Characterization of the Histone H1-binding Protein, NASP, as a Cell Cycle-regulated Somatic Protein*

Nuclear autoantigenic sperm protein (NASP), initially described as a highly autoimmunogenic testis and sperm-specific protein, is a histone-binding protein that is a homologue of the N1/N2 gene expressed in oocytes of Xenopus laevis. Here, we report a somatic form of NASP (sNASP) present in all mitotic cells examined, including mouse embryonic cells and several mouse and human tissue culture cell lines. Affinity chromatography and histone isolation demonstrate that NASP from myeloma cells is complexed only with H1, linker histones. Somatic NASP is a shorter version of testicular NASP (tNASP) with two deletions in the coding region arising from alternative splicing and differs from tNASP in its 5′ untranslated regions. We examined the relationship between NASP mRNA expression and the cell cycle and report that in cultures of synchronized mouse 3T3 cells and HeLa cells sNASP mRNA levels increase during S-phase and decline in G2, concomitant with histone mRNA levels. NASP protein levels remain stable in these cells but become undetectable in confluent cultures of nondividing CV-1 cells and in nonmitotic cells in various body tissues. Expression of sNASP mRNA is regulated during the cell cycle and, consistent with a role as a histone transport protein, NASP mRNA expression parallels histone mRNA expression.

NASP, 1 initially described as a highly autoimmunogenic testis and sperm-specific protein, is present in the nucleus of spermatozoa and spermatogenic cells (1)(2)(3), hence the name nuclear autoantigenic sperm protein. Previous studies (4) have demonstrated that human NASP contains three functional his-tone binding sites: site I (amino acids 116 -127), site II, (amino acids 469 -512), and site III (amino acids 211-244). In vitro, recombinant NASP will bind both linker and core histones (4). 2 As a histone-binding protein (4,5), NASP appears to be a homologue of the N1/N2 gene expressed in oocytes of Xenopus laevis (6 -9) because, in addition to the conserved coding regions, there is an almost 60% identity between the 3Ј untranslated regions of rabbit NASP and Xenopus N1/N2 mRNAs (2). In Xenopus oocytes, the non-chromatin-bound core histones H3 and H4 are associated in a complex with N1/N2 (10,11), providing a mechanism for the storage of histones for DNA replication in the early embryo. Both mouse (12) and sea urchin oocytes (13) also store histones in a non-chromatin-bound form. Other mammalian histone-binding proteins, for example the nucleosome assembly factor NAP-1 (14 -17), the chromatin assembly factor CAF-1 (18 -21), and the transcription proteins HIRA-HIRIP3 (22), not only bind histones but appear to be important for the assembly of chromatin (19,(23)(24)(25)(26). Consequently NASP may be important not only for both the storage and transport of histones but also for the assembly of chromatin.
Here, we report a somatic form of NASP (sNASP) present in almost all tissues examined, including mouse embryonic cells and several mouse and human tissue culture cell lines. Interestingly, an analysis of the histones complexed with NASP in vivo revealed that in mouse myeloma 66-2 cells only H1, linker histones, could be detected. Somatic NASP is a shorter version of the testicular form of NASP (tNASP) with two deletions in the coding region arising from alternative splicing and differing from tNASP in its 5Ј untranslated regions. In both embryonic and transformed cell lines in culture, we found that both sNASP and tNASP are expressed. Unlike spermatogenesis in the testis, in which DNA replication and histone synthesis are uncoupled, in somatic cells histone synthesis and histone mRNA levels are closely coupled to DNA replication (27)(28)(29). Having found that NASP is present in somatic cells, we therefore also examined the relationship between NASP mRNA expression and the cell cycle and report that sNASP mRNA levels increase during S-phase and decline during G 2 , concomitant with histone mRNA levels.

EXPERIMENTAL PROCEDURES
All reagents and chemicals used in this study were of molecular biology grade. Alkaline phosphatase, peroxidase, and fluorescein isothiocyanate-labeled goat anti-rabbit (heavy and light chain) antisera were purchased from Organon Technica, Inc. (West Chester, PA). DNA probes were labeled with either a North2South biotin-DNA labeling kit (Pierce) or [ 32 P-rsqb]dCTP (ICN Biomedicals Inc., Costa Mesa, CA) with Ready To Go DNA labeling beads (Amersham Pharmacia Biotech). All oligonucleotides were synthesized at the University of North Carolina-Chapel Hill (UNC-CH) Nucleic Acids Core Facility. Restriction and modifying enzymes were purchased from New England Biolabs, Inc. (Beverly, MA). Purifications of plasmid DNA and PCR products were carried out using QIAprep Miniprep and QIAquick PCR purification kits, respectively (Qiagen Inc., Valencia, CA).

Library Screening and Sequencing
The mouse testis cDNA library was described previously (30), and the mouse ovary cDNA library was a gift from Dr. Jurrien Dean (National Institutes of Health, Bethesda, MD). Screening by colony hybridization was performed with 32 P-labeled probes using Express-Hyb Solution (CLONTECH, Inc., Palo Alto, CA) according to the recommended protocol. Both strands of positive clones were sequenced by the UNC-CH automated sequencing facility on a model 377 DNA sequencer using the ABI Prism rhodamine terminator cycle sequencing kit (Applied Biosystems Inc., Foster City, CA). The sequence for mouse tNASP ( Fig. 1) was obtained by screening a mouse testis cDNA library with a rabbit tNASP probe (2). The sequence for sNASP ( Fig. 1) was obtained by screening a mouse ovary cDNA library with mouse tNASP cDNA. Sequence manipulations and analyses were performed using DNAsis software (Hitachi Software Inc., San Francisco, CA).

Northern Blots
Northern blots (Fig. 2, a multiple tissue Northern blot; Fig. 3, an RNA master blot (CLONTECH)) were probed with 32 P-labeled fulllength tNASP, generated by PCR primers specific to the coding region of tNASP (nt 92-2414), and hybridized using Express-Hyb Solution according to the CLONTECH protocol. Blots were exposed for 16 h on a phosphorImager screen, scanned using a Molecular Dynamics Storm 860 PhosphorImager (Amersham Pharmacia Biotech), and analyzed and quantitated using either ImageQuant software (ImageQuant, Inc., Sunnyvale, CA) or GelExpert software (Nucleotech, Inc., San Mateo, CA). Northern blots of HeLa cell and mouse 3T3 fibroblast RNA were prepared from 1.2% agarose gels loaded with 5 g/well total RNA and electrophoresed as described for low percentage formaldehyde gels in the Qiagen RNeasy Mini Kit handbook. After electrophoresis, the gels were capillary-transferred to Hybond-N (Amersham Pharmacia Biotech).

PCR Analysis
Gene expression in a variety of tissues was studied using the multiple tissue cDNA panel (CLONTECH). CLONTECH cDNAs are normalized to eight different housekeeping genes. cDNAs from thymus and ovary were prepared in our laboratory, but these were not normalized to different housekeeping genes. Thymus and ovary cDNAs were prepared as follows. RNA was isolated from the tissue according to the manufacturer's instructions using the RNeasy Mini Kit for Total RNA (Qiagen) and cDNA was prepared from an oligo(dT) primer using a SuperScript preamplication system for first-strand cDNA synthesis (Life Technologies, Inc.). PCRs were performed using Ready To Go PCR beads (Amersham Pharmacia Biotech). Cycle conditions were set up as described previously (23) using the following primers: for amplification of the full-length NASP coding region, the sense primer was 5Ј-ATGGCCA-CAGATTCTACAGCC-3Ј and the antisense primer was 5Ј-TTAACATG-CAGTGCTTTT-3Ј; for amplification of a fragment specific for tNASP, the sense primer was 5Ј-AATGAGTGTGGGGAAGCC-3Ј and the antisense primer was 5Ј-TTCACTCTCAGAGGTAGC-3Ј.

Antisera
Rabbits were immunized with purified recombinant proteins as described (31). Expression constructs were produced using PCR, cloned into pQE-30, transformed to bacterial host M15 (pREP-4) for expression, and purified on Ni-NTA resin as described (31). N-terminal (nucleotides 96 -1099) and C-terminal (nucleotides 1100 -2414) constructs from the sequence of mouse tNASP (GenBank TM accession number AF034610) were used for recombinant protein production. Affinitypurified antibodies were prepared as described (31) using immobilized N-terminal recombinant protein.

Western Blots
Proteins were separated on 10 -20% gradient minigels (Bio-Rad) and transblotted to Immobilon-P (Waters Inc., Bedford, MA), either stained for protein or blocked, and probed with anti-NASP antiserum as described (32). In blots used to assess the expression of NASP in HeLa and 3T3 cells, the blots were stained with amido black for protein to ensure equal loading, computer-imaged with a desktop scanner, destained, and immunostained with anti-NASP. Protein lysates of mouse embryo, spleen, and testis were made by grinding the minced tissue in a Dounce homogenizer with PBS and Protease Inhibitor Mixture Set 1 (Calbiochem) at the recommended dilution. The suspension was frozen, thawed, and centrifuged at 12,000 ϫ g. After centrifugation, the supernatant was extracted (twice in chloroform and methanol (3:1)) and the volatile solvents removed by vacuum centrifugation (10 min). Protein concentrations were determined using Micro BCA protein detection reagents (Pierce).

Immunohistochemistry
Mouse tissue sections and embryos were fixed in Bouin's solution, paraffin-embedded, sectioned, and incubated with rabbit anti-recombinant N-terminal NASP (affinity-purified) as primary antiserum. This antiserum recognizes both somatic and testicular NASP protein. Immunostaining was carried out as described (33).

Indirect Immunofluorescence
Cell cultures were grown on microscope slides with wells (Lab-Tek II chamber slides, Nalgene-Nunc, Naperville, IL). Medium was removed by aspiration and the cells washed twice with cold (4°C) PBS. Methanol (Ϫ20°C) was added to fix the cells (10 min, Ϫ20°C). The cells were washed twice with cold PBS, blocked in 2% normal goat serum (20 min), and washed with cold PBS (twice, 5 min each). Rabbit anti-NASP or preimmune serum was added to each well (1:500 in PBS and 0.1% bovine serum albumin), the cells were incubated (45 min) and washed in PBS (3 times, 5 min each), and fluorescein isothiocyanate-labeled goat anti-rabbit antiserum was added (1:1000 in PBS and 0.1% bovine serum albumin) for a 30-min incubation followed by three washes in PBS. The cells were also stained with propidium iodide (0.01 mg/ml, 10 min) and washed (3 times in PBS) before the plastic wells were removed.

Cell Cycle Studies
Thymidine Block-To obtain populations of cells in G 2 -phase, HeLa cells were arrested by double-thymidine block. Cells were blocked for 18 h with 2 mM thymidine, released for 9 h by washing out the thymidine, and blocked again with 2 mM thymidine for 17 h to arrest all of the cells at the beginning of S-phase. The cells were released by washing out the thymidine and total RNA and lysates prepared from samples taken at specific time points after release from thymidine.
Serum Starvation-At 72 h after plating 3T3 cells, complete medium was removed, cells were washed three times with medium alone and incubated with Dulbecco's modified Eagle's medium containing 0.5% calf serum for 48 h. At time 0 h, the low serum medium was removed, and the cells were stimulated by addition of medium containing 10% serum.
The cell cycle positions of the HeLa and 3T3 cells were determined by analyzing the DNA content of propidium iodide-stained cells from each time point at the UNC Flow Cytometry Facility. Total cell RNA was prepared for the cell cycle studies using the Ultraspec RNA isolation system (Biotecx Laboratories Inc., Houston, TX). Cell lysates were prepared by washing the collected cells in PBS followed by lysis in 1% SDS and boiling for 5 min. In all cell cycle studies, equal RNA loading was verified by probing Northern blots with G3PDH, and the position of the cells in the cell cycle was monitored using FACS analysis (shown for Fig. 6).

Affinity Purification
Rabbit antibodies to recombinant mouse NASP were affixed to Reactigel 6X (Pierce) and were used to affinity purify native NASP and its associated histones from lysates. Myeloma 66-2 cell pellets were thawed on ice and resuspended in phosphate-buffered saline (PBS) with the addition of Protease Inhibitor Mixture I and Proteasome Inhibitor I (Calbiochem). Cell debris was pelleted in a microcentrifuge at maximum speed, and the supernatant was mixed with the NASP:Reactigel beads and rocked overnight at 4°C. A column was made from the beads and was washed extensively with several volumes of PBS before elution with Immunopure elution buffer (Pierce). Collected fractions were immediately neutralized with 1 M Tris-Cl, pH 8.5, and elution of protein was tracked at 280 nm. Pooled fractions were dialyzed in 10,000 molecular weight cutoff Slidealyzer cassettes (Pierce) versus several changes of 1/20ϫ PBS and lyophilized.

Isolation of Histones
H1 proteins from mouse myeloma cells or affinity-purified NASP were isolated as described by Brown and Sittman (35). Core histones were isolated from the perchloric acid-insoluble pellet by extraction with 0.2 M H 2 SO 4 , precipitated with trichloroacetic acid or 8 volumes of cold acetone, and washed with acetone, 0.01 M HCl. All procedures were done at 4°C. Mouse histone H1 cDNAs (kindly provided by Dr. D. Sittman, University of Mississippi Medical Center) were expressed from a pET-11d vector. The histone H1 constructs were transformed into the BL21 (DE3) bacterial host cell line and the recombinant proteins purified using the T7 tag affinity purification kit (Novagen, Madison, WI) according to the manufacturer's instructions.

Histone Labeling
Biotin labeling of histones was performed as described (4). The binding of labeled histones to streptavidin-agarose was used to test the efficiency of biotinylation, and bound and unbound fractions were analyzed by SDS-polyacrylamide gel electrophoresis.

Co-immunoprecipitation
For binding experiments, rNASP (2 g) was mixed with biotinlabeled histone H1 recombinants in a 500-l volume of 0.5ϫ Trisbuffered EDTA and incubated for 1 h at room temperature with gentle shaking. Precipitation was carried out with either streptavidin or affinity-purified anti-rNASP antibodies. The samples were loaded on 10 -20% SDS-polyacrylamide gels.

Sequences of Somatic and Testicular
Mouse NASP-In the course of our studies on tNASP, it was of interest to ask whether or not the fertilizing spermatozoon would carry NASP into the unfertilized egg (5). However, contrary to expectations (2), NASP was detected with an anti-NASP antibody in the germinal vesicle of growing oocytes within the ovary (see Fig.  9E). Subsequent screening of an ovarian cDNA library resulted in the isolation of a cDNA encoding sNASP, which differed from tNASP, suggesting that the ovarian cDNA was an alternatively spliced transcript. As a result of this observation, we began an extensive search for the presence of NASP in embryonic and somatic tissues. Fig. 1A shows the relationship between sNASP and tNASP cDNAs. The sNASP is identical to tNASP except for the deletion of the two separate regions, shown in Fig. 1A, reducing the size of sNASP to 421 amino acids (45,752 Da) from the 773 amino acids of tNASP (83,934 Da). The sequence for mouse tNASP (Fig. 1B), obtained by screening a mouse testis cDNA library, yielded a 2543-bp consensus sequence. The sequence for sNASP (Fig. 1B), obtained from the mouse ovary cDNA library yielded a consensus sequence of 1486 bp, which differs from tNASP cDNA in the 5Ј UTR as well as in the deletion of two portions of the coding region. To confirm the sequence of sNASP in other somatic tissues, the complete sNASP coding region from spleen cDNA, generated by PCR, was sequenced and found to be identical to ovary sNASP (data not shown). The lack of two tNASP regions in the sNASP mRNA and the presence of differing sequences in the upstream portion of the 5Ј UTR of sNASP are most likely explained by differential pre-mRNA splicing events and different promoters. Fig. 1C shows the relevant 5Ј UTR sequences for the two NASP forms, indicating how the 5Ј ends of the transcripts are alternatively spliced from the genomic sequence. The 5Ј end of the testis mRNA arises by the splicing of two small exons (T1a and T1b, Fig. 1C) onto the 5Ј end of exon one (E-1). The unique portion of the 5Ј UTR of sNASP, (S1, Fig. 1C) is contiguous with the sequence (E-1) common to both sNASP and tNASP. A 3Ј splice is evident at the end of S1 (Fig. 1C). The splice junctions of the exons have been identified by sequencing genomic DNA (not shown).
The deduced mouse tNASP protein sequence is 81% identical to human tNASP and retains the three histone binding sites (Fig. 1A) previously identified in human tNASP (4). Additionally, the sNASP-deduced protein sequence has the same structural features found in all tNASP sequences (36), including a nuclear localization signal, a leucine zipper, and an ATP/GTP binding site.
Northern Blot Analysis-Probing a multiple tissue Northern blot (CLONTECH) of mouse tissues with the tNASP coding region cDNA revealed a strong signal in mouse testis at 2.7 kb (ϽT) and a weaker signal at 1.7 kb (ϽS) in mouse liver, kidney, spleen, brain, and heart ( Fig. 2A). Probing 66-2 mouse myeloma cell RNA revealed a strong 1.7-kb (ϽS) signal and a weaker 2.7-kb (ϽT) signal (Fig. 2B). To assay a wider variety of tissues for NASP expression, a mouse mRNA master blot was probed with the tNASP coding region. As shown in Fig. 3A, testis mRNA showed the strongest signal, at least five times that found in 11-day-old embryos. Densitometeric analysis of the blot without the testis signal (Fig. 3B) revealed that the other most predominant signals were found in the thymus, spleen, ovary, and all embryonic stages, with peak expression on day 11. Epididymal tissue expressed approximately 20% of the 11-day-old embryo level, whereas all other tissues, including mRNAs from mouse brain, eye, liver, lung, kidney, heart, skeletal muscle, smooth muscle, pancreas, thyroid, submaxillary gland, prostate, and uterus, had weak signals (Ͻ20%).
PCR Analysis of NASP Transcript Tissue Distribution-PCRs were performed with primers specific for the coding region common to both mouse sNASP and tNASP, employing a multiple tissue cDNA panel as the source of cDNA templates, including cDNAs from heart, brain, spleen, lung, liver, skeletal muscle, kidney, testis, and 7-, 11-, 15-, and 17-day embryos (Fig. 4). Agarose gel analysis of the PCR products showed the (M r 45,751). The histone binding domains are single underlined, and the nuclear translocation signal is double-underlined. Nucleotide numbers refer to the tNASP sequence. The GenBank TM accession numbers are AF034610 (tNASP) and AF095722 (sNASP). C, genomic structure of NASP encoding the alternately spliced 5Ј UTRs of somatic and testis NASP. The genomic sequence is diagrammed as the thin centerline with exons shown as wider bands. In tNASP, the 5Ј UTR includes T1a and T1b spliced onto the 5Ј end of exon one (E-1). In sNASP, the 5Ј UTR includes E-1 and an additional 49 bp, S1, not present in tNASP. A TATA box for the somatic form is indicated ϳ30 bp upstream of the transcription initiation site. presence of a 2319-bp tNASP band only in the testis (ϽT, Fig.  4A). However, all tissues, including the testis, produced a 1263-bp sNASP band (ϽS, Fig. 4A). The PCRs of testis RNA were clearly inefficient in this assay, as the Northern blots show very little sNASP compared with tNASP in the testis ( Fig. 2A), probably because of the large size of the amplified testis fragments. Therefore, PCRs were performed using the cDNA panel and primers specific for a 725-bp sequence specific to tNASP from nucleotides 334 to 1059. Agarose gel analysis of these products showed 725-bp bands only from the testis and embryos, with a strong signal on day 11 but with faint signals in the other embryonic stages (ϽT, Fig. 4B). Embryonic stage cDNAs did not give a 2319-bp band when amplified with common tNASP and sNASP coding region primers because of the inefficient amplification of the tNASP cDNA.
Western Blot Analysis of Native NASP-Western blots of lysates (equal protein loading) from mouse testis, spleen, embryos, mouse 66-2 myeloma cell line, and mouse 3T3 cells were probed with anti-NASP antibodies. A strong tNASP band of 138 kDa was present in the testis, in myeloma cells, and in 3T3 cells, with a somewhat weaker band in embryos (Fig. 5). An sNASP band of 62 kDa was present in all of the lysates (Fig. 5). These results support the data presented in Figs. 2-4, which show that tNASP is present in 3T3 cells, myeloma cells, and embryos but not in the spleen (Figs. 2-4). Only sNASP is present in spleen.
The addition of protease inhibitors to the lysates did not prevent the appearance of immunoreactive NASP breakdown products, particularly in the myeloma extract. No significant bands were seen when blots were probed with preimmune serum (Fig. 5). Recombinant 6-His-tagged tNASP also migrated at 138 kDa (data not shown). The anomalous migration of NASP and other similar nuclear proteins after SDS-polyacrylamide gel electrophoresis has been observed and discussed previously (1).
NASP mRNA Regulation during the Cell Cycle-The level of NASP mRNA was assayed in an asynchronous, exponentially growing population of HeLa cells and in double thymidineblocked cells allowed to progress in synchrony through the cell cycle. In double thymidine-locked cells at the G 1 /S border (t ϭ 0, Fig. 6A), there was little, if any, sNASP and relatively little tNASP mRNA present. However, as cells were released from the G 1 /S block and entered S-phase, increased levels of NASP mRNA were detected (t ϭ 2, 4, and 6 h, Fig. 6, A and C). Asynchronously dividing cells expressed an intermediate level of NASP mRNA (AS, Fig. 6A). As expected, histone H4 mRNA levels were low at the G 1 /S border in blocked cells and increased significantly as cells progressed through S-phase (Fig.  6A), particularly relative to the G3PDH controls. Testis and somatic NASP were detected by PCR. cDNAs included are from the heart (He), brain (Br), spleen (Sp), lung (Lu), liver (Li), skeletal muscle (Sk), kidney (Ki), and testis (Te), as well as embryonic (E) stages from day 7, 11, 15, and 17. Additional cDNAs were prepared using thymus (Th) and ovary (Ov), but these were not normalized to the expression of housekeeping genes. In both A panels, primers specific to the common coding regions of testicular (2319 bp) and somatic (1363 bp) NASP were used. In both B panels, primers specific to a 725-bp sequence of tNASP were used. Standards were: S1, X174, HaeIII digest; and S2, gt11, HindIII digest. ϽT, tNASP, ϽS, sNASP.  E11, E15, E17). B, relative image intensities were quantified using Gel Expert software. The intensities are normalized to the day 11 embryo signal. The testis signal (not shown on graph) was more than five times greater than the E11 signal.
In a second experiment, time points were taken from G 1 /S through mitosis. After 6 -6.5 h, NASP mRNA levels began to decrease and were almost undetectable by 10 h, late G 2 /Mphase (Fig. 7, A and B). Similarly, histone H4 mRNA levels began to decrease and were relatively low by 10 h (Fig. 7). In Western blots of lysates from parallel cultures of synchronous HeLa cells, NASP levels remained constant throughout the cell cycle. After release from the thymidine block, protein levels of sNASP and tNASP were not significantly different from those at t ϭ 0 (SϾ, TϾ, Fig. 6B), consistent with NASP being a relatively stable protein.
NASP and histone H4 mRNA levels were also assayed in 3T3 cells. In cells starved for serum for 48 h (t ϭ 0), very low levels of NASP mRNA were detected (Fig. 8). After serum replenishment, some cells entered S-phase at about 12 h, and the ma-jority entered by 16 h, when there were significantly increased levels of both NASP and histone H4 mRNA by 16 h (Fig. 8). Western blots of 3T3 cells growing in culture demonstrated that NASP levels remained constant throughout the cell cycle (Fig. 8B) in the same manner as observed in HeLa cells (Fig.  6B). However, CV-1 cells that had been confluent for 8 days and were not dividing showed little or no NASP present, indicating the eventual loss of NASP protein in nondividing cells (Fig.  10E, see below). To ensure the balanced loading of RNA in each well of each gel shown in Figs. 6 -8, all blots were also probed

FIG. 7. Cell cycle regulation of NASP mRNA in HeLa cells. A,
Northern blots of RNA isolated from double thymidine-blocked cells allowed to progress in synchrony through the cell cycle, probed with full-length tNASP. As in Fig. 6A, t ϭ 0, G 1 /S border. NASP mRNA was assayed over a 10-h period. S, S-phase, G 2 , G 2 -phase, M, mitosis, ϽT, tNASP, ϽS, sNASP. B, bar graph representing the relative intensities of the tNASP (ϽT) and sNASP (ϽS) bands shown in A.

FIG. 8. Cell cycle regulation of NASP expression in mouse 3T3 fibroblasts.
A, the level of NASP mRNA was assayed by probing Northern blots with full-length tNASP from asynchronously growing mouse 3T3 fibroblasts (AS) from a population of cells that were synchronized by serum starvation. In serum-starved cells (t ϭ 0), very low levels of NASP expression were detected. After serum replacement, cells entered S-phase and showed a peak of NASP mRNA by t ϭ 16 h. ϽT, tNASP, ϽS, sNASP. B, Western blot analysis of NASP expression in parallel cultures of synchronized 3T3 cells. ϽT indicates tNASP at 138 kDa, and ϽS indicates sNASP at 62 kDa.
for the housekeeping gene G3PDH. Similar G3PDH bands at all times sampled indicated an equal loading of RNA at each time period.
Localization of NASP- Fig. 9 demonstrates the localization of NASP in embryonic and somatic tissues. In the 11.5-day mouse embryo (Fig. 9, A-D) most cells in the embryo are positive for NASP. Nuclear staining was particularly evident throughout areas of dense cell proliferation as seen in the nasal process (NP) and brain (Fig. 9A) and in the cervical somites (S) and dorsal (D) root ganglia (Fig. 9C). Control sections stained with preimmune serum showed peroxidase-positive blood cells but no specific staining for NASP (Fig. 9, B and D).
In the ovary, sNASP is present in the germinal vesicle ((GV) nucleus) of primary oocytes in both primary and secondary follicles (Fig. 9E). Occasional staining was seen in oocytes within primordial follicles. During the rapid proliferation of granulosa cells in the growing follicle (F), NASP is evident in those cells adjacent to the zona pellucida (corona radiata) and those lining the developing antra (Fig. 9F).
In the spleen, sNASP is present in clustered groups of cells within germinal centers of the white pulp (W, Fig. 9H). The nuclear localization of NASP is clearly evident in these proliferating lymphocytes (Fig. 9G). NASP is also seen in cells scattered throughout the splenic cords of the red pulp (R, Fig. 9H). Control sections showed no staining (Fig. 9I).
In 3T3 cells growing in culture, NASP was primarily localized in the nucleus (Fig. 10, A-D). In these cultures, which were approximately 50% confluent, every cell contained NASP (compare Fig. 10, A and B, C and D). In contrast, cultures of CV-1 cells that had been confluent for 8 days had very few mitotic cells (compare Fig. 10, E and F) and showed little or no NASP present in either the cytoplasm or nucleus (Fig. 10E) of most cells. This finding is consistent with the loss of NASP mRNA as cells exit S-phase and slow degradation of the NASP protein in the arrested cells (Fig. 7). However, the occasional CV-1 cell that entered mitosis in these cultures showed strong staining for NASP (arrow, Fig. 10E), indicating a strong correlation between the presence of NASP and the growth state of the cells.
Histone Interaction with NASP-To characterize the functional role of NASP as a histone-binding protein in somatic cells, the histones bound to native NASP in myeloma 66-2 cells were characterized. Native NASP was isolated from the soluble fraction of mouse myeloma cells by affinity-purified anti-recombinant NASP antibody column chromatography (data not shown). Histones were extracted from the isolated native NASP complex and the histone types present in the affinitypurified NASP complex identified by reverse-phase HPLC. Based on our standard elution profiles of purified mouse myeloma cell core and linker histones, only the linker histones (variants H1a, H1b, H1c, and H1d/e) were found to co-purify with native NASP (Fig. 11A). No core histones were present in the sulfuric acid extracts. To confirm the binding of H1 subtypes, they were expressed in Escherichia coli as recombinants, purified, labeled with biotin, and used to reconstitute the complex with recombinant NASP. Fig. 11B shows that NASP was detected in the streptavidin-precipitated fraction when biotinlabeled recombinant H1c and recombinant NASP were incubated together. Conversely, by immunoprecipitation of the complexes with anti-NASP antibodies, the histone H1c protein was associated with the immunoprecipitated fraction (Fig.  11C).

DISCUSSION
In this study, we have demonstrated that the histone-binding protein NASP is present in dividing somatic cells and coupled to the cell cycle. Importantly, it appears that in myeloma cells NASP is complexed only with H1 linker histones and that the complex can be reconstructed in vitro with recombinant NASP and histones, implying that NASP functions to transport and/or store H1 histones in this cell line. sNASP is a shortened version of the testicular form of NASP, lacking two separate regions of the tNASP protein sequence (Fig. 1). This is a result of alternative splicing of the NASP pre-mRNA, skipping exons 4 and 6 (data not shown). Significantly, sNASP retains its characteristic features, namely two of the three histone binding sites, a nuclear localization signal, a leucine zipper, and an ATP/GTP binding site (36). We have further demonstrated the presence of NASP mRNA and protein in a variety of mouse somatic tissues, including mouse embryonic cells (Figs. 2-5, 9, and 10). Moreover, NASP mRNA accumula-tion is cell cycle-dependent, increasing during S-phase, concomitant with increased histone mRNA accumulation, and declining in G 2 , even in continually cycling cells (Figs. 6 -8).
Consequently, since we have previously demonstrated that NASP functions as a histone-binding protein with a nuclear localization signal (4), our present data suggest that sNASP plays an important role during the cell division cycle, providing for the more rapid and efficient transport of histones into the nucleus. Additionally, NASP might also provide for non-chromatin-bound storage of linker histones.
Additional support for the presence of somatic NASP was found from searches of GenBank TM (37), where EST sequences have been reported from various developmental stages, skin, melanoma, neuroepithelium, and mammary gland. Somatic NASP-like sequences also appear in the human EST data base, and in GenBank TM there is a partial sequence for an "autosomal cDNA variant of human testis nuclear autoantigenic sperm protein from primary human peripheral blood mononuclear cells" (GenBank TM accession number AF035191), which has the same overall structure as mouse sNASP. tNASP EST sequences have been reported from the testis, embryo, and transformed cells.
The coupling of NASP mRNA accumulation to the cell cycle (Figs. 6 -8) and the nuclear localization of NASP in dividing 3T3 cells (Fig. 10) in large lymphocytes (lymphoblasts) in the spleen (Fig. 9G) and in cells in growing embryonic tissues (Fig.  9, A and C) imply that NASP has a role in the cell's normal progression through its division cycle. This is also evident from our observation that nonmitotic CV-1 cells and nonmitotic cells in various tissues throughout the body do not have detectable quantities of NASP. Even though NASP protein levels remained constant in cultured cells when they were arrested for only 1 day, they ultimately did lose their NASP (Fig. 10, E and  F). Hence, NASP remains present in cycling cells for some period of time after cells leave the cycle, but eventually (8 days in the case of CV-1 cells) NASP is no longer detectable.
Although the DNA replication-coupled synthesis of histones has been known for many years, early workers recognized that histone gene expression could occur without DNA replication (38). Synthesis of mouse oocyte histone H4 occurs independently of DNA replication during oogenesis (12), and synthesis of testis-specific histones continues during spermatogenesis, long after DNA synthesis for meiosis has finished (4,39,40). A number of histone H1 proteins are also synthesized independently of DNA replication (11), and H1 mRNAs, particularly H1c and H1e mRNA, are present in nondividing cells (41). The decoupling of DNA and histone synthesis in certain specific cases such as spermatogenesis and oogenesis would seem to support the need for non-chromatin histone storage proteins. The accumulation of NASP is uncoupled from DNA replication during oogenesis (Fig. 9) and spermatogenesis (5), tissues in which histone mRNA and protein synthesis are also uncoupled from DNA replication. In somatic cells, NASP mRNA appears to follow the histone pattern in that it is coupled to DNA replication. Histone mRNA is regulated transcriptionally and post-transcriptionally (41)(42)(43), and NASP mRNA may also be similarly regulated. Indeed, it is quite possible that common regulatory elements exist to ensure the concomitant appearance and disappearance of both histone and NASP mRNAs as well as the persistence of both NASP and histones when they are needed in nondividing cells.
Acknowledgments-We thank Gail Grossman and the Immunohistochemistry Core Facility of the Laboratories for Reproductive Biology for their help and expertise.
FIG. 11. H1 linker histones and mouse myeloma NASP. A, reverse-phase HPLC separation and identification of mouse histones bound to affinity-purified native NASP. Total mouse myeloma H1 variants (gray trace) and NASP bound H1 variants (black trace). B, precipitation of complexes of biotin-labeled recombinant histone H1c and recombinant NASP with streptavidin followed by Western blotting with anti-rNASP antibodies. C, immunoprecipitation of biotin-labeled recombinant histone H1c and rNASP. Pre-assembled complexes (ϩ or Ϫ NASP) were immunoprecipitated with anti-rNASP antibodies. Western blotting was performed using alkaline phosphatase-conjugated avidin. Histone H1c was detected in the antibody-precipitated fraction.