This Article Abstract Full Text (PDF) All Versions of this Article: biophysj.106.086645v1 91/8/2991    most recent Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Borucki, B. Articles by Heyn, M. P. Search for Related Content PubMed PubMed Citation Articles by Borucki, B. Articles by Heyn, M. P.

The Transient Accumulation of the Signaling State of Photoactive Yellow Protein Is Controlled by the External pH

Berthold Borucki ^*, Chandra P. Joshi ^*, Harald Otto ^*, Michael A. Cusanovich  and Maarten P. Heyn ^*

^* Biophysics Group, Department of Physics, Freie Universität Berlin, 14195 Berlin, Germany; and Department of Biochemistry and Molecular Biophysics, University of Arizona, Tucson, Arizona, 85721

Correspondence: Address reprint requests to Maarten P. Heyn, Biophysics Group, Dept. of Physics, Freie Universität Berlin, Arnimallee 14, 14195 Berlin, Germany. Tel: 49-30-83856160; Fax: 49-30-83856299; E-mail: heyn{at}physik.fu-berlin.de.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
The signaling state of the photoreceptor photoactive yellow protein is the long-lived intermediate I₂'. The pH dependence of the equilibrium between the transient photocycle intermediates I₂ and I₂' was investigated. The formation of I₂' from I₂ is accompanied by a major conformational change. The kinetics and intermediates of the photocycle and of the photoreversal were measured by transient absorption spectroscopy from pH 4.6 to 8.4. Singular value decomposition (SVD) analysis of the data at pH 7 showed the presence of three spectrally distinguishable species: I₁, I₂, and I₂'. Their spectra were determined using the extrapolated difference method. I₂ and I₂' have electronic absorption spectra, with maxima at 370 ± 5 and 350 ± 5 nm, respectively. Formation of the signaling state is thus associated with a change in the environment of the protonated chromophore. The time courses of the I₁, I₂, and I₂' intermediates were determined from the wavelength-dependent transient absorbance changes at each pH, assuming that their spectra are pH-independent. After the formation of I₂' (2 ms), these three intermediates are in equilibrium and decay together to the initial dark state. The equilibrium between I₂ and I₂' is pH dependent with a pK_a of 6.4 and with I₂' the main species above this pK_a. Measurements of the pH dependence of the photoreversal kinetics with a second flash of 355 nm at a delay of 20 ms confirm this pK_a value. I₂ and I₂' are photoreversed with reversal times of 55 µs and several hundred microseconds, respectively. The corresponding signal amplitudes are pH dependent with a pK_a of 6.1. Photoreversal from I₂' dominates above the pK_a. The transient accumulation of I₂', the active state of photoactive yellow protein, is thus controlled by the proton concentration. The rate constant k₃ for the recovery to the initial dark state also has a pK_a of 6.3. This equality of the equilibrium and kinetic pK_a values is not accidental and suggests that k₃ is proportional to [I₂'].

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
Photoactive yellow protein (PYP) is significant as the structural prototype for a large and diverse superfamily of signaling proteins that share a common structural motif termed the PAS domain (1,2). PAS domains are found in all kingdoms of life, generally as the sensory component of multi-domain proteins. The structure of PYP has a central six-stranded antiparallel ß-sheet, with four structural features: an N-terminal cap, a PAS core with the first three ß-strands of the central ß-sheet, a helical connector, and a so-called ß-scaffold consisting of the last three ß-strands (3–7). PYP from the halophilic purple phototrophic bacterium Halorhodospira halophila is a small 14-kDa soluble cytoplasmic protein. The physiological function of H. halophila PYP was reported to be in negative phototaxis, resulting in movement away from blue/UV light (8). Its chromophore is p-hydroxycinnamic acid covalently bound via a thioester linkage to cysteine-69. In the dark, the phenol group of the chromophore is deprotonated, and the C₇=C₈ bond is trans. The ionization of the chromophore and its hydrogen bonding to E-46 and Y-42 are mainly responsible for the observed spectral tuning in the dark state (_max 446 nm). Light-induced trans–cis isomerization around the C₇=C₈ bond is rapid (<3 ps) and is followed by a sequence of slower relaxations in the dark that ultimately lead to recovery of the dark state in less than 1 s. This photocycle has already been studied in considerable detail (9–11). The first long-lived intermediate, after the two very short-lived intermediates I₀ and is I₁. I₁ forms in 3 ns and has a red-shifted absorption spectrum (_max 460 nm). In several hundred microseconds it decays to I₂. I₂ has a protonated chromophore and a blue-shifted absorption spectrum that is commonly believed to have its _max value at 355 nm. In I₂ the chromophore phenol is partially exposed to the aqueous medium and hydrogen-bonded to the side chain of R-52 (6). The protonation of the chromophore occurs either intramolecularly from E-46 (12) or from the external medium (13). In several milliseconds I₂ is transformed into I₂'. I₂' is believed to be the signaling state. It also has a protonated chromophore with a _max value similar to that of I₂. The I₂ to I₂' transition is associated with a major global structural change, which has been documented by NMR (14,15), CD (16), small-angle x-ray scattering (17), and FTIR (12,18). Formation of I₂' is associated with exposure of a hydrophobic surface patch (19), presumably the recognition and binding site for a response regulator. Experiments with hydrophobic dyes showed that these bind transiently to I₂' but not to I₂ (13). PYP shares a number of features, such as spectral tuning, photoisomerization, transient chromophore (de)protonation, and photoreversal, with other photoreceptors such as rhodopsin and phytochrome. This together with the availability of high-resolution structures of intermediates (3–7) makes PYP an attractive model system for signal transduction (1).

For an understanding of the mechanism of PYP, a detailed characterization of its photocycle is essential. Such studies have been carried out by both electronic (9–11,20) and vibrational (21,22) spectroscopy. The photocycle kinetics are pH (23–25) and salt (26,27) dependent. Whereas early models assumed a unidirectional sequential mechanism (9,20), it is becoming increasingly clear that back reactions and equilibria between intermediates play essential roles (11,24,25,28). These equilibria are also pH (24–26,28–30) and salt dependent (26). An example of the pH- and salt-dependent equilibrium between I₁ and I₂/I₂' was recently described for the mutant Y98Q (26). At alkaline pH, the I₁'and I₂' intermediates are in equilibrium with a pK_a of 9.9 (25,28). Most photocycle intermediates can be photoreversed when excited with light of the appropriate wavelength at the right time (31–33). Using this double-flash method, we showed that the I₂ and I₂' intermediates are in equilibrium (33). I₂ partly decays to I₂' in 2 ms and then remains in equilibrium with I₂' until the end of the cycle (33). Our laboratory showed from time-resolved fluorescence and photostationary absorption measurements that this I₂/I₂' equilibrium is pH dependent with a pK_a of 6.3 (34). Interestingly the rate constant for the kinetics of the ground state recovery has a bell-shaped pH dependence with a lower pK_a of 6.4 (23), i.e., identical to the value for the I₂/I₂' equilibrium (34). The nature of the group(s) responsible for this pK_a is still not entirely certain, but it is commonly ascribed to E-46 (35,36).

The I₂ and I₂' intermediates were originally characterized by time-resolved FTIR spectroscopy (12,18), but they can also be distinguished by transient absorption spectroscopy in the UV/visible. Indeed, our laboratory recently obtained _max values of 372 and 352 nm for I₂ and I₂', respectively, from measurements of the pH dependence of the photostationary absorbance in the presence of background illumination (34). Here we characterize the equilibrium and spectra of the I₂ and I₂' intermediates by kinetic methods using transient absorption spectroscopy in the UV/visible and the extrapolated difference method (37). These methods allow a direct determination of the time-dependent intermediate populations and their pH dependencies. We find that the transient accumulation of the signaling state (I₂') increases with pH with a pK_a of 6.4.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
Protein production and purification
H. halophila holo-PYP was produced by coexpression with the biosynthetic enzymes TAL and pCL and subsequently purified from Escherichia coli BL 21 (DE3) as described (38).

Transient absorption spectroscopy
Time-resolved absorption spectroscopy with single- and double-flash excitation was performed as described (13,24,33,39). To resolve the photoreversal kinetics, the data acquisition was triggered on the second flash (33,39). SVD methods were used as described (13,24,40–42).

Data analysis
Spectra and time courses of intermediates were obtained using the extrapolated difference method as described (25,37). The transient absorbance change A (, t) is given by a sum of contributions from the spectral intermediates i, whose relative concentration is given by n_i(t). In matrix notation:

(1)

where the columns of A are the spectra of intermediates and each of the identical columns of A_p is the spectrum of the dark state P. Assuming first-order kinetics, A can be represented by a sum of exponentials. The wavelength-dependent coefficients of the exponentials are the amplitude spectra B_i(). The B_i() are given in matrix notation by:

(2)

where C_ij is the weight of the jth exponential in the time dependence of n_i(t).

The amplitude spectra are ordered from low to high apparent time constants in the matrix B. New matrices and are formed from the columns of B and C by adding up columns (25,37). The columns of represent the extrapolated absorption difference spectra; the columns of contain the relative contributions of the intermediates in these difference spectra. In the case of PYP at acid and neutral pH, only three intermediates contribute, as we will see, in the time range investigated: I₁, I₂, and I₂'. Their relative contributions to the ith extrapolated absorption difference spectrum will be called (x_i, y_i, z_i) in the order I₁, I₂, I₂'.

The intermediate spectrum (A)_i can then be calculated from column i of matrix

(3)

To calculate the elements of the following two constraints are introduced:

1. The sum of the relative intermediate concentrations of I₁, I₂, and I₂' in the extrapolated difference spectra is constant at all times before ground state recovery and equals the fraction of molecules cycling, . This means that the sum of x, y, and z for each extrapolated difference spectrum equals or that the sum of the matrix elements of each column of equals , and that of is 1/. This is simply the conservation law for the number of cycling molecules in a sequential cycle.
   
2. The absorption of I₂' is identical to zero for wavelengths larger than or equal to 410 nm. I₂' has its absorption maximum for the longest wavelength S₀-S₁ transition around 355 nm. Because the spectra of all intermediates have similar bandwidths, we can estimate from the spectrum of P, which does not absorb beyond 500 nm, that I₂' does not absorb beyond 410 nm.
   

Matrix calculations were performed with Matlab version R12.1. Fits with sum of exponentials were carried out with Microcal Origin 7.5.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
Spectra and time courses of intermediates at pH 7: extrapolated difference method
Transient absorbance changes were measured at 19 wavelengths, ranging from 330 to 510 nm, in the time domain from 50 ns to 5 s. Fig. 1 A shows data at pH 7. For clarity, time traces at only eight wavelengths are shown. The complete data set was subjected to SVD analysis. The first six singular values were 11.2, 1.8, 0.14, 0.05, 0.03, and 0.02. We consider the first three to be significant, suggesting the presence of only three spectrally distinguishable intermediates. The additional components (s₄, s₅, and s₆) show very noisy time traces and were therefore neglected. The three weighted time traces from SVD were fitted simultaneously starting at 10 µs with a sum of three exponentials with time constants ₁ = 270 µs, ₂ = 2.0 ms, and ₃ = 260 ms. These times are marked by vertical dashed lines in Fig. 1 A. The dotted lines in Fig. 1 A, which can barely be distinguished from the data, represent these fit curves for the individual time traces. From the fit to the SVD time traces and the corresponding basis spectra, the amplitude spectra B_j() were calculated as described (42). These are presented in Fig. 1 B. The amplitude spectra provide considerable insight into the spectra of the intermediates. B₁() clearly describes the transition from I₁ (_max 460 nm) to I₂ with a _max value above 360 nm. B₂() is apparently a transition from an equilibrium of I₁ and I₂ to I₂' with I₂' blue-shifted with respect to I₂. B₃() represents the ground state recovery and suggests that I₂' has its _max value near 350 nm. Comparison of the negative minimum of B₁ with the positive maximum of B₃ suggests that the transition from I₂ to I₂' is associated with a blue shift of the order of 20 nm. So qualitatively the conclusion that there is a blue shift between the early UV intermediate I₂ and the later UV intermediate I₂' is already apparent from the data alone (amplitude spectra) without any model-dependent assumptions. The following quantitative analysis, which does use two plausible constraints, strengthens this conclusion.

View larger version (16K): [in this window] [in a new window]   FIGURE 1  (A) Transient absorption changes after excitation at 430 nm at 19 wavelengths varying from 330 to 510 nm. For clarity, only the traces at the indicated wavelengths are shown. The vertical dashed lines indicate the time constants for a global fit to the weighted SVD time traces with a sum of three exponentials. 1 = 270 µs is the rise time of I2, 2 = 2.0 ms is the rise time of I2', and 3 = 260 ms is the return to P. The dotted lines, only distinguishable from the data in the microsecond time range, are the fits. Conditions: pH 7, 20°C, 50 mM KCl, and 50 mM Tris. PYP concentration 35 µM. (B) Amplitude spectra Bi() calculated from the amplitudes of the exponential fits to the SVD time traces and the corresponding basis spectra of the data in A. The three amplitude spectra correspond to the following time constants: 1 = 270 µs (•), 2 = 2.0 ms (), 3 = 260 ms (). The solid curve is a scaled and inverted ground-state spectrum. (C) Extrapolated difference spectra obtained from the amplitude spectra of B as described in the text: 1 (), 2 (), and 3 ().

We now use the second constraint as described in Materials and Methods, that I₂' does not absorb beyond 410 nm, and consider Eq. 3 only in the range > 410 nm; i.e., we drop the rows for the shorter wavelengths. Then the third column of the reduced matrix A, corresponding to the spectrum of I₂', is the null vector: (A)₃ =. Because (A)₃ is zero, we can solve Eq. 3 for (^–1)₃. In this way, we determine the third column of ^–1. Under the first constraint, the sum of these matrix elements equals ^–1. In this way, we find, for the fraction of molecules cycling, = 0.371. Finally, using and A_p for the whole spectral range allows us to calculate the spectrum of I₂' from (^–1)₃ using Eq. 3. The result is shown in Fig. 2 A (solid squares). The _max value of the spectrum is at 350 ± 5 nm.

View larger version (17K): [in this window] [in a new window]   FIGURE 2  (A) Intermediate spectra I1 (), I2 (), and I2' () calculated from the extrapolated difference spectra of Fig. 1 C. The solid curve represents the spectrum of dark state P for comparison. Vertical dashed lines indicate the wavelengths of the blue (430 nm) and violet (355 nm) excitation flashes used. (B) I2 spectrum for various allowed values of y2 as described in the text. y2 = 0.12 (•), y2 = 0.15 (), y2 = 0.18 (), y2 = 0.22 (), y2 = 0.27 (), and y2 = 0.37 (). (C) Time courses of the relative concentrations of I1 (solid line), I2 (dashed line), and I2' (dotted line) calculated according to Eq. 1. The time course of the sum of the relative concentrations of I1, I2, and I2' is indicated by dashed-dot-dashed line. The vertical dashed lines indicate the time constants from the global SVD fit of Fig. 1 A.

To calculate the spectrum of the third spectral species, I₂, we proceeded as follows. From Fig. 1 B, we note that B₁ reflects a transition between two intermediates with _max values of 460 nm (decay of I₁) and 370 nm (rise of I₂). Because there is apparently no contribution from the more blue-shifted species I₂' (_max 350 nm) in B₁, which is well known to be formed from I₂ in the next transition (12), I₂' does not contribute to ₂ either. Dye-binding experiments also showed that the formation of the signaling state I₂' is delayed with respect to the formation of I₂ (13). Therefore, we conclude that I₂' is not involved in the first transition, and thus, z₂ = 0. The elements x₂, y₂, z₂ of the second column of can now be expressed in terms of x₂, y₂, and with the help of ^– = I. Using the conservation constraint, x₂ + y₂ = , we finally obtain for the elements of the second column of ^–1, . So we have now determined all elements of the second column of ^–1, the only free parameter remaining is y₂. Because x_i, y_i, and z_i can assume only positive values and x₂ + y₂ = , y₂ is restricted to values between 0 and . The spectrum of I₂ can now be calculated from (^–1)₂, , and A_p using Eq. 3. The results are shown in Fig. 2 B for six values of y₂ from 0.12 to 0.37 (). Because the extinction coefficient has to be positive, physically meaningful absorption spectra are obtained only for y₂ 0.22. Of the spectra remaining in Fig. 2 B, we pick the one associated with y₂ = 0.22 because it has the smallest spectral bandwidth. For y₂ considerably larger than 0.22, the spectral bandwidth becomes much larger than for P and I₁, and a secondary absorption maximum develops near 460 nm. This contradicts the original assumption that the UV transitions of I₂ and I₂' are the longest-wavelength transitions of these intermediates, which precludes transitions at higher wavelengths. The spectrum of I₂ for y₂ = 0.22 is redrawn in Fig. 2 A (open circles). Its _max value is at 370 ± 5 nm. We note that the value of _max is independent of the choice of y₂.

With the spectra of I₁, I₂, and I₂', the time courses of the intermediates were calculated from the experimental A(,t) data by matrix inversion of Eq. 1. The time dependence of the relative concentrations of the I₁, I₂, and I₂' intermediates at pH 7 are shown in Fig. 2 C. I₁ partially decays to I₂ in 270 µs. I₁ and I₂ then further decay around 2 ms to an I₁/I₂/I₂' equilibrium. This equilibrium finally decays to P in 260 ms. Also shown is the sum of the relative concentrations of these intermediates (dash-dot line). To a good approximation, this sum is constant over the entire time range before the decay to P, validating the data analysis. Its value is very close to = 0.371, the fraction cycling, showing the internal consistency of the analysis.

pH dependence of photocycle kinetics
To learn more about the nature of the transition between the acid and the neutral pH regimes, the photocycle kinetics were measured at the following 15 pH values: 4.6, 4.8, 5.1, 5.4, 5.7, 6.0, 6.3, 6.6, 6.75, 6.9, 7.35, 7.7, 7.9, 8.1, 8.4. With excitation at 430 nm, time traces were collected at the seven wavelengths 340, 370, 390, 410, 450, 490, and 500 nm over the time range from 50 ns to 50 s. Results for selected wavelengths are shown in Fig. 3. Note that the panels of Fig. 3 have very different vertical scales and correspondingly different signal/noise ratios. The smallest pH-induced absorbance changes are at 500 nm. The initial absorbance change is almost pH independent at every wavelength, suggesting that the amount of I₁ formed is independent of pH in this range. At each pH, the absorbance changes at all wavelengths could be fitted simultaneously with a sum of three exponentials. The first time constant was virtually constant in this pH range, varying between 200 and 350 µs. The second time constant varied between 1.3 (pH 8.4) and 10.6 (pH 5.1) ms. As we saw above, the first transition results the decay of I₁ to I₂, and the second transition is the decay of I₁/I₂ to the I₁/I₂/I₂' equilibrium. The data of Fig. 3 show that the third time constant, the return of the I₁/I₂/I₂' equilibrium to P, is strongly pH dependent, slowing down with decreasing pH.

View larger version (41K): [in this window] [in a new window]   FIGURE 3  pH dependence of the transient absorbance changes after excitation at 430 nm at various wavelengths: (A) 340 nm (characteristic for I2'), (B) 370 nm (characteristic for I2), (C) 390 nm (characteristic for I2), (D) 410 nm (characteristic for I1'), (E) 450 nm (characteristic for P), and (F) 500 nm (characteristic for I1). The color codes for the pH values in each panel are the following: black, pH 8.4; red, pH 6.9; green, pH 6.6; blue, pH 6.0; light blue, pH 5.7; pink, pH 5.4; dark blue, pH 4.8. Conditions: 50 mM MES, 50 mM KCl, 20°C. PYP concentration 53 µM.

To obtain the time courses of the intermediate populations, it was assumed that the spectra of I₁, I₂, and I₂' of Fig. 2 A are pH independent and that no other intermediates contribute in the pH range from 4.6 to 8.4. Equation 1 was then used to calculate the time traces n_i(t) for each intermediate at each pH value from the absorbance changes A(, t) and the spectra A_i() by matrix inversion. The time dependencies of the populations of I₁, I₂, and I₂' at 7 of the 15 pH values are shown in panels A, B, and C of Fig. 4. They confirm what was suggested by the data of Fig. 3: I₁ decays partially to I₂; I₂ then partially decays to I₂'; beyond 10 ms I₁, I₂, and I₂' are in equilibrium and return together to P. Fig. 4 D shows that the sum of the populations is approximately constant in time and equal to the fraction cycling. Whereas the population of I₁ in equilibrium with I₂ and I₂' is only slightly pH dependent (see traces in Fig. 4 A around 10 ms), the concentrations of I₂ and I₂' show a strong and opposite pH dependence. With increasing pH, the amount of I₂' increases at the expense of a corresponding decrease in the I₂ population. To quantify this pH dependence of the equilibrium populations, their concentrations at 10 ms were taken from Fig. 4, B and C (vertical dashed lines). The corresponding concentrations of I₂ and I₂' are plotted in Fig. 5 A as a function of pH. The solid curves are the results of a simultaneous fit with the Henderson-Hasselbalch equation. The fit parameters were pK_a 6.4 and n 0.98. I₂ and I₂' are thus in a pH-dependent equilibrium.

View larger version (24K): [in this window] [in a new window]   FIGURE 4  Time courses of the relative concentrations of the I1 (A), I2 (B), and I2' (C) intermediates at various pH values calculated using Eq. 1 as explained in the text. (D) Time course of the sum of the populations of I1, I2, and I2'. Color code as in Fig. 3.

View larger version (13K): [in this window] [in a new window]   FIGURE 5  (A) pH dependence of the equilibrium concentrations of the I2 () and I2' (•) intermediates at 10 ms derived from Fig. 4, B and C, respectively. The solid curves are simultaneous fits of these titration curves with the Henderson-Hasselbalch equation with pKa = 6.4 and n 0.98. (B) pH dependence of the decay rate k3 of the ground-state recovery. For every pH, the decay rate k3 was derived from the simultaneous fit of the measured transient absorbance changes at 340 nm, 370 nm, 390 nm, 410 nm, 450 nm, 490 nm, and 500 nm with a sum of three exponentials. The solid curve in B is the fit of the decay rate k3 with the Henderson-Hasselbalch equation with pKa = 6.3, n 0.84.

pH dependence of photoreversal kinetics
Recently, we measured and analyzed the kinetics of photoreversal from I₂, I₂', and I₁ at pH 6 in detail (33). The time delay between the two flashes was varied from 1 µs to 3 s, and the photoreversal kinetics were measured at 26 wavelengths from 330 to 510 nm (33). The photoreversal time traces at pH 6 required two exponentials for an adequate fit with well-separated time constants of ₁ = 60 and ₂ = 400 µs. These time constants were assigned to photoreversal from I₂ (60 µs) and I₂' (400 µs), respectively, on the basis of the delay dependence of the amplitudes (33). Moreover, we found that I₂ and I₂' are in equilibrium (33). Here, our focus is on the pH dependence. The photoreversal kinetics were therefore measured at only two wavelengths (340 and 450 nm) and at the fixed delay of 20 ms. At this time delay all three intermediates I₂, I₂', and I₁ are present and in equilibrium (Figs. 2 C and 4).

The photoreversal signals at pH 5.1 and 8.1 are shown in Fig. 6 A. For clarity, only the traces at 2 of the 15 pH values are shown together with their simultaneous fits (solid lines). As at pH 6 (33), the kinetics required two exponentials over the pH range from 4.6 to 6.9. These two phases can be clearly discerned in the data at pH 5.1. From pH 7.3 on, only the slow component was required. The fast time constant was virtually pH independent, varying between 48 and 63 µs. The second time constant varied between 330 and 770 µs in the pH range investigated. Comparison of the time traces in Fig. 6 A makes it clear that there are significant differences between pH 5.1 and 8.1. The total initial photoreversal signal is somewhat larger at pH 5.1 than at 8.1, suggesting that more I₂/I₂' can be photoreversed. The amplitude data provide further insight. The pH dependence of the amplitudes A₁ and A₂ (for the corresponding time constants ₁ and ₂) are plotted in Fig. 6, B (340 nm) and C (450 nm). These figures confirm that, with increasing pH, A₁ becomes smaller, approaching zero around pH 7 at both wavelengths, whereas A₂ shows a corresponding increase. The solid curves are simultaneous fits of the amplitudes at 340 and 450 nm with the Henderson-Hasselbalch equation, with pK_a = 6.1 and n = 1.9. We assigned A₁ and A₂ to photoreversal from I₂ and I₂', respectively (33). The results of Fig. 6, B and C thus suggest that, with increasing pH, the I₂/I₂' equilibrium shifts from I₂ at pH 4.6 to I₂' at pH 8.4. These results, from the pH dependence of the photoreversal amplitudes, thus support our observations from the photocycle, where a pK_a of 6.4 was obtained. We note that in the photocycle experiments, n = 0.98 (I₂/I₂' equilibrium) or 0.8 (recovery rate), whereas we obtained n = 1.9 from the photoreversal kinetics. The photoreversal absorbance changes required various corrections and are, moreover, quite small. The errors are correspondingly large. The photocycle data points, on the other hand, display much less scattering, in particular for k₃, and did not require complex corrections. We believe, therefore, that the number of protons involved is one in both experiments.

View larger version (17K): [in this window] [in a new window]   FIGURE 6  (A) Photoreversal signals, at 340 nm (positive) and 450 nm (negative) at pH 5.1 (black) and 8.1 (red), calculated as described in the text. For clarity, the data obtained at 13 other pH values ranging from 4.6 to 8.4 are not shown. The solid curves represent a simultaneous exponential fit to the 340-nm and 450-nm traces. Conditions: 20°C, 50 mM KCl and 50 mM MES, PYP concentration 53 µM. The delay between first (blue, 430 nm) and second (violet, 355 nm) flashes is 20 ms. (B and C) pH dependence of the photoreversal amplitudes A1 and A2 at 340 nm (B) and 450 nm (C). A1 (•) and A2 () are the amplitudes of the fast (48–60 µs) and slow (300–770 µs) components, respectively, obtained from the simultaneous fit of the 340-nm and 450-nm traces of panel A. The solid curves of panels B and C represent a common fit to the pH dependence of all four amplitudes with the Henderson-Hasselbalch equation, with a pKa of 6.1 (dashed vertical lines) and a Hill coefficient of 1.9.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

View larger version (13K): [in this window] [in a new window]   FIGURE 7  Proposed model for the kinetics of the photocycle and photoreversal of PYP in the pH range from 4 to 8. The and I2'cis intermediates are in a pH-dependent equilibrium and photoreverse to Ptrans with exponential time constants of 57 and 380 µs. The pKa of the equilibrium is 6.4. Note the equilibria among I1, I2, and I2'. These three intermediates decay together to P. For clarity the short-lived intermediates I0 and between P* and I1 are not shown. The proton arrows indicate the transient uptake and release of protons as detected by a pH-sensitive dye in solution (13).

Using the spectra of I₂ and I₂' from Fig. 2 A and assuming that they are pH independent in the pH range from 4.6 to 8.4, we obtained the time dependence of the concentrations of I₁, I₂, and I₂' (Fig. 4). From 5 ms onward, I₁, I₂, and I₂' are in equilibrium and decay together to the initial dark state P. The equilibrium intermediate populations of I₂ and I₂' are pH dependent with a pK_a of 6.4. Below the pK_a, I₂ is the major species; above the pK_a the opposite holds. We note that these transitions are not complete (Fig. 5 A).

The sum of the populations of the I₁, I₂, and I₂' intermediates in Fig. 4 D is not as constant in time for every pH as it should be and as it is at pH 7 (Fig. 2 C). One cause could be that the intermediate spectra are not exactly pH independent over the whole pH range, so that the spectra derived at pH 7 are not quite correct for every pH. So we derived a second set of intermediate spectra from a joint SVD analysis over the whole pH range as in Joshi et al. (25). These are the best pH-averaged intermediate spectra. The differences with the set at pH 7 (Fig. 2 A) are small. A third set was obtained from the scaled subtraction method (25). All three sets of intermediate spectra lead to the same pK_a value. The results differed only somewhat in terms of the end values of the titrations at high pH. The sum of the populations remained slightly time-dependent, however, and other inadequacies developed, such as slightly negative populations. Because the spectra of I₂ and I₂' are quite similar, the analysis is sensitive to the small differences between them. We believe this is the likely source of the error in Fig. 4 D.

Measurements of the pH dependence of the photoreversal kinetics of I₂ and I₂' at a delay of 20 ms (Fig. 6) provide further support for this pK_a. The kinetics are characterized by a fast (50–60 µs) and a slow (300–800 µs) component, which result from photoreversal from I₂ and I₂', respectively (33). The corresponding amplitudes are pH dependent (Fig. 6, B and C) with a pK_a of 6.1, in good agreement with the value of 6.4 obtained from the photocycle kinetics. The amplitude for I₂ goes to zero beyond pH 7.3. The photoreversal data thus suggest that no I₂ remains at alkaline pH, whereas the photocycle data indicate that some I₂ remains because [I₂]/([I₂] + [I₂']) is 0.25 at pH 8.6 (Fig. 5 A). We note that the end value of the I₂ population at high pH in Fig. 5 A is quite sensitive to the exact choice of the I₁ spectrum. In a recent investigation on the intermediates and spectra at alkaline pH (25), we found that only three intermediates, I₁, I₁', and I₂', are present above pH 8, i.e., no I₂.

The pH dependence of the absorption spectrum of a photostationary mixture of P, I₂, and I₂' produced by background illumination was recently analyzed by SVD (34). It was shown that I₂ and I₂' are in a pH-dependent equilibrium with a pK_a of 6.3 and that the spectrum of the high-pH species I₂' is blue-shifted with respect to that of the low-pH species I₂ (34). The I₂ and I₂' intermediates also differ with regard to the fluorescence lifetime of the single tryptophan of PYP, W-119 (34). In I₂, the lifetime is long (0.82 ns). In I₂', the lifetime is much shorter (0.04 ns). Using background illumination, these authors showed that the fluorescence decay in the photostationary state is pH dependent with I₂ dominating at low pH and I₂' at high pH (34). The pK_a was 6.3. Combining the pH dependence of the fluorescence amplitudes with that of the photostationary absorption, absorption spectra were calculated for the I₂ and I₂' species. These had _max values of 372 and 352 nm for I₂ and I₂', respectively (34), in good agreement with the results reported in this work.

The results on the pH dependence of the I₂/I₂' equilibrium (34) were recently confirmed (29). Analysis of the photostationary absorption spectra by a scaled subtraction procedure yielded a pK_a of 6.4 and _max values of 367 and 356 nm for I₂ and I₂', respectively (29). These authors showed, moreover, from CD and small-angle x-ray scattering experiments that the global structural transition occurs between these two intermediates with a pK_a of 6.4. Together with the kinetics results from time-resolved absorption spectroscopy presented here, these complementary methods lead to a comprehensive picture of the I₂-to-I₂' equilibrium.

As is well known (23) and confirmed here (Fig. 5 C), the rate k₃, for the ground state recovery, is also pH dependent with a pK_a of 6.3, i.e., within experimental error equal to that for the I₂/I₂' equilibrium (6.4 and 6.1, from single flash photolysis and photoreversal measurements, respectively). Thus, k₃ seems to be proportional to the I₂' population (compare Fig. 5, A and B). Such a proportionality is expected under the following conditions: 1) the equilibration rates among I₁, I₂, and I₂' are rapid compared to the microscopic rates of return from each intermediate to the ground state; 2) the latter are pH independent; and 3) the rate from I₂' to P is much larger than from the other intermediates. A similar model was recently proposed to explain the pH dependence of the rate of ground state recovery at alkaline pH (25). At alkaline pH, the I₁, I₁', and I₂' intermediates are in equilibrium. The pK_a of the I₁'-to-I₂' equilibrium is 9.9, and the ground state recovery rate constant k₃ has a pK_a of 9.7 and is proportional to the I₂' population. There is thus a striking similarity between the behavior at high and low pH. For both branches of the bell-shaped pH dependence of k₃, it seems that k₃ is proportional to [I₂']. This proportionality is not exact however, because k₃ approaches zero at low pH, whereas [I₂'] approaches a constant value unequal to zero. A more detailed model thus seems to be required. Nevertheless, this symmetry between low and high pH behavior is worth pointing out, and the underlying model provides a lowest-order explanation.

An important question concerns the group responsible for the pK_a of 6.4. The similar pK_a for the recovery rate k₃ is commonly attributed to the carboxyl group of E-46 (35,36). We now need to discuss this pK_a in the context of the underlying I₂/I₂' equilibrium. What is the mechanism whereby a change in protonation of E-46 shifts the equilibrium from I₂ to I₂'. In I₂ the chromophore is already protonated and has moved away from E-46 toward the surface (6). If E-46 is the internal proton donor for the chromophore, its carboxyl group is presumably already deprotonated in I₂, in accordance with some observations from time-resolved FTIR (12). In I₂' the chromophore remains protonated, but the protein structure is changed in a major way. It is unclear how the deprotonated E-46 could affect the I₂/I₂' conformational equilibrium. If, however, E-46 is not the internal proton donor, and the chromophore is protonated from the external medium, as suggested (13), E-46 could remain protonated in I₂ and be deprotonated in I₂'. In other time-resolved FTIR measurements (18), a positive band was observed at 1759 cm^–1 with a risetime of 113 µs (formation of I₂) and assigned to an environmental shift of the protonated E-46. Brudler et al. (18) were not aware of the I₂/I₂' equilibrium and concluded from the fact that the amplitude of the positive band was significantly smaller than that of the negative band due to the initial dark state, that only a fraction of the molecules cycling had a protonated E-46 in I₂. Their experiments were, however, performed in buffer at pH 7. At this pH, the I₂/I₂' equilibrium is far on the side of I₂' (see Fig. 5 A), so that only a minority of molecules would have been in the I₂ state. It is thus consistent with the time-resolved data of (18) to conclude that in I₂ the carboxyl group of E-46 is protonated. Recent photostationary FTIR measurements also indicate that E-46 is at least partially protonated in I₂ (29). A role of E-46 in controlling the I₂/I₂' equilibrium is thus plausible. In the absence of the carboxyl group, in the mutant E46Q, the conformational change in I₂' is much smaller or absent (12), and the absorption maximum at pH 7 is at 368 nm (45), i.e., I₂-like. These results suggest that in the absence of E-46 the I₂/I₂' equilibrium is predominantly or entirely on the side of I₂ and further support the idea that E-46 is responsible for the wild-type pK_a of 6.4. We note that this reinterpretation of the FTIR results is consistent with a mechanism of chromophore protonation from the external medium (13).

Another residue that might be involved is H-108. The pH dependence of the steady-state proton uptake was investigated (46), and from the observed pH dependence, a pK_a of 6.6 was obtained, which was attributed to histidine-108. This residue is located on the central ß-scaffold and may be involved in the interaction between the ß-scaffold and the N-terminal domain. It was recently postulated, on the basis of the observation that the I₂-to-I₂' transition is blocked at low salt concentrations, that the loss of this interaction is a prerequisite for the formation of I₂' (26).

Recently two forms of I₁ could be distinguished on the basis of their resonance Raman spectra, which are in a pH-dependent equilibrium with a pK_a of 6.2 (30). The low-pH form lacks the hydrogen bond with E-46, whereas the high-pH form () has both hydrogen bonds. It is possible that the pH dependence observed here for the I₂/I₂' equilibrium results from the pH dependence of the preceding equilibrium.

In conclusion, we have shown from measurements of the time-dependent intermediate populations that the equilibrium between I₂ and I₂' and the formation of the signaling state I₂' are pH dependent with a pK_a of 6.4. The intracellular pH may thus regulate the amount of active receptor. The value of this pK_a provides an explanation for the similar well-known pK_a of the rate constant for the ground-state recovery. We find, moreover, that I₂' is blue-shifted with respect to I₂ by 20 nm, suggesting a different chromophore environment for the exposed chromophore in the signaling state.

	ACKNOWLEDGEMENTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
This work was supported by the National Institutes of Health (grant GM 66146 to M.A.C.) and the Deutsche Forschungsgemeinschaft (grant GK 788 TP A9 to M.P.H.).

Submitted on April 10, 2006; accepted for publication June 15, 2006.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
1. Cusanovich, M. A., and T. E. Meyer. 2003. Photoactive yellow protein: a prototypic PAS domain sensory protein and development of a common signaling mechanism. Biochemistry. 42:4759–4770.[CrossRef][Medline]

2. Taylor, B. L., and I. B. Zhulin. 1999. PAS domains: internal sensors of oxygen, redox potential, and light. Microbiol. Mol. Biol. Rev. 63:479–506.[Abstract/Free Full Text]

3. Borgstahl, G. E. O., D. R. Williams, and E. D. Getzoff. 1995. 1.4 Å structure of photoactive yellow protein, a cytosolic photoreceptor: unusual fold, active site and chromophore. Biochemistry. 34:6278–6287.[CrossRef][Medline]

4. Rajagopal, S., S. Anderson, V. Srajer, M. Schmidt, R. Pahl, and K. Moffat. 2005. A structural pathway for signalling in the E46Q mutant of photoactive yellow protein. Structure. 13:55–63.[Medline]

5. Ihee, H., S. Rajagopal, V. rajer, R. Pahl, S. Anderson, M. Schmidt, F. Schotte, P. A. Anfinrud, M. Wulff, and K. Moffat. 2005. Visualizing reaction pathways in photoactive yellow protein from nanoseconds to seconds. Proc. Natl. Acad. Sci. USA. 102:7145–7150.[Abstract/Free Full Text]

6. Genick, U. K., G. E. Borgstahl, K. Ng, Z. Ren, C. Pradervand, P. M. Burke, V. Srajer, T. Y. Teng, W. Schildkamp, D. E. McRee, K. Moffat, and E. D. Getzoff. 1997. Structure of a protein photocycle intermediate by millisecond time-resolved crystallography. Science. 275:1471–1475.[Abstract/Free Full Text]

7. Dux, P., G. Rubinstenn, G. W. Vuister, R. Boelens, F. A. A. Mulder, K. Hard, W. D. Hoff, A. R. Kroon, W. Crielaard, K. J. Hellingwerf, and R. Kaptein. 1998. Solution structure and backbone dynamics of the photoactive yellow protein. Biochemistry. 37:12689–12699.[CrossRef][Medline]

8. Sprenger, W. W., W. D. Hoff, J. P. Armitage, and K. J. Hellingwerf. 1993. The eubacterium Ectothiorhodospira halophila is negatively phototactic, with a wavelength dependence that fits the absorption spectrum of photoactive yellow protein. J. Bacteriol. 175:3096–3104.[Abstract/Free Full Text]

9. Hoff, W. D., I. H. M. van Stokkum, H.-J. van Ramesdonk, M. E. van Brederode, A. M. Brouwer, J. C. Fitch, T. E. Meyer, R. van Grondelle, and K. J. Hellingwerf. 1994. Measurement and global analysis of the absorbance changes in the photocycle of the photoactive yellow protein from Ectothiorhodospira halophila. Biophys. J. 67:1691–1705.[Abstract/Free Full Text]

10. Ujj, L., S. Devanathan, T. E. Meyer, M. A. Cusanovich, G. Tollin, and G. H. Atkinson. 1998. New photocycle intermediate in the photoactive yellow protein from Ectorhodospira halophila: picosecond transient absorption spectroscopy. Biophys. J. 75:406–412.[Abstract/Free Full Text]

11. Hendriks, J., I. H. M. van Stokkum, and K. J. Hellingwerf. 2003. Deuterium isotope effects in the photocycle transitions of the photoactive yellow protein. Biophys. J. 84:1180–1191.[Abstract/Free Full Text]

12. Xie, A., L. Kelemen, J. Hendriks, B. J. White, K. J. Hellingwerf, and W. D. Hoff. 2001. Formation of a new buried charge drives a large-amplitude protein quake in photoreceptor activation. Biochemistry. 40:1510–1517.[CrossRef][Medline]

13. Borucki, B., S. Devanathan, H. Otto, M. A. Cusanovich, G. Tollin, and M. P. Heyn. 2002. Kinetics of proton uptake and dye binding by photoactive yellow protein in wildtype and in the E46Q and E46A mutants. Biochemistry. 41:10026–10037.[CrossRef][Medline]

14. Rubinstenn, G., G. W. Vuister, F. A. A. Mulder, P. E. Dux, R. Boelens, K. J. Hellingwerf, and R. Kaptein. 1998. Structural and dynamic changes of photoactive yellow protein during its photocycle in solution. Nat. Struct. Biol. 5:568–570.[CrossRef][Medline]

15. Bernhard, C., K. Houben, N. M. Derix, D. Marks, M. A. van der Horst, K. J. Hellingwerf, R. Boelens, R. Kaptein, and N. A. J. van Nuland. 2005. The solution structure of a transient photoreceptor intermediate: 25 photoactive yellow protein. Structure. 13:953–962.[Medline]

16. Lee, B.-C., P. A. Croonquist, T. R. Sosnick, and W. D. Hoff. 2001. PAS domain receptor photoactive yellow protein is converted to a molten globule state upon activation. J. Biol. Chem. 276:20821–20823.[Abstract/Free Full Text]

17. Imamoto, Y., H. Kamikubo, M. Harigai, N. Shimizu, and M. Kataoka. 2002. Light-induced global conformational change of photoactive yellow protein in solution. Biochemistry. 41:13595–13601.[CrossRef][Medline]

18. Brudler, R., R. Rammelsberg, T. T. Woo, E. D. Getzoff, and K. Gerwert. 2001. Structure of the I₁ early intermediate of photoactive yellow protein by FTIR spectroscopy. Nat. Struct. Biol. 8:265–270.[CrossRef][Medline]

19. Meyer, T. E., G. Tollin, J. H. Hazzard, and M. A. Cusanovich. 1989. Photoactive yellow protein from the purple phototropic bacterium, Ectothiorhodospira halophila: quantum yield of photobleaching and effects of temperature, alcohols, glycerol, and sucrose on kinetics of photobleaching and recovery. Biophys. J. 56:559–564.[Abstract/Free Full Text]

20. Meyer, T. E., E. Yakali, M. A. Cusanovich, and G. Tollin. 1987. Properties of a water-soluble, yellow protein isolated from a halophilic phototrophic bacterium that has photochemical activity analogous to sensory rhodopsin. Biochemistry. 26:418–423.[CrossRef][Medline]

21. Unno, M., M. Kumauchi, J. Sasaki, F. Tokunaga, and S. Yamauchi. 2000. Evidence for a protonated and cis configuration chromophore in the photobleached intermediate of photoactive yellow protein. J. Am. Chem. Soc. 122:4233–4234.[CrossRef]

22. Pan, D., A. Philip, W. D. Hoff, and R. A. Mathies. 2004. Time-resolved resonance Raman structural studies of the pB' intermediate in the photocycle of photoactive yellow protein. Biophys. J. 86:2374–2382.[Abstract/Free Full Text]

23. Genick, U. K., S. Devanathan, T. E. Meyer, I. L. Canestrelli, E. Williams, M. A. Cusanovich, G. Tollin, and E. D. Getzoff. 1997. Active site mutants implicate key residues for control of color and light cycle kinetics of photoactive yellow protein. Biochemistry. 36:8–14.[CrossRef][Medline]

24. Borucki, B., H. Otto, C. P. Joshi, C. Gasperi, M. A. Cusanovich, S. Devanathan, G. Tollin, and M. P. Heyn. 2003. pH dependence of the photocycle kinetics of the E46Q mutant of photoactive yellow protein: protonation equilibrium between the I₁ and I₂ intermediates, chromophore deprotonation by hydroxyl uptake, and protonation relaxation in the dark state. Biochemistry. 42:8780–8790.[CrossRef][Medline]

25. Joshi, C. P., B. Borucki, H. Otto, T. E. Meyer, M. A. Cusanovich, and M. P. Heyn. 2006. Photocycle and photoreversal of photoactive yellow protein at alkaline pH: kinetics, intermediates and equilibria. Biochemistry. 45:7057–7068.[CrossRef][Medline]

26. Borucki, B., J. A. Kyndt, C. P. Joshi, H. Otto, T. E. Meyer, M. A. Cusanovich, and M. P. Heyn. 2005. Effect of salt and pH on the activation of photoactive yellow protein and gateway mutants Y98Q and Y98F. Biochemistry. 44:13650–13663.[CrossRef]