Lam-Yuk-Tseung, S., “Nramp Metal Transporters: Insights into their structure, function, and subcellular targeting”, Doctoral Thesis, 2006.

irondiet40.TheyshowedevidencethatNramp2behavedasaproton-coupleddivalent
metaltransporterwithanunusuallybroadsubstratespecify,inc1udingFe
2+andMn2+.
Thec10ningandcharacterizationofNramp2isdescribedindetailinChapter1.3ofthis
thesis.Indeed,theidentificationandcharacterization
ofnumerousNramporthologsin
speciesrangingfrombacteriatohumanssuggestedapotentialfunctionforNramp1atthe
phagosomalmembraneasapH-dependentdivalentmetaltransporter.Byanalogytothe
knownsubstrates,membraneorganization,anddirection
oftransportestablishedfor
Nramp2,weandothersproposedamodelwherebyNramp1wouldactasaneffluxpump
atthemembrane
ofpathogen-containingphagosomes,restrictingmicrobialaccessto
essentialmetalssuchas
Mn2
+andFe2
+41;42.However,earlystudiessuggestedthat
Nramp1wouldactasametalinfluxpumpatthephagosomalmembranetofacilitatethe
production
ofoxygenradicalsviatheHaber-Weissreaction43;44.Thesestudiesreported
anNrampl-dependantincreaseintheaccumulationlbindingofradiolabeledFe
2+into
isolatedphagosomescontainingeitherLatexbeads
orMyocobacteriumavium43-45.The
phagosomalFe
2+accumulationwasblockedbytheadditionofanti-Nramp1antibodies
suggestingthat
NramplmighttransportcytoplasmicFe2+intophagosomes.However,a
potentiallimitation
ofthesestudieswasthelackofdistinctionbetweenphagosomalmetal
binding(whetherNrampl-dependent
orindependent)andactualmetaltransportacross
thephagosomalmembrane.Alaterstudyusing
NramplmRNAinjectedintoXenopus
laevis
oocytesfoundanincreaseinZn2+-dependentinwardcurrentswhenthecellswere
exposedtoalkalineextracellularpH,whichsuggestedmetaluptakeagainstthe
experimentalprotongradient
46.BasedonadditionalpH-dependenttransportstudies
usingradioabeled
Zn2+,theseauthorsconcludedthatNramplmighttransportmetalsfrom
thecytoplasmintophagosomes
byaprotonldivalent-metalantiportmechanism.They
proposedthatincreasedphagosomalFe
2+couldstimulateoxygenradicalproductionvia
theFentonreactionforincreasedbactericidalactivity
43-47.However,forNramp1to
functionasprotonldivalent-metalantiporteratthephagosomalmembraneitwoulddictate
atransportmechanismdistinctfromotherNramporthologsasweIlasNramp2,bothwith
respecttothedirection
oftransportandutilizationofthetransmembranepHgradient.
8

Thiswasveryunlikelyconsideringthehighaminoacidconservationinthepolypeptide
sequences
ofNramp1andNramp2(64%identifyand78%similarity)1.
RecentexperimentshaveprovidedstrongevidencethatNramp1indeedfunctions
asapH-dependentdivalentmetaleffluxpumpatthephagosomalmembrane,consistent
withthebiochemicaltransportdatareportedinstudies
ofNramporthologs26;27;40;48-51.
First,ametal-sensitivefluorescentdyewascovalentlycoupledtozymosanpartic1es,
whichwerethenfedtoperitonealmacrophagesisolatedfromNrampl-expressing(+/+)
andNrampI-deficient(-/-)mice.Thetransport
ofdivalentmetalsacrossthephagosomal
membranewasmonitored
bytrackingfluorescencequenching.Nramp1+1+phagosomes
showedreducedaccumulationandincreasedefflux
ofMn2+comparedtoNramprl
–
phagosomes.Thisdifferencewaslostwhenphagosomalacidificationwasblockedwith
theV-typeATPaseinhibitorbafilomycin,suggestingthatdivalentmetaltransport
by
NramplwasH+-dependent,likeNramp252.Second,anotherconsequenceofthe
functionalsimilaritybetweenNramp1andNramp2isthatshouldNramp1
beexpressedat
theplasmamembrane,itwouldfunctioncomparablytoNramp2andtransportdivalent
metalsintothecellinapH-dependentmanner.Suchan
Nramplvariantwascreatedby
insertingahemaglutininepitopetaginthefourthpredictedextracytoplasmicloopofthe
proteindelineated
byTM7andTM827.Atthecellsurface,theNramp1variantindeed
causedpH-dependentuptake
ofFe2+andMn2+,asmeasuredbybothfluorescence
quenchingandradioisotopicmetals
27.Interestingly,Nramp1wasfoundtobeamore
efficienttransporterfor
Mn2+thanFe2+,suggestingthatMn2+isthenaturalsubstratefor
Nramp1atthephagosomalmembrane
27.Theseresultsgivestrongcredencetothemodel
thatNramp1controlstheintracellularreplication
ofmycobacteriaandotherpathogensby
actingasaneffluxpumpatthephagosomalmembrane,restrictingmicrobialaccessto
essentialmetals(Figure1).
1.2.5PathogensunderNramplcontrol
InadditiontoSalmonellatyphimurium,Leishmaniadonovani,andseveralspecies
ofMycobacteria
(Mbovis53,Mavium54,Mlepraemurium55,Mintracellulare56),
thereareanumberofotherpathogenstowhichdifferentleve1sofresistanceor
9

Figure1.MetaltransportbyNrampproteinsattheinterfaceofhost­
pathogeninteraction.Inthetwodiagrams,Nramp1(green),Iysosomal
associatedmembraneproteinLamp(yellow),vacuolarW-ATPase(red),and
metalions(blue)arecolourcoded(Ieft)andidentified(right).
Inmacrophages,
bacterium-containingphagosomesmaturebyfusiontovacuolarW-ATPase
positivevesicles,andLamp-1/Nramp1positivelysosomes.ATP-dependent
acidification
ofthephagosomeisrequiredforNramp1-mediatedeffluxof
divalentmetals(Fe2+,Mn2+,Zn2+).Atthephagosomalmembrane(rightpanel),
thevacuolarW-ATPasecreatesaprotongradient([H+]versus
[H+),whichis
necessaryforNramp1-mediatedmetaleffluxfromthephagosomallumen.
Bacteriapresent
inthephagosomeexpressNramphomologues(e.g.MntH)
thatmayimportdivalentmetalsfromthephagosomalspacebyasimilarpH­
dependentmechanism.Themembranetopology
ofNrampproteinsisshown
withrespecttotheamine(N)andcarboxyl
(C)terminus.(From:Lam-Yuk­
Tseung,
S.andGros,P.”Geneticcontrolofsusceptibilitytobacterial
infectionsinmousemodels”.CellularMicrobiology.2003May;5(5):299-313.)

susceptibilityhavebeenassociatedwithpolymorphicvariationsinNrampl.These
inc1udeavariety
ofovine57andavian58strainsofSalmonella,aswellasBrucella
abortus
59,Pasteurellapneumotropica60,Candidaalbicans47,Toxoplasmagondii61;62,
andFrancisellatularensis63;64.Surprisingly,Nramplisnotcriticalforresistanceinmice
to
invivoinfectionwithH37rvMycobacteriumtuberculosis,thecauseofpulmonary
tuberculosisinhumans65-67.Thesereportsareincontrasttothesituationinhumans
wherepolymorphicvariantsat
NRAMP1havebeenassociatedwithtuberculosis68-74.
However,thisapparentparadoxmaybeexplainedbydifferencesbetweenthe
experimentalmodel
oftuberculosisinthemouse(loworhighdoses,i.v.),andthenatural
humansituation(lowdoses,aerosol).Thenextsections
ofthisthesiswilldescribethe
effects
ofNramp1oncontrollingtheintracellularreplicationofMycobacteriumand
Salmonella,sincetheroleofNramplininfectionbytheseparasiteshavebeenrelatively
wellcharacterized.
1.2.6EffectofNramplonintracellularMycobacterium
NrampI-mediatedmetaleffluxfromthephagosomehaspleiotropiceffectsonthe
engulfedpathogen.
Ithaslongbeenestablishedthatmycobacteriasurvivewithin
macrophages
bypreventingthematurationofphagosomesintofullybactericidal
phagolysosomes
75-81.MycobacteriapresentinNrampI-deficientphagosomesareableto
inhibitluminalacidification
byblockingrecruitmentofvacuolarH+-ATPasepumpstothe
phagosomalmembrane,aswellasreducingtheacquisition
oflysosomalmarkerssuchas
Lamp1andCathepsinD
78-81.Rather,thesephagosomesretainsortinglearlyendosome
markerssuchastransferrinreceptor(animportantsource
ofTf-iron)andRab5(required
forearlyendosomefusion).Thisinhibition
ofphagosomalmaturationappearstobean
activeprocessfromthelivepathogensinceitdoesnotoccur
whenmacrophagesengulf
latexbeads
ordeadmycobacteria.InthepresenceofNrampl(Nrampl+/+),mycobacteria
areunabletoblockphagosomalmaturation,resultinginincreasedluminalacidification,
increasedfusionwithlysosomes,enhancedbacterialdamage,anddiminished
mycobacterialreplicationcomparedto
Nrampr/-phagosomes82;83(Figure2).Nrampl+/+
macrophagesalsoproducemorenitricoxide(NO)inresponsetomycobacterialinfection
11

Nramp+/+
Macophage
•Nramp1/Slc11a1
oRab5,TfR
Nramp-/-
LE/Lysosome
Macophage
=Lamp1,H-ATPase,CathepsinD
1Livemycobacterium
eKilledmycobacterium
Figure2.TheeffectofNramp1onintracellularMycobacterium.Inthepresenceof
Nramp1(Nramp1+/+,leftpanel),mycobacteriaareunabletoblockphagosomalmatura­
tion,resultinginincreasedluminalacidification,increasedfusionwithlysosomes,
enhancedbacterialdamage,anddiminishedmycobacterialreplication.Mycobacteria
survivewithinmacrophagesbypreventingthematuration
ofphagosomesintofully
bactericidalphagolysosomes.MycobacteriapresentinNramp1-deficientphagosomes
(rightpanel)inhibit
luminaiacidificationbyblockingrecruitmentofV-ATPasepumpsto
thephagosomalmembrane,asweilasreducingtheacquisition
ofIysosomalmarkers
suchasLamp1andCathepsin
D.Rather,thesephagosomesretainsorting/earlyendo­
somemarkerssuchastransferrinreceptorandRab5.

and/orinterferon-ystimulationthanNramprl-controIs,possiblycontributingto
bacteriostasis84-86.
Themechanismbywhichmycobacteriacaninhibitphagosomalmaturationis
poorlyunderstood.However,Nramp1c1earlyplaysaroleinantagonizingthisprocess.
Onepossibleexplanationisthatinhibition
ofphagosomalmaturationbymycobacteriais
ametal-dependantprocessthatcanbecountered
byNrampI-mediatedmetalefflux.
Indeed,anumber
ofstudieshavedemonstratedtheimportanceofironinmycobacteria­
mediatedinhibition
ofphagosomalmaturationaswellasinthevirulenceofthe
intracellularpathogens
87-90.Inaddition,microarraygeneexpressionanalysisofM
tuberculosisisolatedfromactivatedmacrophages,showedup-regulationofmycobacterial
genesinvolvedinmetaluptake,accountingforover
halfoftheactivatedgenes89.Iron
promotesmycobacterialgrowthin
Nramp1+/+miceandpromotesthedevelopmentof
activetuberculosisinhumans90;91.Likewise,extracellularironcanstimulate
intracellulargrowth
ofMaviuminbothNramp1+/+andNrampr/-macrophages92,
suggestingthatexcesscellularironmayoverwhelmthefunctionofNramp190;92.
Collectively,theseresultssuggestthatironisrequiredbymycobacteriatoarrest
phagosomalmaturationandsurviveintracellularly.
1.2.7EffectofNramplonintracellularSalmonella
Unlikemycobacteria,Salmonelladonotsurviveintracellularlybyinhibiting
phagosomalacidification.Rather,theysequesterthemselvesinsidespecialized
Salmonella-containingvacuoles(SCVs)thatbecomeacidifiedandareabletofusewith
LampI-positivelysosomes
93.InpermissiveNrampr/-macrophages,Salmonellasurvive
byblockingrecruitmentofmannose-6-phosphatereceptor(M6PR)positivevesiclesand
isolatethemselvesfromtheendosomalpathway
37.Innon-permissiveNrampl-expressing
macrophages,SCVsfusewithM6PR-positivevesic1esandmaintainaccesstothe
endosomalpathway,expressingendosomalmarkerssuchasEEAIandfluidphaseFITC­
dextran
37.TheseeffectscouldbemimickedinNrampI-negativemacrophages
13

byusingmembrane-permeantironchelators94.Theseresultssuggestamode!inwhich
Nramp1counteractstheability
ofSalmonellatosequesteritselffromthedegradative
phagosomalpathway
byrestrictingitsaccesstodivalentmetals.
Divalentmetals,suchasFe2
+andMn2+,have
beenshowntobeessentialin
Salmonellaforvirulenceinvivoandreplicationinculturedmacrophages91;95.Salmonella
possessseveralhighandlowaffinityironandmanganesetransporters96-99andseveralof
thesetransportershavebeenshowntocontributetovirulenceinvivo98;100-102.Infact,
mutationsatanumber
ofthesetransporters(suchasfeoB,sitA-D,orMntH)ordivalent
metalchelationhavebeenshowntoreduce
orabrogatevirulenceofSalmonella101.Thus,
NrampI-mediateddepletion
ofmetalsfromtheintra-phagosomalspaceappearstohavea
majorimpact
onintracellularsurvivalandreplicationofSalmonella.Thepathogen,in
tum,reactstothemetaldepletion
byupregulatingexpressionofmetaltransporterssuch
as
MntHandsitA103andincreasingexpressionoftheirPathogenicityIsland2(SPI2)
virulencegenes104.
1.2.8NRAMPIanditsroleinresistancetoinfectioninhumans
Inhumans,NRAMP1hasbeenmappedtochromosomalregion2q35andconsists
of15exonsthatspana13kbregion105;106.Twotranscriptionalstartsiteshavebeen
mappedandregulatorymotifs
intheNRAMP1promoterregionhavebeenidentified.
Theseinc1udeaTATAboxelement,interferon-yresponseelements,andbindingsitesfor
thetranscriptionalfactorsNFKB,
AP-l,andPU.II05-108.Resultsobtainedinmurine
modelshaveraisedthepossibilitythat
NRAMPgenescouldbeimportantdeterminantsof
susceptibilitytocommonhumandiseases.Anumberofstudiesconductedonethnically
andgeographicallydiversepopulationsareingeneralagreementthat
NRAMP1allelesare
riskfactorsfortuberculosis
74;109-116.Indeed,anumberofgenomicDNApolymorphisms
havebeenidentifiedin
ornearNRAMP1thathavebeenassociatedwithsevera!infectious
andinflammatorydiseasessuchasleprosy
117-119,non-tuberculousmycobacterial(NTM)
lungdisease
120,humanimmunodeficiencyvirus(HIV)121,rheumatoidarthritis122-125,
inflammatoryboweldiseases126;127,multiplesc1erosis128,andtype1diabetes129;130.The
numerousillnessesaffected
byNRAMP1polymorphismsreflectthepleiotropiceffectsof
14

Nramp1functionandtheimportanceofdivalentmetalsindiseaseandinfection.
Recently,Kissler
etalusedlentiviraltransgenesisandRNAinterferencetoknock-down
expression
ofNramp1inanonobesediabetic(NOD)mousemodel,andshowedthat
Nramp1-silencingreducedthefrequency
oftype1diabetesintheseanimaIs130.Their
resultsprovidedstrongevidencethattheinsulin-dependentdiabeteslocus5.2
(idd5.2),
whichhasbeenimplicatedintheNODmousemodel,isindeedNrampl.
1.3Nramp2anditsroleinironhomeostasis
1.3.1Cloning
ortheNramp2gene
TwoyearsafterthepositionalcloningofNrampl,asecondmouseNrampgene
wasidentified
bycDNAcross-hybridizationstudiesbasedonitshighsequencehomology
with
Nrampl131.ThenovelgenewasnamedNramp2,andwasmappedtothedistalpart
ofmousechromosome15.ThemurineNramp2geneconsistsof18exonsspreadover
morethana30kbregion.Analysis
ofthenucleotideandpredictedaminoacidsequence
ofNramp2,indicatedthatitwasanovelproteincloselyhomologoustoNramplandthat
thetwoproteinsdefinedanewfamily
ofproteins1.ThetwoNrampproteinsidentified
shareahigh63%sequenceidentityand78%similarity
1;131.However,incontrasttothe
macrophage-specifie
Nrampl,Nramp2rnRNAexpressionisdetectedalmostubiquitously
40;131
1.3.2Nramp2encodesadivalentmetaltransporter
AlthoughthepolypeptidesequencesofNramp1andNramp2suggestedthatboth
proteinsfunctionedasmembranetransporters,thesubstrateandmechanism
oftransport
ofNrampproteinsremainedelusive.PerhapsthefirstindicationofthefunctionofNramp
proteinsoccurredin1996whenSupekandcolleaguesreportedtheidentificationanovel
Saccharomycescerevisiaemutantthatwasunabletogrowinthepresenceofthemetal
chelatorEGTA
132.Themutantcou1dbesuppressedbyoverexpressionoftheyeastgene
SMF1,ageneoriginallyclonedasamulticopysuppressorofatemperature-sensitive
mutant
(mifl-l)defectiveinthefunctionofmitochondrialprocessingpeptidase133.
However,theabilityofSmflptocomplementgrowthinthepresenceofEGTAsuggested
15

thatSMFlpIayedadirectroleinmetalhomeostasisinyeast.Consistentwiththis
hypothesis,Supek
etal.showedthatasm/1nullmutantexhibitedreducedMn2
+uptake.
Interestingly,
Smflpshowedsignificantresemblance(46%similarity)tobothNramp1
andNramp2proteins
1.
FurtherinsightintothefunctionofNrampproteinsoccurredin1997,whentwo
simultaneousindependentstudiesshowed
bydifferentmethodsthatNramp2functionsas
atransporter
ofdivalentcations,includingiron.Inthefirststudy,Nramp2wasisolatedby
expressioncloningfromaratcDNAlibraryinXenopusoocytesbyGunshinand
colleaguesinasearchforirontransportproteins
40.Thegoalofthisstudywastoidentify
theintestinaltransporterresponsibleforabsorption
ofnon-hemedietaryiron.To
accomplishthis,theypreparedacDNAlibraryusingduodenalrnRNAfromratsthatwere
kept
onalow-irondiet.Theproximalduodenumhadbeenpreviouslyidentifiedasthe
mainsite
ofnon-hemeironabsorptionbyphysiologicalironuptakemeasurements134;135.
AsinglecDNAclonecauseda200-foldincreaseinironuptakecomparedtocontrol
uninjectedoocytes.Theclonewasfoundtocontainrat
Nramp2,althoughitwasre-named
DCT1(Diva/entCationTransporter1)inthisreport,andwaslaterre-namedagainto
DMT1forDiva/entMetalTransporter1.AlthoughDMT1isthemostcommonlyused
termfor
Nramp2today,ithasbeenrecentlyre-namedonceagaintoS1c11a2(solute
carrierfamily
Ilmember2).Forclarity,1willusetheNramp2/DMT1nomenclaturein
theremainderofthisthesis.
Expression
ofNramp2inoocyteswasdemonstratedtomediateuptakeofabroad
range
ofdivalentmetalsincludingFe2+,Mn2+,Zn2+,C02+,Cu2+,Ni2+,Pb2+,and
Cd2
+but
not
Ca2+orMg2
+40.MetaltransportwasshowntobedependentonextracellularpHand
thecellmembranepotential.Furthermore,mRNAexpressionstudiesconfirmedthe
ubiquitousexpression
ofNramp2butalsodemonstratedastrikingtissue-specifie
upregulation
ofNramp2rnRNAuponchronicirondepletion.Thisupregulationwas
strongest
intheproximalintestinebutwasalsoseentoalowerextentinthekidney,liver,
brain,heart,lung,andtestis
40.
16

1.3.3AmutationatNramp2causesirondeficiencyinratsandmice
InaparallelstudytoGunshinetal.40,Flemingandcolleaguesusedapositional
c10ningapproachtoidentifyNramp2asthecausativemutationinmicewithmicrocytic
anemia(mk)
136.Homozygousmklmkmicesufferfromseverehypochromic,microcytic
anemiaduetoimpairedintestinalironabsorptionanddefectiveerythroidironutilization
137-140.ThemkmutationarosespontaneouslyinthebreedingstocksoftheJackson
Laboratoryoverthirtyyearsagoandisinheritedasanautosomalrecessivetraitwithfull
penetrance
137.Newbommicehomozygousforthemutation(mklmk)werereadily
distinguishableatbirth
bytheirsmallsizeandpalecoloration,andweresubsequently
maintained
bybreedingontoseveralgeneticbackgrounds137.ThestudybyFlemingetal.
tracedthe
mkphenotypetoamutationinNramp2thatresultsinanon-conservative
glycinetoargininesubstitution(G185R)
inthepredictedTM4oftheprotein136.
Strikingly,itwaslaterdiscoveredthattheidenticalmutationinNramp2(G185R)wasthe
cause
ofdiseaseinaradiation-inducedratmutantthatexhibitedanmk-likephenotype,
theBelgrade(b)rat
141.Likemk,bisinheritedasanautosomalrecessivetraitthatresults
inhypochromic,microcyticanemiaassociatedwithimpairedreticulocyteironuptakeand
gastrointestinalironabsorption
142-145.Together,theseresultsprovidedstrongevidence
thatNramp2functionsasapH-dependentirontransporteressentialforbothnormal
intestinalironabsorptionanderythroidironusage.
InvitrostudieshaveshownthattheNramp2
Gl85Rproteinexpressedinbothmk
miceandBelgraderatsdisplaysmultiplebiosyntheticandfunctionaldefectsthatcombine
tocauseirondeficiencyinrodents.SuandcolleagueshaveshownthattheG185R
mutationreducestheability
ofNramptotransportironintransfectedHEK293kidney
cells
146.Incontrast,proteinexpressionstudiesinwild-typeandmutantanimaIsled
Canonne-Hergaux
etaltoconc1udethatNramp2Gl85Risinadequatelytargetedorretained
bytheplasmamembraneofenterocytesandreticulocytesfrommkmice147.Experiments
performed
byTouretetalintransfectedLLC-PK1epithelialcellsshowedthat
Nramp2
Gl85Risfunctionallyimpaired,lessstable,andinefficientlyprocessedresultingin
defectiveglycosylationandreducedplasmamembraneexpression
148.Consequently,
17

Nramp2G185Rwasaisoshowntoaccumulateintheendoplasmicreticulumwhereitis
rapidlydegraded
byaproteosome-dependentmechanism148.
ThefactthattheG185Rmutationhasoccurredspontaneouslyontwooccasionsin
mkmice136andonceinBelgrade141ratsisstrikingandsuggeststhateitherthecodon
containingG185ishypermutableorthatG185Rrepresentsagain
offunctionthatconfers
sornesurvivaladvantagetotheorganism.Work
byXuandcolleaguessupportthelatter
hypothesis.TheyshowedthatNramp2
G185R
,despitepresentingmultiplebiosyntheticand
functiondeficiencies,
confersanovelCa2+-selectivepermeabilitypathway149.Theyargue
thattheinflux
ofCa2+mightpotentiatetheresidualNramp2iron-transportactivity.The
transferrincycleisessentialforironuptake
byerythroidprecursorcells150andNramp2
mediatestransfer
ofironfromtransferrincycleendosomestothecytoplasm42;51;141.
ElevatedintracellularCa2+hasbeenreportedtoaccelerateironuptakethroughthe
transferrincycle,apparentlythroughactivation
ofproteinkinaseC151.
Recently,mutationsinNRAMP2havebeenidentifiedinthreehumanpatients
sufferingfromseveremicrocyticanemiaandhepaticironoverload
4-6.Theeffectsof
thesemutationsatboththephysiologicalandmolecularlevelsarediscussedfurtherin
section1.3.9.3
ofthischapter.
1.3.4AlternatesplicingofNramp2rnRNA
TheamineacidsequencepredictedfromtheratcDNAisolatedbyGunshinetal.
wasaimostidenticaitothesequencereportedformouseNramp2,exceptattheextremeC
terminus,wherethetwosequencesdiverged
40;131.However,theratNramp2sequence
washomologoustothepublishedC-terminaisequence
ofhumanNRAMP2152.The
discrepancy
intheC-terminaisequenceswasexplainedinsubsequentstudiesreporting
theexistence
oftwoisoformsofNramp2,generatedbyalterativesplicingattwo3’exons
inmice,ratsandhumans40;131;136;152;153.ThisaltematesplicinggeneratestwoNramp2
proteinswithdifferentC-terminaisegments.Interestingly,one
ofthetwoNramp2splice
isoforms,namedisoform1or
+IRE,containsanironresponsiveelement(IRE)inits3’
untranslatedregion(UTR)andproducesa561aminoacidproteininratsandmice.The
otherspliceisoform,namedisoformIlor
-IRE,lacksanIREinits3’UTRandproduces
18

a568aminoacidprotein.IREsareknowntoplayregulatoryfoIesincellulariron
metabolismthroughinteractionswithironregulatoryproteins(IRPs).TheroleIRPsin
regulatingcellularironhomeostasisisdescribedfurtherinsection1.3.8.4
ofthisChapter.
Recently,a
novel5’upstreamexonhasbeenidentifiedinthehurnan,mouse,andrat
Nramp2genes
154.Thisnovelexon(exonlA)isgeneratedbytheusageofanaltemate
transcriptioninitiationsite,andispredictedtoproduceanNramp2proteinbearingan
additional29-31aminoacids(exon
lA)upstreamofthepreviouslyidentifiedstartcodon
ofNramp2isoforms1andII(exonlB).TheusageofexonlAalsoappearstobehighly
tissue-specifie,withenhancedusageinthekidneyandduodenum
154.Thus,atheoretical
total
offourNramp2mRNAsandproteinscanbeproducedbythegene,dependingon
altemate
5′(exonlAvs.lB)and3′(I/+IREvs.ill-IRE)splicing,givingrisetofour
Nrampisoforms:I(A),I(B),II(A),andII(B)(Figure3).However,most
invitrostudieson
Nramp2,inc1udingaIlthestudiesdescribedinthisthesis,havebeenperformedusingthe
originallyidentifiedisoformsI(B)andII(B).Littleiscurrentlyknownaboutthe
importance
ofexonlAortheextra29-31N-terminalresiduesintheregulation,function,
ortargetingofNramp2.
1.3.5StructuralfeaturesoftheNramp2protein
LikeNramp1,Nramp2isexpressedasa90-100kDaintegralmembrane
phosphoglycoproteinintransfectedCHOorLLC-PK1cellsthatispredictedtohave
twelvemembrane-spanningdomains.Thepredictedtopology
ofNramp2inthe
membranehasbeensupported
byseveralepitopeaccessibilitystudiesinintactcells
26;27;51.PicardandcolleaguesusedimmunofluorecenceexperimentsonintactCHOcells
expressinganepitope-taggedNramp2varianttoshowthattheregionconnectingTM7
and8isindeedexofacial.SimilarexperimentswithNramp1andNramp2variants
possessingN
orCterminalepitopetagshavedemonstratedthattheaminoandcarboxyl
terminalsegments
ofbothtransportersfacethecytosol26;27.GlycosylationinNramp2is
extensive,withthetransporterowing
~50%ofitsmolecularmasstosugarmoietiesadded
attwoconsensusN-linkedglycosylationsignaIspresentinthefourthextra-cytoplasmic
loop
42.ExtensiveglycosylationofNramp2mayhelpprotecttheproteinagainst
19

Isoform1(A)
Nram2•
IsoformIl(A)
Nramp2I<;'IRE"'~:I IsoformIl(8) Isoform 1(8) ,Nramp2 Figure3.AlternativesplicingofNramp2pre-mRNAgeneratesfourtheoretical isoforms.Theuseofalternatetranscriptioninitiationsitesatthe5'exon(exon.1Avs. 1 B),generatesdiversityatthe5'endofthemessage.Alternatesplicingatthe3'terminal exongeneratesfurtherdiversity,causingtheinclusion(+IRE)ornot(-IRE)aniron resonpiveelement(IRE) inthe3'untranslatedregion. degradationormaybeimportantforpropermaturationandtargetingofthetransporterto itsite ofaction.TabuchiandcolleagueshaveshownthatdisruptionoftheN-linked glycosylationsitesinNramp2affectthetransporter'spolarizeddistributiontotheapical membrane ofconfluentMDCKkidneycells155.AsinNramp1,thefourthintra­ cytoplasmicloop ofNramp2containsaconservedtransportmotif(CTM)ofunknown funetion1;131. ConsistentwithothermembersoftheNrampfamily,Nramp2possessesnine highlyconserved,yetthermodynamicallydisfavored,ehargedaminoaeidswithinitsTM domains 1.AisopresentaretwohighlyeonservedhistidineresiduesinTM6ofthe protein.Beeause oftheirhighconservation,theseehargedresidueslikelyplayaeritieal roleinthestructureand/orfunctionofNramp2.TheworkdeseribedinChapter2 ofthis thesisidentifiesthreenegativelyehargedresidues(D86,D192,E299)within transmembranedomains 1,4,and7ofNramp2thatareessentialformetaltransport activity 28.Thesenegativeresiduesarethoughttoformpartofametalbindingsiteorline ahydrophilicporeortransportpath.Theseresultshavebeensupportedbysimilar subsequentmutagenesisstudies ofthechargedTMresiduesinthebacterialNramp ortholog, MntH156.SeveralstudieshavebeenaimedatbetterunderstandingthepH­ dependence ofmetaltransportbyNrampproteins.Sacherandcolleagueshaveprovided evidencefromelectrophysiologicalmeasurementsthatNramp2,andotherNramp proteins,mayexhibitaprotonclutchmeehanism,allowingmultipleprotonsto"slip" throughthetransporteratacidic pH50.Theadvantageofsuchaprotonslippage mechanismisnotcurrentlyknown,however,ithasbeenpostulatedtoprotectthe organismagainsttheoverloading ofmetalsinthepresenceofexeessprotons.Atthe molecularlevel,experimentsdescribedinChapter2identifytwoconservedhistidines (H267,H272)inTM6 ofNramp2thatareeritiealforregulatingthepHdependeneeof metaltransport28.SubsequentstudiesonratNramp2expressedinXenopusooctyesusing electrophysiologiealandradioisotopictechniqueshaveshownthatH267andH272are criticalfortheH+-couplingrequiredformetaltransport 157.Finally,electrophysiological studieshaveshownthatasingleaminoacidsubstitution(F227DinTM4 ofNramp2can increasetheratio ofmetalstoprotonstransportedby14-fold158. 21 StudiesinvolvingtheyeastNramporthologs(Smfl-3)havealsoprovidedvaluable insightintothemolecularstructureandfunction ofNramp2.Amutantyeaststrainwith two ofitsNramporthologsinactivated(SmflLYSmj2L1)isunabletogrowinthepresence ofthemetalchelatorEGTA34;132.ExpressionofNramp2inSm/lLYSmj2L1yeastisknown tosuppressthisphenotypeandrestoregrowthinthepresence ofEGTA34.Pinneretal tookadvantageofthistodemonstratethatmutations(Q384E,G394V)athighly conservedresidueswithintheCTM ofNramp2abrogatethefunctionoftheprotein34. Cohenandcolleaguesusedcomplementationassaysinyeastinconjunctionwith e1ectrophysiologicalmeasurementsinoocytestoanalyzethestructuralandfunctional importance ofthefirstextra-cytoplasmicloop(ECI)ofNramp2159.Theyshowedthat criticalsinglemutations inECI,Gl19AandQ126D,resultedinacompletelossofmetal uptakeactivitybyNramp2.However,thenon-functionalQl26Dmutantcouldbe partiallyrestoredbytheintroductionofasecondmutationattwoaminoacidsupstreamof Ql26(D124A).ThedoublemutantD124NQl26Dalsoappearedtoalterthemetalion specificity ofthetransporterinfavorofFe2+whileatriplemutant (Gl19ND124NQ126D)displayednotransportactivityyetpossessedalteredpre-steady­ statecurrents 159.Theseresultssuggestthatthefirstextra-cytoplasmicloopofNramp2 maybeinvolvedinmetalionbindingandproton-couplingoftransport. 1.3.6Nramp2expression Nramp2expressionhasbeendetectedathighlevelsinepithelialcellsaswellas peripheraltissues.Ingeneral,isoform1 ofNramp2appearstobepredominantlyexpressed inepithelialcellswhileisoformIIappearsto beexpressedmostlyinnon-epithelialcells. However,preferentialexpression ofNramp2isoforms1andIIisnotnecessarilymutually exclusive:simultaneousexpression ofbothisoformsatthemRNAslevelhasbeen observedinseveraltissuesincludingkidney,thymus,andliver 40;155;160. Isoform1isexpressedattheapicalmembraneofduodenalbrushborder enterocytes 161,whereitisresponsiblefortheuptakeofdietaryinorganiciron(Figure4). Immunostainingexperiments oftissuesectionsshowthatNramp2expressionislimitedto thevilliandabsent inthecryptscells161-165.Nramp2expressionappearstobestrongest 22 2+••Fe.••F3+•e ~~e 01 l MNramp2/DMT1 •Ferroportin '"Ferrireductase ...Hephaestin -:Iron Figure4.Absorptionofnon-hemeironattheintestinalbrushborderandin reticulocytes.DMT1/Nramp2isoform1isexpressedattheapicalmembraneof intestinalenteracytes,whereitmediatestheabsorptionofnon-hemedivalentiranthat hasfirstbeenreduced byaputativeferrireductase(1).Atthebasolateralmembrane, Ferroportin(Fpn)isresponsiblefortheexport ofiranfromtheenterocyteintocircula­ tion(2).FolrowingoxidationatthebasolateralmembranebytheferroxidaseHephaes­ tin,ironisrapidlyboundtotransferrin(3).Non-epithelialcellssuchaserythroid precursorsuptakeironprimarilythroughinternalization ofthetransferrin-transferrin receptorcomplex(4).Endosomalacidificationfacilitatestherelease ofironfrom transferrinandcreatestheprotongradientrequiredforirontransportbyNramp2 isoformIlacrasstheendosomalmembraneandintothecytosol(5).Ironisthen transportedintothemitochondriafor hemesynthesisorstoredboundtoferritin. Apotransferrin-transferrinreceptorandNramp2isoform Ilaresubsequentlyrecycled backtothe ceUsurface. intheapicaltwo-thirdsofthevilliwhereitislimitedtoenterocytesthatmakeupthe columnarabsorptiveepithelium ofthemucosa.Inthesecells,Nramp2localizesprimarily totheapicalplasmamembraneknownasthebrushborder.Thispattern ofexpressionwas alsoobserved inthehumanintestinalcelllineCaco-2163;166-168.Intheintestine,Nramp2 isoform 1expressionisup-regulatedinresponsetodietaryirondepravationandseems repressedunderironoverloadconditions161.lmmunohistochemicalstudies byCanonne­ Hergauxandconeaguesshowedthatin mk/mkenterocytes,Nramp2expressionis upregulated butisnotproperlytargetedtothebrushborder147. Incontrast,isoformIIofNramp2ispredominantlyexpressedinperipheralornon­ epithelialcellssuchasinerythroidceIls.StudiesinprimaryasweIlastransfectedceIl lineshaveshownthatisoformIIisnotonlyexpressedattheplasmamembrane ofthese censbutalsointransferrinreceptorpositiverecyclingendosomes42;51;146.Inmice, reticulocytesandreticulocyteprecursorsceIls(whichgiverisetomatureoxygen-carrying redbloodcens)arethemajorphysiologicalsite ofisoformIIexpression,wherethe transporterisdetectedasa70 kDaspeciesandhasbeenshowntocolocalizewith transferrinreceptor 169.Duringerythropoiesis,precursorredbloodceIlshavealarge requirementforironforhemebiosynthesisandtheproductionofhemoglobin.Thehigh level ofexpressionofisoformIIinthesecenspermitsthecoordinateduptakeofironinto recyclingendosomesviathetransferrincycleandthetransport ofironacrossthe endosomalmembraneintothecytosol byNramp2(Figure4).AlthoughNramp2 expressioninreticulocytesdoesnotseemto beregulatedbybodyironstatus,treatmentof micewithphenylhydrazineorerythropoietin(toboosterythropoiesis)inducesthe productionofNramp2-expressingreticulocytes 170.However,reticulocytesfrommk/mk miceexpressnodetectableNramp2protein. Macrophagesareanothersource ofsignificantNramp2isoformIIexpression42. Thesephagocyticcellsplayacriticalroleintherecyclingofhemeironthrough phagocytosis ofsenescentredbloodcens.Nramp2mayindeedtransportphagosomaliron intothecytosolafterhemoglobindegradation.Immunoblottingexperimentshaveshown thatNramp2isexpressedasaglycosylated70-90 kDaproteininmacrophages42. ImmunofluorescenceexperimentsrevealedthatNramp2islocalizedprimarilytotheearly 24 endosomecompartmentwithsornepresenceinLampI-positivelate endosomes/lysosomesintransfected RAW264.7macropagesandtwomurineSerolicell lines ofthetestis(TM4and15P-l)171.Immunofluorescencestudieshavealsoshownthat Nramp2can beassociatedwithphagosomalmembranesduringphagocytosis42.More recently,Jabado etalshowedthatNramp2canassociatewitherythrocyte-containing phagosomes intransfectedRAW264.7macrophages171. Inthekidney,Nramp2proteinhasbeendetectedinmicrosomalmembrane fractionsasa70-75 kDamembraneprotein161.Inmousekidneys,Canonne-Hergauxand colleaguesusedimmunohistochemicalstainingwithanaffinitypurifiedanti-Nramp2 antibodytodetectNramp2inthecortexbutnotthemedulIa,andatthebrushborderand apicalpoles ofepithelialcellsoftheproximaltubule172•Inratkidneys,Fergusonand colleaguesdetectedNramp2expressionintheintracellularvesicles ofproximaltubule cells ofS3segment,collectingducts,thickascendinglimbsofHenle'sloopand,more intensely,attheapicalmembrane ofdistalconvolutedtubules173.Recently,Abouhamed etalreportedthedetectionofNramp2inLampl-positivelateendosomesandlysosomes withinratproximaltubulecells 174.Becausethesestudiesusedpolyclonalantibodies recognizingtheNterminus ofNramp2,theexactisoform(1orII)expressedinthekidney couldnot bedetermined.AlthoughthefactCanonne-Hergauxetalcouldnotdetectany significantexpressioninthekidneyusinganantibodyspecificallyrecognizingisoform il stronglysuggeststhatthemajorityofNramp2presentinthekidneyisindeedisoform1 172.Insharpcontrasttotheintestine,Nramp2mRNAandproteinexpressionappearsto beonlymodestlyaffectedbydietaryirondeprevationandisnotaffectedbystimulationof erythropoiesis172;175.Furthermore,whileanemicmkandBelgrademutantsshowstrong upregulation ofNramp2intheintestine,theyshowverylowNramp2expressioninthe kidney 172;176. Inthebrain,workbyGunshinandcolleaguesshowedNramp2mRNAis expressedinmostneuronsatlowlevelsbutnotinglial orependymalceUs40.Stronger Nramp2expressionwasdetectedindenselypackedcellgroups,suchasthehippocampal pyramidalandgranulecelIs,cerebellargranulecelIs,thepreopticnucleusandpyramidal cells ofthepiriformcortex40.StrongNramp2mRNAexpressionwasalsoseeninthe 25 ventralportionoftheanteriorolfactorynucleusandintheepithelialcellsofthechoroid plexus 40.Attheproteinlevel,BurdoandcolleaguesdetectedNramp2inneuronsof striatum,cerebellum,thalamusandintheependymalandvascularcellsthroughouttherat brain 177.ImmunohistochemicalstudiesbyMoosandMorganinratsconfirmedthat Nramp2proteinisexpressedinneuronsandthechoroidplexus, butnotinbraincapillary endothelialcells orinmacro-ormicroglialcells178.Sornestudieshavesuggestedthat Nramp2expressioninneuronalcellsisregulated bythehypoxiaregulator.Recentstudies byLisandcolleagueshaveshownthathypoxiaselectivelyincreasesexpressionof Nramp2,particularlytheexonlAcontainingisoformsofthetransporter,inrat pheochromocytoma (PCI2)cells179. Intheliver,themainsiteofbodyironstorage,Nramp2mRNAexpressionhas beendetectedinanumber ofcellstypes.Usinginsituhybridizationandquantitativereal­ timePCR,ZhangandcolleaguesdetectedNramp2 mRNAexpressioninratliver hepatocytes,Kupffercells,sinusoidalendothelialcells(SECs),andhepaticstellatecells (RSCs) 180.Attheproteinlevel,immunohistochemicalstudiesbyTrinderandcolleagues detectedNramp2 ontheplasmamembranesofrathepatocytesandthisexpression appearedtoincreaseanddecreaseinresponsetoironoverload orirondeficiency, respectively 162.TrinderetalsuggestedthatupregulationofNramp2intheliverduring ironoverloadcouldreducetherisk ofincreasedironintakeandcelldamageelsewherein thebody. Nramp2hasalsobeendetectedinseveralothertissues,althoughthedistribution ofthetransporteratthesesiteshasbeenlessextensivelycharacterized.Inthethymus, Nramp2 mRNAwasdetectedinthecortical,butnotmedullary,thymocytes40.Inthe testis,Nramp2 mRNA40andprotein171expressionhasbeendetectedintheSertolicells ofseminiferoustubules.Nramp2proteinhasalsobeendetectedintheplacenta,bothin thecytoplasmatthejunction offetalmembraneandfetalvessels181.Inthisstudy, GeorgieffandcolleaguessuggestedthatNramp2 maytransportendosomalferrousiron intothecytoplasm ofthehumansyncytiotrophoblasttosupplythefetus181. 26 1.3.7SubcellulartargetingofNramp2 AlthoughbothNramp2isoforms1andIIareexpressedattheplasmamembrane, immunolocalizationstudiesintransfectedcellshaveshownthatisoforms 1andIIexhibit differentsubcellularendosomaltargetingatsteady-state. IntransfectedLLC-PKI,CHO and RAW264.7cells,isoformIIismostlyexpressedinearlyandrecyclingendosomes 42;51.Studiesusinganexofacially-taggedNramp2moleculehaveshownthatisoformII moleculespresentatthecellsurfaceandinrecyclingendosomesareindynamic equilibrium,withsurfacetransportersbeingcontinuouslyinternalizedviaaclathrinand dynamin-dependentprocess 51.Incontrast,isoform1hasbeendetectedinLampl-positive lateendosomesandlysosomesintransfectedHEp-2andLLC-PK Icells155;182.Lysosomal expression ofisoform1inphagocyticcellssuchasmacrophagesmayberequiredforthe recycling ofironfromsenescentredbloodcells.WorkdescribedinChapter6ofthis thesiscomparesNramp2isoforms 1andIIwithrespecttofunction,subcellular localization,endocytosiskinetics,andfateuponinternalizationintransfectedLLC-PK I cells.Thisworkshowsthatisoform1isexpressedathigherlevelsatthecellsurface comparedtoisoformIIduetoaslowerrate ofinternalizedfromtheplasmamembrane.In addition,isoform1isnotefficientlyrecycleduponendocytosisandistargetedtolate endosomesandlysosomes 182.Highersurfaceexpressionisadvantageousforisoform1- expressingepithelialcells,whichmustabsorbFe2+and possiblyotherdivalentmetals acrosstheirplasmamembranes.Thus,alternativesplicing ofNramp2criticallyregulates thesubcellularlocalizationandsite ofmetaltransport. Cytoplasmicmotifssuchastyrosine-based(NPXY, YXXsignaIs,wereshownto
haveanadditionalbutlessdeterminateeffect
onsubcellulartargetingofthetransporter
184
1.3.8Nramp2andcellularironhomeostasis
Ironisanessentialelementforlife,playingavarietyofrolesinorganismsranging
fromthetransportation
ofoxygentothegenerationofanti-microbialreactiveoxygen
species.One
ofthekeypropertiesofironisitsabilitytomediateoxidation-reduction
reactions
byalteringbetweenitsFe3+andFe2+oxidativestates.However,itisthishighly
reactivenaturethatmakesanexcess
of”free”ironverytoxictoorganismsand
necessitatesstrictmechanisms
ofregulation.
1.3.8.1Mechanismsofcellularironabsorption
Highlyreactiveironinitsionicformisnotnormallyfoundathighlevelsin
normalorganisms.Rather,themajority
ofcirculatingiron(Fe3+)existsinaninertform
boundtotheabundantplasmaglycoproteintransferrin(Tf).Mostcellsgainaccesstoiron
byintemalizingtheentirediferric-Tfcomplexthroughtheubiquitously-expressed
transferrinreceptor
1(TfRl),whichresidesinc1athrin-coatedpitsonthesurfaceofcells.
Diferric
TfbindsTfRlattheplasmamembraneandentirecomplexisintemalizedby
clathrin-dependentreceptor-mediatedendocytosisintoearlyendosomes.Endosomal
acidification(pH5.5-6.0)
byvacuolarW-ATPasestrengthenstheinteractionbetweenTf
andTfRlbutweakensthebindingofirontoTf,causingthereleaseofFe3+inthe
endosomallumen.Arecentlyidentifiedferrireductase,
Steap3,reducesFe3+toFe2+inthe
endosomallumen
185;186.Theacidicenvironmentintheendosomealsoprovidesthe
protongradientforNramp2totransportFe
2+acrosstheendosomalmembraneintothe
CytOSOI141;169.Apo-TflTfRlaswellasNramp2arethenrecyc1edbacktothecellsurface,
28

wheretheycanbeusedinfurthercyclesofironuptake51(Figure4).Althoughmostcells
arethoughttouptakeironviatheTf-
TfRlcycle,thisprocessisofparticularimportance
fordevelopingerythrocyteswhichhaveanenormousneed
ofironforhemoglobin
synthesis.Ahomolog
ofTfRl,transferrinreceptor2(TfR2),isexpressedexclusivelyin
erythroidcells,hepatocytes,andduodenalcryptcells
187.TfR2alsofunctionsinthe
uptake
ofTf-iron,althoughitbindsTfwithlessaffinitycomparedtoTiRlandisnot
regulated
bytheIRP/IREregulatorysystem.TfR2clearlyplaysacriticalroleiniron
homeostasisasmutationsatthehuman
TjR2generesultinsystemicironoverload188.
PolarizedepithelialcellsofthekidneyareabletouptakeTf-ironbyamegalin-dependent
mechanismusingtheTf-bindingendocyticreceptorcubilin
189.
ThereareseveralTf-independentcellularmechanismsofironuptake.Brush
borderintestinalenterocytesuptakedietarynon-hemeirondirectlyacrosstheirapical
membranesthroughtheaction
ofNramp2(Figure4).Non-hemeironisfoundin
vegetablesandgrainproductsandexistsinitsinsolubleFe3+form,whichmustfirstbe
convertedtoFe
2+beforeitcanbetransportedbyNramp2.Thisisthoughttooccur
throughtheaction
ofacytochromeb-likeferrireductasecalledDcytb(CybrdJ),whichis
expressedattheapicalenterocytemembrane
190.Otherredundantferrireductasesmay
alsoexistsincestudiesinvolvingtargeteddisruption
oftheCybrdlgeneinmicesuggest
Dcytbisnotessentialforintestinalironabsorptioninmicefedanormalirondiet
191.
Gastricacidsecretionbythestomachpromotesthesolubilityofironcomplexesandalso
providestheprotongradientrequiredforNramp2-mediatedFe
2+uptake.
Ironisrequiredtomaintainproperneuronalcellfunction,however,anexcess
of
ironcausesneurodegeneration.EarlyimmunohistochemicalstudiesbyMoosclearly
showedadiverseexpressionoftransferrinreceptorthroughoutthecentralnervoussystem
(CNS)
192.Morerecently,severalindependentgroupshaveshownthatNramp2isalso
expressedinneuronalcells
40;177;193,suggestingthatTiRandNramp2workin
conjunctioninthesecellstouptaketransferrin-boundiron.Irontransportedthroughthe
blood-brainbarriermayexistinthebraininterstitialfluidinalowmolecularweight
form,suchasiron-citrate.Thepresence
ofnontransferrin-boundironinbrain
29

extracellularfluidssuggeststhatneuronscanalsotakeupironinatransferrin-freeform
194-197.ThismaybemediatedbyNramp2expressionattheneuronalplasmamembrane.
Ironre-absorptioninthekidneyisnotcompletelyunderstoodyetstudies
by
Canonne-Hergauxetal172andFergusonetal173suggestthatapicallyexpressedNramp2
mayfunctionaspart
ofanironre-absorptionsysteminthekidney,possiblytopreventthe
loss
ofpreciousironintheurine.ThisideaissupportedbytheobservationthatNramp2
expressioninthekidneyincreasesapproximatelytwo-foldinmicefedalow-irondiet
172.
Furthermore,Wareingandcolleagueshaveshownusingmicroinjectionand
microperfusiontechniquesthatincreasedNramp2expressioninratkidneywas
accompanied
byadecreaseinurinaryironexcretionrateandviceversa198.Interestingly,
mk/mkmiceandb/bratssufferingfromseveremycrocyticanemiashowalmosta
completeloss
ofNramp2expressioninthekidney172;176.However,anotherstudyby
Fergusonetalshowedthatalthoughb/bratsexhibithigherserumironlevelsthan+/b
rats,bothanimaIsdisplaysimilarurinaryironexcretionrates,suggestingthatNramp2is
notplayacriticalroleinironre-absorption
bythekidney176.
Dietaryhemeironisanimportantnutritionalsourceofironincarnivoresand
omnivoresthatismorereadilyabsorbedthannon-hemeiron.Ironisreleasedfromheme
throughthebreakdownofhemoglobinandmyoglobinpresent
inredmeat.Inmammals,it
hasbeenknownforsometimethatproximalduodenalenterocytesandhepatocytesare
themajorsites
ofhemetransport199;200.Themechanismofhemetransportacrossthe
plasmamembraneremainedelusiveuntilShayeghiandcolleaguesrecentlyidentifieda
duodenalmembraneprotein,hemecarrierprotein1(HCPI),thatmediatescellularheme
uptake
201.HCPIshowssequencesimilaritytoseveralbacterialmetal-tetracyclin
transporters,isupregulatedisresponsetohypoxia,andappearstoberecruitedtothe
plasmamembranefromthecytoplasmuponirondeficiency
201.Upontranslocationacross
theplasmamembrane,hemeappearstolocalizeinmembraneboundvesic1eswithin
thecytoplasm
199;200;202;203;203beforeitisdegradedbytheenzymehemeoxygenase(HO-l
andHO-2)toyieldferrousiron
204.
Smallpeptidesiderophoresaresecretedbymicro-organismstotrapandimport
ironacrossthecellmembrane.Likethisbacterialiron-uptakesystem,mammalianneural
30

gelatinase-associatedlipocalin(NGAL/24p3)hasbeenreportedtocomplexwithironand
beintemalizedandtraffickedtolateendosomes205.RecombinantNGAL/24p3loaded
withironhasbeenshowntodeliverthemetalintoculturedcellsandaffecttheexpression
ofiron-regulatedgenes.ThisNGAL/24p3deliverypathwayappearstobeimportantfor
differentiatingepithelialcellsduringearlyembryonicdevelopment
205;206.
Finally,tissuemacrophagesabsorbironindirectlythroughthephagocytosisofold
ordamagederythrocytes.Engulfedredcellsarelysed,hemoglobinisdegraded(withthe
help
ofhemeoxygenase),andthereleasedironiseitherstoredboundtoferritinor
exportedtoreloadcirculatingapo-Tf.Therecycling
ofironbymacrophagesisahighly
efficientprocessthatrepresentsthemajorsource
ofironforcellularprocessessuchas
erythropoesis.
1.3.8.2Mechanismsofcellularironstorageandusage
Onceimportedintothecell,excessironthatwillnotbeimmediatelyusedis
storedboundtotheubiquitousproteinferritin
207.Ferritinisahighlyconserved
multimericproteinthatisabletosequesterinorganicFe
2+byfirstconvertingittoFe3+
beforestoringitasthechemicallylessreactiveferrihydrite208.Mammalianferritin
consists
of24light(L)andheavy(H)chainsubunitsthatcanaccommodateupto4500
atoms
ofiron209.Whencellsexperienceagreaterrequirementforiron,ferritinis
degradedandironisreleased,howeverthisprocessisnotcurrentlywellunderstood.
Most
oftheironwithincellsisdirectedtothemitochondriaforeitherheme
biosynthesis
ormaturationofiron-sulfur(Fe-S)clusters.Herneissynthesizedbyan
enzymecalledferrochelatase,whichcatalyzestheinsertion
offerrousionsinto
protoporphyrinIX(reviewedin
210).Hernesynthesisisofparticularimportancein
oxygen-carryingerythroidcells,howeverthemechanism
bywhichironisdeliveredto
mitochondriaisnotcurrentlyknown.Recently,Zhangandcolleaguesproposedthata
transientinteractionbetweeniron-richendosomesandmitochondriaisresponsibleforthe
transfer
ofironacquiredfromtransferrintoferrochelatase211.Newlysynthesizedhemeis
rapidlyexportedoutsidethemitochondriaintothecytosolandendoplasmicreticulum,
andassociateswithapo-hemoproteins
212;213.Fe-Sclustersareubiquitousandparticipate
31

inanumberofproccsscsincludingclcctrontransfer,substratebindinglactivation,and
ironlsulfurstorage
214.Themitochondrialmatrixproteinfrataxinisrequiredfor
biogenesis
ofFe-Sproteinsandisthoughttoplayrolesinmitochondrialironexportand
storage
215-217.Mutationsatthefrataxingenecausetheneurodegenerativedisorder
Friedreich’sataxia.PatientswithFriedreich’sataxiapossessadeficiency
inFe-S
containingproteinsanddisplaymitochondrialironoverload
218.
1.3.8.3Mechanismsofironexport
Becauseironissuchapreciouselement,thecurrentdogmaisthatthereisno
excretionmechanismforiron,withlossfromthebodynormallyresultingfrom
desquamation
orthelossofbiologicalfluids(especiallyblood).Atthecellularlevel,in
contrasttothenumerousmechanismsidentifiedforcellularironimport,muchlessis
knownaboutthemechanismsunderlyingcellulariron
eXPOrt’Infact,theonlyputative
mammalianironexporteridentifiedtodateisferroportin(MTP,Ireg
1)219-221.Ferroportin
isexpressedatthebasolateralmembrane
ofduodenalenterocyteswhereitmediatesiron
(Fe2
+)eXPOrtinto
thebloodstreaminconjunctionwiththeferroxidasehephaestin,which
convertsFe2
+to
Fe3+before
itcanbeboundbycirculatingapo_Tf222.Mutationof
hephaestininsex-linkedanemia(Sla)miceleadstoironaccumulationintheepitheliaand
resultsin
Slamicehavingasystemicirondeficiency222.Innon-intestinalcells,iron
exportrequiresoxidation
bytheplasmaproteinceruloplasmin223.Ferroportinisalso
highlyexpressed
inintracellularvesic1esoftissuemacrophages,whereitisthoughtto
playaroleinthe
eXPOrtofironrecyc1edfromengulfedredbloodcells.Whenferroportin
expressionisupregulatedthroughirontreatment
orerythropagocytosis,ferroportin
expressionisstronglyenhancedattheplasmamembrane
ofmacrophages224.Incontrast,
treatmentwiththeironregulatoryproteinhepcidin(seebelow)causesrapid
intemalizationanddegradation
ofthemacrophageironexporter224;225.
1.3.8.4Regulationofcellularironhomeostasis
Proteinsinvolvedinironuptake,usage,storageandexportrequiretight
regulation.Thebestcharacterizedform
ofiron-dependentregulationisatthelevelof
32

mRNAstability.Ironresponsiveelements(IREs)areconservedhairpinstructures
approximately30nuc1eotidesinlengthfoundineitherthe
5’or3’untranslatedregions
(UTRs)
ofmRNAsencodingproteinsinvolvedinironhomeostasissuchasNramp2,TfR,
ferroportin,andferritin.Specializedironregulatoryproteins(IRP
l,IRP2)interactwith
theIREsbasedoncellularlabileironpoollevels.Ironlevelsregulatethebinding
ofIRP1
andIRP2toIREs
bydistinctmechanisms226.Whencellularironlevelsarehigh,iron(as
Fe-Sclusters)bindsIRPIandinhibitsitsbindingtoIREs
byconvertingittoanaconitase,
whichinteraconvertscitrateandisocitrate.Whencellularironstoresarelow,IRPIisable
tofreelybindIREs.IRP/IREinteractionsarealsocontrolled
byotherfactorsincludingthe
presence
ofreactiveoxygenspecies,nitricoxide,andhypoxia.Themechanismbywhich
ironregulatesIRP2islesswellunderstood.UnlikeIRPI,IRP2doesnotbindFe-S
clustersbuthasbeenshowntoaccumulateiniron-starvedcellsandistargetedfor
degradationwhenironlevelsarehigh
227.IRP2isalsosensitivetodegradationinthe
presenceofnitricoxide(NO),whereasIRPIisactivated
byNO228;229.Recentstudies
havesuggestedthatIRP2isthepredominantregulator
ofironhomeostasisinmammalian
cells.Indeed,micedeficientinIRPIappearnormalyetmicedeficientinIRP2show
pronouncedmis-regulation
ofironmetabolismandnervedamage230;231.
SingleIREsarefoundinthe5’UTRsofmRNAsencodingferritin,erythroid
5-aminolevulinicacidsynthase(requiredforhemebiosynthesis),mitochondrialaconitase
(acitratecycleenzyme),andferroportin
220;221;227;232.TheformationofIRP/IRE
complexesonthe
5’UTRofthesetranscriptsinhibitstheirtranslation,leadingtodown­
regulation
oftheprotein233.TheoppositeresultoccurswhenIREsarelocatedinthe3′
UTRoftranscripts.,MultipleIREsarepresentinthe3’UTRofmRNAencodingTfRl
andbinding
ofIRPservestostabilizethetranscript,leadingtoincreasedtranslationand
up-regulation
oftheprotein227,The3’UTRofmRNAencodingNramp2isoform1
containsasingleIREthathasbeenshowntobindIRPI
invitro234,However,thisIRE
hasbeenshowntoonlymarginallyregulateNramp2expressioninculturedcells
234.The
recentlyidentified
5’promoter/exonlAregionofNramp2appearstoworkinconjunction
withthe
3’UTRIREmodulateNramp2expressioninresponsetothecellularand
systemicironlevels
154.
33

Genesinvolvedinironhomeostasisarealsoregulatedatthetranscriptionallevel.
Forexample,cytokineproductionhasbeenreportedtomodulateexpressionofvarious
iron-relatedgenes.Interferon-y
(INF-y),tumornecrosisfactor-a,interleukin-l(IL-l),and
IL-6haveallbeenreportedtostimulateferritinexpressionbutdown-regulate
TfRl207.
INF-yandlipopolysaccharide(LPS)havebeenreportedtoinduceNramp2expressionbut
inhibitferroportinexpressioninactivatedmonocytes
235;236.Theresponseofferroportin
toLPSappearstorequiresignalingthroughtheLPSreceptor,Toll-likereceptor4(TLR-
4)
236,andmaycontributetotheironsequestrationbyphagocyticcellsobservedduring
inflammation.
1.3.9Nramp2andsystemicironhomeostasis
Over65yearsago,studiesonironbalanceinhumansshowedthatvirtuallyno
ironisexcretedandthatstableironlevelsaremaintained
bymodulatingabsorptionof
ironfromthegut.Insufficientorexcessiveaccumulationofironcanbeproblematic,
especiallycoupledwiththefactthatmammalslackanaturalphysiologicalexcretion
systemforiron.Therefore,organismshavedevelopedsophisticatedmechanismstosense
andregulatebodyironlevelsinresponsetothedemandingneeds
ofsystemicprocesses
suchaserythropoesis.Themaintenanceofhealthyironhomeostasisrequiresefficient
communicationbetweenthecellsthatusethemostiron(erythroidprecursors)andthe
cellsthatobtainandstoreiron(duodenalenterocytes,tissuemacrophages,hepatocytes).
1.3.9.1Systemicironregulators
Theprimaryironstoragesitesaretissuemacrophagesandhepatocytes.Ironis
storedboundtoferritininthesecellsandthelevel
ofthesestoresismonitoredand
maintained
byaso-called”storesregulator”237;238.Thestoresregulatorcan,inpart,alter
intestinalironabsorptioninresponsetobodystores.Erythroidcellsarebyfarthemajor
consumers
ofironandtheamounttheironpresentinthebonemarrow,erythroid
precursors,andcirculatingredbloodcellsisnormallygreaterthantheamount
ofiron
presentinthestores.Whensystemicirondemandforerythropoeisisexceedswhatis
availableinthestoragepools,an”erythroidregulator”up-regulatesintestinalironuptake
34

tocompensateandreplenishthestores238.A”hypoxiaregulator”isaIsebelievedteexistl
alteringironhomeostasisinresponsetoinsufficientbloodandtissueoxygenlevels.
Meanwhile,an”inflammatoryregulator”isthoughttomediatetheretention
ofironwithin
tissuemacrophagesasweIlasthereduction
ofintestinalironabsorption,asahost
responsetoinfection
orinflammation.Thisactionisbelievedtopromotehostresistance
byrestrictingtheamountofironavailabletopathogens.
Itisclearthatahierarchyexistsamongtheregulators.ThisisespeciaUyevidentin
diseaseswhereerythropoesisisiron-deficientyetthe
bodyironstoresareoverfilled.
Hypotransferrinemicmice
(Tr!PX),whichpossessamutationdisruptingasplicedonor
siteinthe
transferringene,haveaneartotaldeficiencyincirculatingserumtransferrin
239.Erythroidprecursorsuptakethemajorityoftheirironasdiferric-TfviaTfR!and
thereforeerythropoesisisseverelyimpaired
inTr!pxmice150.However,sinceanumber
ofnon-erythroidceUsandtissuesreadilyuptakenon-transferrinboundiron,ironoverload
developsinvariousorganssuchastheliverandpancreas.Inthiscase,signaIsfromthe
erythroidregulatorseemtooverridesignaIsfromthestoresregulator,up-regulatingthe
absorption
ofdietaryironintheduodenum.Thebodyappearstoprioritize(andrightfully
so)ironneededforerythropoesisovertheneedtoavoidironoverload.
Ananalogous
scenariooccurs
intwoothermousemodelsofironoverloadgeneratedbytargeted
disruption
oftheHfeandf3-2microglobulingenes240;241.
Thesearchforregulatoryeffectorsthatcanmodulateintestinalironabsorption,
ironrecycling
bymacrophages,andironstoragebyhepatocyteshasbeenextensive.This
searchledtotherecentidentification
ofhepcidin(HAMP,LEAP-l),asmaUcysteine-rich
peptidehormonethatisproduced
byhepatocytesandexcretedthroughthekidneys242-244.
Producedfromthecleavageofalargerprecursorprotein,hepcidinshowssequence
similaritytopeptides(defensins)involved
ininnateimmunityandpossessesitselfanti­
microbialproperties
242;243.HepcidincontrolsextraceUularironlevelsbyregulatingits
intestinalabsorption,placentaltransport,recyc1ing
bymacrophages,andreleasefrom
stores.Hepcidin-deficient
(USF2-1-)micedevelopelevatedbodyironstoreswhile
transgenicmiceconstitutively-expressinghepcidineventuallydiefromsevereiron
deficiencyanemia
245;246.Hepcidinalsoappearstobeanacutephaseresponsepeptide
35

induceddirectlybyvariousinflammatorystimuliincludingbacteriallipopolysaccharide
247-250.Severalstudieshavelinkedhepcidintotheerythroid,stores,hypoxia,and
inflammatoryregulators(reviewedin
251).Althoughthesignalingpathwaysofthese
mechanismsarenotcompletelyunderstood,moststudiesare
inagreementthatwheniron
isscarce,decreasedserumhepcidinexpressionisassociatedwithenhancedironrelease
fromintestinalcellsandmacrophages
248;252;253.Conversely,inastateofironoverload,
increasedserumhepcidinexpressionisassociatedwithdecreasedironreleasefrom
intestinalcellsandmacrophages
244;254;255.Recently,hepcidinwasfoundtoinhibit
cellularironefflux
bybindingtotheironexporterferroportinandinducingits
internalizationfromthecellsurfaceanddegradation
224;256.Mutagenesisstudieshave
revealedthattheNterminus
ofhepcidinisresponsibleforitsinteractionwithferroportin
257
1.3.9.2ImportanceofNramp2insystemicironhomeostasis
Becauseofitsdualroleinintestinalironabsorptionaswellasinacquisitionof
transferrin-iron,mutationsinNramp2severelyaffectsystemicironhomeostasis.Bothmk
miceandBelgraderats,whichpossesstheidenticalmutation(G185R)inNramp2,suffer
fromseveremicrocyticanemia
136;141.Theneedforincreasederythropoesisinthese
rodentspromptstheerythroidregulatortoincreaseintestinalironabsorption
byup­
regulatingNramp2expressionattheduodenalbrushborder
147.However,becausethe
G185RmutationattenuatesNramp2function,theanemiapersists
146;148.
RecentworkbyGunshinandcolleagueshasfurtheredourunderstandingofthe
roleofNramp2invivo.TheyshowedthatglobalinactivationofNramp2inmiceresults
inasimilarbutmoreseverephenotypethanthatseenforanimaIshomozygousforthe
G185Rmutation
258.AlthoughNramp2-1
-micewerebornalivewithnoapparent
anatomicalabnormalities,theywerenoticeablypaleratbirthandlaterdisplayed
progressivepost-natalgrowthretardation,withnomicesurvivingmorethan7days
258.
TherelativelynormaldevelopmentoftheNramp2-1
-miceinuterosuggestedthatfetal
Nramp2isnotessentialformaterno-fetalirontransfer.Transplantinghematopoeticstem
cellsfrom
Nramp2-1
-miceintoirradiatedWTmiceresultedinabnormalerythrocyte
36

morphologyanddecreasedhemoglobinlevels,confirmingtheimportanceofNramp2in
erythroidironutilization
258.Furthermore,studiesinvolvingselectiveinactivationof
Nramp2intheintestineconfirmedtheimportanceofNramp2intheabsorptionofdietary
non-hemeiron
258.
1.3.9.3DiseasesassociatedwithNRAMP2mutationsinhumans
Currently,therearethreereportedcasesofmutationsinhumanNRAMP2
associatedwithiron-relateddisorders.Thefirstcasewasidentifiedin2004,when
Priwitzerovaandcolleaguesreporteda20-year-oldfemalepatient
ofCzechorigin,the
product
ofaconsanguineousunion,whosufferedfromliverhemosiderosisandsevere
congenitalhypochromicmicrocyticanemiaduetodefectiveerythroidironuse
259.Ayear
later,Mims
etalreportedthatthispatientwasinfacthomozygousforaGtoC
substitutionatthelastnuc1eotide
ofexon12ofNRAMP2(NRAMP2G1285C)4.This
mutationhadadualeffect,impairingnormalsplicing
ofexon12byanestimated90%
andintroducinganaminoacidsubstitution(E399D)intheremaining
(-10%)properly
splicedtranscriptfoundinthepatient.Theimpairedsplicingisthoughttoresultfrom
alteration
oftheconsensussequenceforbindingoftheUlsnRNPattheexon12/intron12
boundaryandtranslatesintoatruncated,non-functionalNramp2protein
4;260;261.
AlthoughE399Drepresentsaconservativemutation,E399residesinthehighly
conservedfourthintra-cytoplasmicloop
ofNramp2definedastheconservedtransport
motif.WorkdescribedinChapter3
ofthisthesisexaminedtheeffectsoftheE399D
mutationaswellasothermutationsatthatpositionintransfectedLLC-PK
1cells.These
resultsshowedthattheE399Dmutationdoesnotinitselfaffectexpression,function,or
targeting
oftheNramp2protein,andthatthereducedNramp2functioninthepatientis
likelycaused
byaquantitativereductioninNRAMP2mRNAlevelsduetoimproper
splicing
262.Thisresultwasconfirmedbystudiesfromothergroups260;261.
Inearly2006,Iolasconandcolleaguesreportedasecondhumanpatientwhois
compoundheterozygousfortwonovelmutationsin
NRAMP2263.Thepatientisa5-year­
oldmale
ofltalianoriginsufferingfromseverehypochromicmicrocyticanemiawith
hepaticironoverload.Thepatientpossessedtwonovelmutations
inNRAMP2:a3base
37

pairdeletion(deICTT)inintron4andaCtoTtransition(NRAMP2c/246T)inexon13
resultinginanArgtoCyssubstitutionatposition416(R416C)ofNramp2.TheCTT
deletionwasshowntodisruptnonnalsplicing
ofNRAMP2pre-mRNA,causingapartial
(30-35%)skipping
ofexon5263.Thisskippingresultsinatruncated,non-functional
Nramp2proteinthatisinheritedinafullyrecessivemanner.R416representsahighly
conservedresiduelocatedinTM9.WorkdescribedinChapter4
ofthisthesisexamined
theeffects
ofconservativeandnon-conservativemutationsatR416.Theseresultsshowed
thatnon-conservativesubstitutionsatR416(C,A,E)causemultiplefunctional
deficienciesincludingdefectiveproteinprocessing,loss
oftransportactivity,impaired
cellsurfacetargetingandrecyclingthroughendosomes,concomitantwithretention
ofthe
transporterintheendoplasmicreticulum
264.ThesefindingsdemonstratedthattheR416C
mutationrepresentsacompleteloss-of-functionandthataquantitativereductionin
Nramp2expressionisthecause
ofthemicrocyticanemiaandironoverloadinthepatient.
Recently,Beaumontandcolleaguesreportedathirdhumanpatientcompound
heterozygotefortwonovel
NRAMP2mutations,associatedwithmicrocyticanemiaand
progressiveliverironoverload
265.Inthefirstmutation,aGTGdeletioninexon5causes
thein-framedeletion
ofV114inTM2.Thesecondmutation,aG>Ttransitioninexon8,
causesaG212VsubstitutioninTM5.Theeffectsoftheserecentlyidentifiedmutationson
Nramp2functionarenotcurrentlyknownandneedtobestudied.AlthoughV114isnota
highlyconservedamongNramp2orthologs,itsin-framedeletionmaystillaffectNramp2
function.Ontheotherhand,G212inTM5issignificantlyconservedanditislikelythat
thenon-conservativeG212VsubstitutionaffectsNramp2function.
Together,thesethreereportedcasesdefineanewsyndrome
ofcongenital
microcytichypochromicanemiawithliverironoverload.
Inthefirsttworeports,a
quantitativereductioninNramp2expressionwasthecause
ofdisease,indicatingthe
presence
ofaminimumthresholdofNramp2expressionrequiredfornonnal
physiologicalfunction.Thisminimumthresholdappearsto
bebetweenthelowlevelof
activityobservedinbothNRAMP2mutantpatientsandthe50%activityretainedintheir
phenotypicallynonnalheterozygoterelatives
4;263.ResultsfromstudieswithNramp2-/­
micesuggestthatcompletelossofNramp2activitymaynotbecompatiblewithlife258.
38

1.3.9.4Otheriron-overloaddisordersinhumans
Hereditaryhemochromatosis(HH)isanautosomalrecessiveinheritediron
overloaddisordercharacterized
byincreasedintestinalironabsorptionandiron
accumulationinvitalorgans,eventuallyleadingtoorganfailure(reviewedin
266).When
identifiedbeforesevereorgandamagehasoccurred,hemochromatosiscanbetreated
by
phlebotomy(bleeding)toremoveiron-richredbloodcells.Mutationsinfourgeneshave
beenlinkedtoHH:HFE,TFR2,hemojuvelin,andhepcidin.
ThefirstgeneshowntobeassociatedwithHHwasHFE,anatypicalmajor
histocompatibilityclass1proteinthatheterodimerizeswith
(3-2microglobulinbutdoes
notbindasmallpeptide
267;268.Mostpatients(80%)withHFE-linkedhemochromatosis
arehomozygousforamissensemutation(C282Y)thatpartiallydisruptsHFEfunction
267.TheC282YmutationisparticularlycommoninindividualsofNorthemEuropean
descent.Othermutationsat
HFEarerareandtheircontributionstoclinicaldiseasehave
notbeenwellestablished.Onlyafraction
ofpatientswithahemochromatosisgenotype
developclinicalironoverloadsymptomsandenvironmentalfactorsplayasignificantrole
indiseasepenetrance.Despitevastefforts,theprecisemolecularfunction
oftheHFE
proteinhasremainedelusiveandexperimentsexpressingHFE
intransfectedcellshas
yieldedconflictingresults
268-271.However,itisclearthatHFEislinkedtothetransferrin
cycleandcanformahigh-affinitycomplexwith
TfR!,competingwithtransferrintobind
thereceptor
268;272;273.RecentstudiesinHfe-/-micehavelinkedHFEtohepcidinlevels
andsuggestedthetwoproteinsarepart
ofthesamesignalingpathway274-276.
AlthoughlesscommonthanHFEmutations,homozygousmutationsinTfR2have
alsobeenidentifiedinpatientswithhemochromatosis
188;277-282.Theclinicalsymptomsof
TFR2hemochromatosisaresimilarbutmoreseverethantheHFE-relateddisease283.A
definingcharacteristic
ofTFR2hemochromatosisisthepersistenceofhightransferrin
saturation,evenfollowingphlebotomy.
Sornepatientswithjuvenilehemochromatosisarehomozygousformutationsat
HAMP,thegeneencodinghepcidin284.Mutationsidentifiedthusfarcauseeithera
completeinactivation
oftheprotein285orsubstituteoneoftheinvariantcysteinesofthe
39

peptide286-288.However,mostjuvenilehemochromatosispatientspossessmutationsin
thegenehemojuvelin(HFE2orHJV)289.Hemojuvelinencodesa
glycosylphosphatidylinositol(GPI)-linkedprotein,characterized
byanRGDmotifanda
vonWillebrandtypeDdomainthatishighlyexpressedintheliver,skeletalmuscle,and
heart.Over30differentmutationshavebeenidentifiedinthe
HJVgeneinpatientswith
juvenilehemochromatosis(reviewedin
290).HJVhemochromatosissharesnumerous
featureswith
HFEhemochromatosis,butaUtheclinicalmanifestationsdevelopearlier
duetogreaterintestinalironabsorptionandfasterrate
ofironaccumulation.Thefactthat
patientswithcriticalmutationsin
HJVdisplayloworunmeasurablehepcidinlevels289
suggestthattheproteinisacomponentofthehepcidinregulatorypathway.
Adistincttype
ofironoverloadinheritedinanautosomaldominantfashionis
foundinpatientswithcriticalmutations
inferroportin291;292.Thesepatientspresent
variableclinicalphenotypesdependingonthespecifiemutationtheyharbor.
Ferroportin
mutationsessentiaUyfaUintotwomainclasses:(1)mutationsthatdisruptceUsurface
expressionandinhibitironexport
293;and(2)mutationsthatlocalizetotheceUsurface
butareunabletorespondtheregulatorypeptidehepcidin
294.Patientswiththefirsttype
ofmutationshowtypicalferroportindiseasewithlowtransferrinsaturationandiron
accumulationinmacrophages.Patientswiththesecondtype
ofmutationshowhigh
transferrinsaturationandearlyhepatocyteironloadingsimilartoclassichereditary
hemochromatosis.Thedominantinheritance
offerroportin-linkedhemochromatosismay
beexplained
byinvitroexperimentsthatsuggestferroportinfunctionsasamultimerand
thatmutantsactasdominantnegatives,affectingthebehavior
ofthewildtypeprotein295.
However,thisstillremainscontroversialsinceotherstudieshaveshownthatWTand
mutantferroporinsdonotformoligomersintransfectedHEK293ceUs
293;296.
40

PrefacetoChapter2
TheidentificationofNramp2asatransporterofdivalentmetalsin1997provided
newinsightintothefunction
ofNrampproteinsinvivo40;136,Electrophysiological
measurementsindirectlyidentifiedFe
2+,Mn2+,C02+,Cd2+,Cu2+,Ni2+,Pb2+,andZn2+as
substratesforNramp2andshowedthattransportoccurred
byapH-dependentmechanism
40,ThedemonstrationofdivalentmetaltransportbynumerousdistantNramporthologs
highlightedfunctionalconservationintheNrampsuper-family
1;2;297;298,Atthistime,
structure-functionstudiesbecameessentialtounderstandingthemolecularbasisforboth
metaltransportandpH-dependence
byNrampproteins,
Chapter2
ofthisthesisdescribesoneofthefirstsystematicattemptstoexplore
structure-functionrelationshipsinNrampproteins
bysite-specifiemutagenesis.Through
multiplesequencealignments
ofvariousprokaryoticandeukaryotieNrampsequences,
weidentifiedanumber
ofhighlyconservedyetthermodynamicallydisfavoredcharged
aminoacidsresidingwithinthehydrophobicTMdomainsofNramp2.Ourmutagenesis
studiesrevealedseveralchargedresiduesthatareessentialforNramp2-mediatedmetal
transport,aswellastwoinvarianthistidines(H267andH272)thatarecriticalforpH­
dependence.Later,similarmutagenesisstudiesperformedinabaeterialNramportholog,
MntH,confirmedtheimportance
ofTMchargedresiduesinmetaltransportbyNramp
proteins
299,Furthermore,ouridentificationofH267andH272ascriticalforpH­
dependeneewaslatersupportedbyelectrophysiologicalexperimentsinXenopusoocytes
300
41

Chapter2:
IrontransportbyNramp2/DMT1:pHregulation
oftransportbytwohistidinesintransmembrane
domain6
42

ABSTRACT
MutationsatNramplimpairphagocytefunctionandcausesusceptibilityto
infectionswhilemutationsat
Nramp2(DMTl)affectironhomeostasisandcausesevere
microcyticanemia.Structure-functionrelationshipsintheNrampsuperfamilywere
studied
bymutagenesis,followedbyfunctionalcharacterizationinyeastandin
mammaliancells.Thesestudiesidentifythreenegativelychargedandhighlyconserved
residuesintransmembranedomains(TM)1,4,and7asessentialforcationtransport
by
Nramp2/DMTl.Theintroductionofachargedresidue(G185R)inTM4foundinthe
naturally-occurringmicrocyticanemia
mk(mouse)andBe/grade(rat)mutantsisshownto
causeapartialorcompleteloss
offunctioninmammalianandyeastcells,respectively.A
pair
ofmutation-sensitiveandhighlyconservedhistidines(H267,H272)wasidentifiedin
TM6.Surprisingly,inactiveH267andH272mutantscould
berescuedbyloweringthe
pHofthetransportassay.ThisindicatesthatH2671H272arenotdirectlyinvolvedin
metalbindingbutrathertheyplayanimportantrolein
pHregulationofmetaltransportby
Nrampproteins.
43

INTRODUCTION
Nramp2(DMTl,DCTl)isessentialfornutritionaliron(Fe2+)uptakebythe
duodenumbrushborder
40;136;141andforirontransportacrosstheendosomalmembranein
peripheraltissues
42;141;301,Nramp2isanintegralmembraneproteincomposedof12
predictedtransmembrane(TM)domains
1.TheNramp2geneproduces2rnRNAsby
alternativesplicingoftheterminal3’exonthatshowdifferent3’untranslatedregions
containing(isoform
I,+IRE)ornot(isoformII,-IRE)anironresponseelement(IRE),as
wellasdistinctC-terminalproteinsequences
153;302,Innormaltissues,Nramp2protein
(isoform
1)isexpressedatthebrushborderoftheproximalportionoftheduodenum,
whereitisregulated
bydietaryirondeprivation161.Nramp2(isoformII)isalsopresentin
the
Tfreceptorpositiverecyc1ingendosomecompartmentoferythroidprecursorsthatcan
berecruitedinvivobytreatmentwithphenylhydrazineorerythropoietin169.Nramp2
mRNAexpressionwasalsodetectedinthekidney(IsoformsIandII)175andNramp2
proteinexpression(isoform
1)wasdetectedatthebrushborderofthekidneyproximal
tubule
172,DirecttransportstudiesinXenopuslaevisoocytessuggestthatNramp2isa
pH-dependentdivalentmetaltransporterwithbroadsubstratespecificityinc1udingFe
2+,
Mn2+,C02+,Cd2+,Cu2+,Ni2+,Pb2+,andpossiblyZn2+andthatmayfunctionbyaH+co­
transportmechanism
40.Parallelstudieswithmetal-sensitivedyeshaveshownsimilar
transportpropertiesforNramp2expressedattheplasmamembrane
26;168,An
independentlyarisingnaturalmutationinTM4ofNramp2(G185R)isresponsiblefor
microcyticanemia
ofthemkmouse136andtheBelgraderat141,bothassociatedwitha
severedefectinintestinalironabsorptionandimpaireduse
byerythroidcells138,
TransportexperimentsinvitrointransfectedHEK293cells,asweIlassubcellular
localizationstudiesinvivointargettissuesfrommk/mkmicehaveshownthatthe
Nramp2G185Rmutationcausesaseverelossoffunctioncharacterizedbyreducedactivity
andpossiblyimpairedmaturation/targeting
146;147.Together,thesestudiessuggestadual
roleforNramp2astheTf-independentironacquisitionsystem
oftheduodenum,andas
transportingreducedTf-ironacrosstheendosomalmembrane
251303.
TheNramp2homolog,Nrampl1;7isexpressedinthelysosomalcompartmentof
macrophagesandneutrophils18andisrecruitedtothemembraneofpathogencontaining
44

phagosomesfonnedintheseceUs35;304,whereitmayfunctionasaMn2+effluxpump52.
Anaturally-occurringmutationinpredictedTM4ofNrampl(G169D)impairsmaturation
andmembranetargeting
oftheprotein,andcausessusceptibilitytoinfectionwith
unrelatedintracellularpathogens
7;15.NramplandNramp2definealargesuper-familyof
membranetransportershighlyconservedfrombacteriatohumans1;41;305.Demonstration
ofdivalentcationtransportbydistantNramphomologsfrombacteria(MntH,Mramp),
yeast(Sm!1,Smj2,Smf3),fly(M/v),andplant(AtNramp)hashighlightedfunctional
conservationinthisfamily
1;2;297;298,Inaddition,expressionofmouseNramp2proteinin
adoublesmfl/smj2mutantcanrestoretheability
ofthismutanttogrowatalkalinepH
andonmediumcontainingmetalchelators
34;306,Likewise,expressionofhuman
NRAMP1intheflymutantma/voUacorrectsthetastediscriminationdefectofthismutant
2,inamannersimilartothatproducedbyincreasingdietaryFe2+orMn2+307.
Inthepresentstudy,wehaveusedmultiplesequencealignmentstoidentifyhighly
conservedresiduesintheNrampsuper-family.Wehavestudiedthefunctionalrole
of
conservedchargedTMdomainresiduesinsubstrateselectivityandpHregulationof
Nramp2.
45

MATERIALSANDMETHODS
Materials
Calceinacetoxymethylester(calcein-AM;500/lMstocksolutionpreparedin
DMSO)wasobtainedfromMolecularProbes(Eugene,OR).StockFe
2+(2mM)aqueous
solutions
offerrousammoniumsulfate(FAS;Sigma)werealwayspreparedfreshin
degassed,de-ionizedwater.CoClz(Sigma)waspreparedasastocksolution(2mM)in
water.Themembrane-permeableironchelatorsalicylaldehydeisocotinoylhydrazone
(Sm,25mMstocksolution)waspreparedinDMSO,andthemembrane-impermeable
ironchelatorHES-DFO(6desferroxamine/M
r50,000starchmolecule;38mMstock)was
preparedinwaterandstoredat-20°C.
smandHES-DFOweregenerousgiftsfromDr.
P.Ponka(McGillUniversity,LadyDavisInstitute,Montreal,Canada).
PlasmidsandConstructs
AfulliengthmouseNramp2(DMTl)cDNAIsoformIIlackingtheIRE(GenBank™
AccessionNumberL33415)wasmodified
bythein-frameadditionof2antigeniccMyc
epitopetagsatthecarboxyterminus
oftheprotein(N2-2Myc)34.Mutationsatspecific
aminoacidpositionswerecreated
bysite-directedmutagenesisusingarecombinantPCR
protocol308,andusingoligonucleotideprimerslistedinTable1.Themutantswerecloned
intothemammalianexpressionvectorpCB6.Forcomplementationstudies
inyeast,
mutantswereintroduced
intheyeastexpressionvectorpVT.
YeastTransformationandsmfl/smj2ComplementationAssays
TheS.cerevisiaesmf1/smf2doublemutant,isamutantinwhichtheSMF1and
SMF2geneshavebeeninsertionallyinactivated(MATaura3-521eu2-3-112gal2
SMFl::LEU2,SMF2::LEU2)
34.Thismutantcannotgrowonalkalinemediumoron
mediumcontainingmetalchelators
34;306.YeastcellsweretransformedwithpVT/Nramp2
constructs309andura+transformantsweregrownasmasspopulations,andcrude
membranefractions
310werepreparedforNramp2proteinexpressionbyimmunoblotting
usingthemouseanti-cMycmonoclonalantibody9E10(Babco,Berkeley,CA)
161.Totest
46

Table1.Primersusedformutagenesis.
PrimerNucleotidesequence(5’to3′)PositioninNramp2
pCB6Faatgggcggtaggcgtgta848-866ofpCB6
pCB6Raggccaggagaggcactgg1006-1034ofpCB6
N2
D86AFcctacctagCcccaggaaac248-267
N2R119AF
ctgctgctccagGCccttgcag346-367
N2
R146AFcaaggtcccaGCgatcatcctgtg426-449
N2
E154AFctgatggtg9Qgttggcaatca451-472
N2D161AF
cattggttctgCcatgcaggaagtc471-495
N2G184DF
ccctgtgggAcggagtcctca542-563
N2G185RF
gggtccccctgtggggcaGagtcctcatcaccatc536-570
N2
D192AFcaccatcgcagCcacttttgtg564-585
N2
E225AFgtttggatat9Qgtacattac663-684
N2H267AF
atcatgccgGCcaacatgtacct790-825
N2H267CF
atcatgccgTGcaacatgtacct790-825
N2H267RF
atcatgccgcGGaacatgtacct790-825
N2
H272AFaacatgtacctgGCttctgcctta802-825
N2H272CF
aacatgtacctgTGttctgcctta802-825
N2H272RF
atgtacctgcGttctgcctta805-825
N2H267/272AF
atcatgccgGCcaacatgtacctgGCttctgcgtta790-825
N2H267/272CF
atcatgccgTGcaacatgtacctgTGttctgcgtta790-825
N2H267/272RF
ctgtgatcatgccgcGGaacatgtacctgcGttctgc785-828
cttagtc
N2E299AFacttcttcatc9Qgtcctgcatcg883-907
N2
R416AFgtgatcctgaccGCgtctatcgc1234-1256
Findicatesforward;
R,reverse.Codonsofmutatedaminoacidsareunderlined.Uppercaseletiersin
nucleotidesequencesindicatebasesthatdifferfromtheNramp2sequence.
47

forpossiblecomplementationofthegrowthdefectsofthesmJllsmJ2mutantbyNramp2
variants,duplicatealiquots(1mL)ofura+transfonnantswereresuspendedineither
YPDmediumorinalkalineYPDmedium(pH7.9,
OD595=0.02)in96-wellplates(100
~Llwell).Growthwasmeasuredafter24hincubationat30°CusinganELISAmicroplate
reader(Bio-Radmodel450).SensitivitytoalkalinepHwasmeasuredasrelativegrowth
ofeachtransfonnant(expressedasa%)inalkalineYPD7.9comparedtogrowthofthe
sametransfonnantinnonnalYPDmedium(pH5.5-6.0).Forcomplementationstudieson
alkalineagar,cellsgrowntosaturationinnonnalYPDweredilutedtoOD
S95=0.2,0.02,
0.002,0.0002,and0.00002inYPD7.9,and20
~Lofeachdilutionwasspottedon
alkalineYPDagarplatesandgrownfor36to48hoursat30°Cbeforephotography.
CellCultureandTransfection
LR-73ChineseHamsterOvary(CHO)cellswereroutinelygrownina-minimum
essentialmedium(MEM)supplementedwith10%fetalbovineserum,50units/mL
penicillin,and50
~glmLstreptomycin(Invitrogen).AlIpCB6(neo)plasmidconstructs
weretransfectedintocellsascalciumphosphateco-precipitates,accordingtoaprocedure
wehavepreviouslydescribed
311.Clonesofstabletransfectantswereselectedinmedium
containinggeneticin(G418;770
~glmL;Invitrogen)andwerepickedafter8-13daysof
selection.
CrudeMembranePreparationfromCHOTransfectants
Cellpelletswereresuspendedin250
~LofTNEBuffer(100mMNaCI;10mM
Tris-Cl,pH7.0;10mMEDTA)containingproteaseinhibitors(1mMPMSF,1
~M
pepstatin,0.3~Maprotinin,1~Mleupeptin).Cellswerehomogenizedby20passages
througha
25gaugex5/8inneedle,followedbycentrifugation(4°C,2000xg,10
minutes)toeliminatenuc1eiandunbrokencells.Membraneswerethenpelletedfromthe
supematant
byultracentrifugation(75000rpm,TLA-I00rotor(Beckman),4°C),and
wereresuspendedinTNEcontaining30%glycerolandproteaseinhibitors.Recombinant
Nramp2proteinvariantsweredetectedusingthemousemonoclonalanti-c-Mycantibody
9ElO(1:1000,Babco)aspreviouslydescribed
26.
48

CalceinLoadingoftheCellsandDivalentMetalTransportAssay
CHONramp2transfectantswereloadedwiththemetalsensitivefluorescentdye
calcein-AM,aswehavepreviouslydescribed
26.Briefly,CHOtransfectants(lx106cells)
wereincubatedwith0.250
J.LMcalcein-AMfor10minutesat37°Cin1mLofloading
medium(a-minimumessentialmedium,1mg/mLBSA,20mMHEPES,pH7.4).The
cellswerewashedtwice
inandresuspendedin500J.LLoftransportbuffer(150mMNaCI,
20
mMMES,pH5.0-6.5).Thecellsuspensionwastransferredtoastirredthermostated
(37°C)semi-microcuvette,andfluorescencewasrecordedusinganLS-50Bfluorescence
spectrometer(PerkinElmer,Inc.;excitation,488nm;emission,517nm;excitationand
emissionbandpass,5nm;responsetime,6
s;datainterval,0.5s).Divalentmetals(20J.LM
finalconcentrationofFe2+orC02+)wereaddedtothecellsuspensionafterallowing
fluorescencetostabilizefor60
s.Forassaysusingiron,acombinationofmembrane­
permeable
(Slli)andmembrane-impermeable(HES-DFO)ironchelatorswereusedat
varioustimepoints(Fig.SA)todistinguishintracellularfromcell-associatedquenchable
fluorescence:Afterfluorescencewasallowedtostabilizefor60s,20
J.LMFe2+wasadded
tothecellsuspension(Fig.SA,
arrow1).After240s,themembrane-impermeableiron
chelatorHES-DFOwasadded(Fig.SA,
arrow2)toreleasemetal-inducedquenchingof
extracellularcell-associatedcomplexedcalcein.At300s,amembrane-permeableiron
chelator
Slliwasadded(Fig.SA,arrow3),toreleasemetal-inducedquenchingof
intracellularcalceinfluorescence.Initialrateswerecalculatedfromquenchingcurves,
andthesize
oftheintracellularlabileironpoolwasextractedfromdataobtainedwiththe
2metalchelators.
CellSurfaceProteinBiotinylation
CHONramp2transfectantswerewashedthoroughlywithcoldPBS
(supplementedwith1
mMMgCh,0.1mMCaCh)andthenwithcoIdboratebuffer(10
mMboricacid,154mMNaCI,7.2mMKCI,1.8mMCaCh,pH9.0).Cellswerelabeled
for
15minonicewith0.5mg/mLSulfo-NHS-SS-Biotin(Pierce)incoIdboratebuffer.
AfterwashingthreetimeswithcoIdquenchingbuffer(PBS,200
mMglycine),cellswere
49

scraped,collectedandresuspendedin1mLlysisbuffer(1%TritonX-100,150mM
NaCI,2mMEDTA,10mMTris-ClpH7.4,30%glycerol)plusproteaseinhibitors.
Lysateswereincubated
onicefor20minandpelleted(10min,10OOOxg,4°C).
Supematantswerecollectedandproteinlevelswerequantified
byBradfordassay.500Ilg
oftotalproteinlysatewereincubatedovemightat4°Cwith50ilLofimmobilized
streptavidinbeads(Pierce)inafinalvolume
of500ilL(withlysisbufferandprotease
inhibitors).Streptavidinbeadswerewashedthreetimeswithlysisbufferthenoncewith
PBS.Labeledcellsurfaceproteinswereelutedwith50ilL1xSDS-PAGEloadingbuffer
andseparated
bySDS-PAGE.
50

RESULTS
MutagenesisStrategy
Alignmentof28eukaryoticandprokaryoticNrampsequencesidentifyacommon
andconservedhydrophobiecore
of10TMdomains(~30%identitybetweenbacteriaand
man)
1thatcontainfourabsolutelyinvariant(D86,E154,E299,R416)and5highly
conserved
(RI19,R146,D161,D192,E225)chargedresidues(6negative,3positive)in
TMdomains(Fig.1;darkblueresidues).AnotheruniquefeatureoftheNrampfamilyis
thepresence
oftwoinvariantHistidineresidues(H267,H272)inthepredictedTM
domain6(Fig.1;redresidues).Inmice,naturallyoccurringmutationsatadjacent
residuesinTM4
ofNramplandNramp2causesusceptibilitytoinfectionandmicrocytic
anemia,respectively,andtheloss-of-functionphenotype
ofthemicrocyticanemia
Nramp2variant(GI85R)wasstudied.MutantcDNAswereexpressedinyeastS.
cerevisiae,andtestedfortheirabilitytocomplement2phenotypesofasmfllsmj2mutant:
(a)impairedgrowthonmetalchelators
306and(b)impairedgrowthatalkalinepH.34
Mutantsshowingpartialorcompleteloss-of-functionweresubsequentlyexpressedin
CHûceIls,andtestedforFe2+andCo2+transportattheplasmamembrane.26
ConservedChargedResiduesinMembraneDomains
ImmunoblottingindicatedthataIlmutantscouldbestablyexpressedas60-65kDa
immunoreactiveproteins
inyeastmembranefractions(Fig.2A,leftpanel).Theseresults
suggestthatnone
ofthemutationshadamajoreffectonproteinexpressionorstabilityin
yeast.Theability
ofeachmutanttocomplementthenullsmfl/smj2yeastmutantwas
testedinparaIlel
byplatingseriaIdilutionsofeachtransformantonYPDagarpH7.9
(Fig.2B,
leftpanel),andusingagrowthinhibitionassayinliquidYPD(pH7.9).
Routinely,yeast
WTNramp2transformantsshoweda9-foldstimulationforgrowthon
alkalinemediumovernegativecontrols(pVT,
smf-/-;Fig.3,leftpanel).TwoNramp2
mutantspreviouslyshowntoeitherabrogate(G394V)ortohavenoeffect(Q395E)on
Nramp2function
inyeast34wereusedasadditionalcontrols(Fig.3,leftpanel).
Inbothassays,mutantsR119A,R146A,D161A,E225Ashowedfull
complementation
ofthesmfl/smj2growthdefect,whilemutantsD86A,E154A,D192A,
51

314-358
Isoforml(+IRE)
Figure1.SchematicrepresentationofmouseNramp2(DMT1)isoform”(-IRE)and
isoform
1(+IRE).Thetwelvetransmembranedomainsarepredictedfromhydropathyprofiling,
calculations
ofhydrophobiemomentandothercomputer-assistedanalyses6andfromdirect
epitopemappingstudies.13Individualpredictedintraeellularandextracel/ularsegmentsare
identifiedandtheirpositionwithintheprimarysequenceisshown.Aminoacidresiduesdefining
sequencelandmarksandsignaturemotifsaredepictedindifferentcolors,includingnegatively
andpositivelychargedresidueswithinpredictedTMdomains(darkblue),conservedhistidine
residuesinTM6(red),glycineresiduesinTM4alteredinanemicmkiBelgrademutants(G185R),
andmutated(inNramp1)
inmicesusceptibletoinfections(G185D)(yellow).Aisoidentifiedare
Asn-linkedglyçosylationsignaIsintheTM7-TM8extracytoplasmicloop(black),predicted
membranetargeting/sortingmotifs(tyrosine-based,anddi-Ieucine)(green),andconsensus
transportsignaturecommontoNramporthologsandpresentinthecytoplasmicface
ofmem­
braneanchors
ofbacterialperiplasmicpermeases(orange).The2differentC-terminiofthe
proteingeneratedbyalternativemRNAsplicingcontaining
ornotaniron-responseelement
(isoforml,+IRE;isoformIl,-IRE)
inthe3’untranslatedregionareidentified,withcorresponding
numbering.Final/y,thepolarity
oftheproteinandmembranedomainswithrespecttothe
membrane(Iightblue)isindicated(in,out,lumen).

A<:u0::NNN N"-"-"-0-<:<:<:00::<: <:<:u0::<:u0::NNN<:.....................E<:0~.-~L.I"lN"-"-"-NNN"-"-"- ~ro~L.I"l00000"-"-"-....00NNNNNNNNN0-Z0J: J:J: J:J:J: J:J:J: .'. B<:u0::NNN N"-"-"-10-<:<:<:<:0::<:<:<:~LJ.J<:<:u0::<:u0::NNNj-.....................E<:0~.-L.I"lNL.I"l0L.I"l"-"-"-NNN"-"-"- ~....ro.-~L.I"l000N0000.-"-"-"-E....00.-.- .-.-.-.-NNMM~NNN NN NN NN0-VIZ00::0::LJ.J0.:)0LJ.JLJ.J.:).:)J:J: J:J:J:J:J:J:J: Figure 2.Nramp2proteinexpressioninsmf1/smf2mutantyeastandfunctionalcomple­ mentation ofgrowthatalkalinepH.(A).Crudemembranefractionsfromsmf1/smf2yeastcells (pVT)expressingeitherwildtype(Nramp2) orindividualmutantvariantsofNramp2(indicatedon top)wereseparated bySDS-polyacrylamidegelelectrophoresis.Immunoblottingwasperformed usinganaffinitypurifiedrabbitanti-mousepolyclonalanti-Nramp2antibody.Apparentelectropho­ reticmobility oftheimmunoreactivespeciesisconsistentwithamolecularmassof60-65kDa. (8).Functionalcomplementation ofthegrowthdefectofthesmf1/smf2mutantwastestedonsolid YPDagaradjustedatalkalinepH(pH7.9).Seriaiten-folddilutions ofculturescorrespondingto individualNramp2mutants(identified)werespotted(fromtoptobottom)onYPD-agarplates(pH 7.9),followedbyincubationfor48hat30?Candphotography. 140%.------------------------------,------------------~ 120% .!:100% ~-80%(9 ~60%+='~ &40% Figure3.QuantitationofcomplementationoftheSmf1/Smf2yeastmutantby Nramp2variants.IndividualNramp2mutants(identified),togetherwithpositive (Nramp2)andnegative(pVT)controlsweregrownto s~turationanddilutedto00595= 0.02ineitherYPOorYPOpH7.9.CeUswereseededin96-wellplatesandgrowthw~s measuredafter24hincubationat30CbymeasuringOD595.Resultsrepresentrelative growth(expressedasapercentage) ofindividualmutantgrownatpH7.9vs.normal YPO.Errer'barsrepresentstandarddeviations ofthemeans. E299A,R416Afailedtodoso.Inaddition,testingtheseIlmutantsfortheirabilityto complementsusceptibility ofsmfl/smj2mutanttometalchelator(growthinEGTA­ containingliquidmedium)showedrelativeactivitiessimilartothatseenforgrowthat alkalinepH(datanotshown).Theseresultsidentify5 ofthe9chargedresiduesinTM domainsasessentialforNramp2functionand sm/1/smj2complementationinyeast. TheD86A,E154A,D161A,D192A,E299A,R416Amutantswereanalyzedfor metaltransportaftertransfectionin CHûcells(Fig.4A,leftpanel).Transfectantswere screenedbyimmunoblottingforexpression ofthecorrespondingNramp2variantsatthe cellmembrane.MembranesfromcellstransfectedwiththepCB6emptyvectorwereused asnegativecontrols,whiletwopreviouslycharacterizedtransfectants 26expressingeither low(N2GG) orhigh(N2.3)levelsofNramp2proteinwereusedaspositivecontrols. CHûcellclonesstablyexpressingmutantsEl54AandR416A,couldnotbeobtainedin4 independenttransfections(141clonesscreened)forE154Aand3independent transfections(75clonesscreened)forR416A,suggestingpossibleeffect ofthese mutationsonproteinfoldinglprocessingand/ortoxicityforthecells.StableCHû transfectantsexpressingD86A,D192AandE299Acouldbereadilyisolated.Inthese clonesthelevel ofexpressionvariedbutwasintherangeoflevelsseeninpositive controlsexpressingWT(N2.3,N2GG)orD161Aproteins(Fig.4A, leftpanel).The variabilityinexpressioncanbeattributedtoacombination ofthesiteofintegrationofthe vectorandtheeffectoftheintroducedmutation.Inaddition,biotinlabelingexperiments inintactcellsidentifiedcellsurfacereactivity ofallexpressedmutants(D86A,D161A, D192A,E299A),withlevelsapproximatelyproportionatetototalexpressionlevels detectedinmembranefractions byimmunoblotting(Fig.4B,leftpanel).Theseresults indicatethatmutationsatD86,D161,D192,andE299donothaveamajoreffecton Nramp2proteinmaturation ormembranetargeting. Transportproperties ofthemutantswereinvestigatedinintactcellsusinga calceinquenchingassay 26.MetaltransportbyNramp2usingthecalceinquenchingassay andtransportofisotopic 55Fehas beenshowntobecomparableintheidenticalcontrol Nramp2 CHûtransfectants26aswellassorneNramp2mutantvariants(datanotshown). 55 A(9«0:::«‘.j:iro20Qjc::::
0
….……..
….0I..!:)0
Figure5.RelativeironandcobalttransportactivityofNramp2mutationsaffectingconserved
chargedresiduesinTMdomains.(A)ControlCHOcells(pCB6),asweilasCHOtransfectantsexpressing
wildtypeNramp2(N2.3)wereloadedwiththemetalsensitivefluorescentdyecalcein.Cellswerewashed,
resuspendedtransportbuffer(pH6.0),andfluorescencewasrecorded
withafluorescencespectrometer.
Whenfluorescencestabilized,divalentmetals(20
/lMfinalconcentrationofFe2+orCo2+)wereaddedto
thecellsuspension(arrow1),andfluorescencewascontinuouslymonitoredforanadditional3minutes.
Forirontransportstudies(rightpanel),
themembrane-impermeableironchelatorHES-DFO(200/lM)was
addedat4
min(arrow2)torevealcellassociatedextracellularFe-calceincomplexes.At5min,the
membrane-permeantironchelatorSIH(250/lM)wasadded(arrow3),andrevealedintracellularFe-calcein
complexes(seeMaterialsandMethods).Thetransportactivity
ofNramp2wasmeasuredasaninitialrate
(siope)duringtheearlyportionofthecalceinquenchingcurve,andisshownasashadedarea(siope
interval).
(B).Relativetransportactivitiesofwildtype(N2.3,N2GG)andmutantNramp2variants
(identified)werecalculatedfrom
theinitialslopeofcalceinquenchingcurvesasshownin(A),andare
expressed
asarelativetransportactivitywithtransportinthepositivecontrolN2.3setat100%.Errorbars
representstandarderrorson
themeansof3ormoreindependentexperiments.Asterisks(*)marktrans­
portactivitiessignificantlydifferent(Le.morethan2standarddeviations)fromthosedetectedinthe
negativecontrol,vector-transfected(pCB6)cells.

clearlysuggestthattheG185Rmutationonlyattenuatesbutdoesnoteliminatetransport
function
ofNramp2proteininmammaliancells.
ConservedHistidinePairinTM6.
Histidinepairscanformmetalbindingsitesinsoluble312orinmembraneproteins
313;314.ThehighlyconservedHistidinepairfoundinTM6(H267,H272)oftheNramp
super-familywasinitiallystudiedinNramp2
bymutagenesistothesmallneutralresidue
AlanineinmutantsH267A,H272AandinthedoublemutantH267A/H272A.Metal
binding
byHisresiduesinvolvesdonatingimidazolenitrogenlone-pairelectronstothe
unfilledorbitaIs
ofthemetal,andCysteinescansubstituteforHisinthisprocess312;315-
317.Thus,mutantsH267C,H272CandH267CIH272Cwerecreated.Theimidazole
proton
ofHishasapKaof6.6andcanmediatepH-dependenteffectsinproteins.Thus,a
lastset
ofmutantswerecreatedinwhichHiswasreplacedbythepositivelychargedArg
(pKa~12)whichmayfunctionallymimmictheprotonatedHis,albeitwithamuchlarger
bulk.Mutantsweretransformedinyeastcells,andimmunoblottinganalysesshowthatall
9singleanddoublemutantscouldbestablyexpressedinthemembranefraction
ofthe
smfl/smf2mutant(Fig.2A,rightpanel).Complementationstudiesforgrowthatalkaline
pH(Fig.3)andinthepresenceofmetalchelators(datanotshown)indicatethatH267is
highlymutation-sensitivewithreplacementstoAla,CysandArgcausingsevereor
completeloss-of-function.H272waslessmutationsensitivethanH267,withH272Aand
H272CretainingnearWTactivity,andonlysubstitutiontothebulkierArg(H272R)
abrogatedcomplementation.Finally,thethreedoublemutants(H267A/H272A;
H267C1H272C;H267R1H272R)werecompletelyinactiveinyeast.
Immunoblottinganalysis(Fig.4A,
rightpanel)revealedthatallmutantscouldbe
stablyexpressedaftertransfectionin
CHûcells,withtheexceptionofthedouble
H2671H272Cmutant(3independenttransfections,119clonesscreened).Theother8
Nramp2mutantswereexpressedatvaryinglevels,butthatgenerallyfellbetweenthose
seeninthepositivecontrolsexpressinglow(N2GG)andhigh(N2.3)amounts
of
Nramp2.Biotinlabelingexperimentsinintactcellsidentifiedcellsurfacereactivityofall
expressedmutants,withlevelsapproximatelyproportionatetototalexpressionlevels
59

detectedinmembranefractionsbyimmunoblotting(Fig.4B,rightpanel).However,
sornereducedtrafficking
ofproteintotheplasmamembranemayhaveoccurredin
H267CandH272A.Metaltransportactivity
ofthemutantswastestedinthecalcein
quenchingassayforFe
2+andC02+atpH6.0.Mutantsbearingeithersinglemutationsto
Argateitherposition(H267R,H272R),
ordoublemutations(H267NH272A,
H267R1H272R)werecompletelyinactivefortransport(Fig.6).Likewise,mutantH272A
wasalsoinactive
intheseconditions,andmutantH267Conlyretainedlowtransport
activitybutforC0
2+only(Fig.6).InfactonlyH267AandH272Cretainedmodestbut
significanttransportactivitytowardsbothmetalcationsanalyzed
intheassay,with
H267Apossiblyhavingahigherselectivityforcobalt(seebelow).Theseresultsindicate
thatH267andH272playanimportantroleinNramp2metaltransport,bothinyeastand
mammaliancells.
Theapparentchangeissubstrateselectivity(cobaltoveriron)
intheH267A
mutantwasinvestigatedfurtherthroughradioisotopic
55Fetransportmeasurements.
Competitionexperimentswithcoldironandcobaltfailedtorevealasignificant
preferenceforcobaltinH267Acomparedtothewild-typeNramp2.Therefore,thechange
insubstrateselectivityinH267Aobservedinthecalceinquenchingassay,while
statisticallysignificant,couldnotbeconfirmed
byradioactivityandremainsunclear.
Possible
pHeffectsontransportactivityofthemutantswereinvestigated(Fig.7).
Fe
2+(Fig.7A,toprow)andC02+(datanotshown)transportbyNramp2inclonesN2.3
andN2GGapproachedmaximumat
pH~6.0,anddecreasingthepHto5.0hadlittle
effectontransport.Likewise,backgrounduptake
ofmetalinnegativecontrolswasnot
affected
byloweringpH.However,thelowleveloftransportactivitydetectedatpH6.0
inmutantsH267A,H267C,andH272C(Fig.6)couldbesignificantlyenhanced
by
progressivelyloweringpH(Fig.7A,7B).Thisincreaseoccurredgraduallyover
background,pH-insensitivelevelsdetectedincontrolpCB6cells.Surprisingly,mutants
272Aand267A/272Awhichshowedcompleteloss-of-functionat
pH6.0(Fig.6)could
berescued
byloweringthepH.Interestinglyhowever,loweringthepHofthe
extracellularmediumhadnoeffectonC02+(notshown)orFe2+transport(Fig.7B)ofany
ofthehistidinetoargininemutants(H267R,H272R,H267/272R).ThepHeffectseenin
60

theCys/AlamutantswasspecifiefortheconservedHispairofTM6andwasnot
observedinany
oftheotherinactivatingmutationsstudied(e.g.D86A,E299Aor
G185R)(datanotshown).Therefore,mutationsatthetwoconservedHisinTMD6
clearlyaffect
pHsensitivityofNramp2transport,andsuggestthattheseresiduesmaybe
involved
inpHregulationoftransport.
61

120
—:::R0……..**>-.100+-‘.:;:
”;::;80uro+-‘~0600.*VIc:ro40~+-‘CI)
.~20+-‘roCI)a:0coU0.
(V)19N19ZNZ
‘*
«ua:«u……………………NN……..
……..NNN
NN::r:
::r:::r:::r:::r:
a:N……..N::r:
•Cobalt
Oiron
«a:NN…………….NN…………………………..N
N::r:::r:
Figure
6.RelativeironandcobalttransportactivityofNramp2mutationsaffectingconserved
histidineresiduesinTM6.
TransportstudieswereconductedatpH6.5.Resultsarepresentedfor
different
mutantsatH267andH272,asdescribedinthelegendtoFigure5.

A
1.6)
.
1—.
B·1.6
Figure7.
Effect
of
pH
on
transport
activity
of
Nramp2
mutants
at
conserved
histi-
0.8
~
m~
~
dine
residues
in
TM6.(A)Iron
0.6
transportinthevariousnegative
0.4
(pCB6),andpositivecontrols
0.2
pCB6
WT(low)
0.2
(N2.3,N2GG)asweilasin
0
0
CHOclonesexpressing
0
1
2
3
4
5
60
1
2
3
4
5
60
1
2
3
4
56
0
1
2
3
4
5
6
Nramp2mutants
at
theplasma
1.6
1.6
1.41.4
membrane(H267A,H272A,
1.2
1.2
H267/272A,H267C,H272C,
1.0
1.0
H267R,H272R,H267/272R)
~
0.8
~
0.8
wasstudiedbythecalcein
c
0.6
c
quenchingassayasdescribed
Q)
Q)
0.6
inthelegendtoFigure5A,with
u
0.4
~
0.4
11’1
thefollowingmodifications.
~
02
H267A
H272AH267/272A
~
0.2
o
0
0
0
Afterloadingwithcalcein-AMat
::J
0
,
2
3
4
560
12
3
4
5
60
,
2
34
5
6
::J
0
1
2
3
4
5
6
pH
7.4,
ceUs
werewashedand
ü:
ü:
Q)
1.6
Q)
.1.6
resuspendedintransportbuffer
>
.~
1.4
at
differentpH(indicated).
”;::::;
pH
5.0
……
ra
ra
Fluorescencewasallowedto
v
v
stabilizefor1min,followedby
0:
pH
5.5
0:
0.8
addition
of
20
JlM
Fe2+.(B)
0.6
pH
6.0
Summary
of
transportresults
0.4
pH
6.5
0.4
obtainedforthehistidine
0.2
H267C
H272C’
0.2
mutants
at
pH5.0.
0
001
2
3
4
5601
2
34
56
0
1
2
3
4
5
6
1.6
l’L
1.4
-pCB6
1.4
-H267R
1.2
-H272R
1.0
-H267/272R
0.8
0.8
0.60.6
0.40.4
0.2
H267RH272RH267/272R
0.2
00
012
3
4
5
60
,
2
34
56012
34
5
6
0
1
2
3
4
56
Time
(min)
Time(min)

DISCUSSION
WeaimedtostudythefunctionalroleofhighlyconservedchargedresiduesintheTM
domains
ofNrampproteinsonsubstratetransportandpHregulation.Suchresidueswere
mutatedinthebackbone
ofNramp2andcorrespondingmutantsweretestedinyeastand
mammaliancens.Withfewexceptions,therewasgoodgeneralagreementbetween
sm/l/smj2complementationresultsinyeastandtransportdatainmammaliancens.
Interestingly,although
anmutantscouldbeexpressedinthemembranefractionsofyeast
cens(Fig.2A),severalcouldnotbeexpressedinCHOcellmembranes.Theseresults
suggestthatsuch
TMmutations1)affectednormalproteinfolding,maturationand
processing,possiblyleadingtoproteininstabilityand/ordegradationor2)that
overexpression
ofthecorrespondingmutantproteinwassomehowtoxicforthecells.
Theseresultshighlightpossibledifferencesinproteinsorting,maturationandtargeting
mechanisms
inbothcellsystems,butalsostresstheimportanceofmaturationand
membranetargetingforproperNramp2function.
Gly185isinvariantinallNramporthologsfrombacteriatoman.Strikingly,its
mutationtoArgaroseindependentlyin2rodentmodels
ofirondeficiency,themicrocytic
anemiamouse
mk136,andtheanemicBelgrade(h)rat141.TheG185Rmutationshoweda
complexphenotype.Inyeast,G185Rcouldbeexpressedinyeastmembranes,butcould
notcomplementthe
smfl/smj2mutant,suggestingalossoftransportactivity,in
agreementwithresultsfromSuetal
146,andWorthingtonetal318whoreporteda95%
and85%reductionin
55Fe2+transportactivityofG185RinHEK293censandCOS-7
cells,respectively.G185Rcouldalsobeexpressed
inthemembranefractionofCHO
cells(Fig.4),whereitretainedsignificanttransportactivityforbothC0
2+(50%ofWT)
andFe2+(35%ofWT).SimilartransportactivityofG185Rhasalsobeenobservedin
stabletransfectantsintheLLC-PKIporcinekidneycells(datanotshown).This
significanttransportactivitysuggeststhatloss
oftransportfunctionmaynotbethesole
defectresponsibleforthe
mkphenotypeinvivo.Rather,themutationmayaffect
membranetargetinginacell-specificfashion,perhapsinc1udingtargetingtoatransport
incompetentcompartment
inyeastcells.Recentstudiesinvivosupportsuchatissueor
cell-specificeffect.Indeed,immunoblottingstudiesshowrobustexpression
oftheG185R
64

isoforminduodenurnmembranefractionsofmk/mkmice.butimmunohistochemistry
studiesrevealedabsenceofproteintargetingtothebrushborder,thesiteofactive
transport
147.Likewise,mklmkmiceshowastrongreductionofNramp2(GI85R)protein
expressioninthekidney
172,whilemklmkreticulocytesarecompletelydevoidofNramp2
expression169.ParallelanalysisofthediseasesusceptibilityG169DmutationinTM4of
Nrampl(reconstructedinNramp2;datanotshown)associatedwithsusceptibilityto
infectionsalsoindicatedabsence
ofsmfl/smj2complementationinyeast.AIso,wecould
notisolate
CHûclonesstablyexpressingthismutant,asituationsimilartothatseenin
vivoinmacrophagesfromNramplG169Dmousestrains,wherenomatureproteinis
detected
13.Together,theseresultssuggestanimportantroleforthisresidueandTM
segmentfortransportactivity,butalsosuggestthatitsintegrityisrequiredforproper
maturation,foldingand/ortargeting
oftheNrampproteins.
Helicalwheelprojectionsintheconservedhydrophobiccore
ofNrampproteinsreveal
strongamphipathiccharacterforseveralTMdomains,inc1uding
TM3,5and91
.
Sequenceconservationexpressedasavariabilitymomentl,indicatesperiodicitywith
strongconservation
ofthepolarface,withtheapolar”lipid-accessible”faceofthehelix
beingheavilysubstituted.Severalhighlychargedresiduesmaptothepolarside
ofTM
helices.Suchanarrangementischaracteristic
offamilesofiontransportersandion
channels
1.Mutationsatthe9conservedchargesinTMdomainshadeithernoeffect
(Rl19,R146,D161,E225),orcausedpartial(D192)orcompleteloss-of-function(D86,
E154,E299,R416).D86,E154,E299,andR416arethemosthighlyconservedbeing
invariantinmultiplesequencealignmentsperformed(appended).
Itisstrikingthat3of
themhavenegativelychargedsidechains,raisingthepossibilitythattheymaymediate
interactionwiththepositivelychargeddivalentcationsubstrates
ofNramptransporters.
Alternatively,suchresiduesmaybeinvolvedinhydrogenbonding,saltbridgeformation
(dipole),orotherinteractions
intheformationofawater-filledporeortransportpath.
Theabsoluteconservation
oftheHistidinepairH2671H272inTM6ineukaryoticand
prokaryoticNrampsequencessuggestsanimportantrole.Inaddition,studies
byus(data
notshown)andothers
318showthatNramp2-mediatedtransportintransfectedcellsis
sensitivetotheaction
ofthehistidine-specificreagentdiethylpyrocarbonate(DEPC).
65

Here,weshowthatbothresiduesaremutation-sensitive(inparticularH267),with
independentsubstitutionsateitherorbothresiduescausingloss-of-function.Strikingly,
several
poodyactive(H267A,H267C,H272C)orcompletelyinactiveHismutants
(H272A,H272C,H267
AlH272A)atpH6.0couldberescuedbyloweringthepHofthe
transportassay(Fig.7).TheobservedpHeffectwasincremental,withmaximaltransport
attainedatorbelow
pH5.0.However,complete1yinactivehistidinetoargininemutants
(H267R,H272R,H267/272R)couldnotberescued
byloweringthepHofthetransport
assay.This
maybearesultofincreasedsterichindranceassociatedwithreplacementof
histidinetobulkyarginine.Thus,mutationsateitherHisresiduesinTM6shiftedthepH
requiredtoachievemaximaltransporttoamoreacidicvalue.
Severalexplanationscanbeputforwardtoaccountfortheuniqueeffect
ofpHon
transportproperties
oftheNramp2Hismutants.Firstly,H267andH272maybeinvolved
indirectbinding
ofthemetalsubstrateinapH-dependentfashion.MetalbindingbyHis
pairshasbeendocumentedinsolubleproteinssuchastranscriptionfactors
312,andhas
alsobeenused
inmembraneproteinswheretheyhavebeenengineeredtostudyproximity
relationshipsbetweenindividual
TMdomainsbyelectroparamagneticresonance314.
AlthoughH267andH272mayindeedformpartofabindingsiteformetalsinthe
membraneportion
ofNramp2,itappearsunlikely.Indeed,anNramp2mutantlacking
bothHistidines(H267
AlH272A)isstilltransportactiveatpH5.0(Fig.7B).Asecond
possibilityisthatH267,H272orbothparticipate
inH+movementacrossthemembrane
inaH+co-transportmechanismpossiblybyaprotonrelaysystem40.Sucharelaysystem
hasbeendescribedforthelactosepermease
ofE.coli,andinvolvesTMresidues
R3021H322/E325319.SuchaprotonrelaysystemmayexistinNramp2andmayinvolve
conservedresiduessuchas
H2671H272aswellasothernegativelyandpositivelycharged
residuesin
TMdomains.Partialorcompleteinactivationofthissystemwouldbe
predictedtohaveamajoreffectonthepHdependenceoftransport.Athirdexplanation
fortheobserved
pHeffectontransportpropertiesofsingleordoubleHismutants,isthat
H2671H272maybeimplicatedinpHregulationofthetransporterthroughgainorlossof
theimidazoleproton(pK5.5to6.5).Inthisfavoredmodel,protonationofH2671H272
wouldberequiredtomaintaintheprotein
inafunctional,transport-competent
66

conformation.Thiseffectcouldbeeithergeneralandinvolveadditionalresidues,
resultinginapH-dependentglobalconformationalchangefromaninactivestate(neutral
pH)toanactivestate(acidicpH).
Inthismodel,lossofthekeyHisresidueswouldshift
the
pHformaximaltransporttoamoreacidicvalue,requiringprotonationofother
groupsorsidechainstocreatethesameoverallconformationalchange.Altematively,
either
orbothH2671H272couldplayamorespecifieroleincreatingapH-dependent
transportpath
inthetransporter.Forexample,aninteractionwithadjacentandhighly
conservednegative1ychargedresiduesinother
TMdomainscouldbenecessarytoopen
aniontransportpath.LossofH2671H272wou
Idrequireformationofcompensatory
interactions
ofconservednegativelychargedresidueswithotherprotonatedsidechains
and/orwatermoleculesinthetransportpath.Althoughhighlyspeculative,sucha
mechanismappearstoaccountforthepH-dependence
ofaniontransportbytheBand3
transporter
320;321.
ACKNOWLEDGMENTS
TheauthorsareindebtedtoDrs.S.Grinstein(UniversityofToronto,Toronto)and
H.R.Kaback(University
ofCalifomia,LosAngeles)forhelpfuldiscussionsand
suggestionsduringthiswork,andDr.P.Ponka(McGillUniversity,Montreal)forthe
generouspreparationandgift
ofironchelators.
67

PrefacetoChapters3and4
Followingitsinitialfunctionalc10ningin1997,mountingevidencemadeit
increasinglyc1earthatNramp2playedadualroleinironhomeostasis
40;136.Wenow
knowthatNramp2isthetransporterresponsiblefornotonlytheintestinalabsorption
of
nutritionalnon-hemeironbutisalsotheproteinresponsibleforthetransferoftransferrin­
ironfromendosomestothecytosol.Before2005,much
ofwhatwasknownaboutthe
role
ofNramp2insystemicironhomeostasiscamefromstudiesinrodentmodelsofiron
deficiency.BothmkmiceandBelgraderatspossessidenticalmutationsinNramp2
(G185R)leadingtoseverehypochromicmicrocyticanemia136;141.Criticalmutationsin
human
NRAMP2werethoughttocauseasimilarphenotypeinpeople.
In2004,Priwitzerovaandcolleaguesreportedafirsthumanpatientsufferingfrom
aniron-relateddisorderlinkedtoamutationin
NRAMP2.Thepatient,ayoungfemaleof
Czechorigin,sufferedfromseverecongenitalhypochromicmicrocyticanemiaandiron
overload
4;259.Thephenotypewassimilartothatofthemk/Belgraderodentmodels
exceptfortheelevatedhepaticironstoresthatwerespecifictothehumanpatient.
Recently,asecondhumanpatientcompoundheterozygotefortwonovelmutationsin
NRAMP2wasidentified263.Thepatient,a5-year-oldmaleofItalianorigin,sufferedfrom
severehypochromicmicrocyticanemiaandhepaticironoverloadsimilartothefirst
patient.
Inbothhumancases,mutationsinNRAMP2resultedinaminoacidsubstitution
mutationsatconservedresiduesintheNramp2protein.Theeffects
ofthesemutationson
theexpression,function,andsubcellulartargeting
ofNramp2wereunc1earandneededto
bestudiedinordertoexplainthephenotypes
ofthepatientsatthemolecularlevel.The
followingtwochapters
ofthisthesis(Chapters3and4)describethefunctionalproperties
andpossiblecontributiontodisease
ofthetwohumanmutationsexpressedintransfected
LLC-PKIkidneycells.
68

Chapter3:
FunctionalcharacterizationoftheE399D
DMT1/NRAMP2/SLC11A2proteinproducedby
anexon12mutationinapatientwithmicrocytic
anemiaand
itonoverload
69

ABSTRACT
DMTl(Nramp2,Slclla2)mediatesironuptakeattheintestinalbrushborderand
acrossthemembrane
ofacidifiedendosomes.Asinglepatientwithseveremicrocytic
anemiaandironoverloadwasrecentlyreportedtocarryamutationinexon12
ofDMTl
(1285G>C).Themutationhastwoeffects:itseverelyimpairssplicingcausingskipping
ofexon12,andintroducesanaminoacidpolymorphism(E399D)intheproteinencoded
bytheremainingproperlysplicedtranscriptfoundinthepatient.Thefunctional
propertiesandpossiblecontributiontodisease
oftheDMTlE399Dmutationare
unknownandhavebeenstudiedinindependentmutantsatthatposition(E399D,E399Q,
E399A)expressedinLLC-PK
1kidneycells.The3mutantsareshowntobefully
functionalwithrespecttostability,targetingandtraffickingtothemembrane,andare
transportcompetent.Thisindicatesthat
DMT1Gl285Cisnotacompletelossoffunctionbut
ratherthatamodestamount
ofactiveDMTlisproducedinthispatient.Thisactivitymay
explainthedistinguishingironoverloadseeninthispatientinaddition
tomicrocytic
anemia,thatisabsentinparallelrodentmodels
ofDMTldeficiency.
70

INTRODUCTIONStudiesinmousemodels
ofdisordersinironmetabolismhaveidentifiedsorneof
thekeytransporters(DMTlINRAMP2/SLCllA2andferroportin),enzymes(hephaestin),
andregulatorypeptides(hepcidin)thatplayacriticalroleinironhomeostasis251.
DMTlINRAMP2/SLCllA2(hereinreferredtoasDMT1)isexpressedatthebrushborder
oftheduodenumwhereitmediatesuptakeofdietaryiron161,butisalsoexpressedatthe
plasmamembraneandinrecyclingendosomes
ofperipheraltissuesandcells,including
erythroidprecursors,whereitisrequiredfortransport
oftransferrin-associatedironinto
thecytoplasm
169.DMTltransportsseveraldivalentcations(Fe2+,Mn2+,C0
2+,Zn2+)in
a
pH-dependentfashion,anddownaprotongradient
40.Amutation(GI85R)inDMTl
causesmicrocyticanemiaandirondeficiencyinthemkmouseandintheBelgraderat
136;141.ThisG185Rmutationmapsintransmembranedomain4(TM-4)ofDMTland
causesmultiplebiosyntheticandfunctionaldefects,affectingstability,maturation,
targetingandtransport
148.Recently,Priwitzerovaetalidentifiedafirsthumanpatient
homozygousforamutationin
DMTl259.Thepatient,aproductofconsanguineousunion,
presentedseverecongenitalhypochromicmicrocyticanemia,highserumironandserum
transferrinreceptor(TfR)levels,andrequiredseveralbloodtransfusionsafterbirthand
duringinfancy.Shealsoshowserythroidhyperplasiawithdefectivehemoglobinization,
anddevelopedliverhemosiderosis
bytheageof19.Erythroidprogenitorsisolatedfrom
thepatientweresmall,poorlyhemoglobinized,andgrewlessefficientlyinthepresence
oferythropoietincomparedtohealthycontrols.Erythroidgrowth,morphologyand
hemoglobinizationwererescuedinthispatient
bytreatmentwithFe-SIR,anirondelivery
strategythatby-passesthe
DMTl/TfRironpathway259.Intriguingly,anddespiteobvious
similarityinphenotypes,theliverironoverloadnotedinthispatientisnotseeninDMT1-
deficient(G185R)
mkmiceorinBelgraderats.
Thepatienthasintact
FerroportinandTransferrinReceptor(Tfri)genesbutis
homozygousforaGtoCsubstitutionatthelastnucleotide
ofexon12ofDMTl
(DMT1G1285C)4.DMT1G1285CdisruptsnormalsplicingofDMTlpre-mRNAleadingto
skipping
ofexon12,andconcomitantproductionofatruncatedDMTlproteinwhich
lackspredictedTM-8,andisunlikelytobefunctional.Thistruncated
DMTlvariant
71

(~TM-8)appearstorepresentaloss-offunctionandisunlikelytoactasadominant
negativesincethedefectinironmetabolismcaused
byDMT1Gl285Cisinheritedinafully
recessivemanner
4.Althoughskippingofexon12normallyoccursatlowlevels(~10%of
mRNA)incertaintissuesofhealthyindividuals,theDMT1Gl285Cmutationgreatiy
exaggeratesthisprocess
(~90%rnRNA)4.Conversely,asmallamountoffulliength,
properlyspliced,andexon12-containing
DMTlRNAisexpressedinthepatient.Inthis
fulliengthtranscript,however,the
DMT1Gl285CmutationcausesaGlutoAspsubstitution
atposition399(E399D)inthe4
thpredictedintracellularloopofDMTl.
TheconsequenceoftheE399Dmutationonproteinfunction,inc1udingapossible
contributiontotheironmetabolismdisorderseeninthispatient(aloneorincombination
withexon12skipping)remainunknown.Here,wehavecreatedindependentmutationsat
E399andanalyzedtheeffectonproteinstability,targetingandfunctionintransfected
LLC-PKIcells.
72

MATERIALSANDMETHODS
ConstructionandExpressionofDMTlMutants
AmouseDMTl(Nramp2,Slclla2)isoformII(non-IRE)cDNAmodifiedby
insertionofahemagglutininepitope(HA)taginpredictedTM7-TM8extracytoplasmic
loopwasusedformutagenesis
26,usingarecombinantpolymerasechainreaction
strategl8andprimers5′-gtttgtcatggacggattcctga-3′(E399D),5′-tcaggaatccctgcatgacaaac-
3′(E399Q),and5′-gtttgtcatggcgggattcctga-3′(E399A).Mutantswereintroducedin
plasmidvectorpCB6as
SacI/EcoR1fragmentsfoUowedbytransfectionandexpression
inporcineLLC-PKIkidney
ceUsaspreviouslydescribed?8DMT1transportofFe2+and
Co
2+wasmeasuredusingacalceinfluorescencequenchingassay(pH5.0,20~Mfinal
metalconcentration)
26-28.Initialratesofmetaltransportwerecalculatedfromthe
fluorescencequenchingcurves.
Immunojluorescence
LLC-PKIceUsgrownoncoverslipsweretransfectedwithGFP-Syntaxin1351,
fixedwith4%paraformaldehyde,permeabilized(5%non-fatmilk,0.2%saponinin
PBS),andincubated(16hrs,4°C)withanti-Nramp2NTpolyc1onalantibod
y
42(1:300).
Coverslipswerethenwashed,incubatedwithgoatanti-rabbit-Cy3(1:2000)for1h,and
mounted
onglassslides.
CellsurfaceQuantification
CeUsweregrowntoconfluencyin48-weUplatesandfixedin4%paraformaldehyde
(30′).Cellswerewashed(PBScontaining1
mMMgClz,0.1mMCaClz),blocked(5%nonfat
milkinPBS),andincubatedfirstwithanti-HAAb(1:500,90min)andthenwithdonkeyanti­
mouse-HRPsecondaryantibody(1:4000,1h).Fortotal
DMTI-HAexpression,cellswere
permeabilized(5%nonfatmilk,0.2%saponininPBS)priortoincubationwithanti-HAantibody.
PeroxidaseactivitywasmeasuredwithanHRPsubstrate(0.4mg/mLo-phenylenediamine
dihydrochloride),accordingtoconditionsfromthecommercialsupplier(Sigma).Background
absorbancereadingsfromnon-specifiebinding
ofsecondaryantibody,andnon-specifiebinding
ofprimaryantibodytountransfectedcellsweresubtractedforeachsample.Cellsurfacereadings
werenormalizedtototal
DMTI-HAvaluesforeachcloneexpressedasapercentage.
73

RESULTSThe
DMTIG1285CmutationcausesaGlutoAspsubstitutionatposition399
(E399D)inthe4
thpredictedintracellularloopofDMTI(FigurelA).E399formspartof
ahighlyconservedtransportsignaturethatdefinestheNrampfamily.E399isinvariantin
mouse,zebrafish
(cdy),andyeast(Sm/Ip)relativesofhumanDMTI(FigurelB),
suggestinganimportantstructuralorfunctionalrole
1.AmouseDMTI-HAisoformII
(non-IRE)cDNAtemplatewasusedtocreatemutantsE399D(from
DMTIG1285C
mutation),E399Q(lossofcharge,sidechainsizeretained),andE399A(lossofcharge,
smallersidechain).Wildtype(WT)andmutant
DMTIvariantsweretransfectedinLLC­
PKIporcinekidneycells,andcellclonesstablyexpressingindividualproteinswere
selectedforanalysis.LLC-PKIwaschosensinceisderivedfromproximaltubule,asite
ofabundantDMTIproteinexpressioninvivo,andthematuration,targetingandtransport
function
ofDMTIhavebeenextensivelycharacterizedinthesecells51;148.Immunoblot
analysis
ofwholecellextractspreparedfromeitherWTormutant-expressingcells
(Figure2A)usinganantibodydirectedagainstanepitopetagintroducedinallconstructs
identifiedrobustexpression
ofE399DandE399QsimilartoWT,whileE399Awas
expressedatalowerlevel.WTandmutant
DMTIvariantsweredetectedas2
immunoreactivespecies
ofsizes~60and~90kDa.Previousstudieshaveshownthatthe
90kDaspeciescorrespondstothefullyprocessed,”complexglycosylated”form
of
DMT1,whilethe60kDaspeciescorrespondstothefull-Iengthunprocessedor”core
glycosylated”protein
51.Studieswiththetranslationinhibitorcyloheximidefailedto
detectamajoreffect
ofanyofthe3mutationsonproteinstabilityinLLC-PKIcells(data
notshown).TheseresultssuggestthatmutationsatE399(includingE399D)donotaffect
expression,stability
ormaturationofDMT1.
PreviousstudiesfromourlaboratoryhaveshownthatDMTlmutationscanaffect
stability,maturation,targeting,andtransportactivity
28;148.Thus,weinvestigateda
possibleeffect
ofE399mutationsontheseparameters.ThetransportactivityofWTand
mutant
DMTlvariantsforFe2+andC02+wastestedinintactLLC-PKItransfectants,
usingafluorescencequenchingassay
26;28.Transport,measuredasinitialratesfrom
calceinquenchingcurves,wasverysimilarforthe3mutantsandwithintherange
ofthat
74

A
BMouseNramp2
RumanNramp2
Zebrafish
RumanNrampl
MouseNramplWA”‘.1.JJ-ln.n
Drosophila(Mvl)
Yeast(Smflp)
Salmonella(MntR)
CYTOPLASM
KWSRFAR
KWSRFAR
RWSRFAR
RWSRFAR
RWSRFAR
QWPRWCR
KLQPWQR
FHIPLWVR
Figure1.PositionandconservationofE399residueinDMT1.(A)Asche­
maticrepresentation
ofDMT1-HAshowingthepositionoftheE399Dmutation
inthefourthpredictedintracellularloopoftheprotein.(8)Multiplesequence
alignments
ofDMT1orthologsforresiduesformingthefourthpredictedintra­
cellularloop.Theposition
ofE399ishighlighted(black)alongwithresidues
definingthehighlyconservedconsensustransportmotif(grey).

measuredinindependentcellclonesexpressingWTDMTl(Figure2B).Themutations
didnothaveamajoreffect
ofeitherpreferenceforFe2+andC02+ions(Figure2B),oron
thepH-dependence
oftransport(datanotshown)28.Theseresultsdemonstratethat
residueE399isnotcriticalfortransportactivityofDMT1.Thesub-cellularlocalization
ofWTandE399D,E399QandE399AvariantsexpressedinLLC-PKIcellswasnext
analyzed
bydoubleimmunofluorescencelabelingusingananti-
DMTlantibody,andadditionalmarkersofendomembranecompartments42.Having
previouslyshownthatDMTI-HAisoform
ilisexpressedinsyntaxin13-positive
recyclingendosomes
51,weexaminedtheeffectofthemutationsontargetingtothis
compartment.AIlthreemutationsatE399showedstrongco-IocalizationwithGFP­
syntaxin
13(Figure3A;introducedbyco-transfection),andlittleifanyoverlapwiththe
lysosomalmarkerGFP-LamplortheER-markerGFP-Sec
61(notshown).These
findingssuggestthatE399isnotcriticalforpropertargeting
ofDMTltorecycling
endosomes.Finally,weusedanELISA-basedsurfacelabelingmethodtoquantifytheamount
ofWTormutantDMTltargetedtothesurfaceoftransfectedLLC-PK
Icells.
Two
independentclones
ofWTDMTI-HAanalyzedinthismannerwerefoundtoexpress
22.9±1.5%and33.1±4.4%
oftotalDMTI-HAatthecellsurface(Figure3B).Mutants
E399D(26.7±2.5%)andE399A(28.2±6.9%)showedsimilarcellsurfaceexpression
levelscomparedtotheWTprotein(Figure3B).MutantE399Q(49.2±7.0%)showed
slightlyhighercellsurfaceexpressioncomparedtoWT.Theseresultsclearlyshowthat
theE399Dmutationdoesnotaffectcellsurfacetargeting
ofDMT1.
76

A
B
ID+-‘coa:::
..c.0cID:::J
a
107
81
48.7
0.010
0.0080.006
0.004
0.0020
00«0>0>0>0>0>0>
MMMUJUJUJ
~..
“1””””1″””N~
~~a..1()…J
…J
Figure2.StableexpressionandmetaltransportactivityofE399mutants.
(A)EquivalentamountsoftotalcellextractsfromLLC-PK1cellsstablyexpressing
eitherWTDMT-HA
ormutantconstructs(identified)wereseparatedbySDS­
PAGEfollowedbyimmunoblottingwithananti-HAmonoclonalantibody.Size
of
molecularweightmarkers(inkDa)areidentifiedtotheleftoftheblot.(8)Cells
loadedwiththemetal-sensitivefluorescentdyecalcein,wereincubatedwithFe2+
orC02+inacidicbuffer(pH5.0),andtherateoffluorescencequenchingwas
measuredovertime.Theresultsareexpressedastheinitialrate
offluorescence
quenching.Errorbarsrepresentstandarderror
ofthemeansofthreeormore
independentexperiments.

A
B
~~..,”,’-.
1-
,’.:~:/:~;;;~,:_
,(,
80%
t:0(n
60%
(f)~a.xWQ)
40%
(.)-ê::lCf)Q)
20%
ü
0%
..-
C’I
~~
Figure
3.
Subcellular
and
ceU
surface
expression
of
E399
mutants
at
steady-state.
(A)LLC-PK1cellsstablyexpressing
WT
or
mutantDMT1-HAweretransientlytransfectedwithGFP-syntaxin13tolabelrecyclingendosomes.Sevenhourslater,
ceUs
were
fixed,permeabilizedandstainedwithapolyclonalanti-DMT1antibody.DMT1moleculeswerevisualizedusingasecondaryanti­
rabbitantibodycoupledtofluorescentCy3.Imageswereacquiredbyepifluorescencemicroscopy.Insetsshowmagnifications
of
the
areaboxedinthefigure.
(8)
FixedLLC-PK1cellswereincubatedwithanti-HAprimaryantibodywith
or
withoutpriordetergent
permeabilizationasdescribedinMaterialsandMethods.CellswerethenincubatedwithanHRP-coupledsecondaryanti-mouse
antibody,andtheamount
of
boundprimaryantibodypresentwasdeterminedforbothconditionsbyacolorimetriereactionusing
o-phenylenediaminedihydrochloride(OPD)followedbyspectrornetry.Theamount
of
DMT1-HAexpressedatthecellsurface(in
non-permeabilizedcells)isshownasafraction(%)
of
totalproteinexpression(inpermeabilizedcells).

DISCUSSIONTakentogether,ourresultsstronglysuggestthattheE399Dvariantcreatedbythe
DMT1Gl285Cmutationisfunctional.ThisindicatesthatDMT1G1285Cmutationisnota
complete,butratherconstitutesapartialloss-of-function,andthereforethepatientindeed
expressesasmallamount
offunctionalE399D.Thisproteinisproducedinadditiontothe
loss-of-function
~TM-8variantgeneratedbyexon12skipping.Thissmallieveiof
activityisclearlyinsufficienttomaintainminimumphysiologicalfunctioninthe
erythroidcompartment,andcausesmicrocyticanemia.Thisminimumthresholdactivity
ispresumedtobebetweenthe
~10%leveloffulliengthtranscript(E399D)seeninthe
DMT1Gl285Cpatient,andthe~50%levelsobservedinherheterozygotephenotypically
normalsibling
4.
Thesefindingsalsohelpexplaintheapparentphenotypicdifferencesbetween
rodents(microcyticanemia)andhumans(microcyticanemiaplusironoverload)with
mutations
atDMT1.ThesmallamountoffunctionalE399DproducedintheDMT1Gl285C
patientmaybesufficientforacquisitionofnutritionalironatthebrushborderandstorage
inliver(hemosiderosis),butnotsufficientenoughforefficientuseandrecyclinginthe
erythroidcompartment(microcyticanemia).ThisproposaIwouldsuggestthatthe
minimumDMT1activityrequiredforphysiological”sufficiency”intheerythroid
compartmentforhemoglobinproductionishigherthanforacquisition
ofnutritionaliron
atthebrushborderintheformofinorganicmetal(Fe
2+)orofheme-iron.Thissituation
maybeexacerbatedinrodentswherethelaboratorydie!isfree
ofhemeiron.
Finally,thecombinedfindingsthatthe
DMT1Gl285Cpatientistheonlyhuman
knowntocarryanalterationin
DMT1,butyetthatthismutationbehavesasapartialbut
notcompleteloss-of-function,indirectlysuggeststhattruenullmutationsat
DMTlmay
notbecompatiblewithlife.
ACKNOWLEDGEMENTSWewouldliketothankLaurenHamlin-DouglasandRachelBeckermanforvaluable
technicalassistance.
79

Chapter4:
AnovelR416CmutationinhumanDMTl
(SIc11a2)dispIayspIeiotropiceffectson
functionandcausesmicrocyticanemiaand
hepaticironoverload
80

ABSTRACTApatientsufferingfrommicrocyticanemiaandhepaticironoverloadwas
foundtobecompoundheterozygoteforpolymorphismsintheirontransporter
DMTl(Nramp2,SLCllA2),inc1udinga3basepairdeletion(DMTldeICIT)in
intron4thatpartiallyimpairssplicing,andanaminoacidsubstitution
(DMTlCl246T,R416C)ataconservedresidueintransmembranedomain9ofthe
protein.Thefunctionalpropertiesandpossiblecontributiontodisease
ofthe
DMTlR416Cmutationwerestudiedinindependentmutantsatthatposition
(R416C,R416A,R416K,R416E)expressedinLLC-PK
1kidneycells.Non­
conservativesubstitutionsatR416(C,A,E)causemultiplefunctional
deficienciesinc1udingdefectiveproteinprocessing,loss
oftransportactivity,
impairedcellsurfacetargetingandrecyc1ingthroughendosomes,concomitant
withretention
ofthetransporterintheendoplasmicreticulum.Conversely,a
conservativeisoelectricsubstitution(R416K)waslessvulnerable,resultingina
functionaltransporterthatwasproperlyprocessedandtargetedtothecellsurface
andtorecyc1ingendosomes.Weproposethat
DMTlC1246T(R416C)representsa
completeloss-of-functionandthataquantitativereductionin
DMTlexpressionis
thecause
ofthemicrocyticanemiaandironoverloadinthepatient.
81

INTRODUCTIONThemechanisms
bywhichnutritionalironisacquired,re-distributedandrecycled
involveanumber
ofstructurallyandfunctionallydistinctmembranetransporters
includingthe
DivalentMetalTransporter1(DMT1,alsoknownasNramp2orSLCllA2)
251.DMTlisanintegralmembranephosphoglycoproteinexpressedatthebrushborderof
theduodenum,whereitimportsdietaryironinabsorptiveepithe1ialcells161.DMTlis
alsoubiquitouslyexpressedinrecyclingendosomes
ofmanycelltypesandisabundantin
erythroidprecursors,whereitisrequiredfortransport
oftransferrin-assocÏatedironinto
thecytoplasm
169.DMTIrequiresaprotongradienttotransportFe2
+asweIlother
divaientcations(Mn2
+,C02+,Zn2+)in
apH-dependentmanner40.Muchofourknowledge
ofthefunctionofDMTIinvivocornesfromstudiesofrodentmodelsofmicrocytic
anemiaandirondeficiencysuchas
mkinmiceandBelgradeinrats,thatarebothcaused
bythesamenaturaIly-occuringmis-sensemutationinDMTI(GI85R)136;141.Asimilar
butmoreseverephenotypehasrecentlybeenreportedforamousemutantwitha
completeinactivation
ofDMTl(DMTl-I-)258.Furthermore,studiesinvolvingtissue­
specificinactivation
ofDMTlhavedemonstratedthecriticalroleofthetransporterin
intestinalironabsorptionandinerythroidironutilization,butsuggestedalesscriticalrole
ofDMTIinliverironuptake258.
hlhumans,Priwitzerovaetalfirstreportedahumanpatienthomozygotefora
mutationin
DMTI(DMT1G1285c).Theyoungfemalepatientistheproductofa
consanguineousunionwhosufferedfromseverecongenitalhypochromicmicrocytic
anemiaandironoverload
4;259.ThephenotypeissimilartothatofDMTlmousemutants
withthenotableexception
ofelevatedhepaticironstoresthatarespecifictothehuman
patient.Themutationhasadualeffect,partiallyimpairingsplicing
ofexon12ofDMTl
(1285G>C)andintroducinganaminoacidsubstitution(E399D)intheremaining
properlysplicedtranscriptfoundinthepatient.WerecentlyshowedthattheE399D
mutationdoesnotinitselfaffectexpression,function,
ortargetingoftheDMTlprotein
260-262,andthusreducedDMTIfunctioninthispatientislikelycausedbyaquantitative
reductionin
DMTlrnRNAlevelsduetoimpropersplicing.
82

Recently,someofusreportedasecondhumanpatientwhowascompound
heterozygotefortwonovelmutationsin
DMTl263.Thepatientisa5-year-oldmalewho
suffersfromseverehypochromicmicrocyticanemiaandwhohasdevelopedhepaticiron
overloadshortlyafterbirth.Hematologicalvaluesandironmetabolismindicesforthis
patientandforthepatientreported
byPriwitzerovaetalweresimilar:bothhadlow
hemoglobin,lowmeancorpuscularvolume,lowmeancorpuscularhemeoglobin,high
serumironandserumferritin,andelevatedtransferrinsaturation.However,thedegree
of
anemiawasmoresevereinthechildpatientwhorequiredbloodtransfusionsuntil
erythropoietintreatmentsufficientlyamelioratedtheanemiatoallowtransfusion
independence.Thepatientpossessedtwo
nove1mutationsinDMT1:a3basepair
deletion(deICTT)inintron4andaCtoTtransitionatnucleotide1246
(DMTlC1246T)in
exon13thatresultsinanArgtoCyssubstitutionatposition416(R416C)ofDMTl.The
CTTdeletiondisruptsnormalsplicing
ofDMTlpre-mRNAbyaffectingtheconsensus
splicingacceptorsite
ofintron4,causingapartial(30-35%)skippingofexon5263.This
truncated
DMTlvariantlacks40aminoacids,inc1udinganentiretransmembranedomain
(TM2),andappearstorepresentaloss-of-functionthatisinherited
inafullyrecessive
manner.
Incontrast,theeffectsoftheR416CmutationonDMTlexpression,targeting,
andfunctionisunknown.Previousstudies
byourgrouphavesuggestedthatR416maybe
criticalfor
DMTlfunction28.AnalaninesubstitutionmutantatR416(R416A)failedto
restoregrowthundermetallimitingconditionstoayeastmutantdeletedfortwo
ofthe
threeendogenousDMTlhomologs(smfllsmj2)despiterobustmembraneexpressionof
thismutant.TheR416AvariantmaynotbestablyexpressedintransfectedCHûcells
prec1udingitsfunctionalanalysis.
Theeffects
oftheR416CmutationonDMTlfunction,inc1udingapossible
contributiontotheironmetabolismdisorderseeninthispatientremainunknownand
werestudied.WehavecreatedindependentmutationsatR416andanalyzedtheeffecton
proteinexpression,glycosylation,targetingandfunctionintransfectedLLC-PK
1cells.
83

MATERIALSANDMETHODS
Construction,expression,andfunctionalcharacterizationofDMTlMutants
AmouseDMTlisoformII(non-IRE)cDNAbackbonemodifiedby
insertionofahemagglutininepitope(HA)taginthepredictedextracytoplasmic
loopdefined
bytheTM7-TM8interval(DMTI-HA)wasusedformutagenesis26.
MutationsatR4l6wereconstructedusingarecombinantpolymerasechain
reaction(PCR)strategy
28,utilizingprimers5’­
GTGATCCTGACATGTTCTATCGCCATC-3′(R4l6C),5’­
GATCCTGACCAAGTCTATCGCCATC-3′(R416K),and5’­
GATCCTGACCGAGTCTATCGCCATC-3′(R416E).Mutantswereintroduced
intotheplasmidvectorpCB6as
SacI/EcoR1fragments.Constructionofthe
R416AmutationinaDMT1constructbearingtwoc-Mycepitopetagsatthe
carboxylterminushasbeendescribedearlier(R416A-myc)
28.R416A-mycwas
introducedintoDMT1-HAasa
SacI/EcoR1fragment.AlImutantswere
transfectedandstablyexpressedinporcineLLC-PK
1kidneycelIsaspreviously
described
28.DMT1transportofC02+wasmeasuredusingacalceinfluorescence
quenchingassay(pH5.0,20
J..I.Mfinalmetalconcentration)262.Initialratesof
metaltransportwerecalculatedfromtheinitialslopesofthefluorescence
quenchingcurves.
Quantificationofcel!surfaceexpressionbyELISA
QuantificationoftheproportionofDMT1-HAmoleculesexpressedatthe
cellsurfacewas
byanELISAwehavepreviouslydescribed184.Briefly,cellswere
growntoconfluencyin48-wellplatesandfixedwith4%paraformaldehyde(30′).
Cellswerewashed(PBScontaining1
mMMgCh,0.1mMCaCh),blocked(5%
nonfatmilkinPBS),andincubatedfirstwithanti-HAAb(1:500,90min)and
thenwithdonkeyanti-mouse-HRPsecondaryantibody
(1:4000,1h).Fortotal
DMT1-HAexpression,cellswerepermeabilized(0.1%Triton
X-lOOinPBS,30′)
priortoblockingandincubationwithanti-HAantibody.Peroxidaseactivitywas
measuredwithanHRPsubstrate(0.4mg/mLo-phenylenediamine
dihydrochloride),accordingtoconditionsfromthecommercialsupplier(Sigma).
84

Backgroundabsorbancereadingsfromnon-specifiebindingofsecondary
antibody,andnon-specifiebinding
ofprimaryantibodytountransfectedcellswere
subtractedforeachsample.CellsurfacereadingswerenormalizedtototalDMT1-
HAvaluesforeachcloneexpressedasapercentage.
Immunofluorescence
GFPplasmidswerekindgiftsofDr.D.Williams(UniversityofToronto;
GFP-Syntaxin13)andDr.
S.High(UniversityofManchester,Manchester,United
Kingdom;GFP-Sec61).LLC-PK
1cellsgrownoncoverslipsweretransfectedwith
GFPplasmids,fixedwith4%paraformaldehyde,permeabilized(5%non-fatmilk,
0.2%saponin
inPBS),andincubated(1h,20°C)withanti-DMT1NTpolyclonal
antibody(1:300)
42.Coverslipswerethenwashed,incubatedwithgoatanti-rabbit­
Cy3
(1:2000)forlh,andmountedonglassslides.
85

A
B
MouseDMTlLN
HumanDMTlLN
Zebrafish(Chy)LNLRWS
MouseNramplLKLRW
Drosophila(Mvl)LNLQW
Yeast(Smflp)INWKLQ
CYTOPLASM
Figure1.ThepositionandconservationofR416residueinDMT1.
AschematicrepresentationofDMT1-HAshowingthepositionofthe
R416Cmutation
intheninthpredictedtransmembranedomain(TM)of
theproteinisshownin(A).AmultiplesequencealignmentofDMT1
orthologsisshown
in(8)forresiduesneighboringTM9(grey).The
position
ofR416ishighlighted(black).

RESULTS
TheCtoTtransitionatnucleotide1246ofDMTlcausesanArgtoCys
substitution(R416C)inthe9
thpredictedmembrane-spanningsegmentofDMT1(Figure
lA).AlthoughR416ispositivelychargedandthermodynamicallydisfavoredinthe
membranelipidenvironment,R416isinvariantamongmouse,zebrafish
(cdy),yeast
(smflp),andfly(mvl)relativesofhumanDMTl(FigurelB).Thissuggestsanimportant
structuralorfunctionalroleforR416inmetaltransport
byDMT11.Toexplorethe
importance
ofR4l6inDMT1function,weusedamouseDMTI-HAisoformII(non-IRE)
backbonetocreateindependentmutationsatthatposition:
R4l6C(fromDMTlc1246T
mutation),R4l6A(lossofcharge,smallersidechain),R416K(chargeretained),and
R4l6E(negativecharge).Theeffectofthesemutationsonproteinexpression,transport
activity,subceUulardistributionandrecyclingfromthemembraneintotheendosome
compartmentwereexaminedintransfectedmammalian
ceUs.
DMTl-HAbearsanexofacialhemagglutinin(HA)epitopetagthatenablesusto
label
ceUsurfaceexpressedDMT1-HAmoleculesinintactcells(FigurelA).Wehave
previouslyshownthatinsertion
ofanHAtagatthispositionpreservesexpression,
transportactivityorsubcellularlocalization
ofDMTl51.Wildtype(WT)andR4l6
mutantDMTlvariantsweretransfectedintoLLC-PKIporcinekidneyceUsandceU
clonesstablyexpressingindividualproteinswereselectedandexpandedforanalysis.
LLC-PK
1ceUswerechosensincetheyarederivedfromkidneyproximaltubule,asiteof
abundantDMTlproteinexpressioninvivo,andthematuration,targetingandtransport
function
ofDMTlhavebeenextensivelycharacterizedinthesecells51;184;184.Figure2A
showsatypicalimmunoblot
ofcellextractspreparedfromceUstransfectedwitheither
WTormutantDMTl-HA.WT
DMTlwasdetectedastwoimmunoreactivespecies,a
minorspeciesat
~60kDaandamajorspeciesat~90kDa.Previousstudieshaveshown
thattheminorspeciescorrespondstotheprecursor”coreglycosylated”proteinwhilethe
majorspeciescorrespondstothemature”complexglycosylated”protein
51.Formutants
R416C,R416AandR416E,onlyexpression
oftheprecursorproteinwasdetected(Figure
2A).Theslightlyslowerelectrophoreticmobility
oftheR4l6Aprecursorproteinreflects
thepresence
ofthetwoadditionalc-MycepitopetagsinsertedattheCterminusofthis
87

A
B
..-~()«~w0-C1CDCDCDCD40%()..-..-
..-..-….J….r….r….r….r….JIl::Il::Il::Il::
110«l30%190…..1-::E0
51Q)20%0
-ê:::l(/)
(i)10%
0.0125
()
-*0Q)0..!!!.0.0100::::>«-;0.0075…..CIlIl::.s::0.00500cQ)0.0025:::l0
0..-~0-..-1….r()Il::….J….J
Figure2.StableexpressionandmetaltransportactivityofR416mutants.
EquivalentamountsoftotalcellextractsfromLLC-PK1cellsstablyexpressingeither
WTOMT-HAorR416mutantconstructswereseparatedbySOS-PAGEfollowedby
immunoblottingwithananti-HAmonoclonalantibody.Atypicalimmunoblotisshown
in(A).Thesizeofmolecularweightmarkers(inkOa)areidentifiedtotheleftofthe
blot.Cellsloadedwithametal-sensitivefluorescentdye,wereincubatedwithC02+
inacidicbuffer(pH5.0),andtherateoffluorescencequenchingwasmeasuredover
time.
In(8),theresultsareexpressedâsinitialratesoffluorescencequenching.
Errorbarsrepresentstandarderrorofthemeans
ofthreeormoreindependent
experiments.Asterixesindicatequenchratesthataresignificantlyhigherthanfor
untransfectedLLC-PK1cells(p
<0.001).In(C),fixedLLC-PK1cellswereincubated withanti-HAprimaryantibodywith orwithoutpriordetergentpermeabilizationas described inMaterialsandMethods.Theamountofboundprimaryantibodypresent wasdeterminedforbothconditionswithanHRP-coupledsecondaryanti-mouse. antibodyandacolorimetriereactionfollowedbyspectrometry.Theamount of DMT1-HAexpressedatthecellsurface(innon-permeabilizedcells)isshownasa fraction (%)oftotalproteinexpression(inpermeabilizedcells). mutantthatareabsentintheothervariants.TheseresultssuggestthatmutationsatR416 (inc1udingR416C)affectproperprocessingandpost-translationalmodification ofDMT1, inc1udingmaturationtothefullyglycosylatedpolypeptides.Interestingly,onlyforthe mostconservativeR416Ksubstitutionatthatposition,ismaturecomplexglycosylated proteindetected,suggestingthatalargepositivelychargeresidueatposition416is requiredforproperfolding/maturation.Maturation ofR416Kisc1earlyreducedcompared toWT DMTl(Figure2A),demonstratingtheessentialroleofR416inthisprocess. Earlierstudieshaveshownthatconservedchargedresidueswithinthemembrane­ spanningsegments ofDMTlaremutationsensitive,andplayacriticalroleintransport activity,cellsurfaceexpression,andsubcellulartargeting. 28Therefore,weinvestigated thepossibleeffects ofR416mutationsontheseparameters.Wemeasuredthetransport activity ofWTandmutantDMTlvariantsinliveLLC-PK1transfectants,usingacalcein fluorescencequenchingtechniquewheretheinitialrates ofquenchingcurvesareusedas anindicator ofmetaltransportintothecell.MutantsR416C,R416A,andR416E demonstratednosignificantmetaltransportactivity,displayingtransportratessimilarto thatofuntransfectedLLC-PK 1cells(Figure2B).TheseresultssuggestthatR416isnot onlycriticalforprocessingandmaturation ofDMTlbutisalsoimportantformetal transport bythisprotein.Interestingly,mutantR416Kdisplayedmetaltransportactivity (0.002748 ±0.00024AU/sec;meanquenchrate±standarderror)thatwassignificantly higher(p <0.001)thanthatofuntransfectedLLC-PKIcells(0.001472±0.00007 AU/sec)(Figure2B).Significantactivitywasalsodetectedforanadditionalindependent transfectedcellclone ofR416K(datanotshown).However,transportactivitydetectedin R416Kexpressingcellswasclearlylowerthan WTDMT1,whichmaystemfromeither reducedtransportand/orlowerlevel ofexpressionofR416KcomparedtoWT. Nevertheless,theseresultssuggestthata DMTlproteinwithanisoelectricsubstitutionat R416retainspartialmetaltransportactivity. WeusedanELISA-basedsurfacelabelingmethodtodeterminetheamount of WTormutantDMTltargetedtothesurfaceoftransfectedLLC-PKIcells.Cells transfectedwithWT DMTI-HAexpressed30.9±3.8%oftotalDMTI-HAatthecell surface(Figure2C).Thisresultisinagreementwithcellsurface DMTlexpression 89 valuesreportedpreviouslyforisoformIIIwhichwasusedasabackbonetoconstructthe mutants 51;184;262.MutantsR416C(3.6±1.0%),R416A(6.9±1.5%),andR416E(2.5± 0.6%)allshowedseverelyreducedcellsurfaceexpression(Figure2C),suggestingthat R416iscriticalforpropersortingandtrafficking ofDMTltotheplasmamembrane. Theseresultsareconsistentwiththedefectiveprocessingandlack ofactivityobserved forthesemutants(Figure 2Aand2B).Interestingly,ahigherfractionofR416K(16.8± 2.9%)wasdetectedatthecellsurfacecomparedtotheothervariants(Figure2C), consistentwiththenotedpartialmaturation/glycosylationandlowtransportactivity of thismutant.Thisresultsuggeststhatabasic,positivelychargedresidueatposition416is requiredforcellsurfacetargeting ofDMTl. ThesubcellularlocalizationofWTDMTI-HAandtheR416variantswasnext analyzed bydoubleimmunofluorescenceusingarabbitanti-DMTlantiserumaswellas additionalmarkers ofdifferentendomembranecompartments.Co-Iocalizationstudies showthatWT DMTlisoformIIisexpressedinsyntaxin13-positive(GFP-Syntaxin13) recyc1ingendosomes(Figure3),inagreementwithourpreviousstudies 51;184;184;262. However,mutantsR416C,R416A,andR416Ewereexpressedinaperi-nuc1earregion thatshowedlittleoverlapwithGFP-Syntaxin 13(Figure3).Strikingly,R416Kwasthe onlyvarianttoshowstrongco-localizationwiththerecyclingendosomemarker(Figure 3),suggestingthatanisoelectricsubstitutionatR416doesnotaffectpropersubcellular targeting ofDMTl.WhileWTDMTI-HAandvariantR416Kshowedlittleoverlapwith theendoplasmicreticulummarkerGFP-Sec61,variantsR416C,R416A,andR416Eall displayedsignificantco-Iocalizationwiththis ERmarker(Figure4).Theseresultsimply thatalargefraction ofthevariantsR416C,R416A,andR416Earemistargetedandare retainedintheER,consistentwiththeimpairedglycosylation,lack ofactivity,and reducedsurfaceexpressionobservedforthesevariants. 90 Figure3.DoubleimmunofluorescencelabelingofDMT1-HAandtherecy­ clingendosomemarkerGFP-Syntaxin13.LLC-PK1cellsstablyexpressing WTormutantDMT1-HAweretransientlytransfectedwithGFP-syntaxin13 to labelrecyclingendosomes.Thefollowingday,cellswerefixed,permeabilized andstainedwithapolyclonalanti-DMT1antibody. DMT1moleculeswerevisual­ izedusingasecondaryanti-rabbitantibodycoupledtofluorescentCy3.Images wereacquired byepifluorescencemicroscopy.Insetsshowmagnificationsofthe areaboxed inthefigure. Figure4.DoubleimmunofluorescencelabelingofDMT1-HAandtheendo­ plasmicreticulummarkerGFP-Sec61. LLC-PK1cellsstablyexpressingWT ormutantDMT1-HAweretransientlytransfectedwithGFP-Sec 61tolabelthe endoplasmicreticulum.DMT1moleculeswerevisualizedasdescribed inthe legend ofFigure3.Insetsshowmagnificationsoftheareaboxedinthefigure. DISCUSSIONTakentogether,ourdatastronglysuggestthattheR416Cvariantcreated bythe DMTlc1246Tmutationrepresentsacompleteloss-of-functioninwhichthemutantprotein isnotproperlyprocessedandisretainedinthe ERfordegradation.Therefore,wepropose thattheabrogation ofDMTlfunctionobservedinthepatientisduetoacombinationof: a)productionofafulliengthnon-functionalproteinfromanullmutationatoneallele( R416C)andb)amutation (DMTldelcTT)thatpartiallyimpairssplicing(exon5skipping) intheremainingallele,generatinga ~TM2truncatednon-functionalvariant,withasmall butunknownamount offunctionalproteinproducedbyresidualproperlyspliced transcript.Theresultingoveralliowlevel ofDMTlactivityinthepatientisclearly insufficienttomaintainphysiologicalfunctionintheerythroidcompartment,andcauses microcyticanemia. 263TheseresultssuggestthatalowlevelofDMTlactivityis pathologicalandisinsufficienttosupportironmetabolismintheerythroidsystem,yetis sufficienttoallowsornenutritionalironuptakeinthegut,whichinthelongrunleadsto ironoverloadintheliver. Coincidently,thefirstreportedpathologicalhumanmutationin DMTl (DMT1G1285c)alsoresultsinimpropersplicingofthegene,causingskippingofexon12 andthecreationofatruncated(dTM8)non-functionalvariant4.However,the DMT1G1285Csubstitutionmutationleadingtoexon12skippingappearstohaveamore severeeffectonsplicingthanthe3basepairdeletionmutationleadingtoexon5skipping (DMTldelCTT).4;263Indeed,inthefemalepatientwhoishomozygotefortheDMT1G1285C mutation,theremainingsmallieveiofpropersplicingofexon12producesaE399D variantthatshowswildtypeactivity.Inbothpatients,diseaseappearstoresultfroma quantitativereductioninfulliengthfunctional DMTltranscriptbelowacertainminimum threshold,andthesimilaritiesinconditions ofbothpatientsseemtosupportthisidea. Thisminimumthreshold DMTlactivityappearstobebetweenthelowlevelofactivity observedinboth DMTlmutantpatientsandthe-50%activityretainedintheir phenotypicallynormalheterozygoterelatives 4;263.Finally,anewpatientwasrecently reportedwithaV114-inframedeletionintransmembrane2andaG212Vreplacementin transmembrane 5.Thispatienthasaphenotypewhichisverysimilartothepreviously 93 describedsubjects265.However,thefactthatpatientswithacompleteloss-of-functionat both DMTlalleleshavenotyetbeenidentifiedsuggeststhatevenalowlevelofDMTI activity,whileinsufficienttomaintainnormalironhomeostasis,isessentialforlife. Inadditiontoprovidingbetterinsightintotherole ofDMTlinironhomeostasis, characterization oftheR416Cvariantproducedbythepatienthasalsoshedlightintothe structureandfunction oftheDMTlprotein.Ourdatashowsthattheconservedarginine inTM9 ofDMTl(R416)iscriticalfortheproperfoldingandprocessingofthe transporter.Thefactthatnon-conservativemutationsatR416(R416C,R416A,R416E) arenotwelltoleratedandthatanisoelectricsubstitution(R416K)retainspartialfunction stronglysuggeststhatapositivechargeisrequired byDMTlatthatposition,andthatloss ofthatchargeratherthanthegainofacysteineatthatpositionisresponsiblefortheloss offunction.Onepossibilityisthatthispositivechargeinteractswithanothernegatively chargedaminoacide1sewhereintheproteintoformasaltbridgethataidsintheproper foldingortertiarystructureassembly ofDMTl.Indeed,conservedchargedresidues residingwithinthehydrophobiemembranespanningsegments ofproteinshavebeen shownto beimportantforthestructureofothermembranetransporterssuchasCFTRand thelactosepermease 322;323. Hopefully,theidentificationofmorehumanmutationsinDMTlasweIlas additionaldatafrominvitromutagenesisexperimentswillhelpusbetterunderstandthe role ofthistransporterinironhomeostasisandleadtonoveltreatmentsforiron-related disorders. 94 PrefacetoChapter5 Whereasthepreviousthreechaptersfocusedontheimportanceofconserved residuesinNramp2function,thenexttwochaptersfocus onbetterunderstandingits subcellulartargetingandtrafficking.Priortotheworkdescribed inChapters5and6,a number ofstudiesshowedthatNramp2(isoformil)isexpressedattheplasmamembrane andintransferrin-positiverecyclingendosomesatsteady-state 42.StudiesbyTouretetal usinganexofacially-taggedNramp2-HAconstructshowedthatNramp2(isoformil) moleculespresentatthecellsurfaceandinrecyc1ingendosomeswereindynamic equilibrium,withsurfacetransportersbeingcontinuouslyintemalizedviaac1athrinand dynamin-dependentprocess.ThesestudiesgavefurthersupporttotheideathatNramp2 andtransferrinreceptorarefunctionallycoupledtoeffectpH-dependentironuptake acrosstheendosomalmembrane.Examination ofthesequenceofproposedcytoplasmic domains ofNramp2isoformilrevealedthepresenceofthreeputativeintemalization motifs intheaminoandcarboxylterminalregionsofthetransporter.Therefore,we decidedtouseasite-specificmutagenesisapproachtoidentifycytoplasmicmotif(s) responsibleforintemalizationandrecyc1ing ofthetransporterfromtheplasma membrane.ThisworkisdescribedinChapter 5. 95 Chapter5: Carboxylterminusdeterminantsoftheiron transporter DMTl/SLCllA2isoformII (-IRE/lB)mediateinternalizationfromthe plasmamembraneintorecyclingendosomes 96 ABSTRACT MutationsinDMT](Nramp2,S/cl]a2)impairironmetabolismandcause microcyticanemia. DMTlisexpressedattheduodenalbrushborderwhereitcontrols uptake ofdietaryiron,andispresentattheplasmamembraneandinrecyclingendosomes ofmostceIls,whereitisnecessaryforacquisitionoftransferrin-associatediron.Thegoal ofthisstudywastoidentifysignal(s)inthecytoplasmicsegmentsofDMTlresponsible foritssubcellulartargetingandintemalizationfromtheplasmamembraneintorecycling endosomes.Weintroducedmutationsintheaminoterminus(LlNT),carboxylterminus (LlCT),asweIlasin NPAy28-31,YSCF62-65andYLLNT555-559motifsofaDMTlconstruct bearinganexofacialepitopetag,whichallowedlabeling ofthetransporteratthecell surfaceforkineticstudies.Mutantswerestablyexpressed inLLC-PKIkidneyceIlsand werestudiedfortransportactivity,sub-cellularlocalization,cellsurfaceandrecycling pooldistribution,andintemalizationfromtheplasmamembrane.Kineticstudiesshowed thatcarboxylterminusmutants (LlCTandLlYLLNT)hadanincreasedfractionofthe "recyclingpool"thatwasexpressedatthecellsurfaceduetoimpairedintemalization fromtheplasmamembrane.Furthercellsurfacelabelingandimmunofluorescencestudies inintactcellsshowedthatthe LlYLLNTandLlCTmutantsweretargetedtothelysosomal compartmentuponintemalization.Theseresultssuggestthatthemajorsignalfor intemalizationandrecycling ofDMTIisoformII(-IRE/lB)residesinitscarboxyl terminusandthatremoval ofthissignalleadstoadefaultlysosomaltargeting. 97 INTRODUCTION TheDivalentMetalTransporter1(DMT1,alsoknownasNramp2,DeT1, SIcIla2)playsadualroleinintestinalironacquisitionandinironuptakebyperipheral tissues. DMTlispartofalargefamilyofhighlyconservedmetaltransporters.Itisan integralmembranephosphoglycoproteinconsisting oftwelveputativetrans-membrane (TM)domains,withthepredictedaminoandcarboxylterminipositioned ontheintra­ cytoplasmicside ofthemembrane.StudiesinXenopusoocytesandtransfected mammalian celllineshavedemonstratedthatDMTltransportsanumberofdivalent metals(Fe2 +,Mn2+,C02+,Cu2+,Cd2+,Ni 2+,Pb2+,Zn2+)in apH-dependentfashionbya protonco-transportmechanism 26;27;40.DMTlisexpressedatthebrushborderofthe absorptiveepithelium ofduodenalvilli161,whereitisresponsiblefortheuptakeofnon­ hemedietaryironfollowingreductionfromitsferric(Fe3+)to ferrous(Fe2+)statebya ferricreductase. Innon-epithelialcells,ironabsorptionoccursthroughinternalizationof transferrin-boundironviathetransferrinreceptorpathway.DMTlisalsoexpressedatthe plasmamembraneandintransferrin-positiverecyclingendosomes ofmostcellsand tissues 40;42.Itcatalyzestheeffluxofironfromacidifiedendosomes,acrossthe endosomalmembraneintothecytosol 324.VacuolarATPase-mediatedendosomal acidificationsimultaneouslyfacilitatesrelease ofironfromtransferrinandprovidesthe pHgradientrequiredforDMT1function51.Naturally-occurringmutationsinDMTl causesevereirondeficiencyandmicrocyticanemiainmkmiceandBelgraderats136;141. Bothmutantsshowthesamealteration(G185R)inpredictedTM4ofDMT1,which abrogatesbothintestinalironabsorptionandendosomaltransportoftransferrin-iron 137;138;146;155;325.Finally,DMT1isalsoexpressedatthebrushborderofepithelialcellsof proximaltubulesofthekidney172,whereitmayfunctionasare-uptakesystemfor divalentmetals. The DMTlgeneproducesatleasttwodistinctmRNAsbyalternativesplicingof two3'exonsencodingdifferent3'untranslatedregions(UTR)andproteinproductswith distinctC-termini 153;154.DMTlisoforml(+IRE)containsaniron-responsiveelement (IRE)inits3'UTR.DMTlisoformlisexpressedatthebrushborderofduodenal enterocyteswhereitsexpressionisinducedbydietaryirondepravation161;253.DMTl 98 isoform1isalsoexpressedinthekidney161j172.DMTIisoformII(-IRE)lacksanIREand encodesaproteinwhichhasadifferentC-terminal25aminoacidsegment. DMTI isoformIIisexpressedpreferentiallyinnon-epithelialcells,andisveryabundantin reticulocytes 169.Recently,additionalisoformsofDMTImRNAshavebeenidentified basedonaltematepromoterusageat DMTlexon1(exonlAvs.lB)154.Thisaltemate promoterusageispredictedtoproducea DMTIproteinbearinganadditional29amino acids(exon lA)upstreamofthepreviouslyidentifiedstartcodonofDMTIisoforms1 andII(exonIB).Therole oftheseadditionalresiduesintheexpression,function,and targeting ofDMTIhasnotyetbeenstudied.Touretandcolleagueshavedemonstrated that DMTlisoformII,whenexpressedintransfectedCHOandLLC-PK1cells,ispresent attheplasmamembraneandinanacidictransferrinreceptor-positiveendomembrane vesicularcompartment.Kineticstudies ofanexofacially-taggedDMTlmoleculeshowed similaramounts ofproteininendomembranecompartment(s)andatthecellsurface.The twocomponentsareindynamicequilibrium:surfacetransportersbeingcontinuously intemalizedviaac1athrinanddynamin-dependentprocess 51.Therefore,proper physiologicalfunction ofDMTlinironmetabolisminvolvesnotonlypropertargetingto theplasmamembrane,butalsorequiresefficientintemalizationandrecyclingvia recyclingendosomes. Tyrosine-basedmotifs(NPXYand YXX,where=bulkyhydrophobie)have
beenshowntoactasintemalizationsignaIsforcIathrin-mediatedendocytosis
183.The
NPXY
motifwasinitiallydiscoveredasamutation(Y807C)inthecytoplasmicdomain
oftheLDLreceptorthatabolishesintemalizationofthereceptor326inapatientwith
familialhypercholsterolemia.Mutagenesisstudies
327showedthatNPXYisan
intemalizationsignalforsurfaceproteinssuchastheLDLreceptor-relatedprotein1
(LRPl)328,theEGFreceptor329andmegalin330.Severalproteins,inc1udingc1athrin,AP-
2,andDab2,appeartofunctionasrecognitionproteinsfor
NPXYsignals.183TheYXX
motifhasalsobeenshowntoactasaplasmamembraneintemalizationmotiffor
endocyticreceptorssuchasthetransferrinreceptorandiontransporterssuchasthecystic
fibrosistransmembraneconductanceregulator(CFTR)
183.YXXmotifshavealsobeen
implicated
intargetingmembraneproteinssuchasLAMP-IandLAMP-2tolysosomes.
99