Global Genomic Analysis of AlgU ({sigma}E)-Dependent Promoters (Sigmulon) in Pseudomonas aeruginosa and Implications for Inflammatory Processes in Cystic Fibrosis

The conversion of Pseudomonas aeruginosa to the mucoid phenotypecoincides with the establishment of chronic respiratory infectionsin cystic fibrosis (CF). A major pathway of conversion to mucoidyin clinical strains of P. aeruginosa is dependent upon activationof the alternative sigma factor AlgU (P. aeruginosa ${sigma}$ ^E). Herewe initiated studies of AlgU-dependent global expression patternsin P. aeruginosa in order to assess whether additional genes,other than those involved in the production of the mucoid exopolysaccharidealginate, are turned on during conversion to mucoidy. Usinggenomic information and the consensus AlgU promoter sequence,we identified 35 potential AlgU ( ${sigma}$ ^E) promoter sites on the P.aeruginosa chromosome. Each candidate promoter was individuallytested by reverse transcription and mRNA 5'-end mapping usingRNA isolated from algU⁺ and algU::Tc^r mutant cells. A totalof 10 new AlgU-dependent promoters were identified, and thecorresponding mRNA start sites were mapped. Two of the 10 newlyidentified AlgU promoters were upstream of predicted lipoproteingenes. Since bacterial lipoproteins have been implicated asinducers of inflammatory pathways, we tested whether lipopeptidescorresponding to the products of the newly identified AlgU-dependentlipoprotein genes, lptA and lptB, had proinflammatory activity.In human peripheral blood monocyte-derived macrophages the peptidescaused production of interleukin-8, a proinflammatory chemokinetypically present at excessively high levels in the CF lung.Our studies show how genomic information can be used to uncoveron a global scale the genes controlled by a given ${sigma}$ factor (collectivelytermed here sigmulon) using conventional molecular tools. Inaddition, our data suggest the existence of a previously unknownconnection between conversion to mucoidy and expression of lipoproteinswith potential proinflammatory activity. This link may be ofsignificance for infections and inflammatory processes in CF.

INTRODUCTION

Top

Introduction

Cystic fibrosis (CF) is the most common lethal inheritable diseasein Caucasians (60). The primary contributors to the high morbidityand mortality in CF are the chronic respiratory infections causedby bacterial pathogens (59). The predominant CF pathogen isPseudomonas aeruginosa, and over 90% of CF patients eventuallybecome colonized with this organism (20). A classical featureof P. aeruginosa strains infecting CF patients is that theymutate into the mucoid, exopolysaccharide alginate-overproducingform, in a process referred to as conversion to the mucoid phenotype(23). This conversion is concomitant with the establishmentof chronic bacterial colonization (28, 42). The emergence ofmucoid strains also correlates with a poor clinical prognosis(23, 28, 42). The exact mechanisms leading to the worseningof disease coinciding with the conversion to mucoidy in P. aeruginosaare not fully understood, but are believed to stem from excessiveinflammation (4, 27, 29) and associated irreversible lung tissuedamage.

At the genetic level, the conversion to mucoidy in P. aeruginosaoccurs via mutations in a cluster of genes encoding the alternativesigma factor AlgU (35), also known as AlgT (16, 21), and anarray of AlgU regulators: MucA, MucB, MucC, and MucD (5, 7,36, 37). The mutations causing mucoidy in CF isolates most frequentlyoccur in the mucA gene (8, 37). These mutations release AlgUfrom the inhibitory activities of MucA (49, 53). AlgU is theP. aeruginosa ortholog of Escherichia coli and Salmonella ${sigma}$ ^E(63), an alternative sigma factor that directs transcriptionof genes in response to extreme stress conditions (24, 39, 48).Recently, it has been shown that AlgU can also direct transcriptionof the major heat shock sigma factor RpoH (50). As an alternativesigma factor, AlgU is likely to play a role in global gene expression,but the extent of its effects and the exact genes controlled,with the exception of the alginate-specific genes, are not known.

While alginate overproduction by mucoid strains of P. aeruginosahas an established role in pathogenesis (23), it alone cannotaccount for the inflammation and further clinical deteriorationthat correlate with the timing of the emergence of mucoid strains.One hypothesis, which takes into account the likelihood thatAlgU directs transcription of more than just the alginate biosynthesisgenes, includes the possibility of coexpression of toxic orproinflammatory products upon conversion to mucoidy. As a firststep towards testing this hypothesis, we initiated global studiesof AlgU dependent genes using the P. aeruginosa genomic sequenceas a newly available resource (58).

From our previous studies (14, 15, 38, 52) and reports by others(18, 32), a tight consensus sequence [(-35)GAACTT-N_16/17-(-10)TCtgA(invariable residues in capital letters)] for the AlgU and ${sigma}$ ^Epromoters has been derived. Using this consensus sequence, wesearched the P. aeruginosa genome for AlgU promoters and identified35 potential sites. In an experimental follow-up, we carriedout mRNA 5'-end mapping by reverse transcription and establishedAlgU dependence for a number of newly identified promoters.These analyses increased the number of characterized P. aeruginosaAlgU ( ${sigma}$ ^E) promoters from 5 to 15. Our studies also indicate apreviously unappreciated connection between the conversion tomucoidy and expression of Pseudomonas genes encoding lipoproteinswith substantial proinflammatory activity.

MATERIALS AND METHODS

Top

Materials and Methods

Bacterial strains and growth conditions. P. aeruginosa PAO381 and its mucoid derivatives PAO578I (mucA22)and PAO578II (mucA22 sup-2) have been described previously (5,22). The nonmucoid PAO6865 (algU::Tc^r) strain is derived fromPAO578II (6). For RNA isolation, strains were cultured withshaking at 37°C overnight in Luria-Bertani medium, and 1ml of the overnight culture was used to inoculate 100 ml ofLuria-Bertani medium with 0.3 M NaCl (PAO578II and PAO6865)or without the salt supplement (PAO381 and PAO578I) and grownfor 4 h at 37°C.

Tissue culture cells and conditions. Human peripheral blood monocytes were seeded onto glass coverslips and differentiated in RPMI medium with 5% human serum(Sigma, St. Louis, Mo.) to yield confluent monolayers of macrophages.

Genomic searches. The PAO1 genomic sequence was obtained from the PseudomonasGenome Project (www.pseudomonas.com) (58). Data were importedfor analysis by MacVector sequence analysis software (version6.0/7.0; Eastman Kodak Co.). A subsequence search correspondingto the AlgU consensus, GAACTT-N_16/17-TCNNA, was carried outto determine potential AlgU promoter sites in the genome. Regionsstarting 50 bp and ending 1,000 bp downstream of the putativepromoter sites were used in a global BlastX search against theNational Center for Biotechnology Information database to examinepotential open reading frames in the right orientation and position(candidate genes for regulation by the AlgU promoters). Additionally,information from the PseudoCAP annotation database was used(www.pseudomonas.com).

Primer design and DNA methods. We used 16-mer primers (nine G/C and seven A/T) generated foreach of the suspected recognition sites 500 bp upstream anddownstream of the sites. These primers were used in a PCR togenerate a 1-kb fragment from total genomic PAO1 DNA to serveas a sequencing template. A 22-mer primer was designed 60 bpdownstream of each suspected site and oriented to extend backtowards the putative promoter to generate a transcript usingreverse transcriptase in primer extension analyses as well asto sequence the promoter region using a ³³P sequencing kit (Amersham,Piscataway, N.J.). A second primer (E5 primer 2) was designedfor the promoter E5 (lptB) which had its 3" end 65 bp from theAlgU-dependent mRNA start site.

RNA isolation and primer extension analysis. RNA isolation and reverse transcription were carried out asdescribed (26). After rapid cooling in an ethanol-dry ice bath,cells were washed on ice in 50 mM Tris (pH 7.5) and lysed insodium dodecyl sulfate. Total RNA was isolated by centrifugationover a 5.7 M CsCl cushion overnight. The pellet was dried, resuspendedin diethyl pyrocarbonate-treated water, and chloroform extracted.RNA was stored in ethanol at -20°C until just prior to use.Following centrifugation, RNA was resuspended in water, anda spectrophotometry reading at an optical density of 260 wastaken to determine yield.

Primers were end labeled by polynucleotide kinase with [ ${gamma}$ ³²-P]ATPfor 1 h at 37°C. The reaction was stopped with 0.5 M EDTA,diluted with Tris-EDTA, and heat inactivated by incubation at65°C for 5 min. Labeled primers were annealed to 15 µgof total cellular RNA in hybridization buffer (0.5 M KCl, 0.24M Tris-HCl [pH 8.3]) by dissociating at 95°C for 1 min,followed immediately by annealing at 55°C for 2 min andstabilization on ice for 15 min. Primers were extended usingSuperscript II (Invitrogen, Carlsbad, Calif.) according to themanufacturer's instructions and loaded adjacent to a sequencingladder which utilized the same primer (44).

Peptide design and synthesis. Leader sequences denoting lipopeptide modification (45) wereobserved for two ORFs downstream of AlgU-dependent promoters.Two peptides, LPTA(6) Pam₃Cys-DKKEE-OH and LPTB(6) Pam₃Cys-DSQTN-OH,consisting of the palmitylated cysteine (Pam₃Cys) after thecleavage site plus five amino acids from the amino terminusof each lipoprotein, were synthesized (Bio-Synthesis, Inc.,Lewisville, Tex.). The proinflammatory synthetic bacterial lipopeptidePam₃Cys-SKKKK-OH was also synthesized (1, 25).

Cytokine assays. Cells were incubated with lipopolysaccharide (LPS) (1 µg/ml),human tumor necrosis factor alpha (12.5 ng/ml), or bacteriallipopeptides (10 µg/ml) or given no stimulation for 24h. Cell culture supernatants were removed, and any particulatematter was cleared by centrifugation. Supernatants were assayedat a 100-fold dilution for secreted interleukin-8 (IL-8) accordingto the manufacturer's instructions (Quantikine; R&D Systems).

RESULTS

Top

Results

Genomic analysis of the P. aeruginosa PAO1 chromosome for sequences representing potential AlgU ( ${sigma}$ ^E) promoters. The P. aeruginosa PAO1 genome (58) was subjected to a searchfor sequences corresponding to the AlgU ( ${sigma}$ ^E) promoter consensus(14, 18, 32, 38) using computer-assisted subsequence match analysis.The subsequence used to search the database was GAACTT-N_16/17-TCNNA,where N is any base and 16 or 17 is the spacing between the-10 and the -35 regions. This search resulted in 35 sites situatedaround the chromosome (designated A1 to A6, B1 to B9, C1 toC9, D1 to D4, and E1 to E7, based on $>=$ 1-Mb segments [A to F]used to partition the genomic sequence) that represented potentialAlgU promoters (Table 1).

View this table:
[in this window]
[in a new window]
TABLE 1. Predicted AlgU ( ${sigma}$ ^E) promoter sequences in the P. aeruginosa genome^a

Regions including 50 bp upstream and 1,000 bp downstream ofthe potential AlgU promoters were examined for the presenceof previously annotated genes or subjected to global searchesto detect potential open reading frames downstream of the candidatepromoter sites (Table 2). In addition to the algD, algU, andrpoH loci, previously shown to have AlgU-dependent promoters(13, 39, 50), several of the putative promoter sites were foundupstream of genes whose products might aid bacterial adaptationto changing environmental conditions (Table 2). (i) oprF encodesporin F, the major outer membrane protein of P. aeruginosa (61).The levels of OprF have recently been shown to increase duringperiods of stress in an AlgU (AlgT)-dependent manner (34). (ii)The osmC gene encodes an E. coli outer membrane protein thatis transcriptionally induced during hyperosmotic stress (9).(iii) The lptB gene encodes a predicted peptidyl-proly cis/transisomerase (PPIase), most likely functioning in protein folding.(iv) The slyB gene product shows homology to the proposed porinSlyB of Salmonella enterica serovar Typhimurium (33). (v) ThebetT gene is in the bet locus, involved in synthesis of theosmoprotectant glycine betaine from choline (19). (vi) The dksAgene encodes a suppressor that offsets mutations in heat shockgenes such as dnaK, dnaJ, and grpE (2). (vii) The phuR geneencodes a hemin receptor and is controlled by the iron-sensitiveregulator Fur (40). The remaining 24 sites (with the exceptionof TalB [56], a transaldolase that could be involved in generatingthe alginate precursor fructose-6-phosphate from sedoheptulose-7-phosphateand glyceraldehyde-3-phosphate [64]) had no annotated P. aeruginosagenes within 1-kb regions downstream of the putative AlgU promoters.

View this table:
[in this window]
[in a new window]
TABLE 2. Genes downstream of the predicted AlgU ( ${sigma}$ ^E) promoters^a

Mapping of the AlgU promoter upstream of the oprF gene. Recently, porin F levels have been shown to increase in mucoid,mucA mutant P. aeruginosa strains (35). This finding supportedour observation that site A2 was located upstream of the oprFgene, suggesting that this could indeed be an AlgU-dependentpromoter. To test this possibility, we carried out mapping ofthe mRNA 5' end by reverse transcription. A primer was designeddownstream of the predicted mRNA start site. The same primeralso served to generate the corresponding sequencing ladder.Primer extension analysis of the oprF promoter (Fig. 1A) revealedthe presence of a transcript initiating 5 nucleotides downstreamfrom the -10 region of the predicted AlgU-dependent promoter.The oprF transcript was dependent on AlgU, as it was lost inthe algU knockout strain (Fig. 1A). As an additional control,we examined the AlgU-independent promoter of oprF, with themRNA 5' end at -57 from the translational start (see Fig. 1D).The band corresponding to this promoter was not absent, andits intensity was only 10% lower in the algU::Tc^r samples (Fig.1B). Identical relationships (presence of the transcript inthe algU⁺ strain and absence of the corresponding band in algU::Tc^rmutants) were observed with the previously mapped algD promoter(Fig. 1C). These analyses demonstrate that an AlgU-dependentpromoter exists upstream of oprF and most likely contributesto oprF expression.

View larger version (72K):
[in this window]
[in a new window]
FIG. 1. Mapping of the AlgU-dependent oprF promoter. (A) Primer extension mapping of the mRNA 5' end corresponding to the AlgU promoter of oprF. Total RNA was isolated from the algU⁺ mucA22 mucoid strain PAO578II and from its nonmucoid algU knockout derivative (algU::Tc^r; PAO6865) grown under conditions promoting the mucoid phenotype (Materials and Methods). Primer extension products were run adjacent to the sequencing ladder generated with the same primer used for reverse transcription. Vertical bars indicate the -35/-10 regions of the AlgU-dependent promoter of oprF. The nucleotide corresponding to the mRNA 5" end is designated P_AlgU with an asterisk. (B) Control for panel A, showing the presence of the oprF AlgU-independent promoter at -57 upstream of the initiation codon. In this particular gel, the intensity of the band in the algU lane was 17% stronger than in the algU::Tc^r mutant. (C) Primer extension of the previously mapped AlgU-dependent promoter P_D of algD, included here as a control (13). All procedures and designations as in panel A. (D) Maps showing the relative positions of the AlgU-dependent promoters upstream of the algD and oprF genes. A previously mapped oprF promoter, corresponding to the mRNA start site at -40, is dependent on the alternative sigma factor SigX (11). The promoter at -57 is regulated by an unknown sigma factor (17). All numbering is relative to the initiation codon.

Transcriptional analysis of the predicted AlgU ( ${sigma}$ ^E) promoters. Having demonstrated that one of the putative AlgU promotersis indeed an AlgU-dependent transcriptionally active site, wetested all other potential AlgU promoters predicted from ourgenomic search. Nine additional AlgU promoters (for a totalof 10 new AlgU promoters, including the oprF AlgU promoter)were identified, and the corresponding mRNA 5' ends were mapped(Fig. 2) using the same approach as for oprF AlgU promoter mapping.In all 10 cases, the band corresponding to the transcriptionalstart site was downstream of the predicted -35/-10 AlgU promoterregions, with the mRNA 5" end coinciding with a nucleotide located4 to 6 bp downstream of the canonical -10 AlgU promoter sequence(Fig. 2 and Table 3). In all cases, the transcript was completelyabsent in the algU mutant strain, with one exception (Fig. 2,E5).

View larger version (98K):
[in this window]
[in a new window]
FIG. 2. Primer extension analysis of the newly identified AlgU-dependent promoters. Promoter mapping was carried out as described for Fig. 1. Designations are given as in Tables 1 and 2. The mRNA start site(s) has been mapped to the nucleotides shown in Table 3. The start site for E5 (primer 1) was also mapped using an additional oligonucleotide (E5 primer 2), to demonstrate that the lower band (AlgU independent) was not an artifact. The sequencing ladders for D2 and D3 are not shown.

View this table:
[in this window]
[in a new window]
TABLE 3. Mapped AlgU ( ${sigma}$ ^E) promoters in P. aeruginosa and enteric bacteria

In the case of E5, the strong AlgU-dependent top band disappeared,but a second, downstream band, which was weak in the algU⁺ sample,intensified in the algU::Tc^r mutant (Fig. 2, E5, primer 1).The second band was not a reverse transcription artifact, asthe same pattern was observed using a different primer (Fig.2, E5, primer 2). Since P. aeruginosa has close to 20 sigmafactors from the same class as AlgU, we interpret this observationas a possible recognition of the AlgU promoter E5 by anotherAlgU homolog. The presence of this slightly intensified secondband corresponding to this transcript also demonstrates thatthe amounts of RNA loaded were comparable in the algU⁺ and algU::Tc^rlanes in all samples.

For the AlgU promoter sites positively identified by reversetranscription in this work but without previously acknowledgedor annotated genes downstream of the promoter sites, in fourof five cases we could identify an open reading frame downstreamof the mRNA 5" end. The corresponding putative genes, designatedhere as AlgU sigmulon members, have been named asmA6, asmB2,asmC1, asmD2, and asmD3 (Table 2). The distances of these potentialgenes from the transcriptional start sites ranged from 12 to254 nucleotides, which is well within the range of the typicalAlgU promoter-coding sequence distance (Table 2).

The increased expression of the previously characterized AlgU-dependentalginate-specific genes is dependent on inactivation of mucA,a negative regulator of AlgU (37). We next examined how manyof the newly identified promoters were affected by the mucA22mutation. For this analysis, the nonmucoid mucA⁺ strain PAO381and its constitutively mucoid mucA22 derivative (strain PAO578I)were used. The majority of the promoters showed elevated expressionlevels (ranging from 1.2- to 8.5-fold) in the mucA mutant relativeto the mucA⁺ parent (Fig. 3). These findings indicate that themajority of the AlgU promoters are activated upon the loss ofMucA inhibitory activity on AlgU in mucA22 mutants. Nevertheless,basal expression levels were seen in mucA⁺ strains with mostof the AlgU-dependent promoters.

View larger version (35K):
[in this window]
[in a new window]
FIG. 3. Inactivation of MucA, a negative regulator of AlgU, enhances the expression of the majority of the newly identified AlgU-dependent promoters. Total RNA was isolated from algU⁺ mucA⁺ strain PAO381 (lane 1) and its mucoid mucA22 derivative, PAO578I (lane 2), and subjected to primer extension analysis as described for Fig. 1 and 2. Fold induction (determined by densitometry) relative to that in the mucA⁺ background is indicated to the right of each set.

One of the initially detected transcripts, B2 (Fig. 2), didnot show a signal (Fig. 3) in the parental mucA⁺ (PAO381) strainand showed only a weak signal in the mucA22 mutant strain (PAO578I).Since the strains PAO578II (used in the initial screen) andPAO578I differ in their environmental requirements for the expressionof the mucoid phenotype due to the presence of an additionalsup-2 mutation in PAO578II (51), we compared all three strains(PAO381, PAO578I, and PAO578II) grown under conditions promotingmaximal alginate production (Materials and Methods). As shownin Fig. 4, the B2 promoter signal was reproducibly detectedin PAO578II. These observations confirm the notion that theexpression of at least some of the AlgU-dependent promotersdepends on additional factors even in mucA mutants (51). Consequently,we cannot exclude the possibility that at least some of the25 putative sites showing negative transcriptional results representAlgU-dependent promoters but remain silent unless additionalregulators are activated or environmental requirements are met.

View larger version (48K):
[in this window]
[in a new window]
FIG. 4. Expression of some AlgU-dependent promoters is dependent on strain type and environmental conditions. The B2 promoter was tested for activity by reverse transcription in the nonmucoid parent strain PAO381 (algU⁺ mucA⁺), its constitutively mucoid derivative PAO578I (algU⁺ mucA22), and strain PAO578II (algU⁺ mucA22 sup-2; derived from PAO578I), which requires additional environmental stimuli for expression of the mucoid phenotype (see Materials and Methods). A6 is included for comparison. Fold induction levels relative to PAO381 (mucA⁺) are indicated below the lanes. ND, not detectable.

AlgU controls lipoprotein genes: implications for inflammatory processes in CF. A destructive, hyperinflammatory response is one of the hallmarksof CF (3, 47). Detailed analyses of open reading frames downstreamof the newly identified AlgU-directed promoters revealed that2 of the 10 characterized AlgU-dependent genes encoded productswith putative lipoprotein leader sequences (Fig. 5A). Giventhe recent implication of bacterial lipoproteins in Toll-likereceptor-mediated inflammation (1, 10), we examined whetherlipoproteins controlled by AlgU might have proinflammatory activity.

View larger version (35K):
[in this window]
[in a new window]
FIG. 5. Lipopeptides corresponding to the amino termini of the newly identified AlgU-dependent lipoprotein-encoding genes induce IL-8 secretion by human macrophages. (A) Leader peptide sequences corresponding to the predicted products of lptA and lptB. Asterisks indicate residues required for processing. The underlined portion demarcates the hydrophobic residues within the leader peptide. The leader peptide is cleaved (vertical line) just prior to the cysteine residue which is modified by palmitylation (45). (B) Sequence of synthesized lipopeptides corresponding to the first six amino acids of mature LptA [LPTA(6)] and LptB [LPTB(6)]. sBLP, standard synthetic bacterial lipopeptide (1, 25). (C) Human macrophages derived from peripheral blood monocytes were incubated for 24 h with medium alone (bar labeled None), LPS (1 µg/ml), or bacterial lipopeptides (10 µg/ml), as indicated. Supernatants were assayed for IL-8 production by enzyme-linked immunosorbent assay. _*_*, P < 0.001 compared to uninduced control (None).

Figure 5A displays the amino-terminal sequence of the predictedP. aeruginosa lptA and lptB gene products, highlighting (i)the residues critical for lipopeptide sequence modification,(ii) the leader sequences containing typical cleavage sites,and (iii) the lipid-modified cysteine residues. To address thepossibility that LptA and LptB may have proinflammatory activity,we synthesized palmitylated lipopeptides corresponding to thefirst 6 amino acids of the predicted mature LptA and LptB (Fig.5B). The synthesized lipopeptides were assayed for their abilityto induce IL-8 secretion by primary human macrophages derivedfrom peripheral blood monocytes (Fig. 5C). Upon stimulationwith lipopeptides LPTA(6) and LPTB(6), used in standard concentrationsusually applied in such experiments (1, 10) albeit exceedingthat of LPS, significant amounts of IL-8 were detected (P <0.001). The peptides corresponding to the LptA and LptB N terminicaused IL-8 secretion equal in potency to that of LPS and comparableto the standard peptide simulating lipoprotein action (1, 25).These findings suggest that at least some of the genes coexpressedwith the alginate system during conversion to mucoidy may playa role in inflammatory processes in CF.

DISCUSSION

Top

Discussion

Using the newly available P. aeruginosa genomic sequence informationand conventional molecular biology techniques, we have identified10 new P. aeruginosa promoters dependent on the alternativesigma factor AlgU. Of the 35 candidate sites containing theAlgU (P. aeruginosa ${sigma}$ ^E) consensus sequence, one-third were confirmedby mRNA 5"-end mapping and additionally demonstrated to dependon AlgU for transcription. The present work has tripled thenumber of known and mapped P. aeruginosa AlgU promoters fromthe previously defined 5 to a current total of 15. The confirmedpromoters did firm up the AlgU ( ${sigma}$ ^E) promoter consensus further,as shown in Table 3.

In the process of comparing AlgU-controlled genes and correspondinghomologs in S. enterica serovar Typhimurium and E. coli, wealso recognized a previously unappreciated ${sigma}$ ^E promoter in frontof the Salmonella slyB gene encoding an outer membrane porin(33), which coincides with the reported slyB mRNA 5" end (Table3). Thus, the current study also helped find an additional ${sigma}$ ^Epromoter in another organism.

Follow-up studies presented here have uncovered a previouslyunknown link of conversion to the mucoid phenotype in P. aeruginosawith the induction of potential proinflammatory factors. Weidentified two genes, lptA and lptB, that encode putative lipoproteins,as being controlled by AlgU. The promoter activity of lptA andlptB is induced in mucA mutant cells. The synthetic lipopeptidescorresponding to the N termini of the mature, processed LptAand LptB caused IL-8 production in primary human macrophagesderived from peripheral blood monocytes. This observation expandsthe potential impact of conversion to mucoidy in P. aeruginosaon pathogenesis issues. This phenomenon has been ascribed inthe past primarily, if not exclusively, to the production ofalginate, with the following proposed roles for the mucoid coating:(i) overall inhibition of P. aeruginosa clearance from the lungs(8, 62), possibly related to the proposed inhibition of opsonic(54) and nonopsonic phagocytosis (30, 41); (ii) potential radicalscavenging function of alginate in reducing the impact of reactiveoxygen intermediates and hypochlorite generated by immune defensecells (31, 55); and (iii) inhibition of leukocyte function,including chemotaxis, complement activation, and oxidative burst(26, 43, 57). The new observations presented here provide anexpanded view of the conversion to mucoidy in P. aeruginosaby linking it with the inflammatory response in the host. Coinductionof lipoproteins that have proinflammatory potential may be amechanism by which conversion to mucoidy could contribute tothe excessive inflammatory response typically observed in theinfected CF lung.

Not all of the predicted P. aeruginosa AlgU promoters identifiedwith the sequence consensus search were experimentally confirmed.The presence of a band for the promoter designated B2 undervaried conditions raises the possibility that some of the potentialAlgU-dependent promoters that remain unconfirmed may be functionalonly under optimized conditions. In the case of phuR, the predictedAlgU promoter overlaps the two Fur boxes, a region shown tobe under tight repression by the Fur protein (40). Conditionsof low iron could relieve the negative regulation by Fur andallow promoter recognition by AlgU.

In addition to the potential increase in actual promoter numberresulting from the possibility of specific expression requirements,the number of AlgU promoters in P. aeruginosa could be evengreater, considering some known and well-defined AlgU promotersthat depart from the consensus by one residue. Notable examplesare the promoters for algR, which has a substitution in the-35 region, and the P1 promoter for algU, which has a substitutionin the final position of the -10 region. Neither of these promoterswas detected by a subsequence search with the original criteria.Allowing for additional substitutions individually (e.g., -35GCACTT [algR] and -10 TCTAT [algU P1]; underlined residues indicatedepartures from the consensus) revealed 63 additional siteson the P. aeruginosa chromosome.

LptA and LptB were examined in this work in the context of theirlipoprotein structure, in line with our original idea that proinflammatoryfactors may be coinduced with alginate production in mucA mutantmucoid cells. However, at least one of these putative gene productsmost likely plays a role in protein folding. LptB shows stronghomology to FKBP-like peptidyl-prolyl cis/trans isomerases,which catalyze the interconversion of the proline peptide bondsbetween the cis and trans isomers (46). This proposed functionof LptB is in keeping with the nature of the majority of genescontrolled by AlgU (i.e., stress response or protein folding)identified here and elsewhere (14, 18, 32, 38). The corresponding ${sigma}$ ^E-controlled genes in enterics include another peptidyl-prolylisomerase (fkpA) (12).

The role of LptA is more obscure because it encodes a shortpolypeptide of 78 residues and has no homology to any knownsequence. It is possible that LptA participates in some processessimilar to what seems to be the common theme for AlgU-regulatedgenes, i.e., general defense against stress. It may also bethat LptA (or both LptA and LptB) is more narrowly associatedwith alginate production in the extracytoplasmic spaces, asit is coinduced with alginate production. Future inactivationof lptA and lptB genes will address this possibility.

Unlike the case of LptB, there are very few a priori clues regardingthe possible function of LptA, apart form the potential rolein alginate-specific processes discussed above. At present,we do not exclude the possibility that LptA, and perhaps someother molecules of short amino acid stretches with perfect lipoproteinprocessing and modification signals, is made as a bacterialproduct that acts primarily to cause uncontrolled inflammatoryprocesses in the infected host. Should this be the case, wepropose the possibility that such products could be a novelclass of bacterial toxins directed to stimulate host patternrecognition receptors.

Our current data are consistent with the idea that LptA andLptB can act as proinflammatory agents. Based on the data presentedhere, the activation of the AlgU sigmulon may have broader implicationsfor the host with CF in that at least some factors coinducedwith alginate overproduction may act to exacerbate inflammationand lung disease in CF.

ACKNOWLEDGMENTS

We thank Michal H. Mudd for technical expertise and help withthis study and Lisa Tatterson and Jens Poschet for discussions.

This work was supported by NIH grant AI31139.

FOOTNOTES

* Corresponding author. Mailing address: Department of Molecular Genetics and Microbiology, University of New Mexico Health Sciences Center, 915 Camino de Salud, NE, Albuquerque, NM 87131. Phone: (505) 272-0291. Fax: (505) 272-6029. E-mail: vderetic{at}salud.unm.edu.

Present address: Harvard Medical School, Boston, MA 02115.

REFERENCES

Top