Tyrosine sulfation on a PSGL-1 glycopeptide influences the reactivity of glycosyltransferases responsible for synthesis of the attached O-glycan [16]

Full text of DOI:10.1021/ja993820o

742
J. Am. Chem. Soc. 2000, 122, 742-743
Tyrosine Sulfation on a PSGL-1 Glycopeptide
Influences the Reactivity of Glycosyltransferases
Responsible for Synthesis of the Attached O-Glycan
Kathryn M. Koeller, Mark E. B. Smith, and Chi-Huey Wong*
Department of Chemistry, The Scripps Research Institute and
Skaggs Institute for Chemical Biology
10550 N. Torrey Pines Rd., La Jolla, California 92037
ReceiVed October 27, 1999
P-Selectin glycoprotein ligand-1 (PSGL-1) is the primary
counter-receptor for P-selectin during leukocyte extravasation in
the inflammatory response.1 Previously, it has been determined
that the O-glycan attached to threonine 16 and at least one site
of tyrosine sulfation within the N-terminal 19 amino acids are
required for optimal recognition of PSGL-1 by P-selectin.2 In fact,
the N-terminal glycosulfopeptide binds to P-selectin nearly as
efficiently as the full-length dimeric PSGL-1, and the sulfated
glycopeptide binds to the receptor approximately 105 times more
tightly than does the unsulfated glycopeptide ligand (Figure 1a).3
The O-linked glycan structure from PSGL-1 includes a terminal
sialyl Lewis x (sLex) tetrasaccharide extended from a core 2
glycan (Figure 1b).4
To further examine PSGL-1/P-selectin recognition, as well as
to investigate PSGL-1 biosynthesis, the chemoenzymatic synthesis
of a binding determinant of PSGL-1 was undertaken. The target
structure represents a minimal sequence containing a tyrosine
sulfate and the glycosylated threonine residue, corresponding to
amino acid residues 10-17 of the mature PSGL-1 protein. The
synthesis combined solution- and solid-phase methods to arrive
at a disaccharide-linked octapeptide in both sulfated and unsulfated
forms. Glycosyltransferase-catalyzed elaboration of the glycan was
then studied.5 Results indicate that sulfation on tyrosine influences
the reactivity of the glycosyltransferases responsible for the
synthesis of sLex on the attached O-glycan.
Figure 1. (a) The P-selectin/PSGL-1 interaction. (b) The N-terminal
structure of PSGL-1.
Scheme 1. Synthesis of the glyco-Thr Building Block
The synthetic strategy involved incorporation of a protected
disaccharide-linked threonine building block into solid-phase
peptide synthesis (Scheme 1). Disaccharide-threonine conjugate
3 was obtained by reaction of 16 with either glycosyl donor 2a7
under BF3-OEt2 catalysis or donor 2b8 with DMTST as the
activating agent.9 Conversion of 3 to 5 was accomplished by
standard synthetic manipulations.
Building block 5 was incorporated into glycopeptide 6 utilizing
a Rink Amide modified resin as the solid phase (Scheme 2).
Following N-terminal acetylation, sequence 6 was treated with
95% TFA, H2O, and ethane dithiol as a scavenger. These
conditions caused simultaneous liberation of the sequence from
the resin as the C-terminal amide and removal of the tBu ester
a Conditions: (a) BF3-OEt2, -30 ˚C, CH2Cl2 (100%); (b) DMTST,
CH2Cl2, 4 Å MS, 0 ˚C (75%); (c) AcOH/H2O, 45 ˚C; (d) Ac2O/Pyr; (e)
Zn dust, THF/AcOH/Ac2O (80%, 3 steps); (f) TFA/H2O (95:5) (100%).
and ether protecting groups. The crude peptide obtained was
initially purified by ether precipitation and small portions then
further purified using RP-HPLC to give sequence 7.10
Sulfation on the tyrosine residue of glycopeptide 7 was then
accomplished with sulfur trioxide-pyridine complex. A workup
protocol involving a methanol quench and immediate silica gel
chromatography11 allowed the sulfated sequence to be isolated
in much higher yields than has been previously reported.12
Saponification of the acetate esters then gave deprotected gly-
copeptide 8a for the subsequent glycosyltransferase-catalyzed
(1) (a) Kansas, G. S. Blood 1996, 88, 3259. (b) Rosen, S. D.; Bertozzi, C.
R. Curr. Biol. 1996, 6, 261.
(2) (a) Wilkins, P. P.; Moore, K. L.; McEver, R. P.; Cummings, R. D. J.
Biol. Chem. 1995, 270, 22677. (b) DeLuca, M.; Dunlop, L. C.; Andrews, R
K.; Flannery, J. V., Jr.; Ettling, R.; Cumming, D. A.; Veldman, G. M.; Berndt,
M. C. J. Biol. Chem. 1995, 270, 26734. (c) Sako, D.; Comess, K. M.; Barone,
K. M.; Camphausen, R. T.; Cumming, D. A.; Shaw, G. D. Cell 1995, 83,
323. (d) Pouyani, T.; Seed, B. Cell 1995, 83, 333.
(3) Leppanen, A.; Mehta, P.; Ouyang, Y.-B.; Ju, T.; Helin, J.; Moore, K.
L.; van Die, I.; Canfield, W. M.; McEver, R. P.; Cummings, R. D. J. Biol.
Chem. 1999, 274, 24838.
(4) Wilkins, P. P.; McEver, R. P.; Cummings, R. D. J. Biol. Chem. 1996,
271, 18732.
(5) For a previous chemoenzymatic approach to glycopeptide synthesis,
see: Seitz, O.; Wong C.-H. J. Am. Chem. Soc. 1997, 119, 8766.
(6) Mathieux, N.; Paulsen, H.; Meldal, M.; Bock, K. J. Chem. Soc., Perkin
Trans. 1 1997, 2359.
(7) Dullenkopf, W.; Castro-Palomino, J. C.; Manzoni, L.; Schmidt, R. R.
Carbohydr. Res. 1996, 296, 135.
(10) Initial attempts to employ commercially available Fmoc-Tyr(OSO3-)-
OH in SPPS to give the sulfated glycopeptide were unsuccessful. Problems
arose when attempts to remove tBu groups in the presence of the sulfate
failed: (a) Yagami, T.; Shiwa, S.; Futaki, S.; Kitagawa, K. Chem. Pharm.
Bull. 1993, 41, 376. (b) Kitagawa, K.; Futaki, S.; Yagami, T.; Sumi, S.; Inoue,
K. Int. J. Peptide Protein Res. 1994, 43, 190.
(11) The workup procedure employed was previously used for the sulfation
of carbohydrate hydroxyls: Sanders, W. J.; Manning, D. D.; Koeller, K. M.;
Kiessling, L. L. Tetrahedron 1997, 53, 16391.
(12) Marseigne, I.; Roy, P.; Dor, A.; Durieuz, C.; Pelaprat, D.; Reibaud,
M.; Blanchard, J. C.; Roques, B. P. J. Med. Chem. 1989, 32, 445.
(8) Schulz, M.; Kunz, H. Tetrahedron Asymm. 1993, 4, 1205.
(9) Zhang, Z.; Ollmann, I. R.; Ye, X.-S.; Wischnat, R.; Baasov, T.; Wong,
C.-H. J. Am. Chem. Soc. 1999, 121, 734.
10.1021/ja993820o CCC: $19.00 © 2000 American Chemical Society
Published on Web 01/19/2000

Communications to the Editor
J. Am. Chem. Soc., Vol. 122, No. 4, 2000 743
Scheme 2. Solid-Phase Synthesis of Glycopeptides
a (a) Fmoc-AA-OH, HBTU, HOBt, NMM, DMF; (b) Ac2O/pyr; (C) DMF/morpholine; (d) repeat a-c; (e) TFA/H2O/EDT (95:2.5:2.5) (68%, based
on initial loading); (f) SO3-pyr, Pyr, (g) NaOH/MeOH (81%, 2 steps); (g alone) NaOH/MeOH (89%, 1 step).
Scheme 3. Differences in Glycosyltransferase Reactivity with Sulfated vs Unsulfated Glycopeptides
Table 1
substrate
structural changes between glycopeptide sequences 8a and 8b. It
is assumed then that the differences in reactivity toward â1,4-
GalT may be caused by unfavorable electrostatic interactions of
the sulfate with the enzyme. However, the most pronounced vari-
ation between the activity of sulfated and unsulfated glycopeptides
was with R2,3-SiaT. The sulfate appears to significantly interfere
with the transfer of sialic acid from CMP-NeuAc to the glyco-
peptide substrate, as no sialylated product was isolated. Whether
sialyltransferases from other sources can effectively transfer
NeuAc to the glycosulfopeptide is a topic of present investigation.
Current results indicate that sulfation on tyrosine can regulate
the activity of the glycosyltransferases required for the synthesis
of sLex attached to PSGL-1. Work is in progress to investigate
the origin of these reactivity differences, and especially possible
conformational changes of the sulfated PSGL-1 glycopeptide upon
galactosylation. In any case, tyrosine sulfation may play an
important regulatory role in carbohydrate-mediated receptor
recognition. Although the order in which PSGL-1 is posttrans-
lationally modified is currently unknown, previous studies have
shown that both the glycosylated and nonglycosylated PSGL-1
sequences serve as substrates for the sulfotransferase.3 Together
with this study, these results suggest that enzymatic sulfation can
occur at an early biosynthetic stage to prevent the formation of
the sugar ligand, or alternatively, at the final stage to form the
tight-binding glyco-sulfo-heterodimer of PSGL-1.
Km (µM) Vmax (µM/min) kcat (s-1) kcat/Km (µM-1 s-1
)
sulfated 8a
unsulfated 8b
4073
223
2.190
0.339
0.2740
0.0425
6.72 × 10-5
1.90 × 10-4
glycosylations. By this protocol, both sulfated 8a and nonsulfated
8b peptide sequences were readily accessible.
Initial attempts to append galactose in a â1,4-linkage to sulfated
glycopeptide 8a utilizing recombinant bovine â1,4-galactosyl-
transferase (â1,4-GalT) were only marginally successful (Scheme
3). However, product 9a could be isolated if the reaction was
driven to the disappearance of starting material. In contrast, no
product formation was observed when 9a was treated with CMP-
NeuAc in the presence of recombinant rat R2,3-sialyltransferase
(R2,3-SiaT).
For unsulfated glycopeptide 8b, â1,4-GalT, R2,3-SiaT, and
recombinant human R1,3-fucosyltransferase V (R1,3-FucT V) all
accepted the glycopeptide substrates without difficulty. In this
manner, the complete sLex tetrasaccharide was formed attached
to the glycopeptide, yielding 11.
To further characterize the observed reactivity differences, an
assay was developed to determine the kinetic parameters for the
â1,4-GalT reaction with both sulfated 8a and nonsulfated 8b
(Table 1). Notably, sulfation of the tyrosine hydroxyl appears to
interfere with the binding of the substrate to the enzyme (an 18-
fold increase in Km) and thus affects the reaction of â1,4-GalT at
the remote GlcNAc acceptor site of the glycopeptide.
A structural study further indicated that there are no significant
Acknowledgment. This research was supported by the NIH
(GM44154). K.M.K. thanks the American Chemical Society Organic
Division for a fellowship and also Dr. Dee-Hua Huang and Dr. Laura
Pasternack for aid in NMR experiments and analysis.
(13) The 2-dimensional ROESY spectra of 8a and 8b show nearly identical
overlap. For previous studies of the effect of tyrosine sulfation on peptide
conformation see: (a) Durieux, C.; Belleney, J.; Lallemand, J.-Y.; Roques,
B. P.; Fournie-Zaluski, M.-C. Biochem. Biophys. Res. Commun. 1983, 114,
705. (b) Fournie-Zaluski, M.-C.; Belleney, J.; Lux, B.; Durieux, C.; Gerard,
D.; Gacel, G.; Maigret, B.; Roques, B. P. Biochemistry 1986, 25, 3778.
Supporting Information Available: Experimental procedures and
characterization data for 3, 4, 7, 8a, 8b, 10b, and 11 (PDF). This material
is available free of charge via the Internet at http://pubs.acs.org.
JA993820O