A Calcium-Dependent but Calmodulin-Independent Protein Kinase from Soybean' Received for Publication September 3, 1986 and in Revised Form November 17, 1986

Plant Physiol. (1987) 83, 830-837 0032-0889/87/83/0830/08/$01.00/0

A Calcium-Dependent but Calmodulin-Independent Protein Kinase from Soybean' Received for publication September 3, 1986 and in revised form November 17, 1986

ALICE C. HARMON*, CINDY PUTNAM-EVANS, AND MILTON J. CORMIER Department ofBiochemistry, University ofGeorgia, Athens, Georgia 30602

ABSTRACI kinase from suspension-cultured soybean cells (22). This CDPK2 was completely inhibited by 1 mM W-7, a calmodulin-binding A calcium-dependent protein kinase activity from suspension-cultured compound, and was activated a slight amount (30%) by micro- soybean cells (Glycine max L. Wayne) was shown to be dependent on molar concentrations of calmodulin (22). Because the CDPK calcium but not calmodulin. The concentrations offree calcium required preparation was contaminated with calmodulin and because such for half-maximal histone H1 phosphorylation and autophosphorylation high concentrations ofW-7 and calmodulin were required for an were similar (-2 micromolar). The protein kinase activity was stimulated effect, we were unable to conclude whether calmodulin was the 100-fold by >10 micromolar-free calcium. When exogenous soybean or regulator ofCDPK. To demonstrate whether CDPK is dependent bovine brain calmodulin was added in high concentration (1 micromolar) calcium and activity, we have to the purified kinase, calcium-dependent and -independent activities on calmodulin for sensitivity were weakly stimulated (<2-fold). Bovine serum albumin had a similar purified the enzyme 1000-fold. We show here that CDPK directly effect on both activities. The kinase was separated from a small amount binds calcium and is not dependent on calmodulin for activity. of contaminating calmodulin by sodium dodecyl sulfate polyacrylamide gel electrophoresis. After renaturation the protein kinase autophosphor- MATERIALS AND METHODS ylated and phosphorylated histone H 1 in a calcium-dependent manner. Following electroblotting onto nitrocellulose, the kinase bound 45Ca2+ in Chemicals. W-7 and W-5 were synthesized according to Hi- the presence of KCI and MgCl2, which indicates that the kinase itself is daka et al. (9) and Hart et al. (8). Magnesium chloride (Analar a high-affinity calcium-binding protein. Also, the mobility of one of two grade) from BDH Chemical Co. and 100 mm stock calcium kinase bands in SDS gels was dependent on the presence of calcium. solution from Orion were used in calcium buffers. 45CaC12 (335 Autophosphorylation of the calmodulin-free kinase was inhibited by the Ci/mol) was obtained from ICN and [,y-32P]ATP was from calmodulin-binding compound N-(6-aminohexyl)-5-chloro-I-naphthal- Amersham. ene sulfonamide (W-7), showing that the inhibition of activity by W-7 is Plant Material. Soybean (Glycine max L. Wayne) cell suspen- independent of calmodulin. These results show that soybean calcium- sion cultures were originally obtained from Dr. Joe L. Key, dependent protein kinase represents a new class of protein kinase which Department of Botany, University of Georgia. The cells were requires calcium but not calmodulin for activity. subcultured weekly and grown in the dark at 30°C and 175 rpm in Murashige and Skoog medium (16) supplemented with 20 g/ L sucrose, 1 mg/L naphthaleneacetic acid, and 0.5 mg/L kinetin. Proteins. Calcium-dependent protein kinase was purified One of the ways in which Ca2" may play a second messenger 1000-fold (0.9 umol 32p transferred per min per mg of enzyme) role in plants is to activate protein kinases in response to stimuli from suspension-cultured soybean cells by chromatography on such as growth regulators, light, or stress (5). The receptor and DEAE-Cellulose, TAPP-Sepharose, Sephadex G-100, and Ciba- transducer of the calcium signal could be the Ca2"-binding cron Blue-Sepharose. Procedures for chromatography on the first protein calmodulin or the protein kinase itself. There have been three resins have been described (22). The purification and several reports ofcalcium/calmodulin-dependent protein kinases characterization of CDPK will be described elsewhere. (1, 20, 21, 23, 25, 31) and calcium/phospholipid-dependent Apoaequorin was a gift of Miss Judy C. Peek ofour laboratory. protein kinases (17, 26) in plants. Most of the evidence for Sheep anticalmodulin IgG was a gift of Dr. Anthony Means, calmodulin dependence has been indirect, i.e. based on inhibi- Baylor College of Medicine. Bovine brain calmodulin was pre- tion studies with calmodulin antagonists and on activation stud- pared by the method of Charbonneau et al. (4). Soybean cell ies with exogenous calmodulin performed with impure protein calmodulin co-purified with CDPK through the first three chro- kinase preparations. None of these protein kinases has been matography steps. Calmodulin was separated from CDPK by purified to homogeneity and their interaction with calmodulin chromatography on Cibacron Blue-Sepharose. Calmodulin did demonstrated. not bind to the resin and was recovered in the flow-through Previously we partially purified a calcium-dependent protein fractions. This calmodulin was -95% pure as shown by SDS- PAGE and was able to fully activate the calmodulin-dependent enzymes pea NAD kinase and human erythrocyte Ca2t-ATPase ' Supported by grants from the National Science Foundation (PCM (data not shown). 8213177) and the United States Department of Agriculture Competitive Pea seedling NAD kinase was purified through the carboxy- Research Grants Program (86-CRCR-1-1969). methyl cellulose step according to Harmon et al. (7). Histone H 1 2Abbreviations: CDPK, calcium-dependent protein kinase; W-7, N- (Sigma Chemical Co., type III-S) was exhaustively dialyzed (6-aminohexyl)-5-chloro-l-naphthalene sulfonamide; W-5, N-(6-ami- against deionized H20 to remove contaminating calcium. nohexyl)- -naphthalene sulfonamide; TAPP, 2-trifluoromethyl-1OH- 10- Protein Determination. The concentration of a stock solution (3' aminopropyl) phenothiazine; M,, relative mass; Ko.5, concentration of soybean calmodulin in deionized H20 was determined from required for 50% activation; IC50, concentration required for 50% inhi- amino acid analysis of a 24 h hydrolysate. The amino acid bition. composition of spinach calmodulin (14) was used for the calcu- 830 CALCIUM-DEPENDENT, CALMODULIN-INDEPENDENT PROTEIN KINASE 831 lation of concentration. The concentrations of all other protein Detection of Protein Kinase Autophosphorylation and Activity solutions were determined with a Bio-Rad protein assay kit which in SDS-Polyacrylamide Gels. A modification of the method of is based on the method of Bradford (2). Geahlen et al. (6) was used to detect protein kinase autophos- Enzyme Assays. The free calcium concentration in the protein phorylation in SDS-polyacrylamide gels. The samples were kinase and autophosphorylation assays was controlled by a Ca2+/ heated to 80 to 90°C for 2 min in electrophoresis sample buffer EGTA buffer (final concentrations: 0.2 mM EGTA, 10 mM (13) that contained SDS and 2-mercaptoethanol. The samples MgCl2, 50 mm Hepes [pH 7.2], and various amounts of calcium were electrophoresed as described (13) except that 0.1 mg/ml stock solution). The resulting free calcium concentrations were BSA was included in the separating gel before polymerization calculated by a computer program based on the method ofPerrin and the stacking gel was photopolymerized with 0.5 mg/L ribo- and Sayce (19). The stability constants reported by Sillen and flavin as catalyst. After removal of the SDS with five changes of Martell (29) for metal-chelator and metal-ATP complexes were 50 mM Hepes (pH 7.2), the gel was cut into vertical strips, each used in the calculations. containing CDPK and Mr markers. The strips were incubated in Protein kinase activity was determined by measuring incor- 50 ml Ca2+/EGTA buffer, 30 Mm [y-32PJATP (100-200 cpm/ poration of labeled phosphate from [y-32P]ATP into histone H 1. pmol) and either no added calcium, 13 gM-free calcium, or 13 The standard assay mixture contained, in a volume of 0.15 ml, Mm-free calcium and 500 Mm W-7. The slices were washed and 1 mg/ml histone H 1, 470 ng/ml CDPK, and Ca2+/EGTA buffer fixed as described by Geahlen et al. (6). and the reaction was initiated by the addition of 30 uM [ey32p] For measurement of histone-phosphorylating activity, 5 ug ATP (240 cpm/pmol). For some experiments CDPK was incu- CDPK was electrophoresed in each of two adjacent lanes in a bated overnight at 4°C with anticalmodulin IgG or nonimmune 12.5% polyacrylamide gel as described for autophosphorylation IgG in Ca2+/EGTA buffer (no added calcium) prior to assay. above. The lanes were cut vertically with a cutter that left a Assay mixtures were incubated at 30°C for 6 min and the scalloped edge on the gel. The protein in one ofthe gel strips was reactions were terminated by the addition of 0.5 ml of cold 20% stained with Coomassie blue for 30 min, destained for 4h, then TCA containing 0.2% sodium pyrophosphate. The precipitates incubated in 50 mm Hepes (pH 7.2), for 30 min. Meanwhile, the were collected on Whatman GFA glass filters and washed with protein in the other gel strip was renatured as described above 20% TCA followed by ethanol/ether (1:1). The filters were added for autophosphorylation. The paired strips were aligned by to scintillation vials containing 2 ml Scintiverse II (Fisher Sci- matching the scalloped edges and the portion of the gel corre- entific) and counted in a liquid scintillation counter. sponding to the 46 to 51 kD stained proteins was excised. A gel For autophosphorylation assays, CDPK was incubated for 1 slice was prepared in this manner for each assay condition. The min at 30°C in a total volume of 25 ,l. The reaction mixture gel slices were added to 0.4 ml of protein kinase assay buffer in contained Ca2+/EGTA buffer, 1 gM [.y-32P]ATP (7500 cpm/ a 1.5 ml Eppendorf microfuge tube and the gel was crushed with pmol), 0.1 mg/ml BSA, and various concentrations of calcium. a pestle (Kontes Scientific Glassware/Instruments). The tubes Reactions were terminated and 32P-incorporation was deter- were inverted end-over-end for 3 h on a tube rotator (Scientific mined as described for the protein kinase assay. For some exper- Equipment Products). An aliquot of the supernatant was re- iments CDPK was autophosphorylated prior to electrophoresis. moved, added to electrophoresis sample buffer, and subjected to CDPK (5 ,g) was incubated 5 min at 25°C in a total volume of SDS-PAGE. Phosphorylated histone H1 was detected by auto- 25 ,l which contained EGTA buffer, 30 uM [y-32P]ATP (2000 radiography. cpm/pmol). The reaction was terminated by addition ofelectro- phoresis sample buffer (2% SDS, 1 mM EGTA, 2% glycerol, 5% RESULTS 2-mercaptoethanol, 0.001% bromphenol blue, final concentrations). The samples were boiled for 2 min and electrophoresed Purity of CDPK. A CDPK which catalyzes the phosphoryla- as described below. The dried gels were autoradiographed with tion of the artificial substrate Histone HI was purified 1000-fold Kodak XAR-5 film and a DuPont Cronex Lightning Plus inten- from suspension-cultured soybean cells as described in "Materials sifying screen at -80°C. and Methods." CDPK was analyzed by SDS-PAGE to assess its The calmodulin-dependent enzyme NAD kinase was used to purity (Fig. 1). The heavily protein-stained band centered at Mr assay calmodulin. NAD kinase activity was measured as de- 48 kD in lane 2 was shown to be a doublet of Mrs = 46 and 50 scribed by Harmon et al. (7) except that the change in A6W was kD when a lower amount of protein was loaded (lane 1) and in measured with a Varian DMS 100 Spectrophotometer equipped 10 to 15% gradient gels (Fig. 3A). We show below that these with a programmable cell changer. Soybean calmodulin was used bands possess calcium-dependent protein kinase activity. These as a standard. Samples of CDPK, histone H 1, and BSA were bands represent 80% of the protein in the preparation as deter- boiled in EGTA-containing buffer prior to assay for calmodulin. mined by scanning densitometry of the stained gel (data not SDS-PAGE. Proteins were electrophoresed by the method of shown). Other prominently-stained bands observed in CDPK Laemmli (13) in either 0.15 x 8.3 x 6 cm, 10% or 12% gels or (Fig. 1, lane 2) had Mrs = 72, 54, 41, and 32 kD. A faint band in 0.15 x 14 x 12 cm, 10 to 15% gradients gels. Mrs were corresponding to soybean calmodulin (lane 3, Mr = 17 kD) was calculated from a plot of RF versus Log Mr for the following observed in the overloaded CDPK sample in lane 2. marker proteins (Pharmacia low mol wt calibration kit): phos- Calmodulin in the CDPK preparation and in the histone HI phorylase b, 94 kD; BSA, 83kD; ovalbumin, 43 kD; carbonic and BSA used in protein kinase assays was measured with a anhydrase, 30 kD; soybean trypsin inhibitor, 20.1 KD; a-lactal- sensitive enzymic assay based on the activation of NAD kinase. bumin, 14.4 kD. Proteins were detected by staining with Coo- Boiled aliquots of CDPK contained 18 ng calmodulin per Mg of massie brilliant blue R. protein. No calmodulin was detected in histone H1 or BSA. In Calcium-Binding Activity. Calcium-binding proteins were de- the standard protein kinase assay 470 ng/ml ofCDPK was used, tected by 45Ca autoradiography of proteins which had been therefore the calmodulin concentration was 0.5 nM. Assuming a electroblotted onto nitrocellulose following SDS-PAGE as de- molecular mass of 48 kD and 80% purity for CDPK, the con- scribed by Maruyama et al. (15). Following electroblotting the centration of the enzyme in the assay mixture was 8 nM. nitrocellulose was washed three times for 20 min each in 60 mM Calcium-Dependence of Protein Kinase Activity and of Auto- KCI, 5 mM MgC92, 10 mm imidazole (pH 6.8). It was then phosphorylation. The phosphorylation of histone H 1 by CDPK incubated for 10 min in the same buffer to which 1 mCi/L was stimulated 100-fold by the addition of 2-10 Mm-free calcium carrier-free 45CaCl2 had been added, washed in deionized H20 (Fig. 2; Table I). No exogenous calmodulin was required for this for 5 min, and air-dried. Proteins bound to nitrocellulose were effect of calcium. Phosphorylation was half maximal at 2.1 MM- detected by staining with amido black (I15). free calcium. 832 HARMON ET AL. Plant Physiol. Vol. 83, 1987 Autophosphorylation was also stimulated by the addition of free calcium (Fig. 2B and Fig. 3). The Va,> was 10-fold higher with calcium than without. The relative activity in the absence of calcium was higher for autophosphorylation than for histone 1000 - phosphorylation. The concentration ofcalcium (-2 Mm) required for stimulation of autophosphorylation was similar to that re- - quired for histone Hi phosphorylation (Fig. 2A). The autoradi- 800 ogram in Figure 3 shows that calcium promoted the incorpora- E tion of labeled phosphate into at least two protein bands in the E 600- 46 to 51 kD range. Other 32P-labeled bands at 41 and 32 kD were visible when the autoradiogram was overexposed (data not shown). 400 - The 32P label associated with the protein bands of46 to 51 kD was not diminished when, prior to electrophoresis, the autophosphorylated CDPK was successively precipitated with TCA, 200 - washed with acetone, and then treated with hydroxylamine (data not shown). These treatments destroy phosphohistidine and phospholysine, remove phospholipids, and destroy acylphos- 0 - phate, respectively. This observation indicates that the phosphorylated residue(s) is serine, threonine, or tyrosine. 0 10- 10-s 10-4 Effect of Calmodulin on Protein Kinase Activity. Since the Free Ca* (M) amount of calmodulin present in the protein kinase assay was substoichiometric, based on the calculations above, the effect of exogenous calmodulin on CDPK activity was tested. When soybean calmodulin was added in large excess (1 tiM) over the e concentration of CDPK in the presence of calcium, a slight 100 -

80 -

60-

40- 94 - 67 - 20 -to

u Pe....T II I I. I.....I I I I ...... r OT-Tr II II I I II I1 43 - 0 1 0 10- 1074 Free Ca CM) FIG. 2. Effect of calcium on histone Hi phosphorylation and autophosphorylation. A, The phosphorylation of histone Hi by 470 ng/ml CDPK was measured in the presence of various concentrations of free calcium. The on the ordinate is the with no added - point plotted activity 30 calcium. Each point is the average of duplicate assays and the bars indicate the standard deviation. B, The autophosphorylation of CDPK (0.2 mg/ml) was measured in the presence of various concentrations of calcium. The point plotted on the ordinate is the fraction of maximal activity measured with no added calcium. Each point is the average of two determinations in duplicate and the bar indicates the standard 20- deviation.

14 (35%) stimulation of activity was observed (Table I). Increases 1 2A in activity from 30 to 100% were observed with 1 gM calmodulin in other experiments (data not shown). This stimulation ofactivity by calmodulin is small compared to the stimulation (100- fold) of activity by calcium alone. The calcium-independent FIG. 1. SDS-PAGE of CDPK. Lyophilized CDPK (6 pg in lane 1, 36 activity of CDPK was also stimulated by a high concentration of ,gg in lane 2) or soybean calmodulin (I 0 jg in lane 3) were electrophoresed calmodulin (Table I). CDPK activity decreased when a 2.5-fold in a 12% polyacrylamide gel in the presence of SDS. Protein was stained molar excess of soybean calmodulin (20 nM) was added with with Coomassie blue. The positions and Ms in kD ofthe protein markers calcium in the experiment shown in Table I; however, this are indicated. concentration of calmodulin had no effect on activity in other CALCIUM-DEPENDENT, CALMODULIN-INDEPENDENT PROTEIN KINASE 833 Table I. Effect ofCalmodulin or BSA on CDPKActivity Values are the means and standard deviations oftriplicate samples. Free Calcium Addition Activity JM % Experiment 1 13 None 100.0 ± 5.8 0 None 1.0 ±0.1 13 20 nm soybean calmodulin 85.0 ± 2.8 13 1 AM soybean calmodulin 135.3 ± 3.5 0 1 uM soybean calmodulin 2.5 ± 0.4 13 1 ;LM bovine brain calmodulin 130.0 ± 2.5 Experiment 2 13 None 100.0 ± 0.9 13 0.1 mg/ml BSA 174.2 ± 11.9 0 None 1.2 ±0 0 0.lmg/mlBSA 2.5±0.4

experiments. The effect of bovine brain calmodulin on activity A B in the presence of calcium was similar to the effect of soybean calmodulin (Table I). The effect of calmodulin on CDPK activity appears to be nonspecific, since such high concentrations were required for activation and since both basal and Ca2"-stimulated activities were affected. Furthermore, when 1.5 ;iM BSA was added to the protein kinase assay in the second experiment in Table I it also stimulated protein kinase activity in the presence and absence of calcium. 94- Calmodulin-Independence of CDPK Autophosphorylation and Activity. To directly demonstrate that CDPK autophosphoryla- 67- tion and activity were independent of calmodulin, CDPK was separated from calmodulin by electrophoresis in SDS gels, renatured, and either autophosphorylated in the gel or excised from the gel and incubated with the standard protein kinase 43 ~ ~.:...... '.;' ~~~ assay mixture. Figure 4A shows the results of autophosphorylation ofCDPK following SDS-PAGE. The proteins ranging in Mr from 46 to 51 kD were not resolved in this 10% polyacrylamide gel and migrated as a single band of Mr = 49 Md. Labeled phosphate was incorporated into these proteins in a calcium- dependent manner. To show that the labeling with 32P was covalent and not nonspecific, the bands were excised from the 30A dried gel, incubated in electrophoresis sample buffer, and electrophoresed on a second SDS gel. The bands were again detected by autoradiography (data not shown). To further examine which ofthe protein bands had autophosphorylated, CDPK was electrophoresed in a 10 to 15% SDS- gredient gel and then autophosphorylated in the gel (Fig. 4B). 20- Comparison ofthe two bands that were autophosphorylated after SDS-PAGE (Fig. 4B) to the bands that were autophosphorylated when CDPK was incubated with [y-32PJATP prior to electrophoresis (Fig. 3B) shows that the same two bands were labeled in both cases. 14- The 46 to 51 kD proteins renatured following SDS-PAGE were capable of phosphorylating histone H 1 in addition to au- tophosphorylating. When the renatured 46 to 51 kD proteins were excised in a single slice from an SDS gel, and incubated in the standard protein kinase assay mixture, they catalyzed the phosphorylation of histone H 1 in a calcium-dependent manner (Fig. 5). Since no calmodulin was detected in the histone HI or 2 1 2 the BSA, the experiments in Figures 4 and 5 show that both the activity and the autophosphorylation of CDPK are calcium dependent but not calmodulin dependent. FIG. 3. SDS-PAGE ofautophosphorylated CDPK. CDPK (5 ;g) was The calmodulin independence of the histone-phosphorylating autophosphorylated in the presence of 13 iM-free calcium (lane 1) or in activity of CDPK was confirmed in experiments in which anti- the absence of calcium (lane 2) and electrophoresed in a 10 to 15% bodies directed against calmodulin were included in the assay polyacrylamide gradient gel in the presence of SDS. A, Coomassie blue- mixture. CDPK was preincubated with either nonimmune IgG stained gel; B, autoradiogram. or with an amount of anticalmodulin IgG that inhibited NAD- 834 HARMON ET AL. Plant Physiol. Vol. 83, 1987

A 657- sCaP+ -Ca + + Ca7 94.. 3o-- 67- 43 - Oi .8. 30 - 14 :W 20- 234. FIG. 5. Phosphorylation of histone Hi by CDPK which had been renatured following SDS-PAGE. CDPK was electrophoresed, renatured, B 51 and incubated with histone Hi with no added calcium (lane 3) or with 46' 13 AM-free calcium (lane 4) as described in "Materials and Methods." Histones were also incubated in the protein kinase assay mixture with no CDPK added and with either no added calcium (lane 1) or with 13 FIG. 4. Autophosphorylation of CDPK Following SDS-PAGE. riM-free calcium (lane 2). Aliquots of each reaction mixture which con- CDPK (5 Mg) was electrophoresed in a 10% SDS-polyacrylamide gel, tained 20 Ag histone HI were electrophoresed in a 12.5% SDS-polyacryl- renatured in the gel, and autophosphorylated in the presence of 13 zM- amide gel. The autoradiogram of the dried gel is shown. free Ca2" (+Ca2"), in the absence of calcium (-Ca2"), or in the presence of 13 uM Ca2" and 0.5 Mm W-7 (+Ca2", +W-7). A, Autoradiographs of the dried gels. The positions and Mr in kD of the marker proteins are Ca2" or 2 mm EGTA, different mobilities were observed for one indicated. B, CDPK (5 Mg) was electrophoresed in a 10 to 15% SDS- of the Coomassie blue-stained bands (Fig. 7). The mobility of gradient polyacrylamide gel, autophosphorylated, and autoradiographed. this band was greater in the presence of Ca2+ (Mr = 43 kD, Lane The Mrs of the labeled bands are indicated. 2) than in the absence ofCa2" (Mr = 46 kD, lane 3). The mobility of the 50 kD band was unaffected. Calmodulin-Independence of the Inhibition of CDPK Auto- kinase activity in the presence of up to 10 ng/ml soybean phosphorylation and Histone Hl-Phosphorylating Activity by calmodulin. The activity of CDPK was slightly inhibited (10%) W-7 and W-5. The naphthalene sulfonamide W-7, which is a in the presence of calcium and inhibited 50% in the absence of calmodulin-binding compound, and its less potent analog W-5 calcium, whereas nonimmune IgG slightly stimulated activity. were both able to inhibit the phosphorylation of histone HI by The addition of up to 1 Mg/ml soybean calmodulin to the CDPK (Fig. 8). The concentrations ofW-7 and W-5 required for preincubation mixture in the presence ofcalcium did not restore 50% inhibition were 110 Mm and 1 mM, respectively. W-7 also activity (CP-Evans, AC Harmon, MJ Cormier, unpublished ob- inhibited autophosphorylation of CDPK which had been sepa- servations). rated from calmodulin by electrophoresis in an SDS-polyacryl- Ca2"-Binding Activity of CDPK. CDPK and other proteins amide gel (Fig. 4A, +Ca2+ +W-7). In the presence of calcium which had been electroblotted onto nitrocellulose following SDS- and 500 ,uM W-7 the incorporation of labeled phosphate was PAGE were incubated with 45Ca2" (Fig. 6). Protein bands of 46 comparable to the labeling in the absence of calcium (Fig. 4A, kD and 50 kD in CDPK (lane 2) and the positive controls, -Ca2"). This indicates that W-7 inhibits enzyme activity by soybean calmodulin (lane 3) and the Ca2"-dependent photopro- interacting directly with the enzyme and not with calmodulin. tein aequorin (M, = 25 kD, lane 4), produced strong signals on the autoradiograms (lanes marked B). Bands of 61, 51, and 34 DISCUSSION kD in CDPK (lane 2B) produced weak signals on the autoradiogram. The M, marker proteins, which were used as negative The protein bands on one-dimensional SDS-polyacrylamide controls, did not produce signals on the autoradiogram (lane 1), gels that are associated with soybean calcium-dependent protein although in some experiments soybean trypsin inhibitor and a- kinase have been identified. At least two bands in the Mr = 46 lactalbumin produced weak signals (data not shown). BSA did to 51 kD range were shown to autophosphorylate (Fig. 4) and to not bind 45Ca2' even when 15 ug of protein was loaded on the phosphorylate histone H 1 (Fig. 5) in a calcium-dependent man- gel (data not shown). Calmodulin did not stain well with amido ner following SDS-PAGE. Several other bands in the CDPK black (lane 3A). Calmodulin also does not stain well with Pon- preparation incorporated 32P04 when CDPK was incubated with ceau S or India ink (AC Harmon, MJ Cormier, unpublished [,y-32P]ATP prior to SDS-PAGE (Fig. 3). These results indicate observations). that catalytic activity is associated with the 46 to 51 kD proteins Ca2"-Dependent Mobility Change in SDS-PAGE. When and that the other proteins are either contaminants or noncatal- CDPK was electrophoresed in SDS gels in the presence of 2 mm ytic subunits ofCDPK that can serve as substrates ofthe enzyme. CALCIUM-DEPENDENT, CALMODULIN-INDEPENDENT PROTEIN KINASE 835 A B A B A B A B

FIG. 6. 45Ca2+-binding by CDPK. Following electrophoresis in a 10 to 15% SDS-gradient polyacrylamide gel, proteins were electroblotted onto nitrocellulose. The nitrocellulose was incubated in 45CaC12 as described in "Materials and Methods," autoradiographed overnight (lanes marked B), then stained with amido black (lanes marked A). 1, Marker proteins; 2, 15 ug CDPK; 3, 10 gg soybean calmodulin; 4, 10 jg apoaequorin. The marker proteins are from top to bottom: 1.6 jtg phosphorylase b, 2.1 gg BSA, 3.7 Ag ovalbumin, 2.1 usg carbonic anhydrase, 2.0 Mg soybean trypsin inhibitor, and 3.0 jig a-lactalbumin.

I 2 3 4

Whether the 46 to 51 kD proteins are different forms ofthe same two proteins co-purify because the proteins have similar bio- protein or different proteins with similar properties remains to chemical properties and not because they are bound to each be established. Until these proteins can be fully characterized we other. Calmodulin is thought to bind to phenothiazines and will refer to them collectively as CDPK since they do retain naphthalene sulfonamides at a hydrophobic binding site which calcium-dependent activity. The Mr of CDPK calculated from is exposed when it binds calcium. CDPK binds to TAPP sephar- its Stokes radius and its sedimentation coefficient is 59 kD (CP- ose in the presence of calcium and is eluted from the resin when Evans, MJ Cormier, unpublished observation). This observation calcium is removed by EGTA (22). This implies that CDPK also and the fact that the 46 to 51 kD proteins are catalytically active has a hydrophobic binding site that is exposed in the presence of indicate that the active kinase is a monomer. calcium. Both the autophosphorylation and histone H1 phosphoryla- Some reports identifying plant calmodulin-dependent protein tion were stimulated by concentrations of free calcium in the 1 kinases were based in part on the observation that micromolar to 10 jM range (Fig. 2). Although calmodulin-dependent en- levels of calmodulin stimulate protein kinase activity 1.5- to 6- zymes are also activated by concentrations of free calcium in this fold (18, 20, 25, 32). Blowers et al. (1) showed that autophos- range, calmodulin does not confer calcium sensitivity to CDPK. phorylation of a protein kinase, which had been isolated from CDPK activity was slightly inhibited by anticalmodulin, but the pea membranes by electroblotting, was stimulated by the addi- inhibition was not relieved by the addition of calmodulin (data tion of calcium and calmodulin, but they did not show the effect not shown). CDPK, which had been separated from calmodulin of calmodulin alone. Our results indicate that such observations by SDS-PAGE, autophosphorylated in a calcium-dependent could be misleading. We have shown that histone H 1-phospho- manner (Fig. 4), phosphorylated histone H1 in a calcium-de- rylating activity from all stages of CDPK purification (Table I; pendent manner (Fig. 5) and bound 45Ca2+ when incubated in a Refs. 5, 22) was stimulated a small amount (30-100%) by the buffer which contained 60 mM KC 1, 5 mM MgCl2, and micro- addition of micromolar calmodulin regardless of how much molar levels of Ca2+ (Fig. 6). These data show that CDPK endogenous calmodulin is present. Also, CDPK activity was specifically binds calcium, and that calcium directly regulates stimulated 75% by micromolar BSA (Table I). Both BSA and autophosphorylation as well as the phosphorylation of histone calmodulin increased activity in the presence and absence of H 1. Our results also indicate that phospholipids are not required calcium. These results suggest that the stimulation of CDPK for the activity of CDPK, because the purified enzyme is active activity by calmodulin is nonspecific. without the addition of phospholipids and any contaminating Another observation that has been relied upon to identify phospholipids would also be removed from CDPK by SDS- plant calmodulin-dependent protein kinases is the inhibition of PAGE. protein kinase activity by calmodulin antagonists (18, 20, 25, Calmodulin co-purifies with CDPK through DEAE chroma- 32). Our results indicate that these observations could also be tography and TAPP-sepharose chromatography (22). CDPK, like misleading since the inhibition of CDPK by the naphthalene calmodulin, is a high affinity Ca2+-binding protein (Fig. 6) and sulfonamide W-7 was independent of calmodulin (Fig. 4). W-5, interacts with W-7 (Fig. 4). Also, the mobility of one of the two an analog of W-7 that is similar in lipophilicity to W-7 but has CDPK bands in SDS-PAGE was calcium dependent (Fig. 7), as lower affinity for calmodulin than W-7 (1 1), had a higher IC50 is the mobility of calmodulin (3). These results suggest that the than W-7 (Fig. 8). The potency of this pair of inhibitors is in the 836 HARMON ET AL. Plant Physiol. Vol. 83, 1987 by purified CDPK was 110 gM (Fig. 8) and was 400 gM for less pure CDPK. These values are comparable to the IC50s reported for the inhibition ofthe calmodulin-independent protein kinases. Ca2? EGTA Polya and Micucci (21) have shown that Ca2l-dependent protein kinases I and II from wheat germ are inhibited by calmodulin antagonists, inhibitors of protein kinase C, and polyamines. These studies point out the difficulty in using calmodulin inhibitors, especially in vivo and in crude enzyme preparations. The stimulation oftonoplast-associated protein kinase activity by nanomolar levels of calmodulin and the complete inhibition of the activity by 1 gM R2457 1, a calmodulin antagonist, has been reported (31). However, the possibilty that the observed 32P-labeling was due to the incorporation of phosphate into tonoplast membrane ATPase reaction intermediates, rather than to the action of protein kinases, was not eliminated. Ranjeva et al. (23) have deduced, without the use of calmodulin inhibitors, that quinate:NAD' 3-oxidoreductase from carrot cell cultures is phosphorylated in a calcium- and calmodulin- dependent manner. That the activity of the quinate:NAD' 3- oxidoreductase is apparently regulated by a phosphorylation/ dephosphorylation mechanism was indicated by the following observations (23, 24). Quinate:NAD' 3-oxidoreductase activity is lost when impure enzyme preparations are incubated with Mg2+, but inclusion of fluoride, a phosphatase inhibitor, in the extract prevents this loss ofactivity. Addition ofATP to extracts containing inactivated quinate:NAD' 3-oxidoreductase restores * .ys' activity. When extracts containing inactivated quinate:NAD' 3- oxidoreductase are first incubated with ['Y-32P]ATP and then fractionated on an isoelectric focusing column, fractions containing quinate:NAD' 3-oxidoreductase activity also contain 32p. However, it was not shown whether inactive enzyme contains phosphate, whether there is a correlation between the degree of 32P-incorporation and amount of activity, or whether the 32p associated with the active enzyme could subsequently be removed with a concomitant loss of activity. In more recent experiments, it was shown that calcium and calmodulin are required for the reactivation process (23). If either calcium is removed from the extract by chelation with EGTA or calmodulin

1 2 34 800

7. The effect of on the of CDPK in SDS-PAGE. 600 FIG. Ca2" mobility CD Mr markers I and and CDPK lanes 2 and 3) were E (lane 4) (5 Mg, I- electrophoresed in a 1O to 15% SDS-gradient polyacrylamide gel. Either C 2 mm CaC12 (+Ca2+) or 2 mm EGTA was included in the electrophoresis sample buffer. The Mr markers are the same as those listed in the legend to Figure 6. E c proper order to indicate that calmodulin could be involved even 200 though we have shown that the effect ofW-7 is directly upon the enzyme and is independent of calmodulin. Our data shows that inhibition studies using calmodulin-binding compounds must be interpreted with caution. Phenothiazines and naphthalene sul- 0 fonamides have been shown to inhibit other calmodulin-inde- 0 10o5 10-4 1o-3 pendent protein kinases. Trifluoperazine, a phenothiazine, inhibits phospholipid-dependent kinases with IC50s ranging from 38 Concentration (M) to 50 ,uM, whereas it inhibits calmodulin-dependent cyclic nu- FIG. 8. The effect of W-7 and W-5 on the phosphorylation ofhistone cleotide phosphodiesterase with IC50s of 10 to 12 uM (28). W-7 Hi by CDPK. The phosphorylation of histone HI by CDPK (470 ng/ inhibits phospholipid-dependent, cAMP-dependent, and cGMP- ml) was measured in the presence of 13 MM-free Ca2" and various dependent protein kinases with IC50s in the range of 130 to 360 concentrations of W-7 (0) or W-5 (A). Each point is the average of Mm (27, 30) whereas W-7 inhibits the calmodulin-dependent duplicte assays and the bar indicates the standard deviation. The activities enzymes, myosin light chain kinase and cyclic nucleotide phos- plotted on the ordinate were determined in the absence ofinhibitors and phodiesterase, with IC50s of 50 ,M (12) and 26 uM (10), respec- either with 13 gM calcium (100% activity) or without added calcium tively. The IC50 for the inhibition of histone H l-phosphorylation (labeled -+EGTA"). CALCIUM-DEPENDENT, CALMODULIN-INDEPENDENT PROTEIN KINASE 837 is removed by chromatography on an affinity column, quin- Calcium-regulated modulator protein interacting agents inhibit smooth mus- ate:NAD+ 3-oxidoreductase activity is not restored by the addi- cle calcium-stimulated protein kinase and ATPase. Mol Pharmacol 17: 66- 72 tion of ATP. When 0.5 gM bovine brain calmodulin and 1 mm 13. LAEMMLI UK 1970 Cleavage of structural proteins during the assembly of the CaCl2 are added to the inactive calmodulin-depleted extract, head ofbacteriophage T4. Nature 227: 680-685 130% of the original activity is restored. It would be very inter- 14. LUKAS TJ, DB IVERSON, M SCHLEICHER, DM WATTERSON 1984 Structural esting if the calcium- and calmodulin-dependent protein kinase characterization of a higher plant calmodulin. Plant Physiol 75: 788-795 15. MARUYAMA K, T MIKAWA, S EBASHI 1984 implicated in these studies can be isolated and characterized. Detection of calcium binding proteins by 45Ca autoradiography on nitrocellulose membrane after sodium Besides its requirement for calmodulin, this protein kinase is dodecyl sulfate gel electrophoresis. J Biochem 95: 511-519 unlike CDPK in that it does not bind to either DEAE-Sephacel 16. MURASHIGE T, F SKOOG 1962 A revised medium for rapid growth and bioassays (24) or to a calmodulin affinity resin (23). with tobacco tissue cultures. Physiol Plant 15: 473-497 17. MuTo S, K SHIMOGAWARA 1985 Calcium- and phospholipid-dependent phos- Our results show that CDPK binds calcium directly, is acti- phorylation of ribulose-1,5-bisphosphate carboxylase/oxygenase small sub- vated 100-fold by micromolar calcium, is not dependent on unit by a chloroplast envelope-bound protein kinase in situ. FEBS Lett 193: calmodulin or phospholipid for activity, and is directly inhibited 88-92 by W-7. The calcium-dependent protein kinases discovered to 18. PALIYATH G, BW POOVAIAH 1985 Calcium- and calmodulin-promoted phosphorylation of membrane proteins during senescence in apples. Plant Cell date have been dependent on either calmodulin or phospholipids Physiol 26: 977-986 for activity. Therefore, CDPK represents a new class of calcium 19. PERRIN DD, IG SAYCE 1967 Computer calculation of equilbrium concentra- dependent protein kinase. tions in mixtures of metal ions and complexing species. Talanta 14: 833- 842 20. POLYA GM, JR DAVIES, V MIcuccI 1983 Properties of a calmodulin-activated LITERATURE CITED Ca2l-dependent protein kinase from wheat germ. FEBS Lett 150: 167-171 21. POLYA GM, V Micucci 1985 Interaction of wheat germ Ca2l-dependent 1. BLOWERS DP. A HETHERINGTON, A TREWAVAS 1985 Isolation of plasma protein kinases with calmodulin antagonists and polyamines. Plant Physiol membrane-bound calcium/calmodulin-regulated protein kinase from pea 79: 968-972 using western blotting. Planta 166: 208-215 22. PUTNAM-EVANS CL, AC HARMON, MJ CORMIER 1986 Calcium-dependent 2. BRADFORD MM 1976 A rapid and sensitive method for the quantitation of protein phosphorylation in suspension-cultured soybean cells. In AJ Trewa- microgram quantities of protein utilizing the principles of protein-dye bind- vas, ed, Molecular and Cellular Aspects of Calcium in Plant Development. ing. Anal Biochem 72: 248-254 Plenum Publ Corp, New York, pp 99-106 3. BURGESS WH, DK JEMIOLO, RH KRETSINGER 1980 Interaction of calcium and 23. RANJEVA R, G REFENO, AM BOUDET, D MARMt 1983 Activation of plant calmodulin in the presence of sodium dodecyl sulfate. Biochim Biophys Acta quinate:NAD+ 3-oxidoreductase by Ca2 and calmodulin. Proc Natl Acad 623: 257-270 Sci USA 80: 5222-5224 4. CHARBONNEAU H, R HICE, RC HART, MJ CORMIER 1983 Purification of 24. REFENO G, R RANJEVA, AM BOUDET 1982 Modulation of quinate:NAD+ 3- calmodulin by Ca2"-dependent affinity chromatography. Methods Enzymol oxidoreductase activity through reversible phosphorylation in carrot cell 102: 17-39 suspensions. Planta 154: 193-198 5. CORMIER MJ, A HARMON, C PUTNAM-EVANS 1985 Ca2"-dependent regulation 25. SALIMATH BP, D MARME 1983 Protein phosphorylation and its regulation by of NAD kinase and protein phosphorylation in plants. In H Hidaka, DJ calcium and calmodulin in membrane fractions from zucchini hypocotyls. Hartshorne, eds, Calmodulin Antagonists and Cellular Physiology.Academic Planta 158: 560-568 Press, Orlando, FL pp 445-456 26. SCHAFER A, F BYGRAVE, S MATZENAUER, D MARMt 1985 Identification of a 6. GEAHLEN RL, M ANOSTARIO JR, PS Low, ML HARRISON 1986 Detection of calcium- and phospholipid-dependent protein kinase in plant tissue. FEBS protein kinase activity in sodium dodecyl sulfate-polyacrylamide gels. Anal Lett 187: 25-28 Biochem 153: 151-158 27. SCHATZMAN RC, RL RAYNOR, JF Kuo 1983 N-6-aminohexyl)-5-chloro-l- 7. HARMON AC, HW JARRETT, MJ CORMIER 1984 An enzymatic assay for naphthalenesulfonamide (W-7), a calmodulin antagonist, also inhibits phos- calmodulins based on plant NAD kinase activity. Anal Biochem 141: 168- pholipid-sensitive calcium-dependent protein kinase. Biochim Biophys Acta 178 755: 144-147 8. HART RC, MD BATES, MJ CORMIER, GM ROSEN, PM CONN 1983 Synthesis 28. SCHATZMAN RC, BC WISE, JF Kuo 1981 Phospholipid-sensitive calcium- and characterization of calmodulin antagonistic drugs. Methods Enzymol, dependent protein kinase: inhibition by antipsychotic drugs. Biochem Bio- 102: 195-204 phys Res Commun 98: 669-676 9. HIDAKA H, M ASANO, I IWADARE, T TOTSUKA, N AOKI 1978 A novel vascular 29. SILLtEN LG, AE MARTELL 1971 Stability Constants of Metal-ion Complexes, relaxing agent. N-(6-aminohexyl)-5-chloro-1-naphthalene-sulfonamide Ed 2. Burlington House, London which affects vascular smooth muscle actomyosin. J Pharmacol Exp Ther 30. TANAKA T, T OHMURA, T YAMAKADO, H HIDAKA 1982 Two types of calcium- 207: 8-15 dependent protein phosphorylations modulated by calmodulin antagonists: 10. HIDAKA H, M ASANO, T TANAKA 1981 Activity-structure relationship of naphthalenesulfonamide derivatives. Mol Pharmacol 22: 408-412 calmodulin antagonists: naphthalenesulfonamide derivatives. Mol Pharma- 31. TEULIERES C, G ALIBERT, R RANJEVA 1985 Reversible phosphorylation of col 20: 57 1-578 tonoplast proteins involves tonoplast-bound calcium calmodulin-dependent I 1. HIDAKA H, Y SASAKI, T TANAKA, T ENDO, S OHNO, Y FuJi, T NAGATA 1981 protein kinase(s) and protein phosphatase(s). Plant Cell Rep 4: 199-201 N-6-Aminohexyl)-5-chloro- 1 -naphthalenesulfonamide, a calmodulin antag- 32. VELUTHAMBI K, BW POOVAIAH 1986 In vitro and in vivo protein phosphoryl- onist, inhibits cell proliferation. Proc Natl Acad Sci USA 78: 4354-4357 ation in Avena saliva L. coleoptiles. Effects of Ca2", calmodulin antagonists, 12. HIDAKA H. T YAMAKI, N NAKA, T TANAKA, H HAYASHI, R KOBAYASHI 1980 and auxin. Plant Physiol 81: 836-841