HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 46, 44566-44575, November 15, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/46/44566    most recent M205308200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Obin, M. Articles by Taylor, A. Search for Related Content PubMed PubMed Citation Articles by Obin, M. Articles by Taylor, A. Social Bookmarking                      What's this?

Ubiquitylation of the Transducin Subunit Complex

REGULATION BY PHOSDUCIN^*

Martin Obin^§, Bruce Y. Lee^¶, Gretchen Meinke, Andrew Bohm, Rehwa H. Lee^**, Rachelle Gaudet, Johnathan A. Hopp^§§, Vadim Y. Arshavsky^§§¶¶, Barry M. Willardson^¶, and Allen Taylor

From the  Laboratory for Nutrition & Vision Research, JMUSDA-HNRCA at Tufts University and Tufts Center for Vision Research, Boston, Massachusetts 02111; the ^¶ Department of Chemistry and Biochemistry, Brigham Young University, Provo, Utah 84602; the  Boston Biomedical Research Institute, Watertown, Massachusetts 02472; the ^** Jules Stein Eye Institute and Greater Los Angeles Healthcare System at Sepulveda, Sepulveda, California 90095; the  Department of Molecular and Cellular Biology, Harvard University, Boston, Massachusetts 02138, and the ^§§ Department of Ophthalmology, Harvard Medical School and the Massachusetts Eye and Ear Infirmary, Boston, Massachusetts 02114

Received for publication, May 29, 2002, and in revised form, September 4, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

G proteins (G) are essential signaling molecules, which dissociate into G and G upon activation by heptahelical membrane receptors. We have identified the subunit complex of the photoreceptor-specific G protein, transducin (T), as a target of the ubiquitin-proteasome pathway. Ubiquitylated species of the transducin -subunit (T) but not the - or -subunits were assembled de novo in bovine photoreceptor preparations. In addition, T was exclusively ubiquitylated when T was dissociated from T. Ubiquitylation of T on T was selectively catalyzed by human ubiquitin-conjugating enzymes UbcH5 and UbcH7 and was coincident with degradation of the entire T subunit complex in vitro by a mechanism requiring ATP and the proteasome. We also show that T association with phosducin, a photoreceptor-specific protein of unknown physiological function, blocks T ubiquitylation and subsequent degradation. Phosphorylation of phosducin by Ca²⁺/calmodulin-dependent protein kinase II, which inhibits phosducin-T complex formation, completely restored T ubiquitylation and degradation. We conclude that T is a substrate of the ubiquitin-proteasome pathway and suggest that phosducin serves to protect T following the light-dependent dissociation of T.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Heterotrimeric guanine nucleotide-binding proteins (G)¹ constitute a large family of eukaryotic signaling molecules, which transduce signals between seven-helical membrane receptors and intracellular effectors or ion channels (reviewed in Refs. 1-3). Agonist-induced exchange of GDP for GTP on G results in dissociation of G·GTP from G, each of which can interact with effectors in different signaling pathways. Hydrolysis of bound GTP promotes reassociation of G and G and termination of the G protein-mediated signal (1-5). G activity is regulated by at least two classes of additional proteins. They include the G protein-coupled receptor kinases (GRKs) (6) and phosducin (Pd) and phosducin-like proteins (PhLPs) (reviewed in Ref. 7). Phosducin is most highly expressed in the mammalian pineal gland and retina. In the latter it is present in high concentrations (>350 µM) in the cytosol of photoreceptor (i.e. rod) cells (8, 9). Phosducin forms a tight complex with the heterodimer of the photoreceptor-specific G protein, transducin (T) following light-induced transducin dissociation (10, 11). Formation of the phosducin-T (Pd-T) complex reduces the availability of free T for re-association with T·GDP, which is required for subsequent transducin activation. Dark-dependent phosducin phosphorylation by protein kinase A (PKA) or by Ca²⁺/calmodulin-dependent protein kinase II (CaMKII) reduces phosducin binding to T by 3- and 300-fold, respectively (11-15). These observations suggested a mechanism in which the cycle of phosducin phosphorylation/dephosphorylation regulates the light sensitivity of the photoreceptor by controlling the amount of transducin available for activation. Such a mechanism might be important for photoreceptor light adaptation, during which potentially saturating levels of light input to photoreceptor cells are countered by desensitization (reviewed in Ref. 16). However, recent studies (9, 17, 18) clearly demonstrate that most phosducin is localized not in the outer segments of photoreceptors, where the light receptor rhodopsin is present and where phototransduction occurs, but in the inner segments, which are primarily responsible for the metabolic functions of the cell. This inconsistency calls for a revision of the proposed role of phosducin in photoreceptors. A solution might be provided by the observation that, upon continuous illumination, up to 90% of both T and T translocate from the photoreceptor outer segment to the inner segment (Ref. 19 and references therein). However, the functional importance of T interaction with phosducin in the inner segment is unclear.

The ubiquitin (Ub) proteasome pathway (UPP) provides another potential mechanism of G protein regulation (20, 21). The UPP is a conserved pathway of selective protein modification and degradation that controls levels and activities of many highly regulated eukaryotic proteins (reviewed in Refs. 22, 23). Substrates of the UPP are covalently ligated to one or more monomers of the 8.5 kDa protein, ubiquitin, by the sequential activities of three families of thiol enzymes: Ub-activating enzymes (E1), Ub-conjugating enzymes (Ubc), and Ub-isopeptide ligases (E3) (22, 23). Two E1 enzymes, over 30 Ubc enzymes, and more than 100 E3 ligases have been identified. Selectivity in ubiquitylation is accomplished by the specific interplay between Ubc and E3, which interact with distinct but incompletely understood ubiquitylation signals within substrates. Ubiquitylation signals (also called "degrons" when ubiquitylation targets proteins for degradation) consist of two modules: 1) primary determinants for recognition by Ubc, E3 (and occasionally ancillary factors) and 2) secondary determinants, i.e. one or more lysines to which ubiquitin is covalently attached (reviewed in Ref. 24). Multiple ubiquitin molecules can be subsequently attached to a substrate as an isopeptide-linked polyubiquitin chain. Such chains preferentially target the substrate moiety of the Ub-protein conjugate for degradation by the 26 S proteasome, a multicatalytic, ATP-dependent protease (reviewed in Ref. 25). Although the best recognized function of ubiquitylation is selective targeting of proteins for rapid degradation (22, 23), ubiquitylation per se can regulate protein trafficking, phosphorylation, and other nonproteolytic fates (reviewed in Ref. 26).

The most extensive evidence for the regulation of G protein signaling by the UPP has been obtained in S. cerevisiae, where Ub-dependent proteolysis of G controls signaling through the mating pheromone receptor (24, 27). We previously identified T as a UPP substrate in vitro (20) and proposed that T was the ubiquitylated subunit, based on detection of a ~30 kDa Ub protein conjugate in the transducin fraction from bovine photoreceptors (28). Interestingly, phosducin has also been suggested to interact with the UPP. Specifically, two-hybrid screens, reconstitution assays and overexpression studies (29, 30) have demonstrated that phosducin (and PhLP) interacts with p45/Sug1, a subunit of the 26 S proteasome 19 S regulatory complex. This observation has led to a suggestion that phosducin and/or PhLP act as adapters linking G heterodimers to the 26 S proteasome (29-31). Upon delivery to the proteasome, G could be degraded, or alternatively, refolded by the chaperone activity of the 19 S regulatory complex (32).

Here we report that T is ubiquitylated on T and that T ubiquitylation targets the T heterodimer for degradation by the 26 S proteasome in vitro. However, T is completely resistant to ubiquitylation and degradation when complexed with phosducin. These data provide the first example of G subunit complexes being UPP substrates and suggest a novel role for phosducin as a protective factor for T during continuous illumination.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- Materials for electrophoresis were from Bio-Rad laboratories (Hercules, CA). Coomassie Plus protein assay reagent and the Super Signal chemiluminescence detection kit were purchased from Pierce. Na¹²⁵I was supplied by PerkinElmer Life Sciences. Nickel-nitrilotriacetic acid (Ni-NTA) agarose beads were from Qiagen, Inc. (Valencia, CA). Frozen bovine retinas were purchased from W. L. Lawson, Co. (Lincoln, NE). Carbobenzoxyl-leucinyl-leucinyl-leucinal-H (MG132), ubiquitin aldehyde (Ub-aldehyde), His₆-tagged ubiquitin (His₆-Ub), rabbit E1, HeLa cell fraction I (FI), bacterially expressed recombinant human Ubc (UbcH) H2, H6, H7, H9, and H10 and active site CysSer mutants of UbcH5c and UbcH7 were purchased from Boston Biochem. Inc (Cambridge, MA). His₆-UbcH3 was expressed in Escherichia coli and purified on Ni-NTA agarose. (The His₆-UbcH3 plasmid was a generous gift of Dr. Sharon Plon, Texas Children's Cancer Center, Houston). UbcH5c was expressed in E. coli and purified as previously reported (33). (The UbcH5c plasmid was generously provided by Dr. Simon Wing, Department of Medicine, McGill University). Bacterially expressed bovine UbcH1 (34) was a generous gift from Dr. Cecile Pickart (The Johns Hopkins University, Baltimore, MD). Rabbit reticulocytes were purchased from Pel-Freez Biologicals (Rogers, AR). Unless otherwise specified, all other materials were purchased from Sigma and were the highest grade available.

High speed (85,000 × g) supernatants from rabbit reticulocyte lysate and human retinal pigment epithelial (RPE) cells were prepared as described (20, 28). Ubc-depleted RPE supernatant was prepared by spin dialysis through a Centricon 100 microconcentrator (Millipore, Corp, Bedford, MA). Reticulocyte lysate fraction II (FII) was prepared as per Hershko and co-workers (35).

Photoreceptor and Transducin Preparations-- Bovine photoreceptor rod outer segments (ROS) were purified as previously described from dark-adapated retinas on continuous density gradients, and high speed ROS supernatant was obtained (28). A (>90%) mixture of either T·GTPS and T or T·AlF₄ and T was eluted from extensively washed, photolyzed ROS membranes in the presence of 100 µM GTPS (28) or AlF₄ (36), respectively. Transducin was deprenylated by digestion with endoproteinase LysC (endo-LysC), which removes T residues 69-71, including the farnesyl moiety at Cys-71 (37). Deprenylated T was purified by sequential chromatography on Cibacron Blue 3GA-agarose and Mono-Q (Amersham Biosciences) (36). Prenylated (nonproteolyzed) T was purified on Cibacron Blue CL6B (Supelco, Inc., Bellefonte, PA) from a mixture of T·AlF₄ and T (36). T purity (absence of T) was confirmed by immunoblotting. T activity was confirmed by binding to the C-terminal domain of -adrenergic receptor kinase (38), with binding assessed by electrophoretic mobility shift on native gels.² T was iodinated to low specific activity (~30,000 cpm/µg) as previously described (20). The T·GTPS/T mixture, pure T and ¹²⁵I-labeled T were stored at 20 °C in 10 mM Tris-HCl (pH 7.5), 0.5 mM DTT, and 40% glycerol.

Phosducin and Phosducin-T complex-- Pd-T complex was purified from bovine retinas as previously described (39). Association of Pd and T was confirmed by comigration of phosducin and T on native gels (39). Recombinant rat Pd-Myc-His₆ and Pd^S73A-Myc-His₆ were expressed in E. coli, purified on Ni-NTA resin, buffer-exchanged and stored at 20 °C in 20 mM Tris (pH 7.5), 100 mM NaCl, 0.5 mM DTT, and 50% glycerol (14, 37, 40). Pd^S73A-Myc-His₆-T complex was formed in vitro by coincubating Pd^S73A -Myc-His₆ (6 µg) and purified T (4 µg) in 10-12 µl of 20 mM Tris (pH 7.8), 1 mM DTT for 30 min at 22 °C. The ability of recombinant phosducins to bind T was confirmed by inhibition of light-induced binding of T to T (11). Pd-Myc-His₆ was pentaphosphorylated by CaMKII and purified as described (14). The inability of phosphorylated phosducin to bind T was confirmed in T binding assays (as above).

Ubiquitylation Assays-- Assays (25 µl) contained 3-12 mg/ml of either ROS or RPE supernatant, reticulocyte lysate, reticulocyte FII, or combined FII and HeLa cell FI, 2 mM ATP, and an ATP-regenerating system, 80 µM MG132, 4 µM Ub-aldehyde, and 200-400 ng/µl Ub or His₆-Ub (28, 41). Assays also received 2-4 µg of exogenous substrate (i.e. T·GTPS/T, T, or purified or reconstituted Pd-T complex). To identify Ubcs involved in T ubiquitylation, some assays were supplemented with 0.2-0.8 µg of individual recombinant human Ubc or 4 µg of active site CysSer mutants of UbcH5a or UbcH7. To control for differences in Ubc activity, the amount of each Ubc added to assays varied as the inverse of Ubc activity, which was quantitated by autoradiography as the formation of ¹²⁵I-labeled Ub-Ubc thiol esters in the presence of purified rabbit E1 and ATP (41). Phosducin phosphorylation was prevented by inclusion of the kinase inhibitor H-89 (Calbiochem, Novabiochem, San Diego, CA) at a final concentration (400 µM) sufficient to inhibit both PKA (K_i, 0.25 µM) and CaMKII (K_i, 30 µM). Phosphorylated phosducin was stabilized against phosphatase activity by inclusion of microcystin LR (10 µM final concentration).

Ubiquitylation assays were incubated at 37 °C for the times indicated and were terminated by boiling with gel loading buffer containing 2-mercaptoethanol. Ubiquitylated proteins were visualized by Western blotting (41). Primary antibodies included the following: polyclonal IgGs raised against peptide sequences of mammalian G_t1 (sc-389), G_t1 (sc-379), and G_t1 (sc-373) (all purchased from Santa Cruz Biotechnology, Inc., Santa Cruz, CA), monoclonal antibody (mAb) TF-15 raised against bovine G_t1 (42), mAb TF-28 raised against bovine G_t1 (43), and rabbit antiserum raised against the LAP-636 peptide of bovine G_t1 (44) (all generously provided by Dr. Bernard Fung, University of California at Los Angeles), polyclonal IgG (BN-1) raised against a conserved peptide sequence of bovine G_t1 (a generous gift from Dr. Mel Simon, California Institute of Technology, Pasadena, CA), rabbit serum ("Gertie") (13), which recognizes phosducin and PhLP, anti-phosducin mAb 1D6 (a generous gift of Dr. Lawrence Donoso, Wills Eye Hospital, Philadelphia, PA), polyclonal IgG, which recognizes free and conjugated ubiquitin (28) or appropriate preimmune or nonimmune IgG/serum. Specific binding was detected with horseradish peroxidase (HRP)-conjugated anti-rabbit or anti-mouse IgG (Jackson Immunoresearch Laboratories, West Grove, PA) and visualized by ECL. Signal intensities were quantified by digital densitometry (Amersham Biosciences), using standard curves of antigens (in assay mixtures) to determine the linear range of ECL signal detection.

Ni-NTA-agarose Isolation of T-His₆Ub Species Synthesized De Novo-- Ubiquitylation assays (70 µl) were incubated for 1 h and terminated with Buffer I (8 M urea, 0.1% Tween 20, 20 mM 2-mercaptoethanol, 10 mM imidazole, 0.1 M sodium phosphate, pH 8.0, 10 mM Tris-HCl, pH 8.0). Next, 25 µl of washed Ni-NTA beads (50% slurry) were added and incubated (30 min, 22 °C) with gentle rotation. The beads were subjected to 5 cycles of washing with Buffer II (Buffer I, pH 6.5). His₆-Ub-protein conjugates were then eluted with Buffer III (Buffer I, pH 4.5 containing 0.5 M imidazole), boiled in reducing gel buffer, and analyzed by Western blotting.

Proteolysis Assay-- Proteolysis assays were constituted similarly to ubiquitylation assays (above), except that Ub-aldehyde was omitted, and reactions were conducted in the presence and absence of both ATP and MG132. Degradation (i.e. loss) of specific proteins was assessed by densitometry of immunoblotted assay mixtures or, when ¹²⁵I-labeled T was used as a substrate, by -counting of acid-precipitable cpm (20, 28).

Analysis of Phosducin and PhLP Interaction with p45/Sug1 by Glycerol Gradient Sedimentation-- Light-adapted bovine retinas were homogenized in ice-cold 50 mM Tris-HCl/0.5 mM DTT (pH 8.0) and centrifuged (100,000 × g, 20 min, 2 °C) to obtain supernatant. Approximately 1 mg of supernatant was brought to a final concentration of 2 mM ATP, 1.2 mM DTT, and 6 mM MgCl₂, and subjected to sedimentation through a 15-35% glycerol gradient (200,000 × g, 22 h, 4 °C, Beckman SW41Ti). Eleven fractions (2 ml each) were collected, acetone-precipitated and boiled in reducing SDS-PAGE sample buffer. Following electrophoresis and transfer to membrane (41), fractions were probed with one of the following: 1) rabbit serum raised against a common 32-kDa subunit of the 20 S proteasome (45) (a generous gift from Dr. George deMartino, University of Texas Southwest Medical Center, Dallas, TX), 2) rabbit serum raised against Trip1, the human orthologue of yeast p45/Sug1 (a gift of Dr. Richard Young, Massachusetts Institute of Technology, Cambridge, MA), 3) rabbit serum (Gertie) (13), which detects both phosducin and PhLP, or (4) anti-phosducin mAb1D6.

Statistics-- Data are reported as mean ± S.E. Differences between treatments were tested for significance using one-way analysis of variance (Systat v.9).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Identification of Transducin Subunits That Are Ubiquitylated in Vitro-- To identify subunits of transducin that are ubiquitylated in vitro, ubiquitylation assays were conducted with supernatant from gradient-purified, dark-adapted bovine ROS. This preparation contains ubiquitin-conjugating enzyme activities and endogenous, heterotrimeric transducin (T) (28). Proteins were immunoblotted and probed with antibodies raised against T, T, and T. No higher mass (i.e. ubiquitylated) forms of T or T were detected in reaction mixtures supplemented with ATP and Ub (Fig. 1A, lane 2, upper and middle panels, respectively). However, higher mass species of T (T15K, T30K) were detected, and their formation was ATP-dependent, consistent with the requirement of ATP for ubiquitylation (Fig. 1A, lower panel; compare lanes 1 and 2). The apparent molecular mass of T is ~7 kDa, allowing us to rationalize these higher mass forms of T as TUb₁ (15 kDa) and TUb₃ (30 kDa), respectively. However, it is plausible that the 30 kDa ubiquitylated species contains only two ubiquitins (TUb₂), but migrates anomalously (at 30 kDa rather than 22 kDa) in SDS-PAGE gels. An identical mass distribution for T staining was obtained with additional anti-T antibodies (data not shown). These results suggest that T was ubiquitylated exclusively on T, with up to three ubiquitin moieties incorporated.

View larger version (26K): [in this window] [in a new window]   Fig. 1.   De novo ubiquitylation of T by the photoreceptor UPP. A, ATP-dependent synthesis of higher mass species of T in vitro. Ubiquitin-supplemented supernatant from gradient-purified photoreceptor (rod) outer segments was incubated in the presence (lane 2) or absence (lane 1) of ATP (see "Experimental Procedures"). Western blots of reaction mixtures were probed with IgGs directed against T (upper panel), T (middle panel), and T (lower panel). Molecular mass markers are at left. A representative assay of three is shown. B, isolation of His6-ubiquitylated T on nickel agarose. ATP-supplemented photoreceptor ubiquitylation assays (as in A) were supplemented with either native Ub (lanes 2, 4, and 6) or His6-Ub (lanes 1, 3, and 5). Reaction mixtures were incubated with Ni-NTA beads under denaturing conditions, and proteins bound to pelleted beads were subjected to SDS-PAGE and Western blotting with anti-T IgG. Lanes 1 and 2 contain ubiquitylation assay supernatants; lanes 3 and 4 contain cleared ubiquitylation assay supernatants following incubation with and pelleting of Ni-NTA beads; lanes 5 and 6 contain proteins pelleted with Ni-NTA beads. Lanes 5 and 6 contain 10-fold more input than lanes 1-4. The slower migrating mass species of TUb1 in lane 4 is bleed over from lane 5. TUb3 is so designated based on apparent molecular mass (30 kDa), but it may contain only two ubiquitins (see "Results"). Molecular mass markers are at left. A representative assay of two is shown.

T ubiquitylation by the photoreceptor UPP was confirmed by supplementing ROS preparations with His₆-Ub, isolating His₆-ubiquitylated species on Ni-NTA beads under denaturing conditions, and detecting T in the pelleted beads by Western blotting (Fig. 1B). Observations consistent with the ligation of His₆-Ub to T include: 1) retarded electrophoretic migration of TUb₁ and TUb₃ species in assays supplemented with His₆-Ub, reflecting the additional mass of the His₆ tag (Fig. 1B, compare lanes 1 and 2); 2) clearance of TUb₁ and TUb₃ bands from His₆-Ub-supplemented assays by Ni-NTA beads (Fig. 1B, compare lanes 1 and 3); 3) enrichment for TUb₁ and TUb₃ in Ni-NTA pellets from assays containing His₆-Ub (Fig. 1B, lane 5) but not from assays supplemented with native Ub (Fig. 1B, lane 6); and 4) ubiquitin immunoreactivity of 15- and 30-kDa species on Western blots of Ni-NTA pellets from assays containing His₆-Ub (data not shown). No T or T immunoreactivity was detected on Western blots of Ni-NTA pellets (data not shown; note that the T complex was dissociated under the denaturing conditions of these pull-down assays). In summary, these data demonstrate that the higher molecular mass forms of T were ubiquitylated T species and support the notion that transducin ubiquitylation occurred exclusively on T.

T Is Not Required for T Ubiquitylation-- The data from Fig. 1 indicate that T can be ubiquitylated as a part of the soluble T trimer. We next addressed if T could also be ubiquitylated following dissociation of T from T. We incubated a mixture of light-dissociated transducin subunits (T·GTPS/T) in RPE supernatant, a cell-free preparation used as the source of UPP enzymes (20). The assays were supplemented with His₆-Ub and ATP, as well as with MG132 and Ub-aldehyde, inhibitors of the proteasome and deubiquitylating enzymes, respectively. His₆-Ub-protein conjugates were then isolated under denaturing conditions with Ni-NTA agarose beads, and proteins in the supernatants and those attached to the precipitated beads were analyzed by Western blotting for the presence of T. A trace of TUb₁ was detected in reaction mixtures containing transducin but no exogenous ATP (Fig. 2A, lane 2), but this trace amount was insufficient for detectable isolation (Fig. 2A, lane 5). In contrast, significant levels of TUb₁ and some TUb₃ were detected in the ATP-supplemented samples (Fig. 2A, lane 3). In this case, TUb₁ was readily detected in Ni-NTA precipitates of these reactions (Fig. 2A, lane 6). These data indicate that T was ligated to His₆-Ub through an ATP-dependent mechanism. Moreover, since T·GTPS and T were dissociated in these assays, we conclude that T is not required for the ubiquitylation of T within the T subunit complex.

View larger version (29K): [in this window] [in a new window]   Fig. 2.   T is ubiquitylated on T. A, de novo synthesis and purification of His6-ubiquitylated T from light-dissociated transducin (T·GTPS/T). A mixture of T·GTPS and T was prepared from photolyzed bovine photoreceptor membranes (see "Experimental Procedures"). This substrate was incubated with His6-Ub-supplemented RPE supernatant in the presence or absence of ATP, and His6-Ub-protein conjugates were isolated with Ni-NTA beads. Western blots of reaction mixtures (lanes 1-3) and protein pellets (lanes 4-6) were probed with anti- T IgG. Lanes 1 and 4 contain ATP but no transducin; lanes 2 and 5 contain transducin but no ATP; lanes 3 and 6 contain both transducin and ATP. A representative experiment of two is shown. Molecular mass markers are at left. B, de novo synthesis of TUb1 from purified T and stabilization of TUb1 by UPP inhibitors. Bovine T was purified by blue-Sepharose chromatography and incubated with reticulocyte lysate for 30 min in the presence of ATP and Ub. Some assays (lanes 3 and 4) were not supplemented with inhibitors of the proteasome (MG132) and deubiquitylating enzymes (Ub-aldehyde). Western blots of reaction mixtures were probed with anti-T IgG. Molecular mass markers are at left. A representative assay of three is shown.

We further supported this conclusion by conducting ubiquitylation assays with purified T. These experiments used reticulocyte lysate, a classical cell-free UPP system (35), which was supplemented with ATP in the presence or absence of MG132 and Ub-aldehyde. Western blotting confirmed the synthesis of TUb₁, which was detectable in assays containing MG132 and Ub-aldehyde (Fig. 2B, compare lanes 1 and 2). Consistent with results obtained with the G·GTPS/T mixture, ubiquitylation of purified T occurred selectively on T, since no higher molecular mass species of T were detectable on Western blots of ATP-supplemented reactions (data not shown).

T Is a Substrate for Ub-dependent Proteolysis-- In the experiments illustrated in Fig. 2B, we also noted that TUb₁ was only detectable in the presence of MG132 and Ub-aldehyde (Fig. 2B, compare lanes 2 and 4). This observation suggested that ubiquitylated T was a substrate for either the 26 S proteasome and/or de-ubiquitylating enzymes. To confirm this possibility, we conducted UPP proteolysis assays using ¹²⁵I-labeled T as substrate. The hallmarks of ubiquitin-dependent proteolysis are the requirements for ATP and proteasome activity. We therefore monitored the release of acid-soluble radioactivity when ¹²⁵I-labeled T was incubated in reticulocyte lysate in the presence or absence of ATP and MG132. The data (Fig. 3A) indicate that ~13% of T was degraded within 30 min in the presence of ATP, whereas less than 2% was degraded in the absence of ATP. In addition, ATP-dependent proteolysis of ¹²⁵I-labeled T was substantially blocked by MG132 (Fig. 3A). These data suggest that ¹²⁵I-labeled T was degraded almost exclusively by the UPP.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   T is degraded by the ubiquitin-proteasome pathway. A, degradation of 125I-labeled T is ATP- and proteasome-dependent. 125I-labeled T was incubated for 30 min in reticulocyte lysate in the presence and absence of ATP and MG132. Percent degradation was calculated from acid-soluble counts per min after correction for background (<3%) (28). Data are from two assays performed in duplicate. B, both T and T are degraded by the UPP. 125I-labeled T was incubated in reticulocyte lysate in the presence or absence of ATP and MG132 (as in Fig. 3A). Levels of T and T were assessed after 120 min by SDS-PAGE and autoradiography of assay mixtures (20,000 cpm per gel lane). Levels of both 125I-labeled T and 125I-labeled T were reduced in the presence of ATP (compare lanes 1 and 2) but were enhanced when ATP-supplemented assays contained MG132 (compare lanes 2 and 3). Note that T consistently incorporates dramatically less radiolabel as compared with T. C, degradation of native T. Native T was incubated in reticulocyte lysate in the presence of ATP. Levels of T (upper panel) and T (lower panel) were assessed at the start of incubations (T0, lanes 1 and 3) and after 90 min (T90, lanes 2 and 4) in duplicate assays (I, II) by Western blotting with anti-T and anti-T IgGs. A representative experiment of three is shown.

Autoradiography of ¹²⁵I-labeled T degradation assays following 2 h of incubation in reticulocyte lysate (as in Fig. 3A) reveals that levels of both T and T were reduced in the presence of ATP (Fig. 3B, compare lanes 1 and 2) and were stabilized in the presence of MG132 (Fig. 3B, compare lanes 2 and 3). These results suggest that, despite the fact that only T becomes ubiquitylated, the entire T complex is subsequently degraded. To verify that the native (noniodinated) T complex was also degraded by the UPP, we incubated non-radiolabeled T in ATP-supplemented reticulocyte lysate and assessed the levels of T and T, which remained after 90 min by Western blotting (Fig. 3C). The levels of both T and T were reduced (~40-60%) following incubation in ATP-supplemented assays (Fig. 3C, compare lanes 1 and 2 and 3 and 4). Data obtained with radiolabeled and native T therefore support the conclusion that both subunits of the T complex are degraded by the UPP.

Identification of Ubiquitin-conjugating Enzymes That Catalyze T Ubiquitylation and Promote T Degradation-- We then identified Ubc species that can catalyze T ubiquitylation. In an initial screen using T as substrate and RPE supernatant as the UPP source, we found that T ubiquitylation could be enhanced by supplementation of ubiquitylation assays with recombinant human UbcH5a, UbcH5c, or UbcH7 but not with UbcH1, UbcH2, UbcH3, UbcH6, UbcH9, or UbcH10 (50). We then employed an established approach (35, 46) to evaluate the requirement of UbcH5/UbcH7 in T ubiquitylation. In this approach, rabbit reticulocyte lysate was subjected to DEAE chromatography, and UbcH5 isoforms and UbcH7 were removed in the column flow-through (FI). The high salt eluate (FII) contains E1, other Ubcs and E3. Although some T ubiquitylation was observed with FII alone (Fig. 4A, lane 1), ubiquitylation was significantly enhanced by the addition of FI (lane 2; Note that HeLa cell FI was used to avoid the high levels of hemoglobin that are present in reticulocyte FI). Notably, both recombinant UbcH5c (lane 3) and UbcH7 (lane 4) stimulated the formation of TUb₁, TUb₃, and TUb₄. These results demonstrate that T can be ubiquitylated by the actions of UbcH5 and UbcH7. Additional evidence for the role of UbcH5 in catalyzing T ubiquitylation was obtained utilizing dominant negative active site (CysSer) mutant UbcH5a (mUbcH5), which inhibits the function of multiple UbcH5 isoforms in vitro (see "Discussion"). This mutant acts as a dominant negative, because it fails to bind ubiquitin while continuing to interact with E3. When added to ATP-supplemented reticulocyte lysate (Fig. 4B) or to RPE supernatant (not shown), mUbcH5 inhibited the synthesis of both TUb₁ and TUb₃ (Fig. 4B, compare lanes 2 and 3). Interestingly, mUbcH7 had no detectable effect on T ubiquitylation (Fig. 4B, compare lanes 2 and 4). These data suggest that although UbcH7 can catalyze T ubiquitylation in the absence of UbcH5 (Fig. 4A, lane 4), UbcH7 is not required for T in these preparations. Neither UbcH5c nor UbcH7 were able to catalyze T ubiquitylation in reconstitution assays, which contained Ub, ATP, and E1 but which lacked cell supernatant (not shown). This observation strongly suggests that T ubiquitylation requires an E3 and/or ancillary factor. Importantly, Fig. 4C shows that no ubiquitylated species of T were detected in these reconstitution assays, consistent with data obtained using non-fractionated UPP preparations (Fig. 2A). Thus, UbcH5-or UbcH7-mediated ubiquitylation of T occurs exclusively on T.

View larger version (32K): [in this window] [in a new window]   Fig. 4.   Identification of the Ubc that catalyzes T ubiquitylation and promotes T degradation. A, HeLa cell FI, UbcH5, or UbcH7 enhance T ubiquitylation in reticulocyte Fraction II. Reticulocyte Fraction II, which contains diminished levels of UbcH5 and UbcH7 orthologues, was incubated with T in the presence of ATP, His6-Ub, MG132, and Ub-aldehyde. Some incubations also received either HeLa cell Fraction I (FI, lane 2), which is enriched for UbcH5 isoforms and UbcH7, recombinant UbcH5c (lane 3), or recombinant UbcH7 (lane 4) (each at 70 nM final concentration). Ubiquitylation reactions were terminated after 30 min, and proteins were subjected to Western blotting with anti-T IgG. Representative data are from one of three assays. B, an active site CysSer mutant (m) UbcH5, but not mutant Ubch7 inhibits T ubiquitylation. T ubiquitylation assays were conducted in His6-Ub-supplemented reticulocyte lysate in the absence (lane 1) or presence (lanes 2-4) of ATP. Some assay mixtures were preincubated with mUbcH5 (lane 3) or mUbcH7 (lane 4) (each at a final concentration of 0.75 µM). Western blots were probed with anti-T IgG. A representative assay of two is shown. C, the T subunit is not ubiquitylated in the presence of HeLa cell FI, UbcH5, or UbcH7. Ubiquitylation reactions (as in A) were supplemented with HeLa cell FI (lane 1), UbcH5c (lane 2), or UbcH7 (lane 3). Western blots of assay mixtures were probed with anti-T1 serum (LAP636). LAP636 cross-reactivity was assessed in the absence of T (lane 4). nsb, nonspecific band. Representative data are from one of three assays. D, UbcH5 promotes the degradation of 125I-labeled T. 125I-labeled T was incubated for 30 min in Ubc-depleted RPE supernatant in the presence of ATP and Ub. Some assays were additionally supplemented with UbcH5c (35 nM final concentration). Proteolysis (percent acid-soluble cpm) was quantified by gamma counting. Representative data are from one of three experiments performed in duplicate. ***, p < 0.001 for effect of UbcH5. E, UbcH5 promotes the degradation of native T. T was incubated in ATP- and Ub-supplemented reticulocyte FII in the presence (lanes 2 and 4) or absence (lanes 1 and 3) of exogenous UbcH5c (as in A). left panel, levels of T remaining after 90 min were assessed on Western blots probed with anti-T IgG. One of two experiments performed with duplicate T incubations (I, II) is shown. Middle panel, densitometry data from an experiment (as in left panel) performed in triplicate. *, p < 0.05 for effect of UbcH5. Right panel, T standard curve confirming that densitometry data plotted in the middle panel were within the linear range of ECL detection (r2 = 0.996).

To determine if UbcH5-dependent ubiquitylation targets T for degradation, we assessed the effect of exogenous UbcH5c on the release of acid-soluble radiolabel from ¹²⁵I-labeled T (20, 28). ¹²⁵I-labeled T was incubated in ATP- and Ub-supplemented RPE supernatant that had been previously depleted of Ubcs by dialysis (see "Experimental Procedures"). The addition of UbcH5c to Ubc-depleted RPE supernatant was associated with a 3-fold increase in the magnitude of ATP/Ub-dependent degradation of ¹²⁵I-labeled T (p < 0.001; Fig. 4D). In addition, whereas ~40% of ¹²⁵I-labeled T was degraded in 90 min by reticulocyte lysate (as in Fig. 3, A and B), less than 5% of ¹²⁵I-labeled T was degraded during the same period by reticulocyte FII, which is depleted of UbcH5 isoforms (n = 2 assays in duplicate).³ Together, these data support a role for UbcH5 in the degradation of ¹²⁵I-labeled T.

Mammalian UbcH5s and their yeast homologues are implicated in the selective ubiquitylation and degradation of abnormal and damaged proteins (22, 47). To rule out the possibility that UbcH5-dependent degradation of ¹²⁵I-labeled T resulted from T iodination (i.e. oxidation), we also assessed UbcH5-mediated degradation of native (noniodinated) T. T was incubated in ATP- and Ub-supplemented reticulocyte FII in the presence and absence of exogenous UbcH5c. After 90 min, levels of T were assessed by Western blotting (Fig. 4E, left panel) and densitometry (Fig. 4E, middle and right panels). Reaction mixtures supplemented with UbcH5c contained on average ~40% less T than reaction mixtures lacking UbcH5c (p < 0.02) (Fig. 4E, compare lanes 1 and 2 and 3 and 4). These results confirm that UbcH5 promotes the degradation of T in vitro. Considered together, ubiquitylation data (Fig. 4, A-C) and proteolysis data (Fig. 4, D and E) strongly suggest that ubiquitylation of T by UbcH5 destabilizes the T complex, presumably by targeting T to the 26 S proteasome (Fig. 3, A and B).

Phosducin Binding to T Prevents T Ubiquitylation-- Phosducin binds T with high affinity, covering portions of T (37, 48, 49) and putatively inducing conformational changes in the T C terminus (49). To investigate the potential effects of phosducin binding on T ubiquitylation, we conducted two similar groups of experiments. In the first, we purified the phosducin-T complex (Pd-T) from bovine retina and compared its ubiquitylation with ubiquitylation of purified T. As demonstrated above (Fig. 4A), when T was incubated in UbcH5-supplemented FII, T was ubiquitylated in an ATP-dependent manner (Fig. 5A, left panel; compare lanes 1 and 2). In contrast, T was not ubiquitylated when the Pd-T complex was incubated in UbcH5-supplemented FII (Fig. 5A, left panel; compare lanes 2 and 3). The possibilities that T or phosducin were ubiquitylated in these experiments were also investigated, but Western blotting failed to detect higher mass species of either protein (Fig. 5A, middle and right panels). Thus, in contrast to T, the Pd-T complex does not appear to be a substrate for de novo ubiquitylation.

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Effect of phosducin and phosducin phosphorylation on T ubiquitylation. A, the Pd-T complex is resistant to ubiquitylation. T (lanes 1 and 2) and purified bovine Pd-T complex (lane 3) were incubated with Ub- and UbcH5c-supplemented FII in the presence (lanes 2 and 3) or absence (lane 1) of ATP. Ubiquitylation was assessed after 1 h by Western blotting with IgGs raised against T (left panel), T (middle panel), and phosducin (right panel). Molecular mass markers are at left. A representative assay of three is shown. B, preincubation of T with phosducin blocks T ubiquitylation. T was preincubated with bovine serum albumin (lanes 1 and 2) or with phosphorylation-resistant phosducin (PdS73A, lane 3) before addition to Ub-supplemented RPE supernatant in the presence (lanes 2 and 3) or absence (lane 1) of ATP. Assays were terminated after 1 h and immunoblotted for T (as in A). Molecular mass markers are at left. A representative assay of three is shown. C, phosducin does not globally inhibit protein ubiquitylation. Ubiquitylation assays (as in B) were supplemented with 125I-labeled Ub, and total 125I-Ub-protein conjugates were visualized by autoradiography of SDS-PAGE gels. Molecular mass markers are at left. A representative assay of two is shown. D, CaMKII-phosphorylated phosducin does not inhibit T ubiquitylation. Left panel, T was deprenylated with endo-LysC (see "Experimental Procedures") and preincubated with either vehicle (lanes 2 and 3), PdS73A (lane 4) or with CaMKII-phosphorylated phosducin (Pd~p) (lane 5) before addition of ubiquitylation assay mixtures (as in B) supplemented with His6-Ub. Assay mixtures were preincubated with saturating levels of kinase and phosphatase inhibitors (see "Experimental Procedures"). Similar results were obtained in three assays. Molecular mass markers are at left. Right panel, confirmation of Pd~p phosphorylation. After termination, ubiquitylation assays containing PdS73A or Pd~p (as in left panel, lanes 4 and 5) were Western blotted with anti-phosducin IgG. Phosphorylation of Pd~p (lane 2) is confirmed by its retarded electrophoretic mobility relative to PdS73A (lane 1). Similar results were obtained in three assays.

In the second group of experiments, we reconstituted the Pd-T complex in vitro and then compared ubiquitylation of the reconstituted complex with ubiquitylation of native T. In these experiments we used the phosphorylation-resistant phosducin mutant, Pd^S73A (see "Experimental Procedures" for details). As expected, native T was ubiquitylated by RPE supernatant in an ATP-dependent manner, generating TUb₁ and TUb₃ (Fig. 5B, compare lanes 1 and 2). In contrast, T which had been preincubated with Pd^S73A was not ubiquitylated (Fig. 5B, compare lanes 2 and 3). Consistent with results obtained with the Pd-T complex purified from bovine retina (Fig. 5A, middle and right panels), we detected no Ub protein conjugates of T or phosducin (data not shown). Importantly, the ability of Pd^S73A to inhibit ubiquitylation in these assays appears specific for T, as the extent and pattern of global protein ubiquitylation was similar in the presence or absence of Pd^S73A (Fig. 5C, compare lanes 2 and 3). These results confirm that the binding of phosducin to T inhibits T ubiquitylation and does not promote ubiquitylation of either T or phosducin.

Phosphorylated Phosducin Fails to Protect T from Ubiquitylation-- Phosphorylation of phosducin by CaMKII, which is thought to occur in the dark, reduces the affinity of phosducin for T by ~300-fold (14). To assess potential effects of phosphorylation on phosducin's ability to inhibit T ubiquitylation, we compared the ability of Pd^S73A and CaMKII-phosphorylated phosducin (Pd~p) to inhibit T ubiquitylation. As shown in Fig. 5D, in the absence of exogenous Pd^S73A, TUb₁ and TUb₃ were formed in an ATP-dependent manner (compare lanes 2 and 3). As expected, preincubation of T with Pd^S73A abrogated formation of TUb₁ and TUb₃ (Fig. 5D, compare lanes 3 and 4). However, preincubation of T with CaMKII-phosphorylated phosducin had no significant effect on de novo formation of TUb₁ or TUb₃ (Fig. 5D, compare lane 5 with lanes 3 and 2). Thus, in contrast to Pd^S73A, phosphorylated phosducin was permissive for T ubiquitylation. Confirmation that Pd^S73A and CaMKII-phosphorylated phosducin remained differentially phosphorylated throughout the duration of the ubiquitylation assay is shown by the retarded electrophoretic migration of CaMKII-phosphorylated phosducin on SDS-PAGE gels (Fig. 5D, right panel).

Effects of Phosducin Binding on T Proteolysis by the UPP-- Ubiquitylation targets T for degradation by the UPP (Figs. 2B, 3, and 4, D and E). Based on the ability of phosducin to block T ubiquitylation (above), we predicted that formation of the Pd-T complex would protect T from Ub-dependent degradation. Indeed, when purified, retina-derived Pd-T complex was incubated in ATP-supplemented reticulocyte lysate, we observed no loss of either T (not shown) or T (Fig. 6A, compare lanes 1 and 2 and 3 and 4). We also reconstituted the Pd^S73A-T complex in vitro (as above) and added reticulocyte FII, which had been supplemented with Ub, ATP, and UbcH5c (as in Fig. 4E). Levels of T and T remaining after 90 min were visualized by Western blotting. Preincubation of T with Pd^S73A was consistently associated with the retention of more immunoreactive T and T as compared with assays lacking Pd^S73A (Fig. 6B, compare lanes 1 and 2, 3 and 4, 5 and 6). These data suggest that formation of the Pd-T complex stabilizes T.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Phosducin binding blocks degradation of T by the UPP. A, stability of T in the Pd-T complex. Pd-T complex was purified from bovine retina and incubated in ATP-supplemented reticulocyte lysate. Levels of T were assessed on Western blots at the start of reactions (T0) and after 90 min (T90). A representative assay of two performed in duplicate (I, II) is shown. B, T and T are stabilized in the reconstituted Pd-T complex. T was preincubated in the presence (lanes 1, 3, and 5) or absence (lanes 2, 4, and 6) of PdS73A and then incubated for 90 min with reticulocyte FII, which was supplemented with ATP, Ub, and UbcH5c (as in Fig. 4). Levels of T and T remaining were visualized by Western blotting. A representative assay of two performed in triplicate (I, II, III) is shown. C, ATP/Ub-dependent degradation of 125I-labeled T is inhib