All helminth species possess glycoproteins with mannosidic glycans. However, the abundance and diversity of the mannosidic glycans is species-dependent (Hokke and van Diepen, 2017). Mannosidic N-glycans can range from high-mannose type (Man7-9) to Man5 and paucimannosidic N-glycans (Man3). In addition, Man5 or Man3 N-glycans can carry a single branched GlcNAc residue with or without further branch extensions (hybrid or single-branched N-glycans, respectively). Several helminth species, like the trematode Fasciola hepatica and nematode Haemonchus contortus, contain even shorter N-glycans, where additional mannoses are trimmed from the Man3 structure resulting in Man2 glycan structures (Paschinger and Wilson, 2015; Garcia-Campos et al., 2016; Ravidà et al., 2016).

Various N-glycan core modifications have been observed in helminth species and in different developmental stages of a single species (Hokke and van Diepen, 2017). The main modification of the core is fucosylation. In early developmental stages of the trematode Schistosoma mansoni, N-glycans have an α1,3-fucose on their core. In later stages a core α1,6-fucose was found in the presence or absence of core α1,3-fucose. More densely fucosylated N-glycan cores were observed in adults of the nematode H. contortus. A trifucosylated core, with an α1,3-fucose attached to the distal core N-acetyl-glucosamine (GlcNAc) and two core fucoses (α1,3 and α1,6) attached to the proximal core GlcNAc, is the dominant glycan species on the H11 glycoprotein of H. contortus (Haslam et al., 1996). Additionally, galactosylation of the distal core α1,3-fucose and proximal core α1,6-fucose has been observed in H. contortus (Haslam et al., 1996; Hokke and van Diepen, 2017). N-glycan core xylosylation is uncommon in helminths, but has been observed in S. mansoni larval stages during the first three to 6 days after infection of humans, and in mature eggs and miracidia (Smit et al., 2015).

The degree of N-glycan branching is also highly variable between different helminth species and between different developmental stages (Smit et al., 2015; Hokke and van Diepen, 2017). Complex tri- or tetra-antennary N-glycans are often extended with β1,4-linked galactose or N-acetyl-galactosamine (GalNAc) as part of a LacNAc (LN) or LacdiNAc (LDN) motif (Figure 2), respectively (Haslam et al., 2000; Smit et al., 2015). In several helminth species these LN and LDN motifs are further substituted with α1,3-fucose linked to a GlcNAc (part of a Lewis X or LDN-F motif) and is frequently found in the antennae of N-glycans. Additional α1,2-fucoses can be attached to the α1,3-linked fucoses in the antennae to form difucosylated glycan structures, as found on the N-glycans of SmVAL-9 (Yoshino et al., 2014). Another terminal glycan structure, but less common, is a fucosylated terminal poly GlcNAc structure, which has been described for Dirofilaria immitis (Martini et al., 2019). In contrast to human N-glycans extension of antennary galactose with sialic acid is never observed in helminths.

Within the helminth glycome there are several unusual modifications found on N-glycans. The non-carbohydrate moiety phosphorylcholine (PC) was attached to antennary GlcNAc residues or distal core GlcNAc of N-glycans in several helminth species (Pöltl et al., 2007; Paschinger et al., 2012). PC can also be attached to GalNAc combined in a LDN(-F) motif, although this is less common (Wilson and Paschinger, 2016; Martini et al., 2019). PC is found in parasitic nematodes and cestodes, as well as in free-living nematodes such as Caenorhabditis elegans (Lochnit et al., 2000; Hokke and van Diepen, 2017). LDN motifs can also be extended with the unusual sugar moiety glucuronic acid that has been found on N-glycans of Dirofilaria immitis (Martini et al., 2019). Glucuronic acid is even found on antennae that are already modified with PC (Martini et al., 2019). Another rare sugar is tyvelose, which is found on N-glycans in Trichinella spiralis as an extension of LDN-F (Reason et al., 1994). Besides these uncommon moieties as part of helminth N-glycans other post-glycosylation modifications such as methylation, sulfation or phosphorylation were found on different helminth glycans (Ravidà et al., 2016; Jiménez-Castells et al., 2017).

Helminth glycoproteins can also harbor O-glycans and most of them are initiated by the addition of GalNAc (O-GalNAc glycans). O-glycosylation in helminths shows resemblance with mammalian O-GalNAc glycosylation but lacks variants that are capped with sialic acid. The structures encountered are similar to mammalian core 1 and core 2 O-glycan structures (Figure 2; Casaravilla et al., 2003; Hokke and van Diepen, 2017). Both cores can be elongated with a wide variety of glycan motifs, but the observed extensions can differ between core 1 and core 2 O-glycans (Hokke and van Diepen, 2017). Core 1 O-glycans can be elongated with poly-galactose (mainly α1-4 linked), Lewis X or a single fucose moiety as found in Echinococcus multilocularis, S. mansoni and Toxocara canis, respectively. Core 2 O-glycans can be elongated with terminal poly-galactose, Lewis X, F-LDN or LDN-F. Poly-fucosylated versions of LDN with α1,2-linked fucose (di-fucosylated and/or tri-fucosylated motifs) are found in S. mansoni (Smit et al., 2015). The possible combinations of core structures, elongations and terminal structures result in an array of different O-glycans (Talabnin et al., 2013; Smit et al., 2015; Hewitson et al., 2016). For several helminth species unique O-GalNac cores have also been observed. For S. mansoni, a core 1 O-glycan with additional β1,6-linked galactose and other extensions have been observed in several life stages (Smit et al., 2015). H. contortus has a rare O-GalNAc glycan, which is substitute with α1,3-linked galactose (core 8) (van Stijn et al., 2010). Not all helminth O-glycans are GalNAc-based glycans. O-glycans have been detected in S. mansoni and Heligmosomoides polygyrus that were initiated with an uncharacterized hexose (Jang-Lee et al., 2007; Hewitson et al., 2016). These rare O-glycans were further extended with fucose or another hexose residue. Helminth O-glycans can also be modified with post-glycosylation modifications such as methylation. Methylated O-glycans have been found on common core 1 and core 2 O-glycans of helminths (Hokke and van Diepen, 2017). Although the knowledge on the helminth O-glycome is relatively limited compared to the N-glycome, it is receiving more attention, possibly leading to the discovery of structures and post-glycosylation modifications previously unknown to the helminth O-glycome.

To mimic the high-mannose type N-glycans (Man7-9) as found on many helminth glycoproteins, a strategy is needed that prevents maturation of these high-mannose N-glycans. The best known method to achieve this, is the addition of an ER retention signal, such as the C-terminal H/KDEL peptide sequence (Napier et al., 1992; Schouten et al., 1996). Retention of the glycoprotein in the late ER prevents the glycan from being modified by Golgi mannosidases. Alternatively, kifunensine (KIF), an α-mannosidase I inhibitor, has been used as a hydroponic treatment to increase the fraction of Man9 structures in N. benthamiana leaves (Krahn et al., 2017; Roychowdhury et al., 2018). Another inhibitor, swainsonine, has been used as additive in transgenic rice cultures to generate Man5 N-glycans by inhibition of α-mannosidase II (Choi et al., 2018). Man5 structures can also be created through the use of a N-acetyl-glucosaminyltransferase (GnT) I RNAi line of N. benthamiana, but this strategy also seemed to result in shorter Man3 and Man4 structures (Uthailak et al., 2021). Engineering toward specific mannosidic glycans remains challenging, as the necessary glyco-engineering strategies have not yet been developed or require additional optimization.

Native plant N-glycans carry core a α1,3-fucose and β1,2-xylose, which are not always present on helminth N-glycans. Therefore, the inhibition of plant core α1,3-fucosyl- and β1,2-xylosyltransferases to prevent their addition to N-glycans is a prevalent achievement. Targeted interruption of core FucTs and XylTs has also been accomplished through RNA interference (RNAi) in a variety of plant species (Cox et al., 2006; Sourrouille et al., 2008; Strasser et al., 2008; Shin et al., 2011). For N. benthamiana this resulted in the ΔXT/FT RNAi line, which lacked plant-native N-glycan core modifications on the majority of the produced N-glycans (Strasser et al., 2008). The inability of RNAi to completely inactivate core FucT and/or XylT genes was later overcome through CRISPR/Cas9 technology in Nicotiana benthamiana (Jansing et al., 2019). Specific helminths, such as S. mansoni, do synthesize N-glycans that carry core α1,3-fucose with or without β1,2-xylose. The generation of individual knock-out plant lines for either core FucT or XylT genes by Jansing et al. (2019) may prove useful to obtain N-glycans that lack either α1,3-fucose or β1,2-xylose. Taken together, plant lines obtained through the different approaches may all be suitable for glyco-engineering of helminth N-glycans.

To synthesize core α1,6-fucosylated N-glycans in plants a core α1,6-fucosyltransferase (FucT) should be introduced, such as FucT8 from Drosophila melanogaster, Homo sapiens or Mus musculus (DmFucT8, HsFucT8, or MmFucT8, respectively) (Castilho et al., 2011a; Wilbers et al., 2016). In some cases, core α1,3-fucosylation is also desired. When using plants deficient in FucT and XylT activity, reintroduction of core α1,3-fucose is required, which can be achieved by co-expression of an α1,3-FucT from S. mansoni (SmFucTC) or a hybrid Zea mays core α1,3-FucT with the CTS domain of A. thaliana FucT11 (Castilho et al., 2015; van Noort et al., 2020). An aspect that needs attention is glycosyltransferase inhibition that may occur upon simultaneous introduction of multiple FucTs. A strict order of fucosylation is observed for synthesis of double fucosylated N-glycan proximal cores of for instance S. mansoni, with α1,6-fucosylation preceding α1,3-fucosylation (Paschinger et al., 2005, 2019). Therefore, aberrant Golgi localization of transiently introduced (hybrid) FucTs should be carefully monitored to avoid disruption of the activity of introduced FucTs (van Noort et al., 2020).

In plants, the only galactosylated native N-glycan structure is a Lewis A structure. Plants have shown to display the Lewis A glycan motifs on endogenous glycoproteins (Fitchette-Lainé et al., 1997; Strasser et al., 2007b) and in some cases on recombinant glycoproteins (Weise et al., 2007; Castilho et al., 2011b; Wilbers et al., 2016). The presence of Lewis A shows that plants have the machinery to add β1,3-galactose to glycans. Interestingly, β1,3-galactose is always seen in combination with α1,4-fucose in the Lewis A motif. The presence of plant Lewis A epitopes on recombinant glycoproteins can be completely abolished by the disruption of β1,3-galactosyltransferase 1 as has been demonstrated in moss (Parsons et al., 2012).

In contrast to any β1,3-galactose-extended glycans, helminth glycoproteins carry N-glycans that are extended with antennary β1,4-linked galactose (LN). The synthesis of LN-carrying N-glycans has been achieved in N. tabacum upon co-expression of a hybrid β1,4-galactosyltransferase 1 (sialGalT) of Danio rerio or H. sapiens (Anusiem and Beetlestone, 1976; Hesselink et al., 2014). The original CTS domain of GalT was replaced with the CTS domain of Rattus norvegicus α2,6-sialyltransferase (sial). The sial-CTS domain translocates GalT to the trans-Golgi, as earlier addition of galactose can significantly hinder the activity of endogenous N-acetyl-glucosaminyltransferase II (GnTII), mannosidase II (α-Man II) and xylosyltransferase (XylT), leading to the synthesis of hybrid N-glycans (Johnson and Chrispeels, 1987; Bakker et al., 2001; Wilbers et al., 2017). In addition to localization, promotor strength can be exploited to avoid interference with the activity of certain glycosyltransferases to achieve the synthesis of a specific glycan (Wilbers et al., 2017; Kallolimath et al., 2018). For example, strong expression of sialGalT under control of a 35S promoter allows for the synthesis of hybrid N-glycans, whereas sialDrGalT expression under the control of the weaker Gpa2 promoter allows synthesis of single branched non-hybrid N-glycans (Figure 3).

Galactose-extended N-glycans are sometimes fucosylated (Lewis X), as is seen for N-glycans on omega-1 and IPSE/alpha-1 from S. mansoni (Wuhrer et al., 2006; Meevissen et al., 2010). Synthesis of Lewis X was achieved upon co-expression of sialDrGalT and a hybrid α1,3-FucT9a (sialFucT9a) from Tetraodon nigriviridus (Rouwendal et al., 2009; Wilbers et al., 2017). More recently, the fucosyltransferases FucTD and FucTE from the parasite S. mansoni were identified as enzymes able to synthesize Lewis X, where SmFucTD and SmFucTE were most efficient (van Noort et al., 2020). The presence of predominantly monoantennary Lewis X-carrying N-glycans upon engineering suggests that introduction of a sialGalT still interferes with endogenous GnTII activity (Wilbers et al., 2017; van Noort et al., 2020).

While the synthesis of Lewis X is efficient, the synthesis of LN is not always efficient and depends on the protein under investigation (Strasser et al., 2009; Kriechbaum et al., 2020). For this reason, it was postulated that endogenous plant β-galactosidases (BGALs) along the secretory pathway could interfere with the synthesis of galactose-extended N-glycans. Recently, BGAL1 from N. benthamiana (homologous to BGAL8 from A. thaliana) has been characterized as an apoplast localized β-galactosidase and the major contributor to degalactosylation of both N- and O-linked glycans (including endogenous Lewis A). Impairment of NbBGAL1, through RNAi and CRISPR, resulted in a significant increase in galactosylation on various glycoproteins (Kriechbaum et al., 2020). Due to the redundancy of substrate specificity of the BGAL enzyme family and the number of β-galactosidases detected in apoplast fluid, it is likely that residual β-galactosidase activity may still lower the level of galactosylation (Dean et al., 2007; Buscaill et al., 2019). Therefore, additional research into remaining β-galactosidases and their ability to cleave N- and O-glycans is required to optimize glyco-engineering strategies for galactose containing N-glycans.

Helminth N-glycans often contain GalNAc as is seen in LDN and fucosylated LDN (Figure 2). Engineering of LDN carrying N-glycans requires the addition of a GalNAc on a terminal GlcNAc residue. Recently, LDN was synthesized on the N-glycans of S. mansoni kappa-5 in N. benthamiana by introduction of a N-acetyl-galactosaminyltransferase from C. elegans (CeGalNAcT) (Wilbers et al., 2017). GalNAc is never observed in plant glycans, and it was initially expected that incorporation of this sugar into N-glycans would require synthesis of UDP-GalNAc as substrate and facilitate its transport into the Golgi. However, it was observed that co-expression of a C4 epimerase to enhance substrate availability as well as co-expression of UDP-GalNAc transporters for improved Golgi localization was not required (Wilbers et al., 2017). This shows that, even though GalNAc is not native to plant N-glycans, UDP-GalNAc is present within the correct Golgi compartment.

The LDN motif can be further modified by expression of a α1,3-fucosyltransferase (α1,3-FucT; e.g., sialTnFucT9a, SmFucTD or SmFucTE) to yield a LDN-F motif (Wilbers et al., 2017; van Noort et al., 2020). Moreover, the glycan motif F-LDN-F could be synthesized upon co-expression of these enzymes with SmFucTF. The possibility to synthesize fucosylated LDN with SmFucTD, SmFucTE and SmFucTF is a promising finding for the production of (fucosylated) LDN carrying helminth glycoproteins, such as kappa-5. From S. mansoni.

Similar to the synthesis of galactose-extended N-glycans, it seems that fucosylation of LDN, to obtain LDN-F, enhances the extension of N-glycans with GalNAc (Wilbers et al., 2017). Plant-native β-hexosaminidases (HEXOs) activity may limit the generation of LDN glycan motifs on kappa-5 in GalNAcT-expressing Nicotiana benthamiana (Alvisi et al., 2021). HEXOs are hydrolytic enzymes that are able to remove terminal GlcNAc and GalNAc residues from complex N-glycans, thereby generating typical plant paucimannosidic N-glycans (Shin et al., 2017; Alvisi et al., 2021). Plants harbor several HEXO homologs (HEXO1, HEXO2 and HEXO3) that differ by their substrate specificity and sub-cellular localization. HEXO1 and HEXO3 of A. thaliana and N. benthamiana are involved in the processing of plant N-glycans to paucimannosidic N-glycans, where HEXO1 is active in the vacuole and HEXO3 in the apoplast (Strasser et al., 2007a; Shin et al., 2017). N-glycans with non-endogenous GalNAc extended N-glycans can be processed by HEXO3 in the apoplast of N. benthamiana. In addition, HEXO2 was characterized as a membrane-bound enzyme specifically cleaving GalNAc residues from N-glycans (Alvisi et al., 2021). Unfortunately, targeted transient RNAi was not sufficient to improve LDN synthesis in plants (Alvisi et al., 2021). Future efforts should now focus on the generation of a NbHEXO knock-out plants to accelerate helminthisation of N-glycans in a plant expression system.

In contrast to plants, various helminth glycoproteins carry additional branches on their N-glycan structures. This was observed on tri- and tetra-antennary LDN(-F) carrying N-glycans of S. mansoni kappa-5, the tetra-antennary PC-glycans on ES-62 from Acanthocheilonema viteae and the tri- and tetra-antennary N-glycans with tyvelose of T. spiralis (Reason et al., 1994; Harnett et al., 2005; Meevissen et al., 2011). Plants can only synthesize up to two branches on their N-glycans, as plants lack N-acetyl-glucosaminyltransferases (GnT) IV and V (Sheshukova et al., 2016). Currently, the synthesis of N-glycans with LN, Lewis X, LDN and (F-)LDN(-F) is predominantly mono-antennary (Wilbers et al., 2017; van Noort et al., 2020). Co-expression of exogenous GnTII from A. thaliana (AtGnTII) or H. sapiens (HsGnTII) could increase the occurrence of di-antennary N-glycans structures (Schneider et al., 2014; Dicker et al., 2016). In order to synthesize tri- or tetra-antennary N-glycans additional exogenous transferases are required, such as GnTIV and GnTV. Castilho and colleagues demonstrated that co-expression of GnTIV and GnTV from H. sapiens allows synthesis of tetra-antennary N-glycans (Castilho et al., 2011b). This strategy may also be used to produce helminth glycoproteins with tri- and tetra-antennary carrying N-glycans. A potential challenge for increasing the number of branches, is the possible activity of β-hexosaminidases on the initial GlcNAc, and possibly GalNAc extension, of each branch.

Helminth glycoproteins also have been shown to carry complex O-glycans (Hokke and van Diepen, 2017). The synthesis of O-glycans on helminth glycoproteins produced in plants is a next challenge that will allow the study of the role of these proteins in parasitism and their effectiveness as biopharmaceuticals. Just as with N-glycans, glyco-engineering of helminth O-glycans shows resemblance with humanization of the plant production platform. Most helminth O-glycans are initiated with O-GalNAc. O-GalNAc glycans are absent in plants, which allows for a bottom-up approach for the synthesis of such glycans. Two simultaneous, but independent studies have shown that O-GalNAc glycans can be synthesized, either by expression of GalNAcT2 and C1GalT1 leading to mucin core 1 O-GalNAc glycans (Castilho et al., 2012) or the combination of GalNAcT2, GalNAcT4and a GlcNAc C4-epimerase leading to O-GalNAc glycans (Yang et al., 2012). The synthesis of O-GalNAc glycans was possible despite hypothesized interference of prolyl-4-hydroxylases (P4H) and subsequent plant-native O-glycosylation in N. benthamiana (Montero-Morales and Steinkellner, 2018). Plant native O-glycans could pose an issue in terms of immunogenicity when using plant-based pharmaceuticals (Yates et al., 1996; Leonard et al., 2005). In addition, plant native O-glycans could sterically hinder the synthesis of helminth O-glycans. A P4H1 knock-out in Physcomitrella patens prevents proline conversion into hydroxyproline (Hyp) on recombinant EPO (Parsons et al., 2013). In N. benthamiana, P4H1, P4H4, P4H9 and P4H10 have recently been identified as enzymes capable of forming hydroxyproline on recombinant proteins (Mócsai et al., 2021). Unfortunately, transient RNAi of these genes in N. benthamiana was insufficient to reduce P4H activity completely, due to redundancy within the substrate specificity of the P4H genes. However, it remains unknown to what extent plant-native glycosylation interferes with the substrate specificity of O-GalNAc glycans. It is anticipated that the effect of galactosidases and possibly hexosaminidases could be a bigger issue for the synthesis of mucin-type O-GalNAc glycans. BGAL1 from N. benthamiana has been shown to remove the terminal galactose from core 1 O-GalNAc glycans (Kriechbaum et al., 2020), whereas β-hexosaminidases can cleave off HexNAc residues in an elongated O-GalNAc glycan (Cheng et al., 2013). Therefore, more research is required into enzymes involved in O-glycan synthesis and hydrolysis, as this might be required for efficient production of native helminth O-glycosylated proteins.