Figure 1. Flow chart for generation of glycoenzyme expression constructs.

The general workflow for the generation of human glycoenzyme expression constructs is depicted in the flow chart. A master glycogene list was developed as described in Experimental Procedures. Construct designs were generated based on the presence and location of TMD sequences and cleavable signal sequences and available knowledge regarding the location of the enzyme active sites, as described in Experimental Procedures. Truncated or full-length coding regions were incorporated into Gateway “donor” vectors by PCR from either Mammalian Gene Collection clone templates or human cDNA sources or, in some cases, by gene synthesis. Each of the enzyme coding regions contained a TEV protease recognition site appended to either the NH₂- or COOH-terminal end of the open reading frame depending on the respective fusion protein strategy (Supplementary Fig. 1). Thus, a library of all human glycoenzyme constructs (“glycogenes”) was captured in donor vectors that could be transferred to Gateway DEST vectors by Gateway LR recombination (see Experimental Procedures). Custom mammalian or baculovirus DEST vectors were generated harboring additional in-frame fusion tags (Supplementary Figs. 2 and 3) and employed to produce glycogene expression constructs for each recombinant host system. Expression and secretion of recombinant enzymes were examined by batch-mode transient transfection (HEK293 cells) or baculovirus infection (insect cells) followed by separation into cell and media fractions. Secreted products were used for protein purification to produce enzyme preparations with utility for enzymology, structural studies, enzymatic glycan synthesis and numerous other applications.