Planta (1984)161:409~t17
P l a n t a 9 Springer-Verlag 1984
Activation of a pea membrane protein kinase by calcium ions A.M. Hetherington and A. Trewavas Botany Department, University of Edinburgh, Mayfield Road, Edinburgh EH9 3JH, UK
Abstract. Membranes from the buds of P i s u m sativum L. contain a protein kinase which is activated 5- to 15-fold by micromolar levels of calcium. Best calcium activations were found with light-membrane fractions, and on density gradients these band at a similar position to the plasma membrane. Other heavier membranes, however, also contain a calcium-dependent protein kinase. The activity of the calcium-dependent protein kinase is inhibited by added phospholipids and phospholipase, in contrast to protein-kinase C. Calcium-dependent protein-kinase activity can be inhibited by 40% by low concentrations of the calmodulin inhibitor, trifluoperazine, but inhibitions are detected only after prior incubation of the membranes for some hours in ethylene glycol-bis-(flaminoethyl ether)-N,N,N',N'-tetraacetic acid. Substantial calcium-dependent protein-kinase activity remains uninhibited by trifluoperazine indicating that there may be calmodulin-dependent and calmodulin-independent, but calcium-activated, protein kinases in pea membranes. The calcium-activated protein kinase seems to be intrinsically bound to membranes and only slight or partial solubilisation is obtained by the detergents nonidet P-40, (3-[(3-cholamidopropyl)-dimethylammonio]-l-propanesulfonate or octyl glucose. Better solubilisation is obtained by acetone treatment. There is some retention of calcium activation after partial solubilisation. A calcium-independent protein kinase has also been detected in membrane preparations; it has a substrate specificity different from that the calcium-dependent enzyme. Our results indicate, therefore, that there may be at least three protein kinases attached to pea shoot membranes. Abbreviations: EGTA = ethylene glycol-bis(fl-aminoethyl ether)-N,N,N',N'-tetraacetic acid; Hepes = 4-(2-hydroxyethyl)1-piperazineethanesulfonic acid; TFP = trifluoperazine
Key words: Calcium (enzyme activation) - Protein kinase Pisum (membrane protein).
Introduction The regulation of cellular metabolism by calcium ions is probably one of the most fundamental control systems yet discovered. In animals, processes as diverse as fertilization, ciliary movement, muscle contraction and cell divisions are manipulated by controlling the cytoplasmic level of calcium ions (Durham and Walton 1982). Many of the processes utilize calmodulin as a calcium sensor and effector (Cheung 1980; Klee et al. /980). In plants it is likely that an equally diverse array of phenomena are regulated in a similar way. Cell division, cell expansion, cell differentiation and specification and cytoplasmic streaming; all these processes have been shown in one way or another to respond to changes in available cytoplasmic calcium (Hanson 1984). Calmodulin has likewise been isolated from numerous plants (Marine and Dieter 1982). Many of the effects of cytoplasmic calcium can be traced to a modulation of the levels of certain specific enzymes. Amongst these, protein kinases, because of their pleiotropic function, rank as one of the most important (Dabrowska et al. 1977; Yamauchi and Fujisaira 1979), and calcium/calmodulin-dependent (Dalbrowska et al. 1977; Yamauchi and Fujisaira 1979) and calcium/phospholipid-dependent (Takai et al. 1979) protein kinases have been reported in animal tissue. Recently we briefly reported the presence of a calcium-dependent protein kinase in pea shoot preparations (Hetherington and Trewavas 1982) which is not stimulated by cyclic nucleolides. This enzyme was activated three- to fivefold by concentrations of calcium as
410
A.M. Hetherington and A. Trewavas: Pea membrane protein kinase
low as micromolar but the phosphorylation of some protein species could be increased some 20-fold. Activity was only weakly increased by calmodulin and inhibition by trifluoperazine (TFP), an inhibitor of calmodulin-dependent enzymes, was only observed at very high concentrations (Hetherington and Trewavas 1982). As many critical events in plant cell development occur at the membrane level, we have investigated in much greater detail some of the characteristics of this important regulatory enzyme. Subsequent reports have appeared from three other laboratories indicating the presence of calcium-activated protein kinases in plants (Polya and Davies 1982; Ranjeva et al. 1983 and Salimath and Marme 1983).
be calculated. The filters were washed overnight in the 10% TCA mixture, then washed in 5% TCA for 15 min, boiled for 15 min in fresh 10% TCA mix and allowed to cool for 30 rain. They were then washed for 15 s in methanol, and after drying, radioactivity was determined by Cerenkov counting. When Histone H1 was used as an exogenous substrate for the protein kinase, 25% TCA was used throughout the filter procedure and the concentration of calf histone H1 (Sigma) was 150 gg ml 1 during the incubation. Initial velocities were determined from the asymptote of the smoothed curve drawn through the points at zero time. Typically filter discs produced between 200 and 2000 cpm.
Gel electrophoresis. Sodium-dodecyl-sulphate polyacrylamidegel electrophoresis of labelled proteins was as previously described (Hetherington and Trewavas 1982) using 12% gels and detection of labelled bands by autoradiography.
Results and discussion Material and methods Plant material. Peas (Pisum sativum L. cv. Feltham First) were grown in the dark at 22~ for 10 d and the shoot material used. Most experiments used the bud which contained the unexpanded fourth and fifth leaves. In some experiments root material was collected from 7-d-old seedlings. )~-[32P]adenosine 5'-triphosphate (specific activity 111 TBq (mmol 1) was obtained from Amersham International plc (Amersham, Bucks., UK). Trifluoperazine was the kind gift of Smith, Kline and French Labs (Welwyn Garden City, Herts., UK). Trypsin, type X l l 2X crystalized; trypsin inhibitor, type 1-S; histone H I type VS; and other chemicals were obtained from Sigma London Chemical Co. (Poole, Dorset, UK).
Membrane preparation. Fresh bud material (20-30 g) was chilled and ground in 3 vols. ice-cold 0.3 M sorbitol, 50 m M 4-(2-hydroxyethyl)-l-piperazine ethanesulfonic acid (Hepes) pH 7.6, i m M ethylene diamine tetraacetic acid (EDTA), 0.1% phenylmethyl sulfonyl fluoride. After filtration, the solution was centrifuged at 12000 g for 10 min, the pellet discarded and the supernatant recentrifuged at 35 000 g for 30 rain. The pellet was resuspended in a total volume of 20-30 ml of 0.3 M sorbitol, 5 m M Hepes pH 7.0, 0.1 m M EDTA, layered over a cushion of 0.95 1 M sorbitol containing 5 m M Hepes, 0.1 m M EDTA, pH 7, and centrifuged at 150000g for 1 h in a swingout rotor at 2 ~ C. Alternatively, the pellet was resuspended and layered onto a 12-ml, 10-50% (w/w) linear sucrose gradient containing 5 m M Hepes and 0.1 m M E D T A and centrifuged at 80000 g for 16 h. The gradient was divided into 20 fractions, the density of each determined refractometrically and aliquots used for protein-kinase estimation and protein estimation (Lowry et al. 1951). Protein-kinase estimations. The final membrane pellet was resuspended in 6-8 ml 50 m M Hepes and protein-kinase estimations carried out using 7-[32P]ATP at 2 . 1 0 - 6 M as previously described (Hetherington and Trewavas 1982). Assays of each fraction were conducted in the presence or absence of 100 gM free calcium (Hetherington and Trewavas 1982) obtained from Analar CaC12.2H20 (BDH Chemical Co. (Poole, Dorset, UK)). Assays used time periods up to 45 s at 25 ~ C or up to 3 rain at 0 ~ C. The reaction was stopped by pipetting 75-gl aliquots onto 3 ~ filter discs which had been pretreated with 50 ~1 of 10% trichloracetic acid (TCA), 20 m M sodium pyrophosphate, and 10 m M EDTA. Sufficient 75-gl aliquots (usually three) were taken to enable estimates of initial velocity to
Fractionation of the calcium-dependent protein kinase. Table 1 shows that the majority of pellets showing calcium-dependent protein-kinase activity were present in cell fractions requiring 30 min at 20000-40000 g for their sedimentation. Additionally there was detectable activity in the 12000-g pellet. When highly purified mitochondria from this latter fraction (prepared for us by Dr. C.J. Leaver, Department of Botany, University of Edinburgh) were examined, they were found to contain only 10-15% of the total activity of the original fraction and to have a low calcium activation (twofold). The activity in this fraction may, therefore, be caused by large membrane fragments. It can be seen that there is little calcium-dependent protein kinase activity in either the nuclear-rich (1000-g) or 80 000-g supernatant fractions. Table 2 shows the distribution of calcium-deTable 1. Centrifugal distribution of calcium-dependent membrane-bound protein kinase in pea shoot homogenates. Etiolated pea bud tissue (20 g) was collected, homogenised and fractionated at different centrifugal speeds. Pelleted preparations were resuspended in 2 ml 50 m M Hepes medium and time courses of protein-kinase activity determined in the presence or absence of 100 gM free calcium. Initial velocities were estimated as asymptotes to the smoothly drawn curve at zero time Centrifugal force (g average)
1000 12 000 20000 40000 80000 Supernatant
Time (min)
15 10 15 30 30 -
Protein-kinase activity (pmol Pi m i n - 1 g g - 1 protein) + Ca z +
_ Ca 2 +
0.06 0.40 2.22 2.27 0.97 0.132
0.032 0.06 0.25 0.227 0.089 0.116
Activation
1.8 6.8 8.8 10 11 1.2
A.M. Hetherington and A. Trewavas: Pea membrane protein kinase Table 2. The level of calcium-dependent protein kinase in various etiolated pea tissues. Etiolated t0-d-old pea plants were sequentially divided into the apical bud tissue, the apical hook, a ~-cm e~ongat~ng zone, a 3-cm rn~ture zone ptus the node, and a 3-cm apical portion of the main root. Membranes were prepared from each fraction and the initial velocity of bound protein kinase determined in the presence and absence of calcium Tissue
Protein-kinase initial velocity (cpm gg-1 protein) +Ca
Bud Apical hook Elongating zone Mature zone + node Root
0.60 0.182 0.402 0.452 0.26
2+
_Ca
0.065 0.023 0.09 0.078 0.066
411
18
1.165 1.136 d e n s i t y
l
6 o
"o
Activation
8 :>.
2+
6
9.3 8.0 4.5 5.8 3.7
-~
1
2
c
pendent protein kinase in various etiolated pea tissues. As the bud tissue exhibited the greatest calcium activation it has been used in subsequent experiments. Membrane-bound calcium-dependent protein kinase was also detected in cultured carrot cells, maize coleoptiles, tobacco pith and crowngall cells indicating its likely wide distribution. Fractionation o f the membrane-bound proteinkinase preparation by isopycnic centrifugation is shown in Fig. 1. There are at least two components. One is represented by a shoulder and has a density of approx. 1.16 g cm -3, and the second major component has a peak density of 1.136 g c m - a ; it should be noted that the activation ratio ( + Ca 2 + / - Ca a +) is higher in the lightermembrane fraction. This latter density is characteristic of pea epicotyl plasma membrane (Hendrik and Pierce 1980). To confirm this we also used the fully characterised procedure of Rasi-Caldogno et al. (1982) devised for pea tissue for obtaining fractions enriched in plasma membrane, and again assayed the calcium-dependent protein kinase. The results in Table 3 confirm the general picture of the isopycnic-gradient results. The best calcium activations are observed in plasma-membrane-enriched fractions but the enzyme is also located in other cell membranes of differing densities. The calcium-dependent and calcium-independent protein kinase may be different enzymes in pea bud tissues. There are two pieces of evidence which indicate that this may be the case. Figure 2a shows that calf histone HI will act as a substrate for the calcium-activated protein kinase but not the nonactivated form. Neither enzyme will phosphorylate ion-free casein and the other histones (H2-H4) have been found to be inhibitory. Figure 2 b shows
~3 ~- 0
0 0
10 20 F r a c t i o n no.
03
Fig. 1. Isopycnic separation of calcium-activated protein kinase on sucrose density gradients. Pea bud tissue (10 g) was collected, homogenised and a crude microsomal preparation made. The final pellet was suspended in 1 ml 10 mM Hepes and layered on a 14-ml linear sucrose gradient ( 1 ~ 5 0 % , w/v) and centrifuged overnight at 80000 g. The gradient was divided into 20 fractions and aliquots used for density determination, protein estimation and determination of calcium-dependent and calcium-independent protein-kinase activity Table 3. Distribution of calcium-dependent protein kinase amongst membrane fractions using the method of Rasi-Caldogno et al. (1982) for plasma-membrane preparation. Pea bud tissue was collected and the two fractions described by RasiCaldogno et al. (1982) were prepared using the metrizamide step gradients and high bovine-serum-albumin concentrations prepared as described in their paper. Two preparations were made, one at 10 mM Mg-acetate and the other at 1 mM. According to Rasi-Caldogno et al. (1982) this leads to prepalations with markedly different calcium-uptake capabilities. Protein-kinase initial velocities were determined as described in the Material and methods section and in Table 1 Fraction
Plasma-membrane enriched Heavy membrane
Mg 2 + (mM)
l0 1 10
Protein-kinase initial velocity (cpm ktg-1 protein) + C a 2+
_ C a 2+
3.15 2.6 1.65
0.055 0.045 0.052
Acti~ vation
5.7 5.7 3A
autoradiographs o f the peptides phosphorylated 'in vitro' using [32p]ATP by the calcium-activated and the non-activated forms. Whereas the calciumactivated enzyme phosphorylates many proteins, the calcium-independent protein kinase primarily
412
A.M. Hetherington and A. Trewavas: Pea membrane protein kinase
x is
o x
E
5
4
8
6 +Ca
~
|
31
6
cJ
.E
2~
g
3
O r
2c .E
1 _~ ~
__ 9
-
-Ca A
-~
o
D_
,q
,
a
50 100 Hist0ne HI (/ag)
~J
u o
2
150
>
Ik"A----a-
A
it
(D rn
4
9
I
I
I
I
!
J
5 10 15 Added phospho[ipase C (/at)
2a
Fig. 2a, b. Substrate specificity of the calcium-activated and calcium-independent protein kinase of pea shoot membranes, a Variation in pea bud protein-kinase activity with histone H1 concentration. Microsomal membranes were incubated with increasing levels of histone H1 in the presence or absence of 100 gM free calcium and the initial velocity of protein-kinase activity determined. b Gel electrophoresis of pea bud membrane proteins phosphorylated in the presence or absence of 100 gM free Ca 2+. Proteins were phosphorylated for 45 s and the reaction stopped by adding an equal volume of 2.5% sodium lauryl sulphate, 2-amino-2(hydroxymethyl)-l,3-propanediol-chloride (Tris-C1) (0.13 M, pH 6.7) mercaptoethanol (2%) and sucrose (10%). After boiling for 1 min, the solubilised proteins were separated on 12% acrylamide gels and dried and autoradiographed. Arrows mark the position of proteins of M r 40000 and 59 000 specifically phosphorylated in the absence of calcium Fig. 3. Effects of various concentrations of phospholipase C on the calcium-activated protein kinase of pea bud membranes. Microsomal membranes were suspended in 50 mM Hepes (pH 7.8) and to each a designated number of microliters of phospholipase C (Sigma Type V) were added for 2 min at 0 ~ C. The reaction was stopped with 100 gM ethylenediaminetetraacetic acid pH 7.2 (Jolliot et al. 1982). After centrifugation, the membrane pellet was resuspended and calcium-dependent and calcium-independent protein kinases were assayed using histone H1 (150 gg ml-1) as an exogenous substrate
phosphorylates two peptides running at a relative molecular mass (Mr) of 59000 and 40000 with only traces of others.
Inhibition of the calcium-activated protein kinase by phospholipids and diglycerides. A calcium-activated and diglyceride-activated protein kinase (protein kinase C) has been found to be a mediator of growth signals in many animal cells and to be the receptor for the tumour-promoting phorbol esters (Kuo et al. 1980; Takai et al. 1982; Niedel et al. 1983). This enzyme is also membrane-bound. Is the pea bud kinase a plant form of protein kinase C? We have tried to demonstrate diglyceride activation in a number of experiments and Table 4 and Fig. 3 summarise these data. In no case have we achieved activation of the pea bud protein kinase. Routinely we have observed inhibition of between 40-50% both for membrane-bound protein kinases and soluble protein kinases. As this does not rule out the possibili-
ty that sensitivity may be based on an absolute requirement for a particular molecular species of lipid, endogenous diacylglycerols were generated, using the method of Jolliot et al. (1982) by brief exposure of the membrane preparation to phospholipase C. Figure 3 shows that no stimulatory effects were observed, instead the observed inhibition is probably indicative of the hydrophobic nature of the enzyme and indicates lipid-dependence for activity as found in cytidine-diphosphate-choline diacylglycerol phosphorylcholine phosphotransferase (Jolliot et al. 1982).
Inhibition of calcium activation by TFP and tryptic digestion experiments indicate the presence of at least two calcium-regulated protein kinases. Since many calcium-regulated enzymes are calmodulindependent we have examined whether this membrane-bound protein kinase is likewise calmodulindependent. In our previous paper (Hetherington and Trewavas 1982) we reported that added bovine
A.M. Hetherington and A. Trewavas: Pea membrane protein kinase Table 4. Sensitivity of membrane and soluble protein kinase to added lipids. Microsomal membranes were prepared as described in Material and methods. Soluble protein kinase represented a 40000-g supernatant. Phosphatidylinositol and diolein suspensions were prepared by drying the stock solution and then resuspending in incubation buffer with sonication. Resuspension buffer used for the controls was sonicated for the same period
Treatment
M e m b r a n e fraction
Soluble fraction
Initial velocity protein kinase
Ca 2 + activation
Initial velocity protein kinase
Ca 2 § activation
Control
6680
15.5
495
1.8
+ Phosphatidylinositol (10 lag ml-1)
4250
11.9
280
1.5
+ Diolein (1.25 lag m l - 1)
3 750
413
X
q E 3
2
~o
.I
D
.E 09 o
I
n
10.6
70
1.1
calmodulin led to slight increases in assayable protein kinase. These increases were small (20%) but we have found them quite reproducible. Replacing bovine by pea calmodulin (kindly prepared for us by Miss E. Allan, Department of Botany, University of Edinburgh) did not improve this degree of stimulation (data not shown). In addition we found that TFP (an inhibitor of calmodulin-enzyme interactions) only inhibited the calcium protein kinase at millimolar concentrations, an unacceptably high concentration for specific inhibition. We have considerably improved this initially unsatisfactory observation. We have found (like Saitoh et al. (1982)) that prior incubation of membrane preparations in ethylene glycol-bis(fl-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA) for 2 h enables inhibition by lower concentrations of TFP to be demonstrated. Figure 4 summarises these data. Inhibition can be detected at 0.1 m M TFP but there is some sort of levelling out between 0.2 and 0.3 m M TFP when there is still substantial residual calcium-activated protein kinase. Taken at their face value these results indicate that there may be a calmodulin-dependent and a calmodulinindependent, but in both cases calcium-activated, form of protein kinase in these membranes. However, inhibitors are rarely completely specific and the only critical method of establishing this point will be complete solubilisation and separation of the bound protein kinases. The possible presence of two calcium-regulated protein kinases is perhaps more clearly indicated by the data in Fig. 5. Membrane preparations were
9
I
I
0.1
0
I
I
0.2
I
D
0.3
T r i f [ u o p e r a z i n e (raM) Fig. 4. Inhibition of calcium-dependent protein kinase from pea bud membranes by trifluoperazine. Microsomal membrane preparations were prepared and incubated in 2 m M E G T A 50 m M Hepes pH 7.2 for 2 h at 0 ~ C. To appropriate samples, TFP was added at the concentrations shown in the figure and then finally 2.1 m M CaC12 for assessing calcium dependence. Protein-kinase initial velocities were then determined in the usual way using histone H1 (150 lag m1-1) as exogenous substrate
% X | t 1 I
6
; 10
i .2
ul
2~ .E
:>
5
2 ?,
o 0_
P l a n t a 9 Springer-Verlag 1984
Activation of a pea membrane protein kinase by calcium ions A.M. Hetherington and A. Trewavas Botany Department, University of Edinburgh, Mayfield Road, Edinburgh EH9 3JH, UK
Abstract. Membranes from the buds of P i s u m sativum L. contain a protein kinase which is activated 5- to 15-fold by micromolar levels of calcium. Best calcium activations were found with light-membrane fractions, and on density gradients these band at a similar position to the plasma membrane. Other heavier membranes, however, also contain a calcium-dependent protein kinase. The activity of the calcium-dependent protein kinase is inhibited by added phospholipids and phospholipase, in contrast to protein-kinase C. Calcium-dependent protein-kinase activity can be inhibited by 40% by low concentrations of the calmodulin inhibitor, trifluoperazine, but inhibitions are detected only after prior incubation of the membranes for some hours in ethylene glycol-bis-(flaminoethyl ether)-N,N,N',N'-tetraacetic acid. Substantial calcium-dependent protein-kinase activity remains uninhibited by trifluoperazine indicating that there may be calmodulin-dependent and calmodulin-independent, but calcium-activated, protein kinases in pea membranes. The calcium-activated protein kinase seems to be intrinsically bound to membranes and only slight or partial solubilisation is obtained by the detergents nonidet P-40, (3-[(3-cholamidopropyl)-dimethylammonio]-l-propanesulfonate or octyl glucose. Better solubilisation is obtained by acetone treatment. There is some retention of calcium activation after partial solubilisation. A calcium-independent protein kinase has also been detected in membrane preparations; it has a substrate specificity different from that the calcium-dependent enzyme. Our results indicate, therefore, that there may be at least three protein kinases attached to pea shoot membranes. Abbreviations: EGTA = ethylene glycol-bis(fl-aminoethyl ether)-N,N,N',N'-tetraacetic acid; Hepes = 4-(2-hydroxyethyl)1-piperazineethanesulfonic acid; TFP = trifluoperazine
Key words: Calcium (enzyme activation) - Protein kinase Pisum (membrane protein).
Introduction The regulation of cellular metabolism by calcium ions is probably one of the most fundamental control systems yet discovered. In animals, processes as diverse as fertilization, ciliary movement, muscle contraction and cell divisions are manipulated by controlling the cytoplasmic level of calcium ions (Durham and Walton 1982). Many of the processes utilize calmodulin as a calcium sensor and effector (Cheung 1980; Klee et al. /980). In plants it is likely that an equally diverse array of phenomena are regulated in a similar way. Cell division, cell expansion, cell differentiation and specification and cytoplasmic streaming; all these processes have been shown in one way or another to respond to changes in available cytoplasmic calcium (Hanson 1984). Calmodulin has likewise been isolated from numerous plants (Marine and Dieter 1982). Many of the effects of cytoplasmic calcium can be traced to a modulation of the levels of certain specific enzymes. Amongst these, protein kinases, because of their pleiotropic function, rank as one of the most important (Dabrowska et al. 1977; Yamauchi and Fujisaira 1979), and calcium/calmodulin-dependent (Dalbrowska et al. 1977; Yamauchi and Fujisaira 1979) and calcium/phospholipid-dependent (Takai et al. 1979) protein kinases have been reported in animal tissue. Recently we briefly reported the presence of a calcium-dependent protein kinase in pea shoot preparations (Hetherington and Trewavas 1982) which is not stimulated by cyclic nucleolides. This enzyme was activated three- to fivefold by concentrations of calcium as
410
A.M. Hetherington and A. Trewavas: Pea membrane protein kinase
low as micromolar but the phosphorylation of some protein species could be increased some 20-fold. Activity was only weakly increased by calmodulin and inhibition by trifluoperazine (TFP), an inhibitor of calmodulin-dependent enzymes, was only observed at very high concentrations (Hetherington and Trewavas 1982). As many critical events in plant cell development occur at the membrane level, we have investigated in much greater detail some of the characteristics of this important regulatory enzyme. Subsequent reports have appeared from three other laboratories indicating the presence of calcium-activated protein kinases in plants (Polya and Davies 1982; Ranjeva et al. 1983 and Salimath and Marme 1983).
be calculated. The filters were washed overnight in the 10% TCA mixture, then washed in 5% TCA for 15 min, boiled for 15 min in fresh 10% TCA mix and allowed to cool for 30 rain. They were then washed for 15 s in methanol, and after drying, radioactivity was determined by Cerenkov counting. When Histone H1 was used as an exogenous substrate for the protein kinase, 25% TCA was used throughout the filter procedure and the concentration of calf histone H1 (Sigma) was 150 gg ml 1 during the incubation. Initial velocities were determined from the asymptote of the smoothed curve drawn through the points at zero time. Typically filter discs produced between 200 and 2000 cpm.
Gel electrophoresis. Sodium-dodecyl-sulphate polyacrylamidegel electrophoresis of labelled proteins was as previously described (Hetherington and Trewavas 1982) using 12% gels and detection of labelled bands by autoradiography.
Results and discussion Material and methods Plant material. Peas (Pisum sativum L. cv. Feltham First) were grown in the dark at 22~ for 10 d and the shoot material used. Most experiments used the bud which contained the unexpanded fourth and fifth leaves. In some experiments root material was collected from 7-d-old seedlings. )~-[32P]adenosine 5'-triphosphate (specific activity 111 TBq (mmol 1) was obtained from Amersham International plc (Amersham, Bucks., UK). Trifluoperazine was the kind gift of Smith, Kline and French Labs (Welwyn Garden City, Herts., UK). Trypsin, type X l l 2X crystalized; trypsin inhibitor, type 1-S; histone H I type VS; and other chemicals were obtained from Sigma London Chemical Co. (Poole, Dorset, UK).
Membrane preparation. Fresh bud material (20-30 g) was chilled and ground in 3 vols. ice-cold 0.3 M sorbitol, 50 m M 4-(2-hydroxyethyl)-l-piperazine ethanesulfonic acid (Hepes) pH 7.6, i m M ethylene diamine tetraacetic acid (EDTA), 0.1% phenylmethyl sulfonyl fluoride. After filtration, the solution was centrifuged at 12000 g for 10 min, the pellet discarded and the supernatant recentrifuged at 35 000 g for 30 rain. The pellet was resuspended in a total volume of 20-30 ml of 0.3 M sorbitol, 5 m M Hepes pH 7.0, 0.1 m M EDTA, layered over a cushion of 0.95 1 M sorbitol containing 5 m M Hepes, 0.1 m M EDTA, pH 7, and centrifuged at 150000g for 1 h in a swingout rotor at 2 ~ C. Alternatively, the pellet was resuspended and layered onto a 12-ml, 10-50% (w/w) linear sucrose gradient containing 5 m M Hepes and 0.1 m M E D T A and centrifuged at 80000 g for 16 h. The gradient was divided into 20 fractions, the density of each determined refractometrically and aliquots used for protein-kinase estimation and protein estimation (Lowry et al. 1951). Protein-kinase estimations. The final membrane pellet was resuspended in 6-8 ml 50 m M Hepes and protein-kinase estimations carried out using 7-[32P]ATP at 2 . 1 0 - 6 M as previously described (Hetherington and Trewavas 1982). Assays of each fraction were conducted in the presence or absence of 100 gM free calcium (Hetherington and Trewavas 1982) obtained from Analar CaC12.2H20 (BDH Chemical Co. (Poole, Dorset, UK)). Assays used time periods up to 45 s at 25 ~ C or up to 3 rain at 0 ~ C. The reaction was stopped by pipetting 75-gl aliquots onto 3 ~ filter discs which had been pretreated with 50 ~1 of 10% trichloracetic acid (TCA), 20 m M sodium pyrophosphate, and 10 m M EDTA. Sufficient 75-gl aliquots (usually three) were taken to enable estimates of initial velocity to
Fractionation of the calcium-dependent protein kinase. Table 1 shows that the majority of pellets showing calcium-dependent protein-kinase activity were present in cell fractions requiring 30 min at 20000-40000 g for their sedimentation. Additionally there was detectable activity in the 12000-g pellet. When highly purified mitochondria from this latter fraction (prepared for us by Dr. C.J. Leaver, Department of Botany, University of Edinburgh) were examined, they were found to contain only 10-15% of the total activity of the original fraction and to have a low calcium activation (twofold). The activity in this fraction may, therefore, be caused by large membrane fragments. It can be seen that there is little calcium-dependent protein kinase activity in either the nuclear-rich (1000-g) or 80 000-g supernatant fractions. Table 2 shows the distribution of calcium-deTable 1. Centrifugal distribution of calcium-dependent membrane-bound protein kinase in pea shoot homogenates. Etiolated pea bud tissue (20 g) was collected, homogenised and fractionated at different centrifugal speeds. Pelleted preparations were resuspended in 2 ml 50 m M Hepes medium and time courses of protein-kinase activity determined in the presence or absence of 100 gM free calcium. Initial velocities were estimated as asymptotes to the smoothly drawn curve at zero time Centrifugal force (g average)
1000 12 000 20000 40000 80000 Supernatant
Time (min)
15 10 15 30 30 -
Protein-kinase activity (pmol Pi m i n - 1 g g - 1 protein) + Ca z +
_ Ca 2 +
0.06 0.40 2.22 2.27 0.97 0.132
0.032 0.06 0.25 0.227 0.089 0.116
Activation
1.8 6.8 8.8 10 11 1.2
A.M. Hetherington and A. Trewavas: Pea membrane protein kinase Table 2. The level of calcium-dependent protein kinase in various etiolated pea tissues. Etiolated t0-d-old pea plants were sequentially divided into the apical bud tissue, the apical hook, a ~-cm e~ongat~ng zone, a 3-cm rn~ture zone ptus the node, and a 3-cm apical portion of the main root. Membranes were prepared from each fraction and the initial velocity of bound protein kinase determined in the presence and absence of calcium Tissue
Protein-kinase initial velocity (cpm gg-1 protein) +Ca
Bud Apical hook Elongating zone Mature zone + node Root
0.60 0.182 0.402 0.452 0.26
2+
_Ca
0.065 0.023 0.09 0.078 0.066
411
18
1.165 1.136 d e n s i t y
l
6 o
"o
Activation
8 :>.
2+
6
9.3 8.0 4.5 5.8 3.7
-~
1
2
c
pendent protein kinase in various etiolated pea tissues. As the bud tissue exhibited the greatest calcium activation it has been used in subsequent experiments. Membrane-bound calcium-dependent protein kinase was also detected in cultured carrot cells, maize coleoptiles, tobacco pith and crowngall cells indicating its likely wide distribution. Fractionation o f the membrane-bound proteinkinase preparation by isopycnic centrifugation is shown in Fig. 1. There are at least two components. One is represented by a shoulder and has a density of approx. 1.16 g cm -3, and the second major component has a peak density of 1.136 g c m - a ; it should be noted that the activation ratio ( + Ca 2 + / - Ca a +) is higher in the lightermembrane fraction. This latter density is characteristic of pea epicotyl plasma membrane (Hendrik and Pierce 1980). To confirm this we also used the fully characterised procedure of Rasi-Caldogno et al. (1982) devised for pea tissue for obtaining fractions enriched in plasma membrane, and again assayed the calcium-dependent protein kinase. The results in Table 3 confirm the general picture of the isopycnic-gradient results. The best calcium activations are observed in plasma-membrane-enriched fractions but the enzyme is also located in other cell membranes of differing densities. The calcium-dependent and calcium-independent protein kinase may be different enzymes in pea bud tissues. There are two pieces of evidence which indicate that this may be the case. Figure 2a shows that calf histone HI will act as a substrate for the calcium-activated protein kinase but not the nonactivated form. Neither enzyme will phosphorylate ion-free casein and the other histones (H2-H4) have been found to be inhibitory. Figure 2 b shows
~3 ~- 0
0 0
10 20 F r a c t i o n no.
03
Fig. 1. Isopycnic separation of calcium-activated protein kinase on sucrose density gradients. Pea bud tissue (10 g) was collected, homogenised and a crude microsomal preparation made. The final pellet was suspended in 1 ml 10 mM Hepes and layered on a 14-ml linear sucrose gradient ( 1 ~ 5 0 % , w/v) and centrifuged overnight at 80000 g. The gradient was divided into 20 fractions and aliquots used for density determination, protein estimation and determination of calcium-dependent and calcium-independent protein-kinase activity Table 3. Distribution of calcium-dependent protein kinase amongst membrane fractions using the method of Rasi-Caldogno et al. (1982) for plasma-membrane preparation. Pea bud tissue was collected and the two fractions described by RasiCaldogno et al. (1982) were prepared using the metrizamide step gradients and high bovine-serum-albumin concentrations prepared as described in their paper. Two preparations were made, one at 10 mM Mg-acetate and the other at 1 mM. According to Rasi-Caldogno et al. (1982) this leads to prepalations with markedly different calcium-uptake capabilities. Protein-kinase initial velocities were determined as described in the Material and methods section and in Table 1 Fraction
Plasma-membrane enriched Heavy membrane
Mg 2 + (mM)
l0 1 10
Protein-kinase initial velocity (cpm ktg-1 protein) + C a 2+
_ C a 2+
3.15 2.6 1.65
0.055 0.045 0.052
Acti~ vation
5.7 5.7 3A
autoradiographs o f the peptides phosphorylated 'in vitro' using [32p]ATP by the calcium-activated and the non-activated forms. Whereas the calciumactivated enzyme phosphorylates many proteins, the calcium-independent protein kinase primarily
412
A.M. Hetherington and A. Trewavas: Pea membrane protein kinase
x is
o x
E
5
4
8
6 +Ca
~
|
31
6
cJ
.E
2~
g
3
O r
2c .E
1 _~ ~
__ 9
-
-Ca A
-~
o
D_
,q
,
a
50 100 Hist0ne HI (/ag)
~J
u o
2
150
>
Ik"A----a-
A
it
(D rn
4
9
I
I
I
I
!
J
5 10 15 Added phospho[ipase C (/at)
2a
Fig. 2a, b. Substrate specificity of the calcium-activated and calcium-independent protein kinase of pea shoot membranes, a Variation in pea bud protein-kinase activity with histone H1 concentration. Microsomal membranes were incubated with increasing levels of histone H1 in the presence or absence of 100 gM free calcium and the initial velocity of protein-kinase activity determined. b Gel electrophoresis of pea bud membrane proteins phosphorylated in the presence or absence of 100 gM free Ca 2+. Proteins were phosphorylated for 45 s and the reaction stopped by adding an equal volume of 2.5% sodium lauryl sulphate, 2-amino-2(hydroxymethyl)-l,3-propanediol-chloride (Tris-C1) (0.13 M, pH 6.7) mercaptoethanol (2%) and sucrose (10%). After boiling for 1 min, the solubilised proteins were separated on 12% acrylamide gels and dried and autoradiographed. Arrows mark the position of proteins of M r 40000 and 59 000 specifically phosphorylated in the absence of calcium Fig. 3. Effects of various concentrations of phospholipase C on the calcium-activated protein kinase of pea bud membranes. Microsomal membranes were suspended in 50 mM Hepes (pH 7.8) and to each a designated number of microliters of phospholipase C (Sigma Type V) were added for 2 min at 0 ~ C. The reaction was stopped with 100 gM ethylenediaminetetraacetic acid pH 7.2 (Jolliot et al. 1982). After centrifugation, the membrane pellet was resuspended and calcium-dependent and calcium-independent protein kinases were assayed using histone H1 (150 gg ml-1) as an exogenous substrate
phosphorylates two peptides running at a relative molecular mass (Mr) of 59000 and 40000 with only traces of others.
Inhibition of the calcium-activated protein kinase by phospholipids and diglycerides. A calcium-activated and diglyceride-activated protein kinase (protein kinase C) has been found to be a mediator of growth signals in many animal cells and to be the receptor for the tumour-promoting phorbol esters (Kuo et al. 1980; Takai et al. 1982; Niedel et al. 1983). This enzyme is also membrane-bound. Is the pea bud kinase a plant form of protein kinase C? We have tried to demonstrate diglyceride activation in a number of experiments and Table 4 and Fig. 3 summarise these data. In no case have we achieved activation of the pea bud protein kinase. Routinely we have observed inhibition of between 40-50% both for membrane-bound protein kinases and soluble protein kinases. As this does not rule out the possibili-
ty that sensitivity may be based on an absolute requirement for a particular molecular species of lipid, endogenous diacylglycerols were generated, using the method of Jolliot et al. (1982) by brief exposure of the membrane preparation to phospholipase C. Figure 3 shows that no stimulatory effects were observed, instead the observed inhibition is probably indicative of the hydrophobic nature of the enzyme and indicates lipid-dependence for activity as found in cytidine-diphosphate-choline diacylglycerol phosphorylcholine phosphotransferase (Jolliot et al. 1982).
Inhibition of calcium activation by TFP and tryptic digestion experiments indicate the presence of at least two calcium-regulated protein kinases. Since many calcium-regulated enzymes are calmodulindependent we have examined whether this membrane-bound protein kinase is likewise calmodulindependent. In our previous paper (Hetherington and Trewavas 1982) we reported that added bovine
A.M. Hetherington and A. Trewavas: Pea membrane protein kinase Table 4. Sensitivity of membrane and soluble protein kinase to added lipids. Microsomal membranes were prepared as described in Material and methods. Soluble protein kinase represented a 40000-g supernatant. Phosphatidylinositol and diolein suspensions were prepared by drying the stock solution and then resuspending in incubation buffer with sonication. Resuspension buffer used for the controls was sonicated for the same period
Treatment
M e m b r a n e fraction
Soluble fraction
Initial velocity protein kinase
Ca 2 + activation
Initial velocity protein kinase
Ca 2 § activation
Control
6680
15.5
495
1.8
+ Phosphatidylinositol (10 lag ml-1)
4250
11.9
280
1.5
+ Diolein (1.25 lag m l - 1)
3 750
413
X
q E 3
2
~o
.I
D
.E 09 o
I
n
10.6
70
1.1
calmodulin led to slight increases in assayable protein kinase. These increases were small (20%) but we have found them quite reproducible. Replacing bovine by pea calmodulin (kindly prepared for us by Miss E. Allan, Department of Botany, University of Edinburgh) did not improve this degree of stimulation (data not shown). In addition we found that TFP (an inhibitor of calmodulin-enzyme interactions) only inhibited the calcium protein kinase at millimolar concentrations, an unacceptably high concentration for specific inhibition. We have considerably improved this initially unsatisfactory observation. We have found (like Saitoh et al. (1982)) that prior incubation of membrane preparations in ethylene glycol-bis(fl-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA) for 2 h enables inhibition by lower concentrations of TFP to be demonstrated. Figure 4 summarises these data. Inhibition can be detected at 0.1 m M TFP but there is some sort of levelling out between 0.2 and 0.3 m M TFP when there is still substantial residual calcium-activated protein kinase. Taken at their face value these results indicate that there may be a calmodulin-dependent and a calmodulinindependent, but in both cases calcium-activated, form of protein kinase in these membranes. However, inhibitors are rarely completely specific and the only critical method of establishing this point will be complete solubilisation and separation of the bound protein kinases. The possible presence of two calcium-regulated protein kinases is perhaps more clearly indicated by the data in Fig. 5. Membrane preparations were
9
I
I
0.1
0
I
I
0.2
I
D
0.3
T r i f [ u o p e r a z i n e (raM) Fig. 4. Inhibition of calcium-dependent protein kinase from pea bud membranes by trifluoperazine. Microsomal membrane preparations were prepared and incubated in 2 m M E G T A 50 m M Hepes pH 7.2 for 2 h at 0 ~ C. To appropriate samples, TFP was added at the concentrations shown in the figure and then finally 2.1 m M CaC12 for assessing calcium dependence. Protein-kinase initial velocities were then determined in the usual way using histone H1 (150 lag m1-1) as exogenous substrate
% X | t 1 I
6
; 10
i .2
ul
2~ .E
:>
5
2 ?,
o 0_
