isolation and characterization of pseudomonas aeruginosa genes inducible by respiratory mucus...

Isolation and characterization of Pseudomonasaeruginosa genes inducible by respiratory mucusderived from cystic fibrosis patients

Jingyi Wang, 1 Stephen Lory, 2 Reuben Ramphal 3 andShouguang Jin 1*1Department of Microbiology and Immunology, Universityof Arkansas for Medical Sciences, Little Rock, Arkansas72205, USA.2Department of Microbiology, University of Washington,Seattle, Washington 98195, USA.3Department of Infectious Diseases, University of Florida,Gainesville, Florida 32610, USA.

Summary

Pseudomonas aeruginosa , an opportunistic humanpathogen, is a major causative agent of mortality andmorbidity in immunocompromised individuals andthose with cystic fibrosis (CF). In CF patients, the sec-retion of abnormally high amounts of mucus into theairways contributes to their susceptibility to infec-tion by P. aeruginosa . To identify virulence genes ofP. aeruginosa that are important in infection of CFpatients, an in vivo selection system (IVET) was usedto identify promoters that are specifically inducibleby respiratory mucus derived from CF patients. Threegenetic loci that are highly inducible by the mucuswere identified. One of them is a well-characterizedvirulence gene ( fptA ), encoding the receptor for pyo-chelin, which is a P. aeruginosa iron siderophore.Induction of the fptA gene by mucus is suppressedby the addition of exogenous iron, demonstratingthat the mucus is an iron chelator and generates aniron-deficient environment in CF lungs. Therefore, asa part of the host-defence mechanism, the mucuscould also be responsible for induction of iron-regu-lated virulence factors of bacterial pathogens. Thesecond locus, np20 , encodes a peptide that sharessequence homology to a number of transcriptionalregulators. An identical locus was previously identi-fied to be inducible in vivo during infection of miceand was shown to be important in bacterial virulencein a neutropenic-mouse infection model. The third

locus, designated migA (mucus inducible gene), wassequenced and found to encode a 299-amino-acidpeptide which is homologous to glycosyltransferasesof other bacteria, and is involved in the biosynthesisof lipopolysaccharides or exopolysaccharides. Induc-ibilities of the np20 and migA genes are not affectedby iron and the exact nature of the inducing signalsin the mucus is not known. The possible implicationsof the migA inducibility by respiratory mucus is dis-cussed in relation to the P. aeruginosa infection in CF.

Introduction

The common environmental microorganism Pseudomo-nas aeruginosa has a remarkable ability to adapt to andpersist in a variety of environments, including humanhosts, which contributes significantly to the ability of thisbacterium to cause a variety of human infections. Inpatients with cystic fibrosis (CF), extensive colonizationby P. aeruginosa and the accompanying inflammation isresponsible for the severe and often fatal chronic respir-atory disease (Gilligan, 1991). As the conditions of thehuman respiratory tracts are drastically different fromthose of the natural environment, it is reasonable to expectthat as a part of its adaptation during initial colonization, P.aeruginosa would induce expression of genes that encodedeterminants necessary for survival in the host environ-ment, while repressing genes that are unnecessary oreven deleterious. Supporting this hypothesis are clinicalobservations that a number of phenotypic differencesare readily detectable between P. aeruginosa strains iso-lated from CF patients and those isolated from naturalenvironments. For CF isolates, these include overproduc-tion of the alginate exopolysaccharide (Govan, 1988), lossof motility (Mahenthiralingam et al., 1994), alteration in thelipopolysaccharide structure (Hancock et al., 1983), anddownregulation of expression of a number of virulence fac-tors (Woods et al., 1986). These changes may reflectselection of mutant variants within the host environmentor differential gene expression directed by a complex sig-nal-transduction network which is responsible for co-ordi-nate gene expression required for bacterial survivalduring infection.

Molecular Microbiology (1996) 22(5), 1005–1012

Q 1996 Blackwell Science Ltd

Received 14 June, 1996; revised 18 September, 1996; accepted 26September, 1996. *For correspondence. E-mail [email protected]; Tel. (501) 2961396; Fax (501) 6865359.

m

CF patients secrete abnormally high amounts of mucusin their airways, which become colonized by P. aerugi-nosa. The environmental condition of the airways of CFpatients can therefore be partially mimicked in vitro bysolutions of respiratory mucus. The mucus of CF patientsis physically more viscous and chemically different in com-position from that of healthy persons (Carnoy et al., 1993).The mucus not only provides binding sites for bacterialadhesins but also physically blocks the host defence sys-tem by inhibiting normal muco-cilliary clearance mechan-isms and the mobility of vacuolar macrophages (Carnoyet al., 1993; 1994; Marshall and Carroll, 1991). Theseare believed to be the critical factors determining the sus-ceptibility of CF patients to the infection by P. aeruginosa.Because most bacterial virulence genes are specificallyexpressed during infection of their hosts, it is reasonableto predict that any P. aeruginosa genes that are specifi-cally induced by the mucus are likely to play importantroles in the infection of CF patients.

In previous studies, we have described a selection sys-tem for P. aeruginosa, similar to the in vivo expressiontechnology (IVET) first described in Salmonella, and usedit in the identification of P. aeruginosa genes that are spe-cifically induced during infection of neutropenic mice(Wang et al., 1996). In this report, we extend our in vivostudies to in vitro conditions simulating human infections,describing the isolation of P. aeruginosa genes that arespecifically induced by respiratory mucus derived fromCF patients, and the characterization of three such iso-lated genetic loci.

Results

Preparation of the mucus derived from CF patients

To determine the percentage of mucus to be used in theinduction experiments, we tested the growth of a wild-type P. aeruginosa strain PAK-SR in streptomycin-con-taining minimal medium A (see the Experimental proce-dures ) supplemented with 50%, 25%, 10% and 5% ofthe crude mucus, respectively. With 5% or 10% mucus,bacterial cell density reached 109 cells ml¹1 after 16 h,whereas concentrations of mucus exceeding 10% resultedin much lower final bacterial cell densities, i.e. 107 and 108

cells ml¹1 at 50% and 25% mucus, respectively. There-fore, a 10% final concentration of the mucus was used inall experiments described in this article.

In order to determine the selection conditions, the yieldsof the wild-type PAK-SR and the purEK-deletion strainPAK-AR2 on minimal medium A with or without 10%mucus were compared, after growth for 16 h at 378C. Asthe crude mucus samples from CF patients could partiallypromote the growth of PAK-AR2 in minimal medium A(Fig. 1), the crude mucus samples were dialysed toremove low-molecular-weight substrates that promotebacterial growth. With dialysed mucus, no growth of thePAK-AR2 strain was observed (Fig. 1).

Isolation of genes that are inducible by mucus of CFpatients

Initially, 106 cells of the chromosomal co-integrate bank

Q 1996 Blackwell Science Ltd, Molecular Microbiology, 22, 1005–1012

Fig. 1. Growth of a wild-type PAK-SR strainand a purEK-deletion strain PAK-AR2 inminimal medium A (Min A) containing 10%crude or dialysed mucus. Approximately 104

cells were inoculated into 2 ml of liquid mediaand grown at 378C for 16 h. Bacterial celldensities in the final cultures were calculatedby counting the number of colonies growingon L-agar containing streptomycin after platinga serial dilution of the bacterial cultures.

1006 J. Wang, S. Lory, R. Ramphal and S. Jin

(Wang et al., 1996) were grown in 50 ml of minimal med-ium A containing 10% of the dialysed mucus and appro-priate antibiotics. Cell density reached 109 ml¹1 after 48h of incubation, from which 106 cells were removed andused to repeat the above growth cycle for two moretimes. Theoretically, cells expressing the purEK genewere enriched 109-fold over non-purEK expressing onesafter three cycles of the growth (i.e. 103-fold enrichmentin each cycle), therefore cells in the final population shouldhave active promoters controlling the chromosomally inte-grated purEK gene under the above growth condition.From the final culture, 104 cells were plated onto minimalmedium A containing appropriate antibiotics and a limitedamount of adenine (5 mg ml¹1) to support the growth ofdesired cells, i.e. those with the purEK gene under thecontrol of promoters that are turned off in the absence ofmucus. Two different-sized colonies appeared, large andsmall, with a ratio of 200:1, respectively. The originalbank not exposed to mucus had a large to small colonyratio of <1:50. Large colonies represent cells with consti-tutive promoters driving the purEK gene, whereas smallcolonies represent cells having promoters that are specifi-cally inducible by the mucus. A total of 60 small colonieswere examined further.

Individual isolates were re-tested for a growth advan-tage in the presence of mucus. Of the 60 small colonies,22 showed a requirement for mucus when grown at 378Cfor 16 h. These 22 isolates had more than 10-fold higherfinal cell density in minimal medium A containing mucus,compared with the cell density of identical isolates grownin the same minimal medium lacking mucus, indicating

that mucus-inducible promoters control the expression ofthe purEK genes in these strains.

Characterization of the genetic loci inducible by mucus

The location of chromosomal fusions of the purEK gene inthe above 22 mucus-dependent isolates was determinedby Southern hybridization analysis. Chromosomal DNAwas digested with HindIII and probed with a 1.5 kb DNAfragment containing the purEK coding region. As shownin Fig. 2B, all 22 isolates had chromosomal purEK fusionsand they seemed to fall into three different groups, asjudged by the sizes of the hybridizing bands.

To further characterize these mucus-inducible loci, DNAfragments upstream of the chromosomally fused purEKgene were isolated. Each fragment of the chromosomalDNA was digested with EcoRI, self-ligated and transfor-med into E. coli strain DH5a. Because there is a uniqueEcoRI site in the selection vector, re-circularization ofthe vector following EcoRI digestion results in a plasmidwhich carries DNA upstream of the purEK gene, purEKitself, the origin of plasmid replication and the ampicillin-resistance gene (see Fig. 2A). Plasmids from the resultingtransformants, so-called ‘EcoRI-rescued plasmids’, wereused to sequence DNA fragments upstream of the purEKgene (see the Experimental procedures ). Approximately500 bp of sequence was obtained from each plasmid. Thecomparison of sequences from the 22 isolates showedthat the purEK was fused to one of three different genes,tentatively named genes A, B and C. Sixteen out of the22 isolates (no. 1 and nos 8–22) had purEK fusions to


Fig. 2. A. Representative map of thechromosomal co-integrate used in the IVETselection procedure.B. Southern hybridization of the chromosomalDNA from 22 strains isolated by the IVETselection procedure using mucus as aninducer. Chromosomal DNA was digestedwith HindIII and probed with the purEK gene(1.5 kb BamHI fragment from pSJ9437).Lanes 1–22 correspond to the 22 differentstrains.

Mucus-inducible genes of Pseudomonas aeruginosa 1007

gene A, four had fusions to gene B (nos 2, 5, 6 and 7) andthe remaining two (nos 3 and 4) had fusions to gene C.

Homologues of the three genetic loci were searched insequence databases and several relevant sequenceswere identified. The gene A encodes a peptide that hasa high degree of sequence homology to other bacterialgene products involved in the synthesis of lipopolysac-charide (LPS) or exopolysaccharide (EPS) (Fig. 3). There-fore, gene A is a previously unidentified locus, involved inLPS or EPS biosynthesis, and we thus designated it migA(mucus inducible gene).

The DNA sequence of gene B was identical to that of thegene np20 which had previously been identified by in vivoselection using neutropenic mice (Wang et al., 1996). Thislocus encodes a putative transcriptional regulator and dis-ruption of this locus resulted in a drastic reduction in bac-terial virulence in neutropenic mice (Wang et al., 1996).

The DNA sequence of gene C was identical to that of thepyochelin receptor gene (fptA ), a well-studied virulencegene of P. aeruginosa (Ankenbauer, 1992; Ankenbauerand Quan, 1994). The protein FptA is a surface-localizedreceptor which binds iron-bound form of siderophore, pyo-chelin, thus mediating the uptake of iron into the bacterialcell.

Mucus is an iron chelator and generates iron-deficientenvironments

In order to confirm the inducibilities of the three isolatedloci by mucus, plasmids pMS12lacV, pMS13lacV andpMS18lacV, containing the np20, fptA and migA genefusions to a promoterless lacZ gene, respectively, wereintroduced into the wild-type PAK strain and levels ofb-galactosidase were determined after growth in thepresence and absence of mucus. As shown in Fig. 4A,all three loci are highly inducible by mucus.

It is recognized that the genes which specify the

ferripyochelin-uptake system, including fptA, are induciblewhen P. aeruginosa is cultured under conditions of ironlimitation (Ankenbauer, 1992 ). To examine whether theinducibility of this locus by mucus is due to iron limitation,we tested the effect of iron on the inducibility of the fptAlocus by mucus. Wild-type P. aeruginosa strain PAK con-taining the fptA::lacZ fusion construct, pMS13lacV, wastested for the levels of b-galactosidase activity whengrown in minimal medium A containing mucus with or with-out FeCl3. As shown in Fig. 4B, induction of the fptA pro-moter by mucus was totally suppressed when 50 mg ml¹1

of FeCl3 was added, suggesting that the mucus is aniron chelator and generates an iron-limiting environmentin minimal medium A. Further, the effect of mucus canbe mimicked by the addition of a known iron chelatorEDDA (ethylenediamine-di(o-hydroxyphenylacetic acid)),to minimal medium A (Fig. 4B). These data suggest thatthe lungs of CF patients are likely to be iron limiting dueto the presence of high amount of mucus.

Similar analysis using the migA::lacZ and np20 ::lacZfusion constructs, pMS18lacV and pMS12lacV, failed toshow the reversal of the mucus inducibility by iron, thusshowing that migA and np20 genes are not regulated byiron (data not shown). Therefore, besides iron, mucuscontains additional signalling molecules affecting bacterialgene expression.

migA encodes a glycosyltransferase homologue

A cosmid clone, pVKBK19, containing the intact migAlocus was identified by colony hybridization of E. coli con-taining PAK chromosomal DNA gene bank, with a migAprobe derived from the partial upstream sequence inpMS18. Restriction sites of the cosmid were mappedand a 3 kb DNA fragment containing the migA locus wassubcloned (pMSB3A). The 3 kb DNA insert in pMSB3Awas sequenced in both directions (see the Experimental


Fig. 3. MigA shares a sequence homology to glycosyltransferases of other bacteria. IgtD, a glycosyltransferase of Haemophilus influenzae(Fleischmann, 1995); IgtA, a glycosyltransferase of Neisseria gonorrhoeae (Gotschlich, 1994); RfbB, an rfb operon gene product of Yersiniaenterocolitica (Zhang et al., 1993); EpsI, an exopolysaccharide biosynthetic enzyme of Streptococcus thermophilus (Stingele et al., 1996);RfpA, a glycosyltransferase of Shigella dysenteriae (Gohmann et al., 1994).


procedures ). A complete open reading frame (ORF),encoding a 299-amino-acid peptide, was identified (Gen-Bank Accession Number U70729). It is preceded by aputative ribosome binding site and ¹10/¹35 sequencescharacteristic of prokaryotic promoter elements. The origi-nal chromosomal purEK fusion occurred at amino acidposition 134.

As predicted from its partial sequence, the migA-encoded product shares a high degree of amino acidsequence homology with glycosyltransferases of other bac-teria. These include glycosyltransferases from Haemophilusinfluenzae (Fleischmann, 1995), Neisseria gonorrhoeae(Gotschlich, 1994), Shigella dysenteriae (Gohmann et al.,1994) and Yersinia enterocolitica (Zhang et al., 1993) thatare involved in LPS biosynthesis, and those from Strepto-coccus thermophilus (Stingele et al., 1996), Erwinia amy-lovora (Bugert and Geider, 1995) and Bacillus subtilus(Glaser et al., 1993) that are involved in the biosynthesisof EPS. The homology is restricted to the N-terminal halfof the protein and the most homologous enzyme was a gly-cosyltransferase (IgtD) from H. influenzae (Fleischmann,1995), with 43% identity and 74% similarity over a 92-amino-acid region.

Discussion

Using the previously constructed IVET selection system,we have examined the role of human mucus in activatingexpression of P. aeruginosa genes. As the mucus liningof the human respiratory tract is encountered by infectingbacteria very early and during the subsequent chroniccolonization of CF patients (Ramphal et al., 1991), itseems reasonable to speculate that this interaction couldbe an important signalling mechanism for the regulationof specific virulence genes.

Since mucus from non-CF individuals was also capableof inducing the expression of the same genes as the CF-derived mucus, the signalling event may be a general phe-nomenon of pathogen recognition of the host environment,and not a consequence of specific interaction with thehyperhydrated and undersulphated CF mucus.

Three genes that are highly inducible by mucus weredescribed in this report. Two of these, fptA and np20,have been identified previously using the IVET selectionmethod in the neutropenic-mouse model (Wang et al.,1996). The identification of these two loci by two indepen-dent methods may reflect their general role in pathogen-esis and further validates the utility of the IVET selectionmethod to identify new virulence factors that function inmany different tissues. However, there is little doubt thatfor P. aeruginosa, which can cause a wide range of infec-tions, there are signals that result in regulation of only alimited number of genes necessary for colonization of aparticular tissue or organ. Our studies have indeed iden-tified the existence of a set of genes that was induced inthe immunosupressed-mouse model but not expressedin mucus. Moreover, at least one gene, migA, that washighly expressed in mucus was not detected among thepromoters of 22 genes identified during IVET selection inthe neutropenic mouse (Wang et al., 1996).

The pyochelin receptor, FptA, is a bacterial outer


Fig. 4. A. Promoters of migA, np20 and fptA are activated bymucus. Strains of PAK containing pDN19lacV (vector control),pMS18lacV (migA::lacZ fusion), pMS12lacV (np20 ::lacZ fusion)and pMS13lacV (fptA::lacZ fusion), were grown in minimal mediumA with or without 10% mucus for 16 h at 378C.B. Mucus is an iron chelator. PAK containing the fptA::lacZ fusion(pMS13lacV) or vector only (pDN19lacV) were grown for 16 h inminimal medium A with the supplements indicated. Finalconcentrations of the supplements were 10% for mucus, 50 mg ml¹1

for FeCl3 and 100 mM for EDDA.


membrane protein involved in pyochelin-mediated irontransport (Ankenbauer, 1992; Ankenbauer and Quan,1994). In response to iron deprivation, P. aeruginosa pro-duces two unrelated siderophores, pyoverdin and pyo-chelin. Pyochelin stimulates bacterial growth in murineinfections (Cox, 1982), reverses iron deprivation causedby human serum and transferrin (Ankenbauer et al.,1985), and efficiently removes iron from transferrin (Sriyo-sachati and Cox, 1986). Mutants with defects in pyochelin-mediated Fe(III) transport were observed to be markedlyless virulent than wild-type strains of P. aeruginosa(Sokol, 1987). The observation that the inducibility of thefptA locus by mucus is suppressible by the addition offree iron indicates that mucus is an iron chelator, andfurther suggests that the environments of CF lungs arelikely to be iron limiting. This and other studies supportthe hypothesis that iron regulated genes of the P. aerugi-nosa play important roles in the infection of CF patients.Scavenging iron in human tissues is an essential bacterialvirulence trait, because it is required for bacterial growth.Moreover, iron is also a signalling molecule and theexpression of several virulence factors, such as exotoxinA, are influenced by iron availability. The expression ofthe exotoxin A gene, toxA, is controlled by RegA, aniron-regulated transcriptional factor (Bjorn et al., 1978;Frank and Iglewski, 1988). Both the toxA and regAgenes are indeed inducible by mucus (J. Wang and S.Jin, unpublished).

The np20 locus encodes a transcriptional regulator thatshares amino acid sequence similarity to the ferric-uptakeregulators (Fur). A np20-null-mutant strain of P. aerugi-nosa was shown to be much less virulent than the wildtype in neutropenic mice (Wang et al., 1996). The identifi-cation of this locus by two independent methods mayreflect its general role in pathogenesis. As np20 encodesa putative repressor, the function of the gene product mayinvolve direct repression of a gene which, when expres-sed, may interfere with the pathogenesis process. Alterna-tively, the np20 product may be a part of a regulatorycascade, which controls the expression (repression or acti-vation) of one or several genes encoding virulence factors.

We have also identified a mucus-inducible gene that hasnot been previously described. The migA locus encodesan ORF that is homologous to glycosyltransferases fromseveral Gram-negative bacteria. It is conceivable thatthe migA locus is part of a larger operon involved in bio-synthesis or modification of P. aeruginosa LPS or EPS.The rationale for the inducibility of the gene(s) involvedin LPS/EPS biosynthesis, in response to the mucus, isnot clear. Full understanding of the expression of thisoperon will require further analysis of this region, includingthe organization of all of the linked genes. It is interestingto note that LPS and EPS modifications, i.e. conversionof LPS-smooth to LPS-rough variants, together with

non-mucoid to mucoid conversion, are the characteristicphenotypic changes associated with P. aeruginosa recov-ered from the respiratory tracts of chronically infected CFpatients (Govan, 1988; Hancock et al., 1983). It is possiblethat alginate (EPS) and LPS, and perhaps other virulencefactors, are co-ordinately regulated and the initial signal-ling event is the interaction of the bacteria with mucus.Further studies are needed to determine the exact enzy-matic activity of the migA gene product in the synthesisor modification of the bacterial LPS or EPS.

The crude mucus supported a limited growth of thepurEK-mutant strain, and removal of the growth-promotingsubstrates by dialysis indicate that they are low-molecular-weight materials. The origin of this purine source in themucus is not clear. It could have been derived from host-cell secretions or DNA degradation following host or bac-terial cell lysis.

Isolation and characterization of the P. aeruginosagenes specifically inducible by respiratory mucus fromCF patients not only enabled us to identify new virulencegenes but also to probe the in vivo environment of CFlungs. This approach may provide us with alternativeways of controlling P. aeruginosa infections. Our initialscreening of 22 isolates, following induction by mucus,resulted in the identification of only three different promo-ters, which is explainable by the fact that the selection pro-cedure is an enrichment procedure and the extent ofenrichment depends on the promoter strength that con-trols the expression of chromosomally fused purEKgenes. The three isolates represent strongly inducible pro-moters; most of the weakly inducible loci, which, neverthe-less, control expression of important virulence factors,may have been eliminated during the selection process.By screening a larger number of colonies, after reducedrounds of selection on the mucus-containing medium, weshould be able to isolate additional mucus-inducible loci,including some weakly inducible ones which might play cri-tical roles in bacterial infection of CF patients.

Experimental procedures

Bacterial strains and plasmids

Strains and plasmids used in this work are listed in Table 1.The purEK-deleted strain, PAK-AR2, was constructed byreplacing the chromosomal purEK gene of the PAK with anV fragment (Wang et al., 1996). The chromosomal co-integrate bank of cells used for IVET selection has previouslybeen described (Wang et al., 1996). The cosmid-clone bankof the PAK chromosomal DNA was also described elsewhere(Ishimoto and Lory, 1989). The lacZ reporter-gene fusions tonp20, fptA and migA are generated by introducing EcoRI–BamHI fragments from pMS12, pMS13 and pMS18, respec-tively, into the EcoRI–BamHI sites of the lacZ fusion vectorpDN19lacV (Totten and Lory, 1990), generating pMS12lacV,pMS13lacV and pMS18lacV, respectively.



Bacterial culture conditions

E. coli and P. aeruginosa strains were grown in Luria (L)-agarand L-broth at 378C. Minimal medium A (Davis and Mingioli,1950) was also used for the growth of P. aeruginosa.

Antibiotics were used, as required, at the following finalconcentrations. For E. coli : ampicillin, 100 mg ml¹1; spectino-mycin, 50 mg ml¹1; streptomycin, 25 mg ml¹1; and tetracycline20 mg ml¹1. For P. aeruginosa : carbenicillin, 150 mg ml¹1;spectinomycin, 200 mg ml¹1; streptomycin, 200 mg ml¹1; neo-mycin, 200 mg ml¹1; and tetracycline, 100 mg ml¹1.

Adenine was added at 50 mg ml¹1 for a full supplement and5 mg ml¹1 for a limited supplement to support the growth of thepurEK-mutant strain on minimal medium.

DNA sequence analysis

DNA sequencing analysis was performed by polymerasechain reaction (PCR)-mediated Taq DyeDeoxy TerminatorCycle sequence using an Applied Biosystems model 373ADNA sequencer. For sequencing DNA fragments upstreamof the purEK gene, purEK primer (58-CACGCCAACCAG-TGCGCTCATCG-38), complementary to the 58 end of thepurEK coding region, was used. For sequencing the migAlocus, nested deletions in both directions were generated onpMSB3A by the ExoIII–Mung-bean-nuclease method (Mania-tis et al., 1982). DNA restriction sites and ORF analyses wereconducted using the DNA STRIDER program. BLAST and GCG pro-grams were used to analyse DNA and amino acid sequencehomology. The complete sequence of the migA gene isdeposited in GenBank (Accession Number U70729).

Preparation of the mucus samples

Respiratory mucus samples were collected from two CF

patients by expectoration during physical therapy and areconsidered to be 100% (w/v). The mucus samples were dia-lysed against >200× vol. of distilled water overnight at 48C,diluted to 40% (w/v), and frozen at –708C. For inductionassays, minimal medium A was supplemented with 10%final concentration of the above mucus.

Respiratory mucus from CF patients contains bacteria,mainly P. aeruginosa. In order to effectively inhibit the growthof all contaminating bacteria, several antibiotics were tested.Tetracycline, gentamicin and streptomycin all inhibited growthof any bacteria in the mucus sample, as determined by nobacterial colony formation 72 h after plating mucus sampleson minimal medium A containing the individual antibiotics.Streptomycin (at 200 mg ml¹1) was included in the mucus-containing medium to allow the growth of the streptomycin-resistant strains of PAK-SR as well as PAK-AR2.

Miscellaneous methods

Beta-galactosidase activity was measured as described byMiller (1972). As mucus gives turbidity readings, A600 valuesfor the b-galactosidase-activity measurements were adjustedby subtracting the value from a control medium which had nobacterial inoculum. Southern hybridization analyses werecarried out by using the ECL labelling and detection kit fromAmersham.

Acknowledgements

We would like to thank Dr Marie Chow for critical reading ofthis manuscript. This work was supported, in part, by NIHgrants R29AI39524 (to S.J.) and HLBI33622 (to R.R.), anda Pilot and Feasibility grant from the Centre for Gene Therapyof University of Washington (to S.L.). A portion of this work


Table 1. Strains and plasmids used in this work.

Strain/Plasmid Description Source/Reference

Strain

E. coliDH5a endA1 hsdR17 supE44 thi-1 recA1 gyrA96 relA1 D(lacYZA–argF) U169 l-f80

dlacZDM15 ; recipient for recombinant plasmidsBRL

P. aeruginosaPAK Wild-type clinical isolate David BradleyPAK-SR Spontaneous SmR isolate of the PAK Ishimoto and Lory (1989)PAK-AR2 PAK deleted of the purEK gene; SpR, SmR Wang et al. (1996)

Plasmid

pTZ18R E. coli cloning vector; ApR USBpUC19V ‘V’ fragment clone in pUC19; ApR, SpR, SmR Koga et al. (1993)pMS12 Plasmid rescued from the chromosome of isolate no. 2 by EcoRI; ApR This workpMS13 Plasmid rescued from the chromosome of isolate no. 3 by EcoRI; ApR This workpMS18 Plasmid rescued from the chromosome of isolate no. 8 by EcoRI; ApR This workpVKBK19 Cosmid clone that contains a migA locus; KmR This workpMSB3A 3 kb EcoRI fragment, containing migA gene, cloned into pTZ18R; ApR This workpDN19lacV Promoterless lacZ fusion vector; TcR, SpR, SmR Totten and Lory (1990)pMS12lacV np20 promoter fused to lacZ in pDN19lacV; TcR, SpR, SmR This workpMS13lacV fptA promoter fused to lacZ in pDN19lacV; TcR, SpR, SmR This workpMS18lacV migA promoter fused to lacZ in pDN19lacV; TcR, SpR, SmR This workpSJ9347 purEK coding region, 1.5 kb BamHI fragment, cloned into a pTZ18R derivative; ApR Wang et al. (1996)

SpR, spectinomycin-resistance marker; SmR, streptomycin-resistance marker; ApR, ampicillin-resistance marker; KmR, kanamycin-resistancemarker; TcR, tetracycline-resistance marker.


was carried out while one of us (S.J.) held a fellowship fromthe NIH/Cystic Fibrosis-supported RDP at the University ofWashington.

References

Ankenbauer, R.G. (1992) Cloning of the outer membranehigh-affinity Fe (III)-pyochelin receptor of Pseudomonasaeruginosa. J Bacteriol 174: 4401–4409.

Ankenbauer, R.G., and Quan, H.N. (1994) FptA, the Fe (III)-pyochelin receptor of Pseudomonas aeruginosa : a phe-nolate siderophore receptor homologous to hydroxamatesiderophore receptors. J Bacteriol 176: 307–319.

Ankenbauer, R., Sriyosachati, S., and Cox, C.D. (1985)Effects of siderophores on the growth of Pseudomonasaeruginosa in human serum and transferrin. Infect Immun49: 132–140.

Bjorn, M.J., Iglewski, B.H., Ives, S.K., Sadoff, J.C., and Vasil,M.L. (1978) Effect of iron on yields of exotoxin A in culturesof Pseudomonas aeruginosa PA-103. Infect Immun 19:785–791.

Bugert, P., and Geider, K. (1995) Molecular analysis of theams operon required for exopolysaccharide synthesis ofErwinia amylovora. Mol Microbiol 15: 917–933.

Carnoy, C., Ramphal, R., Scharfman, A., Lo, G.J.M.,Houdret, N., Klein, A., Galabert, C., Lamblin, G., andRoussel, P. (1993) Altered carbohydrate composition ofsalivary mucins from patients with cystic fibrosis and theadhesion of Pseudomonas aeruginosa. Am J Respir CellMol Biol 9: 323–334.

Carnoy, C., Scharfman, A., Van, B.E., Lamblin, G., Ramphal,R., and Roussel, P. (1994) Pseudomonas aeruginosaouter membrane adhesins for human respiratory mucusglycoproteins. Infect Immun 62: 1896–1900.

Cox, C.D. (1982) Effect of pyochelin on the virulence ofPseudomonas aeruginosa. Infect Immun 36: 17–23.

Davis, B.D., and Mingioli, E.S. (1950) Mutants of Escherichiacoli requiring vitamin B12. J Bacteriol 60: 17–28.

Fleischmann, R.D. et al. (1995) Whole-genome randomsequencing and assembly of Haemophilus influenzae Rd.Science 169: 496–512.

Frank, D.W., and Iglewski, B.H. (1988) Kinetics of toxA andregA mRNA accumulation in Pseudomonas aeruginosa. JBacteriol 170: 4477–4483.

Gilligan, P.H. (1991) Microbiology of airway disease inpatients with cystic fibrosis. Clin Microbiol Rev 4: 35–51.

Glaser, P., Kunst, F., and Arnaud, M. (1993) Bacillus subtilisgenome project: cloning and sequencing of the 97 kbregion from 325 degrees to 333 degrees. Mol Microbiol 10:371–384.

Gohmann, S., Manning, P.A., Alpert, C.A., Walker, M.J., andTimmis, K.N. (1994) Lipopolysaccharide O-antigen biosyn-thesis in Shigella dysenteriae serotype 1: analysis of theplasmid-carried rfp determinant. Microb Pathog 16: 53–64.

Gotschlich, E.C. (1994) Genetic locus for the biosynthesis ofthe variable portion of Neisseria gonorrhoeae lipooligo-saccharide. J Exp Med 180: 2181–2190.

Govan, J.R.W. (1988) Alginate biosynthesis and unusualcharacteristics associated with the pathogenesisof Pseudo-monas aeruginosa in cystic fibrosis. In Bacterial Infections

of Respiratory and Gastrointestinal Mucosae. Griffiths, E.,Donachie, W., and Stephen J. (ed.). Oxford: IRL Press, pp.67–96.

Hancock, R.E., Mutharia, L.M., Chan, L., Darveau, R.P.,Speert, D.P., and Pier, G.B. (1983) Pseudomonas aeru-ginosa isolates from patients with cystic fibrosis: a class ofserum-sensitive, non-typable strains deficient in lipopoly-saccharide O side chains. Infect Immun 42: 170–177.

Ishimoto, K.S., and Lory, S. (1989) Formation of pilin inPseudomonas aeruginosa requires the alternative sigmafactor (RpoN) of RNA polymerase. Proc Natl Acad Sci USA86: 1954–1957.

Koga, T., Ishimoto, K., and Lory, S. (1993) Genetic andfunctional characterization of the gene cluster specifyingexpression of Pseudomonas aeruginosa pili. Infect Immun61: 1371–1377.

Mahenthiralingam, E., Campbell, M., and Speert, D.P. (1994)Nonmotility and phagocytic resistance of Pseudomonasaeruginosa isolates from chronically colonized patientswith cystic fibrosis. Infect Immun 62: 596–605.

Maniatis, T., Fritsch, E.F., and Sambrook, J. (1982)Molecular Cloning: A Laboratory Manual. Cold SpringHarbor, New York: Cold Spring Harbor Laboratory Press.

Marshall, B.C., and Carroll, K.C. (1991) Interaction betweenPseudomonas aeruginosa and host defences in cysticfibrosis. Semin Respir Infect 6: 11–18.

Miller, J.H. (1972) Experiments in Molecular Genetics. ColdSpring Harbor, New York: Cold Spring Harbor LaboratoryPress.

Ramphal, R., Koo, L., Ishimoto, K.S., Totten, P.A., Lara, J.C.,and Lory, S. (1991) Adhesion of Pseudomonas aeruginosapilin-deficient mutants to mucin. Infect Immun 59: 1307–1311.

Sokol, P.A. (1987) Surface expression of ferripyochelin-binding protein is required for virulence of Pseudomonasaeruginosa. Infect Immun 55: 2021–2025.

Sriyosachati, S., and Cox, C.D. (1986) Siderophore-mediated iron acquisition from transferrin by Pseudomonasaeruginosa. Infect Immun 52: 885–891.

Stingele, F., Neeser, J.R., and Mollet, B. (1996) Identificationand characterization of the eps (exopolysacharide) genecluster from Streptococcus thermophilus Sfi6. J Bacteriol178: 1680–1690.

Totten, P.A., and Lory, S. (1990) Characterization of the typea flagellin gene from Pseudomonas aeruginosa PAK. JBacteriol 172: 7188–7199.

Wang, J.Y., Mushegian, A., Lory, S., and Jin, S. (1996) Largescale isolation of candidate virulence genes of Pseudomo-nas aeruginosa by in vivo selection. Proc Natl Acad SciUSA 93: 10434–10439.

Woods, D.E., Schaffer, M.S., Rabin, H.P., Campbell, G.D.,and Sokol, P.A. (1986) Phenotypic comparison of Pseu-domonas aeruginosa strains isolated from a variety ofclinical sites. J Clin Microbiol 24: 260–264.

Zhang, L., Al-Hendy, A., Toivanen, P., and Skurnik, M. (1993)Genetic organization and sequence of the rfb gene clusterof Yersinia enterocolitica serotype O:3: similarities to thedTDP-L-rhamnose biosynthesis pathway of Salmonella andto the bacterial polysaccharide transport systems. MolMicrobiol 9: 309–321.



isolation and characterization of pseudomonas aeruginosa genes inducible by respiratory mucus...

Documents