Cro Regulatory Protein Specified by Bacteriophage X STRUCTURE, DNA-BINDING, AND REPRESSION OF RNA SYNTHESIS*

The Cro protein specified by bacteriophage A is a repressor of the genes expressed early in phage development and is required for a normal late stage of lytic growth. We have purified Cro protein to virtual homogeneity and analyzed its structure and properties as a DNA-binding protein and repressor of RNA synthesis. To confirm that the protein is the product of the cro gene, we have also shown that a missense mutation in the ero gene leads to a product that is more temperature-and salt-sensitive in its DNA-binding property. As purified, Cro protein is a dimer of identical subunits of molecular weight 8600. The purified protein binds to A-DNA carrying the specific binding sites (operators oL and oli) with an estimated dissociation constant of lo-‘” M to lo-” M; there is also weaker binding to other sites on DNA, as found for other DNA-binding regulatory proteins. In a purified transcription system, the Cro protein is an effective and specific repressor of RNA synthesis from the N and cro genes; thus Cro is an autorepressor which regulates its own synthesis. A comparison of the properties of the two A repressor proteins, CI and Cro, indicates that cI is a “strong repressor” specialized for complete turnoff of lytic functions needed for the maintenance of lysogeny, whereas Cro is a “weak repressor” specialized for a gradual turnoff of early viral genes that potentiates the late stage of lytic development.


The Cro protein specified by bacteriophage
A is a repressor of the genes expressed early in phage development and is required for a normal late stage of lytic growth. We have purified Cro protein to virtual homogeneity and analyzed its structure and properties as a DNA-binding protein and repressor of RNA synthesis. To confirm that the protein is the product of the cro gene, we have also shown that a missense mutation in the ero gene leads to a product that is more temperature-and salt-sensitive in its DNA-binding property. As purified, Cro protein is a dimer of identical subunits of molecular weight 8600. The purified protein binds to A-DNA carrying the specific binding sites (operators oL and oli) with an estimated dissociation constant of lo-'" M to lo-" M; there is also weaker binding to other sites on DNA, as found for other DNA-binding regulatory proteins. In a purified transcription system, the Cro protein is an effective and specific repressor of RNA synthesis from the N and cro genes; thus Cro is an autorepressor which regulates its own synthesis. A comparison of the properties of the two A repressor proteins, CI and Cro, indicates that cI is a "strong repressor" specialized for complete turnoff of lytic functions needed for the maintenance of lysogeny, whereas Cro is a "weak repressor" specialized for a gradual turnoff of early viral genes that potentiates the late stage of lytic development.
The temperate bacteriophage A specifies two repressor proteins, c1 and Cro, which carry out regulatory functions essential for different aspects of the viral life cycle. The c1 protein acts under conditions of stable lysogeny to maintain repression of the integrated viral DNA; Cro protein acts during lytic development to turn off the expression of the phage genes active early after infection (l-6) and thus potentiates the early-late switch in expression of viral genes.
The c1 protein has been purified and extensively characterized in vitro for binding to specific operator sites on X-DNA and ability to repress RNA synthesis initiated at the X promoter sites active early during viral development (7-11) (Fig. 1). In a previous paper, we presented data indicating that Cro protein is a DNA-binding protein which binds to the same operator region of A-DNA as does c1 (12). This report describes the purification of Cro protein to apparent homogeneity, presents a more detailed characterization of its structure and DNA-and bovine serum albumin from Schwarz/Mann. Bacterial and Phage Strains -The bacterial host used to prepare Cro protein was CGOOSu-. The infecting phage for large scale preparations of wild type Cro protein was hNam53ulu3cIaml4Sam7; to characterize the temperature-sensitive mutant protein, parallel infections by hNam53ulu3cro+ and hNam53Num7ulo3cro~ were used, in which the cro-mutation was tof2 (15). The A mutations and the rationale for using them have been described in more detail previously (12). In brief, N-mutation eliminates production of most h proteins besides Cro, ulu3 may increase Cro production, cI-mutation eliminates the DNA-binding activity of c1 protein, and S-mutation prevents cell lysis. Other Mate&Es-Cellulose (CFll) and phosphocellulose (PII) were obtained from Whatman and Sephadex G-75 (140 to 120 CL particle size) from Pharmacia. h-DNA-cellulose was prepared as described by Alberts and Herrick (16 For assays involving purified Cro protein, the chicken blood DNA was omitted.
One unit of DNA-binding activity is defined as the quantity sufficient to retain 1 pg of A-DNA on the filter. The binding values are corrected for a "background" of DNA that is retained on the filter in the absence of binding proteins; this varies from 5 to 15% with DNA preparation.
For the experiments to determine the kinetic and equilibrium binding constants (Figs. 7 to 9), the washing procedure was modified by the use of a 0.5ml wash with 45% ethanol.
This procedure gave higher and more reproducible binding data with lower levels of Cro protein, presumably because dissociation during the washing step did not occur (no decrease in binding was found up to at least 1 ml of wash volume). Binding values similar to those of the ethanol wash were obtained by reducing the volume of the standard washing buffer (0.1 ml), but the data were erratic with a higher level of background binding. RNA Synthesis and Analysis-RNA synthesis was carried out using purified RNA polymerase and hb2-or Ab2imm434-DNA in a reaction in which RNA chains were initiated in the presence of rifampicin; the use of DNA carrying the b2 deletion and of rifampicin results in a larger fraction of RNA from the promoter sites on A-DNA used in viva for early RNA (10,14,17). The standard reaction mixture (0.1 ml) contained: 100 rnM Tris/HCl (pH 7.2), 40 mM NaCl, 10 mM MgCIZ, 1 rnM ATP, GTP, CTP, and 0.1 mM [5JH]-or [(Y-32PlUTP, 1 pg of rifampicin, 0.7 pg of RNA polymerase, and 3 pg of A-DNA.
The A-DNA and Cro protein were first incubated for 8 to 10 min at 17", and RNA polymerase was then added and the mixture incubated for another 20 min at 17"; RNA synthesis was initiated by the addition of the four nucleoside triphosphates together with rifampicin and the mixture was incubated at 30". For measurements of total RNA synthesis, the reaction was terminated after 10 min by the addition of trichloroacetic acid to 5%, and the amount of RNA synthesis was determined by acid-insoluble radioactivity retained on a Millipore filter after filtration and washing with 5% trichloroacetic acid and 70% ethanol.
For analysis of RNAs made, RNA synthesis was carried out in a 0.2-ml reaction mixture in the presence of p factor (4 yglml) and [(u-32PlUTP. After lo-min incubation at 30", 5 pg of pancreatic DNase ("RNase-free") and 150 pg of E. coli tRNA were added and the mixture was kept for 20 min at 0". RNA was purified by phenol extraction and three precipitation steps with 2 volumes of ethanol in the presence of 200 mM potassium acetate (pH 5.5). The precipitated RNA was resuspended in 0.05 ml of 80 mM Tris/HCl (pH 8.3), 2.5 mM EDTA, 8 M urea, and subjected to electrophoresis in a 3.5% polyacrylamide slab gel at 50 V for 10 h in the same buffer. The RNA bands were identified by autoradiography using x-ray film (Kodak RP-54) as described by Rosenberg et al. (18).

Miscellaneous
Methods -Polyacrylamide gel electrophoresis of denatured protein was carried out in 10% polyacrylamide and 0.1% When the A,!,,, of the culture reached 1.0, MgCl, was added to 10 mM and h phage (strain Nam53cIam14ulv3Sum7) was added at a multiplicity of 10 phages/cell. After 60 min at 30", the cells were harvested by centrifugation in a Sharples continuous flow centrifuge, resuspended in 50 ml of 10% sucrose, 50 mM Tris/ HCI (pH 7.8, quick-frozen, and stored at -20". The procedure yielded 150 g of cells.
Preparation of Extract-The frozen cells were thawed and mixed with 80 ml of 2 mg/ml of lysozyme in 250 mM Tris/HCl (pH 7.51, 1 mM EDTA, and 19 ml of 4 M NaCl, 10 mM 2mercaptoethanol. After 30 min at o", lysis was completed by raising the temperature gradually to 32" over a period of 10 to 15 min. Magnesium acetate was added to 10 mM, and pancreatic DNase to 2 pg/ml. After the viscosity was substantially reduced (5 to 10 min), 50 ml of 4 M NaCl were added, and the lysate was centrifuged for 4 h at 30,000 rpm in a Spinco 30 rotor. The supernatant fraction (550 ml) was dialyzed for 3 h (two changes) against 5 liters of 10 mM KPO, (pH 6.41, 0.2 mM EDTA, 0.2 mM dithiothreitol, 5% glycerol (Buffer A) containing 0.1 M KC1 (Fraction I).
Phosphocellulose Chromatography -Fraction I was applied at a flow rate of 3 ml/min to a phosphocellulose column (go-ml bed volume) equilibrated with Buffer A containing 0.1 M KCI. The column was washed with 150 ml of Buffer A containing 0.1 M KCl, and then eluted at a flow rate of 0.8 ml/min with a linear gradient (500 ml total volume) from 0.1 to 1.0 M KC1 in Buffer A. Cro protein, showing the DNA-binding activity specific for A-DNA, eluted at a KC1 concentration of approximately 0.45 M (Fraction II). Binding specificity for A-DNA was checked by parallel assays with both A-DNA and himm434-by guest on March 24, 2020 http://www.jbc.org/ Downloaded from DNA, which lacks the specific binding sites for Cro protein (12).

Sephadex G-75 Gel Filtration-
Fraction II was concentrated by (NH&SO, precipitation (70% saturation at pH 6.0 for 20 min), centrifugation for 20 min at 13,000 r-pm, and resuspension in 3 ml of 10 mM Tris/HCl (pH 7.31, 0.2 mM EDTA, 0.2 rnM dithiothreitol, 5% glycerol (Buffer B) containing 0.2 M NaCl. The concentrated protein solution was applied to a column of Sephadex G-75 (90 x 2.5 cm) equilibrated with Buffer B containing 0.2 M NaCl, and eluted with the same buffer at a flow rate of 15 ml/h. The DNA-binding activity specific for X-DNA eluted between 260 and 300 ml (approximately the position where a marker myoglobin protein elutedl (Fraction III).
DNA-cellulose Chromatography-Fraction III was diluted to 0.1 M NaCl with an equal volume of Buffer B and applied to a h-DNA-cellulose column (lo-ml bed volume; 5 mg of A-DNA) equilibrated with Buffer B containing 0.1 M NaCl. The column was washed with 20 ml of Buffer B containing 0.1 M NaCl and eluted with a linear gradient (50 ml total volume! from 0.1 M to 1.0 M NaCl in Buffer B. The DNA-binding activity specific for A-DNA eluted at a NaCl concentration of approximately 0.3 M (Fig. 2a) (Fraction IV). Fraction IV was diluted to 0.1 M NaCl with Buffer B and rechromatographed on h-DNA-cellulose (5ml bed volume) with a linear gradient (30 ml total volume) from 0.1 to 1.0 M NaCl in Buffer B (Fig. 2b). The pooled fractions, 18 to 21 (Fraction V), were used for the further studies described in this paper. Fraction V was free of DNase activity, as judged by sedimentation of 0.6 pg of A-DNA in an alkaline sucrose gradient after prior incubation with 0.6 pg of Cro protein for 30 min at 30". Fraction V was also free of RNase activity, as judged by no detectable release of acid-soluble radioactivity when 0.01 pg of A-13H]RNA was incubated for 15 min at 30" with 0.5 pg of Cro protein or by no change in the size of 4 S and 6 S RNA analyzed by polyacrylamide gel electrophoresis.
Cro protein was stored at -20" in Buffer B with 0.4 M NaCl and 50% glycerol without loss of activity over a l-year period.
Properties of Altered Cro Protein Produced by cro-Mutation -Although the binding specificity and the conditions for synthesis provided a strong indication that the DNA-binding protein purified above is the product of the cro gene, we have also characterized an altered protein produced by a phage with a missense mutation in the cro gene in order to complete the identification of the Cro protein. For this purpose 2-liter cultures of E. coli CGOOSu-were infected with cro+ or cro-(tof2 mutation)' phage at 30" for 60 min, and a smaller scale puritication was carried out by phosphocellulose chromatography, followed by concentration with dry Sephadex G-75 and sedimentation in a 10 to 30% glycerol gradient in Buffer A containing 0.4 M KCl. Even when assayed at O", the specific activity of the mutant Cro protein thus isolated was about 15 to 20% that of wild type Cro protein, as judged by the amount of Cro protein determined by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate (both preparations contained about 60 to 70% Cro protein).
The normal and mutant Cro proteins were characterized for sensitivity of the DNA-binding reaction to elevated temperature ( Fig. 31 and ionic strength (Fig. 4). There is a more severe inhibitory effect of both temperature and ionic strength on the ' tof2 is a temperature-sensitive cro mutation (15)  DNA-binding activity of the protein produced by the crophage. From these results and others (see Ref. 12 and below), we conclude that the DNA-binding activity specific for A-DNA that we are studying is the product of the A cro gene.
FIG. 4. Salt sensitivitv of Cro urotein suecified bv mutant cro gene. DNA binding assays were-carried out as described under "Methods," except that the binding mixtures included the KC1 concentration indicated on the figure. The nonspecific binding to himm434-DNA has been subtracted from the binding to A-DNA to give the data presented in the figure. The binding to himm434-DNA at 0.03 M KC1 was 3% with the Cro+ preparation and 11% with the Cro-ureuaration: this decreased linearlv with increased salt concen-

Physical and Chemical Properties of Cro Protein
Physical Structure-To estimate the molecular weight of native Cro protein, we carried out velocity sedimentation in a 10 to 30% glycerol gradient (Fig. 5). Two salt concentrations were used in an effort to determine the stability of the subunit structure. In both 0.05 M KC1 and 0.5 M KCl, Cro protein has an estimated sedimentation coefficient of 1.9 to 2.0 S, as judged by the sedimentation of marker proteins of known molecular weight. This indicates a molecular weight of 15,000 to 20,000, assuming that the axial ratio is in a typical range for a globular protein (22).
To determine the monomer molecular weight and estimate the purity of the final preparation, we used polyacrylamide gel electrophoresis in sodium dodecyl suiiate. Only a single protein species was found when 7 pg of the preparation of Cro protein were used (Fig. 6); after electrophoresis of 28 pg, three faint minor bands were discernible. From these results we judge the preparation to be more than 95% Cro protein. The monomer molecular weight of Cro protein is estimated to be approximately 9000 from the migration of marker proteins of known molecular weight.
Amino Acid Composition-The amino acid composition of Cro protein is presented in Table II. The composition is high in lysine and alanine and lacks cysteic acid and tryptophan, showing an interesting similarity to the prokaryotic DNAbinding protein HU (23,241 and to the eukaryotic histone H2B (25). From the amino acid composition, the monomer molecular weight is determined to be 8600. From the combined physical and chemical studies, we conclude that native Cro protein is probably a dimer of identical subunits.

Equilibrium
Binding-Previous experiments have shown that Cro protein binds to the same operator region used by the A c1 protein, the "A repressor" that maintains lysogeny (12).  We wanted to determine the dissociation constant for Cro and compare it to the very low value of lo-l3 M estimated for c1 protein.
In order to analyze the binding data, we needed to know that Cro binding is sufficiently specific for the operator sites on A-DNA for this interaction to dominate the binding curve and that one active Cro protein is sufficient to retain the radioactive A-DNA on the nitrocellulose filter in the standard binding assay. This information is provided by the binding curve of Fig. 7, in which Cro concentration is varied for two DNA substrates, A and Aimm434; Aimm434 is mainly identical with A but lacks the region of A-DNA containing the specific binding sites for c1 and Cro proteins (7, 10, 12). The binding to A-DNA is linear at low concentrations of Cro protein, indicating that 1 active Cro molecule can retain 1 A-DNA mo1ecule.2 The binding is also largely specific for A-DNA. The "nonspecific" binding that we have observed for Aimm434-DNA has also been found for other DNA-binding regulatory proteins (26,271, and presumably represents relatively weak binding interactions that can occur anywhere on a DNA molecule (a similar binding curve to that of Aimm 434 has been found also for (P80 DNA). From the data of Fig. 7, we conclude that the standard * Because of the relatively large dissociation constant of Cro protein, the washing procedure used for the filter assay is important (see "Methods"). DNA-binding assay will allow us to estimate specific binding constants.
A binding curve in which DNA concentration is varied is most appropriate for the measurement of an equilibrium dissociation constant (28,29); the results of this experiment for Cro protein are shown in Fig. 8. The detailed interpretation of the binding curve is complicated by the fact that we do not know precisely how many specific binding sites for Cro are present on a A-DNA molecule. For A c1 protein, there are three binding f"operator") sites on either side of the cI gene, termed ~a,, oa2, and or,:) and oi,,, oL2, and oL3; of these oR, and oL, have substantially higher affinities for c1 than the others (30). For Cro protein, we only know so far that Cro binds to at least two sites in the 0,. and oH region (see Ref. 12 and below). If we assume two binding sites per 3 x lo7 daltons of A-DNA (311, we calculate a dissociation constant of 8 to 9 x 10-l' M. Comparisons with dissociation constants of other proteins are difficult to interpret because the measurements are subject to substantial variation with ionic strength, temperature, and pH. However, Cro does appear to be a substantially weaker DNAbinding protein than other specific regulatory proteins studied so far; approximate values for A c1 protein, lac repressor, and araC protein are lo-l3 M (7), lo-l3 M (281, and lo-'* M (29), respectively.Y

Dissociation
Rate Constant -To estimate the dissociation rate constant, we formed a Cro.X-13'PlDNA complex and measured the loss of X-1321DNA from this complex with time in the presence of a 30-fold excess of unlabeled A-DNA (Fig. 9). Since a Cro molecule should not reassociate to a significant extent with the 132P1DNA once released, the initial decrease in [32P]DNA bound should follow an exponential first order decay curve. Because of the rapid decay found under our standard binding condition of pH 7.3, we also carried out binding assays at pH 6.6. The calculated dissociation rate constant is about 2 x lo-* s-l at pH 7.3 and 5 x 10m3 s-l at pH 6.6. As expected from the equilibrium binding data, the half-life for Cro dissociation at pH 7.3 is much less than that found for c1 protein and lac repressor, and substantially less than the 3-min value found for araC protein.
In an experiment carried out under identical conditions, we found the half-life for c1 dissociation to be greater than 100 min. We have also attempted to measure the dissociation rate for "nonspecific" binding, using DNA lacking the specific operator sites (kimm434-or @80-DNA). The dissociation was very rapid with a half-life of 510 s, which is consistent with the concept that the nonspecific binding is a substantially weaker interaction than the specific binding to the regulatory sites.
We have attempted to determine the association rate constant, but have not been able to do so because Cro protein is unstable at the very high dilution required for this measurement. If the rate constant for Cro is comparable to that estimated for araC protein (2 x 10y M-' SK'), the value calculated for the equilibrium dissociation constant is lo-" M at pH 7.3.

Effect of Cro Protein on RNA Synthesis
From experiments carried out in vivo, we expect Cro protein to act as a specific repressor of RNA synthesis initiated at the early promoter sites of A-DNA, pL and pR (see Refs. 1 to 6 and Fig. 1). To test this expectation, we used purified A-DNA as a template for RNA polymerase and studied the effect of Cro protein on RNA synthesis. To ensure that Cro effects derive from the specific binding reaction, we used Ximm434-DNA as a template in parallel experiments.
The capability of Cro protein to function as a specific repressor in vitro is shown in Fig. 10, in which strong repression of total RNA synthesis from A-DNA occurs in the absence of any repression effect on RNA synthesis from Aimm434-DNA.
The RNA products were analyzed further by separation through polyacrylamide gel electrophoresis and visualization by auto- RNA synthesis was carried out in the presence of p factor using 13*PlUTP. The 13*PlRNA was extracted, subjected to electrophoresis in a 3.5% polyacrylamide slab gel containing 8 M urea, and identified by autoradiography as described under "Methods." a, A-DNA, no Cro protein; b, A-DNA, Cro protein added at 3 pg/ml; c, Aimm434-DNA, Cro protein added at 3 pg/ml. a RNA polymerase was incubated with A-DNA at 17" for 20 min before addition of nucleoside triphosphates and rifampicin (see "Methods").
* Cro protein was added to A-DNA for 10 min at 17" followed by an RNA polymerase incubation as for Line 1. c Cro protein and RNA polymerase were added to A-DNA together followed by an incubation at 17" for 30 min. d RNA polymerase was incubated with A-DNA as for Line 1, and Cro protein was then added for 10 min at 17" before addition of nucleoside triphosphates. radiography ( Fig. 111. The RNA produced in vitro from A-DNA in the presence of p factor is predominantly of four size classes, designated 4 S, 6 S, 8 to 9 S, and 12 S (14,18,32,33).
The 8 to 9 S and 12 S transcripts represent RNA chains initiated at thep, andp, promoters, respectively, and the 4 S and 6 S transcripts represent promoters near the ~11 and Q genes, respectively (Fig. 1). The results of Fig. 11 show that Cro protein represses 8 to 9 S and 12 S RNA synthesis from A-DNA but not from Aimm434-DNA; 6 S RNA synthesis was not repressed by Cro with A-DNA as a template (although not shown at this gel exposure, 4 S RNA was also not repressed by Crol. As expected from the failure of Cro to repress total RNA synthesis from Aimm434-DNA (Fig. lo), there was no difference in the gel pattern of RNA synthesis from Aimm434-DNA in the presence or absence of Cro (data not shown).
If the repression of transcription noted in Fig. 10 and Fig. 11 occurs at the initiation step of RNA synthesis, the binding of