Cloning, Overexpression, and Genomic Mapping of the 14-kDa Subunit of Human Replication Protein A*

Replication protein A (RPA) is a three-subunit pro- tein that plays a central role in eukaryotic DNA replication, recombination, and repair. We have previ- ously reported the cloning and bacterial expression of the 70- and 32-kDa subunits of human RPA (hRPA). We have now cloned the 14-kDa subunit (hRPA3) from a HeLa cell cDNA library. The hRPA3 cDNA is a 692-base pair sequence that contains an open reading frame encoding a protein of 121 amino acids with a calculated molecular mass of 13.6 kDa. The deduced amino acid sequence shows only limited similarity to the small subunit of yeast RPA and is unrelated to any other protein in the cur- rent data banks. homology the Smith-Waterman and Needleman-Wunsch analysis amino acid computer functional protein domains 1992) using Macpattern 1991) on Macintosh IIci personal computer. Construction of E. coli Expression Plasmids-The bacteriophage T7-based expression system (Studier et al., 1990) was used to express the 14-kDa subunit of human RPA in E. coli. PCR primers homologous to the NH, and COOH termini of the open reading frame of the cloned cDNA were synthesized. Both contained an additional 8 bases defining the restriction sites matching target expression vector pET3d sites NcoI and BamHI, which allowed direct insertion of the restricted PCR product into pET3d with retention of the translation initiation signals of T7 gene 10. The resulting plasmid (pCU165) transcribes from the T7 410 promoter, and the first ATG in the mRNA corre- sponds to the human initiator codon, leading to full-length 14-kDa protein without additional amino acids. A construct expressing a protein with an NH2-terminal histidine tag and Factor X cleavage site (pCU232) was obtained by using an alternative 5'-PCR primer encoding amino acids MHHHHHHIEGR- to precede the first native methionine.

Replication protein A (RPA) is a three-subunit protein that plays a central role in eukaryotic DNA replication, recombination, and repair. We have previously reported the cloning and bacterial expression of the 70-and 32-kDa subunits of human RPA (hRPA). We have now cloned the 14-kDa subunit (hRPA3) from a HeLa cell cDNA library.
The hRPA3 cDNA is a 692-base pair sequence that contains an open reading frame encoding a protein of 121 amino acids with a calculated molecular mass of 13.6 kDa. The deduced amino acid sequence shows only limited similarity to the small subunit of yeast RPA and is unrelated to any other protein in the current data banks.
A recombinant protein containing a short histidine tag at the NHa terminus has been purified in good yield from Escherichia coli by metal-chelate affinity chromatography. Antibodies prepared against recombinant hRPA3 recognize the native protein and inhibit SV40 DNA replication in vitro.
We have localized the genes for the 70-, 32-, and 14-kDa subunits to chromosomes 17, 1, and 7, respectively, using polymerase chain reaction amplification of genomic DNA from rodent-human hybrid cell lines. Since RPA appears to be involved in several fundamental cellular processes, the physical mapping of the RPA genes may be useful in identifying possible human genetic defects associated with RPA deficiency or dysfunction.
To study the molecular components and mechanisms involved in mammalian DNA replication, our laboratory developed a model system based on the in vitro replication of DNA templates carrying the papovavirus simian virus 40 (SV40) origin of replication. DNA replication in this model system has properties very similar to those of cellular chromosomal replication (for reviews, see Challberg and , Hurwitz et al. (1990), Kelly (1988), and Stillman (1989)). With the exception of the viral origin-binding protein, the large T antigen, all proteins required for the in vitro replication reaction are derived from HeLa cell extracts and are presumably involved in mammalian DNA replication as well (Li and Grants GM42780 (to T. J. K.) and HG00373 (to E. W. J.). The costs * This work was supported in part by National Institutes of Health of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.  Kelly, 1984Kelly, , 1985. Seven essential cellular replication proteins have been identified and subsequently purified by a number of laboratories (Murakami et al., 1986;Prelich et al., 1987;Syvaoja et al., 1990;Tsurimoto and Stillman, 1989b;Virshup and Kelly, 1989;Weinberg et al., 1990;Weinberg and Kelly, 1989;Wold and Kelly, 1988;Wold et al., 1989;Yang et al., 1987). At the same time, studies in several laboratories have led to insights into the mechanism of SV40 DNA replication (Matsumoto et al., 1990;Tsurimoto et al., 1990;Weinberg et al., 1990). Human replication protein A (hRPA)' is a three-subunit protein that was originally purified as an essential component of the SV40 in vitro replication system (Fairman and Stillman, 1988;Wobbe et al., 1987;Wold and Kelly, 1988). It is a partially sequence-dependent single-stranded DNA-binding protein, with an affinity for pyrimidines 50-fold higher than for purines and preferential binding to the T-rich strand of the replication origin (Kim et al., 1992).

The nucleotide sequence(s) reported in thispaper has
hRPA has been shown to be involved at several critical steps in the replication reaction. It is required for the unwinding of origin-containing duplex DNA by T antigen during the initiation of replication Wold et al., 1987). A stoichiometric DNA unwinding activity can be detected at low ionic strength (Georgaki et al., 1992), suggesting that RPA may prevent reannealing of the melted DNA strands during the T antigen-induced unwinding at the replication origin. Specific protein-protein interactions do not appear to play a role at this stage since a number of heterologous single-stranded DNA-binding proteins can substitute for hRPA (Kenny et al., 1989;Virshup et al., 1989). The subsequent formation of an initiation complex containing T antigen, hRPA, and DNA polymerase a/primase, however, probably involves specific protein-protein interactions since heterologous single-stranded DNA-binding proteins fail to substitute for hRPA (Collins and Kelly, 1991;Erdile et al., 1991aErdile et al., , 1991b. Recent studies with purified proteins have demonstrated specific binding of hRPA to T antigen and polymerase a/primase (Dornreiter et al., 1992). An additional role in suppressing nonspecific priming events was suggested by studies on model templates (Collins and Kelly, 1991;Erdile et al., 1991b). These data suggest that hRPA plays an important role in establishing a replication-competent complex at the replication fork. Following initiation, hRPA also participates in elongation, as indicated by the observed stimulation of both polymerases a and 6 by hRPA (Eki et al., 1992;Kenny et al., 1989;Tsurimoto and Stillman, 1989a).
Although all three hRPA genes have recently been shown to be essential in Saccharomyces cerevisiae Heyer et al., 1990), their specific biochemical functions are less well understood. The single-stranded DNA-' The abbreviations used are: hRPA, human replication protein A; PCR, polymerase chain reaction; bp, base pair(s); PAGE, polyacrylamide gel electrophoresis.

6131
binding activity has been localized to the 70-kDa subunit (hRPA1) (Erdile et al., 1991a;Wold et al., 1989). Bacterially expressed hRPAl can substitute for the holoenzyme in the origin unwinding reaction and is capable of interacting with polymerase a, but not T antigen (Dornreiter et al., 1992;Erdile et al., 1991aErdile et al., , 1991b. Antibodies to hRPAl and hRPA2 have been shown to inhibit the replication reaction as well as interactions with DNA polymerases a and 6 (Erdile et al., 1990;. Little is known about the functions of the two smaller subunits, although a role in replication is likely since the bacterially expressed 70-kDa subunit cannot substitute for hRPA (Erdile et al., 1991a). The cell cycle-dependent phosphorylation of RPA2 makes it a good candidate for a regulatory process modulating RPA function in DNA replication (Din et al., 1990;Dutta et al., 1991;Dutta and Stillman, 1992;Fotedar and Roberts, 1992). RPA2 appears to be phosphorylated during the S phase of the cell cycle as well as during in vitro replication reactions (Dutta and Stillman, 1992;Fotedar and Roberts, 1992). The precise role of phosphorylation, however, and the specific kinase(s) involved still remain to be defined. Phosphorylation of hRPA2 does not appear to be required for the association of hRPA with DNA and may be a consequence of (rather than a prerequisite for) the initiation of replication (Fotedar and Roberts, 1992).
In addition to DNA replication, RPA has been shown to play important roles in human DNA repair and recombination. Nucleotide excision repair was shown to depend on hRPA in a cell-free system, and DNA polymerase E, which may be the principal repair DNA polymerase (Nishida et al., 1988), is stimulated by hRPA (Coverley et al., 1991(Coverley et al., , 1992. hRPA has also been shown to stimulate human homologous pairing protein 1, which catalyzes homologous pairing and strand exchange, in a species-specific fashion . A deeper understanding of DNA replication will depend on a more detailed biochemical characterization of the proteins involved. To this end, we and others have started to clone the genes of the mammalian replication proteins and to express them in a variety of host cells. So far, this has been achieved for human topoisomerase I (D'Arpa et al., 1988), the catalytic subunit of human polymerase a (Wong et al., 1988), the p49 subunit of the murine primase (Prussak et al., 1989), human proliferating cell nuclear antigen (Travali et al., 1989), the catalytic subunit of human polymerase 6 (Chung et al., 1991), and the 40-kDa subunit of human replication factor C (activator 1) (Chen et al., 1992). We have previously reported the primary structure and bacterial expression of the 70-kDa (hRPA1) (Erdile et al., 1991a) and 32-kDa (hRPA2) (Erdile et al., 1990) subunits of hRPA (replication factor A, human single-stranded DNA-binding protein).
We report here the cloning and bacterial expression of the 14-kDa subunit (hRPA3) from a HeLa cDNA library. The cloned cDNA sequence shows only limited similarity to the small subunit of yeast RPA and is unrelated to any other protein in the current data banks. In vitro DNA replication can be specifically inhibited by antibodies raised against hRPA3. In addition, the three hRPA genes have been localized to chromosomes 17 (hRPAl), 1 (hRPA2), and 7 (hRPA3).

MATERIALS AND METHODS
Peptide Sequencing and Oligonucleotide Synthesis-The primary sequences of three tryptic peptides of hRPA3 were determined as described previously (Erdile et al., 1990) and were used to design oligonucleotide primers appropriate for polymerase chain reactions (PCR) (see Table I).
Preparation of Target-enriched cDNA--11.5 pg of total cytoplasmic RNA prepared from exponentially growing HeLa cells (Sambrook et al., 1989) was incubated for 2 h with 400 units of Moloney murine leukemia virus reverse transcriptase (Bethesda Research Laboratories) and 400 ng of gene-specific DNA primer TK73c or TK74c (see Table I), followed by isolation of the cDNA by standard methods (Sambrook et al., 1989). Total cellular cDNA was obtained using the same protocol, except that 15 pg of oligo(dT) ((dT)n-le; Pharmacia LKB Biotechnology Inc.) was used as primer.
Subcloning of PCR Products-PCR products were separated by agarose gel electrophoresis. DNA bands of appropriate size were excised and purified using the glass bead-based Geneclean 11" procedure as described by the manufacturer (BiolOl Inc., La Jolla, CA) and ligated into Bluescript plasmids (Stratagene). Recombinant plasmids were transformed into competent Escherichia coli DH5a (Bethesda Research Laboratories), selected by standard blue/white color assay, and purified from overnight cultures by a boiling miniprep procedure (Ausubel et al., 1991). Medium-and large-scale plasmid preparations were performed by standard CsCl/EtBr equilibrium centrifugation (Ausubel et al., 1991) or using the Qiagen' plasmid kit (Qiagen Inc., Chatsworth, CA).
DNA Sequencing-Sequencing was performed by the dideoxy method (Sanger et al., 1977) using a Sequenase 11" kit (U. S. Biochemical Corp.) according to the manufacturer's instructions. All sequences were confirmed by sequencing both strands using commercially available or custom-made primers derived from cloned sequences. Strategy for Obtaining a cDNA Probe-Oligonucleotide primers TK73c to TK75 were designed using sequence information from the ends of the peptide hRPA3-36 (see Table I), defining a PCR product of known length. Target-enriched cDNA (see above) was used as template. The expected 72-bp DNA species was found only in the reaction using the TK75/TK74c primer pair (data not shown) and was subcloned into plasmid pBSKSII(-) (Stratagene). Sequencing revealed an internal 27-nucleotide nondegenerate sequence (TK76) flanked by sequence derived from the PCR primers.
To isolate a larger DNA fragment, a PCR reaction was performed with all combinations of the degenerate primers using total cellular cDNA as template, followed by agarose gel electrophoresis, Southern transfer, and hybridization with the 32P-labeled nondegenerate TK76 probe (Southern, 1975). Using wash conditions of 2 X SSC (0.3 M NaCl, 30 mM Nas citrate.2H20), 0.1% SDS, specific bands of 180 and 200 bp were found in the PCR reactions containing the TK75/ TK67c and TK75/TK71c primer pairs, respectively (data not shown). The 180-and 200-bp fragments were gel-purified and reamplified by repeating the PCR reaction with the same primer pairs. They were then subcloned into the EcoRV site of pBSKSII(-) for sequencing. The derived plasmids (pCU92 and pCU103) were found to contain the nucleotide sequences matching amino acids 50-103 and 50-120, respectively (see Fig. 1).
Screening of Xgtll Library-A 209-bp PstI/HindIII restriction fragment of pCU92 was gel-purified and labeled with [a-32P]dCTP by random oligonucleotide priming (Feinberg and Vogelstein, 1983) to a specific activity of 5.7 X 10' cpm/pg. Nitrocellulose filters containing -300,000 plaques from a Xgtll library containing HeLa cDNA (Stratagene; a gift of Dr. Hank Ratrie, The Johns Hopkins University Medical School) were screened as previously described (Erdile et al., 1990(Erdile et al., ,1991a. Positive plaques underwent two additional rounds of plaque purification. X phage DNA was prepared by standard methods (Sambrook et al., 1989). EcoRI restrictions of the phage DNA did not free the cDNA inserts due to alterations at the 5'cloning site. The cDNA fragments were excised from the phage DNA using an SstIIEcoRI double restriction digest, cloned into the EcoRI/ Sac1 sites of pBSKSII(-) (pCU145 and pCU155), and sequenced. rived from the cloned hRPA3 cDNA was analyzed for homologous Computerized Sequence Analysis-The amino acid sequence deproteins using the FASTA algorithm (Pearson and Lipman, 1988) available through the GenBank on-line service (c/o Intelligenetics, Inc., Mountain View, Ca). The searches were performed against the protein sequence data bases GENPEPT (release 70) and SWISSPROT (release 19) at a k-tuple setting of 2. Pairwise sequence homology analyses were performed using the Smith-Waterman and Needleman-Wunsch algorithms available in the sequence analysis software package of the Genetics Computer Group (Madison, WI). The predicted amino acid sequence was also compared to a computer data base of functional protein domains (Bairoch, 1992) using the Macpattern program (Fuchs, 1991) on a Macintosh IIci personal computer.
Construction of E. coli Expression Plasmids-The bacteriophage T7-based expression system (Studier et al., 1990) was used to express the 14-kDa subunit of human RPA in E. coli. PCR primers homologous to the NH, and COOH termini of the open reading frame of the cloned cDNA were synthesized. Both contained an additional 8 bases defining the restriction sites matching target expression vector pET3d sites NcoI and BamHI, which allowed direct insertion of the restricted PCR product into pET3d with retention of the translation initiation signals of T7 gene 10. The resulting plasmid (pCU165) transcribes from the T7 410 promoter, and the first ATG in the mRNA corresponds to the human initiator codon, leading to full-length 14-kDa protein without additional amino acids. A construct expressing a protein with an NH2-terminal histidine tag and Factor X cleavage site (pCU232) was obtained by using an alternative 5'-PCR primer encoding amino acids MHHHHHHIEGR-to precede the first native methionine.
Nz'+-Agarose Chelote Affinity Chromatography-Immobilized metal ion affinity chromatography (Hochuli et al., 1987(Hochuli et al., , 1988 was used for single-step chromatographic purification of bacterially expressed recombinant protein. The Niz+-nitrilotriacetic acid complex resin-based agarose matrix used is commercially available (Qiagen Inc.). The following buffers were used buffer B (see above), buffer C (buffer B, pH 6.3), buffer D (buffer B, pH 5.9), and buffer BI (buffer B with 250 mM imidazole C1, pH 7.0). A 2-ml Ni2+-nitrilotriacetic acid-agarose column was equilibrated at 4 "C in buffer B with 1 mM phenylmethylsulfonyl fluoride. The urea-soluble fraction of the pellet of a 100-ml culture was loaded by gravity flow at -2 mg of protein/ ml of column matrix. The column was washed extensively with buffer B and eluted in a stepwise fashion with 5 column volumes of buffers C, D, and BI.
Protein Assays-Protein fractions were quantified using the Bio-Rad protein assay (Bradford, 1976) with bovine serum albumin as standard.
Zmmunoblotting-All Western blotting procedures were performed at room temperature using standard procedures (Erdile et al., 1990(Erdile et al., , 1991a. The nitrocellulose filters were incubated in a 1:10,000-fold dilution of antiserum in Western blot buffer (1 X phosphate-buffered saline, 0.3% Triton X-100) and developed with the ECL (enhanced chemiluminescence) Western blotting analysis system according to the manufacturer's protocol (Amersham Corp.).
Replication Assay-Crude cytoplasmic HeLa cell extracts (Li and Kelly, 1984) were preincubated with protein A-Sepharose bead-purified anti-RPA3 serum, preimmune serum, or antibody/dialysate on ice for 30 min. Replication buffer, SV40 origin-containing DNA, ["PI dCTP, and T antigen were then added and incubated at 37 "C for 2 h (Wold et al., 1989). The samples were then analyzed for incorpo-ration of radioactivity by trichloroacetic acid precipitation as well as agarose gel electrophoresis and autoradiography as previously described (Wold et al., 1989).
Chromosomal Localization-Two independent complete rodent-human somatic cell hybrid panels were screened by PCR with genespecific oligonucleotide primers to determine the chromosomal localization of the three hRPA genes. One panel was National Institute of General Medical Sciences (NIGMS) rodent somatic cell hybrid panel 1, consisting of 17 mouse-human hybrids and one Chinese hamster ovary-human hybrid (Coriell Institute for Medical Research, Camden, NJ). The second panel was the BIOS somatic cell hybrid panel, consisting of 25 Chinese hamster ovary-human hybrids (BIOS Corp., New Haven, CT). Each hybrid cell line contained one or several human chromosomes as determined by standard cytogenetic analysis. PCR amplifications of the three target hRPA genes were performed under standard conditions (see above) in the presence of 50 ng of genomic DNA. The oligonucleotide primer pairs used (see Table I) were derived from the 3"untranslated regions of the three cloned cDNAs (Erdile et al., 1990(Erdile et al., ,1991a to minimize the chance of obtaining products of identical length from human and rodent genes. The presence of human-specific PCR products was scored for each hybrid cell line and compared with the cytogenetic information provided by the manufacturer. Hybrid cell lines containing an individual human chromosome in at least 10% of the cells examined were considered positive for that chromosome. The result of the comparison of the PCR and cytogenetic data was expressed as a percent concordance, which was calculated for each chromosome and gene by the following formula: % concordance = ((cases with PCR product and human chromosome + cases without PCR product and no human chromosome) divided by number of hybrid cell lines) X 100.

RESULTS
Cloning of cDNA for 14-kDa Subunit of hRPA-Tryptic peptides of hRPA3 (Table I) were purified by reverse-phase chromatography and sequenced as described previously (Erdile et al., 1990). The peptide sequences were then used to synthesize degenerate oligonucleotides ( Table I). A PCR using oligonucleotides TK75 and TK74c with target-enriched cDNA as template (see "Materials and Methods") produced the expected 72-bp fragment. A segment of 27 bp of internal sequence not derived from the PCR oligonucleotides was then used as nondegenerate probe (TK76) to detect hRPA3-specific PCR products. Various combinations of PCR primers TK67c to TK75 (Table I) were used to amplify HeLa cell first-strand cDNA. Southern blot analysis revealed PCR products of 180 and 200 bp that hybridized with the TK76 probe. These were sequenced and used to screen -300,000 bacteriophage plaques of a HeLa cDNA library. Three independent phage isolates (Ph5, Ph42, and Ph43) were found to contain the hRPA3 cDNA.
The cloned human RPA3 cDNA is a 692-bp sequence that contains an open reading frame encoding a protein of 121 amino acids, beginning with the first ATG codon at codon 31 and terminating at a TGA codon at nucleotide 394 (Fig. 1). The calculated molecular mass of the predicted protein is 13,559 Da, which is consistent with the 14-kDa band observed on SDS-PAGE of purified hRPA. The nucleotides in the vicinity of the first ATG provide a favorable context for initiation of translation, with an A at position -3 and a G at position +4 relative to the putative first ATG triplet (Kozak, 1991). The adjacent sequence also matches the vertebrate translation initiation site (A/GNCATG) determined by the application of the 50/75 consensus rule (Cavener and Ray, 1991). The 3"untranslated sequence contains the eukaryotic AATAAA polyadenylation signal.
The predicted amino acid sequence of hRPA3 includes all three sequences derived from tryptic peptide fragments. The yeast 14-kDa subunit (RPA3) has recently been published  and contains 122 amino acids. The best-fit alignment with yeast RPA3 obtained using the Needleman-Wunsch algorithm is shown in Fig. 2. Allowing five gaps, the alignment shows 25% sequence identity and TABLE I Oligonucleotide primers used during library screening, expression vector construction, and genomic mapping of hRPA3 All sequences are shown using the IUB nomenclature for nucleic acids and amino acids. The single-letter abbreviations for ambiguous nucleotides are as follows: N = G, A, T, or C; Y = T or C; R = A or G H = A, C, or T; and D = A, G, or T. Peptide sequences hRPA3-19, hRPA3-31, hRPA3-36 were obtained from direct sequencing of products of a tryptic digest of gel-purified hRPA3. A question mark denotes a sequenator cycle with an ambiguous signal. Underlined sequences correspond to the synthetic oligonucleotides. Lower-case bases indicate additional nucleotides used for convenient restriction sites.  frame is shown, with the predicted 121 amino acid sequence displayed below the corresponding codon triplets using the standard three-letter amino acid code. The three underlined peptide sequences were read by direct peptide sequencing of a tryptic protein digest of the purified 14-kDa subunit. 46% sequence similarity. No homologous amino acid sequences were found in the current releases of the GenBank/ GENPEPT and SWISSPROT data bases. No amino acid patterns matching functional protein domains were found when the predicted amino acid sequence was compared to release 8.0 of the Prosite data base (Bairoch, 1992).

1/1 31/1 61/11 TTC CCC GAG CCG CAG TCT TGG ACC ATA ATC ATG GTG GAC ATG ATG GAC TTG CCC AGG TCG CGC ATC AAC GCC
Bacterial Ouerexpresswn of 14-kDa Subunit of hRPA-The cloned hRPA3-coding sequence was amplified by PCR and inserted into the pET3d bacterial expression vector. The oligonucleotide primers used included appropriate restriction sites to allow precise positioning of the coding sequence relative to the bacterial ribosomal attachment site. The resulting translation product does not contain any extra amino acids. A second expression construct contained a sequence of 6 histidines, followed by the Factor X recognition sequence (IEGR) immediately NH2-terminal to the first native methionine. Transfection of both expression plasmids as well as the pET3d vector control into E. coli strain BL21(DE3) produced ampicillin-resistant colonies, suggesting that the 14-kDa gene product was less toxic to E. coli than the 70-kDa subunit (Erdile et al., 1991a). Transformants were induced with isopropyl-1-thio-8-D-galactopyranoside, lysed, and fractionated as described under "Materials and Methods." Fig. 3A is a Coomassie Blue-stained SDS-polyacrylamide gel showing the protein patterns of lysates of E. coli host BL21(DE3) with expression plasmid pET3d and with recombinant plasmid coli that co-migrates with native 14-kDa subunit and cross-reacts with antiserum raised against COOHterminal synthetic peptide. E. coli strain BL21(DE3) and transformants containing the vector control (pET3d) or the cloned cDNA (pCU165) were prepared as described under "Materials and Methods." Induction by isopropyl-lthio-0-D-galactopyranoside is indicated (+ above the lane). Lanes S, supernatant from a 10-min centrifugation at 15,000 X g of total bacterial lysate; lunes P , corresponding pellet resuspended in 0.02 volume of lysis buffer. The last two lanes are samples of pellets resuspended in 0.02 volume of 8 M urea for 2 h, followed by centrifugation as described above. Approximately 1 mg of protein was loaded in each lane, except for the control lane with native hRPA and the isopropyl-lthio-8-D-galactopyranoside (ZPTG)-induced pellet fraction of pCU165 (0.5 mg each), and submitted to SDS-PAGE. A shows the protein gel stained by Coomassie Blue. A duplicate protein gel was transferred to nitrocellulose for Western blot analysis with antiserum raised against a COOH-terminal s-ynthetic pep- pCU165, which carries the unmodified hRPA3-coding sequence. The corresponding Western blot from a parallel SDSpolyacrylamide gel performed with an antiserum raised against a synthetic COOH-terminal peptide of hRPA3 is shown in Fig. 3B. The protein gel shows a prominent 14-kDa band co-migrating with the native hRPA3, visible only in lysates derived from cells carrying pCU165. The antiserum raised against the COOH-terminal peptide reacts with the 14-kDa bands of both native hRPA and the bacterially expressed protein. A marked increase in intensity is seen after isopropyl-1-thio-,!?-D-galactopyranoside induction. The majority of the overexpressed protein is insoluble in the standard lysate buffer, but appears soluble in the presence of 8 M urea. The urea-soluble fraction of a lysate derived from cells carrying pCU232, which codes for a recombinant protein with an NHz-terminal His6 tag, was used for Ni2+-agarose column chromatography. Approximately 50% of the loaded protein was recovered in the fraction eluting at pH 5.9 at a concentration of 1 mg/ml. A 100-ml bacterial preparation yielded 6 mg of purified hRPA3. As shown in Fig. 4, a high degree of purification can be achieved with good yield by this method. Attempts to renature the purified recombinant protein have not been successful so far, with precipitation occurring when the concentration of urea is lowered below 1 M. Since the subunits of native hRPA seem to form a tight complex, it is possible that an exposed hydrophobic domain of the overexpressed protein is impairing its solubility.
Antiserum Raised against Recombinant hRPA3 Inhibits in Vitro DNA Replication-As a first step toward defining the role of the 14-kDa subunit of hRPA, we examined whether in vitro replication could be inhibited by antisera raised specifically against hRPA3. All sera used in inhibition experiments were purified by affinity chromatography using protein A-Sepharose beads. Crude cytoplasmic HeLa extracts were preincubated for 30 min on ice with up to 60 pg of preimmune or immune sera. No inhibition was seen using antisera raised against NHz-and COOH-terminal peptides (data not shown). An antiserum was then raised against Ni'+-agarose-purified bacterially expressed hRPA3, which recognized both bacterial and native hRPA3. As shown in Fig. 5, in vitro DNA replication is inhibited by an order of magnitude by preincubation with this antiserum. The inhibition can be reversed by adding additional purified hRPA. A 4-point titration shows a dose dependence of the inhibition, whereas the preimmune serum had no effect.
Three Genes Encoding hRPA1, hRPA2, and hRPA3 Are Located on Chromosomes 17, 1, and 7, Respectively-hRPA consists of three different subunits, so it was of interest to determine whether they had similar genetic locations or were dispersed over the human genome. Since genomic DNA of rodent-human hybrid cell lines is readily available, we chose to determine the chromosomal location of the three genes by PCR amplification using gene-specific oligonucleotide primers. As shown in Fig. 6A, the resulting predicted fragment lengths of 444 bp (hRPAl), 494 bp (hRPA2), and 254 bp (hRPA3) (lunes +C) were diagnostic of the human genes. No amplification products of identical size were seen in reactions using mouse or Chinese hamster ovary cell genomic DNA templates (lunes -c). The PCR reactions scored as positive for the presence of the human gene are indicated (marked +).
In Fig. 6B, we show the calculated extent of concordance of the PCR results with the cytogenetic data on each hybrid cell line. For the purposes of this calculation, hybrid cell lines containing individual human chromosomes in at least 10% of the cells examined were considered positive for that chromosome. Complete concordance with the data from both panels was achieved only by localizing hRPAl to chromosome 17, hRPA2 to chromosome 1, and hRPA3 to chromosome 7. In one case (NIGMS panel (GM/NA 09929), lune 7), we observed a positive signal for hRPA3 in a cell line that contained Crude HeLa cytoplasmic extract was preincubated as described under "Materials and Methods" with buffer alone or with preimmune or immune serum (indicated by +). The additional reagents required for in vitro replication were then added, and replication was carried out for 2 h at 37 "C. In some reactions, purified hRPA was added along with the reaction mixture. Products were analyzed by agarose gel electrophoresis (A) and autoradiography ( B ) , and incorporation was measured by trichloroacetic acid precipitation (C). -2' denotes the control reaction without T antigen. 0, immune serum (antibody (Ab)); 0, preimmune serum; 0, immune serum with hRPA added back; A, preimmune serum with hRPA added back.
chromosome 7 in only 2% of the cells, which was below our arbitrary limit of 10%. Further analysis of this cell line with a chromosome 7-specific alphoid DNA probe also gave a positive result.'

DISCUSSION
We report here the cloning and bacterial expression of the 14-kDa subunit of hRPA from a HeLa cDNA library. Several factors argue that we have, in fact, cloned the correct cDNA.
The only long open reading frame found within the cloned sequence encodes a protein of 121 amino acids with a calculated molecular mass of 13.6 kDa, which is in good agreement E. Wang Jabs, unpublished data.  6 7 8 9 1 0 1 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 1 9 2 0 -c . c + C M M + C 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 1 9 2 0 2 1 2 2 2 3 2 4 2 5 2 6 2 6. Genes for hRPA1, hRPA2, and hRPA3 map to chromosomes 17,1, and 7, resrntively. Genomic DNA from the NIGMS and BIOS rodent-human hybrid panels was amplified by PCR using primers specific to the 3"untranslated sequences of the three genes as with the electrophoretic behavior of the 14-kDa subunit. The sequence was determined from three separate bacteriophage clones. The cDNA obtained is probably incomplete since a 5"untranslated sequence of only 30 bp would be uncharacteristically short. The longest products obtained from a reverse transcription of poly(A)-purified HeLa cell mRNA using a primer derived from the 5'-end of the cloned cDNA were consistent with a 5"untranslated region of at least 100 bp (data not shown). We do, however, believe that we have the complete coding region of the cDNA based on the favorable context for initiation of translation around the first methionine codon (Cavener and Ray, 1991;Kozak, 1991) and on the biochemical evidence presented in this paper. The length of the amino acid sequence is similar to that of the recently published yeast homolog, which contains 122 amino acids , although only limited conservation of the primary sequence is apparent. The primary sequence contains all three sequences that were obtained by peptide sequencing of the native tryptic peptide fragments. Expression of the cloned cDNA in E. coli encodes a 14-kDa protein that co-migrates with the native hRPA3 on SDS-PAGE. The native hRPA3 protein is recognized by antisera raised against a COOH-terminal synthetic peptide and against the purified recombinant bacterial protein.
The antiserum raised against purified recombinant hRPA3 inhibits in uitro DNA replication in a manner that can be specifically reversed by adding back pure native hRPA. This suggests that hRPA3 is present in the replication-competent complex and may be involved in some of the steps of DNA replication. We cannot, however, exclude the possibility that the antiserum sequesters the hRPA holoenzyme during the preincubation. This would not necessarily indicate that the 14-kDa subunit is present at the replication fork in later stages since the structure of hRPA may well change during replication, as suggested by the phosphorylation of hRPA observed during replication (Fotedar and Roberts, 1992).
The low degree of conservation observed between the human and S. cereuisiue proteins suggests that they lacked the tight evolutionary constraints expected for a catalytic component of a required enzymatic reaction. hRPA3 may instead have evolved to interact with other proteins involved in the DNA metabolism of higher eukaryotes, causing the marked divergence in primary sequence. We suspect that the major role of the 14-kDa subunit may be to interact with other cellular proteins involved in DNA replication, recombination, and/or repair. This may serve to deliver the other functional domains of hRPA, such as the single-stranded DNA-binding domain, to their sites of action. These interactions could also help coordinate the ordered assembly of the large nucleoprotein complexes required for controlled structural modifications of DNA. A protein whose role involves interactions with other proteins would be expected to coevolve with such proteins. Over time, this leads to divergence and loss of functional equivalence. For example, yeast RPA cannot substitute in DNA replication or recombination (Erdile et al., 1991a;Moore et al., 1991). Finally, the PCR amplification of the 3'-untranslated sequence from genomic DNA of the rodent-human cell hybrid panels allowed unequivocal chromosomal localization of all three genes. hRPAl is located on chromosome 17, hRPA2 on chromosome 1, and hRPA3 on chromosome 7. Other human replication proteins that have been mapped to a chromosome so far include topoisomerase I (chromosome 20) (Kunze et al., 1989), proliferating cell nuclear antigen (chromosome 20) (Ku et al., 1989), and the catalytic subunits of polymerases a (X chromosome) (Wang et al., 1985) and 6 (chromosome 19) (Chung et al., 1991). A number of human genetic disorders involving DNA metabolism or its regulation are known. These include ataxia telangiectasia (McKinnon, 1987), xeroderma pigmentosum (Wood et al., 1988), Bloom's syndrome (Langlois et d., 1989), Cockayne's syndrome (Nance and Berry, 1992), and Fanconi's anemia (Chaganti and Houldsworth, 1991). There is considerable genetic heterogeneity in these disorders, each consisting of several known complementation groups, some of which have already been localized with varying degrees of resolution. We are presently refining the localization using various rodent-human hybrid panels and in situ hybridization to investigate potential relationships of the hRPA genes with human genetic disorders.