(Received for publication, January 28, 1997, and in revised form, March 17, 1997)
From the Department of Earth and Space Science,
Graduate School of Science, Osaka University, Toyonaka, Osaka 560, Japan and the ¶ Suntory Institute for Bioorganic Research,
1-1, Wakayamadai, Shimamoto-cho, Osaka 618, Japan
Photoactive yellow protein (PYP) belongs to the
novel group of eubacterial photoreceptor proteins. To fully understand
its light signal transduction mechanisms, elucidation of the
intramolecular pathway of the internal proton is indispensable because
it closely correlates with the changes in the hydrogen-bonding network,
which is likely to induce the conformational changes. For this purpose, the vibrational modes of PYP and its photoproduct were studied by
Fourier transform infrared spectroscopy at 40 °C. The vibrational modes characteristic for the anionic p-coumaryl chromophore
(Kim, M., Mathies, R. A., Hoff, W. D., and Hellingwerf, K. J. (1995) Biochemistry 34, 12669-12672) were observed at 1482, 1437, and 1163 cm
1 for PYP. However, the bands corresponding to
these modes were not observed for PYPM, the blue-shifted
intermediate, but the 1175 cm
1 band characteristic of the
neutral p-coumaryl chromophore was observed, indicating
that the phenolic oxygen of the chromophore is protonated in
PYPM. A 1736 cm
1 band was observed for PYP,
but the corresponding band for PYPM was not. Because it
disappeared in the Glu-46
Gln mutant of PYP, this band was assigned
to the C=O stretching mode of the COOH group of Glu-46. These results
strongly suggest that the proton at Glu-46 is transferred to the
chromophore during the photoconversion from PYP to
PYPM.
Photoactive yellow protein (PYP)1
(max = 446 nm) (1) is proposed to be a photoreceptor
protein for the negative phototaxis observed in the purple phototrophic
bacterium, Ectothiorhodospira halophila (2). PYP belongs to
the novel group of photoreceptor proteins (3, 4) whose structures are
quite different from those of the other photoreceptor proteins studied
so far. Namely, the protein moiety of PYP has an
/
fold structure (5) composed of 125 amino
acids (6, 7). The chromophore is a p-coumaric acid (7-9)
bound to a cysteine residue via a thioester bond.
PYP absorbs a photon and enters the photocycle. We have analyzed the
photocycle of PYP in detail by low temperature spectroscopy and
identified several intermediates (10). Irradiation of PYP at
190 °C yields PYPB (
max = 489 nm) and
PYPH (
max = 442 nm), which are thermally
converted to PYPL (
max = 456 nm) through PYPBL (
max = 400 nm) and PYPHL
(
max = 447 nm), respectively. The two pathways beginning
with PYPB and PYPH join at PYPL and revert to PYP. Flash photolysis at ambient temperature identified two
intermediates, pR (
max = 465 nm) and pB
(
max = 355 nm) (11, 12). pR is formed within 10 ns after
flash excitation. It decays to pB over a submillisecond time scale and
reverts to PYP within 1 s. It has been demonstrated that pR is the
same species as PYPL (10) (In this paper, pR and pB are
called PYPL and PYPM, respectively, to avoid
confusion) and that PYPL is accumulated by irradiation of
PYP at
80 °C (10, 13). However, the precursors of PYPL have not been discovered by flash photolysis at room temperature.
Recent studies have clarified some details of the events that take place during the PYP photocycle. Crystallography at 1.4-Å resolution (7) and resonance Raman spectroscopy (14) have shown that the phenolic oxygen of the chromophore of PYP is deprotonated in the dark state. The photocycle is initiated by photon absorption, which involves isomerization of the ethylenic bond of the chromophore (15). During the photocycle, observed proton uptake and release correspond with the formation and decay of PYPM (16).
The largely blue-shifted absorption spectrum of PYPM
suggests that the chromophore/protein interaction of PYPM
is quite different from that of PYP. Therefore, elucidation of the
chromophore structure of PYPM is essential to understand
the photocycle of PYP. For the spectral blue-shift of PYPM,
the following explanations would be possible: (i) the phenolic oxygen
of the chromophore is protonated; (ii) the conjugated double bond
system is broken by extreme distortion of the chromophore (25); and
(iii) the electrons of the conjugated double bond system are
localized by a nearby positive charge (e.g. Arg-52) (26).
The first explanation is the most simple and plausible, but there has
been no previous experimental evidence to support it, and the others
could not be excluded.
In the dark state, the phenolic oxygen of the chromophore interacts with the OH groups of Tyr-42 and Glu-46 by hydrogen bonds (5). As shown in retinal protein systems, the changes in the hydrogen-bonding network centering on the chromophore closely correlates with the protein conformational changes (17). Therefore, it is of importance to study the internal proton movement around the chromophore and nearby amino acid residues to understand the light signal transduction mechanism of PYP. Recently it has been reported that the C=O stretching mode of the COOH group disappears on the formation of PYPM (18). They reasoned that Glu-46 donates a proton to the chromophore, based on the fact that Glu-46 has a unique COOH group embedded in the protein. This idea arose on elucidation of the tertiary structure of PYP (5) and seems reasonable. However, two vital pieces of experimental evidence required to support this conclusion have never been reported: one is that the chromophore of PYPM is protonated, and the other is assignment of the C=O stretching mode.
Recent progress in experimental techniques made it possible to prepare PYP with an isotope-labeled chromophore (9) and a site-directed mutant of PYP (19, 27). These techniques have enabled assignment of the vibrational modes. In the present study, to study the movement of the proton around the chromophore and nearby amino acid residues, the vibrational modes of the chromophore and COOH group of PYP and its intermediates were analyzed by low temperature FTIR spectroscopy. The first experimental findings indicating proton transfer from Glu-46 to the chromophore during the photocycle are presented.
PYP was isolated from E. halophila BN 9626 according
to previously reported methods (1, 10). The Glu-46 Gln mutant of
PYP (E46Q) was expressed by Escherichia coli and
reconstituted with p-coumaric anhydride (19).
p-Coumaric-8-13C-acid was prepared by
p-hydroxybenzaldehyde and
triethylphosphonoacetate-2-13C followed by alkaline
hydrolysis of the ester. 13C-Labeled PYP was prepared by
reconstitution of PYP with 13C-labeled
p-coumaric anhydride and apoPYP (9). They were then desalted
by dialysis and applied to a small DEAE-Sepharose column (Pharmacia
Biotech Inc.). After washing the column with 10 mM phosphate buffer (pH 7.2), PYP was eluted with a linear gradient of
NaCl (100-200 mM) in the same buffer. PYP was then
concentrated with an ultrafiltration membrane (Centriprep 10, Amicon)
and diluted with 10 mM phosphate buffer. Dilution and
concentration steps were repeated several times to remove NaCl.
Finally, PYP was concentrated to 3-4 mg/ml.
A 20 µl sample was placed on a BaF2 window (10 mm in diameter) and dried under a gentle stream of N2 gas. The dried sample was sealed using a silicon rubber spacer and another BaF2 window and set in the sample cell holder. Before sealing, 0.2 µl of H2O or D2O was put inside the spacer for hydration or deuteration of PYP. The sample cell holder was mounted in an optical cryostat (DN1704, Oxford) connected to a temperature controller (ITC502, Oxford).
Absorption spectra in the UV-visible region were recorded with a
Hitachi U-3210 recording spectrophotometer. Infrared spectra were
recorded with a Horiba FT-210 Fourier transform infrared spectrophotometer equipped with an MCT detector. The sample was irradiated with a 1 kW slide projector (HILUX-HR, Tokyo Master) using
glass optical filters (Y47, Toshiba). The difference FTIR spectra shown
in this paper are the means of four to eight independent recordings,
each of which was the mean of 64 scans (resolution = 2 cm1).
Our recent low temperature spectroscopy results with PYP in 66%
glycerol buffer showed that PYPL is formed by the
irradiation of PYP at 80 °C, but PYPM was not observed
in the thermal reaction (10). This could be due to the effect of the
presence of glycerol at high concentration, because it suppresses the
decay of PYPL but accelerates the decay of PYPM
(20). In the method used here, the film sample can be frozen with no
increase in turbidity, and the addition of glycerol was not necessary
for the low temperature experiments, thus removing impediments to the
formation of PYPM. We tested several irradiation conditions
above
80 °C to accumulate PYPM without
PYPL for FTIR spectroscopy. Upon irradiation of PYP film
with >450-nm light at
40 °C, the absorbance at 350 nm was increased (Fig. 1). Because this product had largely
blue-shifted absorption spectrum similar to pB (Fig. 1,
inset) and directly reverted to PYP (data not shown), it
would correspond to pB (called PYPM hereafter). In this
difference spectrum, the absorbance at 460 nm was not increased,
indicating that only PYPM was formed. Under this
irradiation condition, the difference FTIR spectra were recorded with
hydrated and deuterated PYP films. The difference spectra before and
after irradiation with >450-nm light at
40 °C are the difference
between PYPM and PYP (PYPM/PYP spectrum), in which positive and negative peaks are attributed to PYPM
and PYP, respectively (Fig. 2). The prominent bands of
PYP were observed at 1736, 1560, 1482, 1437, 1301, 1163, 1058, 1041, 983, and 831 cm
1. On the other hand, the intensities of
the bands of PYPM were relatively small, but 1175, 1081, and 994 cm
1 bands were characteristic for
PYPM. Several bands of them were affected by
D2O substitution.
Recent resonance Raman spectroscopy covering the range 1750-1000
cm1 (14) was helpful for the implication of our FTIR
data. It was reported that the prominent vibrational modes of PYP were
observed at 1633, 1558, 1498, 1439, 1288, 1165, 1059, and 1043 cm
1 in H2O buffer (14) and were shifted 0-4
cm
1 by D2O substitution. It was concluded
that the chromophore of PYP is deprotonated; otherwise the chromophore
bands would be D2O-sensitive. It was further pointed out
that the deprotonated chromophore shows bands at 1498, 1439 and 1165 cm
1 but when protonated there is only one band at 1176 cm
1 (14). The PYPM/PYP spectrum reported here
shows corresponding bands: negative at 1482, 1437, and 1163 cm
1 and positive at 1175 cm
1. The bands at
1482, 1437, and 1163 cm
1 were observed only on the
negative side, indicating that PYPM does not have
vibrational modes corresponding to these. Moreover, the fact that the
1175 cm
1 band was strongly affected by D2O
substitution suggests that the chromophore of PYPM is
protonated. To confirm that the 1175 cm
1 band of
PYPM is the vibrational mode of the chromophore,
PYPM/PYP spectrum was recorded using PYP containing
13C-labeled chromophore at 8 position under the same
irradiation condition (>450-nm light,
40 °C) (Fig.
3a). The shift of 1175 cm
1 band
with 13C-label (Fig. 3a) was similar to that
with D2O substitution (Fig. 2). Therefore, this band is
attributed to the chromophore, which proves that the chromophore of
PYPM is protonated.
The negative band at 1736 cm1 should be noted because it
could be the C=O stretching mode of the COOH group of either aspartic acid or glutamic acid. A similar band was found recently by FTIR spectroscopy (18), although the irradiation condition was considerably different from that of the present experiment. The appearance of the
negative band without the complementary positive band indicates deprotonation of the COOH group. Among 19 acidic residues of PYP (7),
the most plausible candidate for the origin of this band would be
Glu-46 because its OH group directly interacts with the phenolic oxygen
of the chromophore and it is buried inside the protein. To test this, a
Glu-46
Gln mutant (E46Q) of PYP was prepared (19), and its
vibrational modes were studied by FTIR spectroscopy.
Whereas the absorption maximum of E46Q in the visible region is
slightly red-shifted from wild type PYP (19), the intermediate corresponding to PYPM accumulated under the same
irradiation conditions (>450-nm light, 40 °C) (data not shown).
FTIR spectroscopy was then carried out using E46Q, and a
PYPM/PYP spectrum was obtained (Fig. 3b). As
expected, the PYPM/PYP spectrum of E46Q did not show a
1736-cm
1 band, indicating that the C=O stretching mode of
the COOH group of Glu-46 is indeed the origin of this band. Namely,
Glu-46 is protonated in PYP and releases its proton in the
PYPM state.
Based on our present observations, the model for proton movement
during the PYP photocycle is shown in Fig. 4. On
absorption of a photon, the chromophore of PYP isomerizes to the
cis form (15). There are two possible models for the
isomerization of the chromophore: one is that the phenol part rotates,
and the other is that the ester part rotates. Xie et al.
(18) proposed that the hydrogen bond between the OH group of Glu-46 and
the phenolic oxygen of the chromophore remains in PYPL and
supported the latter model. However, they irradiated PYP at 80 K to
trap PYPL. At 80 K, irradiation yields a mixture of
PYPB and PYPH (10, 21), so the changes in C=O
stretching mode of PYP observed at 80 K cannot be attributed to
PYPL. Moreover, the intermediate formed readily absorbs a
photon, and it is uncertain whether or not the photoproduct accumulated
at 80 K has the cis chromophore. Therefore, few data are
available to discuss the mechanism of the isomerization of the
chromophore. However, the latter model is likely because the phenolic
oxygen interacts with Glu-46 and Tyr-42, and it is fixed by the
hydrogen bond. In the case of retinal proteins, the protein moiety has
the binding site for the -ionone ring of the chromophore,
and the Schiff base portion (rather than the
-ionone ring
region) rotates on photoisomerization (22). That is, in both PYP and
retinal protein systems, the smaller moiety rotates on isomerization.
Minimization of the conformational change upon isomerization would
contribute to the fast and highly efficient photoreaction of
photoreceptor proteins.
The present experiments have demonstrated the first evidence that the phenolic oxygen of the chromophore is protonated in PYPM. The absorption maximum of PYPM is 355 nm (12), which is close to that of denatured PYP at neutral pH (11, 23). Because the phenolic oxygen of PYP is protonated, the largely blue-shifted absorption spectrum of PYPM is due mainly to protonation of the chromophore. However, the possibility that steric interactions and the nearby positive charge affect the absorption spectrum of PYPM is not completely excluded. These effects will be examined in future research.
In retinal proteins, the chromophore of the dark state is protonated at the nitrogen atom of the Schiff base linkage and positively charged. It is deprotonated and neutral in the near-UV intermediates like M and metarhodopsin II. In contrast, the p-coumaryl chromophore is negatively charged in PYP but is neutral in PYPM. Despite the striking differences in the chromophores, protein moieties, and linkages, the chromophores of both retinal proteins and PYP turn neutral in their near-UV intermediates, which are thought to be formed as the result of large conformational changes (17, 28). In addition, cis/trans isomerization of the chromophore is involved in both systems (15). These are the common events for the protein conformational changes among the photoreceptor proteins and would be the most efficient mechanism. Namely, the charged chromophore is the core of the hydrogen-bonding network that determines its tertiary structure in the dark state. Isomerization of the chromophore on photon absorption results in the loss of its charge. As a result, the hydrogen-bonding network is disordered, and the protein conformational change takes place. In PYP, the phenolic oxygen of the chromophore interacts with Tyr-42 and Glu-46, and the chromophore binding site is Cys-69. The loss of this interaction would make the structure of the chromophore binding site remarkably flexible.
The present work confirmed that the 1736 cm1 band of PYP
is the C=O stretching mode of the COOH group of Glu-46. Absence of this
mode in PYPM indicates that Glu-46 is protonated in PYP but deprotonated in PYPM. Concurrently, the chromophore of
PYPM is protonated. This indicates that the
pKa values of the chromophore and Glu-46 of
PYPM are close to those in solution but that those of PYP
are considerably different due to the hydrophobic environment in the
protein. Therefore, the chromophore binding pocket would be exposed to
the solvent in PYPM as the result of the protein
conformational changes.
Because the protonation states of the chromophore and Glu-46 are complementary, the simplest model for proton pathways would be that the proton at Glu-46 is transferred to the chromophore in PYPM state, recovering to Glu-46 in the recovery step (Fig. 4). In E46Q, the chromophore would uptake proton from the solvent or Tyr-42 would act as the alternative proton donor. This model does not explain the proton release and uptake coupled with the formation and decay of PYPM. Therefore, another amino acid residue and/or water might be involved in the proton transfer, resulting in vectorial movement of the proton, as with bacteriorhodopsin (24). Further FTIR studies on the other intermediates using various site-directed mutants will elucidate the pathway of proton transfer.
We thank Dr. Keizo Shimada of Tokyo Metropolitan University for the kind gift of E. halophila strain BN 9626. We also thank Dr. Ian Gleadall for comments on the manuscript and Yuji Shirahige for technical assistance.