A Single Domain of the Replication Termination Protein of Bacillus subtilis Is Involved in Arresting Both DnaB Helicase and RNA Polymerase*

The current models that have been proposed to explain the mechanism of replication termination are (i) passive arrest of a replication fork by the terminus (Ter) DNA-terminator protein complex that impedes the replication fork and the replicative helicase in a polar fashion and (ii) an active barrier model in which the Ter-terminator protein complex arrests a fork not only by DNA-protein interaction but also by mechanistically significant terminator protein-helicase interaction. Despite the existence of some evidence supporting in vitro interaction between the replication terminator protein (RTP) and DnaB helicase, there has been continuing debate in the literature questioning the validity of the protein-protein interaction model. The objective of the present work was two-fold: (i) to reexamine the question of RTP-DnaB interaction by additional techniques and different mutant forms of RTP, and (ii) to investigate if a common domain of RTP is involved in the arrest of both helicase and RNA polymerase. The results validate and confirm the RTP-DnaB interaction in vitro and suggest a critical role for this interaction in replication fork arrest. The results also show that the Tyr33 residue of RTP plays a critical role both in the arrest of helicase and RNA polymerase.

DNA replication in many prokaryotic chromosomes and at some eukaryotic chromosome regions is arrested at specific sequences called replication termini or Ter sites (6). Ter sites act as polar barriers to fork movement and act essentially as replication traps. In Bacillus subtilis, the Ter DNA interacts with a sequence-specific DNA-binding protein called replication terminator protein (RTP) 1 that acts as a polar contrahelicase; i.e. it impedes helicase-catalyzed DNA unwinding when present in one orientation, whereas it lets the helicase pass through unimpeded in the opposite orientation (1)(2)(3)(4).
The bacterial chromosome is believed to exist in vivo not as naked DNA but as a DNA-protein complex, with most of the DNA-binding proteins remaining bound to the chromosome (5). Despite the fact that some of these proteins bind to DNA with relatively high affinity (e.g. lac repressor), the replication fork apparently has the ability to pass through these complexes unimpeded. The only region of the chromosome that is known to arrest effectively the replication forks is the terminus (6). The preceding observations suggest the following: (i) the replication apparatus apparently has an activity that allows it to pass through most protein-DNA complexes, some of which contain strong DNA-binding proteins, and (ii) since the replication terminus is able to arrest forks effectively, the terminator protein-DNA complex is likely to have special features that enable it to arrest replication forks. Thus, high affinity binding of terminator protein to Ter sites per se does not appear to be sufficient to cause the replication-terminating activity of RTP (12). We have hypothesized that DnaB (or the equivalent helicase of B. subtilis)-RTP interaction plays a key role in fork arrest (1). Despite the existence of in vitro evidence for RTP-DnaB interaction (7,12), the validity of the protein-protein interaction model has been debated (11).
The raison d'etre for carrying out this work was 2-fold: (i) to perform additional experiments, using independent approaches and different mutant forms of RTP, to reexamine the question of RTP-DnaB interaction in vitro and (ii) to investigate whether a common domain of RTP is involved in the arrest of both RNA polymerase and helicase. The observations presented here confirm the biologically meaningful interaction between RTP and DnaB and further extend the result by showing that a common domain of RTP seems to be involved in the arrest of both DnaB helicase and T7 RNA polymerase (and perhaps other RNA polymerases).
The replication termini of B. subtilis ( Fig. 1) consist of overlapping core and auxiliary sites. A RTP dimer first binds to a core and then, by cooperativity, promotes the binding of a second dimer of RTP to the auxiliary site (8,9). Interaction between two dimers is essential for fork arrest with the core end of the Ter site arresting the helicase and the auxiliary end, allowing the helicase to pass through unimpeded (7).
The RTP of B. subtilis is a homodimer belonging to the class of winged helix proteins (8,9). The crystal structure of the RTP apoprotein has been solved at high resolution, and the structure contains four ␣-helices, three ␤ strands, and an unstructured, N-terminal arm ( Fig. 2A) (9). Extensive random and site-directed mutagenesis of RTP had identified the N-terminal arm, the ␣ 3 helix, and the ␤ 2 strand to be the main DNAbinding elements (8), with the ␣ 3 inserting into the major groove and the ␤ 2 into the minor groove of Ter DNA (8). Affinity cleavage analysis that converted RTP to a site-directed chemical nuclease was used to determine amino acid to base contacts, and the results had confirmed that the ␣ 3 helix contacted the major groove and ␤ 2 , to the minor groove of Ter DNA (10, 26). X-ray crystallography had revealed an exposed hydropho-bic patch that was suggested to be a possible docking surface for the helicase (9) (see Fig. 2B).
In this paper, we have used cross-linking, label transfer, and other techniques along with different mutant forms of RTP to present additional evidence in favor of the RTP-DnaB interaction model of replication termination. We also show that a common region of RTP is involved in arresting both the helicase and T7 RNA polymerase. Both helicase and RNA polymerases melt DNA, and the results suggest that there probably is a common structural motif in these enzymes that may be recognized by RTP.
Gel Mobility Shift Assay-The assays were carried out as described in Ref. 13. Briefly, polymerase chain reaction products of Ter 1 and core binding sites were amplified from pUC18BS3 (53-base pair fragment having complete RTP binding site cloned as EcoRI-HindIII cassette) and pUC18core (fragment having the core binding site of RTP cloned as EcoRI-HindIII cassette) using universal M13 forward and reverse primers. These fragments were end-labeled with [␥-32 P]ATP and T4 polynucleotide kinase. 1 fmol of labeled DNA was used in each reaction, which was carried out in 20-l volumes of 40 mM Tris-Cl (pH 7.5), 4 mM MgCl 2 , 50 g/ml bovine serum albumin, 50 mM potassium glutamate, 5 g of calf thymus DNA, and increasing amounts of RTP (0, 1,2,3,4,6,8,12,16,20,40,60,80,100,200, 300, 400, 600, 800, and 1200 fmol). The reactions were carried out at room temperature for 30 min and resolved on 8% native polyacrylamide gels.
Radioactivity from gel bands corresponding to free, singly and doubly shifted 32 P-labeled IR1 DNA was quantified by a PhosphorImager (Molecular Dynamics, Sunnyvale, CA). From these data, the fraction of IR1 oligonucleotide bound (FB) by two dimers of RTP protein was calculated and plotted as a function of the free RTP concentration (see Fig. 7). The data were then replotted as the log [FB/(1 -FB)] versus log [free RTP] (Fig. 7), based upon a logarithmic form of the Hill equation (Equation 1) (21).
where H is the Hill coefficient and K 0.5 is a constant equal to the concentration of RTP at FB ϭ 0.5. Plotted in this format, the line slope for points between 0.1 Ͻ FB Ͻ 0.9 is the Hill coefficient. At FB ϭ 0.5, the following is true, allowing direct calculation of the concentration of RTP required for half-maximal IR1 binding. Hill Analyses of Wild Type, Y33F, and Y33N Replication Terminator Proteins-Listed are the Hill coefficients and the corresponding R 2 values derived for assessment of IR1-binding cooperativity by wild type, Y33F, and Y33N replication terminator proteins. Also listed are the values for log K 0.5H (defined here) and the concentrations of each RTP form required for half-maximal saturation of IR1 DNA (Table I).
Purification of DnaB and RTP-The proteins were purified as described (7, 10).  Helicase Assay-Helicase assays were performed as in Ref. 1. Briefly, 10 pmol of complementary oligonucleotide with RTP binding sites were end-labeled using [␥-32 P]ATP and T4 polynucleotide kinase and annealed to 1 pmol of M13mp18 BS3REV (RTP binding site cloned in nonblocking orientation as EcoRI-HindIII fragment) or M13 mp19BS3 (RTP binding site cloned in blocking orientation as EcoRI-HindIII fragment) single-stranded DNA, respectively. These helicase substrates were purified through a CL4B Sepharose spin column. 10 fmol of substrate were used in each reaction of 20-l volume containing 40 mM Tris-Cl (pH 7.5), 4 mM MgCl 2 , 50 g/ml bovine serum albumin, 2 mM ATP, 50 mM potassium glutamate, 5 mM dithiothreitol, and increasing amounts of RTP (0, 100, 200, 300, 400, 600, and 800 fmol). After 15 min of incubation at room temperature, 100 ng of DnaB helicase was added to the reaction mixture and incubation was continued for 30 min at 37°C. The reaction was stopped with SDS-EDTA-bromphenol blue dye and resolved on 8% native polyacrylamide gels.
In Vitro Replication Assay-In vitro replication assays were carried out according to Ref. 3 with some modifications. Briefly, cell extracts were prepared from E. coli W3110 strain. 1.5 g of pUC18 BS3 rev (RTP binding site cloned in nonblocking orientation) and pUC19 BS4 (RTP binding site cloned in blocking orientation; Ref. 15) plasmid DNAs were used in 50-l reactions of 40 mM HEPES (pH 8.0), 40 mM potassium glutamate, 10 mM magnesium acetate, 0.5 mM each NTP, 20 M each dNTP, 10 mM NAD, 5% polyethylene glycol, 15 mM creatine phosphate, 1 unit of phosphocreatine kinase, and 3 l (ϳ300 g of protein) of cell extract, and a 0 -10-fold excess of exogenous RTP was used in the reaction. RTP was allowed to bind to its recognition site at room temperature for 15 min, and then cell extract was added to initiate replication, and the reaction continued at 30°C for 90 min. The reaction was stopped by adding EDTA to 5 mM and SDS to 0.1%. RNase digestion (2 mg/ml) was carried out at 37°C for 15 min followed by proteinase K digestion (0.4 mg/ml 37°C for 15 min), phenol extraction, and ethanol precipitation of replication products. The replication products were analyzed on 6% denaturing polyacrylamide gels, and an end-labeled 100-base pair ladder was used as a molecular weight marker.
Azidophenacyl Bromide Cross-linking-The azidophenacyl moiety was coupled to mutant and wild type RTP proteins according to published methods (14). Y33C and E56C mutant forms of RTP were generated by using the QuikChange site-directed mutagenesis kit (Stratagene), and proteins were purified according to the procedures described earlier (16). 100 g of protein was used for coupling reaction in a 500-l reaction volume containing 300 mM azidophenacyl bromide, 20 mM Tris-Cl, pH 8.0, 200 mM KCl, 0.1 mM EDTA, 5% glycerol, and 1.7% dimethyl formamide. The reaction mixture was incubated at room temperature for 3 h followed by 15 h at 4°C. Unreacted reagent was removed by dialyzing the reaction mixture against 20 mM Tris-Cl, pH 8.0, 0.1 mM EDTA, 50 mM KCl, and 10% glycerol for 3 h at 4°C. UV cross-linking of derivatized proteins with end-labeled DNA fragment having Ter 1 RTP binding site were carried out in a 50-l reaction  6 -9). Note that only derivatized E56C yielded cross-linked DNA-protein (arrows; C). Little or no cross-linked products were seen when either the wild type or Y33C RTPs were used (B and D).

FIG. 4. Binding of radiolabeled DnaB to wild type RTP-GST and various mutant forms of RTP-GST matrices.
A, autoradiogram of a SDS-polyacrylamide gel showing the binding of increasing amounts of radiolabeled DnaB to wild type RTP-GST, Y33N-GST, and Y33F-GST affinity matrices. B, quantitation of the radioactivity from three sets of gels (values averaged) using a PhosphorImager. Note that all of the mutational substitutions at Tyr 33 cause reduction in affinity for DnaB.
volumes containing a 5-, 10-, 25-, or 50-fold excess of derivatized proteins (wild type RTP, E56C, and Y33C) over 100 fmol of end-labeled DNA, 10 mM MOPS-NaOH, pH 7.3, 200 mM NaCl, and 50 g of bovine serum albumin. The reaction mix was incubated in the dark at room temperature for 45 min. UV irradiation to samples was carried out using the long wavelength range of the UV lamp (model UVGL-25; Ultraviolet Products, San Gabriel, CA) held at a distance of 10 cm for 2 min. The UV-cross-linked samples were resolved on 10% SDS-polyacrylamide gels.
Protein-Protein Interaction Studies by Glutathione S-Transferase (GST) Affinity Column Chromatography-GST affinity chromatography assays were performed according to Refs. 7, 16, and 24. After a 10-min incubation at room temperature, the labeled 125 I-ACT was used directly to react with RTP protein. 60 l of 125 I-ACT (6.7 nmol) from the iodination quench vial were mixed with 10 mol of RTP in 40 l and incubated for 20 min. The reaction mixture was loaded on 2 ml of Bio-Gel column (Bio-Rad) equilibrated previously in 40 mM Tris-Cl, pH 8.0, 100 mM KCl, 1 mM EDTA, 10 mM MgCl 2 , 5% glycerol (AT buffer) and eluted in the same buffer. 100-l fractions containing derivatized 125 I-RTP were pooled and used for photocross-linking. 80 l of 125 I-RTP, 160 ml of 200 nM DnaB helicase, and 560 l of AT buffer were mixed and UV-irradiated using a Rayonet photochemical reactor, model RPR-100, for 20 s at room temperature in the dark. 5.9 mg of iodoacetamide were added for alkylation and reacted for 15 min at room temperature in the dark. The reaction mix was concentrated at 2000 ϫ g to 20 l using Ultrafree-CL centrifugal filters (Millipore Corp.). 20 l of 80 mM iodoacetamide in 15 M urea were added to concentrated sample and incubated for 15 min at room temperature in the dark. Two aliquots of concentrated samples (20 l each) were prepared. To the first, 5 ml of 5ϫ loading dye (0.5 M Tris-Cl, pH 6.8, 50% glycerol, 10% SDS, and 0.1 mg/ml bromphenol blue) was added, and to the second sample, loading dye and 10% ␤-mercaptoethanol was added, and the samples were resolved on 30-cm-long 10% SDS-polyacrylamide gels.

RESULTS
Despite the existence of some evidence for RTP-DnaB interaction in vitro (7) and in vivo (12), there has been continuing debate (11) as to whether RTP-DnaB helicase is really involved in replication fork arrest. This debate prompted us to reexamine this significant mechanistic question by using different approaches and additional mutants, including a conservative Y33F mutant form of RTP. Tyr 33 was chosen for further investigation because previous work had implicated this as a critical residue for helicase arrest (7). We also wished to investigate whether the domain of RTP that is known to be involved in helicase arrest also impedes the elongation of RNA polymerase. First we wished to determine if the Tyr 33 residue of RTP was in contact with Ter DNA in solution as described below.
The Residue Tyr 33 Does Not Contact DNA in an RTP-Ter DNA Complex-The exposed hydrophobic patch of RTP (Fig.  2B) was postulated to be a possible docking surface for DnaB helicase (9), and the Tyr 33 residue, located near the patch, was found to be a critical residue in the RTP-DnaB interaction in previous experiments (18). Since the crystal structure of an RTP-Ter complex is not yet available, we wished to investigate whether the Tyr 33 residue directly contacted Ter DNA in solution. We isolated the Y33C mutation by site-directed mutagenesis and coupled the photodynamic reagent azidophenacyl bromide to wild type RTP (which has a Cys residue in the dimerization domain, at coordinate 110), Y33C and E56C mutant forms of RTP. E56C was chosen for derivatization (Fig. 3A) because the residue is located in the ␣ 3 helix of RTP ( Fig. 2A) FIG. 5. Cross-linking-label transfer using wild type and mutant forms of RTP. A, reaction scheme showing the coupling of radioiodinated ACT to RTP having a cysteine at position 30 (E30C mutant) and photocross-linking to DnaB followed by cleavage of the S-S bond with ␤-mercaptoethanol. B, autoradiogram showing failure to derivatize wild type RTP but successful derivatization of Y33C protein. Lanes 1-4 show derivatized RTP (out of the gel because of low molecular size); derivatized RTP ϩ DnaB, no UV; derivatized RTP ϩ DnaB ϩ UV ϩ ␤-mercaptoethanol; and derivatized RTP ϩ DnaB ϩ UV, no ␤-mercaptoethanol. Note that E30C cross-linked to RTP to yield a labeled product of higher molecular mass than DnaB (lane D, arrow) and upon cleavage with ␤-mercaptoethanol; label transfer occurred to a protein that had a mobility identical to that of unlabeled, kinase-tagged, highly purified DnaB (arrow, lane 3; stained gel not shown). C, autoradiogram showing comparative cross-linking and label transfer of the double E30C,Y33N and the single E30C mutant forms of RTP. Lanes 1-3 show derivatized RTP ϩ DnaB, no UV; derivatized RTP ϩ DnaB ϩ UV ϩ ␤-mercaptoethanol; and derivatized RTP ϩ DnaB ϩ UV, no ␤-mercaptoethanol. The gels were run for a shorter time to get sharper bands without diffusion. The shorter gel run did resolve the small mobility difference between DnaB-RTP complex and the cleaved labeled DnaB (lanes 2 and 3 of E30C; arrow). Note that the Y33N mutation in the double mutant form abolished the cross-linking of RTP to DnaB (lanes 2 and 3 of E30C,Y33N). and is known to contact the major groove of Ter DNA (10,26). This derivative was therefore used as a positive control. Neither derivatized wild type RTP nor the Y33C form yielded significant amounts of cross-linked product with labeled Ter DNA (Fig. 3, B and D). In contrast, E56C protein readily yielded detectable amounts of protein-DNA cross-links, and two cross-linked species, corresponding to DNA bound to a single dimer or with two dimers of RTP, were readily detected (Fig. 3C, arrows). Thus, on the basis of the result presented here, we concluded that the Tyr 33 residue neither contacted DNA in solution nor was located within 11 Å of the Ter DNA (the length of the cross-linker is 11 Å; see Ref. 19).
Interaction between RTP and DnaB as Revealed by Affinity Column Chromatography-For the purpose of critically reevaluating the interaction between RTP and DnaB further, we tagged DnaB with a kinase tag at the N-terminal end and purified the protein to near homogeneity and labeled it with [␣-32 P]ATP and muscle kinase (20). Previous work had used DnaB synthesized by coupled transcription-translation in vitro (7). We used authentic DnaB (kinase-tagged), produced in vivo that had full helicase activity (data not shown) in order to eliminate any chance of binding artifacts caused by misfolded or inactive DnaB produced in vitro. Wild type RTP and the various mutant forms were produced as fusion proteins with GST and immobilized separately onto glutathione-agarose beads. We made sure that equal molar amounts of wild type and each of the mutant forms of RTP were immobilized to a fixed amount of the affinity matrix by removing equal aliquots of each type of RTP affinity beads and estimating the amount of bound proteins by SDS-polyacrylamide gel electrophoresis.
We cleaved off the GST moiety from the wild type fusion protein and performed helicase-blocking assays to make sure that the protein was biologically active. It should be kept in mind that the Tyr 33 residue projects out of the surface of RTP apoprotein (Fig. 2B), and mutational alteration at this site does not seem to affect the folding of the protein as suggested by its chromatographic behavior, solubility properties that were indistinguishable from that of the wild type protein. In fact, we were able to crystallize the mutant forms and collected a partial set of diffraction data but did not attempt to solve the structure.
Increasing amounts of labeled DnaB protein were bound to equivalent amounts of the wild type and mutant forms of the RTP-affinity beads and washed, the labeled protein was stepeluted with increasing concentrations of NaCl, precipitated, resolved in SDS-polyacrylamide gels, and autoradiographed. The amount of bound proteins was quantitated with a Phos-phorImager. The results showed that the wild type RTP matrix retained the labeled DnaB. Considering the fact that RTP is a basic protein, whereas DnaB is acid, the possibility of nonspecific acidic and basic protein interaction was taken into account and eliminated by the following experiment (Fig. 4). We kinasetagged and labeled another acidic protein (namely DnaG) and performed the affinity adsorption experiments with wild type RTP affinity beads. We did not observe any retention of DnaG by RTP-GST affinity matrix (not shown).
Nonspecific charge interaction causing binding artifacts is also not supported by our observation that replacement of an uncharged Tyr 33 residue by another uncharged F or A caused no alteration of the net charge but significantly reduced the protein-protein interaction. In fact, replacement of Tyr 33 with Asn, Ala, Cys, or Phe resulted in a reduction in protein-protein interaction (Fig. 3, A and B). The outcomes of the affinity binding experiments were consistent with the conclusion that there was specific protein-protein interaction between RTP and DnaB in solution and in the absence of Ter DNA and that the Tyr 33 residue played a key role in that interaction.
Cross-linking Label Transfer Experiment Supported a Critical Contribution of Tyr 33 to the Interaction-We also wished to reexamine possible interaction of DnaB to RTP by an independent method. We radioiodinated the bifunctional, photodynamic, azido cross-linker ACT (21) at position 3 to generate the 3 125 Iodo-ACT and coupled it to E30C RTP at the cysteine residue (Figs. 2B and 5A). This location was chosen because the crystal structure showed that both residues 30 and 33 were located close to each other on the protein surface and that both residues projected out at similar angles from the protein surface. Furthermore, derivatization at residue 30 only marginally reduced RTP-DnaB interaction (7). Similarly, wild type RTP (having a cysteine at position 110 in the dimerization domain) and the E30C,Y33N double mutant form of RTP were also derivatized in the dark with radioiodinated ACT. The rationale of the experimental approach is depicted in Fig. 5A. If RTP interacted with DnaB in vitro, the derivatized E30C protein should photocross-link to DnaB. Upon cleavage of the S-S bond by incubation with ␤-mercaptoethanol, the 125 I label should transfer to DnaB. The wild type RTP should neither cross-link nor transfer label to DnaB because the naturally occurring cysteine at position 110 is buried in the dimerization domain in the ␣ 4 helix (Fig. 2A). If the Tyr 33 residue contributed a critical contact with DnaB as suggested by the affinity adsorption experiments described above, then the double mutant form Y33N, E30C, derivatized with labeled ACT at the Cys 30 residue, should neither cross-link nor transfer label to DnaB.
The results of the experiment, in the absence of Ter DNA, are shown in Fig. 5, B and C. The wild type RTP, after incubation with the bifunctional cross-linking agent, showed no labeling of the protein, presumably because the naturally occurring cysteine in the dimerization domain was solvent-inaccessible (Fig.   5B, wild type, lanes 1-4). We incubated derivatized E30C with purified DnaB and irradiated the reaction mixture with UV to activate the azido group. The derivatized 125 I-labeled RTP, cross-linked to DnaB, generated a complex with a characteristic mobility in a SDS-8% polyacrylamide gel (Fig. 5, E30C, lane  4). Treatment of the reaction mixture with ␤-mercaptoethanol caused cleavage at the S-S bond and the 125 I label got transferred from the RTP to DnaB as indicated by a band of lower mobility that co-migrated with purified DnaB in the SDS gel ( Fig. 5B; E30C, lane 3). If the reaction mixture was not irradiated with UV, no label transfer to DnaB occurred (Fig. 5B,  E30C, lane 2). It should be noted that the difference in mobility between the DnaB-RTP complex is rather small, and thus it required a long run in an SDS gel to resolve the difference. A shorter run of the gel produced sharper bands but caused the RTP-DnaB complex to run at a position that was indistinguishable from that of DnaB (Fig. 5C, E30C, lanes 2 and 3).
We wished to determine whether the interaction between the two proteins was specific and if the Tyr 33 residue played a critical role in the interaction. The experimental results showed that when the Y33N,E30C double mutant form of RTP was derivatized by the radioiodinated ICP no label appeared in a DnaB-RTP complex or in DnaB (Fig. 5C, E30C,Y33N, lanes 2  and 3). The single mutant form E30C under identical conditions continued to show label transfer (Fig. 5C, lanes 2 and 3). Thus, the cross-linking, label transfer experiments were consistent with the conclusions derived from the affinity binding data described above that RTP interacted in vitro with DnaB even in the absence of Ter DNA.
The cross-linking in vitro between RTP and DnaB was also carried out in the presence of Ter DNA, and the results were identical to that obtained in the absence of DNA excepting that the amount of label transfer was considerably less (not shown). We wished to determine further the peptide of DnaB that was the recipient of the transferred label but were unable to do so due to the small yield of labeled material.  were performed using 32 P-labeled IR1 (Ter 1) DNA. The results from these experiments are shown in Figs. 6 and 7, plotted as the fraction of doubly occupied IR1 DNA as a function of the logarithm of the free RTP concentration. All three forms of the RTP bound to IR1 DNA with high affinities, with Y33N showing some reduction in affinity (less than 10-fold). The concentrations of free RTP for half-maximal saturation (i.e. two dimers of RTP) of IR1 were 1.9, 1.6, and 0.52 nM for the wild type, Y33F, and Y33N RTP forms, respectively. IR1 binding by the wild type and Y33F RTP forms was highly cooperative, as can be seen from the rapid rise in IR1 fraction bound with increasing RTP concentration and as shown in the logarithmic Hill plots (Fig. 7; see insets). The Hill coefficients for IR1 binding by wild type and Y33F RTPs were 1.8 and 1.9, respectively. In contrast, IR1 binding by the Y33N RTP mutant, although of high affinity, was less cooperative (Hill coefficient ϭ 0.7; see Table I). Thus, the results showed that while the Y33F mutation either did not affect or only minimally affected the cooperativity and binding affinity of the mutant form of the protein to Ter DNA, it significantly reduced RTP-DnaB interaction in vitro. We have examined the off rate of Y33F protein in comparison with that of wild type RTP and have observed that the mutant form had a 3-4-fold higher off rate than that of the wild type protein (data not shown). The Y33F Mutation Abrogated Helicase-arresting Activity in Vitro-Helicase assays were performed to analyze the ability of the wild type and the Tyr 33 mutant form of RTP to block DnaB helicase activity. Single-stranded DNAs from two M13 clones containing blocking and nonblocking orientations of Ter 1 were hybridized to respective radiolabeled complementary oligonucleotides to generate substrate DNAs that contained the Ter site at the duplex region. The ability of DnaB helicase to melt and dislodge the oligonucleotide from the partial duplexes in the presence of increasing concentrations of wild type and the Y33F mutant form of RTP was analyzed on 8% nondenaturing polyacrylamide gels. Wild type RTP, as expected, was able to arrest DnaB in a polar fashion, whereas the Y33F mutant was almost completely defective in this function (Fig. 8). This result showed that position 33 is involved in interaction with DnaB and, as a consequence, also in helicase arrest in vitro. The template that has a Col E1 ori and the Ter 1 site of B. subtilis in either orientation was replicated in the presence of RTP and resolved in a denaturing gel. The appearance of a band corresponding to the leading strand, of characteristic length (610 nt), that extends from the ori to the Ter was taken as positive evidence for replication fork arrest. B, autoradiogram showing replication arrest by wild type RTP in the blocking orientation of Ter 1 (arrow) but lack of arrest by Y33F and Y33N forms of RTP on the same template. The template containing the nonblocking orientation of Ter 1, as expected, did not give an arrested product.  lanes 1 and 2), transcription without added RTP. Bottom, control experiments using the nonblocking orientation of the Ter site (pET22b-IR1 rev) showed no detectable arrest of the RNA chain elongation in the presence of RTP. An increasing range of 0.35, 0.7, and 1.4 fmol of the wild type and each of the mutant forms of the protein were used in the experiments (e.g. in lanes 3-5, 6 -8, 9 -11, and 12-14).
Y33F Mutation Abrogated Replication Arrest in Vitro-In vitro replication assays with E. coli cell extracts were carried out with template DNA having the Ter 1 binding site in both orientations with respect to the ori. Wild type RTP and the Y33F,Y33N RTP mutant forms were tested for their ability to arrest the replication fork movement in vitro. The replication products were analyzed on 6% denaturing gels. The generation of a 610-nt-long (distance from ori to Ter) leading strand of replication intermediate was diagnostic of fork arrest by RTP. The blocking orientation of Ter 1 showed the arrested replication intermediate band in the presence of wild type RTP, but the band was not visible in the nonblocking orientation of Ter (Fig. 9B, Wt. RTP BLK and NBLK, respectively). The Y33F and Y33N RTP mutants were unable to arrest replication forks, showing an almost complete loss of this function (Fig. 9B, Y33F and Y33N). Thus, even the tyrosine to phenylalanine, conservative mutation at the residue 33 of RTP resulted in almost a complete loss of replication fork arresting activity.
Mutations in the Tyr 33 Residue Abrogate Arrest of Transcription Elongation-Previous work had shown that RTP arrested RNA polymerases (T7 and E. coli RNA polymerases) in a polar mode. We wished to investigate whether a common region of RTP was involved in the arrest of both helicases and RNA polymerases. We set up transcription reactions using a template that had an upstream T7 promoter and a downstream Ter site (present in either orientation with respect to the promoter). Transcription initiated from the upstream promoter was arrested in the blocking orientation of the Ter site by wild type RTP. All mutant forms of RTP, Y33A, Y33C, and Y33F, were unable to arrest T7 RNA polymerase in the blocking orientation of the Ter site (Fig. 10, top). Templates having the Ter site in the opposite orientation failed, as expected, to arrest transcription elongation (Fig. 10, bottom).
Thus, it appears that the Tyr 33 residue is a key element not only in arresting helicase-catalyzed DNA unwinding but also in impeding RNA polymerase-catalyzed chain elongation. DISCUSSION Although the ability of replication terminator proteins to arrest replicative helicases (contrahelicase activity) and RNA polymerase has been known for some time (1, 2, 4) and we have published in vitro evidence for RTP-DnaB interaction (6), the details of the mechanism by which RTP arrests replication fork and helicases needed additional investigation. The reinvestigation of the RTP-DnaB interaction by independent methods and different mutants was also prompted in part by continuing debate in the literature as to whether RTP-DnaB interaction was involved in replication fork arrest (11). Another group had claimed that the mutations in the Tyr 33 residue had caused a 100-fold reduction in the DNA binding affinity and that this defect in DNA-protein interaction rather than the observed reduction in RTP-DnaB interaction was responsible for failure to arrest replication forks.
Our present results show that the conservative Y33F mutations elicited strong and cooperative binding of RTP to DNA as shown by the Hill plot. The Y33N mutations did show no more than a 10-fold reduction in affinity for DNA. Can a reduction of this magnitude (even up to 10-fold) of RTP-DNA interaction completely abolish the helicase-arresting activity of RTP? The answer appears to be in the negative from another consideration. Instead of mutating RTP, one could in principle, mutate the Ter site to significantly reduce its affinity for RTP and then ask whether such a variant Ter site is still capable of arresting forks in vivo. We have performed this type of experiment by examining the mechanism of fork arrest at the L1, checkpoint Ter site of B. subtilis. The naturally occurring L1 site binds to DNA with a half-life of the DNA-protein complex of just a few seconds, whereas the normal Ter 1 site-RTP complex has a relative half-life of ϳ180 min under identical conditions. Despite this striking difference in the stability of the two DNA protein complexes, we observed that the L1-Ter was able to arrest replication forks in vivo, suggesting thereby that strong RTP-Ter interaction was not obligatory for fork arrest (12).
It should be noted that almost all of the evidence for RTP-DnaB interaction presented here and elsewhere, is of an in vitro nature (7). To derive in vivo evidence, we have attempted to perform yeast two-hybrid analysis using RTP and DnaB but have been unsuccessful due to lack of expression of RTP in yeast. However, we have been successful in showing the in vivo interaction between Tus terminator protein and DnaB of E. coli using the two-hybrid system. Furthermore, using a reverse two-hybrid analysis, we have isolated mutant forms of Tus that bind to Ter normally but are defective in interaction with DnaB and in helicase arrest. 2 Although the Tus protein and RTP have significantly different crystal structures, the biochemical attributes of both proteins are similar, suggesting that they function by similar mechanisms (4,6,9). The results from both B. subtilis and E. coli systems, taken together, strongly support a mechanism of replication termination that involves not only Ter-terminator protein interaction but also mechanistically significant terminator protein-helicase interaction.
Recent structure-function analysis of the PcrA helicase supports an "inchworm" model of helicase action that involves ATP-dependent DNA melting and helicase translocation on single-stranded DNA (19). Future work should be able to settle whether terminator proteins block both of these steps by contacting the helicase.
Finally, the failure of mutant forms of RTP with amino acid substitutions at Tyr 33 to arrest helicase and RNA polymerase would suggest a common inhibitory surface on RTP. It would be interesting in the future to isolate mutant forms of T7 RNA polymerase that escape arrest by RTP. Such mutant forms of T7 RNA polymerase, considered along with the crystal structure of the enzyme (25), should be a productive avenue for further investigations and should shed light on the mechanism that causes arrest of RNA chain elongation.