Compartmentalization of fucosyltransferase and alpha-L-fucosidase in human milk

BIOCHEMICAL AND MOLECULAR MEDICINE ARTICLE NO.

58, 211–220 (1996)

0051

Compartmentalization of Fucosyltransferase and a-L-Fucosidase in Human Milk GHERMAN YA. WIEDERSCHAIN

AND

DAVID S. NEWBURG

Department of Biochemistry, Shriver Center for Mental Retardation, 200 Trapelo Road, Waltham, Massachusetts 02254; and Harvard Medical School, Boston, Massachusetts 02115 Received March 6, 1996

olism of the fucosyloligosaccharides of human milk that protect against disease. q 1996 Academic Press, Inc.

Several pathogenic agents of pediatric gastroenteritis are inhibited by fucosylated oligosaccharides of human milk. Biosynthesis and degradation of fucosyloligosaccharides is controlled by fucosyltransferase and a- L-fucosidase. The activity of these enzymes varies reciprocally over the course of lactation. We hypothesized that differences in the specific organization of these enzymes in compartments of human milk might contribute to such differences in activity. Therefore, the distribution of these enzymes in various compartments of human milk was investigated. After ultracentrifugation at 120,000g for 2h, the fucosyltransferase activity distributes evenly between the supernatant and the membranous pellet. Ultracentrifugation at 180,000g for 17 h further fractionated the milk into a clear supernatant, a fluff layer from the supernatant, and a pellet. The fluff was visualized by electron microscopy. The distribution of fucosyltransferase activity in colostrum was compared with that of mature milk from the same donor. In mature milk from Day 30 of lactation, most fucosyltransferase activity was in the membranous fluff fraction, while in colostrum from Day 5 of lactation, most of the fucosyltransferase activity was in the supernatant. In contrast to fucosyltransferase, fucosidase activity was found only in the soluble milk fraction; upon prolonged ultracentrifugation, most of this was membrane associated. The nature of human milk fucosidase was studied. This enzyme is glycosylated and exhibits several characteristics common to other fucosidases. Under the conditions found in human milk, it exhibits almost full activity. The variation in compartmentalization of fucosyltransferase activity during lactation may reflect variations in metab-

Human milk contains large amounts and a wide variety of oligosaccharides and glycoproteins (1 – 4). Such fucosyloligosaccharide structures of milk may be homologous to epithelial cell surface carbohydrates that function as pathogen receptors. Milk contains large amounts of fucosyloligosaccharides and fucose-containing glycoconjugates, some of which inhibit adhesion to epithelial surfaces by bacteria and other pathogens; such adhesion is an essential early stage in the pathogenesis of the disease (5,6). For example, oligosaccharides inhibit binding of Streptococcus pneumoniae to target cells in vitro (5). We found that fucosyloligosaccharide(s) of human milk protects against stable toxin of Escherichia coli in vivo (7) and in vitro (8) and that fucosyloligosaccharides inhibit the binding of invasive strains of Campylobacter jejuni to target cells in vitro (9) and in vivo (10). More than 80 oligosaccharides have been isolated from human milk and their structures determined (4). Moreover, mass spectrometry of the human milk oligosaccharide fractions indicate that the number and complexity of unidentified fucosyloligosaccharides greatly exceeds those already known, with structures of up to 15 fucosyl residues and as many as 32 sugars (11). Recognition of the potential functions of such structures is also expanding rapidly. Some of these oligosaccharides contain important epitopes that are found on the erythrocyte cell surface and are antigenic determinants of blood group substances ABO(H) and Lewis systems. For exam211 1077-3150/96 $18.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.

AID

BMM 2517

/

af02$$$161

07-10-96 15:34:39

bmma

AP: BMM

212

WIEDERSCHAIN AND NEWBURG

ple, the Lewisx and Lewisa epitope are frequent motifs in milk oligosaccharides, and both these and their sialylated moieties are typically found as terminal residues of larger oligosaccharides on glycoproteins, proteoglycans, and glycolipids. These structures can be biologically significant, i.e., as ligands for adhesion molecules such as E-, L-, and Pselectins (12). The fucosyl residues of the human milk glycoconjugates can be attached by any of several types of linkages, each determined by one of a family of fucosyltransferases. The fucosyltransferases have been isolated from a number of sources, including human milk (13–15). The major fucosyltransferase of milk is reported to be the a(1-3/4) fucosyltransferase (EC 2.4.1.65), whose gene, located on chromosome 19, has been cloned; this enzyme is referred to as Lewis type fucosyltransferase III (16,17). a-L-Fucosidase (EC 3.2.1.51) can hydrolyze fucose from fucosyloligosaccharides and other fucose-containing glycoconjugates, which, by changing their structure also changes their biological activity (18). Fucosidase is a ubiquitous lysosomal glycosidase found in a wide variety of organisms (19–21). The gene for a-L-fucosidase, located on chromosome 1, was cloned and accounts for approximately 80% of the mature enzyme activity (22). Recently, we found fucosidase activity in human milk from healthy donors. A high degree of individual variability was observed for both the fucosyltransferase and the fucosidase activities of human milk. Fucosyltransferase and fucosidase activities are inversely related in human milk over the course of lactation (23). In this context, it is interesting to note that there are temporal and individual variations in the oligosaccharide composition of human milk (24). As variation in the spectrum of biologically active fucosyloligosaccharides found in milk reflects variation in their synthesis and modification, changes in fucosyltransferase and fucosidase activity over the course of lactation might ultimately help explain variation of host sensitivity to specific pathogens in young infants. One key step toward understanding the regulation of these enzyme activities is to obtain data about the compartmentalization of these milk enzymes and of their substrates. Human milk consists of several compartments as defined by differential centrifugation (25,26). These include the cream, consisting of membrane-bound lipid droplets, and the components of the skim milk, including the casein pellet, a membranous fluffy layer, and a clear

AID

BMM 2517

/

af02$$$162

07-10-96 15:34:39

supernatant (26). Compartmentalization of fucosyltransferase and fucosidase in human milk reflects their origins and may control their functions and activities. Therefore, we investigated the distribution of fucosyltransferase and fucosidase in different compartments of human milk. Unlike fucosyltransferases, there are virtually no data on human milk a-L-fucosidase. Therefore, the glycoprotein nature of this enzyme and some of its biochemical characteristics were studied. MATERIALS AND METHODS Milk Fractionation Human milk, previously stored at 0707C, was fractionated by sequential centrifugation. Milk was centrifuged at 3000g for 30 min at 47C. The 3000g infranatant (skim milk) was then centrifuged at 120,000g for 2 h. Prolonged ultracentrifugation of the skim milk was performed at 180,000g for 17 h at 47C, yielding a supernatant, a fluffy layer, and a pellet. Electron Microscopy Ten milliliters of human skim milk produced a supernatant of 7 ml clear fluid and approximately 2 ml of supernatant containing the fluff material; each was removed by Pasteur pipette. The solid pellet from the bottom of the tube was resuspended in 1 ml of 0.9% NaCl and homogenized in a glass Kontes dual tissue grinder. The fluff solution and pellet were then prepared for electron microscopic studies as described by Stewart et al. (27). Briefly, a portion of the fluff solution and the resuspended pellet were fixed in 3% glutaraldehyde in 0.1 M cacodylate buffer, pH 7.3, for 2 h. The samples were centrifuged, decanted, washed overnight in 0.1 M cacodylate, and postfixed in 1% OsO4 in the same buffer for 1 h. Specimens were dehydrated through graded acetones and embedded in Epon (Polysciences). Silver to silver-gold thin sections were cut onto copper 200 mesh grids and stained with Sato’s lead before examination at 60 kV in a Philips 300 electron microscope. Fucosyltransferase Assay The reaction mixture (100 ml) contained 5 mmol of 3-(N-morpholino) propansulfonic acid (Mops)/NaOH buffer, pH 7.5, 0.5 mmol of MnCl2 , 10 mmol of NaCl, 1 nmol of GDP-L-fucose (0.1 nmol of GDP-L-[14C]fucose, 57,000 cpm [283 mCi/mmol, New England Nuclear, Boston, MA], and 0.9 nmol of unlabeled

bmma

AP: BMM

213

MILK FUCOSYLTRANSFERASE, FUCOSIDASE

GDP-L-fucose [Sigma, St. Louis, MO]), and 24 mmol of lactose. Skim milk was used as a positive control for fucosyltransferase activity. After 1 h of incubation at 377C the reaction was terminated by the addition of 1 ml of ice-cold water with immersion of the sample into an ice bath. The mixture was immediately applied to an anion exchange column (1 ml) of AG-1X8 resin (100–200 mesh, acetate, BioRad, Richmond, CA), the column washed with 1 ml of water, and the combined effluents containing the fucosylated products were collected in scintillation vials. After addition of 5 ml liquid scintillation cocktail (Ready Safe, Beckman), emission of beta particles was quantified in a liquid scintillation counter. Background was determined in reagent blanks incubated without any fucosyltransferase. Fucosyltransferase activity was expressed as picomoles of total fucosylated products per hour per milliliter of milk or per milligram of total milk protein.

a-L-Fucosidase Assay The reaction mixture (100 ml) contained 5 ml skim milk (protein concentration about 1 mg/ml) and 1 mM 4-methylumbelliferyl a-L-fucopyranoside (MUFuc [Sigma]) dissolved in 0.2 M citrate–phosphate buffer, pH 5.0. After 1 h of incubation at 377C the reaction was stopped with 2 ml of 0.25 M glycine– NaOH buffer, pH 10.4. The fluorescence of enzymatically liberated 4-methylumbelliferone was determined (Aminco SPF-500 C) by activation at 365 nm and emission measured at 480 nm. The fucosidase activity was expressed as nanomoles of 4-methylumbelliferyl a-L-fucopyranoside (MUFuc) hydrolyzed per hour per milliliter of milk or per milligram of protein.

tion at 5000g, 30 min, the supernatant was used for isoelectric focusing in a Rotofor apparatus (BioRad) with carrier ampholite (pH 3–10, 5–7, and 4–6, BioRad) at a final concentration of 1%. This milk was separated by high-resolution preparative isoelectric focusing at 12 W for 4 h at 47C. The fucosidase activity was determined in the 20 resulting fractions after adjusting the pH to 5.0 with citrate (0.1 M)–phosphate (0.2 M) buffer in the assay samples. MUFuc in a final concentration of 1 mM was used as substrate. The time of incubation was 1 h. pH Activity and pH Stability The pH–activity curve of human milk a-L-fucosidase was determined with 1 mM MUFuc as substrate in citrate (0.1 M)–phosphate (0.2 M) buffer with various pH values, from pH 3.0 to pH 7.5. The pH–stability was measured after 30 min of incubation of milk samples at 377C from pH 3.0 to pH 7.5. Residual a-L-fucosidase activity was determined after adjustment of all samples to pH 5.0 using citrate (0.1 M)–phosphate (0.2 M) buffer. Thermostability Heat inactivation assays for a-L-fucosidase were performed by incubation of milk samples for 30 min at various temperatures. The activity of the enzyme was determined at its pH optimum, 5.0. Protein was determined using BCA protein assay reagent (Pierce, Rockford, IL) as described (23). The rate of the enzyme reactions was shown to be linear with time and with protein concentration within the parameters of the research described above. RESULTS

Concanavalin A–Sepharose Chromatography One milliliter skimmed human milk was applied to a Concanavalin A–Sepharose column (Vt Å 1 ml) that had been equilibrated in Tris–HCl buffer, pH 7.4, containing 0.5 M NaCl, 1 mM CaCl2 , and 1 mM MnCl2 . The column was washed with the same buffer until the absorbance at 280 nm was close to baseline (approximately 12 ml). Adherent sample was then eluted with 5 ml of the same buffer containing 0.5 M a-methylmannoside. Isoelectric Focusing Fifty milliliters of human skim milk was dialyzed for 48 h against distilled water and after centrifuga-

AID

BMM 2517

/

af02$$$162

07-10-96 15:34:39

Characterization of Human Milk a-L-Fucosidase

a-L-Fucosidase binding to Concanavalin A–Sepharose. The bulk of milk proteins, which lacked aL-fucosidase activity, eluted in the wash preceding the a-methylmannopyranoside elution step. Human milk fucosidase binds to Concanavalin A–Sepharose and elutes with 0.5 M a-methylmannopyranoside (Fig. 1). As Concanavalin A has binding specificity for mannose and glucose glycosides, and a-methylmannoside is a specific inhibitor of lectin-mannoside binding, this indicates that the human milk fucosidase contains mannosyl residues and, therefore, is a glycoprotein enzyme.

bmma

AP: BMM

214

WIEDERSCHAIN AND NEWBURG

shows a typical a-L-fucosidase profile using pH 5–7 ampholytes.

FIG. 1. Elution profile of human milk a-L-fucosidase from a Con A–Sepharose column by 0.5 M methyl-a-D-mannopyranoside. Dark square, fucosidase activity; open diamond, absorption at 280 nm; the arrow shows start of elution by 0.5 M methylmannoside.

Isoelectric focusing of a-L-fucosidase. The isoelectric focusing profile of a-L-fucosidase activity from human milk was performed in the presence of 1% ampholytes. The profiles at pH 3–10, 4–6, and 5–7 consistently displayed a broad asymmetric peak with maximal fucosidase activity at pH 5.7–6.0. The asymmetry displayed two shoulders, suggesting the presence of more than one component. Figure 2D

pH activity of a-L-fucosidase. Figure 2A depicts pH-activity of human milk a-L-fucosidase. The optimal pH for fucosidase activity was at 5.0. At acidic pH 3.0 or 3.5, the residual activity was only about 4 or 11% of that at pH 5.0. At pH 5.5–6.5 the residual fucosidase activity was about 90% of optimal activity at pH 5.0. At neutral pH (7.0), close to the pH of human milk, the activity of fucosidase was 73% of optimal activity, decreasing steadily at pH 7.5 when the enzyme showed approximately 50% from activity at pH 5.0. pH stability of a-L-fucosidase. The pH–stability curve of fucosidase activity is shown in Fig. 2B. The activities of human milk samples incubated at pH 3.0, 3.5, and 4.0 were reduced to 5, 7, and 25%, respectively, following adjustment of the pH to 5.0 and incubation. The enzyme was completely stable at pH 5.0–7.5. Thermostability. Human milk a-L-fucosidase is a comparatively thermostable enzyme, which remains completely active after 30 min of heating at 37 and 507C. Rapid inactivation of enzyme activity was ob-

FIG. 2. Biochemical parameters of human milk a-L-fucosidase. (A) pH-dependence of activity: (B) pH–stability: (C) Thermostability: (D) Isoelectric focusing profile at pH 5.0–7.0: pH, open circles; fluorescence, closed circles.

AID

BMM 2517

/

af02$$$162

07-10-96 15:34:39

bmma

AP: BMM

MILK FUCOSYLTRANSFERASE, FUCOSIDASE

215

FIG. 3. Distribution of fucosyltransferase (A) and a-L-fucosidase (B) from human milk after ultracentrifugation at 220,000g, 2 h. (S), supernatant; (P), pellet. Distribution of fucosyltransferase (C) and a-L-fucosidase (D) in supernatant (S), in fluff layers (F), and in pellets (P) after centrifugation at 180,000g, 17 h. In all cases the activity of each enzyme in skim milk without centrifugation was expressed as 100%. Data shown are mean values from duplicate experiments; replicates were within 5% of each other.

tained at 60 and 707C, as measured by 30 min of subsequent incubation at 377C. Under these conditions, the activity of a-L-fucosidase was 10 and 4%, respectively, of the original activity without any heating (Fig. 2C). Distribution of Fucosyltransferase and a-L-Fucosidase in Human Milk After centrifugation at 120,000g for 2 h, about 50% of the fucosyltransferase activity of human milk and 100% of the fucosidase activity were found in the supernatant fraction. About 30% of the fucosyltransferase activity remained in the pellet. Increasing centrifugation to 220,000g for 2 h results in essentially the same distribution of fucosyltransferase and fucosidase (Figs. 3A and 3B). Treatment of the pellet with 1% Triton X-100 or sonication did not increase the partition of fucosyltransferase activity into the supernatant. Skim milk was subjected to prolonged centrifugation (17 h at 180,000g) from one donor at two stages of lactation, 5 and 30 days. The yield of fucosyltrans-

AID

BMM 2517

/

af02$$$162

07-10-96 15:34:39

ferase activity in the milk sample at 30 days of lactation was highest in the fluff layer (52%) and in the pellet fraction (36%). The supernatant fraction during this period of lactation contained only 12% of total fucosyltransferase activity (data not shown). The yield of fucosyltransferase activity from the milk of the same donor at 5 days of lactation in the supernatant, fluff layer, and pellet were 57, 28, and 14%, respectively (Fig. 3C). In the samples from both 5 and 30 days of lactation the highest fucosidase activity, 45% of the original activity, was found in the fluff layer, whereas the supernatant and pellet fractions contained only 10 to 20% of the original fucosidase activity (Fig. 3D). The origin of the milk fucosyltransferases and fucosidase could be either the Golgi complex of the mammary acinar cell or the leukocytes in milk. To investigate this issue, we obtained fresh human milk. Within 30 min of obtaining the samples, the milk was centrifuged at 7000g for 10 min. Under these conditions the cells sediment to form a pellet; for both enzymes more than 90% of the activity was

bmma

AP: BMM

216

WIEDERSCHAIN AND NEWBURG

casein described in bovine milk by Stewart et al. (27). The fluff layer (Fig. 4B) contained large open vesicles (open arrows) and darkly staining granules (arrowheads) similar to the features described in the human fluff layer by Isaacs and Moretz (28) and by Patton and Huston (29). Differential electron densities were noted among membranes of vesicular structures, similar to that described in bovine milk by Stewart et al. (27). DISCUSSION

FIG. 4. The high speed ‘‘casein’’ pellet (A) had features similar to that seen in portions of the high speed pellet of human milk. The fluff layer (B) contained large open vesicles (open arrows), vesicles containing densely staining materials (closed arrows), and darkly staining granules (arrowheads).

found in the supernatant, indicating that most fucosyltransferase and fucosidase is found in human milk, per se, and not within the intact leukocytes of human milk. The morphological features of the high-speed (180,000g, 17 h) pellet and fluff layers of the milk is shown in Fig. 4. The high-speed pellet (Fig. 4A) had features similar to that seen in portions of the highspeed pellet of human milk described by Isaacs and Moretz (28) and similar to the features of ‘‘dispersed’’

AID

BMM 2517

/

af02$$$163

07-10-96 15:34:39

It is now widely accepted that human milk is a source not only for nutrient components but also consists of many types of defense agents, enzymes, hormones, growth factors, membranes, and cells (30). Among these are human milk components that might affect nervous system development, brain growth, and maturation (31). Many human milk components are compartmentalized. This compartmentalization may be biochemically and physiologically relevant. Compartmentalization of enzymes in human milk may be of special interest because these biological catalysts might play key roles in modifications of different compounds of human milk, including defensive factors against pathogenic agents of pediatric gastroenteritis. Infectious diarrhea causes the death of more than 5 million people annually worldwide. Most of these deaths occur in individuals less than 1 year old (32). Fucose-containing oligosaccharides and glycoconjugates are potent inhibitors of bacterial adhesion to epithelial surfaces and increase resistance to enteric bacterial toxins in vivo. We previously observed reciprocal changes of activity of fucosyltransferase and fucosidase during lactation. The activity of a-Lfucosidase in human milk was described for the first time (23). Biochemical characterization of a-L-fucosidase in the present study showed that this enzyme is a glycoprotein and is comparatively stable to extremes of heat and pH. It has an optimum activity at pH 5.0, which is typical for other lysosomal enzymes and for a-L-fucosidases from other mammalian tissues and biological fluids (20,21)). The characteristics of the pH–activity curve as well as isoelectric focusing data suggest that human milk contains more than one fucosidase, consistent with a-L-fucosidases from different animal and human tissues and biological fluids (21). Differential sialylation of the carbohydrate moieties of a-L-fucosidase accounts for some of the multiple molecular forms found in human tissues and blood serum (33–35). However, hu-

bmma

AP: BMM

MILK FUCOSYLTRANSFERASE, FUCOSIDASE

man milk a-L-fucosidase did not separate into multiple forms by isoelectric focusing, usually a rather effective method for detecting multiple forms of a-Lfucosidase. In ultrafiltration experiments only part of unpurified a-L-fucosidase from skimmed human milk passed through a membrane filter which nominally excludes proteins with molecular weights of about 300,000. The remainder of active enzyme was found in the milk filtrate (unpublished observation). The molecular weight of purified tetrameric fucosidase from different mammalian tissues and liquids is approximately 230–280 kD by sedimentation equilibrium analysis (21), suggesting that aggregate complexes could form in human milk. The nature of these aggregates is unclear. Ultracentrifugation of human milk samples at 120,000g and 220,000g for 2 h indicates that almost all of the fucosidase activity associates with the supernatant. This indicates that fucosidase found in human milk is a soluble glycoprotein, as are the fucosidases from other human tissues and fluids (19–21,36). This is consistent with an origin for milk fucosidases similar to that of fucosidases in other secretions (37–39), that is, synthesis as a glycoprotein in the Golgi complex and then secretion into the fluid compartment of milk during lactation. However, after prolonged centrifugation of 180,000g for 17 h in samples from both 5 and 30 days of lactation, the highest fucosidase activity, 45% of the original activity, was found in the fluff layer, whereas the supernatant and pellet fractions contained only 10 to 20% of the original fucosidase activity. These data might reflect sedimentation of high-molecular-weight fucosidase aggregates or association of fucosidase with other high-molecular-weight components of human milk or with small, light membrane fragments that coprecipitate with the fluff layer. Electron microscopy indicated that the fluff layer was composed primarily of membranes and membrane fragments, consistent with the latter. These speculations need more experimental evidence. We are unaware of any other information about compartmentalization of fucosyltransferases in human milk. Most of the information available describes the purification, properties, substrate specificity, and molecular biology of this enzyme(s) (13–17,40–42). The yield of fucosyltransferase after ultracentrifugation at 120,000g and 220,000g showed that activity of this enzyme was distributed almost equally between the supernatant and pellet fractions. This suggests that human milk fucosyltrans-

AID

BMM 2517

/

af02$$$163

07-10-96 15:34:39

217

ferase is a member of the Golgi complex glycosyltransferase family, which typically is a resident membrane glycoprotein with an amino-terminal cytoplasmic tail, signal anchor domain, and extended stem region followed by a large carboxyterminal catalytic domain (43,44). The soluble form of this enzyme that was observed in milk might be due to cleavage between the catalytic domain and the transmembrane domain by endogenous proteases. If this were the case, then the stem region could be a target site for proteolytic cleavage, and the cleaved catalytic domains could be secreted from the cells, as is known for the two soluble forms of b-1,4 galactosyltransferase secreted into milk (17). The distribution of fucosyltransferase following long periods of centrifugation (180,000g, 17 h) was different in skim milk samples from one donor at 5 days of lactation (colostrum) and 30 days (mature milk). After ultracentrifugation, we observed that the top milk fat globule layer, fluff, and pellet layers were significantly greater than those from the 30-day milk sample. In the 30-day milk sample, most fucosyltransferase activity was in the membranous fluff fraction, while in colostrum most of the fucosyltransferase activity was in the supernatant. This suggests that the different periods of lactation may be characterized by variation in the distribution of fucosyltransferase activity, perhaps reflecting a more general variation in some of the membranous components of human milk. The identification of the membranous components of the fluff layer and pellet by electron microscopy demonstrated that our separation of the milk components was very similar to that obtained by Isaacs and Moretz (28) and Patton and Huston (29). The fluff layer of mature milk is thought to contain membranes of the milk fat globule, and endoplasmic reticulum from the mammary acinar cell is often included in the milk fat globule; these membranous structures are likely to copurify in the fluff layer. The association of fucosyltransferase with the fluff component in mature milk may be related to the association of glycosyltransferases with endoplasmic reticular structures of the mammary epithelium. Human milk is known to contain many types of cells including leukocytes (neutrophils, monocytes, macrophages, and lymphocytes), epithelial cells (secretory, squamous, ductal), and others. Using fresh milk samples we determined that 90% of the activity of both enzymes was in the supernatant, indicating that bulk of these enzymes are present in human milk per se rather than within the cellular compartments.

bmma

AP: BMM

218

WIEDERSCHAIN AND NEWBURG

Furthermore, membrane-bound fucosyltransferase was active, and this activity did not change after Triton X-100 treatment or sonication. This indicates that membrane-bound fucosyltransferase, as found in human milk, is as available to its soluble donor and acceptors as the soluble form of the enzyme and is not inhibited by proximate membranous structures. In conclusion, our data show different compartmentalization of fucosyltransferase and fucosidase, enzymes involved in the synthesis and degradation of fucose-containing compounds. Each of these enzyme activities may have distinct effects on the metabolism of fucosyloligosaccharides and other complex fucosylated carbohydrate moieties, and both enzyme activities change significantly during lactation. Fucosyltransferase requires high concentrations of GDP-fucose as the donor of fucosyl residues. Such concentrations of nucleotide sugars, including GDPfucose, have not been described in human milk. The potential for fucosidase to manifest activity in human milk, on the other hand, is quite high. Human milk fucosidase demonstrates stability to extremes of temperature and pH and demonstrates high activity at the pH of human milk. Generally, mammalian fucosidases show activity with a wide variety of fucosecontaining substrates, including milk oligosaccharides and synthetic fucose-containing oligosaccharides (18,45–47). Fragments of blood group A and H substances (48) and some glycopeptides (49), H antigen glycolipids (50), fucose containing gangliosides (51,52), and a synthetic fluorogenic lipid-like fucoside are all substrates for mammalian a-L-fucosidases (53). The broad substrate specificity of fucosidase to natural substrates shows that practically all of the known human milk fucose-containing oligosaccharides and glycoconjugates known to be protective agents against pediatric gastroenteritis might be potential substrates for this enzyme. Moreover, some types of fucosidases are capable of hydrolyzing hydrophobic fucosyl glycoconjugates without any activator proteins or detergents at pH 3.4 (51). Although biochemical pathways for the synthesis and secretion of many components of milk have been established during the past two decades (54–56), less is known about the synthesis and metabolism of the oligosaccharides and glycoconjugates of human milk, particularly with regard to the relationship of milk enzymes to the metabolism and concentration of their substrates in human milk. Specifically, human milk fucosyltransferase and fucosidase might play some significant role in structural modification of milk fucosyl-

AID

BMM 2517

/

af02$$$163

07-10-96 15:34:39

oligosaccharides in the mammary acinar cell, in fluid milk once secreted, and/or in the alimentary canal of the infant. In this context, it is interesting to speculate that some of these enzymes can be considered candidates for biological modifiers, not only of glycoconjugate receptors of bacteria and toxins in the microvillus membrane (32,57), but also of certain glycoconjugate immunoreactive antigens of human immunodeficiency virus, schistosomes (58), and others, which might be considered as a trigger to the first steps in the pathogenesis of gastroenteritis. Information about the stability of these enzymes in the gastrointestinal tract would be necessary to postulate any possible role of these milk enzymes in modifying different glycoconjugate enterocyte receptors. Further studies are needed to investigate the possibility of these enzymes modifying the structures of oligosaccharides with potential biological activity. ACKNOWLEDGMENTS The authors appreciate the thoughtful input that Dr. Robert H. McCluer has contributed to the manuscript. We thank Dr. James Crandall for preparation of photomicrographs of milk samples after ultracentrifugation and Kathryn Newburg for her expert assistance in the preparation of the manuscript. Supported by NIH Grant HD 13021.

REFERENCES 1. Kobata A. Milk glycoproteins and oligosaccharides. In Mammalian Glycoproteins and Glycolipids (Horowitz MI, Pigman W, Eds.). New York: Academic Press, 1977, Vol 1, pp 423– 440. 2. Egge H, Dell A, Von Nicolai H. Fucose containing oligosaccharides from human milk. I. Separation and identification of new constituents. Arch Biochem Biophys 224:235–253, 1983. 3. Ginsburg V, Robbins PW (Eds.). Biology of Carbohydrates. New York: Wiley, 1984. 4. Newburg DS, Neubauer SH. Carbohydrates in milk. In Handbook of Milk Composition (Jensen RG, Ed.). San Diego: Academic Press, 1995, pp 273–349. 5. Andersson B, Porras O, Hanson LA, Lagergard T, SvanborgEden C. Inhibition of attachment of Streptococcus pneumoniae and Haemophilus influenzae by human milk and receptor oligosaccharides. J Infect Dis 153:232–237, 1986. 6. Newburg DS. Oligosaccharides and glycoconjugates in human milk: Their role in host defense. J Mamm Gland Biol Neoplasia 1: 271–283, 1996. 7. Newburg DS, Pickering LK, McCluer RH, Cleary TG. Fucosylated oligosaccharides of human milk protect suckling mice from heat-stabile enterotoxin of Escherichia coli. J Infect Dis 162:1075–1080, 1990. 8. Crane JK, Azar SS, Stam A, Newburg DS. Oligosaccharides from human milk block binding and activity of the Esche-

bmma

AP: BMM

MILK FUCOSYLTRANSFERASE, FUCOSIDASE richia coli heat-stable enterotoxin (STa) in T84 intestinal cells. J Nutr 124:2358–2364, 1994. 9. Ruiz-Palacios GM, Cervantes LE, Newburg DS, Lopez-Vidal Y, Calva JJ. In vitro models for studying Campylobacter jejuni infections. In Campylobacter jejuni. Current Status and Future Trends (Nachamkin I, Blaser MJ, Tomkins LS, Eds.). Washington, DC: Am. Soc. for Microbiol., 1992, pp 176–183. 10. Cervantes LE, Newburg DS, Ruiz-Palacios GM. a 1–2 Fucosylated chains (H-2 and Lewisb) are the main human milk receptor analogs for Campylobacter. Pediatr Res 37:171A, 1995. 11. Stahl B, Thurl S, Zeng J, Karas M, Hillenkamp F, Steup M, Sawatzki G. Oligosaccharides from human milk as revealed by matrix-assisted laser desorption/ionization mass spectrometry. Anal Biochem 223:218–226, 1994. 12. Lasky LA. Selectin–carbohydrate interactions and the initiation of the inflammatory response. Annu Rev Biochem 64:113–139, 1995. 13. Johnson PH, Watkins WM. Purification of the Lewis bloodgroup gene associated a-3/4-fucosyltransferase from human milk: An enzyme transferring fucose primarily to type 1 and lactose-based oligosaccharide chains. Glycoconjugate J 9:241–249, 1992. 14. Johnson PH, Donald ASR, Feeney J, Watkins WM. Reassessment of the acceptor specificity and general properties of the Lewis blood-group gene associated a-3/4-fucosyltransferase purified from human milk. Glycoconjugate J 9:251– 264, 1992. 15. Eppenberger-Castori S, Lotscher H, Finne J. Purification of the N-acetylglucosaminide a-(1-3/4) fucosyltransferase of human milk. Glycoconjugate J 6:101–114, 1989. 16. Macher BA, Holmes EH, Swiedler SJ, Stults CLM, Srnka CA. Human a 1–3 fucosyltransferases. Glycobiology 1:577– 584, 1991. 17. Narimatsu H. Recent progress in molecular cloning of glycosyltransferase genes of eucaryotes. Microbiol Immunol 38:489–504, 1994. 18. Wiederschain GY, Rosenfeld EL. Two forms of a-L-fucosidase from pig kidney and their action on natural oligosaccharides. Biochem Biophys Res Commun 44:1008–1014, 1971. 19. Wiederschain GY. Multiple forms of human glycosidases and their role in glycoconjugate degradation. Adv Clin Enzymol 2:150–157, 1982. 20. Beyer EM, Wiederschain GY. Human and mammalian fucosidases. In Uspekhi Biol. Khim. (Advances in Biological Chemistry, Russia). Moscow: Nauka, 1982, Vol 23, pp 102– 121. 21. Johnson SW, Alhadeff JA. Mammalian a-L-fucosidases. Comp Biochem Physiol 99B:479–488, 1991. 22. O’Brien JS, Willems PJ, Fukushima H, de Wet JR, Darby JK, Di Cioccio R, Fowler ML, Shows TB. Molecular biology of the alpha-L-fucosidase gene and fucosidosis. Enzyme 38:45–53, 1987. 23. Wiederschain GY, Newburg DS. Human milk fucosyltransferase and a-L-fucosidase activities change during the course of lactation. J Nutr Biochem 6:582–587, 1995. 24. Miller JB, Bull S, Miller J, McVeagh P. The oligosaccharide composition of human milk: Temporal and individual varia-

AID

BMM 2517

/

af02$$$164

07-10-96 15:34:39

219

tions in monosaccharide components. J Pediatr Gastroenterol Nutr 19:371–376, 1994. 25. Patton S, Jensen RG. Biomedical Aspects of Lactation. Oxford: Pergamon Press, 1976. 26. Keenan TW, Patton S. The structure of milk: Implications for sampling and storage. In Handbook of Milk Composition (Jensen RG, Ed.). San Diego: Academic Press, 1995, pp 5– 50. 27. Stewart PS, Puppione DL, Patton S. The presence of microvilli and other membrane fragments in the non-fat phase of bovine milk. Z Zellforsch 123:161–167, 1972. 28. Isaacs CE, Moretz RC. Skim milk membranes in human milk. In Human Lactation: Maternal and Environmental Factors (Hamosh M, Goldman AS, Eds.). New York: Plenum, 1986, Vol 2, pp 617–620. 29. Patton S, Huston GE. Isolation of fat globules from human milk. In Human Lactation: Milk Components and Methodologies (Jensen RG, Neville MC, Eds.). New York: Plenum, 1985, pp 81–83. 30. Goldman AS. Introduction to the International Conference: The effects of human milk upon the recipient infant. In Human Lactation: The Effects of Human Milk on the Recipient Infant (Goldman AS, Atkinson SA, Hanson LA, Eds.). New York: Plenum, 1987, Vol 3, pp 1–7. 31. Lucas A, Morley R, Cole TJ, Lister G, Leeson-Payne C. Breast milk and subsequent intelligence quotient in children born preterm. Lancet 339:261–264, 1992. 32. Chu S-HW, Walker WA. Bacterial toxin interaction with the developing intestine. Gastroenterology 104:916–925, 1993. 33. Alhadeff JA, Janowsky AJ. Human serum a-L-fucosidase. Clin Chim Acta 82:133–140, 1978. 34. Beyer EM, Wiederschain GY. Further evidence of human aL-fucosidase polymorphism. Clin Chim Acta 123:251–259, 1982. 35. Beyer EM, Ivleva TS, Artykova GT, Wiederschain GY. Comparative studies of intracellular activity, secretion and multiple forms spectra of human skin fibroblast a-L-fucosidase in the normal and after sucrose load. Biochem Mol Biol Int 30:367–375, 1993. 36. Beyer EM, Wiederschain GY. Substrate specificity and some properties of free and immobilized animal a-L-fucosidase. Biokhimia (Russ) 5:881–889, 1977. 37. DiCioccio RA, Brown KS. Biosynthesis, processing, and extracellular release of a-L-fucosidase in lymphoid cell lines of different genetic origins. Biochem Genet 26:401–420, 1988. 38. DiCioccio RA, Mahoney CM. Effect of glycosylation inhibitors and acidotropic amines on the synthesis, processing, and intracellular-extracellular distribution of a-L-fucosidase in B-lymphoblastoid cells. Carbohydr Res 197:217–226, 1990. 39. Beyer E, Ivleva T, Artykova G, Wiederschain G. Change of isoforms’ spectra of a-L-fucosidase from human skin fibroblasts in intracellular storage of nonhydrolyzable substances. Biochim Biophys Acta 1270:7–11, 1995. 40. Watkins WM. Glycosyltransferases. Early history, development and future prospects. Carbohydr Res 149:1–12, 1986. 41. Natsuka S, Lowe JB. Enzymes involved in mammalian oligosaccharide biosynthesis. Curr Opinion Struct Biol 4:683– 691, 1994.

bmma

AP: BMM

220

WIEDERSCHAIN AND NEWBURG

42. Field MC, Wainwright LJ. Molecular cloning of eukaryotic glycoprotein and glycolipid glycosyltransferases: A survey. Glycobiology 5:463–472, 1995. 43. Paulson JC, Colley KJ. Glycosyltransferases. Structure, localization, and control of cell type-specific glycosylation. J Biol Chem 264:17615–17618, 1989. [Review] 44. Kleene R, Berger EG. The molecular and cell biology of glycosyltransferases. Biochim Biophys Acta 1154:283–325, 1993. 45. Bach NL, Beyer EM, Wiederschain GY, Bovin NV, Zurabyan SE. A study of properties and substrate specificity of human a-L-fucosidase multiple forms using natural and synthetic fucose-containing oligosaccharides. Bioorgan Chem 7:1024– 1033, 1981. 46. Alhadeff JA, Miller AL, Wenaas H, Vedvick T, O’Brien JS. Human liver a-L-fucosidase. Purification, characterization, and immunochemical studies. J Biol Chem 250:7106–7113, 1975. 47. DiCioccio RA, Barlow JJ, Matta KL. Substrate specificity and other properties of a-L-fucosidase from human serum. J Biol Chem 257:714–718, 1982. 48. Wiederschain GY, Rosenfeld EL. Specificity of pig kidney aL-fucosidase and its action on different fragments of blood group A / H substance. Bull Soc Chim Biol 51:1075–1084, 1969. 49. Carlsen RB, Pierce JG. Purification and properties of an aL-fucosidase from rat epididymis. J Biol Chem 247:23–32, 1972. 50. Dawson G, Tsay G. Substrate specificity of human a-L-fucosidase. Arch Biochem Biophys 184:12–23, 1977.

AID

BMM 2517

/

af02$$$164

07-10-96 15:34:39

51. Hopfer RL, Johnson SW, Masserini M, Giuliani A, Alhadeff JA. Hydrolysis of fucosyl-GM1 ganglioside by purified pelletassociated human brain and human liver a-L-fucosidases without activator proteins or detergents. Biochem J 266:491–496, 1990. 52. Johnson SW, Masserini M, Alhadeff JA. High-performance liquid chromatographic analysis of fucoganglioside hydrolysis by a-L-fucosidase. Anal Biochem 189:209–212, 1990. 53. Wiederschain GY, Kozlova IK, Ilyina GA, Mikhaylova MA, Beyer EM. The use of glycosides of 6- and 8-acylamino-4methylumbelliferone in studies of the specificity and properties of human lysosomal glycolipid hydrolases. Carbohydr Res 224:255–272, 1992. 54. Patton S. Origin of the milk fat globule. J Am Oil Chem Soc 50:178–185, 1973. 55. Neville MC, Neifert MR (Eds.). Lactation: Physiology, Nutrition, and Breast-Feeding. New York: Plenum, 1983. 56. Neville MC. Sampling and storage of human milk. In Handbook of Milk Composition (Jensen RG, Ed.). San Diego: Academic Press, 1995, pp 63–79. 57. Umesaki Y, Okada Y, Matsumoto S, Imaoka A, Setoyama H. Segmented filamentous bacteria are indigenous intestinal bacteria that activate intraepithelial lymphocytes and induce MHC class II molecules and fucosyl asialo GM1 glycolipids on the small intestinal epithelial cells in the ex-germfree mouse. Microbiol Immunol 39:555–562, 1995. 58. Velupillai P, Harn DA. Oligosaccharide-specific induction of interleukin 10 production by B220/ cells from schistosomeinfected mice: A mechanism for regulation of CD4/ T-cell subsets. Proc Natl Acad Sci USA 91:18–22, 1994.

bmma

AP: BMM