Biosynthetic production of anticoagulant heparin polysaccharides through metabolic and sulfotransferases engineering strategies

Abstract

Heparin is an important anticoagulant drug, and microbial heparin biosynthesis is a potential alternative to animal-derived heparin production. However, effectively using heparin synthesis enzymes faces challenges, especially with microbial recombinant expression of active heparan sulfate N-deacetylase/N-sulfotransferase. Here, we introduce the monosaccharide N-trifluoroacetylglucosamine into Escherichia coli K5 to facilitate sulfation modification. The Protein Repair One-Stop Service-Focused Rational Iterative Site-specific Mutagenesis (PROSS-FRISM) platform is used to enhance sulfotransferase efficiency, resulting in the engineered NST-M8 enzyme with significantly improved stability (11.32-fold) and activity (2.53-fold) compared to the wild-type N-sulfotransferase. This approach can be applied to engineering various sulfotransferases. The multienzyme cascade reaction enables the production of active heparin from bioengineered heparosan, demonstrating anti-FXa (246.09 IU/mg) and anti-FIIa (48.62 IU/mg) activities. This study offers insights into overcoming challenges in heparin synthesis and modification, paving the way for the future development of animal-free heparins using a cellular system-based semisynthetic strategy.

Fig. 1. Heparin synthesis employing an in vitro/in vivo combination strategy.

The left diagram illustrates the conventional method for -GlcA-GlcNS- heparin preparation. N-acetylglucosamine (GlcNAc) undergoes polymerization with glucuronic acid (GlcA), catalyzed by heparin backbone synthase. Subsequently, the precursor polysaccharide -GlcA-GlcNS- undergoes deacetylation and N-sulfation under the influence of N-deacetylase/N-sulfotransferase (NDST). Expression of active recombinant NDST has historically been a significant bottleneck in achieving this process in prokaryotic expression systems. The right diagram represents the method proposed in this study. GlcNAc analog N-trifluoroacetylglucosamine (GlcNTFA) is taken up by the E. coli K5ASSH strain and utilized for the production of -GlcA-GlcNTFA- polysaccharides, which are then converted to -GlcA-GlcNS- polysaccharides through chemical enzymatic synthesis using an engineered highly stable N-sulfotransferase (NST) mutant (NST-M8). -GlcA-GlcNS- polysaccharides are ultimately modified into anticoagulant polysaccharides by a cascade sulfation system.

Fig. 2. Integrated strategy for N-sulfated polysaccharide synthesis and analysis.

A By disrupting the E. coli kfiA gene to impede heparosan polysaccharide biosynthesis and the glmS gene to halt the UDP-GlcNAc biosynthesis pathway, the non-natural sugar GlcNTFA is incorporated into capsular polysaccharide via the rescue pathway catalyzed by NahK and AGX1. These enzymes convert GlcNTFA into an activated nucleotide sugar, which is polymerized into a -GlcNTFA-GlcA- structured polysaccharide by PmHS2. PmHS2 Pasteurella multocida heparosan synthase 2, KfiA UDP-N-acetyl-d-glucosamine:heparosan α-1,4-N-acetyl-d-glucosaminyltransferase, KfiC UDP-glucuronic acid: heparosan β-1,4-glucuronosyltransferase, GlmS glutamine-fructose-6-phosphate transaminase, Pgm phosphoglucomutase, GalU UTP-glucose-1-phosphate uridylyltransferase, UgdA UDP-glucose 6-dehydrogenase, GlmM phosphoglucosamine mutase, GlmU bifunctional GlcNAc-1-phosphate uridyltransferase/glucosamine-1-phosphate acetyltransferase, NahK N-acetylhexosamine 1-kinase, AGX1 UDP-GlcNAc pyrophosphorylase 1, NST N-sulfotransferase. B Schematic illustration depicting the synthesis of N-sulfated heparin polysaccharides through an in vitro/in vivo combination strategy. C Analysis of disaccharides resulting from LiOH treatment and degradation of bioengineered heparosan (catalyzed by HepIII) using polyamine-based anion exchange (PAMN)-high-performance liquid chromatography (HPLC) with detection at 232 nm. Retention times: ΔUA-GlcNH₂, ~7.0 min; ΔUA-GlcNAc, ~25.5 min; ΔUA-GlcNS, ~30.5 min. Orange: Polysaccharides from E. coli K5 ASSH fed GlcNAc. Red: Polysaccharides from E. coli K5 ASSH fed GlcNAc; the extracted polysaccharides were treated with LiOH. Purple: Polysaccharides from E. coli K5 ASSH fed GlcNTFA. Blue: Polysaccharides from E. coli K5 ASSH fed GlcNTFA; the extracted polysaccharides were treated with LiOH. Black: Polysaccharides from E. coli K5 ASSH fed GlcNTFA; after treating the extracted polysaccharides with LiOH they were reacted with 3-phosphonoadenosine-5-phosphosulfate (PAPS) and NST. Polysaccharides treated with NST and PAPS showed the disappearance of the ΔUA-GlcNH₂ chromatographic peak and the emergence of a ΔUA-GlcNS peak. D–F Electrospray ionization-mass spectra of disaccharides produced by degradation of bioengineering heparosan with HepIII. D ΔUA-GlcNH₂ (m/z 336.05 ± 0.2 Da [M–H]⁻). E ΔUA-GlcNAc (m/z 378.10 ± 0.2 Da [M–H]⁻). F ΔUA-GlcNS (m/z 416.05 ± 0.2 Da [M–H]⁻). Source data are provided as a Source Data file.

Fig. 3. Schematic representation of development pathway of robust N-sulfotransferase.

A In step 1, we predicted 39 hot-spot amino acid sites through molecular dynamics simulation of glycosylation and multiple sequence alignment. Step 2 involved further analysis through PROSS engineering calculations and activity testing, resulting in the stable first-generation mutant NST-M1. In step 3, virtual saturation mutagenesis was applied to the remaining 24 hot-spot sites to exclude any unreasonable sites and mutations, identifying 15 sites for further optimization. Step 4 involved designing simplified degenerate codons for each of the remaining 15 mutation sites, and screening through mutation libraries. The beneficial single mutations obtained were iterated to achieve a robust eighth-generation mutant, NST-M8. B The specific enzyme activity of protein variants achieved through PROSS engineering. NST-design 3, featuring 15 single-point mutations compared with the wild-type enzyme, was validated and used for subsequent mutation rounds. The experiment was conducted with three independent experimental samples (n = 3). Data are presented as mean values ± SD. C The free energy changes obtained from virtual saturation screening. Residues marked in blue on the left-axis indicate hit amino acid sites, while red signifies unsuitable mutation sites. Dashed-red boxes represent the amino acid hits, and refined degenerate codon libraries were constructed to generate the corresponding protein mutants for activity screening. D The stability of protein variants determined through virtual saturation screening. Blue on the left-axis represents hit amino acid sites, while red indicates unconsidered mutations. Enzyme activity was determined by incubating enzyme at 50 °C for 0.5 h, followed by residual enzyme activity measurement at 37 °C. The activity of transferring 1 μmol of sulfate from the donor to the receptor per hour was defined as 1 IU, with units μmol/h. The experiment was conducted with three independent experimental samples (n = 3). Data are presented as mean values ± SD. E The stability enhancement process of mutants in iterative evolution involves obtaining the most stable mutants through the combination of a combinatorial library selection. The optimal mutations from each group were as follows: Group 1: NST-M2 (T872L/T801I); Group 2: NST-M3 (S742K/D843E/A846K); Group 3: NST-M4 (I761L/Y864H); and Group 4: NST-M5 (R688K/N771S). These mutations were further combined to obtain mutants with enhanced stability. NST-M2 and NST-M3 were combined to form NST-M6 (T872L/T801I/S742K/D843E/A846K), while NST-M4 and NST-M5 were combined to form NST-M7 (I761L/Y864H/R688K/N771S). Finally, NST-M6 and NST-M7 were combined to obtain the most stable mutant, NST-M8 (T872L/T801I/S742K/D843E/A846K/I761L/Y864H/R688K/N771S). The experiment was conducted with three independent experimental samples (n = 3). Data are presented as mean values ± SD. Source data are provided as a Source Data file.

Fig. 4. Semi-rational design of a robust N-sulfotransferase mutant.

A The stable structural basis of the designed variant NST-M8 is depicted, with 24 mutated positions (compared with the wild-type enzyme) distributed throughout and represented as spheres. Spheres of different colors signify various reasons for stability enhancement. Yellow spheres indicate sites of stability improvement due to loop rigidity (S637P, S741P, E839P, T869P). Green spheres represent sites of stability improvement due to surface polarity (G625S, S637P, E660D, R688K, V699I, D721N, S738D, S742K, K743E, I761L, A767Y, N771S). Red spheres represent sites of stability improvement due to core packing (D843E, A846K, Y864K, T869P, T872L). Blue spheres represent sites of stability improvement due to helix capping (E839P, L842E). A thumbnail highlights the stabilizing effect of selected mutations. B The changes in enzyme activity and stability from wild-type NST to engineered mutant M1 and then to M8. Compared with the wild-type, M8 exhibited a 2.5-fold increase in activity and an 11.3-fold increase in stability. The experiment was conducted with three independent experimental samples (n = 3). Data are presented as mean values ± SD. C The stability of free and immobilized NST and its variants at 37 °C. NST-WT and NST-M8 was immobilized on ReliSorb SP400 carrier using enzyme surface tethering via a Zbasic2-tag. The experiment was conducted with three independent experimental samples (n = 3). Data are presented as mean values ± SD. D Radar plot showcases multiparameter progress half-life of enzyme at 37 °C (days), enzyme activity (IU/mg), catalytic efficiency(μM/s), substrate concentration(g/L), synthetic efficiency(g/L) from wild-type NST (gray) to NST-M1 (red) and NST-M8 (blue). Source data are provided as a Source Data file.

Fig. 5. Structural analysis and anticoagulant activity assays of bioengineered heparin polysaccharides.

A The synthesis process of K5 engineered heparin. The N-site sulfated polysaccharides prepared are used to synthesize anticoagulant K5 engineered heparin through a multienzyme cascade system involving C5-epi, 2OST, 6OST−1, and 3OST−1. In this process, C5-epi catalyzes the isomerization reaction of GlcA at the C5 position, leading to the formation of IdoA residues. Subsequently, 2OST catalyzes the sulfation reaction at the 2- position of IdoA, while 6OST-1 facilitates the sulfation reaction at the 6- position of GlcNS. 3OST-1 mediates the sulfation reaction at the 3- position of GlcNS6S. B PAMN-HPLC analysis of disaccharides resulting from HepI, HepII, and HepIII degradation of K5EH. The retention times for various disaccharides were: ΔUA-GlcNAc, ~30.5 min; ΔUA-pNP, ~34.5 min; ΔUA-GlcNS, ~36 min; ΔUA-GlcNAc6S, ~38 min; ΔUA2S-GlcNAc, ~41 min; ΔUA-GlcNS6S, ~56 min; ΔUA2S-GlcNS, ~58.5 min; ΔUA2S-GlcNAc6S, ~63 min; ΔUA2S-GlcNS6S-OMe, ~69.5 min; and ΔUA2S-GlcNS6S, ~74 min. The heparins used included: unfractionated heparin (UFH), heparin sulfate (HS), three low-molecular-weight heparins (Enoxaparin, Nadroparin, and Dalteparin), Fondaparinux (chemically synthesized heparin pentasaccharide), Dekaparin (chemoenzymatically synthesized low-molecular-weight heparin), and K5EH (heparin prepared in this study using E. coli strain K5ASSH, NST-M8, and other sulfotransferases). C Detection principle of anticoagulant factor activity. After binding with antithrombin, heparin enhances its association with coagulation factors Xa or IIa. The binding to FXa requires a specific pentasaccharide sequence, while the binding to FIIa requires a glycan chain structure of >18 sugar residues in length. Subsequently, a specific fluorescent substrate is added to detect the remaining coagulation factors (S-2765 for FXa, S2238 for FIIa). The coagulation factors hydrolyze the fluorescent substrate, releasing p-nitrophenol, which is then detected at 405 nm. D Anti-FXa activity assays. FXa activity was determined by the rate of absorbance increase of p-nitrophenol at 405 nm. Each data point represents the average of three determinations. Data are presented as mean values ± SD. E Anti-FIIa activity assays. FIIa activity was determined by the rate of absorbance increase of p-nitrophenol at 405 nm. Each data point represents the average of three determinations. Data are presented as mean values ± SD. C5-epi Homo sapiens heparin C5 epimerase, 2OST Gallus gallus heparin 2-O-sulfotransferase, 6OST-1 Mus musculus heparin 6-O-sulfotransferase isoform 1, 3OST-1 Homo sapiens heparin 3-O-sulfotransferase isoform 1, IdoA Iduronic acid. Source data are provided as a Source Data file.