Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
List of types of proteins wikipedia , lookup
5-Hydroxyeicosatetraenoic acid wikipedia , lookup
Hedgehog signaling pathway wikipedia , lookup
P-type ATPase wikipedia , lookup
Purinergic signalling wikipedia , lookup
Cooperative binding wikipedia , lookup
NMDA receptor wikipedia , lookup
Signal transduction wikipedia , lookup
G protein–coupled receptor wikipedia , lookup
Biochemical Society Transactions 10. Kawasaki, N., Kawasaki, T. and Yamashina, I. (1985) J. Biochem. (Tokyo) 98,1309-1 320 11. Childs, R. A.,Wright, J. R., Ross, G. F., Yuen, C. T., Lawson, A. M., Chai, W., Drickamer, K. and Feizi, T. (1992)J. Biol. Chem. 267,9972-9979 12. Schweinle, J. E., Ezekowitz, R. A., Tenner, A. J., Kuhlman, M. and Joiner, K. A. (1989) J. Clin. Invest. 84, 1821-1829 13. Matsushita, M. and Fujita, T. (1992)J. Exp. Med. 176, 1497-1502 14. Ihara, S., Takahashi, A., Hatsuse, H., Sumitomo, K., Doi, K. and Kawakami, M. (1991) J. Immunol. 146, 1874-1879 15. Lu, J., Laursen, S. B., Thiel, S., Jensenius, J. C. and Reid, K. B. M. (1993) Biochem. J. 291, in the press 16. Hirani, S., Lambris, J. D. and Muller-Eberhard, H. J. (1985)J. Immunol. 134,1105-1 109 17. Young, N. M. and Leon, M. A. (1987) Biochem. Biophys. Res. Common. 143,645-65 1 18. Loveless, R. W., Feizi, T., Childs, R. A,, Mizuochi, T., Stoll, M. S., Oldroyd, R. G. and Lachmann, P. J. (1989) Biochem. J. 258,109-113 19. Lu, J., Willis, A. C. and Reid, K. B. M. (1992) Biochem. J. 284,795-802 20. Friis-Christiansen,P., Thiel, S., Svehag, S. E., Dessau, R., Svendsen, P., Andersen, O., Laursen, S. €3. and Jensenius, J. C. (1990) Scand. J. Immunol. 31, 453-460 21. Strang, C. J., Slayter, H. S., Lachmann, P. J. and Davis, A. E. (1986) Biochem.J.234,381-389 22. Holmskov, U., Teisner, B., Willis, A. C., Reid, K. B. M. and Jensenius, J. C. (1993) J. Biol. Chem., in the press 23. Weaver, T. E. and Whitsett, J. A. (1991) Biochem. J. 273,249-264 24. van Iwaarden, F, Welmers, B., Verhoef, J., Haagsman, H. P. and van Golde, I,. M. G. (1990) Am. J. Respir. Cell Mol. Biol. 2,9 1-98 25. van Iwaarden, F., van Strijp, J. A. G., Ebskamp, M. J. M., Welmers, A. C., Verhoef, J. and van Golde, I,. M. G. (1991) Am. J. Physiol. 63,692-698 26. Tenner, A. J., Robinson, S. L., Borchelt,J. and Wright, J. R. (1989)J. Biol. Chem. 264,13923-1 3928 27. Manz-Keinke, H., Plattner, H. and Schlepper-Schafer, J, (1992) Eur. J. Cell Biol. 57, 95-100 28. Kuan, S. F., Rust, K. and Crouch, E. (1992) J. Clin. Invest. 90,97-106 29. Voss, T., Eistetter, H., Schafer, K. P. and Engel, J. (1988)J. Mol. Biol. 201,219-2217 30. Haagsman, H.P., Hawgood, S., Sargeant, T., Buckley, D., White, R. T., Drickamer, K. and Benson, B. J. (1987)J. Biol. Chem. 262,13877-13880 31. Person, A., Chang, D. and Crouch, E. (1990) J. Biol. Chem. 265,5755-5760 32. van Iwaarden, J. F., Shimizu, H., van Golde, P. H. M., Voelker, D. R. and van Golde, I,. M. G. (1992) Biochem. J. 286,s-8 33. Bruns, G., Stroh, H., Voldman, G., Latt, S. A. and Floros, J. (1987) Hum. Genet. 76, 58-62 34. Fisher, J. H., Kao, F. T., Jones, C., White, K. T., Benson, B. J. and Mason, R. J. (1987) Am. J. Hum. Genet. 40,503-5 11 35. Kolble, K., Lu, J., Mole, S. E., Kaluz, S. and Reid, K. B. M.(1993) Genomics, in the press 36. Schuffenecker, I., Narod, S. A,, Ezekowitz, R. A. B., Sobol, H., Feunteun, J. and Lenoir. G. M. (1991) Cytogenet. Cell Genet. 56,99-102 37. Katyal, S.L., Singh, G. and Locker, J. (1992) Am. J. Respir. Cell Mol. Biol. 6, 446-452 38. Korfhagen, T. R., Glasser, S. W., Bruno, M. D., McMahan, M. J. and Whitsett, J. A. (1991) Am. J. Respir. Cell Mol. Biol. 4,463-469 39. White, R. T., Damm, D., Miller, J., Spratt, K., Schilling,J., Hawgood, S., Benson, B. and Cordell, €3. (1985) Nature (London) 317, 361-363 40. Taylor, M. E., Brickell, P. M., Craig, R. K. and Summerfield,J. A. (1989) Biochem.J. 262,763-771 41. Lim, B. L., Lu, J. and Reid, K. B. M. (1993) Immunology 78,159- 165 Received 17 December 1992 Recognition of complex carbohydrates by the macrophage mannose receptor Maureen E. Taylor Glycobiology Institute, Department of Biochemistry, University of Oxford, South Parks Road, Oxford OX I 3QU, U.K. Introduction T h e mannose receptor of macrophages and hepatic endothelial cells mediates binding and internalization of glycoconjugates terminating in mannose, fucose or N-acetylglucosamine [ 11. Recognition of Abbreviations used: BSA, bovine serum albumin; CRD, carbohydrate-recognition domain. Volume 2 I carbohydrates by the mannose receptor allows the discrimination of self from non-self, since terminal mannose and N-acetylglycosamine residues are found rarely in mammalian cell-surface or serum glycoproteins, but are a common component of the cell surface of bacteria, fungi and parasites. T h u s the mannose receptor can play a role in the innate immune response against pathogenic micro-organ- Carbohydrates, Shapes and Biological Recognition isms by mediating opsonin-independent phagocytosis [2-41. Some endogenous glycoproteins bearing high-mannose oligosaccharides, including lysosomal enzymes [ 51 and tissue plasminogen activator [6], are also cleared by the mannose receptor. These potentially harmful proteins are often released from cells in response to pathological events. Structure of the mannose receptor Mannose receptors have been isolated from macrophages [7, 81 and from placenta [9]. The receptor from both sources is a glycoprotein consisting of a single subunit with a molecular mass of about 175 m a . Mannose receptors from human placenta [ 101 and human monocyte-derived macrophages [ 111 have been cloned and sequenced. Clones from these two sources have identical sequences, suggesting that the receptor isolated from placenta is also macrophage-derived. The receptor is oriented as a type I transmembrane protein (C-terminus inside the cell) with a 41 amino-acid cytoplasmic tail. The extracellular portion of the receptor consists of three types of cysteine-rich domains. The N-terminal cysteine-rich domain shows no similarity to other known sequences. The second domain resembles the type I1 repeats of fibronectin [ 121. No specific function has been assigned to the type I1 repeats in fibronectin or to related domains found in other proteins, so it is not possible to make predictions about the role of this domain in the mannose receptor. The rest of the extracellular part of the receptor consists of eight segments showing sequence similarity to the Ca2+dependent carbohydrate-recognition domains (Ctype CRDs) of other animal lectins [ 131. C-type CRDs are discussed in detail elsewhere in this colloquium. They bind various different saccharides in a calcium-dependent manner and are characterized by 14 invariant residues and 18 highly conserved residues, spread over approximately 120 amino acids [ 141. They are associated with a variety of effector domains in a large number of proteins [15]. The mannose receptor is the only protein known to have more than one CRD within a single polypeptide. Structure-function analysis of the mannose receptor Since the mannose receptor is known to bind carbohydrates in a Caz+-dependent manner it is reasonable to assume that the carbohydrate-binding activity of the receptor resides within the CRD-like segments. Each of the mannose receptor CRDs contains most of the conserved residues necessary for maintaining the CRD fold seen in the crystal structure of the CRD of rat serum mannose-binding protein [14]. However, it is not possible to predict simply from examination of the sequences, which of the CRDs have binding activity. Analysis of the binding activity of portions of the receptor expressed in vitro, in fibroblasts, in bacteria and in insect cells was undertaken in order to answer the following questions: (1) are the N-terminal cysteinerich domain and the fibronectin-type I1 repeat necessary for binding and internalization of glycoproteins? (2) Which of the CRDs can bind to carbohydrates? (3) Does the multispecificity of the receptor arise from each CRD having specificity for different saccharides? (4) Are several active CRDs necessary for high-affinity binding to oligosaccharides? localization of the binding activity to CRDS4-8 Initial experiments indicated that the N-terminal cysteine-rich domain and the fibronectin type I1 repeat are not essential for binding and endocytosis of glycoproteins. Fibroblasts expressing a truncated receptor with an extracellular portion consisting of CRDs 1-8 are able to bind and endocytose a neoglycoprotein ligand, mannose,,-bovine serum albumin (Man,,-BSA), as efficiently as fibroblasts expressing the intact receptor [16]. The truncated receptor has the same affinity as the intact receptor for three different mannose-terminated ligands. These results indicate that the carbohydrate-binding activity of the receptor resides within the 0 s . The function of the N-terminal cysteine-rich domain and the fibronectin type I1 repeat remains unclear. It is possible that they bind to a non-carbohydrate ligand. An in vitro-binding assay was used to determine which of the CRDs are able to bind sugars [ 161. Portions of DNA coding for one or more CRDs were fused to codons specifying the preproinsulin signal sequence in an SP64 expression vector and transcribed in vitro. The RNAs produced were translated in vitro in the presence of microsomes. The insulin signal sequence directs translation products into the lumen of the microsomes, where disulphide bonds which are necessary for the production of an active CRD, can form. Translation products were solubilized with detergent and tested for their ability to bind mannose immobilized on Sepharose. The results are summarized in Figure 1. I993 469 Biochemical Society Transactions Figure I Carbohydrate-binding activity of CRDs in the macrophage mannose receptor 470 The structure of the CRD-containing portion of the receptor IS summarized at the top with the activity of constructs containing subgroups of CRDs shown below. Activity was measured by retention on mannose-Sepharose following in vitrotranscription and translation [ 161. Adapted from [ I71 with permission. Binding of MMR domains to Man-Sepharose following in vitro translation BINDING DOMAINS 1 2 3 4 5 6 4 5 6 4 5 4 5 5 6 5 6 6 7 8 7 8 7 7 8 + + + + + + - A segment of the receptor consisting of CRDs 1-3 does not bind to mannose-Sepharose, whereas a segment consisting of CN>s 4-8 has binding activity. O f CKIIs 4-8, only CRD 4 can bind to mannose-Sepharose in the absence of other CKI)s. CKI) 5 alone does not show binding activity. However, a segment of the receptor consisting of CRDs 5-8 does show binding activity. Since the removal of CKI) 5 results in loss of binding, it can be concluded that CRD 5 contributes to the binding activity of the receptor. A similar argument can be made for CRI) 7, since a fragment consisting of CKI)s 5-7 binds to mannose-Sepharose, but CKDs 5-6 do not bind. CRDs 1-3 and CRDs 6-8 failed to bind to fucose. N-acetylglucosamine and glucose, as well as to mannose. The results of the in vitro-translation studies suggest that CRI)s 1-3 do not bind carbohydrates and that the binding activity of the receptor resides in CRI)s 4-8.IIowever, since this is not a quantitative assay, it was necessary to confirm the results by further expression studies in fibroblasts [ 161. Cells expressing a truncated receptor consisting of domains 4-8 are able to endocytose "'1Man,,-BSA as efficiently as cells expressing the Volume 2 I intact receptor. This truncated receptor has the same affinity as the intact receptor for three different ligands. Cells expressing a truncated receptor consisting of CRLIs 5-8 can also endocytose "'IMan,,-HSA, but at a much slower rate than cells expressing CIWs 4-8 or the intact receptor, highlighting the importance of CRI) 4.Cells expressing CRDs 6-8 are not able to internalize "'IMan,,-HSA, indicating that CKI) S is also essential for binding and endocytosis. Saccharide-binding activity of CRD 4 Since CRD 4 was the only domain found to show carbohydrate-binding activity in the absence of other CRDs, this domain was produced in a bacterial expression system so that its binding properties could be characterized further [ 161. Correctly folded CRI) 4 was purified from the bacteria by affinity chromatography on mannose-Sepharose. Purified CRI) 4 immobilized on a microtitre plate binds 1251-Man,,-HSA in a saturable fashion. The affinity of the purified domain for various monosaccharides and glycoproteins was determined, based on their ability to inhibit binding to "'1Man,,-HSA. CKI) 4 binds to N-acetylglucosamine, fucose and glucose as well as to mannose with dissociation constants in the millimolar range. The intact receptor also has specificities for these monosaccharides with dissociation constants in the millimolar range. Thus a single CRD can mimic the monosaccharide-binding properties of the whole receptor. Multispecificity for mannose, fucose, Nacetylglucosamine and glucose is also shown by the CRD of rat serum mannose-binding protein, which has recently been crystallized in complex with an oligosaccharide [ 181. It is likely that the interaction between CRD 4 and monosaccharides will be very similar to that seen for the mannose-binding protein, since the residues involved in ligating the sugar are completely conserved [ 191. However, CRD 4 binds only poorly to glycoproteins, such as invertase and mannan, and thus cannot account for the binding of the receptor to natural ligands. As indicated by the expression studies in fibroblasts and in vitro. other CRDs in the 4-8 segment must contribute to the binding of oligosaccharides. Binding to multivalent ligands Hydrodynamic studies of the receptor show that it is a monomer in solution and crosslinking studies indicate that it is a monomer in the membrane [20]. Thus multiple carbohydrates in a single receptor polypeptide must co-operate to achieve high-affinity binding of complex ligands. In order to determine Carbohydrates, Shapes and Biological Recognition how the CRDs cluster to achieve this high-affinity binding, internal fragments of the receptor were produced using the baculovirus expression system [ZO]. Fragments of the receptor consisting of CRDs 4-5,4-6 and 4-7 were purified from the medium of insect cells by affinity chromatography on mannose-Sepharose and were used in proteolysis studies and in assays to determine the affinities for various ligands [201. The solid-phase-binding assay described for CRD 4 was used to determine the affinities of the larger fragments of the receptor for Man,,-BSA and for two natural glycoproteins, invertase and mannan, which bear high-mannose oligosaccharides. The results obtained for the internal fragments of the receptor and those obtained for CRD 4 alone and for CRDs 4-8 expressed in fibroblasts are compared in Figure 2, which shows a plot of affinities for the three ligands as a function of the number of CRDs present. The largest increase in affinity is seen when CRD 5 is combined with CRD 4, with the increase being particularly marked for the two natural glycoproteins. CRDs 4 and 5 together are sufficient to match the affinity of the intact receptor for Man,,-BSA and invertase, suggesting that these two domains form a ligand-binding core essential for high-affinity binding of multivalent ligands. However, the affinity for mannan increases as more CRDs are added. The full affinity of the intact receptor for mannan is only reached when CRDs 4-8 are present, indicating that accessory domains 6, 7 and 8 are necessary for binding to some ligands. Proteolysis experiments support the idea that CRDs 4 and 5 form a ligand-binding core, since these two CRDs cannot be separated by treatment with subtilisin [20]. Resistance to proteolysis in the presence of Ca2+ is a common feature of C-type CRDs [Zl]. However, the fact that proteolysis of a fragment consisting of CRDs 4-5 does not result in the release of individual CRDs suggests that these two domains form extensive contacts and are not simply linked by a flexible tether. Such an arrangement may fix the orientation of the binding sites within these two CRDs. Different modes of clustering of Ctype CRDs . MANNAN The experiments described above indicate that the mannose receptor achieves high-affinity binding to oligosaccharides by clustering of several active CRDs in a single polypeptide. Clustering of CRDs to allow high-affinity binding to oligosaccharides is a common feature of many C-type lectins. These other lectins have a single CRD in each polypeptide and thus must form oligomers to achieve clustering of CRDs. The chicken hepatic lectin forms trimers by association of identical polypeptides in the membrane [21, 221. The soluble mannose-binding proteins also forni trimers. In this case identical polypeptides each with a single CRD are held together by the association of collagen-like domains [23]. In the rat asialoglycoprotein receptor, two different polypeptides each with a single CRD form hetero-oligomers in the membrane "241. Different subunits of the asialoglycoprotein receptor bind to different terminal galactose residues on a glycopeptide, indicating that clustering of CRDs may determine specificity as well as affinity [25]. Analysis of the crystal structure of the mannose-binding protein CRD bound to an oligosaccharide has revealed why clustering of CRDs is necessary to achieve high-affinity binding [ 181. The contact between the sugar and the CRD in the crystal is very limited, with each CRD binding to a single sugar residue by ligation of two hydroxyl residues. I993 47 I Biochemical Society Transactions 472 Thus the interaction between a single CRD and a sugar is weak and several CRDs are needed to bind tightly to oligosaccharide ligands. It is probable that the different forms of clustering of CRDs seen in C-type lectins allows the selection of different types of ligand. The CRDs of each lectin must be arranged to match the geometric configurations of their particular oligosaccharide ligands. The arrangement of CRDs in the asialoglycoprotein receptor (Figure 3 ) may provide optimal binding to the clusters of terminal galactose residues that are found in the desialated complex oligosaccharide chains of serum glycoproteins cleared by this receptor. In contrast the linear arrangement of CRDs in the mannose receptor (Figure 3 ) may be more suitable for high-affinity binding of repeated polymers such as yeast mannan. The asialoglycoprotein receptor binds endogenous ligands with a limited set of structures, while the mannose receptor must recognize a diversity of foreign ligands. In the mannose receptor the presence of multiple CRDs in addition to the CRD 4-5 ligand-binding core may provide the flexibility needed to interact with this diversity of ligands. Conclusions Structure-function analysis of the mannose receptor has revealed that the carbohydrate-binding Figure 3 Different modes of clustering of C-type CRDs seen in the macrophage mannose receptor and the asialoglycoprotein receptor The macrophage mannose receptor (left) contains eight CRDs in a single polypeptide. The two CRDs that form the ligandbinding core are shaded. The N-terminal cysteine-rich domain is represented by the vertical box and the fibronectin type II repeat by the horizontal box. In the asialoglycoprotein receptor (right) three polypeptides, each containing a single CRD, associate in the membrane to form a trimer. MACROPHAGE MANNOSE RECEPT0 R 8 ASIALOGLYCOPROTEIN RECEPTOR activity of the receptor is located in CRDs 4-8. CRDs 4-5 form a protease-resistant ligand-binding core sufficient to bind some ligands with high affinity, but accessory CRDs 6-8 are also required for high-affinity binding of other ligands. A consequence of the organization of the receptor is that both the valency and the geometry of glycoconjugates are important determinants of the binding affinity. Further structural work is required to determine how each of the CRDs interacts with sugars and how the CRDs are arranged spatially to match the geometric configurations of particular oligosaccharide ligands. I thank Kurt Drickamer and Raymond Dwek for comments on the manuscript. The work described here was supported by Grant GM42628 from the National Institutes of Health. The Glycobiology Institute is supported by Monsanto. 1. Stahl, P. D. (1990) Am. J. Respir. Cell Mol. Hiol. 2, 317-318 2. Warr, S. A. (1980) Hiochem. Biophys. Res. Commun. 93,737-745 3. Speert, D. P., Wright, S. D.. Silverstein, S. C. and Mah, B. (1988)J. Clin. Invest. 82,872-879 4. Ezekowitz, R. A. H., Williams, D. J., Koziel, H., Armstrong, M. Y. K., Warner, A., Richards, F. F. and Rose, R. M. (1991) Nature (London) 351, 155-158 5. Stahl, P. D. and Schlesinger, 1’. H. (1980) Trends Hiochem. Sci. 5. 194- 196 6. Otter, M.. Barrett-Bergshoeff,M. M. and Kijken, 1). C. (1991)J. Biol. Chem. 266,13931-13935 7. Haltiwanger, K. S. and Hill, R. I,. (1986) J. Hiol. Chem. 261,7440-7444 8. Wileman, T., Lennartz, M. K. and Stahl, P. D. (1986) Proc. Natl. Acad. Sci. U.S.A. 83,2501-2505 9. Lennartz, M. R., Cole, F. S., Shepherd, V. I,., Wileman, T. E. and Stahl, P. D. (1987)J. Hiol. Chem. 262,9943-9944 10. Taylor, M. E., Conary, J. T., Lennarz. M. K., Stahl, P. D. and Drickamer, K. (1990) J. Hiol. Chem. 265, 12156-12162 11. Ezekowitz, R. A. B., Sastry, K., Hailly, IJ. and Warner, A. (1990)J. Exp. Med. 172,1785-1794 12. Kornblihtt, A. R., Umezawa, K.. Vibe-Pedersen, K. and Haralle, F. E. (1985) EMBO J. 4. 1755-1759 13. Drickamer, K. (1988)J. Hiol. Chem. 263,9557-9560 14. Weis, W. I., Kahn, R., Fourme, R., Drickamer, K. and Hendrickson. W. A. (199 1) Science 254,1608- 1615 15. Bezouska, K., Crichlow, G. V., Rose, J. M., Taylor, M. E. and Drickamer, K. (1991) J. Biol. Chem. 266, 11604-11609 16. Taylor, M. E., Bezouska, K. and Drickamer, K. (1992)J. Biol. Chem. 267, 1719- 1726 17. Weis, W. I., Quesenberry, M. S., Taylor, M. E., Bezouska, K., Hendrickson, W. A. and Drickamer, K. l i \ Volume 21 Carbohydrates, Shapes and Biological Recognition (1993) Cold Spring Harbor Symp. Quant. Biol. 52, in the press 18. Weis, W. I., Drickamer, K. and Hendrickson, W. A. (1992)Nature (London) 360, 127-134 19. Drickamer, K. (1992)Nature (London) 360, 183-186 20. Taylor, M. E. and Drickamer, K. (1993) J. Biol. Chem. 268,399-404 21. Loeb, J. A. and Drickamer, K. (1987)J. Biol. Chem. 262,3022-3029 ~ 22. Verry, F. and Drickamer, K. (1993) Biochem. J., in the press 23. Drickamer, K., Dordal, M. S. and Reynolds, I,. (1986) J. Biol. Chem. 261,6878-6887 24. Spiess, M. (1990)Biochemistry 29, 10009-10018 25. Rice, K. G., Weisz, 0. A., Barthel, T., Lee, R. T. and Lee, Y. C. (1990)J. Biol. Chem. 265,18429-18434 Received 10 December 1992 ___ ~ _ _ _ The role of mannose-binding protein in host defence John A. Summerfield Department of Medicine, St Mary’s Hospital Medical School, Imperial College of Science, Technology and Medicine, London W2 INY, U.K. Mannose-binding protein Human serum contains mannose-binding protein (MBP), a calcium-dependent C-type lectin, secreted by hepatocytes, which binds glycoproteins terminating in mannose or N-acetylglucosamine. MRP occurs in serum as a mixture of oligomers of 9-18 identical polypeptide chains of 32 kDa [I-51. On binding to a mannose-rich surface, MRP activates complement through the classical pathway [6]. MRP binds C l r and Cls to form C1 esterase, which cleaves C 2 and C4, which in turn form a complex on the mannose-rich surface to form C 3 convertase. C 3 convertase cleaves C 3 to form C3b which then binds to the surface and opsonizes the ligand. Recently MBP has been shown to be identical with Ra-reactive factor [7]. Ra-reactive factor is a complement-activating bactericidal protein which binds to Ra-chemotype strains of bacteria and yeasts. MRP is encoded by four exons each of which code for different functional domains of the molecule [8,9]. Exon 1 encodes the signal sequence of a secreted protein, a cysteine-rich domain and seven copies of the motif Gly-Xaa-Yaa, which is typical of a collagen domain. The junction between exon 1 and exon 2 encodes the sequence Gly-Gln-Gly. Exon 2 encodes a further 12 Gly-Xaa-Yaa collagen repeats. Exon 3 encodes a short ‘neck‘ domain and exon 4,the largest exon, encodes the carbohydratebinding domain. The structure of the high-molecular-weight oligomers of MBP that are found in serum can be inferred from the sequence of the MBP gene. Trimers of MBP polypeptides associate by forming Abbreviation used: MHI’, mannose-binding protein. a triple helix between their collagen domains. The interruption of the collagen motif by the sequence Gly-Gln-Gly between exons 1 and 2, by analogy with Clq, is probably the site where the triple-helical chains of MBP appear to bend on electronmicroscopy [ 101. The triple helices are stabilized by disulphide bridges between the cysteine-rich domains. Three to six of the trimers assemble by disulphide bridges into oligomers. This gives the final MBP oligomer the appearance of a bunch of flowers, where the flower heads are the carbohydrate-binding domains and the stalks are the collagen domains. MBP is an acute phase protein [8, 113. The 5’ flanking sequence of the MBP gene contains a heatshock-promoter sequence and two glucocorticoidresponsive promoter sequences typical of acute phase protein genes. Immunodeficiency caused by defective opsonization Infants of 6-18 mth with this immunodeficiency syndrome suffer repeated bacterial and fungal infections [ 12, 131. The immunodeficiency is present in about 25% of children with frequent unexplained infections. Children with the defect suffer about twice as many severe infections as matched control children [ 141. The infections may be severe and five deaths have been reported. Usually relatives of the immunodeficient infants suffered repeated infections while infants. There is a high frequency of atopy in the children and their families [14]. This immunodeficiency is common with an estimated frequency, judged by tests of opsonic function, of 5-7% [13, 15, 161. The sites of infection are varied; otitis media, chronic diarrhoea and meningitis are the most I993 473