Eur. J. Biochem. 209,415-422 (1992) FEES 1992

(c

Substrate specificity of small-intestinal lactase Assessment of the role of the substrate hydroxyl groups Alronso RIVERA-SAGREDO, F. Javier CARADA, OFelia NIETO, Jesis JIMENEZ-BARBER0 and Manuel MARTIN-LOMAS Grupo de Carbohidratos, lnstituto de Quimica Organica, Consejo Superior de Investigaciones Cientificas, Madrid, Spain (Received May 13/July 22, 1992) - EJB 92 0654

Lactase-phlorizin hydrolase is a disaccharidase present in the small intestine of mammals. This enzyme has two active sites, one being responsible for the hydrolysis of lactose. Lactase activity is thought to be selective towards glycosides with a hydrophilic aglycon. In this work, we report a systematic study on the importance of each hydroxyl group in the substrate molecule for lactase activity. For this purpose, all of the monodeoxy derivatives of methyl b-lactoside and other lactose analogues are studied as lactase substrates. With respect to the galactose moiety, it is shown here that HO-3’ and HO-2’ are necessary for hydrolysis of the substrates by lactase. Using these chemically modified substrates, it has been confirmed that lactase does not behave as a typical P-galactosidase, since it does not show an absolute selectivity with respect to substitution and stereochemistry at C4‘ in the galactose moiety of the substrate. However, the glucose moiety, in particular the HO-6, appears to be important for substrate hydrolysis, although none of the hydroxyl groups seemed to be essential. In order to differentiate both activities of the enzyme, a new assay for the phlorizin-hydrolase activity has also been developed.

Small-intestinal disaccharidases in higher animals are located in the luminal membrane of intestinal mucose where they hydrolyze the disaccharides in the diet and the oligosaccharides that arise in the intestinal lumen after a-amiIolysis of starch [l, 21. Lactase is the small-intestinal disaccharidase that splits the major 8-glycoside of the diet, lactose, to give glucose and galactose (Scheme 1). The enzyme has been the subject of considerable investigation because of its involvement in the most frequent genetically based syndrome in man, known as adult-type alactasia or lactose intolerance in the adult resulting in a remarkable decrease in lactase activity in adulthood which affects more than 33% of adult humans [I, 3,4]. A biological clock appears to exist, causing a marked decline of lactase after weaning, except for individuals whose ancestors are dependent on a substantial consumption of milk and milk-derived products. It has been known for some time that, in addition to lactase activity, the enzyme carries a P-glucosidase activity refered to as phlorizin hydrolase [5-81, which is in fact a glycosylceramidase [9]. The enzyme is therefore named lactasephlorizin hydrolase or the intestinal P-glycosidase complex. The cloning and complete sequencing of the cDNA encoding the 8-glycosidase complex from human and from rabbit have recently been carried out [ 101. Evidence indicates the existence Correspondence to M. Martin-Lomas, Grupo de Carbohidratos, Instituto de Quimica Organlca, C. S . I. C., Juan de la Cierva 3, E-28006 Madrid, Spain Ahhreviution LPH, lactase-phlorizin hydrolase. Enzyme. Lactase-phlorizin hydrolase (EC 3.2.1.23). Dedication; To the memory of Professor Alberto Sols.

of two active sites, one for lactase activity, the other for phlorizin-hydrolase activity [I - 31, and both of them are labelled with the affinity label conduritol-B epoxide [lo, 111. This is in agreement with previous findings on the substrate specificity of the P-glycosidase complex, since cellobiose, cellotriose, cellotetraose and even cellulose are hydrolyzed by the enzyme [6]. In contrast, lactulose is not a substrate [12]. However, p nitrophenyl b-galactoside and p-nitrophenyl P-glucoside are similarly hydrolyzed, while phenyl b-galactoside and 2naphthyl 8-galactoside are poor substrates [2]. These results indicate that the aglyconic part of the substrate plays an important role, while the configuration at the C4 of the galactose moiety does not seem to be critical. On this basis, it has been suggested [2] that hydrophilic glycosides, either with the D-gluco or u-galacto configuration, are appropriate as substrates for the lactase site [6], while hydrophobic glycosides, either with the 0-gluco or D-galacto configuration, are suitable as substrates for the phlorizin-hydrolase site [9, 51. We have now carried out a systematic study of the substrate specificity with respect to the lactase site of the bglycosidase complex from sheep small intestine, using chemically modified methyl P-lactoside derivatives. Methyl P-lactoside rather than lactose, has been used as a reference compound in order to simplify the analysis of the reaction mixtures. These chcmically modified substrates have been frequently used in the study of enzyme mechanisms [13 - 161 and have proved extremely useful in probing the combining sites for the elucidation of specific interactions between proteins and carbohydrate ligands [17].

41 6 OH

Lactase OH HO

H2 0

HO ‘ O

h OH o H

+

HO&HO

OH OH

OH

Lactose

Galactose

Glucose

Scheme 1. Hydrolysis of lactose by lactase-phlorizin hydrolase.

compound

lactose 1

6 7 8 9 10 11 12

R’ OH OCH3 OCH, OCH3 OC H3 OC H3 OH OCH3 OH

R2

R3

R5

OH

OH OH OH H OH OCH, OCH, OH OH

CHzOH CHzOH CHzOH CHPOH CH3 CHzOH CHzOH H H

OH H OH OH OH OH OH OH

O ‘H

OCH, R3’

OH compound 1 13

14 15 16

R2’

R3

R4

R5’

OH H OH OH OH

OH OH H OH OH

OH OH OH H OH

CHzOH CHzOH CHzOH CH2OH CH3

HO

OH

Scheme 2. Structure of lactose analogues.

MATERIALS AND METHODS

Source and purification of the enzyme

Materials

Lamb small intestines were used as the source of the enzyme. The intestines were obtained from young lambs immediately after sacrifice in a local slaughterhouse. The intestines were washed with cold 1 M NaC1, sodium phosphate 50 mM NaCI, pH 7.3, then frozen or immediately homogenized. The lactase-phlorizin hydrolase activity was purified according to the method described by Schlegel-Haueter et al. [7] with minor variations. Briefly, the enzymatic activity from the membranes of intestinal-brush-bordm cells was solubilized by proteolytic treatment with papain, then purified by subsequent chromatography on Sepharose 2B, Sephadex G-200 and DEAE-cellulose. The fractions showing lactase activity were collected. Protein concentration was determined by the Lowry method [26] using bovine serum albumin as a standard. The molecular mass was estimated by SDS/PAGE, in the presence of 2-mercaptoethanol, using 7.5% acrylamide gels.

Sepharose 2B, Sephadex G-200 and DEAE-cellulose were obtained from Pharmacia, p-chloromercuribenzoate, Trinder reagent (glucose oxidase kit), galactose dehydrogenase, NAD and papain were obtained from Sigma. Reagents and molecular-mass markers used in SDSjPAGE were products of BioRad .

Substrates and other lactose analogues Lactose, cellobiose, phlorizin, o-nitrophenyl-P-D-galactopyranoside and p-nitrophenyl-P-D-galactopyranoside,2naphtyl-8-D-galactopyranoside, methyl P-glucopyranoside, glucose and galactose were from commercial sources. Phloretin was obtained by acid hydrolysis from phlorizin. The lactose analogues (Scheme 2) methyl P-lactoside (1) [18]; methyl P-cellobioside (2) [19]; methyl a-lactoside (3) [20]; 1,6anhydrolactose (4) [21]; lactal ( 5 ) [22]; methyl 2-deoxy-j?lactoside (6) [23]; methyl 3-deoxy-P-lactoside (7) [23]; methyl 6-deoxy-fi-lactosidc (8) [23]; methyl 3-0-methyl-P-lactoside (9) [23]; 3-0-methyl-fi-lactose (10) [24]; methyl 4 - 0 - P - ~ galactopyranosyl-P-D-xylopyranoside(11) [23] ; 4 - 0 - p - ~ galactopyranosyl-D-xylose (12) [25]; methyl 2’-deoxy-Plactoside (13) [23]; methyl 3’-deoxy-fi-lactoside (14) [23]; methyl 4’-deoxy-P-lactoside (15 ) [23]; methyl 6’-deoxy-j?lactoside ( 3 6) [23] were synthesized as described previously.

Enzymatic assays Lactose hydrolysis was measured using coupled enzymatic assays. The enzyme sample, obtained from chromatographic fractions or the purified preparation, was diluted 3 - 100 times with 50 mM sodium maleate, or 50 mM sodium phosphate or 50mM sodium citrate/phosphate at pH 5.9. Lactose was added to a final concentration of 50mM and the mixture (total volume 50 pl) was incubated for 11)-30 min at 37°C.

41 7 The reaction was stopped by boiling for 2 min. Lactose hydrolysis was determined as the galactose released measured by the galactose-dehydrogenase reaction [27], or as the glucose released, by the glucose-oxidase reaction [28],using a commercial kit ("Trinder" from Sigma) following the instructions of the manufacturer. Phlorizin hydrolysis was determined as the phloretin release. The purified enzyme preparation (0.13 mg protein/ml) in sodium phosphate (50 mM, pH 7.3) was diluted up to 100 times with a warmed phlorizin solution (0.5 pM to 5 mM) in sodium citrate (50 mM, pH 5.9). After 5 min incubation at 37"C, the reaction was stopped by the addition of 200 pl 100 mM trifluoroacetic acid in 30% acetonitrile/water. This mixture (200 pl) was applied to an HPLC system (Waters, NOVA-pack C18 , 30% acetonitrile/lO mM trifluoroacetic acid isocratic, 0.6 mljmin). Phlorizin and phloretin were detected by their absorbance at 300 nm and eluted at 4 min and 8 min, respectively. A calibration curve for phloretin was obtained under the same conditions. Lactose-analogue hydrolysis was measured by gas chromatography of the products as follows. A variable amount of enzyme (3 - 40 pg/ml) was incubated with the corresponding substrate analogue at 1-200 mM final concentration, in 50 mM sodium phosphate, pH 5.9 (other buffers interfered in the subsequent silylation reaction). After 10 min incubation at 37"C, the reaction (25 pl) was stopped and dried by boiling for 15 min. The dry residue was redissolved in pyridine (25 PI), containing 1 mM benzyl P-xyloside as standard, then 25 pl trymethylsilylimidazol was added. The mixture was kept at 60 "C for 30 min, then analyzed by GC (capillary column SE 30, 11 m, temperature gradient 150- 250 "C). The release of galactose or methyl fi-glucoside was measured, depending on the analogue that was being analyzed. Previously, calibration curves of galactose and methyl fi-glucoside were obtained under the same conditions. Preliminary assays were performed with each substrate to determine the amount of enzyme needed to obtain a measurable amount of product. Inhibition experiments Pldorizin-hydrolase activity inhibition

The purified enzymic preparation, diluted to a final concentration of 0.6 pg proteinlml with 50 mM sodium citratej phosphate, pH 5.9, was incubated with phlorizin (10 pM or 500 pM) in the presence of variable amounts of disaccharides (1, 10, 100 mM) at 37°C. The incubation was stopped after 5 min and phloretin was measured as mentioned above. Lactase-activity inhibition

To 20 p1 of a lactose solution in 50 mM sodium phosphate, pH 7.3, 20 pl of solutions of different concentrations of the lactose analogue in the same buffer were added (see figures for final concentrations). The reaction was started by the addition of 20 pl purified enzyme preparation (3 pg/ml final Concentration) in the same buffer. The mixture was incubated for 30 min at 37°C and the reaction was stopped by heating for 2 min at 100°C. The glucose released was measured with the glucose-oxidase kit. Mixed-substrate experiment

Mixtures of lactose and lactal were prepared using the equation:

[substrate A] x K,,

+ [substrate B] x KmA= c,

(1j

where A = lactose, B = lactal, as described by Keleti et al. [29]. Using the independently calculated K, values for lactose and lactal, a c number (c = 410) was chosen in order to use concentrations of both substrates at around their K , values. The purified-enzyme preparation was added to those mixtures in 50 mM sodium phosphate, pH 7.3, at a final concentration of 3 pg protein/ml in 40 pl incubation mixture. After 30 min incubation at 37"C, the reaction was stopped by heating for 2 min at 100°C. The combined rate of hydrolysis was measured as the galactose released by both substrates using the galactose-dehydrogenase method. In all the figures, symbols represent mean values of three determinations and error bars represent the standard deviation of the mean. Error bars not shown are within symbols.

RESULTS AND DISCUSSION

The enzyme was purified following a procedure previously reported [7]. The enzyme was solubilized by proteolysis of the carboxy-terminal-anchoring peptide with papain, thus avoiding the use of detergents. The purified protein had a molecular mass of approximately 140 kDa, as estimated by SDS/PAGE, which is in accordance with the apparent molecular mass of the enzyme from other sources [12]. The lactase activity was obtained after a 100- 150-fold purification with respect to the intestinal homogenate. As some minor bands were observed in the electrophoresis gel, a method developed by A. Dahlqvist [30, 311 was followed to prove that no other intestinal P-galactosidases were present in the preparation. Thus, o-nitrophenyl-fi-u-galactopyranosideand p-nitrophenyl-P-D-galactopyranoside were substrates for the enzymic activity, while 2-napthyl-P-~-galactopyranoside was not a substrate, as should be expected if only lactase activity is present [31]. Equally, p-chloromercuribcnzoate did not inhibit the P-galactosidase activity of this preparation, as has been described for lactase activity [30]. The pH dependence of lactase activity showed a plato of maximal activity over pH 4.5 6, compatible with the pH for maximal activity of the enzyme from other sources (around pH 5 - 6) [7,5]. Furthermore, the preparation contained both lactase and phlorizin-hydrolase activities and cellobiose was also a competent substrate (Table l), as expected for the P-glycosidase complex lactase-phlorizin hydrolase [7, 121. The kinetic parameters obtained for lactose which were a K, of 9 mM and a specific activity of 14.5 pmol . min-l . mg protein (Fig. 1j, were comparable to those of the enzyme from other sources. A K , of 21 mM has been reported for the human enzyme [5]and 40,10,22 and 5.1 mM for those of rat [8], calf [32], monkey [33] and pig [12], respectively. With phlorizin as a substrate, a K , of 2 pM was obtained (Fig. 2). This value is clearly lower than the corresponding K, values reported in the literature of 0.4 mM for the rat enzyme [8] or 0.025mM for the pig enzyme [12]. However, it should be mentioned that, in the case of the porcine enzyme, the assay could not be conducted under linear conditions, and the value reported is an upper limit. We have now developed a more sensitive assay for determining phlorizin hydrolysis, based on the chromatographic separation and ultraviolet-absorbance measurement of the hydrolysis product, phloretin. Using this method, it has been possible to use phlorizin in the low micromolar range and keep the substrate consumption below

418

0

I

I

1

I

20

40

60

80

[Lactose] rnM Fig. I . Hydrolysis of lactose by lactase-phlorizin hydrolase. MichaelisMenten plot of released galactose as a function of lactose concentration. K,, 8.8 2 0.3 mM. See Materials and Methods for experimental details.

20% of the maximum. As expected, this enzymic activity was not inhibited by lactose or cellobiose, but phlorizin was able to inhibit lactose hydrolysis. This is in accordance with the proposed hypothesis for the existence of two different active sites in the enzyme, one for each activity [5, 6, 81. With this enzyme preparation, the substrate specificity of synthetic lactose analogues (Scheme 2), with respect to the lactase activity, was studied. The enzymic assays previously used for measuring lactase activity are based on coupled enzymatic assays [27, 281. In order to avoid differences in the processing, by the coupled enzymes, of thc products released from the hydrolysis of each lactose analogue, it was necessary to develop a more general method that could be applied to all the compounds under similar conditions. We chose a system based on GC, where both products (modified galactoses and glucoses) and substrates, could be detected as their pertrimethylsilyl derivatives. A similar system has been used in the study of the stereochemistry at C1 of the released galactose [34]. The compounds were studied as methyl B-lactoside derivativcs, avoiding, in this way, the anomeric heterogeneity of substrates and products. It has been reported that the a and p isomers of lactose behave differently, the K, for the a anomer being 7.63 mM and that for the fl anomer being 11.38 mM. However, the hydrolysis of the equilibrium mixture followed a Michaelis-type kinetics with an apparent K, of 10mM [32], as the small differenccs in the respective K,,, values were compensated by the differences in the concentration of each isomer at equilibrium. As shown in Fig. 1, lactose hydrolysis also follows Michaelis type kinetics with respect to lactase activity. As a first approach, methyl fl-lactoside and its CI isomer were studied. Although both compounds were substrates, the a isomer seemed to be a slightly better substrate in terms of V,,,/K, (see Table 1). It has been suggested recently that lactose intolerance could be due to the absence of a hypothetical anomerase activity, which would be necessary if lactase were anomer selective [35]. This hypothesis is not supported by the following observations: a-lactose and 8lactose are substrates [32]; @-lactoseis hydrolyzed by lactase, without anomerization at C1 or inversion at Cl’, to p-galactose and a-glucose, respectively [34]; as reported here, the non-anomerizable lactose analogues, methyl a-lactoside and P-lactoside, are both substrates of lactase.

It is evident that modifications at C1 are permissible, and, in fact, it has already been shown that lactase is able to hydrolyze longer oligosaccharides, such as cellotriose, cellotetraose and even cellulose to some extent [12]. To address the issue of the importance of the glucose moiety, a drastic change in the disaccharide structure was considered. The overall shape of 1,6-anhydrolactose, shows some differences compared to that of methyl P-lactoside [36], although it still retains two hydroxyl groups in the D-glucopyranose moiety. When 1,h-anhydrolactose was assayed with lactase, no hydrolysis products could be detected. It should be remembered that lactulose (4-0-~-galactopyranosyl-u-fructose), which, in aqueous solution, is in an equilibrium in which the forms having the furanose conformation predominate [37],is not a substrate for the intestinal P-glycosidase complex [12]. It seems, therefore, that the presence of a polar aglycon bound to galactose, is not a sufficient requisite for lactase substrates, thus suggesting that the D-glucopyranose moiety must be involved in the binding. The 2-desoxy, 3-desoxy and 6-desoxy derivatives were synthesized [23] and assayed with lactase. The HO-2 and HO3 groups seem to have some, although small, influence on substrate recognition, because the K, of the corresponding deoxy derivatives increases, although not significantly, with respect to that of methyl p-lactoside. However, HO-6 is important for recognition, as indicated by the significant increase in the K , for the 6-deoxy derivative (246 mM, Table I). Furthermore, when methyl 4-0-~-~-galactopyranosyl-~-~-xylopyranoside, where the hydroxymethylene group on C-5 is absent, was assayed, a high K,,, (>300 mM) was also obtained. As neither HO-2 nor substitution at position 1 seemed important, a double modification was introduced. When lactal was assayed with lactase, no significant differences, with respect to methyl p-lactoside, were observed. This result indicates that the absence of both hydroxyl groups at positions 1 and 2 is permitted. It should be mentioned that, although the glucopyranosyl ring of lactal should be flattened to adopt the 4H5conformation, the overall shape of the molecule resembles that of the methyl 8-lactoside, as can be determined by conformational analysis [38]. The results for methyl 3-0-methyl-8-lactoside were particularly interesting as well. Comparison of the parameters for methyl-P-lactoside with those for the 3-deoxy derivative and the 3-0-methyl derivative clearly show that replacement of the hydroxyl group at C3 for a methoxy group causes a dramatic effect (Table 1). Thus, although, according to above results, HO-3 may not be directly involved in the recognition events or in the catalytic process, the active site permits, only with difficulty, an increase in the steric volume in this environment. This 3-0-methyl derivative, together with the xyloside derivative, have been important compounds in the development of a non-invasive evaluation method [24,25,39]. In both cases the derivatives 10 and 12 (Scheme 2), which are not substituted at the hydroxyl of C1, behave similarly, both of them being hydrolyzed by lactase, although poorly (Table 1). It is interesting to compare these results with those obtained for p-galactosidase from Escherichia coli [15]. This enzyme is much less selective towards the glucose moiety of the substrate than lactase-phlorizin hydrolase. In fact, the absence of any of the hydroxyl groups from the aglycon part only have minor effects on the hydrolysis of lactose analogues by E. coli 8-galactosidase [15]. A further important question concerns whether these results are meaningful in defining the substrate specificity of the lactase site or if the phlorizin-hydrolase site also contributes

419 Table 1. Substrate specificity of the lactase-phlorizin hydrolase. All enzymic assays were performed at least in duplicate. The kinetic calculations and curve fits were carried out with help of the EnzJiter program [45].For compounds 4 and 14, the compound was neither substrate nor inhibitor; Values given for compounds 8 - 12 are estimates of the kinetic parameters since saturation conditions could not be reached; For compound 13, the compound was not a substrate and behaved as a competitive inhibitor (see Fig. 5). See Scheme 2 for the identification of compounds 1 - 16.

pmol. min- . mg-' 14.50 0.5 0.22 0.002 13.40 0.6 1.36f 0.26 10.03 0.09

mM Lactose Phlorizin

*

1 2 3 4 5 6 7 8 9

-

23.2 k 1.6 50.8 5.4 51.3 f 4.5 246.0 48 153.0 f 25 120.0 f 20 2 300 2 370

* *

10

11 12

13 14

Yo

*

8.8 f 0.3 (2 f 0.06)x 17.4 f 2.6 5.3 2 4.5 k 0.5

-

-

9.90 & 0.33 11.80 f 0.5 16.30 f 0.6 5.70 0.96 0.60

68.3 81.4 112.0 39.3 6.6 4.1 11.0 22.1

1.80

3.20

0.43 0.23 0.32 0.023 0.006 0.005 0.006 0.009

-

-

-

-

-

-

-

3.10& 0.18 4.60 k 0.12

14.4 f 0.6 0.4 f 0.05

16

10

15

20

25

30

0.22 10.50

21.4 31.7

0.0 5

-

-

15

0

1.65 110.00 0.77 0.26 2.23

100.0 1.5 92.0 9.4 69.2

0.1

0.2

0.3

35

[Phlorizin] mM x lo3 Fig. 2. Hydrolysis of phlorizin by lactase phlorizin hydrolase. MichaelisMenten plot of released phloretin as a function of phlorizin concentration. K,, 1.9 f 0.06 mM. See Materials and Methods for experimental details.

to some extent to the hydrolysis of the synthetic substrates. It has been already mentioned that neither lactose nor cellobiose inhibit phlorizin hydrolysis. Methyl P-lactoside, methyl P-cellobioside and lactal were also checked as inhibitors of phlorizin hydrolysis. These compounds showcd some inhibition at high concentrations (100 mM) but preliminary experiments (results not. shown) did not indicate competitive inhibition. Methyl 8-lactoside was also assayed as an inhibitor of the hydrolysis of lactose, which is unequivocally hydrolyzed in the lactase site. A competitive inhibition pattern was obtained (Fig. 3) with a Kiof 14 mM, very close, within experimental error, to the K, of 1 7 m M (Table 1) obtained with methyl P-lactoside alone. These results confirmed that the reference compound, methyl 8-lactoside is mostly hydrolyzed, if not completely, at the lactase site.

1/[Lactose]

(rnM-')

Fig. 3. Methyl p-lactoside inhibition of lactose hydrolysis. LineweaverBurk representation of released glucose from lactose hydrolysis as a function of the lactose concentration in the presence of increasing concentrations of methyl j-lactoside: 0 no methyl B-lactoside added; V 10 mM; V 20 mM; n 32 mM; 0 50 mM. Calculated Ki = 14 f 3mM.

Finally, since according to previous data the hydrophilicity or the hydrophobicity of the aglycon seems to play an important role in determining recognition by either lactase or phlorizin hydrolase site [5, 6, 8, 91, an additional experiment aimed to affirm the validity of the results, with regard to the substrate specificity of the lactase site, was performed with lactal. Lactal lacks the hydroxyl groups at C1 and C2 and it is the less hydrophlic of the substrates used in the present study. Therefore, if it could be proven that lactal was hydrolyzed by the lactase site, it would be reasonable to accept that the methyl p-lactoside analogues used in the present study specifically interact with the lactase site. Thus, a kinetic method elaborated by Keleti et a1 [29] was followed. This method allows one to differentiate whether a single enzyme,

420 [Lactal] rnM 48 14

0

24 I

0.00

0

3

6

9

12

15

0.05

0.10

l/[Lactose]

18

[Lactose] rnM Fig. 4. Mixed-substrates experiment. Released galactose from simultaneous hydrolysis of lactose and lactal is plotted against lactose concentration and lactal concentration. Initial lactose and lactal concentrations were related through Eqn 1 given in Materials and Melhods. Following the critcria described by Keleti et a1 [29], (- - -) corresponds to the calculated theoretical curve when each substrate is processed by a different and independent active site and (-) when both substrates are processed by the same active site.

0.15

0.20

(rnM-’)

Fig. 5. Methyl Z’-deoxy-fXactoside inhibition of lactose hydrolysis. Lineveawer-Burk representation of released glucose from lactose hydrolysis as a function of lactose concentration in the presence of increasing concentrations of methyl 2’-deoxy-fl-lactoside: 0 no methyl 2-deoxy-P-lactoside added; V 10 mM; 7 20 mM; 30 mM. Calculated Ki, 10 2.6 mM.

~

with two active sites, is acting on two different substrates, simultaneously in the same site, or in a different site for each substrate. When this method was applied to appropriate mixtures of lactose and lactal, a linear relationship between the lactose concentration and galactose release was found (Fig. 4). This indicates that both substrates, lactal and lactose, are been hydrolyzed at the same active site. Another approach to differentiate both activities consists in a differential thermal inactivation. This approach has been used successfully with the P-glycosidase complex from other sources [6, 51. In our preparation from sheep, this was not applicable as both active sites showed a similar and surprisingly high thermal stability (75% of both activities still remain after 30 min at 5OOC). In the case of the Simian enzyme, a similar behavior has been observed [33]. All of these methods rely on previous assumptions and therefore full differentiation of both active sites will require their physical separation or independent blocking. In this respect, it is interesting to note that the already-known gene encoding the protein has four segments with internal homology [lo]. The native protein corresponds to two of these segments [lo] and each of the corresponding protein domains has a carboxylate group which is involved in each putative active site [MI. With regard to the P-D-galactopyranosyl moiety, a similar systematic study was carried out. The 2’-deoxy, 3‘-deoxy, 4‘deoxy and 6’-deoxy derivatives were synthesized as described elsewhere [23]. With the 2‘-deoxy derivative, transformation was not observed, within the limits of the GC method, even when high concentrations of enzyme and long incubation times were used. When this compound was assayed as a possible inhibitor of lactase, it inhibited the hydrolysis of lactose as well as of methyl P-lactoside. The Lineveaver-Burk representation (Fig. 5) shows a competitive-inhibition pattern. This is correct, when lactose is used as the substrate, if we consider the approximation made previously that the hydrolysis of an equilibrium mixture of a-lactose and jl-lactose follows

a Michaelis-type kinetics. When methyl P-lactoside was used as a substrate, the same inhibition pattern was observed at low concentrations of both, substrate and inhibitor. However, this experiment could not be completed because, when the GC method was used, the 2’-deoxy derivative decomposed by up to 2% of the total during the sample preparation, giving a product that coeluted with methyl 8-glucoside. When the galactose-dehydrogenase method was used, an unexpected inhibition of the coupled enzyme by high concentrations of methyl b-lactoside interfered with the assay. Fig. 5 indicates that the compound is recognized at the active site of the enzyme, but the absence of the HO-2’ practically blocks hydrolysis. It should be mentioned that, in the case of the P-glucosidase A3 from Aspergillus Wentii, it has been shown that the absence of the corresponding 2-hydroxyl group on the substrate also produces a drastic decrease in the k,,,, with minor modification in the K, [13]. It has been suggested that such a hydroxyl group is necessary to induce the optimal orientation of the catalytic groups on the enzyme with respect to C1’ and the glycosidic oxygen [13]. A similar behavior for the 2’-deoxy derivative has been found using the fi-galactosidase from E. coli [14]. Thus, it has been proposed for this system [14] that the HO-2’ could be involved in the approximation of a carboxylate group, present at the active site [41] and necessary to stabilize the intermediate glycosyloxocarbonium ion or form a covalent galactosyl-enzyme intermediate. In the case of lactase, a similar mechanism which involves an oxocarbonium ion stabilized by a carboxylate has been proposed [lo]. The above result with methy 2’-deoxy-Plactoside allows one to hypothesize a similar role for the HO-2’ on the lactose molecule, with respect to lactase activity, to that proposed in the case of the E. coli enzyme. The 3‘-deoxy derivative was not hydrolyzed either. This compound, however, did not behave as an inhibitor of lactase. Thus, the HO-3’ seems to be a key polar group for recognition and binding. Its absence prevents recognition by the active site of the enzyme and the compound is not bound and, therefore, not hydrolyzed. Again, this behavior is similar to that of P-galactosidase of E. coli [14].

421

.OH

This work ha$ been supported by the Direccibn General de Investigacidn Cientifica y Ticnica (grant PB 87-0367). We thank Professor Juan J. Aragon (Departamento de Bioguimica, Universidad Autdnoma de Madrid) for valuable advice and supervision in the purification of the enzyme and the Ministerio de Educaciiin y Ciencia for fellowships (to A. R.-S. and F. J. C.).

fOH1 Scheme 3. Important hydroxyl groups for lactose hydrolysis by lactase. Hydroxyl groups on C2’ and C3’are essentials; the absence of HO6’ renders a better substrate; the absencc of HO-6 or methylation of 110-3 renders a poor substrates.

REFERENCES

1. Semenza, G. & Auricchio, S. (1989) Small-Intestinal Disaccharidases, in The metabolic basis of inherited diseases (Scrives, C. R., Beaudet, A. L., Sly & Valle, V. S., eds) pp. 2975 -2997, Mc Graw-Hill, New York. 2. Semenza, G. (1987) Glycosidases, in Mammalian ectoenzymes (Kenny, A. I. & Turner, A. J, eds) pp. 265-287, Elsevier, Amsterdam. The HO-4’ is very significant in terms of the absence of selectivity. Contrary to what could be expected for a #?-ga- 3. Kretchmer, N. (1971) Memorial Lecture: Lactose and Lactase A Historical Perspective, Gasfroennterology.61, 805- 813. lactosidase, lactase is not selective towards the axial hydroxyl 4. Simoons, F. J. (1973) Ethnic differences in adult lactose intolcrgroup at C4’. It has been already mentioned that lactase can ance, Am. J . Dig. Dis. 18, 595-611. hydrolyze cellobiose and other glucosides [2, 7, 121. Both 5. Skovbjerg, H., Sjostrom, H. & Noren, 0.(1981) Purification and methyl #?-cellobiosideand the 4’-deoxy derivative were comCharacterisation of Amphiphilic Lactase/Phlorizin Hydrolase from Human Small Intestine, Eur. J. Biochem. 114, 653 -661. petent as substrates (Table 1). In both cases, the compounds 6. Colombo, V., Lorenz-Meyer, H. & Semenza, G. (1973) Small were hydrolyzed less efficiently, in terms of the ratio V,/ Intestinal Phlorizin Hydrolase: The “/I-Glycosidase complex”, K , (Table l), when compared with the reference substrate. Biochim. Biophys. Acta. 327,412-424. However, the decrease in the catalytic efficiency was not as 7. Schlegel-Haueter, S., Hore, P., Kerry, K. R. & Semenza, G. (1972) dramatic as in the case of the #?-galactosidaseof E. coli, where The preparation of Lactase and Glucoamylase of Rat Small the axial HO-4‘ is an essential requirement [14]. Thus, this Intestine, Biochim. Biophys. Acta. 258, 506- 519. lack of selectivity, with respect to the stereochemistry at C4’, 8. Kraml, J., Kolinska, J., Ellederovi, D. & HirSova, D. (1972) confirms that lactase is not a pure P-galactosidase. In fact, 8-Glucosidase (Phlorizin Hydrolase) activity of the lactase fraclactase-phlorizin hydrolase has recently been included, on the tion isolated from small intestinal mucosa of infant rats, and the relationship between fi-glucosidases and p-Galactosidases, basis of sequence homology, in a family of enzymes comprisBiochim. Biophys. Acta. 258, 520- 530. ing several glycosidases, where some of them are cellulases Y. Leese, H. J. & Semenza, G. (1973) On the Tdentity between the and glucosidases [42,43]. Small lntestinal Enzymes Phlorizin Hydrolase and GlycoFinally, the influence of HO-6’ was studied. Previous receramidase, J . Biol. Chem. 248, 8170-8173. ports had shown that o-nitrophenyl-p-D-fucoside and 10. Mantei, N., Villa, M., Enzler, T., Wacker, H., Boll, W., James, p-nitrophenyl-P-D-fucoside were substrates of the lactaseP., Hunziker, W. & Semenza, G. (1988) Complete primary phlorizin hydrolase complex [6, 441, mostly through the structure of the human and rabbit lactase-phlorizin hydrolase: phlorizin site. To compare this result with the activity at the implications for biosynthesis, membrane anchoring and evolother site, the corresponding 6‘-deoxy lactoside derivative (or ution of the enzyme, EMBO J. 7,2705-2713. methyl 4-O-#?-~-fucopyranosy~-~-D-glucopyranoside) was 11. Semenza, G. (1989) The insertion of stalked proteins of the brush border membranes: the state of the art in 1988, Biochem. Int. synthesized and studied as a substrate of lactase. The results 18.15-33. (Table 1) indicate that HO-6 is not important, in fact its absence made of 6‘-deoxylactoside derivative the best sub- 12. Skovbjerg, H., Norkn, O., Sjostrom, H., Danielsen, E. M. & Enevoldsen, B. S. (1982) Further Characterization of Intestinal strate of all the synthetic analogues studied in this work. This Lactase/Phlorizin Hydrolase, Biochim. Biophys. Acta. 707,89 result suggests that the enzyme should present a hydrophobic 97. environment to the C6’ of the substrate. 13. Roeser, K.-R. & Legler, G . (1981) Role of sugar hydroxyl groups A summary of the results is presented in the Scheme 3 . in glycoside hydrolysis. Cleave mechanism of deoxyglucosides The HO-3’ and HO-2’ are outlined as essential polar groups. and related substrates by P-glucosidase A3 from Aspergillus Wentii,Biochim. Biophys. Acta. 657, 321 -333. The first is necessary for binding, and the recognition of the second by the enzyme in the substrate-enzyme complex is 14. Bock, K. & Adelhorst, K. (1990) Derivatives of methyl P-lactoside as substrates for and inhibitors of fi-o-galactosidase from E. important for substrate hydrolysis. Hydroxyl goups at C6’ coli, Carbohydr. Res. 202, 131- 149. and C 4 are dispensable one by one without important loses 15. Bock, K. & Adelhorst, K. (1992) Synthesis of Deoxy Derivatives on the capacity of the corresponding derivatives to be hyof Lactose and their Hydrolysis by j-Galactosidase from E. drolyzed by lactase. Indeed, the absence of HO-6‘ renders an coli, Acta Chem. Scand. 46. 186-193. improved substrate. The glucopyranosyl moiety, although it 16. Sinnott, M. L. (1990) Catalytic Mechanisms of Enzymic Glycosyl does not present any key polar group, seems to be important Transfer, Chem. Rev. 90, 1171 - 1202. for the lactase activity, especially the hydroxyl groups close 17. Rivera-Sagredo, A,, Solis, D., Diaz-Mauriiio, T., Jimenez-Barbero, J. & Martin-Lomas, M. (1991) Studies on the molecular to the glycosidic bond. Thus, the HO-6 has some importance recognition of synthetic methyl p-lactoside analogs by ricin, a as its removal produces a very poor substrate and the enzyme cytotoxic plant lectin, Eur. J . Biochem. 197,217 - 228. does not easily accept bulky substitutions on C3, as can be 18. Bhatt, R. S., Hough, L. & Richardson, A. C. (1977) The Chemisobserved when the HO-3 is methylated. try of Cellobiose and Lactose. Part 7. Selective Benzoylation The results described here will be very helpful in the design of Methyl ,&Lactoside, .I. C. S . Perkin I, 2001 -2005. and synthesis of new lactose analogues, useful for the develop- 19. Newth, F. H., Nicholas, S. D., Smith, F. & Wiggings, L. F. (1949) ment of non-invasive diagnostic methods for lactose intoler3,6,3’,6‘-Dianhydro derivatives of fi-methylcellobioside and of ance in children and adults. 8-methylmaltoside, J. Chem. Soc. 2550 - 2553.

422 20. Fernandez-Mayoralas, A,, Bernabk, M. & Martin-Lomas, M. (1988) The regioselectivity of dibutylstannylene-mediated oxidation of methyl 3’,4,-o-isopropylidene-cc-and B-lactoside. A new synthesis of N-acetyllactosamine, Tetrahedron 44, 4877 4882. 21. Tejima, S. & Chiba, T. (1973) Chemical modification of lactose. 11. Synthesis of 6-acetamide-6-deoxylactose derivatives, Chem. Pharm. BUN.21, 546-551. 22. Haworth, W.N.,Hirst, E. L., Plant, M. M. T. & Reynolds, R. J. W. (1930) The Structure of Carbohydrates and their Optical Rotatory Power. Part 111. 4-Galactoside-a-mannose and its Derivatives, J . Chem. Soc. 2644-2653. 23.Rivera-Sagredo, A., Jimenez-Barbero, J., Martin-Lomas, M., Solis, D. & Diaz-Mauriiio, T. (1992) Studies on the molecular recognition of synthetic methyl /3-lactoside analogues by Ricinus communis agglutinin, Carbohydr. Res., in thc press: 24. Fernhdez-Mayoralas, A,, iMartin-Lomas, M. & Villanueva, D. (1985) 3-O-~-Galactopyranosyl-3-O-methyl-~-glucose: a new synthesis and application to the evaluation of intestinal lactase, Carbohydr. Res. 140, 81 -91. 25. Rivera-Sagredo, A., Fernandez-Mayoralas, A., Jimenez-Barbero, J., Martin-Lomas, M., Villanueva, D. & Aragon, J. J. (1992) 4-O-~-u-Ga~actopyranosyl-~-xy~ose: A new synthesis and application to the evaluation of intestinal lactase, Carbohydr. Res. 228, 129-135. 26. Lowry, 0.H., Rosebrough, N. J., Farr, A. I. & Randall, R. J. (1953) Protein measurement with the Fohn phenol reagent, J . Biol. Chem. 193, 265 - 275. 27. Distler, J. J. & Jourdain, G . W. (1973) The Purification and Propcrties of /3-Galactosidase from Bovine Testis, J . Biol. Chem. 248,6772 - 6780. 28. Dahlqvist, A. (1964) Method for Assay of Intestinal Disaccharidases, Anal. Biochem. 7, 18-25. 29. Keleti, T.,Leoncini, R., Pagani, R. & Marinello, E. (1987) A kinetic method for distinguishing whether an enzyme has one or two active sites for two different substrates. Rat liver Lthreonine dehydratase has a single active site for threonine and serine, Eur. J . Biochem. 170, 179 - 183. 30. Asp, N.-G. & Dahlqvist, A. (1972) Human Small Intestinal 8Galactosidases: Specific Assay of Three Different Enzymes, Anal. Biochem. 47, 527-538. 31. Asp, N.-G. & Dalhqvist, A. (1974) Intestinal P-Galactosidases in Adult Low Lactase Activity and in Congenital Lactase Deficiency, Enzyme 18, 84- 102. 32. Wallenfels, K.& Fischer, J. (1960) Untersuchungen iiber milchzuckerspaltende Enzyme. X. Die Lactase des Kalberdarms, Zeitsch. Physiol. Chem. 321, 223 - 245.

33. Ramaswamy, S.& Radhakrishnan, A. N. (1975) Lactase-Phlorizin Hydrolase Complex from Monkey Small Intestine. Purification, Properties and Evidence for Two Catalytic Sites., Biochim. Biophys. Acta. 403,446-455. 34. Semenza, G., Curtius, H.-C., Raunhardt, 0. & Muller, M. (1969) The Configurations at the Anomeric Carbon of the reaction Products of some Digestive Carbohydrases, Carbohydr. Res. 10,417-428. 35. Mbuyi-Kalala, A., Perraudin, J. P. & Prieels, J. P. (1990) Anomerization and Hydrolysis of Lactose by P-Galactosidase from Saccharomyces lactis, Archv. Biochem. Biophys. 227, 434 -438. 36. Rivera-Sagredo, A. & Jimenez-Barbero, J. (1993) The conformation of 1,6-anhydrolactose and its hexaacetate in solution, Carbohydr. Res. 215,239-250. 37. Angyal, S. J. (1984) The composition of reducing sugars in solution, Adv. Carbohydr. Chem. Biochem. 42, 15 - 68. 38. Rivera-Sagredo, A. & Jimenez-Barbero, J. (1992) The solution conformation of lactal and its hexa-0-acelyl derivative, J . Carbohydr. Chem. in the press. 39. Martinez-Pardo, M., Montes, P. G., Martin-Lomas, M. & Sols, A. (1979) Intestinal lactase evaluation in vivo with 3methyllactose, FEBS Left.98, 99- 102. 40. Trimbur, D. E.,Warren, R. A. J. & Withers, S.G . (1992) Regiondirected mutagenesis of residues surrounding the active site nucleophile in B-glucosidase from Agrobacterium ,faecalis, J . Biol. Chem. 267,10248-10251. 41. Herrchen, M. & Legler, G. (1984) Identification of an essential carboxylate group a t the active site of lacZ P-galactosidase from Escherichia coli, Eur. J . Biochem. 138, 527- 531. 42. Grabnitz, F., Seiss, M., Rucknagel, K. P. & Staudenbauer, W. L. (1991) Structure of the P-glucosidase gene bglA of Clostridiurn thermocellum. Sequence analysis reveals a superfamily of cellulases and jl-glycosidases including human lactase/phlorizin hydrolase, Eur. J. Biochem. 200, 301 - 309. 43. Hcnrissat, B. (1991) A classification of glycosyl hydrolases based on amino acid sequence similarities, Biochem. J. 280,309 - 316. 44. Lau, H.K. F.(1987) Physicochemical characterization of human intestinal lactase, Biochem. J . 241, 567 - 572. 45. Leatherbarrow. R. J. (1987) Enzfitter, Elscvier-BIOSOFT, Cambridge, UK.