Low-temperature Spectroscopy of Met100Ala Mutant of Photoactive Yellow Protein^†‡

Introduction

Photoactive yellow protein (PYP), first isolated from a purple photosynthetic bacterium Halorhodospira halophila (1), is an excellent target for investigating the light-capturing mechanism of photoreceptor proteins at the atomic level (2,3). PYP is a putative photosensor protein for the negative phototaxis of the bacteria (4). PYP is a water-soluble protein composed of 125 amino acid residues, and its tertiary structure at high resolution is available (5–7). It is a typical PAS domain protein, composed of a six-stranded β-sheet and several short α-helices. To absorb visible light, the trans-p-coumaroyl chromophore binds to Cys69 by a thioester bond (8–10). A photon absorbed by the chromophore induces the trans-to-cis isomerization of the vinyl bond, which initiates a photocycle similar to that of bacterial retinal proteins. The PYP photocycle comprises several intermediates, as demonstrated by previous low-temperature spectroscopy studies (11,12). Blue light or yellow light converts PYP_dark (λ_max is 449 nm in glycerol samples) into PYP_B (489 nm) or PYP_H (442 nm), respectively, at 80 K. Both of them are thermally converted to PYP_L (456 nm), which is likely to be the same species as I₁ (=pR) formed on the microsecond time scale at room temperature (13,14). At room temperature, PYP_L is converted to a near-UV intermediate (PYP_M, I₂ or pB) (356 nm) as a consequence of a protein conformational change (15–17), but in glycerol samples at low temperature, PYP_L is directly converted to the dark (ground) state without the formation of PYP_M.

Using site-directed mutants, the roles of various amino acid residues of PYP in the structure and photoreaction have been studied so far. One of the key amino acid residues governing the PYP photocycle is Met100 (18–20), which is located in the loop region connecting the fourth and fifth β-strands (Met100 loop). This region is curled toward the chromophore, and the distance between the chromophore and Met100 side chain is less than 4.5 Å in the dark state (PDB ID 1nwz) (Fig. 1). Mutation of Met100 substantially slows the photocycle (18–20). M100A shows much slower decay of the I₂ state (M100A_M), whereas the decay of the I₁ state (M100A_L) is faster than that of the wild type (WT) (18). As the back-isomerization of the cis-chromophore in M100A_M by light rapidly generates M100A_Dark, Met100 is thought to catalyze the cis-to-trans isomerization of the chromophore during the photocycle (18).

While M100A_M has been extensively characterized, the primary photoreaction of M100A has not been studied in detail. Prof. Fumio Tokunaga and his coworkers have demonstrated that low temperature spectroscopy is a powerful tool to analyze the photoreaction and subsequent thermal reactions of photoreceptor proteins. Using liquid nitrogen as a coolant, one can observe the ultrafast reactions that take place on the picosecond time scale at room temperature using a conventional UV–visible recording spectrophotometer. Heating the sample mimics the elapse of time. Thus, the whole reaction can be easily and consistently covered using a single experimental setup. Here, the photocycle of M100A was studied by low-temperature spectroscopy to clarify the role of Met100 in the early stage of the PYP photocycle.

Materials and methods

Apoproteins of WT PYP and M100A were prepared according to the standard protocol (21). They were reconstituted into holoproteins and purified by DEAE-Sepharose column chromatography as reported previously (10,21). WT and M100A were suspended in 10 mm TAPS buffer at pH 8.0 containing 200 mm NaCl. Two volumes of glycerol were added to the sample for low-temperature spectroscopy (66% glycerol sample).

The sample was put into a sample cell composed of a transparent quartz window, an opal glass window (Sigma Koki, Tokyo, Japan) and a silicon rubber spacer (2 mm in thickness, 6 mm in inner diameter) (22). It was set in an optical cryostat (Oxford Optistat DN) installed in a recording spectrophotometer (Shimadzu UV2400PC). The reference beam of the spectrophotometer was similarly scattered with another opal glass plate. The slit of the measurement beam was set at 2 nm. The sample was irradiated with blue (436 nm) or yellow (>450 nm) light passed through an optical interference filter (Edmund 43161) or a glass cutoff filter (Asahi Techno Glass Y47). A 1-kW slide projector was used as an irradiation light source.

Wild type and M100A were irradiated at 80 K and the subsequent thermal reaction of the photoproduct was induced by warming. The temperature of the sample, which was monitored by the thermocouple attached to the sample cell holder, was increased at 1 K min⁻¹ using an ITC502 temperature controller (Oxford). The absorption spectra were recorded every 10 K.

The absorption spectra of PYP and intermediates are sharpened by cooling (11). Therefore, increasing the temperature induces thermal broadening of the absorption spectra as well as thermal decay of the intermediates. In the previous low temperature spectroscopy, the sample, warmed to induce the thermal reaction, was re-cooled to the base temperature (83 or 193 K) for measurement to avoid the effect of temperature on the absorption spectrum (11). However, repetition of warming and cooling readily shifts the baseline. In the present experiment, the absorption spectrum of the dark state was measured every 10 K, and the difference spectra before and after irradiation were calculated using the dark spectrum and light spectrum measured at the same temperature to minimize the effect of temperature on the absorption spectra.

The PYP sample was irradiated at temperature t₀ (80 K) and then warmed to t₁ to measure the absorption spectrum. The absorption spectrum measured at t₁ () is composed of the spectra of intermediate(s) present at t₁ () and dark state at t₁ () as follows.

(1)

where f is the fraction of PYP undergoing the photocycle. As f was 0.1–0.3 in the present experimental condition, was the major component. In the present analysis, was subtracted from , giving the difference spectrum (Diff) as follows:

(2)

Therefore, the thermal effect of the major component is canceled, and the shape of the difference spectrum is independent of f.

Results

Photoreaction at 80 K

The absorption maximum of M100A in the visible region is located at 446 nm, and it is shifted to 448 nm in 66% glycerol buffer. A similar small redshift by glycerol at high concentration is also observed for WT (11). The absorption spectrum of M100A at 80 K was sharpened and showed a spectral shoulder at 420 nm, which was similar to the sharpening and shoulder observed for WT (Fig. 2). WT and M100A in 66% glycerol buffer were irradiated with blue light (436 nm) or yellow light (>450 nm) at 80 K and the absorption spectra were recorded (Fig. 2).

Irradiation of WT with blue light yielded a significant amount of the redshifted intermediate called PYP_B, which corresponds to I₀ at room temperature (23,24) (Fig. 2, curve 2). After irradiation for 240 s, no further spectral shift was observed upon prolonged irradiation because the photo-steady-state mixture was produced. On the other hand, yellow light caused a 7 nm blueshift in the absorption maximum, indicating the formation of a blueshifted intermediate called PYP_H (curve 3). As the extinction coefficient of PYP_H is smaller than that of PYP_dark throughout the visible region, no absorbance increase was observed in the difference spectrum between them (curve 3′ in inset). Due to the shift in the peak and spectral shoulder, the difference spectra between PYP_H and PYP_dark showed two sharp negative bands (curve 3′ in inset). The presence of a pair of sharp negative bands in the difference spectra is a marker of the formation of PYP_H. Thus, the photo-steady-state mixture produced by blue light contained PYP_H as well as PYP_B (curve 2′ in inset).

Irradiation of M100A with blue light produced a blueshifted intermediate having a maximum at 380 nm (curve 5 in Fig. 2). The spectroscopic properties of this photoproduct were similar to those of PYP_BL (M100A_BL). Concomitantly, a small amount of redshifted intermediate was formed (curves 5 and 5′). However, the absorbance increase at 490 nm indicating the formation of this product (M100A_B′) was much smaller than that of PYP_B. In addition, the absorption band of M100 A_B′ was located in a wavelength region 20–30 nm shorter than that for PYP_B. The broad negative band without a pair of sharp negative bands (curve 5′) shows that the amount of M100A_H in the photo-steady-state mixture produced by the blue light was smaller than that of PYP_H.

Irradiation of M100A with yellow light induced a blueshift in the absorption spectrum (curve 6). The absorbance increase at 390 nm and a pair of negative bands in the difference spectrum (curve 6′) showed the formation of M100A_BL and M100A_H, respectively.

Thermal reactions of photoproducts formed at 80 K

The photo-steady-state mixture produced by blue light or yellow light at 80 K was warmed at a steady rate of 1 K min⁻¹, and the absorption spectra were recorded at every 10 K. Because the mixture was not equilibrated, the spectra of the transient mixture were recorded. The dark-state spectra measured at the same temperature were then subtracted to obtain the difference spectra between the dark state and the photocycle intermediate(s) (Fig. 3). Prior to examining M100A, the photocycle of WT was surveyed to validate the present experimental setup and analysis (broken lines in Fig. 3). Irradiation of WT with yellow light at 80 K yielded PYP_H, as shown by the pair of sharp negative bands and absence of a positive band (top of left-hand column). As a result of warming, the negative band turned broad, indicating the formation of PYP_HL (123–163 K) (11) and the redshifted intermediate (PYP_L) was formed above 183 K. At 213 K, PYP_L reverted to the dark state without the formation of the near-UV intermediate (PYP_M).

The mixture of PYP_B and PYP_H produced by blue light was similarly warmed (right-hand column). Warming the photo-steady-state mixture caused the decay of PYP_B to PYP_BL, which was completed at 133 K. At this temperature, the negative band was broadened due to the conversion from PYP_H to PYP_HL, as shown in the left-hand column. PYP_BL was converted to PYP_L above 173 K, and then PYP_L directly reverted to the dark state. These spectral changes were consistent with the previous low temperature spectroscopy in which the effect of temperature on the absorption spectrum was avoided by measuring at fixed temperature after warming (11).

The thermal reactions of the photoproducts produced from M100A were studied in the same manner (solid lines in Fig. 3). The photo-steady-state mixture produced by yellow light, which contained M100A_BL and M100A_H, was warmed (left-hand column). The sharp negative bands were first broadened (123–163 K), and then M100A_L was formed (183 K) and decayed above 193 K. Concomitant with the decay of M100A_L, the positive 380 nm band was shifted to 360 nm, showing that M100A_BL was converted to M100A_Mvia M100A_L (193–203 K).

M100A_B′ formed at 80 K showed a 20–30 nm blueshifted spectrum relative to that of PYP_B (right-hand column). Along with M100A_B′, M100A_BL was formed at 80 K by blue light. Upon warming, M100A_B′ decayed at 113 K, and M100A_L was formed at 183 K. While the spectral properties of M100A_B′ and M100A_L were very similar to each other, the decay and formation of the 480 nm species clearly showed that they are distinct species (see Fig. 4). At 193 K, M100A_L was converted to M100A_M. It should be noted that PYP_BL was completely converted to PYP_L (difference spectra of WT at 193–203 K), whereas M100A_BL was present over a wide temperature range (83–203 K). The 380 nm band of M100A_BL was shifted to 360 nm, indicating the formation of M100A_M, at 223 K. The difference spectra at 233 K clearly showed the formation of M100A_M, while PYP_M was not observed in this condition.

To illustrate the thermal behavior of intermediates, the heat-induced absorbance changes at typical wavelengths are presented (Fig. 4). For WT, the absorbance at 485 nm indicates the presence of PYP_B and PYP_L, whereas those at 388 and 350 nm indicate that of PYP_BL. The absorbance decrease at 485 nm is complementary to the increase at 388 or 350 nm and vice versa, indicating a simple sequential reaction. However, M100A_BL was present over a wide temperature range despite the decay and formation of M100A_B′ and M100A_L, showing the equilibrium between M100A_BL, M100A_B′ and M100A_L. The photocycle of M100A is illustrated in Fig. 5 in comparison with that of WT.

Discussion

In the present study, the photocycle of M100A was studied by low-temperature UV–visible spectroscopy. M100A was irradiated at 80 K and the subsequent thermal reactions were induced by warming the sample. Warming caused thermal broadening of the absorption spectra as well as decay of the thermolabile species. The percentages of the photocycle intermediate(s) contained in the photo-steady-state mixture produced at 80 K were estimated to be 30% for blue light and 10% for yellow light. Therefore, the effect of temperature on the absorption spectrum of the major component (dark state) is not negligible. To minimize the effect of temperature on the dark-state spectrum, the difference spectra were calculated by subtracting the dark-state spectra measured at the same temperature (Eq. 2). In this method, the effect of temperature on the absorption spectra was small enough to follow the decay-associated spectral changes (Fig. 3).

Above 213 K, PYP_L in 66% glycerol buffer directly reverted to PYP_dark, indicating that the cis-to-trans back-isomerization takes place more readily than the proton transfer from the chromophore to Glu46 (25,26). In contrast, M100A_L decayed at lower temperatures than PYP_L and was converted to M100A_M, suggesting that the equilibrium between M100A_L and M100A_M (27,28) is shifted toward the latter. It is well known that the decay of M100A_M at room temperature is extremely slow, while I₁ of M100A (M100A_L) decays very fast (18). Therefore, the lower decay temperature of M100A_L compared to PYP_L is consistent with the findings at room temperature.

The difference spectra obtained by low-temperature spectroscopy of M100A before and after irradiation were quite different in shape from those of WT. However, the photocycles of M100A and WT were explained by qualitatively similar intermediates by taking the thermal equilibrium into consideration (Fig. 5). The most marked difference was the presence of M100A_BL over a wide temperature range. For WT, PYP_BL is not formed at 80 K—it is thermally formed from PYP_B above 93 K. The apparent decay temperature of M100A_B′ was 20 K higher than that of WT. If the back reaction from M100A_BL to M100A_B′ does not take place, M100A_B′ should be more stable than PYP_B. However, M100A_BL formed at a lower temperature than PYP_BL, indicating the presence of equilibrium between M100A_B′ and M100A_BL. In addition, the decay of M100A_BL and M100A_L took place simultaneously. Therefore, M100A_B′, M100A_BL and M100A_L are in thermal equilibrium, which favors M100A_BL (Fig. 5).

We previously proposed that the progress of the photocycle from PYP_B or PYP_H to PYP_L is a consequence of the relaxation of the twisted double bond of the chromophore (29). The photo-isomerization of the chromophore takes place by flipping the carbonyl oxygen (O1) of the thioester bond (22,25,29–31) (Fig. 6). Although the thermal equilibria between the photocycle intermediates are markedly shifted by mutation of Met100, it is unlikely that Met100 directly interacts with the chromophore, because the distance between C3 of the chromophore and S of Met100 is 4.10 or 4.51 Å in PYP_B (PDB ID 1uwp) (30) and 4.53 Å in PYP_BL (3pyp) (31). On the other hand, trans-to-cis isomerization moves O1 toward the phenyl ring of Phe96, which may cause steric hindrance. In addition, electrostatic repulsion would be present between O1 and the π-electron system of the phenyl ring. Thus, it is likely that PYP_B, PYP_BL, PYP_H, PYP_HL and PYP_Ldiffer as regards to the interactions between O1 and Phe96. S of Met100 and N of Arg52 are located at a distance of 3.88 Å (PDB ID 2D01), suggesting weak interaction via the NH-S hydrogen bond (32). Mutation of Met100 eliminates this interaction, resulting in a structural change in the Met100 loop. As a consequence, M100A_BL would be stabilized by the displacement of the side chain of Phe96. At low temperatures, the reverse reaction from PYP_L to PYP_dark is dominant over the conversion to PYP_M, and would also be driven by the interaction between O1 and Phe96.

Mutation of Phe96 substantially slows the decay of PYP_M. The decay time constant of F96A is 1000 s (33), which is comparable to that of M100A_M. While this suggests the occurrence of the interaction between O1 and Phe96 in PYP_M, it is not likely that the spatial relationship between the chromophore, Phe96 and Met100 demonstrated by the crystallography of PYP_B is maintained in PYP_M, because the Met100 loop undergoes a large conformational change (34–36) that has not been observed by crystallography (37). Thus, the mechanism by which Met100 catalyzes the cis-to-trans isomerization remains unclear because of the lack of the high-resolution structure of PYP_M in the physiologic condition. The lifetimes of M100A_M, M100K_M and M100L_M are 1000 times longer than that of PYP_M, but M100E_M decays 10 times faster than M100A_M (20), suggesting that the proton acceptor for the hydrogen bond at position 100 facilitates the decay of PYP_M. Thus, the NH-S hydrogen bond between Met100 and Arg52 partially contributes to the decay of PYP_M, as indicated by the fact that the lifetime of R52Q_M, in which the distance between S of Met100 and N of Gln52 is 8.57 Å in the dark state (PDB ID 2D02) (38), is five times longer than that of PYP_M.