Volume 6 Number 10 1979
Volume 6 Number 10 1979
Nucleic Acids Research
Nucleic Acids Research
DNA-binding activity of tightly-bound nonhistone chromosomal proteins in chicken liver chromatin David M.Gates and Isaac Bekhor* Department of Basic Sciences, School of Dentistry, Laboratory for Developmental Biology, Ethel Percy Andrus Gerontology Center, University of Southern California, Los Angeles, CA 90007, USA Received 16 April 1979 ABSTRACT We have isolated a nonhistone chromosomal protein fraction from chicken liver chromatin which possesses high affinity and preferential sequence DNA binding. Residually DNA-bound nonhistone chromosomal proteins after 2.0 M NaCl extraction of bulk chromatin are isolated. Bound proteins are released by dissociation of the complexes in 5.0 M urea/3.0 M NaCl. We have investigated the in vitro DNA-binding properties of this class. In contrast to other DNA-binding NHCP whose activities have been studied, direct DNA-binding activity is observed which is not abolished under conditions of high ionic strength (to 3.0 M NaCl). Strong preference in binding fractionated homologous DNA is observed, while binding of heterologous (E. Coli) DNA is negligible. The fractionation of homologous DNA permits the isolation of DNA for which this protein class displays strong binding preference, presumably through a concentration of binding sites. The composite data suggest sequence-specific interaction between this protein class and DNA, which is not abolished by high ionic strength.
INTRODUCTION Because of various lines of evidence suggesting that nonhistone chromosomal proteins (NHCP1) play an important role in eukaryotic gene regulation (2-5), there has been considerable interest in recent years in the DNA-binding relationships of these proteins. DNA-binding proteins which display high affinity binding to control sequences, thereby controlling the genetic readout of contiguous gene sequences are well known for several prokaryotic operons (6-9). A number of different eukaryotic DNA-binding NHCP have been studied. Thomas and Patel (10) found a DNA-binding NHCP from rat liver chromatin which preferentially bound denatured DNA. Sevall et al. (11) reported that a small fraction of rat liver NHCP could be isolated by tandem column chromatography which showed preference in binding rat DNA. Weideli et al. (12) found a DNAbinding protein from unfertilized Drosophila eggs displaying a specific affinity for a defined Drosophila DNA sequence, amplified using recombinant DNA technology. Jagodzinski et al. (13) found that a DNA-binding NHCP fraction
C) Information Retrieval Limited 1 Falconberg Court London Wl V 5FG England
3411
Nucleic Acids Research showed preference in binding middle repetitive sequences in the rat genome. Lesser and Comings (14) reported a NHCP fraction from mouse liver chromatin which showed preference in binding mouse over E. Coli DNA. Recently, Hsieh and Brutlag (15) reported the isolation and partial characterization of a protein from Drosophila embryos which binds in sequence-specific manner to one of the four Drosophila satellite DNAs. This laboratory has been interested in the DNA-binding activities of the tightly-bound nonhistone chromosomal proteins. We designate this class M3, since it can be dissociated from chromatin DNA in our medium three (5.0 M urea - 3.0 M NaCl). Previous work from this laboratory demonstrated that this class may possess a direct DNA-binding activity, without a requirement for the presence of histones (16). Further, this class caused a marked enhancement in the transcriptional capacity of DNA complexed with histones (17). In more recent experimentation, it was demonstrated that the tightly-bound NHCP from chicken reticulocytes markedly concentrate globin gene sequences, but those from liver do not(18). Through the study of reassociation kinetics of those DNA sequences which remain protein-bound after 2.0 M NaCl extraction (which removes histones as well as the vast majority of nonhistones), we found (19) that the tightly-bound NHCP in chicken reticulocyte chromatin are nonrandomly distributed on chromatin DNA, and appear to be exclusively associated with DNA of unique sequence. The composite data suggest that this protein fraction may be capable of DNA sequence recognition, and as such components thereof could be viable candidates for participation in the control of specific gene function In this report, we describe the DNA-binding activity of the tightly-bound NHCP from chicken liver chromatin. Briefly, the experimental approach is to fractionate chromosomal proteins and chromatin DNA by extraction of bulk chromatin with 2.0 M NaCl. After high speed centrifugation, two DNA fractions are recovered: one virtually protein free; the other displaying a relatively high protein/DNA ratio. With subsequent processing, unbound DNA and residual DNAprotein complexes can be separated. The details of this fractionation have been published elsewhere (18,19). Residual protein (the tightly-bound NHCP) can be extracted from the isolated DNA-protein complexes by dissociation in 5.0 M urea - 3.0 M NaCl, followed by dialysis to remove the urea and NaCl. We find that the tightly-bound NHCP isolated in this manner have some very interesting properties. As assayed by nitrocellulose filter binding of proteinDNA complexes (20), and under the proper conditions, we find that (1) these proteins display strong preference for the bound homologous fraction (DNA-P) over homologous unbound DNA (DNA-S), with negligible binding to heterologous 3412
Nucleic Acids Research (E. Coli) DNA, (2) the binding is saturable, (3) the binding reaction is not abolished under conditions of high ionic strength (to 3.0 M NaCl); rather, increasing ionic strength slightly enhances the binding reaction, and (4) the binding reaction displays defined pH and temperature optima. Taken together, the data suggest that this NHCP class may be capable of sequence-specific interaction with homologous DNA, and this sequence recognition is apparently unperturbed by high ionic strength. The observation of DNA binding under conditions of high ionic strength suggest that nonionic forces (including hydrophobic interaction) may be significant. Further, the results suggest that DNA fractionation, resulting in a concentration of putative NHCP binding sites, as described in this report, may be useful in the study of specific protein-DNA interactions in chromatin. MATERIALS AND METHODS Preparation of Chromatin: Chicken liver chromatin was prepared by the procedure of Hymer and Kuff (22) as modified by Simpson (21), with some changes Livers were excised from freshly sacrificed adult chickens and immediately placed on ice. The livers were minced with scissors and placed in grinding buffer (0.25 M sucrose, 1.5 mM MgCl2, 10 mM Tris-HCl, pH 8.0). This suspension was mixed in the blender at low speed for 3 min., until no macroscopic particles were evident. The suspension was filtered through 3 layers of cheesecloth, and cells were pelleted by centrifugation at 2,000 x g for 10 min. in the Sorvall SS-34 rotor. Cells were resuspended in grinding buffer, mixed for 2 min. in the blender at low speed, and Triton X-100 was added to 1% final concentration. Mixing was resumed for 2 min. Nuclei were pelleted by centrifugation at 12,000 x g for 10 min. in the Sorvall SS-34. Nuclei were washed three times in grinding buffer without Triton X-100. After the final wash, nuclei were briefly allowed to swell in deionized water, and collected by centrifugation as before. Nuclei were then suspended in TPD1, homogenized in the Dounce homogenizer with tight-fitting ("A") pestle, and the composite of unlysed nuclei and crude chromatin was pelleted by centrifugation at 12,000 x g for 10 min. in the SS-34 rotor. The TPD homogenization was repeated twice, until the crude chromatin pellet was free of unlysed nuclei as determined by phase contrast microscopy. The final crude chromatin pellet was resuspended in TPD and sedimented
through 20 ml. cushions of 1.7 M sucrose - TPD in the Beckman SW-27 rotor at 26,000 rpm for 90 min. The chromatin pellet was homogenized into TPD and repelleted at 12,000 x g for 10 min. in the SS-34 rotor. This purified chrom3413
Nucleic Acids Research atin was used for further extraction. Chromatin Extraction and DNA Fractionation: Extraction of chromatin with 2.0 M NaCi and DNA fractionation have been described (18,19). This processing gives rise to 2 DNA fractions: protein-free and protein-bound. Residual DNAprotein complexes were dialyzed overnight in TPD against 40 volumes of TPD and collected by centrifugation at 12,000 x g for 10 min. in the Sorvall SS-34 rotor. Protein Extraction: Residually bound protein after 2.0 M NaCl extraction was dissociated from the isolated protein-DNA complexes by homogenization into 5.0 M urea - 3.0 M NaCl in TPD. After 1 hour in the cold, dissociated protein was separated from DNA by centrifugation for 24 hours in the Beckman SW-6OTi rotor at 60,000 rpm. Extracted proteins were dialyzed against 2 changes of 160 volumes of 0.25 M NaCl - TPD, and were frozen in aliquots at -20° C until use. Protein concentration in storage was 80 ig/ml. Under these conditions, it was determined that the proteins remained in solution. DNA pellets were recovered from the centrifuge tubes, purified, and used for further study. DNA Purification and Nick Translation: All DNAs were purified as previously described (23), and displayed spectral scans characteristic of protein-free DNA. E. Coli DNA was purchased from Calbiochem-Behring, and purified as above. Aliquots of each DNA fraction were taken for use in "nicktranslation." The remainder of the unlabelled DNA fractions were sheared to an average length of 550 base pairs by sonication using a Brownwill sonifier. The DNA fractions were labelled with 3H-dCMP by the "nick-translation" reaction (24), as described (19). Specific activities ranged from 0.6 - 1.5 x 106 cpm/Ig of DNA. Labelled DNAs were frozen in 0.1 x SSCI in aliquots until use.
Nitrocellulose Filter Binding Assays: In vitro DNA binding was performed in polypropylene tubes in a final volume of 1.0 ml. Specific conditions for each experiment are described in the appropriate legend. For binding assays, 25 mm diameter/0.45 um pore size nitrocellulose filters (BA85, Schleicher and Schuell) were presoaked in 0.5 N NaOH to reduce background binding (25). The filters were washed once with deionized water, and once with binding buffer. The filters were mounted on a vacuum manifold, and the 1.0 ml. reaction mixture was filtered under gentle vacuum. The filters were then washed with 2.0 volumes of binding buffer, and solubilized by periodic vortexing in 15 mis. of Filtersolv (Beckman) scintillation cocktail. Compensation was made for filter and chemical quenching, and a quenched sample was taken as 1009 input radioactivity. All samples were counted in a Beckman LS-8100 liquid scintil3414
Nucleic Acids Research lation spectrometer (Beckman Instruments). Other Methods: Protein concentrations were estimated by the dye-binding method of Bradford (26). Protein electrophoresis was performed on 15% polyacrylamide-SDS1 gels using the discontinuous buffer system of Laemmli (27). Stained gels were scanned in a Pye-Unicam spectrophotometer equipped with a linear gel transport and X-Y recorder. RESULTS Chromosomal Protein and DNA Fractionation: Fractionation of purified chicken liver chromatin by 2.0 M NaCl extraction, with subsequent processing as described in Methods, is summarized in Table 1. Most of chromosomal DNA (ca. 96%, designated DNA-S) is rendered virtually protein-free. A smaller fraction (ca. 4%, designated DNA-P) remains protein bound at a high protein: DNA mass ratio. This is similar, but quantitatively different, to the yields in these fractions from chicken reticulocyte chromatin (18,19). Compared to reticulocytes, both the yield in DNA-P and the protein:DNA mass ratio is greater in liver. Since DNA-P was found to be enriched in active globin gene sequences in reticulocytes (18), and it would not be unreasonable to expect a similar enrichment in active gene sequences in liver DNA-P, we suspect that the higher yield in these fractions from liver may be the result of more diverse genetic readout in liver, and possibly different protein organization. That the proteins remaining DNA-bound in the DNA-P complexes are truly resistant to 2.0 M NaCl extraction, rather than simply being incompletely removed on first pass, is indicated in Table 2. To verify that those proteins are truly resistant to 2.0 M NaCl extraction, we reextracted the isolated DNA-
Table 1.
DNA and Protein Fractionation from Chromatin by 2.0 M NaCl
Extraction
Component
Protein:DNA
Ratioa
Yieldb
DNA-Sc
O.o04
96.0
DNA-Pd
2.03
4.0
Total chicken liver chromatin was fractionated as described in Methods. amass:mass, determined as described in Methods. bexpressed in terms of percentage of total DNA recovered. cDNA component rendered soluble after 2.0 M extraction and ddialysis into 10 mM Tris-HCl, pH 8.0, isolated as described in Methods. DNA component remaining insoluble after extraction with 2.0 M NaCl and subsequent dialysis into 10 mM Tris-HCl, pH 8.0, isolated as described in Methods. 3415
Nucleic Acids Research Table 2.
Sequential Protein Extraction from Chromatin.
Manipulation
Protein Releasea
1st 2 M NaCl extraction
95.5
2nd 2 M NaCl extraction
0.4
5.0 M urea/3.0 M NaCl extraction
4.1
Purified chromatin was sequentially extracted as indicated above using the procedure described in Methods. aexpressed in terms of percentage of total protein released. protein complexes with 2.0 M NaCl. The fact that there is negligible protein release during the second 2.0 M NaCl treatment, but that there is release using a combination of urea and NaCl (5.0 M urea - 3.0 M NaCl) indicates that the first NaCl extraction is very efficient in removing those proteins which can be removed by this high ionic strength condition; further, the observation that these proteins are resistant to extraction in high ionic strength but can be removed in urea suggests that these proteins may be capable of interaction with DNA through hydrophobic and/or hydrogen bonding. The apparent absence of a significant ionic strength component in the interaction of these proteins with DNA is a point which will be discussed in greater detail. To examine the proteins released during these extractions, we performed SDS-polyacrylamide disc gel electrophoresis. The scans of the electrophoretic mobilities of the proteins released during the two different extractions are presented in Figure 1. Several points are revealed here: first, the presence of histones in the salt-extracted, but not the tightly-bound fraction is evident; second, the mobilities of the proteins in the two fractions are apparently different; finally, it is clear that both fractions are highly heterogeneous. The densitometric profiles indicate that the two fractions are largely, if not entirely, nonidentical in terms of composition. The apparent molecular weight range of the tightly-bound NHCP is from ca. 14,000 - 200,000, as determined from the mobilities of marker proteins, and since these measurements are made under conditions in which subunits are dissociated, it is possible that these proteins in their "native" state could have substantially higher molecular weights. Three predominant bands appear in the tightlybound fraction: these correspond to molecular weights of approximately 31,600, 68,500, and 190,000 daltons. Numerous other bands appear which are of lesser 3416
Nucleic Acids Research
z
I '0 0-
IAI~~~~~~~~~~~~~~
ui I~~~~~~~~I
LIJ
H1 1E~~~~~~~~~~~~~~1.
DISTANCE MIGRATED (cm.) Figure 1. Densitometric Profiles of Chromosomal Protein Fractions. Approximately 150 'g of each protein fraction was run on 15% polyacrylamideSDS gels as described in Methods. Gels were stained with 1% Coomassie Brilliant Blue and destained in methanol/acetic acid. -Gels after destaining were scanned at 595 nm in a Pye-Unicam spectrophotometer equipped with a linear gel transport and X-Y recorder. Symbols: solid line, proteins extracted from chromatin in 2.0 M NaCl; dashed line, tightly-bound nonhistone chromosomal proteins. The molecular weights determined from the mobilities of marker proteins,run on a parallel gel,are indicated. "H" denotes the mobilities of hi stones.
promi nence. DNA Binding Activities: The first question to be considered was whether or not the tightly-bound NHCP as isolated actually possessed DNA-binding activity, due to the possibility that isolation in urea could cause irreversible loss of protein conformation, resulting in loss of DNA-binding activity. However, the corresponding protein fraction from murine Krebs II chromatin isolated in urea was found to retain DNA-binding activity (16), without a requirement for the presence of histones. Initial assays of binding activity were performed under conditions of low ionic strength (5 mM NaCl), and binding activity was quantified by the retention of labelled "nick-translated" DNA on nitrocellulose filters. These results are presented in Figure 2. DNA retention (and therefore, binding) increased nearly linearly with increasing input protein, but reached an apparent plateau at input protein ratios of 3417
Nucleic Acids Research
a I -
a-
a4=r
micrograms M3 Input
Figure 2. DNA-Binding by Tightly-Bound NHCP Under Conditions of Low lonic 3 Strength. 10 1ig of sheared, unlabelled DNA and 2000 cpm of nick translated HdCMP labelled DNA of the same fraction as tracer were incubated in binding buffer (5 mM NaCl, 1.0 mM MgCl2, in TPD at pH 7.4), with varied input NHCP at 22 C for 1 hour. Reaction mixtures were assayed on nitrocellulose filters as described in Methods. Each point represents the average of triplicate determinations, with a typical deviation of 2.1%. Data has been corrected for a background retention of 1.4% for DNA-P, and 2.3% for DNA-S. Symbols: open circles, DNA-P; closed circles, DNA-S. 1.5 - 2.0:1 (protein:DNA, mass:mass) for the binding of DNA-P (presumably a specific interaction). This was not the case for the binding of DNA-S, for which under the range of input protein employed, apparent saturation was not observed. Some preference was seen under these conditions for the binding of DNA-P over DNA-S (approximately two-fold, at intermediate protein/DNA ratios). This indicated that DNA-binding activity was present, and suggested that sequence recognition might take place. Since it was possible that reaction conditions could have profound effects on protein/DNA recognition, these findings led us to explore the effects of reaction conditions on DNA-protein complex formation. Additionally, an understanding of the effects of reaction conditions could provide insight
3418
Nucleic Acids Research into the mechanism of DNA binding by these proteins. In these studies, DNA-P was used to nmnitor "specific" site association, and E. Coli DNA was used to monitor "nonspecific" site association. If such sites indeed exist in the form of specific DNA sequences, the probability of such sequences being shared between such divergent organisms as chicken and E. Coli is extremely low, for this reason E. Coli DNA should be a valid monitor of nonspecific DNA binding. The formation of DNA-protein complexes as a function of binding buffer ionic strength is presented in Figure 3. This study was conducted at a protein:DNA input mass ratio of 2:1, which appeared to be saturating for the binding of DNA-P under conditions of very low ionic strength. It will be seen later that under more stringent conditions, this is well above the saturation level. From Fig. 3 it is apparent that even under very high ionic strength binding conditions, binding activity is not abolished. When contrasted to
W
E
ZUE I
0
E
a
C
-
z 0
e
_.
NaCi concentration (M)
Figure 3. Effect of Ionic Strength on DNA-Binding Activity. Binding reactions were performed as described in Fig. 2, in ba3ic binding buffer containing 10 pg of sheared, unlabelled DNA, 2000 cpm of H-DNA of the same DNA fraction as tracer, and 20 iig of input protein, at varying NaCl concentrations as indicated. Background retention varied with NaCl concentration, showing a maximal level of 4.2% for DNA-P and 3.6% for E. Coli DNA. Data has been corrected to reflect background retention. Points are an average of triplicate determinations, with a typical deviation of 1.8%. Symbols: open circles, DNA-P; closed circles, E. Coli DNA.
3419
Nucleic Acids Research other DNA-binding proteins studied in eukaryotic systems, this is clearly unusual, as other such proteins have been observed to lose binding activity at higher ionic strength. This would be expected if ionic forces were significant in the formation of DNA-protein bonding. However, in this case, there is an apparent enhancement of binding activity with increasing ionic strength. This apparent enhancement might be explained on the basis of solvent effects on molecular interactions (29). On this basis, it might be expected that a buffer of high ionic strength would limit ionic interactions, but concommitantly favor the formation of nonionic ( and more specifically, hydrophobic) macromolecular interactions. Alternatively, it is known that ionic strength can affect DNA conformation (30), and it is possible that increasing ionic strength might shift DNA conformation to that for which these proteins have a higher affinity. We suspect that the comparatively high level of DNA retention at zero input ionic strength (which in reality is not zero, but is very low) is the result of molecular aggregation. Aggregation does not fully explain this, however, for there is an approximately two-fold greater retention of homologous over heterologous DNA even under these conditions. Therefore, it is probably the case that both "correct" interactions and aggregation are taking place. Aggregation in the form of precipitable protein complexes does not take place at ionic strengths above approximately 0.1 M, as assessed by light scattering (data not shown). At any point along the ionic strength continuum, there is preference for binding homologous over heterologous DNA, and this preference becomes more pronounced at higher ionic strength. It is probable that increased ionic strength is effective in limiting "incorrect" binding, while it is equally apparent that this condition does not affect "correct" binding (i.e., the binding of DNA-P). For this reason, we conclude that ionic forces probably are not significant in sequence recognition. pH Effects: To further characterize the binding reaction, we studied the effect of pH on the reformation of protein-DNA complexes. These results are presented in Figure 4. The tightly-bound NHCP display binding activity over a generally broad pH range, being optimal between pH 7.0 - 7.4, in close aggreement with that seen for another DNA-binding protein from rat liver (28). It is clear that pH exerts an effect on nonspecific binding (i.e., retention of E. Coli DNA), with nonspecific interaction decreasing with increasing pH. At any given point within the pH range tested, however, significant preference is shown for homologous DNA. Temperature Effects: The effect of incubation temperature on the binding 3420
Nucleic Acids Research
30
100
5E zEI,
_OI
f_
0
50
V
c
co
za
is_
O
X 0~~~~~~~~~ 6.0
7.0
8.0
9.0
pH
Fioure 4. Effect of pH on DNA-Binding Activity. Binding reactions were performed as described in Fig. 3, in binding buffer containing 0.25 M NaCl. The pH range of 5.8 - 7.0 was buffered by 0.05 M phosphate buffer, and the range of 7.2 - 9.0 was buffered by 10 mM Tris-HCI. Correction in input NaCl was made to insure equal final Na concentrations under both conditions. Points are an average of triplicate determinations, with a typical deviation of 2.3%. Background retention varied with pH, and was maximal at 8.2% for DNA-P and 6.2% for E. Coli DNA. Data has been corrected to reflect background retention. Symbols: closed circles, DNA-P; open circles, E. Coli DNA.
reaction is shown in Figure 5. Under the conditions employed, optimal binding is displayed at intermediate temperature, between 220 C - 370 C. A 370 C binding optimum is compatible with physiological temperature, and has been observed for a Drosophila DNA-binding protein (15). As tested, binding is decreased below 220 C and above 370 C. However, some measure of homologous DNA binding is retained at the extremes of 40 C and 600 C. At all points, under the ionic strength conditions emnloyed (0.25 M NaCl), binding of E. Coli DNA is negligible. Therefore, temperature differences are largely manifested in homologous (specific) interactions. As temperature may affect protein conformation, and as it is possible that specific protein conformation may be required for correct binding, the temperature differences seen may be the result of subtle conformational changes within the binding proteins, induced by temperature.
3421
Nucleic Acids Research
-
100
30
ZEa 2 ~~ z
o 'aX
~
~
~
~
~
~
5
50 a
0-1
10
20
30
40
DEGREES CELSIUS
50
60
(Oc)
Figure 5. Effect of Incubation Temperature on DNA-Binding. Binding reactions were performed in basic binding buffer containing 10 mM Tris-HCI, pH 7.4, 0.25 M NaCi, at the indicated temperatures. 5ach reaction mixture received 10 ag of sheared, unlabelled DNA, 2000 cpm of H-DNA of the same DNA fraction as tracer, and 20 ig of input NHCP. Points are an average of triplicate determinations, with a typical deviation of 3.2%. Background retention varied with temperature, being maximal at 2.7% for DNA-P and 1.8% for E. Coli DNA. Data are corrected to reflect background retention. Symbols: open circles, DNA-P; closed circles, E. Coli DNA.
Protein Titration: The binding profile at intermediate (0.25 M NaCl) ionic strength with increasing protein input is presented in Figure 6. In this experiment, we monitored the retention of homologous and heterologous DNA. In addition, we monitored the retention of homologous DNA fractionated into bound (DNA-P) and unbound (DNA-S) as described in Methods. We find that there is an approximately 20-fold preference in binding homologous "bound" DNA over heterologous E. Coli DNA. Saturation is observed at a lower level of retention under these conditions than at lower ionic strength as seen in Fig. 2 (22% vs. 32%), and at lower input protein:DNA mass ratios (0.5 - 1.0:1 vs. 1.5 - 2.0:1). Under these conditions, retention of E. Coli DNA is barely above the limit of detection, while retention of homologous "unbound" DNA is approximately two-fold higher. It is possible that the higher retention of 3422
Nucleic Acids Research
a
Z, 1I-c
0
10
30
20 micrograms
M3
40
Input
Figure 6. DNA-Binding with Increasing Input Protein. Binding reactions were carried out in basic binding buffer containing 0.25 M NaCl at pH 7.4 at 220 C, with the indicated input protein concentrations. Points are an average of triplicate determinations with a typical deviation of 5.1%. Background retention was 3.0% for DNA-P, 1.9% for DNA-S, and 1.4,% for E. Coli DNA. Data has been corrected to reflect background retention. Symbols: closed circles, DNA-P; closed squares, DNA-S; Closed triangles, E. Coli DNA. DNA-S compared to E. Coli DNA
may
indicate that
some
protein binding sites
sequences) remain in DNA-S, and therefore a portion of DNA-S retention be the result of correct sequence binding. It is known that substantial sequence differences exist between the two DNA fractions in chicken reticulocytes (19), but site depletion may not be complete. It is clear, however, that the striking preference demonstrated for the binding of DNA-P is strongly suggestive of specific DNA sequence recognition on the part of this protein class. To confirm that differences in binding between the different DNA fractions were not the result of artifacts in the DNA fractions themselves, we compared the binding of DNA-P, DNA-S, and E. Coli DNA with a nonspecific DNA-binding protein (calf thymus histone H1 Sigma). We found (data not shown) that the efficiency in binding histone H1 of these three fractions was similar in magnitude and reached maximal levels at an input ratio of 1.0 (mg H1/mg DNA) in 10 mM Tris-HCl, pH 7.4, 0.25 M NaCl, at room temperature. Further, the probes (DNA
may
3423
Nucleic Acids Research are all essentially duplex as judged by resistance to digestion by nuclease S1. We found that the nick translated DNA was 99.24% resistant to this digestion.
DISCUSSION We have presented evidence herein to suggest that the tightly-bound NHCP as we have isolated them from chicken liver chromatin have some very interesting properties. It is becoming evident that this may be a general phenomenon in many systems. From our own work, we know that the tightly-bound NHCP from chicken reticulocytes significantly concentrate globin gene sequences (18), are nonrandomly distributed on chromatin DNA, and appear to be exclusively bound to unique sequences within the chicken genome (19). The comparable protein fraction from human placenta significantly concentrates the active placental lactogen gene, and in chick oviduct, the active ovalbumin gene is concentrated (Norman and Bekhor, Mirell and Bekhor, unpublished). Studies on HeLa cell metaphase chromosomes indicate that a similar protein class may be responsible for the establishment and/or maintenance of the highly ordered metaphase chromosome structure (31-33). Similarly, nonrandom distribution of these proteins on genomic DNA of mouse L cell metaphase chromosomes has been reported (34). This protein fraction is resistant to dissociation in 2.0 M NaCl and appears to be immunologically cell specific (35). A tightly-bound NHCP or component thereof may be a participating element in the specific conferral of the DNAase I-sensitive structure of the globin gene in chicken erythrocytes (36). Residual DNA-protein complexes after 2.0 M NaCl extraction of rat prostate chromatin interact directly and tissue-specifically with steroidreceptor complexes (37), and this protein class may therefore include acceptor activities. These combined lines of evidence, coupled with the data presented herein, indicate that this protein class may be responsible for a variety of functions within eukaryotic chromatin, and the suggestion of sequence-specific binding activity makes this class a likely candidate to include modulators of specific gene functions, though more experimentation will be required to determine if this is indeed the case. We do not as yet know what fraction of this diverse protein class actually participates in DNA-binding activity, nor do we have precise information regarding the DNA sequences to which they bind. It is likely that amplification involving recombinant DNA will be required to rigorously demonstrate sequencespecificity, as well as achieving the characterization of the binding site(s). As correctly observed (12), binding to a unique or poorly reiterated sequence within eukaryotic DNA is difficult to detect in the absence of DNA fractionation 3424
Nucleic Acids Research or amplification. We suspect that DNA fractionation of the type employed in this report, using the specific protein class to fractionate the DNA, may prove to be useful in the study of specific interactions between the tightly-bound nonhistone chromosomal proteins and DNA. It is probable that when the DNA can be more rigorously fractionated to individually defined sequences, more striking binding preferences will be detected. In summary, the data that we have presented herein indicate that a
protein class from chicken liver chromatin possessing direct DNA-binding activity displays: (1) strongly preferential binding to fractionated homologous DNA, with approximately ten-fold preference for the bound fraction, (2) negligible binding to heterologous E. Coli DNA, (3) defined pH, temperature, and ionic strength optima, and (4) a DNA-binding activity which is not adversely influenced by ionic strength. Taken together, sequence-specific DNAbinding is strongly suggested. This presumed sequence recognition may be related to the control of genetic readout. ACKNOWLEDGEMENTS The authors wish to thank Mrs. Mary Gates for skillful preparation of graphics. DG is supported by Training Grant DE-07006 from the National Institute of Dental Research, USPHS. Research reported herein was supported by NIH grant DE-04031, and in part by private funds. To whom correspondence should be directed REFERENCES 1.
Abbreviations: NHCP, nonhistone chromosomal proteins; TPD, 0.01 M TrisHCI, pH 8.0, 0.1 mM phenylmethane sulfonyl fluoride, 0.2 mM dithiothreitol;
2. 3.
4.
SSC, 0.15 M NaCl, 0.015 M Na Citrate; SDS, sodium dodecyl sulfate. Paul, J., and Gilmour, R.S. ¾1968) J. Mol. Biol. 34, 305-316. Stein, G., Park, W., Thrall, C., Mans, R., and Stein, J. (1975) Nature 257, 764-767. Gilmour, R.S., Windass, J.D., Affara, N., and Paul, L. (1975) Cell Physiol.
5.
85, 449-458. Bekhor, I. (1978) in The Cell Nucleus, Busch, H., ed., Vol. V, pp. 137-166, Academic Press, New York.
6. Wilcox, G., Clementson, K.J., Santi, D.V., and Engelsberg, E. (1971) Proc. Natl. Acad. Sci. USA 68, 2415-2418. 7. Maniatis, T., and Ptashne, M. (1973) Proc. Natl. Acad. Sci. USA 70, 15311535.
8. Riggs, A.D., Newby, R.F., Bourgeois, S., and Cohn, M. (1968) J. Mol. Biol. 34, 365-368. 9. Gilbert, W., and Muller-Hill, B. (1967) Proc. Nat]. Acad. Sci. USA 58, 2415-2421. 3425
Nucleic Acids Research 10. Thomas, T.L., and Patel, G. (1975) Proc. Natl. Acad. Sci. USA 73, 4364-
4368.
11. Sevall, J.S., Cockburn, A., Savage, M., and Bonner, J. (1975) Biochemistry
14, 782-789. 12. Weideli, H., Schedl, P., Artavanis-Tsakonas, S., Steward, R., Yuan, R., and Gehring, W.J. (1977) Cold Spring Harbor Symp. Quant. Biol., vol. XLII, part 2, pp. 693-700. 13. Jagodzinski, L.L., Chilton, J.C., and Sevall, J.S. (1978) Nucl. Acids Res. 5, 1487-1499. 14. Lesser, B.H., and Comings, D.E. (1978) Biochim. et Biophys. Acta 521, 117-125. 15. Hsieh, T., and Brutlag, D. (1979) Proc. Natl. Acad. Sci. USA 76, 726-730. 16. Lapeyre, J.-N., and Bekhor, I. (1976) J. Mol. Biol. 104, 24-58. 17. Bekhor, I., and Samal, B. (1977) Arch. Biochem. Biophys. 179, 537-543. 18. Bekhor, I., and Mirell, C. (1979) Biochemistry 18, 609-616. 19. Gates, D.M., and Bekhor, I. (1979) Nucl. Acids. Res. 6, 1617-1630. 20. Riggs, A.D., Bourgeois, S., and Cohn, M. (1970) J. Mol. Biol. 53,
401-417. 21. Simpson, R.T. (1977) in Methods in Cell Biology, Prescott, D.M., ed., Vol. XVI, pp. 437-446, Academic Press, New York. 22. Hymer, W.C., and Kuff, E.L. (1964) J. Histochem. Cytochem. 12, 359-363. 23. Lapeyre, J.-N., and Bekhor, I. (1974) J. Mol. Biol. 89, 137-162. 24. Rigby, P.W.J., Dieckmann, M., Rhodes, C., and Berg, P. (1977) J. Mol. Biol . 113, 237-251. 25. Riggs, A.D., Reiness, G., and Zubay, G. (1971) Proc. Natl. Acad. Sci. USA 68, 1222-1225. 26. Bradford, N.M. (1976) Anal. Biochem. 72, 248-54. 27. Laemmli, U.K. (1970) Nature 227, 680-685. 28. Catino, J.J., Yeoman, L.C., Mandel, M., and Busch, H. (1978) Biochemistry 17, 983-987. 29. Klotz, I.M., and Shimkana, K. (1968) Arch. Biochem. Biophys. 123,
551-557.
Chan, A., Kilkusie, R., and Hanlon, S. (1979) Biochemistry 18, 85-91. Paulson, J.R., and Laemmli, U.K. (1977) Cell 12, 817-828. Adolph, K.W., Cheng, S., and Laemmli, U.K. (1977) Cell 12 805-816. Adolph, K.W., Cheng, S.M., Paulson, J.R., and Laemmli, U.K. (1977) Proc. Natl. Acad. Sci. USA 74, 4937-4941. 34. Razin, S.V., Mantieva, V.L., and Georgiev, G.P. (1978) Nucl. Acids. Res.
30. 31. 32. 33.
5, 4737-4751.
35. Campbell, A.M., Briggs, R.C., Bird, R.E., and Hnilica, L.S. (1979) Nucl. Acids. Res. 6, 205-218. 36. Weisbrod, S., and Weintraub, H. (1979) Proc. Natl. Acad. Sci. USA 76,
630-634. 37. Wang, T.Y. (1978) Biochim. et Biophys. Acta 518, 81-88.
3426
Volume 6 Number 10 1979
Nucleic Acids Research
Nucleic Acids Research
DNA-binding activity of tightly-bound nonhistone chromosomal proteins in chicken liver chromatin David M.Gates and Isaac Bekhor* Department of Basic Sciences, School of Dentistry, Laboratory for Developmental Biology, Ethel Percy Andrus Gerontology Center, University of Southern California, Los Angeles, CA 90007, USA Received 16 April 1979 ABSTRACT We have isolated a nonhistone chromosomal protein fraction from chicken liver chromatin which possesses high affinity and preferential sequence DNA binding. Residually DNA-bound nonhistone chromosomal proteins after 2.0 M NaCl extraction of bulk chromatin are isolated. Bound proteins are released by dissociation of the complexes in 5.0 M urea/3.0 M NaCl. We have investigated the in vitro DNA-binding properties of this class. In contrast to other DNA-binding NHCP whose activities have been studied, direct DNA-binding activity is observed which is not abolished under conditions of high ionic strength (to 3.0 M NaCl). Strong preference in binding fractionated homologous DNA is observed, while binding of heterologous (E. Coli) DNA is negligible. The fractionation of homologous DNA permits the isolation of DNA for which this protein class displays strong binding preference, presumably through a concentration of binding sites. The composite data suggest sequence-specific interaction between this protein class and DNA, which is not abolished by high ionic strength.
INTRODUCTION Because of various lines of evidence suggesting that nonhistone chromosomal proteins (NHCP1) play an important role in eukaryotic gene regulation (2-5), there has been considerable interest in recent years in the DNA-binding relationships of these proteins. DNA-binding proteins which display high affinity binding to control sequences, thereby controlling the genetic readout of contiguous gene sequences are well known for several prokaryotic operons (6-9). A number of different eukaryotic DNA-binding NHCP have been studied. Thomas and Patel (10) found a DNA-binding NHCP from rat liver chromatin which preferentially bound denatured DNA. Sevall et al. (11) reported that a small fraction of rat liver NHCP could be isolated by tandem column chromatography which showed preference in binding rat DNA. Weideli et al. (12) found a DNAbinding protein from unfertilized Drosophila eggs displaying a specific affinity for a defined Drosophila DNA sequence, amplified using recombinant DNA technology. Jagodzinski et al. (13) found that a DNA-binding NHCP fraction
C) Information Retrieval Limited 1 Falconberg Court London Wl V 5FG England
3411
Nucleic Acids Research showed preference in binding middle repetitive sequences in the rat genome. Lesser and Comings (14) reported a NHCP fraction from mouse liver chromatin which showed preference in binding mouse over E. Coli DNA. Recently, Hsieh and Brutlag (15) reported the isolation and partial characterization of a protein from Drosophila embryos which binds in sequence-specific manner to one of the four Drosophila satellite DNAs. This laboratory has been interested in the DNA-binding activities of the tightly-bound nonhistone chromosomal proteins. We designate this class M3, since it can be dissociated from chromatin DNA in our medium three (5.0 M urea - 3.0 M NaCl). Previous work from this laboratory demonstrated that this class may possess a direct DNA-binding activity, without a requirement for the presence of histones (16). Further, this class caused a marked enhancement in the transcriptional capacity of DNA complexed with histones (17). In more recent experimentation, it was demonstrated that the tightly-bound NHCP from chicken reticulocytes markedly concentrate globin gene sequences, but those from liver do not(18). Through the study of reassociation kinetics of those DNA sequences which remain protein-bound after 2.0 M NaCl extraction (which removes histones as well as the vast majority of nonhistones), we found (19) that the tightly-bound NHCP in chicken reticulocyte chromatin are nonrandomly distributed on chromatin DNA, and appear to be exclusively associated with DNA of unique sequence. The composite data suggest that this protein fraction may be capable of DNA sequence recognition, and as such components thereof could be viable candidates for participation in the control of specific gene function In this report, we describe the DNA-binding activity of the tightly-bound NHCP from chicken liver chromatin. Briefly, the experimental approach is to fractionate chromosomal proteins and chromatin DNA by extraction of bulk chromatin with 2.0 M NaCl. After high speed centrifugation, two DNA fractions are recovered: one virtually protein free; the other displaying a relatively high protein/DNA ratio. With subsequent processing, unbound DNA and residual DNAprotein complexes can be separated. The details of this fractionation have been published elsewhere (18,19). Residual protein (the tightly-bound NHCP) can be extracted from the isolated DNA-protein complexes by dissociation in 5.0 M urea - 3.0 M NaCl, followed by dialysis to remove the urea and NaCl. We find that the tightly-bound NHCP isolated in this manner have some very interesting properties. As assayed by nitrocellulose filter binding of proteinDNA complexes (20), and under the proper conditions, we find that (1) these proteins display strong preference for the bound homologous fraction (DNA-P) over homologous unbound DNA (DNA-S), with negligible binding to heterologous 3412
Nucleic Acids Research (E. Coli) DNA, (2) the binding is saturable, (3) the binding reaction is not abolished under conditions of high ionic strength (to 3.0 M NaCl); rather, increasing ionic strength slightly enhances the binding reaction, and (4) the binding reaction displays defined pH and temperature optima. Taken together, the data suggest that this NHCP class may be capable of sequence-specific interaction with homologous DNA, and this sequence recognition is apparently unperturbed by high ionic strength. The observation of DNA binding under conditions of high ionic strength suggest that nonionic forces (including hydrophobic interaction) may be significant. Further, the results suggest that DNA fractionation, resulting in a concentration of putative NHCP binding sites, as described in this report, may be useful in the study of specific protein-DNA interactions in chromatin. MATERIALS AND METHODS Preparation of Chromatin: Chicken liver chromatin was prepared by the procedure of Hymer and Kuff (22) as modified by Simpson (21), with some changes Livers were excised from freshly sacrificed adult chickens and immediately placed on ice. The livers were minced with scissors and placed in grinding buffer (0.25 M sucrose, 1.5 mM MgCl2, 10 mM Tris-HCl, pH 8.0). This suspension was mixed in the blender at low speed for 3 min., until no macroscopic particles were evident. The suspension was filtered through 3 layers of cheesecloth, and cells were pelleted by centrifugation at 2,000 x g for 10 min. in the Sorvall SS-34 rotor. Cells were resuspended in grinding buffer, mixed for 2 min. in the blender at low speed, and Triton X-100 was added to 1% final concentration. Mixing was resumed for 2 min. Nuclei were pelleted by centrifugation at 12,000 x g for 10 min. in the Sorvall SS-34. Nuclei were washed three times in grinding buffer without Triton X-100. After the final wash, nuclei were briefly allowed to swell in deionized water, and collected by centrifugation as before. Nuclei were then suspended in TPD1, homogenized in the Dounce homogenizer with tight-fitting ("A") pestle, and the composite of unlysed nuclei and crude chromatin was pelleted by centrifugation at 12,000 x g for 10 min. in the SS-34 rotor. The TPD homogenization was repeated twice, until the crude chromatin pellet was free of unlysed nuclei as determined by phase contrast microscopy. The final crude chromatin pellet was resuspended in TPD and sedimented
through 20 ml. cushions of 1.7 M sucrose - TPD in the Beckman SW-27 rotor at 26,000 rpm for 90 min. The chromatin pellet was homogenized into TPD and repelleted at 12,000 x g for 10 min. in the SS-34 rotor. This purified chrom3413
Nucleic Acids Research atin was used for further extraction. Chromatin Extraction and DNA Fractionation: Extraction of chromatin with 2.0 M NaCi and DNA fractionation have been described (18,19). This processing gives rise to 2 DNA fractions: protein-free and protein-bound. Residual DNAprotein complexes were dialyzed overnight in TPD against 40 volumes of TPD and collected by centrifugation at 12,000 x g for 10 min. in the Sorvall SS-34 rotor. Protein Extraction: Residually bound protein after 2.0 M NaCl extraction was dissociated from the isolated protein-DNA complexes by homogenization into 5.0 M urea - 3.0 M NaCl in TPD. After 1 hour in the cold, dissociated protein was separated from DNA by centrifugation for 24 hours in the Beckman SW-6OTi rotor at 60,000 rpm. Extracted proteins were dialyzed against 2 changes of 160 volumes of 0.25 M NaCl - TPD, and were frozen in aliquots at -20° C until use. Protein concentration in storage was 80 ig/ml. Under these conditions, it was determined that the proteins remained in solution. DNA pellets were recovered from the centrifuge tubes, purified, and used for further study. DNA Purification and Nick Translation: All DNAs were purified as previously described (23), and displayed spectral scans characteristic of protein-free DNA. E. Coli DNA was purchased from Calbiochem-Behring, and purified as above. Aliquots of each DNA fraction were taken for use in "nicktranslation." The remainder of the unlabelled DNA fractions were sheared to an average length of 550 base pairs by sonication using a Brownwill sonifier. The DNA fractions were labelled with 3H-dCMP by the "nick-translation" reaction (24), as described (19). Specific activities ranged from 0.6 - 1.5 x 106 cpm/Ig of DNA. Labelled DNAs were frozen in 0.1 x SSCI in aliquots until use.
Nitrocellulose Filter Binding Assays: In vitro DNA binding was performed in polypropylene tubes in a final volume of 1.0 ml. Specific conditions for each experiment are described in the appropriate legend. For binding assays, 25 mm diameter/0.45 um pore size nitrocellulose filters (BA85, Schleicher and Schuell) were presoaked in 0.5 N NaOH to reduce background binding (25). The filters were washed once with deionized water, and once with binding buffer. The filters were mounted on a vacuum manifold, and the 1.0 ml. reaction mixture was filtered under gentle vacuum. The filters were then washed with 2.0 volumes of binding buffer, and solubilized by periodic vortexing in 15 mis. of Filtersolv (Beckman) scintillation cocktail. Compensation was made for filter and chemical quenching, and a quenched sample was taken as 1009 input radioactivity. All samples were counted in a Beckman LS-8100 liquid scintil3414
Nucleic Acids Research lation spectrometer (Beckman Instruments). Other Methods: Protein concentrations were estimated by the dye-binding method of Bradford (26). Protein electrophoresis was performed on 15% polyacrylamide-SDS1 gels using the discontinuous buffer system of Laemmli (27). Stained gels were scanned in a Pye-Unicam spectrophotometer equipped with a linear gel transport and X-Y recorder. RESULTS Chromosomal Protein and DNA Fractionation: Fractionation of purified chicken liver chromatin by 2.0 M NaCl extraction, with subsequent processing as described in Methods, is summarized in Table 1. Most of chromosomal DNA (ca. 96%, designated DNA-S) is rendered virtually protein-free. A smaller fraction (ca. 4%, designated DNA-P) remains protein bound at a high protein: DNA mass ratio. This is similar, but quantitatively different, to the yields in these fractions from chicken reticulocyte chromatin (18,19). Compared to reticulocytes, both the yield in DNA-P and the protein:DNA mass ratio is greater in liver. Since DNA-P was found to be enriched in active globin gene sequences in reticulocytes (18), and it would not be unreasonable to expect a similar enrichment in active gene sequences in liver DNA-P, we suspect that the higher yield in these fractions from liver may be the result of more diverse genetic readout in liver, and possibly different protein organization. That the proteins remaining DNA-bound in the DNA-P complexes are truly resistant to 2.0 M NaCl extraction, rather than simply being incompletely removed on first pass, is indicated in Table 2. To verify that those proteins are truly resistant to 2.0 M NaCl extraction, we reextracted the isolated DNA-
Table 1.
DNA and Protein Fractionation from Chromatin by 2.0 M NaCl
Extraction
Component
Protein:DNA
Ratioa
Yieldb
DNA-Sc
O.o04
96.0
DNA-Pd
2.03
4.0
Total chicken liver chromatin was fractionated as described in Methods. amass:mass, determined as described in Methods. bexpressed in terms of percentage of total DNA recovered. cDNA component rendered soluble after 2.0 M extraction and ddialysis into 10 mM Tris-HCl, pH 8.0, isolated as described in Methods. DNA component remaining insoluble after extraction with 2.0 M NaCl and subsequent dialysis into 10 mM Tris-HCl, pH 8.0, isolated as described in Methods. 3415
Nucleic Acids Research Table 2.
Sequential Protein Extraction from Chromatin.
Manipulation
Protein Releasea
1st 2 M NaCl extraction
95.5
2nd 2 M NaCl extraction
0.4
5.0 M urea/3.0 M NaCl extraction
4.1
Purified chromatin was sequentially extracted as indicated above using the procedure described in Methods. aexpressed in terms of percentage of total protein released. protein complexes with 2.0 M NaCl. The fact that there is negligible protein release during the second 2.0 M NaCl treatment, but that there is release using a combination of urea and NaCl (5.0 M urea - 3.0 M NaCl) indicates that the first NaCl extraction is very efficient in removing those proteins which can be removed by this high ionic strength condition; further, the observation that these proteins are resistant to extraction in high ionic strength but can be removed in urea suggests that these proteins may be capable of interaction with DNA through hydrophobic and/or hydrogen bonding. The apparent absence of a significant ionic strength component in the interaction of these proteins with DNA is a point which will be discussed in greater detail. To examine the proteins released during these extractions, we performed SDS-polyacrylamide disc gel electrophoresis. The scans of the electrophoretic mobilities of the proteins released during the two different extractions are presented in Figure 1. Several points are revealed here: first, the presence of histones in the salt-extracted, but not the tightly-bound fraction is evident; second, the mobilities of the proteins in the two fractions are apparently different; finally, it is clear that both fractions are highly heterogeneous. The densitometric profiles indicate that the two fractions are largely, if not entirely, nonidentical in terms of composition. The apparent molecular weight range of the tightly-bound NHCP is from ca. 14,000 - 200,000, as determined from the mobilities of marker proteins, and since these measurements are made under conditions in which subunits are dissociated, it is possible that these proteins in their "native" state could have substantially higher molecular weights. Three predominant bands appear in the tightlybound fraction: these correspond to molecular weights of approximately 31,600, 68,500, and 190,000 daltons. Numerous other bands appear which are of lesser 3416
Nucleic Acids Research
z
I '0 0-
IAI~~~~~~~~~~~~~~
ui I~~~~~~~~I
LIJ
H1 1E~~~~~~~~~~~~~~1.
DISTANCE MIGRATED (cm.) Figure 1. Densitometric Profiles of Chromosomal Protein Fractions. Approximately 150 'g of each protein fraction was run on 15% polyacrylamideSDS gels as described in Methods. Gels were stained with 1% Coomassie Brilliant Blue and destained in methanol/acetic acid. -Gels after destaining were scanned at 595 nm in a Pye-Unicam spectrophotometer equipped with a linear gel transport and X-Y recorder. Symbols: solid line, proteins extracted from chromatin in 2.0 M NaCl; dashed line, tightly-bound nonhistone chromosomal proteins. The molecular weights determined from the mobilities of marker proteins,run on a parallel gel,are indicated. "H" denotes the mobilities of hi stones.
promi nence. DNA Binding Activities: The first question to be considered was whether or not the tightly-bound NHCP as isolated actually possessed DNA-binding activity, due to the possibility that isolation in urea could cause irreversible loss of protein conformation, resulting in loss of DNA-binding activity. However, the corresponding protein fraction from murine Krebs II chromatin isolated in urea was found to retain DNA-binding activity (16), without a requirement for the presence of histones. Initial assays of binding activity were performed under conditions of low ionic strength (5 mM NaCl), and binding activity was quantified by the retention of labelled "nick-translated" DNA on nitrocellulose filters. These results are presented in Figure 2. DNA retention (and therefore, binding) increased nearly linearly with increasing input protein, but reached an apparent plateau at input protein ratios of 3417
Nucleic Acids Research
a I -
a-
a4=r
micrograms M3 Input
Figure 2. DNA-Binding by Tightly-Bound NHCP Under Conditions of Low lonic 3 Strength. 10 1ig of sheared, unlabelled DNA and 2000 cpm of nick translated HdCMP labelled DNA of the same fraction as tracer were incubated in binding buffer (5 mM NaCl, 1.0 mM MgCl2, in TPD at pH 7.4), with varied input NHCP at 22 C for 1 hour. Reaction mixtures were assayed on nitrocellulose filters as described in Methods. Each point represents the average of triplicate determinations, with a typical deviation of 2.1%. Data has been corrected for a background retention of 1.4% for DNA-P, and 2.3% for DNA-S. Symbols: open circles, DNA-P; closed circles, DNA-S. 1.5 - 2.0:1 (protein:DNA, mass:mass) for the binding of DNA-P (presumably a specific interaction). This was not the case for the binding of DNA-S, for which under the range of input protein employed, apparent saturation was not observed. Some preference was seen under these conditions for the binding of DNA-P over DNA-S (approximately two-fold, at intermediate protein/DNA ratios). This indicated that DNA-binding activity was present, and suggested that sequence recognition might take place. Since it was possible that reaction conditions could have profound effects on protein/DNA recognition, these findings led us to explore the effects of reaction conditions on DNA-protein complex formation. Additionally, an understanding of the effects of reaction conditions could provide insight
3418
Nucleic Acids Research into the mechanism of DNA binding by these proteins. In these studies, DNA-P was used to nmnitor "specific" site association, and E. Coli DNA was used to monitor "nonspecific" site association. If such sites indeed exist in the form of specific DNA sequences, the probability of such sequences being shared between such divergent organisms as chicken and E. Coli is extremely low, for this reason E. Coli DNA should be a valid monitor of nonspecific DNA binding. The formation of DNA-protein complexes as a function of binding buffer ionic strength is presented in Figure 3. This study was conducted at a protein:DNA input mass ratio of 2:1, which appeared to be saturating for the binding of DNA-P under conditions of very low ionic strength. It will be seen later that under more stringent conditions, this is well above the saturation level. From Fig. 3 it is apparent that even under very high ionic strength binding conditions, binding activity is not abolished. When contrasted to
W
E
ZUE I
0
E
a
C
-
z 0
e
_.
NaCi concentration (M)
Figure 3. Effect of Ionic Strength on DNA-Binding Activity. Binding reactions were performed as described in Fig. 2, in ba3ic binding buffer containing 10 pg of sheared, unlabelled DNA, 2000 cpm of H-DNA of the same DNA fraction as tracer, and 20 iig of input protein, at varying NaCl concentrations as indicated. Background retention varied with NaCl concentration, showing a maximal level of 4.2% for DNA-P and 3.6% for E. Coli DNA. Data has been corrected to reflect background retention. Points are an average of triplicate determinations, with a typical deviation of 1.8%. Symbols: open circles, DNA-P; closed circles, E. Coli DNA.
3419
Nucleic Acids Research other DNA-binding proteins studied in eukaryotic systems, this is clearly unusual, as other such proteins have been observed to lose binding activity at higher ionic strength. This would be expected if ionic forces were significant in the formation of DNA-protein bonding. However, in this case, there is an apparent enhancement of binding activity with increasing ionic strength. This apparent enhancement might be explained on the basis of solvent effects on molecular interactions (29). On this basis, it might be expected that a buffer of high ionic strength would limit ionic interactions, but concommitantly favor the formation of nonionic ( and more specifically, hydrophobic) macromolecular interactions. Alternatively, it is known that ionic strength can affect DNA conformation (30), and it is possible that increasing ionic strength might shift DNA conformation to that for which these proteins have a higher affinity. We suspect that the comparatively high level of DNA retention at zero input ionic strength (which in reality is not zero, but is very low) is the result of molecular aggregation. Aggregation does not fully explain this, however, for there is an approximately two-fold greater retention of homologous over heterologous DNA even under these conditions. Therefore, it is probably the case that both "correct" interactions and aggregation are taking place. Aggregation in the form of precipitable protein complexes does not take place at ionic strengths above approximately 0.1 M, as assessed by light scattering (data not shown). At any point along the ionic strength continuum, there is preference for binding homologous over heterologous DNA, and this preference becomes more pronounced at higher ionic strength. It is probable that increased ionic strength is effective in limiting "incorrect" binding, while it is equally apparent that this condition does not affect "correct" binding (i.e., the binding of DNA-P). For this reason, we conclude that ionic forces probably are not significant in sequence recognition. pH Effects: To further characterize the binding reaction, we studied the effect of pH on the reformation of protein-DNA complexes. These results are presented in Figure 4. The tightly-bound NHCP display binding activity over a generally broad pH range, being optimal between pH 7.0 - 7.4, in close aggreement with that seen for another DNA-binding protein from rat liver (28). It is clear that pH exerts an effect on nonspecific binding (i.e., retention of E. Coli DNA), with nonspecific interaction decreasing with increasing pH. At any given point within the pH range tested, however, significant preference is shown for homologous DNA. Temperature Effects: The effect of incubation temperature on the binding 3420
Nucleic Acids Research
30
100
5E zEI,
_OI
f_
0
50
V
c
co
za
is_
O
X 0~~~~~~~~~ 6.0
7.0
8.0
9.0
pH
Fioure 4. Effect of pH on DNA-Binding Activity. Binding reactions were performed as described in Fig. 3, in binding buffer containing 0.25 M NaCl. The pH range of 5.8 - 7.0 was buffered by 0.05 M phosphate buffer, and the range of 7.2 - 9.0 was buffered by 10 mM Tris-HCI. Correction in input NaCl was made to insure equal final Na concentrations under both conditions. Points are an average of triplicate determinations, with a typical deviation of 2.3%. Background retention varied with pH, and was maximal at 8.2% for DNA-P and 6.2% for E. Coli DNA. Data has been corrected to reflect background retention. Symbols: closed circles, DNA-P; open circles, E. Coli DNA.
reaction is shown in Figure 5. Under the conditions employed, optimal binding is displayed at intermediate temperature, between 220 C - 370 C. A 370 C binding optimum is compatible with physiological temperature, and has been observed for a Drosophila DNA-binding protein (15). As tested, binding is decreased below 220 C and above 370 C. However, some measure of homologous DNA binding is retained at the extremes of 40 C and 600 C. At all points, under the ionic strength conditions emnloyed (0.25 M NaCl), binding of E. Coli DNA is negligible. Therefore, temperature differences are largely manifested in homologous (specific) interactions. As temperature may affect protein conformation, and as it is possible that specific protein conformation may be required for correct binding, the temperature differences seen may be the result of subtle conformational changes within the binding proteins, induced by temperature.
3421
Nucleic Acids Research
-
100
30
ZEa 2 ~~ z
o 'aX
~
~
~
~
~
~
5
50 a
0-1
10
20
30
40
DEGREES CELSIUS
50
60
(Oc)
Figure 5. Effect of Incubation Temperature on DNA-Binding. Binding reactions were performed in basic binding buffer containing 10 mM Tris-HCI, pH 7.4, 0.25 M NaCi, at the indicated temperatures. 5ach reaction mixture received 10 ag of sheared, unlabelled DNA, 2000 cpm of H-DNA of the same DNA fraction as tracer, and 20 ig of input NHCP. Points are an average of triplicate determinations, with a typical deviation of 3.2%. Background retention varied with temperature, being maximal at 2.7% for DNA-P and 1.8% for E. Coli DNA. Data are corrected to reflect background retention. Symbols: open circles, DNA-P; closed circles, E. Coli DNA.
Protein Titration: The binding profile at intermediate (0.25 M NaCl) ionic strength with increasing protein input is presented in Figure 6. In this experiment, we monitored the retention of homologous and heterologous DNA. In addition, we monitored the retention of homologous DNA fractionated into bound (DNA-P) and unbound (DNA-S) as described in Methods. We find that there is an approximately 20-fold preference in binding homologous "bound" DNA over heterologous E. Coli DNA. Saturation is observed at a lower level of retention under these conditions than at lower ionic strength as seen in Fig. 2 (22% vs. 32%), and at lower input protein:DNA mass ratios (0.5 - 1.0:1 vs. 1.5 - 2.0:1). Under these conditions, retention of E. Coli DNA is barely above the limit of detection, while retention of homologous "unbound" DNA is approximately two-fold higher. It is possible that the higher retention of 3422
Nucleic Acids Research
a
Z, 1I-c
0
10
30
20 micrograms
M3
40
Input
Figure 6. DNA-Binding with Increasing Input Protein. Binding reactions were carried out in basic binding buffer containing 0.25 M NaCl at pH 7.4 at 220 C, with the indicated input protein concentrations. Points are an average of triplicate determinations with a typical deviation of 5.1%. Background retention was 3.0% for DNA-P, 1.9% for DNA-S, and 1.4,% for E. Coli DNA. Data has been corrected to reflect background retention. Symbols: closed circles, DNA-P; closed squares, DNA-S; Closed triangles, E. Coli DNA. DNA-S compared to E. Coli DNA
may
indicate that
some
protein binding sites
sequences) remain in DNA-S, and therefore a portion of DNA-S retention be the result of correct sequence binding. It is known that substantial sequence differences exist between the two DNA fractions in chicken reticulocytes (19), but site depletion may not be complete. It is clear, however, that the striking preference demonstrated for the binding of DNA-P is strongly suggestive of specific DNA sequence recognition on the part of this protein class. To confirm that differences in binding between the different DNA fractions were not the result of artifacts in the DNA fractions themselves, we compared the binding of DNA-P, DNA-S, and E. Coli DNA with a nonspecific DNA-binding protein (calf thymus histone H1 Sigma). We found (data not shown) that the efficiency in binding histone H1 of these three fractions was similar in magnitude and reached maximal levels at an input ratio of 1.0 (mg H1/mg DNA) in 10 mM Tris-HCl, pH 7.4, 0.25 M NaCl, at room temperature. Further, the probes (DNA
may
3423
Nucleic Acids Research are all essentially duplex as judged by resistance to digestion by nuclease S1. We found that the nick translated DNA was 99.24% resistant to this digestion.
DISCUSSION We have presented evidence herein to suggest that the tightly-bound NHCP as we have isolated them from chicken liver chromatin have some very interesting properties. It is becoming evident that this may be a general phenomenon in many systems. From our own work, we know that the tightly-bound NHCP from chicken reticulocytes significantly concentrate globin gene sequences (18), are nonrandomly distributed on chromatin DNA, and appear to be exclusively bound to unique sequences within the chicken genome (19). The comparable protein fraction from human placenta significantly concentrates the active placental lactogen gene, and in chick oviduct, the active ovalbumin gene is concentrated (Norman and Bekhor, Mirell and Bekhor, unpublished). Studies on HeLa cell metaphase chromosomes indicate that a similar protein class may be responsible for the establishment and/or maintenance of the highly ordered metaphase chromosome structure (31-33). Similarly, nonrandom distribution of these proteins on genomic DNA of mouse L cell metaphase chromosomes has been reported (34). This protein fraction is resistant to dissociation in 2.0 M NaCl and appears to be immunologically cell specific (35). A tightly-bound NHCP or component thereof may be a participating element in the specific conferral of the DNAase I-sensitive structure of the globin gene in chicken erythrocytes (36). Residual DNA-protein complexes after 2.0 M NaCl extraction of rat prostate chromatin interact directly and tissue-specifically with steroidreceptor complexes (37), and this protein class may therefore include acceptor activities. These combined lines of evidence, coupled with the data presented herein, indicate that this protein class may be responsible for a variety of functions within eukaryotic chromatin, and the suggestion of sequence-specific binding activity makes this class a likely candidate to include modulators of specific gene functions, though more experimentation will be required to determine if this is indeed the case. We do not as yet know what fraction of this diverse protein class actually participates in DNA-binding activity, nor do we have precise information regarding the DNA sequences to which they bind. It is likely that amplification involving recombinant DNA will be required to rigorously demonstrate sequencespecificity, as well as achieving the characterization of the binding site(s). As correctly observed (12), binding to a unique or poorly reiterated sequence within eukaryotic DNA is difficult to detect in the absence of DNA fractionation 3424
Nucleic Acids Research or amplification. We suspect that DNA fractionation of the type employed in this report, using the specific protein class to fractionate the DNA, may prove to be useful in the study of specific interactions between the tightly-bound nonhistone chromosomal proteins and DNA. It is probable that when the DNA can be more rigorously fractionated to individually defined sequences, more striking binding preferences will be detected. In summary, the data that we have presented herein indicate that a
protein class from chicken liver chromatin possessing direct DNA-binding activity displays: (1) strongly preferential binding to fractionated homologous DNA, with approximately ten-fold preference for the bound fraction, (2) negligible binding to heterologous E. Coli DNA, (3) defined pH, temperature, and ionic strength optima, and (4) a DNA-binding activity which is not adversely influenced by ionic strength. Taken together, sequence-specific DNAbinding is strongly suggested. This presumed sequence recognition may be related to the control of genetic readout. ACKNOWLEDGEMENTS The authors wish to thank Mrs. Mary Gates for skillful preparation of graphics. DG is supported by Training Grant DE-07006 from the National Institute of Dental Research, USPHS. Research reported herein was supported by NIH grant DE-04031, and in part by private funds. To whom correspondence should be directed REFERENCES 1.
Abbreviations: NHCP, nonhistone chromosomal proteins; TPD, 0.01 M TrisHCI, pH 8.0, 0.1 mM phenylmethane sulfonyl fluoride, 0.2 mM dithiothreitol;
2. 3.
4.
SSC, 0.15 M NaCl, 0.015 M Na Citrate; SDS, sodium dodecyl sulfate. Paul, J., and Gilmour, R.S. ¾1968) J. Mol. Biol. 34, 305-316. Stein, G., Park, W., Thrall, C., Mans, R., and Stein, J. (1975) Nature 257, 764-767. Gilmour, R.S., Windass, J.D., Affara, N., and Paul, L. (1975) Cell Physiol.
5.
85, 449-458. Bekhor, I. (1978) in The Cell Nucleus, Busch, H., ed., Vol. V, pp. 137-166, Academic Press, New York.
6. Wilcox, G., Clementson, K.J., Santi, D.V., and Engelsberg, E. (1971) Proc. Natl. Acad. Sci. USA 68, 2415-2418. 7. Maniatis, T., and Ptashne, M. (1973) Proc. Natl. Acad. Sci. USA 70, 15311535.
8. Riggs, A.D., Newby, R.F., Bourgeois, S., and Cohn, M. (1968) J. Mol. Biol. 34, 365-368. 9. Gilbert, W., and Muller-Hill, B. (1967) Proc. Nat]. Acad. Sci. USA 58, 2415-2421. 3425
Nucleic Acids Research 10. Thomas, T.L., and Patel, G. (1975) Proc. Natl. Acad. Sci. USA 73, 4364-
4368.
11. Sevall, J.S., Cockburn, A., Savage, M., and Bonner, J. (1975) Biochemistry
14, 782-789. 12. Weideli, H., Schedl, P., Artavanis-Tsakonas, S., Steward, R., Yuan, R., and Gehring, W.J. (1977) Cold Spring Harbor Symp. Quant. Biol., vol. XLII, part 2, pp. 693-700. 13. Jagodzinski, L.L., Chilton, J.C., and Sevall, J.S. (1978) Nucl. Acids Res. 5, 1487-1499. 14. Lesser, B.H., and Comings, D.E. (1978) Biochim. et Biophys. Acta 521, 117-125. 15. Hsieh, T., and Brutlag, D. (1979) Proc. Natl. Acad. Sci. USA 76, 726-730. 16. Lapeyre, J.-N., and Bekhor, I. (1976) J. Mol. Biol. 104, 24-58. 17. Bekhor, I., and Samal, B. (1977) Arch. Biochem. Biophys. 179, 537-543. 18. Bekhor, I., and Mirell, C. (1979) Biochemistry 18, 609-616. 19. Gates, D.M., and Bekhor, I. (1979) Nucl. Acids. Res. 6, 1617-1630. 20. Riggs, A.D., Bourgeois, S., and Cohn, M. (1970) J. Mol. Biol. 53,
401-417. 21. Simpson, R.T. (1977) in Methods in Cell Biology, Prescott, D.M., ed., Vol. XVI, pp. 437-446, Academic Press, New York. 22. Hymer, W.C., and Kuff, E.L. (1964) J. Histochem. Cytochem. 12, 359-363. 23. Lapeyre, J.-N., and Bekhor, I. (1974) J. Mol. Biol. 89, 137-162. 24. Rigby, P.W.J., Dieckmann, M., Rhodes, C., and Berg, P. (1977) J. Mol. Biol . 113, 237-251. 25. Riggs, A.D., Reiness, G., and Zubay, G. (1971) Proc. Natl. Acad. Sci. USA 68, 1222-1225. 26. Bradford, N.M. (1976) Anal. Biochem. 72, 248-54. 27. Laemmli, U.K. (1970) Nature 227, 680-685. 28. Catino, J.J., Yeoman, L.C., Mandel, M., and Busch, H. (1978) Biochemistry 17, 983-987. 29. Klotz, I.M., and Shimkana, K. (1968) Arch. Biochem. Biophys. 123,
551-557.
Chan, A., Kilkusie, R., and Hanlon, S. (1979) Biochemistry 18, 85-91. Paulson, J.R., and Laemmli, U.K. (1977) Cell 12, 817-828. Adolph, K.W., Cheng, S., and Laemmli, U.K. (1977) Cell 12 805-816. Adolph, K.W., Cheng, S.M., Paulson, J.R., and Laemmli, U.K. (1977) Proc. Natl. Acad. Sci. USA 74, 4937-4941. 34. Razin, S.V., Mantieva, V.L., and Georgiev, G.P. (1978) Nucl. Acids. Res.
30. 31. 32. 33.
5, 4737-4751.
35. Campbell, A.M., Briggs, R.C., Bird, R.E., and Hnilica, L.S. (1979) Nucl. Acids. Res. 6, 205-218. 36. Weisbrod, S., and Weintraub, H. (1979) Proc. Natl. Acad. Sci. USA 76,
630-634. 37. Wang, T.Y. (1978) Biochim. et Biophys. Acta 518, 81-88.
3426
