The direct effect of K on xanthophores of the medaka, Oryzias latipes, was studied using denervated scale preparations and cultured cells. At concentrations of 30 mM or higher, K ions caused almost complete aggregation of pigment in denervated xanthophores. The withdrawal of divalent cations from the perfusing solution, or the addition of Ca2 channel blocker to the medium, made the cells unresponsive to K . When Fluo 3-loaded xanthophores were stimulated by K , a temporary increase in the fluorescence intensity was observed. Experimental elevation of cytosolic Ca2 concentration induced aggregation of pigment in denervated xanthophores pretreated with A23187. Forskolin, an activator of adenylate cyclase, inhibited the K -induced aggregation of pigment in a dose-dependent manner, whereas W-7, a calmodulin antagonist, did not. From these results, it is likely that membrane depolarization caused by elevated K concentration in the perfusing medium may be accompanied by Ca2 entry into the cytoplasm through voltage-dependent Ca2 channels present in the xanthophore membrane. This would result in inhibition of adenylate cyclase activity, since it is well established that decreases in intracellular cAMP level induce pigment aggregation in fish chromatophores.
The rapid physiological color changes of teleost fishes are due to motile activities of integumentary colored cells, termed chromatophores. Their motility, i.e. aggregation or dispersion of pigment granules within them, is well known to be under hormonal and neural regulation (Fujii and Oshima, 1986, 1994). In light-absorbing chromatophores (melanophores, xanthophores and erythrophores) of many fish species, pigment aggregation is also caused by elevated K+ concentration in the perfusing solution, although K+ does not affect directly those cells (Fujii, 1959; Iwata et al., 1959). Rather, K+acts on neural elements surrounding chromatophores to liberate the neurotransmitter (norepinephrine), which in turn induces the centripetal migration of pigment granules mediated by α-adrenoceptors on the cell membrane. The dispersion of light-scattering organelle in leucophores of the medaka via β-adrenoceptors (Iga, 1978) and the coloring response of mo-tile iridophores of the blue damselfish (Oshima et al., 1985) are also elicited by K+ through a similar mechanism.
Using scales of the medaka, Oryzias latipes (wild type and red type), Iga (1969) first observed that denervated xanthophores aggregated their pigment as did innervated xanthophores, when the perfusing solution was changed from physiological saline to isotonic KCl, whereas denervated melanophores and leucophores never responded to the K+-rich solution. Since xanthophores of the medaka also respond to norepinephrine, α-melanophore stimulating hormone (α-MSH), melanin-concentrating hormone, etc., the motile activities of those cells are surely controlled by nervous and hormonal systems (Fujii and Oshima, 1986).
In the present study, the pigment-aggregating action of K+ on medaka xanthophores extraneous to the nervous system was confirmed by the use of cultured cells in addition to denervated xanthophores on scales, and the mechanism of the direct action of K+ was further examined. It may facilitate understanding of delicate and subtle color changes of fish to clarify the difference in characteristics among chromatophore species.
MATERIALS AND METHODS
Medaka (Oryzias latipes, wild type and orange-red variety) of both sexes, 2.5–3.5 cm in length, were used. For denervation, reserpine (1 mg) obtained in commercial ampoules (Daiichi Pharmaceutical) was diluted with tap water (500 ml), and medakas were kept alive for 12 hr in the dark in that solution. Since melanophores (wild type) or amelanotic melanophores (orange-red variety) in the skin tissue of scales from these medakas were subsequently refractory to treatment with isotonic K+ solution (128 mM), chromatophores were considered to be completely denervated (cf. Fujii, 1959).
Scales on the dorsal trunk were isolated in a physiological saline solution (PSS) of the following composition: NaCl 125.3 mM, KCl 2.7 mM, CaCl2 1.8 mM, MgCl2 1.8 mM, d-glucose 5.6 mM, Tris-HCl buffer 5.0 mM (pH 7.2). Scales were immersed for 5 min in Ca2+- and Mg2+- free saline (CMFS; NaCl 130.7 mM, KCl 2.7 mM, d-glucose 5.6 mM, EDTA 3 mM, Tris-HCl buffer 5.0 mM; pH 7.2) and the epidermal layer of skin tissue on each scale was removed with forceps. Scales were fixed on a holder with a fine glass fiber, both sides of which had been attached to a coverslip with epoxy adhesive; the holder was then turned over and placed on a perfusion chamber with a narrow trough filled with saline (Oshima and Fujii, 1984).
Treatment with K+
K+-rich solutions (10, 20, 30, 40, 50, 70, 90, 110, 128 mM) were prepared by substitution of equimolar KCl for NaCl in PSS. After high K+ stimulation for 3 min, denervated scale preparations were treated with 5 μM norepinephrine (NE) for 3 min to obtain perfect (100%) responses of chromatophores. The NE solution was prepared by dilution of an injection fluid of norepinephrine hydrochloride (Sankyo). Responses of xanthophores were continuously photomicrographed and a video image analyzing system, Luzex F (Nireco), was employed for quantitative analysis. In this system, changes in the area occupied by pigment granules that are present inside a rectangular frame on a monitor screen are calibrated (cf. Oshima et al., 1988).
In addition to scale preparations, cultured xanthophores were also used: xanthophores dissociated from scales were cultured for 48 hr at 25°C by methods described in detail previously (Oshima et al., 1996). In case pigment granules in cultured xanthophores were aggregated in PSS, they were dispersed beforehand with 1 nM α-melanophore stimulating hormone (α-MSH, Sigma Chemical).
Effects of divalent cation-free medium and channel blockers
To prepare Ca2+- and Mg2+-free 50 mM K+ solution containing 3 mM EDTA, equimolar NaCl in CMFS was replaced by KCl. After rinsing with CMFS for 7–8 min, denervated xanthophores were stimulated with the divalent cation-free K+-rich solution for 4 min and their responses were photomicrographed. As Ca2+ and Na+ channel blockers, methoxy verapamil (D-600; Sigma) and tetrodotoxin (TTX; Wako Pure Chemical) were used, respectively. Denervated xanthophores pretreated with D-600 or TTX for 5 min were exposed to 50 mM K+ solution for 4 min in the continued presence of each blocker. After washing with PSS for 30-40 min, cells were stimulated again by K+ for 4 min and finally by NE (5 μM) for 3 min.
Measurement of fluorescence signal from Ca2+-indicator
Epidermis-free, denervated scales were treated with 5 μM Fluo 3-AM (Sigma) for 60 min in the dark at room temperature; Fluo 3-AM was dissolved in a small quantity of DMSO and diluted with PSS immediately before use. With its reduced polarity, acetoxymethyl ester of Fluo 3, i.e. Fluo 3-AM, passes through the cell membrane with ease, and is hydrolyzed in the cytoplasm by acetylesterase (Minta et al., 1989; Kao et al., 1989). Thus, cells can be loaded with Fluo 3 by incubation with its acetoxymethyl ester. In the cytoplasm, the Ca2+-indicator binds free Ca2+, leading to an augmentation of fluorescence. The treatment with the dye was followed by multiple washings with PSS, and scale preparations were fixed on the bottom of the perfusion chamber.
Fluorescence signals were monitored with an InSIGHTplus laser scanning confocal microscope system, equipped with an Olympus inverted microscope IMT-2, (Meridian Instruments). The dye has a peak emission at 525 nm when excited by the 488 nm argon laser line.
Manipulation of intracellular concentration of Ca2+
After epidermis-free, denervated scales were rinsed with CMFS for 5–6 min, they were pretreated with 20 μM calcium ionophore A23187 (Sigma) for 3 min at room temperature. The ionophore solution was prepared by diluting a stock solution of A23187 dissolved in DMSO with CMFS. The final maximum concentration of DMSO in the medium was less than 0.5%. Scale preparations were next incubated in a Ca-buffer (pH 7.4) in the presence of A23187; Ca-buffers were prepared according to the equation given by Imai and Takeda (1967). In the present study, the concentration of EGTA (Dojinndo Laboratories) was fixed at 2.5 mM, and the amount of CaCl2 added was changed.
Effects of calmodulin blocker and adenylate cyclase activator
Denervated xanthophores were pretreated with a specific calmodulin antagonist, N-(6-aminohexyl)-5-chloro-naphthalenesulfonamide hydrochloride (W-7; Seikagaku Kogyo), for 5 min and then exposed to K+-rich solution (50 mM) for 4 min in the continued presence of the inhibitor. A saline rinse for 30 min was followed by K+stimulation for 4 min, and finally 5 μM NE was applied to obtain complete (100%) aggregation of pigment. Xanthophores were also incubated with 50 mM K+ solution in the presence of forskolin (Wako Pure Chemical), an activator of adenylate cyclase, for 4 min. After a 30 min wash with PSS, responses to K+ and perfect aggregation evoked by 5 μM NE were photomicrographed.
In physiological saline solution (PSS), xanthophores on scales from denervated medakas maintained a pigment-dispersed state. In isotonic solution containing 10 mM K+, pigment granules remained dispersed, while about 60% aggregation occurred at 20 mM K+ (Fig. 1). Thus, the threshold concentration of K+ must be between 10 and 20 mM in denervated xanthophores. Almost complete aggregation of pigment was induced within 3 min at K+ concentrations higher than 30 mM, and therefore, a 50 mM K+ solution was used in subsequent experiments to stimulate xanthophores. K+ had no effect on denervated melanophores (wild type), amelanotic melanophores (orange-red variety) or leucophores.
Next, responses of xanthophores in primary cell culture were studied. In about half of cultured xanthophores, pigment granules remained dispersed in PSS. Such cells responded within 3 min to 50 mM K+ by pigment aggregation (Fig. 2), and the responses were reversible. The other half of xanthophores in primary culture assumed a pigment-aggregated state in PSS. After pigment granules were fully dispersed by pretreatment with α-MSH (1 nM) for 8 min, such xanthophores were stimulated with 50 mM K+, or rinsed with PSS (control). The application of 50 mM K+ accelerated pigment aggregation; the time required for 50% aggregation was 3.8 min (n = 5), whereas in the control experiment, it took 6.7 min (n = 5) to obtain the same level of aggregation. From these results, it is clear that the pigment-aggregating action of K+ on medaka xanthophores is direct.
In Ca2+- and Mg2+-free K+-rich solution (50 mM), pigment aggregation in denervated xanthophores was not detected (Fig. 3), implying that transport of Ca2+ through the xanthophore membrane may be indispensable for K+ action. D-600, a Ca2+channel blocker (Lee and Tsien, 1983), was employed to investigate the role of Ca2+ channels on K+-induced aggregation of pigment. As indicated in Fig. 4, D-600 inhibited pigment aggregation in a dose-dependent manner, and its effect was reversible. On the other hand, TTX, which is one of the most potent channel blockers of Na+, and which can block Na+ flux at about 1 μM (Catterall, 1984), had no effect on K+-induced pigment aggregation, even at 10 or 50 μM. In addition, as shown in Fig. 1, 128 mM K+ solution prepared by substituting KCl totally for NaCl in standard PSS, aggregated pigment in denervated xanthophores. These results suggest that K+ acts directly on the xanthophore membrane to bring about Ca2+ influx, but not Na+ influx, to the cytoplasm.
Using Fluo 3 (a Ca2+-indicator)-loaded xanthophores in scales from the denervated medaka (orange-red variety), the fluorescence signal was monitored. Figure 5 shows the fluorescence intensity by means of pseudo-color, 0.3, 3.3, 7.3 and 47.3 sec after application of 50 mM K+. In the area surrounded by the white line, many xanthophores were present. The continuous change in the fluorescence intensity averaged on that area is shown in Fig. 6. A temporary increase in intensity occurred immediately after K+ application, and about 6.5 sec later, the level reached the maximum. After measurements of fluorescence intensity, xanthophores were confirmed to be able to assume the pigment-aggregated state.
The experimental elevation of cytosolic Ca2+ concentration actually caused pigment aggregation in medaka (wild type) xanthophores. When denervated scale preparations pretreated with 20 μM A23187 were exposed to 10 nM Ca-buffer in the presence of the ionophore, pigment aggregation within xanthophores occurred, and that aggregation progressed even further in 100 nM Ca-buffer (Fig. 7). In melanophores, however, an aggregation response was not observed.
To analyze the role of cytosolic Ca2+ in pigment aggregation within xanthophores, the effect of W-7 (10 and 50 μM), a specific calmodulin inhibitor, was examined. At either concentration of W-7, pigment-aggregating responses of xanthophores were not inhibited. In the presence of 50 μM W-7, the mean value of the magnitude of pigment aggregation by K+ was 94.8 ± 2.1% (n = 5), whereas it was 97.0 ± 1.3% (n = 5) in the absence of the antagonist. There was no significant difference between those values (t-test, P > 0.05).
Therefore, it is likely that calmodulin is not involved in this signal transduction pathway.
When denervated xanthophores were exposed to 50 mM K+ in the presence of forskolin, an adenylate cyclase (AC) stimulating agent, pigment aggregation was depressed in a dose-dependent manner (Fig. 8). For example, the level of K+-induced aggregation in the presence of 10 μM forskolin was only 58.8% of the complete aggregation evoked by 5 μM NE. Therefore, Ca2+ that influx into xanthophores through Ca2+channels may inhibit the AC activity, leading to a decrease in the cytosolic level of cAMP, which in turn may cause pigment aggregation in medaka xanthophores.
Based on the observation that denervated fish melanophores were unresponsive to K+, Fujii (1959) concluded that K+ does not act directly on melanophores themselves, but on adrenergic nerves innervating them to release the transmitter, norepinephrine (NE). In medaka xanthophores, however, depolarization of the cell membrane elicited by increases in extracellular K+ concentration may also trigger aggregation of pigment. The present results imply that voltage-dependent Ca2+ channels present in the xanthophore membrane are opened to allow Ca2+ to penetrate into the cytoplasm, a phenomenon which resembles the K+-contracture of musculature. Luby-Phelps and Porter (1982) further reported that cultured erythrophores of Holocentrus ascensionis responded to 60 mM K+ solution by pigment aggregation. In that paper, the authors questioned whether the earlier conclusion of Fujii (1959) was right. Now, we can easily understand that characteristics of brightly pigmented cells (erythrophores and xanthophores) of some fish species are very different from those of melanophores. There may be few voltage-dependent Ca2+ channels on the melanophore membrane, because cultured (Oshima et al., 1988) and denervated melanophores (present data) do not respond to K+.
Up to the present, it has not been confirmed whether chromatophore membranes are actually depolarized by certain endogenous mechanisms. However, if such mechanisms are functional, differential responses of melanophores and brightly pigmented cells triggered by membrane depolarization might contribute to revelation of subtle body colors of fish. Using Fundulus, Fries (1931) found that, on a blue background, melanophores were generally dispersed while xanthophores were aggregated, whereas both types were dispersed on a black background.
In permeabilized melanophores of the Nile tilapia, exogenously applied inositol 1,4,5-trisphosphate (IP3) caused pigment aggregation (Fujii et al., 1991). The application of NE was accompanied by an increase in cytosolic Ca2+ in platyfish (Oshima et al., 1988) and tilapia (Toyohara and Fujii, 1992) melanophores. In tilapia cells, the cytosolic level of Ca2+ was doubled (about 200 nM) when melanin aggregation was elicited by NE. Further, W-7, a calmodulin antagonist, inhibited NE-induced aggregation in a dose-dependent manner in blue damselfish (Oshima et al., 1990) and in tilapia melanophores (Oshima and Wannitikul, 1996). Thus, in melanophores of at least some teleost species, a Ca/calmodulin complex (calmodulin activated by Ca2+) may be involved in signal transduction in the pigment-aggregating response to NE. Activated calmodulin is thought to stimulate protein phosphatase 2B, leading to pigment aggregation (Thaler and Haimo, 1990; Oshima and Wannitikul, 1996). In medaka xanthophores, however, K+-induced aggregation was not blocked by W-7, suggesting that calmodulin may not play an important role in intracellular events. Since aggregation was inhibited by forskolin, Ca2+ supplied from outside the cells through Ca2+ channels may regulate the activity of adenylate cyclase (AC). DeBernardi et al. (1991) and Yoshimura and Cooper (1992) reported that low concentrations of Ca2+ (submicromolar Ca2+), independently of calmodulin, inhibited AC (type V) activity in mammalian cells. If also true in medaka xanthophores, decreases in cytosolic cyclic AMP concentration caused by decreases in AC activity may induce pigment aggregation via decreases in activity of protein kinase A (Fujii and Oshima, 1986; Oshima and Wannitikul, 1996), although the type of AC in membranes of medaka xanthophores has not yet been examined.
W-7 also had no inhibitory effect on pigment aggregation by NE in melanophores of the medaka (Oshima, unpublished data). Further, medaka melanophores, different from those of other fish species, were refractory to elevated intracellular Ca2+ concentration shown in Fig. 7. Therefore, Ca2+ and calmodulin may not signal the pigment aggregation response in medaka melanophores.
We thank to Dr. R. Fujii of Toho University for his valuable discussion. This study was supported in part by a grant from the Biomolecular Science Award Foundation of Toho University.