Use of biosynthetic enzymes in heparin and heparan sulfate synthesis

Accepted Manuscript Use of biosynthetic enzymes in heparin and heparan sulfate synthesis Elizabeth P. Chappell, Jian Liu PII: DOI: Reference:

S0968-0896(12)00946-7 http://dx.doi.org/10.1016/j.bmc.2012.11.053 BMC 10460

To appear in:

Bioorganic & Medicinal Chemistry

Please cite this article as: Chappell, E.P., Liu, J., Use of biosynthetic enzymes in heparin and heparan sulfate synthesis, Bioorganic & Medicinal Chemistry (2012), doi: http://dx.doi.org/10.1016/j.bmc.2012.11.053

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Use of biosynthetic enzymes in heparin and heparan sulfate synthesis

Elizabeth P. Chappell and Jian Liu*

Division of Chemical Biology and Medicinal Chemistry, Eshelman School of Pharmacy, University of North Carolina, Chapel Hill, North Carolina 27599

* To whom correspondence should be addressed: Rm 303, Beard Hall, University of North Carolina, Chapel Hill, NC 27599. Tel.: 919-843-6511; E-mail: [email protected]

This work was supported by the National Institutes of Health [R01HL094463 to J.L and F31AG040927-01 to E.P.C.].

1

Summary Heparan sulfate and heparin are highly sulfated polysaccharides consisting of repeating disaccharide units of glucuronic acid or iduronic acid that is linked to glucosamine. Heparan sulfate displays a range of biological functions, and heparin is a widely used anticoagulant drug in hospitals. It has been known to organic chemists that the chemical synthesis of heparan sulfate and heparin oligosaccharides is extremely difficult. Recent advances in the study of the biosynthesis of heparan sulfate/heparin offer a chemoenzymatic approach to synthesize heparan sulfate and heparin. Compared to chemical synthesis, the chemoenzymatic method shortens the synthesis and improves the product yields significantly, providing an excellent opportunity to advance the understanding of the structure and function relationships of heparan sulfate. In this review, we attempt to summarize the progress of the chemoenzymatic synthetic method and its application in heparan sulfate and heparin research.

2

1. Introduction Heparan sulfate (HS) is a widely expressed carbohydrate. As the body’s most negatively charged molecule, it interacts with numerous proteins to influence biological functions ranging from development and coagulation to inflammation and cancer metastasis. Heparin and HS have very similar structures; however, heparin refers to a special form of HS that has more sulfo groups and a higher level of iduronic acid residues. The bioactivity of HS is dependent on its structure, namely the location of electronegative sulfo groups along its backbone and the presence of iduronic acid (IdoA), glucuronic acid (GlcA) and glucosamine (GlcN) residues (Figure 1). HS exists on the surface of animal cells and within the extracellular matrix as a proteoglycan consisting of long carbohydrate chains attached to a core protein. Cell surface HS proteoglycans include syndecans and glypicans; perlecan and agrin are the primary extracellular examples.1-3 The HS chains are composed of repeating disaccharide units of uronic acid and glucosamine. The uronic acid is present as either glucuronic or iduronic acid and can be sulfated at carbon 2 (2-Osulfation). The glucosamine can be 6-O- and 3-O-sulfated, and its amine group can be acetylated, sulfated or unsubstituted (Figure 1). The biological functions of HS proteoglycans are dominated by the HS side chains. This broad structural variation in the location of negative groups, in addition to variation in length and glycosidic bond position, allows HS to interact with different binding protein partners to display many biological functions. Although nonspecific ionic interactions between HS and proteins exist, the binding of HS to proteins can be specific. Thus, the preparation of unique HS chains with defined sulfation patterns and length is highly desirable, as they allow researchers to investigate the substrate specificity of HS-protein interactions and provide numerous therapeutic opportunities. 3

2. Development of the chemoenzymatic method The complexity and size of HS renders chemical synthesis quite challenging. Chemical synthesis requires the coupling of monosaccharide or short oligosaccharides to a larger fragment followed by the installation of sulfo groups. The synthesis typically demands multiple protection/deprotection steps. For example, the chemical synthesis of fondaparinux, a pentasaccharide mimicking the antithrombin-binding region of heparin, is achieved by linking diand trisaccharides in which the eventual -OSO3H positions are fully protected by acetyl groups and the -OH groups of the final product are protected by benzyl ethers.4 Hydrolysis of the acetyl and benzyl ether moieties, followed by selective N-sulfation, is required to yield the desired product. The synthesis of fondaparinux requires about 50 steps with a yield of only ~0.1%;5 the long synthetic route and necessary purification steps make it an impractical method in an academic lab to routinely carry out such complicated syntheses. A chemoenzymatic method utilizing HS biosynthetic enzymes has emerged as a promising alternative to the chemical synthesis of HS. In 2003, the Rosenberg group reported the synthesis of antithrombin (AT)-binding HS polysaccharides using cloned enzymes that mimicked those present in cells.6 Binding to antithrombin was achieved by both classical and non-classical (lacking IdoA2S) HS. Using K5 polysaccharide from E. coli as a starting material, synthesis of an AT-binding pentasaccharide was achieved in just six steps with a yield at least two-fold greater than that using chemical synthesis.7 The work from Rosenberg’s group clearly demonstrated the feasibility of an enzymatic approach to synthesize HS. However, the synthesis was completed in microgram scale, which is not nearly adequate for satisfying the needs of biological studies. Synthesis of structurally heterogeneous polysaccharides using the chemoenzymatic method in milligram scales has also been reported. Early work focused primarily on the preparation of long polysaccharide chains using heparosan as a starting material, and libraries 4

of HS polysaccharides with specific sulfation types and IdoA content were successfully created in a limited number of reaction and purification steps.8, 9 The successful expression of the enzymes in E. coli played an essential role in the larger-scale synthesis of those early studies. Recent breakthroughs have enabled the preparation of sulfated size-defined HS oligosaccharides in milligram quantities.10 These oligosaccharide substrates provide a new frontier in medicinal chemistry for addressing biological questions involving HS. The preparation of structures with all possible combination of sulfations at a given position (N-, 2-O-, 6-O- or 3-

O-) remains challenging, but deeper understanding of the actions of each of the involved enzymes has enabled the production of many HS structures.

3. Heparan sulfate biosynthesis HS biosynthesis occurs in the Golgi apparatus and is carried out by a series of enzymes (Figure 1). Recombinant versions of these enzymes have been successfully expressed and purified in E. coli, yeast or in insect cells via a baculovirus expression system. E. coli expression allows for inexpensive and easy expression with low incidence of contamination. The cells can be spun down for simple purification, whereas yeast and insect cell expression requires purifying large volumes of cell medium. However, enzymes from yeast and insects cells have the benefit of being glycosylated and endotoxin free, and enzymes expressed in yeast exhibit better thermostability. The different expression systems produce enzymes with comparable activity. Table 1 shows the HS biosynthetic enzymes and some bacterial enzymes utilized in chemoenzymatic syntheses.

3.1. Use of biosynthetic enzymes Synthesis of HS has two main components: chain elongation and modification of the individual sugars. In vivo, the HS chain is elongated by the Exostosin genes, Ext1 and Ext2.11 This process can be mimicked in vitro by two bacterial glycosyltransferases: N5

acetylglucosaminyl transferase from the E. coli K5 strain, known as KfiA,12 and heparosan synthase-2 from Pasteurella multocida, or pmHS2.13 KfiA and pmHS2 are incubated with uridine 5’-diphospho-N-trifluoroacetylglucosamine (UDP-NTFA) and UDP-GlcA, respectively; with each cycle of incubation, the oligosaccharide is elongated by one sugar unit.14 GlcNTFA is used because it is easily chemically deacetylated to GlcNH2 for later N-sulfation. The N-deacetylase/N-sulfotransferase (NDST) enzyme is the first to modify an intact HS chain during biosynthesis, and its action is believed to direct the location of all subsequent sulfation reactions.15 The 325-aa C-terminal region (constituting the N-sulfotransferase domain) of NDST is commonly expressed and used for HS synthesis in vitro following chemical deacetylation,16 although recent studies have focused on expression of the entire NDST enzyme.17 Sulfation reactions are carried out by incubating HS with a sulfotransferase and 3’phosphoadenosine 5’-phosphosulfate (PAPS), a natural sulfate donor. After N-sulfation, C5-epimerase (C5-epi) converts some D-glucuronic acid residues to Liduronic acid by altering the configuration of carbon 5.18 A series of O-sulfotransferases then sulfate their respective positions on HS. Heparan sulfate 2-O-sulfotransferase (2-OST) catalyzes the transfer of an -OSO3H group from PAPS to IdoA or GlcA. It is present in one isoform and has approximately five-fold greater affinity for IdoA than for GlcA,19 although mutational analyses have suggested that the preference for IdoA over GlcA can be controlled through site-specific mutations.20 6-OST sulfates both GlcNAc and GlcNS to form GlcNAc6S and GlcNS6S, respectively. 6-OST isoforms, of which there are three, appear to sulfate the same substrates;21 however, placement of 6-O-sulfo groups in oligosaccharides can be controlled somewhat by the enzymatic reaction time and by elongating oligosaccharides already containing 6-O-sulfated glucosamine.22 A combination of 6-OST-1 and -3 was demonstrated to prefer GlcNAc residues close to the reducing end of oligosaccharide substrates, but placement of a single 6-O-sulfo group in an oligosaccharide remains a challenge.22, 23 6

3-OST adds a sulfo group to the 3-OH position of GlcN residues and is present in seven isoforms. In endogenous HS, 3-O-sulfation is a relatively rare modification, sulfating only a small component of disaccharides.24 However, the presence of 3-O-sulfation is critical for multiple types of HS activity, including binding to antithrombin, gD (a herpes simplex virus type 1 entry receptor) and Stabilin-1/2 (heparin clearance receptors).25, 26

3.2. Enzyme substrate recognition Some of the HS biosynthetic enzymes exhibit interesting substrate recognition patterns. Characterization of NDST-1 by Sheng et al. uncovered a unique substrate specificity for this enzyme.27 Treatment of a synthetic dodecasaccharide substrate with NDST-1 produced a variety of N-sulfated products containing clusters of GlcNS, suggesting that NDST-1 binds to HS at a random position, converts consecutive GlcA to GlcNS from the non-reducing to reducing end, then releases the substrate when it is five sugars away from the reducing end. There are a total of four isoforms of NDST; NDST-2, which is highly expressed in mast cells, is proposed to be involved in the synthesis of highly-sulfated heparin but not HS.28 This was principally confirmed by the absence of heparin in mast cells from NDST-2 knockout mice.29 The Ndeacetylase and N-sulfotransferase activities of NDST-2 through -4 have been investigated, but their use in HS synthesis specifically has not been fully explored.30 C5-epimerase was recently shown to exhibit a biphasic catalytic mode: depending on the substitution groups of the surrounding saccharide residues, the epimerization reaction is either reversible or irreversible. Using structurally defined oligosaccharides, Sheng et al. identified that C5-epi will act on a GlcA residue if the residue immediately upstream (towards the non-reducing end) is a GlcNS residue. If the residue three sugars upstream is GlcNS, GlcNH2 or not present, the reaction is reversible; if it is GlcNAc, the reaction is irreversible (Figure 2A).31 This finding will enable researchers to “lock” IdoA sugars in place and synthesize pure HS oligosaccharides containing IdoA during in vitro synthesis. 7

The 3-OST isoforms exhibit greater than 60% homology in their sulfotransferase domains,15 and the substrate recognition of the 3-OSTs generally falls into one of three types, depending on the sugar linked to the non-reducing side of the glucosamine to be modified. 3OST-1 will transfer a sulfo group to a GlcNS that is linked to a GlcA or IdoA at the non-reducing end. 3-OST-5 sulfates a GlcNS that is linked to an IdoA2S, GlcA or IdoA. The remaining 3OSTs (-2, -3A, -3B, -4 and -6) will sulfate a GlcNS linked to an IdoA2S at the non-reducing end (Figure 2B). These different substrate specificities allow researchers to place 3-O-sulfo groups in specific locations depending on the identity of the neighboring residue. Based on their ease of production in E. coli, 3-OST-1, -3 and -5 are most commonly used in the chemoenzymatic affinity activity; 3OST-3 and -5 will prepare HS that binds to gD.25

4. The chemoenzymatic method in practice: synthesis of ultralow-molecular-weight heparins Short homogenous heparin structures such as fondaparinux have several advantages over full-length unfractionated heparin, such as reduced heparin-induced thrombocytopenia side effect and prolonged anticoagulant activity. These attributes, along with problems from naturally sourced heparin such as the distribution of contaminated heparin in 2007, have made new synthetic heparins highly desirable.32 Two ultralow-molecular-weight heparins (ULMWH constructs 1 and 2) were successfully prepared using the chemoenzymatic synthetic method as reported by Xu et al.10 Using a disaccharide isolated from nitrous acid-degraded heparosan as a starting material, KfiA and pmHS2 were incubated with UDP-GlcNTFA and UDP-GlcA, respectively, to elongate the disaccharides to hexasaccharides. One pool of these hexasaccharides was treated with KfiA and UDP-GlcNAc and the second was treated with KfiA and UDP-GlcNTFA; this provided a GlcNAc at the non-reducing end of the final construct 1 and a GlcNS in construct 2. Thus, after chemical deacetylation and successive incubations with NST, C5-epi, 2-OST, 6-OST and 38

OST-1, ULMWH 1 contained the AT-binding motif of porcine heparin and ULMWH 2 contained that of bovine heparin (Figure 3).33 The structures were proven by ESI-MS and NMR, and the two constructs and fondaparinux were shown to have similar anti-factor Xa activity in vitro and similar clearance profiles in rabbits. This study has proven that the chemoenzymatic method can be harnessed to produce heparin constructs with potent in vitro and in vivo activity. Compared with the purely chemical approach, the chemoenzymatic method has shortened the synthesis to 10 to 12 steps from ~50 steps with significant improvement in the recovery yield. There is no doubt that the chemoenzymatic approach has simplified the synthesis of HS oligosaccharides.

5. Using the chemoenzymatic method to answer biological questions Chemoenzymatic synthesis has enabled researchers to explore the bioactivity of structurally defined HS oligosaccharides. This technology has potential therapeutic applications (such as new anticoagulant drugs) as well as basic research benefits, including detailed structure-function studies and the elucidation of heparan sulfate’s role in various physiological processes. Here, we describe clearance receptor binding studies and X-ray crystallography of a ternary complex with an HS biosynthetic enzyme as recent examples.

5.1. Stabilin-1/2 Stabilin-2 was recently identified as a systemic clearance receptor for heparin.34 Stabilin2 is present in sinusoidal endothelial cells of the liver, kidney and spleen, and it takes up glycosaminoglycans for digestion via clathrin pit-mediated endocytosis. Full-length unfractionated heparin and a portion of low-molecular-weight heparin are known to be metabolized by the liver; shorter heparin drugs such as fondaparinux are cleared by the kidneys. Initial studies identified that heparin and low-molecular-weight heparin were able to

9

bind to this receptor and were internalized by Stabilin-2 expressing cells; however, the structures required for HS binding to Stabilin-2 were unknown.35 To determine which HS sulfation types were required for Stabilin binding-2, nine radiolabeled polysaccharide substrates were prepared using the chemoenzymatic method, and their uptake by Stabilin-1 and Stabilin-2-expressing cells was measured. While N-sulfation was required and 2-O- and 6-O-sulfation improved cellular uptake, much greater internalization was seen when 3-O-sulfation was added to the substrates.26 Based on this conclusion, a series of oligosaccharides ranging from 7 to 19 sugar units in length were synthesized using glycosyltransferases and modified with N-, 2-O-, 6-O- and 3-Osulfation. An oligosaccharide of at least a decasaccharide was determined necessary for robust HS binding. This information should be beneficial for designing heparin drugs with controlled clearance rates and routes of elimination.

5.2. Cocrystal structure of 3-OST-1, a heptasaccharide and PAP Crystal structures of several of the HS sulfotransferases have been solved,20, 36-39 providing an intricate understanding of the interface between residues of the enzyme active site and HS sugar units. A structurally defined heptasaccharide was recently synthesized using bacterial glycosyltransferases and was cocrystallized with the catalytic domain of 3-OST-1, providing insight into a binding interaction that is critical for the production of antithrombinbinding HS.40 The heptasaccharide (having a structure of GlcNAc6S-GlcA-GlcNS6S*-IdoA2SGlcNS6S-GlcA-AnMan, *: 3-O-sulfation acceptor site) was shown to interact with the binding cleft of 3-OST-1 through a series of hydrogen bonds. Five of the sugars in the pentasaccharide, including the acceptor sugar and the two adjacent on either side, interacted with this cleft (Figure 4). Comparing this new cocrystal structure with that of 3-OST-3 suggests that the 3-

10

OSTs bind HS in different orientations with a complex substrate recognition system depending on multiple saccharide units.38

6. Conclusions Since its inception, the chemoenzymatic method has evolved so that the synthesis of HS oligosaccharides over 20 units long with specific functional groups is now possible.41 Deeper understanding of the HS-modifying enzymes has enabled researchers to place IdoA residues and sulfo groups in specific locations. Certain challenges remain: for example, the control of the placement of IdoA2S has not been fully established. Still, our present knowledge to date enables the production of HS structures that have broad utility in medical and basic scientific research. In addition, enzymatic synthesis is typically limited to synthesizing naturally occurring structures due to the substrate specificities of HS biosynthetic enzymes. We anticipate that a combination of chemoenzymatic and a pure chemical synthetic method may expand the flexibility of the synthesis of HS oligosaccharides.

Figure Legends Figure 1. HS biosynthetic pathway and structure. The generic structure of HS is shown in the box at left. N-deacetylase/N-sulfotransferase replaces N-acetyl groups with N-sulfo groups. C5epimerase converts glucuronic acid to iduronic acid. 2-O-, 6-O- and 3-O-sulfotransferase add sulfation at their respective positions. Figure 2. Substrate recognition of C5-epimerase and 3OST. (A) The epimerization reaction of C5-epi is reversible or irreversible depending on the GlcN three sugars towards the nonreducing end of the glucuronic acid to be modified. The asterisk indicates a GlcA that could be epimerized if the sugar to the immediate left is GlcNS. (B) Substrates of the seven 3OST isoforms. The position that will carry the 3-O-sulfo group is shown in blue. Figure 3. Synthesis of two antithrombin-binding oligosaccharides. Ultralow-molecular-weight construct 1 contains the porcine AT-binding pentasaccharide and construct 2 contains the bovine pentasaccharide. Reaction steps include: a) KfiA, UDP-GlcNTFA, b) pmHS2, UDP-GlcA, c) KfiA, UDP-GlcNAc, d) triethylamine/CH3OH/H2O, e) NST, PAPS, f) C5-epi/2-OST, PAPS, g) 6-OST, PAPS and h) 3-OST-1, PAPS. Figure 4. Interactions between 3-OST-1 and a synthetic heptasaccharide substrate. Hydrogen bonds are indicated by dashed lines. The arrow indicates the 3-O-sulfation site. 11

Table 1. List of HS biosynthetic enzymes Enzymes

Abbreviated Enzymatic function Names

Expression system

N-deacetylase/Nsulfotransferase -1 N-deacetylase/Nsulfotransferase -2 N-deacetylase/Nsulfotransferase -3 N-deacetylase/Nsulfotransferase -4 N-sulfotransferase1

NDST-1

Converts a GlcNAc to a GlcNS residue

NDST-2 NDST-3

Similar to NDST-1, but prone to synthesize a long cluster of GlcA-GlcNS repeating domains Similar to NDST-1

E. coli (36), S. cerevisiae (17) E. coli (42), insect cells (7)

NDST-4

Similar to NDST-1

NST

Converts a GlcNH2 to a GlcNS residue

C5-epimerase

C5-epi

Converts a GlcA to an IdoA residue

2-O-sulfotransferase

2-OST

Introduces a sulfo group to the 2-OH position of an IdoA or a GlcA residue

6-O-sulfotransferase 1

6-OST-1

Introduces a sulfo group to the 6-OH position of a GlcNS or a GlcNAc residue

6-O-sulfotransferase 2

6-OST-2

Same as 6-OST-1

6-O-sulfotransferase 3

6-OST-3

Same as 6-OST-1

3-O-sulfotransferase 1

3-OST-1

3-O-sulfotransferase 2

3-OST-2

3-O-sulfotransferase 3A

3-OST-3A

3-O-sulfotransferase 3B

3-OST-3B

Introduces a sulfo group to the 3-OH position of a GlcNS±6S residue that is linked to a GlcA (or an IdoA) on the nonreducing end Introduces a sulfo group to the 3-OH position of a GlcNS±6S residue that is linked to an IdoA2S on the nonreducing end Introduces a sulfo group to the 3-OH position of a GlcNS±6S residue that is linked to an IdoA2S on the nonreducing end Same as 3-OST-3A

3-O-sulfotransferase 4

3-OST-4

Same as 3-OST-3A

insect cells (47)

3-O-sulfotransferase 5

3-OST-5

E. coli (39), insect cells (49)

3-O-sulfotransferase 6

3-OST-6

Has both 3-OST-1 and 3-OST-3A substrate specificities Same as 3-OST-3A

from P. multocida from E. coli

Transfers a GlcA and a GlcNAc (or a GlcNTFA) residue to the backbone Transfers a GlcNAc (or a GlcNTFA) residue to the backbone

E. coli (13)

2

pmHS2 KfiA2

1.

2.

E. coli (36), K. lactis (43) E. coli (44), insect cells (6) E. coli (45), K. lactis (43), insect cells (7) E. coli (45), K. lactis (43), insect cells (7) insect cells (7) E. coli (8), K. lactis (43) E. coli (37), K. lactis (43), insect cells (46) insect cells (47)

E. coli (38), insect cells (48) insect cells (47)

E. coli (12)

N-sulfotransferase is the protein that is composed of N-sulfotransferase domain of NDST-1. NST is an unnatural protein, and is used in the chemoenzymatic synthesis to convert a GlcNH2 residue to a GlcNS residue. Both pmHS2 and KfiA are bacteria enzymes, which do not belong to HS biosynthetic enzymes. However, they are used for building the backbone structure of HS during the chemoenzymatic synthesis.

12

References 1. Bernfield, M.; Gotte, M.; Park, P. W.; Reizes, O.; Fitzgerald, M. L.; Lincecum, J.; Zako, M. Annu Rev Biochem 1999, 68, 729. 2. Park, P. W.; Reizes, O.; Bernfield, M. J Biol Chem 2000, 275, 29923. 3. Iozzo, R. V. Annu Rev Biochem 1998, 67, 609. 4. Petitou, M.; van Boeckel, C. A. Angew Chem Int Ed Engl 2004, 43, 3118. 5. M. Petitou, J. C. J., J. Choay, J. C. Lormeau, M. Nassr. U.S. Patent 4,818,816 1989. 6. Kuberan, B.; Beeler, D. L.; Lech, M.; Wu, Z. L.; Rosenberg, R. D. J Biol Chem 2003, 278, 52613. 7. Kuberan, B.; Lech, M. Z.; Beeler, D. L.; Wu, Z. L.; Rosenberg, R. D. Nat Biotechnol 2003, 21, 1343. 8. Chen, J.; Jones, C. L.; Liu, J. Chem Biol 2007, 14, 986. 9. Zhang, Z.; McCallum, S. A.; Xie, J.; Nieto, L.; Corzana, F.; Jimenez-Barbero, J.; Chen, M.; Liu, J.; Linhardt, R. J. J Am Chem Soc 2008, 130, 12998. 10. Xu, Y.; Masuko, S.; Takieddin, M.; Xu, H.; Liu, R.; Jing, J.; Mousa, S. A.; Linhardt, R. J.; Liu, J. Science 2011, 334, 498. 11. McCormick, C.; Duncan, G.; Goutsos, K. T.; Tufaro, F. Proc Natl Acad Sci U S A 2000, 97, 668. 12. Chen, M.; Bridges, A.; Liu, J. Biochemistry 2006, 45, 12358. 13. Sismey-Ragatz, A. E.; Green, D. E.; Otto, N. J.; Rejzek, M.; Field, R. A.; DeAngelis, P. L. J Biol Chem 2007, 282, 28321. 14. Liu, R.; Xu, Y.; Chen, M.; Weiwer, M.; Zhou, X.; Bridges, A. S.; DeAngelis, P. L.; Zhang, Q.; Linhardt, R. J.; Liu, J. J Biol Chem 2010, 285, 34240. 15. Peterson, S.; Frick, A.; Liu, J. Nat Prod Rep 2009, 26, 610. 16. Sueyoshi, T.; Kakuta, Y.; Pedersen, L. C.; Wall, F. E.; Pedersen, L. G.; Negishi, M. FEBS Lett 1998, 433, 211. 17. Saribas, A. S.; Mobasseri, A.; Pristatsky, P.; Chen, X.; Barthelson, R.; Hakes, D.; Wang, J. Glycobiology 2004, 14, 1217. 18. Hagner-McWhirter, A.; Hannesson, H. H.; Campbell, P.; Westley, J.; Roden, L.; Lindahl, U.; Li, J. P. Glycobiology 2000, 10, 159. 19. Rong, J.; Habuchi, H.; Kimata, K.; Lindahl, U.; Kusche-Gullberg, M. Biochemistry 2001, 40, 5548. 20. Bethea, H. N.; Xu, D.; Liu, J.; Pedersen, L. C. Proc Natl Acad Sci U S A 2008, 105, 18724. 21. Smeds, E.; Habuchi, H.; Do, A. T.; Hjertson, E.; Grundberg, H.; Kimata, K.; Lindahl, U.; KuscheGullberg, M. Biochem J 2003, 372, 371. 22. Liu, R.; Liu, J. Biochemistry 2011, 50, 4382. 23. Jemth, P.; Smeds, E.; Do, A. T.; Habuchi, H.; Kimata, K.; Lindahl, U.; Kusche-Gullberg, M. J Biol Chem 2003, 278, 24371. 24. Esko, J. D.; Lindahl, U. J Clin Invest 2001, 108, 169. 25. 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. Cell 1999, 99, 13. 26. Pempe, E. H.; Xu, Y.; Gopalakrishnan, S.; Liu, J.; Harris, E. N. J Biol Chem 2012, 287, 20774. 27. Sheng, J.; Liu, R.; Xu, Y.; Liu, J. J Biol Chem 2011, 286, 19768. 28. de Paz, J. L.; Noti, C.; Seeberger, P. H. J Am Chem Soc 2006, 128, 2766. 29. Forsberg, E.; Pejler, G.; Ringvall, M.; Lunderius, C.; Tomasini-Johansson, B.; Kusche-Gullberg, M.; Eriksson, I.; Ledin, J.; Hellman, L.; Kjellen, L. Nature 1999, 400, 773. 30. Aikawa, J.; Grobe, K.; Tsujimoto, M.; Esko, J. D. J Biol Chem 2001, 276, 5876. 31. Sheng, J.; Xu, Y.; Dulaney, S. B.; Huang, X.; Liu, J. J Biol Chem 2012, 287, 20996. 32. Liu, H.; Zhang, Z.; Linhardt, R. J. Nat Prod Rep 2009, 26, 313. 13

33. Loganathan, D.; Wang, H. M.; Mallis, L. M.; Linhardt, R. J. Biochemistry 1990, 29, 4362. 34. Harris, E. N.; Weigel, J. A.; Weigel, P. H. J Biol Chem 2008, 283, 17341. 35. Harris, E. N.; Baggenstoss, B. A.; Weigel, P. H. Am J Physiol Gastrointest Liver Physiol 2009, 296, G1191. 36. Kakuta, Y.; Sueyoshi, T.; Negishi, M.; Pedersen, L. C. J Biol Chem 1999, 274, 10673. 37. Edavettal, S. C.; Lee, K. A.; Negishi, M.; Linhardt, R. J.; Liu, J.; Pedersen, L. C. J Biol Chem 2004, 279, 25789. 38. Moon, A. F.; Edavettal, S. C.; Krahn, J. M.; Munoz, E. M.; Negishi, M.; Linhardt, R. J.; Liu, J.; Pedersen, L. C. J Biol Chem 2004, 279, 45185. 39. Xu, D.; Moon, A. F.; Song, D.; Pedersen, L. C.; Liu, J. Nat Chem Biol 2008, 4, 200. 40. Moon, A. F.; Xu, Y.; Woody, S. M.; Krahn, J. M.; Linhardt, R. J.; Liu, J.; Pedersen, L. C. Proc Natl Acad Sci U S A 2012, 109, 5265. 41. Xu, Y.; Pempe, E. H.; Liu, J. J Biol Chem 2012, 287, 29054. 42. Duncan, M. B.; Liu, M.; Fox, C.; Liu, J. Biochem Biophys Res Commun 2006, 339, 1232. 43. Zhou, X.; Chandarajoti, K.; Pham, T. Q.; Liu, R.; Liu, J. Glycobiology 2011, 21, 771. 44. Munoz, E.; Xu, D.; Avci, F.; Kemp, M.; Liu, J.; Linhardt, R. J. Biochem Biophys Res Commun 2006, 339, 597. 45. Chen, J.; Avci, F. Y.; Munoz, E. M.; McDowell, L. M.; Chen, M.; Pedersen, L. C.; Zhang, L.; Linhardt, R. J.; Liu, J. J Biol Chem 2005, 280, 42817. 46. Myette, J. R.; Shriver, Z.; Liu, J.; Venkataraman, G.; Rosenberg, R.; Sasisekharan, R. Biochem Biophys Res Commun 2002, 290, 1206. 47. Mochizuki, H.; Yoshida, K.; Shibata, Y.; Kimata, K. J Biol Chem 2008, 283, 31237. 48. Liu, J.; Shriver, Z.; Blaiklock, P.; Yoshida, K.; Sasisekharan, R.; Rosenberg, R. D. J Biol Chem 1999, 274, 38155. 49. Chen, J.; Duncan, M. B.; Carrick, K.; Pope, R. M.; Liu, J. Glycobiology 2003, 13, 785.

14

Fig. 1

N-deacetylase/N-sulfotransferase, PAPS GlcA-GlcN

1. C5-epimerase 2. 2-O-sulfotransferase, PAPS IdoA-GlcN

R = -H or -SO3H R’= -H or -SO3H or -Ac

1. 6-O-sulfotransferase, PAPS 2. 3-O-sulfotransferase, PAPS

Fig. 2 A

* reversible reaction GlcNS, GlcNH2 or absent

epimerization site

irreversible reaction GlcNAc R= -SO3H or -Ac

B

GlcA-GlcNS 6S

IdoA-GlcNS 6S

3OST-5:

3OST-1:

3OST-2, -3A, -3B, -4, -6:

R= -SO3H or -H

IdoA2S-GlcNS 6S

Fig. 3

a, b

c

d, e, f, g, h

ULMWH construct 1 GlcNac6S-GlcA-GlcNS6S3S-IdoA2S-GlcNS6S-GlcA-AnMan

R=

d, e, a, f

d, e, g, h

ULMWH construct 2 GlcNS6S-GlcA-GlcNS6S3S-IdoA2S-GlcNS6S-GlcA-AnMan

Fig. 4

Graphical Abstract

2OST C5-epi NDST

6OST

3OST