Purification and Characterization of Yeast

Yeast RNA polymerase I1 general initiation factor g was purified to near homogeneity on the basis of its function in a reconstituted transcription system. Polypeptides of 30,54, and 105 kDa co-purified with transcriptional activity, forming a complex with a mass of 300 kDa as judged by gel filtration, but only 100 kDa based on sedimentation in glycerol gradients, suggesting an elongated shape. Transcription activity could be reconstituted after separation of the three polypeptides under denaturing conditions; the 54and 105kDa subunits were both essential, while the 30-kDa subunit was slightly stimulatory. Factor g was required for initiation at all promoters tested, including those from Saccharomyces cerevisiae, Schizosaccharomyces pombe, and adenovirus. Factor g can stably associate with RNA polymerase 11, as shown by cosedimentation in a glycerol gradient.

yeast counterpart of human TFIIB. 3 An accompanying paper (27) describes a fractionation scheme that allowed the isolation of factors a, b, e, and RNA polymerase I1 from a whole cell extract, and which revealed the existence of two additional factors, designated f and g. Factor f stimulated transcription with partially purified fractions, but was dispensible in reactions containing highly purified factors. Factor g was required for promoter-directed transcription with purified factors a, b, e, TFIID, and RNA polymerase 11. Here we report the purification and properties of factor g, and show that it can stably associate with highly purified RNA polymerase 11.

EXPERIMENTAL PROCEDURES
Materials and Buffers-Chromatography resins, reagents used in transcription assays, lysis buffer, and buffers A, B, C, D, E, and F were as described (27). Buffer G was the same as buffer E but with 10% glycerol and without Nonidet P-40. Buffers contained protease inhibitors (Sigma) as described (4).
Purification of Factor g-Yeast (active dry which had not been oven-dried, 3 kg) was obtained from Fleischmann's Yeast (Oakland, CA) and washed in cold distilled water. Subsequent manipulations were performed at 0-4 "C. Cells were suspended in 1 liter of 3 X lysis buffer, broken with glass beads as described (5), and centrifuged in a Beckman JA-10 rotor for 20 min at 8,000 rpm. Potassium acetate (one-seventh volume of a 4 M solution, pH 7.8) was added, and the mixture was stirred for 15 min and centrifuged at 42,000 rprn for 90 min in a Beckman 45Ti rotor. The supernatant (whole cell extract, 1.6 liters) was diluted with 4 liters of buffer A-0 and applied at 800 ml/h to a Bio-Rex 70 column (17 X 9 cm) equilibrated in buffer A-0.1 (buffer A containing 0.1 M potassium acetate). The column was washed with 1 liter of buffer A-0.1, and was eluted with 2 liters of buffer A-0.3 and 1 liter of buffer A-0.6. The latter eluate (fraction I, Table I) was dialyzed for 6 h against buffer B-0, diluted with 250 ml of buffer B-0 to a total volume of 770 ml, centrifuged for 10 min at 10,000 rpm in a Sorvall SS-34 rotor, and applied at 300 ml/h to a DEAE-Sephacel column (15 X 5 cm) equilibrated in buffer B-0.1. The column was washed with 300 ml of buffer B-0.1, and was eluted with 600 ml of buffer B-0.2 and 300 ml of buffer B-0.4. The latter eluate (fraction 11) was applied at 1.6 ml/min to a hydroxylapatite column (9.8 X 3 cm) equilibrated in buffer C. The column was washed with 70 ml of buffer C and eluted with a linear gradient (700 ml) to buffer D. Factors b and g co-eluted in a broad peak from 70 to 120 mM potassium phosphate. Fractions containing activity were pooled (fraction 111), dialyzed against buffer E-0.05 to the conductivity of buffer E-0.1, centrifuged at 20,000 rpm for 20 min in a Beckman 45Ti rotor, and applied at 2.5 ml/min to a Bio-Gel DEAE-5-PW HPLC column (150 X 21.5-mm; Bio-Rad) equilibrated in buffer E-0.15. The column was developed with a linear gradient (540 ml) to buffer E-1.0. Factor b and g transcription activities co-eluted at 0.4 M potassium acetate. Half of the pool of peak fractions (40 ml of fraction IV) was dialyzed against buffer F-0.025 to the conductivity of buffer F-0.1, centrifuged at 20,000 rpm for 20 min in a Beckman Ti60 rotor, and applied at 0.5 ml/min to a Mono S HR5/5 column (Pharmacia LKB Biotechnology Inc.) equilibrated in buffer F-0.1. The column was developed with a linear gradient (20 ml) to buffer F-0.5, and the peak of factor g activity eluted at 0.42 M potassium acetate. Active fractions were pooled (fraction V), dialyzed against buffer G-0.025 to the conductivity of buffer G-0.1, centrifuged in a microcentrifuge at 13,000 rpm for 15 min, and applied at 0.3 ml/min to a TSK-heparin-5-PW HPLC column (75 X 7.5 mm; Supelco) equilibrated in buffer G-0.1. The column was developed with a linear gradient (33 ml) to buffer G-0.7, and factor g eluted in a sharp peak at 0.4 M potassium acetate (fraction VI). A portion of fraction VI (5.2 ml) was centrifuged at 13,000 rpm for 15 min in a microcentrifuge, and applied at 0.5 ml/min to a Mono Q HR5/5 column (Pharmacia) equilibrated in buffer E-0.4. The column was developed with a linear gradient (20 ml) to buffer E-1.0, and factor g eluted at 0.8 M potassium acetate. Peak fractions (1.6 ml) were pooled and used for analytical gel filtration and glycerol gradient sedimentation experiments.
In some factor g preparations, an SP-5-PW column (75 X 7.5 mm; Bio-Rad) was included after the DEAE-5-PW step. The DEAE pool (fraction IV) was dialyzed against buffer F-0.05 to the conductivity of buffer F-0.2, centrifuged in a Beckman 45Ti rotor for 20 min at 20,000 rpm, and applied at 0.5 ml/min to an SP-5-PW column equilibrated in buffer F-0.2. The column was developed with a linear gradient (30 ml) to buffer F-0.7, and the peak of factor g eluted at 0.4 M potassium acetate.
A preparation of Mono S factor g (fraction V) was obtained from the yeast laboratory strain BJ926 as described (27), and subjected to chromatography on the TSK-heparin-5-PW and Mono Q columns as described above. Peak fractions from the Mono Q column were pooled, and a portion (1 ml) was centrifuged at 13,000 rpm for 15 min and applied at 0.25 ml/min to a Bio-Gel HPHT cartridge (30 X 4.6 mm; Bio-Rad) equilibrated in buffer C containing 10 p~ calcium chloride. The column was developed with a linear gradient (10 ml) to buffer D containing 10 p~ calcium chloride. Factor g was recovered in a broad peak (3 column volume) from 112 to 125 mM potassium phosphate.
Molecular Weight Estimation of Factor g-Mono Q-purified factor g (0.25 ml) was applied at 0.2 ml/min to a Superose-12 HR10/30 column (Pharmacia) equilibrated in buffer B-0.3, and 0.5-ml fractions were collected. The native mass of factor g was estimated as described for factor a (28) by comparing the elution volume of factor g with those of yeast alcohol dehydrogenase (150 kDa), sweet potato @amylase (200 kDa), horse spleen apoferritin (443 kDa)(Sigma), and blue dextran 2000 (Pharmacia) chromatographed separately in identical conditions.
Glycerol Gradient Sedimentation Analysis-Mono Q-purified factor g (0.25 ml) was dialyzed against 20 mM Tris, pH 7.8,O.l M potassium acetate, 0.5 mM EDTA, 1 mM dithiothreitol, and protease inhibitors until the glycerol concentration was less than lo%, and was layered on top of a 10-25% glycerol gradient (11 ml) formed in the same buffer but containing 0.3 M potassium acetate. The gradient was centrifuged in a Beckman SW41 rotor at 35,000 rpm for 21.5 h a t 4 "C, and fractions (0.4 ml) were collected from the bottom of the tube. Protein molecular weight standards (see above) were sedimented in a parallel gradient of identical composition.
Factor g fraction VI (400 pl; 20 pg) and purified RNA polymerase I1 (400 pl; 50 pg) were dialyzed separately against 20 mM Tris, pH 7.8,0.5 mM EDTA, 0.2 M potassium acetate, 1 mM dithiothreitol, and protease inhibitors for 1 h at 4 "C, and 200 pl of each dialysate were combined, incubated at 4 "C for 10 min, then at 22 "C for 10 min, layered on top of a 10-25% glycerol gradient (11 ml) formed in dialysis buffer, and centrifuged in a Beckman SW41 rotor for 19 h a t 35,000 rpm at 4 "C. Fractions (0.4 ml) were collected from the bottom of the tubes and assayed for transcription activities and proteins as described below. RNA polymerase I1 alone (200 pl) and factor g alone (200 pl) were analyzed separately in parallel gradients. Transcription Assays-The reconstituted transcription assay for factor g contained (in 25 pl) 250 ng of plasmid DNA template, 3 ng of factor a (fraction VI; Ref CYCl promoter (pGAL4CG-or pACG-; 11) were used unless stated otherwise. Other conditions and procedures were as described (27) except that reactions contained 105-115 mM potassium acetate. Factor b was assayed with heat-treated nuclear extract as described (5), with the following modifications. Transcription reactions (25 pl) contained 5-6 pl of heat-treated nuclear extract, 100 ng of yeast TFIID (lo), 0.5-1 pl of column fractions, 250 ng of pACG-template, 40-60 mM Hepes-KOH, pH 7.6, 80 mM potassium acetate, 7.5 mM magnesium acetate, 6 mM EGTA, 25 p~ UTP, 0.8 mM ATP, 0.8 mM CTP, 10 pCi of [a-"P]UTP), 0.2 units of Inhibit-Ace (5-Prime + 3-Prime; Boulder, CO), 1.5 mM dithiothreitol, and 4.8 mM phosphoenolpyruvate. Reactions were incubated for 60 min at 20 "C before the addition of RNase TI (Behring Diagnostics), and products were processed for denaturing gel electrophoresis and autoradiography as described (5).
For purification of factor e for reconstituted transcription assays, factor e fraction IV was prepared from commercial yeast (750 g) as described? This fraction was dialyzed against buffer F-0.025 to the conductivity of buffer F-0.1, and applied at 4 ml/h to a DEAE-Sephadex A-25 column (2 ml; Pharmacia) equilibrated in buffer E-0.1. The flow-through, which contained factor e, was applied at 0.5

TABLE I
Purification of factor g from 3 kg of commerical yeast One unit of activity is the amount needed to produce 0.1 fmol of specifically initiated transcripts in the reconstituted transcription system. Factor g activity in whole cell extract could not be measured due to the presence of inhibitors.

FIG. 3. Co-purification of factor g subunits and activity on a gel filtration column and in a glycerol gradient. A, gel filtration on a
Superose-12 HR 10/30 column. Upper, SDS-10% polyacrylamide gel of column fractions (400 pl) indicated by numbers above the lanes. Molecular weight markers (sizes in kDa) are indicated on the left. Fractions 21, 23, and 24 from Superose-12 represented elutions volumes of horse spleen apoferritin (443 kDa), sweet potato &amylase (200 kDa), and yeast alcohol dehydrogenase (150 kDa), respectively, in a separate calibration run performed under identical conditions. Lower, transcription assays of fractions (7.5 pl), with control reactions from which fractions were omitted (-g) or to which 1.5 pl of material applied to the column (load) was added. R, sedimentation in a 10-25% glycerol gradient. Upper, SDS-13% polyacrylamide gel of gradient fractions (350 pl). In a separate glycerol gradient run in parallel with the gradient analyzed here, fractions 9, 16, and 19 contained the peaks of apoferritin, &amylase, and alcohol dehydrogenase, respectively. Lower, transcription assays of fractions (8 pl), with control reactions as in panel A (2 pl of load). Reactions contained 1.5 pl of the fractions indicated above the lanes, renatured as described under "Experimental Procedures." g (fraction IV) was subjected to chromatography on an SP-5-PW column as described above, purified further on TSK-heparin-5-PW followed by Mono S, acidified with 0.1% trifluoroacetic acid, and applied at 1 ml/min to a HiPore Reversed Phase RP-304 column (250 X 4.6-mm; Bio-Rad) equilibrated in 0.1% (v/v) trifluoroacetic acid. The column was developed with a linear gradient (88 ml) to 100% acetonitrile containing 0.1% trifluoroacetic acid, and fractions were analyzed by SDS-polyacrylamide gel electrophoresis. Fractions containing the 30-, 54-, and 105-kDa polypeptides were evaporated to dryness, separately resuspended with 7 pl of 6 M guanidinium-HCI, incubated at 24 "C for 40 min, diluted with 350 pl of 50 mM Hepes-KOH, pH 7.6,0.5 mM EDTA, 0.1 M potassium acetate, 0.01% Nonidet P-40, 10% glycerol, 1 mM dithiothreitol, and protease inhibitors, dialyzed against the same buffer for 60 min, and assayed for factor g transcription activity.
Other Methods-Protein concentrations were determined by the method of Bradford (12) using bovine serum albumin as standard. Proteins were concentrated for SDS-polyacrylamide gel electrophoresis (13) by precipitation with 10% trichloroacetic acid using sodium deoxycholate as carrier, and were visualized in gels by silver staining (14).

RESULTS AND DISCUSSION
Purification of Factor g-Purification of factor g from yeast whole cell extract was monitored by transcription reconstituted with highly purified factors a, b, e, yeast TFIID, and RNA polymerase 11. Factors b and g co-purified through the first four chromatographic steps, and were then separated on a Mono S fast protein liquid chromatography column (Fig.  LA). The recovery of factor g activity from the Mono S column in the example shown here (Table I, fraction V) was lower than observed in other independent preparations.
To the extent that factor b was limiting in the transcription assay, its presence in fraction IV would lead to an overestimation of the amount of factor g activity in that fraction, and a corresponding underestimation of the recovery of factor g in fraction V. It is also possible that inhibitory contaminants were concentrated along with factor g on Mono S, since subsequent chromatography of fraction V on a heparin-HPLC column gave a 3-fold increase in total activity. Further fractionation on a Mono Q fast protein liquid chromatography column gave Co-sedimentation of factor g with RNA polymerase I1 in a glycerol gradient. A, upper, transcription assays of fractions (7.5 p l ) from glycerol gradient analysis of factor g alone. Middle, transcription assays of fractions (6 p l ) from glycerol gradient analysis of RNA polymerase I1 (pol ZZ) alone. Reactions were as described for factor g except RNA polymerase I1 was omitted and 0.15 pg of factor g (fraction VII) was included. Lower, transcription assays of fractions (8 pl) from glycerol gradient analysis of a mixture of factor g and RNA polymerase 11. Reactions contained all components except those indicated on the right. B, SDS-10% polyacrylamide gel showing fractions from glycerol gradient analysis of a mixture of factor g and RNA polymerase 11. Positions of molecular weight markers (sizes in kDa) are indicated on the left. Arrows and asterisks on the right indicate factor g and RNA polymerase I1 subunits, respectively. a small increase in specific activity but with poor yield, so this step was not routinely used. Starting from the first chromatographic fraction (Bio-Rex 70), factor g activity was enriched about 5000-fold. Since the activity in the whole cell extract was not determined due to the presence of inhibitory material, the full extent of purification is unknown.
Subunit Composition of Factor g-Highly purified factor g contained three major polypeptides of 30, 54, and 105 kDa (Fig. 1B). In the example shown, the largest polypeptide decreased in size slightly between the Mono S and TSKheparin-5-PW steps, possibly due to proteolysis, but this effect was not observed with other preparations of factor g. The three polypeptides precisely co-purified with activity on Mono Q and hydroxylapatite HPLC columns (Fig. 2). The two larger subunits also co-purified with activity on a gel filtration column and in a glycerol gradient (Fig. 3). The smallest subunit, however, was always distributed more broadly, displaced towards the leading edge of the factor g activity peak in the gel filtration column, and trailing into Yeast RNA Polymerase 1 1 General Initiation Factor g fractions that were largely devoid of both the 54-and 105-kDa subunits and transcription activity in the glycerol gradient. These data suggested that the 54-and 105-kDa subunits of factor g were required for transcription, while the role of the 30-kDa subunit was less well defined.
The roles of the individual subunits in transcription were clarified by their separation on a reversed phase HPLC column under denaturing conditions (Fig. 4). The 30-and 105-kDa polypeptides were each recovered in an apparently homogeneous form, but the 54-kDa polypeptide remained slightly contaminated with the 105-kDa polypeptide. (An apparent trace of a 54-kDa polypeptide in the 105-kDa peak fraction was an artifact of silver staining in the gel shown in Fig. 4A, and was not evident in other gels containing the same fraction.) The poiypeptides were renatured separately and assayed for transcriptional activity (Fig. 4B). The combination of the 54-and 105-kDa polypeptides supported transcription (second lanefrom right), and addition of the 30-kDa polypeptide was slightly stimulatory. No other painvise combination or individual polypeptide supported significant transcription. The slight activity of the 54-kDa polypeptide alone or in combination with the 30-kDa polypeptide (third and fifth lanes from left, respectively) was probably due to the small degree of contamination by the 105-kDa polypeptide. We conclude that the 54-and 105-kDa subunits of factor g are necessary and sufficient for transcriptional activity, while the 30-kDa subunit is dispensable.
The native mass of factor g was estimated to be 300 kDa by gel filtration, but the sedimentation rate in glycerol gradients corresponded to a globular 100-kDa protein. Since the three associated polypeptides would form a complex of at least 190 kDa, these results suggest that factor g may have an elongated shape. Further work is required to determine the subunit stoichiometries and conformation of factor g, and to investigate the role of the closely associated 30-kDa subunit.
Interaction of Factor g with RNA Polymerase ZZ-Several yeast and mammalian initiation factors have been shown to associate with RNA polymerase 11 in vitro (15).2,4 The possibility of a factor g-polymerase interaction was investigated by glycerol gradient sedimentation (Fig. 5). When factor g and polymerase were analyzed separately, they sedimented at very different rates, but when analyzed together, they co-sedimented at a rate faster than either component alone, demonstrating the stable association of factor g with the polymerase in the presence of 0.2 M potassium acetate. Factor g also co-sedimented on a glycerol gradient with RNA polymerase I1 lacking the fourth and seventh largest subunits (16): indicating that those two subunits are not required for the interaction between factor g and polymerase. Binding of factor g to RNA polymerase I1 exhibited specificity since both the sedimentation rate of factor g in a glycerol gradient and the elution volume of factor g from the Superose-12 column were unaffected by the presence of apoferritin, @-amylase, or yeast alcohol dehydrogenase.' Conversely, contaminants in factor g failed to co-sediment with polymerase ( Fig. 5B; see fraction 24). Furthermore, a different yeast initiation factor, factor a, H. Serizawa, J. W. Conaway, and R. C. Conaway, manuscript N. Lynn Henry, unpublished observation.

submitted.
also failed to co-sediment with polymerase under identical conditions. 6 Further Characterization of Factor g-Purified factor g is required for transcription directed by the Saccharomyces cerevisiae CYCl, GALIO,ADCl (27) and PYKl promoters, the adenoviral major late promoter, and the Schizosaccharomyces pombe ADHl promoter: demonstrating its role as a general initiation factor for yeast RNA polymerase 11. In view of its action at the adenoviral promoter, factor g seems likely to be related to a general initiation factor identified in human or rat RNA polymerase I1 transcription systems (1,2). As yeast homologs of human TFIID/rat 7 (17,18), TFIIB/a (19,20), and BTF-2/6 (8, 9) have been identified,2s3*5 the remaining candidates appear to be TFIIE/c (18, 21-23) and TFIIF(RAP30/74)//3y (24-26). Both TFIIE/e and TFIIF/@y independently associate with RNA polymerase I1 (15)4, as does factor g. The mammalian factors are each made up of two polypeptides, similar in size to the two smaller subunits of factor g. Since, however, only the two larger subunits of factor g appear to be required for transcription, it remains unclear whether factor g is homologous to any mammalian transcription factor. Cloning and sequencing of the genes for factor g will resolve the issue.