Versatile Action of Escherichia coli ClpXP as Protease or Molecular Chaperone for Bacteriophage Mu Transposition*

The molecular chaperone ClpX of Escherichia coli plays two distinct functions for bacteriophage Mu DNA replication by transposition. As specificity component of a chaperone-linked protease, it recognizes the Mu immunity repressor for degradation by the peptidase component ClpP, thus derepressing Mu transposition functions. After strand exchange has been promoted by MuA transposase, ClpX alone can alter the conformation of the transpososome (the complex of MuA with Mu ends), and the remodeled MuA promotes transition to replisome assembly. Although ClpXP can degrade MuA, the presence of both ClpP and ClpX in the reconstituted transposition system did not destroy MuA essential for initiation of DNA replication by specific host replication enzymes. Levels of ClpXP needed to overcome inhibition by the repressor did not prevent MuA from promoting strand transfer, and ClpP stimulated alteration of the transpososome by ClpX. Apparently intact MuA was still present in the resulting transpososome, promoting initiation of Mu DNA replication by specific replication enzymes. The results indicate that ClpXP can discriminate repressor and MuA in the transpososome as substrates of the protease or the molecular chaperone alone, degrading repressor while remodeling MuA for its next critical function.

Chaperone-linked proteases employ the action of a protease component and a regulatory subunit with ATPase activity. In Escherichia coli, ClpA and ClpX proteins serve as regulatory subunits for the protease subunit ClpP (1)(2)(3). These regulatory subunits can by themselves act as molecular chaperones, breaking up aggregates or changing quaternary interactions of specific protein substrates (4,5). The substrates of the molecular chaperone can become the targets of the chaperone-linked protease when the corresponding protease subunit is present (6). Thus, the molecular chaperones endow the protease with substrate specificity, most likely recognizing the substrate and presenting it to the protease subunit.
ClpP is synthesized as a protein of 207 amino acids (7), of which 14 amino acids at the N terminus are autocatalytically removed to yield the mature enzyme (8). The active protein consists of a tetradecamer, composed of two stacked heptameric rings (9,10). The structure of ClpP is analogous to the 20 S proteasomes of eukaryotes and Archaebacteria, with multiple active sites residing in the interior of the multimeric rings (10). Presumably, the molecular chaperones unfold the substrate and feed it into the proteolytic chamber of the ClpP tetradecamer, leading to the apparently processive degradation of the substrate (11).
ClpX is known to be involved in two distinct stages of the Mu life cycle. As part of the ClpXP protease, it can promote entry of a lysogen into lytic development by degrading the Mu immunity repressor (12,13), which serves to shut down Mu transposition functions for the establishment and maintenance of lysogeny. ClpX also promotes initiation of Mu DNA replication, a process that does not require ClpP (14). It activates MuA's function of promoting transition to DNA synthesis after MuA's role in recombination has been completed (15).
During transposition, the MuA transposase binds to the Mu ends (16) and assembles into a tetramer (17), which catalyzes transfer of each Mu end to target DNA (18 -20). This forms a branched DNA intermediate that contains a potential replication fork at each Mu end. Replication of Mu DNA by specific host enzymes completes transposition to form the cointegrate product (21)(22)(23)(24)(25), and MuA plays a critical role in this process. Upon completing Mu strand transfer, MuA remains very tightly bound to the Mu ends, still holding the Mu ends together in a nucleoprotein complex known as the type II transpososome (or the strand transfer complex 1, STC1) 1 (19,26). Removal of MuA by phenol extraction to create a deproteinized strand transfer product (STP) eliminates the requirement in vitro for specific host replication proteins implicated in Mu replication in vivo (25,27), and if MuA in STC1 is damaged by partial proteolysis, replication cannot be initiated (25).
ClpX remodels MuA in STC1 to form STC2, destabilizing MuA's tight grip on DNA without removing it from the nucleoprotein complex and activating MuA's potential to promote transition to a replisome (15). Yet unidentified factors (Mu replication factor (MRF) ␣ 2 ) displace MuA in STC2 to form a new nucleoprotein complex STC3 (15), which permits initiation of DNA replication only by specific replication proteins including the DNA polymerase (pol) III holoenzyme and constituents of the multiprotein priming apparatus known as the X174type primosome (27).
However, MuA is also a substrate for the ClpXP protease (28) just as other known substrates of ClpX and ClpA as molecular chaperones are substrates of the corresponding chaperonelinked protease (reviewed in Gottesman et al. (29)). Therefore, can ClpX in the presence of ClpP remodel MuA in STC1 without degrading it? Whether a protein substrate is degraded or remodeled may simply be determined by the availability of ClpP when the substrate is bound by the molecular chaperone. Such a model raises the question of how ClpX could perform the two distinct functions for Mu lytic development, first promoting degradation of repressor by ClpP and then remodeling MuA in STC1 for transition to replisome assembly. The first function requires ClpP but the second function could potentially be disrupted by the engagement of the ClpP proteolytic activity. ClpX would not be expected to act in isolation from ClpP in vivo to perform the second function, raising the question of what effect ClpP has on the DNA replication phase of Mu transposition. Here we report that ClpP does not interfere with MuAmediated transition to DNA synthesis even when present in saturating amounts and in fact stimulates ClpX-promoted conversion of STC1 to STC2.

Materials
Bacterial Strains, Plasmids, and Proteins-E. coli BL21(DE3) (30) and plasmid pET-19b were purchased from Novagen. Plasmid pWPC9 (8) including the coding sequences for ClpP and ClpX was obtained from Susan Gottesman (National Institutes of Health). Plasmid pGG215 (19) bearing a mini-Mu element was used for Mu transposition in vitro. Phage f1 RFI DNA and X174 RFI DNA were used as target substrates. Purified preparations of MuA, MuB, Muc ϩ (Rep), Mucvir3060 (Vir), and host proteins required for Mu transposition and replication have been described (15,31). DNA pol III* was purified from MGC1020 (W3110 malE::Tn10, lexA3, uvrD::kan) obtained from Dr. Charles McHenry (University of Colorado Health Sciences Center) as described previously (32). Purification of ClpP is described below. Restriction enzymes, DNA pol I, DNA pol I large fragment (Klenow), and E. coli DNA ligase were purchased from New England BioLabs.
Plasmid pCPX01 expressing ClpP with a polyhistidine tag fused to the N terminus (h-ClpP) was constructed by cloning an NdeI-BamHI fragment of pWPC9 (8) containing the coding sequences for ClpP and ClpX into NdeI-BamHI cloning site of pET-19b, placing them under the control of the T7 10 promoter (Fig. 1A). The site at which ClpP is autocatalytically processed was retained on this expression vector.

Methods
Fluorogenic Assay for ClpP Activity-The protease activity of ClpP alone was measured by its ability to cleave the fluorogenic substrate Suc-Leu-Tyr-AMC (33). Suc-Leu-Tyr-AMC (5 mM) was incubated with ClpP in buffer A (25 l total volume) for 30 min at 37°C. The reaction was stopped by addition of 475 l of ice-cold 0.6 M imidazole (pH 7.8) and transferred to a quartz cuvette suitable for analysis in an AMINCO Bowman Series 2 luminescence spectrometer. To determine the amount of AMC generated, the sample was irradiated at 380 nm, and emission at 460 nm was measured and compared with a standard curve for AMC fluorescence. One unit of ClpP activity was defined as 10 pmol of AMC generated under the conditions described here.
Overexpression of H-ClpP and Purification by Ni-Column Chromatography-Strain BL21(DE3) transformed with plasmid pCPX01 was grown in 10 liters of modified terrific broth (15) in a 14-liter New Brunswick fermenter (37°C, A 595 ϭ 2.0), and ClpP overproduction was induced by the addition of isopropyl ␤-D-thiogalactopyranoside to 0.4 mM. After 30 min, cells were harvested into chilled bottles and collected by centrifugation at 4°C (74 g of wet cells). Cells were resuspended in buffer B to an A 595 of 315 and transferred to Ti45 centrifuge tubes.
All steps for protein fractionation were conducted at 4°C unless otherwise indicated. To the cell suspension, lysozyme was added to 300 g/ml, and cells were placed on ice for 30 min. They were then subjected to three rounds of rapid freeze-thaw using liquid N 2 and three bursts of sonication (10 s each). The lysate was cleared by centrifugation in a Beckman Ti45 rotor (35,000 rpm, 30 min, 4°C), and the supernatant was collected (Fraction I; 140 ml; 10 g of protein). To remove nucleic acids, polyethylenimine was added dropwise to 0.16% with constant stirring over 10 min. Nucleic acids were removed by centrifugation in an SS34 rotor (9,000 rpm, 30 min, 4°C). Protein was precipitated from the supernatant by the addition of ammonium sulfate to 70% saturation and collected by centrifugation in an SS34 rotor (9,000 rpm, 30 min). Protein was resuspended in a minimal volume of buffer B and dialyzed twice (45 min each) against 1 liter of buffer B (Fraction II; 64 ml; 2.6 g).
Fraction II was loaded at 1 column volume/h onto a Ni-NTA column (60-ml bed volume) equilibrated in buffer B. The column was washed extensively with buffer B and then with buffer B containing 25 mM imidazole. The column was then eluted with a 600-ml linear gradient from 25 to 600 mM imidazole in buffer B. Protein fractions containing ClpP (h-ClpP and processed ClpP) were identified by SDS-PAGE and pooled (Fraction III; 225 ml; 76 mg). ClpP was precipitated by the addition of 2 volumes of acetone and collected by centrifugation in an SS34 rotor (9,000 rpm, 30 min). The protein was resuspended in 10 ml of buffer B and dialyzed against 1 liter of buffer B for 60 min. The sample was then incubated at 37°C for 3 h to promote further autocatalytic processing of h-ClpP to the mature form (Fraction IV; 40 ml; 26 mg). Although acetone precipitation was employed in the preparation of Fraction IV to obtain high protein concentrations and to maximize purity, this step could be omitted to maximize yield of ClpP without significant loss of purity. Fraction IV was loaded at 1 column volume/h onto a Ni-NTA column (12-ml bed volume) equilibrated in buffer B. The column was washed with 24 ml of buffer B and then with buffer B containing 50 mM imidazole until no more protein eluted from the column. The majority of mature ClpP eluted during these washes and were pooled. The purified ClpP was dialyzed for 1 h against 1 liter of buffer C (Fraction V; 35 ml; 15 mg) and stored in small aliquots at Ϫ80°C.
Replication products were extracted and linearized with EcoRI (f1 RFI target) or BamHI (X174 RFI target), resolved on a 0.6% agarose gel (alkaline electrophoresis buffer) and stained with SYBR Green I for detection by the Molecular Dynamics Storm 840 system. Gels were then dried for autoradiography or phosphorimagery. Phosphorimagery data was analyzed using ImageQuant software. All images of replication products in figures are from autoradiographs.
Gel Disruption Assay to Detect Conversion of STC1 to STC2-Conversion of STC1 to STC2 was measured by detecting disruption of the synaptic complex of Mu ends by agarose gel electrophoresis. The synaptic complex is unstable in STC2 and is disassembled during electrophoresis (15). Reaction mixtures (25 l) containing STC1 (equivalent of 0.25 g of donor substrate, predigested with PstI (300 units/ml at 37°C for 30 min)), ClpX, and ClpP (concentrations indicated in individual experiments) were incubated at 37°C for 2-30 min under previously described reaction conditions (15). Reactions were stopped with EDTA (20 mM) and heparin (2 mg/ml), and products were resolved on a 0.6% agarose gel (TAE electrophoresis buffer containing 0.5 g/ml EtBr). DNA was stained with SYBR Green I for detection by the Molecular Dynamics Storm 840 system, and bands corresponding to STC1 and the disrupted complex (STC2) were quantitated using ImageQuant software.
Other Methods-Glutaraldehyde fixation of nucleoprotein complexes, Western blot analysis, and isolation of nucleoprotein complexes by gel filtration were performed as described previously (15). Densitometric analysis of Western blots detected by enhanced chemiluminescence was performed using Scan Analysis TM software from Biosoft. ClpP protein concentrations were determined by the method of Bradford (36). Nterminal sequencing of purified ClpP protein was performed by Dr. A. Fowler (UCLA Protein Microsequencing Facility).

Overproduction of Histidine-tagged ClpP and Rapid
Purification of the Mature Protein-In plasmid pCPX01 the clpP and clpX genes are under the control of the phage T7 10 promoter (Fig. 1A). ClpP protein is expressed with a polyhistidine tag fused to the N terminus (h-ClpP) for rapid purification on Ni-NTA agarose columns (see "Experimental Procedures" and Table I). Induction of BL21(DE3) harboring pCPX01 resulted in the overproduction of proteins of molecular mass of 45, 23, and 21 kDa, apparently corresponding to ClpX, h-ClpP, and the mature form of ClpP (Fig. 1B, lane 2). The overproduction of mature ClpP suggested that h-ClpP could be processed to the mature form just as the wild-type precursor ClpP (pre-ClpP) is autocatalytically processed by cleavage of 14 amino acids at the N terminus (8).
Both h-ClpP and mature ClpP were retained on the Ni-NTA column (Fig. 1C, lane 1). The removal of the precursor peptide is not required for ClpP oligomerization (37), and thus mature ClpP associated with h-ClpP should also be retained on Ni-NTA. A third polypeptide retained by the column corresponded to the molecular weight of pre-ClpP. The bound protein fraction from the Ni-NTA column was incubated at 37°C for 3 h to promote further processing of the his-ClpP and then passed through the column a second time. Most of the protein passed through the column or was eluted at low (50 mM) imidazole concentrations. These protein fractions were pooled and were found to be greater than 95% homogeneous for the 21-kDa form; the presence of the 23-kDa form was undetectable (Fig.  1C, lane 2). The N-terminal sequence of the purified 21-kDa form was determined to be identical to that of mature wild-type ClpP.
ClpXP Protease Relieves Inhibition of Mu Transposition in Vitro by Repressor in the ClpXP-sensitive State-The Mu immunity repressor inhibits Mu transposition in vitro by competing for MuA binding sites in the operator that enhances transpososome assembly (38,39). ClpXP protease can degrade the Mu immunity repressor (13), and it potentially plays an important role in inducing Mu DNA replication in lysogens. We determined whether the presence of ClpXP protease allows the reconstituted Mu transposition system to overcome inhibition by the repressor.
Although the wild-type repressor (Rep) is a substrate for the ClpXP protease, it is degraded with a relatively high K m and assumes a ClpXP-resistant state in the presence of DNA (13). A mutant form of the repressor (Vir, repressor encoded by the virulent strain Muvir3060), which has an alteration at the C terminus and is rapidly degraded in vivo by the ClpXP protease (12,14), is degraded by ClpXP with a 20-fold lower K m and is locked into the high affinity state even when bound to DNA (13). Both Rep and Vir are known to bind operator DNA with nearly equal affinity, Vir having a slightly higher dissociation constant (40). They both inhibited Mu strand transfer in the reconstituted system (Fig. 2, cf. lanes 6 and 10 with lane 2), but only inhibition by Vir could be eliminated by ClpXP (cf. lane 13 with 9). And even though MuA is a ClpXP substrate (28), amounts of ClpXP sufficient to remove Vir inhibition did not destroy MuA needed to catalyze strand transfer (Fig. 2, lanes 5  and 13). MuA is degraded slowly compared with Vir (13), permitting MuA to carry out its transposition function at levels of ClpXP that effectively eliminate the repressor.
ClpP Stimulates Action of ClpX on MuA Present in STC1-One method for examining the conversion of STC1 to STC2 is to examine disassembly of the transpososome by agarose gel electrophoresis (15). Treatment of STC1 with ClpX destabilizes tight binding of MuA to the DNA such that it readily disassembles during challenge with heparin and resolution by electrophoresis, and disruption of the synaptic complex of Mu ends results in a change in mobility of the strand transfer product. We used this disassembly assay to compare the kinetics with which ClpX acts on STC1 in the presence or absence of ClpP.
The presence of ClpP increased by 2-3-fold the amount of STC1 disassembled at 30 min by limiting concentrations of ClpX (Fig. 3A). The level of ClpP used was saturating; the ClpP was overproduced as a histidine-tagged fusion protein and purified as described under "Experimental Procedures." Specific activity was determined using the fluorogenic assay described in "Experimental Procedures"; 1 unit corresponds to 10 pmol of fluorescent product released in 30 min at 37°C.  1. Overproduction and purification of ClpP protein. A, pCPX01 contains the clpP gene downstream of an in-frame coding region for a polyhistidine tag in pET-19b (Novagen). In this plasmid both clpP and clpX are under the control of the T7 10 promoter. B, E. coli BL21(DE3) cells transformed with pCPX01 were grown and induced as described under "Experimental Procedures." Equivalent quantities of cells from uninduced and induced cultures were lysed in cracking buffer and subjected to SDS-PAGE on 12% gels. Proteins were visualized by Coomassie Brilliant Blue staining (46). C, samples of ClpP Fraction III and IV (0.5 g) were separated by SDS-PAGE on 12% gels, and proteins were visualized by Coomassie Brilliant Blue staining.
presence of 2-fold higher levels of ClpP did not result in any additional stimulation (data not shown). Time course experiments (Fig. 3, B and C) revealed that the extent of STC1 disassembly was approximately proportional to the amount of added ClpX and that ClpP increased the extent of STC1 disassembly at each ClpX concentration. Even in the presence of ClpP, however, an excess of ClpX monomers over STC1 was required for 100% disassembly, with at least 2.5 pmol of ClpX being required to disassemble 0.05 pmol of STC1 (formed from 250 ng of donor substrate in a reaction mixture containing 55 ng of MuA). If sufficient ClpX was present, ClpP was not essential for complete disassembly of STC1 (e.g. see Fig. 5, lane 7) as shown previously (15,28).
ClpXP Promotes Transition from STC1 to STC2, Leaving Activated MuA Molecules Bound to the Mu Ends-The disassembly assay used to detect ClpX action above does not necessarily measure conversion of STC1 to STC2. STC2 includes activated MuA that remains bound to DNA and plays a critical role in the transition to DNA synthesis. The removal of that MuA from the DNA by artificial means such as phenol extraction eliminates the specific requirement in vitro for replication enzymes that are needed in vivo. For example, on the deproteinized STP, potential primers for leading strand synthesis at the Mu left and right ends can be extended by DNA pol I or Klenow fragment, whereas they cannot be extended by pol I in STC1, STC2, or STC3, even when all components of the Mu replication system except pol III holoenzyme are present (15,27). We determined whether ClpX in the presence of ClpP removes MuA from STC1, leaving the ends of Mu accessible to DNA pol I, or converts STC1 to STC2 in which MuA remains bound to the ends.
Upon treatment of STC1 with ClpXP, the leading strand primers at the Mu left and right ends could not be appreciably extended by the Klenow fragment (Fig. 4B, lanes 5-7). This was true even after incubation for 60 min, twice the time required for 100% of STC1 to be disassembled by the gel electrophoresis assay (Fig. 3B). Identical results were obtained with STC1 treated with ClpX alone (Fig. 4B, lanes 2-4). Although a low level of nucleotide incorporation was detected after prolonged incubation of ClpXP-treated STC with Klenow, this was not associated with any measurable extension of the leading strand primers. Thus, only a limited amount of leading strand synthe-sis if any could have been catalyzed, and most of the incorporation was due to nonspecific DNA synthesis such as the alternating action of 3Ј to 5Ј exonuclease and polymerase activities of Klenow at 3Ј-hydroxyl ends. In contrast, leading strand primers on the deproteinized STP could be readily extended by the Klenow fragment (Fig. 4A, lanes 2-5), and the presence of ClpXP could not inhibit this reaction (data not shown).
The synaptic complex in STC treated with ClpXP could be preserved with glutaraldehyde (Fig. 5, lane 10), just as it is when STC1 is treated with ClpX alone (lane 12). This indicates that MuA maintains the synaptic complex in STC even after treatment with ClpXP. Cross-linking MuA with glutaraldehyde stabilizes the synaptic complex in STC2 such that it is pre- served during agarose gel electrophoresis, and we have verified the presence of MuA in this cross-linked nucleoprotein complex (15). We confirmed that this concentration of ClpX and ClpP was able to convert STC1 to STC2. When STC1 was treated with ClpXP and the resulting nucleoprotein complex was not cross-linked prior to electrophoresis (Fig. 5, lane 5), essentially 100% of the DNA migrated at the position of the deproteinized STP (lane 13). This was also true when STC was treated with a higher level of ClpX alone (lane 7).
Undegraded MuA was still present after treatment of STC with ClpXP. STC1 isolated free of unbound proteins by gel filtration was treated with sufficient ClpXP or ClpX alone so that 100% of the synaptic complexes could be disrupted by gel electrophoresis. We detected apparently intact MuA in the resulting nucleoprotein complex by Western blot analysis (Fig.  6, lane 4). Although no reduction in MuA levels is apparent after treatment with ClpX alone (Fig. 6, lanes 2 and 5), we generally observed some reduction in MuA based on densitometric analysis of the Western blot (no more than 2-fold) when both ClpX and ClpP were present. Essentially 100% of synaptic complexes in ClpXP-treated STC could be preserved with glutaraldehyde (Fig. 5, lane 10), and very little DNA synthesis was catalyzed on these templates by pol I Klenow fragment (Fig.  4B, lanes 5-7). Thus, any MuA degradation is not associated with disassembly of transpososomes but rather the removal of any MuA not essential for STC2 function.
Specific Host Proteins Are Required to Initiate Mu DNA Replication on ClpXP-treated STC-The most important function of MuA bound to STC2 is to allow initiation of Mu DNA replication only by a specific group of host proteins, which include DnaB helicase, DnaC protein (15), and X174-type primosome proteins PriA and DnaT (27). Proteolytic damage to MuA in STC1 can block Mu replication even by these specific proteins (31). However, ClpXP-treated STC allowed Mu DNA synthesis to proceed with kinetics comparable to that on STC treated with ClpX alone (Fig. 7A). We have previously demonstrated that the primers at the ends of Mu in STC are not extended when all components required for replication except PriA or DnaBC are present (27). Likewise, on both ClpXtreated and ClpXP-treated STC, initiation of Mu DNA replication required MRF␣, DnaB, DnaC, and PriA (Fig. 7B, cf. lanes  3 and 4 with 2, and lanes 8 and 9 with 7). Mu DNA synthesis was also dependent upon DnaT, the low levels of Mu DNA synthesis catalyzed without added DnaT (Fig. 7B, lanes 5 and  10) most likely being due to DnaT present in the MRF␣ fractions (27). On the deproteinized STP, limited extension of the leading strand primers at the left and right ends was catalyzed in the absence of these factors (Fig. 7B, lane 11), whereas no limited synthesis was catalyzed on ClpX-treated and ClpXPtreated STC.
Thus our results indicate that the function of STC2 produced by the action of ClpXP is indistinguishable from STC2 produced by ClpX alone (Figs. 4, 5, and 7). ClpP in these reactions can interact functionally with ClpX to relieve inhibition of transposition by Vir (Fig. 2) and stimulate STC1 to STC2 transition promoted by limiting ClpX (Fig. 3). velopment (12)(13)(14). ClpX is also needed for Mu DNA replication (14), activating MuA in the transpososome so that MuA can promote transition to DNA synthesis (15). The latter function does not require ClpP and in theory could be disrupted by its presence. If the presence of ClpP is sufficient to promote degradation of substrates bound by ClpX, MuA in the transpososome would be degraded and would no longer be available to regulate access of host enzymes to the potential Mu replication forks. The results of this work indicate that the ClpX function of activating MuA can be carried out in the presence of the ClpP protein despite the fact that MuA can be a protease substrate (28,41). ClpP stimulates ClpX without degrading crucial MuA needed for STC2 function. Thus, ClpX in the presence of ClpP can simultaneously function as a molecular chaperone as well as a specificity component for the protease.
The mechanism by which substrates are specifically recognized by ClpX and ClpP is not fully understood. Substrates that can be remodeled by molecular chaperone ClpA or ClpX alone can also be degraded when ClpP is also present. For example, aggregates of the phage O protein, which is the initiator for DNA replication, can be disaggregated by ClpX alone (5), and O is degraded by the ClpXP protease (3). MuA is also an example of a protein that is both a substrate of the ClpX molecular chaperone and a substrate of the ClpXP protease (28). Wawrzynow et al. (42) have hypothesized that determinants on the substrate specify whether ClpX will release the substrate to promote protein folding or present the substrate to ClpP for degradation. Determinants necessary for degradation are present in the C-terminal regions of the ClpXP substrates MuA (28,41) and Mu Vir repressor (43). The last 10 amino acids at the C terminus of MuA, when attached to the C terminus of the phage P22 Arc protein, are sufficient to convert Arc into a ClpXP substrate (41). However, when 7 amino acids from the C terminus of Vir3061 are fused to F plasmid proteins CcdA and CcdB, only CcdA becomes a ClpXP substrate (43). This leaves open the possibility that additional determinants present in Arc and CcdA but not in CcdB are also needed to promote degradation of the substrate by the protease.
In the in vitro transposition system, ClpXP degraded Vir to remove the transposition block while leaving MuA in STC2 intact. Although ClpXP can degrade monomeric MuA in solution (41), MuA needed for STC2 function was not degraded even after treating the transpososome with excess ClpXP for an extended period (60 min). This indicates that ClpP will not necessarily degrade all molecules recognized by ClpX and that some property of the ClpX substrate determines whether it is presented to ClpP for degradation. Once ClpX binds to substrate recognition sites on MuA or repressor, other signals on these proteins may determine whether they are degraded. The accessibility of such signals that promote substrate degradation may be determined by the protein's tertiary or quaternary structure or its interactions with other macromolecules (44). DNA binding and quaternary interactions of MuA in the transpososome could shield determinants that promote its presentation to ClpP. ClpXP can discriminate Vir bound to operator and MuA in STC1 as substrates for degradation and for remodeling, respectively, and this plays an important function in determining the course of Mu lytic development.
ClpP only stimulates ClpX function by 2-3-fold for the transition of STC1 to STC2, ClpX being able to perform this function alone when present at sufficiently high levels, and this is consistent with the finding that ClpP is not required for Mu DNA replication in vivo (14). The stimulation of ClpX activity suggests that ClpP is interacting with ClpX; however, how ClpP interacts with ClpX to stimulate chaperone activity is not yet clear. Our results indicate that a molar excess of ClpX monomers over STC1 was required for complete conversion of STC1 to STC2 and that the extent of STC2 formation was proportional to the ClpX concentration. The molar excess of ClpX used in the STC1 disassembly assay is typical of amounts used in previous work (15,28,41), and it is possible that ClpX is not turning over catalytically in this reaction. Through its interaction with ClpX, ClpP may be increasing the effective concentration of active ClpX in the reaction mixture. ClpP could allosterically activate the ClpX molecular chaperone or stabilize an active multimeric configuration of ClpX.
Alternatively, the ClpP proteolytic function could facilitate transition to STC2. STC2 formed in the presence of ClpXP was functionally identical to that formed with ClpX alone, but we consistently noticed some decrease in total MuA upon treatment of STC1 with ClpXP. Transpososomes are made up of a core tetramer of MuA plus additional, more loosely bound MuA protomers that play an auxiliary function (45). It is possible that these auxiliary MuA protomers are degraded during ClpXP treatment, thereby facilitating STC1 to STC2 transition. The oligomeric structure of MuA must nevertheless be retained in the resulting STC2 for Mu ends to be held together in a synaptic complex. When ClpX promotes transition of STC1 to STC2, the oligomeric structure of MuA is preserved (15). FIG. 7. ClpP does not interfere with MuA-promoted replication of STC by specific proteins. A, replication was catalyzed on STC1 (f1 RFI target) in the eight-protein system supplemented with PriA, PriB, PriC, DnaT, ClpX (228 ng), MRF␣, and ClpP (0 or 880 ng). Replication was measured by determining deoxynucleotide (dNMP) incorporation (picomoles). Values are the average of three to four independent trials with standard deviation of the mean given by error bars. B, replication was catalyzed for 30 min on STC1 (lanes 1-10) as in A except that DNA pol I was increased to 0.1 unit and various proteins were omitted as indicated. Replication was also catalyzed for 30 min on the deproteinized STP with 0.1 unit of DNA pol I in a reaction system that also included SSB, DNA gyrase, DnaG protein, and DNA ligase (lane 11). DNA products were extracted, digested with EcoRI, which cuts asymmetrically in the donor vector, and separated by alkaline agarose gel electrophoresis. P, positions of leading strand primers at the Mu ends prior to replication (visualized by SYBR Green I staining); Ex, products resulting from extension of leading strand primers by DNA pol I. Co, full-length cointegrate product; hmw, high molecular weight replication product (34).
The versatility of ClpXP to act either as molecular chaperone or protease may play a key role in determining the course of biochemical reactions such as MuA-catalyzed transposition. While the decision not to degrade MuA in STC1 and to promote transition to STC2 would assure Mu DNA replication, degradation of MuA under certain cellular conditions could permit other pathways for processing the strand transfer intermediate to yield the final recombination product. Thus, the choice of acting as protease or chaperone could influence key decision points in biochemical pathways.