Computational Prediction and Experimental Evaluation of a


Nov 2, 2006 - A survey of analysis software for array-comparative genomic hybridisation studies to detect copy number variation. Anis Karimpour-Fard ,...
1 downloads 11 Views 169KB Size


22991

2006, 110, 22991-22994 Published on Web 11/02/2006

Computational Prediction and Experimental Evaluation of a Photoinduced Electron-Transfer Sensor Matthew E. McCarroll,* Yu Shi, Samantha Harris, Surendra Puli, Irene Kimaru, Ruisong Xu, Lichang Wang,* and Daniel J. Dyer* Department of Chemistry and Biochemistry, Southern Illinois UniVersity, Carbondale, Illinois 62901-4409 ReceiVed: September 8, 2006; In Final Form: October 12, 2006

An approach is presented for the design of photoinduced electron-transfer-based sensors. The approach relies on the computational and theoretical prediction of electron-transfer kinetics based on Rehm-Weller and Marcus theories. The approach allows evaluation of the photophysical behavior of a prototype fluorescent probe/sensor prior to the synthesis of the molecule. As a proof of concept, a prototype sensor for divalent metal ions is evaluated computationally, synthesized, and then analyzed spectroscopically for its fluorescence response to zinc. Calculations predicted that the system would show a competition between electron transfer and fluorescence in the free state. In the zinc-bound state, the compound was predicted to be more highly fluorescent, due to the inhibition of electron transfer. Both predictions were confirmed experimentally. A nonzero fluorescence signal was observed in the absence of zinc and an enhancement was observed in the presence of zinc. Specifically, a 56-fold enhancement was observed over a 10-fold increase in zinc concentration.

Introduction Fluorescence-based sensors and probes can offer significant fundamental advantages over other sensing techniques. A prime example is the pioneering development of calcium probes by Tsein, which have been enormously successful by allowing determination of cellular Ca2+ concentrations with good detection limits and spatial resolution.1 Despite many successes, the development of fluorescent probes and sensors has significantly lagged behind the need for such sensors. One reason for this is the often labor intensive synthetic work that typically follows a design process, often driven by empirical and intuitive considerations. While these approaches are often successful, they are arguably slow and not always predictable. Even more challenging is the development of sensors for molecular species that typically have more complex and often weaker interactions. One effective approach, well-demonstrated by Anslin’s group, is that of using combinatorial chemistry.2-4 The need for accurate sensing of molecular species is ever present, especially with the recent advances made in proteomics and metabonomics.5 Thus, there is wide interest in the development of new design and discovery approaches that are more direct and effective. In this regard we have set out to develop a predictive design approach that relies on combined efforts of computational, synthetic, and analytical chemistries. Recent efforts in our laboratories have focused on developing a systematic approach to developing fluorescence sensors based on photoinduced electron transfer (PET) as a transduction mechanism. Herein we describe the computational modeling, prediction, synthesis, and analytical evaluation of a prototype PET sensor for divalent metal ions. Most PET sensors are based * Corresponding authors. E-mail: (McCarroll) [email protected]

10.1021/jp065876s CCC: $33.50

on a fluorophore-spacer-receptor architecture, where the electron receptor also possesses a recognition site that binds to the analyte. The fluorophore moiety is responsible for the fluorescence signal, which is modulated based on whether the recognition site is occupied. When an electron in the fluorophore frontier orbital is promoted to an excited electronic state by the absorption of a photon, the molecule can return to the ground state either by emission of a photon (fluorescence) or by nonradiative electron transfer to the receptor moiety. Photoinduced electron-transfer-based sensors are typically designed such that electron transfer is the favored deactivation route when the recognition site is unoccupied. When the recognition site is occupied, the receptor molecular orbital (MO) is perturbed such that the electron transfer is no longer favorable and fluorescence becomes the favored process (i.e., “switch-on”). In a welldesigned system, the spectral properties of the fluorescence process are not affected by the binding event since the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) of the fluorophore remain unchanged and only the fluorescence quantum yield is affected, which can be a significant advantage in the development of quantitative sensing systems. An exciting aspect of sensors based on PET is the fact that the transduction mechanism is well understood on a theoretical basis, and is, in principle, predictable.6-8 While many other factors can impact the effectiveness of a PET sensor, a fundamental challenge lies in choosing a system that will undergo efficient electron transfer in the free state and exhibit strong fluorescence properties in the bound state. Molecular modeling is capable of predicting the energy level and geometry of molecular orbitals, and in this work we set out to evaluate the effectiveness of a computational approach in PET sensor design. The sensor used in this study consists of an anthracene © 2006 American Chemical Society

22992 J. Phys. Chem. B, Vol. 110, No. 46, 2006

Figure 1. Operation of a PET sensor (Compound 1) in the (a) free state, and (b) bound state with zinc. The HOMO/LUMO energy levels were calculated using B3LYP with a basis set of 6-31G.

fluorophore tethered to a pyridine receptor via an alkylamino bridge (Figure 1). Sensor 1 was expected to bind to divalent metal ions, which would then alter the HOMO and LUMO levels of the receptor sufficiently to selectively modulate the PET process and turn the fluorescence on and off. Figures 1a and 1b show the results of calculations of the energy levels for Compound 1. These results were obtained using B3LYP with a basis set of 6-31G(d,p), which was implemented in Gaussian03.9 In the absence of zinc, energy level calculations indicate the thermodynamic feasibility of fluorescence switching via a PET mechanism. Specifically, following promotion of the fluorophore electron to the LUMO (-1.65 eV), the receptor HOMO (-5.236 eV) could transfer to the slightly lower fluorophore HOMO (-5.239 eV), which is then followed by electron transfer from the fluorophore LUMO to the receptor HOMO. Upon binding to zinc, the HOMO and LUMO levels of the receptor are both significantly shifted to lower energy levels, such that electron transfer is not favorable and the sensor is switched on (Figure 1b). To test these predictions, Compound 1 was synthesized (see Supporting Information) and evaluated experimentally. The spectral and binding properties of Compound 1 were evaluated in a number of solvent systems. The system was found to behave erratically in very polar systems, such as buffered aqueous solutions and methanol/water. This is likely attributable to the limited solubility of Compound 1 and the complex equilibria that may have resulted. The response that was observed was also limited, due to competitive interactions with the solvent system. Compound 1 was found to behave much more predictably in a predominately acetone solvent system, which increases both the solubility and binding in the system due to the aprotic nature of the solvent. Figure 2a shows the fluorescence spectra of Compound 1 at zinc concentrations ranging from 0 to 330 µM. Compound 1 was prepared at a concentration of 6.7 µM in an acetone/methanol solution (199:1). The fluorescence intensity of the system clearly increases as a function of increased zinc concentration, presumably due to the predicted perturbation of electron transfer in the zinc-bound system. It is interesting to note that Compound 1 was fluorescent in the unbound state, indicating that the electron-transfer efficiency was less than unity. This could be understood in simplistic terms by the fact that the receptor HOMO is only slightly higher in energy than the HOMO of the fluorophore.

Letters

Figure 2. Fluorescence excitation and emission spectra of Compound 1 with increasing concentrations of Zn (II) and emission spectrum of 1 in the absence of Zn (II) (inset).

However, the ultimate determining factor of a successful PET sensor is the relative kinetics of the two major competing processes, that is, (a) fluorescence and (b) electron transfer. It should be noted that the above statement assumes that the electron transfer from the receptor’s HOMO back to the flurophore’s HOMO, also denoted as a charge recombination process, is fast. In this work, we estimated the electron-transfer rate for the electron-transfer process using the Marcus equation,10,11

kET )

( ) 4π3 h2λkBT

1/2

[

‚V2‚ exp -

]

(∆G° + λ)2 4λkBT

(1)

where h is Planck’s constant, kB is the Boltzmann constant, and T is the absolute temperature. As our experiments were performed at room temperature, here T ) 298 K. The free energy, ∆G°, was obtained by using the Rehm-Weller equation,12-14

∆G° ) E°(D+/D) - E°(A/A -) - ∆E0,0 + w

(2)

where E°(D+/D) is the oxidation potential of the electron donor, E°(A/A-) is the reduction potential of the electron acceptor. Here E°(D+/D) - E°(A/A-) was approximated by the HOMO and LUMO energy difference, i.e., 3.586 eV. The excitation energy, ∆E0,0, is taken as 3.589 eV. These values were obtained from B3LYP calculations using a basis set of 6-31G(d,p). The work required to bring the donor and acceptor together, w, was estimated by

w)

1 e2 ‚ 4π0 R

(3)

where 0 and  are the dielectric constant for the vacuum and the acetone solvent, respectively. The value of  of acetone is taken as 20.7.15 Furthermore, e is the electron charge and R is the distance between the donor and acceptor. Using the estimated R value of 5.08 Å, we obtained w as 0.137 eV. Therefore, the estimated free energy, ∆G°, is 0.134 eV. A positive free energy usually indicates that the electron-transfer process is not spontaneous. Hence, this small positive value for ∆G° indicates that electron transfer may not be the completely favored mode of deactivation in this system. In fact, these predictions are in agreement with the fact that Compound 1 is not completely

Letters

J. Phys. Chem. B, Vol. 110, No. 46, 2006 22993

nonfluorescent in the free (e.g., off) state. Thus, the overall fluorescence intensity (quantum yield) is determined by the competing kinetic rates of the fluorescence and electron-transfer processes. Therefore the electron-transfer rate must be fully considered for use as a prescreening tool in sensor design, rather than the more simplistic approach of comparing relative energy levels. The reorganization energy, λ, in eq 1 is estimated by16

λ≈

(

)(

)

e2 1 1 1 1 1 + 4π0 2r1 2r2 R n2 

(4)

where r1 and r2 are the radii of the electron donor and acceptor, respectively. Here, n is the refractive index of the acetone solvent, which is 1.3590.15 Finally, the electronic coupling matrix element, V, in eq 1 was estimated using the generalized Mulliken-Hush method,17

V)

µ12∆E12 (∆µ122 + 4µ122)1/2

(5)

where µ12 is the transition dipole moment connecting the two adiabatic states in the electron transfer, ∆E12 is the energy difference between the initial and the final adiabatic states, and ∆µ12 is the dipole moment difference between the adiabatic states. We assumed the dipole moment difference is small with respect to the transition dipole moment, therefore the coupling matrix element V becomes half the energy difference between two adiabatic states, ∆E12. Using the Marcus equation and approximating ∆E12 to be in the range of 0.01 to 1.0 eV, we estimated the electron-transfer rate would be in the range of 0.0000205 to 0.205 ns-1. Based on the fluorescence lifetime of anthracene and various derivatives,18 we estimated the rate of fluorescence, the inverse of fluorescence lifetime, to be about 0.083-0.143 ns-1. Thus, the fluorescence rate lies within the boundaries of the estimated electron-transfer rates and the two rates are competitive deactivation processes. Therefore, a balance between electron transfer and fluorescence is predicted in the free state. When Compound 1 is in the bound state, however, both the HOMO and LUMO energy levels decrease such that electron transfer is no longer energetically feasible and fluorescence should become the favored process. Experimental data shown in Figure 2 demonstrate the predicted response. In the absence of zinc (Figure 2, inset) the fluorescence is at a minimum, yet is still observed, indicating competition between electron transfer and fluorescence. The addition of Zn2+, however, results in a marked increase due to inhibition of the electron-transfer pathway. The response of the system is more clearly shown in Figure 3. A 56-fold increase was observed, with the majority of the increase occurring over a 10-fold change in concentration. The inset of Figure 3 shows the evaluation of a modified double reciprocal plot to evaluate the strength of binding,19 which was found to be on the order of 3 × 105 M-1 in this system. In this report we have shown the potential of using a computationally based approach in the design of PET based sensors. While the current state of our approach does not preclude a certain amount of trial and error in the development of effective sensors, the potential savings in the time and resources consumed in the synthesis phase of development is significant. It is also apparent from our approach that a comprehensive assessment of the excited-state kinetics must be considered in the design of PET based sensors, rather than the

Figure 3. Fluorescence response curve of 1 to the presence of zinc and the modified double reciprocal plot to evaluate the binding strength (inset).

more simplistic consideration of HOMO/LUMO energy levels. Namely, the electron-transfer rate and the fluorescence rate must be considered to predict the behavior. Most importantly, these are quasi-independent parameters that can be considered in the design of effective PET based sensors. For example, by choosing a fluorescent moiety with a short or long fluorescence lifetime, the balance between radiative luminescence and nonradiative electron transfer can be tuned, offering a new degree of control in the design of PET-based sensors and probes. One aspect of sensing that we have not discussed is selectivity. While this sensor does exhibit some selectivity over competing cations, we are not yet able to predict this. Research in our group is investigating methods for predicting both fluorescence response and selectivity. Acknowledgment. The authors gratefully acknowledge partial support from the donors of the Petroleum Research Fund, administered by the American Chemical Society (PRF # 39264G4 (M.E.M.), the National Science Foundation (CHE-0421012) and the SIU Materials Technology. Additional support was provided by Southern Illinois University in the form of start up funds and the faculty Seed Grant Program. Supporting Information Available: Additional information, synthetic scheme, characterization data, and experimental details. This information is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Minta, A.; Kao, J.; Tsien, R. J. Biol. Chem. 1989, 264, 81718178. (2) Schneider, S. E.; O’Neil, S. N.; Anslyn, E. V. J. Am. Chem. Soc. 2003, 122(3), 542-543. (3) McCleskey, S. C.; Griffin, M. J.; Schneider, S. E.; McDevitt, J. T.; Anslyn, E. V. J. Am. Chem. Soc. 2003, 125, 1114-1115. (4) Cabell, L. A.; Best, M. D.; Lavigne, J. J.; Schneider, S. E.; Perreault, D. M.; Monahan, M.-K.; Anslyn, E. V. J. Chem. Soc., Perkin Trans. 2 2001, 3, 315-323. (5) Lindon, J. C.; Holmes, E.; Nicholson, J. K. Anal. Chem. 2003, 75(17), 384A-391A. (6) Stuchebrukhov, A. A.; Marcus, R. A. J. Phys. Chem. 1995, 99, 7581-7590. (7) Lukas, A. S.; Bushard, P. J.; Weiss, E. A.; Wasielewski, M. R. J. Am. Chem. Soc. 2003, 125, 3921-3930. (8) de Silva, A. P.; McCaughan, B.; McKinney, B. O. F.; Querol, M. Dalton Trans. 2003, 1902-1913. (9) Frisch, M. J.; G. W. T., Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,

22994 J. Phys. Chem. B, Vol. 110, No. 46, 2006 R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Salvador, P.; Dannenberg, J. J.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.; Replogle, E. S. and Pople, J. A. Gaussian 98, Revision A.11; Guassian, Inc.: Pittsburgh, 2001. (10) Fukuzumi, S. Org. Biomol. Chem. 2003, 1, 609-620. (11) Marcus, R. A. Ann. ReV. Phys. Chem. 1964, 15, 155-196. (12) Croney, J. C.; Helms, M. K.; Jameson, D. M.; and Larsen, R. W. Biophys. J. 2003, 84, 4135-4143.

Letters (13) Rau, H.; Frank, R.; and Greiner, G. J. Phys. Chem. 1986, 90, 24762481. (14) Rehm, D. a. W. A. Isr. J. Chem. 1970, 8, 259-271. (15) Lide, D. R. CRC Handbook of Chemistry and Physics, 72nd ed.; CRC Press: Boca Raton, 1991. (16) Fournier, T.; Travender, S. M.; Parker, A. W.; Scholes, G. D.; Philips, D. J. Phys. Chem. A 1997, 101, 5320. (17) Rust, M. J. L.; Cave, R. J. J. Phys. Chem. A 2002, 106, 39303940. (18) Lin, H.-J.; Herman, P.; Kang, J. S.; Lakowicz, J. R. Anal. Biochem. 2002, 294, 118. (19) Dirr, H. W.; Schabort, J. C.; Weitz, C. Biochem. J. 1986, 233(3), 649-653.