Advent of Recombinant Expression and Purification Technology

In contrast to other GPCRs, rhodopsin is readily available from native sources (bovine retina) in very high quantities and this source proved useful in limited early structural studies.31 Significant structural information was obtained also from unpu-rified systems expressing GPCRs, and such studies have played important roles in studies of ligand binding pockets of GPCRs other than rhodopsin. However, high resolution structural studies require significant amounts of recombinant pure protein. A variety of efficient recombinant systems were tested and systematically evaluated for production of rhodopsin, including Escherichia coli, yeasts, oocytes, baculovirus insect cell systems, and mammalian COS, CHO, and HEK293 cells. In E. coli, negligible amounts of rhodopsin were expressed or the expressed protein was mis-folded and not functional.32

Two yeast strains were tested for expression of rhodopsin. In Saccharomyces cerevisiae, 2.0 ± 0.5 mg/L were expressed, of which only 2 to 4% were correctly folded.33 In Pichia pastoris, 0.3/L were produced, of which 4 to 15% were reported functional.34 Rhodopsin is expressed in functional form in Xenopus laevis oocytes35 and in Chinese hamster ovarian (CHO) cells,36 although yields have not been reported. For small scale (microgram range) preparation of rhodopsin, transient transfection in the COS cell system37 is the most widely used method. Typical wildtype expression yields are 25 |j,g protein/15-cm dish.

For production of larger quantities of functional rhodopsin, expression in baculovirus Sf9 and mammalian HEK293 cells has been successful. Whereas insect cell systems produced yields of ~1 mg/L rhodopsin, mammalian cell systems have been the mainstays in terms of providing large quantities of active rhodopsin for biophysical analyses. Stable mammalian cell lines expressing wild-type bovine rhodopsin were originally produced using a suspension-adapted HEK293 cell line and a vector system placing the opsin gene under the control of a constitutively enhanced cytomegalovirus promoter system.38

Yields of wild-type rhodopsin from this system produced in large-scale spinner flask preparations reached ~2 mg/L and this method has been extensively applied to the production of rhodopsin samples suitable for NMR analysis.38-35 Recent optimizations of this stable cell line method include accommodation of a tetracy-cline-inducible promoter system, increasing yields five-fold, achieving levels of ~10 mg/L from bioreactors with enhanced nutrient mixtures, and permitting overexpression of a toxic, constitutively active rhodopsin mutant.46,47

Tetracycline induction also allowed production of nonglycosylated rhodopsin (important for the preparation of samples amenable to biophysical analyses such as protein crystallization and nuclear magnetic resonance) through addition of tunica-mycin immediately prior to induction; although cell viability dropped after only 24 h, overall yields of nonglycosylated rhodopsin were ~3 mg/L.47 Alternatively, a modified (N-acetylglucosamine transferase I-negative) HEK293S cell line with restricted glycosylation engineered and combined with the stable tetracycline induction system yielded more than ~6 mg/L of nonglycosylated rhodopsin.4748 Balancing cost and efficiency of functional expression, the COS and HEK293 cell expression

Cytoplasmic msnfr qE ™Mv

FIGURE 11.2 Amino acids on the cytoplasmic surface of rhodopsin studied by cysteine mutagenesis. Highlighted in black letters are all residues that have been mutated to cysteines one at a time (see single cysteine mutants in Table 11.1, top). All other amino acids are omitted from this graph, except those in the C-terminal tail. Residues studied by cysteine mutagenesis are shown as black letters; those that have not been studied are shown as gray letters. Cysteines that have served as fixed reference points in di-cysteine mutants (see text) are marked in bold. These cysteines were fixed in mutants, where a second cysteine was systematically varied. The resulting pairs are listed in Table 11.1, bottom.

Cytoplasmic msnfr qE ™Mv

II III IV V VI VII

FIGURE 11.2 Amino acids on the cytoplasmic surface of rhodopsin studied by cysteine mutagenesis. Highlighted in black letters are all residues that have been mutated to cysteines one at a time (see single cysteine mutants in Table 11.1, top). All other amino acids are omitted from this graph, except those in the C-terminal tail. Residues studied by cysteine mutagenesis are shown as black letters; those that have not been studied are shown as gray letters. Cysteines that have served as fixed reference points in di-cysteine mutants (see text) are marked in bold. These cysteines were fixed in mutants, where a second cysteine was systematically varied. The resulting pairs are listed in Table 11.1, bottom.

systems have proven to be the most useful in terms of providing microgram to hundreds of milligram quantities of active rhodopsin, respectively. A refined antibody affinity purification scheme allowed one-step purification of rhodopsin from native and recombinant systems.37 Other alternative affinity chromatography methods include concanavalin A affinity purification of rhodopsin from natural sources and immobilized metal affinity chromatography in correlation with insect cell expression work.1249-51 An in-depth review of expression of GPCRs other than rhodopsin can be found in Chapter 8 of this book.

11.2.3 Rhodopsin Structure via Cysteine Mutagenesis

Followed by Biochemical and Biophysical Analyses

The structure and dynamics of the CP loop regions and CP ends of the helices of rhodopsin in solution were investigated using a combination of cysteine mutagenesis followed by biochemical and biophysical studies of the cysteine mutants. Figure 11.2 shows the residues on the CP side of the molecule that were replaced, one at a time, by cysteine. The top section of Table 11.1 lists the single cysteine mutants. In separate experiments, pairs of cysteines (di-cysteine mutants) were introduced, as also shown in Figure 11.2 and Table 11.1.

The unique chemistry of the cysteine sulfhydryl group allows specific derivati-zation of reactive and accessible cysteines with biophysical probes. Because dark-state, wild-type rhodopsin contains two reactive cysteines, Cys140 and Cys316, these cysteines were replaced by serines in the cysteine mutants to avoid ambiguity. The other cysteines in rhodopsin (palmitoylation sites, TM and EC cysteines, see Figure 11.1B) are not reactive in the dark and were therefore not replaced. Figure 11.3A

TABLE 11.1

Cysteine Mutants of Rhodopsin Studied via Biochemical and Biophysical Means

Mutant Type

Reference

Single Cysteine

Asn55 - Ile75 Tyr136 - Met155 Gln225 - Ile256 Tyr306 - Leu321

Lys325, Asn326, Leu328, Asp331, Glu332, Thr335-Thr340

70, 131

Di-Cysteine

Cysteine I (fixed)

Cysteine II (varied) Cys60-Cys74

Cys139 Cys135 Cys338

Cys316 Cys65

Cys246 or Cys250

Cys306-Cys321 Cys311-Cys314 Cys247-Cys252 Cys250

Cys240-Cys250

56 62 119 62 70

lists some of the biochemical and biophysical probes used to study the structure and dynamics of rhodopsin.

11.2.3.1 Cysteine Reactivity as Qualitative Indicator for Structure

Rhodopsin carrying a free sulfhydryl group reacts with 4,4'-dithiodipyridine to form the thiopyridinyl molecular entity derivative shown in Figure 11.3A, molecular entity

I.52 The rate of this reaction is very sensitive to the accessibility of the cysteine. Cysteines buried in the micelle or protein interior are entirely unreactive, and cysteines exposed in the fully accessible aqueous portion of the CP domain react instantly.53 This is shown for cysteines placed at positions 56 through 75 in Figure

II.3B, Panel I. The positions of these exposed cysteines correspond well with those identified by collisions of paramagnetic agents with spin-labeled derivatives (Figure 11.3A, molecular entity 3).54

The resulting thiopyridinyl derivative of rhodopsin (1) is reactive toward free sulf-hydryl reagents by way of disulfide exchange. The rate of this exchange is an extremely sensitive probe for tertiary structure and has been used to prove the presence of light-induced conformational motions resulting in distinct tertiary structure changes,53 even where EPR methods suggest the magnitude of conformational change to be small.54 For example, changes in the rates of reactions in the dark as compared to the light are shown in Figure 11.3B, Panel II, for residues in the 60 through 74 sequence.

The disulfide exchange reaction of thiopyridinyl-rhodopsin (1) with sulfhydryl reagents has practical value in that the nature of the sulfhydryl reagent can be chosen relatively freely, thereby presenting a general scheme for derivatization of cysteine groups on proteins with reporter probes for which no reactive derivatives exist. For

FIGURE 11.3 Biophysical probes used in studies of structure and dynamics of full-length rhodopsin. A: Molecular entities in various derivatization reactions of cysteines. The measurement of cysteine reactivity using 4,4'-dithiodipyridine yielding thiopyridinyl-cysteine derivative in process I, the exchange reaction with sulfhydryl reagents in process II, and the EPR spin-labeling procedure in process III yield molecular entities 1, 2, and 3, respectively. B: Examples for the results of processes I through III from A for the loop connecting helices I and II in rhodopsin. Panel I: Measurement of chemical reactivity of cysteines using 4,4'-dithiodipyridine.53 Panel II: Release of thiopyridone via disulfide exchange using sulfhydryl reagents. Panel III: EPR spectra of spin-labeled cysteine derivatives.54 (Reproduced with permission from Biochemistry. Copyright, American Chemical Society, Washington, D.C., 1999.)

FIGURE 11.3 Biophysical probes used in studies of structure and dynamics of full-length rhodopsin. A: Molecular entities in various derivatization reactions of cysteines. The measurement of cysteine reactivity using 4,4'-dithiodipyridine yielding thiopyridinyl-cysteine derivative in process I, the exchange reaction with sulfhydryl reagents in process II, and the EPR spin-labeling procedure in process III yield molecular entities 1, 2, and 3, respectively. B: Examples for the results of processes I through III from A for the loop connecting helices I and II in rhodopsin. Panel I: Measurement of chemical reactivity of cysteines using 4,4'-dithiodipyridine.53 Panel II: Release of thiopyridone via disulfide exchange using sulfhydryl reagents. Panel III: EPR spectra of spin-labeled cysteine derivatives.54 (Reproduced with permission from Biochemistry. Copyright, American Chemical Society, Washington, D.C., 1999.)

example, it was possible to derivatize rhodopsin with the trifluoroethylthiol group resulting in molecular entity (2), Figure 11.3A, to enable 19F-NMR applications.42,43,45

11.2.3.2 Distance Constraints by Rates of Disulfide Bond Formation

To further define tertiary structure in the CP domain of rhodopsin, disulfide bond formation was used to establish proximity between amino acids in rhodopsin,55,56 a method developed by Falke and Koshland57 in studies of the aspartate receptor.58,59 The method is based on structural analysis of disulfides occurring in protein crystal structures that have shown preferred conformations for their formation.60 The distance between a-carbon atoms across the disulfide ranges from about 4 to 9 A in crystal structures, with 95% of all refined disulfides in the range of 4.4 to 6.8 A. The average distances across left-handed and right-handed disulfides is 5.88 ± 0.49 A and 5.07 ± 0.73 A,61 respectively.

Thus, the presence of a disulfide bond indicates that the a-carbons of the participating cysteines are about 5 to 6 A apart. However, the geometry derived from crystallography may not hold in solution, especially for mobile regions at protein surfaces.61 The formation of a disulfide bond between two cysteines does not imply a time-average proximity of the two residues in the protein structure. Once the disulfide bond is formed, the two cysteines are locked in a conformation that may not necessarily be favored.

The rates of disulfide bond formation were determined for sets of di-cysteine mutants of rhodopsin, in which a cysteine was kept constant at one site while the position of a second cysteine was varied at a proximal region (Table 11.1). The rates of disulfide bond formation were measured for different sets of di-cysteine mutants62 and compared to proximities between amino acids deduced from a crystallographic model.8 The reciprocal of the distances obtained as a function of residue position and the rates of disulfide bond formation are reproduced in Figure 11.4A for disulfide bonds between Cys316 in helix VIII at CP loop 4 and cysteines at positions 55 through 75 in CP loop 1.55 The comparison showed excellent correlation between the rates of disulfide bond formation and the interthiol distances derived from the cysteine replacements in the crystal structure (Figure 11.4B). The three positions that most rapidly formed disulfide bonds with Cys316, H65C, L68C, and V61C, are 4 to 5 A distant from Cys316, and facing it. In order for a disulfide bond to form, however, 3 to 4 A translational movements would be necessary. This requires sufficient flexibility of the amino acids in this region of the rhodopsin CP face. The fact, however, that only those cysteines that faced Cys316 were able to bridge this small gap indicates that no unfolding of the ends of the helices occurs; instead, a movement of intact helices brings residues in CP loop 1 close to Cys316. These results presented the first direct evidence for substantive backbone motion in the CP domain of rhodopsin in solution.

11.2.3.3 Mobility, Accessibility, and Distance Constraints via Electron Paramagnetic Resonance Spectroscopy

Application of EPR spectroscopy to rhodopsin was reviewed earlier.63,64 To enable application of EPR spectroscopy, nitroxide spin labels (molecular entity 3, Figure

FIGURE 11.4 Disulfide bond formation in di-cysteine mutants Tyr60Cys-Tyr74Cys/Cys316.55 A: Rates of disulfide bond formation at pH 7.7, 25oC (dotted lines, open circles). The reciprocals of distances between cysteines in the crystal structure of rhodopsin are shown as black lines, filled circles. (Reproduced with permission from Biochemistry. Copyright, American Chemical Society, Washington, D.C., 1999.) B: Visualization of positions of cysteines placed at sites where fastest disulfide bonds were formed with Cys316 (Leu68, His65 and Val61). All these positions are in closest proximity to Cys316 in the crystal structure model.

FIGURE 11.4 Disulfide bond formation in di-cysteine mutants Tyr60Cys-Tyr74Cys/Cys316.55 A: Rates of disulfide bond formation at pH 7.7, 25oC (dotted lines, open circles). The reciprocals of distances between cysteines in the crystal structure of rhodopsin are shown as black lines, filled circles. (Reproduced with permission from Biochemistry. Copyright, American Chemical Society, Washington, D.C., 1999.) B: Visualization of positions of cysteines placed at sites where fastest disulfide bonds were formed with Cys316 (Leu68, His65 and Val61). All these positions are in closest proximity to Cys316 in the crystal structure model.

11.3A) are introduced either at single reactive cysteine sites (single cysteine mutants) or in pairs (di-cysteine mutants); see Section 11.2.3. EPR spectroscopy of spin-labeled single cysteine mutants is useful for estimating mobility and for obtaining qualitative indicators for tertiary structure contacts surrounding the nitroxides and the cysteines to which the nitroxides are attached. For example, EPR spectra of spin labels at positions 59 through 75 are shown in Figure 11.3B, Panel III. In conjunction with collision measurements, EPR spectroscopy can provide a measure of water accessibility. Applications of these methods to rhodopsin enabled derivation of the orientations of helices, protrusions of helices into the CP surface, relative flexibility of different loops, and qualitative assessment of conformational changes upon light activation.54,65-67 To obtain more quantitative information on conformational changes, di-cysteine mutants were studied. Application of EPR spectroscopy in the presence of two spin labels allowed detection of dipolar interactions among the labels. This resulted in derivation of distance constraints for the sets of di-cysteines listed in Table 11.1. Most significantly, changes in distances upon light activation allowed the derivation of a model for the mechanism of activation of rhodopsin.67-70 The relative increase in the distances between the CP ends of helices suggested an opening of the helical bundle toward the CP face, exposing amino acids critical for interaction with the G protein. Although distance changes were observed in the majority of sites studied, the largest changes in relative distance were observed between helices III and VI, followed by smaller changes in relative distances between helices I and II with respect to helix VII and helix VIII.6364

11.2.4 Solid-State Nuclear Magnetic Resonance Applications to Rhodopsin

11.2.4.1 Chromophore Structure and Schiff Base Linkage

The advent of 13C magic angle spinning (MAS) solid-state NMR provided significant improvements over the original early suspension NMR experiments through increased line resolution by resolving broad chemical shift anisotropy powder patterns into sharp center bands at the isotropic chemical shift with side bands at the spinning frequency.71-73 The first applications of 13C cross polarization (CP) MAS to rhodopsin samples with the aim of observing protein-lipid interactions were limited by rotational diffusion of rhodopsin molecules in the bilayer on a time scale that interfered with decoupling and reduced cross-polarization efficiency.7475

Spurred by success with a motionally restricted bacteriorhodopsin system,76 two reports later demonstrated that restriction of rhodopsin rotational diffusion by application of low temperatures75 or lyophilization77 slowed motion sufficiently to permit effective application of CP MAS to analysis of chromophore structure. Interactions in rhodopsin provided evidence that the Schiff base C-N double bond linkage is protonated and in the anti conformation. Furthermore, these experiments supported the presence of negative counter ion(s) in close proximity to the retinyl moiety.

Ultimately, a combination of molecular orbital calculations, site-directed mutagenesis, and chemical shifts in 13C -MAS spectra for 11-cis retinal carbons physically localized the counter ion to a carboxylate group situated above C12 of the chromophore in rhodopsin.7879 Specifically one of the carboxylate oxygens (O1) of Glu113 was suggested to be ~3A from the C12 position.8081 Combining this with further analyses of the mechanism of the opsin shift,82 the mechanism of energy storage in bathorhodopsin,8083 the unprotonated form of the Schiff base in the meta II rhodopsin intermediate84 and a variety of site-directed mutagenesis studies85-88 provided a coherent picture of the photoactivation mechanism of rhodopsin.89

Subsequently, a novel method for macroscopic alignment of bilayers on solid surfaces termed isopotential spin-dry ultracentrifugation (ISDU) was applied to rhodopsin preparations in native and reconstituted membranes.90 Angles relative to the membrane normal for individual 2H bonds in the chromophore were determined by this method. However, combining the ISDU-oriented bilayer method with 2H MAS NMR91 modified the experiment such that side band intensity became a function of size and orientation of the quadrupolar coupling tensor, allowing more precise measurement of the C-C(2H)3 bond vector orientation with respect to the membrane normal. Such magic angle-oriented sample spinning experiments were applied to labeled chromophores in reconstituted rhodopsin membranes and ultimately provided a three-dimensional structure and orientation of retinal within the binding pocket of rhodopsin in the ground state and upon photoactivation to the meta I intermediate.

Additional details about the chromophore structures were later solved by applications of one-dimensional CP MAS NMR rotational resonance measurement of distances between 13C labeled carbons as well as a novel double quantum hetero-nuclear local field NMR method for determination of torsion angles about 13C-labeled atoms.92-95 Distance values and torsion angles confirmed the 11-cis conformation of retinal in the dark state of rhodopsin with a twisted C10-C13 unit. A relaxed planar all-irans conformation in the meta I rhodopsin intermediate was also confirmed. Most recently, these methods provided corrections to the magic angle oriented spinning model91 of the 11-cis ground state chromophore such that a modest 28-degree twist is present in the C6-C7 torsion angle and the ring remains in the binding site in the meta I intermediate.96,97

As described earlier, dramatic progress was achieved in the field of high-yield recombinant expression of rhodopsin. This progress led to the first use of protein site-specific isotopically labeled recombinantly produced rhodopsin for MAS NMR experiments.

Homogeneously [a,e-15N2]-lysine labeled rhodopsin from a baculovirus Sf9 cell expression system was subjected to CP MAS 15N NMR to investigate the Schiff base linkage.9899 The nitrogen of the lysine contributing to the Schiff base linkage was observed to resonate with an isotropic shift signal at 155.9 ppm relative to ammonium chloride.99 This spectrum was correlated with the model of a protonated Schiff base stabilized by a complex counterion. Homogeneously a,e-15N-lysine and 2-13C-glycine di-labeled rhodopsin produced by overexpression in a stably trans-fected mammalian suspension cell line (HEK293S) was also used to investigate Schiff base linkage.39-40 The 15N chemical shift of the lysine contributing to the Schiff base was observed at 156.8 ppm, a chemical shift characteristic of a protonated Schiff base nitrogen that has only weak counterion interactions.

The distance between the Schiff base nitrogen and the counterion was estimated at approximately 4.0 A. This distance is consistent with a structural water molecule in the binding site hydrogen bonded with both the Schiff base nitrogen and the Glu113 counterion. These reports demonstrate the feasibility of applying solid-state MAS NMR to structural studies on isotopically labeled rhodopsin and potentially other GPCRs.

11.2.4.2 Other Tertiary Interactions

Noncovalent Protein-Chromophore Interactions — Aside from the obvious cova-lent linkage of the chromophore to the rhodopsin molecule by the protonated Schiff base and counterion interaction, the chromophore is expected to make other significant noncovalent interactions with the TM helices as alluded to in early 19F studies.100 Some of these interactions were identified through combinations of site-directed mutagenic analysis and 13C MAS NMR of labeled chromophore and computational methods.89 A model was proposed that highlights direct steric interactions between the retinal chromophore and TM helix III in the region of Gly121 that are modulated during the conversion of 11-cis to all-irans retinal. Subsequently, 11-cis chromophore labeled with 13C at ten sites along most of the entire polyene chain or homogeneously was incorporated into native rhodopsin embedded in membranes.101,102

Application of 2D experiments was made possible by use of a 2D radio frequency-driven dipolar recoupling during the MAS experiments. Chemical shift dif ferences between the free and bound ligand provided an NMR assay of the spatial and electronic structures of the chromophore for assessment of interactions between the chromophore and the binding site including protein-induced conformational restraints or electronic effects. Nonbonding interactions between the protons of the methyl groups of the p-ionone ring and the protein were observed. The interactions were attributed to aromatic residues Phe208, Phe212 (interacting with H16 and H17), and Trp265 (interacting with H18) determined to be in close proximity by crystal-lographic methods.8103

Protein-Lipid Interactions — Rhodopsin is naturally packed and aligned in the ROS disk membranes in a regular parallel array of a-helices such that the overall intact ROS has an average diamagnetic anisotropy allowing orientation of intact ROS in solution with its long axis parallel to even very low magnetic fields.104 105 Observation of 31P spectra of such magnetically oriented intact ROS provided much more specific information than that obtained from earlier sonicated or reconstituted suspension samples (see Section 11.2.1), including resolution of the main phospholipid pools and metabolite phosphate groups present in these relatively complex biological systems.106

Rhodopsin Phosphorylation — Phosphorylation of activated rhodopsin is the first step in desensitization of the receptor. A comparison of 31P spectra collected for phosphorylated and unphosphorylated ROS demonstrated the presence of a 31P resonance specific for the phosphorylated sample.107 Limited proteolysis of the C-terminus of rhodopsin, which removes the predicted phosphorylation sites, eliminated the resonance and determination of the pKa obtained from pH titration, indicating it was likely a serine phosphate.

Single pulse, CP, and rotational echo double resonance (REDOR) MAS techniques were applied to the analysis of phosphorylated rhodopsin in ROS mem-branes.42 The single pulse and CP methods were used to probe the temperature dependence of the 31P signal, which indicated that the C-terminus was highly mobile. REDOR analyses were then applied to measure internuclear distances between the 31P label on the C-terminal tail and 19F labels specifically incorporated (as described in Section 11.2.3.1) at Cys140 (helix III) and Cys316 (helix VIII) of the rhodopsin CP face of the native wild-type (WT) phosphorylated rhodopsin. 31P-19F distances were all estimated to be greater than 12 A, suggesting that the rhodopsin C-terminus does not contact the areas of the CP face implicated in arrestin binding. This implies a change in the average position of the C-terminal tail upon phosphorylation from positions determined by other methods.8,42

11.2.5 Solution NMR

While solution NMR was used in some of the earliest studies of rhodopsin-chro-mophore interactions (Section 11.2.1), overall resolution was very poor for the sonicated membrane suspensions and NMR efforts for the most part shifted to solidstate methods. However, with the advent of higher field instruments, novel pulse sequences, and recombinant expression technology, solution NMR can be applied to structural analyses of rhodopsin in a number of interesting ways.

11.2.5.1 Protein-Protein Interactions

Solution NMR was initially used to study the interaction between rhodopsin and the G protein a subunit known as transducin108 via application of a transferred nuclear Overhauser effect (Tr-NOESY)109 experiment to a system containing a fragment peptide of transducin (residues 340 through 350 with Lys341Arg) known to interact with rhodopsin.108 The observed NOEs provided information on distances between various protons in the bound peptide, which along with distance geometry, NOE-constrained molecular dynamics, simulated annealing and grid searches produced a series of structures of the transducin peptide in association with and the excited meta II intermediate.

Although a number of problems including low signal-to-noise and assignment errors were later identified in this innovative work,110 a similar analysis using Tr-NOESY and TOCSY (total correlation spectroscopy) experiments to compare free transducin peptide (340 to 350) to the meta II intermediate bound form was suc-cessful.111 Disordered peptide was observed in solution which, upon light activation, underwent a dramatic shift in conformation to include a helical turn followed by an open reverse turn along the peptide chain. Comparison of this structure was made to the same segment identified in the crystal structure of transducin, which indicated a mechanism of signal transduction by which activated receptors control G proteins through reversible conformational changes.111115 This model was later confirmed by measurement of residual dipolar couplings.116117

Magnetically oriented bilayer 31P studies indicated that ROS disk membranes, which alone lack the cooperativity in alignment that exists in the intact ROS, actually have sufficiently large total magnetic susceptibility anisotropy to align with their membrane normals parallel to a magnetic field.117 Residual dipolar contributions were observed (identified by smaller 1JNH splittings in 1H-15N HSQC spectra) between light-activated rhodopsin (meta II state) in oriented ROS disk membranes and a selectively 15N di-labeled undecapeptide, closely resembling a fragment (residues 340 to 350) of the transducin G protein a subunit known to interact with activated rhodopsin.117 Information on the conformation of the bound ligand and its orientation relative to the bilayer was determined from the measurement of N-H vector angles of 48 and 40 degrees with the disk normal for the labeled amino acids (Leu5 and Gly9, respectively). Transferred dipolar contributions were observed to return to their isotropic values at a rate determined by the decay of the meta II state of rhodopsin.

This magnetically oriented bilayer NMR system provided significantly more conformational information than in the transferred-NOE experiments.108111 Ultimately, this oriented bilayer-transferred dipolar coupling method (including observation of contributions to both 1JNH and 1JCH splittings), combined with determination of approximate distance restraints by transferred NOEs, was used to solve a high resolution structure of the G protein fragment.116 The fragment was shown to take on a helical conformation upon binding to light-activated rhodopsin.

11.2.5.2 Rhodopsin Structure

The positions and widths of 19F-NMR signals are indicators of chemical environment and mobility118 and are therefore sensitive to light-induced conformational changes in rhodopsin.43 The first solution NMR experiment analyzing the overall structure of full-length rhodopsin was carried out by application of 19F NMR to detergent-solubilized recombinant rhodopsin.43 A number of mutants (produced in HEK293S cells) containing single reactive cysteine residues in different regions of CP faces were labeled with 19F as described above.43 Each of these labeled derivatives specifically located at amino acid residue positions 67, 140, 245, 248, 311, and 316 of the CP face was purified in sufficient quantities for NMR analysis.

Recorded solution NMR spectra of the ground state rhodopsin resolved different chemical shifts for each of the 19F labels with differing line widths representing the different environments and mobilities across the CP surface (Figure 11.5). Upon illumination to produce the meta II intermediate, significant upfield and downfield shifts of the 19F resonances were observed, in particular those at amino acid positions 67 and 140 as well as at 248 and 316, respectively.43 These shifts were entirely consistent with previous models of rhodopsin activation based on EPR analyses

I. Cys-140

II. Cys-316

II. K311C

IV. K245C

V. K67C

VI. K248C

10.6 ppm

10.2 ppm

10.6 ppm

10.2 ppm

10.1 ppm

0.0 ppm

10.1 ppm

0.0 ppm

9.9 ppm A^pm

9.5 ppm

11.511.010.510.0 9.5 9.0 8.5 Chemical shift (ppm relative to TFA)

FIGURE 11.5 [19F]NMR spectra of single cysteine mutants.43 Single cysteine mutants of rhodopsin were derivatized with molecular entity 2, shown in Figure 11.3A. A minimum of 5 mg pure protein was analyzed in dodecyl maltoside micelles. [19F]NMR was collected within 2-min acquisition time (average of 160 scans). Light gray lines indicate NMR spectra in the dark and dark gray lines indicate spectra after illumination.

9.5 ppm

11.511.010.510.0 9.5 9.0 8.5 Chemical shift (ppm relative to TFA)

FIGURE 11.5 [19F]NMR spectra of single cysteine mutants.43 Single cysteine mutants of rhodopsin were derivatized with molecular entity 2, shown in Figure 11.3A. A minimum of 5 mg pure protein was analyzed in dodecyl maltoside micelles. [19F]NMR was collected within 2-min acquisition time (average of 160 scans). Light gray lines indicate NMR spectra in the dark and dark gray lines indicate spectra after illumination.

(Section 11.2.3.3).119 Subsequently, this method was expanded to report on measurements of short distances in the CP surfaces by determination of NOEs between 19F labels in di-labeled rhodopsin mutants containing pairs of cysteines.45 Specifically, labels were incorporated at the ends of helices III and VIII (WT Cys140 and Cys316), helices I and VIII (Cys65 and Cys316), and helices III and VI (Cys139 and Cys251), respectively. Distinct chemical shifts were observed for all the labels in the dark state of rhodopsin.

Analysis of homonuclear 19F NOEs between the labeled cysteines indicated no evidence of proximity for the WT pair, but significant negative enhancements were observed for the other two pairs. The extent of the observed NOE enhancement correlated well with distances measured in the dark state crystal structure (29.8, 3.9, and 3.0 A, respectively) and emphasized the advantage of this method over EPR distance measurement that has a lower distance limit of 8 A, due in part to the size differences of the respective labels (molecular entities 2 and 3, Figure 11.3A).8 45 A discrepancy between observed rates of relaxation and theoretical calculated values is acknowledged and concise quantitation of the distances was not attempted by this method.

Yeagle and co-workers set about solving the complete structure of bovine rhodopsin by independently analyzing the structures of each individual loop and TM segment by solution NMR and assembling the conformational information into a final overall structure.120 The rationale for this method included observation of spontaneous assembly of the independent rhodopsin TM segments into a helical bundle in solution121 and successful application of this method to solving the structure of bacteriorhodopsin in comparison to its previously derived crystal struc-ture.122-125 Structures of all the individual TM and loop segments were solved by solution two-dimensional homonuclear *H NMR (2DQ-COSY and NOESY) in DMSO.126 127 Distance and dihedral angle constraints were obtained from the NMR data directly whereas long distance and interhelical constraints for the intact protein (for both ground and activated states) were obtained from previous structural studies of rhodopsin including methods such as electron microscopy, EPR, site-directed mutagenesis, cross-linking, and solid state NMR to produce the final model.120

The final calculated structure was in rough agreement with the crystal structure of ground state rhodopsin and included a more defined CP surface that was not well resolved in the crystal structure.8120 These experiments combined with modeling and information about the heterotrimeric G protein structure provided a calculated meta II intermediate structure.112,116 117 120 128-130 Differences between the ground and activated states, in particular in the CP face, suggest how that surface is activated to interact with the G protein. Overall, the authors presented a model in which a conformational distortion driven by the energy of binding is induced in transducin when it binds to the activated form of rhodopsin.129

11.2.5.3 Conformational Dynamics in Rhodopsin

Quantitative evidence for conformational fluctuations in rhodopsin came from a solution NMR study of rhodopsin labeled with [a-15N]-lysine (Figure 11.6A).44 Despite the presence of 11 lysines in rhodopsin, only a single sharp signal (signal

I I Cytoplasmic

111 IV V VI VII339

-oocTpavqstet© svt ts

,snfr 231

vvvc

150E

AAQ ak Ea

alw aay

pwq mf gge watmfv AC«.1

"fcy

v"

TV250

aed AspN

rTT EG F

i s ty lyyqPaef, AcN-m ngteg l

100h aec gyfvf p.

PlLwFc gaayv

300Y N

187 Nte

0 ee q cs gi h qc idyytp g hqgsdf

Intradiscal

0 0

Post a comment