Identification of histone H2A.X as a growth factor secreted by an androgen-independent subline of mouse mammary carcinoma cells.

Shionogi carcinoma 115 (SC 115) cells and Chiba subline 2 (CS 2) cells are clones of an androgen-responsive mouse tumor cell line and its autonomous subline, respectively. We have shown previously that CS 2 cells produce a heparin-binding growth factor that stimulates the growth of SC 115 cells as well as the growth of themselves. In this study, a growth factor was purified from serum-free conditioned media of CS 2 cells cultured without testosterone. A heparin-binding fraction showed growth- promoting activity on SC 115 cells and BALB/3T3 cells. The amino acid sequence analysis revealed that the components were identical to histones H2A.1 and H2A.X. Since histone H2A purified from bovine thymus had almost no growth-promoting activity on SC115 cells, histone H2A.X was assumed to be a growth factor. cDNA of histone H2A.X was cloned from a library of CS 2 cells, and its sequence was confirmed. The expressed product of histone H2A.X cDNA in Escherichia coli showed remarkable stimulatory effects on growth of SC 115 cells cultured in the absence of testosterone. These results indicate that histone H2A.X is secreted from CS 2 cells cultured without testosterone and plays a role as a growth factor.

It has been reported that the growth of some hormoneresponsive tumors is controlled by hormone-induced growth factors in the autocrine manner (1). For example, on MCF-7 cells, an estrogen-responsive human breast cancer cell line, transforming ␣-like and insulin-like growth factors are thought to mediate the estrogen-responsive growth of cancer cells (2,3). LNCaP, an androgen-responsive human prostate cancer cell line, is thought to secrete a fibroblast growth factor (FGF) 1 -like peptide in response to androgen stimuli (4). It is generally accepted that hormone-responsive tumors gradually progress to hormone-unresponsive ones; however, the mechanism regulating the growth of the latter seems to be obscure.
Shionogi carcinoma 115 (SC 115) is an androgen-responsive mouse mammary tumor (5). Recently, an androgen-induced growth factor (AIGF) secreted from SC 115 cells in the presence of testosterone was purified, and its cDNA was cloned (6). The structural analysis revealed that AIGF was a novel FGF-like growth factor, which was established as the 8th one in the FGF family. An androgen-independent subline, Chiba subline 2 (CS 2), was derived from SC 115 in our laboratory (7)(8)(9), and a clone from CS 2 cells has subsequently been maintained. We have shown previously that CS 2 cells also produce a heparin binding growth factor that stimulates the growth of SC 115 cells and CS 2 cells without testosterone (10). This factor was thought to be different from AIGF, because AIGF mRNA was not expressed in CS 2 cells (11). To shed light on growthregulatory mechanisms of hormone-unresponsive tumor cells, the present study was undertaken on purification of the heparin-binding growth factor produced by CS 2 cells and on analysis of amino acid sequence.

MATERIALS AND METHODS
Cells-SC 115 cells and CS 2 cells are clones of an androgen-dependent mouse tumor cell line and its autonomous subline, respectively. The methods for cloning and culture of these cells were described previously (12). BALB/3T3 cells were donated from the Japanese Cancer Research Resources Bank (Tokyo, Japan) and were maintained in minimum essential medium containing 10% fetal bovine serum.
Assay of Growth-promoting Activity-The assay was performed by [ 3 H]thymidine incorporation in SC 115 cells (1 ϫ 10 4 cells/well) or in BALB/3T3 cells (2 ϫ 10 4 cells/well) as described previously (10). One unit of activity is defined as half of the maximal stimulation of thymidine incorporation induced by 2 ng/ml basic FGF (R&D Systems, Minneapolis, MN) in the SC 115 cells.
Preparation of Serum-free Conditioned Media and Purification of Growth Factor-CS 2 cells (5 ϫ 10 5 cells/100-mm dish) were plated and cultured as described previously (10). Serum-free culture media from CS 2 cells were filtered through a nylon membrane (0.22 m pore size; Costar Corp., Cambridge, MA) as soon as they were obtained. The filtrate was concentrated (up to 20-fold) by ultrafiltration with M r 10,000 cut-off membrane discs (PM 10; Amicon Inc., Beverly, MA) and dialyzed against 10 mM Tris-HCl buffer (pH7.5) containing 0.1% CHAPS. The concentrated and dialyzed culture media were applied to a 10-ml heparin-Ultrogel column (IBF Biotechnics, Villeneuve-la-Garenne, France) equilibrated with 10 mM Tris-HCl buffer (pH 7.5) containing 0.1% CHAPS. The column was washed with 10 mM Tris-HCl buffer (pH 7.5) containing 0.1% CHAPS extensively until the absorbance returned to base line, and then adsorbed proteins were eluted with a 0.1-3.0 M NaCl gradient in 10 mM Tris-HCl buffer (pH 7.5) containing 0.1% CHAPS. Since the fractions around 1.0 M NaCl showed the greatest growth-promoting activity on SC 115 cells, these fractions were loaded onto a 4.6 ϫ 250-mm YMC C 4 reverse-phase high performance liquid chromatography (RP-HPLC) column (Yamamura Chemical Laboratories, Kyoto, Japan) and developed with a linear gradient of 20 -60% acetonitrile in 0.08% trifluoroacetic acid at a flow rate of 1 ml/min. Each 1 ml was fractionated, and 3 l of each fraction was assayed for [ 3 H]thymidine incorporation in SC 115 cells. Two hundred l of each fraction showing growth-promoting activity was lyophilized and electrophoresed under reducing conditions on a 15% polyacrylamide gel * This study was supported in part by Grants 06281103, 07273108, 07557070, and 06404036 from the Ministry of Education, Science, and Culture, Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Construction of cDNA Libraries and Cloning of Mouse H2A.X cDNA-Subconfluent CS 2 cells (about 5 ϫ 10 6 cells/dish) cultured in minimum essential medium/Ham's F-12 medium (1:1,v/v; serum-free medium) were scraped, and the total cellular RNA was prepared with guanidine isothiocyanate followed by centrifugation in CsCl solution (14). Poly(A) ϩ RNA was isolated by an oligo(dT)-cellulose column (Pharmacia Biotech Inc.).
Double strand cDNA was synthesized with 5 g of poly(A) ϩ RNA prepared from cultured CS 2 cells, using a random hexamer as the first strand primer, by the modified Gubler and Hoffman method (15). The cDNA was size-fractionated, ligated to an EcoRI/NotI adaptor (Pharmacia Biotech), and inserted into a gt10 vector (Stratagene, La Jolla, CA). Ligated DNA was packaged with packaging extract and introduced into host Escherichia coli. To obtain mouse H2A.X cDNA, two 32 P-endlabeled antisense oligonucleotide probes were prepared, AX-1 (5Ј-CAGCTTGTTGAGCTCCTCGTCGTTGCGGAT-3Ј) and X-1 (5Ј-TACTC-CTGAGAGGCCTGCGA-3Ј). They were specific sequences to both mouse H2A.1 and H2A.X and to mouse H2A.X, respectively. Hybridizations were carried out in 4 ϫ SSC (standard saline citrate), 2 ϫ Denhardt's solution, 40 mM sodium phosphate buffer (pH 6.5), 0.1% SDS, 100 g/ml sonicated and denatured salmon testis DNA, and with 1 ϫ 10 6 cpm/ml 32 P-end-labeled probes at 37°C for 16 h. The nylon membranes were sequentially washed in 2 ϫ SSC containing 0.1% SDS at room temperature for 1 h, at 40°C for 30 min, at 45°C for 30 min, and at 50°C for 30 min before autoradiography. DNA sequencing was done by the dideoxynucleotide chain termination procedure (16) after subcloning appropriate DNA fragments into M13 mp18 and mp19 (Takara, Kyoto, Japan) (17).
Purification of Mouse H2A.X Protein Synthesized in E. coli-Plasmid was constructed (pGEX-H2A.X) to express the mouse H2A.X protein fused with a 26-kDa glutathione S-transferase (GST) in HB101 E. coli bacteria (Toyobo, Osaka, Japan), using the EcoRI fragment (1.4-kilobase pairs) of the cloned cDNA of mouse H2A.X. After digestion, the fragment was subcloned into the EcoRI site of the bacterial expression vector pGEX2T (Pharmacia). HB101 E. coli bacteria were transformed with pGEX-H2A.X or with pGEX2T as control. Overnight cultures of E. coli transformed with pGEX-H2A.X or pGEX2T were diluted 1/10 with fresh medium and were incubated for 2 h before addition of isopropyl ␤-D-thiogalactopyranoside to a final concentration of 10 mM and a further 10-h incubation. The cells were then pelleted and resuspended in phosphate-buffered saline containing 1% Triton X-100. The cells were lysed on ice by mild sonication and then centrifuged at 10,000 ϫ g for 5 min at 4°C. At this stage, SDS-PAGE analysis showed that the expressed fusion protein occurred as an insoluble form in the pellet. The pellet was resuspended in 8 M urea and sonicated gently on ice. This solution was left for 1 h at room temperature and then centrifuged at 10,000 ϫ g for 15 min at room temperature. The supernatant was sequentially dialyzed against 6 and 2 M urea and 50 mM Tris-HCl buffer (pH 8.0). Ten g of thrombin (Boehringer Mannheim) was added to 200 l of this solution, and the mixture was allowed to react for 1.5 h at room temperature. Then 2.8 ml of 10 mM Tris-HCl buffer (pH 7.0) was added to the mixture, and this solution was dialyzed twice against 10 mM Tris-HCl buffer (pH 7.0) for 4 h. For further purification, this solution was applied to a heparin-Ultrogel column (gel bed volume, 1 ml) equilibrated with 10 mM Tris-HCl buffer (pH 7). The adsorbed proteins to this column were eluted sequentially with 10 mM Tris-HCl buffer (pH 7.0) containing 0.2 M NaCl, 0.5 M NaCl, and 1.5 M NaCl. After dialysis of each fraction against 10 mM Tris-HCl buffer (pH 7.0), the aliquot was assayed for [ 3 H]thymidine incorporation in SC 115 cells and electrophoresed under reducing conditions on 15% polyacrylamide gel with 0.1% SDS.

RESULTS
Purification of Growth Factor from CS 2 Cells-Ten liters of serum-free conditioned media obtained from CS 2 cells were used for purification of growth factor. The substances with growth-promoting activity were bound to heparin-Ultrogel and were eluted with the buffer containing 1.0 M NaCl. After the heparin-Ultrogel chromatography, bioactive fractions were subjected to RP-HPLC (Fig. 1A). The bioactive fractions obtained from the RP-HPLC showed a single major band with relative molecular masses of 17 kDa on SDS-PAGE under reducing conditions (Fig. 1B). Through these purification steps, 12 g of the purified protein was obtained from 10 liters of serum-free conditioned media, and the specific activity of fraction 37 in Fig. 1 increased up to 1.0 ϫ 10 6 units/mg (Table I). This fraction also showed as potent growth-promoting activity on BALB/3T3 cells as basic FGF (Fig. 2). Sequence analysis was performed directly with an aliquot of this fraction, but no amino-terminal amino acid was detected. Then, this fraction was digested with lysyl endopeptidase, and 13 peptides were isolated with RP-HPLC (Fig. 3). Sequence analysis of each peptide revealed that all sequences of the 13 peptides were completely identical to the sequences in either histones H2A.1 (18) or H2A.X (19) (Fig. 4). In the figure, the sequences of peptides 1 and 2 and of 3, 4, and 5 were identical to the unique carboxyl-terminal sequences of histones H2A and H2A.X, respectively. While the first 119 amino acid residues of the major histone H2A (H2A.1, H2A.2) and H2A.X are homologous (96 -97%), each histone has a unique carboxyl-terminal sequence (20). Consequently the isolated growth factors were thought to be identical to the two histones, H2A.1 and H2A.X. From the recovered amounts and the absorbance of each peptide, the ratio of the amounts of H2A.1 and H2A.X was estimated to be approximately 5 to 1. Histone H2A purified from bovine thymus (Boehringer Mannheim) was the mixture of H2A.1 and H2A.2, the amino acid sequences of which were identical to those of mouse H2A.1 and H2A.2, and did not show any growthpromoting activity on SC 115 cells (Fig. 5). From these results, histone H2A.X was assumed to be the candidate for a growth factor secreted from CS 2 cells.
cDNA Cloning of Mouse H2A.X-Cloning of mouse histone H2A.X cDNA was performed to examine the growth-promoting activity of expressed mouse histone H2A.X protein. A cDNA library was prepared from poly(A) ϩ mRNA of CS 2 cells and screened with two antisense oligonucleotide probes, AX-1 and X-1. From 4 ϫ 10 5 independent clones, 26 positive clones that were hybridized with both probes were obtained. Since their restriction enzyme maps were identical, 5 of the clones were characterized. Each of these clones was digested with EcoRI, and the resulting EcoRI fragment that was hybridized with both probes was subjected to sequence analysis. The sequence analysis revealed that all of these clones included mouse histone H2A.X cDNA.
Production of Mouse H2A.X Protein in E. coli and Its Growthpromoting Activity-To analyze the growth-promoting activity of the mouse H2A.X protein, the EcoRI DNA fragment containing mouse H2A.X cDNA was ligated into the EcoRI site of the pGEX2T expression vector. The entire mouse H2A.X genecoding region and 56-base pair 5Ј-noncoding region were translated as a fusion protein with a 26-kDa GST in E. coli. The SDS-PAGE analysis showed that the expressed fusion protein occurred as aggregates or inclusion bodies after lysing the cells (data not shown). Since the inclusion bodies were in an insoluble and inactive form, they were solubilized with 8 M urea and then refolded with dialysis. Isolation of the inclusion bodies was beneficial to purification of the expressed fusion protein from other solubilized proteins derived from E. coli. The refolded fusion protein was then digested with thrombin to remove GST, and the expressed mouse H2A.X had an additional 23 amino acids at the amino terminus. Using a heparin-Ultrogel column, the expressed H2A.X was further purified (Fig. 6). The expressed mouse H2A.X protein that was eluted from the heparin-Ultrogel column with 1.5 M NaCl showed a single band with relative molecular masses of approximately 18 kDa on SDS-PAGE, which was larger than that of H2A.X purified from CS 2 cells because of an additional 23 amino acids. The 1.5 M NaCl fraction containing expressed mouse H2A.X protein showed remarkable growth-promoting activity on SC 115 cells compared with the solution obtained from E. coli transformed with a vector plasmid, pGEX2T, alone in the same manner (Fig. 5). DISCUSSION The growth of SC 115 cells is stimulated by AIGF (6), acidic and basic FGFs (10, 11), and schwannoma-derived growth factor (21). This study showed that, in addition to these growth Fig. 1 (E) and basic FGF (q) was added to BALB/3T3 cells, and the growth-promoting activity was measured in terms of [ 3 H]thymidine uptake. Each point was based on three wells. Data are shown as means Ϯ S.E. and lie within the symbols.   factors, histone H2A.X secreted from CS 2 cells also had growth-promoting activity on SC 115 cells (Figs. 1A and 5). Moreover, histone H2A.X stimulated the growth of BALB/3T3 cells (Fig. 2).

FIG. 2. Growth-promoting activity of the purified growth factor from CS 2 cells and of basic FGF on BALB/3T3 cells. A sample of fraction 37 in
First, the growth factor secreted from CS 2 cells was purified with a heparin-Ultrogel column, and then the 1.0 M NaCl fractions which exhibited the greatest stimulating activity were applied to an RP-HPLC column (Fig. 1A). Approximately 200 l of each fraction that showed great growth-promoting activity in RP-HPLC was applied to SDS-PAGE analysis, and a single major band was obtained (Fig. 1B). Although the molecular weights of H2A.1 and H2A.X are different and the migrations of these histones on SDS-PAGE seemed to be slightly different, it was thought that the staining band of H2A.X was covered over with that of H2A.1 on the SDS-PAGE analysis because the amount of H2A.1 in the purified fraction was much larger than that of H2A.X, and the gel of SDS-PAGE was strongly developed after silver staining to confirm that the other minor proteins were not contained in this fraction. Subsequent sequence analysis after lysyl endopeptidase digestion showed that no proteins other than H2A.1 and H2A.X were present in the fraction (Figs. 3 and 4). Because histone H2A.1 did not show any growth-promoting activity on SC 115 cells (Fig. 5), histone H2A.X was assumed to be a growth factor secreted from CS 2 cells. The expression experiment confirmed that histone H2A.X is one of the heparin-binding growth factors secreted from CS 2 cells (Figs. 5 and 6).
Histones are small, highly basic proteins that associate with the DNA to form nucleosomes and play a fundamental role in organizing chromatin architecture by compacting DNAs. As for histone H2A, one of the core histones, at least four species of isoproteins are detected, H2A.1, H2A.2, H2A.X, and H2A.Z (20,22). There have been some reports that H2A plays a role as a gonadotropin-releasing hormone-binding inhibitor (23) or as a homeostatic thymus hormone with histone H2B (24). Therefore, the histone H2 group might show some biological functions in addition to its main function in organizing chromatin architecture by compacting DNA. As for human H2A.X (25), the carboxyl-terminal amino acid sequence (Ser-Gln-Glu) of H2A.X is homologous with those of several species of lower eukaryotes, e.g. Saccharomyces cerevisiae H2A.1 and H2A.2 (26) and Aspergillus nidulans H2A (27). In lower eukaryotes, this type of histone is estimated to compose a large fraction of the H2A proteins, and their chromatin largely consists of a transcriptionally active region (28). These facts suggest that the carboxyl-terminal amino acid sequence (Ser-Gln-Glu) of H2A.X plays some role in active chromatin and might influence regulation of proliferation.
It has been reported that H2A.X has two mRNAs of 0.5 and 1.4 kilobases, which are transcribed from a single gene and have different stabilities throughout the cell cycle (19,29). The amount of the shorter mRNA is coupled to DNA synthesis. It is synthesized just after the beginning of S-phase and degraded just after the end of S-phase. On the other hand, the longer mRNA is very stable through the cell cycle. The two mRNAs are present in spleen, thymus and testes of mice and several tissue culture cell lines. The 1.4-kilobase transcript is detected strongly in embryonic cells such as undifferentiated F9 teratocarcinoma cells. These results imply that H2A.X plays a role of not only organizing chromatin architecture during S-phase but also growth regulator through the cell cycle in these tissues and embryonic cells.
There have been some reports about the heparin-binding growth factors that do not belong to the FGF family. As one of these heparin-binding growth factors, a hepatoma-derived growth factor was purified from the conditioned medium of human hepatoma-derived cell line, HuH-7 (30). Its primary sequence shares homology with high mobility group-1 protein.
It was also reported that brain heparin-binding protein (amphotericin), which enhances neurite outgrowth in cerebral neurons, was identical to high mobility group-1 and was found at relatively high levels in transformed cells (31,32). These growth factors do not have any homology with the FGF family and with histones. Histone H2A.X, shown as one of the heparin-binding growth factors in the present study, is not in the FGF family. These results indicate that some of these chromosomal heparin binding proteins have a role to regulate cell cycle.
Since H2A.X has no signal sequence for secretion, it is not yet clear how H2A.X is secreted from the cells. However, hydrophobic amino acids are abundant from the middle of the molecule to the carboxyl terminus of H2A.X, and it is considered that this region is an internal signal sequence (33). It is now important to clarify the mechanism by which H2A.X acts on cell growth.