Biochirnica em3jo~hysica Acta, 1042 (1990) 188-192
188
Elsevier
BBABIQ 53298
Apoli~oprotei~ C-1 inhibits the hydrolysis by phospholipas~ A2 of ~hospholipi~s in liposomes and cell m~rnb~an~s Jutta Poensgen (Received 8 May 1989) (Revised manuscript received 18 August 1989)
Key words: ~hosphoIi~ase A,; Apo~poprotein C-l; ~~ospholipa~ inhibitor; Liposome; Cell membrane; (Hurn~R leukemia (~11)
A small ~ly~ptide isolated from human serum inhibits the action of phospholip~ A, on dip~rnitoylg~y~~~ ~os~~~line vesicles. Sequence analysis revealed the protein to be a~ii~protein C-l, a major com~nent of very light-densi~ lipoprotein. The inhibiting efficiency is increased by one order of magnitude after 10 min preincubation of the protein with the substrates but not the enzyme. It also depends on the con~en~tion of the ~ospholipid. X2, is about 0.5 PM at 0.2 mM DPPC and 1 ,c~M at 1 mM DPFC. A~~i~protein C-l is afso inhibitor in a more physiologic system: in broken human leukemia cells (HI.&@ cefls) it inhibits the release by e~genous p~sphoIi~s of ~achidooic acid from memb~e phospholipi~. The effective ~onceo~ations co~es~nd to those fowd in the serum. It is concluded that apolipoprotein C-l and similar phosp~lipid-binding proteins may act as phospholip~e inhibitors by blocking the access to the substrate.
A role for PLA, has been suggested in several cellular events such as exoeytosis and mitogenesis (1,2] as well as in pathoIo~ca1 situations - e.g., septic shock or inflammation [3,4], where PLA, is thou~t to control the level of free arachidonic acid and thereby of prosta~andin and leukotriene synthesis. But in spite of the key-role of PLA,, it is not realIy underst~ bow its activity is controlled. Several proteins have been found to i~bit the action of PLA, [5-8]. The best known is lipocortin, a 36 kDa protein which has been claimed to mediate the ~t~nfla~ato~ effect of gi~c~urtic~ids. It was assumed that lipocortin binds to PLA, and forms an eirzymatically inactive complex [Q] just like those found with proteinases and their e~doge~ous inhibitors. But recently, two forms of lipocortin have been identified as caipactins [lo] which are known in the presence of Ca2+ to bind to phospholipids and actin, and consequently lipocortin was shown to inhibit PLA~~acti~~y by sequestration of the substrate 1111.
Abbreviations: PLA,, phospho~p~e A,; apo C-1, a~l~p~~rotein C-l ; DPPC, dipaI~toyI~y~rop~~spb~hoIine; HPLC, high-pressure liquid chrumato~apby; BSA, bovine serum aIbu~n. Corr~~onden~:
F.R.G. ~~-~7~/~/$~3.50
J. Poensgen, FlandriscIx
StraI?e 34, D-51
Aache%
Dialyzed human serum exhibits a considerable PLAY-in~biting capacity which cannot be accounted for by lipo~ortin or albumin (~bu~n interferes with some PLA,-assays [12]). Very likely the serum contains other inhibitory proteins as well. I have isolated one of these factors in order to see whether it acts on the substrate or on the enzyme and if it is related to lipocortin. A preli~na~ report was given at the 4th PhosphoIipase Symposium, Universitat Ulm, June 10th to 11th 1988. Materials and Meth~s
PLA, from hog pancreas, RPM1 1640 medium and fetal calf serum were obtained from Boeh~~ger (Mannheim, F.R.G.), [l-‘4C]aracbidonie acid was purchased from Amersham Buchler (Brau~schweig, F.R.G.), af ~u~~~~~~~-l-14C]dipal~toyigly~erophospb~ho~iue was bought from New England Nuclear (Dreieichenh~n, F.R.G.), the unlabeIled compound and DMSO for cell cultures were from Sigma (Taufkirchen, F.R.G.). Nucleosif 5 C,, was a product of ~acherey und Nagel (D&en, F.R.G.), Bio-Gel P-30 was obtained fro,” Bio-Rad (~~nchen, F.R.G.), and silica-gel (63-100 A) was produced by ICN Bi~he~c~s (~s~hwege, F&G.}. For PLA~*i~bition assay with soluble PLA,, dpm per experi[i4C!]DPPC (20 or 100 nmol and 15
0 1990 Elsevier Science Pubfishers B.V. ~~iornedic~ Division)
189 mental point) was dried under N,, taken up in 0.05 M Tris-HC1/4 mM CaCl, (pH 8.0), to a concentration of 0.4 or 2 mM, respectively, and sonicated for 30 min above 37 * C. 50 ~1 liposomes were added to 50 $0.05 M Tris-HCI (pH 8.0), containing 5-10 pmol PLA, or 10 pmol PLA, and inhibitor. After 5 min incubation at 37” C, 10 ~1 1 M HCl was added and the mixtures transferred to small plastic columns with 1.5 ml silica-gel (63-100 A). Free [‘4C]palmitic acid was eluted by four washes with 1.25 ml hexane/ dioxane/ acetic acid (70/30/L v/v), ]13]. For the PLA, assay with HL 60-homogenates, HL 60 cells were grown in RPM1 1640 supplemented with 10% fetal calf serum, 2 mM glutamine, 1 mM sodium pyruvate, 1 mM nonessential amino acids to lo6 cells/ml as described in Ref. 14. 50 U/ml penicillin and 50 pg/ml streptomycin were added routinely. Differentiation was induced by cultivating 3. lo5 cells/ml for 5 days in the medium described above in the presence of 1.25% (v/v) DMSO. During the last 18 h the cells were labelled with 1 ,nCi [1-‘4C]arachidonic acid per 100 ml culture medium. The labelled cells were pelleted, washed once with Hepes/saline/BSA (25 mM Hepes, 125 mM NaCl, 0.7 mM MgSO,, 0.4 mM EGTA, 10 mM glucose + 2 mg/ml fatty-acid-free BSA (pH 7.4), and resuspended at 2 - 10’ cells/ml in Hepes/saline/BSA. They were broken by ultrasonication (3 - 15 s at 0” C in a Branson sonifier B 15 equipped with microtip) and used immediately: 0.2 ml cell homogenate was preincubated at 37 o C with and without apo C-l dissolved in DMSO (final concentration 2% DMSO - controls received 2% DMSO only). Arachidonic acid release was started by addition of 0.6 pmol CaCl, in 0.1 ml Hepes/ saline/BSA. At 15 min intervals the samples were quenched and extracted with 1.5 ml CHCl,/CH,OH/ CH,COOH (2 : 1: 0.05, v/v). The organic phase was dried under N,, the residue was dissolved in 25 ,ul CH,Cl/CH,OH (2 : 1, v/v) and applied to silica-gel TLC plates. Phospholipids and arachidonic acid were separated with CHCl,/CH,OH/ H,O (69 : 27 : 4, v/v) and identified by comparison to standards. The dist~bution of the label was determined by scanning with a Berthold Automatic TLC Linear Analyzer LB 2842. PLA, activity is given as X radioactivity in arachidonic acid per x min. Protein concentration was determined by the Bradford method 1151or, in the case of purified apo C-l, by the amino acid content of its acid hydrolysate: The pure protein was hydrolyzed in 6 M HCl containing 0.1% thioglycolic acid at 110 o C for 20 h. Amino acid composition was determined with the amino acid analyzer LC 6001 from Biotronik (Frankfurt, FRG) using a 0.6 cm X 21 cm column of DC 6A resin. Identification of the amino acids was carried out by conventional ninhydrin analysis combined with an elution pro-
gram controlling the flow with time of the 5 buffers employed [16]. The amino acid analysis given is not corrected for partial destruction (e.g., serine or threonine) or for incomplete release during hydrolysis (e.g., valine or isoleucine). Sequence analysis of the purified protein was performed essentially as described by Edman [17] using an automatic liquid-phase sequencer from Beckman, Model 890M. Phenylthiohydantoin derivatives were analyzed by isocratic HPLC on a superspher CH, column (Merck, Darmstadt) according to Ref. 18. For prepara~on of apo C-l, 1 1 of human serum was cooled to 0 o C and brought to 65% saturation with solid (NH,),SO,. Following centrifugation, the supernatant was ultrafiltered and dialyzed in a stirred cell (Amicon) equipped with a PM 10 membrane until the volume was reduced to 100 ml and the (NH4)*S04 concentration was about 10% of the original. 33 ml each were mixed with 15 ml acetic acid and fractionated on a Bio-Gel P-30 column (4.6 x 80 cm) equilibrated with 30% acetic acid. Inhibitory fractions eluted after the exclusion peak were refractionated twice, first in a medium-sized (3 x 50 cm) and finally in a small (1.6 x 70 cm) column, both filled with Bio-Gel P-30 in 30% acetic acid. The resulting small polypeptides were further separated by reversed-phase HPLC on Nucleosil 5 C,, using the following gradient of 0.05% tri~uoroacetic acid in water (A) and 0.025% trifluoroacetic acid in acetonitrile (B): O-41% B in A in 75 min, 41% B in A for 20 min, 41-60% B in A in 5 min, 60% B in A for 10 min. Flow rate: 1.6 ml/mm, T = 60 o C, detection at 214 nm. The isolated protein was kept in acetonitrile/acetic acid/ water (5 : 3 : 2 v/v) at - 20 o C until used for the experiments described. Then the solvent was evaporated, the residual material was dissolved in DMSO and diluted with 0.1 M Tris-HCl (pH 9.0), yielding a final pH of 8.0. Results
Fractionation of human serum by various chromatographic procedures showed the major part of the inhibitory activity (as assayed by its effect on the hydrolysis by pancreatic PLA, of sonicated DPPC-vesicles) to comigrate with the main protein peaks. However, by gel chromatography about 10% of this activity were eluted with the small molecular weight fraction and even appeared beyond vi. As I wanted to avoid the use of detergents and had ascertained that the inhibitory activity did not decline in 30% acetic acid, I aimed for a small protein and fractioned the serum in 30% acetic acid so as to reduce interactions with the gel matrix. Subsequent to the removal of the majority of the proteins with ammonium sulfate, the remainder was concentrated and fractionated by repeated gel chromatography on B&Gel P-30 in 30% acetic acid, which
190 Isolated
:
T
P
0
V
5 S
S
A
L
D
‘::
L
K
E
F
15 G
N
T
L
E
:
APO C-l
:
T
P
D
V
5 S
5
A
L
D
‘K”
L
K
E
F
‘2
N
T
L
E
200
protein
:
U
A
RELISRI
30 K
Q
s
E
L
3:
A
K
M
R
40 E
K
A
R
30 K
Q
S
E
L
3:
A
K
M
R
“Eo
protein
25 lsoloted
ApOC-1:
E
*?
I
S
R
I
Fig. 1. The sequence of the first 40 amino acid residues of the isolated protein and of apo C-l.
eluent reduces the exclusion limit to 10 kDa. The resulting group of small polypeptides was further purified by reversed-phase HPLC yielding an in~bito~ protein of 6.5 kDa (determined by SDS gel electrophoresis). About 1 mg was obtained from 1 liter of serum. Determination of the amino acid composition revealed that the isolated polypeptide contained neither cysteine nor tyrosine. The sequence of the first 40 amino acids was determined and showed the purified protein to have two NH,-termini - about half the material was shortened at the amino end by two amino acid residues. Comparison to known sequences revealed identity with apo C-l, a major component of very light-density lipoprotein which consists of 57 amino acid residues [19]. Judging from the amino acid composition (given in Table I) and the sequence of 40 out of 57 amino acids (shown in Fig. 1) I took the isolated protein to be apo C-l. Apo (Z-1 is known to bind to phosphatidylcho~ne bilayers [20], and I expected it to interact with the substrate, not the enzyme. For confirmation I preincubated apo C-l either with the DPPC-vesicles or with
PLA, before performing the assay. It turned out that 10 min preincubation with the substrate decreased the concentration of apo C-l needed for half-mammal inhibition by one order of magnitude (Fig. 2, the inset gives an example of the inhibition progress with increasing preincubation times). Preincubation with the enzyme has no such effect. The effect of preincubation is not due to retarded binding of apo C-l to the DPPCvesicles: The initial time-course of the hydrolysis catalyzed by PLA, in the presence or absence of apo C-l is straight and shows that apo C-l exerts its inhibitory action immediately (Fig. 3) (see Discussion).
TABLE I Amino acid analysis of the prorated protein (a) and amino acid composition
ofup0 C-I
(b)
~ 01
1
nM ape C-l
Amino acid
(a)
(b)
Asx Thr Ser Glx
5.3 2.1 7.2 11.2 1.6 3.3 1.5 1.0 2.3 6.0 0.1 3.6 0.6 8.5 3.4 0.8 n.d. a
5 3 7 9 1 3 2 1 3 6
GIY Ala VaI Met Be Leu Tyr Phe His Lys Arg Pro Trp a Not determined,
3 9 3 1 1
but the protein absorbs at 280 nm.
b PM opoC-I
Fig. 2. The inhibitory effect of apo C-l after 10 min preincubation with the substrate. (a) 1 mM DPPC in 0.05 M Tris. HC1/4 mM CaCI (pH 8.0) was incubated for 10 min at 37’ C with and without apo C-l. The assay was started by adding 10 pmol PLA, (final volume: 100 ~1) and stopped with 10 pl 1 M HCI. Free [‘4CIpahnitic acid was separated from DPPC on silica-gel coIumns and counted in a j3-counter. The bars represent standard deviations obtained with triplicate determinations. The data are representative for two experiments carried out independently. (b) PLA, was assayed without preincubation with 1 mM DPPC and apo C-l (IC,, taken from Fig. 4B). Inset: Effect of preincubation time on the inhibition by apo C-l. 2 mM DPPC in 0.05 M Tris-HC1/4 mM CaCI, (pH 8.0) was kept for 20 mitt at 37OC. At f = 0, 15, 17.5 and 20 min. duplicates of 50 pl aliquots received 30 pf = 12.6 pmol apo C-l in Tris-HCI (0.05 M, pH 8.0). 2 controls received buffer only. At t = 20 mitt, 20 pmol PLA, in 20 ~1 0.05 M Tris-HCI, pH 8.0, were added. The incubation was continued for 2 min and stopped with 10 pI 1 M HCI. (A) PLA, only. (B) PLA, +apo C-l without preincubation. (C) PLA, +apo C-l, preincubated for 2.5 min. (D) PLA, +apo C-l, preincubated for 5 min. (E) PLA, + apo C-l, preincubated for 20 min.
191
10-
Fig. 3. Time-course of the hydrolysis of DPPC-vesicles by PLA, in the presence and absence of apo C-l. 0.2 mM DPPC was incubated at 37OC with 0.1 pM PLA, +f0.55 pM apo C-1. Every 2 min, duplicates of 0.1 ml aliquots were quenched with 10 pl 1 M HCl and free pahnitic acid was obtained by elution from silicagel columns. Upper curve ( X ); absence of apo C-l, lower curve (+ ); presence of apo C-l. ,
The interaction of apo C-l with the substrate implies that the inhibitory effect depends on the concentration of the substrate as well as on that of the inhibitor. I compared the concentration of apo C-l needed for 50% inhibition with 0.2 mM and 1 mM substrate. With 0.2 mM DPPC, the hydrolysis is 50% reduced by about 0.5 PM apo C-l, whereas with 1 mM DPPC, 50% inhibition is achieved by 1 FM apo C-l (Fig. 4). Our assay uses an artificial substrate and a secreted PLA,, and the inhibition by apo C-l might well be limited to that system. If any physiological significance is to be ascribed to the inhibition of PLA, activity by apo C-l, an inhibitory effect should also be demonstrable with physiological substrates such as cell membranes and with an endogenous PLA,. We used broken HL 60 cells which had previously been induced to differentiate by DMSO and were labelled with [l“C]arachidonic acid. These homogenates were incubated with 2 mM CaCl, and the time course of arachidonic acid release in the presence and absence of apo C-l was determined. In four experiments I found 30-40% inhibition of arachidonic acid release with 13 to 130 nM apo C-l (Fig. 5). In contrast to the experiments with DPPC-vesicles (cf. Fig. 4}, inhibition was never complete and did not exceed 50% even with 1.3 FM apo C-l (data not shown).
1%; 0
1
2
3
4
j.iM ape C-l
Fig, 4. The i~bition by apo C-l of the hydrolysis of DPPC at two different substrate concentrations. PLA, was incubated for 5 mm at 37 o C with apo C-l and DPPC. A: 0.2 mM DPPC, 0.1 PM PLA,. B: 1 mM DPPC, 0.1 PM PLA, (x) or 0.2 pM PLA, (0). The data are taken from three experiments carried out independently. Each value is the mean of duplicate determinations (error bars were omitted for reasons of clarity).
This is the case with mepacrine, an often used ‘PLAN-~bitor’ and other cationic arnp~p~~c drugs [22,23]. By this mechanism several therapeutic agents 15 t
Discussion
Due to the fact that a soluble enzyme acts on a water-insoluble substrate, there are many pitfalls in the assessment of phospholipase inhibitors, and often the inhibitory effect is predetermined by the choice of the assay system. The activity of PLA, is markedly influenced by the substrate matrix 1211 and anything interfering with the structure of the substrate may inhibit (or even stimulate) PLA, action.
Fig. 5. The inhibition by apo C-l of arachidonic acid release in broken HL 60 calls. 0.2 ml homogenate of (t4C]arachidonic acidlabelled, differentiated HL 60 cells were preincubated for 10 min at 37OC with and without 20 nM apo C-l. After addition of 0.1 ml buffer containing 0.6 pmol CaCl,, the appearance of [t4C$rachidonic acid was followed by extracting the lipids and separating phospholipids and arachidonic acid by TLC in CHCI,/CH,OH/HaO (69/27/4 v/v). The distribution of the “C-label was determined in a TLC-scanner. Upper curve ( X) - control; lower curve ( +) - sample with 13 nM apo C-l.
192 produce a phospholipidosis [24], which proves that interference with the substrate is also effective in biological systems. Sequestration and thus depletion of the substrate is still another mechanism by which lipocortin and other phospholipid-bin~ng proteins may modulate the activity of PLA, [II]. In a recent publication, Conricode and Ochs also described PLAY-in~bition through substrate depletion by proteins 1251. Apo C-l is a protein which complexes phosphatidylcholine and interferes with the binding of PLA, to the substrate, The time-course of the reaction shows inhibition and therefore binding of apo C-l to be immediate, but nonetheless 10 min preincubation with the substrate increases the inhibitory capacity by one order of magnitude. I think that this phenomenon is an artefact caused by apo C-l-induced aggregation of the DPPC-vesicles, since apo C-l-complexed phosphatidylcholine vesicles have been shown to form aggregates [20]. It tallies with this assumption that the progress of i~bition with increasing preincubation times varies with different substrate preparations. Elucidation of these presumably complicated events would require a separate detailed study. Without preincubation 50% inhibition at 0.2 mM DPPC is achieved by about 500 nM apo C-l, and at 1 mM substrate ICY,, is 1 PM apo C-l. On the average, small unilamellar DPPC-vesicles consist of approx. 2500 molecules, two-thirds of which are in the outer leaflet [26]. Thus, 3-6 molecules of apo C-l suffice to shield half the surface of a liposome. Neither in~o~oration of an ‘inhibitor’ into the substrate membrane nor sequestration of the substrate achieve a specific inhibition of PLA,. Lipocortin has been shown to inhibit phospholipase D, too [27]. AS regards different substrates, the inhibition might show a certain selectivity. Probably the various phospholipid-binding proteins have different affinities to the different phospholipid components of cell membranes. Apo~poprote~s A-l and C-l bind to phosphatidylcholine which in erythrocytes is located predominantly in the outer leaflet of the plasma membrane [28]. Thus, these are well suited to shield the membrane of the red blood cell against phospholipases released into the plasma. The experiments with HL SO-homogenates prove that apo C-l in lo-100 nM ~n~ntrat~ons is well able to protect biological membranes against phospholipase attack, although complete inhibition was not attained. Possibly, certain regions of the different cell membranes do not bind apo C-l. Generally, 2 liter of serum yielded 100-150 nmol apo C-l. So the normal concentration at least equals the range found inhibitory in the HL 60 assay. I do not assume that apo~poproteins are the only extracellular inhibitors of phospholipase activity. Actually, all phospholipid-binding proteins should have similar membrane-shielding effects.
I thank Monika Nellessen for skilful assistance and Drs. G. Steffen and B. Wolf (Griinenthal GmbH) for determining the amino acid composition of the purified protein and its sequence, respectively. My thanks are also due to Dr. Habenicht (Universitat Heidelberg) for generously supplying me with HL 60 cells. I am very grateful to Drs. U. Gehring (Universitat Heidelberg) and B. Hamprecht (Universitat Tubingen) for helpful discussions during the preparation of the manuscript.
References 1 2 3 4 5 6 7
8 9 10 11 12 13 14 15 16 17
18 19
20
21 22 23 24 25 26 27 28
Creutz, C.E. (1981) J. Cell. Biol. 91, 247-256. Shier, W.T. (1980) Proc. Natl. Acad. Sci. USA 77, 137-141. Vadas, P. (1984) J. Lab. Clin. Med. 104, 873-881. Vadas, P. and Pruzanski, W. (1984) Adv. Inflam. Res. 7, 51-59. BlackweB, G.J., Carnuccio, R., Di Rosa, M., Flower, R.J., Parente, L. and Per&a, P. (1980) Nature 287, 147-149. Hirata, F., Schiffmann, E., Ve~atasubram~an, K., Salomon, D. and Axelrod, J. (1980} Proc. Natl. Acad. Sci. USA 77, 2533-2536. Clok, J.F., Colard, O., Rothhut, B. and Russo-Marie, F. (1983) Br. J. Pharmacol. 79, 313-321. Haigler, H.T., Schlaepfer, D.D. and Burgess, W.H. (1987) J. Biol. Chem. 262,6921-6930. Hirata, F. (1981) J. Biol. Chem. 256, 7730-7733. Gienney, J.R., Jr., Tack, B. and Powelf, M.A. (1987) J. Cell Biol. 104, 503-511. Davidson, F., Dennis, E.A., Powell, M. and Glenney, J.R., Jr. (1987) J. Biol. Chem. 262, 1698-1705. Vadas, P., Stefanski, E. and Pruzanski, W. (1986) Inflammation 10, 183-193. Cosentino, M.J. and Ellis, L.C. (1981) Prostaglandins 22, 309-322. Billah, M.M., Eckel, S., Myers, R.F. and Siegel, M.J. (1986) J. Biol. Chem. 261.5824-5831. Bradford, M.M. (1976) Anal. Biochem. 72, 248-254. Biotronik (1981) Appliiationsreport ASA 981/2, Munchen. Edman, P. and Henschen, A. (1975) in Molecutar Biology, Biochemistry and Biophysics (Needleman, S.B., ed.), 8, 232-279, Springer, Berlin. Lottspeich, F. (1980) Hoppe-Seyler’s 2. Physiol. Chem. 361, 1829-1834. Jackson, R.L., Sparrow, J.T., Baker, H.N., Morrisett, J.D., Taunton, O.D. and Gotto, A.M., Jr. (1974) J. Biol. Chem. 249, 530885313. Jackson, R.L., Morrisett, J.D., Sparrow, J.T., Segrest, J.P., PownaB, H.J., Smith, L.C., Hoff, H.F. and Gotto, A.M., Jr. (1974) J. Biol. Chem. 249, 5314-5320. Jain, M.K., Streb, M., Rogers, J. and de Haas, G.H. (1984) B&hem. Pharmacol. 33, 2541-2551. Vargaftig, B.B. (1977) J. Pharm. Pharmac. 29, 222-228. Grlbner, R. (1987) B&hem, Pharmacol. 36, 1063-1067. Hostetler, K.J. (1984) Fed. Proc. 43, 2582-2585. Comicode, K.M. and Ochs, R.S. (1989) Biochim. Biophys. Acta 1003, 36-43. Cornell, B.A., Middlehurst, V. and Separovic, F. (1980) Biocbim. Biophys. Acta 598, 405-410. Kobayashi, M., Bansal, VS., Singh, I. and Kaufer, J.N. (1988) FEBS Lett. 236, 380-382. Zwaal, R.F.A., Roelofson, B. and Colle, CM. (1973) Biochim. Biophys. Acta 300, 159-182.
188
Elsevier
BBABIQ 53298
Apoli~oprotei~ C-1 inhibits the hydrolysis by phospholipas~ A2 of ~hospholipi~s in liposomes and cell m~rnb~an~s Jutta Poensgen (Received 8 May 1989) (Revised manuscript received 18 August 1989)
Key words: ~hosphoIi~ase A,; Apo~poprotein C-l; ~~ospholipa~ inhibitor; Liposome; Cell membrane; (Hurn~R leukemia (~11)
A small ~ly~ptide isolated from human serum inhibits the action of phospholip~ A, on dip~rnitoylg~y~~~ ~os~~~line vesicles. Sequence analysis revealed the protein to be a~ii~protein C-l, a major com~nent of very light-densi~ lipoprotein. The inhibiting efficiency is increased by one order of magnitude after 10 min preincubation of the protein with the substrates but not the enzyme. It also depends on the con~en~tion of the ~ospholipid. X2, is about 0.5 PM at 0.2 mM DPPC and 1 ,c~M at 1 mM DPFC. A~~i~protein C-l is afso inhibitor in a more physiologic system: in broken human leukemia cells (HI.&@ cefls) it inhibits the release by e~genous p~sphoIi~s of ~achidooic acid from memb~e phospholipi~. The effective ~onceo~ations co~es~nd to those fowd in the serum. It is concluded that apolipoprotein C-l and similar phosp~lipid-binding proteins may act as phospholip~e inhibitors by blocking the access to the substrate.
A role for PLA, has been suggested in several cellular events such as exoeytosis and mitogenesis (1,2] as well as in pathoIo~ca1 situations - e.g., septic shock or inflammation [3,4], where PLA, is thou~t to control the level of free arachidonic acid and thereby of prosta~andin and leukotriene synthesis. But in spite of the key-role of PLA,, it is not realIy underst~ bow its activity is controlled. Several proteins have been found to i~bit the action of PLA, [5-8]. The best known is lipocortin, a 36 kDa protein which has been claimed to mediate the ~t~nfla~ato~ effect of gi~c~urtic~ids. It was assumed that lipocortin binds to PLA, and forms an eirzymatically inactive complex [Q] just like those found with proteinases and their e~doge~ous inhibitors. But recently, two forms of lipocortin have been identified as caipactins [lo] which are known in the presence of Ca2+ to bind to phospholipids and actin, and consequently lipocortin was shown to inhibit PLA~~acti~~y by sequestration of the substrate 1111.
Abbreviations: PLA,, phospho~p~e A,; apo C-1, a~l~p~~rotein C-l ; DPPC, dipaI~toyI~y~rop~~spb~hoIine; HPLC, high-pressure liquid chrumato~apby; BSA, bovine serum aIbu~n. Corr~~onden~:
F.R.G. ~~-~7~/~/$~3.50
J. Poensgen, FlandriscIx
StraI?e 34, D-51
Aache%
Dialyzed human serum exhibits a considerable PLAY-in~biting capacity which cannot be accounted for by lipo~ortin or albumin (~bu~n interferes with some PLA,-assays [12]). Very likely the serum contains other inhibitory proteins as well. I have isolated one of these factors in order to see whether it acts on the substrate or on the enzyme and if it is related to lipocortin. A preli~na~ report was given at the 4th PhosphoIipase Symposium, Universitat Ulm, June 10th to 11th 1988. Materials and Meth~s
PLA, from hog pancreas, RPM1 1640 medium and fetal calf serum were obtained from Boeh~~ger (Mannheim, F.R.G.), [l-‘4C]aracbidonie acid was purchased from Amersham Buchler (Brau~schweig, F.R.G.), af ~u~~~~~~~-l-14C]dipal~toyigly~erophospb~ho~iue was bought from New England Nuclear (Dreieichenh~n, F.R.G.), the unlabeIled compound and DMSO for cell cultures were from Sigma (Taufkirchen, F.R.G.). Nucleosif 5 C,, was a product of ~acherey und Nagel (D&en, F.R.G.), Bio-Gel P-30 was obtained fro,” Bio-Rad (~~nchen, F.R.G.), and silica-gel (63-100 A) was produced by ICN Bi~he~c~s (~s~hwege, F&G.}. For PLA~*i~bition assay with soluble PLA,, dpm per experi[i4C!]DPPC (20 or 100 nmol and 15
0 1990 Elsevier Science Pubfishers B.V. ~~iornedic~ Division)
189 mental point) was dried under N,, taken up in 0.05 M Tris-HC1/4 mM CaCl, (pH 8.0), to a concentration of 0.4 or 2 mM, respectively, and sonicated for 30 min above 37 * C. 50 ~1 liposomes were added to 50 $0.05 M Tris-HCI (pH 8.0), containing 5-10 pmol PLA, or 10 pmol PLA, and inhibitor. After 5 min incubation at 37” C, 10 ~1 1 M HCl was added and the mixtures transferred to small plastic columns with 1.5 ml silica-gel (63-100 A). Free [‘4C]palmitic acid was eluted by four washes with 1.25 ml hexane/ dioxane/ acetic acid (70/30/L v/v), ]13]. For the PLA, assay with HL 60-homogenates, HL 60 cells were grown in RPM1 1640 supplemented with 10% fetal calf serum, 2 mM glutamine, 1 mM sodium pyruvate, 1 mM nonessential amino acids to lo6 cells/ml as described in Ref. 14. 50 U/ml penicillin and 50 pg/ml streptomycin were added routinely. Differentiation was induced by cultivating 3. lo5 cells/ml for 5 days in the medium described above in the presence of 1.25% (v/v) DMSO. During the last 18 h the cells were labelled with 1 ,nCi [1-‘4C]arachidonic acid per 100 ml culture medium. The labelled cells were pelleted, washed once with Hepes/saline/BSA (25 mM Hepes, 125 mM NaCl, 0.7 mM MgSO,, 0.4 mM EGTA, 10 mM glucose + 2 mg/ml fatty-acid-free BSA (pH 7.4), and resuspended at 2 - 10’ cells/ml in Hepes/saline/BSA. They were broken by ultrasonication (3 - 15 s at 0” C in a Branson sonifier B 15 equipped with microtip) and used immediately: 0.2 ml cell homogenate was preincubated at 37 o C with and without apo C-l dissolved in DMSO (final concentration 2% DMSO - controls received 2% DMSO only). Arachidonic acid release was started by addition of 0.6 pmol CaCl, in 0.1 ml Hepes/ saline/BSA. At 15 min intervals the samples were quenched and extracted with 1.5 ml CHCl,/CH,OH/ CH,COOH (2 : 1: 0.05, v/v). The organic phase was dried under N,, the residue was dissolved in 25 ,ul CH,Cl/CH,OH (2 : 1, v/v) and applied to silica-gel TLC plates. Phospholipids and arachidonic acid were separated with CHCl,/CH,OH/ H,O (69 : 27 : 4, v/v) and identified by comparison to standards. The dist~bution of the label was determined by scanning with a Berthold Automatic TLC Linear Analyzer LB 2842. PLA, activity is given as X radioactivity in arachidonic acid per x min. Protein concentration was determined by the Bradford method 1151or, in the case of purified apo C-l, by the amino acid content of its acid hydrolysate: The pure protein was hydrolyzed in 6 M HCl containing 0.1% thioglycolic acid at 110 o C for 20 h. Amino acid composition was determined with the amino acid analyzer LC 6001 from Biotronik (Frankfurt, FRG) using a 0.6 cm X 21 cm column of DC 6A resin. Identification of the amino acids was carried out by conventional ninhydrin analysis combined with an elution pro-
gram controlling the flow with time of the 5 buffers employed [16]. The amino acid analysis given is not corrected for partial destruction (e.g., serine or threonine) or for incomplete release during hydrolysis (e.g., valine or isoleucine). Sequence analysis of the purified protein was performed essentially as described by Edman [17] using an automatic liquid-phase sequencer from Beckman, Model 890M. Phenylthiohydantoin derivatives were analyzed by isocratic HPLC on a superspher CH, column (Merck, Darmstadt) according to Ref. 18. For prepara~on of apo C-l, 1 1 of human serum was cooled to 0 o C and brought to 65% saturation with solid (NH,),SO,. Following centrifugation, the supernatant was ultrafiltered and dialyzed in a stirred cell (Amicon) equipped with a PM 10 membrane until the volume was reduced to 100 ml and the (NH4)*S04 concentration was about 10% of the original. 33 ml each were mixed with 15 ml acetic acid and fractionated on a Bio-Gel P-30 column (4.6 x 80 cm) equilibrated with 30% acetic acid. Inhibitory fractions eluted after the exclusion peak were refractionated twice, first in a medium-sized (3 x 50 cm) and finally in a small (1.6 x 70 cm) column, both filled with Bio-Gel P-30 in 30% acetic acid. The resulting small polypeptides were further separated by reversed-phase HPLC on Nucleosil 5 C,, using the following gradient of 0.05% tri~uoroacetic acid in water (A) and 0.025% trifluoroacetic acid in acetonitrile (B): O-41% B in A in 75 min, 41% B in A for 20 min, 41-60% B in A in 5 min, 60% B in A for 10 min. Flow rate: 1.6 ml/mm, T = 60 o C, detection at 214 nm. The isolated protein was kept in acetonitrile/acetic acid/ water (5 : 3 : 2 v/v) at - 20 o C until used for the experiments described. Then the solvent was evaporated, the residual material was dissolved in DMSO and diluted with 0.1 M Tris-HCl (pH 9.0), yielding a final pH of 8.0. Results
Fractionation of human serum by various chromatographic procedures showed the major part of the inhibitory activity (as assayed by its effect on the hydrolysis by pancreatic PLA, of sonicated DPPC-vesicles) to comigrate with the main protein peaks. However, by gel chromatography about 10% of this activity were eluted with the small molecular weight fraction and even appeared beyond vi. As I wanted to avoid the use of detergents and had ascertained that the inhibitory activity did not decline in 30% acetic acid, I aimed for a small protein and fractioned the serum in 30% acetic acid so as to reduce interactions with the gel matrix. Subsequent to the removal of the majority of the proteins with ammonium sulfate, the remainder was concentrated and fractionated by repeated gel chromatography on B&Gel P-30 in 30% acetic acid, which
190 Isolated
:
T
P
0
V
5 S
S
A
L
D
‘::
L
K
E
F
15 G
N
T
L
E
:
APO C-l
:
T
P
D
V
5 S
5
A
L
D
‘K”
L
K
E
F
‘2
N
T
L
E
200
protein
:
U
A
RELISRI
30 K
Q
s
E
L
3:
A
K
M
R
40 E
K
A
R
30 K
Q
S
E
L
3:
A
K
M
R
“Eo
protein
25 lsoloted
ApOC-1:
E
*?
I
S
R
I
Fig. 1. The sequence of the first 40 amino acid residues of the isolated protein and of apo C-l.
eluent reduces the exclusion limit to 10 kDa. The resulting group of small polypeptides was further purified by reversed-phase HPLC yielding an in~bito~ protein of 6.5 kDa (determined by SDS gel electrophoresis). About 1 mg was obtained from 1 liter of serum. Determination of the amino acid composition revealed that the isolated polypeptide contained neither cysteine nor tyrosine. The sequence of the first 40 amino acids was determined and showed the purified protein to have two NH,-termini - about half the material was shortened at the amino end by two amino acid residues. Comparison to known sequences revealed identity with apo C-l, a major component of very light-density lipoprotein which consists of 57 amino acid residues [19]. Judging from the amino acid composition (given in Table I) and the sequence of 40 out of 57 amino acids (shown in Fig. 1) I took the isolated protein to be apo C-l. Apo (Z-1 is known to bind to phosphatidylcho~ne bilayers [20], and I expected it to interact with the substrate, not the enzyme. For confirmation I preincubated apo C-l either with the DPPC-vesicles or with
PLA, before performing the assay. It turned out that 10 min preincubation with the substrate decreased the concentration of apo C-l needed for half-mammal inhibition by one order of magnitude (Fig. 2, the inset gives an example of the inhibition progress with increasing preincubation times). Preincubation with the enzyme has no such effect. The effect of preincubation is not due to retarded binding of apo C-l to the DPPCvesicles: The initial time-course of the hydrolysis catalyzed by PLA, in the presence or absence of apo C-l is straight and shows that apo C-l exerts its inhibitory action immediately (Fig. 3) (see Discussion).
TABLE I Amino acid analysis of the prorated protein (a) and amino acid composition
ofup0 C-I
(b)
~ 01
1
nM ape C-l
Amino acid
(a)
(b)
Asx Thr Ser Glx
5.3 2.1 7.2 11.2 1.6 3.3 1.5 1.0 2.3 6.0 0.1 3.6 0.6 8.5 3.4 0.8 n.d. a
5 3 7 9 1 3 2 1 3 6
GIY Ala VaI Met Be Leu Tyr Phe His Lys Arg Pro Trp a Not determined,
3 9 3 1 1
but the protein absorbs at 280 nm.
b PM opoC-I
Fig. 2. The inhibitory effect of apo C-l after 10 min preincubation with the substrate. (a) 1 mM DPPC in 0.05 M Tris. HC1/4 mM CaCI (pH 8.0) was incubated for 10 min at 37’ C with and without apo C-l. The assay was started by adding 10 pmol PLA, (final volume: 100 ~1) and stopped with 10 pl 1 M HCI. Free [‘4CIpahnitic acid was separated from DPPC on silica-gel coIumns and counted in a j3-counter. The bars represent standard deviations obtained with triplicate determinations. The data are representative for two experiments carried out independently. (b) PLA, was assayed without preincubation with 1 mM DPPC and apo C-l (IC,, taken from Fig. 4B). Inset: Effect of preincubation time on the inhibition by apo C-l. 2 mM DPPC in 0.05 M Tris-HC1/4 mM CaCI, (pH 8.0) was kept for 20 mitt at 37OC. At f = 0, 15, 17.5 and 20 min. duplicates of 50 pl aliquots received 30 pf = 12.6 pmol apo C-l in Tris-HCI (0.05 M, pH 8.0). 2 controls received buffer only. At t = 20 mitt, 20 pmol PLA, in 20 ~1 0.05 M Tris-HCI, pH 8.0, were added. The incubation was continued for 2 min and stopped with 10 pI 1 M HCI. (A) PLA, only. (B) PLA, +apo C-l without preincubation. (C) PLA, +apo C-l, preincubated for 2.5 min. (D) PLA, +apo C-l, preincubated for 5 min. (E) PLA, + apo C-l, preincubated for 20 min.
191
10-
Fig. 3. Time-course of the hydrolysis of DPPC-vesicles by PLA, in the presence and absence of apo C-l. 0.2 mM DPPC was incubated at 37OC with 0.1 pM PLA, +f0.55 pM apo C-1. Every 2 min, duplicates of 0.1 ml aliquots were quenched with 10 pl 1 M HCl and free pahnitic acid was obtained by elution from silicagel columns. Upper curve ( X ); absence of apo C-l, lower curve (+ ); presence of apo C-l. ,
The interaction of apo C-l with the substrate implies that the inhibitory effect depends on the concentration of the substrate as well as on that of the inhibitor. I compared the concentration of apo C-l needed for 50% inhibition with 0.2 mM and 1 mM substrate. With 0.2 mM DPPC, the hydrolysis is 50% reduced by about 0.5 PM apo C-l, whereas with 1 mM DPPC, 50% inhibition is achieved by 1 FM apo C-l (Fig. 4). Our assay uses an artificial substrate and a secreted PLA,, and the inhibition by apo C-l might well be limited to that system. If any physiological significance is to be ascribed to the inhibition of PLA, activity by apo C-l, an inhibitory effect should also be demonstrable with physiological substrates such as cell membranes and with an endogenous PLA,. We used broken HL 60 cells which had previously been induced to differentiate by DMSO and were labelled with [l“C]arachidonic acid. These homogenates were incubated with 2 mM CaCl, and the time course of arachidonic acid release in the presence and absence of apo C-l was determined. In four experiments I found 30-40% inhibition of arachidonic acid release with 13 to 130 nM apo C-l (Fig. 5). In contrast to the experiments with DPPC-vesicles (cf. Fig. 4}, inhibition was never complete and did not exceed 50% even with 1.3 FM apo C-l (data not shown).
1%; 0
1
2
3
4
j.iM ape C-l
Fig, 4. The i~bition by apo C-l of the hydrolysis of DPPC at two different substrate concentrations. PLA, was incubated for 5 mm at 37 o C with apo C-l and DPPC. A: 0.2 mM DPPC, 0.1 PM PLA,. B: 1 mM DPPC, 0.1 PM PLA, (x) or 0.2 pM PLA, (0). The data are taken from three experiments carried out independently. Each value is the mean of duplicate determinations (error bars were omitted for reasons of clarity).
This is the case with mepacrine, an often used ‘PLAN-~bitor’ and other cationic arnp~p~~c drugs [22,23]. By this mechanism several therapeutic agents 15 t
Discussion
Due to the fact that a soluble enzyme acts on a water-insoluble substrate, there are many pitfalls in the assessment of phospholipase inhibitors, and often the inhibitory effect is predetermined by the choice of the assay system. The activity of PLA, is markedly influenced by the substrate matrix 1211 and anything interfering with the structure of the substrate may inhibit (or even stimulate) PLA, action.
Fig. 5. The inhibition by apo C-l of arachidonic acid release in broken HL 60 calls. 0.2 ml homogenate of (t4C]arachidonic acidlabelled, differentiated HL 60 cells were preincubated for 10 min at 37OC with and without 20 nM apo C-l. After addition of 0.1 ml buffer containing 0.6 pmol CaCl,, the appearance of [t4C$rachidonic acid was followed by extracting the lipids and separating phospholipids and arachidonic acid by TLC in CHCI,/CH,OH/HaO (69/27/4 v/v). The distribution of the “C-label was determined in a TLC-scanner. Upper curve ( X) - control; lower curve ( +) - sample with 13 nM apo C-l.
192 produce a phospholipidosis [24], which proves that interference with the substrate is also effective in biological systems. Sequestration and thus depletion of the substrate is still another mechanism by which lipocortin and other phospholipid-bin~ng proteins may modulate the activity of PLA, [II]. In a recent publication, Conricode and Ochs also described PLAY-in~bition through substrate depletion by proteins 1251. Apo C-l is a protein which complexes phosphatidylcholine and interferes with the binding of PLA, to the substrate, The time-course of the reaction shows inhibition and therefore binding of apo C-l to be immediate, but nonetheless 10 min preincubation with the substrate increases the inhibitory capacity by one order of magnitude. I think that this phenomenon is an artefact caused by apo C-l-induced aggregation of the DPPC-vesicles, since apo C-l-complexed phosphatidylcholine vesicles have been shown to form aggregates [20]. It tallies with this assumption that the progress of i~bition with increasing preincubation times varies with different substrate preparations. Elucidation of these presumably complicated events would require a separate detailed study. Without preincubation 50% inhibition at 0.2 mM DPPC is achieved by about 500 nM apo C-l, and at 1 mM substrate ICY,, is 1 PM apo C-l. On the average, small unilamellar DPPC-vesicles consist of approx. 2500 molecules, two-thirds of which are in the outer leaflet [26]. Thus, 3-6 molecules of apo C-l suffice to shield half the surface of a liposome. Neither in~o~oration of an ‘inhibitor’ into the substrate membrane nor sequestration of the substrate achieve a specific inhibition of PLA,. Lipocortin has been shown to inhibit phospholipase D, too [27]. AS regards different substrates, the inhibition might show a certain selectivity. Probably the various phospholipid-binding proteins have different affinities to the different phospholipid components of cell membranes. Apo~poprote~s A-l and C-l bind to phosphatidylcholine which in erythrocytes is located predominantly in the outer leaflet of the plasma membrane [28]. Thus, these are well suited to shield the membrane of the red blood cell against phospholipases released into the plasma. The experiments with HL SO-homogenates prove that apo C-l in lo-100 nM ~n~ntrat~ons is well able to protect biological membranes against phospholipase attack, although complete inhibition was not attained. Possibly, certain regions of the different cell membranes do not bind apo C-l. Generally, 2 liter of serum yielded 100-150 nmol apo C-l. So the normal concentration at least equals the range found inhibitory in the HL 60 assay. I do not assume that apo~poproteins are the only extracellular inhibitors of phospholipase activity. Actually, all phospholipid-binding proteins should have similar membrane-shielding effects.
I thank Monika Nellessen for skilful assistance and Drs. G. Steffen and B. Wolf (Griinenthal GmbH) for determining the amino acid composition of the purified protein and its sequence, respectively. My thanks are also due to Dr. Habenicht (Universitat Heidelberg) for generously supplying me with HL 60 cells. I am very grateful to Drs. U. Gehring (Universitat Heidelberg) and B. Hamprecht (Universitat Tubingen) for helpful discussions during the preparation of the manuscript.
References 1 2 3 4 5 6 7
8 9 10 11 12 13 14 15 16 17
18 19
20
21 22 23 24 25 26 27 28
Creutz, C.E. (1981) J. Cell. Biol. 91, 247-256. Shier, W.T. (1980) Proc. Natl. Acad. Sci. USA 77, 137-141. Vadas, P. (1984) J. Lab. Clin. Med. 104, 873-881. Vadas, P. and Pruzanski, W. (1984) Adv. Inflam. Res. 7, 51-59. BlackweB, G.J., Carnuccio, R., Di Rosa, M., Flower, R.J., Parente, L. and Per&a, P. (1980) Nature 287, 147-149. Hirata, F., Schiffmann, E., Ve~atasubram~an, K., Salomon, D. and Axelrod, J. (1980} Proc. Natl. Acad. Sci. USA 77, 2533-2536. Clok, J.F., Colard, O., Rothhut, B. and Russo-Marie, F. (1983) Br. J. Pharmacol. 79, 313-321. Haigler, H.T., Schlaepfer, D.D. and Burgess, W.H. (1987) J. Biol. Chem. 262,6921-6930. Hirata, F. (1981) J. Biol. Chem. 256, 7730-7733. Gienney, J.R., Jr., Tack, B. and Powelf, M.A. (1987) J. Cell Biol. 104, 503-511. Davidson, F., Dennis, E.A., Powell, M. and Glenney, J.R., Jr. (1987) J. Biol. Chem. 262, 1698-1705. Vadas, P., Stefanski, E. and Pruzanski, W. (1986) Inflammation 10, 183-193. Cosentino, M.J. and Ellis, L.C. (1981) Prostaglandins 22, 309-322. Billah, M.M., Eckel, S., Myers, R.F. and Siegel, M.J. (1986) J. Biol. Chem. 261.5824-5831. Bradford, M.M. (1976) Anal. Biochem. 72, 248-254. Biotronik (1981) Appliiationsreport ASA 981/2, Munchen. Edman, P. and Henschen, A. (1975) in Molecutar Biology, Biochemistry and Biophysics (Needleman, S.B., ed.), 8, 232-279, Springer, Berlin. Lottspeich, F. (1980) Hoppe-Seyler’s 2. Physiol. Chem. 361, 1829-1834. Jackson, R.L., Sparrow, J.T., Baker, H.N., Morrisett, J.D., Taunton, O.D. and Gotto, A.M., Jr. (1974) J. Biol. Chem. 249, 530885313. Jackson, R.L., Morrisett, J.D., Sparrow, J.T., Segrest, J.P., PownaB, H.J., Smith, L.C., Hoff, H.F. and Gotto, A.M., Jr. (1974) J. Biol. Chem. 249, 5314-5320. Jain, M.K., Streb, M., Rogers, J. and de Haas, G.H. (1984) B&hem. Pharmacol. 33, 2541-2551. Vargaftig, B.B. (1977) J. Pharm. Pharmac. 29, 222-228. Grlbner, R. (1987) B&hem, Pharmacol. 36, 1063-1067. Hostetler, K.J. (1984) Fed. Proc. 43, 2582-2585. Comicode, K.M. and Ochs, R.S. (1989) Biochim. Biophys. Acta 1003, 36-43. Cornell, B.A., Middlehurst, V. and Separovic, F. (1980) Biocbim. Biophys. Acta 598, 405-410. Kobayashi, M., Bansal, VS., Singh, I. and Kaufer, J.N. (1988) FEBS Lett. 236, 380-382. Zwaal, R.F.A., Roelofson, B. and Colle, CM. (1973) Biochim. Biophys. Acta 300, 159-182.
