HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 276, Issue 44, 40668-40679, November 2, 2001

This Article Abstract Full Text (PDF) All Versions of this Article: 276/44/40668    most recent M106373200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Song, M.-S. Articles by McHenry, C. S. Search for Related Content PubMed PubMed Citation Articles by Song, M.-S. Articles by McHenry, C. S. Social Bookmarking                      What's this?

Carboxyl-terminal Domain III of the ' Subunit of the DNA Polymerase III Holoenzyme Binds ^*

Min-Sun Song, H. Garry Dallmann, and Charles S. McHenry

From the Department of Biochemistry and Molecular Genetics, University of Colorado Health Sciences Center, Denver, Colorado 80262

Received for publication, July 9, 2001, and in revised form, August 16, 2001

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The  and ' subunits are essential components of the DNA polymerase III holoenzyme, required for assembly and function of the DnaX-complex clamp loader (₂'). The x-ray crystal structure of ' contains three structural domains (Guenther, B., Onrust, R., Sali, A., O'Donnell, M., and Kuriyan, J. (1997) Cell 91, 335-345). In this study, we localize the -binding domain of ' to a carboxyl-terminal domain III by quantifying the interaction of  with a series of ' fusion proteins lacking specific domains. Purification and immobilization of the fusion proteins were facilitated by the inclusion of a tag containing hexahistidine and a short biotinylation sequence. Both NH₂- and COOH-terminal-tagged full-length ' were soluble and had specific activities comparable with that of native '.  and ' form a 1:1 heterodimer with a dissociation constant (K_D) of 5 × 10⁷ M determined by equilibrium sedimentation. The K_D determined by surface plasmon resonance was comparable. Domain III alone bound  at an affinity comparable to that of wild type ', whereas proteins lacking domain III did not bind . Using a panel of domain-specific anti-' monoclonal antibodies, we found that two of the domain III-specific monoclonal antibodies interfered with -' interaction and abolished the replication activity of DNA polymerase-III holoenzyme.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Although different species employ divergent DNA polymerases, a number of general features and strategies are conserved. The hallmark traits of all major replicative DNA polymerases are fast elongation rates and processivity. Polymerases from bacterial to eukaryotic sources harness energy-dependent assembly and complex molecular architectures to effect these distinctive properties. For example, Escherichia coli DNA polymerase III holoenzyme (holoenzyme),¹ the phage T4 DNA polymerase, and mammalian DNA polymerase  all require auxiliary subunits and ATP hydrolysis.

The E. coli DNA polymerase III holoenzyme is widely studied as a prototypical replicative complex. It contains three functional units: a polymerase core (pol III), a sliding clamp processivity factor (₂), and a clamp assembly apparatus (DnaX complex). The pol III subassembly consists of the  subunit, which contains the DNA polymerase activity; the  subunit, which provides a 3'-5' exonuclease proofreading activity; and the  subunit, the function of which is not yet well understood (1-5). In E. coli, high processivity is enabled by the  subunit, which dimerizes to form a ring-shaped factor that surrounds DNA, and tethers the core polymerase (6, 7). The DnaX complex loads ₂ onto a primed template in an ATP-dependent reaction (8-10, 18). The  and  subunits are both products of the dnaX gene (11, 12);  is a truncated version of  arising from a 1 ribosomal frameshift (13-16). The DnaX proteins comprise the ATPase that drives the ₂ loading onto a primed template, and facilitate the replication complex assembly (8, 17-19). The carboxyl-terminal extension of , absent from , is responsible for dimerization of pol III (20, 21) and binding to the DnaB helicase, effectively coupling all of the replicative activities of the fork into one complex (22-24).

The  and ' subunits, key components of the DnaX complex (DnaX₃'), are essential in DNA replication both in vivo and in vitro. Bacterial strains bearing chromosomal knock-outs of either holA () or holB (') gene were not viable (25). Biochemical studies showed that in addition to their roles in the assembly and function of the DnaX complex,  and ' are also part of the initiation complex, and enable efficient DNA chain elongation (25). Results from proteolytic digestion experiments suggest that the -' interaction is involved in transducing the ATP driven conformational change of the DnaX complex (28, 30, 49). Despite the import of the clamp loader, little is known about the molecular mechanisms underlying the functions of  and '. We do know that ' interacts with both  and DnaX, holding them together within the DnaX complex (26). The  subunit has a positive synergistic effect on DnaX-' binding (27), and this subunit also interacts with the processivity factor , facilitating its binding to primed DNA (28-30). The ' subunit is a C-shaped molecule comprised of three structural domains determined by x-ray crystallography (31). Although the crystal structure of ' is known, the function of ' domains within the holoenzyme have not been elucidated.

As a first step toward deepening our understanding of the mechanism of clamp loading, we generated fusion proteins corresponding to different ' domains. In this article, we report the use of these reagents to map the portion of the ' subunit responsible for binding the  subunit. Using surface plasmon resonance, we identified domain III (COOH-terminal 127 amino acids) as the  subunit-binding domain of '. A panel of domain-specific mAbs was used to corroborate the assignment of domain III as the  binding portion of '.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Strains-- Initial molecular cloning procedures and plasmid propagation employed E. coli DH5 (F 80dlacM15 (lacZYA-argF) U169 endA1 recA1 hsdR17(r_K mK⁺) deoR thi-1 supE44 gyrA96 relA1), which was purchased from Life Technologies Inc. (Rockville, MD). E. coli BL21 (F ompT hsdS_B(r_Bm_B) gal dcm), and MGC1030PR (DMSO 1422) (mcrA mcrB IN(rrn_DrrnE)1 lexA3 (uvrD)::T_c^r (ompT)::K_m^r) were used for protein expression. E. coli BL21 was from Novagen, Inc. (Madison, WI), and MGC1030PR (DMSO 1422) was previously generated in our laboratory.

Reagents and Materials-- Synthetic oligonucleotides (Table I) were purchased from Life Technologies. SDS-PAGE protein standards were obtained from New England Biolabs (Beverly, MA). Coomassie Plus Protein Assay Reagent was obtained from Pierce Inc. (Rockford, IL). MonoQ HR 5/5, Superose 6 HR10/10 FPLC, and the NAP-25 pre-packed desalting columns were from Amersham Pharmacia Biotech (Piscataway, NJ). Ni²⁺-NTA-agarose, QIAquick Gel extraction kits, QIAquick PCR purification kits, and plasmid preparation kits were purchased from Qiagen (Valencia, CA). Biacore CM5 sensor chips (research grade), P-20 surfactant, N-hydroxysuccinimide, N-ethyl-N'-(3-diethylamino-propyl)carbodiimide, and ethanolamine hydrochloride were obtained from Biacore Inc. (Piscataway, NJ). Protease inhibitors and d-biotin were purchased from Sigma. Transfer membranes (Immobilon-P) were from Millipore (Burlington, MA).

View this table: [in this window] [in a new window]   Table I Oligonucleotides used for amplification of sequences encoding ' fusion proteins lacking specific domains

Proteins, Enzymes, and Antibodies-- Previously described methods were used for the purification of the native DNA polymerase III holoenzyme (36), the pol III core (5), and the - (38), - and - (10), - and '- (25), and - (39) subunits. SSB was prepared using the method of Griep and McHenry (40). BSA was purchased from Sigma. Bovine IgG was obtained from Bio-Rad. Control non-immune mouse IgG (ImmunoPure Mouse IgG) and streptavidin were from Pierce. Rabbit anti-mouse Fc was purchased from Biacore. Lysozyme was from Worthington Biochemical Co. (Lakewood, NJ). Restriction enzymes, T4 DNA kinase, and T4 DNA ligase were obtained from New England Biolabs. Pre-stained molecular mass markers were from Life Technologies. The LMW (low molecular weight)-gel filtration calibration kit was from Amersham Pharmacia Biotech; this kit contains each of the following protein standards: bovine pancreas ribonuclease A (13.7 kDa), bovine pancreas chymotrypsinogen A (25 kDa), hen egg ovalbumin (43 kDa), and bovine serum albumin (67 kDa). Various ' fusion proteins are described under "Construction of Expression Vectors."

Buffers-- Tris sucrose buffer is 50 mM Tris (pH 7.5) and 10% (w/v) sucrose. Buffer A is 50 mM sodium phosphate (pH 7.4), 500 mM NaCl, 10% (v/v) glycerol, 0.5 mM phenylmethylsulfonyl fluoride, and 0.5 mM benzamidine. Buffer B is the same as Buffer A, except that it contains 20% (v/v) glycerol. Buffer C is 20 mM KPO₄ (pH 7.4), 1 mM DTT, and 0.1 mM phenylmethylsulfonyl fluoride. Buffer D is 50 mM Hepes (pH 7.4), 100 mM NaCl, 10% (v/v) glycerol, and 0.5 mM DTT. Buffer E is 20 mM Hepes (pH 7.4), 100 mM NaCl, 0.5% (v/v) glycerol, 10 mM MgCl_2, and 1 mM DTT. Buffer F is 50 mM Hepes (pH 7.4), 100 mM potassium glutamate, 10% (v/v) glycerol, 5 mM DTT, 10 mM Mg(OAc)₂, 200 µg/ml BSA, 0.02% (v/v) Tween 20, and 200 µM ATP. HBS buffer is 10 mM Hepes (pH 7.4), 150 mM NaCl, 3.4 mM EDTA, and 0.005% (v/v) P-20 surfactant. HKGM buffer is 50 mM Hepes (pH 7.4), 100 mM potassium glutamate, 10 mM Mg(OAc)₂, and 5 mM DTT.

Construction of Expression Vectors-- Plasmids expressing fusion proteins corresponding to different portions of the ' subunit preceded or followed by a tripartite, multipurpose tag sequence were generated as described below. The tag sequence contains a 13-amino acid biotinylation sequence, a hexahistidine sequence, and a thrombin cleavage site (41). The hexahistidine sequence enables purification via Ni²⁺-chelate affinity chromatography, and the biotinylation sequence permits specific immobilization as well as blotting-based identification by virtue of streptavidin/biotin binding. The thrombin cleavage site was not employed in these studies. It was included as a precautionary measure, to enable ready separation of the ' sequences from the tag sequences in the event that the latter introduced adverse effects (41). In each case, the tag sequence as well as the P_A1/04/03 promoter/operator (P_A1) were excised from either pDRKN(M) (DMSO 1002) or pDRK-C (DMSO 964) (41), and the gene encoding ' (holB) or portions thereof were derived from pD'DSP.4 (DMSO 965) (42). The nomenclature employed for the expressed fusion proteins reflects the relevant portion of the ' subunit (domains) and the position of the fusion peptide tag (Fig. 2B). A construct corresponding to the entire ' subunit tagged at the carboxyl end is denoted '_C. An amino-terminal tagged fusion protein corresponding to the entire ' subunit (except lacking the initiating Met) is referred to as '_N. Fusion proteins corresponding to the first or third domains of ' followed by the tag are named I_C and III_C, respectively. A protein comprised of the first two domains of ' and a COOH-terminal tag is denoted I + II_C. Finally, a fusion protein bearing the tripartite tag followed by the second and third domains of ' is named II + III_N.

After the construction of each expression plasmid, the sequences of the inserted ' codons and flanking cloning sites were verified by DNA sequencing and restriction enzyme digestion. To generate the plasmid encoding '_C (pMSCd'), a cloning intermediate (pD'DSP.4B) containing holB, but lacking its stop codon, was first generated by replacing the carboxyl-terminal PstI/SalI fragment of the holB sequence within the parental vector (pD'DSP.4) with a PCR-generated fragment (Fig. 1A). The pD'DSP.4 (parental) plasmid also served as the template for the PCR amplification of the replacing fragment with primers P1 and P2. Primer P1 is complementary to the codons for ' amino acids 262-266 (Table I). P2 contains a complementary 3'-region extending from the codons for ' amino acids 330-334, followed by a non-complementary 5'-region containing a SpeI cloning site. The 1050-base pair BamHI/SpeI fragment from the resultant plasmid (pD'DSP.4B) was ligated to the similarly digested pDRK-C vector to generate pMSCd' (Fig. 1A), which was transformed into BL-21 to generate a strain (DMSO 1586) for expression of '_C.

To produce a vector encoding '_N, the PstI/SalI fragment (165 base pairs) of pD'DSP.4, which encodes the carboxyl-terminal portion of the holB sequence, was first ligated to the PstI/SalI digested pDRKN(M) fusion vector to produce pDRKN(M)B (Fig. 1B). The remaining amino-terminal portion of the holB coding sequence was supplied by a PCR-generated fragment, which was generated using primers P3 and P2 with template pD'DSP.4 (Fig. 2B). P3 contained a PstI cloning site in a non-complementary 5'-region, followed by a complementary 3'-region annealed to the holB codons for ' amino acids 2-6 (Table I). The resultant PCR product was digested with PstI to generate an 862-base pair PstI/PstI fragment, dephosphorylated, and ligated to the PstI-digested pDRKN(M)B to produce pMSNd' (Fig. 1B). Orientation of the insert was verified by evaluating the size of the NsiI/SalI restriction fragment of pMSNd'. pMSNd' was transformed into BL-21 to generate a strain (DMSO 1592) for expression of '_N.

To generate a plasmid encoding II + III_N (pMSND2+3), primers P10 and P2 were first used with template pD'DSP.4 to PCR amplify a truncated holB fragment (' amino acids 164-334)(Fig. 2, A and B). P10 contained a PstI site in a non-complementary 5'-region, followed by a complementary 3'-region extending from the codons for ' amino acids 164-168 (Table I). The amplified holB PCR fragment was digested with PstI, resulting in a 371-base pair PstI/PstI fragment. The PCR product was dephosphorylated, then ligated to the PstI-digested pDRKN(M)B fusion vector (Fig. 1B) to produce pMSND2+3. Orientation of the insert was verified by evaluating the size of NsiI/SalI digestion fragment of pMSND2+3. pMSND2+3 was transformed into BL-21 to generate a strain (DMSO 1588) for expression of II+III_N.

To produce a plasmid encoding II + III_C (pMSCD1+2), primers P5 and P6 were first used with template pD'DSP.4 to generate a PCR product containing a partial holB fragment (' amino acids 1-207) (Fig. 2, A and B). PCR primer P5 was annealed to nucleotides 7538-7637 of pD'DSP.4. This stretch is located 104 nucleotides upstream of the vector's BamHI cloning site, which is 18 nucleotides upstream of the holB start codon. P6 contained the holB codons encoding ' amino acids 201-207 in a complementary 5'-region, followed by a SpeI cloning site in a non-complementary 3'-region (Fig. 2B) (Table I). The BamHI/SpeI-digested PCR fragment was ligated to the similarly digested pDRK-C vector to generate pMSCD1+2. Two different I + II_C-expressing strains, DMSO 1585 and DMSO 1483, were generated after transforming this vector into BL-21 or MGC1030PR, respectively.

In the generation of an expression vector for I_C (pMSCD1), primers P5 and P4 were first used with template pD'DSP.4 to generate a partial holB fragment (' amino acids 1-168)(Fig. 2, A and B). The primer P4 contains the holB codons for ' amino acids 164-168 in a complementary 5'-region, followed by a SpeI cloning site in a non-complementary 3'-region (Fig. 2B)(Table I). The BamHI/SpeI-digested PCR fragment was ligated to the BamHI/SpeI-digested pDRK-C fusion vector (Fig. 1B) to generate pMSCD1, which was transformed into BL-21 to generate a strain (DMSO 1594) for expression of I_C.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Construction of plasmids that express ' with amino or carboxyl tags. PA1 is the PA1/04/03 promoter (41). All of the plasmids shown contain the -lactamase gene and lacIq. The tag contains a short biotinylation sequence, a hexahistidine sequence, and a thrombin cleavage site (41). A, construction of COOH-terminal ' fusion expression plasmid with PA1 promoter. B, construction of NH2-terminal ' fusion expression plasmid with PA1 promoter. Details are described under "Experimental Procedures."

To generate a plasmid encoding III_C, primers P7 and P2 were first used with template pD'DSP.4 to generate a partial holB fragment (' amino acids 207-334)(Fig. 2, A and B). P7 contains a BamHI cloning site, a Shine-Dalgarno sequence, a nine-nucleotide spacer, and a start codon in a non-complementary 5'-region, followed by the holB codons for ' amino acids 207-211 in a complementary 3'-region. P2 contains the codons for ' amino acids 330-334 in a complementary 5'-region, followed by a SpeI cloning site in a non-complementary 3'-region (Table I). The BamHI/SpeI-digested PCR fragment was ligated to the similarly digested pDRK-C fusion vector to generate pMSCD3. Two III_C-expressing strains, DMSO 1591 and DMSO 1484, were generated via the transformation of this plasmid into BL-21 and MGC1030PR, respectively.

Cell Growth and Induction-- Transformed strains bearing plasmids pMSCd', pMSNd', pMSND2+3, pMSCD1, pMSCD1+2, or pMSCD3 were grown at 37 °C to an optical density (A₆₀₀) of 0.8 in 6 liters of F-media plus 10% glucose and 200 µg/ml ampicillin. Overexpression was induced by the addition of isopropyl-1-thio--D-galactopyranoside (1 mM final concentration). At the point of induction, d-biotin (10 µM) and ampicillin (200 µg/ml) were added to the cell culture. After 3 h of induction, cells were harvested by centrifugation at 6,000 × g for 15 min at 4 °C and resuspended in Tris sucrose buffer (1 ml/g cells). Cells were frozen in liquid N₂ and stored at 80 °C.

Purification of the ' Fusion Proteins-- The ' fusion proteins were purified using Ni²⁺-NTA ion chelating chromatography (41). BL21 cells containing expression plasmids were treated with lysozyme (2 mg/g cells), 5 mM EDTA, 5 mM benzamidine, and 1 mM phenylmethylsulfonyl fluoride for 1 h on ice followed by a 5-min incubation at 37 °C for the cell lysis. For MGC1030PR cells containing expression plasmids pMSCD1+2 or pMSCD3, the lysis procedure was modified by decreasing the concentrations of lysozyme (1 mg/g of cells), EDTA (1 mM), and the heat treatment time at 37 °C (3 min). Lysates were centrifuged at 23,300 × g at 4 °C for 1 h to remove cellular debris and proteins were precipitated by addition of 0.226 g of ammonium sulfate to each milliliter of the supernatant. The precipitate was collected by centrifugation at 23,300 × g at 4 °C for 1 h. Protein pellets were resuspended to ~10 mg of protein/ml in Buffer A plus 1 mM imidazole, and applied to a pre-equilibrated Ni²⁺-NTA column (1 × 7 cm) at a flow rate of 0.5 column volumes/h. The column load was 30 mg of total protein/ml of resin. The column was washed step wise with 20 column volumes of Buffer A + 1 mM imidazole followed by 20 column volumes of Buffer B + 10 mM imidazole. Fusion proteins were eluted with 10 column volumes of an imidazole gradient in Buffer B. A 10-300 mM imidazole gradient was employed in the purification of three constructs ('_N, '_C, or I + II_C), and a steeper gradient of 10 to 200 mM gradient was used to effect elution of the others (I_C, III_C, or II + III_N), respectively. The peak fractions of '_N or '_C eluted at about 70-90 mM imidazole, and I + II_C eluted at about 80-90 mM imidazole. The peak fractions of I_C, III_C, or II + III_N eluted at about 40-60 mM imidazole. Individual fusion protein preparations were performed at 0-4 °C unless stated otherwise.

MonoQ Ion Exchange Column Chromatography-- Fraction III of III_C was subjected to further purification by using FPLC MonoQ HR 5/5 ion exchange column chromatography. Protein was precipitated by the addition of an equal volume of saturated ammonium sulfate solution and centrifuged at 23,300 × g at 4 °C for 1 h. The protein pellet was dissolved in 2 ml of Buffer C and desalted by using a NAP-25 desalting column (1.5 × 4.9 cm) with Buffer C. Immediately after desalting, the eluted fraction was applied to MonoQ column (1 ml) pre-equilibrated with Buffer C (4 mg of total protein/ml). III_C was eluted with 20 column volumes of a 0-500 mM NaCl gradient in Buffer C. The peak fraction of III_C eluted at a conductivity roughly equivalent to Buffer C + 150 mM NaCl.

Stokes' Radius Determination by Gel Filtration-- Gel filtration chromatography was performed using a FPLC Superose 6 HR 10/10 column equilibrated with Buffer D. Aliquots (300 µl, in Buffer D) of each ' fusion protein (20-50 µg) or each globular standard protein (50 µg) of the LMW-kit (see "Reagents and Materials") were injected onto the column individually. Elution fractions (0.4 ml/tube) were collected and assayed for total protein. The K_av was calculated using the equation K_av = (V_e  V_o)/(V_t  V_o); where V_e is the observed elution volume, V_t is the included volume, and V_o is the exclusion volume. The V_t of the Superose 6 column was 24 ml, and the V_o, determined using M13_goriDNA (~20 MDa) was 7.6 ml. The known Stokes' radii of the standard proteins were first plotted against the determined K_av values of both the standard proteins and the ' fusion proteins. The Stokes' radius of each ' fusion protein was then calculated from this plot by linear fitting with the Stokes' radii of standard proteins using Origin 5.0 program (Microcal Software, Inc., Northampton, MA). This initial graph was then re-plotted to generate Fig. 5, in which the Stokes' radii (x axis) are plotted versus the molecular weights (y axis) of the proteins.

Protein Binding Studies-- Biacore technology was used to measure protein binding. The carboxymethyl-dextran matrix of the CM5 sensor chip was activated by the NHS/EDC coupling reaction as previously described (43). The matrix was activated using a 100-µl injection of a mixture of 0.2 M EDC and 0.05 M NHS in water to maximize the conversion of the carboxyl groups of the sensor chip matrix to NHS esters. For the analysis of the biotinylated fusion proteins, streptavidin and bovine IgG were sequentially captured onto the activated matrix by injecting over the chip in 10 mM sodium acetate buffer (pH 4.5) at 0.2 and 0.1 mg/ml, respectively. Bovine IgG was used to partially block the negatively charged carboxyl groups on the sensor chip surface. For the analysis of the native ' protein, 10 µl of 0.1 mg/ml purified native ' in 20 mM MES (pH 6.0) and 5 mM DTT was injected over the NHS/EDC-activated sensor chip for protein immobilization. For antibody-' binding analysis, 10 µl of 0.1 µg/ml anti-mouse Fc in 20 mM Na(OAc)₂ was injected over the activated sensor chip for immobilization. Unreacted NHS ester groups were inactivated using 1 M ethanolamine-HCl (pH 8.5). The biotinylated ' protein or mAb was then injected over the immobilized streptavidin or anti-mouse Fc, respectively, in HBS buffer or in Buffer E + 0.005% P-20 surfactant. For kinetic analyses, 800-2,000 RU of ' fusion protein was immobilized. Binding studies of native ' or ' fusion domains to the  subunit were performed in HBS buffer + 5 mM DTT or HKGM buffer + 0.005% P-20 surfactant unless otherwise stated. Inclusion of 5 mM DTT was found to be important for maintenance of -' interaction. Antibody-antigen binding was performed in HBS buffer. A flow rate of 5 µl/min was used for kinetic analysis. Kinetic parameters were determined using the BIAevaluation 2.1 software (Biacore). Apparent stoichiometries were estimated using Equation 1,

(Eq. 1)

where RU is the measured response unit obtained at binding saturation and M is molecular weight.

Sedimentation Equilibrium Analysis for , ', and -' Complex-- Sedimentation equilibrium studies were carried out at 4 °C with an Optima XL-A analytical ultracentrifuge (Beckman Coulter, Inc., Fullerton, CA). Experiments employed 6-sector cells at rotor speeds of 14,000, 18,000, and 22,000 rpm.  and ' were sedimented at concentrations of 1, 2, and 4 µM at each rotor speed. Radial absorbance scans (280 nm) were taken at 4-h intervals. Scans used for analysis were taken when repeated scans were invariant (typically > 24 h). Sedimentation equilibrium data for the  and ' subunits were analyzed using Equation 2,

(Eq. 2)

where C_T = observed concentration; C₀ = reference concentration; M = subunit molecular weight (g/mol);  = protein partial specific volume (cm³/g);  = buffer density (g/liter);  = angular velocity (radians/sec); R = gas constant (8.314 × 10⁷ ergs/K/mol); T = temperature (K); r = radial position (cm); r₀ = reference radial position (cm). The data were fit to Equation 2 using MLAB (Civilized Software Inc., Bethesda, MD). For each subunit, all nine data sets (three speeds, three concentrations) were globally fit to Equation 2 to determine the protein partial specific volume ().

The fitted values of  were used in the determination of the -' equilibrium dissociation constant. The  and ' subunits were mixed in equimolar amounts at concentrations of 2 and 4 µM. Sedimentation equilibrium analysis was carried out using the same conditions described above. Data were analyzed using Equation 3,

(Eq. 3)

where C_' is the observed concentration; M, M_', C₀,, C₀,_', , and _' are molecular weights, reference concentrations, and protein partial specific volumes for each individual subunit; and lnK_' is the natural log of the -' association constant in absorbance units. The sedimentation equilibrium data for the ' association were fit to Equation 3 using MLAB. A total of six data sets (three speeds, two concentrations) were globally fit to Equation 3 to determine lnK_'. Because the data were analyzed using the absorbance data directly, the fitted value for lnK_' was converted to a molar association constant according to Equation 4,

(Eq. 4)

where ₂₈₀ is the molar extinction coefficient (280 nm) for the indicated subunit. Both ₂₈₀ and ₂₈₀' were determined as described (46).

Monoclonal Antibodies against Native '-- Monoclonal antibodies directed against the native ' subunits were produced in collaboration with the University of Colorado Cancer Center Tissue Culture and Monoclonal Antibody Core Facility as described (25). The monoclonal antibodies used in this study were from the anti-' cell line McHenry D fusions 229A2, 583H9, 629G10, 874B2, 1406F6, 1755D9, and 1805H11. Anti-' monoclonal antibodies were purified as described (25).

DNA Polymerase III Holoenzyme Reconstitution Assay-- Activities of ' fusion proteins were followed by their abilities to support the reconstitution of holoenzyme activity, which was determined by measuring DNA synthesis from a primed M13_Gori template (25). Isolated DNA polymerase III holoenzyme components (200 fmol of each of pol III, , , , ) and the ' domain deletion fusion protein under study (20-200 fmol) were incubated for 5 min at 30 °C in an assay reaction mixture containing 540 pmol of M13_Gori (as nucleotide), 165 units (40 ng) of DnaG primase, 1.6 µg of E. coli single-stranded DNA-binding protein plus 48 µM nucleotides (dATP, dCTP, and dGTP), and 18 µM [³H]dTTP (specific activity 150 cpm/pmol) in Buffer F. The polymerization reaction was quenched by trichloroacetic acid precipitation, then filtered through GF/C filters as described (10). One unit is defined as the amount of enzyme catalyzing the incorporation of 1 pmol of dNTPs per min at 30 °C.

Protein Determinations-- The concentrations of ' fusion domains were determined by using the Pierce Coomassie Plus Assay Reagent according to the manufacturer's instructions. BSA (fat free, Sigma) was used as a standard. The concentrations of other proteins were determined by their extinction coefficients at A₂₈₀ as described by Pritchard et al. (46).

Biotin Blots-- After separation by SDS-PAGE, proteins were transferred onto the transfer membrane and developed as described (41).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Expression and Purification of the Fusion Proteins of ' Lacking Specific Domains

Variants of ' proteins lacking specific domains were designed based on its crystal structure, which was determined by Guenther et al. (31). To facilitate purification, detection, and immobilization, these reagents were expressed as fusion proteins tagged, on either the amino or carboxyl terminus, with hexahistidine and biotinylation sequences (Fig. 2). As defined under "Experimental Procedures," and diagrammatically depicted in Fig. 2, the panel of ' fusion proteins employed in this study consisted of '_N (40.8 kDa), '_C (40.8 kDa), I_C (22.1 kDa), I + II_C (26.1 kDa), II + III_N (22.2 kDa), and III_C (18.2 kDa). The levels of separately expressed '_N, '_C, I_C, and II + III_N, were each ~1% of total cell protein as determined by densitometric scans of Coomassie-stained SDS-polyacrylamide gels (data not shown). In contrast, the expression levels of I + II_C and III_C were significantly lower (<0.1% each), and not distinguishable on Coomassie-stained gels of cell extracts. However, the biotin blot of these fusion proteins confirmed the induced signal of III_C and I + II_C, respectively (data not shown). Separately expressed ' fusion proteins lacking specific domains were soluble, and were purified by Ni²⁺-NTA metal-chelating column chromatography (Figs. 3 and 4). Biotin blots were used to evaluate individual elution fractions for the presence of the desired fusion proteins. '_N, '_C, I_C, I + II_C, II + III_N, and III_C were the only biotinylated proteins observed in the corresponding eluted fractions (Figs. 3 and 4). Thus, the biotinylated ' fusion proteins were presumed to be the only proteins captured onto the Biacore sensor chip during the immobilization step. Most of fusion proteins of ' were purified to greater than 80-95% purity after Ni²⁺-NTA chromatography (Fig. 4). III_C, which was expressed at a low level, required further purification using MonoQ ion exchange chromatography, yielding preparations of about 70-75% purity (Fig. 4, Table II). Activities of ' proteins lacking specific domains were determined by a holoenzyme reconstitution assay. Both '_N and '_C had specific activities of 5.5 × 10⁷ and 5.1 × 10⁷ units/mg, respectively, comparable to that of native ' (1.0 × 10⁸ units/mg)(Table II). However, none of the other constructs, each lacking either domain I or domain III of ' provided the holoenzyme activity (Table II), even when included at a 100-fold molar excess, or with increased incubation times (5-20 min at 30 °C). These data indicate that the deletion constructs (I_C, I + II_C, III_C, or II + III_N) are not capable of supporting enzymatic activity at levels detectable with our assay system, and suggest that the amino- and carboxyl-terminal domains of the ' subunit are required for holoenzyme activity. However, misfolded fusion proteins would also be expected to yield negative results in reconstitution assays. We used Superose 6 gel filtration chromatography to determine the Stokes' radii of I_C, I + II_C, III_C, or II + III_N as an indication of whether they were properly folded (Fig. 5). Globular proteins with known Stokes' radii as well as native ' were used as standards. Three of the fusion proteins, I_C, I + II_C, and III_C, eluted in positions near the standard proteins of similar molecular weights. A plot of the determined Stokes' radii of these three proteins versus their calculated molecular weights gave a correlation similar to that of the standard proteins (Fig. 5). These results suggest that the ' fusion proteins corresponding to domain I, domain I-II, or domain III were globular, and presumably properly folded. One of the fusion proteins, II + III_N, exhibited an unexpectedly large Stokes radius (Fig. 5). Given that III_C appeared to be properly folded (Fig. 5), it seems reasonable to presume that the aberrant Stokes' radius obtained with II + III_N was likely due to a partially unfolded domain II.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Strategy used for the construction of ' fusion proteins. A, x-ray structure of '. The schematic diagram of ' structure employed the x-ray structure-coordinates of Guenther et al. (31), and was generated using the RasMol program (RasWin Molecular Graphics, UK). -Strands are indicated by arrows, and -helices by coils. The three domains of ' are labeled. N, amino-terminal end of '. C, carboxyl-terminal end of '. Numbers shown indicate the positions of amino acid residues within wild type '. B, strategy used to construct domain deletions in NH2- or COOH-terminal fusion proteins of the ' subunit. As described under "Experimental Procedures," PCR amplification employing the primers listed in Table I was used to generate DNA fragments encoding domain I, domains I plus II, or domain III of the ' subunit. The resultant PCR products were cleaved with the relevant restriction enzymes, and ligated into expression vectors (Fig. 1). C, schematic diagram of the ' fusion proteins lacking specific domains. Black square box indicates the tripartite tag (41).

View larger version (38K): [in this window] [in a new window]   Fig. 3.   Chromatographic purification of 'C fusion protein. A, Ni2+-NTA column chromatography profiles. Purification of 'C from 1,300 mg (Fraction II) in 50 ml of total protein applied to a 7-ml column. Elution fractions (numbers 1-130; 1 ml each) were collected upon application of an imidazole gradient (10 to 300 mM) in 50 mM sodium phosphate (pH 7.4), 500 mM NaCl, 20% glycerol, 0.5 mM phenylmethylsulfonyl fluoride, and 0.5 mM benzamidine. Each fraction was assayed for polymerase activity (closed diamonds) and protein level (open circles). B, Coomassie blue-stained 15% SDS-polyacrylamide gel of the Ni2+-NTA column fractions of 'C. I, fraction II (load protein, 50 µg). II, flow-through (50 µg). III, column wash elute (50 µg). Numbers shown above the gel are elution fraction numbers (13 µl of each loaded). C, biotin blot of the SDS-polyacrylamide gel shown on B. BCCP, a biotin carboxyl carrier protein (20.7 kDa) of acetyl-CoA carboxylase (41).

View larger version (79K): [in this window] [in a new window]   Fig. 4.   Purified ' fusion proteins lacking specific domains. A, 15% SDS-polyacrylamide gel of the purified ' fusion proteins. The gel was stained with Coomassie Brilliant Blue for 3 h and destained in a solution of 10% methanol and 10% acetic acid to visualize proteins. 2-6 µg of each purified protein were loaded into each well and separated at constant voltage (150 V). B, biotin blot of the 10% SDS-polyacrylamide gel of purified ' fusion proteins lacking specific domains. After SDS-PAGE, the gel was subjected to biotin blots as described (41). Lane 1, IC (fraction III); lane 2, I + IIC (fraction III); lane 3, IIIC (fraction IV); lane 4, II + IIIN (fraction III); lane 5, 'C (fraction III); lane 6, 'N (fraction III). st, standard protein molecular weight marker.

View this table: [in this window] [in a new window]   Table II Purification of the ' fusion proteins lacking specific domains

View larger version (19K): [in this window] [in a new window]   Fig. 5.   Gel filtration analysis of the ' domain proteins. Gel filtration of native ', the ' fusion proteins lacking specific domains, and standard proteins were performed using FPLC Superose-6 column chromatography. The figure shown here is a re-plot of the initial graph to reveal the relationship between the Stokes' radius values (x axis) versus molecular weights of the proteins (y axis). Stokes' radii of ' domain proteins were compared with the values of the known globular standard proteins and native '. The initial graph was made by plotting the determined Kav values of both standard proteins and ' domain proteins obtained by gel filtration versus the known Stokes' radius values of standard proteins. Molecular weights of ' domain proteins were from their amino acids calculated by the Omega 1.1 program. RNa, bovine pancreas ribonuclease A (13.7 kDa, 16.4 Å); Chy, bovine pancreas chymotrypsinogen A (25 kDa, 20.9 Å); Ova, hen egg ovalbumin (43 kDa, 30.5 Å); BSA, 67 kDa (35.5 Å). MW, molecular weight.

Binding Analyses of Native  and '

Biochemical studies show that ' binds both  and DnaX (26). The -' interaction is necessary for recruiting  into the DnaX complex (DnaX₃'). Prior to evaluating the binding of ' fusion proteins to , we characterized the interaction of the native proteins via two techniques, surface plasmon resonance and equilibrium sedimentation. These experiments were performed for two reasons. First, the K_D values obtained for the interaction of the native proteins serve as baselines for interpretation of results obtained using the fusion proteins. Second, we wanted to evaluate whether the K_D value of -' interaction obtained using surface plasmon resonance approximated that determined using equilibrium sedimentation. Surface plasmon resonance provides a rapid means of surveying interactions and binding kinetics, but it is not a true equilibrium method. Thus, we sought an experimental confirmation of the K_D determined using surface plasmon resonance by a rigorous method, equilibrium sedimentation.

Surface Plasmon Resonance-- The interaction between native ' and  was first characterized via Biacore methodology. ' was immobilized onto a sensor chip by using the amine coupling method. Varying concentrations of  were passed over the immobilized ', and binding was monitored (Fig. 6). Data obtained using subsaturating levels of  were used to calculate the association rate constants. To preclude artifacts arising from reassociation (47), off-rates were determined after saturating ' with . Data from several experiments were combined to calculate a K_D (k_off/k_on) of 0.86 ± 0.07 µM for the binding of  and immobilized ' (Table III).

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Interaction of  and ' by surface plasmon resonance. Samples containing different concentrations of  (0.5, 1, or 2 µM) in 50 mM Hepes (pH 7.4), 100 mM potassium glutamate, 10 mM Mg(OAc)2, and 5 mM DTT, with 0.005% BIA surfactant were injected over the immobilized native ' on a CM5 sensor chip. Binding signals are shown as RU. Arrows show the injection start point and stop point of . The association rate (kon) and dissociation rate (koff) were calculated from the association and dissociation phases of the -' interaction, followed by the calculation of the equilibrium dissociation constant (KD = koff/kon). Nonlinear curve fitting for the kinetic parameter was performed by BIAevaluation software 2.1.

View this table: [in this window] [in a new window]   Table III Kinetic and equilibrium constants of -' interaction as determined by surface plasmon resonance

Equilibrium Sedimentation-- Before studying the binding of  and ' via equilibrium sedimentation, we first needed to determine the oligomerization status, and then the partial specific volume, of each of these subunits alone. Different concentrations (1, 2, and 4 µM) of native  (Fig. 7A) or ' (Fig. 7B) were evaluated in this first set of sedimentation equilibrium experiments. In each experiment, equilibrium boundary scans were taken every 4 h at rotor speeds of 14,000, 18,000, and 22,000 rpm, and the scans used for analysis were taken when repeated scans were invariant (Fig. 7A for , and Fig. 7B for '). Preliminary analyses of 9 data sets obtained for each of the  and ' subunits were fit into the model of homomeric monomer, dimer, and trimer, respectively. In both cases, the  and ' subunits behaved as ideally sedimenting monomeric species. Only the models assuming monomeric forms fit the data from the equilibrium sedimentation of  (Fig. 7A) or ' (Fig. 7B), with very low residuals distributed around the theoretical curve (<± 0.02 A₂₈₀ units in both cases). These data indicate that  and ' are monomeric when free in solution, confirming the previously reported findings based on gel filtration and other methods (26).

View larger version (25K): [in this window] [in a new window]   Fig. 7.   Interaction of  and ' by equilibrium sedimentation. Equilibrium sedimentation studies of the  and ' subunits were performed using an analytical ultracentrifuge. Representative data, curve fits, and residual plots are shown. A, sedimentation equilibrium analysis of . This profile was obtained after sedimentation of the purified  subunit (4 µM by 280) at 18,000 rpm for 40 h. B, sedimentation equilibrium analysis of '. Data for the purified ' subunit (2 µM by 280) were obtained after centrifugation at 22,000 rpm for 40 h. For both  and ', data obtained in separate experiments employing three different speeds (14,000, 18,000, or 22,000 rpm) and protein concentrations (1, 2, or 4 µM) were fit to Equation 2. C, sedimentation equilibrium of the -' complex. Representative data, curve fits, and residual plots are shown for the -' complex (2 µM) sedimented at 18,000 rpm for 40 h. Data obtained at the three speeds noted at two different concentrations of the -' complex (2 µM or 4 µM) in 50 mM Hepes (pH 7.4), 100 mM potassium glutamate, and 10 mM Mg(OAc)2, were fit to Equation 3. Residuals are expressed in absorbance units (280 nm).

It is important to note that the  value of a given protein varies depending on the degree of hydration of residues in the globular protein versus free amino acids (44). Since the value of  is normally calculated based on the amino acid composition of the protein and values for amino acid partial specific volume in water, there is greater uncertainty in this value than in the calculated monomer molecular weight (45). Therefore the sedimentation equilibrium data were then fit to Equation 2 in order to determine the partial specific volumes () of  and ' experimentally. Our determined  value for  () was 0.738 ± 0.001 cm³/g, and that for '(_') was 0.758 ± 0.001 cm³/g. These values differed from the calculated numbers by 2 to 3%; the calculated values for and ' are 0.746 and 0.741, respectively. We used the experimentally determined values of and ' to analyze the sedimentation equilibrium profile of the -' complex in order to calculate the K_D (Fig. 7C). Using Equation 4, an equilibrium K_D of 0.57 ± 0.03 µM was calculated for the interaction of  and ' under the conditions employed in this study.² The average K_D value obtained in surface plasmon resonance experiments (0.86 ± 0.07 µM) was marginally greater, but reasonably close, to this value. In both cases, the standard deviation was acceptable for our purposes, although somewhat better agreement was obtained using sedimentation equilibrium (±5.2%), than surface plasmon resonance (±8.1%). The sedimentation equilibrium data for the -' complex was fit to models assuming heterodimeric, -trimeric, and -tetrameric forms. Only the heterodimeric model fit the sedimentation equilibrium data. A lower ratio of  to ' (1 to 0.6) was calculated based on the surface plasmon resonance data (Table IV). This lower stoichiometry may have been due to the immobilization of ' to the sensor chip surface via chemical cross-linking; some of the ' molecules are likely positioned in a binding-incompetent orientation. However, despite this limitation, the close agreement between K_D values obtained in sedimentation equilibrium experiments versus Biacore-based analyses indicate that the latter is useful for comparing the relative binding of  to ' fusion proteins.

View this table: [in this window] [in a new window]   Table IV Characterizing the binding of  and ' fusion proteins lacking specific domains

Mapping the -binding Domain of '

The panel of ' proteins lacking specific domains was employed to localize the -binding domain of ' via Biacore methodology. The interaction between  and recombinant proteins corresponding to full-length ' with either an amino- or carboxyl-terminal tag was first assessed. Streptavidin was chemically coupled to a CM5 sensor chip, and '_N or '_C was immobilized via biotin-streptavidin interaction. Varying concentrations of  were injected over and bound to the immobilized '_N or '_C, and the rate constants were obtained as described under "Binding Analyses of Native ' and ." Similar K_D values were calculated for the binding of  to immobilized '_N (0.33 ± 0.04 µM) or '_C (0.42 ± 0.07 µM). These values were comparable to those obtained for the binding of native ' and  (Table IV). Thus, including the