Recombinant Gaussia Luciferase. Overexpression ... - ACS Publications

Johanna M Niers, John W Chen, Grant Lewandrowski, Mariam Kerami, Elisabeth ... Nikolaos C. Deliolanis, Lisa Pike, Johanna M. Niers, Lee-Ann Tjon-Kon-F...
0 downloads 0 Views 141KB Size
Download PDF
Anal. Chem. 2002, 74, 4378-4385

Recombinant Gaussia Luciferase. Overexpression, Purification, and Analytical Application of a Bioluminescent Reporter for DNA Hybridization Monique Verhaegen† and Theodore K. Christopoulos*,‡

Department of Chemistry and Biochemistry, University of Windsor, Windsor, Ontario, Canada N9B 3P4, and Department of Chemistry, University of Patras, Patras, Greece 26500

The cDNA for Gaussia luciferase (GLuc), the enzyme responsible for the bioluminescent reaction of the marine copepod Gaussia princeps, has been cloned recently. GLuc (MW ) 19 900) catalyzes the oxidative decarboxylation of coelenterazine to produce coelenteramide and light. We report the first quantitative analytical study of GLuc and examine its potential as a new reporter for DNA hybridization. A plasmid encoding both a biotin acceptor peptide-GLuc fusion protein as well as the enzyme biotin protein ligase (BPL) is engineered by using GLuc cDNA as a starting template. BPL catalyzes the covalent attachment of a single biotin to the fusion protein in vivo. Purification of GLuc is then accomplished by affinity chromatography using immobilized monomeric avidin. Moreover, the in vivo biotinylation enables subsequent complexation of GLuc with streptavidin (SA), thereby avoiding chemical conjugation reactions that are known to inactivate luciferases. Purified GLuc can be detected down to 1 amol with a signal-to-background ratio of 2 and a linear range extending over 5 orders of magnitude. The background luminescence of coelenterazine is the main limiting factor for even higher detectability of GLuc. Furthermore, the GLuc-SA complex is used as a detection reagent in a microtiter well-based DNA hybridization assay. The analytical range extends from 1.6 to 800 pmol/L of target DNA. Biotinylated GLuc produced from 1 L of bacterial culture is sufficient for 150 000 hybridization assays. Bio- and chemiluminescence, the emission of light from chemically generated excited states, find rapidly expanding and diverse analytical applications in areas requiring the determination of analytes in low concentrations.1 The principal advantage of bio(chemi)luminometric assays is their superior detectability over conventional spectrophotometric and fluorometric ones. Because bio(chemi)luminescence does not require the use of excitation light, problems arising from scattering of excitation radiation, * To whom correspondence should be addressed. Tel: (+30) 610 997130. Fax: (+30) 610 997118.. E-mail: [email protected]. † University of Windsor. ‡ University of Patras. (1) Roda, A.; Pasini, P.; Guardigli, M.; Baraldini, M.; Musiani, M.; Mirasoli, M. Fresenius J. Anal. Chem. 2000, 366, 752-759.

4378 Analytical Chemistry, Vol. 74, No. 17, September 1, 2002

fluorescence from other components of the sample, and photobleaching, are eliminated. Major areas of analytical applications of bio- and chemiluminescence include the following: (i) development of nonradioactive detection systems for nucleic acid hybridization assays.2 In one category of assays, an enzyme is used as a label whose activity is determined by using a chemiluminogenic substrate (e.g., alkaline phosphatase along with a dioxetane derivative). Alternatively, a chemiluminescent compound (e.g., an acridinium ester) may be used as a label; (ii) immunoassays and protein blotting;3 (iii) luciferase-encoding cDNAs used as reporter genes for spatial and temporal monitoring of gene expression and for studying the strength and developmental regulation of promoters/enhancers;4,5 (iv) development of light-emitting biosensors consisting of whole cells (bacteria) as the transducer for direct determination of, for example, pollutants in environmental samples;6,7 (v) proteinprotein interaction assays carried out in living cells and based on bioluminescence resonance energy transfer.8 The rapid growth of applications of bioluminescence has stimulated research activity for investigation and exploitation of new bioluminescent systems. Gaussia princeps (http://www. biolum.org/marine/biolum2/middle/livinglights/llcrustimage. html) is a bioluminescent marine copepod (body length 10 mm) living in 350-1000-m depth.9-12 Bioluminescence originates as a secretion from 30 glands, located in the antennas, cephalothorax, thorax, and abdomen, in response to mechanical, electrical, or light stimuli. The release of a luminous bolus from G. princeps is accompanied by rapid swimming that displaces the copepod away from the bolus. G. princeps has the ability to control both the (2) Christopoulos, T. K. Anal. Chem. 1999, 71, 425R-438R. (3) Kricka, L. J. Methods Enzymol. 2000, 305, 333-345. (4) Bronstein, I.; Fortin, J.; Stanley, P. E.; Stewart, G. S.; Kricka, L. J. Anal. Biochem. 1994, 219, 169-181. (5) Srikantha, T.; Klapach, A.; Lorenz, W. W.; Tsai, L. K.; Laughlin, L. A.; Gorman, J. A.; Soll, D. R. J. Bacteriol. 1996, 178, 121-129. (6) Lewis, J. C.; Feltus, A.; Ensor, C. M.; Ramanathan, S.; Daunert, S. Anal. Chem. 1998, 70, 579A-585A. (7) Ramanathan, S.; Weiping, S.; Rosen, B. P.; Daunert, S. Anal. Chem. 1997, 69, 3380-3384. (8) Xu, Y.; Piston, D. W.; Johnson, C. H. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 151-156. (9) Barnes, A. T.; Case, J. F. J. Exp. Mar. Biol. Ecol. 1972, 8, 53-71. (10) Latz, M. I.; Bowlby, M. R.; Case, J. F. J. Exp. Mar. Biol. Ecol. 1990, 136, 1-22. (11) Bowlby, M. R.; Case, J. F. Mar. Biol. 1991, 110, 329-336. (12) Bowlby, M. R.; Case, J. F. Biol. Bull. 1991, 180, 440-446. 10.1021/ac025742k CCC: $22.00

© 2002 American Chemical Society Published on Web 07/31/2002

number of luminous glands and the strength of discharge of individual glands, thus varying the emission pattern from quick, bright flashes to a fixed pattern of discrete points glowing long after the animal has departed. Bioluminescence in G. princeps probably serves as a defense mechanism. The flashes startle and blind a dark-adapted predator. The glowing decoy provides a spatially defined target to hold the predator’s attention away from the escaping copepod.9-12 While studies on light emission from the whole organism under various stimuli have been carried out for a long time, the cloning of cDNA for Gaussia luciferase, the enzyme responsible for the bioluminescent reaction, was accomplished only recently.13 Gaussia luciferase (GLuc) is a single polypeptide chain consisting of 185 amino acids (MW)19 900). GLuc catalyzes the oxidative decarboxylation of coelenterazine to produce the excited state of coelenteramide, which upon relaxation to the ground state emits blue light (470 nm).13,14 It has been reported that when transfected into mammalian cells, Gaussia luciferase gives high levels of light emission.14 The objective of the present work is to conduct the first quantitative analytical study of Gaussia luciferase and examine its potential as a new bioluminescent reporter molecule for DNA hybridization. It is well known that luciferases are inactivated upon conjugation to other biomolecules, such as DNA probes or antibodies.15 This limits their application to DNA hybridization assays. To avoid inactivation problems, we design an overexpression system that produces in vivo biotinylated GLuc. A vector is constructed that drives the overexpression of both (a) a biotin acceptor peptide-GLuc fusion protein and (b) the biotin protein ligase (BPL). BPL catalyzes the covalent attachment of a single biotin to the fusion protein in vivo. Purification is then accomplished by affinity chromatography using a monomeric avidin column. Besides facilitating purification, the in vivo biotinylation enables subsequent complexation of GLuc with streptavidin (SA), thereby avoiding chemical conjugation reactions. We study the light emission of purified GLuc at various coelenterazine concentrations. The relationship between luminescence and the concentration of GLuc is then investigated in order to establish the detectability of the new reporter. Subsequently, the complexation of SA with GLuc is studied and the complexes are used for determination of hybrids in a microtiter well-based bioluminometric DNA hybridization assay. MATERIALS AND METHODS Instrumentation. The Chemilmager 4400 light-imaging system with AlphaEase software used for photography of ethidium bromide-stained agarose gels and Coomassie blue-stained polyacrylamide gels was from Alpha Innotech Corp. (San Leandro, CA). The Agilent 8453 UV-visible spectrophotometer was purchased from Agilent Technologies (Mississauga, ON, Canada). The dual-vertical minigel unit for SDS-PAGE was from CBS Scientific (Del Mar, CA). Luminescence measurements were carried out using the MLX microtiter plate luminometer controlled by the Revelation software from Dynex Technologies (Chantilly, VA). Polymerase chain reactions (PCR) were performed using the 48-well Perkin-Elmer Cetus DNA thermal cycler (Perkin(13) Bryan, B. J.; Szent-Gyorgyi, C. S. U.S. Patent 6232107, May 2001. (14) Ballou, B.; Szent-Gyorgyi, C.; Finley, G. Abstract from 11th Int. Symp. on Biolumin., & Chemilumin., Asilomar, CA, 2000. (15) Kricka, L. J. Anal. Biochem. 1988, 175, 14-21.

Elmer, Norwalk, CT). The hybridization assays were performed using the Amerlite shaker/incubator from Amersham (Oakville, ON, Canada). The SPECTRAmax GEMINI XS dual-scanning microplate spectrofluorometer was from Molecular Devices Corp. (Sunnyvale, CA). The Eppendorf vacufuge was from VWR (Mississauga, ON, Canada). Materials. Opaque polystyrene Microlite 2 microtiter wells were obtained from Dynatech Laboratories Inc. (Chantilly, VA). Sephadex G-25 gel filtration columns (Nap-5 and Nap-10), alkaline phosphatase (calf intestine), and T4 DNA ligase were from Amersham-Pharmacia Biotech (Baie d’Urfe, Quebec, Canada). Agarose was purchased from Invitrogen Corp. (Burlington, ON, Canada). Digoxigenin-11-dUTP (Dig-dUTP) and anti-digoxigenin antibody (from sheep) were from Roche Diagnostics (Laval, QB, Canada). The Qiaex II kit for purification of DNA from agarose gels was from Qiagen (Mississauga, ON, Canada). Microcon-30 concentrators were from Amicon (Beverly, MA), and Ultrafree15 centrifugal devices were from Millipore (Etobicoke, ON, Canada). Spin-pure G-25 columns were from CPG Inc. (Lincoln Park, NJ). Pfu turbo DNA polymerase was obtained from Stratagene (La Jolla, CA). Soft Link Soft Release avidin resin, PinPoint Xa-1 vector, DNA ladder (100 bp), and Escherichia coli JM109 cells were purchased from Promega (Madison, WI). B-PER II bacterial protein extraction reagent was from Pierce (Rockford, IL). Coomassie brilliant blue R-250, broad range SDS-PAGE molecular weight standards, and 40% acrylamide/bis (37.5:1) solution were from Bio-Rad Laboratories (Mississauga, ON, Canada). The plasmid pGLuc was obtained from Nanolight Technologies (Pinetop, AZ). Restriction enzymes KpnI, BglII, NotI, and SmaI were from MBI Fermentas (Burlington, ON, Canada). Phenylmethanesulfonyl fluoride (PMSF) was from BDH (Toronto, ON, Canada). Recombinant streptavidin, bovine serum albumin (BSA), ampicillin, CaCl2, mineral oil, Bradford protein dye reagent, D-biotin, and isopropyl thiogalactopyranoside (IPTG) were purchased from Sigma-Aldrich (Oakville, ON, Canada). All other general chemicals were also from Sigma-Aldrich. A 10 mmol/L stock coelenterazine solution was prepared by dissolving coelenterazine (Biosynth AG, Staad, Switzerland) in acidified (0.1 mol/L HCl) methanol that was deoxygenated by purging with N2. Working solutions (2.5 mmol/L) were then prepared by diluting the stock in acidified methanol. All coelenterazine solutions were stored at -80 °C for long periods of time or at -20 °C for several days. The working solution was kept on ice during preparation of the substrate solutions. Oligonucleotides used as PCR primers for creation of the birA gene insert were synthesized by Bio-Synthesis (Lewisville, TX) as follows: (a) 5′-T CCC CCG GGA GGA GAT ATA CAT ATG AAG GAT AAC ACC, a 37-mer upstream primer introducing a SmaI site (boldface type region) with the underlined region homologous to codons 1-5 of the birA gene; (b) 5′-A TAG TTT AGC GGC CGC TTA TTT TTC TGC ACT, a 31-mer downstream primer introducing a NotI site (boldface type region) with the underlined region containing a stop codon and complementary to the last four codons of the birA gene. Primers used to create the GLuc insert were synthesized by ACGT Corp. (Toronto, ON, Canada) as follows: (c) 5′-C CGG TAC CGA AAA CCA ACT GAA AAC, a 25-mer upstream primer introducing a KpnI site (boldface type region) and homologous to codons 2-6 of the GLuc gene Analytical Chemistry, Vol. 74, No. 17, September 1, 2002

4379

(underlined region); (d) 5′-GAG ATC TGA TTA ATC ACC ACC GGC ACC, a 27-mer downstream primer introducing a BglII site (boldface type region) and complementary to the last five codons of the GLuc gene with an added stop codon present (underlined region). The target DNA was a 233-bp fragment generated by RT-PCR of prostate-specific antigen mRNA, as previously described.16 The target DNA was biotinylated, through PCR, by using an upstream primer labeled with biotin at the 5′-end. Quantification of the target DNA was carried out fluorometrically by using the PicoGreen dsDNA quantitation kit from Molecular Probes (Eugene, OR). The oligonucleotide used as a probe in this work was complementary to the target region bp 67-90. Buffer A contained 10 mmol/L Tris, pH 7.8, 1 mmol/L EDTA, and 0.60 mol/L NaCl. Buffer B contained 10 mmol/L Tris, pH 7.8, 1 mmol/L EDTA, 0.60 mol/L NaCl, and 20% B-PER II reagent (v/v). Buffer C contained 0.60 mol/L NaCl, 0.10 mol/L potassium phosphate, 0.2 g/L BSA, and 0.06 mmol/L NaN3, pH 7.8. Buffer D consisted of 10 g/L blocking reagent in 0.1 mol/L maleic acid, 0.15 mol/L NaCl, and 1 mmol/L EDTA, pH 7.5. The “wash solution” contained 50 mmol/L Tris, 0.15 mol/L NaCl, 1 mmol/L EDTA, and 0.5 mL/L Tween-20, pH 7.4. The substrate solution for Gaussia luciferase consisted of coelenterazine (5-20 µmol/L) diluted in Buffer A. Polymerase Chain Reaction. PCR was carried out in a total volume of 100 µL, containing 10 mmol/L KCl, 20 mmol/L TrisHCl (pH 8.8), 1 g/L Triton X-100, 2 mmol/L MgSO4, 10 mmol/L (NH4)2SO4, 0.1 g/L BSA, 0.15 mmol/L of each dNTP, 50 pmol of each primer, appropriate DNA template, and 2.5 units of Pfu DNA polymerase. PCR was run for 30 cycles, each consisting of denaturation (95 °C, 1 min), annealing (60 °C, 1 min), and extension (72 °C, 1 min) unless otherwise specified. Mixtures were then incubated at 72 °C for 10 min and cooled to 4 °C until further use. Amplification of Gaussia Luciferase-Coding DNA. The plasmid pGLuc was used as a source of the G. princeps luciferase gene. A 526-bp DNA fragment, containing the GLuc-coding sequence flanked by BglII and KpnI restriction sites, was created through PCR using the oligonucleotides c and d as upstream and downstream primers, respectively. Amplification of birA Gene from E. coli Genomic DNA. Genomic DNA was isolated using a procedure based on ref 17. Briefly, E. coli JM109 cells were grown overnight at 37 °C in LB broth (50-mL culture). The cells were harvested by centrifugation (2000 rpm, Beckman J-6B swinging bucket centrifuge) for 30 min at 4 °C and resuspended in 5 mL of SET buffer (75 mmol/L NaCl, 25 mmol/L EDTA, 20 mmol/L Tris, pH 7.5) with lysozyme added to a final concentration of 1 g/L. Following a 1-h incubation at 37 °C, SDS and proteinase K were added to final concentrations of 10 and 0.5 g/L, respectively. The mixture (5.7 mL) was incubated at 55 °C for 2 h with occasional shaking, followed by the addition of 1/3 volume of 5 mol/L NaCl (1.9 mL) and 1 volume of CHCl3 (7.6 mL), and incubated for 30 min at room temperature with frequent inversion. The phases were separated by centrifugation (3000g, 15 min), and the genomic DNA was precipitated by adding 1 volume of 2-propanol to the aqueous layer. The DNA strands

were resuspended in TE buffer (10 mmol/L Tris, 1 mmol/L EDTA, pH 8.0), followed by ethanol precipitation.18 The birA gene was amplified from E. coli genomic DNA by PCR (35 cycles, as described above with an annealing temperature of 55 °C) using oligonucleotides a and b as upstream and downstream primers, respectively. The primers were based on sequence 9667-10632 from E. coli K12 genome (birA gene, Genbank accession number AAC76951). A 1004-bp DNA fragment was produced. Construction of Plasmid pBGLuc-birA. BirA amplification products were concentrated with Microcon-30 columns and digested with 20 units of NotI (90 min at 37 °C) followed by heat inactivation of the enzyme (65 °C for 20 min). The digested birA PCR product was then purified from a 1% agarose gel using the Qiaex II kit. The PinPoint Xa-1 (pXa) vector (3 µg) was linearized with 30 units of NotI as described above. The buffer was changed using Spin-pure G-25 columns, and dephosphorylation was carried out for 30 min at 37 °C with 0.05 unit of alkaline phosphatase in 10 mmol/L Tris acetate, 10 mmol/L magnesium acetate, and 50 mmol/L potassium acetate, pH 7.5. The enzyme was heat inactivated at 85 °C for 15 min. Both digests were quantified fluorometrically with PicoGreen, and then ligation was carried out in the same buffer for 12 h at 10 °C with 11 units of T4 DNA ligase using 62 fmol of birA insert and 4.2 fmol of PinPoint Xa-1 vector in a total volume of 30 µL. The enzyme was then inactivated at 65 °C for 10 min. The buffer was changed, and a second digestion with 20 units of SmaI was carried out for 2 h at 30 °C followed by heat inactivation at 65 °C for 20 min. The buffer was changed, and the small Smal-digested fragments were removed using a Microcon-30 spin column. Recircularization was carried using the entire digestion reaction product in a total volume of 25 µL with ligase conditions as described above, creating the plasmid pXa-birA (4298 bp). The GLuc PCR products were purified and concentrated with the Wizard PCR Preps DNA purification system and subsequently digested with 10 units of KpnI for 90 min at 37 °C. Spin-pure G-25 columns were used to exchange the buffer followed by digestion with 10 units of BglII for 90 min at 37 °C. The recombinant pXa-birA plasmid was similarly digested with both BglII and KpnI. The digested products were separated by agarose gel (1.2%) electrophoresis, and the appropriate fragments were purified from the gel (using the Qiaex II purification kit) and quantified fluorometrically with PicoGreen. The doubly digested purified GLuc coding sequence (182 fmol) and the pXa-birA fragments (46 fmol) were ligated in a total volume of 20 µL as above to create the plasmid pBGLuc-birA. Bacterial Expression and Purification of in Vivo Biotinylated Gluc (BGLuc). E. coli strain JM109 transformed with the pBGLuc-birA was inoculated into LB broth (3 mL) containing 0.1 g/L ampicillin and supplemented with 2 mg/L biotin. This seed culture was grown overnight at 30 °C, diluted 100-fold, and again grown at 30 °C until an absorbance of 0.8-0.9 (at 600 nm) was reached. At this point, protein synthesis was induced with 1 mmol/L IPTG for 6 h. Cells were harvested by centrifugation at 2000 rpm for 25 min at 4 °C and stored at -20 °C.

(16) Verhaegen, M.; Ioannou, P. C.; Christopoulos, T. K. Clin. Chem. 1998, 44, 1170-1176. (17) Pospiech, A.; Neumann, B. Trends Genet. 1995, 11, 217-218.

(18) Sambrook, J.; Fritsch, E. F.; Maniatis, T. Molecular Cloning. A laboratory manual, 2nd ed.; Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY, 1989.

4380

Analytical Chemistry, Vol. 74, No. 17, September 1, 2002

Frozen cell pellets were washed twice in TBS buffer (25 mmol/L Tris, 2.6 mmol/L KCl, 137 mmol/L NaCl, pH 7.4) to remove excess biotin. Cell lysis and soluble protein extraction were carried out using the B-PER II bacterial protein extraction reagent as per manufacturer’s instructions. PMSF was included at a final concentration of 0.7 mmol/L. The monomeric avidin resin was initially incubated with free biotin (5 mmol/L in 0.1 mol/L sodium phosphate, pH 7.0) for 15 min to saturate any nonreversible tetrameric avidin-binding sites that may have been present. The resin was then regenerated by washing with 8 column volumes of 10% acetic acid, followed by washing with 8 volumes of sodium phosphate (0.1 mol/L, pH 7.0). When the pH reached 6.8, the wash flow was stopped for 30 min to allow time for avidin refolding. Prior to use, the resin (typically 1 mL of slurry/g of cell pellet) was equilibrated with buffer B. The crude cell extract diluted 5 times in buffer A was then applied directly to the slurry, and binding was allowed to occur for 60 min at room temperature with gentle shaking. After spinning at 1000g for 2 min (Beckman, swinging bucket model J6-B), the crude supernatant was removed and the avidin resin containing bound BGLuc protein was transferred to a small column. The column was washed with 10 mL of buffer B, followed by elution with 5 mmol/L biotin in buffer B. Fractions of 1 mL were collected, and the BGluc activity was measured by diluting each fraction 600-fold in buffer C, pipetting 5 µL into each well containing 45 µL of buffer C and dispensing 25 µL of 10 µmol/L coelenterazine diluted in buffer A. The luminescence was integrated for 15 s. Fractions giving high luminescence signals were pooled, concentrated (by ultrafiltration using an Ultrafree-15 centrifugal device), and purified three times from excess biotin by size exclusion chromatography with Nap-5 columns, followed by final concentration using the Eppendorf Vacufuge. Final preparations of the purified BGLuc enzyme were present in buffer A containing 10% B-PER II reagent and supplemented with 10% glycerol. The mass of purified biotinylated GLuc was determined by using the Bradford assay. The solution of BGLuc was aliquoted and stored immediately at -20 °C. Bioluminometric Determination of Gaussia Luciferase. Assays for luminescence activity of Gaussia luciferase were carried out by diluting the purified BGLuc in buffer D and pipetting 50 µL into a microtiter well. The coelenterazine substrate solution (50 µL, 20 µmol/L, diluted in buffer A) was then injected into the well and the luminescence was integrated for 20 s. Labeling of the Oligonucleotide Probe. The probe was tailed enzymically with Dig-dUTP using terminal deoxynucleotidyl transferase (TdT). Tailing reactions were carried out in a total volume of 20 µL consisting of 0.2 mol/L potassium cacodylate, 25 mmol/L Tris-HCl (pH 6.6), 0.25 g/L BSA, 5 mmol/L CoCl2, 50 µmol/L Dig-dUTP, 0.5 mmol/L dATP, 25 units of TdT, and 100 pmol of probe. The reactions were carried out for 60 min at 37 °C. The tailed probe was purified once from excess Dig-dUTP using a Nap-5 column and then concentrated with the Eppendorf vacufuge. The final concentration of the probe was 2 µmol/L. Microtiter Well-Based Bioluminometric Hybridization Assay Using in Vivo Biotinylated Gaussia Luciferase as a Reporter Molecule. Opaque polystyrene microtiter wells were coated, overnight at room temperature, with 50 µL of 5 mg/L antidigoxigenin antibody diluted in 0.1 mol/L carbonate buffer, pH

9.6. The wells were washed 3 times with wash solution, and 50 µL of 4 nmol/L probe, diluted in buffer D, was added into each well. The probe was allowed to bind to the antibody for 60 min at room temperature, followed by washing. Biotinylated target DNA, diluted in buffer D, was denatured at 95 °C for 10 min and immediately placed on ice. A 10-µL aliquot was added to each well containing 40 µL of buffer D, and hybridization was carried out at 42 °C for 30 min followed by washing. A complex of streptavidin with biotinylated GLuc was prepared by mixing 20 nmol/L biotinylated GLuc with 10 nmol/L streptavidin (both diluted in buffer D) and incubating for 30 min at room temperature. The complex was diluted 5 times in buffer D, and a 50-µL aliquot was added to each well. After a 30-min incubation, the wells were washed and the activity of GLuc was measured by adding coelenterazine solution as described above. RESULTS AND DISCUSSION Construction of Plasmid pBGLuc-birA. The plasmid pBGLuc-birA (4815 bp) drives the overexpression, in E. coli, of both the biotin protein ligase and a fusion protein consisting of the biotin acceptor peptide (bap) genetically attached to the amino terminus of Gaussia luciferase. The construction of pBGLuc-birA is illustrated in Figure 1. The Pinpoint Xa-1 vector was used as a source of the sequence encoding the bap from Propionibacterium shermanii transcarboxylase positioned downstream of the tac promoter and a ribosome binding site. The structure of the biotin domain of P. shermanii transcarboxylase is very similar to that of E. coli acetyl-CoA carboxylase, which is the physiological substrate of biotin protein ligase.19 BPL is a 35.5-kDa monomeric enzyme encoded by the birA gene that catalyzes the in vivo biotinylation of the biotin acceptor peptide at a unique site, thus producing a biotinylated bap-GLuc fusion protein. This facilitates (a) the purification of BGLuc, in a single step, from the crude cellular extract by affinity chromatography using a monomeric avidin column and (b) the complexation of purified BGLuc with streptavidin for direct use in bioluminometric hybridization assays, thus avoiding chemical conjugation. Following plasmid isolation, pBGLuc-birA was digested to confirm the presence and size of both the Gaussia luciferase and birA genes. A 1.2% agarose gel electropherogram is shown in Figure 1. Expression and Purification of in Vivo Biotinylated Gaussia Luciferase (BGLuc). Bacterial cells were transformed with pBGLuc-birA and protein synthesis was induced with IPTG. The in vivo biotinylated GLuc was purified from the crude cell extract by affinity chromatography using a monomeric avidin resin. Monomeric avidin has a much lower affinity for biotin (Kd ) 10-7) than the native tetrameric avidin (Kd ) 10-15), thus allowing both the binding of biotinylated molecules and the subsequent elution with free biotin under mild, nondenaturing conditions.20 A typical profile of the protein mass (as determined by Bradford assay) and the luminescence obtained during washing of the monomeric avidin resin and elution of the in vivo biotinylated Gaussia luciferase are shown in Figure 2. The protein mass of the first fraction (corresponding to the removed supernatant) is high whereas the luminescence is low. The total protein then decreases to undetectable levels as soluble proteins from the cell (19) Reddy, D. V.; Rothemund, S.; Shenoy, B. C.; Carey, P. R.; Sonnichsen, F. D. Protein Sci. 1998, 7, 2156-2163. (20) Kohanski, R. A.; Lane, M. D. Methods Enzymol. 1990, 184, 194-200.

Analytical Chemistry, Vol. 74, No. 17, September 1, 2002

4381

Figure 1. Left panel: Schematic presentation of the construction of plasmid pBGLuc-birA suitable for expression of biotinylated Gaussia luciferase. Relative locations of the tac promoter, the ribosome binding site (RBS), the restriction sites, and the gene inserts are indicated. birA, gene encoding the biotin protein ligase in E. coli; ampr, β-lactamase gene conferring ampicillin resistance. Right panel: Ethidium bromidestained 1.2% agarose gel depicting the electrophoretic analysis of plasmid pBGLuc-birA. Lane 1: molecular weight DNA markers. Lane 2: recombinant plasmid pBGLuc-birA (4815 bp) linearized with NotI. Lane 3: plasmid pBGLuc-birA digested with BglII, KpnI, and NotI to excise both the Gaussia luciferase (517 bp) and the birA (992 bp) genes.

Figure 2. Left panel: Typical profile of protein mass and luminescence obtained during purification of in vivo biotinylated Gaussia luciferase from a crude cell extract by affinity chromatography using a monomeric avidin resin. A 1-g cell pellet was used, and 1-mL fractions were collected. The protein mass of each fraction was determined by Bradford assay, and luciferase activity monitored by diluting each fraction 6000-fold (in buffer C) and adding 50 µL to wells. Following the injection of 25 µL of coelenterazine solution (10 µmol/L in buffer A), the luminescence was integrated for 10 s. Right panel: A 12% SDS-PAGE gel. Lane 1: Broad range molecular weight protein markers. Lane 2: 10 µL of crude cell lysate resulting from B-PER II lysis. Lane 3: 10 µL of total soluble protein fraction prior to purification with monomeric avidin resin. Lane 4: approximately 2 µg purified, concentrated in vivo biotinylated Gaussia luciferase. All procedures to obtain samples are described in Materials and Methods. The gel was stained with Coomassie Brilliant blue R-250.

extract are washed through the column. Subsequently, biotinylated Gaussia luciferase is eluted using buffer B supplemented with 5 mmol/L biotin. Most of the GLuc activity is eluted in the first 3-4 mL of elution buffer and pooled for further concentration. Following the affinity chromatography purification step, approximately 40% of the luciferase activity (as determined by the luminescence signal) was recovered in the pooled fractions. The remaining activity was removed with the supernatant or passed through the column during washing. Supplementing the growth 4382 Analytical Chemistry, Vol. 74, No. 17, September 1, 2002

media with increasing concentrations of biotin (up to 10 mg/L) and prolonging the induction time resulted in no improvement in capture of the fusion protein on the avidin resin. The expression system was designed such that the in vivo biotinylation reaction does not rely only on the catalytic activity of the endogenous BPL of the host. Plasmid pBGLuc-birA drives the overexpression of BPL along with GLuc (as a single mRNA molecule). It has been shown that about 90% biotinylation is attained by the combined use of birA expression and biotin supplemented growth media.21

Consequently, the loss of a fraction of GLuc during purification was not due to low BPL activity but rather to the fact that, in this fraction, the folding of the fusion protein either rendered the biotinylation site inaccessible to BPL or it resulted in a limited accessibility (postbiotinylation) of the biotin moiety to the avidin resin. To test whether the presence of B-PER reagent in the cellular extract affected the binding of biotinylated protein to monomeric avidin resin, we diluted the cell extract 5-50 times in buffer A prior to mixing with the resin. There was no improvement of BGLuc binding with decreasing B-PER concentrations. It should be noted, however, that the addition of B-PER to the elution buffer contributed to the stability of purified GLuc. Elution in buffer B (buffer A plus B-PER) resulted in no decrease in luminescence even after 3 weeks at 4 °C (and much longer at -20 °C), whereas purified BGLuc stored in buffer A lost luminescence activity at 4 °C in less than 24 h. Bradford assay of the purified protein, following affinity purification, concentration by ultrafiltration, and biotin removal by size exclusion chromatography (repeated 3 times), gave approximately 90-115 µg of BGLuc/g of cell pellet. This corresponds to a yield of 0.55-0.69 mg of purified protein/L of bacterial culture. Heterologous proteins are often packaged in the cytoplasm of E. coli in the form of inactive aggregates (inclusion bodies) when plasmid-encoded genes are directed to be overexpressed at high rates.22-24 Since both recombinant GLuc and BPL were overexpressed, some inclusion body formation is unavoidable. In this work, only the soluble fraction of biotinylated GLuc was purified. SDS-PAGE analysis was performed to verify the size and purity of the in vivo biotinylated Gaussia luciferase (Figure 2). Following purification, a single band was observed at 33.9 kDa, corresponding to the sum of the 18.8-kDa Gaussia luciferase and 15.1-kDa biotin acceptor domain. Plasmid pGLuc contains the cDNA encoding amino acids 18-185 of the luciferase. Consequently, the expressed BGLuc lacks the first 17 amino acids of the native protein. E. coli contains one biotinylated protein, the biotin carboxy carrier protein (BCCP), one of the three subunits of acetyl-CoA carboxylase,25 which also binds avidin and would have been coeluted with the biotinylated GLuc. However, no band corresponding to BCCP was observed in SDS-PAGE, indicating that BCCP was present at a much lower concentration than Gaussia luciferase. Determination of Gaussia Luciferase by Using Its Bioluminescent Reaction. Data pertaining to the time course of light emission from the in vivo biotinylated Gaussia luciferase-catalyzed reaction, at various coelenterazine concentrations, are presented in Figure 3. The light emission peaks at 1 s and then decreases rapidly with a half-life dependent on coelenterazine concentration (flash-type emission). The decrease is attributed to the inhibition of GLuc from the reaction product (coelenteramide). It has been reported that coelenteramide also inhibits the luciferase from (21) Smith, P. A.; Tripp, B. C.; DiBlasio-Smith, E. A.; Lu, Z.; LaVallie, E. R.; McCoy, J. M. Nucleic Acids Res. 1998, 26, 1414-1420. (22) Georgiou, G.; Valax, P. Curr. Opin. Biotechnol. 1996, 7, 190-197. (23) Strandberg, L.; Enfors, S. O. Appl. Environ. Microbiol. 1991, 57, 16691674. (24) Marston, F. A. O. Biochem. J. 1986, 240, 1-12. (25) Fall, R. R. Methods Enzymol. 1979, 184, 390-398.

Figure 3. Light emission kinetics for biotinylated Gaussia luciferase at various concentrations of coelenterazine. Data are shown for 0.4 fmol of BGLuc (50 µL) diluted in buffer D, followed by injection of 50 µL of coelenterazine diluted in buffer A. Initial light emission peaks at 1 s and rapidly decays. Both peak height and total light production increase with substrate concentration.

Renilla reniformis.26 The light emission was integrated for 20 s, for all subsequent studies. The effect of coelenterazine concentration on the activity of BGLuc was studied in the range of 5-130 µmol/L, and the results are presented in Figure 4. While the luminescence increased with higher coelenterazine concentrations, the background also increased accordingly. The background is defined as the luminescence obtained with all reagents present but zero BGLuc concentration. As a result, the signal-to-background ratio remained relatively constant over the entire range of coelenterazine concentrations studied, with a slight enhancement at 20 µmol/L. It has been reported previously that coelenterazine in aqueous solutions exhibits a low level of luminescence that has been attributed to autooxidation.27 To assess the detectability of in vivo biotinylated GLuc, serial dilutions of the purified protein (from a stock whose protein concentration was determined by Bradford assay) were prepared in buffer D, 50 µL was pipetted into the well, and the luminescence was measured after injecting 50 µL of 20 µmol/L coelenterazine solution. In Figure 4, the luminescence (corrected for the background) is plotted as a function of the amount of biotinylated Gaussia luciferase. As low as 1 amol of Gaussia luciferase can be detected with a signal-to-background ratio of 2.0. Moreover, the linearity extends over 5 orders of magnitude. This experiment was repeated five times over the period of one month, using different stock coelenterazine solutions, and the standard deviations of the signals are presented as error bars in Figure 4. Coelenterazine autooxidation is the impediment to the detection of even lower amounts of GLuc. (26) Matthews, J. C.; Hori, K.; Cormier, M. J. Biochemistry 1977, 16, 85-91. (27) Teranishi, K.; Shimomura, O. Anal. Biochem. 1997, 249, 37-43.

Analytical Chemistry, Vol. 74, No. 17, September 1, 2002

4383

Figure 4. Left panel: effect of coelenterazine concentration on the luminescence (signal) of biotinylated Gaussia luciferase (solid line). The dashed line represents the effect of coelenterazine concentration on the background, i.e., the luminescence obtained in the absence of BGLuc. In both cases, the luminescence was integrated for 20 s. BGLuc diluted in buffer D (0.4 fmol, 50 µL) was pipetted in the well followed by injection of 50 µL of coelenterazine solution. Right panel: luminescence as a function of the attomoles of in vivo biotinylated Gaussia luciferase. Serial dilutions of BGLuc in buffer D (50 µL) were pipetted into microtiter wells, and the luminescence was integrated for 20 s following addition of 50 µL of 20 µmol/L coelenterazine substrate. The error bars represent the standard deviation of the luminescence from graphs obtained on different days over a 1-month period (n ) 5).

Figure 5. Left panel: study of the complexation of in vivo biotinylated Gaussia luciferase with streptavidin. The complexes were formed by mixing 4 nmol/L biotinylated Gaussia luciferase with various concentrations of streptavidin. The complexes were then tested using the DNA hybridization assay configuration described under Materials and Methods. The luminescence and the signal-to-background ratios are plotted as a function of streptavidin concentration. Each experimental point represents the mean value of two assays. Right panel: luminescence as a function of target DNA concentration. The hybridization assay was performed as described under Materials and Methods using in vivo biotinylated Gaussia luciferase as a reporter molecule. The error bars correspond to the standard deviation of the luminescence from graphs obtained on different days (n ) 5).

In Vivo Biotinylated Gaussia Luciferase as a Reporter in Bioluminometric DNA Hybridization Assays. A hybridization assay was developed in which a biotinylated, denatured target DNA was hybridized to an oligonucleotide probe immobilized in microtiter wells through a digoxigenin-antidigoxigenin interaction. The hybrids were then quantified by adding a streptavidinbiotinylated Gaussia luciferase complex and using coelenterazine as the substrate. The removal of excess biotin following purification of the BGLuc from the avidin resin was crucial prior to complexation with streptavidin. Optimization of the biotinylated Gaussia luciferase-to-streptavidin molar ratio was then carried out by incubating a constant concentration of BGLuc (4 nmol/L) with varying concentrations of streptavidin (1-16 nmol/L) and applying the complex (SA-BGLuc) to the wells. As shown in Figure 5, maximum luminescence occurs at a 1:2 molar ratio of streptavidin 4384 Analytical Chemistry, Vol. 74, No. 17, September 1, 2002

to BGLuc. If biotinylated GLuc is in excess, the four biotin-binding sites on streptavidin become saturated and the SA-BGLuc complex is no longer able to bind the immobilized biotinylated hybrids. If streptavidin is in excess, then free streptavidin competes with the SA-BGLuc complex for binding to biotinylated hybrids on the well. Complexation with SA does not inhibit the bioluminescence of BGLuc, because the biotin moiety is attached to the biotin acceptor peptide and not to the primary sequence of BGLuc. This constitutes a significant advantage of in vivo biotinylation or labeling in general. On the contrary, in vitro biotinylation reactions utilizing cross-linking reagents cause modification of several functional groups of the protein (e.g., primary amino groups), which may affect its biological activity. To assess the overall performance of the hybridization assay, various dilutions of biotinylated target DNA were prepared in buffer D and analyzed as described in Materials and Methods. In

Figure 5, the luminescence (corrected for the background) was plotted as a function of target DNA concentration. The background is defined as the luminescence obtained when no target DNA is present in the well. The linearity of the assay extends from 1.6 to 800 pmol/L target DNA. The signal-to-background ratio at 1.6 pmol/L (80 amol/well) is 1.4. The assay was repeated over 1-week period and the signals were averaged (n ) 5). The (day-to-day) standard deviations at each concentration are shown as error bars. To our knowledge, this is the first report describing the use of a luciferase as a label in a DNA hybridization assay. The inactivation of luciferases upon conjugation has prohibited their application as reporter molecules in in vitro assays. At this stage, GLuc-based assays exhibit similar detectabilities with assays using the photoprotein aequorin. Aequorin is a complex of apoaequorin with coelenterazine and exhibits no enzymic turnover. However, GLuc has the potential for a severalfold higher detectability than aequorin if the autooxidation of coelenterazine can be reduced. It is calculated that 1 L of bacterial culture provides enough biotinylated Gaussia luciferase for 150 000 hybridization assays. The entire process, including culturing pBGLuc-birA transformed cells, extraction of total soluble protein, and generation and purification of active biotinylated Gaussia luciferase is completed

in less than 2 days. The hybridization assay on wells containing immobilized probe is complete in 60 min including the complexation reaction. Gaussia luciferase activity is measured rapidly and easily with high sensitivity and a wide linear range. The detectability of GLuc may be further enhanced by suppressing the autooxidation (autoluminescence) of coelenterazine, thus allowing higher substrate concentrations to be used. Work in this direction is currently being performed in our laboratory. Finally, the biotinylated GLuc-streptavidin complex may also be used as a detection reagent in immunoassays in combination with biotinylated specific antibodies. ACKNOWLEDGMENT This work was supported from the National Science and Engineering Research Council of Canada (NSERC) and the University of Windsor. M.V. acknowledges an Ontario Graduate Scholarship (OGS) and an Ontario Graduate Scholarship in Science and Technology (OGST).

Received for review May 1, 2002. Accepted July 5, 2002. AC025742K

Analytical Chemistry, Vol. 74, No. 17, September 1, 2002

4385