High level in vivo mucin-type glycosylation in Escherichia coli.

BACKGROUND: Increasing efforts have been made to assess the potential of Escherichia coli strains for the production of complex recombinant proteins. Since a considerable part of therapeutic proteins are glycoproteins, the lack of the post-translational attachment of sugar moieties in standard E. coli expression strains represents a major caveat, thus limiting the use of E. coli based cell factories. The establishment of an E. coli expression system capable of protein glycosylation could potentially facilitate the production of therapeutics with a putative concomitant reduction of production costs. RESULTS: The previously established E. coli strain expressing the soluble form of the functional human-derived glycosyltransferase polypeptide N-acetylgalactosaminyltransferase 2 (GalNAc-T2) was further modified by co-expressing the UDP-GlcNAc 4-epimerase WbgU derived from Plesiomonas shigelloides. This enables the conversion of uridine 5'-diphospho-N-acetylglucosamine (UDP-GlcNAc) to the sugar donor uridine 5'-diphospho-N-acetylgalactosamine (UDP-GalNAc) in the bacterial cytoplasm. Initially, the codon-optimised gene wbgU was inserted into a pET-derived vector and a Tobacco Etch Virus (TEV) protease cleavable polyhistidine-tag was translationally fused to the C- terminus of the amino acid sequence. The 4-epimerase was subsequently expressed and purified. Following the removal of the polyhistidine-tag, WbgU was analysed by circular dichroism spectroscopy to determine folding state and thermal transitions of the protein. The in vitro activity of WbgU was validated by employing a modified glycosyltransferase assay. The conversion of UDP-GlcNAc to UDP-GalNAc was shown by capillary electrophoresis analysis. Using a previously established chaperone pre-/co- expression platform, the in vivo activity of both glycosyltransferase GalNAc-T2 and 4-epimerase WbgU was assessed in E. coli, in combination with a mucin 10-derived target protein. Monitoring glycosylation by liquid chromatography electrospray ionization mass spectrometry (LC-ESI-MS), the results clearly indicated the in vivo glycosylation of the mucin-derived acceptor peptide. CONCLUSION: In the present work, the previously established E. coli- based expression system was further optimized and the potential for in vivo O-glycosylation was shown by demonstrating the transfer of sugar moieties to a mucin-derived acceptor protein. The results offer the possibility to assess the practical use of the described expression platform for in vivo glycosylations of important biopharmaceutical compounds in E. coli.

Expression of the functional recombinant human glycosyltransferase GalNAcT2 in Escherichia coli.

BACKGROUND: Recombinant protein-based therapeutics have become indispensable for the treatment of many diseases. They are produced using well-established expression systems based on bacteria, yeast, insect and mammalian cells. The majority of therapeutic proteins are glycoproteins and therefore the post-translational attachment of sugar residues is required. The development of an engineered Escherichia coli-based expression system for production of human glycoproteins could potentially lead to increased yields, as well as significant decreases in processing time and costs. RESULTS: This work describes the expression of functional human-derived glycosyltransferase UDP-GalNAc:polypeptide N-acetylgalactosaminyltransferase 2 (GalNAcT2) in a recombinant E. coli strain. For expression, a codon-optimised gene encoding amino acids 52-571 of GalNAcT2 lacking the transmembrane N-terminal domain was inserted into a pET-23 derived vector encoding a polyhistidine-tag which was translationally fused to the N-terminus of the glycosyltransferase (HisDapGalNAcT2). The glycosyltransferase was produced in E. coli using a recently published expression system. Soluble HisDapGalNAcT2 produced in SHuffle® T7 host cells was purified using nickel affinity chromatography and was subsequently analysed by size exclusion chromatography coupled to multi-angle light scattering (SEC-MALS) and circular dichroism spectroscopy to determine molecular mass, folding state and thermal transitions of the protein. The activity of purified HisDapGalNAcT2 was monitored using a colorimetric assay based on the release of phosphate during transfer of glycosyl residues to a model acceptor peptide or, alternatively, to the granulocyte-colony stimulating growth factor (G-CSF). Modifications were assessed by Matrix Assisted Laser Desorption Ionization Time-of-flight Mass Spectrometry analysis (MALDI-TOF-MS) and Electrospray Mass Spectrometry analysis (ESI-MS). The results clearly indicate the glycosylation of the acceptor peptide and of G-CSF. CONCLUSION: In the present work, we isolated a human-derived glycosyltransferase by expressing soluble HisDapGalNAcT2 in E. coli. The functional activity of the enzyme was shown in vitro. Further investigations are needed to assess the potential of in vivo glycosylation in E. coli.

Recombinant strains for the enhanced production of bioengineered rapalogs.

The rapK gene required for biosynthesis of the DHCHC starter acid that initiates rapamycin biosynthesis was deleted from strain BIOT-3410, a derivative of Streptomyces rapamycinicus which had been subjected to classical strain and process development and capable of robust rapamycin production at titres up to 250mg/L. The resulting strain BIOT-4010 could no longer produce rapamycin, but when supplied exogenously with DHCHC produced rapamycin at titres equivalent to its parent strain. This strain enabled mutasynthetic access to new rapalogs that could not readily be isolated from lower titre strains when fed DHCHC analogs. Mutasynthesis of some rapalogs resulted predominantly in compounds lacking late post polyketide synthase biosynthetic modifications. To enhance the relative production of fully elaborated rapalogs, genes encoding late-acting biosynthetic pathway enzymes which failed to act efficiently on the novel compounds were expressed ectopically to give strain BIOT-4110. Strains BIOT-4010 and BIOT-4110 represent valuable tools for natural product lead optimization using biosynthetic medicinal chemistry and for the production of rapalogs for pre-clinical and early stage clinical trials.

Glycosylation engineering of spinosyn analogues containing an L-olivose moiety.

Biosynthetic genes encoding proteins involved in the first steps of deoxyhexose biosynthesis from D-glucose-1-phosphate were expressed in Saccharopolyspora erythraea. The resulting mutant was able to accumulate and utilise TDP-L-olivose. Co-expression of the spinosyn glycosyl transferase SpnP in the resulting mutant endowed upon it the ability to biotransform exogenously added spinosyn aglycones to yield novel spinosyn analogues.

Molecular characterization of macbecin as an Hsp90 inhibitor.

Macbecin compares favorably to geldanamycin as an Hsp90 inhibitor, being more soluble, stable, more potently inhibiting ATPase activity (IC50 = 2 microM) and binding with higher affinity (Kd = 0.24 microM). Structural studies reveal significant differences in their Hsp90 binding characteristics, and macbecin-induced tumor cell growth inhibition is accompanied by characteristic degradation of Hsp90 client proteins. Macbecin significantly reduced tumor growth rates (minimum T/C: 32%) in a DU145 murine xenograft. Macbecin thus represents an attractive lead for further optimization.

Engineered biosynthesis of hybrid macrolide polyketides containing D-angolosamine and D-mycaminose moieties.

The glycosylation of natural product scaffolds with highly modified deoxysugars is often essential for their biological activity, being responsible for specific contacts to molecular targets and significantly affecting their pharmacokinetic properties. In order to provide tools for the targeted alteration of natural product glycosylation patterns, significant strides have been made to understand the biosynthesis of activated deoxysugars and their transfer. We report here efforts towards the production of plasmid-borne biosynthetic gene cassettes capable of producing TDP-activated forms of D-mycaminose, D-angolosamine and D-desosamine. We additionally describe the transfer of these deoxysugars to macrolide aglycones using the glycosyl transferases EryCIII, TylMII and AngMII, which display usefully broad substrate tolerance.

Engineered biosynthesis of novel spinosyns bearing altered deoxyhexose substituents.

Novel spinosyns have been prepared by biotransformation, using a genetically engineered strain of Saccharopolyspora erythraea, in which the beta-D-forosamine moiety in glycosidic linkage to the hydroxy group at C17 is replaced by alpha-L-mycarose.

Optimizing natural products by biosynthetic engineering: discovery of nonquinone Hsp90 inhibitors.

A biosynthetic medicinal chemistry approach was applied to the optimization of the natural product Hsp90 inhibitor macbecin. By genetic engineering, mutants have been created to produce novel macbecin analogues including a nonquinone compound (5) that has significantly improved binding affinity to Hsp90 (Kd 3 nM vs 240 nM for macbecin) and reduced toxicity (MTD > or = 250 mg/kg). Structural flexibility may contribute to the preorganization of 5 to exist in solution in the Hsp90-bound conformation.

Rapamycin biosynthesis: Elucidation of gene product function.

The function of gene products involved in the biosynthesis of the clinically important polyketide rapamycin were elucidated by biotransformation and gene complementation.

Parallel pathways for oxidation of 14-membered polyketide macrolactones in Saccharopolyspora erythraea.

The glycosyltransferases OleG1 and OleG2 and the cytochrome P450 oxidase OleP from the oleandomycin biosynthetic gene cluster of Streptomyces antibioticus have been expressed, either separately or from artificial gene cassettes, in strains of Saccharopolyspora erythraea blocked in erythromycin biosynthesis, to investigate their potential for the production of diverse novel macrolides from erythronolide precursors. OleP was found to oxidize 6-deoxyerythronolide B, but not erythronolide B. However, OleP did oxidize derivatives of erythronolide B in which a neutral sugar is attached at C-3. The oxidized products 3-O-mycarosyl-8a-hydroxyerythronolide B, 3-O-mycarosyl-8,8a-epoxyerythronolide B, 6-deoxy-8-hydroxyerythronolide B and the olefin 6-deoxy-8,8a-dehydroerythronolide B were all isolated and their structures determined. When oleP and the mycarosyltransferase eryBV were co-expressed in a gene cassette, 3-O-mycarosyl-6-deoxy-8,8a-dihydroxyerythronolide B was directly obtained. When oleG2 was co-expressed in a gene cassette together with oleP, 6-deoxyerythronolide B was converted into a mixture of 3-O-rhamnosyl-6-deoxy-8,8a-dehydroerythronolide B and 3-O-rhamnosyl-6-deoxy-8,8a-dihydroxyerythronolide B, confirming previous reports that OleG2 can transfer rhamnose, and confirming that oxidation by OleP and attachment of the neutral sugar to the aglycone can occur in either order. Similarly, four different 3-O-mycarosylerythronolides were found to be substrates for the desosaminyltransferase OleG1. These results provide additional insight into the nature of the intermediates in OleP-mediated oxidation, and suggest that oleandomycin biosynthesis might follow parallel pathways in which epoxidation either precedes or follows attachment of the neutral sugar.

Direct production of ivermectin-like drugs after domain exchange in the avermectin polyketide synthase of Streptomyces avermitilis ATCC31272.

Ivermectin, a mixture of 22,23-dihydroavermectin B1a9 with minor amounts of 22,23-dihydroavermectin B1b 10, is one of the most successful veterinary antiparasitic drugs ever produced. In humans, ivermectin has been used for the treatment of African river blindness (onchocerciasis) resulting in an encouraging decrease in the prevalence of skin and eye diseases linked to this infection. The components of ivermectin are currently synthesized by chemical hydrogenation of a specific double bond at C22-C23 in the polyketide macrolides avermectins B1a 5 and B1b 6, broad-spectrum antiparasitic agents isolated from the soil bacterium Streptomyces avermitilis. We describe here the production of such compounds (22,23-dihydroavermectins B1a 9 and A1a 11) by direct fermentation of a recombinant strain of S. avermitilis containing an appropriately-engineered polyketide synthase (PKS). This suggests the feasibility of a direct biological route to this valuable drug.