Characterization of a complete immunoglobulin heavy-chain variable region germ-line gene of rainbow trout.

A germ-line heavy-chain variable region (VH) gene (RTVH431) has been isolated from a rainbow trout (Salmo gairdneri) and characterized by complete nucleotide sequencing. It is characteristic of VH, as shown by the conserved octamer and TATA motif in the 5' region, the heptamernonamer recombination signal sequence in the 3' region, and the 18-amino-acid-long hydrophobic leader interrupted by an intron. The 98-amino-acid-long VH coding region has 50-70% nucleotide sequence homology and 40-60% amino acid sequence homology with VHS of various vertebrate species. We have also found unique or species-specific amino acid residues in the VHS of rainbow trout, amphibia (Xenopus), reptile (Caiman), and shark (Heterodontus) in our sequence analyses. The RTVH431 has an unusual amino acid in the conserved 34th position in complementarity-determining region 1 of VH. Southern hybridization results suggest the presence of a large gene family related to RTVH431 in the trout genome. The complex evolution of antibody V genes is discussed.     

Evidence for somatic rearrangement of immunoglobulin genes coding for variable and constant regions.

A high-molecular-weight DNA from Balb/c mouse early embryo or from MOPC 321 plasmacytoma (a k-chain producer) was digested to completion with Bacillus amyloliquefaciens strain H restriction enzyme (BamH I). The resulting DNA fragments were fractionated according to size in preparative agarose gel electrophoresis. DNA fragments carrying gene sequences coding for the variable or constant region of k chains were detected by hybridization with purified, 125I-labeled, whole MOPC 321 K MRNA and with its 3'-end half. The pattern of hybridization was completely different in the genomes of embryo cells and of the plasmacytoma. The pattern of embryo DNA showed two components, one of which (molecular weight=6.0 million) hybridized with C-gene sequences and the other (molecular weight=3.9 million) with V-gene sequences. The pattern of the tumor DNA showed a single component that hybridized with both V-gene and C-gene sequences and that is smaller (molecular weight=2.4 million) than either of the components in embryo DNA. The results were interpreted to mean that the Vk and Ck genes, which are some distance away from each other in the embryo cells, are joined to form a contiguous polynucleotide stretch during differentiation of lymphocytes. Such joining occurs in both of the homologous chromosomes. Relevance of these findings with respect to models for V-C gene joining, activation of a specific V k gene, and allelic exclusion in immunoglobulin gene loci is discussed.     

The 48-base-long primordial building block of immunoglobulin light-chain variable regions is complementary to the primordial building block of heavy-chain variable regions.

The ancestral gene for immunoglobulin light-chain variable regions (Ig VLs) of the kappa as well as the lambda class apparently arose from about 12 tandem repeats of the 48-base-long primordial building block sequence TCT-TGC-GCA-GTA-AGT-CCA-CTC-CAG-GTC-ATA-TCC-AGT-CAG-GCT-GCT-GAA. Even today, amino acid residues 67 to 82 of each Ig V kappa L are still specified by a direct descendant in toto of the above-noted primordial building block, whereas amino acid residues 14 to 25 are invariably specified by its truncated copy. The Ig VL primordial building block presently identified is 100% complementary to the Ig VH (heavy-chain variable region) primordial building block previously identified. In the recognition of specific antigenic determinants by antibodies, Ig VL and Ig VH of light-chain--heavy-chain dimers have to complement each other. It is perhaps fitting that the primordial building blocks of the two are represented by the complementary strands of the same 48-base-pair-long DNA sequence.     

Mouse immunoglobulin coding sequences for the heavy-chain variable region arose as repeats of the two short building blocks.

The coding sequence for the 97-amino-acid-residue-long immunoglobulin heavy-chain variable (VH) regions of the mouse apparently arose as repeats of the two short building blocks. Three of the recognizable copies of the one 21-base-long prototype sequence A-C-T-G-G-A-T-A-T-G-A-C-C-T-G-G-A-G-T-G-G are invariably found to occupy the fixed positions within the 5' half of each VH coding sequence. Interestingly, the first and third copies specify the relatively invariant regions represented by the 7th to 13th and 41st to 47th amino acid residues (the first and second framework regions), whereas the second copy specifies the first hypervariable region (31st to 35th amino acid residues). These copies maintain at least 57.2% (12 out of 21) base sequence homology to the above-noted prototype building block. Base sequences of the other 14- to 15-base-long prototype building block differ from each other by as much as 60% between individual VHS. Yet one of its copies invariably occupies the terminal region of each VH coding sequence, thus specifying the very invariant third framework region. Other copies occupy unfixed positions in the VH and its attendant hydrophobic leader coding sequence as well as in adjacent noncoding sequences. The homology thus revealed between the VH coding sequence and its adjacent noncoding sequences suggests their concordant evolution.     

Fine specificity, idiotypy, and nature of cloned heavy-chain variable region genes of murine monoclonal rheumatoid factor antibodies.

We investigated the immunochemical and molecular characteristics of murine monoclonal rheumatoid factors. Study of the fine specificity of 20 monoclonal rheumatoid factor antibodies shows a wide degree of heterogeneity. However, many express an interstrain cross-reactive idiotype. We show that our rheumatoid factors utilize a restricted set of the heavy-chain variable region (VH) repertoire representing the more 3' VH families. Preferential expression of 3' VH families is known to occur early in development. We report the nucleotide sequence of two cloned rheumatoid factor VH genes, Y19-10 (VH J558) and 129-48 (VH 7183) in which no major differences are observed between VH genes encoding the heavy chain of autoantibodies and antibodies against foreign antigens.     

DNA sequences of the joining regions of mouse lambda light chain immunoglobulin genes.

The joining (J) segments of mouse immunoglobulin lambda light chain genes, lambda 2, lambda 3, and a presumptive lambda 4, were cloned, and their sequences were determined and compared with that of lambda 1. Although all the lambda J segments share sequence homology, the J1 and J4 segments and the J2 and J3 segments, respectively, are more homologous. These sequence data, together with the fact that present day lambda genes occur in two clusters, 5' J3C3J1C1 3' and 5' J2C2J4C4 3', further substantiates a probable evolutionary duplication unit, JIICIIJICI, with II the precursor of lambda 3 and lambda 2 and I the precursor of lambda I and lambda 4. From the J4 sequence, we conclude that the lambda 4 gene is most likely nonfunctional (i.e., a pseudogene). The signal nonamer sequence 5' to J3 differs from that of J1 in two consecutive base pairs. This difference could account in part for the lower level of expression of lambda 3 as compared with lambda 1 in mouse serum.     

Identification of the 48-base-long primordial building block sequence of mouse immunoglobulin variable region genes.

Mouse immunoglobulin heavy-chain variable region (Ig VH) genes apparently arose from the approximately 600-base-pair-long (approximately 12 tandem repeats of the 48-base-pair-long primordial building block sequence TTC-AGC-AGC-CTG-ACT-GGA-TAT-GAC-CTG-GAG-TGG-ACT-TAC-TGC-GCA-AGA) that in the original reading frame specified the amino acid sequence Phe-Ser-Ser-Leu-Thr-Gly-Tyr-Asp-Leu-Glu-Trp-Thr-Tyr-Cys-Ala-Arg. The previously identified, shorter prototype building blocks merely represented particular portions of the above primordial sequence. Even today, the direct descendant in toto of this primordial sequence specifies the last one-sixth of each VH coding sequence: the 83rd to 98th amino acid residues. Furthermore, its four truncated derivatives specify the 4th to 14th, 17th to 23rd, 29th to 37th, and 38th to 48th amino acid residues. Accordingly, all three relatively invariant--therefore, conserved--framework regions (FW-1, FW-2, and FW-3) of VHs are specified by recognizable--therefore, conserved--descendants of the primordial sequence.     

Nucleotide sequence of immunoglobulin heavy chain joining segments between translocated VH and mu constant regions genes.

To investigate the mechanism of recombination of immunoglobulin heavy chain variable and constant region genes, we have determined the nucleotide sequence of a large portion of the recombination region between an active C mu gene and its associated VH gene, isolated from an IgM-secreting mouse plasmacytoma, HPC76. By comparison with the sequence of the mu mRNA, we determined the exact boundaries of the intervening sequence between the VH76 and C mu genes. The rearranged VH76 gene encodes up to amino acid 116 without interruption, the 3' 39 nucleotides (the JH76 region) being derived from an embryonic JH segment (JH315) whose sequence was recently determined [Early, P., Huang, H., Davis, M., Calame, K. & Hood, L. (1980) Cell 195, 981-992]. The active JH76 does not use the first two codons of the embryonic JH315 from which it is derived. This indicates that V-J recombination is important in generating diversity within the third hypervariable region of heavy chains. We have identified another JH segment (JHA4), located 336 nucleotides 3' to the rearranged JH76 segment. This JH segment is expressed in the heavy chains of anti-levan myeloma proteins, which have truncated third hypervariable regions. We propose that the nucleotide sequence 5' to JHA4 is important for generating V region genes with shortened third hypervariable regions.     

Complete nucleotide sequence of an immunoglobulin heavy-chain gene and analysis of immunoglobulin gene organization in a primitive teleost species.

The immunoglobulin heavy-chain variable region (VH) locus in a phylogenetically primitive teleost (Elops saurus) has been characterized by a strategy that relied initially on cross-hybridization between genomic VH segments and a murine VH probe. Using a homologous (Elops) VH probe and DNA sequencing, this gene family has been shown to be complex and to contain overt pseudogenes. A homologous probe also has been used to isolate a full copy length cDNA containing constant (CH) as well as joining (JH) and VH regions. Genomic analyses using CH-, JH-, and VH-specific probes have demonstrated the presence of only a single hybridizing CH and several JH elements. JH-CH linkage is less than or equal to 3.6 kilobases (kb) and VH-CH linkage is less than or equal to 100 kb, as estimated by field-inversion gel electrophoresis. An additional VH family sharing less than 50% nucleotide identity with the prototype Elops VH sequence is described. Taken together, these results suggest that the immunoglobulin VH locus in a comparatively primitive teleost resembles the VH locus in mammals, but not that found in the more phylogenetically distant elasmobranchs. The evolutionary radiations of cartilaginous and bony fishes are associated with a dramatic change in the organization and, presumably, regulation of immunoglobulin genes. The origins of the modern VH gene locus can be traced to the primitive teleost fishes.     

Molecular sequence accuracy and the analysis of protein coding regions.

Molecular sequences, like all experimental data, have finite error rates. The impact of errors on the information content of molecular sequence data is dependent on the analytic paradigm used to interpret the data. We studied the impact of nucleic acid sequence errors on the ability to align predicted amino acid sequences with the sequences of related proteins. We found that with a simultaneous translation and alignment algorithm, identification of sequence homologies is resilient to the introduction of random errors. Proteins with greater than 30% sequence identity can be reliably recognized even in the presence of 1% frameshifting (insertion or deletion) error rates and 5% base substitution rates. Incorporation of prior knowledge about the location and characteristics of errors improves tolerance to error of amino acid sequence alignments. Similarly, inclusion of prior knowledge of biased codon utilization by yeast (Saccharomyces cerevisiae) allows reliable detection of correct reading frames in yeast sequences even in the presence of 5% substitution and 1% frameshift errors.     

Mutation affecting the expression of immunoglobulin variable regions in the rabbit.

We have found a variant of the allotype allele a2 in the rabbit, which presumably arose by mutation, that segregates as expected for an allele at the a locus. This allele is called "ali" and the corresponding rabbit strain is called "Alicia." In heterozygous animals (ali/a1 and ali/a3) the concentration of a2 molecules is lower by a factor of 1000 than in standard a2/a2 homozygotes. In homozygous ali/ali individuals the a2 concentration varies with age--i.e., very low in young rabbits and higher in older ones--but it never reaches normal levels. The low level of a2 is compensated by increased amounts of a-negative molecules. Southern blot analysis did not reveal any gross changes in the intron between JH and C mu (joining region of immunoglobulin heavy chain and constant region of immunoglobulin mu chain) or in the number of VH gene segments encoding a locus specificities. We suggest that the ali phenotype is due to a mutation in a control element.     

Kinetoplast DNA minicircles: regions of extensive sequence divergence.

Previous work has shown that the kinetoplast minicircle DNA of Leishmania species exhibits species-specific sequence divergence and this observation has led to the development of a DNA probe-based diagnostic test for leishmaniasis. In the work reported here, we demonstrate that the minicircle is composed of three types of DNA sequences with differing specificities reflecting different rates of DNA sequence change. A library of cloned fragments of kinetoplast DNA (kDNA) from Leishmania mexicana amazonensis was prepared and the cloned subfragments were found to contain DNA sequences with different taxonomic specificities based on hybridization analysis with various species of Leishmania. Four groups of subfragments were found, those that hybridized with a large number of Leishmania sp. as well as sequences unique to the species, subspecies, or isolate. Analysis of nested deletions of a single, full-length minicircle demonstrates that these different taxonomic specificities are contained within a single minicircle. This implies that different regions of a single minicircle have DNA sequences that diverge at different rates. These sequences represent potentially valuable tools in diagnostic, epidemiologic, and ecological studies of leishmaniasis and provide the basis for a model of kDNA sequence evolution.     

Recombined flanks of the variable and joining segments of immunoglobulin genes.

The mechanism of generating immunoglobulin light chain genes by rearrangement of variable (V), joining (J), and constant (C) gene segments is still unknown. It has been discussed mostly in terms of excision and deletion of the DNA between the recombined V and J gene segments. However, the finding of DNA digests from the mouse myeloma T of a fragment (called f-T) that contains the 3' flank of a V kappa and the 5' flank of a J1 gene segment argued against a simple deletion mechanism [Steinmetz, M., Altenburger, W. & Zachau, H. G. (1980) Nucleic Acids Res. 8, 1709--1720]. The origin of fragment f-T has now been investigated by cloning and determining the sequence of the germ-line V gene segment that apparently participated in its formation. Moreover, analogous fragments containing flanking sequences were isolated from the myelomas MOPC 173 and 41 (f-173 and f-41) and studied by sequence analysis. The f fragments appear to be recombination products of V--J rearrangements reciprocal to rearranged kappa genes but, at least in the cases of f-T and f-173, not of the rearranged V genes present in the same tumor cell. This fact is best explained by a sister chromatid exchange mechanism of V--J recombination because, by this model, the rearranged V genes and the reciprocal flank recombination products would segregate into different cells during the following mitosis. The possibility is suggested that there exists in lymphocyte differentiation more than one mechanism of V--J recombination.     

Escherichia coli extract-catalyzed recombination in switch regions of mouse immunoglobulin genes.

We have shown that Escherichia coli extracts catalyze recombination between mouse immunoglobulin mu and alpha genes inserted separately in lambda phage vectors carrying different genetic markers. Most of the recombination sites in the inserts are located in the switch regions of the heavy chain genes, as previously found in the expressed genes of myeloma cells. The recombination took place at relatively high frequency (10(-4)). The recombinational system in E. coli or lambda phage seems to prefer short nucleotide sequences similar to those used in the class switch recombination.     

Identification of a DNA binding protein that recognizes the nonamer recombinational signal sequence of immunoglobulin genes.

Extracts of nuclei from B- and T-lymphoid cells contain a protein that binds specifically to the conserved nonamer DNA sequence within the recombinational signals of immunoglobulin genes. Complexes with DNA fragments from four kappa light-chain joining (J) segments have the same electrophoretic mobility. Nonamer-containing DNA fragments from heavy-chain and light-chain genes compete for binding. Within the 5'-flanking DNA of the J kappa 4 gene segment, the binding site has been localized to a 27-base-pair interval spanning the nonamer region. The binding activity is recovered as a single peak after ion-exchange chromatography. The site of binding of the protein and its presence in nuclei of lymphoid cells suggest that it may function in the assembly of immunoglobulin genes.     

Molecular characterization of J558 genes encoding tight-skin mouse autoantibodies: identical heavy-chain variable genes code for antibodies with different specificities.

Tight-skin mouse, a mutant strain with a single gene defect, develops cutaneous hyperplasia and specific autoantibodies, like humans affected by scleroderma. The autoantibodies produced in the tight-skin mouse are encoded primarily by heavy-chain variable (VH) genes from the J558 family. To understand the genetic basis of production of autoantibodies, we have analyzed the structure of J558 genes encoding these autoantibodies. The results showed that J558 genes encoding these antibodies were not derived from a selected germ-line gene(s) or a single subfamily but were derived from genes belonging to diverse J558 subfamilies. However, two prototype VH genes representing two new subfamilies were found to be repeatedly expressed in their germ-line form in eight independent clones. Autoantibodies with distinct specificities appear to be generated by pairing of similar/identical VH genes with different V kappa genes derived from the same or different families. Fourteen of 18 autoantibodies shared a conserved heptapeptide sequence motif, YNEKFKG, in the second complementarity-determining region of heavy chains. Usage of germ-line genes from diverse J558 subfamilies bearing a common motif to encode autoantibodies suggests a regulatory role for this motif. Thus, selection and expansion of the autoreactive B-cell repertoire in the tight-skin mouse appear to be VH-gene mediated. The frequency of N nucleotide addition at diversity-joining (D-JH) junctions was lower, whereas the frequency of usage of the DFL16 segment was higher. Finally, in contrast to normal and other autoimmune mouse strains, the frequencies of D-D fusions and D inversions were higher in tight-skin mouse total immunoglobulin as well as autoantibody repertoires.