The Pseudomonas aeruginosa Ribbon-Helix-Helix DNA-Binding Protein AlgZ (AmrZ) Controls Twitching Motility and Biogenesis of Type IV Pili

Pseudomonas aeruginosa is an opportunistic pathogen that iscommonly found in water and soil. In order to colonize surfaceswith low water content, P. aeruginosa utilizes a flagellum-independentform of locomotion called twitching motility, which is dependentupon the extension and retraction of type IV pili. This studydemonstrates that AlgZ, previously identified as a DNA-bindingprotein absolutely required for transcription of the alginatebiosynthetic operon, is required for twitching motility. AlgZmay be required for the biogenesis or function of type IV piliin twitching motility. Transmission electron microscopy analysisof an algZ deletion in nonmucoid PAO1 failed to detect surfacepili. To examine expression and localization of PilA (the majorpilin subunit), whole-cell extracts and cell surface pilin preparationswere analyzed by Western blotting. While the PilA levels presentin whole-cell extracts were similar for wild-type P. aeruginosaand P. aeruginosa with the algZ deletion, the amount of PilAon the surface of the cells was drastically reduced in the algZmutant. Analysis of algZ and algD mutants indicates that theDNA-binding activity of AlgZ is essential for the regulationof twitching motility and that this is independent of the roleof AlgZ in alginate expression. These data show that AlgZ DNA-bindingactivity is required for twitching motility independently ofits role in alginate production and that this involves the surfacelocalization of type IV pili. Given this new role in twitchingmotility, we propose that algZ (PA3385) be designated amrZ (alginateand motility regulator Z).

INTRODUCTION

Top

Introduction

Pseudomonas aeruginosa is a formidable opportunistic pathogencausing deadly infections in individuals who are burn patients,undergoing cancer chemotherapy, or have the genetic diseasecystic fibrosis (CF) (27). Initial colonizing strains in theCF lung have a nonmucoid phenotype, but gradually mucoid strainsemerge (16). This mucoid phenotype is due to the overproductionof the exopolysaccharide alginate (14). This trait providesadherence properties and protects bacteria from the immune system,allowing P. aeruginosa to chronically infect the lung. For CFpatients, chronic lung colonization by mucoid strains of P.aeruginosa leads to a potent inflammatory response and tissuedestruction, with the eventual result in the vast majority ofpatients being respiratory failure and death (17).

Most of the biosynthetic genes required for alginate productionare located in a large operon. Activation of algD, the firstgene in this operon, is absolutely required for alginate production,and its expression is tightly controlled (10, 49). A numberof genes have been shown to mediate control of algD transcription,including the alternative sigma factor AlgT (AlgU, ${sigma}$ ^E) that ispart of a family of sigma factors that regulate extracytoplasmicfunctions (13, 28). The mucABCD genes adjacent to algT negativelyregulate the activity of AlgT. Mutations in these genes resultin an active AlgT and subsequent transcription of genes, resultingin mucoidy (3, 15, 29). Regulators of algD that have been shownto be under AlgT control include the response regulator proteinsAlgB and AlgR and the ribbon-helix-helix (RHH) DNA-binding proteinAlgZ (GenBank accession number AF139988 and Pseudomonas aeruginosaGenome Database [http://www.pseudomonas.com/current_annotation.asp]designation PA3385) (2, 49, 50). AlgR and AlgZ regulate algDby directly binding to sequences far upstream of the promoter.Each of these regulatory genes is required, but none appearto be sufficient for expression of the mucoid phenotype (1,49).

In addition to alginate production, another important P. aeruginosavirulence factor is type IV pili (TFP). These polar appendagesare responsible for attachment to epithelia, biofilm development(likely including DNA binding), and twitching motility (TM)(30, 38). TM allows P. aeruginosa to travel across solid surfacesvia the extension and retraction of TFP (4). More than 40 geneshave been identified as involved in TFP function or the regulationof TM. A loss of the TFP complex or the ability to extend andretract TFP results in cells that are avirulent in cytotoxicitywith murine models of infection (9, 30).

Our laboratories have focused on the study of the proposed RHHDNA-binding protein AlgZ (1). Other members of this family,such as Mnt and Arc of Salmonella enterica serovar Typhimuriumphage P22 (8, 39, 40) and NikR (8) of Escherichia coli, arerepressors, while AlgZ was identified as a transcriptional activator(2). Recently, it has been found that AlgZ binds its own promoterand also represses transcription (31). Our search of the literaturehas not identified any other RHH protein that has been identifiedas both an activator and a repressor in vivo. All previous studiesof AlgZ have been undertaken with mucoid P. aeruginosa strains.In the present work, we examined the potential role of AlgZin the control of genes in the nonmucoid strain PAO1. We discoveredthat PAO1-dervied algZ mutants were deficient in TM. The roleof algZ in TM was shown to require its DNA-binding activitybut was independent of its role in algD activation. Transmissionelectron microscopy and antipilin Western blot analysis of surfaceand whole-cell preparations suggest that algZ is required forproper assembly of surface-exposed TFP but not for expressionof the major subunit pilA. Because of its dual role in controllingboth alginate production and TFP-mediated TM, we propose thatalgZ (GenBank accession numbers AF139988 and PA3385) be renamedamrZ (alginate and motility regulator Z).

MATERIALS AND METHODS

Top

Materials and Methods

Bacterial strains, plasmids, and oligonucleotides. All P. aeruginosa strains used in this study originated fromPAO1 and are listed in Table 1. WFPA203 was constructed by insertingan arabinose-inducible amrZ at the neutral attB site with pJF2,a mini-CTX-based plasmid (19), whose construction is describedbelow. Subsequently, pAB3 was used to replace the wild-typeamrZ with an omega tetracycline cassette as described previously(31). Plasmid pJF2 was constructed by cloning the product ofa PCR amplification with template pPJ145, which was previouslydescribed (50), and primers pET1 and pET2 into the EcoRI-SmaIsite of pSW196. This plasmid was then mobilized into PAO1 viaE. coli strain SM10. The plasmid was integrated at the chromosomalattB site via the plasmid-encoded site-specific integrase (19).The cloning site in this vector is flanked by Saccharomycescerevisiae Flp recombinase target sites. These sites facilitatedin vivo excision of the plasmid, including the antibiotic resistancemarker, via the Flp recombinase that was supplied on a differentplasmid (pFLP2) via conjugation. This plasmid was cured viathe plasmid-encoded sacB counterselectable marker (19). ThepPJ145 template used contains the wild-type amrZ and has beenpreviously described (50). pSW196 contains the promoter frompBAD30 (18) and a small multiple cloning site (EcoRI, PstI,SmaI, NotI, and SacI) cloned into mini-CTX (19). Briefly, pSW161,which has been previously described (46), was cut with EcoRIand HindIII, and a corresponding fragment containing algT wasinserted (pSW186); this was then cleaved with EcoRI to removethe small polylinker with a KpnI site (pSW192). The KpnI-HindIIIfragment containing algT with the polylinker and pBAD sequenceswas then cloned into the corresponding sites of mini-CTX (19)(pSW195), and the algT sequences were removed by cleavage withEcoRI and re-ligation (pSW196).

View this table:
[in this window]
[in a new window]
TABLE 1. Strains, plasmids, and primers used in this study

The PAO1 strains with mutations in the putative amrZ DNA-bindingdomain were constructed starting with a PAO1 ${Delta}$ amrZ::tet strain(WFPA205). WFPA205 was constructed using pAB3 to replace thewild-type amrZ with an omega tetracycline cassette as describedpreviously (31). Isogenic strains with the indicated mutationswere constructed by allelic exchange of the omega tetracyclinecassette with the designated amrZ allele. The alleles were constructedusing site-directed mutagenesis (Promega Altered Sites) withthe indicated oligonucleotides. Each mutant oligonucleotidewas used with the plasmid pPJ136 (the 1.8-kb BamHI-HindIII fragmentfrom pDJW585 containing amrZ in pAlter). A chromosomal R22Amutation was introduced as described previously (31). Briefly,pPJ136 and mutagenic oligonucleotide amrZ19 resulted in pPJ150(amrZ R22A). The BamHI-HindIII fragment of this plasmid wascloned into pEX18Ap, resulting in pPJ157 (31). With the mutagenicoligonucleotide amrZ18, pPJ149 (amrZ V20A) was constructed andthe BamHI-HindIII fragment was subcloned into pEX18Ap, resultingin pPJ156. With mutagenic oligonucleotide amrZ17, pPJ148 (amrZK18A) was constructed and the BamHI-HindIII fragment was thencloned into pEX18Ap to produce pPJ155. The wild-type allele(pDJW585) and each construct (pPJ155, pPJ156, and pPJ157) weresubstituted for the ${Delta}$ amrZ:: ${Omega}$ tet allele of WFPA205. This resultedin strains WFPA510 (complemented strain, amrZ ${Delta}$ amrZ⁺), WFPA511(amrZ K18A), WFPA512 (amrZ V20A), and WFPA513 (amrZ R22A), respectively.

Site-directed mutagenesis of pPJ136 was used with mutant oligonucleotideamrZ17 to generate the amrZ17 allele (K18A, pPJ148). In thesame manner, mutant oligonucleotide amrZ18 was used to generatethe amrZ18 allele (V20A, pPJ149). The amrZ19 allele (R22A, pPJ150)was constructed as previously described (31). To generate six-His-taggedalanine mutants of AmrZ, the mutant alleles were cloned intothe BamHI-EcoRI sites of pTrcHisA (Invitrogen) by using primersamrZ26 and amrZ27 and template pPJ148, pPJ149, or pPJ150, resultingin pHL20 (AmrZ K18A), pHL21 (AmrZ V20A), and pHL22 (AmrZ R22A),respectively. Wild-type six-His-tagged AmrZ was constructedas previously described (31).

Cleared extracts derived from E. coli JM109/pDR2 and alaninesubstitution mutants were prepared as previously described (31).In brief, the cleared extract was added to a slurry containingNi-nitrilotriacetic acid magnetic agarose beads (QIAGEN) andbinding buffer (50 mM NaH₂PO₄, 300 mM NaCl, 20 mM imidazole,pH 8.0), and the sample was allowed to mix at 4°C for 1hour. Residual proteins were removed with two wash steps, andthe pure protein was eluted from the beads by the addition ofbinding buffer containing 1 M imidazole and 0.005% Tween. Proteinexpression and purification steps were monitored by sodium dodecylsulfate-polyacrylamide gel electrophoresis and staining withCoomassie brilliant blue (Fisher).

DNA-binding assays and conditions are similar to those previouslyreported (2, 31). In brief, increasing amounts of purified wild-typeN-terminal six-His AmrZ or alanine substitution mutants (AmrZK18A, AmrZ V20A, and AmrZ R22A) were incubated in a DNA-bindingassay with an end-labeled algD fragment amplified from templatepDJW221 by using primers algD5 and algD7 (1, 48). All DNA-bindingassays were performed at 25°C for 10 min in a reaction volumeof 10 µl and processed as previously described (2). Gelswere dried at 80°C for 20 min and exposed to a phosphorimagingscreen for variable amounts of time.

Media and chemicals. All P. aeruginosa strains were cultured in LBNS (10 g liter^–1tryptone and 5 g liter^–1 yeast extract) or on LANS (LBNSwith 15 g liter^–1 agar). E. coli strains were grown inLB (10 g liter^–1 tryptone, 5 g liter^–1 yeast extract,5g liter^–1 sodium chloride) or on LA (LB with 15 g liter^–1agar). Incubations were carried out at 37°C. Antibioticconcentrations (when necessary) for culture of E. coli wereas follows: for tetracycline, 15 µg/ml; for ampicillin,100 µg/ml; and for gentamicin, 15 µg/ml. For P.aeruginosa, antibiotics were used at the following concentrations:tetracycline, 100 µg/ml; carbenicillin, 300 µg/ml;and gentamicin, 100 µg/ml. The medium for culturing ofP. aeruginosa strains after mating was LANS supplemented withIrgasan (25 µg/ml). Sucrose counterselection via sacBwas performed at 30°C with 5% sucrose in LANS. All chemicalswere obtained from Sigma unless otherwise stated.

Twitching motility assays. Twitching motility was assayed by stab inoculating strains througha thin 1% LANS plate containing 0.1% tetrazolium red (with 2%arabinose supplementation where indicated) with 24 to 48 h ofincubation at 37°C under humidified conditions. Twitchingmotility zones were visualized at the agar plate interface (43).

Transmission electron microscopy (TEM). Bacterial strains were inoculated onto LANS plates and placedat 37°C overnight. After 12 h, the plates were flooded with3 ml of LBNS and the colonies were gently suspended. From eachplate, 1.0 ml was collected and placed into microfuge tubes.Copper grids were treated with 100% ethyl alcohol for 1 minute.A fresh suspension of each strain was placed on the correspondinggrid, barely covering the darker side of the grid. After 1 minute,the edge of the grid was wicked with filter paper. The treatedsurface of the grid was washed with cold double-distilled waterthree times. Upon washing, the excess of moisture was wickedfrom the grid. A drop of 2% uracyl acetate was placed on thegrid for 1 minute to negatively stain the sample. The grid wasdried with filter paper from the edge. Visualization up to x25,000was made via a Philips TEM 400 (Micromed, Wake Forest UniversitySchool of Medicine). The images are of a representative bacteriumthat corresponded to 90 to 95% of those observed for each strainanalyzed.

Western blot assays. Whole-cell PilA preparations for Western blotting were preparedby resuspending bacterial cultures from plates in LBNS to anoptical density of 1.0 at 600 nm. Samples (1 ml) were centrifuged,and the pellets were resuspended in 50 µl of 1x loadingbuffer (0.25 M Tris-HCl, pH 6.8, 10% glycerol, 2% sodium dodecylsulfate, 0.0125% bromophenol blue, 5% 2-mercaptoethanol). ChromosomalDNA was sheared by drawing samples through a 27.5-G needle approximately10 to 20 times or until the solution was no longer viscous.The samples were boiled for 4 min, resolved in a 4 to 20% precastgel (Bio-Rad Laboratories) via sodium dodecyl sulfate-polyacrylamidegel electrophoresis, and transferred electrophoretically usinga semidry transfer apparatus (Bio-Rad) to Immobilon-P polyvinylidenedifluoride membranes (Millipore). PilA was detected with anti-PilAantibody (a gift of Randall Irving) used at a concentrationof 1:5,000 in Tris-buffered saline (20 mM Tris base, 137 mMNaCl, pH 7.6) with 0.1% Tween 20, followed by donkey anti-rabbitimmunoglobulin G, horseradish peroxidase-linked antibody (AmershamBiosciences). Chemiluminescent detection reagents (AmershamBiosciences) were used according to the manufacturer's procedure,and CL-XPosure Clear-Blue X-ray film (Pierce Biotechnology)was exposed to the blots and developed. Fimbrial-surface preparationsfor Western blotting were prepared as done previously (45) byscraping whole plates of bacteria into 1.5 ml phosphate-bufferedsaline (8 g liter^–1 NaCl, 1.16 g liter^–1 Na₂HPO₄,0.2 g liter^–1 KH₂PO₄, 0.2 g liter^–1 KCl, pH 7.0).Samples were vortexed to resuspend bacteria and to shear thefimbriae from the cell surface. The cells were centrifuged twice,and the supernatants were transferred to fresh 1.5-ml microfugetubes. MgCl₂ (1 M, 150 µl) was added to precipitate thefimbriae, and the tubes were vortexed briefly and placed at4°C overnight. The TFP were pelleted by centrifugation andresuspended in 100 µl of 1x loading buffer. Samples wereresolved and detected as described above. When included, flagellin-specificantibody was provided at a 1:20,000 dilution. Loading controlgels were stained with Gelcode Blue (Pierce) as instructed bythe manufacturer.

RESULTS

Top

Results

AmrZ (AlgZ) is required for twitching motility, and this role does not involve alginate production. While AmrZ (previously referred to as AlgZ, encoded by algZ[PA3385]) was identified as a DNA-binding protein required forthe production of alginate in mucoid P. aeruginosa isolates(2), it was unclear what role this protein may play in nonmucoidstrains, which are the predominant phenotype found in the environmentand in acute-infection isolates. An amrZ deletion was constructedin the nonmucoid strain PAO1, and this resulted in a domed-colonymorphology lacking a rough "ground glass" edge. This phenotypewas suggestive of a defect in TM (35). Complementation was accomplishedby placing an arabinose-inducible amrZ at the neutral attB site.This resulted in the wild-type spreading, rough-edged colonytype of the parental PAO1 strain upon arabinose induction (datanot shown).

TM assays were performed in order to more precisely analyzethe TM phenotype of these strains (Fig. 1A). This involved stabinoculation of a 1% agar plate and visualization of TFP-mediatedgrowth at the agar plate interface by utilizing tetrazoliumred. In the absence of arabinose, only PAO1 (Fig. 1A) demonstratesa halo of bacterial motility at the agar plate interface. Boththe amrZ deletion strain (WFPA205) and the deletion strain containingan inducible amrZ (WFPA203) are TM deficient. However, in thepresence of arabinose, both PAO1 and WFPA203 demonstrate a TM⁺phenotype (Fig. 1B), as WFPA203 is complemented by the arabinose-inducibleamrZ. These data indicate a previously unrecognized role foramrZ in the control of TM. This was not due to a growth defect,since the amrZ deletion and parental PAO1 strain have similargrowth curves under rich-medium growth conditions (data notshown). Due to its function as a regulator of at least two distinctsurface components in P. aeruginosa, the algZ gene is now giventhe designation amrZ (alginate and motility regulator Z).

View larger version (43K):
[in this window]
[in a new window]
FIG. 1. TM analysis of amrZ and algD strains. A thin 1% agar plate containing LA with no sodium chloride and tetrazolium red without (A) or with (B) 2% arabinose supplementation was stab inoculated to the bottom of the plate and incubated for 24 to 48 h at 37°C. The central growth of the colony on top of the agar is seen in all cases, while TM is shown as a lighter circle surrounding the colony. The following strains were analyzed: strain 1, WFPA205 (amrZ deletion); strain 2, PAO1 (wild type, TM⁺); strain 3, WFPA203 (amrZ deletion with an arabinose-inducible amrZ at attB). (C) TM analysis of algD and amrZ strains performed as described above. The following strains were analyzed: strain 1, WFPA1 (algD deletion); strain 2, WFPA205 (amrZ deletion).

Since amrZ is required both for the production of alginate andfor TM, the simplest explanation would be that alginate maybe required for TM. Perhaps this exopolysaccharide aids in TMby decreasing friction and allowing the bacteria to move moreeasily. In order to address this possibility, an isogenic strain(WFPA1) in which algD was deleted was constructed and the TMphenotypes of the amrZ and algD deletion strains were examined(Fig. 1C). WFPA1 displays a TM⁺ phenotype, compared with WFPA205(amrZ deletion) (Fig. 1C), which is TM^–. This indicatesthat alginate is not required for TM and that the role AmrZplays in TM is independent of algD activation.

TEM reveals that the amrZ deletion strain produces no detectable surface pili. Previously, Bradley found that many TM mutants lacked functionalTFP and were either apiliate or hyperpiliated (4). In orderto determine the phenotype of the amrZ deletion strain withregard to production of pili, the surfaces of the various bacterialstrains were analyzed via TEM. Upon examination of numerousnegatively stained cells, no TFP were visualized on the surfaceof bacteria harboring a deletion in pilA or amrZ, while thewild-type PAO1 strain clearly possessed these appendages (Fig.2). These data suggest that amrZ is required for pilA expression,TFP assembly, or TFP export.

View larger version (52K):
[in this window]
[in a new window]
FIG. 2. TEM analysis of P. aeruginosa strains. Bacteria were harvested from plates cultured overnight, negatively stained using 2% uracyl acetate, and visualized using a Philips TEM 400. Strains analyzed were (A) PAO1 (wild type), (B) AWO (pilA deletion), and (C) WFPA205 (amrZ deletion).

AmrZ is required for proper surface localization of TFP but not for pilA expression. In order to determine whether amrZ plays a role in pilA expression,we prepared whole-cell extracts of PAO1, AWO (pilA deletion),WFPA205 (amrZ deletion), and WFPA203 (amrZ deletion with anarabinose-inducible amrZ at attB) both with and without arabinosesupplementation and analyzed these via Western blotting usingantipilin antibodies (Fig. 3A). Bacteria were harvested fromplates, and the numbers of cells used were normalized by dilutionof the samples to the same optical density at 600 nm. A stainedgel of the samples indicates that similar amounts of total proteinwere in each lane (Fig. 3C). The amounts and masses of PilApresent in the wild-type PAO1 strain and the amrZ deletion strainsare similar, while PilA is not detectable in the pilA deletionstrain AWO (Fig. 3A, compare lanes 1, 3, and 4 to lane 2). Thisindicates that amrZ is not required for pilA expression or PilAmaturation. Since the amount of PilA in the amrZ deletions wassimilar to that of the wild type, we wanted to determine ifthe TM defect was due to PilA export or TFP assembly. To addressthis, we made fimbrial-surface preparations of the above strainsand subjected these to Western blot analysis (Fig. 3B). In orderto confirm that similar amounts of the preparations were loadedinto each lane, antiflagellin antibody was used in additionto anti-PilA antibody in probing the blots of the surface samples.The amount of flagellin detected in each lane is similar, indicatingthat comparable numbers of cells were analyzed (Fig. 3D). Comparedto the wild-type strain, which possesses copious amounts ofsurface PilA, the amrZ deletions show extremely low levels (Fig.3B, compare lane 1 with lanes 3 and 4). When arabinose is suppliedin order to induce amrZ in the WFPA203 strain (amrZ deletionwith an arabinose-inducible amrZ at attB), the result is wild-typeamounts of surface PilA (Fig. 3B, lane 5). This confirms theTEM data and demonstrates that amrZ is required for proper localizationof wild-type amounts of TFP.

View larger version (35K):
[in this window]
[in a new window]
FIG. 3. Analysis of PilA production and localization via Western blotting. Whole-cell extracts (A) with identical samples in a stained-gel control (C) and sheared-surface TFP (B) were visualized via Western blotting with enhanced chemiluminescence (ECL) detection. The primary antipilin antibody was a gift from Randall Irvin and was used at a 1:5,000 dilution. Antiflagellin antibody at 1:20,000 was used as a loading control for the surface samples (D) with detection via ECL reagents. The following strains were analyzed: strain 1, PAO1 (wild type); strain 2, AWO (pilA deletion); strain 3, WFPA205 (amrZ deletion); strain 4, WFPA203 (amrZ deletion with an arabinose-inducible amrZ at attB) without arabinose supplementation; strain 5, WFPA203 with 3% arabinose supplementation.

AmrZ DNA binding is required for TM. Since AmrZ-mediated autoregulation and algD expression requirethe DNA-binding activity of AmrZ (2, 31), we examined if thisis the case with TM. The amino termini of the RHH DNA-bindingproteins Arc and Mnt are responsible for operator recognitionand binding, forming a unique DNA-binding domain (22). The DNA-bindingdomains of the dimeric protein Arc (53 residues, monomer) andthe tetrameric protein Mnt (82 residues, monomer) are similarin structure and assume the shape of an orthogonal bundle (37).The orthogonal bundle is comprised of a beta-strand, which contactsDNA, followed by two alpha helices that form a coiled coil andcomprise a dimerization domain (6). Previous studies have shownthat residues Arg2, His6, Asp8, and Arg10 of Mnt are responsiblefor operator recognition and binding (21). Of the first 26 residuesof the N terminus of AmrZ, residues 13 to 26 share amino acididentity with the beta strands of Arc and Mnt (Fig. 4A) (1).

View larger version (78K):
[in this window]
[in a new window]
FIG. 4. Amino acid residues K18 and R22 are absolutely required for binding of AmrZ to the algD promoter. (A) Amino acid alignment between Arc (residues 1 to 53) and AmrZ (residues 10 to 62). Amino acid residues that correspond to the beta-strand (Arc residues 9 to 13), alpha helix A (Arc residues 16 to 28), or alpha helix B (Arc residues 33 to 47) are indicated in bold. Amino acids conserved between the two proteins are underlined. Numerical references for a portion of the amino acids are listed above their corresponding residues. (B to E) Electrophoretic mobility shift assays with the following amounts of the wild-type (B), K18A (C), V20A (D), or R22A (E) N-terminal domain (NTD) six-His-tagged AmrZ incubated with end-labeled DNA containing the AmrZ-binding site at algD: lane 1, 1.22 pmol; lane 2, 2.44 pmol; lane 3, 3.66 pmol; lane 4, 4.88 pmol; lane 5, 6.10 pmol; lane 6, 12.2 pmol; lane 7, 24.4 pmol; lane 8, 36.6 pmol; lane 9, 48.8 pmol.

We first addressed whether these residues would be requiredfor DNA-binding activity to a published AmrZ-binding target.To test this, we engineered amino acid substitutions in placeof AmrZ residues K18, V20, and R22 by using site-directed mutagenesisand tested purified wild-type and mutant proteins for bindingto an algD DNA fragment containing an AmrZ-binding site (Fig.4B to E).

As the concentration of wild-type six-His AmrZ increases (Fig.4B), the AmrZ-algD complexes were observed and eventually shiftedaway 100% of the free DNA (Fig. 4B, lanes 2 to 10). Distinctcomplex formation was observed at higher concentrations of six-HisAmrZ only (Fig. 4B, compare lanes 5 to 7 with lanes 8 to 10).Similar to wild-type AmrZ, AmrZ V20A bound to algD DNA (Fig.4D), and this mutant protein was capable of shifting 100% ofthe free DNA at high protein concentrations (Fig. 4D, lanes7 to 10). The migration patterns of the complexes formed byAmrZ V20A differ from wild-type AmrZ in that the most distinctcomplex appears to migrate much higher in the gel than thatobserved for the wild type (compare Fig. 4B and D, lanes 2 to7). In contrast to wild-type AmrZ, AmrZ K18A and AmrZ R22A demonstrateda complete loss of DNA-binding activity (Fig. 4C and E, respectively).This suggests that residues K18 and R22 are absolutely requiredfor AmrZ-mediated DNA-binding activity at algD.

To test the effect of these mutations in vivo, we utilized genereplacement strategies to construct isogenic strains expressingthese AmrZ mutant proteins in place of wild-type AmrZ (see Materialand Methods). The amrZ deletion WFPA205 shows no TM (Fig. 5).However, complementation of the amrZ deletion with wild-typeamrZ (WFPA510) restored the TM⁺ phenotype (note that strainWFPA510 demonstrates a twitching zone identical to that of PAO1[Fig. 5]). The strains with mutations in residues required forAmrZ DNA binding and/or activation at algD are shown at thebottom of the plate (Fig. 5). In each situation, there is nodetectable TM, demonstrating that these amino acids are requiredfor TM. Since AmrZ V20A is able to bind the algD DNA in a mannersimilar to that of the wild-type AmrZ protein, this may indicateeither a function for this residue other than DNA binding orthat AmrZ V20A is defective in steps after DNA binding. Transmissionelectron microscopy was used to analyze the surface of the strains.The complemented strain (Fig. 6A) demonstrates wild-type TFP(compare with Fig. 2A). In contrast, bacteria harboring AmrZK18A, AmrZ V20A, or AmrZ R22A do not express detectable surfaceTFP (Fig. 6B to D). This demonstrates that a mutation of anyof these residues results in the apiliate appearance of an amrZnull strain and suggests that the DNA-binding activity of AmrZis essential for TM. Unfortunately, the amount of AmrZ producedby wild-type PAO1 and PAO1-derived strains with the above AmrZvariants are below the level of detection by Western blotting.However, when these constructs were placed in a mucoid strainbackground, AmrZ was readily detectable (P. J. Baynham and D.J. Wozniak, unpublished data).

View larger version (33K):
[in this window]
[in a new window]
FIG. 5. TM analysis of amrZ alleles with a mutation in the DNA-binding domain. Strains analyzed were PAO1 (wild type), WFPA510 (amrZ deletion complemented with WT amrZ), WFPA511 (amrZ K18A), WFPA512 (amrZ V20A), WFPA513 (amrZ R22A), and WFPA205 (amrZ deletion).

View larger version (100K):
[in this window]
[in a new window]
FIG. 6. TEM analysis of P. aeruginosa strains containing DNA-binding-domain mutations. Bacteria were harvested from plates cultured overnight, negatively stained using 2% uracyl acetate, and visualized using a Philips TEM 400. Strains analyzed were all PAO1 derivatives: (A) WFPA510 (complemented strain, amrZ ${Delta}$ amrZ⁺), (B) WFPA511 (amrZ K18A), (C) WFPA512 (amrZ V20A), and (D) WFPA513 (amrZ R22A).

DISCUSSION

Top

Discussion

Previously, we identified AmrZ as a proposed RHH DNA-bindingprotein absolutely required for algD transcription in mucoidP. aeruginosa strain FRD1 (1). Since these strains are mucoiddue to mutations that lead to alginate overproduction, we wantedto examine if amrZ had any role in nonmucoid strains of P. aeruginosa.In these studies, we constructed an amrZ deletion in nonmucoidPAO1 that resulted in a TM^– phenotype. Since an algD deletionremains TM⁺, this provides strong evidence that AmrZ enhancesTM through some mechanism other than algD activation. Furthermore,mutagenesis of three amino acids crucial for AmrZ DNA bindingreveals that the DNA-binding activity is required for its functionin promoting TM. We also examined whether amrZ is involved inpilA expression or PilA maturation. To address this, we performedTEM and Western blot analysis of surface and whole-cell preparations.A deletion in amrZ resulted in no detectable pili on the surfaceof the cells via TEM. The Western blot analysis showed thatPilA present in the whole-cell preparation of the deletion strainswas of the same size and amount as that present in the wildtype. However, there were only trace amounts of surface-localizedPilA in the amrZ mutants. This reveals a defect in TFP surfacelocalization in the amrZ deletion strain. The involvement ofamrZ in TM is confirmed by previous studies in which an elaboratetransposon library was constructed and then screened for mutationsthat result in a loss of TM. Although these studies were notpursued, Jacobs et al. also discovered that a mutation in amrZ(PA3385) exhibited a TM^– phenotype (20).

It is not unusual for regulators of algD to have roles in theregulation of other genes. Since alginate may afford protectionfrom environmental stresses encountered during infection (12,36, 47), having regulators coordinately controlling other genesnecessary for successful infection would be an advantage. Integrationhost factor is required for full algD activation and also functionsin the regulation of enzymes involved in heme biosynthesis,which occurs largely under anaerobic conditions (23, 48). Inaddition to alginate production, AlgR has been shown to regulateTM and cyanide production (7, 43). AmrZ also regulates multiplegenes, functioning as an activator of algD and a repressor ofits own synthesis (1, 2, 31), and data from this study demonstratean additional role for AmrZ in TM.

Both AmrZ and AlgR bind the algD promoter and are required foralgD activation (1, 11), although the exact mechanism of activationis not known. It is not clear whether a precise three-dimensionalstructure is required for RNA polymerase binding or if theremay be specific protein-protein interactions in which one orboth of these regulators participate. Since AlgR and AmrZ bothparticipate in algD activation and regulate TM (43), it is interestingto speculate that they may both be required for expression ofthe gene(s) involved in TFP assembly and localization. Consistentwith this hypothesis is the finding that both amrZ and algRmutants exhibit similar TM phenotypes (45). Recently, it hasbeen found that the TM genes under AlgR control are in the fimTU-pilVWXY1Y2Eoperon (26). A DNA-binding consensus for AmrZ has recently beenproposed: 5'-gGCCAttACCagcc-3 (31). Upon examination of thefimTU-pilVWXY1Y2E operon, there does not appear to be any obviousAmrZ DNA-binding site (P. J. Baynham, unpublished data). AlgRbinding to this promoter has not been demonstrated, and so bothAmrZ and AlgR may regulate this promoter indirectly. An attemptto complement the amrZ deletion with a plasmid (pVDtacPIL, agift of M. J. Schurr) containing the fimTU-pilVWXY1Y2E operonwas unsuccessful, while an algR deletion strain was complementedin tandem, supporting the hypothesis that AmrZ regulates TMin some other manner (data not shown). Since AmrZ has both activatorand repressor functions (1, 31), it is unclear how AmrZ maycontrol TM in P. aeruginosa. Genomics experiments in progressshould elucidate the AmrZ-dependent genes involved in TM andTFP biogenesis.

P. aeruginosa has a number of virulence factors, which includean array of extracellular proteases, toxins, the ability toform biofilms, and type III secretion. Additionally, there area number of regulators that control more than one P. aeruginosavirulence factor. For example, vfr is necessary for exotoxinA and protease production (41). The quorum sensing systems controla plethora of virulence traits, including elastases, alkalineprotease, exotoxin A, and pyocyanin (5, 25, 32, 34). Likewise,a three-component regulatory network controls both biofilm developmentand the type III secretion system (24).

As our knowledge of bacterial gene regulation advances, thecoordinate regulation of virulence factors is elucidated (7,24, 43, 44). An understanding of how this regulation occursis central to determining how to counter these traits. The targetingof important regulators may be an effective strategy after thedetermination of how they function. The examination of the roleof AmrZ in TM is a crucial step in this process.

ACKNOWLEDGMENTS

This work was supported by RUI grant 0113459 (0443186) fromthe National Science Foundation (P.J.B.). D.J.W. was supportedby Public Health Service grants AI-35177 and HL-58334. D.M.R.was supported by American Heart Association predoctoral fellowship0215191U.

We thank Randall Irvin for generously providing the antipilinantibodies. We also appreciate the generosity of Michael Schurrin providing the pVDtacPIL plasmid. Appreciation is expressedto Samuel Woolwine for assistance with his plasmid constructs.We also thank Haiping Lu for technical support.

FOOTNOTES

* Corresponding author. Present address: Department of Biology, St. Edward's University, 3001 South Congress Avenue, Austin, TX 78704. Phone: (512) 233-1675. Fax: (512) 448-8764. E-mail: patricib{at}admin.stedwards.edu.

REFERENCES

Top