Dystrophin associates with caveolae of rat cardiac myocytes: relationship to dystroglycan.

The possibility of an interaction between the cytoskeletal protein dystrophin and cell surface caveolae in the mammalian myocardium was investigated by several techniques. Caveolin (cav)-3-enriched, detergent-insoluble membranes isolated from purified ventricular sarcolemma by density-gradient fractionation were found to contain dystrophin and dystroglycan. Further purification of cav-3-containing membranes by immunoprecipitation using anti-cav-3-coated magnetic beads yielded dystrophin but not always dystroglycan. Electron microscopic analysis of precipitated material revealed caveola-sized vesicular profiles that could be double-labeled with anti-dystrophin and anti-cav-3 antibodies. In contrast, immunoprecipitation of membranes with anti-dystrophin-coated beads yielded both cav-3 and dystroglycan. Electron microscopic analysis of this material showed heterogeneous membrane profiles, some of which could be decorated with anti-cav-3 antibodies. To confirm that dystrophin and cav-3 were closely associated in cardiac myocytes, we verified that dystrophin was also present in immunoprecipitated cav-3-containing membranes from detergent extracts, as well as in sonicated extracts of purified ventricular myocytes. Confocal immunofluorescence microscopy of ventricular and atrial cardiac myocytes showed that the cellular distributions of cav-3 and dystrophin partially overlapped. Immuno-electron micrographs of thin sections of rat atrial myocytes revealed a fraction of dystrophin molecules that are in apparently close apposition to caveolae. These results suggest that a subpopulation of dystrophin molecules interacts with cardiac myocyte caveolae in vivo and that some of the dystrophin is engaged in linking cav-3 with the dystroglycan complex.

Water channel proteins in rat cardiac myocyte caveolae: osmolarity-dependent reversible internalization.

We show by confocal immunofluorescence microscopy that the water channel protein aquaporin-1, not previously identified within cardiomyocytes, localizes at 20 and 37 degrees C to rat cardiomyocyte sarcolemmal caveolar membrane and subsarcolemmal cytoplasm of primary atrial myocyte cultures, dissociated atrial and ventricular myocytes, and in situ cardiomyocytes of atrial and ventricular frozen sections. Confocal immunofluorescence microscopy shows that the normal in situ colocalization of the quasi-muscle-specific caveolar coating protein caveolin-3 with aquaporin-1 is reversibly disrupted by exposing in situ atrial or ventricular myocytes to physiological saline made hypertonic by adding 150 mM sucrose or 75 mM NaCl to isotonic physiological saline. This causes caveolae to close off from the interstitium and swell, while aquaporin-1 is internalized reversibly. At 4 degrees C aquaporin-1 does not colocalize with caveolin-3. We suggest that 1) in vivo, under near-isotonic conditions, caveolae may alternate frequently between brief open and closed-off states; 2) aquaporin-1-caveolin-3 colocalization may be energy dependent; and 3) while closed off from the interstitium, each caveola transiently functions as an osmometer that experiences, monitors, and reacts to net water flow from or into the subcaveolar cytosol of the myocyte.

T-cadherin is a major glycophosphoinositol-anchored protein associated with noncaveolar detergent-insoluble domains of the cardiac sarcolemma.

Sucrose-density flotation analysis of Triton-insoluble membrane domains isolated from highly purified sheep ventricular sarcolemma revealed the presence of two major 120- and 100-kDa proteins. Both species migrated in two-dimensional isoelectric focussing/SDS gels with an apparent pI of approximately 4.3, suggesting that they might be related. Microsequence analysis of peptides derived from the 100-kDa protein yielded amino acid sequences with high homology to T-cadherin, a truncated cadherin lacking a cytoplasmic domain. The similarity was confirmed using antibodies to chicken T-cadherin that reacted with both proteins on immunoblots. T-cadherin was released from the detergent-insoluble sarcolemmal fraction by phospholipase C treatment indicating that it is linked to the membrane by a glycophosphoinositol anchor. T-cadherin could be ADP-ribosylated by a transferase that was also present in the caveolin-enriched Triton-insoluble fraction. T-cadherin-containing membrane fragments cofractionated on sucrose gradients with caveolin-3, a marker protein for myocyte caveolae. However, immunopurified caveolin-3-containing membranes contained no associated T-cadherin. Immunocytochemical analysis of cultured rat atrial myocytes revealed that T-cadherin and caveolin have related but nonoverlapping staining patterns. These results suggest that T-cadherin is a major glycophosphoinositol-linked protein in cardiac myocytes and that it may be located in plasma membrane "rafts" distinct from but possibly adjacent to caveolae.

Type B atrial natriuretic peptide receptor in cardiac myocyte caveolae.

We have previously shown that atrial natriuretic peptide (ANP) is present in caveolae of in situ rat atrial myocytes. To investigate whether intracaveolar ANP of rat atrial myocytes exists within caveolae bound to type B ANP receptors (ANP-RB, a guanylyl cyclase), we have used confocal immunofluorescence microscopy applied to primary cultures of atrial myocytes from adult rats and to freshly dissociated rat atrial myocytes (not cultured). These experimental designs tested whether atrial myocyte ANP-RB colocalizes at the plasmalemma and elsewhere in the cell with the muscle-specific isoform of the caveolar coating protein caveolin-3, and with a fraction of cellular ANP. The experiments showed that cellular caveolin-3, a fraction of cellular ANP-RB, and a fraction of cellular ANP colocalize at the plasmalemma of cultured atrial myocytes and of freshly dissociated atrial myocytes. The observations support the hypothesis that in rat atrial myocytes, intracaveolar ANP is bound to ANP-RB, a protein whose cytosolic amino acid sequences are known to encode guanylyl cyclase activity. We suggest that among the (probably multiple) effects of the cGMP thus generated in the cytoplasmic microdomain underlying atrial myocyte caveolae may be the activation of cGMP-dependent protein kinase, which would thereby inhibit plasma membrane Ca2+ channel activity and contribute to a negative inotropic effect of ANP.

Localization of atrial natriuretic peptide in caveolae of in situ atrial myocytes.

The plasma membrane-associated non-clathrin-coated vesicles called caveolae are multifunctional organelles thought to be implicated in the sequestration and transport of small molecules (potocytosis) as well as in the binding of Ca2+ ions, signal transduction, and processing of hormonal and mechanosensitive signals. We have previously suggested that the apparent contiguity of caveolar and atrial granule membranes observed in electron micrographs of in situ mouse atrial myocytes might reflect externalization of atrial natriuretic peptide through caveolae. Using Tokuyasu's classic technique, we now show by immunoelectron microscopy of glutaraldehyde-fixed and cryosectioned mouse and rat atria that antibody against atrial natriuretic peptide prohormone is present within caveolae of in situ atrial myocytes. We confirm this intracaveolar localization by stereoimaging colloidal gold-labeled antibody to the prohormone in electron micrographs of glutaraldehyde/osmium tetroxide-fixed positively stained atrial thin sections. Because profiles of caveolae were rarely immunolabeled with antibody against atrial peptide unless there was a profile of an immunolabeled atrial granule nearby in the subjacent cytoplasm, we concluded that the intracaveolar hormone was derived predominantly from a direct interaction of atrial granules with caveolae. Perturbations that markedly increase the rate of natriuretic peptide secretion via the regulated pathway, including atrial stretch, contractions, and increased external Ca2+ concentration, failed to alter caveolar immunostaining. These results suggest that atrial peptide can pass from atrial granules into caveolae by transiently open pathways between the interiors of granules and caveolae. The results are interpreted as suggesting the presence of a second pathway for externalization of atrial natriuretic peptide through caveolae in addition to the classic pathway for regulated atrial peptide secretion at noncaveolar plasmalemma.

Endocytosis and uptake of lucifer yellow by cultured atrial myocytes and isolated intact atria from adult rats. Regulation and subcellular localization.

The time course of endocytic uptake of Lucifer yellow (LY) was followed by fluorescence and electron microscopy after exposure of primary cultures of atrial myocytes from adult rats to LY under conditions that prevented transplasmalemmal LY entry via channels or carriers. After a 2-minute exposure to LY at 37 degrees C, electron microscopy revealed classic clathrin-coated vesicles fused to endosomes, which were absent in LY-free medium or at 2 degrees C, suggesting that LY turns on endocytosis or accelerates a slow constitutive endocytosis. Fluorescence microscopy, which detected no LY entry at 2 minutes in LY, showed punctate cytoplasmic fluorescent densities after 10 minutes, which were readily distinguishable from intrinsic perinuclear fluorescence. Fluorescence microscopy after immunostaining with antibodies against clathrin, vacuolar H(+)-ATPase, atrial peptide, or a marker for acidified compartments suggested LY sorting into an acidified prelysosomal pathway. Using absence of punctate fluorescence after 10 minutes in LY as a criterion for inhibition of endocytosis, we showed that endocytosis was inhibited by inhibitors of protein phosphatases 1 and 2A or inhibitors of cAMP-dependent protein kinases 1 and 2, by effects of caffeine on sarcoplasmic reticulum Ca2+ release, and by temperatures below 18 degrees C, but not by staurosporine, phorbol esters, pertussis toxin, thapsigargin, preventing contractions with nifedipine, ryanodine and low [Ca2+]o, or raising cytosolic cAMP concentrations. Both phosphatase inhibitors and caffeine also inhibited a fraction of LY uptake by intact rat atria. We conclude that endocytic uptake of LY is an energy-dependent, specifically regulated process, whose understanding and control are potentially important for the medically relevant problem of introducing drugs and macromolecules into atrial heart muscle cells.

Fluid-phase endocytosis by in situ cardiac myocytes of rat atria.

Fluid-phase endocytosis (FPE) associated with recycling of fused plasmalemma-secretory granules or membranes and/or membrane receptors by in situ cardiac myocytes was studied at 37 degrees C in vitro noncontracting adult rat atrial preparations. Measurements included 1) the volume (VS*) of the compartment consisting of presumptive endocytotic vesicles and the endosomes or lysosomes transiently in continuity with them (S*), which internalizes [14C]-sucrose but is inaccessible to simultaneously measured [methoxy-3H]inulin, 2) the kinetics of [14C]sucrose efflux from S*, and 3) morphometry to quantify interstitial space and non-heart muscle cells. Vs* (0.39 +/- 0.04 ml/g dry atrium for unstretched atria at 37 degrees C) was 1) variable over a 3.7-fold range under various experimental conditions, 2) significantly increased by neomycin or by lowering the temperature to 18 degrees C, and 3) significantly decreased by alpha 1-adrenergic stimulation. Analysis of sucrose efflux kinetics confirmed the presence of an intramyocytic sucrose-containing compartment. A smaller inulin-inaccessible sucrose space (S*) was also present in right ventricle. Thus, during FPE, vesicles and endosomes initially containing high (extracellular) Ca2+ and Cl- concentrations continually enter, circulate within, and undergo exocytosis from myocardial cells.

Rat atrial myocyte plasmalemmal caveolae in situ. Reversible experimental increases in caveolar size and in surface density of caveolar necks.

The structure, size, and surface density of the conspicuous flask-shaped structures called caveolae that are located under the plasma membrane of cardiac myocytes in intact rat atria were studied by electron microscopy after physiological perturbations designed to examine whether caveolae and/or their necks are fixed or mobile and whether the caveolar lumen is always open or can close off from the interstitial space. We showed that, in stretched and unstretched atria, horseradish peroxidase could enter or be washed out of caveolae at 37 degrees, 18 degrees, and 4 degrees C, but this finding does not rule out that caveolae and/or their necks can cycle rapidly between states closed and open to the interstitial space. Electron microscopy of thin sections revealed that exposure of atria at 37 degrees or 18 degrees C to physiological salt solutions made hypertonic by adding 150 mM sucrose or mannitol resulted in a striking enlargement of caveolar profiles within 1 to 5 minutes. Caveolar enlargement was rapidly reversible on return to control saline. After freeze fracture of atria exposed to these hypertonic solutions, quantitative analysis of electron micrographs of the fracture faces revealed statistically significant increases in cross-sectional diameter of cross-fractured caveolar necks and in mean number of caveolar necks penetrating per unit area of plasmalemmal fracture face. These results suggest that atrial myocyte caveolae are dynamic structures whose necks may be reversibly inserted into and withdrawn from the plasmalemma, possibly (but not necessarily) corresponding to states in which caveolae are, respectively, open and closed to the interstitial spaces.

Permeability of rat atrial endocardium, epicardium, and myocardium to large molecules. Stretch-dependent effects.

Atrial distension, which stimulates atrial natriuretic peptide secretion by atrial myocytes, also stretches nonmuscle cells. In a noncontracting in vitro preparation of combined right and left atria we demonstrated by electron microscopy that, at 37 degrees C, transition from zero pressure to a physiological distending pressure of 5.1 mm Hg rapidly rendered atrial endocardial endothelium permeable to the macromolecular probes horseradish peroxidase (HRP; M(r), approximately 40,000) and wheat germ agglutinin-HRP (M(r), approximately 70,000); each probe was introduced at the atrial cavitary endocardial surface. Stretch-dependent permeabilization was also demonstrable in spontaneously contracting atria, was reversed by removing the distending pressure, and was unaffected by varying external Ca2+ concentration from 0.2 to 1.4 mM or by experimental perturbations that markedly decrease ANP secretory rates. Although transendocardial HRP and wheat germ agglutinin-HRP passage required stretch, native ferritin (M(r), = 500,000) could traverse unstretched endocardium. Probes were detected in noncoated endocardial vesicles and intercellular junctions between endocardial cells, but the relative contributions of vesicular transcytosis and paracellular diffusion could not be determined. Although HRP entered plasmalemmal caveolae of myocytes in stretched atria, myocytes did not internalize HRP by fluid-phase endocytosis. Distending pressure also caused apparent flow reversal in thebesian blood vessels, with retrograde transfer of HRP across the endocardium into the myocardium. HRP and ferritin presented at the external surface of the epicardium (visceral pericardium) were endocytosed by mesothelial cells, entered junctions between mesothelial cells, and readily crossed the epicardium of both stretched and unstretched preparations.

Inhibition of atrial peptide secretion at different stages of the secretory process: Ca2+ dependence.

We have used a noncontracting in vitro preparation of stretched and unstretched rat atria to estimate contributions of constitutive and regulated pathways to the rates of stretch-augmented and basal secretion of immunoreactive atrial natriuretic peptide (ANP) and to examine effects of inhibition of the secretory sequence by 1) protein synthesis inhibitors, 2) disruption of forward vesicular traffic between endoplasmic reticulum and Golgi with brefeldin A (BFA, and 3) cellular ATP depletion. Protein synthesis inhibition with cycloheximide for 44 min slowed neither basal nor stretch-augmented ANP secretion but instead accelerated stretch-augmented secretion at low (but not at physiological) external Ca2+ concentration, suggesting that the constitutive component does not contribute substantially to either basal or stretch-augmented secretion. BFA, which disassembled Golgi cisternae, increased the stretch-augmented secretory rate via the regulated pathway and prevented Ca(2+)-dependent inactivation with time. Cellular ATP depletion rapidly and completely inhibited stretch-augmented secretion. We conclude that both basal and stretch-augmented utilize the energy-dependent regulated pathway, drawing on a large reservoir of concentrated prohormone stored in granules that is not detectably depleted during 44 min of stretch-augmented secretion at 37 degrees C.

[The study on ultrastructure of gap junction of rat atrial myocytes in primary culture].

We have studied the ultrastructure of gap junction of rat atrial myocytes of primary culture in situ embedded by ultramicrotomy. We observed a kind of gap junction-associated vesicles (GJAV), a kind of cluster of particles which are surrounded by plasma membrane, and a concatenate GJAV complexes (CGJAVC) in some big clusters. We found that both of GJAV and CGJAVC are very near the plasma membrane at the intercellular space, and at the same time they are usually adjacent the assembled GJ. So we infered that they are the pre-body or midproduct of the assembled GJ probably. This article analysed these observes and probed into the process about how to assemble the gap junction.

Effect of external Ca2+ concentration on stretch-augmented natriuretic peptide secretion by rat atria.

An in vitro noncontracting rat atrial preparation stretched at 37 degrees C by a distending pressure of 5.1 mmHg was used to examine effects of external Ca2+ concentration ([Ca2+]out, 0.05-3.0 mM) on secretion of immunoreactive atrial natriuretic peptide (ANP) in presence of saxitoxin (STX) and in presence or absence of ryanodine. Under these conditions, the time course of the amount (y) of ANP secreted per milligram dry atrium during 44 min could be approximated by a rate coefficient (k) according to the relation y = s[1 - e(-kt)], where s is the maximal amount secreted after a long time (t). Although k, the rate coefficient for stretch-augmented secretion, increased significantly as [Ca2+]out was raised, secretion inactivated progressively in a time- and [Ca2+]out-dependent manner. This time-dependent decrease was not prevented by ryanodine. We conclude that a component of ANP secreted by quiescent atria in vitro is positively modulated by [Ca2+]out and does not require ryanodine-sensitive Ca2+ release from sarcoplasmic reticulum. The [Ca2+]out-sensitive processes underlying time-dependent inactivation of secretion remain undetermined.

Basal and stretch-augmented natriuretic peptide secretion by quiescent rat atria.

Absolute rates of immunoreactive atrial natriuretic peptide (ANP) secretion were measured in vitro at 37 degrees C in noncontracting preparations of combined right and left rat atria at constant distending pressures of 0 or 5.1 mmHg in presence of 0.2 mM extracellular Ca2+ concentration [( Ca2+]o), 10 microM ryanodine, and either 1 microM saxitoxin or 10 microM tetrodotoxin. By systematic deletion of external Na+, K+, Mg2+, Cl-, or HCO3-, and reduction of [Ca2+]o, and by selective ion transport inhibitors, neither net transplasmalemmal fluxes of Na+, Cl-, HCO3-, and Mg2+ nor ryanodine-sensitive Ca2+ release from sarcoplasmic reticulum (SR) was found necessary for stretch augmentation of secretory rate (Ra). Ra 1) was near zero within 20 min when extracellular Na+ concentration = 0 and [Ca2+]o = 20 microM and within 5 min or less after preincubation with caffeine, 8-chlorophenylthioadenosine 3',5'-cyclic monophosphate, or at 18 degrees C; 2) was significantly decreased by Cd2+, Ni2+, the isoquinoline H-7, and trifluoperazine but not 100 microM ryanodine; 3) was increased by neomycin; and 4) had an apparent activation energy of 18.5 +/- 4.1 x 10(3) cal/mol between 23 and 42 degrees C. These experiments strongly implicated transplasmalemmal Ca2+ influx and cAMP but not SR Ca2+ release in control of Ra under the experimental conditions studied.

Cardiac gap junctions and gap junction-associated vesicles: ultrastructural comparison of in situ negative staining with conventional positive staining.

By comparing in situ negative staining of mammalian heart muscle using La(NO3)3 with conventional positive staining by uranium and lead salts, we showed that 1) the membrane area of rat cardiac gap junctions (GJs) at the intercalated disks is threefold to fourfold greater than previously thought; 2) connexon arrays of cardiac GJ are subdivided into microdomains by connexon-free aisles; 3) profiles of GJ-associated vesicles (GJAVs) of plasmalemmal origin (which are present extracellularly and sharply localized at three extracellular sites) are paired to form GJs with each other and with myocyte plasmalemma; 4) some GJAVs contain arrays of assembled connexons; and 5) myocytes contain intracytoplasmic complexes lying within cylindrical or cigar-shaped membranes and consisting of GJs and multiple vesicles apparently dissociating from these GJs.

Tunneling cell processes in myocytes of stretched mouse atria.

Serial section electron micrographs of mouse atria stretched in vitro show that myocytes have cell processes which tunnel into adjacent myocytes for 8 microns or more. The tunneling cell processes (TCP) (diam 4-6.2 microns) lack myofibrils and organelles associated with atrial peptide secretion. The glycogen-rich TCP cytoplasmic matrix contains conspicuous tubules and vesicles originating from endoplasmic reticulum and resembling free sarcoplasmic reticulum (SR). TCP are surrounded by a plasmalemma derived from their myocyte of origin, the plasmalemma of the tunneled myocyte, and an intervening narrow compartment continuous with the interstitial space. Profiles having the characteristics cytoplasmic structure of TCP are also found both in the interstitial space between myocytes and near the longitudinal terminations where myocyte ends about on the interstitial space. We suggest that TCP tubules and vesicles may proliferate and/or transport in response to stretch, might be free SR, and may respond to stretch-activated changes in ionic composition or potential of the surrounding myocyte and narrow intercellular compartment.

Human cardiac gap junctions: isolation, ultrastructure, and protein composition.

Recent experiments from our laboratory have shown that the ultrastructure and protein composition of gap junctions isolated from rat ventricles are tissue specific, i.e., markedly different from gap junctions of liver and lens. The differences include a cytoplasmic surface component characteristic for cardiac gap junctions; this component can be visualized by two ultrastructural techniques: as a fuzzy layer in electron micrographs of thin-sectioned junctional pellets and as cytoplasmic surface particles in deep-etched freeze-fractured junctions. The component corresponds to a Mr 17,500 cytoplasmic surface domain of each of the six (Mr 47,000) rat heart gap junctional channel protein subunits that make up the gap junctional channel hexamer known as a connexon. The cytoplasmic surface component is localized at the carboxy-terminal of the subunit. Within the cytoplasmic surface component, rat cardiac gap junctions are cross-linked by disulfide linkages between subunits of the same connexon and between subunits of adjacent connexons. By contrast, the Mr 28,000 liver gap junctional subunit lacks a comparably large cytoplasmic surface component, cytoplasmic surface fuzz, cytoplasmic surface particles, and intra- and interconnexon disulfide linkages. Most of these unique characteristics of cardiac gap junctions were discovered in junctions isolated from rat ventricles. Unlike liver and lens gap junctions, cardiac gap junctions from humans, non-human primates, or other large mammals have not previously been isolated and characterized. Here we report the isolation of unproteolyzed gap junctions from the ventricle of a 24 year-old man with advanced cardiomyopathy whose heart was removed for replacement by a transplanted heart.(ABSTRACT TRUNCATED AT 250 WORDS)

Ultrastructural features of atrial peptide secretion.

The ultrastructural bases of exocytotic extrusion of atrial peptides were reexamined in electron micrographs of thin-sectioned or freeze-fractured mouse atria. Exocytotic extrusion was demonstrated in both thin-sectioned and freeze-fractured atrial myocytes. Ultrastructural evidence suggested that the necks of plasmalemmal caveolae may constitute preformed pathways for extrusion of secretion from granules fusing with caveolae. Multiple subsarcolemmal foci with peripheral Golgi cisterns and accumulations of granules, indicating peripheral processing of secretory proteins, were striking features of mouse and rat atria, and were present but rare in sheep and dog atria. Conspicuous focal ellipsoidal deposits, a new structure, approximately 1.6-4.6 micron long and approximately 0.8-1.8 micron wide, consisting of amorphous cytoplasmic material that is penetrated peripherally by tubules connecting with secretion-containing "multivesicular bodies," were present in some mouse atrial myocytes, but were absent in myocytes of mouse ventricle and rat, dog, and sheep atria.

Developmental changes in internal structure of chick heart plasma membrane.

Although changes in electrophysiologically measurable membrane properties of chick embryo cardiac plasma membrane have been repeatedly documented during embryonic development, ultrastructural techniques were heretofore too insensitive to detect developmental changes in internal structure of this membrane. We report here significant structural changes detected by applying the quantitative analysis of Kordylewski, Karrison, and Page (Amer. J. Physiol. 245, H992-H997, 1983 and 248, H297-H304, 1985) to stereo imaged electron microscopic negatives of glutaraldehyde-fixed chick embryo hearts, freeze fractured and photographed with a goniometer stage. Between Hamburger-Hamilton stages 12+ (about 48 hr incubation) and 24 (about 96 hr incubation), plasmalemmal P-face particle density of ventricular myocytes increased from 2228 +/- 139 to 3063 +/- 109 (P less than 0.01); thereafter, measurements at stages 30, 37, 40, and 45 (7, 11, 15, and 19 days incubation) showed a slower significant linear increase which gave a least-squares line with a slope of 41 +/- 13 particles/day (P less than 0.01). Just before hatching, (stage 45) the value of 3762 +/- 234 was similar to, though slightly smaller than, the values of 4122 +/- 153 (8 days after hatching) and 4281 +/- 218 (adult chicken). These results indicate striking stage-dependent changes in the population of integral membrane proteins (channels, carriers, receptors, etc.), especially marked during early embryogenesis.

Proteolysis of cardiac gap junctions during their isolation from rat hearts.

Gap junctions (GJ) isolated from rat hearts in presence of the protease inhibitor phenylmethylsulfonylfluoride (PMSF) contain a Mr 44,000 to 47,000 major polypeptide and have a urea-resistant layer of fuzz on their cytoplasmic surfaces, whereas junctions isolated without PMSF are proteolyzed to a Mr 29,500 polypeptide by a serine protease and have smooth cytoplasmic surfaces (C.K. Manjunath, G.E. Goings & E. Page Am. J. Physiol. 246:H865-H875, 1984). Rat liver GJ isolated with or without PMSF contain a Mr 28,000 polypeptide and have smooth cytoplasmic surfaces. Here we examine the origin, type and inhibitor sensitivity of the heart protease; why similar proteolysis is absent during isolation of rat liver gap junctions; and whether the Mr 44,000 to 47,000 cardiac GJ polypeptide is the precursor of the Mr 29,500 subunit. We show that the Mr 44,000 to 47,000 polypeptide corresponds to the unproteolyzed connexon subunit; that proteolysis of this polypeptide occurs predominantly during exposure to high ionic strength solution (0.6 M KI) which releases serine protease from mast cell granules; that this protease is inhibitable with PMSF and (less completely) soybean trypsin inhibitor and chymostatin; and that in vivo degranulation of mast cells by injecting rats with compound 48/80 fails to prevent breakdown of cardiac GJ during isolation. The results support the concept that GJ from rat heart and liver differ in protein composition.

Cytoplasmic surface and intramembrane components of rat heart gap junctional proteins.

Gap junctions were purified from rat hearts in the presence of absence of proteolysis inhibitors and examined by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and electron microscopy of thin sections. In absence of proteolysis inhibitors or in presence of ethylenediaminetetraacetic acid or leupeptin, gap junctions contained a single major protein band at relative molecular weight (Mr) 29,500 and minor bands at Mr 44,000-47,000, 17,750, and 16,500 and showed smooth cytoplasmic surfaces in electron micrographs. SDS-PAGE of junctions prepared with phenylmethylsulfonylfluoride (PMSF) showed markedly decreased intensity of the Mr 29,500 band and increased intensity of bands at Mr 44,000, 45,500, and 47,000; electron microscopy of these gap junctions showed presence of a fuzzy layer on their cytoplasmic surfaces. Urea (8 M) could not remove this fuzzy layer. In electron micrographs of rat ventricular myocytes, cytoplasmic surfaces of gap junctions were fuzzy. We conclude that rat heart gap junction protein consists of an intramembrane component (Mr 29,500) that extends into the "gap" and a cytoplasmic surface component (Mr 14,500-17,500) that corresponds to the fuzzy layer and is hydrolyzable by a serine protease.