The effect of peptides with sequences derived from connexins, the constituent proteins of gap junctions, on mechanically stimulated intercellular Ca2+signaling in tracheal airway epithelial cells was studied. Three peptides with sequences corresponding to connexin extracellular loop regions reversibly restricted propagation of Ca2+ waves to neighboring cells. Recovery of communication began within 10 min of removal of the peptides, with inhibition totally reversed by 20–40 min. The peptides were shown to be more effective in inhibiting Ca2+ waves than glycyrrhetinic acid or oleamide. Inhibition of intercellular Ca2+ waves by connexin mimetic peptides did not affect the Ca2+ response to extracellular ATP. Although the intracellular Ca2+ response of tracheal epithelial cells to ATP was greatly reduced by either pretreatment with high doses of ATP or application of apyrase, mechanically stimulated intercellular Ca2+ signaling was not affected by these agents. We conclude that connexin mimetic peptides are effective and reversible inhibitors of gap junctional communication of physiologically significant molecules that underlie Ca2+wave propagation in tracheal epithelial cells and propose a potential mechanism for the mode of action of mimetic peptides.
- cell communication
- gap junction inhibitors
movement of ions and small molecules between cells can occur directly and rapidly through gap junctions. These ubiquitous cell communication channels in the plasma membrane play important roles in coordinating electrical continuity in excitable tissues and organs and in facilitating intercellular signaling during development, growth regulation, and cell differentiation (33, 47). It has recently been shown in tracheal epithelial cells that propagated Ca2+ changes can coordinate ciliary beat frequency (37) and Cl− secretion (29). Gap junction channels form when hexameric connexon hemichannels contributed by neighboring cells dock (2, 50). Connexons are radially arranged arrays of six connexin proteins that form an internal pore. At present, 21 mammalian connexins have been identified, and they all share a well-conserved topographic arrangement in the lipid bilayer characterized by four-membrane-spanning domains, two extracellular loops, one cytosolic loop, and cytosolic amino and carboxy termini (2). Individual connexins differ in general by a variation in carboxy tail length and are identified by a numeric suffix that is derived from the resulting difference in molecular mass (in kDa).
Gap junctional intercellular communication (GJIC) has traditionally been monitored by transfer of low molecular weight fluorescent dyes or by measurement of electrical conductance. Tracheal airway epithelial cells have been shown to be electrically coupled (45), but experiments to show coupling with lucifer yellow have been unsuccessful (27; also data not shown). Although these techniques are recognized as valuable experimental tools to identify cellular coupling, results from these studies can vary between the chosen dyes or cell types (39,42), and they do not always reflect how the transfer of physiological molecules correlates with the presence of GJIC (20). More recently, mechanical stimulation of cultured cells that results in intercellularly propagated changes of intracellular Ca2+ concentration ([Ca2+]i), or Ca2+ waves, has emerged as an attractive and physiologically significant model system to study intercellular communication (43). Mechanical stimulation can lead to propagation of [Ca2+]i changes that are transmitted through gap junctions (e.g., Ref. 3), are independent of gap junctions (e.g., Ref. 40), or may involve both mechanisms (e.g., Ref. 6). Consequently, many reagents that restrict dye transfer between cells have been used to evaluate pathways for Ca2+ waves. Clear interpretations from these approaches are often not forthcoming because of indirect and secondary effects of commonly used GJIC blockers on Ca2+homeostasis, lack of knowledge of their mechanisms of action, or unexplained cell specificity of the blockers. The development of specific inhibitors to monitor intercellular coupling would thus greatly benefit studies of GJIC.
It is highly likely that strong noncovalent interaction occurring between docked connexons is dependent on the interaction of amino acid sequences within both extracellular loops of individual connexins (16, 20, 52), a view reinforced by high-resolution electron crystallography data of recombinant gap junctions (50). The importance of the extracellular loop regions has been shown by using peptides derived from extracellular loop sequences to restrict gap junction formation in cultured cells [e.g., in myocytes (8) and hepatocytes (15)]. Additionally, in established cell systems, extracellular loop-derived peptides have been used to block lucifer yellow dye transfer in COS cells (10) and to demonstrate a role for gap junctions in the functional coupling of endothelium to smooth muscle (9,30). To further analyze the mechanism of action of extracellular loop connexin peptides, we now show that they reversibly uncouple Ca2+ communication mediated through gap junctions in tracheal airway epithelial cells.
Connexin peptides with sequences derived from rat cDNA (1) were purchased from Genosys (Cambridge, UK) or Severn Biotech (Worcester, UK) and are listed in Table1. Dulbecco's modified Eagle's medium (DMEM), Hanks' balanced saline solution (HBSS), penicillin, streptomycin, and amphotericin B were purchased from GIBCO BRL (Life Technologies, Grand Island, NY). Fura 2 and fura 2-AM were purchased from Calbiochem (La Jolla, CA). Apyrase, ATP, α-glycyrrhetinic acid (GA), β-GA, and oleamide were purchased from Sigma (St. Louis, MO). All other chemicals were purchased from Fisher Scientific (Pittsburgh, PA) or Sigma.
Connexin peptides were dissolved in phosphate-buffered saline at 10 mg/ml; stock α-GA was prepared in ethanol; stock β-GA and oleamide were prepared in DMSO. For the experiments, the agents were diluted in HBSS (1.3 mM CaCl2, 5.0 mM KCl, 0.3 mM KH2PO4, 0.5 mM MgCl2, 0.4 mM MgSO4, 137.9 mM NaCl, 0.3 mM Na2PO4, and 1% glucose additionally buffered with 25 mM HEPES, pH 7.4). Mimetic peptides were diluted to 0.25 mg/ml, which resulted in the following concentrations: Gap 26, 160 μM; Gap 27, 190 μM; and Gap 36, 130 μM. Final α-GA, β-GA, and oleamide concentrations were 50 μM.
Tracheal airway tissue cultures were prepared by methods previously described (46). Tracheae were removed from New Zealand White rabbits, and the mucosal epithelium was dissected and cut into small explants. After transfer to collagen-coated glass coverslips, the explants were placed in DMEM with 10% fetal bovine serum, penicillin, streptomycin, and amphotericin B and cultured at 37°C in a 5% CO2 atmosphere. The cultures formed confluent monolayers and were between 7 and 10 days old when used.
Mechanical stimulation of airway cultures was under piezoelectric control and was performed with a glass micropipette (∼1-μm tip diameter) positioned near the apical membrane of a single cell by a hydraulically driven micromanipulator. The pipette was deflected downward for 150 ms to deform the cell membrane. If cell membranes were broken (as measured by loss of fura 2 dye and/or the lack of a complete recovery of the stimulated cell to resting [Ca2+]i), the experiment was not included in data analysis. This prevented analysis of Ca2+ wave propagation by loss of intracellular contents such as ATP (e.g., Refs.26, 32).
[Ca2+]i was calculated by ratiometric analysis of fura 2 fluorescence (21). Cells were loaded with fura 2 by fura 2-AM incubation. Fura 2 fluorescence was observed on an Olympus IX70 microscope after alternating excitation at 340 and 380 nm by a 75-W xenon lamp linked to a Delta Ram illuminator [Photon Technologies Incorporated (PTI), Monmouth Junction, NJ] and a fiber-optic line. Images of emitted fluorescence above 510 nm were recorded by an intensified charge-coupled device camera (PTI) and simultaneously displayed on a 21-inch Vivitron color monitor. The imaging system was under software control (ImageMaster, PTI) on an IBM clone computer.
A change in [Ca2+]i was considered positive if the cell increased [Ca2+]i to 200 nM or more, a two- to fourfold change over resting values. The stimulated cell was always counted in analysis. Thus a Ca2+wave of one cell represents a response only by the stimulated cell and no intercellular communication. In these experiments, the field of view varied and was limited to ∼20–35 cells. On occasion, wave propagation would encompass >20 cells (or exit the field of view). In such cases, the Ca2+ wave propagation was given a total score of 20 cells. Because maximum numbers were imposed on wave counts, the distance of Ca2+ wave propagation under poor or no blocking conditions may be underrepresented.
Mechanically induced Ca2+ wave propagation.
Mechanical stimulation of a single airway epithelial cell resulted in an immediate increase in [Ca2+]iin that cell and was propagated to surrounding cells (Fig.1 A). The spread of mechanically induced Ca2+ waves was quantified by determining the total number of cells that increased [Ca2+]i to >200 nM. Under these conditions, a typical Ca2+ wave involved ∼15 cells (Fig.2 A).
Mechanically induced Ca2+ waves in the presence of extracellular loop I peptides.
The propagation of Ca2+ signaling in cells exposed to Gap 26 peptide (Table 1) derived from the first extracellular loop sequence of connexin (Cx) 43 was examined. During the first 15 min of exposure to Gap 26 peptide, the number of cells participating in the Ca2+ wave was similar to that seen in the absence of peptide (Fig. 2 B). However, between 16 and 30 min of incubation in the presence of Gap 26 peptide, the number of cells of the Ca2+ wave decreased to less than half of that seen in the absence of the peptide. After 31–60 min of exposure to Gap 26 peptide, the propagation of the Ca2+ wave was confined to one or two adjacent cells. The inhibition of Ca2+ wave propagation was reversible because within 10 min of removal of Gap 26 peptide, the total number of cells participating in Ca2+waves gradually increased, and by 20 min of washout, mechanically stimulated Ca2+ waves propagated to a similar number of cells that participated before Gap 26 peptide application. As with other peptides examined in the present study, Gap 26 did not alter the cellular pathways that increase [Ca2+]i nor did they have an effect on HBSS pH, itself a potential regulator of GJIC (data not shown).
Mechanically induced Ca2+ waves in the presence of extracellular loop II peptides.
Propagation of Ca2+ signaling was also examined in the presence of two peptides derived from sequences in the second extracellular loop of Cx43 (Gap 27 and Gap 36 peptides; Table 1). In the first 15 min of incubation with Gap 27 peptide, the average number of cells (12 cells) of mechanically induced Ca2+ waves was slightly reduced from the normal number (Fig. 2 C). A continuing reduction in total cell participation on Gap 27 incubation was time dependent. By 60 min in Gap 27 peptide, Ca2+ waves were nearly completely inhibited; in many of the experiments, [Ca2+]i changes were restricted to the stimulated cell (Fig. 1 B). As with Gap 26 peptide, inhibition of Ca2+ wave propagation by Gap 27 peptide was reversible. Within 10 min of Gap 27 peptide washout, recovery of mechanically induced Ca2+ waves was evident, and by 20 min after removal, Ca2+ waves returned to normal size (Figs. 1 C and 2 C).
Gap 36 peptide, an extended extracellular loop peptide that included part of the sequence in the Gap 27 peptide (SRPTEK; Table 1) was also tested (Fig. 2 D). This peptide was applied at a lower concentration and was slower in inhibiting Ca2+ wave propagation. Nevertheless, after a 90-min incubation with Gap 27 peptide, Ca2+ waves were restricted to <10 cells. By 120 min, Ca2+ waves averaged less than two cells per experiment. Washout of Gap 36 peptide also resulted in recovery of Ca2+ waves; by 20 min of washout, Ca2+ waves covered nearly 10 cells, and by 40 min of washout, Ca2+waves returned to normal levels.
As a control, airway epithelial cultures were incubated in Des 5 peptide derived from an intracellular loop connexin sequence (Table 1). Little change in Ca2+ wave propagation was observed when cells were exposed to this peptide. After a 90-min incubation in Des 5 peptide, Ca2+ waves still covered 13–15 cells (Fig.2 E). Other intracellularly derived connexin mimetic peptides had no effect on Ca2+ wave propagation when applied extracellularly (data not shown).
Mechanically induced Ca2+ waves in the presence of nonpeptide inhibitors.
The effects of three GJIC inhibitors, α-GA, β-GA, and oleamide, on Ca2+ wave propagation were also tested. Incubation in 50 μM α-GA produced a reduction in the number of cells participating in the Ca2+ wave (Figs. 1 D and3 B). After 60 min, Ca2+ waves were reduced to approximately seven cells on average (significant difference from untreated cells by paired Student's t-test, 0.500 < P < 0.100), which remained constant for a further 60 min (significant difference from untreated cells by paired Student's t-test, 0.100 < P < 0.250). As with the peptides, the inhibition was reversible; within 20 min after washout, Ca2+ waves spread to 13 cells.
β-GA and oleamide used in these studies were far less effective in inhibiting Ca2+ wave propagation than the above agents. β-GA produced a slight reduction in wave propagation, involving 11–12 cells after a 60-min incubation (Fig. 3 C). Washout of β-GA restored the Ca2+ wave to 15 cells (Fig.3 C). Similarly, oleamide had little or no effect on Ca2+ wave propagation; the same number of cells participated in Ca2+ waves before and after 50 μM oleamide application (Fig. 3 D).
It should be noted that the doses of α-GA, β-GA, and oleamide were very high compared with what has been used to uncouple dye transfer in other experimental cell systems (e.g., Ref. 22). Higher doses and a subsequent dose-response curve were uninterpretable due to the compromising effect contributed by ethanol or DMSO, necessary for dilution of the inhibitor into appropriate stock solutions, on the cells.
Effects of ATP on [Ca2+]i in the presence of GJIC inhibitors.
A role for ATP as an extracellular messenger in mechanically induced Ca2+ waves in other cell types has been proposed (5,11, 17, 25). Cultured tracheal airway epithelial cells also respond to externally applied ATP, with an EC50 of ∼0.5 μM (26), although evidence with a purinergic receptor blocker and anti-connexin antibodies argues against extracellularly diffusible factors being a major contributor to the propagation of mechanically induced intercellular Ca2+ waves (3,26)
To further investigate a potential role for ATP, we first compared the response of cells to ATP before and after inhibition by Gap 27. The percentage of cells that responded to 100 nM or 1 μM ATP by increasing [Ca2+]i was similar whether cells were blocked from intercellular communication by Gap 27 incubation or were untreated (Fig. 4 A). These data suggest that despite a lack of propagation of Ca2+ waves after Gap 27 treatment, the ability to respond to ATP is still intact. We attempted to alter the ability for tracheal epithelial cells to respond to extracellular ATP (Fig. 4 B). As expected, pretreatment with high doses of ATP (10 μM) almost completely eliminated [Ca2+]i responses to previously saturating concentrations of ATP (1 μM). Additionally, we applied the ATP-degrading enzyme apyrase (50 U/ml) to separate cultures, and this resulted in the elimination of the ATP response to saturating extracellular ATP (10 μM). However, when mechanical stimulation was used to induce intercellular Ca2+ wave propagation under either of these conditions, there was no effect on the number of cells participating in the wave (Fig.5). Thus changes in [Ca2+]i that occur during mechanically stimulated Ca2+ wave propagation are largely independent of any ATP-mediated pathway.
We show that short connexin extracellular loop mimetic peptides reversibly inhibit the gap junction-dependent propagation of Ca2+ waves between tracheal airway epithelial cells. Three peptides (Gap 26, Gap 27, and Gap 36) that correspond in sequence to short segments of the two highly conserved extracellular loops of connexins were shown to be effective in restricting the propagation of mechanically induced Ca2+ waves. In agreement with previous results indicating that mechanically induced Ca2+waves in tracheal airway epithelial cultures are propagated through gap junctions (3, 4, 26, 44), Gap 27 peptide did not alter ATP-induced intracellular Ca2+ signaling, and dampening the response to extracellular ATP signaling did not alter mechanically induced Ca2+ wave propagation. We also compared the action of the connexin mimetic peptides with three agents used to uncouple cells linked by gap junctions, namely, α-GA, β-GA, and oleamide. These compounds were not as effective in blocking mechanically induced Ca2+ wave propagation, and a measurable degree of Ca2+ wave restriction was only observed with high concentrations of α-GA.
Inhibition of integrative behavior, believed to be underwritten by GJIC by the extracellular loop connexin mimetic peptides, is not limited to the tracheal cell cultures shown in this study. Other peptides with sequences corresponding to those in connexin extracellular loops have been used to disrupt formation of GJIC in myocytes (52) and hepatocytes (15). Additionally, they have been used to dissect the gap junctional component of endothelium-dependent relaxations in rabbit arteries (8, 14, 30). Because connexin mimetic peptides block the transfer of a multitude of physiological signals in a variety of cell systems and appear to operate on amino acid sequences conserved within connexin proteins, they may provide a generic and specific assay for GJIC across cell types.
Because of the strong interaction between opposing connexons (50) and the time course of the inhibition of Ca2+ wave propagation (Fig. 2), we think that it is unlikely that the mimetic peptides are disrupting the operation of existing gap junctions. We propose that the primary action of the connexin mimetic peptides is in preventing the initial docking or the redocking of connexons after disassembly and before cytosolic uptake to prevent completion of intercellular channels. When relative contributions to connexon-connexon interactions by specific regions in the extracellular loop peptides were studied, the conserved QPG and SHVR motifs of extracellular loop I and the SRPTEK motif in extracellular loop II of connexins were suggested to be important in cell coupling (12, 52). Data from these studies are consistent with the SHVR (present in Gap 26 peptide) and SRPTEK (present in Gap 27 and Gap 36 peptides) amino acid regions contributing crucially to the docking of connexons to generate gap junctions. The proposed model is in agreement with the low level of connexins at the plasma membrane in cultured tracheal epithelial cells (3,45) as well as the rapid turnover of gap junction proteins (35), the quick delivery time for connexins to the plasma membrane (18), and the dynamic nature of gap junction plaques demonstrated with connexin-green fluorescent protein (31) that have been described in other cell culture systems. However, we cannot exclude a mechanism of inhibition that involves a more direct interference with assembly or an effect on channel gating induced from interaction of the peptides with external aspects of the cooperating connexons. Because we have measured the formation of functional communication and not the turnover of connexons at the plasma membrane, a direct correlation with the overall making and breaking of gap junctions cannot be made at this time.
Two major pathways have been proposed for mechanically induced Ca2+ waves, GJIC and the release of extracellular signaling molecules, most commonly ATP. In the case of GJIC, mechanical stimulation increases [Ca2+]i in the stimulated cell by inducing an influx of extracellular Ca2+or a release from intracellular stores. Proposed signaling mechanisms involving transfer of inositol 1,4,5-trisphosphate or Ca2+through gap junctions can result in further release of Ca2+from intracellular stores (7, 22, 32, 44, 51). In the second case, mechanical stimulation causes secretion of molecules, including ATP, that can diffuse in the extracellular environment and bind to receptors on adjacent cells where they activate increases in [Ca2+]i (5, 25). In many cases, it has been shown that mechanical stimulation can induce both types of signaling (6, 17, 32, 49). Cultured tracheal epithelial cells respond to extracellular ATP by increasing [Ca2+]i (26, 41, 53). However, transmission of mechanically induced Ca2+ waves in tracheal epithelial cells is not affected by suramin, a purinergic receptor agonist (26), and is blocked by intracellular loading of connexin-specific antibodies (3), both supporting a predominant, if not exclusive, gap junction-mediated pathway. A recent report (11) using several cell lines suggests that connexins may alter ATP release through connexons or unidentified ATP-binding cassette proteins and thus indirectly cause changes in Ca2+ wave propagation independent of their contributions to gap junctions. This conclusion was based in part on the use of gap junction inhibitors such as α-GA to interrupt dye transfer (dicarboxydichlorofluorescein) despite the ineffectiveness of these inhibitors in blocking Ca2+ wave propagation (11). The present data from tracheal airway epithelial cells also indicate the low efficacy of α-GA on Ca2+ wave propagation. In contrast, the peptide mimetics may directly target a key step in gap junction assembly for biogenesis without altering ATP response. Further support against extracellular ATP-mediated Ca2+ waves emerges from the absence of a change in Ca2+ wave propagation after reduction of the extracellular ATP response. These data concur with other reports (3, 4, 26,44) that in cultured airway epithelial cells, the propagation of Ca2+ waves is mediated via gap junctions.
A full analysis of how Ca2+ changes are propagated has been limited by the lack of specific gap junction inhibitors. Lipophilic compounds such as octanol (11), heptanol (22), and the general anesthetic halothane (44) have been suggested to inhibit GJIC by dissolving in the plasma membrane, thereby physically contracting the gap junction pores (48). Although these agents may reduce GJIC, many nonspecific effects on other membrane proteins, including those that lead to alterations of Ca2+ signaling, have been reported (13, 38). More recently, further agents such as the arachidonic acid derivative anandamide (22), α-GA, β-GA (17, 22, 49), and oleamide (22) have been proposed to more specifically target gap junctions and inhibit the operation of gap junction channels. Such compounds share a common drawback with lipophilic agents that block gap junctional conductance; their actions are complex and probably nonspecific, and their primary mechanisms of action remain unknown. For example, it has been suggested that GA derivatives can alter GJIC through phosphorylation of Cx43 proteins (23). However, GA derivatives have been shown to be effective in blocking gap junctions made of Cx26 (19), which does not have the suggested phosphorylation target sites. Also, in cultured alveolar type II cells, α-GA was shown to alter Cx43 phosphorylation as well as Cx43 protein and mRNA expression (24), thus displaying a multitude of effects on lung epithelial cultures.
Other inhibitors used to block GJIC include antibodies directed against specific regions of connexins (1). These have proved to be valuable probes for the study of GJIC in live cells because of their specificity of binding to the target protein and subsequent minimal nonspecific effects on cell signaling systems (1, 36). Connexin antibody probes have been used to block dye and electrical communication (36) as well as Ca2+ waves (3, 28). In astrocytes, extracellular loop antibodies were used to disrupt intercellular Ca2+ signaling (28), and in tracheal epithelial cells, antibodies to the cytosolic loop and carboxy terminus were used to block propagation of Ca2+ waves (3). The peptide inhibitors used in this study have important advantages over the application of antibodies, especially with respect to their size and ease of application. A disadvantage to peptide inhibition is the potential difficulty in blocking gap junctions made up of specific connexin isoforms while allowing gap junctions made up of other isoforms to function. However, because of the limited connexon-connexon pairings among connexons made of different connexins, selection of extracellular loop peptides appears feasible in specific tissues expressing a variety of connexins (34) and may allow for some selectivity of connexin block in future experiments. Nevertheless, connexin mimetic peptides can be used to selectively and reversibly block GJIC-dependent propagation of physiological signals that may underlie coordinated tissue function, including ciliary beat frequency (37) and salt secretions (29), in tracheal epithelial cells.
We thank A. B. Narahara, B. E. Isakson, S. T. Stoddard, and C.E. Olsen for cell culture and helpful comments on imaging sequences.
This work was supported by American Lung Association Research Grant RG-0420N (to S. Boitano) and Medical Research Council Programme Grant G-9305117 (to W. H. Evans).
Address for reprint requests and other correspondence: S. Boitano, Dept. of Zoology and Physiology, PO Box 3166, Univ. of Wyoming, Laramie, WY 82071-3166 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2000 the American Physiological Society