OUP user menu

Proteolysis of connexin43-containing gap junctions in normal and heat-stressed cardiac myocytes

James G Laing , Peter N Tadros , Karen Green , Jeffrey E Saffitz , Eric C Beyer
DOI: http://dx.doi.org/10.1016/S0008-6363(98)00060-1 711-718 First published online: 1 June 1998

Abstract

Objective: The present studies were performed to examine the degradation of connexin43-containing gap junctions by the lysosome or the proteasome in normal and heat-stressed cultures of neonatal rat ventricular myocytes. Methods: Primary cultures were prepared from neonatal rat ventricular myocytes. Connexin43 was detected by immunoblotting, immunofluorescence, or immunoprecipitation. Gap junction profiles were detected by transmission electron microscopy. Results: Immunoblots of whole cell lysates demonstrated increased levels of connexin43 in cultures treated with lysosomal inhibitors (chloroquine, leupeptin, E-64, or ammonium chloride) or proteasomal inhibitors (lactacystin or ALLN). Pulse-chase experiments showed that the half-life of connexin43 was 1.4 h in control cultures, but was prolonged to 2.0 or 2.8 h in cultures treated with chloroquine or lactacystin, respectively. Immunofluorescence and electron microscopy showed a significant increase in the number of gap junction profiles in myocytes treated with either chloroquine or lactacystin. Heat treatment of cultures (43.5°C for 30 min) produced a rapid loss of connexin43 as detected by immunoblotting or immunofluorescence. Heat-induced connexin43 degradation was prevented by simultaneous treatment with lactacystin, ALLN, or chloroquine. Connexin43 levels and distribution returned to normal by 3 h following a heat shock and were resistant to a subsequent repeat heat stress. The heat shock also led to production of HSP70 as detected by immunoblotting. Conclusions: These data suggest that Cx43 gap junctions in myocytes are degraded by the proteasome and the lysosome, that this proteolysis can be augmented by heat stress, and that inducible factors such as HSP70 may protect against Cx43 degradation.

Keywords
  • Gap junction
  • Connexin43
  • Proteasome
  • Lysosome
  • Heat shock
Abbreviations
  • Cx43, connexin43
  • ALLN, N-acetyl-l-leucyl-l-leucyl-norleucinal
  • NH4Cl, ammonium chloride
  • ALLM, N-acetyl-leucyl-leucyl-methioninal
  • PBS, phosphate-buffered saline

Time for primary review 28 days.

1 Introduction

Cardiac gap junctions contain intercellular channels that allow the passage of ions between myocytes facilitating electrical conduction. Despite the intuitive impression that gap junctions might be quite stable, pulse-chase studies performed in cultured cells suggest that they are quite dynamic, since the half-lives of their subunit proteins (connexins) are between 1 and 3 h [1–3]. In the current study, we have investigated the mechanisms of cardiac gap junction degradation.

Studies in other systems suggest that both the lysosome and the proteasome play a role in gap junction degradation. Electron microscopy studies have suggested that gap junctions are degraded in the lysosome, since annular gap junctions have been found within cells or gap junctions have been identified within clathrin-coated vesicles, phagolysosomes, or multi-vesicular complex structures [4–7]. Our studies (conducted primarily in E36 Chinese hamster ovary cells) have shown that Cx43 is degraded by the ubiquitin–proteasome pathway, since its degradation is inhibited at the restrictive temperature in a temperature-sensitive mutant containing a thermolabile ubiquitin activating enzyme, E1, and since Cx43 accumulates and its half-life is prolonged in cells treated with the proteasome inhibitor N-acetyl-leucyl-leucyl-norleucinal (ALLN) [8]. Further biochemical studies in the myocyte-derived cell line BWEM have implicated both the proteasome and the lysosome in Cx43 gap junction proteolysis [9].

Understanding the turnover and degradation of cardiac gap junctions may be important, since these processes appear to be significant components of the myocardial responses to chronic or acute injuries. The chronic consequences may be more obvious, since extensive remodeling of the number and spatial distribution of myocyte gap junctions is associated with healing after acute myocardial infarction or remodeling in response to pressure or volume overload [10–12]. In regions bordering healed infarcts, remodeling of gap junction distributions may contribute to altered anisotropic conduction and to the development of re-entrant ventricular tachycardia or conduction block [13–15]. Acutely, during ischemia, cellular uncoupling may be induced leading to locally slowed conduction producing discoordination of ventricular contraction and increased risk of ventricular arrhythmias. This uncoupling in ischemic myocardium likely involves altered gap junction channel gating due to ionic imbalances in severely injured myocytes [16], but it may also involve stress-induced protein degradation.

Heat stress modulates the degradation of many proteins [17]and mimics some other acute stresses to the heart including ischemia. We have observed that heat treatment causes Cx43 degradation in E36 Chinese hamster ovary cells [8], and therefore we have speculated that heat shock might be an appropriate model to study stress-induced degradation of myocardial Cx43 gap junctions. Cardiac heat shock is of interest, not only because of the injuries induced, but also because of the adaptive response to a sub-lethal heat stress. The response to heat stress shows substantial similarities to the response to ischemic preconditioning [18], including the induction of stress proteins such as members of the HSP70 family. Expression of HSP70 (due to an initial heat stress or due to exogenous over-expression) can be protective against subsequent ischemic insults [19, 20].

A major objective of the present study was to determine which proteases were involved in the degradation of Cx43 gap junctions in cardiac myocytes, since we had previously studied Cx43 degradation only in rapidly dividing immortalized cells. Our general strategy was to treat cultured neonatal rat ventricular myocytes with a panel of different protease inhibitors and to examine changes in the abundance or the half-life of immunologically detected Cx43 or in the numbers of ultrastructurally defined gap junctions. We have also examined the effect of heat treatment on the abundance and degradation of Cx43 in these cultures. The data presented begin to reconcile our recent observations [8, 9]with prior electron microscopy studies suggesting degradation of cardiac gap junctions within lysosomes [7, 21].

2 Methods

The investigations all conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1985).

2.1 Isolation and culture of neonatal rat ventricular myocytes

Primary cultures of neonatal rat ventricular myocytes were prepared according to the procedure described by Darrow et al. [3]. Cell cultures were incubated at 37°C in a humidified incubator containing 5% CO2. Cultures were maintained in culture medium containing a 1:2 mixture of PC-1 and Ham's F-12/DMEM, by which time the myocytes had formed a synchronously contracting syncytium. At the time of harvest (72 h), we estimated that ∼95% of the living cultured cells were myocytes. Heat treatment was performed by incubating cultures in a 43.5°C water bath for 30 min in normal media containing 10 mM HEPES to maintain pH.

2.2 Protease inhibitors

The lysosomotropic amines chloroquine and ammonium chloride (NH4Cl) were purchased from Sigma (St. Louis, MO). Leupeptin and E-64 were purchased from Boehringer-Mannheim Biochemicals (Indianapolis, IN). N-acetyl-leucyl-norleucinal (ALLN) and N-acetyl-leucyl-leucyl-methioninal (ALLM) were obtained from Sigma. Lactacystin was purchased from Proscript (Boston, MA) or from Dr. E.J. Corey (Harvard University, Boston) or Calbiochem. For protease inhibition studies, cells were treated with 100 μM chloroquine, 10 mM NH4Cl, 25 μM ALLM, 100 μM E-64, 25 μM leupeptin, 25 μM ALLN, or 25 μM lactacystin. ALLN, ALLM, E-64, and leupeptin all can inhibit calpains and cathepsin B; however, at the concentrations used, only ALLN inhibits proteasomal degradation [22]; lactacystin is a highly specific inhibitor of the proteasome [23, 24].

2.3 Antibodies

A rabbit antiserum directed against a synthetic peptide representing amino acids 252–271 in Cx43 was produced previously and has been extensively characterized [8, 25]. This antiserum was used for immunoprecipitation in the pulse-chase experiments. A (His)6–Cx43CT fusion protein was expressed after subcloning DNA encoding amino acids 212–384 of rat Cx43 into the pET3A vector (Novagen) and was used to prepare a second polyclonal anti-Cx43 reagent. The (His)6–Cx43CT fusion protein was isolated on a Ni2+–NTA column. A rabbit antiserum directed against this (His)6–Cx43CT fusion protein containing was produced; it was subsequently affinity purified against the (His)6–Cx43CT fusion protein bound to a Sulfolink coupling gel (Pierce, Rockford, IL) and was used in all of the immunofluorescence and immunoblotting experiments, with results indistinguishable from the other anti-Cx43 antibodies. Anti-HSP70 antibody was purchased from Stressgen (Victoria, BC); this reagent recognizes exclusively the inducible protein HSP70. CY3 and peroxidase-conjugated goat anti-mouse IgG or anti-rabbit IgG antibodies were purchased from Jackson Immunoresearch (West Grove, PA).

2.4 Immunoblots

Myocytes were washed with PBS and harvested by scraping. The cells were concentrated by centrifugation (1330×g, 5 min) and ruptured by sonication (4×15 s). The cellular residue was concentrated by centrifugation (14 000×g, 10 min), and then resuspended in PBS. The protein concentration of these extracts was determined [26], and 25-mg aliquots were separated by SDS–PAGE on 12.5% polyacrylamide gels [27]and blotted onto Immobilon-P (Millipore, Bedford, MA) [28]. Membranes were probed with a monoclonal antibody directed against Cx43 as previously outlined in Laing and Beyer [8]or with polyclonal antibodies against (His)6–Cx43CT at a 1:1000 dilution. In some experiments, membranes were probed with anti-HSP70 antibodies at a 1:1000 dilution.

2.5 Immunofluorescent labeling of cultured cells

Cells were cultured in plastic chamber microscope slides (Nunc, Naperville, IL) and fixed in 50% methanol/50% acetone for 2 min at room temperature and permeabilized in 1% Triton X-100/PBS. Cells were incubated with anti-Cx43 antibodies at a 1:400 dilution overnight at 4°C followed by secondary antibodies (CY3-conjugated goat anti-rabbit IgG) at 1:800 dilution for 1 h at room temperature with intervening washes [8]. The cells were observed by epifluorescence using a Nikon Optiphot microscope and a 40× (NA 1.0) or a 60× (NA 1.4) objective and photographed [29].

2.6 Metabolic labeling of cells and immunoprecipitations

In metabolic labeling experiments and pulse-chase experiments, cells were labeled for 2 h in methionine-depleted normal media containing [35S]methionine (100 μCi/ml) at 37°C [8, 29]. Cells were chased in normal media supplemented with 2 mM methionine in the presence of the noted protease inhibitor. Myocyte cultures cells were processed for immunoprecipitation. Densitometric images were analyzed as outlined in Laing and Beyer [8]and Darrow et al. [3]. The first-order decay constant (k) was calculated from the best-fit single exponential decay curves of the form y=e(−kt) generated with the program SigmaPlot (Jandel Scientific, San Raphael, CA). The half-life of each protein was determined according to the formula t1/2=0.693/k.

2.7 Electron microscopy

Cultured myocytes were fixed and prepared for electron microscopy as previously described [10, 30]. Intercalated disk and gap junction profile lengths were measured in randomly photographed intercellular junctions to determine the relative size and number of gap junctions and intercalated disc length per unit cell area, as previously described [10, 31]. Briefly, individual test regions were first photographed at a final print magnification of 5000×. The total area occupied by cardiac myocytes was measured in each individual test area. All portions of each test area containing intercalated disks and gap junction profiles were photographed again at high magnification for further analysis (final print magnification 25 000×). The lengths of intercalated disk and gap junction profiles were measured in these high-power micrographs by use of an electronic pen and digitizing tablet (Houston Systems). The number of fields analyzed was 12 for controls and 9 and 8 for chloroquine and lactacystin-treated cultures respectively. Morphometric data were analyzed by one-way ANOVA using the SigmaStat software (Jandel).

3 Results

To determine if the lysosome or the proteasome were involved in the degradation of Cx43, we treated cultures of neonatal rat ventricular myocytes with several different reagents that would inhibit the lysosome (chloroquine, NH4Cl, ALLM, E-64 or leupeptin) or the proteasome (lactacystin or ALLN), for 3 h and determined Cx43 levels by immunoblotting and densitometry. An example of one of these experiments is shown in Fig. 1. Densitometric quantitation of four similar experiments indicated that cells treated with any of the inhibitors contained approximately twice as much Cx43 as control cells (Table 1). In contrast, treatment of cultures with the calpain inhibitor calpastatin did not alter the abundance of Cx43 (data not shown).

Fig. 1

Immunoblot of total cellular Cx43 in cultures of myocytes treated with various protease inhibitors. Cx43 was detected by SDS–PAGE and immunoblotting in total cellular lysates prepared from control myocytes (lane 1) or from parallel cultures treated for 3 h with 100 μM chloroquine (lane 2), 10 mM NH4Cl (lane 3), 25 μM ALLM (lane 4), 25 μM E-64 (lane 5), 100 μM leupeptin (lane 6), 25 μM ALLN (lane 7), or 25 μM lactacystin (lane 8).

View this table:
Table 1

Accumulation of Cx43 in protease inhibitor-treated cardiac myocytes

TreatmentRelative abundance of Cx43 (mean±s.d.)
Control1.00
Chloroquine2.08±0.21
NH4Cl2.25±0.35
ALLM2.09±0.25
E-642.15±0.40
ALLN1.83±0.32
Lactacystin1.62±0.20
  • Cx43 was detected by immunoblotting as in Fig. 1. n=4 for all inhibitors. Cx43 abundance in lysates prepared from cultures treated with any of the inhibitors differed significantly from controls (P<0.05).

In order to confirm that the accumulation of Cx43 in myocytes was due to impaired degradation, we determined the rates of Cx43 turnover in the presence or absence of protease inhibitors by labeling myocyte cultures with [35S]methionine, chasing with medium containing an excess of unlabeled methionine, and isolating Cx43 by immunoprecipitation. A representative set of experiments is shown in Fig. 2. Densitometric analyses were performed on this and 3 similar experiments, and the data were fit with exponential decay curves. These decay curves had degradation constants of k=0.49±0.11 (mean±s.e.m.) for untreated cultures, k=0.34±0.10 for chloroquine treated myocytes, and k=0.24±0.12 for lactacystin treated cultures. Based on these rate constants, the half-life of Cx43 was 1.4 h in control cultures, which agrees well with previous data [1, 3]. In contrast, treatment with chloroquine extended the Cx43 half-life to 2.0 h, and lactacystin treatment increased the half-life to 2.8 h. (These changes of Cx43 turnover would certainly be sufficient to account for the increased Cx43 abundance detected by immunoblotting.)

Fig. 2

The turnover of Cx43 is prolonged by treatment with chloroquine or lactacystin. Cultures of cardiac myocytes were incubated in the presence of [35S]methionine for 2 h and then chased with medium containing unlabeled methionine and no protease inhibitors or 100 μM chloroquine or 25 μM lactacystin for 0, 2, 4, or 6 h. Cx43 was immunoprecipitated, resolved by SDS–PAGE, and detected by fluorography. A representative experiment is shown.

To determine whether altered turnover of the Cx43 polypeptide also reflected impaired degradation of gap junctions, parallel cultures of myocytes were treated with protease inhibitors, fixed, and examined by immunofluorescence. In control cultures, Cx43 staining was primarily found as broken lines of fluorescence at appositional membranes between cells (Fig. 3A). An increase in the abundance of immunoreactive Cx43 at appositional membranes between myocytes was observed in cultures treated with either chloroquine or lactacystin (Fig. 3B,C).

Fig. 3

Chloroquine and lactacystin increase the abundance of immunoreactive Cx43 between neonatal rat myocytes. Cx43 was detected by immunofluorescence in cultures of neonatal rat ventricular myocytes. Individual panels are: (A) control myocytes; (B) myocytes treated for 3 h with 100 μM chloroquine; and (C) myocytes treated for 3 h with 25 μM lactacystin. Scale bar: 8 μm.

This accumulation of gap junctions was further examined by transmission electron microscopy where gap junction structures were identified and quantitated. A representative example of the gap junction profiles identified in control and chloroquine treated myocytes is shown in Fig. 4. The results of the morphometric analyses are shown in Table 2. The size and abundance of gap junctions in control cultures were very similar to what we have observed previously [30]. Treatment for 3 h with 100 μM chloroquine led to a dramatic increase in both the number and length of myocyte gap junctions. Treatment with 25 μM lactacystin led to a less dramatic increase in gap junctions, but the gap junction length per myocyte area was significantly greater than in controls.

Fig. 4

Chloroquine and lactacystin increase the size and number of gap junctions in neonatal rat myocytes. Representative transmission electron micrographs of junctional membrane regions between control myocytes (A) and myocytes from a culture treated for 3 h with 100 μM chloroquine (B). The intercellular junction region in the control cells contains two gap junction profiles (arrows) each approximately 0.3 μm in length. The junction between chloroquine-treated cells also includes two gap junction profiles (arrows) approximately 0.5 and 0.6 μm in length. Original magnification: ×25 000.

View this table:
Table 2

Accumulation of gap junction profiles in lactacystin- and chloroquine-treated cardiac myocytes

ControlChloroquineLactacystin
Intercalated disc length (μm)/1000 μm2 cell area11.50±3.719.40±8.413.10±7.2
Number of gap junctions per 100 μm intercalated disc length11.50±12.832.00±30.227.80±16.8
Number of gap junctions per 1000 μm2 cell area1.20±1.14.20±1.7a2.50±1.3
Gap junction length (μm) per 100 μm intercalated disc length2.90±3.714.10±12.8a10.60±7.2
Gap junction length (μm) per 1000 μm2 cell area0.30±0.31.90±1.1a0.90±0.5a
Mean gap junction length (μm)0.24±0.130.44±0.280.37±0.1
  • All values are shown as mean±s.d.

    aSignificantly different from controls (P<0.05).

In order to assess if Cx43 degradation was induced by heat treatment, cultures of neonatal cardiac myocytes were treated at 43.5°C for 30 min, and Cx43 was detected by immunoblotting or immunofluorescence (Fig. 5Fig. 6). This heat treatment led to a dramatic loss of Cx43 (compare Fig. 5 lane 2 to lane 1 and compare Fig. 6B to A). To determine the roles of the lysosome and the proteasome in this degradation of Cx43 in heat-treated myocytes, cultures of myocytes were heat treated in the presence of proteasomal inhibitors (ALLN or lactacystin) or the lysosomotropic amine, chloroquine. Cells that were heat treated in the presence of any of these reagents contained substantial amounts of Cx43 (Fig. 5, lanes 3–5 and Fig. 6D–F).

Fig. 6

Immunofluorescence analysis of Cx43 in heat-treated cultures of myocytes. Cx43 was detected by immunofluorescence in cultures of neonatal rat ventricular myocytes. Individual panels are: (A) Control myocytes; (B) myocytes treated for 30 min at 43.5°C; (C) myocytes treated for 30 min at 43.5°C, allowed to recover for 3 h at 37°C, and subsequently re-treated for 30 min at 43.5°C; (D) myocytes treated for 30 min at 43.5°C in the presence of 25 μM ALLN; (E) myocytes treated for 30 min at 43.5°C in the presence of 25 μM lactacystin; and (F) myocytes treated for 30 min at 43.5°C in the presence of 100 μM chloroquine. Scale bar: 8 μm.

Fig. 5

Immunoblot of total cellular Cx43 in heat-treated cultures of myocytes and effects of simultaneous treatment with protease inhibitors. Total cellular lysates of myocyte cultures were resolved by SDS–PAGE and immunoblots were reacted with antibodies to Cx43 or HSP70. Specific treatments of individual cultures were: lane 1, control myocytes; lane 2, myocytes treated for 30 min at 43.5°C and immediately harvested; lane 3, myocytes treated for 30 min at 43.5°C in the presence of 25 μM ALLN; lane 4, myocytes treated for 30 min at 43.5°C in the presence of 25 μM lactacystin; or lane 5, myocytes treated for 30 min at 43.5°C in the presence of 100 μM chloroquine.

The heat-induced Cx43 loss was fully reversible, since after returning cultures to 37°C for 3 h, cultures showed a similar abundance and staining pattern to controls (data not shown). Application of a second 30-min heat shock had little observable effect on either the abundance of Cx43 (Fig. 7 lane 3) or on the pattern of Cx43 staining (Fig. 6C), suggesting that Cx43 gap junctions were protected from proteolysis.

Fig. 7

Immunoblot of total cellular Cx43 and HSP70 in heat-treated cultures of myocytes. Total cellular lysates of myocyte cultures were resolved by SDS–PAGE and immunoblots were reacted with antibodies to Cx43 (upper panel) or the inducible heat-shock protein, HSP70 (lower panel). Specific treatments of individual cultures were: lane 1, control myocytes; lane 2, cultures treated for 30 min at 43.5°C and immediately harvested; and lane 3, myocytes treated for 30 min at 43.5°C, allowed to recover for 3 h at 37°C, and subsequently re-treated for 30 min at 43.5°C.

Identical immunoblots of heat-treated myocyte cultures were reacted with antibodies directed against HSP70. These blots showed a dramatic increase in the abundance of HSP70 in the heat-treated cultures (Fig. 7, lane 3).

4 Discussion

In this report, we have presented evidence that both the lysosome and the proteasome contribute to degradation of Cx43 gap junctions in cardiac myocytes. This conclusion is based on the accumulation of Cx43 as detected by immunoblotting and immunofluorescence, accumulation of gap junction profiles detected by transmission electron microscopy, and prolongation of the turnover of Cx43 in pulse-chase experiments in treated cultures of neonatal rat ventricular myocytes.

These protease pathways may have differing relative importances at different cellular sites or at different times in the life cycle of Cx43, leading to differential effects of the inhibitors. The accumulation of gap junction structures following treatment with either lysosomal or proteasomal inhibitors indicates that both of these protease activities act upon Cx43 located in junctions at the plasma membrane. The greater effect of chloroquine than lactacystin (Table 1) suggests that the lysosome is more important than the proteasome in the normal degradation of mature gap junctions, supporting the conclusions of previous morphological studies [7, 21]. In contrast, lactacystin treatment led to a greater increase of the half-life of Cx43 than chloroquine, suggesting a greater role for the proteasome in turnover of the Cx43 polypeptide. The proteasome participates in a proof-reading process of degradation of newly synthesized, but misfolded or improperly oligomerized proteins located in the endoplasmic reticulum [32–35]. The microscopy and pulse-chase data might be reconciled if the proteasome is similarly involved in such ‘quality control’ during the biosynthesis of Cx43 in addition to a role in the degradation of mature gap junctions.

We have also presented evidence that Cx43 was rapidly degraded in cultures of neonatal rat ventricular myocytes subjected to a moderate heat shock (30 min at 43.5°C). The disappearance of Cx43 was clearly due to proteolysis, since degradation of Cx43 in these stressed cells was inhibited by proteasomal inhibitors (lactacystin or ALLN) or by the lysosomotropic amine, chloroquine. Previous studies have shown that heat shock stimulates proteolysis by inducing components of the ATP/ubiquitin-dependent pathway [36]and the lysosomal-dependent process of microautophagy [37].

There may be significant link between the heat-shock response and mechanisms of Cx43 folding and degradation. We found that a brief heat stress followed by recovery (3 h) protected Cx43 from degradation when the cells were given a repeat heat stress. This observation suggested that the cells were producing a protective factor. A candidate for such a protective factor is the heat-shock protein, HSP70, since we observed a parallel induction of expression of HSP70. Many of the major heat-shock proteins function as such molecular chaperones involved in the folding, assembly, and/or degradation of aggregated, misfolded, or damaged proteins [37, 38]. Recent investigations have demonstrated that proteasomal inhibition can lead to a heat-shock response, expression of chaperones, and thermotolerance [39], suggesting that accumulation of misfolded proteins initiates the stress response. It is interesting to speculate that heat-shock proteins (possibly including HSP70) may protect against Cx43 proteolysis.

Induction of the heat-shock response in the heart has been implicated in prevention from ischemic injury [18–20]. Since heat shock can modulate Cx43 degradation, it is possible that interventions to elicit this response or to block proteolysis might minimize degradation or remodeling of myocardial gap junctions.

Acknowledgements

These studies were supported by NIH Grants HL45466, EY08368, and HL50598. J.G.L. is supported by NIH training Grant HL07275. P.N.T. was supported by HL0934001.

References

View Abstract