Pathways of protein N-glycosylation modification were previously proposed in some microalgae, such as C. reinhardtii, Chlorella vulgaris, and P. tricornutum (4, 41, 42). The putative glycosyltransferases (GTs) and glycosidases (GSs) that participated in the ER and Golgi N-glycosylation pathway were identified from different microalgae (4, 43). The N-glycosylation pathways of protein in microalgae were classified into GnTI-dependent and GnTI-independent pathways. However, the complete and precise pathway in microalgae remain largely unknown. A few studies explored the pathways of the genes involved in protein N-glycosylation in microalgae (4). In C. reinhardtii, a study on the functions of GnTI, mannosidase 1A (Man1A), xylosyltransferases (XylT1A (XTA), XylT1B (XTB)) and fucosyltransferase (FucT) revealed that the N-glycosylation process followed the GnT I-independent pathway (44). The Man1A insertion mutant affected methylation of mannoses and the addition of terminal xylose while the absence of XylT1A resulted in shorter N-glycan structure compared to that of wild type (45). XTA was responsible for the core β-1,2-xylose modification and XTB the β-1,4-xylose modification on the linear branch of the N-glycan and partly the core β-1,2-xylose modification (46). Moreover, knocking down FucT affected Man1A-depedent trimming and fucose transfer on N-glycan, while the knockdown of both XylTs and FucT lead to the formation of N-glycans with strongly diminished core modifications (15). The functions of several genes were also studied in P. tricornutum. N-acetylglucosaminyltransferase I (PtGnTI) could restore the maturation of complex type N-glycans in the Chinese Hamster Ovary (CHO) Lec1 mutant lacking endogenic GnT I, thus demonstrating the functional activity of the diatom N-acetylglucosaminyltransferase (3). A putative GDP-L-fucose transporter (PtGFT) was able to rescue the fucosylation of proteins in the CHO-gmt5 mutant cell line, suggesting the potential transporter function of PtGFT in P. tricornutum (20). A preliminary study of PtFucT1 (54599) function via α-1,3-fucose antibody was also carried out in P. tricornutum, and it showed that the overexpression of FucT increased the N-glycans bearing α-1,3-fucose epitope (20). However, the physiological functions and the effects of PtFucT1 on N-glycoproteins and N-glycan structures were only comprehensively elucidated in this study.
Among 7 glycosyltransferase (GT) families, the GT37 family are certificated to be involved in the fucosylation of plant-specific organelles, such as cell wall, while GT10 family contained α-1,3/4-FucTs for N-glycans in plants (9). The three PtFucTs were clustered in GT10 family, indicating their putative functions as α-1,3/4-FucTs on N-glycans. Among the five conserved motifs of plant FucTs from GT10 family, three PtFucTs were located on the Motif IV and Motif V. Two of these potentially functional motifs were also found in invertebrates α-1,3-FucT via Pfam analysis (47). While the putative domain from the C-terminus of PtFucT2 and PtFucT3 were observed for plant α-1,3-FucT (48). Motif IV (i.e., α-1,3-FucT motif), and can interact with the donor substrate GDP-fucose for fucosyltransferase activity (20, 48). The CXXC motif (Motif V) was located at the C-terminus of the three PtFucTs, which is often involved in the formation of disulfide bonds in plant α-1,3-FucT (41).
The subcellular localization of a protein is important for its function in cell. Until now, GDP-fucose transporter (PtGFT) was located to the Golgi apparatus in the CHO-gmt5 mutant cell line, PtGnTI and PtFucT1 were verified to be medial/ trans-Golgi localization via transmission electron microscope coupled to immune-gold labeling in P. tricornutum (20). The medial/ trans-Golgi located PtFucT1 was further confirmed in our study by eGFP fusion expression and co-expression with Golgi marker PtXylT/PtVps29-mRFP via confocal laser scanning microscopy (49). Additionally, owing to the plastid stroma-like fluorescence (21), PtFucT2 and PtFucT3 lacking plastid targeting information were localized to the plastid stroma. However, it is still unknown how to explain the plastid stroma localization of PtFucT2 and PtFucT3. It was proposed that the plastid of the Chlamydomonas cells does not contain the N-glycosylation machinery (50), indicating that PtFucT2 and PtFucT3 might not participate in the N-glycosylation modification of protein. Therefore, it is interesting to study the function of Golgi-located PtFucT1 during the protein N-glycosylation pathway.
Subsequently, PtFucT1-OE mutant and the knockout mutant of PtFucT1 (PtFucT1-KO) lacking two main motifs (Motif IV and V) were analyzed in physiological and N-glycoproteomic levels. Tubulin proteins are an important constituent of microtubules that build the cytoskeleton (51) and regulate cell elongation and growth (52). Therefore, the significant decrease of cell density was probably due to the down-regulated N-glycosylation of tubulin alpha and beta chains in PtFucT1-OE mutant compared to the wild type cells. PsbA affected primary photochemistry, such as chlorophylls and the transfer of electron in Photosynthesis II (53). Gene rbcL participated in the CO2 fixation and regulated the plant biomass (54). Lhcr14 encoding the fucoxanthin chlorophyll a/c protein is a unique light-harvesting apparatus in diatom (55). Besides, ftsH is important for the growth of leaf and the development of chloroplast (56). All these relevant proteins regulated the photosynthesis efficiency in PtFucT1-OE mutant. Therefore, the reduced Fv/Fm in PtFucT1-OE mutant might be related with the up-regulated N-glycosylation of psaL and down-regulated N-glycosylation of the other nine photosynthesis relevant proteins (ftsH, Lhcr14, PetJ, PsbC, PsbO, rbcL, PsbA, Atp1 and Cytb6f). The increased soluble polysaccharide in PtFucT1-OE mutant was probably due to the down-regulation of the N-glycosylation of GPI_1, Cts and GapC2a. It was already known that these three key enzymes were important during the processes of glycolysis or tricarboxylic acid cycle as they regulated carbon metabolism (57). The DGPs involved in the cytoskeleton were not identified in PtFucT1-KO mutant. However, seven DGPs related with photosynthesis, carbon and nitrogen metabolisms were found in PtFucT1-KO mutant. Especially, glutamine synthetase is a key enzyme for the assimilation of nitrogen (24). The up- or down-regulated N-glycosylation of these proteins might change the proteins’ structure and subsequent function, leading to various physiological phenotypes in PtFucT1-KO mutant.
N-glycan structures are important for the protein folding, structure stability, and function (5). However, the N-glycans of the DGPs discussed in this study were not identified in the N-glycomics. Although previous studies showed that the overexpression of PtFucT1 increased the core fucose modified N-glycoproteins (20), the knockout of PtFucT1 inactivated the α-1,4-fucose modification but not the core α-1,3-fucose modification of N-glycans in the present study. This is consistent with the functional prediction, which showed that PtFucT1 is a non-core fucosyltransferase. The change of complex type N-glycans in wild type to mannose type in PtFucT1-KO mutant indicated that the knockout of PtFucT1 not only inhibited the functions of α-1,4-fucosyltransferase, but also the activity of PtGnTI. This was certificated by the interaction of PtFucT1 and PtGnTI. PtGnTI is responsible for the synthesis of complex N-glycans (3). The knock-out of PtFucT1 inhibited the activity of PtGnTI and then changed the N-glycan structure. The proposed working model of PtFucT1 was shown in Fig. 9.
Interestingly, the core α-1,3-fucose residues of N-glycans were both observed in wild type and PtFucT1-KO mutant, suggesting the existence of core fucosyltransferase in P. tricornutum. For example, Acetyl-CoA carboxylase (ACC, B7GEB5) harboring core α-1,3-fucose residue was observed in both the wild type and PtFucT1-KO mutant. ACC is a key enzyme responsible for the catalysis of carboxylation of acetyl-CoA into malonyl-CoA in plastids (58). However, it is still unknown which protein is responsible for the core fucose modification. The cause of the absence of fucose residue in 13 N-glycopeptides (7.6%) in the wild type while they have fucose residue in the PtFucT1-KO mutant is unknown. We speculate that the difference could be due to the remaining first domain (97-229AA) in PtFucT1-KO mutant or failure to knockout the first domain of PtFucT1 in P. tricornutum. In addition to the reported main mannose type N-glycans (5–9 mannose residues) in P. tricornutum (3, 59), large amount of complex N-glycans with GlcNAc, fucose and xylose residues were identified in this study. Diatoms may initially synthesized mannose type N-glycans, and then the complex type N-glycans (59), with the Glc2Man9GlcNAc2 structure being the lipid-linked oligosaccharide in P. tricornutum (60). Although the main N-glycan structure and lipid-linked oligosaccharide were reported, to our knowledge, this is the first time the total N-glycoproteins and their corresponding precise N-glycans were analyzed using N-glycoproteomic and N-glycomic levels in P. tricornutum. This provided critical data for further functional study of these N-glycoproteins and the mechanism of protein N-glycosylation modification.