- Email: [email protected]
Human milk bile salt-stimulated lipase: Functional and molecular aspects
Human milk bile salt-stimulated lipase: Functional and molecular aspects Olle Hernell, PhD, MD, and Lars Bl~ckberg, PhD From the Department of Pediatrics and the Department of Medical Biochemistry and Biophysics, University of Ume~, Ume6, Sweden In breast-fed infants, digestion of milk triglycerides, the major source of energy and long-chain polyunsaturated fatty acids, is catalyzed by a concerted action of gastric lipase, colipase-dependent pancreatic lipase, and bile salt-stimulated lipase (BSSL). The major part of BSSLis present in the milk and the lesser part originates in the infant's exocrine pancreas. Gastric lipase is important in initiating digestion of milk fat globule triglycerides in the stomach. BSSLshifts the final products of triglyceride digestion from monoglyceride and free fatty acid (the products of colipase-dependent pancreatic lipase) to glycerol and free fatty acid, which may promote efficient absorption. Moreover, BSSL is likely to promote efficient use of milk cholesteryl- and fat-soluble vitaminesters and long-chain polyunsaturated fatty acids (>C18). The cDNA sequence has shown that BSSLhas a unique primary structure. The N-terminal half is highly conserved between species and shows striking homology to typical esterases, for example, acetylcholine esterase. In contrast, the C-terminal half, containing 16 proline-rich repeats of 11 amino acid residues, is unique to BSSL. Using several recombinant variants of BSSL, we have found that these unique repeats and the glycosylation are completely dispensable for activity. Thus all typical properties of BSSL reside in the N-terminal half of the molecule. (J PEDIATR 1994;125:$56-61)
Milk triglycerides constitute approximately 50% of the dietary energy substrates in breast-fed infants. However, the postprandial intraluminal levels of colipase-dependent pancreatic lipase and bile salt, two essential components in triglyceride digestion in adults, are low in newborn infants in general and in preterm infants in particular, and may explain the common occurrence of fat malabsorption in formula-fed neonates. In contrast, human milk triglycerides are efficiently used, even in preterm infants. A major reason for this is the bile salt-stimulated lipase in the milk,
Supported by the Swedish Medical Research Council (19X05708), Astra H~ssle AB, and the Medical Faculty, University of Umeh. Reprint requests: Olle Hernell, PhD, MD, Department of Pediatrics, University of Ume/t, S-901 85 Ume~, Sweden. Copyright © 1994 by Mosby-Year Book, Inc. 0022-3476/94/$3.00 + 0 9/0/59298
$56
which is present in the milk of only a limited number of species (e.g., man, gorilla, dog, cat, ferret, and seal) and is absent in others (e.g., cow, pig, rat, and rhesus monkey). 1 The reason for these differences is not known. However, BSSL has a counterpart in exocrine pancreas, carboxyl ester lipase (detailed comparison to follow), which in some species, for example, leopard shark 2 and cod, 3 is the only triglyceride-hydrolyzing enzyme secreted from the
I
BSSL CEL
Bile salt-stimulated lipase Carboxyl ester lipase
I
pancreas. In the suckling rat CEL is the major pancreatic lipolytic enzyme, and it is only after weaning that colipasedependent lipase becomes the dominating enzyme. 4 This finding emphasizes that BSSL/CEL, whether it is secreted from the pancreas or borne by milk, may be of particular importance in the neonatal period. In addition to BSSL,
The Journal of Pediatrics Volume 125, Number 5, Part 2
Hernell and Bl~ckberg
S 57
Table I. Specific activity of BSSL against different substrates
Substrate
Activity (#mol product • min -1 . mg protein -1)
Trioleate 100 sn-1 (3) monooieate 21 sn-2 monooleate 16 Cholesteryl ester 0.5 Retinyl palmitate 6-8 p-Nitrophenyl acetate 30 Dioctanoyl-phospatidyl choline 80 Modifiedfrom B1/iekbergL., HernellO. FEBS Lett 1983;157:337-41.Additionaldata from Bl~ickberget al. (unpublisheddata, 1983).
gastric lipase, which is secreted by the chief cells of the gastric mucosa, 5 is considered to be of particular importance in the neonatal period. 1 FUNCTIONAL
PROPERTIES
OF BSSL
Substrate and positional specificity. Bile salt-stimulated lipase can hydrolyze a variety of ester substrates (Table I). Whether the substrate is in monomericl micellar, or emulsified form has little influence on the specific activity. It has been shown that BSSL also hydrolyzes certain amide bonds, 6 the physiologic significance of which is not presently understood. From a physiologic viewpoint, several substrates are likely to be attacked by BSSL, for example, triglycerides and partial glycerides and cholesteryl and retinyl esters. In addition to a relative lack of substrate specificity, BSSL has no obvious positional specificity when hydrolyzing a triglyceride. Unlike colipase-dependent pancreatic lipase, BSSL can hydrolyze all three ester bonds; as a result, free fatty acids and free glycerol are the end products of lipolysis7 rather than free fatty acids and sn-2 monoglyceride, which are the typical products of colipasedependent lipase. Wang et al. 8 compared the rates at which triglyceride, diglyceride, and monoglyceride were hydrolyzed. Diglyceride was the preferred substrate, and monoglyceride was hydrolyzed at the lowest rate. Bile salt activation. A characteristic feature of BSSL is its interaction with bile salts, which are required for BSSL activity with a long-chain triglyceride substrate. 9 Therefore, although BSSL is present in high amounts in human milk, it will not act on the milk fat before it reaches the duodenum, where it encounters bile salts. Neither the structural basis for bile salt interaction nor the exact mechanism for activation is completely understood. The activation shows a high degree of specificity. Only primary bile salts activate it, although both primary (e.g., cholate and
(
Activation site
Cholate
Lipid-bindingpromoti nsite g
~
Deoxycholate
Cholate
Fig. 4. Model presentation of the bile salt interaction of bile saltstimulated lipase, including two bile salt-binding sites. Tentative ligands and amino acid residues suggested to participate in the binding are indicated. Y, Tyrosine; R, arginine.
chenodeoxycholate) and secondary (e.g., deoxycholate) bile salts bind to the enzyme. 9 This suggests that the 7ahydroxyl group in primary bile salts is essential for BSSL activation. It has been suggested that primary bile salts activate the enzyme merely by inducing binding to a substrate interface, for example, an emulsion particle. 1° This suggestion is supported by the fact that BSSL can act on watersoluble ester substrates, for example, p-nitrophenylacetate, also in the absence of bile salts. With the use of radiolabeled BSSL we could show, in agreement with Wang et al., that BSSL was unable to bind to a lipid emulsion in the absence of bile salts. 11 Adding cholate to the system induced binding. However, efficient binding was also seen with deoxycholate, but without resulting lipolysis. Hence, promoting lipid binding is only part of the activation mechanism. From these data and other comparisons of cholate, deoxycholate, and cholamidopropyl dimethylammoniopropane sulfate, a cholate analogue with identical backbon e hydroxylation but with a different side chain, we have proposed a model for the activation that is based on two bile salt-binding sites on the enzyme molecule (Fig. 1).11 A nonspecific binding site (the lipid-binding-promoting site) is involved in binding to the lipid-water interface. Binding to a second site (the activation site) that is specific for bile salts or analogues that contain the 7c~-hydroxyl group is also required for activity. Experiments with amino acid-modifying reagents have provided evidence that tyrosyl and arginyl residues are involved in the bile salt interaction. 11-13 In addition to the effects on activity, bile salts also protect the enzyme from degradation and inactivation by pancreatic proteases, for example, trypsin. Unlike activation, this protective effect is seen with primary and secondary bile salts and cholamidopropyl dimethylammoniopropane sulfate,11, 14 Role in the digestion of human milk triglyeeride. Digestion of milk triglyceride is a process in which all three gastrointestinal lipases--gastric lipase, colipase-dependent
S 58
Hernell and Bl?tckberg
The Journal of Pediatrics November 1994
Table II. Characteristics of B S S L / C E L from different species Species
Man
Source
Milk, pancreas Cow Pancreas Rat Pancreas Rabbit Pancreas Salmon Pancreas Data from references21 and 28-31.
No. of residues
No. of repeats
722
16
578 592 556 540
3 4 2 or 3 0
bonds with long-chain polyunsaturated fatty acids, for example, arachidonic acid and eicosapentaenoic acid.18' 20 On the other hand, BSSL does not discriminate between these and other fatty acids that are present in the triglyceride substrate.18, 19 The most efficient release of long-chain polyunsaturated fatty acids was achieved when the two lipases were combined. From these in vitro experiments it is tempting to speculate that one role of BSSL is to ensure adequate levels of long-chain polyunsaturated fatty acids in breast-fed infants. STRUCTURAL
pancreatic lipase, and B S S L - - p l a y important roles. From in vitro studies using purified enzymes and cofactors, and using human milk fat globules as substrate under conditions mimicking those of the postprandial gastric and duodenal contents, respectively, we have put forward a model for how their interaction ensures efficient digestion and subsequent product absorption.15 Human milk fat globule triglyceride is not digested efficiently by either colipase-dependent lipase, even in the presence of its cofactor colipase, or by BSSL in the presence of bile salts. However, if the globules are first digested to only a limited degree by gastric lipase, the resulting droplets are efficiently hydrolyzed by colipasedependent lipase, yielding free fatty acids and sn-2 monoglycerides. If BSSL is included, hydrolysis nears completion (>90% of the ester bonds are hydrotyzed), producing free fatty acids and free glycerol as the end productsJ 5 Under conditions with low intraluminal bile salt concentrations, as is often the case in the neonatal period, such a product pattern may be favorable for absorption. Fat absorption studies in preterm infants have shown that pasteurization of human milk, which inactivates BSSL, reduces the coefficient of fat absorption by as much as one third) 6' 17 Role in the use of long-chain polyunsaturated fatty acids. In addition to its role in overall triglyceride digestion, it has been suggested that BSSL plays a role in the use of longchain polyunsaturated fatty acidsJ 8' 19 These fatty acids, most of which are supplied by the milk triglycerides, are essential not only as structural components, for example, in the brain and the retina, but also as precursors for eicosanoids. In adults these fatty acids are synthesized from precursor fatty acids of the n-6 and n-3 series, linoleic and ~-linolenic acid, respectively, by desaturation and elongation reactions. The capacity for these reactions is not fully developed at birth, however, and thus not only the precursor fatty acids but also the desaturated and elongated derivatives should be considered essential dietary components in the neonatal period. In vitro, colipase-dependent lipase has been shown to have restricted activity against ester
ASPECTS
OF BSSL
Structure of BSSL and its relationship to pancreatic CEL. We have cloned and sequenced eDNA covering the entire coding sequence of human-milk BSSL. 21 The sequence has been confirmed. 22 The deduced amino acid sequence is identical to that of CEL, also known as carboxyl ester hydrolase and cholesterol esterase, which is secreted by the human pancreas. 23 Earlier studies suggested that these two enzymes are one functional entity. 24 When the genomic structure became available it was clear that they are products of the same gene. 25 The previously noted difference in the apparent molecular size24 is, therefore, most likely the result of differences in the degree or type of tissue-specific glycosylation. The peptide chain contains one potential N-linked glycosylation site, located seven amino acid residues to the N-terminal of the tentative active site serine. 21 This site is glycosylated in both the milk and the pancreatic enzyme. In the pancreatic enzyme, the carbohydrate chain is known to be of the biantennary complex type. 26 Both enzymes are also heavily O-glycosylated at the C-terminal end of the molecule. This region of the polypeptide chain is characterized by a unique structure that consists of 16 near-identical proline-rich repeats of 11 amino acid residues. 21 Each repeat contains at least one potential O-glycosylation site. For the pancreatic enzyme the O-linked sugars are heterogeneous. 27 We have found pronounced differences between the milk and the pancreatic enzymes in composition of O-linked sugars (Bl~ickberg et al., unpublished data, 1993), which is a likely explanation for the difference in molecular size. The pancreatic enzyme has been cloned and sequenced in several other species, that is, rat, cow, rabbit, and salmon. 28"31 The main difference between species is in the number and structure of the C-terminal repeats; more than 70% homology exists in the N-terminal region preceding the repeats. With the exception of the salmon enzyme, all enzymes contain at least two repeats (Table II). A variant of human milk BSSL was recently reported that is smaller than the common form. 32 It is tempting to speculate that this shorter form contains fewer repeats and thus is more
The Journal of Pediatrics Volume 125, Number 5, Part 2
Hernell and Bl~ckberg
S I
-
D I
S 59
H I
~-~A A A A A A A A A A A A A A A A ~C
Fig. 2. Schematic presentatio n of human milk bile salt-stimulated lipase. Solid boxes indicate regions of homology to choline esterase. Positions of amino acid residues of the catalytic triad are indicated. Open box represents the region with the proline-rieh repeats. Positions of the disulfide bridges (l__t)and glycosylation sites (A) are indicated. S, Serine; D, aspartate; H, histidine.
similar to the forms found in other species. All species sequenced contain the same N-linked glycosylation site, which suggests that this site is of functional importance. The bovine enzyme contains an additional N-glycosylation site; whether this site is glycosylated is not known. Solely on the basis of sequence homologies, very few similarities between B S S L / C E L and other well-characterized mammalian lipases are found. All mammalian lipases contain the consensus sequence around the active-site serine: G-X-S-X-G. There are, however, homologies when BSSL/ CEL is compared with some lipases of microbial origin, for example, Geotrichum candidum lipase.33 These homologies are restricted to the N-terminal part of BSSL/CEL; none of the microbial enzymes contains any structure that resembles the C-terminal repeats of the human enzyme. Striking homologies in the N-terminal part are found with some typical esterases, for example, acetylcholine esterase from different species and rat liver carboxyl esterase. 21 The locations of these homologies are depicted in Fig. 2. Structure/function relationships in BSSL. An intriguing question is the functional importance of the unique C-terminal repeats. To address this question, we, in collaboration with L. Hansson and M. Str6mqvist (Umefi, Sweden) and L. Lundberg (M(51ndal, Sweden), constructed and expressed a set of BSSL variants by site-directed mutagenesis. 34 By using a bovine papilloma virus-based vector and C- 127 cells, variants were produced that lacked all or all but two of the repeats but contained the tail of 11 residues C-terminal of the last repeat. Both variants, as well as a full-length recombinant enzyme, were produced and secreted into the culture medium. The different enzyme proteins were purified, essentially by heparin-sepharose and immunoaffinity chromatographic methods, and characterized. All variants were active and had approximately the same specific activity when a long-chain triglyceride substrate was used. Furthermore, they all had the same bile salt dependency with regard to both the type and the concentration required for maximal activity. 34 It can thus be concluded that the C-terminal proline-rich, O-glycosylated repeats are essential for neither the catalytic activity of the enzyme nor the bile salt interaction, the lipid binding, or the heparin binding. It has been suggested that the sequence of
the repeats is similar to that of PEST (proline, glutamine, serine, threonine) sequences, which is, supposedly, a signal for proteolytic degradation and a fast turnover. 3° Such a signal is, however, highly unlikely for an enzyme that acts in the milieu of the intestinal contents. On the contrary, BSSL has been shown to be very resistant to inactivation and degradation by several proteases. Thus, why evolution has added these unique proline-rich, O-glycosylated repeats to a typical esterase is an intriguing but unanswered question. Using the same strategy as that described previously, we also investigated the importance of the N-linked sugar chain for catalytic activity.34 A BSSL variant in which the asparagine residue (asn-187) had been altered to a glutamine residue was equally active as the native enzyme. The conclusion that glycosylation, N- or O-linked, is not important for activity or bile salt activation was supported by the fact that we were able to produce active recombinant enzyme in Escherichia coli with the same bile salt dependency as the native milk enzyme. 34 Sequence homologies to other lipases/esterases strongly indicate that B S S L / C E L is a member of the family of a/fl_hydrolases,31, 35 with antiparallel/3 sheets making up the interior of the N-terminal part of the molecule. A catalytic triad typical of esterases/lipases has been suggested from homologies to other lipases/esterases to include ser-194, asp-320, and his-446. 35 The position of the essential serine and aspartate residues has been confirmed by site-directed mutagenesis. 22, 36 Interfacial activation. Some lipases, including pancreatic colipase-dependent lipase, contain a "lid structure. ''37 This short stretch of amino acid residues forms a loop structure that covers the active-site groove. For pancreatic lipase, x-ray crystall0graphy of a complex of lipase and its cofactor protein (colipase), formed in the presence of mixed micelles, has shown that this lid undergoes a conformational change when the lipase interacts with a lipid surface138 This change makes the active-site groove accessible to substrate molecules. Hence, it can explain the interfacial activation typical for colipase-dependent lipase. B S S L / C E L does not have the same interfacial activation, and it is therefore not surprising that no obvious lid-structure sequence can be
S60
Hernel[ and Bli~ckberg
found, The relative lack of substrate specificity of BSSL/ CEL indicates that the active site is easily accessible and supports the view that this enzyme lacks a lid. Moreover, the absence of a lid structure rules out the possibility that the bile salt-activation mechanism involves a conformational change in the lipase that is similar to that observed when colipase-dependent lipase interacts with a substrate interface. The exact residues involved in the interaction with bile salts have not been pinpointed. However, alignments of sequences of similar enzymes (e.g., acetylcholine esterases) that are unaffected by bile salts provide evidence for the tentative arginine and tyrosine residues involved (Fig. 1): CONCLUSION H u m a n milk BSSL is a lipolytic enzyme with several functions ascribed to it. Not only is BSSL important for overall triglyceride use in the breast-fed infant, but it has also been suggested to have specific roles in the use of milk long-chain polyunsaturated fatty acids, cholesterol, and fat-soluble vitamins. On a structural level BSSL belongs to the family of a/~-hydrolases in which many lipases and esterases a r e found. It is, however, very different from Other well-characterized mammalian lipases, with the exception of pancreatic CEL, which is a product of the same gene. From an evolutionary standpoint, the N-terminal part appears to be a product of an ancestral gene that gives rise to several lipases and esterases. The function of a unique repetitive sequence that has been added to this gene remains to be elucidated.
The Journal of Pediatrics November 1994
9.
10.
11.
12.
13.
14.
15:
16.
17.
18. REFERENCES
1. Hernell O, B1/ickberg L. Digestion and absorption of human milk lipids. In: Dulbecco R, ed. Encyclopedia of human biology. New York: Academic Press, 1991;3:47-56. 2. Patton JS, Warner TG, Benson AA. Partial characterization of the bile salt-dependent triacylgl~ceerollipase from the leopard shark pancreas. Biochim Biophys Acta 1977;486:32230. 3. GjeUsevik DR, Lombardo D, Walther BT. Pancreatic bile salt dependent lipase from cod (Gadus morhua): purification and properties. Biochim Biophys Acta 1992;1124:123-34. 4. Bradshaw WS, Rutter WJ. Multiple pancreatic lipases: tissue distribution and pattern of accumulation during embryological development. Biochemistry 1972;1 !:1517-28. 5. Moreau H; Laugier R, Gargouri Y, Ferrato F, Verger R. Human preduodenal lipase is entirely of gastric fundic origin. Gastroenterology 1988;95:1221-6. 6. Hui DY, Hayakawa K, Oizumi J. Lipoamidase activity in normal and mutagenized pancreatic cholesterol esterase (bile salt-stimulated lipase). Biochem J 1993;291:65-9. 7. Hernell O, B1/ickberg L. Digestion of human milk lipids: physiologic significance of sn-2 monoacy!glycerol hydro!ysis by bile salt-stimulated lipase. Pediatr Res 1982;16:882-5. 8. Wang C-S, Hartsuck JA, Downs D. Kinetics of acylglycerol
19.
20.
21.
22.
23.
24.
sequential hydrolysis by human milk bile salt activated lipase and effect of taurocholate as fatty acid receptor. Biochemistry 1988;27:4834-40. Hernell O, Olivecrona T. Human milk lipases, II: Bile salt-stimulated lipase. Biochim Biophys Acta 1974;369:23444. Wang C-S, Lee DM. Kinetic properties of human milk bile salt-activated lipase: studies using long chain triacylglycerol as substrate. J Lipid Res 1985;26:824-30. Blhckberg L, Hernell O. Bile salt-stimulated lipase in human milk: evidence that bile salts induce lipid binding and activation via binding to different sites. FEBS Lett 1993;323:207-10. Lombardo D, Campese D, Multigner L, Lafont H, De Caro A. On the probable involvement of arginine residues in the bilesalt-binding site of human pancreatic carboxylic ester hydrolase. Eur J Biochem 1983;133:327-33. Campese D, Lombardo D, Multigner L, Lafont H, De Caro A. Implication of a tyrosine residue in the unspecific bile salt binding site of human pancreatic carboxylic ester hydrolase. Biochim Biophys Acta 1984;784:147-57. B1/ickberg L, Hernell O. Further characterization of the bile salt-stimulated lipase in human milk. FEBS Lett 1983; 157:33741. Bernb~ick S, Bl~ickberg L, Hernell O. The complete digestion of human milk triacylglycerol in vitro requires gastric lipase, pancreatic colipase-dependent lipase, and bile salt-stimulated lipase. J Clin Invest 1990;85:1221-6. Williamson S, Finucane E, Ellis H, Gamsu HR. Effect of heat treatment of human milk on absorption of nitrogen, fat, sodium, calcium, and phosphorus by preterm infants. Arch Dis Child 1978;53:555-63. Atkinson SA, Bryan MH, Anderson GH. Human milk feeding in premature infants: protein, fat, and carbohydrate balances in the first two weeks of life. J PEDIATR1981;99:61724. Hernell O, Bl~ickbergL, Chen Q, Sternby B, Nilsson ~,. Does the bile salt-stimulated lipase of human milk have a role in the use of the milk long-chain polyunsaturated fatty acids? J Pediatr Gastroenterol Nutr 1993;16:426-31. Chen Q, Bl~ickberg L, Nilsson ~., Sternby B, Hernell O. Digestion of triacylglycerols containing long-chain polyenoic fatty acids in vitro by colipase-dependent pancreatic lipase and human milk bile salt-stimulated lipase. Biochim Biophys Acta 1994;1210:239-43. Chert Q, Sternby B, Akesson B, Nilsson ,~. Effects of human pancreatic lipase-colipase and carboxyl ester lipase on eicosapentaenoic and arachidonic acid ester bonds of triacylglycerols rich in fish oil fatty acids. Bi0chim Biophys Acta 1990; 1044:111-7. Nilsson J, B1/ickberg L, Carlsson P, EnerNick S, Hernell O, Bjursell G. eDNA cloning of human-milk bile-salt-stimulated lipase and evidence for its identity to pancreatic carboxylic ester hydrolase. Eur J Bioehem 1990;192:543-50. Baba T, Downs D, Jackson KW, Tang J, Wang CS. Structure of human milk bile salt activated lipase. Biochemistry 1991; 30:500-10. Reue K, Zambaux J, Wong H, et al. eDNA clon!ngof carboxyl ester lipase from human pancreas reveals a unique proline-rich repeat unit. J Lipid Res 199l;32:267-76. Bl~ickberg L, Lombardo D, Hernell O, Guy O, Olivecrona T. Bile salt-stimulated lipase in human milk and Carboxyl ester
The Journal of Pediatrics Volume 125, Number 5, Part 2
25.
26.
27.
28.
29.
30.
31.
hydrolase in pancreatic juice: are they identical enzymes? FEBS Lett 1981;136:284-8. Lidberg U, Nilsson J, Str6mberg K, et al. Genomic organization, sequence analysis, and chromosomal localization of the human carboxyl ester lipase (CEL) gene and a CEL-like (CELL) gene. Genomics 1992;13:630-40. Sugo T, Mas E, Abouakil N, et al. The structure of N-linked oligosaccharides of human pancreatic bile-salt-dependent lipase. Eur J Biochem 1993;216:799-805. Mas E, Abouakil N, Roudani S, Franc J-L, Motreuil J, Lombardo D. Variation of the glycosylation of human pancreatic bile-salt-dependent lipase. Eur J Biochem 1993;216:80712. Han JH, Stratowa C, Rutter WJ. Isolation of full-length putative rat lysophospholipase cDNA using improved methods for mRNA isolation and cDNA cloning. Biochemistry 1987; 26:1617-25. Kyger EM, Wiegand RC, Lange LG. Cloning of the bovine pancreatic cholesterol esterase/lysophospholipase. Biochem Biophys Res Commun 1989;164:1302-9. Colwell NS, Aleman-Gomez JA, Kumar BV. Molecular cloning and expression of rabbit pancreatic cholesterol esterase. Biochim Biophys Acta 1993;1172:175-80. Gjellesvik DR, Lorens JB, Male R. Computer-modelled tertiary structure of carboxylester lipase from Atlantic salmon
Hernell and Bl~tckberg
32.
33.
34.
35.
36.
37. 38.
S61
(Salmo salar). In: Lipases: structure, function and protein engineering. Elsinore, Denmark: 1993:47. Swan JS, Hoffman MM, Lord MK, Poechmann JL. Two forms of human milk bile-salt-stimulated lipase. Biochem J 1992; 283:119-22. Schrag JD, Li Y, Wu S, Cygler M. Ser-His-Glu triad forms the catalytic site of the lipase from Geotrichum candidum. Nature 1991;351:761-4. Hansson L, Blfickberg L, Edlund M, Lundberg L, Str6mqvist M, Hernell O. Recombinant human milk bile salt-stimulated lipase: catalytic activity is retained in the absence of glycosylation and the unique proline-rich repeats. J Biol Chem 1993; 268:26692-8. Wang C-S, Hartsuck JA. Bile salt-activated lipase: a multiple function lipolytic enzyme. Biochim Biophys Acta 1993; 1166:119. DiPersio LP, Hui DY. Aspartic acid 320 is required for optimal activity of rat pancreatic cholesterol esterase. J Biol Chem 1993;268:300-4. Winkler FK, D'Arcy A, Hunziker W. Structure of human pancreatic lipase. Nature 1990;343:771-4. van Tilbeurgh H, Egloff M-P, Martinez C, Rugani N, Verger R, Cambillau C. Interracial activation of the lipase-procolipase complex by mixed micelles revealed by x-ray crystallography. Nature 1993;362:814-20.
Milk triglycerides constitute approximately 50% of the dietary energy substrates in breast-fed infants. However, the postprandial intraluminal levels of colipase-dependent pancreatic lipase and bile salt, two essential components in triglyceride digestion in adults, are low in newborn infants in general and in preterm infants in particular, and may explain the common occurrence of fat malabsorption in formula-fed neonates. In contrast, human milk triglycerides are efficiently used, even in preterm infants. A major reason for this is the bile salt-stimulated lipase in the milk,
Supported by the Swedish Medical Research Council (19X05708), Astra H~ssle AB, and the Medical Faculty, University of Umeh. Reprint requests: Olle Hernell, PhD, MD, Department of Pediatrics, University of Ume/t, S-901 85 Ume~, Sweden. Copyright © 1994 by Mosby-Year Book, Inc. 0022-3476/94/$3.00 + 0 9/0/59298
$56
which is present in the milk of only a limited number of species (e.g., man, gorilla, dog, cat, ferret, and seal) and is absent in others (e.g., cow, pig, rat, and rhesus monkey). 1 The reason for these differences is not known. However, BSSL has a counterpart in exocrine pancreas, carboxyl ester lipase (detailed comparison to follow), which in some species, for example, leopard shark 2 and cod, 3 is the only triglyceride-hydrolyzing enzyme secreted from the
I
BSSL CEL
Bile salt-stimulated lipase Carboxyl ester lipase
I
pancreas. In the suckling rat CEL is the major pancreatic lipolytic enzyme, and it is only after weaning that colipasedependent lipase becomes the dominating enzyme. 4 This finding emphasizes that BSSL/CEL, whether it is secreted from the pancreas or borne by milk, may be of particular importance in the neonatal period. In addition to BSSL,
The Journal of Pediatrics Volume 125, Number 5, Part 2
Hernell and Bl~ckberg
S 57
Table I. Specific activity of BSSL against different substrates
Substrate
Activity (#mol product • min -1 . mg protein -1)
Trioleate 100 sn-1 (3) monooieate 21 sn-2 monooleate 16 Cholesteryl ester 0.5 Retinyl palmitate 6-8 p-Nitrophenyl acetate 30 Dioctanoyl-phospatidyl choline 80 Modifiedfrom B1/iekbergL., HernellO. FEBS Lett 1983;157:337-41.Additionaldata from Bl~ickberget al. (unpublisheddata, 1983).
gastric lipase, which is secreted by the chief cells of the gastric mucosa, 5 is considered to be of particular importance in the neonatal period. 1 FUNCTIONAL
PROPERTIES
OF BSSL
Substrate and positional specificity. Bile salt-stimulated lipase can hydrolyze a variety of ester substrates (Table I). Whether the substrate is in monomericl micellar, or emulsified form has little influence on the specific activity. It has been shown that BSSL also hydrolyzes certain amide bonds, 6 the physiologic significance of which is not presently understood. From a physiologic viewpoint, several substrates are likely to be attacked by BSSL, for example, triglycerides and partial glycerides and cholesteryl and retinyl esters. In addition to a relative lack of substrate specificity, BSSL has no obvious positional specificity when hydrolyzing a triglyceride. Unlike colipase-dependent pancreatic lipase, BSSL can hydrolyze all three ester bonds; as a result, free fatty acids and free glycerol are the end products of lipolysis7 rather than free fatty acids and sn-2 monoglyceride, which are the typical products of colipasedependent lipase. Wang et al. 8 compared the rates at which triglyceride, diglyceride, and monoglyceride were hydrolyzed. Diglyceride was the preferred substrate, and monoglyceride was hydrolyzed at the lowest rate. Bile salt activation. A characteristic feature of BSSL is its interaction with bile salts, which are required for BSSL activity with a long-chain triglyceride substrate. 9 Therefore, although BSSL is present in high amounts in human milk, it will not act on the milk fat before it reaches the duodenum, where it encounters bile salts. Neither the structural basis for bile salt interaction nor the exact mechanism for activation is completely understood. The activation shows a high degree of specificity. Only primary bile salts activate it, although both primary (e.g., cholate and
(
Activation site
Cholate
Lipid-bindingpromoti nsite g
~
Deoxycholate
Cholate
Fig. 4. Model presentation of the bile salt interaction of bile saltstimulated lipase, including two bile salt-binding sites. Tentative ligands and amino acid residues suggested to participate in the binding are indicated. Y, Tyrosine; R, arginine.
chenodeoxycholate) and secondary (e.g., deoxycholate) bile salts bind to the enzyme. 9 This suggests that the 7ahydroxyl group in primary bile salts is essential for BSSL activation. It has been suggested that primary bile salts activate the enzyme merely by inducing binding to a substrate interface, for example, an emulsion particle. 1° This suggestion is supported by the fact that BSSL can act on watersoluble ester substrates, for example, p-nitrophenylacetate, also in the absence of bile salts. With the use of radiolabeled BSSL we could show, in agreement with Wang et al., that BSSL was unable to bind to a lipid emulsion in the absence of bile salts. 11 Adding cholate to the system induced binding. However, efficient binding was also seen with deoxycholate, but without resulting lipolysis. Hence, promoting lipid binding is only part of the activation mechanism. From these data and other comparisons of cholate, deoxycholate, and cholamidopropyl dimethylammoniopropane sulfate, a cholate analogue with identical backbon e hydroxylation but with a different side chain, we have proposed a model for the activation that is based on two bile salt-binding sites on the enzyme molecule (Fig. 1).11 A nonspecific binding site (the lipid-binding-promoting site) is involved in binding to the lipid-water interface. Binding to a second site (the activation site) that is specific for bile salts or analogues that contain the 7c~-hydroxyl group is also required for activity. Experiments with amino acid-modifying reagents have provided evidence that tyrosyl and arginyl residues are involved in the bile salt interaction. 11-13 In addition to the effects on activity, bile salts also protect the enzyme from degradation and inactivation by pancreatic proteases, for example, trypsin. Unlike activation, this protective effect is seen with primary and secondary bile salts and cholamidopropyl dimethylammoniopropane sulfate,11, 14 Role in the digestion of human milk triglyeeride. Digestion of milk triglyceride is a process in which all three gastrointestinal lipases--gastric lipase, colipase-dependent
S 58
Hernell and Bl?tckberg
The Journal of Pediatrics November 1994
Table II. Characteristics of B S S L / C E L from different species Species
Man
Source
Milk, pancreas Cow Pancreas Rat Pancreas Rabbit Pancreas Salmon Pancreas Data from references21 and 28-31.
No. of residues
No. of repeats
722
16
578 592 556 540
3 4 2 or 3 0
bonds with long-chain polyunsaturated fatty acids, for example, arachidonic acid and eicosapentaenoic acid.18' 20 On the other hand, BSSL does not discriminate between these and other fatty acids that are present in the triglyceride substrate.18, 19 The most efficient release of long-chain polyunsaturated fatty acids was achieved when the two lipases were combined. From these in vitro experiments it is tempting to speculate that one role of BSSL is to ensure adequate levels of long-chain polyunsaturated fatty acids in breast-fed infants. STRUCTURAL
pancreatic lipase, and B S S L - - p l a y important roles. From in vitro studies using purified enzymes and cofactors, and using human milk fat globules as substrate under conditions mimicking those of the postprandial gastric and duodenal contents, respectively, we have put forward a model for how their interaction ensures efficient digestion and subsequent product absorption.15 Human milk fat globule triglyceride is not digested efficiently by either colipase-dependent lipase, even in the presence of its cofactor colipase, or by BSSL in the presence of bile salts. However, if the globules are first digested to only a limited degree by gastric lipase, the resulting droplets are efficiently hydrolyzed by colipasedependent lipase, yielding free fatty acids and sn-2 monoglycerides. If BSSL is included, hydrolysis nears completion (>90% of the ester bonds are hydrotyzed), producing free fatty acids and free glycerol as the end productsJ 5 Under conditions with low intraluminal bile salt concentrations, as is often the case in the neonatal period, such a product pattern may be favorable for absorption. Fat absorption studies in preterm infants have shown that pasteurization of human milk, which inactivates BSSL, reduces the coefficient of fat absorption by as much as one third) 6' 17 Role in the use of long-chain polyunsaturated fatty acids. In addition to its role in overall triglyceride digestion, it has been suggested that BSSL plays a role in the use of longchain polyunsaturated fatty acidsJ 8' 19 These fatty acids, most of which are supplied by the milk triglycerides, are essential not only as structural components, for example, in the brain and the retina, but also as precursors for eicosanoids. In adults these fatty acids are synthesized from precursor fatty acids of the n-6 and n-3 series, linoleic and ~-linolenic acid, respectively, by desaturation and elongation reactions. The capacity for these reactions is not fully developed at birth, however, and thus not only the precursor fatty acids but also the desaturated and elongated derivatives should be considered essential dietary components in the neonatal period. In vitro, colipase-dependent lipase has been shown to have restricted activity against ester
ASPECTS
OF BSSL
Structure of BSSL and its relationship to pancreatic CEL. We have cloned and sequenced eDNA covering the entire coding sequence of human-milk BSSL. 21 The sequence has been confirmed. 22 The deduced amino acid sequence is identical to that of CEL, also known as carboxyl ester hydrolase and cholesterol esterase, which is secreted by the human pancreas. 23 Earlier studies suggested that these two enzymes are one functional entity. 24 When the genomic structure became available it was clear that they are products of the same gene. 25 The previously noted difference in the apparent molecular size24 is, therefore, most likely the result of differences in the degree or type of tissue-specific glycosylation. The peptide chain contains one potential N-linked glycosylation site, located seven amino acid residues to the N-terminal of the tentative active site serine. 21 This site is glycosylated in both the milk and the pancreatic enzyme. In the pancreatic enzyme, the carbohydrate chain is known to be of the biantennary complex type. 26 Both enzymes are also heavily O-glycosylated at the C-terminal end of the molecule. This region of the polypeptide chain is characterized by a unique structure that consists of 16 near-identical proline-rich repeats of 11 amino acid residues. 21 Each repeat contains at least one potential O-glycosylation site. For the pancreatic enzyme the O-linked sugars are heterogeneous. 27 We have found pronounced differences between the milk and the pancreatic enzymes in composition of O-linked sugars (Bl~ickberg et al., unpublished data, 1993), which is a likely explanation for the difference in molecular size. The pancreatic enzyme has been cloned and sequenced in several other species, that is, rat, cow, rabbit, and salmon. 28"31 The main difference between species is in the number and structure of the C-terminal repeats; more than 70% homology exists in the N-terminal region preceding the repeats. With the exception of the salmon enzyme, all enzymes contain at least two repeats (Table II). A variant of human milk BSSL was recently reported that is smaller than the common form. 32 It is tempting to speculate that this shorter form contains fewer repeats and thus is more
The Journal of Pediatrics Volume 125, Number 5, Part 2
Hernell and Bl~ckberg
S I
-
D I
S 59
H I
~-~A A A A A A A A A A A A A A A A ~C
Fig. 2. Schematic presentatio n of human milk bile salt-stimulated lipase. Solid boxes indicate regions of homology to choline esterase. Positions of amino acid residues of the catalytic triad are indicated. Open box represents the region with the proline-rieh repeats. Positions of the disulfide bridges (l__t)and glycosylation sites (A) are indicated. S, Serine; D, aspartate; H, histidine.
similar to the forms found in other species. All species sequenced contain the same N-linked glycosylation site, which suggests that this site is of functional importance. The bovine enzyme contains an additional N-glycosylation site; whether this site is glycosylated is not known. Solely on the basis of sequence homologies, very few similarities between B S S L / C E L and other well-characterized mammalian lipases are found. All mammalian lipases contain the consensus sequence around the active-site serine: G-X-S-X-G. There are, however, homologies when BSSL/ CEL is compared with some lipases of microbial origin, for example, Geotrichum candidum lipase.33 These homologies are restricted to the N-terminal part of BSSL/CEL; none of the microbial enzymes contains any structure that resembles the C-terminal repeats of the human enzyme. Striking homologies in the N-terminal part are found with some typical esterases, for example, acetylcholine esterase from different species and rat liver carboxyl esterase. 21 The locations of these homologies are depicted in Fig. 2. Structure/function relationships in BSSL. An intriguing question is the functional importance of the unique C-terminal repeats. To address this question, we, in collaboration with L. Hansson and M. Str6mqvist (Umefi, Sweden) and L. Lundberg (M(51ndal, Sweden), constructed and expressed a set of BSSL variants by site-directed mutagenesis. 34 By using a bovine papilloma virus-based vector and C- 127 cells, variants were produced that lacked all or all but two of the repeats but contained the tail of 11 residues C-terminal of the last repeat. Both variants, as well as a full-length recombinant enzyme, were produced and secreted into the culture medium. The different enzyme proteins were purified, essentially by heparin-sepharose and immunoaffinity chromatographic methods, and characterized. All variants were active and had approximately the same specific activity when a long-chain triglyceride substrate was used. Furthermore, they all had the same bile salt dependency with regard to both the type and the concentration required for maximal activity. 34 It can thus be concluded that the C-terminal proline-rich, O-glycosylated repeats are essential for neither the catalytic activity of the enzyme nor the bile salt interaction, the lipid binding, or the heparin binding. It has been suggested that the sequence of
the repeats is similar to that of PEST (proline, glutamine, serine, threonine) sequences, which is, supposedly, a signal for proteolytic degradation and a fast turnover. 3° Such a signal is, however, highly unlikely for an enzyme that acts in the milieu of the intestinal contents. On the contrary, BSSL has been shown to be very resistant to inactivation and degradation by several proteases. Thus, why evolution has added these unique proline-rich, O-glycosylated repeats to a typical esterase is an intriguing but unanswered question. Using the same strategy as that described previously, we also investigated the importance of the N-linked sugar chain for catalytic activity.34 A BSSL variant in which the asparagine residue (asn-187) had been altered to a glutamine residue was equally active as the native enzyme. The conclusion that glycosylation, N- or O-linked, is not important for activity or bile salt activation was supported by the fact that we were able to produce active recombinant enzyme in Escherichia coli with the same bile salt dependency as the native milk enzyme. 34 Sequence homologies to other lipases/esterases strongly indicate that B S S L / C E L is a member of the family of a/fl_hydrolases,31, 35 with antiparallel/3 sheets making up the interior of the N-terminal part of the molecule. A catalytic triad typical of esterases/lipases has been suggested from homologies to other lipases/esterases to include ser-194, asp-320, and his-446. 35 The position of the essential serine and aspartate residues has been confirmed by site-directed mutagenesis. 22, 36 Interfacial activation. Some lipases, including pancreatic colipase-dependent lipase, contain a "lid structure. ''37 This short stretch of amino acid residues forms a loop structure that covers the active-site groove. For pancreatic lipase, x-ray crystall0graphy of a complex of lipase and its cofactor protein (colipase), formed in the presence of mixed micelles, has shown that this lid undergoes a conformational change when the lipase interacts with a lipid surface138 This change makes the active-site groove accessible to substrate molecules. Hence, it can explain the interfacial activation typical for colipase-dependent lipase. B S S L / C E L does not have the same interfacial activation, and it is therefore not surprising that no obvious lid-structure sequence can be
S60
Hernel[ and Bli~ckberg
found, The relative lack of substrate specificity of BSSL/ CEL indicates that the active site is easily accessible and supports the view that this enzyme lacks a lid. Moreover, the absence of a lid structure rules out the possibility that the bile salt-activation mechanism involves a conformational change in the lipase that is similar to that observed when colipase-dependent lipase interacts with a substrate interface. The exact residues involved in the interaction with bile salts have not been pinpointed. However, alignments of sequences of similar enzymes (e.g., acetylcholine esterases) that are unaffected by bile salts provide evidence for the tentative arginine and tyrosine residues involved (Fig. 1): CONCLUSION H u m a n milk BSSL is a lipolytic enzyme with several functions ascribed to it. Not only is BSSL important for overall triglyceride use in the breast-fed infant, but it has also been suggested to have specific roles in the use of milk long-chain polyunsaturated fatty acids, cholesterol, and fat-soluble vitamins. On a structural level BSSL belongs to the family of a/~-hydrolases in which many lipases and esterases a r e found. It is, however, very different from Other well-characterized mammalian lipases, with the exception of pancreatic CEL, which is a product of the same gene. From an evolutionary standpoint, the N-terminal part appears to be a product of an ancestral gene that gives rise to several lipases and esterases. The function of a unique repetitive sequence that has been added to this gene remains to be elucidated.
The Journal of Pediatrics November 1994
9.
10.
11.
12.
13.
14.
15:
16.
17.
18. REFERENCES
1. Hernell O, B1/ickberg L. Digestion and absorption of human milk lipids. In: Dulbecco R, ed. Encyclopedia of human biology. New York: Academic Press, 1991;3:47-56. 2. Patton JS, Warner TG, Benson AA. Partial characterization of the bile salt-dependent triacylgl~ceerollipase from the leopard shark pancreas. Biochim Biophys Acta 1977;486:32230. 3. GjeUsevik DR, Lombardo D, Walther BT. Pancreatic bile salt dependent lipase from cod (Gadus morhua): purification and properties. Biochim Biophys Acta 1992;1124:123-34. 4. Bradshaw WS, Rutter WJ. Multiple pancreatic lipases: tissue distribution and pattern of accumulation during embryological development. Biochemistry 1972;1 !:1517-28. 5. Moreau H; Laugier R, Gargouri Y, Ferrato F, Verger R. Human preduodenal lipase is entirely of gastric fundic origin. Gastroenterology 1988;95:1221-6. 6. Hui DY, Hayakawa K, Oizumi J. Lipoamidase activity in normal and mutagenized pancreatic cholesterol esterase (bile salt-stimulated lipase). Biochem J 1993;291:65-9. 7. Hernell O, B1/ickberg L. Digestion of human milk lipids: physiologic significance of sn-2 monoacy!glycerol hydro!ysis by bile salt-stimulated lipase. Pediatr Res 1982;16:882-5. 8. Wang C-S, Hartsuck JA, Downs D. Kinetics of acylglycerol
19.
20.
21.
22.
23.
24.
sequential hydrolysis by human milk bile salt activated lipase and effect of taurocholate as fatty acid receptor. Biochemistry 1988;27:4834-40. Hernell O, Olivecrona T. Human milk lipases, II: Bile salt-stimulated lipase. Biochim Biophys Acta 1974;369:23444. Wang C-S, Lee DM. Kinetic properties of human milk bile salt-activated lipase: studies using long chain triacylglycerol as substrate. J Lipid Res 1985;26:824-30. Blhckberg L, Hernell O. Bile salt-stimulated lipase in human milk: evidence that bile salts induce lipid binding and activation via binding to different sites. FEBS Lett 1993;323:207-10. Lombardo D, Campese D, Multigner L, Lafont H, De Caro A. On the probable involvement of arginine residues in the bilesalt-binding site of human pancreatic carboxylic ester hydrolase. Eur J Biochem 1983;133:327-33. Campese D, Lombardo D, Multigner L, Lafont H, De Caro A. Implication of a tyrosine residue in the unspecific bile salt binding site of human pancreatic carboxylic ester hydrolase. Biochim Biophys Acta 1984;784:147-57. B1/ickberg L, Hernell O. Further characterization of the bile salt-stimulated lipase in human milk. FEBS Lett 1983; 157:33741. Bernb~ick S, Bl~ickberg L, Hernell O. The complete digestion of human milk triacylglycerol in vitro requires gastric lipase, pancreatic colipase-dependent lipase, and bile salt-stimulated lipase. J Clin Invest 1990;85:1221-6. Williamson S, Finucane E, Ellis H, Gamsu HR. Effect of heat treatment of human milk on absorption of nitrogen, fat, sodium, calcium, and phosphorus by preterm infants. Arch Dis Child 1978;53:555-63. Atkinson SA, Bryan MH, Anderson GH. Human milk feeding in premature infants: protein, fat, and carbohydrate balances in the first two weeks of life. J PEDIATR1981;99:61724. Hernell O, Bl~ickbergL, Chen Q, Sternby B, Nilsson ~,. Does the bile salt-stimulated lipase of human milk have a role in the use of the milk long-chain polyunsaturated fatty acids? J Pediatr Gastroenterol Nutr 1993;16:426-31. Chen Q, Bl~ickberg L, Nilsson ~., Sternby B, Hernell O. Digestion of triacylglycerols containing long-chain polyenoic fatty acids in vitro by colipase-dependent pancreatic lipase and human milk bile salt-stimulated lipase. Biochim Biophys Acta 1994;1210:239-43. Chert Q, Sternby B, Akesson B, Nilsson ,~. Effects of human pancreatic lipase-colipase and carboxyl ester lipase on eicosapentaenoic and arachidonic acid ester bonds of triacylglycerols rich in fish oil fatty acids. Bi0chim Biophys Acta 1990; 1044:111-7. Nilsson J, B1/ickberg L, Carlsson P, EnerNick S, Hernell O, Bjursell G. eDNA cloning of human-milk bile-salt-stimulated lipase and evidence for its identity to pancreatic carboxylic ester hydrolase. Eur J Bioehem 1990;192:543-50. Baba T, Downs D, Jackson KW, Tang J, Wang CS. Structure of human milk bile salt activated lipase. Biochemistry 1991; 30:500-10. Reue K, Zambaux J, Wong H, et al. eDNA clon!ngof carboxyl ester lipase from human pancreas reveals a unique proline-rich repeat unit. J Lipid Res 199l;32:267-76. Bl~ickberg L, Lombardo D, Hernell O, Guy O, Olivecrona T. Bile salt-stimulated lipase in human milk and Carboxyl ester
The Journal of Pediatrics Volume 125, Number 5, Part 2
25.
26.
27.
28.
29.
30.
31.
hydrolase in pancreatic juice: are they identical enzymes? FEBS Lett 1981;136:284-8. Lidberg U, Nilsson J, Str6mberg K, et al. Genomic organization, sequence analysis, and chromosomal localization of the human carboxyl ester lipase (CEL) gene and a CEL-like (CELL) gene. Genomics 1992;13:630-40. Sugo T, Mas E, Abouakil N, et al. The structure of N-linked oligosaccharides of human pancreatic bile-salt-dependent lipase. Eur J Biochem 1993;216:799-805. Mas E, Abouakil N, Roudani S, Franc J-L, Motreuil J, Lombardo D. Variation of the glycosylation of human pancreatic bile-salt-dependent lipase. Eur J Biochem 1993;216:80712. Han JH, Stratowa C, Rutter WJ. Isolation of full-length putative rat lysophospholipase cDNA using improved methods for mRNA isolation and cDNA cloning. Biochemistry 1987; 26:1617-25. Kyger EM, Wiegand RC, Lange LG. Cloning of the bovine pancreatic cholesterol esterase/lysophospholipase. Biochem Biophys Res Commun 1989;164:1302-9. Colwell NS, Aleman-Gomez JA, Kumar BV. Molecular cloning and expression of rabbit pancreatic cholesterol esterase. Biochim Biophys Acta 1993;1172:175-80. Gjellesvik DR, Lorens JB, Male R. Computer-modelled tertiary structure of carboxylester lipase from Atlantic salmon
Hernell and Bl~tckberg
32.
33.
34.
35.
36.
37. 38.
S61
(Salmo salar). In: Lipases: structure, function and protein engineering. Elsinore, Denmark: 1993:47. Swan JS, Hoffman MM, Lord MK, Poechmann JL. Two forms of human milk bile-salt-stimulated lipase. Biochem J 1992; 283:119-22. Schrag JD, Li Y, Wu S, Cygler M. Ser-His-Glu triad forms the catalytic site of the lipase from Geotrichum candidum. Nature 1991;351:761-4. Hansson L, Blfickberg L, Edlund M, Lundberg L, Str6mqvist M, Hernell O. Recombinant human milk bile salt-stimulated lipase: catalytic activity is retained in the absence of glycosylation and the unique proline-rich repeats. J Biol Chem 1993; 268:26692-8. Wang C-S, Hartsuck JA. Bile salt-activated lipase: a multiple function lipolytic enzyme. Biochim Biophys Acta 1993; 1166:119. DiPersio LP, Hui DY. Aspartic acid 320 is required for optimal activity of rat pancreatic cholesterol esterase. J Biol Chem 1993;268:300-4. Winkler FK, D'Arcy A, Hunziker W. Structure of human pancreatic lipase. Nature 1990;343:771-4. van Tilbeurgh H, Egloff M-P, Martinez C, Rugani N, Verger R, Cambillau C. Interracial activation of the lipase-procolipase complex by mixed micelles revealed by x-ray crystallography. Nature 1993;362:814-20.