Eur. J. Biochem. 238, 38-47 (1996) 0 FEBS 1996

CTBPlIRBPl, a Saccharomyces cerevisiae protein which binds to T-rich single-stranded DNA containing the 11-bp core sequence of autonomously replicating sequence, is a poly(deoxypyrimidine)-binding protein Masato IKEDA' *, Ken-ichi ARAI' and Hisao MASAI'

' Department of Molecular and Developmental Biology, Institute of Medical Science, University of Tokyo, Japan ' Department of Virology I, National Institute of Health, Japan

(Received 13 February 1996) - EJB 96 0200/2

South-Western screening of a glutathione-S-transferase fusion protein library constructed from the yeast Saccharomyces cerevisiae genomic DNA lead to isolation of core T-rich-strand-binding protein (CTBP) clones that bound to single-stranded DNA containing the T-rich-strand of the 11 -bp core sequence of autonomously replicating sequences. One of these clones, CTBP1, contains a portion of previously described RBPl which is an RNA-binding and single-stranded DNA-binding protein of S. cerevisiae. GST-CTBP1 as well as the full-length fusion protein with RBPl (GST-RBPZ) bind exclusively to the T- rich strand of the core sequence with an apparent dissociation constant of 5 nM, but not to the A-rich strand or double strand of the same sequence. Mutations within the core which reduce the number of T or C residues decrease the affinity of this protein. In keeping with this, binding of GST-CTBPI to the core sequence is efficiently competed by poly(dT), poly(dT-dC) or poly(dC), but not by poly(dA) or poly(dG) to significant extents. Among polyribonucleic acids, GST-CTBPI binds to poly(U) and poly(1) with greatest affinity, whereas GST-RBP1 binds to RNA in a rather non-specific manner. In no cases was affinity for RNA greater than that for DNA. Our results indicate that CTBPURBPI is a polydeoxypyri- midine-binding protein of S. cerevisiae. CTBPl contains two sets of an RNA-recognition motif (RRM) and a glutamine stretch. The binding affinity of the N-terminal or C-terminal set containing one RRM and one glutamine stretch is nearly two orders of magnitude lower than that of the wild-type CTBPI containing both sets. The isolated N-terminal or C-terminal RRM alone (RRM1 and RRM2, respectively) is sufficient for binding nucleic acids with the binding specificity similar to that of the wild-type RRM, although the binding affinity of the isolated RRM2 is nearly two orders of magnitude lower than that of RRMI. Our results indicate that the two RRMs present in CTBPURBPI have differential binding affini- ties and that the high affinity of RRM for polydeoxypyrimidine results from synergy between two lower- affinity RRMs.

Keywords: single-stranded DNA-binding protein; replication origin; glutathione S-transferase fusion pro- tein ; RNA recognition motif; poly(de0xypyrimidine).

Single-stranded DNA-binding proteins are essential for the processes of DNA replication, DNA recombination and DNA repair. In Escherichia coli, tetrameric SSB plays essential roles in all of the above three processes (Lohman and Ferrari, 1994). In replication of duplex DNA, SSB stabilizes the replication forks by binding to the unwound DNA region (Komberg and Baker, 1992). SSB plays essential roles also in the replication of ssDNA replicons such as M13, G4 or 4x174 by stabilizing the secondary structure of the origin region or by generating a higher order structure favored for initiation (Sun and Godson, 1993; Masai and Arai, 1993). In eukaryotic cells, hetero-trimeric

Correspondence to H. Masai, Department of Molecular and Devel- opmental Biology, Institute of Medical Science, University of Tokyo, 4-6-1, Shirokanedai, Minato-ku, Tokyo, Japan 108

Abbreviations. ARS, autonomously replicating sequence ; CTBP, core T-rich-strand-binding protein; EMSA, electrophoretic-mobility- shift assay; GST, glutathione S-transferase; RRM, RNA recognition mo- tif; Q stretch, glutamine stretch.

Note. In this report, the coding frame of RBPl initially isolated as a glutathione S-transferase clone which binds to core T-rich strand, as well as to its deletion derivatives, are shown as CTBPl to distinguish them from the full-length gene which is expressed as RBPl.

RF-A (or RP-A) is an essential ssDNA-binding protein (Still- man, 1989; Challberg and Kelly, 1989). RF-A is required for in vitro simian virus 40 DNA replication reconstituted from puri- fied proteins, and the genes encoding yeast RF-A subunits are essential for viability (Brill and Stillman, 1991). RF-A facilitates DNA replication not only by binding to unwound ssDNA re- gions near the origin but also by interacting with various pro- teins such as DNA polymerase a, transcription factors and possi- bly initiatorhelicase required for initiation (Donreiter et al., 1992; He et al., 1993; Li and Botchan, 1993; Dutta, 1993). RF- A binds to ssDNA in a relatively non-specific fashion as pre- dicted from its role as a general helix-destabilizing protein (Kim et al., 1992). In contrast, other classes of ssDNA-binding pro- teins that bind to specific sequences are also known (Bergemann and Johnson, 1992; Tzfati et al., 1992; Carmichael et al., 1993). These proteins have been implicated in DNA replication, tran- scription and recombination.

Chromosomal replication origins of yeast Saccharomyces cerevisiae have been identified as autonomously replicating se- quences (ARS) which contain the l l -bp core sequences essential for replication (Newlon, 1988). The presence of the core and

Ikeda et al. ( E M J. Biochem. 238) 39

Table 1. Oligonucleotides used in this study. The sequences are derived from ARS307, a chromosomal replication origin on the chromosome I11 of S. cerevisiae (Plazkill and Newlon, 1988). Bold letters indicate the 11-bp core sequence of ARS. Letters in lower case show the base substitutions within the core.

Oligonucleotide Sequence

T-ARS T-ARS - 2 5 T-ARS-mutl T-ARS-mut2 T-ARS-mut3 T-ARS-mut4 T-ARS-mut5 T-ARS-mut6 T-ARS-mut7 T-ARS-mut8 T-ARS-mut9

A-ARS

5’-TATCGA1’TTTTATTTATGTTTTCTTCTTCACACATGGGTTACTGCA-3’ 5’-GATC‘CTTTATTTATGTTTTCTTCTG-3’

5‘-TATCGATTTTTAgTTATGTTTTCTTCTTCACACATGGGTTACTGCA-3’ 5’-TATCGAl1TTTTAgCTATGTTTTCTTCTTCACACATGGGTTACTGCA-3’ 5’-TATCGA1’TTTTAgCgATGTTTTCTTCTTCACACATGGGTTACTGCA-3’ 5’-TATCGATTTTTAgcgcTGTTTTCTTCTTCACACATGGGTTACTGCA-3’ 5‘-TATCGATTTTTATTTATGTTgTCTTCTTCACACATGGGTTACTGCA-3‘ 5’-TATCGATTTTTATTTATGTCgTCTTCTTCACACATGGGTTACTGCA-3’ 5’-TATCGAl1TTTTATTTATGgCgTCTTCTTCACACATGGGTTACTGCA-3’ 5’-TATCGATTTTTATTTATCgCgTCTTCTTCACACATGGGTTACTGCA-3’ 5‘-TATCGATTTTTAgCgctCgcgTCTTCTTCACACATGGGTTACTGCA-3’

5’-TATGCAC:TAACCCATGTGTGAAGAAGAAAACATAAATAAAAATCGA-3’

core-like sequences within the ARS is likely to contribute to initiation of DNA replication through case of unwinding due to low free energy required for strand opening (Umek and Kowal- ski, 1988; Umek et al., 1989). Replisomes or primosomes may be assembled at these core or core-like sequences after initial unwinding or melting within ARS. We have searched for pro- teins that bind to ssDNA sequences derived from a yeast ARS, since they may be components of the replisome assembled at the origin or may regulate the process of initiation. We obtained CTBPl which contained two sets of an RNA-recognition motif (RRM) and a glutamine stretch (Q stretch) as a gene encoding the protein that bound to the T-rich strand of the ARS core se- quence. RRM, a loosely conserved 80-amino-acid sequence do- main, contains two conserved motifs, an octamer sequence termed ribonucleoprotein consensus sequence (RNP-CS or RNP-1) and a less conserved hexamer (RNP-2), and has been identified in a number of proteins from both prokaryotes and eukaryotes that bind to RNA andor ssDNA (Birney et al., 1993; Burd and Dreyfuss, 1994; Kenan et al., 1991; Nagai et al., 1995). Competition assays with various synthetic DNA and RNA revealed that CTBPI as well as the full-length RBPl bound to polydeoxypyrimidine with the highest affinity. We fur- ther dissected the functional domains required for nucleic acid binding. The possible function of CTBPI in DNA replication and in recognition of altered DNA structures on the genome will also be discussed.

MATERIALS AND METHODS

Oligonucleotides and plasmid DNA. Synthetic oligonucle- otides (Table 1) T-ARS and T-ARS-25 are derived from the T- rich strand of S. cerevisiae ARS307 (Plazkill and Newton, 1988). T-ARS-mut-1-9 contain base substitutions in the core sequence of T-ARS, and A-ARS and A-ARS-25 contain the A- rich-strand sequence complementary to T-ARS and T-ARS-25, respectively. The oligonucleotides were 5‘-end labeled with T4 polynucleotide kinase and [y-3’P]ATP, and the labeled oligonu- cleotides were purified by a Nick Column or microspin (Phar- macia). GST-RBPI containing the entire coding frame of RBPl was a gift from Dr Fan-Jen S. Lee (National Taiwan University, Taiwan).

Buffers. The following buffers were used; binding buffer (50 mM Hepes/KOH, pH 8.0, 50 mM potassium glutamate, 0.5 mM EDTA, 0.5 mM dithiothreitol, 0.1 mg/ml BSA and 10% glycerol) ; washing buffer (binding buffer without BSA) ; gluta-

thione S-transferase (GST) buffer (washing buffer containing 0.5 % sodium sarcosine); EMSA binding buffer (binding buffer containing 10 mM magnesium acetate).

Construction and screening of a GST-fusion protein li- brary. Derivatives of pGEX-3X, PMIlOl, pMI102 and pMI103, which have a BglII site in three different coding frames, were constructed by inserting a BglII linker (GAGATCTC, GGA- GATCTCC or GGGAGATCTCCC) at the SmuI site of pGEX- 3X (Smith and Johnson, 1988). The genomic DNA of a yeast strain NNYII (Nomoto et al., 1990), prepared as previously de- scribed, was partially digested with Sau3AI and the DNA frag- ments of 1-2 kb were isolated from an agarose gel by electro- elution, were ligated with mixtures of BglII-digested pMI101, pMI102 and pMI103, and were transformed into Escherichia coli MC1061 (Meissner et al., 1987). The number of indepen- dent colonies in the library was approximately 1 X106. Screening of the library, the details of which will be described elsewhere, was conducted as follows. Colonies grown on Luria-Bertani containing 50 pg/ml ampicillin were transferred to nitrocellulose filters (Schleicher & Schull) which were placed onto Luria-Ber- tani plate containing 1 mM isopropyl-P-D-thiogalactoside for 3 hours to induce expression of fusion proteins. The nitrocellulose filters were first incubated in 6 M guanidine hydrochloride to denature the proteins, then renatured in solutions with decreas- ing concentrations of guanidine hydrochloride and finally in the binding buffer. Filters were incubated at 30°C for 30 min in the same binding buffer containing 50 pg/ml poly(d1-dC) with 0.2 nM 5’-end-labeled oligonucleotide T-ARS, washed in the washing buffer and autoradiographed.

Preparation of GST proteins and cleavage with factorXa. E. coli C600 lon- (Mizusawa et al., 1983) was transformed with each GST-fusion plasmid, and extracts were prepared as will be described elsewhere. Briefly, cells were lysed by sonication in the GST buffer and GST-fusion proteins were purified by gluta- thione- Sepharose-4B affinity column (Pharmacia) and were eluted with the GST buffer containing 5 mM reduced glutathi- one (Frangioni and Neel, 1993). GST fusion proteins were di- gested with factorXa (New England BioLabs) in the binding buffer containing 1 mM CaC1, at 23°C for 12 hours.

Filter-binding assays. An end-labeled oligonucleotide (0.2 nM) were incubated at 30°C for 30 min with a purified GST fusion protein in binding buffer (100 pl) containing 0.5 pg po- ly(d1-dC). The reaction mixtures were filtered through a nitro- cellulose membrane filter (type HA, 0.45 pM, Millipore) soaked in 0.5 M KOH at room temperature for 30 min to reduce nonspe- cific binding of oligonucleotides to the filter (Yokota et al.,

40 Ikeda et al. (EUK J. Biochem. 238)

1979). The filter was washed twice with 1 ml washing buffer and radioactivity retained on the filter was determined with a liquid scintillation counter. In competition assays, a 50-fold mass excess of a competing polydeoxyribonucleotide, polyribonucleo- tide (Sigma) or oligonucleotide was mixed with labeled oligonu- cleotide, then added to the binding reactions.

Electrophoretic mobility shift assays (EMSA). End-la- beled oligonucleotide (0.2 nM) was incubated at 30°C for 30 min with each purified protein in the EMSA-binding buffer (50 PI) containing 0.5 pg poly(d1-dC). The reaction mixtures were applied to 4% (monohis = 19 : l ) polyacrylamide gel and electrophoresis was carried out at 4°C in 1X TBE buffer con- taining 10 mM magnesium acetate. The gel was then dried and autoradiographed by Fuji BAS-2000.

Southwestern blotting. A total extract or 15 pmol each purified GST fusion protein was fractionated by 10% SDS/ PAGE and was transferred to BAS-85 nitrocellulose filter (Schleicher & Schiill). The filter was denatured, renatured as described above for Southwestern screening, incubated at 30°C for 3 hours with "P-labeled T-ARS (0.2 nM) in 5 ml binding buffer containing 10 pg poly(d1-dC), washed with the washing buffer and autoradiographed.

Preparation of yeast extract. Yeast extract was prepared from a strain CB023 (Brenner and Fuller, 1992) by vigorously vortexing the cells in the presence of glassbeads (Dunn and Wobbe, 1993).

RESULTS

Analysis of total yeast extract by Southwestern using an oli- gonucleotide probe containing the T-rich strand of ARS core sequence. The chromosomal origins of budding yeast are char- acterized by the presence of multiple copies of the 11 -bp ARS core and its near matches on both strands of DNA (Plazkill and Newlon, 1988). The AT-rich nature of the core sequence is ex- pected to contribute to strand-opening prerequisite for initiation of DNA replication (Umek et al., 1989). In fact, DNA regions with low free energy necessary for helix unwinding frequently coincide with those sequences where the core and its near matches are present in a cluster (Shirahige et al., 1993). On the assumption that replisomes are constructed on unwound DNA at the origins, we have looked for proteins that bind to ssDNA sequences containing the T-rich strand of the core sequence. Southwestern analysis of total yeast extract with a 32P-labeled oligonucleotide probe (T-ARS) derived from the T-rich strand of a yeast chromosomal origin, ARS307 (Palzkill and Newlon, 1988), revealed that multiple proteins bound to this oligonucleo- tide (Fig. 1 A), some of which failed to bind or showed reduced binding to a mutant oligonucleotide containing eight base substi- tutions within the core sequence (T-ARS-mut9; Fig. 1 B). This result indicates the presence of multiple proteins in the yeast extract which bind to this oligonucleotide with various degrees of affinity for the core sequence.

Southwestern screening of core T-rich-strand-binding pro- teins (CTBP). In order to facilitate cloning of genes encoding proteins that bind to T-ARS, we constructed a GST-fusion pro- tein library from the S. cerevisiae genomic DNA, which could be readily screened by a South-western method (Ikeda, M., Arai, K. and Masai, H., unpublished results). The cell extracts, pre- pared from the five positive clones (core T-rich-strand binding proteins 1-5), were fractionated on an SDS/PAGE and were analyzed by Coomassie brilliant blue staining and by Southwest- ern screening using the '2P-labeled T-ARS. Each extract showed the presence of protein band(s) that are induced after isopropyl-

Fig. 1. Southwestern analysis of yeast total extracts. An extract, pre- pared from yeast S. cerevisiae strain CB023, was run on 12.5% SDS/ PAGE. Proteins were transferred to a nitrocellulose filter and were ana- lyzed by Southwestern as described in Materials and Methods. Probes used were: A, T-ARS; B, T-ARS- mut9.

P-D-thiogalactoside addition, and some of these bands as well as those not identified on Coomassie brilliant blue staining bound to the probe (data not shown). These results show that fusion proteins are actually synthesized from the isolated clones and are responsible for the binding to the probe. The presence of multiple binding proteins in each extract probably reflects pro- teolytic degradation of induced proteins. CTBP2, containing a previously unidentified open reading frame on chromosome 111, directed synthesis of a 38-kDa protein (GST-CTBP2) which bound to T-ARS as well as to A-ARS, an oligonucleotide con- taining the opposite strand, and its binding activity was rather feeble in the extract (data not shown). CTBPS contained a coding frame from SSB1, a previously reported S. cerevisiae ssDNA-binding protein (Jong et al., 1987). It gave a strong signal on Southwestern analysis, but we failed to detect binding in the extract made from E. coli cells containing CTBP5. Cross- hybridization analysis indicated that the other CTBPs (CTBPl , -3 and -4) shared the same DNA sequence. CTBPl fusion protein (GST-CTBP1) was purified from C600 Ion- cells (Mizu- sawa and Gottesman, 1983) harboring CTBP1, and its binding property was examined by filter-binding assays. Binding of GST-CTBP1 to T-ARS was rather weak in the absence of salts and divalent cations and was highly stimulated by the presence of salt; with sodium acetate or potassium acetate, which has low ionic strength, binding was stimulated by more than tenfold at 200 mM ionic strength and stayed constant up to 500 mM ionic strength, whereas NaCl and KCI, which stimulated the binding up to 200 mM concentration, were inhibitory at higher concen- trations. Generally, salts containing the chloride anion inhibited reactions at higher concentrations (data not shown). Magnesium acetate at 10 mM also stimulated binding to the maximum level in the absence of salt (Fig. 2A and B). Addition of salt in the presence of magnesium did not alter the efficiency of binding (data not shown). Therefore, we did all the following reactions in the presence of 10 mM magnesium acetate and 50 mM potas- sium glutamate. GST-CTBP1 bound to T-ARS with an apparent equilibrium dissociation constant ( K J of 5 nM under the opti- mum condition, as determined from the concentration of the pro- tein required for 50% saturation of binding (Fig. 2A; Kelly et al., 1976). In the absence of magnesium, it did not bind to A-ARS in 10mM potassium glutamate or even at higher salt concentrations (Fig. 2A and data not shown). Feeble binding to

Ikeda et al. (EM J. Biochem. 238) 41

1.0 ,

0.0- 0 60 120 180 240 300 2

0.4

0.3

0.2

0.1'

0 . I , , , ,

I 0 1 2 3 4 5 6

CTBP-1 (nM)

Fig. 2. Specific binding of GST-CTBP1 to ARS core-derived T-rich ssDNA. (A) and (B) The labeled oligonucleotide (0.2 nM each; circles, T- ARS; squares, A-ARS; triangles, T-ARS-25) was incubated with increasing concentrations of GST-CTBP1 in the binding buffer in the presence (filled circles) or absence (open circles) of 10 mM magnesium acetate, and filter binding was conducted as described in Materials and Methods. The washing buffer for the former reactions contained 10 mM magnesium acetate. (C) GST-CTBP1 (circles) and GST-RBP1 (squares), either undigested (filled symbols) or digested with FactorXa (open symbols), were incubated with the labeled T-ARS (0.2 nM) under the standard binding conditions containing 10 mM magnesium acetate.

A-ARS was observed in the presence of 10mM magnesium, although the bound fraction did not exceed 5 % of the input DNA at 300 nM of the added protein (Fig. 2A). GST-CTBP1 also bound to T-ARS-25, a 25-nucleotide sequence containing a portion of T-ARS (46 nucleotides), although the affinity was reduced (Kd=20 nM, Fig. 2B). It did not bind to the 25-nucleo- tide A-ARS-25 containing the opposite strand (data not shown). We also tested the binding of CTBPl freed from the GST por- tion by factorXa digestion (Fig. 2C). Free CTBPI bound to T- ARS with a Kd of 5-10 nM, showing that the GST portion does not affect the binding property. The decrease in the extent of binding with the digested proteins may be due to partial inactiva- tion of the proteins during the digestion. These results show that GST-CTBPI binds specifically to the T-rich strand of DNA con- taining the 11-bp core sequence of ARS. We decided to further characterize the structure and binding specificity of this protein.

GST-CTBP1 preferentially binds to polydeoxypyrimidine. A more detailed binding specificity of GST-CTBP1 was examined by competition assays with oligonucleotides containing base substitutions within the 11-nucleotide core sequences of the 46- nucleotide T-ARS. T-ARS-mut9 containing substitutions at eight positions showed dramatically reduced affinity to GST-CTBP1 ; binding was inhibited only by 50% in the presence of this mu- tant oligonucleotide at a 50-fold excess over the radiolabeled probe. Other mutant oligonucleotides, whose base substitutions were known to completely inactivate origin function (van Houten and Newlon, 1990), were only slightly less efficient competitors than the wild-type oligonucleotide. Among them, only T-ARS-mut3, in which T residues 2-4 of the core were replaced with GCG, showed significantly reduced affinity (Fig. 3A). Interestingly, T-ARS-mut4 with an additional A to C substitution at the fifth position was as good a competitor as the wild-type oligonucleotide. These results indicate that, although GST-CTBP1 binds exclusively to the T-rich strand of the core, its binding does not require specific core sequences essential for the origin function. It appeared rather that the binding affinity of CTBPl was generally correlated with the number of T and C residues. Therefore, we next examined various polydeoxyribo- nucleic acids as competitors in binding assays. GST-CTBP1 bound to poly(dT) with the highest affinity, and to poly(dC) with one order of magnitude lower affinity (Fig. 3 A and B). It bound to the alternating polydeoxyribonucleotide poly[dT-dC] with an affinity similar to that to poly(dT), much more inefficiently to poly(dG) and did not bind to poly(dA). We also tested polyribo-

nucleic acids in competition assays. Among polyribonucleic acids tested, it bound to poly(U) and poly(1) with affinities simi- lar to that of poly(dT) and poly(dC), respectively. It showed very little binding to poly(A), poly(G), poly(C) or double-stranded RNA (Fig. 3A and B). In competition assays, GST-free CTBPl exhibited a binding specificity almost identical to that of undi- gested GST-CTBP1 (Fig. 3A).

Domains of CTBPl required for binding to nucleic acids. Nucleotide sequence analysis of CTBPl , -3 and -4 revealed that they contained portions of the coding frame of a S. cerevisiae gene, RBPI, encoding a protein with two RRMs and two Q- stretch sequences located 3' to each RRM (Fig.4A; Lee and Moss, 1993). The cloned segment in CTBPl contained both of the two RRMs and Q stretches, while CTBP4 lacked the 5' half of the N-terminal proximal RRM. CTBP3 was identical to CTBP1. In order to determine the domain of CTBPl required for nucleic acid binding, we constructed a number of GST fusion proteins containing different portions of the CTBP 1, purified each of them (Fig. 4B) and assayed their binding activities in the filter-binding assays. Deletion of a part or all of the C-termi- nal RMM (RRM2; HA, KA and Ed) or of the N-terminal RMM (RRMI ; AA and AE) decreased the affinity by an order of mag- nitude comparable to that of the wild-type GST-CTBP1 (Fig. 5A). If the remaining RRM was disrupted by further dele- tion, binding activity was completely lost (Fig. 5A, AA and AH). AK, which lacked the RRMllQ stretch and the first eight amino acids of RRM2, still bound to DNA, although the entire RNP-2 of RRM2 (LFVGDL) was replaced by a sequence de- rived from the vector (GRGIPG). Although the latter sequence does not contain a conserved aromatic residue (second position) or leucine (sixth position), secondary-structure prediction indi- cates that it could form a P-sheet structure (data not shown). Ad, in which the last nine amino acids of RRMl were (RVAY- ATPRN) were replaced by a vector-derived sequence (KFIVTD), did not bind to nucleic acids, although it contained both RNP-2 and RNP-1 intact. The C-terminal end of RRM con- tributes to the fourth p sheet, which constitutes an integral part of the pa@@ RRM structure. Either of the RRMs was sufficient for binding (Fig. 5B, RRMl and RRM2), indicating that the Q stretches are not required for binding. The isolated RRMl showed a higher affinity for DNA than Ed containing both RRMl and a Q stretch. Under the standard binding conditions, the binding affinity of the isolated RRM2 was lower than that of RRMl by more than an order of magnitude, although AE

lkeda et al. ( E m J. Biochem. 238)

CTBP-1 1 .o - 0.8 -

I

1 .o - 0.8 - 0.6 - 0.4

0.2 - 1 .o - 0.8 - AE

. -

1 n-l I . ." 0.8

0.6

0.4

0.2

0.0

RRMl

RRM2

- POlY[dTl - POlYtdCI - POlY[UI - POlY[lI

0.0 . , . , . , . , . , . 0 10 20 30 40 50 60

Competitor (ratio) Fig. 3. Specificity of nucleic acid binding of GST-CTBPl and its derivatives. (A) Filter-binding assays were conducted as described in Materials and Methods. The reaction mixtures were incubated with 5 , 5 , 30, 30, 10 and 50 nM of CTBP-I, FactorXa-digested CTBPI, Ed, AE, RRMl and RRM2 as indicated and 0.2 nM of the labeled T-ARS, respectively, in the presence of SO-fold mass excess of various competitor nucleic acids as shown. (B) Binding of GST-CTBP1 to T-ARS in the presence of various amounts of competitors. The mass excess of each competitor over the labeled oligonucleotide is shown as a ratio. In both (A) and (B), the extent of the binding is shown as relative binding, values relative to that in the absence of a competitor which is taken as 1 .

containing both RRM2 and the Q stretch bound to the probe with an affinity similar to that of Ed (RRM1 fQ-stretch; Fig. 5B). These results indicate that Q stretches, although not essential for binding, can either enhance (in case of RRM2) or decrease (in case of RRMI) the binding affinity of an RRM. It is also obvious that Ed and dE, containing one set of RRM and

a Q-stretch sequence, bound to the probe with an efficiency one or two orders of magnitude lower than that of GST-CTBPI con- taining two sets of RRM and a Q stretch (Fig. 5B and C). In the presence of a 10-fold excess of poly(d1-dC) (5 pg), binding of the isolated RRM2 was almost completely abolished and that of other deletion derivatives containing one set of RRM and a Q

Ikeda et al. ( E m J. Biochem. 238) 43

Fig. 4. Schematic drawings of structure of CTBPURBPl segments contained in various deletion derivatives of GST-CTBP1. (A) Open box at the top represents the 672-amino-acid coding frame of RBPl. The black and hatched regions in the box indicate the Q stretches and RRMs, respectively, with the numbers (amino acid residues) in the parentheses defining the extent of each region. Thick lines under the box indicate the extent of the coding regions contained in each GST fusion protein, with the numbers showing the amino acid residues at the ends of the inserts. (B) 20 fmol each purified GST fusion proteins, as indicated on top of each lane, was loaded onto 12.5 % SDSPAGE, and proteins were visualized by Coomassie brilliant blue staining. MW, molecular-mass markers. GST indicates the non-fused GST protein.

stretch showed nearly two orders of magnitude lower affinity than the wild-type nucleotide (Fig. 5 C). The DNA-binding ac- tivity of each deletion derivative after digestion with factor Xa was basically similar to that without digestion (data not shown) except that deletions lacking a portion of RRM2 and the Q stretch (HA, KA and Ed) bound to T-ARS with less affinity after digestion than before digestion. We have concluded that RRM2 of CTBP1, although similar to RRMl in its primary structure, is a much weaker binding domain. We also speculate that a com- bination of two low-affinity binding units can create a high-af- finity binding module.

Results obtained in EMSA were similar to those obtained with filter-binding assays (Fig. 6). In EMSA with 5 nM GST-

0.5

CD 0.2 S

E B z 0.0 1. 0 1 10 100

1 .o

0.8

0.6

0.4

0.2

0.0 10 100 0 1

.y

CI .- > Protein (nM) Protein (nM) m

0.8 -1 A

1 .o 1 .o

0.8

0.6

0.4

CD 0.2 S 0.2

E

0.8

0.5

B z 0.0 0.0 a 0 10 100 1 10 100 0 1

Protein (nM) .- > Protein (nM) CI m ." - - CTBP-1 Q) 1.0

+ HA 0.8 -uA

- E A 0.6 -AA

--t AH 0.4 - AK

--+- AE 0.2 + AA

+ RRMl 0.0 + RRMP

a

0 1 10 100 + GST Protein (nM)

Fig. 5. Binding of GST-CTBP1 and its deletion derivatives to T-ARS in the filter binding assays. (A), (B), (C) The labeled T-ARS (0.2 nM) was incubated with vaned concentrations of GST-CTBPI or its deletion derivatives, as shown in the figure, in EMSA binding buffer and filter- binding assays were conducted as described in Materials and Methods except that the washing buffer contained 10 mM magnesium acetate. (C) Ten-fold excess of poly(d1-dC) (5 pglassay) over the standard condition was added in the reactions.

CTBPI, two major protein-DNA complexes were detected. One stayed near the origin of the gel and the other migrated into the gel. The former complex may contain multiple molecules of the proteidmolecule probe. Both complex disappeared upon compe- tition witha 50-fold excess of unlabeled probe DNA but not by the mutant oligonucleotide containing eight nucleotide substitu- tions in the core sequence, indicating that they are specific com- plexes. Competition with A-ARS completely eliminated the complex formation due to generation of duplex DNA, to which GST-CTBPI does not bind. With deletion derivatives containing the N-terminal RRMl (HA, KA, Ed and RRMl), specific com- plexes were detected at both high (100 nM) and low (10 nM) concentrations of the proteins. In contrast, AK, AE and AA con- taining the C-terminal RRM21Q stretch as well as RRM2 gener- ated a smear which migrated between the free probe and the top of the gel. This may indicate that the complexes of the latter deletion derivatives are unstable and dissociate during the elec- trophoresis.

Binding was also examined by Southwestern analysis. Puri- fied GST fusion proteins were transferred to a nitrocellulose fil- ter and Southwestern blotting was performed with the labeled oligonucleotide T-ARS (Fig. 7). In this assay, T-ARS bound to the GST fusion protein containing RRM2 (C-terminal RRM; AA and AE). The wild-type CTBPI containing two RRMs bound more efficiently than AA or AE, consistent with the results of filter-binding assays. Binding to GST fusions containing RRMl (N-terminal RRM; HA, KA and Ed) was very inefficient in Southwestern blots. We do not know the reason for this.

Binding specificity of GST fusion proteins containing one RRM. Although either of the RRMs is sufficient for binding to ssDNA, it is still possible that the specificity of binding may be affected by the number of RRMs or by the presence of a Q

44 Ikeda et al. (Eul: J. Biochem. 238)

- Protein (nM) I .

Fig. 6. Binding of GST-CTBP1 and its deletion derivatives to T-ARS. The end-labeled T-ARS was incubated with the indicated GST fusion proteins in EMSA binding buffer as described in Materials and Methods. (A) 5 nM GST-CTBPI was incubated with the labeled T-ARS in the presence of a 50-fold excess of unlabeled oligonucleotide as indicated, except for the right-most lane where only T-ARS was incubated. (B) Reactions were carried out in the presence of two different concentra- tions of each protein; 10 nM (left lane of each pair) and 100 nM (right). GST indicates the non-fused GST protein. In both (A) and (B), the reac- tion mixtures were applied to 5 % PAGE and DNA-protein complexes were identified by autoradiography. The thick arrows indicate two dis- tinct complexes generated. free and ori indicate the position of uncom- plexed oligonucleotides and the top of the gel, respectively.

~ ~ ~

RBPl

1-

w 05- E 0 C

.-

.- m 0--

.- B digested RBPl -

CI 0 - d '-

0 5-

0

Fig. 8. Binding properties of GST-RBPl. Filter-binding assays were conducted as described in Materials and Methods. (A) GST-CTBPI (cir- cles) and GST-RBPI (squares) were incubated with labeled T-ARS (open symbols) or A-ARS (filled symbols). (B) The reaction mixtures were incubated with 100 nM undigested GST-RBP1 or FactorXa-digested GST-RBP1 and 0.2 nM labeled T-ARS in the presence of a %-fold ex- cess of various competitor nucleic acids as indicated. The extent of the binding is shown as relative binding, values relative to that in the ab- sence of a competitor which is taken as 1.

Fig. 7. Binding of GST-CTBP1 and its deletion derivatives to T-ARS as detected by Southwestern analysis. 1 pg purified proteins, as indi- cated at the top of each lane, were separated on 12.5% SDSIPAGE, transfeffed to a nitrocellulose filter, and incubated with the labeled T- ARS oligonucleotide as described in Materials and Methods. The filter was washed and autoradiographed. GST indicates the non-fused GST protein.

stretch. Therefore, we conducted competition assays using puri- fied GST fusion proteins containing only one RRM (Fig. 3A). The binding specificity of Ed (containing the N-terminal RRMI/ Q-stretch) and dE (containing the C-terminal RRM2Qstretch)

was similar to that of CTBPl. However, both were more specific to poly(dT) or poly(dT-dC); only 25% or 50% inhibition was observed in the presence of a 50-fold excess of poly(dC) with Ed and dE, respectively. Ed had a significantly higher affinity to ssDNA than to RNA; a 50-fold excess of poly(U) and poly(I), which inhibited binding of CTBPI by more than 90%, inhibited binding of Ed by only 50% and 25 %, respectively. The presence of a Q stretch did not significantly affect the binding specifities RRMI and RRM2 (compare Ed with RRM1, LIE with RRM2 in Fig. 3A). We have concluded that each RRM by itself binds to nucleic acids with a specificity similar to the wild-type CTBPI, although subtle differences between the two RRMs and between the polypeptide with a single RRM and that with double RRMs can be detected.

Binding property of full-length RBP1. Although CTBPl con- tains all the major binding determinants of RBPl, other domains of RBPl might affect the nature of binding. Therefore, we puri- fied GST-RBPI containing the entire polypeptide of RBPl and characterized its binding to nucleic acids. GST-RBP1 and free RBPl bound to T-ARS with affinities similar to that of CTBPl, but not to A-ARS (Fig. 2 C and Fig. 8A). Competition assays

Ikeda et al. (Eul: J. Biochem. 238) 45

with various synthetic nucleic acids showed that GST-RBPI also bound to poly(dT) and poly(dC) with highest affinity, as was observed with GST-CTBPI (Fig. 8 B). As GST-CTBPI , GST- RBPl did not show any binding to poly(dA). However, it showed some affinity for poly(dG). Although poly(U) was a most efficient RNA competitor with GST-RBPI, it showed sig- nificant binding to any polyribonucleotides tested. In contrast, GST-CTBPI was rather specific for poly(U) and poly(1). Titra- tion of competitor RNAs indicated that the affinity of GST- RBPI RNA was one order of magnitude lower than that for DNA (data not shown). Free RBPl showed an almost identical binding specificity except that it did not bind to poly(1) (Fig. 8B). In conclusion, the binding property of RBPl is basi- cally reflected in that of CTBPl , although there is subtle differ- ence between the full-length and truncated proteins. These dif- ferences may arise from the auxiliary domains present in the N- terminal and C-terminal regions of RBPl .

DISCUSSION

Yeast ARS core-T-rich-strand-binding protein CTBPl/RBPl is a polydeoxypyrimidine-binding protein. The CTBPURBPl gene product isolated by Southwestern screening of a newly constructed GST fusion protein library binds specifically to ssDNA containing the T-rich strand of the budding yeast ARS core sequence, but not to the opposite A-rich strand or to double- stranded DNA. The binding was reduced by base substitutions within the core sequence. CTBPI bound to poly(dT), poly(dT- dC) and poly(dC) with highest affinity, but not to poly(dA) or poly(dG). Full-length RBPI also bound efficiently to poly(dT) and poly(dC), and to poly(dG) with much reduced affinity, but not to Poly(dA). Consistent with this, point mutations which increased the number of G residues in the core sequence gen- erally reduced the binding affinity, while alterations to C resi- dues did not affect the binding affinity. CTBPl and RBPI also bound efficiently to poly(U) and poly(I), respectively. However, the affinity for RNA was consistently lower than that for DNA in both CTBPl and RBP1. We have concluded that CTBPU RBPI is a polydeoxypyrimidine-binding protein of S. cerevisiae.

The minimum segment of CTBPl required for binding to nucleic acids. The cloned segment of CTBPl which directed efficient binding to ssDNA contained two sets of an RRM and a Q stretch and was identical to a portion of the previously re- ported RBPI, an RNA-binding and ssDNA-binding protein (Lee and Moss, 1993). Analyses of nucleic acid binding of deletion derivatives of GST-CTBPI revealed that each set could sustain binding with a specificity similar to that of the polypeptide con- taining two sets of RRM and Q and that the binding absolutely required the presence of an intact RRM, although a polypeptide containing two sets of RRM and a Q stretch bound to ssDNA with significantly higher affinity. Many RN NssDNA-binding proteins contain multiple sets of RRMs within one polypeptide. It is an interesting possibility that multiplicated RRMs, each of which is a rather weak binding unit, could generate a highly efficient nucleic-acid-binding domain. Requirement of synergy between multiple RRMs on one protein for general or specific nucleic acid binding, has previously been reported (Burd et al., 1991 ; Zamore et al., 1992; Caceres and Krainer, 1993). Integrity of the secondary structure of RRM, rather than specific amino acids sequences, appears to be important for binding, since d K in which the entire RNP2 sequence of RRM2 was replaced by an unrelated sequence could bind as efficiently as other deletions, whereas Ad, whose p4-forming amino acids were replaced by vector-derived sequences, lost binding, It is also possible that

specific sequences in the C-terminal region of RRMl are critical for binding. The importance of the C-terminal tail contiguous from the minimum RRM in specific RNA binding was recently reported for human RNP C proteins and U1A protein (Gorlach et al., 1994; Oubridge et al., 1994). It is not clear at present why the isolated RRM2 binds to ssDNA much less efficiently than RRM1. Comparison of the RRMl and RRM2 of CTBPI with RRM consensus sequences shows that the third residue in RNPI, which is an aromatic residue (Phe or Tyr) in most RRMs and provides a part of the platform for base stacking with bound nucleic acids, is cysteine in RRM2. Furthermore, the loop be- tween p2 and p3 p sheets of RRM2 immediately preceding the RNP-1 is only two residues long, in contrast to a length of 5- 8 residues for the loop found in most RRMS. This loop region was shown to be critical in determination of binding specificity (Scherly et al., 1990). This structural deviation of RRM2 from the consensus RRM may be responsible for the reduced binding activity. Different nucleic-acid-binding properties associated with different isolated RRM sequences from a single protein have been previously reported for Xenopus poly(A)-binding pro- tein, which contains four RRMs in tandem, the first of which did not bind to nucleic acids (Nietfeld et al., 1990).

Role of Q stretches of CTBPl in nucleic acid binding. Many RRM-type proteins are characterized by the presence of distinc- tive auxiliary domains associated with RRM. hnRNP A1 and A2, PABP (Adam et al., 1986; Sachs et al., 1986), splicing factor ASF/SF2 (Krainer et al., 1991) and hnRNP C (Swanson and Dreyfuss, 1988a) contain a glycine-rich domain, a proline- rich sequence, SR (serine and arginine)-rich sequences and an acidic-domain, respectively, in addition to one or more RRM. Each of these RRM proteins binds to nucleic acids with a unique specificity. For example, PABP binds preferentially to poly(A), whereas hnRNP C binds to poly(U) (Swanson and Dreyfuss, 1988 a) or polypyrimidine sequences (Swanson and Dreyfuss, 1988 b). However, functions of these auxiliary sequences have been characterized in only a few cases. SR sequences of ASF/ SF2 were shown to be critical for splicing reactions (Zahler et al., 1992). The glycine-rich domain of hnRNP Al , termed the RGG domain, is responsible for cooperative binding to nucleic acids (Kumar et al., 1990). Our result that an isolated RRM of CTBPI in the absence of a Q stretch was sufficient for nucleic acid binding with specificity similar to the wild-type is consis- tent with previous reports on the activities and specificities of nucleic-acid-binding resides within an isolated RRM. However, the presence of a Q stretch C-terminal to the low-affinity RRM2 increased the affinity of RRM2 to the level comparable to that of the N-terminal RRMUQ-stretch (compare AE and RRM2 in Fig. 5B and C). In contrast, the isolated RRMl in the absence of a stretch showed even higher affinity to T-rich ARS than the RRMUQ-stretch combination (compare Ed and RRMl in Fig. 5B and C). This suggests that a Q stretch could enhance or decrease the binding activity in combination with a distinct RRM. We have also concluded that the binding specificity of RBPI is largely determined the structure of the RRM, although we observed minor differences of binding specificity between isolated RRM and the full-length RBPI. RBPl bound to po- ly(dG) to some extent, to which CTBPl or isolated RRM did not show significant binding. RBPI showed rather non-specific binding to RNA, while CTBPI was specific for poly(U) and poly(1). These differences may be caused by domains outside the RRM of RBPI. There are clusters of methionine and aspara- gine near the C-terminus and two potential long a helices near the N-terminus of RBPI (Lee and Moss, 1993). These domains may affect the binding specificity of the central RRM by directly

46 Ikeda et al. (EM J. Biochem. 238)

interacting nucleic acids or by influencing with the higher-order structure of the RRM.

Multiple ARS core-T-rich-strand binding proteins in S. eere- visiue. Although CTBPURBPI bound specifically to the T-rich strand of the ARS core sequence, correlation between binding of CTBPURBPI and the origin activity of these core mutants was not obvious. All the base substitutions employed in mut-l- 9 are lethal to the origin activity (van Houten and Newlon, 1990). Nevertheless, only mut-3 and mut-9 lost their affinities for CTBPl to a significant extent and other mutants were bound by the protein nearly as efficiently as the wild-type probe. Furthermore, the null mutant of CTBPI/RBPl is viable (our un- published result, Lee and Moss, 1993). These findings render the possibility of this protein playing an essential role in replication unlikely, although auxiliary roles of CTBPliRBPl in the process of DNA replication are still possible; it could stabilize the un- wound duplex DNA at ARS by binding to the T-rich strand of the core and its related sequences, which are found in multiple copies in ARS (Plazkill and Newlon, 1988). The CTBPI/RBPl null cells grow faster than the wild-type cells (Lee and Moss, 1993 ; our unpublished result), suggesting a possible negative regulatory role of CTBPl in DNA replication. Several other T- rich-core-sequence-binding proteins have previously been re- ported (Kuno et al., 1990; Hofniann and Gasser, 1991 ; Schmidt et al., 1991; Zeidler et al., 1993; Yamazoe et al., 1994; Cockell et al., 1994). None of these proteins were shown to be essential for growth of yeast cells, and binding specifities were not neces- sarily correlated with requirements for origin function. The pres- ence of multiple T-rich-strand-binding proteins in yeast extract, which was also shown by Southwestern analysis of an yeast extract (Fig. l), may suggest redundant functions of these pro- teins in initiation of DNA replication.

Possible recognition of unusual DNA structures containing alternating TC sequences by CTBPl/RBPl. CTBPl exhibited strong affinity to alternating poly(dT-dC). The alternating TC sequences [d(TC),, tract] or asymmetric distribution of purine and pyrimidine on the two DNA strands (CT element) can gen- erate non-B-DNA conformations including H DNA or triple- stranded DNA (Wells et al., 1988). These sequences can also adopt an unwound, S1-nuclease-sensitive conformation, indica- tive of the presence of ssDNA region (Dignam et al., 1983; Hoffman-Liebermann et al., 1986). d(TC),& tracts constitute as much as 0.3-0.5% of the total mouse genome, and have been implicated in transcriptional regulation (Manor et al., 1988), re- combination (Weinerb et al., 1990) and DNA replication (Baran et al., 1987; Caddle et al., 1990). The finding that CTBPl/RBPl binds preferentially to a polypyrimidine sequence suggests that this protein may bind specifically to these sequences, which can adopt unusual higher-order structures, to regulate these pro- cesses. A poly(dT-dC)-binding protein or polydeoxypyrimidine- binding proteins from mammals have previously been reported (Yee et al., 1991; Brunel et al., 1991; Muraiso et al., 1992). These proteins may affect the rate of generation and/or stability of triplex DNA or other DNA structures induced by polypurine- polypyrimidine sequences. CTBPl is the first yeast protein which is shown to bind specifically to the polydeoxypyrimidine sequence. It remains to be seen whether CTBPURBPl encodes a yeast homologue of any of these mammalian polydeoxypyri- midine-binding proteins described previously. Further genetic and biochemical characterization of RBPUCTBPl should pro- vide useful information on the functions of this class of se- quence-specific ssDNA-binding proteins.

We thank Dr Fan-Jen S. Lee for generous gift of GST-RBPI fusion plasmid, and Dr Koichi Kawakami for discussion, advice and help on

handling yeast cells and the gift of yeast strains. We also thank Dr Kazuo Yanagi for generous support of M. I.

REFERENCES Adam, S. A,, Nakdgawa, T., Swanson, M. S., Woodruff, T. K. & Drey-

fuss, G. (1986) mRNA polyadenylate-binding protein: gene isolation and sequencing and identification of a ribonucleoprotein consensus sequence, Mol. Cell. Biol. 6 , 2932-2943.

Bergemann, A. D. & Johnson, E. M. (1992) The HeLa Pur factor binds single-stranded DNA at a specific element conserved in gene flank- ing regions and origins of DNA replication, Mol. Cell. Biol. 12,

Birney, E., Kumar, S. & Krainer, A. (1993) Analysis of the RNA-recog- nition motif and RS and RGG domains: conservation in metazoan pre-mRNA splicing factors, Nucleic Acids Res. 21, 5803-5816.

Burd, C. & Dreyfuss, G. (1994) Conserved structures and diversity of functions of RA-binding proteins, Science 265, 615-621.

Burd, C. G., Matunis, E. L. & Dreyfuss, G. (1991) The multiple RNA- binding domains of the mRNA poly(A)-binding protein have dif- ferent RNA-binding activities, Mol. Cell. Biol. 11, 3419-3424.

Brenner, C. & Fuller, R. S. (1992) Structural and enzymatic chardcteriza- tion of a purified prohormone-processing enzyme: secreted soluble Kex2 protease, Proc. Nut1 Acad. Sci. USA 89, 922-926.

Brunel, F., Alzari, P. M., Ferara, P. & Zakian, M. M. (1991) Cloning and sequencing of PYBP, a pyrimidine-rich specific single strand DNA-binding protein, Nucleic Acids Res. 19, 5237-5245.

Caceres, J. & Krainer, A. R. (1993) Functional analysis of pre-mRNA splicing factor SF2/ASF structural domains, EMBO J. 12, 4715- 4726.

Caddle, M. S., Lussier, R. H. & Heintz, N. H. (1990) CHO amplification origin, Mol. Cell. Bid. 211, 19-33.

Carmichael, E. P., Roome, J. M. & Wahl, A. F. (1993) Binding of a sequence-specific single-stranded DNA binding factor to the Simian Virus 40 core origin inverted repeat domain is cell cycle regulated, Mol. Cell. Bid. 13, 408-420.

Challberg, M. D. & Kelly, T. J. (1989) Animal virus DNA replication, Annu. Rev. Biochem. 58, 671 -717.

Cockell, M., Frutiger, S., Hughes, G. J. & Gasser, S. M. (1994) The yeast protein encoded by PUB1 binds T-rich single-stranded DNA, Nucleic Acids Rex 22, 32-40.

Dignam, J. D., Lebovitz, R. M. & Roeder, R. G. (1983) Accurate tran- scription initiation by RNA polymerase I1 in a soluble extract from isolated mammalian nuclei, Nucleic Acids Res. 11, 1475-1489.

Donreiter, I., Erdile, L. F., Gilbert, I. U., von Winkler, D., Kelly, T. J. & Fanning, E. (1992) Interaction between DNA polymerase a primase and cellular replication protein A and SV40 T antigen, EMBO J. 11,

Dunn, B. & Wobbe, C. R. (1993) Preparation of protein extracts from yeast, in Current Protocols in molecular biology (Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G. & Struhl, K., eds) unit 13.13.

1257-126s.

769-776.

Dutta, A. (1993) Trans-plication factors? Current Bid. 3, 709-712. Frangioni, J. V. & Neel, B. G. (1993) Solubilization and purification of

enzymatically active glutathione S-transferase (pGEX) fusion pro- teins, Anal. Biochem. 210, 179-187.

Gorlach, M., Burd, C. G. & Dreyfuss, G. (1994) The determinant of RNA-binding specificity of heterogeneous nuclear ribonucleoprotein C proteins, J. Biol. Chem. 269, 23074-23078.

He, Z., Brinton, B. T., Greenblatt, J., Hassell, J. A. & Ingles, C. J. (1993) The transactivator protein VP16 and Gal4 bind replication factor A, Cell 7-3, 1223-1232.

Hofmann, J. F.-X. & Gasser, S. M. (1991) Identification and purification of a protein that binds the yeast ARS consensus sequence, Cell 64, 951-960.

Hoffman-Liebermann, B., Liebermann, D., Troutt, A,, Kedes, L. H. & Cohen, S. N. (1986) Human homologs of TU transposon sequences: polypurine/polypyrimidine sequence elements that can alter DNA conformation in vitro and in vivo, Mol. Cell. Biol. 6 , 3632-3642.

Jong, A. Y.-S., Michael, W. C., Gilbert, M., Oehm, A. & Campbell, J. L. (1987) Saccharomyces cerevisiae SSBl protein and its relation- ship to nucleolar RNA-binding protein, Mol. Cell. Biol. 7, 2947- 2955.

Ikeda et al. (Eur: J. Biochem. 238) 47

Kelly, R. C., Jensen, D. E. & von Hippel, P. H. (1976) DNA ‘melting’ proteins. IV. Fluorescence measurements of binding parameters for bacteriophage T4 gene 32-protein to mono-, oligo-, and polynucleo- tides, J. Biol. Chem. 251, 7240-7250.

Kenan, D. J., Query, C. C. & Keene, J. D. (1991) RNA recognition: towards identifying determinants of specificity, Trends Biochem Sci.

Kim, C., Snyder, R. 0. & Wold, M. S. (1992) Binding properties of replication protein A from human and yeast cells, Mol. Cell. Bid.

16, 214-220.

12, 3050-3059. Kornberg, A. & Baker, T. (1992) DNA replication, 2nd edn. Krainer, A. R., Mayeda, A,, Kozak, D. & Binns, G. (1991) Functional

expression of cloned human splicing factor SF2 : homology to RNA- binding proteins, U1 70K, and Drosophila splicing regulators, Cell

Kumar, A,, Casa-Finet, J. R., Luneau, C. J., Karpel, R. L., Merrill, B. M., Williams, K. R. & Wilson, S. H. (1990) Mammalian hetero- geneous nuclear ribonucleoprotein A1 : nucleic acid binding proper- ties of the COOH-terminal domain, J. Biol. Chem. 265, 17094- 17 100.

Kuno, K., Murakami, S . & Kuno, S. (1990) Single-stranded-binding factor(s) which interact with ARSl of Saccharomyces cerevisiae, Gene (Amst.) 5, 73-77.

Lee, F.-J. S. & Moss, J. (1993) An RNA binding protein gene (RBPI) of Saccharomyces cerevisiae encodes a putative glucose-responsive protein containing two RNA recognition motifs. J. Bid. Chem. 268,

Li, R. & Botchan, M. R. (1993) The acidic transcriptional activation domains of VP16 and p53 bind the cellular replication protein A and stimulate in vitro BPV-1 DNA replication, Cell 73, 1223-1232.

Lohman, T. M. & Ferrari, M. E. (1994) Escherichia coli single-straned DNA binding protein : multiple DNA-binding modes and cooperativ- ities, Annu. Rev. Biochern. 63, 527-570.

Masai, H. & Arai, K. (1993) Replication and transcription on single- stranded DNA templates: proteins that specifically recognize single- stranded DNA sequences with a higher order structure, Mol. Biol. (Life Sci. Adv.) 13, 19-37.

Manor, H., Rao, B. S. & Martin, R. G. (1988) Abundance and degree of dispersion of genomic d(GA)/(TC)n sequences, J. Mol. Evol. 27,

Meissner, P. S . , Sisk, W. P. & Berman, M. L. (1987) Bacteriophage lambda cloning system for the construction of directional cDNA li- braries, Proc. Natl Acad. Sci. USA 84, 4171 -4 175.

Mizusawa, S. & Gottesman, S . (1983) Protein degradation in Escherichia coli: The Ion gene controls the stability of sulA protein, Proc. Natl Acad. Sci. USA 80, 358-362.

Nagai, K., Oubridge, C., Ito, N., Avis, J. & Evans, P. (1995) The RNP domain: a sequence-specific RNA binding domain involved in pro- cessing and transport of RNA, Trend Biochem. Sci. 20, 235-240.

Newlon, C. S. (1988) Yeast chromosome replication and segregation, Microbiol. Rev. 52, 568-601.

Nietfeld, W., Mentzel, H. & Pieler, T. (1990) The Xenopus laevis po- ly(A) binding protein is composed of multiple functionally indepen- dent RNA binding domains, EMBO J. 9, 3699--3705.

Nomoto, S . , Nakayama, N., Arai, K. & Matsumoto, K. (1990) Regula- tion of the yeast pheromone response pathway by G protein subunits,

Oubridge, C., Ito, N.,Evans, P., Teo, C.-H. & Nagai, K. (1994) Crystal structure at 1.92 A resolution of the RNA-binding domain of the U1A spliceosomal protein complexed with an RNA hairpin, Nature

Plazkill, T. G. & Newlon, C. S. (1988) A yeast replication origin consists of multiple copies of a small conserved sequence, Cell 53,441-450.

Sachs, A. B., Bond, M. W. & Kornberg, R. D. (1986) A single gene from yeast for both nuclear and cytoplasmic polyadenylate-binding proteins: domain structure and expression, Cell 45, 827-835.

66, 383-394.

15080-15087.

96-101.

EMBO J. 9, 691-696.

372,432-438.

Scherly, D., Boelens, W., Dathan, N. A,, van Venrooij, W. J. & Mattaj, I. W. (1990) Major determinants of the specificity of interaction be- tween small nuclear ribonucleoproteins U1A and U2B and their cog- nate RNAs, Nature 345, 502-506.

Schmidt, A. M. A., Herterich, S . U. & Krauss, G. (1991) A single- stranded DNA binding protein from S . cerevisiae specifically recog- nizes the T-rich strand of the core sequence of ARS elements and discriminates against mutant sequences, EMBO J. 10, 981 -985.

Shirahige, K., Iwasaki, T., Rashid, M. B., Ogasawara, N. & Yoshikawa, H. (1993) Location and characterization of autonomously replicating sequences from chromosome VI of Saccharomyces cerevisiae, Mol. Cell. Biol. 13, 5043-5056.

Smith, D. B. & Johnson, K. S. (1988) Single-step purification of poly- peptides expressed in Escherichia coli as fusions with glutathione S- transferase, Gene (Amst.) 6, 731 -740.

Stillman, B. (1989) Initiation of eukaryotic DNA replication in vitro, Annu. Rev. Cell Biol. 5, 197-245.

Sun, W. & Godson, G. N. (1993) Binding and phasing of Escherichia coli single-stranded DNA-binding protein by the secondary structure of phage G4 origin of complementary DNA strand synthesis (G4oric), J . Biol. Chem. 268, 8026-8039.

Swanson, M. S. & Dreyfuss, G. (1988a) Classification and purification of proteins of heterogeneous nuclear ribonucleoprotein particles by RNA-binding specificities, Mol. Cell. Biol. 8, 2237-2241.

Swanson, M. S . & Dreyfuss, G. (1988b) RNA binding specificity of hnRNP proteins: a subset bind to the 3’ end of introns, EMBO J. 7,

Tzfati, Y., Abeliovich, H., Kapeller, I. & Shiomai, J. (1992) A single- stranded DNA-binding protein from Crithidia fasciculata recognizes the nucleotide sequence at the origin of replication of kinetoplast DNA minicircles, Proc. Natl Acad. Sci. USA 89, 6891-6895.

Umek, R. M. & Kowalski, D. (1988) The ease of DNA unwinding as a determinant of initiation at yeast replication origins, Cell 52, 559- 567.

Umek, R. M., Liskens, M. H. K., Kowalski, D. & Huberman, J. A. (1989) New beginnings in studies of eukaryotic DNA replication origins, Biochim. Biophys. Acta 1007, 1-14.

van Houten, J. V. & Newton, C. S . (1990) Mutational analysis of the consensus sequence of a replication origin from yeast chromosome 111, Mol. Cell. Biol. 10, 3917-3925.

Weinerb, A,, Collier, D. A,, Birshtein, B. K. & Wells, R. D. (1990) Left- handed Z-DNA and intramolecular triplex formation at the site of an unequal sister chromatid exchange, J. Biol. Chem. 265, 1352-1359.

Wells, R. D., Collier, D. A,, Hanvey, J. C., Shimizu, M. & Wohlrab, F. (1988) The chemistry and biology of unusual DNA structures adopted by oligopurine/oligopyrimidine sequences, FASEB J. 2,

Yamazoe, M., Shirahige, K., Rashid, M. B., Kaneko, Y., Nakayama, T., Ogasawara, N. & Yoshikawa, H. (1994) A protein which binds pref- erentially to single-stranded core sequence of autonomously replicat- ing sequence is essential for respiratory function in mitochondria of Saccharomyces cerevisiae, J . Biol. Chem. 269, 15 244- 15 252.

Yokota, T., Arai, K. & Kaziro, Y. (1979) Studies on 30s ribosomal pro- tein S1 from E. coli. I. Purification and physiochemical properties, J. Biochem (Tokyo) 86, 1725-1737.

Zahler, A. M., Lane, W. S . , Stolk, J. A. & Roth, M. B. (1992) SR pro- teins: a conserved family of pre-mRNA splicing factors, Genes & Dev. 6, 837-847.

Zamore, P. D., Patton, J. G. & Green, M. R. (1992) Cloning and domain structure of the mammalian splicing factor U2AF, Nature 355, 609- 614.

Zeidler, R., Hobert, O., Johannes, L., Faulhammer. H. & Kraus, G. (1993) Characterization of two novel single-stranded DNA-specific autonomously replicating sequence-binding protein from Saccharo- myces cerevisiae, one of which is adenylosuccinate synthetase, J. Biol. Chem. 268, 20 191 -20 197.

3519-3529.

2939 -2949.