Human salivary gustin is a potent activator of calmodulindependent

Transcription

1 Proc. Nati. Acad. Sci. USA Vol. 84, pp , March 1987 Medical Sciences Human salivary gustin is a potent activator of calmodulindependent brain phosphodiesterase (cyclic nucleotides/taste) JOSEPH S. LAW*, NETA NELSON*, KouICHI WATANABEtt, AND ROBERT I. HENKIN* *Center for Molecular Nutrition and Sensory Disorders, Georgetown University Medical Center, Washington, DC 20007; and tdepartment of Pharmacology, Howard University College of Medicine, Washington, DC Communicated by George K. Davis, November 5, 1986 (received for review September 16, 1986) ABSTRACT Human salivary gustin stimulated activity of brain calmodulin-dependent cyclic nucleotide phosphodiesterase (camp PDEase; 3',5'-cyclic-nucleotide phosphodiesterase, EC ) in a dose-dependent manner in the absence of calmodulin. At physiological levels found in human saliva, gustin activated camp PDEase 5- to 6-fold. Activation of PDEase occurred with as little as 500 ng of gustin. Comparative sensitivity of activation of PDEase by gustin was intermediate between calmodulin and lysophosphatidylcholine with maximal activation and half-maximal activation (indicated in parentheses) at 3 x 10-8 M (4.3 x 10-9 M), 3.4 x 10-6 M (3.4 x 10-7 M), and 2.5 x 10-3,M (4.0 x 10-5 M) for calmodulin, gustin, and lysophosphatidylcholine, respectively. No other major salivary protein activated PDEase. Anticalmodulin antibody completely inhibited calmodulin-activated camp PDEase activity, but the antibody had no effect on gustinactivated camp PDEase activity. A sensitive calmodulin RIA indicated that no calmodulin was detected in any gustin preparation that activated camp PDEase. Both gustin and calmodulin rendered camp PDEase thermally labile to a similar extent and increased V. without affecting the apparent Km for the substrate camp. Activation by gustin and calmodulin was unaffected by lubrol-px, trypsin inhibitor, pepstatin A, or leupeptin. In the presence of 1 mm EGTA, gustin activated camp PDE 5- to 6-fold, but the activating ability was completely lost after gustin was heated at 100'C for 5 min. In contrast, calmodulin lost all activating ability in the presence of 1 mm EGTA, whereas heating calmodulin at 1000C for 5 min did not affect its activation of camp PDEase. Lysophosphatidylcholine-activation of camp PDEase, like gustin activation, was unaffected by EGTA, but lysophosphatidylcholine-activation of camp PDEase, like calmodulin activation, was unaffected by heating at 100'C for 5 min. Gustin, the major zinc-containing protein in human parotid saliva, contributes -3% of the total parotid salivary protein, has a Mr of 37,000 Da (1), and contains one mole of tightly bound zinc per mole of protein (2) and a second mole of loosely bound zinc that can be removed by dialysis and replaced by other metal cations, including copper and cobalt (2). Gustin has been implicated in the growth and nutrition of taste buds (3), has physiological and biochemical properties similar to nerve growth factor (NGF) (3) and specifically displaces nerve growth factor from its binding sites on purified taste bud membranes (4). Studies in patients with hypogeusia, decreased taste acuity, showed that gustin concentration in some of these salivas was much lower than normal and that treatment with zinc, which increased saliva gustin concentration, returned taste function to or toward normal (5). Increasing evidence suggests that camp plays a role in the transduction of taste information (6-8). camp phosphodiesterase (PDEase; 3',5'-cyclic-nucleotide phosphodiesterase, EC ) has also been reported to be involved in taste function (9-12), specifically associated with taste bud membranes (12) and is inhibited by zinc, at relatively high concentrations (13), and thyroxine (14), both of which are required to maintain normal taste function (5, 15, 16). Previous studies have revealed that gustin has no enzymatic activity as a protease, esterase, or phosphatase (2). However, the putative role of camp PDEase in taste function led us to ask whether gustin might play some enzymatic role in this system. To test this hypothesis the effect of gustin on partially purified bovine brain camp PDEase was investigated, and the results demonstrated that gustin, indeed, activated camp PDEase. MATERIALS AND METHODS Purification of calmodulin-dependent camp PDEase from bovine brain was done using chromatography on DEAE- Sephadex and Affi-Gel Blue columns (17, 18). Briefly, bovine brain was homogenized with 25 mm 2-[N-morpholino]- ethanesulfonic acid (Mes), ph 6.7/0.5 mm MgCl2/4 mm EGTA/25% (vol/vol) ethylene glycol (buffer A). After washing the column with 10 volumes of buffer A, camp PDEase was eluted with buffer A and a linear gradient of NaCl (0-2 M). camp PDEase was eluted at =1.5 M NaCl, and the yield was 0.27%. NaDodSO4/PAGE indicated that this protein contained two major bands each having a Mr slightly greater than 55,000 and two minor bands each having Mr >92,000. The eluted camp PDEase was not stimulated by addition of 0.1 mm Ca2' but was stimulated 5- to 6-fold over basal activity with 0.1 mm Ca2' and 58.4 nm calmodulin. camp PDEase activity was measured as described previously (19). The standard incubation mixture contained 100,.g of bovine serum albumin (BSA), 50 mm Tris HCl at ph 7.4, 5 mm MgCl2, 1 gm unlabeled camp, ,uCi [H3]cAMP (1 Ci = 37 GBq), and an estimated 10-20,ug of partially purified calmodulin-dependent PDEase solution in a final volume of 200,ul. To some reaction mixtures various proteins and substances were added, and the mixtures were incubated for 12 min at 350C with gentle shaking. The PDEase assay was completed as described previously (19). The PDEase activity of the partially purified protein ranged from pmol per min per mg of protein; mean ± SEM was 1905 ± 366 pmol from 12 separate, duplicate preparations. When substrate concentration in our system was increased to 1000 um, PDEase activity increased to >200 nmol per min per mg of protein-values similar to those previously reported (20). Effects of each protein or substance added were studied in The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C solely to indicate this fact Abbreviations: BSA, bovine serum albumin; PDEase, phosphodiesterase. tpresent address: Fujimoto Diagnostics Inc., Osaka, 580 Japan.

2 Medical Sciences: Law et al. two to nine experiments on at least two separate occasions. For ease of comparison, data are presented as a percent of control camp PDEase activity determined in each experiment. Statistical significance of differences between the different conditions was assessed by Student's t test. Gustin was purified from saliva, as previously described, by chromatography on Sephacryl S-200, then by chromatography on DEAE-Sephadex A-50, and finally by chromatography on CM-cellulose (1, 2). Lysophosphatidylcholine (Type I) was obtained from Sigma. Lubrol-PX was also obtained from Sigma. Purification of calmodulin was done by modification of the method of Kakiuchi et al. (21) using a fluphenazine-coupled epoxy-activated Sepharose column. Calmodulin was identified as a single band following electrophoresis in a 12.5% acrylamide gel containing 0.1% NaDodSO4. Protein was determined by the method of Lowry et al. (22) with BSA as a standard. Calmodulin was measured by RIA with a kit obtained from New England Nuclear; the technique utilizes a competitive inhibitory system with sufficient sensitivity to detect 0.3 ng of calmodulin. RESULTS Partially purified bovine brain PDEase was stimulated in a dose-dependent manner by gustin (Fig. 1). Maximal stimulation of PDE activity was seen at about 25,ug of gustin; at this concentration camp PDEase was activated %. Half-maximal stimulation was obtained at about 2.5,ug of gustin (0.34,M). From a Hill plot (Fig. 1 Inset) a Hill coefficient (NH = 1.8) was calculated that indicates positive cooperativity. Effects of other salivary proteins and zinc on camp PDEase activation were compared with that of gustin (Table 1). At physiological concentrations found in human saliva, both gustin (3.4 x 10-6 M) and calmodulin (3 x 10-8 M) maximally stimulated PDEase activity (481% vs. 506%, respectively), whereas other major salivary proteins such as amylase (23) and lumicarmine (24) had no effect even at 1.2 x 10-5 M and 5 x 10-4 M, respectively. In contrast, both thyroxine (1 x 10-4 M) and ZnCl2 (1 x 10-4 M), which play T Proc. Natl. Acad. Sci. USA 84 (1987) 1675 Table 1. Effects of various salivary components on partially purified bovine brain camp PDEase Component M activity, %* None 100 Gustin 3 x ± 32 Calmodulin 3 x ± 54 Amylase 1 x ± 3 Lumicarmine 5 x ± 10 Thyroxine 1 x ± 9 Zn2+ 1 X ± 6 *Activity determined in presence of 100 jug BSA and expressed as percent of control mean 1905 ± 366 (SD) pmol per min per mg of protein as determined in triplicate in nine experiments; relative important roles in taste function (5, 15), significantly inhibited PDEase activity (P < 0.01). Because gustin activated camp PDEase at a concentration -100-fold that of calmodulin, any slight calmodulin contamination could yield spurious results. To evaluate this possibility, anticalmodulin antibody, which completely blocked the activation of camp PDEase by 250 ng of calmodulin, was added to duplicate incubation mixtures containing 250 ng of calmodulin in three separate experiments; no activation of PDEase occurred in any of these studies-calmodulin activation was 424% ± 35% (mean ± SD) and calmodulin activation with antibody was 108% ± 11% (where control activity was considered as 100%). In another group of three experiments done in duplicate, 10-25,ug of gustin was added with the same amount of anticalmodulin antibody that had blocked calmodulin activation of camp PDEase in the previous experiments; no change in activation of camp PDEase by gustin was noted in any sample; the gustin activation measured 262% ± 20% of control values and with anticalmodulin antibody measured 255% ± 50%. No calmodulin was detected in any of 10 samples of gustin (20-90,ug) analyzed for calmodulin by RIA, although the gustin in each sample activated camp PDEase from 300 to 600%; 90,tg gustin was the highest concentration used. In chromatograms of purified saliva or gustin from Sephacryl S-200 or CM- DC 400 1NH = u 200 In: a 0~~~~~~.0_ r / ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I ) Gustin,,tg 50 acu) a FIG. 1. Activation of purified bovine brain camp PDEase by gustin at various concentrations. Results are reported as means of duplicates of five experiments at concentrations of 1-25,g and two experiments at 50 jig. Lines above and below solid circles indicate ±1 SEM; circles without lines indicate a SEM within the experimental mean. Studies at high concentrations were limited by the availability of gustin. Inset, Hill plot comprised of logarithmic activation of camp PDEase data on the ordinate and gustin concentration in jug on the abscissa, and the line is drawn by the least-squares method from the data obtained; the Hill coefficient, NH, was 1.8. FIG. 2. Effect of (A) gustin (25 jug, hatched bar), (B) calmodulin (100 ng, dotted bar), and (C) gustin (25 jug) and calmodulin (100 ng) together (hatched and dotted bar) on camp PDEase activity (as % of control). Bars indicate mean ± SD of three experiments done in duplicate.

3 1676 Medical Sciences: Law et al. Table 2. Effect of BSA on gustin-activated camp PDEase activity Compound Ag activity, %* Control 100 BSA ± 17 BSA ± 37 Gustin ± 33 Gustin + BSA ± 23 Gustin + BSA ± 27 *Five determinations in two to four experiments. Relative activities are reported as mean values ± SEM. cellulose columns, protein that activated camp PDEase was found only in the gustin peak (1, 2) and not in any other fraction; eluates from two Sephacryl S-200 and CM-cellulose columns were measured. Effects of gustin and calmodulin together at concentrations at which each alone maximally stimulated camp PDEase were not additive (Fig. 2). The partially purified camp PDEase used in these studies was incubated with and without BSA as a stabilizing agent. To ascertain whether activation of camp PDEase by gustin was due to a possible stabilizing effect of BSA, the camp PDEase activation was also studied in the presence of excess BSA (200,g). Nearly maximal activation of camp PDEase still occurred with 180 Ag of gustin in the presence of this large excess of BSA (Table 2). The effects of divalent ion chelators on activation of camp PDEase by various known activators was also examined (Table 3). In contrast to activation by calmodulin, activation by gustin was unaffected by 1 mm EGTA. Whereas activation of camp PDEase by lysophosphatidylcholine was completely abolished by 0.1% lubrol-px, activation by both calmodulin and gustin was not significantly (P > 0.10) affected by this detergent (Table 4). Because calmodulin and lysophosphatidylcholine are both thermally stable, the influence ofboiling on gustin and these known activators of camp PDEase was compared. In contrast to calmodulin and lysophosphatidylcholine, activation by gustin was completely abolished by heating at 100'C for 5 min (Table 5). Because calmodulin and lysophosphatidylcholine have been shown to decrease the stability of calcium-dependent PDEase (25, 26), the effect of gustin on thermal denaturation of camp PDEase was examined. Exposure ofcamp PDEase to 580C for 10 min in the presence of gustin resulted in >90% decrease in its activation (Fig. 3). The thermal denaturation curve of 10 pag of gustin was similar to that of 100 ng of calmodulin (Fig. 3). The effect of gustin on camp PDEase activity was also examined by using different concentrations of camp in the presence of fixed amounts of gustin and calmodulin. Control camp PDEase activity had an apparent Km of 30,uM and Table 3. Activity of bovine camp PDEase in the presence of gustin, calmodulin, and other substances Substance M activity, %* None 100 Gustin 3 x ± 29 Calmodulin 3 x ± 32 EGTA 1 x ± 8 EDTA 5 x ± 2 Gustin + EGTA 3 x X ± 17 Calmodulin + EGTA 3 x x ± 18 *Four determinations in two to four separate experiments. Relative Proc. Natl. Acad. Sci. USA 84 (1987) Table 4. Effects of lubrol-px on activation of bovine brain camp PDEase by gustin, calmodulin, and lysophosphatidylcholine activity, %* Activator M Control Lubrol-PX, 0.1% None ± 11 Gustin 3 x ± ± 70 Calmodulin 3 x ± ± 40 Lysophosphatidylcholine 2 x ± 9 83 ± 16 *Four determinations in two to four separate experiments. Relative Vmax of 2000 pmol of camp hydrolyzed per min per mg of protein (Fig. 4). In the presence of 5 gg of gustin, Vmax was about twice that without gustin, whereas Km was unchanged (Fig. 3). Similarly, 25 ng of calmodulin did not affect Km but raised Vmax to 5000 pmol of camp hydrolyzed per min per mg of protein. It has been well established that various proteolytic enzymes activate camp PDEase (27-31); therefore, possible proteolytic activation of camp PDEase by gustin was examined by studying effects of various protease inhibitors. Neither the peptide protease inhibitor leupeptin (32) nor the carboxyl protease inhibitor pepstatin A (32) had any effect on gustin activation of camp PDEase (Table 6). Trypsin inhibitor was also ineffective in blocking gustin activation of the enzyme (Table 6). DISCUSSION In some ways activation of camp PDEase by gustin is similar to that by calmodulin. Although the concentration of gustin required to activate the enzyme is two orders of magnitude higher than for calmodulin, activation curves indicate similar dose-response characteristics. Gustin activates camp PDEase at the physiological concentration at which it is found in normal human parotid saliva (25 ± 6,ug per 200,ul of saliva). Both gustin and calmodulin rendered the enzyme less thermally labile to a similar extent and increased Vmax without change in apparent Km for substrate camp. Activation by both proteins was completely abolished by addition of 5 mm of EDTA and was not restored by addition of excessive amounts of Mg2+ (10 mm). Furthermore, activation by gustin and calmodulin was unaffected by lubrol-px, trypsin inhibitor, pepstatin A, leupeptin, dithiothreitol, and 2-mercaptoethanol. In other ways, gustin activation of camp PDEase differs from that of calmodulin. Activation of the enzyme by calmodulin is calcium dependent, whereas that by gustin is not. Calmodulin is heat stable, whereas gustin is heat labile and its ability to activate PDEase is lost by heating at 100 C for 5 min. Table 5. Effect of boiling on activators of bovine brain camp PDEase activity, %* Activator M Control Heatt None Gustin 3 x ± ± 15 Calmodulin 3 x ± ± 10 Lysophosphatidylcholine 2 x ± ± 57 *Three to six determinations from two to four experiments. Relative tboiling at 100 C for 5 min.

4 0 * Time, min at 58 C FIG. 3. Effect ofgustin (e) and calmodulin (o) on thermal stability of camp PDEase. Ten-microgram samples of partially purified camp PDEase with or without 10,g of gustin or 100 ng of calmodulin were incubated at 58 C for various times. At the end of different incubation times all samples were transferred to a 4 C ice bath. To these tubes 80-,ul standard incubation mixture was added, and residual camp PDEase activity was determined in the presence of 100 jig of BSA. o, Control camp PDEase activity exposed to heat but without additions. Results are duplicates of at least two separate experiments. L-._ L- 0. o 20 X 10 8 "a o N -4> 'a c; C" 2 u -C 4.t c) Medical Sciences: Law et al. v: lox 1/S. M-1 FIG. 4. Double reciprocal plot of camp hydrolysis by PDEase with or without 5,ug of gustin or 25 ng of calmodulin. Substrate camp was varied from 1-10 jim. PDEase activity was determined in the presence of 100,ug of BSA. o7, Control camp PDEase activity; *, addition of gustin; and o, addition of calmodulin. Km values for all studies were the same; V,,, for control was 2000 pmol per min per mg of protein, and Va, was 4000 for gustin and 5000 for calmodulin. Results are duplicates of at least two separate experiments. S, camp. 60 Proc. Natl. Acad. Sci. USA 84 (1987) 1677 Table 6. Effects of various enzyme inhibitors on activation of bovine brain PDEase by gustin and calmodulin Compound Conc. activity, %* None 100 Gustin 25,g 428 ± 64 Calmodulin 100 ng 420 ± 60 Trypsin inhibitor 5 /.g 106 ± ± ± 55 Pepstatin A 0.4 Ag 102 ± ± ± 62 Leupeptin 50 /hm 98 ± ± ± 62 *Five to seven determinations from two to four experiments. Gustin (25 Ag) or calmodulin (100 ng) was added to each inhibitor. Relative Activation of camp PDEase by gustin was not due to the presence of small amounts of calmodulin because anticalmodulin antibody completely inhibited calmodulin activation but had no effect on gustin activation. In addition, in those gustin preparations that maximally activated camp PDEase, there was no detectable calmodulin. Activation of camp PDEase by gustin was also not due to stabilization of the enzyme or to any proteolytic effect because excessive amounts of BSA and various protease inhibitors had no effect on gustin-activated PDEase. Activation by gustin exhibited no thiol requirement, as neither dithiothreitol nor 2-mercaptoethanol influenced activation (data not shown). Some salivary compounds-e.g., gustin, zinc, and thyroxine-affected camp PDEase, whereas others, such as amylase and lumicarmine, which have no effect on taste, had no effect on this enzyme activity. Gustin levels have been shown to be depressed in saliva of some patients with hypogeusia (5) and associated with low levels of total zinc in parotid saliva (33) and with pathological changes in taste buds (34). Treatment of some of these patients with zinc increased gustin levels in saliva and was associated with improvement in taste function (5). Thus, the observation that gustin activates PDEase at physiological levels found in human saliva suggests that this activation may be one mechanism by which gustin influences taste function. These studies also emphasize the importance of camp PDEase in the regulation of taste function. The roles that gustin plays in taste may be complex and involved at more than one level of cellular organization. In studies of receptor-rich taste bud membranes, gustin increased camp concentration (35), whereas in the present system it stimulated camp PDEase. These initially appearing divergent effects may relate to the complex role that gustin plays in the modulation of camp activity in taste. 1. Henkin, R. I., Lippoldt, R. E., Bilstad, J. & Edelhoch, H. (1975) Proc. Natl. Acad. Sci. USA 72, Shatzman, A. R. & Henkin, R. I. (1980) Biochim. Biophys. Acta 623, Henkin, R. I. (1978) in Zinc and Copper in Clinical Medicine, eds. Hambidge, K. M. & Nichols, B. L. (Spectrum, New York), pp Lum, C. K. L. & Henkin, R. I. (1977) Clin. Res. 5, 497A (abstr.). 5. Shatzman, A. R. & Henkin, R. I. (1981) Proc. Natl. Acad. Sci. USA 78, Kurihara, K. & Koyama, N. (1974) Biochim. Biophys. Acta 291, Cagan, R. H. (1976) J. Neurosci. Res. 2, Henkin, R. I. & Shatzman, A. R. (1981) Clin. Res. 29, Kurihara, K. (1972) FEBS Lett. 27,

5 1678 Medical Sciences: Law et al. 10. Price, S. (1973) Nature (London) 241, Law, J. S. & Henkin, R. I. (1982) Clin. Res. 30, 410A (abstr.). 12. Law, J. S. & Henkin, R. I. (1982) Res. Commun. Chem. Pathol. Pharmacol. 38, Law, J. S. & Henkin, R. I. (1983) Res. Commun. Chem. Pathol. Pharmacol. 41, Law, J. S. & Henkin, R. I. (1984) Res. Commun. Chem. Pathol. Pharmacol. 43, McConnell, R. J., Menendez, C. E., Smith, F. R., Henkin, R. I. & Rivlin, R. S. (1979) Am. J. Med. 59, Henkin, R. I. (1975) in Anatomical Neuroendocrinology, eds. Stumpf, W. E. & Grant, L. D. (Karger, Basel), pp Klee, C. B. & Krinks, M. H. (1978) Biochemistry 17, Morrill, M. E., Thompson, S. T. & Stellwagen, E. (1979) J. Biol. Chem. 254, Boudreau, R. J. & Drummond, G. I. (1979) Anal. Biochem. 63, Lin, Y. M., Liu, X. P. & Cheung, W. Y. (1974) J. Biol. Chem. 249, Kakiuchi, S., Sobue, K., Yamazaki, R., Kambayashi, J., Sakou, M. & Kosaki, G. (1981) FEBS Lett. 226, Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1951) J. Biol. Chem. 193, Agarwal, R. P. & Henkin, R. I. (1984) Metabolism 33, Proc. Natl. Acad. Sci. USA 84 (1987) 24. Shatzman, A. R. & Henkin, R. I. (1983) Biochem. Med. 29, Wolff, D. J. & Brostrom, C. 0. (1976) Arch. Biochem. Biophys. 173, Pichard, A. L. & Cheung, W. Y. (1977) J. Biol. Chem. 252, Cheung, W. Y. (1967) Biochem. Biophys. Res. Commun. 29, Cheung, W. Y. (1971) J. Biol. Chem. 246, Wells, N. N. & Hardman, J. G. (1977) Adv. Cyclic Nucleotide Res. 8, Moss, J., Manganiello, V. C. & Vaughan, M. (1978) Biochim. Biophys. Acta 541, Sakai, T., Makino, H. & Tanaka, R. (1978) Biochim. Biophys. Acta 522, Umezawa, H. & Aoyagi, T. (1977) in Proteinases in Mammalian Cells and Tissues, ed. Barrett, A. (North-Holland, Amsterdam), pp Henkin, R. I., Mueller, C. & Wolf, R. (1975) J. Lab. Clin. Med. 86, Henkin, R. I., Schechter, R. J., Hoye, R. C. & Mattern, C. F. T. (1971) J. Am. Med. Assoc. 217, Henkin, R. I. (1983) Biol. Trace Elem. Res. 6,