Deciphering the substrate recognition mechanisms of the heparan sulfate 3-O-sulfotransferase-3

The sulfation at the 3-OH position of a glucosamine saccharide is a rare modification, but is critically important for the biological activities of heparan sulfate polysaccharides. Heparan sulfate 3-O-sulfotransferase (3-OST), the enzyme responsible for completing this modification, is present in seven different isoforms in humans. Individual isoforms display substrate selectivity to uniquely sulfated saccharide sequences present in heparan sulfate polysaccharides. Here, we report two ternary crystal structures of heparan sulfate 3-OST isoform 3 (3-OST-3) with PAP (3′-phosphoadenosine 5′-phosphate) and two octasaccharide substrates: non 6-O-sulfated octasaccharide (8-mer 1) and 6-O-sulfated octasaccharide (8-mer 3). The 8-mer 1 is a known favorable substrate for 3-OST-3, whereas the 8-mer 3 is an unfavorable one. Unlike the 8-mer 1, we discovered that the 8-mer 3 displays two binding orientations to the enzyme: productive binding and non-productive binding. Results from the enzyme activity studies demonstrate that 8-mer 3 can contribute to either substrate or product inhibition, possibly attributed to a non-productive binding mode. Our results suggest that heparan sulfate substrates interact with the 3-OST-3 enzyme in more than one orientation, which may regulate the activity of the enzyme. Our findings also suggest that different binding orientations between polysaccharides and their protein binding partners could influence biological outcomes.


Protein Co-crystallization
Crystals of the 3-OST-3/PAP/8-mer 1 complex were obtained using sitting drop vapor diffusion at 20C by mixing 325 nl of 3-OST-3 at 14 mg/ml in 25 mM Tris pH 7.5, 125 mM NaCl, 4 mM PAP, and 5 mM 8-mer 1 with 200 nl of reservoir solution consisting of 200 mM potassium iodide and 20% (w/v) polyethylene glycol (PEG) 3350. For data collection, 1 l of cryo solution consisting of 15% ethylene glycol, 85% reservoir, and 5mM 8-mer 1 was added directly to the drop, followed by transfer of the crystal to 100% cryo solution.
The crystal was then flash frozen in liquid nitrogen. For data collection, 1 l of cryo solution consisting of 26% ethylene glycol and 74% reservoir was added to the drop, followed by the addition of another 1 l. The crystal was then flash frozen in liquid nitrogen.
All data were collected on a Rigaku MicroMax 007HF X-ray generator using a Dectris P200 detector at -173C and processed with HKL3000. The structures were both solved independently in Phenix 3 . The protein component of PDB ID code 1T8U 1 was used as a search model with Phaser 4 . All refinement was carried out with multiple cycles of refinement in Phenix and manual model building in Coot 5,6 (Table 1).
These elongation steps were repeated until a 6-mer structure was obtained. The product of each step was purified using a C 18 column (0.75 x 20 cm; Biotage) and eluted on a linear gradient from 1-100% acetonitrile in H 2 O and 0.1% trifluoroacetic acid (TFA) in 60 minutes with a flow rate of 4 ml/min. Once the chain reached 6 saccharides, the GlcNTFA saccharides were detrifluroacetylated by addition of 0.1 M LiOH and maintenance of pH above 12 at room temperature for 10 min. The formation of the Oligosaccharide purity was determined by HPLC analysis with a ProPac A1 column. All synthesized oligosaccharides were >93% pure.

HPLC Analysis
Reactions were analyzed by a ProPac A1 column using buffer A (20 mM NaOAc, pH 5.0) and buffer B (20 mM NaOAc, 2M NaCl, pH 5.0). A linear gradient from 35-100% buffer B over 60 minutes with a flow rate of 1 ml/min, was used for elution, with monitoring at 310 nm (corresponds to the absorbance of the paranitrophenol on the reducing end of the oligosaccharides).    Displayed are Fo-Fc simulated annealing omit maps (blue) of (A) 8-mer 1 binding to 3OST-3 (B) 8-mer 3 binding in productive mode to 3OST-3 and (C) 8-mer 3 binding in nonproductive mode to 3-OST-3. Octasaccharides colored as in Figures 5 and 6 Nonproductive binding orientation Suppl. Fig. S6. Different conformations of saccharide f of 8-mer 3 in productive and nonproductive orientation when the octasaccharide interacts with 3-OST-3.

Suppl. Fig. S7. Sequence alignment of human 3-OST-3a to mouse 3-OST-1
Sequence alignment of human 3-OST-3a versus mouse 3-OST-1 catalytic domains based on crystal structures. Residues highlighted in red are the catalytic base glutamates (Glu184 for 3-OST-3 and Glu90 for 3-OST-1) and the catalytic acids (Lys162 for 3-OST-3 and Lys68 for 3-OST-1). Conserved residues lining the heparan sulfate binding cleft that greatly affect catalytic activity for 3-OST-3 when mutated are colored magenta 1 . Residues that line the binding cleft that are not conserved and were mutated in this study are colored green.