Replication Protein A
CHARACTERIZATION AND CRYSTALLIZATION OF THE DNA BINDING DOMAIN*

(Received for publication, August 30, 1996)

Richard A. Pfuetzner , Alexey Bochkarev Dagger , Lori Frappier § and Aled M. Edwards §par

From the Cancer Research Group, Institute for Molecular Biology and Biotechnology, Departments of Pathology and Biochemistry, McMaster University, 1200 Main St. West, Hamilton, Ontario L8N 3Z5, Canada

ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
REFERENCES


ABSTRACT

Replication protein A (RPA) is a heterotrimeric single-stranded DNA-binding protein in eukaryotic cells. The DNA binding activity of human RPA has been previously localized to the N-terminal 441 amino acids of the 70-kDa subunit, RPA70. We have used a combination of limited proteolysis and mutational analysis to define the smallest soluble fragment of human RPA70 that retains complete DNA binding activity. This fragment comprises residues 181-422. RPA181-422 bound DNA with the same affinity as the 1-441 fragment and had a DNA binding site of 8 nucleotides or less. RPA70 fragments were subjected to crystal trials in the presence of single-stranded DNA, and diffraction quality crystals were obtained for RPA181-422 bound to octadeoxycytidine. The RPA181-422 co-crystals belonged to the P212121 space group, with unit cell dimensions of a = 34.3 Å, b = 78.0 Å, and c = 95.4 Å and diffracted to a resolution of 2.1 Å.


INTRODUCTION

Replication protein A (RPA)1 is a single-stranded DNA (ssDNA)-binding protein that is highly conserved in eukaryotic cells. RPA was originally identified as a factor essential for the replication of SV40 DNA in vitro and subsequently has also been shown to play an important role in DNA repair and recombination (1, 2, 3, 4). RPA is a heterotrimer comprising 70-, 32-, and 14-kDa subunits, each of which is essential in yeast (5). The 70-kDa subunit (RPA70) binds ssDNA (6) and mediates the interaction of RPA with a number of proteins, including the primase component of DNA polymerase alpha -primase (7), the E2 protein of bovine papilloma virus (8), DNA repair factor XPA (3, 9), the EBNA1 protein of Epstein-Barr virus,2 and the acidic activation domains of VP16, GAL4, p53, and EBNA2 (8, 10, 11). The functions of the 32- and 14-kDa RPA subunits are not yet clear, but they are required for SV40 replication in vitro (6). The 32-kDa subunit is phosphorylated in a cell cycle-dependent manner and therefore may play a role in the regulation of RPA function (12, 13, 14).

The interaction of RPA with ssDNA has been characterized. RPA binds ssDNA with low cooperativity (15), and the affinity of RPA for ssDNA is dependent on both the length and the sequence of the DNA, with pyrimidine residues preferred over purines (16, 17). RPA has been reported to bind ssDNA with a binding site that spans either 90, 30, or 8-10 nucleotides depending on the assay conditions (17, 18, 19), suggesting that RPA may have two binding modes. The DNA binding domain of human RPA70 has been localized to the N terminus of RPA70; using a series of C-terminal truncation mutants, Gomes and Wold (20) showed that residues required for complete DNA binding activity reside between amino acids 1 and 441.

Despite the above-mentioned studies, the mechanism by which RPA70 interacts with ssDNA is unclear. This information can only be obtained from a high resolution structure of the RPA70 DNA binding domain bound to ssDNA. Toward this end we have used a combination of limited proteolysis and mutational approaches to further define the boundaries of the human RPA70 DNA binding domain and have explored conditions in which this domain will crystallize. Here we show that the human RPA70 DNA binding domain is located between amino acids 181 and 422, and that this RPA70 fragment forms co-crystals with octadeoxycytidine that diffract to 2.1 Å and are suitable for diffraction studies.


MATERIALS AND METHODS

RPA Constructs

RPA70 deletion mutants containing amino acids 1-442, 161-442, 181-442, 187-442, 191-442, 181-432, 181-422, and 181-412 were generated by PCR. The primers were designed such that an NdeI restriction site and a BamHI site were added to the 5'- and 3'-ends of the PCR products, respectively. The PCR products were digested with NdeI and BamHI and inserted between the NdeI and BamHI sites of the T7 polymerase expression vector pET15b (Novagen). The resulting proteins were expressed as C-terminal fusions to an N-terminal hexahistidine tag and a thrombin cleavage site. Digestion of each fusion protein with thrombin resulted in the inclusion of four residues, GSHM, at the N terminus of the RPA derivatives.

Protein Purification

Bacterial BL21(DE3)pLysS cells (21) containing the pET15b-RPA constructs were grown at 37 °C to an optical density of 0.8 (A = 600 nm) in 6 liters of Luria broth supplemented with 100 µg/ml ampicillin and 25 µg/ml chloramphenicol. RPA expression was induced by addition of isopropyl-beta -D-thiogalactopyranoside to 0.5 mM. Three hours after induction, cells were harvested by centrifugation, resuspended in 50 mM Tris-HCl (pH 7.5) and 10% sucrose, and frozen at -70 °C. Cells were thawed in the presence of EDTA, PMSF, and benzamidine, each at a final concentration of 1 mM, and NaCl at 300 mM. All subsequent steps were carried out at 4 °C. The cells were lysed by sonication, and the lysate was clarified by centrifugation at 70,000 × g in a Beckman Instruments SW 28 rotor for 30 min. Nucleic acids and other negatively charged contaminants were then removed from the lysate by adsorption to a 25-ml DE-52 column (Whatman) equilibrated in 50 mM Tris-HCl (pH 7.5), 300 mM NaCl, 10% glycerol, 1 mM PMSF, 1 mM benzamidine, and 1 mM DTT. The lysate was dialyzed against 50 mM HEPES (pH 7.4), 0.5 M NaCl, 10% glycerol, 1 mM PMSF, 1 mM benzamidine, and 5 mM imidazole (buffer A), loaded on a 15-ml metal chelating column (Novagen, Madison, WI) charged with 0.2 M nickel sulfate, and equilibrated in buffer A. The column was washed with 10 column volumes of buffer A, followed by three column volumes of buffer A containing 50 mM imidazole, and then developed with buffer A containing 300 mM imidazole. The eluate was dialyzed against 50 mM HEPES (pH 7.4), 150 mM NaCl, and 10% glycerol and then incubated with bovine thrombin (5 units/mg protein) for 8-10 h at 4 °C. After digestion, NaCl was added to 0.5 M, and the protein was passed through a 5-ml metal chelating column (charged with 0.2 M NiSO4 and equilibrated in buffer A containing 5 mM imidazole) to remove the hexahistidine-tagged fragment and other contaminants. The flow-through was dialyzed against 50 mM HEPES (pH 7.4), 50 mM NaCl, 10% glycerol, 1 mM DTT, and 1 mM EDTA (buffer B) and loaded onto a 5-ml Poros HS column (PerSeptive Biosystems, Cambridge, MA) in buffer B. The Poros HS column was developed with a 20-ml linear gradient from 50 to 500 mM NaCl in buffer B. For crystal trials, the proteins were dialyzed against 1 mM HEPES, 1 mM DTT, and 50 mM NaCl and concentrated to 15-20 mg/ml using a Centricon-10 microconcentrator (Amicon).

Proteolysis and Identification of Partial Proteolytic Fragments

10 µg of purified RPA1-442 containing the N-terminal histidine tag was treated with Staphylococcus aureas V8 protease (0.01 µg) for 30 min in a solution containing 50 mM HEPES (pH 7.4), 150 mM NaCl, and 10% glycerol. Two products were generated. The proteolytic fragments, having approximate masses of 30 and 18 kDa, were resolved by Ni-chelate chromatography; the N-terminal 18-kDa fragment bound to the Ni-chelate column, and the 30-kDa fragment passed through the column. Each fragment was assayed for ssDNA binding activity as described below. The 30-kDa fragment, which retained full activity, was blotted onto a polyvinylidene difluoride membrane (5 µg of protein) for N-terminal sequencing (University of British Columbia Biotechnology Facility, Vancouver, British Columbia, Canada). 5 µg of the purified 30-kDa fragment was dialyzed extensively against water, lyophilized, and then analyzed by electrospray mass spectrometry (University of Waterloo Protein Analysis Laboratory, Waterloo, Ontario, Canada).

Electrophoretic Mobility Shift Assays

Fragments of RPA70 were incubated with 16 fmol of 32P-end-labeled oligonucleotide (either a purified (dC)8 or (dC)18) for 20 min at room temperature in 10 µl of 25 mM HEPES (pH 7.4), 200 mM NaCl, 4% glycerol, 1 mM MgCl2, 0.5 mM DTT, and 0.01% Nonidet P-40. Free DNA and RPA·DNA complexes were then resolved by gel electrophoresis at 4 °C on an 8% polyacrylamide gel containing 1% glycerol and 0.5 × TBE (45 mM Tris, 45 mM borate, and 1 mM EDTA) and visualized by autoradiography. Bound and free DNA were quantified using a PhosphorImager (Molecular Dynamics).

Protein Crystallization and Data Collection

Crystals of RPA and of RPA·ssDNA complexes were grown by the method of hanging or sitting drop vapor diffusion.

For RPA161-442, the hanging drop consisted of an equal volume of the protein solution (10-20 mg/ml) and the reservoir solution, which contained 4.2 M NaCl, 50 mM Tris-HCl (pH 8.5), and 5 mM DTT. Crystals grew at room temperature to 0.6 × 0.1 × 0.1 mm over the course of 5-7 days. For RPA181-442, the hanging drop consisted of an equal volume of the protein solution (10-20 mg/ml) and the reservoir solution, which contained 3.9 M NaCl, 50 mM Tris-HCl (pH 8.5), and 10 mM DTT. Crystals again grew at room temperature to 0.6 × 0.1 × 0.1 mm over the course of 5-7 days. For the co-crystals of RPA181-422 with ssDNA, the protein (4-8 mg/ml) was incubated on ice for 20 min in the presence of a 1.5-fold molar excess of either 5'-phosphorylated (dC)6, (dC)8, (dC)10, or (dC)12. The complexes were mixed with an equal volume of a solution containing 24-27% PEG2000monomethyl ether, 20 mM NaCl, 10 mM MgCl2, and 5 mM DTT and buffered with either 100 mM bis-tris (pH 6.7), 100 mM MES (pH 6.8), or 100 mM HEPES (pH 6.9 and 7.0). Crystals of the protein complexed with (dC)8 and (dC)10 grew at 4 °C over the course of 2-3 weeks. Co-crystals of RPA181-442 with (dC)8 were grown in the same way, except the mother liquor contained 28-32% PEG600, 100 mM Tris (pH 7.5-8.5), 10 mM MgCl2, 20 mM NaCl, and 5 mM DTT. Crystals of RPA161-442 and RPA181-442 were mounted in glass capillaries. For the complex of RPA181-422 and (dC)8, the crystals were mounted in a loop made of a dental floss fiber, cryoprotected in the mother liquor supplemented to 35% PEG2000monomethyl ether, and flash frozen in a stream of liquid nitrogen. X-ray diffraction data were collected on a RAXISII image plate area detector equipped with a rotating anode x-ray source. The diffraction data were integrated with DENZO, and the intensities were scaled with SCALEPACK (Z. Otwinowski, Yale University School of Medicine).


RESULTS

Limited Proteolysis Defines a DNA Binding Domain

The goal of these studies was to crystallize a domain of RPA that harbored ssDNA binding activity. Our initial strategy was to use limited proteolysis to identify a stable domain of RPA with ssDNA binding activity and then to subject that domain to crystallization trials. We started with an RPA70 fragment containing amino acids 1-442 (RPA701-442), because this fragment was shown by Gomes and Wold (20) to be soluble and to bind ssDNA. RPA701-442 was produced in Escherichia coli, purified to homogeneity, and incubated with either trypsin, chymotrypsin, or V8 protease. A V8 protease-resistant fragment of RPA701-442 retained the ability to bind ssDNA (data not shown). This fragment was purified and shown by N-terminal sequencing and mass spectrometry to comprise amino acids 161-442 (RPA70161-442). The RPA70161-442 fragment was then subcloned using PCR, expressed in E. coli as a C-terminal fusion to a hexahistidine tag and a thrombin cleavage site, and purified to homogeneity. The DNA binding activities of purified RPA701-442 and RPA70161-442 were compared by electrophoretic mobility shift assays (EMSAs). The RPA701-442 and RPA70161-442 proteins bound (dC)18 with affinities of 2.8 and 0.5 × 10-8 M, respectively (Table I).

Table I.

Properties of RPA70 fragments

Dissociation constants were calculated from the point of half-maximal DNA binding. The data in Fig. 1 were used to calculate the dissociation constants for the (dC)8 oligonucleotide.
RPA70 fragment Solubility Accessibility of thrombin cleavage site (dC)18 (dC)8

1 -442 + + 2.8 1.3
161 -442 + + 0.5 1.5
181 -442 + + 1.2 1.3
187 -442 +  - NDa ND
191 -442  - ND  ND ND
181 -432 + +  ND 2.2
181 -422 + +  ND 1.0
181 -412 +b +  ND 1.3

a  ND, not determined.
b  Soluble only to 3 mg/ml.

Determination of the Size of the ssDNA Binding Site in RPA70161-442

We wished to perform crystallization trials of RPA70161-442 in the presence and absence of DNA. Therefore, we determined the approximate size of the RPA70161-442 ssDNA binding site to provide an indication of the length of the ssDNA oligonucleotides that should be used in the co-crystallization trials. The size of the ssDNA binding site in the RPA heterotrimer is a matter of controversy and has been reported in the literature as 90, 30, and 8-10 nucleotides (17, 18, 19). To determine the size of the binding site for RPA70161-442, we explored the binding of this fragment to oligonucleotides of different lengths. Two observations suggest that RPA70161-442 binds tightly to oligonucleotides as short as 8 mer. First, the affinity of RPA70161-442 for 8 and 18 mers are very similar (Table I). These values are also similar to the affinity reported previously for RPA701-441 for (dT)30 (20). Second, we have observed that two complexes of different electrophoretic mobility form on oligonucleotides as small as 16 mer; the complex of slower mobility was only formed at higher protein concentrations (data not shown). We concluded that the slower migrating complex represents a DNA molecule to which two RPA70161-442 molecules are bound. The DNA binding site in each of the two RPA70161-442 molecules in the complex is therefore 8 nucleotides or less. On the basis of these data, we initiated co-crystal trials of RPA70161-442 with a panel of oligonucleotides ranging from 6-12 mer.

Crystallization of RPA70161-442

Crystallization trials of RPA70161-442 were performed using commercial crystal screens in the presence and absence of ssDNA of various lengths. Crystallization trials of RPA70161-442 complexed with either a 6, 8, 10, or 12 mer of oligo(dC) were performed using a DNA-protein crystal screen developed in the laboratory of S. Burley (Rockefeller University).3 We embarked on crystallization trials with RPA70161-442, because protease fragments have often been used successfully for crystallization. In the presence of DNA, no crystals were obtained. In the absence of DNA, large rod-like crystals of free RPA formed in 3.9 M NaCl at pH 8.5 almost immediately. However, these crystals diffracted poorly (dmin, 5 Å) and were highly mosaic (mosaic spread, 1.4°) and very fragile (Table II). A partial data set was collected, the unit cell was tentatively assigned as C222, and the cell constants were a = 76.4 Å, b = 83.4 Å, and c = 122.2 Å.

Table II.

Crystal data


Crystal
Space group Unit cell
Diffraction limit
RPA70 fragment DNA a b c

Å Å
161 -442 None C222 76.4 83.4 122.2 4.0
181 -442 None C222 75.5 85.6 121.4 3.5
181 -442 (dC)8 ND a ND ND ND ND
181 -422 None  ND ND ND ND ND
181 -422 (dC)8 P212121 34.3 78.0 95.4 2.1

a  ND, not determined.

Determining the N-terminal Border of the ssDNA Binding Domain

We reasoned that the use of proteolysis to define domain boundaries might be limited by the disposition of protease sites in the protein sequence. The RPA70161-442 fragment might therefore have disordered N or C termini that might inhibit formation of well ordered crystals. Inspection of the sequence of the N-terminal portion of RPA70161-442 revealed a paucity of potential protease sites. We therefore chose to further define the N-terminal boundary of the DNA binding domain using deletion analysis. The N-terminal boundary of the DNA binding domain was determined by generating a series of RPA70 fragments starting at residues 171, 181, 187, or 191 and extending to residue 442. Like RPA701-442 and RPA70161-442, these RPA70 fragments were produced in E. coli as hexahistidine fusions. The fragments were initially monitored by level of expression and degree of solubility. All of these constructs produced soluble proteins except the 191-442 fragment, which was largely insoluble (Table I). These solubility characteristics suggest that the N-terminal boundary of the DNA binding domain lay between amino acids 187 and 191. The soluble RPA70 fragments were purified to homogeneity and incubated with thrombin to remove the hexahistidine tag. The thrombin cleavage site fused to the N terminus of RPA70187-442 was found to be resistant to cleavage, suggesting that amino acid 187 is very close to the boundary of the DNA binding domain and that the thrombin cleavage site is masked (Table I). We concluded that the N-terminal border of the ssDNA binding domain was near residue 181 and that RPA70181-442 represented a smaller DNA binding domain. The DNA binding activity of RPA70181-442 was confirmed using an EMSA. RPA70181-442 bound (dC)18 and (dC)8 oligonucleotides with a similar affinity as RPA701-442 (Table I and Fig. 1).


Fig. 1. Binding of the RPA70 fragments to (dC)8. RPA70 fragments containing the amino acids indicated were titrated with 16 fmol of (dC)8 oligonucleotide, and DNA binding was assessed by EMSAs as described under "Materials and Methods". Quantification of DNA bound in each assay is shown.
[View Larger Version of this Image (19K GIF file)]


Crystallization of RPA70181-442

The N-terminal deletion mutant RPA70181-442 was subjected to the same crystallization trials as were used for RPA70161-442. In the presence of DNA, co-crystals of RPA70181-442 with (dC)8 were obtained. The crystals were very thin needles and often found mixed with oily droplets. The small size precluded diffraction analysis. In the absence of DNA, RPA70181-442 crystallized under the exact conditions as did RPA70161-442. X-ray diffraction analysis revealed that the crystals of the two different RPA fragments had very similar unit cell dimensions and likely belonged to the same space group (Table II). However, the crystals of RPA70181-442 were substantially better ordered that those of RPA70161-442. The diffraction limit improved from 4 to 3.5 Å, and the mosaic spread dropped from 1.4 to 0.7°. We concluded from these analyses that RPA70 residues 161-181 were most likely disordered in the RPA70161-442 crystals, did not participate in crystal packing, and likely were the cause of the poorer diffraction quality. However, the crystals of RPA70181-442, although improved over those of RPA70161-442, were not suitable for structure determination. We therefore initiated a C-terminal deletion analysis despite the fact that the C-terminal region of RPA70181-442 is laden with potential protease sites.

Determining the C-terminal Border of the ssDNA Binding Domain

C-terminal truncations of RPA70181-442 were constructed using PCR deletion mutagenesis. Three shorter fragments, RPA70181-432, RPA70181-422, and RPA70181-412, were produced in E. coli as hexahistidine tags and purified to homogeneity. The DNA binding activity of each of the three derivatives was determined using the (dC)8 oligonucleotide and was found to be equivalent to RPA70181-442 (Table I and Fig. 1). However, RPA70181-412 was considerably less soluble than the other derivatives. Each of RPA70181-442, RPA70181-432, and RPA70181-422 was soluble to 20 mg/ml but in the same solution conditions, RPA70181-412 precipitated at 3 mg/ml. We concluded that the C-terminal boundary of the ssDNA binding domain was near residue 422.

Crystallization of RPA70181-422

The RPA70181-422 fragment was crystallized in the presence and absence of DNA. Crystals of RPA70181-422 grew from NaCl under the same conditions as for RPA70181-442 and RPA70161-442; however, the crystals of RPA70181-422 were much smaller and had rounded crystal edges. We were unable to collect any usable diffraction data from these crystals. By contrast, highly ordered crystals of RPA70181-422 complexed with both (dC)8 and (dC)10 oligonucleotides were grown. The crystals were solubilized and resolved on a denaturing polyacrylamide gel; both the protein and DNA were visualized by silver staining. The crystals of RPA70181-422 complexed with (dC)8 gave excellent diffraction (dmin, 2.1 Å) and had low mosaicity (0.25°). The crystal parameters and data collection statistics are shown in Table II. The data from these co-crystals were 97.3% complete and had an Rsym value of 6.9%. A heavy atomic derivative of the crystals was prepared by co-crystallization with a (dC)8 molecule containing a 5-iodocytosine at the second position. Comparing the diffraction from the derivatized and native crystals using difference Patterson analysis confirmed both the space group assignment as P212121 and that the crystals contained ssDNA.


DISCUSSION

In the interest of determining the mechanisms by which proteins bind to ssDNA, we set out to derive crystals of RPA that were suitable for structure determination. To accomplish this, we developed an approach to crystallization that may prove successful for a wide variety of proteins. This approach, which we term the smallest active fragment (SAF) approach, is based on two tenets. First, structural biologists wishing to understand a particular activity need analyze only the fragment of the protein that harbors the biologically relevant activity. Second, eliminating structural or conformational microheterogeneity within a protein or protein domain greatly increases the probability of crystallization. Such heterogeneity might be caused by the motion of unstructured loops or tails. With these tenets in mind, we propose that the careful selection of the borders of protein domains, with the aim of reducing the amount of unstructured protein, will dramatically increase the probability of forming well ordered crystals.

The borders of structural domains have often been established through the use of limited proteolysis (22, 23, 24), since the connections between protein domains are exposed and thus are better protease substrates. However, limited proteolysis can map the borders of protein domains only to an approximation. Proteases have inherent sequence specificity; the extent to which the exact borders of the domain can be mapped is constrained considerably by the sequences surrounding the domain. Therefore, protein domain mapping by proteolysis will likely generate domains that have N- and C-terminal tails that extend from the compact domain. We suggest that these tails, which cause conformational heterogeneity, might inhibit the growth of well ordered crystals. The corollary of this statement is that the formation of well ordered crystals can be achieved by removing the highly mobile parts of the proteins. The SAF method, which uses partial proteolysis to determine the approximate edges of the domain and then deletion analysis to map the domain boundaries more accurately, was designed to derive protein fragments with minimal conformational heterogeneity.

Two examples from our laboratories illustrate the utility of the method. In the first case, we were attempting to crystallize the EBNA1 protein from Epstein-Barr virus (25). The initial construct (EBNA1452-641) had been commonly used for biological assays. We chose this fragment because it expressed well, was both soluble and monodisperse at high concentrations, and was fully active for DNA binding. This fragment formed co-crystals with DNA in our first round of screening. However, the crystals were poorly ordered and diffracted only to about 6 Å. A version of EBNA1 in which the C-terminal acidic tail was removed (EBNA1459-619) was then subjected to crystal trials. This version also formed poor crystals. EBNA1459-619 was then subjected to partial proteolysis, and a smaller, active tryptic fragment was identified (EBNA1470-619). This tryptic fragment, subcloned and expressed in bacteria, formed well ordered, but twinned, crystals. We then removed 12 residues from the C terminus of this fragment because we were aware, from deletion analysis performed in the laboratory of Hayward and co-workers (26, 27), that the C-terminal residues from residue 607-619 were not required for DNA binding activity. This smaller fragment (EBNA1470-607) was expressed, purified, and crystallized. The crystals were well ordered, and from these the structure was solved (28).

For RPA, we began our studies with a fragment of RPA70 (RPA1-442) that was highly soluble and monodispersed and retained ssDNA binding activity. Partial proteolysis generated a fragment (RPA161-442) that formed poorly ordered crystals in the absence of DNA and no crystals in the presence of DNA. As described in this article, we then initiated deletion analysis from both N and C termini and identified the smallest soluble fragment that harbored ssDNA binding activity (RPA181-422). This fragment formed well ordered crystals in the presence of DNA in the first rounds of screening, and the structure was determined using this crystal form (29).

For our RPA analysis, a crucial component to the design of our crystal trials was the determination of the approximate size of the DNA binding site for RPA. Under our solution conditions, the minimal DNA binding domain of RPA70 binds with high affinity to an octanucleotide (8-mer binding domain); therefore, we investigated the co-crystallization of RPA with a range of ssDNAs varying in length from 6 to 12 deoxynucleotides. Our estimation of the binding site size for this domain is consistent with the size of the heterotrimeric RPA DNA binding site, which has also been reported to be 8 nucleotides (18). For the RPA181-422, we observed little difference between the binding affinity for shorter (8 mer) or longer (18 mer) oligonucleotides. By contrast, the RPA heterotrimer apparently can adopt an alternate mode of DNA binding that covers and binds with higher affinity to longer lengths of ssDNA (17, 18, 19). If the RPA181-422 domain is solely responsible for this 30-nucleotide interaction, it must do so in a conformation that is favored in solution conditions that were not explored in this study. Alternatively, other regions of RPA distinct from RPA181-422 might contribute to the binding of longer DNA fragments. The elucidation of the crystal structure of RPA181-422 bound to DNA will clearly contribute to our understanding of the these multiple modes of DNA binding.

In summary, the general utility of the SAF method remains to be established. However, there are already a growing number of examples in which partial proteolytic analysis has provided the key to identifying protein domains that form well ordered crystals. One can assume that the combination of proteolytic analysis and iterative deletion analysis described here can only increase the probability of generating crystals of protein domains.


FOOTNOTES

*   This work was supported by grants from the National Cancer Institute of Canada and the Medical Research Council of Canada (to A. M. E. and L. F.). The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Dagger    Supported by a fellowship from the National Cancer Institute of Canada.
   Research Scientist of the National Cancer Institute of Canada.
par    Research Scholar of the Medical Research Council of Canada.
§   To whom correspondence should be addressed.
1    The abbreviations used are: RPA, replication protein A; ss, single-stranded; PCR, polymerase chain reaction; PMSF, phenylmethylsulfonyl fluoride; EBNA, Epstein-Barr nuclearantigen; dmin, diffraction limit; DTT, dithiothreitol; TBE, Tris borate/EDTA; PEG, polyethylene glycol; MES, 2-(N-morpholino)ethanesulfonic acid; EMSA, electrophoretic mobility shift assay; SAF, smallest active fragment.
2    D. Zhang, L. Frappier, and M. O'Donnell, unpublished data.
3    S. Burley, personal communication.

REFERENCES

  1. Coverly, D., Kenny, M., Munn, M., Rupp, W., Lane, D., and Wood, R. D. (1991) Nature 349, 538-541 [CrossRef][Medline] [Order article via Infotrieve]
  2. Moore, S., Erdile, L., Kelly, T., and Fishel, R. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 9067-9071 [Abstract]
  3. He, Z., Henricksen, L. A., Wold, M. S., and Ingles, C. J. (1995) Nature 374, 566-569 [CrossRef][Medline] [Order article via Infotrieve]
  4. Wold, M. S., and Kelly, T. (1988) Proc. Natl. Acad. Sci. U. S. A. 85, 2523-2527 [Abstract]
  5. Brill, S. J., and Stillman, B. (1991) Genes Dev. 5, 1589-1600 [Abstract]
  6. Erdile, L. F., Heyer, W.-D., Kolodner, R., and Kelly, T. J. (1991) J. Biol. Chem. 266, 12090-12098 [Abstract/Free Full Text]
  7. Dornreiter, I., Erdile, L., Gilbert, I., von Winkler, D., Kelly, T., and Fanning, E. (1992) EMBO J. 11, 769-776 [Abstract]
  8. Li, R., and Botchan, M. R. (1993) Cell 73, 1207-1221 [Medline] [Order article via Infotrieve]
  9. Li, L., Lu, X., Peterson, C. A., and Legerski, R. J. (1995) Mol. Cell. Biol. 15, 5396-5402 [Abstract]
  10. He, Z., Brinton, B. T., Greenblatt, J., Hassell, J. A., and Ingles, C. J. (1993) Cell 73, 1223-1232 [Medline] [Order article via Infotrieve]
  11. Tong, X., Wang, F., Thut, C. J., and Kieff, E. (1995) J. Virol. 69, 585-588 [Abstract]
  12. Din, S.-U., Brill, S. J., Fairman, M. P., and Stillman, B. (1990) Genes Dev. 4, 968-977 [Abstract]
  13. Dutta, A., and Stillman, B. (1992) EMBO J. 11, 2189-2199 [Abstract]
  14. Fotedar, R., and Roberts, J. M. (1992) EMBO J. 11, 2177-2187 [Abstract]
  15. Kim, C., and Wold, M. S. (1995) Biochemistry 34, 2058-2064 [Medline] [Order article via Infotrieve]
  16. Kim, C., Paulus, B. F., and Wold, M. S. (1994) Biochemistry 33, 14197-14206 [Medline] [Order article via Infotrieve]
  17. Kim, C., Snyder, R. O., and Wold, M. S. (1992) Mol. Cell. Biol. 12, 3050-3059 [Abstract]
  18. Blackwell, L. J., and Borowiec, J. A. (1994) Mol. Cell. Biol. 14, 3993-4001 [Abstract]
  19. Alani, E., Thresher, R., Griffith, J. D., and Kolodner, R. D. (1992) J. Mol. Biol. 227, 54-71 [Medline] [Order article via Infotrieve]
  20. Gomes, X. V., and Wold, M. S. (1995) J. Biol. Chem. 270, 4534-4543 [Abstract/Free Full Text]
  21. Studier, F. W., Rosenberg, A. H., Dunn, J. J., and Dubendorff, J. W. (1990) Methods Enzymol. 185, 60-89 [Medline] [Order article via Infotrieve]
  22. Finnin, M. S., Hoffman, D. W., Kreuzer, K. N., Porter, S. J., Schmidt, R. P., and White, S. W. (1993) J. Mol. Biol. 232, 301-304 [CrossRef][Medline] [Order article via Infotrieve]
  23. Hostomska, Z., Matthews, D. A., Davies, J. F., Nodes, B. R., and Hostomsky, Z. (1991) J. Biol. Chem. 266, 14697-14702 [Abstract/Free Full Text]
  24. Kycia, J. H., Biou, V., Shu, F., Gerchman, S. E., Graziano, V., and Ramakrishnan, V. (1995) Biochemistry 34, 6183-6187 [Medline] [Order article via Infotrieve]
  25. Barwell, J. A., Bochkarev, A., Pfuetzner, R. A., Tong, H., Yang, D. S. C., Frappier, L. D., and Edwards, A. M. (1995) J. Biol. Chem. 270, 20556-20559 [Abstract/Free Full Text]
  26. Ambinder, R. F., Mullen, M., Chang, Y., Hayward, G. S., and Hayward, S. D. (1991) J. Virol. 65, 1466-1478 [Medline] [Order article via Infotrieve]
  27. Shah, W. A., Ambinder, R. F., Hayward, G. S., and Hayward, S. D. (1992) J. Virol. 66, 3355-3362 [Abstract]
  28. Bochkarev, A., Barwell, J., Pfuetzner, R., Furey, W. J., Edwards, A. M., and Frappier, L. D. (1995) Cell 83, 39-46 [Medline] [Order article via Infotrieve]
  29. NatureNature in pressBochkarev, A., Pfuetzner, R. A., Edwards, A. M., and Frappier, L. Nature, in press

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.