Bi~'himica ct Biophysica Acre, 11186( 1991 ) 167-172 ~¢~ 1991 Elsevier Science Publishers B.V. All rights reserved flOO5-27fi0/91/$03.50
] fi7
:3BALIP 53787
Fatty acid specificity of bile salt-dependent lipase: enzyme recognition and super-substrate effects D a g R u n e Gjellesvik Dcl~artmcnt ~f lliot'hcntistry. Unit 'er~la" of Bc(gcn. Bergen I Norway ) (Received 8 July Iqgl )
Key words: En~'mc specificity: Bik. sail dependence; Lipasc; ('apelin oil: (('od): ((huht~ morlma): (lluman)
A putative fatty acid specificity of bile salt-dependent lipases (BSDLs) has been re.investigated. The strategy was to use two evolutionally distant, homologous BSDLs (from human and cod), and to investigate their hydrolysis of different fatty acid esters at different assay conditions affecting the physicochemical phase of the substrate. Depending on assay conditions, large variations were seen in the hydrolysis rate for esters of different fatty acids. The two enzymes displayed similar fatty acid specificity patterns, with small, but significant differences that were maintained at various assay conditions. Compared to the human enzyme, the cod enzyme showed a preference for hydrolysis of long-chain polyunsaturated fatty acyl esters (up to 22 carbons in length). On the other hand, the human enzyme hydrolysed esters of shorter chain saturated fatty acids at significantly higher rates compared to the cod enzyme. Changing physicochemical factors affecting the substrate phase induced large changes in fatty acid specificity that affected both enzymes in similar manners. It is concluded that though the aliphatic chains of the fatty acids may not be recognized by the enzymes, these chains indirectly affect the conformation or interracial availability of the carboxyl ester bond in the substrate, and the enzymes show minor specificities for variations in these structures.
Introduction Pancreatic bile salt-dependent lipase (BSDL) is considered to play an important role in the digestion of lipids in mammals. This enzyme is charactcrized by a broad substrate specificity, which includes both solublc and insoluble carboxyl esters [1-3]. For hydrolysis of insoluble fatty acid esters, bile salts are required fl~r activity. The physiological role of BSDL has been considered to be in hydrolysis of cholesteryl and rctinyl fatty acid esters, because these substrates cannot be hydrolysed by other pancreatic enzymes. However, BSDLs also hydrolyse triglycerides, though with a considerably lower specific activity than pancreatic lipasc (EC 3.1.1.3). In mammals, this enzyme constitute 4 ~ of the pancreatic juice proteins [4]. BSDLs are particularly active towards diglyeerides [5], where the rate constant for hydrolysis may be comparable to that of pancreatic
Correspondence: D.R. Gjellesvik, ,~,~ladveien 19. N-5(M~ Bergen. Norway.
lipasc. Thcrefiwc, pancreatic BSDL may also be important in triglyceride digestion in conjunction with gastric and pancreatic lipase. Esters of certain fatty acid esters are more resistant to hydrolysis by pancreatic lipasc [6-8]. These arc especially esters of polyunsaturatcd fatty acids (PUFAs) that contain double bonds in positions J2-J5. These fatty acids are essential nutrients for most vertebrates. Hydrolysis studies using pancreatic lipasc and BSDL have shown that BSDL is necessary for cfficicnt hydrolysis of these esters, especially in diglyccrides, as diglycerides containing PUFA accumulated when pancreatic lipase was thc only lipase present [9,10]. it has therefi~re been proposed that one important function of BSDL is to hydrolyse glycerides containing PUFAs. Earlier studies of fatty acid specificities of BSDL type enzymes have yielded contradictory results. The variances in specificity may be explained I,y different methods in enzyme assays, and that the observed fatty acid specificity is thus not a property of the enzyme, but rather properties of the substrate interface (supersubstrate) [i 1,12]. It has therefore been proposed that BSDLs are non-specific not only towards different lipids, but also towards their fatty acids.
168
Wc recently r~:ported the purification and characterization o f a BSDL from cod pancreas (Gjcllesvik et al., submitted). The enzyme is most probably homologous to mammalian pancreatic BSDLs, and is furthermore the only lipolytic enzyme contained in cod pancreas. This first preparatkm of BSDL from non-mammalian sources allows us to investigate basic properties of this enzyme class in comparative studies, The present study concerns the putative fatty acid specificity of the cod and human BSDLs. The hypothesis fiw this work was that variances in filtty acid specificity shared by both enzymes may be ascribed to substrate properties. On the other hand. differing fiflty acid specificities of the two enzymes at identical conditions may be ascribed to enzyme properties. Only the latter properties reflect "true" fatty acid specificity of the enzymes. The background for this hypothesis is that the cod, feeding on relatively large amotmts of polyunsaturated fills, may have ew~lved a digt'stive lipaso better suited fiw this purpose than terrestrial vertebrates. Hence. the cw~lutionary distance between fish and human may be useful fiw elucidating basic knowledge of enzyme mechanisms.
Materials and Methods
Materials Pancreatic bile salt-dependent lipase from cod was purified as earlier described (Gjellesvik et al., submitted). Immunoaffinity-purified human pancreatic bile salt-dependent lipase was a generous gift from Dr. D. Lombardo. INSERM U260, Marseille. France. Porcine pancreatic lipasc (crude. Type V i i methylated fatty acids and Na-taurocholate 199ci: ) was purchased from Sigma (U.K.). Purified capelin oil was a kind gilt from The Norwegian Herring Oil and Meal Industry Research Institute (Bergen. Norway). TLC plates (Kicselgel 60, thickness: 0.2 ram) was obtained from Merck (Germany). (k~d bile salts were extracted from natural cod bile as previously described [13]. EII2)'ItlC
[IS.4itlYA "
All enzyme assays (except assays with gum arabicemulsified capclin oil) contained 150 mM Tris-HCI (pH 8.5). 15 mM Na-taurocholate and substrate. Unless otherwise indicated, the incubations were made at 18°C. Activity towards emulsified substrate was measured titrimetrically with gum arabic emulsified capelin oil as earlier described for triolein [14], except that the total volume was 5.3 ml. Bile salt was added either as 11.5 ml natural cod bile or 200 mM Na-taurocholate. For hydrolysis of fatty acid methyl esters, substrate (25 mM in ethanol) was added to a final concentration of 0.25 mM in a total volume of 0.5 ml. Alternatively, substrate dissolved in hcxane was evaporated to dry-
ness and solubilized in 100 mM Na-taurocholate, which was added to the incubation mixture to the final concentrations described above. Incubation, cxtractkm and product analysis was done as previously described (Gjellesvik et al.. submitted). For hydrolysis of capelin oil, the substrate was added in cthanolic solution (10 m g / m l ) to a final concentration of 0.5 m g / m l . The solution was turbid and stable. Alternatively. capclin oil in hcxanc (ft~l m g / m l ) was evaporated to dryness, and sonicatcd twice in ftXl mM Na-taurocholate for 311 s with a microtip (Branson Sonifier). The enzyme assays (I m l ) w e r e made up to the concentrations depicted alcove, with 11.5 mg capelin oil. The hydrolysis was started by the addition of enzyme, and the reaction was stopped by the addition of 0.2 ml 2 M HCI. Fatty acids and glycerides were extracted by the addition of 4 ml chloroform / methanol (I :2, v / v ) and I ml of water. The reaction products were separated on TLC with hexane/ethyl acetate/acetic acid (80 : 25 : 11.7,v/v), visualized by brief exposure to iodine vapour and marked. After sublimation of iodine, spots were ~ r a p e d out and extracted with dicthyl ether. Fatty acids were methylated in 12% BF.~ in methanol, and fatty acids in glycerides were transmethylated as previously d e ~ r i b e d [15]. The methyl esters were analyzed by gas chromatography in a Supelcowax I(I f u n d silica column (Supelco, U.S.A.) with helium as carrier gas in an HP 5710A gas chomatograph (Hewlett Packard, U.S.A.). The temperature gradient was 120-250°C, 8 C°/min. The GC was connected to a computer with Boreal chromatography analysis software (Boreal, Paris, France). Peaks were identified from individual fatty acid esters (Sigma) and a standard mixture (PUFA-I, Supelco).
I'ositional determhlation of fatty acids in cape~in oil. Capelin oil triglyceride (11.5 mg in alcoholic solution) was enzymatically hydrolysed at the conditions described above, using 2 mg crude porcine pancreatic lipase (Sigma type VI), and 0.1 mg each of purified cod and human BSDL. After I(I rain of incubation at rtmm temperature, the reaction products were only fatty acids and monoglycerides 12:11 by TLC. The fatty acid compt~itions in the free fatty acid and monoglyceride fractions were determined as described alx~ve, and interpreted as the compositions in 1,3- and 2-glyceryl positions, respectively. Results
Faro' acid methyl esters Hydrolysis of fatty acid methyl esters was measured either on methyl esters added in alcoholic solutions or solubilized in Na-taurocholate. The first gave slightly turbid solutions, while the latter resulted in clear solu-
169 tions, probably with mixed t a u r o c h o l a t e / s u b s t r a t e micelles. The specific activities of the cod BSDL was I / 3 (alcoholic sul'strate) and I / 4 (TC-~lubilized) that of the h u m a n BSDL. For illustrative purposes, the results are shown as activities normalized with respect to the hydrolysis rate of methyl oleate (Fig. 1). The specificity patterns of the two enzymes were relatively similar, with some quantitative differences between the cod and the h u m a n enzyme. The cod enzyme showed a higher relative activity rewards polyunsaturated fatty acyl esters compared to the human enzyme. TC-solubilizod substrate (Fig. 1B) altered these specificity patterns, as the hydrolysis of 2 i 1 : 4 (arachidonic acid) and 20: 5 ( E P A ) esters was dramatically reduced compared to that of the alcoholic substrate experiment (Fig. IA). O n the o t h e r hand, hydrolysis of 2 2 : 4 and 2 2 : 6 esters was relatively increased when the substrates were solubilized with Na-TC.
Hydrolysis o f capelin oil T h e positional specificity of the cod BSDL on triglycerides has been shown to be 1,3-specific [14], (Gjellesvik et ai., submitted). Therefore the composition of fatty acids in this position must be determined and taken into account when establishing a measure for specificity on triglyceride fatty acids using this enzyme. By enzymatic hydrolysis of the 1,3-bonds using porcine pancreatic lipase and the present two BSDLs, a complete hydrolysis of capelin oil into monoglycerides and fatty acids (approx. I M G : 2 FFA. lespec-
20 ~
' ~-~--~---~
,
3
I
A!
IL
>~ l o
" 'llJl.dll[
i
20
B ->
~ to "6
0
o4
~
o4
¢q
~
u3
~
.~
~o
Forty ocid methyl es*,er
Fig. I. Hydrolysis of individual faUy acid methyl osiers by Ihe cod
and human BSDLs. The activities are shown as relative activity compared to specific activity towards oleic acid methyl ester of the cod enzyme. (A) Substrates added in alcoholic solutions. (B) Substrates solubilized in 100 mM Na-taurocholate.
TABLE I
Fatty acid COml~),sitions of calaqin oil Total com~)sition was obtained by transmelhylation of intact capelin oil. The 1,3-1~sition and 2-position com[a)silions were obtained by analysis of liberated fatty acids and monoglyccridcs, respectively. after complete hydrolysisto fatty acids and mom)glyceridcs, respeclively, after c(nnpletc hydrolysisto fatty acid and monoglyccridcsby pancreatic lipa~, cod BSD[. and human BSDL The obtained values for 1.3- and 2-~)sition htty acids are back calculated h) average fatty acid coml~ition and sh(~vnas a control n.d.. no! detected. Fatty acid
Total
1.3-position
2-position
Control
14:ll (myristic) 16:li (palmilic) 16: I (palmilolcici 18:Iw9 (oleic) 18:Iw7 (vaccenic) Ig :2 (linoleic) 18:4 ((~ladecatclraem)ic) 2(1: I (cicosenoic) 20:5 (cicosapentaenoic) 22:1 {doc(~aem)ic) 22:6 (docosahcxaem)ic)
0.5 12.0 1(I.3 I11.1) 2.5 i.+ 5.1 2(I.9 7.0 12.S 6.3
4.1 11.3 7.8 I.i 7 3.4 i.9 5.11 27.5 4.6 17.2 2.b
IS.2 27.2 12.3 6.4 n.d. 2.1 5.1 6.7 &l 3.1 II).S
8.8 16.6 0.3 12.11 2.2 2.0 5.(I 2n.5 5.8 12.5 5.3
tively) was achieved. The fatty acid coml~)sitions in the two fractions were determined by GC-analysis, and are given in Table !. Attempts to produce mixed micellar ~ l u t i o n s of capelin oil and bile salt failed. T h e substrate was added either as an alcoholic solution, or capelin oil was dispersed in the assay mixture by sonication. Both solutions were opaque but stable, indicating a stable dispersion of triglyceride droplets. The cod and human enzymes displayed similar specific activities in these assays, judged by product formation only, as seen by GC-analysis of liberated fatty acids. Two conditions were met for all incubations: First, in order to minimize the effect of acyl migration, all incubation times were kept to a minimum (5 to 2(1 min, depending on incubation temperature). Second, in order to allow discrimination between different fatty acids (especially those of low abundance in the 1,3position), less than 10% of total fatty acids should be liberated. To correct for the varying abundance of the different fatty acid species in the 1,3-p(~ition of the trigiyeerides (Table l), the results are presented as 'relative activity'. This value is the result of the fraction of each fatty acid species divided by the fraction of this species found in the 1,3-position of the substrate, in this way, a fatty acyl esters with "relative activity' greater than 1.0 is hydrolysed at a faster rate than the expected if no discrimination occurred. Fig. 2 shows the results of incubations with capelin oil added in alcoholic solution (A) and dispersed by sonication (B). The figure shows that both enzymes displayed similar patterns of ester hydrolysis. The hu-
170 5
t.o I
- -
0
~
EL
~, 0.5 0.0 2.5
~r
2.0
B 0 >. 5
,
,
,
,
,
.> 4 0.0 9. ~t
o t,o
T. to
.-: eo
.~. eo
~ eo
T. o
.~. o
T. e4
.~. r~
C
~2 .
Fatty acid
.
.
.
.
.
.
.
.~.
..
Fig. 2. Relca~,c o l tany acidr, during initial hydrolysis of capclin oil by human and cod BSI)L. Rclalivc activity means fraction of released fatty acid species divktcd with fraction o f thai laity acid species contained in Ihe 1.3-position of capclin oil. The incubations were done al 18"(" fi)r lil rain using 54 ~ug c(~ BSDL and 38 /.tg human BSI)I_ (A} Substrale added in alcoholic solution. (ID Subslralc dispersed by ~mication.
O ¢..) (.3 ¢D U ¢.~ Folly acid Fig. 3. Re ease (ff fatty acids during initial hydrolysis of capelin oil by human and cod B S D L at different lcmpcrai.urc~.. For dclails, see legend to Fig. 2. Substrate was added as alcoholic solution. ( A ) Incubated at It]PC for ")O rain. (B) Incubated al 18°C for I 0 rain (identical to Fig. 2A}. (C) Incubated at 37°C for 5 rain.
man enzyme showed higher activity towards mediumchain fatty aeyl esters (myristic and palmitic acid). while the cod enzyme generally was more active towards esters of h)ngcr polyunsaturated fatty acids. Both enzymes hydrolyscd the fatty acyl esters of the 22carbon series at significantly slower rates. Since the two assays yielded comparable results, the assay using substrate added in alcoholic solution was used for further investigations. The effect of incubation temperature on relative hydrolysis rates is shown in Fig. 3. Assays were incubated at low (10°( ". panel A). medium (18°C. panel B) and high (37°C. panel C) temperature. The differences in relative fatty acyl ester hydrolysis of the two enzymes wcrc similar to that seen in Fig. 2. but the various incubation temperatures resulted in markedly different hydrolysis rate patterns for both enzymes. This effect was seen as an increased liberation of medium-chain fatty acids with increasing incubatilm temperature. while the hydrolysis of polyunsaturated fatty acyl esters (18:4 and 20:5) declined with temperature. When bile salts obtained from natural cod bile was used in place of pure Na-taurocholate (Fig. 4). the results were similar to that seen in Fig. 2. though the differences between the two enzymes became more
However, the activities of the two enzymes were measured on emulsified capelin oil titrimetrically. The results showed that the cod and the human enzyme behaved very differently towards this substrate, in the presence of 18.8 mM Na-TC, the specific activity of the cod enzyme was 17.1 + 1.2 U / m g , while specific activity of the human enzyme was 4.1 _+ 0.8 U / m g . However, in the presence of 9% natural cod bile, the specific activity of the cod enzyme was increased to 63.7 __ 8.4 U/mg, while specific activity of the human enzyme was very low (i.2 U / m g ) . These results indicate that the largest difference between the cod and human enzyme is in substratephase specificity, and the observed differences in specific activity in the assays
pronounced.
Fatty acid specificity studies using gum arabicemulsified capelin oil was not possible because gum arabic entered the chloroform phase during extraction.
¢D ¢D ¢D t.J
2.0
~
>= ~.5
!
i.o
i
2~o. 0.0 u
~
u
u
u
~)
u
u
u
~
Fatty acid Fig. 4. Release of fatty acids during initial hydrolysis of capelin oil by human and cod BSDL with bile salts from natural c(~J bile. Experimental conditions were identical Io Ihosc d e ~ r i b e d in legend to Fig. 2A.
171 used in this work may be ascribed to different substratc phases. Discussion
The objective of this work was to establish putative fatty acid specificities of bile salt-dependent lipases. It is recognized that the term 'fatty acid specificity' consists of at least two components: One component originates from enzymatic properties (enzymatic specificity -classical substrate fit in active site) and the other component is a result of physicochemical substrate properties (substrate orientation and ester availability at the interface), also called supersubstrate, in order to separate these components experimentally, two evolutionally distant BSDLs (liable to differ in fatty acid specificity) were used under different conditions of substrate presentation. The catalytic characteristics of the two enzymes in study are qualitatively similar, but with ~ m e quantitative differences (Gjellesvik et al., submitted). Shortchain fatty acid 4-nitrophenyl esters (especially soluble esters) were hydrolysed more effectively by the human enzyme, but the specific activities toward long-chain fatty acid esters were similar. Hydrolysis of methyl esters revealed the importance of the substrate phase for fatty acid specificities of the two enzymes, since the hydrolysis rates of certain esters were greatly influenced when the esters were in a dispersed or mixed micellar state. This was especially true for the 20:5 (EPA) ester, which was preferred as snbstrate for both enzymes in the dispersed state. However, enzyme-specific effects were seen, as the 18:2, 22:4 and 22:6 fatty acid esters were more effectively hydrolysed by the cod enzyme in the mixed miceilar state, while the activities for the human cnzyme remained unchanged. The cod BSDL has been shown to be 1,3-spccific for hydrolysis on triolein [14], (Gjellesvik et al., submitted). Therefore the present hydrolysis rates for individual fatty acyl esters are normalized according to tbeir abundance in the 1,3-position. However, mammalian BSDLs have been reported to hydrolyse also the 2-ester bond in triglycerides [1,3,16]. On the other hand, studies on sequential hydrolysis of triolein have shown that monoglyceride hydrolysis of the human milk BSDL (which is identical to the pancreatic enzyme [ 14,17,18]) is rate-limited by acyl migration [5]. Also the human pancreatic BSDL has been found to have low activity towards monoolein, even after long (2 h) incubation [191. There were small differences on the fatty acid specificity patterns if capelin oil was added in alcoholic solution or dispersed by sonication. This is probably due to similar physicochemieal states of the substrate, which gave turbid .solutions in both cases, in both
cases, the cod enzyme showeti a slightly higher activity towards fatty acid esters containing more than two double bonds, while the human enzyme wa.~ more active on the fatty acids of shortest length. An increase in temperature is likely to alter tbe substrate state, (or binding of enzyme to the interface) more than altering an active site-fit of substrate. The effect of temperature does not alter the differences in specificity between the two enzymes, hut I~th enzymes show a generally increased activity t t ~ t r d s esters of medium-chain fatty acids (myristic and palmitic) at the expense of long-chain fatty atvI esters (especially if polyunsaturated). This indicates that temperature affects the substrate, and not specificity of the enzymes. Cod bile consists of primarily taurine-conjugated bi!e salts of di- and tri-hydroxy cholanic acid in a I :5 ratio, [20,21]. Since fatty acid specificity may be dependent on substrate surface structure, it was of interest It) investigate the effect of natural cod bile on fatty acid specificity. The results showed that natural cod bile did not alter the pattern of fatty acid specificity, but the differences between the cod and human enzymes became more apparent. Noteworthy in this context is the ability of natural cod bile to enhance the activity of the c t ~ BSDL towards emulsified capelin oil. At identical conditions, the activity of the human enzyme was reduced, while the activities in the prescacc of taurocholatc were comparable to earlier findings [3]. It should be noted that the human enzyme displayed normal activity when the substratc was added as alcoholic solution. In this case, the substrate concentration was only IV, (d the substrate concentration in the emulsified assay, and the .~)lution was less turbid (partly .,a)lubilizcd) than when Na-'r(" was used. These observations may bc explained by the differcnt physiological roles of the cod and human enzymes. While the major digestivc lipase in humans is pancreatic lipase, this enzyme is absent in the cod (Gjcllcsvik c t a l . , submitted). Thus, the human enzyme can operate in a mixed micellar p h a ~ of partial glyccridcs (generated by pancreatic lipa.~), cholesterol- and rctinol-cstcrs and bile salts [12,19,22,23]. On the other hand, the cod enzyme must bc able lt) hydndy.~ emulsified lipids in the chymc effectively since pancreatic lipase is absent. Natural cod bile .seemed to facilitate hydrolysis of emulgated triglycerides by cod BSDL. It is concluded that the two enzymes s h t ~ mimer, but significant differences in fatty acid specificity that persisted in all assays used. The human enzyme was more effective in hydrolysis of shorter chain fatty acyl esters, while the cod enzyme was more active towards esters of polyunsaturated fatty acids. However, the largest effects were seen on both enzymes wben the substrate phase was altered. It is thus possible that the observed specificities mainly result from the confl~rmation (or exposure) of the carboxyl estcr bond at the
172 interface (or in the micellar phase), in spite of this, the two enzymes respond differently to these structures. ttenee, though aliphatic chains of the fatty acids may not be recognized by the enzymes, these moieties may influence the conformation (or availability) of the ester bond, which in turn may determine the enzyme affinities for particular fatty acids. This indirect effect defines a type of "non-classiear substrate specificity. Acknowledgements This work was financed in part by the Norwegian Fisheries Research Council (NFFR). I am indebted to Dr. D. Lombardo, INSERM U260, Marseillc. fl~r the supply of purified human B S D L and for discussions. References I [~rlan~m. ('. (1975) Stand. J. (;a~.lr(~:nlcrol. I(I. 4IH-408. 2 h~mbardo. I). and (;u). O. (19811) B,g'him Biophys. Acta bl I. 14% =55. 3 Lombardo. D . Fauvcl. J. and Guy. O. (198.) Biochim. Biophy, Acta 611. I.~-14h. 4 Guy. O. and Figarcna. ('. (1981) Scand. J GaslroenteroL I¢) (Suppl. 67). 50-61. 5 Wang. C.S.. llarl~;uck. J.A. and Do~'ns, D. ( 1 9 ~ ) Biochemislr~ 27. 4834-484O.
6 D.)nino. N.. Vandenburg. G.A. and Reiscr. R. (1967) Lipids 2. 489-403. 7 Brockcrhoff. |1. ~I970) Biochim Biophys. Acla 212. 02-101. 8 Ileimermann. W.II.. Ilolman. R T . , Gordon. D.T.. K~n~alyshyn. I).F,. and J c n ~ n . R.G. (1973) Lipids 8. 45-47. 9 ('hen. O.. Stcrnb). B. and Nilsson. /4. (1989) Biochim Biophy~. Acla I|NH. 372 385. I(! Chcn. O.. Sternby. B.. ,~kes~m. B. and Nilsson. ,~. (Iq~)) Biochim. Biophys. Acla 1044. I I I - ( 17. II Bna:kerhoff. It. and Jenscn. R.G. (1974) Liradylic Enzymes. Academk' Pres~, New York. 12 Rudd. I-.A. and Brockman I t . k ( 1985l in Lipases (Borgstr6m, B. and B,K:kman. iI.L,, eds.L pp. 185-~)4, Elsevier. Amsterdam. 13 Panon, J.S., Warner, T.G. anti Benson. A.A. (1977) Biochim. Biophys. Acla 486. 322-3.~). 14 Gjellesvik. D.R.. R a a e . A J . and Walther, B.T. (1989) Aquaculture 79, 177-184. 15 Kennerly. D.A. (1987) J. Biol. ('hem. 262. 16.~)5-16313. 16 Wang. ('.S., Kuksis. A.. Manganaro. F., Myher, J J . . Downs. D. and Bass. H B . (1983)J. Biol. ('hem. 258. 9197-9~)2. 17 Ilui. D.Y. and Kissel. J.A. (19~)) FEBS Lett. 276. 131-134. IS Reue. K.. Zambaux. J.. Wong. H.. Lee. G.. Leete. T.H.. R o n L T.. Shively. R., Slernby, B.. Borgslr6m. B.. Ameis. D. and Schotz. M.('. (1991)J. Lipid Res. 32. 267-276. 19 I ;ndstriim. M.. Slernby. B. and Borgstr6m. B. ((088) B/ochim. IL ~phys. Acta 959. 178-184. 20 H..dewla~l. G.A.D. and Sj/~valL I. (1054) Biochem. J. 57. 1.?.6-!~). 21 Ka~lner. A. ((W~S)Aeta ('hem. Scand. 22. ~ 6 1 - 2 3 7 0 . 22 Patton. LS. and Carey. M.C. (It;70) Science 204, 145-148. 23 Wang. C.S. and Lee. D.M. (1985) J. Lipid Res. 26. 824-8.~L
] fi7
:3BALIP 53787
Fatty acid specificity of bile salt-dependent lipase: enzyme recognition and super-substrate effects D a g R u n e Gjellesvik Dcl~artmcnt ~f lliot'hcntistry. Unit 'er~la" of Bc(gcn. Bergen I Norway ) (Received 8 July Iqgl )
Key words: En~'mc specificity: Bik. sail dependence; Lipasc; ('apelin oil: (('od): ((huht~ morlma): (lluman)
A putative fatty acid specificity of bile salt-dependent lipases (BSDLs) has been re.investigated. The strategy was to use two evolutionally distant, homologous BSDLs (from human and cod), and to investigate their hydrolysis of different fatty acid esters at different assay conditions affecting the physicochemical phase of the substrate. Depending on assay conditions, large variations were seen in the hydrolysis rate for esters of different fatty acids. The two enzymes displayed similar fatty acid specificity patterns, with small, but significant differences that were maintained at various assay conditions. Compared to the human enzyme, the cod enzyme showed a preference for hydrolysis of long-chain polyunsaturated fatty acyl esters (up to 22 carbons in length). On the other hand, the human enzyme hydrolysed esters of shorter chain saturated fatty acids at significantly higher rates compared to the cod enzyme. Changing physicochemical factors affecting the substrate phase induced large changes in fatty acid specificity that affected both enzymes in similar manners. It is concluded that though the aliphatic chains of the fatty acids may not be recognized by the enzymes, these chains indirectly affect the conformation or interracial availability of the carboxyl ester bond in the substrate, and the enzymes show minor specificities for variations in these structures.
Introduction Pancreatic bile salt-dependent lipase (BSDL) is considered to play an important role in the digestion of lipids in mammals. This enzyme is charactcrized by a broad substrate specificity, which includes both solublc and insoluble carboxyl esters [1-3]. For hydrolysis of insoluble fatty acid esters, bile salts are required fl~r activity. The physiological role of BSDL has been considered to be in hydrolysis of cholesteryl and rctinyl fatty acid esters, because these substrates cannot be hydrolysed by other pancreatic enzymes. However, BSDLs also hydrolyse triglycerides, though with a considerably lower specific activity than pancreatic lipasc (EC 3.1.1.3). In mammals, this enzyme constitute 4 ~ of the pancreatic juice proteins [4]. BSDLs are particularly active towards diglyeerides [5], where the rate constant for hydrolysis may be comparable to that of pancreatic
Correspondence: D.R. Gjellesvik, ,~,~ladveien 19. N-5(M~ Bergen. Norway.
lipasc. Thcrefiwc, pancreatic BSDL may also be important in triglyceride digestion in conjunction with gastric and pancreatic lipase. Esters of certain fatty acid esters are more resistant to hydrolysis by pancreatic lipasc [6-8]. These arc especially esters of polyunsaturatcd fatty acids (PUFAs) that contain double bonds in positions J2-J5. These fatty acids are essential nutrients for most vertebrates. Hydrolysis studies using pancreatic lipasc and BSDL have shown that BSDL is necessary for cfficicnt hydrolysis of these esters, especially in diglyccrides, as diglycerides containing PUFA accumulated when pancreatic lipase was thc only lipase present [9,10]. it has therefi~re been proposed that one important function of BSDL is to hydrolyse glycerides containing PUFAs. Earlier studies of fatty acid specificities of BSDL type enzymes have yielded contradictory results. The variances in specificity may be explained I,y different methods in enzyme assays, and that the observed fatty acid specificity is thus not a property of the enzyme, but rather properties of the substrate interface (supersubstrate) [i 1,12]. It has therefore been proposed that BSDLs are non-specific not only towards different lipids, but also towards their fatty acids.
168
Wc recently r~:ported the purification and characterization o f a BSDL from cod pancreas (Gjcllesvik et al., submitted). The enzyme is most probably homologous to mammalian pancreatic BSDLs, and is furthermore the only lipolytic enzyme contained in cod pancreas. This first preparatkm of BSDL from non-mammalian sources allows us to investigate basic properties of this enzyme class in comparative studies, The present study concerns the putative fatty acid specificity of the cod and human BSDLs. The hypothesis fiw this work was that variances in filtty acid specificity shared by both enzymes may be ascribed to substrate properties. On the other hand. differing fiflty acid specificities of the two enzymes at identical conditions may be ascribed to enzyme properties. Only the latter properties reflect "true" fatty acid specificity of the enzymes. The background for this hypothesis is that the cod, feeding on relatively large amotmts of polyunsaturated fills, may have ew~lved a digt'stive lipaso better suited fiw this purpose than terrestrial vertebrates. Hence. the cw~lutionary distance between fish and human may be useful fiw elucidating basic knowledge of enzyme mechanisms.
Materials and Methods
Materials Pancreatic bile salt-dependent lipase from cod was purified as earlier described (Gjellesvik et al., submitted). Immunoaffinity-purified human pancreatic bile salt-dependent lipase was a generous gift from Dr. D. Lombardo. INSERM U260, Marseille. France. Porcine pancreatic lipasc (crude. Type V i i methylated fatty acids and Na-taurocholate 199ci: ) was purchased from Sigma (U.K.). Purified capelin oil was a kind gilt from The Norwegian Herring Oil and Meal Industry Research Institute (Bergen. Norway). TLC plates (Kicselgel 60, thickness: 0.2 ram) was obtained from Merck (Germany). (k~d bile salts were extracted from natural cod bile as previously described [13]. EII2)'ItlC
[IS.4itlYA "
All enzyme assays (except assays with gum arabicemulsified capclin oil) contained 150 mM Tris-HCI (pH 8.5). 15 mM Na-taurocholate and substrate. Unless otherwise indicated, the incubations were made at 18°C. Activity towards emulsified substrate was measured titrimetrically with gum arabic emulsified capelin oil as earlier described for triolein [14], except that the total volume was 5.3 ml. Bile salt was added either as 11.5 ml natural cod bile or 200 mM Na-taurocholate. For hydrolysis of fatty acid methyl esters, substrate (25 mM in ethanol) was added to a final concentration of 0.25 mM in a total volume of 0.5 ml. Alternatively, substrate dissolved in hcxane was evaporated to dry-
ness and solubilized in 100 mM Na-taurocholate, which was added to the incubation mixture to the final concentrations described above. Incubation, cxtractkm and product analysis was done as previously described (Gjellesvik et al.. submitted). For hydrolysis of capelin oil, the substrate was added in cthanolic solution (10 m g / m l ) to a final concentration of 0.5 m g / m l . The solution was turbid and stable. Alternatively. capclin oil in hcxanc (ft~l m g / m l ) was evaporated to dryness, and sonicatcd twice in ftXl mM Na-taurocholate for 311 s with a microtip (Branson Sonifier). The enzyme assays (I m l ) w e r e made up to the concentrations depicted alcove, with 11.5 mg capelin oil. The hydrolysis was started by the addition of enzyme, and the reaction was stopped by the addition of 0.2 ml 2 M HCI. Fatty acids and glycerides were extracted by the addition of 4 ml chloroform / methanol (I :2, v / v ) and I ml of water. The reaction products were separated on TLC with hexane/ethyl acetate/acetic acid (80 : 25 : 11.7,v/v), visualized by brief exposure to iodine vapour and marked. After sublimation of iodine, spots were ~ r a p e d out and extracted with dicthyl ether. Fatty acids were methylated in 12% BF.~ in methanol, and fatty acids in glycerides were transmethylated as previously d e ~ r i b e d [15]. The methyl esters were analyzed by gas chromatography in a Supelcowax I(I f u n d silica column (Supelco, U.S.A.) with helium as carrier gas in an HP 5710A gas chomatograph (Hewlett Packard, U.S.A.). The temperature gradient was 120-250°C, 8 C°/min. The GC was connected to a computer with Boreal chromatography analysis software (Boreal, Paris, France). Peaks were identified from individual fatty acid esters (Sigma) and a standard mixture (PUFA-I, Supelco).
I'ositional determhlation of fatty acids in cape~in oil. Capelin oil triglyceride (11.5 mg in alcoholic solution) was enzymatically hydrolysed at the conditions described above, using 2 mg crude porcine pancreatic lipase (Sigma type VI), and 0.1 mg each of purified cod and human BSDL. After I(I rain of incubation at rtmm temperature, the reaction products were only fatty acids and monoglycerides 12:11 by TLC. The fatty acid compt~itions in the free fatty acid and monoglyceride fractions were determined as described alx~ve, and interpreted as the compositions in 1,3- and 2-glyceryl positions, respectively. Results
Faro' acid methyl esters Hydrolysis of fatty acid methyl esters was measured either on methyl esters added in alcoholic solutions or solubilized in Na-taurocholate. The first gave slightly turbid solutions, while the latter resulted in clear solu-
169 tions, probably with mixed t a u r o c h o l a t e / s u b s t r a t e micelles. The specific activities of the cod BSDL was I / 3 (alcoholic sul'strate) and I / 4 (TC-~lubilized) that of the h u m a n BSDL. For illustrative purposes, the results are shown as activities normalized with respect to the hydrolysis rate of methyl oleate (Fig. 1). The specificity patterns of the two enzymes were relatively similar, with some quantitative differences between the cod and the h u m a n enzyme. The cod enzyme showed a higher relative activity rewards polyunsaturated fatty acyl esters compared to the human enzyme. TC-solubilizod substrate (Fig. 1B) altered these specificity patterns, as the hydrolysis of 2 i 1 : 4 (arachidonic acid) and 20: 5 ( E P A ) esters was dramatically reduced compared to that of the alcoholic substrate experiment (Fig. IA). O n the o t h e r hand, hydrolysis of 2 2 : 4 and 2 2 : 6 esters was relatively increased when the substrates were solubilized with Na-TC.
Hydrolysis o f capelin oil T h e positional specificity of the cod BSDL on triglycerides has been shown to be 1,3-specific [14], (Gjellesvik et ai., submitted). Therefore the composition of fatty acids in this position must be determined and taken into account when establishing a measure for specificity on triglyceride fatty acids using this enzyme. By enzymatic hydrolysis of the 1,3-bonds using porcine pancreatic lipase and the present two BSDLs, a complete hydrolysis of capelin oil into monoglycerides and fatty acids (approx. I M G : 2 FFA. lespec-
20 ~
' ~-~--~---~
,
3
I
A!
IL
>~ l o
" 'llJl.dll[
i
20
B ->
~ to "6
0
o4
~
o4
¢q
~
u3
~
.~
~o
Forty ocid methyl es*,er
Fig. I. Hydrolysis of individual faUy acid methyl osiers by Ihe cod
and human BSDLs. The activities are shown as relative activity compared to specific activity towards oleic acid methyl ester of the cod enzyme. (A) Substrates added in alcoholic solutions. (B) Substrates solubilized in 100 mM Na-taurocholate.
TABLE I
Fatty acid COml~),sitions of calaqin oil Total com~)sition was obtained by transmelhylation of intact capelin oil. The 1,3-1~sition and 2-position com[a)silions were obtained by analysis of liberated fatty acids and monoglyccridcs, respectively. after complete hydrolysisto fatty acids and mom)glyceridcs, respeclively, after c(nnpletc hydrolysisto fatty acid and monoglyccridcsby pancreatic lipa~, cod BSD[. and human BSDL The obtained values for 1.3- and 2-~)sition htty acids are back calculated h) average fatty acid coml~ition and sh(~vnas a control n.d.. no! detected. Fatty acid
Total
1.3-position
2-position
Control
14:ll (myristic) 16:li (palmilic) 16: I (palmilolcici 18:Iw9 (oleic) 18:Iw7 (vaccenic) Ig :2 (linoleic) 18:4 ((~ladecatclraem)ic) 2(1: I (cicosenoic) 20:5 (cicosapentaenoic) 22:1 {doc(~aem)ic) 22:6 (docosahcxaem)ic)
0.5 12.0 1(I.3 I11.1) 2.5 i.+ 5.1 2(I.9 7.0 12.S 6.3
4.1 11.3 7.8 I.i 7 3.4 i.9 5.11 27.5 4.6 17.2 2.b
IS.2 27.2 12.3 6.4 n.d. 2.1 5.1 6.7 &l 3.1 II).S
8.8 16.6 0.3 12.11 2.2 2.0 5.(I 2n.5 5.8 12.5 5.3
tively) was achieved. The fatty acid coml~)sitions in the two fractions were determined by GC-analysis, and are given in Table !. Attempts to produce mixed micellar ~ l u t i o n s of capelin oil and bile salt failed. T h e substrate was added either as an alcoholic solution, or capelin oil was dispersed in the assay mixture by sonication. Both solutions were opaque but stable, indicating a stable dispersion of triglyceride droplets. The cod and human enzymes displayed similar specific activities in these assays, judged by product formation only, as seen by GC-analysis of liberated fatty acids. Two conditions were met for all incubations: First, in order to minimize the effect of acyl migration, all incubation times were kept to a minimum (5 to 2(1 min, depending on incubation temperature). Second, in order to allow discrimination between different fatty acids (especially those of low abundance in the 1,3position), less than 10% of total fatty acids should be liberated. To correct for the varying abundance of the different fatty acid species in the 1,3-p(~ition of the trigiyeerides (Table l), the results are presented as 'relative activity'. This value is the result of the fraction of each fatty acid species divided by the fraction of this species found in the 1,3-position of the substrate, in this way, a fatty acyl esters with "relative activity' greater than 1.0 is hydrolysed at a faster rate than the expected if no discrimination occurred. Fig. 2 shows the results of incubations with capelin oil added in alcoholic solution (A) and dispersed by sonication (B). The figure shows that both enzymes displayed similar patterns of ester hydrolysis. The hu-
170 5
t.o I
- -
0
~
EL
~, 0.5 0.0 2.5
~r
2.0
B 0 >. 5
,
,
,
,
,
.> 4 0.0 9. ~t
o t,o
T. to
.-: eo
.~. eo
~ eo
T. o
.~. o
T. e4
.~. r~
C
~2 .
Fatty acid
.
.
.
.
.
.
.
.~.
..
Fig. 2. Relca~,c o l tany acidr, during initial hydrolysis of capclin oil by human and cod BSI)L. Rclalivc activity means fraction of released fatty acid species divktcd with fraction o f thai laity acid species contained in Ihe 1.3-position of capclin oil. The incubations were done al 18"(" fi)r lil rain using 54 ~ug c(~ BSDL and 38 /.tg human BSI)I_ (A} Substrale added in alcoholic solution. (ID Subslralc dispersed by ~mication.
O ¢..) (.3 ¢D U ¢.~ Folly acid Fig. 3. Re ease (ff fatty acids during initial hydrolysis of capelin oil by human and cod B S D L at different lcmpcrai.urc~.. For dclails, see legend to Fig. 2. Substrate was added as alcoholic solution. ( A ) Incubated at It]PC for ")O rain. (B) Incubated al 18°C for I 0 rain (identical to Fig. 2A}. (C) Incubated at 37°C for 5 rain.
man enzyme showed higher activity towards mediumchain fatty aeyl esters (myristic and palmitic acid). while the cod enzyme generally was more active towards esters of h)ngcr polyunsaturated fatty acids. Both enzymes hydrolyscd the fatty acyl esters of the 22carbon series at significantly slower rates. Since the two assays yielded comparable results, the assay using substrate added in alcoholic solution was used for further investigations. The effect of incubation temperature on relative hydrolysis rates is shown in Fig. 3. Assays were incubated at low (10°( ". panel A). medium (18°C. panel B) and high (37°C. panel C) temperature. The differences in relative fatty acyl ester hydrolysis of the two enzymes wcrc similar to that seen in Fig. 2. but the various incubation temperatures resulted in markedly different hydrolysis rate patterns for both enzymes. This effect was seen as an increased liberation of medium-chain fatty acids with increasing incubatilm temperature. while the hydrolysis of polyunsaturated fatty acyl esters (18:4 and 20:5) declined with temperature. When bile salts obtained from natural cod bile was used in place of pure Na-taurocholate (Fig. 4). the results were similar to that seen in Fig. 2. though the differences between the two enzymes became more
However, the activities of the two enzymes were measured on emulsified capelin oil titrimetrically. The results showed that the cod and the human enzyme behaved very differently towards this substrate, in the presence of 18.8 mM Na-TC, the specific activity of the cod enzyme was 17.1 + 1.2 U / m g , while specific activity of the human enzyme was 4.1 _+ 0.8 U / m g . However, in the presence of 9% natural cod bile, the specific activity of the cod enzyme was increased to 63.7 __ 8.4 U/mg, while specific activity of the human enzyme was very low (i.2 U / m g ) . These results indicate that the largest difference between the cod and human enzyme is in substratephase specificity, and the observed differences in specific activity in the assays
pronounced.
Fatty acid specificity studies using gum arabicemulsified capelin oil was not possible because gum arabic entered the chloroform phase during extraction.
¢D ¢D ¢D t.J
2.0
~
>= ~.5
!
i.o
i
2~o. 0.0 u
~
u
u
u
~)
u
u
u
~
Fatty acid Fig. 4. Release of fatty acids during initial hydrolysis of capelin oil by human and cod BSDL with bile salts from natural c(~J bile. Experimental conditions were identical Io Ihosc d e ~ r i b e d in legend to Fig. 2A.
171 used in this work may be ascribed to different substratc phases. Discussion
The objective of this work was to establish putative fatty acid specificities of bile salt-dependent lipases. It is recognized that the term 'fatty acid specificity' consists of at least two components: One component originates from enzymatic properties (enzymatic specificity -classical substrate fit in active site) and the other component is a result of physicochemical substrate properties (substrate orientation and ester availability at the interface), also called supersubstrate, in order to separate these components experimentally, two evolutionally distant BSDLs (liable to differ in fatty acid specificity) were used under different conditions of substrate presentation. The catalytic characteristics of the two enzymes in study are qualitatively similar, but with ~ m e quantitative differences (Gjellesvik et al., submitted). Shortchain fatty acid 4-nitrophenyl esters (especially soluble esters) were hydrolysed more effectively by the human enzyme, but the specific activities toward long-chain fatty acid esters were similar. Hydrolysis of methyl esters revealed the importance of the substrate phase for fatty acid specificities of the two enzymes, since the hydrolysis rates of certain esters were greatly influenced when the esters were in a dispersed or mixed micellar state. This was especially true for the 20:5 (EPA) ester, which was preferred as snbstrate for both enzymes in the dispersed state. However, enzyme-specific effects were seen, as the 18:2, 22:4 and 22:6 fatty acid esters were more effectively hydrolysed by the cod enzyme in the mixed miceilar state, while the activities for the human cnzyme remained unchanged. The cod BSDL has been shown to be 1,3-spccific for hydrolysis on triolein [14], (Gjellesvik et al., submitted). Therefore the present hydrolysis rates for individual fatty acyl esters are normalized according to tbeir abundance in the 1,3-position. However, mammalian BSDLs have been reported to hydrolyse also the 2-ester bond in triglycerides [1,3,16]. On the other hand, studies on sequential hydrolysis of triolein have shown that monoglyceride hydrolysis of the human milk BSDL (which is identical to the pancreatic enzyme [ 14,17,18]) is rate-limited by acyl migration [5]. Also the human pancreatic BSDL has been found to have low activity towards monoolein, even after long (2 h) incubation [191. There were small differences on the fatty acid specificity patterns if capelin oil was added in alcoholic solution or dispersed by sonication. This is probably due to similar physicochemieal states of the substrate, which gave turbid .solutions in both cases, in both
cases, the cod enzyme showeti a slightly higher activity towards fatty acid esters containing more than two double bonds, while the human enzyme wa.~ more active on the fatty acids of shortest length. An increase in temperature is likely to alter tbe substrate state, (or binding of enzyme to the interface) more than altering an active site-fit of substrate. The effect of temperature does not alter the differences in specificity between the two enzymes, hut I~th enzymes show a generally increased activity t t ~ t r d s esters of medium-chain fatty acids (myristic and palmitic) at the expense of long-chain fatty atvI esters (especially if polyunsaturated). This indicates that temperature affects the substrate, and not specificity of the enzymes. Cod bile consists of primarily taurine-conjugated bi!e salts of di- and tri-hydroxy cholanic acid in a I :5 ratio, [20,21]. Since fatty acid specificity may be dependent on substrate surface structure, it was of interest It) investigate the effect of natural cod bile on fatty acid specificity. The results showed that natural cod bile did not alter the pattern of fatty acid specificity, but the differences between the cod and human enzymes became more apparent. Noteworthy in this context is the ability of natural cod bile to enhance the activity of the c t ~ BSDL towards emulsified capelin oil. At identical conditions, the activity of the human enzyme was reduced, while the activities in the prescacc of taurocholatc were comparable to earlier findings [3]. It should be noted that the human enzyme displayed normal activity when the substratc was added as alcoholic solution. In this case, the substrate concentration was only IV, (d the substrate concentration in the emulsified assay, and the .~)lution was less turbid (partly .,a)lubilizcd) than when Na-'r(" was used. These observations may bc explained by the differcnt physiological roles of the cod and human enzymes. While the major digestivc lipase in humans is pancreatic lipase, this enzyme is absent in the cod (Gjcllcsvik c t a l . , submitted). Thus, the human enzyme can operate in a mixed micellar p h a ~ of partial glyccridcs (generated by pancreatic lipa.~), cholesterol- and rctinol-cstcrs and bile salts [12,19,22,23]. On the other hand, the cod enzyme must bc able lt) hydndy.~ emulsified lipids in the chymc effectively since pancreatic lipase is absent. Natural cod bile .seemed to facilitate hydrolysis of emulgated triglycerides by cod BSDL. It is concluded that the two enzymes s h t ~ mimer, but significant differences in fatty acid specificity that persisted in all assays used. The human enzyme was more effective in hydrolysis of shorter chain fatty acyl esters, while the cod enzyme was more active towards esters of polyunsaturated fatty acids. However, the largest effects were seen on both enzymes wben the substrate phase was altered. It is thus possible that the observed specificities mainly result from the confl~rmation (or exposure) of the carboxyl estcr bond at the
172 interface (or in the micellar phase), in spite of this, the two enzymes respond differently to these structures. ttenee, though aliphatic chains of the fatty acids may not be recognized by the enzymes, these moieties may influence the conformation (or availability) of the ester bond, which in turn may determine the enzyme affinities for particular fatty acids. This indirect effect defines a type of "non-classiear substrate specificity. Acknowledgements This work was financed in part by the Norwegian Fisheries Research Council (NFFR). I am indebted to Dr. D. Lombardo, INSERM U260, Marseillc. fl~r the supply of purified human B S D L and for discussions. References I [~rlan~m. ('. (1975) Stand. J. (;a~.lr(~:nlcrol. I(I. 4IH-408. 2 h~mbardo. I). and (;u). O. (19811) B,g'him Biophys. Acta bl I. 14% =55. 3 Lombardo. D . Fauvcl. J. and Guy. O. (198.) Biochim. Biophy, Acta 611. I.~-14h. 4 Guy. O. and Figarcna. ('. (1981) Scand. J GaslroenteroL I¢) (Suppl. 67). 50-61. 5 Wang. C.S.. llarl~;uck. J.A. and Do~'ns, D. ( 1 9 ~ ) Biochemislr~ 27. 4834-484O.
6 D.)nino. N.. Vandenburg. G.A. and Reiscr. R. (1967) Lipids 2. 489-403. 7 Brockcrhoff. |1. ~I970) Biochim Biophys. Acla 212. 02-101. 8 Ileimermann. W.II.. Ilolman. R T . , Gordon. D.T.. K~n~alyshyn. I).F,. and J c n ~ n . R.G. (1973) Lipids 8. 45-47. 9 ('hen. O.. Stcrnb). B. and Nilsson. /4. (1989) Biochim Biophy~. Acla I|NH. 372 385. I(! Chcn. O.. Sternby. B.. ,~kes~m. B. and Nilsson. ,~. (Iq~)) Biochim. Biophys. Acla 1044. I I I - ( 17. II Bna:kerhoff. It. and Jenscn. R.G. (1974) Liradylic Enzymes. Academk' Pres~, New York. 12 Rudd. I-.A. and Brockman I t . k ( 1985l in Lipases (Borgstr6m, B. and B,K:kman. iI.L,, eds.L pp. 185-~)4, Elsevier. Amsterdam. 13 Panon, J.S., Warner, T.G. anti Benson. A.A. (1977) Biochim. Biophys. Acla 486. 322-3.~). 14 Gjellesvik. D.R.. R a a e . A J . and Walther, B.T. (1989) Aquaculture 79, 177-184. 15 Kennerly. D.A. (1987) J. Biol. ('hem. 262. 16.~)5-16313. 16 Wang. ('.S., Kuksis. A.. Manganaro. F., Myher, J J . . Downs. D. and Bass. H B . (1983)J. Biol. ('hem. 258. 9197-9~)2. 17 Ilui. D.Y. and Kissel. J.A. (19~)) FEBS Lett. 276. 131-134. IS Reue. K.. Zambaux. J.. Wong. H.. Lee. G.. Leete. T.H.. R o n L T.. Shively. R., Slernby, B.. Borgslr6m. B.. Ameis. D. and Schotz. M.('. (1991)J. Lipid Res. 32. 267-276. 19 I ;ndstriim. M.. Slernby. B. and Borgstr6m. B. ((088) B/ochim. IL ~phys. Acta 959. 178-184. 20 H..dewla~l. G.A.D. and Sj/~valL I. (1054) Biochem. J. 57. 1.?.6-!~). 21 Ka~lner. A. ((W~S)Aeta ('hem. Scand. 22. ~ 6 1 - 2 3 7 0 . 22 Patton. LS. and Carey. M.C. (It;70) Science 204, 145-148. 23 Wang. C.S. and Lee. D.M. (1985) J. Lipid Res. 26. 824-8.~L
