Studies of Nuclear Acidic Proteins EFFECTS

Methods are described for the extraction, separation, and electrophoretic analysis of a class of acidic nuclear proteins from various tissues of the rat. Such proteins occur in the chromatin and their distribution in diverse nuclear types is tissue specific. Many of the chromosomal acidic proteins are phosphoproteins which incorporate 32P-orthophosphate in vivo with the formation of phosphoserine and phosphothreonine residues. The patterns of phosphorylation of individual nuclear proteins vary from one tissue to another. Materials and Methods


Methods
are described for the extraction, separation, and electrophoretic analysis of a class of acidic nuclear proteins from various tissues of the rat.
Such proteins occur in the chromatin and their distribution in diverse nuclear types is tissue specific.
Many of the chromosomal acidic proteins are phosphoproteins which incorporate 32P-orthophosphate in vivo with the formation of phosphoserine and phosphothreonine residues.
The patterns of phosphorylation of individual nuclear proteins vary from one tissue to another.

Materials and Methods
Many of the acidic proteins prepared from liver and kidney nuclei of the rat form complexes with rat liver DNA. Binding to the DNA of a closely related species (mouse) also occurs, but to a lesser extent.
Little or no binding is observed when rat liver phosphoproteins are added under the same conditions to DNA's  Orangeburg, New York), neutralized with 0.2 N NaOH and adjusted to 0.9% with NaCl, was injected intraperitoneally at a dosage of 2 mCi/lOO g of rat body weight.
The animals were killed 90 min later and tissues were removed for isolation of nuclei and preparation of the phosphoprotein fractions.
Nuclear phosphoproteins stimulate transcription in a cellfree RNA-synthesizing system. Correlations are observed between DNA binding and enhancement of RNA synthesis. DNA-protein complexes have been separated by density gradient centrifugation after "annealing" has occurred. The nature of the proteins present in the complex depends on the tissue from which the nuclear proteins were isolated. The experiments to bo described deal with the isolation and characterization of a complex set of acidic proteins associated with DNA in the nuclei of animal tissues. Evidence will be presented to show that nuclear acidic protein fractions prepared from various tissues of the rat differ in their composition and patterns of phosphorylation.
DNA binding by such proteins is selective, and when complex formation occurs, an enhancement of RNA synthesis in vitro can be shown.
Liver nuclei were prepared as follows: the livers from two to four normal or adrenalectomized rats were chilled and finely minced with scissors.
All subsequent steps in the isolation were performed in the cold. The minced tissue was homogenized in 5 volumes of 0.32 M sucrose-3 mM MgClz with a Teflon-to-glass homogenizer of 55-ml capacity, llO-mm grinding length, 44-mm pestle head length, and a clearance of 0.15 to 0.22 mm (Arthur H. Thomas, Inc., Philadelphia, Pennsylvania), rotating at 1560 rpm for 20 strokes.
The homogenate was filtered through double napped flannellette cloth, and cold deionized water was added to the filtrate to reduce the total sucrose concentration to 0.25 M.
Aliquots of the homogenate were taken for determination of total DNA and protein content.
The remainder was transferred to centrifuge tubes and 0.25 volume of 0.32 M sucrose-3 mM MgC& was pipetted into each tube to form a layer beneath the broken cell suspension. The tubes were centrifuged at 1100 x g for 10 min.
The crude nuclear pellet was resuspended in 10 volumes of 2.4 M sucrose-l mM MgC& and the nuclei were sedimented by centrifugation at 100,000 X g for 1 hour. Aliquots of the nuclear pellet were taken for determination of DNA and protein content.
Of the DNA in the original liver homogenate 52y0 was recovered in the purified nuclear fraction (Table I). The urea was removed and the salt concentration was lowered by dialysis against 0.01 BI Tris-HCl, pH 8.0-0.01 M NaCl for 2 hours.
After dialysis, the sample, 0.5 ml, was carefully layered on top of a 4-ml 5 to 25c/, sucrose gradient in 0.01 M Tris-HCl, The tubes were centrifuged at 358,000 x g for 2+ hours.
They were punctured at the base and 0.2-ml fractions were collected.
Each fraction was diluted to 0.4 ml and its optical density at 260 rnp was determined as a measure of its DNA content.
The distribution of 32P labeled protein in the gradient was determined by adding 10 ml of Bray's solution (16) to each fraction and measuring radioactivit>y by scintillation spectrontetry.
For recovery and characterization of the DNA-bound proteins, the scale of the experiments Was increased Gfold. The DNA-protein peak was recovered from the gradient and dialyzed against 0.1 M Tris-HCl, pH 8.4-0.01 M EDTA-0.14 M 2-mcrcaptoethanol (Buffer A) to remove the sucrose. The protein was then re-extracted with 2 ml of phenol saturated with Buffer A, overnight, and processed by the standard isolation procedure prior to analysis by polyacrylamide gel electrophoresis. In some experiments, the phosphoprotein fraction was treated with ribonuclease in an attempt to deterrnine whether binding to DNA could be influenced by residual (but nondetectable) amounts of contaminating chromosomal RNA's (15). which had been previously heated at 60" for 15 min to remove any residual deoxyribonuclease activity. After incubation for 2 hours at 37", the phosphoprotein fraction was "annealed" to 200 pg of rat liver DNA, subjected to gradient dialysis, and fractionated bg sucrose density gradient centrifugation, as described.
Thr peak containing the nucleoprotein complex was isolated and the ratio of protein to DNA was determined.
In the RNase-treated samples, the protein to DNA ratio was 0.114 (13.2 pg of protein per 115 pg of DNA), while the corresponding figure for control proteins incubated and annealed in the absence of ribonuclcase was 0.117, ilrdicating that enzymatic digestion had lit'tle effect on the protein-DNAbinding react,ion. The formation of l~l~osphoprotein-DNB Vol. 246,No. 11 complexes in the presence of RNase, which is a basic protein, also indicates that this amount of enzyme did not block all the reactive regions of the DNA. Other experiments established that ribonuclease treatment did not remove 32P activity from the isolated nuclear phosphoprotein fraction after labeling in viva with 32P-orthophosphate, since no counts were lost when 117 pg of liver nuclear acidic protein was incubated with 20 pg of RNase for 2 hours at 37". Parallel experiments were carried out to assure that ribonuclease is functional in the 2 M NaCl-5 M urea-O.01 M Tris-HCl buffer employed in the first stage of the binding experiments.

Properties of Isolated
Total rat liver RNA, labeled in viva with i4C-erotic acid and isolated by the phenol procedure, was incubated with 50 pg of RNase under the conditions described. Over 88% of the counts were lost in a 2-hour incubation, a result fully equivalent to that observed when the digestion was carried out in 0.1 M phosphate buffer at pH 8.0. It follows that the absence of a ribonuclease effect on the binding reaction is not caused by a loss of activity of the enzyme in the high ionic strength urea buffer.
RNA Polymerase Assay-The effects of nuclear phosphoproteins on transcription were tested in vitro, with the RNA polymerase assay as described by Burgess (17) (17)5 were "primed" with 2 to 4 pg of free DNA or with the equivalent amount of DNA-protein complex.
In some experiments the phosphopro-i&n-DNA complex was isolated by sucrose density gradient centrifugation as described for the DNA-binding experiments. In other tests, the entire protein-DNA mixture was tested after the annealing and dialysis steps but prior to centrifugation.
Triplicate samples were incubated for 10 min at 37". The tubes were then chilled in ice and 1.0 ml of DNA carrier solution (10 pg of DNA per ml) was added, followed by 5 ml of 5yG trichloroacetic acid containing 0.01 M sodium pyrophosphate. After 30 min the precipitates were collected on Millipore filters (0.45 p pore size) and washed with 25 ml of 5% trichloroacetic acid containing 0.01 M sodium pyrophosphate. The filters were dried and transferred to counting vials containing 5 ml of toluene-Liquifluor (New England Nuclear, Inc., Boston, Massachusetts). 3H-activity was measured on a Nuclear Chicago Mark I liquid scintillation spectrometer.

Isolation
and Properties of Nuclear Acidic Proteins-A new method for the preparation of nuclear acidic proteins, many of which have an affinity for rat DNA, has been applied to the liver, kidney, and other tissues of the rat.
The isolation procedure is based on the nuclear localization and characteristic solubility properties of these proteins, which were extracted from nuclei that had been purified by centrifugation through sucrose density barriers to remove whole cells and cytoplasmic contamination.
A differential extraction of nuclear proteins was 5 The E. coli RNA polymerase was the generous gift of Dr. Teh-Sheng Chan and Dr. Peter Model of the Rockefeller University. then carried out, with 0.14 M NaCl to remove 0.14 ELI NaClsoluble components, 0.25 N IICl to remove histones, and phenol to solubilize a class of nuclear phosphoproteins which show strong indications of involvement in the control of chromosomal RNA synthesis.
The use of phenol to solubilize proteins associated with nucleic acids has precedents in the study of plant viruses (18) and bacteriophages (2), and phenol has been used previously to extract proteins with high rates of synthesis from the nuclei of mammalian cells (19,20). The present procedure combines phenol extraction with electrophoretic separations of individual proteins from different nuclear types. It permits the demonstration that many of the acidic proteins of the nucleus are t,issue-specific phosphoproteins.
Some of these combine selectively with DNA and stimulate transcription in an in vitro RNA-synthesizing system.
Localization-The phenol-soluble proteins occur in the nucleus and comprise a quantitatively important fraction of the total nuclear protein.
In liver nuclei, for example, where the ratio of total protein to DNA is 4.2 mg of protein per mg of DNA, the phenol-soluble protein fraction is about 12% of the total protein, and the ratio of phenol-soluble proteins to DNA is at least 0.45: 1.0 (Table I).
Thus, in this metabolically active tissue, there is nearly half as much acidic nuclear protein as there is DNA.
The proteins under consideration appear to be components of the chromosomes rather than the nuclear sap. This is indicated by the fact that the proteins remain associated with the DNA during the isolation of chromatin. Electrophoretic analyses of the acidic proteins prepared from liver chromatin show close resemblances to the corresponding fractions prcpared from intact nuclei (Fig. 1).
Amino Acid Composition-Amino acid analyses of the phenolsoluble protein fractions of liver and kidney nuclei clearly indicate their acidic nature, i.e. the content of aspartic plus glutamic acids greatly exceeds that of the basic amino acids, lysine, arginine, and histidine (Table  JT). Although the extent, to which the acidic amino acids occur in the polypeptide chains as free acids or as the corresponding amides is not known, the electrophoretic mobility of the proteins at neutral pH values indicates that they have isoelectric points well below pll 7.4. Electrophoresis of 32P-labeled nuclear phosphoproteins in $1 gradients which permit isoelectric focusing reveals that most of the radioactive protein bands occur in regions of the gradient with pH values below 7.0.
Heterogeneity and Tissue SpeciJicity-The phenol-soluble protein fraction is complex.
It comprises a mixture of proteins which differ in size, electrophoretic mobility, degree of phosphorylation, and affinity for DNA. The mixture has been resolved into many of its components by electrophoresis in polyacrylamide gels. Electrophoresis in the presence of sodium dodecyl sulfate separates the proteins on the basis of molecular size (2)(3)(4)(5)(6). Subsequent staining with Amido black 1OB reveals the presence of many proteins with a broad range of molecular weights (Fig. 2). Judging by the mobilities of individual bands, as compared with the mobilities of 1251-labeled marker proteins of known molecular weights, most, of the liver nuclear acidic prot,eins have molecular weights above 35,000.
A molecular weight range of 13,000 to 80,000 is conmonly observed under these conditions (20). The number of protein bands and their relative intensities are highly reproducible in the phenol-soluble protein fractions pre-pared from any given tissue, such as the liver. The electrophoretie pattern is not altered if the acidic protein fraction is reextracted in phenol and reisolated. Estimates of the relative amounts of individual components can be made by (a) cutting out the bands and extracting the dye from each gel slice in dimethylsulfoxide (7) (the amount of dye recovered from each band is directly proportional to the protein content of that band,2) and (b) by quantitative densitometry and computer analysis of the densitometer tracing.2*4 The complexity of the phenol-soluble protein fraction of rat liver nuclei is indicated by the multiple banding pattern shown in Fig. 2; the presence of over 26 components is indicated. A densitometer tracing of the gel photograph is aligned with the banding pattern and shown in the upper part of the figure. This curve has been resolved into 38 components by computer analysis.
The banding patterns of the nuclear acidic proteins vary from one tissue to another. This is clearly indicated by a comparison of the stained gels and densitometer tracings of preparations of liver ( Fig. 2A) and kidney acidic proteins (Fig. 2B). Smaller differences are evident in electrophoretic analyses of the phenolsoluble proteins of spleen nuclei (Fig. 3A) and brain nuclei (Fig. 3B). Thus, tissue specificity, which should be one of the characteristics of a class of DNA-associated proteins concerned with the control of transcription in differentiated cells, is observed in the phenol-soluble protein fraction. The tissue specificity of the nuclear acidic proteins is also indicated by differences in their rates of phosphorylation in vivo.
Fro. 1. Comparison of electrophoretic banding patterns of acidic proteins extracted from intact rat liver nuclei and from isolated rat liver chromatin. SDS complexes of the phenolsoluble proteins were fractionated by electrophoresis in 10% polyacrylamide gels at pH 7.4, as described under "Materials and Methods." The gels were stained with Amido black 10B to indicate the positions and relative concentrations of individual protein bands. The similarities in banding patterns indicate a ohromosomal origin for many of the proteins in the whole nuclear extract.
Phosphmylation-The presence of phosphate in the acidic nuclear protein fraction has been exhibited in three ways: (a) by analysis of the phenol-soluble proteins for alkali-labile phosphate content, (b) by acid hydrolysis of the protein followed by chromatographic separations of phosphoserine and phosphothreonine, and (c) by measuring the incorporation of 32P-labeled orthophosphate in vivo into individual phosphoproteins of rat liver and kidney nuclei.
Phosphate esterified to serine and threonine residues in the nuclear phosphoproteins was released by a brief hydrolysis in 1 N NaOH at 100". The resulting inorganic phosphate was measured by calorimetric analysis of the phosphomolybdate complex after its extraction in isobutyl alcohol-benzene (10). By this test, the average phosphorus content of the phenol-soluble protein fraction is 0.94$$ by weight, a figure which is to be compared with earlier estimates of 1.14% phosphorus in rat liver nuclear phosphoproteins prepared by a combination of salt extraction and chromatographic purification (21).
The presence of the phosphorylated amino acids, phosphoserine and phosphothreonine, was established by hydrolysis of the a2P-labeled protein from liver nuclei and chromatographic separation of the isotopically labeled amino acid esters, using the method of Schaffer et al. (11) to separate phosphoserine from phosphothreonine and from inorganic phosphate. Phosphoserine is the major phosphorylated amino acid in the phenolsoluble protein fraction. 32P-labeled phosphothreonine is also present.
The incorporation and release of phosphate by nuclear phosphoproteins i n vivo is a major aspect of their metabolism (10, 21-29) .2 The present studies verify earlier observations on extensive 32P-orthophosphate uptake into nuclear acidic proteins, and extend them to a consideration of 3zP incorporation into individual proteins of liver and kidney nuclei. Rats were injected with a2P-orthophosphate and the nuclei of liver and kidney were isolated 90 min later. The phenol-soluble protein fraction was prepared autl separated into its components by polyacrylamide gel electrophorcsis. The 32P activity in each baud was measured by slicing the gel transversely and counting individual bands. Clear differences are evident in the extent of phosphorylat~ion of different prot>eins from the same tissue (Fig. 4A).
The pattern of phosphorylation of kidney and liver proteins also appears to be tissue specific (compare Fig.  4, .I and B).
A complication in such studies is the fact that the rate of phosphorylation of individual proteins in liver nuclei is subject to hormonal regulation. I Tydrocortisone, for example, iuflucnces both the rate of synthesis (20) and the kinetics of 32P incorporation into individual proteins of the phenol extract of liver nuclei. 2 In any case, the presence of isotopic phosphate in so many of the nuclear proteins of liver and kidney cells iutlicates both the extreme complrsity and the high metabolic activity of the phosphoprotein fraction. .4 number of t,ests have been carried out to ensure that the a21' activity observed is not the result of contamination by nucleic acids or phospholipids.
When rats were injected with erotic acid-6r4C (15 mCi per rat; specific activity 36.5 mCi per mmole) as a marker for newly synthesized nucleic acids, the nuclear RNS of the liver had a specific activity of over 130,000 cpm per mg after 30 min.
The total phosphoprotein fraction had only 20 cpm per mg, and no 1% activity could be detected in any of the phosphoprotein bands in the polyacrylamide gels. This is in accord with the expectation that contamination by ribonucleic acids would be minimal, because RNA's would be espect,ed to remain in the aqueous phase when the nuclear acidic proteins are extracted in phenol. Similarly, contamination by lipids is minimized by extracting t,he nuclei with chloroform-methanoLHC1 before dissolving the phosphoproteins in phenol. The effective removal of lipids by this procedure was tested by injecting rats with glycerol-2-VT (250 MC1 per rat; specific activity 200 mCi per mmole) and measuring incorporation after 3 hours. l'hc total activity in the lipids extracted by chloroform-methanol-HCl was 9900 cpm, while the total activity in the phenol-soluble fract,ion of liver nuclei was only 3';/, of that figure.
Thus, it is unlikely that incorporation of radioactive phosphate into the phenol-soluble fraction represents contamination of the phosphoproteins by ribonucleic acids or by phospholipids.
Sekctive DNA-binding by Nuclear Phosphoproteins-There are several indications that interactions between nuclear phosphoproteins and DNA are highly selective. Under the proper conditions, the formation of complexes between phosphoproteins aud DNA is surprisingly species dependent. The specificity of the DNA-binding reaction has been studied with prepara-tions of 32P-labeled phosphoproteins from rat liver and kidney nuclei.
The procedure is based on a slow annealing reaction in which samples of DNA and protein, each dissolved in 2 M NaCl-5 M urea-O.01 M Tris-HCl, pH 8.0, were mixed and dialyzed together against a progression of salt solutions of decreasing concentration.
After removal of the urea by dialysis against 0.01 M Tris-HCl, pH 8.0-0.01 M NaCl, the samples were layered over a 5 to 25y0 buffered sucrose gradient and centrifuged.
The distribution of DNA in the gradient was determined by its absorption at, 260 rnp, while the position of the phosphoproteins was indicated by 32P radioactivity. A binding of proteins to DNA is accompanied by a shift of 32P activity downward into denser regions of the gradient.
Proteins which are not bound to DNA remain in the light, upper portion of the gradient. The application of this procedure for the separat,ion of DNAprotein complexes is shown for rat liver phosphoproteins and rat liver DNA in Fig. 5A. The binding reaction is dependent upon salt concentration. At low ionic strengths (e.g. 0.01 M NaCl) the cosedimentation of 32P-labeled phosphoprotein with DNA is evident (Fig. 5A), but binding is not observed at 0.05 M NaCl ( Fig. 5B) nor at 0.15 M NaCl (Fig. 5C). At low ionic strengths, phosphoproteins from rat kidney nuclei also combine with DNA isolated from rat liver (Fig. 6A).
This type of association appears to be relatively selective for the DNA of the species from which the phosphoprotein fraction was prepared.
No comparable binding of the phosphoproteins of rat kidney nuclei to calf thymus DNA (Fig. 6B) or to pneumococcal DNA (Fig. 6C) was observed. Similar restrictions apply to the binding of rat liver phosphoproteins.
Combination with rat DNA is readily indicated (Fig. 5A), but neither calf thymus DNA (Fig. 7A) nor pneumococcal DNA (Fig. 7B) form soluble complexes with the 32P-labeled proteins of rat liver nuclei.
DNA-binding studies have been extended to DNA's prepared from human placenta, and from the livers of the dog, mouse, chicken, and duck.
No appreciable binding of rat liver phosphoproteins to human DNA or dog DNA was observed, but binding of rat liver phosphoproteins to mouse DNA does occur. However, the extent of complex formation between the rat nuclear acidic protein and the DNA of this closely related species is less than that seen when both the protein and the DNA come from the rat.
For example, the protein to DNA ratio in the peak region of the gradient was 0.11 pg of protein per 1 pg of DNA for rat DNA and 0.08 pg of protein per 1 pg of mouse DNA tested under identical conditions. The basis for selectivity in phosphoprotein-DNA interactions FIG. 2 (top). Electrophoretic separation of acidic proteins from rat liver and kidnev nuclei.
Sodium dodecvl sulfate comnlexes of the phenol-soluble proteins were fractionated by electropdoresis in 10% polyacrylamide gels at pH 7.4, as described under "Materials and Methods." The gels were stained with Amido black IOU IO indicat.e the positions and relative concentrations of individual protein bapds.
A densitometer tracing of the gel photograph is aligned with t,he banding patt,ern in the upper part of each figure.

(cenfer).
Electrophoretic separations of acidic proteins from rat spleen and brain nuclei.
SDS complexes of the phenolsoluble proteins were fractionated by electrophoresis in 10% polyacrylamide gels at pH 7.4, as described under "Materials and Met.hods." The gels were stained with Amido black 10B to indicate t,he positions and relat,ive concentrations of individual protein bands.
A densitometer tracing of the gel photograph is aligned with the banding pattern in the upper part of each figure. A, rat spleen nuclear phosphoprotein fraction; B, rat brain nuclear phosphoprotein fraction. FIG. 4 (bottom).
Radioactive orthophosphate incorporation into individual components of the nuclear acidic protein fractions of rat liver and kidney nuclei.
After labeling in vivo with 32Porthophosphate, the proteins were extracted and separated as SDS complexes in oolvacrvlamide gels. The nosition of the individual bands after ele&ophore& is shown by staining with Amido black 10B. Radioactivity in each band was determined by transverse sectioning of the gel and counting individual l-mm slices. The radioactivity is plotted as a function of the distance of migration in the upper part of each figure. This is aligned with the banding pattern in the lower part of each figure.
A, a%P distribution in rat liver nuclear phosphoproteins; B, 32P distribution in rat kidney nuclear phosphoproteins.  ---).
Not,e the formawas isolated and "annealed" with various DNAs as described tion of complexes between rat kidney phosphoproteins and rat under "Materials and Methods." Each mixture was layered over liver DNA (A), but not with the DNA of calf thymus (R) or with a 5 to 25% sucrose gradient and centrifuged at 358,000 X 9 for pneumococcal DNA (C).
is not understood, and a few exceptions have been noted, such It should be noted that the differences in binding efficiency as interactions between avian liver DNB's and rat liver phos-that we have observed are not likely to be the result of differences phoproteins.
However, the binding affinities of rat nuclear in mode of preparation or molecular size of the different 1)NA acidic proteins for DNA's of widely divergent species are usually samples. Dog, human, and rat DNA's were isolated by the minimal.
Independent evidence for species specificity in same procedure, but only the latter combines with rat nuclear DNA binding is provided by the recent work of Kleinsmith, phosphoproteins.
Both the rat liver DNA and the pneumo-Heidema, and Carroll (30) who used DNA-cellulose column coccal DNA used in these studies were found to have a comchromatography to show that a fraction of the nonhistone parable range of sedimentation coefficients, but no binding to the proteins from rat liver chromatin binds to rat DNA but not to bacterial DNA was observed. We conclude, tentatively, t,hat DNA's from salmon sperm or from E. coli. In this respect, the acidic proteins in the phenol-soluble fraction of the mammalian cell nucleus would resemble other proteins for which specific DNA affinities have been shown, such as the Zac operon repressor (31, 32), the C, gene repressor of the bacteriophage X (33), and RNA polymerase (35).
The question arises as to whether all or only a fraction of the nuclear phosphoproteins combine selectively with DNA in this manner.
The problem was investigated by recovering the proteins from the DNA-protein complex.
The DNA-protein peak was selected from the gradient and re-extracted with phenol. After the usual isolation procedure, the phenol-soluble proteins were analyzed by polyacrylamide gel electrophoresis. The results are shown in Fig. 8A, which compares the banding pattern of the DNh-bound protein fraction with that of the total phenol-soluble proteins of rat liver nuclei. It is clear that many of the liver nuclear acidic proteins have the capacity to combine with DNA to form a soluble complex under these conditions.
The absence of some bands from the DNA-protein complex recovered from the gradient shows that some of the acidic proteins of liver nuclei do not combine as effectively as do others with the DNA of the species of origin.
It is not known whether their failure to do so indicates that they have other functions, or whether it reflects changes in molecular configuration of the proteins during the rigorous isolation procedure. A similar specificity in DNA binding is evident in the analysis of the soluble DNA-protein complex isolated after the annealing of rat kidney nuclear proteins to rat liver D?JA (Fig. 8B).
In this case, one of the major components of the phenol-soluble fraction (the fastest moving band in the gel electrophoretic pattern) binds to DNA. The stoichiometry of the interaction between nuclear phosphoproteins and the appropriate DNA depends on the conditions of annealing and on the ionic strength during centrifugation. Under the conditions described, the reconstituted acidic protein-DNA complex of rat kidney has a protein to DNA ratio of 0.13: 1 (14.9 pg of protein and 115 pg of DNA in the peak region of the gradient).
About 13% of the total 32P-labeled protein applied to the sucrose gradient appears in the peak, and 8% of the protein originally added to the annealing mixture is recovered by re-extraction of the DNAprotein complex with phenol. The comparable figures for reconstituted liver phosphoprotein-DNA complexes indicate a protein to DNA ratio of 0.114: 1, and binding of 13% of the total protein added.
Considering the possibilities for denaturation of the proteins, and the limitations of the reconstitution experiments, it, is likely that these figures are deceptively low. It has been suggested that the regulation of transcription in the cells of higher organisms involves certain low molecular weight chromosomal RNA's which, by combining with histones or other nuclear proteins, permit a specific interaction with the DNA template (e.g. 15, 36, 37). Accordingly, we have tested to see whether the interaction between nuclear phosphoproteins and DNA requires the participation of RNA.
To date, the evidence is negative on three counts: (a) no contamination of the nuclear phosphoproteins by radioactive RNA was detected in labeling experiments with erotic acid-6-W, Obviously, the presence of small amounts of the enzyme, which is a basic protein, did not block all the acidic protein-combining sites of the appropriate DNA.) Although these results do not rule out the possible presence of very small amounts of nonradioactive RNA which may be shielded from nuclease action, there is no compelling reason to invoke an RNA-DNA hybridization mechanism to account for the selective binding of the nuclear phosphoproteins to the DNA of the species of origin.
It seems more probable that the specificitv resirlrs in t)he Properties of Isolated Nuclear Acidic Proteins Vol. 246,No. 11 FIG. 8. Evidence for the heterogeneity of DNA-binding acidic proteins from rat liver and kidney nuclei.
The phenol-soluble fraction was prepared and "annealed" with rat DNA as described under "Materials and Methods." The DNA-protein complex was separated by sucrose density-gradient cent.rifugation and the protein bound to DNA was released by extraction in phenol.
amino acid sequence and structural conformation of the polypeptide chains, as it does for the specific DNA binding of the Zuc repressor (31, 32) and the Cl gene repressor of the bacteriophage X (33). E$ects on Transcription-The binding studies show that many of the acidic proteins of rat liver and kidney nuclei have the capacity to combine with DNA in vitro, and that this combination is selective with respect to both the DNA and the protein components of the complex. Evidence that phosphoproteins are closely associated with DNA in tivo is provided by the presence of many of the proteins in isolated chromatin (Fig. l), and by autoradiographic studies by Benjamin and Goodman (38) which show a localization of azP-labeled proteins along the salivary gland chromosomes of Sciura larvae.
The functional significance of the binding of acidic nuclear proteins to DNA has been investigated in a cell-free RNAsynthesizing system. Reaction mixtures containing phosphoproteins from rat liver or rat kidney nuclei and DNA's from different sources were incubated in the presence of the four ribonucleoside triphosphates and the soluble RNA polymerase enzyme of E. coli (17). The utilization of aH-UTP was taken as a measure of RNA synthesis.
The results, summarized in centrifugation. For example, RNA synthesis directed by the soluble acidic protein-DNA complex of the liver is 78 to 90% higher than that observed with rat liver DNA alone. Similarly, the kidney acidic protein-DNA complex is 72% more active in transcription than is the corresponding amount of free rat DNA. The stimulation of RNA synthesis by nuclear acidic proteins is also observed in unfractionated mixtures of the phenol-soluble proteins and rat DNA, but the increase is less than that seen with ,gradient-isolated complexes (Table III).
This suggests that some components of the mixture may act to suppress transcription.
An important point is that stimulation of transcription does not occur for all DNA's. Mixtures of rat liver nuclear phosphoproteins and calf thymus DNA are not more active than thymus DNA alone (Table III).
These results support the DNA-binding studies shown in Fig. 6B and 7A which indicate that soluble nucleoprotein complexes are not formed when rat nuclear acidic proteins are added to bovine DNA. Calf thymus DNA recovered from the peak region of the gradient after admixture with rat nuclear proteins is not a better template for RNA synthesis than the original DNA preparation (Table III).
Thus, the interactions between nuclear acidic proteins and DNA's are selective. When complex formation occurs, as it does for rat nuclear proteins and rat DNA, transcription is enhanced. When complex formation fails to occur, as in the case of calf thymus DNA and rat liver phosphoproteins, They are present in high concentrations in the chromatin of metabolically active tissues (1,(50)(51)(52)(53)(54), and they are preferentially localized in those regions of the chromatin that are most active in RNA synthesis (21,55,56). Since many of the acidic proteins of the nucleus are phosphoproteins (10, 21-29) whose rates of phosphorylation are increased at times of gene activation (29) and which are relatively deficient in cell types that are not actively engaged in RNA synthesis (57), one may surmise that such proteins are likely to be involved in posilive control mechanisms, rather than in the suppression of genetic activity.

RNA
Evidence that acidic nuclear proteins may counteract the inhibition of RNA synthesis by histones has been obtained repeatedly in many laboratories (e.g., 21, 24, 5.5, 58-60). Studies of RNA synthesis by reconstituted chromatin fractions also indicate that selectivity in transcription is influenced by the nonhistone proteins added to the mixture (61,62). Our present findings show that the acidic protein fraction includes tissuespecific components which are highly selective in their interactions with DNA and which increase rates of RNA synthesis from the appropriate DNA templates. We propose that their function is to regulate transcription.
The differentiated cells of higher organisms synthesize different populations of messenger RNA molecules, and they would be expected to differ in their contents of proteins which control this aspect of chromosomal activity.
The varied electrophoretie patterns of the phenol-soluble nuclear proteins from different tissues indicate that each differentiated cell type does contain a specific population of DNA-associated nonhintone proteins.
Differences exist both in the nature and the relative concentrations of individual protein bands in the acrylamide gel pattern (63).
On the other hand, all cells synthesize certain RKA's in common, such as ribosomal, 5S, and aminoacyl t,ransfer RXA's, and one would expect to find a broad distribution of proteins concerned with the transciption of the more ubiquitous RSA t,ypes. Comparisons of the acrylamide gel patterns indicate that similar phosphoproteins exist in diverse nuclear types. The similarities in nonhistone chromosomal proteins have been noted and emphasized by other investigators (e.g. 64). They can be interpreted in terms of common structural proteins and enzymes concerned with nucleic acid biosynthesis and histone metabolism.
It is not yet known whether any of the components of the phenol-soluble nuclear protein fraction are enzymatically active in RNA synthesis or RNA methylntion, or in the modification of histone structure by acetylation, methylation, or phosphorylation. Clarification of this problem will require isolation and tests for function of individual components of the acidic nuclear protein fraction (assuming that enzymatic activity will not be irreversibly lost during the isolation procedure).
The isolation and properties of two nuclear proteins, one from liver, and one from kidney, will be described shortly.
A major aspect of the metabolism of many of the nuclear acidic proteins is the incorporation of 32P-orthophospl~ate into phosphoserine and phosphothreonine residues in the polpept,ide chains.
ATP is the phosphoryl group donor in this esterification reaction (10, 21, 27). The phosphorylation isenzymatic, it occurs in the nucleus, and it takes place after protein synthesis has occurred; i.e. 32P-phosphate uptake into nuclear phosphoproteins is not affected when 14C-serine incorporation is completely blocked by puromycin (10). The significance of the phosphorylation of proteins associated with DKA in the chromatin has yet to be determined, but recent studies relate phosphorylation to RNA synthesis. The pattern of phosphorylnt'ion of individual acidic proteins of the liver nucleus is altered in a complex way when hormones, such as cortisol, stimulate RNA synthesis in the liver.
The changes in phosphate uptake are rapid, specific, and not the result of fluctuations in ATP pool sizes.l Taken together with earlier evidence that nuclear phosphoprotein metabolism is greatly accelerated during gene activation in lymphocytes (29, 65), the present findings in liver and kidney nuclei support the view that phosphorg-lation of the Drl'A -bincliny Propedies of Isolated Nuclear Acidic Proteins Vol. 246,h?o. 11 l>NA-associated proteins in chromatin is somehow related to the specific regulation of RNA synthesis. An important point is that phosphorylation is reversible, and that the turnover of phosphate groups is also increased at times of gene activation (29). The reversibility of the phosphorylation reaction in viva suggests a cyclical binding and release mechanism for the attachment of control proteins to DNA templates. Conclusions-The nuclear phosphoproteins have many characteristics of proteins concerned with the specific control of genetic activity in the cells of higher organisms. They are localized in active regions of the chromatin.
Their distribution in diverse cell types is tissue specific.
Their activity, particularly with regard to phosphate group turnover on the hydroxylated amino acids, reflects the RNA synthetic capacity of the cell. They combine selectively with the DNA of the appropriate species. Transcription from the DN&phosphoprotein complex, once formed, exceeds that from DNA alone.
We conclude that the nuclear phosphoprotein fraction includes components which are involved in the positive control of RNA synthesis. It is likely that such components recognize and combine with specific polynucleotide sequences to promote transcription at particular gene loci. Phosphorylation seems to be a critical variable in the interaction between the phosphoproteins, DNA, and the RNA polymerase.
In many respects, the nuclear phosphoprotein control system is analogous to the u factor control of RNA polymerase activity in bacterial systems (66)(67)(68)(69)(70). The u factors appear to confer specificity on the polymerase by facilitating its attachment to specific initiation or "promoter" sites on the DNL4.
After initiation, the u factor is released from the enzyme and it can participate in a new round of initiation. Phosphorylation of E. coli u factors by protein kinases from animal tissues has been found to stimulate RNA synthesis (71). The u-phosphorylation reaction is itself stimulated by cyclic AMP (71). The resemblance of the u-phosphorylation system, particularly with respect to hormonal (cyclic AMP) response, to the phosphorylation mechanisms known to exist for chromosomal proteins in higher cells is strikingly close. Further developments in this area promise to reveal protein phosphorylation as a general mechanism for influencing the interaction between DNA and proteins which regulate transcription in prokaryotic and eukaryotic cells.