A. R. N. Arulanandam

Conformation and function of the N-linked glycan in the adhesion domain of human CD2.

The adhesion domain of human CD2 bears a single N-linked carbohydrate. The solution structure of a fragment of CD2 containing the covalently bound high-mannose N-glycan [-(N-acetylglucosamine)2-(mannose)5-8] was solved by nuclear magnetic resonance. The stem and two of three branches of the carbohydrate structure are well defined and the mobility of proximal glycan residues is restricted. Mutagenesis of all residues in the vicinity of the glycan suggests that the glycan is not a component of the CD2-CD58 interface; rather, the carbohydrate stabilizes the protein fold by counterbalancing an unfavorable clustering of five positive charges centered about lysine-61 of CD2.

Structure of the glycosylated adhesion domain of human T lymphocyte glycoprotein CD2

BACKGROUND CD2, a T-cell specific surface glycoprotein, is critically important for mediating adherence of T cells to antigen-presenting cells or target cells. Domain 1 of human CD2 is responsible for cell adhesion, binding to CD58 (LFA-3) expressed on the cell to which the T cell binds. Human CD2 domain 1 requires N-linked carbohydrate to maintain its native conformation and ability to bind CD58. In contrast, rat CD2 does not require N-linked carbohydrate, and binds to a different ligand, CD48. RESULTS The three-dimensional structure of the glycosylated form of domain 1 of human CD2 has been determined by NMR spectroscopy. The overall structure resembles the typical beta-barrel of an immunoglobulin variable domain. Nuclear Overhauser enhancement contacts between the protein and the N-linked glycan have been tentatively identified. CONCLUSION Based on our results, we propose a model showing how the N-linked glycan might be positioned in the human CD2 domain 1 structure. The model provides an explanation for the observed instability of deglycosylated human CD2, and allows residues that are important for CD58 binding to be differentiated from those affecting conformational stability via interactions with the glycan.

CD2: an exception to the immunoglobulin superfamily concept?

gene. 11. Genomic Southern blots of eral-1 DNA digested with Eco RI and probed with right-border (RB) TDNA produced three bands (13, 7.0, and 8 kb) that could be accounted for by a contiguous double insertion of T-DNA in which one of the copies has been inverted to give two flanking RB regions attached to plant genomic DNA. Subsequent analysis with other restriction enzymes verified that the 7and 8-kb bands contained the insertion points of T-DNA and flanking plant DNA. DNA pools enriched for these fragments were ligated to the X-ZAPII arms according to the manufacturers instructions (Stratagene) and five positive plaques were identified that hybridized with the RB probe. Two plasmids (pSC10 and pSC1 1), each containing about 1.2 kb of T-DNA attached to different flanking regions, detected a single Eco RI 1.5-kb fragment in wild-type DNA by Southern analysis. Both clones therefore contain genomic sequences that are contiguous in wild-type DNA, indicating that the RB fragments in the era 1-1 mutant represent the ends of a single-site T-DNA insertion. pSC10 was used as a probe to screen an Arabidopsis cDNA library, PRL2 X-ZipLox (ABRC, stock CD4-7); and five positive cDNAs were identified and sequenced. The longest cDNA, pZL51 (1.45 kb), was used to screen Columbia (X-ZAPII, Stratagene) and Lansberg genomic (A-FIX, ABRC stock CD4-8) libraries and four positive clones were identified. One clone spanning 6 kb and encompassing the entire pZL51 clone was completely sequenced. A larger genomic insert (14 kb) was used to size deletions in the fast-neutron mutants (era l-2 and era 1 -3). 12. B. W. Shirley, S. Hanley, H. M. Goodman, Plant Cell 4, 333 (1992). 13. L. E. Goodman, C. M. Perou, A. Fujiyama, F. Tamanoi, Yeast 4, 271 (1988); W-J. Chen, D. A. Anders, J. L. Goldstein, D. W. Russell, M. S. Brown, Ce// 66, 327 (1991); Z. Yang, C. L. Cramer, J. C. Watson, Plant Physiol. 101, 667 (1993). 14. W. R. Schafer and J. Rine, Annu. Rev. Genet. 30, 209 (1992). 15. To assay farnesylation, 1 g (fresh weight) of wildtype or mutant flower buds was homogenized in extraction buffer [50 mM Hepes (pH 7.5), 1 mM MgCI2, 1 mM EGTA, 5 mM dithiothreitol (DTT), leupeptin (2 .tg/ml) aprotinin (2 ,ug/ml), and 1 mM phenylmethylsulfonyl fluoride]. The synthetic heptapeptides GGCCAIM (CAIM) and GGCCAIL(CAIL) (19) were prepared as described [B. He et a/., Proc. Natl. Acad. Sci. U.S.A. 88, 11373 (1991)]. Target peptide sequences were chosen according to activities measured in tobacco culture [S. K. Randall et al., Plant Cell 5, 433 (1993)]. Extracts were clarified at 4°C at 10,000g for 10 min and then at 1 00,000g for 30 min. Soluble protein extract (100 ,g) was incubated at 300C for 40 min in 25 ,u of reaction buffer [50 mM Hepes (pH 7.5), 5 mM MgCI2, 5 mM DTT, 50 FtM peptide, and 0.5 pM [3H]farnesyl diphosphate (FPP) (1 7.0 Ci/mmol; Amersham)]. Reactions were terminated with EDTA (at a final concentration of 50 mM), spotted onto Silica Gel 60 thin-layer chromatography plates (Millipore), and developed with n-propanol and water (7:3 v/v) for 4 to 5 hours. The plates were dried, sprayed with En3Hance (New England Nuclear), and exposed to Kodak X-OMAT AR film at -700C for 4 days. 16. J. Hancock, H. Paterson, C. Marshall, Cell 43, 133 (1990); S. Leevers, H. F. Paterson, C. J. Marshall, Nature 369, 411 (1994); D. Stokoe, S. G. Macdonald, K. Cadwallader, M. Symons, J. F. Hancock, Science 264, 1463 (1994); P. J. Casey, Biochem. Soc. Trans. 23,161 (1995). 17. P. C. Sternweis, Curr. Opin. Cell Biol. 6, 198 (1994). 18. J. Inglese, W. J. Koch, M. G. Caron, R. J. Lefkowitz, Nature 359, 147 (1992). 19. Abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, lie; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; 5, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr. 20. We thank K. Keith for thoughtful discussion and extensive contributions in the writing of this paper. We thank S. Sarkar for critical reading of the manuscript and D. Skalamera for technical assistance with photography. Many of the strains and DNA libraries used were provided by the Arabidopsis Stock Center at

Response: CD2: An Exception to the Immunologlobulin Superfamily Concept?

Response: In our report, we stated that the functional form of human CD2 (hCD2) is maintained by the stabilizing effect of its N-linked glycan in domain 1. Earlier we had observed that (i) whereas soluble nonglycosylated human CD2 domain 1 (hsCD2dl) expressed in Escherichia coli was not functional, Chinese hamster ovary cell derived glycosylated hsCD2dl was fully active, and (ii) complete removal of the single N-glycan in domain 1 of hCM2, either by Nglycanase treatment of hsCD2dl or by mutation of the Asn65-Xxx-Thr67 sequence of hCD2, completely abrogated CD58 and anti-CD2 monoclonal antibody binding (1, 2). Our recent structUral studies on glycosylated hsCD2dl (3, 4) revealed that this N-glycan is opposite to the CD58 binding site in hCD2 and is not involved in CD58 recognition. To determine why this N-glycan is nevertheless critical for hCD2 adhesion function, we stuLdied both enzymatically treated hsCD2dl and a series of fulllength hCD2 mutants (4). From these studies we concluded that the N-glycan crucially stabilizes the folded protein structure, and showed that the first N-acetylglucosamine residue (GIcNAc-1) is sufficient to keep the glycoprotein folded (4). Several recent studies on glycoproteins have shown that carbohydrates globally stabilize the polypeptide fold (5-11 ) and in two cases, an increase in the stability of about 1.2 kcal/ mol was measured (6, 8). In two examples, this stabilizing function was achieved by a single sugar unit like in hCD2 (5, 6). However, the mechanism of this stabilization is not clear yet. On the basis of our mutations, we have not been able to identify a particulalr stabilizing interaction between glycan and protein in the folded state (4). On the other hand, the presence of the bulky, conformationally restricted GlcNAc-I ring is likely to lower the entropy of the unfolded state favoring the folded form. The marginal stability of hCD2 has been shown to be due to the unfavorable clIstering of five lysines centered aroUnd Lys-61 (4). Compared to hCD2, rat CD2-which does not require a glycan for maintenance of its stable fold-contains a glutamic acid at the corresponding position of Lys-61. A nonglycosylated hCD2 mutant containing a K61E mutation is stable and shows full CD58 binding activity. This suggests that in this hCD2 mutant, Glu-61 removes the unfavorable clustering of positive charges and forms a stabilizing salt bridge with Lys69, eliminating the need for a stabilizing glycan (4). Davis and van der Merwe have done independent work on CD58 and anti-CD2 monoclonal antibody binding of enzymatically treated hsCD2dld2 using surface plasmon resonance (SPR) analysis (12). First, they state that the N-linked glycans do not stabilize the folded conformation of hsCD2dl, but may enhance its overall solubility. However, it is unlikely that the stabilizing effect demonstrated for hsCD2d 1 should not contribute to the stability of the two-domain fragment, in particular because similar stabilization effects of glycans have been found in a number of other glycoproteins (5-11). Consistent with our view, they observed a marked reduction of binding activity for several anti-CD2 monoclonal antibodies compared to the results of endoglycosidase H treatment, when in addition, the GIcNAc-1 of the three N-glycans of hsCD2dld2 were removed with N-glycanase treatment (12). Furthermore, with N-glycanase treatment, hsCD2dld2 needs to he used in their SPR analysis of CD58 binding activity immediately to avoid protein aggregation. Exposure of hydrophobic groups on removal of the three GIcNAc-1 of hsCD2dld2 per se is unlikely in this case, given the nature of the hsCD2dld2 surface defined by the crystal structure in the vicinity of the three N-glycans (13). The observed aggregation is more likely a consequence of partial unfolding, a common phenomenon, as hydrophobic protein core interiors will be exposed. Although N-glycanase treated hsCD2dld2 binds CD58 with wild-type affinity, this does not necessarily mean that it is equally stable as wild-type hCD2. The fact that N-glycanase-treated hsCD2dld2 binds several conformationally-sensitive antibodies to a much lower degree than wild-type hCD2 and its high tendency to aggregate clearly show that this molecule is no longer fully native-like, although it may still be able to provide an intact CD58 binding site for some amount of time. Second, Davis and van der Merwe suggest that the N-glycan, within the adhesion domain of hCD2, influences the initial folding ofhCD2 rather than determining its post-folding stability. However, the CD2 copy number of nonglycosylated hCD2 mutants on the cell surface is around 50% or greater of wild-type hCD2. Misfolded proteins are usually recognized and degraded by the quality control system in the endoplasmic reticulum, and hence are expressed at very low levels relative to wild-type proteins (14). Nevertheless, besides its role in post-folding stability, the N-glycan may also play an important role in the conformational maturation of hCD2. Ellis L. Reinherz Jing Li Alex Smoylar Laboratory of Immunobiology, Dana-Farber Cancer Institute and Department of Medicine, Harvard Medical School, Boston, MA 02115, USA Daniel F. Wyss Maria H. Knoppers Kevin J. Willis Antonio R. N. Arulanandam Procept, Inc., Cambridge, MA 02139, USA Johnathan S. Choi Gerhard Wagner Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School, Boston, MA 02115, USA

NMR Studies of Proteins Involved in Cell Adhesion Processes

Many biological processes on the molecular level are associated with membrane-bound receptors or other integral membrane proteins. We have been working with proteins that are domains of receptors, or inhibit their function. The first topic are the T-cell surface glycoprotein receptor CD2 and its counter receptor CD58. The other topic is about antagonists of the integrin adhesion receptor glycoprotein IIbIIIa (GPIIbIIIa) which is found on platelet surfaces. Human CD2 is a glycoprotein, and the carbohydrate of its adhesion domain is crucial for adhesion function. The platelet receptor GPIIbIIIa, a Ca2+ dependent heterodimeric glycoprotein from the integrin family, binds fibrinogen and mediates the aggregation of platelets to form a blood clot. Natural protein antagonists of this receptor have primarily been found in the venum of various snakes, which have been termed disintegrins, and in the saliva of blood-sucking leeches. They contain an Arg-Gly-Asp (RGD) sequence in their active site. Due to the potent antiplatelet effect of these RGD proteins, the structures of their active sites have been of considerable interest for the design of antithromotic drugs.