonds, we observed that up to 75% of the initially bound GFP-YGH6 remained covalently bound to the bead while retaining its structure and activity. Dityrosine cross-linking was confirmed by quenching the reaction with free tyrosine. The method may find particular utility in the construction and optimization of protein microarrays.

Introduction Surface immobilization of proteins is essential in several rapidly expanding areas of biomedical science and technology. These areas include proteomics, protein arrays for clinical diagnostics, and surface modification of biomedical devices, to name a few broad categories.1-4 An ideal immobilization technology would target a specific protein site to allow immobilization of the protein with a defined orientation, would be supported on a proteinresistant substrate to retain the structure and activity of the immobilized protein while simultaneously preventing nonspecific binding of nontarget proteins, and would be based on robust covalent cross-links to prevent leaching of the protein from the surface upon dilution or during extensive washing. Early methods of protein immobilization based on nonspecific adsorption or traditional conjugation chemistries fall short on several of these specifications. The major drawback of nonspecific adsorption is that many proteins are immobilized with a random orientation to the surface, possibly blocking the ligand-binding site or active site. Nonspecific adsorption also frequently results in denaturation and inactivation of the adsorbed proteins. Approaches involving covalent coupling of proteins to derivatized surfaces using conventional protein modification chemistries have similar drawbacks.4 Many of these approaches rely on reagents that, if the solution conditions are right, react somewhat specifically with a particular class of nucleophilic functional group, such as the thiol of cysteines or the primary amine of lysines. Specificity is limited because all accessible functional groups of a particular class can be modified nonselectively. Nonspecific reactions lead to random orientations and possibly denaturation or inactivation * To whom correspondence [email protected].

should

be

addressed.

E-mail:

(1) Cretich, M.; Damin, F.; Pirri, G.; Chiari, M. Biomol. Eng. 2006, 23, 7788. (2) Aoki, J.; Serruys, P. W.; van Beusekom, H.; Ong, A. T. L.; McFadden, E. P.; Sianos, G.; van der Giessen, W. J.; Regar, E.; de Feyter, P. J.; Davis, H. R.; Rowland, S.; Kutryk, M. J. B. J. Am. Coll. Cardiol. 2005, 45, 1574-1579. (3) Haab, B. B.; Dunham, M.; Brown, P. O. Genome Biol. 2001, 2, 1-13. (4) MacBeath, G.; Schreiber, S. L. Science 2000, 289, 1760-1762.

during surface immobilization if the functional group is in or near the active site of the target protein. Methods based on genetically fusing target proteins with ligandbinding proteins, such as glutathione S-transferase, or the addition of metal-binding peptide tags, such as hexahistidine, allow proteins to be immobilized through a specific site.5,6 Because the bonds are reversible, one limitation of this method is that the target protein can be eluted from the surface by extensive dilution or washing or changes in solution conditions, such as ionic strength or pH. Recent innovations overcome this limitation by genetically fusing the target protein with enzymes that form covalent adducts with surface-immobilized suicide inhibitors, referred to as activesite-directed capture ligands (ASDCLs).7,8 In the first example, Hodneland et al.8 fused target proteins with cutinase, a monomeric MW ) 22000 serine esterase that forms covalent adducts with a phosphonate suicide inhibitor. Cutinase fusion proteins were specifically and covalently surface immobilized by tethering the phosphonate capture ligand to a surface. A similar approach was reported by Kindermann et al.7 The human DNA repair protein O6-alkylguanine-DNA alkyltransferase (hAGT) forms covalent adducts at its active site with O6-benzylguanine. Proteins fused to hAGT were specifically and covalently coupled to surfaceimmobilized O6-benzylguanine. Like cutinase, the MW of hAGT is around 21000. The active-site-directed capture ligand methods combine the specificity of recombinant fusion proteins with the stability of covalent immobilization. Phenolic oxidative cross-linking has been widely observed in nature, including protein cross-linking through the phenolic side chains of tyrosine residues. Dityrosine cross-linking occurs in structural proteins such as elastin, silk, plant cell wall extensin, and hardened fertilization membranes of insect and sea urchin (5) Jung, J.-W.; Jung, S.-H.; Kim, H.-S.; Yuk, J. S.; Park, J.-B.; Kim, Y.-M.; Han, J.-A.; Kim, P.-H.; Ha, K.-S. Proteomics 2006, 6, 1110-1120. (6) Chaga, G. S. J. Biochem. Biophys. Methods 2001, 49, 313-334. (7) Kindermann, M.; George, N.; Johnsson, N.; Johnsson, K. J. Am. Chem. Soc. 2003, 125, 7810-7811. (8) Hodneland, C. D.; Lee, Y.-S.; Min, D.-H.; Mrksich, M. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5048-5052.

10.1021/la0618216 CCC: $33.50 © 2006 American Chemical Society Published on Web 11/18/2006

11306 Langmuir, Vol. 22, No. 26, 2006

Endrizzi et al.

Experimental Section

Figure 1. (A) Two-step covalent protein immobilization. GFPYGH6 associates specifically with surface-immobilized NTA-Ni(II). Addition of an oxidant initiates dityrosine cross-linking by the colocalized catalytic GGHYK-Ni(II) complex. Metal ions are removed by washing with the chelating agent EDTA. Covalently immobilized GFP-YGH6 remains bound. (B) Proposed mechanism of oxidative cross-linking of tyrosines by a catalytic GGHYKNi(II) peptide complex. A colocalized GGH-Ni(II) complex, 1, upon addition of an oxidant, catalyzes the formation of tyrosine radicals 4 that couple into dityrosine cross-links 5.16-18

eggs.9-12 Natural oxidative cross-linking between tyrosine residues is catalyzed by a structurally and mechanistically diverse group of metalloenzymes that includes tyrosinases, peroxidases, and laccases that use dioxygen or hydrogen peroxide as terminal electron acceptors.13-15 Dityrosine coupling is also catalyzed by metal-binding peptides in the presence of oxidants.16-20 For example, Brown and co-workers17 demonstrated that Ni(II) complexed by the tripeptide NH2-Gly-Gly-His-COOH (GGH) catalyzed protein cross-linking through tyrosine residues. Here, we report a method for specific and covalent surface immobilization of proteins that mimics in some respects the natural metal-catalyzed dityrosine cross-linking. Target proteins are first reversibly surface associated through the well-known interaction of His6 peptide tags and NTA-Ni(II) complexes.6 Upon addition of an oxidant, the colocalized GGHYK-Ni(II) complex catalyzes the formation of cross-links between tyrosine residues strategically placed adjacent to the His6 peptide tag and the catalytic peptide complex (Figure 1). (9) LaBella, F. S.; Keeley, F. W.; Vivian, S.; Thornhill, D. P. Biochem. Biophys. Res. Comm. 1967, 26, 748-753. (10) Raven, D. J.; Earland, C.; Little, M. Biochim. Biophys. Acta 1971, 251, 96-99. (11) Brady, J. D.; Sadler, I. H.; Fry, S. C. Biochem. J. 1996, 315, 323-327. (12) Foerder, C. A.; Shapiro, B. M. Proc. Natl. Acad. Sci. U.S.A. 1977, 74, 4214-4218. (13) Michon, T.; Chenu, M.; Kellershon, N.; Desmadril, M.; Gue´guen, J. Biochemistry 1997, 36, 8504-8513. (14) Sanchez-Ferrer, A.; Rodriguez-Lopez, J. N.; Garcia-Canovas, F.; GarciaCarmona, F. Biochim. Biophys. Acta 1995, 1247, 1-11. (15) Thurston, C. F. Microbiology 1994, 140, 19-26. (16) Brown, K. C.; Yang, S.-H.; Kodadek, T. Biochemistry 1995, 34, 47334739. (17) Brown, K. C.; Yu, Z.; Burlingame, A. L.; Craik, C. S. Biochemistry 1998, 37, 4397-4406. (18) Fancy, D. A.; Kodadek, T. Biochem. Biophys. Res. Commun. 1998, 247, 420-426. (19) Fancy, D. A.; Kodadek, T. Tetrahedron 1997, 53, 11953-11960. (20) Gill, G.; Richter-Rusli, A. A.; Ghosh, M.; Burrows, C. J.; Rokita, S. E. Chem. Res. Toxicol. 1997, 10, 302-309.

Materials. H-Lys(Z)-OH and Fmoc-Lys(Alloc)-OH were purchased from Bachem (Torrance, CA). Bromoacetic acid, methacryloyl chloride, and TEMED were purchased from Sigma-Aldrich (St. Louis, MO). Wang resin, Fmoc-Tyr(tBu)-OH, Fmoc-His(Trt)-OH, FmocFly-OH, and Boc-Gly-OH were all purchased from Novabiochem (San Diego, CA). HOBT was purchased from Oakwood Products (West Columbia, SC). Acrylamide was purchased from Polysciences (Warrington, PA). N′,N′-methylenebisacrylamide was purchased from ICN Laboratories, Inc. (Aurora, OH). Ammonium persulfate was purchased from Bio-Rad Laboratories (Hercules, CA). The ultrafiltration membrane for the protein was purchased from Millipore (Billarica, MA). All enzymes and cell lines used for plasmid construction and protein expression were purchased from Invitrogen (Carlsbad, CA). Chemicals not listed were purchased from Acros (Geel, Belgium). Synthesis of MABNTA. A polymerizable nitrilotriacetic acid (NTA) group, (2-methacrylamidobutyl)nitrilotriacetic acid (MABNTA), was synthesized as previously described.21 Bromoacetic acid (8.34 g, 60 mmol) and N′-(benzyloxycarbonyl)-L-lysine (H-Lys(Z)-OH) (8.4 g, 30 mmol) were used to prepare NR,NR-bis(carboxymethyl)-N-(benzyloxycarbonyl)-L-lysine (Z-NTA). The Z protecting group was removed by hydrogenating 9 g of Z-NTA in 150 mL of MeOH at 40 psi overnight with 10% Pd/C catalyst to obtain NR,NR-bis(carboxymethyl)-L-lysine (ABNTA). Purified ABNTA (2.62 g, 10 mmol) was dissolved in 15 mL of 2 N NaOH, and a small amount of tert-octylpyrocatechine was added. This solution was cooled to between -5 and 0 °C, and while the solution was stirred, distilled methacryloyl chloride (1.3 mL, 13.4 mmol) was added dropwise. After a slight delay, 10 mL of 1 N NaOH was added, and the solution was stirred at room temperature for an additional 2 h. The organic layer was removed, and the aqueous layer was washed twice with 15 mL of diethyl ether. The pH of the aqueous layer was decreased to pH 2 using 6 N HCl. MABNTA was extracted with ethyl acetate (15 mL × 15). This was collected and dried overnight with anhydrous magnesium sulfate. The drying agent was filtered off, and solvent was removed by rotary evaporation. The product was redissolved in water and lyophilized: yield 46%; 1H NMR (DMSO-d ) δ 1.3-1.6 (m, 6H), 1.8 (s, 1H), 3.0 (m, 2H), 6 3.3 (t, 3H), 3.5 (s, 4H), 5.3 and 5.6 (s, 2H), 7.6 (t, 1H), 12.4 (s, 3H). Synthesis of GGHYK Monomer. The GGHYK monomer was synthesized on Wang resin using standard Fmoc chemistry as shown in Figure 2. Wang resin (0.5 g) was weighed into a 6 mL filter tube and preswelled in DMF. Fmoc-Lys(Alloc)-OH (631 mg, 13.9 mmol) was weighed into a vial and dissolved in 3 mL of dry DMF. Pyridine (110 mg, 13.9 mmol) and DIC (176 mg, 13.9 mmol) were added, and then the resulting solution was mixed with the preswelled resin. The coupling reaction continued until a Kaiser test confirmed that Lys had been coupled with 1. Unreacted groups on the resin were capped using acetic anhydride, DMAP, and triethylamine. After two 30 min reactions with acetic anhydride, the resin was washed with DMF. The Fmoc group on the Lys was removed by mixing twice for 20 min with 20% piperidine in DMF. The resin was washed three times with DMF. Fmoc-Tyr(tBu)-OH (653 mg, 13.9 mmol) and HOBT (213 mg, 13.9 mmol) were dissolved in 3 mL of dry DMF. DIC (176 mg, 13.9 mmol) was added to the vial, and this solution was added to the deprotected resin. The reaction proceeded until the resin tested negative for free amines in a Kaiser assay. Using the same procedure, Fmoc-His(Trt)-OH (3), Fmoc-Gly-OH (4), and Boc-Gly-OH (5) were coupled to the resin. The Lys -amine Alloc group was removed by mixing twice with 3 molar equivalents of borane-dimethylamine dissolved in 3 mL of DCM and 0.05 molar equivalent of tetrakis(triphenylphosphine)palladium(0) also dissolved in 3 mL of DCM. Both deprotection reactions proceeded for 10 min with a DCM wash in between. Deprotection of the amine group was confirmed by a Kaiser test. Six molar equivalents of methacrylic acid was added to 3 molar equivalents of HOBT in 3 mL of dry DMF. DIC was added, and (21) Tang, A.; Wang, C.; Stewart, R. J.; Kopecek, J. J. Controlled Release 2001, 72, 57-70.

Specific CoValent Immobilization of Proteins

Langmuir, Vol. 22, No. 26, 2006 11307

Figure 2. Solid-phase synthesis of the catalytic NH2-GGHYK methacrylate monomer. this solution was applied to the resin. The reaction proceeded until a negative Kaiser test indicated the -amine group had fully reacted. The methacrylated peptide was cleaved from the resin with 5:5:90 TIS/DCM/TFA and a spatula tip of tert-octylpyrocatechine to inhibit polymerization. After 3 h, the cleavage reaction was drained into 10-fold excess cold ether, and the resin was rinsed with TFA. The peptide 6 was titurated several times in cold ether and centrifuged. After five washes, the precipitate was dried under vacuum. The peptide monomer was analyzed by HPLC and positive ion MALDI mass spectrometry. Microbead Polymerization. The GGHYK peptide monomer was copolymerized with MABNTA and acrylamide at a molar ratio of 1:1:98 into microbeads by inverse microemulsion polymerization.22 The GGHYK (4.4 mg, 7 µmol) and MABNTA (2.3 mg, 7 µmol) monomers, acrylamide (50 mg, 0.7 mmol), ammonium persulfate (APS; 4 mg, 17.5 µmol), and 5 mol % N′,N′-methylenebisacrylamide (5.4 mg, 35 µmol) as the cross-linker were added to 200 µL of 0.1 M PBS (pH 7.4). The aqueous phase was added to 2 mL of cyclohexane containing 60 mg of a 3:1 Span 80/Tween 80 mixture. The solution was sonicated for 15 s and then deoxygenated by bubbling N2 for 30 min with stirring, after which 15 µL of TEMED was added. N2 was bubbled into the solution for an additional 15 min, after which the reaction was capped and left to polymerize overnight. After polymerization, the beads were washed first with hexane to remove the surfactant, methanol, and finally acetone. The beads were dried under vacuum for 2 h. The beads were weighed and resuspended in water before binding experiments. The beads were examined by bright field microscopy. They were polydisperse with a diameter ranging from 1 to 5 µm. Construction and Purification of GFP-YGH6. As a model protein, green fluorescent protein (GFP) was genetically tagged with a C-terminal Tyr-Gly-His6 (YGH6) peptide. The GFP-YGH6 was constructed by PCR amplifying the GFP coding region from an enhanced GFP gene contained in a pET20b plasmid.23 The T7 primer was used as the forward primer. The reverse primer contained additional sequence encoding the YGH6 tag and an XhoI restriction site for cloning. The purified PCR product was digested with XbaI and XhoI, ligated into the pET28c expression vector, and transformed into DH5R cells. The construct was verified by DNA sequencing. For expression, the modified GFP gene was transformed into an Escherichia coli strain, BL21(DE3), by chemical transformation. (22) Goh, S. L.; Nurthy, N.; Xu, M.; Frechet, J. M. J. Bioconjugate Chem. 2004, 15, 467-474. (23) Patterson, G. H.; Knobel, S. M.; Sharif, W. D.; Kain, S. R.; Piston, D. W. Biophys. J. 1997, 73, 2782-2790.

Single colonies from agar plates containing 50 µg/mL kanamycin were grown for 5 h at 26 °C in 5 mL of LB medium with 50 µg/mL kanamycin according to published procedures for correct GFP folding.24 LB cultures (500 mL) inoculated with the 5 mL cultures containing 50 µg/mL kanamycin were also grown overnight at 26 °C. The cells were centrifuged at 4500 rpm for 15 min at 4 °C. The cell pellets were stored at -70 °C. The frozen cell pellets were resuspended in 15 mL of binding buffer (20 mM Tris, 500 mM NaCl, pH 7.4) and 15 µL of Triton X. This was sonicated at power level 4 on ice three times for 20 s each time with 1 min breaks in between. The supernatant was mixed with 2.5 mL of Ni2+-NTA agarose beads equilibrated with binding buffer, allowed to settle in a column, then drained by gravity. The column was washed with 4 column volumes of binding buffer, 3 column volumes of binding buffer containing 60 mM imidazole, and last 3 column volumes of binding buffer containing 300 mM imidazole to elute the GFP. The purities of the collected fractions were determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using 12% polyacrylamide gels. After the protein was desalted into a phosphate-buffered saline (PBS; pH 7.4) on PD 10 Sepharose columns, it was concentrated using a 3000 MW cutoff ultrafiltration membrane. The concentrated protein was stored in 50% glycerol at -20 °C. Stability of GFP during Oxidative Tyrosine Cross-Linking. Green fluorescent protein was added to a solution containing 3.8 mM N-acetyl-L-tyrosine ethyl ester (ATEE) and 38 µM GGHNi(II). APS (15 mM) was added to the solution to initiate the formation of dityrosine catalyzed by the GGH-Ni(II) complex. Dityrosine formation was followed by measuring the increase in dityrosine fluorescence at 420 nm.25 GFP fluorescence (509 nm) was monitored simultaneously during the oxidative formation of dityrosine. The GFP fluorescence was not diminished during the reaction, demonstrating that GFP is not damaged by APS, the reactive intermediates, or products of the dityrosine cross-linking reaction. GFP-Binding Assays. For protein binding studies, the beads were first metalated by mixing with an excess of NiSO4. Unchelated Ni(II) was removed by washing the beads with water. The metalated beads were then mixed for 25 min with an excess of GFP-YGH6 in PBS (pH 7.4). Unbound GFP-YGH6 was removed by washing the beads three times for 15 min with PBS (pH 7.4). GFP-YGH6 specifically bound to NTA-Ni(II) was removed by washing the beads three (24) Merkel, J. S.; Regan, L. J. Biol. Chem. 2000, 275, 29200-29206. (25) Malencik, D. A.; Sprouse, J. F.; Swanson, C. A.; Anderson, S. R. Anal. Biochem. 1996, 242, 202-213.

11308 Langmuir, Vol. 22, No. 26, 2006

Figure 3. GFP stability during oxidative tyrosine cross-linking catalyzed by GGH-Ni(II) in solution. times with the chelating agent EDTA (100 mM, pH 7.4) to disrupt the NTA-Ni(II) complexes. The fluorescence of GFP-YGH6 (509 nm) and dityrosine (420 nm) bound to 0.9 mg of polyacrylamide beads was determined in a fluorescence plate reader. Covalent immobilization through dityrosine cross-linking was initiated by adding an oxidant, 15 mM APS, to beads with GFPYGH6 bound to NTA-Ni(II). After 15 min, the cross-linking reaction was quenched by washing the beads three times with 100 mM EDTA to remove the APS and to disrupt the Ni(II) complexes. Covalently bound GFP remained attached to the beads. Cross-Linking Specificity. The specificity of the cross-linking reaction was tested by adding 5 mM free amino acids to the solutions before the oxidative cross-linking reaction was initiated with 15 mM APS.

Results GFP-YGH6. GFP was a convenient model protein for these studies for several reasons. First, its intrinsic fluorescence allowed protein immobilization to be accurately quantitated by fluorimetry. Second, the fluorescence is dependent on the structural integrity of the protein and therefore indicated that the protein was not denatured or damaged by the immobilization procedure. Third, GFP is biochemically well-behaved in E. coli expression systems. GFP was tagged on the carboxy terminus with the sequence YGH6 on the basis of our earlier observations that the position of the tyrosine residue in relation to the His6 residues had a pronounced effect on the efficiency of dityrosine cross-linking.26 The YGH6 peptide tag was used to purify GFP by immobilized metal affinity chromatography using standard procedures. To test the stability of GFP under the conditions of oxidative cross-linking, GFP fluorescence was monitored as dityrosine formation was catalyzed by GGH-Ni(II) from free tyrosine in the presence of 15 mM APS. As shown in Figure 3, GFP fluorescence was not diminished during dityrosine formation in solution, demonstrating that GFP is not damaged directly by APS or by the products or intermediates of oxidative dityrosine cross-linking. Surface Immobilization of GFP-YGH6. Polyacrylamide beads polymerized with either the NTA or GGHYK monomers (Figure 4, bar 1) bound little GFP-YGH6 relative to the beads containing 1 mol % each of the NTA and GGHYK ligands (Figure 4, bar 3). Beads with 1 mol % GGHYK (bar 2) had background binding of GFP-YGH6 comparable to that of the polyacrylamideonly control beads (bar 1). GFP-YGH6 bound to the co-NTA(26) Stayner, R. S.; Min, D.-J.; Kiser, P. F.; Stewart, R. J. Bioconjugate Chem. 2005, 16, 1617-1623.

Endrizzi et al.

Figure 4. Fluorescence of GFP-YGH6 immobilized on polyacrylamide beads. The red bars represent bead-bound GFP fluorescence before washing with EDTA, and the yellow bars represent beadbound GFP fluorescence after extensive washing with EDTA to disrupt metal coordination bonds. Bars represent the average ( SD of at least three binding experiments.

Ni(II), co-GGHYK-Ni(II) beads could be eluted down to background levels with the Ni(II) chelator EDTA (bar 4). Together, these results demonstrate that there was little nonspecific binding of GFP-YGH6 to the polyacrylamide substrate and that the initial binding of GFP-YGH6 was predominately specific and reversible binding to the NTA-Ni(II) complexes. After treatment with APS for 15 min, the bound GFP-YGH6 could not be removed from the beads by extensive washing with 100 mM EDTA, demonstrating that the protein was covalently bound to the bead surface (Figure 4, bars 5-7). Covalently bound GFP-YGH6 increased slightly with increasing APS. At 15 mM APS, ∼75% of the GFP-YGH6 initially bound to NTA-Ni(II) became irreversibly bound. In the present studies, the NTANi(II) to catalytic peptide molar ratio was 1:1. It may be possible to increase the cross-linking efficiency of GFP-YGH6 beyond 75% by fine-tuning the NTA-Ni(II) to GGHYK-Ni(II) ratio. The surface density of the conjugation sites can be controlled by varying the concentration (mol %) of metal ligands copolymerized with acrylamide. Reaction Specificity. Dityrosine has a characteristic fluorescence emission at 420 nm which was monitored during the oxidative cross-linking reaction. Dityrosine fluorescence rose concomitantly with the increase in covalently bound GFP (Figure 4). If cross-linking occurred through dityrosine formation, free tyrosine in solution should competitively inhibit the cross-linking reaction. Indeed, 5 mM tyrosine inhibited covalent immobilization of GFP by about 50% (Figure 5). Addition of free glycine, serine, or lysine had no effect on the extent of covalent coupling. The readily oxidizable amino acid tryptophan quenched the crosslinking reaction but to a lesser extent than tyrosine. Histidine, an oxidizable amino acid, could not be tested in this assay because it disrupted the binding of the His6 tag to the NTA-Ni(II).

Discussion The results demonstrate proteins can be efficiently coupled to surfaces through dityrosine cross-links in a process analogous to natural protein dityrosine cross-linking. In our scheme, an immobilized catalytic peptide-Ni(II) complex mimics the metalloenyzmes that catalyze dityrosine formation in nature. Two lines of evidence suggest that the GFP-YGH6 is indeed covalently bound through dityrosine cross-links. First, dityrosine has a characteristic fluorescent emission at 420 nm. Addition of APS resulted in a 4-fold increase above the background in 420 nm

Specific CoValent Immobilization of Proteins

Figure 5. Inhibition of covalent GFP immobilization by free amino acids. The indicated amino acid (5 mM) was added before the crosslinking reaction was initiated by the addition of APS (15 mM). Bars represent bead-bound GFP after extensive washing with EDTA (100 mM) to disrupt metal coordination bonds. Bars represent the average ( SD of at least three binding experiments.

fluorescence (Figure 4, blue bars). Second, the addition of free N-acetyl-L-tyrosine ethyl ester before the addition of APS inhibited the covalent cross-linking of GFP-YGH6 to the beads by approximately 50% without diminishing the dityrosine fluorescence (Figure 5). Of several other amino acids tested, only the readily oxidizable tryptophan showed inhibition but to a lesser extent than tyrosine. Cysteine and histidine were not tested, but it is likely that they would also be oxidized in the vicinity of the catalytic peptide and inhibit cross-linking. The metal-peptide-catalyzed oxidative cross-linking system (Figure 1) circumvents most of the major limitations of conventional approaches to chemical and genetic protein surface immobilization described earlier. The system has several advantages. (1) The conjugation site is preestablished as a specific, noncovalent metal complex before chemically reactive groups are created under mild oxidizing conditions. Reactive intermediates of the tyrosine cross-linking reaction are created only in the vicinity of the colocalized catalytic peptide-Ni(II) complexes, limiting unwanted side reactions that could damage or inactivate the target proteins. As a result, proteins are covalently coupled through a specific site and have a defined orientation to the surface determined by the position of the YGH6 peptide tag. (2) Covalent coupling through tyrosines is orthogonal to other covalent conjugation chemistries that target nucleophilic side chains. (3) Copolymerization of the metal-binding monomers with acrylamide creates a convenient, protein-resistant hydrogel substrate that limits nonspecific binding of nontarget proteins and preserves the target protein activity. The hydrogel substrate is adaptable to several configurations in addition to the microspheres reported here, including thin coatings on magnetic microspheres, glass slides, microwell plates, and other flat surfaces. (4) The method can be adapted to several other surface modification chemistries. For example, the metal-binding ligands can be synthesized on PEGylated alkanethiols27 that would allow creation of mixed self-assembling monolayers or surfactants such as pluronics that would allow proteins to be immobilized on hydrophobic interfaces.28 (5) The genetic modification is a structurally benign peptide tag that can be placed at either the C- or N-termini of the target protein. In comparison, the much larger (MW ) 20000) enzymes used as fusion partners in the ASDCL methods could interfere in some cases with expression, protein-protein interactions, or assembly of multimeric pro(27) Prime, K. L.; Whitesides, G. M. Science 1991, 252, 1164-1167. (28) Ho, C.-H.; Limberis, L.; Caldwell, K. D.; Stewart, R. J. Langmuir 1998, 14, 3889-3894.

Langmuir, Vol. 22, No. 26, 2006 11309

teins.7,8 (6) Immobilized metal affinity chromatography based on His6 tags is a widely accepted and practiced methodology. In fact, human cDNA expression libraries have been created in which every library member is His6-tagged.29 The multitudes of existing His6-tagged proteins and peptides can be easily converted to H6GY-tagged proteins without affecting expression and purification protocols. (7) Because the YGH6 tag is genetically encoded, the target protein does not need to be purified before modification. Rather, H6GY-tagged proteins can be specifically and covalently immobilized in a single step directly from a crude lysate of cells expressing the gene. (8) Like the traditional His6 tag and unlike the ASDCL enzymes, the H6GY tag can be used under nonphysiological conditions, such as high salt concentrations, and in denaturing solutions, such as 8 M urea. (9) Finally, the reagents and ligands are inexpensive, widely available, and stable. The method may be particularly useful in the continuing development of protein arrays for medical diagnostics and as basic biomedical research tools. In medical diagnostics, protein arrays are widely expected to have a dramatic impact on human health care in the coming years.3,4,30,31 The proteome is much more complex than the genome because of posttranscriptional modification of RNA, posttranslational modification of proteins, and complex networks of protein interactions. For diagnostic purposes the proteome therefore contains much more useful information about physiological conditions and potential disease states than the genome. Eventually it will be possible to routinely screen physiological samples in highly multiplexed tests for early indicators of cancers, cardiovascular disease, and metabolic disorders, such as type II diabetes.32-36 As basic biomedical research tools, protein microarrays have been used to discover and validate disease biomarkers, to monitor the response of intracellular signaling proteins to external signals, to monitor protein posttranslational modifications, to map protein-to-protein connections of the human proteome, for high-throughput protein functional analysis such as ligand binding, and much more.37-40 Our system meets the needs of optimized array developments low nonspecific binding to increase the signal-to-noise ratio, preservation of protein activity that allows higher array feature densities and efficient use of potentially expensive protein reagents, robust attachment to withstand washing procedures (29) Bussow, K.; Nordhoff, E.; Lubbert, C.; Lehrach, H.; Walter, G. Genomics 2000, 65, 1-8. (30) Kodadek, T. Trends Biochem. Sci. 2002, 27, 295-300. (31) Petrik, J. Vox Sang. 2001, 80, 1-11. (32) Qiu, J.; Madoz-Gurpide, J.; Misek, D. E.; Kuick, R.; Brenner, D. E.; Michalidis, G.; Haab, B. B.; Omenn, G. S.; Hanash, S. J. Proteome Res. 2004, 3, 261-267. (33) Sanchez-Carbayo, M.; Socci, N. D.; Lozano, J. J.; Haab, B. B.; CordonCardo, C. Am. J. Pathol. 2006, 168, 93-103. (34) Gao, W.-M.; Kuick, R.; Orchekoski, R. P.; Misek, D. E.; Qiu, J.; Greenberg, A. K.; Rom, W. N.; Brenner, D. E.; Omenn, G. S.; Haab, B. B.; Hanash, S. M. BMC Cancer 2005, 5, 110. (35) Horn, S.; Lueking, A.; Murphy, D.; Staudt, A.; Gutjarh, C.; Schulte, K.; Konig, A.; Landsberger, M.; Hans, L.; Felix, S. B.; Cahill, D. J. Proteomics 2006, 6, 605-613. (36) Quintana, F. J.; H., H. P.; Elizur, G.; Merbl, Y.; Domany, E.; Cohen, I. R. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 14615-14621. (37) Mattoon, D.; Michaud, G.; Merkel, J.; Schweitzer, B. Expert ReV. Proteomics 2005, 2, 879-889. (38) Nielsen, U. B.; Cardone, M. H.; Sinskey, A. J.; MacBeath, G.; Sorger, P. K. Proc. Natl. Acad. Sci. U.S.A. 2003, 16. (39) Gembitsky, D. S.; Lawlor, K.; Jacovina, A.; Yaneva, M.; Temps, P. Mol. Cell. Proteomics 2004, 3, 1102-1118. (40) Ptacek, J.; Devgan, G.; Michaud, G.; Zhu, H.; Zhu, X.; Fasolo, J.; Guo, H.; Jona, G.; Breitkreutz, A.; Sopko, R.; McCartney, R. R.; Schmidt, M. C.; Rachidi, N.; Lee, S.-J.; Mah, A. S.; Meng, L.; Stark, M. J. R.; Stern, D. F.; Virgilio, C. D.; Tyers, M.; Andrews, B.; Gerstein, M.; Schweitzer, B.; Predki, P. F.; Snyder, M. Nature (London) 2005, 438, 679-684.

11310 Langmuir, Vol. 22, No. 26, 2006

during array manufacture or screening, and potential for efficient, cost-effective manufacturing.

Conclusions Adapting the well-known His6 metal-binding peptide tag to a covalent immobilization scheme based on oxidative dityrosine formation provides a convenient, versatile, and flexible technology. The method combines the advantages of specific interactions, such as gentle and oriented immobilization, with the advantages

Endrizzi et al.

of covalent cross-linking, mainly stability during extensive washing or high shear stresses under flow. These attributes suggest that the method may be broadly applicable to protein surface immobilization and in particular to the further optimization of protein microarrays. Acknowledgment. This work was supported by a grant from the NIH (GM070826). LA0618216