Purification and structural characterization of transcriptional regulator Leu3 of yeast.

The transcriptional regulatory protein Leu3 of Saccharomyces cerevisiae was enriched approximately 70-fold above wild type level in yeast cells carrying a pGAL1-LEU3 expression vector. Sustained overproduction of Leu3 following induction by galactose required elevated intracellular levels of alpha-isopropylmalate, a leucine pathway intermediate known to act as transcriptional co-activator. Starting with galactose-induced cells, the Leu3 protein was purified about 3,500-fold (i.e. 245,000-fold over wild type level) by a procedure that included treatment of the cell-free extract with polyethylenimine, fractionation with ammonium sulfate, heat treatment, and DNA affinity chromatography. Highly purified preparations still showed two protein bands when subjected to polyacrylamide electrophoresis under denaturing conditions. Their apparent molecular masses were about 104,000 and 110,000 kDa. The smaller of these values was very close to the maximum molecular weight obtained previously for Leu3 protein translated in vitro in a rabbit reticulocyte lysate. (The molecular weight deduced from the open reading frame of the LEU3 gene is 100,162.) Both protein bands reacted with antibodies raised against different portions of the Leu3 molecule and were, therefore, likely to represent two forms of Leu3. Treatment with calf intestinal phosphatase quantitatively converted the slower moving band into the faster moving one. Conversion was prevented by inorganic phosphate, a phosphatase inhibitor. These experiments showed that the two bands very likely correspond to phosphorylated and nonphosphorylated forms of Leu3. Phosphorylation did not appear to affect the DNA binding function of Leu3, but (indirect) effects on the activation function or effects on the modulation by alpha-isopropylmalate have not been ruled out. Electrophoretic mobility shift assays were used to estimate the apparent dissociation constants of the two specific Leu3-DNA complexes routinely seen in these assays. The values obtained were 1.1 and 2.6 nM. Finally, using size exclusion chromatography, native Leu3 protein was shown to have dimeric structure, irrespective of the state of phosphorylation.

Purification and Structural Characterization of Transcriptional Regulator Leu3 of Yeast* (Received for publication, May 1, 1992) Ji-Ying Sze and Gunter B. Kohlhaw$ From the Department of Biochemistry, Purdue University, West Lafayette, Indiana 47907-1153 The transcriptional regulatory protein Leu3 of Saccharomyces cereviaiae was enriched approximately 70-fold above wild type level in yeast cells carrying a pGAL1-LEU3 expression vector. Sustained overproduction of Leu3 following induction by galactose required elevated intracellular levels of a-isopropylmalate, a leucine pathway intermediate known to act as transcriptional co-activator. Starting with galactoseinduced cells, the Leu3 protein was purified about 3,500-fold (i.e. 245,000-fold over wild type level) by a procedure that included treatment of the cell-free extract with polyethylenimine, fractionation with ammonium sulfate, heat treatment, and DNA affinity chromatography. Highly purified preparations still showed two protein bands when subjected to polyacrylamide electrophoresis under denaturing conditions. Their apparent molecular masses were about 104,000 and 110,000 kDa. The smaller of these values was very close to the maximum molecular weight obtained previously for Leu3 protein translated in vitro in a rabbit reticulocyte lysate. (The molecular weight deduced from the open reading frame of the LEU3 gene is 100,162.) Both protein bands reacted with antibodies raised against different portions of the Leu3 molecule and were, therefore, likely to represent two forms of Leu3. Treatment with calf intestinal phosphatase quantitatively converted the slower moving band into the faster moving one. Conversion was prevented by inorganic phosphate, a phosphatase inhibitor. These experiments showed that the two bands very likely correspond to phosphorylated and nonphosphorylated forms of Leu3. Phosphorylation did not appear to affect the DNA binding function of Leu3, but (indirect) effects on the activation function or effects on the modulation by a-isopropylmalate have not been ruled out.
Electrophoretic mobility shift assays were used to estimate the apparent dissociation constants of the two specific Leu3-DNA complexes routinely seen in these assays. The values obtained were 1.1 and 2.6 nM. Finally, using size exclusion chromatography, native Leu3 protein was shown to have dimeric structure, irrespective of the state of phosphorylation.
The Leu3 protein of yeast is a transcriptional regulator * This work was supported in part by National Institutes of Health Research Grant GM15102 and by a Purdue Research Foundation grant (to J. S. ). This is Journal Paper 13416 of the Agricultural Experiment Station, Purdue University. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisernent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$To whom correspondence should be addressed. Tel.: 317-494-1616. whose principal effect is on branched chain amino acid biosynthesis (1)(2)(3)(4). It interacts with a specific 12-base pair DNA sequence (UASL)' located upstream of the regulated genes (5-9). The Leu3 protein has several outstanding features. First, it is unique among known eukaryotic regulators in that its function as activator depends entirely on a low molecular weight metabolite (10). The metabolite, a-isopropylmalate (a-IPM), is an early intermediate in leucine biosynthesis whose intracellular concentration is influenced by the branched chain amino acid pools, the general amino acid control system, and other factors (1,3,4). Second, in the absence of a-IPM, Leu3 actually functions as a repressor of transcription (3,lO). Third, a recently developed, yeast-derived in vitro transcription system (11) faithfully executes Leu3-dependent activation, repression, and modulation by a-IPM (lo), opening the door for a more definitive study of the mechanisms by which Leu3 operates.
Manipulation of the LEU3 gene has defined several functional regions within the 886 amino acid residues of the Leu3 protein (12-15): a DNA binding region (residues 17-110) that contains a Zn(II)2 Cyss binuclear cluster typical of an entire family of lower eukaryotic transcriptional regulators whose members include the gene products of G A U , PPRI, HAPI, PDRl, PUT3, Lac9, and qa-1F (12,13); an acidic-hydrophobic transcriptional activation region (residues 855-886) (14); and a "modulation region," which encompasses a large central portion of the Leu3 protein (12,14,15). Removal of the central portion creates a strong constitutive activator that no longer responds to a-IPM. Evidence has been obtained showing that certain individual residues of the activation region (e.g. Trp-864, Trp-861) are also involved in modulation, and a model has been proposed, which postulates that interaction of the Leu3 protein with a-IPM results in the disruption of contacts between the activation region and the central portion of the protein (14).
In order to facilitate further structural and functional analyses such as complete in uitro reconstitution of the Leu3dependent transcription process and interaction of Leu3 protein with a-IPM, we decided to purify Leu3. Here we report a method to rapidly isolate stable Leu3 protein from overproducing yeast cells. Interestingly, highly purified preparations of Leu3 display two protein species upon polyacrylamide gel electrophoresis under denaturing conditions. The apparent molecular weight of the two species, their immunoreactivity, and their response to phosphatase treatment suggest that they represent phosphorylated and nonphosphorylated forms of Leu3. This observation adds yet another layer of complexity to the structure and possibly the function of the Leu3 protein.
The abbreviations used are: UAS, upstream activating sequence; IPM, isopropylmalate. 1 Material to be analyzed was in each case obtained by preparing cellfree extracts, collecting the protein that precipitated upon saturation with (NH&SO, to a final concentration of 46% (4 "C), and redissolving the precipitated protein in buffer A (see "Experimental Procedures"). Lanes I, uninduced; lanes 2, induced for 6 h; lanes 3, induced for 8 h; lanes 4, induced for 1'7 h. Designations a, b, and f refer to the two shifted bands a and b routinely seen with Leu3 preparations (12)(13)(14) and to free DNA probe, respectively. Material at the origin is seen to varying degrees in most of our band shift assays and probably represents aggregate or insoluble forms of probe-protein complexes (12, 13).
Plasmid pGALI-LEU3 (see Fig. 1) was generated from pKZ5 (12) and pRM272 (17). First, a unique BamHI site was created at the filled in XhoI site of pKZ5 by inserting a RarnHI linker (d[GGGATCCC]). The 3130-base pair RamHI-SrnaI fragment of pKZ5, which contains the entire LEU3 gene, was then ligated to the large BamHI-Sal1 fragment of pBM272, with the Sal1 site filled in. In this way, the LEU3 gene was placed directly behind the GALI promoter.
At an Amnm of about 1.0, the cultures were diluted 12.5-fold with YPD medium (16), and growth was continued. At an Amnrn of about 0.5, induction of the GALI promoter was initiated by adding Dgalactose to a final concentration of 2%. Cells were harvested a t various times after induction. XK16O/pCALl-LEU3 cells grown for the purpose of purifying Leu3 were routinely harvested 14 h after induction at an Amnm of 6-7, which yielded 20-24 g of cells (wet weight)/liter. Cell pellets, which had been washed once with deionized water, were stored a t -70 "C. To prepare cell-free extract, frozen pellets were thawed on ice and suspended in extraction buffer (200 mM Tris-HCI buffer, pH 7.9; 400 mM (NH,)2S04; 5 mM MgCk 50 *M ZnSO,; 1 mM EDTA; 4 mM dithiothreitol; 20% (v/v) glycerol; 2 mM phenylmethylsulfonyl fluoride; 2 mM benzamidine; and 2 pM pepstatin) a t a ratio of 1.5 ml/g of cells (wet weight). Suspensions were passed twice through a French pressure cell a t 138 megapascals. The resulting extract was clarified by centrifugation for $0 min at 41,000 X g (4 "C).

E. coli cells were grown a t 37 "C in L-broth (19) containing ampicillin (100 pg/ml).
Purification of Leu3 Protein-Except for the heat step, all procedures were performed a t 0-4 "C. Cell-free extract was prepared as outlined above from 40-50 g of XK16O/pCALI-LEU3 cells (wet weight). Polyethylenimine was then added dropwise with stirring from a 5% solution to a final concentration of 0.033%. After standing on ice for 10 min, the resulting suspension was centrifuged a t 12,000 X g for 10 min. The supernatant solution was adjusted to an (NH4),S04 concentration of 46% using a saturated (NH,),S04 solution, which also contained 25 mM HEPES.NaOH buffer, pH 7.9, and 1 mM EDTA and taking into account the amount of (NH,),SO, present in the extraction buffer. After overnight incubation, the precipitate was collected by centrifugation at 12,000 X g for 1 h and was redissolved in 25 mM HEPES. NaOH buffer, pH 7.9, containing 5 mM MgCI,, 50 p~ ZnS04, 1 mM EDTA, 20% (v/v) glycerol, 2 mM each of phenylmethylsulfonyl fluoride and benzamidine, and 0.02% Nonidet P-40 ("buffer A"). The solution was divided into 0.5-ml aliquots. To each aliquot, an equal volume of buffer A pre-equilibrated a t 85 "C was added, and the mixture was further incubated a t 85 "C for 2 min before being rapidly cooled on ice. Denatured protein was pelleted by centrifugation a t 31,000 X g for 20 min. The supernatant solution was dialyzed against several changes of "buffer R" (identical to buffer A except that dithiothreitol was replaced with 0.1% 8mercaptoethanol and MgCI, was omitted) containing 80 mM NaC1. To trap proteins that react with DNA nonspecifically, sonicated salmon sperm DNA was added a t 12.5 pg/ml dialysate. The mixture was incubated for 20 min a t 0 "C and then loaded on a UASL-Sepharose column (at a ratio of up to 100 mg of total protein/ml of column volume) that had been equilibrated with "buffer C" (identical to buffer A except that MgCh was omitted) containing 80 mM NaCI. The column material was prepared by first ligating pieces of U A S L .~~ DNA (see below, section on "Electrophoretic Mobility Shift Assays") to form concatemers (ranging from 3 to 23 units), which were then coupled to CNRr-activated Sepharose 4B (Pharmacia-LKB Biotechnology Inc. ). The affinity matrix contained approximately 34 pg of DNA/ml of resin. After allowing the protein to interact with the affinity matrix (1 h a t 4 "C), the column was eluted successively with 10 column volumes of 80,200, and 300 mM NaCl in buffer C, followed by 6 column volumes each of 1 and 2 M NaCl in buffer C. The appearance of Leu3 protein was monitored by electrophoretic mobility shift assays and by immunoblotting. Fractions containing the major-

Purification and Characterization of
Leu3 2507 3 Protein concentration was determined by the method of Bradford (38) except for affinity-purified preparations whose protein concentration was estimated from silver stained gels (39) using bovine serum albumin as a standard.
' Activity units were measured by performing electrophoretic mobility shift assays on polyacrylamide gels (see "Experimental Procedures").
Material from the different purification steps was incubated with the same batch of "'P-labeled DNA probe, and aliquots were applied to the same gel slab. Following electrophoresis, the two Leu3-specific bands were excised from the gel, combined, and subjected to scintillation counting. One unit is defined as 1 cpm detected under the conditions of the experiment.
Stepwise elution with NaCl FIG. 3. UASL-Sepharose affinity chromatography. Ten ml of a heat-treated, dialyzed preparation of Leu3 containing 164 mg of total protein were loaded onto a UASL.2.-Sepharose affinity column (diameter, 17 mm; height, 20 mm). After equilibration and collection of approximately 10 ml of "flow-through," 33 fractions of 3 ml each were collected while increasing the NaCl concentration in a stepwise fashion (see "Experimental Procedures"). The figure shows electrophoretic mobility shift assay results obtained with the material loaded on the column, the flow-through, and fractions representing the stepwise elution. Strong, Leu3-specific band shifting was seen only after the addition of 1 M NaCl, with insignificant amounts eluting at the 300 mM step. The designations a, b, and f refer to the positions of shifted bands a and b and to free DNA probe, respectively.
ity of the Leu3 protein were pooled and stored a t -70 "C. Under these conditions, the Leu3 protein is stable for a t least 6 months.
Transformation Procedures-Yeast cells were transformed as described by Ito et al. (20). E. coli cells were transformed by the calcium chloride procedure essentially as described by Maniatis et al. (21).
Oligonucleotide Synthesis and Purification-Oligonucleotides were synthesized on an Applied Biosystems model 380A DNA synthesizer by the Laboratory for Macromolecular Structure, Purdue University. Purification was done as described (3).
Electrophoretic Mobility Shift Assays-DNA binding was monitored as follows. A given amount of protein was incubated with 25 mM HEPES.NaOH buffer, pH 7.9, containing 80 mM NaC1, 1 mM EDTA, 4 mM dithiothreitol, 5% (v/v) glycerol, 1 pg of poly(d1-dC), and 0.8 ng of DNA probe (unless indicated otherwise) for 15 min at complementary strands were 5' end-labeled using [yY2P]ATP with a specific radioactivity of about 5,000 Ci/mmol. Following incubation, reaction mixtures were loaded onto a 4% polyacrylamide gel (crosslinking ratio, 40:l) containing 80.7 g of Tris base, 36.7 g of boric acid, and 372.2 g of Na2EDTA. 2HpO/liter. Following electrophoresis, gels were dried and subjected to autoradiography. When needed, shifted bands were excised individually, and the radioactivity in each slice was determined by scintillation counting, using a Beckman model LS8100 counter and ICN Ecolume as scintillation fluid.
Antibody Preparation and Immunoblotting-In order to prepare antibodies against the COOH-terminal "activation domain" of Leu3 (14), a peptide corresponding to the 28 COOH-terminal amino acid residues of the Leu3 protein (859-AGWDNWESDMVWRDVDILM-NEFAFNPKV-886) was synthesized by the Laboratory for Macromolecular Structure, Purdue University. It was purified by reversephase high performance liquid chromatography, and its structure was confirmed by mass spectrometry and partial amino acid sequence Purification and Characterization of Leu3 analysis. The purified peptide was then modified with Traut's reagent, 2-iminothiolane HCI, to introduce sulfhydryl groups (22). At pH 9, the pH at which the reaction was performed, the preferred target would be the amino terminus of the peptide. Next, the modified peptide was conjugated to maleimide-activated keyhole limpet hemocyanin, following the protocol of the supplier (Pierce Chemical Co.). After purification by gel filtration, the peptide-hemocyanin conjugate was used directly to immunize rabbits. Antiserum obtained after the second boost was partially purified by the method of Steinbuch and Audran (23) as modified by Dankert et al. (24).
Immunoblotting was performed according to Burnette (25) using electrophoresis under denaturing conditions on 7.5% polyacrylamide slab gels to resolve the proteins. Powdered milk was used to block nonspecific binding by antibodies. Primary antibodies were detected by horseradish peroxidase-conjugated anti-rabbit antibodies from donkey using the enhanced chemiluminescence system of Amersham Corp.

RESULTS
Overproduction of the Leu3 Protein-Overproduction of the Leu3 protein was accomplished by placing the LEU3 gene behind the strong GAL1 promoter of yeast (Fig. 1). The plasmid carrying the GALl-LEU3 construct was expressed in two hosts that differed in their ability to synthesize a-IPM. Both host strains had a regl-501 gall genetic background. The regl-501 mutation eliminates glucose repression of GAL genes, while the gall mutation prevents galactose from being metabolized (17). Cells can, therefore, be grown to very high densities in glucose-containing media, with galactose serving as gratuitous inducer. Host strain 334 had a normal complement of leucine genes and, therefore, produced normal levels of a-IPM. Host strain XK160 was leul and LEU4'b'. A strain with this genetic background secretes at least 15 times as much a-IPM into the surrounding medium as does wild type (3). The leul mutation prevents the metabolism of a-IPM; the LEU4%' mutation causes constitutive overproduction of a-IPM due to loss of feedback inhibition of LEU4-encoded a-IPM synthase (1). The production of the Leu3 protein in strains 334 and XK160, transformed with pGAL1-LEU3, was examined by electrophoretic band shift assays (Fig. 2). In both strains, the level of Leu3 increased for 8 h after induction with galactose. Thereafter, the level of Leu3 protein fell off in strain 334, while the level in strain XK160 continued to increase for at least 9 more h. This result suggests that a-IPM has a stabilizing effect in the later stages of induction. Quantitation of the induction by determining the radioactivity of cut out gel slices corresponding to shifted bands a and b indicated that the Leu3 protein level increased about 70-fold upon induction of strain XKlGOIpGALI-LEU3 for 17 h. (Note that uninduced strain XKlGOIpGALl-LEU3 still expresses its single genomic copy of LEU3 and should, therefore, represent "wild type" with respect to Leu3 protein level.) Purification of the Leu3 Protein from "Enriched" Crude Extract-Scheme I shows the strategy adopted for the purification of Leu3. Since there is only one chromatography step (UASL-affinity chromatography), the procedure can be completed in about 2 days. Table I shows that affinity chromatography was by far the most efficient step with respect to overall purification. The results of a typical UASL-affinity chromatography step are shown in Fig. 3. The affinity column (prepared as described under "Experimental Procedures") was eluted stepwise with increasing concentrations of NaC1. Leu3 protein appeared after raising the NaCl concentration to 1 M. Recovery of Leu3 protein after heat treatment fluctuated from preparation to preparation. I t was nevertheless essential to retain the heat step as part of the purification procedure, since it significantly improved the subsequent stability of the Leu3 protein. The overall purification factor was close to 3,500.
Electrophoretic mobility shift assays showed that material from all stages of purification was capable of forming the two UASL-Leu3 complexes a and b observed before with crude yeast preparations (Fig. 4). The relative mobility of the complexes formed with affinity-purified material was indistinguishable from that of less purified preparations.
Affinity of Leu3 Protein for UASL-Since the mobility shift assays clearly separated complexes a and b from each other and from DNA, attempts were made to determine the apparent dissociation constants (Fig. 5). Under the experimental conditions chosen, Koa,,,, was 2.6 nM for complex a and 1.1 nM for complex b. The range of these values is very close to that reported for estrogen and thyroid hormone receptors (26,27) and for the yeast transcriptional factors Gcn4 and Gal4 (28,29). The finding that the protein component of complex a has a slightly lesser affinity for UASL than the protein component of complex b is consistent with previous estimations of relative association and dissociation rate constants, which suggested thatwomplex a was somewhat less stable than complex b and formed much more slowly (13).

Highly Purified Preparations Contain Two Forms of Leu3
Protein That Can Be Converted into One by Phosphatase Treatment-Polyacrylamide gel electrophoresis under denaturing conditions revealed that affinity-purified Leu3 protein contained two protein species, designated a and fl (Fig. 6A).
The apparent molecular weights of these two proteins were about 104,000 and 110,000. The smaller of these values was within 2% of the maximum molecular weight obtained previously for Leu3 protein translated from full-length LEU3 mRNA in a rabbit reticulocyte lysate (41). In that experiment,

Purification and Characterization of Leu3
FIG. 5. Determination of apparent dissociation constants for the two Leu3-UASL complexes a a n d b.
Values of K,),,pp were obtained by quantitation of electrophoretic mobility shift assays (see "Experimental Procedures" for general method). A constant concentration of affinity-purified Leu3 protein (0.8 ng/30 p i ) was incubated with radiolabeled UAS,. DNA a t concentrations ranging from 0. 16    a strong single band appeared a t a position corresponding to a molelcular weight of 106,000. All of these apparent molecular weights are slightly above the value of 100,162 deduced for Leu3 protein monomer from the DNA sequence (2). When highly purified Leu3 preparations were subjected to immunoblotting with antibodies raised against different portions of Leu3, both species reacted. Fig. 6B shows the results of an experiment with polyclonal antibodies raised against a hemocyanin-linked peptide consisting of the 28 COOH-terminal amino acid residues of Leu3 (residues 859-886).' Similar results were obtained when polyclonal antibodies raised against a peptide containing residues 17-147 of Leu3 were used (data not shown).
Incubation of highly purified Leu3 protein with calf intestinal phosphatase in the presence of 0.2% sodium dodecyl sulfate caused a dramatic disappearance of the a band with a concomitant intensification of the p band (Fig. 7). The presence of 10 mM sodium phosphate, a phosphatase inhibitor, prevented the disappearance of the a band. These results strongly suggest that the slower moving band represents a phosphorylated form of the Leu3 protein.
Site Exclusion Chromatography of Native Leu3 Protein-For this experiment, a partially purified preparation from induced XK16OIpGALl -LEU3 cells was placed on a calibrated fast protein liquid chromatography Superose 12 column. Fractions were collected and tested for the presence of Leu3 protein using electrophoretic mobility shift assays (Fig. 8).
There was only one peak of material capable of forming specific complexes with UASL; it eluted at a position corresponding to an apparent molecular weight of about 220,000. This result suggests that DNA complex formation-competent Leu3 protein exists as a dimer in solution.

DISCUSSION
Until now, our knowledge of the structure and function of the Leu3 protein has come exclusively from molecular genetic studies of the LEU3 gene and from work with crude Leu3 preparations (2,3,(12)(13)(14)(15). A detailed characterization of the Leu3 protein and of the mechanisms by which it makes contact with DNA, the transcription machinery, and modulating factor(s) has been hampered by the very low abundance of the Leu3 protein in wild type yeast cells on the one hand and, on the other hand, by difficulties in expressing recombinant Leu3 protein in E. coli in which it tends to be degraded? Similar difficulties were reported for other eukaryotic regulatory factors, most notably Gal4 (30)(31)(32). In this paper, we have presented a simple and rapid method for isolating Leu3 protein from a Leu3-overproducingyeast strain. The key steps in the purification procedure were heat treatment and affinity chromatography. Although the heat treatment caused an obvious loss of Leu3 binding activity, this disadvantage was far outweighed by two important advantages: (i) the heat step made possible the fast isolation of active Leu3 protein by J. Sze

FIG. 8. Size exclusion chromatography of Leu3 protein.
Cell-free extract from XK160/pGALl-LEU3 cells was subjected directly to ammonium sulfate fractionation. Material precipitating a t 46% saturation (4 "C) was collected, dissolved in buffer A (see "Experimental Procedures"), and loaded on a Superose 12 column (total volume, 23 eliminating the need for further chromatographic procedures; and (ii) it caused a dramatic increase in the stability of the Leu3 protein, which we attribute to the inactivation of contaminating proteinases. Without heat treatment, the Leu3 protein loses its activity within a few days; with heat treatment, it is stable for a t least 6 months a t -70", even when it is repeatedly thawed and refrozen.
How functionally intact is the purified Leu3 protein? The electrophoretic mobility shift assays used to monitor the purification of the Leu3 protein gave practically identical results irrespective of whether crude extract or highly purified protein was assayed, indicating that no auxiliary factors were lost during purification and that the DNA binding function remained intact. We have also begun to examine Leu3 protein-dependent transcriptional activation in vitro (10). For these experiments, whole cell yeast extract capable of supporting accurate transcription by RNA polymerase I1 (11) was used. Extract was prepared from cells from which the genomic LEU3 gene had been deleted (3). The template used contained a Leu3 binding site upstream from the CYCl TATA box linked to a G-less cassette. In this system, affinitypurified Leu3 protein strongly activated transcription. Significantly, activation was absolutely dependent on the addition of a-IPM (10). In fact, in the absence of a-IPM, Leu3 caused repression of transcription. Taken together, these results strongly argue that the Leu3 protein purified by the protocol presented here is active with respect to DNA binding, tran-scriptional regulation, and response to the metabolite a-IPM, and that its in vitro behavior faithfully mimics that observed in vivo. The stage is therefore set for an in vitro study of interactions between components of the transcription machinery and Leu3 under both activating and repressing conditions, which may have important repercussions for our understanding of transcriptional regulation in general.
Highly purified Leu3 protein displayed a doublet of bands when subjected to polyacrylamide gel electrophoresis under denaturing conditions. The apparent molecular weight of the two protein bands was slightly above that predicted by the sequence of the LEU3 gene, making it highly unlikely that the two species were the result of some proteolytic event.
Both protein species co-eluted from a UASL-Sepharose affinity column, and both reacted with antibodies prepared against a peptide containing the 28 COOH-terminal residues of Leu3 as well as with antibodies prepared against a peptide containing the DNA binding region of Leu3. Thus, they were both Leu3 proteins. It is interesting in this context that the size of the smaller species was very close to that of the largest protein observed in earlier experiments in which full-length LEU3 mRNA was used to direct Leu3 protein synthesis in a rabbit reticulocyte lysate (41). That protein appeared as a singlet with an apparent molecular weight of 106,000, suggesting that the larger species seen in the present work (the 110,000 species) was generated by a yeast-specific process.
Since alkaline phosphatase treatment cleanly converted the slower moving band into the faster moving one, we believe that the two Leu3 bands represent phosphorylated and nonphosphorylated forms of Leu3. It is well known that the electrophoretic mobility of a protein can be reduced upon phosphorylation (33)(34)(35). Phosphorylated residues of Leu3 have not yet been identified, but inspection shows that the protein contains several potential sites for enzymes with the specificity of serine-and threonine-specific protein kinase C (36) and serine-and threonine-specific casein kinase I1 (37). Dephosphorylation of Leu3 by the calf intestine phosphatase required at least mildly denaturing conditions (i.e. it did not occur in the absence of sodium dodecyl sulfate) suggesting that the phosphorylated residue(s) could not be reached by this phosphatase as long as the Leu3 protein was in its native configuration. Because it was partially denatured, phosphatase-treated Leu3 protein was not suited for functional tests. It is nevertheless possible to point to some functions for which the phosphorylation/dephosphorylation process is probably irrelevant. For example, the fact that both forms of Leu3 bind equally well to the UASL affinity column would suggest that phosphorylation does not significantly affect DNA binding. Also, since both in vivo and in vitro experiments have shown that activation by the Leu3 protein does not occur unless a-IPM is present and since the purified protein used in the in vitro experiments contained both species of Leu3, it may be concluded that neither the phosphorylated nor the nonphosphorylated form of Leu3 is transcriptionally active by itself. Whether phosphorylation alters the properties of Leu3 in a more subtle way, e.g. by influencing the interaction between Leu3 and a-IPM, or whether Leu3 is phosphorylated simply as a consequence of being a transcriptional activator (as has been postulated for Gal4 (33)) remains to be settled by future experiments.
Another question that remains unanswered is the identity of the two Leu3-UASL complexes routinely seen in our electrophoretic mobility shift assays. Does the Leu3 protein in these complexes correspond to the phosphorylated and nonphosphorylated forms? Or do the two complexes represent different states of aggregation of the Leu3 protein? Size exclusion chromatography of native Leu3 protein strongly suggests that, in the absence of DNA, Leu3 exists in dimer form, irrespective of the state of phosphorylation (the material applied to the column contained both phosphorylated and nonphosphorylated forms). It has also been shown, by methylation interference footprinting, that Leu3 makes symmetrical contact with the decanucleotide core of its target DNA, which itself exhibits dyad symmetry ( 7 ) . While these observations are consistent with the idea that Leu3 protein binds DNA as a dimer, they do not rule out the existence of other oligomeric states of Leu3 when it is in contact with DNA. Attempts at protein-protein cross-linking of DNA-bound Leu3 protein have so far led to inconclusive results. Attempts at dissociating the DNA-protein complexes and subsequent isolation of the protein moieties have also been unsuccessful because of very poor recoveries. More definitive information on the DNA complex-forming ability as well as on other properties of the Leu3 protein is expected to become available once phosphorylated and nonphosphorylated forms of Leu3 can be obtained and studied as separate, nondenatured entities. To obtain nondenatured Leu3 protein of different states of phosphorylation may require identification of the phosphorylation site(s) and the use of site-directed mutagenesis to eliminate them.