Circadian time series proteomics reveals daily dynamics in cartilage physiology.

OBJECTIVE: Cartilage in joints such as the hip and knee experiences repeated phases of heavy loading and low load recovery during the 24-h day/night cycle. Our previous work has shown 24 h rhythmic changes in gene expression at transcript level between night and day in wild type mouse cartilage which is lost in a circadian clock knock-out mouse model. However, it remains unknown to what extent circadian rhythms also regulate protein level gene expression in this matrix rich tissue. METHODS: We investigated daily changes of protein abundance in mouse femoral head articular cartilage by performing a 48-h time-series LC-MS/MS analysis. RESULTS: Out of the 1,177 proteins we identified across all time points, 145 proteins showed rhythmic changes in their abundance within the femoral head cartilage. Among these were molecules that have been implicated in key cartilage functions, including CTGF, MATN1, PAI-1 and PLOD1 & 2. Pathway analysis revealed that protein synthesis, cytoskeleton and glucose metabolism exhibited time-of-day dependent functions. Analysis of published cartilage proteomics datasets revealed that a significant portion of rhythmic proteins were dysregulated in osteoarthritis and/or ageing. CONCLUSIONS: Our circadian proteomics study reveals that articular cartilage is a much more dynamic tissue than previously thought, with chondrocytes driving circadian rhythms not only in gene transcription but also in protein abundance. Our results clearly call for the consideration of circadian timing mechanisms not only in cartilage biology, but also in the pathogenesis, treatment strategies and biomarker detection in osteoarthritis.

A dominant TRPV4 variant underlies osteochondrodysplasia in Scottish fold cats.

OBJECTIVE: Scottish fold cats, named for their unique ear shape, have a dominantly inherited osteochondrodysplasia involving malformation in the distal forelimbs, distal hindlimbs and tail, and progressive joint destruction. This study aimed to identify the gene and the underlying variant responsible for the osteochondrodysplasia. DESIGN: DNA samples from 44 Scottish fold and 54 control cats were genotyped using a feline DNA array and a case-control genome-wide association analysis conducted. The gene encoding a calcium permeable ion channel, transient receptor potential cation channel, subfamily V, member 4 (TRPV4) was identified as a candidate within the associated region and sequenced. Stably transfected HEK293 cells were used to compare wild-type and mutant TRPV4 expression, cell surface localisation and responses to activation with a synthetic agonist GSK1016709A, hypo-osmolarity, and protease-activated receptor 2 stimulation. RESULTS: The dominantly inherited folded ear and osteochondrodysplasia in Scottish fold cats is associated with a p.V342F substitution (c.1024G>T) in TRPV4. The change was not found in 648 unaffected cats. Functional analysis in HEK293 cells showed V342F mutant TRPV4 was poorly expressed at the cell surface compared to wild-type TRPV4 and as a consequence the maximum response to a synthetic agonist was reduced. Mutant TRPV4 channels had a higher basal activity and an increased response to hypotonic conditions. CONCLUSIONS: Access to a naturally-occurring TRPV4 mutation in the Scottish fold cat will allow further functional studies to identify how and why the mutations affect cartilage and bone development.

Genomic sequence of mouse COL1A1 encoding the collagen propeptides.

The nucleotide sequences of the mouse pro alpha 1(I) gene regions coding for the N- and C-propeptides is reported. The exon-intron structure was highly homologous to human COL1A1 and the deduced amino acid sequences of the N- and C-propeptides showed 67% and 91% identity with the human sequence. This gene sequence information will allow the production of specific gene mutations by site-directed mutagenesis to study the structure and function of these important propeptide domains.

Chemical cleavage method for the detection of RNA base changes: experience in the application to collagen mutations in osteogenesis imperfecta.

We discuss the definition of mutations in osteogenesis imperfecta (OI) using a chemical cleavage method for detecting mismatched bases in patient mRNA: control cDNA heteroduplexes. The method is based on the increased chemical modification of cytosines (Cs) by hydroxylamine and thymines (Ts) by osmium tetroxide when they are not paired with their complementary base. The DNA is then cleaved at the modified base with piperidine and the use of radioactively labeled DNA probes allows the position of the mismatched base to be determined by electrophoresis of the cleavage-product. The precise mutations are then determined by specific amplification and sequencing of the region containing the mismatched base. In perinatally lethal OI (OI type II) mismatches have been detected in all 17 cases studied; 12 of these have been fully characterized. In 7 of these 12 cases the mismatches were point mutations in the genes for pro alpha 1(I) or pro alpha 2(I) which resulted in glycine substitutions in the triple helical region of the protein. Sequence variation was detected in addition to the glycine substitutions in 2 cases. In 2 cases the RNA mismatch resulted from changes in the amino acid sequence of the C-propeptide domain. In the 3 remaining cases the mismatch resulted from silent nucleotide sequence variants. In the less severe forms of OI we have studied, mismatches have been detected and characterized in 8 of 12 cases. In 4 of these 8 cases the mismatch resulted from presumably neutral sequence variation and in the other 4 cases mutations have been defined.(ABSTRACT TRUNCATED AT 250 WORDS)

In vitro expression analysis of collagen biosynthesis and assembly.

While the generalised pathway of collagen biosynthesis is well understood, the specific molecular interactions that drive chain recognition and assembly and the formation of tissue-specific extracellular supramolecular structures have not been elucidated. This review focuses on the use of in vitro collagen expression systems to explore some of these fundamental questions on the molecular basis of normal and mutant collagen assembly. Three in vitro expression/assembly systems are discussed. Firstly, a simple cell-free transcription/translation system to study the initial stages of collagen chain assembly. Secondly, a novel T7-driven high level expression system, using a recombinant vaccinia virus expressing T7 RNA polymerase, in transiently transfected cells which allows appropriate postranslational modification and collagen folding. Thirdly, the more complex questions of normal and mutant collagen extracellular matrix assembly are addressed by stable transfection and expression in cells which allow the formation of a 'tissue equivalent' matrix during long-term culture.

Diagnosis and etiology of congenital muscular dystrophy.

OBJECTIVE: We aimed to determine the frequency of all known forms of congenital muscular dystrophy (CMD) in a large Australasian cohort. METHODS: We screened 101 patients with CMD with a combination of immunofluorescence, Western blotting, and DNA sequencing to identify disease-associated abnormalities in glycosylated alpha-dystroglycan, collagen VI, laminin alpha2, alpha7-integrin, and selenoprotein. RESULTS: A total of 45% of the CMD cohort were assigned to an immunofluorescent subgroup based on their abnormal staining pattern. Abnormal staining for glycosylated alpha-dystroglycan was present in 25% of patients, and approximately half of these had reduced glycosylated alpha-dystroglycan by Western blot. Sequencing of the FKRP, fukutin, POMGnT1, and POMT1 genes in all patients with abnormal alpha-dystroglycan immunofluorescence identified mutations in one patient for each of these genes and two patients had mutations in POMT2. Twelve percent of patients had abnormalities in collagen VI immunofluorescence, and we identified disease-causing COL6 mutations in eight of nine patients in whom the genes were sequenced. Laminin alpha2 deficiency accounted for only 8% of CMD. alpha7-Integrin staining was absent in 12 of 45 patients studied, and ITGA7 gene mutations were excluded in all of these patients. CONCLUSIONS: We define the distribution of different forms of congenital muscular dystrophy in a large cohort of mixed ethnicity and demonstrate the utility and limitations of current diagnostic techniques.

The type I collagen pro alpha 1(I) COOH-terminal propeptide N-linked oligosaccharide. Functional analysis by site-directed mutagenesis.

The C-propeptides of the pro alpha 1(I) and pro alpha 2(I) chains of type I collagen are each substituted with a single high-mannose N-linked oligosaccharide. Conservation of this motif among the fibrillar collagens has led to the proposal that the oligosaccharide has structural or functional importance, but a role in collagen biosynthesis has not been unambiguously defined. To examine directly the function of the pro alpha 1(I) C-propeptide N-linked oligosaccharide, the acceptor Asn residue was changed to Gln by site-directed mutagenesis. In transfected mouse Mov13 and 3T6 cells, unglycosylated mutant pro alpha 1(I) folded and assembled normally into trimeric molecules with pro alpha 2(I). In biosynthetic pulse-chase experiments mutant pro alpha 1(I) were secreted at the same rate as wild-type chains; however, following secretion, the chains were partitioned differently between the cell layer and medium, with a greater proportion of the mutant pro alpha 1(I) being released into the medium. This distribution difference was not eliminated by the inclusion of yeast mannan indicating that the high-mannose oligosaccharide itself was not binding to the matrix or the fibroblast surface after secretion. Subtle alterations in the tertiary structure of unglycosylated C-propeptides may have decreased their affinity for a cell-surface component. Further support for a small conformational change in the mutant C-propeptides came from experiments suggesting that unglycosylated pro alpha 1(I) chains were cleaved in vitro by the purified C-proteinase slightly less efficiently than wild-type chains. Mutant and normal pro alpha 1(I) were deposited with equal efficiency into the 3T6 cell accumulated matrix, thus the reduced cleavage by C-proteinase and altered distribution in the short pulse-chase experiments were not functionally significant in this in vitro extracellular matrix model system.

A mouse 3T6 fibroblast cell culture model for the study of normal and protein-engineered collagen synthesis and deposition into the extracellular matrix.

Mouse 3T6 fibroblasts deposited an organized collagenous extracellular matrix during long-term culture in the presence of ascorbic acid. The matrix produced by the cells had a similar distribution of collagen types as the mouse dermal matrix, comprising predominantly type I with smaller amounts of types III and V collagens. By day 8 of culture more than 70% of the collagen in the 3T6 matrix was involved in covalent crosslinkages and required pepsin digestion for extraction. Incorporation of NaB3H4 into reducible crosslinks and aldehydes directly demonstrated the involvement of the alpha 1 (I)CB6 and alpha 2(I)CB3.5 in crosslinks. The pattern of reducible crosslinks in the in vitro 3T6 matrix was similar to that in mouse skin suggesting a comparable fibril organization. Processing of procollagen to collagen occurred efficiently throughout the culture period and the rate of collagen production was unaltered during 15 days of culture, indicating that the development of a collagenous matrix does not directly play a role in procollagen processing or biosynthetic regulation. The existence of a preformed matrix did however, increase the efficiency with which newly synthesised collagen was incorporated into the pericellular matrix. At day 0, when there was no measurable matrix present, 29% of the collagen synthesised was deposited, while by day 15, 88% of the collagen was laid down in the matrix. The development of this 3T6 culture system, where collagen is efficiently incorporated into an organized extracellular matrix, will facilitate detailed studies on matrix organization and regulation and provide a system in which protein-engineered mutant collagens can be expressed to determine their effects on the production of a functional extracellular matrix.

A type I collagen reporter gene construct for protein engineering studies. Functional equivalence of transfected reporter COL1A1 and endogenous gene products during biosynthesis and in vitro extracellular matrix accumulation.

A type I collagen reporter gene construct, designed to facilitate detailed analysis of the consequences of introduced structural and regulatory mutations on collagen biosynthesis and participation in the extracellular matrix, was produced by site-directed mutagenesis of the mouse COL1A1 gene. The reporter construct, pWTCI-Ile822, carried a single base change which converted the codon for amino acid 822 of the triple helix from methionine to isoleucine. This change allowed the reporter protein, [Ile822]alpha 1(I), to be distinguished from the wild-type alpha 1(I), and quantified, by its altered CNBr cleavage pattern. In mouse Mov13 cells, which synthesize no endogenous pro alpha 1(I), reporter chains associated with endogenous pro alpha 2(I), formed pepsin-stable triple helices and were secreted efficiently from the cell. The thermal stability of wild-type molecules and molecules containing the reporter [Ile822]alpha 1(I) chains was identical. The biosynthetic characteristics of wild-type and reporter chains were directly compared in stably transfected 3T6 cells. These cells did not make a distinction between reporter and endogenous alpha 1(I) chains, which were secreted from the cells at the same rate and were processed and deposited into the 3T6 cell in vitro accumulated extracellular matrix with equal efficiency. These data demonstrate that the helical sequence alteration in the reporter protein is functionally neutral and that the reporter construct, pWTCI-Ile822, is a suitable vector for the analysis of the biochemical effects of site-directed mutations in the putative COL1A1 functional domains.

Procollagen folding and assembly: the role of endoplasmic reticulum enzymes and molecular chaperones.

Procollagen assembly occurs within the endoplasmic reticulum, where the C-propeptide domains of three polypeptide alpha-chains fold individually, and then interact and trimerise to initiate folding of the triple helical region. This highly complex folding and assembly pathway requires the co-ordinated action of a large number of endoplasmic reticulum-resident enzymes and molecular chaperones. Disease-causing mutations in the procollagens disturb folding and assembly and lead to prolonged interactions with molecular chaperones, retention in the endoplasmic reticulum, and intracellular degradation. This review focuses predominantly on prolyl 1-hydroxylase, an essential collagen modifying enzyme, and HSP47, a collagen-specific binding protein, and their proposed roles as molecular chaperones involved in fibrillar procollagen folding and assembly, quality control, and secretion.

Proteasomal degradation of unassembled mutant type I collagen pro-alpha1(I) chains.

We have previously shown that type I procollagen pro-alpha1(I) chains from an osteogenesis imperfecta patient (OI26) with a frameshift mutation resulting in a truncated C-propeptide, have impaired assembly, and are degraded by an endoplasmic reticulum-associated pathway (Lamandé, S. R., Chessler, S. D., Golub, S. B., Byers, P. H., Chan, D., Cole, W. G., Sillence, D. O. and Bateman, J. F. (1995) J. Biol. Chem. 270, 8642-8649). To further explore the degradation of procollagen chains with mutant C-propeptides, mouse Mov13 cells, which produce no endogenous pro-alpha1(I), were stably transfected with a pro-alpha1(I) expression construct containing a frameshift mutation that predicts the synthesis of a protein 85 residues longer than normal. Despite high levels of mutant mRNA in transfected Mov13 cells, only minute amounts of mutant pro-alpha1(I) could be detected indicating that the majority of the mutant pro-alpha1(I) chains synthesized are targeted for rapid intracellular degradation. Degradation was not prevented by brefeldin A, monensin, or NH(4)Cl, agents that interfere with intracellular transport or lysosomal function. However, mutant pro-alpha1(I) chains in both transfected Mov13 cells and OI26 cells were protected from proteolysis by specific proteasome inhibitors. Together these data demonstrate for the first time that procollagen chains containing C-propeptide mutations that impair assembly are degraded by the cytoplasmic proteasome complex, and that the previously identified endoplasmic reticulum-associated degradation of mutant pro-alpha1(I) in OI26 is mediated by proteasomes.

Substitution of arginine for glycine 664 in the collagen alpha 1(I) chain in lethal perinatal osteogenesis imperfecta. Demonstration of the peptide defect by in vitro expression of the mutant cDNA.

Structurally abnormal type I collagen was identified in tissues and cultured fibroblasts from a case of lethal perinatal osteogenesis imperfecta. Two-dimensional gel electrophoresis of the CNBr peptides demonstrated that the alpha 1(I)CB7 peptide from the alpha 1(I) chain of type I collagen existed in a normal form and a mutant form with a more basic charge distribution (Bateman, J. F., Mascara, T., Chan, D., and Cole, W. G. (1987) J. Biol. Chem. 262, 4445-4451). Sequencing of cloned alpha 1(I) cDNAs prepared using mRNA from the patient's fibroblasts demonstrated that one clone had a single base substitution of A for G which resulted in the substitution of arginine for glycine 664 within the alpha 1(I)CB7 peptide. To determine whether this mutation was responsible for the peptide map abnormality, in vitro transcription of mRNA from the mutant cDNA was performed using an SP6 vector system. The mRNA was then translated into mutant protein in a rabbit reticulocyte lysate. Peptide analysis of the protein produced from the mutant cDNA demonstrated the same altered charge distribution of the alpha 1(I)CB7 peptide as observed with tissue- and cell-derived mutant collagen peptides. This finding confirmed that the arginine for glycine 664 sequence abnormality defined in the cDNA clone was the mutation causing the observed protein peptide map defect. This mutation is consistent with the functional abnormalities of collagen observed in this case such as reduced helical stability, reduced secretion, increased degradation, and excessive posttranslational modification of lysine.

Detection and localization of base changes in RNA using a chemical cleavage method.

The detection of base changes in DNA and RNA is of central importance in genetic research. Mismatched cytosines and thymines in heteroduplex DNA molecules show increased chemical reactivity with hydroxylamine and osmium tetroxide, respectively, and the DNA can then be specifically cleaved at the modified nucleotides. We show here that mismatched cytosines and thymines can be detected and located directly in RNA: DNA heteroduplex molecules. In order to detect guanosine and adenosine base changes the complementary cDNA strand must be analyzed. In addition, the sensitivity of the technique can be increased by employing the polymerase chain reaction. To test the fidelity of this method a number of known or predicted mutations were analyzed. These include single point mutations in the human collagen alpha 1(I) and rat phenylalanine hydroxylase mRNA, two engineered point mutations in a mouse collagen alpha 1(I) mRNA, and a deletion in a human collagen alpha 2(I) mRNA. All known base changes were detected and correctly localized. In addition, the predicted base changes were confirmed.

Regulation of procollagen synthesis and processing during ascorbate-induced extracellular matrix accumulation in vitro.

Procollagen biosynthesis and matrix deposition were studied in long-term human skin fibroblast cultures exposed to ascorbic acid. Ascorbic acid specifically stimulated types I and III collagen synthesis, reaching a maximum at day 2 and maintaining a specific high rate of production until day 10 of ascorbate exposure, after which collagen production declined. The increased level of collagen synthesis after different exposure times could also be achieved by only brief treatment (10 h) of parallel scorbutic (ascorbic-acid-deficient) cultures with ascorbic acid. This brief exposure did not result in increased collagen mRNA, thus demonstrating that the ascorbate-induced increase in collagen synthesis at all stages of ascorbic acid exposure was due to post-transcriptional mechanisms, most likely a rapid increase in type 1 collagen mRNA translational efficiency. This mechanism, rather than the transcriptional activation, was the primary response and is adequate to explain the ascorbate-induced increase in collagen synthesis. These data also demonstrate that the presence of a collagenous extracellular matrix was not involved in this collagen biosynthetic regulation. During long-term exposure (18 days) to ascorbic acid, a substantial cross-linked collagenous matrix formed, following an approximately sigmoidal time course. The most rapid matrix deposition occurred during the later days of exposure when the rate of collagen synthesis was decreasing, suggesting that the presence of a pre-existing matrix is important for further collagen accumulation. Procollagen was also efficiently processed to collagen during this phase, demonstrating that efficient procollagen processing is an important regulatory event in collagen matrix deposition.

Clinical and radiographic features of a family with autosomal dominant amelogenesis imperfecta with taurodontism.

This paper describes the clinical features of a family of four generations with autosomal dominant amelogenesis imperfecta with taurodontism (ADAIT). Considerable variation in phenotype was seen, both between individuals and within the dentition of some individuals. Many of the adults had received extensive dental restorative work. These findings re-enforce previous observations of variable phenotype in this and other forms of the condition and add to the argument for a revision of methods of classification. This history of this large family draws further attention to the restorative demands of this group of dental anomalies and, by their generous co-operation, will prove an invaluable help in the investigation by molecular genetic techniques of this disfiguring condition.

Characterization of point mutations in the collagen COL1A1 and COL1A2 genes causing lethal perinatal osteogenesis imperfecta.

Type I collagen mutations in a group of patients with lethal perinatal osteogenesis imperfecta were identified in fibroblast RNA by a new method which can detect, by chemical modification and cleavage, single mismatched bases in heteroduplexes formed between mRNA and normal cDNA probes. Control cDNA probes spanning the area of the pro-alpha 1(I) and pro-alpha 2(I) chains likely to contain the mutations were radioactively labeled and used to form heteroduplexes with total patient RNA. Treatment of these heteroduplexes with hydroxylamine followed by cleavage of the cDNA strand at reactive bases by piperidine identified mismatches in the pro-alpha 1(I) cDNA in four patients. In the fifth patient a mismatch was detected in the pro-alpha 2(I) cDNA. To characterize these mutations the regions containing the mismatches were amplified by the polymerase chain reaction, cloned, and sequenced. All were heterozygous single base mutations which led to the substitution of glycine residues in the helical region of the pro-alpha-chains. The substitutions were pro-alpha 1(I) Gly973 and Gly1006 to Val, Gly928 to Ala, Gly976 to Arg, and pro-alpha 2(I) Gly865 to Ser. These mutations emphasize the importance of the Gly-X-Y repeating amino acid sequence for normal collagen helix formation and function in the extracellular matrix.

A frameshift mutation results in a truncated nonfunctional carboxyl-terminal pro alpha 1(I) propeptide of type I collagen in osteogenesis imperfecta.

A codon frameshift mutation caused by a single base (U) insertion after base pair 4088 of prepro alpha 1(I) mRNA of type I procollagen was identified in a baby with lethal perinatal osteogenesis imperfecta. The mutation was identified in fibroblast RNA by a new method that allows the direct detection of mismatched bases by chemical modification and cleavage in heteroduplexes formed between mRNA and control cDNA probes. The region of mismatches was specifically amplified by the polymerase chain reaction and sequenced. The heterozygous mutation in the amplified cDNA most likely resulted from a T insertion in exon 49 of COL1A1. The frameshift resulted in a truncated pro alpha 1(I) carboxyl-terminal propeptide in which the amino acid sequence was abnormal from Val1146 to the carboxyl terminus. The propeptide lacked Asn1187, which normally carries an N-linked oligosaccharide unit, and was more basic than the normal propeptide. The distribution of cysteines was altered and the mutant propeptide was unable to form normal interchain disulfide bonds. Some of the mutant pro alpha 1(I)' chains were incorporated into type I procollagen molecules but resulted in abnormal helix formation with over-hydroxylation of lysine residues, increased degradation, and poor secretion. Only normal type I collagen was incorporated into the extracellular matrix in vivo resulting in a tissue type I collagen content approximately 20% of that of control (Bateman, J. F., Chan, D., Mascara, T., Rogers, J. G., and Cole, W. G. (1986) Biochem. J. 240, 699-708).

The N-terminal N5 subdomain of the alpha 3(VI) chain is important for collagen VI microfibril formation.

Collagen VI assembly is unique within the collagen superfamily in that the alpha 1(VI), alpha 2(VI), and alpha 3(VI) chains associate intracellularly to form triple helical monomers, and then dimers and tetramers, which are secreted from the cell. Secreted tetramers associate end-to-end to form the distinctive extracellular microfibrils that are found in virtually all connective tissues. Although the precise protein interactions involved in this process are unknown, the N-terminal globular regions, which are composed of multiple copies of von Willebrand factor type A-like domains, are likely to play a critical role in microfibril formation, because they are exposed at both ends of the tetramers. To explore the role of these subdomains in collagen VI intracellular and extracellular assembly, alpha 3(VI) cDNA expression constructs with sequential N-terminal deletions were stably transfected into SaOS-2 cells, producing cell lines that express alpha 3(VI) chains with N-terminal globular domains containing modules N9-N1, N6-N1, N5-N1, N4-N1, N3-N1, or N1, as well as the complete triple helix and C-terminal globular domain (C1-C5). All of these transfected alpha 3(VI) chains were able to associate with endogenous alpha 1(VI) and alpha 2(VI) to form collagen VI monomers, dimers, and tetramers, which were secreted. Importantly, cells that expressed alpha 3(VI) chains containing the N5 subdomain, alpha 3(VI) N9-C5, N6-C5, and N5-C5, formed microfibrils and deposited a collagen VI matrix. In contrast, cells that expressed the shorter alpha 3(VI) chains, N4-C5, N3-C5, and N1-C5, were severely compromised in their ability to form end-to-end tetramer assemblies and failed to deposit a collagen VI matrix. These data demonstrate that the alpha 3(VI) N5 module is critical for microfibril formation, thus identifying a functional role for a specific type A subdomain in collagen VI assembly.

The role of the alpha3(VI) chain in collagen VI assembly. Expression of an alpha3(VI) chain lacking N-terminal modules N10-N7 restores collagen VI assembly, secretion, and matrix deposition in an alpha3(VI)-deficient cell line.

Collagen VI is a microfibrillar protein found in the extracellular matrix of virtually all connective tissues. Three genetically distinct subunits, the alpha1(VI), alpha2(VI), and alpha3(VI) chains, associate intracellularly to form triple-helical monomers, which then assemble into disulfide-bonded dimers and tetramers before secretion. Although sequence considerations suggest that collagen VI monomers composed of all three chains are the most stable isoform, the precise chain composition of collagen VI remains controversial and alternative assemblies containing only alpha1(VI) and alpha2(VI) chains have also been proposed. To address this question directly and study the role of the alpha3(VI) chain in assembly, we have characterized collagen VI biosynthesis and in vitro matrix formation by a human osteosarcoma cell line (SaOS-2) that is deficient in alpha3(VI) production. Northern analysis showed an abundance of alpha1(VI) and alpha2(VI) mRNAs, but no detectable alpha3(VI) mRNA was apparent in SaOS-2 cells. By day 30 of culture, however, small amounts of alpha3(VI) mRNA were detected, although the level of expression was still much less than alpha1(VI) and alpha2(VI). Collagen VI protein was not detected in SaOS-2 medium or cell layer samples until day 30 of culture, demonstrating that despite the abundant synthesis of alpha1(VI) and alpha2(VI), no stable collagen VI protein was produced without expression of alpha3(VI). The alpha1(VI) and alpha2(VI) chains produced in the absence of alpha3(VI) were non-helical and were largely retained intracellularly and degraded. The critical role of the alpha3(VI) chain in collagen VI assembly was directly demonstrated after stable transfection of SaOS-2 cells with an alpha3(VI) cDNA expression construct that lacked 4 of the 10 N-terminal type A subdomains. The transfected alpha3(VI) N6-C5 chains associated with endogenous alpha1(VI) and alpha2(VI) and formed collagen VI dimers and tetramers, which were secreted and deposited into an extensive network in the extracellular matrix. These data demonstrated that alpha3(VI) is essential for the formation of stable collagen VI molecules and subdomains N10-N7 are not required for molecular assembly.