195

Biochimica et Biophysica Acta, 1038 (1990) 195-198 Elsevier BBAPRO 33619

Regulation of the calcium-activated neutral proteinase (CANP) of bovine brain by myelin lipids A r u n K. C h a k r a b a r t i , S o m s a n k a r D a s g u p t a , N a r e n L. B a n i k a n d E d w a r d L. H o g a n Department of Neurology, Medical University of South Carolina, Charleston, SC (U.S.A.) (Received 4 October 1989)

Key words: Calpain; Phospholipid; Galactolipid; Myelin; Autolysis; Enzyme activation

Since calcium-activated neutral proteinase (CANP; calpain) activation occurs at the plasmalemma and the enzyme is found in myelin, we examined myelin lipid activation of brain CANP. Purified lipids were dried, sonicated and incubated with purified myelin CANP. The CANP was assayed using [14C]azocasein as substrate and the Ca2÷ concentration ranged from 2 pM for pCANP to 5 mM for mCANP. Phosphatidylinositol (PI), phosphatidylserine (PS) and dioleoylglycerol stimulated the mCANP activity by 193, 89 and 78%, respectively. PI stimulated both m- and /tCANP in a concentration-dependent manner, while phosphatidylcholine was least effective. Cerebroside and sulfatide at higher concentrations (750 pM) were stimulatory. The phospholipid (PL)-mediated activation was inhibited by the PL-binding drug trifluoperazine. PI reduced the Ca2+ requirement for CANPs significantly (20-fold). These results suggest that acidic lipids and particularly acidic phospholipids activate membrane CANP.

Introduction Two forms of calcium-activated neutral proteinase (CANP; calpain), /~CANP active at #M Ca 2+ and mCANP active at mM Ca 2+ concentrations have been found in muscle, kidney, liver, heart, brain and other tissue [1,2]. Initial studies allocated CANP to the cytosol, but subsequently membrane association (hydrophobic binding) of mCANP has also been established [3-6]. The mCANP in the central nervous system (CNS) is both cytosolic and membrane bound, with the majority associated with myelin [7-10]. The physiological role of mCANP must be explained in terms of its activation requirement for mM Ca 2÷ concentration, which far exceeds the intracellular Ca 2+ concentration. To explore its biological function, we have examined the activation of CNS mCANP at a physiological intracellular Ca 2+ concentration. Autolysis of skeletal muscle CANP i n the presence of phosphatidylinositol (PI) in vitro greatly reduces the Ca 2÷ concentration for CANP

Abbreviations: PI, phosphatidylinositol; PS, phosphatidylserine; PL, phospholipid; CANP, calcium-activated neutral proteinase, calpain; CNS, central nervous system; buffer A, 0.5 mM Tris-acetate buffer (pH 7.5). Correspondence: N.L. Banik, Department of Neurology, Medical University of South Carolina, Charleston, SC 29425, U.S.A.

activation and suggests one possible means for CANP regulation in cell membranes [11-13]. In view of this, and because myelin is enriched in mCANP [8,9] and is 70% lipid [14], we studied several myelin lipids including PI, and for the first time galactolipids, as regulators of CNS myelin mCANP. The CANP in myelin may be involved in the turnover of myelin proteins and in the myelinolysis occurring in demyelinating diseases. Preliminary accounts of this study have been presented [15]. Materials and Methods [14C]Formaldehyde was purchased from American Radiolabeled Chemicals (St. Louis, MO). Phospholipids, galactocerebroside, sulfatide and sphingosine prepared from bovine brain were of highest quality grade (98-99%) and obtained from Sigma Chemical Co. (St. Louis, MO). The galactocerebroside contained ahydroxy fatty acid, and PI contained stearic and arachidonic acids. Azocasein, chlorpromazine and trifluoperazine were also obtained from Sigma Chemical Co. All other chemicals were obtained from Fisher Scientific Co. (Fair Lawn, N J). Methods

Lipids were stored at 1-10 mg/ml in chloroform/ methanol (2:1, v/v) at - 2 0 ° C . Lipids (appropriate amounts) were dried on the bottom of assay tubes

0167-4838/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)

196 under a stream of nitrogen and sonicated in buffer A (0.5 mM Tris-acetate buffer, p H 7.5) prior to assay. This procedure produced mainly single unilamellar vesicles [16]. Enzyme purification, m C A N P and # C A N P were purified from myelin isolated from bovine brain according to Chakrabarti and Banik [17]. /~CANP was also purified from brain cytosol. Fractions containing the 100 mM NaC1 extract from the first DEAE-cellulose column were pooled, adjusted to 500 mM NaC1 concentration and passed through a phenyl-Sepharose column. /~CANP was eluted with 1% ethylene glycol in a buffer containing 20 mM Tris-acetate (pH 7.5), 1.5 mM EDTA, 2.5 mM fl-mercaptoethanol and 1 mM N a N 3. E n z y m e assays. Substrate [14C]azocasein was prepared by reductive alkylation with [14C]formaldehyde [18], and assays were according to the method described earlier [9]. The assay was initiated by adding an appropriate amount of purified enzyme (0.5-1.0/~g) to 65 /~1 of buffer A (with or without lipid) containing 4 m g / m l of [14C]azocasein (approx. 59.103 cpm/12.5/~1) and 10 mM KC1, 1.0% fl-mercaptoethanol along with 2 /~M and 5 mM CaC12 for determining/~- and m C A N P activity, respectively. EGTA (2/~M and 5 mM) replaced Ca 2+ in the control samples. After 30 min incubation at 37°C, the reaction was terminated with 3 /~1 of 6% unlabeled azocasein and 100/~1 of 20% ( w / v ) trichloroacetic acid and it was then centrifuged at 4000 rpm for 1 h. Following centrifugation the acid-soluble radioactivity was determined, and that in the acid-soluble fraction less that of the E G T A control was taken as the C A N P activity, which is expressed as cpm- 10 -3 per h per mg protein. Results

The effects of phospholipids and glycolipids upon purified myelin C A N P were studied. A concentrationdependent stimulation of C A N P activity was produced by PI, dioleoylglycerol, phosphatidylserine (PS) and phosphatidylcholine (PC) (Fig. 1). Maximum stimulation occurred with 150/~M lipid and the activity did not increase at higher concentrations (300 /~M), except for PI which stimulated activity to 300 /~M. PC was least active with only a modest activity increase at 150 ~tM. PI stimulated most (193%), and PS and dioleoylglycerol increased C A N P activity by 89 and 78%, respectively. We found that cerebroside and sulfatide did not affect enzyme activity at low concentrations (75 /~M), but stimulated activities (145 and 111%) significantly at higher concentrations (750 I~M of cerebroside and sulfatide, respectively). To our knowledge this is the first report showing such a stimulation of C A N P activity by myelin specific galactolipids. Other lipids (sphingosine, sphingomyelin, ethanolamine, arachidonic acid, L - a - p h o s p h a t i d i c acid a n d 1 2 - O - t e t r a -

bc

150

'7

100

z, < (..)

50

'

l:f9 rX

o

/.

0

-" ,

•

*

200

400

600

~

800

Lipid Concentration

1000

(/~M)

Fig. 1. The effect of phospholipids on myelin mCANP activity. O, Phosphatidylinositiol; A, dioleoyiglycerol;II, phosphatidyiserine; and O, phosphatidylcholine. All values are expressed as mean + S.E. of five observations. S.E. values are too small to be visible in the figure. Experimental conditions are described in the text. Specific activity of enzyme was 10.55:0.8 cpm.10-6 nag protein in absence of any added lipid.

250

%" O3

~

200 150

-~

:_S IO0 ~8 ~L

7

50 0

t~""'A

0

i

i

i

i

i

100 200 500 400 500 600 Phosphatidylinositol Concentration (/~M)

Fig. 2. The effect of phosphatidylinositol on #- and mCANP activity. A, /tCANP; and o, mCANP. All values are expressed as mean + S.E. of five observations. Assay +0.1 and 10.2+0.9 cpm.10-6/mg protein, respectively, in absence of any added lipid.

TABLE I Effect of trifluoperazine on lipid mediated activation of mCANP Values are expressed as mean + S.E. of five different observations. Concentrations of PI, DG and sulfatide are 300, 100 and 750 #M, respectively. Equimolar amounts of TFP was used in each case. TFP concentration was 750 /~M for enzyme alone. Percent change was calculated compared to enzyme activity alone. Specific activity of enzyme was 10.2+0.9 cpm. 10-6/mg protein in the absence of any lipid.

Enzyme+ Enzyme+ Enzyme+ Enzyme+ Enzyme+ Enzyme+ Enzyme+

Conditions

% Stimulation

Trifluoperazine(TFP) Phosphatidylinositol(PI) PI and TFP Dioleoylglycerol(DG) DG and TFP Sulfatide Sulfatideand TFP

4.9 + 0.8 193.5+ 10.7 7.2+ 1.2 71.3 + 4.2 0.4+ 0.1 111.7+ 12.6 7.3+ 1.0

197 30 >~ >

r,3

z

25

'o20

CL

z

x

15

0

E

\\

u 10

o

5 0

\',

0.25

0.50

Calcium

0.75

1.00

Concentration

'

'

2.50

5.00

(mM)

Fig. 3. The effect of phospholipids o n C a 2+ sensitivity of myelin mCANP. (3, Control (enzyme only); • phosphatidylserine; I1, dioleoylglycerol; and o, phosphatidylinositol.All values are expressed as mean + S.E. of five observations. The concentration of PI and dioleoylglycerolwas 100/~M each, and that of PS was 300 #M.

decanoylphorbol/13-acetate) produced marginal or no change (results not shown). For example, arachidonic acid at lower concentrations (75/~M) inhibited mCANP activity (17.7%) whereas at higher concentrations (750 #M), this inhibition was abolished. The effect of phosphatidic acid and phorbol acetate at higher concentration was to some extent inhibitory. The effect of PI on /tCANP was also studied, and a comparative concentration-dependent curve for stimulation of #- and mCANP is seen in Fig. 2. The stimulation of both reached a plateau at 300/~M, with greater stimulation of mCANP than /~CANP at all concentrations of PI. Stimulation with mono- and diphosphatidylinositol was comparable that with triphosphatidylinositol (results not shown). There was also a significant stimulation of CANP activity when total lipid (1 mg, chloroform/methanol, 2 : 1 ( v / v ) extract) of myelin was incubated with purified enzyme. N o enzyme activity or calcium was found associated with lipids in controls. 20

>

x 10

~\

(.D

E&

5

0.25 0.50 Q75

1.00 2.00 3.00 4.00 5.00

Calcium Concentration

('mM)

Fig. 4. The effect of cerebrosideand sulfatide o n C a 2+ sensitivityof myelin mCANP, o, Enzyme only; zx, sulfatide; and rq, cerebroside. All values are expressedas mean 5: S.E. of five observations.

Trifluoperazine, a kinase C inhibitor a n d / o r phospholipid binding antagonist, did not alter CANP activity, but did prevent the stimulation of CANP found with PI, dioleoylglycerol and sulfatide (Table I). Acidic lipids, mainly PS, PI, and metabolites of PI, dioleoylglycerol reduced significantly the Ca 2+ requirement for CANP activation. The extent of this reduction for conversion of m C A N P into the more Ca 2+-sensitive form varied. There was a 20-fold reduction in Ca 2+ requirement with PI and reductions of 15- and 10-fold with dioleoylglycerol and PS, respectively (Fig. 3), whereas sulfatide and cerebroside reduced the Ca 2+ requirement to the extent of 20- and 10-fold, respectively (Fig. 4). Discussion

PI and to some extent dioleoylglycerol (both lipids at 75 /~M concentration) modulate the Ca 2+ dependence of a muscle neutral proteinase activated by Ca 2+ (mCANP) [11,12]. We confirm this for brain CANP and find that myelin phospholipids (25-300/~M) and glycolipids (cerebroside and sulfatide at 750 /~M concentration) can also modulate the brain enzyme. At lower concentration both glycolipids are ineffective. The greatest effects are found with PI, PS and dioleoylglycerol and the order of efficacy for phospholipid activation is PI > PS > dioleoylglycerol > PC. It is important to note, however, that stimulation of CANP by PI is much greater (about 2.5-fold) than that of galactocerebroside. This is because the stimulation obtained by PI (175%) at a lower concentration (300 /~M), while cerebroside stimulated to the same extent at a 2.5-fold higher concentration. The effects of cerebroside and sulfatide at a concentration of 750 /~M were greater than with PS and dioleoylglycerol at this concentration. Although the cause of this effect is not yet clear, it may be due to the carbohydrate moiety binding and subsequent stimulation of CANP as reported earlier by Zimmerman and Schlaepfer [19]. Free arachidonic acid or other constituent moieties (sphingosine) of stimulatory lipids (e.g., cerebroside and sulfatide) have either very little or no effect on CANP, suggesting that intact lipid is necessary. Our present finding of slight inhibition of brain CANP by arachidonic acid is consistent with that of Zalewska et al. [20]. Autolytic activation of m C A N P is unlikely to occur in the cell since a mM concentration of free calcium is necessary. A mechanism by which the mCANP may be converted, in the presence of Ca 2+ and other factors, into a Ca2+-sensitive form active at physiological Ca 2+ level has been proposed [13]. It implies that the membrane m C A N P binds with membrane lipids (i.e., phospholipid a n d / o r glycolipids). Our experiments with total myelin lipid also suggest such contact of enzyme with lipids in the membrane. An interaction of the N-termi-

198 nal hydrophobic region of the 30 kDa subunit of mCANP and PI has been reported [12]. PI is a good candidate to mediate the reduction in Ca 2+ requirement for autolysis of muscle mCANP into #M Ca2+-sensitive form active at intracellular Ca 2÷ concentration [11]. Cerebroside and sulfatide might also regulate mCANP in this fashion. The carbohydrate affinity revealed by CANP binding to an agarose column [19] raises the possibility that the carbohydrate moiety of these giycolipids bind to the enzyme. Our study shows conversion of brain mCANP into a more Ca2+-sensitive form in the presence of PI. The mechanism of /~CANP interaction with PI and resultant change in catalytic activity may be same as the conversion of mCANP since the structure of 30 kDa (binding and regulatory subunit) subunits of the/~- and mCANPs are identical [21]. It is noteworthy that the requirement of PI for the conversion of mCANP to the /~ form is much less than for #CANP stimulation. Trifluoperazine, a phospholipid binding antagonist and a protein kinase C inhibitor, had no effect upon mCANP activity but did inhibit lipid (PI, dioleoylglycerol and sulfatide) modulation of mCANP. This suggests that lipid-mediated activation involves binding (hydrophobic) of lipids to the enzyme. We also found that the Ca 2+ requirement of brain (myelin) mCANP is reduced significantly by lipids (PI, PS, dioleoylglycerol, cerebroside and sulfatide) varying with lipid structure. This reduction in Ca 2÷ requirement is consistent with a lipid (PI) action to regulate autolysis of muscle mCANP to a Ca2+-sensitive #M form active at a physiological intracellular Ca 2÷ concentration [11,12,15]. Myelin contains CANP, phospholipids and glycolipids [8,9,14] and therefore has the catalytic potential for conversion of mCANP into the more Ca2+-sensitive /~CANP form active at physiological Ca 2÷ concentration. Although in our experiments the reduction of Ca 2+ concentration in the presence of lipids did not reach intracellular level (1 #M or less), the Ca 2÷ requirement for mCANP was significantly reduced. On the other hand, the amount of PI (0.009 #mol) and cerebroside (0.076 /~mol) which stimulated CANP activity are comparable to the estimated amount of these lipids (0.076 /~mol and 0.48 #mol/mg myelin lipid, respectively) in the myelin membrane. In addition to Ca 2+, lipids and calpastatin, there may be other factors involved in the regulation of C A N P that have not yet been defined.

Acknowledgements This work was supported in part by grants NS-21353 and NS-11066 from NIH-NINDS. We thank Mrs. Elaine Terry and Denise Lobo for their technical help, and Ms. Vickie Edwards, Ms. Johanna Gehlken and Mrs. Joan Kingsley for their excellent secretarial assistance.

References 1 Murachi, T. (1984) Biochem. Soc. Symp. 49, 149-167. 2 Suzuki, K. (1987) Trends Biochem. Sci. 12 (March), 103-105. 3 Banik, N.L., McAlhaney, W.W. and Hogan, E.L. (1985) J. Neurochem. 45, 587-588. 4 Baudry, M., Bundman, M.C., Smith, E.K. and Lynch, G.S. (1987) Science 212, 937. 5 Malik, M.N., Fenko, M.D. and Wisniewski, H.M. (1984) Neurochem. Res. 9, 233-240. 6 Sato, S. and Miyatake, T. (1982) Biomed. Res. 3, 461-464. 7 Banik, N.L., Chakrabarti, A.K. and Hogan, E.L. (1987) Life Sci. 41, 1089-1095. 8 Yanagisawa, K., Sato, S., O'Shannessy, D.J., Quarles, R.H., Suzuki, K. and Miyatake, T. (1988) J. Neurochem. 51, 803-807. 9 Chakrabarti, A.K., Yoshida, Y., Powers, J.M., Singh, I., Hogan, E.L. and Banik, N.L. (1988) J. Neurosci. Res. 20, 351-358. 10 DeRosbo, N.K., Carnegie, P.R. and Bernard, C.C. A (1986) J. Neurochem. 47, 1007-1012. 11 Coolican, S.A. and Hathaway, D.R. (1984) J. Biol. Chem. 259 (19), 11627-11630. 12 Imajoh, S., Kawasaki, H. and Suzuki, K. (1986) J. Biochem. 99, 1281-1284. 13 Mellgren, R.L. (1987) FASEB J. 1, 110-115. 14 Morell, P., Quarles, R.H., and Norton, W.T. (1989) in Basic Neurochemistry (Siegel, G.J., Albers, R.W., Katzman, R., Agranoff, B.W. and Molinoff, P.B., eds.), Vol. 4, pp. 109-136, Raven Press, New York. 15 Chakrabarti, A.K., Dasgupta, S., Banik, N.L. and Hogan, E.L. (1988) J. Cell Biol. 107, 393a. 16 Garret, C., Cottin, P., Dufourcq, J. and Ducastaing, A. (1988). FEBS Lett. 227, 209-214. 17 Chakrabarti, A.K. and Banik, N.L. (1988) Neurochem. Res. 13, 127-134. 18 Dottavio-Martin, D. and Ravel, J.M. (1978) Anal. Biochem. 87, 562-565. 19 Zimmerman, U.-J.P. and Schlaepfer, W.W. (1988) J. Biol. Chem. 263 (24), 11609-11612. 20 Zalewska, T., Strosznajder, J. and Kawashima, S. (1988) Neurochem. Pathol. 8, 79-89. 21 Kawasaki, H., Imajoh, S., Kawashima, S., Hayashi, H. and Suzuki, K. (1986) J. Biochem. (Tokyo) 99, 1525-1532.