Isolation, Characterization, and Molecular Cloning of a Protein (Abp2) That Binds to a Schizosaccharomyces pombe Origin of Replication (ars3002)

Abstract

The autonomously replicating sequence (ARS) element ars3002 is associated with the most active replication origin within a cluster of three closely spaced origins on chromosome III of Schizosaccharomyces pombe. A 361-bp portion of ars3002 containing detectable ARS activity includes multiple near matches to the S. pombe ARS consensus sequence previously reported by Maundrell et al. (K. Maundrell, A. Hutchison, and S. Shall, EMBO J. 7:2203–2209, 1988). Using a gel shift assay with a multimer of an oligonucleotide containing three overlapping matches to the Maundrell ARS consensus sequence, we have detected several proteins in S. pombe crude extracts that bind to the oligonucleotide and ars3002. One of these proteins, ARS binding protein 1, was previously described (Abp1 [Y. Murakami, J. A. Huberman, and J. Hurwitz, Proc. Natl. Acad. Sci. USA 93:502–507, 1996]). In this report the isolation, characterization, and cloning of a second binding activity, designated ARS binding protein 2 (Abp2), are described. Purified Abp2 has an apparent molecular mass of 75 kDa. Footprinting analyses revealed that it binds preferentially to overlapping near matches to the Maundrell ARS consensus sequence. The gene abp2 was isolated, sequenced, and overexpressed in Escherichia coli. The DNA binding activity of overexpressed Abp2 was similar to that of native Abp2. The deduced amino acid sequence contains a region similar to a proline-rich motif (GRP) present in several proteins that bind A+T-rich DNA sequences. Replacement of amino acids within this motif with alanine either abolished or markedly reduced the DNA binding activity of the mutated Abp2 protein, indicating that this motif is essential for the DNA binding activity of Abp2. Disruption of the abp2 gene showed that the gene is not essential for cell viability. However, at elevated temperatures the null mutant was less viable than the wild type and exhibited changes in nuclear morphology. The null mutant entered mitosis with delayed kinetics when DNA replication was blocked with hydroxyurea, and advancement through mitosis led to the loss of cell viability and aberrant formation of septa. The null mutant was also sensitive to UV radiation, suggesting that Abp2 may play a role in regulating the cell cycle response to stress signals.

The best-characterized eukaryotic origins of DNA replication are those of the budding yeast Saccharomyces cerevisiae. These replication origins were identified as chromosomal sequences that support autonomous replication of plasmids (called autonomously replicating sequence [ARS] elements). Subsequently, two-dimensional gel electrophoretic methods showed that ARS elements colocalize with replication initiation sites both in plasmids and on chromosomes (reviewed in references 17 and 30). Mutational analyses of several S. cerevisiae ARSs have defined two essential domains, A and B (19, 21, 35). Domain A contains a match to the 11-bp S. cerevisiae ARS element consensus sequence, while domain B includes three or four subdomains, referred to as B1, B2, B3, and/or B4 (19, 21).

A complex of six polypeptides, the origin recognition complex, binds to domains A and B1 in an ATP-dependent manner (6, 21). Biochemical and genetic studies indicate that the origin recognition complex participates in the initiation of DNA replication (11). Similar studies suggest that additional proteins, including Cdc6 and the minichromosome maintenance family of proteins, are also essential for initiation (reviewed in references 5 and 30).

Replication origins in animal cells are not as well understood as those of S. cerevisiae. There is controversy concerning the distribution of the initiation sites at animal cell replication origins, and little is known about the cis-acting sequences affecting these origins (7, 18).

The fission yeast Schizosaccharomyces pombe resembles animal cells in some respects (such as centromere structure) to a greater extent than does S. cerevisiae and has a number of experimental advantages, like S. cerevisiae. For these reasons, studies concerned with replication origins and origin-binding proteins in S. pombe may contribute to the understanding of the initiation of replication in animal cells.

We have used nucleotide sequences from the ARS element ars3002 which correspond to the most active replication origin in a cluster of three closely spaced origins near the ura4 gene on S. pombe chromosome III (12) as a DNA substrate to identify ARS binding proteins. Deletion analysis of ars3002 previously defined a region of 361 bp that supports replication, albeit at a low level (37). This region contains multiple sequences similar to the consensus sequence (A/T)(A/G)TTTATTTA(A/T) found by Maundrell et al. (22) in most S. pombe ARS elements. Within this 361-bp region, two sequences of 30 to 55 bp (sequences α and β [see Fig. 1A]) were shown to be essential for ARS activity (13). In S. pombe ars1, one region of about 30 bp was shown to be essential for ARS activity (9). This region resembles both the α and β elements of ars3002 (13). The α and β sequences in ars3002 both contain multiple matches to the Maundrell ARS consensus sequence (see Fig. 1A).

FIG. 1.

(A) Structure of a 361-bp region of S. pombe ars3002. Deletion analysis within this ARS element defined a minimal region of 361 bp that supports replication (37). The regions indicated by thick arrows contain matches of 11 of 11 bases to the consensus sequence of Maundrell et al. (S. pombe ARS consensus sequence [22]); the regions indicated by medium arrows contain matches of 10 of 11 bases, and regions indicated by thin arrows possess matches of 9 of 11 bases. The arrows indicate the 5′-to-3′ orientation of the consensus match. Mutational analysis of this 361-bp region identified two essential sequences, indicated by the letters α (bp 127 to 175) and β (bp 256 to 285), that colocalize with some matches to the S. pombe ARS consensus sequence. In the case of the α sequence, the thin lines indicate a region (nt 127 to 145) in which linker substitution had a smaller effect on ARS activity (13). In the β sequence, the thin line and arrow indicate that the left boundary of this essential region has not been completely defined. The α and β sequences were shown to be essential by linker replacement (13). These sequences correspond to the linker substitutions 7d, 7e, and 8a-f (for α) and 10 a-f (for b) described in reference 13. (B) Identification of complexes formed with S. pombe crude extracts in the gel mobility assay in the presence of labeled MMACS. Two amounts of S. pombe cell extract (lanes 1 to 3) were incubated with the labeled MMACS tetramer and subjected to PAGE. DNA-protein complexes I and II are indicated. The lower panel shows the sequence of the MMACS oligonucleotide used. The oligonucleotide was annealed, restricted, and ligated to form the tetramer (27) used in this study. Thick and medium lines indicate a perfect match or one base mismatch, respectively, to the S. pombe ARS consensus sequence (27). Bases capitalized correspond to the 28-bp region present in ars3002 (nt 3371 to 3398 [37]).

To identify possible ARS binding proteins, we used gel shift assays with multimers of a double-stranded oligonucleotide, called MMACS (for multiple Maundrell ARS consensus sequence), based on a 28-bp sequence of ars3002 that contains three overlapping matches to the Maundrell ARS consensus sequence (27). Multiple complexes were detected when a dimeric or tetrameric MMACS was incubated with crude extracts from S. pombe (27). We have previously purified a 60-kDa protein, ARS binding protein 1 (Abp1), responsible for the formation of one of the complexes (27). In this report we describe the purification and cloning of a second protein, ARS binding protein 2 (Abp2), which is responsible for the formation of another complex.

MATERIALS AND METHODS

Escherichia coli strains.

E. coli XL-1 Blue (Stratagene) was used for plasmid maintenance. E. coli DH10 (obtained from L. Guarente), y1080r (Clontech), and BL21(DE3) (Novagen) were used for cDNA library maintenance, as the host for maintaining the genomic library, and for the expression of the recombinant protein, respectively.

Yeast strains and diploid strain construction.

S. pombe extracts were prepared from strain 972 h⁻ (FCY1; American Type Culture Collection). KGY246 (h⁻ ade6-216 ura4-D18 leu1-32) was crossed to KGY249 (h⁺ ade6-210 ura4-D18 leu1-32), and the ade6⁺ diploid was selected as previously described (1). The cut5-T401 strain was used for UV radiation experiments (31).

Gel shift assay and DNA substrates.

Gel shift assays and DNA substrates were described previously (27). A 361-bp segment from ars3002 (nucleotides 3371 to 3730 of the sequence described by Zhu et al. [37]) was divided by PCR into three separate regions of 120 bp each, using oligonucleotide primers (35-mers) 5′ and 3′ to the ends of the regions from bp 1 to 120, 121 to 240, and 241 to 361 (the sequences are shown in Fig. 5). The three amplified DNA sequences were cloned into pBluescript (Stratagene) and sequenced according to the protocol of the U.S. Biochemical Corp. (USB) (20). Construction of the multimeric MMACS oligonucleotide (see Fig. 1B) was described previously (27).

FIG. 5.

DNase I footprint of Abp2 bound to ars3002. DNase I footprint analysis with duplex DNA containing nt 1 to 120 of the 361-bp ars3002 DNA (Materials and Methods) (A), nt 121 to 240 (B), and nt 241 to 360 (C). Purified Abp2 was added to 5 fmol of labeled DNA. Reaction mixtures were prepared and sequencing gel electrophoresis was carried out as described in Materials and Methods. Control reactions with no protein are shown in lanes 1, 5, and 9. The hatched boxes alongside the panels indicate the regions protected. The essential regions, α and β (Fig. 1) (13), are indicated. The sequence of the DNA analyzed is indicated below each panel. The region indicated with a thick line contains a match of 11 of 11 bases to the consensus sequence of Maundrell et al. (22), those indicated with medium lines contain matches of 10 of 11 bases, and those indicated with thin lines possess matches of 9 of 11 bases. The sequence overlined with dashed lines are some of many containing a match of 8 of 11 bases. Other matches of 8 of 11 bases in ars3002 are not shown due to their high abundance. Hatched boxes under the sequence indicate regions protected from DNase I attack by Abp2.

Purification of Abp2.

During Abp2 purification, DNA binding activity was measured by gel mobility shift assays with the MMACS tetramer as a substrate (see Table 1). As previously described, S. pombe extracts (from 0.8 kg of cells) were prepared and subjected to S-Sepharose (Pharmacia) chromatography (27). The 0.25 M KCl eluate from the column (3 liters) contained Abp2. The subsequent steps used to purify Abp2 were identical to those described for the isolation of Abp1, with the following modifications (27). The active fractions eluted from the dimeric MMACS affinity column were pooled and diluted to 0.1 M KCl with buffer H (50 mM HEPES-KOH [pH 7.5], 5 mM magnesium acetate, 1 mM EDTA, 1 mM EGTA, 0.02% Nonidet P-40, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 10 mM NaHSO₃, 0.5 μg of leupeptin and antipain per ml). This fraction (9.5 ml, 130 μg) was loaded onto a Mono Q HR 5/5 column (1 ml) equilibrated with buffer H plus 0.1 M KCl and then eluted with a 7-ml salt gradient (0.1 to 0.8 M KCl) in buffer H; 35 fractions of 0.2 ml each were collected. The most active fractions (assayed by gel shift) were eluted at 0.3 M KCl and were pooled. For the glycerol gradient fractionation step, 10.5 μg of protein in 0.13 ml was loaded onto a 15 to 35% glycerol gradient (5 ml, containing buffer H plus 0.3 M KCl [21]). After centrifugation at 4°C for 20 h at 45,000 rpm (SW50.1 Ti rotor; Beckman), 30 fractions (170 μl each) were collected. Protein concentrations were determined by the Bio-Rad protein assay unless otherwise indicated.

TABLE 1.

Purification of Abp2

Fraction	Vol (ml)	Amt of protein (mg)	Total (U)a	Sp act (mU/mg)
Whole-cell extract	1,400	8,100	330,000	0.041
S-Sepharose	1,100	946	215,000	0.23
Q-Sepharose	80	136	141,200	1.04
First MMACS-agarose	12.5	1.4	23,600	16.9
Second MMACS-agarose	2.8	0.13	6,000	46.2
Mono Q HR 5/5	0.4	0.031	4,400	142
Glycerol gradientb	0.45	0.005c	1,510	302

^aOne unit was defined as the amount of protein required to convert 50% of the substrate (MMACS tetramer) to complexes in the standard gel mobility shift assay described in Materials and Methods. 

^bIn this step, one-third of the Mono Q HR 5/5 fraction was used; the results presented were calculated assuming that the entire fraction was used. 

^cEstimated by silver staining. 

Footprinting assay.

Reaction mixtures (20 μl) contained 40 mM HEPES-NaOH (pH 7.5), 5 mM magnesium chloride, 1 mM calcium chloride, 2 mM dithiothreitol, 5% glycerol, 2% polyethylene glycol (molecular weight, 20,000), 0.2 μg of poly(dA-dC) (5,000 bp in length; 61.3 fmol) as a nonspecific competitor, 4 mM ATP, 2 μg of bovine serum albumin, 5 fmol of MMACS dimer labeled at the 5′ end with ³²P (sequence presented in Fig. 4; 0.5 × 10⁴ to 1 × 10⁴ cpm/fmol), or 5 fmol of one of the three separate 120-bp regions of ars3002 DNA (see Fig. 5). Purified fractions of Abp2, as indicated, were added, and the mixture was incubated for 15 min at 30°C, after which time 2 mU of DNase I (Worthington Biochemical Corp.) was added. After 1 min at 30°C, the reaction was terminated with 1 volume of phenol-chloroform (1:1), and the mixture was centrifuged for 3 min at 13,000 rpm in a Fisher Scientific microcentrifuge. Yeast tRNA (5 μg) was added as a carrier, and the mixture was precipitated with 3 volumes of ethanol. After centrifugation, pellets were resuspended in 3 μl of TE buffer (10 mM Tris, 1 mM EDTA [pH 8.0]) and 2 μl of loading buffer (0.5% bromophenol blue, 0.5% xylene cyanol, 90% formamide). Samples were incubated for 2 min at 90°C and then electrophoresed through DNA sequencing gels (6% polyacrylamide, 7 M urea), as described previously (20). After electrophoresis, gels were dried and autoradiographed. Mixtures were electrophoresed in parallel with G/A chemical sequencing reactions of the DNA (20).

FIG. 4.

Binding of Abp2 to ars3002 DNA. Increasing amounts of purified Abp2 were incubated with 2 fmol of the labeled 361-bp ars3002 DNA and then subjected to 1.5% agarose gel electrophoresis, as described in Materials and Methods. A control lacking Abp2 is indicated in lane 5.

Cloning of Abp2.

Five distinct tryptic peptide sequences were obtained from purified preparations of Abp2 (Microchemistry Core Facility, Memorial Sloan-Kettering Cancer Center). Based on their amino acid sequences, two degenerate oligonucleotides (30-mers) were synthesized and used to amplify by PCR DNA from both an S. pombe λgt11 genomic library (Clontech) and S. pombe genomic DNA. A PCR product of 520 bp was obtained with both templates and was cloned into a plasmid vector (pUC19). The PCR product, sequenced from both ends, contained the sequence of the two degenerate oligonucleotides. Screening of an S. pombe cDNA library (15) with the radiolabeled PCR products as probes yielded seven positive clones. Inserts from three clones, each containing a full-length cDNA of approximately 2.8 kb, were subcloned into pBluescript (Stratagene), and the entire sequence of each was determined by using Sequenase according to protocols described by the manufacturer (USB). The sequences of all three cloned fragments were identical and contained a long (1,581-nucleotide [nt]) open reading frame that encoded all of the peptide sequences obtained from Abp2.

Purification of GST-Abp2 protein.

The Abp2 protein was overexpressed in bacteria as a glutathione S-transferase (GST)-Abp2 fusion protein by using the pET system (33). The expression vector used to express GST-Abp2 was pET19 (Novagen), modified by the method of Müller et al. (26). The cDNA of Abp2, except for the portion encoding the 9 amino acids of the N terminus, was subcloned into the AvrII restriction site of the vector pET19GST. Growth and induction were carried out as described by Studier et al. (34). Briefly, E. coli BL21(DE3) cells were induced at 25°C for 4 h. Pelleted cells (5 g from 3 liters) were resuspended in 50 ml of 25 mM Tris-HCl buffer (pH 7.5)–5 mM EDTA–0.1 M NaCl–2 mM dithiothreitol–0.5% Nonidet P-40–1 mM phenylmethylsulfonyl fluoride–1 mg of lysozyme per ml and incubated for 1 h on ice. After sonication, the lysate was centrifuged at 20,000 × g for 45 min in a Sorvall SS34 rotor. The supernatant (45 ml, 0.5 g) was loaded onto a glutathione-Sepharose column (2 ml, 2 by 1.0 cm; Pharmacia), and the GST-Abp2 fusion protein was purified as described by Müller et al. (26). Approximately 50% of the GST-Abp2 fusion protein bound to the affinity column. The column was washed with 10 volumes of buffer H plus 0.8 M NaCl, and the GST-Abp2 fusion protein (4 to 5 mg in 2.5 ml) was eluted with 20 mM glutathione in buffer H. This material was further purified with a Superose 12 (Pharmacia) column (25 ml, 25 by 1.0 cm). This procedure yielded 1.2 mg of homogeneous GST-Abp2 protein (100 kDa) which was dialyzed for 5 h against 0.5 liter of buffer H plus 0.25 M NaCl–20% glycerol and stored as aliquots at −80°C.

Site-directed mutagenesis of the sequence encoding the GRP motif.

To create mutations by site-directed mutagenesis in the sequence encoding the GRP motif of Abp2, we used the Chameleon double-stranded site-directed mutagenesis kit of Stratagene. The mutations were made in the double-stranded vector pET19GST-abp2, described above (see “Purification of GST-Abp2 protein”). The mutagenic oligonucleotide primer A1 (5′-GCAGGAGTCCCTCGTAAAGCCGGGCGTCCGCCAGGAGCT-3′) was used to convert the codon for Arg 331 (CGC) in the abp2 coding frame to one for an Ala (GCC; underlined above). The oligonucleotide primer A2 (5′-GGAGTCCCTCGTAAACGCGCGCGTCCGCCAGGAGCTCGT-3′) was used to replace the codon for Gly 332 (GGG) in the coding frame with a codon for an Ala (GCG; underlined above), and the mutagenic oligonucleotide primer A3 (5′-GTCCCTCGTAAACGCGGGGCTCCGCCAGGAGCTCGTAAC-3′) was used to change the codon for Arg 333 (CGT) to a codon for Ala (GCT; underlined above). The mutagenic oligonucleotide primer A4 (5′AATGAAGCAGGAGTCCCTAAGAAACGCGGGCGTCCGCCA-3′) was used to replace the codon for Arg 329 (CGC) in the abp2 coding frame with a codon for Lys (AAG). The selection oligonucleotide primer PT1 (5′-GACACCACGATGCCGGCGGCAATGGCAACAACG-3′) was used to convert the PstI restriction site located in the ampicillin resistance gene in the vector (CTGCAG) to a noncleavable sequence (CGGCGG).

Site-directed mutagenesis was performed as described in the instruction manual provided with the kit. Briefly, the denatured plasmid pET19GST-abp2 was annealed simultaneously to the selection primer (PT1) and the mutagenic primer. The annealed primers were extended with T7 DNA polymerase and ligated with T4 DNA ligase. Next, the plasmid DNA was restricted with PstI to linearize the residual parental plasmid, leaving the mutant plasmid undigested. The digested DNA preparation was used to transform the repair-deficient mutS strain of E. coli. Since the selection of the “correct” strand is random in this strain, half of the isolated plasmids contained the desired mutation. The DNA was purified from a pool of transformants by a miniprep procedure and digested with PstI, and the digested DNA was used to transform the XL-1 Blue strain. The DNA isolated from different clones was screened for the absence of the PstI restriction site. Clones devoid of this site were sequenced with Sequenase according to the protocol provided by the manufacturer (USB). About 50 to 80% of the isolated clones contained the desired mutation. To facilitate the purification of the recombinant protein, pET19GST plasmids containing the mutated abp2 gene were used to transform E. coli BL21(DE3) cells. Induction and purification of the mutated GST-Abp2 proteins were carried out as described for the purification of the GST-Abp2 protein.

Gene disruption.

A one-step gene replacement to disrupt the abp2⁺ gene was done as follows (3, 11, 16). A 1.3-kb XbaI-PvuI fragment containing most of the coding region of abp2⁺ was isolated and the ends were made blunt with Klenow polymerase and deoxynucleoside triphosphates (20). The resulting fragment was ligated to the SmaI site of plasmid pUC18. The 0.35-kb MstI-BbvII fragment in the coding region of abp2⁺ was replaced with the 2.2-kb S. cerevisiae LEU2⁺ gene, and the resulting plasmid was linearized by double digestion with PstI and ScaI. The linear fragment containing the disrupted gene was isolated and introduced into the chromosome of a diploid strain by homologous recombination. The diploid cells were transformed by electroporation as previously described (28). Diploids containing the disrupted gene were selected based on the presence of the LEU2 marker gene. DNA was isolated (1) from diploid cells containing the disrupted gene and subjected to PCR analysis with primers Pa1 (5′-TTCAACCCCTGACTTTCTTTGGGTGA) and Pa2 (5′-CCAATTCTGTCTTTGCTGCAATCCCT) (26-mers). PCRs were performed according to protocols described by the manufacturer (Takara Inc., Otsu, Japan). The 5′ end of primer Pa1 is located at nt 1761 of the Abp2-coding sequence, and the 5′ end of primer Pa2 is located at nt 914 of the Abp2-coding sequence but oriented in the opposite direction to primer Pa1 (see Fig. 8B). Reaction products were subjected to electrophoresis on a 0.8% agarose gel, followed by staining with ethidium bromide. The formation of a 2.85-kb product indicated disruption of the abp2 gene, whereas the presence of a 0.85-kb product indicated the presence of a wild-type copy of the gene. Heterozygous diploids were sporulated and the resulting tetrads were dissected as described previously (1). Chromosomal DNAs isolated from spores were analyzed by PCR.

FIG. 8.

Disruption of abp2⁺. (A) A one-step replacement method was used for the disruption of the abp2⁺ gene (see Materials and Methods). The disruptants were designated Δabp2 A and B (see Materials and Methods). A linear DNA fragment containing the disrupted abp2 gene was introduced into a diploid strain by homologous recombination. The heterozygous diploid strain was sporulated, and the resulting tetrads were dissected. (B) PCR analysis of heterozygous diploid cells and the tetrads resulting from the disruption yielding Δabp2 A. PCRs were performed with chromosomal DNA isolated from diploid cells and isolated tetrads as described in Materials and Methods. Heterozygous diploid cells yielded a 2.8-kbp band from the disrupted gene and a 0.85-kbp band from the wild-type gene (lane 1). The dissected tetrads showed a 2:2 segregation of the disrupted gene to the wild-type gene (lanes 2 to 5). Lanes 6 and 7 show controls for the PCR in which S. pombe haploid chromosomal DNA and control DNA were used, respectively; lane 8 contains wild-type DNA with no primers. Molecular size (MW) markers are indicated at the right.

For the second deletion of the abp2 gene, a 1.06-kb BspMI-NcoI fragment in the coding region of abp2⁺ was replaced with the 2.2-kb S. cerevisiae LEU2⁺ gene, and the resulting plasmid was linearized by double digestion with XbaI and PflMI (see Fig. 8A). The linear fragment containing the disrupted gene was isolated and introduced into the chromosome of a diploid strain by homologous recombination. The haploid cells were transformed by electroporation as previously described (28), and cells containing the disrupted gene were selected based on the presence of the LEU2 marker gene. DNA was isolated (1) from haploid cells containing the disrupted gene and subjected to PCR analysis with primers Pa3 (5′-CCCTTGCTATACGTGCCATATAGCTTA) and Pa4 (5′-AATTGTCCTTGATGGAACGGTCCAAAT) (27-mers). The 5′ end of primer Pa3 is located at nt 310 of the Abp2-coding sequence, and the 5′ end of primer Pa4 is located at nt 2150 of the Abp2-coding sequence but oriented in the opposite direction to primer Pa3 (see Fig. 8A). PCRs were performed as previously described, and reaction products were subjected to electrophoresis. The formation of a 3.01-kb product indicated disruption of the abp2 gene, whereas the presence of a 1.85-kb product indicated the presence of a wild-type copy.

Analysis of the phenotype of cells containing the disrupted abp2 gene.

Wild-type cells and cells with a disrupted abp2 gene were grown to an optical density at 595 nm of 0.1 in YE standard medium (1) (plus supplements) at 28°C and then shifted to 36°C. Aliquots of cells were taken at 0, 4, 8, 18, 26, and 44 h for analysis of cell number and viability (1).

Fluorescence-activated cell sorter (FACS) analysis and DAPI and Calcofluor staining.

A Becton Dickinson FACScan was used to estimate cellular DNA content by procedures described previously (1). In brief, cells (0.5 × 10⁷) were collected, washed once with 1 ml of distilled water, and then resuspended in 70% ethanol. The cell were stored at 4°C for more than 24 h. After one wash with 0.5 ml of 50 mM sodium citrate (pH 7.0) and resuspension in the same buffer, RNase A (Sigma) was added to a final concentration of 0.5 mg/ml. Following incubation at 37°C for 1 h, propidium iodide (Sigma) was added to a concentration of 12.5 μg/ml. The stained cells were filtered through a nylon mesh (35 μm; Small Parts Inc.) and then analyzed. 4′,6-Diamidino-2-phenylindole (DAPI) and Calcofluor staining procedures were employed as described previously (1).

UV and hydroxyurea survival analysis.

A known density of cells (1 × 10³ to 2 × 10³) were plated onto minimal medium (MM) agar plates, exposed to a dose of UV light determined by the setting on an Ultra-Lum (Carson, Calif.) UVC-508 UV cross-linker, and then incubated for 2 to 3 days. Colonies were counted and survival was expressed as a percentage of colonies formed on equivalent plates not exposed to UV. Survival in hydroxyurea was determined with growing asynchronous cultures (optical density at 600 nm of 0.1) in supplemented MM in the presence of 10 mM hydroxyurea. After 4 h of incubation in 10 mM hydroxyurea, the hydroxyurea concentration was increased to 14 mM to completely arrest cell growth. Thereafter, aliquots were removed every 4 h, diluted, counted, and plated. Survival was expressed as a percentage of colonies formed on equivalent plates incubated without hydroxyurea.

Nucleotide sequence accession number.

The nucleotide sequence reported here is entered in GenBank with accession no. U73044.

RESULTS

Purification of Abp2.

ars3002 DNA, which contains an origin of replication in S. pombe, was used as a substrate to isolate proteins that bind to this DNA. As described previously (27), we employed a mobility shift assay with a labeled MMACS tetramer as a substrate. The sequence of the MMACS monomer contains one match of 11 of 11 bases and two matches of 10 of 11 bases of the S. pombe ARS (Fig. 1B). Incubation of increasing amounts of crude extract with the labeled MMACS tetramer resulted in the formation of multiple protein-DNA complexes (Fig. 1B). Complex II was previously shown to be due to the interaction of the MMACS tetramer with the 60-kDa Abp1. With this gel shift assay, the DNA binding activity responsible for complex I formation was purified approximately 10⁴-fold, with a recovery of 0.46% (Table 1). Mobility shift assays (Fig. 2A) and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analyses (Fig. 2B) were performed in parallel with the fractions obtained from the last step of the purification procedure (glycerol gradient sedimentation). Though multiple protein bands were detected, the MMACS binding activity correlated best with a 75-kDa protein which possessed a sedimentation coefficient of 6.7 (Fig. 2A, fractions 7 and 9). The 75-kDa protein consistently comigrated with the peak of DNA binding activity during separate steps used for the isolation of Abp2. From these results, we concluded that the 75-kDa protein was probably responsible for the binding activity. We have designated this protein Abp2.

FIG. 2.

Glycerol gradient sedimentation reveals the presence of a 75-kDa protein responsible for complex I formation. A representative glycerol gradient of the peak fractions obtained from the second Mono Q HR 5/5 column (see Materials and Methods) is shown. (A) Results of the gel retardation assay using 2 μl of each fraction of the glycerol gradient (as described in Materials and Methods) with the labeled MMACS tetramer as a substrate; (B) silver staining profile of the SDS-PAGE analysis of the fractions (30 μl) obtained from the glycerol gradient. The protein concentrations of the peak fractions of the glycerol gradient were estimated to be 5 to 8 ng/μl (based on silver staining). The positions of molecular mass markers (in kilodaltons) are shown to the right. The arrow indicates the 75-kDa protein that comigrated with the peak of DNA binding activity shown in panel A.

Characterization of Abp2 binding activity.

A total of 4 mM ATP and 10 mM Mg²⁺ were used in the DNA binding assays described above. In the presence of ATP, binding of Abp2 to MMACS was maximally stimulated (four- to fivefold) by 5 mM Mg²⁺; higher Mg²⁺ levels inhibited the reaction (data not shown). ATP was not required for the binding of Abp2 to MMACS; however, at 10 mM Mg²⁺, DNA binding activity was stimulated two- to threefold in the presence of 2 to 5 mM ATP. We attribute this stimulation to the sequestration of Mg²⁺ by ATP, which decreases the inhibitory effect of high Mg²⁺ levels. Similar effects were observed with Abp1 (27). In the most purified fractions of Abp2, no ATPase activity or single-stranded-DNA binding activity was detected (data not shown).

To define more precisely the sequences in the MMACS dimer bound by Abp2, DNase I footprinting analyses were carried out. The results in Fig. 3 show that increasing concentrations of Abp2 (lanes 1 to 3) protected regions within the perfect match to the Maundrell ARS consensus sequence but not its flanking sequences. Binding of Abp2 generated DNase I-hypersensitive sites in nt 30 to 32 and 61 to 62 of the MMACS dimer. The precise boundaries of DNA protected by Abp2 were difficult to define because the regions flanking the Maundrell ARS consensus sequence were not cleaved by DNase I even in the absence of Abp2 (Fig. 3, lane 4). However, these results suggest that Abp2 binds preferentially to the Maundrell ARS consensus sequence in the dimeric substrate.

FIG. 3.

DNase I footprint of the MMACS dimer complexed with Abp2. Increasing amounts of purified Abp2 were incubated with 5 fmol of labeled MMACS dimer and then treated with DNase I as described in Materials and Methods. A control reaction with no Abp2 protein is shown in lane 4. The sequence of the labeled DNA strand of the MMACS dimer used in the footprinting assay is shown. Thick and medium lines indicate a perfect match or one base mismatch, respectively, to the S. pombe ARS consensus sequence. The hatched boxes indicate the DNA regions protected by Abp2 binding. The perfect matches to the S. pombe ARS consensus sequence are indicated; they span nt 13 to 23 and nt 44 to 54. The position marked at nt 32 indicates the beginning of the second MMACS monomer.

The interaction between Abp2 and ars3002 as seen in the gel shift assay is shown in Fig. 4. Several different complexes formed as the level of protein was increased, probably reflecting multiple Abp2 binding sites in ars3002. To determine more precisely the sites in ars3002 recognized by Abp2, DNase I footprinting studies were carried out (Fig. 5). In order to simplify these studies, the 360-bp ars3002 DNA was divided into three 120-bp regions (from bp 0 to 120 [Fig. 5A], bp 120 to 240 [Fig. 5B], and bp 240 to 361 [Fig. 5C]) which were used individually as substrates in the footprinting experiment. Increasing concentrations of Abp2 protected multiple regions with various lengths. The most pronounced effect was observed between nt 242 to 251 and nt 256 to 268 (Fig. 5C). Protection of this region was observed at the lowest concentration of Abp2 added (15 ng). Two other protected regions spanned nt 10 to 19 (Fig. 5A) and nt 149 to 163 (Fig. 5B). Weak protection was also observed in other regions only at the highest level of Abp2 added (nt 58 to 72 [Fig. 5A] and nt 323 to 331 [Fig. 5C]).

The two protected sites located between nt 149 to 163 and nt 256 to 268 include part of the α and β sequences, regions previously found to be essential for the in vivo function of ars3002 (13). The other region protected, nt 10 to 19, contains an overlapping cluster of matches to the Maundrell S. pombe ARS consensus sequence.

These results suggest that Abp2 binds at or near regions possessing multiple overlapping matches to the ARS consensus sequence. However, it is evident that this specificity is not exacting.

The behavior of Abp2 in competition experiments with synthetic homopolymers differed from that of Abp1. The binding of Abp2 to MMACS was inhibited completely by a 50-fold molar excess of poly(dI-dC), poly(dA-dT), or poly(dA) · poly(dT). In contrast, the formation of the Abp1-MMACS complex was unaffected by the addition of poly(dI-dC) and poly(dA) · poly(dT) (27), whereas poly(dA-dT) competed Abp1 binding (data not presented). These results suggest that Abp2 has a binding specificity distinct from that of Abp1.

Cloning of Abp2.

To facilitate detailed biochemical and genetic analysis of the function of Abp2, the gene encoding Abp2 was cloned. The sequences of five peptides obtained from tryptic digests of purified Abp2 were used to design degenerate primers. The primers were then used to generate a probe, amplified (by PCR) from genomic DNA, for the isolation of clones containing abp2 cDNA from an S. pombe cDNA library (see Materials and Methods). We also screened a cosmid DNA library covering the entire genome of S. pombe. This procedure localized the abp2 gene to contig 4 of chromosome II (data not shown) (25).

The cDNA nucleotide sequence of abp2 and the predicted amino acid sequence are shown in Fig. 6. Analysis of the 845-nt 5′ untranslated region reveals two upstream AUG (beginning at nt 157 and 467) codons, neither of which contains an adenine nucleotide in the −3 position that is highly conserved in the translation of the initiation start site in higher eukaryotes (36). They code for 5 and 19 amino acids, respectively. The true initiation start site (846 nt) of the Abp2 reading frame contains an adenine in the −3 position.

FIG. 6.

(A) cDNA sequence and deduced amino acid sequence of Abp2. The open reading frame starts at nt 846 and ends at the termination codon, corresponding to nt 2427. Thin underlines indicate tryptic peptides derived from the purified 75-kDa Abp2 protein. Thick underlines indicate peptide sequences used for PCR. The boxed region shows the conserved GRP motif. (B) Homologous GRP domain common to Abp2 and proteins that bind AT-rich sequences. Asterisks indicate the positions of exact matches between the consensus sequence for the GRP box and Abp2.

The cDNA sequence contains an open reading frame encoding 527 amino acids corresponding to a protein of 60.8 kDa. The amino acid sequence included the five peptides obtained from the tryptic digests of Abp2 (Fig. 6).

Analysis of the amino acid sequence revealed significant similarity (Fig. 6B) to a proline-rich motif (GRP) present in several proteins that bind AT-rich DNA sequences (e.g., HMGI and MIF2 [8, 24]). A search of the GenBank database indicated that Abp2 is not similar to any reported protein.

Expression of GST-Abp2.

We cloned the Abp2 gene into a modified expression vector, pET19GST (see Materials and Methods), and expressed the resulting GST-Abp2 fusion protein in E. coli. The fusion protein was purified as described in Materials and Methods. The peak of binding activity (data not shown) comigrated with an 100-kDa protein which migrated more slowly than expected for the GST-Abp2 fusion protein (85 kDa). Abp2, expressed in E. coli as a histidine fusion protein, also migrated slower in gels than expected based on its molecular weight (data not presented). Competition experiments with different polymers indicated that the binding specificity of the fusion protein was identical to that of the native protein isolated from extracts of S. pombe (data not shown).

Site-directed mutagenesis of the sequence encoding the GRP motif in Abp2.

We tested whether the single GRP motif present in Abp2 is important for DNA binding by altering specific amino acids adjacent to and within this motif (RKRGRPPG) by site-directed mutagenesis. We replaced the codon for Arg 331 with one for Ala (clone A1-1), that for Gly 332 with one for Ala (clone A2-11), that for Arg 333 with one for Ala (clone A3-3), and that for Arg 329 with one for Lys (clone A4-6) in the coding frame of a pET19GST-abp2 expression vector. Each mutated GST-Abp2 fusion protein was purified to homogeneity, as described in Materials and Methods, and assayed for DNA binding activity with the MMACS tetramer substrate (Fig. 7). Conversion of Gly 332 or Arg 333 to Ala completely abolished the DNA binding activity. Even at high levels (800 ng) of protein, no binding activity was detected. Replacement of Arg 331 with Ala (clone A1-1) decreased the DNA binding activity of the mutated protein at least 40-fold compared to that of the wild-type GST-Abp2 protein (Fig. 7A and B). The binding activity of GST-Abp2 mutated in Arg 331 was sensitive to salt (50 to 150 mM NaCl) in comparison to wild-type GST-Abp2 (data not shown). Similar results were obtained when ars3002 DNA was used as a substrate (data not presented). Conserved replacement of Arg 329 located outside the GRP motif with Lys did not affect the DNA binding activity (data not presented). These results suggest that Gly 332 and Arg 333 are essential for DNA binding activity and that Arg 333 markedly stimulates the DNA binding activity of Abp2.

FIG. 7.

Comparison of the binding of wild-type and mutant GST-Abp2 fusion proteins to MMACS tetramer DNA. (A) Increasing amounts of purified wild-type GST-Abp2, A1-1, A2-11, and A3-3 proteins were incubated with 5 fmol of the labeled MMACS tetramer DNA and then subjected to 1.5% agarose gel electrophoresis, as described in Materials and Methods. A control lacking Abp2 is indicated in lane 15. (B) Gel shift results were quantified by PhosphorImager analysis (Fuji), and the percentage of the input DNA which was converted to a protein-DNA complex was determined and plotted as a function of the amount of protein added. (C) SDS-PAGE analysis of purified wild-type GST-Abp2 (WT) and mutant proteins. Each lane contained approximately 400 ng of protein. After SDS-PAGE, gels were stained with Coomassie blue. The positions of molecular mass markers (in kilodaltons) are shown to the right.

Disruption of the abp2 gene.

To determine whether Abp2 is essential for cell growth, abp2 null mutants were constructed. Two disruption mutants were prepared; in the first construct, Δabp2 A, the abp2 gene was disrupted between the codons for amino acids 172 and 255, whereas in the second disruption the codons for amino acids 32 to 385 were disrupted (Δabp2 B) (see Materials and Methods) (Fig. 8A). The abp2⁺ allele was replaced with a disrupted abp2 gene containing the S. cerevisiae LEU2⁺ gene (Δabp2 A [Fig. 8A]). To verify the disruption of one of the two abp2⁺ alleles of the diploid, PCR analysis was used with specific set of primers located within the abp2 gene (see Materials and Methods). We expected a 2.8-kbp product from the disrupted gene (abp2::LEU2 Δabp2 A) and a 0.85-kbp product from the wild-type gene. PCR yielded two bands of the expected size (Fig. 8B, lane 1), indicating that one abp2 allele was disrupted in the diploid cells. Heterozygous diploid cells were sporulated, and the resulting tetrads were dissected and analyzed by PCR as described above. Two spores yielded a 2.85-kbp band, indicative of the disrupted gene, and two spores yielded a 0.85-kbp band, indicative of the wild-type copy of the abp2 gene (Fig. 8B, lanes 2 to 5), as expected for a 2:2 segregation pattern. Tetrad analysis also indicated that the LEU2 marker cosegregated with the disrupted abp2 gene. PCR analysis of disruption yielding Δabp2 B, though the results are not shown, is detailed in Materials and Methods.

All spores analyzed gave rise to colonies, indicating that the abp2 gene is not essential for cell viability. However, though the haploid strain containing the disrupted abp2 gene Δabp2 A was able to grow between 15 and 36°C, its growth rate at 36°C was reduced compared to that of the wild-type strain. When incubated at 36°C, the Δabp2 B disruptant revealed a phenotype similar to that observed with Δabp2 A. To examine the phenotype of the Δabp2 A strain at high temperature, the viability and septation index were analyzed and compared to those for the wild-type strain grown at the same temperature (Fig. 9). When the strain containing the disrupted abp2 gene (Δabp2 A) and the wild-type strain were grown in liquid medium (YE medium plus supplements at 36°C), a substantial loss of viability was observed with Δabp2 A cells whereas the wild type was hardly affected (Fig. 9A). Staining with Calcofluor showed that 42% of the Δabp2 A cells contained septa at 26 h whereas 5% of wild-type cells contained septa (Fig. 9B and C). DAPI staining indicated that approximately 30% of the Δabp2 A cells contained an abnormal nuclear DNA content at 44 h. FACS analysis of Δabp2 A cells showed that most cells contained a 4 N DNA content at 18 to 24 h whereas the wild-type cells showed a 2 N DNA content (data not shown). A similar but more dramatic phenotype was observed after incubation of the Δabp2 A strain at 36°C on YE plates (Fig. 10). A fraction (20%) of the mutant cells had an abnormal shape and increased size compared to that of wild-type cells. Similar results were obtained with the Δabp2 B cells. These results suggest that though the abp2 gene is not essential for viability, at high temperatures the null mutants display a number of anomalies.

FIG. 9.

Comparison of abp2 and the wild-type null mutant cells cultured at 36°C. (A) Δabp2 A and wild-type (WT) cells were grown at 36°C in YE medium (plus supplements). Cells were collected at the indicated times and plated for viability as described in Materials and Methods. (B) At the same times, cells were collected and stained with Calcofluor to measure the septation. The percentage of cells containing a septum were plotted as a function of time at 36°C. (C) Calcofluor staining of wild-type and Δabp2 A cells. Cells were collected after 20 h, fixed, and stained with Calcofluor. Cells were observed with Nomarski optics and by Calcofluor staining fluorescence. Bar = 8 μm.

FIG. 10.

DAPI staining of wild-type and Δabp2 A cells. Cells were grown on YE plates (plus supplements) at 36°C for 48 h. They were then collected, stained with DAPI as described in Materials and Methods, and observed with Nomarski optics and by DAPI fluorescence. Bar = 8 μm.

To further study the function of Abp2, we examined whether these disruptants entered mitosis when DNA replication was blocked with hydroxyurea. The wild type and the two disruptants were incubated in MM in presence of 10 mM hydroxyurea. Samples were taken at various times and examined for viability and septation index. Analysis of septation index (Fig. 11A) at 4 h indicated that the wild-type and disruptant cells did not show septum formation, which suggested that the cells were arrested. FACS analysis indicated that the cells were arrested with a 1 N DNA content (data not shown). At this point the concentration of hydroxyurea was raised to 14 mM and the cells were incubated further. After 12 h of incubation with hydroxyurea, 35 and 85% of the Δabp2 A and Δabp2 B cells were septated, respectively, whereas less than 1% of the wild-type cells contained septa. The Δabp2 B cells were abnormally elongated compared to the wild-type and Δabp2 A cells. Many Δabp2 A and Δabp2 B cells contained septa that were abnormally positioned. Concomitant with the increase in number of septated cells, both disruptants suffered a loss of viability, but it was more pronounced in Δabp2 B cells (Fig. 11B). To examine whether the mutant cells grown in the presence of hydroxyurea entered mitosis in the absence of replication, cells were stained with DAPI (Fig. 12). Abnormal elongated cells were visible, with dark bands visualized by DAPI staining that represented septa. Most cells contained two nuclei which were abnormally located, whereas nuclei from the wild-type cells were normal. These results suggest that both Δabp2 A and Δabp2 B cells enter mitosis in the presence of hydroxyurea, leading to a loss of cell viability. In the case of Δabp2 B cells the phenotypes were more dramatic than those observed with Δabp2 A disruptant. Both disruptants enter mitosis with a delayed kinetics and do not undergo cytokinesis. The effects of hydroxyurea described above were observed only at the high levels of hydroxyurea used. At 5 mM hydroxyurea, no differences between wild-type and mutant cells were noted.

FIG. 11.

Analysis of wild-type, Δabp2 A, and Δabp2 B cells incubated in the presence of hydroxyurea. All three strains were incubated in MM in the presence of hydroxyurea at 30°C. (A) Cells were collected at the indicated times and stained with Calcofluor to measure the septation index. The percentage of cells containing septa was plotted as a function of time. (B) At the times indicated, cells were collected, diluted, and plated for cell viability as described in Materials and Methods and the percentage of viable cells was plotted as a function of the time for which the cells had been incubated in the presence of hydroxyurea. Experiments were also carried out at 36°C, and no differences from those carried out at 30°C were noted.

FIG. 12.

Observation of wild-type, Δabp2 A, and Δabp2 B cells incubated in the presence of hydroxyurea by DAPI staining and with Nomarski optics. All three strains were incubated in MM in the presence of hydroxyurea. Cells were collected at 12 h, fixed, and stained with DAPI. Bar = 8 μm.

We tested whether disruption of the abp2 gene altered resistance to UV radiation. Wild-type, Δabp2 A, and Δabp2 B cells were plated, irradiated with different levels of UV, and examined for viability. The cut5-T401/rad4 mutant, a temperature-sensitive mutant that is UV radiation sensitive, was used as a control (31). Both types of mutant cells were more sensitive to high doses (30 J/m²) of radiation than the wild type (Fig. 13). The cut5-T401/rad4 mutant showed a sensitivity similar to that observed with the disruptants.

FIG. 13.

UV radiation sensitivity of wild-type, Δabp2 A, and abp2 B cells. Cells were irradiated in the range of 10 to 50 J of UV radiation per m², and the survival relative to that of unirradiated controls was determined and plotted as a function of the UV dose. Experiments were carried out at 30°C. Identical results were obtained at 36°C.

DISCUSSION

We have purified and cloned a gene encoding a 75-kDa protein (Abp2). This protein binds to multimers of an oligonucleotide (MMACS) containing three overlapping near matches to the S. pombe ARS consensus sequence. DNase I footprint analysis indicated that Abp2 protected preferentially the Maundrell ARS consensus sequence (22) in the MMACS dimer. The binding of Abp2 to DNA was stimulated by Mg²⁺ and was independent of ATP.

Abp2 binds to the ars3002 sequence, forming several complexes detectable by gel shift analysis. DNase I footprint studies showed that Abp2 protected at least five regions within a 360-bp stretch of the ars3002 region. Abp2 bound preferentially at or near positions in ars3002 containing overlapping matches (at 9 of 11 bases or greater) to the Maundrell S. pombe ARS consensus sequence. Other regions containing a single match of 9 or 8 of 11 bases were not protected. The matches of 8 of 11 bases are very abundant in ars3002 since it is an AT-rich sequence. These results indicate that Abp2 recognizes sequences and/or structural elements provided by overlapping Maundrell ARS consensus sequences or sequences resembling this consensus. Portions of the essential α and β elements (13) of ars3002 were protected by Abp2, suggesting that Abp2 might be one of a number of distinct proteins that bind to these essential elements. Protection of the asymmetric AT sequence (consensus sequence of Zhu et al. [37]) in the β element could not be determined since this region is not cleaved by DNase I. Attempts to carry out hydroxy radical footprint analysis with ferrous ammonium sulfate were unsuccessful because this reagent blocked the interaction of Abp2 with MMACS and with ars3002.

We cloned the Abp2 cDNA and found that it contained a single open reading frame encoding a protein with a calculated molecular mass of 60 kDa, and we expressed it in E. coli as a fusion protein with GST. Upon removal of the GST tag with thrombin, the Abp2 protein product migrated slower than the predicted molecular mass (60 kDa), migrating as a 75-kDa protein, similar to the apparent molecular mass of the protein isolated from S. pombe extracts (data not presented). The basis for this discrepancy is unknown.

The 5′ untranslated region of Abp2 is 845 nt long, which is unusually long for an S. pombe cDNA (which are usually 100 to 200 nt long). This region contains two AUG start codons, neither of which contains an adenine residue in the −3 position, in contrast to the third AUG (the initiation start site of Abp2), which does. An adenine nucleotide in the −3 position is highly conserved in translation of initiation start sites in higher eukaryotes. At least four cDNAs of S. pombe in which the first AUG is not translated (the second is used instead) because of poor context have been reported (36).

The isolated cDNA and genomic clone encode the Abp2 protein. This conclusion is based on the following observations: (i) the amino acid sequence corresponding to the open reading frame contained the five distinct tryptic peptide sequences derived from purified Abp2, (ii) the GST-Abp2 protein binds AT-rich sequences in a manner similar to that observed with the purified protein, and (iii) polyclonal antibodies against the E. coli-expressed Abp2 recognized a 75-kDa protein following SDS-PAGE of a highly purified fraction of Abp2 (glycerol gradient fraction) that contained MMACS DNA binding activity.

Abp2 possesses a short region with strong homology to an AT-rich DNA binding domain found in HMGI and MIF2, the GRP motif (8, 24). A synthetic peptide (11-mer [TPKRPRGRPKK]) corresponding to the binding domain of HMGI specifically binds AT-rich DNA substrates in a manner similar to that seen with the intact protein, and it has been suggested that this structure resembles a hook (29). This GRP motif is also a functional part of the DNA binding domain of Hin recombinase, and deletion of the arginine residue within this motif abolished DNA binding activity (32).

Since the Abp2 protein contains the GRP motif, we examined whether this motif is necessary for DNA binding activity. For this purpose, we changed Arg 331 to Ala, Gly 332 to Ala, or Arg 333 to Ala, alterations that should disrupt the hook structure of the GRP motif. We also mutated Arg 329 to Lys, a change that should not alter the GRP binding domain. Abp2 containing Ala 332 or Ala 333 was devoid of DNA binding activity, whereas replacement of Arg 331 with Ala substantially decreased the DNA binding activity compared to that of the wild-type protein. A conserved change of Arg 329 to Lys did not affect the DNA binding activity. These experiments indicate that this GRP motif is essential for the DNA binding activity of Abp2.

Crystallographic studies of AT-rich DNA oligomers have shown that this sequence tends to be straight and possesses a narrow minor groove and a high degree of twist between base pairs (10). The interaction between the AT hook DNA binding domain of HMGI with DNA appears to be mediated by the recognition of the structure of the narrow minor groove of AT regions, rather than the nucleotide sequence (29). The antitumor drug netropsin and the dye Hoechst 33258, which contain secondary structures similar to that of the BD peptide (11-mer that contains the GRP motif from HMGI), also bind to the minor groove and are able to specifically compete the AT hook peptide (29). Since Abp2 contains an AT hook DNA binding domain, it is possible that its broad binding specificity is due to its ability to recognize the structure of the narrow minor groove, rather than to its interaction with a specific nucleotide sequence.

Though the disruption of the abp2 gene was not lethal, incubation of cells lacking the Abp2 protein at an elevated temperature (36°C) resulted in pleiotropic morphological changes. These included formation of multinucleated cells, fragmented nuclei, and septated cells with aberrant chromosome separation.

The loss of viability of Δabp2 B cells at high temperature was prevented by the transfection of a plasmid containing the wild-type abp2 gene (data not presented). The background expression of Abp2 in pREP1 vectors was sufficient to prevent the abnormal phenotypes observed at high temperature (4, 23). These results support the conclusion that the phenotypes observed with the cells with a disrupted abp2 gene at high temperature are due to the absence of the Abp2 protein. Furthermore, when extracts isolated from abp2 null mutant strains were incubated with labeled MMACS tetramer, only the Abp1-DNA complex (complex II [Fig. 1B]) was detected. No Abp2-DNA complex was observed.

We examined whether cells containing the disrupted abp2 gene entered mitosis when replication was blocked with hydroxyurea. After 12 h of incubation with hydroxyurea, 35 to 85% of the disruptant cells contained aberrantly positioned septa and were unusually long compared to wild-type cells. Concomitant with formation of septa, cell viability decreased substantially. Both types of disruptants showed aberrant DNA distribution, suggesting that they were subjected a mitotic catastrophe. These results suggest that when replication is blocked with hydroxyurea, both types of disruptants entered mitosis with delayed kinetics compared to that of S. pombe hus mutants (2, 3, 14) and did not undergo cytokinesis, suggesting that disruption of the abp2 gene deregulates the control of the entry into mitosis.

The hus mutants lack a checkpoint control that protects cells from undergoing mitosis with unreplicated or damaged DNA in the presence of hydroxyurea (2, 3, 14). The loss of this checkpoint results in a mitotic catastrophe leading to missegregation of chromosomal DNA that is trapped between two cells (cut phenotype). The results described above suggest that abp2 null cells resemble some hus mutants that enter mitosis with delayed kinetics with the exception that they do not undergo cytokinesis, typically observed in hus mutants after a mitotic catastrophe. Thus, in abp2 null cells the cytokinesis checkpoint is functional. Furthermore, abp2 null cells, like hus mutants, are sensitive to UV radiation, suggesting that Abp2 may play a role in radiation-induced cell cycle delay.

Though Abp2 was identified and purified based on its ability to bind to the ARS consensus sequence, its role in DNA replication is unclear. The firing of the ars3002 origin in both wild-type cells and cells in which the abp2 gene is disrupted was examined by two-dimensional gel analysis. No differences were noted (data not presented). The stabilities of plasmids containing the ars3002 origin were identical in both abp2 null and wild-type strains. These results suggest that Abp2 protein is not essential for origin replication.

Studies using affinity-purified antibodies to Abp2 indicate that Abp2 is localized to the nucleus and displays a punctated pattern (data not presented). Further genetic and biochemical analyses of this novel protein, which binds to S. pombe ars3002, should help define its role in DNA replication or in other processes.