Effects of Periplocoside P from Periploca sepium on the Midgut Transmembrane Potential of Mythimna separata Larvae

Periplocoside P (PSP) isolated from the root bark of Periploca sepium contains a pregnane glycoside skeleton and possesses high insecticidal properties. Preliminary studies indicated that PSP disrupts epithelial functions in the midgut of lepidopteran larvae. In the present study, we examined the effects of PSP on the apical and basolateral membrane voltages, Va and Vbl, respectively, of cells from (1) midguts isolated from the larvae of the oriental armyworm Mythimna separata that were in vitro incubated with toxins and (2) midguts isolated from M. separata larvae force-fed with PSP. We compared the effects of PSP with the effects of the Bacillus thuringiensis toxin Cry1Ab and inactive periplocoside E (PSE) on the midgut epithelial cells. The results showed that Va rapidly decreased in the presence of PSP in a time- and dose-dependent manner, similar to the effects of Cry1Ab. By contrast, PSE did not affect the Va and Vbl. Additionally, PSP did not influence the Vbl. Given these results, we speculate that PSP may modulate transport mechanisms at the apical membrane of the midgut epithelial cells by inhibiting the V-type H+ ATPase.

Scientific RepoRts | 6:36982 | DOI: 10.1038/srep36982 membrane of the midgut goblet cells [25][26][27][28] . In the present study, we used standard microelectrodes to measure the membrane potential in epithelial cells of the midguts isolated from the larvae of M. separata as previously described [29][30][31][32] . The membrane potential of epithelial cells reflects the transport properties of the epithelium; hence, we measured the effects of PSP on the membrane potential of apical and basolateral cells.
Our results indicated that PSP inhibits the apical membrane voltage of the midgut epithelium of M. separata and might be the first direct evidence that the apical membrane of the midgut epithelium is the action site of PSP. These findings revealed potential targets of the actions of periplocosides and laid a foundation for the further elucidation of the molecular mechanism of action of this botanical insecticide.

Results
Effects of orally administered PSP on midgut transmembrane potential. The apical membrane potential (V a ) and basolateral membrane potential (V bl ) of M. separata midgut epithelial cells were measured after force-feeding the sixth-instar larvae with 1 μ L of dimethyl sulfoxide (DMSO) containing 20 μ g of either test compound or 1 μ L of the 32 K solution containing 1 μ g of activated Cry 1Ab for 2 h to 12 h (Fig. 1).
The Cry1Ab-fed M. separata larvae exhibited similar effects on the V a of midgut epithelial cells (Fig. 1B). Moreover, the V a was − 81.1 ± 5.2 mV (n = 12) and stable for up to 8 h in the larvae fed with only the 32 K solution (Fig. 1B). However, after the larvae ingested Cry 1Ab, the V a decreased rapidly. After 2, 4, 6, and 8 h of force-feeding with 32 K containing Cry1Ab, V a decreased to − 30.1 ± 7.8 mV (n = 10), − 21.7 ± 5.4 mV (n = 12), − 19.0 ± 7.7 mV (n = 11), and − 21.1 ± 6.2 mV (n = 11), respectively (Fig. 1B). By contrast, periplocoside E (PSE), which does not hold any insecticidal activity against the M. separata larvae 32 , did not influence the V a (Fig. 1C) and displayed trends similar to those in the control group. Sixth instar larvae were fed with 1 μ L of DMSO containing 20 μ g of either test compound or 1 μ L of the 32 K solution containing 1 μ g activated Cry 1Ab, and the V a and V bl were measured with time. The control was fed only with 1 μ L DMSO or 32 K solution. A-C display the V a of PSP, Cry 1Ab, and PSP, respectively; D shows the V bl of PSP. V a or V bl refers to the mean membrane potential measured when the values remain stable over 5 min. The values are expressed as the mean ± SEM for 7 to 28 independent experiments. Three asterisks indicate significant differences by Student's t-test (P < 0.001).
Scientific RepoRts | 6:36982 | DOI: 10.1038/srep36982 The basolateral membrane voltage (V bl ) of the midguts isolated from the larvae under control conditions was − 29.2 ± 2.9 mV (n = 14) and did not significantly change in the larvae that ingested PSP (Fig. 1D). Similarly, Cry1Ab exerted no effect on V bl as previously reported 27,31 .
Accordingly, there was no significant change in the transepithelial voltages (V t ) when only solution was added (DMSO or 32 K). The V t following the different treatments are shown in Table 1. The V t gradually decreased with time and approached zero for both the PSP and Cry1Ab treatments. The inactive PSE did not significantly influence the V t . This result further illustrates that the insecticidal activity of PSP is correlated to V t changes. Figure 2 shows the effects of the different concentrations of PSP given by bath incubation on the V a and V bl of the isolated midgut epithelial cells from M. separata larvae. Three concentrations of 0.13 mg/mL to 0.4 mg/mL PSP caused a concentration-dependent depolarization of V a ( Fig. 2A), with the membrane potential depolarized by approximately 15.9 ± 3.4%, 26.3 ± 4.7%, and 31.7 ± 4.9% after 10 min, and 20.9 ± 1.1%, 32.7 ± 4.5%, and 35.8 ± 4.3% after 15 min. By contrast, PSP only slightly influenced the V bl, even at 0.4 μ g/mL (Fig. 2B).

Effects of PSP incubated with midgut cells on transmembrane potential.
As the positive control, activated Cry1Ab affected the V a and V bl to a similar extent as the force-fed PSP. The activated Cry1Ab also depolarized the midgut epithelial cell membrane potential V a at a concentration of 5 μ g/mL ( Fig. 2A). The rate of the depolarization of the apical membrane potential induced by Cry1Ab was higher than that of PSP; the necessary time to reduce membrane potential by 50% was reached after approximately 9.8 ± 0.4 min. However, PSP was faster in response time than Cry1Ab.
Alternatively, the 10 μ g/mL concentration of Cry1Ab applied to the basolateral side of the epithelium cells did not influence V bl (Fig. 2B). Moreover, 0.5 mg/mL PSE, known to lack insecticidal activity, did not affect either the V a or V bl ( Fig. 2A,B).

Discussion
In the present study, we examined the effects of PSP on the apical and basolateral membrane potentials. The V a and V bl of the midgut epithelial cells under both force-feeding and in vitro incubation with PSP were obtained and compared with the effects of PSE and the Bt toxin Cry1Ab. PSE possesses the same molecular structure as PSP, except for one substituent, and holds no insecticidal activity against lepidopteran larvae 22

Figure 2.
Effects of incubated PSP on the V a (A) and V bl (B) of midgut cells from M. separata larvae. When the membrane was stable for over 5 min, 0.4 mL of 32 K solution containing the test compounds of interest was directly added to the bath to replace a similar volume solution. The bath concentrations of PSP were 0.0, 0.13, 0.27 or 0.40 mg/mL, and that of PSE was 0.50 mg/mL. Cry1Ab was adopted as the positive control; its concentration was 5.0 μ g/mL for V a or 10.0 μ g/mL for V bl . V refers to V a or V b measured at the indicated times, and V o is the mean membrane potential measured during the 5 min stable condition before the addition of the test compounds. Data are presented as the mean ± SEM for 5 to 18 independent experiments.
that PSP inhibits the apical membrane voltage of the midgut epithelial cells of M. separata, providing the first direct evidence that the apical membrane of midgut epithelial cells plays a critical role in PSP toxicity.
In general, the transmembrane electrical potential difference is 140 mV or greater and can be measured at the apical side of the columnar cells of lepidopteran larvae 33 . Our results showed that the normal V a of the M. separata larval midgut cells bathed in the standard 32 K solution was approximately − 80 mV and remained stable for up to 30 min. The V a from our study was consistent with previous reports of transmembrane voltages in lepidopteran larvae [29][30][31][32] .
The effects of Cry1Ab on the V a of M. separata larvae midgut epithelial cells were also consistent with the previous results in Lymantria dispar, Bombyx mori, and Manduca sexta larval midguts 29,30 namely, Cry1Ab significantly depolarized the apical membrane potential. After incubation with 5 μ g/mL of Cry1Ab, the V a depolarized by approximately 50% after about 10 min. Cry1Ab exerted no significant effect on the basolateral membrane potential. The activated toxin Cry1Ab is well known to bind to specific receptors and form pores in the apical membranes of midgut epithelial cells. These pores abolish ionic gradients and cause the lysis of epithelial cells, leading to insect death 29,34 .
Similar to the effect of Cry1Ab, the effect of the depolarization of the apical membrane potential by the PSP of midgut epithelial cells from M. separata larvae was concentration-and time-dependent. Notably, PSP was more potent when applied on the apical membrane rather than on the basal membrane of the midgut epithelial cells. This finding suggests that PSP may function by interacting with the apical membrane of the midgut epithelium.
Compared with PSP, PSE is inactive against M. separata. We found that PSP significantly affected V a under both force-feeding and in vitro incubation, but PSE did not influence V a even at high concentrations. Consistent with the bioassay result, PSP acted on the apical membranes of the midgut epithelium when the insects were fed with the natural insecticide. PSP influenced the transmembrane ion transport of the insect's apical membrane, resulting in abnormal transmembrane potentials, which then affected the normal physiological and biochemical functions of the cells. The effects eventually led to ion balance disorder on both sides of the apical membrane, causing midgut cell edema, abdominal swelling, and intestinal rupture until insect death. Because active PSP, rather than inactive PSE, can directly affect the ionic permeability of the apical membrane, membrane potential measurements provide a rapid and sensitive in vitro assay for insecticidal potency in their normal target cells.
In accordance with the mode of action of Cry1A toxins, the activated Cry1Ab binds to aminopeptidase N or alkaline phosphatase, which is involved in driving the activated Cry1Ab into the apical membrane where the pre-pore complex is converted into a membrane-inserted pore that finally leads to ion leakage, cell lysis, and insect death [35][36][37] . PSP is a small molecular compound containing a pregnane glycoside skeleton. Thus, PSP-induced pore formation in the apical membrane, such as that by Bt toxin, is theoretically impossible. In general, the transmembrane electrical potential difference and ion/nutrition transport are based on the function of V-type H + ATPase and V-type H + ATPase coupled with the K + /H + antiporter in the apical membrane of the midgut epithelium goblet cells of lepidopteran larvae 22,25,26,33 . Our results showed that PSP could directly act on the apical membrane and depolarize the transmembrane potential. A previous study showed that PSP inhibits transepithelial electrolyte and fluid secretion in the Malpighian tubules of the yellow fever mosquito Aedes aegypti by inhibiting the V-type H + ATPase 38 . PSP could significantly inhibit the total ATPase activity similar to the inhibition by bafilomycin, which is a known inhibitor of the V-type H + ATPase. Therefore, based on the above analysis and discussion, we can speculate that the most important target of PSP may be the V-type H + ATPase. PSP may potentially disrupt the ionic and chemical homeostasis in midgut epithelial cells and eventually the whole midgut through the cooperative activity of goblet and columnar cells, causing irreversible damage to the midgut.
Alternatively, Bt has been reported to potentially act on the V-type H + ATPase and directly block various types of transmembrane transporters 39 . By comparing the effects of Cry1Ab and PSP on the transmembrane potential of the apical membrane of M. separata, we found that the effects of PSP incubated with the midgut cell membrane were less efficient than those of Cry1Ab. This finding further illustrates that the effect of Cry1Ab on the membrane potential mainly lies in pore formation in the midgut apical membrane. Such a phenomenon facilitates the ion imbalance across the apical membrane and more easily depolarizes the potential. By contrast, PSP may act on the V-type H + ATPase. Although PSP's in vitro efficiency is less than that of Cry1Ab, the former's response time in directly acting on the apical membrane potential is faster, and its effects on physiological function are particularly significant. The similarity of the changes in the apical membrane potentials at different times after feeding with PSP and the changes observed in the Bt bioassay also illustrate this point.
To our knowledge, PSP is the first small-molecule compound reported as an insecticide that targets the V-type H + ATPase. Even so, we cannot rule out the role of PSP in other important insect targets, such as its previously reported ability to activate trypsin activity 22 . However, additional efforts are needed to elucidate whether PSP involves multiple targets or only acts on only the midgut apical membrane transporter to cause the subsequent effects.
Given our results, we consider that PSP may act on the larval midgut apical membrane of M. separata, particularly the V-type H + ATPase of midgut goblet cells, to disrupt the electrochemical gradient and thus abolish the capacity of the cells to transport solutes, which is essential in maintaining the equilibration of the pH between the cytoplasm and the highly alkaline content of the lumen. Such disturbances may lead to the disruption of the midgut epithelium.

Materials and Methods
Insects. Laboratory-adapted M. separata were provided by the Institute of Pesticide Science, Northwest A&F University (NWAFU), China. The population was reared under laboratory conditions (25 °C, 75% to 80% relative humidity, and a 16 h light: 8 h dark photoperiod) and was not exposed to any pesticides prior to the experiment. The larvae were reared on wheat leaves. The actively feeding sixth-instar larvae were used for electrophysiological experiments.
Solutions. Experiments were conducted in the standard 32 K solution, which was composed of 32 mM KCl, To investigate the direct effects of PSP on the larvae midgut, the larvae were fasted for 8 h prior to the in vitro study.
Stock solutions of PSP and PSE were diluted with 32 K solution before use. The final DMSO concentration was less than 0.1% in the bath chamber.
Midgut isolation. Midgut isolation was performed as described previously 31 . The larvae were pinned with the ventral side down on an ice bag, and the body wall was cut with fine microdissection scissors on the dorsal side along with the longitudinal axis from the last abdominal segment to the first thoracic segment. The midguts were transected at each end and rinsed with the 32 K solution. The peritrophic membrane with its food content and the Malpighian tubules were removed using forceps.
Membrane potential measurements. Experiments were conducted as described previously by Peyronnet et al. 27 with minor modifications. When the midgut was cut transversely, both ends tended to curl back onto themselves. The bathing medium was oxygenated by vigorous O 2 bubbling for 30 min prior to use. To measure the apical membrane potential (V a ), the isolated midgut was aspirated from its posterior end into a glass pipette until its anterior end curled around the pipette tip, thus exposing the apical surface (facing the lumen) of the epithelial cells. The pipette was lowered to near the bottom of the testing chamber, and the midgut epithelial cells were impaled with a glass microelectrode filled with 1 M KCl. To measure the basolateral membrane potential (V bl ), each end of the midgut was aspirated into a holding pipette, and the epithelial cells were impaled from the basal side (facing the hemolymph). The electrode resistance was 50 to 150 MΩ . A disc electrode with lead wires (E242; Warner Instruments, Hamden, CT, USA) served as the ground electrode placed in the bath solution. The midgut membrane potentials were measured with an Axoclamp 900 A (Axon Instruments, Foster City, CA, USA) and a Digidata 1440 A interface (Axon Instruments); the pCLAMP 10.2 software (Axon Instruments) was used for data acquisition and analysis. In all experiments, the bathing solution was adjusted to maintain a potassium concentration equal to that of the 32 K solution. All experiments were performed at room temperature.
The midgut preparation was immersed in 3 mL of the 32 K bath solution in vitro. The membrane voltages were considered stable when V a and V bl did not change by more than 0.5 and 0.1 mV/min, respectively. After recording a stable control membrane voltage for 5 min, 0.4 mL of the 32 K solution was collected with a pipette, and an equal volume of the 32 K solution containing test compounds of interest was directly added to the bath. Measured voltages (V) were normalized relative to the voltage measured after the first 5 min (V 0 ). After different treatment periods (0, 2, 4, 6, 8, and 12 h), the force-fed larvae were dissected. The midgut was isolated and rinsed with 3 mL of the 32 K solution. When the membrane voltage remained stable for 5 min, we recorded the membrane potential V a or V bl . The transepithelial voltage V t is the difference between V bl and V a . Electrophysiological data are presented as the means ± SEM for more than seven independent experiments. Statistical analyses were conducted using two-tailed unpaired Student's t test, and P < 0.05 was considered statistically significant.