The polymeric hemoglobin molecule of Artemia. Interpretation of translated cDNA sequence of nine domains.

Translated cDNA for Artemia hemoglobin provided sequence data for almost nine domains, from the fourth residue of the A helix of one domain through 1405 residues to a stop codon after the ninth domain. The domain sequences were all different (homology between pairs 17-38%) but aligned well with each other and with conventional globins, satisfying the requirements for Phe at CD1, His at F8 and most other highly conserved features of globins including His at E7. Features found to be characteristic of Artemia globin and present in all nine domains were Phe at B10, Tyr at C4, Gly at F5, Phe at G5 and Gly at H22. Approximately 14 residues including a consensus -Val-Asp-Pro-Val-Thr-Gly-Leu- were available to form the linker between each pair of domains. The Artemia sequence data were compared with the crystal structures of Chironomus thummi thummi erythrocruorin III and sperm whale myoglobin in order to identify features of structural similarity and to examine the consequences of the differences. The Artemia sequences were compatible with the main helices and critical features of the globin fold. Possible modifications to the C helix, FG turn, and GH turn were studied in terms of molecular coordinates.     

Structural interpretation of the amino acid sequence of a second domain from the Artemia covalent polymer globin

Artemia has a complex extracellular hemoglobin of Mr 260,000 comprising two globin chains (Mr 130,000) each of which is a polymer of eight covalently linked domains of Mr 16,000. The primary structure of this polymeric globin was studied to understand how globin folded domains are ordered within a globin chain and, in turn, how the latter associate into a functional hemoglobin molecule. Here we report the amino acid sequence of a second domain, E7 (Mr 16,081, excluding the heme), and interpretations of sequence data by computer-assisted alignment and modeling. This clearly shows that, as with domain E1 (Moens, L., Van Hauwaert, M.-L., De Smet, K., Geelen, D., Verpooten, G., Van Beeumen, J., Wodak, S., Alard, P., & Trotman, C. (1988) J. Biol. Chem. 263, 4679-4685), domain E7 is compatible with a globin folded structure of the beta-type chain. Several specific differences of domains E7 and E1 from the classic globins are identified. They possibly can be interpreted in terms of specific requirements for a double octameric functional molecule.     

Structural basis of human cytoglobin for ligand binding

Cytoglobin (Cgb), a newly discovered member of the vertebrate globin family, binds O(2) reversibly via its heme, as is the case for other mammalian globins (hemoglobin (Hb), myoglobin (Mb) and neuroglobin (Ngb)). While Cgb is expressed in various tissues, its physiological role is not clearly understood. Here, the X-ray crystal structure of wild type human Cgb in the ferric state at 2.4A resolution is reported. In the crystal structure, ferric Cgb is dimerized through two intermolecular disulfide bonds between Cys38(B2) and Cys83(E9), and the dimerization interface is similar to that of lamprey Hb and Ngb. The overall backbone structure of the Cgb monomer exhibits a traditional globin fold with a three-over-three alpha-helical sandwich, in which the arrangement of helices is basically the same among all globins studied to date. A detailed comparison reveals that the backbone structure of the CD corner to D helix region, the N terminus of the E-helix and the F-helix of Cgb resembles more closely those of pentacoordinated globins (Mb, lamprey Hb), rather than hexacoordinated globins (Ngb, rice Hb). However, the His81(E7) imidazole group coordinates directly to the heme iron as a sixth axial ligand to form a hexcoordinated heme, like Ngb and rice Hb. The position and orientation of the highly conserved residues in the heme pocket (Phe(CD1), Val(E11), distal His(E7) and proximal His(F8)) are similar to those of other globin proteins. Two alternative conformations of the Arg84(E10) guanidium group were observed, suggesting that it participates in ligand binding to Cgb, as is the case for Arg(E10) of Aplysia Mb and Lys(E10) of Ngb. The structural diversities and similarities among globin proteins are discussed with relevance to molecular evolutionary relationships.     

Bacterial and archaeal globins — A revised perspective

A bioinformatics survey of putative globins in over 2200 bacterial and some 140 archaeal genomes revealed that over half the bacterial and approximately one fifth of archaeal genomes contain genes encoding globins that were classified into three families: the M (myoglobin-like), and S (sensor) families all exhibiting the canonical 3/3 myoglobin fold, and the T family (truncated myoglobin fold). Although the M family comprises 2 subfamilies, flavohemoglobins (FHbs) and single domain globins (SDgbs), the S family encompasses chimeric globin-coupled sensors (GCSs), single domain Pgbs (protoglobins) and SSDgbs (sensor single domain globins). The T family comprises three classes TrHb1s, TrHb2s and TrHb3s, characterized by the abbreviated 2/2 myoglobin fold. The Archaea contain only Pgbs, GCSs and TrHb1s. The smallest globin-bearing genomes are the streamlined genomes (~ 1.3 Mbp) of the SAR11 clade of alphaproteobacteria and the slightly larger (ca. 1.7 Mbp) genomes of Aquificae. The smallest genome with members of all three families is the 2.3 Mbp genome of the extremophile Methylacidiphilum infernorum (Verrumicrobia). Of the 147 possible combinations of the eight globin subfamilies, only 83 are observed. Although binary combinations are infrequent and ternary combinations are rare, the FHb + TrHb2 combination is the most commonly observed. Of the possible functions of bacterial globins we discuss the two principal ones — nitric oxide detoxification via the NO dioxygenase or denitrosylase activities and the sensing of oxygen concentration in the environmental niche. In only few cases has a physiological role been demonstrated in vivo. This article is part of a Special Issue entitled: Oxygen Binding and Sensing Proteins.     

Variation of folded polypeptide surface area with probe size

Three types of polypeptide surface area (contact, accessible, and molecular) have been studied as a function of the radius of a probe sphere used to map the surface. The surfaces are: (1) three alpha-helices, the H-helix of myoglobin, the E-helix of leghemoglobin, and an artificial polyalanine helix, each with 26 residues; (2) two globins, myoglobin and leghemoglobin, each with 153 residues; and (3) a two-center model system for which the three types of surface area have been calculated analytically. The two globin helices have almost identical surface areas as a function of probe size as do the two globins. The polyalanine helix surface area is smaller but similar in shape to the globin helix areas. All three helix contact areas tend to the same limit as the probe size increases, and the globin contact areas behave similarly. Fractal dimensions were calculated for the helix and globin contact and molecular surfaces. All fractal dimensions showed strong dependence on probe size. The contact fractal dimension peaks at larger values for both the helices and globins. Most residues do not make contact with large probes (15 A).     

Amino acid sequence of the dimeric hemoglobin (Hb I) from the deep-sea cold-seep clam Calyptogena soyoae and the phylogenetic relationship with other molluscan globins

The deep-sea cold-seep clam Calyptogena soyoae has two homodimeric hemoglobins (Hbs I and II) in erythrocytes. The complete amino acid sequence of Hb I has been determined. It is composed of 144 amino acid residues, has a high content of hydrophobic residues, and a calculated molecular weight of 16,350 including a heme group. The sequence of Calyptogena Hb I showed high homology (42% identity) with that of Calyptogena Hb II (Suzuki, T., Takagi T. and Ohta, S. (1989) Biochem. J. 260, 177-182), although it has a long insertion of seven residues in the C-terminal region compared with Hb II. On the other hand, it showed low homology (12-20% identity) with other molluscan globins. As well as Hb II, Calyptogena Hb I lacked the N-terminal extension of 7-9 residues characteristic of molluscan intracellular hemoglobins, and the distal (E7) histidine was replaced by glutamine. A phylogenetic tree was constructed from 13 molluscan globins belonging to the five families Aplysiidae, Galeodidae, Potamididae, Arcidae and Vesicomyidae. The globin sequences of Calyptogena (Vesicomyidae) were found to be rather distant from other globin sequences, suggesting that they might conserve a primitive form of molluscan globins.     

Divergence pattern and selective mode in protein evolution: The example of vertebrate myoglobins and hemoglobin chains

Summary The evolutionary relation of vertebrate myoglobin and the hemoglobin chains including the agnathan hemoglobin chain is investigated on the basis of a new view of amino acid changes that is developed by canonical discriminant analysis of amino acid residues at individual sites. In contrast to the clear discrimination of amino acid residues between myoglobin, hemoglobin a chain, and hemoglobin chain in warm-blood vertebrates, the three types of globins in the lower class of vertebrates show so much variation that they are not well discriminated. This is seen particularly at the sites that are ascertained in mammals to carry the amino acid residues participating in stabilizing the monomeric structure in myoglobin and the residues forming the subunit contacts in hemoglobin. At these sites, agnathan hemoglobin chains are evaluated to be intermediate between the myoglobin and hemoglobin chains of gnathostomes. The variation in the phylogenetically lower class of globins is also seen in the internal region; there the amino acid residues of myoglobin and hemoglobin chains in the phylogenetically higher class exhibit an example of parallel evolution at the molecular level. New quantities, the distance of sequence property between discriminated groups and the variation within each group, are derived from the values of discriminant functions along the peptide chain, and this set of quantities simply describes an overall feature of globins such that the distinction between the three types of globins has been clearer as the vertebrates have evolved to become jawed, landed, and warm-blooded. This result strongly suggests that the functional constraint on the amino acid sequence of a protein is changed by living conditions and that severe conditions constitute a driving force that creates a distinctive protein from a less-constrained protein.Offprint requests to: J. Otsuka     

A globin gene of ancient evolutionary origin in lower vertebrates: evidence for two distinct globin families in animals

Hemoglobin, myoglobin, neuroglobin, and cytoglobin are four types of vertebrate globins with distinct tissue distributions and functions. Here, we report the identification of a fifth and novel globin gene from fish and amphibians, which has apparently been lost in the evolution of higher vertebrates (Amniota). Because its function is presently unknown, we tentatively call it globin X (GbX). Globin X sequences were obtained from three fish species, the zebrafish Danio rerio, the goldfish Carassius auratus, and the pufferfish Tetraodon nigroviridis, and the clawed frog Silurana tropicalis. Globin X sequences are distinct from vertebrate hemoglobins, myoglobins, neuroglobins, and cytoglobins. Globin X displays the highest identity scores with neuroglobin (approximately 26% to 35%), although it is not a neuronal protein, as revealed by RT-PCR experiments on goldfish RNA from various tissues. The distal ligand-binding and the proximal heme-binding histidines (E7 and F8), as well as the conserved phenylalanine CD1 are present in the globin X sequences, but because of extensions at the N-terminal and C-terminal, the globin X proteins are longer than the typical eight alpha-helical globins and comprise about 200 amino acids. In addition to the conserved globin introns at helix positions B12.2 and G7.0, the globin X genes contain two introns in E10.2 and H10.0. The intron in E10.2 is shifted by 1 bp in respect to the vertebrate neuroglobin gene (E11.0), providing possible evidence for an intron sliding event. Phylogenetic analyses confirm an ancient evolutionary relationship of globin X with neuroglobin and suggest the existence of two distinct globin types in the last common ancestor of Protostomia and Deuterostomia.     

Fragmentation and reduction of bovine secretory component. Preparation of a biologically active fragment and some evidence for a multiple-domain structure.

A tryptic fragment (A) of Mr 25000 was prepared from bovine secretory component. The fragment binds polymeric immunoglobulin, although 9 times less effectively than secretory component on a molar basis. The fragment has four buried half-cystine residues and two exposed half-cystine residues. It gives rise to two fragments of Mr 11000-13000 on prolonged digestion with trypsin, and these do not bind polymeric immunoglobulin. It is proposed that fragment A consists of two immunoglobulin-like domains. Bovine secretory component was found to have 9-11 buried half-cystine residues and four exposed half-cystine residues. Reduction and alkylation of the exposed residues decreases the binding of polymeric immunoglobulin by 3-fold. Initial tryptic cleavage of bovine secretory component gives a fragment (Q) disulphide-bridged to a further fragment (T). Fragment Q is similar in size to a three-domain immunoglobulin fragment, and fragment T is similar in size to a two-domain immunoglobulin fragment. The two-domain fragment A is derived from fragment Q by further tryptic cleavage. The results are compatible with the proposal by Mostov, Friedlander & Blobel [(1984) Nature (London) 308, 37-43] that secretory component consists of multiple immunoglobulin-like domains. The results also indicate that optimal binding of polymeric immunoglobulin involves several domains stabilized by an exposed disulphide bridge.     

Isolation and cDNA-derived amino acid sequences of hemoglobin and myoglobin from the deep-sea clam Calyptogena kaikoi

The heterodont clam Calyptogena kaikoi, living in the cold-seep area at a depth of 3761 m of the Nankai Trough, Japan, has abundant hemoglobins and myoglobins in erythrocytes and adductor muscle, respectively. Two types of hemoglobins (Hb I and Hb II) were isolated, and the complete amino acid sequences of Hb I (145 residues) and Hb II (137 residues) were obtained with combination of cDNA and protein sequencing. The amino acid sequences of C. kaikoi Hbs I and II differed from homologous chains of the congeneric clam Calyptogena soyoae in eight and five positions, respectively. The distal (E7) His, one of the functionally important residues in hemoglobin and myoglobin, was replaced by Gln in hemoglobins of C. kaikoi. A phylogenetic analysis of clam hemoglobins indicates that the evolutionary rate of Calyptogena hemoglobins is rather faster than those of other clams, suggesting that the mutation rate might be accelerated in the deep-sea animals around the areas of cold seeps or hydrothermal vents. On the other hand, it was found unexpectedly that two myoglobins Mbs I and II, isolated from the red adductor muscle, are identical in amino acid sequence Hbs I and II, respectively. Thus it was assumed that genes for Hbs I and II are also expressed in the muscle of C. kaikoi in substitution for myoglobin gene. This suggests that the major physiological role of globins in C. kaikoi is storage of oxygen under the low oxygen conditions, rather than circulating of oxygen.     

The 1.7 A crystal structure of BPI: a study of how two dissimilar amino acid sequences can adopt the same fold

We have extended the resolution of the crystal structure of human bactericidal/permeability-increasing protein (BPI) to 1.7 A. BPI has two domains with the same fold, but with little sequence similarity. To understand the similarity in structure of the two domains, we compare the corresponding residue positions in the two domains by the method of 3D-1D profiles. A 3D-1D profile is a string formed by assigning each position in the 3D structure to one of 18 environment classes. The environment classes are defined by the local secondary structure, the area of the residue which is buried from solvent, and the fraction of the area buried by polar atoms. A structural alignment between the two BPI domains was used to compare the 3D-1D environments of structurally equivalent positions. Greater than 31% of the aligned positions have conserved 3D-1D environments, but only 13% have conserved residue identities. Analysis of the 3D-1D environmentally conserved positions helps to identify pairs of residues likely to be important in conserving the fold, regardless of the residue similarity. We find examples of 3D-1D environmentally conserved positions with dissimilar residues which nevertheless play similar structural roles. To generalize our findings, we analyzed four other proteins with similar structures yet dissimilar sequences. Together, these examples show that aligned pairs of dissimilar residues often share similar structural roles, stabilizing dissimilar sequences in the same fold.     

O2 migration pathways are not conserved across proteins of a similar fold

Recent advances in computational biology have made it possible to map the complete network and energy profile of gas migration pathways inside proteins. Although networks of O(2) pathways have already been characterized for a small number of proteins, the general properties and locations of these pathways have not been previously compared between proteins. In this study, maps of the O(2) pathways inside 12 monomeric globins were computed. It is found that, despite the conserved tertiary structure fold of the studied globins, the shape and topology of O(2) pathway networks exhibit a large variability between different globins, except when two globins are nearly identical. The locations of the O(2) pathways are, however, found to be correlated with the location of large hydrophobic residues, and a similar correlation is observed in two unrelated protein families: monomeric globins and copper-containing amine oxidases. The results have implications for the evolution of gas pathways in proteins and for protein engineering applications involving modifications of these pathways.     

Neuroglobin and cytoglobin in search of their role in the vertebrate globin family

Neuroglobin and cytoglobin are two recent additions to the family of heme-containing respiratory proteins of man and other vertebrates. Here, we review the present state of knowledge of the structures, ligand binding kinetics, evolution and expression patterns of these two proteins. These data provide a first glimpse into the possible physiological roles of these globins in the animal's metabolism. Both, neuroglobin and cytoglobin are structurally similar to myoglobin, although they contain distinct cavities that may be instrumental in ligand binding. Kinetic and structural studies show that neuroglobin and cytoglobin belong to the class of hexa-coordinated globins with a biphasic ligand-binding kinetics. Nevertheless, their oxygen affinities resemble that of myoglobin. While neuroglobin is evolutionarily related to the invertebrate nerve-globins, cytoglobin shares a more recent common ancestry with myoglobin. Neuroglobin expression is confined mainly to brain and a few other tissues, with the highest expression observed in the retina. Present evidence points to an important role of neuroglobin in neuronal oxygen homeostasis and hypoxia protection, though other functions are still conceivable. Cytoglobin is predominantly expressed in fibroblasts and related cell types, but also in distinct nerve cell populations. Much less is known about its function, although in fibroblasts it might be involved in collagen synthesis.     

Glomerular basement membrane: evidence for collagenous domain of the alpha 3 and alpha 4 chains of collagen IV

A collagenous component(s) of Mr = 60K was extracted from glomerular basement membrane with urea and was purified. Upon digestion, it yielded a collagenase-resistant fragment(s) of Mr = 23.5K. Both component and fragment showed immunochemical identity with the noncollagenous domains of the new alpha 3 & alpha 4 chains of collagen IV. The component is characterized by a collagenous domain of about 280 residues and a noncollagenous domain of about 250 residues. These findings further establish these new chains as distinct entities of collagen IV.     

Crystal structure of Escherichia coli RNase D, an exoribonuclease involved in structured RNA processing

RNase D (RND) is one of seven exoribonucleases identified in Escherichia coli. RNase D has homologs in many eubacteria and eukaryotes, and has been shown to contribute to the 3' maturation of several stable RNAs. Here, we report the 1.6 A resolution crystal structure of E. coli RNase D. The conserved DEDD residues of RNase D fold into an arrangement very similar to the Klenow fragment exonuclease domain. Besides the catalytic domain, RNase D also contains two structurally similar alpha-helical domains with no discernible sequence homology between them. These closely resemble the HRDC domain previously seen in RecQ-family helicases and several other proteins acting on nucleic acids. More interestingly, the DEDD catalytic domain and the two helical domains come together to form a ring-shaped structure. The ring-shaped architecture of E. coli RNase D and the HRDC domains likely play a major role in determining the substrate specificity of this exoribonuclease.     

Three-dimensional structure of cyclodextrin glycosyltransferase from Bacillus circulans at 3.4 A resolution

Cyclodextrin glycosyltransferase (EC 2.4.1.19) from Bacillus circulans has been purified, crystallized and analyzed by X-ray diffraction. The enzyme is monomeric. SDS/polyacrylamide gel electrophoresis gave an Mr of 73,600(+/- 1000), corresponding to 670(+/- 10) amino acid residues. The structure of the crystalline enzyme has been elucidated at a resolution of 3.4 A, using multiple isomorphous replacement and solvent flattening for phase determination. The resulting electron density map allowed tracing of the polypeptide chain; 664 residue positions have been assigned. The chain fold has been subdivided into five domains. The N-terminal domain forms a (beta alpha)8-barrel, which contains the second domain of about 55 residues as an insert after the third beta-strand. The three remaining domains form almost exclusively beta-pleated sheet structures and consist of about 90, 80 and 95 residues. The chain fold of the three N-terminal domains of 492 residues resembles closely the two known structures of alpha-amylases. This geometric similarity corresponds to the observed amino acid sequence homology. On the basis of the sequence homology with alpha-amylases, the active center can be located. The fourth domain has an immunoglobulin fold and is far away from the active center, while the fifth domain participates in the formation of the broad depression at the active center. Accordingly, the cyclodextrin glycosyltransferase chain fold can be considered as an alpha-amylase chain fold with two additional domains.     

Arabidopsis proteins containing similarity to the universal stress protein domain of bacteria

We have collected a set of 44 Arabidopsis proteins with similarity to the USPA (universal stress protein A of Escherichia coli) domain of bacteria. The USPA domain is found either in small proteins, or it makes up the N-terminal portion of a larger protein, usually a protein kinase. Phylogenetic tree analysis based upon a multiple sequence alignment of the USPA domains shows that these domains of protein kinases 1.3.1 and 1.3.2 form distinct groups, as do the protein kinases 1.4.1. This indicates that their USPA domain structures have diverged appreciably and suggests that they may subserve distinct cellular functions. Two USPA fold classes have been proposed: one based on Methanococcus jannaschii MJ0577 (1MJH) that binds ATP, and the other based on the Haemophilus influenzae universal stress protein (1JMV), highly similar to E. coli UspA, which does not bind ATP. A set of common residues involved in ATP binding in 1MJH and conserved in similar bacterial sequences is also found in a distinct cluster of Arabidopsis sequences. Threading analysis, which examines aspects of secondary and tertiary structure, confirms this Arabidopsis sequence cluster as highly similar to 1MJH. This structural approach can distinguish between the characteristic fold differences of 1MJH-like and 1JMV-like bacterial proteins and was used to assign the complete set of candidate Arabidopsis proteins to one of these fold classes. It is clear that all the plant sequences have arisen from a 1MJH-like ancestor.     

Crystal structure of a lysine biosynthesis enzyme,LysX, from Thermus thermophilus HB8

The thermophilic bacterium Thermus thermophilus synthesizes lysine through the alpha-aminoadipate pathway, which uses alpha-aminoadipate as a biosynthetic intermediate of lysine. LysX is the essential enzyme in this pathway, and is believed to catalyze the acylation of alpha-aminoadipate. We have determined the crystal structures of LysX and its complex with ADP at 2.0A and 2.38A resolutions, respectively. LysX is composed of three alpha+beta domains, each composed of a four to five-stranded beta-sheet core flanked by alpha-helices. The C-terminal and central domains form an ATP-grasp fold, which is responsible for ATP binding. LysX has two flexible loop regions, which are expected to play an important role in substrate binding and protection. In spite of the low level of sequence identity, the overall fold of LysX is surprisingly similar to that of other ATP-grasp fold proteins, such as D-Ala:D-Ala ligase, PurT-encoded glycinamide ribonucleotide transformylase, glutathione synthetase, and synapsin I. In particular, they share a similar spatial arrangement of the amino acid residues around the ATP-binding site. This observation strongly suggests that LysX is an ATP-utilizing enzyme that shares a common evolutionary ancestor with other ATP-grasp fold proteins possessing a carboxylate-amine/thiol ligase activity.     

Recombinant protein sequences can trigger methylation of N-terminal amino acids in Escherichia coli.

Recombinant human hemoglobin rHb1.1 has been genetically engineered with the replacement of the wild-type valine residues at all N-termini with methionine, an Asn 108 Lys substitution on the beta globins, and a fusion of the two alpha globins with a glycine linker. When rHb1.1 was expressed in Escherichia coli, methylation of the N-terminal methionine of the alpha globin was discovered. Another mutant has been engineered with the alpha globin gene coding for N-terminal methionine followed by an insertion of alanine. Characterization of expressed hemoglobin from this variant revealed a methylated N-terminal alanine that occurred through two posttranslational events: initial excision of the N-terminal methionine, followed by methylation of alanine as the newly generated N-terminus. No methylation was observed for variants expressed with wild-type valine at the N-terminus of the alpha globin. The methylation of N-terminal amino acids was attributed to a specific protein sequence that can trigger methylation of proteins expressed in E. coli. Here we demonstrate that proline at position 4 in the protein sequence of alpha globin seems an essential part of that signaling. Although N-terminal methylation has been observed previously for native E. coli proteins with similar N-terminal sequences, methylation of the recombinant globins has allowed further delineation of the recognition sequence, and indicates that methylation of heterologous proteins can occur in E. coli.     

Identification of cDNA clones encoding different domains of the basement membrane heparan sulfate proteoglycan

We have used antibodies to the basement membrane proteoglycan to screen lambda gt11 expression vector libraries and have isolated two cDNA clones, termed BPG 5 and BPG 7, which encode different portions of the core protein of the heparan sulfate basement membrane proteoglycan. These clones hybridize to a single mRNA species of approximately 12 kilobases. Amino acid sequences obtained on peptides derived from protease digests of the core protein were found in the deduced sequence, confirming the identity of these clones. BPG 5 spanned 1986 base pairs and has an open reading frame of 662 amino acids. The amino acid sequence deduced from BPG 5 contains two cysteine-rich domains and two internally homologous domains lacking cysteine. The cysteine-rich domains show homology to the cysteine-rich domains of the laminin chains. A globule-rod structure, similar to that of the short arms of the laminin chains, is proposed for this region of the proteoglycan. The other clone, BPG 7, is 2193 base pairs long and has an open reading frame of 731 amino acids. The deduced sequence contains eight internal repeats with 2 cysteine residues in each repeat. These repeats show homology to the neural-cell adhesion molecule N-CAM and the plasma alpha 1B-glycoprotein. Looping structures similar to these proteins and to other proteins of the immunoglobulin gene superfamily are proposed for this region of the proteoglycan. The sequence DSGEY was found four times in this domain and could be heparan sulfate attachment sites.