Biochimie 83 (2001) 261−267 © 2001 Société française de biochimie et biologie moléculaire / Éditions scientifiques et médicales Elsevier SAS. All rights reserved. S0300908400012141/FLA

The GemA protein of phage Mu and the GyrB gyrase subunit of Escherichia coli: The search for targets and interactions leading to the reversion of Mu-induced mutations Chiraz Abbes, Danièle Joseleau-Petit, Jean-Claude Liébart, Richard D’Ari*, Guennadi Sezonov Institut Jacques-Monod, CNRS, Université Paris 6, Université Paris 7, 75251 Paris cedex 05, France (Received 22 November 2000; accepted 7 December 2000) Abstract — The mutant bacteriophage Mugem2(Ts), known to synchronize the division of infected cells, to relax DNA supercoiling and, as prophage, to give rise to precisely excised revertants, has been thought to overexpress the gemA-mor operon, and genetic evidence suggests that the B subunit of DNA gyrase (GyrB) is the target of action of GemA. In two different double hybrid tests presented here, we find no evidence of GemA-GyrB protein-protein interaction. We do observe a GemA-GemA interaction, however, indicating that GemA can dimerize. In lacZ::Mu lysogens, overexpression of the gemA-mor operon from a plasmid, under control of the L-arabinose inducible paraBAD promoter, does not permit the recovery of Lac+ revertants. These observations suggest that GyrB is not the direct target of GemA action and that the various phenotypes of Mugem2(Ts) are not caused by overexpression of the gemA-mor operon. © 2001 Société française de biochimie et biologie moléculaire / Éditions scientifiques et médicales Elsevier SAS Mugem2(Ts) / Mu semi-essential region / mor / gemB / protein-protein interaction

1. Introduction The mutant bacteriophage Mugem2(Ts) has several interesting properties. After infection at 42 °C (a temperature non-permissive for plaque formation), it synchronizes the division of the surviving bacteria and causes waves of DNA relaxation followed by reintroduction of negative supercoiling at the time of the synchronous division [1–3]. As prophage, bacterial mutations due to the insertion of Mugem2(Ts) give rise to revertants at frequencies around 10–6 [4, 5], whereas mutations due to the insertion of a wild type Mu in a gene are never observed to revert (< 10–10). Several observations have suggested that the target of Mugem2(Ts) infection is the B subunit of DNA gyrase, product of the gyrB gene. First of all, hic mutants, resistant to Mugem2(Ts) killing, carry a mutation tightly linked to gyrB that confers resistance to the GyrB inhibitors coumermycin and novobiocin, causes relaxation of plasmid DNA and, in one case at least, results in four-fold lower gyrase activity [2, 3]. Second, some gyrB mutants can no longer be synchronized by Mugem2(Ts) infection [1, 3]. Third, Mugem2(Ts) lysogens stored in stabs tend to acquire mutations tightly linked to gyrB that increase DNA supercoiling [6]. Finally, in Mugem2(Ts) lysogens, the bacterial DNA is highly relaxed (cited in [7]), and * Correspondence and reprints. E-mail address: [email protected] (R. D’Ari).

expression of the gemA gene from a plasmid induces relaxation of a reporter plasmid (cited in [8]). The gem2(Ts) mutation of phage Mu is located in the promoter region of the gemA gene, near the end of the semi-essential region [9]. This is the first gene of a two-gene operon, followed by the gene mor [10], also called gemB [11]. The mor gene product is a transcriptional activator of gene C, which in turn activates the late Mu promoters [10, 11]. There is evidence that the gemAmor operon is expressed constitutively in the presence of Mu repressor [7, 12]. These observations led to an attractive model for the mechanism of synchronisation of cell division by Mugem2(Ts). It was speculated that the GemA protein interferes directly with GyrB activity, thereby leading to DNA relaxation [2]. The gem2(Ts) mutation, located in the promoter, would cause overproduction of GemA, with stronger DNA relaxation. This in turn was suggested to lead to a reprogramming of the cell cycle, hinting that normal cell cycle regulation may involve changes in superhelical density [3]. In the present work we present evidence that GemA and GyrB do not in fact interact directly, and we show that GemA interacts with itself, indicating that it forms an oligomer. We further show that overexpression of the gemA-mor operon is not per se sufficient to permit the reversion of mutations due to the insertion of a wild type Mu prophage.

262

Abbes et al.

Table I. Bacterial strains. Strain

Genotype

Source or reference

MT5 JC223 AT408 AT397 GC2793 XL1-Blue

ilv his rpsL lN7N53 CI+cro+::gal+ MT5 CI/pSC101lacIq relA1 araD lacZ::Mu MG1655 ∆ara lacZ::Mu ∆cya-854 recA1 endA1 gyrA96(Nalr) thi relA1 − + hsdR17 共 rK mK 兲 supE44 lac/F’::Tn10 + + q proA B lacI ∆ (lacZ)M15

[27] This work L. Paolozzi This work [28] [29]

2. Materials and methods 2.1. Bacterial strains, phage, yeast strain The bacterial strains used, all derivatives of E. coli K12, are described in table I. A Gal+ derivative of strain MT5 was isolated and transformed with plasmid pSC101lacIq to form strain JC223. Strain AT408 (RS54 with a wild type Mu phage integrated in the lacZ gene) was given to us by L. Paolozzi. The bacteriophages used were Mu wild type, P1vir for transduction and λ and λimm434 to test λ immunity. The Saccharomyces cerevisiae strain used in double hybrid experiments was Y190, genotype MATa ura3-52 his3-200 lys2-801 ade2-101

trp1-901 leu2-3 gal4∆ gal80∆ cyhr2 LYS2::GAL1UASHIS3TATA-HIS3 URA3::GAL1UAS-GAL1TATA-lacZ [13]. 2.2. Plasmids Plasmid pAS2 codes for the DNA-binding domain of the yeast GAL4 protein and the TRP1 gene and pACTII codes for the activating domain of the yeast GAL4 protein and the LEU2 gene [13]; PRP9 and PRP21 code for interacting yeast splicing factors cloned in plasmids pAS2 and pACTII [14]; pCR2.1-TOPO is described in the Invitrogen Instruction Manual ‘TOPO TA Cloning’; pT25 codes for amino acids 1–224 of the Cya protein of Bordetella pertussis and pT18 codes for amino acids 225 to 399 [15]; pAG111 carries the gyrB gene under control of the ptac promoter [16]; pC132 codes for the N-terminal (DNA-binding) domain of the λCI repressor under plac control [17]; pBAD24 carries the araBAD promoter and the araC gene, coding for the L-arabinose-dependent activator [18]; pSC101lacIq carries the lacI gene with an up promoter mutation [19]. Construction of the gemAcarrying plasmids is described in table II and figure 1; all inserts were verified by direct DNA sequencing. Construction of the gyrB-carrying plasmids is described in table III. 2.3. Media and growth conditions For E. coli, the media used were LB, MacConkey (with 1% lactose, maltose or galactose), M63 and M9 [20]; the

Table II. Construction of gemA-carrying plasmids*. Plasmid pT18gemA pT25gemA pAS2gemA pACTIIgemA pC132gemA pBAD24gemA pBAD24gemAlong pBAD24gemAgemB pBAD24gemAlonggemB

Leading strand oligonucleotide

Lagging strand oligonucleotide

DNA source

tcGGTACCgATGAGTCGCACATCC Asp718 ⇓8233 atCTGCAGgcATGAGTCGCACATC PstI ⇓8233 atgCCATGGATGAGTCGCACATCCCTG NcoI ⇓8233 cgcGGATCCATGAGTCGCACATCCCTG BamHI ⇓8233 cgcGTCGACcATGAGTCGCACATCC SalI ⇓8233 ggtGGTACCATGAGTCGCACATCCCTG KpnI ⇓8233 cggGAATTCGTGTTGCAGCACCGCCCCCTTTC EcoRI ⇓8197 ggtGGTACCATGAGTCGCACATCCCTG KpnI ⇓8233 cggGGTACCGTGTTGCAGCACCGCCCCCTTTC KpnI ⇓8197

gcATCGATccTTGTGCCATCCTCC ClaI ⇓8781 atGGTACCgcTTGTGCCATCCTCC Asp718 ⇓8781 cgcGGATCCTCATTGTGCCATCCTCCG BamHI ⇓8784 gCTCGAGTCATTGTGCCATCCTCCG XhoI ⇓8784 cgcGGATCCTCATTGTGCCATCCTCCG BamHI ⇓8784 aaCTGCAGTCATTGTGCCATCCTCCG PstI ⇓8784 aaCTGCAGTCATTGTGCCATCCTCCG PstI ⇓8784 aaCTGCAGTTACAGAAGTGAGGGCTG PstI ⇓9170 aaCTGCAGTTACAGAAGTGAGGGCTG PstI ⇓9170

pBAD24gemA pBAD24gemA pBAD24gemA pBAD24gemA pBAD24gemA Muc62(Ts) Muc62(Ts) Muc62(Ts) Muc62(Ts)

* DNA from the indicated source was PCR amplified using the oligonucleotides shown. These include a restriction enzyme site (underlined capital letters with the enzyme name underneath) followed by a sequence homologous to gemA or mor (capital letters). The nucleotide coordinates in the Mu DNA sequence are shown under the first homologous base (in bold). The amplified fragment was digested with the two enzymes indicated and cloned in the vector digested with the same enzymes. The plasmids designed to code for a fusion protein (first five plasmids) do not include the gemA stop codon in the lagging oligonucleotide.

The GemA protein of phage Mu and the GyrB gyrase subunit of E. coli

263

Figure 1. The Mu gemA-mor operon. The two genes are shown together with the coordinates of the two potential start points for the gemA coding sequence, the start point for mor, and the ends of the two stop codons. The coordinates are from the Mu sequence in the EMBL data base (accession number AF 083977); they are 2 bp higher than the coordinates in the book (Phage Mu) [26].

latter was supplemented with 0.2% carbon source and 100 µg/mL casamino acids. For S. cerevisiae, the rich medium was YPGA and minimal medium was W0 supplemented with adenine and uracil, 20 µg/mL each [21]. Antibiotics were used, as needed, at the following concentrations: 30 µg/mL chloramphenicol, 50 or 100 µg/mL ampicillin, 50 µg/mL spectinomycin. X-Gal was used at 40 µg/mL. In the yeast double hybrid tests, strain Y109 has two reporter genes, HIS3 and lacZ. After double transformation on medium lacking histidine, the His+ colonies were transferred to X-Gal-containing medium (400 µg/mL) and tested for coloration. 2.4. Selection of non-dimerizing gemA mutants To isolate non-dimerizing mutants of gemA, we PCR amplified the gemA gene from pBADgemA, using mutagenic conditions as described [22]. The fragment was then digested by SalI and BamHI, cloned in plasmid pC132 and introduced into strain JC223. Alternatively, we

mutagenized plasmid pC132gemA in vitro with nitrous acid, as described [23], then subcloned the gemA moiety (SalI-BamHI fragment) in plasmid pC132 and transformed strain JC223. Transformants with the mutagenized plasmids were selected on MacConkey galactose plates containing ampicillin, spectinomycin and IPTG (5 × 10–3 M). With the unmutated gemA gene, the colonies are white (Gal–). Red (Gal+) colonies were characterized further; they were red in the absence of IPTG, and they did not express λ immunity even in the presence of IPTG.

3. Results 3.1. Test for GemA-GyrB interaction using double hybrid assays To test the hypothesis that GemA and GyrB interact, we used the double hybrid system in S. cerevisiae, in which

Table III. Construction of gyrB-carrying plasmids*. Plasmid pT18gyrB pT25gyrB pAS2gyrB pACTIIgyrB

Leading strand oligonucleotide cggGGTACCgatgtcgaattcttatgac Asp718 aaCTGCAGgcatgtcgaattcttatgac PstI cggtcCATATGtcgaattcttatgac NdeI ccgGGATCCggatgtcgaattcttatgac BamHI

Lagging strand oligonucleotide cccAAGCTTgctcgcatggttagc HindIII cggGGTACCgctcgcatggttagc Asp718 gcgGGATCCgctcgcatggttagc BamHI gcgCTCGAGgctcgcatggttagc XhoI

* The original source of gyrB DNA was pAG111. The gene was recloned in pCR2.1-TOPO, which was then used for PCR amplification. This DNA was PCR amplified using the oligonucleotides shown, which include a restriction enzyme site (shown in underlined capital letters with the enzyme name underneath) followed by a sequence homologous to gyrB. The amplified fragment was digested with the two enzymes indicated and cloned in the appropriate vector (the first part of the plasmid name) digested with the same enzymes.

264 two interacting proteins will reconstitute a functional GAL4 transcriptional activator. The gyrB gene and the gemA gene were each fused in frame to the sequences coding for the GAL4 DNA binding domain (plasmid pAS2) and for the activation domain (plasmid pACTII). We transformed the yeast reporter strain Y109 either with pAS2gemA and pACTIIgyrB or with pAS2gyrB and pACTIIgemA. Although His+ colonies were obtained in both cases, none were LacZ+, indicating that neither combination of GemA and GyrB resulted in a functional GAL4 activator. As positive control, sequences coding for the interacting yeast splicing factors PRP9 and PRP21, cloned on the two vectors [14], induced strong expression of β-galactosidase among His+ double transformants. To carry this analysis further, we wished to test the possibility of a GemA-GyrB interaction in a different organism. We therefore turned to a double hybrid system developed in E. coli, in which an interaction between the two proteins to be tested provides an active adenyl cyclase, which in turn synthesizes cyclic-AMP, effector of the transcriptional activator Crp [15]. Active Crp-cAMP complexes are readily detected, for example, by the Mal+ phenotype they confer. In this system, we cloned the entire gyrB gene fused to the sequence coding for one domain of adenyl cyclase and gemA to the sequence coding for the other cyclase domain. We transformed the E. coli ∆cya strain GC2793 either with pT18gemA and pT25gyrB or with pT18gyrB and pT25gemA, then tested the phenotype on MacConkey maltose plates. (The use of a ∆cya deletion allele completely eliminated the Mal+ background.) Neither combination gave a Mal+ phenotype. As control, a sequence coding for a leucine zipper cloned in each vector gave a clear Mal+ phenotype. We also observed a clear Mal+ phenotype in the presence of pT18gemA and pT25gemA together, indicating that GemA can dimerize. 3.2. Dimerization of GemA The ability of GemA to interact with itself was confirmed in a recently developed in vivo system for detecting dimerization [23]. The protein to be tested is fused to the N-terminal (DNA-binding) moiety of λ repressor, replacing the C-terminal (dimerization) domain; if it can dimerize, active λ repressor will result, detectable by the immunity it confers to λ infection. The fusion protein is expressed from plac and is thus IPTG-inducible. The gemA coding sequence was fused to the sequence coding for the N-terminal part of λ repressor and the resulting plasmid, pC132gemA, was introduced into the λ sensitive strain XL1-Blue. The strain expressed λ immunity in the presence of IPTG but not in its absence. This again indicates that GemA can dimerize. A truncated GemA protein, lacking the first 28 amino acids, was also able to dimerize, showing that the N-terminal region is not required for dimerization. We

Abbes et al. isolated mutants of plasmid gemA no longer able to confer λ immunity, after PCR- or nitrous acid mutagenesis (see 2.Materials and methods). Of 30 mutant plasmids obtained, DNA sequencing revealed that six had single single amino acid changes in GemA (the other mutants had multiple changes). The single alterations were R14W, T32S, T41S, K69R, H123Q, and H174Q. These mutations are not clustered within the gemA gene and clearly include alleles that affect overall GemA structure or stability, since at least one altered amino acid lies in the N-terminal domain, dispensable for dimerization. 3.3. Can overexpression of GemA (or GemA + Mor) permit the reversion of Mu-induced mutations? As mentioned in the 1.Introduction, the gemA-mor operon is expressed constitutively, in the presence of Mu repressor, although we were unable to detect GemA protein by immunoblot in a Mu lysogen (data not shown). A Mugem2(Ts) prophage is thought to overexpress the gemA-mor operon, and mutations due to the insertion of this mutant prophage give rise to revertants, unlike mutants due to the insertion of a wild type Mu prophage. We reasoned that overexpression of the gemA-mor operon in trans should allow reversion of mutants due to the insertion of a wild type Mu prophage. We therefore looked to see whether we could detect revertants of lacZ::Mu lysogens, with a wild type Mu inserted in the lacZ gene, in the presence of a plasmid carrying the gemA gene under control of the araBAD promoter, to allow overexpression of GemA. Two different lysogens were used, strains AT408 and AT397. Each has a wild type Mu inserted within the lacZ gene, in different genetic backgrounds (see table I). The strains were transformed with plasmids expressing gemA under control of paraBAD.The selection for Lac+ bacteria was on minimal lactose plates with L-arabinose at concentrations ranging from 1.5 × 10–5 to 1 × 10–3 M. No Lac+ revertants were obtained at any arabinose concentration (< 2 × 10–9; table IV). The transcriptional initiation site of the gemA operon was identified by primer extension to be at position 8217 from the Mu left end, although the probable promoter lies some 29 bp upstream [7]. Since this region can form a stem-loop structure, it seemed possible that the 5’ end of the gemA mRNA corresponded to a site of RNA cleavage. If this is the case, then, accepting the promoter position, translation could start at the GTG (position 8197; see figure 1) located 12 codons upstream of the proposed ATG start (position 8233), in which case our cloned gemA gene would produce a truncated, possibly inactive GemA protein, 12 amino acids shorter than the product ‘GemAlong’. Analysis of codon usage in GemAlong (programme Frame plot 2.3, cf. http://www.kazusa.or.jp/codon) is consistent with the hypothesis that the 12 codons upstream of the ATG suggested by Fabozzi et al. are part of the gemA

The GemA protein of phage Mu and the GyrB gyrase subunit of E. coli Table IV. Frequency of Lac+ revertants of lacZ::Mu mutants in the presence of GemA*. Plasmids pBAD24 pBAD24 gemA pBAD24 gemAlong pBAD24 gemAgemB pBAD24 gemAlonggemB

AT408

AT397

< 2 × 10–9 < 2 × 10–9 < 3 × 10–9 < 1 × 10–9 < 3 × 10–9

< 5 × 10–9 < 6 × 10–10 < 2 × 10–9 < 2 × 10–9 < 6 × 10–10

* Overnight cultures in LB ampicillin were centrifuged and washed three times, then 2 × 108 to 2 × 109 bacteria were plated on minimal lactose plates supplemented with 100 µg/mL casaminoacids, 100 µg/mL ampicillin and L-arabinose at concentrations ranging from 1.5 × 10–5 M to 1 × 10–3 M (or 0 M). Bacteria were assayed on the same plates containing glucose instead of lactose. For each culture, the same number of cells were plated at each arabinose concentration, so the upper limits shown are valid for all concentrations ≤ 1 × 10–3 M.

gene. We therefore recloned the gemA gene from the GTG site and introduced this plasmid into the two lacZ::Mu lysogens. We then selected for Lac+ revertants as above and again found none at any arabinose concentration (< 3 × 10–9; table IV). Since gemA and mor form an operon, if a Mugem2(Ts) prophage overexpresses gemA, it should also overexpress mor, and the ability to give rise to Lac+ revertants might require overexpression of both proteins. We therefore cloned gemA-mor and gemAlong-mor under control of the araBAD promoter and selected for Lac+ revertants; none were obtained (< 3 × 10–9; table IV). Lac+ revertants were also screened for on MacConkey lactose plates and on LB plates containing X-Gal, using both lysogens and all four plasmids; again none were found (data not shown).

4. Discussion The GemA protein of bacteriophage Mu has been implicated in the interesting properties of the mutant phage Mugem2(Ts). It has been suggested that the mutant overexpresses the gemA-mor operon, both as prophage and after infection, and that excess GemA interacts with GyrB, thereby interfering with gyrase activity and leading in both cases to DNA relaxation, followed, after infection, by reintroduction of supercoiling and a synchronous division of the cells [1], and, in lysogens, by the recovery of revertants of mutations created by the insertion of a Mugem2(Ts) prophage [24]. These ideas were supported by the observation that three revertants of Mugem2(Ts) that could once again form plaques at 42 °C and had recovered the wild type sequence in the gemA promoter region had simultaneously lost the ability to synchronize the cells they infected [9].

265

In the present work we looked for direct evidence of a protein-protein interaction between GemA and GyrB, as suggested by the genetic data. Using two different double hybrid systems, one in yeast and the other in bacteria, we detected no interaction. While this does not formally prove that the two proteins do not interact, it certainly suggests that there is not a strong association constant between them. Furthermore, we found that massive overproduction of GemA, confirmed by immunoblot (data not shown), has no noticeable effect on cell growth or morphology, showing that GemA cannot be a strong gyrase inhibitor. Interestingly, it was reported recently that a Mu phage in which the gemA gene is inactivated by a non-polar deletion exhibits completely normal lytic growth [25], establishing that GemA is not required for replicative transposition of Mu. The function of GemA remains unknown. In the present work, we found that GemA can dimerize, as detected in two different in vivo systems. Lysogens for wild type Mu frequently have a host gene disrupted by the prophage, and these mutations have never been observed to revert (frequency < 10–10), in striking contrast to mutations due to other transposable elements. This apparent paradox has been shown to reflect loss of cells in which precise excision occurs, rather than to lack of precise excision (C. Abbes, G. Sezonov, D. JoseleauPetit, R. D’Ari and J.-C. Liébart, submitted). When the prophage carries the gem2(Ts) mutation, however, revertants are obtained at frequencies around 10–6 [6, 24]. We reasoned that if the primary effect of the gem2(Ts) mutation is indeed to overexpress the gemA-mor operon, it should be possible to obtain revertants of mutants due to insertion of a wild type Mu prophage simply by providing these products in trans. Using lacZ::Mu lysogens transformed with plasmids carrying either gemA alone or gemA and mor cloned under control of the araBAD promoter, no Lac+ revertants were recovered (< 2 × 10–9) over a broad range of GemA-Mor expression. We also recloned these genes in order to include a second potential initiation codon 12 codons upstream of the suggested translational start; again, overexpression of GemAlong or GemAlong plus Mor did not allow us to detect Lac+ revertants. We conclude that mere overexpression of GemA or GemA plus Mor is not sufficient to permit the reversion of mutants due to the integration of a wild type Mu prophage. This in turn implies that a Mugem2(Ts) prophage has some other property that cannot be mimicked by the production of GemA (plus Mor) in trans. How then can one explain the phenotypes of Mugem2(Ts)? One can question the hypothesis that the mutated gem2(Ts) promoter causes overproduction of the gemA-mor operon. The genetic arguments that suggested this were generally based on the assumption that GemA interferes with gyrase activity, whereas the results presented here, lack of interaction in two double hybrid systems, lack of effect on growth and morphology of

266 massive overproduction of GemA, seriously question this assumption. It is possible that the gem2(Ts) promoter causes underexpression of the gemA-mor operon rather than overexpression, with the absence of plaque formation at 42 °C reflecting insufficient expression of Mor, necessary for late gene expression [10]. This is consistent with two earlier observations. First, a Mumor(Am) mutant is not complemented by Mugem2(Ts) [11], strongly suggesting that the latter phage has low mor expression. And second, the Mugem2(Ts) mutant can make plaques at 42 °C in the presence of a plasmid carrying the region of Mu DNA from 8027 to 9947 bp, which includes the mor gene (cited in [4]). According to this hypothesis, the other phenotypes of the Mugem2(Ts) mutant (synchronization of cell division and DNA relaxation) could, for example, result from underexpression of one or more late Mu genes whose expression depends on Mor. A different (but not mutually exclusive) hypothesis is that the Mugem2(Ts) phage carries a second mutation, in addition to the base change found in the gemA promoter region, with both mutations being required for the observed phenotypes. If this is the case, the second mutation is not in the gemA-mor operon [9]. In this study we have critically tested the hypothesis that the phenotypes of the Mugem2(Ts) mutant have as primary cause the overexpression of the Mu gemA-mor operon. We show that expression of gemA-mor from a plasmid does not permit the recovery of Lac+ revertants of a lacZ::Mu lysogen. Other, indirect effects, such as activation of gene C by Mor or of Mu late operons by C, should be reproducible in trans since an entire Mu prophage was present in our strains. Thus the lack of reversion strongly suggests that the gem2(Ts) mutation, which allows reversion of lacZ::Mugem2(Ts) lysogens, does not simply cause overexpression of gemA-mor. This in turn suggests that the other phenotypes of Mugem2(Ts), viz. synchronization of cell division and waves of DNA relaxation, are also not caused by overexpression of the gemA-mor operon.

Acknowledgments We extend our deepest thanks to Danielle Thévenet for her participation in this work. C.A. was recipient of fellowship from the Ministère de l’Education Supérieure Tunisienne. This work was supported in part by grant no. 9981 from the Association pour la Recherche sur le Cancer (France).

References [1] Paolozzi L., Nicosia A., Liébart J.C., Ghelardini P., Synchronous division induced in Escherichia coli K12 by gemts mutants of phage Mu, Mol. Gen. Genet. 218 (1989) 13–17.

Abbes et al. [2] Ghelardini P., Liébart J.C., Marchelli C., Pedrini A.M., Paolozzi L., Escherichia coli K-12 gyrB gene product is involved in the lethal effect of the ligts2 mutant of bacteriophage Mu, J. Bacteriol. 157 (1984) 665–668. [3] Ghelardini P., Lauri P., Ruberti I., Orlando V., Paolozzi L., Synchronous division induced in Escherichia coli K12 by phage Mu: analysis of DNA topology and gene expression during the cell cycle, Res. Microbiol. 142 (1991) 259–267. [4] Ghelardini P., Liébart J.C., Di Zenzo G., Micheli G., D’Ari R., Paolozzi L., A novel illegitimate recombination event: precise excision and reintegration with the Mu gem mutant prophage, Mol. Microbiol. 13 (1994) 709–718. [5] Ghelardini P., Liébart J.C., Fabozzi G., Tomassini B., D’Ari R., Paolozzi L., Reversal of Mu gem2ts-induced mutations, FEMS Microbiol. Rev. 17 (1995) 171–176. [6] Di Lallo G., Fabozzi G., Ghelardini P., Paolozzi L., A spontaneous Escherichia coli K12 mutant which inhibits the excisionreintegration process of Mu gem2ts, FEMS Microbiol. Lett. 154 (1997) 59–64. [7] Fabozzi G., Paolozzi L., Ghelardini P., Regulation of the bacteriophage Mu gem operon, Virology 241 (1998) 73–79. [8] Paolozzi L., Ghelardini P., A case of lysogenic conversion: modification of cell phenotype by constitutive expression of the Mu gem operon, Res. Microbiol. 143 (1992) 237–243. [9] La Valle R., Fabozzi G., Ghelardini P., Paolozzi L., Bacteriophage Mu gem2ts, which induces synchronous cell division in the infected host, is mutated in the gem operon promoter, Res. Microbiol. 147 (1996) 615–618. [10] Mathee K., Howe M.M., Identification of a positive regulator of the Mu middle operon, J. Bacteriol. 172 (1990) 6641–6650. [11] Giusti M., Di Lallo G., Ghelardini P., Paolozzi L., The bacteriophage Mu gem gene: a positive regulator of the C operon required for normal levels of late transcription, Virology 179 (1990) 694–700. [12] Stoddard S.F., Howe M.M., Localization and regulation of bacteriophage Mu promoters, J. Bacteriol. 171 (1989) 3440–3448. [13] Durfee T., Becherer K., Chen P.L., Yeh S.H., Yang Y., Kilburn A.E., Lee W.H., Elledge S.J., The retinoblastoma protein associates with the protein phosphatase type 1 catalytic subunit, Genes Dev. 7 (1993) 555–569. [14] Legrain P., Chapon C., Galisson F., Interactions between PRP9 and SPP91 splicing factors identify a protein complex required in prespliceosome assembly, Genes Dev 7 (1993) 1390–1399. [15] Karimova G., Pidoux J., Ullmann A., Ladant D., A bacterial two-hybrid system based on a reconstituted signal transduction pathway, Proc. Natl. Acad. Sci. USA 95 (1998) 5752–5756. [16] Hallett P., Grimshaw A.J., Wigley D.B., Maxwell A., Cloning of the DNA gyrase genes under tac promoter control: overproduction of the gyrase A and B proteins, Gene 93 (1990) 139–142. [17] Longo F., Marchetti M.A., Castagnoli L., Battaglia P.A., Gigliani F., A novel approach to protein-protein interaction: complex formation between the p53 tumor suppressor and the HIV Tat proteins, Biochem. Biophys. Res. Commun. 206 (1995) 326–334. [18] Guzman L.M., Belin D., Carson M.J., Beckwith J., Tight regulation, modulation and high-level expression by vectors containing the arabinose pBAD promoter, J. Bacteriol. 177 (1995) 4121–4130. [19] Bernardi A., Bernardi F., Complete sequence of pSC101, Nucleic Acids Res. 12 (1984) 9415–9426. [20] Miller J.H., A Short Course in Bacterial Genetics, Cold Spring Harbor Laboratory Press, Plainview, NY, 1992. [21] Dujardin G., Pajot P., Groudinsky O., Slonimski P.P., Long range control circuits within mitochondria and between nucleus and mitochondria, Mol. Gen. Genet. 179 (1980) 469–482. [22] Diaz J.J., Rhoads D.D., Roufa D.J., PCR-mediated chemical mutagenesis of cloned duplex DNAs, Biotechniques 11 (1991) 204–211. [23] Di Lallo G., Anderluzzi D., Ghelardini P., Paolozzi L., FtsZ dimerization in vivo, Mol. Microbiol. 32 (1999) 265–274.

The GemA protein of phage Mu and the GyrB gyrase subunit of E. coli [24] Ghelardini P., Liébart J.C., Di Zenzo G., Micheli G., D’Ari R., Paolozzi L., A novel illegitimate recombination event: precise excision and reintegration with the Mu gem mutant prophage, Mol. Microbiol. 13 (1994) 709–718. [25] Pato M.L., Banerjee M., Genetic analysis of the strong gyrase site (SGS) of bacteriophage Mu: localization of determinants required for promoting Mu replication, Mol. Microbiol. 37 (2000) 800–810. [26] Priess H., Bräuer B., Schmidt C., Kamp D., Sequence of the left end of Mu, in: Symonds N., Toussaint A., van de Putte P., Howe M.M. (Eds.), Phage Mu, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1987, pp. 277–296.

267

[27] Toman Z., Dambly-Chaudière C., Tenenbaum L., Radman M., A system for detection of genetic and epigenetic alterations in Escherichia coli induced by DNA-damaging agents, J. Mol. Biol. 186 (1985) 97–105. [28] D’Ari R., Jaffé A., Bouloc P., Robin A., Cyclic AMP and cell division in Escherichia coli, J. Bacteriol. 170 (1988) 65–70. [29] Bullock W.O., Fernandez J.M., Short J.M., XL1-Blue: A high efficiency plasmid transforming recA Escherichia coli strain with β-galactosidase selection, BioTechniques 5 (1987) 376–378.