Selected Amplification of the Cell Division Genes ftsQ-ftsA-ftsZ in Escherichia coli

Corresponding author: Daniel Vinella, Institut Jacques Monod (CNRS, Université Paris 7, Université Paris 6), 2 place Jussieu, 75251 Paris Cedex 05, France., vinella{at}ijm.jussieu.fr (E-mail)

Communicating editor: P. L. FOSTER

	ABSTRACT

TOP ABSTRACT MATERIAL AND METHODS RESULTS DISCUSSION LITERATURE CITED

Rapidly growing Escherichia coli is unable to divide in the presence of the antibiotic mecillinam, whose direct target is penicillin-binding protein 2 (PBP2), responsible for the elongation of the cylindrical portion of the cell wall. Division can be restored in the absence of PBP2 activity by increasing the concentration of the cell division proteins FtsQ, FtsA, and FtsZ. We tried to identify regulators of the ftsQ-ftsA-ftsZ operon among mecillinam-resistant mutants, which include strains overexpressing these genes. By insertional mutagenesis with mini-Tn10 elements, we selected for insertions that conferred mecillinam resistance. Among 15 such mutants, 7 suppressed the thermosensitivity of the ftsZ84(Ts) mutant, strongly suggesting that they had increased FtsZ activity. In all 7 cases, however, the mutants resulted from a duplication of the ftsQAZ region. These duplications seemed to result from multiple events, suggesting that no simple insertional inactivation can result in a mutant with sufficiently amplified ftsQAZ expression to confer mecillinam resistance. The structure of the duplications suggests a general method for constructing directed duplications of precise sequences.

GROWTH of rod-shaped bacteria such as Escherichia coli takes place in two different modes, lateral extension of the cylindrical portion of the rod and formation, at midcell, of a septum, or cross wall, which becomes the new pole of each daughter cell. The elongation phase of growth distinguishes bacilli from cocci, which grow and divide by pure septation. In E. coli, elongation specifically requires penicillin-binding protein 2 (PBP2; SPRATT 1975 ). During rapid growth, however, a block of the elongation process brought about by inactivation of PBP2 also results in a specific arrest of cell division (VINELLA et al. 1993 ) and consequent cell death (MATSUHASHI et al. 1974 ; JAMES et al. 1975 ). Division in the absence of PBP2 activity can be restored if the cell division proteins FtsQ, FtsA, and FtsZ are all overexpressed (VINELLA et al. 1993 ; JOSELEAU-PETIT et al. 1994 ). These proteins thus seem to participate in a sort of cell division potential that has to be increased to allow cell division in spherical E. coli cells growing in rich medium.

The ftsQ, ftsA, and ftsZ genes are adjacent to each other on the chromosome (ROBINSON et al. 1984 , ROBINSON et al. 1986 ; YI et al. 1985 ), within the 16-gene cluster of cell wall and cell division genes located at 2 min. These genes, governed by multiple promoters, are known to respond to several transcriptional regulators—^S, SdiA, and RcsB (for review, see JOSELEAU-PETIT et al. 1999 ). In the hope of identifying additional transcriptional regulators of the ftsQ-ftsA-ftsZ operon, we selected mutants in which the expression of these three genes was increased enough to permit cell division in the absence of PBP2 activity.

Inactivation of PBP2 can be brought about either genetically, by mutations in the structural gene pbpA, or by use of the highly specific ß-lactam mecillinam (LUND and TYBRING 1972 ; SPRATT and PARDEE 1975 ; SPRATT 1977 ). E. coli can be made resistant to mecillinam not only by overproduction of the division proteins FtsQ, FtsA, and FtsZ (VINELLA et al. 1993 ; JOSELEAU-PETIT et al. 1994 ), as mentioned above, but also by a moderate increase in the concentration of ppGpp (VINELLA et al. 1992 ; JOSELEAU-PETIT et al. 1994 ). This nucleotide is the effector of the stringent response (see CASHEL et al. 1996 for review) observed when relA⁺ cells are starved for amino acids (CASHEL 1969 ; CASHEL and GALLANT 1969 ). Under these conditions, the RelA protein, bound to the ribosomes, responds whenever an uncharged cognate tRNA is presented by synthesizing a compound derived from GTP, called pppGpp or "magic spot," which is rapidly hydrolyzed to ppGpp, the probable effector of the stringent response. This nucleotide binds to RNA polymerase (CHATTERJI et al. 1998 ) and alters its specificity; during amino acid starvation, the high level of ppGpp produced rapidly blocks the transcription of ribosomal operons, thus arresting ribosome formation. E. coli has two ppGpp synthetases, as shown by the fact that relA mutants still have almost normal pools of the nucleotide (LAZZARINI et al. 1971 ; RYALS et al. 1982 ; METZGER et al. 1989 ). The identity of the second ppGpp synthetase was unknown until the observation that relA spoT strains are devoid of detectable ppGpp (HERNANDEZ and BREMER 1991 ; XIAO et al. 1991 ), strongly suggesting that the SpoT protein is the second synthetase, although no in vitro synthetase activity has been detected so far.

We selected mutants that had become resistant to mecillinam through insertion of a mini-Tn10. Initially, we expected two classes of mutants, those with increased levels of FtsQ, FtsA, and FtsZ and those with increased ppGpp levels. Of 12 independent mutants isolated and characterized, 5 remained mecillinam-resistant in the complete absence of ppGpp (i.e., in a relA spoT genetic background). They were thus good candidates for affecting the level of expression of the ftsQ-ftsA-ftsZ operon. We present here a detailed analysis of these 5 mutants and 2 similar mutants previously selected. We show that their mecillinam resistance results from gene duplication via a novel mechanism that in principle could be used for specific, directed amplification of virtually any region of the chromosome.

	MATERIAL AND METHODS

TOP ABSTRACT MATERIAL AND METHODS RESULTS DISCUSSION LITERATURE CITED

Bacterial strains and phages:
All strains used in this work are E. coli K12 derivatives; the genotypes of the principal strains are given in Table 1. The strains carrying various Tn10 transposons, kindly provided by D. Touati, are from the Singer collection (SINGER et al. 1989 ). Strain CF6301 was constructed as follows: the P1_rrnB::lacZ fusion was first introduced from strain VH270 (HERNANDEZ and BREMER 1990 ) into a lacZ argA::Tn10 derivative of MG1655, selecting for kanamycin resistance (which is associated with the fusion), then relA251::kan was introduced by cotransduction with Arg⁺, donor strain CF1651, selecting on minimal glucose plates and screening for transductants sensitive to serine, methionine, and glycine, a sensitivity associated with relA mutants (UZAN and DANCHIN 1976 ). Strain DV250 [relA251::kan lacIpoZ(Mlu) P1_rrnB::lacZ spoT::kan] was constructed by transducing a pyrE zib-563::Tn10 derivative of strain CF6301 with a P1 stock grown on a relA1 spoT206::kan pyrE⁺ strain and selecting Pyr⁺ transductants on minimal glucose medium supplemented with Casamino acids; strain DV250 is a Pyr⁺ Tc^s clone that cannot grow on minimal glucose plates lacking amino acids, indicating the absence of ppGpp (XIAO et al. 1991 ). The ftsZ84(Ts) mutation was brought into a leu::Tn10 derivative of CF1742, selecting Leu⁺ transductants on minimal glucose plates; strain DV262 is a Leu⁺ clone unable to grow at 42° on Luria broth (LB) plates lacking NaCl. Phage NK1324 (KLECKNER et al. 1991 ) was used for random insertional mutagenesis; it carries the mini-Tn10 (Cm^r) transposable element and the Ptac-ats1 ats2 gene, which provides transposase in trans. The selection of insertional mutants was carried out as described previously (VINELLA et al. 1996 ), except that kanamycin was replaced by chloramphenicol in this work.

Media and growth conditions:
The rich and minimal media used in this work were, respectively, LB broth (containing 10 g/liter NaCl unless otherwise indicated) and M63 (MILLER 1992 ). Glucose (0.4%) and amino acids (100 µg/ml) were added when required. Solid media contained 1.5% agar. Antibiotics were used at the following concentrations: 20 µg/ml chloramphenicol (Cm), 50 µg/ml kanamycin (Km), 1 or 10 µg/ml mecillinam (Mec), 20 µg/ml tetracycline (Tc), 50 µg/ml spectinomycin (Spc), and 100 µg/ml ampicillin (Amp).

DNA techniques and plasmids:
Plasmids were extracted and transformations were carried out as described by SAMBROOK et al. 1989 . Sequence determinations and PCR amplifications were carried out using a sequenase kit (United States Biochemical, Cleveland) and Taq polymerase (Perkin-Elmer, Norwalk, CT), respectively. The cloning vectors used in this work were pKS+ (high copy, Amp^r; Stratagene, La Jolla, CA) and pCL1920 (low copy, Spc^r; LERNER and INOUYE 1990 ). To sequence the chromosomal DNA at the ends of the insertional element, we used the following synthetic primers (5' to 3'): CmD, TTATTCTGCCTCCCAGAGCC and CmF, AACGGCAAAAGCACCGCCGG for the chloramphenicol-resistant mutants; and primers ISL, CTACCTTAACTTAATGATTTTGATA and ISR, CCACCTTAACTTAATGATTTTTACCfor the two kanamycin-resistant mutants.

Southern blots:
Chromosomal DNA extracted from the parental strain MG1655 and its mutant derivatives was completely digested by EcoRI and, after agarose (0.9%) electrophoresis of 5 µg, transferred onto hybond-N⁺ nylon membranes (Amersham Life Science, Rockville, MD) using the VacuGene- XL vacuum blotting system (Pharmacia Biotech, Piscataway, NJ). After DNA fixation by UV irradiation, the membranes were probed in tubes with four different probes using ECL gold buffer (Amersham Life Science) following the exact protocol described in the manual (primary wash solution without urea at 55°, secondary wash solution containing 0.2% SSC). The primers used to generate the probes by PCR were (5' to 3') fesUP, GGTCTACATCACTGGTGTGA; fesDO, CGCAT GATCATCGGCTCGCG; ftsZUP, GCGGTAAATACCGATGCA CAAGC; ftsZDO, CATTCGGCGGGCCAGTTTAG; yjjMUP, GCAAGGGTGGCATCAAGGTC; yjjMDO, GACATGATTCGG CTCTCCAC; ddlBUP, TCGCGCTACACGGTCGCGGCGGTG; and ddlBDO, CGTACTACCAACTGCGAGAAGCTC. The probes for the fes (899 bp), ftsZ (1088 bp), yjjM (980 bp), and ddlB (721 bp) genes were expected to hybridize with EcoRI-EcoRI fragments of 3048, 2232, 1812, and 838 bp, respectively. The probes were labeled with [³²P]ATP using T4 polynucleotide kinase (SAMBROOK et al. 1989 ); 1 pmol of each probe was used for one blot. The hybridized membranes were exposed in a storm 860 (Molecular Dynamics, Sunnyvale, CA) and the intensity of each band was measured using the Image Quant program. Using MG1655 DNA, the response was linear for each probe over the range 1.7–7.5 µg DNA.

	RESULTS

TOP ABSTRACT MATERIAL AND METHODS RESULTS DISCUSSION LITERATURE CITED

Isolation of mini-Tn10 insertions conferring mecillinam resistance:
To avoid isolating mutants whose mecillinam resistance resulted from increased RelA-dependent synthesis of ppGpp, we carried out our selection in the relA strain CF6301, derived from the wild-type strain MG1655. We made random mini-Tn10 (Cm^r) inserts and selected simultaneously for resistance to chloramphenicol and mecillinam (at 1 or 10 µg/ml), as described previously (VINELLA et al. 1996 ). Of 69 Cm^r Mec^r colonies, 52 grew on purification on the same medium, 30 with 1 µg/ml and 22 with 10 µg/ml mecillinam. The insertions from these clones were transduced back into strain CF6301, and the Cm^r transductants were tested for mecillinam resistance. We retained for further study only the 12 mutants for which mecillinam resistance was 100% cotransducible with chloramphenicol resistance (48 Cm^r transductants tested). They were provisionally named mcr-4::cat to mcr-15::cat. One Cm^r transductant of each mutant was chosen and used for further analysis.

Stability of the mutants:
We tested the stability of the mini-Tn10 insertions by growing the Cm^r transductants at 37° for at least 10 generations in LB in the absence of chloramphenicol. The strains with insertions mcr-6, -7, -9, -10, -11, -13, and -14::cat did not segregate Cm^s clones in these conditions (0/48), whereas the five mutants with alleles mcr-4, -5, -8, -12, and -15::cat did (15–71%), indicating that these insertions were unstable. Loss of chloramphenicol resistance was systematically accompanied by loss of mecillinam resistance.

We previously described the isolation of three similar mecillinam-resistant mutants, selected in a relA1 strain by insertion of mini-Tn10 (Km^r). Two of these mutants, mcr-1::kan and mcr-2::kan, were similarly unstable, and their instability was shown to be RecA dependent (VIN- ELLA et al. 1996); the recA1 mutation also stabilized the insertions mcr-4, -5, -8, -12, and -15::cat.

Mecillinam resistance and ppGpp:
Our insertions, isolated in a relA mutant, also conferred mecillinam resistance on the wild-type strain MG1655 (Table 2). It was possible that their mecillinam resistance was due to a SpoT-dependent increase of the ppGpp pool. The relA strain cannot grow on minimal medium supplemented with glucose in the presence of serine, methionine, and glycine (SMG^s; UZAN and DANCHIN 1976 ) or in the presence of aminotriazole (AT^s; RUDD et al. 1985 ), conditions in which the cells are starved for isoleucine and valine or for histidine, respectively. Both SMG^s and AT^s phenotypes can be suppressed by spoT alleles that decrease the ppGpp hydrolase activity of the protein, thereby increasing the ppGpp concentration (SARUBBI et al. 1988 ). None of the mcr::cat insertions suppressed the SMG^s or AT^s phenotypes of the relA mutant CF6301, suggesting that they did not sufficiently increase the ppGpp pool. However, we could distinguish in the CF6301 strain two classes of insertions on the basis of the expression of the P1_rrnB::lacZ fusion, which is strongly repressed by ppGpp (Table 2): one class of mutations (mcr-7, -9, -10, and -14) strongly decreased the intensity of the blue coloration of the colonies obtained on LB plates containing X-gal, as if they increased the ppGpp pool, while none of the five unstable insertions (mcr-4, -5, -8, -12, and -15::cat) significantly affected the expression of the P1_rrnB::lacZ fusion, indicating that they did not affect the ppGpp level.

We transduced the insertions into a relA spoT206::kan derivative of MG1655 that is completely devoid of ppGpp and therefore auxotrophic for several amino acids (HERNANDEZ and BREMER 1991 ; XIAO et al. 1991 ). We recovered Cm^r transductants in all cases but one, in which the mcr-9::cat mutant was donor; this insertion is apparently lethal in the absence of ppGpp. None of the Cm^r insertions tested restored prototrophy in the absence of ppGpp, indicating that they have not opened an alternative pathway of ppGpp synthesis. All 12 insertions conferred mecillinam resistance on the relA strain (original selection) and also on transductants of the wild-type strain MG1655 (Table 2). In the ppGpp-deficient relA spoT206::kan strain, however, only the 5 unstable insertions conferred mecillinam resistance, whereas none of the 6 stable, viable insertions tested did so (Table 2). We conclude that the mecillinam resistance due to the unstable class of insertions mcr-4, -5, -8, -12, and -15::cat does not require ppGpp.

In conclusion, we found three classes of insertional mutants: class 1, mutants that had an increased ppGpp pool and required the presence of the nucleotide to be mecillinam resistant; class 2, mutants that seemed to have a normal ppGpp pool but, nevertheless, required the nucleotide to be mecillinam resistant; and class 3, mutants for which mecillinam resistance did not require an increased ppGpp pool and, furthermore, did not require the nucleotide at all. We have previously reported the characterization of mutants of the first two classes. Aminoacyl-tRNA synthetase mutants belong to the first class: the mecillinam resistance of these mutants (argS, alaS, and leuS) is due to a RelA-dependent increase of the ppGpp pool (VINELLA et al. 1992 , VINELLA et al. 1993 ). The aroK mutant (mcr-3::kan) belongs to the second class: its mecillinam resistance requires SpoT-dependent ppGpp synthesis (VINELLA et al. 1996 ). We showed that the AroK protein, one of two shikimate kinases in E. coli, has a second activity involved in mecillinam sensitivity and we proposed that it might be the phosphorylation of protein(s) involved in cell division (VIN- ELLA et al. 1996). The third class, represented by the five unstable insertion mutants, is a new class of mecillinam-resistant mutants, possibly comprising overproducers of the division proteins FtsZ, FtsA, and FtsQ. We thus focused our work on the further characterization of this class, described below.

Mecillinam resistance and FtsZ activity:
To test whether the mecillinam resistance conferred by the unstable insertions is associated with an increase in FtsZ activity, we first looked to see whether the insertions suppress the phenotype of an ftsZ84(Ts) mutant, since increased production of the mutant FtsZ84 protein is known to restore division in nonpermissive conditions (WANG et al. 1991 ). The insertions were transduced into the strain DV262 [relA1 ftsZ84(Ts)] and the plating efficiency of the mutant strains was measured on LB or LB lacking NaCl (LB*) plates at 30°, 37°, and 42°. The relA1 ftsZ84 mutant, as previously reported (WANG et al. 1991 ), grew on LB plates at all three temperatures (plating efficiency >95%) but was unable to form colonies on LB* even at 30° (plating efficiency 1 x 10^-2 at 30°, 2.5 x 10^-4 at 37°, and <1.3 x 10^-4 at 42°). All seven unstable insertions (five mcr::cat and two mcr::kan) suppressed the growth defect of the ftsZ84(Ts) mutant on LB* at 30° and 37° (plating efficiency >56%), suggesting that these mutations increase FtsZ activity, possibly by affecting regulators of FtsZ. These results are consistent with the hypothesis that the unstable class 3 insertions confer mecillinam resistance by increasing the concentration of the cell division proteins FtsZ, FtsA, and FtsQ.

Genetic mapping of the unstable insertions:
We first carried out Hfr mapping of the five mcr::cat unstable insertions. We transduced the mcr::cat alleles into Hfr strains injecting an early Tc^r marker (WANNER 1986 ) and crossed these with MG1655 relA::kan, selecting either Tc^r Km^r or Cm^r Km^r exconjugants on appropriately supplemented LB plates. All five Cm^r markers cotransferred efficiently with the thr::Tn10 marker (0 min) of strain BW6164 (HfrRa2), with >64% linkage between Cm^r and Tet^r, whereas no cotransfer occurred, for example, with the ilv::Tn10 marker (85 min) of strain BW6159 (KL14; <2% linkage). The insertions thus all seemed to be near 0 min. The mcr-1::kan and mcr-2::kan insertions isolated previously were also found to be linked to thr (data not shown).

We next transduced the insertions into strains carrying Tn10 markers regularly spaced (every 1–1.5 min) from 89.4 min (argE::Tn10) to 8.7 min (aroL::Tn10), selecting transductants on chloramphenicol- or kanamycin-containing LB plates. We never obtained Tc^s clones among at least 48 Cm^r or (in the case of mcr-1 and mcr-2) 48 Km^r clones analyzed for each transduction.

Determination of the insertion sites:
To determine the location of the unstable mini-Tn10 insertions, we first cloned their Cm^r marker. Chromosomal DNA of strains carrying the unstable class 3 insertions was extracted, completely digested with the restriction enzyme PstI, which does not cut within the transposable element, and cloned in the vector pCL1920 (Spc^r). These genomic libraries were used to transform strain XL1Blue, selecting on LB plates containing spectinomycin and chloramphenicol. Plasmid DNA was extracted from the Spc^r Cm^r clones and used to transform the same strain, XL1Blue, to verify that the plasmids carried both the Spc^r and the Cm^r markers. These plasmids were then partially sequenced using primers CmD and CmF to determine the joint between vector DNA and the cloned insert and to identify the adjacent chromosomal sequence. The sequences adjacent to the insertions, determined in this way, are shown on the left of Fig 1. At least four of the mutants appeared to result from an event more complex than simple insertion, involving an IS1 element.

View larger version (24K): [in this window] [in a new window] [Download PPT slide]   Figure 1. Physical structure of the joints between the mini-Tn10 inserts and flanking chromosomal sequences in the unstable mcr::cat and mcr::kan mutants and extent of diploidy. On the left are shown the genes flanking each mini-Tn10 insert and their orientation. The solid triangles represent the mini-Tn10. On the right, the genetic map is shown (not drawn to scale), with the chromosomal order and orientation of these genes. The heavy lines indicate the extent of the duplication, as judged by diploidy for the markers in the top line and by the end points defined by the flanking regions shown on the left.

The mcr-1::kan and mcr-2::kan inserts were similarly analyzed by cloning a fragment containing the kan determinant (Km^r) and adjacent chromosomal DNA in the vector pKS+. The right and left parts of the insertional element were then separately subcloned using the unique EcoRI restriction site present in the mini-Tn10 element. The joint between the insertional element and chromosomal DNA, determined using primers ISR and ISL, indicated that the mcr-1::kan and mcr-2::kan mutants also had complex genomic rearrangements (Fig 1).

Transduction of the insertions creates duplications:
In the course of strain constructions, we transduced the leu::Tn10 marker (1.8 min) from the strain MG1655 leu::Tn10 (phenotype Tc^r Leu^-) into the unstable mcr::cat derivatives of strain CF6301, selecting for Tc^r transductants. To our surprise, all Cm^r Tc^r transductants were able to grow on minimal glucose plates lacking leucine (Table 3), suggesting that the recipient strains carried a duplication of the leu operon. To get an idea of the extent of the putative duplications, we transduced the mcr::cat strains to argE::Tn10 (89.4 min), thr::Tn10 (0 min), proA::Tn10 (5.7 min), and aroA::Tn10 (20.7 min; Table 3). We also analyzed the two mcr::kan mutants for diploidy at these loci. The extent of the duplications, deduced from the results reported in Table 3, is shown in the right of Fig 1; all mutants except mcr-1::kan appeared to be diploid for at least two markers, and two displayed apparent duplication of three loci, from thr to proA, which are separated on the genetic map by 5.6 min (260 kb). Such long duplications cannot be transduced by P1, which encapsidates only 2 min (95 kb) of DNA. The apparent transduction of these mcr::cat alleles is explained below.

When P1 was grown on MG1655 leu::Tn10 (Tc^r Leu^-) and used to transduce derivatives of the wild-type (relA⁺) strain MG1655 carrying the insertions (mcr-2::kan, mcr-4, -5, -8, -12, and -15::cat) to Tc^r, diploidy for the leu operon could again be demonstrated: all Tc^r Cm^r transductants remained prototrophic (Leu⁺). This experiment was done with five transductants of MG1655 for each insertion. Our results indicate that the introduction of the five mcr::cat and the mcr-2::kan insertions by transduction was invariably correlated with the appearance of a duplication in the recipient strain. Moreover, we showed for at least one transductant of each insertion that the extent of the duplications was identical to that in the original mutant.

In reciprocal crosses, we transduced the mcr::cat and mcr::kan insertions from the CF6301 background into derivatives of the wild-type (relA⁺) strain MG1655 carrying either a thr::Tn10, leu::Tn10, or pro::Tn10 marker and phenotypically Thr^-, Leu^-, or Pro^-, respectively. As mentioned above, the Cm^r and Km^r transductants all remained tetracycline-resistant; we moreover observed that they also all remained auxotrophic (96 Cm^r or Km^r clones were tested for each transduction), indicating that the wild-type genes that were duplicated in the donor strains were not cotransduced with chloramphenicol or kanamycin resistance.

The fact that duplicated donor genes are not cotransduced with the mcr::cat markers presumably reflects the size of the duplications, which are longer than the 2 min that P1 can transduce. Nevertheless, introduction of the insertions seemed systematically to produce Cm^r transductants with duplications. We hypothesized that introduction of the insertions induced duplications of resident genes in the recipient strain. As just mentioned, when we transduced the mcr::cat insertions into strain MG1655 leu::Tn10, all Cm^r transductants remained Tc^r Leu^-. Our hypothesis would predict that they are in fact diploid for leu::Tn10. To test this, five such transductants, carrying the five mcr::cat insertions, were transduced to Leu⁺ with P1 grown on a wild-type donor. Leu⁺ Cm^r transductants were readily obtained with all five recipients, and indeed they remained Tc^r, confirming that these strains were diploid for the leu operon (cf. Fig 2).

View larger version (25K): [in this window] [in a new window] [Download PPT slide]   Figure 2. Genetic demonstration of diploidy at the leu and ftsZ loci in mcr::cat strains. Donor strains for P1 transductions are shown in boxes. Selected phenotypes are in parentheses. The unselected phenotypes found correspond to the expectation for diploid strains (left).

The mcr::cat mutants are diploid for ftsZ:
Since the mcr::cat strains are all diploid for the leu operon, which is closely linked to the ftsQ-ftsA-ftsZ operon, we wished to see whether they were also diploid for these cell division genes, potentially providing a mechanism for their mecillinam resistance (see Introduction). To test this, we took advantage of strain VIP205 (PALACIOS et al. 1996 ), in which the ftsZ gene has been cut off from its natural promoters by a Km^r insert that provides a lac promoter governing ftsZ expression. Since the FtsZ protein is absolutely required for division, this strain can grow only in the presence of an inducer of the lac operon. We introduced the 2-min region of VIP205 into the mcr::cat strains that were diploid for leu::Tn10, selecting transductants on minimal glucose plates lacking leucine and containing chloramphenicol and the lac operon inducer isopropyl thiogalactoside (IPTG) at 2 x 10^-5 M. Leu⁺ Cm^r transductants were readily obtained in each case. They were all Tc^r, and about half had also received the Km^r allele associated with the P_tac::ftsZ fusion. However, all these Tc^r Km^r Cm^r cotransductants were able to grow in the absence of IPTG, indicating the presence of a wild-type ftsZ allele in addition to the P_tac::ftsZ fusion (Fig 2).

These results show that the introduction of the mcr::cat insertions, in addition to duplicating the leu::Tn10 allele, also duplicated the ftsZ⁺ allele of the recipient strain. Since the ftsQ and ftsA genes lie between leu and ftsZ, they are presumably duplicated as well.

Physical evidence for amplification of the fts QAZ region in the mutants:
Cultures of all seven unstable mutants were grown in LB containing mecillinam; the control strain MG1655 was grown in LB. DNA was extracted from overnight cultures and analyzed by quantitative Southern blots (see MATERIALS AND METHODS), using four probes. From the genetic evidence, two of these, from genes ftsZ and ddlB, were expected to be amplified in all mutants. The third probe, from gene yjjM at 99.0 min, was expected to be amplified in mutants mcr-5, mcr-15, and possibly mcr-4. The fourth probe was from the fes gene at 13 min, which is not included in any of the duplications; it was used to normalize the data for each culture. The results (Fig 3 and Table 4) show that the ddlB and ftsZ genes are duplicated in all mutants and that the yjjM gene is duplicated only in mcr-5, mcr-15, and mcr-4. These data are in full agreement with the genetic data.

View larger version (90K): [in this window] [in a new window] [Download PPT slide]   Figure 3. Physical evidence for partial diploidy. The seven unstable mcr mutants were grown in LB containing mecillinam; the control strain MG1655 was grown in LB. DNA was extracted from overnight cultures and Southern blots were carried out with four probes, as described in MATERIALS AND METHODS.

Is a duplication sufficient to confer mecillinam resistance?
We have shown that overexpression of the FtsZ, FtsA, and FtsQ proteins confers mecillinam resistance (VINELLA et al. 1993 ; NAVARRO et al. 1998 ). Resistance was observed when the amount of FtsZ was increased sevenfold, but we did not determine the minimum amplification necessary for this phenotype. Since the mcr::cat insertions seemed to duplicate the ftsZ gene (and, presumably, the entire ftsQAZ operon), this could be the basis for their mecillinam resistance.

We tested whether duplication of the ftsQAZ operon conferred mecillinam resistance by introducing into MG1655 leu::Tn10 the F'104 episome, which carries the chromosomal region between 97 min and 6 min and is present at about one copy per chromosome. Leu⁺ Tc^r exconjugants were purified and tested for mecillinam resistance. They were sensitive, indicating that doubling the copy number of the ftsQAZ operon is not sufficient to confer mecillinam resistance.

Since the mcr::cat mutants all have duplications of the ftsQAZ operon, they might further amplify it by unequal crossing over (HAACK and ROTH 1995 ). To test this, we looked for evidence of triploidy of the leu operon. We transduced the leu::Tn10 marker into the mcr-2::kan mutant; the resulting Tc^r Km^r strain remained Leu⁺, as previously observed, indicating the presence of leu::Tn10 and leu⁺ alleles (Fig 2). We then transduced a leu::cat (Cm^r) marker into this strain, selecting the transductants on plates containing kanamycin and chloramphenicol. If the recipient strain had only two copies of the leu locus, we would expect to replace one of the resident alleles, giving Km^r Cm^r transductants that are either Tc^r Leu^- or Tc^s Leu⁺, according to whether the leu⁺ or the leu::Tn10 allele had been replaced, respectively. Indeed, we obtained these two types of transductants, but in addition we found Km^r Cm^r Tc^r Leu⁺ clones, indicating that in the culture of the recipient strain, at least 50% of the cells had three or more leu operons. An explanation of these results is presented in Fig 4.

View larger version (21K): [in this window] [in a new window] [Download PPT slide]   Figure 4. Genetic demonstration of triploidy at the leu locus in the mcr-2::kan mutant. An mcr-2::kan strain heterozygous for leu+/leu::Tn10 (phenotype Leu+ Tcr) was transduced to chloramphenicol resistance from a leu::Tn9 donor strain by selecting on LB plates containing chloramphenicol and kanamycin (to maintain the mcr-2::kan allele). The unselected phenotypes Leu+/- and Tcr/s were then tested. On the left are shown the possible phenotypes in the case of a diploid recipient; on the right are the expected phenotypes with a triploid recipient. Since all four possible phenotypes were found, at least part of the recipient population must be triploid at the leu locus.

Model:
These duplications are all unstable, they are stabilized in recA strains, they are diploid for markers in the ftsQAZ region near 2 min on the genetic map and covering as much as 250 kb, transduction of the mini-Tn10 inserts by P1 creates duplications of resident genes in the recipient strain, and the chromosomal sequences flanking the mini-Tn10 inserts are from different genes on the left and right. To explain these results, we propose a model (Fig 5) in which the initial event was a double transposition event or a transposition associated with a homologous recombination event, creating a tandem duplication with the mini-Tn10 located at the joint. When the mini-Tn10 is transduced by P1, the surrounding DNA carries the end points of the duplication, with the right end to the left and the left end to the right (Fig 5). This DNA fragment can then recombine with two sister chromosomes in the recipient cell to recreate the same duplication as in the donor chromosome, except that it is the recipient genes that are duplicated. Further amplification of the duplicated region can then occur by unequal crossing over.

View larger version (30K): [in this window] [in a new window] [Download PPT slide]   Figure 5. Model for the formation and transduction of duplications. In the diagram, the mini-Tn10 element, called cat, first transposes to a position between genes C and D. This chromosome then recombines with a sister chromosome in the region of homology between the mini-Tn10 and an IS1 sequence between genes K and L (the homology is represented by a cross-hatched bar), resulting in a duplication covering the genes from D to K, including the ftsQAZ operon. If this strain is used as a transductional donor, the phage can pick up the mini-Tn10 and flanking sequences (top box) and, by homologous recombination with two sister chromosomes, recreate the same duplication as in the donor strain. This duplication, by unequal crossing over between sister chromosomes (bottom box), can generate a triploid chromosome, which in turn can further amplify the number of copies of the duplicated region.

This model makes several predictions. First of all, the duplications created by transducing the mini-Tn10 inserts should have exactly the same end points as those in the donor strains. We have shown above that for the thr, leu, and proA operons, the duplications in transductants have the same extent as in those in the donors. Second, on the basis of the sequences assumed to define the duplication end points, one predicts that the mcr-5::cat mutant should have a duplication of the purA gene while mcr-15::cat should not (Fig 1). When we transduced these mutants with a P1 stock grown on a purA45 zjd::Tn10 donor strain, Tc^r Pur^- cotransductants were never found when the recipient was mcr-5::cat (<1%) but readily recovered with mcr-15::cat (30%).

Third, it should be possible to cotransduce donor markers that are very close to the mini-Tn10, and indeed we found this with our smallest duplication, mcr-1::kan. In one experiment, with a leu::Tn10 recipient, we found one Tc^s clone among 125 Km^r transductants. With an ftsI23(Ts) recipient, we found 90% cotransduction of temperature resistance (ftsI⁺) and Km^r. Note that the leu operon is not duplicated in mcr-1::kan and is 1.6 min from the mini-Tn10, whereas the ftsI gene, between leuO and ftsQ, is covered by the duplication with the two copies located 0.2 and 1.3 min from the mini-Tn10 (cf. Fig 1). We therefore expected the ftsI⁺ cotransductants to be diploid and, for most of them, heterozygous. Indeed, when the donor strain was ftsI23(Ts) mcr-1::kan and the recipient wild type, only three temperature-sensitive clones were found among 144 Km^r transductants. These had presumably received both donor genes and were thus homozygous for the ftsI23(Ts) allele. Their temperature sensitivity suggests that further amplification of this allele by unequal crossing over does not restore temperature-resistant division, unlike the ftsZ84(Ts) allele.

The instability of these tandem duplications arises from RecA-dependent recombination between the duplicated sequences. As a fourth prediction of our model, one would expect heterozygous leu⁺ leu::Tn10 diploids to segregate two types of haploid, Tc^r Leu^- and Tc^s Leu⁺, according to where the crossover takes place. Heterozygous diploids of this sort, carrying the mcr-2::kan allele or any one of the five mcr::cat alleles, were grown and plated without antibiotics, and Km^s or Cm^s segregants were looked for. All six mutants gave rise to both types of segregants; the ratio of Tc^s Leu⁺ to Tc^r Leu^- colonies ranged from 0.08 to 27, according to the mcr allele.

Fifth, the mechanism proposed for recreation of the duplications by P1 transduction requires at least three crossover events, compared to a double crossover in the normal transduction process, so transduction of our mini-Tn10 inserts would be expected to be less efficient than normal transduction. Indeed, we consistently found some 10-fold fewer transductants when selecting one of these mini-Tn10 inserts as compared to the number of transductants obtained in straightforward selections for markers such as leu::Tn10.

A sixth prediction of our model concerns the presumed further amplification by unequal crossing over. As shown above, cultures of the mcr-1::kan mutant can be at least triploid for the leu operon. However, this amplification, which is absolutely necessary for mecillinam resistance, should require RecA-dependent recombination. We found that the recA1 mutation, in addition to stabilizing the mini-Tn10 element, also suppressed the mecillinam resistance of all seven unstable mcr mutants.

	DISCUSSION

TOP ABSTRACT MATERIAL AND METHODS RESULTS DISCUSSION LITERATURE CITED

In this article, using insertional mutagenesis, we carried out a selection for mutants that overproduce the division proteins FtsQ, FtsA, and FtsZ, in the hope of identifying regulators of this complex operon. Of 12 new and 3 previously selected mecillinam-resistant mutants, 7 were of this type. However, in all cases the amplification was produced by increasing the gene copy number rather than by increasing the rate of transcription initiation. Since these duplications seemed to result from multiple events, it seems unlikely that a single insertional mutation can increase ftsQAZ transcription sufficiently to confer mecillinam resistance.

Gene amplification was brought about by creating a tandem duplication with the mini-Tn10 at the joint between the repeated sequences (Fig 5). The duplications are too long to be encapsidated by the transducing phage P1, but they can be recreated by transduction. The mechanism proposed requires encapsidating the deletion end points and the adjacent mini-Tn10, which carries the selective marker (cf. Fig 5).

It is striking that four of the seven duplications described here have one end point in an IS1 element. On examining the sequence of the mini-Tn10, we noted that it has a 58-bp stretch that is homologous to one end of IS1 (one difference compared with IS1B and IS1C). Thus the initial formation of the duplications in these cases seems to have involved a transposition event, determining one duplication end point, and a homologous recombination event between the mini-Tn10 and an IS1, determining the other end point. In Salmonella typhimurium, duplications have been shown to form between separate IS200 elements, presumably by unequal crossing over (HAACK and ROTH 1995 ). In the mcr-4::cat mutant, the sequences flanking the mini-Tn10 are both from the open reading frame yaeT; however, the strain is diploid for thr and leu, showing that the structure, more complex than a simple insertion in yaeT, includes a duplication in the 2-min region, as in the other mutants.

Our model for duplication formation suggests a general method for constructing strains with a precise tandem duplication of virtually any sequence on the chromosome. To do this, the deletion end points should first be cloned in a plasmid vector, in the proper orientation, next to a selective marker (Tn, for example). If one wants to duplicate, say, the sequence DEFGHIJK, the right end point JK should be cloned to the left of the Tn marker and the left end point DE should be cloned to the right, taking care to maintain the normal chromosomal orientation of each sequence. The cloned fragment JK-Tn-DE can then be separated from plasmid replication functions and used in linear transformation, where, by the mechanism shown in Fig 5, it will create a tandem duplication having the structure AB CDEFGHIJK-Tn-DEFGHIJKLM.

In our system, a mere doubling of the number of ftsQAZ operons is not enough to confer mecillinam resistance. Nevertheless, essentially all cells carrying a tandem duplication of the operon succeed in forming a colony on mecillinam plates, apparently through further amplification by unequal crossing over. A selection of this sort can be included in the construction of a designed duplication. For this, one need only include, together with the selectable Tn element, genes whose further amplification can be selected for. The ftsQAZ genes would be one possibility; growth in the presence of mecillinam would then produce a population with more than two copies of the designed duplication. Other drug resistance markers could also be used, as well as metabolic functions with insufficient expression.

Our selection produced the phenotype we expected: 7 of our 15 mecillinam-resistant mutants overproduce FtsQ, FtsA, and FtsZ (Table 2, class 3), and at least 4 of the others seem to have a higher than normal ppGpp pool (Table 2, class 1). The fact that our class 3 overproducers did not define regulators of the ftsQAZ operon illustrates once again that there are many ways to skin a cat, and E. coli is more clever than we are by eluding our attempt to define regulators by insertion mutagenesis. We are currently completing the characterization of the mutants of classes 1 and 2 and selecting new mutants that become mecillinam resistant via overproduction of specific gene products.

	ACKNOWLEDGMENTS

We thank Bénédicte Gagny for her participation in the analysis of the mcr-1::kan and mcr-2::kan mutants, Peter Kuempel for stimulating discussions, and Evelyne Maillet for her precious help in the Southern blot experiments. D. Vinella was supported in part by a Fogarty grant International Fellowship. This work was supported in part by grant 9981 from the Association pour la Recherche sur le Cancer.

Manuscript received March 14, 2000; Accepted for publication August 14, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIAL AND METHODS RESULTS DISCUSSION LITERATURE CITED

BACHMANN, B. J., 1996 Derivations and genotypes of some mutant derivatives of Escherichia coli K-12, pp. 2460–2488 in Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology, edited by F. C. NEIDHARDT, J. L. INGRAHAM, K. B. LOW, B. MAGASANIK, M. SCHAECHTER and H. E. UMBARGER. American Society for Microbiology, Washington, DC.

BULLOCK, W. O., J. M. FERNANDEZ, and J. M. SHORT, 1987  XL1-Blue: a high efficiency plasmid transforming recA Escherichia coli strain with beta-galactosidase selection. BioTechniques 5:376-378.

CASHEL, M., 1969  The control of ribonucleic acid synthesis in Escherichia coli. IV. Relevance of unusual phosphorylated compounds from amino acid-starved stringent strains. J. Biol. Chem. 244:3133-3141[Abstract/Free Full Text].

CASHEL, M. and J. GALLANT, 1969  Two compounds implicated in the function of the RC gene of Escherichia coli.. Nature 221:838-841[Medline].

CASHEL, M., D. R. GENTRY, V. J. HERNANDEZ and D. VINELLA, 1996 The stringent response, pp. 1459–1496 in Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology, edited by F. C. NEIDHARDT, J. L. INGRAHAM, K. B. LOW, B. MAGASANIK, M. SCHAECHTER and H. E. UMBARGER. American Society for Microbiology, Washington, DC.

CHATTERJI, D., N. FUJITA, and A. ISHIHAMA, 1998  The mediator for stringent control, ppGpp, binds to the ß-subunit of Escherichia coli RNA polymerase. Genes Cells 3:279-287[Abstract].

HAACK, K. R. and J. R. ROTH, 1995  Recombination between chromosomal IS200 elements supports frequent duplication formation in Salmonella typhimurium.. Genetics 141:1245-1252[Abstract].

HERNANDEZ, V. J. and H. BREMER, 1990  Guanosine tetraphosphate (ppGpp) dependence of the growth rate control of rrnB P1 promoter activity in Escherichia coli.. J. Biol. Chem. 265:11605-11614[Abstract/Free Full Text].

HERNANDEZ, V. J. and H. BREMER, 1991  Escherichia coli ppGpp synthetase II activity requires spoT.. J. Biol. Chem. 266:5991-5999[Abstract/Free Full Text].

JAMES, R., J. Y. HAGA, and A. B. PARDEE, 1975  Inhibition of an early event in the cell division cycle of Escherichia coli by FL 1060, an amidinopenicillanic acid. J. Bacteriol. 122:1283-1292[Abstract/Free Full Text].

JOSELEAU-PETIT, D., D. THÉVENET, and R. D'ARI, 1994  ppGpp concentration, growth without PBP2 activity and growth rate control in Escherichia coli.. Mol. Microbiol. 13:911-917[Medline].

JOSELEAU-PETIT, D., D. VINELLA, and R. D'ARI, 1999  Metabolic alarms and cell division in Escherichia coli.. J. Bacteriol. 181:9-14[Free Full Text].

KLECKNER, N., J. BENDER, and S. GOTTESMAN, 1991  Uses of transposons with emphasis on Tn10.. Methods Enzymol. 204:139-180[Medline].

LAZZARINI, R. A., M. CASHEL, and J. GALLANT, 1971  On the regulation of guanosine tetraphosphate levels in stringent and relaxed strains of Escherichia coli.. J. Biol. Chem. 246:4381-4385[Abstract/Free Full Text].

LERNER, C. G. and M. INOUYE, 1990  Low copy number plasmids for regulated low-level expression of cloned genes in Escherichia coli with blue/white insert screening capability. Nucleic Acids Res. 18:4631[Free Full Text].

LUND, F. and L. TYBRING, 1972  6ß-amidinopenicillanic acid—a new group of antibiotics. Nat. New Biol. 236:135-137[Medline].

MATSUHASHI, S., T. KAMIRYO, P. M. BLUMBERG, P. LINNETT, and E. WILLOUGHBY et al., 1974  Mechanism of action and development of resistance to a new amidino penicillin. J. Bacteriol. 117:578-587[Abstract/Free Full Text].

METZGER, S., G. SCHREIBER, E. AIZENNMAN, M. CASHEL, and G. GLASER, 1989  Characterization of the relA1 mutation and comparison of relA1 with new relA null alleles in Escherichia coli.. J. Biol. Chem. 264:21146-21152[Abstract/Free Full Text].

MILLER, J. H., 1992 A Short Course in Bacterial Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

NAVARRO, F., A. ROBIN, R. D'ARI, and D. JOSELEAU-PETIT, 1998  Analysis of the effect of ppGpp on the ftsQAZ operon in Escherichia coli.. Mol. Microbiol. 29:815-823[Medline].

PALACIOS, P., M. VICENTE, and M. SÁNCHEZ, 1996  Dependency of Escherichia coli cell-division size, and independency of nucleoid segregation on the mode and level of ftsZ expression. Mol. Microbiol. 20:1093-1098[Medline].

ROBINSON, A. C., D. J. KENAN, G. F. HATFULL, N. F. SULLIVAN, and R. SPIEGELBERG et al., 1984  DNA sequence and transcriptional organization of essential cell division genes ftsQ and ftsA of Escherichia coli: evidence for overlapping transcriptional units. J. Bacteriol. 160:546-555[Abstract/Free Full Text].

ROBINSON, A. C., D. J. KENAN, J. SWEENEY, and W. D. DONACHIE, 1986  Further evidence for overlapping transcriptional units in an Escherichia coli cell envelope-cell division gene cluster: DNA sequence and transcriptional organization of the ddl ftsQ region. J. Bacteriol. 167:809-817[Abstract/Free Full Text].

RUDD, K. E., B. R. BOCHNER, M. CASHEL, and J. R. ROTH, 1985  Mutations in the spoT gene of Salmonella typhimurium: effects on his operon expression. J. Bacteriol. 163:534-542[Abstract/Free Full Text].

RYALS, J., R. LITTLE, and H. BREMER, 1982  Control of rRNA and tRNA syntheses in Escherichia coli by guanosine tetraphosphate. J. Bacteriol. 151:1261-1268[Abstract/Free Full Text].

SAMBROOK, J., E. F. FRITSCH and T. MANIATIS, 1989 Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SARUBBI, E., K. E. RUDD, and M. CASHEL, 1988  Basal ppGpp level adjustment shown by new spoT mutants affects steady state growth rates and rrnA ribosomal promoter regulation in Escherichia coli.. Mol. Gen. Genet. 213:214-222[Medline].

SINGER, M., T. A. BAKER, G. SCHNITZLER, S. M. DEISCHEL, and M. GOEL et al., 1989  A collection of strains containing genetically linked alternating antibiotic resistance elements for genetic mapping of Escherichia coli.. Microbiol. Rev. 53:1-24[Abstract/Free Full Text].

SPRATT, B. G., 1975  Distinct penicillin binding proteins involved in the division, elongation, and shape of Escherichia coli K-12. Proc. Natl. Acad. Sci. USA 72:2999-3003[Abstract/Free Full Text].

SPRATT, B. G., 1977  The mechanism of action of mecillinam. J. Antimicrob. Chemother. 3(Suppl. B):13-19.

SPRATT, B. G. and A. B. PARDEE, 1975  Penicillin-binding protein and cell shape in E. coli.. Nature 254:515-517[Medline].

UZAN, M. and A. DANCHIN, 1976  A rapid test for the relA mutation in E. coli.. Biochem. Biophys. Res. Commun. 69:751-758[Medline].

VINELLA, D., R. D'ARI, A. JAFFÉ, and P. BOULOC, 1992  Penicillin-binding protein 2 is dispensable in Escherichia coli when ppGpp synthesis is induced. EMBO J. 11:1493-1501[Medline].

VINELLA, D., D. JOSELEAU-PETIT, D. THÉVENET, P. BOULOC, and R. D'ARI, 1993  Penicillin-binding protein 2 inactivation in Escherichia coli results in cell division inhibition, which is relieved by FtsZ overexpression. J. Bacteriol. 175:6704-6710[Abstract/Free Full Text].

VINELLA, D., B. GAGNY, D. JOSELEAU-PETIT, R. D'ARI, and M. CASHEL, 1996  Mecillinam resistance in Escherichia coli is conferred by loss of a second activity of the AroK protein. J. Bacteriol. 178:3818-3828[Abstract/Free Full Text].

WANG, X., P. A. J. DE BOER, and L. I. ROTHFIELD, 1991  A factor that positively regulates cell division by activating transcription of the major cluster of essential cell division genes of Escherichia coli.. EMBO J. 10:3363-3372[Medline].

WANNER, B. L., 1986  Novel regulatory mutants of the phosphate regulon in Escherichia coli K-12. J. Mol. Biol. 191:39-58[Medline].

XIAO, H., M. KALMAN, K. IKEHARA, S. ZEMEL, and G. GLASER et al., 1991  Residual guanosine 3',5'-bispyrophosphate synthetic activity of relA null mutants can be eliminated by spoT null mutations. J. Biol. Chem. 266:5980-5990[Abstract/Free Full Text].

YI, Q.-M., S. ROCKENBACH, J. E. J. WARD, and J. LUTKENHAUS, 1985  Structure and expression of the cell division genes ftsQ, ftsA, and ftsZ.. J. Mol. Biol. 184:399-412[Medline].

This article has been cited by other articles:

				B. A. Legaree, C. B. Adams, and A. J. Clarke Overproduction of Penicillin-Binding Protein 2 and Its Inactive Variants Causes Morphological Changes and Lysis in Escherichia coli J. Bacteriol., July 15, 2007; 189(14): 4975 - 4983. [Abstract] [Full Text] [PDF]

				B. A. Legaree, K. Daniels, J. T. Weadge, D. Cockburn, and A. J. Clarke Function of penicillin-binding protein 2 in viability and morphology of Pseudomonas aeruginosa J. Antimicrob. Chemother., March 1, 2007; 59(3): 411 - 424. [Abstract] [Full Text] [PDF]

				A. Costanzo and S. E. Ades Growth Phase-Dependent Regulation of the Extracytoplasmic Stress Factor, {sigma}E, by Guanosine 3',5'-Bispyrophosphate (ppGpp) J. Bacteriol., July 1, 2006; 188(13): 4627 - 4634. [Abstract] [Full Text] [PDF]

				K. Potrykus, D. Vinella, H. Murphy, A. Szalewska-Palasz, R. D'Ari, and M. Cashel Antagonistic Regulation of Escherichia coli Ribosomal RNA rrnB P1 Promoter Activity by GreA and DksA J. Biol. Chem., June 2, 2006; 281(22): 15238 - 15248. [Abstract] [Full Text] [PDF]

				H. Nicoloff, V. Perreten, L. M. McMurry, and S. B. Levy Role for Tandem Duplication and Lon Protease in AcrAB-TolC- Dependent Multiple Antibiotic Resistance (Mar) in an Escherichia coli Mutant without Mutations in marRAB or acrRAB. J. Bacteriol., June 1, 2006; 188(12): 4413 - 4423. [Abstract] [Full Text] [PDF]

				Z. Ding and P. J. Christie Agrobacterium tumefaciens Twin-Arginine-Dependent Translocation Is Important for Virulence, Flagellation, and Chemotaxis but Not Type IV Secretion J. Bacteriol., February 1, 2003; 185(3): 760 - 771. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Vinella, D. Articles by D'Ari, R. Search for Related Content PubMed