Biochimica Elsevier

BBALIP

et Biophysics

Acia, 1043 (1990) 75-82

53325

Competitive analogues

inhibition of lipolytic enzymes. III. Some acylamino of phospholipids are potent competitive inhibitors of porcine pancreatic phospholipase A 2

G.H. de Haas I, R. Dijkman ’ Laboratory of Biochemistry

C. B. L. E., Utrecht (The Netherlands)

(Revised

Key words:

‘, M.G. van Oort ’ and R. Verger

Phospholipase

and .’ C. N. R. S. - C. B. M. 5, 13402 Marseille (France)

(Received 9 June 1989) manuscript received 23 October

A,; Acylamino

analog;

2

Inhibition;

1989)

Lipolytic

enzyme;

(Porcine

pancreas)

Competitive inhibition of porcine pancreatic phospholipase A z was studied in mixed-micellar systems composed of longand medium-chain substrates, potential inhibitors and detergents. A number of positional and stereoisomeric monoacylamino, acyloxyglycerophospholipids were investigated for their inhibitory properties, using as substrates the corresponding diacyl-sn-glycero-3-phospholipids possessing the same polar headgroup and identical acyl chain lengths. Based on a kinetic model applicable to water-insoluble inhibitors (see accompanying paper I), which allows a quantitative comparison of the inhibitory power (Z ) of the various phospholipid analogues, the following results were obtained: Substitution of a single acylester bond in a diacylglycerophospholipid by an acylamino group can transform the substrate molecule in a potent competitive inhibitor. This property is acquired only when this substitution occurs on the phospholipase-susceptible ester bond of the substrate. If the acylamino group replaces an ester bond which cannot be attacked by the highly positional and stereospecific phospholipase, the resulting molecule binds with similar affinity to the active site of the enzyme as the parent substrate molecule. Because of its positional and stereospecificity, this so-called inhibitory ‘amide effect’ suggests that these inhibitors behave as substrate-derived analogues. The inhibitory ‘amide-effect’ observed with several medium- and long-chain monoacyloxy-, monoacylamino-deoxyglycerophosphatides is completely lost upon specific alkaline hydrolysis of the single acylester bond. Reesterification of the free glycerol OH group in these lysoacylaminophosphoglycerides, even with an acetyl residue, restores the inhibitory properties. These observations indicate that specific binding of phospholipids to the active site of pancreatic phospholipase A,, requires the presence of two chains in substrate or inhibitor structure and suggest that those results obtained with lysophospholipids and single-chain analogues may be questionable.

Introduction Competitive inhibition studies have significantly contributed to our present knowledge of intimate details of catalysis of many enzymes. Application of these techniques to .lipolytic enzymes such as lipases and phospholipases has been retarded for a long time because heterogeneous catalysis of water-insoluble, aggregated lipids by soluble enzymes does not follow classical Michaelis-Menten kinetics [l]. 25 years ago we determined the minimal substrate requirements of phospholipase A, [2]. Based on this result, a series of closely related phospholipids were investigated as potential competitive inhibitors of the pancreatic enzyme [3]. These studies, limited to short-

Correspondence: G.H. de Haas, Laboratory of Biochemistry, Tram III, Padulaan 8, 3584 CH, Utrecht, The Netherlands. 0005-2760/90/$3.50

0 1990 Elsevier Science Publishers

C.B.L.E.

B.V. (Biomedical

chain, micelle forming phospholipid substrates and analogues, showed that 2-acylamino-2-deoxyphosphatidylcholines are potent inhibitors of the enzyme. At this time, our primitive understanding of interfacial enzyme kinetics [4] prevented further exploration. The renewed interest in phospholipase A, inhibition during the last decade stems predominantly from potential therapeutical applications of strong competitive inhibitors to modulate the ‘arachidonate cascade’. The successful kinetic studies using mixed phospholipids and Triton X-100 micelles and Nuja nuju phospholipase by Robson and Dennis [5] initiated studies by several groups in which a series of potentially active competitive inhibitors of phospholipase A, were synthesized. Hajdu and co-workers [6-S], as well as Magolda et al. [9], developed new or improved chemical syntheses of several acylamino analogues of phospholipids and confirmed the potent inhibitory properties of these substrate-derived compounds. Yuan et al. [lo] extended the Division)

76 possibilities to competitively inhibit phospholipase Az by synthesizing a series of substrate-derived fluoroketones and sulfur-substituted phospholipid analogues. Taking into account our current knowledge of interfacial enzyme kinetics (see accompanying Paper I) and the availability of various assay systems, a renewed investigation of a series of N-acylaminodeoxyglycerophospholipids as potential inhibitors of pancreatic phospholipase A 2, seems to be justified. The more so because such compounds are essential to detailed studies of enzyme-lipid interactions by high-resolution Xray diffraction and *H NMR. In this first contribution on phospholipase inhibition, long- and medium-chain phospholipids and analogues were investigated in a mixed-micellar system with bile salts, using the pH-stat method. The behavior of the medium-chain compounds (C,, analogues) which can be analyzed also in the absence of detergents by the monomolecular surface film technique will be subject of a following publication. TABLE

The key concepts of this paper were presented by G.H. de Haas at the Christmas meeting of the Biochemical Society, London, December 1987. Materials and Methods As long-chain and medium-chain substrates and potential inhibitors of porcine pancreatic phospholipase A,, the compounds listed in Table I have been used. Porcine pancreatic prophospholipase A z was isolated, purified and converted into the corresponding phospholipase A, as described by Nieuwenhuizen et al. [12]. Sodium deoxycholate was purchased from Merck, sodium taurodeoxycholate from Sigma. Phospholipuse

assay

Enzymatic activity on the various phosphatidylcholine substrates (pmol/,min per mg protein) was determined by the pH-stat method at pH 8. Conditions: 5 mM Tris-HCl; 10 mM CaCl,; T = 25” C. Substrate

I

(A) Long-chain

Compounds (l-4) and the etherphosphatidylcholines other compounds are given in the preceding paper,

(9,10), were prepared

Substrates (1) (2) (3) (4)

as described

earlier

[II]. The syntheses

and structural

II.

Short-hand

l-palmitoyl-2-myristoyl-sn-glycero-3-phosphocholine l-myristoyl-2-palmitoyl-sn-glycero-3-phosphocholine l-myristoyl-3-palmitoyl-sn-glycero-2-phosphocholine l-palmitoyl-3-myristoyl-sn-glycero-2-phosphocholine

notation

16/14/PN 14/16/PN 14/PN/16 16/PN/14

Inhibitors

Short-hand

(5) (6) (7) (8) (9) (10) (11) (12) (13) (14)

rat-16/N-14/PN rat-N-14/16/PN N-14/PN/16 16/PN,‘N-14 rat-1 6/O-14/PN rat-O-14/16/PN rat-OH/N-14/PN rat-N-14/PN/OH rat-N-14/0H/PN rue-OH/N-14/PG

rac-l-palmitoyl-2-myristoylamino-2-deoxyglycero-3-phosphocholine, rac-l-myristoylamino-2-palmitoyl-l-deoxyg~ycero-3-phosphocholine l-myristoylamino-3-pal~toyl-l-deoxy-sn-glycero-2-phosphocholine l-palmitoyl-3-myristoylamino-3-deoxy-sn-~ycero-2-phosphocholine rac-l-palmitoyl-2-tetradecylglycero-3-phosphocholine rac-l-tetradecyl-2-palmitoylglycero-3-phosphocholine rac-2-myristoylamino-2-deoxyglycero-3-phosphocholine rac-l-myristoylamino-l-deoxyglycero-2-phosphocholine rat-1 -myristoylamino-1-deoxyglycero-3-phosphocholine rac-2-myristoylamino-2-deoxyglycero-3-phosphoglycol

notation

(B) Medrum charn

The synthesis

of (15-25)

is described

in the preceding

paper,

II.

Substrates

Short-hand

(15) 1,2-dilauroyl-sn-glycero-3-phosphocholine (16) 1,2-Dilauroyl-sn-glycero-3-phosphoglycol,

12/12/PN 12/12/PG

Inhibitors

Short-hand

(17) 2,3-dilauroyl-sn-glycero-1-phosphocholine (18) 1-lauroyl-2-lauroylamno-2-deoxy-sn-glycero-3-phosphocholine (19) 3-lauroyl-2-lauroylamino-2-deoxy-sn-glycero-1 -phosphocholine (20) I-lauroyl-2-lauroylamino-2-deoxy-sn-glycero-3-phosphoglyco~ (21) 3-1auroyl-2-lauroylamino-2-deoxy-sn-glycero-l-phosphoglycol (22) 2-lauroylamino-2-deoxy-sn-glycero-3-phosphoglycol (23) 2-lauroylamino-2-deoxy-sn-glycero-l-phosphoglycol (24) 1-myristoyl-sn-3-glycerophosphocholine (25) ra~-l-acetyl-2-myristoylamino-2-deoxyglycerophosphocholine

PN/12/12 12/N-12/PN PN/N-12/12 12/N-12/PG PC/N-12/12 OH/N-I2/PG PC/N-I2/OH 14/0H/PN rat-2/N-14/PN

notation

notation

formulas

of the

77 I

concentration: 3 mM. Sodium deoxycholate and sodium taurodeoxycholate concentration: 3 mM. The inhibitory properties of the various inhibitors compiled in Tables II and III were determined by similar titration assays. Enzymatic velocity in the presence of inhibitor was measured as function of the mole fraction a( = I/I + S). The sum of the substrate and in~bitor concentration (1 + S) was kept constant at 3 mM in 3 mM detergent (deoxy- or taurodeoxycholate). These solutions are optically clear and provide such a large micellar interface that all enzyme is present in the lipid/water interface. Determination of critical micelle concentration Critical rnicelle concentrations (CMC) of various water-soluble potential inhibitors were measured by the Wilhelmy plate method 1131 and/or by means of the soluble fluorescent probe 8-anilinonaphthalenesulfonic acid (ANS) as described by de Vendittis et al. [14]. ANS was excited at 370 nM and emission was recorded at 470 nM. T= 25°C. Direct binding studies with Monomeric Substrate Analogues The affinity of porcine pancreatic phospholipase A, for several lysophospholipid monomers was studied using ultraviolet absorbance difference spectroscopy at 25°C. The difference spectra, characterized by peaks at 282 and 289 nm, are produced by lipid binding close to Tyr-69 of the enzyme molecule. From measurements of 6 A at 289 nm as a function of lipid concentration, the number of binding sites N, the ultra-violet difference extinction coefficient C and the dissociation constant K, were determined (De Araujo et al. [15], Hille et al.

Pa Differential scanning calorimetry The liquid-crystalline transition the various synthetic phospholipids described by Aranda et al. 1171.

temperatures (T,) of were determined as

Competitive inhibition at interfaces As described in the accompanying paper, I, the general formula describing the steady-state velocity under bulk conditions in the presence of a mixed micellar system containing a detergent-solubilized inhibitor and substrate reads: 0)

If we define R, as the ratio of steady-state velocities in the case where the area per molecule of inhibitor and substrate are identical then: R,=

velocity in the presence

of an inhibitor

with Ki* = K,*

velocity in the presence

of an inhibitor

with K,* + K,*

B 27

n +l4

r

r** El.6

-

ir 51.4

-

5 at.2

-

8 cn 3

I

I

I

I

I

I

O-

o-o

.m -+-. .+\. -+\

l-

0

I

I

20

40

I

I

I

60 80 100 TIME (P&n)

I

I

120 140

Fig. 1. Effect of micellar sodiumtaurodeoxycholate (5 mM) on the inactivation rate of porcine pancreatic phospholipase A, by pbromophenacyl bromide. Conditions: 10 mM Tris-HCL; 0.15 M NaCI; 10 mM Ca*+; T = 40 o C; pH = 8.0. o -0, blank without inhibitor and without detergent; + +, inhibitor (0.72 mM) without detergent; n -----a, inhibitor (0.72 mM) with 5 mM detergent.

and a as the molar

fraction

of inhibitor

I

(a=I+s 1 then Eqn. I can be written

R,=l+a

K,

as:

=1+olz

(2)

The slope of the straight line, R, = f(a) represents the inhibitory power Z of the inhibitor molecule tested. As indicated in formula (l), in a mixed micellar system containing detergent (D), substrate (S) and inhibitor (I), it is obvious that the detergent itself might behave as a second competitive inhibitor. In order to select an appropriate detergent with the lowest interfacial affinity for the enzyme (l/Kg), we applied the technique developed by Volwerk et al. [18] for the irreversible inhibition of phospholipase A, by pbromophenacyl bromide. This latter inhibitor is an active site-directed inhibitor and protection by lipids can be used as a probe for active-site binding. Fig. 1 shows that micellar sodium taurodeoxychoIate (5 mM) does not interfere with p-bromophenacyl bromide inactivation indicating no active site binding by the detergent. Also the observation that identical Z values were obtained using either sodiumdeoxycholate or sodiumtaurodeoxycholate as detergent, indicates that bile salts have no affinity for the active site of pancreatic phospholipase. Similar results have been reported for carboxyl ester lipases [20]. Finally, preliminary inhibition experiments using the monomolecular surface film technique, in the absence of detergents, yielded again Z values similar to those obtained in mixed micellar systems with detergents.

78 Results

28 R”

In Table II the inhibitory power Z of a number of phosphatidylcholine analogues is recorded. Each potential inhibitor is composed of one C,, ester and one C,, chain, the latter chain being attached via an ester, ether or amide linkage. For every phosphatidylcholine analogue, the inhibitory power (Z) was measured in combination with a strongly resembling, optically pure (Y- or P-phosphatidylcholine as substrate containing one myristoyl and one palmitoyl chain. The inhibition results of the couples 5-l and 7-3 clearly show that inhibitors containing the myristoylamino linkage in the same position as the PLA,-susceptible myristoyl ester bond in the corresponding substrates, are characterized by elevated Z-values. In contrast, if the myristoylamino bond is located in a position different from the PLA,-susceptible ester bond in the corresponding substrate (combinations 6-2 and E-4), Z values close to 0 are found, indicating weak inhibition (K,* = KG). The high affinity of the enzyme for phosphatidylcholine analogues containing an amide linkage in the scissile position is observed both for (Y- and /3-phosphatidylcholine inhibitors, notwithstanding the fact that the pure fi-phosphatidylcholine substrates are hydrolyzed with much lower V,,,,x values than the isomerit a-phosphatidylcholines. The behaviour of the combinations 7-2 and 7-1 (Z values S-10) shows that a P-phosphatidylcholine analogue containing the amide bond in position 1, which is the PLA,-susceptible position in P-phosphatidylcholine, has a considerably higher affinity for the enzyme than the corresponding diacyla-phosphatidylcholine. Again when the amide bond occupies the sn-3-position in the P-phosphatidylcholine analogue (combinations 8-2 and E-l), low Z values are

TABLE

H 24

0

0.5 lnhibltor

1.0 concentration:

0

0.5

1.0

n ( molar tractionl

Fig. 2. (A) Hydrolysis of 1.2-didodecanyl-sn-glycero-3-phosphocholine by porcine pancreatic phospholipase A, in the presence of either l-dodecanoyl-2-N-dodecanoylamino-deoxy-s~-glycero-3-phosphocholine (0 ~ o) or 3-dodecanoyl-2-N-dodecanoylamino-2-deoxysn-glycero-I-phosphocholine (X x ). For experimental conditions, see Materials and Methods. (B) R, as function of the mole fraction of inhibitor (a) for the same analogues and substrate as in Fig. 2A. The inhibitory power of the stereoisomeric ‘amide phosphatidylcholines’ is given by the slopes of the straight lines (see Materials z=40; xp and Methods). o -0. x, z=o.

found. The lower Z values found for the couples 7-2 and 7-l as compared with the combination 5-l indicate that the substrate analogue intermediate is formed more easily from a secondary ester-amide bond than from primary ester-amide linkage in /3-phosphatidylcholines. The high value (Z = 20) of the couple 7-3 can be understood if one considers that both inhibitor and substrate belong to the ~2-2 family. The kinetic results of the combinations 9-l and 10-2 demonstrate that acyl/ alkyl phosphatidylcholines are weak inhibitors of the enzyme. They bind to the active site of PLA, with a similar affinity as the diacylphosphatidylcholine substrates. Finally in agree-

II

Schemarrc structural formulas

of long-chain phosphatrdylcholine

anaiogues and of corresponding

diacylphosphcrtidylchoiines

used as .substrute

and pure analogues ( VA,,) are Full names and stereoconfiguration are given in Materials and Methods. V,,, values for the pure substrates (V,,,) as detergent. T, and TA are the transition in pmol.min-’ per mg of porcine pancreatic phospholipase A, measured with sodium deoxycholate temperatures of the pure substrates and pure analogues. respectively. The Z values represent the inhibitory power of the substrate analogue. For further experimental details see Materials and Methods.

Phosphatidylcholine (5) rat-16/N-14/PN (6) rat-N-14/16/PN (7) N-14/PN/16 (8) 16/PN/N-14 (7) N-14/PN/16 (7) N-14/PN/16 (X) 16/PN/N-14 (8) 16/PN/N-14 (9) rrrc-16/O-14/PN (10) rut-O-14/16/PN (3) 14/PN/16 (4) 16/PN/14

analogue

Substrate

V&X

GUI.

Cl

T,:

Z

(1) (2) (3) (4) (2) (1) (2) (1) (1) (2) (1) (2)

108 81 4.8 2.1 81 108 81 108 108 81 108 81

0 20 0 Cl 0 0 (1 il 0 27 4.8 2.7

29.3 34.1 25.6 27.1 34.1 29.3 34.1 29.3 29.3 34.1 29.3 34.1

25.6 33.2 26.6 26.2 26.6 26.6 26.6 26.6 32 38 25.6 27.1

22 0.1 20 0.1 8 10 0.3 0.3 0.1 0 0.1 0.1

16/14/PN 14/16/PN 14/PN/16 16/PN/14 14/16/PN 16/14/PN 14/16/PN 16/14/PN 16/14/PN 14/16,‘PN 16/14,‘PN 14/16/PN

79 TABLE

III

Schematic structural formulas of medium-chain phospholipid analogues and of the corresponding diacylphospholipids used as substrates Full names and stereoconfiguration are given under Materials and Methods. The I’,,,_ values for the pure substrates are in pmol/min per mg porcine pancreatic phospholipase A, measured with sodium taurodeoxycholate as detergent. The Z values represent the inhibitory power of the substrate analogues. Inhibitor

Substrate

Vmax

Z

(18) (19) (17) (20) (21)

12/N-12/PN PN/N-12/12 PN/12/12 12/N-12/PG PG/N-12/12

(15) (15) (15) (15) (15)

12/12PN 12/12PN 12/12PN 12/12PN 12/12PN

130 130 130 130 130

40 0 0 1100 22

(18) (19) (20) (21)

12/N-12/PN PN/N-12/12 12/N-12/PG PG/N-12/12

(16) (16) (16) (16)

12/12/PG 12/12/PG 12/12/PG 12/12/PG

950 950 950 950

17 -0.5 400 12

(22) OH/N-lZ/PG (23) PG/N-lZ/OH

(15) 12/12/PN (15) 12/12/PN

130 130

0.7 0.3

(11) (12) (13) (25)

(15) (15) (15) (15)

130 130 130 130

0 0 0 6.6

rat-OH/N-14/PN rat-N-14/PN/OH rat-N-14/0H/PN rat-2/N-14/PN

12/12/PN 12/12/PN 12/12/PN 12/12/PN

ment with previous reports [3], the results of the couples 3-l and 4-2 show that /?-phosphatidylcholines have the same affinity to the enzyme as L-a-phosphatidylcholines when the acyl chains are identical. It should be noted that one of the most potent inhibitors of Table II, MC-16N/14/PN (5) with a Z-value of 22, contains two optical antipodes. If some of these ‘amide-phosphatidylcholines’ behave indeed as tight-binding substrate analogues, one must expect that the m-3 isomer will be a much stronger inhibitor than the corresponding sn-1 antipode. To investigate this possibility, both optical antipodes of 1-dodecanoyl-2N-dodecanoylamino-2-deoxyglycero-3-phosphocholine were synthesized and their inhibitory properties were measured using 1,2-didodecanoyl-sn-glycero-3-phosphocholine as substrate in sodium-taurodeoxycholate as micellar detergent. The results are shown in Fig. 2. The inhibitory power of the pure sn-3 inhibitor, Z = 40 (see Fig. 2B) is indeed nearly twice as high as that of the racemic mixture (combination 5-l in Table II). Taking into account that the sn-1 inhibitors bind with the same affinity to the enzyme as the substrate, as shown in Table III (combinations 19-15 and 17-15) it seems evident that for these latter inhibitors no tightbinding intermediates exist and that the enzyme-phospholipid affinity is determined mainly by hydrophobic interactions. The kinetic behavior of the inhibitor-substrate combinations 20-15 and 21-15 clearly demonstrates that,

besides hydrophobic interactions, ionic forces also play an important role in binding of phospholipids to the active site of pancreatic phospholipase A 2: substitution of the choline moiety in the inhibitor molecules 18 and 19 by a glycol group transforms the zwitterionic phosphatidylcholine analogues into anionic phospholipids and results in a considerable increase in inhibitory power Z. From a comparison of the Z values, 18-15 with 19-15 and 20-16 with 21-16, having an identical polar headgroup and acyl chain length in inhibitor and substrate, it can be concluded that a correctly placed amido function in the inhibitor molecule results in a 30-50-times stronger inhibition, because of a more favourable transition state. Introduction of a negative charge in the inhibitor molecule produces a further considerable rise in Z value (compare the combinations 18-15 with 20-15), if the neutral phosphatidylcholine is used as substrate. In line with these observations are the considerably lower Z values of the combinations 18-16 and 19-16 where the zwitterionic amide-inhibitor has to compete with a negatively charged substrate. Of considerable interest are the very low Z-values for the stereoisomeric lysophospholipids 22 and 23. Notwithstanding a negative charge and a correctly placed amide function in the inhibitor 22, these compounds seem to bind with a rather similar affinity to the active site of the enzyme as the neutral diacylphosphatidylcholine substrate molecule. Of course, the presence of only one acyl chain in the inhibitor molecule 22 will considerably diminish the hydrophobic interactions with the enzyme and this effect could partly explain the low Z value of the, inhibitor 22. On the other hand, the similar Z value for the stereoisomer 23, which has the amide function in the wrong stereochemical position, indicates that in these lysophospholipids, the inhibitory ‘amide-effect’ is lost. The loss of the inhibitory ‘amide-effect’ in lysophospholipid analogues was confirmed by inhibi-

TABLE

IV

Binding data of isomeric acylamino~sophosphatidylcholine analogues to porcine pancreatrc phospholipase A_, obtamed by ultraviolet absorbance dijference spectroscop.~~ Conditions: 50 mM acetate (pH 5.4); 50 mM Ca++; 0.1 M NaCl. For reasons of comparison. data of 1-myristoyl-sn-3Iysophosphatidylcholine (24) are included. Lysophosphatidylcholine analogue

CMC(pM)

N9

Ch

K, (mM)

(11) (12) (13) (24) (25)

115 120 100 50 50

1.1 1.0 1.0 1.0 1.0

400 500 500 500 207

115 111 136 60 10

rat-OH/N-14/PN rat-N-14/PN/OH rat-N-14/0H/PN 14/0H/PN rat-2/N-14/PN.

a Molar ratio of enzyme to lipid monomers in the final complex. ’ Expressed as M-‘.crn-’ at 289 nm.

80

tion studies using the racemic dylcholine analogues:

isomeric

lysolphosphati-

C-OH

I C-NH-CO--C,3H,, I C-P--N+

-.

(11) RAC OH/N-14/PN C-NH-CO-C,,H,,

I C-P--N+ I C-OH (12) RAC N-14/PN/OH C-NH-CO-C,,H,,

1

-2 2

C-OH I C-P--N+ (13) RAC N-14/0H/PN

All three lysophosphatidylcholine analogues yield a Z value close to 0 in inhibition experiments with 1,2dilauroyl-sn-glycero-3-phosphocholine as substrate, whereas the corresponding palmitoyl-acylated derivatives of these lysophosphatidylcholine analogues are characterized by Z values of 22, 20 and 0.1, respectively (see Table 11). As could be expected, the lysophosphatidylcholine analogues are highly water-soluble and possess very similar critical micellar concentrations (loo-120 PM). This behavior allowed direct binding studies of monomeric phospholipids to the enzyme using ultraviolet absorbance difference spectroscopy. The results are summarized in Table IV. Although the K, values cannot be determined with an high accuracy because the lipid titrations are restricted to concentrations below the CMC value, it is evident that all three lysophosphatidylcholine analogues form 1 : 1 complexes with the enzyme with very similar dissociation constants. Notwithstanding the fact that one of the optical antipodes of rut-OH/N-14/PN and of rut-N-14/PN/ OH contain an amide bond in the right stereochemical position to form in principle a strong complex with the enzyme, no significant difference was observed in K, value with compound rat N-14/OH/PN, which cannot form such an intermediate. To further explore this possibility, rat-OH/N-14/ PN (11) was acetylated and the inhibitory properties of the resulting rat-1-acetyl-2-myristoylamino-2-deoxyglycero-3-phosphocholine (25) were measured, using as substrates a series of 1,2-diacyl-sn-glycero-3-phosphocholines. The results are shown in Fig. 3. The sum of the concentrations of inhibitor and substrate was always kept constant at 3 mM in 3 mM sodium-taurodeoxycholate. V,,, situations were checked by increasing the amount of mixed micelles containing diacylphosphatidylcholine as substrate.

4

x : ACYLCHAIN

-100 6 LENGTH

6

10

12

14

OF THE SUBSTRATE

Fig. 3. The inhibitory power Z of rut-2/N-14/PN and (m -m) of rat -OH/N-14/PN (o ----00) as function of the acyl chain length X in the substrate used. X = 7 diheptanoyl-sn-3-phosphatidylcholine; X = 8, dioctanoyl-sn-3-phosphatidylcholine; X = 9, dinonanoyl-sn-3-phosphatidylcholine; X = 10. didecanoyl-sn-3-phosphatidylcholine; X = 12, didodecanoyl-sn-3-phosphatidylcholine; X = 14, ditetradecanoyl-sn-3-phosphatidylcholine; X = 15, l-tetradecanoyl-2hexadecanoyl-sn-3-phosphatidylcholine. Experimental conditions: pH-stat titration at 25OC. Buffer: 5 mM Tris-HCl; 0.15 M NaCl; 10 mM Ca’+; pH 8.0. The right ordinate (* * ) represents the specific activities (units/mg of protein) of porcine pancreatic phospholipase A, on the various mixed micellar substrates in the absence of inhibitor.

It is evident that the lysophosphatidylcholine analogue rut-OH/N-14/PN (11) is unable to compete efficiently with the diacylphosphatidylcholines used as substrates. With the exception of didodecanoylphosphatidylcholine, in every case a negative Z value was found, indicating a much weaker binding of the lysophosphatidylcholine as compared with the substrate (O-O). The introduction of the acetyl group in (11) however, restores the inhibitory properties considerably (filled squares) with optimal inhibition in combination with didodecanoylphosphatidylcholine as substrate. The observed Z value of 6.6 for the rat-inhibitor (which would mean 13.2 for the pure sn-3 compound) is more remarkable, taking into account the total number of carbon atoms in the acyl chains of substrate and inhibitor: 24 and 16, respectively. Discussion The hydrolysis of water-insoluble phospholipids by water-soluble phospholipases does not follow classical Michaelis-Menten kinetics and special techniques are required to quantitate the action of these enzymes. For a review see Ref. 1 and for a preliminary report see Ref. 21. A convenient bulk assay system consists of an optically clear mixed-micellar solution of natural or

81 synthetic phospholipids with a detergent. Hydrolysis rates can be measured by pH-stat titration. As discussed in Paper I of this series, this system can also be used to study competitive inhibition of the phospholipases under certain limiting conditions: (1) Under the assay conditions, the total lipid concentration [I] + [S] must be held constant for all mole fractions of inhibitor ((Y = Z/( Z + S)); (2) Substrate and inhibitor should have similar molecular areas and their mixtures must behave ideally; (3) The mixed-micellar interface (substrate + inhibitor + detergent) should always be large enough to allow the binding of all the enzyme to the interface. We think that optimal fulfillment of the second condition was obtained by a careful choice of substrate and potential inhibitor combinations. In most cases we have used identical fatty acid chain lengths and polar headgroups in substrate and inhibitor, the only difference between both phospholipids being the presence of one -NH-CO group instead of an -O-CO function. Table II shows that the phase transition temperatures r, of the various phospholipids used are rather close. Moreover, the force-area curves determined by monomolecular surface film techniques at the air/ water interface are nearly identical for several couples of substrate and inhibitor (data not shown). As regards the third requirement, this could easily be verified. Maximal and constant enzymatic velocities were observed at 1, 2 and 4 mM of (I + S) and detergent in equimolar mixture, indicating interfacial saturation. Under these conditions, equation II (Materials and Methods) can be used, which predicts that R, as function of the mole fraction of inhibitor (CX) should be a straight line, the slope of which gives a numerical value for the inhibitory power, Z, of the inhibitor under investigation. The results of Fig. 2B are in agreement with this prediction. From the results compiled in Table II and Fig. 2, it is clear that diacylphospholipids, in which one particular ester bond is replaced by an amide linkage, are highly effective competitive inhibitors of pancreatic phospholipase A,. The observation that high Z values are observed only when the amide-bond occupies the same position as the PLA,-susceptible ester bond in the corresponding substrate, is in agreement with the well known positional specificity of this enzyme and suggests that these derivatives behave as tight-binding substrate analogues. Taking into account also the highly stereospecific properties of phospholipase A,, it is not surprising that only 2-acylaminophosphatidylcholine analogues belonging to the naturally occurring sn-3 configuration, are able to form a transition state-like intermediate. The experimentally found Z values of 40 and 0 for the stereoisomeric amidophosphatidylcholines (See Fig. 2) support the hypothesis that a correctly placed amide bond forms a strong complex more easily than the

corresponding ester function. From Table III referring to combinations 18-15 and 19-15, 20-16 and 21-16, it can be concluded that, if polar headgroups and fatty acid chainlengths are identical in substrate and inhibitor, the 2-amide-bond in the sn-3-inhibitor gives rise to a 40-fold tighter binding to the active site as compared to the affinity of the corresponding sn-1 analogue. We propose to call the ratio of inhibitory power of the two stereoisomeric 2-amidophosphatidylcholines the specific inhibitory ‘amide-effect’. The much higher Z value (1100) observed with the combination 20-15 of Table III is partially caused by the inhibitory ‘amide-effect’. In order to explain the difference between Z values of 1100 and 40, one has to remember the well-known preference of pancreatic phospolipase A, for anionic substrates (compare the combinations 21-15 and 19-15). The lower Z values observed with the combinations 18-16 and 19-16 as compared with those of 18-15 and 19-15, also show that the inhibitory ‘amide-effect’ is counteracted by the combination of a neutral inhibitor and a negatively charged substrate. It should be noted however, that the combinations of substrates and inhibitors, containing different polar headgroups might interfere with our limiting conditions! We do not know, for example, whether the combination of a negatively charged amide-phospholipid with a zwitterionic substrate, could give rise to phase-separation phenomena induced by Ca’+. Most surprising is the weak inhibitory power observed in the combination 22-15 of Table III. Notwithstanding the correct positional and stereochemical configuration of the acylamino group and the presence of an anionic headgroup in the inhibitor, the Z value of 0.7 suggests that this lysophospholipid analogue cannot effectively compete with the neutral diacyllecithin to occupy the active site of the enzyme. It might be argued, however, that the critical micelle concentrations of these stereoisomeric lauroylaminolysoglycerophosphoglycols could be so high that even at 3 mM concentration, most of the phospholipid is still present as monomer in solution and not in mixed micelles with the detergent. For this reason we initiated inhibition experiments with the three isomeric myristoylaminolysophosphatidylcholines, 11, 12 and 13, given in Table IV. Their CMC values (around 100 PM) make sure that under our experimental conditions (I + S = 3 mM in 3 mM detergent) the potential inhibitor is present in the lipid/water interface as mixed micelles. With dilauroyl-sn-3-phosphatidylcholine as substrate, all three myristoylamino-lysophosphatidylcholine analogues gave Z values close to 0, indicating again complete loss of the specific ‘amide-effect’. Also, the results of direct binding studies given in Table IV shows similar binding constants for the three lysophosphatidylcholine analogues. Why do the three isomeric lysophosphatidylcholine analogues display a

82 nearly identical binding behavior and what is the reason for the total loss of the specific inhibitory ‘amide-effect’ in these single-chain lysophospholipids? The results of Fig. 3, where the inhibitory properties of one of the myristoylaminolysophosphatidylcholines are compared with those of its acylated derivative, using several diacylphosphatidylcholines as substrates, might provide a clue to the above questions. This experiment shows that the introduction of a second acylchain in the myristoylaminolysophosphatidylcholine analogue dramatically improves the inhibitory power Z. The observation that this inhibitory effect is obtained already after acetylation of 2-myristoylaminolysophosphatidylcholine makes clear that the absence of the specific inhibitory ‘amideeffect’ in the lysoamidephosphatidylcholines is not primarily caused by diminished hydrophobic interactions with the active site in the enzyme. Our working hypothesis at this moment is that productive binding of a substrate or competitive inhibitor molecule to the active site of the enzyme, absolutely requires the presence of two chains in the phospholipid molecule. In line with these observations are the very poor substrate properties of 2-acyl-sn-3-lysophospholipids for pancreatic phospholipases A, [19]. Acknowledgements This work has been supported financially by the Biotechnology Action Program of the E.E.C. (B.A.P.00071-NL). The authors are indebted to Mrs. J.G. Mandersloot for help in the transition temperature measurements. They thank colleagues at Fisons (U.K.) for carefully reading the manuscript. References 1 Verger, R. and De Haas, G.H. (1976) Annu. 5.77-117.

Rev. Biophys.

Bioeng.

Van Deenen, L.L.M. and De Haas, G.H. (1963) B&him. Biophys. Acta 70. 538-553. Bonsen, P.P.M., De Haas, G.H., Pieterson, W.A. and Van Deenen, L.L.M. (1972) Biochim. Biophys. Acta 270, 364-382. Zografi, G., Verger, R. and De Haas, G.H. (1971) Chem. Phys. Lipids 7, 185-206. Robson, R.J. and Dennis, E.A. (1983) Act. Chem. Res. 16. 251-258. Chandrakumar, N.S. and Hajdu, J. (1981) Tetrahedron Lett. 22, 294992952. Chandrakumar, N.S. and Hajdu, J. (1983) J. Org. Chem. 48. 1197-1202. Davidson, F.F., Hajdu. J. and Dennis. E.A. (1986) Biochem. Biophys. Res. Commun. 137, 587-592. Magolda, R.L., Confalone, P. and Johnson, P.R. (1987) US. Pat. No. 4.702.864. IOaYuan, W., Berman, 109, 8071-8081.

R.J. and Gelb, M.H. (1987) J. Am. Chem. Sot.

lObYuan, W., Fearon, 906-910.

K. and Gelb.

11 De Haas, G.H. and Van Deenen, 22. 7-11.

M.H. (1989) J. Org. Chem. L.L.M. (1960) Tetrahedron

54, Lett.

12 Nieuwenhuizen, W., Kunze, H. and De Haas, G.H. (1976) Methods Enzymol. 32B, 147-154. 13 Davies, J.T. and Rideal, E.K. (1961) in Interfacial Phenomena, pp. 46-47, Academic Press. New York. 14 De Vendittis, E., Palumbo, 0.. Parlato. G. and Bochino, V. (1981) Anal. Biochem. 115, 2788286. 15 Soares de Araujo, P.. Rosseneu, M.Y., Kremer, J.M.H., Van Zoelen, E.J.J. and De Haas, G.H. (1979) Biochemistry 18, 580-586. 16 Hille. J.D.R., Donne-Op den Kelder, G.M., Sauve, P., De Haas, G.H. and Egmond, M. (1981) Biochemistry 20, 4068-4073. 17 Aranda, F.J., Killian, J.A. and De Kruijff, B. (1987) Biochim. Biophys. Acta 901, 217-228. 18 Volwerk, J.J.. Pieterson, W.A. and De Haas. G.H. (1974) Biochemistry 13. 1446-1454. 19 Slotboom, A.J., De Haas. G.H., Burbach-Westerhuis, G.J. and Van Deenen, L.L.M. (1970) Chem. Phys. Lipids 4, 30-36. 20 Tsujita, T., Mizuno, N.K. and Brockman, H.L. (1987) J. Lipid Res. 28, 1434-1443. 21 De Haas, G.H., Van Oort, M.G., Dijkman, R. and Verger, R. (1989) Biochem. Sot. Transact. 17, 274-276.