Analysis of virulence plasmid gene expression deﬁnes three ...€¦ · Analysis of virulence plasmid gene expression deﬁnes three classes of effectors in the type III secretion

Analysis of virulence plasmid gene expressiondefines three classes of effectors in the typeIII secretion system of Shigella flexneri

Tony Le Gall,1,2 Maria Mavris,3 Maria Celeste Martino,4

Maria Lina Bernardini,4 Erick Denamur1 and Claude Parsot3

Correspondence

Claude Parsot

[email protected]

1INSERM E0339, Faculte de Medecine Xavier Bichat, 16 rue Henri Huchard, F-75018 Paris,France

2Laboratoire de Bacteriologie, Faculte de Medecine de Brest, F-29285 Brest Cedex, France

3Unite de Pathogenie Microbienne Moleculaire, INSERM U389, Institut Pasteur, 25 rue du DrRoux, F-75724 Paris Cedex 15, France

4University Roma La Sapienza, Sez Sci Microbiol, Dipartimento Biol Cellulare & Sviluppo,Via Sardi 70, Rome, I-00185 Italy

Received 22 September 2004

Revised 22 November 2004

Accepted 23 November 2004

Proteins directly involved in entry and dissemination of Shigella flexneri into epithelial cells

are encoded by a virulence plasmid of 200 kb. A 30-kb region (designated the entry region)

of this plasmid encodes components of a type III secretion (TTS) apparatus, substrates of this

apparatus and their dedicated chaperones. During growth of bacteria in broth, expression of

these genes is induced at 37 6C and the TTS apparatus is assembled in the bacterial envelope

but is not active. Secretion is activated upon contact of bacteria with host cells and is

deregulated in an ipaB mutant. The plasmid encodes four transcriptional regulators, VirF,

VirB, MxiE and Orf81. VirF controls transcription of virB, whose product is required for

transcription of entry region genes. MxiE, with the chaperone IpgC acting as a co-activator, controls

expression of several effectors that are induced under conditions of secretion. Genes under the

control of Orf81 are not known. The aim of this study was to define further the repertoires of

virulence plasmid genes that are under the control of (i) the growth temperature, (ii) each of the

known virulence plasmid-encoded transcriptional regulators (VirF, VirB, MxiE and Orf81) and

(iii) the activity of the TTS apparatus. Using a macroarray analysis, the expression profiles of 71

plasmid genes were compared in the wild-type strain grown at 37 and 30 6C and in virF, virB,

mxiE, ipaB, ipaB mxiE and orf81 mutants grown at 37 6C. Many genes were found to be under

the control of VirB and indirectly of VirF. No alteration of expression of any gene was detected in

the orf81 mutant. Expression of 13 genes was increased in the secretion-deregulated ipaB

mutant in an MxiE-dependent manner. On the basis of their expression profile, substrates of the

TTS apparatus can be classified into three categories: (i) those that are controlled by VirB,

(ii) those that are controlled by MxiE and (iii) those that are controlled by both VirB and MxiE.

The differential regulation of expression of TTS effectors in response to the TTS apparatus

activity suggests that different effectors might be required at different times following contact

of bacteria with host cells.

INTRODUCTION

Bacteria of Shigella species are responsible for shigellosis,a human disease characterized by the destruction of thecolonic epithelium. Shigellae and enteroinvasive Escherichiacoli both contain a virulence plasmid (VP) of approximately200 kb that encodes most proteins directly involved in

entry and dissemination of bacteria into epithelial cells.Sequence analysis of the VP pWR100 (Buchrieser et al.,2000) and its derivative pWR501 (Venkatesan et al., 2001)from a Shigella flexneri strain of serotype 5 and the VPpCP301 (Jin et al., 2002) from an S. flexneri strain ofserotype 2a indicated that the VP is composed of a mosaicof approximately 100 genes and numerous insertionsequences. The VP genes exhibit a G+C content from30 to 60 mol%, suggesting that they were acquired fromdifferent sources.Abbreviations: TTS, type III secretion; VP, virulence plasmid.

0002-7639 G 2005 SGM Printed in Great Britain 951

Microbiology (2005), 151, 951–962 DOI 10.1099/mic.0.27639-0

Genes required for entry of bacteria into epithelial cellsand the induction of apoptosis in infected macrophagesare clustered on a 30 kb region (designated the entry region)of the VP. This region encodes components of a type IIIsecretion (TTS) apparatus (the Mxi-Spa TTS apparatus),substrates of this secretion apparatus (the translocatorsIpaB and IpaC and the effectors IpaD, IpgB1, IpgD and IcsB)and their dedicated chaperones (IpgA, IpgC, IpgE andSpa15) and two transcriptional activators (VirB and MxiE).The current model of the TTS pathway proposes that, uponcontact of bacteria with host cells, translocators insert intothe membrane of the host cell to form a pore throughwhich effectors transit to reach the cell cytoplasm (Hueck,1998). Other substrates of the TTS apparatus are encodedby genes scattered throughout the VP, such as virA, ospB,C, D, E, F and G and ipaH genes (Buchrieser et al., 2000).Several of these putative effectors are encoded by multigenefamilies, with five ipaH, four ospC, three ospD and two ospEgenes carried by the VP. Genes encoding components ofthe TTS apparatus and its substrates exhibit a similar lowG+C content (approx. 34 mol%), suggesting that theentire TTS system was acquired once by lateral transfer. Inaddition, the VP encodes at least five other proteins thatare involved in virulence, including IcsA, IcsP, VirK, MsbB2and SepA. IcsA (VirG) is an outer-membrane proteindirectly involved in promoting actin polymerization atone pole of intracellular bacteria (Bernardini et al., 1989;Goldberg & Theriot, 1995; Lett et al., 1989). IcsP (SopA) isan outer-membrane protease involved in the release of acertain proportion of surface-exposed IcsA (Egile et al.,1997; Shere et al., 1997). VirK is required for productionof IcsA by an unknown mechanism (Nakata et al., 1992).MsbB2 is an acyl transferase that, in conjunction with theproduct of the chromosomal msbB1 gene, acts to producefull acyl-oxy-acylation of the myristate at the 39 position ofthe lipid A glucosamine disaccharide (d’Hauteville et al.,2002). SepA is a secreted serine protease of the IgA1 proteasefamily whose substrate is not known (Benjelloun-Touimiet al., 1995, 1998).

The regulation of expression of VP genes has been thesubject of extensive studies [see Dorman & Porter (1998)for a review]. Expression of genes of the entry region isregulated by the growth temperature, these genes beingexpressed at 37 uC but not at 30 uC (Maurelli et al., 1984),and by a cascade involving two VP gene products, VirFand VirB. VirF, a member of the AraC family of trans-criptional activators, is required for transcription of virB(Sakai et al., 1986; Tobe et al., 1993) and VirB, a memberof the ParB family of partitioning proteins, is requiredfor transcription of genes of the entry region (Adler et al.,1989; Tobe et al., 1991). Control by temperature is depend-ent upon the chromosomally encoded protein H-NS, asinactivation of hns leads to expression of genes of the entryregion at both 30 and 37 uC (Maurelli & Sansonetti, 1988;Prosseda et al., 1998). The current model proposes thatbinding of H-NS to the virF and virB promoters, as well asto promoters controlled by VirB, prevents transcription of

these genes at 30 uC (Beloin & Dorman, 2003). At 37 uC,changes in DNA conformation at virF and virB promoterslead to an increased transcription of virF and to theactivation of the virB promoter by VirF (Dorman et al.,2001). Production of VirB, in turn, leads to activation ofpromoters of the entry region by a mechanism that hasnot yet been elucidated but that might involve displace-ment of H-NS (Adler et al., 1989; Beloin et al., 2002; Beloin& Dorman, 2003; McKenna et al., 2003). During growthof bacteria in broth at 37 uC, the TTS apparatus is assembledin the bacterial envelope but is only weakly active (Menardet al., 1994a) and substrates of the TTS apparatus are storedin the cytoplasm in association with specific chaperones(Menard et al., 1994b; Niebuhr et al., 2000; Ogawa et al.,2003; Page et al., 2001, 2002). The TTS apparatus is acti-vated upon both contact of bacteria with epithelial cells(Menard et al., 1994a) and exposure of bacteria to the dyeCongo red (Bahrani et al., 1997; Parsot et al., 1995). TheTTS apparatus is deregulated, i.e. constitutively active inbacteria growing in broth, by inactivation of ipaB or ipaD(Menard et al., 1994a). Conditions of active or deregulatedsecretion lead to increased transcription of members ofthe ipaH family, ospF, virA and ospC1, but not genes ofthe entry region (Demers et al., 1998; Mavris et al., 2002a).Transcription of genes regulated by the TTS apparatusactivity involves both MxiE, a transcriptional activator ofthe AraC family encoded by the entry region, and IpgC,the chaperone for the IpaB and IpaC translocators, actingas a co-activator of MxiE (Mavris et al., 2002a). Underconditions of non-secretion, IpaB and IpaC are asso-ciated with and titrate IpgC and MxiE is not active. Underconditions of secretion, IpaB and IpaC are secreted andIpgC is liberated in the cytoplasm, where it activates MxiEby an unknown mechanism likely to involve an interactionbetween MxiE and IpgC. Much less is known about theregulation of expression of VP genes that are not part ofthe TTS system. Production of IcsA (VirG) was notaffected in a virB mutant (Adler et al., 1989) and trans-cription of an icsA–lacZ fusion carried by a reporter plasmidwas decreased 2?5-fold in a virF mutant compared withthe wild-type strain (Sakai et al., 1988), suggesting thatexpression of icsA was directly under the control of VirF.Northern blot analysis indicated that expression of icsAwas moderately (2-fold) modulated by the temperature ofgrowth (Porter & Dorman, 1997). Expression of sepA wasnot modulated by temperature (Benjelloun-Touimi et al.,1995) and expression of icsP was shown to be controlled byVirB (Berlutti et al., 1998; Wing et al., 2004). In additionto VirF and MxiE, pWR100 encodes a third protein ofthe AraC family, Orf81, whose function has not yet beeninvestigated (Buchrieser et al., 2000).

The numerous studies that have been performed on theregulation of expression of VP genes have provided a wealthof information on the nature of some genes that areregulated and the mechanism(s) of their control. However,these studies used a variety of techniques and differentstrains and growth conditions and, as a consequence, their

952 Microbiology 151

T. Le Gall and others

results are difficult to compare. To obtain a global viewon expression of VP genes, we performed a macroarrayanalysis using membranes carrying PCR fragments corres-ponding to most VP genes. To define further the repertoiresof genes that are under the control of (i) the growth tem-perature, (ii) each of the known transcriptional regulatorsencoded by the VP, i.e. VirF, VirB, MxiE and Orf81, and (iii)the activity of the TTS apparatus, i.e. activated in an ipaBmutant, we analysed the expression profiles in the wild-type strain grown at 37 and 30 uC and in virF, virB, mxiE,ipaB, ipaB mxiE and orf81 mutants grown at 37 uC. For thesake of clarity, results are presented in the reverse orderof the regulatory cascade presented above, i.e. from theanalysis of an ipaB mutant at 37 uC, in which most genesare expressed, to that of the wild-type strain at 30 uC, inwhich most genes are not expressed.

METHODS

Bacterial strains. The S. flexneri wild-type strain M90T-Sm andthe ipaB4, mxiE and ipaB4 mxiE mutants were described previously(Allaoui et al., 1992; Mavris et al., 2002a). The orf81 mutant wasconstructed by allelic exchange, inserting an aphA3 cassette at codon25 of the orf81 gene carried by the VP. The virF mutant was con-structed by integration of a suicide plasmid carrying a DNA frag-ment corresponding to codons 27–118 of virF into the virF geneof pWR100. The virB mutant was constructed by allelic exchange,replacing codons 135–158 of virB by a cassette conferring resistanceto chloramphenicol.

Macroarray design. We developed a set of membranes each carry-ing PCR fragments corresponding to most pWR100 genes (exceptthose encoded by insertion sequences) and four chromosomal genes,adk, cysK, udp and tufB. For each gene, primer pairs were designedto amplify internal DNA fragments of approximately 600 bp specificto each gene, with the following exceptions. In the case of the smallmxiH and mxiI genes, the PCR fragment included both genes andwas designated mxiHI. The sequences of ipaH1 and ipaH2, ospE1and ospE2, and ospC2, ospC3 and ospC4 are identical or too similarto be distinguished by hybridization and the corresponding PCRfragments were designated ipaH1/2, ospE1/2 and ospC2/3/4, respec-tively. The ipaH1.4, ipaH2.5, ipaH4.5, ipaH7.8 and ipaH9.8 genes aredesignated here ipaH1, ipaH2, ipaH4, ipaH7 and ipaH9, respectively.PCR products were spotted in duplicate copies onto nitrocellulosemembranes at the Genopole at the Pasteur Institute (Paris). Asnegative controls, S. flexneri genomic DNA at the same concentra-tion as that used for PCR amplifications of plasmid genes andmouse genomic DNA were also spotted onto the membranes.

DNA and RNA extraction and labelling. Bacteria were grownaerobically at 37 uC in 869 medium with constant shaking. Totalgenomic DNA from the wild-type strain was isolated from 1 mlculture in stationary phase using the standard protocol of theWizard Genomic DNA purification kit (Promega). Five hundrednanograms of total genomic DNA was used as a template for incor-poration of [a-33P]dCTP (Perkin Elmer) in a polymerization reactionusing random hexamer primers and the Klenow fragment of DNApolymerase I (Roche) according to the manufacturers’ recommenda-tions. For RNA extraction, an overnight culture of bacteria wasdiluted 1 : 10 000 into 10 ml fresh medium and incubated at 37 uCuntil the mid-exponential phase was reached. Six millilitres of thebacterial suspension was centrifuged and the bacterial pellet wasresuspended in 6 ml of a solution of RNAplus (Qbiogen) (pre-warmed at 65 uC) by vigorous shaking and then frozen rapidly on

dry ice. RNA was extracted as recommended in the RNAplus pro-tocol and the purification was performed using the Nucleospin RNAII kit (Macherey Nagel), which included a DNase treatment step.cDNA synthesis was performed using 20 mg total RNA usingrandom hexamer primers (Amersham Biosciences), the reversetranscriptase Superscript II (Invitrogen) and [a-33P]dCTP. Unincor-porated nucleotides were removed from labelled probes by gel filtra-tion through G-25 Sephadex columns (Roche).

Macroarray hybridizations. Membranes on which PCR productshad been spotted were rinsed in 66 SSC solution (Invitrogen) andprehybridized in 10 ml ExpressHyb hybridization solution (Ozyme)for 1 h at 68 uC. Labelled cDNA or labelled genomic DNA probeswere denatured in 10 ml ExpressHyb hybridization solution andused for hybridization with membranes for 15–18 h at 68 uC.Membranes were then washed three times with a solution containing0?56 SSPE and 0?2% SDS at room temperature for 3 min, followedby three washes in the same solution at 65 uC for 20 min each,rinsed in 66 SSC solution, slightly dried on a Whatman paper,wrapped and sealed in a damp plastic 25 mm film and exposed to aPhosphorImager screen ADC plate MD30 (AGFA) for 48–72 h.Before rehybridization, membranes were stripped in 500 ml boiledbuffer (10 mM Tris/HCl, pH 7?6, 1 mM EDTA, 1% SDS) for25 min at 100 uC. Complete removal of radioactivity was checked byovernight exposure to a PhosphorImager screen. DNA hybridiza-tions were repeated three times and cDNA hybridizations wererepeated with different RNA extractions at least four times, exceptfor the wild-type strain grown at 30 uC (three times).

Macroarray data acquisition. Exposed PhosphorImager screenswere scanned on a Storm 860 PhosphorImager (Molecular Dynamics)at a pixel size of 100 mm. The software package XDotsReader (Cose)was used to grid the phosphorimaging image and to record thepixel densities. The background signal was locally quantified andsubtracted from each spot intensity. The output data were exportedto a Microsoft Excel spreadsheet for subsequent manipulations.

Data processing and statistical analysis. Following hybridiza-tion with genomic DNA, each dot intensity was normalized accord-ing to the mean value measured over all spots on each DNA array.Dots showing heterogeneity between different arrays were not con-sidered in the analysis presented below. Moreover, as an internalquality control of each array, the signal variation (SV) over eachspot of a membrane was calculated as SV=dot signal/mean ofoverall dot signal on that membrane. Generally, these values werebetween 0?7 (stbA) and 1?3 (orf82), with two exceptions: the SVvalue was 1?5 for ipaH1/2 and 1?7 for ospC2/3/4, consistent with thefact that ipaH1 and ipaH2 genes and ospC2, ospC3 and ospC4 genescould hybridize with the same PCR fragments on the membranes.In contrast, although the opsE1 and ospE2 genes are identical, the SVvalue for the ospE1/2 spots was not increased (SV=0?9).

For transcription profiles, standardization was performed by expres-sing the intensity value of individual spots relative to the meancalculated over the intensities obtained from the two tufB spotssuch that the tufB intensities in each experiment were close to 1. Theaccuracy and reproducibility of the results were analysed by compar-ing the distribution of intensity in each spot between duplicate dotsand between hybridization repeats. In every possible comparison, theresults showed a high degree of correlation (r2>0?90). The expressionprofiles of mutants were compared to those of the wild-type strainby calculating the consistency of differential expression across replicatehybridization values using the Wilcoxon two-sample test (Z-test). Thisnon-parametric statistical method contained in the StatView 5.0.1package (SAS Institute Inc.) tests the hypothesis that one of thepaired variables is larger or smaller than the other variable regardlessof the magnitude of the difference and is appropriate for the analysisof small samples. This allowed us to identify genes that were

http://mic.sgmjournals.org 953

Transcriptome analysis of virulence plasmid genes

differentially expressed between two strains: only genes with the lowestpossible Z-test P value were considered for analysis. All raw and pro-cessed data obtained in this study can be downloaded in tabular formatfrom the web sites http://www.pasteur.fr/recherche/unites/Pmm/contents/gb/08_suppldata/suppldata.html or http://www.bichat.inserm.fr/equipes/Emi0339/publications_excel/pWR100transcriptome.xls

RESULTS AND DISCUSSION

Design of the array and overview of results

The VP pWR100 carries approximately 100 genes, notincluding those of insertion sequences (Buchrieser et al.,2000). Pairs of primers were designed for 75 of these plasmidgenes and four chromosomal genes, including tufB, toamplify PCR fragments of approximately the same size(500–600 bp) specific for each gene, except for ipaH1and ipaH2, ospE1 and ospE2, and ospC2, ospC3 and ospC4,which are identical or too similar to be distinguished byhybridization. PCR fragments for these latter genes weredesignated ipaH1/2, ospE1/2 and ospC2/3/4. PCR frag-ments were spotted onto nitrocellulose membranes andeach membrane was subjected to hybridizations using thelabelled genomic DNA as a probe. Spots corresponding to71 plasmid genes gave identical signals over all membranesand were considered in the transcriptional analysis pres-ented below. These genes are representative of most, if notall, transcriptional units of the VP. For all these genes,except ipaH1/2 and ospC2/3/4, signals obtained using thelabelled genomic DNA as a probe were not significantlydifferent, indicating that differences in the G+C contentof the various genes had no influence on the efficiencyof hybridization under these experimental conditions.Stronger signals were detected for ipaH1/2 and ospC2/3/4,consistent with the presence of two or three copies ofthese genes on the VP. For the transcriptional analysis,the intensity of each spot was standardized relative to themean value calculated over intensities obtained for thetufB spots, since expression of that chromosomal gene wassimilar in the wild-type strain and in its derivative thathad been cured of the VP (T. Le Gall and E. Denamur,unpublished observation). Similar results were obtainedusing sepA, a plasmid gene whose expression is not con-trolled by temperature (Benjelloun-Touimi et al., 1995),instead of tufB to standardize the data (data not shown).

We analysed the transcription profiles of 71 plasmid genesin the wild-type strain grown at 37 and 30 uC and in virF,virB, mxiE, ipaB, ipaB mxiE and orf81 mutants grown at37 uC. Fig. 1 shows the expression profiles of the wild-typestrain grown at either 30 uC (open bars) or 37 uC (lightshaded bars) and an ipaB mutant grown at 37 uC (darkshaded bars). Since signal intensities obtained upon hybri-dization of membranes with genomic DNA were similar(except for ipaH1/2 and ospC2/3/4 as indicated above), thecomparison of signal intensities obtained upon hybridiza-tion of cDNA suggests that genes for which the amountsof mRNA were the most abundant in the wild-type straingrown in broth at 37 uC are those of the entry region,

Expression level relative to the ET

Fig. 1. Schematic representation of expression levels ofplasmid genes in the wild-type strain grown at 30 6C (openbars) and 37 6C (light shaded bars) and in the ipaB mutantgrown at 37 6C (dark shaded bars). Genes are indicatedaccording to their position on the VP pWR100 (Buchrieseret al., 2000). Expression levels (in arbitrary units) are indicatedrelative to the expression threshold (ET). The ET was definedas the mean of signals obtained on non-spotted areas plus twostandard deviations.



especially the ipa operon, and virA. Under conditions ofderegulated secretion, such as in the ipaB mutant, strongsignals were also detected for ipaH and some osp genes.In most cases, genes either demonstrated or postulatedfrom the genetic organization to belong to the same operonexhibited signals of similar intensities. The few cases wheredifferent intensities were observed for genes belonging tothe same operon, e.g. ospB and phoN2 or within the mxioperon, might be due to different stabilities of parts ofthe mRNA resulting from mRNA processing. Expressionlevels of each gene for each experimental condition areshown in Fig. 2 along with the ratios calculated for pair-wise comparisons between a mutant and the wild-typestrain or between two mutants. Differences in gene expres-sion leading to a ratio >0?5 and <2 were not consideredas significant, although this threshold is arbitrary.

Genes regulated by the activity of secretion andMxiE

Previous analyses indicated that transcription of a numberof genes encoding substrates of the TTS apparatus isinduced when the TTS apparatus is deregulated in ipaBand ipaD mutants constructed either by insertion of anon-polar cassette in ipaB or ipaD or by integration of asuicide plasmid in ipaB (Demers et al., 1998; Mavris et al.,2002a, b). To determine the repertoire of genes whoseexpression is increased when the activity of secretion isderegulated, the transcription profile of a strain carryingthe ipaB4 allele (corresponding to the inactivation of ipaBby integration of a suicide plasmid) was compared withthat of the wild-type strain. As expected, expression of ipaC,ipaD and ipaA, which are located downstream from ipaB,was substantially decreased in this mutant compared withthe wild-type strain, as a consequence of the polar effectresulting from insertion of the suicide plasmid into ipaB.Genes whose expression was increased in the ipaB mutantcompared with the wild-type strain include ospB, phoN2,ospF, ipaH7, ipaH4, ospD3, ospC1, virA, ipaH9, ospG, ipaH1/2 and ospE1/2 (Fig. 2; ratio ipaB : w.t. 37 uC).

To determine whether the increased transcription of thesegenes in the ipaB mutant was dependent on MxiE, weanalysed the transcription profile of an ipaB mxiE mutantin which the activity of secretion is also deregulated butthat does not produce MxiE. Expression of all genes thatwere activated in the ipaBmutant was decreased in the ipaBmxiE mutant at a level that was similar to that observedin the wild-type strain (Fig. 2; ratio ipaB mxiE : w.t. 37 uC),confirming that activation of these genes in response to thederegulation of the secretion activity is MxiE dependent.

We also analysed the transcription profile of an mxiEmutant that lacks MxiE but in which the activity of secre-tion is not deregulated, i.e. the TTS apparatus is not activeduring growth of bacteria in broth. Expression of a fewgenes was slightly increased in the mxiE mutant comparedwith the wild-type (Fig. 2; ratio mxiE : w.t. 37 uC); however,the ratios for these genes ranged from 1?2 to 1?5. This

indicates that no genes are strongly controlled, eitherpositively or negatively, by MxiE under conditions of non-secretion. This is consistent with the previous observationthat the activity of MxiE is dependent on the presence ofthe co-activator IpgC, which is not available under con-ditions of non-secretion as it is titrated by IpaB and IpaC(Mavris et al., 2002a).

The cis-acting element required for increased transcriptionof ospC1, virA and ipaH9 promoters was shown to be a17 bp motif (59-GTATCGTTTTTTTANAG), designatedthe MxiE box, which is located between positions 249and 233 with respect to the transcription start site(Mavris et al., 2002b). The increased expression of ospF,ipaH4, ipaH7, ospC1, virA and ipaH9 detected in the presentanalysis is consistent with previous results obtained usinglacZ transcriptional fusions (Demers et al., 1998; Mavriset al., 2002b) and with the presence, in the 59 region ofthese genes, of a putative MxiE box exhibiting at least 14matches out of 16 positions with the MxiE box consensussequence (Fig. 3). The control of ospD3 transcription byMxiE in the absence of an MxiE box in the ospC1–ospD3intergenic region suggests that ospD3 is part of the sameoperon as ospC1, which has an MxiE box in its promoterregion. Similarly, control of ospG transcription by MxiEsuggests that ospG is expressed from the promoter ofthe upstream ipaH9 gene. There is evidence that ospE2 iscontrolled by MxiE (Kane et al., 2002), which is consistentwith the presence of an MxiE box upstream from ospE2.Since ipaH2 is separated from ospE2 by an insertionsequence, it seems unlikely that this gene is expressed, andthe signal detected on ipaH1/2 spots should correspondonly to expression of ipaH1. Since there is no MxiE boxin the ospE1–ipaH1 intergenic region, regulation of ipaH1transcription by the activity of secretion and MxiE sug-gests that transcription of ipaH1 is regulated by the MxiEbox that is present in the 59 region of ospE1. Expressionof ospB and phoN2, which belong to the same operon(Santapaola et al., 2002), was increased in the ipaB mutantand reduced to a wild-type level in the ipaB mxiE mutant.The DNA region located between positions 2106 and 289with respect to the ospB translation start site exhibits 13matches (out of 16 positions) with the sequence of theMxiE box and is likely to be involved in the MxiE-dependent activation of the ospB–phoN2 operon (Fig. 3).The VP carries two other regions that exhibit 13 matcheswith the sequence of the MxiE box, at coordinates 125 388and 144 574; however, these regions are unlikely to representfunctional MxiE boxes, since the former is located down-stream from and in the opposite orientation to virA andthe latter is located within spa47, whose expression, asthat of downstream genes, was not increased in the ipaBmutant (Fig. 2).

Using reporter plasmids in which the 59 region of a numberof genes encoding secreted proteins was driving expressionof the green fluorescent protein, Kane et al. (2002) showedthat ospB, ospF, ospC1, virA, ipaH9 and ospE2 promoters



Gene Expression levels Expression ratios

Name Map

w.t

. 37

˚Cip

aB

ipaB

mxi

E m

xiE

vir

B v

irF

w.t

. 30

˚C o

rf81

ipaBw.t. 37 ˚C

w.t. 30 ˚Cw.t. 37 ˚C w.t. 37 ˚C w.t. 37 ˚C w.t. 37 ˚C w.t. 37 ˚C w.t. 37 ˚C

ipaB mxiE mxiE virBvirB

virF virF orf81

Gene expression range colour code (relative to the expression threshold):

>1 ; 3> >3 ; 5> >5 ; 7> >7 ; 9> >9 ; 11> >11 ; 13> <1 ≥13



had been activated upon entry of bacteria into epithelialcells, as shown previously for ipaH1, ipaH4, ipaH7, ipaH9and virA genes (Demers et al., 1998), and that this activa-tion was dependent on the presence of MxiE. No MxiE-dependent expression of ospG was observed from thereporter plasmid (Kane et al., 2002), which is consistentwith the hypothesis presented above that ospG is expressedfrom the ipaH9 promoter.

In addition to the five ipaH genes carried by the VP dis-cussed above, other ipaH genes are carried by the chromo-some (Buysse et al., 1995). There are five chromosomalipaH genes in the S. flexneri strain M90T of serotype 5and we showed previously that at least some of thesegenes are controlled by MxiE and IpgC (Mavris et al.,2002b). The recent elucidation of the genomic sequenceof two S. flexneri strains of serotype 2a, CP301 and 2457T,revealed the presence of seven chromosomal ipaH genesin these strains, including four complete copies and three

copies inactivated by insertion sequences or frameshiftingmutations (Jin et al., 2002; Wei et al., 2003). Sequenceanalysis revealed the presence of an MxiE box upstreamfrom each chromosomal ipaH gene copy of strain CP301(Fig. 3) and strain 2457T (data not shown). Sequencesexhibiting two mismatches with the MxiE box consensussequence are also present upstream from genes encodingmannose-6-phosphate isomerase and L-aspartate oxidase;however, these sequence are unlikely to represent regula-tory sites as they are also present at the same position inthe E. coli strain MG1655, which does not contain a TTSsystem.

In conclusion, comparison of transcription profiles of thewild-type, mxiE, ipaB and ipaB mxiE strains indicates thatexpression of 13 genes (ipaH1, ipaH4, ipaH7, ipaH9, ospB,phoN2, ospC1, ospD3, ospE1, ospE2, ospF, ospG and virA)was activated under conditions of deregulated secretionin an MxiE-dependent manner. Transcription of operons

Fig. 3. Comparison of MxiE boxes carried by the virulence plasmid pWR100 and the chromosome of strain CP301.Sequences located upstream from the translation start site (ATG) of the various genes were aligned with respect to the regionexhibiting similarity with the MxiE box consensus sequence shown at the bottom. Genes ipaHa–e correspond to ORFsSF0722, SF2610, SF1383, SF1880 and SF2022, respectively, of S. flexneri strain CP301 (Jin et al., 2002). For the sake ofclarity, SF2202 and SF0887, which are respectively identical to SF2022 and SF2610, are not shown. In each case,nucleotides identical to those present in the MxiE box consensus sequence are shown in upper-case bold characters andnucleotides identical to those in the ”10 box consensus sequence are shown in upper-case. The nucleotides correspondingto the transcription start sites determined for ospC1, virA and ipaH9.8 (Mavris et al., 2002b) are underlined. The numbers ofnucleotides present between the determined or putative transcription start and the translational start are indicated.

Fig. 2. Expression profiles of the wild-type strain grown at 30 and 37 6C and virF, virB, ipaB, mxiE, orf81 and ipaB mxiE

mutants grown at 37 6C. Genes are listed in the left column according to their map position on pWR100 (Buchrieser et al.,2000). For each gene, the direction of transcription is indicated by the orientation of the corresponding arrowhead and theG+C content of that gene is indicated by the colour of the arrow head: red, below 40 mol%; blue, 40–50 mol%; green,above 50 mol%. Arrows indicate the direction and extent of potential operons, as suggested by the genetic organization. Theposition of insertion sequences is indicated by dashed black lines. Expression levels are indicated by colour-coded squares,using the code indicated at the bottom of the figure. The expression threshold was defined as the mean of signals obtained onnon-spotted areas plus two standard deviations. Ratios for salient pairwise comparisons were calculated for each gene andonly those for which the lowest possible Wilcoxon Z-test P value was obtained are shown. In these comparisons, expressionratios concerning the gene that was inactivated in a given mutant, e.g. the virB gene in the virB mutant, were not consideredand are replaced by a solidus (/).



carried by the entry region was not affected in mxiE, ipaBandmxiE ipaBmutants, indicating that neither the secretionactivity nor MxiE regulates expression of components ofthe TTS apparatus and translocators. Control of promoteractivity by MxiE correlates with the presence of nine MxiEboxes on the VP. MxiE boxes are also present in the 59region of chromosomal ipaH genes, consistent with theobservation that these genes are also controlled by MxiEand IpgC (Mavris et al., 2002b).

Genes controlled by VirB

VirB is necessary for expression of genes of the entryregion, as well as icsP, phoN2 and virA (Adler et al., 1989;Berlutti et al., 1998; Day & Maurelli, 2001; Santapaolaet al., 2002; Uchiya et al., 1995; Watanabe et al., 1990; Winget al., 2004). To determine whether other VP genes arecontrolled by VirB, we analysed the transcription profile ofa virB mutant. As expected, transcription of entry regiongenes, icsP, phoN2 and virA, was markedly decreased in thismutant compared with the wild-type strain. In addition,transcription of ospD2, ospF, orf13, ospD1, ospC2/3/4, ospC1,orf81, ipaJ, orf131b, orf137 and icsA was decreased in thismutant (Fig. 2; ratio virB : w.t. 37 uC). Since orf131b islocated immediately downstream from spa40, it is probablypart of the spa operon (Buchrieser et al., 2000; Sasakawaet al., 1993). The decreased expression of icsA in the virBmutant detected here is not consistent with the initialobservation that production of IcsA (VirG) was notaffected in a virB mutant, although a 2-fold reduction inthe amount of IcsA might not have been detected byimmunoblot analysis (Adler et al., 1989). Accordingly, thepresent analysis suggests that, in addition to genes of theentry region, icsP, phoN2 and virA, the repertoire of VirB-controlled genes includes ospD2, ospF, orf13, ospD1, ospC2/3/4, ospC1, orf81, ipaJ, orf137 and icsA.

Expression of phoN2, ospF, ospC1 and virA was decreasedin the virB mutant compared with the wild-type strain andwas increased in the ipaB mutant in an MxiE-dependentmanner (see above). The decreased expression of thesegenes in the virB mutant is not due to the decreased pro-duction of MxiE (whose expression was indeed decreasedin the virB mutant), since no differences in expression ofthese genes were observed between the mxiE mutant andthe wild-type strain (Fig. 2; ratio mxiE : w.t. 37 uC). Thisindicates that these genes are expressed in a VirB-dependent,MxiE-independent manner under conditions of non-secretion (i.e. in the wild-type and mxiE strains) and thattheir expression is increased in an MxiE-dependent mannerunder conditions of secretion (i.e. in the ipaB mutant).

Genes controlled by VirF

To investigate genes controlled by VirF, we compared thetranscription profile of a virF mutant with those of thewild-type and virB strains (Fig. 2; ratios virF : w.t. 37 uCand virF : virB). Expression of virB was markedly decreasedin the virF mutant compared with the wild-type strain and

expression of all genes that were less expressed in the virBmutant was also decreased in the virF mutant, consistentwith previous results (Adler et al., 1989; Porter & Dorman,2002; Sakai et al., 1988; Watanabe et al., 1990). In additionto virB (and genes that were less expressed in the virBmutant), genes that were less expressed in the virF mutantcompared with the wild-type strain included ipgB2, ipaH4,orf82, orf94, orf136, phoN1, orf163, repA and ospE1/2. In allthese cases, differences between the wild-type strain andthe virF mutant were at most 2-fold. The decreased expres-sion of icsA in the virF mutant is consistent with previousreports using an icsA–lacZ translational fusion that showeda 2?5-fold reduction in b-galactosidase activity producedby a virF mutant compared with the wild-type strain (Sakaiet al., 1988). A comparison of the virF and virB mutants(Fig. 2; ratio virF : virB) did not reveal any significant differ-ences between the expression profiles of these two strains.Since genes under the control of VirF should correspond tothose genes that are less expressed in the virF mutant(compared with the wild-type strain) and are not affectedin the virB mutant, this comparison suggested that onlyvirB is solely under the control of VirF.

Genes controlled by the temperature

To investigate which genes are controlled by the tempera-ture of growth, we determined the transcription profileof the wild-type strain grown at 30 uC (Fig. 2; ratio w.t.30 uC :w.t. 37 uC). Expression of virF was slightly decreasedat 30 uC, although the difference from the wild-type straingrown at 37 uC was less than 2-fold. Previous studies usingreporter genes and Northern blot analysis indicated thatthe range of regulation of virF expression by the growthtemperature is low, from 2- to 4-fold (Durand et al., 2000;Porter & Dorman, 1997; Prosseda et al., 2004). Other genesthat were less expressed in the wild-type strain grown at30 uC include all genes whose expressionwas decreased in thevirF mutant (at 37 uC), consistent with the weak expressionof VirF and VirB at 30 uC. In addition, ipaH7, ushA, orf169b,ipaH9, ospG, ipaH1/2 and orf212 were expressed slightlyless in the wild-type strain at 30 uC compared with thewild-type and virF strains at 37 uC. This suggests thatexpression of these genes might be slightly modulated bythe temperature of growth independently of VirF.

Genes controlled by Orf81

In addition to VirF and MxiE, the VP encodes a thirdprotein belonging to the AraC family of transcriptionalactivators, Orf81 (Buchrieser et al., 2000). As indicatedabove, expression of orf81 was decreased in the virF andvirB mutants, suggesting that Orf81 was produced by thewild-type strain and was potentially active during growthin broth. To determine which genes might be under thecontrol of Orf81, we compared the expression profilesof the wild-type strain and an orf81 mutant (Fig. 2; ratioorf81 : w.t. 37 uC). Expression of a few genes was slightlyincreased in the orf81 mutant; however, the differenceswere approximately 1?5-fold. This suggests that inactivation



of orf81 did not affect strongly, either positively or nega-tively, the expression of any VP gene. Since the activity oftranscriptional activators of the AraC family is oftenmodulated in response to external signals, it is possiblethat, in the wild-type strain, Orf81 was not active underthe growth conditions used in the present study. Alter-natively, orf81 might be a remnant of a regulatory circuitthat is no longer used by S. flexneri. This hypothesis isconsistent with the observation that orf81 is inactivated byan insertion sequence in an S. flexneri strain of serotype2a (Jin et al., 2002) and could not be amplified by PCR from astrain of serotype 6 (Lan et al., 2003).

Concluding remarks

The comparison of transcription profiles of most VP genesin the wild-type strain and in mutants for each of theknown regulatory genes carried by the VP presented hereprovides a general view of the repertoires of genes that areunder the control of each regulator. Data are remarkablyconsistent with previous observations, obtained using awide variety of methods, concerning particular genes thathad been shown to be controlled by VirF, VirB, MxiEand the activity of secretion. These results also extend thelist of genes that are controlled by VirB and MxiE and giveinformation on genes that are not affected in the variousmutants and, hence, on the specificity of each regulatorycircuit. Differences in gene expression leading to a ratio>0?5 and <2 were considered cautiously, although thisthreshold is arbitrary. Within these limits of interpreta-tion, it seems that only virB is under the direct control ofVirF and that no genes are controlled by Orf81 and MxiEin the wild-type strain growing in broth. In addition, itappears that a number of genes known to be involvedin virulence, including sepA, virK and msbB2, are notcontrolled by any of the VP-encoded regulators. Amongthe other genes that are not controlled by VP-encodedregulators or by temperature are genes whose products areinvolved in plasmid replication, maintenance or transfer,such as parB, stbA, ccdAB, mvpA, trbH, finO, repA, orf159band orf163.

Numerous genes located outside the entry region appearto be under the control of VirB (and indirectly of VirF).Many of these genes exhibit a low G+C content that issimilar to genes of the entry region, suggesting that theymight have the same origin as and possibly be part of theTTS system. Such genes include ospC1, ospC2/3/4, ospD1,ospD2, ospF and virA, the products of which are indeedsubstrates of the TTS apparatus, since they were identifiedby N-terminal sequencing of proteins secreted by asecretion-deregulated strain (or exhibited sequence simi-larity with a secreted protein) (Buchrieser et al., 2000).Other VirB-controlled genes that exhibit a low G+Ccontent include orf13, orf137, orf81 and ipaJ, the latterbeing located within the entry region, as well as icsP. Thetwo VirB-controlled genes that exhibit a G+C contenthigher than genes of the TTS system are phoN2 and icsA.Control of VirB over expression of these genes might be a

consequence of their localization on the VP, as phoN2 islocated downstream from and transcribed from the samepromoter as ospB (encoding a TTS substrate) and icsA islocated upstream from and is transcribed in the oppositeorientation to virA (also encoding a TTS substrate).

The mechanism by which VirB activates transcription isstill not completely understood. In vivo binding and oligo-merization of VirB on target DNA has been reported, aswell as DNA binding of the purified protein (Beloin et al.,2002; McKenna et al., 2003). A short DNA sequence (59-ATTTCAT-39) located upstream from the icsB promoterhas been shown to be required for activity of this promoterin vivo and for binding of a GST–VirB protein in vitro(Taniya et al., 2003). However, there are 56 occurrences ofthis sequence on pWR100, including within the numerousinsertion sequences that are scattered along the plasmid,and a similar motif with one mismatch is present morethan 1000 times on the plasmid, suggesting that thissequence alone does not define the VirB target. Althoughnot all genes with a low G+C content are controlled byVirB, promoter regions of genes that were less expressed inthe virB mutant exhibit a low G+C content. Accordingly,the possible constraint that this G+C content mightimpose on the local conformation of the DNA (e.g. bend-ing) might be part of the VirB recognition motif. It hasbeen proposed that activation of VirB-controlled promotersmight involve displacement of H-NS by VirB at thesepromoters (Beloin & Dorman, 2003). The present analy-sis, by further defining the repertoire of genes that are (orare not) under the control of VirB, might help futurestudies aimed at determining its mode of action.

The repertoire of genes that were induced under conditionsof deregulated secretion includes 13 genes. Activation ofall these genes in the ipaB mutant was dependent on MxiEand their regulated expression correlates with the presenceof an MxiE box in the 59 region of each regulated gene(or operon). Some of these genes, such as ospF, ospC1, ospB,phoN2 and virA, were expressed in a VirB-dependentmanner under conditions of non-secretion and induced inan MxiE-dependent manner under conditions of secre-tion. Other genes, such as ipaH genes, ospG and ospE1/2,were not expressed (or expressed at a low level that wasindependent of VirB) under conditions of non-secretionand were activated in an MxiE-dependent manner underconditions of secretion. Accordingly, substrates of the TTSapparatus might be classified into three categories on thebasis of their expression profiles (Fig. 4): (i) those that arecontrolled by VirB (i.e. their transcription was decreasedin the virB mutant and was not increased in the ipaBmutant), (ii) those that are controlled by MxiE (i.e. theirexpression was barely detectable in the wild-type strain,was not decreased in the virB mutant and was increased inthe ipaB mutant) and (iii) those that are controlled byboth VirB and MxiE (i.e. their expression was decreasedin the virB mutant and was increased in the ipaB mutant).In addition, it appears that homologous effectors do not



all belong to the same expression class. For example, ospD1and ospD2 are controlled by VirB, whereas ospD3 is con-trolled by MxiE. Likewise, expression of ospC2/3/4 iscontrolled by VirB whereas that of ospC1 is controlled byboth VirB and MxiE. Since effectors encoded by genescontrolled by VirB are expressed whether or not the TTSapparatus is active, these proteins are produced and pre-sumably stored in the cytoplasm prior to activation ofthe TTS apparatus and, therefore, can transit through theTTS apparatus immediately upon contact of bacteria witheukaryotic cells. In contrast, effectors encoded by genesthat are regulated by MxiE are produced only after acti-vation of the TTS apparatus and are likely to correspond toa second wave of effectors that are involved at a later stageor, possibly, in different cells. Effectors that are controlledby both VirB and MxiE are produced both before and afterthe TTS is activated and would be involved in bothsituations. Accordingly, the differential regulation of genesencoding TTS effectors is evidence that different effectorsmight be required at different times following contact ofbacteria with host cells. The function of each of theseeffectors, initially identified as substrates of the TTSapparatus, within eukaryotic cells and the cell type(s) inwhich these effectors have an important role remain to beidentified.

Regulation of the TTS apparatus activity by external signalsand control of the activity of some promoters by the TTSapparatus activity is not specific to the S. flexneri TTSsystem. In pathogenic Salmonella enterica, the pathogenicityisland 1 (SPI-1) encodes a TTS system very similar to theone encoded by the entry region of the S. flexneri VP.SPI-1 encodes several transcriptional activators, includingHilA (a member of the OmpR family) and InvF (a memberof the AraC family). HilA regulates transcription of genesencoding components of the TTS apparatus, whereas InvFcontrols transcription of genes encoding effectors locatedboth within and outside SPI-1 (Eichelberg & Galan, 1999).The activity of InvF is dependent on the co-activator SicA,which is homologous to IpgC (Darwin & Miller, 2000,2001). Modulation of the activity of InvF by the activityof secretion has not yet been documented, since there areno conditions known to modulate the activity of the TTSapparatus in Salmonella enterica. Nevertheless, the Salmo-nella enterica InvF/SicA and S. flexneri MxiE/IpgC systemsappear as clear variations on the same theme, despite the

fact that InvF, but not MxiE, controls transcription ofthe operon encoding its co-activator and translocators. Inpathogenic Yersinia species, a feedback regulation on tran-scription of yop genes by the TTS apparatus activity hasbeen extensively studied (Michiels et al., 1991; Petterssonet al., 1996). In contrast to the Shigella and Salmonellasystems, the VP-encoded TTS system of Yersinia speciescontains only one transcriptional activator, VirF (also amember of the AraC family) (Cornelis et al., 1998). Further-more, in Yersinia species, expression of genes encodingboth components of the TTS apparatus and effectors isincreased under conditions of active secretion (Lambertde Derouvroit et al., 1992; Rimpilainen et al., 1992). Themechanism by which the TTS apparatus activity is trans-mitted to VirF has not yet been elucidated, but also involvesTTS chaperones (Francis et al., 2001; Wulff-Strobel et al.,2002). Control of gene transcription by the TTS apparatusactivity is likely to be a general feature of TTS systems andis strong evidence that a temporal regulation of expressionof effectors is a key element in the function of TTS systems(Francis et al., 2002; Miller, 2002). Differences betweenregulatory circuits operating in each system might reflectthe specific constraints imposed by the strategy used byeach pathogen to infect its host(s).

ACKNOWLEDGEMENTS

We are pleased to acknowledge Giulia Nigro for help in strain con-struction, Elysabeth Couve and Jean-Yves Coppee for spotting PCRfragments at the Genopole of the Pasteur Institute, Lionel Frangeulfor sequence analysis and John Rohde for critical reading of themanuscript. This work was supported in part by grants from theItalian Ministero dell’Istruzione, Universita e Ricerca (COFIN 2002).

REFERENCES

Adler, B., Sasakawa, C., Tobe, T., Makino, S., Komatsu, K. &Yoshikawa, M. (1989). A dual transcriptional activation system forthe 230-kb plasmid genes coding for virulence-associated antigens ofShigella flexneri. Mol Microbiol 3, 627–635.

Allaoui, A., Mounier, J., Prevost, M. C., Sansonetti, P. J. & Parsot, C.(1992). icsB: a Shigella flexneri virulence gene necessary for thelysis of protrusions during intercellular spread. Mol Microbiol 6,1605–1616.

Bahrani, F. K., Sansonetti, P. J. & Parsot, C. (1997). Secretion ofIpa proteins by Shigella flexneri: inducer molecules and kinetics ofactivation. Infect Immun 65, 4005–4010.

Beloin, C. & Dorman, C. J. (2003). An extended role for the nucleoidstructuring protein H-NS in the virulence gene regulatory cascade ofShigella flexneri. Mol Microbiol 47, 825–838.

Beloin, C., McKenna, S. & Dorman, C. J. (2002). Molecular dissec-tion of VirB, a key regulator of the virulence cascade of Shigellaflexneri. J Biol Chem 277, 15333–15344.

Benjelloun-Touimi, Z., Sansonetti, P. J. & Parsot, C. (1995).SepA, the major extracellular protein of Shigella flexneri: autono-mous secretion and involvement in tissue invasion. Mol Microbiol17, 123–135.

Benjelloun-Touimi, Z., Si Tahar, M., Montecucco, C., Sansonetti,P. J. & Parsot, C. (1998). SepA, the 110 kDa protein secreted by

Fig. 4. Repertoires of TTS effector genes controlled by VirBand/or MxiE.



Shigella flexneri: two-domain structure and proteolytic activity.

Microbiology 144, 1815–1822.

Berlutti, F., Casalino, M., Zagaglia, C., Fradiani, P. A., Visca, P. &Nicoletti, M. (1998). Expression of the virulence plasmid-carried

apyrase gene (apy) of enteroinvasive Escherichia coli and Shigella

flexneri is under the control of H-NS and the VirF and VirB

regulatory cascade. Infect Immun 66, 4957–4964.

Bernardini, M. L., Mounier, J., d’Hauteville, H., Coquis-Rondon, M.& Sansonetti, P. J. (1989). Identification of IcsA, a plasmid locus

of Shigella flexneri that governs bacterial intra- and intercellular

spread through interaction with F-actin. Proc Natl Acad Sci U S A 86,

3867–3871.

Buchrieser, C., Glaser, P., Rusniok, C., Nedjari, H., d’Hauteville, H.,Kunst, F., Sansonetti, P. & Parsot, C. (2000). The virulence plasmid

pWR100 and the repertoire of proteins secreted by the type III

secretion apparatus of Shigella flexneri. Mol Microbiol 38, 760–771.

Buysse, J. M., Hartman, A. B., Strockbine, N. & Venkatesan, M.(1995). Genetic polymorphism of the ipaH multicopy antigen gene

in Shigella spps and enteroinvasive Escherichia coli. Microb Pathog 19,

335–349.

Cornelis, G. R., Boland, A., Boyd, A. P., Geuijen, C., Iriarte, M.,Neyt, C., Sory, M. P. & Stainier, I. (1998). The virulence plasmid of

Yersinia, an antihost genome. Microbiol Mol Biol Rev 62, 1315–1352.

Darwin, K. H. & Miller, V. L. (2000). The putative invasion protein

chaperone SicA acts together with InvF to activate the expression of

Salmonella typhimurium virulence genes. Mol Microbiol 35, 949–960.

Darwin, K. H. & Miller, V. L. (2001). Type III secretion chaperone-

dependent regulation: activation of virulence genes by SicA and

InvF in Salmonella typhimurium. EMBO J 20, 1850–1862.

Day, W. A., Jr & Maurelli, A. T. (2001). Shigella flexneri LuxS quorum-

sensing system modulates virB expression but is not essential for

virulence. Infect Immun 69, 15–23.

Demers, B., Sansonetti, P. J. & Parsot, C. (1998). Induction of type

III secretion in Shigella flexneri is associated with differential control

of transcription of genes encoding secreted proteins. EMBO J 17,

2894–2903.

d’Hauteville, H., Khan, S., Maskell, D. J. & 7 other authors (2002).Two msbB genes encoding maximal acylation of lipid A are required

for invasive Shigella flexneri to mediate inflammatory rupture and

destruction of the intestinal epithelium. J Immunol 168, 5240–5251.

Dorman, C. J. & Porter, M. E. (1998). The Shigella virulence gene

regulatory cascade: a paradigm of bacterial gene control mechanisms.

Mol Microbiol 29, 677–684.

Dorman, C. J., McKenna, S. & Beloin, C. (2001). Regulation of

virulence gene expression in Shigella flexneri, a facultative intra-

cellular pathogen. Int J Med Microbiol 291, 89–96.

Durand, J. M. B., Dagberg, B., Uhlin, B. E. & Bjork, G. R. (2000).Transfer RNA modification, temperature and DNA superhelicity

have a common target in the regulatory network of the virulence of

Shigella flexneri: the expression of the virF gene. Mol Microbiol 35,

924–935.

Egile, C., d’Hauteville, H., Parsot, C. & Sansonetti, P. J. (1997).SopA, the outer membrane protease responsible for polar localiza-

tion of IcsA in Shigella flexneri. Mol Microbiol 23, 1063–1073.

Eichelberg, K. & Galan, J. E. (1999). Differential regulation of

Salmonella typhimurium type III secreted proteins by pathogenicity

island 1 (SPI-1)-encoded transcriptional activators InvF and HilA.

Infect Immun 67, 4099–4105.

Francis, M. S., Lloyd, S. A. & Wolf-Watz, H. (2001). The type III

secretion chaperone LcrH co-operates with YopD to establish a

negative, regulatory loop for control of Yop synthesis in Yersinia

pseudotuberculosis. Mol Microbiol 42, 1075–1093.

Francis, M. S., Wolf-Watz, H. & Forsberg, A. (2002). Regulation of

type III secretion systems. Curr Opin Microbiol 5, 166–172.

Goldberg, M. B. & Theriot, J. A. (1995). Shigella flexneri surfaceprotein Icsa is sufficient to direct actin-based motility. Proc Natl

Acad Sci U S A 92, 6572–6576.

Hueck, C. J. (1998). Type III protein secretion systems in bac-

terial pathogens of animals and plants. Microbiol Mol Biol Rev 62,379–433.

Jin, Q., Yuan, Z., Xu, J. & 30 other authors (2002). Genome sequence

of Shigella flexneri 2a: insights into pathogenicity through compar-ison with genomes of Escherichia coli K12 and O157. Nucleic Acids

Res 30, 4432–4441.

Kane, C. D., Schuch, R., Day, W. A., Jr & Maurelli, A. T. (2002). MxiE

regulates intracellular expression of factors secreted by the Shigellaflexneri 2a type III secretion system. J Bacteriol 184, 4409–4419.

Lambert de Derouvroit, C., Sluiters, C. & Cornelis, G. R. (1992).Role of the transcriptional activator, VirF, and temperature in theexpression of the pYV plasmid genes of Yersinia enterocolitica. Mol

Microbiol 6, 395–409.

Lan, R. T., Stevenson, G. & Reeves, P. R. (2003). Comparison of

two major forms of the Shigella virulence plasmid pINV: positiveselection is a major force driving the divergence. Infect Immun 71,

6298–6306.

Lett, M. C., Sasakawa, C., Okada, N., Sakai, T., Makino, S.,Yamada, M., Komatsu, K. & Yoshikawa, M. (1989). virG, a plasmid-

coded virulence gene of Shigella flexneri: identification of the VirGprotein and determination of the complete coding sequence.

J Bacteriol 171, 353–359.

Maurelli, A. T. & Sansonetti, P. J. (1988). Identification of a chromo-

somal gene controlling temperature-regulated expression of Shigellavirulence. Proc Natl Acad Sci U S A 85, 2820–2824.

Maurelli, A. T., Blackmon, B. & Curtiss, R., III (1984). Temperature-

dependent expression of virulence genes in Shigella species. InfectImmun 43, 195–201.

Mavris, M., Page, A. L., Tournebize, R., Demers, B., Sansonetti, P.& Parsot, C. (2002a). Regulation of transcription by the activity of

the Shigella flexneri type III secretion apparatus. Mol Microbiol 43,1543–1553.

Mavris, M., Sansonetti, P. J. & Parsot, C. (2002b). Identification of

the cis-acting site involved in activation of promoters regulatedby activity of the type III secretion apparatus in Shigella flexneri.

J Bacteriol 184, 6751–6759.

McKenna, S., Beloin, C. & Dorman, C. J. (2003). In vitro DNA-

binding properties of VirB, the Shigella flexneri virulence regulatoryprotein. FEBS Lett 545, 183–187.

Menard, R., Sansonetti, P. & Parsot, C. (1994a). The secretion of

the Shigella flexneri Ipa invasins is activated by epithelial cells andcontrolled by IpaB and IpaD. EMBO J 13, 5293–5302.

Menard, R., Sansonetti, P., Parsot, C. & Vasselon, T. (1994b).Extracellular association and cytoplasmic partitioning of the IpaB

and IpaC invasins of S. flexneri. Cell 79, 515–525.

Michiels, T., Vanooteghem, J. C., Lambert de Derouvroit, C.,China, B., Gustin, A., Boudry, P. & Cornelis, G. R. (1991). Analysis ofvirC, an operon involved in the secretion of Yop proteins by Yersiniaenterocolitica. J Bacteriol 173, 4994–5009.

Miller, V. L. (2002). Connections between transcriptional regulationand type III secretion? Curr Opin Microbiol 5, 211–215.

Nakata, N., Sasakawa, C., Okada, N., Tobe, T., Fukuda, I., Suzuki, T.,Komatsu, K. & Yoshikawa, M. (1992). Identification and character-ization of virK, a virulence-associated large plasmid gene essential

for intercellular spreading of Shigella flexneri. Mol Microbiol 6,2387–2395.



Niebuhr, K., Jouihri, N., Allaoui, A., Gounon, P., Sansonetti, P. J. &

Parsot, C. (2000). IpgD, a protein secreted by the type III secretion

machinery of Shigella flexneri, is chaperoned by IpgE and implicated

in entry focus formation. Mol Microbiol 38, 8–19.

Ogawa, M., Suzuki, T., Tatsuno, I., Abe, H. & Sasakawa, C. (2003).IcsB, secreted via the type III secretion system, is chaperoned by

IpgA and required at the post-invasion stage of Shigella patho-

genicity. Mol Microbiol 48, 913–931.

Page, A. L., Fromont-Racine, M., Sansonetti, P., Legrain, P. &Parsot, C. (2001). Characterization of the interaction partners of

secreted proteins and chaperones of Shigella flexneri. Mol Microbiol

42, 1133–1145.

Page, A. L., Sansonetti, P. & Parsot, C. (2002). Spa15 of Shigella

flexneri, a third type of chaperone in the type III secretion pathway.

Mol Microbiol 43, 1533–1542.

Parsot, C., Menard, R., Gounon, P. & Sansonetti, P. J. (1995).Enhanced secretion through the Shigella flexneri Mxi-Spa translocon

leads to assembly of extracellular proteins into macromolecular

structures. Mol Microbiol 16, 291–300.

Pettersson, J., Nordfelth, R., Dubinina, E., Bergman, T.,Gustafsson, M., Magnusson, K. E. & Wolf-Watz, H. (1996).

Modulation of virulence factor expression by pathogen target cell

contact. Science 273, 1231–1233.

Porter, M. E. & Dorman, C. J. (1997). Differential regulation of the

plasmid-encoded genes in the Shigella flexneri virulence regulon. Mol

Gen Genet 256, 93–103.

Porter, M. E. & Dorman, C. J. (2002). In vivo DNA-binding and

oligomerization properties of the Shigella flexneri AraC-like trans-

criptional regulator VirF as identified by random and site-specific

mutagenesis. J Bacteriol 184, 531–539.

Prosseda, G., Fradiani, P. A., Di Lorenzo, M., Falconi, M., Micheli, G.,Casalino, M., Nicoletti, M. & Colonna, B. (1998). A role for H-NS

in the regulation of the virF gene of Shigella and enteroinvasive

Escherichia coli. Res Microbiol 149, 15–25.

Prosseda, G., Falconi, M., Giangrossi, M., Gualerzi, C. O., Micheli, G.

& Colonna, B. (2004). The virF promoter in Shigella: more than just

a curved DNA stretch. Mol Microbiol 51, 523–537.

Rimpilainen, M., Forsberg, A. & Wolfwatz, H. (1992). A novel

protein, LcrQ, involved in the low-calcium response of Yersinia

pseudotuberculosis shows extensive homology to YopH. J Bacteriol

174, 3355–3363.

Sakai, T., Sasakawa, C., Makino, S. & Yoshikawa, M. (1986). DNAsequence and product analysis of the virfF locus responsible for

Congo red binding and cell invasion in Shigella flexneri 2a. Infect

Immun 54, 395–402.

Sakai, T., Sasakawa, C. & Yoshikawa, M. (1988). Expression of four

virulence antigens of Shigella flexneri is positively regulated at the

transcriptional level by the 30 kilodalton VirF protein. Mol Microbiol

2, 589–597.

Santapaola, D., Casalino, M., Petrucca, A., Presutti, C., Zagaglia, C.,Berlutti, F., Colonna, B. & Nicoletti, M. (2002). EnteroinvasiveEscherichia coli virulence-plasmid-carried apyrase (apy) and ospBgenes are organized as a bicistronic operon and are subject todifferential expression. Microbiology 148, 2519–2529.

Sasakawa, C., Komatsu, K., Tobe, T., Suzuki, T. & Yoshikawa, M.(1993). Eight genes in region 5 that form an operon are essential forinvasion of epithelial cells by Shigella flexneri 2a. J Bacteriol 175,2334–2346.

Shere, K. D., Sallustio, S., Manessis, A., D’Aversa, T. G. & Goldberg,M. B. (1997). Disruption of IcsP, the major Shigella protease thatcleaves IcsA, accelerates actin-based motility. Mol Microbiol 25,451–462.

Taniya, T., Mitobe, J., Nakayama, S., Mingshan, Q., Okuda, K. &Watanabe, H. (2003). Determination of the InvE binding site requiredfor expression of IpaB of the Shigella sonnei virulence plasmid:involvement of a ParB boxA-like sequence. J Bacteriol 185, 5158–5165.

Tobe, T., Nagai, S., Okada, N., Adler, B., Yoshikawa, M. &Sasakawa, C. (1991). Temperature-regulated expression of invasiongenes in Shigella flexneri is controlled through the transcriptionalactivation of the virB gene on the large plasmid. Mol Microbiol 5,887–893.

Tobe, T., Yoshikawa, M., Mizuno, T. & Sasakawa, C. (1993). Tran-scriptional control of the invasion regulatory gene virB of Shigellaflexneri: activation by Virf and repression by H-NS. J Bacteriol 175,6142–6149.

Uchiya, K., Tobe, T., Komatsu, K., Suzuki, T., Watarai, M., Fukuda, I.,Yoshikawa, M. & Sasakawa, C. (1995). Identification of a novelvirulence gene, virA, on the large plasmid of Shigella, involved ininvasion and intercellular spreading. Mol Microbiol 17, 241–250.

Venkatesan, M. M., Goldberg, M. B., Rose, D. J., Grotbeck, E. J.,Burland, V. & Blattner, F. R. (2001). Complete DNA sequence andanalysis of the large virulence plasmid of Shigella flexneri. InfectImmun 69, 3271–3285.

Watanabe, H., Arakawa, E., Ito, K. I., Kato, J. I. & Nakamura, A.(1990). Genetic analysis of an invasion region by use of a Tn3-lactransposon and identification of a second positive regulator gene,invE, for cell invasion of Shigella sonnei: significant homology ofInvE with ParB of plasmid P1. J Bacteriol 172, 619–629.

Wei, J., Goldberg, M. B., Burland, V. & 14 other authors (2003).Complete genome sequence and comparative genomics of Shigellaflexneri serotype 2a strain 2457T. Infect Immun 71, 2775–2786.

Wing, H. J., Yan, A. W., Goldman, S. R. & Goldberg, M. B. (2004).Regulation of IcsP, the outer membrane protease of the Shigella actintail assembly protein IcsA, by virulence plasmid regulators VirF andVirB. J Bacteriol 186, 699–705.

Wulff-Strobel, C. R., Williams, A. W. & Straley, S. C. (2002). LcrQand SycH function together at the Ysc type III secretion system inYersinia pestis to impose a hierarchy of secretion. Mol Microbiol 43,411–423.



Documents

Analysis of virulence plasmid gene expression deﬁnes three ...€¦ · Analysis of virulence plasmid gene expression deﬁnes three classes of effectors in the type III secretion