Nucleotide Dependent Binding of the GTPase Domain of the Signal Recognition Particle Receptor β -Subunit to the α -Subunit.

of the amino-terminal transmembrane of SR a translocation that overlaps the GTPase domain. We that the domain of SR α that binds SR β does by the nucleotide form of the GTPase of SR β A mutant SR β XTP Our results on of


4
SRβ GTPase activating protein (10). However, SRβ also functions as a membrane anchor for SRα (8,11)via an interaction with an independently folded domain (SRX2) at the amino terminus of SRα (11). A translational pause encoded in the SRα mRNA ensures efficient targeting of SRα to the ER membrane by allowing the SRX2 domain to fold and interact with SRβ before translation of SRα has completed (12). Once formed, the dimer is stable in 1% non-ionic detergent and 500 mM KOAc.
Therefore, it was surprising that in yeast the transmembrane domain of SRβ is not required for SR function or for dimerization with SRα (13). Moreover, SR binding to the ER membrane was dramatically decreased, but it was not abolished, in yeast expressing a mutant SRβ lacking the transmembrane region (13).
We have examined the role of SRβ in membrane assembly and heterodimerization of SR using a mammalian cell free system. Our data demonstrate that sequences amino terminal of the SRβ GTPase are required for membrane binding and for orienting SRβ with respect to the membrane. Moreover, this region of the molecule partially suppresses a cryptic ER translocation signal that overlaps the SRβ GTPase domain. Surprisingly, the SRβ GTPase domain is both necessary and sufficient for interaction with SRα via the SRX2 domain. Mutation of the SRβ GTPase to favour XTP over GTP allowed us to demonstrate that in the nucleotide free state, SRβ is unable to bind SRα. Along with the recent observation that the ribosome increases the GTPase activity of SRβ (10), these data suggest that binding of SRα to the ER membrane may be regulated during targeting and transfer of nascent polypeptides to the translocon.

EXPERIMENTAL PROCEDURES
Materials and General Methods-General chemical reagents were obtained from either by guest on  http://www.jbc.org/ Downloaded from 6 Plasmid pMAC853 encodes SR-TM, a fusion of the carboxyl terminal 206 amino acids of canine SR with an amino terminal HA epitope tag (MYPYDVPDYAA) 2 . To assemble this plasmid the coding sequence for SR md in pMAC690 was replaced by a PCR product generated by amplifying the appropriate region of the coding sequence from pMAC455 using an amino terminal primer CATGCCATGGCTAAGTTCATCCGGAGCAGA and a carboxyl terminal primer complementary to the T7 promoter. The PCR product was digested with NcoI and EcoRI and subcloned into the vector cut with the same enzymes.
pMAC747 encodes SRC1, a version of SR md lacking the carboxyl terminal six amino acids. The coding region for SRC1 was amplified by PCR from pMAC455 using an amino terminal primer complementary to the SP6 promoter, and a carboxyl terminal primer GCTCTAGACCTACTTCTCCAGGTCCTGGATG. The PCR product was cut with NcoI and XbaI and subcloned into pMAC690 in place of the coding sequence for SR md . pMAC1300 encodes SRC1-ûTM, and was created by digesting pMAC747 (encoding SRC1) with BspEI and BamHI, and subcloning cloning it into the corresponding region of pMAC853 (encoding SR-ûTM).
pMAC1056 encodes SRD4, the carboxyl terminal 195 amino acids of canine SR. The coding region for SR was amplified by PCR from pMAC853 using an amino terminal primer CATGCCATGGCTGTTCTTCTTGTTGGC and a carboxyl terminal T7 promoter primer. The PCR product was cut with NcoI and BamHI and used to replace the SR md coding region in pMAC690 digested with the same enzymes. pMAC1057 encodes SRD5, the carboxyl terminal 182 amino acids of canine SR. The appropriate coding region was amplified from pMAC853 using an amino terminal primer 7 CATGCCATGGGATTACTGTTTGTCAGGTTGTTAAC and a carboxyl terminal T7 promoter primer. This fragment was cloned into pMAC690 using NcoI and BamHI as above.
pMAC1082 encodes SR-ûloop, a version of SR-ûTM containing a deletion of amino acids 185 to 219. To prepare the plasmid for PCR, an endogenous HindIII site was removed from a noncoding region of the plasmid by cutting pMAC853 with HindIII, treating with Klenow fragment and ligating. The resulting plasmid (pMAC853-HindIII) was amplified with a 5' primer

A A T T A A G C T T G G A A A G A A A G G C A A A G A A T T T G A G T an d a 3 ' p ri m er
AATTAAGCTTTGCGGATTTTGCCATTGTAATG. The PCR product (the entire plasmid except for the deleted region) was digested with HindIII and ligated to yield pMAC1082. pMAC1363 encodes SRloop2, a version of SRûTM with amino acids 185-219 replaced with the pseudo-random sequence RSTISLQQASPLTGTPDKSGRSATVLAQQLALNKL. This sequence was created as a pair of oligonucleotides containing a 5' BglII and a 3' HindIII site. The oligos were ligated into pSPUTK to give pSPUTK-loop2. The carboxyl terminal region of SR and T7 RNA polymerase promoter were amplified from pMAC853-HindIII using a 5' primer containing a HindIII site and a 3' primer containing a NheI site, and ligated into pSPUTK, giving pSPSR3'. pSPUTK-loop2 was digested with BglII and HindIII, and the fragment was ligated into pSRSR3' to give pSPSR3'loop2. The amino terminal region of pMAC853 including the SP6 RNA polymerase promoter and coding regions for the HA tag and the amino terminal part of SR were amplified using a 5' primer containing an AflIII site and a 3' primer containing a BglII site, and ligated into pSPSR3'loop2 to assemble the coding region for SRloop2.
All of the GTPase mutants were created by oligonucleotide-directed point mutation and PCR using a method described previously (18) using an ApaI site. All plasmids were sequenced to 8 confirm the presence of the desired mutations.
Quantification of in vitro translation products-1 µL of translation reaction products were added to 50 µL 200 mM NaOH and 2.5 µL H 2 O 2 , and spotted onto glass microfibre filters (Whatman). The filters were soaked in 10% w/v trichloroacetic acid (TCA) for 10 minutes, washed 3 times in 5% TCA, once in 95% ethanol, and allowed to dry. The dry filters was added to scintillation fluid, and the measured radioactivity was converted to fmoles of protein by the following equations: , 2 2 10 12 where Ci is the number of curies per sample, %P is the percentage of counts from the translated protein, M is the number of methionines in the protein, and EF is the efficiency factor of the scintillation counter.

( )
where f is the radioactive decay factor and SA 0 is the specific activity of the isotope. Periodically, 0.5 µL of reaction products were run on SDS-PAGE gels and subjected to Phosphorimager analysis to determine the percentage of radioactivity arising from the product of interest (%P), as compared to non-specific translation products. The determination of fmoles of protein rather than simply measuring radioactivity and correcting for the number of methionines was used to permit repeat experiments to be performed with identical quantities of in vitro translation products irrespective of the specific activity of the 35 S methionine.
Immunoprecipitations and Proteolysis-For immunoprecipitation of dimers consisting of SRα and the specified SRβ mutant, equimolar amounts of the two molecules were incubated for 15 minutes at 24C. The reactions were then diluted in TXSWB (100 mM Tris-Cl, pH 8.0, 500 mM membranes, microsomes were added to 20 microliter translation reactions. After translation for 60 minutes the membranes and membrane bound proteins were isolated by gel-filtration chromatography and analyzed by proteolysis as described previously (19).
GTP depletion-Depletion of small molecules from translation reactions was performed essentially as described (20). Briefly, translation reactions were loaded onto a 20X bed volume of Sephadex G25 resin equilibrated in 250 mM sucrose, 25 mM HEPES-KOH, pH 7.5, 10 mM KOAc, 1 mM DTT, 5% glycerol and centrifuged at low speed at 4C. The flow through was collected and the procedure was repeated. To determine how effectively this procedure removed nucleotides a control reaction containing 35 S labeled ATP was analyzed. For nucleotide titration experiments, the appropriate amount of nucleotide was added to 10 µL of depleted SR reaction products and incubated for 20 minutes on ice prior to mixing with translation reactions for SRα for 15 minutes at 24C.

RESULTS
To examine membrane assembly and identify the SRα-binding domain within SRβ, a series of deletion mutants were created ( Figure 1). These mutants were synthesized in a reticulocyte lysate cell free system and assayed for both correct assembly into ER membranes and for heterodimerization with SRα. As a starting point for this analysis we used the SRβ md chimera consisting of the first ten amino acids of murine SRβ followed by residues 10 to the carboxylterminus (residue 265) of canine SRβ (11). This chimera, referred to here as SRβ, was used because it has been previously demonstrated to form heterodimers with canine SRα that are indistinguishable from the canine proteins found on microsomes (11). A glycosylation site introduced at the extreme amino terminus of SRβ was efficiently glycosylated when the protein was synthesized in the presence of RMs (data not shown), confirming the hypothesis that SRβ spans the membrane as a typical type I integral membrane protein.
Therefore, the amino acids at the amino-terminus of SRβ (approximately residues 1-40) reside within or span the ER membrane, where they are unlikely to come into contact with SRα. Thus, the sequence containing a series of positive charges (amino acids 40-58) following the transmembrane domain, is the most amino-terminal region of SRβ that is a candidate for binding to SRα.
Deletion of these residues from full length SRβ md (SRβ-∆ch) inverts the orientation of some of the type I molecules in the ER membrane. As a result, the topology of some of the non- Since the GTPase domain of SRβ appears to be both necessary and sufficient for binding to SRα, a series of GTPase point mutations (Table I) were created in SRβ−∆TM to determine whether a functional GTPase domain is required for SR dimerization. Similar GTPase mutations have also been examined for functional complementation of SRβ deletion in yeast (13).
Coimmunoprecipitation experiments performed with these GTPase mutants illustrate that only a subset were competent for SR dimerization (Figure 4 and Table I). Identical results were obtained when binding to SRX2 was assayed (Figure 4). The mutants that bind to SRα include the G118L and H119L mutations believed to reduce the GTPase activity of GTP binding domains (21)(22)(23)(24) and D181N that alters the nucleotide binding preference from GTP to XTP (25,26). Although the D181N mutation alters binding preference it does not eliminate GTP binding. Therefore, since our in vitro translation reactions contain at least 1 mM GTP it is likely that when synthesized in vitro, the D181N mutant is still primarily in the GTP bound form. Of the mutants that do not bind SR.
K75I is predicted to have much reduced nucleotide affinity (27) while N178K is believed to be structurally unstable (24). Thus the common feature of all of the mutants that bind SRα is that they are all expected to bind nucleotide.
To determine whether the GTP-bound, GDP-bound form or the empty form of SRβ binds to SRα, small molecules, including GTP, were removed from in vitro translation reactions by gel filtration chromatography. Since the in vitro translation reactions contain very small amounts of SRβ, (approximately 100 fMoles) there was a practical limit on the size of gel filtration column that could be used, before there was unacceptable loss of SRβ due to non-specific binding. As a result, we were unable to reduce the amount of GTP in the reaction mixture to significantly less than 100 nM. Since SRβ has a K d for GTP of approximately 20 nM (10) sufficient GTP/GDP remains tightly bound to SRβ that gel-filtration chromatography prior to incubation with SRα did not reduce binding of SRβ−∆TM to SRα ( Figure 5). However, after gel filtration chromatography the D181N mutant no longer dimerized with SRα. The D181N mutant is expected to possess a lower than wildtype affinity for GTP therefore, we presume that gel filtration chromatography reduced the concentration of GTP sufficiently to unload the D181N mutant, which was then unable to bind SRα ( Figure 5). These data suggest that the empty form of the SRβ GTPase domain is unable dimerize with SRα. To confirm this result and examine the nucleotide specificity for binding, this experiment was repeated and dimerization was assayed after adding back specific nucleotides.
When dimer formation was assayed with added ATP (up to 1 mM) co-precipitation of D181N and SRα was not observed. In contrast, adding GTP, GDP, XTP or XDP restored some SRα-binding by D181N ( Figure 6). The effect of the xanthosine nucleotides is more pronounced than the guanosine nucleotides, consistent with the predicted preference of D181N for XTP compared to GTP. Surprisingly, there seems to be little preference for the di-or tri-phosphate forms of either nucleotide. Addition of GTP or XTP to 10 mM did not increase dimer formation further while 10 mM GDP or XDP resulted in only a marginal increase in dimer formation. We were unable to restore dimer formation to the levels observed prior to nucleotide depletion, suggesting that similar to other small molecular weight GTPases, D181N may be destabilized in the absence of bound nucleotide (28). Taken together, these data reveal that an intact and nucleotide bound SRβ GTP binding domain is both necessary and sufficient for binding to the SRX2 region of SRα.

DISCUSSION
Every step in the translocation of nascent polypeptides into and across the endoplasmic reticulum membrane is highly regulated. Evidence of the extent of this regulation includes the identification of three GTPases, all with putative roles in the targeting of ribosomes to the translocon.
Regulation begins when the signal peptide emerges from the ribosome where it is bound by SRP54, the GTPase in SRP. Transfer of the nascent polypeptide to the translocon occurs after the GTPbound forms of SRP54 and SRα bind to each other. This binding interaction is unusually tight with a half-life for dissociation in vitro in excess of 6 hours (5). It has been proposed that transfer of the nascent polypeptide to the translocon is unidirectional because in the absence of GTP hydrolysis, binding of SRP54 to SRα is essentially permanent. Only after transfer of the nascent polypeptide to the translocon is GTP hydrolyzed by both SRP54 and SRα to release this interaction (5,7,29).
The role of the SRβ GTP binding domain has remained enigmatic, although the recent demonstration that the ribosome may be a GTPase activating protein for SRβ suggests a role in transfer of the ribosome to the Sec61 complex (10). Our results suggest an additional role for nucleotide binding in the regulation of dimerization of SRα and SRβ.
We have shown previously that SR. is anchored to the ER membrane through an interaction between the amino terminal SRX2 domain and SR (11). Our current data demonstrate that the minimal GTPase domain of SR is necessary and sufficient for binding to SR. via SRX2 (Figures GDP (K75I) or destabilize the GTPase fold (N178K) (Sigal et al., 1986;Pai et al., 1990).
However, the most convincing evidence that dimerization is regulated by nucleotide binding comes from the mutant D181N. This mutant is expected to have reduced affinity for GTP but a preference to bind XTP and XDP. In the absence of nucleotide this mutant was no longer able to bind to SRα . However, adding nucleotide back restored binding of SRβ to SRα. Surprisingly, GTP, GDP, XTP and XDP all supported dimerization suggesting that regulation is via the empty state.
There is precedent for regulation of low molecular weight GTPases by guanine nucleotide exchange factors (30). However, a major difference between ras type and SRP type GTPases is that the empty state is transient for ras family GTPases while SRP GTPases are stable in the empty state (5,31,32).
The similarity of the SRβ GTP binding domain to ras family GTPases suggests that nucleotide exchange by SRβ will be regulated by an as yet unidentified guanine nucleotide exchange factor.
At present there are no candidates for such an exchange factor associated with the translocon.
Release of nucleotide from SRβ is predicted to dissociate the heterodimeric SR on the endoplasmic reticulum membrane. Is there any precedent for regulated binding of membranes by SR? In E. coli, the SR is composed of a single protein, FtsY, that is found in both the cytoplasm and on the plasma membrane (33). Membrane binding by FtsY is due in part to a direct interaction with lipids in the E. coli membrane (34). However, there is also evidence that FtsY membrane binding is regulated. First, since lipids are in a huge molar excess over FtsY, non-regulated binding to lipids is not compatible with the observed distribution of FtsY between membrane bound and soluble pools. Second, a large fraction of FtsY proteins are specifically cleaved after binding to the plasma membrane thereby limiting the total amount of FtsY bound to membrane (35). Finally, evidence has been presented for a membrane bound FtsY receptor (34).    Proteolysis of selected SRβ mutants.
SRβ mutants were translated in the absence (lanes 1, 5,9,13) or presence (lanes 2-4, 6-8, 10-12, 14-16) of canine microsomes (RM). Membranes were isolated from the translation reactions by gel filtration chromatography and divided into three aliquots. Digestion with proteinase K (PK) was used to examine topology with respect to the membrane and to assay protein folding. Triton X-100, (TX100) was added to solubilize the membrane to permit digestion of protease sensitive fragments otherwise protected from the protease by the membrane. Triangles indicate type I transmembrane molecules including a protease resistant core (dots) and the transmembrane amino-terminus protected by the membrane. Asterisks indicate the transmembrane aminoterminus protected by the membrane. One microliter of total translation products were analyzed in lanes 1,5,9,13. All other lanes correspond to 7 microliters of the original translation reaction.
The migration positions of molecular weight standards are indicated to the left of the panel.