Structure of the chicken link protein gene: exons correlate with the protein domains

Directed evolution of proteins by exon shuffling

Evolution of eukaryotes is mediated by sexual recombination of parental genomes. Crossovers occur in random, but homologous, positions at a frequency that depends on DNA length. As exons occupy only 1% of the human genome and introns about 24%, by far most of the crossovers occur between exons, rather than inside. The natural process of creating new combinations of exons by intronic recombination is called exon shuffling. Our group is developing in vitro formats for exon shuffling and applying these to the directed evolution of proteins. Based on the splice frame junctions, nine classes of exons and three classes of introns can be distinguished. Splice frame diagrams of natural genes show how the splice frame rules govern exon shuffling. Here, we review various approaches to constructing libraries of exon-shuffled genes. For example, exon shuffling of human pharmaceutical proteins can generate libraries in which all of the sequences are fully human, without the point mutations that raise concerns about immunogenicity.

Domain organization, genomic structure, evolution, and regulation of expression of the aggrecan gene family

Proteoglycans are complex macromolecules, consisting of a polypeptide backbone to which are covalently attached one or more glycosaminoglycan chains. Molecular cloning has allowed identification of the genes encoding the core proteins of various proteoglycans, leading to a better understanding of the diversity of proteoglycan structure and function, as well as to the evolution of a classification of proteoglycans on the basis of emerging gene families that encode the different core proteins. One such family includes several proteoglycans that have been grouped with aggrecan, the large aggregating chondroitin sulfate proteoglycan of cartilage, based on a high number of sequence similarities within the N- and C-terminal domains. Thus far these proteoglycans include versican, neurocan, and brevican. It is now apparent that these proteins, as a group, are truly a gene family with shared structural motifs on the protein and nucleotide (mRNA) levels, and with nearly identical genomic organizations. Clearly a common ancestral origin is indicated for the members of the aggrecan family of proteoglycans. However, differing patterns of amplification and divergence have also occurred within certain exons across species and family members, leading to the class-characteristic protein motifs in the central carbohydrate-rich region exclusively. Thus the overall domain organization strongly suggests that sequence conservation in the terminal globular domains underlies common functions, whereas differences in the central portions of the genes account for functional specialization among the members of this gene family.

Gene structure of chick cartilage chondroitin sulfate proteoglycan (aggrecan) core protein

Large aggregating chondroitin sulfate proteoglycan (CSPG/aggrecan) is one of the major extracellular matrix components in cartilage. The core protein is also large, over 200 kDa, and modular with a distinct correspondence between protein structural domains and the encoding exons. Here we report the isolation, using chick CSPG cDNA probes and the ensuing sequencing, of genomic clones containing exons encoding the chick CSPG core protein. The 5' two globular domains, G1 and G2, are encoded by four and three exons, respectively, and the interglobular domain is encoded by a single exon. The chondroitin sulfate attachment domain is encoded by the largest exon, 3,216 bp, which is approximately 50% of the total coding sequence. Combined with the previous report (Tanaka, T., Har-el, R. Tanzer, M.L. 1988 J. Biol. Chem. 263, 15831-15835), these data reveal that the chick CSPG gene contains at least 18 exons spanning a genome which is greater than 30 kb. No evidence was obtained for multiple genes for aggrecan in the chick genome. Elucidation of the chick genomic structure allows comparison of the avian and mammalian link protein genes to the homologous portions of avian and mammalian core protein genes (hyaluronate binding domain) with respect to their origins and paths of duplication and divergence.

Structure of the human aggrecan gene: exon-intron organization and association with the protein domains

The complete exon-intron organization of the human aggrecan gene has been defined, and the exon organization has been compared with the individual domains of the protein core. A yeast artificial chromosome containing the aggrecan gene was selected from the Centre d'Etude du Polymorphisme Humaine yeast artificial chromosome library. A cosmid sulibrary was created from this, and direct sequencing of individual cosmids was used to provide the exon-intron organization. The human aggrecan gene was found to be composed of 19 exons ranging in size from 77 to 4224 bp. Exon 1 is non-coding, whereas exons 2-19 code for a protein core of 2454 amino acids with a calculated mass of 254379 Da. Intron 1 of the gene is at least 13 kb. Overall, the sizes of the 18 introns range from 0.5 to greater than 13 kb. Each intron begins with a GT and ends with an AG, thus obeying the GT/AG rule of splice-junction sequences. The entire coding region is contained in 39.4 kb of the gene. The organization of exons is strongly related to the specific domains of the protein core. The A loop of G1 and the interglobular domain are encoded by exons 3 and 7 respectively. The B and B' loops of G1 are encoded by exons 4-6, and those of G2 are encoded by exons 8-10. These sets of exons, coding for the B and B' loops, are identical in size and organization. This is supported by the intron classes associated with these exons. Exon 11 codes for the 5' half of the keratan sulphate-rich region, and exon 12 codes for the 3' half of the keratan sulphate-rich region as well as the entire chondroitin sulphate-rich region. G3 is encoded by exons 13-18, including the alternatively spliced epidermal growth factor-like and complement regulatory protein-like domains. The correspondence between the exon organization and the protein domains argues strongly for modular assembly of the aggrecan gene.

Cartilage proteoglycans: structure and potential functions

Hyaline cartilage contains five well-characterized proteoglycans in its extracellular matrix, and it is likely that others exist. The largest in size and most abundant by weight is aggrecan, a proteoglycan that possesses over 100 chondroitin sulfate and keratan sulfate chains. Aggrecan is also characterized by its ability to interact with hyaluronic acid to form large proteoglycan aggregates. Both the high anionic charge on the individual aggrecan molecules endowed by the sulfated glycosaminoglycan chains and the localization within the matrix endowed by aggregate formation are essential for aggrecan function. The molecule provides cartilage with its osmotic properties, which give articular cartilage its ability to resist compressive loads. The other proteoglycans are characterized by their ability to interact with collagen. They are much smaller than aggrecan in size but may be present in similar molar amounts. Decorin, biglycan, and fibromodulin are closely related in protein structure but differ in glycosaminoglycan composition and function. Decorin and biglycan possess one and two dermatan sulfate chains, respectively, whereas fibromodulin bears several keratan sulfate chains. Decorin and fibromodulin both interact with the type II collagen fibrils in the matrix and may play a role in fibrillogenesis and interfibril interactions. Biglycan is preferentially localized in the pericellular matrix, where it may interact with type VI collagen. Finally, type IX collagen can also be considered as a proteoglycan, as its alpha 2(IX) chain may bear a glycosaminoglycan chain. It may serve as a bridge between the collagen fibrils or with the interspersed aggrecan network.

The expression of functional link protein in a baculovirus system: analysis of mutants lacking the A, B and B' domains

Functional recombinant human link protein has been produced using a baculovirus expression system. In addition to the intact link protein, three mutant forms have also been expressed. Each mutant bears a deletion equivalent to the protein encoded by one exon in the gene. These deletions represent the A domain, which is thought to be responsible for interaction with aggrecan, and the B or B' domains, which are associated with the interaction with hyaluronate. Such deletions split codons spanning exon boundaries, but maintain the reading frame of the protein and result in the correct amino acid being present at the splice junction. All the recombinant proteins appear as two components upon SDS/PAGe, though the abundance of the two forms does vary between preparations, as a result of variable substitution by N-linked oligosaccharides. The recombinant intact link protein was able to interact with both hyaluronate and aggrecan, showing that the baculovirus system is able to produce functional molecules. All of the recombinant mutant link proteins were also able to interact with hyaluronate, indicating that both the B and B' domains can function independently. The recombinant mutant link proteins were also able to interact with aggrecan, with the exception of the mutant lacking the A domain, confirming that this ability resides entirely within this domain.

The link proteins

Aggregates of chondroitin-keratan sulfate proteoglycan (aggrecan) and hyaluronic acid (hyaluronan) are the major space-filling components of cartilage. A glycoprotein, link protein (LP; 40-48 kDa) stabilizes the aggregate by binding to both hyaluronic acid and aggrecan. In the absence of LP, aggregates are smaller (as estimated by rotary shadowing of electron micrographs) and less stable (they dissociate at pH 5) than they are in the presence of LP. The proteoglycan aggregate, including LP, is dissociated in the presence of chaotropes such as 4 M guanidine hydrochloride. On removal of the chaotrope, the complex will reassociate. This forms the basis of the isolation of LP from cartilage and has been described in detail elsewhere. Tryptic digestion of the proteoglycan aggregates results in a high molecular weight product that consists of hyaluronic acid to which is bound LP and the N-terminal globular domain of aggrecan (hyaluronic acid binding region; HABR) in a 1:1 stoichiometry. The amino acid sequences of LP and HABR are surprisingly similar. The amino acid sequence can be divided into three domains; an N-terminal domain that falls into the immunoglobulin super-family and two C-terminal domains that are similar to each other. The DNA structure echoes this similarity, in that the major domains are reflected in three separate exons in both LP and HABR. The two C-terminal domains are largely responsible for the association with HA and are related to two recently described hyaluronate-binding proteins, CD44 and TSG-6. A variety of approaches, including analysis of the forms of LP in vivo, rotary shadowing and analysis of the sequence in the immunoglobulin-like domain, have shed considerable light on the structure-function relationships of LP. This review describes the structure and function of LP in detail, focusing on what can be inferred from the similarity of LP, HABR and related molecules such as immunoglobulins and lymphocyte HA-receptors.

The link proteins

Aggregates of chondroitin-keratan sulfate proteoglycan (aggrecan) and hyaluronic acid (hyaluronan) are the major space-filling components of cartilage. A glycoprotein, link protein (LP; 40-48 kDa) stabilizes the aggregate by binding to both hyaluronic acid and aggrecan. In the absence of LP, aggregates are smaller (as estimated by rotary shadowing of electron micrographs) and less stable (they dissociate at pH 5) than they are in the presence of LP. The proteoglycan aggregate, including LP, is dissociated in the presence of chaotropes such as 4 M guanidine hydrochloride. On removal of the chaotrope, the complex will reassociate. This forms the basis of the isolation of LP from cartilage and has been described in detail elsewhere. Tryptic digestion of the proteoglycan aggregates results in a high molecular weight product that consists of hyaluronic acid to which is bound LP and the N-terminal globular domain of aggrecan (hyaluronic acid binding region; HABR) in a 1:1 stoichiometry. The amino acid sequences of LP and HABR are surprisingly similar. The amino acid sequence can be divided into three domains; an N-terminal domain that falls into the immunoglobulin super-family and two C-terminal domains that are similar to each other. The DNA structure echoes this similarity, in that the major domains are reflected in three separate exons in both LP and HABR. The two C-terminal domains are largely responsible for the association with HA and are related to two recently described hyaluronate-binding proteins, CD44 and TSG-6. A variety of approaches, including analysis of the forms of LP found in vivo, rotary shadowing and analysis of the sequence in the immunoglobulin-like domain, have shed considerable light on the structure-function relationships of LP. This review describes the structure and function of LP in detail, focusing on what can be inferred from the similarity of LP, HABR and related molecules such as immunoglobulins and lymphocyte HA-receptors.

Versican gene expression in human articular cartilage and comparison of mRNA splicing variation with aggrecan

The chondrocytes in human articular cartilage from subjects of all ages express mRNAs for both of the aggregating proteoglycans aggrecan and versican, although the level of expression of versican mRNA is much lower than that of aggrecan mRNA. Aggrecan shows alternative splicing of the epidermal growth factor (EGF)-like domain within its C-terminal globular region, but there is no evidence for a major difference in situ in the relative expression of this domain with age. At all ages studied from birth to the mature adult, a greater proportion of transcripts lacked the EGF domain. The relative proportions of the two transcripts did not change upon culture and passage of isolated chondrocytes. In contrast, the neighbouring complement regulatory protein (CRP)-like domain was predominantly expressed irrespective of age, but cell culture did result in variation of the splicing of this domain. Versican possesses two EGF-like domains and one CRP-like domain, but at all ages the three domains were predominantly present in all transcripts. This situation persisted upon culture and passage of the chondrocytes. Thus, unlike aggrecan, the versican expressed by human articular cartilage does not appear to undergo alternative splicing of its C-terminal globular region, either in cartilage in situ or in chondrocytes in culture.

Molecular modeling of the multidomain structures of the proteoglycan binding region and the link protein of cartilage by neutron and synchrotron X-ray scattering

The interaction of proteoglycan monomers with hyaluronate in cartilage is mediated by a globular binding region at the N-terminus of the proteoglycan monomer; this interaction is stabilized by link protein. Sequences show that both the binding region (27% carbohydrate) and the link protein (6% carbohydrate) contain an immunoglobulin (Ig) fold domain and two proteoglycan tandem repeat (PTR) domains. Both proteins were investigated by neutron and synchrotron X-ray solution scattering, in which nonspecific aggregate formation was reduced by the use of citraconylation to modify surface lysine residues. The neutron and X-ray radius of gyration RG of native and citraconylated binding region is 5.1 nm, and the cross-sectional RG (RXS) is 1.9-2.0 nm. No neutron contrast dependence of the RG values was observed; however, a large contrast dependence was seen for the RXS values which is attributed to the high carbohydrate content of the binding region. The neutron RG for citraconylated link protein is 2.9 nm, its RXS is 0.8 nm, and these data are also independent of the neutron contrast. The scattering curves of binding region and link protein were modeled using small spheres. Both protein structures were defined initially by the representation of one domain by a crystal structure for a variable Ig fold and a fixed volume for the two PTR domains calculated from sequence data. The final models showed that the different dimensions and neutron contrast properties of binding region compared to link protein could be attributed to an extended glycosylated C-terminal peptide with extended carbohydrate structures in the binding region.(ABSTRACT TRUNCATED AT 250 WORDS)

Complex pattern of alternative splicing generates unusual diversity in the leader sequence of the chicken link protein mRNA

We report here the isolation of the 5' end and the promoter region of the gene for chicken cartilage link protein, and demonstrate extensive heterogeneity of the leader sequence arising from differential utilization of multiple splice sites within the 5'-most exon. The 500-base pairs (bp) exon 1 consists of solely untranslated sequence and is followed by an intron greater than 33 kilobase pairs (kb). Together, the five exons predict a gene size longer than 100 kb. Multiple transcription initiation sites were mapped 34, 46, 56, 66 and 76 bp downstream of a TATA-like motif. Sequence analysis revealed that in addition to the non-spliced variant, multiple mRNA species were generated by alternative splicing resulting in the exclusion of 92, 166, 170, 174 and 263 nucleotides (nt), respectively, from exon 1. Polymerase chain reaction confirmed the existence of various splice forms, and showed cell type- and developmental stage-specific expression for one group of them. Secondary structure predictions indicated that the leaders of the splice forms could form stable hairpin structures with different free energies of formation (up to delta G = -110 kcal/mol), suggesting translational control. The splice variant detected in the largest amount had the least stable predicted hairpin (delta G = -31.7 kcal/mol).

Characterization of the promoter for the rat and human link protein gene

We have isolated the 5'-end of the gene for the rat and human link protein by screening genomic libraries with oligonucleotides corresponding to the 5'-cDNA sequence. Several overlapping clones were isolated for the human link protein gene, while only one clone was obtained for the rat. All the clones contained a single exon of which the sequence was identical to the most 5'-end of the rat and human cDNAs. Transcription initiation sites for the rat link gene were identified by primer extension and S1 protection analysis using total RNA from the rat Swarm chondrosarcoma. Transcriptional initiation sites for the human link gene were determined by specific primer extension of RNA from human fetal cartilage. Comparison of 1500 bp of 5'-flanking sequence between the rat and human link protein genes showed strong sequence conservation near the start site of transcription with 80% overall identity. Analysis of the 5'-flanking regions also revealed a large inverted repeat consisting of repeating purine-pyrimidine, which has the potential to form left-handed Z-DNA. Transcriptional regulation of the link protein gene was studied by coupling either 7.0 kb or 0.85 kb of 5'-flanking rat DNA to the chloramphenicol acetyltransferase (CAT) gene followed by transfection into chick embryonic chondrocytes (CEC) and HeLa cells. Both constructs had considerable CAT activity in CEC cells and less activity in HeLa cells. Furthermore, inclusion of a DNA fragment from the first intron increased relative CAT activity in both of these cell types. The increased activity from the first intron was shown to be orientation independent in CEC. These results indicate the presence of positive cisacting regulatory elements in both the promoter and first intron of the rat gene for link protein.

Interaction of the NG2 chondroitin sulfate proteoglycan with type VI collagen

The NG2 chondroitin sulfate proteoglycan is a membrane-associated molecule of approximately 500 kD with a core glycoprotein of 300 kD. Both the complete proteoglycan and a smaller quantity of the 300-kD core are immunoprecipitable with polyclonal and monoclonal antibodies against purified NG2. From some cell lines, the antibodies coprecipitate NG2 and type VI collagen, the latter appearing on SDS-PAGE as components of 140 and 250 kD under reducing conditions. The immunoprecipitation of type VI collagen does not seem to be due to recognition of the collagen by the antibodies, but rather to binding of the collagen to NG2. Studies on the NG2-type VI collagen complex suggest that binding between the two molecules is mediated by protein-protein interactions rather than by ionic interactions involving the glycosaminoglycans. Immunofluorescence double labeling in frozen sections of embryonic rat shows that NG2 and type VI collagen are colocalized in structures such as the intervertebral discs and arteries of the spinal column. In vitro the two molecules are highly colocalized on the surface of several cell lines. Treatment of these cells resulting in a change in the distribution of NG2 on the cell surface also causes a parallel change in type VI collagen distribution. Our results suggest that cell surface NG2 may mediate cellular interactions with the extracellular matrix by binding to type VI collagen.

Identification of positive and negative regulatory regions controlling expression of the cartilage matrix protein gene

A complex pattern of regulation of the cartilage matrix protein gene was revealed by transient expression experiments. A minimal promoter from positions -15 to +64 functioned in chondrocytes and fibroblasts. An enhancer located in the first intron exerted chondrocyte-specific stimulation on the minimal promoter activity. The same fragment, however, had a negative effect in fibroblasts. Between -334 and -15, a silencer was found which inhibited the gene expression driven from its homologous as well as heterologous promoters both in chondrocytes and fibroblasts. Additional positive and negative control regions were mapped further upstream of the promoter.

Identification of positive and negative regulatory regions controlling expression of the cartilage matrix protein gene

A complex pattern of regulation of the cartilage matrix protein gene was revealed by transient expression experiments. A minimal promoter from positions -15 to +64 functioned in chondrocytes and fibroblasts. An enhancer located in the first intron exerted chondrocyte-specific stimulation on the minimal promoter activity. The same fragment, however, had a negative effect in fibroblasts. Between -334 and -15, a silencer was found which inhibited the gene expression driven from its homologous as well as heterologous promoters both in chondrocytes and fibroblasts. Additional positive and negative control regions were mapped further upstream of the promoter.

The 5' splice site: phylogenetic evolution and variable geometry of association with U1RNA

The 5' splice site sequences of 3294 introns from various organisms (1-672) were analyzed in order to determine the rules governing evolution of this sequence, which may shed light on the mechanism of cleavage at the exon-intron junction. The data indicate that, currently, in all organisms, a common sequence 1GUAAG6U and its derivatives are used as well as an additional sequence and its derivatives, which differ in metazoa (G/1GUgAG6U), lower eucaryotes (1GUAxG6U) and higher plants (AG/1GU3A). They all partly resemble the prototype sequence AG/1GUAAG6U whose 8 contigous nucleotides are complementary to the nucleotides 4-11 of U1RNA, which are perfectly conserved in the course of phylogenetic evolution. Detailed examination of the data shows that U1RNA can recognize different parts of 5' splice sites. As a rule, either prototype nucleotides at position -2 and -1 or at positions 4, 5 or 6 or at positions 3-4 are dispensable provided that the stability of the U1RNA-5' splice site hybrid is conserved. On the basis of frequency of sequences, the optimal size of the hybridizable region is 5-7 nucleotides. Thus, the cleavage at the exon-intron junction seems to imply, first, that the 5' splice site is recognized by U1RNA according to a "variable geometry" program; second, that the precise cleavage site is determined by the conserved sequence of U1RNA since it occurs exactly opposite to the junction between nucleotides C9 and C10 of U1RNA. The variable geometry of the U1RNA-5' splice site association provides flexibility to the system and allows diversification in the course of phylogenetic evolution.