Using engineered 6-O-sulfotransferase to improve the synthesis of anticoagulant heparin

Abstract

Heparan sulfate (HS) and heparin are sulfated polysaccharides exhibiting diverse physiological functions. HS 6-O-sulfotransferase (6-OST) is a HS biosynthetic enzyme that transfers a sulfo group to the 6-OH position of glucosamine to synthesize HS with desired biological activities. Chemoenzymatic synthesis is a widely adopted method to obtain HS oligosaccharides to support biological studies. However, this method is unable to synthesize all possible structures due to the specificity of natural enzymes. Here, we report the use of an engineered 6-OST to achieve fine control of the 6-O-sulfation. Unlike wild type enzyme, the engineered 6-OST only sulfates the non-reducing end glucosamine residue. Utilizing the engineered enzyme and wild type enzyme, we successfully completed the synthesis of five hexasaccharides and one octasaccharide differing in 6-O-sulfation patterns. We also identified a hexasaccharide construct as a new anticoagulant drug candidate. Our results demonstrate the feasibility of using an engineered HS biosynthetic enzyme to prepare HS-based therapeutics.

Graphical Abstract

Heparan sulfate (HS) is a sulfated polysaccharide found on the surface of mammalian cells and in the extracellular matrix. HS displays a wide range of biological functions including inflammatory responses, blood coagulation, angiogenesis, embryonic development and viral/bacterial infection^1,2. HS is also involved in the initial attachment/entry of SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2) to the host, and potentially attenuates lung damage in COVID-19 (coronavirus disease 2019) patients³. HS polysaccharide chains consist of a disaccharide repeating unit of glucuronic acid (GlcA) or iduronic acid (IdoA) and glucosamine (GlcN). Both IdoA and GlcN saccharide units carry sulfo groups at the 2-OH position of IdoA and at N-, 3-OH and 6-OH positions of GlcN. Specific sulfation patterns and saccharide sequences dictate the biological activity of HS polysaccharides⁴. HS and heparin isolated from biological sources are a heterogenous mixture of polysaccharides with different sugar chain lengths and sulfation patterns, complicating the efforts to investigate the structure/function relationship for HS.

Lack of HS molecules with uniform length and sulfation patterns has been a major impediment for HS-related research. Although the synthesis of HS oligosaccharides using a purely chemical approach is possible and has made progress in recent years^5–8, it is a labor intensive and costly process, especially to make longer than octasaccharides in large quantity to support biological studies. Chemoenzymatic synthesis of HS recently emerged as an alternative. This method utilizes a series of HS biosynthetic enzymes mimicking the biosynthesis of HS in vivo⁹. Several HS sulfotransferases are needed to synthesize HS oligosaccharides using the chemoenzymatic method. These enzymes include N-sulfotransferase, 2-O-sulfotransferase, 6-O-sulfotransferase (6-OST) and 3-O-sulfotransferase¹⁰. One can control the site of N-sulfation of GlcN and 2-O-sulfation of IdoA using an unnatural UDP-sugar nucleotide and following a special sequence of enzymatic modification steps. However, the control of 6-O-sulfation using the 6-OST enzyme has been incomplete^2,11. So far, there is no effective method to allow installation of 6-O-sulfation on the non-reducing end, while leaving the reducing end glucosamine residues unsulfated.

There is an urgent public healthcare need for preparing a safe high quality synthetic alternative for animal-sourced heparin, a common anticoagulant drug used to treat blood clotting disorders. Heparin and HS can have the same disaccharide repeating units, but heparin has a higher degree of sulfations. Chemoenzymatic synthesis can be used to make synthetic heparin cost-effectively¹². Harvested from porcine intestine, the production of heparin requires a long but poorly regulated supply chain, which is a serious concern for regulatory agencies. Batches of contaminated heparin that entered the worldwide market in late 2007 were responsible for many deaths, making this one of the worst incidents of a toxic drug product reaching the market in US history, underscoring the drawbacks of animal-sourced heparin¹³. Fondaparinux, a fully synthetic, is an alternative to animal-sourced heparin¹⁴, but its production cost is substantially higher than low-molecular weight heparins. A less expensive synthetic substitute to fondaparinux would be welcomed by healthcare providers. The 6-O-sulfation is one of four suflation steps in the chemoenzymatic synthesis of anticoagulant heparin. The investigation of biochemistry 6-OST contributes to the effort to design better anticoagulant heparin constructs.

Here, we report the design of 6-OST mutant enzyme engineered to achieve fine control of the 6-O-sulfation site for HS synthesis using the chemoenzymatic method. The mutant 6-OST enzyme, designated as 6-OST Mt-4, was obtained based on the crystal structure of 6-OST-3 from zebra fish¹⁵. Results from substrate specificity analysis revealed that 6-OST Mt-4 sulfates only the non-reducing end glucosamine residue. We used this mutant to prepare an octasaccharide that carries two 6-O-sulfated glucosamine units and a library of five hexasaccharides. The library of hexasaccharides was used to investigate the contribution of 6-O-sulfation to the anticoagulant activity. A hexasaccharide was identified as a new anticoagulant drug candidate, the anticoagulant activity of which proved to be reversible by andexanet, reducing the risk of bleeding side effect. In addition, the synthesis of this hexasaccharide can be carried out at higher concentration suitable for a scalable synthesis. Taken together, the results from our study demonstrate for the first time the use of an engineered HS biosynthetic enzyme to expand the chemoenzymatic synthesis approach for preparing new heparin-based therapeutics and long-sought after research reagents.

Results

6-OST has a positively charged pocket that binds to the GlcN unit located at the non-reducing end side of the acceptor site.

To achieve the control of 6-O-sulfation, we conducted site-directed mutagenesis based on the crystal structure of 6-OST. Examining the crystal structure of the ternary complex of 6-OST-3/PAP/heptasaccharide (PDB: 5T0A)¹⁵, we identified a positively charged pocket that is located on the non-reducing end side of the acceptor/catalytic site. The pocket appears capable of accommodating a glucosamine unit beyond the non-reducing end of the heptasaccharide substrate in the crystal structure. We extended the heptasaccharide to an octasaccharide by manually modeling a N-sulfo glucosamine (GlcNS) unit (Fig 1A). The modeled GlcNS unit is positioned along the binding pocket lined with positively charged residues, Arg112, Arg206 and Arg329 (Fig 1A). This placement suggests that Arg112 could interact with an N-sulfo group, while Arg 206 and 329 are in position to interact with a 6S moiety (Fig 1A). We prepared four 6-OST mutants to replace the arginine residues with negatively charged glutamate residues. One of the mutants contains a single arginine-residue substitution (6-OST Mt-1; R112E), two contain double arginine-residue substitutions (6-OST Mt-2; R112E/R206E and 6-OST Mt-3; R112E/R329E), and one contains a triple arginine-residue substitution (6-OST Mt-4; R112E/R206E/329E). Substitutions of arginines with glutamates would presumably preclude binding of GlcNS or GlcNS6S to this region due to electrostatic and/or potential steric clashes between the negatively charged sulfates and glutamate residues (Fig 1B). Based on the position of the glutamate mutations from manually modeling, they should not alter the acceptor GlcNS binding site, thus maintaining sulfotransferase activity for the non-reducing terminal GlcNS (Fig1B). Results from activity measurement using a structurally heterogeneous polysaccharide substrate revealed that all four mutants possessed sulfotransferase activity (Table 1).

Fig 1. Interaction of 6-OST with an octasaccharide substrate.

Panel A) shows an enlarged surface rendering of the active site from the crystal structure (pdb 5T0A) of zf6OST isoform 3 with heptasaccahride (cyan) bound, with an additional glucosamine modeled on the non-reducing end. The position of the acceptor 6-OH is marked with a green asterisk. Arginine residues surrounding the modeled non-reducing end GlcNS6S are labeled. The positive charge of the binding pocket at the non-reducing end supports binding of the negatively charged octasaccharide (GlcNS6S-GlcA-GlcNS-GlcA-GlcNS-IdoA2S-GlcNS-GlcA-pNP). The proposed interactions of three arginine residues to the non-reducing end GlcNS6S are depicted on the top. Only three saccharide units from the crystal structure (black) and one modeled unit (red) are displayed. Panel B) shows electrostatic surface of the modeled 6-OST Mt-4 (from pdb 5T0A) with mutated glutamates (labeled green) surrounding the non-reducing end of the octasaccharide. The mutations result in a greatly reduced/inverted charge on the surface for the non-reducing end of the binding pocket as depicted on the top.

Table 1.

Sulfotransferase activities for 6-OST wild-type and mutant enzymes

Name	Mutation sites	Sulfotransferase activity (pmoles/μg/h)*
6-OST WT	Not applicable	72.3
6-OST Mt-1	6-OST R112E	40.2
6-OST Mt-2	6-OST R112E/R206E	27.8
6-OST Mt-3	6OST R112E/R329E	73.5
6-OST Mt-4	6-OST R112E/R206E/R329E	12.6

^*Sulfotransferase activity was determined by measuring the amount of ³⁵S-labeled sulfate transferred from [³⁵S]PAPS to completely desulfated/N-sulfated (CDSNS) heparin, a structurally heterogeneous polysaccharide substrate.

6-OST Mt-4 preferably sulfates the non-reducing end N-sulfo glucosamine (GlcNS) unit

The substrate specificities of 6-OST mutants were evaluated using structurally homogeneous oligosaccharide substrates. For example, we examined the 6-O-sulfation of a hexasaccharide substrate with three possible sulfation sites (comp 1, Fig 2) modified by wild type or mutant enzymes. The products were analyzed by strong anion exchange high performance liquid chromatography (SAX-HPLC), a high-resolution method to resolve all seven products. The products include three mono-6-O-sulfated hexasaccharides (m-6S), three di-6-O-sulfated hexasaccharides (d-6S) and one tri-6-O-sulfated hexasaccharide (t-6S). The HPLC chromatograms of comp 1 modified by the 6-OST variants are distinct (Fig 3, Fig S1 and Table S1), suggesting that wild type enzyme and Mt-2 and Mt-4 generated different 6-O-sulfated products. After 7-h incubation, the mono-6-O-sulfated product GlcNS6S-GlcA-GlcNS-IdoA2S-GlcNS-GlcA-pNP that eluted at 28.0 min from SAX-HPLC was observed as the major peak when comp 1 was incubated with 6-OST Mt-4, but different SAX-HPLC profiles were observed when the same substrate was incubated with 6-OST WT and 6-OST Mt-2 (Fig 3 and Table S1). The structure of the mono-6-sulfated product was confirmed by the disaccharide analysis method using liquid chromatography coupled mass spectrometry (LC-MS, Figs. S2 and S4), confirming that the 6-O-sulfation is located at the F unit of comp 1 (Fig. 2 and Figs. S3)¹⁶. This specific 6-O-sulfated product represented only 5% of total comp 1 modified by 6-OST WT; whereas this product represented 21.9% and 74.4% for 6-OST Mt-2 and Mt-4 modified comp 1, respectively. Notably, the mono-6-O-sulfated hexasaccharide was generated nearly to homogeneity by 6-OST Mt-4 after 50-h incubation (Fig 3). Under the same condition, comp 1 was completely converted to the tri-6-O-sulfated product by 6-OST-WT where the B, D and F units were all 6-O-sulfated (Fig 3 and Table S1). The results suggest that 6-OST Mt-4 sulfates only the nonreducing end GlcNS unit (F unit), having lost its ability to sulfate internal and reducing-end GlcNS (B or D units of comp 1). Our results, therefore, demonstrate that the location of the three arginine residues (Arg-112, Arg-206 and Arg-329) play a critical role in installing multiple 6-O-sulfation sites.

Fig 2. Chemical structures of key oligosaccharides synthesized in this study.

Eight oligosaccharides were synthesized for this study. Comp 1 was used as a substrate to probe the substrate selectivity of different 6-OST mutants. The synthesis and full characterization of comp 1 and comp 8 were reported previously^18,23. The site of 6-O-sulfation by 6-OST Mt-4 is circled for clarity. The synthesis of comp 2 to 7 have not been previously reported. Fondaparinux was purchased from Cardinal Health.

Fig 3. SAX-HPLC chromatograms of comp 1 modified by 6-OST WT, 6-OST Mt-2 and 6-OST Mt-4.

Comp 1 was incubated with equal amounts 6-OST wild type and mutants proteins based on the sulfotransferase activity, and the same amount of each enzyme was added after 24-h incubation independently. The reaction mixture by SAX-HPLC was analyzed after 7-h incubation and 50-h incubation. Elution positions of comp 1, mono-6-O-sulfated hexasaccharides (m-6S), di-6-O-sulfated hexasaccharide (d-6S), and tri-6-O-sulfated hexasaccharide (t-6S) are indicated in the upper-left chromatogram. * indicates the elution position of the 6-O-sulfated hexasaccharide product with a structure of GlcNS6S-GlcA-GlcNS-IdoA2S-GlcNS-GlcA-pNP. The structure of this hexasaccharide was confirmed by disaccharide analysis using LC/MS (Supplementary Figs S1 and S2).

We also investigated the substrate specificities of 6-OST mutants and wild type enzymes towards three additional oligosaccharide substrates. These substrates included GlcNS-GlcA-GlcNS-GlcA-GlcNS-GlcA-pNP (GlcA-containing hexasaccharide), GlcA-GlcNS-IdoA2S-GlcNS-GlcA-pNP (IdoA2S-containing pentasaccharide), and GlcA-GlcNS-GlcA-GlcNS-GlcA-pNP (GlcA-containing pentasaccharide) (Tables S2–S4 and Fig S5–S9). For all three oligosaccharides, the conclusion from the SAX-HPLC analysis was essentially the same as those from using comp 1 as a substrate. Namely, 6-OST Mt-4 only sulfates the GlcNS unit located at the nonreducing end even when a terminal GlcA was present. The structures of the mono-6-sulfated intermediate products of these three oligosaccharide substrates were confirmed by a disaccharide analysis and NMR methods.

Control of 6-O-sulfation using 6-OST Mt-4 and 6-OST WT

We next used both 6-OST Mt-4 and 6-OST WT to control the 6-O-sulfation site during the synthesis (Fig 4). First, we used 6-OST Mt-4 to synthesize comp 2, an octasaccharide that contains two GlcNS6S units at the non-reducing end (Fig 4A). The data suggest that multiple rounds of saccharide addition and sulfation of GlcNS6S units on the non-reducing end is feasible using 6-OST Mt-4 and the presence of a terminal GlcA does not disrupt the sulfation of this GlcNS. Second, we used the combination 6-OST Mt-4 and 6-OST WT to synthesize six different hexasaccharides (Fig 4B and Fig S10). Those hexasaccharides were then further modified by 3-O-sulfotransferase 1 to yield comp 3 through 8, which contain the 3-O-sulfation that is critically important for anticoagulant activity¹⁷. The purity of each synthesized compound was confirmed by SAX-HPLC, and the chemical structure was characterized by electrospray ionization mass spectrometry (ESI-MS) (Table S5 and Figs S11–S12) and NMR (Table S6 and Figs S13–S33). The results clearly demonstrate that strategic use of 6-OST Mt-4 or 6-OST WT enzymes offers an effective method to control the location of the 6-O-sulfation along the oligosaccharide.

Fig 4. Synthetic schemes of different oligosaccharides using 6-OST Mt-4 and 6-OST WT.

Panel A shows the synthesis of comp 2, a partially 6-O-sulfated octasaccharide that contains two GlcNS6S units at the nonreducing end. The 6-O-sulfation step was completed using 6-OST Mt-4. Panel B shows the schemes to synthesize four hexasaccharides, including comp 3, comp 4, comp 5 and comp 8. Depending on the 6-O-sulfation patterns in the final products, both 6-OST WT and 6-OST Mt-4 were used at the 6-O-sulfation step as shown. To maintain clarity, only shorthand symbols are shown in the figures, and the synthetic schemes for comp 6 and comp 7 are shown in Supplementary Fig. S3. pmHS2 represents (heparosan synthase 2 from pasteurella multocida; GlcNTFA represents trifluoroacetylated glucosamine; and 3-OST-1 represents 3-O-sulfotransferase 1.

6-O-sulfation of nonreducing end GlcNS unit is essential for the anticoagulant activity

The successful synthesis of comp 3 through 8 offered an opportunity to study the contribution of 6-O-sulfation to the anticoagulant activity. Fondaparinux, an FDA (US Food and Drug Administration) approved anticoagulant drug, was used as a positive control in the study. To this end, we measured the potency of factor Xa inhibition by these compounds, or anti-FXa activity (Fig 5A)¹⁸. We also measured the binding affinity (K_d) between the hexasaccharides and antithrombin (Fig 5A). Our results revealed that comp 4, 5, 7 and 8 have comparable anti-FXa activity as well as display similar binding affinity to antithrombin (Fig 5A) to those of fondaparinux. A representative inhibition curve of the activity of FXa by comp 4 and fondaparinux is shown in Fig 5B. Notably, neither comp 3 nor comp 6 inhibited the activity of FXa or bound to antithrombin, suggesting that both compounds are non-anticoagulant hexasaccharides. Comparing the structures of comp 3 and 6 with others, we discovered that both hexasaccharides lack 6-O-sulfation at the non-reducing end GlcNS unit. This finding is consistent with the previously published conclusion that this particular 6-O-sulfo group is essential for the anticoagulant activity^14,17,19.

Fig 5. Evaluation of the anticoagulant property of comp 4 and other synthesized hexasaccharides.

Panel A shows the anti-FXa activity of comp 3 to comp 8 and fondaparinux and their binding affinities to antithrombin. Panel B shows representative inhibition curves for the anti-FXa experiment. Panel C shows the comparison of conversion rates to 6-O-sulfated intermediates from comp 1 under different concentrations by 6-OST Mt-4 (47 U mL⁻¹) and 6-OST WT (51 U mL⁻¹). Incubation comp 1 with 6-OST Mt-4 yielded mono-6-S intermediate, whereas incubation comp 1 with 6-OST WT yielded tri-6-S intermediate (Fig 4B). The percentage of the 6-O-sulfated intermediates was calculated based on the SAX-HPLC analysis of the reaction after 2-h incubation. Panel D shows the reversibility of the anti-FXa activity of fondaparinux and comp 4 by andexanet. The experiment was carried out in an in vitro assay.

Comp 4 is amenable for large-scale production as an anticoagulant drug candidate

We compared the efficiency for the synthesis of 6-O-sulfated intermediates to make comp 4 and comp 8, respectively. Previously, we reported the synthesis of comp 8 as an anticoagulant drug candidate¹⁸. The synthesis of comp 4 and comp 8 is completed from a common starting hexasaccharide (comp 1) using 6-OST Mt-4 and 6-OST WT, respectively, followed by 3-O-sulfation (Fig 4B). To this end, we compared the efficiency for the synthesis of 6-O-sulfated hexasaccharide intermediates and mono-6-S and tri-6-S intermediates. The conversion to mono-6-S intermediate using 6-OST Mt-4 reached 88.1% completion in two hours at 1.5 mM of comp 1; however, only 14.9 % comp 1 was converted to tri-6-S intermediate using 6-OST WT under the same substrate concentration (Fig 5C). Extending the incubation time to 14 hours improved the conversion to tri-6-S intermediate, but the conversion was still substantially lower than mono-6-S intermediate (Fig S34). Results from kinetic analysis of 6OST-WT and 6OST-Mt-4 revealed that mutant protein synthesizes mono-6-S intermediate 3-fold more efficiently than the wild type protein synthesizes tri-6-S intermediate (Fig S35). Both mono-6-S intermediate and tri-6-S intermediate were then converted to comp 4 and comp 8, respectively, by 3-O-sulfation with similar efficiency. The results suggest that enzymatic conversion of comp 1 to comp 4 is more efficient than to comp 8. Indeed, using 6-OST Mt-4, we prepared 848 mg comp 4, demonstrating the scalability of the synthesis.

We examined the reversibility of comp 4 in vitro using andexanet, an FDA approved drug to reverse uncontrolled bleeding effect caused by direct FXa inhibitors^20,21. Andexanet was previously demonstrated to neutralize fondaparinux²². Similar to fondaparinux, we found that the anti-FXa activity of comp 4 was fully reversed by andexanet (Fig 5D). Because comp 4 is a short oligosaccharide, we found that the anti-FXa effect from comp 4 was not reversed by protamine, a polypeptide that reverses the anticoagulant activity of unfractionated heparin (Supplementary Fig S36). The reversibility by andexanet reduces the bleeding risk of comp 4.

Discussion

Here, we demonstrated use of an engineered 6-OST enzyme to achieve the control of 6-O-sulfation in HS oligosaccharide synthesis. Increasing evidence indicates that chemoenzymatic synthesis of HS oligosaccharide is a promising alternative to purely chemical synthesis, especially in synthesizing oligosaccharides longer than octasaccharides^9,12,23,24. The key reagents in the chemoenzymatic approach are recombinant HS biosynthetic enzymes. Many of the enzymes exhibit extremely high selectivity to the substrates required for generating specific products, making the method efficient and powerful. However, the high substrate specificity of the enzymes imposes limits on the numbers and types of structures that can be synthesized. Yet more than one hundred oligosaccharides have been synthesized and are commercially available (www.glycantherapeutics.com). However, all these oligosaccharides are prepared within the boundary of the natural substrate specificities of wild type HS biosynthetic enzymes. The next challenge for the chemoenzymatic synthesis is to expand the diversity of existing HS oligosaccharide libraries. Obtaining enzymes with additional sulfation ability/selectivity that is beyond the capability of wild type proteins is highly desirable. In this manuscript, we demonstrate that an engineered 6-OST mutant exhibits a substrate specificity distinct from the wild type enzyme. The mutant enzyme, 6-OST Mt-4, expresses well in E. coli, and the amount of the enzyme is suitable for large-scale synthesis. The mutant enzyme was used to specifically sulfate the non-reducing end glucosamine unit, enhancing control of 6-O-sulfation. We also used 6-OST Mt-4 to synthesize and characterize a potential new ultra-low molecular weight heparin construct. This hexasaccharide is tailored to contain only the 6-O-sulfo group that is critical for antithrombin binding based on the crystal structure of a pentasaccharide analog with antithrombin (Fig S37)²⁵.

The unique substrate specificity of 6-OST Mt-4 provides insight into how the enzyme recognizes the acceptor/sulfation site using the positively charged binding pocket. A crystal structure of zebra fish 6-OST-3 was solved with a bound heptasaccharide positioned with the non-reducing end GlcNS at the active site.¹⁵ Modeling of an additional GlcNS to the terminal GlcA of the heptasaccharide suggests this unit interacts with a binding pocket consisting of residues Arg329 and Arg206 on one side and Arg112 on the other side of the pocket when an internal or reducing end glucosamine is situated at the acceptor position. Replacement of glutamates at these positions likely excludes the binding of negatively charged glucosamines on the non-reducing end side of the acceptor glucosamine as in the case of 6-OST Mt-4. Arg329, Arg206 and Arg112 are conserved in all three isoforms of 6OST in human, mouse, zebrafish, chicken, and frog, with the structurally equivalent residues to R112 and R329 also being conserved in the two isoforms from Drosophila (Table S7). These findings suggest that using a positively charged substrate binding pocket to accommodate the negative charge from non-reducing end saccharides is a generic mechanism among 6-OSTs to achieve consecutive 6-O-sulfation.

The discovery of comp 4 offers a new drug candidate for anticoagulant ultra-low molecular weight heparin. Compared to comp 8, it has fewer number of sulfo groups, which makes the 6-O-sulfation step easier, with shorter reaction times required for completion and a lower concentration of PAPS (3’-phosphoadenosine 5’-phosphosulfate) during the 6-O-sulfation step. A lower PAPS concentration at the 6-O-sulfation step also reduces product inhibition of the reaction. During the 6-O-sulfation reaction, 6-OST transfers the sulfo group from PAPS to the oligosaccharide substrate, and transforms PAPS to PAP (3’-phosphoadenosine 5’-phosphate), an inhibitor of 6-OST²⁶. To prepare comp 8, three PAPS molecules are needed to prepare one molecule of the tri-6S intermediate, which is then converted to comp 8. To minimize the inhibition effect in the 6-O-sulfation step, only low concentrations of comp 1 (< 1 mM) have been used to prepare tri-6S intermediate, which will limit the size for industrial scale synthesis. To prepare comp 4, only one molecule of PAPS is needed to prepare the mono-6S intermediate, suggesting that 6-O-sulfation can be performed under a higher concentration of substrate, and reduced PAP inhibition. Indeed, we discovered that the conversion of mono-6-S intermediate (for making comp 4) is 6-fold more effective than that of tri-6-S intermediate (for making comp 8) at 3 mM substrate concentration under 2-h incubation time, and greater than 2-fold efficiency under 14-h incubation time.

In summary, using engineered enzymes to prepare biologically active compounds is a fast-growing research field with a wide range of applications²⁷, but potential in HS synthesis has not yet been reported. Although previous reports showed that mutations in HS 2-O-sulfotransferase and 3-O-sulfotransferase can alter the substrate specificities^28,29, the substrate specificities from these mutants diminish under a high substrate concentration. Therefore, the mutants currently have limited functionality in the synthesis of HS oligosaccharides. The findings from our study on 6-OST mutants clearly demonstrate the feasibility to expand the chemoenzymatic synthesis of HS oligosaccharides for the investigation of their biological functions and designing heparin-based therapeutics.