Investigation of the interaction of m-calpain with phospholipids: calpain-phospholipid interactions

BB

ELSEVIER

Biochimica et Biophysica Acta 1293 (1996) 201-206

Biochi~ic~a et BiophysicaA~ta

Investigation of the interaction of m-calpain with phospholipids: calpain-phospholipid interactions J. Simon C. Arthur *'l, Catherine Crawford Laboratory of Molecular Biophysics. Department of Biochemistry, Unicersit3"of OxJbrd. The Rex Richard's Building. South Parks Road, Oxford, OXI 3QU, UK Received 4 September 1995; revised 10 November 1995; accepted 23 November 1995

Abstract Phosphatidyl inositol, pho:~phatidyl choline, phosphatidyl glycerol, phosphatidyl serine, phosphatidyl ethanolamine, phosphatidic acid and sphingomyelin were all found to be effective at reducing the Ca 2÷ requirement for m-calpain autolysis. In the absence of phospholipid, pig kidney m-calpain required 1.4 mM Ca 2+ for 50% autolysis under the assay conditions used. Phospholipids caused a reduction in this Ca 2- requirement to a value between 0.45 mM Ca 2÷ for phosphatidyl glycerol and 1.1 mM Ca 2+ for phosphatidyl ethanolamine. Previous studies (Crawford, C., Brown, N.R. and Willis, A.C. (1990) Biochem. J. 265,575-579) have shown that the most probable site for phospholipid interaction in calpain is the N-terminal region between residues 39 to 62 of the small subunit of calpain (G~vTAMRILGG). In this study we examine the possible role of this Gj7TAMRILGG region. Three synthetic peptides corresponding to parts of this sequence were used to examine the phospholipid binding sequence. Analysis of the phospholipid vesicle binding properties of these peptides suggested that both the TAMRIL and polyglycine sequences were required for binding to phosphatidyl inositol vesicles. Keywords: Calpain; Autolysis; Phospolipid; Peptide; Peptide-phospholipid interaction

1. Introduction Calpains are Ca2+-dependent cysteine proteinases consisting of two subunits. The large subunit (80 kDa) contains the active-site domain, which shows some similarity to papain-like cysteine proteinases, and a C-terminal Ca2+-binding domain similar to calmodulin. The small subunit (30 kDa) has a gly.:ine rich N-terminal domain and a calmodulin like C-termiaal Ca 2+ binding domain. Two distinct forms of calpain have been well characterised, m-calpain and /x-calpain, which have different large subunits but identical small subunits. The two forms differ in their requirement for Ca 2+ : m-calpain can be activated in vitro by millimolar level,; of Ca 2+ while /,-calpain requires micromolar Ca 2÷ levels [ I - 4 ] . Both forms undergo

Abbreviations: PI, phosphati:tyl inositol; PC, phosphatidyl choline; PG, phosphatidyl glycerol; PS, phosphatidyl serine; PE, phosphatidyl ethanolamine; PA, phosphatidic acid, • Corresponding author. Fax: + 1 (613) 5452497; e-mail: [email protected]. i Present address: Department of Biochemistry, Queens University, Kingston, Ontario, Canada, K7L 3N6. 0167-4838/96/$15.00 © 1996 F~lsevierScience B.V. All rights reserved SSDI 0 1 6 7 - 4 8 3 8 ( 9 5 ) 0 0 2 4 3 - X

autolysis in the presence of C a 2+. The initial autolytic cleavages remove the N-termini from both subunits and generate an enzyme with a lower Ca 2. requirement than the intact calpain [5-7]. Further cleavages occur if exposure to Ca 2+ is prolonged, leading to an eventual loss of activity [8]. In order to understand the role of m-calpain in vivo it is important to know how the enzyme is activated. As mcalpain requires millimolar levels of Ca 2- for both autolysis and proteolytic activity in vitro, the Ca 2+ concentrations found in normal cells are unlikely to be able to activate m-calpain directly. This suggests that other factors are involved in activating m-calpain in vivo. Previous reports have suggested that certain phospholipids can lower the Ca 2 + requirement for calpain autolysis; however, there is some disagreement as to which phospholipids have this effect. Coolican and Hathaway [9] found that PI could decrease the Ca 2- requirement for autolysis of m-calpain while PS, PC, PE and PA did not. In contrast Pontremoli et al. [10] found that PC, PI and PE and to a lesser extent PS could decrease the Ca -,+ requirement for autolysis in mcalpain while sphingomyelin did not. In addition they also suggest that unsaturated fatty acid chains must be present

202

J.S.C. Arthur, C. Cra~ford / Biochimica et Biophysica Acta 1293 (1996) 201-206

in the phospholipid molecule to allow the interaction with calpain to occur. Charakbarti et al. [11] showed that PI, PS and PC could cause an increase in calpain activity while PE and PA could not. The N-terminal region of the small subunit is essential for the interaction of calpain with phospholipid. A region between 39 and 62 amino acids from the N-terminal has been identified as being required for the interaction of calpain with phospholipid [12,13] In this paper we define which phospholipids affect the Ca 2- requirement of m-calpain. In addition we identify a phospholipid binding site in calpain using peptides homologous to parts of N-terminal domain of the small subunit.

2. Materials

and

methods

2.1. Materials Highly purified phospholipids were obtained as chloroform/methanol stocks from Lipid Products (South Nut-

A

1

2

3

4

5

6

field, Redhill, Surrey, UK). The biological sources of the lipids were as follows: PI from wheatgerm, PS and sphingomyelin from bovine spinal chord, PC and PE from egg yolk and PA and PG as derivatives of PC. The lipids were prepared and stored under nitrogen; butylated hydroxytoluene (0.05%) was added to the stock solutions to minimise subsequent oxidation of the lipid. Sephadex G-10 and Sepharose 4B gel filtration media were obtained from Pharmacia (Milton Keynes, Bucks•, UK). Tritium labelled phosphatidyl inositol was obtained from Amersham (Lincon Green, Aylesbury, Bucks., UK).

2.2• Method~ Purification of m-calpain. Calpain was purified from pig kidney as previously described [13]. Preparation of phospholipid cesicles. Phospholipid (1 /xmol) was dried down under nitrogen from the stock solution and resuspended in 1 ml of 0.193 mM Tris-HC1, 0.187 mM KCI (pH 7.5). To disperse the lipid the solution was sonicated for four 30 second bursts on a MSE soniprep

B

1

2

3

4

5

6

i-~,.:':-::~--.:.,i.: :.~.~:.~;;i~-'.'.~::.4~,'(.~-*..,,~-'.~-~;"~. • ~,7,, .,".~-~'.... ~:; 94

94

68

:

,, : ' =

. rr .~jr~.~...::

.. ",~ ' ,. ,

-:t."'.:..,,.~,.'

~

68

45

" "

.

~,~'i...

::2.,'~',.(.~

, ;.i=r "., ~4:,v,

- h -:'..%.!.

-

~

..?:.. :~:'~ ~ :'._~.: •

"..4

~-.

: ::~

:....

45 u

~

-

•

•

35

..........

~

....

35

?... ..

"

"

';.

r

¢%

" :i7".:.:.

.

: .y~..21 ,,....-.

. ~.~

~'[;~-:'

.

" .

:....

.

•

.. •

~

.,=,...;.:.;-:.

"..

~

_

" . ....

"..

.

.".i

"

: :.;,

•

..-

..:

..

,

:[:OmII, ~:~,."

;~,-;'.~d;.~r'.

"

. .

".~.

? ~ -". . . . . ;!%:,

.. •

.

.....i.....,

.-'-~ ~ •

.

.., ;~!':,~ "

.

:- ... -

, ' : ..'5":~.

,.~_~-,~--::.'F~;~. ,~:"" : ~:..'~47".~ ~

.

.... "~

Fig. 1. Effect of phospholipid on m-calpain autolysis• Calpain was incubated, at 30°C, in the presence of 0.0 (lane 1), 0.25 (lane 2). 0.50 (lane 3), 0.75 (lane 4), 1.0 (lane 5) and 2.5 mM Ca 2+ (lane 6). Autolysis was stopped after 1 min by addition of gel sample buffer. SDS gel electrophoresis was then used to determine the extent of autolysis. (a) Extent of autolysis of the small subunit after 1 min in the absence of phospholipid at the various Ca 2÷ concentrations. (b) Extent of autolysis of the small subunit after I min in 0.5 mM phosphatidyl glycerol at the various Ca 2÷ concentrations.

J.S.C. Arthur, C. Crawford / Biochimica

150, using power level 8 at room temperature. All the vesicles were prepared in this way except for those containing PE. PE dispersions were produced by sonication in water as PE aggregated in the above buffer. PI vesicles for the coelution experiments were prepared in 25 mM Tris-HCl, 100 raM NaCI, 5 mM EDTA (pH 7.5) by using a bath sonicator. Electrophoresis. S D S / p o l y a c r y l a m i d e - g e l electrophoresis was performed using 12.5% ( w / v ) polyacrylamide gels and the buffer system as previously described [14]. Samples were prepared by heating in a boiling water bath for 3 min with an equal volume of 125 mM Tris-HCl buffer (pH 6.8), containing 20% ( w / v ) glycerol, 10% 100

et

Biophysica Acta 1293 (1996) 201-206 100

8O

6o in '~

40

0.0

,,,o t-, e~

203

0.5

1.0

1.5

2.0

2.5

C ~ rcOncentratiOn (raM)

Fig. 3. Effect of phospholipid concentration on calpain autolysis. PS vesicles were prepared by bath sonication of a I mM PS dispersion in 0.193 mM Tris-HCl, 0.187 mM KCI (pH 7.5). Their effect on calpain autolysis was determined by incubating the vesicles in the presence of 4.5 /lg of calpain and either 0.0, 0.25, 0.50, 0.75, 1.5 or 2.5 mM Ca 2+. Autolysis was stopped after 1 min by addition of gel sample buffer. SDS

E 5O

gel electrophoresis was then used to determine the extent of autolysis. Final PS concentrations of 0 mM (O), 0.16 mM ( • ), 0.33 mM ( v ) and 0.5 mM ([]) were tested. Error bars represent the S.E.M. of 4 points.

20 S I

O

Ov 0.0

/

ntral

I 1.0

0.5

I 1.5

I 2.0

I 2.5

Ca2+ c o n c e n t r a t i o n (mM) I00

~e-

90

----

70

~

6o

-

(b)

/l/

© PC

~ /.L//

'~ PP. ,

0 0.0

oo

~ 0.5 10 1.5 2.0 Ca2+ : o n c e n t r a t i o n (rnM)

ol

2.5

Fig. 2. Effect of phospholipi,.t on m-calpain autolysis. Calpain was incubated at 30°C in the 0.0, 0.25, 0.50, 0.75, 1.0 or 2.5 mM Ca 2+ for 1 min in the presence 0.5 mM concentrations of various phospholipids. The extent of autolysis was determined as described in the results section. (a) Percentage conversion of the small subunit in the presence of PG (©), PS (O), P1 ( v ) , PA ( • ) and no ~hospholipid control ( • ) . (b) Percentage conversion of the small subunit PC (O), sphingomyelin (O), PE ( ~7) and no phospholipid control ( • ) . Error bars represent the S.E.M. of 8 points.

( w / v ) SDS, 5% ( v / v ) saturated Bromophenol blue and 2% monothiolglycerol. Gels were stained with Coomassie brilliant blue and scanned on a LBK 2202 Ultroscan laser densitometer. Scintillation counting. Radioactivity was counted with a LBK Wallac 1215 Rackbeta counter, using a 2,5-diphenyloxalzole/triton/xylene (18 mg:l litre:3 litre) scintillant. '4C was counted with a window from 110 to 176, tritium with a window from 8 to 80. Purification ofpeptides. Peptides labelled by N-acetylation with ['4C]acetic anhydride were obtained from Alta Bioscience (School of Biochemistry, University of Birmingham, UK). Three peptides were made: GIoTAMRILGG, GGTAMRILGG and G,0TAMQILGG. The first two peptides (Gt0TAMRILGG and GGTAMRILGG) were dissolved in 100 txl trifluoroacetic acid (100%) and then diluted to 1 ml with 50 mM HCI. Low molecular weight contaminants were removed by purification on a Sephadex G-10 gel filtration column (58 × 1 cm) run in 50 mM HCI. The third peptide was purified in a similar way except it was dissolved in 50 mM Hepes (pH 7.5) and the G-10 column was run in 50 mM Hepes (pH 7.5), in order to prevent any acid degradation of the glutamine. The concentration and composition of the purified peptide pools were determined by amino-acid analysis. This was performed as follows. The peptide was hydrolysed in concentrated HCI for 22 h. It was then applied to an ABI 420A derivatiser/analyser which utilises pre-column derivatisation with phenylisothiocyanate to form phenylthiocarbamyl

204

J.S. C. Arthur, C. Crawford / Biochimica et Biophysica Acta 1293 (1996)201-206

amino acids. These were then automatically sampled to a narrow-bore HPLC system for analysis. A sample was also counted for ~4C counts and the specific activity defined as the number of counts per min per nmole of peptide.

peptides (GIoTAMRILGG and GGTAMRILGG) corresponding to different parts of this region, was investigated. A third peptide (G~0TAMQILGG) was also tested. In this

100

3. Results

2.0

3.1. The effect of phospholipids on m-calpain autolysis The effect of PI, PS, PE, PC, PA, PG and sphingomyelin on m-calpain autolysis were investigated by incubating 4.5 /zg of calpain for 1 min at 30°C in the presence of 0.5 mM phospholipid and Ca 2 + concentrations from 0 to 2.5 mM. The final volume was 30 /xl with a buffer concentration of 100 mM Tris-HCl (pH 7.5), 93.3 mM KCI and 6.7 mM NaCI. Autolysis was stopped by the addition of 15 /zl of SDS gel sample buffer. The extent of autolysis was determined by SDS-PAGE followed by Coomassie brilliant blue staining (Fig. 1). A laser densitometer was used to scan the gels and determine the proportions of the intact small subunit (30 kDa) and its autolytic product (18 kDa). All the phospholipids tested, except PE, caused a significant reduction in the Ca 2+ concentration required for calpain autolysis (Fig. 2). PE was found to have a much reduced effect on caipain autolysis. The effect of varying the phospholipid concentration was also examined. The same assay as above was used but the phospholipid concentration was varied while keeping the Ca 2 + concentration constant (Fig. 3). The data shows that at 0.5 mM phospholipid, small variations in lipid concentration do not significantly affect calpain autolysis.

3.2. Phospholipid binding peptides The most probable phospholipid binding site in calpain is the small subunit N-terminus. Previous work suggested that a region, with the sequence GITTAMRILGG, between 38 and 62 amino acids from the N-terminus was responsible [13]. The binding to PI vesicles of two synthetic

25

(a)

80 E

1.5

C

E

60

C

{3_

O.

~

1.0 ~.

4o

(3.

cl

0.5

20

O0

10

20

30

40

50

60

Elution volume (ml) 140

2.5

(b) 120

2.0 100 0

E 3

1.5

8O

E C

C~

o

6o

~-

'to

~.o ~_ Ck

0.5 20

0 -

0

__

,I

10

0.0 20

30

40

50

60

Fl,,tlc~n v~l,Jrn~, ( m l ~

(c)

7 6

Fig. 4. Coelution of peptides and phosphatidyl inositol vesicles. Peptide (5 to 20 nmoles in 0.2 ml of buffer) was added to tritium labelled phosphatidyl inositol vesicles (0.5 ml of 1 mM phospholipid) and incubated for 45 min at room temperature. The mixture was then run on a Sepharose 4B column. Fractions (0.95 ml) were collected and 300 p,l aliquots were counted for ~4C and tritium counts. The molar amounts of phosphatidyl inositol (C)) and peptide ( O ) were the calculated using specific activities. GIoTAMQILGG (Fig. 4c) was found to have a much lower specific activity than the other two peptides. This made co-elution experiments with tritium labelled vesicles impractical. To overcome this unlabelled vesicles were used. This allowed a wider 14C window to be used for counting, thus increasing the specific activity of the peptide. In addition the entire fraction was counted. (a) Peptide GGTAMRILGG, (b) peptide GIoTAMR/LGG and (c) peptide GIoTAMQILGG.

oJ

I

¢v C

O.

2

1 0 0

10

20

30

40

Elutlon volume (m])

50

60

J.S. C. Arthur, C. Crawford / Biochimica et Biophysica Acta 1293 (1996) 201-206

peptide the only charged residue in the region (Arg) was replaced by an uncharged one (Gln); glutamine was chosen as a replacement for arginine as it is the most similar uncharged amino acid. Gel filtration was used to examine the interaction of three peptides to phospholipid. Coelution of J4C labelled peptide w!th tritium labelled vesicles on a Sepharose 4B column would show the peptide had bound. Peptide was added to I).5 ml of I mM tritium labelled PI vesicles and incubated for 45 rain at room temperature. The mixture was then ran on a Sepharose 4B column (58 × 1 cm). The buffer ased was 50 mM Tris-HCl, 100 mM NaCI, 1 mM EDTA, 0.01% ( w / v ) sodium azide (pH 7.5). The results are shown in Fig. 4. Control runs were also perlbrmed to check the elution points of both the peptides and vesicles on their own. This showed that the three peptides were eluted as expected at the end of the column run and did not aggregate or bind to the column. The main phospholipid peak occurred in the void volume with a smaller peak eluting after the void volume. Neither of these peaks coincided with the peptides peaks. Of the three peptides only one (GIoTAMR1LGG) was found to bind to the PI vesicles. A ratio of approximately 1 peptide molecule to 300 phosphatidyl inositol molecules was observed. Peptides GGTAMRILGG and Gt0TAMQILGG showed no interaction wi':h the vesicles.

4. Discussion

All the phospholipids tested caused a reduction in the Ca 2÷ concentration required for autolysis (Fig. 2). PE was the least effective phospholipid, with a reduction in the Ca 2- requirement for 50% autolysis from 1.4 mM to 1.2 mM. The remaining phospholipids gave values between 0.48 mM (PG) and 0.76 mM (PC) for the Ca 2+ concentration required for 50% aatolysis. Previous conflicting reports [9-11] have suggested that only certain phospholipids are able to interact with calpain. More recently it has been proposed that only acidic phospholipids have the effect of enhancing caipain autolysis [15]. The data presented here shows that both acidic and neutral phospholipids can interact with calpain, although acidic phospholipids were found to be more effective. If the interaction between phospholipid artd calpain depends solely on the binding of calpain to the phospholipid head group, calpain would require a head-group binding site. To explain the results in Fig. 2 this binding site would best accommodate glycerol, serine and inositol, however, choline and hydrogen groups would also bind. This is unlikely as the binding site in calpain would require considerable flexibility to bind such diverse structures. Comparison of the effects of sphingomyelin and phosphatidyl choline provides further evidence that head groups are not the only important factor in this system. Sphingomyelin and phosphatidyl choline have different effects on calpain but both have choline head groups. In addition it has previously been shown that

205

all of the phospholipid molecule is required for an effect on calpain, the head group alone is not sufficient [9]. Given the lack of head-group specificity observed in Fig. 2 it is likely that calpain interacts by binding with the diacylglycerol or phosphate regions in the phospholipid bilayer. If this is the case, it is necessary to explain why different phospholipids produce quantitatively different effects. There are many complex factors which may contribute to the differences in behaviour of the different phospholipids. These arise in part from the influence of the head group on the physical nature of the vesicles formed. For example; the ratio of the multilamellar to unilamellar vesicles present after sonication may vary, resulting in the available surface area of the phospholipid, variations in the packing of the phospholipid molecules and fluidity of the membrane and variation in the effects of Ca 2+ on the vesicles. The small effect seen with PE is readily explained on this basis. PE has a strong tendency to form hexagonal HII structures in aqueous solution [16,17] and so does not mimic a membrane bilayer. The observation of Pontremoli et al. [10] that phospholipids with completely saturated fatty acids are not able to activate calpain can be explained in a similar way. The two saturated lipids used have transition temperatures above the temperature at which the experiments were performed. Thus any bilayers present would be in the rigid gel state rather than the more normal fluid liquid-crystal state [17]. The results can therefore be explained as a lack of suitable vesicles being present to activate the calpain. In conclusion, the data presented here shows that all the phospholipids tested can decrease the Ca 2+ concentration required for autolysis. In addition they suggest that the variation in the effectiveness of the different phospholipids is due to differences in their ability to form bilayers suitable for calpain interaction rather than the existence of a specific head-group binding site in calpain. Previous work has shown that the most probable phospholipid binding site in calpain is the N-terminus of the small subunit and that a region between 39 and 62 amino acids (GIvTAMRILGG) from the N-terminus is required [12,13]. The peptide work presented here is the first direct demonstration that a sequence in calpain can bind to phospholipid. The results shown in Fig. 4 demonstrate that the TAMRIL sequence alone is not sufficient to bind to phospholipid; however, in conjunction with a polyglycine sequence, binding does occur. Polyglycine is unlikely to have a very defined secondary structure and it is hard to envisage it forming a specific binding site. It may however be able to interact with the inside of the bilayer, or affect the structure or orientation of the adjacent TAMRIL sequence. The failure of the GIoTAMQILGG peptide to bind shows a role for the arginine residue. This is the only charged residue in the region and it is possible that the arginine interacts with the negatively charged phosphate group in the phospholipid. Arginine has been shown to have a role in binding to phospholipids in other systems

206

J.S.C. Arthur, C. C r a ~ f o r d / Biochimica et Biophysica Acta 1293 (1996) 201-206

distinct lbrm calpain, although this often involves groups of arginines rather than single residues [18]. The work presented in this paper supports the model proposed by Suzuki [2] for calpain activation in which calpain binds to phospholipid prior to autolysis. Presumably during calpain activation binding of Ca 2÷ to high affinity Ca 2+ sites causes a confbrmational change which allows the G~7TAMRILGG sequence, possibly in conjunction with other sequences in calpain, to interact with phospholipid bilayers. This interaction then causes another conlbrmational change which either promotes autolysis directly or causes an increase in the affinity of the remaining sites for Ca -,+ allowing them to bind Ca 2÷ and initiate autolysis. The results show that a phospholipid bilayer can lower the Ca"- levels required for the autolysis of calpain, a factor which may be important during calpain action in vivo. However the reduction in Ca -,+ requirement observed here is not enough on its own to explain calpain activation in vivo, suggesting that other factors are involved

Acknowledgements We thank Antony Willis lor performing the amino-acid analysis and Nick Brown for help with the preparation of calpain. This work was supported by a Medical Research Council Project Grant held by C.C. and an Agriculture and Food Research Council Co-operative Award Research Studentship held by S.A.

References [1] Murachi, T. (1989) Biochem. Int. 18, 263-294. [2] Suzuki, K., Ohno, S., Emori, Y. and Kawasaki, H. (1987) Prog. Clin. Biochem. Med. V 43-65. [3] Mellgrcn, R.L. (1987) FASEB J. 1, 110-115. [4] Suzuki, K. (1987)Trends Biochem. Sci. 12, 103-105. [5] Drayton. W.R. (1982) Biochim. Biophys. Acta 709, 166-72. [6] Suzuki, K., Tsuji, S., Ishiura, S., Kubota, S., Kimura, Y. and Imahori. K. (1981)J. Biochem. 90, 1787-1793. [7] Inomata, M., Hayashi, M, Nakamura, M.. Imahori, K. and Kawashima. S. (1985) J. Biochem. 98, 407-416. [8] Crawford, C.. Willis, A.C. and Gagon. J. (1987) Biochem. J. 248, 579-588. [9] Ccmlican, S.A. and Hathaway, D.R. (1984) J. Biol. Chem. 259, 11627-11630. [10] Pontremoli, S., Melloni, E., Sparatore, B.. Salamino, F., Michetti, M.. Sacco. O. and Horecker, B.L. (1985) Biochem. Biophys. Res. Commun. 129, 389-395. [11] Chakrabarti, A. K.. Dasgupta. S., Banik, N. 1,. and Hogan, E. (1990) Biochim. Biophys. Acta. 1038, 195-198. [12] Imajoh, S., Kawasaki, H. and Suzuki, K., (1986) J. Biochem. 99. 1281-1284. [13] Crawford, C., Brown, N.R. and Willis, A.C. (1990) Biochem. J. 265, 575-579. [14] Laemmli, U.K. (1970) Nature (London)227, 680-685. [15] Saido, T.C., Shibata, M., Takenawa, T., Murofushi, H. and Suzuki, K. (1992) J. Biol. Chem., 267. 24585-24590. [16] New, R.R.C. (1990) Liposomes. A Practical Approach. IRL press, Oxtord. [17] De Gier, J., Mandersloot,J.G., Van Deenen,L.LM. (1968) Biochim. Biophys. Acta. 150. 666-675. [18] Mosior, M. and Malaughlin,S. (1992) Biochemistry31, 1767-1773.