A monoclonal antibody distinguishes between two IgM heavy chain isotypes in Atlantic salmon and brown trout: Protein characterization, 3D modeling and epitope mapping

(1)

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies are encouraged to visit:

http://www.elsevier.com/copyright

(2)

ContentslistsavailableatScienceDirect

Molecular Immunology

j our na l h o me p ag e :w w w . e l s e v i e r . c o m / l o c a t e / m o l i m m

A monoclonal antibody distinguishes between two IgM heavy chain isotypes in Atlantic salmon and brown trout: Protein characterization, 3D modeling and epitope mapping

Atif Kamil

^a

, Knut Falk

^b

, Animesh Sharma

^c

, Arnt Raae

^d

, Frode Berven

^e

, Erling Olaf Koppang

^f

, Ivar Hordvik

^a,∗

aUniversityofBergen,DepartmentofBiology,N-5020Bergen,Norway

bNorwegianVeterinaryInstitute,DepartmentofDiagnostics,N-0106Oslo,Norway

cUniversityofBergen,DepartmentofInformatics,N-5020Bergen,Norway

dUniversityofBergen,DepartmentofMolecularBiology,N-5020Bergen,Norway

eUniversityofBergen,DepartmentofBiomedicine,PROBE,N-5020Bergen,Norway

fTheNorwegianSchoolofVeterinaryScience,BasAm,N-0033Oslo,Norway

a r t i c l e i n f o

Articlehistory:

Received8March2011

Receivedinrevisedform4May2011 Accepted9May2011

Available online 31 May 2011

Keywords:

IgM Salmon Trout Salmo Teleost Tetraploidy

a b s t r a c t

Atlanticsalmon(Salmosalar)andbrowntrout(Salmotrutta)possesstwodistinctsubpopulationsofIgM whichcanbeseparatedbyanionexchangechromatography.Accordingly,therearetwoisotypic␮genes inthesespecies,relatedtoancestraltetraploidy.Inthepresentworkitwasverifiedbymassspectrometry thatIgMofpeak1(subpopulation1)haveheavychainspreviouslydesignatedas␮BtypewhereasIgM ofpeak2(subpopulation2)haveheavychainsof␮Atype.Twoadjacentcysteineresiduesarepresent neartheC-terminalpartof␮B,incontrasttoonecysteineresiduein␮A.SalmonIgMofbothpeak1and peak2containlightchainsofthetwomostcommonisotypes:IgL1andIgL3.Incontrasttosalmonand browntrout,IgMofrainbowtrout(Oncorhynchusmykiss)iselutedinasinglepeakwhensubjectedto anionexchangechromatography.Surprisingly,amonoclonalantibodyMAb4C10againstrainbowtrout IgM,reactedwith␮Ainsalmon,whereasinbrowntroutitreactedwith␮B.Itisplausibletoassume thatDNAhasbeenexchangedbetweentheparalogousAandBlociduringevolutionwhilemaintaining thetwosub-variants,withandwithouttheextracysteine.MAb4C10wasconjugatedtomagneticbeads andusedtoseparatecells,demonstratingthat␮transcriptsresidingfromcapturedcellswereprimarily ofAtypeinsalmonandBtypeinbrowntrout.Ananalysisofaminoacidsubstitutionsin␮Aand␮Bof salmonandbrowntroutindicatedthatthethirdconstantdomainisessentialforMAb4C10binding.This wassupportedby3DmodelingandwasfinallyverifiedbystudiesofMAb4C10reactivitywithaseriesof recombinant␮3constructs.

1. Introduction

IgMistheprimarysystemicantibodyinteleostﬁsh.TeleostIgM istypicallyatetramer(Actonetal.,1971),andeachmonomercon- sistsoftwoidenticalheavychainsandtwoidenticallightchains.

Theheavychain(␮)of secretedIgMconsistsofone variableIg domainandfourconstantIgdomains(␮1,␮2,␮3and␮4).The

∗Correspondingauthor.Tel.:+4755584538;fax:+4755584450.

E-mail addresses: [email protected](A. Kamil),[email protected] (K.

Falk),[email protected] (A.Sharma), [email protected] (A. Raae), [email protected](F.Berven),[email protected](E.O.Koppang), [email protected](I.Hordvik).

membraneanchoredformofIgM,i.e.,theB-cellreceptor,isone Igdomainshorterthanthesecretedformasaresultofaspecial splicingpatterninteleostswhichexcludes␮4(Rossetal.,1998).

AJ-chainhomologhasbeenrevealedinrepresentativesofall vertebrates exceptcyclostomes and bonyﬁsh(Klimovichet al., 2008).Thus, presenceofaJ-chainappearstocorrelatewiththe abilitytoformIgMpentamers;inmammals,amphibians,reptiles andcartilaginousﬁshes.

Purification ofserum IgMfromsalmonid fishisusually per- formedbyacombinationofanionexchangechromatographyand gel filtration,or by affinity chromatography employing specific antibodies against the IgM of interest (Kobayashi et al., 1982;

doi:10.1016/j.molimm.2011.05.005

(3)

1860 A.Kamiletal./MolecularImmunology48 (2011) 1859–1867

rainbowtroutIgMwasfoundtobindtoStaphylococcalprotein- A(Estevezetal.,1993).Inadditiontobeingthemajorantibody inserum,IgMhasalsobeendetectedinskinmucusandeggsof salmonidﬁsh(Hattenetal.,2001;OlsenandPress,1997).

Early studies in our laboratoryshowed that IgM of Atlantic salmon(Salmosalar)canbeseparatedintotwodistinctsubpopula- tionsbyanionexchangechromatography(Haavarsteinetal.,1988).

Accordingly,twodistincttypesofcDNAswereisolatedandshown torepresentisotypicgenesnamed␮Aand␮B(Hordviketal.,1992, 1997).Acomparativestudyof browntrout(Salmotrutta),rainbowtrout(Oncorhynchusmykiss)andArcticchar(Salvelinusalpinus) showedthatonlyIgMofbrowntroutwasseparatedintotwopeaks byanionexchangechromatography,likeIgMofsalmon(Hordvik etal.,2002).

AsinmostteleostﬁshestheIgheavychaingenecomplexin Atlanticsalmonencodesthreemaintypesofheavychains:␮,␦ and ␶, corresponding to theclasses IgM, IgDand IgT(Hordvik etal.,1992,1997,1999;Tadisoetal.,2011;Yasuikeetal.,2010).

AtlanticsalmonbelongtothefamilySalmonidae.Duetoancestral tetraploidy,membersofthisfamilyoffishesveryoftenpossesstwo similarsub-variantsofproteinsencoded byparalogousloci.The generalviewisthatsalmonidfishdescendfromatetraploidances- torand thatmembersofthis fishfamily arestillgoingthrough adiploidisation process (Allendorf and Thorgaard,1984).It has beensuggestedthatthegeneraSalmo,OncorhynchusandSalveli- nusradiated12–16millionyearsago(Anderssonetal.,1995)and thatthetetraploideventoccurredabout25–100millionyearsago (AllendorfandThorgaard,1984).Asaresultofancestraltetraploidy therearetwoIgheavychaingenecomplexes,AandB,inAtlantic salmon,encodinghighlysimilarsub-variantsofIgM,IgDandIgT (Hordvik,1998,2002;Solemetal.,2001;Tadisoetal.,2011;Yasuike etal.,2010).

LikeinAtlanticsalmon,two␮isotypesinbrowntroutweredes- ignatedasAandBtype,respectively(Hordviketal.,2002).Since IgMsubpopulations of salmon and browntrout showedhighly similarelutionproﬁlesfromanionexchangechromatographywe expectedthattheyhadsimilarpIfeatures.Somewhatunexpected, theIgMheavychainsinbrowntroutdifferedbyonly0.14pIunits (theoretically),whileinAtlanticsalmonthedifferencewas0.67.

IsoelectricfocusingofIgMfromAtlanticsalmonandbrowntrout wasinagreementwiththetheoreticalvalues(Hordviketal.,2002).

OnlyonecommonresidueischaracteristicfortheBtypeinbrown troutandAtlanticsalmon;thisisanextracysteineresiduenearthe C-terminalpartoftheheavychain(Hordviketal.,2002).Atlantic salmonpossessat leastthree isotypes ofimmunoglobulin light chains(IgL).ThemostabundanttranscriptsencodeIgL1andIgL3, respectively(SolemandJorgensen,2002).

A molecule homologous to the polymeric immunoglobulin receptor (pIgR) is present in teleost ﬁsh,and canbe boundto mucosalIgM(Fengetal.,2009;Hamuroetal.,2007;Romboutetal., 2008).CharacterizationofapIgRhomologinsalmonisinprogress (TadisoandHordvik,unpublisheddata).Inmammals,thepIgRhas afundamentalrole inthetransportofIgA(andIgM)acrossthe epithelialcelllayerintothemucus.ApartofthepIgR(secretory component)isboundtotheantibodyandprotectsitfromdegra- dationinthehostilemucosalmilieu.ApIgRhomologinrainbow troutwasfoundtobeassociatedwithpolymericIgTingutmucus, andtheconcentrationsofgutIgTweredoublethoseinserum,indi- catingthatthisIgclassisspecializedinmucosalimmunity(Zhang etal.,2010).

The aim of the present study was tocharacterize IgM sub- populationsinAtlanticsalmonandbrowntroutinmoredetail.A monoclonalantibodyMAb4C10,originallyraisedagainstrainbow troutIgM(Thuvanderetal.,1990)showedtobeusefulasitreacted exclusivelywith␮Ainsalmonandexclusivelywith␮Binbrown trout.MAb4C10hasbeenappliedforvariouspurposesbyseveral

researchgroupsandhasbeenusedinatleast55of100studies referringtoThuvanderetal.(1990).

2. Materialsandmethods 2.1. Fish

AtlanticsalmonwereobtainedfromTheIndustrialandAquatic LaboratoryattheHighTechnologyCenterinBergen.Rainbowtrout wereprovidedfromthemarineresearchstationatMatre(Institute ofMarineResearch).Browntroutwerecaughtinamountainlake nearBergen(Bergsdalen).

2.2. PuriﬁcationofIgMfromAtlanticsalmon,browntroutand rainbowtrout

IgM from serum were purified essentially as described in Haavarsteinetal.(1988).SalmonIgMwasfirstpartlypurifiedby gelfiltration(Superdex2001660).TheIgMrichlow-throughfrac- tionwasloadedontoananionexchanger(MonoQ)andIgMwas subsequentlyseparatedintotwoseparatepeaks.

2.3. MonoclonalantibodyagainstrainbowtroutIgM

MAb4C10:amouseIgG1antibodyagainstrainbowtroutIgMhas beendescribedpreviously(Thuvanderetal.,1990).Inthepresent study, supernatant wasused ifnot otherwise stated. ProteinG- puriﬁedMAb4C10wasappliedinsomeexperiments.

2.4. ImmunomagneticpuriﬁcationofsalmonIgM

IgMwaspuriﬁedfromgelﬁltratefractionsofAtlanticsalmon serum using Dynabeads^® M-450 Epoxy coated with MAb4C10 accordingtotheprovidedmanual(Invitrogen).

2.5. Precipitationandup-concentrationofproteinsamples

Proteinsampleswereprecipitatedwith3×volicecoldacetone overnightat−20^◦Candcentrifugedat15,000×gat4^◦Cfor20min topellettheproteins.Acetonewasremovedandpelletswereair- driedandre-suspendedin1×SDSsamplebuffer.Proteinsamples wereup-concentratedwithAmicon^®Ultra-1510,000MWCOcen- trifugalﬁlterdevices(Millipore).

2.6. Proteindeglycosylation

Approximately3␮gproteinwasdissolvedin15␮lofdenatu- rationsolution(5%SDSwith10%2-mercaptoethanol)andheated at 100^◦C for 5min.Aftercooling,1.5␮lof 10× PNGase F reac- tionbufferwasadded(500mMammoniumbicarbonatewith10%

NP-40). Deglycosylation was performed with 1 unit PNGase F (Sigma–Aldrich)per2␮gofproteinsampleat37^◦Covernight.

2.7. SDS-PAGE,Westernblotandimmunodetection

SDS-PAGEwasperformedaccordingtothemethoddescribed by Laemmli (1970). Proteinsamples mixed with 1× SDS load- ing bufferwereboiledfor 5minat95^◦C beforeloadingonthe polyacrylamide gel (4% stacking gel and 12.5% separating gel).

The gel electrophoresis was carried out at 180V for approxi- mately1h. Thegelwaseither precededforCoomassie Brilliant BlueR-250(Sigma)stainingandde-staining,orWesternblotting;

100Vfor1hat4^◦C(BioRad systemandAmersham Hybond^TM- PPVDFMembrane).Afterelectro-blotting, thePVDFmembrane wasblockedat roomtemperaturefor 1hin5% drymilk in1× PBST,andincubatedovernightwith1:50dilutionofMAb4C10at

(4)

4^◦C. The nextdaythemembrane waswashed fourtimes with 1× PBST,each for 5minat roomtemperatureona rocker,and was incubated for 1h with HRP-conjugated anti-mouse IgG in 1:3000dilutionatroomtemperature.Themembranewaswashed again four times with 1× PBST each for 5min at room tem- peratureand developedusing ECLreagents asdescribedby the manufacturer(ECLPlusWesternBlotDetection,GEHealthcareLife Sciences).

2.8. Massspectrometryproteinanalysis

ThesamplesenrichedinIgMafteranionexchangechromatog- raphywereacetoneprecipitatedandsolubilisedin1×SDSloading buffer,boiledfor5minat95^◦C,andloadedandseparatedonaNu PAGE4–12%BisTrisgel.Theproteinbandscorrespondinginmass totheheavyandlightIgMchainswereexcisedfromthegel,andthe proteinsinthegelpiecewerereduced/alkylatedanddigestedby trypsinasdescribedelsewhere(http://www.uib.no/ﬁlearchive/in- gel-proteindigestion.pdf).Theresulting peptides extracted from thegelpieceweredissolvedin0.1%FA,andinjectedintoanano- HPLCsystem.ThesettingsfortheLCseparationwere:trapcolumn:

2%ACN,0.1%FAwithaﬂowrateof25␮l/min.Analyticalcolumn:

theanalyticalcolumnwasafused-silicacapillarycolumn(15cm long,75␮mi.d.)packedwithReprosil–Pur3␮mC18resin (Dr.

Maisch,Ammerbuch-Entringen,Germany).SolventAwas0.1%FA andSolventBwas90%ACN,0.1%FA.Theflowratewas0.300␮l/min withthefollowinggradient:5–10%SolventBin2min,10–40%Sol- ventBin43min,40–95%SolventBin1min,95%SolventBwas keptconstantfor5min,95–5%SolventBin3min,andregener- ationofthecolumn for21min.Thenano-HPLCsystem(Dionex, Ultimate,Sunnyvale,CA,USA) wascoupledonlinetoanUltima GlobalESI-Q-TOFmassspectrometer(Waters,Wilford,MA,USA), andthepeptideswereanalyzedbythemassspectrometerduring continuouselutionfromtheanalyticalcolumn.Thescanareaforthe MSsurveyscanwasm/z300–1500withautomaticfragmentation ofthethreeionswithhighestintensity.Allthedatawasacquiredin datadependentmode.Theresultingdatawassearchedagainstthe NCBIndatabaseusingMascot.Taxonomychosenforthesearchwas Metazoa(animals),withcarbamidomethylationofcysteineasfixed modificationandoxidationofmethionineasvariablemodification.

2.9. Isolationoflymphocytes

The fish were killed by a sharp blow to the head. Blood wereimmediatelycollectedusingasyringewith5mlvacutainer tubes (containing heparin) and kept on ice for a maximum of three hours before further use. The amount of blood isolated from each fish varied from 2 to 8ml depending on the fish size and the success of the blood collecting. One ml of blood wasmixedwith3mlBalancedSalt Solution(SolutionA:Anhy- drousd-glucose;0.1%,CaCl₂×2H₂O;5.0×10⁻⁵M,MgCl₂×6H₂O;

9.8×10⁻⁴M,KCl; 5,4×10⁻³M,Tris;0.145M,Solution B:NaCl;

0.14M). 3ml Ficoll-Paque Plus (Amersham Biosciences) were addedto10mlcentrifugetubes.Thedilutedbloodwascarefully layeredontopof theFicoll-PaquePlus.Centrifugationwascar- riedoutinatemperaturerangebetween8and15^◦Cfor40min at400×g.Aftercentrifugation theplasmalayerwasdrawn off before collecting the band of lymphocytes. Lymphocytes from thesameﬁshbutfromdifferenttubeswerepooledandwashed twice in 3 volumes of Balanced Salt Solution with centrifuga- tion at 100×g for 20min tocollect thecells in between each wash. The cells were ﬁnally suspended in 500␮l of PBS/0.1%

BSA.

2.10. Immunomagneticseparationofcells

Dynabeads M-450 Goat anti-Mouse IgG (Dynal) were uti- lized for separation of cells using the direct technique by pre-coatingtheDynabeadswithMAb4C10.1×10⁷DynabeadsM- 450 were washed twice in PBS/0.1% BSA and re-suspended in 500␮l PBS/0.1% BSA before adding 1.5␮g of rinsed MAb4C10.

The mixture was incubated by gentle tilting and rotation for 60minatroomtemperature.ThecoatedDynabeadswerewashed 4× in PBS/0.1% BSA utilizing a magnetic particle concentra- tor and ﬁnally suspended in PBS/0.1% BSA. Cells suspended in PBS/0.1% BSA weremixed withthe pre-coated Dynabeads in a total volume of 1ml and subsequently incubated at 4^◦C on a rotor for 20min. Estimation of number of cells was done utilizing a Brinkmanncell chamber. Theratiosof Dynabeads/cells varied (between1:9 and3:1). Aftertheincubation ofthe Dyn- abeads/Cell mix the Dynabeads were washed 5× in PBS/0.1%

BSA.

2.11. Scanningelectronmicroscopy(SEM)

LeukocytesfixedwithKarnowskywerewashedin0.2Msodium phosphatebuffer,followedby1hfixationin1%aqueoussolutionof osmiumtetroxide(OsO₄).ThecellswerewashedinPBSanddehy- dratedwithcoldacetone;(1)60%acetone,(2) 90%acetone,and (3)100%acetone.Thecellswerefinallyattachedtoanobjectglass and coatedwithgold–palladium(PolaronSC502Sputter Coater, FisonInstruments).Thecellswereexaminedbyscanningelectron microscopy(ZEISSSupra55VP).

2.12. IsolationofRNAandsynthesisofcDNA

RNAwasisolatedusingTrizolReagent(LifeTechnologies,USA).

First strandcDNAwassynthesizedbyoligo-dTprimingontotal RNAwithMMLVreversetranscriptase(Promega,Madison,USA).

2.13. Polymerasechainreaction(PCR)

ChemicalsandTaqpolymeraseforPCRwerepurchasedfrom Pharmacia.Followingproﬁlewasused:94^◦C,30s,55^◦C,30s,72^◦C, 1min,35cycles,linkedto72^◦C,10min.

2.14. RelativeabundanceofAandBmRNAincellscaptured byMAb4C10/Dynabeads

cDNA descending from the cell fraction captured by the MAb4C10/DynabeadsaswellascDNAdescendingfromcellsthat werenotcapturedbytheMAb4C10/Dynabeads,wereusedastem- plate inPCR withprimers J-sense(TTTGACTACTGGGGGAAAGG) and ␮3-antisense (CCCATTGCTCCAGTCCTCAT). A characteristic EcoRI site in salmon ␮A1, which is lacking in ␮B1 cDNA, was employed to decide whether the cells captured by the MAb4C10/Dynabeads had transcripts for either ␮A or ␮B.

A characteristic Sau3A site in brown trout ␮B3, which is lacking in ␮A3, was employed to decide whether cells captured by MAb4C10/Dynabeads had transcripts for either ␮A or ␮B. PCR-products were puriﬁed by the use of QIAquick spin columns (Qiagen) before being subjected to restriction digestion. Selected PCR products were cloned into TA-vector (Invitrogen).

2.15. Constructionof3expressionplasmids

DNA-fragments encoding salmon ␮A3, salmon ␮B3, brown trout␮A3,browntrout␮B3andrainbowtrout␮3weregenerated

(5)

μ1 Consensus ASSTAPTLFPLAQCGSGTGDMMTLGCIATGFTPASLTFKWNEEGGNSLTDFVQYPAVQTGGSYMGVSQLRVKRADWDSKxFECAVEHSAGSKxVPVKKQ Salmon A ---------------Q------S----------I------T--- Salmon B ---------V---------Q------S----------I------T--L--- Trout A ---------------Q-------V---T-------I------T--- Trout B ---------------Q------T-------I------T--- Rainbow --------------D-------------K----K---

μ2 Consensus xEYLQHPSLYVMTPSKEEMAENKTASFACFANDFSPRTHTIKWMRMEKGTExEVVSDFKSSCESEKKSEKTLYSTTSYLRVNESEWKSEEVTFTCVFENKAGNVRRTVGYTSSD Salmon A A--------M----------Q-I-K---D----------A--------- Salmon B V--------------I-K----------S-----K------ Trout A V--------------I-K-------------S--------- Trout B V--------------I-K-------------S--------- Rainbow P----Q-------S---------Q------T---------------

μ3 Consensus AGPVHxHSVVIxIxPPSLEDMLMNKKAELVCDVxELVPGFMSVKWENDNGKTLTSRKGVTDKIAILDITYEDWSNGTVFYCAVDHLENLGxLVKKAYKRETG Salmon A -----A-----K-T-------E----------R---------S----P------ Salmon B -----A---N-I--------K---T--------M------------S----P------ Trout A -----V---N-I--------K----------R---------T----P------ Trout B -----A-----K-T----S------E---------A--R---------T----P------ Rainbow -----G---T-IE---Q---N----L------------M----D------

μ4 Consensus GDPQRPSVFLLAPAEQTSDNTVTLTCYVKDFYPKDVLVAWLVDDEPVERTSSSALYQFNTTSQIQSGRTYSVYSQLTFSNDLWKNEEVVYSCVVYHESMIKSTKILMRTIDRTS Salmon A -------K------E------I--------T-------K--------- Salmon B -------K------E------I--------------------- Trout A ------P---------E------I---L--------------------- Trout B ------------E------I-------------K--------- Rainbow -V-----------------------K-------N-I------ Consensus NQPNLVNLSLNVPQSCKAQ

Salmon A ---Y------- Salmon B ------C---- Trout A ---------- Trout B ------C---- Rainbow -------R-M--

Fig.1.Alignmentof␮sequencesfromAtlanticsalmon(Hordviketal.,1992),browntrout(Hordviketal.,2002)andrainbowtrout(Leeetal.,1993;Hansenetal.,1994) showingpeptidesidentiﬁedbymassspectrometry.Trypsincleavagesitesareinboldandthepeptidesthatwereidentiﬁedbymassspectrometryareindicatedwithgrey.

byreverse transcription (RT)-PCR,utilizingsense primer EcoRI- IgMs(GGAATTCACTGTGGGCTACACTTCATCA)and reverseprimer EcoRV-IgMa(GATATCATCATTTCACCTTGATGGCAGT).ThePCRfrag- mentswere cloned into TA-vector (Invitrogen) and sequenced.

Subsequently,plasmidpreparationsweredigestedwithEcoRIand EcoRVandinsertswerepuriﬁedfromagarosegelbeforebeinglig- atedintopcDNASpFLAGvectorasdescribed in(Koppangetal., 2010).

2.16. Transfectionandimmunostaining

Approximately35%confluentSH-SY5Ycellsweretransfected in DMEM media (Sigma–Aldrich) with plasmids using Lipofec- tamine 2000 (Invitrogen) or Metafectin PRO (Biontex, Planegg, Germany) transfectionreagent accordingto themanufacturer’s protocols.After6hthemediumwaschangedtoDMEMcontain- ingampicillin/streptomycinand10% serum,and wasincubated for further 48h. When the cells were approximately 80% con- fluentthey werefixed oncover slips by incubation for 20min at room temperature in 2% formaldehyde in PBS, and thereafter washed three times with PBS. For permeabilization the cells wereincubated with0.2% Triton X-100in PBS for 10min and thereafter washed. Thenthe cells were blocked for 1h at roomtemperature with 10% BSA in PBS. Cells were immunos- tainedovernightat4^◦Cusingprimaryantibodymousemonoclonal anti-Flag(1:1000)orMAb4C10(1:40)inPBSwith3%BSA.After the cells were washed, they were incubated for 1h in dark atroomtemperature,withsecondaryantibodyFITCanti-mouse (1:5000)in PBSwith3%BSA.The cellswerewashed againand thecoverslipsweremountedonanobjectglasswithadropof mounting solution ProLong^® Gold antifade with DAPI (Invitro- gen).

2.17. SequencingandanalysisofDNA

DNAsequencingwasperformedbyuseoftheBigDyeSequenc- ing kit (Amersham Life Science, Cleveland, USA). Sequences wereanalyzedwithVectorNTISuite(Informax,Inc.),CLUSTALW (www.ebi.ac.uk/services)andBLAST(www.ncbi.nih.nlm.gov).

2.18. Homologymodelingandstructuralanalysis

Primary sequence analysis was done using BLAST (www.ncbi.nih.nlm.gov)andCLUSTALW(www.ebi.ac.uk/services).

Modeller (http://www.salilab.org/modeller/) and SWISS-MODEL (http://swissmodel.expasy.org) were used to model the 3D structures. Cysteine bridge analysis was done using WHAT-IF (http://swift.cmbi.ru.nl/servers/html/listcys.html). Visualiza- tion and presentation of the models were done using Rasmol (http://openrasmol.org/)withcommand-lineinteraction.

2.19. Bioinformaticsanalysisand3Dstructureprediction

PDBID2W59chainAwithstructuralresolutionof1.75angstrom wasfoundtobeagoodtemplate(ID%∼25%)formodelingusing BLASTand SWISS-MODEL.Thistemplatesatisfiedtheadditional constraintofpossibledisulfidebondbetweenthecysteineresidues inthestudiedproteinsequence.Inthemodeledstructures,WHAT- IFanalysisshowedthepossibilityofcysteinebridges.TheAnolea andGromosscoreswereinfavorablenegativerangeformostof themodeledresidueswithfinalenergy∼1900KJ/mol.Furtherthe ModellershowedmeanDOPEscoreof∼8000andGA341∼0.5for allthemodels.ThemodelswerefurtherprocessedusingRasMol 2.7.5windows.

(6)

Peak 1 Atlantic salmon

AF273014-L1 ---IFIWIFALHLQESRGQVTVTQTPAVKTISVGDLVSLSCKTSSAVYSD-RHGQRLAWYQQKPGGAPELLIYLAKTLQSGIPSRFSGSGTGSD AF273017-L1 -MTFITIFIWTLALCLQESRGQVTVTQTPAVKSVSVGNSVSLSCKTSSAVYSD-SNGHYLHWYQQKPGGAPELLIYWAKTLQSGTPSRFSGSGSGSD AF273020-L1 -MTFIMSFVWILMSLIHESRGQVTVTQTPAVKAVLTGQTVPLNCKTSSDVYQAGTSSPRLAWYQQKPGEAPKLLIYYATTLQSGTPSRFSGSGTHSD AF273012-L1 -MTFITIFIWTLAFCFQESRGQITVTQTPTVKAVVSGQTVSLNCKTSSDVHAN----VYVAWYQQKPGGAPELLIYTATSLQSGTPFRFSGSGSGSD AF406957-L3 MMMSLTLLLGTLGLLVQESSGDIILTQSPKSQSVRPGETVSISCTASSSTYNN---LQWYLQKPGEAPKLLVYSTTNRQSGIPGRFSGSGSGSS ACI68649-L3 MMMSLILLLWTLGLLVQESSADIILTQSPKSQSVRLGETVSISCTASSSTGNN---LHWYLQKPGEAPKLLVYSTTSRQSGIPGRFSGSGSGSS ACI70011-L3 -MTFITIFIWTLVCYTQGSWGQYVLTQS-AAQSVQPGQTVSIDCSASSKVNQYSGSRYYLAWYHQTFGEAPKLLIYYTSDRFTGVSTRFSGSGRGNG AF273014-L1 ---FTLTISGVQAEDAGDYYCQSLHNPNSVWVFTFGSGTRLDVGSNSAPTLTVLPPSSEELSSTTTATLMCLANKGFPSDWTMSWKVDGNSKK---- AF273017-L1 ---FTLTISGVQAEDAGDYYCQSYHSG---YVYTFGSGTRLDVGSNSAPTLTVLPPSSEELSSTTTATLMCLANKGFPSDWTMSWKVDGTSKK---- AF273020-L1 ---FTLTISGVQAEDAGDYYCQSFHYPNSKYVFTFGSATRLDVGSNSAPTLTVLPPSSEELSSTTTTTLMCLANKGFPSDWTMSWKVDGNSKK---- AF273012-L1 ---FTLTISGVQAEDAGDYYCQSLHNPNSVWVYTFGSGTRLDVGSNSAPTLTILPPSSEELSSTTTATLTCLANKGFPSDWTMSWKVDGTSKN---- AF406957-L3 YTQFTLTISGVQAEDAGDYYCQQGYS----TPYTFGGGTRLDIGSDVRPTLTVLPPSSVELQ-QGKATLMCLANKGFPSDWKLGWKVDGSSSS-TWE ACI68649-L3 YTHYTLTISGVQAEDAGDYYCQQGNS----SPWTFGGGTKLSVGSDVRPTLTVLPPSSVELQ-QGKATLMCLANKGFPSDWKLSWKVDGSSSSNTWE ACI70011-L3 -IDFTLTISKVQAEDTGVYYCQSYHSG---TVLTFGGGTKLSVGSDVRPTMTVLPPSSVELQ-QEKATLMCLANKGFPSDWKLSWKVDGSSSS-TWE AF273014-L1 QEASPGVLEKDGLYSWSSTLTLTAQEWTKAGEVTCEAQQISQTPVTKTLRRADCSG 236

AF273017-L1 QEASPGVLEKDGLYSWSSTLTLTAQEWTKAGEVTCEAQQISQTPVTKTLRRADCSG 238 AF273020-L1 QEASPGVLEKDGLYSWSSTLTLTAQEWTKAGEVTCEAQQISQTPVTKTLRRADCSG 242 AF273012-L1 QKTSPGVLEKDRLYSWSSTLTLTGQEWTKAGEVTCEAQQNSQTPVTKTLRRADCSG 238 AF406957-L3 VTGSPGVLEKDGHYSWSSTLTFPVDQWKKVGSVVCEATQGSQSPLSETLRRDQCSD 238 ACI68649-L3 VTGSPGFQEKDGHYSWSSTLTLPVDQWKKVGSVVCEATQGSQSPLSETLRRDQCSD 239 ACI70011-L3 VTGSPGVLEKDGHYSWSSTLTLPVDQWRKVGSVTCEATQGTQTPLSETLRRDQCSD 242

Peak 2 Atlantic salmon

AF273020-L1 -MTFIMSFVWILMSLIHESRGQVTVTQTPAVKAVLTGQTVPLNCKTSSDVYQAGTSS---PRLAWYQQKPGEAPKLLIYYATTLQSGTPSRFSGSGTHSD AF273018-L1 -MTFIMSFVWILMSLIHESRGQVTVTQTPAVKAVLTGQTVPLNCKTSSDVYQAGTSS---PRLAWYQQKPGEAPKLLIYYATTLQSGTPSRFSGSGTHSD ACI66923-L1 -MTFIMSFVWILMSLIHESRGQATVTQTPAVKAVLTGQTVPLNCKTSSDVYQAGTSS---PRLAWYQQKPGEAPKLLIYYATTLQSGTPSRFSGSGTHSD AF273014-L1 ---IFIWIFALHLQESRGQVTVTQTPAVKTISVGDLVSLSCKTSSAVYSDRHG---QRLAWYQQKPGGAPELLIYLAKTLQSGIPSRFSGSGTGSD AF273017-L1 -MTFITIFIWTLALCLQESRGQVTVTQTPAVKSVSVGNSVSLSCKTSSAVYSDSNG---HYLHWYQQKPGGAPELLIYWAKTLQSGTPSRFSGSGSGSD ACI68406-L1 -MTSITIFIWTLALCFKDSRGQITVTQTPAVKAVLPGQAVYLTCKTSSSVFGDCHNGQSWGHQCLSWYQQKPGENPKLMMLPGNTLYSGTPSRFSGSGSGSD AF273012-L1 -MTFITIFIWTLAFCFQESRGQITVTQTPTVKAVVSGQTVSLNCKTSSDVHAN---VYVAWYQQKPGGAPELLIYTATSLQSGTPFRFSGSGSGSD ACI67959-L1 -MTFITIFIWTLVSCFQEARGQYVLTQTPAVKAVVPGQQVSLNCKTSSDVYND---NCLAWYQQKPGGAPKLLIYYATTLQSGTPSRFSGSGSRSD AF406957-L3 MMMSLTLLLGTLGLLVQESSGDIILTQSPKSQSVRPGETVSISCTASSSTYNN---LQWYLQKPGEAPKLLVYSTTNRQSGIPGRFSGSGSGSS ACI68649-L3 MMMSLILLLWTLGLLVQESSADIILTQSPKSQSVRLGETVSISCTASSSTGNN---LHWYLQKPGEAPKLLVYSTTSRQSGIPGRFSGSGSGSS ACI70011-L3 -MTFITIFIWTLVCYTQGSWGQYVLTQS-AAQSVQPGQTVSIDCSASSKVNQYSGSR---YYLAWYHQTFGEAPKLLIYYTSDRFTGVSTRFSGSGRGNG AF273020-L1 ---FTLTISGVQAEDAGDYYCQSFHYPNSKYVFTFGSATRLDVGSNSAPTLTVLPPSSEELSSTTTTTLMCLANKGFPSDWTMSWKVDGNSKK----QEASP AF273018-L1 ---FTLTISGVQAEDAGDYYCQSFHYPNSKYVYTFGSATRLDVGSNSAPTLTVLPPSSEELSSTTTATLMCLANKGFPSDWTMSWKVDGNSKK----QEASP ACI66923-L1 ---FTLTISGVLAEDAGDYYCQSYHYINSKNVYTFGSATRLDVGSNSAPTLTVLPPSSEELSSTTTATLMCLANKGFPSDWTMSWKVDGNSKK----QEASP AF273014-L1 ---FTLTISGVQAEDAGDYYCQSLHNPNSVWVFTFGSGTRLDVGSNSAPTLTVLPPSSEELSSTTTATLMCLANKGFPSDWTMSWKVDGNSKK----QEASP AF273017-L1 ---FTLTISGVQAEDAGDYYCQSYHSG---YVYTFGSGTRLDVGSNSAPTLTVLPPSSEELSSTTTATLMCLANKGFPSDWTMSWKVDGTSKK----QEASP ACI68406-L1 ---FTLTISGVQAEDAGDYYCQSWHSGN---VFTFGSATRLDVESNSAPTLTVLPPSSEELSSTTTATLMCLANKGFPSDWTMSWKVDGNSKK----QEASP AF273012-L1 ---FTLTISGVQAEDAGDYYCQSLHNPNSVWVYTFGSGTRLDVGSNSAPTLTILPPSSEELSSTTTATLTCLANKGFPSDWTMSWKVDGTSKN----QKTSP ACI67959-L1 ---FTLTISRVQAEDAGDYYCQSFHYPNYKYVYTFGSGTRLDVGSNSAPTLTVLPPSSEELSSTTTATLTCLANKGFPSDWTMSWKVDGTSKN----QETSP AF406957-L3 YTQFTLTISGVQAEDAGDYYCQQGYS----TPYTFGGGTRLDIGSDVRPTLTVLPPSSVELQ-QGKATLMCLANKGFPSDWKLGWKVDGSSSS-TWEVTGSP ACI68649-L3 YTHYTLTISGVQAEDAGDYYCQQGNS----SPWTFGGGTKLSVGSDVRPTLTVLPPSSVELQ-QGKATLMCLANKGFPSDWKLSWKVDGSSSSNTWEVTGSP ACI70011-L3 -IDFTLTISKVQAEDTGVYYCQSYHSG---TVLTFGGGTKLSVGSDVRPTMTVLPPSSVELQ-QEKATLMCLANKGFPSDWKLSWKVDGSSSS-TWEVTGSP

AF273020-L1 GVLEKDGLYSWSSTLTLTAQEWTKAGEVTCEAQQISQTPVTKTLRRADCSG 242 AF273018-L1 GVLEKDGLYSWSSTLTLTAQEWTKAGEVTCEAQQISQTPVTKTLRRADCSG 242 ACI66923-L1 GVLEKDGLYSWSSTLTLTAQEWTKAGEVTCEAQQISQTPVTKTLRRADCSG 242 AF273014-L1 GVLEKDGLYSWSSTLTLTAQEWTKAGEVTCEAQQISQTPVTKTLRRADCSG 236 AF273017-L1 GVLEKDGLYSWSSTLTLTAQEWTKAGEVTCEAQQISQTPVTKTLRRADCSG 238 ACI68406-L1 GVLEKDGLYSWSSTLTLTAQEWTKAGEVTCEAQQISQTPVTKTLRRADCSG 244 AF273012-L1 GVLEKDRLYSWSSTLTLTGQEWTKAGEVTCEAQQNSQTPVTKTLRRADCSG 238 ACI67959-L1 GVLEKDGLYSWSSTLTLTGQEWTKAGEVTCEAQQKSQTPVTKTLRKADCSG 238 AF406957-L3 GVLEKDGHYSWSSTLTFPVDQWKKVGSVVCEATQGSQSPLSETLRRDQCSD 238 ACI68649-L3 GFQEKDGHYSWSSTLTLPVDQWKKVGSVVCEATQGSQSPLSETLRRDQCSD 239 ACI70011-L3 GVLEKDGHYSWSSTLTLPVDQWRKVGSVTCEATQGTQTPLSETLRRDQCSD 242

Fig.2.AlignmentoflightchainsequencesidentiﬁedwithintheAtlanticsalmonIgMpopulationselutedinpeak1andpeak2,respectively.Trypsincleavagesitesareinbold andthepeptidesthatwereidentiﬁedareindicatedwithgrey.GenBankaccessionnumbersareshowninadditiontowhichisotypefamilythesequencebelongto.L1andL3 arethemostabundantlightchainisotypesinAtlanticsalmon(Solemetal.,2001).

3. Results

3.1. Massspectrometryproteinidentiﬁcationofheavyandlight chainsinIgMsamplespuriﬁedfromAtlanticsalmon,browntrout andrainbowtrout

Serawerepuriﬁedbygelﬁltrationfollowedbyanionexchange chromatography. IgM subpopulations of Atlantic salmon and browntrout wereelutedintwo distinctpeaks,whereas IgMof

rainbowtroutwaseluted inone peak.Proteinidentiﬁcationby massspectrometryconﬁrmedthatIgMofpeak1containedheavy chainsof␮BtypewhereasIgMofpeak2containedheavychains of ␮A type;inboth Atlanticsalmonandbrowntrout,asprevi- ouslyhypothesized(Hordviketal.,2002).Analysisoflightchain bandsfrompeak1andpeak2ofsalmonshowedthatbothfrac- tionscontainedthetwomostcommonlightchainisotypes;IgL1 and IgL3 (Solem et al., 2001). BLAST matches are indicated in Figs.1and2.

(7)

Fig.3. Cross-reactivitybetweenMAb4C10andIgMofsalmon,browntroutandrain- bowtrout.(A)Mab4C10reactedinWesternblotswithrainbowtrout␮,salmon␮A, browntrout␮B,butnotwithsalmon␮Borbrowntrout␮A.(B)MAb4C10reacted alsowithdeglycosylatedproteininWesternblots;thereactionwithsalmon␮Ais shown.

3.2. MAb4C10reactswithAinAtlanticsalmonwhereasit reactswithBinbrowntrout

MAb4C10wasfoundtoreactinWesternblotswithsalmon␮A, butnotwithsalmon␮B.Inbrowntroutitwasopposite:itreacted with␮B,butnotwith␮A.MAb4C10alsoreactedwiththedeglyco- sylatedpolypeptide(Fig.3).Identiﬁcationoftheimmunopuriﬁed nativesalmonIgMshowedthatitwastheIgM-Asubpopulationthat wascapturedbymagneticbeadscoatedwithMAb4C10(resultsnot shown).

3.3. MAb4C10captureoflymphocytesfromAtlanticsalmonand browntrout

AsillustratedinFig.4A,MAb4C10/Dynabeadscapturedleuko- cyteswiththeexpectedsizeandformofB-cells.Theabundanceof

␮Aversus␮BtranscriptsincellscapturedbyMAb4C10/Dynabeads wasanalyzedbyaPCRapproach.AnEcoRIrestrictionsite in␮1 isuniqueforisotype␮Ainsalmon,allowingaroughestimateof whichtranscriptsaremostabundant.EcoRIrestrictionofPCRprod- uctsthatoriginatedfromcDNAofcellsthathadbeencapturedby theMAb4C10/Dynabeadsindicatedthatthe␮Atranscriptswere predominant(Fig.4B).Cloningandsequencing ofPCR products weredonetoverifytheﬁndings.ForbrowntroutaSau3Arestric- tionsitein␮3isuniquefor␮B.Theprocedurewasrepeatedfor browntrout,showingthatMAb4C10/Dynabeadscapturedcellshad primarilytranscriptsfor␮B(resultsnotshown).

3.4. SearchingforapossibleMAb4C10bindingepitope

The fact that MAb4C10 reacts with membrane bound IgM stronglyindicatesthatthereactiveepitopemustbelocatedin␮1-

␮2-␮3,since␮4isnotpartoftheB-cellreceptor.Alignmentof aminoacidsubstitutionsin␮Aand␮BofAtlanticsalmoncompared to␮Aand␮Bofbrowntroutshowsthattheonlycommonandchar- acteristicresiduein␮BisanextracysteineneartheC-terminal partof␮4.Interestingly,threesubstitutedpositionsin␮3were

Fig.4. CaptureoflymphocyteswithMAb4C10/Dynabeads,andrestrictionanalysis of␮RT-PCRproducts.(A)EMpictureoflymphocytecapturedbyMAb4C10.(B)RT- PCRproductsdigestedwithEcoRI.Lymphocytesfromfoursalmonindividualswere capturedbyMAb4C10/Dynabeads.RNAwaspuriﬁedfromcapturedcellsandfrom cellsthatwerenotcaptured.ThesamplesweresubjectedtoRT-PCR,followedby restrictionwithEcoRI.Untreatedleukocyteswereusedasareference(left).Cleavage ofthemajorfractionofthePCRproductsindicatedthat␮Aismostabundantinthe capturedcellswhereas␮Bismostabundantinthecellsthatwerenotcaptured.

thesameinsalmon␮Aandbrowntrout␮B,pointingtoapossible regionforinteractionwithMAb4C10(Fig.5).Modelingof␮3gave furthersupportforthishypothesis:thethreeresiduesareexposed onthesurfaceoftheIgfoldwhichcouldaccountforthereactiv- ityagainstnativeIgM(Fig.6).However,MAb4C10reactivityinthis partof␮3isnotobvioussincethecorrespondingresiduesinrain- bowtroutaredifferent.Still,asdiscussedbelow,␮3wasthought tobethebestcandidateforMAb4C10reactivitywhenconsidering thephysiochemicalpropertiesoftheactualaminoacids.

3.5. ExperimentalevidenceforbindingbetweenMAb4C10and3

TheeffectofDNA-fragmentsencodingsalmon␮A3,salmon␮B3, browntrout␮A3,browntrout␮B3andrainbowtrout␮3were analyzedaftertransfectionintoSH-SY5Ycells.Inaccordancewith thehypothesisdiscussedinSection3.4,MAb4C10showedreac- tivitywithsalmon␮A3,browntrout␮B3andrainbowtrout␮3, butnoreactivitywithsalmon␮B3andbrowntrout␮A3(Table1).

Althoughasubjectiveobservation,thereactivityintransfectedcells appeared tobesomewhatstrongerwithsalmon␮A andbrown trout␮Bcomparedtorainbowtrout␮3.

Table1

Transfectionandimmunostaining.

pcDNAplasmid Reactivityagainst FlagAb

Reactivityagainst MAb4C10

Atlanticsalmon␮A3 + +

Atlanticsalmon␮B3 + −

Browntrout␮A3 + −

Browntrout␮B3 + +

Rainbowtrout␮3 + +

(8)

μ2 μ1

μ3

μ4

22

μA μB

salmon trou t salmon trou t

56 60 64 96 100 122 147 168 191 197 219 225 227 237 247 251 270 272 275 304 325

401 381 361

433 444 331

M S A

V A M

M Q D

A K A E

T N E P K V R S L K P T K Y S

M G V

V V T

K K E

V N S E

I N K P K V R T P Q L S E N S

V S A

L V M

K K E

A N S K

I N K T M V K S L K P S K N C

M G A

V V T

K K E

A K S E

T S E P K A R T L Q P S K N

*

C

Fig.5. Aminoacidsubstitutionsin␮Aand␮BofAtlanticsalmonandbrowntrout:

searchforapossibleMab4C10reactiveepitope.Residuesthatareidenticalinsalmon

␮Aandbrowntrout␮B,andatthesametimedifferentfromsalmon␮Bandbrown trout␮Aareindicatedwitharrows.

4. Discussion

ThepresentstudyhasshownthattwoIgMsubpopulationsin Atlanticsalmonandbrowntroutcorrespondto␮Aand␮B,previ- ouslycharacterizedbycDNAcloning(Hordviketal.,1992;Hordvik etal.,2002).TheIgMfractionwhichwaselutedﬁrstontheanion exchangercontained␮B,deﬁnedbyanextracysteineresiduenear theC-terminalendofthepolypeptide.Theextracysteineresiduein

␮4istheonlyresiduethatiscommonforsalmonandbrowntrout

␮Bandatthesametimeisdifferentfrombothsalmonandbrown trout␮A.

Somewhatunexpectedwefoundthatthemonoclonalantibody MAb4C10, originally raised against rainbow trout IgM, reacted exclusivelywith␮AinAtlanticsalmonandexclusivelywith␮B inbrowntrout.The␮Btranscriptsinsalmonwerepreviouslyesti- matedtoconstituteabout60%oftotal␮mRNAinleukocytesand immune organs of healthy ﬁsh.This estimate is also in agree- ment withtheratioof IgM-Bversus IgM-Apredictedfrom the anionexchangeelutionproﬁles(Hordviketal.,2002).Duringmany yearswehaveobservedthatbothIgM-Aand IgM-Barepresent ineveryexaminedindividual.Thus,itisplausibletoassumethat

MAb4C10 reactswith40–60%of theIgMpopulationinAtlantic salmon and browntrout, and accordingly, with40–60% of the IgM+lymphocytesinhealthyﬁsh.

As illustrated in Fig. 5, substitutions in ␮3 might explain MAb4C10 reactivity with salmon␮A and browntrout ␮B, and absenceof reactivitywithsalmon␮Bandbrowntrout␮A.The patternofsubstitutionsindicatesthatrecombinationhasoccurred betweentheparalogousAandBlociineithersalmonorbrowntrout aftertheintroductionoftherelevantmutations.Inrainbowtrout, thecorrespondingpositionsareoccupiedbyaminoacidsT₂₂₅,I₂₂₇ andN₂₄₇,whicharedifferentfromthoseinsalmon␮Aandbrown trout␮B,i.e.,K225,T227andE247.The227positionisoccupiedby anisoleucineinrainbowtrout,correspondingtothatinsalmon␮B andbrowntrout␮A(N₂₂₅,I₂₂₇andK₂₄₇).

In general,theteleost␮3sequencesshowhigherdivergence ratesthan␮1,␮2and␮4.Twoslightlydifferentallelicvariantsof rainbowtrout␮,differinginposition247(NversusI,respectively) havebeenreportedtothedatabases(AnderssonandMatsunaga, 1993;Leeetal.,1993;Hansenetal.,1994).However,allelicdif- ferencesarenotrelevantinthepresentstudysincetransfection constructs were sequenced and veriﬁed to be identical to the previously reportedsequences showninFig.1.Whenconsider- ingfeatures oftheaminoacidsinrainbowtrout,thepattern of MAb4C10reactivityisstillreasonable.Itistemptingtospeculate onthepossibilitythatthenegativelychargedE247insalmon␮Aand browntrout␮BiscompatiblewithMAb4C10reactivitywhereasthe positivelychargedK₂₄₇insalmon␮Bandbrowntrout␮Amight repeltheinteractionwithMAb4C10.

Whereas the putative MAb4C10 binding epitope must be exposedonthesurface,aputativeN-glycosylationsiteislocated on the opposite site of the Ig fold, i.e., towards the apparent core of the IgM monomer (Fig. 6F). This site is conserved in salmon, brown trout, rainbow trout and char (Hordvik et al., 2002)andcouldrepresentabindingsite forcarbohydrate moi- etiesinvolvedinstabilizationofthemolecule.Anexactdeﬁnition oftheepitope–MAb4C10interactionmightprovideusefulinfor- mationwithregardtofurtherexperimentsandunderstandingof theIgM-AversusIgM-Bstructure.Thepresentstudydidnotreveal anyassociatedmoleculesthatcouldpossiblyexplainwhyIgM-A andIgM-Bareelutedintwodistinctpeaksbyanionchromatogra- phy.Thus,westillholdtothehypothesisthattheextracysteinein theC-terminalpartof␮Bhassomemajorimpactonthepolymer structureandthatthisleadstothecharacteristicelutionproﬁleof IgM-AandIgM-B(Hordviketal.,2002).

Thepresentstudyhasnotaddressedfunctionalaspectsrelated tothepresenceofIgM-AandIgM-BinAtlanticsalmonandbrown trout. However,since ␮Aand ␮Bhave continuedto exist over a long time during evolution it is likely that this variety has somebiologicalsignificance.Arecentstudyshowedaconnection betweengreaterantibodyaffinityandincreaseddisulfidepolymer- ization inrainbowtrout:itwasdemonstratedthathighaffinity B-cellsproducemorehighlypolymerizedIgM,andthatthehigh- affinity, highlypolymerized antibodiespossess longerhalf-lives thanlower-affinityantibodies(Yeetal.,2010).Variabilityininter- heavychainpolymerizationoftheIgMtetrameriscommonamong teleostfish(Kaattarietal.,1998;Bromageetal.,2006;Yeetal., 2011).Inchannelcatfish,anadditionalcysteineresidueintheC- terminalsequenceof␮wasfoundtobeessentialforestablishinga seriesofcovalentlyinter-bondedformsofIgM(GhaffariandLobb, 1989;Getahunetal.,1999).Inthiscontextitisstrikingthatthe␮A and␮Bvariantsofsalmonandbrowntroutdifferwithrespectto thepresenceofanextracysteineneartheC-terminalpart.Recom- binationbetweentheparalogouslocicouldeasilyallowoneofthe variantstotakeover,butbothvariantshavebeenmaintaineddur- ingevolution.Thus,therearereasonstobelievethatthepresence ofbothIgM-AandIgM-Bisbeneficial.Amongothers,aninterest-

(9)

Fig.6. Modelingof␮3variantsinAtlanticsalmon,browntroutandrainbowtroutrevealssurfaceexposureofputativeMab4C10reactiveaminoacids.Keyresiduesindicated witharrowsinFig.5aredepictedinred(hydrophobic),blue(negativelycharged),brown(positivelycharged)andgreen(neutral),respectively.(A)Atlanticsalmon␮A3,(B) Atlanticsalmon␮B3,(C)Browntrout␮B3,(D)Browntrout␮A3,(E)Rainbowtrout␮3and(F)Rainbowtrout␮3;aconservedN-glycosylationsiteinrainbowtrout,Atlantic salmonandbrowntrout(N288)ontheoppositesideoftheIgfoldandtheputativeMab4C10bindingsiteisindicated.

ingtopicforafollowupstudywillbetofindoutwhetherthereisa differencetheinter-heavychainpolymerizationofIgM-AandIgM- B,andifthiscanbecorrelatedtohighaffinityversuslowaffinity antibodies.

Acknowledgements

We thank Christian deVries Lindstrøm for his contributions during the initial immunomagnetic separation and analysis of lymphocytes,KunioriWatanabeforexcellenttechnicalassistance withtheelectronmicroscopyanalysis,andDr.KariFladmarkfor accesstolabfacilities allthroughthetransfectionexperiments.

ThepresentworkwassupportedbygrantsfromtheMeltzerFund, UniversityofBergen.

References

Acton,R.T.,Weinheimer,P.F.,Hall,S.J.,Niedermeier,W.,Shelton,E.,Bennett,J.C., 1971.Tetramericimmunemacroglobulinsinthreeordersofbonyﬁshes.Proc.

Natl.Acad.Sci.U.S.A.68,107–111.

Allendorf,F.W.,Thorgaard,G.H.,1984.Tetraploidyandtheevolutionofsalmonid ﬁshes.In:Turner,B.J.(Ed.),EvolutionaryGeneticsofFishes.PlenumPress,New York,pp.1–55.

Andersson,E.,Matsunaga,T.,1993.CompletecDNAsequenceofarainbowtrout IgMgeneandevolutionofvertebrateIgMconstantdomains.Immunogenetics 38,243–250.

Andersson,E.,Peixoto,B.,Tormanen,V.,Matsunaga,T.,1995.Evolutionofthe immunoglobulinM constantregion genesofsalmonid ﬁsh,rainbow trout (Oncorhynchusmykiss)andArcticchar(Salvelinusalpinus):implicationscon- cerningdivergencetimeofspecies.Immunogenetics41,312–315.

Bromage,E.S.,Ye,J.,Kaattari,S.L.,2006.Antibodystructuralvariationinrainbow troutﬂuids.Comp.Biochem.Physiol.BBiochem.Mol.Biol.143,61–69.

Estevez,J.,Sanches,C.,Dominguez,J.,Leiro,J.,Sanmartin,L.,Ubeira,F.M.,1993.

Protein-A binding characteristics of rainbow trout (Oncorhynchus mykiss) immunoglobulins.Comp.Biochem.Physiol.106B,173–180.

Feng,L.N.,Lu,D.Q., Bei,J.X.,Chen,J.L.,Liu,Y.,Zhang,Y.,Liu,X.C.,Meng,Z.N., Wang,L.,Lin,H.R.,2009.Molecularcloningandfunctionalanalysisofpoly- mericimmunoglobulinreceptorgeneinorange-spottedgrouper(Epinephelus coioides).Comp.Biochem.Physiol.154B,282–289.

Fuda, H., Soyano,K., Yamazaki,F.,Hara, A., 1991. SerumimmunoglobulinM (IgM)duringearlydevelopmentofMasusalmon(Oncorhynchusmasou).Comp.

Biochem.Physiol.99A,637–643.

Ghaffari,S.H.,Lobb,C.J.,1989.Cloningandsequenceanalysisofchannelcatﬁshheavy chaincDNAindicatephylogeneticdiversitywithintheIgMimmunoglobulin family.J.Immunol.142,1356–1365.