Endoproteolytic Processing of Sst2, a Multidomain Regulator of G Protein Signaling in Yeast*

Regulators of G protein signaling (RGS proteins) con-stitute a large family of G protein-binding proteins. All RGS proteins contain a conserved core domain that can accelerate G protein GTPase activity. In addition, many family members contain a unique N-terminal domain of unknown function. Here, we demonstrate that the RGS protein in yeast, Sst2, is proteolytically processed in vivo to yield separate but functional N-terminal and RGS core domain fragments. In whole cell lysates, the full-length SST2 product (82 kDa) as well as a prominent 36-kDa species are specifically recognized by antibodies against the C terminus of the Sst2 protein. Purification and chemical sequencing of the 36-kDa species revealed cleavage sites after Ser-414 and Ser-416, just preceding the region of RGS homology. Expression of a mutationally truncated form of the protein (C-Sst2) could not restore function to an sst2 D mutant strain. In contrast, co-expression of C-Sst2 with the N-terminal domain (N-Sst2) partially restored the ability to regulate the growth arrest response but not the transcription induction response. Whereas the full-length protein was localized to the microsomal and plasma membrane fractions, the N-Sst2 species was predominantly in the microsomal fraction, and C-Sst2 was in the soluble fraction. Mutations that block proteasome or vacuolar protease function, or mutations in the cleavage site Ser residues of Sst2, did not alter processing. However, Sst2 processing did require WCG4a ura3 leu2-3, WCG4-11/22a (WCG4a, pre1-1 pre2-2 MHY753 MAT a his3- D leu2 ura3-52 lys2-801 D MHY754 (MHY753, 3-112 trp1-1 Plasmids and Mutagenesis— Expression plasmids used in this study are pRS315 ( CEN , amp R , LEU2 ), pRS423 (2 m m, amp R , HIS3 ), pRS425 (2 m m, amp R , LEU2 ) pRS316-ADH ( CEN , amp R , URA3 , ADH1 promoter/terminator) (41), and pAD4M (2 m m, amp R , LEU2 , ADH1 promoter/terminator) (from P. McCabe, Onyx Pharmaceutical). pAD4M-SST2 was constructed by digesting SST2 with Sal I (mutant site, 2 35 nt relative to the initiator ATG) and Sac I and ligating into the corresponding sites of pAD4M (42). The construction of pAD4M-SST2-his was described previously (25). pRS316-ADH-SST2 was constructed by digesting pRS315-GAL-SST2 (42) with Bam HI (mutant site, 2 29 nt relative to the initiator ATG; multiple cloning site adjacent to the SST2 Hin dIII at nucleotide 3539) and ligating into the corresponding site in pRS316-ADH. pRS316-ADH-N-SST2 m M 250 ml per 11 liters Cells were further disrupted using a steel bead- beater (Biospec) and salt, with 3 30-s pulses, once every 90 s. The remaining procedures were carried out at tem- perature. The disrupted cells were rocked for 75 min and then centrifugation 3840 3 g for 20 min and filtration The soluble material was mixed with 3 ml of equilibrated Superflow nickel-nitrilotriacetic acid resin (Qiagen) for 60–90 min, into an HR 10/10 (Amersham Pharmacia Biotech) column, and washed using 10 column volumes of Urea Buffer, 250 m M NaCl, and 15 m M imidazole at 1.5 ml/min, followed by 10 column volumes of Urea Buffer at 1 ml/min. Sst2 was eluted in 10 column volumes of Urea Buffer and 75 m M imidazole at 1 ml/min. Peak fractions were pooled, concentrated, and desalted using an Ultrafree-10 (Millipore) filter. The final purified product (in 75 m l) was resolved by 11% SDS-PAGE and transferred to polyvinylidene difluoride (ProBlott, Applied Biosystems). visualized using a Coomassie Blue stain (0.025% 50% The 36-kDa band was excised and for N-terminal sequencing using an Applied Biosystems Procise on-line

The actions of a vast array of chemical and sensory stimuli are mediated through G protein-coupled receptors. In the yeast Saccharomyces cerevisiae, the ␣-factor pheromone binds a receptor (Ste2), which activates a G protein and triggers a cas-cade of events leading to cell fusion and mating. G protein activation entails GTP binding to the ␣ subunit (Gpa1), dissociation of Gpa1 from the ␤␥ subunits (Ste4/Ste18), and activation of effector molecules (Ste5, Cdc24, and Ste20) that propagate the signal. Upon GTP hydrolysis, the G protein subunits reassociate and signaling stops. The RGS 1 protein Sst2 attenuates G protein signaling by accelerating GTP hydrolysis and promoting subunit reassociation (1). RGS activity is essential for normal signal regulation in vivo. A disruption of the SST2 gene can increase pheromone sensitivity by 100 -300-fold. Conversely, overexpression of SST2 can dampen the pheromone response substantially (2).
The mechanism of RGS action has been well characterized through detailed biochemical and biophysical analysis of purified components (3)(4)(5). RGS proteins act by binding and stabilizing three "switch" regions that undergo conformational change upon GTP hydrolysis. Stabilization of the transition state conformation appears to lower the energy of activation, leading to a 10 -1000-fold increase in the rate of the reaction (6 -9).
All RGS proteins have a common, conserved "RGS core domain" of ϳ120 amino acids, which, for several RGS proteins, has been shown to be necessary and sufficient for their GTPase accelerating protein (GAP) activity. For instance, Wilkie and colleagues (10) have demonstrated that the RGS domains of RGS4, RGS10, and GAIP retain full GAP activity for G i ␣ in vitro. Several other RGS proteins are considerably larger and contain additional domains or motifs that may be recognized by proteins other than G␣ (11)(12)(13)(14)(15)(16). For example, the RGS protein p115RhoGEF has one domain that acts as a GAP for G 13 ␣ and a second domain that acts as a GDP-GTP exchange factor for RhoA (17,18). These findings underscore the view that some RGS proteins are not simply GAPs but have separate functions that link them to other signaling pathways.
Recent studies have addressed the question of how RGS proteins are themselves regulated. Several mechanisms have been established, such as alternative splicing (19), regulation of transcription (20,21), altered localization (22)(23)(24), phosphorylation (25)(26)(27), palmitoylation (28,29), and binding of regulatory proteins (14, 16, 30 -32). Sst2, in particular, has been shown to be regulated by transcription and phosphorylation. SST2 mRNA levels increase by at least 5-fold in response to pheromone stimulation (21). This translates to a comparable increase in protein expression levels (2). Also in response to pheromone, Sst2 is stoichiometrically phosphorylated at Ser-539. This phosphorylation leads to an electrophoretic mobility shift, from 82 to 84 kDa, and appears to slow the overall rate of degradation of the phosphorylated 84-kDa species (25).
Here we present a novel and previously undescribed mechanism by which cells can post-translationally regulate RGS function, proteolytic processing. In the course of our studies on Sst2 phosphorylation, we noted that the mobility shift of fulllength Sst2 paralleled that of a much smaller protein also recognized by our Sst2 antibodies. This smaller protein corresponds in size to the RGS core domain in Sst2 (ϳ36 kDa). The experiments described here were aimed at testing the possibility that this Sst2 fragment is expressed and functional in vivo. We show that the 36-kDa product is the result of endoproteolytic processing of the full-length protein, that processing is regulated, and that this processing event leads to profound alterations in the activity and subcellular distribution of Sst2.
pAD4M-SST2 was constructed by digesting SST2 with SalI (mutant site, Ϫ35 nt relative to the initiator ATG) and SacI and ligating into the corresponding sites of pAD4M (42). The construction of pAD4M-SST2his was described previously (25). pRS316-ADH-SST2 was constructed by digesting pRS315-GAL-SST2 (42) with BamHI (mutant site, Ϫ29 nt relative to the initiator ATG; multiple cloning site adjacent to the SST2 HindIII at nucleotide 3539) and ligating into the corresponding site in pRS316-ADH. pRS316-ADH-N-SST2 was constructed by PCR so as to include a mutant BamHI site at position Ϫ1 nt with respect to the initiator ATG, SST2 codons 1-392, a Myc epitope tag (DKLDLEEQKLI-SEEDLLRK-STOP), and an EcoRI site three nucleotides after the stop codon. The resulting PCR product was cloned into the BamHI and EcoRI sites of pRS316-ADH. pRS315-ADH-C-SST2 (also known as ADHleu-C-SST2) was constructed by PCR so as to include a mutant BamHI site at position Ϫ1 nt with respect to Met-411, codons 411-698 of SST2, and an EcoRI site immediately following the stop codon. The PCR product was cloned into the BamHI and EcoRI sites of pRS316-ADH and then transferred as a PvuI-PvuI cassette into the corresponding sites of pRS315. pPEC-GST-SST2 was obtained from K. Madura, (Rutgers University); its construction was described previously (referred to as pPEC9) (43).
pRS423-FUS1-lacZ was constructed by inserting a HindIII-HindIII fragment containing the FUS1-lacZ cassette from pMD56 (44) into the HindIII site of pRS425 to yield pRS425-FUS1-lacZ. This product was digested with XhoI and EagI and ligated into the corresponding sites of pRS423.
All PCR-amplified products were confirmed by DNA sequencing (W. M. Keck Biotechnology Resource Laboratory, Yale University).
Preparation of Whole Cell Lysates-Cells were grown at 30°C in selective media to mid-log phase and treated with 2.5 M ␣-factor pheromone for 1 h, unless otherwise indicated. Temperature-sensitive mutants were grown at 24°C to early-log phase and then shifted to 37°C for 3 h, in the last hour they were treated with 2.5 M ␣-factor. Cells were harvested by centrifugation at 2000 ϫ g for 10 min at 24°C, then resuspended (1.5 ϫ 10 6 cells/l) in 1ϫ SDS-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer (62.5 mM Tris-HCl, pH 6.8, 10% glycerol, 2% SDS, 1% 2-mercaptoethanol, 0.0005% bromphenol blue), and boiled for 10 min. In some cases, cells were washed and stored briefly in 10 mM NaN 3 , on ice. The cells were disrupted by glass bead vortex homogenization (Sigma, G-8772) for 4 min and centrifuged at 16,000 ϫ g for 2 min. The supernatant was collected and stored at Ϫ20°C. Lysates were reheated at 37°C for 20 min before SDS-PAGE and transfer to nitrocellulose. Immunoblots using antibodies to Sst2 (2), Pma1 (45), GST (from J. Steitz, Yale University), the Myc epitope tag (46), and Pgk1 (47) were carried out as described (48).
Purification-YDM400 cells transformed with pAD4M-SST2-his were grown to A 600 nm ϳ1.0, chilled, mixed with 10 mM NaN 3 , and harvested by centrifugation. They were then washed once in 10 mM NaN 3 and rapidly frozen in liquid nitrogen. On the day of the purification, cells were thawed in Urea Buffer (6 M urea, 100 mM Na 2 H 2 PO 4 , 10 mM Tris, and 10 mM 2-mercaptoethanol, pH 8.0), 250 mM NaCl, and 15 mM imidazole (at 250 ml per 11 liters of original culture) at room temperature. Cells were further disrupted using a stainless steel beadbeater (Biospec) packed in ice and salt, with 10 ϫ 30-s pulses, once every 90 s. The remaining procedures were carried out at room temperature. The disrupted cells were rocked for 75 min and then clarified by centrifugation at 3840 ϫ g for 20 min and paper filtration (Whatman No. 1). The soluble material was mixed with 3 ml of equilibrated Superflow nickel-nitrilotriacetic acid resin (Qiagen) for 60 -90 min, packed into an HR 10/10 (Amersham Pharmacia Biotech) column, and washed using 10 column volumes of Urea Buffer, 250 mM NaCl, and 15 mM imidazole at 1.5 ml/min, followed by 10 column volumes of Urea Buffer at 1 ml/min. Sst2 was eluted in 10 column volumes of Urea Buffer and 75 mM imidazole at 1 ml/min. Peak fractions were pooled, concentrated, and desalted using an Ultrafree-10 (Millipore) filter. The final purified product (in 75 l) was resolved by 11% SDS-PAGE and transferred to polyvinylidene difluoride (ProBlott, Applied Biosystems). Protein was visualized using a Coomassie Blue stain (0.025% in 50% methanol). The 36-kDa band was excised and submitted for N-terminal sequencing using an Applied Biosystems Procise 494 cLc instrument equipped with an on-line high performance liquid chromatograph (W. M. Keck Biotechnology Resource Laboratory, Yale University).
Pheromone Response Assays-Halo and reporter-transcription assays were performed as described (49) with minor modifications. For the pheromone-dependent growth inhibition assay (halo assay), cultures were grown to saturation (2-3 days), and 100 l was diluted with 2 ml of sterile water, followed by the addition of 2 ml of 1% (w/v) dissolved agar (60°C). This mixture was then poured onto an agar plate containing selective medium. Sterile filter discs were spotted with synthetic ␣-factor pheromone (5 and 15 g for each plate) and placed onto the nascent lawn. The resulting zone of growth-arrested cells was documented after 2 days.
For pheromone-dependent reporter-transcription assays, cells were transformed with pRS423-FUS1-lacZ and grown to mid-log phase. Cultures were then aliquoted (90 l) to a 96-well plate and mixed with 10 l of ␣-factor, in triplicate. Final ␣-factor concentrations ranged from 0 to 100 M. After 90 min at 30°C, ␤-galactosidase activity was measured by adding 20 l of a freshly prepared solution of 83 M fluorescein di-␤-D-galactopyranoside (Molecular Probes, 10 mM stock in Me 2 SO), 137.5 mM PIPES, pH 7.2, and 2.5% Triton X-100 and incubating for 90 min at 37°C. The reaction was stopped by the addition of 20 l of 1 M Na 2 CO 3 , and the resulting fluorescence activity was measured with a multiwell plate reader using 485 nm excitation and 530 nm emission. All determinations were carried out at least twice with similar results, unless otherwise indicated.
Membrane Fractionation-Methods for cell membrane fractionation have been described in detail elsewhere (41). Briefly, cells were grown at 30°C in selective media to mid-log phase, centrifuged at 500 ϫ g for 10 min, and resuspended in rich media (YPD) at A 600 nm ϭ 0.5/ml. Cells were grown for one doubling period, the last hour of which they were treated with 2.5 M ␣-factor. Growth was then stopped by addition of NaN 3 to 10 mM. Approximately 3 ϫ 10 9 cells were centrifuged and washed once with SK buffer (1.2 M sorbitol, 0.1 M KPO 4 , pH 7.5). Spheroplasts were then prepared by resuspending the cells in 10 ml of SK buffer containing 1 mg of zymolyase 100T (Kirin Brewery) and 28.8 mM 2-mercaptoethanol for 45 min at 30°C. All subsequent manipulations were performed at 0 -4°C. Spheroplasts were centrifuged at 500 ϫ g for 10 min, washed once with SK buffer and once with lysis buffer (0.8 M sucrose, 20 mM triethanolamine, pH 8, 1 mM EDTA acid, 1 mM dithiothreitol, and a protease inhibitor mixture containing 1 mM [4-(2-aminoethyl)-benzenesulfonyl fluoride, HCl] and 10 g/ml each of leupeptin, pepstatin, and benzamidine (final concentrations)). Cell pellets were resuspended in 1 ml of lysis buffer and disrupted with 25 strokes of a motor-driven Potter-Elvehjem homogenizer. The lysate was cleared of unbroken cells and debris by centrifuging twice at 500 ϫ g for 10 min. 200 l of this lysate ("cleared lysate") (total, "T") was added to an equal volume of 2ϫ SDS-PAGE sample buffer and boiled for 10 min. For isolation of total cell membranes, approximately 300 l of the cleared lysate was centrifuged at 100,000 ϫ g for 30 min. The top 100 l of supernatant ("S" fraction) was diluted with 100 l of 2ϫ SDS-PAGE sample buffer and boiled for 10 min. The pellet ("P") was resuspended in lysis buffer to the original volume, mixed with an equal volume of 2ϫ SDS-PAGE sample buffer, and boiled for 10 min. For resolution of cell membrane compartments, 606 mg of sucrose was added to 650 l of the cleared lysate (T) and dissolved (final sucrose concentration 70% w/v). The sample was transferred to a Beckman thin walled polypropylene tube and overlaid with 1-ml sucrose solutions of 60, 50, 40, and 30% (w/v) in 10 mM triethanolamine, pH 8, respectively. The samples were then centrifuged in a Beckman SW40Ti swinging bucket rotor for 16 h at 190,000 ϫ g in a Beckman L-80 ultracentrifuge. Sixteen samples of 300 l each were collected from the bottom of the gradient into 100 l of 4ϫ SDS-PAGE sample buffer and boiled for 10 min. Fractions 1-14 were resolved by SDS-PAGE and immunoblotted as described above.

Sst2 Is Proteolytically Processed in Vivo-
The SST2 gene encodes a 698-amino acid protein with a predicted mass of 79,696 Da. Immunoblots of whole cell lysates revealed a protein near the predicted size of the full-length product (ϳ82 kDa) as well as a number of prominent lower molecular weight species, one of which migrates at ϳ36 kDa ( Fig. 1, p36). Since our antibodies are directed to the last 365 residues of Sst2, this low molecular weight species probably corresponds to a C-terminal, proteolytically processed form of the protein. It is unlikely to be derived from another gene product, since it is absent in an SST2-deficient strain (Fig. 1, sst2⌬) and is more abundant in cells that overexpress SST2 from a plasmid (Fig. 1, sst2⌬ ϩ Sst2 o.e.). It is also not the result of alternative mRNA splicing, since SST2 is encoded by a single exon. The multiple bands in the 36-kDa region correspond to phosphorylated and unphos-phorylated forms of two slightly different sized fragments of Sst2 (see below).
Significantly, p36 corresponds in size to the RGS core domain of Sst2 (residues ϳ417-698). For some mammalian RGS proteins, the core domain is sufficient for GTPase activating function in vitro (10), and for signal attenuating activity in cell culture (50,51). Thus we considered whether this naturally occurring fragment of Sst2 is functional as an RGS protein in vivo.
We first set out to confirm the identity of the 36-kDa species. Initially, we examined whether this product extends completely to the C terminus of Sst2. Our approach here was to determine if a similarly sized fragment could be detected using an antibody directed to the extreme C terminus of the protein. Accordingly, a hexahistidine tag (ϪHis) was appended to the 3Ј end of the full-length open reading frame and expressed. The Sst2-His fusion is fully functional, as determined by its ability to complement the sst2⌬ gene disruption mutant (25). Lysates were prepared from cells expressing the Sst2-His fusion and analyzed by immunoblotting. In this case, anti-His antibodies recognize a discrete product of 37 kDa, corresponding in size to the original 36-kDa product, plus the His tag and a short linker sequence (data not shown). Thus it appears that Sst2 is cleaved internally in a manner that leaves the end most C-terminal residues intact.
To determine the site of internal cleavage, we purified and sequenced the N terminus of the 37-kDa Sst2-His product. Cells were disrupted in a denaturing buffer containing 6 M urea so as to solubilize completely Sst2 and to prevent further (nonspecific) proteolysis. Sst2-His was purified by nickel-nitrilotriacetic acid affinity chromatography, resolved by gel electrophoresis, and transferred to a polyvinylidene difluoride membrane. The 37-kDa protein was then visualized by Coomassie Blue staining, excised, and subjected to Edman degradation sequencing. As indicated in Fig. 2A, two overlapping sequences were obtained. The first, Ser-Ser-Asn-Leu-Asn-Lys-Leu-Asp, begins at position 415; the second, Asn-Leu-Asn-Lys-Leu-Asp-Tyr-Val, begins at position 417. Significantly, Asn-417 represents the beginning of the RGS core domain in Sst2 (Fig. 2B).
C-Sst2 Is Functional in Vivo-Once having defined the boundaries of the 36-kDa fragment, we examined whether this truncated form of Sst2 could function in vivo. Specifically, we examined whether expression of the C-terminal portion of Sst2 could regulate pheromone responsiveness in the same manner as the full-length protein. A plasmid containing the C-terminal 288 residues (C-Sst2) was constructed to initiate translation at a naturally occurring Met at position 411 just preceding the site of cleavage. Expression of C-Sst2 was confirmed by immunoblotting (Fig. 3A, C-Sst2). Then activity was tested by two bioassays of the pheromone response. First, a reporter transcription assay was performed using the lacZ gene under the control of the pheromone-inducible promoter from FUS1. As shown in Fig. 3B, expression of C-Sst2 yielded ␤-galactosidase activities comparable to those seen with the expression of vector alone. C-Sst2 was then tested using the pheromone-dependent growth inhibition (halo) assay. Cells were plated and exposed to different amounts of synthetic ␣-factor spotted onto filter disks. As shown in Fig. 3C, cells expressing C-Sst2 exhibit a clear zone of growth inhibition, again similar to that seen with vector alone. These results indicate that, although C-Sst2 is stably expressed, it is not able to complement the loss of SST2 in vivo.
Since cleavage of Sst2 yields a discrete C-terminal product, the same cleavage event most likely produces a complementary N-terminal fragment. To test this, we examined if a fragment of Lysates were prepared from cells expressing the GST-Sst2 fusion and analyzed by immunoblotting using anti-GST antibodies. As shown in Fig. 3D, a 72-kDa band is expressed, corresponding to the N-terminal proteolytic product (46 kDa) plus the GST tag (26 kDa).
We then examined whether the N-Sst2 fragment could function in vivo. A second plasmid encoding the 46-kDa species was prepared, encoding residues 1-392 plus a Myc epitope tag (to monitor expression) in place of residues 393-414. The N-Sst2 construct alone had little or no effect on pheromone responsiveness, by either the transcription induction or growth inhibition assays (Fig. 3, B and C). To examine whether both cleavage products are required for function, we co-expressed N-Sst2 and C-Sst2 in an sst2⌬ mutant strain (51). Co-expression of both products failed to restore Sst2 function by the transcriptional reporter assay (Fig. 3B). However, co-expression of N-and C-Sst2 did yield a substantial reduction in pheromone-mediated growth arrest, more similar to that seen in cells bearing the full-length SST2 gene (Fig. 3C). This uncoupling of the two mating responses is highly unusual and reveals that proteoly-sis dramatically alters the way in which Sst2 functions.
To examine further the effect of proteolysis on Sst2 function, we set out to assess the activity of only the full-length protein.
To do this, we attempted to create a protease-resistant form of Sst2. A series of triple-amino acid substitution mutants were prepared, replacing the residues preceding each of the two cleavage sites (Ser-414 and Ser-416) as well as the intervening residue (Ser-415). Each Ser (a small polar amino acid) was replaced by the chemically distinct residues Phe (large hydrophobic), Ala (small hydrophobic), or Gln (large polar). These mutants were then expressed in an sst2⌬ strain and probed by immunoblotting. In every case, however, the mutants resembled wild-type Sst2 with regard to proteolytic processing, transcriptional induction, and growth arrest response (data not shown).
Processing Alters Sst2 Localization-We then considered whether the change in activity that accompanies the processing of Sst2 is due to a change in subcellular localization. Such a model was proposed to explain enhanced signaling by nonpalmitoylated G␣ (Gpa1), which becomes "mislocalized" to microsomal membrane fractions and thus is less able to sequester G␤␥ (52). First, we used high speed centrifugation to resolve membrane and cytosolic fractions. As shown previously, full- FIG. 2. Sequencing of the p36 fragment. A, the ϳ37-kDa Sst2-his product was purified and subjected to Edman degradation sequencing. Top, values represent the amount (in picomoles) of each amino acid (single letter code) obtained for each of the 8 cycles of sequencing. Bottom, two overlapping sequences from Sst2 were detected, beginning at Ser-415 (Sequence 1) and Asn-417 (Sequence 2). B, CLUSTALW 1.8 was used to align Sst2 (NCBI accession number AAB67534) with representative mammalian RGS proteins, GAIP (NCBI accession number NP_005864), RGS1 (NCBI accession number Q08116), RGS3 (NCBI accession number S78089), RGS4 (NCBI accession number S78221), and RGS7 (NCBI accession number AAD34290). Numbers indicate the first amino acid residue in the alignment. Asterisks are above the N-terminal residues of the two purified p36 species. Boxes indicate residues that are identical (black) or similar (gray) in at least 4 of the aligned sequences. Numbers within parentheses indicate amino acids omitted from the alignment.
length Sst2 is predominantly in the pellet (membrane) fraction. In contrast, the processed 36-kDa form of the protein is predominantly in the supernatant (cytosol). The C-Sst2 protein is also in the supernatant fraction, and N-Sst2 is largely in the pellet fraction (Fig. 4A). Co-expression of N-Sst2 and C-Sst2 does not substantially alter the distribution of either protein.
We then examined if N-Sst2 is specifically associated with the plasma membrane or a subcellular membrane compartment. Cell lysates were subjected to centrifugation through a 70 -30% sucrose flotation gradient, and each fraction was re-solved by SDS-PAGE and immunoblotting. As shown in Fig.  4B, full-length Sst2 is present in the plasma membrane fractions and in fractions that contain cytosol and Golgi/other microsomal membranes (2). In contrast, N-Sst2 is largely absent from the plasma membrane and is instead enriched within the microsomal membrane fractions (Fig. 4C, Myc Ab). C-Sst2 and p36 are found in fractions that contain cytosol (Fig. 4, B  and C). Again, co-expression of N-Sst2 and C-Sst2 does not alter the distribution of either product. These results suggest that Sst2 processing leads to the release of a soluble C-terminal FIG. 3. Co-expression of N-and C-Sst2 partially restores Sst2 function. Strain YDM400 was transformed with plasmids containing no insert (vector; pRS316-ADH, pRS315), full-length Sst2 (wt; pRS316-ADH-SST2, pRS315), N-Sst2 alone (N-Sst2; pRS316-ADH-N-Sst2, pRS315), C-Sst2 alone (C-Sst2; pRS316-ADH, pRS315-ADH-C-Sst2), or both N-Sst2 and C-Sst2 (N-Sst2 ϩ C-Sst2: pRS316-ADH-N-Sst2, pRS315-ADH-C-Sst2). A, immunoblot analysis. Cell lysates were prepared and subjected to 11% SDS-PAGE and immunoblotting with anti-Myc antibodies (N-Sst2) or 8% SDS-PAGE and immunoblotting with anti-Sst2 antibodies (C-Sst2/p36). For N-Sst2, cleared lysates were used in order to improve detection with the Myc antibody, as described in Fig. 4. B, reporter transcription assays. Cells were transformed with an additional plasmid containing a pheromone-responsive promoter and the lacZ reporter gene (pRS423-FUS1-lacZ). Cells were treated with the indicated concentrations of ␣-factor for 90 min and assayed for ␤-galactosidase activity. The modest rightward shift in the EC 50 for cells expressing N-Sst2 (versus vector or C-Sst2 alone) was not observed in all experiments. C, halo assays. Paper disks containing either 5 or 15 g of ␣-factor were placed on a nascent lawn of cells, which was allowed to grow at 30°C for ϳ48 h. Co-expression of N-and C-Sst2 produced considerably smaller halos than N-Sst2 or C-Sst2 alone but larger than full-length wild type Sst2. D, immunoblot of N-terminal cleavage product. Strain YDM400 was transformed with either pPEC GST-Sst2 (GST-Sst2) or pRS316-ADH-SST2 (Sst2). Cell lysates were subjected to 8% SDS-PAGE and immunoblotting with anti-GST antibodies. Specific immunoreactive bands include GST-Full-length (ϳ106 kDa) and GST-N terminus (ϳ72 kDa), which represents the N-terminal cleavage product of Sst2. E, schematic of the two constructions designed to mimic the Sst2 cleavage products. N-Sst2 contains residues 1-392 and has a C-terminal Myc epitope tag. C-Sst2 contains residues 411-698. wt indicates full-length Sst2. fragment and a redistribution of the N-terminal fragment to a non-plasma membrane compartment. This redistribution of Sst2 away from the plasma membrane may be responsible for the functional differences between processed and full-length Sst2.
Regulation of Processing-The evidence presented above suggests that endoproteolytic processing can regulate Sst2 localization and function. To help exclude the possibility that the 36-kDa species is merely an intermediate in the degradation of Sst2, we examined processing in a number of protein degrada- FIG. 4. Localization of proteolyzed fragments. wt, N-Sst2, C-Sst2, and N-Sst2ϩC-Sst2 strains are as described in Fig. 3. A, cell fractionation. Cells were disrupted, and the membrane and cytosolic fractions were separated by high speed centrifugation. Cleared total lysate (T), membrane (pellet, P), and cytosolic (soluble, S) fractions were resolved by 11% SDS-PAGE and immunoblotting to detect full-length Sst2, p36, and C-Sst2 (Sst2 Ab). The membranes were then stripped and re-probed with antibodies to detect N-Sst2 (Myc Ab). Full-length Sst2 and N-Sst2 are primarily in the membrane fraction, and C-Sst2 and p36 are predominantly in the supernatant fraction. The fractionation pattern does not change substantially when N-Sst2 and C-Sst2 are co-expressed. B, subcellular fractionation of full-length and processed Sst2. Cleared lysates from wild type cells were subjected to sucrose density gradient centrifugation, and the fractions were resolved by 8% SDS-PAGE and immunoblotting with anti-Sst2 antibodies (top) or anti-Pma1 antibodies (plasma membrane marker, bottom). A portion of the full-length Sst2 fractionates with the plasma membrane, whereas p36 does not. C, subcellular fractionation of N-Sst2 and C-Sst2. Cleared lysates from cells expressing C-Sst2 alone (top left), N-Sst2 alone (bottom left), or N-Sst2 and C-Sst2 (right) were fractionated by sucrose density gradient centrifugation, resolved by 10% SDS-PAGE, and immunoblotted with anti-Sst2 (C-Sst2), anti-Myc (N-Sst2), or anti-Pma1 antibodies. The fractionation pattern does not change when N-Sst2 and C-Sst2 are co-expressed. tion mutants, including conditional alleles affecting the vacuolar proteases (pep4⌬, prb1⌬), the 20 S core proteasome (pre1-1, pre2-2), and the 26 S proteasome (cim3-1). These mutants and their isogenic wild-type strains were transformed with an Sst2 plasmid, grown at 24°C, and then shifted to 37°C (the restrictive temperature for the mutants) for 3 h. As shown in Fig. 5, the abundance of full-length Sst2 is dramatically increased in each of the proteasome mutants but is unaffected by the vacuolar protease mutant. This suggests that Sst2 turnover is mediated by the 26 S proteasome. However, the relative abundance of the 36-kDa product is unaltered in each of the three mutant strains tested. This suggests that Sst2 processing is not mediated by any of the known components of the protein degradation machinery.
Finally, we examined whether Sst2 processing is regulated by pheromone stimulation or by some intracellular component of the signaling cascade. As shown previously (25), Sst2 is phosphorylated in response to pheromone treatment, and this modification leads to an electrophoretic mobility shift (from 82 to 84 kDa) (Fig. 6A). The pheromone-dependent mobility shift of the 36-kDa form of Sst2 mirrors that of the full-length protein (Fig. 6A, p36 versus p36-P). Pheromone stimulation does not appear to influence processing of Sst2, however, since the abundance of p36 is not dramatically altered. To determine whether Sst2 processing requires some other component of the signaling cascade, lysates were prepared from cells lacking the receptor (ste2⌬), the G protein ␤ or ␥ subunits (ste4⌬, ste18⌬), components of the downstream kinase cascade (ste20⌬, ste11⌬, kss1⌬/fus3⌬), or the downstream transcription factor (ste12⌬). In every case, expression of the 36-kDa species is greatly diminished compared with the wild-type strain (Fig. 6B). Expression of the full-length species is in some cases reduced but never to the extent of the 36-kDa species. These observations indicate that Sst2 processing, although not pheromone-regulated, nevertheless requires an intact signal transduction apparatus. DISCUSSION A fundamental aspect of cell regulation is the ability to modulate protein activity (53). One way proteins can be modulated is through post-translational modifications, most notably through phosphorylation, glycosylation, and lipidation. Another common mechanism is allosteric regulation, through binding of ligands, substrates, or products. Finally, there is growing interest in regulation through limited proteolysis.
Here we present data showing that Sst2 undergoes endopro- The PRE1 and PRE2 gene products are required for 20 S proteasome activity; PEP4 and PRB1 are required for vacuolar protease activity; CIM3 is required for 26 S proteasome activity. Cells were grown to early log phase at 24°C and then shifted to 37°C (restrictive temperature) for 3 h. Whole cell lysates were resolved by 11% SDS-PAGE and immunoblotting with anti-Sst2 antibodies. The proteasome mutants exhibit elevated levels of full-length Sst2 and the p36 species, indicating that the proteasome is needed for Sst2 turnover but not for production of p36.
FIG. 6. Expression of p36 is regulated. A, strain YPH499 was transformed with an Sst2 overexpression plasmid (pRS316-ADH-SST2, Sst2) or the empty vector (pRS316-ADH, vector). Cells were treated with 2.5 M ␣-factor for 1 h, except as indicated (Ϫ ␣-factor). Whole cell lysates were resolved by 11% SDS-PAGE and immunoblotting with anti-Sst2 antibodies. Pheromone stimulation leads to phosphorylation and reduced mobility of the full-length and p36 species. However, pheromone is not required for the production of p36. B, strain YPH499 was transformed with an empty vector (pRS316-ADH, vector) or an Sst2 overexpression plasmid (pRS316-ADH-SST2, wt). YPH499-derived mutants lacking the ␣-factor receptor (ste2⌬), G␤ subunit (ste4⌬), G␥ subunit (ste18⌬), the downstream p21-activated protein kinase homologue (ste20⌬), the MAP kinase kinase kinase (ste11⌬), two functionally redundant MAP kinases (fus3⌬/kss1⌬), or the transcription factor (ste12⌬) were also transformed with the Sst2 plasmid. Whole cell lysates were resolved by 11% SDS-PAGE and immunoblotting to detect Sst2 (Sst2 Ab) or Pgk1 (Pgk1 Ab, to confirm equal loading). All of the signaling components tested appear important for production of p36 but not full-length Sst2 or the Pgk1 control. teolytic processing and redistribution within the cell. This processing event is regulated, in so far as it requires an intact G protein signaling pathway. Moreover, the processing event is functionally significant, because it profoundly alters cell responsiveness to pheromone.
There may be some particular advantages to proteolysis as a means of regulating RGS action. First, proteolysis is irreversible, in contrast to phosphorylation or palmitoylation. This is particularly important for inherently irreversible cell processes, such as mitosis or the mating of haploid yeast cells. Second, large reservoirs of intact protein can accumulate and become rapidly activated as needed. Consequently, proteolysis may be considerably faster than gene transcription or protein translation.
There is ample precedent for regulation of cell signaling events through limited proteolysis. In some cases, proteolytic processing is essential for protein activity or maturation. Many proteases are themselves synthesized as inactive precursors (zymogens) and then proteolytically processed to form active proteases (54). One prominent example is the serum protease thrombin, which is activated through cleavage of pro-thrombin. Activated thrombin in turn catalyzes the proteolytic removal of an extracellular portion of the thrombin receptor (55). In platelets, thrombin receptor proteolysis leads to activation of a G protein, cell aggregation, and clot formation. As another example, the calcium-dependent protease calpain can cleave arrestin, a factor that binds and uncouples receptors from the G protein. Only the receptor-bound form of arrestin is a substrate, and once proteolyzed it is less able to dissociate from the receptor, apparently leading to prolonged receptor desensitization (56). Calpain has also been shown to cleave the G s ␣ subunit, leading to persistent activation of adenylyl cyclase (57).
Calpain can modulate cell signaling of non-G protein-coupled receptors as well. In fibroblasts, limited proteolysis of talin, ezrin, and focal adhesion kinase diminishes their ability to promote cell adhesion. Moreover, cleavage occurs in a polarized manner at the trailing edge, thereby allowing the cell to move in a unidirectional manner (see Ref. 58).
There are some intriguing parallels between processing of cell adhesion molecules and the processing of Sst2. Fibroblast motility requires an intact MAP kinase signaling pathway, as well as a functional calpain protease (58). Likewise, deleting any component of the MAP kinase signaling cascade in yeast can block the action of the protease that cleaves Sst2. In yeast, as in fibroblasts, there are circumstances when signaling proteins are regulated in a spatially (or temporally) restricted manner. For instance, the receptor (Ste2) and effector kinase (Ste20) are localized predominantly to the tip of the mating projection where cell fusion takes place (Fig. 7, top) (59,60). Since these components are required for cleavage of Sst2, processing might also occur primarily within the mating projection (Fig. 7, bottom). This could represent a positive feedback loop, allowing enhanced signaling at the site of cell fusion and reduced signaling at the more distal regions of the cell.
While the N-and C-Sst2 fragments are clearly sufficient for some aspects of RGS activity in vivo, important differences do exist. Most notably, the fragments restore Sst2 regulation of the growth arrest response but not the transcriptional induction response. A similar pattern of selective regulation has previously been described in yeast lacking the G␤␥-binding protein Plp1; these mutants exhibit slightly enhanced transcriptional induction but no change in long term growth arrest (61). These observations raise the possibility that new genetic screens for mutants in which transcriptional induction and growth arrest are uncoupled may yield additional novel regulators of G protein function.
The differences in function exhibited by full-length and processed forms of Sst2 may stem from differences in their subcellular distribution. A similar model has been proposed for the p35/Nck5a protein, where limited proteolysis results in its mislocalization to the cytoplasm and a diminished ability to inhibit Cdk5 kinase. Persistent activation of Cdk5 leads to constitutive phosphorylation and aggregation of Tau, the principal component of neurofibrillary tangles in Alzheimer's patients (62). Another possibility is that proteolysis diminishes Sst2 GAP activity. This seems less likely, given that mutationally truncated mammalian RGS proteins are fully active in vitro (10,63). In vitro GAP assays comparing the full-length and truncated forms of Sst2 are also under way. However, interpretation of these experiments may be complicated by the presence of Sst2 degradation products in preparations of purified recombinant protein (64) and the present inability to measure accurately small differences in Sst2 GAP activity.
A challenge for the future is to identify the protease that degrades Sst2. There are a number of candidates encoded by the yeast genome, including 19 Cys proteases, 15 metalloproteases, 11 Ser proteases, and 8 Asp proteases (65). So far we have only ruled out the proteasome and vacuolar proteases as being directly involved. Of the remaining candidates, the relevant protease most likely is cytoplasmic and has broad sequence specificity; it must also be capable of limited proteolysis. We are currently investigating whether any of the known proteases are required for Sst2 proteolysis, and we are paying particular attention to any Cys proteases that resemble calpain or caspase. Once the Sst2 protease has been identified, we can begin to characterize its function in vivo and determine how other signaling components influence its activity.
In summary, RGS proteins are well known to promote G protein GTPase activity and desensitization. Our objective here was to determine how one regulator of G protein signaling is itself regulated post-translationally. Given the striking similarities in yeast and human signaling pathways, it is likely that the novel regulatory mechanism described here will be recapitulated for other multidomain RGS proteins in other organisms.