Glycobiology

This Article Abstract FREE Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (19) Request Permissions Disclaimer Google Scholar Articles by Boeggeman, E. Articles by Qasba, P. K. Search for Related Content PubMed PubMed Citation Articles by Boeggeman, E. Articles by Qasba, P. K. Social Bookmarking          What's this?

Glycobiology, 2002, Vol. 12, No. 7 395-407
© 2002 Oxford University Press

Studies on the metal binding sites in the catalytic domain of ß1,4-galactosyltransferase

Elizabeth Boeggeman^2,3 and Pradman K. Qasba^1,2

² Structural Glycobiology Section and ³ Intramural Research Support Program-SAIC, Laboratory of Experimental and Computational Biology, NCI-CCR, Building 469, Room 221, Frederick, MD, 21702-1201, USA

Received on December 6, 2001; revised on February 6, 2002; accepted on February 24, 2002

	Abstract

Top Abstract Introduction Results Discussion Materials and methods Acknowledgments Abbreviations References

 
The catalytic domain of bovine ß1,4-galactosyltransferase (ß4Gal-T1) has been shown to have two metal binding sites, each with a distinct binding affinity. Site I binds Mn²⁺ with high affinity and does not bind Ca²⁺, whereas site II binds a variety of metal ions, including Ca²⁺. The catalytic region of ß4Gal-T1 has DXD motifs, associated with metal binding in glycosyltransferases, in two separate sequences: D²⁴²YDYNCFVFSDVD²⁵⁴ (region I) and W³¹²GWGGEDDD³²⁰ (region II). Recently, the crystal structure of ß4Gal-T1 bound with UDP, Mn²⁺, and -lactalbumin was determined in our laboratory. It shows that in the primary metal binding site of ß4Gal-T1, the Mn²⁺ ion, is coordinated to five ligands, two supplied by the phosphates of the sugar nucleotide and the other three by Asp254, His347, and Met344. The residue Asp254 in the D²⁵²VD²⁵⁴ sequence in region I is the only residue that is coordinated to the Mn²⁺ ion. Region II forms a loop structure and contains the E³¹⁷DDD³²⁰ sequence in which residues Asp318 and Asp319 are directly involved in GlcNAc binding. This study, using site-directed mutagenesis, kinetic, and binding affinity analysis, shows that Asp254 and His347 are strong metal ligands, whereas Met344, which coordinates less strongly, can be substituted by alanine or glutamine. Specifically, substitution of Met344 to Gln has a less severe effect on the catalysis driven by Co²⁺. Glu317 and Asp320 mutants, when partially activated by Mn²⁺ binding to the primary site, can be further activated by Co²⁺ or inhibited by Ca²⁺, an effect that is the opposite of what is observed with the wild-type enzyme.

Key words: aspartate motifs/bovine ß4Gal-T1/catalytic domain/metal-binding sites/substrate binding

	Introduction

Top Abstract Introduction Results Discussion Materials and methods Acknowledgments Abbreviations References

 
Members of the Golgi-resident glycosyltransferase family of enzymes transfer the sugar residue from a sugar-nucleotide donor to the oligosaccharide moiety of glycoproteins and glycolipids. These Golgi enzymes are membrane-bound and have a type II membrane protein topology consisting of a cytoplasmic tail, a hydrophobic transmembrane domain, a stem region, and a globular catalytic domain that faces the Golgi lumen (Paulson and Colley, 1989). ß1,4-Galactosyltransferase (ß4Gal-T1) from milk catalyzes the transfer of galactose from UDP--D-galactose (UDP-Gal) to either free- or glycan-bound N-acetylglucosamine (GlcNAc), producing either free- or oligosaccharide-bound GlcNAc (Galß1-4GlcNAc) (Hill, 1979). With -lactalbumin (LA), a mammary gland–specific protein, it forms the lactose synthase complex that transfers galactose from UDP-Gal to free glucose to produce lactose (Galß1-4Glc) (Morrison and Ebner, 1971a).

Although the bovine ß4Gal-T1 was the first glycosyltransferase to be isolated and cloned (Narimatsu et al., 1986; D'Agostaro et al., 1989; Shaper et al., 1986; Masibay and Qasba, 1989; Russo et al., 1990), it was not until recently that a family (ß4Gal-T2 to ß4Gal-T7) of related human and mouse ß4Gal-Ts was identified (Lo et al., 1998; Amado et al., 1999). These glycosyltransferases differ in the oligosaccharide acceptor specificities. Only ß4Gal-T1 and ß4Gal-T2 have lactose synthase activity in the presence of LA (Almeida et al., 1997). ß4Gal-T4 has a weak interaction with LA and is involved in the synthesis of glycosphingolipids (Schwientek et al., 1998). ß4Gal-T5 has 37% sequence identity to human ß4Gal-T1 but has no lactose synthase activity in the presence of LA (Sato et al., 1998a,b). ß4Gal-T6, which has a 94% sequence identity with the rat lactosylceramide synthase and transfers galactose from UDP-Gal to glucosylceramide, shows 39% sequence identity to the catalytic domain of the mouse ß4Gal-T1 (Nomura et al., 1998). The newly discovered ß4Gal-T7 shows 38% sequence identity to the Caenorhabditis elegans sqv-3 gene (Okajima et al., 1999) and is reported to be involved in the synthesis of proteoglycans (Okajima et al., 1999; Almeida et al., 1999).

The specificities of glycosyltransferases are determined not only with respect to donor and acceptor substrates but also with regard to the nature of the glycosidic linkage ( or ß) that they form. The ß4Gal-T family members are classified as "inverting" enzymes because they transfer galactose from UDP--D-galactose to GlcNAc with an "inversion" of the configuration at the anomeric carbon atom of galactose, generating a ß1-4-linked product (Sinnott, 1990). Bovine milk ß4Gal-T1 requires Mn²⁺ for the enzymatic reaction (Morrison and Ebner, 1971b; Khatra et al., 1974; Powell and Brew, 1974; Bell et al., 1976). Kinetic studies showed that the enzyme has two metal binding sites and that the occupation of both sites is required for maximum activation (Powell and Brew, 1976). One of the sites binds Mn²⁺ with high affinity but does not bind Ca²⁺. The other is a low-affinity site that can also bind Ca²⁺ (Powell and Brew, 1976; O'Keeffe et al., 1980). Investigations of the metal ion requirements of the enzyme have shown that Zn²⁺, Cd²⁺, Fe²⁺, and Co²⁺ can also induce activation with lower activities at saturation than that of Mn²⁺. The DXD motifs, conserved in the family members of the glycosyltransferases (Breton et al., 1998; Yuan et al., 1997; Boeggeman and Qasba, 1998; Wiggins and Munro, 1998; Busch et al., 1998; Shibayama et al., 1998, 1999; Zhang et al., 1999; Hodson et al., 2000; Unligil et al., 2000), have been shown to participate in metal binding and catalysis (Boeggeman and Qasba, 1998; Wiggins and Munro, 1998; Busch et al., 1998; Shibayama et al., 1999; Zhang et al., 1999; Hodson et al., 2000; Unligil et al., 2000).

In the catalytic domain of the ß4Gal-T family members, two sequence regions, D²⁴²YDYNCFVFSDVD²⁵⁴ (region I) and W³¹²GWGGEDDD³²⁰ (region II), contain the DXD motifs. The D²⁴²YD²⁴⁴ in region I is not a conserved DXD motif in all the ß4Gal-T family members; on the other hand, D²⁵²VD²⁵⁴ and E³¹⁷DDD³²⁰ are highly conserved motifs. The kinetic mechanism of ß4Gal-T1 has been suggested to be an equilibrium-ordered mechanism, where UDP-Gal and Mn²⁺ form a complex with the enzyme prior to the binding of the sugar acceptor (Khatra et al., 1974; Powell and Brew, 1974). It has also been suggested that the binding of the substrates to the enzyme–Mn²⁺ complex occurs via a random equilibrium mechanism (Bell et al., 1976). The interaction of the enzyme–Mn²⁺ complex with UDP-Gal has been studied by circular dichroism, and the data indicate that a conformational change occurs on the binding of UDP-Gal to the enzyme-Mn²⁺ complex (Geren et al., 1975) protecting one Trp residue from UV light–induced destruction (Clymer et al., 1976).

Although the crystal structure of the catalytic domain of bovine ß4Gal-T1, with and without UDP, has been determined at a 2.4 Å resolution (Gastinel et al., 1999), it did not provide the exact location of the sugar acceptor or that of Mn²⁺. In our laboratory, we have determined the crystal structure of the complexes of bovine ß4Gal-T1 and mouse -LA in the presence of Mn²⁺ with various acceptor substrates, at 2 Å resolution (Ramakrishnan and Qasba, 2001). These show that the Mn²⁺ ion is coordinated with five atoms; two of these are the pyrophosphate ionic oxygen atoms of the sugar nucleotide, and the remaining three atoms are the C of Asp254, the N2 atom of His347, and the S atom of Met344 of ß4Gal-T1. The structures reveal a conformational change to have occurred within the region comprising residues 345–365 of ß4Gal-T1 from the structure obtained by Gastinel et al. (1999) (conformation I) to the structure (conformation II) we found (Ramakrishnan and Qasba, 2001). This change positions His347 of the region in such a way that it can bind the Mn²⁺ ion, which subsequently can bind to the phosphates of the UDP-Gal. The second metal binding site was never identified in the LS, UDP-Gal and Mn²⁺ complex even though 17 mM MnCl₂ was used during crystallization. The sequence region II, W³¹²GWGGEDDD³²⁰, has a loop structure that is similar in conformations, I and II, with subtle differences. In conformation II, the side chain orientation of Trp314 allows it to interact with GlcNAc. Asp318 and Asp319 of the E³¹⁷DDD³²⁰ sequence motif, localized in the loop structure, are also involved in GlcNAc binding. Asp320 does not appear to directly bind GlcNAc (Ramakrishnan and Qasba, 2001).

In the present study, using site-directed mutagenesis we investigated the involvement of residues Asp242, Asp244, Asp252, Asp254, His347, and Met344 in metal-ion binding and catalysis. Mutation of any of these residues, except Asp242 and Asp244, affect the activity of the enzyme. The kinetic properties of the wild-type and mutated ß4Gal-T1 show that among the residues Asp254, His347, and Met344, which coordinate Mn²⁺ (Ramakrishnan and Qasba, 2001), only Asp254 and His347 are important for metal binding. Met344 can be replaced with alanine or glutamine with only a partial loss of enzyme activity. In addition, Glu317 and Asp320 mutants show an effect on their activities that is opposite to the wild-type enzyme when partially activated by Mn²⁺ and subsequently treated with Ca²⁺ or Co²⁺ ions.

	Results

Top Abstract Introduction Results Discussion Materials and methods Acknowledgments Abbreviations References

 
Expression and site-directed mutagenesis of ß4Gal-T1
The catalytic domain of the N-terminal deleted bovine ß4Gal-T1, residues 130–402 (recombinant ß4Gal-T1), and its mutants were expressed in Escherichia coli as nonfusion proteins with an 11 amino acid N-terminal extension using standard procedures (Ausubel et al., 1987) as previously described (Boeggeman et al., 1993, 1995). The final yield of the purified folded mutant proteins was comparable to the yield of recombinant ß4Gal-T1. Sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) analysis of the purified and refolded mutant proteins showed a single protein band corresponding to a molecular weight of 33 kDa of ß4Gal-T1. Secondary structure analysis by Fourier-transformed infrared spectroscopy (Keiderling et al., 1999) showed that the mutants and the wild-type ß4Gal-T1 have the same secondary structure (data not shown).

Involvement of residues Asp254, Met344 and His347 in metal binding
In the absence of metal ions, ß4Gal-T1 does not catalyze the transfer of galactose from UDP-Gal to GlcNAc. The structure of ß4Gal-T1·UDP·LA·Mn²⁺ (Ramakrishnan and Qasba, 2001) showed that of the two Asp residues of the D²⁵²VD²⁵⁴ motif, only the carboxylate oxygen of Asp254 coordinates with Mn²⁺. On the other hand, Asp252 binds only to the galactose moiety of UDP-Gal in the donor substrate binding pocket of ß4Gal-T1 (Ramakrishnan et al., 2002). Also, in the ß4Gal-T1·UDP-Glc·LA·Mn²⁺ complex, Asp252 binds to the glucose moiety of UDP-Glc (Ramakrishnan et al., 2001). Although the mutation of Asp252 completely abolishes the catalytic activity of the enzyme, the mutant D252N binds to an UDP-agarose column like the enzymatically active wild type. Also, the mutants of Asp254, D254E and D254N barely retain 0.01% of the catalytic activity of ß4Gal-T1 (Table I). Due to the low levels of activity, the kinetic parameters for these mutants could not be obtained with confidence. The results are consistent with the finding that these residues, Asp252 and Asp254, are highly conserved among all the members of the ß4Gal-T family (Breton et al., 1998).

View this table: [in this window] [in a new window]   Table I. Metal ion-dependent activity of the mutants of the primary metal binding site of ß4Gal-T1

 
M344A mutant, in the presence of Mn2+, retained 54% of the wild-type enzyme activity, whereas the mutant M344Q retained only 15% of the activity (Table I). In the presence of Co²⁺, however, M344Q was fourfold more active than ß4Gal-T1. Comparison of the sequences of the catalytic domain of the family members of ß4Gal-T shows that Met344 is only present in Gal-T1, T2, T3, and T4 (Lo et al., 1998), and is substituted by serine in other ß4Gal-T family members. The substitution of His347 either by glutamic acid, glutamine, aspartic acid, or asparagine abolishes the activity of the enzyme almost completely when Mn2+ was used (Table I). However, Co²⁺ then becomes the preferred metal ion when His347 is substituted either by glutamic acid, glutamine, asparagine or aspartic acid (Table I). Comparison of the sequences of this region among the family members of ß4Gal-T shows that His347 is also a highly conserved residue.

Effect of Mn²⁺ ion on the enzymatic activity of Met344 mutants
Enzyme activities of the wild-type and the Met344 mutants were measured at Mn²⁺ concentrations from 0 to 5 mM and at fixed concentrations of GlcNAc and UDP-Gal. The Mn2+ ion–dependent activation of M344A and M344Q showed a sigmoidal behavior (Figure 1). At least 20 µM Mn²⁺ is required to detect the activation of Met mutants, compared to about 4 µM for the wild-type enzyme. The mutants clearly require a higher Mn2+ concentration for activation than the wild-type protein (Figure 1). The enzymatic activities of ß4Gal-T1 and M344A at Mn²⁺ concentrations from 0 to 5 mM fitted well to a rate equation describing two metal binding sites, I (high affinity) and II (low affinity). Mutant M344A, compared to the wild-type, shows a fourfold lower binding affinity for site I and no effect on the affinity for site II (Table II). The data for M344Q did not fit well to Equation 3 (see Materials and methods), which describes two metal binding sites. However, the rate equation for a single metal binding site, Equation 4 (see Materials and methods), gave a better fit (Table II), showing higher affinity for Mn²⁺ but with a sevenfold reduction in the associated velocity.

View larger version (18K): [in this window] [in a new window]   Fig. 1. Mn2+-dependent enzyme activity of the wild-type ß4Gal-T1 and mutated proteins. Assays were carried out in duplicates at fixed concentrations of GlcNAc (25 mM) and UDP-Gal (0.2 mM) and at varied concentrations of Mn2+. The insert shows the sigmoidal dependence of enzyme activity at low concentrations of Mn2+ (0–40 µM) for ß4Gal-T1 and the methionine mutants M344A and M344Q. The enzymatic activity of the histidine (H347X) mutants is very low compared to ß4Gal-T1.

 

View this table: [in this window] [in a new window]   Table II. Mn2+ or Co2+ activation parameters of ß4Gal-T1 and mutated proteins

 
Effect of Co on the enzymatic activity of M344 and H347 mutants
We tested Co²⁺, Zn²⁺, and Cd²⁺ for the activation of the wild-type and mutant enzymes and found that Co2+ activated the mutants more than the other metal ions. The Met344 and His347 mutants are activated to some extent by Co²⁺ (Table I). In the presence of either Mn²⁺ or Co²⁺, His347 mutants show drastically reduced enzymatic activity. However, the mutants H347E, H347Q, and H347N show four- to fivefold increased activity in the presence of Co²⁺ as compared to Mn²⁺ (Table I). The M344A mutant is 10% less efficient in the presence of Co²⁺ than in the presence of Mn²⁺ (Table I). Surprisingly, M344Q is twofold more active in the presence of Co²⁺ than in the presence of Mn²⁺ and is fourfold more active than recombinant ß4Gal-T1. The enzyme activities of mutants were measured at Co²⁺ concentrations from 0 to 5 mM, at fixed concentrations of UDP-Gal (0.2 mM) and GlcNAc (25 mM). Under these conditions, Co²⁺ was observed to enhance the activity of ß4Gal-T1 and of the mutant, M344A, with increasing concentrations up to 250 µM. Above 0.5 mM, it was inhibitory (Figure 2A). In contrast, the histidine mutants show activation when Co²⁺ concentrations were between 200 µM and 5 mM (Figure 2B). The mutant M344Q at Co²⁺ concentration of 3 mM attains at least three times more activity than ß4Gal-T1 before any inhibition in the enzyme activity is observed (Figure 2A). The inhibition by Co²⁺ at higher concentrations may be the result of its binding to more than one site or a denaturation of the proteins at these concentrations. The activity profile with Co²⁺ between the concentrations of 0–400 µM for ß4Gal-T1 and the mutants M344A and M344Q fitted well to a rate equation describing a single metal binding site (Table II). The binding affinity for Co²⁺ (K_d) and the associated velocity (V) were similar for the mutant M344A and the wild-type ß4Gal-T1. In contrast, for M344Q the K_d for Co²⁺ was 38-fold higher. Thus, it appears that the substitution of Met344 with Ala neither affects the dissociation constant (K_d) for Co²⁺ nor the associated velocity (V), whereas substituting it with Gln reduces the Co²⁺ affinity by 42-fold but increases the velocity by 7-fold.

View larger version (19K): [in this window] [in a new window]   Fig. 2. The effect of increasing concentrations of Co2+ on N-acetyllactosamine synthesis. (A) The methionine mutants; (B) the histidine mutants. The insert in A shows the activation of methionine mutants and ß4Gal-T1 at 0–40 µM Co2+. The insert in B shows the activation of the histidine mutants at 0–1 mM Co2+. Enzyme assays were carried out at fixed concentrations of GlcNAc (25 mM) and UDP-Gal (0.2 mM) and varied concentrations of Co2+ (0 µM to 5 mM).

 
Binding of wild-type ß4Gal-T1 and mutants to the UDP-agarose column in the presence of different concentrations of Mn²⁺
To determine the interactions of the residues Met344, Asp254, and His 347 with Mn²⁺ in ß4Gal-T1, we tested the Mn²⁺-dependent binding of the mutants of these residues on UDP-agarose columns and compared the results with that of the enzymatically active protein ß4Gal-T1. The mutant proteins, compared to the wild type, required higher concentrations of Mn²⁺ to bind to the UDP-agarose column. In the presence of 25 mM Mn²⁺, H347D binds to the UDP-column with lesser affinity than the enzymatically active ß4Gal-T1 (Figure 3). Moreover, H347D does not bind to the affinity column if the concentration of the metal is reduced in the equilibration buffer to either 5 or 1 mM (Figure 3). In contrast, M344A binds to UDP-agarose as tightly as the wild type only in the presence of 5 and 25 mM Mn²⁺ and not in the presence of 1 mM Mn²⁺ (Figure 3). The mutant D254N does not bind to the UDP-agarose column at any of the Mn²⁺ concentrations tested (Figure 3). On the other hand, the mutant D252N shows Mn²⁺-dependent binding to the UDP-agarose like the wild type ß4Gal-T1 (Figure 3), suggesting that Asp 252 is not involved in Mn²⁺ binding.

View larger version (41K): [in this window] [in a new window]   Fig. 3. SDS–PAGE analysis of wild-type ß4Gal-T1 and mutated proteins. The renatured proteins of wild-type ß4Gal-T1, M344A, H347D, D254N, and D252N were bound to and eluted from UDP-agarose columns as described under Material and methods. Aliquots of each fraction were analyzed on gels and proteins visualized by Coomassie blue staining. I is the input sample, U is unbound, W is wash, and E is the eluted sample. The amounts of the input sample (I) loaded on the columns were the same under the three equilibration buffers tested, which contained either 1, 5, or 25 mM MnCl2. The wild type remains bound at all concentrations of Mn2+ tested and it is eluted (E) in the presence either of 1, 5 or 25 mM Mn2+. M344A at 1 mM Mn2+ binds less efficiently, and a portion of the protein remains in the unbound and the wash fraction (U and W). H347D does not bind to the UDP-agarose column at either 1 or 5 mM Mn2+ (U and W) and shows reduced binding at 25 mM Mn2+ (U, W, and E). D254N does not bind to the UDP-agarose column (U and W). On the other hand, the mutant D252N binds to the column as the wild type does, and it is eluted (E) in the presence either of 1, 5, or 25 mM Mn2+.

 
UDP-Gal kinetic parameters for ß4Gal-T1 and M344A
The bisubstrate reactions can utilize either sequential (ordered or random) or ping-pong (double displacement) mechanisms, which are distinguished by steady-state analyses in which the concentrations of two substrates are independently varied. Because M344Q and His347 mutants showed lower levels of Mn²⁺-dependent activity and the kinetic parameters could not be determined with confidence, the parameters for only ß4Gal-T1 and M344A were obtained (Table III). Galactose transfer to GlcNAc by ß4Gal-T1 and M344A was measured at a fixed concentration of GlcNAc (25 mM) and at UDP-Gal concentrations varying from 0.025 mM to 0.2 mM and different fixed concentrations of MnCl₂. The double reciprocal plots of velocity versus UDP-Gal at various fixed concentrations of MnCl₂ gave a series of intersecting lines as previously observed (Powell and Brew, 1974) (not shown). In each case, the data fitted best to Equation 1 (see Materials and methods), indicating a symmetrical sequential or random equilibrium mechanism. Table III shows that mutating Met344 to Ala344 does not significantly change the true K_m value for UDP-Gal, whereas the K_m for Mn²⁺ increases about threefold and the k_cat is reduced by a third.

View this table: [in this window] [in a new window]   Table III. Kinetic parameters for ß4Gal-T1 and M344A

 
Mutants of E³¹⁷DDD³²⁰ sequence that affect ß4Gal-T1 enzymatic activity
The E³¹⁷DDD³²⁰ motif is a part of the sequence of region II, W³¹²GWGGEDDD³²⁰, conserved among the ß4Gal-T family members. Site-directed mutagenesis of glutamic acid at position 317 and aspartic acid residues at positions 318, 319, and 320 showed that all these residues are important for catalysis (Table IV). Substituting Asp318 or Asp319 with asparagine, alanine, or glutamic acid abolishes the enzymatic activity almost completely (legend to Table IV); the kinetic parameters for these mutants could not be determined.

View this table: [in this window] [in a new window]   Table IV. Metal ion–dependent activity of the mutants of the putative secondary metal binding site of ß4Gal-T1

 
Effect of Mn²⁺ on the enzymatic activity of mutants Glu317 and Asp320
A change of glutamic acid residue at position 317 lowers the activity of ß4Gal-T1; specifically, a change to Asp reduces the enzyme activity to 9% compared to ß4Gal-T1 (Table IV). Substitution of Glu317 by Ala or Gln abolished the enzymatic activity completely, indicating that both the carboxylate group and the length of the side chain are critical for the activity. A conservative substitution of Asp with Glu at position 320 results in the retention of about 18% of the Mn²⁺-dependent enzymatic activity of the ß4Gal-T1. Introduction of Ala or Asn at the same position lowered the Mn²⁺-dependent enzyme activity by 92% and 99%, respectively. Thus, the correct length of the side chain carrying the carboxylate group at position 320 is also important for activity. The Mn²⁺-dependent activity curves of ß4Gal-T1 and of the mutants Glu317 and Asp320 are also sigmoidal in behavior (Figure 4A). The mutants, E317D, D320E, and D320N, show reduced response at lower metal ion concentration compared to M344 mutants and require higher Mn²⁺ ion concentration to activate the enzyme compared to the wild type (Figure 4A). The mutants are not activated by 20 µM Mn²⁺, whereas the activity of the wild type can be measured even at 4 µM Mn²⁺ concentration (Figure 1). The Mn²⁺ concentration-dependent (0–5 mM) activity data of the mutants, at fixed concentrations of UDP-Gal (0.2 mM) and GlcNAc (25 mM), did not fit well into a rate equation describing the two metal binding sites in the case of E317D and D320E mutants (Table V). The D320N mutant, however, showed five- and eightfold increase in the K₁ and K₂ values, respectively, over the values for ß4Gal-T1 (data not shown). The equation describing a single binding site for Mn²⁺ gave a better fit for all the mutants (Table V). Mutants E317D, D320E, and D320N showed a 4- to 17-fold increase in the metal dissociation constant, K_d, compared to the wild type.

View larger version (19K): [in this window] [in a new window]   Fig. 4. Mn2+- and Co2+-dependent enzyme activity of ß4Gal-T1 and the mutants of the E317DDD320 region. Assays were carried out at fixed concentrations of GlcNAc (25 mM) and UDP-Gal (0.2 mM) and at varied concentrations of MnCl2 (0 µM to 5 mM) (A) or at varied concentrations of CoCl2 (B) (0 µM to 4 mM). The insert in A shows the sigmoidal dependence of enzyme activity at low concentrations of Mn2+ (0–40 µM) on Glu317 and Asp320 mutants.

 

View this table: [in this window] [in a new window]   Table V. Mn2+ or Co2+ activation parameters of ß4Gal-T1 and mutated proteins of E317DDD320 region

 
Effect of Co²⁺ on Glu317 and Asp320 mutants
A change in the glutamic acid residue at position 317, as mentioned, lowers ß4Gal-T1 activity. Mutant E317D retains about 2% of the Mn²⁺-dependent activity in the presence of Co²⁺ (Table IV). A conservative substitution of Asp by Glu at position 320 also retains about 5% of the Mn²⁺-dependent enzymatic activity of ß4Gal-T1. A change to Asn, however, results in a higher enzyme activity in the presence of Co²⁺ than in the presence of Mn²⁺ (Table IV and Figure 4B). The enzyme activities at Co²⁺ concentrations between 0 mM and 4 mM, at fixed concentrations of UDP-Gal (0.2 mM) and GlcNAc (25 mM) did not fit well to the rate equation describing two metal binding sites. The rate equation describing a single metal binding site did fit better to the data (Table V). The K_d values for E317D do not change significantly compared to wild type; however, the metal binding constant, K_d, for D320E and D320N shows an18-fold and 30-fold increase, respectively.

Effect of Co²⁺ and Ca²⁺ at a low Mn²⁺ concentration on the activation of Glu317 and Asp320
It has been shown previously that at low Mn²⁺ ion concentrations, Co²⁺ inhibits the activity of milk ß4Gal-T1 (Powell and Brew, 1976). In the presence of a fixed Mn²⁺ concentration (20 µM), the recombinant ß4Gal-T1 is inhibited with increasing concentrations of Co²⁺ behaving exactly like milk ß4Gal-T1 (Figure 5A). Displacement of Mn²⁺ by Co²⁺ in the wild-type enzyme produces an enzyme–metal complex with lower activity causing enzyme inhibition (Powell and Brew, 1976). With increasing concentrations of Co²⁺, Mn²⁺ is displaced from the high-affinity binding site, thus inhibiting the enzyme (Figure 5A). In strong contrast, Co²⁺ activates the mutants D320E, D320N, and E317D at 100 µM Mn²⁺ (Figure 5A). These results indicate that Co²⁺ is a more effective cofactor for these mutants compared to wild-type ß4Gal-T1.

View larger version (16K): [in this window] [in a new window]   Fig. 5. The effect of Co2+ and Ca2+ at low concentrations of Mn2+ on the activity of ß4Gal-T1 and mutants. The assays were carried out at fixed concentrations of GlcNAc (25 mM) and UDP-Gal (0.2 mM) in the presence of low concentrations of Mn2+ and at varied concentrations of CoCl2 (A) or CaCl2 (B). A shows inhibition of ß4Gal-T1 by Co2+ in the presence of 20 µM Mn2+ and activation of E317D, D320E, and D320N by Co2+ at 100 µM Mn2+. B shows the activation of ß4Gal-T1 and D244N by Ca2+ at 10 µM Mn2+. At higher Mn concentrations (100 µM), Ca2+ inhibits the activity of E317D and D320E.

 
Ca²⁺ has been shown to bind to the low-affinity metal binding site and not to the high affinity binding site of milk ß4Gal-T1 (Powell and Brew, 1976). It cannot by itself stimulate milk ß4Gal-T1 (Powell and Brew, 1976). The presence of low concentrations of Mn²⁺ is essential for Ca²⁺ to stimulate milk ß4Gal-T1 (Powell and Brew, 1976) (Figure 5B). The presence of at least two metal binding sites in recombinant ß4Gal-T1 is confirmed by the observation that at low Mn²⁺ concentrations (10–20 µM), ß4Gal-T1 and the mutant D244N (which behaves like wild type) could be activated by Ca²⁺ (Figure 5B). In contrast, Ca²⁺ did not activate mutants E317D and D320E at low Mn²⁺ concentrations (10–20 µM) (data not shown). In fact, Ca²⁺ had an inhibitory effect on the activity of E317D and D320E at 100 µM Mn²⁺ (Figure 5B). Clearly, in E317D and D320E, the Ca²⁺-binding site is altered, because the metal ion cannot activate the enzyme even when Mn²⁺ is present at low concentrations.

Kinetic parameters for GlcNAc and UDP-Gal for Glu317 and Asp320 mutants in the presence of Mn²⁺
The initial velocity data for ß4Gal-T1 and for the mutants were obtained at saturating concentrations of Mn²⁺ (10 mM) with varying concentrations of UDP-Gal and a series of fixed concentrations of GlcNAc (Table VI). The data were analyzed using the equations (see Materials and methods) that represent the sequential symmetrical initial velocity pattern (Equation 1) or the asymmetric initial velocity pattern (Equation 2). The data for the mutant E317D did not fit well to Equation 1, which is associated with an ordered or random equilibrium mechanism. Equation 2, associated with a ping-pong mechanism, gave a better fit where the value for K_i was zero. The K_m for GlcNAc does not change in the mutant E317D, whereas the K_m for UDP-Gal increases 2-fold compared to ß4Gal-T1 (Table VI). Using Eq. 1, the true K_m values for UDP-Gal (K_md) and GlcNAc (K_ma) were determined for the D320 mutants and compared to the wild type. The K_m for UDP-galactose increased fivefold for D320N, but it decreased to a third in the mutant D320E (Table VI). The equilibrium dissociation constant, K_id, for UDP-Gal increased by about 1.5- to 3-fold in these mutants. The K_m for GlcNAc for the mutants D320E and D320N, compared to the wild type, showed values 11- and 151-fold higher, respectively. The turnover number, k_cat, however, was reduced, resulting in a 240- to 700-fold decrease in the catalytic efficiency (k_cat/K_ma * K_id). k_cat, can be influenced by both the rate-limiting acceptor binding and the product release steps; therefore, the specificity constants (k_cat/K_md * K_ia or k_cat/K_ma * K_id), which combine rate and binding, are a measure of the stability of the transition state complex. Thus, extending the side chain length of Asp320 indirectly affects the binding of the acceptor and the binding of UDP-Gal to a lesser degree. It also affects the interaction between the enzyme and the substrate in the transition state during the catalytic mechanism.

View this table: [in this window] [in a new window]   Table VI. Kinetic parameters for ß4Gal-T1 and E317DDD320 mutants

 

	Discussion

Top Abstract Introduction Results Discussion Materials and methods Acknowledgments Abbreviations References

 
Many glycosyltransferases require exogenous divalent cations for their activities and possess specific binding sites for such ions. The DXD motifs in this class of enzymes were suggested to be involved in the metal binding and were also implicated in sugar nucleotide binding (Boeggeman and Qasba, 1998; Busch et al., 1998; Wiggins and Munro, 1998; Unligil et al., 2000). Alignment of the sequences of the catalytic domain of ß4Gal-T family members shows that DXD sequence motifs are located within the two sequences, region I, D²⁴²YDYNCFVFSDVD²⁵⁴, and region II, W³¹²GWGGEDDD³²⁰ (Figure 6). Mutagenesis and kinetic analysis of the mutants presented herein showed that Asp242 and Asp244 mutants in region I gave similar activities as the wild-type protein, whereas only the D²⁵²VD²⁵⁴ motif from region I (Table I) and E³¹⁷DDD³²⁰ from the sequence region II (Table IV) are important for ß4Gal-T1 activity (Boeggeman and Qasba, 1998;).

View larger version (21K): [in this window] [in a new window]   Fig. 6. Sequence comparisons of the two DXD regions in the catalytic domain of ß4Gal-T family members. Regions I (residues 242–255) and II (residues 311–321) of bovine ß4Gal-T1 and seven human homologs. The conserved residues are boxed and the gaps in the sequence are shown as dashes. The DXD motifs are shown below the black boxes.

 
This study, using kinetics and binding affinity analysis, shows that of the three protein residues, Asp254, Met344 and His347, that have been identified by X-ray crystallography to coordinate the Mn²⁺ ion (Figure 7) (Ramakrishnan and Qasba, 2001), Met344 coordinates the metal less strongly. This situation is similar to the metal binding site of azurins where Cu²⁺ is coordinated to three strong protein ligands (two histidines and one cysteine) and weakly to a fourth ligand (methionine) (Messerschmidt et al., 1998). Substituting Met344 with alanine or glutamine does not completely abrogate the activity of the protein. It reduces the activity by 45% or 85%, respectively. The mutants M344A and M344Q have about a third to one-fifth affinity for manganese at site I (Table III). Also, the mutant M344A, compared to the wild type, required higher concentrations of Mn²⁺ to bind to the UDP-agarose column (Figure 3), showing it has a weaker affinity for Mn²⁺. Although M344A has a lower affinity for Mn2+, the equilibrium dissociation constant for UDP-Gal binding to the enzyme–Mn²⁺ complex, the K_id, is decreased and there is a small but significant decrease in the k_cat (Table III).

View larger version (42K): [in this window] [in a new window]   Fig. 7. Stereo diagram of the flexible loop of ß4Gal-T1 in conformations I and II. The overall structure of ß4Gal-T1 in the two conformations, I and II, is the same and for comparison, the orientation of Tyr 286 and Tyr 289 in the two structures (A and B) are shown (purple ribbon). However, the structures differ in the flexible loop of ß4Gal-T1, residues 345–365 (green ribbon) and the loop region (residues 311–320) (orange ribbon). The E317DDD320 sequence region in the loop is shown in red. A shows the structures of these regions in conformation I (PDB file; 1FGX). The flexible loop is folded in a way that it covers the sugar binding site and does not show the metal binding pocket in ß4Gal-T1. B shows the structure of these regions in conformation II (PDB file; 1J8X) (Ramakrishnan and Qasba, 2001). The flexible loop (green ribbon) folds in a way that it creates the UDP-Gal, GlcNAc, and metal binding sites in ß4Gal-T1. In this conformation, Mn2+ coordinates with UDP through two phosphate ionic oxygen atoms and also coordinates with the side chains of the ligands, D254, H347, and M344. The residues Asp318 and Asp319, localized in the loop region (orange ribbon), bind directly to GlcNAc. In the loop region, Tyr 311 is hydrogen bonded to Asp320 in both conformations, I and II, while the orientation of the side chain of Trp314 is different in the two conformations. A portion of the hydrophobic pocket surrounding the acceptor is shown in purple (residues Arg359 and Phe360).

 
From the metal binding studies on ß4Gal-T1, it is apparent that Co²⁺ and Mn²⁺ bind to site I in a similar way, as evidenced by the K₁ values (Table II). Mn²⁺ binding to site II results in activation (Figure 1), whereas Co²⁺ binding results in an inhibition at higher concentrations of the metal ion (Figure 2A). Mutation of Met344 to Gln344 perturbs the properties by decreasing the specific affinity for Co²⁺ while increasing the velocity V₁ about sevenfold (Tables I and II). A change of Met344 to Gln344 restricts the accommodation of the preferred Mn²⁺ ion at the primary metal binding site. However, the mutation may favor local Co²⁺ coordination in a way that stabilizes the transition state complex during catalysis, increasing the velocity of the reaction. The data show that mutating Asp254 and His347 almost completely abolish the enzymatic activity of the protein (Table I), and the mutants D254N and H347D do not bind to the UDP-agarose column in the presence of 1 and 5 mM Mn²⁺ remaining in the unbound fraction (Figure 3). These results show that these residues are strongly coordinating to Mn²⁺. We crystallized both the mutant proteins, H347E and M344A, in the presence of Co2+, and their structures are being determined.

The E³¹⁷DDD³²⁰ sequence localized in the conserved sequence W³¹²GWGGEDDD³²⁰ (region II) also contains the DXD motif, which is a potential site for the coordination of divalent cations (Boeggeman and Qasba, 1998). Substituting Asp318 or Asp319 with either asparagine or alanine or glutamic acid almost totally abrogated the activity (Boeggeman and Qasba, 1998) (Table IV), and it was not possible to analyze the kinetic parameters for these mutants. The results presented by Zhang et al. (1999) also show that mutating Asp318 with either asparagine or glutamic acid results in 0.1% of the activity of the wild-type protein. This region in the two structures, conformation I and II, also shows differences as evidenced by the side chain orientation of Trp314 (Figures 7A and 7B). In conformation II, the orientation of the side chain of Trp314 is such that it can make hydrophobic interactions with GlcNAc (Figure 7B), as well as with the galactose moiety of UDP-Gal (not shown) and a hydrogen bond with the oxygen atom of the phosphate group of the sugar nucleotide (Ramakrishnan and Qasba, 2001). An earlier study showed that Trp314 is very important for catalysis (Aoki et al., 1990), which is consistent with the structural evidence. The crystal structures of ß4Gal-T1and LA in complex with GlcNAc or Glc show that the carboxylate oxygens of Asp 318 and Asp 319 are hydrogen bonded to the monosaccharide (Ramakrishnan and Qasba, 2001). The carboxylate oxygen of Asp318 forms hydrogen bonds with O3 and O4 of GlcNAc (or Glc) (Figure 7B) (Ramakrishnan and Qasba, 2001). Furthermore, the interactions of N-butanoyl-GlcN with the Gal-T1 molecule in the N-butanoyl-GlcN-Gal-T1-LA complex are quite similar to the interactions of GlcNAc in the Gal-T1-LA-GlcNac complex (Ramakrishnan et al., 2001).

In recent experiments using photoaffinity-labeled GlcNAc, Asp318 has been identified as the residue that binds the sugar acceptor (Hatanaka et al., 2001). Loss of activity associated with the substitution of Asp318 or Asp319 with either asparagine or alanine or glutamic acid is also consistent with the structural evidence (Ramakrishnan and Qasba, 2001) and the mutational data of Zhang et al. (1999). However, the interpretation of the kinetic data presented by Zhang et al. (1999), which states that Asp318 does not directly affect GlcNAc binding, is not consistent with the structural and photoaffinity-labeling data. Mutation of Asp320 also affects the activity; in particular when it is substituted by asparagine. It then prefers Co2+ above Mn2+ as a metal ion. Substitution of Asp320 with glutamic acid, however, results in the retention of about 18% of the Mn²⁺-dependent activity of the wild type (Table IV). The mutation of Asp320 lowers the affinity for UDP-Gal and more strongly reduces the affinity for GlcNAc. This mutation, however, does not affect the metal binding to the high affinity site (K₁) for Mn2+. In the crystal structure, Asp320 is not shown to make any contact with GlcNAc (Figure 7B). The carboxylate group of Asp320 is hydrogen-bonded with Tyr311 in both conformations I and II (Figures 7A and 7B). Substituting Tyr311 with glycine results in an increase in K_m for GlcNAc by 1,000-fold and to a lesser extent (35-fold) the K_m for UDP-Gal (Aoki et al., 1990). The hydrogen bond between Asp320 and Tyr311 is likely to help stabilize the loop structure of the sequence region W³¹²GWGGEDDD³²⁰, the region that is directly involved in GlcNAc binding and catalysis. Because the D320E mutant retains only 18% of the enzyme activity, the carboxylate group of the glutamic acid may still form a weak hydrogen bond with the hydroxyl group of Tyr311, stabilizing (albeit less efficiently) the loop structure.

Though our data are in agreement with the conclusion that the residues of the E³¹⁷DDD³²⁰ motif do not constitute the high-affinity metal binding site (Zhang et al., 1999), they do however show that mutating Glu317 and Asp320 affect the Ca²⁺- or Co²⁺-dependent activation. In the absence of Mn²⁺, Ca²⁺ does not stimulate the enzyme activity of either the wild type or the mutants of Asp242 or Asp244. The wild-type protein and the mutants where the E³¹⁷DDD³²⁰ site has not been altered can however be activated by a Ca2+ ion in the presence of low amounts of Mn2+ (Figure 5B). In contrast, Ca²⁺ inhibits the mutants of the E³¹⁷DDD³²⁰ region, E317D and D320E at higher Mn²⁺ concentrations (Figure 5B). Because the ionic radius of Ca²⁺ is larger (0.99 Å) than that of Mn²⁺ (0.80 Å), Ca²⁺ may not fit in the mutated E³¹⁷DDD³²⁰ region. Co²⁺, on the other hand, may be fitting better in the mutated region because it is seen to activate these mutants when low concentrations of Mn²⁺ are present (Figure 5A). One water molecule has been seen in the center of the loop structure that contains the sequence E³¹⁷DDD³²⁰ hydrogen-bonded to Glu317 and the amino nitrogen of Asn356, which could be a potential second metal ion site (Ramakrishnan and Qasba, 2001). Interestingly, there is some sequence identity between the sequence motif of W³¹⁴GGEDDDIY³²² of ß4Gal-T and the sequence motif W³⁶⁸GGAGKDIF³⁷⁶ of Serratia protease. The latter also adopts a loop conformation in the protein, and has been shown to coordinate a Ca²⁺ ion (Figure 8) (Baumann et al., 1995). At present we do not have any structural data for metal binding to the W³¹⁴GGEDDDIY³²² region, but our data show that E317D and D320E or N behave differently toward Ca²⁺ and Co²⁺ activation compared to the wild type. Because the second metal binding site of ß4Gal-T1 is a low-affinity binding site, either high concentrations of Mn²⁺ or high concentrations of Ca²⁺ with low concentrations of Mn²⁺ may be required during crystallization to trap and localize the second metal ion. The mutants described herein are being investigated further for the localization of the second metal binding site of ß4Gal-T1.

View larger version (21K): [in this window] [in a new window]   Fig. 8. Structures of the putative second metal binding site in ß4Gal-T1 (left) and of the Ca2+ binding site in Serratia protease (right). The conserved sequence region II of ß4Gal-T1, W314GGEDDDIY322, and the sequence of the Ca2+ binding loop of Serratia protease, W368GGAGKDIF376 (PDB file; 1SRP), show similar residues (underlined) and loop structures.

 

	Materials and methods

Top Abstract Introduction Results Discussion Materials and methods Acknowledgments Abbreviations References

 
Restriction enzymes were from New England Biolabs (Beverly, MA). The Plasmid Mini Preparation Kit was from Qiagen (Santa Clarita, CA). Ampicillin , UDP-Gal, and UDP-agarose were from Sigma (St. Louis, MO). pET23a vector and BL21 (DE-3) / pLysS competent cells were from Novagen (Madison, WI). E. coli XL2-Blue ultracompetent cells were from Stratagene (La Jolla, CA). AG1-X8 resin, chloride form, 200–400 mesh was from Bio-Rad (Hercules, CA). Taq DNA polymerase and polymerase chain reaction nucleotide mixes were obtained from Boehringer Mannheim.

Bacterial strains and growth conditions
Bacterial growth and transformation and plasmid isolation were carried out by standard procedures (Ausubel et al., 1987). Plasmid pEGT-d129, which encodes the catalytic domain (residues 130–402) of bovine ß4Gal-T1, was used as the plasmid template for point mutations. Both wild-type and mutant ß4Gal-T1 were expressed in E. coli XL2-Blue ultracompetent cells. All the clones were grown in Luria-Bertani medium (LB) containing 50 µg/ml ampicillin at 37°C (Boeggeman et al., 1993, 1995).

Site-directed mutagenesis
The protocol used for site-directed mutagenesis was from Clontech Laboratories (Transformer Site-Directed Mutagenesis Kit). All the mutants were constructed using the plasmid pEGT-d129, containing a Bam HI / Eco RI fragment of ß4Gal-T1, cloned into pET23a (Boeggeman et al., 1993). The transformation process included a selection primer, pNotApa, which was employed to introduce a unique restriction site in pEGT-d129 for further selection. The mutagenic primers were used to introduce the desired mutations. Amplifications of both the parental and mutated strands were accomplished in a repair-deficient E. coli strain, BMH71-18 muts. After digesting the parental plasmid with the selection-specific restriction enzyme (Not I), the mutant plasmids were amplified and cloned in E. coli XL2-Blue ultracompetent cells and BL21 (DE3)/pLysS competent cells (Boeggeman et al., 1993, 1995). The mutagenic primers were from Genosys Biotechnologies or Recombinant DNA Laboratory (NCI, Frederick). The nucleotide sequences of the mutagenic primers used for mutagenesis are:

• D242N: 5' pGAGGCCTTGAAGAACTATGACTACAACTGC-3'

• D242A: 5' pGAGGCCTTGAAGGCCTATGACTACAACTGC-3'

• D242E: 5' pGAGGCCTTGAAGGAATATGACTACAACTGC-3'

• D244N: 5' pGAGGCCTTGAAGGACTATAACTACAACTGC-3'

• D244A: 5' pGAGGCCTTGAAGGACTATGCCTACAACTGC-3'

• D244E: 5' pGAGGCCTTGAAGGACTATGAATACAACTGC-3'

• D252N: 5' pGCTTTGTGTTTAGCAATGTCGACCTCATCCCA-3'

• D252A: 5' pGTGTTTAGCGCTGTCGACCTCATCCCAATG-3'

• D252E: 5' pGTGTTTAGCGAAGTCGACCTCATCCCAATG-3'

• D254N: 5' pGTGTTTAGCGATGTGAACCTCATCCCAATGA-3'

• D254A: 5' pGTGTTTAGCGATGTCGCCCTCATCCCAATGA-3'

• D254E: 5' pGTGTTTAGCGATGTCGAACTCATCCCAATGA-3'

• E317Q: 5' pGGGGCTGGGGAGGTCAAGATGATGACATTTATAAC-3'

• E317A: 5' pGGCTGGGGAGGTGCAGATGATGACATTTAT-3'

• E317D: 5' pGGCTGGGGAGGTGACGATGATGACATTTAT-3'

• D318N: 5' pGGGGCTGGGGAGGTGAAAATGATGACATTTATAAC-3'

• D318A: 5' pGGCTGGGGAGGTGAAGCTGATGACATTTAT-3'

• D318E: 5' pGGCTGGGGAGGTGAAGAAGATGACATTTAT-3'

• D319N: 5' pGGGGAGGTGAAGATAATGACATTTATAACAGATTAGC-3'

• D319A: 5' pGGCTGGGGAGGTGAAGATGCTGACATTTAT-3'

• D319E: 5' pGGCTGGGGAGGTGAAGATGAAGACATTTAT-3'

• D320N: 5' pGGGGAGGTGAAGATGATAACATTTATAACAGATTAGC-3'

• D320E: 5' pGGGGAGGTGAAGATGATGAAATTTATAACAGATTAGC-3'

• M344A: 5' pGTGATCGGGAAGACGCGCGCGATCCGCCACTCGAGAGAGAAGAAA-3'

• M344Q: 5' pGTGATCGGGAAGACGCGCCAGATCCGCCACTCGAGAGACAAGAAA-3'

• H347E: 5' pGTGATCGGGAAGACGCGCATGATCCGCGAATCGAGAGACAAGAAA-3'

• H347D: 5' pGTGATCGGGAAGACGCGCATGATCCGCGACTCGAGAGACAAGAAA-3'

• H347N: 5' pGTGATCGGGAAGACGCGCATGATCCGCAACTCGAGAGACAAGAAA-3'

• H347Q: 5' pGTGATCGGGAAGACGCGCATGATCCGCCAGTCGAGAGACAAGAAA-3'

The mutated residues are shown in italics. The potential mutants were selected after observing changes in the restriction enzyme patterns. Mutations were confirmed by DNA sequencing using Sequenase version 2.0 (USB, Cleveland, OH).

Preparation and purification of wild-type and mutant proteins
The E. coli XL2-Blue ultracompetent and BL21 (DE3)/pLysS competent cells were transformed with the pET vector derivatives per manufacturer’s protocols. The transformed cells were grown overnight with shaking at 37°C in an LB broth containing 50 µg ml^–1 ampicillin. The overnight cultures were induced with 0.4 mM ispropylthiogalactoside (IPTG) for 4 h. After IPTG induction, the inclusion bodies were isolated from the bacterial pellet essentially as described (Boeggeman et al., 1993). After washing four times with a suspension buffer containing 25% (w/v) sucrose, the inclusion bodies were washed with 1x phosphate buffered saline buffer. From a liter of induced bacterial culture, one generally gets ~80–100 mg of purified inclusion bodies. Our laboratory has developed a protocol to generate active recombinant ß4Gal-T1 from these inclusion bodies using oxido-shuffling agents during renaturation (Boeggeman et al., 1993). On dilution with guanidine-HCl and subsequent dialysis, a portion of the folded protein was found to precipitate in the form of aggregates. To increase the protein yields, we modified this protocol as follows: 100 mg of inclusion bodies were dissolved in 10 ml of 5 M guanidine-HCl containing 0.3 M sodium sulfite. One milliliter of a 50 mM disodium 2-nitro-5-thiosulfobenzoate solution is added to sulfonate the free thiols in the proteins. Completion of sulfonation was judged by a color change from deep orange to pale yellow. The sulfonated protein was diluted 10-fold in water, precipitated, collected by centrifugation at 10,000 x g, and washed four times with water to remove any remaining sulfonating agent. The protein was redissolved in 5 M guanidine-HCl, to a protein concentration of 1 mg/ml, with OD₂₇₅ of 1.9–2.0. The protein solution was diluted 10-fold in a solution containing 100 µg/ml ß4Gal-T1, 50 mM Tris–HCl (pH 8.0), 5 mM ethylenediamine tetra-acetic acid, 0.5M guanidine-HCl, 4 mM cysteamine, and 2 mM cystamine. The protein was renatured for 48 h at 4°C, dialyzed, and concentrated using ultrafiltration membranes (Amicon, Beverly, MA). The advantage of this method is that the precipitated protein can be resulfonated to generate more of active ß4Gal-T1. Using this procedure, the yield was 3–5 mg of pure and active ß4Gal-T1 from 100 mg sulfonated protein per liter of folding solution. The precast Novex gels were used for SDS–PAGE analysis, and the protein bands visualized with Coomassie blue. Protein concentrations were measured with the Bio-Rad protein dye reagent with bovine serum albumin as the standard.

ß4Gal-T1 enzyme assays and kinetic analysis
The in vitro assay for ß4Gal-T1 enzyme activity was performed as described (Boeggeman et al., 1995). The activities were measured using ³H-labeled-UDP-Gal and the sugar acceptor GlcNAc. Typically, a 100-µl incubation mixture contained 25 mM GlcNAc, 5 mM MnCl₂, 0.05 mM UDP-Gal, 5 mM Tris–HCl (pH 8.0), 0.5% Triton X-100, 10–20 ng of ß4Gal-T1, and 0.5 µCi ³H-UDP-Gal. A reaction without GlcNAc was used as a control. The incubation was carried out at 30°C for 15 min and terminated by the addition of 200 µl cold water. The incubation mixture was added to a column containing 0.5 ml bed volume AG1-X8 cation resin and was washed with cold water. The flow-through was diluted in 20 ml Biosafe scintillation fluid and the radioactivity measured in a Beckman counter. The mutated proteins showed reduced but detectable enzymatic activities as compared to the wild-type recombinant ß4Gal-T1.

The kinetic parameters for UDP-Gal and GlcNAc were determined using double substrate kinetics at a fixed Mn²⁺ concentration of 10 mM, with concentrations of UDP-Gal varying between 0.005 and 2 mM and at fixed GlcNAc concentration of 1–50 mM. Data were analyzed for a two-substrate system by fitting to an equation for sequential symmetrical initial velocity pattern, Equation 1 (Zhang et al., 1999), (ordered or random equilibrium mechanism); or to an equation for asymmetric initial velocity pattern, Equation 2 (Zhang et al., 1999), for a double-displacement or ping-pong mechanism.

(1) = (V_max* [A] * [D]) / (K_id * K_md) + (K_md * [A]) + (K_ma * [D]) +

([A] * [D])

(2) = (V_max * [A] * [D]) / (K_ma * [D]) + (K_md * [A]) + ([A] * [D])

In the equations, [D] represents the concentration of the donor; [A] the concentration of the acceptor; K_ma and K_md, the true K_m for the acceptor and the donor, respectively; and K_id the dissociation constant for the donor. The Enzfitter (Biosoft), a nonlinear curve-fitting program for Windows was used to obtain the kinetic parameters from the fitted curves using Equations 1 and 2.

For determining metal binding constants the rate of N-acetyllactosamine synthesis was measured at various concentrations of metal ions at fixed GlcNAc (20 mM) and UDP-Gal concentrations (0.2 mM). Mn²⁺ and Co²⁺ concentrations were varied between 0 and 5 mM. The data were fitted to Equation 3 (Zhang et al., 1999), describing two-metal binding sites, or to Equation 4 (Zhang et al., 1999), describing a single-metal binding site.

(3) = [M] * [(V₁ * K₂) + (V₂ * [M])] / [(K₁ * K₂) + (K₂ * [M] +[M]²)]

(4) = (V * [M]) / (K_d +[M])

The metal binding to site 1 with high affinity (K₁) is associated with low velocity (V₁) and the binding to the low-affinity site 2 with an apparent K_d of K₂ is associated with high velocity (V₂). [M] is the concentration of either Mn²⁺ or Co²⁺.

Binding of recombinant proteins to UDP-affinity column
The binding to the UDP-agarose column was carried out at 4°C. Columns containing 1 ml bed volume of UDP-agarose were preequilibrated with a 25 mM cacodylic buffer, pH 7.6, containing variable amounts of MnCl₂ (25, 5, or 1 mM). The renatured protein solutions (0.5 mg) were adjusted to the preequilibration buffer conditions and applied to the columns. After loading of the samples, the pass-through was recycled again. The pass-through was analyzed for the unbound protein (U). The columns were washed 3x with 1-ml portions of equilibration buffer (W). The elution buffer consisted of 25 mM cacodylic buffer, pH 7.6, containing 25 mM ethylenediamine tetra-acetic acid and 1 M NaCl. For elution of the proteins, 0.5-ml fractions (E) were collected from the columns, and aliquots were analyzed by SDS–PAGE.

	Acknowledgments

Top Abstract Introduction Results Discussion Materials and methods Acknowledgments Abbreviations References

 
We thank B. Ramakrishnan for many helpful suggestions and discussions during the course of this work and S. Kumar and P.V. Balaji for critical readings of the manuscript. The content of this publication does not necessarily reflect the view or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organization imply endorsement by the U.S. government. This project was funded in part with federal funds from the NCI, NIH, under contract no. NO1-CO-12400.

	Abbreviations

Top Abstract Introduction Results Discussion Materials and methods Acknowledgments Abbreviations References

 
Gal-T, galactosyltransferase; GlcNAc, N-acetylglucosamine; IPTG, ispropylthiogalactoside; LA, lactalbumin; LB, Luria-Bertani; SDS–PAGE, sodium dodecyl sulfate–polyacrylamide gel electrophoresis.

	Footnotes

 
¹ To whom correspondence should be addressed; E-mail: qasba{at}helix.nih.gov

	References

Top Abstract Introduction Results Discussion Materials and methods Acknowledgments Abbreviations References

 
Almeida, R., Amado, M., David, L., Levery, S.B., Holmes, E.H., Merkx, G., van Kessel, A.G., Rygaard, E., Hassan, H., Bennett, E., and Clausen, H. (1997) A family of human beta4-galactosyltransferases. Cloning and expression of two novel UDP-galactose:beta-N-acetylglucosamine beta1, 4-galactosyltransferases, beta4Gal-T2 and beta4Gal-T3. J. Biol. Chem., 272, 31979–31991.[Abstract/Free Full Text]

Almeida, R., Levery, S.B., Mandel, U., Kressel, H., Schwientek, T., Bennett, E.P., and Clausen, H. (1999) Cloning and expression of a proteoglycan UDP-galactose:beta-xylose beta1, 4-galactosyltransferase I. A seventh member of the human beta4-galactosyltransferase gene family. J. Biol. Chem., 274, 26165–26171.[Abstract/Free Full Text]

Amado, M., Almeida, R., Schwientek, T., and Clausen, H. (1999) Identification and characterization of large galactosyltransferase gene families: galactosyltransferases for all functions. Biochim. Biophys. Acta, 1473, 35–53.[Medline]

Aoki, D., Appert, H.E., Johnson, D., Wong, S.S., and Fukuda, M.N. (1990) Analysis of the substrate binding sites of human galactosyltransferase by protein engineering. EMBO J., 9, 3171–3178.[ISI][Medline]

Ausubel, F.M., Brent, R., Kinstom, R.E., Moore, D.D., Smith, J.A., Seidman, J.G., and Struhl, K. (1987) Current protocols in molecular biology. Wiley, New York.

Baumann, U., Bauer, M., Letoffe, S., Delepelaire, P., and Wandersman, C. (1995) Crystal structure of a complex between Serratia marcescens metallo-protease and an inhibitor from Erwinia chrysanthemi. J. Mol. Biol., 248, 653–661.[CrossRef][ISI][Medline]

Bell, J.E., Beyer, T.A., and Hill, R.L. (1976) The kinetic mechanism of bovine milk galactosyltransferase.