INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

pharaonis phoborhodopsin (ppR) from Natronobacterium pharaonis is a member of the archaeal rhodopsins (Kamo et al., 2001; Sasaki and Spudich, 2000). ppR activates a cognate transducer protein upon light absorption, leading to negative phototaxis. It possesses a retinal chromophore that is embedded within seven-transmembrane helices, like the well-studied proton-pump protein bacteriorhodopsin (BR) (Kamo et al., 2001; Sasaki and Spudich, 2000; Spudich and Lanyi, 1996). In ppR or BR, the retinal forms a Schiff base linkage with Lys-205 or Lys-216, respectively, and the protonated Schiff base is stabilized by a negatively charged counterion, Asp-75 or Asp-85, respectively. Light absorption of ppR triggers trans-cis photoisomerization of the retinal chromophore in its electronically excited state (Kandori et al., 2002b), followed by rapid formation of the ground-state species such as the K intermediate (Lutz et al., 2001). This process is also the case in BR. Relaxation of the primary intermediates eventually leads to functional processes during their photocycles (Kamo et al., 2001; Sasaki and Spudich, 2000; Spudich and Lanyi, 1996).

Comparative investigation of ppR and BR is a powerful method to understand their molecular mechanisms. We started such comparative studies by means of low-temperature FTIR spectroscopy. The results on the primary K intermediate revealed the structural similarity between ppR and BR on the polyene chain of the chromophore (Kandori et al., 2001b), and hydrogen bonds of internal water molecules (Kandori et al., 2001a). These observations were consistent with the similar crystallographic structures of ppR (Luecke et al., 2001; Royant et al., 2001) and BR (Belrhali et al., 1999; Luecke et al., 1999). In contrast, the structure of the K state after photoisomerization was more extended in ppR than in BR (Kandori et al., 2001b), which was probably correlated with the high thermal stability of ppR_K (Hirayama et al., 1992). In fact, ppR_K was stable even at 170 K, where the L intermediate was formed in BR (Kandori et al., 2001b).

Accompanying the relaxation of ppR_K and BR_L, the M intermediates appear by deprotonation of the Schiff base. The M intermediates of ppR and BR are functionally important in transducer activation and proton pumping, respectively (Kamo et al., 2001; Sasaki and Spudich, 2000; Spudich and Lanyi, 1996). ppR_M is formed in tens of microseconds, like BR_M, and decays in 1-2 s (Imamoto et al., 1992), which is two orders of magnitude longer than the lifetime of BR_M. The long-lived M-state in ppR must be advantageous in interaction with the transducer. The longer lifetime of ppR_M than BR_M predominantly originates from the lack of the proton-donating groups (Asp-96-Thr-46 in BR) to the Schiff base, because the M intermediate of the F86D/L40T (Phe-86 and Leu-40 in ppR correspond to Asp-96 and Thr-46 in BR, respectively) mutant of ppR decays as fast as BR_M (Iwamoto et al., 1999). By means of spin-labeling, Wegener et al. observed the opening of the F-helix at the cytoplasmic side during M formation (Wegener et al., 2000, 2001), as was the case for BR. These facts suggested similar structural changes at the M states in ppR and BR.

Despite functional importance, molecular understanding of ppR_M has been much less than that of BR_M. Engelhard et al. reported the difference infrared spectra between ppR_M and ppR, whereas vibrational bands have not been assigned (Engelhard et al., 1996). A positive band appeared in the carboxylic C==O stretching region upon M formation (Engelhard et al., 1996), suggesting the proton transfer from the Schiff base to a carboxylate. Because the structure of ppR is similar to that of BR (Belrhali et al., 1999; Luecke et al., 1999, 2001; Royant et al., 2001), the proton acceptor is likely to be Asp-75, a corresponding amino acid of Asp-85 in BR. Indeed, the D75N mutant experiment strongly suggests that Asp-75 is the proton acceptor (Schmies et al., 1998). Nevertheless, it has not been assigned to date. Protein structural changes are much less known for ppR_M, while extensive studies have been reported for BR_M (Lanyi et al., 2000).

One of the notable aspects in ppR is that the N state has not been well identified in its photocycle. It is reasonable because the corresponding amino acid residue of Asp-96 in BR is Phe-86 in ppR (Iwamoto et al., 1999). In BR, the M intermediate decays accompanying a proton transfer from Asp-96 to the Schiff base, and the formed N intermediate possesses the protonated Schiff base and deprotonated Asp-96 with largely changed protein structure. The N-specific protein structure can be described by the highly dichroic strong amide-I vibrations at 1671 (), 1663 (+), and 1649 (+) cm¹ in the BR_N minus BR difference infrared spectrum (Kandori, 1998). The frequency of the C==O stretch of Asp-85 is shifted from 1762 in BR_M to 1754 cm¹ in BR_N (Braiman et al., 1991; Hessling et al., 1993; Kandori, 1998; Ormos et al., 1992; Pfefferlé et al., 1991). Great changes in amide-I vibrations presumably correspond to the opening of the F-helix at the cytoplasmic side in BR_N, which was probed by diffraction (Dencher et al., 1989; Kamikubo et al., 1996; Subramaniam et al., 1999; Vonck, 1996) and spin-labeling (Rink et al., 2000; Thorgeirsson et al., 1997) experiments. When Asp-96 is replaced to Asn in BR, the N state is not observed and the M state is highly stabilized at alkaline pH. Previous FTIR spectroscopy of the D96N protein of BR revealed the appearance of the M_N state after the M state, where the chromophore was M-like (deprotonated) but the protein structure was N-like (largely changed) (Sasaki et al., 1992). This fact suggested that protonation of the chromophore was not a prerequisite for formation of the N-like protein structure in BR. Observation of the M_N state in D96N of BR then raised a question on ppR; does the M_N-like state appear during the photocycle of ppR?

In this paper we report the structural changes occurring upon formation of ppR_M by means of FTIR spectroscopy. By use of mutant proteins, we were able to assign the C==O stretch of Asp-75 and the C==O stretch of Asn-105 in the ppR_M minus ppR spectrum. These data provided similar protein structural changes between ppR and BR at the M state. In contrast, the present study revealed that the photocycle of ppR lacked the N-like protein structure, which was in clear contrast to that of BR. Lack of the N-like (M_N-like) structure in the photocycle of ppR may be substantial to its functional processes. Protein structural changes in the M intermediates are discussed on the basis of the present FTIR data.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Preparation of the ppR sample

The wild-type and mutant proteins of ppR were prepared as described previously (Kandori et al., 2001a, b; Shimono et al., 1997). Briefly, the ppR proteins possessing histidine tag at the C-terminus were expressed in Escherichia coli, solubilized with 1.5% n-dodecyl--D-maltoside (DM), and purified by Ni-column. The purified ppR sample was then reconstituted into L--phosphatidylcholine (PC) liposome by dialysis, where the molar ratio of the added PC was 50 times that of ppR.

FTIR spectroscopy

FTIR spectroscopy was applied as described previously (Kandori et al., 2001a, b). The ppR sample in the PC liposome was washed twice by buffers at pH 7 (2 mM phosphate) or 9 (2 mM borate). Five mM NaCl was added to test the chloride effect. A 90 µl sample of the ppR was dried on a BaF₂ window with a diameter of 18 mm. After hydration by either H₂O or D₂O, the sample was placed in a cell, which was mounted in an Oxford DN-1704 cryostat (Oxon, England) equipped in the Bio-Rad FTS-40 spectrometer (Cambridge, MA).

Illumination with >480 nm light (VY-50, Toshiba, Shizuoka, Japan) provided by a 1 kW halogen-tungsten lamp at 250 K for 90 s converted ppR to ppR_M. Because the ppR_M completely reverted to ppR upon illumination with a UV light (UG-5, Melles Griot, Irvine, CA) for 90 s, as evidenced by the same but inverted spectral shape, the cycles of alternative illuminations with >480 nm light and UV light were repeated a number of times. The difference spectrum was calculated from the spectra constructed with 64 interferograms after minus before the illumination. Twenty-four spectra obtained in this way were averaged for the ppR_M minus ppR spectrum.

Because photointermediates decay rapidly at 290 K, a slightly different experimental setup was applied to study the M_N state in ppR, as described previously (Chon et al., 1999). In this case, the sample film was tilted 45^o relative to the probe light, and the >480 nm light from a 150 W xenon lamp was focused on the sample at an angle of 90^o with respect to the probe light. During illumination infrared spectra were obtained, and the difference spectrum was calculated from the spectra constructed with 128 interferograms during minus before the illumination. Ten spectra obtained in this way were averaged at 250, 270, and 290 K.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Infrared spectral changes of ppR upon formation of the M photointermediate

We tested various conditions such as pH, temperature, and illumination wavelength to accumulate ppR_M. It was found that ppR and ppR_M can be photoconverted with each other at 250 K, as shown by the mirror images of the difference IR spectra. Photoreversion of ppR_M to ppR was previously reported by Balashov et al. (2000). The dotted line in Fig. 1 a shows the typical ppR_M minus ppR spectrum measured at 250 K for the hydrated film at pH 9. There were no spectral differences between pH 9 (dotted line) and 7 (solid line), which was consistent with the previous report (Engelhard et al., 1996). In contrast, the O intermediate appeared at higher temperature and neutral pH, whereas only ppR_M was observed at pH 9 even at higher temperatures (data not shown). Thus, alkaline pH is favorable to accumulate ppR_M, as was reported by the flash photolysis (Miyazaki et al., 1992). Such pH dependence is also the case for BR. FTIR studies on ppR_O will be reported elsewhere.

View larger version (29K): [in this window] [in a new window]   FIGURE 1   (a) ppRM minus ppR spectra measured at pH 7 (solid line) and 9 (dotted line) in the 1820-760 cm1 region. The spectra are measured at 250 K for the hydrated film with H2O. ppR is converted to ppRM with >480 nm light, while being reverted with a UV-pass filter. (b) ppRM minus ppR spectra measured in the absence (dotted line) and the presence (solid line) of NaCl at 250 K and pH 9. The dotted line is reproduced from the dotted line in a. One division of the y axis corresponds to 0.01 absorbance unit.

Fig. 1 b exhibits a spectral comparison of ppR_M between the absence (dotted line) and presence (solid line) of NaCl. Recent structural determination of ppR showed the presence of a chloride ion in the extracellular side of the chromophore (Royant et al., 2001). It is known that chloride does not affect the absorption of ppR (Shimono et al., 2000), while it is intriguing to determine whether the structure of ppR_M is influenced by chloride ions. Fig. 1 b shows that both spectra are identical in the 1500-800 cm¹ region. Frequencies in the 1800-1500 cm¹ were also the same between the absence and presence of chloride ions, though amplitudes of some bands were different. Thus, we concluded that chloride ions do not affect the protein structural changes between ppR and ppR_M.

The ppR_M minus ppR spectrum (Fig. 1 a) exhibits a negative peak at 1545 cm¹ in the ethylenic C==C stretching region. The value is identical to that in the native ppR (Engelhard et al., 1996) and also in good agreement with the previous resonance Raman spectroscopy of the native ppR in DM solution (1548 cm¹) (Gellini et al., 2000). Negative bands at 1244, 1202, and 1164 cm¹ are attributable to the CC stretching vibrations of the retinal chromophore. The 1244 and 1202 cm¹ bands were also observed in the ppR_K minus ppR spectrum, and tentatively assigned as a mixture of C12C13 stretch and NH in-plane bending, and C14C15 stretch in ppR, respectively, from the analogy to BR (Kandori et al., 2001b). The 1164-cm¹ band was not observed in the ppR_K minus ppR spectrum (Kandori et al., 2001b), and newly appeared in the ppR_M minus ppR spectrum. It is likely that the 1164-cm¹ band originates from C10C11 stretch, because the corresponding band of BR appears at 1170 cm¹ (Smith et al., 1985). This suggests that there are no structural changes at the C10C11 moiety between ppR_K and ppR, and some changes in ppR_M. It may also be possible that the chromophore structure is not altered at position C10C11 between ppR and ppR_M, whereas deprotonation in ppR_M weakens the absorbance of the IR band so that it appeared as a negative band. Another possibility is that the appearance of a different band at 1164 cm¹ cancels out the negative band at the same frequency in the ppR_K minus ppR spectrum, possibly because of the extended structural changes upon K formation (Kandori et al., 2001b).

Because the band corresponds to the C10C11 stretching vibration of the retinal chromophore, these results can be interpreted in terms of no structural changes between ppR_K and ppR, and some changes in ppR_M. It may also be possible that the chromophore structure is not altered at position C10C11 between ppR and ppR_M, whereas deprotonation in ppR_M weakens the absorbance of the IR band so that it appeared as a negative band.

The appearance of the positive 1765-cm¹ band is characteristic of the M intermediate (Engelhard et al., 1996; Maeda, 1995; Rath et al., 1996). In addition, various peaks were observed in the >1600 cm¹ region; 1707 (+), 1701 (), 1686 (), 1663 (), 1644 (+), 1634 (), and 1622 (+) cm¹. These bands originate from protein vibrations. The C==N stretching vibration of the protonated Schiff base at 1650 cm¹ is present as the shoulder of the strong negative band at 1663 cm¹. The negative 1663-cm¹ band is located at the typical frequency region of amide-I vibration in the II-helix.

Assignment of the C==O stretching vibrations of Asp-75 in ppR_M

The appearance of a positive band at 1770-1760 cm¹ upon formation of the M intermediate is characteristic among BR, ppR, and phoborhodopsin of Halobacterium salinarum (pR), and interpreted in terms of protonation of a carboxylate by the proton transfer from the Schiff base. By means of FTIR spectroscopy, the proton acceptors were identified as Asp-85 (Maeda, 1995) and Asp-73 (Bergo et al., 2000) in BR and pR, respectively, whereas the acceptor has not been assigned for ppR. Fig. 2 a shows the positive bands at 1765 and 1753 cm¹ in H₂O and D₂O, respectively, in the carboxylic C==O stretching region.

View larger version (31K): [in this window] [in a new window]   FIGURE 2   (a) ppRM minus ppR spectra of the wild-type protein in H2O (solid line) and D2O (dotted line) in the 1820-900 cm1 region. The spectra are measured at 250 K and pH 9. (b) Difference IR spectra of the wild-type (solid line) and the D75E mutant (dotted line) proteins in H2O. The latter spectrum is measured at 220 K and pH 9. One division of the y axis corresponds to 0.01 absorbance unit.

To assign the positive 1765-cm¹ band in the ppR_M minus ppR spectrum, we replaced Asp-75 to Glu. HPLC analysis revealed that the D75E mutant protein possesses 40% all-trans and 60% 13-cis forms in the dark, like BR, whereas the wild-type ppR possesses only the all-trans form (Hirayama et al., 1995; Shimono et al., 2001). We found that, unlike BR, D75E does not show light-adaptation, possibly because of the slow photocycle of the all-trans form. Therefore, we searched the suitable illumination conditions to accumulate the M intermediate of the D75E protein. Absorption spectra in the UV and visible region showed that the M intermediate is formed by illumination with >480 nm light at 220 K, being a lower temperature than for the wild type (not shown). Thus, we measured the difference IR spectrum for the D75E protein. The obtained spectrum exhibited negative bands at 1244 and 1202 cm¹ and a positive band at 1179 cm¹ in the fingerprint region (Fig. 2 b). The negative bands imply that the photocycle of the all-trans form was predominantly involved in the spectrum, whereas the appearance of the positive band at 1179 cm¹ may originate from formation of another product possessing a protonated Schiff base.

In the carboxylic C==O stretching region, a positive band was observed at 1729 cm¹ (Fig. 2 b). In BR, the C==O stretching vibrations of Asp-85 and Glu-85 appeared at 1760 and 1724 cm¹, respectively (Braiman et al., 1988). These facts strongly suggest that the positive 1765-cm¹ band originates from the C==O stretching vibration of Asp-75, which down-shifts to 1729 cm¹ in the D75E mutant protein of ppR. The C==O stretching vibrations of the aspartic acids at 1760 and 1765 cm¹ indicate that Asp-85 in BR and Asp-75 in ppR, respectively, exist in the highly hydrophobic environment (Dioumaev and Braiman, 1995). In addition, it is likely that the environment of the position 75 in ppR_M is more hydrophobic than that of the corresponding position 85 in BR_M, because higher frequency represents a more hydrophobic environment for carboxylates.

Assignment of the C==O stretching vibrations of Asn-105 in ppR_M

The ppR_M minus ppR spectrum has a characteristic peak pair at 1707 (+)/1701 () cm¹ (Fig. 1 a). In the ppR_K minus ppR spectrum, there is a peak pair at 1704 ()/1700 (+) cm¹, which is not shifted in D₂O (Fig. 3 a) (Kandori et al., 2001b). We recently assigned the bands as the C==O stretching vibration of Asn-105 in D-helix (Kandori et al., 2002a). In addition, from the analysis of the amplitude of the C==O stretches, we provided experimental evidence that photoisomerization yields more extended protein structural changes in ppR than in BR (Kandori et al., 2002a). Fig. 3 b shows the ppR_M minus ppR spectra measured in H₂O (solid line) and D₂O (dotted line). The bands at 1707 (+) and 1701 () cm¹ are not shifted in D₂O, while the positive 1695-cm¹ band disappeared in D₂O.

View larger version (21K): [in this window] [in a new window]   FIGURE 3   (a) ppRK minus ppR spectra of the wild-type protein in H2O (solid line) and D2O (dotted line). The spectra measured at 77 K are reproduced from Kandori et al., 2001b. (b) ppRM minus ppR spectra of the wild-type protein in H2O (solid line) and D2O (dotted line) measured at 250 K and pH 9. The solid line is reproduced from the dotted line in Fig. 1 a. (c) ppRM minus ppR spectra of the N105D mutant protein in H2O (solid line) and D2O (dotted line). One division of the y axis corresponds to 0.003 absorbance unit.

The bands at 1707 (+) and 1701 () cm¹ completely disappeared in the N105D mutant protein, where the D₂O-sensitive 1695-cm¹ band remained (Fig. 3 c). Instead, a negative peak newly appeared at 1739 cm¹, which shifts to 1728 cm¹ in D₂O. This band is close in frequency to the negative 1744-cm¹ band in the ppR_K minus ppR spectrum (Kandori et al., 2002a). Thus, we concluded that the bands at 1707 (+) and 1701 () cm¹ originate from the C==O stretch of Asn-105. In the difference spectrum of N105D, it seems that there is a positive peak at 1729 cm¹ in addition to the positive one at 1746 cm¹, which may suggest the structural heterogeneity in the M state of the mutant.

In ppR, the C==O stretching frequency of Asn-105 is down-shifted in ppR_K and up-shifted in ppR_M, indicating that the hydrogen bond is strengthened upon photoisomerization, and weakened upon primary proton transfer. In BR, it is known that the C==O stretching frequency of Asp-115 is down-shifted in BR_K and BR_L, and up-shifted in BR_M (Maeda, 1995; Sasaki et al., 1994). Thus, the structural changes at position 105 in ppR are similar to those at the corresponding position 115 in BR through light absorption and M formation.

Lack of the N-like (M_N-like) structure in ppR

In BR, the M intermediate decays accompanying a proton transfer from Asp-96 to the Schiff base, and the formed N intermediate possesses 1) a 13-cis chromophore; 2) a protonated Schiff base; 3) deprotonated Asp-96; and 4) largely changed protein structure. The N-specific protein structure can be described in the BR_N minus BR difference infrared spectrum by the highly dichroic strong amide-I vibrations at 1671 (), 1663 (+), and 1649 (+) cm¹ (Kandori, 1998), and the frequency shift of the C==O stretch of Asp-85 from 1762 (BR_M) to 1754 (BR_N) cm¹ (Braiman et al., 1991; Hessling et al., 1993; Kandori, 1998; Ormos et al., 1992; Pfefferlé et al., 1991). As described in the Introduction, protonation of the chromophore is not prerequisite for the formation of an N-like structure in BR. When Asp-96, an internal proton donor to the chromophore, is replaced to Asn, the M state is highly stabilized at alkaline pH. Previous FTIR spectroscopy of the D96N protein of BR revealed the appearance of the M_N state after M, where the chromophore is M-like (deprotonated) but the protein structure is N-like (largely changed) (Sasaki et al., 1992).

Observation of the M_N state in BR provides an implication for ppR. In ppR, the N state has not been well identified in its photocycle. It is reasonable because the corresponding amino acid residue of Asp-96 in BR is Phe-86 in ppR. Therefore, it is an intriguing question whether the M_N-like state is present during the photocycle of ppR. To answer this question, we measured infrared spectral changes of the alkaline film of ppR at various temperatures. BR_N is normally trapped at 273 K (Kandori, 1998). In the present study, we also examined higher temperatures (290 K). Consequently, we had to modify the experimental setup as described in Materials and Methods because photointermediates decayed rapidly at room temperature. Fig. 4 a shows the ppR_M minus ppR spectrum measured at 250 K. It is noted that the amplitudes of some bands were different in Fig. 4 a from those in Fig. 1 a, though frequencies were identical. Such difference presumably originates from the fact that the ppR molecules were partially oriented in the film. In Fig. 4 the film sample was tilted by 45^o, so that vibrations whose dipole moments were parallel to the membrane normal were enhanced in amplitude. In fact, the C==O stretch of Asp-75 and various amide-I vibrations whose dipole moments tend to be parallel to the membrane normal were likely to be enhanced in Fig. 4, while C==C and CC stretches of the chromophore whose dipole moments are parallel to the membrane were reduced in intensity from Fig. 1.

View larger version (33K): [in this window] [in a new window]   FIGURE 4   ppRM minus ppR spectra measured in the 1820-1090 cm1 region at 250 (a), 270 (b), and 290 (c) K. Dotted lines in b and c reproduce the spectrum of a for comparison. The spectra are measured for the hydrated film at pH 9. One division of the y axis corresponds to 0.008 absorbance unit.

Fig. 4 shows that spectral shape and amplitude do not change among 250, 270, and 290 K. This is clear contrast to the D96N mutant of BR, where the M state was observed at lower temperature, and the M_N state was observable at higher temperature (Sasaki et al., 1992). Formation of the M_N state accompanied 1) appearance of the prominent amide-I bands at 1669 () and 1649 (+) cm¹; and 2) frequency shift of the C==O stretch of Asp-85 from 1762 to 1755 cm¹ (Sasaki et al., 1992). In ppR, the positive 1765-cm¹ band of Asp-75 did not change its frequency at 250-290 K. In addition, there are no significant amplitude changes in the amide-I vibrations (1686 (), 1664 (), and 1644 (+) bands), which can be observed in N and M_N (Sasaki et al., 1992; Kandori, 1998). Thus, we concluded that the N-like (M_N-like) structure was not involved in the photocycle of ppR at alkaline pH.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

In this paper we studied the structure of ppR_M by means of FTIR spectroscopy. Balashov et al. reported that ppR_M was formed at ~220 K in their low-temperature visible spectroscopy (Balashov et al., 2000). However, we observed spectral contamination of ppR_K at 220-240 K in the present FTIR measurement (data not shown). The difference between the two experiments probably originates from the sample condition; the ppR molecule was solubilized by DM in the previous visible spectroscopy (Balashov et al., 2000), while the protein was reconstituted into the PC liposome in this work. Thus, it is likely that ppR_K is more stabilized in membrane than in the DM solution. A similar observation was reported for visual rhodopsin, where metarhodopsin II, having the deprotonated Schiff base like ppR_M, was highly stable in the DM solution (Hofmann et al., 1995).

The present study assigned the positive 1765-cm¹ band as the C==O stretch of the protonated Asp-75 in ppR_M (Fig. 2). Upon formation of ppR_M, Asp-75 receives a proton from the Schiff base, while BR_M formation accompanies a proton transfer from the Schiff base to Asp-85. Thus, such mechanism is common between ppR and BR. In addition, this study assigned the bands at 1707 (+)/1701 () cm¹ as the C==O stretches of Asn-105 (Fig. 3). Together with the previous report on Asn-105 in ppR_K (Kandori et al., 2002a), it was revealed that hydrogen-bonding alterations at position 105 in ppR (115 in BR) are common between ppR and BR, first strengthened upon photoisomerization, followed by weakened upon primary proton transfer from the Schiff base to Asp-75 in ppR (Asp-85 in BR). These observations provided the experimental evidence for the similar structural changes. However, there is a certain difference between ppR and BR. In the absence of transducer, ppR can pump protons, whereas the pumping efficiency is much lower in ppR than in BR (Schmies et al., 2001; Sudo et al., 2001a). This has to be explained in terms of structural factors.

The present study revealed that the photocycle of ppR lacked the N-like structure at alkaline pH. This observation was in clear contrast to that of BR. In the D96N mutant protein of BR, the N-like protein structure is formed with the deprotonated Schiff base (called M_N), even though the protein does not have an internal proton donor to the Schiff base (Sasaki et al., 1992). Thus, lack of the N-like (M_N-like) structure implies different structural changes between ppR and BR. They may be substantial to the low efficiency in the proton pumping of ppR (Schmies et al., 2001). It is, however, noted that the recent spin-labeling experiment of ppR reported the opening of F-helix in ppR and in BR (Wegener et al., 2000, 2001). This fact may suggest that the N-specific amide vibrations in BR_N are not directly correlated with the outward motion of F-helix. An alternative explanation is that the mechanism of the F-helix opening is different between ppR and BR.

The lack of the N-like (M_N-like) structure in ppR also suggests the mechanism of the interaction with its transducer protein. The protein structure characteristic of ppR_M is likely to activate its transducer protein. It is, however, noted that the association between ppR and its transducer is weakened upon formation of ppR_M (Sudo et al., 2001b, 2002). This may suggest that the structure of ppR_M is not important in the complex formation with the transducer. Rather, a ppR-transducer complex forces the transducer in the non-active state, and light-induced dissociation of the complex may be an essence of the transducer activation. Further experiments will lead to better understanding of the structural changes in the transducer activation.

The present study also provided implications for the chloride binding site. Royant et al. showed the presence of the chloride binding site inside ppR according to their x-ray structure (Royant et al., 2001), which was not visible in the x-ray structure of Luecke et al. (2001). The present results clearly showed little chloride effect on the ppR_M minus ppR spectrum (Fig. 1 b), suggesting that there are no structural alterations at the chloride binding site present in the extracellular side. In contrast, chloride ions influence the structure of ppR_O according to our FTIR spectroscopy (Furutani et al., manuscript in preparation). Thus, chloride ions are likely to influence the last step in the photocycle of ppR, which is our next focus.