Close this search box.

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

Synthesis of Ac4GlcNTFA

1,3,4,6-tetra-O-acetyl-β-D-glucosamine hydrochloride (2.04 g; Samuel, Jinan, China) was mixed with anhydrous sodium carbonate (1.00 g) and dissolved in 10 mL of anhydrous methanol; 3.5 mL of ethyl trifluoroacetate (Sigma) was added. The mixture was stirred at room temperature for 12 h. The residue was purified by flash column chromatography (gradient cyclohexane/ethyl acetate v/v 3:1–1:1) to obtain 2.32 g of product with a yield of 92%.

Preparation of E. coli strain K5 polysaccharide analog

The K5ASSH strain32 of E. coli was cultured with chloramphenicol (34 μg/mL), kanamycin (50 μg/mL), and GlcNAc (100 μg/mL). Cells were activated in liquid LB medium at 225 rpm and 37 °C for 12 h. After consuming GlcNAc for 1 h, Ac4GlcNTFA (500 μg/mL) was added to the culture medium and cultivation was continued at 37 °C and 225 rpm until OD600 (the optical density at 600 nm) reached 0.6. At that time, IPTG was added to the culture medium (to 0.2 mM). Cultivation was continued at 22 °C for 12 h. After centrifugation to remove bacteria, the 3-times volume of ethanol was added to the supernatant, which was precipitated overnight at 4 °C. The crude polysaccharide obtained was purified by using a Bio-Gel P-10 column to yield the K5 polysaccharide analog.

Analysis of K5 polysaccharide analog

After treating 1 mg/mL of K5 polysaccharide analog in 0.1 M LiOH for 1 h, 50 mM MES (4-morpholine ethanesulfonic acid) was added, and the pH was adjusted to approximately 7.0. Subsequently, 0.2 mg/mL NST and 2 mg/mL PAPS were added, and reaction was carried out at 37 °C for 12 h before purification of polysaccharide using a DEAE column. This reaction converted all -GlcNTFA-GlcA- into -GlcNS-GlcA-. Subsequently, 0.1 mg/mL Hep III64 and 5 mM CaCl2 were added to the reaction overnight, and the product was analyzed by PAMN-HPLC. The lysed disaccharide product was purified by DEAE column chromatography and desalinated by using a Bio-Gel P-2 column before being subjected to HPLC-MS analysis.

MD simulations

The crystal structure of N-sulfotransferase (PDB: 1NST) was meticulously refined using Schrödinger. Additionally, the GlcNAc2Man3 pentasaccharide structure, extracted from the human C5-epimerase crystal structure (PDB: 6HZZ49), was carefully grafted onto the N-linked site (residue N667) of NST. This procedure enabled the simulation of the glycosylated form of the wild-type N-sulfotransferase (NST) structure. We conducted Molecular Dynamics (MD) simulations using the advanced Amber software suite. The Amber ff14SB force field was specifically employed for the N-sulfonyltransferase, in both its glycosylated and non-glycosylated states65. The respective force field parameters for the ligand were generated using the Amber antechamber module in collaboration with the General Amber Force Field (GAFF)66. We immersed the protein structure in a truncated octahedral box, which extends 8 Å beyond the solute, and used the TIP3P water model while applying periodic boundary conditions. The MD simulations, lasting for 20 ns, were performed under the NPT (Number of Particles, Pressure, Temperature) condition. All MD results were subsequently analyzed using the ptraj module, an indispensable part of the Amber software suite. The variation in root-mean-square fluctuation (RMSF) value due to the presence or absence of glycosylation was calculated as ΔRMSF.

Multiple sequence alignment

A search of the amino acid sequence HMM (Hidden Markov Model) was conducted using the HMMER webpage ( The alignment was subsequently employed to construct a HMM profile, courtesy of the “hmmbuild” function. For the purpose of visualization, the HMM logo was forged via the Skylign online tool67 ( The frequencies of each letter in the designated target position were calculated, founded on the relative entropy or Kullback-Leibler distance.

PROSS engineering of wild-type NST

In identifying candidate amino acid residues for mutagenesis, to ensure reliability while covering a comprehensive array of residues, a Δroot-mean-square fluctuation value of <0.6 in the MD simulations before and after glycosylation, along with a score of <8 for the residue in sequence conservation analysis (values range from 0 to 10, with 10 indicating complete amino acid conservation at that position) were applied as thresholds. A total of 39 amino acids were hit: G625, S637, E660, T669, D682, V685, R688, A691, K698, V699, T701, N705, D721, S738, S741, S742, K743, I761, A767, Y768, H769, A770, N771, Q772, K779, E784, K787, M791, T801, H805, T807, G823, E839, L842, D843, A846, Y864, T869, and T872.We used the Protein Repair One Stop Shop (PROSS)42. server, a bioinformatics strategy, to design a series of mutants with increasing surface polarity. The crystal structure of NST (1NST) served as the template for the PROSS algorithm. It is worth noting that only above 39 specific sites were targeted for engineering, which were identified through dynamic simulations and multiple sequence alignment. The algorithm suggested potential amino acid substitutions based on sequence homologs that shared at least 75% similarity with the template 1NST. None of these substitutions was located within the active or substrate-binding sites of the enzyme. From the suggested mutants, we selected four for expression named NST-design 1-4.

Virtual saturation mutation calculation

After activity testing, the most stable NST-design 3 (NST-M1) include 15 mutations: G625S, S637P, E660D, V699I, D721N, S738D, S741P, K743E, A767Y, K779Q, E784D, T807H, E839P, L842E, and T869P. Out of the original 39 candidate sites, there are 24 remaining sites: T669, D682, V685, R688, A691, K698, T701, N705, S742, I761, Y768, H769, A770, N771, Q772, K787, M791, T801, H805, G823, D843, A846, Y864, and T872. Based on the crystal structure of NST (PDB ID: 1NST), energy calculations with Rosetta_ddg47 were performed; these 24 sites were mutated in silico to all proteinogenic amino acids.

Preparation of point mutant library using one-step PCR method

Construction of the point mutant library was carried out using the Vazyme Mut Express II Fast Mutagenesis Kit V2, and specific operations can be found in the instruction manual. In short, PCR amplification of the NST-M1-encoding plasmid was performed using specific primers with simplified codons. The selection of simplified codons is shown in Fig. 3C. The amplified product was digested by DpnI and cyclized by ClonExpress recombination, and then directly transformed into DH5a to complete site-specific mutagenesis.

Preparation of combinatorial mutation library through gene recombination

The Mut Express MultiS Fast Mutagenesis Kit V2 (Vazyme, Nanjing, China) was used for the construction, and specific operations can be found in the instruction manual. In brief, the template plasmid (pGEX-4T-1-NST-M1) was divided into three segments based on the target mutation sites. Partially complementary primer mixtures (including mutant and wild-type genotypes) were designed for each target mutation site. Using the template plasmid (pGEX-4T-1-NST-M1) as a template, amplification and recovery were performed separately for each segment. After digestion with DpnI, the amplified products were subjected to recombination and ligation. The recombined products were directly transformed into DH5a to complete the construction of the mutant library. All primers used for this step can be found in Supplementary Data 1.

General methods for purification of NST proteins

The NST (UniProt P52848.1, residues L557–R882) and its variants were genetically engineered into vector pGEX-4T-161 (GE Healthcare), before being chemically introduced into E. coli Origami B (DE3) cells (Novagen). The transformed cells were incubated with shaking at 225 rpm at 37 °C for 4 h, until OD600 reached approximately 0.8; then, IPTG was added to a final concentration of 0.2 mM. Protein expression was induced at 22 °C and 225 rpm for 16–18 h. The NST was purified using a glutathione S-transferase column supplied by Sangon Biotech (Shanghai, China).

General method for determining NST stability by PAMN-HPLC

The experiment involved the use of deacetylated heparin pentasaccharide GlcA-GlcNH2-GlcA-GlcNH2-GlcA-pNP12 as the acceptor substrate and PAPS as the donor substrate. Initially, the enzyme was subjected to heat treatment at 50 °C for 0.5 h. The reaction system was comprised of 50 mM MES (pH 7.0), 0.2 mM pentasaccharide acceptor substrate, 0.6 mM PAPS, and 0.05 mg/mL protein. The reaction was conducted in a water bath, maintained at 37 °C, for 0.5 h, followed by inactivation of the enzyme by boiling for 5 min. The reaction solution was filtered using a 0.22-μm membrane and subjected to PAMN-HPLC detection. The pNP group of the monosaccharide acceptor showed specific absorption at 310 nm. One IU of enzymatic activity was defined as the amount that transferred 1 μmol of sulfate to the acceptor substrate per hour (unit, μmol/h).

General method for determining NST stability using microporous plates

Synthesis of primers for mutation was performed by Genscript (Nanjing, China), and the mutation library was constructed using the above methods. Recombinant plasmids were chemically transformed into E. coli Origami B (DE3) cells. A single colony was picked and placed in 1 mL of sterilized LB medium containing 50 μg/mL carbenicillin. This was then cultured at 37 °C and 225 rpm. The overnight culture broth was transferred to 2 mL of fresh LB medium (2% v/v inoculum); the newly inoculated culture was incubated at 37 °C, 225 rpm, for approximately 4 h until OD600 reached approximately 0.8. Then, 0.2 mM IPTG was added, and protein expression was induced at 22 °C and 225 rpm for 16–18 h. The bacterial cells were collected by centrifugation, washed twice with phosphate-buffered saline (PBS), and subjected to two cycles of freezing and thawing. The cellular material was then resuspended in PBS that contained 0.1 mg/mL lysozyme and incubated at 37 °C for 1 h. After centrifugation, the supernatant was collected. A standard reaction system was set up that included 50 mM MES (pH 7.0), 0.2 mM pentasaccharide receptor substrate, 0.6 mM PAPS, 2 mM pNPS, 0.2 mg/mL ASTIV, and 5 μL supernatant. The absorption at 410 nm was then measured at 37 °C for 90 min. The screening results of all mutants can be found in Supplementary Data 2.

Enzyme immobilization

A total of 60 mg ReliSorb SP400 carrier (Biokal, Netherlands) was weighed out and put in an Eppendorf tube. The carrier was washed three times with water, followed by two washes with PBS. Approximately 1 mL of purified NST-WT or NST-M8 protein was added and incubated at room temperature with agitation at 40 rpm for 2 h. The supernatant was removed and the carrier was washed with PBS. The enzyme fixation was analyzed by SDS-PAGE. NST or immobilized NST was placed at 37 °C for 1–9 days; enzyme activity determination was performed according to the general method for determination of NST activity by PAMN-HPLC.

Differential scanning fluorimetry determination of protein melting temperature

The thermal denaturation assay using differential scanning fluorimetry (DSF) was conducted for NST-WT and NST-M8. Protein samples were diluted to a final concentration of 5 µM in a 20 mM buffer with varying pH values and 150 mM NaCl. DSF data was collected using a CFX96 RT-PCR instrument (Bio-Rad) with the “HEX” channel for fluorescence excitation and emission. The protein samples were dispensed into a 96-well Frame Star PCR plate and covered with a clear thermal-seal film to prevent evaporation. The temperature was increased in increments of 0.5 °C with a 5-second equilibration hold at each temperature. As the temperature increased, the protein unfolded, exposing its hydrophobic core to the solvent. The hydrophobic sites were bound by SYPRO Orange dye (Sigma), leading to an increase in fluorescence. Fluorescence intensity was monitored at 570 nm, allowing for simultaneous and independent readings of all 96 wells. The fluorescence intensity data was fitted to a Boltzmann sigmoidal curve using Prism software to determine the melting temperatures of NST-WT and NST-M8.

Preparation of bioengineered heparin

In previous work, we constructed expression strains for Homo sapiens heparin C5 epimerase (C5-epi5), Gallus gallus heparin 2-O-sulfotransferase (2OST55), Mus musculus heparin 6-O-sulfotransferase isoform 1 (6OST-115), and Homo sapiens heparin 3-O-sulfotransferase isoform 1 (3OST-156). All proteins were recombinantly expressed in, and purified from E. coli. -GlcNDFA-GlcA- polysaccharide analog was prepared according to the protocol “Preparation of K5 polysaccharide analogue”. To prepare -GlcNS-GlcA- polysaccharide analog, 1 mg/mL of the K5 polysaccharide analog was treated with 0.1 M LiOH for 1 h. Then, 50 mM MES was added and the pH was adjusted to approximately 7.0. Subsequently, 0.2 mg/mL NST and 2 mg/mL PAPS were added, and reaction was carried out at 37 °C for 12 h. This reaction converted all -GlcNTFA-GlcA- structures into -GlcNS-GlcA-. After the reaction, purification was performed using a DEAE column. For reaction involving 1 mg/mL of -GlcNS-GlcA- polysaccharide, 10 mg/mL PAPS and 0.2 mg/mL of each of C5-epi, 2OST, 6OST, and 3OST-1 were added, and reaction was carried out overnight at 37 °C. The product was then purified using a DEAE column and desalted using a Bio-Gel P-10 column. For analysis, a mixture of HepI, HepII, and HepIII was added to the engineered heparin, and then analysis of the disaccharide degradation products was performed using PAMN-HPLC.

Determination of polysaccharide molecular weight

The molecular weight and purity of polysaccharides were determined using high-performance gel permeation chromatography. The chromatographic method employed a mobile phase of 0.2 M NaCl solution, a BRT105-103-101 series gel column (8 × 300 mm), a flow-rate of 0.8 mL/min, a column temperature of 40 °C, an injection volume of 25 μL, a refractive index detector RID-10A, and an analysis time of 60 min.

Determination of anti-Xa and anti-IIa activity

The measurement of anti-Xa and anti-IIa activity was based on methods outlined in the Chinese Pharmacopoeia. Purified heparin was diluted stepwise to prepare solutions with a series of concentrations. Forty microliters of each concentration of the test sample solution was added to a 96-well enzyme-linked immunosorbent assay plate. Subsequently, 40 μL of antithrombin (ATIII; 1 IU/mL) was added to each well and incubated at 37 °C for 2 min. Then, 40 μL of Xa or IIa (8 μg/mL) was added and incubated at 37 °C for 2 min. Next, 40 μL of substrate S-2765 or S-2238 (0.8 mg/mL) was added and incubated at 37 °C for 2 min. Finally, 80 μL of 20% acetic acid was added to stop the reaction, and absorbance at 405 nm was measured immediately. A standard curve was established using an ungraded heparin standard. Nonlinear regression analysis was performed with absorbance as the y-axis and the concentration of the test sample solution as the x-axis to calculate IC50 values for the inhibition of factor Xa or IIa by the heparin.

Heparin disaccharide analysis

Unfractionated heparin, heparin sulfate, enoxaparin, nadroparin, dalteparin, and fondaparinux were purchased from Macklin. Dekaparin (chemoenzymatically synthesized low-molecular-weight heparin) was synthesized using chemoenzymatic method. Heparinases I, II, and III sourced from Flavobacterium heparinum were recombinantly expressed in E. coli. For disaccharide composition analysis of all heparins, a mixture of heparinases I, II, and III (each at 0.1 mg/mL) was used to completely digest samples overnight in 50 mM ammonium acetate buffer containing 2 mM CaCl2. After boiling and centrifugation of the reaction mixtures (each initially containing 50 μg of heparin), the filtrate was subjected to HPLC analysis.

HPLC for disaccharide composition analysis

PAMN-HPLC was used to analyze the disaccharide composition of heparin. The column was equilibrated with 1.8 mM sodium dihydrogen phosphate (mobile phase A, pH 3.0), followed by elution with a gradient of 1.8 mM sodium dihydrogen phosphate and 1 M potassium dihydrogen phosphate (mobile phase B, pH 3.0). The column temperature was maintained at 30 °C. The flow was set for a low-pressure gradient with a flow-rate of 0.5 mL/min. The gradient for mobile phase B was as follows: 0–6 min: 0%, 6–110 min: 0%–100%, 110–120 min: 0%; stopped at 120 min. The chromatograms were recorded using a UV detector at 232 nm. The injection volume was 10 μL. Disaccharide standards were purchased from Iduron (Manchester, UK).

HPLC for NST activity analysis

The deacetylated heparin pentasaccharide GlcA-GlcNH2-GlcA-GlcNH2-GlcA-pNP was used as the substrate for NST. PAMN-HPLC was employed to analyze the reaction products. For PAMN-HPLC, the column used was a Polyamine II-HPLC, 4.6 × 250 mm, from YMC. The elution process involved a linear gradient of KH2PO4 from 0 to 1 M over 40 min, followed by 1 M KH2PO4 for 30 min, at a flow-rate of 0.5 mL/min. The chromatograms were recorded using a UV detector at 310 nm. The injection volume was 10 μL.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.