Biochimica et Biophysica Acta 1788 (2009) 2563–2574

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

journal homepage: www.elsevier.com/locate/bbamem

Solid-state NMR study of proteorhodopsin in the lipid environment: Secondary structure and dynamics

Lichi Shi a,b, Evelyn M.R. Lake a, Mumdooh A.M. Ahmed a,b, Leonid S. Brown a,b,⁎, Vladimir Ladizhansky a,b,⁎ a Department of Physics, University of Guelph, 50 Stone Rd. E. Guelph, Ontario, Canada N1G 2W1 b Biophysics Interdepartmental Group, University of Guelph, 50 Stone Rd. E. Guelph, Ontario, Canada N1G 2W1 article info abstract

Article history: Proteorhodopsins are typical -binding light-driven proton pumps of heptahelical architecture widely Received 20 July 2009 distributed in marine and freshwater . Recently, we have shown that green proteorhodopsin (GPR) Received in revised form 16 September 2009 can be prepared in a lipid-bound state that gives well-resolved magic angle spinning (MAS) NMR spectra in Accepted 21 September 2009 samples with different patterns of reverse labelling. Here, we present 3D and 4D sequential chemical shift Available online 30 September 2009 assignments identified through experiments conducted on a uniformly 13C,15N-labelled sample. These experiments provided the assignments for 153 residues, with a particularly high density in the Keywords: ∼ Solid-state NMR transmembrane regions ( 74% of residues). The extent of assignments permitted a detailed examination Membrane protein of the secondary structure and dynamics in GPR. In particular, we present experimental evidence of mobility Proteorhodopsin of the protein's termini and of the A–B, C–D, and F–G loops, the latter being possibly coupled to the GPR ion- Secondary structure transporting function. Protein dynamics © 2009 Elsevier B.V. All rights reserved. Carboxyl ionization

1. Introduction last decade, we have been witnessing a rapid development of magic angle spinning (MAS) solid-state NMR-based approaches to protein While membrane proteins constitute approximately a third of all structure determination. SSNMR is well suited for studies of insoluble proteins, their high resolution structures are significantly underrep- proteins, and has been used extensively to characterize amyloid fibrils resented in the protein data bank, with only about 200 unique and globular proteins [20–33]. It is also actively employed to study structures known to date (http://blanco.biomol.uci.edu/). Structural both the structure and dynamics of membrane proteins in their studies of membrane proteins are particularly important because of native-like lipid-bound state [18,34–41]. A few membrane proteins medical relevance, as more than half of membrane proteins are drug had their resonances at least partially assigned by solid-state NMR. targets. X-ray crystallography is the primary tool for structure The most complete study to date is that of DsbB [34,35], where the determination of membrane proteins, but solution NMR has been majority of resonances have been assigned from three-dimensional increasingly applied to membrane proteins as well [1–12].In solid-state NMR spectra. particular, two solution structures, for a membrane-bound disulfide In this study we apply solid-state NMR to characterize a recently bond forming , DsbB [13], and diacylglycerol kinase [14] discovered ubiquitous eubacterial light-driven retinal-binding proton (DAGK) represent the latest examples of successful application of pump, green-absorbing proteorhodopsin (GPR) [42]. The structure of solution NMR to multispan α-helical membrane proteins. Yet, it is fair GPR is not known, as it does not produce well-diffracting 3D crystals to say that there are still no universal methodologies for preparation [43]. Although GPR shares many conserved residues with its well- of well-diffracting crystals for X-ray studies, nor solubilization studied archaeal homolog (BR), especially in the conditions for solution NMR. Therefore, structure determination of retinal binding pocket, it also contains many unique residues in the membrane proteins remains one of the biggest challenges in membrane core, and lacks the extracellular proton-releasing complex, structural biology. consisting of the Glu194–Glu204 in BR. Many of the novel details of Solid-state NMR (SSNMR) established itself as a promising and proton translocation in GPR are poorly understood [44–49], which powerful tool for studies of membrane proteins in their lipid warrants further structural studies. environment (see reviews [15–19] and references therein). In the We have recently reported partial backbone and side-chain spectroscopic assignments obtained from GPR reconstituted in lipids [50]. We have used three-dimensional chemical shift correlation experiments, performed on two samples with reversely ⁎ Corresponding authors. Fax: +1 519 836 9967. E-mail addresses: [email protected] (L.S. Brown), labelled F, L, Y and W, H, Y, I, F amino acids (referred to as FLY [email protected] (V. Ladizhansky). and WHYIF samples, respectively). These amino acids were

0005-2736/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.bbamem.2009.09.011 2564 L. Shi et al. / Biochimica et Biophysica Acta 1788 (2009) 2563–2574 L. Shi et al. / Biochimica et Biophysica Acta 1788 (2009) 2563–2574 2565 selected for reverse labelling based on three criteria: abundance in Liposomes were prepared by hydrating dried DMPC and DMPA the intramembrane regions with expected high spectral degener- mixed in 9:1 ratio (w/w), and mixed with purified GPR solubilized in acy, difficulty of the detection due to magnetization diffusion in detergent, at a lipid:protein ratio of 1:2 (w/w). The mixture was bulky sidechains, and proven success in reverse labelling. In the incubated for 1 h at room temperature before adding Bio-beads for current study, we have extended multidimensional heteronuclear detergent removal. Proteoliposomes were collected by ultracentrifu- NMR to the samples with uniformly incorporated 13C and 15N gation at 150,000×g for 1 h. The presence of lipids in the reconstituted isotopic labels (UCN sample in the following). 3D and 4D chemical sample was reconfirmed by weighting as well as by mass spectros- shift correlation spectroscopy has led to assignments for a total of copy. The uniformity of lipid reconstitution was assured by the 153 out of 238 residues, with a particularly high density of narrow linewidth of the carbon and nitrogen NMR spectra. The assignments for transmembrane regions, where ∼74% of the functionality of the protein in our samples was tested using visible residues have been identified in the spectra. This allowed and FTIR spectroscopy as was explained previously [50]. identification of additional structural perturbations in the trans- membrane domain, detection of ionization states of important 2.3. MAS SSNMR spectroscopy carboxylic acids, as well as elucidation of regions of high mobility. Our discussion focuses on three major topics: (i) solid-state NMR Approximately 12 mg of UCN GPR sample was transferred and methodology and practical considerations of the 3D MAS NMR center-packed in a thin wall 3.2 mm rotor. All SSNMR experiments chemical shift correlation spectroscopy; (ii) The significance of were performed on a Bruker Avance III spectrometer operating at reverse labelling strategies in the assignment process; (iii) 800.230 MHz equipped with 3.2 mm E-free 1H–13C–15N probe (Bruker Dynamics in the loops. In our previous publication, we used 2D USA, Billerica, MA). The MAS frequency was 14.3 kHz and the sample 13C–13C correlation spectra to assess the completeness of our solid- temperature was maintained at 5 °C in all experiments. Three- state NMR data. We estimated that most of the protein residues dimensional NCOCX, NCACX, and CONCA chemical shift correlation could be seen in the 2D spectra, but variation of cross peak intensities experiments were performed using previously described pulse indicated a significant degree of mobility in GPR. The extent of sequences [50,51]. chemical shift assignments obtained in the current study allowed us to Four-dimensional CONCACX chemical shift correlation experi- refine this information, and provided clear evidence of mobility, ments were recorded as described previously [35,52]. The 4D time especially in three of the extra-membranous loops and in the termini. domain matrix was 24(t1)×36(t2)×48(t3)×1024(t4, complex), with t1, t2, t3, t4 increments of 190 μs, 212.5 μs, 95.5 μs, and 20 μs, 2. Materials and methods respectively. The carrier frequency was placed at 176.5, 116.5, 57.0,

and 120.0 ppm for nitrogen and carbon chemical shift evolution in t1 2.1. Materials (CO), t2 (N), t3 (CA), and t4 (CX) dimensions, respectively. The typical π/2 pulses were 2.5 μs for 1H, 4 μs for 13C, and 6 μs for Common chemicals of reagent grade were purchased from either 15N. 1H/15N cross-polarization (CP) [53] contact time was 1 ms, with Fisher Scientific (Unionville, Ontario, Canada) or Sigma-Aldrich the constant radio-frequency (r.f.) field of 50 kHz on nitrogen, and 15 (Oakville, Ontario, Canada). The isotopically labelled NH4Cl and with the proton lock field ramped linearly around n=1 Hartmann- 13 13 15 1 13 C6-glucose, and algal extract of 16 C, N-labelled amino acids were Hahn condition [54] (10% ramp, optimized experimentally). H/ C obtained from Cambridge Isotope Laboratories (Andover, MA). The CP contact time was 2 ms, with the constant r.f. field of 64 kHz on 13C Ni2+-NTA (nitrilotriacetic acid) agarose resin was purchased from and with the proton field ramped linearly around n=1 Hartmann– Qiagen (Mississauga, Ontario, Canada). Lipids were purchased from Hahn condition (10% ramp). 15N/13Cα and 15N/13C′ transfers [55] Avanti Polar Lipids (Alabaster, AL). were implemented with a contact time of 6 ms. For 15N/13Cα CP, a

constant lock field of 2.5×νr (νr =ωr/2π spinning frequency) field 2.2. Expression, purification and reconstitution of GPR strength was applied on 15N, while the 13C field was ramped linearly 15 13 (10% ramp) around 1.5×νr. For N/ C′ transfer, a constant lock field 13 15 The protocols for expressing GPR were described previously [50]. of 3.5×νr field strength was applied on C, while N field was Briefly, BL21-Codonplus-RIL cells were transformed with a plasmid ramped linearly (10% ramp) around 2.5×νr. The CW decoupling encoding C-terminally 6×His-tagged GPR, and cultured in during 13C/15N CP steps was always 90 kHz. The length of the 13 M9 minimal medium at 30°C, using 4 g of uniformly labelled C6- Gaussian selective pulse was 136 μs. SPINAL64 decoupling [56] of glucose, 1 g of 15N-ammonium chloride, and 250 mg of algal extract of 80 kHz was used during indirect and direct chemical shift evolutions. 16 13C,15N-labelled amino acids per litre of culture as the carbon and A single NCACX experiment was recorded with 50 ms DARR mixing, nitrogen sources. The expression of GPR was induced by 1 mM IPTG and two NCOCX experiments were performed with 50 ms and 100 ms when the cell density reached A600 = 0.4 OD. Retinal at a DARR mixing times, respectively. 4D CONCACX experiments were concentration of 7.5 μM was added exogenously at the time of recorded with 35 ms DARR mixing time. The recycle delay was 1.8 s in induction to regenerate the expressed . The harvested cells were all 3D and 4D experiments, with 16 scans per point. pre-treated with lysozyme (12 mg per litre of culture) and DNAse I (600 units per litre of culture), and subsequently broken by 2.4. Data analysis sonication. The membrane fraction was solubilized in 1 % Triton X- 100 at 4 °C, and purified following the batch procedure described in The carbon chemical shifts of the UCN sample were indirectly the Qiagen Ni2+-NTA resin manual. Approximately 15 mg of GPR was referenced to 2,2-dimethyl-2-silapentane-5-sulfonic acid (DSS) purified from one litre of culture. The molar amount of GPR was through the 13Cadamantanedownfield peak resonating at determined by the absorbance of opsin-bound retinal, using the 40.48 ppm [57]. The nitrogen chemical shifts were referenced − 1 − 1 extinction coefficient of 44,000 M cm [45]. indirectly [58] by using the ratio of gyromagnetic ratios γN/

13 13 1 Fig. 1. (A) 2D C– C chemical shift correlation spectrum of the UCN sample at 800 MHz H frequency with 1.26 ms of SPC53 mixing. (B) Comparison of the 60/40 ppm-centered 1 enlarged region between the UCN, WHYIF, and FLY [50] samples. Acquisition parameters were as follows: 100 kHz H CW decoupling during SPC53 mixing, 11.1 ms t1 acquisition time (2048×5.4 μs) with TPPI phase-sensitive detection, a 20.48 ms t2 acquisition time (1024×20 μs), and 1.8 s recycle delay. Data were processed with Lorentzian-to-Gaussian apodization functions and zero filled to 16384 (t1)×4096 (t2) prior to taking the Fourier transform. 40 Hz of Lorentzian line narrowing and 80 Hz of Gaussian line broadening were applied in both the indirect and direct dimensions. The first contour is cut at 5×σ, with each additional level multiplied by 1.3. 2566 L. Shi et al. / Biochimica et Biophysica Acta 1788 (2009) 2563–2574

γC =0.402979946, taken without temperature factor correction [59]. 3. Results and discussion To minimize slight systematic variations in peak positions between spectra from different samples (these variations are most likely due to 3.1. Efficiency of reverse labelling slight variations in the sample preparation conditions, e.g., protein: lipid ratio, and/or experimental settings such as temperature, and First, we analyze the efficiency of the reverse labelling. The usually do not exceed ±0.2 ppm), all spectra from FLY and WHYIF importance of reverse labelling as an efficient way of simplification of samples were referenced using resolved and intense CO[T177]-N solid-state NMR spectra has long been recognized [38,62]. We are [A178]-CA[A178], N[A178]-CA[A178]-CX[A178], and N[A178]-CO particularly interested in the reverse labelling efficiency because its [T177]-CX[T177] cross-peaks as internal references in the CONCA, completeness is a critical assumption in our assignment strategy, i.e., NCACX, and NCOCX spectra, respectively. wrong assumptions about the degree of unlabelling could result in Experimental data were processed with NMRPipe [60]. The erroneous assignments. 13 13 processing parameters for 3D data were described previously [50]. In Fig. 1A, we show the aliphatic region of the 2D C– CSPC53 The 4D data was apodized with Lorentzian-to-Gaussian function spectrum collected from the UCN sample, with all peaks labelled and zero filled to 128(t1)×256(t2)×256(t3)×2048(t4, complex) according to our assignments. Only four of the peaks assigned from prior to taking the Fourier transform. 16 Hz of Lorentzian line the 3D experiments are not observed in this spectrum: F205 CA/CB narrowing and 48 Hz of Gaussian line broadening were applied to at 58.7/36.3 ppm, I112 CG1/CD1 at 26.5/15.8 ppm, M146 CB/CG at 15 the t2 N indirect dimension, 40 Hz of Lorentzian line narrowing 38.0/32.1 ppm, and I193 CB/CG2 at 36.5/19.0 ppm. F205 is located 13 and 100 Hz of Gaussian line broadening were applied to the t1 CO at the edge of the F helix and is probably quite mobile. The 3D 13 and t3 CA indirect dimensions, 40 Hz of Lorentzian line narrowing correlations for F205 are also quite weak, likely because of the 13 and 80 Hz of Gaussian line broadening were applied to the t4 CX dynamics at the interface (see below for further discussion on direct dimension. Peak picking and noise analysis were performed dynamics). The remaining three missing cross peaks all correspond in the CARA environment [61]. to side chain correlations. Two of the residues, M146 and I112, are

Fig. 2. A comparison of spectral resolution of the 2D NCACX and 3D NCACX spectra. Certain amino-acid specific regions are indicated in the 2D NCACX. The assignments shown in the 3D spectra are according to the previously published data [50]. L. Shi et al. / Biochimica et Biophysica Acta 1788 (2009) 2563–2574 2567 located in a loop or at the interface, and are probably affected by correlation spectra indicates a very complicated transamination motion as well. pattern, with all reversely labelled amino acids being susceptible With the number of current assignments obtained from the to transamination. combined analysis of the WHYIF, FLY, and UCN samples, we can confirm that in the BL21-Codonplus-RIL strain the 3.2. Spectral resolution in two-dimensional and reverse labelling for carbon atoms in Trp, His, Phe, Tyr, Leu, and Ile is three-dimensional spectroscopy nearly 100% efficient. This is corroborated by examining the 2D 13C– 13C correlation spectra shown in Fig. 1B, which compare the spectral Lipid-reconstituted samples of GPR yield NMR spectra with typical regions around 60/40 ppm along the F1/F2 dimensions in the UCN, line widths on the order of 0.5–0.6 ppm for carbon, and ∼0.7 ppm for FLY, and WHYIF samples. Most CA/CB correlations for Phe, Ile, Tyr, and nitrogen resonances. An example of the 2D NCACX correlation Leu are expected to be in this region. Most peaks corresponding to the spectrum collected on the WHYIF sample is shown in Fig. 2. This unlabelled amino acids completely disappear from the spectra of spectrum has many well-resolved resonances, some of which can be reversely labelled samples, indicating the high efficiency of reverse easily identified by residue type, e.g. threonines, glycines, etc. Yet, labelling. Some visible exceptions occur when the cross peaks of there is a considerable degree of spectral overlap and degeneracy unlabelled residues overlap with those of labelled residues. For which makes the application of 2D-based assignment strategies very example, L41 CA/CB correlation at 58/39 ppm overlaps with difficult for GPR. The straightforward approach to simplify the data unassigned Asp or Asn cross peaks. Furthermore, the relative analysis is to extend the experiment to a third dimension. As was intensities of resonances of unaffected amino acids are generally not discussed in a number of publications [64–66], the minimal set of 3D altered by the reverse labelling of W, H, Y, I, F, L, indicating minimal experiments includes NCACX, NCOCX, and CONCA correlations. scrambling between amino acids types, as expected from a number of Although the sensitivity of these 3D experiments is lower than that earlier studies cited above. of 2D–the 2D NCACX in Fig. 2 was collected in 8.7 h, while the 3D took The reverse labelling of nitrogen is generally more complicated. ∼3.5 days–the benefits are overwhelming. Many spin systems can be In some simple cases, it can be assessed directly from the 1D 15N uniquely identified with sufficiently high sensitivity, as can be clearly spectra (not shown). For example, we found that the ND1/NE2 seen in the lower panel of Fig. 2, where nearly the entire side chain of atoms of His are almost entirely unlabelled. The effect of Lys172 is observed with up to 3 bond transfers. Likewise, the entire transamination from alanine to valine amides has been described side chains for Leu113 and Leu135 are detected, which makes building in the literature [38,62,63], but little is known about other types spin systems and their assignment to an amino-acid type a of amino acids. Our analysis of the heteronuclear NCOCX straightforward task. Thus, the spectral simplification offered by 3D

Fig. 3. An example of two-dimensional planes of the (A) 3D NCACX and (B) NCOCX experiments with 50 ms of DARR mixing performed on the UCN sample. Most of the peaks are assigned in these planes, but only a few assignments are indicated for simplification purposes. In both spectra, the first contour is cut at 4σ, with each additional contour level multiplied by 1.2. 2568 L. Shi et al. / Biochimica et Biophysica Acta 1788 (2009) 2563–2574

Fig. 4. Sequential walk from I193 to I199 through the NCOCX, CONCA, and NCACX spectra. Each strip is an N-C 2D plane cut at the indicated chemical shift. The 15N axis is 2 ppm wide. Each block of the three spectra, NCOCX, CONCA, and NCACX, represents an extended CX[i-1]-N[i]-CX[i] spin system. For example, the I199 NCACX peaks are associated to A198–I199 CONCA peaks through matching N and CA shifts, the A198 NCOCX peaks are associated to A198–I199 CONCA peaks through matching N and CO shifts. It is linked to the preceding system through the entire three carbon shifts of A198 that are present in both NCACX and NCOCX spectra. L. Shi et al. / Biochimica et Biophysica Acta 1788 (2009) 2563–2574 2569 spectroscopy results in a much more reliable spectral analysis, greatly A strip plot for the fragment of 7 residues from I193 to I199 reducing chances of misassignment. assigned in the UCN sample is shown in Fig. 4. The linking for the most In our previous study, we have sequentially assigned 103 residues part relies on matching shifts of three carbon resonances or more, and in a series of three-dimensional chemical shift correlation experiments, is therefore very reliable. The only visible exception is the linking of performed on FLY and WHYIF samples. Many more spin systems were W197 with G196 and A198. Generally, identification and linking of identified but could not be assigned, largely because of the interrup- aromatic residues is more difficult because of the heavy overlap in tions in the backbone walk introduced by the reverse labelling. their NCA and NCO regions, and since the signals from these residues In this study, the experiments have been extended to a sample are weaker because of the magnetization diffusion to the ring carbons with uniformly incorporated 13C,15N labels. The established protocol [68]. In this , linking G196–W197–A198 requires additional for 3D-based sequential assignment involves building extended CX[i- information from the reverse labelling and deserves a special 1]-N[i]-CX[i] systems and the assignment of the spin systems to discussion. amino acid types, linking the extended systems into contiguous Fig. 5 compares 2D strips from the CONCA, NCACX, and NCOCX fragments, and mapping the constructed fragments to the protein spectra recorded in the UCN and WHYIF samples (the UCN strips amino acid sequence. In the first step, the CO[i-1]-N[i]-CA[i] peaks are are the same as those in Fig. 4, and shown here for direct detected and picked in the 3D CONCA spectrum. These peaks are comparison). It is clear from the NCACX spectra in Fig. 5A that the matched to the N[i]-CO[i-1]-CX[i-1] and N[i]-CA[i]-CX[i] series of CB peak is missing in the W197 pattern, most likely because of the peaks in the NCOCX and NCACX spectra, through matching N[i]-CO[i- magnetization diffusion to the bulky indole ring [68]. It is also 1] and N[i]-CA[i] pairs of shifts, respectively. The NCOCX and NCACX apparent that W197 has degenerate N, CO, and CA shifts in both experiments have high sensitivity and allow detection of nearly entire NCOCX and NCACX planes. The information on the W197 chemical side chains for the aliphatic residues. Two examples of the 2D planes shifts can be obtained from the resolved peaks in the CONCA extracted from the 3D NCOCX and NCACX experiments are shown in spectrum: the CO[W197]-N[A198]-CA[A198] peak identifies the Fig. 3. For instance, the entire side chains are clearly detectable for shift of W197C' whereas the CO[G196]-N[W197]-CA[W197] peak I106, I194, and V241 in the NCACX spectrum in Fig. 3A, and for T29, gives the shift of W197CA. The W197–Ala198 linking can still be V229, and V45 in the NCOCX spectrum in Fig. 3B. performed based on the pair of W197 C' and CA shifts. Further High signal-to-noise ratio of the spectra and detection of full side confirmation of the sequential assignment of W197 resonances chains greatly facilitate identification of the amino acid type and comes from the comparison with the spectra of the reversely building contiguous fragments. The fragments are built by simulta- labelled samples. In particular, the intense and well-resolved CO neously matching CX[j-1] and CX[i] shifts between the two extended [G196]-N[W197]-CA[W197] and CO[W197]-N[A198]-CA[A198] systems, CX[i-1]-N[i]-CX[i] and CX[j-1]-N[j]-CX[j], which is performed cross-peaks found in the spectra of the UCN sample disappear in using Autolink [67] in the CARA environment [61]. When CX[j-1] and the spectra of the WHYIF sample shown in Fig. 5B. These peaks CX[i] shifts match, the two extended spin systems are merged into a remain intact in the spectra of the FLY sample (not shown), three residue fragment by setting j=i+1. supporting the assignment of the W197 resonances.

Fig. 5. 2D strip plots for W197 in the (A) UCN and (B) WHYIF samples. UCN strips are reproduced from Fig. 4. Only peaks relevant to the discussion are labelled. 2570 L. Shi et al. / Biochimica et Biophysica Acta 1788 (2009) 2563–2574

3.3. Four-dimensional chemical shift correlation spectroscopy procedure described above, only a few minor conflicts with the previously published results [50] were found and corrected. First, the A combined analysis of the 3D experiments conducted in the three assignments of T29–G30 and T206–G207 fragments were corrected: samples, FLY, WHYIF, and UCN, has resulted in backbone and side T206–G207 pair could now be assigned unambiguously as part of a chain assignments for a total of 153 residues, but some spectral long fragment. Second, an assignment of M190C' was removed as not regions were found to be not sufficiently resolved. In an attempt to being reliable because of the spectral overlap, and 13C shifts of T44C' resolve this remaining ambiguity and to evaluate the feasibility of 4D and P103C' misassigned in the previous work were corrected. Third, spectroscopy in membrane proteins, we acquired a 4D CONCACX chemical shift positions of 9 13C resonances (S43 CA, S43 CB, L113 CA, spectrum on the UCN sample. The performance of the 4D CONCACX M134 CB, L135 CB, M190 CG, G203 C, K231 CG, and N240 CA) have experiment was analyzed in detail in a small globular protein GB1 been adjusted by 0.3–0.5 ppm, which was made possible by the [52], and in a much more challenging case of the membrane- improved signal-to-noise ratio and better defined line shapes in the embedded DsbB [35]. The main challenge for 4D spectroscopy in a U–13C,15N spectra. Finally, a backbone 15N shift of Asn224 was membrane protein is its relatively poor sensitivity. In GPR, the adjusted by 0.4 ppm. The assigned chemical shifts have been CONCACX experiment took about 2 weeks. A total of 83 spin systems submitted to BioMagResBank (BMRB entry 15955). could be identified, matching well with the 3D spectra, and confirming many of the assignments. In most cases, the CONCACX 3.5. Protein secondary structure and ionization states of carboxylic acids polarization transfer scheme has resulted in reliable cross-peaks corresponding to one bond transfers (i.e. CA-CB and CA-CO). In some The chemical shift-based secondary structure analysis is shown in cases, however, two bond transfers could be seen, e.g., for Val133 and Fig. 7A. In addition to the previously detected proline kink at L100 in I154 in Fig. 6. Further S/N improvement will likely be required for 4D the helix C, and a non-proline π-bulge at the Schiff base-forming K231 spectroscopy to become routinely feasible in SSNMR of membrane in the helix G, we can identify a number of other distortions. Some of proteins. In the case of GPR, this can probably be achieved by them occur at the edges of the helices, e.g., G141, but many are buried conducting experiments at lower temperature, provided that appro- inside the transmembrane core, including those at F73, W98, L135, priate freezing conditions that do not lead to substantial line and W197. broadening can be found. This optimization is currently under way The two Trp residues, W98 and W197, are the homologs of the in our laboratory. retinal-flanking W86 and W182 in BR. The unusual conformation of W98 and W197 in GPR may be caused by a steric conflict with the 3.4. Summary of the resonance assignments retinal . Additionally, W98 and L135 may be in a close contact as both are located one helical turn from the site of a tight Based on the analysis of the 3D and 4D experiments performed on interhelical interaction, predicted from the homology with BR. In GPR, the UCN, WHYIF, and FLY samples, we have assigned 153 residues, or the interacting hydrogen-bonded pair T90–D115 is replaced by the 727 15N and 13C atoms (Table S1), with a particularly high density of V102–S131 pair, which may allow for a closer interhelical contact, assignments for intramembrane amino acids. Besides the subtle causing a steric clash between W98 and L135 with local distortion of differences (within 0.3 ppm) due to the internal referencing the helices.

Fig. 6. Representative two-dimensional F3–F4 planes of the 4D CONCACX experimental data. L. Shi et al. / Biochimica et Biophysica Acta 1788 (2009) 2563–2574 2571

Fig. 7. (A) Chemical Shift Index analysis for GPR. Positive indices indicate α-helical structure, while negative indices show β-strand elements and other deviations from helicity. Helices, predicted by the homology modeling, and refined from the CSI values, are shown as rectangles on top. (B) Overall transmembrane architecture of GPR. The general topology was taken from the homology modeling, and refined using the CSI analysis. The signal peptide cleavage site was predicted using SignalP 3.0 Server [80, 81] (http://www.cbs.dtu.dk/ services/SignalP/). Assigned residues are shown as filled circles.

The secondary structure analysis further confirms that the helix E deprotonated Asp85 in BR [70], with the protonated Asp85 resonating of GPR is quite short, comprised of only 18 residues, much shorter ∼4 ppm lower [71]. The anomalously low value can be explained by than in BR, but similar to that in XR [69]. The helix F, connected to it by the strong hydrogen bond between Asp97 and water molecules an extra-membranous peripheral helix, is quite long, containing 26 present in the [47]. Interestingly, the absolute positions and residues. the differences in the chemical shifts of the two counterion carboxyls The current set of assignments allows assessing the ionization (Asp97 and Asp227 of GPR, and Asp85 and Asp212 of BR), are almost states of carboxylic acids. In addition to earlier assignment of identical for the two proteins, despite the observed differences in their functionally important Asp227 [50], which forms a part of the Schiff proton affinities [44] and the of the two proteins base counterion, we have determined the chemical shift of Asp97, its [47,72]. primary proton acceptor. As expected from homology with BR and Additionally, we have determined the ionization state of Glu165, FTIR measurements [47], both of these residues are deprotonated, which is located close to the cytoplasmic surface and appears to be with Asp227 Cγ signal at 178.1 ppm, and that of Asp97 at 175.4 ppm. deprotonated. This is what can be expected from rather polar The latter value, although somewhat low for a deprotonated aspartic character of the cytoplasmic end of the helix E, which is in high acid, is very close to the value measured for the homologous contrast to the homologous non-polar region in BR. 2572 L. Shi et al. / Biochimica et Biophysica Acta 1788 (2009) 2563–2574

Fig. 8. 2D skyline N-CA projection of 3D NCACX and CONCA spectra of the UCN sample. The first contour is cut at 5×σ, with each additional level multiplied by 1.2. Detected but unassigned cross peaks are labelled as Gxx1, Gxx2, etc.

3.6. Mobility and sensitivity such as proteorhodopsins [74]. In particular, in GPR the Glu194– Glu204 pair is replaced with the hydrophobic Leu209–Leu219 Fig. 7B summarizes the extent of assignments obtained in GPR. residues, located in the F–G loop, and in the helix G, respectively. Most of the unassigned residues comprise contiguous fragments, The F–G loop contains three glycines and a serine, which may mainly located in the tails, loops (A–B, C–D, F–G), or at the edges of impart it a structural flexibility, in accordance with our previous the helices. This suggests that local dynamics may play a significant conclusion that both glycyl and seryl residues located in the loops role [38,73] and result in partial or complete averaging of dipolar are mobile. Our hypothesis is that the F–G loop may act as a interactions. hydrophobic plug, transiently exposing a hydrophilic cavity in the Information on dynamics can be deduced from the analysis of extracellular half of the proton translocation pathway in GPR. This is relative peak intensities. Glycines are distributed throughout both in line with a recent X-ray structure of a distant carotenoid-binding transmembrane and loop regions and are good candidates for such homolog of GPR, XR [69], where such a cavity was observed, while analysis because we have assigned all transmembrane glycine most of the F–G loop was disordered, possibly due to its high signals. There are 23 glycines in GPR, 21 of which are detected in mobility. Based on homology modeling, this cavity in GPR could the 3D spectra. Fig. 8 shows a 2D NCA projection of the NCACX include such polar residues as Glu142 and Tyr223, and extend deep spectrum, in which all ten glycines located in the transmembrane into the intramembrane region to reach Arg94. regions and six extra-membranous glycines are shown with their In summary, we have demonstrated an efficient approach to the assignments. chemical shift assignments of membrane proteins based on the The remaining five cross peaks corresponding to glycyl residues, combined use of multidimensional (3D, 4D) chemical shift correlation detected in the 3D spectra but unassigned, are denoted as Gxx1, Gxx2, spectroscopy and reverse labelling schemes. The major hurdle for etc. These peaks must correspond to the residues located either in the completing assignments is the mobility of the protein, which results C–DorF–G loops, or in the N-terminus. They are generally of lower in decreased efficiency of dipolar-driven polarization transfers for the intensity in the 3D NCACX spectrum, consistent with increased mobile residues. In this respect, we would like to propose two mobility of the loops and the termini. Compared to NCACX, the CONCA approaches. Firstly, in our observation, the use of sparse labelling transfer involves an extra dipolar-driven polarization transfer step, helps improve sensitivity of the measurements, and can result in and is expected to be more sensitive to motion. Thus, the attenuated many additional assignments. Secondly, the application of hetero- relative intensities of Gxx1, Gxx4, and Gxx5, cross peaks of extra- nuclear J-spectroscopy [38,75–79] is of a great interest, and will likely membranous G81, G87, and G207, and of the interfacial G30 in the result in additional assignments for mobile residues, especially in the CONCA spectrum (Fig. 8) serve as further evidence of the increased loops and tails. The J-driven methods can become even more dynamics in the loop and termini regions. On the other hand, the beneficial if combined with extensive deuteration of GPR thus relative intensities of the G144, G169, and G171 peaks are much less improving its relaxation properties. This work is in progress in our lab. affected in the CONCA because their mobility is restricted by the local peripheral secondary structure. Acknowledgements Similar information can be retrieved from the analysis of serine correlation patterns: there are nine unassigned serines, three of which This research was supported by the University of Guelph, the are located in the termini, three in the loops, and the remaining three Natural Sciences and Engineering Research Council of Canada are in the helices. In a combined analysis of the three 13C–13C (Discovery Grant RG298480-04 to V.L.; Discovery Grant DG250202 correlation spectra from the UCN, FLY, and WHYIF samples shown in to L.S.B.), the Canada Foundation for Innovation and Ontario the Supplemental information (Figure S1), we can identify only six Ministry for Research Innovation. V.L. holds Canada Research very weak CA/CB serine correlations, again indicating a high degree of Chair in Biophysics, and is a recipient of an Early Researcher dynamics affecting the efficiency of dipolar-driven polarization Award from the Ontario Ministry of Research and Innovation. L.S. transfers for these residues. is a recipient of Ontario Graduate Scholarship. M.A. is a recipient of The mobility of the F–G loop may be coupled to the GPR ion- a Doctoral Studentship from the Ministry of Higher Education and transporting function. In comparison to its structural and functional Scientific Research of Egypt. We thank Drs. Werner Maas and archaeal homolog, bacteriorhodopsin, GPR lacks the pair of Jochem Struppe for their assistance with the optimization of the glutamates which form part of the proton release complex thin-wall rotor performance, Ms. Valerie Robertson and Mr. Peter (Glu194 and Glu204). This pair is typical for halobacterial and Scheffer of the University of Guelph NMR Centre for their technical fungal , but quite uncommon in eubacterial rhodopsins assistance. L. Shi et al. / Biochimica et Biophysica Acta 1788 (2009) 2563–2574 2573

Appendix A. Supplementary data [28] S.G. Zech, A.J. Wand, A.E. McDermott, Protein structure determination by high- resolution solid-state NMR spectroscopy: application to microcrystalline ubiqui- tin, J. Am. Chem. Soc. 127 (2005) 8618–8626. Supplementary data associated with this article can be found, in [29] W.T. Franks, D.H. Zhou, B.J. Wylie, B.G. Money, D.T. Graesser, H.L. Frericks, G. the online version, at doi:10.1016/j.bbamem.2009.09.011. Sahota, C.M. Rienstra, Magic-angle spinning solid-state NMR spectroscopy of the beta 1 immunoglobulin binding domain of protein G (GB1): N-15 and C-13 chemical shift assignments and conformational analysis, J. Am. Chem. Soc. 127 References (2005) 12291–12305. [30] D.H. Zhou, J.J. Shea, A.J. Nieuwkoop, W.T. Franks, B.J. Wylie, C. Mullen, D. Sandoz, [1] K. Oxenoid, H.J. Kim, J. Jacob, F.D. Sonnichsen, C.R. Sanders, NMR assignments for a C.M. Rienstra, Solid-state protein-structure determination with proton-detected helical 40 kDa membrane protein, J. Am. Chem. Soc. 126 (2004) 5048–5049. triple-resonance 3D magic-angle-spinning NMR spectroscopy, Angew. Chem. Int. [2] T.P. Roosild, J. Greenwald, M. Vega, S. Castronovo, R. Riek, S. Choe, NMR structure Ed. Engl. 46 (2007) 8380–8383. of Mistic, a membrane-integrating protein for membrane protein expression, [31] D.H. Zhou, G. Shah, M. Cormos, C. Mullen, D. Sandoz, C.M. Rienstra, Proton- Science 307 (2005) 1317–1321. detected solid-state NMR spectroscopy of fully protonated proteins at 40 kHz [3] C.M. Franzin, P. Teriete, F.M. Marassi, Structural similarity of a membrane protein magic-angle spinning, J. Am. Chem. Soc. 129 (2007) 11791–11801. in micelles and membranes, J. Am. Chem. Soc. 129 (2007) 8078–8079. [32] W.T. Franks, B.J. Wylie, H.L.F. Schmidt, A.J. Nieuwkoop, R.M. Mayrhofer, G.J. Shah, [4] J.R. Schnell, J.J. Chou, Structure and mechanism of the M2 proton channel of D.T. Graesser, C.M. Rienstra, Dipole tensor-based atomic-resolution structure influenza A virus, Nature 451 (2008) U512–U591. determination of a nanocrystalline protein by solid-state NMR, Proc. Natl. Acad. [5] E.V. Bocharov, K.S. Mineev, P.E. Volynsky, Y.S. Ermolyuk, E.N. Tkach, A.G. Sobol, V.V. Sci. U. S. A. 105 (2008) 4621–4626. Chupin, M.P. Kirpichnikov, R.G. Efremov, A.S. Arseniev, Spatial structure of the [33] A.E. McDermott, Structural and dynamic studies of proteins by solid-state NMR dimeric transmembrane domain of the growth factor receptor ErbB2 presumably spectroscopy: rapid movement forward, Curr. Opin. Struct. Biol. 14 (2004) corresponding to the receptor active state, J. Biol. Chem. 283 (2008) 6950–6956. 554–561. [6] A. Gautier, J.P. Kirkpatrick, D. Nietlispach, Solution-state NMR spectroscopy of a [34] Y. Li, D.A. Berthold, H.L. Frericks, R.B. Gennis, C.M. Rienstra, Partial C-13 and N-15 seven-helix receptor: backbone assignment, secondary chemical-shift assignments of the disulfide-bond-forming enzyme DsbB by 3D structure, and dynamics, Angew. Chem. Int. Ed. Engl. 47 (2008) 7297–7300. magic-angle spinning NMR spectroscopy, Chembiochem 8 (2007) 434–442. [7] P.M. Hwang, W.Y. Choy, E.I. Lo, L. Chen, J.D. Forman-Kay, C.R.H. Raetz, G.G. Prive, [35] Y. Li, D.A. Berthold, R.B. Gennis, C.M. Rienstra, Chemical shift assignment of the R.E. Bishop, L.E. Kay, Solution structure and dynamics of the outer membrane transmembrane helices of DsbB, a 20-kDa integral membrane enzyme, by 3D enzyme PagP by NMR, Proc. Natl. Acad. Sci. U. S. A. 99 (2002) 13560–13565. magic-angle spinning NMR spectroscopy, Protein Sci. 17 (2008) 199–204. [8] C. Fernandez, C. Hilty, G. Wider, P. Guntert, K. Wuthrich, NMR structure of the [36] A. Goldbourt, B.J. Gross, L.A. Day, A.E. McDermott, Filamentous phage studied by integral membrane protein OmpX, J. Mol. Biol. 336 (2004) 1211–1221. magic-angle spinning NMR: resonance assignment and secondary structure of the [9] M.U. Johansson, S. Alioth, K.F. Hu, R. Walser, R. Koebnik, K. Pervushin, A minimal coat protein in Pf1, J. Am. Chem. Soc. 129 (2007) 2338–2344. transmembrane beta-barrel platform protein studied by nuclear magnetic [37] A. Lange, K. Giller, S. Hornig, M.F. Martin-Eauclaire, O. Pongs, S. Becker, M. Baldus, resonance, Biochemistry-US 46 (2007) 1128–1140. Toxin-induced conformational changes in a potassium channel revealed by solid- [10] B.Y. Liang, L.K. Tamm, Structure of outer membrane protein G by solution NMR state NMR, Nature 440 (2006) 959–962. spectroscopy, Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 16140–16145. [38] M. Etzkorn, S. Martell, O.C. Andronesi, K. Seidel, M. Engelhard, M. Baldus, [11] S. Hiller, R.G. Garces, T.J. Malia, V.Y. Orekhov, M. Colombini, G. Wagner, Solution Secondary structure, dynamics, and topology of a seven-helix receptor in native structure of the integral human membrane protein VDAC-1 in detergent micelles, membranes, studied by solid-state NMR spectroscopy, Angew. Chem. Int. Ed. 46 Science 321 (2008) 1206–1210. (2007) 459–462. [12] M. Bayrhuber, T. Meins, M. Habeck, S. Becker, K. Giller, S. Villinger, C. Vonrhein, C. [39] C. Ader, R. Schneider, S. Hornig, P. Velisetty, E.M. Wilson, A. Lange, K. Giller, I. Griesinger, M. Zweckstetter, K. Zeth, Structure of the human voltage-dependent Ohmert, M.F. Martin-Eauclaire, D. Trauner, S. Becker, O. Pongs, M. Baldus, A anion channel, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 15370–15375. structural link between inactivation and block of a K+ channel, Nat. Struct. Mol. [13] Y.P. Zhou, T. Cierpicki, R.H.F. Jimenez, S.M. Lukasik, J.F. Ellena, D.S. Cafiso, H. Biol. 15 (2008) 605–612. Kadokura, J. Beckwith, J.H. Bushweller, NMR solution structure of the integral [40] M. Etzkorn, H. Kneuper, P. Dunnwald, V. Vijayan, J. Kramer, C. Griesinger, S. membrane enzyme DsbB: Functional insights into DsbB-catalyzed disulfide bond Becker, G. Unden, M. Baldus, Plasticity of the PAS domain and a potential role for formation, Mol. Cell 31 (2008) 896–908. signal transduction in the histidine kinase DcuS, Nat. Struct. Mol. Biol. 15 (2008) [14] W.D. Van Horn, H.J. Kim, C.D. Ellis, A. Hadziselimovic, E.S. Sulistijo, M.D. Karra, C.L. 1031–1039. Tian, F.D. Sonnichsen, C.R. Sanders, Solution nuclear magnetic resonance structure [41] A. Bockmann, 3D protein structures by solid-state NMR spectroscopy: Ready for of membrane-integral diacylglycerol kinase, Science 324 (2009) 1726–1729. high resolution, Angew. Chem. Int. Ed. 47 (2008) 6110–6113. [15] M. Hong, Structure, topology, and dynamics of membrane peptides and proteins [42] O. Beja, L. Aravind, E.V. Koonin, M.T. Suzuki, A. Hadd, L.P. Nguyen, S. Jovanovich, from solid-state NMR Spectroscopy, J. Phys. Chem. B 111 (2007) 10340–10351. C.M. Gates, R.A. Feldman, J.L. Spudich, E.N. Spudich, E.F. DeLong, Bacterial [16] A. Ramamoorthy, Beyond NMR spectra of antimicrobial peptides: Dynamical : evidence for a new type of phototrophy in the , Science 289 images at atomic resolution and functional insights, Solid State Nucl. Magn. 35 (2000) 1902–1906. (2009) 201–207. [43] P. Gourdon, A. Alfredsson, A. Pedersen, E. Malmerberg, M. Nyblom, M. Widell, R. [17] M. Baldus, Molecular interactions investigated by multi-dimensional solid-state Berntsson, J. Pinhassi, M. Braiman, O. Hansson, N. Bonander, G. Karlsson, R. Neutze, NMR, Curr. Opin. Struct. Biol. 16 (2006) 618–623. Optimized in vitro and in vivo expression of proteorhodopsin: a seven- [18] A. McDermott, Structure and dynamics of membrane proteins by magic angle transmembrane , Protein Expr. Purif. 58 (2008) 103–113. spinning solid-state NMR, Annu. Rev. Biophys. 38 (2009) 385–403. [44] A.K. Dioumaev, L.S. Brown, J. Shih, E.N. Spudich, J.L. Spudich, J.K. Lanyi, Proton [19] H. Saito, A. Naito, NMR studies on fully hydrated membrane proteins, with transfers in the photochemical reaction cycle of proteorhodopsin, Biochemistry-US emphasis on bacteriorhodopsin as a typical and prototype membrane protein, 41 (2002) 5348–5358. Biochim. Biophys. Acta 1768 (2007) 3145–3161. [45] T. Friedrich, S. Geibel, R. Kalmbach, I. Chizhov, K. Ataka, J. Heberle, M. Engelhard, E. [20] A.T. Petkova, Y. Ishii, J.J. Balbach, O.N. Antzutkin, R.D. Leapman, F. Delaglio, R. Bamberg, Proteorhodopsin is a light-driven proton pump with variable vectori- Tycko, A structural model for Alzheimer's beta-amyloid fibrils based on ality, J. Mol. Biol. 321 (2002) 821–838. experimental constraints from solid state NMR, Proc. Natl. Acad. Sci. U. S. A. 99 [46] R. Partha, R. Krebs, T.L. Caterino, M.S. Braiman, Weakened coupling of conserved (2002) 16742–16747. to the proteorhodopsin chromophore and its counterion implies [21] A.T. Petkova, R.D. Leapman, Z.H. Guo, W.M. Yau, M.P. Mattson, R. Tycko, Self- structural differences from bacteriorhodopsin, Biochim. Biophys. Acta-Bioenerg. propagating, molecular-level polymorphism in Alzheimer's beta-amyloid fibrils, 1708 (2005) 6–12. Science 307 (2005) 262–265. [47] D. Ikeda, Y. Furutani, H. Kandori, FTIR study of the retinal Schiff base and internal [22] C. Wasmer, A. Lange, H. Van Melckebeke, A.B. Siemer, R. Riek, B.H. Meier, Amyloid water molecules of proteorhodopsin, Biochemistry-Us 46 (2007) 5365–5373. fibrils of the HET-s(218-289) prion form a beta solenoid with a triangular [48] V.B. Bergo, O.A. Sineshchekov, J.M. Kralj, R. Partha, E.N. Spudich, K.J. Rothschild, J.L. hydrophobic core, Science 319 (2008) 1523–1526. Spudich, His-75 in proteorhodopsin, a novel component in light-driven proton [23] C.P. Jaroniec, C.E. MacPhee, V.S. Bajaj, M.T. McMahon, C.M. Dobson, R.G. Griffin, translocation by primary pumps, J. Biol. Chem. 284 (2009) 2836–2843. High-resolution molecular structure of a peptide in an amyloid fibril determined [49] N. Pfleger, A.C. Worner, J. Yang, S. Shastri, U.A. Hellmich, L. Aslimovska, M.S.M. by magic angle spinning NMR spectroscopy, Proc. Natl. Acad. Sci. U. S. A. 101 Maier, C. Glaubitz, Solid-state NMR and functional studies on proteorhodopsin, (2004) 711–716. Biochim. Biophys. Acta-Bioenerg. 1787 (2009) 697–705. [24] J.J. Helmus, K. Surewicz, P.S. Nadaud, W.K. Surewicz, C.P. Jaroniec, Molecular [50] L. Shi, M.A.M. Ahmed, W. Zhang, G. Whited, L.S. Brown, V. Ladizhansky, Three- conformation and dynamics of the Y145Stop variant of human prion protein, Proc. dimensional solid-state NMR study of a seven-helical integral membrane proton Natl. Acad. Sci. U.S.A. 105 (2008) 6284–6289. pump—structural insights, J. Mol. Biol. 386 (2009) 1078–1093. [25] F. Castellani, B. van Rossum, A. Diehl, M. Schubert, K. Rehbein, H. Oschkinat, [51] L. Shi, X. Peng, M.A. Ahmed, D.A. Edwards, L.S. Brown, V. Ladizhansky, Resolution Structure of a protein determined by solid-state magic-angle-spinning NMR enhancement by homonuclear J-decoupling: application to three-dimensional spectroscopy, Nature 420 (2002) 98–102. solid-state magic angle spinning NMR spectroscopy, J. Biomol. NMR 41 (2008) [26] T.I. Igumenova, A.J. Wand, A.E. McDermott, Assignment of the backbone 9–15. resonancesformicrocrystallineubiquitin,J.Am.Chem.Soc.126(2004) [52] W.T. Franks, K.D. Kloepper, B.J. Wylie, C.M. Rienstra, Four-dimensional hetero- 5323–5331. nuclear correlation experiments for chemical shift assignment of solid proteins, [27] T.I. Igumenova, A.E. McDermott, K.W. Zilm, R.W. Martin, E.K. Paulson, A.J. Wand, J. Biomol. NMR 39 (2007) 107–131. Assignments of carbon NMR resonances for microcrystalline ubiquitin, J. Am. [53] A. Pines, M.G. Gibby, J.S. Waugh, Proton-enhanced NMR of dilute spins in solids, Chem. Soc. 126 (2004) 6720–6727. J. Chem. Phys. 59 (1973) 569–590. 2574 L. Shi et al. / Biochimica et Biophysica Acta 1788 (2009) 2563–2574

[54] S.R. Hartmann, E.L. Hahn, Nuclear double resonance rotating frame, Phys. Rev. 128 [69] H. Luecke, B. Schobert, J. Stagno, E.S. Imasheva, J.M. Wang, S.P. Balashov, J.K. (1962) 2042–2053. Lanyi, Crystallographic structure of xanthorhodopsin, the light-driven proton [55] M. Baldus, A.T. Petkova, J. Herzfeld, R.G. Griffin, Cross polarization in the tilted pump with a dual chromophore, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) frame: assignment and spectral simplification in heteronuclear spin systems, Mol. 16561–16565. Phys. 95 (1998) 1197–1207. [70] G. Metz, F. Siebert, M. Engelhard, High-resolution solid-state C-13 NMR of [56] B.M. Fung, A.K. Khitrin, K. Ermolaev, An improved broadband decoupling bacteriorhodopsin—characterization of [4-C-13]Asp resonances, Biochemistry-US sequence for liquid crystals and solids, J. Magn. Reson. 142 (2000) 97–101. 31 (1992) 455–462. [57] C.R. Morcombe, K.W. Zilm, Chemical shift referencing in MAS solid state NMR, [71] G. Metz, F. Siebert, M. Engelhard, Asp(85) is the only internal aspartic-acid that J. Magn. Reson. 162 (2003) 479–486. gets protonated in the M-intermediate and the purple-to-blue transition of [58] J. Cavanagh, W.J. Fairbrother, A.G. Palmer, N.J. Skelton, Protein NMR Spectroscopy: bacteriorhodopsin—a solid-state C-13 CP-MAS NMR investigation, FEBS Lett. 303 principles and practice, Academic Press, San Diego, 1996. (1992) 237–241. [59] J.L. Markley, A. Bax, Y. Arata, C.W. Hilbers, R. Kaptein, B.D. Sykes, P.E. Wright, K. [72] V. Bergo, J.J. Amsden, E.N. Spudich, J.L. Spudich, K.J. Rothschild, Structural changes Wuthrich, Recommendations for the presentation of NMR structures of proteins in the photoactive site of proteorhodopsin during the primary photoreaction, and nucleic acids—IUPAC-IUBMB-IUPAB inter-union task group on the standard- Biochemistry-US 43 (2004) 9075–9083. ization of data bases of protein and nucleic acid structures determined by NMR [73] H. Saito, J. Mikami, S. Yamaguchi, M. Tanio, A. Kira, T. Arakawa, K. Yamamoto, S. spectroscopy, Eur. J. Biochem. 256 (1998) 1–15. Tuzi, Site-directed C-13 solid-state NMR studies on membrane proteins: strategy [60] F. Delaglio, S. Grzesiek, G.W. Vuister, G. Zhu, J. Pfeifer, A. Bax, NMRPIPE—a and goals toward revealing conformation and dynamics as illustrated for multidimensional spectral processing system based on UNIX pipes, J. Biomol. bacteriorhodopsin labeled with [1-C-13]amino acid residues, Magn. Reson. NMR. 6 (1995) 277–293. Chem. 42 (2004) 218–230. [61] R. Keller, The Computer Aided Resonance Assignment Tutorial, first ed. CANTINA [74] L.S. Brown, K.H. Jung, Bacteriorhodopsin-like proteins of eubacteria and fungi: the Verlag, Goldau, 2004. extent of conservation of the haloarchaeal proton-pumping mechanism, Photo- [62] H. Heise, W. Hoyer, S. Becker, O.C. Andronesi, D. Riedel, M. Baldus, Molecular-level chem. Photobiol. Sci. 5 (2006) 538–546. secondary structure, polymorphism, and dynamics of full-length alpha-synuclein [75] M. Baldus, B.H. Meier, Total correlation spectroscopy in the solid state. The use of fibrils studied by solid-state NMR, Proc. Natl. Acad. Sci. U. S. A. 102 (2005) scalar couplings to determine the through-bond connectivity, J. Magn. Reson. Ser. 15871–15876. A. 121 (1996) 65–69. [63] M.J.S. Kelly, C. Krieger, L.J. Ball, Y.H. Yu, G. Richter, P. Schmieder, A. Bacher, H. [76] O.C. Andronesi, S. Becker, K. Seidel, H. Heise, H.S. Young, M. Baldus, Determination Oschkinat, Application of amino acid type-specific H-1-and N-14-labeling in a H- of membrane protein structure and dynamics by magic-angle-spinning solid- 2-, N-15-labeled background to a 47 kDa homodimer: potential for NMR structure state NMR spectroscopy, J. Am. Chem. Soc. 127 (2005) 12965–12974. determination of large proteins, J. Biomol. NMR 14 (1999) 79–83. [77] L. Zhong, V.V. Bamm, M.A. Ahmed, G. Harauz, V. Ladizhansky, Solid-state NMR [64] B.Q. Sun, C.M. Rienstra, P.R. Costa, J.R. Williamson, R.G. Griffin, 3D N-15-C-13-C-13 spectroscopy of 18.5 kDa myelin basic protein reconstituted with lipid vesicles: chemical shift correlation spectroscopy in rotating solids, J. Am. Chem. Soc. 119 Spectroscopic characterisation and spectral assignments of solvent-exposed (1997) 8540–8546. protein fragments, Biochim. Biophys. Acta 1768 (2007) 3193–3205. [65] M. Hong, Resonance assignment of C-13/N-15 labeled solid proteins by two- and [78] L.L. Chen, R.A. Olsen, D.W. Elliott, J.M. Boettcher, D.H.H. Zhou, C.M. Rienstra, L.J. three-dimensional magic-angle-spinning NMR, J. Biomol. NMR 15 (1999) 1–14. Mueller, Constant-time through-bond C-13 correlation spectroscopy for assigning [66] C.M. Rienstra, M. Hohwy, M. Hong, R.G. Griffin, 2D and 3D N-15-C-13-C-13 NMR protein resonances with solid-state NMR spectroscopy, J. Am. Chem. Soc. 128 chemical shift correlation spectroscopy of solids: assignment of MAS spectra of (2006) 9992–9993. peptides, J. Am. Chem. Soc. 122 (2000) 10979–10990. [79] L. Chen, J.M. Kaiser, T. Polenova, J. Yang, C.M. Rienstra, L.J. Mueller, Backbone [67] J.E. Masse, R. Keller, AutoLink: automated sequential resonance assignment of assignments in solid-state proteins using J-based 3D Heteronuclear correlation biopolymers from NMR data by relative-hypothesis-prioritization-based simu- spectroscopy, J. Am. Chem. Soc. 129 (2007) 10650–10651. lated logic, J. Magn. Reson. 174 (2005) 133–151. [80] J.D. Bendtsen, H. Nielsen, G. von Heijne, S. Brunak, Improved prediction of signal [68] M. Hiller, V.A. Higman, S. Jehle, B.J. van Rossum, W. Kuhlbrandt, H. Oschkinat, [2,3- peptides: SignalP 3.0, J. Mol. Biol. 340 (2004) 783–795. (13)C]-labeling of aromatic residues—getting a head start in the magic-angle- [81] H. Nielsen, J. Engelbrecht, S. Brunak, G. vonHeijne, Identification of prokaryotic spinning NMR assignment of membrane proteins, J. Am. Chem. Soc. 130 (2008) and eukaryotic signal peptides and prediction of their cleavage sites, Protein Eng. 408–409. 10 (1997) 1–6.