WO2006003456A2 - Glycosylation of antibiotics - Google Patents

Abstract

It has been demonstrated that macrolide glycosyltransferase enzymes can be recombinantly expressed at high levels. It has also been demonstrated that that these enzymes are able to accept a range of planar, cyclic, hydrophobic molecules in the aglycone binding site such that they may be used to glycosylate a range of substrates. It has been further demonstrated that these enzymes display a plasticity in donor sugar recognition such that they may be utilitised in introducing novel sugar decorations into bioactive molecules. Accordingly, the invention provides a method of attaching a sugar to a substrate comprising contacing the sugar with a macrolide glycosyltransferase enzyme in the presence of a sugar donor. The invention also provides a method of recombinantly expressing a macrolide glycosyltransferase enzyme at a high level comprising expressing a polynucleotide encoding the enzyme in a bacterial host cell.

Description

GLYCOSYLATION OF ANTIBIOTICS

Field of the invention

The invention relates to the glycosylation of various substances. In particular, the invention relates to the glycosylation of antibiotics and sugars using macrolide glycosyltransferases.

Background of the invention

Macrolides constitute an important group of antibiotics that kill primarily Gram-positive prokaryotes. These bioactive molecules are used extensively in the treatment of a range of bacterial infections and collectively have been classed as "the last line of defence" against rapidly emerging strains of multiply resistant pathogens (C. Mendez, J. A. Salas, Trends Biotechnol. 2001, 19, 449). Macrolide antibiotics comprise a macrocyclic backbone to which saccharide moieties are appended. These glycan moieties play a critical role in the function of these antibiotics, and changes to the nature and extent of the sugar decoration alters the activity or even the specificity of these antimicrobial agents. Indeed, glycan alteration is one of the most common resistance mechanisms in the intracellular inactivation of macrolide antibiotics and is used by the Gram-positive streptomycetes that produce these molecules as a protective mechanism from the action of their endogenous antibiotics (L. M. Quiros, I. Aguirrezabalaga, C. Olano, C. Mendez, J. A. Salas, Molec. Microbiol. 1998, 28, 1177). Moreover, it has been suggested that for macrolide antibiotics, in particular, glycosylation state may have an impact on resistance induction (S. Douthwaite, Clin. Microbiol. Infect. 2001, 7, 11). A fuller understanding of these resistance processes is thus vital.

Furthermore, antibiotic glycan alteration or iteration (so-called grycorandomization), which includes decoration of antibiotics with non-natural sugar variants, is also a potentially powerful strategy in combating emerging bacterial resistance (W. A. Barton, J. Lesniak, J. B. Biggins, P. D. Jeffrey, J. Jiang, K. R. Rajashankar, J. S. Thorson, D. B. Nikolov, Nature Struct. Biol. 2001, 8, 545; W. A. Barton, J. B. Biggins, J. Jiang, J. S. Thorson, D. B. Nikolov, Proc. Natl Acad. ScL, USA 2002, 99, 13397; C. Albermann, A. Soriano, J. Jiang, H. Vollmer, J. B. Biggins, W. A. Barton, J. Lesniak, D. B. Nikolov, J. S. Thorson, Org. Lett. 2003, 5, 933; and X. Fu, C. Albermann, J. Jiang, J. Liao, C. Zhang, J. S. Thorson, Nature Biotechnol. 2003, 21, 1467). There are some elegant examples of in vitro glycan modification of antibiotics (largely focussed on cyclic non-ribosomal peptides such as vancomycin) (X. Fu, C. Albermann, J. Jiang, J. Liao, C. Zhang, J. S. Thorson, Nature Biotechnol. 2003, 21, 1467; M. Ge, Z. Chen, H. R. Onishi, J. Kohler, L. L. Silver, R. Kerns, S. Fukuzawa, C. Thompson, D. Kahne, Science 1999, 284, 507; K. C. Nicolaou, S. Y. Cho, R. Hughes, N. Winssinger, C. Smethurst, H. Labischinski, R. Endermann, Chem. Eur. J. 2001, 7, 3798; and H. C. Losey, J. Jiang, J. B. Biggins, M. Oberthur, X.-Y. Ye, S. D. Dong, D. Kahne, J. S. Thorson, C. T. Walsh, Chem. Biol. 2002, 9, 1305). However, the difficulties of performing glycosylation chemistry in richly functionalised molecules, including the need for high chemo- and regio-selectivity, has greatly hampered this process. So much so in fact that although in vivo approaches, so-called biosynthetic engineering, have been explored to date, no such in vitro studies have been performed using macrolide antibiotics (S. Gaisser, J. Reather, G. Wirtz, L. Kellenberger, J. U. Staunton, P. F. Leadlay, MoI. Microbiol. 2000, 36, 391; S. Gaisser, R. LiIl, G. Wirtz, F. Grolle, J. Staunton, P. F. Leadlay, MoL Microbiol. 2001, 41, 1223; L. Tang, R. McDaniel, Chem. Biol. 2001, 8, 547; L. Rodriguez, I. Aguirrezabalaga, N. Allende, A. F. Brana, C. Mendez, J. A. Salas, Chem. Biol. 2002, 9, 111; S. Gaisser, C. J. Martin, B. Wilkinson, R. M. Sheridan, R. E. LiIl, A. J. Weston, S. J. Ready, C. Waldron, G. D. Grouse, P. F. Leadlay, J. Staunton, Chem. Commun. 2002, 618).

In nature, these glycosylation reactions are generally catalysed by Leloir-type glycosyltransferases that use nucleotidediphosphate-activated sugars as glycosyl donor substrates. The observation that glycosylation can modulate the function of these bioactive molecules, indicates that glycosyltransferases can be exploited as "tool kits" to produce biological molecules that display novel carbohydrate decorations, which may confer new activities. The exquisite substrate specificity typically displayed by glycosyltransferases, however, severely curtails their application to only appending the preferred sugar donor to preferred acceptor (D. H. G. Crout, G. Vic, Curr. Opin. Chem. Biol. 1998, 2, 98; M. M. Palcic, Curr. Opin. Biotechnol. 1999, 10, 616; and R. Ohrlein, Top. Curr. Chem. 1999, 200, 227). Indeed, while some antibiotic modifying glycosyltransferases, such as GtfE show a good degree of variance in the sugar substrates that they can use (H. C. Losey, J. Jiang, J. B. Biggins, M. Oberthur, X.-Y. Ye, S. D. Dong, D. Kahne, J. S. Thorson, C. T. Walsh, Chem. Biol. 2002, 9, 1305) they are by the same token more stringent in the molecules to which they transfer sugars. Moreover, attempts to identify other flexible glycosyltransferases have met with limited success and such studies have instead highlighted the often stringent specificity of such enzymes (C. Albermann, A. Soriano, J. Jiang, H. Vollmer, J. B. Biggins, W. A. Barton, J. Lesniak, D. B. Nikolov, J. S. Thorson, Org. Lett. 2003, 5, 933). There is therefore a strong need for glycosyltransferases capable of broad-ranging substrate tolerance as powerful synthetic tools in antibiotic remodelling and for methods that allow their ready identification and characterization.

The glycosyltransferases that catalyse "resistance" glycosylation reactions typically append sugars to the 6-deoxyhexose moieties that are components of antibiotics. Primary examples include the macrolide glycosyltransferase (MGT) from Streptomyces lividans (G. Jenkins, E. Cundliffe, Gene 1991, 108, 55) and two enzymes from the oleandomycin-producing bacterium Streptomyces antibioticus designated OleD and Olel (L. M. Quirόs, J. A. Salas, J. Biol. Chem. 1995, 270, 18234; and L. M. Quirόs, R. J. Carbajo, A. F. Brana, J. A. Salas, J. Biol. Chem. 2000, 275, 11713). Initial assessments have indicated that MGT (G. Jenkins, E. Cundliffe, Gene 1991, 108, 55) and OleD (L. M. Quirόs, R. J. Carbajo, A. F. Brana, J. A. Salas, J. Biol. Chem. 2000, 275, 11713) are possibly non-specific enzymes that are able to transfer glucose (GIc) molecules to a range of macrolides. In contrast, Olel is apparently specific in its transfer of GIc to oleandomycin (G. Jenkins, E. Cundliffe, Gene 1991, 108, 55; L. M. Quirόs, J. A. Salas, J. Biol Chem. 1995, 270, 18234; and L. M. Quirόs, R. J. Carbajo, A. F. Brana, J. A. Salas, J. Biol. Chem. 2000, 275, 11713). These GIc transfers from α-glucosyl uridinediphosphate (UDP-GIc) occur with inversion of anomeric configuration to create β-glucosides, which is entirely consistent with their location in glycosyltransferase family 1 (GT-I; P. M. Coutinho, E. Deleury, G. J. Davies, B. Henrissat, J. MoI. Biol. 2003, 328, 307).

The differing specificity displayed by these GTs suggested that they could be exploited as a "tool kit" for antibiotic remodelling. Little is known, however, about either the full range of aglycone acceptors that these enzymes can utilise, or their specificity for nucleotidediphosphate sugar donors.

Summary of the invention It has been demonstrated that macrolide glycosyltransferase enzymes can be recombinantly expressed at high levels in bacterial cells. This is surprising because the vast majority of glycosyltransferases are normally very poorly expressed in heterologous hosts. The high level of production of these enzymes greatly facilitates their exploitation as tools for remodeling of antibiotic glycosylation on a synthetically useful scale.

It has also been demonstrated that that macrolide glycosyltransferase enzymes are able to accept a range of planar, cyclic, hydrophobic molecules in the aglycone binding site such that they may be used to glycosylate a range of substrates. It has been further demonstrated that these enzymes display a plasticity in donor sugar recognition such that they may be utilitised in introducing novel sugar decorations into bioactive molecules.

Accordingly, the invention provides a method of attaching a sugar to a substrate comprising contacting the substrate with a macrolide glycosyltransferase enzyme in the presence of a sugar donor. The invention further provides: a method of producing a macrolide glycosyltransferase enzyme, which comprises culturing a bacterial cell containing a heterologous polynucleotide encoding the enzyme under conditions in which the enzyme is expressed, and recovering the enzyme; - a bacterial cell which expresses a heterologous polynucleotide encoding a macrolide glycosyltransferase enzyme; and a bacterial expression vector comprising a nucleic acid sequence encoding a macrolide glycosyltransferase enzyme

Description of the Figures

Figure 1 shows typical single-step purification of OleD (2-5 and 7), together with purified Olel (9-12 and 14) and MGT (16-19 and 21). Proteins were subjected to SDS-PAGE using a 12.5 % (w/v) acrylamide gel. The lanes contained the following samples: 1, 8, 13 and 15, 20 μl aliquot of the low molecular weight Dalton Mark VII-L SDS-PAGE standard (Sigma #SDS-7; the M_x (kDa) 66, 45, 36, 29, 24, 20.1 and 14.2); 2, 9 and 16, 4 μl of whole cells expressing OleD, Olel or MGT; 3, 6 10, 17, 20 and 22, soluble cell-free extract from E. coli expressing OleD, Olel or MGT; 4, 11 and 18, material that did not bind to the Talon column; 5, 12 and 19, Talon column wash; 7, purified Olel; 14, purified Olel; 21, purified MGT.

Figure 2 shows illustrative mass spectra of GT-catalyzed oleandomycin glycosylates (a) using UDP-GIc + OleD: Product peaks for 1-Glc at 884, 886 [M + Cl^"]^" (b) using UDP-5S-Glc + OleD: 1-5S-GIc at 900, 902 [M + CF]^" (c) using UDP-

GaI and Olel: 1-Gal at 884, 886 [M + CF]^" (d) total TIC time course for 1-Glc peak in (a) at [UDP-GIc] = 100 μM.

Figure 3 shows the Lineweaver-Burke Plot for OleD with UDPGIc and Oleandomycin as substrates. Lineweaver-Burke intersection in the second quandrant indicated that K_IA > KA-

Figure 4 shows Lineweaver-Burke Plot for OleD with UDP5SGlc and Oleandomycin as substrates.

Figure 5 shows the Lineweaver-Burke Plot for Olel with UDPGIc and Oleandomycin as substrates. Figure 6 shows the Lineweaver-Burke Plot for Olel with UDP5SGlc and

Oleandomycin as substrates.

Figure 7 shows the Lineweaver-Burke Plot for MGT with UDPGIc and Oleandomycin as substrates. Lineweaver-Burke intersection in the third quandrant indicated that K₁A < KA- Figure 8 shows the Lineweaver-Burke Plot for MGT with UDP5SGlc and

Oleandomycin as substrates.

Figure 9 shows ¹H and ¹³C NMR data of glycosylated oledomycins (a) 1-Gal and (b) 1-5SGIc in CD₃OD.

Figure 10 shows the acceptor library. 1) Oleandomycin, 2) Baicalein, 3) Umbelliferone, 4) 4-Methyl-umbelliferone, 5) Sinapic acid, 6) 2-hydroxy-benzoic acid, 7) a-cyano-4-hydroxyl-cinamic acid, 8) 3,4-dichloroaniline, 9) 3,4- dihydroxylbenzoic acid, 10) 2,5-dihydroxylbenzoic acid, 11) D-glycerate, 12) GIcNAc, 13) Indole 3-acetate, 14) Gibberellin A3, 15) Gibberellin A4, 16) (±> Jasmonic acid, 17) (±)- cis, trans Abscisic acid, 18) Kinetin, 19) Zeatin, 20) Luteolin, 21) Quercetin, 22) Fisetin, 23) Kaempferol, 24) Cinnamic acid, 25) 4-hydroxy cinnamic acid, 26) 3,4-dihyroxy cinnamic acid, 27) 4-hydroxy 3-methoxy cinnamic acid, 28) 2-hydroxy cinnamic acid, 29) 3 -hydroxy cinnamic acid, 30)

7-hydroxy 6-methoxy coumarin (Scopoletin), 31) 6,7-dihydroxy coumarin (Esculetin), 32) Threonine, 33) Glucose, 34) Dihydrojasmonic acid, 35) Ser-Phe 36) Ser-Leu, 37) BocCysThrOMe, 38) l-Thio-S-cyanomethyl-N-acetyl-D-glucosamine, 39) l-Thio-5'-(16-hexadecanoic acid)-N-acetyl-D-glucosamine, 40) 1-Thio-S- ^J cyanomethyl-N-acetyl-D-lactosamine, 41) MUGIcNAc, 42) Trans-dihydroquercetin (DHQ), 43) Cyanidin chloride (CYN), 44) Vancomycin, 45) Novobiocin, 46) Coumermycin Al, 47) 7-hydroxycoumerin 3-carboxylic acid, 48) 7- hydroxycoumerin 4-acetic acid, 49) Chloramphenicol, 50) Kanamycin, 51) Carbencillin disodium salt, 52) B-GIcOBn, 53) α-ManOBn, 54) α -ManOPh, 55) α- ManOCH2Bn, 56) α -ManOPMP, 57) α-ManOBn(pNO2), 58) α-ManOPhF5, 59) α- ManOBnF5, 60) ManSTol, 61) Man (6α) Man(3 α )Man(α)OPh, 62) Z)-erythro- spingosine, 63) Erythromycin, 64) Tylosin

Figure 1 IA shows GAR results of UDPGIc and library acceptors using (a) OleD (b) Olel (c) MGT. Figure 1 IB shows the GAR results of Figure 1 IA in tabular form.

Figure 12A shows another set of GAR results of UDPGIc and library acceptors using (a) OleD (b) Olel (c) MGT.

Figure 12B shows the GAR results of Figure 12 A in tabular form.

Figure 13 A shows the GAR results of varied donors with selected acceptors using (a) OleD (b) Olel (c) MGT.

Figure 13B shows the GAR results of Figure 13 A in tabular form. A = UDP- GIc, B = UDP-5S-Glc, C - UDP-GIcNAc, D = UDP-Man, E - UDP-GaI, F = GDP- Man and G = GDP-Fuc.

Figure 14A shows another set of GAR results of varied donors with selected acceptors using (a) OleD (b) Olel (c) MGT. Figure 14 B shows the GAR results of Figure 14A in tabular form. B = UDP- 5S-GIc, C = UDP-GIcNAc, D = UDP-Man, E = UDP-GaI, F = GDP-Man and G = GDP-Fuc.

Figure 15 A shows GAR results of donor library with (a) Oleandomycin using OleD (1), Olel (2) and MGT (3); (b) Erythromycin using OleD (IA) and MGT (3A) and Tylosin using OleD (IB) and MGT (3B). All the enzymes are restricted in the C2 and C3 configuration, and tolerate of the change of ring oxygen to sulfur.

Figure 15B shows the GAR results of Figure 15A in tabular form.

Figure 16 shows enzyme catalyzed syntheses of remodeled antibiotics GaI- Oleandomycin (1-Gal) and 5SGlc-Oleandomcyin (1-5SGIc). For 1-Gal, Olel, 1OmM TRIS pH 7.8, 37⁰C, 2d, 61% (X=O, A=H, B=OH). For 1-5SGLc, OleD, ImM MES pH 6.5, 37⁰C, 4d, 82% (X=S, A=OH, B-H).

Figure 17 shows a proposed intermediate of OleD, Olel and MGT. Positive charged Nitrogen from Oleandomycin must act as a recognization centre for the binding of donor and stabilize the intermediate. The tighter restrictions on C2 and C3 configuration suggest that there is strong binding between donor sugars and enzymes In contrast, C4 substituent variation and very loose C6 requirements suggests the binding is less strong.

Figure 18 shows the donor library. A: α -UDP-/J>-Glucose, B: α-UDP-D- GIcNAc , C: α -UDP-5-thio-D-glucose, D: α -UDP-TJ-Galactose, E: α -UDP-D- Mannose, F: α -GDP-D-Mannose, G: α -GDP-Z-Fucose, H: GDP-D-Glucose, I: UDP-Z-Fucose, J: UDP-6-deoxyl-6-fluoro-.D-Galactose, K: UDP-6-O-methyl-/J>- galactose, L: UDP-Z-Arabinose, M: UDP-S-deoxy-S-fluoro-D-Galactose, N: UDP- D-Glucosamine, P: UDP-2-deoxy-2-fluoro-/>Galactose, Q: UDP-5-thio-I- Arabinose, R: dTDP-D-Xylose, S: UDP-D-Xylose.

Figure 19 shows 1-Gal antibiotic activity. Top left area contains Oleandomycin (50 μg/ml), bottom left contains 1-Glc (50 μg/ml), top right contains 1-Gal (50 μg/ml) and bottom right contains 1-Gal (50 μg/ml) and IPTG (1 mM).

Figure 20 shows 1-Gal uptake test by E. coli. (a) BL21. (b) Tuner. Description of the Sequence Listing

SEQ ID NO: 1 shows the polynucleotide sequence that encodes the macrolide glycosyltransferase (MGT) from Streptomyces lividans (GenBank No. M74717). MGT is encoded by residues 819-2075. SEQ ID NO: 2 shows the amino acid sequence of the macrolide glycosyltransferase (MGT) from Streptomyces lividans (GenBank No. M74717).

SEQ ID NO: 3 shows the polynucleotide sequence that encodes Olel from Streptomyces antibioticus (GenBank No. AF055579). Olel is encoded by residues 9957-11231. SEQ ID NO: 4 shows the amino acid sequence of Olel from Streptomyces antibioticus (GenBank No. AF055579).

SEQ ID NO: 5 shows the polynucleotide sequence that encodes OleD from Streptomyces antibioticus (GenBank No. Z22577). OleD is encoded by residues 2044-3336. SEQ ID NO: 6 shows the amino acid sequence of OleD from Streptomyces antibioticus (GenBank No. Z22577).

Detailed description of the invention

Method of glycosylating, a substrate

The present invention relates to a method of attaching a sugar to a substrate comprising contacting the substrate with a macrolide glycosyltransferase enzyme in the presence of a sugar donor. The invention therefore relates to a method of glycosylating a substrate using a macrolide glycosyltransferase enzyme in the presence of a sugar donor.

The method of the invention is typically carried out in vitro. The method may be carried out using (i) an isolated, naturally-occurring macrolide glycosyltransferase enzyme, (ii) an isolated, recombinant macrolide glycosyltransferase enzyme or (iii) a cell expressing a naturally-occurring or recombinant macrolide glycosyltransferase enzyme. The method may employ a cell of the invention.

The method of the invention is carried out under conditions that are standard for the glycosyltransferase enzymes. Such conditions are well in the art (For example, L. M. Quiros, I. Aguirrezabalaga, C. Olano, C. Mendez, J. A. Salas, Molec. Microbiol. 1998, 28, 1177). The reaction conditions preferably includes the use of buffers and optionally surfactants in aqueous solution with metals and coenzymes if applicable with the appropriate sugar donor and acceptor to be glycosylated. The Examples disclose typical reaction conditions. Glycosylation of the substrate may be monitored using methods that are common in the art (for example, L. M. Quiros, I. Aguirrezabalaga, C. Olano, C. Mendez, J. A. Salas, Molec. Microbiol. 1998, 28, 1177). For instance, HTS mass spectrometry may be used to monitor glycosylation of the substrate. The method of the invention involves the culturing of a bacterial cell containing a heterologous polynucleotide encoding a macrolide glycosyltransferase under conditions in which the enzyme is expressed. This is described in more detail below. The method also involves recovering the enzyme. This can by done using standard methods in the art. The recovering step typically includes purifying the enzyme. The enzyme may be purified using standard methods in the art. The enzyme is preferably appended with a His-tag and purified using a single step affinity purification strategy.

In one embodiment, attachment of a sugar to the substrate in accordance with the invention results in increased internalization of the substrate into cells. This is due to glycotargeting. For instance, attachment of galactose to oleandomycin results in an increased internalization of the glycosylated oleandomycin by cells.

Macrolide glycosyltransferases

The method of the invention is carried out using a macrolide or leiloir-type glycosyltransferase. These enzymes catalyze the addition of the glycosyl group from a UTP-sugar to a hydrophobic molecule. In particular, they specifically inactivate macrolide antibiotics via 2'-O-glycosylation using UDP-glucose. Suitable enzymes include, but are not limited to, those expressed in Streptomyces. Preferred enzymes for use in the method of the invention include, but are not limited to, the macrolide glycosyltransferase (MGT) from Streptomyces lividans (GenBank No. M74717), Olel from Streptomyces antibioticus (GenBank No. AF055579) and OleD from Streptomyces antibioticus (GenBank No. Z22577). MGT, Olel and OleD in accordance with the invention include the polypeptides shown in SEQ ID NOs: 2, 4 and 6 and variants thereof.

A variant is an enzyme having an amino acid sequence which varies from that of MGT (SEQ ID NO: 2), Olel (SEQ ID NO: 4) or OleD (SEQ ID NO: 6) but retains macrolide glycosyltransferase activity.

A variant of any of SEQ ID NOs: 2, 4 or 6 may be a naturally occurring variant which is expressed by organism, for example another strain oiStreptomyces. Such variants may be identified by looking for macrolide glycosylatransferase activity in those strains which have a sequence which is highly conserved compared to any of SEQ ID NOs: 2, 4 or 6. Such proteins may be identified by analysis of the polynucleotide encoding such a protein isolated from another organism, for example, by carrying out the polymerase chain reaction using primers derived from portions of any of SEQ ID NOs: 1, 3 or 5.

Variants of any of SEQ ID NOs: 2, 4 or 6 include sequences which vary from SEQ ID NO: 2 but are not necessarily naturally occurring MGT, Olel or OleD. Over the entire length of the amino acid sequence of any of SEQ ID NOs: 2, 4 or 6, a variant will preferably be at least 35% homologous to that sequence based on amino acid identity. More preferably, the polypeptide may be at least 40%, at least 50%, at least 60%, at least 70%, at least 80% or at least 90% and more preferably at least 95%, 97% or 99% homologous based on amino acid identity to the amino acid sequence of any of SEQ ID NOs: 2, 4 or 6 over the entire sequence. There may be at least 80%, for example at least 85%, 90% or 95%, amino acid identity over a stretch of 40 or more, for example 60, 100 or 120 or more, contiguous amino acids ("hard homology"). Amino acid substitutions may be made to the amino acid sequence of any of

SEQ ID NOs: 2, 4 or 6, for example up to 1, 2, 3, 4, 5, 10, 20 or 30 substitutions. Conservative substitutions may be made, for example, according to the following table. Amino acids in the same block in the second column and preferably in the same line in the third column may be substituted for each other:

One or more amino acid residues of the amino acid sequence of any of SEQ ID NOs: 2, 4 or 6 may alternatively or additionally be deleted. Up to 1, 2, 3, 4, 5, 10, 20 or 30 residues may be deleted, or more. Variants of any of SEQ ID NOs: 2, 4 or 6 include fragments of those sequences. Such fragments retain macrolide glycosyltransferase activity. Fragments may be at least 100, 200, 250, 300, 350 or 400 amino acids in length. Such fragments may be used to produce chimeric enzymes as described in more detail below. A fragment preferably comprises the catalytic domain of any of SEQ ID NOs: 2, 4 or 6. Variants of any of SEQ ID NOs: 2, 4 or 6 include chimeric proteins comprising fragments or portions of any of SEQ ID NOs: 2, 4 or 6.

One or more amino acids may be alternatively or additionally added to the polypeptides described above. An extension may be provided at the N-terminus or C-terminus of the amino acid sequence of any of SEQ ID NOs: 2, 4 or 6 or polypeptide variant or fragment thereof. The extension may be quite short, for example from 1 to 10 amino acids in length. Alternatively, the extension may be longer, for example up to 50 or 100 amino acids. A carrier protein may be fused to an amino acid sequence according to the invention.

Standard methods in the art may be used to determine homology. For example the UWGCG Package provides the BESTFIT program which can be used to calculate homology, for example used on its default settings (Devereux et al (1984) Nucleic Acids Research 12, p387-395). The PILEUP and BLAST algorithms can be used to calculate homology or line up sequences (such as identifying equivalent residues or corresponding sequences (typically on their default settings)), for example as described in Altschul S. F. (1993) J MoI Evol 36:290-300; Altschul, S.F et al (1990) J MoI Biol 215:403-10.

Software for performing BLAST analyses is publicly available through the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/). This algorithm involves first identifying high scoring sequence pair (HSPs) by identifying short words of length W in the query sequence that either match or satisfy some positive-valued threshold score T when aligned with a word of the same length in a database sequence. T is referred to as the neighbourhood word score threshold (Altschul et al, supra). These initial neighbourhood word hits act as seeds for initiating searches to find HSP's containing them. The word hits are extended in both directions along each sequence for as far as the cumulative alignment score can be increased. Extensions for the word hits in each direction are halted when: the cumulative alignment score falls off by the quantity X from its maximum achieved value; the cumulative score goes to zero or below, due to the accumulation of one or more negative-scoring residue alignments; or the end of either sequence is reached. The BLAST algorithm parameters W, T and X determine the sensitivity and speed of the alignment. The BLAST program uses as defaults a word length (W) of 11, the BLOSUM62 scoring matrix (see Henikoff and Henikoff (1992) Proc. Natl. Acad. ScL USA 89: 10915-10919) alignments (B) of 50, expectation (E) of 10, M=5, N=4, and a comparison of both strands.

The BLAST algorithm performs a statistical analysis of the similarity between two sequences; see e.g., Karlin and Altschul (1993) Proc. Natl. Acad. Sci. USA 90: 5873-5787. One measure of similarity provided by the BLAST algorithm is the smallest sum probability (P(N)), which provides an indication of the probability by which a match between two amino acid sequences would occur by chance. For example, a sequence is considered similar to another sequence if the smallest sum probability in comparison of the first sequence to the second sequence is less than about 1, preferably less than about 0.1, more preferably less than about 0.01, and most preferably less than about 0.001. Acceptor substrates

The acceptor substrate is glycosylated by the macrolide glycosyltransferase in the method of the invention. The sugar substrate and/or the acceptor substrate are generally not the natural substrates for the enzyme. The enzyme attaches or appends a sugar to the substrate in the method of the invention. The enzyme generally appends the sugar to a hydroxy! (-OH) group on the substrate. However, the enzyme may also append the sugar to other groups on the substrate. Examples of such groups include NH and SH. The enzyme may also append the sugar to an aromativc ring of the substrate to form C-glycosided. The enzyme preferably appends the sugar to a sugar moiety on the substrate. The enzyme typically appends the sugar to a 6- deoxyhexose moiety in the substrate.

The substrate is generally an antibiotic. At least part of the substrate is preferably a planar, cyclic and/or hydrophobic molecule that fits into the aglycone binding site of the macrolide glycosyltransferase. At least part of the substrate is preferably planar such that its atoms are substantially aligned in a single plane (i.e. at least part of the substrate is preferably flat). At least part of the substrate is preferably cyclic such that it comprises one or more moieties, preferably ring structures, that are arranged in a circular pattern. At least part of the substrate is also preferably hydrophobic as measured using conventional methods in the art. In one embodiment, the method of the invention is carried out using a polyketide as the substrate. Polyketides are compounds containing alternating carbonyl and methylene groups (β-polyketones) which are biogenetically derived from repeated condensation of acetyl coenzyme A (via malonyl coenzyme A) and the compounds typically derived from them by further condensations. Suitable polyketides include, but are not limited to, fatty acids, polyproprionates and aromatic polyketides. Suitable polyproprionates include, but are not limited to, polyether antibiotics, macrolides and spiroketals. Suitable macrolides include, but are not limited to, polyoxomacrolides, polyene macrolides, ionophore macrolides and ansamycin macrolides Suitable spiroketals include, but are not limited to, steroidal glycosides, sprioketal enolethers and polyether ionophores.

The acceptor substrate used in the method of the invention is preferably a macrolide. A macrolide is a compound comprising a large lactone ring with few (e.g. one, two or three) or no double bonds, which is linked glycosidically to one or more (e.g one, two or three) sugar moieties. In one embodiment, the acceptor substrate is of the formula I:

wherein X comprises C_3-20 alklyene or C_3-20 alkenylene, and may further comprise one or more (e.g. one, two or three) nitrogen, sulphur or oxygen atoms; and wherein the C_3-20 alkylene or C_3-20 alkenylene group in X is substituted by one or more -OR¹ groups (e.g. one, two or three such groups) wherein R¹ is a sugar and is optionally further substituted by one or more substituents (e.g. 1 to 20 substituents, preferably 5 to 15 substituents) selected from:

C_1-6 alkyl; - C_1-6 alkenyl;

=0

-OR wherein R is selected from H, C_1-6 alkyl and C_1-6 alkenyl;

-(CH₂)_n-COR³ wherein n = 0-6 and wherein R³ is selected from H and C_1-6 alkyl; and

-O(CH₂)_n- wherein n = 1 -6 and wherein the O and the terminal carbon atom are both bonded to the same carbon atom of the C_3-20 alkylene or C_3-20 alkenylene group.

The alkylene or alkenylene groups in X preferably have 8, 9, 10, 11, 12, 13,

14 or 15 carbon atoms.

R¹ is a sugar as described in more detail below (the sugar that constitutes R¹ may be selected from the sugars described below in the context of donor substrates).

R¹ is preferably selected from desoaminyl, cladinosyl, mycarosyl-mycaminosyl and mycinosyl.

R is preferably C_1-4 alkyl or alkenyl and more preferably C_1-2 alkyl or alkenyl. In preferred embodiments, n is 1-4 or 1-2.

Examples of macrolides include azithromycin, brefeldin A, erythromycin A, leucomycin Al, methymycin, oleandomycin, pikromycin, tylosin and zearalenon. Suitable polyketides for use in the method of the invention include, but are not limited to, aflatoxin, amphotericin B, ascomycin, avermectin, azaspiracid, azithromycin, brefeldin A, brevetoxin B, calcimycin (A-23187), callipeltoside A, chlortetracycline, clarithromycin, concanamycin A, epothilone B, erythromycin A, etheromycin, ionomycin, lasalocid A, leucomycin Al, lovastatin, maitotoxin, methymycin, 6-methylsalicyclic acid, monensin, maytansine, nonactin, nystatin Al, okadaic acid, oleandomycin, oxytetracycline, pamamycin, pikromycin, rapamycin, reservatrol, reveromycin A, reveromycin B, rifamycin, roxithromycin, tsukubaenolide FK506, tylosin, virginiamycin and zearalenon. Preferred polyketides for use in the method of the invention include, but are not limited to, oleandomycin, erythromycin and rapamycin. In another embodiment, the method of the invention is carried out using a cyclic non-ribosomal peptide as the substrate. Suitable cyclic non-ribosomal peptides include, but are not limited to, balhimycin, teicoplanin and vancomycin. In another embodiment, the method of the invention is carried out using a coumarin-class antibiotic as the substrate. Suitable coumarin-class antibiotics include, but are not limited to, coumermycin Al and novobiocin.

In another embodiment, the method of the invention is carried out using an aminoglycoside as the substrate. Suitable aminoglycosides include, but are not limited to, amikacin, framycetin, gentamycin, kanamycin, neomycin, netilmicin, spectinomycin, streptomycin and tobramycin. Preferred aminoglycosides for use in the method of the invention include, but are not limited to, genatmycin, neomycin or streptomycin.

In another embodiment, the method of the invention is carried out using an antracycline antibiotic as the substrate. Suitable antracyclines include, but are not limited to, daunomycin (rubidomycin) and doxorubicin. In another embodiment, the method of the invention is carried out using an enediyne antibiotic as the substrate. Suitable enediynes include, but are not limited to, calicheamicin, esperamicin-Aj: and lidamycin. In another embodiment, the method of the invention is carried out using a sugar as the substrate. The sugar substrate may be a monosaccharide, disaccharide or polysaccharide. Suitable sugar substrates for use in the method of the invention include, but are not limited to, those sugars described in more detail below (the sugar that constitutes the acceptor may be selected from the sugars described below in the context of donor substrates).

Sugar donors

The method of the invention requires a sugar donor. The macrolide glycosyltransferase enzyme transfers the sugar from the sugar donor to the acceptor substrate. A sugar donor is a molecule comprising a sugar that is capable of donating the sugar to the substrate in the presence of a macrolide glycosyltransferase. The sugar donor may contain any nucleobase or nucleoside attached to the sugar. Suitable nucleobases include purine and pyrimidines and more specifically adenine, guanine, thymine, uracil and cytosine. Suitable nucleosides include adenosine, guanosine, thymidine, uridine, cytidine and inosine. The sugar donor may contain a nucleobase or a nucleoside monophosphate, diphosphate or triphosphate. The sugar donor preferably comprises a nucleoside diphosphate (e.g. uridinediphosphate (UDP)) attached to the sugar. The sugar donor most preferably comprises UDP and is of the formula II:

Sugars include but, but are not limited to, monosaccharides, disaccharides, polysaccharides, substances derived from the monosaccharides by oxidation of one or more terminal groups to carboxylic acids and substances derived from the monosaccharides by replacement of one or more hydroxy groups by a hydrogen atom, an amino group, a thiol group or similar heteroatomic groups.

In one embodiment, the sugar donor comprises a monosaccharide. Monosaccharides are polyhydroxy aldehydes (H-(CHOH)n-CHO) or polyhydroxy ketones (H-(CHOH)n-CO-(CHOH)m-H) with three (triose), four (terose), five

(pentose), six (hexose), seven (heptose) or more carbon atoms. The monsaccharide may have a terminal, i.e. aldehydic, (potential) carbonyl group (aldose) or a non¬ terminal, i.e. ketonic, (potential) carbonyl group (ketose). The monosaccharide have more than one (potential) carbonyl group, i.e. may be a dialdose, diketose or aldoketose. A potential carbonyl group is a hemiacetal group that arises from the formation of a ring structure.

The monosaccharide may form a ring structure and be a cyclic hemiacetal or hemiketal. Cyclic forms include oxiroses (C₃), oxetoses (C₄), furanoses (C₅), pyranoses (C₆), septanoses (C₇) and octanoses (C₈). The position at which the ring closes may vary.

Monosaccharides may also be modified at various positions. Such modifications include, but are not limited to, substitution of an alcoholic hydroxy group with hydrogen (deoxy sugar), substitution of an alcoholic hydroxy group or a ring oxygen with an amino group (amino sugar), substitution of a hydroxyl group by thiol or a ring oxygen with sulphur (thiosugar), substitution with selenium (a seleno sugar), substitution of a ring carbon with nitrogen (aza sugar), substitution of a aldehydic group with a carboxyl group (aldonic acid), substitution of a carbonyl group with an alcoholic group (aldonic acid or ketoaldonic acid) and oxidizing an aldose or aldose derivative to a carboxyl group (uronic acid). In one embodiment, the monosaccharide is of the formula III:

(III)

wherein Y is O, S or NH; wherein R⁴ is selected from:

H;

C_1-6 alkyl;

C_1-6 alkenyl;

NH₂;

-(CH₂)_n-OH wherein n = 0 to 6;

OR⁵ wherein R⁵ is H or C_1-6 alkyl;

-(CH₂)_n-COR⁶ wherein n = 0 to 6 and wherein R⁶ is selected from H and C_1-6 alkyl; and wherein R⁷ is selected from H, OH and -(CH₂)_n-OH wherein n = 0 to 6. Preferably, one of the R groups at each position is H and the other R group at the same position is other than H. In another embodiment, the monosaccharide is of the formula IV:

wherein wherein Y is O, S or NH; wherein R⁸ is selected from:

H;

C_1-6 alkyl;

C_1-6 alkenyl;

NH₂;

-(CH₂)_n-OH wherein n = 0 to 6;

OR⁹ wherein R⁹ is H or C_1-6 alkyl;

-(CH₂)_n-COR¹⁰ wherein n = 0 to 6 and wherein R . ⁱ1^υ0 is selected from H and C_1-6 alkyl; and wherein R¹¹ is selected from H, OH and -(CH₂)_n-OH wherein n = O to 6.

Preferably, one of the R groups at each position is H ad the other R group at the same position is other than H.

Suitable monosaccharides include, but are not limited to, arabinose, ribose, ribulose, xylose, xyulose, lyxose, allose, altrose, glucose (GIc), N-acetylglucosamine (GIcNAC), fructose (Frc), galactose (Gal), fucose (Fuc), gulose, idose, mannose (Man), sorbose, talose and tagatose. Preferred sugar donors include, but are not limited to, UDP-GIc, UDP-5S-Glc, UDP-GIcNAc and UDP-Man. These monosaccharides and sugar donors may be modified as set forth above. In another embodiment, the sugar donor comprises a disaccharide.

Disaccharides may be derived from the combination of any two of the monosaccharides described above. Suitable disaccharides include, but are not limited to, sucrose (Sue), lactose (Lac), maltose (MaI), isomaltose (Isomal), trehalose (Tre) and cellobiose. In another embodiment, the sugar donor comprises a polysaccharide.

Polysaccharides are derived from the combination of three or more (for example, 20, 30, 50, 100, 200 or more) monosaccharide units. Suitable polysaccharides include, but are not limited to, starch, amylase, amylopectin, glycogen, inulin, cellulose, chitin, glycosaminoglycans, agar, carrageenan, pectin, xantham gum and glucomannan.

Method of expressing a glycosyltransf erase

The present invention also provides a method of expressing a macrolide glycosyltransferase enzyme at a high level comprising expressing a nucleic acid sequence encoding the enzyme in a bacterial cell. The invention therefore concerns a method of producing a high amount of macrolide glycosyltransferase enzyme in a bacterial cell using recombinant techniques.

The macrolide glycosyltransferase enzyme is typically expressed at a level greater than 20 mg/L of culture, 30 mg/L of culture, 40 mg/L of culture, culture, 50 mg/L of culture, 60 mg/L of culture, 70 mg/L of culture or 80 mg/L of culture. The maximum expression level may be 1000 mg/L, 500 mg/L, 200 mg/L or 100 mg/L.

Thus, the macrolide glycosyltransferase enzyme is typically expressed a level from 20 to 1000 mg/L of culture, from 30 to 500mg/L of culture, or from 40 to 200 mg/L of culture. The amount of expression of the macrolide glycosyltransferase enzyme may be measured using standard methods in the art. For instance, the amount of macrolide glycosyltransferase enzyme protein may be measured using SDS-PAGE. The amount of the enzyme may also be determined by measuring the absorbance of the proteins at their molar extinction coefficient after they have been purified, for instance by immobilized metal ion affinity chromatography. OleD, Olel and MGT have a molar extinction coefficient at 280 nm based on the number of tryptophan and tyrosine amino acids they contain. The method of the invention is carried out using a (macrolide or leiloir-type) glycosyltransferase enzyme nucleic acid sequence. Suitable enzymes include, but are not limited to, those expressed in Streptomyces. Preferred enzymes for use in the method of the invention include, but are not limited to, the macrolide glycosyltransferase (MGT) from Streptomyces lividans (SEQ ID NO: 2), Olel from Streptomyces antibioticus (SEQ ID NO: 4) and OleD from Streptomyces antibioticus (SEQ ID NO: 2) and variants thereof (as described above). The method may employ any of the macrolide glycosyltransferase sequences described above.

The host cell and polynucleotide are heterologous with respect to each other. A heterologous polynucleotide is one that is derived from a different species to the host cell. In one embodiment, the host cell is bacterial, preferably from Escherichia coli.

Nucleic acid sequences encoding a macrolide glycosyltransferase enzyme may be isolated and replicated using standard methods in the art. These methods are described in more detail below with reference to the vectors of the invention. Nucleic acid sequences encoding a macrolide glycosyltransferase enzyme may be expressed in a bacterial host cell using standard techniques in the art. The nucleic acid sequence encoding a macrolide glycosyltransferase enzyme may be cloned into suitable expression vector. Suitable expression vectors are also described in more detail below. The expression vector may then be introduced into a suitable host cell. Thus the method of the invention may be carried out by introducing a nucleic acid sequence encoding a macrolide glycosyltransferase enzyme into an expression vector, introducing the vector into a compatible bacterial host cell, and growing the host cell under conditions which bring about expression of the nucleic acid sequence encoding the macrolide glycosyltransferase enzyme.

Host cells The present invention also provides a bacterial host cell that recombinantly expresses a macrolide glycosyltransferase enzyme at a high level. The host cell has therefore been transformed with a nucleic acid sequence encoding a macrolide glycosyltransferase enzyme. The host cell may comprise a vector of the invention and may be used in the expression method of the invention. The macrolide glycosyltransferase enzyme is typically expressed by the host cell at a level described above.

The host cells of the invention comprise a (macrolide or leiloir-type) glycosyltransferase enzyme nucleic acid sequence. Suitable enzymes include, but are not limited to, those expressed in Streptomyces. Preferred enzymes include, but are not limited to, the macrolide glycosyltransferase (MGT) from Streptomyces lividans (SEQ ID NO: 2), Olel from Streptomyces antibioticus (SEQ ID NO: 4) and OleD from Streptomyces antibioticus (SEQ ID NO: 2) and variants thereof.

Host cells transformed with a nucleic acid sequence encoding a macrolide glycosyltransferase enzyme will be chosen to be compatible with the expression vector used to transform the cell. In a preferred embodiment, the host cell is

Escherichia coll Any cell with a λ DE3 lysogen, for example C41 (DE3), BL21 (DE3), JMl 09 (DE3), B834 (DE3), TUNER, Origami and Origami B, can express a vector comprising the T7 promoter. In a preferred embodiment, C41(DE3) is used. Escherichia coli JM 109 and NM554 are preferably used for vectors comprising the trc promoter.

Vectors

The invention also provides a vector comprising a nucleic acid sequence encoding a macrolide glycosyltransferase enzyme. The macrolide glycosyltransferase enzyme may be any of those described above. The vector may be a cloning vector or an expression vector. A cloning vector of the invention may be used to produce nucleic acid sequences encoding a macrolide glycosyltransferase enzyme. An expression vector may be used in the expression method of the invention or to produce a host cell of the invention.

Nucleic acid sequences encoding a macrolide glycosyltransferase enzyme may be isolated and replicated using standard methods in the art. Chromosomal DNA may be extracted from a macrolide glycosyltransferase enzyme-producing organism, such as Sfreptomyces. The gene encoding the macrolide glycosyltransferase enzyme may be amplified using PCR involving specific primers. The amplified sequence may then be incorporated into a recombinant replicable vector such as a cloning vector. The vector may be used to replicate the nucleic acid in a compatible host cell. Thus nucleic acid sequences encoding a macrolide glycosyltransferase enzyme may be made by introducing a polynucleotide encoding a macrolide glycosyltransferase enzyme into a replicable vector, introducing the vector into a compatible host cell, and growing the host cell under conditions which bring about replication of the vector. The vector may be recovered from the host cell. Suitable host cells for cloning of polynucleotides encoding a macrolide glycosyltransferase enzyme are described in more detail below.

In an expression vector, the nucleic acid sequence encoding a macrolide glycosyltransferase enzyme is typically operably linked to a control sequence which is capable of providing for the expression of the coding sequence by the host cell. Such expression vectors can be used to express a macrolide glycosyltransferase enzyme in accordance with the invention or produce a host cell of the invention.

The term "operably linked" refers to a juxtaposition wherein the components described are in a relationship permitting them to function in their intended manner. A control sequence "operably linked" to a coding sequence is ligated in such a way that expression of the coding sequence is achieved under conditions compatible with the control sequences. Multiple copies of the same or different macrolide glycosyltransferase enzyme genes may be introduced into the vector.

Such vectors may be transformed into a suitable host cell to provide for expression of a macrolide glycosyltransferase enzyme. Thus, a macrolide glycosyltransferase enzyme polypeptide can be obtained by cultivating a host cell transformed or transfected with an expression vector as described above under conditions to provide for expression of the polypeptide, and recovering the expressed polypeptide.

The vectors may be for example, plasmid, virus or phage vectors provided with an origin of replication, optionally a promoter for the expression of the said nucleic acid sequence and optionally a regulator of the promoter. The vectors may contain one or more selectable marker genes, for example a tetracycline resistance gene. Promoters and other expression regulation signals may be selected to be compatible with the host cell for which the expression vector is designed. A T7, trc, lac, ara or λ^ promoter is typically used.

The following Example illustrates the invention:

Example

1. Methods

Bacterial strains, culture conditions and plasmid vectors

The E. coli strains used in this study were C41(DE3) (gift from Prof. A. R. Fersht at the Medical Research Council, Cambridge, UK) and XLl -Blue, while the streptomycetes species employed were Streptomyces lividans (ATCC DSMZ No.

46482) and Streptomyces endibioticus (ATCC DSMZ No. 40868). The bacteria were cultured in Luria broth (LB) at 37⁰C with aeration unless otherwise stated. Media were supplemented with 50 μg ml^'1 ampicillin to select for and maintain recombinant plasmids. The plasmid vectors used in this study were and Minipreseta (gift from Prof. A. R. Fersht at the Medical Research Council, Cambridge, UK). Proteins encoded by recombinant plasmids derived from the expression vector Minipreseta were induced as follows: 20 colonies from a fresh agar plate were inoculated into 800ml LB in 2 litre flasks and cultures incubated at 37⁰C at 180 rpm to an A600 °f 0.4 at which point the cultures were maintained at 16⁰C with aeration for 1 h and after the addition of isopropyl-β-D-thiogalactopyranoside (IPTG) to a final concentration of 60 μM incubation was continued for 16 h at 16⁰C. Cloning of glycosyltransferase genes into E. coli expression vectors

The plasmid vector used in this study was Miniprseta (gift from Prof. A. R. Fersht at the Medical Research Council, Cambridge, UK). Chromosomal DNA was extracted from Streptomyces lividans (DSMZ No.46482) and Streptomyces antibioticus (DSMZ No.40868) using the QIAGEN Genomic-tip 100/G kit (#10243). The genes encoding MGT, OleD and Olel were amplified in the standard 50μl reaction mixture (advised by Novagen) containing 0.5 μg of the appropriate genomic DNA, the primers listed in Table 1 and Kod Hot Start polymerase (Novagen). The PCRs were carried out using a Techne PHC-3 Thermocycler at the following temperatures; 94 ⁰C for 2 min followed by 35 cycles at 94 °C for 30 sec and 68 °C for 2 min with a final single extended elongation phase at 68 °C for 10 min. The amplified DNA was digested with BamHl and EcoRI (Olel and MGT) or BgRl and EcoRI (OleD) and ligated into the polylinker region of Minipreseta (precut using BamHl and EcoRI). The ligated plasmids were transformed into the Escherichia coli strain XLl -Blue (Stratagene).

The cloned genes were sequenced from two independent recombinant plasmids by MWG using the T7 promoter and T7 terminater sequencing primers (supplied by MWG). The data showed that both plasmids for each gene contained identical sequences, but differed in several places from the respective sequence deposited in GenBank. The corrected sequences have now been deposited with GenBank (OleD - Z22577, Olel - AF055579 and MGT - M74717 - (original numbers)).

Table 1 - Oligonucleotide primers for gene amplification

Protein expression and purification

The glycosyltransferases were expressed in E.coli cell strain C41(DE3) (B. Miroux, J. E. Walker, J. MoI. Biol. 1996, 260, 289) using LB medium supplemented with ampicillin at 50 μg mL^"1. They were initially grown at 37°C to an A_6Oo_nm value of 0.4 before overnight induction was initiated with the addition of isopropyl-β-D- thiogalactopyranoside (60 μM) at 16 ⁰C (MGT and OleD) or 30 °C (Olel). The cells obtained from 1.6 1 of each culture were re-suspended in 80 mL of 20 mM Tris-HCl buffer, pH 8.0, containing 500 mM NaCl (Buffer A). The cells were disrupted by ultrasonication and the insoluble material was removed by centrifugation at 30000g for 30 min at 4°C. The cell-free extract was passed through 4 mL of TALON resin (Clontech), which was then washed with 60 mL of Buffer A and the purified glycosyltransferases eluted in 15 mL of Buffer A containing 100 mM imidazole. Purity was assessed by SDS-PAGE (Figure 1) and the amounts of each protein determined using the appropriate molar-extinction coefficient (OleD is 57040 M^"1 cm^"1, Olel is 44380 M^"1 cm^"1 and MGT is 62730 M^"1 cm^"1).

Full kinetic and HTS mass spectrometric screening

The methods used are taken from M. Yang, M. Brazier, R. Edwards, B. G. Davis, submitted 2004. Briefly, Waters ZMD-MS (ESF), Waters 600 HPLC system with Waters 2700 sampler were operated by Micromass Masslynx 3.3 and data processed using Masslynx 3.5, Microsoft Excel 2002 and Origin 7. The HPLC/auto- sampler control was divided into two stages: stage 1) internal standard NDP (0.1 mM) injection (MeCN : H₂O (50 : 50); 0.12 mL/min; isocratic for 0.1 min; injection volume: 10.0 μL; 150-1000 ESF for 0.1 min, scan 0.2 min); 2) 2. sample injection (MeCN : H₂O (50 : 50); 0.12 mL/min; isocratic for 5.5 min; injection volume: 10.0 μL; 150-1000 ESF for 5. min, TIC of single ion peaks monitored, dwell 0.1 s). Typical reactions: [UDPGIc] 20, 40, 60, 80, 100 μM; [acceptor] 20, 60, 100 μM; 10 μL enzyme; total volume 300 μL TRIS (1 mM, pH 7.8, 1 mM DTT). For the GAR screen a 96 well plate, each well containing TRIS buffer.O mM, pH 7.8, 100 μl), donor (10 mM, 5 μl), acceptor (10 mM, 5 μl), enzymes (lmg/mL, 5 μl) was incubated at 37°C overnight and monitored by MS (full scan from 150-1100 Da). Enzyme catalyzed synthesis of 1 -Gal

UDPGaI disodium salt (10.1 nig, 0.0166 mmol), oleandomycin phosphate (15.7 mg, 0.02 mmol) and Olel (5 mg/mL, 500 μL) were dissolved in TRIS buffer (10 niM, pH 7.8, 20 mL) and incubated at 37° for 2 days. The formation of product was monitored by MS. The reaction was terminated by heating to 90°C for 10 min, centrifuged, the pH of supernatant adjusted to 10 and extracted with CHCl₃ (20 mL). The organic layer was reduced, MeOH (0.5 mL) added and purified by preparative TLC using n-propanol : ethyl acetate : water : acetic acid (PEWA, 40:31:20:1, v/v/v/v) as the mobile phase. Bands at Rf 0.16 were collected to give 1-Gal (8.5 mg, 61%). ¹H NMR (400 MHz, CD₃OD): δ - 5.37 (dq, J= 1.7, 6.4 Hz, IH, H13), 4.97 (d, J= 2.3 Hz, IH, Hl"), 4.40 (d, J= 7.2 Hz, IH, Hl'), 4.38 (d, J= 9.1 Hz, IH, Hl'"), 3.87 (d, J= 3.6 Hz, IH, H4'"), 3.85 (d, J= 4.6 Hz, IH, H3), 3.81 (dd, J=4.0, 7.1 Hz, IH, H6'"), 3.80 (dd, J= 2.4, 4.6 Hz, IH, HI l), 3.72 (m, IH, H5"), 3.70 (m, IH, H5'), 3.70 (dd, J= 4.5, 9.4 Hz, IH, H6"'), 3.62 (t, J= 8.6 Hz, IH, H2'"), 3.58 (m, IH, H5), 3.53 (m, IH, H5'"), 3.49 (m, IH, H3'"), 3.48 (m, IH, H2'), 3.45 (m, IH, H3"), 3.45 (s, 3H, H7"), 3.20 (dd, J= 1.7, 5.1 Hz, IH, HlO), 3.05 (t, J= 9.6 Hz, IH, H4"), 2.95 (m, IH, H3'), 2.83 (d, J= 5.5 Hz, IH, H18), 2.78 (d, J= 5.7 Hz, IH₅ H18), 2.77 (m, IH, H2), 2.58 (dd, J= 3.4, 12.3 Hz, IH, H7), 2.45 (s, 6H, H7', 8'), 2.41 (m, IH, H2"), 1.86 (m, IH, H4), 1.92 (m, IH, H6), 1.90 (m, IH, H4'), 1.67 (t, J = 7.6 Hz, IH, H12), 1.61 (dd, J= 10.2, 10.9 Hz, IH, H7), 1.49 (ddd, J= 3.7, 13.0,

13.3 Hz, IH, H2"), 1.38 (m, IH, H4'), 1.33 (d, J= 6.2 Hz, 3H, H14), 1.28 (d, J= 6.8 Hz, 3H, H6"), 1.24 (d, J= 5.4 Hz, 3H, H6'), 1.23 (d, J= 7.2 Hz, 3H, H16), 1.21 (d, J= 6.8 Hz, 3H, H15), 1.18 (d, J= 6.5 Hz, 3H, H17), 1.02 (d, J= 6.9 Hz, 3H, H20), 1.00 (d, J= 6.9 Hz, 3H, H19); ¹³C NMR (100 MHz, CD₃OD): δ = 209.5 (C9), 179.1 (Cl), 106.2 (Cl'"), 102.5 (Cl'), 98.5 (Cl"), 82.4 (C5), 81.5 (CT), 79.5 (C3), 78.2 (C3"), 77.3 (C5'"), 77.3 (C4"), 74.3 (C3'"), 73.6 (C2'"), 72.2 (C13), 70.2 (CI l), 69.8 (C5"), 69.5 (C5'), 68.2 (C4'"), 67.1 (C3'), 65.0 (C8), 61.5 (C6'"), 57.3 (C7"), 49.5 (Cl 8), 46.3 (ClO), 45.5 (C4), 45.2 (C2), 42.4 (C12), 39.5 (CT), 39.5 (C8'), 34.9 (C2"), 34.0 (C6), 33.5 (Cl), 30.0 (C4'), 21.2 (C6'), 19.2 (C17), 18.0 (C6"), 17.6 (C14), 12.5 (C15), 9.5 (C16), 7.5 (C19), 9.0 (C20); ¹³C-¹H coupling constants from gradient phase sensitive HSQC: C1"-H1" 168.8 Hz, indicated Hl" is α linked; Cl'- Hl' 155.6 Hz indicated Hl' is β linked; C1'"-H1"' 150.5 Hz indicated Hl'" is β linked. HMBC, Hl'" (4.38 ppm) coupled with C2' (81.5 ppm) confirmed galactosylation of 0-2'. HRMS (ESI-) m/z 884.4412 ([M + Cl^"]^") (calc. 884.4411).

Enzyme catalyzed synthesis of 1-5SGIc α-UDP5SGlc (0.0059 mmol), oleandomycin phosphate (9.1 mg, 0.0119 mmol), OleD (8 mg/mL, 2.5mL) were dissolved in MES buffer (1 mM, pH 6.5, 400 mL) and incubated at 37 °C for 3d. Purification followed a very similar procedure to that described above except: extraction with CHCl₃ (400 mL) and the Rf of product 1-5SGIc is 0.24 (30% MeOH in CHCl₃). Note that it's slightly higher than Oleandomycin). Product obtained (4.3 mg, 82%). ¹H NMR (400 MHz, CD₃OD): δ = 5.38 (dq, J= 1.8, 7.2 Hz, IH, H13), 4.97 (d, J= 2.7 Hz, IH, Hl"), 4.60 (d, J= 9.2 Hz, IH, Hl'"), 4.30 (d, J= 7.9 Hz, IH, Hl'), 4.01 (dd, J= 3.7, 10.9 Hz, IH, H4'"), 3.88 (d, J= 6.7 Hz, IH, HI l), 3.81 (d, J= 5.3 Hz, IH, H3), 3.69 (m, IH, H3'"), 3.67 (m, IH, H5"), 3.65 (m, IH, H6'"), 3.60 (t, J= 9.1 Hz, IH, H2'"), 3.56 (m, IH, H5'), 3.54 (d, J= 10.2 Hz IH, H5), 3.44 (m, IH, H6'"), 3.43 (s, 3H, H7"), 3.43 (m, IH, H3"), 3.30 (m, IH, H2'), 3.21 (d, J= 6.0 Hz, IH, HlO), 3.04 (t, J= 8.9 Hz, IH, H4"), 2.81 (m, IH, H5'"), 2.81 (m, IH, H18), 2.77 (m, IH, H2), 2.77 (m, Hz, IH, H18), 2.77 (m, IH, H3'), 2.62 (dd, J= 4.1, 15.1 Hz, IH, H7), 2.45 (dd, J= 6.2, 14.1 Hz, IH, H2"), 2.30 (s, 6H, H7', 8'), 1.93 (m, IH, H4), 1.91 (m, IH, H6), 1.83 (m, IH, H4'), 1.70 (dd, J= 12.6, 15.8 Hz, IH, H7), 1.65 (m, IH, H12), 1.44 (ddd, J= 4.1, 13.2, 14.2 Hz, IH, H2"), 1.33 (d, J= 5.9 Hz, 3H, H14), 1.31 (m, IH, H4'), 1.28 (d, J= 6.9 Hz, 3H, H16), 1.27 (d, J= 6.2 Hz, 3H, H6"), 1.24 (d, J= 6.8 Hz, 3H, H6'), 1.22 (d, J= 6.8 Hz, 3H, H17), 1.17 (d, J= 7.3 Hz, 3H, H15), 1.03 (d, J= 6.6 Hz, 3H, H20), 1.01 (d, J= 6.2 Hz, 3H, H19); ¹³C NMR (100 MHz, CD₃OD): δ = 209.3 (C9), 179.5 (Cl), 101.8 (Cl'), 99.7 (Cl"), 82.5 (C2'), 82.4 (Cl'"), 81.8 (C5), 79.5 (C3), 79.2 (C2'")₃ 79.1 (C4"), 78.2 (C3"), 72.3 C13), 70.6 (CI l), 70.0 (C5"), 69.5 (C6"^!) 69.1 (C5')₅ 66.1(C3'), 65.1 (C8), 62.0 (C4'"), 61.9 (C3'")_; 57.3 (C7"), 49.3 (C18), 47.3 (C5'"), 46.4 (ClO), 45.8 (C2), 45.2 (C4), 42.3 (C12), 39.7 (C7'), 39.7 (C8'), 35.0 (C2"), 33.1 (C6), 32.8 (C7), 29.9 (C4'), 21.3 (C6'), 19.2 (C17), 18.1 (C6"), 17.8 (C14), 13.1 (C15), 9.6 (C16), 9.2 (C20), 7.3 (C19);MS (ESI-); HMBC, Hl'" (4.60 ppm) coupled with C2' (82.5 ppm) confirmed thioglucosylation of O-2'. HRMS m/z 900.4174 ([M + Cl^"]^"), (calc. 900.4182). 2. Results

Enzyme expression

The respective MGT₅ OleD and Olel, genes were inserted into the Miniprseta E. coli expression vector and the resultant recombinant plasmids were transformed into E. coli C41(DE3). IPTG-induced recombinant protein expression followed by fractionation using SDS-PAGE showed proteins of expected mass and all three proteins were purified by (His)₆-tag cobalt affinity chromatography (Talon, 100 mM imidazole eluant) to >95 % homogeneity as judged by SDS-PAGE at levels in excess of ~60 mg/L of culture (Figure 1). This expression of glycosyltransferases using E. coli at high levels is highly unusual as the vast majority of glycosyltransferases belonging to GT-I (and other GT families) are normally very poorly expressed in heterologous hosts as they frequently contain transmembrane regions leading to protein aggregation and lethality. The production of these enzymes to a high level of purity greatly facilitated their exploitation as "tool kits" in remodelling of antibiotic glycosylation on a synthetically useful scale.

Kinetic parameters

Full kinetic parameters were determined using mass spectrometric monitoring (Figure 2) coupled with pseudo-spiking (M. Yang, M. Brazier, R. Edwards, B. G.

Davis, submitted 2004) calibration as a robust and rapid technique allowing the ready acquisition of biocatalytic data (Table 2). Reciprocal regression analysis of the data employed the rapid equilibrium assumption and assumed no a priori role for donor or acceptor e.g., 1, in the reaction i.e., conducted for both substrate A = I₅ B = UDP- GIc; and B = 1 , A = UDP-GIc.

Table 2- Full Kinetic Parameters for OleD, Olel and MGT with Oleandomycin

[a] Taken from L. M. Quirόs, J. A. Salas, J. Biol. Chem. 1995, 270, 18234.

Low K_M values shown by OleD for both 1 and UDP-GIc indicate that they are both good substrates for this enzyme. In addition, the low K_A/K_B ratio between UPDG and oleandomycin (K_A/K_B = 1.1) is interpreted as being consistent with a random Bi-Bi mechanism with almost no preference of the enzyme for either substrate (M. Yang, M. Brazier, R. Edwards, B. G. Davis, submitted 2004). No previous kinetic data for the OleD-catalyzed transfer of GIc to oleandomycin 1 exists, although previous analysis of the transfer to the alternative macrolide lankomycin suggested in that case an ordered Bi-Bi mechanism (L. M. Quirόs, R. J. Carbajo, A. F. Brana, J. A. Salas, J. Biol. Chem. 2000, 275, 11713). For MGT data, including KA/KB = 2.6, are most consistent with a random Bi-Bi mechanism, with UDP-GIc as only a slightly preferred substrate. K_M values for each substrate that are much larger than those of OleD and Olel correlate with the behaviour of MGT as a macrolide glycosyltransferase that is not specific to oleandomycin 1 and suggest that the low affinity of the enzymes for the full range of the substrates might be compensated in vivo by high turnover rate operate and/or higher ambient substrate concentrations. This is also consistent with the notion that Olel has been evolutionary fine tuned for the specific glycosylation of endogenous oleandomycin during oleandomycin production, whilst MGT has not.

Surprisingly, OleD shares 47% amino acid identity with Olel and 80% with MGT, it is thought to be employed in the glycosylation of exogenous macrolides (like MGT) and yet the K_M values were more akin to Olel than MGT. This may reflect evolution of OleD under the selective pressure exerted by continuous re- exposure to endogenous oleandomycin.

Additional kinetic results are shown in Tables 3 to 10 (below) and Figures 3 to 9.

Table 3 - Regression for OleD with the hypothesis of Oleandomycin as A (see Figure 3)

Equation Y=1/(P1/A*B/(P2*P4+P3*B+P4/A+B/A))

V_max Pl 0.00095 ± 0.00022 niM-min^"1

Ki(oleandomycin) P2 165 ± 35 μM

Km(oleandomcyin) P3 32 ± 8 μM

K_m(uDPG) P4 36 ± 15 μM

Table 4 - Regression for OleD with the hypothesis of UDPGIc as A (see Figure 3)

Equation Y=1/(P1/A*B/(P2*P4+P3*B+P4/A+B/A))

V_max Pl 0.00095 ± 0.00022 rnM-min^"1

Km(oleandomcyin) P4 32 ± 8 μM

Table 5 - Kinetic parameters of OleD with UDP5SGlc and Oleandomycin as substrates (see Figure 4)

V_max 2.12 E-4 niM-min^'1

KM(UDPSSG) 37.81 μM k ^■c_ca_at_t ^: =Vmax/En 0.00328 s^"1 k_cat/K_M (UDP₅SG) 86.70 M^" S^"

Table 6 - Regression for Olel with the hypothesis of Oleandomycin as A (see Figure 5)

Equation Y=1/(P1/A*B/(P2*P4+P3*B+P4/A+B/A))

V_max Pl 0.00093 ± 0.00004 niM-min^"1

^(oleandomycin) P2 18 ± 0.3 μM

K_M(oleandomcyin) P3 4.8 ± 0.5 μM

KM(UDPG) P4 97 ± 6 μM

KM(UDPG)/KM(oieandomcyin) ⁼ 20, indicates an Ordered Bi Bi mechanism with the enzyme binding Oleandomycin first. k_Cat =V_max/Eo 0.042 s^'1 kcat/KM (Oleandomycin) 8719 M^" s^" k_cat/K_M (UDPG) 437 M^-1S^"1

Table 7 - Kinetic parameters for Olel with UDP5SGlc and Oleandomycin as substrates (see Figure 6).

KM(UDP₅SG) 129 μM k_cat =Vmax/Eo 0.013 s^"1 WKM (UDP₅SG) 100 M⁴S^"1 Table 8 - Regression for MGT with the hypothesis of Oleandomycin as A (see Figure 8)

Equation Y=1/(P1/A*B/(P2*P4+P3*B+P4/A+B/A))

V_max Pl 0.024 ± 0.017 mM-niin ¹

Ki_{(0leand0mycin)} P2 172 ± 8 μM

K₁(UDPG) P2 65 ± 0.4 μM

KM(oleandomcyin) P3 1305 ± 885 μM

Table 9 - Kinetic parameters for MGT with UDP5SGlc and Oleandomycin as substrates (see Figure 9)

» max 0.023 mM-min^"

KM(UDP₅SG) 200 μM k_cat =Vmax/E₀ 1.8 s-¹ kcat/KM (UDP₅SG) 8990 M-¹S^"1

Table 10 - ¹H and ¹³C NMR data of glycosylated oleandomycins 1-Gal and 1- 5SGIc in CD₃OD (see Figure 10)

1-Gal 1-5SGIc

iH iH

13C I ppm J(H, 13C I ppm J(H,

Site /ppm Multiplicity lppm H)(Hz) Multiplicity H)(Hz)

1 179.1 - 179.5 -

2 45.2 2.77 m 45.8 2.77 m

3 79.5 3.85 d 4.6 79.5 3.81 d 5.3

4 45.5 1.86 m 45.2 1.93 m

5 82.4 3.58 m 81.8 3.54 d 10.2

6 34.0 1.92 m 33.1 1.91 m

7 33.5 1.61 dd 10.2, 10.9 32.8 1.77 dd 12.6,15.8

2.58 dd 3.4, 12.3 2.62 dd 4.1 , 15.1

8 65.0 - 65.1 -

9 209.5 - 209.3 -

10 46.3 3.20 dd 1.7, 5.1 46.4 3.21 d 6.0

11 70.2 3.80 dd 2.4, 4.6 70.6 3.88 d 6.7

12 42.4 1.67 t 7.6 42.3 1.65 m

13 72.2 5.37 dq 1.7, 6.4 72.3 3.58 dq 1.8, 7.2

14 17.6 1.33 d 6.2 17.8 1.33 d 5.9

15 12.5 1.21 d 6.8 13.1 1.17 d 7.3

16 9.5 1.23 d 7.2 9.6 1.28 d 6.9

17 19.2 1.18 d 6.5 19.2 1.22 d 6.8

18 49.5 2.78 d 5.7 49.3 2.77 m

2.83 d 5.5 2.81 m

19 7.5 1.00 d 6.9 7.3 1.01 d 6.2

20 9.0 1.02 d 6.9 9.2 1.03 d 6.6 r 102.5 4.40 d 7.2 101.8 4.30 d 7.9

2' 81.5 3.48 m 82.5 3.30 m

3¹ 67.1 2.95 m 66.1 111 m

4' 30.0 1.38 m 29.9 1.31 m

1.90 m 1.83 m

5' 69.5 3.70 m 69.1 3.56 m

6¹ 21.2 1.24 d 5.4 21.3 1.24 d 6.8

T,

39.5 2.45 S 39.7 2.30 S 8'

1" 98.5 4.97 d 2.3 99.7 4.97 d 2.7

3.7, 13.0,

2" 34.9 1.49 4.1 , 13.2, ddd 35.0 1.44 ddd 13.3 14.2

2.41 m 2.45 dd 6.2, 14.1

3" 78.2 3.45 m 78.2 3.43 m

4" 77.3 3.05 t 9.6 79.1 3.04 t 8.9

5" 69.8 3.72 m 70.0 3.67 m

6" 18.0 1.28 d 6.8 18.1 1.27 d 6.2 7" 57.3 3.45 _s 57.3 3.43 S

1 106.2 4.38 d 9.1 82.4 4.60 d 9.2

2 73.6 3.62 t 8.6 79.2 3.60 t 9.1

3 74.3 3.49 _m 61.9 3.69 n

4 68.2 3.87 d 3.6 62.0 4.01 d( 3.7. 10.9

5 77.3 3.53 m 47.3 2.81 m

6 61.5 3.70 dd 4.5, 9.4 69.5 3.44 m

3.81 dd 4.0, 7.1 3.65 m

Substrate specificity

To investigate the full extent of the substrate specificity of the three glycosyltransferases, a high throughput library screening format was used to assess the ability of these three enzymes to transfer to a variety of potential acceptors (Figures 11 and 12) using a "Green- Amber-Red" qualitative notation based on the relative signal/noise ratio of expected product Total Ion Count (TIC) peaks [Green S/N >10, Amber S/N 1-10] (M. Yang, M. Brazier, R. Edwards, B. G. Davis, submitted 2004). The results of these studies are shown in Figures 1 IA, 1 IB, 12A and 12B. Figures 1 IA and 1 IB show the results from one experiment, while Figures 12A and 12B show the results from another. Figures 1 IA and 12A show the "Green- Amber-Red" results. Figures 1 IB and 12B show the same results as Figures 1 IA and 12A respectively in written form (H = high/green, M = medium/amber and L = low/red). Broadly, these indicated that oleandomycin, flavanols, coumarins, and other aromatics, such as 3,4-dichloroaniline 8, are all acceptors for OleD, Olel and MGT. These data are very surprising as they indicate that Olel displays a broad acceptor specificity similar to that of OleD and MGT. This is in contrast to previous studies, which have pointed to Olel exhibiting tight specificity for the aglycone substrate, albeit with other macrolide acceptors (L. M. Quirόs, J. A. Salas, J. Biol. Chem. 1995, 270, 18234).

The specificity of the three GTs suggest that the enzymes have not evolved to recognise the basic macrolide as substrate but rather are able to accept a range of planar, cyclic, hydrophobic molecules in the aglycone binding site. Indeed, the reaction of oleandomycin 1 involved transfer to OH-2 of the unusually hydrophobic deoxysugar desosamine, which other than OH-2 contains only the hydrophobic substituents CH₃ at C-5 and N(CH₃)₂ at C-3. All 3 enzymes failed to transfer to other more hydrophilic carbohydrate acceptors 12, 33 and 37. This is despite the fact that the latter contains a large hydrophobic aglycone akin to oleandomycin: presumably desosamine is sufficiently hydrophobic whereas GIcNAc is not.

In this context it is interesting to note the striking similarity between this novel acceptor specificity of Olel, OleD and MGT for coumarins and coumarin-like molecules discovered here and the activity of the aminocoumarin glycosyltransferase NovM found in the biosynthetic pathway of novobiocin (C. L. F. Meyers, M. Oberthur, J. W. Anderson, D. Kahne, C. T. Walsh, Biochemistry 2003, 42, 4179). Indeed, this has allowed us here to use OleD, Olel and MGT to create a 3-Glc analogue of a recently NovM-synthesized 3-noviose novobiocin analogue (C. L. F. Meyers, M. Oberthur, J. W. Anderson, D. Kahne, C. T. Walsh, Biochemistry 2003, 42, 4179) all with enhanced kinetic efficiency. The kinetic data in C. L. F. Meyers, M. Oberthur, J. W. Anderson, D. Kahne, C. T. Walsh, Biochemistry 2003, 42, 4179 uses only pseudo single substrate conditions and therefore comparison was only possible on this basis rather through full bisubstrate kinetic parameters.

Sugar donor plasticity

To explore the specificity these powerful enzymes further, a donor molecule GAR screen was employed to determine the plasticity in both sugar and nucleotide recognition with a selected number of substrates from our model acceptor array. The results of these studies are shown in Figures 13A, 13B, 14A, 14B, 15A and 15B. Figures 13, 14 and 15 each show results from different experiments. Figures 13 A, 14A and 15B show the "Green- Amber-Red" results. Figures 13B, 14B and 15B show the same results in written form (H = high/green, M = medium/amber and L = low/red). This revealed that the non-natural donor, UDP-5 S-GIc (M. Yang, M. Brazier,

R. Edwards, B. G. Davis, submitted 2004; and M. Yang, PhD thesis, University of Huddersfield 2003) is also a substrate for OleD, Olel and MGT. Using a fixed concentration of oleandomycin 1, pseudo-single substrate enzyme kinetic data were obtained (OleD: V_max 0.21 μM-min^"1, K_M(_UDP_5SG) 37.8 μM, k_cat 0.0033 s^"1; Olel: V_max 0.19 μM-min^"1, K_M(UDP5SG) 129 μM, k_cat 0.013 s^"1; MGT: V_max 23 μM-min^"1, K_M(UDPSSGIC) 200 μM, k_cat 1.8 s^"1, k_cat/K_M (UDPSSGIC) 8990 M^'V¹ all at [1]= 50 μM). For OleD and Olel, the similar K_M values for UDP-GIc and UDP-5 S-GIc indicate that despite the change in endocyclic heteroatom the binding affinities of OleD and Olel for UDP-GIc and UDP-5SGlc are similar.

Intriguingly the 2.5-fold lower K_M for UDP-5S-Glc than for UDP-GIc indicates better binding under these conditions by MGT to UDP-5 S-GIc. More dramatic effects were observed on k_cat; OleD k_cat(UDP-5 S-GIc) is more than an order of magnitude lower while that for Olel is reduced to a third of its UDP-GIc value. In contrast, k_cat(UDP-5S-Glc) for MGT is twice that for UDP-GIc. The transition state of transfer is thus better stabilised by an α-sulfur in the order MGT>OleI>OleD and importantly highlights the utility of UDP-5 S-GIc as a mechanistic probe in addition to it providing a source of non-natural glycan in GT-catalysed glycosylations. These results are consistent with OleD and Olel's more stringent specificities and suggest a less electropositive, early transition state for MGT.

Other sugars screened included UDP-GIcNAc, UDP-GaI, UDP-Man (M. Yang, M. Brazier, R. Edwards, B. G. Davis, submitted 2004), GDP-Man and GDP- Fuc. These screens revealed that OleD also transfers UDP-GIcNAc (to acceptors 2, 4 and 23 but not oleandomycin) and UDP-GaI (to only 23), but not UDP-Man, GDP- Fuc, or GDP-Man (Figure 13A and 13B(a)). Olel, although showing strong activity using UDP-5S-Glc with oleandomycin as a co-subtrate in contrast only showed minimal activity using UDP-5 S-GIc towards a range of other acceptor substrates (Figure 13A and 13B(b)).

Excitingly and surprisingly Ole I also showed strong activity using UDP-GaI, again only with oleandomycin 1 as a substrate. All other donors either did not function as substrates or gave minimal activity for this more stringent enzyme. MGT transferred 5-S-Glc to oleandomycin 1 and showed some low activity towards flavanols 20,21,23,30,31 but showed little or no activity for all other donors screened. Intriguingly, OleD and MGT showed some low but noticeable activity when using UDP-Man as a donor substrate for the acceptors luteolin 20 and baicalein 2, respectively. It should be noted that the non-natural donor UDP-Man (M. Yang, M. Brazier, R. Edwards, B. G. Davis, submitted 2004) by virtue of its nucleotide U is accepted by MGT whereas the natural GDP-Man (and indeed GDP-Fuc) is not.

For the results shown in Figure 15, all the enzymes are restricted in the C2 and C3 configuration, and tolerate of the change of ring oxygen to sulfur. MGT and Olel cannot (or very weakly) use other bases rather than Uridine, indicating the donor binding pocket of MGT and Olel are quite similar but different from that of OleD, which uses U, G and TDP sugars. OleD also tolerates the changes on C6 and C4. These donor results indicate that there is tolerance by all enzymes for ring heteroatom alteration, by OleD for varied functionality at C-2 (although the relatively low capacity to incorporate UDP-Man compared to UDP-GIc highlights the need for correct configuration at C-2) and by OleD and Olel for configurational flexibility at C-4 (as demonstrated by their ability to transfer Gal). The observation that the non-natural variant UDP-Man but not GDP-Man (or GDP-Fuc) can act as a substrate for OleD and MGT, albeit with low activity, demonstrates the important role of the nucleotide as a molecular determinant of substrate specificity for this enzyme. Interestingly, these results also parallel broad sugar moiety plasticity observed for powerful transferase inhibitors in which the donor sugar moiety was successfully replaced by aromatic groups (B. Muller, C. Schaub, R. R. Schmidt, Angew. Chem. Intl Ed. 1998, 37, 2893; and A. K. Bhattacharya, F. Stolz, R. R. Schmidt, Tetrahedron Lett. 2001, 42, 5393).

The breadth of substrate tolerance, especially for acceptors, demonstrated here for these macrolide glycosyltransferases is highly unusual for a class of enzymes normally regarded as highly stringent in their specificity and highlights their strong potential in the chemical synthesis of novel polyketide (N. L. Pohl, Curr. Opin. Chem. Biol. 2002, 6, 773) and coumarin (C. L. F. Meyers, M. Oberthur, J. W. Anderson, D. Kahne, C. T. Walsh, Biochemistry 2003, 42, 4179) antibiotics. Moreover, initial promising indications in donor tolerance have allowed the transfer of GIcNAc, 5S-GIc and Gal to a number of acceptor substrates and notably the novel macrolides oleandomycin-5 S-GIc (1-5S-GIc), oleandomycin-Gal (1-Gal) were synthesized on a preparative scale in 82% and 61% yields, using OleD and Olel, respectively (Figure 16). Gradient phase sensitive HSQC coupling constants [¹J C₁- Hr, (-160-165 Hz); C₁-H₁- (-150-155 Hz); Cr-H₁- (-150-155 Hz) and ³J H₁ - H₂- (~9 Hz)] along with H₁-C_2" HMBC coupling established the exclusive and excellent stereoselectivity (C₁- α, Ci» β, C₁'" β) and regioselectivity (O-2" glycosylation) of the transformations, respectively. We are currently exploring the potency of these unnatural, remodelled macrolide variants as potential antibiotics (or pro-antibiotics) with different modes of action and the exciting prospect that nonnatural glycosylated macrolides 1-5S-GIc and 1-Gal may also be inhibitors of the glycosidase that is critical to the protection of S. antibioticus because thio sugars such as found in 1-5S-GIc are known inhibitors of glycosidases (L. M. Quiros, I. Aguirrezabalaga, C. Olano, C. Mendez, J. A. Salas, Molec. Microbiol. 1998, 28, 1177).

Antibiotic activity An LB-agar plate containing Kanamycin 50 μg/ml was divided into 4 parts

(as shown in Figure 19). Top left area contains Oleandomycin (50 μg/ml), bottom left contains 1-Glc (50 μg/ml), top right contains 1-Gal (50 μg/ml) and bottom right contains 1-Gal (50 μg/ml) and IPTG (1 mM). E. CoIi. BL21 transformed with a- galactoside DNA (ssaG) overnight media (5 μL) was plated on each part and incubated at 37⁰C for overnight. The bottom right area shown inhibited growth of cells.

Antibiotic uptake

25 ml of ssβG overnight cells was harvested by centrifuge (3000 rpm, 4⁰C, 30min) and resuspend in TRIS buffer (5 ml, 10 mM, pH 7.0). Cells were starved by shaking at 37⁰C, 200 rpm for 2 hours and Gal-Ole was added to make the final concentration of 50 μg/ml and continue shake at the same condition. Aliquots (20 μL) were taken out every 15 or 30 min. Each aliquot was added with EtOH (20μL) and centrifuged (14000rpm, 10 min) to remove cells. Supernatant (30 μL) was mixed with TRIS buffer (30 μL) and analyzed by LCMS using the method reported (M. Yang, M. Brazier, R. Edwards, B. G. Davis, ChemBioChem accepted). After 540 min, cells media (2 ml) was centrifuged (14000 rpm, lOmin) and cells were washed carefully with TRIS buffer (3 >< 1 ml) to remove any contamination. Cells were harvested, resuspended in 0.5 ml TRIS buffer and broken by sonication (3 ^χ l min, O⁰C). Supernatant was obtained by centrifuge and analyzed by LC/MS.

Figure 20 shows the disappearance of 1-Gal in the solution. Cell extraction at 540 min show that there was 1-Gal equal to 14 μg/ml while worked from the average in the above figure, 13.5 μg/ml was expected.

The experiment was done the same as above expect using Tuner, which lacks lactose permease lacy element, instead of ssβG. The results are shown in Figure 21. Uptake of only ~4 μg/ml suggest that lactose permease's recognistion of GaI- appended structures is responsible for the active uptage of Gal-Ole. Once internalized Gal-Ole is hydrolyzed by the recombinant galactosidase to Ole; this combined active uptake and "uncapping" causes enhanced antibacterial action.

Minimum Inhibition Concentration (MIC) test A series of tubes containing LB base, Kanamycin (50 μg/ml), IPTG (0.1 mM). One set of tubes contains Gal-Ole, 0, 3 5, 25, 50, 100 μg/ml respectively, total volume ImI; another set of tubes contains Glc-Ole, 0, 3, 5, 25, 50, 100 μg/ml espectively, total volume 2ml. Galactose was kept in the same condition but with total volume of 5 ml. Glucose concentration was kept at 3, 5, 25, 50, 100 μg/ml with the total volume of 5 ml. Oleandomycin was kept at 0, 50, 100, 250, 500, 1000 μg/ml with the total volume of 5 ml. All tubes were inoculated with 1/10,000 overnight media and shake at 37°C, 250 rpm for overnight. Media were diluted 10⁵ and 10⁶ times and plated on LB Agar plate, then incubated at 37⁰C for overnight. Single clonies were counted. The results are shown in Tables 11 and 12.

Table 11- OD600 data

Concentration (/μg/ml) 0 3 5 25 50 100

Gal-Ole 1.367* 1.141 1.146 1.012 1.045 1.101

Glc-Ole 1.014* 0.991 0.967 1.026 1.050 1.043 Galactose 0.637* 0.631 0.627 0.616 0.594 0.609

Glucose - 0.634 0.625 0.643 0.635 0.655

Concentration (/μg/ml) 0 50 100 250 500 1000

Oleandomycin 0.624* 0.623 0.581 0.529 0.472 0.307

* The difference of the OD600 with no added antibiotics was caused from the different volumes.

These results show Glc-Ole has no effect on the cells growth, so do galactose and glucose. Oleandomycin give relative strong inhibition starting from 250-500 μg/ml. As for Gal-Ole, although it's not very clear, an inhibition can be found starting from 25 μg/ml. Single colony count was used to confirm this result.

Table 12 - Single colony counts

Oleandomycin (/μg/ml) 0 50 100 250 500 1000

10⁵ dilution 389 303 215 67 88 78

10⁶ dilution 40 49 23 32 6 2

Gal-Ole (/μg/ml) 0 3 5 25 50 100

10⁵ dilution 430 482 508 187 175 188

10⁶ dilution 92 103 155 53 39 40

Glc-Ole (/μg/ml) 0 3 5 25 50 100

10⁵ dilution 238 192 209 215 218 237

10⁶ dilution 58 38 41 43 43 53 This data is consistent with the OD600 results. Glc-Ole has no antibiotic activity. Oleandomycin inhibits cell growth at 250 μg/ml, but GalOle starts to inhibit the cell growth at 25 μg/ml.

Claims

1. A method of attaching a sugar to a substrate comprising contacting the substrate with a macrolide glycosyltransferase enzyme in the presence of a sugar donor.

2. A method according to claim 1, wherein the macrolide glycosyltransferase enzyme is the macrolide glycosyltransferase (MGT) from Streptoniyces lividans.

3. A method according to claim 1 , wherein the macrolide glycosyltransferase enzyme is Olel from Streptomyces antibioticus.

4. A method according to claim 1 , wherein the macrolide glycosyltransferase enzyme is OleD from Streptomyces antibioticus.

5. A method according to any one of claims 1 to 4, wherein the substrate is a planar, cyclic, hydrophobic molecule.

6. A method according to any one of claims 1 to 5, wherein the substrate is a polyketide antibiotic.

7. A method according to claim 6, wherein the polyketide antibiotic is a macrolide.

8. A method according to any one of the preceeding claims, wherein the substrate is of the formula I:

O oλ ^(I) wherein X comprises C_3-2O alklyene or C_3-20 alkenylene, and may further comprise one or more nitrogen, sulphur or oxygen atoms; and wherein the C_3-20 alkylene or C_3-20 alkenylene group in X is substituted by one or more -OR¹ groups wherein R¹ is a sugar and is optionally further substituted by one or more substituents selected from:

C_1-6 alkyl; C_1-6 alkenyl; =O

-OR wherein R is selected from H, C_1-6 alkyl and C_1-6 alkenyl;

-(CH₂)_n-COR wherein n = 0-6 and wherein R is selected from H and C_1-6 alkyl; and

-O(CH₂)_n- wherein n = 1 -6 and wherein the O and the terminal carbon atom are both bonded to the same carbon atom of the C_3-2O alkylene or C_3-2O alkenylene group.

9. A method according to claim 6, wherein the polyketide is oleandomycin, erythromycin or rapamycin.

10. A method according to any one of claims 1 to 5, wherein the substrate is a cyclic non-ribosomal peptide antibiotic.

11. A method according to claim 10, wherein the non-ribosomal cyclic peptide is balhimycin, teicoplanin or vancomycin.

12. A method according to any one of claims 1 to 5, wherein the substrate is a coumarin-class antibiotic.

13. A method according to claim 12, wherein the coumarin-class antibiotic is coumermycin Al or novobiocin.

14. A method according to any one of claims 1 to 5, wherein the substrate is an aminoglycoside antibiotic.

15. A method according to claim 14, wherein the aminoglycoside antibiotic is genatmycin, neomycin or streptomycin.

16. A method according to any one of claims 1 to 5, wherein the substrate is a sugar.

17. A method according to any one of the preceding claims, wherein the sugar donor comprises uridinediphosphate (UDP).

18. A method according to claim 17, wherein the sugar donor is of the formula:

S

19. A method according to claim 17 or 18, wherein the sugar donor comprises a monosaccharide.

20. A method according to claim 19, wherein the monosaccharide is of the formula: In one embodiment, the monosaccharide is of the formula III:

wherein Y is O, S or NH; wherein R⁴ is selected from:

H;

C_1-6 alkyl;

C_1-6 alkenyl;

NH₂;

-(CH₂)_n-OH wherein n = 0 to 6;

OR⁵ wherein R⁵ is H or C_1-6 alkyl;

-(CH₂)_n-COR⁶ wherein n = 0 to 6 and wherein R⁶ is selected from H and C_1-6 alkyl; and wherein R⁷ is selected from H, OH and -(CH₂)_n-OH wherein n = 0 to 6.

21. A method according to claim 19, wherein the monosaccharide is of the formula IV:

wherein wherein Y is O, S or NH; wherein R is selected from:

H;

Ci_-6 alkyl;

C_1-6 alkenyl;

NH₂;

-(CH₂)_n-OH wherein n = 0 to 6;

OR⁹ wherein R⁹ is H or C_1-6 alkyl; -(CH₂)_n-COR¹⁰ wherein n = 0 to 6 and wherein R¹⁰ is selected from H and C_1-6 alkyl; and wherein R¹¹ is selected from H, OH and -(CH₂VOH wherein n = 0 to 6.

22. A method according to any one of claims 19 to 21 , wherein the monosaccharide is glucose (GIc), N-acetylglucosamine (GIcNAC), fructose (Frc), galactose (Gal) or fucose (Fuc).

23. A method according to any one of claims 17 to 22, wherein the sugar donor is UDP-GIc, UDP-5S-Glc or UDP-GIcNAc.

24. A method according to claim 17 or 18, wherein the sugar donor comprises a disaccharide.

25. A method according to claim 24, wherein the disaccharide is sucrose

(Sue), lactose (Lac), maltose (MaI), isomaltose (Isomal), trehalose (Tre) or cellobiose.

26. A method of producing a macrolide glycosyltransferase enzyme, which comprises culturing a bacterial cell containing a heterologous polynucleotide encoding the enzyme under conditions in which the enzyme is expressed, and recovering the enzyme.

27. A method according to claim 26, wherein the glycosyltransferase enzyme is expressed at a level greater than 20mg/L of culture.

28. A method according to claim 26 or 27, wherein the bacterial cell is Escherichia coli.

29. A method according to any one of claims 26 to 28, wherein the glycosyltransferase enzyme is the macrolide glycosyltransferase (MGT) from Streptomyces lividans.

30. A method according to any one of claims 26 to 28, wherein the glycosyltransferase enzyme is Olel from Streptomyces antibioticus.

31. A method according to any one of claims 26 to 28, wherein the glycosyltransferase enzyme is OleD from Streptomyces antibioticus.

32. A bacterial cell which expresses a heterologous polynucleotide encoding a macrolide glycosyltransferase enzyme.

33. A cell according to claims 32, wherein the glycosyltransferase enzyme is expressed at a level greater than 20mg/L of culture.

34. A cell according to claim 32 or 33, which is an Escherichia coli cell.

35. A cell according to any one of claims 32 to 34, wherein the glycosyltransferase enzyme is the macrolide glycosyltransferase (MGT) from Streptomyces lividans.

36. A cell according to any one of claims 32 to 34, wherein the glycosyltransferase enzyme is Olel from Streptomyces antibioticus.

37. A cell according to any one of claims 32 to 34, wherein the glycosyltransferase enzyme is OleD from Streptomyces antibioticus.

38. A bacterial expression vector comprising a nucleic acid sequence encoding a macrolide glycosyltransferase enzyme.

39. A vector according to claim 38, wherein the glycosyltransferase enzyme is the macrolide glycosyltransferase (MGT) from Streptomyces lividans.

40. A vector according to claim 38, wherein the glycosyltransferase enzyme is Olel from Streptomyces antibioticus.

41. A vector according to claim 38, wherein the glycosyltransferase enzyme is OleD from Streptomyces antibioticus.