BTS1 encodes a geranylgeranyl diphosphate synthase in Saccharomyces cerevisiae.

Protein prenylation utilizes different types of isoprenoids groups, namely farnesyl and geranylgeranyl, to modify proteins. These lipophilic moieties attach to carboxyl-terminal cysteine residues to promote the association of soluble proteins to membranes. Most prenylated proteins are geranylgeranylated. Geranylgeranylation is catalyzed by two different prenyltransferases, the type I and type II geranylgeranyl transferases, both of which utilize geranylgeranyl diphosphate as a lipid donor. In the yeast Saccharomyces cerevisiae, the BET2 gene encodes the β-subunit of the type II geranylgeranyl transferase. Mutations in this gene cause a defect in the geranylgeranylation of small GTP-binding proteins that mediate vesicular traffic. In an attempt to analyze those genes whose products may interact with Bet2, we isolated a suppressor of the bet2-1 mutant. This suppressor gene, called BTS1, encodes the yeast geranylgeranyl diphosphate synthase. BTS1 is not essential for the vegetative growth of cells; however, disrupting it impedes the geranylgeranylation of many cellular proteins and renders cells cold sensitive for growth. Our findings imply that BTS1 suppresses the bet2-1 mutant by increasing the intracellular pool of geranylgeranyl diphosphate.

geranylgeranyl (20 carbons), are post-translationally attached to proteins (Epstein et al., 1991). Farnesyl is added to proteins that terminate in a CAAX motif (where C is cysteine, A is an aliphatic amino acid, and X can be methionine, cysteine, alanine, glutamine, phenylalanine, or serine), while geranylgeranyl is transferred onto proteins that end in CAAL (where L is leucine), CC, or CXC motifs (X is any amino acid) (Reiss et al., 1990Seabra et al., 1991Seabra et al., , 1992. Most known prenylated proteins are geranylgeranylated (Epstein et al., 1991).
Farnesyl and geranylgeranyl groups are attached to proteins from all-trans farnesyl diphosphate (FPP) 1 and all-trans geranylgeranyl diphosphate (GGPP), respectively (Casey, 1992). These lipid precursors are intermediates in the isoprenoid biosynthetic pathway . This pathway consists of a series of reactions by which mevalonate is converted into a diverse family of lipophilic molecules that contain a repetitive five-carbon structure. The isoprenoids are subsequently incorporated into a large number of end products, which includes: sterols, ubiquinones, dolichols, tRNAs, and prenylated proteins .
FPP is the product of the farnesyl diphosphate synthase. This enzyme, which is the most abundant and widely occurring prenyltransferase, catalyzes the formation of FPP by the sequential addition of isopentenyl diphosphate (IPP) to dimethylallyl diphosphate (DMAPP), and geranyl diphosphate (GPP) (Anderson et al., 1989;Bartlett et al., 1985;Sheares et al., 1989). In some organisms, GGPP is synthesized by a GGPP synthase that catalyzes stepwise additions of IPP to DMAPP, GPP, and FPP. This type of GGPP synthase activity has been detected in mammalian tissue. However, eukaryotic geranylgeranyl diphosphate synthases are known that synthesize GGPP by the addition of a single molecule of IPP to FPP (McCaskill and Croteau, 1993;Sagami et al., 1992Sagami et al., , 1993Sagami et al., , 1994. But, due to its low activity and the problems in separating this enzyme from FPP synthase, its purification has proven to be difficult (Runquist et al., 1992;Sagami et al., 1985Sagami et al., , 1993Sagami et al., , 1994. GGPP is the substrate for two different protein prenyltransferases, the type I (GGTase-I) and type II (GGTase-II) geranylgeranyl transferases (Jiang and Ferro-Novick, 1994;Seabra et al., 1991Seabra et al., , 1992. GGTase-I catalyzes the transfer of a geranylgeranyl group from GGPP onto proteins that terminate in a CAAL motif, while GGTase-II attaches geranylgeranyl to terminal CC or CXC residues. Its protein substrates include members of the Rab family of small GTP-binding proteins (Jiang and Ferro-Novick, 1994;Seabra et al., 1992). In the yeast Saccharomyces cerevisiae, the GGTase-II is composed of three subunits, which are encoded by BET2, BET4 (formerly called MAD2), and MRS6, respectively (Rossi et al., 1991;Jiang and Ferro-Novick, 1994;Jiang et al., 1993;Li et al., 1993). The BET2 gene product binds to the product of the BET4 gene to form the catalytic component of this enzyme. MRS6 encodes the accessory protein that binds the protein substrate. Mutations in these genes abolish the geranylgeranylation of Ypt1p and Sec4p, two small GTP-binding proteins that mediate intracellular membrane traffic (Jiang and Ferro-Novick, 1994;Li et al., 1993;Rossi et al., 1991). The bet2-1 gene is a recessive temperature-sensitive mutant allele that fails to grow at 37°C (Newman and Ferro-Novick, 1987). In this mutant, a failure to geranylgeranylate Ypt1p and Sec4p leads to a defect in the membrane association of these proteins. This deficiency results in a block in intracellular membrane trafficking (Rossi et al., 1991). In an attempt to identify new genes that may interact genetically with BET2, BET4, or MRS6, we isolated a suppressor of the bet2-1 mutant. This suppressor gene, named BTS1, suppresses the growth defect of bet2-1 when expressed on a low (CEN) or high (2 m) copy vector. Sequence analysis revealed a significant homology between BTS1 and the geranylgeranyl diphosphate synthase from Neurospora crassa, suggesting that BTS1 encodes the homologue of this gene in S. cerevisiae. In accordance with this proposal, the BTS1 gene product was found to be required for the membrane attachment of Ypt1p and Sec4p, a process that is known to require geranylgeranylation. When BTS1 was expressed in bacterial cells, it generated an activity that was able to convert FPP to GGPP, thereby conclusively demonstrating that the BTS1 gene product is the yeast geranylgeranyl diphosphate synthase. This enzyme is a previously unidentified component of the yeast isoprenoid biosynthetic pathway.
Isolation of BTS1-The yeast genomic DNA library used in this study was prepared by ligation of genomic DNA that was prepared from ANY119. This DNA was partially digested with Sau3A and inserted into the BamHI site of pRS316 (CEN, URA3). The library was used to transform the bet2-1 mutant (ANY119), and the transformants were (1 ϫ 10 5 ) selected on minimal medium lacking uracil. After a 3-day incubation at 25°C, the cells were stamped onto YPD plates and incubated overnight at 37°C. 11 positive transformants were obtained. Plasmids (pS1-pS11) retrieved from these transformants were amplified in Escherichia coli and retested in ANY119.
DNA Sequence and Analysis-The DNA sequence of the BTS1 gene was determined by the dideoxynucleotide chain termination method (Sanger et al., 1977). The reactions were performed using the Sequenase (U. S. Biochemical Corp.) protocol, and the data were analyzed with GCG software. Homology searches were performed with the EMBO/GenBank and Swiss-prot data bases.
Disruption of the BTS1 Gene-To disrupt BTS1, a 1.7-kb DraI-NruI fragment containing the BTS1 gene was excised from pS8 and cloned into the PvuII site of pUC118 to generate pSJ30. The plasmid-borne disruption of BTS1 was constructed by replacing a 0.65-kb SacI-EcoRI fragment in pSJ30 with a 1.2-kb SacI-EcoRI fragment containing the URA3 gene. The resulting plasmid (pSJ31) was digested with SspI and BglII and transformed into NY648. The transformants were sporulated, and tetrad analysis was performed. After 3 days at 25°C, all 48 tetrads examined displayed two large and two small colonies. The large colonies were Ura Ϫ , and the small ones were Ura ϩ .
DNA Hybridization Analysis-Yeast genomic DNA prepared from NY180 or SFNY368 was examined by DNA-DNA hybridization as described before (Naumvoski and Friedberg, 1984). Genomic DNA digested with BglII was fractionated on a 0.8% agarose gel and transferred to a BioTrans membrane (ICN). The blot was probed with a radiolabeled 0.65-kb SacI-EcoRI fragment, containing BST1, prepared by random-primer labeling and visualized by autoradiography.
Cell Fractionation Studies-Wild type (NY180) and the ⌬BTS1 strain (SFNY368) were grown overnight at 30°C in YPD medium to early exponential phase. 1 aliquot of cells (150 A 599 units) was pelleted and washed once with ice-cold 10 mM sodium azide. The remaining cells were shifted to 14°C, and the incubation was continued for 12 h before the cells were harvested. To generate spheroplasts, cells were resuspended in 0.7 ml of 10 mM ice-cold sodium azide and mixed with an equal volume of 2 ϫ spheroplast medium (2.8 M sorbitol, 100 mM Tris-HCl (pH 7.5), 20 mM sodium azide) containing 100 units of zymolyase. After a 1-h incubation at 25°C, the spheroplasts were harvested by centrifugation in a clinical centrifuge during a spin at 1400 rpm for 5 min, washed, and lysed in 1.4 ml of ice-cold lysis buffer (0.8 M sorbitol, 10 mM triethanolamine (pH 7.2), 1 mM EDTA) as described before (Rossi et al., 1991). Cell debris was removed during a 3-min spin at 450 ϫ g, and the supernatant from this spin was centrifuged at 100,000 ϫ g for 1 h to generate a soluble fraction. The pellet was resuspended in a volume of lysis buffer equal to the supernatant. Samples were electrophoresed and subjected to Western blot analysis using anti-Ypt1p or anti-Sec4p antibodies (1:2000 dilution).
Expression of the BTS1 Gene in E. coli-To express BTS1 in E. coli, the BTS1 open reading frame sequence was generated by polymerase chain reaction using two primers that overlapped the initiation codon or the region 100 base pairs downstream from the stop codon. EcoRI and ClaI sites were also incorporated into the 5Ј-and 3Ј-ends, respectively. The polymerase chain reaction product was digested with EcoRI and ClaI and cloned into the pUC118 expression vector. The resulting gene fusion encodes a Bts1 protein with six additional NH 2 -terminal amino acids from ␤-galactosidase. This construct was then transformed into JM101 bacterial cells and expressed as described before (Jiang and Ferro-Novick, 1994).
Synthase Assay and Product Analysis-The standard assay mixture contained 20 mM BHDA buffer (pH 7.0), 10 mM ␤-mercaptoethanol, 1 mM MgCl 2 , 0.1% (w/v) bovine serum albumin, 200 M DMAPP or FPP, 20 M [1-14 C]IPP (10 Ci/mol purchased from Amersham), and 70 -80 g of protein in a total volume of 200 l. DMAPP, FPP, and GGPP were synthesized by the method of Davisson et al. (1986). After 10 min at 37°C, 200 l of CH 3 OH-HCl (4:1) was added, and the incubation was continued for 30 min. The reaction mixture was extracted with 1 ml of ligroin, and 0.5 ml of the ligroin layer was mixed with 10 ml of Cytoscint-ES (ICN) for the measurement of radioactivity in a Packard Tri-Carb 4530 liquid scintillation spectrometer. Products were analyzed using HPLC. For the product analysis, bovine serum albumin was omitted from the standard assay mixture, but 10 mM sodium fluoride was present to suppress phosphatase activity. After a 1-h incubation at 37°C, the reaction was terminated by the addition of EDTA (12.5 mM, final concentration). Unlabeled GGPP (25 g) was added, and 150 l of the mixture was injected onto a Shodex Asahipak ODP-50 column (4.6 mm (inner diameter) ϫ 250 mm). 2-min fractions were collected, and the radioactivity in each fraction was determined by liquid scintillation counter after the addition of 15 ml of Cytoscint-ES.
Preparation of Yeast Extracts and Protein Prenylation Assay-Yeast cells were grown in YPD medium at 25°C to late log phase. The cells were harvested, lysed with glass beads, and centrifuged at 100,000 ϫ g for 45 min. The soluble fraction was collected and assayed for GGTase-II activity. Prenylation assays were performed in a 50-l reaction that contained 50 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 , 5 mM dithiothreitol, 25 g of extract, 0.4 M of recombinant Ypt1p, and varying concentrations of [ 3 H]GGPP (American Research Lab, 17,500 dpm/ pmol). The reaction mixture was incubated at 30°C for 30 min before it was terminated with 1 M HCl in ethanol (1 ml) and filtered on a Whatman GF/A filter as described before (Jiang et al., 1993).

RESULTS
Isolation of Suppressors of the bet2-1 Mutant-bet2-1 is a temperature-sensitive mutant that grows at 25°C (permissive temperature) but dies at 37°C. To isolate genes whose products may interact with the Bet2 protein (Bet2p), we screened a yeast genomic library that was prepared from the bet2-1 mutant for plasmids that conferred growth at 37°C. After screening 1 ϫ 10 5 transformants, 11 positive colonies were obtained and retested. The growth of mutant cells containing six of these plasmids (group A) was indistinguishable from that of wild type at 37°C (data not shown). The other five (group B), however, did not suppress as well. Restriction analysis indicated that the plasmids in group A contained the BET2 structural gene. Since the genomic library was prepared from bet2-1 mutant cells, the restoration of growth observed at 37°C is not true complementation. Plasmids in group B contained an overlapping 2.0-kb region of DNA. Therefore, the gene that suppresses the bet2-1 mutant is located within this 2.0-kb fragment.
The smallest group B plasmid (pS8) that we isolated contained a 2.8-kb insert (Fig. 1b). To analyze the ability of this insert to suppress bet2-1, we cloned this fragment into a high copy URA3 vector (pRS426) to generate pSJ28. When pSJ28 was transformed into bet2-1 mutant cells, suppression was significantly enhanced (Fig. 1, compare b and c to the mutant alone in a). In fact, growth of the mutant was restored to that of wild type (Fig. 1, compare c and d), suggesting that suppression was gene dosage dependent.
Plasmid pSJ28 Increases the Membrane-bound Pool of Ypt1p and Sec4p in bet2-1 Mutant Cells-Previous studies have shown that the membrane association of Ypt1p and Sec4p is defective in bet2 mutant cells (Rossi et al., 1991). This defect is a consequence of the failure to geranylgeranylate these proteins (Jiang et al., 1993;Rossi et al., 1991). Thus, the lethal phenotype of the bet2-1 mutant is likely to be a consequence of the inability of these proteins to attach to membranes. Since plasmid pSJ28 suppresses the growth defect of the bet2-1 mutant at 37°C, it may also cure the membrane attachment defect observed in these cells. To address this possibility, we transformed pSJ28 into bet2-1. When the distribution of Ypt1p and Sec4p was examined in these transformants and compared to the mutant and wild type, pSJ28 was found to enhance the membrane association of these small GTP-binding proteins (Fig. 2, compare the amount in the lysate (T) to the supernatant (S) and pellet (P) fractions). The presence of pSJ28 did not lead to an increase in the residual GGTase-II activity that can be measured in bet2 mutant cells (Jiang et al., 1993). Thus, the restoration of the membrane association of Ypt1p and Sec4p is not a consequence of increasing GGTase-II activity.
Cloning and Sequencing the Suppressor Gene-To localize the suppressor within the 2.8-kb genomic fragment described above, subclones of pSJ8 were constructed and inserted into pRS316 (URA3, CEN). Suppression studies revealed that the SacI site contained within this fragment is critical for its activity. The smallest region of DNA capable of suppressing bet2-1 was found to be a 1.6-kb SspI-NruII fragment. This region of DNA was sequenced in both directions using the strategy shown in Fig. 3. An open reading frame of 1005 base pairs that spans the SacI site was identified. We called the gene that encodes this open reading frame BTS1 (Bet Two Suppressor). BTS1 is predicted to encode a protein of 335 amino acids with a calculated molecular mass of 38,627 daltons (Fig. 4). Overall, the amino acid composition of the Bts1p is hydrophilic, and no significant hydrophobic stretches were observed.
Bts1p Is Homologous to Known Prenyltransferases-Comparison of the predicted Bts1p amino acid sequence with the Swiss-Prot protein sequence data base revealed a significant similarity between Bts1p and the N. crassa albino-3 gene product (Carattoli et al., 1991). These proteins are 40% identical at the amino acid level with the most conserved region localized to the middle of these proteins (Fig. 5). The albino-3 gene encodes a geranylgeranyl diphosphate synthase in the carotenoid biosynthetic pathway of N. crassa (Nelson et al., 1989). Bts1p also contains five conserved regions found in other FPP and GGPP synthases (Chen and Poulter, 1994). These comparisons suggest that BTS1 encodes GGPP synthase, an unidentified prenyltransferase of S. cerevisiae.
Disruption of the BTS1 Gene-To investigate if BTS1 is required for the vegetative growth of yeast cells, we disrupted one copy of this gene in diploid cells and then performed tetrad analysis. This disruption was constructed by exchanging a 0.65-kb segment of the coding region with the URA3 gene. The disrupted gene was then transformed into NY648 to replace one of the chromosomal copies. The Ura ϩ transformants were selected, sporulated, and analyzed. In all of the 48 tetrads examined, four viable spores were obtained. However, two of the colonies in each of the tetrads displayed a growth defect at 25°C. The small colonies were Ura ϩ , indicating that they contained the disrupted BTS1 gene. This was confirmed by Southern blot analysis using BTS1 as a probe (see "Materials and Methods"). Thus, BTS1 is not essential for the vegetative growth of yeast cells. But in its absence, growth is impaired. The growth of the disrupted strain (SFNY368 or ⌬BTS1) was examined further at different temperatures. As shown in Fig.  6, ⌬BTS1 cells (Fig. 6, a and d) grew as well as wild type at 30°C (Fig. 6, b and c). However, at lower temperatures (25 and 14°C) a growth defect emerged. Only small colonies appeared Wild type (W.T.) (SFNY26 -6A) and bet2-1 mutant cells (ANY119) that contained pSJ28 were grown to early exponential phase in minimal medium that was supplemented with the appropriate amino acids and 2% gluocse. Cells were harvested, converted to spheroplasts, lysed, and the membrane (P) and soluble (S) fractions were recovered by centrifuging the lysates (T) at 100,000 ϫ g. Samples were then subjected to Western blot analysis using anti-Ypt1p and anti-Sec4p antibodies.
after 3 days at 25°C (Fig. 6, a and d), while at 14°C, the cells did not survive (Fig. 6, a and d). This result clearly demonstrated that SFNY368 is cold sensitive for growth.
The BTS1 Gene Product Is Required for the Membrane At-tachment of Ypt1p and Sec4p-Ypt1p and Sec4p are two small GTP-binding proteins that regulate intracellular membrane traffic . Like many small GTPbinding proteins, they are synthesized in the cytosol but be- come membrane-bound to perform their function (Rossi et al., 1991). The ability of Ypt1p and Sec4p to bind to membranes is conferred by the addition of the 20-carbon, geranylgeranyl moiety (Jiang et al., 1993). The geranylgeranylation of these proteins is catalyzed by a protein prenyltransferase that utilizes GGPP as a lipid donor. If BTS1 encodes GGPP synthase, disruption of this gene should result in the depletion of GGPP. Consequently, the geranylgeranylation of Ypt1p and Sec4p will be abolished. To test this hypothesis, we examined the membrane association of these proteins in the ⌬BTS1 strain. SFNY368 was grown at 30°C for 12 h until the A 600 was 1.0 prior to shifting the cells to 14°C for another 12 h. Aliquots of cells were removed at each time point, converted to spheroplasts, lysed, and centrifuged at 450 ϫ g to remove unbroken cells and nuclei. Subsequently, these lysates were centrifuged at 100,000 ϫ g for 1 h to obtain supernatant and pellet fractions, and the distribution of Ypt1p and Sec4p was examined in each of these fractions by Western blot analysis. In wild type cells (Fig. 7, compare the amount in the lysate (T) to the supernatant (S) and pellet (P)), most of Ypt1p and Sec4p was membrane-bound at both time points, and the change in temperature did not affect their membrane association (Fig. 7, compare 14 and 30°C). However, in SFNY368, most of the Ypt1p and Sec4p was soluble at both temperatures (Fig. 7, compare the amount in the lysate (T) to the supernatant (S) and pellet (P)), although this defect was more pronounced at 14°C. Thus, the membrane association of these small GTPbinding proteins is defective in ⌬BTS1 cells.
Prenyltransferase Activity of Crude Extracts-To test the hypothesis that BTS1 encodes a geranylgeranyl diphosphate synthase, we cloned the gene into a pUC118 vector to express it in E. coli. Crude extracts of E. coli containing pUC118 (control) or pUC118/BTS1 were assayed for prenyltransferase activity in the presence of [1-14 C]IPP, using DMAPP or FPP as the allylic substrate, and the reaction mixture was analyzed by HPLC. The prenyltransferase activity observed was dependent FIG. 5

. A comparison of the BTS1 and albino-3 (Al-3) gene products.
The amino acid sequence of Bts1p is compared to Al-3 using the Bestfit program (GCG software package). Identity is indicated by a line, and conserved changes are marked by two dots (two corresponding bases in a codon) or one dot (one corresponding base in a codon) between the sequences. The gaps are designated by dots within a sequence. Bts1p and Al-3 share 40% identity at the amino acid level.
FIG. 6. SFNY368 (⌬BTS1) is cold sensitive for growth. Diploid cells with one copy of BTS1 disrupted were sporulated and subjected to tetrad analysis. In each tetrad, two wild type spores (b and c) and two spores containing the disrupted BTS1 gene (a and d) were germinated, purified, and grown at various temperatures (14, 25, and 30°C). The 25 and 30°C plates were incubated for 3 days, while the 14°C plate was incubated for 7 days.
upon the presence of FPP, since no counts were obtained when the pUC118/BTS1 extract was assayed in the absence of FPP (not shown). The radioactive product of this incubation coeluted with unlabeled synthetic GGPP, indicating that it is GGPP (Fig. 8). No conversion of FPP to GGPP was seen with the pUC118 control. Both extracts also showed low levels of activity in the conversion of DMAPP to an acid-labile product. However, because the extent of conversion was the same for both samples, this activity could not be due to Bts1p (not shown). These findings confirm that BTS1 encodes a geranylgeranyl diphosphate synthase.
bet2-1 Mutant Extracts Have a Lower Affinity for GGPP-We next investigated the mechanism by which the overexpression of BTS1 suppresses the lethality of the bet2-1 mutant. One possibility is that BTS1 suppresses by increasing the intracellular pool of GGPP, thereby compensating for a mutant GGTase-II that has a lower affinity for GGPP. To test this hypothesis, we measured the GGTase-II activity of wild type and bet2-1 mutant extracts in the presence of varying concentrations of GGPP. As a control, we also assessed the activity of bet4 -2 mutant extracts. BET4 encodes the ␣-subunit of the GGTase-II (Li et al., 1993;Jiang et al., 1993), and extracts prepared from this mutant are devoid of GGTase-II activity. Unlike bet2-1, the overexpression of BTS1 does not suppress the temperature-sensitive growth defect of the bet4 -2 mutant (not shown). As shown in Fig. 9, the GGTase-II activity of the wild type extract was saturated at ϳ0.8 M of GGPP. At this concentration, the activity of the bet2-1 mutant extract was approximately 5-10% of wild type. This activity was significantly enhanced when the GGPP concentration was increased beyond 2 uM, and saturation was achieved at 6 M. In contrast, the GGTase-II activity of the bet4 -2 mutant extract could not be compensated for by increasing the concentration of GGPP. The calculated K m values of GGTase-II in the bet2-1 mutant and wild type were ϳ3.6 and 0.4 M, respectively. Therefore, it appears that GGTase-II in the bet2-1 mutant has a reduced affinity for GGPP, which results in a decrease in prenylation activity. By increasing the amount of GGPP that is added to the assay, prenylation activity is efficiently restored. This result provides a clear explanation for the suppression of bet2-1 by BTS1. DISCUSSION Previously, we have shown that the yeast GGTase-II is composed of three subunits (BET2, BET4, and MRS6). Bet2p, the ␤-subunit of this enzyme complex, forms a complex with Bet4p, the ␣-subunit (Jiang et al., 1993). Mrs6p is an escort protein that presents protein substrate to the Bet2p-Bet4p complex (Jiang et al., 1994). During geranylgeranylation, the Bet2p-Bet4p complex binds to and transfers GGPP to Ypt1p, Sec4p, and other small GTP-binding proteins. In an attempt to identify new genes whose products may interact with Bet2p, we isolated a suppressor of the bet2-1 mutant. Our data demonstrates that this suppressor gene, called BTS1, encodes a geranylgeranyl diphosphate synthase, an unidentified prenyltransferase of the yeast isoprenoid biosynthetic pathway. The BTS1 gene product functions on this pathway to convert FPP to GGPP.
The function of BTS1 was revealed by analyzing the sequence of this gene. The predicted amino acid sequence of Bts1p was found to be significantly homologous to the albino-3 gene product, the N. crassa GGPP synthase (Carattoli et al., 1991). Upon a closer examination, we also found that Bts1p contains five highly conserved motifs that are present in all known FPP and GGPP synthases (Chen and Poulter, 1994), including the aspartate-rich sequences proposed to be involved FIG. 7. The membrane attachment of Ypt1p and Sec4p is defective in ⌬BTS1 cells. Cells were grown at 30°C to exponential phase. 1 aliquot of cells was removed, pelleted, converted to spheroplasts, and lysed. The remaining cells were shifted to 14°C and grown for 12 h before they were harvested and lysed. Lysates (T) were centrifuged at 100,000 ϫ g to obtain pellet (P) and supernatant (S) fractions. Samples were electrophoresed on a 12.5% SDS-polyacrylamide gel and subjected to Western blot analysis using anti-Ypt1p (A) or anti-Sec4p (B) antibodies. W.T., wild type. in binding and catalysis (Ashby and Edwards, 1990;Joly and Edwards, 1993;Song and Poulter, 1994). This finding suggested that BTS1 encodes the yeast GGPP synthase. To confirm this hypothesis, we expressed the BTS1 gene in bacteria. Bacterial lysates that express Bts1p were found to contain an activity that synthesizes GGPP from IPP and FPP.
The suppression of the bet2-1 mutant by BTS1 could be explained in several ways. The BTS1 gene product may itself have GGTase-II activity, or it could directly interact with GGTase-II to stimulate its activity. In either situation, the overexpression of BTS1 would be expected to increase GGTase-II activity. However, this was not observed. Alternatively, suppression may simply be a consequence of increasing the intracellular pool of GGPP. Since in vitro prenylation studies have demonstrated that mutant GGTase-II has a low affinity (increased K m ) for GGPP, which is compensated for by higher concentrations of GGPP, this alternate possibility is most likely. According to this model, additional copies of BTS1 should result in higher intracellular concentrations of GGPP and enhanced suppression of bet2-1, thus explaining why the suppression of bet2-1 by BTS1 is gene dosage dependent.
Because each of the subunits of the GGTase-II are essential, we anticipated that BTS1 would also be required for the vegetative growth of yeast cells. To our surprise, the ⌬BTS1 strain was only cold sensitive for growth. Furthermore, the growth of this strain was not impaired at 30°C or higher temperatures. When the membrane association of Ypt1p and Sec4p was examined in ⌬BTS1 cells grown at 30°C, a small fraction of each of these proteins was membrane bound. Thus, BTS1-depleted cells are able to prenylate proteins at a level that is sufficient to sustain cell growth at higher temperatures. When these cells were shifted to 14°C, less membrane-bound Ypt1p and Sec4p was detected, implying that growth ceases as a consequence of the failure to prenylate these essential proteins.
Since BTS1 is not essential for the growth of yeast cells, the synthase gene may be duplicated. Preliminary DNA hybridization experiments, however, argue against this possibility. Another explanation for the dispensability of BTS1 is that GGTase-II might utilize FPP as an alternate substrate. However, since GGTase-II cannot transfer FPP to Ypt1p, this possibility seems unlikely (Jiang et al., 1993). Furthermore, extracts prepared from ⌬BTS1 cells do not support the transfer of [ 3 H]FPP onto Ypt1p. Thus, it is more likely that another prenyltransferase, such as hexaprenyl diphosphate synthase, might pro-duce small amounts of GGPP as an intermediate product during the elongation of FPP to longer polyisoprenoid chains. In the ⌬BTS1 strain, GGPP may be formed in this way, enabling yeast cells to survive at certain temperatures in the absence of the geranylgeranyl synthase.