Photochemistry of a Dual-Bacteriorhodopsin System in Haloarcula

Jun 18, 2014 - 4, Roosevelt Road,. Taipei 10617, Taiwan. §. Institute of Biotechnology, College of Bio-Resources and Agriculture, National Taiwan Uni...
0 downloads 0 Views 9MB Size
Download PDF
Article pubs.acs.org/JPCB

Photochemistry of a Dual-Bacteriorhodopsin System in Haloarcula marismortui: HmbRI and HmbRII Fu-Kuo Tsai,†,# Hsu-Yuan Fu,‡,# Chii-Shen Yang,‡,§ and Li-Kang Chu*,† †

Department of Chemistry, National Tsing Hua University, 101, Sec. 2, Kuang-Fu Road, Hsinchu 30013, Taiwan Department of Biochemical Science and Technology, College of Life Science, National Taiwan University, 1, Sec. 4, Roosevelt Road, Taipei 10617, Taiwan § Institute of Biotechnology, College of Bio-Resources and Agriculture, National Taiwan University, 1, Sec. 4, Roosevelt Road, Taipei 10617, Taiwan ‡

S Supporting Information *

ABSTRACT: Recently, a dual-bacteriorhodopsin system, containing HmbRI and HmbRII, has been found in Haloarcula marismortui (Mol. Microbiol. 2013, 88, 551−561), and the light-driven proton pump activities were intrinsically different in a wide pH range. Compared with bacteriorhodopsin in H. salinarum (HsbR), the identical steady-state absorption contours of HsbR and HmbRs in the visible range indicated similarities in the retinal pocket. In addition, other reactive residues, including the proton relay channel, proton release group, and proton collecting funnel at the cytoplasm, were mostly conserved. We employed transient absorption spectroscopy and global analysis to characterize the photocycle intermediates and kinetics of HmbRI and HmbRII in the pH range of 4−8. The features of the time-resolved difference spectra of HmbRI indicated that the photocycle of HmbRI mainly followed the conventional pathway, including intermediates M, N, and O. A minute bypassed pathway from intermediate M needed to be included to better match the experimental data. The corresponding intermediate M′ is attributed to the all-trans deprotonated Schiff base retinal, indicating the occurrence of retinal reisomerization prior to the reprotonation of the deprotonated Schiff base following the decay of intermediate M. Regarding HmbRII, its photocycle only followed the intermediates M and N, without intermediate O. The plausible molecular mechanisms, including the effects of the lengths of the loops and the distribution of the charged residues in the bacterio-opsin interior, were proposed to explain the differences in the photocycles. The pH-dependent photocycles were also investigated, and the results supported our proposed mechanism. Unravelling the photocycles of the HmbRs in the Haloarcula marismortui provided evidence that not only expanding the functional pH ranges but also the turnover kinetics are the strategies of the dual-bR system in the evolution of microbes in extreme environments.

1. INTRODUCTION Due to the successful conjugation of neuron cells with photosensitive rhodopsin proteins, such as halorhodopsin,1,2 channelrhodopsin,3 and so on, optogenetics has become a promising method of manually controlling biobehaviors with light.4,5 The various light-driven ions, such as chloride1,2 and sodium,3 serve as the charge carriers of the signaling.6 The first proton pump, bacteriorhodopsin (bR), was found in the halophilic archaea H. salinarum in the 1970s.7 The photoexcitation of bR leads to a transient proton gradient between the cytoplasm and extracellular side via vectorial proton translocation.7 In the intervening decades, many bR-like proton pump proteins from various halophiles, such as cruxrhodopsin,8 archaerhodopsin,9 proteorhodopsin,10 and xanthorhodopsin,11 have been found, and the corresponding photochemistry and biological functions have been extensively investigated. In most cases, the organism is composed of a single photoactive protein that is responsible for single ion-pump activity. Based on a previous study, the known system containing two light-gated cation channels was © XXXX American Chemical Society

only found in unicellular green algae, such as Chlamydomonas reinhardtii, comprising channelrhodopsin-1 (ChR1) and channelrhodopsin-2 (ChR2).12 Turning to microbes, an old archaea, Haloarcula marismortui,13 was investigated recently and was found to contain two proton pump-type bRs at the same time, named HmbRI and HmbRII.14,15 Because the ChR1 and ChR2 had different cation specificities and action spectra, H. marismortui turned out to be the first-known microbe containing two isochromatic light-driven proton pumps simultaneously. H. marismortui is a limited survivor in the Dead Sea,16 but details of its biological and photoinduced features are still lacking. In the Dead Sea, the average pH is around 5.5, but the optimal growth pH of H. salinarum is 7.5.17,18 Previous studies showed the lightdriven proton pumping activity under pH 4−9 in H. marismortui on the basis of both functional assay and retinal-binding pocket Received: April 14, 2014 Revised: June 3, 2014

A

dx.doi.org/10.1021/jp503629v | J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

Table 1. Comparison of the Interactive Pockets of HsbR, HmbRI, and HmbRII HsbR protonated Schiff basea retinal pocketa

proton relay channelb

proton collecting funnel on the cytoplasmic sidec

HmbRI

K216 M20 A53 T89 D115 M118 W138 S141 Y185 P186 D85e D96f D212 R82g E194g E204g D36 D38 D102 D104 E161

HmbRII

K219 M17 A50 T86 D112 M115 W141 S144 Y188 P189 D82 D93 D215 R79 E197 E207 D33 R35 G99 D101 A164

K221 M24 A57 T94 D120 M123 W143 S146 Y190 P191 D90 D101 D217 R87 E199 E209 D40 E42 G107 S109 S166

interior negatively charged residuesd

cytoplasm D38, E166

E42, E44, E45, D112, E171

K30, K40, K41, K172, R225

extracellular E6, D82, E207 middle D215 cytoplasm R27, R35, R36, K38, R102, R171, K175

R82, R134

R79, R137

E9, D85, E204 D212 interior positively charged residues (except for the protonated Schiff base retinal Lysine)d

E13, D82, D90, E209 D217 K34, R111, R230

extracellular R87, R139 middle R175 a

R178

R180

ref 20. bref 28. cref 53. dref 54, 55. eproton acceptor. fproton donor. gproton release group.

property.14,15 However, it was not clear why H. marismortui needs two bRs at the same time because HmbRII itself could deal with the relative acidic condition.14,15 In order to understand the roles of HmbRs, it is of great importance to investigate the photochemical properties of these two light-driven proton pump proteins in H. marismortui. The sequences of HmbRI and HmbRII are highly homologous with HsbR.19 The overall amino acid identity is above 50%, based on ClustalW (Slow/Accurate, Gonnet), and HmbRI has higher similarity to HsbR than HmbRII does. The sequence alignments are shown in Figure S1 in the Supporting Information, and some localized interactive pockets are listed in Table 1. The residues in the pockets of the protonated Schiff base retinal in HmbRI and HmbRII are mostly conserved,20 suggesting similar absorption contours of the protonated Schiff base retinal in the visible region. Moreover, all the residues correlating to proton translocation are conserved, involving Asp85, Asp96, Glu194, and Glu204 denoted in HsbR,21 indicating that both HmbRI and HmbRII could be light-driven proton pumps. The only significant disagreement has been found on the N-terminal, the C-terminal, and both extracellular and cytosolic loops. Those regions are related flexibly and are believed to form not a significant secondary structure but only loop beta-sheets or a conformational change-induced coil structure.

In addition to the structural and biological similarities to HsbRs, the optical modulation of HmbRs at 530 and 410 nm upon pulsed excitation indicate similar photocycle behaviors.14,15 A conventional photocycle of HsbR has been adapted to describe the photoresponse of the retinal Schiff base in the visible region during the photocycle,22−24 hv

0.5ps

3ps

∼ 2μs

50μs

bR 568 → I460 ⎯⎯⎯⎯→ J625 ⎯→ ⎯ K 590 ⎯⎯⎯⎯→ L550 ⎯⎯⎯⎯→ M412(M1 → M 2) 1ms

2ms

8ms

⎯⎯⎯→ N560 ⎯⎯⎯→ O640 ⎯⎯⎯→ bR 568

The positions of the absorption maximum and the periods of the transition are shown in the subscripts and in the superscripts of the arrows, respectively. Upon the excitation of light-adapted HsbR, the all-trans retinal isomerizes to the 13-cis form25−27 and triggers the following steps.21 The proton migration from the protonated Schiff base to the proton acceptor, deprotonated Asp85, results in the generation of intermediate M.28,29 Then the reprotonation of the deprotonated Schiff base, accepting the proton from the proton donor, the protonated Asp96, leads to the generation of intermediate N.30,31 Consequently, the reisomerization of the protonated Schiff base retinal from 13-cis to all-trans,32 coupled with the reprotonation of Asp96,30 gives rise to the generation of intermediate O, in which the all-trans retinal is slightly twisted.32 Structural relaxation B

dx.doi.org/10.1021/jp503629v | J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

optical path length of 1 cm was employed in the measurements aforementioned. 2.3. Transient Absorption in 370−710 nm upon 532 nm Pulsed Excitation. The module of the transient absorption over a wide detection spectral range was composed of one pulse laser, optical dispersion elements, and a data acquisition system. An optical parametric oscillator (basiScan-M/120/HE, Spectra-Physics) pumped by a frequency-tripled Nd:YAG laser (INDI-40-10, Spectra-Physics) provided a 532 nm pulse with a bandwidth of 3 ns. The excitation flux was controlled at 0.3 mJ cm−2 within a standard deviation of energy fluctuation of less than 5%. A mechanical shutter (VS25S2T1, Uniblitz) was employed to decouple the 10 Hz irradiation to 5, 1, 2, and 0.5 Hz to avoid overshooting, depending on the recovery rates of various bRs at different pH upon photoexcitation. The laser pulse was introduced via fiber optics (Model 76842, Newports) to the front of the sample cuvette. A dim broadband beam from the tungsten halogen lamp (DH-2000, Ocean Optics), perpendicular to the excitation laser, was employed for continuous probing, coupled with a monochromator (Model 218, McPherson) to define the optical modulation at 370−710 nm at 10 nm intervals. A 532 nm notch filter (XNF-532.0-25.0M, CVI) was positioned in front of the PMT to reduce the scattering of the excitation laser. The photon signal was collected by a photomultiplier tube (R928, Hamamatsu), and the data were acquired with a 200 MHz oscilloscope (WS24MXs-B, LeCroy). The length of the data acquisition was 100 000 points per waveform. The temporal profiles were averaged from 1500 laser shots. The evolution of the absorbance difference was derived by the equation −ΔA = log(St/S0), where St and S0 represent the dc-coupled voltages in the presence and absence of the excitation laser, respectively. 2.4. Data Analysis. A number of mathematical methods have been proposed to analyze the photocycle of bacteriorhodopsin, mainly singular value decomposition44−46 and global exponential analysis.47,48 In order to obtain the kinetics of the photocycles and the corresponding spectra of the intermediates, a global fitting program, Glotaran, was employed to solve the time-resolved difference spectra,49−51 and the brief description was supplemented in the Supporting Information. A conventional notation of the photocycle of the bacteriorhodopsin analogues, using M, N, and O intermediates, was utilized to construct the kinetics model.22−24

of the skeleton drives the intermediate O back to the original status, coupled with the deprotonation of Asp85. The molecular mechanism has also been thoroughly discussed.21 The photocycle kinetics of HsbR in PM at different pH have been extensively studied.22,33−37 As pH increases from slightly acid to alkaline, the generation of intermediate M is accelerated,38 and its decay is decelerated.39 The following N intermediate decay, in association with the generation of intermediate O, exhibits intrinsic pH-dependent kinetics in the pH range of 3−9, in association with the population alteration.37 In addition to HsbR in the form of PM, monomeric bR, or solubilized bR in surfactant micelles, has also been extensively studied.31,40,41 Although the photocycle kinetics of the solubilized HsbR is slightly different from that in PM, most photocycle features, in terms of intermediates M, N, and O, can be observed in the photocycle.31,41 In this study, we employed transient absorption spectroscopy and global analysis to unveil the photocycle kinetics of HmbRs and the corresponding difference spectra of intermediates. We found that the photocycle of HmbRI is completed through M, N, and O intermediates, coupled with a minute bypassed pathway from the decay of intermediate M, in which the retinal reisomerization takes place prior to the protonation of the retinal Schiff base. In contrast, the photocycle of HmbRII is completed through merely M and N intermediates. These intrinsic differences in the photocycles of HmbRs might result from constituent residues that possess different charge distributions on the cytoplasmic side.

2. MATERIALS AND METHODS 2.1. Materials Preparation. 2.1.1. HmbRI and HmbRII. The heteroexpressed protein sample was prepared as described in a previous study.14 Briefly, corresponding protein expression was induced in transformed Escherichia coli C43(DE3) cells, and the membrane fraction was harvested by ultracentrifugation (100 000g) and solubilized in 50 mM Tris, 4 M NaCl, pH 7.8, 0.2 mM phenylmethylsulfonyl fluoride, 14.7 mM β-mercaptoethanol, 1.5% (w.t./vol) n-dodecyl-β-D-maltoside (DDM) for 16 h at 4 °C. After a purification procedure using Ni-nitrilotriacetic acid (NTA) agarose, the target rhodopsins were then eluted from the column with 50 mM Tris, 4 M NaCl, 0.05% DDM, 250 mM imidazole, pH 7.8. Buffer replacement of target fractions was done through dialysis against the condition in this study. 2.1.2. Monomeric (Solubilized) HsbR. Purple membrane was prepared from Halobacterium salinarum S9 according to a previous method.42 Monomeric bR from H. salinarum, named HsbR, was prepared by solubilizing purple membrane (PM) suspension in Triton X-100 based on the method by Wang et al.43 PM suspension was mixed with Triton X-100 at a ratio 1:7 (w/w) and kept in the dark for more than 48 h. The HsbR monomer in Triton X-100 was further extracted after 10 min of centrifugation at 18 407g and slight sediment was discarded. HsbR monomer solution was prepared in 25 mM phosphate buffer solution at pH ∼7.0, and the UV−vis absorption spectrum maximum accordingly shifted from 568 to 553 nm. 2.2. Steady-State Absorption Spectroscopy. Steady-state spectroscopy was employed to characterize the bR samples before and after the photoexcitation experiments and at different pHs. Before the transient absorption measurements, bRs were light adapted upon white light illumination for 30 min. UV−vis absorption spectra were monitored with a spectrometer (USB4000-UV-NIR, Ocean Optics). A cuvette of

3. RESULTS AND DISCUSSION 3.1. Sequence Analysis. Surveying the sequence alignments of HsbR and HmbRs, as shown in Figure S1 in the Supporting Information, we found several similarities and dissimilarities in different portions. The overall identity was above 50% with respect to HsbR, and HmbRI had higher similarity to HsbR than HmbRII did. We compared the sequences on the basis of structural subunits, and these comparisons are summarized in Table 1. The retinal pocket20 and proton relay channel21,28 are identical in three bRs. However, their charge distributions at the cytoplasmic side52−55 are intrinsically different. In addition, the D−E loop at extracellular side of HmbRI is slightly longer than that of HsbR.56−59 The detailed comparisons were supplemented in the Supporting Information. 3.2. Steady-State Absorption Spectra. The absorption contour of HsbR is mainly composed of intense bands at approximately 560 and 280 nm, which are attributed to the all-trans protonated Schiff base retinal and tryptophans, respectively.60 The normalized UV−vis absorption spectra at pH 5.8 of C

dx.doi.org/10.1021/jp503629v | J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

and HmbRII at different pH are shown in Figures 1b, c, and d, respectively. The retinal band of the monomerized HsbR was characterized at 553 nm,43 and the absorption maxima did not change significantly as the pH varied in the range of 4 to 7, as shown in the inset of Figure 1b. Moreover, the feature of the hydrolyzed62,63 or deprotonated Schiff base retinal at 380 nm was not observed. In contrast, the absorption contours of HmbRs slightly changed as the pH changed. The band maximum of HmbRI at pH 4 and 5.8 peaked at 553 nm and slightly shifted to 550 nm as the pH increased to 8, with a slight increase in the absorbance at 390 nm, plausibly resulting from minute hydrolysis or deprotonation of the retinal Schiff base. Regarding HmbRII, the pH change from 4 to 8 resulted only in a negligible spectral shift of the retinal from 556 to 555 nm, without observable spectral alteration at 380 nm. In addition, no red-shifted contours were found in the pH range of 4−8, indicating that no acid bluelike bR was generated,64 and the pH-dependent photocycle kinetics in the later sections was attributed to the sole species in each bR system. 3.3. Time-Resolved Difference Spectra of HsbR and HmbRs in the 370−710 nm Range. Time-resolved difference spectra were collected in the 370−710 nm range. The signal-to-noise ratio was not completely satisfying because the excitation laser flux was restricted, controlled below 0.3 mJ cm−2, to avoid photobleaching and excitation saturation. The contours of the observed time-resolved difference spectra of HsbR, HmbRI, and HmbRII at different pHs are shown in Figures 2a, b, and c, respectively. Because monomeric HsbR became unstable at higher pH,65,66 the transient absorption measurements of HsbR were performed at pH 4.0, 5.8 and 7.0, at which no photobleaching was observed after photolysis experiments. In order to present the contours in the same time scale, partial data of HmbRs in the prolonged period were truncated, but the data in whole temporal range were included in the global analysis. Although most spectral characteristics were similar in tens of microseconds to tens of milliseconds, the intrinsically different temporal behaviors indicated the unique kinetics and photocycle mechanisms in each bR system. Instead of the experimental data, the comparison of the active residues listed in Table 1 will be employed to differentiate the origins of the intrinsic kinetics behaviors. 3.3.1. HsbR. Before the difference spectra were collected, the laser repetition rates were examined to avoid overshooting. The photocycle of HsbR was believed to be finished within 100 ms.22−24 The temporal profiles of the recovery of HsbR at different pH upon excitation at 532 nm using different repetition rates were shown in Figure S2a in the Supporting Information. The highest repetition rates without causing the overshooting were used to initiate the photocycles at different values of pH to reduce the time for data acquisition, as the temporal profiles were independent of the repetition rates before returning to their corresponding original states. The laser repetition rates for acquiring the contours of the time-resolved difference spectra are listed in Table S1 in the Supporting Information. Upon photoexcitation at 532 nm, the contours of the timeresolved difference spectra of HsbR exhibited the conventional intermediate characteristics, including intermediates M, N, and O,22−24 as shown in Figure 2a. At pH 4, instantaneous depletion of HsbR was observed at 550 nm, followed by the generation of intermediate M at 410 nm. A later generation at 650 nm was observed and was attributed to the difference spectrum of intermediate O with respect to the HsbR original

Figure 1. (a) Normalized UV−vis absorption spectra of light-adapted HsbR and HmbRs, with respect to the band at around 553 nm and at pH 5.8. (b), (c), and (d) represent the normalized contours, with respect to the 280 nm band, of HsbR, HmbRI, and HmbRII at different values of pH, respectively.

light-adapted HsbR and HmbRs, with respect to the band at around 553 nm are shown in Figure 1a. The 280 nm band of HmbRI is more intense than that of HmbRII because HmbRI contains three more tryptophans than HmbRII does. Regarding HsbR, the much more intense 280 nm band is attributed to the solubilization agent Triton X-100, which is composed of phenyl rings and overlaps with the absorption contour of HsbR’s constituent tryptophans. Besides the aromatic resides, the absorption contours in the visible region, attributed to the retinal Schiff base, were almost identical in all bR samples. This similarity indicates the similar chemical environments of the retinal pockets.20 In comparison with the aligned sequence, listed in Table 1, the three bRs contained same residues in the retinal pocket,20 further supporting the similarity in the absorption contours of the three bRs, attributed to the1Bu-like → S0 electronic transition of the all-trans protonated retinal Schiff base.61 As pH changed, the absorption contours of the retinal moiety of HsbR and the HmbRs changed slightly. The normalized contours, with respect to the 280 nm band, of HsbR, HmbRI, D

dx.doi.org/10.1021/jp503629v | J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

Figure 2. Two-dimensional contours of the time-resolved difference absorption spectra of (a) HsbR, (b) HmbRI, and (c) HmbRII at different pH. The flux of the incident laser at 532 nm was controlled at 0.3 mJ cm−2. The representative difference absorption spectra at given time points are shown in the right-hand column in each subfigure. The unit of the ordinate is 10−2 O.D.

the downward feature was less intense because of compensation of the generation of intermediate N, and an upward red-shifted feature was attributed to intermediate O. The pH-dependent experiments were carried out, and the resultant temporal profiles of the evolution of the intermediates at different pH are shown in Figure 5a. It is clear that the kinetics of N/O decelerated when the pH was changed from acidic to neutral, consistent with the photocycles of HsbR in PM at different values of pH.37 The transition from N to O involved proton translocation from the bulk to Asp96 on the cytoplasmic side21,37 and the reisomerization of the retinal from 13-cis to all-trans, which requires a protonated Asp96.21,54,55 As the pH was increased, the extent of the protonation of Asp96 on the cytoplasmic side was weakened, coupled with the decelerated thermal retinal isomerization from 13-cis to all-trans. Therefore, the kinetics of intermediates N and O were retarded. 3.3.2. HmbRI. The laser repetition rates were examined, and the corresponding profiles of the recovery of HmbRI at 568 nm at different pHs after 532 nm laser irradiation are shown in Figure S2b in the Supporting Information. The following transient absorption measurements were performed at 2, 1, and 0.5 Hz laser repetition at pH 4.0, 5.8, and 8.0, respectively, to avoid double excitation when the photocycle was not yet completed.

state. As the pH was adjusted to 5.8 and 7, the contours of the bR recovery at 550 nm and the intermediate O changed. An asymmetric recovery contour, on the short-wavelength side of the depletion band at 550 nm after 1 ms delay, referred to intermediate N, whose absorption contour was slightly blueshifted with respect to the depletion of HsbR.67 A simplified mechanism, adapting the conventional reversible kinetics model,31,36,37 was established for analysis of the observed timeresolved difference spectra using global analysis (Scheme 1)49,50 kM → N,O

kN,O → bR ⎯⎯⎯⎯⎯⎯→ HsbR → M N, O ⎯⎯⎯⎯⎯⎯⎯→ HsbR ←⎯⎯⎯⎯⎯⎯⎯ kN,O → M

(Scheme 1)

From surveying the observed and the manipulated contours, shown in Figure 3a, and comparing the observed and fitted temporal profiles at certain given wavelengths, shown in Figure S3a in the Supporting Information, satisfactory agreement was achieved. The difference spectra of intermediates with respect to initial-state HsbR at pH 5.8 are shown in Figure 4a. The difference spectrum of intermediate M exhibited an intense upward feature at 410 nm. The downward feature was attributed to the depletion of the HsbR. Regarding the difference spectrum of the mixture of intermediates N and O, E

dx.doi.org/10.1021/jp503629v | J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

Figure 3. Comparison of the observed and manipulated contours of difference spectra of (a) HsbR, (b) HmbRI, and (c) HmbRII at different values of pH. The time and difference absorbance scales were the same in each comparative subplot.

with HsbR, the decay of intermediate M at pH 5.8 was accelerated, with a minute retarded decay. The prolonged asymmetric recovery contour of the HmbRI after approximately 1 ms in the range of 500−600 nm implied the generation of an N-like intermediate which possessed a long lifetime. A bathochromic species in the range of 600−680 nm that appeared during the prolonged period (ca. 1 ms) could be attributed to an O-like intermediate. In addition to the similarity of the contours of the difference spectra, the constituent residues of proton acceptor (Asp82) and proton donor (Asp93), proton release group, and proton relay channel were conserved in HmbRI, as shown in Table 1. As a result, we employed HsbR-like kinetics to illustrate the photocycle of HmbRI kM → N,O

HmbRI → M

kN,O → bR ⎯⎯⎯⎯⎯⎯→ N, O ⎯⎯⎯⎯⎯⎯⎯→ HmbRI ←⎯⎯⎯⎯⎯⎯⎯ kN,O → M

(Scheme 2a)

However, the above mechanism was not able to satisfactorily explain the retarded decay at 410 nm after 1 ms. A closer examination of the temporal profile at 410 nm upon excitation of HmbRI at pH 5.8 with 8000 laser pulses, shown in Figure 6a, revealed that a fast kinetics dominated the temporal profile in the early stage, coupled with a less-intense decay component. As a result, a bypassed route, in accordance with an intermediate consisting with deprotonated Schiff base, featured below 430 nm and named M′, was introduced to modify the mechanism in Scheme 2a. It is presented schematically in Figure 7a

Figure 4. Difference spectra of the photocycle intermediates (X−bR, X = M, M′, and N/O) of (a) HsbR, (b) HmbRI, and (c) HmbRII at pH 5.8 after global analysis.

kM → M ′

kM ′→ bR

M ⎯⎯⎯⎯⎯→ M′ ⎯⎯⎯⎯⎯⎯→ HmbRI

The contours of the time-resolved difference spectra of HmbRI at different pHs, shown in Figure 2b, exhibited intrinsic differences from those of HsbR. The contours were mainly composed of the characteristics of intermediates M, N, and O. In comparison

(Scheme 2b)

Using the mechanism including Scheme 2a and Scheme 2b to analyze the time-resolved difference spectra of HmbRI at different pH with global analysis, the observed and the F

dx.doi.org/10.1021/jp503629v | J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

Figure 5. Temporal profiles of the evolution of the intermediates in (a) HsbR, (b) HmbRI, and (c) HmbRII at different pH.

Figure 6. Temporal profiles at 410 nm of (a) HmbRI and (b) HmbRII upon 532 nm excitation with 8000 laser shots.

Figure 7. Schematics of the photocycles of (a) HmbRI and (b) HmbRII.

manipulated contours (Figure 3b) and the comparison between the observed and fitted temporal profiles at certain given wavelengths (Figure S3b in the Supporting Information) were consistent with each other. This result implied that the proposed mechanism closely described the photocycle of HmbRI.

The difference spectra of intermediates with respect to initialstate HmbRI at pH 5.8 are shown in Figure 4b. The difference spectra of M and M′ were characterized in the spectral region 460−370 nm, referring to the deprotonated Schiff base retinal. G

dx.doi.org/10.1021/jp503629v | J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

generation of intermediate O, was associated with the significant movement of helix G toward the intramonomer location at extracellular side and the inside-to-outside displacement of helix E, coupled with modest structural alteration of helix F and the displacement of helix D toward the proton channel.70,71 In HsbR, an antiparallel beta-sheet existed in the B−C interhelical loop to stabilize the tertiary structure. The β sheet of HmbRI had a similar length and constituent residues to those of HsbR.56 However, the lengthened and unstructured loop D−E in HmbRI might enhance the structural flexibility for the 13-cis-toall-trans retinal isomerization happening in the deprotonated Schiff base. Moreover, more positively charged environment at the cytoplasm, containing one aspartate, five arginines, and two lysines in HmbRI, could partially lead to a quick protonation of Asp93, which could trigger the retinal isomerization from 13-cis to all-trans prior to the reprotonation of the Schiff base,21,54,55,72 referring to the generation of intermediate M′. In addition to the kinetics at pH 5.8, the difference spectra of intermediates for the HsbR-like photocycles, M and N/O, and the corresponding temporal profiles of the evolution of the intermediates at different pHs are shown in Figure S4b and Figure 5b, respectively. The contours of the difference spectra of intermediate M and N/O in the conventional photocycle were not significantly different in the spectral shift and the relative difference in the absorptivity. Instead, increasing the pH increased the population of the intermediate M′ and retarded the decay of intermediate N/O, as shown in Figure 5b. Ludmann et al. found that the increase in pH leads to the decelerated generation of intermediate N in HsbR.33 In our proposed mechanism, the generations of intermediate N and M′ were competitive, branching from intermediate M. As the pH increased, the slow generation of intermediate N could lead to an increase in the probability of the generation of intermediate M′. This interpretation supported our observation that the transient population of intermediate M′ was increased as the pH was increased. Moreover, the transition from M′ to the HmbRI original state was associated with the reprotonation of the Schiff base. It implied that the decay of intermediate M′ should be decelerated as the pH was increased. As expected, our observation confirmed the slowed transition from intermediate M′ to HmbRI. These pH-dependent results further supported our proposed mechanism. 3.3.3. HmbRII. The laser repetition rates were determined to be 2, 2, and 0.5 Hz at pH 4.0, 5.8, and 8.0 according to the temporal profiles of the recovery of HmbRII at different values of pH, as shown in Figure S2(c) in the Supporting Information. The contours of the time-resolved difference spectra of HmbRII, shown in Figure 2c, exhibited a distinct difference from HsbR and HmbRI. A long-lived N-like intermediate directly returned to the initial HmbRII without passing the intermediate O. Comparing the temporal profiles at 410 nm with HmbRI shown in Figure 6a, a pair of fast rise and decay dominated the profile in the early period, and no retarded component was observed in the later stage (Figure 6b). In other words, a bypassed pathway (Scheme 2b) was absent in the photocycle of HmbRII. Moreover, the residues responsible for the proton acceptance (Asp90) and donation (Asp101), homologous to Asp85 and Asp96 in HsbR and responsible for the kinetics of intermediate M, were conserved. As a result, a modified HsbR-like photocycle mechanism, without intermediate O, was proposed to describe the HmbRII’s photocycle and is shown in Figure 7b schematically

The contours of mixed features of intermediates N and O exhibited similarities to those in HsbR, showing downward and upward features at 550 and 620 nm, respectively. In HmbRI, the deprotonation from the protonated 13-cis Schiff base retinal to Asp82 gives rise to the generation of intermediate M. The reprotonation of deprotonated 13-cis Schiff base retinal from its counterion pair Asp93 leads to the decay of intermediate M and the generation of intermediate N. Comparing the distributions of the charged residues of HsbR and HmbRI on the cytoplasmic side and the constituent residues of the proton collection funnel53 in Table 1 reveals high similarity. As a result, a HsbR-like mechanism for the generation of intermediate N and O can be adapted in that the retinal reisomerization from 13-cis protonated Schiff base retinal to all-trans configuration, coupled with the reprotonation of Asp93 through the bulk, results in the transition from intermediate N to O. However, a minor bypassed route, from the intermediate M, existed in the photocycle of HmbRI. After the global analysis, we found that the difference spectrum of intermediate M′ was slightly blueshifted with respect to that of intermediate M by roughly 10 nm, as shown in Figure 4b. Comparing the absorption maximum of the ground state of HsbR in PM at 568 nm with that of intermediate K peaked at 590 nm,68 the absorption maximum of the all-trans retinal Schiff base at protonated status is more hypsochromic than the 13-cis. (It is reasonable to assume that structures of the bacterio-opsin moiety in ground state and intermediate K were unchanged because the generation of intermediate K is too rapid to lead to the structural alteration.) In addition, Kalisky et al. performed the second flash excitation of intermediate M and found the generation of the all-trans retinal deprotonated Schiff base intermediate, named as M′, which exhibited a blue-shifted absorption maximum at 390 nm.69 As a result, it was reasonable to assign the long-lived species M′ in HmbRI to be the all-trans deprotonated Schiff base retinal on the basis of the absorption contour. The comparisons of the retinal configuration and the protonation status of the Schiff base with the corresponding absorption maxima are summarized in Table 2. Table 2. Absorption Characteristics of Parts of the Photocycle Intermediates of HsbR and HmbRI intermediate

retinal configuration

ground state K M M′

all-trans 13-cis 13-cis all-trans

M M′

13-cis all-trans

Schiff base status HsbR protonated protonated deprotonated deprotonated HmbRI deprotonated deprotonated

λmax (nm)

ref

568 590 410 390

60 68 22 69

ca. 410