Chicken Histone Deacetylase-2 Controls the Amount of the IgM H-chain at the Steps of Both Transcription of Its Gene and Alternative Processing of Its Pre-mRNA in the DT40 Cell Line*

Histone deacetylases (HDACs) are involved in the deacetylation of core histones, which is an important event in transcription regulation in eukaryotes through alterations in the chromatin structure. We cloned cDNAs and genomic DNAs encoding two chicken HDACs (chHDAC-1 and -2), which are preferentially localized in nuclei. Treatment with trichostatin A reduced the HDAC activities in immunoprecipitates obtained with anti-chHDAC-1 and -2 antisera. Using gene targeting techniques, we generated homozygous DT40 mutants, ΔchHDAC-1 and -2, devoid of two alleles of the chHDAC-1and -2 genes, respectively. The protein patterns on two-dimensional PAGE definitely changed for ΔchHDAC-2, and the amounts of the IgM H- and L-chains increased in it. Of the two IgM H-chain forms, the secreted form (μs) increased in ΔchHDAC-2, but the membrane-bound form (μm) decreased. The IgM H-chain gene was transcribed more in ΔchHDAC-2 than in DT40 cells. In the mutant, the alternative processing of IgM H-chain pre-mRNA preferentially occurred, resulting in an increase in the amount of μs mRNA, whereas the stability of the two types of mRNA, μs and μm, was unchanged. In DT40 cells, treatment with trichostatin A increased both the amounts of IgM H-chain mRNAs and the switch from μm to μs mRNAs. Based on these results, we propose a model for a role of chHDAC-2 in both the transcription and alternative processing steps, resulting in control of the amount of the μs IgM H-chain in the DT40 cell line.

Alterations in the chromatin structure have been thought to participate in the regulation of gene expression in eukaryotes. In recent years, knowledge concerning the involvement of the acetylation of core histones in the regulation of gene expression has rapidly accumulated (1)(2)(3). This covalent modification, that reduces how tightly histones are associated with DNA in the chromatin (nucleosomes), should be precisely controlled through the cooperative actions of histone acetyltransferase(s) and deacetylase(s) (formerly HD, renamed HDAC 1 according to the GDB Genetic Nomenclature Guide). First, unexpectedly, a Tetrahymena histone acetyltransferase enzyme was reported to be very similar to a yeast protein called Gcn5p (general controlled nonrepressed protein) (4). Successively, a large cellular protein (p300/CBP) that forms a complex with certain transcription factors, a p300/CBP-associated factor, and a protein (TAF II 230/250) that is part of a large transcription factor (TFIID) were found to possess the ability to acetylate histones (5)(6)(7). These histone acetyltransferases have somewhat different substrate specificities; i.e. p300/CBP can acetylate all core histones (H2A, H2B, H3, and H4), but Gcn5p can only add acetyl groups to H3 and H4 (8).
In addition, with regard to HDAC too, exciting results were reported in rapid succession (2,3,9,10). The yeast transcriptional corepressors, Sin3 and Rpd3, negatively regulate the global expression of genes (11). One human Rpd3 homolog is HDAC-1, one of the reported human HDACs (12). Furthermore, in higher eukaryotes, transcriptional repression by Mad-Max heterodimers requires the interaction of Mad with mSin3, the mammalian homolog of Sin3 (13)(14)(15). The corepressor, mSin3, exists in a complex with HDAC-1 and -2 (16). Both the Mad repression and HDAC activity of mSin3 immunocomplexes are abolished by trichostatin A (TSA), a specific inhibitor of HDAC (17). The transcriptional corepressors, SMRT and N-CoR, act as silencing mediators for retinoid and thyroid hormone receptors (18 -20). Transcriptional repression by nuclear receptors is mediated by a complex containing SMRT or N-CoR, mSin3, and HDAC. Recently, mammalian HDAC-1 and -2 were reported to be recruited by the retinoblastoma protein to repress transcription (21,22). These results, therefore, indicate that a complex containing HDAC deacetylates nucleosomal core histones, producing alterations in the chromatin structure that block transcription.
Chickens appear to have one of the lowest histone gene repetitive frequencies among higher eukaryotes. Thirty-nine of the 44 core and H1 genes belong to a major gene cluster of 110 kb, and the 42 sequenced genes encode a set of amino acid sequences, as follows: six H1 variants, three H2A variants, four H2B variants, two H3 variants, and an H4 protein (23). In addition, the DT40 chicken B cell line incorporates foreign DNA by targeted integration at frequencies similar to those for random integration (24,25). Using gene targeting techniques, therefore, we generated several DT40 mutants, respectively, devoid of one or two particular histone genes (26 -28), one allele of the gene cluster of 110 kb (29), an approximately 57-kb segment of the cluster (30), and 11 of the 12 H1 gene copies (31). Systematic analyses of the resultant mutants led us to some notable conclusions, as follows (32). First, all of the histone gene families have the inherent ability to compensate for the deletion of approximately half of their own constituents and to maintain the amount of each of the histone subtypes in stoichiometric balance, based on increases in the expression of the remaining genes, regardless of the disruption of one allele of the major gene cluster or approximately half-segments of the two alleles (29,30). Therefore, one allele of the major histone gene cluster is enough for cell proliferation. Second, the deletion of 11 of the 12 H1 gene copies does not affect cell functions, i.e. the growth rate and global chromatin structure, indicating that only one copy of the H1 genes is enough for cell proliferation (31). Third, the protein patterns are altered in the mutants, respectively, lacking particular H1 and H2B variants (26,28,30,31), and the variable proteins are specific for the corresponding mutants, suggesting that the H1 and core variants should be individually involved in regulation of the expression of specific genes. These results, together with those obtained for a number of different genes (33)(34)(35), suggest that the DT40 cell line should be most useful for an in vivo understanding of the roles of important genes in higher eukaryotes.
In this study, we first cloned cDNAs and genomic DNAs encoding two chicken HDACs, chHDAC-1 and -2, which are preferentially localized in nuclei. The HDAC activities in chH-DAC-1 and -2 immunoprecipitates decreased on treatment with TSA. Next, we generated two homozygous DT40 mutants, ⌬chHDAC-1 and -2, devoid of two alleles of chHDAC-1 and 2, respectively. The protein patterns on two-dimensional PAGE for ⌬chHDAC-2 were obviously distinct from those for the wildtype cell line. The amounts of IgM H-and L-chains, respectively, accumulated in the mutant. In ⌬chHDAC-2, the secreted form (s) of the IgM H-chain increased, but the membranebound form (m) of it decreased. The IgM H-chain gene was transcribed more in ⌬chHDAC-2 than in DT40 cells. Furthermore, the alternative processing from m to s mRNA of IgM H-chain pre-mRNA preferentially occurred in the mutant, resulting in increased amounts of s mRNA, whereas the stabilities of these two types of mRNAs did not change. The increases in both the amounts of IgM H-chain mRNAs and the switch from m to s mRNA were mediated by TSA in DT40 cells. Based on these results, we propose a model for the dual participation of chHDAC-2 in the transcription and alternative processing steps, resulting in the control of the amount of the s IgM H-chain in the DT40 cell line.

EXPERIMENTAL PROCEDURES
Cell Cultures-DT40 cells and all subclones were cultured essentially as described (24,25). To study the stability of IgM H-chain mRNA, exponentially growing DT40 subclones were further cultured for various times in medium containing 4 g of actinomycin D per ml (30). To study the effect of TSA, DT40 cells were further cultured for 24 or 48 h in the presence of the drug at 0, 500, or 1000 nM.
Cloning of cDNAs and Genomic DNAs of chHDAC-1 and -2-Based on conserved amino acid sequences (IDIDIHH and WTYETAV) in Xenopus, human, and mouse HDAC-1s deduced from their cDNAs (12,36,37) and sense and antisense degenerate oligonucleotide primers, respectively, containing sequences 5Ј-AT(T/C/A)GA(T/C)AT(T/C/A)GA(T/ C)AT(T/C/A)CA(T/C)CA-3Ј and 5Ј-ACIGT(T/C)TC(A/G)TAIGTCCA-3Ј, were constructed. A PCR product of 437 bp, corresponding to a part of the coding regions of mammalian HDAC cDNAs, was prepared from DT40 cDNAs using the two degenerate primers. To obtain full-length chHDAC cDNAs, using the resultant PCR product as a probe, we screened a DT40 ZAP II cDNA library, constructed essentially as described (38), under low stringency conditions. The entire cDNA sequences of both strands of two different full-length cDNAs were sequenced by the dye terminator method (Applied Biosystems Division, Perkin-Elmer).
Using the AflII/FspI fragment of 1.0 kb of chHDAC-1 cDNA, which is highly homologous to that of chHDAC-2 cDNA, as a probe, two different genomic DNAs were isolated on the screening of a DT40 genomic DNA library in a phage, which was also constructed as described (38). The organization of the resultant genomic DNAs was determined according to the PCR-sequencing protocol (Amersham Pharmacia Biotech).
Generation of Antibodies against chHDAC-1 and -2-To generate antisera specific for chHDAC-1 and -2, cDNA fragments encoding two C-terminal regions, amino acids 372-480 of chHDAC-1 and 373-488 of chHDAC-2, were subcloned into the pGEX-2T plasmid (CLONTECH) in frame, and then chHDAC-1 and -2 C-terminal peptide-GST fusion proteins were synthesized in E. coli, extracted, and purified to more than 95% homogeneity. According to a standard immunization protocol, New Zealand White rabbits were immunized, and then preimmune and immune sera were collected. To remove the anti-GST antibodies, the crude antisera were passed through a column of GST bound to glutathione-agarose beads. To avoid the cross-reaction between chHDAC-1 and -2, anti-chHDAC-1 and -2 antisera were inversely passed through columns of chHDAC-2 and -1 C-terminal peptide-GST-conjugated glutathione-agarose beads, respectively.
The in vitro HDAC activity in the immunoprecipitates obtained with anti-chHDAC-1 or -2 antiserum was assayed essentially as described (12). Cells (4 ϫ 10 7 ) were lysed in 2 ml of phosphate-buffered saline containing 1% Nonidet P-40, and the aliquots (500 l) of the extracts were immunoprecipitated and washed with HD buffer (20 mM Tris-HCl, pH 8.0, 150 mM NaCl, 10% glycerol). The resultant immunocomplexes were incubated for 1 h at 37°C with 100 g (2 ϫ 10 5 dpm) of acidsoluble histones isolated from [ 3 H]acetate-labeled DT40 cells. The released [ 3 H]acetic acid was extracted with ethyl acetate and then measured. Pretreatment of immunoprecipitated samples with 100 nM TSA was performed for 2 h at 4°C, prior to the addition of the labeled histones.
Immunofluorescence Microscopy and GFP Fusion Protein Experiments-Immunofluorescence experiments were carried out with a tyramide signal amplification system (NEN Life Science Products) according to the manufacturer's protocol. DT40 cells on slides were fixed in 4% formalin for 20 min at room temperature. Primary antibodies (rabbit anti-chHDAC-1 or -2 antiserum) were incubated at a 1:100 dilution for 1 h in a moist chamber, followed by incubation for 1 h with horseradish peroxidase-conjugated goat anti-rabbit IgG (Dako) as a secondary antibody at a 1:1000 dilution. After incubation with fluorescein isothiocyanate-conjugated tyramide (FITC), the cells were stained with 4Ј,6diamidino-2-phenylindole. The FITC-labeled chHDAC-1 and -2 and nuclei were examined by fluorescence microscopy. The same fields were also photographed.
To obtain the pchHDAC-1/neo or pchHDAC-1/hisD construct, we ligated the 1.5-kb AccI/BamHI fragment, corresponding to the upstream of exon 2 of chHDAC-1, to the 5Ј-end of the cassette carrying neo or hisD and then the 3.0-kb Eco47III fragment, corresponding to the 3Ј part of exon 5 to exon 10 of the gene, to the 3Ј-end of either cassette. In the resultant targeting vectors, therefore, genomic sequences, corre-sponding to exons 2-5 of chHDAC-1, were replaced with the neo or hisD-carrying cassette. To obtain the pchHDAC-2/neo or pchHDAC-2/ hisD construct, we replaced the StuI fragment between exon 3 and intron 6 of chHDAC-2 of the parental plasmid, which contains the 7.0-kb BamHI fragment covering exons 2-10 of the gene, with the cassette carrying neo or hisD.
Probe A was the XhoI/AccI fragment of 0.5-kb derived from the upstream of exon 1 of chHDAC-1. Probe D was the 0.5-kb fragment derived from the upstream of the BamHI site of exon 2 of chHDAC-2. Probes B and C were derived from neo and hisD, as described (26,28).
Transfection was carried out essentially as described (24,26). To obtain chHDAC-1-deficient mutants, transfectants with the pchHDAC-1/neo construct were selected in medium containing 2 mg of G-418 per ml. We transfected the pchHDAC-1/hisD construct into clones in which one of two chHDAC-1 alleles had already been disrupted and selected stable transfectants in medium containing 2 mg of G-418 and 800 g of histidinol per ml, respectively. To disrupt chHDAC-2, the pchHDAC-2/ neo and pchHDAC-2/hisD constructs were used, instead of the pchH-DAC-1/neo and pchHDAC-1/hisD constructs.
Southern Blotting-Genomic DNAs were digested with the indicated enzymes, separated in a 0.8% agarose gel, transferred to a Hybond Nϩ membrane, and then hybridized with 32 P-labeled probe A, B, C, or D, as described (44).
Two-dimensional Gel Electrophoresis and Amino Acid Sequence Analysis-Total cellular proteins were prepared from exponentially growing DT40 subclones (38) and separated with an automated apparatus for two-dimensional PAGE, Multiphor II (Amersham Pharmacia Biotech), as described (29,31). Immobiline DryStrip (pH 4 -7) and ExcelGel XL SDS (gradient 12-14) gels were used for the first dimensional IEF and second dimensional SDS-PAGE, respectively. The SDS-PAGE gel was stained by the fluorostaining method with SYPRO Red as recommended by the supplier (Molecular Dynamics, Inc., Sunnyvale, CA).
Amino acid sequences were determined as follows. A large amount of total cellular proteins was prepared from the chHDAC-2-deficient mutant, subjected to two-dimensional PAGE (29,31), and then electrotransferred to a polyvinylidene difluoride membrane. The blotted membrane was briefly stained with Coomassie Blue. After extensive washing with water, membrane pieces containing the proteins of interest were excised and then examined with a Procise 492 gas phase sequencer (Applied Biosystems Division, Perkin-Elmer).
The amounts of the s IgM H-chain were measured as follows. The final culture media of DT40 subclones were used as conditioned media, containing the s IgM H-chain secreted. Cells (1 ϫ 10 7 ) were treated with 10% trichloroacetic acid and then lysed in 100 l of SDS buffer. Following the addition of 100 l of loading buffer, aliquots (4 l) of 1:50 dilutions of the resultant whole cell extracts, together with the conditioned media, were analyzed, essentially as mentioned above, except for the use of 7.5% SDS-PAGE and goat anti-chicken IgM -chain antibodies (Betyl Laboratories Inc.).
Staining for the m IgM H-chain on the Surface of DT40 Subclones-Staining for the presence of the m IgM H-chain on the surface of DT40 subclones was performed, essentially as described (45). Cells (1 ϫ 10 5 ) were incubated for 30 min on ice in 200 l of staining buffer (phosphatebuffered saline, 2% bovine serum albumin, 0.02% sodium azide, 0.5 mM EDTA), containing a 10 g/ml concentration of the goat anti-chicken IgM -chain antiserum. After washing, the cells were incubated for 30 min on ice in a 1:1000 dilution of FITC-labeled rabbit anti-goat IgG (American Qualex). After further washing, the cells were analyzed with a FACS Vantage instrument (Becton-Dickinson).
RNase Protection Method-For the quantification of total IgM Hchain mRNAs, the amplified 442-bp PCR fragment, corresponding to the V, D, and J regions of the IgM H-chain gene (46 -48), was inserted into the EcoRV site of Bluescript II. [ 32 P]CMP-labeled antisense RNAs were synthesized with phage T7 RNA polymerase after cutting with BamHI and used as probe VH. This antisense RNA probe consisted of a 388-nt fragment of the IgM H-chain gene and an 84-nt fragment from the plasmid vector.
To separately measure the amounts of s and m IgM H-chain mRNAs, an antisense RNA probe was prepared as follows. A PCR fragment of 258 bp, containing the 221-bp 4 and 37-bp s poly(A) regions of the IgM H-chain gene, was inserted into the EcoRV site of Bluescript II. 32 P-Labeled antisense RNAs were synthesized with phage T7 RNA polymerase after cutting with BamHI and used as probe Cmu. This probe was 342 nt long, since it consisted of a 258-nt fragment from the IgM H-chain gene and an 84-nt fragment from the vector.
The steady-state levels of total, s, and m IgM H-chain mRNAs were determined by the RNase protection method with 32 P-labeled RNA probes and an Ambion RPAII kit according to the manufacturer's protocol as described (26,27). After electrophoresis in a denaturating polyacrylamide gel, autoradiography was carried out. The intensities of the radioactivity of protected fragments were quantified with a Fuji BAS 1000 Image Analyzer and normalized as to those of protected fragments for GAPDH (28,49).
Nuclear Run-off Transcription Assay-The nuclear run-off transcription assay was performed essentially as described (50). The nuclei were isolated from DT40 subclones and incubated with 32 P-labeled UTP and three unlabeled XTPs, and then the 32 P-labeled RNAs were purified. IgM H-chain mRNAs were detected by hybridizing the 32 P-labeled RNAs to DNA fragment VH, corresponding to the V, D, and J regions, spotted onto a nitrocellulose membrane.

RESULTS
Cloning of cDNAs of chHDAC-1 and -2-To identify unequivocally the chicken HDACs as novel enzymes, we have cloned and sequenced their cDNAs. Based on conserved amino acid sequences in the Xenopus (37), human (12), and mouse (36) HDAC-1 s, we first prepared the 437-bp fragment, corresponding to a part of cDNAs encoding the mammalian enzymes, by PCR using cDNAs from DT40 cells with degenerate primers, i.e. a sense primer and an antisense primer. Our screening, using the resultant PCR product, of a DT40 ZAP II cDNA library yielded 41 positive cDNA clones. These clones were classified into three groups based on their restriction enzyme patterns. Sequence analyses of the two largest cDNA inserts of two of the three groups, designated as clones 2 and 20, revealed that they, respectively, contained both an initiation codon and a 3Ј poly(A) tail and thus appeared to contain the full-length sequences of two different chicken HDAC cDNAs. The amino acid sequences deduced from nucleotide sequences of these cDNAs are shown, together with those of other organisms, in Fig. 1. The proteins encoded by clones 20 and 2, respectively, comprise 480 and 488 amino acid residues, including a putative initiation Met, and exhibit 93.8 and 97.1% identities in amino acid sequence to human HDAC-1 and -2. The consensus sequence, SDKRI(S/A)CDEEFSDSE, for the retinoblastoma protein binding site (21,22) is located at positions 410 -425 in both proteins. Thus, these two proteins are chicken homologs of HDAC-1 and -2 and are designated as chHDAC-1 and -2. Within the almost constant background, a considerable difference was observed in the C-terminal sequence of approximately 50 amino acids.
In immunofluorescence microscopy experiments, involving the antibodies against chHDAC-1 and -2, we examined their subcellular localization in DT40 cells. As shown in Fig. 3A, chHDAC-1 and -2 were preferentially localized in nuclei. To confirm these results, we constructed chimeric plasmids expressing the C-terminal regions of chHDAC-1 and -2 fused to GFP and transiently expressed the GFP-chHDAC-1C and -2C fusion proteins in DT40 cells. As shown in Fig. 3B, inspection of the resultant cells by fluorescence microscopy revealed that chHDAC-1 and -2 were localized in nuclei.
Organization of Genes Encoding chHDAC-1 and -2 in the Genome-Using the 1.0-kb AflII/FspI fragment of chHDAC-1, which exhibits high homology to the nucleotide sequence of chHDAC-2, we screened a DT40 genomic library in a phage and isolated two different genomic DNAs specific for chH-DAC-1 and -2, respectively. Detailed analyses, involving the PCR method followed by sequencing, revealed that the genomic DNAs of both chHDAC-1 and -2 comprise 14 exons and are 12-14 kb in length, as schematically shown in the upper panels in Fig. 4, A and C. Moreover, sequence analyses also showed that the 5Ј-upstream regions of these two genes were rich in GC, and typical TATA boxes were not observed in these regions (data not shown).
Generation of Homozygous chHDAC-1-and chHDAC-2-Deficient Mutants-To generate chHDAC-1-deficient mutants, we first transfected DT40 cells with the pchHDAC-1/neo construct (Fig. 4, A and B). As expected, after integration of this targeting vector into the chHDAC-1 locus, in one of the stable transfectants selected with neomycin (Geneticin; G-418), probe A, originating from the left side of exon 1, hybridized to the 6.7-kb BamHI fragment, in addition to the endogenous 4.5-kb BamHI fragment. Probe B, derived from neo, hybridized to the 6.7-kb BamHI fragment. One of the G-418-resistant clones (ϩ/Ϫ) was chosen for transfection of the pchHDAC-1/hisD construct. As expected, after integration of the targeting vector into the remaining chHDAC-1 allele, in two of the analyzed clones Ϫ/Ϫ, probe A newly hybridized to the 18.0-kb BamHI fragment, in addition to the 6.7-kb BamHI fragment, and probe C derived from hisD also hybridized to this 18.0-kb fragment.
To disrupt chHDAC-2, the pchHDAC-2/neo and pchHDAC- FIG. 3. Subcellular localization of  chHDAC-1 and -2. A, immunofluorescence experiments. DT40 cells were prepared for indirect immunofluorescence detection of chHDAC-1 or -2 by formalin fixation and probed with rabbit anti-chH-DAC-1 or -2 antiserum as the primary antibody, followed by the use of horseradish peroxidase-conjugated goat anti-rabbit IgG as the secondary antibody. Preimmune serum was used as a control. Cells were incubated with fluorescein isothiocyanate-conjugated tyramide and then stained with DAPI to visualize DNA. The same fields were photographed. B, GFP fusion protein experiments. DT40 cells were transfected with two chimeric plasmids (pGFPchHDAC1C and pGFPchHDAC2C), expressing GFP-chHDAC-1 and -2 C-terminal peptide fusion proteins, respectively, fixed in formalin, and then stained with DAPI. The GFP-chHDAC-1 and -2 fusion proteins and nuclei were examined by fluorescence microscopy. 2/hisD constructs were sequentially introduced into DT40 cells essentially as described above (Fig. 4, C and D). As expected, in the clone ϩ/Ϫ, probe D, originating from the left side of exon 2, newly hybridized to the 12.6-kb EcoRV fragment in addition to the endogenous 14.0-kb EcoRV fragment. Probe B also hybridized to the 12.6-kb EcoRV fragment. In the case of clones Ϫ/Ϫ, probe D newly hybridized to the 15.0-kb EcoRV fragment, with the disappearance of the endogenous 14.0-kb fragment. Probe B gave no additional bands, but probe C newly hybridized to the 15.0-kb EcoRV fragment.
To determine whether or not chHDAC-1 and -2 were really disrupted in the corresponding DT40 mutants, we measured the steady-state levels of chHDAC-1 and -2 mRNAs by Northern blotting using probes chHDAC-1-3Ј and -2-3Ј, corresponding to the 3Ј-regions of chHDAC-1 and -2 cDNAs, respectively (Fig. 2D). The steady-state levels of chHDAC-1 mRNAs in the heterozygous 1/2⌬chHDAC-1 mutant, appearing as one band (denoted by chHDAC-1), were about 50% those in DT40 cells. For the homozygous ⌬chHDAC-1 mutant, no band was detected. On the other hand, the steady-state levels of chHDAC-2 mRNAs in the heterozygous 1/2⌬chHDAC-2 mutant, appearing as one band (denoted by chHDAC-2), were about 90%, instead of 50% those in DT40 cells, indicating that the chHDAC-2 gene, like histone genes (32), should have the ability to compensate for the disruption of one allele of the gene. No transcript of chHDAC-2 was detected in the homozygous ⌬chHDAC-2 mutant. As expected, the disappearance of chHDAC-1 or -2 in ⌬chHDAC-1 or -2 was confirmed by Western blotting with the corresponding antibodies ( Fig. 2A). In addition, the radioimmunoprecipitation experiments revealed that the positive 52and 50-kDa proteins, which were detected in the chHDAC-1 and -2 immunoprecipitates of DT40 cells, were not detected in ⌬chHDAC-1 and -2, respectively (Fig. 2B). The polypeptides that were co-precipitated with the chHDAC-1 and -2 immunocomplexes in DT40 cells also disappeared in ⌬chHDAC-1 and -2, confirming the results obtained in block ϩ experiments. Finally, as shown in Fig. 2C, the HDAC activities present in the anti-chHDAC-1 and -2 immunoprecipitates prepared from DT40 cells decreased to background levels in ⌬chHDAC-1 and -2.
Differential Changes in the Protein Patterns on Two-dimensional PAGE of ⌬chHDAC-1 and -2-We first examined the possible influence of the disruption of chHDAC-1 or -2 on both the growth rate and global chromatin structure. The growth rates of all the mutants, 1/2⌬chHDAC-1 and -2, and ⌬chH-DAC-1 and -2, were essentially identical with that of the wildtype cell line, indicating that the deletion of either of chH-DAC-1 or -2 had no effect on the growth rate of DT40 cells (Fig.  4E). Similarly, the nucleosome ladders of all of the mutants examined were virtually the same as those of DT40 cells, indicating that the global chromatin structure was not altered, even when either chHDAC-1 or -2 was disrupted (data not shown).
To clarify the role of chHDAC-1 or -2 in vivo, we next compared the total cellular proteins of ⌬chHDAC-1 or -2 with those of the wild-type cell line. As shown in Fig. 5A, the electrophoretic patterns on two-dimensional PAGE of the proteins from ⌬chHDAC-1 were virtually the same as those for the control cell line, except for the increase in protein 1, indicating that the disruption of chHDAC-1 resulted in insignificant changes in the amounts of most major endogenous proteins under the culture conditions used. On the other hand, several noticeable alterations were observed in ⌬chHDAC-2 (Fig. 5B). Some proteins (numbered 2-7), in addition to protein 1, newly appeared or clearly increased in this mutant. The proteins that varied did not correspond to chHDAC-1 and -2, or to the two exogenous products of 56 and 50 kDa from neo and hisD.
Similar results were obtained for three other clones each of the chHDAC-1-and chHDAC-2-deficient mutants, respectively, which were independently generated, indicating that these variations in the protein patterns were not simply due to clonal deviation. Thus, our results established that the disruption of chHDAC-2 caused increases in the amounts of a set of numbered proteins.
Accumulation of the IgM H and L-Chains in ⌬chHDAC-2-To characterize the proteins that increased in ⌬chHDAC-2, we determined their amino acid sequences. A large amount of total cellular proteins was prepared from the mutant, resolved on two-dimensional PAGE, electroblotted, stained, and then destained. The N-terminal amino acid sequences of the excised protein spot 3-6 materials were directly determined, as shown in Table I.
The N-terminal amino acid sequences of proteins 3 and 4 were, respectively, virtually identical with the stretch comprising positions 22-40 of the amino acid sequence predicted from the nucleotide sequence of the chicken IgM L-chain (51), and the stretch comprising positions 1-21 of this deduced amino acid sequence should comprise a signal peptide. The alteration in the amino acid sequence (Ala instead of Ser at position 27 of proteins 3 and 4) should be due to gene conversion, which continues at a high rate in the DT40 cell line (45). In addition, the difference in molecular weight between proteins 3 and 4 (see Fig. 5B) should be due to proteolytic cleavage(s) in the C-terminal region of the former.
The N-terminal amino acid sequence of protein 5 was completely identical with the stretch comprising positions 20 -39 of the amino acid sequence deduced from the nucleotide sequence of the chicken IgM H-chain (52). The stretch, positions 1-19, prior to position 20 of this deduced amino acid sequence of the IgM H-chain also corresponds to a putative signal peptide. On the other hand, the N-terminal amino acid sequence of protein 6 was slightly distinct. The amino acid residue corresponding to position 26 was Gln, instead of Ser, and the residue corresponding to position 29 could not be determined but should be Cys or Trp. This undefined amino acid residue was followed by the sequence Gln-Thr-Pro, which was identical with the stretch comprising positions 31-33 of protein 5 (and the IgM H-chain). Thus, these results indicated the co-occurrence of both the substitution and deletion of amino acids surrounding position 30 of protein 6, probably due to the gene conversion mentioned above.
Moreover, we determined the amino acid sequences of proteins 1, 2, and 7, but they exhibited no homology to any known proteins (data not shown).

Increase or Decrease in the Secreted Form or Membranebound Form of the IgM H-chain in ⌬chHDAC-2-It has been
reported that there are two types of the IgM H-chain, the secreted (s) and membrane-bound (m) forms, during the B cell development in various organisms, including the chicken (46 -48, 53, 54).
To determine which form, s or m, is preferentially related to the accumulation of the IgM H-chain in ⌬chHDAC-2, after a final cell passage, we performed Western blotting of both whole cell extracts and conditioned media. As shown in Fig. 6A, in ⌬chHDAC-2 (cl-33-12 and cl- , the cytoplasmic level of the IgM H-chain increased, the amount being about 10-fold higher than that in DT40 cells. These results were consistent with those obtained on two-dimensional PAGE mentioned above. In addition, the amount of the IgM H-chain (probably the s form) in the conditioned medium from ⌬chHDAC-2 was about 3 times that from DT40 cells (Fig. 6B).
To examine the alteration in the amounts of the m IgM H-chain on the surface of the DT40 and ⌬chHDAC-2 cell lines, we next performed FACS analysis on both cell lines. As shown in Fig. 6C, conversely, in ⌬chHDAC-2 the amounts of the m IgM H-chain on the surface decreased, the magnitude of the decrease being about 70%.
Increases of Total IgM H-chain mRNA and s mRNA or Decrease of m mRNA in ⌬chHDAC-2-We first measured the steady-state levels of total IgM H-chain mRNA in the DT40 and ⌬chHDAC-2 cell lines. Total RNAs were analyzed by the RNase protection method using antisense RNA probe VH, correspond-ing to the 388-bp V, D, and J regions of IgM H-chain pre-mRNA. Since probe VH is common for the coding regions of s and m mRNAs, it could completely protect the 5Ј-coding sequences of 388 nt of the two types of mRNAs (Fig. 7A). As shown in Fig. 7B, the steady-state level of total IgM H-chain mRNA increased in ⌬chHDAC-2, the increase being about 2.5fold, but it was unchanged in ⌬chHDAC-1.
On the other hand, probe Cmu, corresponding to the 258-bp 4 and s regions of IgM H-chain pre-mRNA, should differentially protect the 3Ј-coding sequence of 258 nt of s mRNA and that of 221 nt of m mRNA, respectively (Fig. 7A). In DT40 cells, as expected, this probe could distinguish s mRNA, appearing as a band of 258 nt, from m mRNA, appearing as a band of 221 nt (see Fig. 7C). Moreover, the use of probe Cmu revealed that in DT40 cells the amount of s mRNA was approximately half that of m mRNA. Although no difference was observed in ⌬chHDAC-1, analysis of ⌬chHDAC-2 with this probe gave two noticeable results, as follows. In the mutant, the steady-state level of m mRNA decreased, the magnitude . SDS-PAGE in the second dimension was performed on an ExcelGel XL SDS gel (gradient [12][13][14], followed by the fluorostaining method. Proteins 1-7 are the proteins that increased in the mutants, and the regions containing them are magnified. of the decrease being about 30%. Conversely, the steady-state level of s mRNA increased, the increase being about 8-fold. Thus, the sum of m and s mRNAs in ⌬chHDAC-2 was about 2.5 times that in DT40 cells, and this result virtually agreed with that obtained for total IgM H-chain mRNA mentioned above. These results, together, showed that the ratio of s to m mRNAs in the mutant increased, the magnitude of the increase being about 8-fold. Increased Expression of the IgM H-Chain Gene in ⌬chHDAC-2-To determine whether or not the IgM H-chain gene is transcribed more in ⌬chHDAC-2 than in DT40 cells, we first carried out a nuclear run-off transcription assay on DT40 subclones. 32 P-Labeled RNAs were purified from nuclei from the DT40 and ⌬chHDAC-2 (cl-33-27 and cl-33-30) cell lines, and then hybridized to the VH DNA fragment, corresponding to the V, D, and J regions of the IgM H-chain gene, together with GAPDH cDNA and Bluescript II plasmid as positive and negative controls, respectively (Fig. 8A). As compared with the wild-type cell line, the level of transcription of the IgM H-chain gene increased in the mutant, the magnitude of the increase being about 2-fold (see table in Fig. 8A).

IgM L-chain a 22 A L T Q P S S V S A N P G E T V K I T 40 Protein 3 A L T Q P A S V S A N P G E T V K I T Protein 4 A L T Q P A S V S A N P G E T V K I T
Next, to exclude the possibility that the increase in the steady-state level of s mRNA (and total IgM H-chain mRNA) was simply due to the increase in the stability in ⌬chHDAC-2, using the RNase protection method with probe Cmu, we measured the amounts of s and m mRNAs in the DT40 and ⌬chHDAC-2 cell lines in the presence of actinomycin D. Fig. 8B shows that in every experiment with treatment with actinomycin D, by 12 h the ratios of s to m mRNAs in DT40 and ⌬chHDAC-2 were about 0.5 and 5.0, respectively. These ratios agreed with those mentioned above (see Fig. 7B). The time courses of the relative amounts of s and m mRNAs in DT40 and ⌬chHDAC-2 cells are presented in Fig. 8C. The half-life of s or m mRNA in ⌬chHDAC-2 was about 5 or 4 h, which is identical to that of the two types of mRNAs in DT40 cells, indicating insignificant alterations in the stability of s and m mRNAs in the mutant.
These results thus indicated that the increase in the steadystate level of total IgM H-chain mRNA (equal to the sum of s and m mRNAs) in ⌬chHDAC-2 was mainly based on the increased transcription of the IgM H-chain gene.

TSA Increases both Transcription of the IgM H-chain Gene and Alternative Processing of IgM H-chain Pre-mRNA in DT40
Cells-To confirm the participation of chHDAC-2 in the increases in both the transcription of the IgM H-chain gene and alternative processing from m to s mRNA, we studied the effect of TSA on these steps in DT40 cells.
Total RNA from DT40 cells cultured for 24 or 48 h in the presence of TSA at various concentrations was analyzed by the RNase protection method using probe VH or Cmu (Fig. 9), essentially as described above. The use of probe VH revealed that the steady-state level of total IgM H-chain mRNA increased in the presence of TSA (by 1000 nM), the magnitude of the increase being about 2-fold. In addition, probe Cmu showed that the steady-state level of s mRNA increased in the presence of TSA (by 1000 nM), and the ratio of s to m mRNA was about 2.0, compared with that (about 0.5) of the two types of mRNAs in the absence of the drug.
These results thus revealed that the increases in the steadystate levels of both total IgM H-chain mRNA and s mRNA in ⌬chHDAC-2 were certainly related to the chHDAC activity. of goat anti-chicken IgM -chain antiserum, followed by incubation in a 1:1000 dilution of FITC-labeled rabbit anti-goat IgG and then analyzed with a FACS Vantage. The symbols for the cell lines are shown in the inset. There is general agreement that the level of acetylation of nucleosomal core histones is related to the transcriptional activity, and the acetylation induces an open chromatin conformation that allows the transcription machinery access to promoters. Only recently have many studies revealed an attractive model for dynamic changes in the chromatin structure, based on both the acetylation and deacetylation of core histones, respectively, catalyzed by histone acetyltransferases and HDACs (1-3, 9, 10). It has been reported that there are three human (12,55) and two mouse (36,56) HDACs. In this study, we revealed the presence of two chicken HDACs, chHDAC-1 and -2, and then clarified differences in the characteristics of the two enzymes (Figs. 1-3). Furthermore, our recent screening of cDNA and genomic DNA libraries revealed that one more HDAC, named chHDAC-3, exists in the chicken (data not shown). Therefore, these HDAC family members precisely control, in combination with one another and/or with histone acetyltransferase family members, the level of chemical modification with acetyl groups of core histones, which is related to transcription activity. To understand the overall picture of this regulated transcription in higher eukaryotes, therefore, it should be essential to assess the individual, particular roles of these multiple HDACs in vivo.
To clarify the difference in the roles of chHDAC-1 and -2, we generated two homozygous chicken DT40 mutants, ⌬chH-DAC-1 and -2, respectively, devoid of two alleles of genes encoding the two enzymes (Fig. 4), and then compared total cellular proteins of the two mutants with those of the wild-type cell line (Fig. 5). The protein patterns on two-dimensional PAGE definitely changed for ⌬chHDAC-2, but the changes were insignificant for ⌬chHDAC-1. Interestingly, our analyses, involving amino acid sequence determination of varying proteins, revealed that the amounts of the IgM H and L-chains (Table I) increased in ⌬chHDAC-2, whereas three other varying proteins exhibited no homology to any known proteins. Moreover, the amino acid sequences of about 30 other major proteins in DT40 cells, which were unchanged in quantity in both ⌬chHDAC-1 and -2, were determined. All of them were well known predominant proteins, including the tubulin and heat-shock proteins, but did not correspond to any known B cell-specific proteins (data not shown). Together, these results or Cmu (C), together with total yeast tRNAs (tRNA lane). Probe GAPDH was used as a control, and the labeled probes were also run. After electrophoresis in denaturating polyacrylamide gels, autoradiography was carried out.
revealed that chHDAC-1 and -2 are definitely distinct in their roles in certain cell functions; i.e. the involvement of the former in protein pattern changes is slight, but the latter predominantly participates in the accumulation of the IgM H-and L-chains and other particular proteins. The magnitude of the increase in the amount of the IgM H-chain with the deletion of chHDAC-2 was far superior to the capacity to secrete it (see Figs. 5B and 6).
Both chHDAC-1 and -2 exhibit extensive homology (ϳ94%) in approximately 430 N-terminal amino acid sequences, but considerably low homology (ϳ50%) in approximately 50 Cterminal amino acid sequences (Fig. 1). The consensus sequence for the retinoblastoma protein binding site, which is conserved in the Xenopus (37), mouse (36,56), and human (12, 55) HDACs, is involved in the N-terminal regions of chHDAC-1 and -2. The difference in the participation of chHDAC-1 and -2 in the accumulation of the IgM H-and L-chains, therefore, should be due to the difference in their C-terminal sequences. Although some of the polypeptides co-precipitated with the anti-chHDAC-1 and -2 antisera (Fig. 2B) possibly correspond to chicken homologs of mSin3A, SMRT, or N-CoR (13-16, 18 -20), the roles of these co-precipitated polypeptides in the chHDAC-2 immunoprecipitate are unknown.
In eukaryotes, gene expression can be controlled through multiple steps, including the transcription and alternative pre-mRNA processing. The regulation of IgM H-chain synthesis during B cell development is known as an example of the latter mechanism (57-59). There is a regulated switch from the membrane-bound (m) to secreted (s) form of IgM H-chain mRNA (46,47). Several factors, including CstF-64, participate cooperatively in this alternative processing of IgM H-chain pre-mRNA (48). Our findings indicate that the chHDAC-2 mutation influenced the expression of the IgM H-chain gene (Figs. 7B and 8A) but not the stabilities of s and m mRNAs (Fig. 8,  B and C). In addition, the mutation predominated the switch from m to s mRNA ( Fig. 7C) but had an insignificant effect on the steady-state level of CstF-64 mRNA, measured by both RT-PCR and the RNase protection method with an antisense RNA probe, corresponding to a part of CstF-64 cDNA (kindly provided by Yoshio Takagaki and James Manley) (data not shown). This participation of chHDAC-2 was confirmed by the following findings. Treatment with TSA resulted in increases in both the steady-state level of total IgM H-chain mRNA and the switch from m to s mRNA in DT40 cells (Fig. 9). These influences of TSA should be related only to chHDAC-2, since the chHDAC-1 mutation had no effect on the amounts of the IgM H-chain and its mRNAs (see Figs. 5A and 7), and an insignificant alteration in the amount of the protein was observed in the conditional DT40 null mutant, devoid of chH-DAC-3, which is lethal for DT40 cells (data not shown).
Based on our results obtained for ⌬chHDAC-2, together with those reported (46 -48, 53), we propose a model for a role of chHDAC-2 in the control of the amount of the s IgM H-chain in the DT40 cell line (Fig. 10). The transcription regulation mediated by chHDAC-2 probably occurs in conjunction with hypothetical signal(s), and the chHDAC-2 activity is reduced or completely abolished by it. The decreased activity should directly change the chromatin structure restricted to the narrow region surrounding the IgM H-chain gene, probably due to the remaining acetyl groups of particular Lys residue(s) of core histones, resulting in the increased transcription of this target gene. In addition, the increased transcription of the IgM Hchain gene is possibly accompanied by that of the gene(s) encoding putative switch-related factor(s), including a factor that stimulates the CstF-64 binding to its binding site, even when the amounts of CstF-64 are unchanged (53), and/or a putative factor that is considered to be induced by lymphokines (60). Despite the insignificant increase in CstF-64, the putative factor(s) should promote the switch from m to s mRNA. Thus, chHDAC-2 dually controls the amounts of the s IgM H-chain  9. Effects of treatment with TSA on the amounts of total, s, and m IgM H-chain mRNAs in DT40 cells. Total RNAs were isolated from DT40 cells cultured in the presence of TSA (500 or 1000 nM) for 24 or 48 h and then analyzed by the RNase protection method using 32 P-labeled antisense RNA probe VH or Cmu, together with that from DT40 cells (TSA, 0 nM) and yeast tRNAs (tRNA). Probe GAPDH was used as a control, and the labeled probes were also run.
at the steps of both the transcription and alternative pre-mRNA processing, through alterations in the chromatin structure.
Most or all of the specific functions of chHDAC-2 cannot be compensated for by any remaining chHDAC enzymes, all of which should be different from the former in substrate specificity, i.e. in core histone subtypes, histone variants, or deacetylatable Lys residues. Furthermore, chHDAC-2 should be predominantly involved in the control of the transcription of genes encoding a set of particular proteins, including the IgM Lchain, and proteins 1, 2, and 7. The overall picture concerning the participation of chHDAC-2 in the transcription regulation of B cell-specific genes should be clarified by further studies.