Functional Analysis of Nucleosome Assembly Protein, NAP- 1

A nucleosome assembly protein (NAP-1) of Saccha- romyces cerevisiae facilitates the association of histones with DNA to form nucleosomes in vitro at physiological ionic conditions. The cloned gene was ex- pressed in Escherichia coli using a T7 expression system, and the protein (417 amino acid residues) was purified by Mono Q column chromatography. Various deletion fragments of NAP-1 protein were also produced, and their nucleosome assembly activity was examined by supercoiling assay. The internal fragment containing the residues 43-365 was necessary and sufficient for the activity, and a long stretch of nega- tively charged region near the carboxyl terminus was dispensable. This minimal size fragment could form the 12 S NAP-l-histone complex as the whole protein could, whereas deleted fragments on either side could bind with core histones only to form aggregates. binding

In eukaryotes, DNA and histones are assembled in a repeating unit called a nucleosome. A nucleosome consists of two sets of the four core histones, H2A, H2B, H3, and H4, and approximately 145 base pairs of DNA (1,2). Nucleosomes can be reconstituted by mixing DNA and histones dissolved in >1 M NaCl and 5 M urea and then dialyzing the mixture t o remove salts and denaturant gradually (3). At a physiological ionic strength, DNA and histones form precipitates, and intact nucleosome structure can rarely be formed.
Nucleoplasmin isolated from the eggs of Xenopus laeuis is the first example of such proteins that can promote nucleosome assembly in uitro at physiological conditions (4). N1/N2 were subsequently isolated in a form of complex with histones from the eggs of X . laeuis (5). Dilworth et al. (6) reported that histones H2A and H2B were associated with nucleoplasmin while histones H3 and H4 were associated with N1 and that both complexes were required for the maximal nucleosome assembly. Both nucleoplasmin and N1 contain clusters of negatively charged amino acids (7, 8). Recently, Kleinschmidt et al. (9) showed that the deletion of the larger acidic domain of N1 drastically reduced the histone binding. I t was also reported that the negatively charged polyanions, such as polyglutamic acid or RNA, could assist nucleosome assembly in uitro (10,11). These results suggested that such * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisenent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
To whom correspondence should be addressed. Tel.: 427-24-6229; Fax: 427- 29-1252. a long consecutive stretch, as many as 20 residues, of negatively charged amino acids might be directly involved in nucleosome assembly.
NAP-1 has the same activity as nucleoplasmin and N1/N2 and was first isolated from mammalian cells (12,13) and found in most eukaryotic cells (14). The yeast NAP-1 gene was cloned, and the gene product expressed in Escherichia coli could assist nucleosome assembly in uitro (14). Deduced from the DNA sequence data, NAP-1 contains three negatively charged regions, and the longest one, with 15 glutamic and 13 aspartic residues out of 38 residues, is located near the COOH terminus. In this report, we constructed various deletion derivatives of the yeast NAP-1 gene, expressed them in E. coli, and determined the region essential for nucleosome assembly. We also showed the relation between the 12 S NAP-l-histone complex formation and the activity of nucleosome assembly.

MATERIALS AND METHODS
Construction of Expression Plasmids-An expression plasmid for NAP-1 was constructed as shown in Fig. 1. The Sau3AI-ApaI fragment that codes for amino acid residues 1-42 and the ApaI-Hind111 fragment containing the residues 43-417 were inserted into the T7 expression vector pET3C (15) digested with BamHI and HindIII. The resulting plasmid pTN2 codes for an in-frame fusion product of the whole NAP-1 with the first 11 amino acids of the T7 gene 10 product at the NH, terminus.
Each deletion derivative of NAP-1 was constructed as shown in Fig. 4. The COOH-terminal deletions were made by cleaving each site by an appropriate restriction enzyme and then inserting the SpeI linker which causes termination of translation by creating nonsense codons in all three reading frames. Both NHp-terminal and internal deletion derivatives were made by digesting pTN2 with restriction enzymes and inserting appropriate BamHI linkers to connect the remaining two DNA in-frame.
Expression in E. coli-Each expression plasmid was introduced into E. coli BL21 (DE3), a strain carrying the T7 RNA polymerase gene under the control of the lac UV5 promoter. When the AeW of the culture was 0.5, T7 RNA polymerase was induced by the addition of isopropyl-0-D-thiogalactopyranoside to the final concentration of 0.4 mM. The cells were harvested after 3 h of induction. 50 pg/ml ampicillin was included in all of the culture media (LB) to avoid the loss of recombinant plasmids. Every NAP-1 derivative was produced as the most abundant component in the E. coli crude extract.
Partial Purification of NAP-I-Cells producing NAP-1 and its derivatives in 0.5 liter of culture were harvested after induction and suspended in 5 ml of the buffer containing 100 mM Tris-HC1, pH 7.5, 20% sucrose, 0.5 mM EDTA, 2 mM DTT,' 1 mM phenylmethylsulfonyl fluoride and 200 pg/ml lysozyme. After 20 min a t 0 "C, Brij 58 and KC1 were added to the final concentration of 0.1% and 75 mM, respectively.
For the complete NAP-1 and its deleted forms that were soluble by this extraction method, the suspensions were centrifuged at 140,000 X g for 90 min (Beckman 60 Ti rotor) (16), and the supernatant was dialyzed against 25 mM Tris-HC1, pH 7.5, 2 mM DTT, 10% glycerol, and 50 mM NaC1. Then, proteins were fractionated by ammonium sulfate precipitation. The fraction of 35-65% saturation was saved, dialyzed against 20 mM Tris-HC1, pH 7.5, 0.5 mM EDTA, 1 mM DTT, 10% glycerol, and 0.1 m M phenylmethylsulfonyl fluoride (buffer A) containing 100 mM NaCl, and loaded on a Mono Q column (HR10/10) connected to an FPLC system (Pharmacia LKB Biotechnology Inc.) in buffer A + 0.1 M NaCl. Proteins bound to the column were eluted by the linear gradient of 0.1-1 M NaCl in buffer A.
For the fragments of NAP-1 in (Fig. 4, lanes [4][5][6][7][8] that could not be solubilized by the lysozyme/Brij extraction method, cell suspensions were sonicated at 10-s intervals (20 kHz, 130 watts) for 5 min in Cellruptor (Olympus, Japan) at 4 "C, and the homogenate was centrifuged at 140,000 X g for 30 min. The pellet was dissolved in 50 mM Tris-HC1, pH 8.0,5 M urea, 50 mM NaCl, and 1 mM EDTA, and centrifuged at 140,000 X g for 30 min. The supernatant was dialyzed against buffer A + 0.1 M NaC1, and the NAP-1 fragment was purified by Mono Q column chromatography as described above.
Purification of Histones-Core histones were prepared by the method of Simon and Felsenfeld (17). HeLa chromatin was applied to a hydroxylapatite column, and core histones were eluted from the column with 2 M NaC1. To split core histones into two fractions, H2A +H2B and H3+H4 were eluted from the column by stepwise elution with 0.93 and 2 M NaC1, respectively.
Core histones (0.6 pg), which had been prepared from HeLa cells, were preincubated with NAP-1 derivatives (1.0 pg) for 15 min in the same buffer solution (20 pl). These two reaction mixtures were combined and incubated further for 45 min. SDS and proteinase K were added to the final concentration of 0.2% and 100 pg/ml, respectively, and incubated for 15 min. DNA was purified by phenol/ chloroform extraction and electrophoresed in 1% agarose gel in 89 mM Tris, 89 mM borate, 2.5 mM EDTA buffer system (TBE). Those DNA that could form nucleosomes gain negative superhelical turns and those that failed to form nucleosomes remain as relaxed circles (28). Zmmunoblot Analysis-Proteins were electrophoresed in 12.5% polyacrylamide gel containing 0.1% SDS according to the method of Laemmli (18). Proteins were stained with Coomassie Brilliant Blue or transferred electrophoretically to a nitrocellulose membrane. After the nitrocellulose membrane was incubated in TBS (50 mM Tris-HCl, pH 7.5, 150 mM NaC1) containing 5% horse serum for 1 h, it was further incubated for 1 h at 37 "C with the spent medium of hybridoma cells (clone 4A8) that secrete a monoclonal antibody against HeLa NAP-1 (14,19) or with the rabbit polyclonal antibodies against yeast NAP-1. The nitrocellulose membrane was washed with TBS containing 0.1% Triton X-100 and was soaked in the TBS solution containing peroxidase-conjugated goat anti-mouse IgG (Bio-Rad) for the monoclonal antibody or peroxidase-conjugated goat antirabbit IgG (Bio-Rad) for polyclonal antiserum. After the nitrocellulose membrane was washed, proteins that reacted with the anti-NAP-1 antibody were detected by chloronaphthol (Konica immunostain HRP kit). Polyclonal antibodies were raised against yeast NAP-1 prepared from E. coli using a T7 expression system. This polyclonal antiserum was found to be specific to yeast NAP-1 because it did not cross-react to human NAP-1 protein.

RESULTS
Nucleosome Assembly Activity of NAP-1 Expressed in E. coli-The yeast NAP-1 was overproduced in E. coli using a T7 expression vector. An expression plasmid coding for the intact NAP-1, pTN2, was constructed as shown in Fig. 1. It could express the whole NAP-1 protein (amino acid residues 1-417) fused at its amino terminus with the first 11 amino acids of the T7 gene 10 product, and two more amino acids were created when ligating these two fragments. When total protein extract from E. coli cells harboring pTN2 was analyzed by SDS-polyacrylamide gel electrophoresis, a unique band of about 60 kDa could be seen as one of the major bands, which was not present in the control extract from cells that contained the expression vector without any insert (data not shown). The 35-65% ammonium sulfate fraction of the cell lysate was applied to Mono Q column chromatography, and each fraction was examined by SDS-polyacrylamide gel electrophoresis. The fraction, eluted at about 0.5 M NaC1, contained the 60-kDa protein as a major constituent (Fig. 2a). By immunoblot analysis, this protein was recognized by the monoclonal anti-human NAP-1 antibody (Fig. 26) and by polyclonal antibodies against yeast NAP-1 (Fig. 2c), indicating that the purified 60-kDa protein was yeast NAP-1.
Nucleosome assembly activity of this NAP-1 preparation  (14), a 719-base pair DNA fragment was isolated by digestion with EcoRI. It was further digested with Sau3AI and ApaI, and the resulting 130-base pair fragment that encoded amino acids 1-42 of NAP-1 was obtained. From clone 3, a 2.0-kilobase pair DNA fragment that encoded amino acids 43-417 was isolated by digestion with ApaI and HindIII. These two fragments were ligated into the T7 RNA polymerase-dependent expression vector pET3C (15) that had been digested with BamHI and HindIII. The resulting plasmid pTN2 formed an in-frame fusion product of whole NAP-1 with the first 11 amino acids of the T7 gene 10 product. E, A, H, B, and S represent EcoRI, ApaI, HindIII, BamHI, and Sau3A1, respectively. was examined by the supercoiling assay. NAP-1 was preincubated with histones, and then relaxed circular DNA was added to the reaction mixture, and the reaction mixture was incubated in the presence of topoisomerase I. When the DNA was recovered as a negatively supercoiled form, it indicated that nucleosomes were formed, otherwise DNA remained as a relaxed circle. When 0.2 pg of pBR322 DNA was reacted with 0.6 pg of core histones in the presence of 1 pg of NAP-1, the DNA was recovered as a fully supercoiled form. The nucleosome assembly reaction was dependent on the presence of both core histones and NAP-1 (Fig. 3a). These results  5 ( l a n e 4 ) , and 1.0 ( l a n e 5 ) pg or a fixed amount (1.0 pg) of NAP-1 (lanes [6][7][8][9] was incubated with decreasing amounts (0.4-0 pg) of histones: 0.4 ( l a n e 6), 0.2 ( l a n e 7), 0.1 ( l a n e 8), and 0 ( l a n e 9) pg for 15 min. Relaxed pBR322 DNA (0.3 pg) was added to the mixture, and the reaction mixtures were further incubated for 45 min in the presence of topoisomerase I. indicated that yeast NAP-1 expressed in E. coli using T7 expression vector had an activity of nucleosome assembly comparable to that of NAP-1 purified from mouse cells (13). Next, NAP-1 was mixed with either histones H2A+H2B or H3+H4 instead of core histones and each supercoiling activity was examined (Fig. 3b). When NAP-1 was reacted with only H2A+H2B histones, no superhelical turn was introduced at any amount. On the other hand, when NAP-1 was incubated with 0.3 or 0.6 pg of H3+H4 histones, superhelical turns were introduced into DNA. But no superhelical changes were observed in the presence of 0.9 pg of histones H3+H4. In this case, the amount of NAP-1 becomes limiting over the excess amount of H3+H4 histones. Free histones tend to bind to DNA in a nonspecific manner to form an aberrant structure and inhibit the nucleosome assembly process. These results suggest that NAP-1 binds with H3 and H4 histones and transfers them to DNA to make the basal structure of nucleosomes.
Analysis of Short Fragments of NAP-1-To determine the regions necessary for nucleosome assembly, a part of DNA coding for NAP-1 was deleted as shown in Fig. 4. These deletion fragments were expressed in E. coli using a T7 expression vector, and each protein was partially purified by  Fig. 4. NAP-I derivative (1 pg) was added to each reaction mixture; no protein was added in lane 15. Mono Q column chromatography in the same way as the intact NAP-1, and they were examined by SDS gel electrophoresis (Fig. 2a). The size of these proteins almost coincides with the value calculated from their sequence data. These proteins were all recognized by polyclonal antibodies against yeast NAP-1, which were prepared by immunizing rabbits with partially purified yeast NAP-1 product expressed in E. coli (Fig. 2c). Monoclonal antibody also recognized all of these fragments but two (lanes 5 and 7), which might be devoid of the epitope for this antibody (Fig. 2b). These results indicated that purified proteins were deletion fragments of yeast The nucleosome assembly activity of each protein was examined by the supercoiling assay (Fig. 5). When we deleted the 52 residues from the COOH terminus, the resulting fragment devoid of the largest negatively charged region at the COOH terminus had the full activity comparable to the intact NAP-1 protein. But the shorter fragment lacking further 21 residues from the COOH terminus did not support the activity, although it still had the intact NH2 terminus and almost all of the two other negatively charged regions located in the center of the protein. Thus, we concluded that the longest negatively charged region at the COOH terminus was not essential for nucleosome assembly.

NAP-1.
The fragment lacking the first 42 amino acid residues at the NH2 terminus also retained the nucleosome assembly activity. Since the 52 residues from the COOH terminus were not essential, we constructed fragments depleted of both the NH2-terminal and the COOH-terminal regions. A deletion fragment that lacked the residues 1-42 and 366-417 had the nucleosome assembly activity, but the further deletion from the NH2 terminus, removing residues 43-126, abolished the activity. These results suggest that the region between residues 43 and 365 is necessary and sufficient for the nucleosome assembly activity.
To know the role of internal region, we constructed internal deletion fragments. As shown in Fig. 5, none of the internal deletion fragments (lanes 6-10) had the nucleosome assembly activity comparable with that of complete NAP-1, although a small amount of fully supercoiled DNA was formed under the conditions used (lanes 7 and 9). Some internal deletion fragments (lanes 6, 8, and 10) were inactive even though they carried all three negatively charged regions. These results (summarized in Fig. 4) suggest that the existence of negatively charged regions is not sufficient for the nucleosome assembly activity.
Binding of NAP-1 Derivatives with Core Histones-We next examined whether the fragments that did not have the assembly activity could bind the core histones. These deletion derivatives of NAP-1, inactive for nucleosome assembly reaction, could bind the core histones as effectively as intact NAP-1 in the enzyme-linked immunosorbent assay (26) (data not shown). As reported previously (26), when native NAP-1 was mixed with the core histones under physiological conditions, a 12 S complex was formed that contained NAP-1 and four kinds of histones in equal amounts. When the fragment lacking both residues 1-42 and 366-417 (Fig. 5, lane 12) was reacted, large aggregates (>15 S) were formed in addition to the 12 S complex (Fig. 6a). As summarized in Table I, such an aggregate was not formed when the intact NAP-1 (Fig. 4,  lane 1 ) and NAP-1 derivatives (lanes 3 and 11 ) were mixed with histones. Thus the fragments in lane 12 might be less    Fig. 4, and molecular mass of the NAP-1 fragments calculated from their sequences is indicated.   1-126 and 366-417 ( l a n e 13) or a fragment lacking residues 335-417 ( l a n e 4 ) was reacted (Fig.  6, b and c), only the aggregates were formed. Under the same condition, free core histones or the NAP-1 fragments sedi- mented at the 1-3 S position (data not shown). We tested the complex formation of the NAP-1 fragments as listed in Table   I. The data indicate that all of the fragments that formed the 12 S complex had the nucleosome assembly activity comparable with the native NAP-1 and that all of the fragments that made aggregates did not have the nucleosome assembly activity. We concluded that the formation of the 12 S complex was essential for the activity and that the fragment between residues 43 and 365 was indispensable for the formation of the 12 S complex. Epitope to Monoclonal Antibody-Although the fragment of NAP-1 missing residues 335-417 reacted with the monoclonal antibody against NAP-1 purified from HeLa cells, the deletion fragment missing the residue 257-417 did not. The internal deletion of NAP-1 missing the residue 257-334 did not react with monoclonal antibody either (Fig. 2b). These results indicate that the epitope against the monoclonal antibody is located between residues 257 and 334. This monoclonal antibody reacted with a 50-60-kDa protein in human, mouse, Xenopus, fruit fly, and yeast cell lysates (14). Therefore, a part of this region must be highly conserved among these species and is located in the region required for the NAP-1 function.

DISCUSSION
Clusters of negatively charged regions are found in N1, nucleoplasmin, and HMG-1, all of which could promote nucleosome assembly in uitro. Kleinschmidt et al. (9) suggested that the acidic domain in the center of protein N1 played a major role in binding histones. This is supported by the result that polyglutamic acid could promote nucleosome assembly at a physiological ionic condition.
We have already shown that three negatively charged re-TABLE I1 Putative epitope sequence for NAP-1 monoclonal antibody Nucleotide sequence of clone 734-2 and its translated sequence in-frame with the lacZ sequence of the Xgtll vector is given (EcoRI cloning site is underlined). This open reading frame gave a strong positive signal for immunoblotting analysis by NAP-1 monoclonal antibody, which was confirmed by recloning it in other expression vectors, for example, the Gemex system (Promega) using T7 gene 10 to make a fusion protein. Amino acid sequence of yeast NAP-1 from residues 305 to 320 (14) is compared with the corresponding region deduced from the nucleotide sequence of mouse NAP-1 clone (4AS-9). *, stop codon.

734-2
lacZ-GAATTCCCGTGTTGCTCAAGTACTCTCCAGTTCAATTTCCCCATCTGA E F P C C S S T L Q F N F P I . Yeast NAP-1 (305) K I T P I E S F F N F F D P P K ( 3 2 0 ) Mouse NAP-1 (4AS-9)

K T V S N D S F F N F F A P P E
gions are present in yeast NAP-1 protein (14). The NAP-1 fragment (lane 3 in Fig. 4) missing the largest negatively charged region of 38 amino acids at the COOH terminus retained the nucleosome assembly activity. The fragments (Fig. 4, 'lane 4 ) lacking the region with a helix-turn-helix motif (330-356) did not have the activity. These results suggest that a long stretch of negatively charged regions might be dispensable for the nucleosome assembly but other structural features might be important. Since the fragment lacking the COOH-terminal negatively charged region is still acidic (PI = 4.9), it is possible that the acidity of the protein as a whole is one of the important factors for the nucleosome assembly. We also found that the fragments active in the nucleosome assembly could precedently form a 12 S histone-NAP-1 complex. Other inactive derivatives formed large aggregates, when mixed with core histones, which sedimented at the bottom of sucrose gradient. These results indicate that the formation of the 12 S complex by NAP-1 and core histones is correlated with the nucleosome assembly. It has been shown that nucleosomes are formed when this 12 S complex is mixed with DNA (13). Some deleted fragments were not soluble by the lysozyme/ Brij extraction method. We solubilized these proteins in urea to purify them. These proteins were inactive or less active for nucleosome assembly. We cannot exclude the possibility that treatment with urea changes the proper conformation of the proteins and renders them inactive, although the treatment of purified NAP-1 protein with urea does not affect the activity at all.
In vivo studies have demonstrated that during DNA replication, histones H3+H4 became associated with newly replicated DNA before assembly of histones H2A+H2B (20-24). N1/N2 bind H3+H4 histones, and this complex can induce supercoiling of DNA (25). So, we examined whether NAP-1 could induce superhelical turns when it was reacted with only H3+H4 histones (Fig. 3b). With this combination, negative supercoils were introduced to the same extent as in the reaction containing all four core histones. Ishimi et al. (26) showed that an 8 S complex was formed by mixing H3+H4 histones with mouse NAP-1. These results suggest that NAP-1 binds with H3+H4 histones and transfers them to DNA to form the backbone structure of nucleosomes with a n arginine-rich kernel (29). When deletion fragments of NAP-1 were reacted with only histones H3+H4, those that could promote nucleosome assembly with core histones could also introduce supercoiling of DNA in this system (data not shown). These results suggest that supercoiling of DNA depends on the interaction of NAP-l with histones H3+H4. Kleinschmidt et al. (25) indicated that histones H3+H4 bound to N1/N2 could introduce supercoiling of DNA at a histone:DNA ratio of 2:1, but could not at a histone:DNA ratio of 6:l. Our results in Fig. 3 seem to coincide with the results of Kleinschmidt, et al. (25) and suggest that the ratio of histones to NAP-1 in the reaction mixture is important for nucleosome assembly. Zucker and Worcel (27) indicated that DNA supercoiling induced by the histones H3+H4/N1 complex was caused by the subnucleosomal particles containing histones H3+H4 and 65 base pairs of DNA. We do not know whether supercoiling induced by histones H3+H4 and NAP-1 complex is caused by the subnucleosomal particles.
An intriguing observation is that polyclonal antibodies raised against yeast protein recognize NAP-1 from yeast origin but not from any other organisms. Since the monoclonal antibody raised against HeLa NAP-1 cross-reacted with yeast NAP-1 as well as that of Xenopus and Drosophila, its epitope must be in the very conserved region. In the course of screening the Xgtll cDNA library of mouse spermatocyte by using the monoclonal antibody, we picked up one false clone in addition to the authentic mouse NAP-1 From the sequence analysis of the false clone, it was found that oligopeptide (13 residues as shown in Table 11, clone 734-2) was fused to the end of &galactosidase in the clone, which gave an extremely strong positive sign to its fusion product by immunoblotting analysis (better than any other positive clones). In Table 11, the sequence of this fusion peptide is compared with the epitope regions of yeast NAP-1 and the corresponding mouse NAP-1 sequence. Apparently, FNF sequence is common to all the three clones, suggesting that this stretch of a few amino acids may be the epitope of monoclonal antibody.
Comparing mouse NAP-1 sequence with that of yeast, the putative epitope sequence and another KGIPEFWLT sequence (14) are strictly conserved (8 residues out of 9 in both circumstances), while other parts are considerably diverted (homology of amino acid sequences is less than 30%). This may explain the failure of polyclonal antibodies to cross-react with NAP-1 from any other organisms than of yeast origin.
The NAP-1 protein could be produced and purified in quantity, which is vely useful to reconstitute chromosomes in uitro. These chromosomes can serve as a template for transcription or replication in uitro. Since the chromosomal structure in the initiation site of replication or transcription is very important, the reconstruction of chromosomes with or without initiator proteins could greatly accelerate the future studies of regulation of transcription and replication of chromosomes.