An in vitro one-pot synthetic biology approach to simulating diverging Golgi O-glycosylation of tumor-associated MUC1 from normal tissue MUC1

An in vitro one-pot synthetic biology approach to simulating diverging Golgi O-glycosylation of tumor-associated MUC1 from normal tissue MUC1 | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article An in vitro one-pot synthetic biology approach to simulating diverging Golgi O-glycosylation of tumor-associated MUC1 from normal tissue MUC1 Kevin Naidoo, Abdullateef Nashed, Kyllen Dilsook, Tharindu Senapathi This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5783651/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 22 Apr, 2026 Read the published version in Nature Communications → Version 1 posted You are reading this latest preprint version Abstract Peptide O-glycosylation relies on the coordinated action of glycosyltransferases (GTs) within the endoplasmic reticulum (ER) and Golgi apparatus. An in vitro one-pot synthetic biology approach was developed to investigate the specificity and kinetics of GT O-GalNAc glycosylation that leads to tumor antigen glycoforms of mucin 1 (MUC1). The focus is to experimentally simulate the divergent glycosylation pathways that lead to the synthesis of cancer-associated antigens (Tn, T) and their sialylated derivatives. First, the biosynthetic details of the defining first step of GALNT relocation from the ER to the Golgi was modeled using the one-pot method. Our findings reveal that an ER enriched with GALNTs results in complete Galnac (Tn) MUC1 site occupancy. This comes about as a function of two processes that are i) extended GALNT reaction time and ii) prevention of inhibition by subsequent glycosylation enzymes such as C1GALT1. The modeling confirms that B3GNT6 has negligible specificity for MUC1 Tn, explaining the absence of core 3 and core 4 structures in MUC1 in both normal and cancerous breast cell lines. Moreover, ST6GALNAC1, and not ST6GALNAC2, is primarily responsible for α-2-6 sialylation of Tn and T antigens. Computer reaction dynamic simulations combined with kinetic experimental analysis show that ST6GALNAC1 prefers fully glycosylated MUC1 and more importantly that its preference is to sialyate the S9 and T13 sites in the SAPDTR motif. This is especially the case when the MUC1 concentration is high (i.e., high-level of expression), suggesting that sTn upregulation on MUC1 in cancer is linked to the occupancy status of S9 and T13 glycosylated sites, that were previously found to be cancer-associated. The results from the one-pot synthesis approach presented here demonstrate its ability to simulate cellular glycosylation within the Golgi-ER. This systems modelling unpacks the molecular details of enzyme localization and substrate glycan occupancy that is fundamental to the regulatory mechanisms that gives rise to tumor-associated MUC1 antigens. Biological sciences/Chemical biology/Glycobiology Biological sciences/Systems biology/Synthetic biology In vitro synthetic biology Cancer MUC1 antigens systems chemical glycobiology reaction dynamics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1. Introduction The post-translational event of peptide or protein glycosylation is a non-template-driven process that relies on more than just glycoenzyme gene expression data or even the glycoenzyme expression levels themselves. 1 A case in point is that while glycosyltransferase (GT) gene expression can be used to classify cancer, 2 this genomic level data cannot be used to directly infer the difference between cancerous and healthy glycoconjugate expression. Specifically, the characteristically high degree of sialylation observed in tumor tissues 3 and the associated structural modifications of glycans cannot be directly correlated with the genes that express the sialyltransferases. This is because the complex glycosylation pathways within cells are intimately connected and intertwined with their critical metabolic and regulatory networks. 4 This complexity dissociates the high degree of sialylation observed in tumor tissues 3 and structural modifications of glycans directly from gene and protein expression. Consequently, a critical first step in systems biology modelling is the construction of a developmental model to mimic sequential biosynthesis processes within the ER and Golgi apparatus. Site specific chemoenzymatic synthesis is the standard method for producing model glycopeptides, where the initial sugar peptide bond is chemically synthesised. An enzymatic synthesis of the glycan ensues one sugar residue at a time. 5 Advances in understanding enzyme specificities and mechanisms have enabled the synthesis of more complex glycopeptides, 5 expanding the dimensions of glycopeptide arrays. In vitro methods that produce single, purifiable, and spectroscopically verifiable structures are more suitable than synthetic biology methods since the intention is to measure the kinetics of GTs as well as map out their selectivity and mechanistic action. The localization of GTs, such as GALNTs, has been found to be a regulatory mechanism involved in cancer phenotypes by altering O-GalNAc glycan structures and levels. 6 , 7 Consequently, in vitro methods must have the capability to mimic the alterations of in vivo glycosylation resulting from the spatial-temporal rearrangement of the distribution and presentation of GTs to the substrate in the ER-Golgi system. 8 , 9 The first step in mucin-type O-linked glycosylation is the addition of GalNAc to serine or threonine residues facilitated by several N-acetylgalactosaminyltransferases (GALNTs), forming the Thomsen-nouvelle (Tn) antigen (Fig. 1A). Following this the addition of galactose to the Tn antigen through T synthase (C1GALT1) is modified to form the T antigen (core 1). Alternative to this, the core 3 can be made by the addition of GlcNAc via b-1,3-N-acetylglucosaminyltransferase 6 (B3GNT6). These foundational structures undergo further branching and elongation with successive additions of monosaccharides such as GlcNAc and galactose, generating diverse glycan chains. Sialyltransferases mediate the sialylation of Tn and T antigens (left pathways in Fig. 1A). These sialylated forms (sTn and sT) terminates chain progression. Clinically, Tn, T, sTn, and sT antigens are significant through their role in establishing the hallmarks of cancer such as tumor progression, immune evasion, and metastasis. 10 , 11 The glycosylation of proteins takes place in the ER-Golgi system mostly through glycosyltransferases (GTs) and in some instances in combination with glycosidases. These glycoenzymes are distributed across specific cisternae. 8 , 9 , 12 The localization of GTs along the ER-Golgi axis is dynamic and they are constantly shuffled in both directions via a complex tightly regulated system involving COP-I and COP-II vesicles. 12 , 13 This localization across various cisternae has led to an assembly line of compartments performing sequential glycosylation to build glycans on target proteins. In the case of an organism disease state a protein’s glycan is often altered when there is deregulation of this localization, such as the relocation of GALNTs from the cis Golgi to the ER in tumour formation. 6 , 7 (Fig. 1, B). Here we use Mucin 1 (MUC1), as a peptide glycosylation prototype systems model. The aim is to resolve reasoning underlying differences in enzymatic construction in normal glycosylated MUC1 compared with tumor-associated (TA) MUC1. An in vitro method is developed to simulate ER-Golgi conditions for O-GalNAc glycosylation of the MUC1 and TA-MUC1 peptides. Specifically, we illustrate: i) the kinetic parameters governing glycosyltransferase (GT) activities and substrate specificities using the UGC assay 14 , ii) the molecular mechanisms governing the GTs site specificities, and iii) the effect of their expression and distribution along the ER-Golgi axis. Following this, the experimental model along with advanced computer reaction dynamics simulations, were used to reveal peptide site specificity of each of the GTs involved in the synthesis of Tn, T, sTn, and sT antigens at the five unique MUC1 glycosylation sites. 2. Results 2.1 The Sequential One-pot synthesis method: To capture the spatial-temporal segregation of glycosylation pathways a fusion tag protein carrying the MUC1 peptide was designed as an assembly conveyor (Fig. 2A). The conveyor vehicle (fusion protein and tags) was tested for biosynthesis interference (Fig. 2B). In the assembly line design for a glycan biosynthesis, the kinetics of the GT-catalyzed reactions and the associated intermediate glycan products are analyzed at every point of construction along the assembly line (Fig. 2C). The data obtained at each point informs subsequent steps, supporting model construction and iterative optimization of the synthesis. A fusion protein containing the core MUC1 peptide and a carrier protein, superfolder green fluorescent protein (sfGFP), expressed in E. coli. sfGFP was selected for its folding efficiency, minimized dimerization, and enhanced solubility. 15 To minimize the possibility of interaction with the MUC1 peptide, sfGFP was separated from the peptide using a linker (Linker 2) comprising of three rigid and three flexible units (from the N to C direction). Additionally, a rigid linker (Linker 1) was incorporated to improve steric presentation to the N-terminus His-tag, enhancing affinity-based purification. Central to the design are the work functions of the linkers. Firstly, the rigid region on Linker 2 must maximize the peptide sfGFP distance. Secondly, the rigid linker 1 must maximize the presentation of the His-tag for later TEV protease cleavage when salvaging the glycosylated MUC1. Here the predicted low conformational and structural predicted confidence around the TEV protease cleavage region signifies flexibility and so the designed accessibility of the protease. The GalNAc acceptor functions of the 27-mer in its fusion form and cleaved form (plus tag) were tested for each of the GALNT enzymes: GALNT1, GALNT2, GALNT4, and GALNT7 (Fig. 2B). The tag proved not to interfere with the enzyme activity or substrate specificity since the glycosylation rates were identical for the tagged and untagged MUC1 for the GTs. The fusion protein glycosylation carrier function was therefore optimized while preserving the inherent glycan recipient functions of the target peptide (MUC). 2.1.2 In vitro healthy vs. tumor GT distribution and concentration models The GalNAc-glycosylated sites are subject to either sialylation (addition of Sia) via ST6GALNAC1, galactosylation (addition of Gal) via C1GALT1 and its chaperone C1GALT1C1 (Cosmc), or N-acetylglucosaminylation (addition of GlcNAc) via B3GNT6 (Fig. 1A). In healthy contexts, GALNTs and C1GALT1 are localized in the cis-Golgi, while the ST6GALNAC1 is distributed across all the cisternae of the Golgi (Fig. 1, B). 8 , 9 , 12 No specific localization of B3GNT6 has been reported. The localization of only the GALNTs were reported to be altered in response to the EGF stimulation of SRC (the proto-oncogene in cancer) via COP-1 mediated retrograde from cis Golgi to the ER. 16 This relocation results in the overexpression of Tn in the ER where a fraction of this Tn transits to the cell surface without modification, while another fraction transits with modification to T antigen. In patient samples, the same study found that the mean expression of Tn in breast cancer tissue samples was 4.5 folds higher than in normal tissues. From the samples with high Tn expression, 70% of the samples showed ER localization of GALNTs (inferred indirectly from ER localization of Tn), whereas no significant loss of C1GALT1 was detected in these samples, pointing to the ER localization as the driving factor of the observed Tn overexpression. 6 , 7 Accompanying the relocation of GALNTs, the expression of O-GalNAc glycosylation enzymes is altered in cancer compared to normal cells. Furthermore, ST6GALNAC1 was reported to be upregulated in almost all cancer types 17 and C1GALT1 downregulated, mainly due to Cosmc mutation or epigenetic alteration of both C1GALT1 and Cosmc. 18 , 19 The objective here is to build a one-pot in vitro synthesis model that will be representative of the impact of the redistribution of GALNTs in altering glycan structures independently of enzyme levels (Fig. 1B and Fig. 2C). To achieve this, all activities and kinetics experiments are performed at a standardized enzyme concentration of 250 nM. The performance of the 23-mer and 27-mer peptides was compared by measuring the relative reactivity of GALNT1, GALNT2, GALNT4, and GALNT7 individually and in various combinations using the UGC assay 14 (Fig. 3A). The primary observation from both experiments is that GALNT4 alone does not react with either peptide, confirming it relies on a strictly lectin-dependent mechanism. GALNT7 shows no reactivity with either peptide in any combination. On the other hand, GALNT1 and GALNT2 could individually react with both peptides due to their direct catalytic mechanism. The rates of glycosylation via GALNT1 and GALNT2 were generally lower for the 23-mer than the 27-mer. This is explained by the depletion of the direct glycosylation-specific sites T13 and T20 for GALNT1 and GALNT2 respectively, in the 23-mer case, in addition to the absence of secondary lectin-assisted sites in the preferred direction. In contrast, the 27-mer permits the continuation of glycosylation via the lectin-assisted sites (Fig. 3B). When GALNT4 was tested in combination with GALNT1 or GALNT2, using the 27-mer construct, it showed no activity. However, GALNT4 exhibited significant activity using the 23mer peptide when combined with GALNT2, as indicated by the enhanced glycosylation rate when compared with the reaction of GALNT2 individually. GALNT2 is known to have a preferred specificity to T20 (in the PGST sequence) via direct catalytic domain recognition. In the 23-mer, T20 is at the N-terminal to T13 and S9 (GALNT4 specific sites), and pre-glycosylation at T20 left to T13 and S9 sites is the prerequisite for GALNT4 lectin-dependent specificity (Fig. 3B). 20 The collective evidence presented here confirms the efficiency of the fusion protein design to mimic natural MUC1 while its tag components do not affect the catalytic function of the GALNTs. Additionally, 23-mer has proved valid to capture the reaction mechanism of GALNT4 and the direct mechanism for both GALNT1 and GALNT2. However, the reported lectin-dependent mechanisms for GALNT1 and GALNT2 can be better studied using the 27mer. 2.2. In vitro synthesis of the Tn antigen In vitro synthesis of the Tn antigen was carried out to explore the possible structures enabled by extensive glycosylation (higher enzyme concentrations for a longer period) using varying combinations of GALNTs (Fig. 3C I). LC-MS results confirmed the structure and purity of the peptide (i). The catalytic action of GALNT1 results in glycosylation of 2 or 3 sites, and 3 sites in the case of GALNT2. The glycosylation of T8 and T20 is consistent with a direct GALNT1 and GALNT2 mechanism respectively. However, the observed glycosylation of the remaining two sites cannot be attributed to the lectin-dependent mechanisms as previously reported for the GALNT1 and GALNT2 activity on MUC1. 20 , 21 This is evident from short-term recorded activity on the remaining sites (Fig. 3, A) since the 23mer does not provide the directionality required for this mechanism. A randomized peptide sequences study found that GALNT1 can catalyze a long-range N terminal lectin-dependent mechanism while GALNT 2 can participate in both N and C long-range lectin-dependent mechanism. 22 Another study confirmed the bi-directionality of the lectin-dependent mechanism for both enzymes with preference given to N and C long range lectin-dependent mechanisms for GALNT1 and GALNT2, respectively 23 . No further glycosylation beyond three sites was observed when GALNT1 and GALNT2 were combined, suggesting the absence of synergy between the two enzymes and that the two remaining sites do not conform to the specificity of these GTs. When the product resulting from GALNT1 and GALNT2 activity was glycosylated with GALNT4, all five sites were glycosylated. These results indicate that the remaining two sites are GALNT4 specific and are glycosylated via a lectin-dependent mechanism. The consensus from all previously reported results is that T8 and T20 are specific sites for GALNT1 and GALNT2 respectively, and S9 and T13 are GALNT4 specific. Taken together, it can be concluded that GALNT1, GALNT2, or their combination can glycosylate S19, T20, and T8 by utilizing direct and lectin-dependent mechanisms. However, GALNT1 is less efficient than GALNT2 in completing the glycosylation of either S19 or T20 (Fig. 3C III). It is widely accepted that the GalNAc site occupancy increases with GALNTs over expression and their ER relocation. 6 , 7 The effect of these two factors was simulated here in the One-pot biosynthesis model where extensive glycosylation was performed to interrogate the action of GALNTs in isolation of other GTs. It was seen that sites, such as T20, S19, and T8 were not selective and can be glycosylated by multiple GALNTs, such as GALNT1 and GALNT2 as shown here, as well as GALNT3 that was previously reported. 20 , 21 We now refer to these sites that are the first ones to be glycosylated as GALNT_D-sites. The S9 and T13 sites are strictly lectin-dependent and thus are dependent on prior glycosylation of T20, S19, and T8 which we refer to now as GALNT_L-sites (Fig. 4A). The lectin-dependent mechanism is therefore responsible for high-density GalNAc-O-glycosylation. 24 This differential site occupancy was also found to be associated with cancer transformation. The site saturation was observed in MUC1 expressed in tumor cells compared with normal MUC1 in breast milk. 25 The GalNAc3-23mer (3 glycosylated sites at GALNT_D-sites only) and GalNAc5-23mer (5 glycosylated sites at GALNT_D-sites and GALNT_L-sites), synthesized here will be used as models for site occupancy, semi-glycosylated and completely-glycosylated, respectively. 2.3 In vitro synthesis of the T and sTn antigen and core 3 The sequential addition of one or two sugars to the Tn antigen generates diverse glycan core structures, with eight different cores have been identified. The most common structures are cores 1–4 (Fig. 1A) while cores 5–8 are rare. 26 Core 1, also known as T antigen, is formed by adding galactose to the Tn antigen via a β1–3 bond, a process catalyzed by T synthase (C1GALT1) with the help of its chaperone C1GALT1C1 (Cosmc). Within this context the core 3 is synthesized by B3GNT6 catalyses of the GlcNAc covalent β1–3 bond of to the Tn antigen. The ST6GALNAc1 sialylates Tn antigen to sialyl-Tn (sTn) that terminates the synthesis preventing the structures from undergoing further glycosylation via C1GALT1 or B3GNT3. Core 2 and Core 4 are synthesized by extending Core 1 and core 3 structures via GCNTs enzymes that can be extended to more complex structures (Fig. 1A). The reactivity of the three enzymes were compared for GalNAc3-23mer (semi-glycosylated) and GalNAc5-23mer (completely-glycosylated) as models for site occupancy. Serial dilutions of both substrates were prepared by normalizing the concentrations to the number of GalNAc-glycosylated sites (Fig. 4A). While the overall activity of C1GALT1 was much higher than that of ST6GALNAC1, the results show that ST6GALNAC1 preferably sialylates the GalNAc5-23mer MUC1 whereas C1GALT1 preferably glycosylates the GalNAc3-23mer MUC1. The difference in the selectivity of both C1GALT1 and ST6GALNAC1 increases with an increase in MUC1 concentration. The kinetics parameters derived from the dose-response of ST6GALNAC1 for both semi- and completely-glycosylated MUC1 indicates that the enzyme has a lower affinity (Km value of 0.114 mM for GalNAc5-23mer vs 0.062 mM for GalNAc3-23mer) but higher turnover (Vmax for GalNAc5-23mer is 180% of the Vmax of GalNAc3-23mer) of the fully glycosylated MUC1 compared with the semi glycosylated MUC1 (table in Fig. 4A). On the other hand, no significant difference in the Km values of C1GALT1 were observed between the semi-saturated and completely saturated MUC1. This indicates that C1GALT1 prefers glycosylation of the GALNT_D-sites over the GALNT_L-sites at any concentration of MUC1 and ST6GALNAC1 prefers glycosylation of the GALNT_L-sites over the GALNT_D-sites at high concentrations of MUC1. The activity of ST6GALNAC2 on both acceptors was tested as well, and no significant glycosylation was detected despite the confirmation of expression of ST6GALNAC2 in active form when tested with asialofetuin. 2.3.1 Evaluating ST6GALNAC1 Tn site specificity The formation of the sTn MUC1 is a key antigen in several cancers, consequently a detailed molecular description of the location of this epitope on the MUC1 frame is essential for drug discovery as well as vaccine development. In the section detailing Michaelis-Menten kinetics experiments (Fig. 4A) it was revealed that the GALNT1_D-sites are slowly sialylated alongside the rapidly sialylated GALNT1_L-sites A computational study focused on three distinct MUC1 reaction configurations: (I) the sialylation of the T20 residue on a partially glycosylated peptide, (II) the sialylation of the T20 residue on a fully glycosylated peptide, and (III) the sialylation of the T13 residue on a fully glycosylated peptide (Fig. 5A). Typical MUC1-ST6GALNAC1 poses taken from Free Energies of Adaptive Reaction Coordinate Forces (FEARCF) 27 , 28 reaction dynamics trajectories are shown (Fig. 5B). Free energy reaction profiles were extracted in the form of minimum energy pathways for each reaction so providing the energetic details for the molecular transformation of reactants through two transition states to products (Fig. 5C). While the reaction mechanisms are common to all three sialylation processes (Fig. 5D), these simulations revealed critical insights into the differences in each of the peptide enzyme binding as well as the molecular reaction kinetics and mechanisms. The minimum energy pathways (MEPs) were determined as one-dimensional (1D) reaction coordinates (Fig. 5C) defined in Fig. 5D. The calculated MEPs are consistent with the desiccation-driven mechanism. 29 The MEPs for each sialylation reaction reveals that the formation of an intermediate after the cleavage of CMP is a mechanistic feature common to all three setups. This intermediate is pivotal to the two-step reaction mechanism and reinforces the highly coordinated series of substrate catalytic domain interactions engineered by ST6GALNAC1. The sialylation of the T20 residue on partially glycosylated peptides (I) exhibited an energetic profile that is distinct from its fully glycosylated (II) counterpart (Fig. 5C). Similarly, the T13 residue sialylation on fully glycosylated peptides presented unique energetic and kinetic characteristics. The sialylation of the T20 on the semi-glycosylated peptide formed the oxocarbenium intermediate (OC) with an energy of 19.42 kcal/mol after surmounting a transition state 1 (TS1) energy barrier of 21.24 kcal/mol. The sialylated product results after the Michaelis complex overcomes a second transition state 2 (TS2) barrier of 24.82 kcal/mol. In the case when Tn is sialylated at the T20 on a completely-glycosylated peptide (II), product formation was observed following transition state energy barriers of 22.12 kcal/mol (TS1) and 25.21 kcal/mol (TS2). The elevated reaction energy profiles of the sialyation at the T20 GALNT1_D-site for both reaction configurations is consistent with the slower reactivity observed experimentally and detailed in 2.3 above. The sialylation of T13 on the completely glycosylated peptide (III) proceeded via a transition state 1 (TS1) energy barrier of 16.88 kcal/mol to form a stable oxocarbenium intermediate (OC) with an energy of 11.39 kcal/mol. The formation of products was observed after overcoming a transition state 2 (TS2) energy barrier of 17.46 kcal/mol. This confirms the preference ST6GALNAC1 for GALNT1_L-sites over GALNT1_D-sites observed in the Michaelis-Menten kinetics experiments (section 2.3 ). The molecular reasons for this are that the sugar conformation necessary for the nucleophilic attack is critical to the sialylation reaction. If the sugar conformation is not within the near attack conformation cone angle, the histidine may covalently bind the anomeric carbon after the disassociation of the phosphate. This is the scenario in the T20 I and II T20 case where a side product forms when the catalytic histidine (HIS657) is covalently bonded to the anomeric carbon of the sialic acid lead. For the reaction to proceed without side product formation, the primary alcohol group must be sandwiched between the anomeric carbon of the sialic acid and the proton accepting nitrogen of the catalytic histidine. The most stable sandwiched structure occurs for the Michaelis complex at the T13 site. This leads to an efficient conversion of reactants to products, and with no side product formation. These molecular details explain the observed differences in rates that are experimentally measured and discussed in 2.3 (Table in Fig. 4A). 2.3.2 In vitro synthesis of Core 1 (T antigen) The synthesis of Core 1 via C1GALT1 was performed using different one-pot designs to simulate the varying distribution of GALNTs across the ER-Golgi system between healthy and tumor settings. Two scenarios were modeled: (1) when C1GALT1 competes with GALNTs, mimicking their co-localization in the cis-Golgi, and (2) when GALNTs act on the peptide first in isolation, representing their re-localization to the ER. When C1GALT1 was mixed with GALNT1 (Fig. 4B), the chromatogram displayed two peaks, indicating the incorporation of Gal-GalNAc at one or two sites. When C1GALT1 was mixed with GALNT2, a uniform product with two-site occupancy was observed. These findings demonstrate that mixing either GALNT1 or GALNT2 with C1GALT1 reduces site occupancy by one compared to when GALNTs act alone. This suggests that after GalNAc is added to T8 or T20 via the direct mechanism of GALNT1 or GALNT2, respectively, C1GALT1 competes with these enzymes’ lectin domains for the modified sites. This competition results in the synthesis of Core 1 structures and inhibits subsequent GalNAc glycosylation at other sites via the lectin-dependent mechanism of GALNTs. In a one-pot reaction containing GALNT1, GALNT2, and C1GALT1, a major peak corresponding to a peptide with two Gal-GalNAc modifications was produced. This represents a reduction of one glycosylation site compared to the reaction with GALNT1 and GALNT2 alone (Fig. 3C). These results suggest that GalNAc glycosylation of S19 by GALNT1 and/or GALNT2 occurs exclusively via a lectin-dependent mechanism. A similar inhibition was observed for GALNT4: When C1GALT1 and GALNT4 reacted with the product of glycosylation by GALNT1 and GALNT2, only three sites with Core 1 structures were identified (Fig. 4B), indicating that the lectin-dependent mechanism of GALNT4 is also inhibited. Finally, when simulating the ER localization of GALNTs (i.e., when the product of GALNT1, 2 and 4 in combination was incubated with C1GALT1), C1GALT1 generated five sites occupied by Core 1 structures. The impact of GALNT co-localization with C1GALT1 on site occupancy can be extended to their co-localization with ST6GALNAC1. This is supported by a study showing that overexpression of ST6GALNAC1 reduced site occupancy by 25% in Chinese Hamster Ovary (CHO) cells. 12 2.4 Core 1 Sialylation Core 1 can be sialylated via ST3GAL1 to form the sialyl-3-T antigen or via the ST6GALNAC family to form the sialyl-6-T antigen (Fig. 4C). However, despite the reported activities of the three enzymes on the T antigen, the specificities of these enzymes were not tested comprehensively using standardized substrates. Additionally, details of ST6GALNAC1 vs ST6GALNAC2 specificities for Tn and T are conflicting across studies and the models (in vitro vs in vivo). 30 , 31 Both ST6GALNAC1 and ST6GALNAC2 show no reactivity with the stand alone GalNAc or Gal-GalNAc acceptor, which confirms the peptide core requirement for both enzymatic activities. 30 , 31 That was confirmed when both showed significant reactivity with asialofetuin (data not shown). We showed above that ST6GALNAC1 (but not ST6GALNAC2) reacts with Tn antigen. The same observation was true for the T antigen (Fig. 4, C). When compared to ST6GALNAC1, ST3GAL1 showed significantly higher reactivity with the T antigen, thus the results suggest that sialyl-3-T is more predominant than sialyl-6-T. 3. Discussion The in vitro reconstruction of the MUC1 GalNAc-O-glycosylation pathway delivered new insights and previously unknown details of a) the competitive specificity of each enzyme to the substrate’s MUC1 locale, b) the possible glycosylated combinations of MUC1 (glycoforms) produced by each enzyme as well as the predominance of the glycoforms, c) the effect of the sequence of the GTs in the glycosylation procession and d) the effect of GT compartmentation (co-localization) on the glycosylation products profile. To illustrate the utility of the approach, the reported effect of GALNTs relocating to the ER, on site occupancy was confirmed. However, insight into the nature of the GALNT relocation effect showed that two mechanisms are at the root of this effect. Firstly, the isolation of the GALNT in the ER extends the exposure to the substrate and so the time to react which leads to complete glycosylation and the saturation of the MUC1 glycosylation target sites through lectin-dependent mechanisms. Secondly, further glycosylation of the GALNT_D-sites by C1GALT1 and ST6GALNAC1 leads to the inhibition of the lectin-dependent mechanism of GALNTs, so preventing complete saturating MUC1 GALNT_L-sites with a primary GalNAc. The in vitro enzyme specificity and competition at each step of the synthesis revealed that B3GNT6 specificity to MUC1 Tn sites are negligible compared with C1GALT1 and ST6GALNAC1. This finding corresponds with previous summations that core 3 and core 4 structures are less predominant in MUC1. Previously a comparison between normal epithelial breast cell lines and breast cancer cell lines recorded the absence of core 3 and core 4 structures from both types of cells. 32 Of greater note, the hypothesis that ST6GALNAC1 and ST6GALNAC2 may have an equal propensity to Tn and T antigens was invalidated by the observation that the relative specificity of ST6GALNAC1 toward Tn and T antigens is significantly greater than ST6GALNAC2. The conclusion is that ST6GALNAC1 and not ST6GALNAC2 is responsible for the α-2 sialylation of these two antigens. The reactivity of T antigen catalyzed by ST3GAL1 compared with ST6GALNAC1 showed that the former is more reactive, suggesting that route to ST via sialyl-3-T is the more likely, than via sialyl-6-T. A comparison of the activity of C1GALT1 and ST6GALNAC1 in the fully glycosylated vs the semi glycosylated MUC1 Tn antigen reactions revealed greater C1GALT1 activity in catalyzing the galactose bond at the GALNT_D-sites (associated with T8, S19, and T20) while ST6GALNAC1 was more active in the sialyation of the completely glycosylated GALNT_L-sites (associated with S9 and T13). Revisiting the chemoenzymatic synthesis approach undertaken by Yoshimura et al. 33 it is apparent that ST6GALNAC1 displayed significant specificity to catalyzing the reaction at threonine T13 compared with T8 and T20. In the same study, S19 was not sialylated by ST6GALNAC1 while limited activity was detected at S9. The reaction dynamics simulations revealed the molecular rationale for this observation. A model rationalizing the differences observed in normal epithelial cell glycosylation compared with tumor epithelial cell glycosylation is now possible (Fig. 6). In the normal case, the co-localization of GALNTs, C1GALT1 and ST6GALNAC1 in the cis-Golgi results in GALNTs competing with ST6GALNAC1 preventing sluggish GALNT glycosylation to form GALNT_L-sites making only GALNT_D-sites. In contrast in tumor cells, the localization of GALNTs in the ER, isolated from the competitive cis-Golgi GTs, allows slower GalNAc catalysis to form GALNT_L-sites as well. This alteration leading to either MUC1(GALNT_D-sites) or TA-MUC1 (GALNT_D-sites + GALNT_L-sites) is necessary for the ST6GALNAC1 and C1GALT1 activity. The activity of C1GALT1 is greater toward at the GALNT_D-sites, which leads to enrichment of glycans with an extended Core1 found in normal MUC1 epithelial cells. Whereas the presence of GALNT_L-sites in tumor-associated epithelial cells provides the opportunity for ST6GALNAC1 to sialylate TA-MUC1 at GALNT_L-sites and so the synthesis of the tumor-associated sTn antigen (at high density). 4. Methods 4.1 MUC1 model peptide design: A tobacco etch virus (TEV) protease recognition sequence, was added to enable peptide cleavage at any step of the synthesis, retaining the native MUC1 sequence (Fig. 2A and S1B). These features were designed to facilitate in vitro enzymatic glycosylation and enable simple one-step purification (Fig. 1C). The carrier protein, along with the linkers and the His-tag, is collectively referred to as the “tag” throughout the manuscript. The design of the fusion protein, ensuring sufficient separation between the fusion protein vehicle and the glycosylation target peptide to prevent interference in the synthesis regime was achieved through the assistance of AlphaFold structure prediction tools. The rigid regions have greater conformational and structural predicted confidence compared with the flexible regions (Fig. 2A). The biosynthesis of GalNAc O-linked glycans (the Tn antigen), are initiated through GALNTs’ catalysis of the reactions forming the α-linkage between GalNAc and Serine or Threonine residues. Each GALNT has a catalytic and lectin-binding domain. There are three mechanisms through which GALNT glycosylation can occur, (i) glycosylation of the naked peptide using only the catalytic domain, (ii) glycosylation of pre-glycosylated peptides through the lectin domain, and (ii) a combination of the two mechanisms occurring sequentially. The lectin-binding domain functions as an anchor binding to pre-glycosylated sites limiting diffusional forces to make focused access to neighboring serine or threonine residues possible for the catalytic domain. Several structural elements including the subunits comprising the lectin domain, the properties of the linker between the catalytic and lectin domain, and the structure of the catalytic domain, determine the specificity and the mechanism of the sequential glycosylation for each GALNT. 23 , 34 , 35 Previously, the specificity of the isoforms GALNT1, 2, and 4 were extensively investigated either using the MUC1 peptide containing multiple tandem repeats or a single repeat with an extension of the N terminal to the first threonine residue, known as TAP-24 peptide (Figure S1 , A). 20 , 21 , 23 , 24 , 36 These designs revealed the specificities of these enzymes. However, we discovered and show here that the TAP24 MUC1 construct employed in these studies is not a representative repeat of the natural tandem repeat able to illustrate the chemical biological stepwise glycosylation process. The optimal sequence must be inclusive of all the variables that determine GALNT specificity, and each variable must only be represented once in the peptide sequence. The commonly used TAP24 MUC1 construct is therefore unsuitable for the quantification of enzyme specificities and the biosynthesis of glycosylated MUC1. The following criteria was therefore set for the optimal MUC1 peptide sequence: (1) includes all the five unique potential glycosylation sites of the MUC1 tandem repeat, and each site is only represented once in the sequence, (2) none of the sites is located at the peptide terminus, and the tandem repeat must be sufficiently extended in both directions of the glycosylation site to account for the motif specificity of the catalytic domain, (3) the frame of the sequence must be optimized for the position of each site in relation to the rest of the sites to accommodate the direction specificity of the lectin domain in GALNTs. To construct the optimal MUC1 peptide model, two peptides were designed, as illustrated in Figure S1 , A, and Fig. 3, A: the 27-mer and the 23-mer. While each peptide independently meets the first two criteria, their combined design satisfies the third criterion. 4.2 Expression of MUC1 peptide fusion protein and tag removal Gene sequences encoding MUC1 peptide fusion proteins were synthesize and inserted in the pET-21b(+) E. coli expression vector by BIOMATIK (Ontario, Canada). The sequence of the synthesized fusion protein is provided in the supplementary information (Figure S1 ). Transformation of E. Coli. BL21 (Sigma-Aldrich, Cat. no. CMC0014) was carried out using the heat shock protocol and grown on Luria Broth (LB) agar plates supplemented with ampicillin to a final concentration of 50 µg/ml at 37°C. Expression was carried out in Terrific Broth (TB) medium overnight at 16°C with shaking at 150 rpm. Glucose was added to a concentration of 2% wt/vol at all stages of expression. Expression was induced by isopropyl β-D-thiogalactoside (IPTG, Sigma-Aldrich, Cat. No. 16758) to a final concentration of 1 mM at OD600 of 0.4–0.6. The cell pellets from E. coli expression were lysed by incubating at 4°C for 4 hours in an IMAC binding buffer consisting of 20 mM Tris-HCl, 5 mM Imidazole-HCl, 500 mM NaCl, 0.05% (w/v) sodium azide, and 10% (v/v) glycerol at pH 7.9. This buffer was supplemented with a protease inhibitor tablet (complete, Sigma Aldrich, Cat. no. 11873580001) and 20 mg of lysozyme per 10 ml of buffer. The cell lysate was then clarified through centrifugation at 48,000 RCF for 30 minutes followed by filtration using a 0.45 µM sterile filter. IMAC was conducted on a protein liquid chromatography (FPLC) system ÄKTA Start utilizing 1 ml HiTrap Chelating High-Performance columns (Cytiva, Cat. no. 17-0408-01). Proteins were eluted with a gradient ranging from 5 to 500 mM imidazole-HCl over 15 minutes at a flow rate of 1 ml/min. The eluted fractions were collected, and SDS-PAGE was used to verify protein purity and size. The purified enzymes were quantified with the Bradford protein assay kit (ThermoFisher, A55866) and stored at -80°C in a freezing buffer comprising 20 mM Tris-HCl, 150 mM NaCl, and 10% (v/v) glycerol at pH 7.6. TEV protease was expressed from the expression plasmid pRK793 (Addgene plasmid #8827; http://n2t.net/addgene:8827 ; RRID: Addgene_8827), in E. coli. BL21(DE3)-RIL cells as described previously. 14 Briefly, Tag removal was performed overnight at 4°C by incubating the TEV protease with the fusion proteins at an optimized ratio (Supportive Information, Figure S2) in a buffer of 50 mM Tris-HCl, 0.5 mM EDTA, 1 mM DTT, pH 8.0 Expression of the 23mer and the 27mer was carried out in E. coli BL21 (DE3). A single-step purification using IMAC yielded around 60 mg of pure fusion protein from 250 ml cell culture. Tag cleavage using TEV protease was performed and SDS-PAGE gel confirmed complete cleavage of the naked peptide and the peptide displaying glycosylation on various sites (Figure S2). 4.3 Expression of Glycosyltransferases HEK293F cells (R79007, Thermo Fisher) were a gift from E. Sturrock (University of Cape Town, South Africa). All expression plasmids for GTs used in this work were purchased from the plasmid repository DNASU: HsCD00522282 for GALNT1, HsCD00413124 for GALNT2, HsCD00413161 for GALNT4, HsCD00413129 for GALNT7, HsCD00413109 for ST6GALNAC2, HsCD00413042 for C1GALT1C1, HsCD00413169 for ST3GAL1. As for the pGEn2 vectors of C1GALT1, ST6GALNAC1 and B3GNT6, they were obtained directly from Kelley Moremen (University of Georgia, The USA). GTs expressed from pGEn2 vectors are N-terminally tagged with signalling peptide-8X His-Avi tag-super folder GFP-Tev protease recognition sites. To remove the tag and obtain the enzymes in their native sequence, purified (tagged) proteins were incubated overnight at 4°C with TEV protease at a ratio of 1:5 (TEV protease: fusion protein) in TEV protease buffer (50 mM Tris-HCl, 0.5 mM EDTA, 1 mM DTT, pH 8.0). The mixture was then processed by an IMAC to collect the untagged enzymes from the flow through fractions. Enzymes were stored at − 80°C in freezing buffer (20 mM Tris-HCl, 150 mM NaCl and 10% (v/v) Glycerol, pH 7.6). 4.4 Enzyme assay The enzymatic reactions were monitored using the previously developed UGC assay, 14 the components of the assay are: l-Lactic dehydrogenase (LDH, L2500), PK (P1506), donors such as CMP-sialic acid (CMP-Neu5Ac, C1006), phosphoenolpyruvic acid monopotassium salt (PEP-K, 860077), adenosine 5′-triphosphate disodium salt trihydrate (ATP, 10519979001), β-Nicotinamide adenine dinucleotide, reduced disodium salt hydrate (NADH, 10128023001), bovine serum albumin (BSA, A3059), N-(2-hydroxyethyl) piperazine-N′-(2-ethanesulfonic acid), 4-(2-hydroxyethyl) piperazine-1-ethanesulfonic acid (HEPES, H3375), all purchased from Sigma Aldrich. The 27-mer peptide used as a standard naked peptide in Fig. 2, B was synthesized at GL Biochem, Shanghai. Nucleoside diphosphate kinase (NDK) and cytidylate kinase (CMK) were expressed in house as described in the original protocol. Peptides used as fusion proteins were quantified using Bradford assay. The concentration of the GT enzymes was standardized in all the kinetics and activity assays at 250 nM. Donor concertation was standardized at 2mM. The induction of reaction and preparation of reaction mixtures were performed following the standard protocol. 14 Briefly, Master mixtures were prepared in CMK-NDK-PK-LDH format for sialyltransferase enzymes and in NDK-PK-LDH for the other enzymes. In addition to the coupling enzymes, master mixtures included: reaction buffer, BSA, NADH, phosphoenolpyruvate, ATP, donor, and the constant reaction components (enzyme or acceptor). The variable reaction component was distributed to the wells in triplicate and the induction was carried out by adding the master mixture to the well simultaneously. Corresponding controls including reactions without the variable component were also included. Fluorescence signals were recorded in real time and initial rates were calculated from the slope of progress curves within the initial linear range and normalized to the corresponding blank controls. Rates were converted from fluorescence unit to molarity/minute using the standard curved obtained for the nucleotide corresponding to the donor used in the reaction as explained in the original protocol. 4.5 Preparative glycosylation The glycosylation of the peptides in their fusion form was performed in a one-pot reaction mixture containing the peptide fusion protein at a concentration of 100 µM, donors at 2 mM, GT enzymes at 560 nM. The buffer used was 50 mM HEPES-NaOH buffer, 50 mM KCl, 10 mM MnCl2, 10 mM MgCl2, and 0.02 vol/vol Triton X-100 pH 7.4. A total of 1 ml reaction mixtures were prepared and incubated for 8 hours at 37°C. After completion of the reaction, the glycosylated peptides in their fusion form were purified in a one-step IMAC. The eluted fusion proteins were then subjected to buffer exchange to either freezing buffer (20 mM Tris-HCl, 150 mM NaCl and, 10% (v/v) Glycerol, pH 7.6) when a subsequent glycosylation reaction is performed or to TEV protease cleavage buffer (50 mM Tris-HCl, 0.5 mM EDTA, 1 mM DTT, pH 8.0) when LC-MS was performed. To ensure that the tag elements were inert and did not interfere with glycosylation, a fusion protein containing a MUC1 27-mer peptide was designed and expressed (Figure S1 ). The protein was cleaved using TEV protease, yielding the 27-mer peptide and the tag. The tag cleavage was confirmed by SDS-PAGE (Figure S2 B). 4.6 LC-MS analysis LC-MS analysis was performed using a Q-Exactive quadrupole-Orbitrap mass spectrometer (Thermo Fisher Scientific, USA) coupled with a Dionex Ultimate 3000 nano-UPLC system, and data was acquired with Xcalibur v4.1.31.9, Chromeleon v6.8 (SR13), Orbitrap MS v2.9 (build 2926), and Thermo Foundations 3.1 (SP4). Peptides were prepared in a solution containing 0.1% (v/v) formic acid (FA) and 2% (v/v) acetonitrile (ACN). Final concentrations of the peptides were estimated to be around 10 nM, with a volume equivalent to 50 fmol of peptide injected per sample. Samples were trapped on a PepMap100 C18 column and separated using a ReproSil-Pur 120 C-18-AQ column with a multi-step gradient of Solvent A (0.1% FA in LC water) and Solvent B (0.1% FA in ACN). The mass spectrometer was operated in positive ion mode at a capillary temperature of 320°C and an electrospray voltage of 1.95 kV. Full scan and data-dependent MS/MS settings were used to determine MS1 and MS2 m/z distributions, details of sample preparation and procedure are in the Supporting Information. 4.7 Statistics and Regression All the data points were repeated three times. Shapiro–Wilk normality tests were performed to confirm the normal distribution of the data. Normality tests and Michaelis-Menten regression were performed using GraphPad Prism 8 software as described previously. 14 The data are presented as (mean value) ± (standard deviation). 4.9 Computational methods The ST6GALNAC1 structure was obtained from the SWISS-MODEL 37 database and used to prepare the Michaelis complex. The protein sequence of the CMP-binding position was determined by structural alignment with the crystal structure of ST6GALNAC2 (PDB ID: 6APL), which was subsequently modified to CMP-Neu5Ac. The binding position of the 23mer MUC1 peptide was determined using the HADDOCK 38 webserver to perform protein-protein docking. HADDOCK generated 10 top clusters, each with 4 representative poses. The lowest-energy cluster placed T13 of the MUC1 23mer peptide closest to the acceptor GalNAc residue in the catalytic pocket. The glycosidic bond between T13 and the GalNAc was modeled, and the resulting glycopeptide structure was energy-minimized. Additional GalNAc residues were modeled, producing glycopeptide III (Fig. 5A). Subsequent mutations, peptide extensions, and glycosylations yielded glycopeptides I and II (Fig. 5A). The Michaelis complex for each system was modelled with the CHARMM36 force field 39 . All systems were solvated in TIP3P water molecules in a cubic water box with a 12Å buffer from the edge of the protein. The system was neutralized and NaCl ions were added to a concentration of 0.15M. Initial energy minimization was performed for 1,000 steps using the ABNR method followed by a 1000 step minimization using the SD method. Thermal equilibration was performed for 100 ps with an NPT ensemble and 100 ns classical molecular dynamics simulation was performed in an NVT ensemble. Particle-mesh Ewald (PME) summation was used to calculate electrostatic interactions. The Verlet cutoff scheme was applied to van der Waals (vdW) interactions with a cutoff distance of 12 Å and a switch distance of 10 Å. SHAKE algorithm is used in all simulations to constrain hydrogen bonds. Simulations were carried out with a time step of 2 fs. All simulations were performed at a temperature of 300 K and under a pressure of 1 bar. The reaction dynamics simulations were modeled using hybrid quantum mechanical/molecular mechanical methods (QM/MM). 40 The QM region comprised HIS567 (catalytic base) sidechain atoms, HIS552 (stabilizing the CMP phosphate group), the Neu5Ac moiety, and the CMP phosphate group. The nonpolar C-C bonds were treated with hydrogen link atoms, and the polar C-O bonds were treated with the Simple Link Atom Saccharide Hybrid (SLASH) 41 method. Free energy simulations were conducted along two reaction coordinates to monitor bond formation and breaking using the Free Energies of Adaptive Reaction Coordinate Forces (FEARCF) 27 , 28 method. These were expressed as a linear combination of bond forming (C2-O6) and bond breaking (C2-O2) primary reaction coordinates (Fig. 5D). Histograms for each reaction coordinate were generated using 110 bins spanning a sampling range of 0.5–6 Å. Twelve FEARCF iterations were performed, each comprising 120 simulations of 30 ps duration that was preceded by a 2 ps equilibration run culminating in production runs each totalling of 43.2 ns. Simulations were performed in a 37 Å water sphere using stochastic boundaries. A buffer shell (30–37 Å from the system origin) was applied. The reaction region was modeled with the mio1-1 DFTB3 parameter set, which includes corrections for hydrogen bonding and dispersion interactions. Forces between partial charges beyond 12 Å were zeroed using an atom-wise force-shifting function. Dynamics in the system were modeled using Langevin dynamics with a 2 fs time step. Declarations Funding Sources The National Research Foundation (NRF) CPPR 466624 and the South African Medical Research Council Self-Initiated Grant (SAMRC SIG) 416090. ACKNOWLEDGMENT This work is based in part upon research supported by the National Research Foundation (NRF) CPPR 466624 grant (K.J.N.) and the South African Medical Research Council Self-Initiated Grant (SAMRC SIG) 416090 grant. A.N. Thanks to the Scientific Computing Research Unit (SCRU) for graduate fellowship funding. We thank the Centre for High Performance Computing (CHPC) for computational resources (CHEM0840). References Pothukuchi, P. et al. Translation of genome to glycome: role of the Golgi apparatus. FEBS Lett 593, 2390–2411 (2019). Ashkani, J. & Naidoo, K.J. Glycosyltransferase Gene Expression Profiles Classify Cancer Types and Propose Prognostic Subtypes. Scientific Reports 6, 26451 (2016). Pinho, S.S. & Reis, C.A. Glycosylation in cancer: mechanisms and clinical implications. Nat Rev Cancer 15, 540–555 (2015). Murrell, M.P., Yarema, K.J. & Levchenko, A. The Systems Biology of Glycosylation. ChemBioChem 5, 1334–1347 (2004). Goth, C.K. et al. Chemokine binding to PSGL-1 is controlled by O-glycosylation and tyrosine sulfation. Cell Chemical Biology 30, 893–905.e7 (2023). Gill, D.J., Chia, J., Senewiratne, J. & Bard, F. Regulation of O-glycosylation through Golgi-to-ER relocation of initiation enzymes. J Cell Biol 189, 843–58 (2010). Gill, D.J. et al. Initiation of GalNAc-type O-glycosylation in the endoplasmic reticulum promotes cancer cell invasiveness. Proc Natl Acad Sci U S A 110, E3152-61 (2013). Bui, S., Mejia, I., Diaz, B. & Wang, Y. Adaptation of the Golgi Apparatus in Cancer Cell Invasion and Metastasis. Front Cell Dev Biol 9, 806482 (2021). Rabouille, C. et al. Mapping the distribution of Golgi enzymes involved in the construction of complex oligosaccharides. J Cell Sci 108 (Pt 4), 1617–27 (1995). Taylor-Papadimitriou, J., Burchell, J., Miles, D.W. & Dalziel, M. MUC1 and cancer. Biochimica et Biophysica Acta (BBA) - Molecular Basis of Disease 1455, 301–313 (1999). Nath, S. & Mukherjee, P. MUC1: a multifaceted oncoprotein with a key role in cancer progression. Trends in Molecular Medicine 20, 332–342 (2014). Sewell, R. et al. The ST6GalNAc-I sialyltransferase localizes throughout the golgi and is responsible for the synthesis of the tumor-associated Sialyl-Tn -glycan in human breast cancer. Journal of Biological Chemistry 281, 3586–3594 (2006). Pantazopoulou, A. & Glick, B.S. A Kinetic View of Membrane Traffic Pathways Can Transcend the Classical View of Golgi Compartments. Frontiers in Cell and Developmental Biology 7(2019). Nashed, A. & Naidoo, K.J. Universal Glycosyltransferase Continuous Assay for Uniform Kinetics and Inhibition Database Development and Mechanistic Studies Illustrated on ST3GAL1, C1GALT1, and FUT1. ACS Omega 9, 17518–17532 (2024). Pedelacq, J.D., Cabantous, S., Tran, T., Terwilliger, T.C. & Waldo, G.S. Engineering and characterization of a superfolder green fluorescent protein. Nat Biotechnol 24, 79–88 (2006). Luo, J. et al. SRC kinase-mediated signaling pathways and targeted therapies in breast cancer. Breast Cancer Res 24, 99 (2022). Hugonnet, M., Singh, P., Haas, Q. & von Gunten, S. The Distinct Roles of Sialyltransferases in Cancer Biology and Onco-Immunology. Front Immunol 12, 799861 (2021). Ju, T. et al. Human tumor antigens Tn and sialyl Tn arise from mutations in Cosmc. Cancer Res 68, 1636–46 (2008). Zeng, J. et al. Promoters of Human Cosmc and T-synthase Genes Are Similar in Structure, Yet Different in Epigenetic Regulation. J Biol Chem 290, 19018–33 (2015). Coelho, H. et al. Atomic and Specificity Details of Mucin 1 O-Glycosylation Process by Multiple Polypeptide GalNAc-Transferase Isoforms Unveiled by NMR and Molecular Modeling. JACS Au 2, 631–645 (2022). Wandall, H.H. et al. Substrate specificities of three members of the human UDP-N-acetyl-alpha-D-galactosamine:Polypeptide N-acetylgalactosaminyltransferase family, GalNAc-T1, -T2, and -T3. J Biol Chem 272, 23503–14 (1997). Revoredo, L. et al. Mucin-type O-glycosylation is controlled by short- and long-range glycopeptide substrate recognition that varies among members of the polypeptide GalNAc transferase family. Glycobiology 26, 360 – 76 (2016). Collette, A.M. et al. An unusual dual sugar-binding lectin domain controls the substrate specificity of a mucin-type O-glycosyltransferase. Sci Adv 10, eadj8829 (2024). Wandall, H.H. et al. The lectin domains of polypeptide GalNAc-transferases exhibit carbohydrate-binding specificity for GalNAc: lectin binding to GalNAc-glycopeptide substrates is required for high density GalNAc-O-glycosylation. Glycobiology 17, 374 – 87 (2007). Hanisch, F.G. & Muller, S. MUC1: the polymorphic appearance of a human mucin. Glycobiology 10, 439 – 49 (2000). Wang, S. et al. Chemoenzymatic modular assembly of O-GalNAc glycans for functional glycomics. Nat Commun 12, 3573 (2021). Naidoo, K.J. FEARCF a multidimensional free energy method for investigating conformational landscapes and chemical reaction mechanisms. Science China Chemistry 54, 1962–1973 (2011). Bruce-Chwatt, T. & Naidoo, K.J. Molecular mechanisms from reaction coordinate graph enabled multidimensional free energies illustrated on water dimer hydrogen bonding. Journal of Computational Chemistry 43, 1802–1813 (2022). Hansen, T. et al. Defining the SN1 Side of Glycosylation Reactions: Stereoselectivity of Glycopyranosyl Cations. ACS Central Science 5, 781–788 (2019). Kono, M. et al. Redefined substrate specificity of ST6GalNAc II: a second candidate sialyl-Tn synthase. Biochem Biophys Res Commun 272, 94–7 (2000). Marcos, N.T. et al. Role of the human ST6GalNAc-I and ST6GalNAc-II in the synthesis of the cancer-associated sialyl-Tn antigen. Cancer Res 64, 7050–7 (2004). Lloyd, K.O., Burchell, J., Kudryashov, V., Yin, B.W. & Taylor-Papadimitriou, J. Comparison of O-linked carbohydrate chains in MUC-1 mucin from normal breast epithelial cell lines and breast carcinoma cell lines. Demonstration of simpler and fewer glycan chains in tumor cells. J Biol Chem 271, 33325–34 (1996). Yoshimura, Y. et al. Products of Chemoenzymatic Synthesis Representing MUC1 Tandem Repeat Unit with T-, ST- or STn-antigen Revealed Distinct Specificities of Anti-MUC1 Antibodies. Scientific Reports 9, 16641 (2019). Pratt, M.R. et al. Deconvoluting the functions of polypeptide N-alpha-acetylgalactosaminyltransferase family members by glycopeptide substrate profiling. Chem Biol 11, 1009–16 (2004). de Las Rivas, M., Lira-Navarrete, E., Gerken, T.A. & Hurtado-Guerrero, R. Polypeptide GalNAc-Ts: from redundancy to specificity. Curr Opin Struct Biol 56, 87–96 (2019). Hassan, H. et al. The lectin domain of UDP-N-acetyl-D-galactosamine: polypeptide N-acetylgalactosaminyltransferase-T4 directs its glycopeptide specificities. J Biol Chem 275, 38197–205 (2000). Waterhouse, A. et al. SWISS-MODEL: homology modelling of protein structures and complexes. Nucleic Acids Research 46, W296-W303 (2018). Honorato, R.V. et al. Structural Biology in the Clouds: The WeNMR-EOSC Ecosystem. Frontiers in Molecular Biosciences 8(2021). Huang, J. et al. CHARMM36m: an improved force field for folded and intrinsically disordered proteins. Nature Methods 14, 71–73 (2017). Field, M.J., Bash, P.A. & Karplus, M. A Combined Quantum Mechanical and Molecular Mechanical Potential for Molecular Dynamics Simulations. J. Comput. Chem. 11, 700–733 (1990). Crous, W., Field, M.J. & Naidoo, K.J. Simple Link Atom Saccharide Hybrid (SLASH) Treatment for Glycosidic Bonds at the QM/MM Boundary. Journal of Chemical Theory and Computation 10, 1727–1738 (2014). Additional Declarations There is NO Competing Interest. Supplementary Files SupportingInformationOnePotMuc1BiosynthesisF.docx Supporting Informaction: An in vitro one-pot synthetic biology approach to simulating diverging Golgi O-glycosylation of tumor-associated MUC1 from normal tissue MUC1 Cite Share Download PDF Status: Published Journal Publication published 22 Apr, 2026 Read the published version in Nature Communications → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5783651","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":400559913,"identity":"0c6cb08a-804f-4b59-9329-c08e95032e1b","order_by":0,"name":"Kevin Naidoo","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA60lEQVRIie3PsQrCMBCA4StCXYJdTwR9hRRBcFBfpRLoVMFJHEQQIV2Ks5uv0hDQRXwCB0Xo5CAI0klMqoNLq24O+SFwBD6SAzCZ/jBUJ9aDDR7A5WdiLZ8Xn8krD0rkG1INI1cOYd+olJk4dXh3CmV5sNJJPqmRLZVLSFxOEtYccIZAfAq4zid19D1JQFocg1ZtwGP1sQCA2gWkkWSkl5G2Js5Z7XUv+BiWYk36GbE0QfWK4AXrR0wRKpnexY12rMoxoWK2yCe4EfMrGcvOKmTikI66juOw4zG95ZNn9G3Wi8efgMlkMpmKewBeKEfwJerIvgAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0002-9898-3708","institution":"University of Cape Town","correspondingAuthor":true,"prefix":"","firstName":"Kevin","middleName":"","lastName":"Naidoo","suffix":""},{"id":400559914,"identity":"b5f4a1ac-5a2e-4692-8bcc-848fa43d0287","order_by":1,"name":"Abdullateef Nashed","email":"","orcid":"","institution":"University of Cape Town","correspondingAuthor":false,"prefix":"","firstName":"Abdullateef","middleName":"","lastName":"Nashed","suffix":""},{"id":400559915,"identity":"b147b0f2-cdf5-4705-955d-3e43b696749a","order_by":2,"name":"Kyllen Dilsook","email":"","orcid":"","institution":"University of Cape Town","correspondingAuthor":false,"prefix":"","firstName":"Kyllen","middleName":"","lastName":"Dilsook","suffix":""},{"id":400559916,"identity":"f50edcc8-dff8-470d-b06e-c2148bb3ae51","order_by":3,"name":"Tharindu Senapathi","email":"","orcid":"","institution":"University of Cape Town","correspondingAuthor":false,"prefix":"","firstName":"Tharindu","middleName":"","lastName":"Senapathi","suffix":""}],"badges":[],"createdAt":"2025-01-07 18:30:48","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5783651/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5783651/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41467-026-72151-y","type":"published","date":"2026-04-22T04:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":73872991,"identity":"8b8e9b60-61d3-4c48-9c05-d76a69eb8b3f","added_by":"auto","created_at":"2025-01-15 12:33:51","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1009210,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMucin O-GalNAc glycosylation.\u003c/strong\u003e A. O-GalNAc glycosylation pathway showing the reactions that lead to the formation of the cancer-associated Tn and T antigens and their sialylated forms sTn and sT, respectively. B. Differential localization of GALNTs between normal epithelial tumor epithelial cells attributed to COP-1 mediated retrograde activation in cancer shown alongside the corresponding in vitro synthesis design.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-5783651/v1/e6ad1a2cba7af0279c2087e6.png"},{"id":73873000,"identity":"509878de-af88-4d1b-9e53-ee773cb5795d","added_by":"auto","created_at":"2025-01-15 12:33:51","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1039805,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe one-pot biosynthesis method.\u003c/strong\u003e A. An AlphaFold structure of the fusion protein for peptide expression. The fusion protein design connected to the MUC1 frame selected from the peptide tandem repeat. B. The initial reaction rates for the 27-mer MUC1 peptide in its tag-fused and unfused forms in addition to the tag used as acceptors for GalNAc catalysed by different GALNTs. C. The pipeline of expression and enzymatic glycosylation of the MUC1 peptide used to construct the pathway model.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-5783651/v1/9c319bbc4ab4a6548ef7fa8f.png"},{"id":73872992,"identity":"db5a17ce-63ff-4fcc-9a2f-8d728222c97a","added_by":"auto","created_at":"2025-01-15 12:33:51","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":797999,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eO-GalNAc glycosylation pathway initiation (Synthesis of Tn antigens).\u003c/strong\u003e A. Reactivity of the 23-mer and the 27-mer peptides with the different GALNT enzymes. Initial rates were calculated from the linear range of the progress curves of the reaction and reported as (Mean ± SD). B. A schematic illustration of the direct and lectin-dependent mechanisms of the glycosylation reactions of GALNT1, GALNT2 and GALNT4 with the 23-mer and the 27-mer peptides. C. In vitro synthesis of GalNAc-glycosylated 23mer: I. Brief description of synthesis designs and steps. II. Abundance of the products of each reaction as measured by liquid chromatography. III. Structures of the products of each reaction as determined from the LC-MS analysis and the synthesis pathway leads to these structures as concluded from the preceding Data. Details of the LC-MS data and analysis are in the Supportive Information.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-5783651/v1/6f26dffcada85a239b2309e9.png"},{"id":73873008,"identity":"88d42767-25b2-4ab6-af69-256e2af95b2e","added_by":"auto","created_at":"2025-01-15 12:33:52","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":866768,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAssessment of the first layer of the structure extension after Tn synthesis.\u003c/strong\u003e A. Effect of GalNAc site occupancy on the specificity of ST6GALNAC1, C1GALT1 and B3GNT6. Serial dilution of concentrations of GalNAc3-23mer and GalNAc5-23mer were calculated per GalNAc site occupancy. Data is presented as (mean ± SD, n = 3). B. Core 1 Synthesis: Effect of GALNTs vs C1GALT1 competition on site occupancy. I. synthesis designs of different combinations of sequential synthesis reflecting C1GALT1 competition with GALNTs in isolation or in different combinations. II. shows the LC peaks and structures deduced from analyzing mass spectrometry results (Supportive Information). C. Sialylation of Core 1 via ST3GAL1, ST6GALNAC1, and ST6GALNAC2. Bars in panels A and C are represented as means ± SD, n = 3), pairs marked with asterisk indicates p value \u0026lt; 0.01.\u003c/p\u003e","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-5783651/v1/ddef8760f99c102276be3da8.jpeg"},{"id":73873013,"identity":"12c7ca1f-cfec-4b1e-bd15-11ada9bed07e","added_by":"auto","created_at":"2025-01-15 12:33:52","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":4925259,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComputational modeling of the reaction site specificity of ST6GALNAC1\u003c/strong\u003e A. Schematic representation of the reactions modeled. I. Sialylation of T20 on a partially glycosylated peptide. II. Sialylation of T20 on a fully glycosylated peptide. III. Sialylation of T13 on a fully glycosylated peptide. B. Snapshot of 23mer MUC1 (red ribbon) undergoing glycosylation with its GalNac residues (green ball and stick) binding to the ST6GALNAC1 surface (colored white) harboring the CMP-Neu5Ac donor (blue). C. 1D representation of the minimum energy pathway determined from the free energy surface showing the free energy associated with the bond forming and bond breaking in going from the reactants (R) to the products (P) via two transition states (TS1 \u0026amp; TS2) and an oxocarbenium intermediate (OC) D. SN1-like reaction mechanism of ST6GALNAC1. The position of link atoms added to prepare the system for QM/MM simulations are marked in green and red. Enzyme amino acids drawn in blue\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-5783651/v1/5ed7cecd7512a0259456a4c4.png"},{"id":73873010,"identity":"afb33304-cfdb-4d40-b323-882e037e2876","added_by":"auto","created_at":"2025-01-15 12:33:52","extension":"jpeg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1009608,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eA proposed model of the effect of the localization of GALNTs on site occupancy and sialylation of the Tn antigen.\u003c/strong\u003eThe left panel illustrates normal cellular MUC1 glycosylation, while the right panel illustrates tumor cellular MUC1 glycosylation.\u003c/p\u003e","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-5783651/v1/a788dc2de222b224b4614ef4.jpeg"},{"id":107603813,"identity":"5364d8a2-b70c-4fe7-9f78-e83d4b5ab366","added_by":"auto","created_at":"2026-04-23 07:11:04","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":9686411,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5783651/v1/e1802a5a-218f-45bb-966c-0bc5aaef1262.pdf"},{"id":73872994,"identity":"c0ffe0ba-1162-4e03-99bb-3bb580378b8b","added_by":"auto","created_at":"2025-01-15 12:33:51","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":3196067,"visible":true,"origin":"","legend":"Supporting Informaction: An in vitro one-pot synthetic biology approach to simulating diverging Golgi O-glycosylation of tumor-associated MUC1 from normal tissue MUC1","description":"","filename":"SupportingInformationOnePotMuc1BiosynthesisF.docx","url":"https://assets-eu.researchsquare.com/files/rs-5783651/v1/30f26b8ff48ef3639880678a.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"An in vitro one-pot synthetic biology approach to simulating diverging Golgi O-glycosylation of tumor-associated MUC1 from normal tissue MUC1","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eThe post-translational event of peptide or protein glycosylation is a non-template-driven process that relies on more than just glycoenzyme gene expression data or even the glycoenzyme expression levels themselves.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e A case in point is that while glycosyltransferase (GT) gene expression can be used to classify cancer,\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e this genomic level data cannot be used to directly infer the difference between cancerous and healthy glycoconjugate expression. Specifically, the characteristically high degree of sialylation observed in tumor tissues\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e and the associated structural modifications of glycans cannot be directly correlated with the genes that express the sialyltransferases. This is because the complex glycosylation pathways within cells are intimately connected and intertwined with their critical metabolic and regulatory networks.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e This complexity dissociates the high degree of sialylation observed in tumor tissues\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e and structural modifications of glycans directly from gene and protein expression. Consequently, a critical first step in systems biology modelling is the construction of a developmental model to mimic sequential biosynthesis processes within the ER and Golgi apparatus.\u003c/p\u003e\n\u003cp\u003eSite specific chemoenzymatic synthesis is the standard method for producing model glycopeptides, where the initial sugar peptide bond is chemically synthesised. An enzymatic synthesis of the glycan ensues one sugar residue at a time.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e Advances in understanding enzyme specificities and mechanisms have enabled the synthesis of more complex glycopeptides,\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e expanding the dimensions of glycopeptide arrays. In vitro methods that produce single, purifiable, and spectroscopically verifiable structures are more suitable than synthetic biology methods since the intention is to measure the kinetics of GTs as well as map out their selectivity and mechanistic action. The localization of GTs, such as GALNTs, has been found to be a regulatory mechanism involved in cancer phenotypes by altering O-GalNAc glycan structures and levels.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e Consequently, in vitro methods must have the capability to mimic the alterations of in vivo glycosylation resulting from the spatial-temporal rearrangement of the distribution and presentation of GTs to the substrate in the ER-Golgi system.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e8\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\n\u003cp\u003eThe first step in mucin-type O-linked glycosylation is the addition of GalNAc to serine or threonine residues facilitated by several N-acetylgalactosaminyltransferases (GALNTs), forming the Thomsen-nouvelle (Tn) antigen (Fig.\u0026nbsp;1A). Following this the addition of galactose to the Tn antigen through T synthase (C1GALT1) is modified to form the T antigen (core 1). Alternative to this, the core 3 can be made by the addition of GlcNAc via b-1,3-N-acetylglucosaminyltransferase 6 (B3GNT6).\u003c/p\u003e\n\u003cp\u003eThese foundational structures undergo further branching and elongation with successive additions of monosaccharides such as GlcNAc and galactose, generating diverse glycan chains. Sialyltransferases mediate the sialylation of Tn and T antigens (left pathways in Fig.\u0026nbsp;1A). These sialylated forms (sTn and sT) terminates chain progression. Clinically, Tn, T, sTn, and sT antigens are significant through their role in establishing the hallmarks of cancer such as tumor progression, immune evasion, and metastasis.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e10\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e The glycosylation of proteins takes place in the ER-Golgi system mostly through glycosyltransferases (GTs) and in some instances in combination with glycosidases. These glycoenzymes are distributed across specific cisternae.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e8\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e9\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e The localization of GTs along the ER-Golgi axis is dynamic and they are constantly shuffled in both directions via a complex tightly regulated system involving COP-I and COP-II vesicles.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e12\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e This localization across various cisternae has led to an assembly line of compartments performing sequential glycosylation to build glycans on target proteins. In the case of an organism disease state a protein\u0026rsquo;s glycan is often altered when there is deregulation of this localization, such as the relocation of GALNTs from the cis Golgi to the ER in tumour formation.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e (Fig.\u0026nbsp;1, B).\u003c/p\u003e\n\u003cp\u003eHere we use Mucin 1 (MUC1), as a peptide glycosylation prototype systems model. The aim is to resolve reasoning underlying differences in enzymatic construction in normal glycosylated MUC1 compared with tumor-associated (TA) MUC1. An in vitro method is developed to simulate ER-Golgi conditions for O-GalNAc glycosylation of the MUC1 and TA-MUC1 peptides. Specifically, we illustrate: i) the kinetic parameters governing glycosyltransferase (GT) activities and substrate specificities using the UGC assay\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e, ii) the molecular mechanisms governing the GTs site specificities, and iii) the effect of their expression and distribution along the ER-Golgi axis. Following this, the experimental model along with advanced computer reaction dynamics simulations, were used to reveal peptide site specificity of each of the GTs involved in the synthesis of Tn, T, sTn, and sT antigens at the five unique MUC1 glycosylation sites.\u003c/p\u003e"},{"header":"2. Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\n\u003ch2\u003e\u003cstrong\u003e2.1\u003c/strong\u003e The Sequential One-pot synthesis method:\u003c/h2\u003e\n\u003cp\u003eTo capture the spatial-temporal segregation of glycosylation pathways a fusion tag protein carrying the MUC1 peptide was designed as an assembly conveyor (Fig.\u0026nbsp;2A). The conveyor vehicle (fusion protein and tags) was tested for biosynthesis interference (Fig.\u0026nbsp;2B). In the assembly line design for a glycan biosynthesis, the kinetics of the GT-catalyzed reactions and the associated intermediate glycan products are analyzed at every point of construction along the assembly line (Fig.\u0026nbsp;2C). The data obtained at each point informs subsequent steps, supporting model construction and iterative optimization of the synthesis.\u003c/p\u003e\n\u003cp\u003eA fusion protein containing the core MUC1 peptide and a carrier protein, superfolder green fluorescent protein (sfGFP), expressed in E. coli. sfGFP was selected for its folding efficiency, minimized dimerization, and enhanced solubility.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e To minimize the possibility of interaction with the MUC1 peptide, sfGFP was separated from the peptide using a linker (Linker 2) comprising of three rigid and three flexible units (from the N to C direction). Additionally, a rigid linker (Linker 1) was incorporated to improve steric presentation to the N-terminus His-tag, enhancing affinity-based purification. Central to the design are the work functions of the linkers. Firstly, the rigid region on Linker 2 must maximize the peptide sfGFP distance. Secondly, the rigid linker 1 must maximize the presentation of the His-tag for later TEV protease cleavage when salvaging the glycosylated MUC1. Here the predicted low conformational and structural predicted confidence around the TEV protease cleavage region signifies flexibility and so the designed accessibility of the protease.\u003c/p\u003e\n\u003cp\u003eThe GalNAc acceptor functions of the 27-mer in its fusion form and cleaved form (plus tag) were tested for each of the GALNT enzymes: GALNT1, GALNT2, GALNT4, and GALNT7 (Fig.\u0026nbsp;2B). The tag proved not to interfere with the enzyme activity or substrate specificity since the glycosylation rates were identical for the tagged and untagged MUC1 for the GTs. The fusion protein glycosylation carrier function was therefore optimized while preserving the inherent glycan recipient functions of the target peptide (MUC).\u003c/p\u003e\n\u003cdiv id=\"Sec4\" class=\"Section3\"\u003e\n\u003ch2\u003e\u003cstrong\u003e2.1.2\u003c/strong\u003e In vitro healthy vs. tumor GT distribution and concentration models\u003c/h2\u003e\n\u003cp\u003eThe GalNAc-glycosylated sites are subject to either sialylation (addition of Sia) via ST6GALNAC1, galactosylation (addition of Gal) via C1GALT1 and its chaperone C1GALT1C1 (Cosmc), or N-acetylglucosaminylation (addition of GlcNAc) via B3GNT6 (Fig.\u0026nbsp;1A). In healthy contexts, GALNTs and C1GALT1 are localized in the cis-Golgi, while the ST6GALNAC1 is distributed across all the cisternae of the Golgi (Fig.\u0026nbsp;1, B).\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e8\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e9\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e No specific localization of B3GNT6 has been reported. The localization of only the GALNTs were reported to be altered in response to the EGF stimulation of SRC (the proto-oncogene in cancer) via COP-1 mediated retrograde from cis Golgi to the ER.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e This relocation results in the overexpression of Tn in the ER where a fraction of this Tn transits to the cell surface without modification, while another fraction transits with modification to T antigen. In patient samples, the same study found that the mean expression of Tn in breast cancer tissue samples was 4.5 folds higher than in normal tissues. From the samples with high Tn expression, 70% of the samples showed ER localization of GALNTs (inferred indirectly from ER localization of Tn), whereas no significant loss of C1GALT1 was detected in these samples, pointing to the ER localization as the driving factor of the observed Tn overexpression.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\n\u003cp\u003eAccompanying the relocation of GALNTs, the expression of O-GalNAc glycosylation enzymes is altered in cancer compared to normal cells. Furthermore, ST6GALNAC1 was reported to be upregulated in almost all cancer types\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e and C1GALT1 downregulated, mainly due to Cosmc mutation or epigenetic alteration of both C1GALT1 and Cosmc.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e The objective here is to build a one-pot in vitro synthesis model that will be representative of the impact of the redistribution of GALNTs in altering glycan structures independently of enzyme levels (Fig.\u0026nbsp;1B and Fig.\u0026nbsp;2C). To achieve this, all activities and kinetics experiments are performed at a standardized enzyme concentration of 250 nM.\u003c/p\u003e\n\u003cp\u003eThe performance of the 23-mer and 27-mer peptides was compared by measuring the relative reactivity of GALNT1, GALNT2, GALNT4, and GALNT7 individually and in various combinations using the UGC assay\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e (Fig.\u0026nbsp;3A). The primary observation from both experiments is that GALNT4 alone does not react with either peptide, confirming it relies on a strictly lectin-dependent mechanism. GALNT7 shows no reactivity with either peptide in any combination. On the other hand, GALNT1 and GALNT2 could individually react with both peptides due to their direct catalytic mechanism. The rates of glycosylation via GALNT1 and GALNT2 were generally lower for the 23-mer than the 27-mer. This is explained by the depletion of the direct glycosylation-specific sites T13 and T20 for GALNT1 and GALNT2 respectively, in the 23-mer case, in addition to the absence of secondary lectin-assisted sites in the preferred direction. In contrast, the 27-mer permits the continuation of glycosylation via the lectin-assisted sites (Fig.\u0026nbsp;3B). When GALNT4 was tested in combination with GALNT1 or GALNT2, using the 27-mer construct, it showed no activity. However, GALNT4 exhibited significant activity using the 23mer peptide when combined with GALNT2, as indicated by the enhanced glycosylation rate when compared with the reaction of GALNT2 individually. GALNT2 is known to have a preferred specificity to T20 (in the PGST sequence) via direct catalytic domain recognition. In the 23-mer, T20 is at the N-terminal to T13 and S9 (GALNT4 specific sites), and pre-glycosylation at T20 left to T13 and S9 sites is the prerequisite for GALNT4 lectin-dependent specificity (Fig.\u0026nbsp;3B).\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e The collective evidence presented here confirms the efficiency of the fusion protein design to mimic natural MUC1 while its tag components do not affect the catalytic function of the GALNTs. Additionally, 23-mer has proved valid to capture the reaction mechanism of GALNT4 and the direct mechanism for both GALNT1 and GALNT2. However, the reported lectin-dependent mechanisms for GALNT1 and GALNT2 can be better studied using the 27mer.\u003c/p\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\n\u003ch2\u003e\u003cstrong\u003e2.2.\u003c/strong\u003e In vitro synthesis of the Tn antigen\u003c/h2\u003e\n\u003cp\u003eIn vitro synthesis of the Tn antigen was carried out to explore the possible structures enabled by extensive glycosylation (higher enzyme concentrations for a longer period) using varying combinations of GALNTs (Fig.\u0026nbsp;3C I). LC-MS results confirmed the structure and purity of the peptide (i). The catalytic action of GALNT1 results in glycosylation of 2 or 3 sites, and 3 sites in the case of GALNT2. The glycosylation of T8 and T20 is consistent with a direct GALNT1 and GALNT2 mechanism respectively. However, the observed glycosylation of the remaining two sites cannot be attributed to the lectin-dependent mechanisms as previously reported for the GALNT1 and GALNT2 activity on MUC1.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e This is evident from short-term recorded activity on the remaining sites (Fig.\u0026nbsp;3, A) since the 23mer does not provide the directionality required for this mechanism. A randomized peptide sequences study found that GALNT1 can catalyze a long-range N terminal lectin-dependent mechanism while GALNT 2 can participate in both N and C long-range lectin-dependent mechanism.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e Another study confirmed the bi-directionality of the lectin-dependent mechanism for both enzymes with preference given to N and C long range lectin-dependent mechanisms for GALNT1 and GALNT2, respectively\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. No further glycosylation beyond three sites was observed when GALNT1 and GALNT2 were combined, suggesting the absence of synergy between the two enzymes and that the two remaining sites do not conform to the specificity of these GTs.\u003c/p\u003e\n\u003cp\u003eWhen the product resulting from GALNT1 and GALNT2 activity was glycosylated with GALNT4, all five sites were glycosylated. These results indicate that the remaining two sites are GALNT4 specific and are glycosylated via a lectin-dependent mechanism. The consensus from all previously reported results is that T8 and T20 are specific sites for GALNT1 and GALNT2 respectively, and S9 and T13 are GALNT4 specific. Taken together, it can be concluded that GALNT1, GALNT2, or their combination can glycosylate S19, T20, and T8 by utilizing direct and lectin-dependent mechanisms. However, GALNT1 is less efficient than GALNT2 in completing the glycosylation of either S19 or T20 (Fig.\u0026nbsp;3C III).\u003c/p\u003e\n\u003cp\u003eIt is widely accepted that the GalNAc site occupancy increases with GALNTs over expression and their ER relocation.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e The effect of these two factors was simulated here in the One-pot biosynthesis model where extensive glycosylation was performed to interrogate the action of GALNTs in isolation of other GTs. It was seen that sites, such as T20, S19, and T8 were not selective and can be glycosylated by multiple GALNTs, such as GALNT1 and GALNT2 as shown here, as well as GALNT3 that was previously reported.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e We now refer to these sites that are the first ones to be glycosylated as GALNT_D-sites. The S9 and T13 sites are strictly lectin-dependent and thus are dependent on prior glycosylation of T20, S19, and T8 which we refer to now as GALNT_L-sites (Fig.\u0026nbsp;4A). The lectin-dependent mechanism is therefore responsible for high-density GalNAc-O-glycosylation.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e This differential site occupancy was also found to be associated with cancer transformation. The site saturation was observed in MUC1 expressed in tumor cells compared with normal MUC1 in breast milk.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e The GalNAc3-23mer (3 glycosylated sites at GALNT_D-sites only) and GalNAc5-23mer (5 glycosylated sites at GALNT_D-sites and GALNT_L-sites), synthesized here will be used as models for site occupancy, semi-glycosylated and completely-glycosylated, respectively.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\n\u003ch2\u003e\u003cstrong\u003e2.3\u003c/strong\u003e In vitro synthesis of the T and sTn antigen and core 3\u003c/h2\u003e\n\u003cp\u003eThe sequential addition of one or two sugars to the Tn antigen generates diverse glycan core structures, with eight different cores have been identified. The most common structures are cores 1\u0026ndash;4 (Fig.\u0026nbsp;1A) while cores 5\u0026ndash;8 are rare.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e Core 1, also known as T antigen, is formed by adding galactose to the Tn antigen via a \u0026beta;1\u0026ndash;3 bond, a process catalyzed by T synthase (C1GALT1) with the help of its chaperone C1GALT1C1 (Cosmc). Within this context the core 3 is synthesized by B3GNT6 catalyses of the GlcNAc covalent \u0026beta;1\u0026ndash;3 bond of to the Tn antigen. The ST6GALNAc1 sialylates Tn antigen to sialyl-Tn (sTn) that terminates the synthesis preventing the structures from undergoing further glycosylation via C1GALT1 or B3GNT3. Core 2 and Core 4 are synthesized by extending Core 1 and core 3 structures via GCNTs enzymes that can be extended to more complex structures (Fig.\u0026nbsp;1A).\u003c/p\u003e\n\u003cp\u003eThe reactivity of the three enzymes were compared for GalNAc3-23mer (semi-glycosylated) and GalNAc5-23mer (completely-glycosylated) as models for site occupancy. Serial dilutions of both substrates were prepared by normalizing the concentrations to the number of GalNAc-glycosylated sites (Fig.\u0026nbsp;4A). While the overall activity of C1GALT1 was much higher than that of ST6GALNAC1, the results show that ST6GALNAC1 preferably sialylates the GalNAc5-23mer MUC1 whereas C1GALT1 preferably glycosylates the GalNAc3-23mer MUC1. The difference in the selectivity of both C1GALT1 and ST6GALNAC1 increases with an increase in MUC1 concentration. The kinetics parameters derived from the dose-response of ST6GALNAC1 for both semi- and completely-glycosylated MUC1 indicates that the enzyme has a lower affinity (Km value of 0.114 mM for GalNAc5-23mer vs 0.062 mM for GalNAc3-23mer) but higher turnover (Vmax for GalNAc5-23mer is 180% of the Vmax of GalNAc3-23mer) of the fully glycosylated MUC1 compared with the semi glycosylated MUC1 (table in Fig.\u0026nbsp;4A). On the other hand, no significant difference in the Km values of C1GALT1 were observed between the semi-saturated and completely saturated MUC1. This indicates that C1GALT1 prefers glycosylation of the GALNT_D-sites over the GALNT_L-sites at any concentration of MUC1 and ST6GALNAC1 prefers glycosylation of the GALNT_L-sites over the GALNT_D-sites at high concentrations of MUC1. The activity of ST6GALNAC2 on both acceptors was tested as well, and no significant glycosylation was detected despite the confirmation of expression of ST6GALNAC2 in active form when tested with asialofetuin.\u003c/p\u003e\n\u003cdiv id=\"Sec7\" class=\"Section3\"\u003e\n\u003ch2\u003e\u003cstrong\u003e2.3.1\u003c/strong\u003e Evaluating ST6GALNAC1 Tn site specificity\u003c/h2\u003e\n\u003cp\u003eThe formation of the sTn MUC1 is a key antigen in several cancers, consequently a detailed molecular description of the location of this epitope on the MUC1 frame is essential for drug discovery as well as vaccine development. In the section detailing Michaelis-Menten kinetics experiments (Fig.\u0026nbsp;4A) it was revealed that the GALNT1_D-sites are slowly sialylated alongside the rapidly sialylated GALNT1_L-sites A computational study focused on three distinct MUC1 reaction configurations: (I) the sialylation of the T20 residue on a partially glycosylated peptide, (II) the sialylation of the T20 residue on a fully glycosylated peptide, and (III) the sialylation of the T13 residue on a fully glycosylated peptide (Fig.\u0026nbsp;5A). Typical MUC1-ST6GALNAC1 poses taken from Free Energies of Adaptive Reaction Coordinate Forces (FEARCF)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e27\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e reaction dynamics trajectories are shown (Fig.\u0026nbsp;5B). Free energy reaction profiles were extracted in the form of minimum energy pathways for each reaction so providing the energetic details for the molecular transformation of reactants through two transition states to products (Fig.\u0026nbsp;5C). While the reaction mechanisms are common to all three sialylation processes (Fig.\u0026nbsp;5D), these simulations revealed critical insights into the differences in each of the peptide enzyme binding as well as the molecular reaction kinetics and mechanisms.\u003c/p\u003e\n\u003cp\u003eThe minimum energy pathways (MEPs) were determined as one-dimensional (1D) reaction coordinates (Fig.\u0026nbsp;5C) defined in Fig.\u0026nbsp;5D. The calculated MEPs are consistent with the desiccation-driven mechanism.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e The MEPs for each sialylation reaction reveals that the formation of an intermediate after the cleavage of CMP is a mechanistic feature common to all three setups. This intermediate is pivotal to the two-step reaction mechanism and reinforces the highly coordinated series of substrate catalytic domain interactions engineered by ST6GALNAC1. The sialylation of the T20 residue on partially glycosylated peptides (I) exhibited an energetic profile that is distinct from its fully glycosylated (II) counterpart (Fig.\u0026nbsp;5C). Similarly, the T13 residue sialylation on fully glycosylated peptides presented unique energetic and kinetic characteristics.\u003c/p\u003e\n\u003cp\u003eThe sialylation of the T20 on the semi-glycosylated peptide formed the oxocarbenium intermediate (OC) with an energy of 19.42 kcal/mol after surmounting a transition state 1 (TS1) energy barrier of 21.24 kcal/mol. The sialylated product results after the Michaelis complex overcomes a second transition state 2 (TS2) barrier of 24.82 kcal/mol. In the case when Tn is sialylated at the T20 on a completely-glycosylated peptide (II), product formation was observed following transition state energy barriers of 22.12 kcal/mol (TS1) and 25.21 kcal/mol (TS2). The elevated reaction energy profiles of the sialyation at the T20 GALNT1_D-site for both reaction configurations is consistent with the slower reactivity observed experimentally and detailed in 2.3 above.\u003c/p\u003e\n\u003cp\u003eThe sialylation of T13 on the completely glycosylated peptide (III) proceeded via a transition state 1 (TS1) energy barrier of 16.88 kcal/mol to form a stable oxocarbenium intermediate (OC) with an energy of 11.39 kcal/mol. The formation of products was observed after overcoming a transition state 2 (TS2) energy barrier of 17.46 kcal/mol. This confirms the preference ST6GALNAC1 for GALNT1_L-sites over GALNT1_D-sites observed in the Michaelis-Menten kinetics experiments (section \u003cspan class=\"InternalRef\"\u003e2.3\u003c/span\u003e). The molecular reasons for this are that the sugar conformation necessary for the nucleophilic attack is critical to the sialylation reaction. If the sugar conformation is not within the near attack conformation cone angle, the histidine may covalently bind the anomeric carbon after the disassociation of the phosphate. This is the scenario in the T20 I and II T20 case where a side product forms when the catalytic histidine (HIS657) is covalently bonded to the anomeric carbon of the sialic acid lead. For the reaction to proceed without side product formation, the primary alcohol group must be sandwiched between the anomeric carbon of the sialic acid and the proton accepting nitrogen of the catalytic histidine. The most stable sandwiched structure occurs for the Michaelis complex at the T13 site. This leads to an efficient conversion of reactants to products, and with no side product formation. These molecular details explain the observed differences in rates that are experimentally measured and discussed in 2.3 (Table in Fig.\u0026nbsp;4A).\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec8\" class=\"Section3\"\u003e\n\u003ch2\u003e\u003cstrong\u003e2.3.2\u003c/strong\u003e In vitro synthesis of Core 1 (T antigen)\u003c/h2\u003e\n\u003cp\u003eThe synthesis of Core 1 via C1GALT1 was performed using different one-pot designs to simulate the varying distribution of GALNTs across the ER-Golgi system between healthy and tumor settings. Two scenarios were modeled: (1) when C1GALT1 competes with GALNTs, mimicking their co-localization in the cis-Golgi, and (2) when GALNTs act on the peptide first in isolation, representing their re-localization to the ER. When C1GALT1 was mixed with GALNT1 (Fig.\u0026nbsp;4B), the chromatogram displayed two peaks, indicating the incorporation of Gal-GalNAc at one or two sites. When C1GALT1 was mixed with GALNT2, a uniform product with two-site occupancy was observed. These findings demonstrate that mixing either GALNT1 or GALNT2 with C1GALT1 reduces site occupancy by one compared to when GALNTs act alone. This suggests that after GalNAc is added to T8 or T20 via the direct mechanism of GALNT1 or GALNT2, respectively, C1GALT1 competes with these enzymes\u0026rsquo; lectin domains for the modified sites. This competition results in the synthesis of Core 1 structures and inhibits subsequent GalNAc glycosylation at other sites via the lectin-dependent mechanism of GALNTs.\u003c/p\u003e\n\u003cp\u003eIn a one-pot reaction containing GALNT1, GALNT2, and C1GALT1, a major peak corresponding to a peptide with two Gal-GalNAc modifications was produced. This represents a reduction of one glycosylation site compared to the reaction with GALNT1 and GALNT2 alone (Fig.\u0026nbsp;3C). These results suggest that GalNAc glycosylation of S19 by GALNT1 and/or GALNT2 occurs exclusively via a lectin-dependent mechanism. A similar inhibition was observed for GALNT4: When C1GALT1 and GALNT4 reacted with the product of glycosylation by GALNT1 and GALNT2, only three sites with Core 1 structures were identified (Fig.\u0026nbsp;4B), indicating that the lectin-dependent mechanism of GALNT4 is also inhibited.\u003c/p\u003e\n\u003cp\u003eFinally, when simulating the ER localization of GALNTs (i.e., when the product of GALNT1, 2 and 4 in combination was incubated with C1GALT1), C1GALT1 generated five sites occupied by Core 1 structures. The impact of GALNT co-localization with C1GALT1 on site occupancy can be extended to their co-localization with ST6GALNAC1. This is supported by a study showing that overexpression of ST6GALNAC1 reduced site occupancy by 25% in Chinese Hamster Ovary (CHO) cells.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\n\u003ch2\u003e\u003cstrong\u003e2.4\u003c/strong\u003e Core 1 Sialylation\u003c/h2\u003e\n\u003cp\u003eCore 1 can be sialylated via ST3GAL1 to form the sialyl-3-T antigen or via the ST6GALNAC family to form the sialyl-6-T antigen (Fig.\u0026nbsp;4C). However, despite the reported activities of the three enzymes on the T antigen, the specificities of these enzymes were not tested comprehensively using standardized substrates. Additionally, details of ST6GALNAC1 vs ST6GALNAC2 specificities for Tn and T are conflicting across studies and the models (in vitro vs in vivo).\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e30\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e Both ST6GALNAC1 and ST6GALNAC2 show no reactivity with the stand alone GalNAc or Gal-GalNAc acceptor, which confirms the peptide core requirement for both enzymatic activities.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e30\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e That was confirmed when both showed significant reactivity with asialofetuin (data not shown). We showed above that ST6GALNAC1 (but not ST6GALNAC2) reacts with Tn antigen. The same observation was true for the T antigen (Fig.\u0026nbsp;4, C). When compared to ST6GALNAC1, ST3GAL1 showed significantly higher reactivity with the T antigen, thus the results suggest that sialyl-3-T is more predominant than sialyl-6-T.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"3. Discussion","content":"\u003cp\u003eThe in vitro reconstruction of the MUC1 GalNAc-O-glycosylation pathway delivered new insights and previously unknown details of a) the competitive specificity of each enzyme to the substrate\u0026rsquo;s MUC1 locale, b) the possible glycosylated combinations of MUC1 (glycoforms) produced by each enzyme as well as the predominance of the glycoforms, c) the effect of the sequence of the GTs in the glycosylation procession and d) the effect of GT compartmentation (co-localization) on the glycosylation products profile. To illustrate the utility of the approach, the reported effect of GALNTs relocating to the ER, on site occupancy was confirmed. However, insight into the nature of the GALNT relocation effect showed that two mechanisms are at the root of this effect. Firstly, the isolation of the GALNT in the ER extends the exposure to the substrate and so the time to react which leads to complete glycosylation and the saturation of the MUC1 glycosylation target sites through lectin-dependent mechanisms. Secondly, further glycosylation of the GALNT_D-sites by C1GALT1 and ST6GALNAC1 leads to the inhibition of the lectin-dependent mechanism of GALNTs, so preventing complete saturating MUC1 GALNT_L-sites with a primary GalNAc.\u003c/p\u003e\n\u003cp\u003eThe in vitro enzyme specificity and competition at each step of the synthesis revealed that B3GNT6 specificity to MUC1 Tn sites are negligible compared with C1GALT1 and ST6GALNAC1. This finding corresponds with previous summations that core 3 and core 4 structures are less predominant in MUC1. Previously a comparison between normal epithelial breast cell lines and breast cancer cell lines recorded the absence of core 3 and core 4 structures from both types of cells.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e Of greater note, the hypothesis that ST6GALNAC1 and ST6GALNAC2 may have an equal propensity to Tn and T antigens was invalidated by the observation that the relative specificity of ST6GALNAC1 toward Tn and T antigens is significantly greater than ST6GALNAC2. The conclusion is that ST6GALNAC1 and not ST6GALNAC2 is responsible for the \u0026alpha;-2 sialylation of these two antigens. The reactivity of T antigen catalyzed by ST3GAL1 compared with ST6GALNAC1 showed that the former is more reactive, suggesting that route to ST via sialyl-3-T is the more likely, than via sialyl-6-T.\u003c/p\u003e\n\u003cp\u003eA comparison of the activity of C1GALT1 and ST6GALNAC1 in the fully glycosylated vs the semi glycosylated MUC1 Tn antigen reactions revealed greater C1GALT1 activity in catalyzing the galactose bond at the GALNT_D-sites (associated with T8, S19, and T20) while ST6GALNAC1 was more active in the sialyation of the completely glycosylated GALNT_L-sites (associated with S9 and T13). Revisiting the chemoenzymatic synthesis approach undertaken by Yoshimura et al.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e it is apparent that ST6GALNAC1 displayed significant specificity to catalyzing the reaction at threonine T13 compared with T8 and T20. In the same study, S19 was not sialylated by ST6GALNAC1 while limited activity was detected at S9. The reaction dynamics simulations revealed the molecular rationale for this observation.\u003c/p\u003e\n\u003cp\u003eA model rationalizing the differences observed in normal epithelial cell glycosylation compared with tumor epithelial cell glycosylation is now possible (Fig.\u0026nbsp;6). In the normal case, the co-localization of GALNTs, C1GALT1 and ST6GALNAC1 in the cis-Golgi results in GALNTs competing with ST6GALNAC1 preventing sluggish GALNT glycosylation to form GALNT_L-sites making only GALNT_D-sites. In contrast in tumor cells, the localization of GALNTs in the ER, isolated from the competitive cis-Golgi GTs, allows slower GalNAc catalysis to form GALNT_L-sites as well. This alteration leading to either MUC1(GALNT_D-sites) or TA-MUC1 (GALNT_D-sites\u0026thinsp;+\u0026thinsp;GALNT_L-sites) is necessary for the ST6GALNAC1 and C1GALT1 activity. The activity of C1GALT1 is greater toward at the GALNT_D-sites, which leads to enrichment of glycans with an extended Core1 found in normal MUC1 epithelial cells. Whereas the presence of GALNT_L-sites in tumor-associated epithelial cells provides the opportunity for ST6GALNAC1 to sialylate TA-MUC1 at GALNT_L-sites and so the synthesis of the tumor-associated sTn antigen (at high density).\u003c/p\u003e"},{"header":"4. Methods","content":"\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e\u003cb\u003e4.1\u003c/b\u003e MUC1 model peptide design:\u003c/h2\u003e \u003cp\u003eA tobacco etch virus (TEV) protease recognition sequence, was added to enable peptide cleavage at any step of the synthesis, retaining the native MUC1 sequence (Fig.\u0026nbsp;2A and S1B). These features were designed to facilitate in vitro enzymatic glycosylation and enable simple one-step purification (Fig.\u0026nbsp;1C). The carrier protein, along with the linkers and the His-tag, is collectively referred to as the \u0026ldquo;tag\u0026rdquo; throughout the manuscript. The design of the fusion protein, ensuring sufficient separation between the fusion protein vehicle and the glycosylation target peptide to prevent interference in the synthesis regime was achieved through the assistance of AlphaFold structure prediction tools. The rigid regions have greater conformational and structural predicted confidence compared with the flexible regions (Fig.\u0026nbsp;2A).\u003c/p\u003e \u003cp\u003eThe biosynthesis of GalNAc O-linked glycans (the Tn antigen), are initiated through GALNTs\u0026rsquo; catalysis of the reactions forming the α-linkage between GalNAc and Serine or Threonine residues. Each GALNT has a catalytic and lectin-binding domain. There are three mechanisms through which GALNT glycosylation can occur, (i) glycosylation of the naked peptide using only the catalytic domain, (ii) glycosylation of pre-glycosylated peptides through the lectin domain, and (ii) a combination of the two mechanisms occurring sequentially. The lectin-binding domain functions as an anchor binding to pre-glycosylated sites limiting diffusional forces to make focused access to neighboring serine or threonine residues possible for the catalytic domain. Several structural elements including the subunits comprising the lectin domain, the properties of the linker between the catalytic and lectin domain, and the structure of the catalytic domain, determine the specificity and the mechanism of the sequential glycosylation for each GALNT. \u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e,\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e,\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003ePreviously, the specificity of the isoforms GALNT1, 2, and 4 were extensively investigated either using the MUC1 peptide containing multiple tandem repeats or a single repeat with an extension of the N terminal to the first threonine residue, known as TAP-24 peptide (Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e, A).\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e,\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e,\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e These designs revealed the specificities of these enzymes. However, we discovered and show here that the TAP24 MUC1 construct employed in these studies is not a representative repeat of the natural tandem repeat able to illustrate the chemical biological stepwise glycosylation process. The optimal sequence must be inclusive of all the variables that determine GALNT specificity, and each variable must only be represented once in the peptide sequence. The commonly used TAP24 MUC1 construct is therefore unsuitable for the quantification of enzyme specificities and the biosynthesis of glycosylated MUC1.\u003c/p\u003e \u003cp\u003eThe following criteria was therefore set for the optimal MUC1 peptide sequence: (1) includes all the five unique potential glycosylation sites of the MUC1 tandem repeat, and each site is only represented once in the sequence, (2) none of the sites is located at the peptide terminus, and the tandem repeat must be sufficiently extended in both directions of the glycosylation site to account for the motif specificity of the catalytic domain, (3) the frame of the sequence must be optimized for the position of each site in relation to the rest of the sites to accommodate the direction specificity of the lectin domain in GALNTs. To construct the optimal MUC1 peptide model, two peptides were designed, as illustrated in Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e, A, and Fig.\u0026nbsp;3, A: the 27-mer and the 23-mer. While each peptide independently meets the first two criteria, their combined design satisfies the third criterion.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e\u003cb\u003e4.2\u003c/b\u003e Expression of MUC1 peptide fusion protein and tag removal\u003c/h2\u003e \u003cp\u003eGene sequences encoding MUC1 peptide fusion proteins were synthesize and inserted in the pET-21b(+) E. coli expression vector by BIOMATIK (Ontario, Canada). The sequence of the synthesized fusion protein is provided in the supplementary information (Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). Transformation of E. Coli. BL21 (Sigma-Aldrich, Cat. no. CMC0014) was carried out using the heat shock protocol and grown on Luria Broth (LB) agar plates supplemented with ampicillin to a final concentration of 50 \u0026micro;g/ml at 37\u0026deg;C. Expression was carried out in Terrific Broth (TB) medium overnight at 16\u0026deg;C with shaking at 150 rpm. Glucose was added to a concentration of 2% wt/vol at all stages of expression. Expression was induced by isopropyl β-D-thiogalactoside (IPTG, Sigma-Aldrich, Cat. No. 16758) to a final concentration of 1 mM at OD600 of 0.4\u0026ndash;0.6.\u003c/p\u003e \u003cp\u003eThe cell pellets from E. coli expression were lysed by incubating at 4\u0026deg;C for 4 hours in an IMAC binding buffer consisting of 20 mM Tris-HCl, 5 mM Imidazole-HCl, 500 mM NaCl, 0.05% (w/v) sodium azide, and 10% (v/v) glycerol at pH 7.9. This buffer was supplemented with a protease inhibitor tablet (complete, Sigma Aldrich, Cat. no. 11873580001) and 20 mg of lysozyme per 10 ml of buffer. The cell lysate was then clarified through centrifugation at 48,000 RCF for 30 minutes followed by filtration using a 0.45 \u0026micro;M sterile filter. IMAC was conducted on a protein liquid chromatography (FPLC) system \u0026Auml;KTA Start utilizing 1 ml HiTrap Chelating High-Performance columns (Cytiva, Cat. no. 17-0408-01). Proteins were eluted with a gradient ranging from 5 to 500 mM imidazole-HCl over 15 minutes at a flow rate of 1 ml/min. The eluted fractions were collected, and SDS-PAGE was used to verify protein purity and size. The purified enzymes were quantified with the Bradford protein assay kit (ThermoFisher, A55866) and stored at -80\u0026deg;C in a freezing buffer comprising 20 mM Tris-HCl, 150 mM NaCl, and 10% (v/v) glycerol at pH 7.6.\u003c/p\u003e \u003cp\u003eTEV protease was expressed from the expression plasmid pRK793 (Addgene plasmid #8827; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://n2t.net/addgene:8827\u003c/span\u003e\u003cspan address=\"http://n2t.net/addgene:8827\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e; RRID: Addgene_8827), in E. coli. BL21(DE3)-RIL cells as described previously. \u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003eBriefly, Tag removal was performed overnight at 4\u0026deg;C by incubating the TEV protease with the fusion proteins at an optimized ratio (Supportive Information, Figure S2) in a buffer of 50 mM Tris-HCl, 0.5 mM EDTA, 1 mM DTT, pH 8.0\u003c/p\u003e \u003cp\u003eExpression of the 23mer and the 27mer was carried out in E. coli BL21 (DE3). A single-step purification using IMAC yielded around 60 mg of pure fusion protein from 250 ml cell culture. Tag cleavage using TEV protease was performed and SDS-PAGE gel confirmed complete cleavage of the naked peptide and the peptide displaying glycosylation on various sites (Figure S2).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e\u003cb\u003e4.3\u003c/b\u003e Expression of Glycosyltransferases\u003c/h2\u003e \u003cp\u003eHEK293F cells (R79007, Thermo Fisher) were a gift from E. Sturrock (University of Cape Town, South Africa). All expression plasmids for GTs used in this work were purchased from the plasmid repository DNASU: HsCD00522282 for GALNT1, HsCD00413124 for GALNT2, HsCD00413161 for GALNT4, HsCD00413129 for GALNT7, HsCD00413109 for ST6GALNAC2, HsCD00413042 for C1GALT1C1, HsCD00413169 for ST3GAL1. As for the pGEn2 vectors of C1GALT1, ST6GALNAC1 and B3GNT6, they were obtained directly from Kelley Moremen (University of Georgia, The USA). GTs expressed from pGEn2 vectors are N-terminally tagged with signalling peptide-8X His-Avi tag-super folder GFP-Tev protease recognition sites. To remove the tag and obtain the enzymes in their native sequence, purified (tagged) proteins were incubated overnight at 4\u0026deg;C with TEV protease at a ratio of 1:5 (TEV protease: fusion protein) in TEV protease buffer (50 mM Tris-HCl, 0.5 mM EDTA, 1 mM DTT, pH 8.0). The mixture was then processed by an IMAC to collect the untagged enzymes from the flow through fractions. Enzymes were stored at \u0026minus;\u0026thinsp;80\u0026deg;C in freezing buffer (20 mM Tris-HCl, 150 mM NaCl and 10% (v/v) Glycerol, pH 7.6).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e\u003cb\u003e4.4\u003c/b\u003e Enzyme assay\u003c/h2\u003e \u003cp\u003eThe enzymatic reactions were monitored using the previously developed UGC assay,\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e the components of the assay are: l-Lactic dehydrogenase (LDH, L2500), PK (P1506), donors such as CMP-sialic acid (CMP-Neu5Ac, C1006), phosphoenolpyruvic acid monopotassium salt (PEP-K, 860077), adenosine 5\u0026prime;-triphosphate disodium salt trihydrate (ATP, 10519979001), β-Nicotinamide adenine dinucleotide, reduced disodium salt hydrate (NADH, 10128023001), bovine serum albumin (BSA, A3059), N-(2-hydroxyethyl) piperazine-N\u0026prime;-(2-ethanesulfonic acid), 4-(2-hydroxyethyl) piperazine-1-ethanesulfonic acid (HEPES, H3375), all purchased from Sigma Aldrich. The 27-mer peptide used as a standard naked peptide in Fig.\u0026nbsp;2, B was synthesized at GL Biochem, Shanghai. Nucleoside diphosphate kinase (NDK) and cytidylate kinase (CMK) were expressed in house as described in the original protocol. Peptides used as fusion proteins were quantified using Bradford assay. The concentration of the GT enzymes was standardized in all the kinetics and activity assays at 250 nM. Donor concertation was standardized at 2mM. The induction of reaction and preparation of reaction mixtures were performed following the standard protocol.\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e Briefly, Master mixtures were prepared in CMK-NDK-PK-LDH format for sialyltransferase enzymes and in NDK-PK-LDH for the other enzymes. In addition to the coupling enzymes, master mixtures included: reaction buffer, BSA, NADH, phosphoenolpyruvate, ATP, donor, and the constant reaction components (enzyme or acceptor). The variable reaction component was distributed to the wells in triplicate and the induction was carried out by adding the master mixture to the well simultaneously. Corresponding controls including reactions without the variable component were also included. Fluorescence signals were recorded in real time and initial rates were calculated from the slope of progress curves within the initial linear range and normalized to the corresponding blank controls. Rates were converted from fluorescence unit to molarity/minute using the standard curved obtained for the nucleotide corresponding to the donor used in the reaction as explained in the original protocol.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e\u003cb\u003e4.5\u003c/b\u003e Preparative glycosylation\u003c/h2\u003e \u003cp\u003eThe glycosylation of the peptides in their fusion form was performed in a one-pot reaction mixture containing the peptide fusion protein at a concentration of 100 \u0026micro;M, donors at 2 mM, GT enzymes at 560 nM. The buffer used was 50 mM HEPES-NaOH buffer, 50 mM KCl, 10 mM MnCl2, 10 mM MgCl2, and 0.02 vol/vol Triton X-100 pH 7.4. A total of 1 ml reaction mixtures were prepared and incubated for 8 hours at 37\u0026deg;C. After completion of the reaction, the glycosylated peptides in their fusion form were purified in a one-step IMAC. The eluted fusion proteins were then subjected to buffer exchange to either freezing buffer (20 mM Tris-HCl, 150 mM NaCl and, 10% (v/v) Glycerol, pH 7.6) when a subsequent glycosylation reaction is performed or to TEV protease cleavage buffer (50 mM Tris-HCl, 0.5 mM EDTA, 1 mM DTT, pH 8.0) when LC-MS was performed.\u003c/p\u003e \u003cp\u003eTo ensure that the tag elements were inert and did not interfere with glycosylation, a fusion protein containing a MUC1 27-mer peptide was designed and expressed (Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). The protein was cleaved using TEV protease, yielding the 27-mer peptide and the tag. The tag cleavage was confirmed by SDS-PAGE (Figure S2 B).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e\u003cb\u003e4.6\u003c/b\u003e LC-MS analysis\u003c/h2\u003e \u003cp\u003eLC-MS analysis was performed using a Q-Exactive quadrupole-Orbitrap mass spectrometer (Thermo Fisher Scientific, USA) coupled with a Dionex Ultimate 3000 nano-UPLC system, and data was acquired with Xcalibur v4.1.31.9, Chromeleon v6.8 (SR13), Orbitrap MS v2.9 (build 2926), and Thermo Foundations 3.1 (SP4). Peptides were prepared in a solution containing 0.1% (v/v) formic acid (FA) and 2% (v/v) acetonitrile (ACN). Final concentrations of the peptides were estimated to be around 10 nM, with a volume equivalent to 50 fmol of peptide injected per sample. Samples were trapped on a PepMap100 C18 column and separated using a ReproSil-Pur 120 C-18-AQ column with a multi-step gradient of Solvent A (0.1% FA in LC water) and Solvent B (0.1% FA in ACN). The mass spectrometer was operated in positive ion mode at a capillary temperature of 320\u0026deg;C and an electrospray voltage of 1.95 kV. Full scan and data-dependent MS/MS settings were used to determine MS1 and MS2 m/z distributions, details of sample preparation and procedure are in the Supporting Information.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e\u003cb\u003e4.7\u003c/b\u003e Statistics and Regression\u003c/h2\u003e \u003cp\u003eAll the data points were repeated three times. Shapiro\u0026ndash;Wilk normality tests were performed to confirm the normal distribution of the data. Normality tests and Michaelis-Menten regression were performed using GraphPad Prism 8 software as described previously.\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e The data are presented as (mean value) \u0026plusmn; (standard deviation).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003e\u003cb\u003e4.9\u003c/b\u003e Computational methods\u003c/h2\u003e \u003cp\u003eThe ST6GALNAC1 structure was obtained from the SWISS-MODEL\u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e database and used to prepare the Michaelis complex. The protein sequence of the CMP-binding position was determined by structural alignment with the crystal structure of ST6GALNAC2 (PDB ID: 6APL), which was subsequently modified to CMP-Neu5Ac. The binding position of the 23mer MUC1 peptide was determined using the HADDOCK\u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e webserver to perform protein-protein docking. HADDOCK generated 10 top clusters, each with 4 representative poses. The lowest-energy cluster placed T13 of the MUC1 23mer peptide closest to the acceptor GalNAc residue in the catalytic pocket. The glycosidic bond between T13 and the GalNAc was modeled, and the resulting glycopeptide structure was energy-minimized. Additional GalNAc residues were modeled, producing glycopeptide III (Fig.\u0026nbsp;5A). Subsequent mutations, peptide extensions, and glycosylations yielded glycopeptides I and II (Fig.\u0026nbsp;5A).\u003c/p\u003e \u003cp\u003eThe Michaelis complex for each system was modelled with the CHARMM36 force field\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e. All systems were solvated in TIP3P water molecules in a cubic water box with a 12\u0026Aring; buffer from the edge of the protein. The system was neutralized and NaCl ions were added to a concentration of 0.15M. Initial energy minimization was performed for 1,000 steps using the ABNR method followed by a 1000 step minimization using the SD method. Thermal equilibration was performed for 100 ps with an NPT ensemble and 100 ns classical molecular dynamics simulation was performed in an NVT ensemble. Particle-mesh Ewald (PME) summation was used to calculate electrostatic interactions. The Verlet cutoff scheme was applied to van der Waals (vdW) interactions with a cutoff distance of 12 \u0026Aring; and a switch distance of 10 \u0026Aring;. SHAKE algorithm is used in all simulations to constrain hydrogen bonds. Simulations were carried out with a time step of 2 fs. All simulations were performed at a temperature of 300 K and under a pressure of 1 bar.\u003c/p\u003e \u003cp\u003eThe reaction dynamics simulations were modeled using hybrid quantum mechanical/molecular mechanical methods (QM/MM).\u003csup\u003e40\u003c/sup\u003e The QM region comprised HIS567 (catalytic base) sidechain atoms, HIS552 (stabilizing the CMP phosphate group), the Neu5Ac moiety, and the CMP phosphate group. The nonpolar C-C bonds were treated with hydrogen link atoms, and the polar C-O bonds were treated with the Simple Link Atom Saccharide Hybrid (SLASH)\u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e method.\u003c/p\u003e \u003cp\u003eFree energy simulations were conducted along two reaction coordinates to monitor bond formation and breaking using the Free Energies of Adaptive Reaction Coordinate Forces (FEARCF)\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e,\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e method. These were expressed as a linear combination of bond forming (C2-O6) and bond breaking (C2-O2) primary reaction coordinates (Fig.\u0026nbsp;5D). Histograms for each reaction coordinate were generated using 110 bins spanning a sampling range of 0.5\u0026ndash;6 \u0026Aring;. Twelve FEARCF iterations were performed, each comprising 120 simulations of 30 ps duration that was preceded by a 2 ps equilibration run culminating in production runs each totalling of 43.2 ns.\u003c/p\u003e \u003cp\u003eSimulations were performed in a 37 \u0026Aring; water sphere using stochastic boundaries. A buffer shell (30\u0026ndash;37 \u0026Aring; from the system origin) was applied. The reaction region was modeled with the mio1-1 DFTB3 parameter set, which includes corrections for hydrogen bonding and dispersion interactions. Forces between partial charges beyond 12 \u0026Aring; were zeroed using an atom-wise force-shifting function. Dynamics in the system were modeled using Langevin dynamics with a 2 fs time step.\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003eFunding Sources\u003c/p\u003e\n\u003cp\u003eThe National Research Foundation (NRF) CPPR 466624 and the South African Medical Research Council Self-Initiated Grant (SAMRC SIG) 416090.\u003c/p\u003e\u003ch2\u003eACKNOWLEDGMENT\u003c/h2\u003e \u003cp\u003eThis work is based in part upon research supported by the National Research Foundation (NRF) CPPR 466624 grant (K.J.N.) and the South African Medical Research Council Self-Initiated Grant (SAMRC SIG) 416090 grant. A.N. Thanks to the Scientific Computing Research Unit (SCRU) for graduate fellowship funding. We thank the Centre for High Performance Computing (CHPC) for computational resources (CHEM0840).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003ePothukuchi, P. et al. Translation of genome to glycome: role of the Golgi apparatus. FEBS Lett 593, 2390\u0026ndash;2411 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAshkani, J. \u0026amp; Naidoo, K.J. Glycosyltransferase Gene Expression Profiles Classify Cancer Types and Propose Prognostic Subtypes. Scientific Reports 6, 26451 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePinho, S.S. \u0026amp; Reis, C.A. Glycosylation in cancer: mechanisms and clinical implications. Nat Rev Cancer 15, 540\u0026ndash;555 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMurrell, M.P., Yarema, K.J. \u0026amp; Levchenko, A. The Systems Biology of Glycosylation. ChemBioChem 5, 1334\u0026ndash;1347 (2004).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGoth, C.K. et al. Chemokine binding to PSGL-1 is controlled by O-glycosylation and tyrosine sulfation. Cell Chemical Biology 30, 893\u0026ndash;905.e7 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGill, D.J., Chia, J., Senewiratne, J. \u0026amp; Bard, F. Regulation of O-glycosylation through Golgi-to-ER relocation of initiation enzymes. J Cell Biol 189, 843\u0026ndash;58 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGill, D.J. et al. Initiation of GalNAc-type O-glycosylation in the endoplasmic reticulum promotes cancer cell invasiveness. Proc Natl Acad Sci U S A 110, E3152-61 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBui, S., Mejia, I., Diaz, B. \u0026amp; Wang, Y. Adaptation of the Golgi Apparatus in Cancer Cell Invasion and Metastasis. Front Cell Dev Biol 9, 806482 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRabouille, C. et al. Mapping the distribution of Golgi enzymes involved in the construction of complex oligosaccharides. J Cell Sci 108 (Pt 4), 1617\u0026ndash;27 (1995).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTaylor-Papadimitriou, J., Burchell, J., Miles, D.W. \u0026amp; Dalziel, M. MUC1 and cancer. Biochimica et Biophysica Acta (BBA) - Molecular Basis of Disease 1455, 301\u0026ndash;313 (1999).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNath, S. \u0026amp; Mukherjee, P. MUC1: a multifaceted oncoprotein with a key role in cancer progression. Trends in Molecular Medicine 20, 332\u0026ndash;342 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSewell, R. et al. The ST6GalNAc-I sialyltransferase localizes throughout the golgi and is responsible for the synthesis of the tumor-associated Sialyl-Tn -glycan in human breast cancer. Journal of Biological Chemistry 281, 3586\u0026ndash;3594 (2006).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePantazopoulou, A. \u0026amp; Glick, B.S. A Kinetic View of Membrane Traffic Pathways Can Transcend the Classical View of Golgi Compartments. Frontiers in Cell and Developmental Biology 7(2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNashed, A. \u0026amp; Naidoo, K.J. Universal Glycosyltransferase Continuous Assay for Uniform Kinetics and Inhibition Database Development and Mechanistic Studies Illustrated on ST3GAL1, C1GALT1, and FUT1. ACS Omega 9, 17518\u0026ndash;17532 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePedelacq, J.D., Cabantous, S., Tran, T., Terwilliger, T.C. \u0026amp; Waldo, G.S. Engineering and characterization of a superfolder green fluorescent protein. Nat Biotechnol 24, 79\u0026ndash;88 (2006).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLuo, J. et al. SRC kinase-mediated signaling pathways and targeted therapies in breast cancer. Breast Cancer Res 24, 99 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHugonnet, M., Singh, P., Haas, Q. \u0026amp; von Gunten, S. The Distinct Roles of Sialyltransferases in Cancer Biology and Onco-Immunology. Front Immunol 12, 799861 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJu, T. et al. Human tumor antigens Tn and sialyl Tn arise from mutations in Cosmc. Cancer Res 68, 1636\u0026ndash;46 (2008).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZeng, J. et al. Promoters of Human Cosmc and T-synthase Genes Are Similar in Structure, Yet Different in Epigenetic Regulation. J Biol Chem 290, 19018\u0026ndash;33 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCoelho, H. et al. Atomic and Specificity Details of Mucin 1 O-Glycosylation Process by Multiple Polypeptide GalNAc-Transferase Isoforms Unveiled by NMR and Molecular Modeling. JACS Au 2, 631\u0026ndash;645 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWandall, H.H. et al. Substrate specificities of three members of the human UDP-N-acetyl-alpha-D-galactosamine:Polypeptide N-acetylgalactosaminyltransferase family, GalNAc-T1, -T2, and -T3. J Biol Chem 272, 23503\u0026ndash;14 (1997).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRevoredo, L. et al. Mucin-type O-glycosylation is controlled by short- and long-range glycopeptide substrate recognition that varies among members of the polypeptide GalNAc transferase family. \u003cem\u003eGlycobiology\u003c/em\u003e 26, 360\u0026thinsp;\u0026ndash;\u0026thinsp;76 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCollette, A.M. et al. An unusual dual sugar-binding lectin domain controls the substrate specificity of a mucin-type O-glycosyltransferase. Sci Adv 10, eadj8829 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWandall, H.H. et al. The lectin domains of polypeptide GalNAc-transferases exhibit carbohydrate-binding specificity for GalNAc: lectin binding to GalNAc-glycopeptide substrates is required for high density GalNAc-O-glycosylation. \u003cem\u003eGlycobiology\u003c/em\u003e 17, 374\u0026thinsp;\u0026ndash;\u0026thinsp;87 (2007).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHanisch, F.G. \u0026amp; Muller, S. MUC1: the polymorphic appearance of a human mucin. \u003cem\u003eGlycobiology\u003c/em\u003e 10, 439\u0026thinsp;\u0026ndash;\u0026thinsp;49 (2000).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, S. et al. Chemoenzymatic modular assembly of O-GalNAc glycans for functional glycomics. Nat Commun 12, 3573 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNaidoo, K.J. FEARCF a multidimensional free energy method for investigating conformational landscapes and chemical reaction mechanisms. Science China Chemistry 54, 1962\u0026ndash;1973 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBruce-Chwatt, T. \u0026amp; Naidoo, K.J. Molecular mechanisms from reaction coordinate graph enabled multidimensional free energies illustrated on water dimer hydrogen bonding. Journal of Computational Chemistry 43, 1802\u0026ndash;1813 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHansen, T. et al. Defining the SN1 Side of Glycosylation Reactions: Stereoselectivity of Glycopyranosyl Cations. ACS Central Science 5, 781\u0026ndash;788 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKono, M. et al. Redefined substrate specificity of ST6GalNAc II: a second candidate sialyl-Tn synthase. Biochem Biophys Res Commun 272, 94\u0026ndash;7 (2000).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMarcos, N.T. et al. Role of the human ST6GalNAc-I and ST6GalNAc-II in the synthesis of the cancer-associated sialyl-Tn antigen. Cancer Res 64, 7050\u0026ndash;7 (2004).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLloyd, K.O., Burchell, J., Kudryashov, V., Yin, B.W. \u0026amp; Taylor-Papadimitriou, J. Comparison of O-linked carbohydrate chains in MUC-1 mucin from normal breast epithelial cell lines and breast carcinoma cell lines. Demonstration of simpler and fewer glycan chains in tumor cells. J Biol Chem 271, 33325\u0026ndash;34 (1996).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYoshimura, Y. et al. Products of Chemoenzymatic Synthesis Representing MUC1 Tandem Repeat Unit with T-, ST- or STn-antigen Revealed Distinct Specificities of Anti-MUC1 Antibodies. Scientific Reports 9, 16641 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePratt, M.R. et al. Deconvoluting the functions of polypeptide N-alpha-acetylgalactosaminyltransferase family members by glycopeptide substrate profiling. Chem Biol 11, 1009\u0026ndash;16 (2004).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ede Las Rivas, M., Lira-Navarrete, E., Gerken, T.A. \u0026amp; Hurtado-Guerrero, R. Polypeptide GalNAc-Ts: from redundancy to specificity. Curr Opin Struct Biol 56, 87\u0026ndash;96 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHassan, H. et al. The lectin domain of UDP-N-acetyl-D-galactosamine: polypeptide N-acetylgalactosaminyltransferase-T4 directs its glycopeptide specificities. J Biol Chem 275, 38197\u0026ndash;205 (2000).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWaterhouse, A. et al. SWISS-MODEL: homology modelling of protein structures and complexes. Nucleic Acids Research 46, W296-W303 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHonorato, R.V. et al. Structural Biology in the Clouds: The WeNMR-EOSC Ecosystem. Frontiers in Molecular Biosciences 8(2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang, J. et al. CHARMM36m: an improved force field for folded and intrinsically disordered proteins. Nature Methods 14, 71\u0026ndash;73 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eField, M.J., Bash, P.A. \u0026amp; Karplus, M. A Combined Quantum Mechanical and Molecular Mechanical Potential for Molecular Dynamics Simulations. J. Comput. Chem. 11, 700\u0026ndash;733 (1990).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCrous, W., Field, M.J. \u0026amp; Naidoo, K.J. Simple Link Atom Saccharide Hybrid (SLASH) Treatment for Glycosidic Bonds at the QM/MM Boundary. Journal of Chemical Theory and Computation 10, 1727\u0026ndash;1738 (2014).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"In vitro synthetic biology, Cancer, MUC1 antigens, systems chemical glycobiology, reaction dynamics","lastPublishedDoi":"10.21203/rs.3.rs-5783651/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5783651/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003ePeptide O-glycosylation relies on the coordinated action of glycosyltransferases (GTs) within the endoplasmic reticulum (ER) and Golgi apparatus. An in vitro one-pot synthetic biology approach was developed to investigate the specificity and kinetics of GT O-GalNAc glycosylation that leads to tumor antigen glycoforms of mucin 1 (MUC1). The focus is to experimentally simulate the divergent glycosylation pathways that lead to the synthesis of cancer-associated antigens (Tn, T) and their sialylated derivatives. First, the biosynthetic details of the defining first step of GALNT relocation from the ER to the Golgi was modeled using the one-pot method. Our findings reveal that an ER enriched with GALNTs results in complete Galnac (Tn) MUC1 site occupancy. This comes about as a function of two processes that are i) extended GALNT reaction time and ii) prevention of inhibition by subsequent glycosylation enzymes such as C1GALT1. The modeling confirms that B3GNT6 has negligible specificity for MUC1 Tn, explaining the absence of core 3 and core 4 structures in MUC1 in both normal and cancerous breast cell lines. Moreover, ST6GALNAC1, and not ST6GALNAC2, is primarily responsible for α-2-6 sialylation of Tn and T antigens. Computer reaction dynamic simulations combined with kinetic experimental analysis show that ST6GALNAC1 prefers fully glycosylated MUC1 and more importantly that its preference is to sialyate the S9 and T13 sites in the SAPDTR motif. This is especially the case when the MUC1 concentration is high (i.e., high-level of expression), suggesting that sTn upregulation on MUC1 in cancer is linked to the occupancy status of S9 and T13 glycosylated sites, that were previously found to be cancer-associated. The results from the one-pot synthesis approach presented here demonstrate its ability to simulate cellular glycosylation within the Golgi-ER. This systems modelling unpacks the molecular details of enzyme localization and substrate glycan occupancy that is fundamental to the regulatory mechanisms that gives rise to tumor-associated MUC1 antigens.\u003c/p\u003e","manuscriptTitle":"An in vitro one-pot synthetic biology approach to simulating diverging Golgi O-glycosylation of tumor-associated MUC1 from normal tissue MUC1","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-01-15 12:33:46","doi":"10.21203/rs.3.rs-5783651/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"8e882efb-502c-4a66-8e54-120ba9eb052b","owner":[],"postedDate":"January 15th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":42694110,"name":"Biological sciences/Chemical biology/Glycobiology"},{"id":42694111,"name":"Biological sciences/Systems biology/Synthetic biology"}],"tags":[],"updatedAt":"2026-04-23T07:09:13+00:00","versionOfRecord":{"articleIdentity":"rs-5783651","link":"https://doi.org/10.1038/s41467-026-72151-y","journal":{"identity":"nature-communications","isVorOnly":false,"title":"Nature Communications"},"publishedOn":"2026-04-22 04:00:00","publishedOnDateReadable":"April 22nd, 2026"},"versionCreatedAt":"2025-01-15 12:33:46","video":"","vorDoi":"10.1038/s41467-026-72151-y","vorDoiUrl":"https://doi.org/10.1038/s41467-026-72151-y","workflowStages":[]},"version":"v1","identity":"rs-5783651","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5783651","identity":"rs-5783651","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}