Structural and Functional Studies of a Newly Grouped Haloquadratum walsbyi Bacteriorhodopsin Reveal the Acid-resistant Light-driven Proton Pumping Activity.

Retinal bound light-driven proton pumps are widespread in eukaryotic and prokaryotic organisms. Among these pumps, bacteriorhodopsin (BR) proteins cooperate with ATP synthase to convert captured solar energy into a biologically consumable form, ATP. In an acidic environment or when pumped-out protons accumulate in the extracellular region, the maximum absorbance of BR proteins shifts markedly to the longer wavelengths. These conditions affect the light-driven proton pumping functional exertion as well. In this study, wild-type crystal structure of a BR with optical stability under wide pH range from a square halophilic archaeon, Haloquadratum walsbyi (HwBR), was solved in two crystal forms. One crystal form, refined to 1.85 Å resolution, contains a trimer in the asymmetric unit, whereas another contains an antiparallel dimer was refined at 2.58 Å. HwBR could not be classified into any existing subgroup of archaeal BR proteins based on the protein sequence phylogenetic tree, and it showed unique absorption spectral stability when exposed to low pH values. All structures showed a unique hydrogen-bonding network between Arg(82) and Thr(201), linking the BC and FG loops to shield the retinal-binding pocket in the interior from the extracellular environment. This result was supported by R82E mutation that attenuated the optical stability. The negatively charged cytoplasmic side and the Arg(82)-Thr(201) hydrogen bond may play an important role in the proton translocation trend in HwBR under acidic conditions. Our findings have unveiled a strategy adopted by BR proteins to solidify their defenses against unfavorable environments and maintain their optical properties associated with proton pumping.

Introduction

Retinal bound integral membrane proteins, rhodopsins, in diverse species of life utilize solar energy for various functions such as ion translocation, photosensing, and channel activity (1, 2). Several types of microbial rhodopsins that are proton pumps have been found in diverse microorganisms: bacteriorhodopsin (BR), blue proteorhodopsin (BPR), green proteorhodopsin (GPR), actinorhodopsin (ActR), Volvox carteri rhodopsin (VChR), Exiguobacterium sibiricum rhodopsin (ESR), and Salinibacter ruber xanthorhodopsin (XR) (3–8). Recently, some of these rhodopsins have been applied as important tools for optogenetic control of cells, tissue, and animals (9).

The light-driven proton pumps feature a seven-transmembrane α-helical region with a lysine-bound retinal that serves as a chromophore responsive to light. These BR proteins respond to ∼550 nm light and exert outward proton pumping, resulting in a proton gradient in the extracellular region (10–13). These proteins consequently facilitate the inflow of protons back into the cell through ATP synthase to generate ATP (14). The first and most well studied BR from Halobacterium salinarum, HsBR (10), was shown to be optically and functionally durable under heat and high salinity conditions (15), making it one of the most stable membrane proteins. However, BR proteins have a well established and significant property, a red-shifted activity spectrum at acidic pH, wherein a maximum red-shift of ∼55 nm in λmax for HsBR is from protonation of an aspartate residue at the retinal Schiff base (Asp85 in HsBR) (16–18). At acidic pH, the lack of proton transport is due to the protonated aspartate, which should be the proton acceptor for the Schiff base during the photocycle; in the absence of an acceptor, the proton transfer cannot take place, and a critical step in the transport does not occur. Those results are expected because releasing a proton from a protein into an environment of low pH is not chemically favored, and a red-shifted action spectrum is a conventional indicator for the protonated aspartate in the Schiff base binding pocket.

Most BR proteins are not functional under acidic conditions. After we reported a BR from Haloarcula marismortui, HmBRII, BR proteins started to surface in the past few years. HmBRII showed high optical stability in acidic conditions even down to pH of 1.6 and maintained its light-driven proton pumping activity at pH of 4.0 (19). This observation was extended when we identified another BR from Haloquadratum walsbyi, HwBR, which showed optical durability in acidic conditions.

After the structures of HsBR were solved by electron microscopy in 1996 (20) and by x-ray in 1997 (21), more than 70 HsBR structures of different length proteins, intermediates, mutants, and binding statuses have been reported, and the molecular mechanism of light-driven proton transportation was described in detail (11, 13, 22). Furthermore, six BR-like crystal structures were determined in the last decade, including bacteriorhodopsin (bR) from H. salinarum (21, 23, 24), archaerhodopsin-1 and -2 (aR-1 and aR-2) from Halorubrum sp. aus-1 and -2 (25, 26), deltarhodopsin-3 (dR3) from Haloterrigena thermotolerans (27), and cruxrhodopsin-3 (cR-3) from Haloarcula vallismortis (28). Moreover, structural information also became available for seven light-driven proton translocators identified from bacteria and eukaryota, xanthorhodopsin (XR) from S. ruber (29), Acetabularia rhodopsin (ARII) from the marine plant Acetabularia acetabulum (30), channelrhodopsin (ChR) chimera from Chlamydomonas reinhardtii (31), and both blue and green proteorhodopsin (BPR and GPR) (32–34). However, none of them has a relatively consistent activity spectrum in broad pH conditions.

Here, we report the atomic structure, sequence analysis, and photochemical properties of a BR protein, HwBR, and we propose that HwBR belongs to a new subfamily of BRs that we have named qR. The crystal structures of HwBR revealed that a unique arginine residue stabilizes the extracellular loop region by forming hydrogen bonds with a threonine residue located in the membrane edge of extracellular region. The importance of this local structure, which shields the interior environment of HwBR from the low pH extracellular area, was further validated by the mutagenesis approach.

Experimental Procedures

Phylogenomic Analysis

Thirteen BR-like amino acid sequences were used for phylogenomic analysis. The unweighted pair group method with arithmetic mean (UPGMA) algorithm was employed in this work, with the Kimura protein distance measure from CLC sequence Viewer 6.9. The bootstrap is based on 100 replicates. The abbreviations are as follows: Hr.sodomense_aR3 (archaerhodopsin-3 in Halorubrum sodomense, National Center for Biotechnology Information (NCBI) GenBankTM BAA09452.1), Hr.sp.aus-1_aR1_1UAZ (archaerhodopsin-1 in Halorubrum sp. aus-1 or also known as Halorubrum chaoviator, Protein Data Bank (PDB) ID 1UAZ), Hr.sp.aus-2_aR2_1VGO (archaerhodopsin-2 in Halorubrum sp. aus-2, PDB ID 1VGO), Hq.walsbyi_BR (bacteriorhodopsin in H. walsbyi, NCBI Gene ID 4193772), Ha.marismortui_BRII (HmBRII in H. marismortui, NCBI Gene ID 3128157), Hb.salinarum_bR_1C3W (bacteriorhodopsin in H. salinarum, PDB ID 1C3W), Ha.japonica_cR (cruxrhodopsin in Haloarcula japonica, NCBI GenBank BAA81816.1), Ha.argentinensis_cR1 (cruxrhodopsin-1 in Haloarcula argentinensis, NCBI Gene ID 1060883), Ha.vallismortis_cR3 (cruxrhodopsin-3 in H. vallismortis, NCBI Gene ID 1769808), Ha.marismortui_BRI (HmBRI in H. marismortui, NCBI Gene ID 3128463), Ha.hispanica_BR (bacteriorhodopsin in Haloarcula hispanica, NCBI Gene ID 11049305), Ha.sp.arg-2_cR2 (cruxrhodopsin-2 in Haloarcula sp. Arg-2, NCBI Gene ID 2499387), and Ha.thermotolerans_dR3 (deltarhodopsin-3 in H. thermotolerans, PDB ID 4FBZ). Sequence information for the other rhodopsins was obtained from GenBank and the Kyoto Encyclopedia of Genes and Genomes (KEGG) database.

Bacterial Strains and Expression of BR from H. walsbyi

Routine DNA manipulations were carried out according to standard molecular cloning procedures. The HwBR gene was cloned into a pUC57 vector using directed synthesis. The final sequence was CCCTXXXX753GACCTCGAG (underlining indicates the restriction sites for NcoI and XhoI, respectively, and g indicates a modified base.) The DNA fragments were obtained by NcoI and XhoI and were then inserted into the NcoI and XhoI sites of the pET-21d vector (Novagen). Consequently, a plasmid encoding hexahistidines at the C terminus was constructed. HwBR gene mutants were generated using the QuikChange site-directed mutagenesis method (Stratagene). The constructed plasmids were confirmed to have the expected nucleotide sequence using an automated sequencer.

Protein Expression in Escherichia coli and Purification

HwBR protein with hexahistidines at the C terminus was expressed in E. coli C43(DE3). The protein was purified using nickel-nitrilotriacetic acid resin chromatography (GE Healthcare) as described previously (35) and was solubilized in 0.05% n-dodecyl-β-d-maltoside (DDM).

UV-visible Spectroscopy

The purified sample was concentrated and exchanged to a buffer containing 4 m NaCl, 50 mm Tris-HCl, and 0.05% DDM using an Amicon apparatus (Millipore). UV-visible spectra were recorded using a U-1900 spectrophotometer (Hitachi). pH-dependent spectra and a titration curve were conducted as described in previous research (36). The temperature was maintained at 298 K.

Photocurrent Measurement

The electrochemical cell was designed by Chu et al. (37) and modified in our previous work (19). Photocurrent measurements of purified proteins were carried out by using a modulated continuous wave (CW) 532-nm green laser as the excitation light source and controlled by a data acquisition card.

Light-driven Proton Transport Activity

Light-driven proton transport activity was measured by monitoring light-induced pH changes using a glass electrode in real time. E. coli cells expressing the target rhodopsin were harvested by centrifugation (4,800 × g for 10 min). They were then washed three times and resuspended in measurement buffer (10 mm NaCl, 10 mm MgSO4, and 100 μm CaCl2). The concentration of the cell suspension was adjusted to obtain an A600 ∼2.0; the suspension was maintained in the dark and then illuminated with a green CW laser at 1 watt (532 nm). A parallel experiment with 10 μm carbonyl cyanide m-chlorophenyl hydrazone (CCCP) was conducted to confirm the proton-specificity of the assay.

Protein Preparation for Crystallography

To screen for the optimal HwBR crystallization conditions, purified HwBR was analyzed for monodispersity using a size exclusion column, and its absorption was monitored at 280 and 552 nm. The protein was loaded onto a size exclusion column (Superdex 200 10/30 GL; GE Health Sciences) using Buffer A (50 mm CH3COONa, pH 4.5, 200 mm NaCl) and Buffer B (20 mm Tris, pH 7.0, 150 mm NaCl) as the elution buffer in the presence of 0.05–0.15% DDM, n-octyl-β-d-glucopyranoside (OG), or n-decyl-β-maltoside (DM). After dialysis in buffer with 0.15% DM, the elution pattern showed a monodispersed peak as well. In our detergent screening experiment, the HwBR protein showed monodisperse peaks in both buffers in 0.15% DM. Therefore, we used HwBR protein in Buffer A with 0.15% DM as the sample for crystallization.

Crystallization and X-ray Diffraction Data Collection

The purified HwBR protein was concentrated to ∼17 mg/ml, as estimated by ultraviolet absorbance, and it was mixed with 1-oleoyl-rac-glycerol (monoolein; Sigma-Aldrich) at a 2:3 (w/w) protein-to-lipid ratio using the twin-syringe mixing method. The volume of each drop was 0.2 μl of protein-lipid mixture plus 1 μl. HwBR crystals of the trimeric form were grown in 0.05 m sodium citrate, pH 5.5, 0.05 m NaCl, and 15% (v/v) PEG 400, and antiparallel dimeric crystals were grown in 0.1 m ammonium sulfate, 0.1 m sodium chloride, 0.01 m sodium acetate, pH 4.0, and 16.5% (v/v) PEG 200. The size of the crystals reached about 50 × 50 × 5 μm within 2–30 days at 20 °C. The HwBR trimeric form crystals were soaked in 30% (v/v) glycerol as a cryoprotectant before harvest.

X-ray diffraction data were collected at BL15A1 of the National Synchrotron Radiation Research Center (NSRRC), Hsinchu, Taiwan and at 44XU of SPring-8, Sayo, Japan. The data were processed using HKL2000 (53). We obtained the phases by molecular replacement using archaerhodopsin-2 as a template (1VGO) (25). The PHENIX (38), refmac5 (39), and COOT (40) programs were used for molecular replacement, structural refinement, and structural viewing, respectively. All structure figures were prepared in PyMOL (Schrödinger, LLC).

Results

Sequence Analysis of BR-like Proteins

A previous study (41) compared the BR protein sequences of several species living in different environments, including salterns (42), spring areas (43), and others (44–46), and reported BR subgroups designated aR to dR (47). Multiple alignment of BR amino acid sequences (Fig. 1A) shows that the BRs share 50–80% identity. However, the amino acid sequence of HwBR from a quadrate-shaped bacterium, H. walsbyi, constitutes a novel group with HmBRII according to the phylogenetic tree analysis (Fig. 1B). We named this unidentified and distinct superfamily qR.

FIGURE 1.

Sequence alignment of the light-driven proton pumps in halobacteria. A, 13 amino acid sequences of BR-like proteins were aligned. The key residues are annotated with different symbols. Circle, retinal-binding pocket; diamond, proton reuptake residue; square, proton releasing group. The key residue, Arg82 (HwBR), in this study is marked by the red box. The secondary-structural information of HwBR is shown above the alignment. B, phylogenomics analysis of the amino acid sequences of the light-driven proton pumps in halobacteria. The analysis classified HwBR from a quadrate-shaped bacteria into a new separate superfamily, qR.

Purification of Monodispersed HwBR Proteins

The HwBR gene was constructed with a C-terminal hexahistidine tag and expressed in E. coli C43 (DE3) as described previously (35, 48). The purified HwBR protein in buffer with detergent had a λmax at 552 nm, which is almost consistent with the absorbance value of HsBR monomer, and as a reference, the absorption peak of HsBR trimer, known as purple membrane, showed a peak value at 568 nm, which was red-shifted 13 nm when compared with monomer BR (Fig. 2A) (49). Three commonly used detergents, DDM, DM, and OG, in different buffer conditions under pH 4.5 and 7.0, were tested for monodispersity properties of HwBR. The HwBR protein showed two very close peaks in both buffers with 0.02% DDM. In buffers with 0.2% OG, the HwBR protein formed broad peaks. In our detergent screening experiment, HwBR protein showed monodispersed peaks in both buffers with 0.15% DM. After the protein in OG was dialyzed against buffer A with 0.15% DM and loaded onto the column, the elution pattern was restored from the broad peaks to a monodispersed peak in DM (Fig. 3). Therefore, we selected HwBR protein in Buffer A containing 0.15% DM as the sample for crystallization, which yielded purple crystals.

FIGURE 2.

Optical property and overall structure of wild-type and D93N A, the UV-visible spectra of purple membrane (trimeric HsBR) (gray dashed line), monomeric HsBR (black dashed line), and HwBR (purple solid line). The UV-visible spectra were measured in the buffer solution containing 50 mm MES (pH 5.8), 4 m NaCl, 0.02% DDM. AU, absorbance units. B, overall structure of monomeric HwBR. Nter, N terminus; Cter, C terminus. C, top view of wild-type trimeric structure. D, top view of wild-type antiparallel dimeric structure. E, three-dimensional structure alignment of BR-like proteins. Superimposition of aR-1 (1UAZ; green), aR-2 (1VGO; yellow), bR (1C3W; cyan), dR3 (4FBZ; purple), cR-3 (4L35; blue), and qR (4QI1; pink) structures is shown.

FIGURE 3.

Buffer and detergent selection of The HwBR protein was loaded into the size exclusion column (Superdex 200 10/30 GL) and eluted by six combinations of two buffers and three detergents. The gray line is the absorption spectra at 280 nm, and the purple line is at 552 nm. AU, absorbance units.

Overall Structures and Proton Translocation Path of HwBR

Previously, we crystallized the protein using the vapor diffusion method, and the crystals diffracted to around 7 Å (48). In this work, HwBR proteins were crystallized using the in meso method (50), and the proteins packed into parallel trimeric and antiparallel dimeric crystals diffracted to 1.85 and 2.58 Å, respectively (Table 1). The structures of the monomers were almost identical and consisted of seven transmembrane α-helices, two β-strands in the BC loops, and a prosthetic group all-trans retinal bound to Lys224 via a Schiff base (Fig. 2B). The values of root mean square deviation (r.m.s.d.) between monomers of the antiparallel dimer and of the trimer are ∼0.36 Å. The trimeric structure showed that lipids (1-monooleoyl-rac-glycerol, MPG) surrounded each monomer to induce the formation of a self-assembled trimeric structure (Fig. 2C). In previous studies, it was shown that lipids control the trimeric structure, conformational flexibility, and photocycle activity of BR (51).

TABLE 1

Data collection and refinement statistics for

	WT HwBR	WT HwBR
Data collection statistics
Beamline	BL15A1, NSRRC	BL44XU, Spring-8
Wavelength (Å)	1.000	1.000
Space group	C 2	C 2
Cell dimensions
a, b, c (Å)	106.23, 61.26, 119.19	131.94, 29.80, 124.97
α, β, γ (°)	90.00, 116.01, 90.00	90.00, 118.77, 90.00
Resolution (Å)	30.0–1.85 (1.88–1.85)a	30.0–2.58 (2.67–2.58)
Rmerge (%)	6.3 (52.4)	16.8 (79.1)
I/σI	17.7 (2.41)	8.2 (2.0)
Completeness (%)	83.7 (90.1)	97.8 (98.0)
Redundancy	2.9 (3.2)	3.5 (3.1)
Refinement statistics
Resolution (Å)	26.6–1.85	28.87–2.58
No. of reflections	48,537	13,107
Rwork/Rfree, (%)	19.53/22.80	20.71/22.94
Average B-factors (Å2) (No. of atoms)
Protein	17.3 (6,354)	50.4 (3,487)
Ligand	14.8 (60) (retinal)	38.1 (40) (retinal)
	39.4 (216) (MPG)	57.0 (24) (MPG)
	42.1 (6) (glycerol)	52.1 (8) (ACT)b
Water	38.8 (743)	57.1 (102)
r.m.s.d. bonds (Å)	0.018	0.015
r.m.s.d. angles (°)	1.78	1.62
Ramachandran statistics (%)
Favored	99.7	96.9
Allowed	0.3	2.9
Disallowed	0	0.2
PDB	4QI1	4QID

Highest resolution shell is shown in parentheses.

ACT, acetate.

We used 30% glycerol as the cryoprotectant when harvesting the crystals. The presence of glycerol improved the resolution from 4 to 1.85 Å. In the structure, one glycerol molecule was found to be bound with chain B (Fig. 2C). The antiparallel dimeric structure had one MPG and two acetate molecules bound around the dimeric protein. A top view of the dimer shows that helices A and B from both monomers form a four-helix bundle for antiparallel dimer formation (Fig. 2D). In the overall structures, the BC and FG loops (Fig. 2E) have some variations that might control the differences between the superfamilies of BR-like proteins (r.m.s.d. ∼0.3–0.4 Å).

The proton translocation path could be divided into three areas (Fig. 4). On the cytoplasmic side, Asp104 (Fig. 4C) is the proton uptake accelerator, as seen for all other BR proteins, known as Asp96 in HsBR. The D104N/HwBR mutant constructed in this work showed retarded proton uptake during the light-driven proton pumping cycle when compared with the wild-type HwBR (Fig. 5). In the photocycle process, once the retinal binding site was fully protonated, a proton was translocated from Asp93 to the proton releasing group, which is composed of Arg90, Glu202, and Glu212, through the hydrogen bond rearrangement matching those observed in other BRs (Fig. 4B) (11, 18). In our 1.85 Å resolution structure, all structural waters for the proton translocation path are conserved in HwBR when compared with the 1.55 Å HsBR (1C3W). The proton outward cap (BC loop) at the extracellular site composed of a β hairpin motif was partially sealed by a pair of amide-carbonyl hydrogen bonds formed between Arg82 and Thr201 (Fig. 4A).

FIGURE 4.

The structure and proton translocation path of A, the proton outward cap region is drawn in a blue box, and residues Arg82 and Thr201 are shown as sticks. The hydrogen bonds are represented by black dashed lines. B, the residues involved in the proton-releasing group are represented by sticks in a green box. C, the retinal binding pocket and proton re-uptake residue Asp93 are shown in a red box. The proton-pumping flow is directed from the cytoplasmic site through the Schiff base to the proton-releasing complex, with protons exiting from the proton outward cap. The waters are shown as red sphere. A–C, have enlarged view on the right site labeled with key residues and waters. D, electron density maps of retinal and the surrounding region. The 2F − F electron density map contoured at 1 σ is shown in blue. The all-trans retinal is shown in orange stick, and the key residues surrounded the binding pocket are shown in magenta stick.

FIGURE 5.

Light-driven proton translocation activity assay using photocurrent measurements. A and B, an indium tin oxide-based photocurrent device was adopted to measure the light-driven photocurrent generation in both wild-type HwBR (A) and D104N/HwBR (D96N/HsBR-corresponding mutant) (B) at pH 5.8 with 0.1% DDM. A continuous 532-nm green laser was turned on at 0 s and turned off at 1.85 s while the photocurrent was continuously recorded. The light green shading indicates the light-on duration. The recovery of photocurrent traces started at time 1.85 s represented the proton reuptake step during the light-driven proton pumping. The recovery half-time (t½) values of the wild type and D104N/HwBR were around 0.05 and 0.75 s, respectively.

In both structures of HwBR solved in this study, two guanidinium nitrogen atoms of Arg82 located at the BC loop formed hydrogen bonds with the carboxyl group of the main chain of Thr201 in the FG loop in the extracellular region. This β hairpin of the BC loop forms a cap covering the proton translocation channel exit site (Fig. 4A). These hydrogen bond connections at this position have never been observed in any other known BR protein (Fig. 1A). In our HwBR structures from different crystal packing forms, Arg82 was located at the center of the β hairpin in both structures, forming hydrogen bond connections with the main chain atoms of Thr201 at the C terminus of helix F.

To further investigate the structural residue corresponding to Arg82 in other BR proteins, all resolved BR structures were aligned (Fig. 4E). Alignment of the BC loop region in HwBR with other structures of aR-1, aR-2, bR, cR-3, and dR3 (23, 25–27) showed that the residues corresponding to Arg82 are glutamic acid in all other BR proteins, except in the aR-2, where threonine is the corresponding residue, and no hydrogen bond formation in the corresponding position was observed (Fig. 6). However, there is a hydrogen bond network formed between Arg74–Thr197 in cR-3 structure. One of the hydrogen bonds in cR-3 linked to the hydrogen atom of the ϵ nitrogen, but two hydrogen bonds in HwBR were formed on the hydrogen atom on η nitrogen. The η nitrogen is sensitive to the micro-chemical environment, and a flexible cap could be formed during proton translocation process.

FIGURE 6.

Comparison of proton outward caps. The proton outward caps of structures from five BR proteins are presented for qR (magenta), bR (cyan), aR-1 (green), aR-2 (yellow), dR3 (purple), and cR-3 (pink). The residues related to Arg82 and Thr201 of HwBR are shown as blue sticks.

HwBR Is an Optically Stable BR under Wide pH Range and Mutagenesis of Arg82 Impairs Optical Stability of HwBR

An important feature of rhodopsin is the pH dependence of the maximum absorbance, λmax, or activation spectrum, which reflects the micro-environment of the retinal protonation state. Here, we found that HwBR has high optical durability under acidic conditions. E. coli-expressed HsBR and HwBR were pre-equilibrated with buffered solution at a pH of 2.0 (Fig. 7A, red) or 8.0 (Fig. 7A, blue) for spectral scanning over 250–750 nm. A mere ∼9-nm red-shift in λmax was recorded for HwBR (Fig. 7A), significantly less than the ∼55-nm red-shift observed in HsBR (19) under the same conditions. The much smaller red-shift at a pH of 2.0 represents an unusual level of optical stability that has not been observed in BR proteins other than HmBRII (19). To further obtain spectra of the fully protonated counterion states for HwBR, D93N/HwBR, the D85N/HsBR-corresponding mutant proteins, were prepared for comparison (Fig. 7A, brown). D93N/HwBR showed a red-shifted spectrum with peak at 581 nm similar to that of D85N/HsBR, being red-shifted 20 nm farther than wild-type HwBR in acidic conditions. The result hinted that the micro-environment of the fully protonated retinal-binding pocket was similar between HwBR and HsBR.

FIGURE 7.

pH-dependent transitions of wild type and R82E/ 1% DDM and 100 mm NaCl. In A and B, the red curves (pH 2) and blue curves (pH 8) indicate the spectra of wild type and R82E/HwBR, respectively. The spectra of the putative fully protonated mutant D93N/HwBR (brown curve) and R82E/D93N/HwBR (magenta curve) are shown in panels A and B, respectively. AU, absorbance units. C, pH dependence of absorption maximum of HwBR (solid circle) and R82E/HwBR (open circle) upon increasing the pH from 1.3 to 8, respectively. Each spectrum was obtained at pH 1.3, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 5.5, 6, 6.5, 7, 7.5, 8. The pH under 1.3 was inapplicable because of the protein denaturation.

To directly corroborate the importance of the Arg82–Thr201 hydrogen-bonding network in the BC loop of HwBR, a pair of mutants (R82E/HwBR and R82E/D93N/HwBR) was constructed to examine their optical stability under different pH conditions when the hydrogen-bonding network would be disrupted. A 21-nm red-shift was observed for R82E/HwBR, with a λmax value of 568 nm under a pH of 2.0 and a value of 547 nm under a pH of 8.0 (Fig. 7B). The fully protonated counterion state of R82E/HwBR was represented by R82E/D93N/HwBR, and its value remained at 581 nm (Fig. 7B, magenta). To evaluate the spectral shift upon protonation of the Schiff base proton acceptor, titration curves for both HwBR and R82E/HwBR were determined via pH-dependent spectra. HwBR showed a single titration curve with a pK of 1.97 (Fig. 7C, solid circle), which is lower than HsBR and similar to HmBRII (19). Replacement of the arginine with glutamate caused the pK to increase to 2.24 (Fig. 7C, open circle). This result supports our hypothesis that the Arg82–Thr201 hydrogen-bonding network might have some effect on the retinal Schiff base counterion.

A light-driven proton pump activity assay was also conducted on R82E/HwBR to confirm that the mutant does not interfere with the light-driven proton pumping activity. E. coli cells transformed with rhodopsins of interest were measured for their light-driven pH change property. The proton pump activity was directly monitored by a pH electrode during light and dark periods. Both wild type and R82E/HwBR showed a pH decrease upon illumination, which was eliminated by the protonophore CCCP, an inhibitor of proton motive force. This result indicated that R82E/HwBR still retains its overall proton translocation ability (Fig. 8).

FIGURE 8.

The light-driven proton pump activity assay of wild-type A and B, light-driven proton transport by HwBR (A) and by R82E/HwBR (B) in E. coli cells. Red and green lines indicate the pump activities measured before and after the addition of CCCP, respectively. The arrows show the 1-watt 532-nm continuous green laser stimulation period. The results indicated that the R82E mutant did not interrupt the overall light-driven proton pump activity in HwBR.

R82E Mutation Slightly Changes the pH-dependent Thermal Stability of HwBR

The diffusion according to the extracellular proton concentration influences the protonation state of the Schiff base in the ground state. Arg82 in HwBR locates on the BC loop and forms a cap above the proton translocation path so that the cap may shield the retinal-binding pocket from outside environment. To investigate this hypothesis, a time-dependent denaturation assay was conducted with slight modifications (43), and the experiments were monitored via spectroscopy at the corresponding λmax of wild type and R82E/HwBR under pH 4 and pH 8 at 75 °C. Both wild type and R82E/HwBR showed a similar time-dependent decrease at pH 4 in 30 min, but R82E/HwBR exhibited a faster denaturation pattern than wild-type HwBR at pH 8 within 15 min (Fig. 9). The results also suggested that the cap together with residue Arg82 had a slight effect on the protonation state in the retinal binding pocket.

FIGURE 9.

pH-dependent thermal stability of wild type and R82E/ The absorbance of wild type and R82E/HwBR was determined at 0, 5, 15, 30, 45, 60, 90, 120, and 180 min at 75 °C in the buffers at pH 4 and 8, respectively. The time versus residual pigment was plotted for wild-type (solid line) and R82E (dashed line) at pH 4 (yellow) and pH 8 (blue) to determine the k value. AU, absorbance units.

Discussion

H. walsbyi, a square halophilic archaeon, was first discovered by A. E. Walsby in 1980 (54). Because this unique square morphology halobacterium is abundant in salt lakes around the world, it plays an important role in ecology. Based on the phylogenetic tree from BR protein sequences, new subgroup qR was firstly named as part of HwBR and HmBRII, which were studied by our research team (Fig. 1B) (35). Although the overall structure of HwBR is similar to most solved BR structures, a special hydrogen-bonding network located at the extracellular region of the proton pumping path was first found.

The significance of the Arg82–Thr201 hydrogen-bonding network within the overall protein surface was summarized by an electrostatic state analysis between HwBR and other BR proteins (Fig. 10A). The key Arg82–Thr201 hydrogen-bonding network sits in the center from the top view. The location of the Arg82–Thr201 hydrogen-bonding network might shield the retinal-binding pocket from the outside proton equilibrium and protect the protonation condition of both the interior channel and the retinal-binding pocket. In other words, this hydrogen network might prevent the proton acceptor of the Schiff base from outside influences in the ground/resting state, thereby leading to the pH-independent activity spectrum. Worth et al. (52) reported that polar and certain charged side chains form hydrogen bonds to main chain atoms in the core of proteins, which is conserved in evolution. For instance, arginine exhibited the highest propensity to form capping interactions that are both conserved and buried at the C termini of α-helices.

FIGURE 10.

Electrostatic analysis of different BR structures and the proposed schematic of important proton translocation path residues and BC loop effect of A, top (extracellular) and bottom (cytoplasmic) views of known BR structures analyzed with respect to their electrostatic charge distribution, using PyMOL. The residues shown in sticks are important during light-driven proton pumping, including the proton-releasing group (Arg82, Glu194, Glu204-corresponding residues), Asp85, Asp212-corresponding residues, retinal residues, and Asp96-corresponding residues numbered in HsBR. B, electrostatic analysis of Glu74 and Ser193 for HsBR and Arg82 and Thr201 for HwBR. A negatively charged hook region in the BC loop was observed in HsBR but not in HwBR. A flat region in the HwBR BC loop showed a positive charge corresponding to the Arg82–Thr201 hydrogen-bonding region. C, a schematic of HsBR and HwBR with their proton translocation path-related residues under acidic conditions. The red solid and dashed arrows indicate better or lower accessibility for protons, respectively. Under low pH conditions, the high proton concentration in the extracellular region might access the outlet of proton releasing group (Arg82, Glu194, Glu204) in HsBR (left) via diffusion, whereas the Arg82–Thr201 hydrogen-binding network in HwBR (right) can shield the access from the outlet of proton releasing group (Arg90, Glu202, Glu212), thus maintaining the protonation status of the interior and the retinal-binding pocket.

When compared with the extracellular side, the cytoplasmic side of HwBR shows a negatively charged region (Fig. 10A, bottom view, area colored in red) with a significantly enlarged surface area among all BRs. Driving the re-uptake of proton from the cytoplasm by the negatively charged region could potentially increase the proton uptake efficiency. Taken together, HwBR has adopted a straightforward approach to achieve a negatively charged region with an enlarged surface area on the cytoplasmic side and a minimized region regulated by the Arg82–Thr201 hydrogen-bonding network on the extracellular side protecting the retinal-binding pocket micro-environment from the extracellular proton concentration direct influence (Fig. 10, B and C). Together with these two properties, HwBR seems like a highly efficient machine for proton pumping in acidic condition.

In summary, we characterized the overall important structural and photochemical features in HwBR and in comparison with other known BRs. This study demonstrated how the Arg82–Thr201 hydrogen-bonding network cap gives HwBR a stable optical property in a wide pH range. The stable optical property might lead to broaden functional pH range of light-driven proton pump activity. This protein property might play an important role in the abundance of H. walsbyi cells in salt lakes around the world.

Author Contributions

M. F. H., H. Y. F., C. S. Y., and A. H. J. W. designed the study and wrote the paper. M. F. H. crystallized and solved the structures. H. Y. F. cloned and purified proteins and performed all activity assays. C. J. C. and H. P. Y. performed light-driven proton pumping activity assay. C. S. Y. and A. H. J. W. supervised the entire project. All authors reviewed the results and approved the final version of the manuscript.