Photoactivation perturbs the membrane-embedded contacts between sensory rhodopsin II and its transducer.

The photoactivation mechanism of the sensory rho-dopsin II (SRII)-HtrII receptor-transducer complex of Natronomonas pharaonis was investigated by time-resolved Fourier transform infrared difference spectroscopy to identify structural changes associated with early events in the signal relay mechanism from the receptor to the transducer. Several prominent bands in the wild-type SRII-HtrII spectra are affected by amino acid substitutions at the receptor Tyr(199) and transducer Asn(74) residues, which form a hydrogen bond between the two proteins near the middle of the bilayer. Our results indicate disappearance of this hydrogen bond in the M and O photointermediates, the likely signaling states of the complex. This event represents one of the largest light-induced alterations in the binding contacts between the receptor and transducer. The vibrational frequency changes suggest that Asn(74) and Tyr(199) form other stronger hydrogen bonds in the M state. The light-induced disruption of the Tyr(199)-Asn(74) bond also occurs when the Schiff base counterion Asp(75) is replaced with a neutral asparagine. We compared the decrease in intensity of difference bands assigned to the Tyr(199)-Asn(74) pair and to chromophore and protein groups of the receptor at various time points during the recovery of the initial state. All difference bands exhibit similar decay kinetics indicating that reformation of the Tyr(199)-Asn(74) hydrogen bond occurs concomitantly with the decay of the M and O photointermediates. This work demonstrates that the signal relay from SRII to HtrII involves early structural alterations in the deeply membrane-embedded domain of the complex and provides a spectroscopic signal useful for correlation with the downstream events in signal transduction.

tight intermolecular complex with its cognate transducer protein HtrII in the cell membrane. The transducer, like its homologous methyl-accepting chemotaxis transducers, possesses two transmembrane helices and a large cytoplasmic domain that binds at its distal end a His-kinase that phosphorylates a flagellar motor switch regulator (1,9).
Interactions of HtrII with the SRII receptor are localized to the transducer transmembrane and membrane-proximal domains (10,11). A crystal structure of SRII bound to an Nterminal HtrII fragment containing the transmembrane helices (TM1 and TM2) shows tight van der Waals interaction and three hydrogen bonds between TM2 and SRII helices F and G (12). An atomic structure of the membrane-proximal domain of the transducer is not available; however, fluorescent probe accessibility and Förster resonance energy transfer measurements show interaction of this domain with the cytoplasmic E-F loop of the receptor (13).
The signal relay mechanism from SRII to HtrII in the complex has become a focus of interest in the past several years, in part because of its importance to the general understanding of interaction between integral membrane proteins. In accord with the unified model for transport and signaling by microbial rhodopsins (14,15), the key event in the transducer activation is an outward tilt of the receptor helix F during the lifetime of its M intermediate, which has been directly detected by EPR of paramagnetic probes in the free receptor (16) and in its complex with transducer (17). In a response to the helix F movements, the TM2 of HtrII is displaced from its initial position (17). Additional structural changes were observed in the cytoplasmic membrane proximal region by fluorescent probe accessibility measurements (13).
Fourier transform infrared (FTIR) difference spectroscopy has been used extensively in the past to elucidate structural changes in the photocycles of several microbial rhodopsins (18 -27). This technique measures changes in the infrared absorption of protein groups and enables studies of light-induced conformational changes without the necessity of introducing potentially structure-perturbing probes. Because of its sensitivity to small changes in hydrogen bonding, it is especially well suited for the study of interactions of polar protein groups. Here we have applied time-resolved FTIR difference spectroscopy to examine the changes in interaction between SRII Tyr 199 and HtrII Asn 74 during the light-activated photocycle. These residues form a hydrogen bond in the dark state that functions as an interhelical bridge between helix G of SRII and TM2 of HtrII embedded near the middle of the membrane interior (12). Assessing the environment of this hydrogen bond would answer whether the movements of helix F and TM2 previously detected near the cytoplasmic region extend more deeply into the membrane interior domain of the complex.

MATERIALS AND METHODS
Plasmid Construction and Site-directed Mutagenesis-The wild-type construct FP120 encoded a fusion protein in which full-length SRII and the 120 N-terminal residues of HtrII containing an additional His 6 tag at the C terminus are joined by the flexible linker ASASNGASA. The fusion gene was placed in plasmid pET21d (Novagen) under control of the T7 promoter. The plasmid was transformed into the Escherichia coli BL21 (DE3) strain. Single amino acid substitutions were performed using QuikChange II site-directed mutagenesis kit (Stratagene). The mutagenesis primer sequences were 5Ј-GGCGCTTATCGTCTTCCTTG-ACCTCGTCAC-3Ј and 5Ј-GTGACGAGGTCAAGGAAGACGATAAGCG-CC-3Ј for the Tyr 199 3 Phe mutation, 5Ј-CCTGCTCGGGATCGACCT-CGGGCTCGTTGC-3Ј and 5Ј-GCAACGAGCCCGAGGTCGATCCCGA-GCAGG-3Ј for Asn 74 3 Asp mutation, and 5Ј-GTCCCCCGGTACATC-AACTGGATTCTCACAACC-3Ј and 5Ј-GGTTGTGAGAATCCAGTTGA-TGTACCGGGGGAC-3Ј for the Asp 75 3 Asn mutation. The double mutant D75N/Y199F was constructed by two-step mutagenesis.
Protein Expression and Purification-The cells were grown in LB medium ϩ ampicillin, 50 g/ml, to an absorbance at 600 nm of ϳ0.4, and the protein synthesis was induced by addition of 1 mM isopropyl ␤-D-thiogalactopyranoside and 5 M all-trans-retinal. After the induction period, the cells were centrifuged at 1000 ϫ g, resuspended in 50 mM Tris/HCl, pH 7.0, 5 mM MgCl 2 , 1 mM phenylmethylsulfonyl fluoride buffer and disrupted by a microfluidizer (Microfluidics Corp., Newton, MA), and the membranes were then harvested by ultracentrifugation. The membranes were solubilized in 300 mM NaCl, 10 mM imidazole, 50 mM potassium phosphate, pH 7.6, 1.5% octyl glucoside. After centrifugation of the solubilized membranes, the supernatant was incubated with nickel-nitrilotriacetic acid-agarose (Qiagen), and the His-tagged protein was eluted with a gradient of imidazole in 50 mM potassium phosphate, pH 7.6, 300 mM NaCl, 0.8% octyl glucoside in a Biologic Duoflow system (Bio-Rad).
Proteoliposome Reconstitution-Purified proteins were reconstituted in phospholipids by the dialysis procedure previously described (28). In the work reported here the protein-to-lipid ratio was 1:6.5 (w/w). Extraction of halobacterial phospholipids used in the reconstitution procedures was performed as described previously (22).
FTIR Difference Spectroscopy-The sample preparation and FTIR measurement procedures have been described previously (27). Rapidscan time-resolved FTIR spectroscopy was performed using a BRUKER IFS 66 v/s FTIR spectrometer (Bruker Optics, Billerica, MA) operating at 4 cm Ϫ1 spectral resolution and 240 kHz scanner velocity corresponding to a data acquisition window of 18 ms.

Spectral Changes in the SRII-HtrII Complex Due to
Tyr 199 3 Phe and Asn 74 3 Asp Substitutions-Difference spectra of the receptor-transducer complex were recorded using rapid scan FTIR difference spectroscopy and averaged over a 30 -85-ms period after photoexcitation ( Fig. 1). This time range is associated predominantly with accumulation of the M intermediate (24), which is a signaling state for the transducer activation (29). Several prominent bands appear at 1163 (Ϫ), 1200 (Ϫ), 1240 (Ϫ), 1544 (Ϫ), and 1567 (ϩ) cm Ϫ1 in the spectra of wild-type SRII-HtrII complex (top trace), which have been previously assigned to the major vibrations of the retinylidene chromophore in the all-trans (negative bands) and 13-cis (positive bands) states as the receptor undergoes a transition from the dark state to the M intermediate (24,30). The presence of M is also evident from the strong positive band near 1764 cm Ϫ1 due to protonation of the Schiff base counterion Asp 75 and prominent bands in the amide I region at 1644 (ϩ) and 1664 (Ϫ) cm Ϫ1 (24,25). The higher intensity of amide I peaks compared with previously reported spectra (27) may result from a shorter length (120 residues compared with 157) of the transducer segment used in this study.
The Tyr 199 3 Phe and Asn 74 3 Asp substitutions (Fig. 1, middle and bottom traces, respectively) do not affect the major chromophore or protein difference bands. These results indicate that the structural changes that the chromophore and receptor undergo during the transition from the dark to the M photointermediate states are very similar for the wild-type and the mutants. On the other hand, changes are observed in spectral regions associated with tyrosine and asparagine side chain vibrations.
(i) A negative shoulder near 1265 cm Ϫ1 is present in wildtype SRII-HtrII but not in the mutants (Fig. 2). This effect is similar to the difference observed previously between the spectra of wild-type SRII-HtrII complex and free receptor (27). In the spectra of tyrosine model compounds, a characteristic band appears between 1255 and 1275 cm Ϫ1 due to the C-O stretching mode 7Јa (CO) of a tyrosine side chain (31), whose frequency is sensitive to the hydrogen-bonding environment (32). On this basis we assign the 1265 cm Ϫ1 (Ϫ) negative feature to Tyr 199 in the dark state. A careful examination of this region also reveals a decrease in intensity at higher frequency near 1275 cm Ϫ1 induced by the Y199F substitution. This indicates that a positive band near this frequency also arises from vibrations of Tyr 199 in the M intermediate of wild-type pigment. The inferred 1265/1275 (Ϫ/ϩ) cm Ϫ1 bands are clearly seen in the double difference spectra obtained by subtraction of the Y199F mutant from the wild-type spectra. The apparent frequency upshift of the Tyr 199 vibration upon photoactivation indicates that a change in the hydrogen bonding strength of Tyr 199 occurs during formation of M (see "Discussion").
(ii) A prominent negative band near 1518 cm Ϫ1 has reduced intensity in N74D and disappears in the Y199F spectrum. A very strong ring mode vibration, which is sensitive to the protonation state of tyrosine, appears at 1518 cm Ϫ1 in the spectra of tyrosine recorded in HCl and near 1480 cm Ϫ1 in the spectra of tyrosinate recorded in NaOH (32). From the frequency of this vibration, we conclude that Tyr 199 is protonated in the receptor dark state. Note that the position of this band in the difference spectra may be affected by spectral overlap with other peaks in this region.
(iii) A negative peak seen in the double difference spectrum (Fig. 2) near 1694 cm Ϫ1 was previously identified in the SRII-HtrII complex but not in the receptor alone (27). It is also absent in both mutant spectra. As noted previously (27), its frequency is typical for the CϭO stretching mode of an asparagine side chain. This vibration is found near 1678 cm Ϫ1 in the model compound spectra and depending on the hydrogen bonding environment can shift significantly, with the lower frequency corresponding to a stronger hydrogen bond (33). We therefore assign the 1694 cm Ϫ1 peak to the transducer Asn 74 . The frequency indicates that this group is in a moderately hydrophilic environment in the SRII-HtrII resting state. A search for a positive counterpart of the 1694 cm Ϫ1 peak revealed a band near 1670 cm Ϫ1 in the double difference spectrum, thereby indicating formation of a stronger hydrogen bond.
(iv) A negative peak appears at 1732 cm Ϫ1 in the N74D spectrum but is not present in WT or Y199F spectra. The CϭO vibrations of protonated aspartic acid side chains appear in the 1730 -1765 cm Ϫ1 range and exhibit strong dependence on the hydrogen-bonding environment (33). We assign the 1732 cm Ϫ1 peak to the CϭO stretching vibration of the carboxylate group introduced in the Asn 74 3 Asp mutant. No positive bands due to Asp 74 are found in this region, which would be indicative of a change in hydrogen bonding of this group during the formation of M. However we observed spectral changes near 1380 (ϩ) and 1595 (ϩ) cm Ϫ1 in the double difference spectra between the wild-type and N74D but not Y199F (data not shown), which are typical for deprotonated carboxylate groups (31). These results suggest that Asp 74 undergoes at least a partial deprotonation during the light activation.
In contrast to the SRII-HtrII complex, the spectrum of the transducer-free receptor does not exhibit peaks near 1265, 1518, and 1694 cm Ϫ1 (27). The absence of tyrosine difference bands indicates that Tyr 199 in the transducer-free receptor exists in a very similar environment in the dark and photointermediate states. The lack of the 1694 cm Ϫ1 peak is also in agreement with its assignment to the transducer. Furthermore, no significant changes were detected between the wildtype and Y199F SRII spectra (data not shown).
We also note that in contrast to the recently reported low temperature FTIR measurements of the SRII-HtrII complex  (36). Despite the photocycle differences, a comparison of the D75N and D75N/Y199F samples (Fig. 3, bottom traces) reveals that the spectral changes, which we have assigned to disruption of the interaction between Tyr 199 and Asn 74 in the wild-type complex (see above), also appear in the D75N/Y199F mutant. Note, however, that the negative band assigned to Asn 74 at 1694 cm Ϫ1 in the wild-type complex is at a lower frequency near 1688 cm Ϫ1 indicating an altered hydrogen bonding environment due to the D75N mutation. This peak is superimposed on a pair of intense positive/negative bands at 1700/1685 cm Ϫ1 assigned to Asn 75 in the D75N mutant receptor (36).
The Kinetics of Difference Bands Arising from the Tyr 199 / Asn 74 Interaction and the Receptor during the Recovery of the Initial State-To test whether the return of HtrII to the initial position occurs with different kinetics than the decay of the SRII signaling state, we compared changes in the amplitude of individual difference bands measured between 50 and 250 ms (Fig.  4). In this time window all peaks exhibit a gradual decrease in intensity reflecting the return of complex to the unphotolyzed state. The majority of peaks including the negative retinal vibrational modes at 1200, 1240, and 1545 cm Ϫ1 and the carboxylic stretching mode at 1764 cm Ϫ1 exhibit an exponential decay (data not shown) with a time decay constant similar to that previously reported (38). Because there is no noticeable delay between the decay of these receptor vibrations and the 1694 cm Ϫ1 band assigned to Asn 74 from the transducer, we conclude that the recovery of the Tyr 199 -Asn 74 hydrogen bonding occurs at a similar rate as SRII returns to the unphotolyzed state. DISCUSSION In an earlier study, we compared the structural changes occurring in the transducer-free receptor SRII and a SRII-HtrII fusion complex using time-resolved FTIR difference spectroscopy (27). Several spectral differences were found between the two systems including bands characteristic of asparagine and tyrosine vibrations. In the present study, we focused on the interaction between Tyr 199 and Asn 74 that is observed in the crystal structure of SRII-HtrII in its dark state (12). These two residues form a hydrogen bond that functions as an interhelical pin between helix G of SRII (Tyr 199 ) and TM2 of HtrII (Asn 74 ) embedded near the middle of the membrane interior (12). As such, changes in the vibrational spectra of Tyr 199 and Asn 74 represent a native probe of possible movements that involve these two helices in the middle of the membrane.
To study the structural changes that specifically involve these two residues we utilized site-directed mutagenesis in combination with time-resolved FTIR. The SRII-HtrII fusion construct used here included the full-length receptor and the first 120 amino acids of the transducer, which constitute the transmembrane and membrane-proximal domains of HtrII. The use of a truncated transducer allowed us to focus on the mechanism of signal transduction from SRII to HtrII while excluding downstream events associated with signal propagation within the transducer.
The results of this work show that the major spectral differences observed between the truncated SRII-HtrII complex and free receptor (27) can be ascribed to structural changes of the receptor Tyr 199 and transducer Asn 74 groups. Considering the sensitivity of infrared spectroscopy to detect very small alterations in the hydrogen bonding of polar groups (39 -41), it is likely that perturbations of these residues represent one of the most significant changes in the hydrogen-bonding structure of the receptor-transducer complex. These results, however, do not exclude the possibility of additional structural changes in the complex. One example is the displacements of transducer helices TM1 and TM2 in response to the receptor conformational changes. These movements are expected to give rise to difference bands in the amide I and amide II spectral regions. In fact, several amide I bands in the receptor spectra are altered in the presence of transducer (27). The detailed analysis of vibrations in this region is a subject of a future study.
Our finding that the replacement of either Tyr 199 or Asn 74 alters light-induced negative bands arising from both of these groups establishes that they interact in proteoliposomes in the SRII-HtrII dark state, in agreement with the crystallographic structure ( Fig. 5) (12). The frequencies of negative bands arising from the side chain vibrations of Tyr 199 and Asn 74 are also consistent with the presence of a hydrogen bond between them. The x-ray diffraction data does not resolve unambiguously whether the hydroxyl group of Tyr 199 is hydrogen-bonded to the oxygen or amide nitrogen of Asn 74 side chain (12,34). The frequency of the Asn 74 peak (1694 cm Ϫ1 ) detected in this study is considerably higher compared with the asparagine model compounds (1677 cm Ϫ1 ) indicating that the side chain carbonyl group is only weakly hydrogen-bonded and therefore the interaction with the Tyr 199 hydroxyl group is more likely through the amide nitrogen.
The FTIR results provide insight regarding the environment of Tyr 199 and Asn 74 in the light-activated state. The frequency downshift from 1694 to 1670 cm Ϫ1 for Asn 74 points to the formation of a considerably stronger hydrogen bond in the photointermediate. This shift is larger than normally associated with environmental changes of hydrogen-bonded asparagine groups. For example, relatively small differences of 6 and 8 cm Ϫ1 were observed between the initial and M states for Asn 105 in SRII (25) and Asn 230 in proteorhodopsin (26), respectively. On the other hand, the data for HtrII are consistent with formation of a strong hydrogen bond involving the CϭO group of Asn. A large upshift of the Tyr 199 vibrational mode initially near 1265 cm Ϫ1 to above 1270 cm Ϫ1 also indicates significant changes in the hydrogen bonding interaction. In particular, a shift of similar amplitude was observed for this mode in the case of Tyrosine D in photosystem II upon the replacement of nearby His 189 with a Gln (32). Therefore, both Tyr 199 and Asn 74 undergo large structural alterations during the photoactivation as a result of the movements of helices F and TM2 (17). The observed FTIR signals in the dark and intermediate states most likely reflect disruption of the Tyr 199 -Asn 74 interaction and formation of new stronger hydrogen bonds. However, we do not exclude the possibility that these residues still interact in the M intermediate. Such interaction may be indirect, for example through a network of hydrogen-bonded groups. Such a weak interaction is in agreement with the observation that the replacement of Tyr 199 with Phe has smaller effect on the transducer affinity for the receptor in the M intermediate compared with the dark state (42).
In conclusion, we demonstrated that FTIR difference spectroscopy in combination with site-directed mutagenesis can be successfully used to probe specific conformational changes that arise from interaction between two membrane proteins. In the case of the SRII-HtrII complex reported here, changes in the transmembrane domain occur during the early steps in trans- ducer activation. The extension of this work including mutations in the cytoplasmic and extracellular protein domains should be able to provide a more comprehensive picture of the SRII-HtrII interaction and ultimately lead to the identification of the specific protein residues responsible for the transfer of the phototaxis signal from the receptor to the transducer. In addition to the peak assignment, the site-directed mutagenesis can be also used to correlate protein structural changes with the cell phototaxis response assessed from functional studies. An example of such an approach is demonstrated in the case of the receptor D75N mutant. The difference spectra of the mutant complex revealed the presence of a Tyr 199 -Asn 74 interaction in the dark state and its perturbation upon photoactivation despite the absence of the Schiff base counterion. The similar conformational changes between the wild-type and mutant complex in the transmembrane region suggested by these results are in agreement with the ability of D75N SRII to activate the transducer.
Comparison of the kinetic spectra recorded during the last photocycle transition reveals that the transducer return to the initial state, measured by disappearance of the Asn 74 signal, occurs concomitantly with the decay of the photoactivated receptor. This result differs from the observation of Wegener et al. (17) of slower kinetics for the recovery of paramagnetic label attached to the transducer residue Val 78 . The difference may be due to several factors including different measurement conditions, probing different positions in the transducer and possible additional structural changes resulting from the introduced label in the case of EPR.