SOS involvement in stress-inducible biofilm formation

PLEASE SCROLL DOWN FOR ARTICLE

This article was downloaded by: [Weitao, Tao]On: 5 July 2010Access details: Access Details: [subscription number 923826533]Publisher Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

BiofoulingPublication details, including instructions for authors and subscription information:http://www.informaworld.com/smpp/title~content=t713454511

SOS involvement in stress-inducible biofilm formationH. Gotohab; N. Kasaranenia; N. Devinenia; S. F. Dalloa; T. Weitaoa

a Department of Biology, The University of Texas at San Antonio, San Antonio, TX, USA b QualityDivision, Research Center for Biologicals, The Kitasato Institute, Kitamoto-shi, Saitama, Japan

First published on: 05 July 2010

To cite this Article Gotoh, H. , Kasaraneni, N. , Devineni, N. , Dallo, S. F. and Weitao, T.(2010) 'SOS involvement in stress-inducible biofilm formation', Biofouling, 26: 5, 603 — 611, First published on: 05 July 2010 (iFirst)To link to this Article: DOI: 10.1080/08927014.2010.501895URL: http://dx.doi.org/10.1080/08927014.2010.501895

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial orsystematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply ordistribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae and drug dosesshould be independently verified with primary sources. The publisher shall not be liable for any loss,actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directlyor indirectly in connection with or arising out of the use of this material.

SOS involvement in stress-inducible biofilm formation

H. Gotoha,b, N. Kasaranenia, N. Devinenia, S.F. Dalloa and T. Weitaoa*

aDepartment of Biology, The University of Texas at San Antonio, One UTSA Circle, San Antonio, TX 78249 USA; bQualityDivision, Research Center for Biologicals, The Kitasato Institute, 6-111 Arai, Kitamoto-shi, Saitama 364-0026, Japan

(Received 21 December 2009; final version received 11 June 2010)

Bacterial biofilm formation can be induced by antimicrobial and DNA damage agents. These agents trigger the SOSresponse, in which SOS sensor RecA stimulates auto-cleavage of repressor LexA. These observations lead to ahypothesis of a connection between stress-inducible biofilm formation and the RecA-LexA interplay. To test thishypothesis, three biofilm assays were conducted, viz. the standard 96-well assay, confocal laser scanning microscopy,and the newly developed biofilm-on-paper assay. It was found that biofilm stimulation by the DNA replicationinhibitor hydroxyurea was dependent on RecA and appeared repressed by the non-cleavable LexA of Pseudomonasaeruginosa. Surprisingly, deletion of lexA led to reduction of both normal and stress-inducible biofilm formation,suggesting that the wild-type LexA contributes to biofilm formation. The decreases was not the result of poorgrowth of the mutants. These results suggest SOS involvement in hydroxyurea-inducible biofilm formation. Inaddition, with the paper biofilm assay, it was found that degradation of the biofilm matrix DNA by DNase Iappeared to render the biofilms susceptible to the replication inhibitor. The puzzling questions concerning the rolesof LexA in DNA release in the biofilm context are discussed.

Keywords: Pseudomonas aeruginosa; bacterial biofilm; DNA replication inhibitor; DNA damage; recombination;SOS response

Introduction

Biofilms are multicellular communities of unicellularorganisms enveloped in an extracellular polymericmatrix that can coat abiotic and biotic surfaces innatural environments and hospital settings (Sauer et al.2002; Stoodley et al. 2002). This extracellular matrixcontributes to antimicrobial resistance (Stewart andCosterton 2001), persistent infections (Costerton et al.1999), and combats wound infections (Dallo and Weitao2010a). Specifically, the Pseudomonas aeruginosa biofilmis responsible for chronic infections in patients withcystic fibrosis (Costerton et al. 1999; Donlan andCosterton 2002; Jackson et al. 2003; Ramsey andWozniak 2005) and in immune-compromised hosts(Costerton et al. 1999). Moreover, having attractedconsiderable attention, P. aeruginosa has emerged as aprevalent nosocomial pathogen not only for hospital-acquired (Obritsch et al. 2005) and medical device-related infections (Costerton et al. 1999; Donlan andCosterton 2002) but also for burn (Bielecki et al. 2008)and war-wound infections (Calhoun et al. 2008).

With biofilms causing 65% of human infections indeveloped countries (Hall-Stoodley et al. 2004),infections resulting from biofilms are posing a for-midable challenge to health care. This is becausebiofilms complicate antibacterial treatment. Such a

complication, on the one hand, stems from the concernthat biofilms aggravate the problem of multi-drugresistance (Costerton et al.1978), whilst on the otherhand, some antimicrobial drugs, such as DNA dama-ging agents, stimulate biofilm formation. For instance,DNA replication inhibitors including quinolone anti-biotics and hydroxyurea induce P. aeruginosa biofilmformation (Takahashi et al. 1995; Gotoh et al. 2008;Weitao 2009a). Aminoglycoside antibiotics enhancebiofilm formation by P. aeruginosa and Escherichia coli(Hoffman et al. 2005; Linares et al. 2006). Bile acidsinduce Vibrio cholerae biofilm formation (Hung et al.2006). Hydrogen peroxide stimulates biofilm forma-tion by Mycobacterium avium (Geier et al. 2008). Sincesuch formation of biofilms is induced under antibac-terial stress, it is termed stress-inducible biofilmformation. The biofilm response to these agents notonly challenges effective treatment but also incurs riskof resistance and virulence emergence, especially whenthe regulatory mechanisms underlying the response areunclear. The mechanisms may possibly involve thebacterial SOS response (Walker 1984) as more andmore antimicrobial agents have been found to induceSOS while triggering stress-inducible biofilm forma-tion. SOS is a transcriptional response, in which thelexA and recA genes control at least 40 SOS genes in

*Corresponding author. Email: [email protected] online 1 July 2010

Biofouling

Vol. 26, No. 5, July 2010, 603–611

ISSN 0892-7014 print/ISSN 1029-2454 online

� 2010 Taylor & Francis

DOI: 10.1080/08927014.2010.501895

http://www.informaworld.com

Downloaded By: [Weitao, Tao] At: 18:53 5 July 2010

E. coli and (Fernandez de Henestrosa et al. 2000;Courcelle et al. 2001; Khil and Camerini-Otero 2002)and 15 in P. aeruginosa (Cirz et al. 2006). Themechanisms of the SOS response in P. aeruginosaand E. coli share the following steps. In the absence ofthe single-stranded DNAs or the SOS signals resultingfrom DNA damage, LexA blocks the transcription ofthe SOS genes (Walker 1984). When the SOS signalsare generated during replication inhibition, RecAcoprotease senses the signals and binds to the single-stranded DNAs to assume an active conformation(Sassanfar and Roberts 1990). Activated RecA stimu-lates the auto-catalytic cleavage of LexA (Little 1991).Consequently, LexA repression of the SOS genes isdismissed by this cleavage. Such derepression inducesthe SOS genes, leading to activation the SOS response.These genes are involved in chromosome recombina-tion, replication, repair, and segregation during celldivision (Cox 1998; Sherratt 2003).

The fact that antimicrobial agents trigger both SOSand stress-inducible biofilm formation hints at theirconnection though direct evidence is lacking. DNAreplication inhibitors, especially hydroxyurea, inducebiofilm formation in P. aeruginosa (Takahashi et al.1995; Gotoh et al. 2008; Weitao 2009a) and the SOSresponse in E. coli (Barbe et al. 1987; VanBogelen et al.1987). As the SOS machinery of P. aeruginosa is similarto that of E. coli, it is believed that hydroxyurea inducesthe SOS response in P. aeruginosa (Gotoh et al. 2008),and it is tempting to assume that the response plays arole in biofilm stimulation. Aligned with this premise is astudy showing that during the SOS response, Staphylo-coccus aureus attachment to fibronectin, an initial step ofbiofilm formation, is mediated through the interplay ofRecA and LexA (Bisognano et al. 2004). Besides,hydrogen peroxide induces the SOS response (Imlayand Linn 1987) and biofilm formation in M. avium(Geier et al. 2008). While these data point to a linkbetween the SOS response and stress-inducible biofilmformation, direct experimental evidence is needed. Thisstudy aims to investigate the connection and fill the gap.

Materials and methods

Bacterial strains, media, and chemicals

P. aeruginosa PAO1 was obtained from the Pseudo-monas Genetic Stock Center (strain PAO0001). Non-polar recA mutant and recA complemented P.aeruginosa (Boles et al. 2004) were kind gifts of DrPradeep K. Singh. The lexA::aacC1 strain was kindlyprovided by Dr Mark. D. Sutton. This strain wasconstructed as described previously (Sanders et al.2006). Briefly, the lexA gene was PCR-amplified fromPAO1 genomic DNA, and an antibiotic resistancecassette was cloned into the site within the lexA coding

sequence to generate DlexA::aacC1. The plasmidcarrying the allele was introduced into P. aeruginosa.Homologous recombination occurred so that the wild-type lexA in P. aeruginosa PAO1 was replaced withDlexA::aacC1.

The LexA non-cleavable strain, gratefully re-ceived from Dr Floyd E. Romesberg, was con-structed by replacement of the catalytic serine ofLexA with alanine as described in Cirz et al. (2006)so that the SOS regulon is repressed in the presenceof DNA damage. In particular, the LexA non-cleavable strain was constructed in three steps. Thefirst was to assemble the lexA allelic exchangecassette that carried the DNA sequences homologousto the regions flanking the lexA open reading frameand the antibiotic resistance marker. The second stepwas to clone the cassette and introduce the S125Amutation into the lexA gene with QuikChangesite-directed mutagenesis kit (Stratagene). Thethird step was to introduce the cassette vector intoP. aeruginosa so that the chromosomal lexA genewas replaced with the S125A allelic exchangecassette.

All the strains were grown at 378C in Luria-Bertani(LB, purchased from Fisher Scientific) with 2 mMhydroxyurea (Sigma-Aldrich) (MIC ¼ 3mM). Theexperiments started with overnight cultures derivedfrom the 1-day-old single colonies grown on LB plates;experiments with colonies older than 3 days might notbe reproducible.

Biofilm formation assays in 96-well plates

Biofilm formation was examined in the assays asdescribed previously (Gotoh et al. 2008). Briefly, theovernight cultures were diluted 50 folds with LB,and the diluted cultures were grown with shaking for3 h in test tubes. Then the cultures were diluted toan OD600 nm of 0.02 with LB and grown withoutshaking with and without hydroxyurea for 24 h in96-well plates. Each of the strains treated withhydroxyurea and the untreated controls was testedin a set of eight wells, and the OD values wereaveraged.

Confocal laser scanning microscopy and imageacquisition

Biofilms were established on a microscopiccoverslip as described previously (Gotoh et al.2008). Briefly, a 10-ml volume culture was placedinto a 50-ml tube, and glass coverslips (24mm 630mm) were placed vertically inside the tube. Thecoverslips were incubated without shaking with andwithout 2 mM hydroxyurea for 24 h. After three

604 H. Gotoh et al.

Downloaded By: [Weitao, Tao] At: 18:53 5 July 2010

washes with saline (0.9% NaCl) the biofilmedcoverslip was covered with a 100-ml volume of10 mg ml71 40, 6-diamidino-phenylindole (DAPI,from Sigma-Aldrich, 1 mg ml71 stock in 30%glycerol, v/v, stored at 7208C). All microscopicobservations and the image acquisitions were per-formed with a Zeiss LSM 510 CLSM (Carl Zeiss,Jena, Germany) equipped with detectors and filtersets (385–470 nm wavelength) for monitoring ofDAPI fluorescence. Images were obtained using206 /1.3 objective. Simulated three-dimensionalimages and sections were generated by using theIMARIS software package (Bitplane AG, Zurich,Switzerland).

Illustrated in Figure 4 are representative images ofthree independent experiments. In each experiment,multiple fields representing the entire biofilm wereinspected as shown in the cartoon (Figure 4b). Thepeak views of the biofilms are presented in Figure 4c–j.Two to three fields at the biofilm peak view wererecorded as indicated in Figure 4b. Statisticallysignificant differences in biofilm thickness betweentest and control (n ¼ 3) were determined by the one-tailed Student’s t-test. The experiment was repeatedthree times to ensure reproducibility.

Paper biofilm assay

A 5-ml volume of the cultures prepared as describedabove was transferred to a glass bottle containing aWhatman CHR paper (3 cm 6 2.5 cm) that stood halfin the culture (Figure 5a). The bottle was kept at 378Cwithout shaking for 18 h for biofilm formation. Then theplanktonic cells were aspirated, the biofilms on the paperwere washed with cold saline covering the whole paper,followed by aspiration, and the wash was repeated threetimes on ice. After 5-ml of cold saline were added, thebiofilmed paper was vortexed vigorously to an homo-genous slurry with the aid of forceps to tear up the paper.Finally, the slurry was loaded onto a 3-ml syringe. Clotsof the paper debris formed to block the paper fibers butto allow cells to flow through, thus separating the paperfibers from the cells. The cell mass was measured foroptical density at a wavelength of 600-nm.

DNase I susceptibility assay

After the planktonic cells were removed, the 24-hbiofilms on the paper were washed with pre-warmed

Figure 1. (a) HU-induced biofilm formation by the wild-type PAO1 strain (WT, white), the recA- strain (gray), andthe recA-complemented strain (black) cultured in 96-wellpolystyrene microplates. (b) Biofilm re-formation abilities ofthe HU-induced and the non-induced wild-type biofilm cellsin the absence of HU. Error bars ¼ SDs (n ¼ 8).* ¼ statistical significance. Illustrated are representativedata from three or four independent experiments. Eachtreatment was tested and measured in eight wells, and themeans and SDs were calculated.

Figure 2. The HU-induced biofilm formation by the wild-type (wt), the recA, the LexA non-cleavable (LexA-N), andthe lexA strains on 96-well microplates. The Figure showsrepresentative data from three or four independentexperiments. Error bars ¼ SDs (n ¼ 8). * ¼ statisticalsignificance. Means and SDs for each treatment werecalculated from eight wells.

Figure 3. Growth of the SOS mutants is not impaired byHU. The wild-type, the recA, the LexA non-cleavable (LexA-N), and the lexA strains were grown in flasks containing LBwith shaking at 378C. Open symbols and dashed lines, ¼ noHU treatment; filled symbols and solid lines ¼ 2 mM HUtreatment. The growth data were based on three independentexperiments.

Biofouling 605

Downloaded By: [Weitao, Tao] At: 18:53 5 July 2010

saline three times. The biofilmed paper was incubatedvertically and without shaking with a 1-ml volume ofDNase 1 (Invitrogen) at 10 mg ml71 in 10 mM Tris-ClpH 7.6, with 2.5 mM MgCl2, and 0.5 mM CaCl2 for0.5 h at 378C, with regular rotating of the tubehorizontally to insure full contact of the paper with theenzyme solution. Following addition of 4-ml LBcontaining hydroxyurea pre-warmed to 378C, the stillincubation continued for a further 2 h. Then the bottlewas placed on ice, and the procedures of washing, cellextraction, filtration and OD measurement were fol-lowed as above.

Statistical analyses

Statistically significant differences (p 5 0.05) betweentest and control were determined by the one-tailedStudent’s t-test as described by Gotoh et al. (2008).Means and standard deviations (SDs) were calculatedeither from the data of the representative experimentsor the data from three or four independent experi-ments, as indicated. They are expressed as mean +SDs. All the analyses were conducted using thestatistic functions provided by Microsoft Office Excel2003 software.

Results

Biofilm repression by preclusion of SOS

Replication inhibitors such as hydorxyurea activate theSOS response (Barbe et al. 1987; VanBogelen et al.1987) and stimulate stress-inducible biofilm formationby P. aeruginosa (Gotoh et al. 2008). These findings ledto a hypothesis that such stress-inducible biofilmformation is regulated by the interplay of the SOSregulators RecA and LexA. To test this hypothesis, astudy was first made to determine whether recA-inactivation to preclude the SOS response would leadto repression of biofilm stimulation. The recA mutantwith the SOS sensor inactivated was used and biofilmformation was quantified by the 96-well assay asdescribed previously (Gotoh et al. 2008).While biofilmformation is hydroxyurea (HU)-stimulated for thewild-type as reported (Gotoh et al. 2008), the biofilmmass in the presence of HU was reduced by more thanfour-fold in the recA strain compared with that of thewild-type (p 5 0.001), and this defect was comple-mented by chromosomally inserted recA (p 5 0.01;Figure 1a). Thus, HU-inducible biofilm formationappeared to be dependent on the SOS sensor RecA, aresult that suggested SOS involvement.

As P. aeruginosa undergoes genetic diversificationduring biofilm formation via a recA-dependent me-chanism and the self-generated diversity increases theability of biofilms to resist environmental stress (Boleset al. 2004), it was tempting to propose that biofilm

Figure 4. Biofilm formation by the SOS mutants asexamined by CLSM. Illustrated are the representativeimages of three independent experiments. (a) HU-inducedwild-type biofilms formed on glass coverslips for 24 h forCLSM. (b) A cartoon with arrows illustrating multiple fieldsof the biofilm for CLSM examination. (c, e, g, i) No HU. (d,f, h, j) 2 mM HU. (c, d) The wild-type. (e, f) recA. (g, h)LexA non-cleavable. (i, j) lexA. 20 6 /1.3 objective. Scalebar ¼ 100 mm.

Figure 5. Biofilm formation by the SOS mutants asexamined by the paper biofilm assay. The bacterial cultureswere incubated at 378C with a standing Whatman paper in aglass bottle for 18 h. The bacterial cells formed biofilms onthe paper and biofilm cell clumps at the air-mediuminterphase while the rest of cells remained planktonic in theliquid medium, (a) in the presence or in the absence of 2 mMHU. Arrow ¼ biofilm clumps. (b) The wild-type biofilmclumps (arrow heads) after washes. (c) DIC microscopy ofthe wild-type biofilm cells extracted from the paper biofilmsand planktonic cells from the shaking culture. Scalebar ¼ 5 mm. (d) Quantification of the biofilm cell massextracted from the paper biofilms treated with or withoutHU. Error bars ¼ SDs (n ¼ 5). * ¼ statistical significance.Means and SDs were calculated, based on the data from fiveindependent experiments.

606 H. Gotoh et al.

Downloaded By: [Weitao, Tao] At: 18:53 5 July 2010

induction by HU was due to generation of rapid-biofilm-forming variants during 8 or 24 h of growth.However, it seemed unlikely because recA-dependentproduction of variants in planktonic culture was notobserved during first 3 days after incubation (Boleset al. 2004). To confirm this, the ability of the cellsfrom the original biofilms induced by HU to re-formbiofilm in the absence of HU as described previously(Hoffman et al. 2005) was examined. The biofilm re-formation abilities of the HU-induced and the non-induced biofilm cells were equivalent (p 4 0.1) in theabsence of HU (Figure 1b). Hence, the HU-inducedbiofilm formation appeared to be not merely attributedto genetic change(s) generated during the length oftime used in this study.

Biofilm repression by non-inducibility of SOS

Since RecA is known to facilitate auto-cleavage ofLexA during the SOS response, the biofilm repressionin the recA strain may result from the action of theuncleaved LexA. Thus, the non-cleavable LexAmutant in which SOS is non-inducible may manifestthe same repression. To test this possibility, biofilmformation by the LexA non-cleavable strain wasexamined in three ways (see Materials and methods).First, the standard 96-well assay was conducted asdescribed previously (Gotoh et al. 2008). In theabsence of HU, while the recA biofilm mass displayedno significant changes compared to the wild-type(p � 0.4), the biofilm mass of the LexA non-cleavablestrain exhibited a substantial reduction (p 5 0.01;Figure 2). Since the catalytic serine of LexA wasreplaced with alanine in the LexA non-cleavable strain,this observation suggested that the mutant LexAprotein with such a replacement was involved in thedecrease in the biofilm mass. In the presence of HU,the HU-stimulation of the wild-type biofilms wasconsistently observed; however, the biofilm mass ofthe LexA non-cleavable strain and the recA straindecreased significantly compared with that of the wild-type (p 5 0.001; Figure 2). Hence, the HU-induciblebiofilm formation appeared repressed by non-cleavableLexA.

It was tempting to argue that the biofilm repressionmight be due to the reduced growth rate of the LexAnon-cleavable strain grown with hydroxyurea. To testthis hypothesis, the growth of the strains was examinedin the presence and in the absence of HU (Figure 3). Inthe absence of HU, the LexA non-cleavable strain andthe other mutant strains did not manifest impairedgrowth compared to the wild-type strain. While thewild-type strain exhibited a reduced growth rate inthe presence of HU, the LexA non-cleavable strain andthe other mutant strains did not display a reduction,

compared to the wild-type strain. Clearly, biofilmrepression in the recA and the LexA non-cleavablestrains was unlikely to result from a decrease in thegrowth rate.

Secondly, CLSM of the biofilms was performed ona glass surface (Figure 4a). The biofilms wereestablished on a microscope coverslip as describedpreviously (Gotoh et al. 2008), and multiple fields asindicated in Figure 4b representing the entire biofilmtopologies were inspected. The peak views of thebiofilms are shown in Figure 4c–j. The confluent celllayers of the biofilms were observed in the culturestreated without or with HU (Figure 4c and d). In thepresence of HU, the wild-type biofilms appearedthicker (45 mm + 3.4 SD), compared to those in theabsence of HU(31 mm + 2.3 SD; p 5 0.05). Theresult was consistent with the previous finding thatbiofilm formation was HU-stimulated (Gotoh et al.2008). When the recA-biofilms treated and untreatedwith HU were compared, the untreated biofilmsappeared slightly disrupted (Figure 4e). Under HUtreatment, when the recA-biofilm was compared withthe wild-type biofilm, the recA-biofilm appearedscattered and disrupted and the thickness was reducedsignificantly (11 mm + 1.2 SD; p 5 0.01; Figure 4f).These results were consistent with those of the 96-wellassay showing RecA-dependent stress-inducible bio-film formation. Since RecA facilitated LexA auto-cleavage during the SOS response, these observationssuggested the LexA-involvement in biofilm repression.This involvement was confirmed by the resultsindicating that the biofilm layers of the LexA non-cleavable strain appeared similar to those of the recA,showing the scattered, disrupted and thin morphology(12 mm + 1.4 SD; Figure 4g and h), as compared tothose of the wild-type (Figure 4c and d; p 5 0.01).Thus, the non-cleavable LexA appeared to repressbiofilm induction.

Thirdly, a paper biofilm assay was performed.While the wild-type and the mutant strains growing onboth the polystyrene and glass surfaces exhibiteddifferential biofilm masses, this further test was carriedout to determine whether these strains displayed thesame biofilm phenotypes on the filter paper. The paperbiofilm assay could reveal additional biofilm propertiesthat the typical biofilm assays could not. For example,biofilm cell clumps became apparent when the wild-type culture was treated with HU (Figure 5a and b).Specifically, in the paper biofilm assay, a filter paperwas placed vertically into the biofilm culture, so thatbiofilm cell clumps formed and attached to the paperbecause of porosity and particle retention nature of thefilter paper. After washes to remove planktonic cells,the clumps were seen attached to the paper (Figure 5b).The biofilm cells on the paper were extracted and

Biofouling 607

Downloaded By: [Weitao, Tao] At: 18:53 5 July 2010

examined by microscopy. As shown in Figure 5c, themorphology of the cells extracted from biofilms onpaper appeared intact, compared to those in plank-tonic culture. However, no other differences wereobserved between the morphology of the cells frombiofilms on paper and those on the glass surface (datanot shown). Thus, cell extraction from paper biofilms,though it inevitably destroyed the biofilm structure,did not affect the cell morphology. When the mass ofthe cells extracted from the paper biofilms wasquantified by measuring OD as shown in Figure 5dthe total yield was approximately 2 6 108 cells perpaper. Consistent with the results from the 96-wellassay, the biofilm mass of the wild-type increased inthe presence of HU (p 5 0.05), and such stimulationappeared repressed in both the recA and the LexAnon-cleavable strains (p 5 0.05).

Role of lexA in biofilm formation and stimulation

Since non-cleavable LexA appeared to repressbiofilm induction, it seemed plausible that deletionof the lexA gene would lead to loss of biofilmrepression. To test this possibility, lexA biofilmformation was examined in the presence and in theabsence of HU. While they appeared to disapprovethe hypothesis, the findings suggested new roles oflexA in biofilm formation. In the absence of HU, the96-well assay indicated that the lexA biofilmdecreased compared with the wild-type (Figure 2;p 5 0.01). CLSM showed the thin biofilm morphol-ogy of the lexA mutant (Figure 4i) with a thicknessof 20 mm + 1.6 SD against 31 mm + 2.3 SD(p 5 0.05) for the wild-type. These observationssuggested that lexA played a positive role in biofilmformation. In the presence of HU, the lexA biofilmmass was reduced as determined by the 96-well assay(Figure 2; p 5 0.001) and by the paper assay (Figure5d; p 5 0.01). With CLSM, the lexA biofilmsappeared scattered and disrupted (Figure 4J), andthe lexA-biofilm thickness (16 mm + 1.1 SD) ap-peared decreased compared with the wild-type(45 mm + 3.4 SD; p 5 0.05). Hence, deletion oflexA seemed to result in a reduction of both normaland stress-inducible biofilm formation rather than ina loss of biofilm repression. The decreases did notresult from poor growth because the lexA strain didnot manifest the reduced growth rate compared tothe wild-type (Figure 3). Therefore, LexA seemed tocontribute to biofilm formation while the non-cleavable LexA appeared to repress it. The seemingparadox of LexA suggests that delicate control ofbiofilm-related gene expression by the cleavable andnon-cleavable LexA contributes to biofilm develop-ment as discussed below.

Cleavage of extracellular DNA contributing to biofilmsusceptibility to HU

Given the findings that extracellular DNA is involvedin biofilm formation (Murakawa 1973; Whitchurchet al. 2002; Nemoto et al. 2003) and that biofilmproduction is induced by DNA replication stress(Gotoh et al. 2008), it was decided to test whetherbiofilm formation becomes susceptible to HU ifextracellular DNA is cleaved. The 24-h biofilms weretreated with DNase I that was known to cleaveextracellular DNA in the biofilm matrix (Whitchurchet al. 2002). Treatment of the biofilms forming onpolystyrene and glass surfaces had been previouslytested by the authors, but the biofilm yields were toolow to be detected. To increase the yields, biofilms wereestablished on filter papers and validated using theassay described above. Then the biofilms on the paperwere treated with DNase 1 as described in theMaterials and methods. The biofilms were thenexposed to HU for 2 h. While it was reduced slightlyby DNase I cleavage alone (p 4 0.1), the biofilm masswas decreased by a combination of the enzyme and thesubsequent HU treatment (p 5 0.05; Figure 6). Hence,degradation of the biofilm matrix DNA by DNase Iappeared to render the biofilms susceptible to thereplication inhibitor.

Discussion

Whether stress-inducible biofilm formation is regulatedby the interplay of the SOS regulators RecA and LexAof P. aeruginiosa was tested. The recA mutant, inwhich LexA auto-cleavage is prevented, exhibitedreduced biofilm formation in response to HU

Figure 6. Biofilm susceptibility to HU by DNase I cleavage.Biofilms of the wild-type strain established on Whatmanpaper were treated with DNase I for 0.5 h followed by HUfor 2 h as indicated. After the planktonic cells were removedand washed off, the biofilm cells on the paper were extracted,and the mass was measured as indicated. Error bars ¼ SDs(n ¼ 5). * ¼ statistical significance. Means and SDs werebased on the data from five independent experiments.

608 H. Gotoh et al.

Downloaded By: [Weitao, Tao] At: 18:53 5 July 2010

treatment. The observations suggested repression ofbiofilm stimulation by non-cleavable LexA. Thispremise received support from the LexA non-cleavablemutant that manifested the repression. However, thelexA-knock out mutant displayed impaired biofilmformation, the result suggesting the positive role ofLexA in biofilm formation. Thus, the cleavable LexAand the non-cleavable LexA seemed to influencebiofilm formation in different ways. The mechanismbehind the difference remains an interesting questionfor further investigation.

The mechanism may reasonably entail repressionof the SOS-regulated genes that are believed to beexpressed in biofilms. In the genome-wide transcrip-tion assay, these genes of P. aeruginosa have beenidentified (Cirz et al. 2006). Some do show increases inthe biofilm-related gene expression profiles. The SOSgenes encoding RecA, UvrD, UvrC, and RecX areinduced in M. smegmatis biofilms (Anil and Hatfull2007). Additionally, the SOS genes including dinI,dinP, dinG, sbmC, recN, sulA are overexpressed in E.coli biofilms, and so the recA mutant with SOSprecluded manifestation of defective biofilms (Beloinet al. 2004). These SOS genes apparently play roles inbiofilm formation, hence, repression of those genes byLexA is expected to hinder biofilm induction. Thepresent data support this premise and stimulate furtherstudies into the mechanisms of SOS-regulated biofilmformation.

It is not clearly understood yet why deletion oflexA leads to biofilm reduction. As biofilm-relatedgene expression seems delicately regulated and pro-grammed during biofilm development (Whiteley et al.2001; Schembri et al. 2003; Beloin et al. 2004;Southey-Pillig et al. 2005), it can be proposed thatknocking out the transcription repressor LexA mayderepress expression of these genes, consequentlydisrupting the programmatic gene expression forbiofilm formation. Therefore, biofilm-related geneexpression seems to be controlled subtly by thecleavable and non-cleavable LexA for biofilmdevelopment.

The paper biofilm assay was developed to addressthe challenge of the low yield of the 96-well assay sothat interventions in biofilm formation using variousdrugs could be readily studied on a reasonable scale.This assay was validated by comparison with thestandard 96-well assay. However, when the estab-lished biofilms were treated with HU for 2 h, thepaper biofilm assay showed no biofilm stimulation(Figure 6), seemingly contradicting the results shownin Figure 5d. They did not. Biofilm formation isknown to be a dynamic process, starting from cellattachment to the surface, matrix production and celldispersion. It had previously been found that HU

stimulates cell attachment at the initial stage thoughit causes DNA damage to the cells (Gotoh et al.2008). When biofilms are established, however, thesurface attachment may not occur as frequently atthe established stage as at the initial stage. Under-standably, treatment of established biofilms with HUthat stimulates early surface attachment is unlikely toincrease their biofilm mass since the establishedbiofilms have passed the initial stage of surfaceattachment. The established biofilms may protect thecells from DNA damage by HU; such protectiveproperties have been reported previously (Elasri andMiller 1999). Hence, the results shown in Figure 6do not contradict those in Figure 5d, because Figure5d addresses the de novo HU-stimulation of biofilmformation, while Figure 6 addresses biofilm protec-tion against HU.

The paper biofilm assay was used to examine theeffect of cleavage of the matrix DNA by DNase I onbiofilm susceptibility to HU. The observation ofbiofilm reduction by exposure to both DNase I andHU is aligned with the well established findings thatextracellular DNA holds cells together in biofilms(Murakawa 1973; Whitchurch et al. 2002; Nemotoet al. 2003; Allesen-Holm et al. 2006). As extra-cellular DNA released from P. aeruginosa has beenmapped in the biofilm matrix (Allesen-Holm et al.2006) and the matrix is protective against DNA-damage agents (Elasri and Miller 1999), cleavage ofthe DNA may reasonably render the biofilmssusceptible to DNA damage agents. As HU stimu-lates biofilm formation and the stimulation appearsto involve SOS, DNA release may be SOS-con-trolled, a hypothesis currently under investigation bythe authors.

Collectively, these data, show for the first time thatthe SOS regulators are involved in regulation ofbiofilm stimulation by a DNA damage agent. Thediscovery of the SOS-biofilm link has a profoundimpact in multiple areas. Antimicrobial agents triggerstress-inducible biofilm formation and the SOS re-sponse, and the response leads to mutagenesis andgenetic instability. When SOS-induced mutagenesisoccurs in biofilms, the stimulated biofilms may becomea haven for molecular evolution where genes forresistance and virulence are frequently mutated sothat new traits may emerge. These evolutionary eventsmay be responsible for the emergence and re-emer-gence of antimicrobial resistance and virulence on theone hand, whilst on the other, the SOS-biofilm link hasadvanced a novel idea of evolution of bacterialanticancer phenotypes under SOS (Dallo and Weitao2010b) and will open a new avenue for the develop-ment of a biofilm-anticancer strategy (Weitao 2009b;Dallo and Weitao 2010b).

Biofouling 609

Downloaded By: [Weitao, Tao] At: 18:53 5 July 2010

Acknowledgements

The authors thank Dr Pradeep K. Singh for the recA strains,Dr Mark D. Sutton for the lexA strains, and Dr Floyd E.Romesberg for the LexA non-cleavable strain. They also thankDr Colleen Witt and Ms Yan Wu for CLSM at UTSA RCMI,Drs James Chambers and Luke Duam for assisting with settingup the laboratory, Drs Jilani Chaudry and Richard G.LeBaron and Ms M. M. Navarro for kind technical support,Reshma Maredia and Ruby Gonzalez for assistance with thepaper biofilm assays, and Ms Maria Macias and Aijie Liu forlaboratory maintenance. Finally, they are grateful to the USAir Force at Brooks City Base in San Antonio for kind supportof facilities. This work was supported by the start-up fundprovided by the University of Texas at San Antonio, anEntrepreneur Initiative Grant from Brooks Air Force Base inSan Antonio, and the San Antonio Area Foundation.

References

Allesen-Holm M, Barken KB, Yang L, Klausen M, Webb JS,Kjelleberg S, Molin S, Givskov M, Tolker-Nielsen T.2006. A characterization of DNA release in Pseudomonasaeruginosa cultures and biofilms. Mol Microbiol59:1114–1128.

Anil O, Hatfull GF. 2007. The role of iron in Mycobacteriumsmegmatis biofilm formation: the exochelin siderophoreis essential in limiting iron conditions for biofilmformation but not for planktonic growth. Mol Microbiol66:468–483.

Barbe J, Villaverde A, Guerrero R. 1987. Induction of theSOS response by hydroxyurea in Escherichia coli K12.Mutat Res 192:105–108.

Beloin C, Valle J, Latour-Lambert P, Faure P, KzreminskiM, Balestrino D, Haagensen JAJ, Molin S, Prensier G,Arbeille B, et al. 2004. Global impact of mature biofilmlifestyle on Escherichia coli K-12 gene expression. MolMicrobiol 51:659–674.

Bielecki P, Glik J, Kawecki M, Martins dos Santos Vt. 2008.Towards understanding Pseudomonas aeruginosa burnwound infections by profiling gene expression. Biotech-nol Lett 30:777–790.

Bisognano C, Kelley WL, Estoppey T, Francois P, SchrenzelJ, Li D, Lew DP, Hooper DC, Cheung AL, Vaudaux P.2004. A RecA-LexA-dependent pathway mediates cipro-floxacin-induced fibronectin binding in Staphylococcusaureus. J Biol Chem 279:9064–9071.

Boles BR, Thoendel M, Singh PK. 2004. From the cover: self-generated diversity produces ‘‘insurance effects’’ in biofilmcommunities. Proc Natl Acad Sci 101:16630–16635.

Calhoun J, Murray C, Manring M. 2008. Multidrug-resistantorganisms in military wounds from Iraq and Afghani-stan. Clin Orthop Relat Res 466:1356–1362.

Cirz RT, O’Neill BM, Hammond JA, Head SR, RomesbergFE. 2006. Defining the Pseudomonas aeruginosa SOSresponse and its role in the global response to theantibiotic ciprofloxacin. J Bacteriol 188:7101–7110.

Costerton JW, Geesey GG, Cheng KJ. 1978. How bacteriastick. Sci Am 238:86–95.

Costerton JW, Stewart PS, Greenberg EP. 1999. Bacterialbiofilms: a common cause of persistent infections.Science 284:1318–1322.

Courcelle J, Khodursky A, Peter B, Brown PO, HanawaltPC. 2001. Comparative gene expression profiles follow-ing UV exposure in wild-ype and SOS-deficient Escher-ichia coli. Genetics 158:41–64.

Cox MM. 1998. A broadening view of recombinational DNArepair in bacteria. Genes Cells 3:65–78.

Dallo SF, Weitao T. 2010a. Insights into Acinetobacter war-wound infections, biofilms and control. Adv Skin WoundCare 23:169–174.

Dallo SF, Weitao T. 2010b. Bacteria under SOS evolveanticancer phenotypes. Infect Agents Cancer 5:3.

Donlan RM, Costerton JW. 2002. Biofilms: survival mechan-isms of clinically relevant microorganisms. Clin Micro-biol Rev 15:167–193.

Elasri MO, Miller RV. 1999. Study of the response of abiofilm bacterial community to UV radiation. ApplEnviron Microbiol 65:2025–2031.

Fernandez de Henestrosa AR, Ogi T, Aoyagi S, Chafin D,Hayes JJ, Ohmori H, Woodgate R. 2000.Identification of additional genes belonging to theLexA regulon in Escherichia coli. Mol Microbiol35:1560–1572.

Geier H, Mostowy S, Cangelosi GA, Behr MA, Ford TE.2008. Autoinducer-2 triggers the oxidative stress responsein Mycobacterium avium, leading to biofilm formation.Appl Environ Microbiol 74:1798–1804.

Gotoh H, Zhang Y, Dallo SF, Hong S, Kasaraneni N,Weitao T. 2008. Pseudomonas aeruginosa under DNAreplication inhibition tends to form biofilms via Arr. ResMicrobiol 159:294–302.

Hall-Stoodley L, Costerton JW, Stoodley P. 2004. Bacterialbiofilms: from the natural environment to infectiousdiseases. Nature Rev Microbiol 2:95–108.

Hoffman LR, D’Argenio DA, MacCoss MJ, Zhang Z, JonesRA, Miller SI. 2005. Aminoglycoside antibioticsinduce bacterial biofilm formation. Nature 436:1171–1175.

Hung DT, Zhu J, Sturtevant D, Mekalanos JJ. 2006. Bileacids stimulate biofilm formation in Vibrio cholerae.Molec Microbiol 59:193–201.

Imlay JA, Linn S. 1987. Mutagenesis and stress responsesinduced in Escherichia coli by hydrogen peroxide. JBacteriol 169:2967–2976.

Jackson K, Keyser R, Wozniak DJ. 2003. The role ofbiofilms in airway disease. Semin Respir Crit Care Med24:663–670.

Khil PP, Camerini-Otero RD. 2002. Over 1000 genes areinvolved in the DNA damage response of Escherichiacoli. Mol Microbiol 44:89–105.

Linares JF, Gustafsson I, Baquero F, Martinez JL. 2006.Antibiotics as intermicrobial signaling agents instead ofweapons. Proc Natl Acad Sci 103:19484–19489.

Little JW. 1991. Mechanism of specific LexA cleavage:autodigestion and the role of RecA coprotease. Biochi-mie 73:411–421.

Murakawa T. 1973. Slime production by Pseudomonasaeruginosa. 3. Purification of slime and its physicochem-ical properties. Jpn J Microbiol 17:273–281.

Nemoto K, Hirota K, Murakami K, Taniguti K, Murata H,Viducic D, Miyake Y. 2003. Effect of varidase (strepto-dornase) on biofilm formed by Pseudomonas aeruginosa.Chemotherapy 49:121–125.

Obritsch MD, Fish DN, MacLaren R, Jung R. 2005.Nosocomial infections due to multidrug-resistant Pseu-domonas aeruginosa: epidemiology and treatment op-tions. Pharmacotherapy 25:1353–1364.

Ramsey DM, Wozniak DJ. 2005. Understanding the controlof Pseudomonas aeruginosa alginate synthesis and theprospects for management of chronic infections in cysticfibrosis. Mol Microbiol 56:309–322.

610 H. Gotoh et al.

Downloaded By: [Weitao, Tao] At: 18:53 5 July 2010

Sanders LH, Rockel A, Lu H, Wozniak DJ, Sutton MD.2006. Role of Pseudomonas aeruginosa dinB-encodedDNA polymerase IV in mutagenesis. J Bacteriol188:8573–8585.

Sassanfar M, Roberts JW. 1990. Nature of the SOS-inducingsignal in Escherichia coli: the involvement of DNAreplication. J Mol Biol 212:79–96.

Sauer K, Camper AK, Ehrlich GD, Costerton JW, DaviesDG. 2002. Pseudomonas aeruginosa displays multiplephenotypes during development as a biofilm. J Bacteriol184:1140–1154.

Schembri MA, Kjaergaard K, Klemm P. 2003. Global geneexpression in Escherichia coli biofilms. Mol Microbiol48:253–267.

Sherratt DJ. 2003. Bacterial chromosome dynamics. Science301:780–785.

Southey-Pillig CJ, Davies DG, Sauer K. 2005. Characteriza-tion of temporal protein production in Pseudomonasaeruginosa biofilms. J Bacteriol 187:8114–8126.

Stewart PS, Costerton WJ. 2001. Antibiotic resistance ofbacteria in biofilms. Lancet 358:135–138.

Stoodley P, Sauer K, Davies DG, Costerton JW. 2002.Biofilms as complex differentiated communities. AnnuRev Microbiol 56:187–209.

Takahashi A, Yomoda S, Ushijima Y, Kobayashi I, InoueM. 1995. Ofloxacin, norfloxacin and ceftazidime increasethe production of alginate and promote the formation ofbiofilm of Pseudomonas aeruginosa in vitro. J AntimicrobChemother 36:743–745.

VanBogelen RA, Kelley PM, Neidhardt FC. 1987. Differ-ential induction of heat shock, SOS, and oxidation stressregulons and accumulation of nucleotides in Escherichiacoli. J Bacteriol 169:26–32.

Walker GC. 1984. Mutagenesis and inducible responses todeoxyribonucleic acid damage in Escherichia coli. Micro-biol Mol Biol Rev 48:60–93.

Weitao T. 2009a. Multicellularity of a unicellular organism inresponse to DNA replication stress. Res Microbiol160:87–88.

Weitao T. 2009b. Bacteria form biofilms against cancermetastasis. Med Hypotheses 72:477–478.

Whitchurch CB, Tolker-Nielsen T, Ragas PC, Mattick JS.2002. Extracellular DNA required for bacterial biofilmformation. Science 295:1487.

Whiteley M, Bangera MG, Bumgarner RE, Parsek MR,Teitzel GM, Lory S, Greenberg EP. 2001. Geneexpression in Pseudomonas aeruginosa biofilms. Nature413:860–864.

Biofouling 611

Downloaded By: [Weitao, Tao] At: 18:53 5 July 2010