Structure and Photoreaction of Photoactive Yellow Protein, a Structural Prototype of the PAS Domain Superfamily

Authors


  • This invited paper is part of the Symposium-in-Print: Photobiology in Asia.

*email: imamoto@ms.naist.jp (Yasushi Imamoto)

Abstract

Photoactive yellow protein (PYP) is a water-soluble photosensor protein found in purple photosynthetic bacteria. Unlike bacterial rhodopsins, photosensor proteins composed of seven transmembrane helices and a retinal chromophore in halophilic archaebacteria, PYP is a highly soluble globular protein. The α/β fold structure of PYP is a structural prototype of the PAS domain superfamily, many members of which function as sensors for various kinds of stimuli. To absorb a photon in the visible region, PYP has a p-coumaric acid chromophore binding to the cysteine residue via a thioester bond. It exists in a deprotonated trans form in the dark. The primary photochemical event is photo-isomerization of the chromophore from trans to cis form. The twisted cis chromophore in early intermediates is relaxed and finally protonated. Consequently, the chromophore becomes electrostatically neutral and rearrangement of the hydrogen-bonding network triggers overall structural change of the protein moiety, in which local conformational change around the chromophore is propagated to the N-terminal region. Thus, it is an ideal model for protein conformational changes that result in functional change, responding to stimuli and expressing physiological activity. In this paper, recent progress in investigation of the photoresponse of PYP is reviewed.

Introduction

Visible light can be utilized as energy by many kinds of organisms but short wavelength light is harmful, therefore detection of the light environment is essential for survival. A photon is captured by the prosthetic group (chromophore) embedded in the photosensor protein. Light energy is converted to enzymatic activity by a protein conformational change, and then the signal is processed in the cell. Therefore, understanding the signal transduction mechanisms equates with understanding the basic mechanism of life. Investigation of photoactive yellow protein (PYP), an excellent photoreceptor model, is therefore reviewed here.

PYP was first found in the purple photosynthetic bacterium Halorhodospira halophila (Ectothiorhodospira halophila) as a soluble, yellow-colored acidic protein that adsorbed to DEAE cellulose (1,2). Soon after its discovery, PYP was found to be photoactive: light stimulus bleaches PYP, however, its color recovers in the dark within 1 s (3). As PYP is a small soluble protein, its high-resolution structure was established by crystallography in 1995 (4). Before crystallographic studies of visual rhodopsin (5,6), the crystal structure of PYP was the only characterized example of photoreceptor protein that utilizes light as a signal. Because of the characterization of the high-resolution structure and reversible color change likely to parallel the conformational change, PYP is an excellent candidate for understanding the signal transduction mechanisms of sensory proteins at the atomic level.

To date, PYP-like proteins (PYPs) and PYP-related domains (PYP-domains) have been found in at least seven species. PYPs from Hr. halophila, Rhodothalassium salexigens and Halochromatium salexigens are proposed to function as photosensors for the photophobic response of these bacteria (7). In contrast, PYPs from Rhodobacter capsulatus and Rhodobacter sphaeroides are proposed to be involved in the regulation of cell buoyancy (8). PYP-domains of Ppr (PYP/Phytochrome Related) from Rhodospirillum centenum and Ppd (PYP/bacteriophytochrome/diguanylate cyclase/phosphodiesterase) from Thermochromatium tepidum are the sensing domains of bacteriophytochromes with histidine kinase and diguanylate cyclase/phosphodiesterase domains, respectively (9,10). While PYP from Hr. halophila shows a single absorption band in the visible-near UV region, PYPs from Rb. sphaeroides and Rb. capsulatus show two absorption bands (Fig. 1). The PYP-domain of Ppr shows a single band but the absorption maximum is slightly blue-shifted.

Figure 1.

 Absorption spectra of PYPs from Hr. halophila, Rb. capsulatus and PYP-domain from Rs. centenum.

Structure of PYP

PYP is composed of 125 amino acids. It is folded into an α/β fold structure with a central six-stranded β-sheet and five short α-helices (4,11) (Fig. 2). This structure is typical of the per arnt sim (PAS) domain superfamily, widely distributed, from humans to bacteria. Many PAS domain proteins function as sensors for various stimuli, such as light, oxygen, voltage and oligomerization state. In most cases, they are present as part of a multidomain protein, however, PYPs function with PAS domain alone, suggesting it as the structural prototype for PAS domain proteins.

Figure 2.

 Structure of PYP. Left: overall structure of PYP divided into four segments: N-terminal cap (purple), PAS core (orange), helical connector (green) and β-scaffold (blue) (12). Right: chromophore of PYP. The p-coumaric acid chromophore (pCA), Tyr42, Glu46, Thr50, Arg52, Phe96 and Met100 are shown.

From similarity to other PAS domain proteins and possibly related to function, the protein moiety of PYP is divided into four segments: the N-terminal cap, PAS core, helical connector and β-scaffold (12) (Fig. 2). The β-scaffold segment is composed of relatively long β-strands (β4–β6) and the loops connecting them. The helical connector segment is composed of the C-terminal half of the chromophore binding loop and a relatively long α-helix (α5). The PAS core segment is composed of relatively short β-strands (β1–β3), short α-helices (α3–α4) and the N-terminal half of the chromophore binding loop. The amino acid sequence of this segment shows high homology between PAS domain sequences. The N-terminal cap is composed of short α-helices (α1–α2) and loops, with the chromophore located on the other side of the β-sheet.

For visible light absorption, PYP has a chromophore with a conjugated double bond system. Based on the similarity of photocycle kinetics to sensory rhodopsins from halobacteria, the chromophore of PYP was first thought to be a retinal, but then determined to be a p-coumaric acid (4-hydroxy cinnamic acid) binding to Cys69 via a thioester bond (13–15) (Fig. 2). Thioester is readily hydrolyzed in aqueous solution in general, however, the thioester linkage of the chromophore in native PYP is very stable. It is cleaved by hydroxylamine, dithiothreitol, performic acid and high pH (15,16). The p-coumaroyl chromophore of PYP has one isomerizable double bond and one ionizable oxygen atom (phenolic oxygen, O4′). In the dark state (PYPdark), O4′ of the chromophore is deprotonated and the double bond is in the trans form (13,17), which is isomerized to cis form on photon absorption (18–20). The absorption maximum of PYP from Hr. halophila is located at 446 nm (Fig. 1), resulting in its bright yellow color. Deprotonation of O4′ significantly contributes to the redshift of the absorption spectrum of p-coumaric acid. For PYP the pKa of O4′ (∼2.5) is significantly lower than free p-coumaric acid (∼9) because of the hydrogen bonds with Tyr42 and Glu46. The PYP-domain of Ppd which possesses Leu instead of Glu46 has the protonated chromophore and shows an absorption band in the near-UV region (absorption maximum = 358 nm) (21). However, absorption maximum of denatured PYP at alkaline pH is ∼400 nm, indicating that the yellow color of PYP is not solely because of deprotonation of the chromophore. In native PYP, distortion of the chromophore as well as the hydrogen-bonding network involving O4′ and hydroxy groups of Tyr42, Glu46 and Thr50 also contribute to the characteristic yellow color. Mutations of Tyr42, Glu46 and Thr50 result in a shift in absorption maximum (22–24). Arg52 is located near the chromophore and is likely to be the counter-ion of the negatively charged chromophore (4). However, mutation of Arg52 results in only small changes in absorption maximum. Theoretical analysis suggests that the PYP spectrum is tuned by a counter-ion effect, hydrogen-bonding effect, and medium effect of the protein matrix (25).

Tandem expression experiments have shown that for the biosynthesis of PYP, p-coumaric acid is produced from tyrosine by tyrosine-ammonia lyase; it is then attached to the thiol group of coenzyme A with p-hydroxycinnamic acid ligase (26,27). Formation of a thioester bond between p-coumaric acid and cysteine in vitro is not as easy as the Schiff base bond between retinal and lysine, however, p-coumaric anhydride, p-coumaric thiophenyl ester (15) and carbonyl diimidazole derivative (23) react with the thiol group of the cysteine residue to reconstitute PYP. These reconstitution techniques enabled the preparation of PYP using an Escherichia coli over-expression system. Wild-type PYP plus its mutants and PYP analogs with artificial chromophores can also be prepared in the same system. Currently, 100 mg PYP is obtained from 1 L of E. coli culture medium, while only 0.3 mg PYP is obtained from 1 L of Hr. halophila culture medium.

PYP Photocycle

Light-induced reversible bleaching of PYP

Transient spectroscopy, measured over a microsecond time scale, first indicated that on photon absorption, the absorption spectrum of PYP redshifts in microseconds, then bleaches as a result of the large blueshift to the near-UV region, then rapidly reverts to the dark state (PYPdark) in 100 ms time scale (3). Red-shifted and blue-shifted intermediates are referred to as I1 and I2 (Table 1). The rate of the photocycle (decay rate of I2) is affected by pH, denaturant, salt, detergent and other reagents (3,28). A lifetime of 100 ms is long enough to trigger the enzymatic activity of the partner if the near UV intermediate (I2) is the physiologically active state (signaling state) like meta (M) intermediates of the retinal proteins. The photocycle of PYP has been extensively studied in detail using room temperature transient spectroscopy (29–31) as well as low-temperature trapping spectroscopy (32,33). In the early stage, correspondence between transient intermediates at room temperature and steady-state intermediates trapped at low temperature was not definitive, therefore several lines of nomenclature were proposed. The current nomenclature and correspondence, discussed here later, are summarized in Table 1.

Table 1.   Correspondence of transient intermediates to steady-state intermediates.
SeriesDark stateTime region (s)
10−1110−9 10−4 10−310−1 
  1. Series 1 and 2 are based on transient spectroscopy, whereas Series 3 is based on steady-state spectroscopy.

1PYPI0I0I1I1I2I2 
2PYP-pR pB′pBpBdeprot
3PYPPYPBPYPBLPYPL PYPMacidPYPMalkali
(PYPM′)
 

Low-temperature spectroscopy

Figure 3 shows the photocycle of PYP determined from low-temperature spectroscopy (32,33). Several photocycle intermediates of PYP are trapped at low temperature. Although the excited state is not detected by low-temperature trapping experiments, a red-shifted photoproduct was found at 80 K (Fig. 4). This intermediate was first identified as I1 (pR), a red-shifted intermediate found over a microsecond time scale by transient spectroscopy at room temperature (32). Subsequently, more detailed analysis at low temperature showed that two distinct red-shifted intermediates are formed at 80 and 190 K (33). To discuss how these intermediates correspond to transient intermediates at room temperature, it was proposed to name the photoproducts found in the low-temperature trapping spectroscopy after photobleaching intermediates of visual rhodopsin (bathorhodopsin, hypsorhodopsin, lumirhodopsin and metarhodopsin) (33). Thus the red-shifted intermediate of PYP trapped at 80 K was called PYPB. The slightly blue-shifted intermediate (PYPH) is simultaneously produced with PYPB at 80 K. PYPdark, PYPB and PYPH are photoreversible at very low temperature, suggesting that their difference is in configuration of the chromophore. Structures of PYPB and PYPH were studied using low-temperature Fourier transform infrared (FTIR) spectroscopy (34) (Fig. 5). The C = O stretching mode of the carboxylic acid of Glu46 is present at 1730–1740 cm−1 (35,36). It is located at 1739 cm−1 for PYPdark, but shifted to 1732–1733 cm−1 for PYPB and PYPH, indicating that Glu46 is protonated and forms a hydrogen bond stronger than PYPdark in these intermediates. The band at 1300 cm−1 for PYPdark indicates that its chromophore is in trans form, whereas those for PYPB and PYPH at 1286–1288 cm−1 indicate their chromophores are in cis form. Resonance Raman spectroscopy for PYP intermediates suggested that 994–995 cm−1 bands of PYPB and PYPH are markers of the cis chromophore (37). The structural basis of PYPB and PYPH has been proposed from cryotrapped crystallography (34).

Figure 3.

 Photocycle of PYP indicated by steady-state spectroscopy (33,41).

Figure 4.

 Absorption spectra for the photocycle intermediates of PYP (33,41).

Figure 5.

 Difference FTIR spectra between PYPdark (negative) and intermediates (positive) (34,41). PYPB/PYP measured at 80 K, PYPH/PYP measured at 80 K, PYPL/PYP measured at 190 K, PYPMacid/PYP measured at 274 K, pH 5.0 and PYPMalkali/PYP measured at 274 K, pH 8.0 (from top to bottom).

PYPB is converted to a blue-shifted intermediate (PYPBL), then to another red-shifted intermediate (PYPL). PYPL is stable at 193 K and shows a similar absorption spectrum to I1 (33). The 1732 cm−1 band for PYPL observed using low-temperature FTIR spectroscopy showed that Glu46 of PYPL is still protonated and forms a hydrogen bond comparable to PYPB and PYPH (Fig. 5). The intense 1169 cm−1 band indicates that the chromophore of PYPL is deprotonated. Resonance Raman spectroscopy also confirmed that PYPL has a deprotonated cis chromophore (37). In 66% glycerol medium, the near-UV intermediate (PYPM) was not produced (33), however, PYPM was formed in a hydrated dry film at 233 K (34,36). FTIR spectroscopy showed that PYPM has deprotonated Glu46 and a protonated cis chromophore (35,36), strongly suggesting that a proton at Glu46 is transferred to O4′ of the chromophore upon decay of PYPL. The change in pKa of the chromophore that causes proton transfer is partially explained by displacement of the side chain of Arg52 (39). Absorbance change in the amide mode (1500–1700 cm−1) during the formation of PYPM was small at 233 K, suggesting that protein conformational change in the frozen sample is restricted to changes possible in the crystalline form (40).

Although PYPM is unstable at room temperature, a photo–steady-state mixture containing PYPM is produced by continuous illumination. Steady-state spectroscopy for this mixture showed that two states of PYPM are present and they are in pH-dependent equilibrium (PYPMacid and PYPMalkali) (41), with absorption maxima at 367 and 356 nm, respectively (Fig. 4). Difference FTIR spectra in the amide mode (1500–1700 cm−1) (Fig. 5), circular dichroism spectra in the far-UV region and radius of gyration (Rg) values estimated by small-angle X-ray scattering (SAXS) measurement showed that structural change during the formation of PYPMacid is less than PYPMalkali.

The pKa value of equilibrium between PYPMacid and PYPMalkali was estimated to be 6.4 (41). Characteristics of PYPM vary according to the solvent condition. pH-dependence of time constants for the recovery of PYPdark follows the Henderson–Hasselbalch equation, with a pKa of 6.4 (23). Proton uptake per PYP during the photocycle also follows the Henderson–Hasselbalch equation, with a pKa of 6.6 (42). The absorption spectrum of E46QM is significantly red-shifted from PYPM at pH 7.0 (43). These findings are consistently explained by a shift in this equilibrium (41). PYPM is also in equilibrium between PYPL (44–48), which is enriched at alkaline pH (apparent pKa = 10.4). At alkaline pH, PYPM is deprotonated to show the absorption maximum at 410 nm (apparent pKa = 10.2).

Transient spectroscopy at room temperature

On photon absorption the π-electron of the p-coumaroyl chromophore is excited. Unlike low-temperature experiments, the excited state is observed using ultrafast spectroscopy. Extensive studies on excited state dynamics have been carried out using fluorescence up-conversion, UV-visible or infrared absorption spectroscopy to understand the highly efficient isomerization mechanisms of PYP chromophore (49). The excited electron is rapidly relaxed to the ground state (30,31,50) to form the first photoproduct. The quantum yield during the formation of the photoproduct is estimated to be 0.35 (51)–0.64 (28). Experimental (52,53) and theoretical (54) analyses showed that amino acid residues near the chromophore accelerate the reaction. Ultrafast infrared spectroscopy confirmed that the primary photochemical event, i.e. isomerization of the chromophore from trans to cis occurs (55,56). The first ground-state photoproduct detected at room temperature is I0. It is a red-shifted intermediate like the first photoproduct of retinal proteins (photo [J] or batho [K] intermediates). The spectral characteristics of I0 are similar to PYPB. It is converted to I0 then I1. The absorption spectrum of I0 is very similar to I0 but detected kinetically (31), whereas the absorption maximum of I1 is located between PYPdark and I0. Time-resolved FTIR (40,57) and optical rotatory dispersion (58) measurements demonstrate that the structural change of the protein moiety during the formation of I1 is not so large, but a small volume decrease was observed (59). I1 is stable over a microsecond time scale, then bleaches in two steps. In the first stage, the chromophore is protonated, however, global conformational change in the protein moiety does not take place at this stage (I2 or pB′) (40,57). The rate constant for the protonation of the chromophore agrees with that of deprotonation of Glu46, strongly suggesting that Glu46 is a proton donor to the chromophore (40,57). However, water molecule possibly donates a proton because the chromophore is efficiently protonated also in E46Q and E46A mutants (60). A large conformational change then takes place upon the formation of I2′ (pB) as demonstrated by a large absorbance change in the amide mode region of the difference FTIR spectra (40,57,61). Transient spectroscopy has shown that the intermediates from I1 through I2′ are in pH-dependent equilibria (45–48) as well as steady-state spectroscopy (41,44). Other techniques like Raman spectroscopy or transient grating have shown that spectroscopically indistinguishable intermediates are present in these steps (62,63). Therefore, it is likely that progress of the photocycle for the chromophore and protein moiety are not necessarily concerted.

The key amino acid residue for recovery of PYPdark from I2′ is Met100 (64,65), located in β4–β5 loop (100 loop). Importance of other residues in the 100 loop have also been proposed (47). Theoretical analysis suggests that the 100 loop is highly flexible and involved in conformational change (66,67). Thermal recovery from I2 intermediates of mutants in which Met100 is replaced is significantly slower, but UV light that induces cis–trans isomerization of the chromophore readily recovers the dark state (64). Similarly, UV light substantially accelerates the photocycle of native PYP (46). Hence, the rate-limiting step for the PYP photocycle is re-isomerization of the chromophore catalyzed by Met100, rather than refolding the overall protein moiety. Interaction between aromatic rings of the chromophore and Phe96 is also required for rapid re-isomerization of the chromophore (68) (Fig. 6), suggesting that a hydrophobic environment facilitates re-isomerization.

Figure 6.

 Key intramolecular interactions for the stability and photocycle of PYP.

The photocycle of the PYP-domain of Ppr (Ppr-PYP), in which Met100 is conserved, is significantly slow (10), which is explained by the distal location of Met100 from the chromophore (69). The signaling state interacts with the downstream transduction system in the cell. Large differences in photocycle kinetics would correlate with differences in physiological function (70). Ppr-PYP is important as a model system for interdomain interaction of PAS domains as it is a part of a multidomain protein, like other PAS domains.

Correspondence of steady-state intermediates to transient intermediates

As the photoreaction at low temperature is essentially the same as at room temperature for most retinal proteins, the correspondence of PYP intermediates detected by the transient spectroscopy to those by low-temperature trapping spectroscopy will be discussed. The first photoproducts at both room temperature and 80 K (I0 and PYPB, respectively) are red-shifted ones. Absorption maxima of I0 and PYPB are estimated to be 510 nm (30,31) and 489 nm (33), respectively. Recently, difference FTIR spectra before and after excitation over a picosecond time scale were reported (56). Although there was a significant difference in temperature of the measurements, I0/PYP spectrum resembled the PYPB/PYP spectrum (Fig. 5) (34). Both I0 and PYPB showed a characteristic mode at ∼1000 cm−1, however, intensity of the other modes was small. These findings strongly suggest that I0 is trapped at 80 K (PYPB). The other red-shifted intermediates, I1 and PYPL, showed similar visible absorption (29,33) and difference FTIR spectra (34,57). These results indicate that PYPB corresponds to I0 and PYPL to I1 (Table 1). The presence of PYPH at room temperature has not been established, but the absorbance increase for the formation of I0 relative to the absorbance decrease for the loss of PYPdark varies according to experimental conditions. This cannot be explained without the presence of an intermediate other than PYPdark and PYPB (30). Therefore, it is likely that the ratio of PYPB and PYPH is dependent on experimental conditions such as excitation wavelength, temperature and/or pH. I0 is formed between I0 and I1, and PYPBL is formed between PYPB and PYPL. However, I0 and PYPBL are clearly in distinct states as their absorption maxima are quite different. Rather PYPB may be a mixture of I0 and I0. At room temperature, two near-UV intermediates (I2 and I2′) are sequentially formed that revert to PYPdark (46,71). Transient spectroscopy has shown that I2 and I2′ are in pH-dependent equilibrium (46,48). Thus, PYPMacid and PYPMalkali correspond to I2 and I2′, respectively.

Structural change during the photocycle

Structural changes in crystal form and solution

As crystals of high quality are available, light-induced protein structural change of PYP has been extensively studied by using time-resolved crystallography over millisecond (19) and nanosecond (72) time scales or cryotrapped crystallography (73). Photon absorption induces isomerization of the chromophore and rearrangement of the nearby amino acid residues. Further detailed studies have been reported (74,75) and events that take place in crystalline form have been clearly observed. These reports are informative in understanding the photocycle of PYP, but they do not satisfactorily explain of photoreactions deduced from several kinds of spectroscopic studies in solution. Therefore, photoreaction in the crystal should be carefully examined to determine whether it reflects the physiological conditions.

The first notable issue is the direction of the chromophore in intermediates. In PYPdark, the O4′ of the chromophore is hydrogen bonded with Glu46 and Tyr42, plus carbonyl oxygen of the thioster bond (O1) is hydrogen-bonded with the main-chain nitrogen of Cys69 (Fig. 2). Cryotrapped crystallography for PYPBL showed that photo-isomerization of the chromophore occurs by flipping O1, with the result that the hydrogen bonds of O4′ are conserved, but the hydrogen bond of O1 is disrupted (73). In contrast, time-resolved crystallography showed that the hydrogen bond of O1 is conserved but the hydrogen bonds of O4′ are disrupted in PYPL (I1) and PYPM (I2) (19,72). As the absorption spectrum of PYP is sensitive to the hydrogen binding network involving O4′, the chromophore binding site of the intermediates was probed using Y42F, E46Q, T50V, and R52Q mutants, in which the hydrogen-bonding network is altered, in low temperature (43) and room temperature (76) spectroscopy. These results demonstrate that rank order of the redshift for PYPB, PYPH and PYPL by these mutations is the same as PYPdark, suggesting that the environment of O4′ is not altered in intermediates up to PYPL. FTIR spectroscopy also demonstrated the presence of hydrogen bond involving Glu46 (34,40,57).

Absorption spectra of R52QM, Y42FM and T50VM are the same as PYPM (43). The absorption maximum of E46QM is significantly red-shifted but it is explained by the shift of equilibrium between PYPMacid and PYPMalkali toward the former (41). These findings suggest that the hydrogen-binding network of O4′ is substantially weakened when O4′ is protonated. However, PYPMalkali shows pKa of the chromophore to be 10.2, greater than free p-coumaric acid by 1.6. Absorption maximum of PYPMalkali is 5 nm red-shifted from acid-bleached state. Also, rapid thermal cis–trans isomerization, never seen for free cis-p-coumaric acid, takes place in the chromophore binding site of PYPM. Therefore, chromophore/protein interaction is likely to remain in PYPM.

Assuming that these crystallographic intermediates are successively formed, the whole chromophore rotates on conversion from PYPBL to PYPL (pR), with this rotation volume-demanding. Flipping of H3 (Hula-twist) switches the crystallographic chromophore structure of PYPL (pR) and PYPM (I2) to that of PYPdark (76). This process is volume-saving but efficiency is likely to be low. Raman spectroscopy has suggested that thioester carbonyl oxygen is involved in the hydrogen bond in PYPL and PYPM (37), although it is not certain that the partner is the same as in the dark state.

The second issue is the magnitude of protein conformational change in PYPM. Although structural change of the protein backbone is small in crystallography (19), FTIR experiments demonstrate that further structural change should take place in solution (40,57,61,77). SAXS experiments also demonstrate a substantial increase in Rg (78,79). Various experiments have shown that PYPdark is in a well-ordered state, but PYPM is in a partially unfolded state (80–82) and its characteristics are similar to the molten globule state. Packing in the crystalline form would prevent PYP from the large protein structural change. However, recent studies have established that PYPM is in an equilibrium composed of two distinct states (PYPMacid and PYPMalkali). As the large protein conformational change occurs upon the transition from I2 (PYPMacid) to I2′ (PYPMalkali) (40,57), PYPMacid has a protein structure comparable to PYPdark (41). The difference in protein structure of PYPM between the crystalline form and solution may be partially explained by the shift of the equilibrium towards I2 (PYPMacid) in the crystal.

Structure of putative signaling state

Although the signaling pathway downstream of PYP has not been identified, the substantially altered protein structure and near-UV absorption spectrum of PYPM (PYPMalkali) suggests that it is a signaling state on the analogy of retinal proteins. During the formation of PYPM, surface properties are altered. Reactivity to hydrophobic reagents and lipid increases, suggesting that the hydrophobic region is exposed by the conformational change (83,84). Time-resolved crystallography suggest that Arg52 is flipped in I2 (PYPM) state and exposed to the solvent (19). The surface of PYPM in solution was probed using organic anions like citrate. Citrate binds not to PYPdark but to PYPM, probably through exposed Arg52 (78). These changes in the surface would result in the formation of an interface to react with the transducer protein.

Photocycle kinetics are largely affected by the N-terminal region (77,85,86). The N-terminal cap is specifically digested by chymotrypsin at the C-terminal side of Phe6, Leu15, and Leu23 (85). As net charge is changed by this cleavage, these fragments are separated using DEAE-Sepharose column chromatography. The lifetime of PYPM is decelerated according to the number of truncated amino acids. It is notable that truncation of only 6 amino acid residues in the N-terminus slows the photocycle by ∼100 times. Recently, it was shown that the key interaction between the N-terminal cap and β-sheet is a weak CH/π interaction between the phenyl ring of Phe6 and alkyl chain of Lys123 (87). These observations suggest that the light signal is propagated from the chromophore to the N-terminal region through intramolecular interactions, resulting in the global conformational change (Fig. 6). Relationship between the N-terminal structural change and large-scale motion is theoretically simulated using a coarse-grained model (88). The long-range intramolecular signaling mechanism is also studied using long-range intramolecular energy transfer pathway analysis, in terms of residue–residue energy conductivities (94).

The structure of PYPM in solution has been studied using NMR (89–92). However, structural analysis of PYPM at high resolution by NMR is difficult because of its disordered structure. Recently, the solution structure of PYPM formed from Δ25, which lacks the N-terminal 25 amino acids and forms stable PYPM, was determined using NMR (92). This model shows that the β-sheet is stable, but all α-helices are unfolded in PYPM. Using atomic coordinates from NMR (PDB 1XFN and 1XFQ) plus Crysol software (93), Rg of Δ25 and its photoproduct were calculated. Rg for Δ25 in the dark state is calculated to be 13.4 Å. Twenty structures of Δ25 M in PDB have given Rg of 16.1–17.4 Å, thus the increase in Rg from light is 2.7–4.0 Å. These increases in Rg are, however, significantly larger than the experimental value for wild-type (0.5–1.0 Å) or T23 (0.7 Å). The rate-limiting step for recovery of dark state is re-isomerization of the cis-chromophore, catalyzed by Met100. Thus, it is likely that the Met100 loop is more proximate to the chromophore in wild-type PYPM than in Δ25 M. The long distance between the chromophore and Met100 is likely to be the main cause for the extremely long lifetime of Δ25 M. Therefore, PYPM of wild-type would assume a more ordered structure than Δ25 M.

Several models have been proposed as the mechanism of large structural change of PYP. The light-induced infrared absorbance change in amide I mode as well as changes in the chemical shift of the mutant with an unchargeable residue at position 46 (E46Q) is significantly smaller than wild-type (40,41,91). Therefore, the generation of negative charge at position 46 is essential for structural change (protein quake model). The chromophore is embedded in the hydrophobic pocket. Time-resolved crystallography suggests that the chromophore is moved out of this pocket in I2 (19). If this is the case, the pocket would collapse, leading to global conformational change (hydrophobic collapse model) (2). Recent series of crystallography using wild-type and E46Q suggests that the structural change is transferred to Glu46, which is located in α3, then to N-terminus via the side chain of Asn43 (75). It is in contrast to our model that a weak CH/π interaction between the phenyl ring of Phe6 and alkyl chain of Lys123 mediates the structural change (87).

Conclusion

PYP is an excellent model of light capturing mechanisms of sensor proteins. Using a combination of high-resolution structure and spectroscopic techniques, the mechanism of light-induced protein conformational change in PYP is being elucidated. The key feature of the intramolecular signal transduction mechanism in PYP is the propagation from the chromophore to the N-terminal cap. The transducer protein that interacts with PYP has not been discovered, but its characterization is likely to open the way to elucidate the intermolecular signal transduction mechanism at the atomic level.

Acknowledgement— We thank Dr. John A. Kyndt of Ghent University for critical reading of the manuscript.

Ancillary