Theoretical Study of r-84 Phycocyanobilin Chromophore from theThermophilic Cyanobacterium Synechococcus elongatus

Costantino Zazza,*,†,‡ Nico Sanna,† and Massimiliano Aschi‡

Consorzio InteruniVersitario per le Applicazioni di Supercalcolo per UniVersita e Ricerca (CASPUR),Via dei Tizii 6b, 00185 Rome, Italy, and Dipartimento di Chimica, Ingegneria Chimica e Materiali,UniVersita di L’Aquila, Via Vetoio (Coppito 1), 67010 L’Aquila, Italy

ReceiVed: February 5, 2007

Time-dependent density functional theory (TD-DFT) calculations were performed to obtain vertical excitationenergies from the ground state to different low-lying singlet excited states of the protonatedR-84phycocyanobilin chromophore (R-84 PCBH+). It clearly emerges that three gradient-corrected approximationfunctionals (B3LYP, PBE0, and PBEPBE) show a similar description, confirming the proposed valenceassignment of the strongest UV-vis absorption band at 618 nm. Moreover, our results show that there arenot appreciable differences, in terms of excitation wavelength of the main peak, between theR-84 PCBH+

chromophore and a model system in which the two propionic chains have not been taken into account. Finally,with the precise aim of investigating the effects ofR-84 PCBH+ conformational fluctuations on its electronicproperties, vertical excitation energies obtained for the potential energy local minimum structure were alsorefined using a recently proposed TD-DFT/principal component analysis/Car-Parrinello molecular dynamicscomputational approach. Interestingly, and in line with previous results on another photosensitive complex,this study essentially suggests that interaction with the surrounding environment (protein matrix plus solventmolecules) coupled with the large amplitude fluctuation of the whole C-Phycocyanin (C-PC) pigment proteincan affect the electronic properties of theR-84 PCB chromophore and therefore its biological activity.

1. Introduction

Light harvesting in cyanobacteria and red algae is a functionof phycobilisomes that are located on the surface of photosyn-thetic membranes1 and has the important role of transferringlight energy to a photosynthetic reaction center very efficiently.2-5

The phycobilosomes are essentially formed by assemblingbiliproteins.6 C-Phycocyanin (C-PC) is one of the mostimportant biliproteins that has covalently bound bilin chro-mophores known as phycocyanobilins (PCBs)7 (shown in Figure1a). Crystallographic results have identified three different PCBstermed asR-84 PCB,â-84 PCB, andâ-155 PCB according totheir position along the primary structure of the protein.8-12 Thephotophysical properties of these chromophores are of crucialimportance for the biological activity of the reaction center. Forthis reason, in the few last years, we have witnessed a veryactive interest, in particular, in the electronic properties ofR-84PCB and on the effect of the protein and solvent environment.13-16

According to UV and CD experimental data,15,16 R-84 PCB ofC-PC shows quite a large absorption interval in the rangebetween 500 and 650 nm with a maximum at 618 nm, whicharises from a valence1π f π* electronic transition, and a smallshoulder at 570 nm. High-level theory quantum chemicalcalculations13 and semiempirical INDO-CI data14 have sug-gested that the shoulder peak at 570 nm comes from a charge-transfer (CT) transition from a nearby amino acid residue,Asp87, to the sameπ* molecular orbital involved in the mainabsorption of theR-84 PCB molecule. However, despite such

investigations, there are still some aspects that need to be betterclarified. In particular,R-84 PCB is a relatively flexible moleculewhose electronic properties may be heavily affected by con-formational transitions.17 All the previous calculations13,14havebeen carried out on theR-84 PCB conformation equal or similarto the one present in the complex with biliprotein. The primarygoal of the present study is to computationally characterize the

* Corresponding author. E-mail: costantino.zazza@caspur.it. Tel.:+3906 44486720.

† Consorzio Interuniversitario per le Applicazioni di Supercalcolo perUniversitae Ricerca.

‡ Universitadi L’Aquila.

Figure 1. (a) Geometry ofR-84 PCBH+ chromophore. Red ballsrepresent O, blue N, bronze C, and white H atoms. (b) Geometry ofR-84 s-PCBH+ model system (s-FO-PCBH+).

5596 J. Phys. Chem. B2007,111,5596-5601

10.1021/jp070994g CCC: $37.00 © 2007 American Chemical SocietyPublished on Web 04/27/2007

intrinsic electronic properties of a flexibleR-84 PCB moleculeusing a recently proposed computational technique,17 whichcombines TD-DFT calculations18 with principal componentanalysis (PCA)19 and Car-Parrinello molecular dynamics(CPMD).20 This study also aimed at extending previoustheoretical studies, using different levels of theory (differentcombination of density functionals and basis sets), for estimatingthe effect of the biological environment (protein and solvent)on R-84 PCB electronic properties.

2. Materials and Methods

2.1. Geometry Optimizations and Vertical ExcitationEnergies.The structure ofR-84 PCB was obtained from thecrystal structure of C-PC isolated from the thermophiliccyanobacteriumSynechococcus elongatus,10 and the hydrogenatoms were added to the molecule and semiempirically mini-mized by the AM1 method.21 Although the crystal structure doesnot give any evidence for the protonation state of the chro-mophore, the highly conserved Asp87 residue is supposed totransfer a proton toR-84 PCB. Therefore, and in line withprevious computational studies,13,14 we have considered theprotonated state ofR-84 PCB, hereafter indicated asR-84PCBH+. Geometry optimizations employing the DFT22,23withingeneralized gradient-corrected approximation (GGA) were thenaccomplished forR-84 PCBH+ relaxing first only the hydrogenpositions and then all the remaining internal coordinates. Forindicating the last situation, we hereafter use the FO (fullyoptimized) acronym. For this purpose, the B3LYP24,25 andPBE026-28 hybrid functionals were adopted in combination withthe 6-31G* and 6-311++G* atomic basis sets.29 TD-DFTcalculations were then performed onR-84 PCBH+ to obtainvertical excitation energies from the ground to the lowest singletelectronic excited states. Note that with the development of alinear scaling procedure, the TD-DFT approach is probably,nowadays, the most accurate theoretical approach that can beapplied to large and complex systems such as the present one.18

Finally, to roughly evaluate also environmental effects on thespectroscopic response, vertical absorptions were also reevalu-ated by carrying out the same TD-DFT calculations correspond-ing to a chemical environment described by the polarizablecontinuum model (PCM).31 In this respect, to take into accountthe effects of the C-PC protein, we have considered the proteinas an external electric field (with a dielectric constant of 4.0),32

which acts onR-84 PCBH+. To further refine the local structuremodeling of the environment around the chromophore, we alsocomputed TD-DFT/PCM absorption energies in the presenceof the highly conserved Asp87 amino acid residue and explicitsolvent molecules.

2.2. TD-DFT/PCA/CPMD Methodology and Software.Forintroducing, to a qualitative extent, the thermal effects forstudying the relationship between local semiclassical fluctuationsand electronic properties ofR-84 PCBH+, we carried outdamped second-order ab initio molecular dynamics33a withinthe CP scheme20,33band PCA procedure19 indicated as the TD-DFT/PCA/CPMD methodology. According to this approach, theconfigurational space accessible to a molecule in its electronicground state is spanned by means of-CPMD. The trajectoryis then analyzed using essential dynamics.19 In this study, wecarried out CPMD simulations in the gas phase. Such a choice,although preventing any quantitative interpretation of the results,is the only computationally accessible procedure for qualitativelyevaluating the effects of internal fluctuations on the electronicproperties of the chromophore and, consequently, also forgaining some information on the mechanical role of the protein

matrix with a computational strategy that has been recentlysuccessfully applied to similar systems.17 With this purpose, acovariance matrixC of the positional fluctuations of the atomswas constructed and diagonalized. This procedure provides anew set of generalized coordinates associated with the eigen-vectors of the matrix. The value of the associated eigenvalues(i.e., the fluctuations along the eigenvectors) allows us toseparate the intramolecular large collective essential motions(large eigenvalues) from the remaining small amplitude fluctua-tions (i.e., in constrained motions). This space defined by theessential motions, finally, allows us to identify the most sampledconformations of the molecule. In the present case, the problemis complicated by the presence of propionic chains whoseflexibility would require very long simulation times. For thesereasons, we carried out TD-DFT calculations as described inthe previous section also on anR-84 PCBH+ system withoutsuch groups, hereafter termed asR-84 s-PCBH+ (depicted inFigure 1b). This choice may be justified by the rather lowinfluence17 (see section 3.1 of this paper) of the propionic chainson the backbone conformational fluctuations. For these calcula-tions, we adopted a plane-wave (PW) pseudo-potential (PPs)approach in conjunction with ultrasoft self-consistent PPs;34 theelectronic structure calculation is based on density functionaltheory, using the Perdew-Burke-Ernzerhof (PBE) GGA forexchange and correlation energy.26,27The PW calculations havebeen performed in an orthorhombic cell where the chargeneutrality has been ensured by a uniform background charge.35a,35b

Energy cutoffs of 25 and 200 Ry have been applied to truncatethe PW expansion of pseudo-wave functions and of theaugmented charge density, respectively.35 A time step of 10 au(0.242 fs) has been applied to integrate the equations of motion.The model system was then heated from 100 to 300 K, and asimulation of 15 ps was carried out in the canonical (NVT)ensemble. Such a simulation time is certainly too short for anyquantitative evaluation. However, significantly longer timescales (i.e., on the order of nanoseconds) are computationallyrather expensive, and simulations in the range of tens ofpicoseconds are nowadays commonly used for semiquantitativemolecular property evaluations of systems even more complexthan the present one.36 However, it is important to stress that,for the goals of the present theoretical investigation, such arelatively short time turned out to be sufficient enough toestimate the semiclassical internal motions characterized by afrequency of about 5 cm-1. The temperature was kept constantby the Nose´ thermostat.37 The CP trajectory was performed withthe CP-Vanderbilt code released within the Espresso package.33c

We consider this method applicable to gas-phase conditions,liquid-phase organic and organometallic chemistry, and surfaces(in particular, nanostructured systems). PCA decomposition wasdone using the Gromacs package utilities,19b whereas all theTD-DFT calculations were performed using Gaussian 03.30

Calculations were run on a four-way HP Proliant DL585 serverwith a Dual-Core AMD Opteron Processor running at 2.4 GHz.Since the PCA decomposition substantially is very fast, thecomputational cost of this method is that generally required byCPMD simulation and time-dependent calculations carried outtaking into account only the selected structures along the largestamplitude fluctuations.

3. Results and Discussion

3.1. Geometry Optimizations and Low-Lying SingletExcited States.The singlet electronic ground state geometriesof freeR-84 PCBH+ andR-84 s-PCBH+ were optimized at theB3LYP/6-31G*, PBE0/6-31G*, B3LYP/6-311++G*, and PBE0/

Study ofR-84 Phycocyanobilin Chromophore J. Phys. Chem. B, Vol. 111, No. 20, 20075597

6-311++G* level of theory, respectively. The geometricalparameters did not result in being appreciably sensitive to thelevel of the computations. TD-DFT calculations were carriedout using the same functional basis set combination. InTable 1, we only report, for the sake of brevity, results withthe largest 6-311++G* atomic basis set, and the results arepresented and compared with experimental data.

From the comparison of UV and CD experimental data ofC-PC15,16 and considering previous semiempirical (INDO-CI)14 and DFT calculations (TD-B3LYP),13 it clearly emergesthat, irrespective of the employed atomic basis set, B3LYP andPBE0 provide a similar description of the main excitation bothshowingπ(HOMO) f π*(LUMO) character along the wholearomatic moiety. It is worth noting that the agreement withprevious TD-DFT calculations published by Wan et al.13a

confirms that the structure of theR-84 PCB antenna complexis similar in two organisms,S. elongatus10 and Fremyelladiplosiphon.11 However, both density functionals seem tomoderately overestimate the experimental transition energy valueeven though the B3LYP functional seems to perform slightlybetter. Moreover, our TD-DFT results provide transition energydifferences between s-FO-PCBH+ and FO-PCBH+ within 9nm; such a result is most likely due to theπ f π* character ofthe main peak, which involves molecular orbitals essentiallylocalized within the aromatic skeleton. Furthermore, corre-sponding to theR-84 PCBH+ optimized structure, at the B3LYP/6-311++G* the level of theory, the electric dipole momentsof the lowest excited states were also evaluated. If comparedwith the ground state, it is quite evident that the lowest valenceelectronic transition increases the norm of the electric dipolemoment of the chromophore (from 5.3 to 8.2 D). Unfortunately,no experimentally determined dipoles are available in theliterature for evaluating the quality of the present results.

Nevertheless, this result suggests the importance of the interac-tion between the light-harvesting center and its biologicalenvironment, which allows a more realistic description of theC-PC system in comparison with isolated QM calculations onthe R-84 PCBH+ subsystem. As a matter of fact, in a recentpaper,13a always using the DFT approach, the authors haveshown that the effect of the C-PC protein environment is veryimportant. The model system consisted of the PCBH+ + Asp87complex embedded in a chemical environment described by thePCM model.31 In this respect, the authors considering the gas-phase geometry of the PCBH+ + Asp87 complex and using adielectric constant of 4.032 to take into account the medium effectof the protein have shown that the excitation wavelength of themain peak, with respect to the ideal gas-phase condition, is red-shifted by 29 nm (from 565 to 594 nm, see Tables 1 and 2).The same trend had already been observed by Kikuchi and co-workers.14 In line with previous finding,13aour data reported inTable 2 show that the inclusion of the PCM model shifts thetransition energy of the main peak toward the experimentalassignment. In fact, with respect to gas-phase results, a red-shift (approximately 26 nm for the FO-PCBH+ chromophoreand 18 nm for the s-FO-PCBH+ system, respectively) wasalways obtained after geometry optimization within the PCMmodel. From the same table, we wish also to remark that at theTD-B3LYP/6-311++G*/PCM level of theory, the s-FO-PCBH+ model system and the FO-PCBH+ chromophoreprovide the same excitation wavelength (i.e., 610 nm, seeTable 2). This result suggests that, at least at the levels of theoryapplied, the complexity (in terms of nuclear and electronicdegrees of freedom) ofR-84 PCBH+ can be reduced, neglectingthe inclusion of two propionic chains to evaluate its electronicproperties. Furthermore, the inclusion of diffuse functions inthe basis set improves the agreement between theory andexperimental assignment; as a matter of fact, it is evident thatin comparison with the 6-311++G* basis set, the 6-31G* onesystematically underestimates the excitation wavelength of thelowest 1π f π* (data not shown) transition. This findingindicates that for a more realistic quantum mechanics descrip-tion of the photosensitive active site in the C-PC pigmentprotein, a basis set including diffuse functions should be used.

Since the latter model does not include the microscopicstructure of the environment around the solute molecule, similarto previous studies,13a we have also computed the absorptionenergies in the presence of the Asp87 residue and explicit watermolecules. The TD calculations carried out on the FO-PCBH+

+ Asp87 model provide shifts in agreement with thosepreviously calculated by Wan et al.13a It is also interesting toremark, from the same table, that at our best level of theory(i.e., TD-B3LYP/6-311++G*/PCM), the lowest CT electronic

TABLE 1: Calculated and Experimental ExcitationWavelength (∆E, in nm) of the Lowest Valence1π f π*Electronic Transition with Corresponding OscillatorStrength (f) at Different Levels of Theorya

method model ∆E f

TD-B3LYP/6-31+G*13a FO-PCB 539 1.09TD-B3LYP/6-31+G*13a FO-PCBH+ 582 1.32TD-B3LYP/6-31+G*13a FO-PCBH+ + Asp87 565INDO-CI14 PCBH+ 588 1.30TD-B3LYP/6-311++G* FO-PCBH+ 584 1.33TD-PBE0/6-311++G* FO-PCBH+ 568 1.35TD-B3LYP/6-311++G* s-FO-PCBH+ 592 1.40TD-PBE0/6-311++G* s-FO-PCBH+ 577 1.42exptl value15,16 C-PC 618

a FO: fully optimized R-84 PCBH+ structure (Figure 1a) and s:simpler PCB structure (without the two propionic groups in Figure 1b).

TABLE 2: Calculated TD-B3LYP/PCM Vertical Excitation Wavelength ( ∆E-PCM, in nm) of the Lowest Valence1π f π* andCT (Asp87 f r-84 FO-PCBH+) Electronic Transitions with Corresponding Oscillator Strength (f-PCM) at Different Levels ofTheory, as well as Corresponding Excitation Wavelength (∆E-GP, in nm) in Ideal Gas-Phase Conditions

method model ∆E-GP f-GP ∆E-PCM f-PCM1π f π*(Exp 618 nm)15,16

TD-B3LYP/6-31+G*13a FO-PCBH+ + Asp87 565 1.32 594 1.55TD-B3LYP/6-311++G* FO-PCBH+ 584 1.33 610 1.46TD-B3LYP/6-311++G* FO-PCBH+ + Asp87 569 1.34 593 1.47TD-B3LYP/6-311++G* s-FO-PCBH+ 592 1.40 610 1.54TD-B3LYP/6-311++G* s-FO-PCBH+ + 2H2O(A) 593 1.42 610 1.57TD-B3LYP/6-311++G* s-FO-PCBH+ + 2H2O(D) 594 1.43 611 1.56TD-B3LYP/6-311++G* s-FO-PCBH+ + 4H2O(A+D) 596 1.44 614 1,57

CT (Exp 570 nm)15,16

TD-B3LYP/6-31+G*13a FO-PCBH+ + Asp87 583 539 0.0005TD-B3LYP/6-311++G* FO-PCBH+ + Asp87 622 567 0.002

5598 J. Phys. Chem. B, Vol. 111, No. 20, 2007 Zazza et al.

transition (Asp87f R-84 FO-PCBH+) drops at 567 nm, inagreement with the experimental assignments at 570 nm.15,16

Finally, hydrogen-bonded complexes, including from two(1:2) to four (1:4) solvent molecules, were also considered forgeometry optimizations and TD calculations in the PCM model.In these complexes, the water molecules would roughly simulatethe inner solvent shell, interacting with each carbonyl group ofthe light-harvesting complex. Interestingly, for the 1:2 micro-solvated structures, we note essentially the same behaviorcharacterized by a solvatochromic shift in the range previouslypredicted considering only the mean field approach. On the otherhand, the calculations carried out on the highest coordinationmodel (1:4, see Figure 2) provide a stronger solvatochromicshift corresponding to an excitation energy value (614 nm atthe TD-B3LYP/6-311++G*/PCM level) in excellent agreementif compared to the experimental data of 618 nm.15,16 In thisrespect, our calculations provide, up to now, the most accurate(in terms of atomic basis set and level of theory) evaluation ofthe relative low-lying valence state excitation energy presentin the literature.

In line with previous computational data,13a,14our results showthat the effects on the low-lying valence PCB excited statescan be summarized as being due to (i) protonation of the PCBchromophore, (ii) interaction with the highly conserved Asp87residue, and (iii) long-range interaction with the protein environ-ment, and (iv) hydrogen-bonding network with explicit solventmolecules.

3.2. Effect of Conformational Transitions. The previoussection indicates that the employed computational methodsprovide a rather suitable tool for investigating electronictransitions in PCB. This allows us to safely address the centralquestion of the present study as to whether and to what extentlocal conformational motions of the photosensitive center affectits own spectroscopic features. The TD-DFT/PCA/CPMDmethodology was then applied to the simplifiedR-84 s-PCBH+

system whose low-lying valence excitation energies were ratherclose toR-84 PCBH+ (see Table 1). CP simulation of 15 psfollowed by essential dynamics show that the conformationaltransitions of gaseousR-84 s-PCBH+ are characterized by afluctuation along a single eigenvector. From a mechanical pointof view, this motion may be described as a torque-like motionin which the terminal rings provide the largest contribution (seeFigure 3).

In Figure 4a, we reported the1π f π* (TD-B3LYP/6-311++G*) vertical excitation wavelength calculated for dif-ferent structures sampled along the first essential eigenvector;the related oscillator strengths are reported in Figure 4b. Fromthese figures, one can observe that, although the oscillator

strength undergoes rather small fluctuations, the semiclassicalconformational transitions sharply alter the excitation wave-length that spans a region as large as 30 nm. In Figure 4a,b, wealso indicate the corresponding value previously calculated forthe gas-phase FO-R-84 s-PCBH+ (592 nm at the same level oftheory, i.e., TD-B3LYP/6-311++G*). This finding clearlydemonstrates the dramatic influence of the semiclassical mo-lecular fluctuations onto the first lowest valence electronictransition of such a chromophore.

Figure 2. Geometry of (1:4) hydrogen-bondedR-84 PCBH+ complex.Red balls represent O, blue N, bronze C, and white H atoms.

Figure 3. Extreme configurations (-1, 1) selected along the firstessential eigenvector as obtained by Car-Parrinello molecular dynamicssimulation. Zero is the averaged structure of theR-84 PCBH+ modelsystem. Red balls represent O, blue N, and bronze C atoms.

Figure 4. TD-B3LYP/6-311++G* vertical excitation wavelength(dashed line) (a) and oscillator strengths (b) for the lowest valence1πf π* electronic transition (strongest UV-vis absorption band in therange between 300 and 700 nm) of theR-84 s-PCBH+ model systemin ideal gas-phase conditions along the first essential eigenvector ascompared to the excitation wavelength (solid line) at the local minimumstructure. Zero is the averaged structure of theR-84 s-PCBH+ modelsystem.

Study ofR-84 Phycocyanobilin Chromophore J. Phys. Chem. B, Vol. 111, No. 20, 20075599

In conclusion, our results on the present TD-DFT/PCAcalculations show that the largest backbone internal fluctuation(reported in Figure 3) tunes the electronic properties of the PCBmolecule, lowering the vertical excitation energy of the strongestabsorption band in the range between 300 and 700 nm.

As a final step in the investigation, we also evaluated theactual gas-phase spectrum (i.e., the excitation energy distributionand spectral transition weighted for the statistically relevantconformations ofR-84 s-PCBH+ at 300 K according to ourCPMD simulation). To this end, we carried out cluster analysison the same CP trajectory. The TD-B3LYP/6-311++G*calculations were performed for eight structures with a RMSDvalue (mass weighted, neglecting the hydrogen atoms) between0.14 and 0.58 nm. The weight of each structure was then foundby using the relative Helmholtz free energy calculated withrespect to a reference structure as-RT ln(Pi/Pref), wherePi isthe occurrence frequency of theith structure andPref is theoccurrence frequency of the reference. The results of thisanalysis, reported in Figure 5, indicate that the most stableconformation absorbs at a higher energy, whereas the conforma-tion resembling the crystal structure (i.e., the local free energyminimum at +4.4 kJ/mol) absorbs at about 592 nm as waspreviously indicated. According to our results, we can assessthat the protein matrix, in this case, should act not only as aperturbation on the electronic properties but also as a mechanicalconstraint, allowingR-84 s-PCBH+ to adopt a structure foundin the crystal that actually represents a high-energy conforma-tion. It is worth noting that, if the selected structures along thefirst essential eigenvector as well as those extracted from thecluster analysis were locally fully relaxed, the main absorptionband was systematically found to be equal to its previouslyobtained value taking into account only the local minimumconformation; the agreement between the fully and the locallyrelaxed conformations suggests that the conformational samplingis essentially confined within the region characterized by thelocal minimum energy structure. The same trend was observedstarting from the protein constrained (PC) conformation, eventhought the X-ray structure of theR-84 PCB antenna complexwas basically never sampled during the CP simulation. Thislast finding, at least within the range explored, indicates thatthe C-PC protein probably modulates the PCB chromophorein its configurational space and, consequently, in its electronicresponses.

Finally, in Figure 6, we show the corresponding spectrumfrom which we may infer a blue-shift of 40 nm with respect tothe experimental detection in the protein matrix in solution.Therefore, when the PCB molecule is confined within the localminimum conformation, our calculations nicely reproduce theexperimental observation if the inclusion of the environmentperturbation is considered (see Table 2). If, on the other hand,the chromophore is left to fluctuate, its intrinsic spectroscopicfeatures are rather altered (see Figure 6); as a matter of fact,during CPMD simulation at 300 K, the PCB chromophore israther flexible, and its electronic properties cannot be addressedby only considering the potential energy absolute minimum (thenneglecting the inclusion of thermal effects). We wish also toremark that the use of a different functional (PBE) does notalter the picture.

Although further investigations are obviously needed, thecombined use of the TD formalism and essential dynamicsclearly shows that the spectroscopic features of theR-84 PCB(not occurring in crystal but in physiological conditions whereinternal fluctuations are conceivable) cannot be completelyaddressed by only using a static approach. In fact, an interestingcorrelation between electronic properties and conformationalfluctuations inR-84 PCBH+ has emerged, suggesting a possibleconformational regulation mechanism of the photochemicalactivity. The surrounding protein, from this study, is supposedto act both as a perturbation of the electronic properties of thechromophore as well as a conformational constraint.

4. Conclusion

In this article, we used the density functional theory withinits time-dependent formalism to investigate the influence of thesurrounding environment on vertical excitation energies fromthe ground state to different low-lying singlet excited states oftheR-84 phycocyanobilin chromophore. Moreover, for the firsttime, a direct computational evaluation of theR-84 PCBchromophore internal motion on the spectroscopic features ofthis system has been carried out and may be of some relevancealso to experimentalists. Our TD-DFT results being in agreementwith the experimental value and in analogy with previous resultsobtained by means of semiempirical and ab initio calculationsshow that (i) even for the evaluation of low-lying excitationtransitions, in combination with TD-DFT formalism, a ratherlarge basis set is mandatory. (ii) Both environmental effects dueto C-PC protein and semiclassical fluctuations of theR-84 PCBchromophore are found to have a decisive influence on thevalence excited state electronic properties. (iii) The environment

Figure 5. 300 K Helmholtz free energy (dashed line) as a function ofTD-B3LYP/6-311++G* vertical excitation wavelength of the lowestvalence1π f π* electronic transition (strongest UV-vis absorptionband in the range between 300 and 700 nm) of theR-84 s-PCBH+

model system in ideal gas-phase conditions as obtained by clusteranalysis.

Figure 6. TD-B3LYP/6-311++G* UV-vis spectrum for gas-phaseR-84 s-PCBH+ at 300 K (see text).

5600 J. Phys. Chem. B, Vol. 111, No. 20, 2007 Zazza et al.

(protein matrix and solvent) is suggested to act not only as anelectrostatic perturbation on chromophore electronic propertiesbut also as a mechanical guide, allowing the chromophore itselfto adopt a relatively high-energy conformation.

It is also obvious that our data intrinsically suffer for thelimitation of our CP-MD simulation. However, in this respect,our next work will be essentially based on a recently developedQM/MM methodology, where the effect of the environment ismodeled as an external electric field acting on the unperturbedeigenstates of the chromophore.38

Acknowledgment. The authors gratefully acknowledge Prof.V. Barone for stimulating discussions about density functionaltheory and the polarizable continuum model. Thanks are alsodue to the Supercomputing Center for University and Research(CASPUR) for the support given for this computational study.

References and Notes

(1) Glazer, A. N.; Lundell, D. J.; Yamanaka, G.; Williams, R. C.Ann.Microbiol. (Paris) 1983, 134, 159.

(2) Danks, S. M.; Evans, E. H.; Whittaker, P. A.PhotosyntheticSystems; Wiley: New York, 1984.

(3) Feher, G.; Allen, J. P.; Okamura, M. J.; Rees, D. C.Nature1989,339, 111.

(4) Poter, G.; Tredwell, C. J.; Searle, C. J.; Barber, G. F. W.J. Biochim.Biophys. Acta1978, 501, 232.

(5) Mimuro, M.; Yamazaki, I.; Murao, T.; Yamazaki, T.; Yoshihara,J.; Fujita, Y. AdVances in Photosynthetic Research; Sibesma, C., Ed.;Mijhoff, M. and Junk W. Publishers: Hague, The Netherlands, 1984; Vol.I, p 21-28.

(6) Glazer, A. N.Biochim. Biophys. Acta1984, 768, 29.(7) Lundell, D. J.; Williams, R. C.; Glazer, A. N.J. Biol. Chem.1981,

256, 3580.(8) Deisenhofer, J.; Michel, H.; Huber, R.Trends Biochem. Sci. 1985,

10, 243.(9) Dale, R. E.; Teale, F. W.J. Photochem. Photobiol., A 1970, 12,

99.(10) Nield, J.; Rizkallah, P. J.; Barber, J.; Chayen, N. E.J. Struct. Biol.

2003, 141, 149.(11) Guerring, M.; Schmidt, G. B.; Huber, R.J. Mol. Biol. 1991, 217,

577.(12) Stec, B.; Troxler, R. F.; Teeter, M. M.Biophys. J.1999, 76, 2912.(13) (a) Wan, J.; Xu, X.; Ren, Y.; Yang, G.J. Phys. Chem. B2005,

109, 11088. (b) Ren, Y.; Wan, J.; Xu, X.; Zhang, Q.; Yang, G.J. Phys.Chem. B.2006, 110, 18665.

(14) Kikuchi, K.; Sugimoto, T.; Mimuro, M.Chem. Phys. Lett. 1997,274, 460.

(15) Mimuro, M.; Fuglistaller, P.; Rumbeli, R.; Zuber, H.Biochim.Biophys. Acta1986, 848, 155.

(16) Mimuro, M.; Rumbeli, R.; Fuglistaller, P.; Zuber, H.Biochim.Biophys. Acta1986, 851, 447.

(17) Spezia, R.; Zazza, C.; Palma, A.; Amadei, A.; Aschi, M.J. Phys.Chem. A2004, 108, 6763.

(18) (a) Hirata, S.; Head-Gordon, M.Chem. Phys. Lett. 1999, 302, 375.(b) Stratmann, R. E.; Scuseria, G. E.; Frisch, M. J.J. Chem. Phys. 1998,109, 8218.

(19) (a) Amadei, A.; Linsenn, A. B. M.; Berendsen, H. J. C.Proteins1993, 17, 412. (b) van der Spoel, D.; Berendsen, H. J. C.; van Buuren, A.R.; Apol, M. E. F.; Meulenhoff, P. J.; Sijbers, A. L. T. M.Gromacs UserManual; Nijenborgh 4, 9747 AG: Groningen, The Netherlands, 1995.

(20) Car, R.; Parrinello, M.Phys. ReV. Lett. 1985, 55, 2471.(21) Dewar, M. J. S.; Reynolds, C. H.J. Comput. Chem. 1986, 2, 140

and references therein.(22) Hohenberg, P.; Kohn, W.Phys. ReV. B:Condens. Matter Mater.

Phys.1964, 136, 864.(23) Kohn, W.; Sham, L. J.Phys. ReV. A: At., Mol., Opt. Phys.1965,

140, 1133.(24) Becke, A. D.J. Chem. Phys. 1986, 84, 4524.(25) Lee, C.; Yang, W.; Parr, R. G.Phys. ReV. B 1988, 37, 785.(26) Perdew, J. P.; Wang, Y.Phys. ReV. B 1992, 45, 13244.(27) Perdew, J. P.; Burke, K.; Ernzerhof, M.Phys. ReV. Lett.1996, 77,

3865.(28) Ernzerhof, M.; Scuseria, G.J. Chem. Phys.1999, 110, 5029.(29) Krishnan, R.; Binkley, J. S.; Seager, R.; Pople, J. A.J. Chem. Phys.

1980, 72, 650.(30) Frisch, M. J., et al.Gaussian 03, Gaussian, Inc.: Wallingford, CT,

2004.(31) (a) Miertus, S.; Scrocco, E.; Tomasi, J.J. Chem. Phys.1981, 55,

117. (b) Cossi, M.; Barone, V.J. Phys. Chem. A2000, 104, 10614.(32) Blomberg, M. R. A.; Siegbahn, P. E. M.; Babcock, G. T.J. Am.

Chem. Soc. 1998, 120, 8812.(33) (a) Tassone, F.; Mauri, F.; Car, R.Phys. ReV. B: Condens. Matter

Mater. Phys.1994, 50, 10561. (b) Pasquarello, A.; Laasonen, K.; Car, R.;Lee, C.; Vanderbilt, D.Phys. ReV. Lett. 1992, 69,1982. (c) Espresso packageis available on-line at the following address: http://www.pwscf.org/.

(34) Vanderbilt, D.Phys. ReV. B: Condens. Matter Mater. Phys.1992,46, 7892.

(35) (a) Palma, A.; Pasquarello, A.; Cicciotti, G.; Car, R.J. Chem. Phys.1998, 108, 9933. (b) Zazza, C.; Palma, A.; Aschi, M.Chem. Phys. Lett.2006, 428, 152. (c) Zazza, C.; Meloni, S.; Palma, A.; Knupfer, M.; Fuentes,G. G.; Car, R.Phys. ReV. Lett. 2007, 98, 46401.

(36) See, for example: Cascella, M.; Magistrato, A.; Tavernelli, I.;Carloni, P.; Rothlisberger, U.Proc. Natl. Acad. Sci. U.S.A.2006, 103, 19641.

(37) Nose, S. J. Chem. Phys.1984, 81, 511.(38) (a) Zazza, C.; Amadei, A.; Sanna, N.; Grandi, A.; Chillemi, G.; Di

Nola, A.; D’Abramo, M.; Aschi, M.Phys. Chem. Chem. Phys.2006, 12,1385. (b) Aschi, M.; Zazza, C.; Spezia, R.; Bossa, C.; Di Nola, A.; Paci,M.; Amadei, A.J. Comput. Chem.2004, 25, 974.

Study ofR-84 Phycocyanobilin Chromophore J. Phys. Chem. B, Vol. 111, No. 20, 20075601