Genetically Modified Semisynthetic Bioluminescent Photoprotein

Oct 20, 2008 - Department of Chemistry, University of Kentucky, Lexington, Kentucky 40506, and Department of Chemistry and Chemical Biology, Indiana U...
0 downloads 13 Views 598KB Size
Anal. Chem. 2008, 80, 8470–8476

Genetically Modified Semisynthetic Bioluminescent Photoprotein Variants: Simultaneous Dual-Analyte Assay in a Single Well Employing Time Resolution of Decay Kinetics Laura Rowe,† Kelly Combs,† Sapna Deo,‡ C. Ensor,† Sylvia Daunert,*,† and Xiaoge Qu† Department of Chemistry, University of Kentucky, Lexington, Kentucky 40506, and Department of Chemistry and Chemical Biology, Indiana University Purdue University Indianapolis, Indianapolis, Indiana 46202 Progress in the miniaturization and automation of complex analytical processes depends largely on increasing the sensitivity, diversity, and robustness of current labels. Because of their ubiquity and ease of use, fluorescent, enzymatic, and bioluminescent labels are often employed in such miniaturized and multiplexed formats, with each type of label having its own unique advantages and drawbacks. The ultrasensitive detection limits of bioluminescent reporters are especially advantageous when dealing with very small sample volumes and biological fluids. However, bioluminescent reporters currently do not have the multiplexing capability that fluorescent labels do. In an effort to address this limitation, we have developed a method of discriminating two semisynthetic aequorin variants from one another using time resolution. In this work we paired two aequorin conjugates with different coelenterazine analogues and then resolved the two signals from one another using the difference in decay kinetics and half-life times. Utilizing this time-resolution, we then developed a simultaneous, dual-analyte, single well assay for 6-keto-prostaglandin-FI-r and angiotensin II, two important cardiovascular molecules. As the push toward faster and more automated analytical techniques has increased, so has the push toward improving and expanding miniaturized multiplexing technologies. However, uniting the goals of multiplexing, miniaturization, and increased speed creates unique challenges when developing assays. The underlying challenges being that the small volumes used in miniaturization accentuates the need for extremely sensitive detection methods, while multiplexing requires that the signal from multiple analytes can be discriminated from one another simultaneously.1,2 Bioluminescent labels are promising labels to meet such goals, in terms of speed and sensitivity, as their signal generation is usually faster than enzymatic-based labels, and they are more * To whom correspondence should be addressed. † University of Kentucky. ‡ Indiana University Purdue University Indianapolis. (1) Madou, M. Fundamentals of Microfabrication: The Science of Miniaturization, 2nd ed.; CRC Press: Boca Raton, FL, 2002. (2) Daunert, S.; Deo, S. Photoproteins in Bioanalysis; Wiley-VCH: Weinheim, Germany, 2006.

8470

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

sensitive than fluorescent reporters.3 Dual- or multianalyte assays have been developed previously with bioluminescent labels, but bioluminescence does not currently have near the multiplexing capacity as fluorescence. Examples of multianalyte detection with bioluminescence include pairing fluorescent or enzymatic labels with bioluminescent labels in order to create dual analyte assays whose signals are emitted due to separate stimuli (calcium and excitation light, etc.)4,5 Purely bioluminescent multianalyte assays have also been developed utilizing spectral resolution of several different luciferase mutants, in order to create a triple color mammalian cell based assay, for example.6,7 In an effort to expand the multiplexing capacity of bioluminescent reporters, we describe herein a novel method of discriminating the bioluminescent signal of two aequorin variants using time resolution of the bioluminescent decay profiles. Aequorin is a 22 kDa, globular bioluminescent photoprotein that sequesters the imidazopyrazinone chromophore, coelenterazine, in its hydrophobic core.8 The binding of calcium to aequorin causes a conformational change in the protein’s structure. This structural change leads to the oxidation of the internal coelenterazine molecule, which then relaxes by emitting bioluminescent light at 471 nm.9 It has been found that by altering the functionality of the coelenterazine chromophore and combining these altered coelenterazines with various aequorin mutants, “semi-synthetic” aequorin mutants can be produced which exhibit significantly different decay half-lives and emission maximas.10-16 Although the different colors provided by shifted emission maxima are often (3) Lewis, J.; Daunert, S. Fresenius J. Anal. Chem. 2000, 366, 760–769. (4) Verhaeden, M.; Christopoulos, T. Anal. Chem. 1998, 70, 4120–4125. (5) Konstantou, J.; Ioannou, P.; Christopoulos, T. Anal. Bioanal. Chem. 2007, 388, 1747–1754. (6) Michelini, E.; Cevenini, L.; Mezzanotte, L.; Ablamsky, D.; Southworth, T.; Branchini, B. R.; Roda, A. Photochem. Photobiol. Sci. 2008, 7, 212–217. (7) Michelini, E.; Cevenini, L.; Mezzanotte, L.; Ablamsky, D.; Southworth, T.; Branchini, B. R.; Roda, A. Anal. Chem. 2008, 80, 260–267. (8) Head, J.; Inouye, K.; Teranishi, K.; Shimomura, O. Nature 2000, 405, 372– 376. (9) Prendergast, F. Nature 2000, 405, 291–293. (10) Shimomura, O.; Musicki, B. Biochem. J. 1988, 251, 405–410. (11) Shimomura, O.; Musicki, B.; Kishi, Y. Biochem. J. 1989, 261, 913–920. (12) Shimomura, O.; Inouye, S.; Musicki, B.; Kishi, Y. Biochem. J. 1990, 270, 309–312. (13) Shimomura, O.; Musicki, B.; Kishi, Y.; Inouye, S. Cell Calcium 1993, 14, 373–378. (14) Hirano, T.; Ohmiya, Y.; Maki, S.; Niwa, H.; Ohashi, M. Tetrahedron Lett. 1998, 39, 5541–5544. (15) Zheng, J.-L.; Chen, F.; et al. Bull. Chem. Soc. Jpn. 2000, 73, 465–469. 10.1021/ac801209x CCC: $40.75  2008 American Chemical Society Published on Web 10/21/2008

utilized for multiplexing with fluorescent proteins, spectral resolution with bioluminescent proteins possess some disadvantages. For example, bioluminescent proteins have a very broad halfmaximum width, causing bioluminescent spectral resolution to require postdetection analysis in order to eliminate spectral crosstalk.17 Moreover, simultaneous spectral resolution requires the use of more expensive detection devices, such as charge coupled device (CCD) cameras. The alternative of time resolution, however, eliminates such disadvantages, since the discrete time signals can be separated by utilizing different channels and the signal can be detected with more affordable devices such as photomultiplier tubes (PMTs). As previously mentioned, recent work has shown that combining certain aequorin mutants with coelenterazine analogues results in semisynthetic aequorin mutants whose decay half-lives are significantly altered.14 The difference in decay half-lives between several of these semisynthetic aequorin mutants varies by more than 20 s, indicating the possibility of time resolution. Thus, in this work we demonstrate how these semisynthetic aequorins can, in fact, be discriminated from one another using time resolution in a dual-analyte, single well bioluminescence assay. We first combined a 6-keto-prostaglandin-F1R-aequorin (6keto-PGF1R) conjugate and an angiotensin II-aequorin fusion protein with two different coelenterazine analogues to form semisynthetic aequorin constructs. Next, we established that the relevant multiplexing property of these semisynthetic aequorin constructs (decay half-life) remained constant in various environmental conditions. Finally, we developed a dual-analyte, single well, bioluminescence immunoassay for 6-keto-PGF1R and angiotensin II using the semisynthetic aequorin constructs (Figure 1). MATERIALS AND METHODS Reagents. Luria Bertrani broth (LB) was purchased from Difco (Sparks, MD), and all acids and bases were purchased from EM Science (Gibbstown, NJ). POROS Self-pack 20 HQ strong anionic exchange and butyl sepharose beads were purchased from Applied Biosystems (Foster City, CA). Tris(hydroxymethyl)amino methane (Tris) free base, ethylenediaminetetraacetic acid (EDTA) sodium salt, glucose, sodium dodecyl sulfate (SDS), ampicillin, Tween-20, sodium phosphate, sodium chloride, bovine serum albumin (BSA), angiotensin II standard, and all other reagents were purchased from Sigma (St. Louis, MO). Coelenterazine i and coelenterazine ntv were all purchased from Biotium (Hayward, CA). Rabbit polyclonal antiserum specific to human 6-keto-PGF1R and the 6-keto-PGF1R standard were purchased from Assay Designs (Ann Arbor, MI), and rabbit monoclonal antibodies specific to human angiotensin II were purchased from AssayPro (St. Charles, MO). Reacti-Bind Goat Anti-Rabbit IgG Coated White 96 well plates were purchased from Pierce Biotechnology (St. Louis, MO). Apparatus. Bacteria expressing the aequorin proteins were incubated in a Fisher Scientific incubating orbital shaker (Fair Lawn, NJ) and collected using a Beckman J2-M1 centrifuge (Palo Alto, CA). The aequorin and aequorin-angiotensin II fusion peptide were both chromatographically purified using a BioCAD SPRINT (16) Rowe, L.; Rothert, A.; et al. Protein Eng., Des. Sci. 2008, 21, 73–81. (17) Roda, A.; Guardigli, M.; Ziessel, R.; Mirasoli, M.; Michelini, E.; Musiani, M. Microchem. J. 2007, 85 (1), 5–12.

Figure 1. Schematic of the time-resolved assay design. Microtiter plate wells are coated with goat antirabbit secondary antibodies (green Ys) and rabbit anti-6-keto-PGF1R (yellow Ys), and rabbit antiangiotensin II (pink Ys) antibodies are first allowed to bind to these secondary antibodies. The 6-keto-PGF1R-aequorin-ctz ntv (yellow arrow with blue star) and angiotensin II-aequorin-ctz i (pink arrow with blue star) conjugates then compete with native 6-keto-PGF1R (yellow arrow) and angiotensin II (pink arrow) for these limited antibody sites. Increasing amounts of native 6-keto-PGF1R and angiotensin II will correspond to decreasing amounts of bioluminescence in the respective time channel. The time resolution aspect is possible because 6-keto-PGF1R-aequorin-ctz ntv conjugate has a short decay half-life (0.5 s) so that its bioluminescence signal is only present in the 0-6 s time channel, whereas the angiotensin II-aequorin-ctz i conjugate’s bioluminescence can be detected in the 6.01-25 s time channel, since it has a decay half-life of over 11 s.

perfusion chromatography system from PE Biosciences (Framingham, MA). Coomassie blue staining with sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) was utilized to determine protein purity. Bioluminescence measurements were obtained using a Polarstar Optima luminometer from BMG Labtech (Durham, NC) and analyzed with GraphPad Prism 4.0 (San Diego, CA). Construction of Semi-Synthetic Aequorin Constructs: 6-keto-PGF1r-Aequorin-Coelenterazine Native (ctz ntv) and Angiotensin II-Aequorin-Coelenterazine i (ctz i). A cysteinefree aequorin mutant was constructed by mutating all three native cysteine residues to serine residues with site-directed mutagenesis. See ref 16 for details.18 The aequorin mutant was then expressed under a lacZ promoter in an OmpA containing pIN4 plasmid transformed into Top10 Escherichia coli cells. Following induction with IPTG, the aequorin mutant was purified to >95% purity using ion exchange followed by butyl sepharose chromatography. An exhaustive description of the protein expression and purification technique used herein can be found in ref 14. The 6-keto-prostaglandin F1R was conjugated to the purified cysteine free aequorin mutant through the primary amine groups of aequorin’s lysine residues using an NHS ester conjugation. For details concerning the conjugation reaction see ref 16. The 6-ketoPGF1R-aequorin conjugate was charged with ctz ntv by incubating the conjugate with a 5 M ctz ntv excess (100 µL of ctz ntv/mL MeOH) for 16 h at 4 °C. The conjugate was then lyophilized and (18) Lewis, J. C.; Cullen, L. C.; Daunert, S. Bioconjugate Chem. 2000, 11, 140– 145.

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

8471

stored at -80 °C until needed. The angiotensin II-aequorin fusion protein was constructed by adding a six amino acid spacer sequence and the eight amino acid sequence of angiotensin II (DRVYIHPF) to the N terminus of the cysteine free aequorin mutant using primers and PCR amplification and purified using ion exchange chromatography (See ref 17 for method details.).19 The purified angiotensin II-aequorin fusion protein was then charged with a 5 M excess of ctz i for 16 h at 4 °C. The angiotensin II-aeq-ctz i constructs were stored at 4 °C until needed. Decay Half-Life of Semisynthetic Aequorin Constructs: pH Study. For decay half-life determinations, the semisynthetic aequorin constructs were diluted with a 30 mM Tris, 150 mM NaCl, 2 mM EDTA, 1 mg/mL BSA, pH 7.5 buffer (buffer A) until their maximal bioluminescent intensity was less that 500 000 and greater than 50 000 relative light units (RLUs) when analyzed using the following protocol on the PolarStar luminometer. A volume of 5-50 µL of the semisynthetic aequorin construct was added to the well of a microtiter plate, and 50 µL of a 100 mM CaCl2, 30 mM Tris buffer (buffer D) was injected into the well immediately prior to bioluminescence readings. Light intensity measurements were collected on the PolarStar luminometer at 0.1 s intervals, for a period of 25 s. All measurements were done in triplicate, and the mean of the three measurements is the reported decay half-life. Bioluminescent decay half-life was calculated using a one phase exponential decay curve fit equation on GraphPad Prism 4.0 and the first order decay kinetics half-life equation. The bioluminescent decay curve was fit starting at the maximal light intensity signal, and the plateau was constrained to zero in all cases. For the pH study, the pH of buffer A aliquots were adjusted using 5 M NaOH or concentrated HCl in order to make the following pH 5, 5.5, 6, 6.5, 7, 7.5, 8, 8.5, and 9 buffer A solutions. Next, the 6-keto-PGF1R-aequorin-ctz ntv and the angiotensin II-aequorin-ctz i semisynthetic constructs were diluted with each of these pH differing buffers until their maximal RLU signal ranged from between 50 000 and 500 000. All measurements were made in triplicate, and the decay half-life time was calculated in the manner described above. Additionally, this and all subsequent experiments were done at room temperature (25 °C) unless otherwise noted. It is likely that the decay kinetics of these semisynthetic aequorin constructs will change slightly at extreme temperatures, with very cold temperature leading to slightly extended decay half-life times and very hot temperature leading to slightly shorter decay half-life times. However, to the best of our knowledge a thorough study on the effect of environmental temperature on the decay half-life time of photoproteins has yet to be undertaken, and in the current experiments the variation caused by a few degrees was negligible. Discrimination of Semi-Synthetic Aequorin Constructs at Two Discrete Time Intervals: Dilution Study. Two time intervals were selected for signal discrimination, 0-6 s and 6.01-25 s. These values were determined based on the decay kinetic profiles of the two semisynthetic aequorin constructs, and the time intervals started with the injection of the bioluminescence triggering calcium buffer (buffer D). The 6-keto-PGF1R-aequorinctz ntv had a decay half-life of approximately 0.5 s and was therefore resolved in the 0-6 s time interval channel. Angiotensin (19) Qu, X.; Deo, S. K.; Dikici, E.; Ensor, M.; Poon, M.; Daunert, S. Anal. Biochem. 2007, 371 (2), 154–161.

8472

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

II-aequorin-ctz i had a decay half-life of approximately 11 s, and its signal was therefore resolved in the 6.01-25 s time interval channel. Both semisynthetic constructs were logarithmically diluted with buffer A, the 6-keto-PGF1R-aequorin-ctz ntv was diluted from 7.6 × 10-5 M to a 7.6 × 10-8 M concentration, and the angiotensin II-aequorin ctz i was diluted from 2.2 × 10-6 M to a 2.2 × 10-13 M concentration. A volume of 50 µL of each dilution was added to a microtiter plate well, and 100 µL of buffer D was added just prior to bioluminescence collection. The bioluminescence signal was collected over a 25 s time interval under the conditions previously listed, and all measurements listed are the mean of triplicate measurements. Several of the more concentrated semisynthetic construct dilutions overloaded the photomultiplier tube detector in the Polarstar luminometer and were therefore excluded from the graphs. The two time channels were separated for both aequorin constructs, and the sum of the RLUs from 0-6 and 6.01-25 s were reported separately for both 6-keto-PGF1Raequorin-ctz ntv and angiotensin II-aequorin-ctz i. Calibration Plots of Semisynthetic Aequorin Constructs. The 6-keto-PGF1R-aequorin-ctz ntv and angiotensin II-aequorinctz i were serially diluted with buffer A over a range of (2.2 × 10-6)-(2.2 × 10-11.5) M angiotensin II-aequorin-ctz i and (7.6 × 10-8)-(7.6 × 10-13.5) M 6-keto-PGF1R-aequorin-ctz ntv. A volume of 50 µL of each dilution was added to a microtiter plate well, and 100 µL of buffer D was injected just prior to bioluminescence light collection using the Polarstar luminometer. Light was collected over a 5 s time period, and the mean of triplicate measurements was plotted. Binder Dilution Curves of Semi-Synthetic Aequorin Constructs. Rabbit anti-6-keto-PGF1R antiserum and rabbit antiangiotensin II antibodies were serially diluted in 100 mM Na2HPO4, 150 mM NaCl, 2 mM EDTA, 1 mg/mL BSA, pH 7.5 buffer (buffer B). The anti-6-keto-PGF1R antiserum was purchased from a proprietary kit from Assay Designs (Ann Arbor, MI) and suspended with BSA, so that the exact concentration of antibodies is unknown. The angiotensin II antibodies were purchased at an original concentration of 0.15 ng in 0.15 mL. Microtiter plates were purchased whose wells were precoated with Fc-specific antirabbit secondary antibodies. These 96 well secondary antibody coated plates were utilized in order to bind both the human angiotensin II and human 6-keto-PGF1R specific antibodies, both of which originated from rabbit. The wells of these secondary antibody coated microtiter plate were washed 3× with 100 µL of 30 mM Tris, 150 mM NaCl, 2 mM EDTA, 0.05% Tween-20 wash buffer (buffer C). A volume of 50 µL of the anti-6-keto-PGF1R antiserum was pipetted into a well of the antibody coated microtiter plate and incubated with shaking at 4 °C for 12 h. The antiserum solution was then dumped, the wells were washed 3× with 100 µL of buffer C, and 50 µL of the 7.6 × 10-10 M concentration of the 6-keto-PGF1R-aequorin-ctz ntv construct was then added to the well. The plate was incubated with orbital shaking at 4 °C for an additional 12 h. The plate was then dumped, and the well was washed 3× with 100 µL of buffer C. Finally, 100 µL of buffer D was injected into the well immediately prior to the collection of bioluminescence light emission. Light was collected on the Polarstar luminometer for 5 s. All reported measurements are the average of three replicates. The identical procedure was followed for the angiotensin II-aequorin-ctz i binder dilution study, except

a range from 1 × 10-4 to 1 × 10-8 µg/mL concentration of the antiangiotensin II Ab and a 2.2 × 10-9 M concentration of the angiotensin II-aequorin-ctz i construct was added to the wells instead of the 6-keto-PGF1R counterparts. Dose-Response Curves of Semisynthetic Aequorin Constructs at Discrete Time Intervals. The wells of a goat antirabbit IgG coated plate were washed 3× with 200 µL of buffer C. A volume of 50 µL of the 1 × 10-1 dilution of the 6-keto-PGF1R antiserum and 50 µL of the 1 × 10-5 µg/mL of the angiotensin II Ab, both diluted in buffer B, was added to the wells and allowed to incubate with shaking for 12 h at 4 °C. The wells were then washed 3× with 200 µL of buffer C. For the 6-keto-PGF1R dose-response curve, 50 µL of 7.6 × 10-10 M 6-keto-PGF1Raequorin-ctz ntv construct was added to the well with a dilution of a 6-keto-PGF1R standard diluted in buffer A ((2.2 × 10-7)-(4.2 × 10-10) M). This was incubated with shaking for 12 h at 4 °C. The wells were then washed 3× with 200 µL of buffer C. A volume of 100 µL of buffer D was injected into the well immediately prior to light collection. Light was collected using the Polarstar luminometer in the manner described in the Discrimination of Semi-Synthetic Aequorin Constructs at Two Discrete Time Intervals: Dilution Study section of the Materials and Methods section. Points in the dose-response curves at both discrete time intervals (0-6 s and 6.01-25 s) are the mean of five replicates. Outliers were rejected on the basis of the q-test, CL ) 80%. The angiotensin II dose-response curve was constructed in the same manner, using 50 µL of the angiotensin II standard dilution ((9.7 × 10-6)-(9.7 × 10-12) M) and 50 µL of 2.2 × 10-9 M angiotensin II-aequorin-ctz i construct instead of the 6-keto-PGF1R counterparts. This concentration range of angiotensin II standard was selected based on previous studies in our laboratory in developing a solitary aequorin-based angiotensin II assay, and the 6-ketoPGF1R standard concentration range was selected because this range was given as the optimal range in the kit from which the 6-keto-PGF1R antiserum and standard were purchased and corresponded to physiologically relevant concentrations for human plasma analysis. Simultaneous Discrimination of Two Signals from Semisynthetic Aequorin Constructs: Spiking Study. For this study, 200 µL of buffer C was first added to the microtiter plate wells of a goat antirabbit IgG coated plate 3× and dumped. Next, a volume of 50 µL of 1 × 10-1 6-keto-PGF1R antiserum and 50 µL of 1 × 10-5 µg/mL angiotensin II antibody was added to the wells and allowed to incubate at 4 °C with orbital shaking for 12 h. The wells were then washed 3× with 200 µL of buffer C and 50 µL of 2.7 × 10-8 M 6-keto-PGF1R standard, 50 µL of the 7.6 × 10-10 M 6-keto-PGF1R-aequorin-ctz ntv construct, and 50 µL of serial angiotensin II-aequorin-ctz i construct dilution ((1 × 10-10)-(1 × 10-12) M) was then added to the wells. These solutions were allowed to incubate at 4 °C for 12 h with orbital shaking. The solutions were then dumped, and the wells were washed 3× with 200 µL of buffer C. A volume of 100 µL of buffer D was then injected into the well just prior to bioluminescence measurement on the Polarstar luminometer. Light was collected for 25 s, at 0.1 s intervals, and the two discrete time intervals (0-6 and 6.01-25 s) were separated and plotted as described previously. All reported values were the mean of triplicate measurements. The control contained no angiotensin II-aequorin-ctz i construct.

RESULTS AND DISCUSSION The analytes 6-keto-PGF1R and angiotensin II were chosen for this study because they are both important cardiovascular hormones which would be amenable to incorporation as a clinically relevant panel on a multiplexed “lab-on-a-chip” or “labon-a-CD” diagnostic platform.20,21 The overall purpose of diversifying the spectral properties of bioluminescent proteins, such as aequorin, is to facilitate the practical development of precisely such multiplexed panels. The rapid signal and low detection limits of bioluminescent proteins gives them a significant advantage over fluorescent and enzymatic labels when developing such miniaturized, high-throughput devices, and the advantages of bioluminescent reporters would be even further enhanced if the signal of several bioluminescent labels could be discretely and easily resolved from one another. Additionally, aequorin-based immunoassays have been previously developed for these two analytes, indicating their suitability for a semisynthetic aequorin based multianalyte assay.17,22 The detailed methods utilized to create and purify the 6-keto-PGF1R-aequorin conjugate and the angiotensin II-aequorin fusion peptide can be found elsewhere (see refs 17 and 20). In brief, 6-keto-PGF1R was conjugated to a cysteinefree aequorin mutant through the aequorin’s lysine residues using NHS-ester conjugation chemistry, and the angiotensin II-aequorin fusion peptide was created by genetically fusing the sequence coding for angiotensin II to the N terminus of a cysteine free aequorin mutant via a six amino acid spacer using primers and PCR. Both of the aequorin-analyte constructs were chromatographically purified with an ion exchange column followed by a butyl sepharose column. The purity of the proteins subsequently utilized in this study was >95% as verified by SDS-PAGE (see Supporting Information). The aequorin constructs were charged with two different coelenterazine analogues rendering the semisynthetic aequorin conjugates. Specifically, the 6-keto-PGF1Raequorin was charged with ctz ntv, and the angiotensin II-aequorin was charged with ctz i. Previous studies have shown that aequorin derivatives charged with ctz i exhibit a red-shifted emission maxima and a much longer decay half-life than aequorin derivatives charged with ctz ntv.14 The replacement of native coelenterazine’s hydroxyl group with the halogen iodine in ctz i is believed to be the underlying cause of these spectral and decay half-life shifts. Specifically, structural studies undertaken on wild type aequorin showed that the larger size of the halogen in the C2 position causes the C2 of the coelenterazine to insert itself in between the helices of EF hand 3 and 4 of aequorin. It has been hypothesized that this insertion interferes with the Ca2+-induced conformational change, since Ca2+ binds to the loop connecting the EF hands, thereby decreasing the rate of the bioluminescence reaction and increasing the decay half-life time. Therefore, for the first step of this study we determined the decay half-life of both aequorin-analyte conjugates and analyzed whether or not their decay half-life was constant in conditions that may be encountered in bioanalytical and imaging applications. The decay half-life time, and bioluminescent activity, of the aequorinanalyte conjugates was determined at different pH levels and at (20) Pace-Asciak, C. Mol. Neurobiol. 2005, 32 (1), 19–26. (21) Smith, D. Am. J. Hypertension 2002, 15, 108S–114S. (22) Desai, U. A.; Deo, S. K.; Hyland, K. V.; Poon, M.; Daunert, S. Anal. Chem. 2002, 74, 3892–3898.

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

8473

Figure 2. Bioluminescence decay kinetics profiles of 6-keto-PGF1Raequorin-ctz ntv (black line) and angiotensin II-aequorin-ctz i (gray line) conjugates showing the flash versus glow-type decay kinetics of the two conjugates. The graph demonstrates how all of the bioluminescent signal of the 6-keto-PGF1R-aequorin-ctz ntv conjugate is expired by the end of the 0-6 s time channel, leaving only the angiotensin II-aequorin-ctz i signal in the 6.01-25 s time channel.

varying concentrations. The pH level study was undertaken because of the various pH levels the semisynthetic aequorinanalyte constructs might encounter when utilized in biosensors and in imaging.23 For this study, the 6-keto-PGF1R-aequorin-ctz ntv construct and the angiotensin II-aequorin-ctz i construct were both suspended in buffer A. The pH of buffer A was altered with concentrated HCl and 5 M NaOH in order to make a total of nine solutions, of pH 5.0, 5.5, 6.0, 6.5, 7.0, 7.5, 8.0, 8.5, and 9.0. The bioluminescence decay emission profile of the constructs was then measured over a 25 s time period, at 0.1 s intervals. The decay half-life of the constructs was determined from this decay profile by curve fitting with a one phase exponential decay equation, and since the decay exhibited first order kinetics, t1/2 ) 0.693/k was utilized for half-life determination. No significant change in the decay half-life occurred in either conjugate across this pH range (Supporting Information, Table 1). As a result of the destabilizing and denaturing effects of extremely basic or acidic pH on the tertiary structure of proteins, the bioluminescence activity of the aequorin conjugates did decrease at the more extreme ends of the pH range (data not shown). However, the overall stability of the aequorin conjugates’ signals throughout such a large pH range indicated that these labels are attractively robust. In order to demonstrate the discrimination of the two semisynthetic aequorin conjugates with time resolution, a proof of principle dual analyte assay was developed next. Two discrete time intervals were selected based on the difference in the decay kinetics profile of the two constructs. As can be seen in Figure 2, after 6 s all the bioluminescence signal emitted by the 6-ketoPGF1R-aequorin-ctz ntv conjugate has disappeared, and only the angiotensin II-aequorin-ctz i conjugate exhibits bioluminescence during the 6.01-25 s time interval. These two discrete time intervals were therefore selected for further study. For a preliminary test, 6-keto-PGF1R-aequorin-ctz ntv and angiotensin IIaequorin-ctz i were logarithmically diluted with buffer A, and the (23) Shrestha, S.; Deo, S. Anal. Bioanal. Chem. 2006, 386 (3), 515–524.

8474

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

Figure 3. Dilution study of semisynthetic constructs: Angiotensin II-aequorin-ctz i and 6-keto-PGF1R-aequorin-ctz ntv conjugates were logarithmically diluted and the total bioluminescence signal of the decay curve summed for both time channels, the 0-6 s time interval (black line) and the 6.01-25 s time interval (gray line). No significant change in the bioluminescence signal in the 6.01-25 s time interval occurred for the 6-keto-PGF1R-aequorin-ctz ntv, unlike the angiotensin II-aequorin-ctz ntv conjugate. All reported values are the mean of triplicate measurements ( 1 standard deviation, values on x axis ) n.

bioluminescence intensity of the two conjugates was determined at the two discrete time intervals (0-6 and 6.01-25 s). The results depicted in Figure 3 show that the bioluminescence signal of 6-keto-PGF1R-aequorin-ctz ntv changes in a dose-dependent manner only in the 0-6 s time interval, whereas there is no change in signal, regardless of concentration, in the 6.01 to 25 s time interval. However, the bioluminescence intensity of the angiotensin II-aequorin-ctz i conjugate increases with increasing concentrations in both time intervals. The bioluminescence assays for 6-keto-PGF1R and angiotensin II were then developed using goat antirabbit antibody precoated microtiter plates. This type of antibody coated plate was selected because there were rabbit Fc-specific antibodies or antiserum commercially available for both 6-keto-PGF1R and angiotensin II. The schematic of the assay can be seen in Figure 1. Both 6-ketoPGF1R antiserum and angiotensin II antibodies are first allowed to bind to the goat antirabbit secondary antibodies that are coated on the microtiter plate wells. The two conjugates, 6-keto-PGF1Raequorin-ctz ntv and angiotensin II-aequorin-ctz i then compete for these antibody sites with native, free 6-keto-PGF1R and angiotensin II. The two signals from a single well are discriminated from one another using time resolution, with the angiotensin II signal being calculated from the 6.01-25 s time interval and the 6-keto-PGF1R signal being calculated from the 0-6 s time interval. Calibration curves (Supporting Information, Figure 1) were first obtained for both constructs in order to determine the minimum concentration of 6-keto-PGF1R-aequorin-ctz ntv and angiotensin II-aequorin-ctz i that exhibited a bioluminescence signal significantly over the blank signal. From the results of the calibration curves a 7.6 × 10-10 M concentration of 6-keto-PGF1R-aequorinctz ntv and a 2.2 × 10-9 M concentration of angiotensin II-aequorinctz i was then selected for further studies. Binder dilution curves were next obtained for both constructs in order to determine the minimum amount of antibody or antiserum needed to bind the 6-keto-PGF1R-aequorin-ctz ntv and angiotensin II-aequorin-ctz i (Supporting Information, Figure 2). For the binder dilution studies, both antibodies, for 6-keto-PGF1R and angiotensin II, were first serially diluted. Next, 50 µL of 6-keto-PGF1R-aequorin-ctz ntv or angiotensin II-aequorin-ctz i was added to the wells containing the serially diluted antibodies and allowed to bind for 12 h.

Figure 4. Dose-response curves for 6-keto-PGF1R, showing the lack of dose response in the 6.01-25 s time channel and the expected dose response in the 0-6 s time channel. All reported values are the mean of five measurements ( 1 standard deviation.

Following incubation and washing to remove all excess, unbound substrates, the bioluminescence intensity of the wells were obtained following calcium injection. The 6-keto-PGF1R binder dilution curve did not plateau, indicating a saturation of antibody binding, probably due to a relatively dilute initial sample (as stated in the Materials and Methods, the antiserum was purchased from a kit and was not very concentrated). Nonetheless, at 1 × 10-1 and 1 × 10-2 6-keto-PGF1R antiserum dilutions, a significant increase in bioluminescence signal could be detected above the blank, so that the 1 × 10-1 6-keto-PGF1R antiserum concentration was selected for subsequent dose-response curves. The increased concentration of the angiotensin II antibody allowed the angiotensin II-aequorin-ctz i to produce the more traditional binder dilution curve shape, and the 1 × 10-5 µg/mL angiotensin II antibody concentration was selected for subsequent studies. In order to obtain dose-response curves for both analytes and ensure that the time resolution was functioning correctly, separate curves were obtained for both analytes, at both time intervals. For the dose-response curves, both human specific 6-keto-PGF1R antiserum and human specific angiotensin II antibodies were first allowed to bind to the secondary antibody coated microtiter plate wells for 12 h at 4 °C, with 50 rpm orbital shaking. Following this incubation, the wells were washed 3× with 100 µL of buffer C in order to remove all unbound antibodies. Next, the 6-keto-PGF1Raequorin-ctz ntv construct and a serial dilution ((2.2 × 10-7)-(4.2 × 10-10) M) of a 6-keto-PGF1R standard was added to the wells and allowed to incubate at 4 °C, 50 rpm, for 12 h. The wells were then washed 3× with 100 µL of buffer B in order to remove all unbound moieties. Following the injection of 100 µL of buffer D, the bioluminescence profile was collected for 25 s, at 0.1 s intervals. The two discrete time intervals were then separated by summing all points from 0-6 s for interval 1 and all points from 6.01-25 s for interval 2. As can be seen in Figure 4, a dose response was obtained for 6-keto-PGF1R only during the first time interval, which was as expected since 6-keto-PGF1R-aequorin-ctz ntv has a decay half-life of only 0.5 s. The same protocol was used for the angiotensin II-aequorin-ctz i, except angiotensin II-aequorinctz i and a serial dilution of an angiotensin II standard ((9.7 × 10-6)-(9.7 × 10-12) M) were added to the wells instead of the 6-keto-PGF1R counterparts. Figure 5 shows, as theoretically expected, that there is a response for angiotensin II in both the

Figure 5. Dose-response curves for angiotensin II, showing the expected dose response in both the 0-6 s and 6.01-25 s time channels. All reported values are the mean of five measurements ( 1 standard deviation.

first and second time intervals, which is logical since the angiotensin-aequorin-ctz i conjugate has a decay half-life of over 11 s. The dose-response curves indicate that the two signals of angiotensin II-aequorin-ctz i and 6-keto-PGF1R-aequorin-ctz ntv can, indeed, be discriminated from one another with time resolution. However, in order to demonstrate that both signals can be discriminated from each other simultaneously, and within the same microtiter plate well, the following spiking study was performed. Wells were prepared in the exact same manner as previously stated in the generation of the 6-keto-PGF1R doseresponse curve. However, in the last step, a 50 µL volume corresponding to a serial dilution of the angiotensin II-aequorinctz i conjugate ((2.2 × 10-10)-(2.2 × 10-12) M) was added to the wells in addition to the 6-keto-PGF1R-aequorin-ctz ntv conjugate and 6-keto-PGF1R standard (2.7 × 10-8 M). The decay curve was separated into two discrete time intervals again, and the bioluminescence in the 6.01-25 s time interval this time increased in a dose-dependent manner in response to increasing concentrations of the spiked angiotensin II-aequorin-ctz i conjugate (Supporting Information, Figure 3). The results of these studies indicate how two analytes, 6-keto-PGF1R and angiotensin II, can be simultaneously detected in a single well by temporally resolving the bioluminescent decay profiles of two distinct semisynthetic aequorin conjugates. It is important to note that for labels as the ones used in this work where the decay kinetics are significantly different, there is no need for complex deconvolution. However, in cases where the decay kinetics are closer, then signal deconvolution algorithms can be used, so that the short decay component of the long decay conjugate can be reliably subtracted from the short decay conjugate. Please see Supporting Information for a more complete mathematical description concerning the isolation and subtraction of the long decay component from the short decay component necessary for simultaneous signal resolution. CONCLUSIONS Miniaturization and multiplexing are becoming increasingly relevant in analysis. As such, affordable and robust labels are required that satisfy the needed sensitivity and discrimination Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

8475

requirements. This study demonstrates how the ultrasensitive bioluminescent protein aequorin can be used to simultaneously detect two analytes in a single well using time resolution. This type of resolution posits some advantages over bioluminescent spectral resolution due to the simplicity of detection equipment and the lack of spectral cross-talk and concomitant data analysis necessary for signal discrimination. We envision that this type of bioluminescent signal resolution will be utilized to develop, for example, assay arrays which can simultaneously detect multiple, clinically relevant molecules. Moreover, this time resolution can be combined with the spectral resolution of a variety of semisynthetic aequorin conjugates in order to further increase the multiplexing capability of bioluminescent proteins. In addition to miniaturization and assay development in vivo temporal resolution of bioluminescent reporters may prove useful for cellular imaging technologies, as an alternative to the extremely rapid decay kinetics of current fluorescence in vivo imaging methods. The ability of the bioluminescent reporters to be expressed inside of cells in a controlled manner, instead of requiring injection like many fluorescent reporters used in time-resolved in vivo imaging, is a distinct advantage for the analysis of low-abundance and transient molecules. Thus, the evidence provided herein demonstrating the temporal resolution of two aequorin constructs

8476

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

expands the labeling potential of bioluminescent proteins to include both in vitro and in vivo simultaneous multianalyte detection. ACKNOWLEDGMENT This work was supported in part by the National Institutes of Health (Grant GM 467917). S.D. is grateful to the Office of the Vice-President for Research at the University of Kentucky for a University Research Professorship. S.D. is also thankful for a Gill Eminent Professorship. We would like to acknowledge and thank Nitin Chopra for his guidance concerning signal deconvolution in the Supporting Information. L.R. is indebted to Predoctoral Fellowships from the National Science Foundation-IGERT and the National Institutes of Health. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

Received for review June 13, 2008. Accepted September 9, 2008. AC801209X