Innovative use of a bacterial enzyme involved in sialic acid degradation to initiate sialic acid biosynthesis in glycoengineered insect cells

The terminal sialic acids on the carbohydrate side chains of glycoproteins have a major impact on their overall properties, including their circulatory half-lives in vivo (Ngantung et al., 2006). Therefore, recombinant glycoprotein biologics intended for human therapeutic applications are usually manufactured in mammalian cell systems, which have protein glycosylation pathways that can provide terminal sialylation (Walsh, 2010). However, one disadvantage of mammalian protein glycosylation pathways is that they sometimes produce recombinant glycoproteins with immunogenic carbohydrate side chains containing either the αGal epitope (Chung et al., 2008) or the relatively rare sialic acid, N-glycolylneuraminic acid (Ghaderi et al., 2010). In contrast, the baculovirus/insect cell system does not produce these immunogenic carbohydrate epitopes. The baculovirus/insect cell system is also generally considered to be safer than mammalian systems because it is less likely to harbor adventitious agents that could infect humans or other mammals (reviewed by Jarvis, 2009). Unfortunately, the unmodified baculovirus/insect cell platform cannot be used to manufacture recombinant glycoproteins for in vivo therapeutic applications because it does not provide terminal sialylation.

This limitation has been addressed at the basic research level by metabolic engineering, which has been used to extend the endogenous processing capabilities of the insect cell protein N-glycosylation pathway (reviewed by Geisler and Jarvis, 2009, Harrison and Jarvis, 2006, Jarvis, 2009, Shi and Jarvis, 2007). These glycoengineering efforts have yielded genetically modified baculovirus/insect cell systems capable of producing N-glycoproteins with complex, mammalian-type carbohydrate side chains (N-glycans) terminated with the most common sialic acid, N-acetylneuraminic acid (Neu5Ac; Aumiller et al., 2003, Aumiller et al., 2012, Hill et al., 2006, Hollister et al., 2002, Hollister and Jarvis, 2001, Jarvis et al., 2001, Seo et al., 2001). One important step taken to achieve this goal was to metabolically engineer insect cells to produce Neu5Ac and CMP-Neu5Ac, which are required for glycoprotein sialylation, but undetectable in non-engineered insect cell lines (Hooker et al., 1999, Tomiya et al., 2001). This was accomplished by introducing mammalian sialic acid synthase (SAS) and CMP-sialic acid synthetase (CMAS) genes into the baculovirus/insect cell system, which enabled the cells to utilize the key sialic acid precursor, N-acetylmannosamine (ManNAc), and produce Neu5Ac and CMP-Neu5Ac (Aumiller et al., 2003, Hill et al., 2006; Fig. 1). While this engineering strategy was successful, the resulting baculovirus/insect cell systems had to be cultured in growth media supplemented with ManNAc in order to efficiently produce Neu5Ac, CMP-Neu5Ac, and sialylated recombinant glycoproteins. This substantially increased the cost of growth media and revealed an acute need for additional glycoengineering to circumvent this expensive media supplementation requirement.

In mammalian cells, a bifunctional enzyme known as UDP-N-acetylglucosamine 2′-epimerase/N-acetylmannosamine kinase (GNE) initiates sialic acid production by converting UDP-N-acetylglucosamine (UDP-GlcNAc) to N-acetylmannosamine-6-phosphate (ManNAc-6-P; Hinderlich et al., 1997, Stäsche et al., 1997). Thus, the most obvious way to circumvent the ManNAc supplementation requirement was to introduce mammalian GNE. In fact, it had already been shown that insect cells infected with recombinant baculovirus vectors encoding human GNE and SAS could produce Neu5Ac in the absence of any media supplements (Viswanathan et al., 2003). However, in the context of our broader goal of glycoengineering insect cells for glycoprotein sialylation, we were concerned that GNE overexpression might deplete intracellular UDP-GlcNAc, thereby reducing one or more N-acetylglucosaminyltransferase activities and diminishing the overall efficiency of N-glycan processing. This concern was heightened by a report showing that the level of UDP-GlcNAc is a critical factor regulating N-acetylglucosaminyltransferase V activity in vivo (Sasai et al., 2002).

With this in mind, we conceived a more innovative way to circumvent the ManNAc supplementation requirement for recombinant glycoprotein sialylation in glycoengineered baculovirus/insect cell systems. Though it initially seemed counterintuitive, our approach focused on E. coli GlcNAc-6-P 2′-epimerase (GNPE), which normally converts ManNAc-6-P to GlcNAc-6-P in a bacterial sialic acid degradation pathway (Vimr et al., 2004). Although insect cells have no detectable ManNAc-6-P, they convert glucose to GlcNAc-6-P as a part of their basal metabolism. Realizing that the GNPE reaction might be reversible, we hypothesized that GNPE could drive the reverse of its normal reaction in insect cells, producing ManNAc-6-P from GlcNAc-6-P and initiating sialic acid biosynthesis in the absence of exogenous ManNAc.

In this study, we tested this hypothesis and demonstrated that a eukaryotic recombinant protein production platform can be glycoengineered with a bacterial gene, that a bacterial enzyme normally involved in sialic acid degradation can be innovatively used to initiate sialic acid biosynthesis, and that insect cells expressing this enzyme can produce sialylated glycoproteins in the absence of media supplementation, which will reduce the cost of recombinant glycoprotein sialylation in glycoengineered insect cell systems.

The transgenic insect cells described in this study were produced using various immediate early expression plasmids, which can be used to express foreign genes constitutively in uninfected insect cells under the transcriptional control of the baculovirus ie1 promoter and hr5 enhancer elements (Table 1; Jarvis and Finn, 1996; Jarvis et al., 1990). pIE1Hygro, pIE1GlcNAcTII, pIE1HRGalT, pIE1ST6, and pIE1-hCSAT are described in the references given in Table 1. pIE1MmSAS and pIE1MmCSAS are new

The most commonly used hosts for baculovirus-mediated recombinant glycoprotein production are established lepidopteran insect cell lines, such as IPLB-SF-21 (Sf21; Vaughn et al., 1977), Sf9 (Summers and Smith, 1987), expresSF+ (Protein Sciences Corporation), and BTI-Tn-5B1-4 (High Five™; Wickham et al., 1992). Each of these cell lines has metabolic pathways that support recombinant glycoprotein biosynthesis and processing. However, these insect cell pathways are simpler than the corresponding

Terminal sialic acids are required for the therapeutic efficacy of many recombinant glycoprotein biologics, such as erythropoietin and some antibodies (Ngantung et al., 2006). This requirement restricts the utility of the unmodified baculovirus/insect cell system as a glycoprotein biologics production platform, which does not support recombinant glycoprotein sialylation (reviewed by Geisler and Jarvis, 2009, Harrison and Jarvis, 2006, Jarvis, 2009, Marchal et al., 2001, Shi and Jarvis, 2007).

This work was supported by Award no. R01GM49734 from the National Institute of General Medical Sciences and Award no U54AI-065357 from the National Institute of Allergy and Infectious Diseases. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of General Medical Sciences, National Institute of Allergy and Infectious Diseases, or National Institutes of Health. We thank Dr. Ann Toth (Department of Molecular