restricted distribution of connexin40, a gap junctional protein, in mammalian heart

Briand, AF Moorman and HJ JongsmaD Gros, T Jarry-Guichard, I Ten Velde, A de Maziere, MJ van Kempen, J Davoust, JP

heartRestricted distribution of connexin40, a gap junctional protein, in mammalian

ISSN: 1524-4571 Copyright © 1994 American Heart Association. All rights reserved. Print ISSN: 0009-7330. Online

TX 72514Circulation Research is published by the American Heart Association. 7272 Greenville Avenue, Dallas,

1994, 74:839-851Circulation Research

http://circres.ahajournals.org/content/74/5/839located on the World Wide Web at:

The online version of this article, along with updated information and services, is

http://www.lww.com/reprintsReprints: Information about reprints can be found online at [email protected]. E-mail:

Fax:Kluwer Health, 351 West Camden Street, Baltimore, MD 21202-2436. Phone: 410-528-4050. Permissions: Permissions & Rights Desk, Lippincott Williams & Wilkins, a division of Wolters http://circres.ahajournals.org//subscriptions/Subscriptions: Information about subscribing to Circulation Research is online at

by guest on July 15, 2011http://circres.ahajournals.org/Downloaded from

http://circres.ahajournals.org/

839

Restricted Distribution of Connexin40, a GapJunctional Protein, in Mammalian Heart

Daniel Gros, Therese Jarry-Guichard, Ingrid Ten Velde, Ann de Maziere,Marjan J.A. van Kempen, Jean Davoust, Jean Paul Briand,

Antoon F.M. Moorman, Habo J. Jongsma

Abstract Connexin40 (Cx40) is a member of the connexinfamily of gap junction proteins. Its mRNA, abundant in lung,is also present in mammalian heart, although in lower amount.Rabbit antipeptide antibodies directed to the COOH terminus(residues 335 to 356) of rat Cx40 were characterized toinvestigate the distribution of Cx40 in rat and guinea pigcardiac tissues. The affinity-purified antibodies detect specifi-cally a major protein (Mr, 40000) in immunoblots of totalextracts from rat lung and rat and guinea pig heart. In sectionsof guinea pig atrial tissue treated for immunofluorescence, astrong labeling associated with myocytes was seen with adistribution consistent with that of intercalated disks. Theresults of immunoelectron microscopy carried out with guineapig atrial tissue showed that epitopes recognized by theseantibodies were exclusively associated with gap junctions.These results, added to those of control experiments, demon-strate that antibodies 335-356 are specific for Cx40. Double-labeling experiments carried out with lung sections usinganti-factor VIII and anti-Cx40 antibodies suggest that Cx40 isexpressed in blood vessel endothelial cells. In guinea pig andrat heart sections, investigated using both immunofluorescence

and immunoperoxidase techniques, a signal was also found tobe associated with vascular walls. In guinea pig heart, onlyatrial myocytes are Cx4O-positive. No labeling was detected inventricular myocytes, including those of the His bundle andthe bundle branches, which otherwise do express connexin43(Cx43). In rat heart Cx4O -expressing myocytes are localized inthe conduction system, ie, the His bundle, the bundlebranches, and the Purkinje fibers. Cx43 is not detected eitherin the His bundle or in the proximal parts of the bundlebranches, and consequently, Cx4O is the first connexin dem-onstrated in this region of the rat conduction system. Cx40 wasnot detected in the working ventricular myocytes. Double-labeling experiments carried out with hen anti-Cx43 antibod-ies and rabbit anti-Cx4O antibodies demonstrated that, intissues expressing both Cx43 and Cx4O, these two connexinswere localized in the same immunoreactive sites. A few sites,however, appear to contain only one or the other of these twoconnexins. (Circ Res. 1994;74:839-851.)Key Words * connexins * gap junctions * heart -

conduction system

A djacent cells are directly coupled, metabolically andelectrically, through intercellular (or junctional)channels that aggregate to form gap junctions.

Each junctional channel consists of two hemichannels (orconnexons) joined together in the extracellular junctionalspace separating two adjacent cells.12 Connexons arehexamers built from proteins that belong to the connexinfamily.3 Twelve mammalian connexin genes have beencloned and sequenced so far.4-18 Predicted molecularmasses of connexins range from 26 to 70 kD, and topolog-ical investigations carried out with connexin32 (Cx32)19-22and connexin43 (Cx43)23-25 have demonstrated that theseproteins have cytoplasmic amino and carboxyl termini andcomprise four transmembrane domains bound by one

Received June 25, 1993; accepted December 29, 1993.From the Laboratoire de G6n6tique et Physiologie du Develop-

pement (D.G., T.J.-G.), UMR C99-43, and Centre d'Immunologie(J.D.), CNRS-INSERM, Facult6 des Sciences de Luminy, Univer-site d'Aix-Marseille II, France; the Department of Physiology(I.T.V., A. de M., H.J.J.) and the Department of Anatomy andEmbryology (M.J.A. van K., A.F.M.M.), Faculty of Medicine,University of Amsterdam, The Netherlands; and Structure desMacromol6cules Biologiques et Mecanismes de Reconnaissance(J.P.B.), UPR CNRS 9002, Strasbourg, France.Correspondence to Dr Daniel Gros, Laboratoire de G6n6tique

et Physiologie du D6veloppement, Facult6 des Sciences de Lu-miny, Avenue de Luminy, Case 907, 13288, Marseille Cedex 9,France.

cytoplasmic loop and two extracellular loops. These do-mains are linked by an intercellular disulfide bridge.26Y27

In excitable tissues such as myocardium and smoothmuscle, electrical coupling through junctional channelsis responsible for action potential propagation.28 Cx43 isthe major protein of mammalian heart junctional chan-nels,6,23,29-31 but it has not been detected in all cardiactissues. Thus, in rat heart, Cx43 cannot be detectedeither in the sinoatrial and atrioventricular nodes32 or inthe His bundle and the proximal part of the bundlebranches.32,33 In human and bovine heart, Cx43 cannotbe detected in the sinoatrial and atrioventricularnodes,34,35 but in contrast to rat heart, it is expressed inthe His bundle and the bundle branch. From electronmicroscopy studies, it is known that gap junctions arepresent in the aforementioned cardiac tissues,36,37 whichcannot be stained immunohistochemically. This contra-diction can only be resolved by hypothesizing that atleast one connexin other than Cx43 is expressed in thesetissues.mRNAs for connexin46 (Cx46), connexin45 (Cx45),

connexin40 (Cx40), and connexin37 (Cx37) have beendetected in low amounts in rat, mouse, and dogheart.8"-'5'7 The following questions remain: withwhich type(s) of cardiac cells (myocytes, fibroblasts,smooth muscle cells, or endothelial cells)38 are theseconnexins associated, and furthermore, if they are asso-ciated with myocytes, what is the distribution in myo-



840 Circulation Research Vol 74, No S May 1994

cardium of junctional channels composed of these con-nexins? Cx40 from dog, rat, and mouse has recentlybeen cloned and sequenced,'3-15,17 and it has beenshown that cDNAs coding for rat and mouse Cx40express functional junctional channels.'31739 To findanswers to some of these questions, we have raised andcharacterized antipeptide antibodies directed to thecarboxyl terminal domain of rat Cx40. These antibodieswere used to investigate the expression and spatialdistribution of Cx40 in guinea pig and rat heart. Theresults obtained demonstrate that Cx40 is differentlydistributed in the heart of these species. This distribu-tion is also different from that recently described in thedog. 40,41

Materials and MethodsPreparation and Purificationof Site-Directed AntibodiesThe peptide corresponding to amino acids 335 to 356

(YHSDKRRLSKASSKARSDDLSV) of rat Cx4014,17 wassynthesized according to Merrifield42 in 1963 and conjugatedto keyhole limpet hemocyanin using bis-diazobenzidine.43 Thesequence YHSDKRRLSKASSKARSDDLSV is conserved inmouse Cx4013; histidine (H) is replaced by glutamine (0) inthe corresponding sequence of dog Cx40.'5 Immunization ofrabbits and the collection of serum have been previouslydescribed in detail.29 Enriched-IgG fractions were purifiedfrom sera using a Trisacryl M-DEAE column; site-directedantibodies were affinity-purified from the previous fractionsusing AH-Sepharose 4B coupled to the peptide by means of1-ethyl-3-(3-dimethylaminopropyl)carbodiimide HCl.29 Pre-immune sera were subjected to the same purification steps asimmune sera. The collected fractions, used in control experi-ments, will be referred to as preimmune fractions.

Antibodies raised in rabbit to the peptide SAEQNRMGQY(residues 314 to 322 of Cx43,6 to which a tyrosine residue wasadded for coupling convenience) were previously shown to bespecific for Cx43 by immunoblotting, immunofluorescence,and immunoelectron microscopy.3T44 The same peptide wasused to raise antibodies in hen. This peptide was conjugated tobovine serum albumin (BSA) using bis-diazobenzidine asabove. The immunization protocols for hens were similar tothose used for rabbits. IgY fractions were extracted from eggyolks according to Polson et al.45 Hen site-directed antibodieswere affinity-purified from the previous fractions using AH-Sepharose 4B coupled to the immunogenic peptide. Preim-mune yolks were subjected to the same purification steps asimmune yolks to obtain preimmune fractions.

In enzyme-linked immunosorbent assays,29 purified rabbitand hen antibodies reacted strongly with the synthetic pep-tides to which they were raised, with no evidence of cross-reactivity between antibodies and the unrelated peptideYPSSRASSRASSRPRPDDLQI.

ImmunoblottingPartially purified gap junction fractions were prepared from

rat heart in the presence of phenylmethylsulfonyl fluoride (1mmol/L). The procedure used was similar to that previouslydescribed,29 but the last purification step on sucrose gradientwas omitted. Samples were solubilized for 30 minutes in 62.5mmol/L Tris-HCI (pH 6.8), 2% sodium dodecyl sulfate, 5%2-mercaptoethanol, 10 mmol/L EDTA, and 5% glycerol andfractionated by electrophoresis. Gels were stained withCoomassie brilliant blue R-250. Other samples were electro-phoresed and electrotransferred onto nitrocellulose mem-brane.29 Immunoreplicas were incubated at 4°C for 8 hourswith the saturation solution (40 mmol/L Tris, pH 7.5, 0.1%Tween 20, and 4% nonfat dry milk) and then overnight witheither purified hen antibodies (20 gg/mL) or rabbit anti-Cx43

antibodies (1 gig/mL).3' After several washes, the replicaswere incubated for 1 hour with biotinylated-goat anti-henantibodies (1:500, Kirkegaard and Perry Laboratories, Gaith-ersburg, Md) or biotinylated-goat anti-rabbit antibodies(1:1000, Jackson Immunoresearch Laboratories, West Grove,Pa), washed, and incubated for 1 hour with peroxidase-labeledstreptavidin (1 1000, Jackson Immunoresearch Laboratories).Peroxidase activity was revealed as previously described.29Lungs and whole hearts from adult rats and guinea pigs as

well as atria isolated from guinea pig hearts were frozen andpulverized in liquid nitrogen before freeze drying. Freeze-dried samples were solubilized, fractionated by electrophore-sis. and then electrotransferred onto nitrocellulose membranesas described above. Saturated immunoreplicas were incubatedovernight at 4°C with rabbit anti-Cx40 antibodies (3 to 10,tg/mL) or rabbit anti-Cx43 antibodies (1 jzg/mL),3144 washed,and then incubated with '251-labeled protein A. Exposure to

_) m

ay r b c d e f

94=-67_-

43_ 3

29.

21._ 0

14._

FiG 1. Characterization by immunoblotting of rabbit antibodiesraised to residues 335 to 356 of rat connexin40 (antibodies335-356). Relative molecular weight (Mr) of proteins detected inthe immunoreplicas was determined by comparison with molec-ular weight standards and cardiac connexin43 (lane a). Mobilityof molecular mass standards is indicated in kilodaltons on theleft of the figure (phosphorylase a, Mr 94 000; bovine serumalbumin, Mr 67 000; ovalbumin, Mr 43 000; carbonic anhydrase,Mr 29 000; soybean trypsin inhibitor, Mr 21 000; and lysozyme,Mr 14 000). The primary antibodies were revealed using 1251-labeled protein A. The arrow indicates the top of immunorepli-cas. Lane a shows rat heart total extracts. This replica wasprobed with rabbit anti-connexin43 antibodies.31 A major proteincorresponding to connexin43 is detected in this immunoreplica.Lanes b, c, d, e, and f show rat lung total extracts. The wells ofelectrophoresis gels corresponding to these lanes were loadedwith 10 times more protein than the well corresponding to lane a.Immunoreplicas a, b, and c, on one hand, and immunoreplicasd, e, and f, on the other hand, come from two separate series ofexperiments. Lanes b and d show replicas probed with rabbitantibodies 335-356 (5 gg/mL). A major protein of 40 kD isdetected in both immunoreplicas. Lane c shows replica probedwith a preimmune fraction (see "Materials and Methods" for thepreparation of preimmune fractions). No protein was detected inreplicas treated with preimmune fractions. Lanes e and f showreplicas probed with rabbit antibodies 335-356 preincubatedeither with a peptide (75 ug/mL) unrelated to the immunogenicpeptide (lane e) or with the immunogenic peptide (75 ,ug/mL)(lane f). The 40-kD band is not detected in the replica shown inlane f.



Gros et al Distribution of Cx4O in Heart 841

Hyperfilm (Amersham International, Buckinghamshire, UK)was carried out at -90°C with intensifying screens.

Specificity of the labelings due to rabbit anti-Cx40 antibod-ies or hen antibodies was checked by (1) omission of antibod-ies, (2) substitution of antibodies for a preimmune fraction,and (3) substitution of antibodies for the corresponding anti-bodies preincubated overnight at 4°C with either the immuno-genic peptide (YHSDKRRLSKASSKARSDDLSV or SAE-QNRMGQY, 50 to 100 jig/mL) or the unrelated peptideLQLLREAEKAAKEAGGTRY (100 gig/mL).

ImmunofluorescenceIsolated whole hearts and heart and lung samples from adult

guinea pigs (Dunkin Hartley) and rats (Wistar) were eitherdirectly frozen (in isopentane precooled at -30°C or liquidnitrogen) or fixed before freezing. Fixation was carried outwith either 2% paraformaldehyde in phosphate-buffered sa-line (PBS), pH 7.4, or periodate/lysine/paraformaldehyde46from 1 hour to overnight. After gradual infiltration with 20%sucrose/Tissue Teck OCT compound (2:1) (Miles Laborato-ries, Elkhart, Ind) or 20% sucrose, the samples were frozen inliquid nitrogen. Frozen sections were collected on poly-L-lysine-coated slides and stored at -90°C until use.

After equilibration to room temperature, sections wererehydrated in PBS and saturated with 2% BSA in PBS. Onlysections of fixed samples were incubated with 0.2% TritonX-100 in PBS for 1 hour and then with 0.5 mol/L NH4C1 for 15minutes before saturation. Sections were incubated overnightat room temperature or at 4°C with rabbit anti-Cx40 antibod-ies (5 to 15 ,ag/mL), hen anti-Cx43 antibodies (5 gg/mL), orrabbit anti-Cx43 antibodies3' (1 to 3 gg/mL) diluted in thesaturation solution. Sections were rinsed with PBS and thenincubated for 1 hour with tetraethyl rhodamine isothiocyanate(TRITC)-labeled goat anti-rabbit IgGs (Jackson Immunore-search Laboratories) or fluorescein isothiocyanate (FITC)-labeled goat anti-hen antibodies (Kirkegaard and Perry Lab-oratories). For the treatment of guinea pig sections, thesecondary antibodies, diluted to 1:10 with the saturationsolution, were mixed for 30 minutes with 10% normal guineapig serum and centrifuged (15 minutes at 15 000 rpm). Thesupernatant was collected and used for labeling. The controlexperiments carried out to check the specificity of labelingswere as follows: (1) omission of primary antibodies, (2)substitution of primary antibodies for a preimmune fraction,and (3) substitution of primary antibodies for primary anti-bodies preincubated overnight at 4°C with immunogenic pep-tide (15 to 50 ,ug/mL).For double-labeling experiments, unfixed frozen sections

from rat heart, guinea pig left atrium, and rat lung were firstincubated overnight at 4°C with rabbit anti-Cx40 antibodiesand then either with hen anti-Cx43 antibodies (8 gg/mL)(heart sections) or with goat anti-factor VIII antibodies (di-lution, 1:1000; Nordic Immunological Laboratories, Tilburg,The Netherlands) (rat heart and lung sections) for 2 to 6 hoursat room temperature. The sections were rinsed with PBS andincubated for 1 hour at room temperature with PBS/BSAcontaining either both TRITC-labeled goat anti-rabbit IgGsand FITC-labeled goat anti-hen antibodies or both TRITC-labeled donkey anti-rabbit IgGs and FITC-labeled donkeyanti-goat antibodies (Jackson Immunoresearch Laboratories).The sections were examined with a Zeiss III or a NikonDiaphot light microscope equipped for epifluorescence withthe appropriate filters to distinguish FITC emission fromTRITC emission.

Double-labeled heart sections were also analyzed by confo-cal laser scanning (CLS) microscopy using a Leica instrumentbased on a Leitz Diaplan microscope interfaced with an argonion laser (model 2020, Spectra-Physics, Mountain View, Calif)adjusted to 488 nm and with an air-cooled helium/neon laserproviding a 543-nm line. Horizontal optical sections weremonitored first for fluorescein by using 488-nm excitation via a

CC

ya b 'c d e f

94-67...

43-

29:

21.-

14_

FIG 2. Characterization by immunoblotting of rabbit antibodiesraised to residues 335 to 356 of rat connexin40 (antibodies335-356). Relative molecular weight (Mr) of proteins detected inthe immunoreplicas was determined by comparison with molec-ular weight standards and cardiac connexin43 (lane a). Mobilityof molecular mass standards is indicated in kilodaltons on theleft of the figure (see Fig 1). The primary antibodies wererevealed using 1251-labeled protein A. The arrow indicates the topof immunoreplicas. Lanes a and b show rat heart total extracts;lanes c and d, guinea pig heart total extracts; and lanes e and f,total extracts of atria isolated from guinea pig heart. The wells ofelectrophoresis gels corresponding to lanes b to f were loadedwith 10 times more protein than the well corresponding to lane a.In lane a, replica was probed with rabbit anti-connexin43 anti-bodies.31 A major protein corresponding to connexin43 is de-tected in this immunoreplica. In lanes b, c, and e, replicas wereprobed with rabbit antibodies 335-356 (5 gg/mL). A majorprotein of 40 kD is detected in these immunoreplicas. In lanes dand f, replicas were probed with a preimmune fraction (lane d)(see "Materials and Methods" for the preparation of preimmunefractions) or with antibodies 335-356 preincubated with theimmunogenic peptide (50 tg/mL) (lane f). In both cases, noprotein is detected.

510-nm dichroic mirror, an FITC band-pass 520- to 560-nmbarrier, and a green-sensitive photomultiplier. The same sec-tions were then scanned to detect the rhodamine signal byusing 543-nm excitation via a 580-nm dichroic mirror, along-pass barrier filter >580 nm, and a red-sensitive photo-multiplier. The two 8-bit-encoded 512x512 pixel images fromoptical sections were combined and visualized with a comput-er-generated green and red pseudocolor scale adapted forfluorescein and rhodamine.

Immunoperoxidase LabelingHearts isolated from adult guinea pigs and rats were fixed

overnight with a freshly made 4% paraformaldehyde solution.After dehydration with a graded series of ethanol followed by1-butanol or chloroform, tissue was embedded in ParaplastPlus (Monoject, Kildare, Ireland). Serial sections 7 gim thickwere collected on either poly-L-lysine or 3-aminopropyltri-ethoxysilane-coated slides and then stored at 4°C until use.After deparaffination of the sections, immunologic stainingwas carried out using the unconjugated peroxidase/antiperox-idase method.47 Primary antibodies used were either rabbitantipeptide antibodies to Cx40 or Cx433' or mouse monoclo-nal antibodies against human desmin (Monosan Sambio,




A, 0

FIG 3. Characterization by immunofluorescence of rabbit antibodies raised to residues 335 to 356 of rat connexin40 (CX40, antibodies335-356) in unfixed rat lung. Panels a (phase-contrast micrograph) and b (fluorescence micrograph, CX40) display a cross-sectionedblood vessel. Note the labeling of the vascular wall. Panels c (factor VilI) and d (CX40) show the distribution of factor Vil andimmunoreactive sites to antibodies 335-356, respectively, in another vessel (double-labeling experiment). Small arrows in panels c andd show that the immunoreactive sites to antibodies 335-356 are associated with the borders of factor Vill-positive cells. Bar=25 gum inpanels a and b; bar=20 gm in panels c and d.

Uden, The Netherlands). Sections were examined with aNikon Diaphot light microscope.

Sections are usually incubated for 20 to 30 minutes with a3,3'-diaminobenzidine (DAB)/H202 solution to reveal peroxi-dase activity.47 However, because more desmin is expressed inthe conduction system cells compared with the ventricularcells,48 any given incubation time produces a more intensestaining in the conduction system than in the surroundingventricular tissue. This differential staining was used to illumi-nate the topography of the conduction system within theventricles. A 10-minute reaction time with DAB/H102 solutionwas found to be the best time to get this differential staining.

Immunoelectron MicroscopyGuinea pig hearts were fixed by perfusion with 4% parafor-

maldehyde in sodium cacodylate (0.1 mol/L, pH 7.4). Smalltissue samples (<0.5 mm3n) from atria were immersed for 1.5 to2 hours at 4°C in the same fixative. Samples from guinea pigright atrium were also fixed directly by immersion with perio-date/lysine/paraformaldehyde.46 After quenching with 0.5mol/L NH4C1 or 50 mmol/L glycine for 1 hour, the sampleswere dehydrated and embedded in Lowicryl K4M at 4°C or at-35 C.49 Ultrathin sections were saturated with 5% normalgoat serum and 1% BSA in PBS for 10 minutes and thenincubated overnight with antibodies 335-356 (8 gg/mL) di-luted in the saturation solution. After rinsing with PBS/BSA,sections were incubated for 4 hours with 5-nm gold-labeledgoat anti-rabbit IgGs (Amersham) diluted to 1:10 with PBS/BSA, extensively rinsed, and stained with uranyl acetate andlead. Sections were examined with a Philips 420 electronmicroscope. Control experiments were similar to those carriedout for immunofluorescence experiments.

ResultsCharacterization of Site-Directed RabbitAnti-Cx40 AntibodiesImmunoblotting

Affinity-purified antibodies raised to residues 335 to356 of rat Cx40 (antibodies 335-356) were first used to

probe replicas of total extracts of rat lung. 12 I-protein Areveals in these replicas a single protein with Mr of40 000 that migrates just below cardiac Cx43 (Fig 1,lanes a, b, and d). This protein is not detected inreplicas incubated either with preimmune fractions (Fig1, lane c) or with antibodies 335-356 preincubated withthe immunogenic peptide YHSDKRRLSKASSKARS-DDLSV (Fig 1, lane e). In contrast, it is detected inreplicas probed with antibodies preincubated with thepeptide LQLLREAKAAKEAGGTRY, unrelated to theimmunogenic peptide (Fig 1, lane e). Identical resultswere obtained with guinea pig lung (not shown). Inreplicas of total extracts of rat and guinea pig heart,probed with antibodies 335-356, 1251-protein A reveals a

major band that migrates just below that of cardiac Cx43,as shown above (Fig 2, lanes a, b, and c). This band,which is heavily labeled in guinea pig atria replicas (Fig 2,lane e) and is usually rather weak in rat heart, is notdetected in replicas probed either with preimmune frac-tions (Fig 2, lane d) or with antibodies preincubated withthe immunogenic peptide (Fig 2, lane f).

ImmunofluorescenceThe results were identical whatever the method of

fixation used. In lung sections treated with antibodies335-356, the main immunoreactive sites are associatedwith blood vessel walls (Fig 3a and 3b). Double-labelingexperiments with antibodies 335-356 and anti-factorVIII antibodies suggest that these immunoreactive sitesare localized between factor VIII-positive cells (Fig 3cand 3d). Immunoreactive sites are also seen at highmagnification between alveolar cells (not shown), but noattempt was made to identify these cells. In sections ofguinea pig atrial tissue treated with antibodies 335-356,a strong labeling associated with myocytes was observed(Fig 4a). Distribution of this labeling is consistent with

Rat Iunb

s

t. v

8125 '94

Al

A

'

I-

'tk




FIG 4. Characterization by im-munofluorescence of rabbit anti-bodies raised to residues 335 to356 of rat connexin40 (CX40, an-tibodies 335-356) in unfixedguinea pig heart. Panels a (CX40)and b (connexin43 [CX43]) dis-play two adjacent sections incu-bated with antibodies 335-356and rabbit anti-CX43 antibodies,31respectively. The sections gothrough the left atrium (LA) andthe upper part of the left ventricu-lar wall (LV). The distribution ofCX43 in these tissues makesthem easily distinguishable (b).Arrows indicate a cross-sectionedblood vessel (a) or its location (b)in the left atrium. Arrowheads in-dicate the localization of endocar-dium. Epitopes recognized byrabbit antibodies 335-356 arestrongly expressed in the atrium(a). They are associated with bothvessel wall (arrow) and atrial myo-cytes. In these latter cells, thedistribution of immunoreactivesites is similar to that of CX43 (b).Note that the ventricular myo-cytes, which express CX43, donot express the epitopes recog-nized by antibodies 335-356.Bar-S50 gm.

that of intercalated disks, which otherwise also containCx43 (Fig 4b). Immunoreactivity to antibodies 335-356was also seen in blood vessel walls (Fig 4a). No labelingwas seen in the lung and cardiac tissue section subjectedto the various control experiments (see "Materials andMethods"), and in particular, labeling was abolished bypreincubation of antibodies 335-356 with the immuno-genic peptide (15 gg/mL).

Immunoelectron MicroscopyExamination of ultrathin sections from guinea pig atrial

tissue treated for immunoelectron microscopy shows thatcolloidal gold particles are exclusively associated with gapjunctions (Fig 5). Identical results were obtained whateverthe method of fixation used. No labeling was seen in thesections subjected to the control experiments.

Distribution of Myocyte-Associated Cx4O in GuineaPig and Rat HeartTo detect Cx4O and Cx43 in the same sections by

double-labeling experiments, antibodies to residues 314to 322 of rat Cx43 were raised in the hen. Character-ization of these antibodies is presented in Figs 6 and 7.The hen antibodies detect a major protein of 45 to 43kD in replicas of partially purified rat heart gap junc-tions (Fig 6, lane d). This protein has the same electro-phoretic mobility as Cx43 (Fig 6, lane c) as judged fromcomparison with the labeling of twin replicas incubatedwith previously characterized rabbit anti-Cx43 antibod-ies.144 In frozen sections of rat ventricular tissue incu-bated with the hen antibodies and examined by fluores-cence microscopy, the pattern of labeling is consistentwith the distribution of intercalated disks and indistin-




t-,-,~~~~~~~-.~~~1~r..L4'

'4 wt. a

FIG 5. Characterization by immunoelectron microscopy of rabbit antibodies raised to residues 335 to 356 of rat connexin40 (antibodies335-356) in paraformaldehyde-fixed guinea pig heart. Ultrathin sections from guinea pig heart atrial tissue were treated as described in"Materials and Methods." Note that gap junctions (arrows), easily identifiable as double-membrane profiles, are strongly and exclusivelylabeled with colloidal gold particles. Arrowhead indicates a desmosome; m, mitochondria; and my, myofilaments. Bar=400 nm.

guishable from that obtained with rabbit anti-Cx43antibodies3' (Fig 7). In sections subjected to controlexperiments, no fluorescence was detected.

Distribution of Cx40 in rat and guinea pig heart wasinvestigated in serial sections by immunofluorescenceand immunoperoxidase labeling. The results obtainedwere similar whatever the method of fixation and inves-tigation (immunofluorescence versus immunoperoxidaselabeling) used. In guinea pig heart, only atrial myocyteswere found to express Cx40 (Fig 4a). Cx40 was notdetected either in the ventricular lateral walls (Fig 4a) orin any regions of the interventricular septum, althoughthis connexin is expressed in blood vessel walls of thesetissues. In contrast, Cx43 was detected both in atrial andventricular myocytes (Fig 4b) and also in the His bundleand the bundle branches (Fig 8). Distribution of bothCx40 and Cx43 was investigated by CLS microscopy inthe guinea pig atrial tissue (Fig 9). Analysis of the resultsshows that both connexins are colocalized in most of theimmunoreactive sites. However, in some sites, only Cx43or Cx40 is detected (Fig 9).Cx40 was found to be expressed in rat heart ventric-

ular tissue, but its expression is restricted to a fewwell-defined regions: (1) the upper apical region of theinterventricular septum (Fig 10), topographically cor-responding to the His bundle as shown by desminstaining, (2) two subendocardial regions extendingalong the septum and localized on both sides of theseptum (these regions correspond to the localizationof the His bundle branches) (Fig 11), (3) the falsetendons (chordae tendineae spuriae),50 joining theinterventricular septum to the lateral ventricular wallsand containing Purkinje fibers50 (Fig 11), and (4) somesubendocardial regions of the lateral walls, mainly intheir distal parts. In these latter regions, Cx40 isassociated with some groups of cells, forming short

cablelike arrangements. Cx40 was not detected in themyocytes of any other parts of the ventricular tissue,although this connexin is expressed in the wall of

a b c d

94

67 - --

43

29

21

14fl

FIG 6. Characterization by immunoblotting of hen antibodiesraised to residues 314 to 322 of rat connexin43 (CX43). Lanes aand b show Coomassie blue-stained gels of standard proteins(molecular masses in kilodaltons are indicated on the left) andpartially purified rat heart gap junctions, respectively. Note inlane b a major band of M, of 45 000 to 43 000 corresponding toCX43. Lanes c and d show immunoreplicas of partially purifiedgap junctions probed with rabbit anti-CX43 antibodies31,44 (1gg/mL) and hen antibodies (20 gg/mL) raised to residues 314 to322 of rat CX43, respectively. These primary antibodies wererevealed using biotinylated secondary antibodies and peroxi-dase-labeled streptavidin. CX43 detected in lanes c and d isindicated by arrowhead. No protein was detected in replicas ofcontrol experiments. Arrow indicates the top of separating gelsand immunoreplicas.




FIG 7. Characterization by immunofluorescence of unfixed ven-

tricular tissue from rat heart incubated with hen antibodies raised

to residues 314 to 322 of rat connexin43 (CX43). a, Phase-

contrast micrograph. b, Fluorescence microscopy. The frozen

section shown in a was incubated with the purified hen antibod-

ies (5 gg/mL) then with fluorescein isothiocyanate-labeled goat

anti-hen antibodies. The pattern of labeling is indistinguishablefrom that obtained with previously characterized rabbit anti-CX43

antibodies.3 Bar= 100 gi.

certain blood vessels irrigating these regions (Fig 12).Subendocardial groups of cells localized in the lateral

walls contain both Cx4O and Cx43. The distribution of

both connexins was investigated in these cells by CLS

microscopy. Examination of sections shows that, as in

guinea pig atrial myocytes, most of the immunoreac-

tive sites contain both Cx4O and Cx43, whereas some

of them contain only one or the other of the connexins

(not shown). In the rat atrial tissue, besides thelabeling of vascular walls, immunoreactivity to rabbitanti-Cx40 antibodies was seen as a very few scatteredfluorescent dots.

Distribution of Cx40 and Cx43 in guinea pig and ratheart is summarized in the Table.

DiscussionAntibodies 335-356 Specifically Recognize aProtein With M, of 40 000 Localized in GapJunctions: Cx40Cx40 mRNA has beeni shown to be present in

various rat and- mouse organs and tissues. This mRNAis expressed in large quantities in the lung and insmaller quantities in the heart.'3' 4,17 For this reason,characterization of antibodies 335 -356 was carried outin the lung first. In total extracts of rat lung probedwith antibodies 335-356, a major protein with M, of40 000 was specifically detected (Fig 1). A protein withthe same electrophoretic mobility was also detected,specifically, in rat and guinea pig hearts (Fig 2).Immunofluorescence experiments showed that theepitopes recognized by antibodies 335-356 are presentonly in certain regions of the guinea pig and rat heart(eg, the guinea pig atria [Fig 4] or subendocardialregions of rat heart interventricular septum [Fig 11]).In these regions, the pattern of labeling is similar tothat of intercalated disks, where Cx43 junctional chan-nels have been also demonstrated to be localized (Figs4 and 9 and Reference 32). Finally, immunoelectronmicroscopic analysis of guinea pig atrial tissue showedthat antibodies 335 -356 recognized myocyte-associ-ated gap junctions (Fig 5). These data demonstratethat the site-directed antibodies we purified are spe-cific for Cx40. For the first time, the M, of endogenousCx40 is reported. This M, is similar to that deducedfrom immunoprecipitation experiments carried outeither with Xenopus oocytes injected with rat Cx40cRNA39 or with cells transfected with cDNA codingfor Cx40.41

Besides certain myocytes, Cx40 was found to be alsoassociated to blood vessel walls. The regularly spacedlabeling seen in cross-sectioned vessels (Figs 4 and 12)and its close association with the borders of factorVIII-positive cells51 (Fig 3) suggest that Cx40 might beexpressed in the endothelial cells lining the lumen ofvessels. Endothelial cells are connected in situ throughfunctional gap junctions,5253 and immunofluorescence

FIG 8. Distribution of connexin43 (CX43) in unfixed guinea pigheart. Sections were incubated with rabbit anti-CX43 antibod-ies31 and then with tetraethyl rhodamine isothiocyanate-labeledsecondary antibodies. This micrograph illustrates the distribu-tion of CX43 in the anterior region of the left branch of the Hisbundle. Many small immunoreactive sites are evenly distributedin this part of the conduction system. In contrast, in the myocytesof the interventricular septum (and also in the myocytes of theventricle lateral walls), CX43 is concentrated in the intercalateddisks. Distribution pattern of CX43 in the bundle branch isillustrated at higher magnification in insert. In sections adjacentto that displayed in this micrograph and incubated with rabbitantibodies 355-356, no labeling was seen in the bundlebranches although immunoreactivity was detected in bloodvessel walls. IVS indicates interventricular septum. Bar-50 gm(insert, bar-i10,in).




FIG 9. Distribution of connexin40 (CX40) and connexin43 (CX43)in guinea pig atrial myocytes (unfixed organ). Double-labelingexperiment was analyzed by confocal laser scanning microscopy.Sections were incubated successively with rabbit anti-CX40 anti-bodies (antibodies 335-356), then with hen anti-CX43 antibodies.These primary antibodies were revealed with tetraethyl rhodamineisothiocyanate-labeled goat anti-rabbit antibodies and fluoresceinisothiocyanate-labeled goat anti-hen antibodies, respectively. Dis-tribution of CX43 and CX40 is illustrated by micrographs a and b,respectively. Micrograph c is a merged picture of the fluorescencesignals shown in a and b. Examination of the micrographs showsthat CX43 and CX40 are colocalized in most of the immunoreac-five sites. However, in a few sites, only one or the other of theconnexins seems to be expressed. The two arrows in a and cindicate immunoreactive sites where CX43 only is detected. Thetwo arrowheads in b and c indicate sites where CX40 only isdetected. Bar=20 pm.- - - - - - - - - - - - - - - - - - - - - - - - - - - ^ v v v v - - - - - - - -




CX43

A 4 7'

.1

44d'

:'1g ..

lvs

0

/0t

Rat heart

b

CX40

t, ''8 a cFIn 10. Distribution of myocyte-associated connexin40 (CX40) in paraformaldehyde-fixed rat heart with immunoperoxidase labeling. a,b, and c, Micrographs correspond to three serial sections incubated with a mouse monoclonal anti-desmin antibody, rabbitanti-connexin43 (CX43) antibodies, and rabbit anti-CX40 antibodies, respectively. The section displayed in panel a shows desminreactivity in the interatrial (IAS) and interventricular (IVS) septa. Differential labeling of desmin allows tracing of the His bundle, at the topof the interventricular septum (boxed area), and the right branch of the bundle, which is only partially visible in this section. Panels b andc are enlargements of the region boxed in panel a in two sections adjacent to the section shown in panel a. As previously reported,32CX43 is not detected in the His bundle (b), but the myocytes of this same bundle do express CX40 (c). AVC indicates atrioventricularcanal; RA, right atrium. Bar= 100 gm in panel a; bar=20 gm in panels b and c.

experiments have already led authors39 to conclude thatendothelial cells of rat heart blood vessels express Cx40.The presence of Cx40 mRNA has been also demon-strated in primary cultures of bovine aortic endothelialcells.17 Thus, results from various sources converge tosuggest that Cx40 might be expressed in vascular endo-thelial cells, at least in the heart and lung. However,conclusive confirmation of the localization of Cx40 inendothelial cell gap junctions by immunoelectron mi-croscopy remains to be provided.

Cx40-Expressing Myocytes Are Localized in GuineaPig Heart Atrial Tissue and Rat HeartConduction System

Expression of Cx40 in myocytes is controversial:Kanter and colleagues15,411)41 have demonstrated thatCx40 is associated with myocyte gap junctions of dogheart ventricles, whereas Bruzzone et a139 have reportedthat rat heart ventricular myocytes do not express Cx40.This is only an apparent contradiction. We demonstratein the present study that guinea pig heart atrial myo-cytes express Cx40 (Fig 4) but that, in contrast, Cx40could not be detected in ventricular myocytes, which,like atrial myocytes, do express Cx43 (Figs 4 and 8). Wealso show that in guinea pig heart the His bundle andthe bundle branches abundantly express Cx43 (Fig 8),as in human and bovine heart,35 but do not express Cx40(Table). A quite different situation is apparent in ratheart (Table). In this species, myocytes expressing Cx40are localized only in a few regions of ventricular tissue.The topography of these well-defined regions (Figs 10and 11) matches the conduction system of the rat heart

(including the Purkinje fibers), as described in detail bylight and electronic microscopy.5t054,55 In particular,Purkinje fibers of type I and type II have been describedin false tendons and in subendocardial regions of ratventricular walls.5(1 Expression of Cx40 in these cells isin agreement with the recent results reported for dogheart.4041 In contrast Cx40 was not detected in ratcontractile ventricular myocytes, whereas it was de-tected in the dog.4tt,41 Absence of Cx40 mRNA in thesemyocytes, as shown by in situ hybridization (M. vanKempen, D. Paul, D. Gros, T. Vermeulen, A. Moor-man, W. Lamers, unpublished data), strengthens ourresults. As previously reported,32 Cx43 was not detectedeither in the His bundle or in the proximal regions of thebundle branches of rat heart, and consequently Cx40 isthe first connexin whose expression has been demon-strated in this part of the rat conduction system.Cx42 is the bird analogue of mammalian Cx40, and

there is 70% amino acid homology between these twoconnexins.56 Recent studies57 have suggested that thedistribution of Cx42 in avian heart is similar to that ofCx40 in rat heart. Despite this similarity, one shouldnot forget that Cx40 is differently distributed in ratand guinea pig hearts. Thus, the distribution of thisconnexin in the myocardium of one species is notnecessarily a basis for drawing conclusions about itsdistribution in the myocardium of another species.This also applies to the expression of Cx43 in theproximal regions of the mammalian heart conductionsystem (His bundle and its branches), as shown in thepresent study and others.32,35

L.g.,,3t

, ^




FIG 11. Distribution of myocyte-associated con-nexin40 (CX40) in unfixed rat heart as detected byimmunofluorescence. The longitudinal section dis-played in panel a (fluorescence microscopy) andpanel b (phase-contrast microscopy) was incu-bated with rabbit anti-CX40 antibodies and thenwith tetraethyl rhodamine isothiocyanate-labeledgoat anti-rabbit IgGs. The region of the interven-tricular septum (IVS) shown in this figure is local-ized in the right distal part of the septum. The longimmunoreactive region running down along theseptum corresponds to the localization of the rightbranch of the His bundle. Regular distribution ofCX40 in this tissue obviously matches that ofintercalated disks. Note that in the interventricularseptum only the subendocardial region is labeled.Cross-sectioned false tendons50 (stars) are par-tially visible in the right ventricular cavity (RVC).Purkinje fibers, which compose this tissue,50 ex-press CX40. Bar=50 gm.

Colocalization of Cx40 and Cx43 in

Immunoreactive Sites

Hen antibodies to residues 314 to 322 of rat Cx43 were

affinity-purified and characterized by immunoblotting andimmunofluorescence. These antibodies specifically recog-

nize a single protein of the same electrophoretic mobilityas Cx43 (Fig 6) localized in intercalated disks (Figs 7, 9,and 12). These results provide reasonable evidence for thespecificity of these antibodies for Cx43.Hen anti-Cx43 antibodies and rabbit anti-Cx40 anti-

bodies were used to detect simultaneously Cx43 andCx40 (double-labeling experiments) in guinea pig and ratheart sections. Analyses by CLS microscopy were focusedon guinea pig atrial tissue and subendocardial regions ofrat heart lateral ventricular walls. These latter regionscontain cell groups expressing both Cx40 and Cx43 thatlikely correspond to subendocardial Purkinje fibers de-scribed in the rat>511 Only results obtained with guinea pigatrial tissue are illustrated (Fig 9), but in both cases theresults are similar: Cx40 and Cx43 are colocalized inmost of the immunoreactive sites, a finding in agreementwith the colocalization of Cx40 and Cx43 in gap junctionsof dog Purkinje fibers.4' Some sites remain monoreactive(Fig 9), suggesting that they only contain Cx40 or Cx43.However, the major connexin type could be "contaminat-

ed" by a very few molecules of the other type, undetect-able by CLS microscopy in spite of the sensitivity of thistechnique. Colocalization of two connexins as describedabove is not unknown (both liver connexin32 and con-nexin26 were identified in the same gap junctions58), butit raises the question of their interactions. There is nojunctional communication between paired Xenopusoocytes when one of them expresses Cx4O and the otherexpresses C43.39 This inability of Cx43 to interact withCx4O, if it is confirmed in other biological systems,reduces the number of possible arrangements betweenthe two connexins.

Cx40 and the Conduction of Action PotentialsThe presence of both Cx43 and Cx4O in certain

cardiac tissues (which does not preclude the expressionof other types of connexins in these same tissues) shouldoffer additional possibilities for the regulation of con-duction velocity. The unitary conductance of Cx43junctional channels of myocytes has been estimated tobe 40 to 60 picosiemens.5960 A recent investigationcarried out with HeLa cells transfected with Cx40cDNA reported the unitary conductance of Cx40 chan-nels to be -150 picosiemens.61 If it is assumed that thekinetics of both Cx43 and Cx4O channels are approxi-




FIG 12. Distribution of connexin40 (CX40) and connexin43(CX43) in unfixed rat heart ventricular tissue as detected bydouble labeling. Micrographs are as follows: a, phase-contrastmicroscopy; b, fluorescence microscopy (tetraethyl rhodamineisothiocyanate [TRITC] channel); and c, fluorescence micros-copy (fluorescein isothiocyanate [FITC] channel). The sectionshown in panel a was successively incubated with rabbitanti-CX40 antibodies (antibodies 335-356) (b) and then withhen anti-CX43 antibodies (c). These primary antibodies wererevealed with TRITC-labeled goat anti-rabbit lgGs and FITC-labeled goat anti-hen antibodies, respectively. Arrow indicatesa longitudinally sectioned blood vessel. Note in panel b thatthe labeling is exclusively associated with the vascular wall.CX40 is not detected in the contractile myocytes that surroundthe blood vessel. The distribution of gap junctions in thesesame myocytes is shown in panel c by the use of henanti-CX43 antibodies. Insert in panel b displays a cross-sectioned vessel observed in a region very close to that shownin this figure. Bar=50 prm.

mately the same at the small transjunctional voltagedifferences normally found between adjacent heart cellsunder physiological conditions, the amount of currentflowing through Cx40 channels will be greater than thatflowing through Cx43 channels. Consequently, cellsconnected with Cx40 channels should conduct actionpotentials more rapidly than cells connected with Cx43channels. However, analysis of the conduction process,which is a very complex phenomenon, has to take into

Distribution of Connexin43 and Connexin40 in Rat andGuinea Pig Hearts

Rat Heart Guinea Pig Heart

Cx43* Cx40t Cx43t Cx40tAtria + -/+ + +

Ventricles + +

Conduction system

His bundle + +

Bundle branches + +

Purkinje fibers + + ND ND

Cx43 indicates connexin43; Cx40, connexin40; +, presence;absence; and ND, not determined.*Data are from Reference 32.tData are from the present study.

account other factors, such as cell geometry, longitudi-nal resistance, and the number of channels per junction.Clearly, further investigations are needed to analyze theconduction process in molecular terms before a firmstatement can be made about the physiological signifi-cance of the pattern of connexin distribution in cardiacgap junctions.Note added in proof: While this article was in proof,

two articles appeared dealing with the distribution ofCx4O in rat heart. Gourdie et al,62 used antibodiesraised to residues 360 to 379 of chick Cx42 to demon-strate that Cx40 is present in all main components of therat conduction system, whereas it is not detected in theworking ventricular tissue. In contrast, Bastide et al,63using antibodies raised to residues 337 to 358 of mouseCx4O (residues that are conserved in rat Cx40), re-ported the coexpression of Cx4O and Cx43 in workingventricular myocytes, although they noted a preferentialexpression of Cx4O in the ventricular conductive myo-cardium. The discrepancy between these results andours is unknown.

AcknowledgmentsThis study was supported by grants from the EEC (SCl-

CT91-0644) (Drs Gros and Jongsma), the Fondation pour laRecherche Medicale (Dr Gros), the Association Frangaisecontre les Myopathies (Dr Gros), the Institut National de laSante et de la Recherche Medicale (INSERM 92-0409) (DrGros), the Direction de la Recherche et des Etudes Tech-niques (DRET 92-159) (Dr Gros), the Netherlands Organiza-tion for Scientific Research (NWO 900-516-093) (DrJongsma), the Royal Netherlands Academy of Arts and Sci-ences (Dr de Maziere), and the Dutch Heart Foundation(NSH 88-092) (Dr van Kempen). We wish to thank HenrietteAubert for secretarial help and Michel Berthoumieu. LucAnalbers, and Berend de Jonge for expert technical assistance.

References1. Bennett MVL, Barrio LC, Bargiello TA, Spray DC, Hertzberg E,

Saez JC. Gap junctions: new tools, new answers, new questions.Neuron. 1991;6:305-320.

2. Somogyi R, Kolb HA. Biochemical and biophysical analysis ofcell-to-cell channels and regulation of gap junctional permeability.Rev Physiol Biochem PharmacoL 1991;1 18:1-47.

3. Beyer EC, Paul DL, Goodenough DA. Connexin family of gapjunction protein. J Membr BioL 1990;1 16:187-194.

4. Paul DL. Molecular cloning of cDNA for rat liver gap junctionprotein. J Cell Biol 1986;103:123-134.



850 Circulation Research Vol 74, No 5 May 1994

5. Kumar NM, Gilula NB. Cloning and characterization of humanand rat cDNAs coding for a gap junction protein. J Cell Bio.1986;103:767-776.

6. Beyer EC, Paul DL, Goodenough DA. Connexin 43, a protein fromrat heart homologous to gap junction protein from liver. J Cell Biol.1987;195:2621-2629.

7. Zhang JT, Nicholson BJ. Sequence and tissue distribution of asecond protein of hepatic gap junctions, CX26, as deduced from itscDNA. J Cell Biol. 1989;109:3391-3401.

8. Willecke K, Heynkes R, Dahl E, Stutemkemper R, Henneman H,Jungbluth S, Suchyna T, Nicholson BJ. Mouse connexin 37: cloningand functional expression of a gap junction gene highly expressedin lung. J Cell Biol. 1991;114:1049-1057.

9. Nishi M, Kumar NM, Gilula NB. Developmental regulation of gapjunction gene expression during mouse embryonic development.Dev Biol. 1991;146:117-130.

10. Hoh JH, John S, Revel JP. Molecular cloning and characterizationof a new member of the gap junction gene family, connexin 31.J Biol Chem. 1991;266:6524-6531.

11. Paul DL, Ebihara L, Takemoto LI, Swenson KI, Goodenough DA.Connexin 46, a novel lens gap junction protein, induces voltage-gatedcurrents in nonjunctional plasma membrane of Xenopus oocytes.J Cell Biol. 1991;115:1077-1089.

12. Hennemann H, Schwarz HJ, Willecke K. Characterization of gapjunction genes expressed in F9 embryonic carcinoma cells:molecular cloning of mouse connexin 31 and -45 cDNAs. EurJ CellBiol. 1992;57:51-58.

13. Hennemann H, Suchyna T, Lichtenberg-Fratd H, Jungbluth S,Dahl E, Schwarz HJ, Nicholson BJ, Willecke K. Molecular cloningand functional expression of mouse connexin 40, a second gapjunction gene preferentially expressed in lung. J Cell Biol. 1992;117:1299-1310.

14. Haefliger JA, Bruzzone R, Jenkins NA, Gilbert DJ, Copeland NG,Paul DL. Four novel members of the connexin family of gapjunction proteins: molecular cloning, expression and chromosomemapping. J Biol Chem. 1992;267:2057-2064.

15. Kanter LH, Saffitz JE, Beyer EC. Cardiac myocytes expressmultiple gap junction proteins. Circ Res. 1992;70:438-444.

16. White T, Bruzzone R, Goodenough DA, Paul DA. Mouse CX50, afunctional member of the connexin family of the gap junctionproteins, is the lens fiber protein MP70. Mol Biol Cell. 1992;3:711-720.

17. Beyer EC, Reed KE, Westphale EM, Kanter HL, Larson DM.Molecular cloning and expression of rat connexin 40, a gapjunction protein expressed in vascular smooth muscle. J MembrBiol. 1992;127:69-76.

18. Hennemann H, Dahl E, White JB, Schwarz HJ, Lalley PA, Chang S,Nicholson BJ, Willecke K. Two gap junction genes, connexin 31.1 and30.3, are closely linked on mouse chromosome 4 and preferentiallyexpressed in skin. J Biol Chem. 1992;267:17225-17233.

19. Zimmer D, Green CR, Evans WH, Gilula NB. Topological analysisof the major protein in isolated intact rat liver gap junctions andgap-derived single membrane structures. J Biol Chem. 1987;262:7751-7763.

20. Milks LC, Kumar NM, Houghten R, Unwin N, Gilula NB.Topology of the 32-kd liver gap junction protein determined bysite-directed antibody localizations. EMBO J. 1988;7:2967-2975.

21. Goodenough DA, Paul DL, Jesaitis L. Topological distribution oftwo connexin 32 antigenic sites in intact and split rodent hepa-tocyte gap junctions. J Cell Biol. 1988;107:1817-1824.

22. Hertzberg EL, Disher RM, Tiller AA, Zhou Y, Cook RG.Topology of the Mr 27,000 liver gap junction protein. J Biol Chem.1988;263:19105-19111.

23. Yancey BJ, John SA, Lal R, Austin BJ, Revel JP. The 43-kDpolypeptide of heart gap junctions: immunolocalization, topologyand functional domains. J Cell Biol. 1989;108:2241-2254.

24. Laird DW, Revel JP. Biochemical and immunological analysis of thearrangement of connexin 43 in rat heart gap junction membrane.J Cell Sci. 1990;97:109-117.

25. Yeager M, Gilula NB. Membrane topology and quaternary struc-ture of cardiac gap junction ion channels. J Mol Biol. 1992;223:929-948.

26. Rahman S, Evans WH. Topology of connexin 32 in rat liver gapjunctions: evidence for an intramolecular disulfide linkage con-necting the two extracellular peptide loops. J Cell Sci. 1991;100:567-578.

27. El Aoumari A, Dupont E, Fromaget C, Jarry T, Briand JP,Kreitman B, Gros D. Immunolocalization of an extracellulardomain of connexin 43 in rat heart gap junctions. Eur JCell Biol.1991;56:391-400.

28. Jongsma HJ, Gros D. The cardiac connection. News Physiol Sci.1991;6:34-40.

29. Dupont E, El Aoumari A, Fromaget C, Briand JP, Gros D. Immu-nological characterization of rat cardiac gap junctions: presence ofcommon antigenic determinant in heart of other vertebrate speciesand in various organs. J Membr Bio. 1988;104:119-128.

30. Beyer EC, Kistler J, Paul DL, Goodenough DA. Antisera directedagainst connexin 43 peptides react with a 43-kD protein localizedto gap junctions in myocardium and other tissues. J Cell Biol.1989;108:595-605.

31. El Aoumari A, Fromaget C, Dupont E, Reggio H, Durbec P,Briand JP, Boller K, Kreitman B, Gros D. Conservation of acytoplasmic carboxyl terminal domain of CX43, a gap junctionalprotein, in mammal heart and brain. J Membr Biol. 1990;115:229-240.

32. Van Kempen MA, Fromaget C, Gros D, Moorman AFM, LamersWH. Spatial distribution of connexin 43, the major cardiac gapjunction protein, in the developing and adult rat heart. Circ Res.1991;68:1638-1651.

33. Gourdie RG, Green CR, Severs NJ, Thompson RP. Immunola-beling patterns of gap junction connexins in the developing andmature rat heart. Anat Embryol. 1992;185:363-378.

34. Oosthoek PW, Viragh S, Mayen AEM, Van Kempen MJA, LamersWA, Moorman AFM. Immunohistochemical delineation of the con-duction system, I: the sinoatrial node. Circ Res. 1993;73:473-481.

35. Oosthoek PW, Viragh S, Lamers WH, Moorman AFM. Immuno-histochemical delineation of the conduction system, II: the atrio-ventricular node and Purkinje fibers. Circ Res. 1993;73:482-491.

36. Masson-P6vet M, Bleeker WK, Gros D. The plasma membrane ofleading pacemaker cells in the rabbit sinus node: a qualitative andquantitative ultrastructural analysis. Circ Res. 1979;45:621-629.

37. Suji Y, Hirakow R. Freeze-fracture studies of the sino-atrial andatrioventricular nodes of the caprine heart, with special referenceto the nexus. Cell Tissue Res. 1986;245:273-279.

38. Nag AC. Study of non-muscle cells of the adult mammalian heart:a fine structural analysis and distribution. Cytobios. 1980;28:41-61.

39. Bruzzone R, Haefliger JA, Gimlich RL, Paul DL. Connexin 40, acomponent of gap junctions in vascular endothelium, is restricted inits ability to interact with other connexins. Mol Biol CelL 1993;4:7-20.

40. Kanter HL, Laing JG, Beau SL, Beyer EC, Saffitz JE. Distinctpatterns of connexin expression in canine Purkinje fibers and ven-tricular muscle. Circ Res. 1993;72:1124-1131.

41. Kanter HL, Laing JG, Beyer EC, Green KG, Saffitz JE. Multipleconnexins colocalize in canine ventricular myocyte gap junctions.Circ Res. 1993;73:344-350.

42. Merrifield R. Solid-phase peptide synthesis, 1: the synthesis oftetrapeptide. J Am Chem Soc. 1963;85:2149-2154.

43. Briand JP, Muller S, Van Regenmortel MHV. Synthetic peptidesas antigens: pitfalls of conjugation methods. J Immunol Methods.1986;78:59-69.

44. Giaume C, Fromaget C, El Aoumari A, Cordier J, Glowinski J,Gros D. Gap junctions in cultured astrocytes: single-channelcurrents and characterization of channel-forming protein. Neuron.1991;6:133-143.

45. Polson A, von Wechmar BM, Van Regenmortel MHV. Isolation ofviral IgY antibodies from yolks of immunized hens. ImmunolCommun. 1980;9:475-493.

46. McLean IW, Nakane IK. Periodate-lysine paraformaldehyde fix-ative: a new fixative for immunoelectron microscopy. J HistochemCytochem. 1974;22:1077-1083.

47. Sternberger LA. Immunohistochemistry. New York, NY: JohnWiley & Sons Inc; 1986.

48. Thornell LE, Johnsson B, Ericksson A, Virtanen I. Intermediatefilaments and associated proteins in the human heart: an immun-ofluorescente study of normal and pathological hearts. Eur Heart J.1984;5(suppl F):F231-F241.

49. Altman LG, Schneider BG, Papermaster DS. Rapid embedding oftissues in Lowicryl K4M for immunoelectron microscopy. JHistochemCytochem. 1984;32:1217-1223.

50. Cantin M, Thibault G, Haile-Meskel H, Ding J, Milne RW, BallakM, Charbonneau C, Nemer M, Drouin J, Garcia R, Genet J. Atrialnatriuretic factor in the impulse conduction system of rat cardiacventricles. Cell Tissue Res. 1989;256:309-325.

51. Handin RI, Wagner DD. The molecular and cellular biology ofvonWillebrand factor. Prog Hemost Thromb. 1989;9:233-259.

52. Larson DM. Intercellular junctions and junctional transfer in theblood vessel wall. In: Ryan US, ed. Endothelial Cells. Boca Raton,Fla: CRC Press; 1988;3:75-88.

53. Beiy JL, Gribi F. Dye and electrical coupling of endothelial cellsin situ. Tissue Cell. 1989;21:797- 802.




54. Viragh S, Porte A. Structure fine du tissu vecteur dans le coeur derat. Z Zellforsch. 1961;55:263-281.

55. Viragh S, Challice CE. The impulse generation and conductionsystem of the rat. In: Challice CE, Viragh S, eds. Ultrastructure oftheMammalian Heart. New York, NY: Academic Press Inc; 1973:43-89.

56. Beyer EC. Molecular cloning and developmental expression oftwo chick embryo gap junction proteins. J Biol Chem. 1990;265:14439-14443.

57. Gourdie RG, Green C, Severs NJ, Anderson RH, Thompson RP.Evidence for a distinct gap-junctional phenotype in ventricularconduction tissues of the developing and mature avian heart. CircRes. 1993;72:278-279.

58. Nicholson BJ, Dermietzel R, Teplow D, Traub 0, Willecke K,Revel JP. Two homologous protein components of hepatic gapjunctions. Nature. 1987;329:732-734.

59. Rook MB, Jongsma HJ, van Ginneken ACG. Properties of singlegap junctional channels between isolated neonatal rat heart cells.Am J Physiol. 1988;255:H770-H782.

60. Burt JM, Spray DC. Single-channel events and gating behavior ofthe cardiac gap junction channel. Proc NatlAcad Sci USA. 1988;85:3431-3434.

61. Hfilser DF, Eckert R, Zempel G, Rottler R. Biophysical andbiochemical characterization of gap junction channels in culturedmammalian cells. In: Communication at the Worshop on CellularNetworks; September 1992; Palanga, Lithuania.

62. Gourdie RG, Severs NJ, Green CR, Rothery S, Germroth P,Thompson RP. The spatial distribution and relative abundance ofgap-junctional connexin40 and connexin43 correlate to functionalproperties of components of the cardiac atrioventricular con-duction system. J Cell Sci. 1993;105:985-991.

63. Bastide B, Neysse L, Ganten D, Paul M, Willecke K, Traub 0.Gap junction protein connexin40 is preferentially expressed invascular endothelium and conductive bundles of rat myocardiumand is increased under hypertensive conditions. Circ Res.1993;73:1138-1149.



restricted distribution of connexin40, a gap junctional protein, in mammalian heart

Documents