| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,
* Department of Pharmacology, School of Medicine, Case Western Reserve University, Cleveland, Ohio, USA;
M.E. Müller Institute for Microscopy, Biozentrum, University of Basel, Basel, Switzerland; and
Institute of Biochemistry and Molecular Medicine, University of Berne, Berne, Switzerland
1Correspondence: Department of Pharmacology, School of Medicine, Case Western Reserve University, Wood Bldg., 10900 Euclid Ave., Cleveland, OH 44106-4965, USA. E-mail: bxj27{at}case.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
S. Removal of all-trans-retinylidene from Rho*-Gte had no effect on the stability of the Rhoe*-Gte complex. Moreover, opsin in the Rhoe*-Gte complex with an empty nucleotide-binding pocket in Gt and an empty retinoid-binding pocket in Rho was regenerated up to 75% without complex dissociation. These results indicate that once Rho* couples with Gt, the chromophore plays a minor role in stabilizing this complex. Moreover, in complexes regenerated with 9-cis-retinal/11-cis-retinal, Rho retains a conformation similar to Rho* that is stabilized by Gte apo-protein.—Jastrzebska, B., Golczak, M., Fotiadis, D., Engel, A., and Palczewski, K.. Isolation and functional characterization of a stable complex between photoactivated rhodopsin and the G protein, transducin.
Key Words: sucrose gradient • detergent • rhodopsin regeneration
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The structures of key phototransduction proteins such as rhodopsin (Rho) (2)
, photoactivated Rho (Rho*) (3)
, G
β protein (transducin, Gt-GDP) (4)
, arrestin (5)
, regulators of G-protein signaling (RGS) proteins (6)
, and other regulatory proteins have been solved and subsequently improved (7)
. However, a comprehensive understanding of phototransduction requires further structural work on complexes of the individual components. Signaling via the Rho*-Gtβ (Rho*-Gte) and other related Rho–Gt complexes is the focus of this study.
Absorption of a photon leads to isomerization of the Rho chromophore, 11-cis-retinylidene, to all-trans-retinylidene, and induces a conformational change in Rho that in turn permits Gt binding. This coupling between photoactivated Rho* (also termed Meta II) and Gt is the first step in the visual signal transduction pathway. In its inactive state, Gt is a membrane-associated complex consisting of β-subunits and noncovalently bound GDP. The Rho*-Gt interaction causes opening of the nucleotide-binding site on Gt, allowing the release of GDP and its replacement by GTP. Activated Gt GTP then is released to interact with the -subunits of phosphodiesterase 6 (PDE6), removing an inhibitory constraint from the catalytic - or β-subunits (1)
.
Although the structures of both Rho and Gt are known, the precise molecular mechanism of this receptor-mediated G protein activation is not fully defined. In the dark, Gt likely is precoupled to Rho as their relative concentrations are 5 mM (8)
and 0.5 mM (9)
, respectively, and they have reported affinities for each other as high as 60 nM (10)
or as low as 10 µM (11)
. But the affinity of Gt for Rho* increases to 1 nM (10)
. The transient complex of Rho*-Gte formed just after GDP release can thus be trapped in ROS membranes. Also Rho*-Gte exists as a tight complex in the absence of free GTP, wherein retinal is blocked from dissociating from Rho* (12)
. This critical observation could potentially be exploited to isolate the Rho*-Gte complex in detergent for further structural characterization.
Much work has been done to define the interacting surfaces between Rho* and Gt that might produce greater insight into the nature of the activated complex. First, the Rho*-Gte complex interface was interrogated in a competition assay involving Rho peptides (e.g., ref. 13
). However, small fragments of Rho may not properly mimic the physiological conformation of the loops of Rho, as evidenced by the high concentrations of peptides needed for these experiments. Khorana’s group (14)
has pioneered the use of site-directed Cys mutagenesis and crosslinking for this purpose. Their studies showed that the second and third loops as well as cytoplasmic helix 8 on the Rho* surface are involved in Gt binding. Gt is docked to Rho* by the extreme C terminus and residues within the 4-β6 loop of Gt. Other studies involving competition and crosslinking of the C-terminal peptide of Gt (residues 340–350) confirmed that the Gt C-terminal region is important for binding to the Rho* receptor and that it stabilizes Rho* in the Meta II conformation (15
, 16)
. Biochemical studies have shown that the Gtβ heterodimer directly contacts different parts of Rho cytoplasmic helix 8 (17)
. Notably, the interface engaged in the complex likely encompasses the entire cytoplasmic surface of Rho. Thus, focusing on specific short regions of Gt or Rho could lead to oversimplification of this interaction.
In view of the above findings, large conformational changes were expected in the structure of Rho* and Gt after initiation of binding if the ratio between Rho* and Gt in the complex were 1:1 (18)
. Because both the C-terminal tails of Gt and the Gt subunits can bind to and stabilize Meta II (19)
, "a sequential fit model" was also proposed wherein several complexes form sequentially (20)
. But the latter would require that a protein "remembers" preceding encounters to produce the next complex, a hysteresis phenomenon virtually unknown in biology unless a post-translational reaction is involved. Another possibility based on the sizes of the Rho and Gt surfaces (21)
, Meta II stabilization by C-terminal fragments of Gt (15)
and the Gt subunits, and the presence of Rho oligomers in native photoreceptor and other membranes (22
, 23)
is that a Rho*Rho dimer might be the functional unit activating G proteins (21)
. To get a clear answer as to the surfaces of Rho and Gt involved in the complex, the crystal structure of the receptor-G protein complex must be solved. Toward this goal, we show in the present study that a stable, detergent-solubilized Rho*-Gte complex can be isolated by sucrose gradient centrifugation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. For complex formation between Rho* and Gt, ROS membranes (0.2 mg/ml of protein) were suspended in 20 mM bis-Tris propane (BTP), pH 6.9, containing 120 mM NaCl, 0.1 mM MgCl2, 1 mM dithiotreitol (DTT), and 1 mM phenylmethylsulphonyl fluoride (PMSF), bleached (150 W lamp) for 10 min under a Fiber-Light covered with a 480–525 nm band-pass filter (Chroma Technology, Rockingham, VT, USA) at a distance of 10 cm and incubated for 30 min on ice. For removal of soluble proteins present in ROS and endogenous GDP released from the Gt nucleotide-binding site, ROS membranes were washed 5 times with 5 mM BTP, pH 6.9, containing 0.1 mM MgCl2, 1 mM DTT, and 1 mM PMSF (12)
. Rho*-Gte then was solubilized with 20 mM BTP, pH 6.9, containing 120 mM NaCl, 0.1 mM MgCl2, 1 mM DTT, 1 mM PMSF, and 20 mM n-dodecyl-β-D-maltopyranoside (DDM) for 1 h on ice. The Rho*-Gte complex was enriched by using a 750–900–900–750-µl step gradient composed of 20–30–40–50% sucrose in 20 mM BTP, pH 6.9, containing 120 mM NaCl, 0.1 mM MgCl2, 1 mM DTT, 1 mM PMSF, and 0.3 mM DDM. The density of each fraction was determined from its refractive index. Detergent-solubilized Rho*-Gte complex (300 µl, 2 mg/ml of protein) was loaded on top of the gradient and centrifuged at 269,000 g for 16 h, and fractions (250 µl each) were collected from the top to the bottom of the gradient. Protein content was analyzed by SDS-PAGE and immunoblotting, and fractions containing the Rho*-Gte complex were saved at 4°C for further experiments. The Rhoe*-Gte complex was generated just as Rho*-Gte, with one exception: ROS membranes were illuminated in the presence of freshly neutralized 50 mM NH2OH. Then excess nucleophile was removed with 4 washes of 5 mM BTP, pH 6.9, containing 0.1 mM MgCl2, 1 mM DTT, and 1 mM PMSF (2 ml each). The resulting Rhoe*-Gte complex was solubilized and isolated as described above for the Rho*-Gte complex.
As controls for the above experiments, 50 µg of purified Rho (25)
and 50 µg of purified Gt (26)
was suspended in 20 mM BTP, pH 6.9, containing 120 mM NaCl, 0.1 mM MgCl2, 1 mM DTT, and 1 mM PMSF (100 µl total volume) and subjected to sucrose gradient centrifugation either in the dark or after being exposed to illumination by light.
Dissociation of Rho*-Gte complex by GTPS
The Rho*-Gte complex was dissociated by adding 200 µM GTPS to the sample either after the Rho*-Gte complex was formed in ROS membranes or immediately after its purification. Both preparations also were tested after 1 or 3 wk of storage at 4°C. Complex disruption was checked by subsequent sucrose gradient centrifugation.
Regeneration of Rho in the Rhoe*-Gte complex
For regeneration of Rho from the Rhoe*-Gte complex with chromophore, an equimolar amount of an ethanolic solution of either 9-cis-retinal or 11-cis-retinal was added under dim red light to the isolated Rhoe*-Gte complex. The final concentration of ethanol did not exceed 0.5%. Formation of Rho was monitored by UV-visible spectroscopy at 0, 5, 15, 30, and 45 min after chromophore addition. The yield of regenerated Rho was quantified based on the changes in the absorption ratio at 280 nm/500 nm for 11-cis-retinal or at 280 nm/495 nm for 9-cis-retinal. The absorption ratio at 280 nm/500 nm for fully regenerated Rho was assumed to be 1.8. Formation of the Schiff base in the Rhoe*/9-cis-retinal-Gte complex was verified by acid denaturation as described in the UV-visible spectroscopy paragraph of Supplemental Methods.
Opsin was regenerated in a control experiment. First, opsin was prepared by 10 min bleaching of Zn2+-extracted Rho from bovine ROS membranes under a Fiber-Light covered with a 480–525 nm band-pass filter. Then this bleached protein was left to decay for 1 h at room temperature and subsequently was aged for 20 h at 4°C, the same time needed for the Rhoe*-Gte preparation.
Quantification of the molar ratio between Rho* and Gt in the isolated Rho*-Gte complex
Concentrations of Rho* and Gte in the fractions enriched in the Rho*-Gte complex were calculated by using different amounts of both immunoaffinity-purified Rho (25)
and purified Gt (26)
as standards. Rho concentrations were quantified by absorption at 500 nm and use of the absorption coefficient 
= 40,600 M–1 cm–1 (27)
. Gt concentrations were determined by Bradford ULTRA (Novexin, Cambridge, UK) with bovine serum albumin used as a standard and confirmed by the amount of GDP released from Gt, as described in Supplemental Methods. SDS-PAGE gels containing different amounts of free Rho (0.25, 0.5, 0.75 and 1 µg), Gt (0.2, 0.4, 0.6, 0.8, and 1 µg), and Rho*-Gte complex from 4 different preparations corresponding to fraction 7 in Fig. 1
were stained with Sypro-Ruby stain (Molecular Probes, Inc., Eugene, OR, USA) and protein bands were quantified by Molecular Imager FX (Bio-Rad, Richmond, CA, USA). Typical standard curves for Rho and Gt gave straight lines (R2>0.99). Results of calculating the ratios between Rho* and Gte in fractions of the isolated Rho*-Gte complex are shown in Supplemental Fig. S5A. In another set of quantifications we also calculated the ratios between Gt and Gtβ and between Rho* and Gtβ in fractions 6, 7, and 8 from gels treated with Coomassie blue stain using the Image J program (U.S. National Institutes of Health, Bethesda, MD, USA). Results of these calculations are shown in Supplemental Fig. S5B.
|
Transmission electron microscopy (TEM) of negatively stained samples
Isolated Rho*-Gte complexes (corresponding to fraction 7 in Fig. 1
) at 7.5 µg/ml in 20 mM BTP, 120 mM NaCl, 0.1 mM MgCl2, 1 mM DTT, 
2 mM DDM (concentration determined as described in Supplemental Methods paragraph dealing with determination of detergent concentrations in protein fractions collected from sucrose gradients), pH 6.9, and purified Gt-GDP protein at 8.7 µg/ml in 20 mM Tris-HCl, 100 mM NaCl, 5 mM MgCl2, 1 mM DTT, 1 mM PMSF, pH 7.5, were adsorbed for 10 s to parlodion carbon-coated copper grids rendered hydrophilic by glow discharge at low pressure in air. Grids were washed with three drops of double-distilled water and stained with 2 drops of 0.75% uranyl formate. Electron micrographs of Rho*-Gte and Gt-GDP preparations were recorded with a Hitachi H-7000 transmission electron microscope operated at 100 kV (Hitachi, Tokyo, Japan).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
).
After overnight ultracentrifugation, the Rho*-Gte complex was found in fractions 6–8 containing 30–40% sucrose (Fig. 1A
). Although the sample was not 100% pure, fractions 6–8 were highly enriched in this complex. To determine whether the Rho*-Gte complex was biochemically active, 200 µM GTPS was added to the sample before loading on the sucrose gradient. Most of the complex dissociated, and Rho* as well as Gte-GTPS was present mainly in fractions 3–5, containing 20–25% sucrose (Fig. 1B
), the same fractions that contained free Rho* and free Gt-GDP in the control experiment (Fig. 1C
). In another control experiment, we mixed purified Rho with purified Gt-GDP at a molar ratio of 2:1 and subjected the mixture to sucrose gradient ultracentrifugation; these mixtures were either kept in the dark or were light-illuminated (Fig. 1E
). After overnight centrifugation both unilluminated Rho and Gt-GDP were present in fractions 3–6, characteristic of free Rho and Gt-GDP. Surprisingly we found little change in sucrose gradient migration when these samples were exposed to light, which suggests that complex formation between Rho* and Gt from purified components in detergent solution is less efficient than in membranes. To confirm that the isolated Rho*-Gte complex was biochemically active, we performed a Gt activation fluorescence assay (Supplemental Fig. S2A). After adding GTPS to the protein sample, we observed an increase of intrinsic fluorescence emanating from the dissociated Gt subunit, with an activation rate of 1.2 x 10–3 s–1, whereas no changes were detected without addition of GTPS.
Dissociation of nucleotides in the Rho*-Gte complex
Gt loses GDP from its nucleotide-binding pocket after binding to Rho* resulting in a transient complex between Rho* and Gte free of nucleotide bound to its -subunit (28)
. As a result, Gte cannot dissociate from Rho* until GTP binds to this vacant nucleotide-binding site. But even without GTP in solution, released GDP might possibly bind back and cause partial dissociation of the Rho*-Gte complex. To prevent Rho*-Gte complex dissociation, we washed the ROS membranes thoroughly before determining the nucleotide content of isolated Rho*-Gte samples by reverse-phase HPLC, according to the procedure described in Supplemental Methods. No nucleotides were detected in the Rho*-Gte complex (Fig. 2A
, top panel) as compared with GTP and GDP standards (Fig. 2A
, bottom panel) or with GDP found in isolated Gt (not shown).
|
Heterogeneity of Rho* photoproducts in the Rho*-Gte complex
A UV-visible spectrum of the isolated Rho*-Gte complex revealed the presence of at least two photoproducts with a predominant absorption at 375 nm and lower absorption at longer wavelengths around 485 nm (Fig. 2B
, solid line spectrum). This observation suggested the presence of Meta II, unprotonated retinylidene Schiff base, and possibly Meta I, photoproducts with a protonated retinylidene Schiff base. However, Meta I should be highly disfavored by Gte binding. Therefore, in the Rho*-Gte sample, other Rho intermediates besides Meta II that absorb at visible wavelengths are most likely present. To distinguish between the Meta II state and free retinal, which also absorbs in the range of 360–390 nm, we used an acid denaturation method. Under acidic conditions the Schiff base of the Meta II photoproduct should become protonated, and its maximum absorption peak would appear at 440 nm, whereas absorption of free retinal should not change. After acidification of the Rho*-Gte complex, Rho photoproducts were converted to 440-nm-absorbing species (Fig. 2B
, dashed line), suggesting that the predominant photoproduct in the Rho*-Gte complex is Meta II with an unprotonated Schiff base. Moreover, the photoproduct absorbing at 485 nm could be converted to the Meta II species after additional illumination of isolated fractions containing the complex, as shown in Fig. 2B
(dotted line). To determine the composition of the 485 nm peak, we added freshly prepared 20 mM neutral NH2OH to the Rho*-Gte complex to cleave the Schiff base that covalently attaches the chromophore to Lys-296. The Schiff base between opsin and isomerized retinal (but not 11-cis-retinal) is hydrolyzed by NH2OH such that the maximum peak of absorption at 360 nm due to free retinal is observed. The dashed-dotted spectrum in Fig. 2B
obtained after addition of NH2OH to the Rho*-Gte sample shows that most of the photoproducts were converted to species absorbing at 360 nm, but minor absorption at longer wavelengths also occurred. This result might suggest that some unbleached Rho was present along with the Rho*-Gte complex, or rather that chemical isomerization of all-trans-retinal to its cis-isomers, favored by the presence of phosphoethanolamine, had occurred.
We used retinoid analysis (as described in Supplemental Methods) to determine the photoproduct composition of the enriched Rho*-Gte preparation. The major fraction of retinoids found (57.3±3.4%) consisted of all-trans-retinal, a component of Rho Meta II as well as Meta I. A smaller fraction (14.2±2.5%) consisted of 11-cis-retinal, a chromophore characteristic of dark-adapted Rho. Despite the use of a Fiber-Light covered with a narrow 480–525 nm band-pass filter (Chroma Technology), 9-cis-retinal (16.2±1.3%) and 13-cis-retinal (12.3±0.7%) were also present in the isolated sample (Fig. 2C
, top panel). Thus, Rho cis-isomers most likely account for the observed absorption peak at 485 nm (Fig. 2C
, top panel). In the additionally illuminated Rho*-Gte sample, we observed conversion of 11-cis- to all-trans-retinal as well as a reduction in 9-cis-retinal (Fig. 2C
, bottom panel).
Stability of the Rho*-Gte complex
To determine the stability of the Rho*-Gte complex, we kept a freshly isolated sample at 4°C and subjected aliquots to a second gradient centrifugation after 1 or 3 wk. After 1 wk, most of the complex was present in the 30–40% sucrose fractions (6
7
8)
, where the Rho*-Gte complex migrates. A minor amount of free Rho* appeared at a lower sucrose density (fraction 5), suggesting either that some of the Rho*-Gte complex had dissociated or, more likely, that an excess of free Rho* was present in the original sample since free Gte was not detected (Fig. 2D
, top panel). Stored for 1 wk at 4°C, the Rho*-Gte complex was still functional since addition of GTPS before sucrose gradient centrifugation caused dissociation of Gte-GTPS from Rho*, such that free Rho* as well as free Gte-GTPS was present in lower-density fractions 4 and 5 (Fig. 2D
, bottom panel). Fractions containing the Rho*-Gte complex were collected and concentrated, after which retinoid analysis and UV-visible spectra were used to determine the composition of the Rho photoproducts.
The UV-visible spectrum of the Rho*-Gte complex kept at 4°C for 1 wk was similar to that of the fresh sample. It displayed two peaks, one at 375 nm and the other at longer wavelengths around 485 nm (Supplemental Fig. S1A, solid line). This result suggests that Meta II as well as other Rho* intermediates were present in the isolated sample. To distinguish between the Meta II state and free retinal, we analyzed photoproducts of the Rho*-Gte complex after acid denaturation. This shifted the 375 nm absorption peak to 440 nm, suggesting that most of the photoproduct had an unprotonated Schiff base (Supplemental Fig. S1A, dashed line). Therefore, Meta II was still present in the Rho*-Gte sample after 7 days. The photoproduct absorbing at 485 nm could be converted to the Meta II species after additional illumination as shown in the spectrum in Supplemental Fig. S1A (dotted-line spectrum). After addition of fresh 20 mM neutral NH2OH to the 1-wk-old Rho*-Gte sample most of the photoproducts were converted to species absorbing at 360 nm, but the minor absorption noted at longer wavelengths suggested the presence of cis Rho isomers (Supplemental Fig. S1A, dashed-dotted spectrum).
The content of retinoids was similar to that found in the freshly isolated sample (Supplemental Fig. S1B). The major fraction consisted of all-trans-retinal; 11-cis-retinal, 9-cis-retinal and 13-cis-retinal were present as well.
Formation and isolation of the Rhoe*-Gte complex
NH2OH hydrolyzes the Schiff base bond (retinylidene) between retinal and opsin at the Meta II and later stages, whereas dark Rho and Rho* in the Meta I state are unaffected (29)
. The released retinal oxime has an absorption maximum at 360 nm. To test whether removal of retinal from the chromophore-binding pocket influences Rhoe*-Gte complex formation and activity, we prepared the complex with empty retinal and nucleotide-binding pockets and isolated it by sucrose gradient ultracentrifugation. Both Rhoe* and Gte proteins were found in fractions characteristic of Rho*-Gte complex migration (Fig. 3A
). In this Rhoe*-Gte complex, Gte could bind GTPS, causing disruption of the complex and appearance of Gte-GTPS in earlier fractions (Fig. 3B
). The functional activity of the isolated Rhoe*-Gte complex was confirmed by the Gt activation fluorescence assay. Addition of GTPS caused an increase of intrinsic fluorescence emanating from dissociated Gt with an activation rate of 3.1 x 10–3 s–1 (Supplemental Fig. S2B). The spectrum of the isolated Rhoe*-Gte complex showed the presence of copurified free retinal oximes that absorb at 360 nm. Addition of H2SO4 had no influence on this spectrum, confirming the absence of a Schiff base in the isolated complex (Fig. 3C
). Retinoid analysis of this Rhoe*-Gte complex showed that virtually all of the Rho chromophore (11-cis-retinal) was converted to all-trans-retinal oximes (Fig. 3D
).
|
Regeneration and stability of the Rhoe*-Gte complex
To determine whether Rhoe* in the Rhoe*-Gte complex could be regenerated, we added an equimolar amount of 9-cis-retinal or 11-cis-retinal to the isolated complex and monitored formation of Rho immediately afterward by recording UV-visible spectra at 0, 5, 15, 30, and 45 min. (Fig. 4A
, left and middle panels). In a control experiment, 20 h-aged opsin was regenerated with 9-cis-retinal or 11-cis-retinal (Fig. 4B
, left and middle panels). Rhoe* regeneration was observed as an increase of the maximum peak at 495 or 500 nm, when 9-cis-retinal or 11-cis-retinal was added. Although only 10% of added free opsin bound to the chromophore (Fig. 4B
, right panel), 75% of the Rhoe* in the Rhoe*-Gte complex was regenerated with 11-cis-retinal and 52% of Rhoe* in the Rhoe*-Gte complex was regenerated with 9-cis-retinal (Fig. 4A
, right panel).
|
No Schiff base was present before or immediately after addition of 9-cis-retinal; however, formation of the Schiff base was observed 30 min after regeneration of the Rhoe*/9-cis-retinal-Gte complex (Supplemental Fig. S3A). To test the ability of the regenerated Rhoe*/11-cis-retinal-Gte complex to form light-activated Rho species, we illuminated this sample (Supplemental Fig. S3B, long dashed line) for 10 min through a 480–525 nm band-pass filter and observed a transition of the absorption peak at 485 nm to 380 nm (Supplemental Fig. S3B, solid line), suggesting formation of Meta II.
Isolated by sucrose gradient centrifugation, regenerated Rhoe*/9-cis-retinal-Gte and Rhoe*/11-cis-retinal-Gte complexes as well as nonregenerated complexes were subjected to a second sucrose gradient ultracentrifugation to examine their stability. Both complexes with exogenously incorporated chromophore migrated in sucrose gradients similarly to the nonregenerated Rhoe*-Gte complex, and both Rhoe* and Gte were located mostly in fractions 6 and 7 characteristic of the complex. Both the regenerated complexes as well as the empty-empty Rhoe*-Gte complex had the ability to bind GTPS. This caused disruption of the complexes and appearance of Gte-GTPS in the lower-density sucrose fractions (fractions 3 and 4) and increase of its amount in fraction 5 (Fig. 5
).
|
To determine whether excess exogenous chromophore might disrupt the Rhoe*-Gte complex, we added a 3-fold higher concentration of 9-cis-retinal to the empty-empty Rhoe*-Gte complex formed in ROS membranes and incubated the mixture for 30 min on ice before loading it on the sucrose gradient. UV-visible spectra of the Rhoe*-Gte complex and free opsin mixture taken before addition of 9-cis-retinal showed no chromophore attached to protein (Supplemental Fig. S4A, solid line), but after 30 min incubation with 9-cis-retinal, the spectrum changed to one suggesting formation of the Schiff base between the protein and chromophore (Supplemental Fig. S4A, dashed line). The isolated Rhoe*/9-cis-retinal-Gte complex also displayed an absorption spectrum with two peaks, one at 495 nm due to Rho regenerated with 9-cis-retinal and the other at 360 nm due to free retinal (Supplemental Fig. S4B). We did not observe any changes either in the migration of the Rhoe*/9-cis-retinal-Gte complex after sucrose gradient centrifugation or in the GTPS binding properties of this complex in the presence of excess free 9-cis-retinal (Supplemental Fig. S4C).
Visualization of Rho*-Gte complex particles by TEM
We used TEM and negative staining to inspect the homogeneity of the Rho*-Gte complex preparation (Fig. 6
). Two populations of particles, as well as bigger copurified vesicular structures (asterisks), were noted. The larger particles, with diameters of 12.5 ± 0.6 nm (n=42; arrowheads), most likely are particles of the Rho*-Gte complex. Magnified images of these particles are displayed in the inset gallery (bottom left). The smaller particles with diameters of 8.6 ± 0.6 nm (n=42, arrows) had sizes characteristic of Rho dimers (30)
. For comparison, purified free Gt-GDP protein also was visualized by TEM, and these protein particles are shown in the right gallery. Gt-GDP particles had a diameter of 9.4 ± 0.6 nm (n=42) and were composed of two elongated densities, similar to Rho dimer. Because both Gte and Rho* dimers had a similar mass and appearance, the content of these proteins in the smaller particles could not be differentiated at the resolution achieved. Free Gte and Rho* dimers may have arisen either from excess free Rho* dimers present in the sample or from Rho*-Gte complexes that fell apart during adsorption to the grid and negative staining.
|
Statistical analysis of the negatively stained protein particles indicated the presence of more than 60% particles with a diameter of 12.5 nm. Considering the sizes of negatively stained Rho (31)
and Gt (this work), the 12.5 nm particles correspond to Rho*-Gte complexes. Thus, besides some heterogeneity our sample is enriched in one population of particles, i.e., Rho*-Gte complexes. The percentage of this major population may even be underestimated considering that complex disruption may happen on sample adsorption to the carbon film and staining (leading to the formation of smaller particles). Therefore, TEM of negatively stained preparations not only identified a certain level of heterogeneity of the preparation but also helped estimating the amounts of intact Rho*-Gte complexes.
Quantification of the Rho*:Gte molar ratio in the isolated Rho*-Gte complex
To quantify the ratio between Rho* and Gte in the isolated Rho*-Gte complex we used two different gel-staining techniques as described in Materials and Methods. Calculated as 2 Rho* molecules per molecule of the Gte and β subunit, this ratio was similar in selected fractions 6, 7, and 8 from several different preparations of Rho*-Gte (Supplemental Fig. S5). The molar ratio between Gt and Gtβ subunits of Gte present in the isolated complex was estimated as 1:1 (Supplemental Fig. S5B).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
subunit (32)
. Over the years, phototransduction with its receptor Rho has served as a prototype for many different GPCR-mediated signal transduction events, so understanding this process is broadly applicable to other signal transduction cascades (1)
. More work is needed to advance knowledge of the mechanism of Rho 
Rho* activation and signal transfer to Gt. The full 3-dimensional interface between Rho* and Gt in the complex has yet to be defined, as well as how the signal is transmitted from Rho* to cause nucleotide exchange on the Gt subunit. Direct proof of the activation process could arise from structural studies of the complex between Rho* and Gt. Before this goal can be achieved, however, the complex needs to be isolated and characterized.
Isolation of Rho-Gt complexes
From very early studies of phototransduction, it was known that Gt can be extracted from dark-adapted ROS membranes with low-ionic-strength buffer, whereas photoactivation of Rho led to retention of Gt on these membranes. Moreover Gt was released from Rho* in such membranes only when GTP was added (33)
. This work was extended to develop a routine procedure for Gt isolation (34)
. Bornancin et al. (12)
showed that GDP dissociated from the Rho*-Gt complex when the latter was diluted by a buffer and that the resulting Rho*-Gte complex was stable for days. Here we developed a protocol that enabled us to solubilize this complex with detergent and partially purify it from excess Rho*/Rho. The method is based on sucrose gradient ultracentrifugation, an inexpensive and readily scalable procedure. By this method proteins can be separated according to the differences in their sizes and the differences in density between the protein/lipid/micelles complexes and the medium. The Rho*-Gte complex in mixed detergent-endogenous lipid micelles likely has a higher density than Rho itself because of different ratios of phospholipids to protein in the complex (lower) and free receptor (higher). Intense washing of Gte bound to Rho* in membranes eliminated the vast majority of soluble proteins and released GDP from the binding site of Gt. Because Rho constitutes >90% of the protein in disk membranes (1)
, the purity of our preparation typically was greater than 95% with a variable excess of Rho*. Initial studies of these enriched Rho* complexes provided new insights into Gt activation by Rho* (see below).
Interplay between nucleotide- and chromophore-binding sites
Once the Rho*-Gt complex is formed, GDP dissociates from the nucleotide-binding pocket (Fig. 7
) and the complex can be solubilized, isolated, and characterized. Addition of the slowly hydrolysable analog of GTP, GTPS, was used to assess whether both proteins were functional under the experimental conditions used. Although addition of GTPS changed the migration of Gt subunits along the gradient, one could be present in the same sucrose fractions as Gt heterotimer. This is because heterotrimeric Gt as well as dissociated Gt and Gtβ most likely bind to lipids/detergent micelles through their myristoyl and farnesyl groups, thus affecting migration of dissociated Gt subunits in the sucrose gradient in a similar way as heterotrimeric Gt (Fig. 1B
comparing with Fig. 1C
). Heterotrimeric Gt migrated in the sucrose gradient as a purified protein in the absence of additional lipids.
|
In fact, the isolated Rho*-Gte complex was found to be stable for more than 7 days without any appreciable decrease in activity under various storage conditions. Although it is known that opsin can be activated by the surplus of soluble all-trans-retinal, this is unlikely to explain the activity we observed in our Rho*-Gte complex preparation, because we documented the presence of the Schiff base in the chromophore binding pocket (35
, 36)
. Really impressive was that the all-trans-retinylidene chromophore was stable in its binding site and that decay of Rho* in the isolated detergent-solubilized Rho*-Gte complex was totally blocked. However, addition of GTPS to the Rho*-Gte sample stimulated its rapid decay from the Meta II to the Meta III photoproduct and eventually to inactive opsin (not shown). Surprisingly, the Schiff base between apo-protein and all-trans-retinylidene was accessible to NH2OH hydrolysis and the Rhoe*-Gte complex still was very stable and active in the GTPS activation assay. This finding suggests that opsin in the complex is conformationally stabilized by the presence of Gte, an unexpected observation because opsin reportedly is unstable in detergent (37
38
39)
. Since free opsin cannot activate Gt or at least activates it very slowly, we cannot exclude that the observed Gt activation in the Rhoe*-Gte complex might be caused by the presence of all-trans-retinal in the isolated sample. As shown by other studies, all-trans-retinal in a complex with opsin can activate Gt robustly (35
, 36)
.
In contrast, when we tested formation of a complex by mixing two purified proteins the yield of the complex was poor. One of the factors that may have affected the efficient coupling of Gt to purified Rho is the presence of detergent micelles. Although Rho* can activate Gt in detergent, the formation of the complex in high yield is not very efficient.
Free all-trans-retinal was detected in the isolated Rhoe*-Gte samples. Released from the chromophore-binding pocket, all-trans-retinal may associate either with membranes or the Rho surface by noncovalent interaction with their Lys residues exposed to the solvent. However, it should not bind back to the retinoid-binding site and inhibit the formation of Rho during its regeneration (40)
. The presence of multiple cis-retinoids suggests that once chromophore is photoisomerized to all-trans-retinylidine, it can flip back to various cis- and di-cis configurations. These cis compounds that constitute <45% of total retinoids are unlikely to be formed by photoisomerization of Meta II because a 480–525 nm band-pass filter was used during our purification procedure. However, these contaminating retinoids might be formed due to nonenzymatic isomerization of all-trans-retinal catalyzed by components of rod outer segments, including phosphoethanolamine (phosphoethanolamine must be present in the isolated Rho*-Gte sample) as previously reported (41)
.
Regeneration of Rho-Gt complexes
Further evidence for the stability of isolated Rhoe*-Gte was its regeneration with 9-cis-retinal/11-cis-retinal. This finding supports the notion that the regenerated Rho-Gte complex was stable and active, even in the presence of 9-cis-retinylidene/11-cis-retinylidene. The same complex also was active in the GTPS activation assay. The latter observation suggests that, once the complex is formed, Rho is conformationally stabilized by the presence of Gte (Fig. 7
). Oprian and colleagues (42)
proposed the highly insightful hypothesis that transformation of Rho Rho* involves disruption of the all-trans-retinylidene-Glu113 salt bridge. This concept may be valid still, but reconstituting this salt bridge between the retinal protonated Schiff base is not sufficient for Gt dissociation once the Rhoe*-Gte complex is formed. This in turn suggests that Rho in the complex is conformationally stabilized by the presence of Gte and can activate Gt. The same finding also demonstrates that changes in the absorption of the chromophore can be decoupled from Gt activation. Thus, the question can be raised as to whether the observed spectral changes in UV-visible spectra measured by ultrafast spectroscopy after bleaching of Rho at low temperatures are fully relevant to Gt activation. These results also are consistent with the idea that G proteins modify the conformation of GPCRs (43)
. Along the same lines, conformational changes during the Rho Rho* transition can be monitored by UV-visible spectroscopy because of the presence of bound chromophore (44)
. After coupling to protein, 11-cis-retinal shifts the absorption maximum from 360 nm to 440 nm when it forms a protonated Schiff base and this absorption is further modified in the opsin transmembrane binding sites. For example, highly homologous cone opsins yield a 
max that can range from 360–640 nm. This max is a sensitive indicator of conformational changes of photoactivated Rho, and specific photointermediates can be trapped by manipulations of pH, salt, sucrose or temperature. However, it is quite intriguing that large differences in the absorbance spectra of Rho and its intermediates arise from minimal structural changes in this protein molecule (3)
. The active MII-like state can be attained without major conformational changes as indicated by recent crystallographic data (3)
. Moreover, a "straightjacketed" Rho created by disulfide bonding of mutant Cys residues introduced into the Rho sequence can still activate Gt and display proper spectral changes despite restricted movement of its helices (45)
.
Heterogeneity of Rho-Gt complexes
As shown by negative stain TEM and SDS-PAGE, the sample of Rho*-Gte complex prepared by sucrose gradient ultracentrifugation was somewhat heterogeneous with an excess of free opsin/Rho*, contaminating unrelated proteins and a small amount of larger membrane particles. The relevant visual components appear to consist of a mixture of free opsin along with complexes with a 1:1 Rho to Gt and a 2:1 Rho to Gt stoichiometry, suggesting that the second Rho/Rho* is not very tightly bound in the complex. It should be noted that stability of the Rho dimmer is critically affected by the concentration of DDM (26
, 31)
.
We describe herein the development of a sucrose density centrifugation method that produces stable, detergent solubilized, functionally active Rho*-Gte, Rhoe*-Gte and 9-cis-retinal/11-cis-retinal regenerated Rho-Gte complexes. These results demonstrate that once Rho* couples with Gt, the chromophore plays only a minor role in the stability of these complexes, regardless of their conformation. This method could be useful, with future modifications, to obtain high-resolution structures of the Rho*-Gte complex.
|
ACKNOWLEDGMENTS |
|---|
Received for publication June 2, 2008. Accepted for publication September 4, 2008.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
subunits by receptors through G-Gβ and G-G interactions. Trends Biochem. Sci. 28,13-17[CrossRef][Medline]This article has been cited by other articles:
|
|
 |
|
  P. Scheerer, M. Heck, A. Goede, J. H. Park, H.-W. Choe, O. P. Ernst, K. P. Hofmann, and P. W. Hildebrand Structural and kinetic modeling of an activating helix switch in the rhodopsin-transducin interface PNAS, June 30, 2009; 106(26): 10660 - 10665. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
