Structural and Genetic Analyses Reveal a Key Role in Prophage Excision for the TorI Response Regulator Inhibitor*

TorI (Tor inhibition protein) has been identified in Escherichia coli as a protein inhibitor acting through protein-protein interaction with the TorR response regulator. This interaction, which does not interfere with TorR DNA binding activity, probably prevents the recruitment of RNA polymerase to the torC promoter. In this study we have solved the solution structure of TorI, which adopts a prokaryotic winged-helix arrangement. Despite no primary sequence similarity, the three-dimensional structure of TorI is highly homologous to the λXis, Mu bacteriophage repressor (MuR-DBD), and transposase (MuA-DBD) structures. We propose that the TorI protein is the structural missing link between the λXis and MuR proteins. Moreover, in vivo assays demonstrated that TorI plays an essential role in prophage excision. Heteronuclear NMR experiments and site-directed mutagenesis studies have pinpointed out key residues involved in the DNA binding activity of TorI. Our findings suggest that TorI-related proteins identified in various pathogenic bacterial genomes define a new family of atypical excisionases.

TorI (Tor inhibition protein) has been identified in Escherichia coli as a protein inhibitor acting through protein-protein interaction with the TorR response regulator. This interaction, which does not interfere with TorR DNA binding activity, probably prevents the recruitment of RNA polymerase to the torC promoter. In this study we have solved the solution structure of TorI, which adopts a prokaryotic winged-helix arrangement. Despite no primary sequence similarity, the three-dimensional structure of TorI is highly homologous to the Xis, Mu bacteriophage repressor (MuR-DBD), and transposase (MuA-DBD) structures. We propose that the TorI protein is the structural missing link between the Xis and MuR proteins. Moreover, in vivo assays demonstrated that TorI plays an essential role in prophage excision. Heteronuclear NMR experiments and site-directed mutagenesis studies have pinpointed out key residues involved in the DNA binding activity of TorI. Our findings suggest that TorI-related proteins identified in various pathogenic bacterial genomes define a new family of atypical excisionases.
Lambdoid phages, exemplified by itself, constitute a wide family of temperate phages, which genomes are found either as a circular doublestranded DNA, or as a prophage integrated into the host chromosome. The phage-encoded integrase ( Int) catalyzes both integration and excision reactions helped in these functions by several accessory proteins (1). Among them two are absolutely required, the host-encoded integration host factor (IHF) 3 is required for integration and excision, whereas the phage-encoded excisionase ( Xis) is necessary for excision only. Excisionase proteins are also called recombination directionality factors (RDFs), (2) since their role is to control the activity of the integrase and to direct the reaction toward excision. Xis plays a critical role during excision by allowing the formation of a specific complex called intasome together with the integrase, IHF, and a third accessory protein Fis (3). Simultaneously to the excision of the prophage genome, Xis also inhibits reintegration by converting the phage attachment site (attP) into a catalytically inactive structure (4). Xis functions require binding to the integrase as well as to DNA in the attR region at two tandemly arranged binding sites (X1 and X2). Binding to these sites promotes sharp bending of DNA and assist the integrase DNA binding to allow intasome formation (5,6).
The TorI protein was first identified as a TorR response regulator inhibitor (7). TorR is part of the two-component system TorS/TorR required for torCAD operon expression in Escherichia coli (8,9). In response to the presence of trimethylamine-N-oxide (TMAO) in the environment, the TorS sensor kinase autophosphorylates and transfers a phosphoryl group to TorR through a four-step phosphorelay leading to the induction of the torCAD operon, which encodes the TMAO reductase respiratory system (10). The complex phosphorelay occurring between TorS and TorR led us to suspect the presence of intermediate checkpoints in the TMAO signal transduction pathway. Indeed, the torI gene has been identified as a negative regulator of the torCAD operon using a genetic multicopy approach. The negative effect was due to a previously unidentified small open reading frame (66 amino acids) that we called torI for Tor inhibition. Further studies showed that TorI does not interfere with the phosphorelay but rather acts at the level of the TorR response regulator. Interestingly, we showed that TorI binds to the C-terminal domain of TorR without affecting its DNA binding capacity, and we proposed that TorI prevents the recruitment of RNA polymerase to the torC promoter (7). So far, TorI is a unique case of response regulator inhibitor acting through protein-protein interaction with the DNA binding domain of a response regulator without interfering with its DNA binding activity.
A look for TorI homologues on finished and unfinished bacterial genomes led us to find two categories of homologous proteins. The first one contained proteins that show 100% identity with TorI. These proteins are the products of gene hkaC in the coliphage HK620 genome and gene 18 in the genome of the Shigella flexneri phage Sf6 (11,12). To date, no biological function has been assigned to these predicted proteins. In the second category of homologous proteins were found several proteins with 25-35% identity to TorI, present in various pathogenic bacteria such as uropathogenic E. coli O157:H7, Yersinia pseudotuberculosis, Vibrio cholerae, and S. flexneri. Analysis of the DNA sequences surrounding the genes of the TorI homologues revealed that most of these genes are located near a phage integrase encoding gene, in a pathogenic island, or in a characterized prophage region. The analysis of the genetic context of torI indicated that it also belonged to the defective prophage KplE1 genome sequence (7,13). All of these proteins share a small size (less than 80 residues) and a high proportion of basic residues, which are typical characteristics of RDF proteins (2). Indeed, two homologues of TorI have been recently described as pathogenic island excisionases, namely Hef and Rox in Y. pseudotuberculosis, and S. flexneri, respectively (14,15). All of these data led us to suspect a role for TorI in the excision of the KplE1 cryptic prophage in E. coli. In this study, we show that TorI is capable of excisionase activity in vivo and presents a three-dimensional fold highly similar to that of both Xis and Mu repressor proteins. We also define the structural features of a new family of atypical excisionases.

MATERIALS AND METHODS
DNA Manipulations-Small-scale plasmid extractions were carried out by using the Miniprep plasmid kit (Promega), and DNA fragments were purified with the Qiagen PCR purification kit (Qiagen Inc.). DNA sequencing was performed on purified plasmids and PCR products at MWG Biotech.
Strain Construction-Strain LCB970 is a derivative of strain MC4100 (Casadaban) and was constructed by insertion of the chloramphenicol acetyl transferase (cat) gene in the KplE1 prophage between yfdO and yfdP according to the method of Datsenko and Wanner (16). Briefly, the cat gene was PCR-amplified using pKD3 as a template with the following primers: KplE1-Cm1 (5Ј-CAGGCGAATTTCGTTTGCCCAGGC-TGTCCAGTTCGGTTCTGTGTAGGCTGGAGCTGCTTC) and KplE1-Cm2 (5Ј-AGCAGGCCGCCGAATGTGACGGCGAGGTGG-TTCGTCCCAACATATGAATATCCTCCTTAG), where the underlined sequences are homologous to the DNA sequence within the KplE1 prophage to allow site-specific recombination by the -Red recombination system. The resulting PCR product was then transformed into a strain containing pKD20, which encodes the Red recombinase under an arabinose inducible promoter. The presence of the cat gene was then verified by PCR amplification of the chromosomal region (the sequence of the primers is available upon request to the authors).
Excision Test-Strain LCB970 carrying torI encoding plasmids pJFi (7) was grown in LB medium until the OD 600 reached 0.5 units, and IPTG (1 mM) was added for 2 h at 37°C under agitation. Culture dilutions were prepared and plated onto rich medium containing either 50 g/ml ampicillin or 5 g/ml chloramphenicol. Numeration of the colonies plated on both antibiotics was performed and the ratio of ampicillin-resistant/chloramphenicol-resistant colonies was calculated. Values represent the average of at least three independent determinations. To confirm prophage DNA excision, a PCR test was performed on randomly chosen colonies plated onto ampicillin. A control colony containing the empty vector was included in the PCR assay. A basic PCR amplification was performed (30 s denaturation at 94°C, 30 s annealing at 55°C, 30 s elongation at 72°C) using the GoTaq DNA polymerase (Promega) and the following primers: ptorI1 (5Ј-GAGCCATACAGC-CTCACACTCGATGAGG) inside KplE1 prophage and Ext3ЈKplE1 (5Ј-CTTATTCGGCCTGCTAGTTCG) outside of the prophage.
Site-directed Mutagenesis-Mutagenesis of residues Tyr 28 and Arg 45 was performed as described previously (17). Briefly, the entire pJFi plasmid (pJF119EHtorI) was PCR-amplified with divergent overlapping primer pairs, one primer of each pair carrying the desired mutation, and a high fidelity thermostable DNA polymerase (Expand High Fidelity DNA, Roche Diagnostic). Mutations of Tyr 28 into Phe or Ser and Arg 45 into Gln or Lys were performed by introducing degenerate codons TYC and MAA (where Y indicates A or C, and M indicates C or T), respectively, at the desired positions in one primer per pair. PCR reaction was conducted with 200 ng of template plasmid, 200 M concentration of each dNTPs, and 5 units of DNA polymerase, in the presence of 5% Me 2 SO, and only 10 amplification cycles were performed to avoid nondesired mutation. The template plasmid was then hydrolyzed by addition of 20 units of the DpnI enzyme (Biolabs), and the purified PCR products were directly transformed into a recA strain (JM109) to allow homologous recombination on both sides of the PCR product to generate a mutated circular plasmid. The resultant plasmids were then purified and sequenced (MWG Biotech) to check for the presence of the desired mutations and the absence of additional mutation in the rest of the torI gene and transformed into the test strain LCB970. To check that TorI mutants were produced and stable in strain LCB970, crude extracts of LCB970 overproducing TorI wild type and mutants were prepared and submitted to Western blot analysis using a TorI antiserum.
TorI Labeling and Purification-TorI protein was produced in BL21(DE3) cells harboring plasmid pETsI (7). To obtain the doublelabeled protein, cells were grown in M9 minimal medium supplemented with 2 g/liter [ 13 C]glucose (Eurisotop) and/or 1 g/liter [ 15 N]NH 4 Cl until the OD 600 reached 0.8 units, and IPTG (1 mM) was added for 2 h at 37°C. French-pressed cell lysate extract was equilibrated with 40 mM Tris buffer (pH 7.4) and loaded onto a HiTrap SP column (Amersham Biosciences). The protein was eluted with a step gradient of KCl and was found in the 0.3 M KCl-containing fraction.
NMR Measurements, Assignment, and Interaction with DNA-The HSQC spectra were first compared at 278, 283, 288, and 298 K to pick up the best stability/intensity ratio. The NMR experiments were then carried out at 278K on a 500 MHz Bruker DRX spectrometer and on a 800 MHz Varian Inova spectrometer equipped with a triple resonance ( 1 H, 15 (18 -21). 3 J HN␣ coupling constants were measured using three-dimensional HNHA experiments (22). NOE distance constraints were obtained from a two-dimensinoal NOESY, three-dimensional NOESY ( 15 N, 1 H)-HSQC spectra with mixing times of 100 and 120 ms. All NMR spectra were processed using XWin-NMR (Bruker) and analyzed using Felix (Accelrys). The DNA sequence used for the titration experiments corresponds to 5Ј-GGGTAAAATA (Fig. 3). Two 10-base complementary oligonucleotides (MWG Biotech) were annealed in 10 mM Tris-HCl, 300 mM NaCl, and ethanol-precipitated. The double strand DNA was then resuspended in 40 mM phosphate (pH 5.9), 100 mM NaCl at a concentration of 0.1 M. The TorI-DNA complex formation was monitored by recording a series of two-dimensional 1 H- 15  Chemical Shift-derived Restraints-The C␣, C␤, Co, H␣, and N chemical shifts of 36 residues served as input for the TALOS program (23). TALOS derives information on the and backbone dihedral angles from a comparison of secondary chemical shift corresponding to known conformations. A conservative approach was chosen requiring that all 10 best matches agree for a prediction to be accepted. The TALOS predictions were converted into dihedral angle restraints as the average and angles Ϯ 2 S.D. or a minimum of Ϯ 10°.
Structure Calculations-Distance restraints were obtained from twodimensional NOESY and three-dimensional 15 N-edited NOESY experiments (with mixing times of 100 and 120 ms). angle dihedral restraints were estimated from 3 J HN-H␣ coupling constants and use of the Karplus equation. TALOS-derived dihedral angle constraints were used as described above. Structure calculations were performed with CNS/ARIA (default parameter with the Amersham Bioscience steps parameter *8 to increase convergence). The best 17 structures were selected based on the lowest total restraint energy and analyzed using Procheck and Procheck_NMR (24).
Coordinates-The Protein Data Bank accession number for the coordinates is 1Z4H.

RESULTS AND DISCUSSION
The TorI Response Regulator Inhibitor Is a Winged Helix Protein-To solve the three-dimensional structure of TorI we produced the recombinant protein in the presence of either [ 13 C]glucose or [ 15 N]NH 4 Cl or both labeled substrates. After recording a series of HSQC spectra at different temperatures, we finally picked up 278 K as the best stability/   intensity ratio. The structure of TorI was then solved at 278 K (40 mM NaPO 4 (pH 5.9)) using conventional homonuclear two-and multidimensional heteronuclear NMR spectroscopy on a 500 MHz Bruker DRX spectrometer and at high resolution field (800 MHz Varian Inova spectrometer), making use of uniformly 15 N-and 13 C-labeled protein.
The NMR assignment of 1 H, 15 (18 -21). The global fold was established using unambiguous assigned long range NOEs. Several long range NOEs between the helices and the ␤-strands were obtained from a three-dimensional NOESY ( 1 H-15 N)-HSQC spectrum. The unassigned NOEs with multiple possible assignments were used in ARIA (25) as ambiguous restraints. The final ensemble of structures was calculated using non-bonded interaction for simulated annealing and refinement of the final structures in an explicit water box. A total of 1341 restraints have been used to calculate the structures. 1173 distance restraints were identified from the two-dimensional and three-dimensional NOESY; 68 dihedral angle constraints were obtained from TALOS (23) on the basis of backbone chemical shift values; 100 dihedral angle constraints were estimated from the chemical shift index (CSI), 3 J HN-H␣ coupling constant, and from the use of the Karplus equation (26). TABLE ONE summarizes the experimental restraints and the structural statistics of the 17 best structures. A superimposition of the final ensemble of 17 simulated annealing structures is shown in Fig. 1A; a view of the representative (closest to average) structure is shown in Fig. 1B, and the electrostatic surface potential of the TorI protein overlaid with the ribbon diagram is presented in Fig. 1C.
The conformers within the ensemble do not exhibit any NOE, dihedral angle, or scalar coupling constant violations greater than 0.4 Å, 5°, or 2 Hz, respectively. Residues Gln 6 -Arg 63 are well structured in solution and the coordinates of their backbone and heavy atoms can be superimposed to the average structure with a root mean square deviation (r.m.s.d.) of 0.69 (Ϯ0.25) Å and 1.18 (Ϯ0.24) Å, respectively.
The winged helix structural motif is a derivative of the usual helixturn-helix DNA binding motif (27)(28)(29)(30). It is known for some winged helix proteins that DNA binding occurs through two kinds of interactions: one ␣-helix interacts with the major groove of the DNA target, while the wing penetrates the minor groove of the DNA target. This wing can be seen as an anchor that increases the specificity of the interaction. In the TorI structure the winged helix is composed of helix ␣2 and of the wing in between ␤2 and ␤3. This motif thus constitutes a potential DNA binding site on the TorI protein.
Structural Similarity of TorI with Other Proteins-Interestingly a search for homologous structure using the programs DALI (www. ebi.ac.uk/dali/) and SSM (www.ebi.ac.uk/msd-srv/ssm) allowed us to identify prophage excisionase-type molecules of the Xis family (r.m.s.d. of 2.63 Å with a Qscore of 0.26 with the structure of full-length excisionase from bacteriophage HK022, Protein Data Bank code 1pm6 and r.m.s.d. of 2.7 Å with a Z-score of 2.5 with the structure of the Xis protein, Protein Data Bank code 1lx8) (27,28) (Fig. 2). These proteins belong to the RDF family of proteins that comprises a diverse group of proteins involved in controlling the directionality of integrase-mediated site-specific recombination (2). In lambdoid phages the Xis proteins are the essential partners of the tyrosine family of site-specific recombinases (Int) for the excision of the prophage DNA during the lytic cycle. Indeed, the integrase plays a key role in both integration and excision reactions together with the host factor IHF, but the directionality of recombination is driven by the excisionase that interacts with the integrase and DNA during prophage excision (3,31). The three-dimensional conformation of Xis has been described to be similar to the DNA binding domain of the Mu bacteriophage repressor (MuR-DBD), to the DNA binding domain of the Mu transposase (MuA-DBD) protein and recently to the excisionase protein from the conjugative transposon Tn916 (32, 33) (see below for discussion).
Identification of TorI DNA Target and in Vivo Excisionase Activity-We previously found that the torI gene was actually part of a cryptic prophage genome, the prophage KplE1 (or CPS53) that extends

Excision activity of TorI and mutants
Excision test have been performed in strain LCB970 containing pJF119EH derivatives as described under "Materials and Methods." pJFi encodes the wild-type TorI protein, and the mutations of the variants are indicated. All activities were measured in the presence of 1 mM IPTG as the inductor for torI gene expression. Excision percentages of the mutants are calculated using the TorI wild-type activity as the reference (100%).  (7,13). The torI gene is located at the 3Ј end of this prophage, whereas the intS gene, encoding a putative tyro-sine integrase, is located at the other extremity (Fig. 3). The KplE1 cryptic prophage was identified by DNA sequence homology to S. flexneri bacteriophages Sf6 and V and coliphage HK620, all of them being -type temper- ate phages. We thus looked for the characteristic sequences involved in prophage excision at the 3Ј end of the KplE1 prophage (attR). As expected, we could predict all of the sequences necessary for the intasome structure formation (Fig. 3). This region indeed comprises the integrase binding sites, composed of a core sequence and two arm-type sequences (O, P1 and P2, respectively), two IHF binding sites (H1, H2), and two perfect repeats of a 10-bp motif that are good candidates for being TorI binding sites (I1, I2).

Ap R /Cm
To demonstrate the in vivo activity of TorI as a prophage excisionase we constructed a test strain in which the chloramphenicol acetyltransferase encoding gene was inserted in a non-coding region of the prophage KplE1. We then checked for the effect of TorI overexpression on KplE1 excision by estimating the number of bacteria (colony-forming units) that lost the ability to grow on chloramphenicol. Upon IPTG induction of the torI gene, most of the colonies proved to be chloramphenicol sensitive compared with the control condition in the presence of the vector alone (TABLE  TWO). PCR amplifications of the 3Ј end of the KplE1 DNA were then performed on a sample of randomly chosen colonies plated on rich medium to confirm the disappearance of the KplE1 prophage DNA from the bacterial chromosome upon TorI overexpression. As expected, no amplification product could be obtained in cells overproducing TorI (Fig. 4,  lanes 1-15), whereas a PCR product at the expected size (490 bp) was obtained in control cells containing the vector alone (Fig. 4, lane C). These results show (i) that the KplE1-defective prophage can be excised and (ii) that TorI overexpression promotes KplE1 prophage excision and allows the in vivo validation of the structural homology of TorI with -type excisionases. Based on similarity with the Xis protein and identification of the winged helix DNA binding motif identified in the three-dimensional structure, we anticipate that TorI should also interact with a specific DNA target in the attR region of the prophage.
TorI Residues Involved in DNA Binding Activity-To investigate the molecular basis of TorI interaction with DNA and to locate the residues important for DNA binding, we performed a titration assay of TorI by adding increasing amounts of the putative TorI DNA target, using NMR chemical shift perturbations. This method consists of recording the twodimensional 1 H-15 N HSQC spectrum of TorI upon successive additions of the 10-bp putative target DNA (5Ј-GGGTAAAATA) (Fig. 3). This method detects residues that interact directly with DNA or that are indirectly affected by its binding (Fig. 5). Upon DNA binding, which leads to modifications of the chemical and/or magnetic environments, the resonances of the unbound TorI underwent chemical shift changes (Fig. 5B). The free and DNA-bound proteins are in fast exchange on the NMR time scale. The largest effects in the 1 H and 15 N resonances occurred in the N terminus, the first turn, the helix 2, and the wing (reverse turn connecting strands ␤2 and ␤3). More precisely, the largest chemical shifts changes were detected for Gly 23 /Lys 24 (end of the first turn) and Arg 45 (in the wing) whereas none of the C-terminal residues of TorI were affected by DNA binding (Fig. 5, B and C). Remarkably, these residues are located in the HTH motif and in the wing, which can thus be defined as the TorI DNA binding motif.
To confirm this model of interaction we decided to change by site-directed mutagenesis two highly conserved residues (Tyr 28 and Arg 45 ) that underwent chemical shifts upon DNA binding (Figs. 5 and 6). We first checked that all TorI mutants were produced and stable in strain LCB970 by Western blot analysis using a whole TorI antiserum (data not shown). As indicated in TABLE TWO, both substitutions of Tyr 28 with either Ser or Phe completely abolished the excisionase activity of TorI. On the other hand, the TorI proteins where Arg 45 was mutated to either Gln or Lys, were still able to promote some excision of the prophage, although the activity was strongly impaired by these mutations (about 7 to 10% of the wild-type activity). These results thus indicate that Tyr 28 in helix ␣2 and Arg 45 in the wing are involved in TorI excisionase activity. However, Tyr 28 seems absolutely required as suggested by the phenotype observed with both conservative mutations, whereas Arg 45 when changed to Gln or Lys retains some excisionase activity, suggesting that DNA binding through the wing is less stringent than through helix ␣2.
TorI, the Structural Missing Link between Xis and MuR Proteins?-According to its three-dimensional structure and to the mutagenesis study, TorI is a prokaryotic winged helix protein, which contains a HTH motif (helix ␣1-turn1-helix ␣2) and a loop (wing) contacting the DNA. Moreover, TorI is closely related to the structure of Xis and to the excisionase of the conjugative transposon Tn916 (32,34) but also to the DNA binding domains of the Mu bacteriophage repressor (MuR-DBD) and transposase (MuA-DBD) proteins (30,33) (Fig. 2). The MuA and Tn916 transposases are similar to the recombination directionality factors in functioning to modulate the efficiency of the transposition of Mu bacteriophage and Tn916, respectively, whereas the MuR protein establishes lysogeny by shutting down phage transposition functions and by competing with the transposase in the operator region (35).
The structures of the excisionase-DNA complex of Xis and of the Mu-repressor-DNA complex have been determined (28,36). These two complexes both reveal interactions between the helix ␣2 and the wing of the proteins and the bound target DNA. In the Xis complex structure the helix ␣2 (equivalent to the TorI helix 2) is inserted into the major groove, while the wing contacts the adjacent minor groove. This model is consistent with several mutagenesis studies where mutants with amino acid substitutions within the helix ␣2 show a decreased excisionase activity in vivo and are defective in DNA binding (28,37). Moreover, NMR chemical shift mapping data on MuR with DNA (30) showed that the largest deviations in the 1 H and 15 N shifts occur in the helix-turnhelix unit (H1-T-H2) and the flexible loop between strands ␤2 and ␤3 (wing). Interestingly, the most drastic 1 H and 15 N chemical shift changes occur in the amide proton of Gly 30 (in turn T) and the amide nitrogen of Ala 57 (in the wing connecting strands ␤2 and ␤3), respectively. These two residues correspond to the Gly 23 (turn T1) and Arg 45 (wing) of TorI that were strongly affected by the binding of DNA (Fig. 5B). Therefore the similar fold between these proteins (TorI, Xis, and MuR/MuA) seems to be intricately correlated to their DNA binding properties.
Despite these convergent structural similarities, the TorI protein presents new structural features compared with the excisionase Xis. The major difference is located at the C-terminal end of TorI where 12 residues form a well structured alpha helix (residues Tyr 51 -Ser 63 ) (Fig. 2), whereas the Xis protein contains a ␤-sheet at its C terminus that was shown to interact with the integrase (3,38). In contrast to Xis, the Mu repressor protein harbors a C-terminal helix similar to TorI, which is also comprised of 12 residues (residues Thr 67 -Gly 79 ) (30). These data suggest that TorI is more related to the Mu repressor than to the excisionase Xis. However, the Mu repressor possesses a large wing (residues Glu 50 -Lys 56 ) compared with TorI and Xis (residues Ile 42 -Arg 45 and Asp 37 -Glu 40 , respectively). The size of the wing has been described to be intrinsically correlated to its dynamic properties. Indeed, the authors have shown that the large wing of MuR undergoes a major structural rearrangement upon DNA binding (36). In fact, in the absence of DNA, the MuR wing is mostly disordered, but upon DNA binding, insertion into the minor groove severely dampens rapid motions within the wing. In the case of Xis the short wing is ordered in the absence of DNA (34). By this way the Xis protein avoids the necessary entropic cost to block the wing to save the energy needed for DNA distortion (28). Through the structural similarities of Xis and MuR, the authors suggested that the unrelated enteric lambda and Mu bacteriophages RDF proteins could be derived from a common ancestor (28) that has yet to be defined. Since TorI possess a short wing similar to that of Xis and a well defined C-terminal helix as found in MuR, we propose that TorI represents the structural missing link between the Xis and MuR proteins.
A Family of Atypical Excisionases-The TorI protein has been identified as a response regulator inhibitor that binds to the TorR response regulator (7). Moreover, this study shows that TorI has an excisionase activity even if no relevant primary sequence homology with Xis was observed. Interestingly, the closest homologues of TorI are of phage origin and belong to different pathogenic species such as V. cholerae, Y. pseudotuberculosis, or E. coli O157:H7 species. The sequence alignment of TorI with its homologues (Fig. 6) shows the conservation of several patches of residues. Among them, Tyr 28 (in helix 2) of TorI is strictly conserved throughout this family. Thr 20 and Gly 21 are also highly conserved, and these two residues constitute the turn of the helix-turn-helix motif of TorI. Pro 37 is also strictly conserved and might be involved in the optimization of the extended structure of the polypeptide chain just before the ␤2-strand. The other conserved residues in this family, His 43 and Arg 45 , are part of the wing. As shown above both Tyr 28 and Arg 45 proved to be crucial for the excisionase activity of TorI. According to these sequence similarities we thus predict that all of these proteins share the same secondary structures and fold, as well as the same function, which is consistent with the in vivo excisionase role of at least three of them, TorI, Hef, and Rox in E. coli, Y. pseudotuberculosis, and S. flexneri, respectively (14,15). Since TorI is also involved in gene regulation, we propose that this particular family of RDF proteins may be also involved in regulatory processes. The TorI structure thus provides a structural basis for the understanding of the dual role of this family of excisionases.