Identification d’inhibiteurs protéiques contre les protéines de division cellulaire FtsZ et FtsA de Pseudomonas aeruginosa en tant que nouvelle classe potentielle d’antimicrobiens

L’ère révolutionnaire des antibiotiques est aujourd’hui dépassée par la capacité évolutive des microorganismes à développer des mécanismes de résistance. D’ailleurs, la résistance accrue du pathogène opportuniste P. aeruginosa diminue l’efficacité de traitement et met en danger les patients immono-déprimés ou souffrants de fibrose kystique. Dans le but d’identifier de nouveaux agents antimicrobiens, nous utilisons la machinerie de division cellulaire bactérienne en tant que cible. Nous étudions plus particulièrement les protéines FtsZ et FtsA de P. aeruginosa ; deux protéines de division cellulaire essentielles et hautement conservées. Les gènes ftsZ and ftsA ont tout d’abord été clonés et les protéines ont été respectivement purifiées par chromatographie et par renaturation de corps d’inclusion purifiés. Le séquençage en N-terminal a confirmé l’identité des protéines et un essai de chromatographie sur couche mince a également confirmé l’activité enzymatique respective de FtsZ et FtsA. Un essai de fixation à l’UV a démontré que FtsZ lie préférentiellement le GTP et que FtsA fixe préférentiellement l’ATP parmi les 4 nucléotides marqués. Les enzymes purifiées ont ensuite été utilisées pour identifier des peptides inhibiteurs avec la technique de présentation phagique. Les phages possédant une affinité spécifique pour FtsZ ou FtsA ont été élués selon 4 conditions spécifiques; la glycine à pH acide, le substrat de l’enzyme, un analogue non-hydrolysable du substrat et avec FtsA pour FtsZ et vice-versa. Nous avons identifié 3 consensus peptidiques contre ftsZ et 2 contre FtsA. La spécificité de l’interaction entre les peptides sélectionnés et FtsZ ou FtsA a été évaluée par ELISA et cela a parmi d’identifier les peptides les plus affins. Les 3 consensus peptidiques sélectionnés contre FtsZ ont été synthétisés et testés sur l’activité GTPase de FtsZ. Les 2 peptides C7C inhibent l’activité enzymatique de FtsZ avec des valeurs de CI50 de 0.45 mM et 1.2 mM alors que le peptide 12-mer montre une CI50 de 5 mM. Ainsi, la technique de présentation phagique utilisée sur de nouvelles cibles bactériennes a permi la découverte de peptides inhibiteurs et le peptidomimétisme devrait permettre le développement d’une nouvelle classe d’agents antimicrobiens.

Identification of Pseudomonas aeruginosa FtsZ and FtsA Peptide Inhibitors as a Potential Novel Class of Antimicrobials

Catherine Paradis-Bleau, François Sanschagrin and Roger C. Levesque*

Centre de recherche sur la fonction, structure et ingénierie des protéines, Faculté de médecine, pavillon Charles-Eugène-Marchand, 1 Université Laval, Sainte-Foy, Québec, Canada G1K 7P4

Running title: Novel FtsZ and FtsA peptide inhibitors

*Corresponding author. Mailing address: Centre de recherche sur la fonction, structure et ingénierie des protéines, Faculté de médecine, pavillon Charles-Eugène-Marchand, Université Laval, Sainte-Foy, Québec, Canada, G1K 7P4. Phone: (1) (418) 656-3070. Fax: (1) (418) 656-7176. E-mail: rclevesq@rsvs.ulaval.ca.

The recent development in genomics, proteomics and bio-informatics areas allows the identification of many uncommon bacterial targets {Projan, 2002 #22;Breithaupt, 1999 #20}. The perfect bacterial target should be essential for bacterial survival and pathogenesis, highly conserved in bacterial evolution but not in the animal kingdom, easily accessible and sufficiently expressed during the infection in the host. To meet these criteria, we selected the prokaryotic cell division machinery as a tool to identify the most attractive new targets. This choice is based upon the intrinsic properties of the cell division process encoding essential proteins and giving a lethal phenotype when inhibited. These proteins are highly conserved in bacterial species but absent in eukaryotic cells and extremely sensitive to inhibition because the process depends on specific protein cascade recruitment {Errington, 2003 #51;Projan, 2002 #22;Dai, 1992 #29}. No new antimicrobial agent specifically targets this process and it will be of interest to see if the prokaryotic kingdom possesses any resistance mechanisms against cell division inhibitors. An additional observation that confirms the validity of the cell division machinery as targets is that intrinsic division inhibitors MinC {Errington, 2003 #51} and SulA {Cordell, 2003 #63} which constitute active regulator of cell division do not select resistance. However, the bacterial cell division process possesses some limitations as drug target due to its cytoplasmic localization. Among bacterial division proteins, we chose FtsZ and FtsA as specific targets. These two most conserved proteins {van den Ent, 2001 #50} presumably constitute key bacterial targets because of their essential enzymatic activities {Errington, 2003 #51} that can be exploited to screen and analyse inhibitory molecules. FtsZ is the most important cell division protein at the top of hierarchic recruitment in the divisosome and its polymerisation into the Z-ring allows the physically separation of daughter cells {Lutkenhaus, 1997 #30}. The FtsA protein is also essential for protein recruitment in the divisosome {Pichoff, 2002 #32} and for the constriction process of the Z-ring {Begg, 1998 #38}. Accumulating evidences suggest that FtsA plays the role of a motor protein in providing energy for constriction {Nanninga, 1998 #31;Feucht, 2001 #27;Errington, 2003 #51}. Finally, the essential FtsZ-FtsA protein-protein interaction {Yan, 2000 #39} represent a more sensitive and specific target than the GTPase FtsZ and ATPase FtsA alone {Haney, 2001 #41}. Thus, these cytoplasmic essential enzymes and interaction remain to be fully characterized in the aim of developing antibacterials with novel modes of action.

In order to identify specific inhibitors of FtsZ and FtsA enzymes, we used phage-display which represents a powerful technique for the selection of short peptide ligands having high binding affinities to proteins of interest among a large pool of random peptide permutations {Christensen, 2001 #57;Sidhu, 2000 #56}. This approach was already proved useful for the detection of various enzyme inhibitors {Hyde-DeRuyscher, 2000 #64} and hence became one of the valuable tools for antibacterial drug discovery.

In this paper, we first describe the purification and biochemical characterization of FtsZ and FtsA cell division proteins. We then systematically exploit the use of two different phage display libraries and identification of specific peptides. Three rounds of phage display screening led to the identification of clear consensus peptide sequences presumably active against FtsZ and FtsA. High affinity binding peptides were also identified. The three FtsZ related synthesized peptides showed specific inhibition of its GTPase enzymatic activity. In order to obtain promising lead compounds with appropriate pharmacologically properties, inhibitory peptides will undergo many chemical modifications and their sequences will constitute the core for the synthesis of libraries of peptidomimetic molecules.

Plasmid construction, bacterial strains, reagents and media

All reagents were purchased from Sigma Aldrich (Oakville, Ontario, Canada) unless otherwise indicated. Restriction endonucleases and T4 ligase were obtained from New England Biolabs (Mississauga, Ontario, Canada). Agarose gel electrophoresis and plasmid DNA preparations were performed according to published procedures {Sambrook, 1989 #60}. Recombinant plasmids containing P. aeruginosa PAO1 ftsZ and ftsA genes were conserved in Escherichia coli NovaBlue, endA1 hsdR17 ( rK12 - mK12 +) supE44 thi-1 recA1 gyrA96 relA1 lac [F’ proA + B + laclqZ∆ M15 ::Tn 10 ] (Novagen, Madison, WI, USA) prior to protein synthesis in E. coli BL21, F- ompT hsdSB ( rB - mB -) gal dcm (DE3) (Novagen). TBS was 50 mM Tris-HCl (pH 7.5), 150 mM NaCl. TBST was TBS + 0.1% [v/v] Tween-20. PEG/NaCl was 20% [w/v] polyethylene glycol-8000, 2.5 M NaCl. Agarose Top (per liter) was: 10 g Bacto-Tryptone, 5 g yeast extract, 5 g NaCl, 1 g MgCl2•6H2O, 7 g agarose. IPTG/Xgal was 1.25 g isopropyl β-D-thiogalactoside (Roche Diagnostics, Laval, Québec, Canada) and 1 g 5-bromo-4-chloro-3-indolyl- β-D-galactoside dissolved in 25 mL dimethyl formamide. LB/IPTG/Xgal plates were prepared by autoclaving 1 L of LB broth (EM Science, Gibbstown, NJ, USA) + 15 g agar, the medium was cooled to < 70°C, 1 mL IPTG/Xgal was added, plates were stored at 4°C in the dark. Host strain used for production, amplification, and determination of the titer of bacteriophage was E. coli ER2537, F’ laclq∆ ( lacZ ) M15 proA + B + / fhuA2 supE thi ∆ ( lac-proAB ) ∆ ( hsdMS-mcrB ) 5 ( rK - mK - McrBC-), (New England Biolabs).

Cloning of P. aeruginosa ftsZ, ftsA and DNA sequencing

Polymerase chain reaction (PCR) cloning was used to obtain FtsZ and FtsA proteins with a His-tag at their C-terminal by deleting the original stop codons and adding 6 histidine amino acids before a termination codon. The ftsZ gene was amplified from genomic DNA of P. aeruginosa PAO1 (70 ng) with forward and reverse primers FtsZ- Nde I 5’-GGA GAG GGC ATA TGT TTG AAC TGG-3’ and FtsZ- Xho I 5’-TTA CTT CAC TCG AGC TGA CGA CGC-3’ designed to contain appropriate restriction sites Nde I and Xho I. The ftsZ gene was amplified in frame with the His-tag changing the two last amino acids of the C-terminal alanine-aspartic acid for leucine-glutamic acid. PCR was done with primers at 0.1 µM each, dNTP (Amersham Biosciences) at 0.2 mM each and MgCl2 at 2 mM in a final volume of 50 µL. PCR conditions were optimized as follows: after 7 min at 95°C, 2.5 U of the Expand Hight Fidelity TAQ polymerase (Roche Diagnostics) was added and DNA amplification was done using 30 cycles of denaturation at 95°C for 1 min, annealing at 60°C for 1 min and extension at 73°C for 1.5 min. The ftsA gene was amplified from genomic DNA of P. aeruginosa PAO1 (70 ng) with reverse and forward primers FtsA-Ctforw 5’-GTA ATA CAT ATG GCA AGC GTG CAG A-3’ and FtsA-Ctrev 5’-TTG AAG CTT AAT TGC CCT GGA CCC-3’ designed to contain appropriate restriction sites Hind III and Nde I. Amplification of ftsA insert was done with primers at 0.1 µM each, dNTP (Amersham Biosciences) at 0.2 mM each, MgCl2 at 1.5 mM and 2.5 U of Hot start TAQ polymerase (Qiagen, Mississauga, Ontario, Canada) in a final volume of 50 µL. PCR conditions were optimized as follows: 30 cycles, denaturation at 94°C for 1 min, annealing at 56.3°C for 1 min, extension at 72°C for 1 min and a final step at 72°C for 10 min after addition of the Hot start 15 min at 95°C. PCR products were purified using Qiaquick PCR purification kit (Qiagen) and were digested with appropriate restriction enzymes. DNA insert containing ftsZ was cloned into the corresponding sites of the expression vector pET24b (Novagen) and the ftsA insert was cloned into the expression vector pET30a (Novagen); both under the control of the bacteriophage T7 promoter. The recombinant plasmids pMON2020 containing the ftsZ insert and pMON2023 with ftsA insert were electroporated into competent E. coli NovaBlue and hept at -80°C. The DNA inserts in recombinant plasmids were sequenced using the T7 promoter primer and the T7 terminator primer (Novagen). Sequence data were compared with the ftsZ and ftsA P. aeruginosa PAO1 sequences using the programs from the Wisconsin Package Version 10.3 (Accelrys Inc., San Diego, CA, USA).

Overexpression and purification of FtsZ

The recombinant plasmid pMON2020 was introduced into the competent E. coli host strain BL21 (λDE3) (Novagen) by classic transformation for expression of FtsZ with an His-Tag at the C-terminus. Two cultures of 950 mL of Luria-Bertani (LB) Broth (Difco Laboratories, Detroit, MI, USA) supplied with 0.05 g/L of kanamycin were inoculated with an overnight culture of 50 mL of E. coli BL21 (λDE3) containing pMON2020 in LB broth with 0.05 g/L of kanamycin and were incubated at 37°C under agitation at 250 rpm until an optical density (O.D.) of 0.8 at 600nm. The induction was done by adding 1 mM of IPTG and overexpression was done for 4 h at 37°C under agitation at 250 rpm. Cells were centrifuged for 15 min at 7 000 g at 4°C and the cell pellets were immediately stored at -80°C. Pellets were resuspended 3.33 times their weight values in sonication buffer (50mM Tris-HCl pH 8.6 and 2 mM EDTA) and treated with lysosyme with 100 μg/mL for 15 min at 30°C. Cells were lysed by sonication for 30 sec/mL using a Virsonic digital 475 ultrasonic cell disrupter (The Virtis Company, Gardiner, New-York, USA) and a protease inhibitor cocktail was immediately added as recommended by the manufacturer (Roche Diagnostics). Cells and cellular debris were then removed by centrifugation at 17 000 g 30 min. Recombinant FtsZ protein was purified to homogeneity on an affinity His-bind nickel resin chromatography (Novagen) as recommended by the manufacturer and the elution was optimized at 150 mM imidazole, 0.5 M NaCl, 20 mM Tris-HCl pH 7.9. Proteins fractions were visualized by SDS-PAGE analysis {Sambrook, 1989 #60} and fractions with sufficient purity were pooled, dialysed in conservation buffer (20 mM Tris-HCl pH 7.6, 10 mM NaCl and 1 mM EDTA) {Zhulanova, 1998 #61} with Slide-A-Lyser Mini Dialysis (Pierce, Rockford, IL, USA) and conserved in aliquots in 50% (v/v) Glycerol at -80°C. FtsZ protein concentration was measured by the Bradford method (BioRad, Mississauga, Ontario, Canada).

Overexpression of FtsA

The recombinant plasmid pMON2023 was introduced into the competent E. coli host strain BL21 (λDE3) (Novagen) by classic transformation for expression of FtsA with His-tag at the C-terminus. Two cultures of 930 mL of LB Broth (Difco) supplied with 0.05 g/L of kanamycin were each inoculated with an overnight culture of 70 mL of E. coli BL21 (λDE3) carrying pMON2023 in LB broth with 0.05 g/L of kanamycin and were incubated at 37°C under agitation at 260 rpm until it reach an O.D. of 0.65 at 600nm. The induction was done by adding 1 mM of IPTG and overexpression was carried out during 6 h. Cells were centrifuged and the pellets were immediately stored at -80°C.

Purification of inclusion bodies and refolding of FtsA

Pellets were resuspended in 1B Wash buffer from the Protein Refolding Kit (Novagen) as recommended by the manufacturer and treated with lysosyme at 100 μg/mL 15 min at 30°C. Cells were lysed by sonication as for FtsZ and protease inhibitor cocktail tablets were immediately added as recommended by the manufacturer (Roche Diagnostics). Purification of inclusion bodies containing FtsA was optimized as follows; disrupted cells were centrifuged for 10 min at 5 000 g ; pellets were resuspended in 1B Wash buffer (Novagen) as recommended by the manufacturer and centrifuged for 10 min at 5 000 g and pellets were resuspended again in 1B Wash buffer (Novagen), centrifuged for 10 min at 10 000 g and inclusion bodies were stored at -80°C. Solubilization of inclusion bodies was performed by gently resuspending pellets in 1X 1B Solubilization buffer from the Protein Refolding Kit (Novagen) with 0.3% N-laurylsarcosine and 0.1% dithiothreithol (DTT) to a final concentration of 5 mg/mL. The suspension was incubated 30 min at room temperature; centrifuged for 30 min at 25 000 g at room temperature and the supernatant was stored at 4°C. Protein fractions were visualized by SDS-PAGE {Sambrook, 1989 #60} and fractions having sufficient purity were pooled and conserved at 4°C for less then 4 months. Refolding of FtsA was carried out by multistep dialysis of solubilized inclusion bodies containing purified FtsA using the Slide-A-Lyser Mini Dialysis (Pierce). The suspension was first dialysed in 75 volumes of 1X Dialysis buffer from the Protein Refolding Kit (Novagen) with 0.1% (v/v) DTT 4 h at 4°C and overnight at 4°C with fresh buffer. The suspension was then dialysed in 100 volumes of 1X Dialysis buffer (Novagen) with 10 mM MgCl2 overnight at 4°C and 5 h at 4°C with fresh buffer. The refolded FtsA protein concentration was measured by the Bradford method (BioRad) and conserved at 4°C for less then 2 months.

FtsZ and FtsA N-terminal Sequencing

Preparation of protein samples was done as follows: after SDS-PAGE {Sambrook, 1989 #60}, 2 μg of both FtsZ and FtsA were transferred on PVDF membranes using a transfer buffer without glycine (10 mM CAPS pH 11 and 10% (v/v) methanol). The membrane was washed twice with water for 5 min and colored with Ponceau Red for 1-2 min. FtsZ and FtsA bands were cut out and abundantly washed with water and N-terminal sequencing was done by Edman degradation at the Biotechnology Research Institute (National Research Concil Canada, Montreal, Québec, Canada).

Biochemical characterization of FtsZ and FtsA

The GTPase and ATPase enzymatic activity of FtsZ and FtsA was measured using a thin layer chromatography (TLC) assay with P32-labelled nucleotides as substrates. The FtsZ GTPase assay was performed using 12 μM of fresh FtsZ in reaction buffer Z (50 mM Bis-Tris propane pH 7.4, 10 mM MgCl2 and 2.5 mM DTT) and 1 μL of GTP32 10 μCi/μL (PerkinElmer, Woodbridge, Ontario, Canada) in a final volume of 20 μL. The FtsA ATPase assay was done using 6.5 μM of fresh FtsA in reaction buffer A (50 mM Tris-HCl pH 7.2, 50 mM potassium acetate, 10 mM MgCl2 and 1 mM DTT) and 1 μL of ATP32 10 μCi/μL (PerkinElmer) in a final volume of 20 μL {Feucht, 2001 #27}. Both mixtures were then incubated 1 h at 37°C and 2 μL of each sample was deposited on a TLC along with negative controls without enzyme. After separation by TLC, hydrolysis of radioactive substrates was measured by autoradiography with a Phosphorimager (Fuji, Stanford, CA, US). An UV cross-link specific nucleotide binding assay was performed with FtsZ and FtsA against the four radioactive nucleotides. Three μg of FtsZ or FtsA were mixed with each of the four P32-labelled nucleotides (PerkinElmer) and the UV cross-link assay was carried out as described previously {Feucht, 2001 #27}. Briefly, the samples were incubated 30 min at 0°C and irradiated 10 min at 254 nm in wells of a microtiter plate on top of a chilled lead brick in iced water. The protein samples were then washed and purified prior to analysis by SDS-PAGE and visualized by autoradiography.

Affinity selection of phage-displayed peptides against FtsZ and FtsA

Phage display screening was carried out with the PH.D.-12 and PH.D.-C7C phage libraries (New England Biolabs, Beverly, MA, USA) containing ~ 2.7 x 109 12-mer and ~ 3.7 x 109 C7C random peptide sequences, respectively. Peptides are fused to the minor coat protein pIII of M13 phage via a flexible linker glycine-glycine-glycine-serine. The target proteins FtsZ and FtsA were immobilized directly on the charged plastic surface of a 96-well microtiter plate. Two 150 µL aliquots of a 100 µg/mL FtsZ and FtsA solutions in 0.1 M NaHCO3 pH 8.6 were incubated in a microtiter plate overnight at 4°C. Protein solutions were then discarded; coated wells were incubated with 200 µL Blocking buffer (0.1 M NaHCO3 pH 8.6, BSA 5 g/L, 0.02% (v/v) NaN3) for 1 h at 4°C and washed 6 times with TBST. Ten µL of each original library containing 4 x 1010 phages were diluted with 100 µL TBST and incubated in the wells for 1 h at room temperature. Non specific binding phages were discarded and wells were washed 10 times with TBST. Bound phages were eluted with 100 µL of 0.2 M glycine-HCl pH 2.2 and 1 mg/mL BSA 5 min at room temperature. Each eluted sample was then collected, neutralized with 15 µL of 1 M Tris-HCl pH 9.1 and 2 µL aliquots were titered as described below. The rest of the phage eluate was amplified by adding it to a 20 mL E. coli ER2537 culture at early-log phase (O.D.600nm ~ 0.3) and incubated at 37°C for 4.5 h under agitation (250 rpm). Each culture was then centrifuged at 14 500 g for 10 min at 4°C. Each supernatant containing amplified phages was transferred to a fresh tube and centrifuged again. The upper 80% of the supernatant (~ 16 mL) was transferred to a fresh tube and 1/6 volume of PEG/NaCl was added for phage precipitation overnight at 4°C. Precipitated phages were centrifuged at 14 500 g for 15 min at 4°C; the supernatant was decanted, re-centrifuged for 2 min and residual supernatant was removed with a pipette. The phage pellet was suspended in 1 mL TBS, transferred to a microtube and centrifuged for 5 min at 4°C to remove residual cells. The supernatant was transferred to a fresh microtube, re-precipitated with 1/6 volume of PEG/NaCl (~ 167 µL), incubated on ice for 60 min prior centrifugation for 10 min at 4°C. The supernatant was discarded, phages were re-centrifuged for 1 min and residual supernatant was removed with a micropipette. The purified phage pellet was suspended in 200 µL TBS and 0.02% NaN3, centrifuged for 1 min and the supernatant containing the amplified eluate was conserved in a fresh tube at 4°C. Ten µL aliquots of each amplified eluate were titered as described below and phage titers were used to calculate phage input for the second round of biopanning. The second round was carried out using each of the first round of amplified eluates as input phage and raising the Tween concentration for the washing steps to 0.5% (v/v). 2 x 1011 phages of each amplified eluate diluted with TBST was incubated in wells coated overnight with 150 µL aliquots of 100 µg/mL FtsZ and FtsA solutions in 0.1 M NaHCO3 pH 8.6 for 1 h at room temperature. Bound phages were eluted with 100 µL of 0.2 M glycine-HCl pH 2.2 and 1 mg/mL BSA for 5 min at room temperature, neutralized with 15 µL 1 M Tris-HCl pH 9.1; 2 µL phage aliquots were titered as described below. The rest of each phage eluate was amplified and purified as described above and 10 µL of amplified phages was then titered. The third round of biopanning was carried out using each of the second round of amplified phages as input phage and 0.5% (v/v) Tween was used in the washing steps. Eight 150 µL aliquots of each 100 µg/mL FtsZ and FtsA solutions in 0.1 M NaHCO3 pH 8.6 were coated in a 96-well microtiter plate overnight at 4°C. Four aliquots of 2 x 1011 phages per library were incubated in wells for 10 min (instead of 1 h) at room temperature. Specifically bounded phages from each library were eluted using 4 different conditions. Phages adsorbed on FtsZ were eluted with 1) glycine as previously described, 2) with 100 µL of 1 mM GTP in TBS for 30 min, 3) with 100 µL of 1 mM 5’-guanylylimidodiphosphate (a non-hydrolysable analogue of GTP) for 30 min and 4) with 100 µL of 1 mM FtsA in TBS for 30 min. Phages adsorbed on FtsA were eluted with 1) glycine as previously described, 2) with 100 µL of 1 mM ATP in TBS for 30 min, 3) with 100 µL of 1 mM 5’-adenylylimi-dodiphosphate (a non-hydrolysable analogue of ATP) for 30 min and 4) with 100 µL of 1 mM FtsZ in TBS for 30 min. Finally, 2 µL of phage from the third round of unamplified eluate was titered as described below and the rest was conserved at 4°C.

Phage tittering

For unamplified eluates, 101 to 104 from binary dilutions in LB broth were used and 108 to 1011 dilutions were used for amplified phage eluates. Ten µL of each dilution was added to 200 µL fresh E. coli ER2537 culture at mid-log phase (OD600 ~ 0.5). The mixture was vortexed and incubated 5 min at room temperature. Infected bacterial cultures were added to a culture tube containing 3 mL of melted Agarose Top maintained at 45°C, vortexed and immediately poured onto pre-warmed LB/IPTG/Xgal plates. Plates were allowed to solidify, inverted, incubated at 37°C overnight and phage plaques counted.

Phage DNA preparation and sequencing

Phage amplification from the third round of biopanning began by inoculating culture tubes containing 10 mL of LB broth with 100 µL E. coli ER2537 from an overnight culture. For each of the two proteins and phage libraries eluted by the 4 elution methods, twelve phage plaques from each of the third round titering plates were randomly used to infect E. coli ER2537. Culture tubes were incubated at 37°C for 4.5 h under agitation at 250 rpm. Phages were centrifuged and phage DNA was then prepared using the Qiaprep spin M13 kit (Qiagen) and sequenced with a -96 gIII sequencing primer (New England Biolabs). Sequencing was done by the Big dye terminator cycle sequencing technique with Ampli Taq DNA polymerase (PerkinElmer) and DNA fragments were separated with an ABI Stretch 373 system (PerkinElmer). DNA sequence analyses were performed on a SGI Origin 2000 computer using programs from the Wisconsin Package Version 10.3 (Accelrys Inc). A FASTA search was carried out with all the deduced amino acid sequences.

Affinity ELISA

The principle of the experiment is described by Carettoni et al .{Carettoni, 2003 #62} and carried out with the following modifications. Briefly, 150 μL of 100 µg/mL his-FtsZ and his-FtsA solutions in TBS were adsorbed overnight at 4°C on Ni-NTA strips (Qiagen) along with 1% BSA control. Strips were washed four times with 200 μL of TBS and blocked with 200 μL of 1% BSA 1 h at room temperature. After a brief wash with TBS, 2 X 109 pfu/well of each amplified phage samples were incubated 1 h at room temperature in triplicates with protein target-coated wells. TBS was added in triplicate to BSA-coated wells for evaluation of non-specific signals. Strips were washed six times with TBS 1% Triton, two times for 10 min and again six times with TBS 1% Triton. One hundred fifty μL of biotine-labeled anti-fd rabbit polyclonal antibodies diluted 1:2 500 in TBS 1% Triton was incubated for 1 h at room temperature in each well. After four 1 min washes with TBS 1% Triton, 150 μL of 1 µg/mL HRP-labeled streptavidine (Roche Diagnostics) in TBS 1% Triton was added to strips for 30 min at room temperature and the excess was removed with four 1 min TBS 1% Triton washes. Strips were developed by incubating 20 min at room temperature with 100 μL of ABTS (Roche Diagnostics) and reading at O.D.405nm. The average was calculated for each experiment done in triplicate and the BSA non-specific signal was subtracted from all values. Specific affinity ratios for phages from the third round of biopanning were determined by dividing the value with the non-specific control phage value.

Choice and synthesis of selected peptides

Consensus peptide sequences were synthesized as they could represent specific inhibitors of FtsZ. Among selected peptides against FtsA, the consensus sequences were synthesized along with the peptide demonstrating higher ratio of specific affinity. Peptides were synthesized on an ABI 433A Peptide Synthesizer using FastMoc chemistry. The activation was carried out with HBTU/DIEA. The N-terminal amino group was protected by Fmoc and side chain functional groups were protected by t-bu (Asp,Glu,Ser, Thr and Tyr), Boc (Lys and Trp), Trt (Asn and Gln) and Pmc (Arg). The peptides were cleaved with TFA/thioanisole/water/EDT (90:5:2.5:2.5) for 2-3 hrs at room temperature, precipitated with ether prior to lyophilization. The peptides were purified on a Vydak 22 x 250 mm C18 reverse-phase HPLC column using a 0.1% TFA/acetonitrile gradient at 10 mL/min.

FtsZ and FtsA inhibitory enzymatic assays

The enzymatic activities of both GTPase FtsZ and ATPase FtsA were measured by detecting products of hydrolysis as described above. The inhibitory capacities of the 12-mer and C7C-mer synthesized peptides were determined by pre-incubating both enzymes in their respective buffers Z or A lacking DTT and P32-labelled nucleotides substrates with various concentrations of peptide buffered solutions for 20 min at room temperature. P32-labelled GTP or ATP was then added and mixtures were immediately incubated for 1 h at 37°C. Then, 2 μL of each sample was separated by TLC along with positive and negative controls. After TLC separation, the percentage of hydrolysis of radioactive substrates was measured by autoradiography with Phosphorimager (Fuji). The 50% inhibitory concentrations (IC50s) of peptides were obtained by plotting the percentage of enzymatic residual activity as a function of increasing peptide concentration. By way of additive controls, inhibitory capacity of non-specific peptide was tested on GTPase FtsZ and ATPase FtsA. A competitive rescue assay was also done with BSA at twice the enzyme concentrations. DTT effect on enzyme inhibition with specific peptides was analysed by testing the inhibition of peptides with or without 2.5 mM DTT and including non-specific peptides as controls.

Purification of P. aeruginosa FtsZ and FtsA

P. aeruginosa ftsZ and ftsA genes were efficiently cloned in pET expression vectors with six histidine at their C-termini as confirmed by DNA inserts sequencing (data not shown). FtsZ protein was overexpressed in the soluble cytoplasmic fraction and FtsA was overexpressed and concentrated in inclusion bodies (Fig. 1 A and B). FtsZ was purified to homogeneity in a single chromatographic step on an affinity nickel column with a yield of 20 mg/L. Overexpressed FtsA was purified by multistep solubilization and the renaturation of purified inclusion bodies with a yield of 2.5 mg/L. The purified proteins with expected molecular weigh were analysed by N-terminal sequencing of the first 15 amino acid residues and confirmed to be FtsZ and FtsA (100% identities in both cases), respectively (data not shown).

Biochemical characterization of FtsZ and FtsA

The GTPase and ATPase enzymatic activity of FtsZ and FtsA was measured using a TLC assay with P32-labelled nucleotides as substrates. FtsZ autoradiography showed that 12 μM of the enzyme hydrolysed 85% of GTP32 in GDP32 in 1 h at 37°C (Fig. 2 A). FtsA autoradiography demonstrated that 6.5 μM of the enzyme hydrolyse 43% of ATP32 distributed in 10% ADP32, 3% AMP32 and 30% Pi32 in 1 h at 37°C (Fig. 2 B). This data confirmed that both enzymes were biologically active and that FtsA was properly refolded. An UV cross-link binding assay was performed to study the FtsZ and FtsA binding properties. The autoradiography showed that FtsZ binds preferentially GTP and that FtsA binds preferentially ATP amongst the four P32-labelled nucleotides used as substrates (Fig. 3). This firmly supports the corresponding enzymatic activity of FtsZ and FtsA.

Affinity selection of phage-displayed peptides against FtsZ and FtsA

Two phage-display libraries expressing randomized 12-mer and C7C-mer peptides were exploited to identify specific binding peptides having affinity with FtsZ or with FtsA. The specificity of the three rounds of biopanning was raised by increasing the stringency of the washes and by decreasing the time of contact between the peptides and the proteins. Elution of tightly bond phages was done by non-specific disruption of binding interactions using glycine and by competitive elution in biopanning using the respective enzyme substrates and their non-hydrolysable analogues. Since FtsZ interacts with FtsA (and vice versa) 7, we decided to use FtsA and FtsZ as competitors in biopanning. We assumed that elution with a known ligand of the target protein such as FtsZ for FtsA would compete with the corresponding bound phage and would yield phage-displaying peptides with probable binding affinities to FtsZ and FtsA active sites. The phage-display technique and the unique biopanning used identified two C7C-mer and one 12-mer consensus peptide sequence against FtsZ (Fig. 4). After the third round of biopanning DNA sequencing of randomly selected phages from each elution group showed a clear consensus peptide sequence for each library against FtsA. We noted the presence of a perfect 12-mer consensus with the FtsZ elution against FtsA (Fig. 5). Among the 12 sequenced phage DNAs from each elution group (Fig. 4 and 5), lost peptide revertants were selected therefore diminishing the number of peptide sequences. To identify known peptides and proteins found in databases, a FASTA search was carried out with the deduced amino acid sequences from each peptide identified in Fig. 5. No peptides or proteins interacting with FtsZ or with FtsA and containing the peptide sequences could be identified.

Affinity ELISA

The relative affinity of chosen phages expressing the selected peptide sequences were tested by ELISA to evaluate the strength of their binding to FtsZ or FtsA. The relative affinity ratios defined as the value of specific/ non-specific phages obtained against FtsZ varied from 2.0 to 4.6 (Table I). For example, a 4.6 relative affinity ratio would indicate that the specific phage had 4.6 times more affinity for the target protein as compared to FtsZ binding of random peptide phages. It can be noted that some phage peptide sequences had higher relative affinity for FtsZ then others and that the three consensus peptides synthesized did not give the highest ratio values. The mean relative affinity ratio of 12-mer peptides was 2.65 ± 0.8 and was non-significantly different from the mean ratio of the C7C-mer peptides of 3.5 ± 0.9. The relative affinity ratio values obtained against FtsA varied from 2.7 to 43.1 (Table II). We noted that the two consensus peptide sequences did not give the highest ratio values and that the 12-mer peptides mean ratio was 24.7 ± 14 and was more significant than the C7C-mer mean ratio of 11.2 ± 6. An interesting fact is that the two best binding peptides with relative affinity ratios of 43.1 and 42.4 have the same G P H conserved motif at their N-termini and both end with a proline. The second and third strongest binding peptide with relative affinity ratios of 42.4 and 32 contain a G M motif at their center and end with a R P motif (Table II). When comparing the FtsZ and FtsA results, we noted that the FtsZ relative affinity ratio values are lower then those of FtsA and present a lower variation range. This difference can be explained by a weak affinity of the non-specific phage for FtsZ that biased the results by diminishing the relative affinity ratio values and lowering the variation range. In fact, the average absorbence at O.D.405nm of the non-specific phage was 0.28 against FtsZ and 0.048 against FtsA.

Choice of synthesized peptides

In order to analyse the inhibitory effect of selected peptides against FtsZ GTPase and FtsA ATPase activities, the peptides were synthesized on an ABI 433A Peptide Synthesizer and HPLC purified. The FtsZ peptide was selected on the basis of consensus sequences even if they did not give the highest affinity ratio values. The first C7C-mer consensus sequence (CSYEKRPMC) was FtsZp1, the second C7C-mer consensus (CLTKSYTSC) was FtsZp2 and the 12-mer consensus (GAVTYSRISGQY) was FtsZp3. In the case of FtsA, the peptide with the highest relative affinity was synthesized along with the C7C-mer and 12-mer consensus sequences that had low and high ratio values, respectively. This was done in an attempt to compare the impact of relative affinity ratios and the frequency of recovery when searching for peptide inhibitors. FtsAp1 was the C7C-mer consensus sequence (CLAPSPSKC), FtsAp2 was the 12-mer consensus (SVSVGMKPSPRP) and FtsAp3 was the (GPHHYWYHLRLP) peptide that had the highest relative affinity ratio value (Table II) .

Inhibition of FtsZ and FtsA enzymatic activities by synthesized peptides

The GTPase enzyme activity of FtsZ and the ATPase activity of FtsA were measured by detecting substrate nucleotide products of hydrolysis. The inhibitory capacities of the synthesized peptides were estimated by calculating the percentage of residual enzyme activities when using various peptide concentrations. This data was obtained by comparing the reaction rate of enzymes in presence of each peptide concentration to the reaction rate of the uninhibited enzyme assay. Both synthesized peptides were able to inhibit GTPase activity of FtsZ. The IC50s obtained from the measurements were 0.45 mM, 1.5 mM and 5 mM for FtsZp1, FtsZp2 and FtsZp3, respectively (Fig. 6). We noted that the two C7C-mer peptides FtsZp1 and FtsZp2 showed significantly lower IC50 values and higher inhibitory capacities then the IC50 values of 12-mer peptides. The BSA competitive rescue assay gave similar IC50 values (data not shown) indicating that the synthesized peptides have a specific binding affinity with FtsZ. DTT had no effect on inhibitory capacity of the 12-mer peptide FtsZp3 (data not shown) but reduced significantly the inhibitory potential of the two C7C peptides FtsZp1 and FtsZp2. In fact, the inhibitory capacity of theses C7C-mer peptides dropped less than 50% with DTT and this suppose that the disulfire bond between the flanked cysteine residues and the consequent loop conformation was important for their inhibitory potential. Analysis of FtsZ GTPase activity using random peptides at high concentration showed that the reaction rate was not reduced by the addition of any of the two peptides (data not shown). This result indicated that GTPase activity of FtsZ was not inhibiting by non-specific C7C or 12-mer peptides.

The purification of FtsZ was efficiently done in a single chromatographic step (Fig. 1 A) giving an appreciable and often superior yield to those previously reported {Lowe, 1998 #46;Romberg, 2001 #43}. The purification of biologically active FtsA was a challenge because this protein is difficult to overexpress and accumulates in inclusion bodies. We developed an adapted multistep solubilization and renaturation protocol that gives purified active FtsA at an acceptable yield (Fig. 1 B). This purification method is faster and easier than those previously described {Feucht, 2001 #27}. The method describes here contain simple purification steps based upon the easy purification of inclusion bodies which contains FtsA at >99% homogeneity. This protocol may be useful for any insoluble protein when taking into consideration the presence of disulfide bonds and the nature of the cofactor or stabilising molecule for the protein expressed in inclusion bodies.

Enzymatic activities of purified biologically active enzymes expressing GTPase for FtsZ and ATPase for FtsA (Fig. 2) were confirmed. Indeed, we corroborate the GTPase activity (Fig. 2 a) as reported previously {Stricker, 2002 #34;Errington, 2003 #51;Scheffers, 2002 #47}. The ATPase activity of FtsA has been demonstrated by another group only{Feucht, 2001 #27} and we are the first to show that FtsA hydrolysed ATP to ADP, AMP and inorganic phosphate (Fig. 2 b). It has previously been shown that FtsZ binds GTP {Diaz, 2001 #65} and that FtsA binds ATP {Feucht, 2001 #27}. Here we clearly demonstrate that FtsZ binds preferentially GTP and that FtsA binds preferentially ATP among the 4 P32-labelled nucleotides used as substrates (Fig. 3). This firmly confirms their respective activities and gives additional information about their enzyme binding specificity.

It is strongly suggested that peptides that adhere to protein would bind preferentially to functional sites, either their substrate binding sites, at their interaction sites with other proteins in the protein-protein interaction context or at allosteric regulatory sites, rather than interacting randomly with the protein surface {Hyde-DeRuyscher, 2000 #64}. Consequently, peptides could exert inhibition of the biological function of the target protein. The powerful technique of phage display has been used by Hyde-DeRuyscher et al. to isolate a series of peptide ligands that bound specifically to a broad range of well-characterized enzymes. The majority of tested peptides were found to be specific inhibitors of enzyme function; kinetic analysis did not reveal a general trend in the peptide inhibitory mechanism {Hyde-DeRuyscher, 2000 #64}. Based on these principles, we used M13 phage libraries to identify peptide sequences with high binding affinities and hence with potential biological interactions with the FtsZ and FtsA activities essential for bacterial survival. Sequencing of phage DNA revealed at least 3 predominant consensus sequences against FtsZ (Fig. 4) and two against FtsA (Fig. 5) with conserved amino acid motifs which could be important for their binding to the target protein. Since FtsZ interacts with FtsA in the divisosome, we noted the presence of a perfect 12-mer consensus against FtsA when eluting with FtsZ (Fig. 5). This firmly suggests that this consensus peptide binds with FtsA at the FtsZ interaction site. This site is not well characterize yet but there is evidence that FtsZ binds FtsA at the C-termini {Yan, 2000 #39;Haney, 2001 #41} and that the FtsZ-FtsA protein interaction is essential for the bacterial cell division process {Yan, 2000 #39}. We noted that for both proteins the frequency of recovery of phages containing each consensus sequence and the nature of amino acids sequences vary in function of elution conditions. This demonstrates that it is useful to employ several different stringent elution conditions during biopanning so as to obtain a repertoire of peptides. The glycine elution leads to recovery of binding phages except for those with very high binding properties. The time of contact between glycine and the phage will not favour selection of these phages {Hoess, 2001 #58}. Competitive elution during the biopanning scheme permitted the recovery of peptides binding to their respective competitive sites. The non-hydrolysable substrate analogues 5’-guanylylimidodiphosphate and 5’-adenylylimidodiphosphate bind more tightly to the enzyme and we supposed that it would allow the recovery of phage with a tighter binding capacity than the hydrolysable substrate which binds lower. We noted that the substrates or non-hydrolysable analogues elution during biopanning did not reveal a characteristically different amino acid consensus sequence then the FtsA or FtsZ elution. This may suggests that the nucleotide binding site of both enzymes are implicated in the FtsZ-FtsA protein-protein interaction. Bio-informatics analysis revealed no relevant peptide homologues or protein motifs in databases. This may be due to the tridimensional structure of peptides which must share tridimensional homology when interacting with the targeted protein. The FtsZ protein interacts with itself during the polymerisation process {Romberg, 2001 #43} and this self-interaction is essential for the enzymatic activity of FtsZ {Scheffers, 2002 #49}. FtsZ interacts also with FtsA and ZipA {Yan, 2000 #39}, MinC {Errington, 2003 #51}, SulA {Cordell, 2003 #63} and MukB {Den Blaauwen, 1999 #45}. The FtsA protein interacts with itself, FtsZ, PBP3, FtsQ and FtsN {Yim, 2000 #37} (Fig. 7). Indeed, the conformation of some peptides obtained could relate from those of mentioned proteins. It would be interesting to use other elution conditions and analyse the peptides selected. A FtsZ elution against FtsZ and a FtsA elution against FtsA could give peptides that inhibit these essential protein-protein interactions {Errington, 2003 #51}.

The affinity ELISA experiments gave useful information about the selected peptides in the perspective of using them as inhibitors. We noted that the FtsZ relative affinity ratio values are lower and present a lower variation range then those of FtsA. We explained this fact by a weak intrinsic affinity of the non-specific phage for FtsZ that biased the results and the intrinsic nature of both proteins. This bias could be avoided if the affinity of the non-specific phage was tested prior to the selection process. Indeed, the results that we obtained confirmed that some peptide sequences obtained had higher relative affinity for FtsZ than those synthesized. The tridimensional structure of FtsZ obtained at 2.8 Angströms resolution indicates a compact and globular structure {Lowe, 1998 #46}; we hypothesised that the C7C-mer peptides would have more affinity than the 12-mer for this protein which contains a sterical hindrance. The results obtained indicate that the mean relative affinity ratio of 12-mer peptides and C7C-peptides are not different when binding to C7C-mer on FtsZ. We did not identify conserved amino acids motifs among the peptides that have the higher relative affinity ratios (Table 1) as we did for FtsA. There is evidence that biopanning gave a collection of more diverse peptides against FtsZ than FtsA. FtsZ may present several distinct protein binding sites; FtsZ interacts with 6 proteins {Romberg, 2001 #43;Yan, 2000 #39;Errington, 2003 #51;Cordell, 2003 #63;Den Blaauwen, 1999 #45} while FtsA interacts with 5 proteins {Yim, 2000 #37} (Fig. 7). The FtsA affinity experiments for peptides gave higher ratios of relative affinity then those of FtsZ. We supposed that the intrinsic nature of both proteins should have a significant effect on the relative affinity ratio results. An interesting point is that the 12-mer peptides mean ratio surpassed significantly the C7C-mer mean ratio. This correlates with our hypothesis that 12-mer peptides would bind more efficiently on FtsA than the C7C-mer because the crystallography of FtsA resolved at 1.9 Angströms shows a scattered and elongated tridimensional structure {van den Ent, 2000 #28}. Indeed, 12-mer peptides should take more physical space than the C7C-mer because cysteines spontaneously form a disulfide bond constraining the peptide in a disulfide loop in the absence of reducing agents. An important fact to note is the conserved amino acids motifs identified amongst the peptides with higher ratios of relative affinity (Table II). These conserved motifs must be important for the binding of 12-mer peptides to FtsA. We noted that the selected consensus peptide sequences did not have the highest ratio values and that the relative affinity values did not correlate with the phage frequency of recovery. This can be explained by several factors that affect phage recovery apart from the inherent binding property such as replication advantage or disadvantage affected by the pIII fusion peptide, protein translocation problems and folding bias, pIII coat stability and retained phage infection efficiency {Carettoni, 2003 #62}. We present evidence that non-hydrolysable substrate analogues used during the elution in biopanning will allow the recovery of better phage ligands than the hydrolysable substrate. However, the affinity ratio results did not confirm this affirmation.

The inhibitory capacities of synthesized peptides on GTPase activity of FtsZ were evaluated by estimating the percentage of residual activity of enzymes when using peptides at various concentrations. The three peptides synthesized were able to inhibit specifically FtsZ when compared to the BSA protein competitive assay and the non-specific peptide inhibitory assay. We noted that the two C7C-mer peptides FtsZp1 and FtsZp2 showed the highest inhibitory capacities compared to the 12-mer peptide FtsZp3. The inhibitory potential of the three peptides does not correlate with their frequency of recovery but it correlates among the C7C-mer peptides. In fact, the best peptide inhibitor FtsZp1 has been isolated 14 times and FtsZp2 has been obtained 9 times. We noted that the peptide IC50 values do not correlate well with the data from the relative affinity assays. Indeed, peptide inhibitory capacity depends essentially on the localization of the interaction and on the binding strength. Therefore weaker binding peptides may have better inhibitory potential than a high affinity peptide because it presumably binds directly in the active site. This observation implies choice of interesting peptides as inhibitors among the selected is still hazardous because neither the frequency of phage recovery nor the relative affinity ratio data well reflect their inhibitory capacity.

This study allowed the identification of three specific inhibitory peptides against the essential cell division protein FtsZ and some promising peptides against FtsA. In perspective, it will be useful to characterize the inhibition mechanism of the inhibitory peptides and their Ki. It will also be necessary to analyse the inhibitory capacity of these peptides on other bacterial GTPases to analyse their specificity as potential antimicrobial agents. The therapeutic use of peptides is limited due to their potential enzymatic degradation, low permeability and unsuitability for oral administration. Ideally, a major goal of medicinal chemistry is to discover novel structures that circumvent the multiple export and metabolism mechanisms that exist to control levels of active peptides in vivo {Bursavich, 2002 #66}. Inhibitory molecules should be able to cross bacterial cell membranes, avoid bacterial efflux pumps, reach its target at a sufficiently high concentration so as to inhibit a vital bacterial function and not be targeted for bacterial modifying or hydrolysing enzymes that could inactivate it. Also, for economic and diagnostic reasons, antimicrobial agents should have a broad spectrum of antibacterial activity. Finally, it should reach inhibitory concentrations at the site of infection in the organism, have little or no toxicity and have minimal side effects in humans {Hughes, 2003 #21}. In order to obtain promising lead compounds that fulfill these objectives, inhibitory peptides will undergo many chemical modifications and their sequences will constitute the core for the synthesis of libraries of peptidomimetic molecules. Compounds of non-peptide nature would emulate the three-dimensional structure of the original peptides but possess much more useful pharmacodynamic properties as leads for new antibacterials, being less subject to degradation and of higher bioavailability {Nefzi, 1998 #52}. Peptides have been shown to be useful as screening tools for the detection of small-molecule inhibitors of enzyme function by a competitive binding assay {Hyde-DeRuyscher, 2000 #64}. The conservation of FtsZ and FtsA {van den Ent, 2001 #50} among almost all bacterial species indicates the potential for development of specific inhibitors. The ZipA protein is present only in gram-negative bacterial species and demonstrate overlapping functions with FtsA {Pichoff, 2002 #32}. Indeed, inhibiting FtsA in gram-negative bacteria may not give a lethal phenotype since ZipA may replace partially FtsA.

2. Cohen, M. L. (2000). Changing patterns of infectious disease. Nature 406, 762-767.

3. Fauci, A. S. (2001). Infectious diseases: considerations for the 21st century. Clin Infect Dis 32, 675-685.

4. Projan, S. J. (2002). New (and not so new) antibacterial targets - from where and when will the novel drugs come? Curr Opin Pharmacol 2, 513-522.

5. Breithaupt, H. (1999). The new antibiotics. Nat Biotechnol 17, 1165-1169.

6. Errington, J., Daniel, R. A. & Scheffers, D. J. (2003). Cytokinesis in bacteria. Microbiol Mol Biol Rev 67, 52-65, table of contents.

7. Dai, K. & Lutkenhaus, J. (1992). The proper ratio of FtsZ to FtsA is required for cell division to occur in Escherichia coli. J Bacteriol 174, 6145-6151.

8. Cordell, S. C., Robinson, E. J. & Lowe, J. (2003). Crystal structure of the SOS cell division inhibitor SulA and in complex with FtsZ. Proc Natl Acad Sci U S A 100, 7889-7894.

9. van den Ent, F., Amos, L. & Lowe, J. (2001). Bacterial ancestry of actin and tubulin. Curr Opin Microbiol 4, 634-638.

10. Lutkenhaus, J. & Addinall, S. G. (1997). Bacterial cell division and the Z ring. Annu Rev Biochem 66, 93-116.

11. Pichoff, S. & Lutkenhaus, J. (2002). Unique and overlapping roles for ZipA and FtsA in septal ring assembly in Escherichia coli . EMBO J 21, 685-693.

12. Begg, K., Nikolaichik, Y., Crossland, N. & Donachie, W. D. (1998). Roles of FtsA and FtsZ in activation of division sites. J Bacteriol 180, 881-884.

13. Nanninga, N. (1998). Morphogenesis of Escherichia coli . Microbiol Mol Biol Rev 62, 110-129.

14. Feucht, A., Lucet, I., Yudkin, M. D. & Errington, J. (2001). Cytological and biochemical characterization of the FtsA cell division protein of Bacillus subtilis . Mol Microbiol 40, 115-125.

15. Yan, K., Pearce, K. H. & Payne, D. J. (2000). A conserved residue at the extreme C-terminus of FtsZ is critical for the FtsA-FtsZ interaction in Staphylococcus aureus . Biochem Biophys Res Commun 270, 387-392.

16. Haney, S. A., Glasfeld, E., Hale, C., Keeney, D., He, Z. & de Boer, P. (2001). Genetic analysis of the Escherichia coli FtsZ.ZipA interaction in the yeast two-hybrid system. Characterization of FtsZ residues essential for the interactions with ZipA and with FtsA. J Biol Chem 276, 11980-11987.

17. Christensen, D. J., Gottlin, E. B., Benson, R. E. & Hamilton, P. T. (2001). Phage display for target-based antibacterial drug discovery. Drug Discov Today 6, 721-727.

18. Sidhu, S. S. (2000). Phage display in pharmaceutical biotechnology. Curr Opin Biotechnol 11, 610-616.

19. Hyde-DeRuyscher, R., Paige, L. A., Christensen, D. J., Hyde-DeRuyscher, N., Lim, A., Fredericks, Z. L. , et al. (2000). Detection of small-molecule enzyme inhibitors with peptides isolated from phage-displayed combinatorial peptide libraries. Chem Biol 7, 17-25.

20. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular cloning : a laboratory manual , 2nd edn. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.

21. Zhulanova, E. & Mikulik, K. (1998). Characterization of ftsZ gene and its protein product from Streptomyces collinus producing kirromycin. Biochem Biophys Res Commun 249, 556-561.

22. Carettoni, D., Gomez-Puertas, P., Yim, L., Mingorance, J., Massidda, O., Vicente, M. , et al. (2003). Phage-display and correlated mutations identify an essential region of subdomain 1C involved in homodimerization of Escherichia coli FtsA. Proteins 50, 192-206.

23. Lowe, J. & Amos, L. A. (1998). Crystal structure of the bacterial cell-division protein FtsZ. Nature 391, 203-206.

24. Romberg, L., Simon, M. & Erickson, H. P. (2001). Polymerization of Ftsz, a bacterial homolog of tubulin. is assembly cooperative? J Biol Chem 276, 11743-11753.

25. Stricker, J., Maddox, P., Salmon, E. D. & Erickson, H. P. (2002). Rapid assembly dynamics of the Escherichia coli FtsZ-ring demonstrated by fluorescence recovery after photobleaching. Proc Natl Acad Sci U S A 99, 3171-3175.

26. Scheffers, D. J., de Wit, J. G., den Blaauwen, T. & Driessen, A. J. (2002). GTP hydrolysis of cell division protein FtsZ: evidence that the active site is formed by the association of monomers. Biochemistry 41, 521-529.

27. Diaz, J. F., Kralicek, A., Mingorance, J., Palacios, J. M., Vicente, M. & Andreu, J. M. (2001). Activation of cell division protein FtsZ. Control of switch loop T3 conformation by the nucleotide gamma-phosphate. J Biol Chem 276, 17307-17315.

28. Hoess, R. H. (2001). Protein design and phage display. Chem Rev 101, 3205-3218.

29. Scheffers, D. J. & Driessen, A. J. (2002). Immediate GTP hydrolysis upon FtsZ polymerization. Mol Microbiol 43, 1517-1521.

30. Den Blaauwen, T., Buddelmeijer, N., Aarsman, M. E., Hameete, C. M. & Nanninga, N. (1999). Timing of FtsZ assembly in Escherichia coli . J Bacteriol 181, 5167-5175.

31. Yim, L., Vandenbussche, G., Mingorance, J., Rueda, S., Casanova, M., Ruysschaert, J. M. , et al. (2000). Role of the carboxy terminus of Escherichia coli FtsA in self-interaction and cell division. J Bacteriol 182, 6366-6373.

32. van den Ent, F. & Lowe, J. (2000). Crystal structure of the cell division protein FtsA from Thermotoga maritima . EMBO J 19, 5300-5307.

33. Bursavich, M. G. & Rich, D. H. (2002). Designing non-peptide peptidomimetics in the 21st century: inhibitors targeting conformational ensembles. J Med Chem 45, 541-558.

34. Hughes, D. (2003). Microbial genetics: Exploiting genomics, genetics and chemistry to combat antibiotic resistance. Nat Rev Genet 4, 432-441.

35. Nefzi, A., Dooley, C., Ostresh, J. M. & Houghten, R. A. (1998). Combinatorial chemistry: from peptides and peptidomimetics to small organic and heterocyclic compounds. Bioorg Med Chem Lett 8, 2273-2278.

Table 1. Determination of relative affinity of selected phages against FtsZ by ELISA. Results are indicated as a ratio of specific/ non-specific phages.

1 Consensus peptide FtsZp1 2 Consensus peptide FtsZp2 3 Consensus peptide FtsZp3

Table 2. Determination of relative affinity of selected phages against FtsA by ELISA. Results are indicated as a ratio of specific/ non-specific phages.

1 Consensus peptide FtsAp1 2 Consensus peptide FtsAp2 3 Peptide that had the highest relative affinity ratio value FtsAp3 X Conserved motifs among the peptides with higher relative affinity ratios

Figure 1. SDS-PAGE showing the overexpression and purification of the P. aeruginosa (A) FtsZ and (B) FtsA proteins in E. coli BL21 (λDE3) cell lysates. Lanes: 1, overexpressed proteins in (A) soluble cytoplasmic fraction and (B) insoluble fraction; 2 , homogeneity purified proteins.

Figure 3. UV cross-link specific nucleotide binding assay showing that (A) FtsZ binds preferentially GTP and that (B) FtsA binds preferentially ATP among the 4 nucleotides.

Figure 4. Consensus peptide sequences of the Ph.D.-12 and Ph.D.-C7C phage libraries against FtsZ obtained by sequencing of phage eluted after the third round of biopanning. Acidic amino acids (D, E) are in green, polar amino acids (Q, N) are in light green, basic amino acids (K, R, H) are in blue, hydrophobic amino acids (I, L, M, V) are in rose, hydrophobic aromatic amino acids (F, Y, W) are in red, small amino acids (A, S, C, T) are in pink, G (tiny amino acid) is in orange, and P (leading to turn formation) is in black (classified according to the Venn diagram for defining the relationships between amino acids).

Figure 5. Consensus peptide sequences of the Ph.D.-12 and Ph.D.-C7C phage libraries against FtsA obtained by sequencing of phage eluted after the third round of biopanning. Amino acids are classified according to Venn diagram as described at Figure 4.

Figure 6. IC50 determinations for the phage display derived peptide inhibitors of FtsZ GTPase. The residual enzymatic activity was measured as a function of the concentration of the C7C peptides (A) FtsZp1, (B) FtsZp2 and (C) the 12-mer FtsZp3.

Figure 7. Schematic representation of the bacterial divisosome showing the FtsZ interactions with itself, FtsA, ZipA, SulA, MinC and MukB along with the FtsA interactions with itself, FtsZ, PBP3, FtsN and FtsQ.

Figure 1

Figure 2

Figure 3

Figure 6

Figure 7