Absorption Spectrum of the Firefly Luciferin Anion Isolated in Vacuo

Feb 25, 2011 - and produces the firefly's bioluminescence. Both luciferin and oxyluciferin are expected to display charge- transfer (CT) character in ...
1 downloads 0 Views 998KB Size
Download PDF
ARTICLE pubs.acs.org/JPCA

Absorption Spectrum of the Firefly Luciferin Anion Isolated in Vacuo Kristian Støchkel,†,* Bruce F. Milne,‡,* and Steen Brøndsted Nielsen† † ‡

Department of Physics and Astronomy, Aarhus University, Ny Munkegade, DK-8000 Aarhus C, Denmark Centre for Computational Physics, Physics Department, University of Coimbra, Rua Larga, 3004-516, Coimbra, Portugal

bS Supporting Information ABSTRACT: The excited-state physics of the firefly luciferin anion depends on its chemical environment, and it is therefore important to establish the intrinsic behavior of the bare ion. Here we report electronic absorption spectra of the anion isolated in vacuo obtained at an electrostatic ion storage ring and an accelerator mass spectrometer where ionic dissociation is monitored on a long time scale (from 33 μs and up to 3 ms) and on a short time scale (0-3 μs), respectively. In the ring experiment the yield of all neutrals (mainly CO2) as a function of wavelength was measured whereas in the single pass experiment, the abundance of daughter ions formed after loss of CO2 was recorded to provide action spectra. We find maxima at 535 and 265 nm, and that the band shape is largely determined by the sampling time interval, which is due to the kinetics of the dissociation process. Calculations at the TD-B3LYP/TZVPPþþ level predict maximum absorption at 533 and 275 nm for the carboxylate isomer in excellent agreement with the experimental findings. The phenolate isomer lies higher in energy by 0.22 eV, and also its absorption maximum is calculated to be at 463 nm, which is far away from the experimental value. Our data serve to benchmark future theoretical models for bioluminescence from fireflies.

’ INTRODUCTION Luciferin (LH2, Figure 1) is the substrate molecule for the enzymatic reaction responsible for the characteristic yellow light emission from fireflies.1 In aqueous solution the carboxylic acid moiety of the molecule is deprotonated at physiological pH, with the anion carrying a unit negative charge. LH2 undergoes a Mg2þmediated reaction with adenosine triphosphate in the active site of the luciferase enzyme, releasing inorganic pyrophosphate. The resulting luciferyl adenylate compound reacts with molecular oxygen, liberating one molecule each of adenosine monophosphate and carbon dioxide. This final step yields the product molecule, oxyluciferin, in a singlet excited state, and it is this excited product molecule that emits a photon of light as it relaxes to the ground state and produces the firefly’s bioluminescence. Both luciferin and oxyluciferin are expected to display chargetransfer (CT) character in the excited state where charge density has been moved between the five-membered thiazolyl ring and the fused benzothiazolyl ring system. Oxyluciferin is extremely unstable and has only recently been synthesized, isolated and studied spectroscopically by Naumov et al.2 In their study, the absorption and fluorescence spectra of oxyluciferin were obtained in a range of solvents. An emission wavelength of 553 nm was obtained for aqueous solution, in good agreement with the Photinus pyralis bioluminescent maximum of 562 nm, while in organic (less-polar) solvents this value was reduced by more than 80 nm. Even though different species of bioluminescent beetles all use the same luciferin, the emission spectrum can vary greatly, e.g., r 2011 American Chemical Society

the light emitted from the firefly Photuris pennsylvanica is green (538 nm) while that emitted by the railroad worm Phrixotrix hirtus is red (623 nm).3 Probably the most studied species is the North American firefly P. pyralis which produces a yellow-green light having a wavelength of 562 nm.3 This variability is most likely linked to alterations in the tertiary structure of the luciferase enzyme arising from interspecies differences in the corresponding amino acid sequence,4 although a number of other factors have been found to play a role in determining the emission color for the luciferase/luciferin reaction including temperature, pH and the presence of divalent metal ions such as Cu2þ.4-6 To better understand the bioluminescence phenomenon and the influence of a chemical environment, detailed theoretical work has been carried out on both luciferin and oxyluciferin, isolated in vacuum or in solvents7-12 or, more recently, embedded in the protein environment.13-16 However, to date there have been no experimental gas-phase data to compare with and thus no reference spectra free of microenvironmental effects with which to evaluate the theoretical methods used. Especially, information on the intrinsic electronic properties of the ligand molecules is relevant as the importance of nearby molecules or charges can then be directly inferred. Received: October 22, 2010 Revised: January 25, 2011 Published: February 25, 2011 2155

dx.doi.org/10.1021/jp110120c | J. Phys. Chem. A 2011, 115, 2155–2159

The Journal of Physical Chemistry A

ARTICLE

laser power at different wavelengths was less than 10%. The configuration of the ion and laser beams was, however, perpendicular. An electrostatic analyzer allowed ions that had lost CO2 (vide infra) to reach a channeltron detector where they were counted.

Figure 1. Structure of the luciferin anion: carboxylate and phenolate isomers.

Here we report the first experimental absorption spectrum of deprotonated luciferin anions in vacuo together with electronic structure calculations. Our spectra serve as a benchmark for future theoretical descriptions since a model for the excited-state physics of oxyluciferin ought to be able to correctly predict excitations in the electronic structure of deprotonated luciferin.

’ EXPERIMENTAL SECTION It is difficult to measure the absorption spectrum of ions in vacuo since there are too few absorbing species to cause a measurable change in light intensity. Instead, absorption is identified from ionic dissociation (i.e., action spectroscopy). Gas-phase absorption spectroscopy on fragile macromolecular ions has been made possible in our laboratory from the combination of an electrospray ion source, an electrostatic ion storage ring, and pulsed lasers; see Figure 2.17,18 Briefly, ions produced by electrospray ionization were accumulated in a 22-pole ion trap. Here they experienced collisions with helium buffer gas kept at room temperature. An ion bunch was accelerated to 22-keV kinetic energies and ions of interest mass-to-charge were selected by a bending magnet. These were injected into the ring that is based on purely electrostatic deflectors and focusing elements. The ions circulated until they changed their mass-to-charge ratio as a result of either collisions with residual gas in the ring (pressure of about 10-10 mbar) or photoexcitation. Neutrals produced on the injection side of the ring are not influenced by the electric fields and hit a microchannel plate detector located at the end of the straight section. The rate of neutrals hitting the detector is a measure of the number of ions circulating in the ring. A small part of the ions was vibrationally excited already when they entered the ring (around 0.5% of the total beam), which is evident from Figure 3a. Luciferin anions were photoexcited after about 36.5 ms to ascertain that these vibrationally excited ions had decayed. The third harmonic (355 nm) from a Nd:YAG laser was used to pump an optical parametric oscillator (OPO) (EKSPLA laser). The visible output from the OPO was frequency doubled in a crystal providing UV light. Neutral density filters were used to attenuate the laser power. The repetition rate of the experiment was 10 Hz. Lifetimes were measured with respect to dissociation. ELISA was also operated as a mass spectrometer by quick switching of ring voltages directly after photoexcitation at 460 and 560 nm.19 At another setup operating at a repetition rate of 20 Hz, the yield of fragment ions as a function of excitation wavelength was monitored up to 3 μs after photoexcitation (see Figure 2). Ions were again accumulated in a multipole ion trap (14 poles). An ion bunch was accelerated to 50-keV energies and appropriate ions were selected by a magnet. These were photoexcited by light from a similar laser system as used at ELISA. The difference in

’ THEORETICAL CALCULATIONS To model the excitations, we carried out electronic structure calculations at the density functional theory (DFT) level of theory. These were performed with version 2.7.0 r1730 of the ORCA electronic structure package obtained from the University of Bonn (http://www.thch.uni-bonn.de/tc/orca/).20 The geometry of the luciferin anion was optimized at the B3LYP level using the augmented polarized double-ζ Karlsruhe basis set (SVPþþ).21-23 Default integration grids were used but more stringent convergence criteria of 10-8 Eh (SCF) and 10-4 Eh Å-1 (gradient) were imposed. Both the carboxylate and the phenolate anions were optimized. Only the low energy trans-N,N rotamer was considered as the cis-conformation has been shown to lie 0.24 eV higher in energy and is therefore not expected to contribute to the experimental spectra in the present work.24 TDDFT calculations of the absorption spectra were performed using the B3LYP exchange correlation functional with the same tightened SCF convergence criteria as used in the geometry optimization.25 This functional was selected as it has seen extensive use in previous work on aspects of firefly bioluminescence.7-12,16 The Karlsruhe triple-ζ TZVPPþþ basis set was chosen for the TDDFT calculations.21-23 The polarization functions provided by ORCA were obtained from the TurboMole basis set library under ftp.chemie.unikarlsruhe.de/ pub/basen. A single diffuse function was included on all atoms. ’ RESULTS AND DISCUSSION A time spectrum obtained after photoexcitation by 535 nm light in the ring is shown in Figure 3b from which it is evident that dissociation is almost complete after 1 ms. The neutrals that gave rise to the count rate were produced ∼33 μs after photoexcitation in the first instance (half a revolution corresponding to the first point in the time spectrum) and then after successive rotations in the ring (revolution time of 67 μs). The yield of photoneutrals increases linearly with the energy of the laser pulse until saturation sets in after about 0.5 mJ per pulse (Figure S1, Supporting Information). In accordance with this, the ion beam depletion also increases linearly with pulse energy until saturation is reached (Figure S2, Supporting Information), which indicates that a single photon is enough to cause the dissociation. At high photon fluxes a significant part of the ions in the interaction region has absorbed a photon, and two-photon absorption may not be negligible. The experiment was therefore performed with pulse energies of less than 0.5 mJ at all wavelengths. Our data indicate that the dissociation monitored is due to the absorption of a single photon. According to mass spectra obtained after either photoexcitation or high-energy (50 keV) collisions, the dominant dissociation pathway is loss of CO2 (Figure S3, Supporting Information). In the high-energy collision experiment an envelope of electronic transitions is accessible, which results in a broad internal energy distribution after internal conversion. The Gibbs free energy change for CO2 loss at 298.15 K (calculated at the B3LYP/TZVPP level of theory) was found to be 0.88 eV, which is less than the energy of one photon and therefore supports the results from the power dependence study. If two photons were absorbed, the dissociation time 2156

dx.doi.org/10.1021/jp110120c |J. Phys. Chem. A 2011, 115, 2155–2159

The Journal of Physical Chemistry A

ARTICLE

Figure 2. (a) Electrospray ion source combined with a linear ion trap. (b) Accelerator mass spectrometer used for single pass experiments. (c) The electrostatic ion storage ring in Aarhus, ELISA. See text for more details.

Figure 3. Time spectra of luciferin anions in ELISA. (a) The first few milliseconds after injection into the ring, decay of vibrationally excited ions dominate. These are probably due to heating in the extraction of ions from the trap. Collisions with residual gas in the ring occur with a time constant of 500 ms. The CID decay is proportional to the number of ions in the ring. Photoexcitation leads to hot ions that dissociate within about 1 ms (b).

constant of such ions would be far too short for them to survive the travel time to the region where dissociation is monitored. The time spectrum is an exponential with some tailing due to the width of the energy distribution and possible population of long-lived triplet states (Figure 3b).26,27 In our analysis described below we use an exponential to account for the tail. A third exponential associated with a time constant of 500 ms is included to describe the decay due to collision-induced dissociation. From exponential fits to the time spectra, it is possible to obtain a number proportional to the absorption cross section.28 The count rate in the MCP detector is the number of neutrals produced within a certain time, that is, it represents a decay rate, -dN*(t)/dt = [N*(t=0)/t*] exp(-t/τ*), where t is the time after photoexcitation. Hence the number of photoexcited ions at time zero, N*(t=0), is simply the preexponential factor multiplied by

the time constant τ*. This number was divided by the neutrals yield prior to photoexcitation and by the number of photons in the laser pulse for each excitation wavelength to correct for variations in ion beam intensity and photon flux. These absorption cross sections are shown in Figure 4 as a function of excitation wavelength, and they make up the action spectrum. This spectrum is identical to the gas-phase absorption spectrum if luminescence does not occur or if its quantum yield is independent of wavelength. A band is evident with a tail to lower wavelengths. A fit to two Gaussians provide absorption band maxima at 500 and 535 nm (band widths of 22 nm). Results from the other setup where the abundance of daughter ions formed after loss of CO2 was measured as a function of wavelength are shown in Figure 4. Again corrections were done for fluctuations in ion and laser beam intensities. The same two Gauss functions as before can be used to fit the spectrum, but their relative importance is changed in favor of the high-energy band. We believe this is due to the fact that at higher photon energies more ions dissociate within the finite sampling time interval. In other words, the dissociation is not complete within 3 μs in accordance with the time spectra recorded at ELISA. Two-photon absorption cannot be completely ruled out in this particular experiment but the effect on the absorption spectrum would be negligible due to low variation in laser power with excitation wavelength. As deprotonation could, in principle, occur at two positions in the molecule, calculations were performed for both the carboxylate and the phenolate anions (see Figure 1), and from the results we find that the carboxylate isomer is lower in energy than the phenolate by 0.22 eV. The two lowest energy transitions were predicted to occur at 533 nm (carboxylate) and 463 nm (phenolate). The theoretical results for the carboxylate are shown as vertical bars in Figure 4, and they are in excellent agreement with experiment. Finally, action spectra were also recorded in the ultraviolet from 210 to 350 nm at ELISA (see Figure 4), and maximum absorption was found to occur at 265 ( 8 nm. A complication in this region is that direct photodetachment can occur since the electron binding energy of the anion is calculated to be 3.76 eV (vertical ionization energy obtained from ΔSCF calculations at the B3LYP/TZVPPþþ level), and the results should therefore 2157

dx.doi.org/10.1021/jp110120c |J. Phys. Chem. A 2011, 115, 2155–2159

The Journal of Physical Chemistry A

ARTICLE

Figure 4. Gas-phase absorption spectra of the luciferin anions measured at ELISA (squares) and the accelerator mass spectrometer SEP1 (open circles). The panel to the left is the spectrum recorded in the ultraviolet region while that to the right is the spectrum in the visible. Gaussian functions have been fit to the data (red curves). The bars represent calculated values for the carboxylate isomer.

be taken with some caution. However, TD-DFT calculations predict this maximum to be at 275 nm in very good agreement with the measurement.

’ CONCLUSIONS In conclusion, the gas-phase absorption spectrum of the luciferin anion has been measured using two different instruments with different time windows for sampling of dissociation. In one experiment, ionic dissociation occurring from 33 μs and up to 3 ms was monitored while in the other the time range was 0-3 μs. Since the time constant for dissociation is longer than 3 μs, the high-energy band is stronger in the latter spectrum relative to the first. Predictions made by TD-B3LYP were found to be in very good agreement with the experimental data, assuming the anion to be in its carboxylate form. More complete theoretical studies are under way to investigate the convergence behavior of TD-DFT calculations on this molecule with regard to choice of functional and basis set. ’ ASSOCIATED CONTENT

bS

Supporting Information. Figures showing power dependence of the laser excitation and photo fragmentation spectrum. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: K.S., [email protected]; B.F.M., bruce@teor.fis.uc.pt.

’ ACKNOWLEDGMENT S.B.N. thanks Lundbeckfonden for financial support. B.F.M. thanks the Portuguese Foundation for Science and Technology for funding (PTDC/FIS/73578/2006). We thank Rui M. M. Brito of the Department of Chemistry, University of Coimbra, Portugal, for providing the luciferin used in this study and Fernando Nogueira of the Centre for Computational Physics, Department of Physics, University of Coimbra, Portugal, for his comments and advice regarding the theoretical portion of this work.

’ REFERENCES (1) Shimomura, O. Bioluminescence. In Chemical Principles and Methods; World Scientific: Singapore, 2006. (2) Naumov, P.; Ozawa, Y.; Ohkubo, K.; Fukuzumi, S. J. Am. Chem. Soc. 2009, 131, 11590. (3) Viviani, V. R.; Arnoldi, F. G. C.; Neto, A. J. S.; Oehlmeyer, T. L.; Bechara, E. J. H.; Ohmiya, Y. Photochem. Photobiol. Sci. 2008, 7, 159. (4) Seliger, H. H.; Buck, J. B.; Fastie, W. G.; McElroy, W. D. J. Gen. Physiol 1964, 48, 95. (5) Lee, R. T.; Denburg, J. L.; McElroy, W. D. Arch. Biochem. Biophys. 1970, 141, 38. (6) Lundovskikh, I.; Dementieva, E.; Ugarova, N. Appl. Biochem. Biotechnol. 2000, 88, 127. (7) Orlova, G.; Goddard, J. D.; Brovko, L. Y. J. Am. Chem. Soc. 2003, 125, 6962. (8) Ren, A.; Goddard, J. D. J. Photochem. Photobiol. B: Biol. 2005, 81, 163. (9) Yang, T.; Goddard, J. D. J. Phys. Chem. A 2007, 111, 4489. (10) Liu, Y.; De Vico, L.; Lindh, R. J. Photochem. Photobiol. A: Chem. 2008, 194, 261. (11) Li, Z.; Ren, A.; Guo, J.; Yang, T.; Goddard, J. D.; Feng, J. J. Phys. Chem. A 2008, 112, 9796. (12) Min, C.; Ren, A.; Guo, J.; Li, Z.; Zou, L.; Goddard, J. D.; Feng, J. ChemPhysChem 2010, 11, 251. (13) Nakatani, N.; Hasegawa, J. Y.; Nakatsuji, H. J. Am. Chem. Soc. 2007, 129, 8756. (14) Nakatani, N.; Hasegawa, J.; Nakatsuji, H. Chem. Phys. Lett. 2009, 469, 191. (15) Tagami, A.; Ishibashi, N.; Kato, D.; Taguchi, N.; Mochizuki, Y.; Watanabe, H.; Ito, M.; Tanaka, S. Chem. Phys. Lett. 2009, 472, 118. (16) Min, C.; Ren, A.; Guo, J.; Zou, L.; Goddard, J.; Sun, C. ChemPhysChem 2010, 11, 2199. (17) Møller, S. P. Nucl. Instrum. Methods Phys. Res. A 1997, 394, 281. (18) Andersen, J. U.; Hvelplund, P.; Brøndsted Nielsen, S.; Tomita, S.; Wahlgreen, H.; Møller, S. P.; Pedersen, U. V.; Forster, J. S.; Jørgensen, T. J. D. Rev. Sci. Instrum. 2002, 73, 1284. (19) Støchkel, K.; Kadhane, U.; Andersen, J. U.; Holm, A. I. S.; Hvelplund, P.; Kirketerp, M.-B. S.; Larsen, M. K.; Lykkegaard, M. K.; Brøndsted Nielsen, S.; Panja, S.; Zettergren, H. Rev. Sci. Instrum. 2008, 79, 023107. (20) Neese, F. ORCA - an ab initio, Density Functional and Semiempirical program package, Version 2.5; University of Bonn: Bonn, Germany 2008. (21) Sch€afer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571. (22) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 1980, 72, 650. (23) R. Ahlrichs and co-workers, personal communication. 2158

dx.doi.org/10.1021/jp110120c |J. Phys. Chem. A 2011, 115, 2155–2159

The Journal of Physical Chemistry A

ARTICLE

(24) YaJun, L.; WeiHai, F. Sci. China Ser. B 2007, 50, 725. (25) Runge, E.; Gross, E. K. U. Phys. Rev. Lett. 1984, 52, 997. (26) Andersen, J. U.; Cederquist, H.; Forster, J. S.; Huber, B. A.; Hvelplund, P.; Jensen, J.; Liu, B.; Manil, B.; Maunoury, L.; Brøndsted Nielsen, S.; Pedersen, U. V.; Rangama, J.; Schmidt, H. T.; Tomita, S.; Zettergren, H. Phys. Chem. Chem. Phys. 2004, 6, 2676. (27) Calvo, M. R.; Andersen, J. U.; Hvelplund, P.; Brøndsted Nielsen, S.; Pedersen, U. V.; Rangama, J.; Tomita, S.; Forster, J. S. J. Chem. Phys. 2004, 120, 5067. (28) Wyer, J. A.; Brøndsted Nielsen, S. J. Chem. Phys. 2010, 133, 084306.

2159

dx.doi.org/10.1021/jp110120c |J. Phys. Chem. A 2011, 115, 2155–2159