Global Conformational Transitions in Escherichia coli Primary Replicative Helicase DnaB Protein Induced by ATP, ADP, and Single-stranded DNA Binding MULTIPLE CONFORMATIONAL STATES OF THE HELICASE HEXAMER*

The direct evidence of dramatic conformational changes of the DnaB hexamer, induced by nucleotide binding, and the presence of multiple conformational states of the enzyme have been obtained by using analytical sedimentation equilibrium, sedimentation veloc- ity studies, and the rigorous fluorescence titration tech-nique. Equilibrium sedimentation measurements show that in the presence of the ATP nonhydrolyzable analog, AMP-PNP, the DnaB helicase fully preserves its hexa- meric structure. However, in the presence of the saturating concentration of AMP-PNP, the sedimentation coefficient of the hexamer is s 20, w (cid:53) 11.9 (cid:54) 0.2 compared to the sedimentation coefficient s 20, w (cid:53) 10.5 (cid:54) 0.2 of the free DnaB helicase hexamer. This large sedimentation coefficient change indicates dramatic global conforma- tional transitions of the hexamer, encompassing all six subunits , upon binding the ATP analog. In the presence of ADP, the sedimentation coefficient is s 20, w (cid:53) 11.4 (cid:54) 0.2, indicating that the conformation of the ADP form of the hexamer is different from the ATP form. The sedimentation coefficient of the ternary complex DnaB-(AMP- PNP)-d (cid:101) A(p (cid:101) A) 19 , s 20, w (cid:53) 12.4, suggests that the DnaB helicase undergoes further conformational changes upon binding single-stranded DNA (ssDNA). The large global structural changes correlate with the functional

X174 phage replication (Wickner et al., 1973;McMacken et al., 1977). DnaB protein is the primary E. coli replicative helicase, i.e. the factor responsible for unwinding the DNA duplex in front of the replication fork. It is the only helicase required to reconstitute DNA replication in vitro from the chromosomal origin of replication (oriC) (Kornberg and Baker, 1992;LeBowitz and McMacken, 1986;Baker et al., 1987). The enzyme is involved in both the initiation and elongation stages of DNA replication (Matson and Kaiser-Rogers, 1990;Kornberg and Baker, 1992;Marians, 1992).
The native DnaB helicase forms a hexamer composed of six identical subunits (Reha-Krantz and Hurwitz, 1978;Bujalowski et al., 1994). Analytical sedimentation studies have shown that the DnaB helicase exists as a stable hexamer over a large protein concentration range with magnesium ions playing a crucial structural role in stabilizing the hexameric structure of the helicase. Hydrodynamic data indicate that six protomers aggregate with cyclic symmetry in which the protomerprotomer contacts are limited to only two neighboring subunits (Bujalowski et al., 1994).
Physiological functions of the DnaB helicase are related to the ability of the protein to interact with ss-and dsDNA 1 under the control of ATP binding and hydrolysis Kornberg, 1981a, 1981b;LeBowitz and McMacken, 1986). Studies of nucleotide binding to the DnaB helicase have established that the hexamer has six nucleotide binding sites, presumably one on each protomer (Arai and Kornberg, 1981b;Bujalowski and Klonowska, 1993, 1994a, 1994b. The binding process is biphasic, resulting from the negative cooperative interactions limited to neighboring subunits (Bujalowski and Klonowska, 1993, 1994a, 1994b. Interactions of the DnaB helicase with ssDNA and the structure of the formed complexes have only recently been quantitatively studied (Bujalowski and Jezewska, 1995). On the basis of thermodynamically rigorous fluorescence titrations, we have established that in the presence of the ATP nonhydrolyzable analog, AMP-PNP, stoichiometry of the DnaB hexamer complex with the polymer ssDNA (site size) is 20 Ϯ 3 nucleotides. Binding studies performed with ssDNA oligomers have shown that the hexamer has only a single, strong ssDNA binding site. Moreover, photo-cross-linking experiments indicate that only a limited set of subunits, most probably only one, is engaged in the complex with the nucleic acid. These results indicate that long-range allosteric interactions occur on the level of the quaternary structure of the hexameric enzyme, leading to the selection of a limited set of subunits as a binding site for ssDNA. Such interactions require significant conformational changes of the hexamer, beyond the nearest neighboring protomers.
In this communication, we present the first direct evidence that, in the presence of the ATP nonhydrolyzable analog, AMP-PNP, the DnaB helicase undergoes global conformational transitions which encompass all six subunits of the hexamer. The sedimentation coefficient of the hexamer increases by ϳ14% when compared to the free hexamer. Binding of ADP to the DnaB helicase causes similar, albeit smaller, global conformational changes. Thus, the nucleotide binding and the ATP/ADP switch, which takes place after every ATP hydrolysis step, are accompanied by significant conformational changes of the entire DnaB hexamer. We report the first quantitative analysis of the allosteric effect of nucleotide cofactors on the DnaB helicase-ssDNA complex formation.
Sedimentation Velocity Measurements-Equilibrium sedimentation and sedimentation velocity experiments were performed using a Spinco Model E analytical centrifuge, and the analysis of the sedimentation runs was performed, as we described previously (Bujalowski et al., 1994). The total concentration at radial position r is defined by Equation 1 (Cantor and Schimmel, 1980) where c b , , and M are the concentration (absorption) at the bottom of the cell, partial specific volume, and molecular weight of the protein, respectively, is the density of the solution, is the angular velocity, and b is the base line error term. Equilibrium sedimentation profiles were fitted to Equation 1 with M and b as fitting parameters. The reported values of the sedimentation coefficients were corrected to the standard conditions, s 20,w , for solvent density and viscosity (Cantor and Schimmel, 1980).
Analysis of the DnaB Hexamer-Polymer ssDNA Binding Isotherm-McGhee and von Hippel (1974) derived two explicit equations for noncooperative and cooperative binding of a large ligand to a one-dimensional, homogeneous lattice, with overlapping potential binding sites. Previously, we obtained a single generalized equation for the McGheevon Hippel model which can be applied to both cooperative and noncooperative binding (Bujalowski et al., 1989). The Scatchard form of the generalized equation is described by where K is the intrinsic binding constant, n is the number of nucleotides covered by the protein in the complex (site size), is the parameter characterizing cooperativity, and R ϭ {[1 Ϫ (n ϩ 1)] 2 ϩ 4(1 Ϫ n)} 0.5 . Binding of the DnaB helicase to the polymer ssDNA has been analyzed using Equation 2.

RESULTS
Allosteric Effect of the Nucleotide Cofactors on the DnaB Helicase Binding to ssDNA-Fluorescence titrations of poly(d⑀A) with the DnaB helicase ( ex ϭ 325 nm, em ϭ 410 nm) in buffer T2 (pH 8.1, 10°C) containing 50 mM NaCl, in the presence of 1 mM AMP-PNP, 1 mM ADP, and in the absence of nucleotide cofactors, are shown in Fig. 1. It is evident that the high affinity of the helicase toward ssDNA is observed only in the presence of AMP-PNP. In the absence of any nucleotide cofactors, the affinity is very low, and only an upper limit of the intrinsic binding constant can be estimated. In the presence of ADP, the intrinsic affinity is higher, but still very low, when compared to the affinity manifested in the presence of AMP-PNP. The solid lines in Fig. 1 are computer fits using the McGhee-von Hippel model, as described by the generalized Equation 2 (see "Experimental Procedures"). The obtained intrinsic binding constants are 1.2 Ϯ 0.2 ϫ 10 5 M Ϫ1 , 5.9 Ϯ 0.3 ϫ 10 2 M Ϫ1 , and ϳ6 ϫ 10 1 M Ϫ1 for binding in the presence of AMP-PNP, ADP, and in the absence of nucleotide cofactors, respectively.
Analytical Sedimentation Studies of the DnaB Helicase-(AMP-PNP) and DnaB Helicase-ADP Complexes-Analytical ultracentrifuge experiments provide direct information about the global conformational properties of a protein which are reflected in the hydrodynamic properties of the macromolecule (Cantor and Schimmel, 1980). Sedimentation velocity profiles of the DnaB helicase in buffer T2 (pH 8.1, 20°C) containing 50 mM NaCl and 5 ϫ 10 Ϫ4 M AMP-PNP (monitored at 292 nm) are shown in Fig. 2a. At this concentration of the ATP analog, all six ATP binding sites of the DnaB hexamer are saturated with the nucleotide (Bujalowski and Klonowska, 1993, 1994a, 1994b. There is a single, well-defined moving boundary throughout the sedimentation process which has a sedimentation coefficient of s 20,w ϭ 11.9 Ϯ 0.2. It should be pointed out that in our studies the error in the measurement of the sedimentation coefficients, for each particular system, is a standard deviation determined from 8 to 12 independent sedimentation velocity experiments using two different protein concentrations. The obtained value of s 20,w of the DnaB-(AMP-PNP) complex is ϳ14% higher than the sedimentation coefficient of s 20,w ϭ 10.5 Ϯ 0.2 of the free DnaB hexamer (Bujalowski et al., 1994). The magnitude of this increase of the sedimentation coefficient can be realized by recalling that a simple dimerization of two hexamers into a dodecamer would FIG. 1. DnaB helicase binding to ssDNA. Fluorescence titrations of poly(d⑀A) with the DnaB protein monitored by the increase of the nucleic acid fluorescence in buffer T2 (pH 8.1, 10°C) containing 50 mM NaCl, in the presence of 1 mM AMP-PNP (f), 1 mM ADP (E), and in the absence of nucleotide cofactors (q), respectively. The nucleic acid concentration is 2 ϫ 10 Ϫ5 M (nucleotide). Solid lines are computer fits of the binding isotherms, using Equation 2, with intrinsic binding constants K ϭ 1.3 ϫ 10 5 M Ϫ1 , 5.9 ϫ 10 2 M Ϫ1 , and 60 ϫ 10 1 M Ϫ1 for isotherms obtained in the presence of AMP-PNP, ADP, and in the absence of nucleotides, respectively. For the titration in the presence of AMP-PNP, the cooperativity parameter ϭ 3.5 and ⌬F max ϭ 3.5. For titrations in the presence of ADP and the absence of nucleotide cofactors, the value of the maximum increase of the nucleic acid fluorescence, ⌬F max , was assumed to be 3.5, the same as determined for the titration in the presence of AMP-PNP. lead to an ϳ58% increase of its sedimentation coefficient (Cantor and Schimmel, 1980). To exclude the possibility that the nucleotide binding affects the oligomeric state of the DnaB hexamer, we performed sedimentation equilibrium measurements of the DnaB helicase in the presence of saturating concentrations of AMP-PNP. A set of DnaB protein concentration profiles (recorded at two different wavelengths, 285 and 292 nm) as a function of the square of the radius at sedimentation equilibrium, in buffer T2 (pH 8.1, 10°C) containing 50 mM NaCl and 5 ϫ 10 Ϫ4 M AMP-PNP, are shown in Fig. 2b. The solid lines are nonlinear least-squares fits to a single exponential function (Equation 1). The excellent agreement between the theoretical lines and the experimentally obtained concentration profiles indicates that there is a single species in the solution. The analysis of the concentration profiles provides the molecular weights of 303,000 Ϯ 10,000 and 301,000 Ϯ 10,000 for scans recorded at 285 and 292 nm, respectively. Comparison between the obtained data and the known molecular weight of the DnaB hexamer (313,590) shows that the DnaB helicase fully preserves its hexameric structure in the presence of AMP-PNP (Bujalowski et al., 1994). Therefore, since the six bound molecules of AMP-PNP per hexamer constitute only ϳ1% of the molecular weight of the protein, a large increase of the sedimentation coefficient in the presence of AMP-PNP indicates dramatic conformational changes of the entire hexamer induced by the binding of the ATP analog (see "Discussion").
The thermodynamic studies of the binding of the DnaB helicase to ssDNA indicate that ADP also induces conformational transitions in the enzyme; however, this leads to a modest increase of the binding affinity toward the ssDNA (Fig. 1). Analogous sedimentation velocity experiments (data not shown) of the DnaB hexamer in the presence of the saturating concentration of ADP (5 ϫ 10 Ϫ4 M) show that the sedimentation coefficient of the helicase is s 20,w ϭ 11.4 Ϯ 0.2, a value significantly higher than the sedimentation coefficient of the free enzyme. However, this value is lower when compared to s 20,w ϭ 11.9 Ϯ 0.2 obtained in the presence of the ATP analog, AMP-PNP, and suggests conformational differences between these two complexes (see "Discussion").
It is very interesting to determine how the binding of the ssDNA affects the conformational transitions of the DnaB hexamer induced by nucleotide binding. Because the etheno-derivatives of the nucleic acids have significant absorption above 310 nm, where there is practically no contribution of the protein spectrum, this allows us to monitor only one component, the nucleic acid, during the sedimentation process. Sedimentation velocity profiles of the mixture of the DnaB helicase and d⑀A(p⑀A) 19 in the 1.25:1 molar excess of the enzyme over the nucleic acid are shown in Fig. 3. Recall, the 20-mer exactly spans the site size of the DnaB helicase-ssDNA complex (Bujalowski and Jezewska, 1995). Because the [DnaB] and [d⑀A(p⑀A) 19 ] are Ͼ Ͼ1/K 20 , all the nucleic acid should be complexed with the helicase (Bujalowski and Jezewska, 1995). As a result, only a single, well-defined boundary of the complex is observed throughout the entire sedimentation process and has a sedimentation coefficient of s 20,w ϭ 12.4 Ϯ 0.2. This value is ϳ17% higher than the sedimentation coefficient of the free DnaB hexamer and ϳ3 and ϳ9% higher than the sedimentation coefficient of the DnaB helicase, when saturated with the ATP analog or ADP. This suggests that the binding of ssDNA, most probably, introduces additional conformational changes in the ternary complex, DnaB-(AMP-PNP)-d⑀A(p⑀A) 19 (see "Discussion"). nucleic acids in the presence of nucleotide cofactors, is that the enzyme cycles in a vectorial fashion through a number of conformational states controlled by ATP binding and hydrolysis in which its affinity for ss-and dsDNA dramatically changes (Wong and Lohman, 1993). Yet, evidence of the physical existence of different conformations or intramolecular transitions within the oligomeric structure of the hexameric enzyme, induced by nucleotide binding, so far has not been obtained.

Binding of the ATP Analog, AMP-PNP, and ADP Induces Major Global Conformational Changes in the DnaB
Hydrodynamic properties, including the sedimentation coefficient, are directly sensitive to the global conformational properties of the macromolecules (Cantor and Schimmel, 1980). Using analytical sedimentation velocity experiments and the rigorous fluorescence titration technique, we present direct evidence that the E. coli primary replicative helicase DnaB protein undergoes dramatic conformational transitions induced by binding of the ATP nonhydrolyzable analog, AMP-PNP, and ADP. The sedimentation coefficient of the free DnaB hexamer increases from s 20,w ϭ 10.5 to s 20,w ϭ 11.9 for the hexamer in the presence of 5 ϫ 10 Ϫ4 M AMP-PNP and to s 20,w ϭ 11.4 in the presence of 5 ϫ 10 Ϫ4 M ADP. At these concentrations of AMP-PNP and ADP, all six binding sites are saturated with the nucleotide (Bujalowski and Klonowska, 1993).
The very large increase of the sedimentation coefficient does not result from the dimerization of the DnaB hexamer induced by nucleotide binding. Equilibrium sedimentation measurements show that, in the presence of the saturating concentration of AMP-PNP, the hexameric structure of the enzyme is fully preserved (Fig. 2b). In the absence of oligomerization, the sedimentation coefficient of the protein can change as a result of changes of its partial specific volume or frictional coefficient, or both. Partial specific volume of the ATP molecule is ϭ 0.44 ml/g compared with ϭ 0.732 ml/g of the DnaB protein (Howlett and Schachman, 1977;Bujalowski et al., 1994). However, the six bound nucleotide molecules constitute only ϳ1% of the molecular weight of the hexamer and the trivial effect of a lower of the bound nucleotides on the measured s 20,w , would amount to ϳ0.004; thus, it is negligible. Moreover, the additional data suggest that the global conformational changes are predominantly induced by the binding of only the first three nucleotide molecules to the three high affinity binding sites of the hexamer. 2 Therefore, regardless of the nature of the observed conformational transitions, the large increase of the sedimentation coefficient indicates large global structural changes of the DnaB hexamer, which are induced by ATP analog binding, and these structural transitions lead to intrinsic changes of the partial specific volume of the hexamer and/or its frictional coefficient. Although the exact nature of the conformational transition induced by nucleotide binding is still unknown, it should be mentioned that dynamic light scattering data indicate that the diffusion coefficient of the hexamer is increased in the presence of the saturating concentration of AMP-PNP, indicating that the frictional coefficient of the hexamer is decreased, in agreement with the sedimentation data reported here. 2 Binding of ADP to the DnaB helicase also induces global conformational changes in the protein hexamer. The sedimentation coefficient, s 20,w ϭ 11.4, is increased by ϳ9% when compared to the free DnaB hexamer; however, this value is smaller than the 11.9 obtained in the presence of AMP-PNP. The difference is larger than the error of the determination of s 20,w (Ϯ 0.2) estimated using multiple scans at different protein concentrations (see above). Thus, the sedimentation data suggest that the conformation of the hexamer-ADP complex differs from that of the free DnaB protein and is also different from the DnaB helicase-(AMP-PNP) complex. As a result, the affinity of the hexamer toward ssDNA drops by ϳ3 orders of magnitude when compared to the affinity of the DnaB-(AMP-PNP) complex (see above). The difference between the effect of AMP-PNP and ADP does not result from the weaker binding of ADP to the helicase. As we determined previously, using a series of fluorescent nucleotide analogs, the ADP and ADP analogs bind with higher affinity to the DnaB helicase than the ATP analogs (Bujalowski and Klonowska, 1993, 1994a, 1994b. Thus, the allosteric effect of AMP-PNP on the DnaB global conformation, which leads to the dramatically increased affinity of the helicase to ssDNA as opposed to the dramatic drop in the affinity in the presence of ADP, must result from the specific interactions of ␥-phosphate in the nucleotide binding site. The sedimentation coefficient of the DnaB-(AMP-PNP)-d⑀A(p⑀A) 19 ternary complex, s 20,w ϭ 12.4 Ϯ 0.2, indicates that conformational changes induced by AMP-PNP binding are preserved in the ternary complex. It should noted that the increase of the sedimentation coefficient of the ternary complex exceeds the value expected from the combined trivial effects of the different partial specific volume of the nucleic acid and the higher molecular weight of the DnaB hexamer-20-mer complex. The bound 20-mer constitutes only an additional ϳ2% of the molecular weight of the complex. The trivial effect of the lower partial specific volume of the nucleic acid molecule ( ϭ 0.531 ml/g, Pearce et al. (1975)) and the increased molecular weight of the complex would contribute ϳ0.008 and ϳ0.2 to the measured sedimentation coefficient, respectively (Cantor and Schimmel, 1980). Moreover, binding of the shorter 10-mer ssDNA fragment (data not shown) causes the same increase of the sedimentation coefficient of the hexamer (s 20,w ϭ 12.5 Ϯ 0.2) as the 20-mer, although it has 10 less nucleotide residues than the 20-mer and constitutes only ϳ1% of the molecular weight of the hexamer. If the trivial effects were mainly responsible for the observed increase of s 20,w of the ternary complex, then the effect of the 10-mer should be half of the one observed for the 20-mer, but this is not what is experimentally observed. Thus, the increased value of s 20,w of the ternary complex suggests that, most probably, the DnaB helicase undergoes further conformational changes upon binding ssDNA, although the obtained hydrodynamic data indicate that the major conformational transition of the helicase is induced by AMP-PNP or ADP binding.
Multiple Conformational States of the DnaB Hexamer-Hydrodynamic studies reported in this work provide the first direct evidence of the dramatic global conformational changes of the DnaB helicase hexamer, induced by nucleotide binding, and the existence of the multiple, structural states of the enzyme. Moreover, the different conformations correlate well with the functional properties of the enzyme, i.e. its interactions with DNA, as determined by thermodynamic studies. In the absence of the nucleotide cofactors, the DnaB hexamer, built of six chemically identical subunits, has a very low affinity toward ssDNA. In this "closed" state, all subunits of the hexamer are equivalent and capable of initiating binding of ATP and ADP (see above, Bujalowski and Klonowska (1993)). Upon binding the ATP analog, the global rearrangement of the protomers within the hexamer is induced, resulting in a "tense" state with a high affinity for ssDNA. The transition leads to the selection of only a limited set of subunits, most probably only one, as a binding site for ssDNA (Bujalowski and Jezewska, 1995). The large increase of the sedimentation coefficient suggests a more compact structure of the hexamer in a tense state. Although ADP also induces similar large global changes in the structure of the DnaB hexamer, this state of the enzyme is different from the one induced by ATP. A smaller value of s 20,w suggests a less compact structure of the hexamer in this "relaxed" state as compared to the tense state.
The mechanism of the dsDNA unwinding by the replicative helicase and the mechanism of the enzyme translocation on the nucleic acid lattice is still unknown. In general, after breaking hydrogen bonds between the base pairs of the duplex DNA in the replication fork, the enzyme must be released from the formed ss nucleic acid and move, in an unidirectional translocation event, toward dsDNA (Hill and Tsuchiya, 1981). Our recent results show the existence of only a single, strong binding site and a very low stoichiometry of the DnaB-ssDNA complex. These results are not compatible with the models of hexameric helicase translocation along the nucleic acid lattice in which all six protomers and/or multiple binding sites are involved in ssDNA binding (Bujalowski and Jezewska, 1995).
The existence of different conformational states of the hexameric helicase described in this work strongly suggests that the mechanism of translocation and nucleic acid unwinding might rely on global, not local, conformational changes in the hexamer which are induced by the ATP/ADP switch and/or nucleic acid binding. Such conformational transitions were postulated as necessary elements of the helicase mechanism by Hill and Tsuchiya (1981). Thus, because the transition from the tense to the relaxed state of the helicase must accompany each ATP hydrolysis step, when the ATP/ADP switch takes place in the nucleotide binding site, this transition is most probably responsible for the partial release of the ssDNA from the nucleic acid binding site (Arai and Kornberg, 1981b). After the release of ADP, the global rearrangement of the protomers within the entire DnaB hexamer, from a closed conformation to a tense one upon rebinding ATP, would allow the enzyme to translocate and encompass the subsequent fragment of dsDNA, within the active site of the helicase.