- Email: [email protected]
Absorption and fluorescence spectroscopic characterization of BLUF domain of AppA from Rhodobacter sphaeroides
Chemical Physics 315 (2005) 142–154 www.elsevier.com/locate/chemphys
Absorption and fluorescence spectroscopic characterization of BLUF domain of AppA from Rhodobacter sphaeroides P. Zirak a, A. Penzkofer
a,*
, T. Schiereis b, P. Hegemann b, A. Jung c, I. Schlichting
c
a
c
Institut II – Experimentelle und Angewandte Physik, Universita¨t Regensburg, Universita¨tstrasse 31, D-93053 Regensburg, Germany b Institut fu¨r Biologie, Experimentelle Biophysik, Humboldt-Universita¨t zu Berlin, Invalidenstrasse 42, D-10115 Berlin, Germany Max-Planck-Institut fu¨r medizinische Forschung, Abteilung Biomolekulare Mechanismen, Jahnstrasse 29, D-69120 Heidelberg, Germany Received 26 December 2004; accepted 13 April 2005 Available online 13 May 2005
Abstract The BLUF domain of the transcriptional anti-repressor protein AppA from the non-sulfur anoxyphototrophic purple bacterium Rhodobacter sphaeroides was characterized by absorption and emission spectroscopy. The BLUF domain constructs AppA148 (consisting of amino-acid residues 1–148) and AppA126 (amino-acid residues 1–126) are investigated. The cofactor of the investigated domains is found to consist of a mixture of the flavins riboflavin, FMN, and FAD. The dark-adapted domains exist in two different active receptor conformations (receptor states) with different sub-nanosecond fluorescence lifetimes (BLUFr,f and BLUFr,sl) and a small non-interacting conformation (BLUFnc). The active receptor conformations are transformed to putative signalling states (BLUFs,f and BLUFs,sl) of low fluorescence efficiency and picosecond fluorescence lifetime by blue-light excitation (light-adapted domains). In the dark at room temperature both signalling states recover back to the initial receptor states with a time constant of about 17 min. A quantum yield of signalling state formation of about 25% was determined by intensity dependent transmission measurements. A photo-cycle scheme is presented including photo-induced charge transfer complex formation, charge recombination, and protein binding pocket reorganisation. 2005 Elsevier B.V. All rights reserved. Keywords: BLUF domain; AppA; Blue-light photoreceptor; Rhodobacter sphaeroides; Absorption spectroscopy; Fluorescence spectroscopy; Fluorescence up-conversion; Photo-cycle; Flavoprotein; FAD; FMN; Riboflavin
1. Introduction The blue-light response of biological organisms is an active field of research (for recent reviews see [1–10]). There are three classes of blue-light receptors, the cryptochromes [4,6,11,12] with sensing Cry domains, the phototropins Phot with LOV domains [5,13], and the regulatory proteins with BLUF domains [8,10,14–18]. The sensing domains non-covalently bind flavin-adenine-dinucleotide (FAD) in the case of Cry domains * Corresponding author. Tel.: +49 0 941 943 2107; fax: +49 0 941 943 2754. E-mail address: [email protected] (A. Penzkofer).
0301-0104/$ - see front matter 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.chemphys.2005.04.008
and BLUF domains, and flavin-mononucleotide (FMN) in the case of LOV domains. BLUF domains (sensor for blue light using FAD) are the light sensors of the multi-domain protein AppA from the purple non-sulfur anoxyphototrophic proteobacterium Rhodobacter sphaeroides [15], of PAC (photo-activated adenylylcyclase) from the unicellular flagellate Euglena gracilis [18], of Slr1694 from the cyanobacterium Synechocystis sp. PCC6803 [19], and of flavoproteins of at least 15 other microorganisms [8]. Blue-light excitation of BLUF domains leads to a slight red-shift of the absorption band and a recovery to the initial absorption behaviour in the dark [15,16,18–22].
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
The flavoprotein AppA [8,14–17] from the non-sulfur anoxyphototropic purple bacterium R. sphaeroides controls photosynthetic gene expression in response to blue light exposure as well as in response to changes in the cellular redox state by functioning as an antirepressor of the photosynthesis repressor protein PpsR [8,16,17,19,23]. Blue-light exposure of AppA in R. sphaeroides represses puf and puc operon expression in this bacterium [17,24]. AppA consists of an N-terminal BLUF domain and a cystein-rich C-terminal domain. The amino acid sequence of the BLUF domain of AppA is found in [8,15,20,21]. The photo-cycle dynamics of AppA by forming a long-living intermediate with slightly red-shifted absorption band was first described in [16]. In [20] a detailed spectroscopic and mutational analysis of the BLUF domain of AppA1-156 (short: AppA156, consisting of amino residues 1–156) was carried out including ultraviolet and visible spectroscopy, fluorescence measurements, nuclear magnetic resonance (NMR) spectroscopy, laser flash photolysis, and gel filtration chromatography. The photo-dynamics is interpreted in terms of flavin stacking with the amino acid Tyr-21 leading to fluorescence reduction [20]. It is proposed that light exposure strengthens a hydrogen bond between flavin and Tyr-21 leading to a stable local conformational change in AppA1-156 [20]. In [21] the photo-cycle of AppA5-125 was studied by UV–Vis spectroscopy, Fourier-transform infrared spectroscopy, pH measurements and site-directed mutagenesis. It is expected that AppA5-125 in its dark-adapted state is protonated [N(5)-H], that it becomes exposed to solvent by blue-light exposure (light-adapted state, signalling state), and that an intra-molecular proton transfer from N(5) to anionic Tyr-21 forms the basis for the stabilisation of the signalling state [21]. Very recently – after finishing this paper – ultrafast time-resolved absorption and fluorescence spectroscopic studies on a femtosecond to nanosecond timescale and laser flash-photolysis studies on a nanosecond to microsecond timescale have been performed [25] to explore the photo-cycle dynamics of AppA5-125. In the BLUF domain Slr1694 from Synechocystis sp. PCC6803 light-induced Fourier transform infrared (FTIR) spectroscopy indicated a weakening of the C(4)@O and C(2)@O bonding and a strengthening of the N(1)C(10a) and C(4a)N(5) bonding [19]. In investigating a recent heterologous expression of the BLUF domain from AppA [22] it was found that riboflavin, FMN, and FAD non-covalently bind to the domain and all three homologues have very similar blue-light photo-cycle dynamics. In this paper the BLUF domains AppA148 (domain consists of amino acid residues 1–148 with C-terminal His-tag) and AppA126 (amino acid residues 1–126) from R. sphaeroides are characterized by absorption and emission spectroscopy. The flavin chromophore compo-
143
sition is analysed by thin-layer chromatography and fluorescence quantum yield measurement. Two active BLUF domain conformations and a third small conformation which does not form an intermediate are revealed by fluorescence lifetime and fluorescence quantum yield analysis. The fluorescence quantum yields and fluorescence lifetimes of the domains in the dark-state and in the light-adapted state are determined. The photo-cycle dynamics of signalling state formation from the receptor state by light exposure and of the dark recovery of the signalling state to the receptor state at room temperature are studied. The quantum yields of photo-induced signalling state formation are determined. All spectroscopic studies have been carried out at room temperature.
2. Experimental A DNA-fragment encoding the BLUF protein AppA (Acc: L42555) from R. sphaeroides 2.4.1. was kindly provided by Dr. G. Klug (Giessen, Germany). Two fragments encoding amino acid 1–148 and 1–126, respectively, were amplified by PCR. AppA 1–126 was identified to be what one could call the proteinÕs minimum BLUF domain by limited proteolysis of AppA 1–148 with Trypsin. Proteolytic fragments were analyzed by MALDI mass spectrometry and Edmann Nterminal sequencing. The two fragments were inserted between the NdeI and SacI restriction sites of a pET28a+ vector (Invitrogen, Karlsruhe), respectively. While the shorter construct only comprises an Nterminal His6-tag, the longer construct possesses both, N- and C-terminal His6-tag, rendering the purification procedure more effective. The C-terminal tag is attached to the AppA-coding region via the amino acid sequence AAPE and a short linker of vector-derived amino acid residues. Escherichia coli cells (strain BL21) were transformed with these two constructs. Selected clones were grown in LB plus Kanamycine over night at 30 C until the optical density at 600 nm had reached 0.5. The cultures were cooled down to 18 C, and production of the AppA fragments was initiated by addition of 0.7 mM IPTG. Protein expression proceeded over night at 18 C. The proteins were purified via Ni-NTA resin according to the instructions of the supplier (Quiagen, Hilden, Germany). They were subsequently exchanged in storage buffer (10 mM phosphate buffer, pH 8, 10 mM NaCl, 100 lM phenylmethylsulfonyl fluoride (PMSF)). The flavin cofactors non-covalently bound to the AppA BLUF domains were determined by chromophore extraction from the protein, followed by thinlayer chromatography and fluorescence analysis. The chromophore was extracted from the protein following
144
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
the procedure described in [22]: The samples were boiled in 70 vol.% ethanol water solution for 1 min, and then they were centrifuged, the supernatants were lyophilized to dryness and re-suspended in 35 vol.% ethanol. The fluorescence quantum yield, /F, of the extract solution of AppA126 (/F = 0.27) was determined and compared to the fluorescence quantum yields of FAD (/F,FAD = 0.127), FMN (/F,FMN = 0.335), and riboflavin (/F,RF = 0.285) in 35 vol.% ethanol. Neglecting the fluorescence quantum yield difference between FMN and riboflavin, the mole-fraction, xFAD, of FAD in the extract was determined from the relation /F,extract = xFAD/F,FAD + (1 xFAD)/F,RF to be xFAD(AppA126) 0.15. The chromophore composition was visualized by thin-layer chromatography of FAD, FMN, riboflavin, AppA126, and AppA148 on a silica gel 60 F254 aluminium foil (Merck # 105554, Darmstadt) with nbutanol/acetic acid/water 3:1:1 (v/v) as carrier medium. The fluorescence traces photographed under a 254 nm UV lamp are displayed in Fig. 1(a). Comparison with the references traces indicates that there are present in AppA126 and AppA148 a non-movable isoalloxazine component, riboflavin (dominant part), FMN and FAD. The spots were analysed by measuring their time-integrated fluorescence signal with a photomultiplier tube after picosecond laser pulse excitation at 400 nm. The determined mole-fractions of riboflavin, FMN, FAD, and non-moveable isoalloxazine moiety are listed in Table 1. The structural formulae of the flavins are shown in Fig. 1(b). The absorption cross-section spectra of dark-adapted AppA148 and AppA126 are determined by transmission measurements with a commercial spectrophotometer (Beckman type ACTA M IV). The transmission spectra, T(k), are converted to absorption coefficient spectra, a(k), by the relation T(k) = exp[a(k)‘], where ‘ is the sample path length. The absorption-coefficient spectra contain a scattering contribution, i.e. a(k) = aa(k) + as(k), where aa is the pure absorption coefficient and as is the scattering coefficient. The scattering contribution is approximated by as(k) = as(k0) (k0/k)4 [26–28] where k0 is a wavelength in the transparency region. The absorption cross-section spectra, ra(k), are calculated from the absorption coefficient spectra, aa(k), by calibration to the absorption cross-section spectrum of riboflavin at neutral pH, i.e. the absorption cross-section integrals extending over the S0–S1 and S0–S2 absorption band (k > 310 nm) are set equal since the same isois present [29] (ra ð~mÞ ¼ að~mÞ Ralloxazine chromophore R ra;R ð~mÞd~m= aa ð~mÞd~m, where ra,R(k) is the absorption cross-section spectrum of riboflavin, ~m ¼ 1=k is the wavenumber). The absorption cross-section spectra of the lightadapted samples (signalling state) are determined by exciting small-volume samples (1.5 · 1.5 · 3.5 mm3) with a high-pressure mercury lamp at kexp = 428 nm
(light intensity at sample Iexp 0.01 W cm2, exposure time Dtexp > 3 min), and probing the transmission with an attenuated tungsten lamp using a spectrometer – diode-array detection system. The absolute absorption cross-section spectra are extracted by exploiting the presence of isobestic points in the transmission spectra of dark-adapted and light-adapted samples (there the absorption cross-sections are the same). The spectral fluorescence behaviour was studied with a self-assembled fluorimeter in front-face collection arrangement [30,31]. The absolute intrinsic fluorescence quantum distributions [32], EF(k), and absolute intrinsic fluorescence quantum yields, /F, were determined by using coumarin 314T in ethanol (fluorescence quantum yield, /F,R = 0.87 [technical data sheet of Kodak]) or riboflavin in water buffered to pH 7 (/F,R = 0.26 [33,34]) as reference. The samples were excited by vertical polarized light and the fluorescence was detected under magic-angle polarizer orientation (polarizer transmission direction under an angle of 54.7 to the vertical). The degree of fluorescence polarization [35], PF = (SF,k SF,?)/(SF,k + SF,?), was determined by measuring the fluorescence signal polarized parallel (SF,k) and polarized perpendicular (SF,?) to the excitation light. For dark-adapted fluorescence quantum distribution measurement the excitation intensity was reduced to Iexp 1.6 · 103 W cm2 (excitation wavelength 428 nm) and the fluorescence accumulation time was extended to Dtexp = 10 min, whereby during measurement the solution in a 4 mm thick cell was circulated with a magnetic stirrer. In the case of light-adapted fluorescence quantum distribution measurement, before accumulation of fluorescence light a small-volume cell (1.5 · 1.5 · 3.5 mm3) was excited for several minutes at an intensity of Iexp 0.01 W cm2. The temporal fluorescence behaviour was studied by short laser pulse excitation at 400 nm with a Ti:sapphire femtosecond oscillator-amplifier laser system (laser system Hurricane from Spectra-Physics). For the time range in the sub-nanosecond to nanosecond region, the laser was operated at a pulse-duration of DtL = 4 ps, and the fluorescence signal was detected with a microchannel-plate photomultiplier (Hamamatsu, type R1564-U01) together with a high-speed digital oscilloscope (LeCroy, type DSO 9362). For the time-range in the sub-picosecond to picosecond region, the laser was operated at a pulse-duration of 110 fs and fluorescence-up-conversion technique was applied for time resolved signal detection [36,37]: The samples were excited with frequency doubled femtosecond pulses at 400 nm, and the generated fluorescence signals (frequency mF) were frequency up-converted with the fundamental laser pulses at 800 nm (frequency mG) in a non-linear optical crystal (BBO crystal of 0.2 mm thickness [38]) by non-collinear phase-matched sum-frequency generation
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
145
Fig. 1. (a) Thin-layer chromatography traces of FAD, FMN, riboflavin, AppA126, and AppA148. (b) Structural formulae of riboflavin, flavinmononucleotide (FMN), and flavin-adenine-dinucleotide (FAD).
(frequency mS, type II phase-matching: mF(e) + mG(o) ! mS(e), o: ordinary polarized light, e: extraordinary polarized light [38]). For time resolution the gating fundamental laser pulse was time-delayed relative to the second-harmonic excitation pulse with a stepper-motor driven linear translation stage. The up-converted fluorescence signal passed through a broad band filter
(Schott glass UG11 of 10 mm thickness, transmission range from 270 to 380 nm) and was detected with a photomultiplier tube (Valvo, type PM2254B) and a highspeed digital oscilloscope (see above). The laser was operated at 1 Hz repetition rate. For dark-adapted measurements the solution in a 4 mm path-length cell was circulated by a magnetic stirrer. For light-adapted
146
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
Table 1 Parameters of AppA BLUF domains in aqueous solution at pH 8 and at room temperature AppA126
Comments
Parameter
AppA148
Chromophore composition xFAD xFMN xRF xiso
0.085 ± 0.03 0.22 ± 0.04 0.54 ± 0.04 0.155 ± 0.03
0.15 ± 0.04 0.12 ± 0.03 0.67 ± 0.05 0.06 ± 0.02
Domain composition xf xsl xnc
0.776 ± 0.03 0.212 ± 0.02 0.012 ± 0.004
0.548 ± 0.03 0.429 ± 0.03 0.023 ± 0.004
Fig. 7 Fig. 7 Fig. 7
Photo-cycle characterization /s sd,rec (min) Wb/(hc0) (cm1)
0.24 ± 0.03 16.7 ± 0.5 7554 ± 10
0.25 ± 0.03 18.8 ± 0.5 7579 ± 10
Fig. 4 Fig. 5(b) Eq. (8)
State dependent parameters /F /F,f /F,sl /F,nc PF sF,f (ps) sF,sl (ps) sF,nc (ps) kElT,f (s1) kElT,sl (s1) /ElT,f /ElT,sl /ElT /ISC,f /ISC,sl /ISC,nc /ISC ra,exc (cm2) ra,p (cm2) /D
Dark-adapted 0.02 ± 0.002 0.012 0.049 0.059 0.385 ± 0.01 232 ± 20 938 ± 50 1113 ± 100 4.11 · 109 8.66 · 108 0.954 0.812 0.922 0.0125 0.05 0.06 0.022 3.39 · 1017 9.3 · 1018
Light-adapted (9.1 ± 0.5) · 104 0.00015 0.0025 0.059 0.37 ± 0.01 2.77 ± 1 47.6 ± 15 1113 ± 100 3.61 · 1011 2.08 · 1010 0.9994 0.9905 0.997 0.00015 0.0026 0.06 0.0014 2.76 · 1017 2.62 · 1017 <1 · 105
Dark-adapted 0.0235 ± 0.002 0.0064 0.046 0.082 0.345 ± 0.01 121.7 ± 20 867 ± 50 1557 ± 100 8.02 · 109 9.53 · 108 0.976 0.827 0.909 0.0065 0.047 0.084 0.027 3.22 · 1017 9.2 · 1018
Fig. Fig. Fig. Fig.
Light-adapted (2.2 ± 0.2) · 103 0.00012 0.0017 0.082 0.33 ± 0.01 2.2 ± 1 31.6 ± 10 1557 ± 100 4.54 · 1011 3.14 · 1010 0.9996 0.9937 0.997 0.00012 0.0017 0.084 0.0028 3.15 · 1017 2.54 · 1017 <1 · 105
1(a) 1(a) 1(a) 1(a)
Fig. 6 Eq. (5) Eq. (5) Eq. (5) Fig. 7 Fig. 7 Fig. 7a Eq. (6) Eq. (6) Eq. (7a) Eq. (7a) Eq. (7b) Eq. (7a) Eq. (7a) Eq. (7a) Eq. (7b) Fig. 2 Fig. 2 Eq. (12), Fig. 5a
Abbreviations: xFAD, xFMN, xRF, xiso, mole-fractions of FAD, FMN, riboflavin, and non-movable isoalloxazine moiety, respectively. xf, xsl, xnc, mole-fractions of BLUF domain conformations with fast fluorescence lifetime and convertible to signalling state, slow fluorescence lifetime and convertible to signalling state, and not convertible to signalling state, respectively. /s, quantum yield of signalling state formation. sd,rec, recovery time of BLUF domains after light exposure from signalling state to receptor state in the dark. Wb/(hc0), energy barrier from signalling state to receptor state (h is Planck constant, c0 is vacuum light velocity). /F is total fluorescence quantum yield. /F,f, /F,sl, /F,nc are fluorescence quantum yields of fast conformation, slow conformation, and non-convertible conformation of the BLUF domains, respectively. PF is degree of fluorescence polarization. sF,f, sF,sl, sF,nc, fluorescence lifetimes of fast, slow, and non-convertible BLUF conformations, respectively. kElT,f, kElT,sl, rate constants of electron transfer for BLUF domain conformations with fast and slow fluorescence lifetime, respectively. /ElT,f, /ElT,Sl, quantum efficiencies of photo-induced electron transfer of fast and slow BLUF domain conformation, respectively. /ElT is total quantum efficiency of photo-induced electron transfer. /ISC,f, /ISC,sl, /ISC,nc, estimated quantum yields of triplet formation for fast, slow, and non-convertible conformations of the BLUF domains. /ISC is estimated total quantum yield of triplet formation. ra,exc, ra,p, absorption cross-sections at excitation wavelength kexc = 428 nm and at probing wavelength kp = 493.1 nm, respectively. /D is quantum yield of photo-degradation. kexc = 428 nm; kp = 493.1 nm. Applied parameters from riboflavin in neutral aqueous solution: radiative rate constant, krad krad,RF 5.28 · 107 s1, intersystem-crossing rate constant kISC kISC,RF 5.37 · 107 s1 [55]. Error bars are estimates taking experimental accuracy into account.
measurements the solution in a 1 mm path-length cell (volume 1 · 10 · 10 mm3) was irradiated with a filtered high-pressure xenon lamp (Schott type BG38 broadband filter of 5 mm thickness transmitting from 350 to 600 nm and broadband interference filter with FWHM transmission from 350 to 440 nm, excitation intensity 0.02 W cm2). The photo-excitation and relaxation dynamics (photo-cycle behaviour) of the samples was studied by measuring the wavelength dependent transmission of a
probing tungsten lamp as a function of (i) the excitation light intensity at 428 nm, (ii) the long-time high-intensity exposure at 428 nm, and (iii) the temporal behaviour (recovery) after excitation shut-off.
3. Results The absorption cross-section spectra of dark-adapted AppA148 (AppA148,r) and riboflavin in aqueous solution
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
at pH 8 are shown in Fig. 2. The absorption cross-section spectrum of riboflavin is taken from [34]. The dark-adapted AppA BLUF domain spectrum is calibrated to the riboflavin spectrum byR equating their absorption cross-section integrals, ra ð~mÞd~m where ~m ¼ k1 is the wavenumber, in the wavelength range above 310 nm. The shapes of the absorption cross-section spectra of AppA148,r and riboflavin are quite similar, only below 310 nm the absorption cross-sections of the AppA BLUF domain is larger because of the absorption contribution of some amino acid residues of the AppA protein (mainly Trp, Tyr and Phe [39]). The vibronic structures of the S0–S1 absorption band peaking at 444 nm and of the S0–S2 absorption band peaking at 374 nm are more pronounced for the AppA BLUF domain than for riboflavin. Within our experimental accuracy the same absorption cross-section spectrum was obtained for AppA126,r as for AppA148,r (curve not shown). The absorption coefficient spectra at different excitation light intensities are shown in Fig. 3 for AppA148. Quite similar spectra were obtained for AppA126 (curves not shown). Part (a) shows the total absorption spectra, aðkÞ ¼ aa ðkÞ þ as ðkÞ, including absorption, aa ðkÞ, and scattering, as(k), together with the approximate scattering contributions, as(k) (k4 dependence). In part (b) the true absorption coefficient spectra, aa ðkÞ, are shown. The absorption coefficient spectra at the highest excitation intensity represent the absorption coefficient spectra of the photo-induced meta-stable intermediates (lightadapted BLUF domain, signalling state). Part (c) show the difference absorption spectra of exposed samples –
147
unexposed samples. Below 380 nm the pure absorption spectra and the difference spectra are less accurate because of some uncertainty in the extrapolated scattering contribution. The absolute absorption cross-section spectra of the signalling state of AppA148 (AppA148,s) and AppA126 (AppA126,s) are extracted from the isobestic point at 454 nm. There the absorption cross-sections of the dark-adapted and light-adapted forms are the same. The obtained spectrum for AppA148,s is displayed in Fig. 2 (dashed line). The S0–S1 absorption bands of the completely light-adapted forms are approximately 16 nm red-shifted compared to the dark-adapted forms. The excitation intensity dependence of the absorption coefficients, aa,p, at kp = 493.1 nm (wavelength position of largest absorption change between exposed and unexposed samples) is displayed by the dot-connected circles in Fig. 4 for AppA148. A similar dependence was obtained for AppA126 (data points are not shown). The rise in absorption will be exploited below to extract the quantum efficiency of signalling state formation by numerical simulations. The temporal development of the absorption coefficients, ap, in the case of high-intensity irradiation with Iexp 0.01 W cm2 at kexp = 428 nm is displayed in Fig. 5(a) for AppA148 (scattering contribution is in(a)
(b)
(c)
Fig. 2. Absorption cross-section spectra of riboflavin, APPA148,r, and APPA148,s in aqueous solution buffered to pH 8. APPA spectra are calibrated to riboflavin spectrum by equating the absorption crosssection integrals in the wavelength range above 310 nm.
Fig. 3. Dependence of absorption spectra of AppA148 on light exposure. Excitation wavelength kexp = 428 nm; cell thickness ‘ = 1.5 mm; cell area, 1.5 mm · 3.5 mm. Exposure time, texp = 1 min. Time interval between exposures is 30 min for Iexp 6 1.6 · 103 W cm2, and 40 min for Iexp > 1.6 · 103 W cm2. (a) Lengthaveraged total absorption spectra, aðkÞ ¼ ln½T ðkÞ=‘ ¼ aa ðkÞ þ as ðkÞ, including pure absorption, aa ðkÞ, and scattering contribution, as(k), together with approximate scattering spectra, as(k). (b) Pure absorption spectra, aa ðkÞ. (c) Pure difference absorption spectra, ad ðkÞ ¼ aa ðk; I exp Þ aa ðk; 0Þ.
148
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
cluded, same dependence obtained for AppA126). The signalling state is formed quickly. The absorption during long-time exposure remains constant indicating a high photo-stability of the light-adapted form. The time-dependent change of absorption at kp = 493.1 nm in the dark after light exposure at kexp = 428 nm is shown in Fig. 5(b) by circles for AppA148 (similar results obtained for AppA126). The absorption reduces gradually in the relaxation from the signalling state (light-adapted system) to the receptor state (dark-adapted system). The curve in Fig. 5(b) is calculated using the relation t aðkp ; tÞ ¼ aðkp ; 1Þ þ ½aðkp ; 0Þ aðkp ; 1Þ exp ; sd;rec
(a)
(b)
ð1Þ where sd,rec is the intermediate recovery time in the dark at room temperature. The best fitting recovery time is sd,rec = 16.7 min for AppA148. For AppA126 a recovery time of sd,rec = 18.8 min was fitted (see Table 1). The fluorescence quantum distributions, EF(k), of dark-adapted and light-adapted AppA148 in aqueous solution at pH 8 together with riboflavin and FAD in aqueous solution at pH 7 are displayed in Fig. 6. For AppA126 qualitatively similar fluorescence quantum distributions have been measured as for AppA148 (curves not shown). The fluorescence quantum yields, /F, are calculated from EF(k) using the relation /F = EF(k)dk.
Fig. 4. Intensity dependent increase of absorption coefficients of APPA148 at wavelength kp = 493.1 nm due to light exposure at kexp = 428 nm. Experimental data are taken from Fig. 3(b). Curves are calculated by use of Eqs. (9)–(11) with parameters of Fig. 3 and Table 1. Quantum yield, /s, of signalling state formation is varied: (1) /s = 1, (2) 0.8, (3) 0.6, (4) 0.4, (5) 0.3, (6) 0.2, (7) 0.1, (8) 0.05, and (9) 0.01.
Fig. 5. (a) Temporal absorption coefficient increase at kp = 493.1 nm due to light exposure at kexp = 428 nm with an intensity of Iexp = 0.0108 W cm2 for AppA148. (b) Temporal absorption coefficient recovery in the dark at kp = 493.1 nm after light exposure at kexp = 428 nm with Iexp = 0.0108 W cm2. Data points are measured and curve is calculated (Eq. (1)) with recovery time constant, sd,rec, as fit parameter.
They are listed in Table 1. The fluorescence spectra of the dark-adapted AppA BLUF domains are approximately 16 nm blue-shifted compared to the spectra of riboflavin and FAD. The fluorescence quantum yield of riboflavin at pH 7 is approximately 26% [33,34], for FAD the fluorescence quantum yield at pH 7 is reduced to approximately 3.7% because of the presence of fluorescent un-stacked conformation and nearly non-fluorescent stacked conformation [40,41]. The AppA BLUF domains in the dark state have only a fluorescence quantum efficiency of about 2%. In the lightadapted state this low efficiency reduces further by a factor of 10 for AppA126 and by a factor of 20 for AppA148. The measured degrees of fluorescence polarization, PF, for the dark-adapted and light-adapted AppA BLUF domains are listed in Table 1. They are reasonably high indicating that the non-covalently bound flavin chromophores are hindered in free rotation by the protein binding pocket. The degree of fluorescence polarization of riboflavin in neutral aqueous solution is PF 0.0146 [42] indicating a molecular reorientation time short compared to the fluorescence lifetime [43]. For the light-adapted forms the degree of fluorescence polarization is somewhat lower than for the darkadapted forms. Below it will be shown that a small fraction of the flavins remains in a state of long fluorescence
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
Fig. 6. Fluorescence quantum distribution, EF(k), of dark-adapted and light-adapted APPA148 in aqueous solution at pH 8 together with fluorescence quantum distribution of riboflavin and FAD in aqueous solution at pH 7.
lifetime even under light exposure (this flavin fraction is likely adsorbed to AppA protein). This fraction has a lower degree of fluorescence polarization and dominates the fluorescence contribution under light-adapted conditions. The radiative lifetime of a chromophore is coupled to the absorption strength by the Strickler–Berg formula [44–46] R Z 8pc0 n3F em EF ðkÞdk ra ðkÞ 1 R dk; ð2Þ srad ¼ 3 nA E ðkÞk dk abs k em F where nF is the average refractive index in the fluorescence region, and nA is the average refractive index in the region of the first absorption band. The integrals extend over the fluorescence region (em) and over the S0–S1 absorption band (abs). For a single-component chromophore the radiative lifetime is alternatively given by sF srad ¼ . ð3Þ /F The radiative lifetime of riboflavin in water at pH 7 was determined to be srad,RF 19 ns [34]. This same value, i.e. srad,r 19 ns, is expected for the dark-adapted BLUF domains AppA148,r and AppA126,r by inspection of Eq. (2), the absorption cross-section spectra of Fig. 2, and the shapes of fluorescence quantum distributions of Fig. 6. The radiative lifetimes of the light-adapted forms, AppA148,s and AppA126,s, are estimated to be srad,s 20 ns by use of Eq. (2) with ra,s(k) from Fig. 2 and EF,s(k) of Fig. 6. The temporal fluorescence signal behaviour of darkadapted and light-adapted AppA148 is displayed in
149
Fig. 7(a)–(c). A similar fluorescence-signal-behaviour was obtained for AppA126 (curves not shown). In Fig. 7(a) the temporal fluorescence signals were recorded by picosecond pulse excitation (kL = 400 nm, DtL = 4 ps) and fluorescence detection with a microchannel-plate photomultiplier and a fast digital oscilloscope. The response function of the detection system is shown by the dotted curve (approximately Gaussian shape with 1/e-time constant of sresp 370 ps). Outside the response-function region the fluorescence signals fit reasonable well to single exponential decays. The best fitting decay times are given in the figure and are listed in Table 1 together with the results for AppA126. They are about sF,r,sl 900 ps for the dark-adapted forms, and sF,nc 1.1–1.6 ns for the light-adapted forms. The fluorescence signal height of the light-adapted AppA BLUF domains is only 0.075 times the fluorescence signal height of the dark-adapted forms. The long fluorescence lifetimes and the low fluorescence quantum yields indicate static fluorescence quenching, i.e. the presence of a multi-component system with different fluorescence quenching surroundings (at least one component with moderate quenching surrounding and one component with strong quenching surrounding have to be present, see below) [32,36,47,48]. For AppA148 the dark-adapted (thin solid curve in Fig. 7(a)) and light-adapted fluorescence signal measurements (thin dashed curve in Fig. 7(a)) have been repeated after 100-fold dilution of the solution with pH 8 phosphate buffer. The dark-adapted fluorescence signal is not affected by the dilution within our experimental accuracy. But after light adaptation for the diluted sample the peak fluorescence signal height is reduced and a longer lasting fluorescence signal is observed than for the undiluted sample. It is thought that the small fraction of non-interacting flavin molecules is desorbed from the protein and is staying in the solvent. There the free riboflavin, FMN, and un-stacked FAD molecules have longer fluorescence lifetimes. The peak signal reduction indicates some lifetime shortening of the cofactor (flavins) in the signalling state (stronger quenching-effect of the BLUF domains undisturbed by neighbouring domains). In Fig. 7(b) (dark-adapted BLUF domain) and Fig. 7(c) (light-adapted BLUF domain) the temporal fluorescence development was recorded by non-linear optical up-conversion of the fluorescence light with variable time-delayed fundamental femtosecond laser pulses in a BBO crystal [36,37]. The line-connected circles were measured. The dotted curves show the response function of the measurement system. It is nearly triangular shaped with a decay half-width of sresp 300 fs. The dark-adapted AppA148 sample in Fig. 7(b) reveals a component with a short fluorescence lifetime of sF,r,f 230 ps (for AppA126 sF,r,f 120 ps) in addition to the component with a slow fluorescence lifetime of
150
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
(a)
(b)
(c) Fig. 7. (a) Normalized temporal fluorescence signals, SF(t)/SF,max,dark, of dark-adapted (thick solid curves) and light-adapted (thick dashed curves) AppA148 in aqueous solution at pH 8. The thin solid curve (dark-adapted) and the thin dashed curve (light-adapted) belong to AppA148 diluted in pH 8 phosphate buffer by a factor of 100. Fluorescence excitation with femtosecond laser system (kL = 400 nm, pulse duration DtL = 4 ps, beam diameter at sample 3.5 mm, pulse energy attenuated to micro-channel-plate photomultiplier signal height in the 5–50 mV range). Dotted line shows experimental response function of the detection system. Dash-dotted lines are single-exponential decay fits according to SF(t) = SF(t0)exp[(t t0)/ sF]. The fit parameters are given in the figure. (b) Femtosecond laser up-converted fluorescence signal versus femtosecond pulse delay time, t, for dark-adapted AppA148 in aqueous solution at pH 8 and at room temperature (pulse duration DtL = 110 fs). Line-connected points are measured. Dash-dotted curve is a three-component exponential fit according to Eq. (4) with parameters given in the figure. Dotted line gives response function of the system. (c) Femtosecond laser up-converted fluorescence signal versus femtosecond pulse delay time, t, for light-adapted AppA148 in aqueous solution at pH 8 and at room temperature. Line-connected points are measured. Dash-dotted curve is a three-component exponential fit according to Eq. (4) with parameters given in the figure. Dotted line gives response function of the system.
sF,r,sl 940 ps (for AppA126 sF,r,sl 870 ps) which was detected in Fig. 7(a). The light-adapted AppA148 sample in Fig. 7(c) shows fluorescence lifetimes of sF,s,f 2.8 ps and sF,s,sl 48 ps
(for AppA126 sF,s,f 2.2 ps and sF,s,sl 32 ps),. A small component with a fluorescence lifetime of sF,nc 1.1 ns (for AppA126 sF,nc 1.6 ns) range was found in Fig. 7(a).
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
The dash-dotted curves in Fig. 7(b) and (c) show three-component exponential fits using the relation X S F ðtÞ ¼ S F ðt0 Þ xi exp½ðt t0 Þ=sF;i ; ð4Þ i¼f;sl;nc
of component i with fluowhere xi is the mole-fraction P rescence lifetime sF,i ( i¼f;sl;nc xi ¼ 1). The fit parameters xi and sF,i are listed in Fig. 7(b) and (c) and are collected in Table 1. The three-conformation composition of the AppA BLUF domains implies that the fluorescence quantum efficiency, /F, of the domains is composed of three contributions according to X X sF;i xi /F;i ¼ xi . ð5Þ /F ¼ s rad i¼f;sl;nc i¼f;sl;nc In the regression fits of Fig. 7(b) and (c) the mole fractions, xi, of AppA conformations are assumed to be the same in the dark-adapted state and the lightadapted state (three distinct components). The timeconstant, sF,nc, of the small longest-time component is assumed to be the same for the dark-adapted and the light-adapted state (no formation of a signalling state, time constant taken from Fig. 7(a)). The other two components are thought to make conformational changes (likely formation of hydrogen bonds) in going from the dark-adapted state to the light-adapted state. The time constant, sF,r,sl, of the slow dark-adapted form is taken from Fig. 7(a). Eq. (5) is used to implicitly determine the mole-fraction of one component in the darkadapted and the light-adapted state. The other parameters are determined in the fits.
4. Discussion In earlier experiments the cofactor of the flavoprotein AppA from R. sphaeroides was found to be FAD [8,15– 17,49]. In [22] it is shown, that upon heterologous expression of the BLUF domain of AppA from R. sphaeroides in E. coli, the BLUF domain may noncovalently bind all naturally occurring flavins, namely riboflavin, FMN, and FAD. Our thin-layer chromatography and fluorescence analysis reveals the presence of riboflavin, FMN, FAD, and a non-moving isoalloxazine moiety in AppA148 and AppA126 (see Table 1). The dominant chromophore was riboflavin. The absorption and fluorescence spectroscopic behaviour and the photo-cycle dynamics seems to be independent of the specific flavin chromophore. The low fluorescence quantum yield and the long fluorescence lifetime, observed in measurements with 370 ps response function time constant, indicates the presence of static fluorescence quenching [32,36,47,48], i.e. the presence of a heterogeneous system with chromophores surrounded by moderate fluorescence quenchers and with chromophores surrounded by
151
strong fluorescence quenchers. The results of the fluorescence quantum yield measurements and of the timeresolved fluorescence measurements are best described by a three-component holo-protein system consisting of two photo-cycle active components of mole-fractions, xf and xsl, and a small component of mole-fraction, xnc, which does not undergo a photo-cycle. For this inactive component the flavin chromophores are likely adsorbed to the domain surface and not incorporated into the binding pocket (diffusion into the solvent at strong dilution, see above). For the two active fractions the isoalloxazine rings are non-covalently bound in the binding pocket of the protein with somewhat different distances to the quenching centre(s) (possibly Tyr-21 [20]). The fluorescence is already quenched in the dark state to sF,r,f = 120–240 ps (fraction xf) and sF,r,sl = 860–940 ps (fraction xsl). The photo-excitation causes a conformational rearrangement of the protein chromophore system leading to absorption spectroscopic shifts (ca. 16 nm red shift) and fluorescence life-time shortening to sF,s,f = 2.2–2.8 ps and sF,s,sl = 30–50 ps. After light switch-off the light-adapted conformations slowly relax back to the initial dark-adapted conformations with a time constant of sd,rec = 16–19 min. The fluorescence quenching in the receptor state and in the signalling state is likely to be due to photoinduced charge transfer complex formation by electron transfer from an amino acid residue (possibly Tyr-21) to the excited flavin chromophore. The charge-transfer complex lifetime (charge recombination time sch,rec) is expected to be in the sub-nanosecond range [50,51]. The fluorescence quenching in the signalling state is stronger than in the receptor state (shorter fluorescence lifetime in signalling state). It is known that the amino acids Tyr, Trp, His, and Cys form charge-transfer complexes with flavins by photoinduced reductive electron transfer [51–54]. For the AppA BLUF domains Tyr-21 is discussed to be responsible for fluorescence quenching [20]. Replacing Tyr-21 by Phe or Leu abolished the photo-cycle dynamics [20]. A schematic drawing of the photo-cycle dynamics of the two photo-active BLUF domain conformations of AppA are show in the right part of Fig. 8. The left part shows the photo-excitation dynamics of the non-active flavins, which are like adsorbed to the AppA BLUF domains. It is thought that after receptor state excitation a conformational reorganisation of the BLUF domains occurs towards signalling state formation. In the reorganisation process it is thought that hydrogen bonds are formed between amino acid residues (like asparagines and glutamine) and hydrogen bond active states of flavin like C(4)@O, N(3)–H, and C(2)@O [19] (causing spectral red shift). The rate constants of singlet excited-state deactivation of BLUFr,sl, BLUFr,f, BLUFs,sl, and BLUFs,f are given by
152
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
k F ¼ k rad þ k IC þ k ISC þ k ElT ¼ k F;0 þ k ElT ;
ð6Þ
where kF = sF1 is the inverse fluorescence lifetime, krad is the radiative rate, kIC is the rate of internal conversion, kISC is the rate of intersystem crossing, and kElT is the rate of electron transfer. The photo-physical relaxation rate, kF,0 = krad + kIC + kISC, is approximately given by the inverse fluorescence lifetime of riboflavin in 8 neutral aqueous solution, i.e. k F;0 s1 F;RF 2 10 1 s . The radiative rate, internal conversion rate, and intersystem crossing rate are also thought to be approximately given by the values of riboflavin, i.e. krad krad,RF 5.28 · 107 s1, kIC kIC,RF 7.1 · 107 s1, kISC kISC,RF 5.37 · 107 s1 [55]. The estimated intersystem-crossing and electron-transfer rate constants (Eq. (6)) for the BLUF domains are included in Table 1. The quantum efficiency, /i,j, of a process i in the subsystem j may be calculated by the relation /i;j ¼
k i;j . k F;j
ð7aÞ
The total quantum efficiency for a process i is given by /i ¼
X
xj /i;j .
ð7bÞ
j
The quantum efficiencies, /F,nc, /F,f,, /F,sl, /F, /ElT,f, /ElT,sl, /ElT, /ISC,nc, /ISC,f, /ISC,sl, and /ISC are included in Table 1. The slow dark recovery to the initial conformation in the dark at room temperature indicates some groundstate barrier. The barrier height may be estimated using an Arrhenius-type equation for the rate of transfer from the signalling state to the receptor state [56,57]
k d;rec ¼
1 sd;rec
Wb ¼ k 0 exp ; kB#
where k0 is the attempt frequency of barrier crossing, Wb is the energy barrier height from the signalling groundstate to the receptor ground-state, kB is the Boltzmann constant, and # is the temperature. The attempt frequency is given by the oscillation frequency of the potential well of the signalling state along the signalling-state to receptor-state coordinate. It is typically of the order of 1012–1013 Hz [58,59]. Using an attempt frequency of k0 = 1013 s1, barrier heights of Wb(AppA148) = 1.5016 · 1019 J (7554 cm1) and Wb(AppA126) = 1.5064 · 1019 J (7579 cm1) are calculated. It should be noted that the exact attempt frequency influences only slightly the barrier height since the barrier height depends logarithmically on the attempt frequency. Photo-induced conformational changes leading to enlarged hydrogen bonding [16] and deprotonation [21] have been discussed to stabilize the signalling state. The quantum yield of photo-induced signaling-state formation, /s, describes the efficiency of signalling state formation. /s, is extracted from the excitation intensity dependent absorption behaviour (Fig. 4). The absorption coefficient, aa,p(t), at exposure time, t, and probe wavelength, kp, is given by aa;p ðtÞ ¼ ½N r ðtÞ þ N nc ra;p þ N s ðtÞra;s;p ;
ð9Þ
where Nr(t) = N0 Ns(t) Nnc is the number density of chromophores in the receptor state, Ns(t) is the number density of chromophores in the signalling state, Nnc = xncN0 is the number density of non-convertible chromophores, and N0 is the total number density of chromophores. N0 is determined from the dark-adapted absorption coefficient, aa and the absorption cross-section ra according to N0 = aa/ra. ra,p is the absorption cross-section of the dark-adapted chromophores at kp, and ra,s,p is the absorption cross-section of the lightadapted chromophores at kp. The temporal development of the active receptor and signalling number densities is given by the following rate equation system: oN s ra;exc I exc Ns ¼ /s ðN 0 N nc N s Þ ; ot hmexc sd;rec oI exc ¼ ra;exc ðN 0 N s ÞI exc oz ra;s;exc N s I exc as;exc I exc ;
Fig. 8. Photo-cycle scheme of photo-active BLUF domains of AppAcon (right), and photo-excitation scheme of BLUF domains, AppAnc, not taking part in receptor state – signalling state photo-cycle dynamics (left).
ð8Þ
ð10Þ
ð11Þ
where z is the coordinate along the propagation direction. ra,exc is the absorption cross-section of the darkadapted chromophores at the excitation wavelength kexc, and ra,s,exc is the absorption cross-section of the lightadapted chromophores at kexc. sd,rec is the signal state recovery time in the dark. as,exc is the scattering coefficient at the excitation wavelength. The absorption cross-section values of the dark and light-adapted
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
samples at kexc = 428 nm and kp = 493.1 nm are listed in Table 1. The solid curves in Fig. 4 are calculated for different quantum yields, /s, by numerical solution of the equation system (10), (11) and application of Eq. (9). The best fit to the experimental data gives /s(AppA148) = 0.24 and /s(AppA126) = 0.25 (curves not shown). The results are listed in Table 1. The temporal development of the absorption spectra under long-time intense light exposure allows the determination of the quantum yield of photo-degradation of the chromophores in the light-adapted state (signalling state), /D,s. This temporal development at kp = 493.1 nm is shown in Fig. 5(a). The quantum yield of photo-degradation is defined as the number of degraded molecules to the number of absorbed photons. It is determined by the relation /D;s ¼
Ns;D . nph;abs
ð12Þ
½ aa;s;p ðt1 Þ aa;s;p ðt2 Þ‘ ; ra;s;p
ð13Þ
Rl Thereby Ns;D ¼ 0 N s;D dz is the length-integrated number density of degraded signalling state molecules, ‘ is the sample length, Ns,D, is the number density of degraded molecules, and nph,abs is the density of absorbed photons at the excitation wavelength kexc. Assuming that the degraded molecules do not absorb at the probing wavelength, then Ns,D is given by Ns;D ¼
ðt2 t1 ÞI exc ð1 T Þ ; hmexc
dark-adapted and light-adapted cofactor – protein conformation.
5. Conclusions The BLUF domains AppA148 and AppA126 from R. sphaeroides have been studied by continuous absorption spectroscopy, continuous fluorescence spectroscopy, and time-resolved fluorescence spectroscopy. The dominant chromophore, non-covalently bound to the binding pocket of the domains, was found to be riboflavin. Two photo-active receptor states and signalling states were found from fluorescence lifetime and fluorescence quantum yield measurements. A small fraction of flavins (1–2%) was found not to take part in photo-cycle dynamics. It is likely adsorbed to the domain surface. The quantum yield for photo-induced signalling state formation was determined by an analysis of the light intensity dependent transmission changes. A quantum efficiency of about 25% was determined. A detailed crystal structure analysis of dark-adapted and light-adapted AppA BLUF domains would be needed to clarify the domain conformations, and more detailed time-resolved absorption and emission spectroscopic analysis would help to understand the quencher – chromophore interactions.
Acknowledgements
a;s;p ðt2 Þ ¼ where aa;s;p ðt1 Þ ¼ ln½T s;p ðt1 Þ=‘ and a ln½T s;p ðt2 Þ=‘ are the length-averaged absorption coefficients at time t1 and t2. The number of absorbed photons per cross-sectional area within the time interval t2 t1 is nph;abs ¼
153
ð14Þ
where Iexc is the input light intensity, and T ¼ ½T ðt1 Þ þ T ðt2 Þ=2 is the average transmission at kexc during the time period t2 t1. In Fig. 5(a), after signalling state formation, no absorption coefficient decrease is observed. Taking the experimental accuracy into account, an upper limit of the quantum of yield of photo-degradation of /D,s,max 1 · 105 is determined for AppA148,s, and the same upper limit is obtained for AppA126,s (curves not shown). The chromophore composition, domain conformation composition, and the dark-adapted and lightadapted fluorescence lifetimes of the domains AppA148 and AppA126 are not exactly the same. The larger domain (AppA148) has longer receptor state and signalling state fluorescence lifetime and a somewhat faster signalling state recovery time. The protein size influences the
We thank Anja Merkel for excellent technical assistance. We thank Dr. G. Klug, Giessen for sending us the appA gene and the Deutsche Forschungsgemeinschaft (DFG) for support in the Research Group FOR 526 ‘‘Sensory Blue Light Receptors’’, which enabled this collaborative work.
References [1] B.L. Taylor, I.B. Zhulin, Microbiol. Mol. Biol. Rev. 63 (1999) 469. [2] W.L. Briggs, E. Huala, Annu. Rev. Cell Dev. Biol. 15 (1999) 33. [3] P. Hegemann, M. Fuhrmann, S. Kateriya, J. Phycol. 37 (2001) 668. [4] J.M. Christie, W.R. Briggs, J. Biol. Chem. 276 (2001) 11457. [5] W.R. Briggs, J.M. Christie, Trends Plant Sci. 7 (2002) 204. [6] C. Lin, The Plant Cell (Supplement) (2002) S207. [7] S. Crosson, S. Rajagopal, K. Moffat, Biochemistry 42 (2003) 2. [8] M. Gomelsky, G. Klug, Trends Biochem. Sci. 27 (2003) 497. [9] E. Liscum, D.W. Hodgson, T.J. Campbell, Plant Physiol. 133 (2003) 1229. [10] M.A. van der Horst, K.J. Hellingwerf, Acc. Chem. Res. 37 (2004) 13. [11] M. Ahmad, N. Grancher, M. Heil, R.C. Black, B. Giovani, P. Galland, D. Lardemer, Plant Physiol. 129 (2002) 774. [12] T. Kleine, P. Lockhart, A. Batschauer, Plant J. 35 (2003) 93. [13] K. Huang, C.F. Beck, PNAS 100 (2003) 6269.
154 [14] [15] [16] [17] [18]
[19] [20] [21] [22] [23] [24] [25]
[26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37]
P. Zirak et al. / Chemical Physics 315 (2005) 142–154 M. Gomelsky, S. Kaplan, J. Bacteriol. 177 (1995) 4609. M. Gomelsky, S. Kaplan, J. Biol. Chem. 273 (1998) 35319. S. Masuda, C.E. Bauer, Cell 110 (2002) 613. S. Braatsch, M. Gomelsky, S. Kuphal, G. Klug, Mol. Microbiol. 45 (2002) 827. M. Iseki, S. Matsunaga, A. Murakami, K. Ohno, K. Shiga, K. Yoshida, M. Sugai, T. Takahashi, T. Hori, M. Watanabe, Nature 415 (2002) 1047. S. Masuda, K. Hasegawa, A. Ishii, T. Ono, Biochemistry 43 (2004) 5304. B.J. Kraft, S. Masuda, J. Kikuchi, V. Dragnea, G. Tollin, J.M. Zaleski, C.E. Bauer, Biochemistry 42 (2003) 6726. W. Laan, M.A. van der Horst, I.H. van Stokkum, K.J. Hellingwerf, Photochem. Photobiol. 78 (2003) 290. W. Laan, T. Bednarz, J. Heberle, K.J. Hellingwerf, Photochem. Photobiol. Sci. 3 (2004) 1011. C. Bauer, S. Elsen, L.R. Swem, D.L. Swem, S. Masuda, Phil. Trans. R. Soc. Lond. B 358 (2002) 147. H. Shimada, K. Iba, K. Takamiya, Plant Cell Physiol. 33 (1992) 471. M. Gauden, S. Yeremenko, W. Laan, I.H.M. van Stokkum, J.A. Ihalainen, R. van Grondelle, K.J. Hellingwerf, J.T.M. Kennis, Biochemistry 44 (2005) 3653. G.F. Bohren, D.R. Huffman, Absorption and Scattering of Light by Small Particles, John Wiley and Sons, New York, 1983. H. Gratz, A. Penzkofer, P. Weidner, J. Non.-Cryst. Solids 189 (1995) 55. R. Weigand, F. Rotermund, A. Penzkofer, Chem. Phys. 220 (1997) 373. W. Holzer, A. Penzkofer, M. Fuhrmann, P. Hegemann, Photochem. Photobiol. 75 (2002) 479. A. Penzkofer, W. Leupacher, J. Lumin. 37 (1987) 61. W. Holzer, M. Pichlmaier, A. Penzkofer, D.D.C. Bradley, W.J. Blau, Chem. Phys. 246 (1999) 445. Th. Fo¨rster, Fluoreszenz Organischer Verbindungen, Vandenhoeck und Ruprecht, Go¨ttingen, Germany, 1951. G. Weber, F.W.J. Teale, Trans. Faraday Soc. 53 (1957) 646. P. Dro¨ssler, W. Holzer, A. Penzkofer, P. Hegemann, Chem. Phys. 282 (2002) 429. F. Do¨rr, Angew. Chem. 78 (1966) 457. B. Valeur, Principles of Fluorescence Spectoscopy, Wiley-VCH, Weinheim, 2002, p. 353. R. Schanz, S.A. Kovalenko, V. Kharlanov, N.P. Ernsting, Appl. Phys. Lett. 79 (2001) 566.
[38] V.G. Dmitriev, G.G. Gurzadyan, N.N. Nikogosyan, Handbook of Nonlinear Optical Crystals, Springer Verlag, Berlin, 1991. [39] A.L. Lehninger, D.L. Nelson, M.M. Cox, Principles of Biochemistry, second ed., Worth Publishers, New York, 1993. [40] S.D.M. Islam, T. Susdorf, A. Penzkofer, P. Hegemann, Chem. Phys. 295 (2003) 137. [41] P.A.W. van den Berg, K.A. Feenstra, A.E. Mark, H.J.C. Berendsen, A.J.W.G. Visser, J. Phys. Chem. B 106 (2002) 8858. [42] G. Weber, in: E.C. Slater (Ed.), Flavins and Flavoproteins, B.B.A. Library, vol. 8, Elsevier, Amsterdam, The Netherlands, 1966, p. 14. [43] C.A. Parker, Photoluminescence of Solutions, Elsevier, Amsterdam, The Netherlands, 1968. [44] S.J. Strickler, R.A. Berg, J. Chem. Phys. 37 (1962) 814. [45] J.B. Birks, D.J. Dyson, Proc. R. Soc. London Ser A 257 (1963) 135. [46] A.V. Deshpande, A. Beidoun, A. Penzkofer, G. Wagenblast, Chem. Phys. 148 (1990) 141. [47] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, second ed., Kluwer Academic/Plenum Publishers, New York, 1999 (Chapter 8). [48] P. Dro¨ssler, W. Holzer, A. Penzkofer, P. Hegemann, Chem. Phys. 286 (2003) 409. [49] S. Braatsch, G. Klug, Photosynth. Res. 79 (2004) 45. [50] D. Zhong, S.K. Pal, C. Wan, A.H. Zewail, PNAS 98 (2001) 11873. [51] D. Zhong, A.H. Zewail, PNAS 98 (2001) 11867. [52] N. Mataga, H. Chosrowjan, Y. Shibata, F. Tanaka, Y. Nishina, K. Shiga, J. Phys. Chem. B 104 (2001) 10667. [53] P.A.W. van den Berg, A.J.W.C. Visser, in: B. Valeur, J.-C. Brochon (Eds.), New Trends in Fluorescence Spectroscopy: Applications to Chemical and Life Sciences, Springer Verlag, New York, 2001, p. 457. [54] C.W.M. Kay, E. Schleicher, A. Kuppig, H. Hofner, W.E. Ru¨diger, M. Schleicher, M. Fischer, A. Becher, S. Weber, G. Richter, J. Biol. Chem. 278 (2003) 10973. [55] Sh.D.M. Islam, A. Penzkofer, P. Hegemann, Chem. Phys. 291 (2003) 97. [56] G.R. Fleming, Chemical Applications of Ultrafast Spectroscopy, Oxford University Press, Oxford, 1986. [57] J. Schmidt, A. Penzkofer, J. Chem. Phys. 91 (1989) 1403. [58] L. Eckertova´, in: Physics of Thin Films, second ed., Oxford University Press, Oxford, 1982, p. 874. [59] J. Schmidt, A. Penzkofer, Chem. Phys. 117 (1987) 265.
Absorption and fluorescence spectroscopic characterization of BLUF domain of AppA from Rhodobacter sphaeroides P. Zirak a, A. Penzkofer
a,*
, T. Schiereis b, P. Hegemann b, A. Jung c, I. Schlichting
c
a
c
Institut II – Experimentelle und Angewandte Physik, Universita¨t Regensburg, Universita¨tstrasse 31, D-93053 Regensburg, Germany b Institut fu¨r Biologie, Experimentelle Biophysik, Humboldt-Universita¨t zu Berlin, Invalidenstrasse 42, D-10115 Berlin, Germany Max-Planck-Institut fu¨r medizinische Forschung, Abteilung Biomolekulare Mechanismen, Jahnstrasse 29, D-69120 Heidelberg, Germany Received 26 December 2004; accepted 13 April 2005 Available online 13 May 2005
Abstract The BLUF domain of the transcriptional anti-repressor protein AppA from the non-sulfur anoxyphototrophic purple bacterium Rhodobacter sphaeroides was characterized by absorption and emission spectroscopy. The BLUF domain constructs AppA148 (consisting of amino-acid residues 1–148) and AppA126 (amino-acid residues 1–126) are investigated. The cofactor of the investigated domains is found to consist of a mixture of the flavins riboflavin, FMN, and FAD. The dark-adapted domains exist in two different active receptor conformations (receptor states) with different sub-nanosecond fluorescence lifetimes (BLUFr,f and BLUFr,sl) and a small non-interacting conformation (BLUFnc). The active receptor conformations are transformed to putative signalling states (BLUFs,f and BLUFs,sl) of low fluorescence efficiency and picosecond fluorescence lifetime by blue-light excitation (light-adapted domains). In the dark at room temperature both signalling states recover back to the initial receptor states with a time constant of about 17 min. A quantum yield of signalling state formation of about 25% was determined by intensity dependent transmission measurements. A photo-cycle scheme is presented including photo-induced charge transfer complex formation, charge recombination, and protein binding pocket reorganisation. 2005 Elsevier B.V. All rights reserved. Keywords: BLUF domain; AppA; Blue-light photoreceptor; Rhodobacter sphaeroides; Absorption spectroscopy; Fluorescence spectroscopy; Fluorescence up-conversion; Photo-cycle; Flavoprotein; FAD; FMN; Riboflavin
1. Introduction The blue-light response of biological organisms is an active field of research (for recent reviews see [1–10]). There are three classes of blue-light receptors, the cryptochromes [4,6,11,12] with sensing Cry domains, the phototropins Phot with LOV domains [5,13], and the regulatory proteins with BLUF domains [8,10,14–18]. The sensing domains non-covalently bind flavin-adenine-dinucleotide (FAD) in the case of Cry domains * Corresponding author. Tel.: +49 0 941 943 2107; fax: +49 0 941 943 2754. E-mail address: [email protected] (A. Penzkofer).
0301-0104/$ - see front matter 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.chemphys.2005.04.008
and BLUF domains, and flavin-mononucleotide (FMN) in the case of LOV domains. BLUF domains (sensor for blue light using FAD) are the light sensors of the multi-domain protein AppA from the purple non-sulfur anoxyphototrophic proteobacterium Rhodobacter sphaeroides [15], of PAC (photo-activated adenylylcyclase) from the unicellular flagellate Euglena gracilis [18], of Slr1694 from the cyanobacterium Synechocystis sp. PCC6803 [19], and of flavoproteins of at least 15 other microorganisms [8]. Blue-light excitation of BLUF domains leads to a slight red-shift of the absorption band and a recovery to the initial absorption behaviour in the dark [15,16,18–22].
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
The flavoprotein AppA [8,14–17] from the non-sulfur anoxyphototropic purple bacterium R. sphaeroides controls photosynthetic gene expression in response to blue light exposure as well as in response to changes in the cellular redox state by functioning as an antirepressor of the photosynthesis repressor protein PpsR [8,16,17,19,23]. Blue-light exposure of AppA in R. sphaeroides represses puf and puc operon expression in this bacterium [17,24]. AppA consists of an N-terminal BLUF domain and a cystein-rich C-terminal domain. The amino acid sequence of the BLUF domain of AppA is found in [8,15,20,21]. The photo-cycle dynamics of AppA by forming a long-living intermediate with slightly red-shifted absorption band was first described in [16]. In [20] a detailed spectroscopic and mutational analysis of the BLUF domain of AppA1-156 (short: AppA156, consisting of amino residues 1–156) was carried out including ultraviolet and visible spectroscopy, fluorescence measurements, nuclear magnetic resonance (NMR) spectroscopy, laser flash photolysis, and gel filtration chromatography. The photo-dynamics is interpreted in terms of flavin stacking with the amino acid Tyr-21 leading to fluorescence reduction [20]. It is proposed that light exposure strengthens a hydrogen bond between flavin and Tyr-21 leading to a stable local conformational change in AppA1-156 [20]. In [21] the photo-cycle of AppA5-125 was studied by UV–Vis spectroscopy, Fourier-transform infrared spectroscopy, pH measurements and site-directed mutagenesis. It is expected that AppA5-125 in its dark-adapted state is protonated [N(5)-H], that it becomes exposed to solvent by blue-light exposure (light-adapted state, signalling state), and that an intra-molecular proton transfer from N(5) to anionic Tyr-21 forms the basis for the stabilisation of the signalling state [21]. Very recently – after finishing this paper – ultrafast time-resolved absorption and fluorescence spectroscopic studies on a femtosecond to nanosecond timescale and laser flash-photolysis studies on a nanosecond to microsecond timescale have been performed [25] to explore the photo-cycle dynamics of AppA5-125. In the BLUF domain Slr1694 from Synechocystis sp. PCC6803 light-induced Fourier transform infrared (FTIR) spectroscopy indicated a weakening of the C(4)@O and C(2)@O bonding and a strengthening of the N(1)C(10a) and C(4a)N(5) bonding [19]. In investigating a recent heterologous expression of the BLUF domain from AppA [22] it was found that riboflavin, FMN, and FAD non-covalently bind to the domain and all three homologues have very similar blue-light photo-cycle dynamics. In this paper the BLUF domains AppA148 (domain consists of amino acid residues 1–148 with C-terminal His-tag) and AppA126 (amino acid residues 1–126) from R. sphaeroides are characterized by absorption and emission spectroscopy. The flavin chromophore compo-
143
sition is analysed by thin-layer chromatography and fluorescence quantum yield measurement. Two active BLUF domain conformations and a third small conformation which does not form an intermediate are revealed by fluorescence lifetime and fluorescence quantum yield analysis. The fluorescence quantum yields and fluorescence lifetimes of the domains in the dark-state and in the light-adapted state are determined. The photo-cycle dynamics of signalling state formation from the receptor state by light exposure and of the dark recovery of the signalling state to the receptor state at room temperature are studied. The quantum yields of photo-induced signalling state formation are determined. All spectroscopic studies have been carried out at room temperature.
2. Experimental A DNA-fragment encoding the BLUF protein AppA (Acc: L42555) from R. sphaeroides 2.4.1. was kindly provided by Dr. G. Klug (Giessen, Germany). Two fragments encoding amino acid 1–148 and 1–126, respectively, were amplified by PCR. AppA 1–126 was identified to be what one could call the proteinÕs minimum BLUF domain by limited proteolysis of AppA 1–148 with Trypsin. Proteolytic fragments were analyzed by MALDI mass spectrometry and Edmann Nterminal sequencing. The two fragments were inserted between the NdeI and SacI restriction sites of a pET28a+ vector (Invitrogen, Karlsruhe), respectively. While the shorter construct only comprises an Nterminal His6-tag, the longer construct possesses both, N- and C-terminal His6-tag, rendering the purification procedure more effective. The C-terminal tag is attached to the AppA-coding region via the amino acid sequence AAPE and a short linker of vector-derived amino acid residues. Escherichia coli cells (strain BL21) were transformed with these two constructs. Selected clones were grown in LB plus Kanamycine over night at 30 C until the optical density at 600 nm had reached 0.5. The cultures were cooled down to 18 C, and production of the AppA fragments was initiated by addition of 0.7 mM IPTG. Protein expression proceeded over night at 18 C. The proteins were purified via Ni-NTA resin according to the instructions of the supplier (Quiagen, Hilden, Germany). They were subsequently exchanged in storage buffer (10 mM phosphate buffer, pH 8, 10 mM NaCl, 100 lM phenylmethylsulfonyl fluoride (PMSF)). The flavin cofactors non-covalently bound to the AppA BLUF domains were determined by chromophore extraction from the protein, followed by thinlayer chromatography and fluorescence analysis. The chromophore was extracted from the protein following
144
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
the procedure described in [22]: The samples were boiled in 70 vol.% ethanol water solution for 1 min, and then they were centrifuged, the supernatants were lyophilized to dryness and re-suspended in 35 vol.% ethanol. The fluorescence quantum yield, /F, of the extract solution of AppA126 (/F = 0.27) was determined and compared to the fluorescence quantum yields of FAD (/F,FAD = 0.127), FMN (/F,FMN = 0.335), and riboflavin (/F,RF = 0.285) in 35 vol.% ethanol. Neglecting the fluorescence quantum yield difference between FMN and riboflavin, the mole-fraction, xFAD, of FAD in the extract was determined from the relation /F,extract = xFAD/F,FAD + (1 xFAD)/F,RF to be xFAD(AppA126) 0.15. The chromophore composition was visualized by thin-layer chromatography of FAD, FMN, riboflavin, AppA126, and AppA148 on a silica gel 60 F254 aluminium foil (Merck # 105554, Darmstadt) with nbutanol/acetic acid/water 3:1:1 (v/v) as carrier medium. The fluorescence traces photographed under a 254 nm UV lamp are displayed in Fig. 1(a). Comparison with the references traces indicates that there are present in AppA126 and AppA148 a non-movable isoalloxazine component, riboflavin (dominant part), FMN and FAD. The spots were analysed by measuring their time-integrated fluorescence signal with a photomultiplier tube after picosecond laser pulse excitation at 400 nm. The determined mole-fractions of riboflavin, FMN, FAD, and non-moveable isoalloxazine moiety are listed in Table 1. The structural formulae of the flavins are shown in Fig. 1(b). The absorption cross-section spectra of dark-adapted AppA148 and AppA126 are determined by transmission measurements with a commercial spectrophotometer (Beckman type ACTA M IV). The transmission spectra, T(k), are converted to absorption coefficient spectra, a(k), by the relation T(k) = exp[a(k)‘], where ‘ is the sample path length. The absorption-coefficient spectra contain a scattering contribution, i.e. a(k) = aa(k) + as(k), where aa is the pure absorption coefficient and as is the scattering coefficient. The scattering contribution is approximated by as(k) = as(k0) (k0/k)4 [26–28] where k0 is a wavelength in the transparency region. The absorption cross-section spectra, ra(k), are calculated from the absorption coefficient spectra, aa(k), by calibration to the absorption cross-section spectrum of riboflavin at neutral pH, i.e. the absorption cross-section integrals extending over the S0–S1 and S0–S2 absorption band (k > 310 nm) are set equal since the same isois present [29] (ra ð~mÞ ¼ að~mÞ Ralloxazine chromophore R ra;R ð~mÞd~m= aa ð~mÞd~m, where ra,R(k) is the absorption cross-section spectrum of riboflavin, ~m ¼ 1=k is the wavenumber). The absorption cross-section spectra of the lightadapted samples (signalling state) are determined by exciting small-volume samples (1.5 · 1.5 · 3.5 mm3) with a high-pressure mercury lamp at kexp = 428 nm
(light intensity at sample Iexp 0.01 W cm2, exposure time Dtexp > 3 min), and probing the transmission with an attenuated tungsten lamp using a spectrometer – diode-array detection system. The absolute absorption cross-section spectra are extracted by exploiting the presence of isobestic points in the transmission spectra of dark-adapted and light-adapted samples (there the absorption cross-sections are the same). The spectral fluorescence behaviour was studied with a self-assembled fluorimeter in front-face collection arrangement [30,31]. The absolute intrinsic fluorescence quantum distributions [32], EF(k), and absolute intrinsic fluorescence quantum yields, /F, were determined by using coumarin 314T in ethanol (fluorescence quantum yield, /F,R = 0.87 [technical data sheet of Kodak]) or riboflavin in water buffered to pH 7 (/F,R = 0.26 [33,34]) as reference. The samples were excited by vertical polarized light and the fluorescence was detected under magic-angle polarizer orientation (polarizer transmission direction under an angle of 54.7 to the vertical). The degree of fluorescence polarization [35], PF = (SF,k SF,?)/(SF,k + SF,?), was determined by measuring the fluorescence signal polarized parallel (SF,k) and polarized perpendicular (SF,?) to the excitation light. For dark-adapted fluorescence quantum distribution measurement the excitation intensity was reduced to Iexp 1.6 · 103 W cm2 (excitation wavelength 428 nm) and the fluorescence accumulation time was extended to Dtexp = 10 min, whereby during measurement the solution in a 4 mm thick cell was circulated with a magnetic stirrer. In the case of light-adapted fluorescence quantum distribution measurement, before accumulation of fluorescence light a small-volume cell (1.5 · 1.5 · 3.5 mm3) was excited for several minutes at an intensity of Iexp 0.01 W cm2. The temporal fluorescence behaviour was studied by short laser pulse excitation at 400 nm with a Ti:sapphire femtosecond oscillator-amplifier laser system (laser system Hurricane from Spectra-Physics). For the time range in the sub-nanosecond to nanosecond region, the laser was operated at a pulse-duration of DtL = 4 ps, and the fluorescence signal was detected with a microchannel-plate photomultiplier (Hamamatsu, type R1564-U01) together with a high-speed digital oscilloscope (LeCroy, type DSO 9362). For the time-range in the sub-picosecond to picosecond region, the laser was operated at a pulse-duration of 110 fs and fluorescence-up-conversion technique was applied for time resolved signal detection [36,37]: The samples were excited with frequency doubled femtosecond pulses at 400 nm, and the generated fluorescence signals (frequency mF) were frequency up-converted with the fundamental laser pulses at 800 nm (frequency mG) in a non-linear optical crystal (BBO crystal of 0.2 mm thickness [38]) by non-collinear phase-matched sum-frequency generation
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
145
Fig. 1. (a) Thin-layer chromatography traces of FAD, FMN, riboflavin, AppA126, and AppA148. (b) Structural formulae of riboflavin, flavinmononucleotide (FMN), and flavin-adenine-dinucleotide (FAD).
(frequency mS, type II phase-matching: mF(e) + mG(o) ! mS(e), o: ordinary polarized light, e: extraordinary polarized light [38]). For time resolution the gating fundamental laser pulse was time-delayed relative to the second-harmonic excitation pulse with a stepper-motor driven linear translation stage. The up-converted fluorescence signal passed through a broad band filter
(Schott glass UG11 of 10 mm thickness, transmission range from 270 to 380 nm) and was detected with a photomultiplier tube (Valvo, type PM2254B) and a highspeed digital oscilloscope (see above). The laser was operated at 1 Hz repetition rate. For dark-adapted measurements the solution in a 4 mm path-length cell was circulated by a magnetic stirrer. For light-adapted
146
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
Table 1 Parameters of AppA BLUF domains in aqueous solution at pH 8 and at room temperature AppA126
Comments
Parameter
AppA148
Chromophore composition xFAD xFMN xRF xiso
0.085 ± 0.03 0.22 ± 0.04 0.54 ± 0.04 0.155 ± 0.03
0.15 ± 0.04 0.12 ± 0.03 0.67 ± 0.05 0.06 ± 0.02
Domain composition xf xsl xnc
0.776 ± 0.03 0.212 ± 0.02 0.012 ± 0.004
0.548 ± 0.03 0.429 ± 0.03 0.023 ± 0.004
Fig. 7 Fig. 7 Fig. 7
Photo-cycle characterization /s sd,rec (min) Wb/(hc0) (cm1)
0.24 ± 0.03 16.7 ± 0.5 7554 ± 10
0.25 ± 0.03 18.8 ± 0.5 7579 ± 10
Fig. 4 Fig. 5(b) Eq. (8)
State dependent parameters /F /F,f /F,sl /F,nc PF sF,f (ps) sF,sl (ps) sF,nc (ps) kElT,f (s1) kElT,sl (s1) /ElT,f /ElT,sl /ElT /ISC,f /ISC,sl /ISC,nc /ISC ra,exc (cm2) ra,p (cm2) /D
Dark-adapted 0.02 ± 0.002 0.012 0.049 0.059 0.385 ± 0.01 232 ± 20 938 ± 50 1113 ± 100 4.11 · 109 8.66 · 108 0.954 0.812 0.922 0.0125 0.05 0.06 0.022 3.39 · 1017 9.3 · 1018
Light-adapted (9.1 ± 0.5) · 104 0.00015 0.0025 0.059 0.37 ± 0.01 2.77 ± 1 47.6 ± 15 1113 ± 100 3.61 · 1011 2.08 · 1010 0.9994 0.9905 0.997 0.00015 0.0026 0.06 0.0014 2.76 · 1017 2.62 · 1017 <1 · 105
Dark-adapted 0.0235 ± 0.002 0.0064 0.046 0.082 0.345 ± 0.01 121.7 ± 20 867 ± 50 1557 ± 100 8.02 · 109 9.53 · 108 0.976 0.827 0.909 0.0065 0.047 0.084 0.027 3.22 · 1017 9.2 · 1018
Fig. Fig. Fig. Fig.
Light-adapted (2.2 ± 0.2) · 103 0.00012 0.0017 0.082 0.33 ± 0.01 2.2 ± 1 31.6 ± 10 1557 ± 100 4.54 · 1011 3.14 · 1010 0.9996 0.9937 0.997 0.00012 0.0017 0.084 0.0028 3.15 · 1017 2.54 · 1017 <1 · 105
1(a) 1(a) 1(a) 1(a)
Fig. 6 Eq. (5) Eq. (5) Eq. (5) Fig. 7 Fig. 7 Fig. 7a Eq. (6) Eq. (6) Eq. (7a) Eq. (7a) Eq. (7b) Eq. (7a) Eq. (7a) Eq. (7a) Eq. (7b) Fig. 2 Fig. 2 Eq. (12), Fig. 5a
Abbreviations: xFAD, xFMN, xRF, xiso, mole-fractions of FAD, FMN, riboflavin, and non-movable isoalloxazine moiety, respectively. xf, xsl, xnc, mole-fractions of BLUF domain conformations with fast fluorescence lifetime and convertible to signalling state, slow fluorescence lifetime and convertible to signalling state, and not convertible to signalling state, respectively. /s, quantum yield of signalling state formation. sd,rec, recovery time of BLUF domains after light exposure from signalling state to receptor state in the dark. Wb/(hc0), energy barrier from signalling state to receptor state (h is Planck constant, c0 is vacuum light velocity). /F is total fluorescence quantum yield. /F,f, /F,sl, /F,nc are fluorescence quantum yields of fast conformation, slow conformation, and non-convertible conformation of the BLUF domains, respectively. PF is degree of fluorescence polarization. sF,f, sF,sl, sF,nc, fluorescence lifetimes of fast, slow, and non-convertible BLUF conformations, respectively. kElT,f, kElT,sl, rate constants of electron transfer for BLUF domain conformations with fast and slow fluorescence lifetime, respectively. /ElT,f, /ElT,Sl, quantum efficiencies of photo-induced electron transfer of fast and slow BLUF domain conformation, respectively. /ElT is total quantum efficiency of photo-induced electron transfer. /ISC,f, /ISC,sl, /ISC,nc, estimated quantum yields of triplet formation for fast, slow, and non-convertible conformations of the BLUF domains. /ISC is estimated total quantum yield of triplet formation. ra,exc, ra,p, absorption cross-sections at excitation wavelength kexc = 428 nm and at probing wavelength kp = 493.1 nm, respectively. /D is quantum yield of photo-degradation. kexc = 428 nm; kp = 493.1 nm. Applied parameters from riboflavin in neutral aqueous solution: radiative rate constant, krad krad,RF 5.28 · 107 s1, intersystem-crossing rate constant kISC kISC,RF 5.37 · 107 s1 [55]. Error bars are estimates taking experimental accuracy into account.
measurements the solution in a 1 mm path-length cell (volume 1 · 10 · 10 mm3) was irradiated with a filtered high-pressure xenon lamp (Schott type BG38 broadband filter of 5 mm thickness transmitting from 350 to 600 nm and broadband interference filter with FWHM transmission from 350 to 440 nm, excitation intensity 0.02 W cm2). The photo-excitation and relaxation dynamics (photo-cycle behaviour) of the samples was studied by measuring the wavelength dependent transmission of a
probing tungsten lamp as a function of (i) the excitation light intensity at 428 nm, (ii) the long-time high-intensity exposure at 428 nm, and (iii) the temporal behaviour (recovery) after excitation shut-off.
3. Results The absorption cross-section spectra of dark-adapted AppA148 (AppA148,r) and riboflavin in aqueous solution
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
at pH 8 are shown in Fig. 2. The absorption cross-section spectrum of riboflavin is taken from [34]. The dark-adapted AppA BLUF domain spectrum is calibrated to the riboflavin spectrum byR equating their absorption cross-section integrals, ra ð~mÞd~m where ~m ¼ k1 is the wavenumber, in the wavelength range above 310 nm. The shapes of the absorption cross-section spectra of AppA148,r and riboflavin are quite similar, only below 310 nm the absorption cross-sections of the AppA BLUF domain is larger because of the absorption contribution of some amino acid residues of the AppA protein (mainly Trp, Tyr and Phe [39]). The vibronic structures of the S0–S1 absorption band peaking at 444 nm and of the S0–S2 absorption band peaking at 374 nm are more pronounced for the AppA BLUF domain than for riboflavin. Within our experimental accuracy the same absorption cross-section spectrum was obtained for AppA126,r as for AppA148,r (curve not shown). The absorption coefficient spectra at different excitation light intensities are shown in Fig. 3 for AppA148. Quite similar spectra were obtained for AppA126 (curves not shown). Part (a) shows the total absorption spectra, aðkÞ ¼ aa ðkÞ þ as ðkÞ, including absorption, aa ðkÞ, and scattering, as(k), together with the approximate scattering contributions, as(k) (k4 dependence). In part (b) the true absorption coefficient spectra, aa ðkÞ, are shown. The absorption coefficient spectra at the highest excitation intensity represent the absorption coefficient spectra of the photo-induced meta-stable intermediates (lightadapted BLUF domain, signalling state). Part (c) show the difference absorption spectra of exposed samples –
147
unexposed samples. Below 380 nm the pure absorption spectra and the difference spectra are less accurate because of some uncertainty in the extrapolated scattering contribution. The absolute absorption cross-section spectra of the signalling state of AppA148 (AppA148,s) and AppA126 (AppA126,s) are extracted from the isobestic point at 454 nm. There the absorption cross-sections of the dark-adapted and light-adapted forms are the same. The obtained spectrum for AppA148,s is displayed in Fig. 2 (dashed line). The S0–S1 absorption bands of the completely light-adapted forms are approximately 16 nm red-shifted compared to the dark-adapted forms. The excitation intensity dependence of the absorption coefficients, aa,p, at kp = 493.1 nm (wavelength position of largest absorption change between exposed and unexposed samples) is displayed by the dot-connected circles in Fig. 4 for AppA148. A similar dependence was obtained for AppA126 (data points are not shown). The rise in absorption will be exploited below to extract the quantum efficiency of signalling state formation by numerical simulations. The temporal development of the absorption coefficients, ap, in the case of high-intensity irradiation with Iexp 0.01 W cm2 at kexp = 428 nm is displayed in Fig. 5(a) for AppA148 (scattering contribution is in(a)
(b)
(c)
Fig. 2. Absorption cross-section spectra of riboflavin, APPA148,r, and APPA148,s in aqueous solution buffered to pH 8. APPA spectra are calibrated to riboflavin spectrum by equating the absorption crosssection integrals in the wavelength range above 310 nm.
Fig. 3. Dependence of absorption spectra of AppA148 on light exposure. Excitation wavelength kexp = 428 nm; cell thickness ‘ = 1.5 mm; cell area, 1.5 mm · 3.5 mm. Exposure time, texp = 1 min. Time interval between exposures is 30 min for Iexp 6 1.6 · 103 W cm2, and 40 min for Iexp > 1.6 · 103 W cm2. (a) Lengthaveraged total absorption spectra, aðkÞ ¼ ln½T ðkÞ=‘ ¼ aa ðkÞ þ as ðkÞ, including pure absorption, aa ðkÞ, and scattering contribution, as(k), together with approximate scattering spectra, as(k). (b) Pure absorption spectra, aa ðkÞ. (c) Pure difference absorption spectra, ad ðkÞ ¼ aa ðk; I exp Þ aa ðk; 0Þ.
148
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
cluded, same dependence obtained for AppA126). The signalling state is formed quickly. The absorption during long-time exposure remains constant indicating a high photo-stability of the light-adapted form. The time-dependent change of absorption at kp = 493.1 nm in the dark after light exposure at kexp = 428 nm is shown in Fig. 5(b) by circles for AppA148 (similar results obtained for AppA126). The absorption reduces gradually in the relaxation from the signalling state (light-adapted system) to the receptor state (dark-adapted system). The curve in Fig. 5(b) is calculated using the relation t aðkp ; tÞ ¼ aðkp ; 1Þ þ ½aðkp ; 0Þ aðkp ; 1Þ exp ; sd;rec
(a)
(b)
ð1Þ where sd,rec is the intermediate recovery time in the dark at room temperature. The best fitting recovery time is sd,rec = 16.7 min for AppA148. For AppA126 a recovery time of sd,rec = 18.8 min was fitted (see Table 1). The fluorescence quantum distributions, EF(k), of dark-adapted and light-adapted AppA148 in aqueous solution at pH 8 together with riboflavin and FAD in aqueous solution at pH 7 are displayed in Fig. 6. For AppA126 qualitatively similar fluorescence quantum distributions have been measured as for AppA148 (curves not shown). The fluorescence quantum yields, /F, are calculated from EF(k) using the relation /F = EF(k)dk.
Fig. 4. Intensity dependent increase of absorption coefficients of APPA148 at wavelength kp = 493.1 nm due to light exposure at kexp = 428 nm. Experimental data are taken from Fig. 3(b). Curves are calculated by use of Eqs. (9)–(11) with parameters of Fig. 3 and Table 1. Quantum yield, /s, of signalling state formation is varied: (1) /s = 1, (2) 0.8, (3) 0.6, (4) 0.4, (5) 0.3, (6) 0.2, (7) 0.1, (8) 0.05, and (9) 0.01.
Fig. 5. (a) Temporal absorption coefficient increase at kp = 493.1 nm due to light exposure at kexp = 428 nm with an intensity of Iexp = 0.0108 W cm2 for AppA148. (b) Temporal absorption coefficient recovery in the dark at kp = 493.1 nm after light exposure at kexp = 428 nm with Iexp = 0.0108 W cm2. Data points are measured and curve is calculated (Eq. (1)) with recovery time constant, sd,rec, as fit parameter.
They are listed in Table 1. The fluorescence spectra of the dark-adapted AppA BLUF domains are approximately 16 nm blue-shifted compared to the spectra of riboflavin and FAD. The fluorescence quantum yield of riboflavin at pH 7 is approximately 26% [33,34], for FAD the fluorescence quantum yield at pH 7 is reduced to approximately 3.7% because of the presence of fluorescent un-stacked conformation and nearly non-fluorescent stacked conformation [40,41]. The AppA BLUF domains in the dark state have only a fluorescence quantum efficiency of about 2%. In the lightadapted state this low efficiency reduces further by a factor of 10 for AppA126 and by a factor of 20 for AppA148. The measured degrees of fluorescence polarization, PF, for the dark-adapted and light-adapted AppA BLUF domains are listed in Table 1. They are reasonably high indicating that the non-covalently bound flavin chromophores are hindered in free rotation by the protein binding pocket. The degree of fluorescence polarization of riboflavin in neutral aqueous solution is PF 0.0146 [42] indicating a molecular reorientation time short compared to the fluorescence lifetime [43]. For the light-adapted forms the degree of fluorescence polarization is somewhat lower than for the darkadapted forms. Below it will be shown that a small fraction of the flavins remains in a state of long fluorescence
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
Fig. 6. Fluorescence quantum distribution, EF(k), of dark-adapted and light-adapted APPA148 in aqueous solution at pH 8 together with fluorescence quantum distribution of riboflavin and FAD in aqueous solution at pH 7.
lifetime even under light exposure (this flavin fraction is likely adsorbed to AppA protein). This fraction has a lower degree of fluorescence polarization and dominates the fluorescence contribution under light-adapted conditions. The radiative lifetime of a chromophore is coupled to the absorption strength by the Strickler–Berg formula [44–46] R Z 8pc0 n3F em EF ðkÞdk ra ðkÞ 1 R dk; ð2Þ srad ¼ 3 nA E ðkÞk dk abs k em F where nF is the average refractive index in the fluorescence region, and nA is the average refractive index in the region of the first absorption band. The integrals extend over the fluorescence region (em) and over the S0–S1 absorption band (abs). For a single-component chromophore the radiative lifetime is alternatively given by sF srad ¼ . ð3Þ /F The radiative lifetime of riboflavin in water at pH 7 was determined to be srad,RF 19 ns [34]. This same value, i.e. srad,r 19 ns, is expected for the dark-adapted BLUF domains AppA148,r and AppA126,r by inspection of Eq. (2), the absorption cross-section spectra of Fig. 2, and the shapes of fluorescence quantum distributions of Fig. 6. The radiative lifetimes of the light-adapted forms, AppA148,s and AppA126,s, are estimated to be srad,s 20 ns by use of Eq. (2) with ra,s(k) from Fig. 2 and EF,s(k) of Fig. 6. The temporal fluorescence signal behaviour of darkadapted and light-adapted AppA148 is displayed in
149
Fig. 7(a)–(c). A similar fluorescence-signal-behaviour was obtained for AppA126 (curves not shown). In Fig. 7(a) the temporal fluorescence signals were recorded by picosecond pulse excitation (kL = 400 nm, DtL = 4 ps) and fluorescence detection with a microchannel-plate photomultiplier and a fast digital oscilloscope. The response function of the detection system is shown by the dotted curve (approximately Gaussian shape with 1/e-time constant of sresp 370 ps). Outside the response-function region the fluorescence signals fit reasonable well to single exponential decays. The best fitting decay times are given in the figure and are listed in Table 1 together with the results for AppA126. They are about sF,r,sl 900 ps for the dark-adapted forms, and sF,nc 1.1–1.6 ns for the light-adapted forms. The fluorescence signal height of the light-adapted AppA BLUF domains is only 0.075 times the fluorescence signal height of the dark-adapted forms. The long fluorescence lifetimes and the low fluorescence quantum yields indicate static fluorescence quenching, i.e. the presence of a multi-component system with different fluorescence quenching surroundings (at least one component with moderate quenching surrounding and one component with strong quenching surrounding have to be present, see below) [32,36,47,48]. For AppA148 the dark-adapted (thin solid curve in Fig. 7(a)) and light-adapted fluorescence signal measurements (thin dashed curve in Fig. 7(a)) have been repeated after 100-fold dilution of the solution with pH 8 phosphate buffer. The dark-adapted fluorescence signal is not affected by the dilution within our experimental accuracy. But after light adaptation for the diluted sample the peak fluorescence signal height is reduced and a longer lasting fluorescence signal is observed than for the undiluted sample. It is thought that the small fraction of non-interacting flavin molecules is desorbed from the protein and is staying in the solvent. There the free riboflavin, FMN, and un-stacked FAD molecules have longer fluorescence lifetimes. The peak signal reduction indicates some lifetime shortening of the cofactor (flavins) in the signalling state (stronger quenching-effect of the BLUF domains undisturbed by neighbouring domains). In Fig. 7(b) (dark-adapted BLUF domain) and Fig. 7(c) (light-adapted BLUF domain) the temporal fluorescence development was recorded by non-linear optical up-conversion of the fluorescence light with variable time-delayed fundamental femtosecond laser pulses in a BBO crystal [36,37]. The line-connected circles were measured. The dotted curves show the response function of the measurement system. It is nearly triangular shaped with a decay half-width of sresp 300 fs. The dark-adapted AppA148 sample in Fig. 7(b) reveals a component with a short fluorescence lifetime of sF,r,f 230 ps (for AppA126 sF,r,f 120 ps) in addition to the component with a slow fluorescence lifetime of
150
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
(a)
(b)
(c) Fig. 7. (a) Normalized temporal fluorescence signals, SF(t)/SF,max,dark, of dark-adapted (thick solid curves) and light-adapted (thick dashed curves) AppA148 in aqueous solution at pH 8. The thin solid curve (dark-adapted) and the thin dashed curve (light-adapted) belong to AppA148 diluted in pH 8 phosphate buffer by a factor of 100. Fluorescence excitation with femtosecond laser system (kL = 400 nm, pulse duration DtL = 4 ps, beam diameter at sample 3.5 mm, pulse energy attenuated to micro-channel-plate photomultiplier signal height in the 5–50 mV range). Dotted line shows experimental response function of the detection system. Dash-dotted lines are single-exponential decay fits according to SF(t) = SF(t0)exp[(t t0)/ sF]. The fit parameters are given in the figure. (b) Femtosecond laser up-converted fluorescence signal versus femtosecond pulse delay time, t, for dark-adapted AppA148 in aqueous solution at pH 8 and at room temperature (pulse duration DtL = 110 fs). Line-connected points are measured. Dash-dotted curve is a three-component exponential fit according to Eq. (4) with parameters given in the figure. Dotted line gives response function of the system. (c) Femtosecond laser up-converted fluorescence signal versus femtosecond pulse delay time, t, for light-adapted AppA148 in aqueous solution at pH 8 and at room temperature. Line-connected points are measured. Dash-dotted curve is a three-component exponential fit according to Eq. (4) with parameters given in the figure. Dotted line gives response function of the system.
sF,r,sl 940 ps (for AppA126 sF,r,sl 870 ps) which was detected in Fig. 7(a). The light-adapted AppA148 sample in Fig. 7(c) shows fluorescence lifetimes of sF,s,f 2.8 ps and sF,s,sl 48 ps
(for AppA126 sF,s,f 2.2 ps and sF,s,sl 32 ps),. A small component with a fluorescence lifetime of sF,nc 1.1 ns (for AppA126 sF,nc 1.6 ns) range was found in Fig. 7(a).
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
The dash-dotted curves in Fig. 7(b) and (c) show three-component exponential fits using the relation X S F ðtÞ ¼ S F ðt0 Þ xi exp½ðt t0 Þ=sF;i ; ð4Þ i¼f;sl;nc
of component i with fluowhere xi is the mole-fraction P rescence lifetime sF,i ( i¼f;sl;nc xi ¼ 1). The fit parameters xi and sF,i are listed in Fig. 7(b) and (c) and are collected in Table 1. The three-conformation composition of the AppA BLUF domains implies that the fluorescence quantum efficiency, /F, of the domains is composed of three contributions according to X X sF;i xi /F;i ¼ xi . ð5Þ /F ¼ s rad i¼f;sl;nc i¼f;sl;nc In the regression fits of Fig. 7(b) and (c) the mole fractions, xi, of AppA conformations are assumed to be the same in the dark-adapted state and the lightadapted state (three distinct components). The timeconstant, sF,nc, of the small longest-time component is assumed to be the same for the dark-adapted and the light-adapted state (no formation of a signalling state, time constant taken from Fig. 7(a)). The other two components are thought to make conformational changes (likely formation of hydrogen bonds) in going from the dark-adapted state to the light-adapted state. The time constant, sF,r,sl, of the slow dark-adapted form is taken from Fig. 7(a). Eq. (5) is used to implicitly determine the mole-fraction of one component in the darkadapted and the light-adapted state. The other parameters are determined in the fits.
4. Discussion In earlier experiments the cofactor of the flavoprotein AppA from R. sphaeroides was found to be FAD [8,15– 17,49]. In [22] it is shown, that upon heterologous expression of the BLUF domain of AppA from R. sphaeroides in E. coli, the BLUF domain may noncovalently bind all naturally occurring flavins, namely riboflavin, FMN, and FAD. Our thin-layer chromatography and fluorescence analysis reveals the presence of riboflavin, FMN, FAD, and a non-moving isoalloxazine moiety in AppA148 and AppA126 (see Table 1). The dominant chromophore was riboflavin. The absorption and fluorescence spectroscopic behaviour and the photo-cycle dynamics seems to be independent of the specific flavin chromophore. The low fluorescence quantum yield and the long fluorescence lifetime, observed in measurements with 370 ps response function time constant, indicates the presence of static fluorescence quenching [32,36,47,48], i.e. the presence of a heterogeneous system with chromophores surrounded by moderate fluorescence quenchers and with chromophores surrounded by
151
strong fluorescence quenchers. The results of the fluorescence quantum yield measurements and of the timeresolved fluorescence measurements are best described by a three-component holo-protein system consisting of two photo-cycle active components of mole-fractions, xf and xsl, and a small component of mole-fraction, xnc, which does not undergo a photo-cycle. For this inactive component the flavin chromophores are likely adsorbed to the domain surface and not incorporated into the binding pocket (diffusion into the solvent at strong dilution, see above). For the two active fractions the isoalloxazine rings are non-covalently bound in the binding pocket of the protein with somewhat different distances to the quenching centre(s) (possibly Tyr-21 [20]). The fluorescence is already quenched in the dark state to sF,r,f = 120–240 ps (fraction xf) and sF,r,sl = 860–940 ps (fraction xsl). The photo-excitation causes a conformational rearrangement of the protein chromophore system leading to absorption spectroscopic shifts (ca. 16 nm red shift) and fluorescence life-time shortening to sF,s,f = 2.2–2.8 ps and sF,s,sl = 30–50 ps. After light switch-off the light-adapted conformations slowly relax back to the initial dark-adapted conformations with a time constant of sd,rec = 16–19 min. The fluorescence quenching in the receptor state and in the signalling state is likely to be due to photoinduced charge transfer complex formation by electron transfer from an amino acid residue (possibly Tyr-21) to the excited flavin chromophore. The charge-transfer complex lifetime (charge recombination time sch,rec) is expected to be in the sub-nanosecond range [50,51]. The fluorescence quenching in the signalling state is stronger than in the receptor state (shorter fluorescence lifetime in signalling state). It is known that the amino acids Tyr, Trp, His, and Cys form charge-transfer complexes with flavins by photoinduced reductive electron transfer [51–54]. For the AppA BLUF domains Tyr-21 is discussed to be responsible for fluorescence quenching [20]. Replacing Tyr-21 by Phe or Leu abolished the photo-cycle dynamics [20]. A schematic drawing of the photo-cycle dynamics of the two photo-active BLUF domain conformations of AppA are show in the right part of Fig. 8. The left part shows the photo-excitation dynamics of the non-active flavins, which are like adsorbed to the AppA BLUF domains. It is thought that after receptor state excitation a conformational reorganisation of the BLUF domains occurs towards signalling state formation. In the reorganisation process it is thought that hydrogen bonds are formed between amino acid residues (like asparagines and glutamine) and hydrogen bond active states of flavin like C(4)@O, N(3)–H, and C(2)@O [19] (causing spectral red shift). The rate constants of singlet excited-state deactivation of BLUFr,sl, BLUFr,f, BLUFs,sl, and BLUFs,f are given by
152
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
k F ¼ k rad þ k IC þ k ISC þ k ElT ¼ k F;0 þ k ElT ;
ð6Þ
where kF = sF1 is the inverse fluorescence lifetime, krad is the radiative rate, kIC is the rate of internal conversion, kISC is the rate of intersystem crossing, and kElT is the rate of electron transfer. The photo-physical relaxation rate, kF,0 = krad + kIC + kISC, is approximately given by the inverse fluorescence lifetime of riboflavin in 8 neutral aqueous solution, i.e. k F;0 s1 F;RF 2 10 1 s . The radiative rate, internal conversion rate, and intersystem crossing rate are also thought to be approximately given by the values of riboflavin, i.e. krad krad,RF 5.28 · 107 s1, kIC kIC,RF 7.1 · 107 s1, kISC kISC,RF 5.37 · 107 s1 [55]. The estimated intersystem-crossing and electron-transfer rate constants (Eq. (6)) for the BLUF domains are included in Table 1. The quantum efficiency, /i,j, of a process i in the subsystem j may be calculated by the relation /i;j ¼
k i;j . k F;j
ð7aÞ
The total quantum efficiency for a process i is given by /i ¼
X
xj /i;j .
ð7bÞ
j
The quantum efficiencies, /F,nc, /F,f,, /F,sl, /F, /ElT,f, /ElT,sl, /ElT, /ISC,nc, /ISC,f, /ISC,sl, and /ISC are included in Table 1. The slow dark recovery to the initial conformation in the dark at room temperature indicates some groundstate barrier. The barrier height may be estimated using an Arrhenius-type equation for the rate of transfer from the signalling state to the receptor state [56,57]
k d;rec ¼
1 sd;rec
Wb ¼ k 0 exp ; kB#
where k0 is the attempt frequency of barrier crossing, Wb is the energy barrier height from the signalling groundstate to the receptor ground-state, kB is the Boltzmann constant, and # is the temperature. The attempt frequency is given by the oscillation frequency of the potential well of the signalling state along the signalling-state to receptor-state coordinate. It is typically of the order of 1012–1013 Hz [58,59]. Using an attempt frequency of k0 = 1013 s1, barrier heights of Wb(AppA148) = 1.5016 · 1019 J (7554 cm1) and Wb(AppA126) = 1.5064 · 1019 J (7579 cm1) are calculated. It should be noted that the exact attempt frequency influences only slightly the barrier height since the barrier height depends logarithmically on the attempt frequency. Photo-induced conformational changes leading to enlarged hydrogen bonding [16] and deprotonation [21] have been discussed to stabilize the signalling state. The quantum yield of photo-induced signaling-state formation, /s, describes the efficiency of signalling state formation. /s, is extracted from the excitation intensity dependent absorption behaviour (Fig. 4). The absorption coefficient, aa,p(t), at exposure time, t, and probe wavelength, kp, is given by aa;p ðtÞ ¼ ½N r ðtÞ þ N nc ra;p þ N s ðtÞra;s;p ;
ð9Þ
where Nr(t) = N0 Ns(t) Nnc is the number density of chromophores in the receptor state, Ns(t) is the number density of chromophores in the signalling state, Nnc = xncN0 is the number density of non-convertible chromophores, and N0 is the total number density of chromophores. N0 is determined from the dark-adapted absorption coefficient, aa and the absorption cross-section ra according to N0 = aa/ra. ra,p is the absorption cross-section of the dark-adapted chromophores at kp, and ra,s,p is the absorption cross-section of the lightadapted chromophores at kp. The temporal development of the active receptor and signalling number densities is given by the following rate equation system: oN s ra;exc I exc Ns ¼ /s ðN 0 N nc N s Þ ; ot hmexc sd;rec oI exc ¼ ra;exc ðN 0 N s ÞI exc oz ra;s;exc N s I exc as;exc I exc ;
Fig. 8. Photo-cycle scheme of photo-active BLUF domains of AppAcon (right), and photo-excitation scheme of BLUF domains, AppAnc, not taking part in receptor state – signalling state photo-cycle dynamics (left).
ð8Þ
ð10Þ
ð11Þ
where z is the coordinate along the propagation direction. ra,exc is the absorption cross-section of the darkadapted chromophores at the excitation wavelength kexc, and ra,s,exc is the absorption cross-section of the lightadapted chromophores at kexc. sd,rec is the signal state recovery time in the dark. as,exc is the scattering coefficient at the excitation wavelength. The absorption cross-section values of the dark and light-adapted
P. Zirak et al. / Chemical Physics 315 (2005) 142–154
samples at kexc = 428 nm and kp = 493.1 nm are listed in Table 1. The solid curves in Fig. 4 are calculated for different quantum yields, /s, by numerical solution of the equation system (10), (11) and application of Eq. (9). The best fit to the experimental data gives /s(AppA148) = 0.24 and /s(AppA126) = 0.25 (curves not shown). The results are listed in Table 1. The temporal development of the absorption spectra under long-time intense light exposure allows the determination of the quantum yield of photo-degradation of the chromophores in the light-adapted state (signalling state), /D,s. This temporal development at kp = 493.1 nm is shown in Fig. 5(a). The quantum yield of photo-degradation is defined as the number of degraded molecules to the number of absorbed photons. It is determined by the relation /D;s ¼
Ns;D . nph;abs
ð12Þ
½ aa;s;p ðt1 Þ aa;s;p ðt2 Þ‘ ; ra;s;p
ð13Þ
Rl Thereby Ns;D ¼ 0 N s;D dz is the length-integrated number density of degraded signalling state molecules, ‘ is the sample length, Ns,D, is the number density of degraded molecules, and nph,abs is the density of absorbed photons at the excitation wavelength kexc. Assuming that the degraded molecules do not absorb at the probing wavelength, then Ns,D is given by Ns;D ¼
ðt2 t1 ÞI exc ð1 T Þ ; hmexc
dark-adapted and light-adapted cofactor – protein conformation.
5. Conclusions The BLUF domains AppA148 and AppA126 from R. sphaeroides have been studied by continuous absorption spectroscopy, continuous fluorescence spectroscopy, and time-resolved fluorescence spectroscopy. The dominant chromophore, non-covalently bound to the binding pocket of the domains, was found to be riboflavin. Two photo-active receptor states and signalling states were found from fluorescence lifetime and fluorescence quantum yield measurements. A small fraction of flavins (1–2%) was found not to take part in photo-cycle dynamics. It is likely adsorbed to the domain surface. The quantum yield for photo-induced signalling state formation was determined by an analysis of the light intensity dependent transmission changes. A quantum efficiency of about 25% was determined. A detailed crystal structure analysis of dark-adapted and light-adapted AppA BLUF domains would be needed to clarify the domain conformations, and more detailed time-resolved absorption and emission spectroscopic analysis would help to understand the quencher – chromophore interactions.
Acknowledgements
a;s;p ðt2 Þ ¼ where aa;s;p ðt1 Þ ¼ ln½T s;p ðt1 Þ=‘ and a ln½T s;p ðt2 Þ=‘ are the length-averaged absorption coefficients at time t1 and t2. The number of absorbed photons per cross-sectional area within the time interval t2 t1 is nph;abs ¼
153
ð14Þ
where Iexc is the input light intensity, and T ¼ ½T ðt1 Þ þ T ðt2 Þ=2 is the average transmission at kexc during the time period t2 t1. In Fig. 5(a), after signalling state formation, no absorption coefficient decrease is observed. Taking the experimental accuracy into account, an upper limit of the quantum of yield of photo-degradation of /D,s,max 1 · 105 is determined for AppA148,s, and the same upper limit is obtained for AppA126,s (curves not shown). The chromophore composition, domain conformation composition, and the dark-adapted and lightadapted fluorescence lifetimes of the domains AppA148 and AppA126 are not exactly the same. The larger domain (AppA148) has longer receptor state and signalling state fluorescence lifetime and a somewhat faster signalling state recovery time. The protein size influences the
We thank Anja Merkel for excellent technical assistance. We thank Dr. G. Klug, Giessen for sending us the appA gene and the Deutsche Forschungsgemeinschaft (DFG) for support in the Research Group FOR 526 ‘‘Sensory Blue Light Receptors’’, which enabled this collaborative work.
References [1] B.L. Taylor, I.B. Zhulin, Microbiol. Mol. Biol. Rev. 63 (1999) 469. [2] W.L. Briggs, E. Huala, Annu. Rev. Cell Dev. Biol. 15 (1999) 33. [3] P. Hegemann, M. Fuhrmann, S. Kateriya, J. Phycol. 37 (2001) 668. [4] J.M. Christie, W.R. Briggs, J. Biol. Chem. 276 (2001) 11457. [5] W.R. Briggs, J.M. Christie, Trends Plant Sci. 7 (2002) 204. [6] C. Lin, The Plant Cell (Supplement) (2002) S207. [7] S. Crosson, S. Rajagopal, K. Moffat, Biochemistry 42 (2003) 2. [8] M. Gomelsky, G. Klug, Trends Biochem. Sci. 27 (2003) 497. [9] E. Liscum, D.W. Hodgson, T.J. Campbell, Plant Physiol. 133 (2003) 1229. [10] M.A. van der Horst, K.J. Hellingwerf, Acc. Chem. Res. 37 (2004) 13. [11] M. Ahmad, N. Grancher, M. Heil, R.C. Black, B. Giovani, P. Galland, D. Lardemer, Plant Physiol. 129 (2002) 774. [12] T. Kleine, P. Lockhart, A. Batschauer, Plant J. 35 (2003) 93. [13] K. Huang, C.F. Beck, PNAS 100 (2003) 6269.
154 [14] [15] [16] [17] [18]
[19] [20] [21] [22] [23] [24] [25]
[26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37]
P. Zirak et al. / Chemical Physics 315 (2005) 142–154 M. Gomelsky, S. Kaplan, J. Bacteriol. 177 (1995) 4609. M. Gomelsky, S. Kaplan, J. Biol. Chem. 273 (1998) 35319. S. Masuda, C.E. Bauer, Cell 110 (2002) 613. S. Braatsch, M. Gomelsky, S. Kuphal, G. Klug, Mol. Microbiol. 45 (2002) 827. M. Iseki, S. Matsunaga, A. Murakami, K. Ohno, K. Shiga, K. Yoshida, M. Sugai, T. Takahashi, T. Hori, M. Watanabe, Nature 415 (2002) 1047. S. Masuda, K. Hasegawa, A. Ishii, T. Ono, Biochemistry 43 (2004) 5304. B.J. Kraft, S. Masuda, J. Kikuchi, V. Dragnea, G. Tollin, J.M. Zaleski, C.E. Bauer, Biochemistry 42 (2003) 6726. W. Laan, M.A. van der Horst, I.H. van Stokkum, K.J. Hellingwerf, Photochem. Photobiol. 78 (2003) 290. W. Laan, T. Bednarz, J. Heberle, K.J. Hellingwerf, Photochem. Photobiol. Sci. 3 (2004) 1011. C. Bauer, S. Elsen, L.R. Swem, D.L. Swem, S. Masuda, Phil. Trans. R. Soc. Lond. B 358 (2002) 147. H. Shimada, K. Iba, K. Takamiya, Plant Cell Physiol. 33 (1992) 471. M. Gauden, S. Yeremenko, W. Laan, I.H.M. van Stokkum, J.A. Ihalainen, R. van Grondelle, K.J. Hellingwerf, J.T.M. Kennis, Biochemistry 44 (2005) 3653. G.F. Bohren, D.R. Huffman, Absorption and Scattering of Light by Small Particles, John Wiley and Sons, New York, 1983. H. Gratz, A. Penzkofer, P. Weidner, J. Non.-Cryst. Solids 189 (1995) 55. R. Weigand, F. Rotermund, A. Penzkofer, Chem. Phys. 220 (1997) 373. W. Holzer, A. Penzkofer, M. Fuhrmann, P. Hegemann, Photochem. Photobiol. 75 (2002) 479. A. Penzkofer, W. Leupacher, J. Lumin. 37 (1987) 61. W. Holzer, M. Pichlmaier, A. Penzkofer, D.D.C. Bradley, W.J. Blau, Chem. Phys. 246 (1999) 445. Th. Fo¨rster, Fluoreszenz Organischer Verbindungen, Vandenhoeck und Ruprecht, Go¨ttingen, Germany, 1951. G. Weber, F.W.J. Teale, Trans. Faraday Soc. 53 (1957) 646. P. Dro¨ssler, W. Holzer, A. Penzkofer, P. Hegemann, Chem. Phys. 282 (2002) 429. F. Do¨rr, Angew. Chem. 78 (1966) 457. B. Valeur, Principles of Fluorescence Spectoscopy, Wiley-VCH, Weinheim, 2002, p. 353. R. Schanz, S.A. Kovalenko, V. Kharlanov, N.P. Ernsting, Appl. Phys. Lett. 79 (2001) 566.
[38] V.G. Dmitriev, G.G. Gurzadyan, N.N. Nikogosyan, Handbook of Nonlinear Optical Crystals, Springer Verlag, Berlin, 1991. [39] A.L. Lehninger, D.L. Nelson, M.M. Cox, Principles of Biochemistry, second ed., Worth Publishers, New York, 1993. [40] S.D.M. Islam, T. Susdorf, A. Penzkofer, P. Hegemann, Chem. Phys. 295 (2003) 137. [41] P.A.W. van den Berg, K.A. Feenstra, A.E. Mark, H.J.C. Berendsen, A.J.W.G. Visser, J. Phys. Chem. B 106 (2002) 8858. [42] G. Weber, in: E.C. Slater (Ed.), Flavins and Flavoproteins, B.B.A. Library, vol. 8, Elsevier, Amsterdam, The Netherlands, 1966, p. 14. [43] C.A. Parker, Photoluminescence of Solutions, Elsevier, Amsterdam, The Netherlands, 1968. [44] S.J. Strickler, R.A. Berg, J. Chem. Phys. 37 (1962) 814. [45] J.B. Birks, D.J. Dyson, Proc. R. Soc. London Ser A 257 (1963) 135. [46] A.V. Deshpande, A. Beidoun, A. Penzkofer, G. Wagenblast, Chem. Phys. 148 (1990) 141. [47] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, second ed., Kluwer Academic/Plenum Publishers, New York, 1999 (Chapter 8). [48] P. Dro¨ssler, W. Holzer, A. Penzkofer, P. Hegemann, Chem. Phys. 286 (2003) 409. [49] S. Braatsch, G. Klug, Photosynth. Res. 79 (2004) 45. [50] D. Zhong, S.K. Pal, C. Wan, A.H. Zewail, PNAS 98 (2001) 11873. [51] D. Zhong, A.H. Zewail, PNAS 98 (2001) 11867. [52] N. Mataga, H. Chosrowjan, Y. Shibata, F. Tanaka, Y. Nishina, K. Shiga, J. Phys. Chem. B 104 (2001) 10667. [53] P.A.W. van den Berg, A.J.W.C. Visser, in: B. Valeur, J.-C. Brochon (Eds.), New Trends in Fluorescence Spectroscopy: Applications to Chemical and Life Sciences, Springer Verlag, New York, 2001, p. 457. [54] C.W.M. Kay, E. Schleicher, A. Kuppig, H. Hofner, W.E. Ru¨diger, M. Schleicher, M. Fischer, A. Becher, S. Weber, G. Richter, J. Biol. Chem. 278 (2003) 10973. [55] Sh.D.M. Islam, A. Penzkofer, P. Hegemann, Chem. Phys. 291 (2003) 97. [56] G.R. Fleming, Chemical Applications of Ultrafast Spectroscopy, Oxford University Press, Oxford, 1986. [57] J. Schmidt, A. Penzkofer, J. Chem. Phys. 91 (1989) 1403. [58] L. Eckertova´, in: Physics of Thin Films, second ed., Oxford University Press, Oxford, 1982, p. 874. [59] J. Schmidt, A. Penzkofer, Chem. Phys. 117 (1987) 265.