Portal de Periódicos da CAPES

Interactions of the CCAAT-binding Trimer NF-Y with Nucleosomes

1999; Elsevier BV; Volume: 274; Issue: 3 Linguagem: Inglês

10.1074/jbc.274.3.1326

1083-351X

Maria Carla Motta, Giuseppina Caretti, G. Badaracco, Roberto Mantovani,

RNA and protein synthesis mechanisms

NF-Y is a sequence-specific evolutionary conserved activator binding to CCAAT boxes with high affinity and specificity. It is a trimer formed by NF-YA and two putative histone-like subunits, NF-YB and NF-YC, showing similarity to histones H2B and H2A, respectively. We investigated the relationships between NF-Y and chromatin using an Artemia franciscana chromatin assembly system with plasmids containing the Major HistoCompatibility complex class II Ea promoter. The NF-Y trimer, but not single subunits, protects the Y box in the presence of reconstituted chromatin, and it can bind the target sequence during and after assembly. Using reconstitution assays with purified chicken histones, we show that NF-Y associates with preformed nucleosomes. Translational analysis of various Ea fragments of identical length in which the CCAAT box is at different positions indicated that the lateral fragment was slightly more prone to NF-Y binding. In competition experiments, NF-Y is able to prevent formation of nucleosomes significantly. These data support the idea that NF-Y is a gene-specific activator with a built-in capacity to interface with chromatin structures. NF-Y is a sequence-specific evolutionary conserved activator binding to CCAAT boxes with high affinity and specificity. It is a trimer formed by NF-YA and two putative histone-like subunits, NF-YB and NF-YC, showing similarity to histones H2B and H2A, respectively. We investigated the relationships between NF-Y and chromatin using an Artemia franciscana chromatin assembly system with plasmids containing the Major HistoCompatibility complex class II Ea promoter. The NF-Y trimer, but not single subunits, protects the Y box in the presence of reconstituted chromatin, and it can bind the target sequence during and after assembly. Using reconstitution assays with purified chicken histones, we show that NF-Y associates with preformed nucleosomes. Translational analysis of various Ea fragments of identical length in which the CCAAT box is at different positions indicated that the lateral fragment was slightly more prone to NF-Y binding. In competition experiments, NF-Y is able to prevent formation of nucleosomes significantly. These data support the idea that NF-Y is a gene-specific activator with a built-in capacity to interface with chromatin structures. Regulation of gene expression is a complex set of events controlled by gene-specific trans-acting factors and general transcription proteins recognizing cis-acting elements in promoters and enhancers and operating in the context of higher order chromatin structures (1Tjian R. Maniatis T. Cell. 1994; 77: 5-8Abstract Full Text PDF PubMed Scopus (954) Google Scholar). The fundamental chromatin unit is the nucleosome, a DNA-protein complex formed by core histones (H2A, H2B, H3, H4) wrapped around 146 base pairs of DNA. Histones are among the most conserved proteins in evolution; analysis of their quasi-invariant COOH-terminal sequences revealed a 65-amino acid histone fold motif shared by all histone proteins, with low sequence identity, 14/18%, and high structural resemblance (2Arents G. Moudrianakis E.N. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 11170-11174Crossref PubMed Scopus (287) Google Scholar). It is composed of three α-helices: a long central one of 28 amino acids flanked by two short ones separated by short loops/strand regions; this structure enables histones to dimerize with the companion subunit and to make non-sequence-specific contacts with the DNA (3Luger K. Mader A.W. Richmond R.K. Sargent D.F. Richmond T.J. Nature. 1997; 389: 251-260Crossref PubMed Scopus (6846) Google Scholar). The latter function results from tetramerization of H3-H4, which first nucleates the wrapping of DNA and subsequently promotes the association of two H2A-H2B dimers. Recent computational analysis of protein data banks identified additional polypeptides involved in the process of transcriptional regulation which contain putative histone fold domains (4Baxevanis A.D. Arents G. Moudrianakis E.N. Landsman D. Nucleic Acids Res. 1995; 23: 2685-2691Crossref PubMed Scopus (181) Google Scholar): (i) dTAFII60-hTAFII80, dTAFII40-hTAFII31, and hTAFII20-dTAFII30, part of the general TFIID complex; crystallization of dTAFII60-dTAFII40 dimers detailed their histone-like structures (for review, see Ref. 5Burley S.L. Roeder R.G. Annu. Rev. Biochem. 1997; 65: 769-799Crossref Scopus (621) Google Scholar); (ii) the two subunits of NC2 (also called Dr1/DRAP1) which bind TBP and repress transcription (6Goppeldt A. Steltzer G. Lottspeich F. Meisterernst M. EMBO J. 1996; 15: 3105-3116Crossref PubMed Scopus (129) Google Scholar, 7Mermelstein F. Yeung K. Cao J. Inostroza J.A. Erdjument-Bromage H. Eagelson K. Landsman D. Tempst P. Reinberg D. Genes Dev. 1996; 10: 1033-1048Crossref PubMed Scopus (113) Google Scholar); (iii) two subunits of NF-Y, a ubiquitous CCAAT-binding heteromeric complex formed by three proteins, all necessary for DNA binding. The CCAAT box is a widely distributed regulatory sequence, present in 25% of promoters and enhancers, very often at position −60/−80 (8Bucher P. J. Mol. Biol. 1990; 212: 563-578Crossref PubMed Scopus (973) Google Scholar). Functional experiments indicate that the CCAAT box plays an important role in essentially all such promoters. NF-Y, originally identified as the protein binding to the major histocompatibility complex class II Ea promoter Y box, has an almost absolute requirement for these five nucleotides (9Dorn A. Bollekens J. Staub A. Benoist C. Mathis D. Cell. 1987; 50: 863-872Abstract Full Text PDF PubMed Scopus (473) Google Scholar) and has been implicated in the activation of more than 100 promoters (10Mantovani R. Nucleic Acids Res. 1998; 26: 1135-1143Crossref PubMed Scopus (445) Google Scholar). NF-Y genes have been cloned in different species, including yeast, plants, and parasites. Protein alignments evidenced highly conserved domains across evolution (11–14 and refs. therein). NF-YA has a 56-amino acid region that can be split into two short separable parts, responsible for contacting NF-YB-NF-YC and DNA (12Xing Y. Fikes J.D. Guarente L. EMBO J. 1993; 12: 4647-4655Crossref PubMed Scopus (137) Google Scholar,15Mantovani R. Li X.-Y. Pessara U. Hooft van Huijsduijnen R. Benoist C. Mathis D. J. Biol. Chem. 1994; 269: 20340-20346Abstract Full Text PDF PubMed Google Scholar). NF-YB and NF-YC have conserved domains of 90 and 84 amino acids containing putative histone fold motifs (4Baxevanis A.D. Arents G. Moudrianakis E.N. Landsman D. Nucleic Acids Res. 1995; 23: 2685-2691Crossref PubMed Scopus (181) Google Scholar). NF-YB and NF-YC are bound tightly, and their dimerization is a prerequisite for NF-YA association and CCAAT box binding (13Sinha S. Maity S.N. Lu J. deCrombrugghe B. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 1624-1628Crossref PubMed Scopus (250) Google Scholar). Mutagenesis of both proteins showed that the integrity of their histone fold motifs is strictly required for subunit interactions and DNA binding (16Sinha S. Kim I.-S. Sohn K.Y. deCrombrugghe B. Maity S.N. Mol. Cell. Biol. 1996; 16: 328-337Crossref PubMed Scopus (144) Google Scholar, 17Kim I.-S. Sinha S. deCrombrugghe B. Maity S.N. Mol. Cell. Biol. 1996; 16: 4003-4013Crossref PubMed Scopus (127) Google Scholar). They also share a particularly high resemblance to NC2 subunits, extending beyond the histone fold motifs, within the larger yeast/human conserved domains. NF-YB and NC2β belong to the H2B-like subtype, whereas NF-YC and NC2α are closer to H2A (4Baxevanis A.D. Arents G. Moudrianakis E.N. Landsman D. Nucleic Acids Res. 1995; 23: 2685-2691Crossref PubMed Scopus (181) Google Scholar, 6Goppeldt A. Steltzer G. Lottspeich F. Meisterernst M. EMBO J. 1996; 15: 3105-3116Crossref PubMed Scopus (129) Google Scholar, 7Mermelstein F. Yeung K. Cao J. Inostroza J.A. Erdjument-Bromage H. Eagelson K. Landsman D. Tempst P. Reinberg D. Genes Dev. 1996; 10: 1033-1048Crossref PubMed Scopus (113) Google Scholar). Because of this peculiar histone-like structure, we felt it important to investigate the relationships between NF-Y and chromatin. In this study, we use an in vitro chromatin reconstitution system from the brine shrimp Artemia franciscana (18Motta M.C. Landsberger N. Merli C. Badaracco G. J. Biol. Chem. 1998; 273: 18028-18039Abstract Full Text Full Text PDF PubMed Scopus (2) Google Scholar) and nucleosome assembly assays with purified chicken histones to understand the NF-Y-CCAAT interactions in chromatin contexts. A. franciscana dry cysts (U. Ghent, Belgium) were rehydrated in synthetic sea water and developed at 24 °C for 20 h. Embryos (30 g) were rinsed in water, collected, resuspended in 50 ml of extraction buffer (50 mm Tris-HCl, pH 8.0, 30 mm NaCl, 250 mm sucrose, 5 mmβ-mercaptoethanol, 1% dimethyl sulfoxide, 1 mmphenylmethylsulfonyl fluoride, and 1 mg/ml pepstatin) and homogenized. Nuclei were pelletted by centrifugation for 5 min at 8,000 rpm and resuspended in 12 ml of extraction buffer. Nuclei were disrupted by increasing the NaCl concentration to 2 m. The resulting lysate (20 ml) was clarified by centrifugation for 2 h at 60,000 rpm. Aliquots of the supernatant were stored at −80 °C. The protein concentration was usually 10–20 mg/ml. In the assembly reaction (25 μl), 1 μg of plasmid DNA was incubated at 30 °C for 30 min with 10/15 μg of extract in 160 mmNaCl, 25 mm Tris-HCl, pH 8.0, 2 mmMgCl2, 1 mm ATP, 50 ng/ml poly-l-glutamic acid (Fluka), 20 mm disodium creatine phosphate (Sigma), 1 mg/ml creatine phosphokinase (Sigma). Recombinant NF-Y subunits (300 ng) were added to each reaction as indicated in the figure legends. Micrococcal nuclease digestions were performed by adding to the reconstitution reactions 5 CaCl2and micrococcal nuclease (1 unit/mg assembled DNA) (Sigma). The reactions were incubated at 30 °C for 1, 2, and 4 min, stopped by adding SDS and EDTA to a final concentration of 0.4% and 20 mm, respectively, and the mixture incubated for 1 h at 37 °C with proteinase K (2 mg/ml final concentration). The DNA was phenol-chloroform extracted, ethanol precipitated, and run on a 1.6% agarose gel. The hybridization analyses were carried out by standard procedures with the 32P-labeled oligonucleotides specified in the figures. The pE3 plasmid harbors the Major Histocompatibility Complex class II Ea proximal promoter sequences fused to the rabbit β-globin reporter gene, and pE3m16 contains a 10-bp 1The abbreviations used are: bp, base pair(s); wt, wild type. mutation in the Y box (19Viville S. Jongeneel V. Koch W. Mantovani R. Benoist C. Mathis D. J. Immunol. 1991; 146: 3211-3217PubMed Google Scholar). Plasmids were prepared by double banding on CsCl gradients. The labeled fragments used for nucleosome assembly were generated by polymerase chain reaction from pE3, using oligonucleotides mapping to the positions indicated in Figs. 3 and 4. Fragment 2m, generated from pE3m16, contains a 10-bp mutation in the Y box (19Viville S. Jongeneel V. Koch W. Mantovani R. Benoist C. Mathis D. J. Immunol. 1991; 146: 3211-3217PubMed Google Scholar). Oligonucleotides were labeled with T4 kinase, and the amplified fragments were purified on 7% polyacrylamide gels.Figure 4Binding of NF-Y to translationally moved Y boxes. A, scheme of the DNA fragments of identical length used for the translational analysis. Fragment 1 is identical to the one described in Fig. 3. Fragments 4–6 are derived by polymerase chain reaction, designing oligonucleotides so that the CCAAT box is moved progressively toward the center of the fragment. B, nucleosomized fragments are in lanes 1, 9,17, and 25. Free DNAs are in lanes 5,13, 21, and 29. Dose response, 0.2, 0.5, and 1 ng, of NF-Y on nucleosomized DNAs (fragment 1, lanes 2–4; fragment 4, lanes 10–12; fragment 5, lanes 18–20; fragment 6, lanes 26–28) are compared with naked DNA (lanes 6–8, 14–16, 22–24,30–32 for fragments 1 and 4–6, respectively). The nucleosome, NF-Y, and nucleosome/NF-Y bands are indicated.View Large Image Figure ViewerDownload (PPT) NF-Y subunits were produced and purified on nickel nitrilotriacetic acid-agarose (Quiagen) according to standard protocols and dialyzed against buffer BC100 (20Bellorini M. Zemzoumi K. Farina A. Berthelsen J. Piaggio G. Mantovani R. Gene (Amst.). 1997; 193: 119-125Crossref PubMed Scopus (0) Google Scholar). Histones H2A-H2B, H3-H4 were prepared from chicken erythrocyte nuclei using hydroxylapatite columns by the procedure of Simon and Felsenfeld (21Simon R.H. Felsenfeld G. Nucleic Acids Res. 1979; 6: 689-696Crossref PubMed Scopus (292) Google Scholar). Nucleosomes were reconstituted by adapting the method described in Ref. 22Godde J. Nakatani Y. Wolffe A.P. Nucleic Acids Res. 1995; 23: 4557-4564Crossref PubMed Scopus (83) Google Scholar. Unlabeled DNA (500 ng of sonicated salmon sperm DNA) and approximately 2 ng of end-labeled probe (105 cpm) were mixed with 1 μg of purified histones in 1 m NaCl, 10 mm Tris-HCl, pH 7.5, and 1 mmβ-mercaptoethanol (20-μl reaction volume). The reaction was incubated at room temperature, and the salt concentration was lowered to 0.1 m by stepwise addition of TE (10 mmTris-HCl, pH 7.8, 1 mm EDTA) as described (22Godde J. Nakatani Y. Wolffe A.P. Nucleic Acids Res. 1995; 23: 4557-4564Crossref PubMed Scopus (83) Google Scholar); a final volume of 300 μl was obtained. Reconstituted nucleosomes (15 μl of reconstitution mixture) were incubated for 5 min at room temperature in NF-Y binding buffer (20 mm HEPES, pH 7.9, 5 mmMgCl2, 100 μg/ml bovine serum albumin, 30 mmKCl, 1 mm dithiothreitol, and 0.25 mmphenylmethylsulfonyl fluoride) with recombinant NF-Y. After the addition of 3% glycerol, samples were electrophoresed in a 4% polyacrylamide gel (acrylamide/bisacrylamide, 30:1) containing 3.5% glycerol, 0.5 mm dithiothreitol, 0.5 × Tris borate EDTA for 2 h at 250 V. Gels were dried and exposed. Fragment 3 was end labeled by polymerase chain reaction, purified, and reconstituted to 90% with chicken histones; when indicated, 10 ng of NF-Y was added and incubated in NF-Y binding buffer for 20 min at 20 °C. DNase I was added, and the reactions were placed at 30 °C for 1 min, stopped by adding a final concentration of 5 mm EDTA, 0.4% SDS, phenol-chloroform, extracted, ethanol precipitated, and analyzed on a 6% sequencing gel. We have recently developed an in vitro chromatin reconstitution system using whole cell extracts from the brine shrimp A. franciscana; physiological spacing was obtained by the addition of ATP (18Motta M.C. Landsberger N. Merli C. Badaracco G. J. Biol. Chem. 1998; 273: 18028-18039Abstract Full Text Full Text PDF PubMed Scopus (2) Google Scholar). We decided to use it to study the binding capacity of NF-Y on a plasmid containing major histocompatibility complex class II Ea promoter with the high affinity Y box (plasmid pE3, see Ref. 19Viville S. Jongeneel V. Koch W. Mantovani R. Benoist C. Mathis D. J. Immunol. 1991; 146: 3211-3217PubMed Google Scholar). After reconstitution, we treated samples with micrococcal nuclease and separated DNA fragments on agarose gels; hybridization of the resulting blots with oligonucleotide probes mapping to different parts of the plasmid gives clues about the nucleosome patterns on a given plasmid sequence. For all of the following experiments we employed pure recombinant NF-Y, consisting of the yeast/human homology domains of the three subunits; the resulting trimer contains all of the information necessary for efficient and sequence-specific CCAAT box binding (20Bellorini M. Zemzoumi K. Farina A. Berthelsen J. Piaggio G. Mantovani R. Gene (Amst.). 1997; 193: 119-125Crossref PubMed Scopus (0) Google Scholar). We first verified the specificity of the interactions between NF-Y and the DNA in our system, using as a probe an oligonucleotide mapping to the Amp resistance gene, some 2 kilobase pairs away from the Ea promoter sequences. As expected from experiments with other templates, incubation of the extract with pE3 leads to the formation of a regular array of physiologically spaced nucleosomes (Fig.1 A, top panel,lanes 1–3). No differences are observed upon preincubation of DNA with single NF-Y subunits (lanes 4–12); the NF-YB-NF-YC dimer (lanes 13–15) and the NF-Y trimer (lanes 16–18) gave minor destructurations of the nucleosomal pattern. Essentially the same results were obtained with a probe in the β-globin reporter gene, 140 bp downstream from the Y box (Fig. 1 A, middle panel). Using as a probe an oligonucleotide corresponding to the Ea Y box (9Dorn A. Bollekens J. Staub A. Benoist C. Mathis D. Cell. 1987; 50: 863-872Abstract Full Text PDF PubMed Scopus (473) Google Scholar), we observed no change in the pattern with NF-YA (Fig. 1 A, bottom panel, lanes 1–6), NF-YB (lanes 7–9), or NF-YC (lanes 10–12), and few general modifications with the NF-YB-NF-YC dimer (lanes 13–15). However, preincubation of the NF-Y trimer provoked an almost complete disappearance of dinucleosomes, a decrease in the intensity of the mononucleosomes, while an intense subnucleosomal signal was apparent (Fig.1 A, compare lanes 16–18 with lanes 1–3 and lanes 16–18 in the three panels). Hypersensitivity to micrococcal nuclease (MNase) generates such subnucleosomal bands that are usually caused by alterations of the nucleosomal structure induced by binding of DNA-binding proteins, a clear indication that NF-Y binds to the Ea CCAAT box in this context. The lack of such bands with the Amp and β-globin probes strongly suggests that NF-Y binding is specific. To verify this point further, we reconstituted chromatin using a template pEm16, which contains a 10-bp mutation in the CCAAT box (19Viville S. Jongeneel V. Koch W. Mantovani R. Benoist C. Mathis D. J. Immunol. 1991; 146: 3211-3217PubMed Google Scholar). As shown in Fig. 1 B, the addition of NF-YA, NF-YB-NF-YC, or the NF-Y trimer does not modify the nucleosomal pattern significantly; in particular, the subnucleosomal bands observed previously with the trimeric NF-Y were completely absent on this CCAAT-less plasmid (Fig. 1 B, compare lanes 1–3 with 10–12). From these experiments we conclude that subnucleosomal bands indeed result from interactions between the trimeric NF-Y and an intact CCAAT box. We then modified the order of the NF-Y addition to the reconstitution system (see scheme in Fig. 2) by incubating it before, during, or after chromatin assembly. With the control Amp probe the nucleosomal pattern was again similar (not shown); however, when probed with the Y box, we observed that although inhibition of mono- and dinucleosomes was more profound when NF-Y was incubated before or together with template DNA (Fig. 2, comparelanes 4–6 with 7–9), the subnucleosomal bands were evident in all conditions (Fig. 2, compare lanes 1–6and 10–12 with lanes 7–9). We conclude that CCAAT box binding by NF-Y is effective even after chromatin reconstitution. Altogether, these results indicate that in our dynamic chromatin assay (i) the single subunits of NF-Y or the NF-YB-NF-YC histone fold-containing dimer have no wide unspecific effects; (ii) NF-Y binds to DNA in the presence of preformed nucleosomes; (iii) it does not seem to have a completely obstructive role, possibly because of a non-mutually exclusive association with nucleosomes. One possible explanation for these results is that by using Artemia extracts, most likely containing at least some of the chromatin-rearranging machineries recently described in man, Drosophila, and yeast (23Cairns B.R. Trends Biochem. Sci. 1998; 23: 20-25Abstract Full Text PDF PubMed Scopus (152) Google Scholar), the observed effects might be caused by indirect facilitation of chromatin rearrangement by such remodeling activities. To verify whether NF-Y associates with nucleosomal DNA, we switched to a different in vitro system. Purified core histones prepared from chicken erythrocytes in separated couples (Fig.3 A) were assembled with four end-labeled Ea promoter fragments in which the CCAAT box is translationally moved or mutated (see scheme in Fig. 3 A). Labeled DNA and a vast (250-fold) excess of cold sonicated salmon sperm DNA were added to histones in high salts; samples were then diluted progressively to decrease salt concentration, and aliquots were finally loaded on 4% polyacrylamide gels. An efficient shift in DNA mobility was seen only when stoichiometric amounts of H2A-H2B and H3-H4 were added (not shown). This complex was stable for >1 week at 4 °C and represents nucleosomized DNA, as shown by DNase I footprinting experiments (see below). We first controlled whether any of the NF-Y subunits, or the NF-Y complex, was able to bind to nucleosomal DNA in the absence of a target CCAAT box. For this, we used fragment 2m (Fig.3 A) containing the same mutated CCAAT box described in Fig.1 B. As evidenced in Fig. 3 B, neither the single NF-Y subunits nor the trimeric complex was able to associate with DNA, either nucleosome-bound (lanes 1–8) or free (lanes 9–16). As an internal control for such an experiment, we used the corresponding naked wt fragment 2; as expected, this DNA was shifted by increasing amounts of the NF-Y trimer (lanes 17–19). We then turned to wt Ea fragments, employing templates that were only 30% nucleosome-bound, reasoning that free DNA in the reaction would help us compare NF-Y binding affinity for naked versus nucleosomal DNA. The dose responses of single subunits and of the trimer on fragments 1, 2, and 3 are shown in Fig. 3 C. Upper bands of slower mobilities were apparent in all fragments only when the NF-Y trimer was added to nucleosomal DNA (Fig. 3 C, lanes 4–6) but absent in the nucleosome complex (lane 1), when NF-YA or NF-YB-NF-YC was added (lanes 2 and3), and, most notably, in samples with naked DNA (lanes 7–11). Three aspects of these upper complexes should be considered: (i) their intensities increased as a function of trimer concentration; (ii) they appear already under conditions where free DNA is still available for NF-Y to bind (see lanes 5, for example); and (iii) as expected, the mobilities of the NF-Y complexes are different among the naked fragments because of NF-Y-induced DNA bending (24Ronchi A. Bellorini M. Mongelli N. Mantovani R. Nucleic Acids Res. 1995; 23: 4565-4572Crossref PubMed Scopus (85) Google Scholar, 25Liberati C. Ronchi A. Lievens P. Ottolenghi S. Mantovani R. J. Biol. Chem. 1998; 273: 16880-16889Abstract Full Text Full Text PDF PubMed Scopus (43) Google Scholar). In fragment 3, where the CCAAT box is central, the NF-Y complex is slower than nucleosomal DNA, but it is faster in fragment 1 (Fig.3 C, compare lanes 4–6 and 9–11). Note that the upper complexes also show slightly dissimilar electrophoretic behaviors. To define better the influence of translational positioning on the affinity of NF-Y, we decided to move progressively by 20 bp the position of the CCAAT box on fragments of identical length; in addition to fragment 1, we labeled and reconstituted the three additional fragments depicted in Fig. 4 A. A parallel dose response of NF-Y indicates that nucleosomized fragments 4–6 are slightly less prone to bind NF-Y compared with the lateral fragment 1 (compare lanes 1–4, 28–32). Moreover, the latter and the central fragment 6 showed a reduction in the intensities of the nucleosomal bands. These data suggest that, with the exception of fragment 1, in which the CCAAT box is eccentric and NF-Y binding can partially overlap with nucleosomes, the three more central positions are recognized with similar affinity. Because of the putative histone-like structure of NF-YB-NF-YC, we wanted to investigate whether the upper complexes contain all NF-Y subunits. It is possible, in fact, that they might result from association of NF-YB-NF-YC with H3-H4, H2A-H2B, or both. Since it is known that only the trimer has CCAAT binding specificity (11Li X.-Y. Mantovani R. Hooft van Huijsduijnen R. Andre I. Benoist C. Mathis D. Nucleic Acids Res. 1992; 20: 1087-1091Crossref PubMed Scopus (104) Google Scholar), we performed competition experiments on the upper complexes; if the CCAAT-binding trimer is present, we expect it to dissociate from the nucleosomal DNA. In these experiments, we used fragment 2, 30% (Fig.5, lanes 1–6) and 60% reconstituted (lanes 7–12). Note that in the latter case the nucleosomal band is shifted completely at the lowest dose of NF-Y, and the upper complexes are readily seen (Fig. 5, compare lanes 2–4 with 7–9). Incubation of cold short oligonucleotides containing the wt but not the mutant Y box after the addition of NF-Y to nucleosomes resulted in the disappearance of the upper complexes as well as of the NF-Y band (Fig. 5, lanes 4–6 and 10–12), while the nucleosomal bands increased in intensity (Fig. 5, compare lanes 4 and 5,10 and 11). Parallel competition on NF-Y alone on naked DNA abolished, as expected, the NF-Y band (Fig. 5, lanes 10–12). We further characterized the upper nucleosome-NF-Y complexes on fragment 3 by DNase I footprinting (Fig.6). A clear footprint over the CCAAT box was seen when the NF-Y trimer was incubated with DNA (Fig. 6, comparelanes 1 and 2); when DNase was added after nucleosome reconstitution, protection of some sites and emergence of hypersensitive sites with a typical 10-bp period were observed (lane 3). The addition of NF-Y after reconstitution provoked no changes in the DNase cutting pattern outside the CCAAT box region but caused a clear decrease of the major hypersensitive site in the CCAAT box and disappearance of the two neighboring sites (Fig. 6, compare lanes 3 and 4). Overall, the pattern is consistent with the simultaneous presence on DNA of nucleosomes and NF-Y and suggests that NF-Y locally modifies histone-DNA interactions. Altogether, these data prove that the NF-Y trimer can bind to nucleosomal DNA with some (positive) influence of the translational position of the CCAAT box in case the latter is lateral and overlapping the nucleosome(s) border. We felt that our in vitro nucleosome reconstitution system could help us answer an important question: what is the assembly efficiency of nucleosomes when reconstitution is performed in the presence of NF-Y? To address this point, we reconstituted nucleosomes by adding together core histones (Fig. 7, lane 1), increasing concentrations of the NF-Y trimer (lanes 2–5) or NF-YA, with core histones (lanes 6 and7). The NF-Y-DNA complex is formed in the presence of H2A-H2B-H3-H4. Comparison with parallel mock reconstitutions of NF-Y and NF-YA in the absence of histones indicates that the efficiency of NF-Y-CCAAT binding is not reduced and indeed is increased slightly (Fig. 7, compare lanes 2–5 with 8–11). Note that at relatively high NF-Y concentrations, still insufficient to shift completely free DNA, the nucleosomal band was essentially not formed. We take these results as an indication that NF-Y can compete successfully with histones for binding to DNA. Histones associate with DNA in a highly stable and compact way. The resulting structure, the nucleosome, is originated and stabilized by a number of strong protein-protein interactions between histones and by multiple ionic interactions with DNA (2Arents G. Moudrianakis E.N. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 11170-11174Crossref PubMed Scopus (287) Google Scholar, 3Luger K. Mader A.W. Richmond R.K. Sargent D.F. Richmond T.J. Nature. 1997; 389: 251-260Crossref PubMed Scopus (6846) Google Scholar). This structure is generally inhibitory for the binding of transcriptional proteins to promoter sequences. Therefore, one key question in the physiological activation of gene expression is how gene-specific activators and the general transcription factors can reach their target sites when they are embedded in chromatin structures. Consequently, the relationships between different transcription factors and nucleosomes have been investigated in several in vitro studies. The ability of sequence-specific transcriptional activators to associate with nucleosomal DNA is dependent upon several circumstances. The number of binding site(s), their translational and rotational positions, the presence of intact histone tails or their hyperacetylation, and the activity of remodeling proteins are all important factors. The two systems used in this study are to a large extent complementary: the Artemia extract is a dynamic situation requiring energy-consuming machines for proper spacing of nucleosomes (18Motta M.C. Landsberger N. Merli C. Badaracco G. J. Biol. Chem. 1998; 273: 18028-18039Abstract Full Text Full Text PDF PubMed Scopus (2) Google Scholar), whereas reconstitution with purified histones on short DNA fragments rules out possible intervention of additional polypeptides and focuses on the fundamental chromatin unit. In chromatin reconstitutions, NF-Y clearly binds the Ea Y box; protection appears to be relatively limited, as a probe 100 bp downstream shows no sign of nucleosome displacement. Moreover, binding leads to incomplete inhibition of nucleosome formation, possibly because of non-mutually exclusive DNA binding. That this is the case is also suggested by the dose response, competition, and footprinting experiments with purified histones, indicating that NF-Y is able to form complexes with preformed nucleosomes. With the Artemia extract this could be influenced by chromatin-rearranging machineries, but this is certainly not the case with purified histones. In both systems NF-Y requires only a single CCAAT box to bind to nucleosomal DNA. The translational position appears to be important only in the case of the most lateral fragment, which shows a slightly higher affinity, probably because the nucleosome position only partially overlaps with the NF-Y binding site. How does this capacity compare with other transcription factors? The affinity of a given specific DNA-binding activator f