Targeted Protein Functionalization Using His-Tags - Bioconjugate

This approach provides: (1) specificity in that only His-tagged targets are modified, (2) regio-specific control ... View: PDF | PDF w/ Links | Full T...
1 downloads 0 Views 264KB Size
Bioconjugate Chem. 2004, 15, 969−982

969

Targeted Protein Functionalization Using His-Tags Gavin D. Meredith,*,† Hayley Y. Wu, and Nancy L. Allbritton* Department of Physiology & Biophysics, University of California - Irvine, Irvine, California 92697-4560. Received May 1, 2004; Revised Manuscript Received July 6, 2004

With the impressive growth in gene sequence data that has become available, recombinant proteins represent an increasingly vast source of molecular components, with unique functional and structural properties, for use in biotechnological applications and devices. To facilitate the use, manipulation, and integration of such molecules into devices, a controllable method for their chemical modification was developed. In this approach, a trifunctional labeling reagent first recognizes and binds a His-tag on the target protein’s surface. After binding, a photoreactive group on the trifunctional molecule is triggered to create a covalent linkage between the reagent and the target protein. The third moiety on the labeling reagent can be varied to bring unique chemical functionality to the target protein. This approach provides: (1) specificity in that only His-tagged targets are modified, (2) regio-specific control in that the target is modified proximal to the His-tag, the position of which can be varied, and (3) stoichiometric control in that the number modifications is limited by the binding capacity of the His-tag. Two such labeling reagents were designed, synthesized, and used to modify both N- and C-terminally His-tagged versions of the enzyme murine dihydrofolate reductase (mDHFR). The first reagent biotinylated the enzyme,while the second served to attach an oligonucleotide to yield a proteinDNA conjugate. In all cases, modification in this manner brings new functionality to the protein while leaving the enzymatic activity intact. The protein-DNA conjugate was used to specifically immobilize the active enzyme through DNA hybridization onto polystyrene microspheres, a step toward creating a functional protein microarray.

INTRODUCTION

Reagents and methods to selectively chemically modify proteins are becoming increasingly important for the study and utilization of protein functions and interactions, particularly due to the explosive growth in our knowledge of protein sequences from genome sequencing efforts and the drive to conduct proteomic research. Robust technologies for the rational chemical modification of proteins are needed for subsequent immobilization to create protein arrays for screening of function and interactions. The chemical attachment of prosthetic groups is also important for the creation of new proteinbased biosensors and reporter molecules for life science research, biodetection assays, and therapeutics. Looking further ahead, controllable modification of proteins should help to foster a new era in biomaterials design and manufacture and semi-synthetic molecular engineering. In general, when chemically modifying a protein, the goal is to introduce novel chemical functionality while preserving its native function and properties. Ideally, one would be able to covalently attach a precise number of unique chemical groups to one or more specific locations on the surface of any target protein in solution. This would provide maximal versatility and control over subsequent handling of the target while minimizing the likelihood of disrupting its native function. It is also important that the approach integrate with established technologies for recombinant protein production. Cur* Corresponding authors. G.D.M.: phone (760) 476-6127; fax (508) 546-7742; e-mail [email protected]. N.L.A.: phone (949) 824-6493; fax (949) 824-8540; e-mail [email protected]. † Current Address: Invitrogen Corporation, 1600 Faraday Avenue, Carlsbad, CA 92008.

rently, the most general approach is to use electrophilic reagents that can react with nucleophilic groups on the surface of a target protein; however, this approach tends to yield a heterogeneous mixture of products since almost all proteins carry a number of nucleophilic groups on their surfaces. The thiol side-chain of cysteine is the most nucleophilic group that naturally occurs on a protein’s surface and thiol-specific modifying reagents are numerous. Hence, to date, the best approach to engineer a site for selective chemical modification into a protein has been to genetically introduce a non-native cysteine at the desired site. After production and purification, the recombinant protein is covalently modified through thiolspecific chemistry. This approach is often successful, but can fail for a number of reasons. Foremost, the number and positions of native cysteines can be vitally important for the structural or functional integrity of many proteins; insertion of a non-native cysteine or removal of native cysteine can lead to misfolding or loss of function. Similarly, inappropriately modified native cysteines can lead to protein inactivation. This is particularly likely for enzymes that utilize a cysteine in their active site, including cysteine-based oxidoreductases, dehydrogenases, proteases, and some metal-binding proteins (1). Similar difficulties can arise from methods that rely on reaction between a terminal cysteine and a terminal thioester as typified by the native chemical ligation and expressed protein ligation approaches (2, 3). In general, the site of modification by these elegant methods is restricted to a protein terminus; however, there are examples where multistep syntheses have permitted the introduction of modifications at internal positions (4, 5). Oxidation-mediated protein labeling is also generally restricted to the termini (6-8). Another strategy to engineer sites for specific chemical modification in pro-

10.1021/bc0498929 CCC: $27.50 © 2004 American Chemical Society Published on Web 08/10/2004

970 Bioconjugate Chem., Vol. 15, No. 5, 2004

teins has been to develop expression systems that allow for the incorporation of unnatural amino acids that can serve as unique targets for covalent modification; however, these strategies have yet to prove widely applicable (9-11). An alternative strategy to alter a protein’s surface chemistry involves introducing unique non-native stretches of amino acids into a recombinant protein’s native sequence (12). Such stretches of amino acids serve as integral “tags” that provide novel chemical functionality or can serve as targets for molecular recognition. A number of such tags have been described; they have been used principally for noncovalent immobilization or to allow antibody recognition. One notable exception has been the development of a tag that can serve as the target for novel biarsenical labeling reagents (13). Another genetic-engineering based approach is to link the sequence of two (or more) protein domains to yield novel recombinant “fusion” or “chimeric” proteins. In general, such tags have not provided a means for further covalent modification; however, recently a specific 207-amino-acid fusion tag has been developed that can serve as a site for selective covalent modification (14, 15). In protein research, the most versatile and widely used genetically encoded tag is the “polyhistidine” or “His-tag”, which typically consists of a string of 2-6 consecutive histidine residues (16, 17). Introduction of a His-tag at either the N- or C-terminus of a recombinant protein was first described as a convenient and generally nonperturbative means to facilitate purification of recombinant protein by immobilized metal-affinity chromatography. Functional His-tags can also be placed internally within the sequence of recombinant proteins, their sequences can be varied to include nonhistidine residues, and they can even take the form of engineered histidine “patches”, consisting of histidines that are close in space though not in sequence within a folded protein structure (18-23). Immobilization takes place through coordination of two or more of the histidine residues with a metal ion (usually Ni(II)) that is incompletely coordinated by a substrateattached chelating group such as nitrilotriacetic acid (NTA). Since its introduction, variants of this basic molecular recognition technology have been developed and utilized for a number of applications beyond protein purification; examples include sensitive target protein detection, protein structure studies, and protein immobilization for binding studies using surface plasmon resonance or atomic force microscopy (24-28). Since in its simplest form, as either an N- or C-terminal extension, the His-tag can be easily incorporated into expression vectors, genome-scale expression libraries of His-tagged proteins have been generated (29-33). In a groundbreaking study by Snyder and co-workers, 5800 different His-tagged yeast proteins, corresponding to about 80% of the yeast proteome, were expressed, purified, and arrayed on nickel-coated glass slides for subsequent probing (32). While this noncovalent immobilization served the purposes of the investigators, we reason that such arrays could be made more robust if the His-tag could serve as the target for covalent surface attachment. One solid-phase based approach to such affinity-directed covalent capture has been reported by Guisan and colleagues where His-tagged proteins have been bonded to chromatography resins that simultaneously carry metal-chelating groups and reactive epoxide groups (34). We envision a broader use of the His-tag as a unique affinity target to direct proximal covalent bonding to a recombinant protein in solution. Solution-phase chemical modification is important to allow subsequent handling

Meredith et al.

of the modified proteins and is required for the creation of protein-based reagents such as soluble therapeutic antibody-enzyme conjugates used in antibody-directed enzyme prodrug therapy (35). We judge the His-tag to have the best ensemble of features since as noted it is small relative to the size of most proteins of interest, it is generally nondisruptive to its host protein’s structure and function, it can be placed at either terminus and even internally within its host’s sequence, and it is widely used, particularly in the construction of expression libraries. Similar reasoning has led others to use the Histag as a solution-phase target for the noncovalent tethering of bifunctional NTA-containing conjugates to protein (36, 37). To direct covalent modification in solution, we have synthesized two different water-soluble trifunctional molecules based on shared design principles. Ultimately, we envision this as a step toward the creation of encoded libraries of proteins (38) and a competitive technology for the creation of self-assembling protein arrays, engineered protein-containing molecular assemblies and nanostructures, and the integration of proteins into devices such as biosensors. EXPERIMENTAL PROCEDURES

Instrumentation. The following instrumentation was used: an Agilent (Palo Alto, CA) 1050 and an 1100 HPLC system, each equipped with quartenary gradient pump, UV absorbance detector, fluorescence detector, and Chemstation software, a Cary 20 UV/vis absorbance spectrophotometer with kinetics software (Varian, Palo Alto, CA), a PerSeptive Voyager STR-DE Biospectrometry matrix-assisted laser deionization time-of-flight (MALDITOF) mass spectrometer (Applied Biosystems, Foster City, CA), a Micromass LCT electrospray ionization mass spectrometer (Waters Corp., Milford, MA), a DRX-500 nuclear magnetic resonance spectrometer (Bruker Biospin, Rheinstetten, Germany), a Perseptive Cytofluor 4000 microplate reader (Applied Biosystems), a Rotavaporator R-200 (Bu¨chi, Flawil, Switzerland) with vacuum pump (Fisher Scientific, Hampton, NH), a Blakray B-100AP handheld long-UV lamp, a French-press, a Beckman preparative J-21 centrifuge (Beckman Coulter, Fullerton, CA), a Sorvall ultracentrifuge and tabletop low-speed centrifuge (Kendro, Asheville, NC), a tabletop 5415C microcentrifuge (Eppendorf AG, Hamburg, Germany), a mini-Protean II gel apparatus with Powerpac 300 power supply (Bio-Rad, Hercules, CA), and a Milli-Q water purification system (Millipore, Billerica, MA). The following columns were used: 250 × 4.60 mm Jupiter C4 packed with 5 µm beads, 250 × 10 mm Jupiter C4 packed with 15 µm beads, 250 × 4.60 mm Jupiter C18 packed with 5 µm beads, and a 250 × 10 mm Jupiter C18 packed with 10 µm beads with a 50 × 10 mm guard column. All Jupiter columns had 300 Å pores and were from Phenomenex (Torrance, CA). Also used: a 4.6 × 250 mm Zorbax GF-250 column with 4 µm beads (Agilent) and a HiPrep 16/60 Sephacryl S-100 HR column (Amersham Biosciences, Buckinghamshire, England). Materials. DHFR expression plasmids, Escherichia coli host strains, Ni-NTA resins, and DAPase were all obtained from Qiagen (Valencia, CA). Bacterial culture media, IPTG, and ampicillin came from USBiological (Swampscott, MA). Buffers, salts, chemicals, biochemicals, and solvents were obtained from Sigma-Aldrich (St. Louis, MO) unless otherwise noted. Herring sperm DNA was from Promega (Madison, WI). HPLC solvents and methanol were purchased from EM Sciences (Hatfield, PA). N-(5-amino-1-carboxypentyl)iminodiacetic acid (NTA-

Targeted Protein Functionalization

amine) came from Toronto Research Chemicals (North York, On., Canada). Sulfo-SBED, monomeric avidin, D-biotin, Immunopure HABA, Coomassie protein assay reagent, Gelcode Blue stain, and Gelcode SilverSNAP stain were from Pierce Chemical (Rockford, IL). β-Maleimido-proprionic acid was purchased from Molecular Biosciences (Boulder, CO). SPOS resin, PyBOP, and HOBt were from NovaBiochem (EMD Biosciences, San Diego, CA). Fmoc-Bpa-OH was purchased from Bachem (Bubendorf, Switzerland). NAP-10 gel filtration columns were from Amersham. Centricon Plus-20 PLGC devices and ZipTips came from Millipore. Precast minigels and protein electrophoresis standards were from Bio-Rad. Polymeric microspheres were from Polysciences (Warrington, PA). DNA-bind plates were from Corning (Corning, NY). Plastic, untreated, flat-bottom 96-well plates were from Nalge Nunc (Rochester, NY). The following synthetic modified oligonucleotides were obtained from Integrated DNA Technologies (Coralville, IA); all linkers were C6-type. Modifications included 5′terminal thiol (SH-) groups, 5′-terminal amine (NH2-) groups, 3′-terminal biotin, and terminal carboxyfluorescein (FAM) or hexachlorofluorescein (HEX) fluorophores. Thiolated oligonucleotides used for protein labeling (the “sense” sequence is underlined): Oligobiotin: 5′SH-TTTTTCGTTGATAACCTGTCCATCTCTA-biotin OligoFAM: 5′SH-TTTTTCGTTGATAACCTGTCCATCTCTA-FAM Aminated oligonucleotides used to create capture surfaces: Complementary: 5′ NH2-TTTTTTTTTTAGAGATGGACAGGT Noncomplementary: 5′ NH2-TTTTTTTTTCCCTAAGAAGACGGA Fluorescent oligonucleotides used to test capture surfaces: FAM-sense: 5′ FAM-ACCTGTCCATCTCTA HEX-anti-noncomplementary: 5′ HEX-TCCGTCTTCTTAGGG Expression and Purification of N- or C-terminally His-Tagged Murine Dihydrofolate Reductase (His6mDHFR and mDHFR-His6, Respectively) and Tagless Murine Dihydrofolate Reductase (mDHFR). To produce the His6-mDHFR and mDHFR-His6 proteins, competent E. coli (M15 [pREP4]) were separately transformed with the respective expression constructs pQE40 and pQE-16 using the source company’s protocols. The amino acid sequence of His6-mDHFR differs from the sequence that can be obtained from the Swissprot database, DYR_MOUSE (accession number P00375), in that it has 14 additional residues at its N-terminus, MRGSHHHHHHGSGI-, has the C-terminal aspartate replaced by 17 residues, -GSRSACGTPGRPAAKLN, and has the cysteine in the seventh position replaced by a serine. The amino acid sequence of mDHFR-His6 differs from DYR_MOUSE in that it has six additional residues at its N-terminus, MRGSGI-, has the C-terminal aspartate replaced by 10 residues, -GSRSHHHHHH, and has the cysteine in the seventh position replaced by a serine. Large-scale cultures were grown in LB medium containing neomycin (25 µg/mL) and ampicillin (100 µg/mL) at 37 °C, induced with 1 mM IPTG, harvested by centrifugation for 15 min at 6000 rpm in a Beckman J-10 rotor in a Beckman JA-21 centrifuge, frozen, and stored at -70 °C based on the Qiagen protocol. For protein purification 20-30 g of pelleted cells were resuspended and brought to a volume of 100 mL in lysis buffer (50 mM potassium phosphate (pH 8.9), 300 mM NaCl, 10 mM imidazole, 1×

Bioconjugate Chem., Vol. 15, No. 5, 2004 971

protease inhibitors). The cells were passed through a French press twice at a pressure of 1200-1600 psi. Cleared cellular lysate was obtained after centrifugation at 4 °C at 100000g for 1 h using an F-28/36 rotor in a Sorvall ultracentrifuge. Protease inhibitors were added at 1× concentration to the cleared lysate prior to loading onto a 10 mL Ni-NTA Superflow column. The column was preequilibrated in lysis buffer and peristaltically pumped at 1.2 mL/min at 4 °C. Following loading, the column was washed with 50 mL of low-imidazole buffer (20 mM imidazole, 300 mM NaCl, 50 mM potassium phosphate (pH 8.3)). Both His-tagged mDHFRs elute as a single peak at approximately 100 mM imidazole in a linear gradient of increasing imidazole. Peak fractions were pooled, dialyzed against mDHFR buffer (20 mM NaCl, 20 mM sodium phosphate (pH 7.6), 20% glycerol), and stored at 4 °C. Untagged mDHFR was generated from His6-mDHFR by proteolytic digest with the enzyme DAPase followed by repurification on NTA-agarose resin according to the manufacturer’s protocols. Assay of Enzyme Activity in Solution. At room temperature, the oxidation of β-NADPH to NADP+ that is coupled with reduction of DHFA to tetrahydrofolate (THFA) by mDHFR, was monitored as a decrease in absorbance at 340 nm with a Cary 20 UV/vis spectrophotometer (39). The reaction buffer, prepared immediately before use, consisted of either DHFA and β-NADPH at 67 µM with 13 mM β-mercaptoethanol and 1 mg/mL BSA, or DHFA and β-NADPH at 50 µM with 20 mM β-mercaptoethanol and 1 mg/mL BSA, in degassed 100 mM potassium phosphate buffer (pH 7.5). After recording the steady absorbance of 1 mL of the reaction buffer alone over 2 min, 1-2 µg of mDHFR or mDHFR-oligo conjugate was added to the cuvette, mixed by several inversions, and returned to the spectrophotometer. Changing absorbance at 340 nm was recorded for >10 min. Reaction velocities were determined from a 4-min linear portion of the recordings that typically began 30 s to 3 min after mixing. Activities were compared based on reaction velocity per microgram of added mDHFR. Synthesis and Purification of NAB. The synthesis of NAB is shown in Scheme 2. NTA-amine (4 mg, 15 µmol) was dissolved/suspended in 1 mL of 50% anhydrous, deaminated dimethyl formamide (DMF), 50% anhydrous dimethyl sulfoxide (DMSO) containing triethylamine (TEA) (10 µL, 69 µmol). To this was added Sulfo-SBED (10 mg, 11 µmol) dissolved in 1:1 DMF/ DMSO. The reaction was stirred under nitrogen overnight, in the dark, at room temperature. After the reaction, the mixture was reduced to a yellowish oily concentrate under high vacuum. The concentrate was redissolved in DMF or dimethyl acetamide (DMA) and purified by RP-HPLC using a C4 column (250 × 4.6 mm) pumped at 1 mL/min. Solvent A was water/0.01% trifluoroacetic acid (TFA) and solvent B was acetonitrile/ 0.01% TFA. The column was loaded and pumped for 10 min in 10% solvent B, followed by a brief linear gradient, from 10 to 27% solvent B over 5 min, then followed by 15 min at 27% solvent B. NAB elutes as the second major peak with a retention time of 28.5 min. Since NAB is light sensitive, preparative purification was conducted with the detector light sources turned off. Peak fractions were identified by UV absorbance spectroscopy of samples diluted into water. Peak fractions were pooled and dried down or used immediately in labeling reactions. The yield under these conditions was 50%. Synthesis and Purification of NBzM. The solidphase synthesis of NBzM is shown in Scheme 3. All steps were at room temperature unless otherwise noted. All

972 Bioconjugate Chem., Vol. 15, No. 5, 2004

amine deprotection and coupling steps were monitored with the TNBS test as described in the NovaBiochem catalog (TNBS is 2,4,6-trinitrobenzenesulfonic acid). In a 25-mL manual peptide synthesis reaction vessel, FmocLys(Mtt)-Wang resin (2.75 g, 1.98 mmol) was swollen in dichloromethane (DCM) for 1 h. Fmoc deprotection was effected by six successive 15 min treatments of the resin with 20% piperidine in DCM. After deprotection, the resin was washed extensively with DCM. To the washed resin was added a 20-fold excess of tert-butyl-bromoacetate (6.4 mL, 40 mmol) with an equal amount of diisopropylethylamine (DIPEA) (7.2 mL, 41 mmol) and 2 mL of DCM. This reaction vessel was shaken at 350 rpm at 45 °C for 27 h. At this point, a TNBS test indicated that there were some residual free amines present on the resin beads. Fresh reagents were added and the vessel was further incubated at 45 °C with shaking for 18.5 h. The beads were washed successively with DCM, ethyl acetate, and DCM again. The Mtt group was removed by five 3-min treatments with 10 mL of 1% TFA, 5% triisopropyl silane (TIS) in DCM. After extensive washing with DCM, the beads were washed twice with 10 mL of 1:1 DCM/DMA. Fmoc-BPA-OH (1.95 g, 3.96 mmol), PyBOP (2.06 g, 3.96 mmol), HOBt (0.606 g, 3.96 mmol), and DIPEA (1.38 mL, 7.92 mmol) were added to the beads in 10 mL of 1:2 DCM/DMA. The beads were agitated with bubbling dry nitrogen for 17.5 h. Fresh reagents were added in half the previous amounts and allow to further react for 3 h. The beads were washed with DCM/DMA followed by DCM prior to Fmoc deprotection as previously carried out. After washing with DCM, the beads were washed twice with 10 mL of 1:1 DCM/DMA. β-Maleimido-proprionic acid (0.375 g, 2.22 mmol), PyBOP (1.155 g, 2.22 mmol), HOBt (0.340 g, 2.22 mmol), and DIPEA (780 µL, 4.44 mmol) were added to the beads in 12 mL of 1:2 DCM/DMA. The beads were allowed to react with agitation for ∼48 h. After washing of the sample, an equivalent amount of fresh reagents were added and allowed to react for seven more hours. The beads were washed successively with DCM, ethyl acetate, methanol, DCM again, and dried briefly under vacuum. Products were cleaved from the resin with several 5-min treatments with 10 mL of 95% TFA, 5% anisole. The material was reduced to near dryness with rotary evaporation, precipitated with ice-cold diethyl ether, and dried under high vacuum. The dried off-white material was dissolved in DMA and purified by semipreparative RP-HPLC over a Jupiter C18 column (50 × 10 mm guard, 250 × 10 mm main) pumped at 2.5 mL/min. Solvent A was water/0.01% TFA and solvent B was acetonitrile/0.01% TFA. The column was loaded in 10% solvent B and immediately linearly ramped up to 28.7% solvent B over 5 min. This was followed by a shallow linear gradient from 28.7 to 29.3% solvent B over 20 min, followed by a 5-min linear ramp to 95%. NBzM elutes as the forth major peak that starts to elute after 24 min. Peak fractions were pooled, dried by lyophilization, and stored dry in the dark. The yield, relative to the amount of Fmoc-Lys(Mtt) on the resin, was 12%. Coupling of NBzM to Thiolated Oligonucleotides. Prior to reaction with NBzM, thiolated synthetic oligonucleotides were dissolved in Milli-Q water and reduced by overnight incubation with 2% TEA (v/v) and 50 mM DTT as recommended by the manufacturer. After reduction, thiolated oligonucleotides were separated from small molecules by gel filtration over a NAP-10 column preequilibrated with 50 mM sodium phosphate (pH 7.0), 300 mM NaCl. A 30-fold excess of dry, weighed NBzM was

Meredith et al.

immediately added to the reduced thiolated oligonucleotide and allowed to react at room temperature overnight. Reduced thiolated oligonucleotide was almost entirely converted to NBz-oligo and was purified by RP-HPLC at neutral pH with a semipreparative C4 column (250 × 10 mm) pumped at 2.5 mL/min (40). Solvent A was 100 mM ammonium acetate (pH 7.0), and solvent B was pure acetonitrile. The column was loaded in 5% solvent B and immediately linearly ramped to 20% B over 30 min. NBz-oligoFAM elutes as the second major peak with a retention time of 25.2 min. MALDI-MS was used to confirm the mass of the NBz-oligo product. The peak fractions that contained the desired mass were pooled, lyophilized to dryness, and resuspended in sterile, Milli-Q water. The concentration of conjugate was determined by absorbance at 260 nm and a calculated extinction coefficient based on the nucleotide sequence and adjusted to account for the presence of benzophenone (41). Yields were typically ∼50% relative to the amount of starting thiolated oligo. To charge the conjugate molecules with Ni(II), concentrated MOPS buffer, pH 7.9, was added to 20 mM followed by addition of NiSO4 to 1.1-fold excess over the conjugate concentration (typically, 0.5-1 mM) (24). Ni(II) charged conjugates (Ni:NBz-oligo) were allowed to equilibrate at room temperature for several hours and then stored at 4 °C. Photocrosslinking. For photoconjugation, purified target protein (10-200 µM) was preincubated from 1 to 12 h at 4 °C with an equimolar amount of trifunctional labeling reagent in 20 mM sodium phosphate (pH 7.68.0) and 300 mM NaCl. Following preincubation, the mixture was adjusted to 10 mM imidazole by addition of 1 M imidazole (pH 8.0) 30-60 min prior to irradiation. The reaction mixture, typically 50-500 µL, was contained in an uncapped microcentrifuge tube tightly sealed with Saran-wrap affixed by an elastic band. The tube(s) were placed in a cooled aluminum block and irradiated with long-UV light (Blak-Ray B-100AP lamp, ∼365 nm) from above (∼6 cm from source) at 4 °C for 3-18 h with stirbar agitation. This corresponded to an irradiation energy of 0.28-1.7 MJ/cm2. Purification of Protein Conjugates. Biotinylated protein conjugates, either NAB-labeled or NBz-oligobiotinlabeled, were purified at room temperature by batch affinity chromatography using an Immunopure Immobilized Monomeric Avidin kit as per the manufacturer’s instructions. Eluted biotinylated proteins were dialyzed, using Spectropore 10 kDa Mw cutoff membranes, extensively against 20 mM NaCl, 20 mM sodium phosphate, 20% glycerol for storage. After photoconjugation, mDHFRoligoFAM conjugates were purified at room temperature by size-exclusion chromatography on a HiPrep 16/60 Sephacryl S-100 HR column installed on an Agilent 1100 series HPLC system equipped with a quartenary pump and UV-visible variable wavelength absorbance and fluorescence detectors. The column was equilibrated and run in 600 mM NaCl, 50 mM sodium phosphate (pH 7.6) at 0.25 mL/min. One-milliliter fractions were collected. Peak fractions were pooled and concentrated and exchanged into storage buffer by centrifugal ultrafiltration using Centricon Plus-20 PLGC devices in a room temperature tabletop centrifuge as per the manufacturer’s instructions. Absorbance Spectroscopy. All concentrations were determined based on absorbance measurements made with a quartz cuvette in a Cary-20 UV-visible spectrophotometer. The amount of biotin present in a pure sample of NAB was determined by a spectrophotometric assay using HABA and avidin (42). This quantity was

Targeted Protein Functionalization

used along with absorbance measurements to determine a molar extinction coefficient at 270 nm, 270, of NAB in water. This extinction coefficient was used for all subsequent determinations of NAB concentration. Oligonucleotide concentrations were determined according to Beer’s law based on absorbance at 260 nm and sequencedependent calculated extinction coefficients provided by Integrated DNA Technologies. Pure mDHFR concentrations were determined from absorbance measurements at 280 nm in a quartz cuvette and a calculated extinction coefficient (280 ) 25 440) based on amino acid content (43). mDHFR-oligo conjugate concentrations were determined by a microscale Bradford assay using a commercial Coomassie Protein Assay reagent and calibrated relative to staining of pure mDHFR with the same reagent; the oligo portion did not stain with this reagent. Mass Spectrometry of Proteins, Oligos, and Conjugates. Protein samples were desalted using ZipTipC18 devices as per the manufacturer’s instructions. Proteins were eluted in ∼2 µL, diluted to 10 µL in sinapinic acid solution (as per the instrument manual), and spotted directly onto a sample plate for interrogation with a Voyager-DE STR MALDI mass spectrometer. Oligonucleotides and NBz-oligo conjugates were desalted using ZipTipC4 devices and spotted with 3-hydroxypicolinic acid as the matrix. For positive-ion ESI-MS, pure NAB, Ni: NAB, NBzM, and Ni:NBzM were dissolved in pure water, neutralized with ammonium bicarbonate buffer (pH 8.0), and diluted 5-fold with neat spectroscopic-grade methanol. NMR. For NMR spectroscopy, approximately 22 mg of pure NBzM was dissolved in ∼800 µL of DMSO-d6. 1H and 13C NMR spectra were collected. Polyacrylamide Gel Electrophoresis. Precast TrisHCl Ready Gel mini-gels were run using a Mini-Protean II apparatus with a PowerPac 300 power supply at 200 V in 25 mM Tris, 192 mM glycine, and 0.1% SDS. For reducing conditions, protein samples were suspended and heat-denatured in 1× Laemmli buffer supplemented with 100 mM DTT; for nonreducing conditions no DTT was added. Prestained protein standards were either Precision- or Precision Plus-type from Bio-Rad. Carboxyfluorescein fluorescence from unstained gels was imaged, with a 2 s exposure, onto Polaroid type 667 film using a SYBR-Gold (Molecular Probes, Eugene, OR) gelatin filter with a Polaroid camera mounted over a UV transilluminator. Gels were stained using either GelCode Blue Stain or GelCode SilverSNAP. Preparation of DNA-Modified Microplate Wells. The wells of DNA-bind plates were modified with aminated oligonucleotides as per the manufacturer’s protocol. Oligonucleotides were used at a concentration of 200 pmol per well and were coupled over 3 h at 37 °C. Ethanolamine was used for blank wells and blocking. Plates were washed and stored in 50 mM sodium phosphate, pH 7.4, 150 mM NaCl and 0.02% sodium azide at 4 °C. Specificity of hybridization was confirmed with the fluorescent oligonucleotides FAM-sense and HEX-anti-noncomplementary. Hybridization of Protein-Oligo Conjugate on Microplate Wells. mDHFR-oligo conjugate and unmodified mDHFR were diluted with water and 2× plate hybridization buffer (1.2 M NaCl, 120 mM sodium citrate, 0.02% Tween-20, 20 mM EDTA, 1 mg/mL Ac-BSA, 0.2 mg/mL Herring Sperm DNA, 100 mM TrisCl (pH 7.5)) to 500 nM in 1× plate hybridization buffer. Specificity of hybridization was confirmed by using the fluorescent oligonucleotides FAM-sense and HEX-anti-noncomplementary similarly diluted to 1 µM. Three separate 100

Bioconjugate Chem., Vol. 15, No. 5, 2004 973

µL volumes of each sample type were dispensed into microplate wells modified with either 5′-aminated complementary, noncomplementary, or no capture oligonucleotide. On-plate hybridization occurred overnight at 37 °C in a shaking-platform incubator. After hybridization, supernatants were removed and saved, and the microwells were washed once with 200 µL of plate wash buffer (300 mM NaCl, 30 mM sodium citrate, 0.01% Tween-20, 50 mM TrisCl (pH 7.5)), again with 150 µL of the same buffer, and finally with 150 µL of 50 mM TrisCl (pH 7.5), 150 mM NaCl, 0.01% Tween-20. Each wash step was at 37 °C for 5 min. After washing, the microwells were filled with 100 µL of 50 mM sodium phosphate (pH 7.0), 100 mM NaCl, and the fluorescence was measured at room temperature with a Cytofluor 4000 fluorescence microplate reader. FAM fluorescence was measured using a 485/20 excitation filter and a 530/25 emission filter. HEX fluorescence was measured using a 530/25 excitation filter and a 620/40 emission filter. Preparation of DNA-Modified Microspheres. Polybead carboxylate microspheres (4.34 ( 0.24 µm diameter) were modified with 5′-aminated oligonucleotides by one-step coupling as described by Weimer and colleagues (44). Briefly, ∼81 mg of microspheres were washed twice with 10 mM NaOH, three times with Millipure water, three times with freshly prepared 100 mM MES (pH 4.7), and divided equally into three microcentrifuge tubes. After removing excess buffer, each of the ∼27 mg portions of microspheres was resuspended to 600 µL with 100 mM 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (EDAC) in 100 mM MES (pH 4.7). At this point, 100 µL of 1 mM aminated oligonucleotide in water, or for blank microspheres, 100 mM ethanolamine in 100 mM MES (pH 4.7), was added to each portion of microspheres. The microfuge tubes of microspheres were then incubated in the dark, at room temperature, with mixing by rotation for ∼72 h. This was followed by a 1 h exposure to 10 mM ethanolamine in 50 mM sodium phosphate, 1 mM EDTA (pH 8.5). The microspheres were washed 4 times with storage buffer (50 mM TrisCl, pH 7.5, 50 mM NaCl, 10 mM EDTA, 0.01% Tween-20, 0.02% sodium azide) and stored in the same buffer at ∼25 mg/mL at 4 °C. The beads were characterized for their binding capacity by hybridization with fluorescently labeled complementary oligonucleotides and measurement of the residual fluorescence left in solution. Hybridization of Protein-Oligo Conjugate on Microspheres. For mDHFR-oligoFAM conjugate hybridization experiments, three separate 50 µL aliquots (∼1.25 mg) of each type of microsphere (bearing complementary, noncomplementary, or no DNA) were pelleted by centrifugation at 14 000 rpm at 4 °C in an Eppendorf 5415C microcentrifuge, resuspended, and incubated for 30 min with mixing, at room temperature in 100 µL of prehybridization solution (22.5 mM NaCl, 25 mM TrisCl (pH 7.5), 10 mM sodium phosphate, 2.5 mM EDTA, 0.01% Tween-20, 10% glycerol, 1 mg/mL acetylated bovine serum albumin (Ac-BSA)). Following repelleting of the sample, and removal of 40 µL of supernatant, the microspheres were resuspended with 20.4 µL of BSA buffer (25 mM NaCl, 50 mM TrisCl (pH 7.5), 5 mM EDTA, 0.02% Tween-20, 2 mg/mL Ac-BSA). To these suspensions was added 20.4 µL of mDHFR-oligo conjugate (100 pmol). Parallel incubations of conjugate without microspheres served as controls. Hybridization occurred at room temperature with mixing by rotation over 20 min. After hybridization and repelleting, supernatants (∼40 µL) were removed carefully to avoid drying any of the microspheres. The microspheres were washed once

974 Bioconjugate Chem., Vol. 15, No. 5, 2004 Scheme 1

for 5 min at room temperature with 100 µL of prehybridization solution, repelleted, and resuspended in 40 µL of fresh prehybridization solution after removal of the wash solution. The resultant ∼50 µL suspensions of microspheres, along with 50 µL samples of supernatants, wash solutions, and control reactions, were pipetted into separate wells of a flat-bottom, untreated, 96-well microplate for fluorescence measurement and mDHFR activity assay with a Cytofluor 4000 fluorescence microplate reader. FAM fluorescence was measured using a 485/20 excitation filter and a 530/25 emission filter. Assay of mDHFR Activity on Immobilized on Microspheres. All measurements were obtained at room temperature. After recording the FAM fluorescence, 50 µL of reaction buffer (100 µM β-NADPH, 100 µM DHFA, 20 mM β-mercaptoethanol in prehybridization solution) was added to the microplate wells. The final concentration of both substrate (DHFA) and cofactor (β-NADPH) was 50 µM, well above their sub-micromolar Km values (45). Mixing of microspheres and substrates for 30 s every cycle was achieved using the instrument’s built-in agitation mechanism. The fluorescence due to β-NADPH was periodically measured from each well at a frequency of once every 97 s (46). Fluorescence was measured using a 360/40 excitation filter and a 460/40 emission filter. Fluorescence measurements were averaged between three wells and background corrected for total fluorescence at t ) 8.1 min. Linear fitting was error weighted and calculated using the program Origin 7.0 (OriginLab Corp., Northampton, MA). RESULTS

The goal of this work was to develop reagents that can covalently label recombinant proteins selectively in solution. The strategy, involving a trifunctional labeling reagent that recognizes the His-tag, is illustrated in Scheme 1. First, the labeling reagent noncovalently coordinates to the His-tag on the protein through its Ni: NTA group. After this, long-wavelength UV light is used

Meredith et al.

to create a covalent bond between the juxtaposed photoreactive group and protein surface. Affinity-driven docking of the labeling reagent ensures that the covalent bond is made proximal to the His-tag. Principal reaction products are expected to be selectively monofunctionalized proteins that retain native activity. Design and Synthesis of NAB. To test whether a protein’s His-tag could serve as a recognition element to guide a subsequent photoconjugation reaction, a trifunctional labeling reagent was synthesized in one step from commercially available precursors (Scheme 2). It is designated NAB since its three branches end in an NTA His-tag targeting group, a phenyl Azide photoreactive group, and a Biotin functional group. Electrospray ionization mass spectrometry (ESI-MS) of RP-HPLC purified NAB demonstrated masses at 925.3 and 947.3 corresponding to the protonated, [M + H]+, and sodiated, [M + Na]+, molecular ions. High-resolution ESI-MS of NAB following incubation with an excess of Ni(II) at pH 8.0 yielded a mass of 981.2538 as compared to an expected mass of 981.2567, corresponding to [(M - 2H + Ni) + H]+ molecular ions; thus each NAB molecule can coordinate a Ni(II) ion to form a 1:1 complex. The molar extinction coefficient at 270 nm, 270, of NAB in water is ∼26 900 M-1 cm-1 based on a quantitative assay of D-biotin. Labeling of mDHFR with NAB and Characterization. NAB coordinated to Ni(II) was assayed for its capacity to selectively biotinylate His-tagged proteins. mDHFR was chosen as a test enzyme for modification by NAB. This small, ∼23 kDa, monomeric enzyme is equally active when expressed with a His-tag at either its N-terminus or C-terminus and it is easily assayed in solution. A crystallographically determined model of the atomic structure of the highly homologous (87% sequence identity) human version of this enzyme shows that both termini occur at the surface of the molecule and are remote from its active-site (47). Thus, we reason that targeted modification of mDHFR guided by a His-tag at either the N- or C-terminus will leave the enzyme active. Upon exposure to UV light, reactions of NAB with either N- or C-terminally His-tagged mDHFR (His6mDHFR and mDHFR-His6, respectively) produce singly and doubly modified enzyme molecules. Under identical conditions, NAB fails to label untagged mDHFR. NABmodified proteins could be detected by a small shift in mobility on nonreducing SDS-PAGE gels (data not shown). The covalent attachment of biotin to mDHFR could also be detected on Western blots (data not shown). Matrix-assisted laser desorption and ionization mass spectrometry (MALDI-MS) confirmed the presence of singly and doubly labeled species (Figure 1). Since a Histag of six consecutive histidines was used, the occurrence of doubly modified mDHFR molecules may have resulted from the simultaneous or sequential coordination of two NAB molecules on a single His-tag; only two appropriately positioned histidine residues are necessary to coordinate a Ni:NTA moiety. We intend to test whether shorter His-tags change the distribution of modified products in future studies. Characterization of NAB-Labeled mDHFR. NABlabeled mDHFR remained active and its covalently attached biotin group was accessible by avidin. During purification by biotin-affinity chromatography, covalently NAB-labeled mDHFR remained immobilized on avidincoated particles in the presence of 200 mM imidazole, a concentration sufficient to disrupt the noncovalent NTA: Ni(II):His-tag interaction. Elution of NAB-labeled mDHFR from the avidin-coated support yielded mixed popu-

Targeted Protein Functionalization

Bioconjugate Chem., Vol. 15, No. 5, 2004 975

Scheme 2

Figure 1. MALDI mass spectrometry of NAB-modified mDHFRs. Depicted are linear-mode MALDI-TOF spectra of both Nand C-terminally His-tagged versions of mDHFR prior to labeling with NAB (top), after labeling but before purification (middle) and after affinity purification of biotinylated species (bottom). Discrete peaks corresponding to singly and doubly modified mDHFR molecules are clearly present in both reactions demonstrating that labeling can be directed through a His-tag on either terminus of this enzyme. Affinity purification yields samples of almost entirely modified enzyme. The specific activity of both preparations remained close to that of the unmodified enzyme (see text).

lations of singly and doubly modified protein molecules containing very little (99.99% of the conjugates will be coordinated to Ni(II) at equilibrium (63).

Targeted Protein Functionalization

Figure 2. NBz-oligo labeling of mDHFR is selective for the presence of the His-tag. Hybrid molecules of covalently linked protein and DNA, the products of photoconjugation between NBz-oligo and His-tagged mDHFR in solution, are resolved as a discrete species on reducing and denaturing 12% polyacrylamide gels by SDS-PAGE. Shown are lanes from the same silverstained gel (irrelevant lanes have been excised) where labeling reactions (+) between C- and N-terminally His-tagged mDHFRs are separated parallel to control reactions lacking NBz-oligo. The final two lanes are from reactions with mDHFR lacking a His-tag. The middle arrow, marked with an asterisk, indicates the position of the shifted species with a mobility equivalent to that of a 35 kDa protein. In similar experiments in which the oligo carries a terminal fluorophore, the fluorescence of this shifted band can be detected on unstained gels. The lower arrow indicates the mobility of the unmodified mDHFR molecules, with molecular masses ranging from 23.3 kDa for the C-tagged species to 24.7 kDa for the N-tagged species. All mDHFR preparations contain, to a variable extent, the ∼50 kDa protein (small arrow) as a contaminant.

Ni:NBz-Oligo Labeling of mDHFRs. NBz-oligo coordinated to Ni(II) was assayed for its capacity to label mDHFR. Reactions of Ni:NBz-oligo with both His6mDHFR and mDHFR-His6 yielded protein molecules linked to a single oligo, while reactions with mDHFR lacking a His-tag showed no evidence of modification (Figure 2). Reactions between Ni:NBz-oligo and Histagged mDHFR were significantly inhibited by the presence of either 150 mM imidazole or 5 mM EDTA (data not shown). Oligo-modified proteins could be detected by a significant shift in mobility on reducing SDS-PAGE gels (Figure 2). In all such labeling reactions, the NBzoligo carried an additional label, either biotin or 6-carboxyfluorescein (FAM), linked to the end of the oligo opposite the NBz-group. When NBz-oligobiotin was used, biotin on the shifted protein was detected on Western blots (data not shown) and when NBz-oligoFAM was used, the fluorescence of the shifted protein was observed directly in the unstained gel (inset, Figure 3). No suitable conditions could be found to measure the molecular weight of these oligo-protein molecules by mass spectrometry. Characterization of Oligo-Labeled mDHFR. Preparative-scale conjugation reactions were conducted between mDHFR-His6 and either NBz-oligobiotin or NBzoligoFAM. Although the incorporation of biotin permitted significant enrichment of mDHFR-oligo conjugates over unmodified mDHFR through biotin affinity chromato-

Bioconjugate Chem., Vol. 15, No. 5, 2004 977

Figure 3. Size-exclusion chromatography allows purification of NBz-oligo-modified mDHFR. A 100 nmol scale reaction between C-terminally His-tagged mDHFR and the fluorescently labeled modification reagent NBz-oligoFAM was separated by high-resolution size exclusion chromatography. The elution profile is shown as a function of absorbance at 280 nm, a wavelength absorbed by protein and DNA. Collected fractions across peaks a and b were pooled separately and concentrated by ultrafiltration. The inset shows Coomassie-blue protein staining (lanes 2, 3, and 5) of reducing SDS-PAGE separations of unmodified C-tagged DHFR and the concentrated components of peaks a and b, respectively. Lanes 4 and 6, fluorescent images of the lanes 3 and 5, respectively, show the fluorescent species visible prior to staining. All lanes are from the same gel. Peak a is highly enriched with a fluorescent protein of high molecular weight (lanes 3 and 4): the NBz-oligoFAM-mDHFR conjugate, indicated by the uppermost arrow labeled with an asterisk. Peak b appears to be a mixture of unmodified mDHFR (lane 5, middle arrow) and unattached fluorescent NBz-oligo (lane 6, lowest arrow). When assayed for mDHFR activity, peak a demonstrated a specific activity of about 80% that of unmodified mDHFR, while peak b had a specific activity equivalent to that of unmodified mDHFR.

graphy, size-exclusion methods permitted purification to near homogeneity (Figure 3). Under the conditions tested, NBz-oligo labeling yields were low, from 1 to 2%. In solution, purified mDHFR-oligo conjugates exhibited on average 72% (n ) 4 preparations, ranging from 56 to 81%) of the specific activity of unmodified mDHFR indicating that the covalent attachment of oligo and subsequent manipulations did not greatly reduce the enzyme’s native activity. When mDHFR-oligo conjugate molecules were incubated with either complementary or noncomplementary fluorescently labeled oligos in solution, analytical HPLC gel-filtration experiments showed the presence of new high-mobility species only with the complementary DNA (data not shown). This suggests that the mDHFR-oligo conjugate molecules could specifically hybridize with a complementary DNA oligonucleotide in solution. Under identical conditions, there was no evidence of hybridization to either of two different noncomplementary oligonucleotides. To test whether the protein-oligo conjugates can be used to make self-assembling protein arrays their capacity to specifically hybridize to DNA-coated surfaces was assayed. For these experiments, the DNA portion of the conjugate was 28 bases long and linked to the mDHFR through its 5′ end. The 3′ end of the DNA consisted of a 15-base hybridization sequence termed “sense”, end-

978 Bioconjugate Chem., Vol. 15, No. 5, 2004

Figure 4. Scale cartoon of mDHFR-oligo conjugate used in hybridization experiments. The mDHFR-oligoFAM conjugate and its capture partners are depicted as fully extended molecules where the DNA segments have the length expected for duplex B-DNA. In the hybridization/immobilization experiments described, the conjugate is presented with either a complementary or noncomplementary capture strand that has been affixed through its 5′-end to a substrate. A 9-base spacer of thymidine residues links the capture sequences to the substrate to minimize crowding and to accommodate the fluorescent endlabel of the conjugate. The nonspacer 15-base capture sequences are shown as shapes that either match or fail to match the shape of the conjugate 15-base “sense” sequence.

labeled with a FAM fluorescent group (Figure 4). This design was chosen to allow the sense sequence to extend well away from the protein surface, to minimize selfcomplementarity, and to allow the conjugates to be directly detected by fluorescence measurements. To prepare surfaces that display capture DNA in the best orientation for hybridization, synthetic 5′-aminated oligos were coupled through primary amine-specific chemistry. Following the 5′-terminal amine, the capture oligo sequences consisted of nine deoxythymidine residues, which serve as an economical spacer sequence, followed by 15 bases that were complementary (5′-TAG AGA TGG ACA GGT-3′) or noncomplementary (5′-CCC TAA GAA GAC GGA-3′). To preserve the structure and activity of the mDHFR, the capture sequences were chosen to allow hybridization to their respective complementary sequences at room temperature at low ionic strength. The sequences were also selected to minimize self- and crosscomplementarity. Specific hybridization of mDHFR-oligo conjugates could be observed on both oligo-modified flat plastic microwell surfaces and polymer microspheres. On microplate, noncomplementary wells captured only 12.1% as much conjugate, as measured by fluorescence, as did complementary ones (n ) 3). Similarly, noncomplementary microspheres retained only 1.4% as much FAM fluorescence as compared to microspheres coated with complementary capture oligo (n ) 3). Precise determination of the changes in fluorescence due to interaction with the microspheres is not within the scope of the current work. To measure the enzymatic activity of the immobilized conjugates, the amount of mDHFR activity retained on microspheres after hybridization and washing was measured. In the assay, oxidation of the cofactor β-NADPH to β-NADP+, which is coupled to the reduction of dihy-

Meredith et al.

Figure 5. Schematic diagram of experiment to measure the enzyme activity of specifically immobilized conjugates. (A) The various microsphere types, either blank (non-DNA bearing), noncomplementary, or complementary beads are schematically shown in separate microplate wells under buffer, after having been incubated with mDHFR-oligoFAM conjugate (see Figure 4) and washed. The conjugate fluorescence is measured at this point. (B) A concentrated mixture of the substrate, DHFA, and cofactor β-NADPH are added to the wells. The enzyme mDHFR couples the oxidation of β-NADPH to NADP+ with the reduction of DHFA to THFA.

Figure 6. mDHFR immobilized through sequence-specific hybridization shows significant activity. Measurements of β-NADPH fluorescence were recorded as described in the methods and depicted in Figure 5. Filled squares, 9, correspond to complementary DNA-coated microspheres after hybridization to mDHFR-oligoFAM conjugate. Open circles, O, correspond to noncomplementary DNA-coated microspheres after identical treatment. Open triangles, 4, correspond to uncoated microspheres treated likewise. Linear least-squares fitting yielded the respective solid, dashed, and dotted lines. Subtracting the slope of the line for the unmodified beads from the slopes of the lines of the DNAcoated beads indicates that the activity difference due to sequence specificity is approximately 10-fold. Each datapoint is the average of measurements from three separate microwells. Error bars correspond to standard error of the mean. The mean background fluorescence of the first time-point used (t ) 8.1 min after addition of substrates) was subtracted from all points for each dataset. Prior to the first time point used, there was significant scatter, most likely due to settling of the beads and inefficient mixing within the microplate reader.

drofolate (DHFA) to tetrahydrofolate (THFA), is monitored as a decrease in the fluorescence attributable to β-NADPH. The assay is depicted schematically in Figure 5. The amount of conjugate retained on the microspheres after hybridization was determined by measuring the FAM fluorescence of the supernatants and wash solutions

Targeted Protein Functionalization

Bioconjugate Chem., Vol. 15, No. 5, 2004 979

Table 1. Quantitative Results from Microsphere-Immobilized mDHFR-Oligo Conjugates

microsphere type complementary noncomplementary

calculated apparent measured quantity mass equivalent quantity of of mDHFRactivity, relative mDHFR-oligo oligo retained on to nonimmobilremaining in microspheres ized conjugate solution (pmol) (pmol) (pmol) 58.5 100

41.5 0

13.9 1.4

and comparing this to the fluorescence of known amounts of conjugate in the same buffer solution (Table 1). This was done because binding to the beads clearly altered the fluorescence of the conjugate molecules. By this measure, aliquots of 1.25 mg of complementary DNAmodified microspheres retained an average of 41.5 pmol of mDHFR-oligo conjugate, whereas equivalent aliquots of noncomplementary DNA-modified and unmodified microspheres failed to retain quantifiable amounts of conjugate. This value closely matches estimates of the binding capacity of the microspheres, 28-40 pmol/mg, as determined with fluorescent complementary oligonucleotides. Addition of the substrate DHFA and the cofactor β-NADPH at concentrations well above their respective Km values allowed the activity of the immobilized mDHFR to be measured over time (Figure 6). With this assay, microspheres carrying specifically immobilized mDHFR showed approximately 10-fold greater activity than control microspheres (Table 1). The activity of the specifically immobilized mDHFR-oligo conjugate is 33.5% of that determined for an equivalent amount of conjugate in solution; this decrease may be due to nonideal mixing, molecular crowding and steric effects, or limits to substrate diffusion to a surface. No effort was made to optimize the kinetic behavior of the immobilized conjugate, nor was the long-term stability or reusability of the immobilized conjugate assessed. DISCUSSION

This is the first report demonstrating the use of the His-tag as a guide for the covalent modification of a protein in solution. This approach has several advantages over conventional labeling methods in that (i) it utilizes the widely used His-tag, making it compatible with common expression systems and established libraries of His-tagged proteins, (ii) it presumably restricts the point of modification to the proximity of the His-tag, (iii) it restricts the stoichiometry of modification to yield largely 1:1 conjugates, (iv) the recognition element used, the Histag, can be located at either protein terminus and possibly at internal sites, (v) a wide spectrum of modifications, from small molecules to macromolecules, are possible. We have shown that when this strategy is used to link a DNA oligonucleotide to a model enzyme, the protein’s functionality remains intact both in solution and upon immobilization through sequence-specific hybridization. Thus, this technology may serve in the future creation of self-assembling, functional protein arrays. With regard to the attachment of nucleic acid to protein for the creation of protein arrays, it should be noted that others have made progress using alternative methods. Such a goal was first expressed by Niemeyer and coworkers who chemically linked oligonucleotides to streptavidin (64). The oligo-streptavidin conjugate molecules were subsequently coupled to functional biotinylated antibodies through the strong noncovalent streptavidinbiotin interaction. Using this approach, the authors demonstrated the “translation”, via hybridization, of a

microtiter plate-based DNA array into a functional antibody array. Recently, the same group used expressed protein ligation to produce covalent conjugates between polyamide nucleic acid (PNA) and Ras protein and between PNA and the Ras binding domain of the c-Raf-1 protein (59). The authors showed that these C-terminally linked PNA-protein conjugates could hybridize to convert a simple glass-slide based DNA microarray into a protein array with retention of specific protein binding properties; however, no immobilized catalytic activity was demonstrated. Kuimelis and colleagues used mRNAprotein fusions to convert a DNA microarray chip to a simple protein microarray (57). The methodology described relies on in vitro translation and is restricted to fusions between the C-terminus of the protein and the 3′-end of its encoding mRNA. The array produced consisted of fragments of full-length proteins that were functional only in that they could be detected by specific antibodies. The approach we describe uses more readily produced biologically expressed full-length proteins, provides for modification in a variety of positions, and is compatible with many different sequences of the more chemically robust DNA or other nucleic acid analogues. Furthermore, the enzyme we immobilized was shown to retain catalytic activity. More recently, Oleinikov and co-workers also used cellfree biosynthesis to produce lysine-biotinylated proteins (58). A series of steps allowed these researchers to produce protein-streptavidin-oligo conjugates that could be used to produce a simple self-assembled protein array. Importantly, both the proteins used, enhanced green fluorescent protein and the enzyme luciferase, retained their described functions. This approach is promising but also depends on the products of in vitro translation reactions. The biotinylation results we describe using the NAB reagent offer an alternative to this cell-free approach that could be easily integrated with their subsequent steps. The random incorporation of biotinylatedlysines by Oleinikov and colleagues eliminates the possibility of controlling the orientation of the proteins within the array. With our strategy, one can imagine varying the position of the His-tag in a given protein to control its resultant orientation upon modification and immobilization. Since protein arrays are sought largely as tools for the screening of molecular interactions, orientational control may prove very important. Further, orientational control is considered important for optimal functionality of immobilized enzymes (reviewed in ref 65). There are a number of ways that our results may be extended and improved. Foremost, greater efficiency of the photoreaction is expected if the affinity of the labeling reagent for the His-tag were to be increased. Drawing a lesson from the results of Ebright and co-workers, the affinity can be expected to be increased 10-100 fold simply by incorporation of two appropriately spaced NTA groups into the labeling reagent (37). Chemistries that yield covalent bonds other than through photolysis may increase efficiency as well. Combinatorial methods may allow co-optimization of the labeling reagent and variations of the His-tag that improve both specificity and affinity. Mapping the exact locations of protein modification produced with these reagents should shed light on their degree of regioselectivity. Other affinity-tag-directed cross-linking reagents are envisioned, such as one that incorporates glutathione to target the commonly used glutathione-S-transferase (GST) fusion tag. Use of this oligo-labeling strategy against a battery of proteins will allow a true self-assembling protein array, either beador chip-based to be demonstrated.

980 Bioconjugate Chem., Vol. 15, No. 5, 2004

Of particular interest is the prospect of harnessing this molecular recognition based technology to achieve an unprecedented degree of control over the artificial assembly of molecular complexes. In recent years, it has become increasingly apparent that biological processes are mediated largely by exquisitely organized assemblies of functional macromolecules (66). Typically, such assemblies are multifunctional molecular machines capable of carrying out a series of physical and chemical transactions in a coordinated and efficient manner. Examples include the transcriptional apparatus that synthesizes RNA from genomic DNA in a highly regulated and processive manner, the ribosome which synthesizes proteins from mRNA also in a regulated and processive manner, and G-protein coupled receptor complexes which assemble and disassemble in response to signals at the cell-surface. In these systems, the identities, number, and relative orientations of the proteins are precisely defined, but by architectural rules that are currently poorly understood. The engineered assembly of proteins into higher-order multifunctional molecular machines requires a means to simplify these rules. Along these lines, significant progress in molecular assembly, programmed through the rules of nucleic acid interaction, has developed over the last two decades (see refs 67 and 68 for recent reviews). With this approach, remarkable molecular objects have been created from both DNA alone and protein-DNA conjugates. It is believed that the technology presented here may offer a predictable and practical means to modify recombinant proteins with precision sufficient to enable the design and assembly of novel multifunctional molecular machines held together and organized by structural elements such as nucleic acid. This or a similar approach, coupled with our everincreasing knowledge of protein sequences, structures, and functions may form the basis for new engineered semi-synthetic biomolecular materials, catalysts, and nanobiotechnological devices. ACKNOWLEDGMENT

This work was supported by NIH Grants F32NS11063 (G.D.M.) and GM57015 (N.L.A.). The authors thank John Greaves for help with mass spectrometry, Philip Dennison for NMR spectroscopy, Dennis Ta and Larry Vickery for help with molecular biology, J. Denis Heck for microarray support, Stephen Hou for help with beadbased fluorescence measurements, David van Vranken, Casey McComas, Josh Yoburn, and Frank Rossi for advice and help with organic synthesis, and Joseph Soughayer and Kwan Ng for helpful discussions. LITERATURE CITED (1) Jacob, C., Giles, G. L., Giles, N. M., and Sies, H. (2003) Sulfur and selenium: The role of oxidation state in protein structure and function. Angew. Chem., Int. Ed. 42, 47424758. (2) Dawson, P. E., Muir, T. W., Clark-Lewis, I., and Kent, S. B. H. (1994) Synthesis of Proteins by Native Chemical Ligation. Science 266, 776-779. (3) Muir, T. W., Sondhi, D., and Cole, P. A. (1998) Expressed protein ligation: A general method for protein engineering. Proc. Natl. Acad. Sci. U.S.A. 95, 6705-6710. (4) Cotton, G. J., Ayers, B., Xu, R., and Muir, T. W. (1999) Insertion of a Synthetic Peptide into a Recombinant Protein Framework: A Protein Biosensor. J. Am. Chem. Soc. 121, 1100-1101. (5) Becker, C. F. W., Hunter, C. L., Seidel, R. P., Kent, S. B. H., Goody, R. S., and Engelhard, M. (2001) A sensitive fluorescence monitor for the detection of activated Ras: total

Meredith et al. chemical synthesis of site-specifically labeled Ras binding domain of c-Raf1 immobilized on a surface. Chem. Biol. 8, 243-252. (6) Geoghegan, K. F., and Stroh, J. G. (1992) Site-Directed Conjugation of Nonpeptide Groups to Peptides and Proteins Via Periodate-Oxidation of a 2-Amino Alcohol-Application to Modification at N-Terminal Serine. Bioconjugate Chem. 3, 138-146. (7) Fancy, D. A., and Kodadek, T. (1997) Site-directed oxidative protein cross-linking. Tetrahedron 53, 11953-11960. (8) Gaertner, H. F., Rose, K., Cotton, R., Timms, D., Camble, R., and Offord, R. E. (1992) Construction of Protein Analogues by Site-Specific Condensation of Unprotected Fragments. Bioconjugate Chem. 3, 262-268. (9) van Hest, J. C. M., and Tirrell, D. A. (1998) Efficient introduction of alkene functionality into proteins in vivo. FEBS Lett. 428, 68-70. (10) Cornish, V. W., Hahn, K. M., and Schultz, P. G. (1996) Sitespecific protein modification using a ketone handle. J. Am. Chem. Soc. 118, 8150-8151. (11) Sharma, N., Furter, R., Kast, P., and Tirrell, D. A. (2000) Efficient introduction of aryl bromide functionality into proteins in vivo. FEBS Lett. 467, 37-40. (12) Johnsson, N., and Johnsson, K. (2003) A fusion of disciplines: Chemical approaches to exploit fusion proteins for functional genomics. ChemBioChem 4, 803-810. (13) Griffin, B. A., Adams, S. R., and Tsien, R. Y. (1998) Specific covalent labeling of recombinant protein molecules inside live cells. Science 281, 269-272. (14) Keppler, A., Gendreizig, S., Gronemeyer, T., Pick, H., Vogel, H., and Johnsson, K. (2003) A general method for the covalent labeling of fusion proteins with small molecules in vivo. Nat. Biotechnol. 21, 86-89. (15) Kindermann, M., George, N., Johnsson, N., and Johnsson, K. (2003) Covalent and selective immobilization of fusion proteins. J. Am. Chem. Soc. 125, 7810-7811. (16) Hochuli, E., Dobeli, H., and Schacher, A. (1987) New Metal Chelate Adsorbent Selective for Proteins and Peptides Containing Neighboring Histidine-Residues. J. Chromatogr. 411, 177-184. (17) Hochuli, E., Bannwarth, W., Dobeli, H., Gentz, R., and Stuber, D. (1988) Genetic Approach to Facilitate Purification of Recombinant Proteins with a Novel Metal Chelate Adsorbent. Bio/Technol. 6, 1321-1325. (18) Lu, Z. J., DiBlasio-Smith, E. A., Grant, K. L., Warne, N. W., LaVallie, E. R., Collinsracie, L. A., Follettie, M. T., Williams, M. J., and McCoy, J. M. (1996) Histidine patch thioredoxins-mutant forms of thioredoxin with metal chelating affinity that provide for convenient purifications of thioredoxin fusion proteins. J. Biol. Chem. 271, 5059-5065. (19) Lin, J. T., Kormanec, J., Homerova, D., and Kinne, R. K. H. (1999) Probing transmembrane topology of the highaffinity sodium/glucose cotransporter (SGLT1) with histidinetagged mutants. J. Membr. Biol. 170, 243-252. (20) Voss, J., Salwinski, L., Kaback, H. R., and Hubbell, W. L. (1995) A method for distance determination in proteins using a designed metal ion binding site and site-directed spin labeling: Evaluation with T4 lysozyme. Proc. Natl. Acad. Sci. U.S.A. 92, 12295-12299. (21) Bailey, J., and Manoil, C. (2002) Genome-wide internal tagging of bacterial exported proteins. Nat. Biotechnol. 20, 839-842. (22) BD HAT Protein Expression and Purification System User Manual, pp 1-31, Clontech, 2002. (23) Xu, Z., and Lee, S. Y. (1999) Display of Polyhistidine Peptides on the Escherichia coli Cell Surface by Using Outer Membrane Protein C as an Anchoring Motif. Appl. Environ. Microbiol. 65, 5142-5147. (24) McMahan, S. A., and Burgess, R. R. (1996) Single-step synthesis and characterization of biotinylated nitrilotriacetic

Targeted Protein Functionalization acid, a unique reagent for the detection of histidine-tagged proteins immobilized on nitrocellulose. Anal. Biochem 236, 101-106. (25) Kubalek, E. W., LeGrice, S. F. J., and Brown, P. O. (1994) Two-Dimensional Crystallization of Histidine-Tagged, HIV-1 Reverse Transcriptase Promoted by a Novel Nickel-Chelating Lipid. J. Struct. Biol. 113, 117-123. (26) O’Shannessy, D. J., O’Donnell, K. C., Martin, J., and Brigham-Burke, M. (1995) Detection and Quantification of Hexa-Histidine-Tagged Recombinant Proteins on Western Blots and by a Surface Plasmon Resonance Biosensor Technique. Anal. Biochem 229, 119-124. (27) Sigal, G. B., Bamdad, C., Barberis, A., Strominger, J., and Whitesides, G. M. (1996) A Self-Assembled Monolayer for the Binding and Study of Histidine-Tagged Proteins by Surface Plasmon Resonance. Anal. Chem 68, 490-497. (28) Schmitt, L., Ludwig, M., Gaub, H. E., and Tampe, R. (2000) A Metal-Chelating Microscopy Tip as a New Toolbox for Single-Molecule Experiments by Atomic Force Microscopy. Biophys. J. 78, 3275-3285. (29) Christendat, D., Yee, A., Dharamsi, A., Kluger, Y., Gerstein, M., Arrowsmith, C. H., and Edwards, A. M. (2000) Structural proteomics: prospects for high throughput sample preparation. Prog. Biophys. Mol. Biol. 73, 339-345. (30) Heinemann, U., Frevert, J., Hofmann, K. P., Illing, G., Maurer, C., Oschkinat, H., and Saenger, W. (2000) An integrated approach to structural genomics. Prog. Biophys. Mol. Biol. 73, 347-362. (31) Bussow, K., Nordhoff, E., Lubbert, C., Lehrach, H., and Walter, G. (2000) A human cDNA library for high-throughput protein expression screening. Genomics 65, 1-8. (32) Zhu, H., Bilgin, M., Bangham, R., Hall, D., Casamayor, A., Bertone, P., Lan, N., Jansen, R., Bidlingmaier, S., Houfek, T., Mitchell, T., Miller, P., Dean, R. A., Gerstein, M., and Snyder, M. (2001) Global analysis of protein activities using proteome chips. Science 293, 2101-2105. (33) Heyman, J. A., Cornthwaite, J., Foncerrada, L., Gilmore, J. R., Gontang, E., Hartman, K. J., Hernandez, C. L., Hood, R., Hull, H. M., Lee, W. Y., Marcil, R., Marsh, E. J., Mudd, K. M., Patino, M. J., Purcell, T. J., Rowland, J. J., Sindici, M. L., and Hoeffler, J. P. (1999) Genome-scale cloning and expression of individual open reading frames using topoisomerase I-mediated ligation. Genome. Res. 9, 383392. (34) Mateo, C., Fernandez-Lorente, G., Cortes, E., Garcia, J. L., Fernandez-Lafuente, R., and Guisan, J. M. (2001) Onestep purification, covalent immobilization, and additional stabilization of poly-His-tagged proteins using novel heterofunctional chelate-epoxy supports. Biotech. Bioeng. 76, 269276. (35) Melton, R. G., and Sherwood, R. F. (1996) Antibody-enzyme conjugates for cancer therapy. J. Natl. Cancer Inst. 88, 153165. (36) Hainfeld, J. F., Liu, W. Q., Halsey, C. M. R., Freimuth, P., and Powell, R. D. (1999) Ni-NTA-gold clusters target histagged proteins. J. Struct. Biol. 127, 185-198. (37) Kapanidis, A. N., Ebright, Y. W., and Ebright, R. H. (2001) Site-specific incorporation of fluorescent probes into protein: Hexahistidine-tag-mediated fluorescent labeling with (Ni2+: Nitrilotriacetic acid)(n)-fluorochrome conjugates. J. Am. Chem. Soc. 123, 12123-12125. (38) Brenner, S., and Lerner, R. A. (1992) Encoded Combinatorial Chemistry. Proc. Natl. Acad. Sci. U.S.A. 89, 53815383. (39) Hillcoat, B. L., and Blakely, R. L. (1966) Dihydrofolate Reductase of Streptococcus faecalis. J. Biol. Chem. 241, 29953001. (40) Becker, C. R., Efcavitch, J. W., Heiner, C. R., and Kaiser, N. F. (1985) Use of a C-4 Column for Reversed-Phase HighPerformance Liquid-Chromatographic Purification of Synthetic Oligonucleotides. J. Chromatogr. 326, 293-299. (41) Bosca, F., and Miranda, M. A. (1998) Photosensitizing drugs containing the benzophenone chromophore. J. Photochem. Photobiol. B 43, 1-26.

Bioconjugate Chem., Vol. 15, No. 5, 2004 981 (42) Savage, M. D., Mattson, G., Desai, S., Nielander, G. W., Morgensen, S., and Conlin, E. J. (1996) Avidin-Biotin Chemistry: A Handbook, Pierce Chemical Co., Rockford, IL. (43) Pace, C. N., Vajdos, F., Fee, L., Grimsley, G., and Gray, T. (1995) How to measure and predict the molar absorption coefficient of a protein. Protein Sci. 4, 2411-2423. (44) Walsh, M. K., Wang, X., and Weimer, B. C. (2001) Optimizing the immobilization of single-stranded DNA onto glass beads. J. Biochem. Biophys. Methods 47, 221231. (45) Blakely, R. L. (1995) Eukaryotic Dihydrofolate Reductase. Adv. Enzymol. 70, 23-102. (46) Held, P. (2001) Determination of NADH or NADPH Concentrations with the FL600 using Fluorescence or Absorbance Modes. Bio-Tek Application Note, http://www.biotek.com/products/tech_res_detail.php?id)57. (47) Lewis, W. S., Cody, V., Galitsky, N., Luft, J. R., Pangborn, W., Chunduru, S. K., Spencer, H. T., Appleman, J. R., and Blakley, R. L. (1995) Methotrexate-resistant variants of human dihydrofolate reductase with substitutions of leucine 22-kinetics, crystallography, and potential as selectable markers. J. Biol. Chem. 270, 5057-5064. (48) Hermanson, G. T. (1996) Bioconjugate Techniques, Academic Press, San Diego. (49) Dorman, G., and Prestwich, G. D. (1994) Benzophenone photophores in biochemistry. Biochemistry 33, 56615673. (50) Corey, D. R., and Schultz, P. G. (1987) Generation of a Hybrid Sequence-Specific Single-Stranded Deoxyribonuclease. Science 238, 1401-1403. (51) Ghosh, S. S., Kao, P. M., and Kwoh, D. Y. (1989) Synthesis of 5′-Oligonucleotide Hydrazide Derivatives and Their Use in Preparation of Enzyme Nucleic-Acid Hybridization Probes. Anal. Biochem 178, 43-51. (52) Roberts, R. W., and Szostak, J. W. (1997) RNA-peptide fusions for the in vitro selection of peptides and proteins. Proc. Natl. Acad. Sci. U.S.A. 94, 12297-12302. (53) Doi, N., and Yanagawa, H. (1999) STABLE: protein-DNA fusion system for screening of combinatorial protein libraries in vitro. FEBS Lett. 457, 227-230. (54) Tighe, H., Takabayashi, K., Schwartz, D., Marsden, R., Beck, L., Corbeil, J., Richman, D. D., Eiden, J. J., Spiegelberg, H. L., and Raz, E. (2000) Conjugation of protein to immunostimulatory DNA results in a rapid, long-lasting and potent induction of cell-mediated and humoral immunity. Eur. J. Immunol. 30, 1939-1947. (55) Rajur, S. B., Roth, C. M., Morgan, J. R., and Yarmush, M. L. (1997) Covalent protein-oligonucleotide conjugates for efficient delivery of antisense molecules. Bioconjugate Chem. 8, 935-940. (56) Kurz, M., Gu, K., Al-Gawari, A., and Lohse, P. A. (2001) cDNA-protein fusions: covalent protein-gene conjugates for the in vitro selection of peptides and proteins. ChemBioChem 2, 666-672. (57) Weng, S., Gu, K., Hammond, P. W., Lohse, P., Rise, C., Wagner, R. W., Wright, M. C., and Kuimelis, R. G. (2002) Generating addressable protein microarrays with PROfusion (TM) covalent mRNA-protein fusion technology. Proteomics 2, 48-57. (58) Oleinikov, A. V., Gray, M. D., Zhao, J., Montgomery, D. D., Ghindills, A. L., and Dill, K. (2003) Self-assembling protein arrays using electronic semiconductor microchips and in vitro translation. J. Proteome Res. 2, 313-319. (59) Lovrinovic, M., Seidel, R., Wacker, R., Schroeder, H., Seitz, O., Engelhard, M., Goody, R. S., and Niemeyer, C. M. (2003) Synthesis of protein-nucleic acid conjugates by expressed protein ligation. Chem. Commun. 2003, 822-823. (60) Shih, H. C., Kassahun, H., Burrows, C. J., and Rokita, S. E. (1999) Selective association between a macrocyclic nickel complex and extrahelical guanine residues. Biochemistry 38, 15034-15042. (61) Hansma, H. G., Pietrasanta, L. I., Golan, R., Sitko, J. C., Viani, M. B., Paloczi, G. T., Smith, B. L., Thrower, D., and

982 Bioconjugate Chem., Vol. 15, No. 5, 2004 Hansma, P. K. (2000) Recent highlights from atomic force microscopy of DNA. J. Biomol. Struct. Dyn. 271-275. (62) Abrescia, N. G. A., Huynh-Dinh, T., and Subirana, J. A. (2002) Nickel-guanine interactions in DNA: crystal structure of nickel-d. J. Biol. Inorg. Chem. 7, 195-199. (63) Dawson, R. M. C., Elliot, D. C., Elliot, W. H., and Jones, K. M. (1986) Data for Biochemical Research, Clarendon Press, Oxford. (64) Niemeyer, C. M., Sano, T., Smith, C. L., and Cantor, C. R. (1994) Oligonucleotide-directed self-assembly of proteins: semisynthetic DNA-streptavidin hybrid molecules as connectors for the generation of macroscopic arrays and the construction of supramolecular bioconjugates. Nucleic Acids. Res. 22, 5530-5539.

Meredith et al. (65) Turkova, J. (1999) Oriented immobilization of biologically active proteins as a tool for revealing protein interactions and function. J. Chromatogr. B 722, 11-31. (66) Baumeister, W., and Steven, A. C. (2000) Macromolecular electron microscopy in the era of structural genomics. Trends Biol. Sci. 25, 624-631. (67) Seeman, N. (2003) At the Crossroads of Chemistry, Biology, and Materials: Structural DNA Nanotechnology. Chem. Biol. 10, 1151-1159. (68) Niemeyer, C. M., and Adler, M. (2002) Nanomechanical Devices Based on DNA. Angew Chem, Int. Ed. 41, 37793783.

BC0498929