differential interaction of the transcription factor prfa and the prfa-activating factor (paf) of...

Differential interaction of the transcription factor PrfAand the PrfA-activating factor (Paf) of Listeriamonocytogenes with target sequences

Carmen Dickneite, † Regine Bo ckmann, † AndreaSpory, Werner Goebel * and Zeljka SokolovicTheodor-Boveri-Institut fur Biowissenschaften derUniversitat Wurzburg (Lehrstuhl fur Mikrobiologie), AmHubland, 97074 Wurzburg, Germany.

Summary

The interaction of the purified PrfA transcription fac-tor with the regulatory sequences located upstreamof the PrfA-dependent listeriolysin ( hly ) and inter-nalin ( inlA ) genes was studied in the presence andin the absence of Paf ( PrfA- activating factor)-contain-ing extracts. It is shown that PrfA protein is able to bind,independently of additional factors, to a 109 bp DNAfragment including the entire hly promoter sequencewith the anticipated PrfA binding site (‘PrfA-box’).PrfA alone, but not in combination with Paf, can alsobind to a shorter target sequence of 28 bp comprisingessentially the PrfA-box of the hly promoter. Theaddition of a Paf-containing extract does not lead tosignificant protein binding to these two hly targetsequences in the absence of PrfA but converts thecomplex (CIII) consisting of PrfA and the 109 bp hlyDNA fragment to a slower migrating PrfA–Paf–DNAcomplex (CI). Incubation of cell-free extracts of wild-type Listeria monocytogenes with the 109 bp DNAfragment leads to the formation of CI. The additionof polyclonal PrfA antibodies causes a supershift ofCIII. Purified PrfA and PrfA–Paf also bind to a DNAfragment containing the PrfA-dependent promoterP2 of inlA , albeit at a lower rate when compared withthe corresponding hly sequence. In contrast to thehly target DNA, the inlA promoter sequence efficientlybinds Paf alone, and this Paf binding reduces that ofPrfA and PrfA–Paf to the inlA target DNA. DNase Ifootprint experiments show that purified PrfA pro-tects sequences of dyad symmetry previously pro-posed as PrfA binding sites in the hly and in theinlA promoter regions.

Introduction

Listeria monocytogenes is a Gram-positive, facultative,intracellular bacterium that causes a severe foodborneinfection (listeriosis) of humans and animals with symp-toms such as septicaemia, encephalomeningitis and abor-tion (Gellin and Broome, 1989). All known virulence genesare located on the chromosome, and six of these genesare clustered between the genes for lactate dehydrogen-ase (ldh) and phosphoribosyl synthetase (prs) (Vazquez-Boland et al., 1992; Gouin et al., 1994; Kreft et al., 1995).This gene cluster is present in all clinical L. monocyto-genes isolates tested. The products determined by thesix genes (prfA, plcA, hly, mpl, actA and plcB) includetwo type C phospholipases (PlcA and PlcB), listeriolysin(Hly/LLO, an SH-activated cytolysin, which is responsiblefor the haemolytic phenotype of L. monocytogenes), ametalloprotease (Mpl), the actin polymerization proteinActA and the transcription activator PrfA (for reviews, seePortnoy et al., 1992; Sheehan et al., 1994; Kuhn andGoebel, 1995). PrfA is not only essential for the transcrip-tion of these clustered virulence factors but also activatesthe expression of other genes, including the inlA–inlBoperon (Leimeister-Wachter et al., 1990; Mengaud et al.,1991; Dramsi et al., 1993) and the recently describedinlC gene (Engelbrecht et al., 1996; Lingnau et al., 1996)located outside of the gene cluster. While InlA and InlBare involved in the invasion of normally non-phagocytichost cells, such as epithelial cells (Gaillard et al., 1991)and hepatocytes (Dramsi et al., 1995), the function ofInlC remains to be elucidated.

PrfA is a member of the CRP/FNR family of prokaryotictranscription factors, which have mainly been isolated andcharacterized from Gram-negative bacteria. PrfA sharessubstantial sequence and structural similarity with theextensively studied cAMP binding protein CRP of Escheri-chia coli (Lampidis et al., 1994; Kreft et al., 1995). Recentstudies have shown that the DNA binding domains of CRPand PrfA, typical helix–turn–helix structures, are function-ally highly related (Sheehan et al., 1996). Major differencesbetween these two transcription activators are (i) an extraC-terminal domain in PrfA possessing a putative leucinezipper motif (Lampidis et al., 1994), which is importantfor the activator function of PrfA (A. Bubert, personal com-munication); and (ii) the lack of the cAMP binding domain

Molecular Microbiology (1998) 27(5), 915–928

Q 1998 Blackwell Science Ltd

Received 11 August, 1997; revised 1 December, 1997; accepted 5December, 1997. †These authors contributed equally to this paper.*For correspondence. Tel. (931) 8884401; Fax (931) 8884402.

m

present in CRP. There is a conserved sequence of dyadsymmetry (comprising 14 bp or less) in front of all knownPrfA-dependent genes and operons (Mengaud et al.,1989; Domann et al., 1991; Vazquez-Boland et al., 1992;Dramsi et al., 1993), centred at around position ¹40from the transcriptional start sites. It has been proposedthat this sequence represents the binding site for PrfA(‘PrfA-box’), and recent studies (Freitag et al., 1993; Shee-han et al., 1996; Bockmann et al., 1996) indeed show thatthis site is essential for the binding of PrfA to DNA frag-ments derived from the hly and actA promoters.

Most L. monocytogenes wild-type strains (type 1 strains)express most PrfA-dependent virulence genes weaklywhen growing at 378C in rich culture medium, such asbrain–heart infusion (BHI) medium (Ripio et al., 1996). Ashift of such L. monocytogenes cultures into minimalessential medium (MEM) induces transcription of thesevirulence genes, with the exception of inlA–inlB, the tran-scription of which is rather reduced under these conditions(Sokolovic et al., 1993; Bohne et al., 1994; 1996). A similareffect has been observed in charcoal-treated BHI medium(Ripio et al., 1996), while the presence of cellobiose ratherinhibits expression of hly and possibly other PrfA-depen-dent genes (Park and Kroll, 1993). Other environmentalparameters, such as elevated temperature, have alsobeen shown to induce transcription of some PrfA-regu-lated genes in these strains (Leimeister-Wachter et al.,1992; Ripio et al., 1997).

There are also L. monocytogenes strains (type 2 strains)that seem to express prfA and the PrfA-regulated genesconstitutively at a high level. In these strains, PrfA-depen-dent gene expression seems to be independent of theparameters that induce transcription of the PrfA-regulatedgenes in type 1 strains. It has been shown recently thattype 2 strains carry a defined mutation in the prfA gene,which seems to render the mutant PrfA independent ofauxiliary factors. Interestingly, this mutation is located ina position of PrfA that, in the homologous CRP, leads toa cAMP-independent, transcriptionally active conforma-tion (Ripio et al., 1997).

We and others (Freitag et al., 1993; Bockmann et al.,1996; Sheehan et al., 1996) have recently used purifiedPrfA to investigate directly the binding of this transcriptionactivator to DNA probes containing the ‘PrfA-box’. Sur-prisingly, no proper binding of purified PrfA was detectedby us, although whole-cell extracts of L. monocytogenes

exhibited PrfA-specific binding to these DNA probes (Bock-mann et al., 1996; Sheehan et al., 1996).

A factor termed Paf (for PrfA-activating factor) wasidentified in PrfA-free extracts (obtained from a prfA dele-tion mutant of L. monocytogenes EGD), which seems tointeract with PrfA to form a stable complex (CI) with thehly promoter sequence (Bockmann et al., 1996). In extractsof L. monocytogenes EGD, a type 1 strain exhibiting strin-gent PrfA-mediated regulation of the virulence genes(Bohne et al., 1996; Ripio et al., 1997), the Paf–PrfA inter-action appears to depend on the iron concentration in themedium (Bockmann et al., 1996). A second specific butPrfA-independent complex (CII) was often detected whenthe hly target DNA was used. CII exhibited the same bind-ing properties as the complex attributed to Paf binding tothe target DNA.

In this communication, we extend these observations,showing that purified PrfA protein alone can bind to andspecifically protect the symmetrical PrfA-boxes of the hlypromoter and the PrfA-dependent inlA promoter P2 andthat Paf influences PrfA binding to these two PrfA-boxesdifferently. The implication of this finding for the differentialregulation of virulence genes in L. monocytogenes isdiscussed.

Results

Binding of purified PrfA protein to hly promotersequences containing the PrfA-box

Recently, we reported (Bockmann et al., 1996) that puri-fied PrfA protein with a [His]6-tag at the N-terminus(termed P2-7) requires the addition of (PrfA-free) listerialprotein extract for specific binding to the hly promotersequence. This recombinant PrfA protein was used in allsubsequent experiments as it was not possible to removethe [His]6-tag efficiently by enterokinase cleavage withoutsignificant loss of activity. The use of the N-terminal [His]6-tagged PrfA protein for the binding studies appears to bevalid for the following reasons: (i) complementation of theL. monocytogenes prfA deletion mutant (prfA¹) with mul-tiple gene copies of either the wild-type prfA gene fromstrain EGD (prfA¹ ×prfA, wild-type prfA carried on theplasmid pERL3-501) or the prfA gene with the 6× Hisaffinity tag sequences (prfA¹ ×h-prfA, also carried onpERL3-501) resulted in the same haemolytic activities

Q 1998 Blackwell Science Ltd, Molecular Microbiology, 27, 915–928

Table 1. Haemolytic activities in CHUa ofwhole-cell suspensions and supernatants of L.monocytogenes strains after shift into MEMmedium.

Strain prfA¹ EGD prfA¹ ×prfA prfA¹ ×h-prfA

Whole-cell suspension 2 32 128 128Supernatant NDb 1 8 8

a. Activities are given in complete haemolytic activities (CHU) (Bohne et al., 1996).b. ND, not detectable.

916 C. Dickneite et al.

(Table 1); (ii) the same amount of PrfA protein was detectedby immunoblotting in both strains (Fig. 1). These data indi-cate that the positively charged N-terminal [His]6-tag doesnot affect the function of the PrfA protein.

The hly promoter is PrfA dependent and controls theexpression of the listeriolysin (hly) gene of L. monocyto-genes (Leimeister-Wachter et al., 1990). The putativecomponent supporting PrfA-mediated binding was namedPaf (for PrfA-activating factor; Bockmann et al., 1996).

With a modified binding protocol (see Experimental pro-cedures), which essentially used higher concentrations ofthe synthetic co-polymer dI-dC, we now obtained specificbinding of the purified PrfA protein to DNA probes derivedfrom the upstream regulatory sequence of the hly gene(Fig. 2A). Both hly DNA probes contained the presump-tive PrfA binding site (‘PrfA-box’), a 14 bp sequence ofdyad symmetry (Mengaud et al., 1989).

The following observations were made with the two hlypromoter probes. (i) With a 109 bp fragment carrying theentire hly promoter in addition to this PrfA-box, a specificPrfA–DNA complex (CIII) was predominantly formed(Fig. 3A). The CIII complex showed a significantly fastermobility in the polyacrylamide gel than the PrfA-specificcomplex CI, which was observed using the same 109 bptarget DNA and a cell-free protein extract preparedfrom the (prfA-positive) L. monocytogenes strain EGD

(Fig. 3A, lane 1). The formation of the CI complex has pre-viously been shown to result from binding of PrfA togetherwith Paf to this target DNA (Bockmann et al., 1996). (ii)With a shorter 28 bp DNA fragment comprising essentiallythe PrfA-box of the hly promoter sequence, binding ofpurified PrfA was also observed (Fig. 3B). PrfA bindingto both hly probes was specific, as shown by competitionwith increasing amounts of the unlabelled hly probe andnon-specific herring sperm DNA (Fig. 3A and B). Basedon these competition data, binding of PrfA seems tooccur at a similar strength to both target DNAs.

Addition of polyclonal anti-PrfA antibodies to the bindingassay containing the 28 bp probe and purified PrfA causeda supershift of the CIII complex (Fig. 3B). The super-shifted complex is less intense in this case than CIII, sug-gesting that the 28 bp DNA probe cannot bind the largeantibody–PrfA complex efficiently. With the 109 bp frag-ment, a similar supershift was observed (data not shown).

PrfA-specific binding to the 28 bp fragment was notdetected with extracts from L. monocytogenes EGD that,in contrast, generated the CI complex predominantly withthe 109 bp fragment as target DNA (see below and com-pare Fig. 3A, lane 1, and Fig. 3B, lane 1). PrfA-specificbinding (resulting in the formation of the CIII complex)was observed only after the addition of purified PrfA tothis binding assay (Fig. 3B, lane 3). It is obvious, however,that formation of CIII by the same amount of purified PrfAwas significantly more efficient in the absence of theextracts than in their presence (compare lanes 4 and 3 inFig. 3B), further suggesting that PrfA forms complex(es)with components in these extracts, which then do notbind to the 28 bp hly probe. It is unlikely that this phenom-enon is caused by proteolytic degradation of PrfA, as dataobtained by immunoblots showed that the amount of PrfAprotein remained unchanged after the addition of extracts(data not shown).

Shift of the CIII complex to CI was demonstrated directlyby the addition of increasing amounts of Paf-containingextracts (prepared from the prfA deletion mutant grownin either BHI or MEM) to the binding assay after preincuba-tion of the 109 bp hly probe with purified PrfA (Fig. 4A).Note that addition of the MEM extract yielded about fivetimes more CI complex (at low extract concentration)than the BHI extract. This is in accordance with the pre-viously observed higher Paf binding activity in MEMextracts compared with that in BHI extracts (Bockmannet al., 1996).

Starting with the CIII complex consisting of PrfA boundto the 28 bp hly probe, the addition of increasing amountsof the same Paf-containing extracts slightly reduced theamount of CIII complex (Fig. 4B). But a slower migratingPrfA–Paf–DNA complex corresponding to CI was not gen-erated, suggesting that this shorter DNA probe, althoughpossessing the entire PrfA-box, lacks sequences necessary


Fig. 1. Immunoblotting with polyclonal anti-PrfA antibodies of thetotal cellular proteins from L. monocytogenes strains EGD, theisogenic prfA deletion mutant (prfA¹) and the prfA deletion mutantcomplemented with either the wild-type prfA (prfA¹ ×prfA) or theprfA gene with the 6× His affinity tag sequences (prfA¹ ×h-prfA)after shift of the cultures into MEM. Positions of marker proteins inkDa are indicated on the lefthand side of the figure. A lane with100 ng of the purified recombinant PrfA protein is included as acontrol (PrfA).

Differential interaction of the transcription factor PrfA 917

for the binding of PrfA–Paf and hence for the formation ofstable CI complex.

PrfA binds to DNA probes containing the PrfA-box ofthe upstream regulatory sequence of the internalin( inlA) gene

Transcription of the internalin gene (inlA) of L. monocyto-genes by the P2 promoter is also under the control of

PrfA, but the mode of regulation seems to be differentfrom that of the hly promoter. In particular, PrfA-depen-dent transcription of inlA is rather reduced in MEM, whilethat of hly is highly induced (Bohne et al., 1996).

Several DNA probes from the upstream regulatorysequence of inlA were used for studying the binding ofPrfA to this target sequence (Fig. 1B). Two probes of53 bp and 101 bp contained the entire P2 promoter, includ-ing the putative but rather degenerated PrfA-box (Dramsi


Fig. 2. Characteristics of the different promoter DNA probes used.A. The PrfA-box-containing region, the ¹10 box and the transcriptional start site (S) of the hly gene are shown as open boxes. Position of the109 bp and the 28 bp hly probes are indicated as grey bars. The palindromic sequence of the PrfA-box is shown underneath.B. The inlA probes used. The anticipated PrfA binding site, the ¹10 box and the transcriptional start point (S2) of the PrfA-dependent inlA P2are shown as open boxes. Regions of the PrfA-independent inlA promoters, P1 and P3, and the corresponding start sites (S1 and S3) areshown in light and dark grey respectively. The palindromic sequence of the PrfA-box of inlA P2 is shown underneath.


et al., 1993). The 101 bp probe differed from the 53 bpprobe essentially by 27 additional basepairs 58 upstreamfrom this PrfA-box and 21 additional basepairs down-stream of the ¹10 box. A third 41 bp probe was includedin this study, which lacked the PrfA-box but containedthe ¹10 region of the P2 promoter and the entire sequenceof the PrfA-independent P1 promoter. The first two probes,but not the third probe, showed binding of purified PrfA.Binding to the shorter 53 bp probe required at least 1.5-fold the amount of PrfA than binding to the 101 bp frag-ment (data not shown). For most of the subsequentlydescribed experiments, we therefore used the 101 bpinlA probe. PrfA binding to this probe was leading to aCIII-analogous complex (Fig. 5). Specific competition wasobserved, but this required the presence of at least fivefoldlarger amounts of unlabelled specific competitor than wasneeded for the same specific competition of PrfA whenbound to the hly probe (compare Fig. 3A and Fig. 5). Atthis high competitor concentration, reduction in PrfA bind-ing to the 101 bp inlA probe was even observed with the

non-specific herring sperm DNA, suggesting that bindingof PrfA to the inlA target sequence is weaker than bindingof PrfA to the hly target sequence. The CIII complex con-sisting of PrfA and the 101 bp inlA probe exhibits a super-shift in the presence of polyclonal anti-PrfA antibodies,proving the specificity of this complex (Fig. 5).

Paf competes with PrfA for binding to the inlApromoter

Next, we tested the binding of PrfA-containing and PrfA-free extracts from L. monocytogenes EGD and the iso-genic prfA deletion mutant, respectively, to the 101 bpinlA probe. Extracts were prepared from bacteria grown ineither BHI, MEM or MEM supplemented with 100 nM Fe3þ

citrate, as described previously (Bockmann et al., 1996).Surprisingly, complex formation was already observed withthe PrfA-free BHI extract, which was further enhancedwith MEM extract and slightly reduced again with theiron-supplemented MEM extract (Fig. 6A), thus exhibiting


Fig. 3. Specific binding of purified PrfA to the two DNA probes derived from the hly promoter both of which contain the PrfA-box. Competitionof PrfA binding with a 50- to 200-fold molar excess of cold specific hly promoter fragment and non-specific herring sperm DNA is shown.Binding with a PrfA-containing cell-free extract of L. monocytogenes EGD as a control is shown in the first lane. Arrowheads indicate thepositions of the different PrfA–DNA complexes (CI and CIII).A. PrfA-specific binding to the 109 bp hly promoter fragment.B. PrfA-specific binding to the 28 bp hly promoter fragment. Binding with an extract of the prfA deletion mutant (prfA¹) from strain EGDwithout (C) and with 200 ng of purified PrfA (þPrfA) is shown in lanes 2 and 3. Supershift caused by the addition of anti-PrfA antibodies (Ab)but not of preimmune serum (P) to the binding assay is shown in the right lanes.


the same binding properties that have previously beenshown to be caused by Paf. Formation of the same com-plex was observed when a Paf-containing fraction of theMEM extract separated by sucrose gradient centrifuga-tion, as described previously (Bockmann et al., 1996), wasused in the bandshift experiment (Fig. 7), suggesting thatthis complex (CII) is generated by Paf binding to the inlADNA probe. In contrast, with the 109 bp hly promoter frag-ment, Paf activity in the same fractions could only bedetected after the addition of purified PrfA, leading tothe formation of CI in the binding assay. Using the inlAP2 promoter fragment, a slower migrating band wasobtained in addition to CII after incubation of purifiedPrfA with extracts of the prfA deletion mutant (Fig. 6).The position of this band suggested the formation of aPrfA–Paf–DNA complex (CI). This assumption was sup-ported by the increased occurrence of this CI band whenMEM extracts were used compared with BHI extracts.With the iron-containing MEM extract, the amount of CIcomplex dropped to the level obtained with the BHI extract.Both complexes, CI and CII, were competed out by increas-ing amounts of the unlabelled inlA fragment but not withunspecific DNA (Fig. 6), suggesting that not only bindingof PrfA–Paf but also of Paf alone to this inlA probe isspecific. Note that CIII complex formation is enhancedafter the addition of a 50-fold molar excess of specificcompetitor compared with the negative control (Fig. 6,

compare lanes 4 and 7), which might be explained bythe stronger binding affinity of Paf to the specific inlA P2competitor DNA. At this concentration, Paf is titrated out,releasing PrfA from CI, thus leading to an increased for-mation of CIII. Addition of polyclonal anti-PrfA antibodiesabolished binding of the CI complex, whereas CII remainedunchanged (Fig. 6). To improve the separation of the twocomplexes, this experiment was carried out with the shorter53 bp inlA probe.

As shown above, Paf-containing extracts converted theCIII complex to CI, which binds to the 109 bp hly promoterfragment. A similar experiment was carried out with the101 bp inlA probe (Fig. 8). As expected, the CIII complexdisappeared with increasing amounts of the extract, butthere was much less CI complex observed than with thehly probe, and the level of the CI complex remained almostunchanged even with increasing amounts of MEM extract.The level of the CII complex increased even more than CIwith increasing amounts of extract, suggesting that Pafalone binds more strongly to the inlA probe than PrfA–Paf.

PrfA protects the PrfA-box in hly and inlA target DNAsagainst DNase I digestion

To define the precise binding site of PrfA in the regulatoryregions upstream of the hly and inlA genes, DNase I foot-print experiments were carried out with the purified PrfA


Fig. 4. Effect of increasing amounts(5–30 mg) of Paf-containing protein extracts(prepared from the prfA deletion mutantgrown in BHI or MEM) on the preformedcomplex (CIII) between PrfA (200 ng) and the109 bp hly promoter fragment (A) or the 28 bphly promoter fragment (B). BHI extracts of theprfA deletion mutant without purified PrfAwere used as control (C). Arrowheads indicatethe positions of the differentPrfA–DNA complexes (CI and CIII).


and two specific hly and inlA probes, both carrying thecorresponding PrfA-boxes.

Using the 109 bp hly probe and increasing amounts ofPrfA protein, we observed (at a concentration of 100 ngof PrfA) a protected region between positions ¹58 and¹33 from the transcriptional start site in the 58 upstreamregion of hly (Domann et al., 1993). This sequence com-prises the anticipated 14 bp PrfA-box of the hly promoterand, in addition, 10 nucleotides upstream and two nucleo-tides downstream of the palindrome. A hypersensitivenucleotide (A) was observed at position ¹40 near thecentre of the palindrome (Fig. 9A). The variations in theintensity of a band in position ¹67 were independent ofthe presence and amount of PrfA and therefore seemedto be unspecific.

To determine the PrfA-protected region within the inlAregulatory sequence, a similar footprint analysis was car-ried out using purified PrfA protein and the 101 bp inlA P2probe. This DNA fragment contains the entire P2 promotersequence including the anticipated PrfA-box, which differsby two basepair substitutions from the PrfA-box of the hlypromoter (see Fig. 2). In addition, the length of symmetryof the proposed inlA PrfA-box (Dramsi et al., 1993) is only12 bp compared with 14 bp for the hly PrfA-box. Detect-able protection of this inlA probe by PrfA required higherPrfA concentrations than with the previously described109 bp hly probe when the same probe concentrationwas used. With 200 ng of PrfA, a DNase I-hypersensitivenucleotide (T) was observed at position ¹39 (from thetranscriptional start site of the P2 promoter; Dramsi etal., 1993). A clear footprint was only observed with 400 ngof PrfA. The protected sequence covered the region from


Fig. 5. Specific binding of purified PrfA to the 101 bp inlA promoterfragment. Competition of PrfA binding with a 10- to 200-fold molarexcess of cold specific inlA promoter fragment and non-specificherring sperm DNA is indicated. Supershift caused by preincubationwith anti-PrfA antibodies (Ab) but not with preimmune serum (P) isshown on the right. The position of the PrfA–DNA complex isindicated (CIII).

Fig. 6. Influence of iron on PrfA-independentand PrfA-dependent binding of proteinextracts to the 101 bp inlA probe. Binding with30 mg of ‘BHI’, ‘MEM’ and ‘MEM þ Fe3þ

citrate (100 nM)’ extracts (B, M, MþFe) of theL. monocytogenes prfA deletion mutant(prfA¹) without and with the addition ofpurified PrfA is shown. The competition wascarried out using a BHI extract of the prfAdeletion mutant (as a source for Paf) andpurified PrfA with a 50- to 600-fold molarexcess of the specific inlA promoter fragmentand with unspecific herring sperm DNA.PrfA-dependent (CI and CIII) andPrfA-independent complexes (CII) areindicated. To improve the separation of CIand CII DNA binding of protein extracts of theprfA deletion mutant (10 mg of ‘BHI’ extract)with 200 ng of purified PrfA to the 53 bp inlAprobe after the addition of anti-PrfA antibodies(Ab) and preimmune serum (P) in a finaldilution of 1:15 is shown as an insert on theright.


¹55 to ¹33 from the transcriptional start site (Fig. 9B). Thisregion includes the palindromic sequence of the PrfA-boxof the inlA P2 promoter and nine additional nucleotidesupstream and two nucleotides downstream of this sequence.The downstream extension of the protection could not bedefined precisely in this experiment, but this has been

shown clearly by DNase I footprints using the 53 bp inlAprobe (data not shown).

These data showing the rather precise protection of thetwo analysed PrfA-boxes, sequences of dyad symmetry,suggest that PrfA binds to its target sites as a homodimer.To test whether PrfA is already present in solution in a


Fig. 7. Relative binding intensities (rel. BI) ofPaf-containing extracts (prepared from theprfA deletion mutant) separated by sucrosegradient centrifugation. Each fraction wasanalysed directly by EMSA using the 109 bphly (A) and 53 bp inlA (B) promoter fragmentsas target DNAs (K), leading to the formationof complex CII only with the inlA probe.(B) Relative intensity of CI complex formationafter the addition of purified PrfA to thebinding assay. The relative binding intensitieswere determined by densitometric scanning.The arrow denotes the sedimentation positionof human haemoglobin (4.5 S), which wasused as an internal marker.


homodimeric form, purified PrfA (as used in the bindingand protection studies) was run on a sucrose gradient(data not shown). The major portion of PrfA banded ataround 3.8 S. The molecular mass calculated from thissedimentation constant is clearly higher than expectedfor monomeric PrfA (27 kDa) and corresponds approxi-mately to a dimeric form of PrfA, suggesting that PrfAmay already exist in solution as a homodimer.

Discussion

The reported data show that purified PrfA protein alone isable to bind to target sites containing the anticipated PrfA-boxes of the PrfA-dependent hly and the inlA promotersconsisting of 14 and 12 bp respectively (Mengaud et al.,1989; Dramsi et al., 1993). Furthermore, footprint experi-ments indicate that PrfA (probably as a homodimer) indeedprotects these sequences of dyad symmetry, thus definitelyproving them as the PrfA binding sites. Binding of purifiedPrfA to these sites can occur in the absence of additional

listerial factors and is more efficient with the PrfA-box ofthe hly promoter than with that of the PrfA-dependentinlA promoter P2. This result is not unexpected consider-ing the substantial deviation in the nucleotide sequenceof the inlA PrfA-box compared with the ‘ideal’ PrfA bindingsequence realized in the PrfA-box of the hly promoter.These findings seem to support a model proposed pre-viously for the differential regulation of PrfA-controlledgenes in L. monocytogenes, which explains differentialPrfA-mediated gene regulation by varying binding effi-ciency of PrfA to the PrfA binding sites (PrfA-boxes) dif-fering from each other by defined nucleotide exchanges(Freitag et al., 1992; 1993; Freitag and Portnoy, 1994;Sheehan et al., 1995). All PrfA-boxes are located in frontof the PrfA-controlled genes or operons at position ¹40to ¹41 upstream of the transcriptional start sites (Men-gaud et al., 1989; Domann et al., 1991; Vazquez-Bolandet al., 1992; Dramsi et al., 1993). However, our previousresults (Bockmann et al., 1996) and those reported hereindicate that cell-free extracts of L. monocytogenes donot contain free PrfA and do not form the same simplePrfA complexes with suitable target DNA probes asobserved with the purified PrfA (complex CIII).

With a DNA fragment carrying the entire hly promoterregion (including the PrfA-box), cell-free, PrfA-containingL. monocytogenes extracts rather form a complex (CI)that migrates considerably slower than complex CIII, sug-gesting that PrfA is present in these extracts in a form ofhigher complexity and binds more efficiently in this com-plexed form the target DNA than to PrfA alone. A proteinfactor, termed Paf (for PrfA-activating factor) has beenidentified recently (Bockmann et al., 1996), which, togetherwith PrfA, seems to participate in CI complex formation.

Our new data show that Paf-containing extracts (obtainedfrom a prfA mutant) eliminate complex CIII, while generat-ing a new slower migrating complex, which moves to thesame position as complex CI obtained directly with PrfA-and Paf-containing L. monocytogenes extracts. Increas-ing amounts of the Paf-containing extract lead to a propor-tional decrease in preformed CIII, irrespective of the targetDNA (hly or inlA probes) to which PrfA is bound and inde-pendently of whether the Paf-containing extracts derivefrom bacteria grown in BHI or MEM. However, the amountof CI complex formed concomitantly depends on the targetDNA and the source of the Paf-containing extracts. Con-siderably more CI complex is formed with the hly probeas target DNA than with the inlA probe. Furthermore, theamount of CI complex is higher with Paf-containing ‘MEMextracts’ than with Paf-containing ‘BHI extracts’. Previousdata have already shown (Bockmann et al., 1996) that lowiron concentration in the growth medium (in MEM <0.05 mMas opposed to 12 mM in BHI) enhances Paf activity and/orinduces synthesis of Paf, leading to increased CI formation,while higher iron concentration represses CI formation.


Fig. 8. Addition of cell-free protein extracts of the L.monocytogenes prfA deletion mutant to the preformed CIII complexof PrfA (200 ng) and the 101 bp inlA promoter fragment. Theamount of extract (prepared from BHI or MEM) used is indicated ontop of each lane (in mg). A MEM extract of the prfA deletion mutantwithout the addition of PrfA was included as control (C). Positionsof the different complexes are indicated by arrowheads (CI, CII andCIII).


The precise effect of iron on Paf is not yet known: theincreased Paf activity in media with low iron may becaused by a larger amount of Paf and/or by the conver-sion of a Paf form (possibly Paf-Fe) unable to bind toPrfA into another one (possibly free Paf), which binds toPrfA efficiently. The PrfA–Paf generated thus could theninteract with the target DNA forming the CI complex. Inthis way, Paf would act as ‘sensor’ of the environmentaliron concentration and transforms it into differential PrfA-dependent expression of the virulence genes, but it cannotbe excluded from our data that Paf itself is modified byanother iron-dependent factor. Expression of at leastsome of these genes is also influenced by other environ-mental factors (Leimeister-Wachter et al., 1992; Park andKroll, 1993; Ripio et al., 1996; Milenbachs et al., 1997),and it remains to be seen whether these environmentalparameters also act via Paf or whether other factors mayinteract with PrfA in these cases.

Formation of the CI complex apparently requires a moreextended target sequence than just the PrfA-box, as CI isnot formed with PrfA-containing extracts and the 28 bp hlyprobe containing essentially the PrfA-box of the hly pro-moter. It is in line with this assumption that Paf-containingextracts are unable to shift the CIII complex, formed withPrfA and this short hly promoter probe, to a correspondingCI complex. Increasing amounts of Paf-containing extracts,which eliminate the preformed complex CIII consisting ofPrfA and the 109 bp hly probe, lead to only a slight decreasein CIII when formed between PrfA and the 28 bp hly probe,suggesting that a stable PrfA–Paf complex can only beformed in the presence of a suitable target DNA. Bindingof polyclonal PrfA antibodies leads to a supershift of theCIII complex (PrfA bound to the hly or the inlA probes)similar to that observed for many other complexes betweenDNA binding proteins and their target DNA sequences.

Compared with the related CRP protein, PrfA contains


Fig. 9. DNase I footprinting of PrfA bound to the 109 bp hly (A) and 101 bp inlA P2 (B) promoter fragments. The amount of PrfA used in thebinding reaction is indicated on top of each lane (in ng). Vertical black bars refer to the protected region. The G þ A lane was a sequencereaction of the DNA probes used as a marker. The numbers refer to the start sites of transcription. On the left side of the figures, thesequences of the protected regions are indicated with the palindromic sequences shown in bold letters and the hypersensitive nucleotidesboxed in. Both DNA probes were radioactively labelled in the (¹) strand.


an extra domain (LZD) exhibiting a leucine zipper motif(Lampidis et al., 1994). A PrfA mutant protein that haslost this LZD can no longer form the CI complex with the109 bp hly probe (M. Goetz, unpublished results). It istherefore tempting to speculate that PrfA interacts withPaf via LZD.

The interaction with second regulatory elements has alsobeen demonstrated for CRP (reviewed by Kolb et al., 1993).Activation of the two promoters, malEp and malKp, dependson the synergistic action of MalT and CRP. In this case, thebinding of these two regulatory factors to the promoterregion results in the formation of a higher order structureresponsible for malEp and malKp activation. A mode ofrepression is mediated by protein interaction betweenthe Crp–cAMP complex and the CytR repressor.

Purified PrfA protein can also bind to DNA fragmentscontaining the putative PrfA-box of the inlA P2 promoter,which is responsible for the partial PrfA-dependent expres-sion of inlA. The amount of PrfA protein necessary forefficient binding to and protection of this target site is con-siderably higher than that needed for the binding and pro-tection of the hly PrfA-box. This is not unexpected as thisPrfA-box shows considerable deviation from the consen-sus sequence. More important is the observation that aspecific complex (CII) is formed between this inlA targetDNA and Paf-containing fractions in the absence of PrfA,suggesting direct binding of Paf to this target site. An effi-cient complex formation was not observed with the hlytarget DNAs. Owing to the present lack of purified Paf,we are unable to determine precisely the sequence thatis protected by Paf, but we expect that Paf binding (e.g.formation of the CII complex) requires a more extendedsequence than PrfA binding. Paf binding to the inlAsequence seems to inhibit subsequent binding of PrfA toits target site, as CI complex formation is less efficientwith the inlA target DNA than with the hly promoterfragment.

Thus, binding of Paf and PrfA to the PrfA-box containingregulatory the inlA sequence is clearly different from thebinding of PrfA and Paf to the corresponding hly regula-tory sequence. Paf seems to play a dual role in the regula-tion of PrfA-dependent genes. While it might activatetranscription of some virulence genes, such as hly, plcA,actA and inlC, in the presence of low iron concentration,it represses the expression of inlA (and inlB) under thesame conditions. This may also explain why transcriptionof the former genes is highly induced, whereas transcrip-tion of the inlA-B operon is almost completely repressedwhen L. monocytogenes replicates within mammalian hostcells (where the iron concentration is low) (Engelbrecht etal., 1996). An opposite transcription pattern of these genesprobably occurs outside the host cells (where the iron con-centration may be higher than within the host cells). Here,L. monocytogenes depends on the expression of inlA-B

for the invasion of host cells, while the expression of hlyand the other PrfA-dependent genes required for intracel-lular replication of the bacteria is less necessary. Underthese ‘extracellular’ conditions, lower amounts of Paf arepresent, and the inlA-B operon is expressed from P2 asshown recently (Bohne et al., 1996).

Experimental procedures

Bacterial strains and culture conditions

The L. monocytogenes strain EGD used in this study wasprovided by S. H. E. Kaufmann, University of Ulm (Germany).The isogenic PrfA deletion mutant (prfA¹) has been des-cribed previously (Bockmann et al., 1996). Bacteria weregrown in brain–heart infusion broth (BHI; Difco) at 378Cwith aeration.

Purification of PrfA protein

The [His]6-tagged PrfA protein (P2-7) was purified on Ni-NTAresin (Qiagen) as described by Bockmann et al. (1996). Pro-tein purity was analysed by Coomassie staining of SDS poly-acrylamide gels according to Laemmli (1970). Yields wereabout 6 mg for 1 l of culture after 4 h of IPTG induction.

Immunoblotting

Total cellular proteins from an equal number of cells wereseparated by 13% SDS–PAGE, and the proteins were trans-ferred from the gel to nitrocellulose filters (Schleicher &Schuell) by the semi-dry electroblotting method of Kyhse-Andersen (1984). After incubation with guinea pig anti-PrfAantiserum (1:1000 dilution), PrfA was detected by a peroxi-dase reaction using 4-chloro-1-naphthol (0.5 mg ml¹1) as thesubstrate.

Determination of haemolytic activity

Bacteria were grown to an OD600 ¼ 1 and shifted for 30 mininto MEM with Earle’s salt (Gibco). The supernatants andwhole-cell suspensions (100 ml of each) were tested for theirhaemolytic activities by titre assays. The haemolytic assaywas performed as described by Kathariou et al. (1987),except that no reducing agent (dithioerythrol) was added.The haemolytic activities are given in complete haemolyticunits (CHU), defined as the reciprocal of the last dilution atwhich complete haemolysis was observed.

Electrophoretic mobility shift assays (EMSAs) withpurified PrfA

The various DNA probes of the hly and inlA promoter regions(see Fig. 2) were either synthetic double-stranded oligo-nucleotides with 58 protruding termini or fragments obtainedby polymerase chain reaction (PCR) amplification using theprimer pairs listed in Table 2. Sites for restriction endo-nucleases are underlined. The 109 bp hly fragment and the101 bp inlA fragment contained sites for XbaI and Bgl II/Bcl Irestriction endonucleases respectively.



After digestion and gel purification, the PCR fragments werequantitated by measuring the optical density at 260 nm. All frag-ments were labelled at both ends using the Klenow polymerase(Pharmacia) with [32P]-dATP and [32P]-dCTP (3000Ci mmol¹1;Amersham).

The binding reactions contained 200 ng of purified PrfA pro-tein, 140 000 c.p.m. of [32P]-labelled DNA, 3 ml of 5× bindingbuffer [100 mM HEPES, pH 7.9, 5 mM dithiothreitol (DTT),5 mM EDTA, 250 mM KCl, 15% Ficoll], 5 mM MgSO4, 1 mgml¹1 BSA and 0.07 mg ml¹1 poly-(dI-dC) non-specific compe-titor DNA in a final volume of 15 ml. Reactions were preincu-bated for 3 min at 378C and then an additional 27 min on ice.

For CI complex formation, Paf-containing protein extractsof the prfA deletion mutant (up to 30 mg) were added to thebinding assay containing PrfA bound to the DNA, and thebinding reactions were further incubated for 5 min at 378Cand 10 min on ice.

Binding reactions with protein extracts alone were per-formed as described previously (Bockmann et al., 1996).

The DNA–protein complexes were separated on native 5%polyacrylamide gels in low-ionic-strength buffer (0.4× Trisborate–EDTA) at 200 V for 3–5 h at room temperature witha prerun of 1.5 h. After electrophoresis, the gels were fixedin 10% acetic acid, vacuum dried and visualized byautoradiography.

PrfA-specific binding in the presence of anti-PrfAantiserum

Anti-PrfA antiserum raised against purified [His]6-tagged PrfAprotein (Bockmann et al., 1996) was incubated with purifiedPrfA or crude protein extracts in binding buffer for 20 min onice before the labelled oligonucleotides were added. The bind-ing reaction, with a final dilution of the antibodies of 1:15, wascontinued for an additional 25 min on ice before the sampleswere loaded onto the gel.

Preparation of protein extracts

Overnight cultures of L. monocytogenes were diluted 1:10into fresh BHI medium and grown to an optical density at600 nm of 1.0. After centrifugation, the cells were eithershifted for 30 min into MEM (Gibco) or resuspended directlyinto 1/200 volume of cold buffer A [10 mM HEPES, pH 7.9,10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM DTT,

0.5 mM phenylmethylsulphonyl fluoride (PMSF)]. The bacteriawere frozen in liquid nitrogen and disrupted in a mortar. Aftersonication (three times for 15 s at 48C), the cell debris wasremoved by centrifugation (20 min, 8000 × g, 48C). Total pro-tein concentrations were determined with the Bio-Rad ProteinMicroassay, a modified Bradford method, and the super-natants were frozen in aliquots at ¹808C.

Sucrose gradient centrifugation

Total protein extracts (250 ml, 1 mg of protein) were layered onthe top of a 5–20% sucrose gradient prepared with bindingbuffer (20 mM HEPES, pH 7.9, 50 mM KCl, 1 mM EDTA,1 mM DTT). Centrifugation was carried out for 20 h in a SW50.1 rotor at 42 000 r.p.m. and 48C, and human haemoglobin(4.5 S) was used as a sedimentation reference. Fractions (12drops) were collected, and 20 ml of each fraction was analysedfor binding activation by electrophoretic mobility shift assays.

DNase I footprint experiments

DNase I footprint experiments were modified after Zu et al.(1996).

The 109 bp hly promoter fragment and the 101 bp inlA pro-moter fragment were cloned into pSK and pUC18 respectively.The plasmids were linearized by EcoRI digestion (Bgl II for theinlA fragment) and end-labelled with [g-32P]-ATP and T4 poly-nucleotide kinase. Subsequently, the plasmid was restrictedwith BamHI (EcoRI for inlA) to produce fragments that areuniquely labelled at one terminus. The DNA was gel purified,and approximately 70 000 c.p.m. was used for footprintingexperiments. Protein–DNA complexes were formed in 50 mlof footprint buffer [20 mM HEPES, pH 7.9, 20 mM KCl, 5 mMCaCl2, 2 mM MgCl2, 1 mM EDTA, 1 mM DTT, 0.2 mg ml¹1

BSA, 0.07 mg ml¹1 poly-(dI-dC)] for 25 min at 258C. DNase I(0.75 U) in footprint buffer with 10 mM MgCl2 was added,and the incubation was continued for 1 min at 258C. Afterthe addition of 140 ml of stop solution (192 mM sodium ace-tate, 32mM EDTA, 0.14% SDS and 64 mg ml¹1 yeast tRNA),the mixture was treated with phenol–chloroform and precipi-tated with ethanol. The pellets were resuspended in sequenceloading buffer, denatured, loaded on 6% polyacrylamidesequencing gels and autoradiographed. The G þ A reactionwas performed according to the rapid method of Maxam–Gil-bert sequencing (Sambrook et al., 1989).


Table 2. Primer pairs used in this study.

Primer Sequence 58→38 DNA probe Source

BM1BM2

1275TCCTATCTAGAAGTGACTTTTATGTT1300

1395GCTTCTCTAGATGAAACGCAATATTA1370hly 109 bp Mengaud et al. (1989)

PEGD1PEGD2

1300TGAGGCATTAACATTTGTTAACG1322

1327ATCGTCGTTAACAAATGTTAATG1305hly 25 bp Mengaud et al. (1989)

iA1iA2

1050CACACCCAGATCTGTTATTTTGAACATAAAGGG1082

1165TCTATATGATCACCCCTTTCAAAGCCC1139inlA 101 bp Gaillard et al. (1991)

AS1AS2

1085GAGGATTAACATAAGTTAATTCTTTTTTTTGGAAAAATAGTTATTAT1130

1090TTAAATAATAATAACTATTTTTCCAAAAAAAAGAATTAACTTATGTTA1137inlA 53 bp Gaillard et al. (1991)

AS5AS6

1104TCTTTTTTTTGGAAAAATAGTTATTATTATTTA1136

1108AAGCCCATTAAATAATAATAACTATTTTTCCAAAAAA1144inlA 41 bp Gaillard et al. (1991)


Acknowledgements

We are grateful to V. Scarlato for help and advise with theDNase I footprint technique. We also thank R. Gross, J. Kreft,E. Ng and J. Daniels for critical reading of the manuscript andfor helpful discussions. This work was supported by grantsfrom the Deutsche Forschungsgemeinschaft (SFB 165-B4).

References

Bockmann, R., Dickneite, C., Middendorf, B., Goebel, W.,and Sokolovic, Z. (1996) Specific binding of the Listeriamonocytogenes transcriptional regulator protein PrfA totarget sequences requires additional factor(s) and is influ-enced by iron. Mol Microbiol 22: 643–653.

Bohne, J., Sokolovic, Z., and Goebel, W. (1994) Transcrip-tional regulation of prfA and PrfA-regulated virulencegenes in Listeria monocytogenes. Mol Microbiol 11: 1141–1150.

Bohne, J., Kestler, H., Uebele, C., Sokolovic, Z., and Goebel,W. (1996) Differential regulation of the virulence genes ofListeria monocytogenes. Mol Microbiol 20: 1189–1198.

Domann, E., Leimeister-Wachter, M., Goebel, W., and Chak-raborty, T. (1991) Molecular cloning, sequencing andidentification of a metalloprotease gene from Listeriamonocytogenes that is species specific and physicallylinked to the listeriolysin gene. Infect Immun 59: 65–72.

Domann, E., Wehland, J., Niebuhr, K., Haffner, C., Leimeister-Wachter, M., and Chakraborty, T. (1993) Detection of aprfA-independent promoter responsible for listeriolysingene expression in mutant Listeria monocytogenes strainslacking the PrfA regulator. Infect Immun 61: 3073–3075.

Dramsi, S., Kocks, C., Forestier, C., and Cossart, P. (1993)Internalin-mediated invasion of epithelial cells by Listeriamonocytogenes is regulated by the bacterial growth state,temperature, and the pleiotropic activator prfA. Mol Micro-biol 9: 931–941.

Dramsi, S., Biswas, I., Maguin, E., Braun, L., Mastroeni, P.,and Cossart, P. (1995) Entry of Listeria monocytogenesinto hepatocytes requires expression of InlB, a surface pro-tein of the internalin multigene family. Mol Microbiol 16:251–261.

Engelbrecht, F., Chun, S., Ochs, C., Hess, J., Lottspeich, F.,Goebel, W., et al. (1996) A new PrfA-regulated gene of Lis-teria monocytogenes encoding a small, secreted proteinwhich belong to the family of internalins. Mol Microbiol21: 823–837.

Freitag, N.E., and Portnoy, D.A. (1994) Dual promoters of theListeria monocytogenes transcriptional activator appearessential in vitro but are redundant in vivo. Mol Microbiol12: 845–853.

Freitag, N.E., Youngman, P., and Portnoy, D.A. (1992) Tran-scriptional activation of the Listeria monocytogenes hemo-lysin gene in Bacillus subtilis. J Bacteriol 174: 1293–1298.

Freitag, N.E., Rong, L., and Portnoy, D.A. (1993) Regulationof the prfA transcriptional activator of Listeria monocyto-genes: multiple promoter elements contribute to intracellu-lar growth and cell-to-cell spread. Infect Immun 61: 2537–2544.

Gaillard, J.L., Berche, P., Frehel, C., Gouin, E., and Cossart,P. (1991) Entry of Listeria monocytogenes into cells is

mediated by internalin, a repeat protein reminiscent of sur-face antigens from Gram-positive cocci. Cell 65: 1127–1141.

Gellin, B.G., and Broome, C.V. (1989) Listeriosis. J Am MedAssoc 261: 1313–1320.

Gouin, E., Mengaud, J., and Cossart, P. (1994) The virulencegene cluster of Listeria monocytogenes is also present inListeria ivanovii, an animal pathogen and Listeria seeligeri,a non-pathogenic species. Infect Immun 62: 3550–3553.

Kathariou, S., Metz, P., Hof, H., and Goebel, W. (1987)Tn917-induced mutations in the hemolysin determinantaffecting virulence of Listeria monocytogenes. J Bacteriol169: 1291–1297.

Kolb, A., Busby, S., Buc, H., Garges, S., and Adhya, S. (1993)Transcriptional regulation by cAMP and its receptor protein.Annu Rev Biochem 62: 749–795.

Kreft, J., Bohne, J., Gross, R., Kestler, H., Sokolovic, Z., andGoebel, W. (1995) Control of Listeria monocytogenes viru-lence by the transcriptional regulator PrfA. In Signal Trans-duction and Bacterial Virulence. Rappuoli, R., Scarlato, V.,and Arico, B. (eds). R.G. Landes Company, Austin, pp.129–142.

Kyhse-Andersen, J. (1984) Electroblotting of multiple gels: asimple apparatus without buffer tank for rapid transfer ofproteins from polyacrylamide to nitrocellulose. J BiochemBiophys Methods 10: 203–209.

Kuhn, M., and Goebel, W. (1995) Molecular studies on thevirulence of Listeria monocytogenes. Genet Eng 17: 31–51.

Laemmli, U.K. (1970) Cleavage of structural proteins duringthe assembly of the head of bacteriophage T4. Nature227: 680–685.

Lampidis, R., Gross, R., Sokolovic, Z., Goebel, W., and Kreft,J. (1994) The virulence regulator protein of Listeria ivanoviiis highly homologous to PrfA from Listeria monocytogenesand both belong to the Crp-Fnr family of transcription reg-ulators. Mol Microbiol 13: 141–151.

Leimeister-Wachter, M., Haffner, C., Domann, E., Goebel, W.,and Chakraborty, T. (1990) Identification of a gene thatpositively regulates expression of listeriolysin, the majorvirulence factor of Listeria monocytogenes. Proc Natl AcadSci USA 87: 8336–8340.

Leimeister-Wachter, M., Domann, E., and Chakraborty, T.(1992) The expression of virulence genes in Listeria mono-cytogenes is thermoregulated. J Bacteriol 174: 947–952.

Lingnau, A., Chakraborty, T., Niebuhr, K., Domann, E., andWehland, J. (1996) Identification and purification of novelinternalin-related proteins in Listeria monocytogenes andListeria ivanovii. Infect Immun 64: 1002–1006.

Mengaud, J., Vincente, M.F., and Cossart, P. (1989) Tran-scriptional mapping and nucleotide sequence of the Lis-teria monocytogenes hlyA region reveal structural featuresthat may be involved in regulation. Infect Immun 57:3695–3701.

Mengaud, J., Dramsi, S., Gouin, E., Vazquez-Boland, J.A.,Milon, G., and Cossart, P. (1991) Pleiotropic control of Lis-teria monocytogenes virulence factors by a gene that isautoregulated. Mol Microbiol 5: 2273–2283.

Milenbachs, A.A., Brown, D.P., Moors, M., and Youngman,P. (1997) Carbon-source regulation of virulence geneexpression in Listeria monocytogenes. Mol Microbiol 23:1075–1085.



Park, S.F., and Kroll, R.G. (1993) Expression of listerio-lysin and phosphatidylinositol-specific phospholipase C isrepressed by the plant-derived molecule cellobiose in Lis-teria monocytogenes. Mol Microbiol 8: 653–661.

Portnoy, D.A., Chakraborty, T., Goebel, W., and Cossart, P.(1992) Molecular determinants of Listeria monocytogenespathogenesis. Infect Immun 60: 1263–1267.

Ripio, M.-T., Dominquez-Bernal, G., Suarez, M., Brehm, K.,Berche, P., and Vazquez-Boland, J.-A. (1996) Transcrip-tional activation of virulence genes in wild-type strains ofListeria monocytogenes in response to a change in theextracellular medium composition. Res Microbiol 147:311–384.

Ripio, M.-T., Dominquez-Bernal, G., Lara, M., Suarez, M.,and Vazquez-Boland, J.-A. (1997) A Gly145Ser substitu-tion in the transcriptional activator PrfA causes constitutiveoverexpression of virulence factors in Listeria monocyto-genes. J Bacteriol 179: 1533–1540.

Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989) Molecu-lar Cloning: a Laboratory Manual. 2nd edn. Cold SpringHarbor, NY: Cold Spring Harbor Laboratory Press.

Sheehan, B., Kocks, C., Dramsi, S., Gouin, F., Klarsfeld,A.D., Mengaud, J., et al. (1994) Molecular and genetic

determinants of the Listeria monocytogenes infectiousprocess. Curr Topics Microbiol Immunol 192: 187–216.

Sheehan, B., Klarsfeld., A., Msadek, T., and Cossart, P.(1995) Differential activation of virulence gene expressionby PrfA, the Listeria monocytogenes virulence regulator.J Bacteriol 177: 6469–6476.

Sheehan, B., Klarsfeld, A., Ebright, R., and Cossart, P.(1996) A single substitution in the putative helix–turn–helix motif on the pleiotropic activator PrfA attenuates Lis-teria monocytogenes virulence. Mol Microbiol 20: 785–797.

Sokolovic, Z., Riedel, J., Wuenscher, M., and Goebel, W.(1993) Surface associated, PrfA-regulated proteins of Lis-teria monocytogenes synthesized under stress conditions.Mol Microbiol 8: 219–227.

Vazquez-Boland, J., Kocks, C., Dramsi, S., Ohayon, H.,Geoffroy, C., Mengaud, J., et al. (1992) Nucleotide sequenceof the lecithinase operon of Listeria monocytogenes andpossible role of lecithinase in cell-to-cell spread. InfectImmun 60: 219–230.

Zu, T., Manetti, R., Rappuoli, R., and Scarlato, V. (1996) Dif-ferential binding of BvgA to two classes of virulence genesof Bordetella pertussis directs promoter selectivity by RNApolymerase. Mol Microbiol 21: 557–565.



differential interaction of the transcription factor prfa and the prfa-activating factor (paf) of...

Documents