Characterization of Inhibitory Effectiveness in Hyperpolarization-Activated Cation Currents by a Group of ent-Kaurane-Type Diterpenoids from Croton tonkinensis

Croton is an extensive flowering plant genus in the spurge family, Euphorbiaceae. Three croton compounds with the common ent-kaurane skeleton have been purified from Croton tonkinensis. Methods: We examined any modifications of croton components (i.e., croton-01 [ent-18-acetoxy-7α-hydroxykaur-16-en-15-one], croton-02 [ent-7α,14β-dihydroxykaur-16-en-15-one] and croton-03 [ent-1β-acetoxy-7α,14β-dihydroxykaur-16-en-15-one] on either hyperpolarization-activated cation current (Ih) or erg-mediated K+ current identified in pituitary tumor (GH3) cells and in rat insulin-secreting (INS-1) cells via patch-clamp methods. Results: Addition of croton-01, croton-02, or croton-03 effectively and differentially depressed Ih amplitude. Croton-03 (3 μM) shifted the activation curve of Ih to a more negative potential by approximately 11 mV. The voltage-dependent hysteresis of Ih was also diminished by croton-03 administration. Croton-03-induced depression of Ih could not be attenuated by SQ-22536 (10 μM), an inhibitor of adenylate cyclase, but indeed reversed by oxaliplatin (10 μM). The Ih in INS-1 cells was also depressed effectively by croton-03. Conclusion: Our study highlights the evidence that these ent-kaurane diterpenoids might conceivably perturb these ionic currents through which they have high influence on the functional activities of endocrine or neuroendocrine cells.


Introduction
The genus croton (Euphorbiaceae) includes about 300 species that are widely distributed throughout tropical regions. C. tonkinensis Gagnep. (Euphorbiaceae), known as Kho sam cho la in Vietnamese, is a tropical shrub native to northern Vietnam and has been used in Vietnam for the treatment of various types of disorders [1]. Previous phytochemical investigations have shown that C. tonkinensis is a rich source of diterpenoids [1]. Anti-inflammatory and cancer chemopreventive activities of C. tonkinensis extracts have been recently reported. Those effects are thought to be linked to their ability to depress the transcription nuclear factor κB [1].
Previous studies have demonstrated that intravenous injection with the essential oil of C. zehntneri or nepetaefolius could induce rapid and dose-dependent hypotension and bradycardia [2][3][4].

Effects of Croton-03 and Other ent-Kaurane Diterpenoids on Hyperpolarization-Activated Cation Current (I h ) Identified in Pituitary GH 3 Cells
In an initial stage of experiments, the modifications of I h caused by croton-03 were evaluated in GH 3 cells which were immersed in Ca 2+ -free Tyrode's solution, and the recording pipette used was filled with K + -containing solution. As the whole-cell current recordings were firmly established, we maintained the examined cells at the level of −40 mV, and a long-lasting membrane hyperpolarization to −110 mV was thereafter applied. Under this experimental protocol, a slowly activating current elicited by sustained hyperpolarization was readily evoked and has hence been referred as an I h [14,18,24]. Notably, as cells were exposed to croton-03, the amplitudes of I h by maintained hyperpolarization was progressively diminished together with a considerable slowing in activation time course of the current ( Figure 1A). For example, cell exposure to 3 µM croton-03 decreased current amplitude at the end of hyperpolarizing pulse from 309±21 to 168 ± 17 pA (n = 8, p < 0.05). After washout of the agent, current amplitude returned to 287 ± 19 pA (n = 8). In addition to the decreased I h amplitude, the value of activation time constant (τ act ) of the current in response to maintained hyperpolarization was evidently raised as cells were exposed to croton-03 ( Figure 1B,C). For example, the presence of 3 µM croton-03 was able to increase the τ act value to 1158 ± 113 msec (n = 8, p < 0.05) from a control value of 567 ± 54 msec (n = 8). membrane hyperpolarization were further examined and compared in this study. The concentrationdependent relationships among the inhibitory effects of these agents on Ih are illustrated in Figure  1D. The IC50 value of croton-01, croton-02, or croton-03 that were required for the inhibition of Ih amplitude was calculated to be 2.89, 6.25, or 2.84 μM, respectively; however, the Hill coefficients obtained among their effects on Ih did not differ significantly. Therefore, croton-01 or croton-03 tends to be more potent in depressing Ih than croton-02.  The effects of these diterpenoids (i.e., croton-01, croton-02 and croton-03) on I h in response to membrane hyperpolarization were further examined and compared in this study. The concentration-dependent relationships among the inhibitory effects of these agents on I h are illustrated in Figure 1D. The IC 50 value of croton-01, croton-02, or croton-03 that were required for the inhibition of I h amplitude was calculated to be 2.89, 6.25, or 2.84 µM, respectively; however, the Hill coefficients obtained among their effects on I h did not differ significantly. Therefore, croton-01 or croton-03 tends to be more potent in depressing I h than croton-02.

Effects of Croton-03 on I-V Relationship and Steady-State Activation Curve of I h in GH 3 Cells
We further examined the effect of croton-03 on I h taken at different levels of membrane potentials. The I h was elicited when the cell was hyperpolarized from −40 to a series of voltage steps ranging between −110 and −30 mV ( Figure 2B). Croton-03 at a concentration of 3 µM was able to depress the amplitude of I h measured at the end of each hyperpolarizing step. The overall I-V relationships of I h with or without addition of croton-03 are shown in Figure 2B. Figure 2C shows the steady-state activation curve of I h taken with or without addition of croton-03 (3 µM). In control (i.e., in the absence of croton-03), V 1/2 = −96.2 ± 2.6 mV, q = 3.2 ± 0.8 e (n = 7), while in the presence of 3 µM croton-03, V 1/2 = −84.3 ± 2.4 mV, q = 3.3 ± 0.7 e (n = 7). Therefore, it is clear from the present results that croton-03 not only led to a conceivable reduction in I h magnitude, but it also significantly shifted the activation curve along the voltage axis to a more negative potential by approximately 11 mV, irrespectively of minimal change in the gating charge (q) of the curve.

Effects of Croton-03 on I-V Relationship and Steady-State Activation Curve of Ih in GH3 Cells
We further examined the effect of croton-03 on Ih taken at different levels of membrane potentials. The Ih was elicited when the cell was hyperpolarized from −40 to a series of voltage steps ranging between −110 and −30 mV ( Figure 2B). Croton-03 at a concentration of 3 μM was able to depress the amplitude of Ih measured at the end of each hyperpolarizing step. The overall I-V relationships of Ih with or without addition of croton-03 are shown in Figure 2B. Figure 2C shows the steady-state activation curve of Ih taken with or without addition of croton-03 (3 μM). In control (i.e., in the absence of croton-03), V1/2 = −96.2 ± 2.6 mV, q = 3.2 ± 0.8 e (n = 7), while in the presence of 3 μM croton-03, V1/2 = −84.3 ± 2.4 mV, q = 3.3 ± 0.7 e (n = 7). Therefore, it is clear from the present results that croton-03 not only led to a conceivable reduction in Ih magnitude, but it also significantly shifted the activation curve along the voltage axis to a more negative potential by approximately 11 mV, irrespectively of minimal change in the gating charge (q) of the curve. In this set of voltage-clamp current recordings, the conditioning voltage pulses with a duration of 2 sec to potential ranging from -120 and -30 mV. After each conditioning pulse, a test pulse to -110 mV was applied to evoke Ih for measurement of the steady-state activation curve of the current. The upper part of (A) indicates the voltage protocol applied. An example of current traces obtained by this two-pulse protocol is illustrated in (A). (B) Averaged I-V relationship of Ih in the control (■) and during the exposure to 3 μM croton-03 (□). Current amplitude was measured at the end of each membrane (i.e., conditioning) potential applied. Each point represents the mean ± SEM (n = 7 for each point). (C) Effect of croton-03 on the steady-state activation curve of Ih in the absence (•) and presence (Ο) of 3 μM croton-03 in GH3 cells. The Ih amplitudes were normalized with those measured at the end of each test pulse and then constructed against the conditioning potentials. The continuous line was least-squares fitted by the Boltzmann equation (see text for details). Please note that the steadystate activation curve of Ih in GH3 cells was significantly shifted toward hyperpolarizing voltage during the exposure to croton-03. In this set of voltage-clamp current recordings, the conditioning voltage pulses with a duration of 2 sec to potential ranging from −120 and −30 mV. After each conditioning pulse, a test pulse to −110 mV was applied to evoke I h for measurement of the steady-state activation curve of the current. The upper part of (A) indicates the voltage protocol applied. An example of current traces obtained by this two-pulse protocol is illustrated in (A). (B) Averaged I-V relationship of I h in the control ( ) and during the exposure to 3 µM croton-03 ( ). Current amplitude was measured at the end of each membrane (i.e., conditioning) potential applied. Each point represents the mean ± SEM (n = 7 for each point). (C) Effect of croton-03 on the steady-state activation curve of I h in the absence (•) and presence (O) of 3 µM croton-03 in GH 3 cells. The I h amplitudes were normalized with those measured at the end of each test pulse and then constructed against the conditioning potentials. The continuous line was least-squares fitted by the Boltzmann equation (see text for details). Please note that the steady-state activation curve of I h in GH 3 cells was significantly shifted toward hyperpolarizing voltage during the exposure to croton-03.

Effect of Croton-03 on the Voltage-Dependent Hysteresis of I h Elicited in Response to Long-Lasting Triangular Ramp Pulse
The voltage-dependent hysteresis of ionic currents (e.g., I h ) has been previously demonstrated to exert a significant impact on electrical behaviors such as the firing of action potentials [25][26][27]. For this reason, we further explored whether there is possible voltage-dependent hysteresis existing in I h recorded from GH 3 cells. In this series of experiments, we delivered a long-lasting triangular ramp pulse with a duration of 2 sec (i.e., ±0.11 V/sec), as whole-cell mode was firmly established. It is evident from Figure 3A that the trajectories of I h elicited by the upsloping (i.e., depolarizing from −150 to −40 mV) and downsloping (hyperpolarizing from −40 to −150 mV) ramp pulse as a function of time were distinguishable between them. The current amplitude elicited by the upsloping limb of triangular voltage ramp was significantly greater than that by the downsloping limb, indicating that there is a voltage hysteresis for this current as depicted in Figure 3B, namely the relationship of I h versus membrane potential. As the ramp speed became reduced, the hysteresis degree for I h was greatly raised. Of note, as the examined cell was exposed to croton-03 (3 µM), I h amplitude evoked in the upsloping limb of long-lasting triangular ramp was noted to be decreased to a greater extent than that measured from the downsloping limb. For example, in controls, I h at the level of -100 mV elicited upon the upsloping and downsloping ends of triangular ramp pulse were 213 ± 23 and 97 ± 12 pA (n = 7), respectively, the values of which were found to differ significantly between them (p < 0.05). Moreover, in the presence of 3 µM croton-03, the amplitudes of forward and backward I h taken at the same level of membrane potential were significantly reduced to 136 ± 16 and 58 ± 9 pA (n = 7, p < 0.05), respectively. However, the magnitude of croton-03-induced current inhibition at the upsloping and downsloping limbs of triangular ramp did not differ significantly, i.e., the presence of 3 µM croton decreased I h amplitude (at −100 mV) in the upsloping and downsloping limb by about 40%.

Effect of Croton-03, YS-035, Zatebradine, Croton-03 Plus SQ-22536, and Croton-03 Plus Oxaliplatin on Ih Amplitude
The extracts from C. zehntneri such as anethole and estragole were demonstrated to produce an increase in the level of intracellular cyclic AMP in the copora cavernosa smooth muscle [28]. In the next set of experiments, we examined and compared the effects of croton-03, YS-035, zatebradine, We next quantified the degree of voltage-dependent hysteresis based on the difference in area under the curve in the forward (upsloping) and reverse (downsloping) direction as described by the arrows in Figure 3B. It was seen that for I h identified in GH 3 cells, the degree of voltage hysteresis increased with slower ramp speed, and that the presence of croton-03 led to a considerable reduction in the amount of such hysteresis. For example, as the duration of triangular ramp was prolonged from 2 to 3 sec, the area under the curve between forward and backward current traces (i.e., ∆area) increased from 23.9 ± 4.2 to 34.1 ± 4.9 mV·pA, (n = 7, p < 0.05). Figure 3C illustrates summary of the data showing effects of croton-03 at the different concentrations on the area under such curve. For example, apart from its depression of I h magnitude, addition of croton-03 (3 µM) significantly decreased the area by about 50% elicited in response to such long-lasting triangular voltage ramp, as demonstrated by a significant decrease of ∆area from 23.9 ± 4.2 to 11.1 ± 2.2 mV·nA (n = 7, p < 0.05).

Effect of Croton-03, YS-035, Zatebradine, Croton-03 Plus SQ-22536, and Croton-03 Plus Oxaliplatin on I h Amplitude
The extracts from C. zehntneri such as anethole and estragole were demonstrated to produce an increase in the level of intracellular cyclic AMP in the copora cavernosa smooth muscle [28]. In the next set of experiments, we examined and compared the effects of croton-03, YS-035, zatebradine, croton-03 plus SQ-22536, and croton-03 plus oxaliplatin on I h amplitude in GH 3 cells. YS-035 or zatebradine was recognized as an inhibitor of I h [19,29], oxaliplatin was recently demonstrated to stimulate I h [30,31], and SQ-22536 is an inhibitor adenylate cyclase. As illustrated in Figure 4, similar to the effect of croton-03 on I h described above, the addition of either YS-053 (3 µM) or zatebradine (3 µM) was effective at depressing I h amplitude. Subsequent application of SQ-22536 (10 µM), still in the presence of croton-03, had minimal effect on I h produced by croton-03; however, further addition of oxaliplatin (10 µM) significantly attenuated croton-03-mediated inhibition of I h . Alternatively, chlorotoxin (1 µM), an inhibitor of Clchannels, did not depress I h amplitude effectively (data not shown). Consequently, the experimental results led us to indicate that croton-03-mediated depression of I h appears to be independent of its increase in intracellular cyclic AMP, though it is sensitive to be reversed by further addition of oxaliplatin.

Effect of Croton-03 on erg-Mediated K + Current (IK(erg)) in GH3 Cells
In another set of current recordings, we investigated the possible modification of croton-03 on IK(erg). In these experiments, we bathed cells in high-K + , Ca 2+ -free solution and, during the measurements, we filled the pipette solution by using K + -containing solution. As the whole-cell current recordings were achieved, the examined cell was maintained at the level of -10 mV and a long-step membrane hyperpolarization was applied. As illustrated in Figure 5, the addition of croton-03 at a concentration of 3 or 10 μM slightly depressed the amplitude of deactivating Ik(erg), the

Effect of Croton-03 on erg-Mediated K + Current (I K(erg) ) in GH 3 Cells
In another set of current recordings, we investigated the possible modification of croton-03 on I K(erg) . In these experiments, we bathed cells in high-K + , Ca 2+ -free solution and, during the measurements, we filled the pipette solution by using K + -containing solution. As the whole-cell current recordings were achieved, the examined cell was maintained at the level of -10 mV and a long-step membrane hyperpolarization was applied. As illustrated in Figure 5, the addition of croton-03 at a concentration of 3 or 10 µM slightly depressed the amplitude of deactivating I k(erg) , the biophysical or pharmacological properties of which were previously described [32]. However, minimal change in the decaying time course of I K(erg) in response to membrane hyperpolarization was demonstrated in the presence of croton-03. The concentration-dependent inhibition of croton-03 on I K(erg) was then constructed with an IC 50 value of 35.3 µM, which is higher than that for the inhibition of I h described above. As such, the I k(erg) identified in GH 3 is relatively resistant to depression by the presence of croton-03.

Effect of Croton-03 on sag Potential Measured from GH3 Cells
We further performed current-clamp voltage recordings in attempts to test if croton-03 could produce any perturbations on sag potential in these cells. Sag potential in response to hyperpolarizing current stimulus has been described to be intimately linked to the occurrence of Ih in different types of central neurons [20,33,34]. As shown in Figure 6, under our experimental condition, when the whole-cell voltage recordings were firmly established, a long-step hyperpolarizing current injection with the amplitude of around 25 pA was found to induce sag

Effect of Croton-03 on Sag Potential Measured from GH 3 Cells
We further performed current-clamp voltage recordings in attempts to test if croton-03 could produce any perturbations on sag potential in these cells. Sag potential in response to hyperpolarizing current stimulus has been described to be intimately linked to the occurrence of I h in different types of central neurons [20,33,34]. As shown in Figure 6, under our experimental condition, when the whole-cell voltage recordings were firmly established, a long-step hyperpolarizing current injection with the amplitude of around 25 pA was found to induce sag potential (i.e., drop down to a lower level in the membrane potential upon hyperpolarizing current stimuli). The addition of ivabradine (3 µM), an inhibitor of I h , was effective at depressing the amplitude of sag potential. Moreover, the presence of 1 or 3 µM croton-03 produced an effective depression of sag potential elicited by hyperpolarizing current stimulus. For example, croton-03 at a concentration of 3 µM decreased the amplitude of sag potential from 29.2 ± 2.1 to 18.3 ± 1.5 mV (n = 8, p < 0.05). Therefore, the sag potential identified in GH 3 cells is connected with the magnitude of I h and the depression of such potential caused by the presence of croton-03 could result largely from its inhibitory effect on I h observed in GH 3 cells.

Inhibitory Effect of Croton-03 on Ih in INS1 Insulin-Secreting Cells
The biophysical or pharmacological properties of Ih observed in GH3 cells may be distinguishable from those in other types of endocrine cells. Previous studies have demonstrated the presence of Ih or delayed-rectifier K + current in pancreatic β-cells [35,36]. INS-1 cells have retained some characteristics of normal pancreatic β-cells [37]. In another set of whole-cell current recordings, we thus further attempted to test whether croton-03 is capable of perturbing ionic currents (e.g., Ih or IK(erg)) present in INS-1 insulin-secreting cells. As illustrated in Figure 7, indistinguishable from the observations made in GH3 cells, the presence of croton-03 effectively depressed the amplitude of Ih in response to long-step membrane hyperpolarization. Moreover, subsequent addition of oxaliplatin (10

Inhibitory Effect of Croton-03 on I h in INS1 Insulin-Secreting Cells
The biophysical or pharmacological properties of I h observed in GH 3 cells may be distinguishable from those in other types of endocrine cells. Previous studies have demonstrated the presence of I h or delayed-rectifier K + current in pancreatic β-cells [35,36]. INS-1 cells have retained some characteristics of normal pancreatic β-cells [37]. In another set of whole-cell current recordings, we thus further attempted to test whether croton-03 is capable of perturbing ionic currents (e.g., I h or I K(erg) ) present in INS-1 insulin-secreting cells. As illustrated in Figure 7, indistinguishable from the observations made in GH 3 cells, the presence of croton-03 effectively depressed the amplitude of I h in response to long-step membrane hyperpolarization. Moreover, subsequent addition of oxaliplatin (10 µM), still in the presence of croton-03 (3 µM), could significantly attenuate croton-03-mediated inhibition of the I h amplitude. Oxaliplatin was previously demonstrated to stimulate I h amplitude [30,31].

Effect of Croton-03 on IK(erg) Identified in INS1 Cells
The IK(erg) identified in INS-1 cells was also tested to evaluate whether croton-03 has any effect on this current. Cells were bathed in high-K + , Ca 2+ -free solution, and the recording pipette was filled with K + -containing solution. As illustrated in Figure 8, when the membrane potential was hyperpolarized from −10 to −140 mV, the deactivating IK(erg) was readily evoked in these cells. Moreover, within 2 min of exposing to croton-03 at a concentration of 10 or 30 μM, the Ik(erg) amplitude elicited by step hyperpolarization was robustly reduced. For example, the addition of croton-03 (30 μM) decreased current amplitude measured at the beginning of hyperpolarizing pulse from 984 ± 43 to 713 ± 28 pA (n = 8, p < 0.05). Hence, in keeping with the observations described above in GH3 cells, the IK(erg) identified in INS-1 cells was subject to be mildly but significantly depressed by croton-03.

Effect of Croton-03 on I K(erg) Identified in INS1 Cells
The I K(erg) identified in INS-1 cells was also tested to evaluate whether croton-03 has any effect on this current. Cells were bathed in high-K + , Ca 2+ -free solution, and the recording pipette was filled with K + -containing solution. As illustrated in Figure 8, when the membrane potential was hyperpolarized from −10 to −140 mV, the deactivating I K(erg) was readily evoked in these cells. Moreover, within 2 min of exposing to croton-03 at a concentration of 10 or 30 µM, the I k(erg) amplitude elicited by step hyperpolarization was robustly reduced. For example, the addition of croton-03 (30 µM) decreased current amplitude measured at the beginning of hyperpolarizing pulse from 984 ± 43 to 713 ± 28 pA (n = 8, p < 0.05). Hence, in keeping with the observations described above in GH 3 cells, the I K(erg) identified in INS-1 cells was subject to be mildly but significantly depressed by croton-03.

Discussion
The principal findings in this study were as follows: (1) the presence of croton-01, croton-02, or croton-03 produced concentration-dependent inhibition of Ih in pituitary GH3 cells with effective IC50 of 2.89, 6.25 or 2.84 μM, respectively.; (2) croton-03 shifted the steady-state activation curve of Ih toward a more negative potential; (3) croton-03 decreased the voltage-dependent hysteresis of Ih and further addition of SQ-22536 did not alter croton-03-mediated inhibition of Ih; (4) croton-03 mildly depressed the amplitude of IK(erg); (5) croton-03 decreased the amplitude of sag potential elicited by hyperpolarizing current stimulus; and (6) croton-03 also depressed Ih amplitude in INS-1 insulinsecreting cells.
A previous report has demonstrated the ability of extracts from C. zehntneri (i.e., anethole and estragole) to increase the levels of cyclic AMP in corpora cavernosa smooth muscle [28]. However, in continued presence of croton-03, further addition of SQ-22536, an inhibitor of adenylate cyclase, was unable to attenuate croton-03-mediated depression of Ih seen in GH3 cells. Therefore, croton-03induced block of Ih is unlikely to ascribe largely from changes in the level of intracellular cyclic AMP.
In light of the steady-state activation curve of Ih during the exposure of croton-03, the voltage for half-maximal activation was noted to be in the range of the firing of action potentials. The activation time course of Ih observed at the different level of voltages became slower in the presence of croton-03. Therefore, it is conceivable that the croton-03 molecule has a higher affinity toward the open state of HCN channels than toward the closed or resting state of the channels identified in GH3 or INS-1 cells. As such, the magnitude of croton-03-induced block on Ih tends to be voltagedependent, and any changes in Ih amplitude caused by croton-03 would fairly rely on the croton-03 concentration applied, the firing of action potentials, and the pre-existing level of membrane potential.
The voltage-dependent hysteresis of Ih has been shown to exhibit substantial role in influencing electrical behavior of electrically excitable cells such as GH3 cells. In keeping with previous In these experiments, we bathed in Ca 2+ -free, high-K + solution and, during the measurements, we filled the pipette by using K + -containing solution. (A) Representative I K(erg) traces elicited in response to 1-sec step hyperpolarization from −10 to −140 mV (as indicated in the upper part). The current trace labeled a is control, and those labeled b or c were obtained in the presence of 10 or 30 µM croton-03, respectively. Arrowhead indicates the zero current level and calibration mark shown in the right lower side applies all current traces. (B) Summary bar graph of croton-03 effects on the amplitude of deactivating I K(erg) elicited by membrane hyperpolarization (mean ± SEM; n = 8 for each bar). Current amplitude was taken at the beginning of each hyperpolarizing pulse. * Significantly different from control (p < 0.05) and † significantly different from croton-03 (10 µM) alone group (p < 0.05).

Discussion
The principal findings in this study were as follows: (1) the presence of croton-01, croton-02, or croton-03 produced concentration-dependent inhibition of I h in pituitary GH 3 cells with effective IC 50 of 2.89, 6.25 or 2.84 µM, respectively.; (2) croton-03 shifted the steady-state activation curve of I h toward a more negative potential; (3) croton-03 decreased the voltage-dependent hysteresis of I h and further addition of SQ-22536 did not alter croton-03-mediated inhibition of I h ; (4) croton-03 mildly depressed the amplitude of I K(erg) ; (5) croton-03 decreased the amplitude of sag potential elicited by hyperpolarizing current stimulus; and (6) croton-03 also depressed I h amplitude in INS-1 insulin-secreting cells.
A previous report has demonstrated the ability of extracts from C. zehntneri (i.e., anethole and estragole) to increase the levels of cyclic AMP in corpora cavernosa smooth muscle [28]. However, in continued presence of croton-03, further addition of SQ-22536, an inhibitor of adenylate cyclase, was unable to attenuate croton-03-mediated depression of I h seen in GH 3 cells. Therefore, croton-03-induced block of I h is unlikely to ascribe largely from changes in the level of intracellular cyclic AMP.
In light of the steady-state activation curve of I h during the exposure of croton-03, the voltage for half-maximal activation was noted to be in the range of the firing of action potentials. The activation time course of I h observed at the different level of voltages became slower in the presence of croton-03. Therefore, it is conceivable that the croton-03 molecule has a higher affinity toward the open state of HCN channels than toward the closed or resting state of the channels identified in GH 3 or INS-1 cells. As such, the magnitude of croton-03-induced block on I h tends to be voltage-dependent, and any changes in I h amplitude caused by croton-03 would fairly rely on the croton-03 concentration applied, the firing of action potentials, and the pre-existing level of membrane potential.
The voltage-dependent hysteresis of I h has been shown to exhibit substantial role in influencing electrical behavior of electrically excitable cells such as GH 3 cells. In keeping with previous observations [25][26][27], the I h identified in GH 3 cells was found to undergo either a hysteresis in its voltage dependence, or mode shift in which the voltage sensitivity of gating charge movements relies on the previous state [26,27]. In this study, we also examined possible perturbations of croton-03 on such non-equilibrium property of I h in GH 3 cells. Our results clearly demonstrated that the presence of this diterpenoid (croton-03) was capable of diminishing such hysteresis involved in the voltage-dependent elicitation of I h .
In this study, the effective IC 50 of croton-03 required for the inhibition of I h seen in GH 3 cells was 2.84 µM, a value that is similar to that of croton-01, but slightly lower than that of croton-02. This discrepancy remains unclear; however, it could be related to the possibility that croton-03 is assumed to be more hydrophobic than croton-01 or croton-02. The presence of croton-03 was also noted to depress the activation time course of I h as well as to shift the activation curve of the current to a more depolarized potential. Regardless of the detailed mechanism of its inhibitory actions, the I h (or HCN-encoded current) could conceivably be an important target of these diterpenoids such as croton-3, which can interact with the HCN channel to decrease the amplitude of I h in endocrine cells such as GH 3 and INS-1 cells. The present results also imply the possibility of its perturbations on different cell types which express functional HCN channels such as sensory neurons [38,39]. Indeed, our results demonstrated that subsequent addition of oxaliplatin, but still in the presence of croton-03, could significantly attenuate croton-03-mediated inhibition of I h in GH 3 and INS-1 cells. Oxaliplatin, an activator of I h , was recognized as a chemotherapy-induced neuropathic pain [30,31].
It needs to be mentioned that the IC 50 value (i.e., 2.84 µM) estimated for croton-03-mediated inhibition of I h tends to be so high that it may be reached anywhere in the organism upon consumption of even large amounts of C. tonkinensis. Therefore, to what extent the results observed herein have a real physiological, nutrition-related relevance remains to be further investigated.
The I h has been recognized to be carried by channels of the HCN gene family, namely HCN1, HCN2, HCN3 and HCN4 [10]. It has been recently demonstrated that HCN1, HCN2 or HCN3 channels play essential roles in sensory excitability and pain sensation [30,38,40]. Regardless of the detailed mechanism of their inhibitory actions, there is likely to be a relevant link between the effects of these croton derivatives on electrically excitable cells and their inhibitory effects on HCN channel activity, though croton-03-mediated effect on HCN channels might not be isoform-specific. It is tempting to speculate that the blocking by these ent-kaurane-type diterpenoids of I h plays an important role in modulating the functional activities of neurons and neuroendocrine or endocrine cells [5,6]. Therefore, croton derivatives and other structurally related ent-kaurane-type diterpenoids would be the harbinger of intriguing compounds that use the open/activated state of the HCN channels as a substrate. Our study strongly suggests that these ent-kaurane diterpenoids should somehow act on cellular mechanisms through which they exert high influence on the functional activities of endocrine or neuroendocrine cells.

Purification, Extraction, and Fractionation of Plant Materials
The general procedures follow those shown in the previous research [41]. The whole C. tonkinensis Gagnep (Euphorbiaceae) plant has been collected in Vietnam, and the plant material was identified and authenticated by Vu Xuan Phuong, Institute of Ecology and Biological Resources, Vietnamese Academy of Science and Technology, Hanoi, Vietnam. A voucher specimen (Viet-TSWu-2009-0901-001) was deposited in the herbarium of the Institute of Ecology and Biological Resources, Vietnamese Academy of Science and Technology.
Extraction and fractionation were performed as described previously [41]. Briefly, air-dried and powdered whole plants of C. tonkinensis were extracted with methanol and concentrated to form brown syrup. The n-hexane soluble fraction was subjected to silica gel column chromatography eluted with a step gradient to afford 19 fractions (H1-H19). Minor fractions in subfraction H13-4, H15-2-2 and H15-3-3 were recrystallized with chloroform methanol to yield croton-01, croton-03, and croton-02, respectively.

Electrophysiological Measurements
Shortly before each experiment, cells (i.e., GH 3 or INS-1 cells) were dissociated and a few drops of cell suspension was transferred to a home-made chamber mounted on the fixed stage of an inverted Diaphot-200 microscope (Nikon, Tokyo, Japan). They were immersed at room temperature (20-25 • C) in normal Tyrode's solution, the composition of which is described above. We fabricated the recording electrode from Kimax-51 glass capillaries (#34500; Kimble, Vineland, NJ) using a PP-830 vertical puller (Narishige, Tokyo, Japan) in which a two-step pull mechanism was applied, and their tips were fire-polished with a microforge (MF-83, Narishige). During the measurements, the electrode with tip resistance ranging from 3 to 5 MΩ, which was firmly inserted into holder, was maneuvered by use of a WR-98 micromanipulator (Narishige). Patch-clamp experiments operated under voltageor clamp-clamp mode were carried out by using an RK-400 patch-clamp amplifier (Bio-Logic, Claix, France) connected with a personal computer [43]. Shortly before giga-seal formation was achieved, the potentials were commonly corrected for the liquid junction potential that generally developed at the pipette tip, as the composition of internal solution was different from that in the bath.

Data Recordings
The data comprising both potential and current traces were collected online and stored in an HP Pavilion ×360 touchscreen laptop computer (14-cd1053TX; Hewlett-Packard, Palo Alto, CA, USA) at 10 kHz equipped with the 12-bit Digidata 1440A interface (Molecular Devices, Sunnyvale, CA, USA). The latter device was widely accepted and used for efficient analog-to-digital/digital-to-analog conversion. During the experiments, the data acquisition system was driven by pCLAMP 10.7 software (Molecular Devices) run under Windows 10 (Redmond, WA, USA), and the current/voltage signals were also simultaneously monitored on LCD monitor (MB169B+; ASUS, Taipei, Taiwan) through a USB type-C connection. Current signals were low-pass filtered at 2 kHz with an FL-4 four-pole Bessel filter (Dagan, Minneapolis, MN, USA). Through digital-to-analog conversion, the pCLAMP-generated voltage-clamp patterns with different rectangular, triangular or linear ramp waveforms were designed beforehand, in attempts to determine the current-voltage (I-V) relationships for different types of ionic currents such as I h , and the steady-state activation curve of I h .

Data Analyses
To determine percentage inhibition of croton-01, croton-02, or croton-03 on I h , each cell was hyperpolarized from −40 to −120 with a duration of 2 s, while to evaluate that of croton-03 on I K(erg) , the cell was hyperpolarized from −10 to −90 mV. The I h or I K(erg) amplitudes respectively measured at the end or beginning of each hyperpolarizing step in the presence of different concentrations of compounds were compared with the control value (i.e., in the absence of any agent). The concentration-dependent effects of these compounds on the inhibition of I h or I K(erg) amplitude seen in GH 3 cells were fitted with the goodness of fit to a modified Hill function by 64-bit OriginPro 2016 (OriginLab); that is, where [C] is the concentration of croton-01, croton-2, or croton-03; IC 50 and n H are the concentration required for a 50% inhibition and the Hill coefficient, respectively; and E max is the maximal inhibitions of I h or I k(erg) amplitude produced by each tested compound.
To evaluate the effect of croton-03 on the steady-state activation of I h , the relationships between the normalized amplitude of I h and the conditioning potentials obtained in the absence and presence of 3 µM croton-03 were least-squares fitted with a Boltzmann function of the following form: where I max is the maximal amplitude of I h taken with or without addition of 3 µM croton-03 at the conditioning potential of −30 mV, V 1/2 the voltage at which half-maximal activation occurs, q the apparent gating charge, F Faraday's constant, R the universal gas constant, and T the absolute temperature.

Statistical Analyses
Linear or nonlinear curve fitting to data shown here was performed by using either OriginPro 2016 (OriginLab, Northampton, MA) or the Solver subroutine embedded in Excel 2016 (Microsoft) [44]. The data are presented as mean±standard errors of the mean (SEM) and error bars were plotted as SEM. Values of n indicate the number of cells obtained. The paired or unpaired Student's t-test and one-way analysis of variance (ANOVA) followed by the least significance difference method for multiple comparisons were used for the statistical evaluation of differences among means. Non-parametric Kruskal-Wallis test would be implemented, as the assumption of normality underlying ANOVA was violated. Statistical analyses were generally made using Statistical Package for the Social Sciences version 20 (SPSS, IBM Corp., Armonk, NY, USA). The level of statistical significance was set at p < 0.05.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.