Shigella IpaD has a dual role: signal transduction from the type III secretion system needle tip and intracellular secretion regulation.

1.4 MB·English

by  Roehrich, A Dorothea

MolecularMicrobiology(2013)87(3),690–706 (cid:2) doi:10.1111/mmi.12124
Firstpublishedonline11January2013
Shigella IpaD has a dual role: signal transduction from the
type III secretion system needle tip and intracellular
secretion regulation
A.DorotheaRoehrich,1EnoraGuillossou,1 Introduction
ArielJ.Blocker1,2*andIsabelMartinez-Argudo1†
TypeIIIsecretionsystems(T3SSs)arecommonvirulence
1SchoolofCellular&MolecularMedicine,Universityof
factors among Gram-negative bacteria. They are protein
Bristol,UniversityWalk,BristolBS81TD,UK.
transport devices used for injecting effector proteins into
2SchoolofBiochemistry,UniversityofBristol,University
host cells to modulate their responses in favour of the
Walk,BristolBS81TD,UK.
bacterium (Cornelis, 2006). To study conserved aspects
of type III secretion, we use Shigella flexneri as a model
Summary system. It is the causative agent of human bacillary dys-
entery, a form of inflammatory diarrhoea, and uses the
Type III secretion systems (T3SSs) are protein injec-
T3SStopromoteitsinvasionintogutepithelialcells.The
tion devices essential for the interaction of many
T3SS consists of a basal body spanning both bacterial
Gram-negative bacteria with eukaryotic cells. While
membranesandahollowextracellularneedleservingas
ShigellaassemblesitsT3SSwhentheenvironmental
secretionchannel.Thedistalendoftheneedleistopped
conditions are appropriate for invasion, secretion is
by the tip complex, which is required for pore formation
onlyactivatedafterphysicalcontactwithahostcell.
and protein translocation into host cells. In Shigella, the
First,thetranslocatorsaresecretedtoformaporein
apparatus is assembled when the environmental condi-
the host cell membrane, followed by effectors which
tionsareappropriateforinvasion,butsecretionisblocked
manipulatethehostcell.Secretionactivationistightly
until physical contact with a host cell generates an acti-
controlled by conserved T3SS components: the
vationsignal(reviewedinSchroederandHilbi,2008).
needletipproteinsIpaDandIpaB,theneedleitselfand
Proteintranslocationintohostcellsrequiresthreepro-
the intracellular gatekeeper protein MxiC. To further
teins known as translocators (IpaB, IpaC and IpaD in
characterize the role of IpaD during activation, we
Shigella). Of these, IpaD is hydrophilic and constitutively
combinedrandommutagenesiswithageneticscreen
present atop mature needles (Espina etal., 2006; Sani
to identify ipaD mutant strains unable to respond to
etal., 2007; Veenendaal etal., 2007). It probably serves
host cell contact. Class II mutants have an overall
asascaffoldforthemostlyhydrophobicproteinsIpaBand
defect in secretion induction. They map to IpaD’s
IpaC,whichlaterformtheporeinthehostcellmembrane.
C-terminalhelixandlikelyaffectactivationsignalgen-
IpaBisalsoconstitutivelypresentatneedletips,whilethe
eration or transmission. The Class I mutant secretes
other hydrophobic translocator, IpaC, is only recruited to
translocatorsprematurelyandisspecificallydefective
the needle tip upon activation (Veenendaal etal., 2007;
in IpaD secretion upon activation. A phenotypically
Roehrich etal., 2010; Shen etal., 2010). In Yersinia, the
equivalent mutant was found in mxiC. We show that
needle tip is formed by the hydrophilic translocator LcrV:
IpaDandMxiCactinthesameintracellularpathway.In
in this species no hydrophobic translocator has been
summary, we demonstrate that IpaD has a dual role
found constitutively present atop needles (Mueller etal.,
and acts at two distinct locations during secretion
2005;2008;Blockeretal.,2008).
activation.
TypeIIIsecretionisatightlyregulatedprocessinitiated
only after host cell contact. As type III proteins are
expressedwhentheenvironmentalconditionsareappro-
priateforinvasion,secretionhastobepreventedpriorto
Accepted 4 December, 2012. *For correspondence. E-mail ariel.
[email protected]; Tel. [phone]; Fax [phone] cell contact. Recently, we proposed a model featuring
091.†Presentaddress:ÁreadeGenética,FacultaddeCienciasAmbi- gatingmechanismslocatedinthecytoplasmthatprevent
entalesyBioquímica,CampusTecnológicodelaFábricadeArmas,
secretion in the absence of activation signals (Martinez-
Universidad de Castilla-La Mancha,Avda. Carlos III, s/n, E-45071,
Toledo,Spain. Argudo and Blocker, 2010). We hypothesized that under
©2013BlackwellPublishingLtd
CharacterizationofipaDandmxiCpointmutants 691
non-activated conditions, a mechanism that represses helicesofthecoiledcoil(Johnsonetal.,2007;Veenendaal
translocatorsecretion,directlyorindirectlyinvolvingIpaD etal., 2007). The C-terminal globular domain is a site of
and/or IpaB, is put in place. In addition, the gatekeeper interactionwithIpaB,whiletheN-terminalglobulardomain
protein MxiC prevents premature effector secretion (Bot- of IpaD has been proposed to act as a self-chaperone
teaux etal., 2009; Martinez-Argudo and Blocker, 2010) (Johnsonetal.,2007).IntheLcrVclassofhydrophilictip
anditrequiresasecretionsignaltodoso(Botteauxetal., proteins,thechaperoneisformedbyaseparatepolypep-
2009). After reception of the activation signal, such tidecalledLcrG(Lawtonetal.,2002).
repression mechanisms would be counteracted and TheneedletipwasfirstvisualizedintheYersiniaT3SS
secretion happens in an ordered way: translocator pro- (Mueller etal., 2005). We proposed that the tip contains
teins are secreted first, then MxiC (Kenjale etal., 2005; fourmoleculesofIpaDandonemoleculeofIpaB(Johnson
Martinez-ArgudoandBlocker,2010)followedbyearlyand etal.,2007;Veenendaaletal.,2007;Blockeretal.,2008).
late effectors (Mavris etal., 2002; Parsot, 2009). As a The crystal structure of IpaD allowed construction of a
DmxiC mutant shows weak and delayed secretion of pentamer based on dimer contacts in one of the crystal
translocators in response to activation when compared formswithahelicalrisesimilartothatoftheMxiHneedle
with the wild-type strain, we suggested that, in Shigella, (Cordesetal.,2003;Johnsonetal.,2007).Intheproposed
MxiCalsoplaysaroleinactivatingtranslocatorsecretion IpaD pentamer, the position between the fourth and first
(Martinez-ArgudoandBlocker,2010). IpaDmoleculesisstructurallydifferentfromallotherposi-
MxiCbelongstoafamilyofT3SSproteinsthatincludes tions. This final position might thus be filled by IpaB
Yersinia YopN/TyeA, enteropathogenic Escherichia coli (Johnson etal., 2007; Veenendaal etal., 2007; Blocker
SepLandSalmonellaInvEandSsaL(Pallenetal.,2005). etal.,2008).Themodelputstheproposedcentralcoiled
Besides similar sequences, the proteins have conserved coil and the C-terminal globular domain of IpaB in an
domain topologies and structures (Schubot etal., 2005; equivalentconformationtothatofIpaD,positioningIpaB’s
Deaneetal.,2008;2010)andallblockeffectorsecretionin domainhomologoustopore-formingcolicinsanditsputa-
theabsenceofanactivationsignal.However,mutantsdo tive transmembrane helices in the ideal place to interact
notshowthesamephenotypewithrespecttothecontrolof withthehostcellmembrane(Johnsonetal.,2007;Barta
translocatorsecretion.InDinvE,DssaL,DsepLandDmxiC etal., 2012a). Together with mutagenesis data this sug-
mutants,secretionoftranslocatorsisimpaired(Kuboriand gests that IpaB could be the host cell sensor and that a
Galan,2002;Coombesetal.,2004;O’Connelletal.,2004; conformationalchangeinthetipcomplexmightbesignal-
Deng etal., 2005), whereas it is constitutive in DyopN or ling event triggering T3SS activation (Veenendaal etal.,
DtyeA mutants (Iriarte etal., 1998; Ferracci etal., 2005). 2007;Roehrichetal.,2010;Shenetal.,2010).
How the gatekeepers perform their tasks is unclear, but Congored(CR)isasmallamphipathicdyemoleculethat
Salmonella InvE interacts with chaperone–translocator wasoriginallyusedtoassessinfectivityofShigellastrains:
complexes (Kubori and Galan, 2002) and Shigella MxiC only colonies binding CR on agar plates were infective,
interactswiththeATPaseinvitro(Botteauxetal.,2009). while ‘white’ colonies were not (Payne and Finkelstein,
Despiteintenseresearchwestilldonotunderstandhow 1977).ItwaslaterfoundthatCRspecificallyactivatestype
theactivationsignalgeneratedaftercontactwiththehost IIIsecretion(Parsotetal.,1995)andthatitstargetislikely
cell is transduced to the cytoplasmic side of theT3SS to in the tip complex (Veenendaal etal., 2007). All T3SS
initiatesecretion.ThefirstpartoftheT3SSthatestablishes component-knockout(KO)strainsthatarenotinducibleby
contactwiththehostcellisprobablytheneedletip.There- CRarealsonon-invasive,suggestingthatCRandhostcell
fore, it is likely that the tip complex senses the host cell. contactinduceT3SSactivationusingarelatedmechanism
ExperimentaldatafromtheShigellaandYersiniasystems (Menardetal.,1993;1994;Blockeretal.,2001).
suggestthattheneedleisdirectlyinvolvedintransmitting To test our hypothesis that, as part of the tip complex,
the activation signal. Particular single amino acid muta- IpaD is involved in conformational changes crucial for
tionsintheneedleproteinleadtoalteredsecretionpheno- signal transduction, we sought to identify ipaD point
types and tip complex composition (Kenjale etal., 2005; mutantsunabletorespondtohostcellcontact.Foreaseof
Torruellasetal.,2005;Veenendaaletal.,2007). analysis,wechosetosetupageneticscreenformutants
TheShigellaneedleproteinMxiHisessentiallyacoiled insensitivetoinductionbyCR.Allmutantsisolatedshowed
coilhairpinthatpolymerizesintothehelicalneedleusing a wild-type-like tip complex composition. Yet, six out of
bothofitstermini(Cordesetal.,2003;Kenjaleetal.,2005; seven mutants were impaired in host cell sensing and
Deaneetal.,2006;Fujiietal.,2012;Loquetetal.,2012). inductionoftypeIIIsecretion.Thepositionofthesemuta-
SimilartotheneedleproteinMxiH,IpaDcontainsalonger tions highlights the importance of the C-terminal helix of
centralcoiledcoilandrequiresitsextremeC-terminusfor IpaD in signal transduction. The remaining ipaD mutant
binding to the needle. In addition, it has two globular wasunabletopreventsecretionundernon-inducingcon-
domains, one at the N-terminus and one linking the two ditions. Parallel analysis of a mxiC mutant with a similar
©2013BlackwellPublishingLtd,MolecularMicrobiology,87,690–706
692 A.D.Roehrich,E.Guillossou,A.J.BlockerandI.Martinez-Argudo (cid:2)
Table1. ipaDmutantsisolatedinthisstudyandtheirphenotypes.
Phenotypicproperties
Timesthemutant
Class ipaDmutation wasfound Leakage CRinduction IpaDexpression Needletipcomposition Invasion
– Wild-type – WT WT WT WT WT
– Knockout – ++ (–)a NA –– ––
I L99P,N255D 1 + (WT)b – WT WT
IIa N186Y 1 WT –– WT WT –
IIa N273I,Q277L 1 WT –– WT WT –
IIa K291I(orTorN) 7c WT –– WT WT –
IIb N292Y 1 WT – WT WT WT
IIb T296I 11c WT – WT WT –
IIb Y104L,Q299L 1 WT – WT WT –
Themutationresponsiblefortheobservedphenotypeisindicatedinboldletters.‘WT’indicateswild-type-likeproperties,+or++anincreaseand
–or––adecrease.‘NA’meansnotapplicable.
a. DipaDisnotinduciblebyCR,butneverthelessproteinsaredetectableasitisafastconstitutivesecreter.
b. MutantL99PonlyhasadefectinIpaDsecretionafterCRinduction.
c. MutantsK291I/T/NandT296Iwerefoundindependentlyintwoandthreeseparatelibrariesrespectively.
phenotype further supports the regulatory roles of both containingCR.Wild-typeShigellacoloniesstainlightred
IpaD and MxiC in translocator secretion.Taken together, onCR-containingplates,whilemutantsunabletosecrete
these data demonstrate that IpaD has a dual role during effectors are white (Maurelli etal., 1984). Constitutive
secretionactivation. secretor mutant colonies, such as those containing bac-
teria that have lost their needle tip, are bright red on CR
Results plates,probablybecauseofthesecretionoflateeffectors
(Parsot etal., 1995). Upon random PCR mutagenesis of
FortheShigellaT3SS,‘induction’describestheburstofIpa
theipaDgene,themutantDNAlibrarieswereclonedand
protein secretion upon host cell sensing (Menard etal.,
amplifiedasdescribedinExperimentalprocedures.After
1994)oradditionofCR,asmallamphipathicdyemolecule
transformation of the mutant libraries into the Shigella
which acts as an artificial inducer of its T3SS (Bahrani
DipaD mutant, we screened for white colonies in plates
etal.,1997).Inductionisaveryfasteventasaround50%
containing CR (Fig.S1). Around 500000 transformants
of the Ipa proteins are secreted at 37°C in the 15min
were screened and as expected a substantial proportion
following CR addition (Bahrani etal., 1997; Magdalena
ofthemwerebrightred(10–20%,indicatingexpressionof
etal.,2002;Parsotetal.,2005).Priortoinductionthereis
non-functionalipaDalleles).Onehundredandten‘white’
a slow, low level of Ipa protein secretion known as
colonies were selected and their plasmids isolated and
‘leakage’, where around 5% of Ipa proteins are secreted
retransformed into the DipaD mutant to ensure the white
(Magdalenaetal.,2002;Parsotetal.,2005).‘Constitutive
colour was not due to loss-of-function mutations else-
secretion’of Ipa proteins represents an unregulated and
wherewithintheT3SS-encodingoperons.
higher-levelsecretionthanleakageandinvolvesnotonly
Twenty-two plasmids were confirmed as giving rise
the Ipa proteins but also the ‘late effectors’, which are
to white colonies and their ipaD gene was sequenced
involvedinlaterstagesofinfection.IntheDipaBandDipaD
(Table1). Mutant ipaD(L99P, N255D) was not white but
mutant strains, constitutive secretion is very fast and
slightly more red than wild-type when retransformed into
detectable in minutes. It is thus named ‘fast constitutive
theDipaDmutant;however,itshowedaninterestingphe-
secretion’(Veenendaaletal.,2007).Werecentlyshowed
notype(seebelow)andwasfurtherinvestigated.Incases
thatconstitutivesecretionhasadifferentcauseinaDipaB
where more than one mutation was found in the ipaD
versus a ipaBD3 mutant (Martinez-Argudo and Blocker,
gene, individual mutations were cloned independently in
2010).Asthelattermutantisinaconstitutivelyactivated
order to identify which was responsible for the mutant
state,wenowcallitsphenotype‘prematuresecretion’.
phenotype and only the mutation generating the mutant
phenotypewasfurtherinvestigated.
IsolationofipaDmutantsaffectedinsecretion
TotestourhypothesisthatIpaDisinvolvedintherecep-
ipaDmutantstrainshavedifferentiallyaltered
tionortransductionoftheactivationsignal,wesoughtto
secretionprofiles
identify ipaD mutant strains that are unable to secrete
after induction by CR. For this, we used a screening TocharacterizetheipaDmutants,weanalysedtheirsecre-
methodbasedonthecolourofShigellacoloniesinplates tionprofileduringexponentialgrowthinnon-inducingcon-
©2013BlackwellPublishingLtd,MolecularMicrobiology,87,690–706
CharacterizationofipaDandmxiCpointmutants 693
Fig.1. AllipaDmutantsisolatedcausepartialdefectsinsecretion.
A.ExponentialleakageofShigellawild-type,DipaDmutant,complementedstrain(DipaD/ipaD+)andipaDmutants(intheDipaDbackground)
isolatedfromamutantlibrary.SampleswerecollectedasdescribedinExperimentalproceduresandSilver-stained.
B.ProteinsecretioninresponsetoartificialinducerCongored(CR)induction.SampleswerecollectedasdescribedinExperimental
proceduresandSilver-stained(toppanel)andWestern-blottedwithantibodiesagainstthetranslocatorproteinsIpaB,IpaC,IpaDandthe
regulatorMxiC(bottompanels).
C.Totalproteinexpressionlevels.SampleswerecollectedasdescribedinExperimentalproceduresandWestern-blottedwiththeindicated
antibodies.
Resultsshownarerepresentativeofatleasttwoindependentexperiments.
ditions(Fig.1A)andfollowingCRinduction(Fig.1B).We mented strain (Fig.1B). In this 15min assay, increased
found two novel types of secretion profile. First, mutant secretionundernon-inducingconditionsisonlyvisibleifit
ipaDL99PsecretedamuchhigherlevelofT3Sproteinsin is ‘fast’as for example in the DipaD mutant. In a second
non-inducing conditions when compared to the DipaD/ groupofmutants(ipaDN186Y,ipaDK291andipaD(N273I,
ipaD+ complemented strain, but still less than the DipaD Q277L)),inductionofsecretionwasimpaired:incompari-
mutant (Fig.1A, c. 4h incubation). Additionally, mutant sontotheDipaD/ipaD+complementedstraintheysecreted
ipaDL99P showed a defect in IpaD secretion (approxi- only approximately 40%, 20%, 10% and 10% of IpaB,
mately 20% of the complemented strain) following CR IpaC, IpaD and MxiC respectively. Mutants ipaDN292Y,
induction,whilesecretionofIpaB,IpaC,MxiCandeffectors ipaDT296I and ipaDQ299L showed a similar secretion
was still induced to levels comparable with the comple- pattern, although the defects in secretion were less pro-
©2013BlackwellPublishingLtd,MolecularMicrobiology,87,690–706
694 A.D.Roehrich,E.Guillossou,A.J.BlockerandI.Martinez-Argudo (cid:2)
Fig.2. TheipaDmutantsshowdefectsinHeLainvasionbuthavenormalneedletipcompositions.
A.InvasionofHeLacellswasmeasuredwithagentamicinprotectionassayasdescribedinExperimentalprocedures.Everyexperimentwas
normalizedtowild-typeandthedataareaveragesofatleastthreeindependentexperimentsperformedatleastinduplicate.Errorbars
indicatethestandarddeviation.WheretheinvasionissignificantlydifferentfromthecomplementedstrainDipaD/ipaD+(P<0.001,ANOVAwith
Tukey’sposthoctest),ipaDmutantstrains(intheDipaDbackground)aremarkedwith***.
B.Needlesfromwild-type,DipaDmutant,complementedstrain(DipaD/ipaD+)andipaDmutants(intheDipaDbackground)overexpressingthe
needleproteinMxiHwereisolatedasdescribedinExperimentalproceduresandWestern-blottedwithantibodiesagainstIpaBandIpaD
(sampleswerenormalizedfortheamountofMxiHpresentinpreparationsasdetectedbySilverstain,bottompanel).Resultsshownare
representativeofatleasttwoindependentexperiments.Forthecomplementedstrain(DipaD/ipaD+)resultsfromtwoexperimentsareshownto
indicateexperimentalvariability.Incomparisontothecomplementedstrain,thewild-typecontainsc.50(cid:2)10%IpaD(average(cid:2)standard
deviation),ipaDL99Pcontainsc.70(cid:2)30%IpaD,ClassIIamutantscontainc.90(cid:2)20%IpaDandClassIIbmutantscontainc.90(cid:2)40%
IpaD.ForIpaB,thesevaluesare70(cid:2)40%forwild-type,100(cid:2)70%foripaDL99P,150(cid:2)100%forClassIIamutantsand130(cid:2)90%for
ClassIIbmutants.TherearenosignificantdifferencesbetweensamplesinanANOVA(P>0.05)forbothIpaDandIpaB.
C.TotalproteinexpressionlevelsinovernightculturesfromstrainsoverexpressingtheneedleproteinasinB.Sampleswerecollectedas
describedinExperimentalproceduresandWestern-blottedwithantibodiesagainstIpaBandIpaD.
nounced(approximately80%,60%,40%and20%ofIpaB, stitutive secretion has a different cause in ipaDL99P
IpaC, IpaD and MxiC respectively; Fig.1B). We named (Fig.S2).Therefore,allIpaDmutantsisolatedarepartially
theseClassesI,IIaandIIbrespectively,startingthenum- defectiveinsecretionregulation.
beringwiththemostN-terminalmutation.Whenweana-
lysedtheexpressionoftheIpaDmutantproteins,thelevel
AllIpaDmutantproteinsaresecretedinaconstitutive
of the mutant proteins was similar to the complemented
secretorbackground
strain with the exception of the IpaDL99P, which was
presentataround60%ofthewild-typeallele(Fig.1C).As AsallthenewlyisolatedipaDmutantsshowedadefectin
deletion of ipaD leads to constitutive secretion, we were IpaDsecretion,wewonderedifthemutationsmadethem
concernedthattheincreasedsecretioninipaDL99Punder ‘unfit’ for secretion, for example by affecting their con-
non-inducing conditions might be due to a decrease in formation in the cytoplasm. To test this possibility we
IpaDexpression.However,theipaDL99Pmutantsecreted checkediftheIpaDmutantsweresecretedinaconstitu-
moreIpaBandIpaCthanastrainexpressingmuchlower tivesecretorbackground.Weandothershavepreviously
levels of the wild-type IpaD protein, indicating that con- shown that the translocator proteins IpaC and IpaD are
©2013BlackwellPublishingLtd,MolecularMicrobiology,87,690–706
CharacterizationofipaDandmxiCpointmutants 695
Fig.3. ClassIIipaDmutantsareseverely
affectedininductionofsecretion.
A.Analysisofproteinssecretedbythe
indicatedstrainsinresponsetoCongored
(CR)induction.SamplesfromShigella
wild-type,DipaDmutant,complementedstrain
(DipaD/ipaD+)andipaDmutants(intheDipaD
background)werecollectedasdescribedin
ExperimentalproceduresandSilver-stained.
B.Totalproteinexpressionlevels.Samples
werecollectedasdescribedinExperimental
proceduresandWestern-blottedwithan
antibodyagainstIpaD.
Resultsshownarerepresentativeofatleast
twoindependentexperiments.
secreted constitutively in a DipaB background (Menard the control strain DipaB DipaD/ipaD+ (Fig.S3). This indi-
etal., 1994; Martinez-Argudo and Blocker, 2010). We catesthatourmutationsarenotimpairingIpaD’sabilityto
generated a DipaB DipaD double KO strain and after besecretedbytheT3SS.
transformingalltheipaDmutantplasmidsindividuallyinto
thisstrainwecomparedtheirsecretionphenotypetothat
ClassIIipaDmutantsshowdefectsininvasion
oftheKOstraintransformedwithawild-typecopyofthe
ofculturedcells
ipaD gene. Analysis of the secretion profile by Western
blot established that the IpaD mutant proteins were con- Totestwhetherthedefectinproteinsecretionafteraddi-
stitutively secreted by the DipaB DipaD/ipaD mutant tionofCRintheipaDmutantscorrelateswithadefectin
strains, as efficiently as wild-type IpaD was secreted by signal reception/transduction after host cell contact, we
analysedwhethertheipaDmutantsarestillabletoinvade
HeLacells.Gentamicinprotectionassayswereperformed
as described in Experimental procedures. In this assay,
the DipaD mutant is completely non-invasive (Menard
etal., 1993).All Class II ipaD mutant strains, except for
ipaDN292Y,showedaseveredefectininvasion(c.20%
of the complemented strain, Fig.2A). Mutant ipaDL99P
was as invasive as the complemented strain. These
results indicate that most of the mutations identified in
ipaDthatcauseadefectinCR-inducedsecretionleadto
adefectininvasionofculturedcells.
Fig.4. ImpairedsecretionofIpaDL99Pisduetolackofregulation
byMxiC. ipaDmutantstrainshavenormalneedle
A.SecretionofIpaDafterCongored(CR)inductionbyaDipaB tipcompositions
DipaDdoublemutantcomplementedwitheitheripaDaloneora
combinedipaBD3ipaDplasmid.Sampleswerecollectedas AsallisolatedipaDmutantshadapartialdefectinsecre-
describedinExperimentalproceduresandWestern-blottedwithan
tion and as IpaD forms part of the tip complex, one pos-
antibodyagainstIpaD.
B.Totalproteinexpressionlevels.Sampleswerecollectedas sible explanation is that the mutants have a defect in
describedinExperimentalproceduresandWestern-blottedwithan the needle tip composition. Altered tip compositions are
antibodyagainstIpaD.
known to alter secretion regulation (Veenendaal etal.,
Resultsshownarerepresentativeofatleasttwoindependent
experiments. 2007).Therefore,weexaminedthetipcomplexcomposi-
©2013BlackwellPublishingLtd,MolecularMicrobiology,87,690–706
696 A.D.Roehrich,E.Guillossou,A.J.BlockerandI.Martinez-Argudo (cid:2)
Fig.5. ThemutationsL99PandK291Iaffect
differentIpaDfunctions.
A.ExponentialleakageoftheDipaDmutant,
complementedstrain(DipaD/ipaD+)andipaD
mutants(intheDipaDbackground).Samples
werecollectedasdescribedinExperimental
proceduresandWestern-blottedwiththe
indicatedantibodies.
B.ProteinsecretioninresponsetoCongored
(CR)induction.Sampleswerecollectedas
describedinExperimentalprocedures,
Silver-stained(toppanel)andWestern-blotted
withtheindicatedantibodies(bottompanels).
C.Totalproteinexpressionlevels.Samples
werecollectedasdescribedinExperimental
proceduresandWestern-blottedwiththe
indicatedantibodies.
Resultsshownarerepresentativeofatleast
twoindependentexperiments.
tion of the ipaD mutants. Needles were purified from the wascompletelyuninducible.Totestwhetherthecombina-
differentipaDmutantstrainsasdescribedinExperimental tionofmutationswouldproduceastrongerphenotype,we
procedures and analysed by Western blot for the pres- constructed two new mutant ipaD plasmids combining
ence of the tip proteins IpaD and IpaB. All mutants had ClassIIamutationsN186YandK291IorClassIIbmuta-
wild-typelevelsofbothIpaDandIpaBwithinthelevelsof tions N292Y, T296I and Q299L. Mutant ipaD(N186Y,
experimental variability (Fig.2B).This was also the case K291I)considerablyreducedthesecretionlevelsfollowing
in the ipaDL99P mutant, which was expressed at lower CRinductioncomparedtosingleN186YandK291Imuta-
levels (Fig.2C). This indicates that none of the mutants tions,whilethetriplemutantipaD(N292Y,T296I,Q299L)
hasadefectinneedletipcomplexcomposition. showed a phenotype similar to Class IIa single mutants
(Fig.3).Justliketheoriginalsinglemutants,thesecombi-
nationsshowednormalleakage(notshown).
ClassIIaandIIbipaDmutantsareaffectedinsignalling
Our previous results indicate that the gatekeeper
Induction of secretion was impaired in mutants from protein MxiC facilitates secretion of the translocators
Classes IIa and IIb. We have hypothesized that, as part IpaB,IpaCandIpaD,astranslocatorsecretionisonlyvery
of the tip complex, IpaD is involved in conformational weaklyinducibleinaDmxiCmutant(Martinez-Argudoand
changes essential for the transduction of the activation Blocker,2010).Wealsofoundthattranslocatorsecretion
signal (Veenendaal etal., 2007; Martinez-Argudo and is MxiC-independent in a DipaB or DipaD constitutive
Blocker, 2010). Therefore, we should be able to isolate secretorbackground,butthatitisstillMxiC-dependentin
mutantsunabletotransmitthesignal.However,nomutant aconstitutively‘on’ipaBD3strain,asinthewild-typestrain
©2013BlackwellPublishingLtd,MolecularMicrobiology,87,690–706
CharacterizationofipaDandmxiCpointmutants 697
Fig.6. mxiC(E201K,E276K,E293K)hasa
phenotypesimilartoipaDL99P.
A.Exponentialleakage.Sampleswere
collectedasdescribedinExperimental
proceduresandWestern-blottedwiththe
indicatedantibodies.
B.ProteinsecretioninresponsetoCongored
(CR)induction.Sampleswerecollected
describedinExperimentalprocedures,
Silver-stained(toppanel)andWestern-blotted
withtheindicatedantibodies(bottompanels).
C.Totalproteinexpressionlevels.Samples
werecollectedasdescribedinExperimental
proceduresandWestern-blottedwiththe
indicatedantibodies.
ThecomplementedstrainDmxiC/mxiC+was
grownwith25mMIPTG,whilethe
DmxiC/mxiC(E201K,E276K,E293K)mutant
wasgrownwith10mMIPTGbecauseofits
highermxiCexpressionlevel.Resultsshown
arerepresentativeofatleasttwoindependent
experiments.
(Martinez-Argudo and Blocker, 2010; Roehrich etal., ground. Thus, impaired secretion of these IpaD mutant
2010).IntheipaBD3mutantIpaBisexpressedlackingits proteinsisnotduetoalackofregulationbyMxiC.Moreo-
lastthreeC-terminalaminoacids.Thismutantislockedin ver,theconstitutivesecretionpatternofallthestrainswas
asecretion‘on’conformation(i.e.similartothesecretion similar to that of the ipaBD3 DipaD/ipaD+ strain (not
activated wild-type), as it is a constitutive secretor, inva- shown), indicating that the ipaBD3 mutation is epistatic
sive and recruits some IpaC to the needle tip (Roehrich overtheipaDClassIImutations.
etal.,2010).Toruleoutthepossibilitythatthedecreased These results, together with the fact that the composi-
secretion of Class II IpaD mutant proteins was due to tion of the needle tip is wild-type-like for those mutants
lack of regulation by MxiC, we investigated if they were (Fig.2B), indicate that Class IIa and IIb mutants are
secreted in an ipaBD3 background. We made a plasmid affected in signalling from the needle tip and not in a
expressingbothipaBD3andipaDintowhichallthemuta- downstreamactivationstep.
tionsweresubsequentlycloned.Allplasmidsweretrans-
formed into the strain DipaB DipaD and their secretion
LackofsecretionoftheClassIIpaDL99Pmutant
profileanalysed.FormutantsbelongingtoClassII,secre-
proteinisduetolackofregulationbyMxiC
tion of IpaD became constitutive in all the mutants (not
shown). This was also the case when non-inducible In the ipaDL99P mutant strain, only IpaD secretion was
mutant ipaD(N186Y, K291I) was expressed in this back- impaired after activation (Fig.1B). As mentioned above,
©2013BlackwellPublishingLtd,MolecularMicrobiology,87,690–706
698 A.D.Roehrich,E.Guillossou,A.J.BlockerandI.Martinez-Argudo (cid:2)
Fig.7. IpaDandMxiCarebothinvolvedin
theregulationofthemechanismthatprevents
secretion.
A.ExponentialleakageoftheDipaDmutant
andaDipaDDmxiCdoublemutant
complementedwithipaDandmxiCcontaining
plasmids.Sampleswerecollectedas
describedinExperimentalproceduresand
Western-blottedwiththeindicatedantibodies.
Allsamplesdetectedwiththesameantibody
wereanalysedonthesamegel.
B.ProteinsecretioninresponsetoCongored
(CR)induction.Sampleswerecollectedas
describedinExperimentalprocedures,
Silver-stained(toppanel)andWestern-blotted
withthedesignatedantibodies(bottom
panels).
C.Totalproteinexpressionlevels.Samples
werecollectedasdescribedinExperimental
proceduresandWestern-blottedwiththe
indicatedantibodies.
Bacteriaweregrownwith1%arabinoseand
25mMIPTGasrequiredforexpressionof
ipaDandmxiCrespectively.Mutant
mxiC(E201K,E276K,E293K)isabbreviated
mxiC***.Resultsshownarerepresentativeof
atleasttwoindependentexperiments.
secretion of the translocator proteins from the cytoplasm conditions by Western blot and found that the mutant
is promoted by MxiC in wild-type and ipaBD3 back- secreted IpaB, IpaC, MxiC and effector proteins (Fig.5A
grounds(Martinez-ArgudoandBlocker,2010).Inorderto and not shown). This suggested that in the ipaDL99P
examineifsecretionoftheIpaDL99Pmutantproteinwas mutantaregulatoryfunctionofIpaDisalteredandthat,as
stilldependentonMxiC,weanalyseditssecretioninthe a result, the mechanism that prevents secretion before
ipaBD3background.WhencomparedwithipaBD3DipaD/ activationsignalreceptionisimpaired.Ontheotherhand,
ipaD+,strainipaBD3ipaDL99Pshowedadifferentpattern ClassIImutantsdidnotshowadefectinthepreventionof
of secretion: IpaDL99P secretion was impaired in the secretion under non-induced conditions (Fig.1A).There-
ipaBD3 background, but constitutive in the DipaB back- fore,ClassIandClassIImutantscouldaffecttwodifferent
ground (Fig.4). Therefore, the impairment in IpaDL99P aspectsofIpaDfunction.Totestthishypothesis,wemade
secretion is probably due to a lack of intracytoplasmic adoublemutantcombiningmutationipaDK291I(ClassIIa)
regulationbyMxiC. andipaDL99P(ClassI).Ifdifferentfunctionswereaffected
intheseclasses,onewouldexpecttoseebothphenotypes
inthecombinedmutant.However,ifbothclassesaffected
IpaDhasadualroleinT3SSregulation
the same function, dominance of one of the phenotypes
TheipaDL99PstrainsecreteshigherlevelsofT3Sproteins would be more likely. The mutant ipaD(L99P, K291I)
innon-inducingconditionscomparedtothecomplemented showedbothphenotypes:itsecretedIpaB,IpaCandMxiC
strain(Fig.1A).Tofurtherinvestigatethis,weanalysedthe under non-inducing conditions and was impaired in CR
supernatant of exponential cultures under non-induced induction (Fig.5).Together with the fact that secretion in
©2013BlackwellPublishingLtd,MolecularMicrobiology,87,690–706
CharacterizationofipaDandmxiCpointmutants 699
Fig.8. PrematuresecretioninipaDL99Pis
MxiC-dependent.
A.ExponentialleakageofaDipaDDmxiC
doublemutantcomplementedwitheitheripaD
oripaDL99PaloneoradditionallywithmxiC.
Sampleswerecollectedasdescribedin
ExperimentalproceduresandWestern-blotted
withtheindicatedantibodies.
B.Totalproteinexpressionlevels.Samples
werecollectedasdescribedinExperimental
proceduresandWestern-blottedwiththe
indicatedantibodies.
Bacteriaweregrownwith1%arabinoseand
25mMIPTGasrequiredforexpressionof
ipaDandmxiCrespectively.Resultsshown
arerepresentativeofatleasttwoindependent
experiments.
thedifferentmutantclassesisdifferentiallyaffectedbyan secretion. To distinguish between these possibilities, we
additionalipaBD3mutation,thisresultsuggeststhatIpaD analysed whether IpaB and IpaC were secreted from a
hasadualroleinregulationofT3SSsecretion. DipaDDmxiC/ipaDL99Pstrainundernon-inducingcondi-
tions(Fig.8).WefoundprematuresecretionintheDipaD
DmxiC/ipaDL99P strain was reduced in comparison to
IpaDandMxiCarebothinvolvedintheregulationofthe
ipaDL99P alone. Thus, premature secretion of IpaB and
mechanismthatcontrolssecretion
IpaCinipaDL99PisdependentonthepresenceofMxiC.
IntracytoplasmicMxiCisrequiredforsecretionoftranslo-
cator proteins after activation (Martinez-Argudo and
Discussion
Blocker,2010).Intriguingly,wegeneratedamutantinmxiC
with a very similar phenotype to the ipaDL99P mutant: Type III secretion systems are activated by host cell
mxiC(E201K, E276K, E293K) is a triple mutant in a con- contact which we proposed leads to a conformational
served negatively charged patch on the surface of MxiC change in the needle tip complex that triggers secretion
thathasbeensuggestedtobeinvolvedinproteininterac- from the cytoplasm (Veenendaal etal., 2007; Blocker
tions(Deaneetal.,2008).ThismutantalsosecretesIpaB, etal., 2008; Martinez-Argudo and Blocker, 2010). To
IpaC, MxiC and effector proteins in non-inducing condi- investigate the role of the major component of the tip
tions. Furthermore, it is also impaired in IpaD secretion complexIpaDintransductionoftheactivationsignalinto
afterinduction,butnotaffectedininductionofanyofthe thecytoplasm,weusedageneticscreentoidentifyipaD
otherproteins(Fig.6andnotshown).Ourresultssuggest mutants unable to secrete after induction by CR.All pre-
thatboththeipaDL99PandmxiC(E201K,E276K,E293K) viousmutationsdescribedinIpaDaresilentorleadtoat
mutationscauseprematureactivationofsecretion. leastpartialloss-of-functionandhenceconstitutivesecre-
When IpaDL99P and MxiC(E201K, E276K, E293K) tion(Pickingetal.,2005;Bartaetal.,2012b;Epleretal.,
were expressed together, we observed an increase in 2012). Here we describe several mutations in ipaD that
secretionundernon-inducedconditionsincomparisonto lead to separate defects in secretion activation (Class I
thesinglemutants(Fig.7).Theseresultsstronglysuggest andClassII)andusethemtoconcludethatIpaDhastwo
that as MxiC, IpaD is involved in the mechanism that rolesinthisprocess.
controlssecretion.
IpaDisinvolvedinsignaltransductionfromthe
PrematuresecretionintheipaDL99Pmutantis needletip
MxiC-dependent
All our results suggest that Class II ipaD mutants are
WethuswonderedwhetherIpaDisinvolvedinthesame affected in signalling from the tip complex. Their very
activationrouteasMxiCorwhetherithasanindependent identification supports the existence of at least two func-
role in controlling secretion. If the former was true, one tional states of the needle tip.All Class II mutants have
would expect the ipaDL99P mutant to require MxiC for wild-type-likeneedletipcomplexcompositionandshowa
prematuresecretion.However,ifIpaDwasdirectlyaffect- decrease in protein secretion after induction by CR. Fur-
ing secretion of the translocators IpaB and IpaC, thermore,thedefectinsecretionofIpaDClassIImutantsis
ipaDL99P should lead to MxiC-independent constitutive notduetotheirinabilitytobesecretedbytheT3SSortoa
©2013BlackwellPublishingLtd,MolecularMicrobiology,87,690–706