Báo cáo Y học: Holliday junction binding and processing by the RuvA protein of Mycoplasma pneumoniae

Eur. J. Biochem. 269, 1525–1533 (2002) (cid:211) FEBS 2002

Holliday junction binding and processing by the RuvA protein of Mycoplasmapneumoniae

Stuart M. Ingleston1, Mark J. Dickman2, Jane A. Grasby3, David P. Hornby2, Gary J. Sharples4 and Robert G. Lloyd1 1Institute of Genetics, University of Nottingham, Queen’s Medical Centre, Nottingham, UK; 2Transgenomic Research Laboratory, Krebs Institute, Department of Molecular Biology and Biotechnology, University of She(cid:129)eld, UK; 3Krebs Institute, Centre for Chemical Biology, University of She(cid:129)eld, UK; 4Department of Biological Sciences, University of Durham, UK

nificantly, it binds duplex DNA more readily. However it does not support branch migration mediated by E. coli RuvB and when bound to junction DNA is unable to pro- vide a platform for stable binding of E. coli RuvC. It also fails to restore radiation resistance to an E. coli ruvA mutant. The data presented suggest that the modified pin region retains the ability to promote junction-specific DNA bind- ing, but acts as a physical obstacle to linear duplex DNA rather than as a charge barrier. They also indicate that such an obstacle may interfere with the binding of a resolvase. Mycoplasma species may therefore process Holliday junc- tions via uncoupled branch migration and resolution reac- tions.

Keywords: recombination; DNA repair; RuvABC resolva- some.

The RuvA, RuvB and RuvC proteins of Escherichia coli act together to process Holliday junctions formed during recombination and DNA repair. RuvA has a well-defined DNA binding surface that is sculptured specifically to accommodate a Holliday junction and allow subsequent loading of RuvB and RuvC. A negatively charged pin pro- jecting from the centre limits binding of linear duplex DNA. The amino-acid sequences forming the pin are highly con- served. However, in certain Mycoplasma and Ureaplasma species the structure is extended by four amino acids and two acidic residues forming a crucial charge barrier are missing. We investigated the significance of these differences by analysing RuvA from Mycoplasma pneumoniae. Gel retar- dation and surface plasmon resonance assays revealed that this protein binds Holliday junctions and other branched DNA structures in a manner similar to E. coli RuvA. Sig-

ducing a pair of symmetrically related incisions at specific sequences as the DNA strands move through the complex [8–10].

The formation and subsequent processing of Holliday junctions are key stages in recombination and DNA repair that provide the means to repair broken DNA molecules, generate recombinants in genetic crosses and rescue repli- cation forks stalled at lesions in the template strands [1–4]. Once formed, these four-way branched DNA structures are targeted by junction-specific DNA helicases and endonuc- leases that act, respectively, to move the branch point along the DNA (branch migration) and to cut specific DNA strands at or near the crossover, thus releasing duplex DNA products (resolution). In Escherichia coli, the resolution reaction appears to be coupled to branch migration via the formation of a specialized molecular machine composed of three protein subunits, RuvA, RuvB and RuvC [5,6]. A tetramer of RuvA binds one face of an open Holliday junction to form a specific complex that supports the loading of two RuvB ring helicases on opposing duplexes and of a dimer of RuvC endonuclease on the other face of the junction in the space between the RuvB rings [7,8]. The RuvAB proteins catalyse junction branch migration while RuvC resolves the structure to duplex products by intro-

The RuvA protein plays a pivotal role in processing Holliday junctions. It functions as a specificity factor for junction binding, provides a RuvA-junction scaffold for assembly of RuvB and RuvC, and actively participates in the branch migration and resolution reactions [11,12]. The atomic structure of RuvA reveals four L-shaped monomers comprising a fourfold symmetrical platform uniquely adapted for binding four-way branched DNA molecules [13,14]. Grooves on the concave surface of the tetramer accommodate each duplex arm of the junction in an open square conformation [13–16]. Two helix-hairpin-helix motifs from each monomer make contacts with the phosphodiester backbone on the minor groove side of each duplex arm of the junction [16,17]. The junction can be bound by a single tetramer of RuvA [15,16] or enclosed between two tetramers [18]. It is not known if binding of a single tetramer of RuvA is sufficient for branch migration by RuvAB. This reaction may require a double tetramer of RuvA or the assembly of a RuvABC resolvasome to anchor the complex [18]. In the case of the octameric RuvA complex, one of the tetramers would need to be released to permit loading of RuvC for Holliday junction resolution.

At the intersection of the grooves, negatively charged pins consisting of Glu55 and Asp56 from each monomer project towards the centre of the Holliday junction [13,14]. The four pairs of negative charges prohibit binding of duplex DNA across the centre of the tetramer and ensure high junction

Correspondence to R. G. Lloyd, Institute of Genetics, University of Nottingham, Queen’s Medical Centre, Nottingham, NG7 2UH, UK. Fax: + 0115 9709906, Tel.: + 0115 9709406, E-mail: bob.lloyd@nottingham.ac.uk Abbreviations: bio, biotin; SA, streptavidin; RU, response units. (Received 1 November 2001, revised 3 January 2002, accepted 22 January 2002)

1526 R. G. Lloyd et al. (Eur. J. Biochem. 269)

(cid:211) FEBS 2002

specificity [12]. Both acidic residues may also participate directly in the branch migration reaction by forming water- mediated contacts with unpaired bases [16]. Mutations that alter the charge on these residues stimulate the rate of branch migration and attenuate the enhanced junction resolution observed with the RuvABC complex [12].

been described [23,24], as have those for Y junction and linear duplex DNA substrates [20,24]. Gel retardation and branch migration assays used substrates in which a single strand had been 5¢ 32P-labelled using [c-32P] ATP and polynucleotide kinase prior to annealing. For SPR analysis the following oligonucleotides were used to make a 50-bp static junction, J0, labelled with biotin (bio) at the 5¢ end of one strand: 1 (bio-AAAAATGGGTCAACGTGGGCAA AGATGTCCTAGCAATGTAATCGTCTATGACGTT), 2 (GTCGGATCCTCTAGACAGCTCCATGTTCACTG GCACTGGTAGAATTCGGC), 3 (TGCCGAATTCTA CCAGTGCCAGTGAAGGACATCTTTGCCCACGTTG ACCC), 4 (CAACGTCATAGACGATTACATTG CTAC ATGGAGCTGTCTAGAGGATCCGA). A three-strand junction was made by omitting strand 4 and a 37-bp duplex DNA by annealing oligonucleotides 5 (bio-AATGCTA CAGTATCGTCCGGTCACGTACAACATCCAG) and 6 (CTGGATGTTGTACGTGACCGGACGATACTGT AGCATT).

The negatively charged pin of RuvA is conserved in almost all bacteria with the exception of three species, Mycoplasma pneumoniae, M. genitalium and Ureaplasma urealyticum. Mycoplasmas belong to the class Mollicutes and are most closely related to Gram-positive bacteria, although their genomes have experienced a drastic com- pression in size. In this work we have examined the properties of M. pneumoniae RuvA (MpRuvA) protein to determine the function of the modified pin. The interaction between MpRuvA and branched DNA substrates and its inability to form heterologous complexes with E. coli Ruv proteins reveal important differences in junction binding and processing by Mycoplasma species.

M A T E R I A L S A N D M E T H O D S

Gel retardation assays Binding mixtures (20 lL) contained 0.2 ng 32P-labelled J11, Y junction, or linear duplex DNA in 50 mM Tris/HCl pH 8.0, 5 mM EDTA, 1 mM dithiothreitol, 100 lg/mL BSA and 6% (v/v) glycerol. Samples were incubated on ice with RuvA protein for 10 min prior to loading onto a 4% polyacrylamide gel in low ionic strength buffer (6.7 mM Tris/HCl pH 8.0, 3.3 mM sodium acetate, 2 mM EDTA). In RuvAC-junction assays, RuvA was added prior to the addition of RuvC. Electrophoresis was at 160 V for 90 min with continuous buffer circulation. Gels were dried and analysed by autoradiography and phosphorimaging.

Strains and plasmids E. coli K-12 strains AB1157 (ruv+), SR2210 (ruvA200), TNM1208 (DruvAC65 rus-1) have been described previ- ously [25,29]. Strain SI171, a DruvAC65 derivative of BL21 (DE3), was used for overexpression of RuvA proteins [17]. EcRuvA was overexpressed from the pT7-7 construct, pAM159 [17]. The Mycoplasma pneumoniae M129 [30] ruvA gene was recovered by PCR from genomic DNA obtained from R. Herrmann (Universita¨ t Heidelberg, Germany). (5¢-AAACTAAGGCATATGATTGCT Oligonucleotides TCAA-3¢ and 5¢-TGCGCCTTATGGATCCGGGACG CTT-3¢) were designed to amplify the gene and provided NdeI and BamHI sites (underlined) for cloning the PCR product in pT7-7. The resulting construct, pSI66, was used for overexpression of MpRuvA. Cells were grown in LB medium supplemented with ampicillin (50 lgÆmL)1) as required for maintenance of plasmids. Sensitivity to UV light was measured as described [25].

Protein purification

Branch migration assays Reaction mixtures (20 lL) contained 0.2 ng of 32P-labelled J12 in 20 mM Tris/HCl pH 7.5, 5 mM EDTA, 2 mM dithiothreitol, 100 lgÆmL)1 BSA. RuvA protein was added before RuvB and reactions incubated at 37 (cid:176)C for 30 min and terminated by the addition of 5 lL of stop mix (2.5% SDS, 200 mM EDTA, 10 mgÆmL)1 proteinase K) with incubation for a further 10 min. Reaction products were separated on 10% PAGE in Tris/borate/EDTA buffer (89 mM Tris/HCl, pH 8.0, 89 mM borate, 2.5 mM EDTA) at 160 V for 90 min and analysed as described above.

Surface plasmon resonance

Purification of MpRuvA followed a similar protocol to that described for EcRuvA [17,31]. RuvB and RuvC proteins were overexpressed and purified as described previously [32,33]. Protein concentrations were estimated by a modified Bradford assay using a Bio-Rad assay kit and bovine serum albumin as standard. Amounts of RuvA, RuvB and RuvC are expressed as moles of the monomeric protein.

DNA substrates

Oligonucleotide synthesis was performed on an Applied Biosystems 394 DNA synthesiser using cyanoethyl phos- phoramidite chemistry. The biotin phosphoramidite was obtained from Glen Research. DNA substrates were prepared by annealing appropriate oligonucleotides follow- ing the procedure described by Parsons [34]. The sequence of oligonucleotides used for the 50 bp junctions J11 and J12, containing mobile cores of 11 and 12 bp, respectively, have

Surface plasmon resonance was performed using a BIAcore 2000TM (Uppsala, Sweden). Oligonucleotides were diluted in buffer [10 mM Hepes pH 7.4, 150 mM NaCl, 3 mM EDTA, 0.05% (v/v) surfactant P20] to a final concentration of 1 ngÆmL)1 and passed over a streptavidin (SA) sensor chip at a flow rate of 10 lLÆmin)1 until approximately 100– 200 response units (RU) of the oligonucleotide was bound to the sensor chip surface. Proteins were also diluted in Hepes/NaCl/Pi/EDTA/P20 and a range of concentrations (4–8000 nM) were injected over the DNA-charged sensor chip at a flow rate of 20 lLÆmin)1 for 3 min and allowed to dissociate for 5 min. Bound protein was removed by injecting 10 lL of 1 M NaCl. This regeneration procedure did not alter the ability of EcRuvA to bind Holliday

Mycoplasma RuvA protein (Eur. J. Biochem. 269) 1527

(cid:211) FEBS 2002

Interaction of MpRuvA with a Holliday junction

junction. Analysis of the data was performed using BIA- EVALUATION software. To remove the effects of the bulk refractive index change at the beginning and end of injections (which occur as a result of a difference in the composition of the running buffer and the injected protein), a control sensorgram obtained over the streptavidin surface was subtracted from each protein injection.

Kinetic analysis

The dissociation rate constants were calculated using linear regression analysis assuming a zero order dissociation using the equation:

dR=dt (cid:136) kdR0e(cid:255)kd(cid:133)t(cid:255)t0(cid:134) where dR/dt is the rate of change of the SPR signal, R and R0, is the response at time t and t0. kd is the dissociation rate constant.

Equilibrium binding analysis

To investigate the effect of these alterations in pin structure on the DNA binding properties of RuvA we purified the Mycoplasma pneumoniae RuvA protein and compared its activity with that of EcRuvA. The protein was overex- pressed in a DruvAC derivative of E. coli BL21 (DE3) to prevent contamination with EcRuvA and purified using the procedure devised previously. A synthetic Holliday junction containing an 11-bp mobile core was used as a substrate in gel retardation assays to assess the ability of the protein to bind junction DNA. Both MpRuvA and EcRuvA bound the junction. Each formed two distinct complexes (Fig. 2A). In the case of the E. coli protein, the two complexes represent the binding of either a single tetramer of RuvA (complex I) or of two tetramers (complex II). The data presented indicate that MpRuvA has the ability to form similar complexes. However, MpRuvA complex II appears less stable as most of the junction is found in complex I (Fig. 2A, lanes l–t). In both cases, 100 nM of protein was sufficient to bind all of the junction DNA molecules (Fig. 2A, lane j and t). Further quantitative studies revealed that MpRuvA may have a slightly higher affinity for junction DNA than EcRuvA (Fig. 2B). The kd values estimated from these data were 18 nM for MpRuvA and 42 nM for EcRuvA.

Specificity of Holliday junction binding by MpRuvA

BIAcore equilibrium binding experiments were performed as described [35] with minor modifications. The instrument was equilibrated at 25 (cid:176)C with 10 mM Hepes, pH 7.4, 150 mM NaCl, 3 mM EDTA, 0.05% (v/v) surfactant p20 at a flow rate of 100 lLÆmin)1. Baseline data were collected for 45 min at the start of the experiment, before the incorpor- ation of the protein into the running buffer. After equilib- rium binding profiles had been generated, the responses from the four flow cells were baseline corrected during the initial washing phase. The response from the reference flow cell was subtracted from the other three flow cells to correct for refractive index changes, nonspecific binding and instrument drift.

R E S U L T S

The modified pin structure of MpRuvA

The E. coli RuvA protein targets four-way junctions with high specificity [19,20]. Mutations that reduce the net charge on each subunit result in a significant increase in affinity for duplex DNA [12]. We investigated the junction specificity of MpRuvA by analysing its binding to a Y-shaped junction and to linear duplex DNA. Like the E. coli protein, MpRuvA formed two complexes with a Y junction. However, as with the four-way junction, only small amounts of complex II were detected, which again suggests that the loading of two tetramers is less favoured (Fig. 3A). No binding to linear duplex DNA was detected with EcRuvA (Fig. 3B, lanes b and c) in keeping with its high selectivity for branched molecules. However, traces of two complexes were detected with MpRuvA, even at relatively low concentrations of protein (Fig. 3B, lanes d and e).

The negatively charged pin of E. coli RuvA (EcRuvA) has two acidic residues (Glu55 and Asp56) flanked by b sheets [13,14]. This arrangement is conserved in the RuvA sequences from 45 other bacterial species [12] (Fig. 1A and data not shown), which suggests that the pin architectures are probably very similar, as demonstrated for Mycobacte- rium leprae RuvA [18]. However, three bacterial species (M. pneumoniae, M. genitalium, and Ureaplasma urealyti- cum) carry RuvA orthologs in which the sequences forming the pin region differ significantly from this pattern (Fig. 1B). These RuvA proteins have an additional four amino acids and lack acidic residues at the apex of the intervening loop. Acidic residues that may potentially compensate for the loss of the negative charge are located nearby in the two Mycoplasma sequences but are positioned in the region corresponding to b sheet 6 in the EcRuvA structure [17]. The global structure of the two RuvA proteins would have to be radically altered to accommodate these residues in the same position as in EcRuvA. In addition, only one acidic residue is retained in U. urealyticum RuvA (Fig. 1B). However, con- servation of sequences in the flanking b sheets suggests that the general architecture of the pin is probably maintained. Thus, the likely net effect of the altered sequence between b5 and b6 is to produce an extended and uncharged pin.

To analyse the structure specificity of MpRuvA more quantitatively we made use of surface plasmon resonance. Biotinylated DNA substrates [a Holliday junction (J0) lacking a homologous core, a three-strand derivative of J0, and duplex DNA] were immobilized on different flow cells on a streptavidin sensor chip. The binding of EcRuvA and MpRuvA to these substrates was examined and the results are shown in Fig. 3C,D. EcRuvA showed the expected preference for Holliday junction DNA over both three- strand and duplex DNA as evident from the gradient of dissociation illustrated on the sensorgram (Fig. 3C). Disso- ciation rate constants were calculated using the equation described in Materials and methods. Whilst this equation may not fit the entire range of protein concentrations under all of the experimental conditions described here, it repre- sents the best case scenario, as the analysis is comparative in nature and describes the net stability of the protein:DNA complex. The rate constants reveal a three to fourfold difference between the linear duplex/three-strand substrates

1528 R. G. Lloyd et al. (Eur. J. Biochem. 269)

(cid:211) FEBS 2002

gel retardation assays confirming that MpRuvA has a reduced specificity for Holliday junctions. SPR analysis also shows that the EcRuvA and MpRuvA bind to the DNA with fast association rate constants (ka). This results in mathematical models that poorly fit the data, and calcula- tions using ka and kd to obtain the equilibrium dissociation constant would be erroneous.

Equilibrium binding analysis was performed to further analyse the interaction of MpRuvA with Holliday junction and duplex DNA (Fig. 4). RuvA protein was placed directly in the running buffer and continually passed over the sensor chip surface containing duplex or Holliday junction attached to different flow cells. The binding profile of the MpRuvA interaction with these DNA substrates is shown in Fig. 4A. The sensorgram reveals that MpRuvA protein, like EcRuvA (Fig. 4B), binds with high affinity to the

and Holliday junction bound by EcRuvA (Table 1), illustrating the additional stability of the Holliday junc- tion-RuvA complex. The binding of the MpRuvA is shown in Fig. 3D and shows little difference in the dissociation rate constants for the three different DNA-MpRuvA complexes, demonstrating that these complexes have equal stabilities. Figure 3E shows a direct comparison of the binding of EcRuvA and MpRuvA to linear duplex DNA and shows the additional stability of the MpRuvA bound DNA complex compared to the EcRuvA bound DNA complex. But the results also show an increase in the amount of MpRuvA binding to duplex DNA compared to EcRuvA, as indicated by the response (Figs 3C–E). MpRuvA also formed a complex with a short 24 bp duplex that was not bound detectably by EcRuvA (data not shown). The SPR data are broadly consistent with the results obtained from

Fig. 1. The modified pin structure of MpRuvA. (A) Structure of the EcRuvA- Holliday junction DNA complex [15]. A tetramer of RuvA (opposing monomers are in shades of grey) binds the Holliday junction in an open square conformation. The duplex arms of the junction are bound in grooves on the concave surface of the protein and converge at a centrally located pin structure formed by Glu55 and Asp56 (red) in each RuvA subunit. (B) Alignment of bacterial RuvA proteins showing conservation of the pin region. Residues 42–65 of EcRuvA are aligned with homologous sequences from selected bacterial species. Residues conserved in the majority of RuvA sequences from 46 bacteria (data not shown) are highlighted in bold. Arrows denote the position of b sheets 5 and 6 in the EcRuvA structure [14,17]. Acidic pin residues are highlighted in red, as are negatively charged residues located nearby in the RuvA sequences from M. pneumoniae, M. genitalium, and U. urealyticum.

Mycoplasma RuvA protein (Eur. J. Biochem. 269) 1529

(cid:211) FEBS 2002

Holliday junction at relatively low concentrations of protein (0.112 and 1.12 nM). Binding to duplex DNA is not observed until a concentration of 11.2 nM is passed over the sensor chip surface (Fig. 4A). Significantly, these results reveal that MpRuvA has a higher affinity for duplex DNA than the EcRuvA protein. Binding of EcRuvA to duplex DNA is not evident until a concentration of 90.4 nM is reached (Fig. 4B). Thus MpRuvA bound to the duplex at a 10-fold lower concentration and assuming the mechanism and mode of binding is the same, the MpRuvA has a 10-fold higher affinity for duplex DNA. Despite this difference, MpRuvA retains Holliday junction specificity with similar kinetics and stoichiometry as EcRuvA.

MpRuvA is unable to interact with E.coliRuvB and RuvC proteins

Fig. 2. Holliday junction binding by MpRuvA. (A) Gel retardation assay showing the formation of complexes I and II with junction J11. Binding mixtures contained 0.2 ng 32P-labelled J11 DNA and 0, 0.1, 0.5, 1, 2, 5, 10, 20, 50, and 100 nM of EcRuvA (lanes a–j) or MpRuvA (lanes k–t) proteins. (B) Titration of MpRuvA and EcRuvA showing the relative binding of J11. Values are the mean of two independent experiments and are based on the fraction of the total DNA bound.

The coupling of branch migration and resolution medi- ated by the E. coli RuvABC resolvasome complex requires the binding of RuvA to one face of the junction and RuvC to the other [8,9]. Complexes formed by the loading of both RuvA and RuvC on a synthetic junction can be detected using a gel retardation assay [23]. We used such an assay to

RuvA and RuvB mediate the branch migration of Holliday junctions and in vitro promote the dissociation of synthetic junction substrates to yield flayed duplex products [19]. We examined MpRuvA to see if it could form a branch migration complex with E. coli RuvB. Heterologous branch migration activity has previously been demonstrated using M. leprae RuvA with E. coli RuvB [21] and E. coli RuvA with Thermus thermophilus RuvB [22]. MpRuvA was incubated with E. coli RuvB and synthetic Holliday junc- tion J12 in reactions containing Mg2+ and ATP (Fig. 5A, lanes j–p). In contrast to reactions containing EcRuvA (Fig. 5A, lanes b–h), no unwinding of the synthetic Holliday junction was detected in reactions containing MpRuvA. Similar results were obtained using other junctions differing in sequence and length of mobile core (data not shown). The results indicate that MpRuvA is unable to form a functional branch migration complex with E. coli RuvB.

Fig. 3. Interaction of MpRuvA with branched DNA structures and lin- ear duplex molecules. (A) Gel retardation assay showing binding of RuvA proteins to a Y-junction DNA substrate. Reactions contained 0.2 ng 32P-labelled DNA and RuvA at 2 nM (lanes b and d) or 20 nM (lanes c and e). (B) Gel retardation assay showing binding of RuvA to linear duplex DNA. Reactions contained 0.2 ng proteins 32P-labelled DNA and RuvA at 10 nM (lanes b and d) or 100 nM (lanes c and e). (C) Surface plasmon resonance sensorgram showing binding of EcRuvA (8 lM) to duplex, three-strand and Holliday junction DNA. (D) Surface plasmon resonance sensorgram showing binding of MpRuvA (6.4 lM) to duplex, three-strand and J0 DNA. (E) Surface plasmon resonance sensogram showing the binding of EcRuvA (6 lM) and MpRuvA (4 lM) to duplex DNA.

1530 R. G. Lloyd et al. (Eur. J. Biochem. 269)

(cid:211) FEBS 2002

Table 1. Dissociation rates for MpRuvA and EcRuvA-DNA complexes.

Dissociation rate constant (kd) (1/s) (cid:139) SDa

DNA MpRuvA EcRuvA

a Determined using surface plasmon resonance analysis.

Holliday junction Three-strand junction Duplex 6.1 · 10)4 (cid:139) 2.2 · 10)5 7.4 · 10)4 (cid:139) 3.0 · 10)5 6.2 · 10)4 (cid:139) 2.3 · 10)5 5.5 · 10)4 (cid:139) 4.2 · 10)5 17 · 10)4 (cid:139) 1.9 · 10)4 19 · 10)4 (cid:139) 2.2 · 10)4

Fig. 5. Interactions between RuvA and either RuvB or RuvC. (A) Branch migration assay showing the dissociation of Holliday junction to flayed duplex products. Reactions contained 0.2 ng 32P-labelled J12 DNA and proteins as indicated. (B) Gel retardation assay showing the formation of RuvAC-junction complexes. Binding mixes contained 0.2 ng 32P-labelled J12 DNA and proteins as indi- cated.

encoding MpRuvA or EcRuvA into E. coli strains SR2210 (ruvA200) and the ruv+ control, AB1157. The plasmid expressing MpRuvA (pSI66) failed to improve the UV sensitivity of the ruvA mutant SR2210 (Fig. 6A), which is not surprising given that MpRuvA fails to form productive interactions with E. coli RuvB or RuvC. Indeed, survival was actually reduced. This negative effect is most likely due to MpRuvA blocking the access of other junction process- ing enzymes such as RecG [24]. Expression of MpRuvA also reduced survival of the ruv+ AB1157 strain (Fig. 6B). However, the effect was rather modest and we conclude that overexpression of MpRuvA does not interfere significantly with junction processing by the resident E. coli RuvABC system.

investigate whether E. coli RuvC could bind a junction already bound by MpRuvA. With 200 nM RuvC and low concentrations of EcRuvA, a RuvA/junction/RuvC com- plex was visualized (Fig. 5B, lanes c and d). No such complex could be detected using MpRuvA (Fig. 5B, lanes l–r). The only complexes seen were those formed by the binding of RuvC alone or of a double tetramer of MpRuvA (complex II). The absence of MpRuvA complex I may be significant, especially as this is the predominant complex formed in the absence of RuvC (Fig. 2A). It is possible that such complexes do bind RuvC but that such binding destabilizes the RuvA–junction interaction.

Effect of MpRuvA on DNA repair in E.coliruv mutants

To further investigate the ability of MpRuvA to block junction processing in vivo, we made use of strain TNM1208 (DruvAC rus-1). This strain lacks the RuvABC resolution pathway due to deletion of the ruvA and ruvC genes. However, it is resistant to UV light because the rus-1 mutation activates an alternative resolvase (RusA) that is able to process Holliday junctions very efficiently in the

The ability of MpRuvA protein to promote DNA repair in vivo was investigated by introducing plasmid constructs

Fig. 4. Equilibrium binding profiles of MpRuvA and EcRuvA on Holl- iday junction (J0) and linear duplex DNA substrates. (A) MpRuvA was incorporated in the running buffer at concentrations of 0.0112 nM (a), 0.112 nM (b), 1.12 nM (c) and 11.2 nM (d). (B) EcRuvA was incor- porated in the running buffer at concentrations of 0.00904 nM (a), 0.0904 nM (b), 0.904 nM (c), 9.04 nM (d) and 90.4 nM (e). The arrows indicate the time at which the concentration of the protein was altered.

Mycoplasma RuvA protein (Eur. J. Biochem. 269) 1531

(cid:211) FEBS 2002

MpRuvA. However the octameric complex (complex II) appears less stable than that formed with EcRuvA. As the pin region of MpRuvA contains an additional four amino acids it is likely that the pin is extended and this extension could cause steric clashes across the central cavity of the open Holliday junction that interfere with stable binding of a tetramer on both faces. The reduced stability of the octamer complex may explain the modest negative effect of MpRuvA compared with EcRuvA on DNA repair mediated by the RusA resolvase in strain TNM1208 (Fig. 6C). This protein forms a very stable octameric complex and when overexpressed is therefore much more likely to prevent RusA gaining access to a Holliday junction. Single tetramers of EcRuvA and MpRuvA bind junction DNA with similar affinities. However, such complexes are less likely to inhibit RusA as one face of the junction would remain free of protein and this may be sufficient for RusA to load on the DNA and resolve the structure.

absence of RuvABC [25–27]. The introduction of a plasmid expressing EcRuvA into this strain increases sensitivity to UV light (Fig. 6C), presumably by blocking Holliday junction resolution by RusA [17]. The plasmid encoding MpRuvA also increases sensitivity to UV, but the effect is considerably less severe (Fig. 6C). This finding suggests that MpRuvA is less able to inhibit the processing of Holliday junctions in vivo than EcRuvA despite the fact that both bind synthetic Holliday junctions with similar affinities in vitro (Fig. 2A).

Fig. 6. Survival of UV-irradiated Escherichia coli strains carrying plasmids expressing either MpRuvA or EcRuvA proteins. (A) Strain SR2210 (ruvA200). (B) Strain AB1157 (ruv+). (C) Strain TNM1208 (ruvAC rus-1). The plasmid constructs used are identified in (B). Values are the mean of at least two independent experiments.

D I S C U S S I O N

We found that MpRuvA is unable to promote DNA repair in E. coli ruvA mutants. This is most likely a consequence of its failure to assemble a functional branch migration complex with E. coli RuvB. Certain conserved residues in domain III of EcRuvA (Leu167, Leu170, Tyr172 and Leu199) are known to participate in protein–protein interactions with EcRuvB [11,14]. MpRuvA has the first three of these residues but differs in the replacement of Leu199 with isoleucine. It is possible that this subtle change accounts for the inactivity of the hybrid MpRuvA-EcRuvB branch migration motor, although other differences affect- ing the architecture of MpRuvA domain III cannot be excluded. Mycobacterium leprae RuvA, which retains a conserved leucine at this position, forms an active branch migration complex with EcRuvB [21]. Compensatory changes in the MpRuvB sequence should correspond to the alterations in MpRuvA that prevent heterologous contacts with EcRuvB. Isoleucine residues at positions 148 and 150 in EcRuvB are critical for the formation of complexes with EcRuvA [28]. In MpRuvB these amino acids are replaced by the alternative hydrophobic residues, valine and methionine, respectively. These substitutions at the MpRuvA–RuvB interface are likely to be responsible for blocking the formation of functional complexes between MpRuvA and EcRuvB.

The negatively charged central pin on the DNA binding surface of RuvA plays a crucial role in junction targeting and processing. It constrains the rate of branch migration by RuvAB and influences resolution by RuvABC [12]. The importance of this structure is reflected in the high conservation of the sequences forming the pin in the majority of bacteria with the exception of two Mycoplasma species and one of Ureaplasma. In the RuvA proteins from these organisms the pin sequence is extended by four residues and lacks negatively charged residues at the apex of the structure. We investigated the properties of the RuvA protein from M. pneumoniae to see how these modifications affected its interaction with DNA.

The MpRuvA protein bound the four-way branched Holliday junction structure with a high affinity. However, relative to EcRuvA, it displayed an increased affinity for Y-shaped duplex DNA structures, three-strand junctions and linear duplex DNA. Its affinity for linear duplex DNA is approximately 10-fold higher than the E. coli protein. The results suggest that the modified pin influences the ability to bind duplex DNA and is consistent with observations by Ingleston et al. [12] showing that mutations in EcRuvA that reduce the net negative charge on the pin, or which add positive charges, result in an increase in binding to duplex DNA.

As with the E. coli protein, we found that a synthetic Holliday junction can bind either one or two tetramers of

We also found that E. coli RuvC was unable to form a complex with a junction already bound by MpRuvA, at least not one stable enough to be detected in a gel retardation assay. In common with other Gram-positive bacteria, M. pneumoniae lacks a homologue of RuvC [6]. It is therefore possible that branch migration and resolution are uncoupled in these species [18]. The assembly of a RuvABC complex is necessary for efficient resolution of Holliday junctions in E. coli and presumably imposes constraints on the evolution of each Ruv protein. In particular, RuvA may have to maintain a compact acidic pin that does not project at the junction core so that the conformation of the RuvA-bound junction allows stable loading of RuvC. In the absence of a RuvC, the constraints on MpRuvA would be reduced and limited to those factors necessary for junction binding and the loading of RuvB. However, several Gram-positive bacteria that lack RuvC apparently retain the conserved pin architecture of EcRuvA [6,12]. In fact, M. pulmonis RuvA has a pin that more

1532 R. G. Lloyd et al. (Eur. J. Biochem. 269)

(cid:211) FEBS 2002

15. Hargreaves, D., Rice, D.W., Sedelnikova, S.E., Artymiuk, P.J., Lloyd, R.G & Rafferty, J.B. (1998) Crystal structure of E. coli RuvA with bound DNA Holliday junction at 6 A˚ resolution. Nat. Struct. Biol. 5, 441–446.

closely resembles the standard pattern rather than its closely related Mollicutes (Fig. 1B). In addition, M. leprae RuvA, which has an apparently identical pin to EcRuvA, also fails to form junction complexes with EcRuvC in a gel retarda- tion assay, perhaps suggesting that there are stabilizing contacts across the junction that are independent of pin structure [21]. Clearly there are subtle differences in the way Holliday junctions are processed by Mycoplasmas. Further insights into the mechanism of Holliday junction branch migration and resolution await the identification and characterization of the novel resolvase employed in Gram- positive eubacteria.

16. Ariyoshi, M., Nishino, T., Iwasaki, H., Shinagawa, H & Morikawa, K. (2000) Crystal structure of the Holliday junction DNA in complex with a single RuvA tetramer. Proc. Natl Acad. Sci. USA 97, 8257–8262.

17. Rafferty, J.B., Ingleston, S.M., Hargreaves, D., Artymiuk, P.J., Sharples, G.J., Lloyd, R.G & Rice, D.W. (1998) Structural simi- larities between Escherichia coli RuvA and other DNA-binding proteins and a mutational analysis of its binding to the Holliday junction. J. Mol. Biol. 278, 105–116.

A C K N O W L E D G E M E N T S

18. Roe, S.M., Barlow, T., Brown, T., Oram, M., Keeley, A., Tsaneva, I.R & Pearl, L.H. (1998) Crystal structure of an octameric RuvA-Holliday junction complex. Mol. Cell. 2, 361– 372.

We thank Richard Herrmann for Mycoplasma pneumoniae genomic DNA. This work was supported by grants from the Biotechnology and Biological Sciences Research Council, the Wellcome Trust, the Royal Society, and the Medical Research Council. M. J. D. was in receipt of a Prize Studentship from the Wellcome Trust.

19. Parsons, C.A., Tsaneva, I., Lloyd, R.G & West, S.C. (1992) Interaction of E. coli RuvA and RuvB proteins with synthetic Holliday junctions. Proc. Natl Acad. Sci. USA 89, 5452–5456. 20. Lloyd, R.G & Sharples, G.J. (1993) Processing of recombination intermediates by the RecG and RuvAB proteins of Escherichia coli. Nucleic Acids Res. 21, 1719–1725.

R E F E R E N C E S

1. Seigneur, M., Bidneko, V., Ehrlich, S.D & Michel, B. (1998) RuvAB acts at arrested replication forks. Cell 95, 419–430. 21. Arenas-Licea, J., van Gool, A.J., Keeley, A.J., Davies, A., West, S.C & Tsaneva, I.R. (2000) Functional interactions of Myco- bacterium leprae RuvA with Escherichia coli RuvB and RuvC on Holliday junctions. J. Mol. Biol. 301, 839–850.

2. McGlynn, P & Lloyd, R.G. (2000) Modulation of RNA poly- merase by (p)ppGpp reveals a RecG-dependent mechanism for replication fork progression. Cell 101, 35–45.

22. Yamada, K., Fukuoh, A., Iwasaki, H & Shinagawa, H. (1999) Novel properties of the Thermus thermophilus RuvB protein, which promotes branch-migration of Holliday junctions. Mol. Gen. Genet. 261, 1001–1011. 3. Kowalczykowski, S.C. (2000) Initiation of genetic recombination and recombination-dependent replication. Trends Biochem. Sci. 25, 156–165.

23. Whitby, M.C., Bolt, E.L., Chan, S.N & Lloyd, R.G. (1996) Interactions between RuvA and RuvC at Holliday junctions: inhibition of junction cleavage and formation of a RuvA–RuvC– DNA complex. J. Mol. Biol. 264, 878–890. 4. Singleton, M.R., Scaife, S & Wigley, D.B. (2001) Structural ana- lysis of DNA replication fork reversal by RecG. Cell 107, 79–89. 5. West, S.C. (1997) Processing of recombination intermediates by the RuvABC proteins. Annu. Rev. Genet. 31, 213–244.

24. Lloyd, R.G & Sharples, G.J. (1993) Dissociation of synthetic Holliday junctions by E. coli RecG protein. EMBO J. 12, 17–22.

6. Sharples, G.J., Ingleston, S.M & Lloyd, R.G. (1999) Holliday junction processing in bacteria: insights from the evolutionary conservation of RuvABC, RecG, and RusA. J. Bacteriol. 181, 5543–5550.

25. Mandal, T.N., Mahdi, A.A., Sharples, G.J & Lloyd, R.G. (1993) Resolution of Holliday intermediates in recombination and DNA repair: indirect suppression of ruvA, ruvB and ruvC mutations. J. Bacteriol. 175, 4325–4334. 7. Parsons, C.A., Stasiak, A., Bennett, R.J & West, S.C. (1995) Structure of a multisubunit complex that promotes DNA branch migration. Nature 374, 375–378. 8. Davies, A.A & West, S.C. (1998) Formation of RuvABC-Holliday junction complexes in vitro. Curr. Biol. 8, 725–727.

9. Zerbib, D.M., E` zard, C., George, H & West, S.C. (1998) Co-ordinated actions of RuvABC in Holliday junction processing. J. Mol. Biol. 281, 621–630.

26. Sharples, G.J., Chan, S.C., Mahdi, A.A., Whitby, M.C & Lloyd, R.G. (1994) Processing of intermediates in recombination and DNA repair: identification of a new endonuclease that spe- cifically cleaves Holliday junctions. EMBO J. 13, 6133–6142. 27. Mahdi, A.A., Sharples, G.J., Mandal, T.N & Lloyd, R.G. (1996) Holliday junction resolvases encoded by homologous rusA genes in Escherichia coli K-12 and phage 82. J. Mol. Biol. 257, 561–573. 10. Eggleston, A.K & West, S.C. (2000) Cleavage of Holliday junc- tions by the Escherichia coli RuvABC complex. J. Biol. Chem. 275, 26467–26476.

11. Nishino, T., Iwasaki, H., Kataoka, M., Ariyoshi, M., Fujita, T., Shinagawa, H & Morikawa, K. (2000) Modulation of RuvB function by the mobile domain III of the Holliday junction recognition protein RuvA. J. Mol. Biol. 298, 407–416. 28. Han, Y.-W., Iwasaki, H., Miyata, T., Mayanagi, K., Yamada, K., Morikawa, K & Shinagawa, H. (2001) A unique b-hairpin pro- truding from AAA+ ATPase domain of RuvB motor protein is involved in the interaction with RuvA DNA recognition protein for branch migration of Holliday junctions. J. Biol. Chem. 276, 35024–35028.

12. Ingleston, S.M., Sharples, G.J & Lloyd, R.G. (2000) The acidic pin of RuvA modulates Holliday junction binding and processing by the RuvABC resolvasome. EMBO J. 19, 6266–6274. 29. Sargentini, N.J & Smith, K.C. (1989) Role of ruvAB genes in UV- and c-radiation and chemical mutagenesis in Escherichia coli. Mutation Res. 215, 115–129.

30. Himmelreich, R., Hilbert, H., Plagens, H., Pirkl, E., Li, B.C & Herrmann, R. (1996) Complete sequence analysis of the genome of the bacterium Mycoplasma pneumoniae. Nucleic Acids Res. 24, 4420–4449. 13. Rafferty, J.B., Sedelnikova, S.E., Hargreaves, D., Artymiuk, P.J., Baker, P.J., Sharples, G.J., Mahdi, A.A., Lloyd, R.G & Rice, D.W. (1996) Crystal structure of DNA recombination protein RuvA and a model for its binding to the Holliday junction. Science 274, 415–421.

(1998) Functional analyses of

31. Sedelnikova, S.E., Rafferty, J.B., Hargreaves, D., Mahdi, A.A., Lloyd, R.G & Rice, D.W. (1997) Crystallisation of E. coli RuvA gives insights into the symmetry of a Holliday junction/protein complex. Acta Crystallogr. D53, 122–124. 14. Nishino, T., Ariyoshi, M., Iwasaki, H., Shinagawa, H & Morikawa, K. the domain structure in the Holliday junction binding protein RuvA. Structure 6, 11–21.

Mycoplasma RuvA protein (Eur. J. Biochem. 269) 1533

(cid:211) FEBS 2002

34. Parsons, C.A., Kemper, B & West, S.C. (1990) Interaction of a four-way junction in DNA with T4 endonuclease VII. J. Biol. Chem. 265, 9285–9289. 32. Dunderdale, H.J., Sharples, G.J., Lloyd, R.G & West, S.C. (1994) Cloning, overexpression, purification and characterization of the Escherichia coli RuvC Holliday junction resolvase. J. Biol. Chem. 269, 5187–5194.

35. Myszka, D.G., Jonsen, M.D & Graves, B.J. (1998) Equilibrium analysis of high a(cid:129)nity interactions using BIAcore. Anal. Biochem. 265, 326–330. 33. Tsaneva, I.R., Illing, G.T., Lloyd, R.G & West, S.C. (1992) Purification and physical properties of the RuvA and RuvB pro- teins of Escherichia coli. Mol. Gen. Genet. 235, 1–10.