In mammals, sialic acids are usually found at the non-reducing terminal position of various glycoconjugates, which play essential roles in many biological processes [1], [2], [3]. Sialyltransferases are key enzymes involved in the biosynthesis of these sialic acid-containing oligosaccharides and glycoconjugates. They catalyze the transfer of a sialic acid residue from its activated sugar nucleotide donor cytidine 5’-monophosphate sialic acid (CMP-sialic acid) to different acceptors, such as galactose (Gal), N-acetylgalactosamine (GalNAc), N-acetylglucosamine (GlcNAc), or another sialic acid residue. Various linkages including α2,3 or α2,6 linked to a Gal residue, or α2,6 linked to a GalNAc residue or a GlcNAc residue can be formed. Sialic acids are also found in α2,8 or α2,9 linkage to another sialic acid residue [4], [5], [6]. Based on protein sequence homology, all sialyltransferases reported to date have been grouped into five carbohydrate-active enzymes (CAZy) glycosyltransferase (GT) families (GT29, GT38, GT42, GT52, and GT80). All known eukaryotic sialyltransferases belong to a single CAZy GT29 family, while nonmammalian sialyltransferases are clustered in CAZy families GT38, 42, 52, and 80 [7], [8]. Like many other glycosyltransferases, sialyltransferases from mammals are found primarily in the Golgi apparatus of cells and are involved in post-translational glycosylation of proteins [9], [10]. Based on substrate specificity, mammalian sialyltransferases include one type of galactoside α2,6-sialyltransferase (ST6Gal I), six types of N-acetylgalactosaminide α2,6-sialyltransferases (ST6GalNAc I-VI), six types of galactoside α2,3-sialyltransferases (ST3Gal I-VI), and five types of α2,8-sialyltransferases (ST8Sia I-VI). N-acetylgalactosaminide α2,6-sialyltransferases ST6GalNAc I-VI catalyze the transfer of a sialic acid residue onto 6-OH of GalcNAc in different substrates. ST6GalNAc I, II, and IV catalyze the formation of α2,6-linkage to GalNAc residues of O-glycans, and three others (ST6GalNAc III, V, and VI) catalyze the addition of sialic acid residues onto GalNAc residues of gangliosides [7].

ST6GalNAc V was first cloned from a cDNA library of mouse brains and could catalyze the synthesis of GD1α from GM1b (Supplementary Table S1) [11]. Mouse ST6GalNAc VI was expressed in a wide range of mouse tissues, such as the colon, liver heart, spleen, and brain. ST6GalNAc VI appears to be specific for glycolipid acceptors and can use GM1b, GD1a, and GT1b as the acceptor substrates but not glycoproteins as the acceptor substrates, catalyzing synthesis of all α-series gangliosides defined so far [12]. Since these two enzymes have been identified, they have been applied in the synthesis of many sialylated glycoconjugates. For example, hST6GalNAc V and hST6GalNAc VI can also catalyze the synthesis of disialyl-lactotetraosylceramide (disialyl Lc4) using sialyl-lactotetraosylceramide (sialyl Lc4) as an acceptor substrate [13]. In addition, disialylgalactosylgloboside (DSGG) is a disialyl glycosphingolipid with a globo-serie core structure and has been detected in renal cancer cells [14]. hST6GalNAc VI is also responsible for the biosynthesis of DSGG from monosialylgalactosylgloboside (MSGG), while hST6GalNAc V did not catalyze the same reaction from MSGG [15]. Disialosyl globopentaosylceramide (DSGb5) is the oligosaccharide moieties of these glycolipids, hST6GalNAc V has also been used to synthesize DSGb5 from monsialosyl globopentaosylceramide (MSGb5), but no product formation was detected when using recombinant hST6GalNAc VI [16]. It’s worth mentioning that these two recombinant enzymes mentioned above were all expressed in eukaryotic cells.

DSLNT (disialyllacto-N-tetraose) is one of the most abundant acidic human milk oligosaccharides (HMOs) [17] and was identified as a specific HMO component that is effective for preventing necrotizing enterocolitis (NEC) in a neonatal rat model [18]. In clinical experiments, it has been observed that a low concentration of DSLNT in the mother’s milk corresponded to an increased risk of NEC in preterm infants [19], [20], [21]. Due to the limited availability of human milk and the absence or the low abundance of DSLNT in bovine milk, it is impractical to obtain DSLNT at a large scale for potential clinical therapeutic applications. Furthermore, the chemical synthesis of DSLNT has not been reported [6]. Prudden and co-workers used a soluble, secreted fusion protein human hST6GalNAc V which was expressed in mammalian HEK293 cells to successfully synthesize DSLNT [22]. However, hST6GalNAc V was expressed in HEK293 cells, which can only be obtained at a high cost and would restrict their large-scale applications. Yu and co-workers used microbial α2,6-sialyltransferase from Photobacterium damselae (Pd2,6ST) to synthesize a series of disialyl HMOs analogues, where sialic acids weren’t attached to C6-OH of GlcNAc and were transferred C6-OH of Gal instead due to substrate specificity of Pd26ST [23], [24]. Even though they have NEC-preventing effects but may cause potential safety issues since these structures are not identified in human milk. To date, there have been no reports of microbial sialyl transferases or homologous counterparts of hST6GalNAc V that can selectively install an α2,6-linked sialyl at GalNAc or GlcNAc of Gal-β1,3-GalNAc/GlcNAc. E. coli is the most popular organism for the production of recombinant proteins due to the well-known advantages it offers over eukaryotic expression systems [25]. Therefore, the expression of functional human sialyltransferases hST6GalNAc V and hST6GalNAc VI in E. coli is one of the potential solutions to overcome the challenges of obtaining DSLNT at a low cost and large scale.

In previous reports [16], [22], a truncated version of hST6GalNAc V lacking the N-terminal 50 amino acids, was expressed as soluble, active, and specific by human embryonic kidney (HEK293) cells, it can successfully synthesize important compounds DSLNT in breast milk. In this study, a truncated sequence of hST6GalNAc V and hST6GalNAc VI without N-terminal transmembrane sequence was constructed for soluble expression in E. coli. And they showed the activity of synthesizing DSLNT. To the best of our knowledge, these two enzymes were successfully expressed in soluble and active form in bacteria for the first time.