The 3-O-sulfation of N-sulfated glucosamine is the last event in the biosynthesis of heparin/heparan sulfate, giving rise to the antithrombin-binding pentasaccharide sequence AGA*IA, which is largely associated with the antithrombotic activity of these molecules. The aim of the present study was the structural and biochemical characterization of a previously unreported AGA*IA*-containing octasaccharide isolated from the very-low-molecular-mass heparin semuloparin, in which both glucosamine residues of the pentasaccharide moiety located at the non-reducing end bear 3-O-sulfate groups. Two-dimensional and STD (saturation transfer difference) NMR experiments clearly confirmed its structure and identified its ligand epitope binding to antithrombin. The molecular conformation of the octasaccharide–antithrombin complex has been determined by NMR experiments and docking/energy minimization. The presence of the second 3-O-sulfated glucosamine in the octasaccharide induced more than one order of magnitude increase in affinity to antithrombin compared to the pentasaccharide AGA*IA.

INTRODUCTION

Heparin and HS (heparan sulfate) are the most extensively studied polysaccharides belonging to the glycosaminoglycan family. They interact, mainly through their sulfate and carboxylate groups, with many proteins, such as enzymes, growth factors, plasma glycoproteins, cytokines and extracellular matrix proteins, influencing a large number of biological processes [1,2].

The backbone of heparin/HS chains comprises alternating 1-4 linked β-D-glucuronic acid (G) or α-L-iduronic acid (I) and α-D-N-glucosamine (A) residues. The A residues can be N-sulfated (ANS) or N-acetylated (ANAc) at position (P-) 2, O-sulfated at P-6 (A6S) and, more rarely, at P-3 (3SA; A* signifies 3SA6SNS). The uronic acid residues, I and more rarely G, can be O-sulfated at P-2 (I2S/G2S). In heparin the most prevalent disaccharide is -I2S-A6SNS-, whereas, in HS highly sulfated domains alternate with sparsely sulfated domains, consisting predominantly of -G-ANAc/NS- [3].

Heparin and HS are biosynthesized as proteoglycans consisting of a unique core protein (serglycin) to which multiple polysaccharide chains (Mr 60000–10000) are covalently attached. In contrast to HS, which is expressed and secreted by most mammalian cells and located on cell surfaces and in the extracellular matrix, heparin is produced exclusively in mast cells. Following synthesis, heparin chains are cleaved at random points by an endo-β-D-glucuronidase to generate shorter heparin chains (Mr 25000–5000) that are stored in the cytoplasmic secretory granules of mast cells.

The biosynthesis of heparin and HS takes place in the Golgi network and is initiated by the construction of the tetrasaccharide-linkage region (-GlcA-β1→3-Gal-β1→3-Gal-β1→4-Xyl-β1→O-Ser), which is assembled by the transfer of monosaccharides from respective UDP (uridine diphosphate) sugars to a serine residue of the core proteins. Next, the sequential addition of the G and ANAc units to the non-reducing end extends the chain. In close association with chain elongation, a series of reactions involving N-deacetylation/N-sulfation of ANAc residues, epimerization of G to I and O-sulfation at different positions occur to generate the final heparin/HS polymer. However, the mechanism of the non-random distribution of the low-sulfated and high sulfated disaccharide units in HS is still unknown [4].

A particularly important enzyme acting in heparin/HS biosynthesis is the 3-OST (3-O-sulfotransferase), which catalyses the transfer of an O-sulfate to P-3 of A with the formation of the antithrombin-binding pentasaccharide -A6SNAc-G-3SA6SNS-I2S-A6SNS- (AGA*IA). This pentasaccharide is responsible for the interaction with serine protease inhibitor AT (antithrombin), resulting in a conformational change and subsequent inhibition of major coagulation cascade proteases, including thrombin (Factor IIa) and Factor Xa. The biosynthesis of the AT-binding region has been the subject of numerous studies [5,6]. Earlier studies proposed that incorporation of the critical 3-O-sulfate group concludes the biosynthetic process adding the final structural component required for the AT binding [7]. Further studies demonstrated that 3-OST exists in seven different isoforms, which are present in different tissues and have distinct functional roles [8,9]. It was demonstrated that 3-OST-1 preferentially transfers a sulfo group to P-3 of ANS or A6SNS linked at the non-reducing end to G, generating AT-binding sites, whereas, 3-OST-2 is able to 3-O-sulfate ANS in the disaccharide -G2S/I2S-ANS-, producing a different protein-binding sequence [10,11]. For instance, it was proved that 3-OST-3 modification of HS HSV-1 (herpes simplex virus type-1)-resistant cells made the cells susceptible to HSV-1 infection [12]. Moreover, the A6SNS-I2S-A* sequence has been identified to be present in substantial amounts in the glomerular basement membrane HS [13], whereas high levels of G2S in HS isolated from various sections of human brain was reported [14]. Finally, the extremely high content of 3OSG (3-O-sulfated glucosamine), found in both AT-binding and non-binding human follicular fluid HS sequences [15], and the observation of a tetrasulfated disaccharide unit in HS from various tissues [16] supports the observation that this residue is not exclusively present in the pentasaccharide sequence.

In spite of that, oligosaccharides so far isolated from heparin through chemical or enzymatic depolymerization contain 3OSG only in the -G-A*-I- sequence of the AT-binding pentasaccharide. Besides the presence of N-acetyl or N-sulfate groups at P-2 of the non-reducing glucosamine residue, and the possible lack of the sulfo group at P-6 of the A* residue (hexasaccharide isolated from semuloparin; results not shown), such oligosaccharides containing intact, or portions of, the AT-binding sequence do not show any variability in their pentasaccharide part [17,18]. Only synthetic heparin pentasaccharide analogues containing an extra 3OSG at the reducing end have been reported to elicit a higher affinity to AT [19].

In contrast greater structural variability can occur in residues flanking the AGA*IA sequence, influencing both AT affinity and the anti-Factor Xa activity of these oligosaccharides. The role of residues elongating the pentasaccharide sequence towards both ends was recently demonstrated for octasaccharides isolated from enoxaparin. Although the presence of some residues, such as the G preceding the AGA*IA sequence or 1,6-anhydro-hexosamine at the reducing end of the pentasaccharide, can be generated by the depolymerization process, all other structural variations reflect the natural heterogeneity of heparin [2022].

In the present study we have isolated and characterized an AT-binding heparin octasaccharide containing two 3-O-sulfated glucosamine residues. Like all fragments generated by β-elimination cleavage of heparin chains, OCTA-7, ΔU-A6SNAc-G-A*-I2S-A*’-I2S’-A6SNS(red), terminates at the non-reducing end with a 4,5-unsaturated uronic acid residue (ΔU). In OCTA-7 the AT-binding pentasaccharide sequence contains a second 3OSG at the reducing end of the pentasaccharide moiety and is accordingly designed as AGA*IA* (Figure 1). This pentasaccharide has a structure similar to that of a synthetic pentasaccharide described previously, which possessed an N-sulfated glucosamine instead of a N-acetylated glucosamine residue at its non-reducing side [19].

Structure of octasaccharides OCTA-7 and OCTA-1

Figure 1
Structure of octasaccharides OCTA-7 and OCTA-1

The AGA*IA* and AGA*IA sequences are highlighted by a dark grey and a grey rectangle respectively.

Figure 1
Structure of octasaccharides OCTA-7 and OCTA-1

The AGA*IA* and AGA*IA sequences are highlighted by a dark grey and a grey rectangle respectively.

The structure of the octasaccharide was unambiguously defined by controlled enzymatic digestion and NMR spectroscopy (see the Supplementary Online Data at http://www.biochemj.org/bj/449/bj4490343add.htm) and its binding region, which is involved in the interaction with AT, was determined by STD (saturation transfer difference) experiments. The conformational properties of OCTA-7 in the free state and in complex with AT was also investigated by NMR [NOE (nuclear Overhauser effect) and tr-NOE (transferred NOE)] spectroscopy, docking simulations and energy minimization. OCTA-7 showed higher affinity to AT with respect to the synthetic pentasaccharide Fondaparinux (AGA*IA) and the analogous OCTA-1, ΔU-A6SNAc-G-A*-I2S-A6SNS-I2S’-A6SNS(red), lacking the 3-O-sulfate group at the reducing glucosamine of the pentasaccharide sequence [21].

EXPERIMENTAL

Isolation and structural characterization by enzymatic sequencing of OCTA-7

OCTA-7 was isolated and characterized as described previously [22] using a combination of AT affinity, CTA (cetyltrimethylammonium)-SAX (strong anion-exchange) and SAX Carbopack AS11 (Dionex) chromatography, starting from an octasaccharide GPC fraction of Semuloparin. The structure of the OCTA-7 was determined using a controlled digestion of the sample with heparitinase (heparinase I, EC 4.2.2.7) followed by HPLC analysis according to the procedure described previously [23]. Briefly, a pH 7 aqueous solution of reduced OCTA-7, by NaBH4 in aqueous solution, was treated with heparitinase, giving at first a mixture of two tetrasaccharides, ΔU-A6SNAc-G-A* and the reduced ΔU2S-A*-I2S-A6SNS. The latter was quickly digested into two disaccharides, ΔU2S-A* and the reducing ΔU2S-A6SNS, giving the proposed structure ΔU-A6SNAc-G-A*-I2S-A*'-I2S'-A6SNS. Only the configuration of the tetrasulfated disaccharide had to be determined by NMR. A detailed description can be found in Supplementary Online data.

Fluorescence titration

The equilibrium dissociation constant (Kd) for the interaction between AT and the OCTA-7 or AGA*IA, was assessed at 25°C in 0.05 M Hepes (pH 7.4) and 0.1 M NaCl. The Kd value was obtained by monitoring the enhancement of the intrinsic fluorescence of the serpin upon its reaction with increasing concentrations of the oligosaccharides [24,25].

The fluorescence intensities (λex=280 nm and λem=340 nm) were measured using a RF5000 Shimadzu fluorospectrophotometer equipped with a temperature controlled cell holder. The best estimate of the Kd value was determined by fitting the fluorescence data to the quadratic equilibrium binding equation eqn (1):

 
formula

where, [AT]0 and [ols]0 are the AT and oligosaccharide initial concentrations respectively, Δf and Δfmax are the absolute change of fluorescence intensity (ff0) for a given oligosaccharide concentration and that of the maximum fluorescence intensity change [(ff0)max] respectively, and n is the binding stoichiometry. In the present study n was set to 1 since the two analysed oligosaccharide preparations were considered to contain a single product, each molecule of which is capable of binding to one molecule of AT. Non-linear regression analysis was conducted with the GraFit software (Erithacus Software).

NMR spectroscopy

NMR spectra were recorded using a Bruker Avance III spectrometer (600 MHz) equipped with a high-sensitivity 5 mm TCI cryoprobe at 33°C. After exchanging the samples three times in 2H2O, samples were dissolved in 0.2 ml 10 mM phosphate buffer, 0.5 M NaCl (pH 7.4) and 3 mM EDTA in 2H2O (99.996%) and placed in 3-mm NMR tubes. The samples used to detect NH signals (proton, STD and spectra) were dissolved in 90:10 H2O/2H2O mixture in the same buffer. For the tr-NOE experiments the samples were prepared by dissolving 1 mg of AT and 180 μg of the octasaccharide in phosphate buffer reaching a molar ratio of 1:3.5 AT/octasaccharide. Proton spectra were recorded with pre-saturation of the residual water signal with a recycle delay of 12 s and 256 scans. Bidimensional DQF (double-quantum-filter)-COSY and 2D (two dimensional)-TOCSY spectra were acquired using 32 scans per series of 2048×512 data points with zero filling in F1 (4096×2048) and a shifted (π/3) squared cosine function was applied prior to Fourier transformation. HSQC (heteronuclear single-quantum coherence) spectra were obtained in phase-sensitivity enhanced pure-absorption mode with decoupling in the acquisition period. The matrix size of 1024×320 data points was zero filled to 4096×2048 by application of a squared cosine function prior to Fourier transformation. All 2D-NOESY and 2D tr-NOESY (transferred NOESY) experiments were performed in a similar way. A total of 32 scans were collected for each free-induction decay (matrix 2048×512 points), the data were zero-filled to 4096×2048 points before Fourier transformation, and mixing time values of 100, 200 and 400 ms were used.

Monodimensional STD experiments

Octasaccharide–AT samples were prepared so that the final protein concentration was 3.4×10−6 M with a ligand/AT molar ratio of 40:1. The pulse sequence used for the monodimensional STD NMR experiments includes a 30 ms spin-lock pulse to eliminate the broad resonances of the protein. A train of 12, 20 and 40 Gaussian-shaped pulses of 50 ms each were applied to produce selective saturations. The on-resonance irradiation was performed at the low-field protein resonances (≫7.2 p.p.m.), whereas the off-resonance control irradiation was performed at 24 p.p.m. The STD spectrum was obtained by phase cycling subtraction of the on-resonance and off-resonance data acquired in interleaved mode. The number of scans and dummy scans were 2048 and 16 respectively.

Computational studies on the octasaccharide–AT complexes

The octasaccharide models were created using MAESTRO, I2S and I2S' of OCTA-7 were set in the 2S0 conformation, whereas in OCTA-1 they were set as 2S0 and 1C4 respectively [22]. The A and G residues were configured as 4C1 [22]. The OCTA-1–AT and OCTA-7–AT complexes were built starting from the AGA*IA–AT structure published previously [26], the complexes were minimized using MacroModel bmin (500–2000 steps). The AMBER* force field was used for the simulation, with the extended cut-off: 20.0, 8.0 and 4.0 Å (1 Å=0.1 nm) for electrostatic, dispersive and hydrogen bond interactions respectively. No restraints were applied on the complex atoms. Before and after 2000 steps of energy minimization the OCTA-7–AT complex geometry was tested, comparing selected simulated tr-NOESY signals with the corresponding experimental data. CORCEMA (complete relaxation and conformational exchange matrix) analysis was applied to simulate the tr-NOESY signals starting from the minimized OCTA-7–AT geometries [27]. For the CORCEMA analysis a qualitative selection of Kd=0.24 μM (experimental value in 0.5 M NaCl) and koff=8.0 s−1 was used. In this condition the exchange times for ligand (τL), protein (τE), and the complex (τEL) were, 0.512, 3.01×10−4, and 0.25 s, respectively, while the correlation time for the ligand and the complex was set to 0.8 and 46.0 ns [22]. Protons selected for the CORCEMA analysis belong to the ligand and the AT sequence Gly2–Ser25/Lys39–Lys53/Thr110–Lys139. The tr-NOESY H1–H2 intensity of the four glucosamines of OCTA-7 were used as a qualitative check of the relative position of the ligand to the AT-binding site; after 2000 minimization steps these simulated intensities fitted qualitatively the experimental values for the bound OCTA-7 at mixing times of 100, 200 and 400 ms.

The R-factor, was calculated using eqn (2):

 
formula

where NOEsim is simulated tr.NOE and NOEexp is experimental tr.NOE.

RESULTS AND DISCUSSION

Equilibrium dissociation constants

The interaction of OCTA-7 with AT resulted in a 40% increase in the protein's intrinsic fluorescence compared with that of AGA*IA interacting with AT. The Kd value measured in 0.05 M Hepes (pH 7.4) and 0.1 M NaCl, for the OCTA-7–AT complex was 1.1 nM, an affinity much higher than that between OCTA-1–AT (208 nM) and AGA*IA–AT (21 nM) (a 1:1 binding stoichiometry was used to determine the best estimate of the Kd value [21]).

NMR characterization of OCTA-7

Proton and carbon resonances (Table 1) were assigned using standard COSY, TOCSY and HSQC (heteronuclear single-quantum coherence) pulse sequences, with NOESY experiments used to sequentially connect saccharide ring systems (Supplementary Figures S2 and S3 at http://www.biochemj.org/bj/449/bj4490343add.htm). The proton spectrum of OCTA-7 is shown in Supplementary Figure S1 (at http://www.biochemj.org/bj/449/bj4490343add.htm).

Table 1
OCTA-7 proton/carbon chemical shifts (p.p.m.) and 3JHH coupling constants (Hz) (30°C, phosphate buffer 10 mM, pH 7.4, and 0.5 M NaCl)
Parameter ΔU ANAc A* I2S A*' I2SANSred 
H1/C1 5.206/103.6 5.452/99.5 4.653/101.4 5.571/98.4 5.246/103.7 5.481/99.5 5.243/102.2 5.483/93.7 
3JH1-H2 5.8 3.8 7.9 3.5 4.0 3.5 3.6 3.6 
H2/C2 3.866/72.9 3.997/56.0 3.421/79.7 3.489/59.2 4.410/76.1 3.525/59.5 4.357/79.3 3.305/60.6 
3JH2-H3   9.3 10.8 7.6 10.8 7.0 10.3 
H3/C3 4.279/69.4 3.844/71.8 3.738/73.4 4.413/78.9 4.157/79.0 4.448/78.3 4.233/72.5 3.745/72.1 
3JH3-H4 3.6    4.1  3.9  
H4/C4 5.865/110.3 3.899/80.6 3.835/78.8 4.012/75.6 4.245/79.1 4.072/74.4 4.184/79.2 3.813/79.7 
3JH4-H5     3.4  3.1  
H5/C5  4.083/71.6 4.974/73.1 4.255/72.2 4.972/79.1 4.123/72.5 4.806/72.5 4.162/71.1 
3JH5-H6a         
3JH5-H6b         
H6a/H6b/C6  4.484/4.238/68.5  4.532/4.318/68.7  4.528/4.342/68.7  4.453/4.334/69.4 
3JH6a-H6b         
-COCH3  24.8       
Parameter ΔU ANAc A* I2S A*' I2SANSred 
H1/C1 5.206/103.6 5.452/99.5 4.653/101.4 5.571/98.4 5.246/103.7 5.481/99.5 5.243/102.2 5.483/93.7 
3JH1-H2 5.8 3.8 7.9 3.5 4.0 3.5 3.6 3.6 
H2/C2 3.866/72.9 3.997/56.0 3.421/79.7 3.489/59.2 4.410/76.1 3.525/59.5 4.357/79.3 3.305/60.6 
3JH2-H3   9.3 10.8 7.6 10.8 7.0 10.3 
H3/C3 4.279/69.4 3.844/71.8 3.738/73.4 4.413/78.9 4.157/79.0 4.448/78.3 4.233/72.5 3.745/72.1 
3JH3-H4 3.6    4.1  3.9  
H4/C4 5.865/110.3 3.899/80.6 3.835/78.8 4.012/75.6 4.245/79.1 4.072/74.4 4.184/79.2 3.813/79.7 
3JH4-H5     3.4  3.1  
H5/C5  4.083/71.6 4.974/73.1 4.255/72.2 4.972/79.1 4.123/72.5 4.806/72.5 4.162/71.1 
3JH5-H6a         
3JH5-H6b         
H6a/H6b/C6  4.484/4.238/68.5  4.532/4.318/68.7  4.528/4.342/68.7  4.453/4.334/69.4 
3JH6a-H6b         
-COCH3  24.8       

The conformation of the two iduronate residues, I2S and I2S', agrees with previous data, as they adopt multiple equi-energetic conformations such as 1C4, 2S0 and, less frequently in heparin sequences, 4C1 [28]. The 3JHH values (Table 1) for the unbound OCTA-7 indicates that the prevalent conformation of the I2S residue is 2S0 (75–80%), whereas I2S' is almost equally distributed between 2S0 and 1C4 (55% and 45% respectively) [29]. The smaller 3JHH values measured for both I2S residues (Table 1) compared with the pentasaccharide synthetic model [19] are due to the high salt concentration present in the buffer (see the Materials and methods section in the Supplementary Online Data). The electrostatic repulsion between the anionic charges of neighbouring glucosamine residues, which is the driving force that determines the conformational equilibrium of the I2S, may be partially attenuated by the presence of counter-ions promoting the 1C4 conformer [29]. The presence of the extra 3OSG increases the 2S0 conformation population of both iduronate moieties as demonstrated by the comparison with 3JH-H values measured on a synthetic pentasaccharide also bearing two 3-O-sulfated glucosamine residues [19].

The NOESY data were in agreement with the above results, since the presence of characteristic intraresidue cross-peaks between H2 and H5 were detected for both iduronate residues. This NOE contact is the marker of the 2S0 conformation, since it cannot take place in the 1C4 conformer. Similar magnitudes of experimental H5-H4 and H5-H2 NOEs were found for the iduronate residue of the A*-I2S-A*’ sequence present exclusively in OCTA-7 (Table 2) and present results suggest a similar distance between I5-I4 and I5-I2 protons, characteristic of the almost pure 2S0 form. Interglycosidic NOEs were similar to those measured in the analogue OCTA-1, which lacks the 3-O-sulfo modification at the reducing-end glucosamine [21] (Supplementary Table S1 at http://www.biochemj.org/bj/449/bj4490343add.htm).

Table 2
Iduronate residue (H5-H4/H5-H2) NOE ratios for AGA*IA [30] and OCTA-7 (free and bound states; mixing time=200 ms)
Ligand Sequence H5-H4/H5-H2 ratio (free) H5-H4/H5-H2 ratio (bound) 
AGA*IA -A*-I2S-A6SNS 2.3 1.1 
OCTA-7 -A*-I2S-A*'− 1.9 1.0 
 -A*'-I2S-A6SNS 2.2 1.3 
Ligand Sequence H5-H4/H5-H2 ratio (free) H5-H4/H5-H2 ratio (bound) 
AGA*IA -A*-I2S-A6SNS 2.3 1.1 
OCTA-7 -A*-I2S-A*'− 1.9 1.0 
 -A*'-I2S-A6SNS 2.2 1.3 

OCTA-7 temperature coefficients

Proton NMR spectra of OCTA-7 were measured with temperatures ranging from 277 K to 298 K. Using COSY and TOCSY spectra it was possible to assign the sulfamate (NHSO3) signals of A*, A*' and A6SNS(red), whereas the amidic NHCOCH3 proton of ANAc was not observed, indicating a fast exchange rate with water, even at low temperature (Supplementary Figures S4 and S5 at http://www.biochemj.org/bj/449/bj4490343add.htm). The NHSO3 signals belonging to A* and A*' were sharp, even at 298 K, unlike the NHSO3 signal of reducing-end glucosamine, A6SNS(red), suggesting their probable involvement in hydrogen bonding. The presence of a persistent hydrogen bond between the sulfamate NH and the adjacent 3-O-sulfo group of A* has been described for AGA*IA [31]. Lower temperature coefficients compared with those of the reducing hexosamine (−4.4 p.p.b./K) were observed for the NHSO3 protons of both 3-O-sulfated glucosamine residues (0.9 p.p.b./K, −2.9 p.p.b./K for the NHSO3 of A*' and A* respectively), which was also observed by Langeslay et al. [31]. In particular A*' showed the lowest temperature coefficients, 9 p.p.b./K, indicating a stronger hydrogen bond involving this proton (Supplementary Figure S6 at http://www.biochemj.org/bj/449/bj4490343add.htm).

STD experiments

STD NMR spectroscopy is a powerful method for the study of interactions between ligands and macromolecular receptors. The experiment is based on the transfer of saturation from the macromolecule to the bound ligand. The protons of the ligand that have the ‘strongest’ contact to the macromolecule will show the most intense STD signals, enabling the mapping of the ligand's binding epitope [32].

Both OCTA-1–AT and OCTA-7–AT were studied (OCTA-1–AT was previously measured, but the experiments were repeated in the same conditions as OCTA-7–AT [21]); the STD intensities were normalized to the most intense STD signal of the spectrum, anomeric protons of A* (Table 3).

Table 3
Relative STD intensities for OCTA-7 and OCTA-1 when bound to AT

The STD NMR values are normalized using H1 of A* as a reference. The corresponding fractional STD effect values of A* are shown in parenthesis [34].

Ligand H1 ΔU2S H4 ΔU2S H1 ANAc H2 G H1 A* H2 A* H1 I2S H5 I2S H1 A*'/ANS H2 A*'/ANS H1 I2SH5 I2SH1 Ared H2 Ared 
OCTA-7 72 86 87 100 100 (0.091) 98 68§ 84 72† 98 68§ 62 72† 33 
OCTA-1 71 77 99‡ 100 100 (0.21) 99 94 80 99‡ 99 38 46 32 25 
Ligand H1 ΔU2S H4 ΔU2S H1 ANAc H2 G H1 A* H2 A* H1 I2S H5 I2S H1 A*'/ANS H2 A*'/ANS H1 I2SH5 I2SH1 Ared H2 Ared 
OCTA-7 72 86 87 100 100 (0.091) 98 68§ 84 72† 98 68§ 62 72† 33 
OCTA-1 71 77 99‡ 100 100 (0.21) 99 94 80 99‡ 99 38 46 32 25 

Overlapping signals of H1 A*' and H1 Ared.

Overlapping signals of H1 ANAc and H1 ANS.

§

Overlapping signals of H1 I2S and H1 I2S'.

For OCTA-1–AT the reducing-end disaccharide I2S'-A6SNS(red) is less involved in the interaction with AT compared with the AGA*IA region [21], this is shown by the weaker STD signals corresponding to H1 and H2 of A6SNS(red) and H1 of I2S' (Table 3 and Figure 2a). Although, some of the anomeric signals of OCTA-7 are superimposed in the spectrum, weaker STD signals observed for the reducing disaccharide indicates that there is a greater distance between those residues and AT. Saturation transfer difference signals were slightly greater for H2 of A6SNS(red) and H5 of I2S' in OCTA-7 compared with OCTA-1 (Table 3), suggesting that the extra 3OSG of OCT-7, A*', increases the contact between the reducing disaccharide, I2S'-A6SNS(red), and AT. STD intensities were also estimated by using the STD-AF (STD amplification factor) [33]. Diagrams reporting the STD-AF and saturation times are shown for both the OCTA-7 and OCTA-1 in Supplementary Figures S7 and S8 (at http://www.biochemj.org/bj/449/bj4490343add.htm). The OCTA-1 STD-AF signals belonging to I2S'-A6SNS(red) (1.8<STD-AF<4.1 at 2 s) are smaller than those of the other residues (STD-AF>6 at 2 s). In the case of OCTA-7, a relatively high STD-AF value was observed for H5 of I2S' (STD-AF of 2.7 at 2 s), which was comparable with H1 of ΔU. This confirms that the extra 3-O-sulfo group of OCTA-7, A*', increases the contact between the reducing disaccharide of OCTA-7 and AT, principally interacting through I2S'. It is worth noting that significantly higher STD-AF values were observed for OCTA-1 compared with OCTA-7, since the two studied complexes are very similar (in terms of structures, binding regions and motional properties). The intensity of STD signals depends mainly on the different dissociation off-rate constant, which is lower for OCTA-7.

1H NMR spectrum (lower traces) and STD spectrum (upper traces) of OCTA-1:AT (a) and OCTA-7:AT (b) (molar ratio 40:1, 35°C, 2H2O buffer)

Saturation transfer difference experiments were also conducted in H2O/2H2O, which made it possible to observe the interaction between the glucosamine amide groups and AT (Figure 3). In both octasaccharides the non-reducing end NH group of N-acetyl glucosamine was not observed because of the faster exchange rate with H2O. Owing to the different koff values, the STD-AF values are higher for the NH groups of OCTA-1 compared with those measured for OCTA-7, as observed in the experiment conducted in 2H2O. The OCTA-7 NH moieties of A* and A*' have strong STD signals, whereas the NH signal belonging to the reducing glucosamine, A6SNS(red), for both octasaccharides has no STD signal, confirming that they play a minor role in the interaction with AT (Table 4).

1H NMR spectrum (lower traces) and STD spectrum (upper traces) of OCTA-1:AT (a) and OCTA-7:AT (b) (molar ratio 40:1, 7°C, in 90:10 H2O/2H2O)

Table 4
STD amplification factors for the NH signals of OCTA-1 and OCTA-7 at 1 s saturation time measured in H2O/2H2O solution buffer at 7°C

NH chemical shifts are shown in parenthesis. nd, not determined.

Ligand NH A* NH ANS NH A*' NH Ared 
OCTA-1 3.40 (5.54 p.p.m.) 2.19 (5.45 p.p.m.) – nd (5.52 p.p.m.) 
OCTA-7 0.68 (5.72 p.p.m.) – 0.60 (5.59 p.p.m.) nd (5.50 p.p.m.) 
Ligand NH A* NH ANS NH A*' NH Ared 
OCTA-1 3.40 (5.54 p.p.m.) 2.19 (5.45 p.p.m.) – nd (5.52 p.p.m.) 
OCTA-7 0.68 (5.72 p.p.m.) – 0.60 (5.59 p.p.m.) nd (5.50 p.p.m.) 

Conformation of the OCTA-7–AT complex

On the addition of AT to OCTA-7, the line width of signals in the 1H spectrum increased by 2–3 Hz, indicating an interaction between the two (Supplementary Figure S9 at http://www.biochemj.org/bj/449/bj4490343add.htm). The small protein-induced shift observed in the spectrum is consistent with an equilibrium regulated by the slow to intermediate dynamic exchange between the free and bound states. Owing to the cross-relaxation rates differing considerably between the free and bound state, there is variation in the cross-peak intensities in the two-dimensional NOESY (free state) and tr-NOESY (bound state) spectra (Supplementary Table S1). Iduronic acid conformations were investigated by analysis of the tr-NOESY spectra (mixing time=100, 200 and 400 ms). A significant change in the ratio between H5-H4 and H5-H2 NOE magnitudes for both I2S and I2S' residues compared with the free state, was observed for the OCTA-7–AT complex (Table 2).

As expected, the interaction between OCTA-7 and AT drives I2S towards an almost pure 2S0 conformation, as confirmed by NOESY ratios between H5-H4/H5-H2, which change from 1.9 to 1.0 after binding to AT. This has also been observed for AGA*IA [30], where the AT-binding surface seems to force I2S to assume the 2S0 conformation. In contrast with those observed in the interaction of OCTA-1 with AT, where the reducing iduronate, I2S', principally maintains a chair conformation [22] the I2S' of OCTA-7 assumes a prevalent 2S0 conformation. Modelling of the OCTA-7–AT complex suggests that when I2S' is in the 2S0 form it allows an extra contact between the 2-O-sulfo group and Arg24 (3.9 Å, Figure 4), this is not possible for I2S', which assumes the 1C4 form in OCTA-1. This difference in conformation for I2S' residues within OCTA-1 and OCTA-7 is the primary difference between the two carbohydrates when they are bound to AT.

Side view of the OCTA-7:AT (blue) and OCTA-1:AT (cyan) complexes superimposed

Figure 4
Side view of the OCTA-7:AT (blue) and OCTA-1:AT (cyan) complexes superimposed

The OCTA-7 I2S' residue assumes a 2S0 conformation, with its 2-O-sulfo-group oriented towards the Arg24 (3.9 Å). This contact is missing for OCTA-1:AT where I2S', which is in the 1C4 form, orients the 2-O-sulfo-group away from the protein surface.

Figure 4
Side view of the OCTA-7:AT (blue) and OCTA-1:AT (cyan) complexes superimposed

The OCTA-7 I2S' residue assumes a 2S0 conformation, with its 2-O-sulfo-group oriented towards the Arg24 (3.9 Å). This contact is missing for OCTA-1:AT where I2S', which is in the 1C4 form, orients the 2-O-sulfo-group away from the protein surface.

The interaction between OCTA-7 and AT does not affect the glycosidic linkage between A*-I2S, this is also observed for OCTA-1. In contrast, there is a significant increase in the magnitude of the NOE signal arising from A*'(H1)-I2S'(H4) with respect to that of A*'(H1)-I2S'(H3), which is not observed for the OCTA-1 [22], suggesting that the extra 3-O-sulfo group affects the corresponding glycosidic geometry of OCTA-7 in the bound state (Supplementary Table S1).

The superimposition of the two modelled complexes OCTA-7–AT and OCTA-1–AT show significant differences in terms of their contacts between the sugars and the protein. In the OCTA-7–AT complex the 3-O-sulfo group of A*' is pointed towards Arg46 and Arg47 of AT, forming an electrostatic interaction which was also observed for a synthetic pentasaccharide analogue [26], whereas in the OCTA-1–AT complex this electrostatic interaction seems to be replaced by a hydrogen bond involving the hydroxyl group now at P-3 of A6SNS (Figure 4). The distances between Arg46 and Arg47 and the 3-O-sulfo group of A*' (OCTA-7) is 7.5 and 5.8 Å respectively, compared with the hydroxyls at P-3 of A6SNS OCTA-1 which are 8.4 and 4.8 Å away from Arg46 and Arg47 respectively.

To validate the proposed models, theoretical tr-NOEs were computed for both OCTA-7–AT complexes using the CORCEMA program [27]. The models in Figure 4 show a good fit between theoretical and experimental intraresidue tr-NOEs. The relative position of OCTA-7 in the AT active site is in agreement with the experimental data, as shown by the R-factor calculated on the H1-H2 tr-NOESY of the glucosamines (Table 5). Also interglycosidic tr-NOEs gave relatively good R-factors ranging from 0.31 to 0.52. However, because small variations of glycosidic torsions give rise to strong variations of inter-glycosidic proton distances (and consequently in theoretical tr-NOE magnitudes), the models obtained so far provide useful indications of the ‘true’ octasaccharide geometry.

Table 5
R-factor between experimental and calculated tr-NOESY signals for the OCTA-7–AT complex
tr-NOESY R-factor 
A6SNAc (H1-H2) 0.25 
A* (H1-H2) 0.14 
A*' (H1-H2) 0.05 
A6SNS(red) (H1-H2) 0.41 
ΔU(H1)-A6SNAc(H4) 0.38 
A*(H1)-I2S(H3) 0.52 
A*(H1)-I2S(H4) 0.51 
A*'(H1)-I2S'(H3) 0.41 
A*'(H1)-I2S'(H4) 0.40 
I2S'(H1)-A6SNS(red)(H4) 0.31 
tr-NOESY R-factor 
A6SNAc (H1-H2) 0.25 
A* (H1-H2) 0.14 
A*' (H1-H2) 0.05 
A6SNS(red) (H1-H2) 0.41 
ΔU(H1)-A6SNAc(H4) 0.38 
A*(H1)-I2S(H3) 0.52 
A*(H1)-I2S(H4) 0.51 
A*'(H1)-I2S'(H3) 0.41 
A*'(H1)-I2S'(H4) 0.40 
I2S'(H1)-A6SNS(red)(H4) 0.31 

In conclusion, the existence of heparin/HS structures with additional 3OSG residues appears highly plausible in view of the substrate specificities defined for various members of the 3-OST family [35]. In fact, several oligosaccharides containing 3OSGs located in sequences distinct from the AT-binding site have been identified in different HS samples [10,13]. Some of these structures are not recognized by AT, but, presumably, they are recognised by other HS binding proteins [36]. In fact, 3OSG residues were also found in heparin fractions which had no affinity for AT, suggesting that this residue can be present in different sequences that do not contribute to AT-mediated anticoagulant activity [37]. However, all fully characterized oligosaccharides so far isolated from heparin have had 3OSG residues exclusively in the -G-A*- sequence, present within the pentasaccharide AGA*IA motif [17,18,2022]. Isolation of AGA*IA-containing octasaccharides showed the important role of the reducing and non-reducing AGA*IA extension for the interaction with AT, both providing additional contacts with the protein, enhancing the electrostatic interactions between the pentasaccharide moieties and the protein [21]. In the present study an octasaccharide isolated from the ultra-low-molecular-mass heparin semuloparin [38], containing two 3-O-sulfated-N-sulfo-glucosamine residues within the active pentasaccharide, was described for the first time. Fluorescence titration experiments indicated that the affinity to AT of the OCTA-7 is characterized by an equilibrium dissociation constant 20-fold lower than that measured for the AGA*IA–AT complex. Models generated from docking and energy minimization elucidated that the extra 3-O-sulfated group increases the contact between OCTA-7 and AT, the interaction being through the positively charged amino acids (Arg46 and Arg47) of AT. It is interesting to note that similar contacts were also observed in AT co-crystallized with an AGA*IA analogue bearing an extra 3-O-sulfated group on the reducing glucose (i.e. AGA*IA*) [26]. It was speculated that these contacts could be the cause of the higher affinity of AT to a synthetic pentasaccharide differing from the AGA*IA* moiety of the OCTA-7 by the presence of a N-sulfated group instead of a N-acetylated group at the non-reducing glucosamine [19,39]. STD NMR experiments indicated that the -I2S-A6SNS(red) reducing extension of OCTA-7 lies slightly closer to the AT-binding site compared with that of OCTA-1 [21], indicating that the 3-O-sulfated group influences the position of such disaccharides. The conformation of the I2S' residue, which is driven towards the 2S0 form by the presence of AT, probably contributes to the stability of this assembly. Notably, the same residue of OCTA-1 remained in the 1C4 form in the bound state, suggesting that the conformational plasticity of I2S' contributed to the formation of the ‘proper’ three-dimensional structure in the protein-binding site, enhancing the electrostatic interactions between AGA*IA and AT.

The discovery of OCTA-7 revises the role of the different enzymes participating in the HS/heparin biosynthetic pathway. The possibility of having two contiguous disaccharides bearing 3-O-sulfated glucosamine residues in heparin/HS chains has to be controlled by the substrate specificities of different 3-OST isoforms, each of them exhibiting unique substrate specificities [4042]. Whereas the 6-O-sulfation of glucosamine and the carboxylate groups of the -ANAc-G- disaccharide located at the non-reducing end of the 3-O-sulfation site are necessary for the interaction with the 3-OST-1, the 2-O-sulfation of the uronate residue, always located at the non-reducing end, is essential for the interaction with the 3-OST-3. Moreover, the N-sulfo group of the glucosamine located two residues toward the reducing end appear to be crucial for the structural stability of the 3-OST-3–substrate complex [40]. The unusual sequence of OCTA-7 demonstrates that a single heparin chain might act as substrate for more than one 3-OSTs, supporting the hypothesis that the 3OSG can be located also in I2S-3SA6SNS-I2S sequences, not necessarily involved in the interaction with AT [43].

Abbreviations

     
  • AT

    antithrombin

  •  
  • CORCEMA

    complete relaxation and conformational exchange matrix

  •  
  • 2D

    two dimensional

  •  
  • HS

    heparan sulfate

  •  
  • HSQC

    heteronuclear single-quantum coherence

  •  
  • HSV-1

    herpes simplex virus type-1

  •  
  • NOE

    nuclear Overhauser effect

  •  
  • 3OSG

    3-O-sulfated glucosamine

  •  
  • 3-OST

    3-O-sulfotransferase

  •  
  • SAX

    strong anion-exchange

  •  
  • STD

    saturation transfer difference

  •  
  • STD-AF

    STD amplification factor

  •  
  • tr-NOE

    transferred NOE

  •  
  • tr-NOESY

    transferred NOESY

AUTHOR CONTRIBUTION

Marco Guerrini, Christian Viskov and Giangiacomo Torri designed the study. Stefano Elli performed computational studies. Pierre Mourier isolated, purified and characterized the octasaccharide. Christian Boudier performed fluorescence titration experiments. Davide Gaudesi and Marco Guerrini performed the NMR experiments. Marco Guerrini, Christian Viskov, Timothy Rudd, Stefano Elli, Benito Casu and Giangiacomo Torri analysed the data and reviewed the paper before submission. Marco Guerrini and Christian Viskov wrote the paper.

We thank Dr G. Cassinelli and Dr Edwin A. Yates for useful discussion.

FUNDING

This work was supported by Sanofi-Aventis and the Ronzoui Foundation.

References

References
1
Lindahl
U.
Li
J. P.
Interactions between heparan sulfate and proteins-design and functional implications
Int. Rev. Cell. Mol. Biol.
2009
, vol. 
276
 (pg. 
105
-
159
)
2
Casu
B.
Naggi
A.
Torri
G.
Heparin-derived heparan sulfate mimics to modulate heparan sulfate-protein interaction in inflammation and cancer
Matrix Biol.
2010
, vol. 
29
 (pg. 
442
-
452
)
3
Casu
B.
Garg
H. G.
Linhardt
R. J.
Hales
C. A.
Structure and active domains of heparin
Chemistry and Biology of Heparin and Heparan Sulfate
2005
Oxford
Elsevier
(pg. 
1
-
28
)
4
Carlsson
P.
Presto
J.
Spillmann
D.
Lindahl
U.
Kjellen
L.
Heparin/heparan sulfate biosynthesis: processive formation of N-sulfated domains
J. Biol. Chem.
2008
, vol. 
283
 (pg. 
20008
-
20014
)
5
Linhardt
R. J.
Wang
H. M.
Loganathan
D.
Bae
J. H.
Search for the heparin antithrombin III-binding site precursor
J. Biol. Chem.
1992
, vol. 
267
 (pg. 
2380
-
2387
)
6
Zhang
L.
Schwartz
J. J.
Miller
J.
Liu
J.
Fritze
L. M.
Shworak
N. W.
Rosenberg
R. D.
The retinoic acid and cAMP-dependent up-regulation of 3-O-sulfotransferase-1 leads to a dramatic augmentation of anticoagulantly active heparan sulfate biosynthesis in F9 embryonal carcinoma cells
J. Biol. Chem.
1998
, vol. 
273
 (pg. 
27998
-
28003
)
7
Kusche
M.
Backstrom
G.
Riesenfeld
J.
Petitou
M.
Choay
J.
Lindahl
U.
Biosynthesis of heparin. O-sulfation of the antithrombin-binding region
J. Biol. Chem.
1988
, vol. 
263
 (pg. 
15474
-
15484
)
8
Shworak
N. W.
Liu
J.
Petros
L. M.
Zhang
L.
Kobayashi
M.
Copeland
N. G.
Jenkins
N. A.
Rosenberg
R. D.
Multiple isoforms of heparan sulfate D-glucosaminyl 3-O-sulfotransferase. Isolation, characterization, and expression of human cdnas and identification of distinct genomic loci
J. Biol. Chem.
1999
, vol. 
274
 (pg. 
5170
-
5184
)
9
Xia
G.
Chen
J.
Tiwari
V.
Ju
W.
Li
J. P.
Malmstrom
A.
Shukla
D.
Liu
J.
Heparan sulfate 3-O-sulfotransferase isoform 5 generates both an antithrombin-binding site and an entry receptor for herpes simplex virus, type 1
J. Biol. Chem.
2002
, vol. 
277
 (pg. 
37912
-
37919
)
10
Liu
J.
Shworak
N. W.
Sinay
P.
Schwartz
J. J.
Zhang
L.
Fritze
L. M.
Rosenberg
R. D.
Expression of heparan sulfate D-glucosaminyl 3-O-sulfotransferase isoforms reveals novel substrate specificities
J. Biol. Chem.
1999
, vol. 
274
 (pg. 
5185
-
5192
)
11
Xu
D.
Moon
A. F.
Song
D.
Pedersen
L. C.
Liu
J.
Engineering sulfotransferases to modify heparan sulfate
Nat. Chem. Biol.
2008
, vol. 
4
 (pg. 
200
-
202
)
12
Shukla
D.
Liu
J.
Blaiklock
P.
Shworak
N. W.
Bai
X.
Esko
J. D.
Cohen
G. H.
Eisenberg
R. J.
Rosenberg
R. D.
Spear
P. G.
A novel role for 3-O-sulfated heparan sulfate in herpes simplex virus 1 entry
Cell
1999
, vol. 
99
 (pg. 
13
-
22
)
13
Edge
A. S.
Spiro
R. G.
Characterization of novel sequences containing 3-O-sulfated glucosamine in glomerular basement membrane heparan sulfate and localization of sulfated disaccharides to a peripheral domain
J. Biol. Chem.
1990
, vol. 
265
 (pg. 
15874
-
15881
)
14
Lindahl
B.
Eriksson
L.
Lindahl
U.
Structure of heparan sulphate from human brain, with special regard to Alzheimer's disease
Biochem. J.
1995
, vol. 
306
 (pg. 
177
-
184
)
15
de Agostini
A. I.
Dong
J. C.
de Vantery Arrighi
C.
Ramus
M. A.
Dentand-Quadri
I.
Thalmann
S.
Ventura
P.
Ibecheole
V.
Monge
F.
Fischer
A. M.
, et al. 
Human follicular fluid heparan sulfate contains abundant 3-O-sulfated chains with anticoagulant activity
J. Biol. Chem.
2008
, vol. 
283
 (pg. 
28115
-
28124
)
16
Mochizuki
H.
Yoshida
K.
Shibata
Y.
Kimata
K.
Tetrasulfated disaccharide unit in heparan sulfate: enzymatic formation and tissue distribution
J. Biol. Chem.
2008
, vol. 
283
 (pg. 
31237
-
31245
)
17
Shriver
Z.
Raman
R.
Venkataraman
G.
Drummond
K.
Turnbull
J.
Toida
T.
Linhardt
R.
Biemann
K.
Sasisekharan
R.
Sequencing of 3-O sulfate containing heparin decasaccharides with a partial antithrombin III binding site
Proc. Natl. Acad. Sci. U.S.A.
2000
, vol. 
97
 (pg. 
10359
-
10364
)
18
Toida
T.
Hileman
R. E.
Smith
A. E.
Vlahova
P. I.
Linhardt
R. J.
Enzymatic preparation of heparin oligosaccharides containing antithrombin III binding sites
J. Biol. Chem.
1996
, vol. 
271
 (pg. 
32040
-
32047
)
19
van Boeckel
C. A. A.
Beetz
T.
van Aelst
S. F.
Synthesis of a potent antithrombin activating pentasaccharide: a new heparin-like fragment containing two 3-O-sulphated glucosamines
Tetrahedron Lett.
1988
, vol. 
29
 (pg. 
803
-
806
)
20
Guerrini
M.
Elli
S.
Gaudesi
D.
Torri
G.
Casu
B.
Mourier
P.
Herman
F.
Boudier
C.
Lorenz
M.
Viskov
C.
Effects on molecular conformation and anticoagulant activities of 1,6-anhydrosugars at the reducing terminal of antithrombin-binding octasaccharides isolated from low-molecular-weight heparin enoxaparin
J. Med. Chem.
2010
, vol. 
53
 (pg. 
8030
-
8040
)
21
Guerrini
M.
Guglieri
S.
Casu
B.
Torri
G.
Mourier
P.
Boudier
C.
Viskov
C.
Antithrombin-binding octasaccharides and role of extensions of the active pentasaccharide sequence in the specificity and strength of interaction. Evidence for very high affinity induced by an unusual glucuronic acid residue
J. Biol. Chem.
2008
, vol. 
283
 (pg. 
26662
-
26675
)
22
Guerrini
M.
Guglieri
S.
Beccati
D.
Torri
G.
Viskov
C.
Mourier
P.
Conformational transitions induced in heparin octasaccharides by binding with antithrombin III
Biochem. J.
2006
, vol. 
399
 (pg. 
191
-
198
)
23
Mourier
P. A.
Viskov
C.
Chromatographic analysis and sequencing approach of heparin oligosaccharides using cetyltrimethylammonium dynamically coated stationary phases
Anal. Biochem.
2004
, vol. 
332
 (pg. 
299
-
313
)
24
Olson
S. T.
Shore
J. D.
Binding of high affinity heparin to antithrombin III. Characterization of the protein fluorescence enhancement
J. Biol. Chem.
1981
, vol. 
256
 (pg. 
11065
-
11072
)
25
Lin
P.
Sinha
U.
Betz
A.
Antithrombin binding of low molecular weight heparins and inhibition of factor Xa
Biochim. Biophys. Acta
2001
, vol. 
1526
 (pg. 
105
-
113
)
26
Jin
L.
Abrahams
J. P.
Skinner
R.
Petitou
M.
Pike
R. N.
Carrell
R. W.
The anticoagulant activation of antithrombin by heparin
Proc. Natl. Acad. Sci. U.S.A.
1997
, vol. 
94
 (pg. 
14683
-
14688
)
27
Moseley
H. N.
Curto
E. V.
Krishna
N. R.
Complete relaxation and conformational exchange matrix (CORCEMA) analysis of NOESY spectra of interacting systems: two-dimensional transferred NOESY
J. Magn. Reson. B
1995
, vol. 
108
 (pg. 
243
-
261
)
28
Ferro
D. R.
Provasoli
A.
Ragazzi
M.
Torri
G.
Casu
B.
Gatti
G.
Jacquinet
J. C.
Sinay
P.
Petitou
M.
Choay
J.
Evidence for conformational equilibrium of the sulfated L-iduronate residue in heparin and in synthetic heparin mono- and oligo-saccharides: NMR and force-field studies
J. Am. Chem. Soc.
1986
, vol. 
108
 (pg. 
6773
-
6778
)
29
Ferro
D. R.
Provasoli
A.
Ragazzi
M.
Casu
B.
Torri
G.
Bossennec
V.
Perly
B.
Sinay
P.
Petitou
M.
Choay
J.
Conformer populations of L-iduronic acid residues in glycosaminoglycan sequences
Carbohydr. Res.
1990
, vol. 
195
 (pg. 
157
-
167
)
30
Hricovini
M.
Guerrini
M.
Bisio
A.
Torri
G.
Petitou
M.
Casu
B.
Conformation of heparin pentasaccharide bound to antithrombin III
Biochem. J.
2001
, vol. 
359
 (pg. 
265
-
272
)
31
Langeslay
D. J.
Young
R. P.
Beni
S.
Beecher
C. N.
Mueller
L. J.
Larive
C. K.
Sulfamate proton solvent exchange in heparin oligosaccharides: evidence for a persistent hydrogen bond in the antithrombin-binding pentasaccharide Arixtra
Glycobiology
2012
, vol. 
22
 (pg. 
1173
-
1182
)
32
Mayer
M.
Meyer
B.
Characterization of ligand binding by saturation transfer difference NMR spectroscopy
Angew. Chem. Int. Ed.
1999
, vol. 
38
 (pg. 
1784
-
1788
)
33
Mayer
T.
Meyer
M.
Janning
A.
Schiedel
A. C.
Barnekow
A.
A mutant form of the rho protein can restore stress fibers and adhesion plaques in v-src transformed fibroblasts
Oncogene
1999
, vol. 
18
 (pg. 
2117
-
2128
)
34
Angulo
J.
Hricovini
M.
Gairi
M.
Guerrini
M.
de Paz
J. L.
Ojeda
R.
Martin-Lomas
M.
Nieto
P. M.
Dynamic properties of biologically active synthetic heparin-like hexasaccharides
Glycobiology
2005
, vol. 
15
 (pg. 
1008
-
1015
)
35
Ten Dam
G. B.
Kurup
S.
van de Westerlo
E. M.
Versteeg
E. M.
Lindahl
U.
Spillmann
D.
van Kuppevelt
T. H.
3-O-sulfated oligosaccharide structures are recognized by anti-heparan sulfate antibody HS4C3
J. Biol. Chem.
2006
, vol. 
281
 (pg. 
4654
-
4662
)
36
Zhang
L.
Beeler
D. L.
Lawrence
R.
Lech
M.
Liu
J.
Davis
J. C.
Shriver
Z.
Sasisekharan
R.
Rosenberg
R. D.
6-O-sulfotransferase-1 represents a critical enzyme in the anticoagulant heparan sulfate biosynthetic pathway
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
42311
-
42321
)
37
Casu
B.
Torri
G.
Structural characterization of low molecular weight heparins
Semin. Thromb. Hemostasis
1999
, vol. 
25
 (pg. 
17
-
25
)
38
Viskov
C.
Just
M.
Laux
V.
Mourier
P.
Lorenz
M.
Description of the chemical and pharmacological characteristics of a new hemisynthetic ultra-low-molecular-weight heparin, AVE5026
J. Thromb. Haemostasis
2009
, vol. 
7
 (pg. 
1143
-
1151
)
39
Grootenhuis
P. D. J.
Van Boeckel
C. A. A.
Constructing a molecular model of the interaction between antithrombin III and a potent heparin analog
J. Am. Chem. Soc.
1991
, vol. 
113
 (pg. 
2743
-
2747
)
40
Moon
A. F.
Xu
Y.
Woody
S. M.
Krahn
J. M.
Linhardt
R. J.
Liu
J.
Pedersen
L. C.
Dissecting the substrate recognition of 3-O-sulfotransferase for the biosynthesis of anticoagulant heparin
Proc. Natl. Acad. Sci. U.S.A.
2012
, vol. 
109
 (pg. 
5265
-
5270
)
41
Moon
A. F.
Edavettal
S. C.
Krahn
J. M.
Munoz
E. M.
Negishi
M.
Linhardt
R. J.
Liu
J.
Pedersen
L. C.
Structural analysis of the sulfotransferase (3-O-sulfotransferase isoform 3) involved in the biosynthesis of an entry receptor for herpes simplex virus 1
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
45185
-
45193
)
42
Edavettal
S. C.
Lee
K. A.
Negishi
M.
Linhardt
R. J.
Liu
J.
Pedersen
L. C.
Crystal structure and mutational analysis of heparan sulfate 3-O-sulfotransferase isoform 1
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
25789
-
25797
)
43
Bisio
A.
Vecchietti
D.
Citterio
L.
Guerrini
M.
Raman
R.
Bertini
S.
Eisele
G.
Naggi
A.
Sasisekharan
R.
Torri
G.
Structural features of low-molecularweight heparins affecting their affinity to antithrombin
Thromb. Haemostasis
2009
, vol. 
102
 (pg. 
865
-
873
)

Author notes

2

Present address: Beamline 23 (Circular Dichroism), Diamond Light Source Ltd, Diamond House, Harwell Science and Innovation Campus, Harwell, Oxfordshire OX11 0DE, U.K.

3

Present address: Dulbecco Telethon Institute, Biomolecular NMR Laboratory, c/o Ospedale S. Raffaele, via Olgettina 58, 20132 Milan, Italy.

Supplementary data