Article pubs.acs.org/biochemistry
Solid-State Nuclear Magnetic Resonance Structural Study of the Retinal-Binding Pocket in Sodium Ion Pump Rhodopsin Arisu Shigeta,† Shota Ito,‡ Keiichi Inoue,‡,§,∥ Takashi Okitsu,⊥ Akimori Wada,⊥ Hideki Kandori,‡,§ and Izuru Kawamura*,† †
Graduate School of Engineering, Yokohama National University, Hodogaya-ku, Yokohama 240-8501, Japan Department of Frontier Materials, Nagoya Institute of Technology, Showa-ku, Nagoya 466-8555, Japan § OptoBioTechnology Research Center, Nagoya Institute of Technology, Showa-ku, Nagoya 466-8555, Japan ∥ PRESTO, Japan Science and Technology Agency (JST), 4-1-8 Honcho Kawaguchi, Saitama 332-0012, Japan ⊥ Department of Organic Chemistry for Life Science, Kobe Pharmaceutical University, Higashinada-ku, Kobe 658-8558, Japan ‡
S Supporting Information *
ABSTRACT: The recently identified Krokinobacter rhodopsin 2 (KR2) functions as a light-driven sodium ion pump. The structure of the retinal-binding pocket of KR2 offers important insights into the mechanisms of KR2, which has motif of Asn112, Asp116, and Gln123 (NDQ) that is common among sodium ion pump rhodopsins but is unique among other microbial rhodopsins. Here we present solid-state nuclear magnetic resonance (NMR) characterization of retinal and functionally important residues in the vicinity of retinal in the ground state. We assigned chemical shifts of retinal C14 and C20 atoms, and Tyr218Cζ, Lys255Cε, and the protonated Schiff base of KR2 in lipid environments at acidic and neutral pH. 15N NMR signals of the protonated Schiff base showed a twist around the N−Cε bond under neutral conditions, compared with other microbial rhodopsins. These data indicated that the location of the counterion Asp116 is one helical pitch toward the cytoplasmic side. In acidic environments, the 15N Schiff base signal was shifted to a lower field, indicating that protonation of Asp116 induces reorientation during interactions between the Schiff base and Asp116. In addition, the Tyr218 residue in the vicinity of retinal formed a weak hydrogen bond with Asp251, a temporary Na+-binding site during the photocycle. These features may indicate unique mechanisms of sodium ion pumps.
F
structural changes comprise a cyclic pathway that follows formation of K, L⇄M, and O intermediates through flashphotolysis in the presence of Na+.2 Mutational and crystallographic studies suggested that an interaction between the protonated Schiff base and the counterion Asp116 plays a key role in the pumping function. In addition, Fourier transform infrared spectroscopy (FTIR) studies showed differing frequencies of N−D stretching of the KR2 Schiff base compared with that from BR, indicating a slightly stronger hydrogen bond between Asp116 and the Schiff base.12 Formation of the J intermediate proceeds quickly after photoexcitation and is involved in the isomerization of chromophores.13 Subsequently, the Schiff base proton is transferred to Asp116 at the M intermediate, followed by flipping of the Asp116 side chain to form hydrogen bonds with Ser70 and Asn112. Sodium ions are then taken up from the cytoplasmic side, while the electrostatic barrier around the
irst-found sodium ion-pumping rhodopsin Krokinobacter rhodopsin 2 (KR2) is a microbial rhodopsin (type I)1 extracted from marine flavobacterium Krokinobacter eikastus.2,3 This bacterium possesses two rhodopsins, KR1 and KR2, which pump protons and sodium ions, respectively, but the latter also pumps protons at low NaCl concentrations.4 Although other sodium ion-pumping rhodopsins have been reported,5−8 crystal structures have been determined for only KR2.9,10 Functions of rhodopsins are related to sequences that are proximal to retinal. Bacteriorhodopsin (BR)-like proton pumps have DTD motifs with Asp85, Thr89, and Asp96 residues.11 Whereas Asp85 works as a counterion of the Schiff base, Asp96 takes a proton from the cytoplasmic side. The corresponding residues in sodium ion pumps are known as Asn, Asp, Gln (NDQ) motifs and comprise Asn112, Asp116, and Gln123 in KR2.11 In contrast with those in BR, Asp116 functions as a counterion and is located one helical pitch toward the cytoplasmic side (Figure 1). Retinal absorbs light and triggers retinal conformational changes, followed by structural changes of functional residues around retinal that result in pumping of sodium ions. These © 2017 American Chemical Society
Received: September 30, 2016 Revised: December 29, 2016 Published: January 1, 2017 543
DOI: 10.1021/acs.biochem.6b00999 Biochemistry 2017, 56, 543−550
Article
Biochemistry
Figure 1. Structure of the retinal-binding pocket. (a) Top view and (c) side view of Krokinobacter rhodopsin 2 (KR2, Protein Data Bank entry 3X3C). One of two rotomers of Asp116 is shown. (b) Top view and (d) side view of bacteriorhodopsin (BR, Protein Data Bank entry 1C3W).
Asp116 residues faced the Schiff base and the other 65% took the same conformation that they did under acidic conditions.9 Hence, conformations of Asp116 are highly dependent on pH, although the ensuing mechanisms remain poorly characterized. Gushchin et al. recently showed that Asp116 is flipped away from the Schiff base in the ground state.16 Flipping of the counterion suggests that this conformational change is characteristic of KR2 during photocycles. The second Asp residue near the Schiff base is Asp251 and is located opposite Asp116 (Figure 1a). The side chain of Asp251 is less mobile under varying pH conditions and forms hydrogen bonds with Tyr218 and Arg109, either directly or via water molecules.9,10 Mutational studies indicated losses of the pumping ability of D251N similar to those observed for D116N,2 which is unusual considering that BR D212N remained active.17,18 Therefore, unlike in other rhodopsins, Asp251 is essential for the sodium ion pumping function of KR2. Solid-state nuclear magnetic resonance (NMR) has been used in structural analyses of less mobile proteins, such as membrane proteins. Hence, structural determinations of whole protein and functionally related local conformational changes have been performed. Among these, structural determinations of Anabaena sensory rhodopsin (ASR) were performed using three-dimensional chemical shift correlation techniques and showed the whole structure in a lipid environment.19 In addition, Park et al. determined the structure of seventransmembrane chemokine receptor CXCR1 in a lipid environment.20 To observe local changes in conformation, retinal and other residues are used as probes for local structural changes during photocycles. Accordingly, Lakshmi et al. demonstrated chemical shifts of retinal in BR in the ground state and in photointermediates.21 Similarly, Ahuja et al. and Eilers et al. studied TyrCζ in bovine rhodopsin and showed changes in hydrogen bond strengths of some Tyr residues in the active state.22,23 These studies indicate the utility of TyrCζ NMR signals as probes for determining hydrogen bond strength.24−26 Thus, in the study presented here, we observed a hydrogen bonding strength of Tyr218 with Asp251 in KR2 using TyrCζ-labeled samples.
Schiff base is removed and bind temporarily to Asp251 and Asn112 at the O intermediate.5,11 Protonated Asp116 then returns the proton to the Schiff base, thus preventing the back flow of sodium ions to the cytoplasmic side. Transportation of sodium ions might be hampered by the positive charge of Arg109, although additional inversions may allow the transfer of sodium ions to the extracellular side. KR2 functions both as a sodium ion pump and as a proton pump, and these pump functions are in competition. Specifically, the rate constant for Na+ uptake is much smaller than that of H + uptake. However, at higher sodium concentrations, sodium ion pumping becomes the major activity.4 Following mutation of the NDQ motif in KR2 to a DTD motif with additional residues at extracellular Na+-binding sites, the N112D/D116T/Q123D/D102N KR2 mutant lost Na+ pumping functions and was converted to a H+ pump in the presence of high concentrations of NaCl.14 Hence, the NDQ motif and sodium ion-binding site at the extracellular side are essential for the Na+ pumping function. Ion selectivity is also controlled at the cytoplasmic side. Konno et al. showed that pore formation with Gly263 and Asn61 at the cytoplasmic side prevented entry of larger ions into the protein. In contrast, the G263F/N61L mutant allowed uptake of larger ions such as cesium.15 Hence, mutations at the cytoplasmic side changed ion selectivity. Microbial rhodopsins have functionally important residues around retinal that transmit signals of changes in retinal conformation to the protein side. Previous crystallographic studies revealed the presence of Asp116 and Asp251 near the retinal Schiff base. In subsequent investigations, mutation of Asp116 to Asn caused a wide shift in the maximal absorption wavelength from 524 to 558 nm.9 In addition, a D116N mutant lost pumping activity,2 indicating that Asp116 is required for the function of this transporter. In addition, X-ray crystallography data demonstrated that Asp116 takes varying conformations at neutral and acidic pH, and Asp116 formed hydrogen bonds with Ser70 and Asn112 at pH 4.0 and faced away from the Schiff base. At low pH, the counterion is protonated and shows an M-like structure. In contrast, after samples had been soaked in pH 7.5−8.5 buffer, ∼35% of 544
DOI: 10.1021/acs.biochem.6b00999 Biochemistry 2017, 56, 543−550
Article
Biochemistry
Figure 2. 13C−13C DARR spectra of wild-type KR2 in a POPE/POPG membrane in a Tris-NaCl solution. (a) Correlation peaks of labeled residues at pH 8.0. Cross peaks and diagonal peaks are represented by the gray line. (b) Comparison of cross peaks at pH 8.0 (top, blue), pH 6.0 (middle, green), and pH 4.0 (bottom, orange). All columns show correlations with retinal C20. Cross peaks and diagonal peaks are represented by the gray line.
Herein, we report the first NMR analyses of the retinalbinding pocket of KR2 using solid-state NMR. We focused on retinal and residues around the retinal Schiff base, including retinal C20, retinal C14, Lys255, and Tyr218, in neutral and acidic environments using 13C cross-polarization magic angle spinning (CP-MAS), 13C−13C dipolar assisted rotational resonance (DARR), and 15N CP-MAS techniques. We successfully assigned all residues in the ground state. On the basis of the corresponding X-ray crystallographic data, the ensuing chemical shifts demonstrate a unique structure of the retinal-binding pocket of KR2. To describe sodium ion pumping mechanisms fully, photoinduced structural changes of the retinal-binding pocket require characterization in addition to the data presented here. We report a method for trapping photointermediates using in situ photoirradiation solid-state NMR. This method allows continuous light irradiation of microbial rhodopsin samples during measurements and traps intermediates steadily.27−29 In KR2, the relaxation time constant is 1.0 ms for the M intermediate and 7.9 and 112 ms for the O intermediate in the DOPC membrane,2 and this lifetime is sufficient for observation by in situ photoirradiation NMR. The data from these experiments may indicate structures of the retinal-binding pocket in photointermediates and suggests sodium ion pumping mechanisms that must be examined in the future.
[7-15N]Lys WT KR2. For comparison with acidic WT, [14,20-13C]retinal, [phenol-4-13C]Tyr, and [7-15N]Lys D116N KR2 were prepared. Induction was performed after an 8 h culture using 100 mg of each labeled amino acid and 500 μL of 10 mM labeled retinal. Main cultures were then stopped at 22 h, and the protein was purified using Co-NTA. Samples were separately reconstituted into widely used 1-palmitoyl-2-oleoylsn-glycero-3-phosphoethanolamine (POPE)/1-palmitoyl-2oleoyl-sn-glycero-3-phospho(1′-rac-glycerol) (POPG) membranes (3:1 POPE:POPG ratio, 1:20 protein:lipid ratio) for the study of the bacterial membrane proteins as previous work characterizing KR2.4,30,31 Subsequently, samples were suspended in 10 mM Tris-H3PO4 buffers at pH 8.0, 6.0, 5.0, or 4.0 with 100 mM NaCl or in Tris-H3PO4 buffer at pH 8.0 with 100 mM CsCl. Solid-State Nuclear Magnetic Resonance Measurement. Reconstituted samples at pH 8.0, 6.0, 5.0, and 4.0 were concentrated by centrifugation and separately packed into 4.0 mm zirconia rotors. 13C or 15N NMR experiments were then performed at 278 K, and the magic angle spinning (MAS) speed was adjusted to 10.0 kHz on a Bruker Avance III spectrometer operated at 14.1 T (600 MHz as the 1H Larmor frequency) with an E-free probe. 13C CP-MAS, 15N CP-MAS, and 13C−13C dipolar assisted rotational resonance (DARR) two-dimensional NMR32,33 were applied with a mixing time of 500 ms34,35 at each pH, and 64 points in the f1 dimension and 2048 points in the f 2 dimension were acquired. Data for the sample with Tris-CsCl buffer were obtained using protondriven spin diffusion (PDSD) at 283 K. Spinal 64 proton decoupling36 of 80 kHz was employed during acquisition. 13C chemical shifts were referenced to the carbonyl resonance of glycine powder at 176.03 ppm [tetramethylsilane (TMS) at 0.0
■
MATERIALS AND METHODS Sample Preparation. Wild-type (WT) KR2 was overexpressed in Escherichia coli strain C41(DE3) cultured in M9 minimal medium and then purified. To perform NMR analyses, we labeled samples with stable isotopes to obtain [14,20-13C]retinal, [U-13C]Asp, [phenol-4-13C]Tyr, [6-13C]Lys WT KR2 and [14,20-13C]retinal, [1,4-13C]Asp, [phenol-4-13C]Tyr, 545
DOI: 10.1021/acs.biochem.6b00999 Biochemistry 2017, 56, 543−550
Article
Biochemistry ppm], and 15N chemical shifts were referenced to Gly powder at 11.59 ppm (NH4NO3 at 0.0 ppm37).
■
RESULTS Signal Assignment of 13C−13C Correlation Peaks and Retinal Configuration. To obtain structural information about the retinal-binding pocket of KR2, we applied DARR and observed magnetization transfer between 13C nuclei.32,33 A mixing time of 500 ms was sufficient to obtain signals of labeled atoms at a distance of ∼7 Å and to observe 13C−13C correlation peaks of retinal with Tyr, Asp, and Lys residues. Although all Tyr, Lys, and Asp residues in KR2 are labeled during overexpression in E. coli, this technique allows selective and specific observations of structure in the vicinity of retinal by detecting the transfer of magnetization from retinal to nearby amino acids.35 Figure 2a shows correlation peaks between retinal C20, retinal C14, TyrCζ, and LysCε in the ground state at pH 8.0, and these were assigned as Tyr218 and Lys255 (Table S1), respectively, on the basis of crystallographic data. However, no clear correlation peaks of Asp were observed. Formerly, the correlation peak between His75 and Asp97 in [U-13C]His- and [1,4-13C]Asp-labeled green proteorhodopsin (GPR) was successfully observed despite scrambling of Asp to other residues.38 Hence, although we detected this correlation peak using NMR, our observations were limited by reduced isotope labeling efficiency due to the scrambling effect. In addition to amino acid labeling, we used [14,20-13C]retinal as a C20 probe for 13-cis/trans conformational changes and a C14 probe for 15-syn/anti conformations.39 Typical chemical shifts of retinal C20 are ∼13.0 ppm for all-trans-retinal and ∼22.0 ppm for 13-cis,15-syn-retinal in the ground state. As for C14, retinal with the all-trans conformation and 13-cis,15syn-retinal typically show peaks at ∼120.0 and ∼110.0 ppm, respectively.40 In agreement, the DARR spectra in Figure 2a show correlation peaks of C20 and C14 at 14.2 and 120.5 ppm, respectively. Hence, we assigned the retinal conformation of KR2 in the ground state to the all-trans,15-anti form. Moreover, the chemical shift of retinal did not differ very much between neutral and acidic pH. Additionally, we observed the D116N mutant for comparison with acidic WT, which showed retinal peaks at 13.2 and 119.8 ppm for C20 and C14, respectively (Figure S1). DARR spectra indicated [13Cε]Lys255 chemical shifts of 52.1 ppm at pH 8.0. Moreover, no significant changes in chemical shifts were observed with changes in pH. In the case of BR, alltrans,15-anti-retinal and 13-cis,15-syn-retinal showed peaks at 53 and 48 ppm, respectively, in the ground state.41 These data indicate that the retinal conformation of KR2 is all-trans,15-anti, but that the chemical shift is 1 ppm toward that of the 15-syn conformation. Chemical shifts of retinal C20, C14, and Lys255Cε did not change with acidification (Figure 2b), and these data correspond to observations of BR in which changes in chemical shifts were
