Use of the polymerase chain reaction to clone and sequence a cDNA encoding the bovine alpha 3 chain of type IV collagen

A novel type IV collagen, alpha 3(IV), has previously been isolated from a collagenase digest of bovine and human glomerular and lens basement membranes. The cloning and sequencing of a cDNA encoding the alpha 3(IV) chain is described here. Using the polymerase chain reaction, with primers derived from the known 27-residue bovine alpha 3(IV) amino acid sequence, a 68-base pair bovine genomic fragment (KEM68) which encodes the known peptide sequence, was synthesized. KEM68 was then used to screen a bovine lens cDNA library and a 1.5-kilobase partial cDNA clone obtained, encoding 471 residues of the bovine alpha 3(IV) chain: 238 residues from the triple helical collagenous domain and all 233 residues of the noncollagenous domain. The collagenous repeat sequence has three interruptions, coinciding with those in the alpha 1(IV) chain. The noncollagenous domain has 12 cysteine residues in identical positions to those of other type IV collagens and 71, 61, and 70% overall similarity with the human alpha 1(IV), alpha 2(IV), and alpha 5(IV) chains. The noncollagenous domain of alpha 3(IV) is of particular interest as it appears to be the component of glomerular basement membrane that reacts maximally with the Goodpasture antibody. Furthermore, such antigenicity is absent from collagenase digests of the glomerular basement membrane of some patients with Alport syndrome. The alpha 3(IV) cDNA clone described here now permits study of the molecular pathology of COL4A3 in Alport syndrome.     

The complete primary structure of the alpha 2 chain of human type IV collagen and comparison with the alpha 1(IV) chain

The complete primary structure of the human type IV collagen alpha 2(IV) chain has been determined by nucleotide sequencing of cDNA clones. The overlapping cDNA clones cover 6,257 base pairs with a 5'-untranslated region of 283 base pairs, the 5,136-base pair open reading frame, and the 3'-untranslated region of 838 base pairs. The predicted amino acid sequence demonstrates that the complete translation product consists of 1,712 residues corresponding in molecular weight to 167,560. The translated polypeptide has a signal peptide of 36 amino acids, an amino-terminal noncollagenous part of 21 residues, a 1,428-residue collagenous domain with 23 interruptions, and a carboxyl-terminal noncollagenous (NC) domain of 227 residues. The calculated molecular mass of the mature human alpha 2(IV) chain is 163,774 Da.     

Complete primary structure of the alpha 1-chain of human basement membrane (type IV) collagen

We have determined the primary structure of the alpha 1(IV)-chain of human type IV collagen by nucleotide sequencing of overlapping cDNA clones that were isolated from a human placental cDNA library. The present data provide the sequence of 295 amino acids not previously determined. Altogether, the alpha 1(IV)-chain contains 1642 amino acids and has a molecular mass of 157625 Da. There are 1413 residues in the collagenous domain and 229 amino acids in the carboxy-terminal globular domain. The human alpha 1(IV)-chain contains a total of 21 interruptions in the collagenous Gly-X-Y repeat sequence. These interruptions vary in length between two and eleven residues. The alpha 1(IV)-chain contains four cysteine residues in the triple-helical domain, four cysteines in the 15-residue long noncollagenous sequence at the amino-terminus and 12 cysteines in the carboxy-terminal NC-domain.     

Substitution of arginine for glycine 325 in the collagen alpha 5 (IV) chain associated with X-linked Alport syndrome: characterization of the mutation by direct sequencing of PCR-amplified lymphoblast cDNA fragments.

A large kindred with adult-type X-linked Alport syndrome was studied with regard to a defect in the recently described COL4A5 collagen gene. Southern blot analysis with COL4A5 cDNA probes showed loss of a MspI restriction site. Direct sequencing of cDNA amplified from lymphoblast mRNA demonstrated a single-base substitution converting a glycine codon to arginine at position 325 in the alpha 5 chain of type IV collagen. The triple-helical collagenous domain of alpha 5(IV), characterized by a Gly-X-Y repeat sequence, is interrupted 22 times by noncollagenous sequences. The mutation creates an additional interruption in the Gly-X-Y repeat motif, between interruptions 4 and 5. It is interesting that such glycine substitutions inside the COL1A1 or COL1A2 genes have been associated with many cases of osteogenesis imperfecta. This gly325-to-arg substitution presumably alters the triple-helix formation, and, in turn, modifies the ultrastructural and functional characteristics of the type IV collagen network inside the glomerular basement membrane.     

Head-to-head arrangement of murine type IV collagen genes

The genes for the two dissimilar subunits of type IV collagen are organized in a head-to-head manner with their translation initiation codons within 874 base pairs. Murine genomic clones which contain portions of both genes have been isolated and characterized. These clones contain the first exon of the alpha 1(IV) chain and the first 3 exons for the alpha 2(IV) chain within a 1.7-kilobase HindIII fragment. The intergenic region appears not to resemble previously described bidirectional promoters. The HindIII fragment is present as a single copy in the mouse genome ruling out the presence of one of these gene fragments as a pseudogene. These findings agree with linkage studies of these two genes and differ from the known organization of human fibrillar collagen genes which have been found to be dispersed within the human genome.     

Isolation and characterization of two types of cDNA for mitochondrial adenylate kinase and their expression in Escherichia coli

Two cDNA clones for mitochondrial adenylate kinase were isolated from a cDNA library of bovine liver poly(A)+ RNA by using synthetic oligodeoxynucleotides as probes. The clone containing a 0.9-kilobase insert had the reading frame for a 241-residue protein (AK2A), while the other clone containing a 1.6-kilobase insert had the frame for a 234-residue protein (AK2B). Nucleotide sequences of these two clones were the same in the 5' portion up to the coding sequence for the 233rd residue, but different in the remaining 3' portions. The reported amino acid sequence of mitochondrial adenylate kinase from bovine heart corresponded to AK2A. Neither AK2A nor AK2B had a cleavable NH2-terminal presequence as that found in other imported mitochondrial proteins. RNA blot analysis of poly(A)+ RNAs from bovine liver and heart revealed three species of mRNA with approximate sizes of 0.9, 1.4, and 1.7 kilobases. The 1.7- and 1.4-kilobase species were specific for AK2B, whereas the 0.9-kilobase species was specific for AK2A. In the liver, the 1.7-kilobase mRNA was more abundant, whereas in the heart the 0.9-kilobase mRNA was predominant. The 1.4-kilobase mRNA was present only in the heart. The AK2A- and AK2B-coding sequences were expressed in Escherichia coli cells under the control of trc promoter. Both the products reverted the temperature-sensitive phenotype of the adenylate kinase mutant of E. coli.     

Evidence for a type-IV-related collagen in Drosophila melanogaster. Evolutionary constancy of the carboxyl-terminal noncollagenous domain

Type IV collagen, a major structural component of basement membrane, has been characterized only in vertebrates. It is unique among the collagenous proteins in that it forms specific lattice networks by end-to-end interactions. In particular, in mammals the C-terminal noncollagenous domain (NCl) of collagen IV was shown to be one of the major cross-linking sites in the network assembly. Here, we give the first direct evidence of type-IV-related collagen in invertebrates by sequence analysis of cDNA and genomic DNA clones for the 3'-end of a previously characterized Drosophila collagen gene. The data describe the C-terminal 190 amino acid residues of the triple helix and the entire noncollagenous domain (231 amino acids) of the chain encoded for by this gene. Comparison with data reported for human and mouse alpha 1(IV) chains reveals that triple-helix regions are quite different, while NC1 structures are very similar. This suggests different constraints on triple-helix and NC1 domains during evolution. Present data support the assumption that the NC1 structure originated from duplication of an ancestral sequence; the extent of both interspecies and intramolecular homologies suggests the maintenance in vertebrates and invertebrates of an ancestral specific function.     

Nucleotide sequence of the luxB gene of Vibrio harveyi and the complete amino acid sequence of the beta subunit of bacterial luciferase

The nucleotide sequence of the 1.30-kilobase EcoRI/BglII fragment from Vibrio harveyi carrying the majority of the luciferase beta subunit coding region (luxB gene) has been determined. The EcoRI/BglII fragment was derived from a 4.0-kilobase HindIII fragment carrying both luxA and luxB which was detected in a genomic clone bank based on the expression of bioluminescence from colonies of Escherichia coli carrying V. harveyi HindIII fragments in plasmid pBR322 (Baldwin, T. O., Berends, T., Bunch, T. A., Holzman, T. F., Rausch, S. K., Shamansky, L., Treat, M. L., and Ziegler, M. M. (1984) Biochemistry 23, 3663-3667). The entire alpha subunit coding sequence (luxA gene) and the amino-terminal 13 codons of the beta subunit sequence (luxB gene) were contained on a 1.85-kilobase EcoRI fragment, the sequence of which has been reported (Cohn, D. H., Mileham, A. J., Simon, M. I., Nealson, K. H., Rausch, S. K., Bonam, D., and Baldwin, T. O. (1985) J. Biol. Chem. 260, 6139-6146). The beta subunit coding sequence was found to terminate 972 bases past the start of the luxB coding sequence. The beta subunit had a calculated molecular weight of 36,349 and comprised a total of 324 amino acid residues; the alpha beta dimer had a molecular weight (alpha + beta) of 76,457. There were 27 base pairs separating the stop codon of the beta subunit structural gene and a 340-base open reading frame extending to (and beyond) the distal BglII site. Approximately two-thirds of the beta subunit was sequenced by protein chemical techniques. The amino acid sequence predicted from the DNA sequence, with few exceptions, confirmed the chemically determined sequence, and the measured amino acid composition was in excellent agreement with the composition implied from the DNA sequence.     

Localization of the human procollagen alpha 1(IV) gene to chromosome 13q34 by in situ hybridization.

Type IV (alpha 1 and alpha 2 chains) appears to be the only procollagen present in basement membranes. The structure of this protein is highly divergent from the interstitial and type V procollagens as exemplified by the interruptions in the Gly-X-Y region and unprocessed amino and carboxyl noncollagenous peptides. To expand our knowledge concerning the primary sequence of type IV and to investigate the factors influencing its unique distribution, we recently isolated cDNA clones coding for part of the human alpha 1(IV) chain. To determine if the alpha 1(IV) gene was cytologically linked to other procollagen genes that have been assigned to autosomes 17, 12, 7, and 2, overlapping clones covering 2.6 kilobases (kb) of the alpha 1(IV) mRNA were used together for in situ hybridization to human metaphase chromosomes. Here, we show precise localization of alpha 1(IV) at the telomere of 13q, thereby defining a fifth chromosome that contains members of this large and surprisingly dispersed multigene family.     

Cloning and nucleotide sequence analysis of the dog insulin gene. Coded amino acid sequence of canine preproinsulin predicts an additional C-peptide fragment

A 4.0-kilobase HindIII/EcoRI-cleaved dog genomic DNA fragment was shown to contain the dog insulin gene by restriction mapping using a human insulin cDNA probe. This fragment was subsequently cloned in a lambda vector, and the nucleotide sequence of the dog insulin gene was determined. As in several other species, the insulin gene of the dog is interrupted by two intervening sequences, one of 151 base pairs located in the 5' untranslated region and the other of 264 base pairs occurring within the codon of the 7th amino acid of the C-peptide. Translation of the nucleotide sequence in one frame revealed the primary structure of canine preproinsulin. An interesting feature of the coded amino acid sequence is that it predicts a C-peptide of 31 amino acids, 8 residues longer than that reported by Peterson et al. (Peterson, J. D., Nehrlich, S., Oyer, P. E., and Steiner, D. F. (1973) J. Biol. Chem. 247, 4866-4871). The additional octapeptide sequence, Glu-Val-Glu-Asp-Leu-Gln-Val-Arg, is located NH2-terminal to the 23-residue C-peptide sequence described in the earlier report. Its coding sequence is interrupted by the second intervening sequence. The arginine at position 8 suggests that a trypsin-like cleavage may separate the NH2-terminal octapeptide from the remainder of the C-peptide during the post-translational processing of dog proinsulin in the pancreas. The revised C-peptide sequence suggests that the proinsulin C-peptide is more highly conserved in length and overall sequence than was previously supposed.     

Mendel's stem length gene (Le) encodes a gibberellin 3 beta-hydroxylase.

We describe the isolation of the Le gene of pea, which controls internode elongation and originally was described by Mendel. Heterologous screening of a pea cDNA library yielded a partial clone that was 61% identical to coding regions of the putative Arabidopsis gibberellin 3 beta-hydroxylase gene, GA4. DNA gel blot analysis with this cDNA revealed a HindIII restriction fragment length polymorphism between pea isolines differing at Mendel's Le locus. Genomic clones of the GA4-related gene were isolated from the Le and le isolines. Polymerase chain reaction combined with restriction fragment length polymorphism analysis were used to show that the gene mapped to the Le locus. A cDNA containing a complete open reading frame of the pea GA4-related gene was amplified by polymerase chain reaction from each isoline. Recombinant expression in Escherichia coli demonstrated that the product of the Le cDNA was a gibberellin 3 beta-hydroxylase that is able to convert GA20 to the bioactive GA1. Substantially reduced levels of gibberellin 3 beta-hydroxylase activity were measured, after expression of the le cDNA, by using identical methods. This reduced activity was associated with an alanine-to-threonine substitution in the predicted amino acid sequence of the enzyme near its proposed active site.     

cDNA clones completing the nucleotide and derived amino acid sequence of the alpha 1 chain of basement membrane (type IV) collagen from mouse

Six cDNA clones add 3549 nucleotides to the DNA sequence of the alpha 1 chain of basement membrane (type IV) collagen. Thus the complete nucleotide and derived amino acid sequence of the alpha 1 type IV collagen with 5007 nucleotides coding for 1669 amino acids with a calculated Mr of 160,827 is known. The six cDNA clones cover the putative N-terminal signal peptide, the 7 S region and two thirds of the helical region extending into the previously published murine nucleotide sequence [(1986) Gene 43, 301]. The protein sequence for 289 amino acids of the helical region adjacent to the 7 S region has not previously been published for any species.     

Nucleotide sequence coding for the human type IV collagen alpha 2 chain cDNA reveals extensive homology with the NC-1 domain of alpha 1 (IV) but not with the collagenous domain or 3'-untranslated region

We have isolated two overlapping cDNA clones that provide the complete nucleotide sequence coding for the NC-1 domain and 3'-untranslated region of the alpha 2 chain of human type IV collagen as well as a sequence encoding 232 residues of the collagenous domain. An extensive homology was observed between the sequences of the NC-1 domain of the alpha 1(IV) and alpha 2(IV) chains, but considerably less between the sequences encoding collagenous and 3'-untranslated regions. There were four interruptions in the collagenous sequence studied whereas the comparable region of the alpha 1(IV) chain had only two. A potential oligosaccharide attachment site was found in a 6-residue long interruption of the collagenous domain but none in the NC-1 domain.     

A novel ochre mutation in the beta-thalassemia gene of a Thai. Identification by direct cloning of the entire beta-globin gene amplified using polymerase chain reactions

The beta-globin genes from a Thai patient compound heterozygous for beta-thalassemia and HbE disease were investigated. The 3.0-kilobase fragment containing the entire beta-globin gene was amplified by polymerase chain reaction, using Taq DNA polymerase followed by direct cloning of the amplified product into plasmid DNA. Sequence analysis of the thalassemia gene revealed only one base change, a C-A transversion within codon for an amino acid 35. This new mutation creates a premature terminator, TAA, an ochre codon, and results in a beta 0-thalassemia phenotype. The same result was obtained when this mutation was analyzed using a conventional cloning technique, direct sequencing of the amplified product, and hybridization with allele-specific oligonucleotide probes. No misincorporation was detected in the sequence analysis of the 3.0-kilobase insert of five clones of the amplified products obtained from genomic DNA of a normal individual. This approach is a rapid and accurate method for molecular cloning of the beta-globin gene and also other genes, the partial nucleotide sequences of which are known.     

Identification and procaryotic expression of the gene coding for the highly immunogenic 28-kilodalton structural phosphoprotein (pp28) of human cytomegalovirus.

Human cytomegalovirus contains a structural polypeptide that is 28 kilodaltons in apparent molecular size and is reactive in Western blot (immunoblot) analysis with the majority of human sera. The gene coding for this polypeptide was mapped on the genome of human cytomegalovirus strain AD169. A monoclonal antibody specific for the 28-kilodalton polypeptide was used to screen a cDNA library constructed from poly(A)+ RNA of human cytomegalovirus-infected cells in the procaryotic expression vector lambda gt11. Hybridization of cDNA with cosmid and plasmid clones mapped the gene to the HindIII R fragment. The gene was transcribed into a late 1.3-kilobase RNA. The nucleotide sequence of the coding region was determined. Parts of the 28-kilodalton polypeptide were expressed in Escherichia coli as hybrid proteins fused to beta-galactosidase. In Western blots these proteins were recognized by human sera. Antibodies raised against the hybrid proteins reacted specifically with the viral antigen in immunoprecipitations and Western blots. In vitro phosphorylation of HCMV virions and immunoprecipitation showed that the 28-kilodalton polypeptide was phosphorylated.     

Molecular cloning and characterization of cDNA for human myeloperoxidase

Partial amino acid sequence of human myeloperoxidase was determined, and a 41-base oligonucleotide containing deoxyinosines at four positions was chemically synthesized. By using the oligonucleotide as a probe, cDNA clones for human myeloperoxidase were isolated from a cDNA library constructed with mRNA from human promyelocytic leukemia HL-60 cells. One of the clones containing a 2.6-kilobase insert was subjected to nucleotide sequence analysis. The sequence was found to contain an open reading frame, 2,235 nucleotides coding for a protein of 745 amino acids with a calculated Mr of 83,868. The heavy chain of myeloperoxidase, consisting of 467 amino acids, was located on the COOH terminus half of the protein. The RNA specified by the cDNA was prepared using SP6 RNA polymerase and translated in rabbit reticulocyte lysates, and the product was identified as human myeloperoxidase by immunoprecipitation with rabbit anti-human myeloperoxidase antibody. By Northern hybridization analysis of RNA from leukemic cells, it was shown that myeloperoxidase mRNA is abundantly expressed in human promyelocytic HL-60 and mouse myeloid leukemia NFS-60 cells. Furthermore, the results of Southern hybridization analysis of human genomic DNA suggest that there are one or two genes for myeloperoxidase in the human haploid genome.