Inhibition of wortmannin activities by amino compounds
Abstract
Wortmannin caused normal and strong inhibition on catecholamine secretion from bovine adrenal chromaffin cells and in vitro phosphoinositide 3-kinase activity in NaCl-, Na isethionate-, choline Cl-, Na acetate-, and N-acetyl glycine-based media. However, brief preincubation of wortmannin with the media containing amino compounds such as glutamate, aspartate, lysine, and glycine resulted in the prevention of the inhibitory effects of wortmannin on the above responses as two indexes of wortmannin activities. On the other hand, the amino compounds also caused several rapid changes in wortmannin medium; the changes in absorption spec- trum of the medium; and the changes in the retention time of the peak on the HPLC chromatogram using a reverse-phase C-18 column and in the pattern of absorption spectrum of the peak. These changes were not observed in the cases of NaCl, Na isethi- onate, choline Cl, Na acetate or N-acetyl glycine. Another amino compound Tris, which was commonly used as a pH buffer, was unique in time course and induced the slow but parallel changes and reached maximal up to about 24 h. These results taken together indicate that the amino compounds markedly inhibit the activities of wortmannin presumably through the binding of wortmannin to amino group.
Keywords: Wortmannin; Chromaffin cell; Catecholamine secretion; PI3K; Phosphoinositide 3-kinase; Amino compound; Myosin light chain kinase
Phosphoinositide 3-kinase (PI3K) plays an important role in regulation of a variety of cell signalings such as cellular mobility, proliferation, survival, and carbohy- drate metabolism. Therefore, the compounds inhibiting this enzyme have potential as therapeutic drugs such as anticancer drugs and antiinflammatory agents [1]. To study in detail the aspect of the action of the PI3K inhib- itor wortmannin may provide information for these drugs.
Wortmannin, which was an antifungal antibiotic from Talaromyces wortmanni, was at first recognized as an inhibitor of myosin light chain kinase (MLCK) and was later found to be a more potent inhibitor against PI3K [2,3]. Since then, this compound has been widely used as a powerful tool to examine the role of MLCK and PI3K in a variety of intracellular signaling pathways.
It is known that wortmannin shows the inhibitory ef- fect on secretory responses in various types of cells. In mast cells and basophilic leukemia cells, the inhibitory effects on secretion have been reported to result from the inactivation of PI3K [4,5]. In neuronal cells, the inhi- bition of neurotransmitter release has been attributed to the inactivation of MLCK [6]. On the other hand, wort- mannin has been also shown to inhibit catecholamine secretion in bovine adrenal chromaffin cells. However, the inhibition of catecholamine secretion by wortman- nin has been reported to be due to the inactivation of MLCK, whereas the objection against this conclusion has been also reported [7,8]. Our original purpose was to study this discrepancy. In the course of such experi- ment, we found by accident that the inhibitory effects by wortmannin strongly depended on the composition of medium and were markedly prevented by the pres- ence of amino compounds.
Materials and methods
Dissociation and culture of bovine adrenal chromaffin cells. Bovine adrenal chromaffin cells were isolated by sequential collagenase digestions and were purified by a Percoll (Amersham Biosciences) gradient method as described previously [9]. The isolated cells were suspended in Dulbecco’s modified Eagle’s medium (Gibco-BRL) sup- plemented with 10 mM Hepes/NaOH (pH 7.4), 10% heat-inactivated newborn calf serum, 100 U/ml penicillin, 100 lg/ml streptomycin, 5 lM cytosine arabinoside, and 5 lM 5-fluoro-20-deoxyuridine and were cultured as a monolayer in 24-well plates at a density of 7 · 105cells/well [10].
Catecholamine secretion. Wortmannin or vehicle (0.5% DMSO) was pretreated for the indicated time with the media containing 20 mM Na-Hepes (pH 7.4) and the indicated concentrations of various com- ponents shown in figures. Cultured bovine adrenal chromaffin cells in 24-well plates were preincubated for 20 min with 0.5 ml of the various media pretreated with wortmannin or vehicle as above. The preincu- bation medium was then changed with 0.5 ml of permeabilization medium consisting of 150 mM Na glutamate, 10 mM Na-Pipes (pH 7.0), 2 mM MgATP, 5 mM EGTA or 5 mM EGTA + 4.75 mM CaCl2 (0.95 lM free Ca2+), 10 lM digitonin, and no wortmannin and the cells were then stimulated for 10 min at room temperature. Catechol- amines released into medium and cellular catecholamines were ex- tracted with 0.4 N perchloric acid and were estimated by the ethylenediamine condensation method using a fluorescence spectro- photometer (HITACHI 650-10S) with an excitation wavelength of 420 nm and an emission wavelength of 540 nm [11]. Secretion induced by Ca2+ stimulus was expressed as a percentage of total cellular cat- echolamines. Basal catecholamine release in the absence of free Ca2+ was 4–6% of total catecholamines and was subtracted from data. Wortmannin had little effect on the basal release. Because wortmannin itself has a weak but detectable fluorescence, and because wortmannin has the ability to react with ethylenediamine having amino group, the presence of wortmannin has a possibility to affect the measurement of catecholamine amounts. Therefore, wortmannin was omitted from the permeabilization medium into which catecholamines released. Although catecholamines also have amino group in the molecule, wortmannin did not affect catecholamine measurements in the per- meabilization medium containing Na glutamate.
In vitro PI3K assay. Recombinant human PI3K p110c (Alexis Biochem.) was used as the enzyme for the in vitro PI3K assay to test the inhibitory effect of wortmannin in various media. PI3K activities were measured according to manufacturer’s protocol with slight modifications. Wortmannin (100 nM) or vehicle (1% DMSO) was pretreated for the indicated time with 100 mM various media shown in figures. The pretreated media were then diluted (1:1) with 2· assay solution for the assay of PI3K activity. The assay solutions (10 ll) include 2 lg phosphatidylinositol, 50 ng PI3K, 20 mM Na-Hepes (pH 7.4), 4 mM MgCl2, 0.5 mM EGTA, and 50 mM various compounds pretreated with wortmannin or vehicle as above. Reaction was initi- ated by the addition of 20 lM [c-32P]ATP (1 lCi). After 15 min of incubation at room temperature (22 °C), the reaction was terminated with 100 ll of 1 N HCl. Lipid was then extracted with 100 ll chloro- form/methanol (2:1). The radioactivities in chloroform layer were measured by a liquid scintillation counter after the elimination of chloroform by evaporation.
Absorption spectrum. Wortmannin (100 lM) was treated for the indicated time with the media (100 ll) consisting of 20 mM Na-Hepes (pH 7.4) and 100 mM various compounds. Absorption spectrum (600– 200 nm) was then measured using UV1600 spectrophotometer (Shimadzu).
High performance liquid chromatography. Wortmannin (100 lM) was treated for 1 h with the media (100 ll) consisting of 20 mM Na- Hepes (pH 7.4) and 100 mM various compounds. Waters 2690 HPLC equipped with Waters 996 photodiode array detector (PDA) was used for the isolation and identification of wortmannin (and its derivatives). Analysis was carried out with a reverse-phase C-18 column (150 · 4.6 mm) with an acetonitrile/water gradient solution (acetoni- trile percentage, 0–100%; flow rate, 1 ml/min). As indicated by PDA profiles of elution, hydrophilic compounds (also Hepes buffer) that did not bind to wortmannin were easily eluted out during 0–4 min. The elution profiles at 254 nm was shown in figure. At this wavelength, only absorbance derived from wortmannin (and its derivatives) was detected.
Statistics. All experiments were repeated at least three times (means ± SE). The statistical significance was assessed by one-way or two-way analysis of variance followed by Fisher’s LSD test. The un- paired Student’s t test was used for comparisons between the two groups.
Materials. Wortmannin was purchased from Kyowa Medex (Tokyo, Japan). Other chemicals and materials were also obtained from commercial sources.
Results
Wortmannin has been reported to inhibit catechol- amine secretion from bovine adrenal chromaffin cells [7,8]. Surprisingly, in spite of these reports, wortmannin even at the higher concentration than that reported to induce almost complete inhibition had little effects on catecholamine secretion from the cells in our early at- tempt. In Na glutamate medium, which was usually used in the experiments in digitonin-permeabilized cells, wortmannin at the concentration of up to 100 lM failed to inhibit catecholamine secretion (Fig. 1A). However, interestingly, replacement of Na glutamate into NaCl, which was usually used in the experiments in intact cells, resulted in wortmannin-induced inhibition of catechol- amine secretion as reported previously. The inhibitory effect of wortmannin on catecholamine secretion was concentration-dependent, and a significant inhibition was observed at the concentrations of above 10 lM. Fig. 1B shows the effect of 100 lM wortmannin on cat- echolamine secretion from the cells in various concen- trations of Na glutamate. Na glutamate reversed wortmannin-induced inhibition in a concentration-de- pendent manner, whereas it had little effect on the con- trol secretion without wortmannin. Reversal by Na glutamate reached almost maximal at 100 mM.
From the above data, it seems that the effect of wort- mannin on catecholamine secretion depends on the com- position of extracellular fluids. To confirm this hypothesis, the effect of wortmannin on catecholamine secretion was examined in various media that have been frequently used in the experiments of the cells. As shown in Fig. 2, wortmannin-induced inhibition was found to be prevented by brief preincubation with the medium including amino group such as Na glutamate, Na as- partate, and lysine Cl, but not NaCl, Na isethionate,
choline Cl or Na acetate. But Tris Cl was exceptional and had little effect on wortmannin-induced inhibition under these experimental conditions, although this com- pound also has amino group. Thus, the effect of various preincubation times of wortmannin with Tris Cl on wortmannin-induced inhibition was investigated to examine whether Tris Cl as well as the other compounds having amino group has the ability to prevent wortman- nin activity. Prolonged preincubation of wortmannin with Tris Cl resulted in the prevention of wortmannin- induced inhibition as shown in other media containing amino group, whereas similar prolonged preincubation with NaCl did not cause significant changes in wortman- nin-induced inhibition (Fig. 3).
Wortmannin is well known to be a more specific inhibitor of PI3K and inhibits this enzyme at the lower concentrations than the case of the effect on catechol- amine secretion. In addition, the mechanism by which wortmannin inhibits catecholamine secretion is still un- clear. Thus, the effect of wortmannin on in vitro PI3K activity as another index of wortmannin activity was examined in various media (Fig. 4A). Wortmannin-in- duced inhibition on in vitro PI3K activity was markedly prevented by the medium including amino group such as Na glutamate, Na aspartate, and lysine Cl, but not NaCl, Na isethionate, choline Cl or Na acetate. More- over, reversal by Tris Cl required long-term preincuba- tion as in the case of catecholamine secretion. In addition, prolonged preincubation with NaCl did not cause significant changes in wortmannin-induced inhibition. These results are in good agreement with those shown in the case of catecholamine secretion. Here, further experiments were performed to study pre- cisely whether the inhibitory effect on wortmannin activ- ities results from the presence of amino group in medium components. As shown in Fig. 4B, glycine, which was also an amino derivative of acetate, pre- vented wortmannin-induced inhibition on PI3K activity, whereas N-acetyl glycine failed to prevent it, indicating that amino group in glycine molecule plays a key role in the prevention of wortmannin activity.
Another intriguing observation concerning wortman- nin medium was that a color of medium turned into yel- low after the incubation of wortmannin with Na glutamate, Na aspartate, lysine Cl, and Tris Cl that have amino group, whereas no alterations were observed in NaCl, Na isethionate, choline Cl, Na acetate that had little effect on wortmannin action. Thus, we measured the absorption spectrum (600–200 nm) of wortmannin in various media. The absorption spectra of the typical samples are shown in Fig. 5. Wortmannin itself had two absorption peaks (257 and 295 nm) in water (data not shown, but similar to the case of NaCl). Mixture of this compound with Na glutamate caused three major changes in an absorption spectrum as the following: (1) the wavelength shift to 319 nm of the absorption peak (295 nm); (2) the wavelength shift to 248 nm of the other absorption peak (257 nm) and an increase in the height of this peak; and (3) the induction of the new absorption peak at 408 nm (yellow in color). The reaction was rapid and these changes in the absorption peaks were already observed just after mixing of this compound with Na glutamate solution, and reached maximal at about 5 min (data not shown). Incubation of wortmannin with Na aspartate, lysine Cl, and glycine as well as Na gluta- mate resulted in the rapid changes in color of the med- ium and adsorption spectrum, whereas NaCl, Na isethionate, choline Cl, Na acetate, and N-acetyl glycine caused little changes (Fig. 5, data not shown). More- over, these changes in Tris Cl medium developed at a slow rate similar to the case of the effect on catechol- amine secretion and in vitro PI3K activity (Fig. 5), indi- cating that the changes in color of the medium and adsorption spectrum are closely correlated with de- creases in wortmannin activities.
Interaction of wortmannin with various compounds such as Na glutamate was further examined by HPLC analysis using C-18 reversed-phase column to examine whether wortmannin binds to the amino compounds, because wortmannin is well known to covalently bind to lysine residue of catalytic subunit of PI3K [12,13]. Wortmannin alone showed a peak with a retention time of about 8.5 min under the experimental conditions de- scribed in Materials and methods (data not shown, but similar to the case of NaCl). Addition of Na glutamate induced a new peak with a retention time of about 4.4 min, while the peak by wortmannin itself disap- peared almost completely (Fig. 6B). Moreover, PDA result shows that Na glutamate caused a change in UV–Vis absorption spectrum of the peak as shown in the inset of figure. On the other hand, NaCl changed neither peak position nor the pattern of absorption spec- trum (Fig. 6A). Similarly, Na isethionate, choline Cl, Na acetate or N-acetyl glycine changed neither peak posi- tion nor the pattern of absorption spectrum, whereas Na aspartate, lysine Cl, and glycine induced new peaks at 4.4, 6.0, and 5.0 min, respectively (data not shown). In accordance with the above experiments, long-term preincubation was necessary for Tris-induced changes in peak position and the pattern of absorption spectrum (data not shown). These results indicate that wortman- nin binds to the amino compounds in a substantially irreversible manner.
Discussion
The amino compounds such as glutamate, aspartate, lysine, and Tris caused several changes in wortmannin medium; prevention of wortmannin-induced inhibition in catecholamine secretion and in vitro PI3K activity; the changes in color and absorption spectrum of the medium; and the changes in the retention time of the peaks on the HPLC chromatogram and in the pattern of absorption spectrum of the peak. Tris is unique in time course and it required about 24 h to cause the max- imal changes as described above. In contrast, NaCl, Na isethionate, choline Cl, and Na acetate had little effects on the above-described wortmannin-induced inhibition, absorbance spectrum or chromatogram. Thus, the pre- vention of wortmannin-induced inhibition on catechol- amine secretion and in vitro PI3K activity induced by the amino compounds was always parallel to the changes in the retention time of the peak of the chro- matogram and in absorption spectra of the peaks.
It is well known that the electrophilic C21 site of wortmannin irreversibly and covalently binds to the primary amine of lysine residue of the PI3K catalytic subunit and inhibits the enzyme [12–14]. In the present study, glycine (aminoacetic acid) as well as the other tested amino acids also caused the changes in wortman- nin medium; prevention of wortmannin-induced inhibi- tion of in vitro PI3K activity; the changes in color and absorption spectrum of the medium; and the changes in the retention time of the peaks on the HPLC chro- matogram and in the pattern of absorption spectrum of the peak, whereas acetate (deamino analogue of gly- cine) and N-acetyl glycine failed to cause the changes (Figs. 4–6). The mechanism by which the introduction of acetyl group into amino group of glycine resulted in the depression of the binding of this amino group into wortmannin remains still uncertain. But one possible explanation is that the introduction of the electron-attracting group (acetyl group) reduced nucleophilicity of the amino group of glycine, which resulted in a decrease in electrophilic attack of wortmannin to glycine. An- other possibility is that this introduction changes the primary amine of glycine into secondary amine. Although the mechanism remains to be clarified, these results strongly indicate that the prevention of wortman- nin action results from the presence of amino group. Therefore, it can be speculated that wortmannin binds to the amino group of the medium components in an irreversible manner such as covalent binding and rapidly loses its activity to inhibit the enzymes such as PI3K.
Wortmannin has been believed to be unstable in aqueous solution as frequently described [15]. In the present study, this compound in the tested solution without amino group was found to be rather stable (T1/2 > 24 h) in the experiments of catecholamine secre- tion, in vitro PI3K activity assay, medium color, absorp- tion spectra, and chromatogram (Figs. 3–5). However, the presence of the amino compounds resulted in a rapid degradation of wortmannin activities.
Baker and Knight [16] developed the electropermea- bilized cell technique to control the intracellular chemi- cal environments such as Ca2+ concentrations. In leaky cells, glutamate ion was found to be more effective in maintaining Ca2+-dependent exocytosis than Cl ion, which was usually used in the intact cells. Since then, glutamate ion has been preferably utilized in the exper- iments using leaky cells. However, caution should be exercised in the experiments using wortmannin, because the action of wortmannin is markedly prevented by the medium component glutamate. As another example, in addition to the common usage as a pH buffer, Tris has been frequently used as the substituent for Na+, when Na+ in medium is omitted [17,18]. On the other hand, the covalent binding of wortmannin to amino group has already been reported and the several derivatives of wortmannin were synthesized by the incubation of wortmannin with amino compounds in the attempt to obtain novel drugs [19]. But, in the present study, wort- mannin was shown to rapidly bind to the amino com- pounds in usual experimental and physiological conditions. Therefore, in addition to the medium com- ponents, at least, care should also be taken in the exper- iments using or measuring chemical transmitters and hormones such as amines, amino acids or peptides, which have amino group, because wortmannin might di- rectly affect the measurement system itself.
Compared to the amino acids such as glutamate, Tris has proved to slowly bind to wortmannin, although the maximal degree was comparable to the case of the other amino compounds. The reason why the reaction velocity is slow in the case of Tris still remains uncertain, but this may be related with that the carbon atom next to the nitrogen atom of amino group is of quaternary struc- ture, which might lead to a steric hindrance. In support of this assumption, alanine (methyl derivative of glycine) as well as the other amino acids including glycine was fast in reaction velocity, while 2-methylalanine (2-ami- noisobutyric acid; dimethyl derivative of glycine) was slow and comparable with the case of Tris (data not shown), although methyl group was electron-releasing. In addition, other Tris analogues such as 2-amino-2- methyl-1-propanol and 2-amino-2-methyl-1,3-propane- diol, which was of quaternary structure, was also slow in reaction velocity (data not shown), being consistent with this assumption. But it requires further examina- tions to provide concrete evidence.
In conclusion, these results taken together indicate that amino compounds markedly inhibit the activities of wortmannin presumably through the binding of wort- mannin to amino group.