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

Identification of a GATA-overlapping sequence within the enhancer of the murine GPIIb promoter that induces transcriptional deregulation in human K562 cells

The human and the murine glycoprotein platelet IIb (GPIIb) promoters are megakaryocyte specific in human and murine cell systems, respectively. Here we show that the murine promoter is, however, highly active when transfected in K562 human cells in which the human promoter is almost inactive. A murine promoter, in which the enhancer element was replaced by the human, retrieves its megakaryocytic specificity in human cell lines. The human and murine GATA-binding sites located in the enhancer region display slight sequence divergence next to the consensus GATA core sequence. Gel shift experiments show that, although the murine and the human GATA sequences both bind GATA-1, the murine sequence alone forms an additional complex (B) not detected with the human sequence. When the murine GATA-containing region is replaced by the human in the context of the murine GPIIb promoter, megakaryocyte specificity is restored in the human cell lines. A G nucleotide 3′ to GATA appears crucial because its substitution abrogates B but not GATA-1 binding and restores megakaryocyte specificity to the murine promoter. Conversely, substitution of the human GATA-1 binding sequence by its murine homologue that binds both GATA-1 and complex B induces an abnormal activity for the human promoter in K562 cells. Altogether, our data suggest that limited changes in the GATA-containing enhancer of the GPIIb promoter can induce the recruitment of accessory proteins that could be involved in alteration of a megakaryocyte-restricted gene activation program.

Identification of a GATA-overlapping sequence within the enhancer of the murine GPIIb promoter that induces transcriptional deregulation in human K562 cells

The human and the murine glycoprotein platelet IIb (GPIIb) promoters are megakaryocyte specific in human and murine cell systems, respectively. Here we show that the murine promoter is, however, highly active when transfected in K562 human cells in which the human promoter is almost inactive. A murine promoter, in which the enhancer element was replaced by the human, retrieves its megakaryocytic specificity in human cell lines. The human and murine GATA-binding sites located in the enhancer region display slight sequence divergence next to the consensus GATA core sequence. Gel shift experiments show that, although the murine and the human GATA sequences both bind GATA-1, the murine sequence alone forms an additional complex (B) not detected with the human sequence. When the murine GATA-containing region is replaced by the human in the context of the murine GPIIb promoter, megakaryocyte specificity is restored in the human cell lines. A G nucleotide 3′ to GATA appears crucial because its substitution abrogates B but not GATA-1 binding and restores megakaryocyte specificity to the murine promoter. Conversely, substitution of the human GATA-1 binding sequence by its murine homologue that binds both GATA-1 and complex B induces an abnormal activity for the human promoter in K562 cells. Altogether, our data suggest that limited changes in the GATA-containing enhancer of the GPIIb promoter can induce the recruitment of accessory proteins that could be involved in alteration of a megakaryocyte-restricted gene activation program.

V(D)J recombinational signal sequence DNA binding activities expressed by fetal bovine thymus

V(D)J recombination, or immunoglobulin gene rearrangement is a developmentally regulated, cell type specific, site directed recombination event that brings either immunoglobulin or T-cell receptor gene segments together to form mature, expressible Ig or TCT genes. This DNA recombination is directed by the recombinational signal sequences or RSS elements present adjacent to Ig and TCR gene segments. The RSS element is composed of a conserved nonamer element and a conserved heptamer element separated by a conserved length spacer region. In this report, we examine the expression of DNA binding proteins that interact with the RSS element in the bovine fetal thymus using EMSA assays. Our data indicates that the nonamer portion of the RSS element is the primary site of recognition for RSS binding proteins expressed in the bovine fetal thymus. We also show that these proteins are expressed from early stages of bovine fetal development through to full term development.

Characterization of a 3'-5' exonuclease associated with VDJP

VDJP (V(D)J RSS Dependent DNA Joining Protein) was cloned based on binding to the nonamer portion of the V(D)J recombinational signal sequence (RSS), and genetic analysis revealed that VDJP is encoded by the same gene as the large subunit of Replication Factor C (RF-C). Recombinant VDJP has a site directed DNA joining activity and is capable of forming a covalent bond between DNA fragments containing an RSS element near their ends and exhibits 3' to 5' exonuclease activity. In this report, we examine the biochemical properties of the VDJP exonuclease activity such as directionality of nuclease action (3' to 5' or 5' to 3'), single-strand substrate preference, cleavage products, dependence on cofactors and metal cations, and optimal reaction conditions. From this analysis, we conclude that VDJP has an intrinsic 3'-5' exonuclease activity that produces mononucleotide products.

Stromal cells and cytokines in the induction of recombination activating gene (RAG) expression in a human lymphoid progenitor cell

The activation of recombination activating genes (RAGs) plays critical roles in the V(D)J gene recombination machinery and lymphocyte repertoire formation. However, the regulation of RAG gene expression in humans as well as animals is poorly understood. We show that RAG gene expression is activated in a human lymphoid progenitor cell line (FL8.2.4.4) by coculturing them on a bone marrow-derived stromal cell line (PA6) in the presence of cytokines. The RAG transcripts become detectable in 12 hours after initiation of culture, and the increased level is sustained at 24 hours. Among the cytokines, IL-3, IL-6, and IL-7, but not IL-2, IL-4, SCF, GM-CSF induces RAG activation. IL-3, IL-6, and IL-7 exert their effect synergistically on RAG activation. A cognate interaction between FL8.2.4.4 cells and PA6 stromal cells seems to be prerequisite for RAG activation. RAG transcripts are inducible in FL8.2.4.4 cells when cocultured on paraformaldehyde fixed-PA6 stromal cells in the presence of cytokines. These data indicate that two separate signals are both required for induction of RAG activation in lymphoid progenitors; one from the cell surface molecule(s) on stromal cells, and the other from recombinant cytokine(s). The expression of RAG mRNA in FL8.2.4.4 cells is concomitant with induction of recombinase activity. Thus, this system may provide a useful means for further understanding of the mechanisms controlling RAG activation and lymphocyte development in human system.

rag-1 and rag-2 are components of a high-molecular-weight complex, and association of rag-2 with this complex is rag-1 dependent

Despite the essential and synergistic functions of the rag-1 and rag-2 proteins in V(D)J recombination and lymphocyte development, little is known about the biochemical properties of the two proteins. We have developed cell lines expressing high levels of the rag proteins and specific, sensitive immunological reagents for their detection, and we have examined the physical properties of the rag proteins in vitro and their subcellular localizations in vivo. rag-1 is tightly associated with nuclear structures, requires a high salt concentration to maintain its solubility, and is a component of large, heterogeneously sized complexes. Furthermore, the presence of rag-1 alters the behavior of rag-2, conferring on it properties similar to those of rag-1 and changing its distribution in the nucleus. We demonstrate that rag-1 and rag-2 are present in the same complex by coimmunoprecipitation, and we provide evidence that these complexes contain more molecules of rag-2 than of rag-1. The demonstration of intracellular complexes containing rag-1 and rag-2 raises the possibility that interaction between these proteins is necessary for their biological function.

Transcription factor GATA-1 regulates human HOXB2 gene expression in erythroid cells

The human HOXB2 gene is a member of the vertebrate Hox gene family that contains genes coding for specific developmental stage DNA-binding proteins. Remarkably, within the hematopoietic compartment, genes of the HOXB complex are expressed specifically in erythromegakaryocytic cell lines and, for some of them, in hematopoietic progenitors. Here, we report the study of HOXB2 gene transcriptional regulation in hematopoietic cells, an initial step in understanding the lineage-specific expression of the whole HOXB complex in these cells. We have isolated the HOXB2 5'-flanking sequence and have characterized a promoter fragment extending 323 base pairs upstream from the transcriptional start site, which, in transfection experiments, was sufficient to direct the tissue-specific expression of HOXB2 in the erythroid cell line K562. In this fragment, we have identified a potential GATA-binding site that is essential to the promoter activity as demonstrated by point mutation experiments. Gel shift analysis revealed the formation of a specific complex in both erythroleukemic lines K562 and HEL that could be prevented by the addition of a specific antiserum raised against GATA-1 protein. These findings suggest a regulatory hierarchy in which GATA-1 is upstream of the HOXB2 gene in erythroid cells.

The SCID mouse: relevance as an animal model system for studying human disease

The simultaneous description some 5 years ago of two methods for the partial reconstitution of a human immune system in severe combined immune-deficient (SCID) mice (collectively, human:SCID mice) was met with great enthusiasm. At the time, it was hoped that human:SCID mice would provide experimental animal model systems for studying human disease and the human immune system. Many of these hopes have been borne out. Importantly, the experimental results obtained from these chimeric human/animal studies appear to be relevant to human disease and immune function. In spite of these glowing achievements, the SCID mouse may not represent the optimal experimental system with which to address these questions. The incomplete penetrance ("leakiness") of the scid mutation and the recent discovery that the mutation is not lymphoid specific, but rather affects a general DNA repair pathway, will only serve to complicate the interpretation of already complex biological interactions. Recently other immune-deficient mice have been described that appear to overcome one or both of these problems and thus these mice could represent improved hosts for the adaptive transfer of a human immune system. The current status of the SCID mouse in light of these new findings is discussed.

RAG-1 and RAG-2 are not sufficient to direct all phases of immunoglobulin gene rearrangement in pre-B-cell lines

To probe the factors controlling immunoglobulin heavy-chain gene rearrangement, we analyzed Abelson virus-transformed pre-B-cell lines that fail to undergo VH-to-DJH joining at an appreciable frequency. Despite this feature, some of these cell lines (rechi) rearrange an extrachromosomal recombination substrate at levels normal for transformed pre-B cells. Others (reclo) rearrange these substrates at levels characteristic of nonlymphoid hematopoietic cells. The DJH rearrangements from a representative rechi cell line were aberrant, suggesting that these cells probably fail to complete heavy-chain gene assembly because some of the necessary cis-acting signals are missing. In contrast, both DJH rearrangements from a reclo cell line appeared normal in structure, indicating that trans-acting factors necessary for recombination might be missing. Introduction of the RAG-1 and RAG-2 genes, genes encoding two such factors, failed to confer a rechi phenotype to these cells. However, fusion of the reclo cells to a rechi cell line generated a high frequency of rechi hybrids. In addition, most of the hybrids rearranged the endogenous kappa light-chain locus. Neither the rechi phenotype nor kappa-chain rearrangement correlated with levels of RAG-1 and RAG-2 expression in all of the hybrids. Thus, both gene transfer and cell fusion experiments indicate that RAG-1 and RAG-2 are not sufficient to activate immunoglobulin gene recombination in at least some pre-B-cell lines. In addition, the fusion experiments suggest that two gene products in addition to RAG-1 and RAG-2 may be required for kappa-gene rearrangement.

RAG-1 and RAG-2 are not sufficient to direct all phases of immunoglobulin gene rearrangement in pre-B-cell lines

To probe the factors controlling immunoglobulin heavy-chain gene rearrangement, we analyzed Abelson virus-transformed pre-B-cell lines that fail to undergo VH-to-DJH joining at an appreciable frequency. Despite this feature, some of these cell lines (rechi) rearrange an extrachromosomal recombination substrate at levels normal for transformed pre-B cells. Others (reclo) rearrange these substrates at levels characteristic of nonlymphoid hematopoietic cells. The DJH rearrangements from a representative rechi cell line were aberrant, suggesting that these cells probably fail to complete heavy-chain gene assembly because some of the necessary cis-acting signals are missing. In contrast, both DJH rearrangements from a reclo cell line appeared normal in structure, indicating that trans-acting factors necessary for recombination might be missing. Introduction of the RAG-1 and RAG-2 genes, genes encoding two such factors, failed to confer a rechi phenotype to these cells. However, fusion of the reclo cells to a rechi cell line generated a high frequency of rechi hybrids. In addition, most of the hybrids rearranged the endogenous kappa light-chain locus. Neither the rechi phenotype nor kappa-chain rearrangement correlated with levels of RAG-1 and RAG-2 expression in all of the hybrids. Thus, both gene transfer and cell fusion experiments indicate that RAG-1 and RAG-2 are not sufficient to activate immunoglobulin gene recombination in at least some pre-B-cell lines. In addition, the fusion experiments suggest that two gene products in addition to RAG-1 and RAG-2 may be required for kappa-gene rearrangement.

Recombination sequence-binding protein in thymocytes undergoing T-cell receptor gene rearrangement

Rearrangement of T-cell antigen receptor and immunoglobulin genes occurs in immature lymphoid cells by an unknown mechanism. To identify components of the rearrangement machinery, we isolated a population of murine thymocytes enriched for rearranging pre-T cells. In the nuclear fraction of these cells, we detected a protein that specifically bound the recombination sequences that flank T-cell receptor and immunoglobulin genes and are required for their rearrangement. This protein recognized both heptamer and nonamer motifs of the recombination sequence, separated by either 12 or 23 bp. The protein complexed with the recombination sequence oligonucleotide had an apparent molecular mass of 30 kDa. The binding characteristics of the protein and its presence in rearranging thymocytes and cell lines suggest that it could serve as the recognition unit of a recombinase complex.

Purification of core-binding factor, a protein that binds the conserved core site in murine leukemia virus enhancers

The Moloney murine leukemia virus causes thymic leukemias when injected into newborn mice. A major genetic determinant of the thymic disease specificity of the Moloney virus genetically maps to two protein binding sites in the Moloney virus enhancer, the leukemia virus factor b site and the adjacent core site. Point mutations introduced into either of these sites significantly shifts the disease specificity of the Moloney virus from thymic leukemia to erythroleukemia (N. A. Speck, B. Renjifo, E. Golemis, T. Frederickson, J. Hartley, and N. Hopkins, Genes Dev. 4:233-242, 1990). We have purified several polypeptides that bind to the core site in the Moloney virus enhancer. These proteins were purified from calf thymus nuclear extracts by selective pH denaturation, followed by chromatography on heparin-Sepharose, nonspecific double-stranded DNA-cellulose, and core oligonucleotide-coupled affinity columns. We have achieved greater than 13,000-fold purification of the core-binding factors (CBFs), with an overall yield of approximately 19%. Analysis of purified protein fractions by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis reveals more than 10 polypeptides. Each of the polypeptides was recovered from an SDS-polyacrylamide gel, and those in the molecular size range of 19 to 35 kDa were demonstrated to have core-binding activity. The purified CBFs were shown by DNase I footprint analyses to bind the core site in the Moloney virus enhancer specifically, and also to core motifs in the enhancers from a simian immunodeficiency virus, the immunoglobulin mu chain, and T-cell receptor gamma-chain genes.

Purification of core-binding factor, a protein that binds the conserved core site in murine leukemia virus enhancers

The Moloney murine leukemia virus causes thymic leukemias when injected into newborn mice. A major genetic determinant of the thymic disease specificity of the Moloney virus genetically maps to two protein binding sites in the Moloney virus enhancer, the leukemia virus factor b site and the adjacent core site. Point mutations introduced into either of these sites significantly shifts the disease specificity of the Moloney virus from thymic leukemia to erythroleukemia (N. A. Speck, B. Renjifo, E. Golemis, T. Frederickson, J. Hartley, and N. Hopkins, Genes Dev. 4:233-242, 1990). We have purified several polypeptides that bind to the core site in the Moloney virus enhancer. These proteins were purified from calf thymus nuclear extracts by selective pH denaturation, followed by chromatography on heparin-Sepharose, nonspecific double-stranded DNA-cellulose, and core oligonucleotide-coupled affinity columns. We have achieved greater than 13,000-fold purification of the core-binding factors (CBFs), with an overall yield of approximately 19%. Analysis of purified protein fractions by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis reveals more than 10 polypeptides. Each of the polypeptides was recovered from an SDS-polyacrylamide gel, and those in the molecular size range of 19 to 35 kDa were demonstrated to have core-binding activity. The purified CBFs were shown by DNase I footprint analyses to bind the core site in the Moloney virus enhancer specifically, and also to core motifs in the enhancers from a simian immunodeficiency virus, the immunoglobulin mu chain, and T-cell receptor gamma-chain genes.

HMG1-related DNA-binding protein isolated with V-(D)-J recombination signal probes

In order to isolate cDNA clones for DNA-binding components of the V-(D)-J recombinase, phage libraries from a pre-B-cell line were screened with a radiolabeled probe containing recombination signal sequences (RSS). Among prospective clones, cDNA T160 was analyzed further. It produced a protein of 80.6 kDa which bound to DNA containing RSS but not to DNA in which the RSS had been mutated. A search of a data base revealed that the T160 protein has significant sequence homology (56%) to the nonhistone chromosomal protein HMG1 within the C-terminal region of 80 amino acids. DNA-binding analysis with truncated proteins showed that the HMG homology region is responsible for DNA binding. Using restriction fragment length polymorphisms, the T160 gene was mapped at the proximal end of mouse chromosome 2. Evidence was obtained for genetic linkage between the T160 gene and the recombination activator genes RAG-1 and RAG-2.

HMG1-related DNA-binding protein isolated with V-(D)-J recombination signal probes

In order to isolate cDNA clones for DNA-binding components of the V-(D)-J recombinase, phage libraries from a pre-B-cell line were screened with a radiolabeled probe containing recombination signal sequences (RSS). Among prospective clones, cDNA T160 was analyzed further. It produced a protein of 80.6 kDa which bound to DNA containing RSS but not to DNA in which the RSS had been mutated. A search of a data base revealed that the T160 protein has significant sequence homology (56%) to the nonhistone chromosomal protein HMG1 within the C-terminal region of 80 amino acids. DNA-binding analysis with truncated proteins showed that the HMG homology region is responsible for DNA binding. Using restriction fragment length polymorphisms, the T160 gene was mapped at the proximal end of mouse chromosome 2. Evidence was obtained for genetic linkage between the T160 gene and the recombination activator genes RAG-1 and RAG-2.

Cloning and characterization of a protein binding to the J kappa recombination signal sequence of immunoglobulin genes

A protein with molecular weight of 60,000 that binds to the recombination signal sequence (RS) of the immunoglobulin J kappa segment was purified from the nuclear extract of a murine pre B cell line 38B9. This binding protein was found in lymphoid cell lines but not in non-lymphoid cell lines. The Kd value of the J kappa RS binding protein to the J kappa RS was 1 nM. The cDNA clone (RBP-2) was isolated based on partial amino-acid sequence of this protein. This cDNA encodes 526 amino-acid residues, and its sequence does not show extensive overall homology with any known proteins, but displays an interesting homology to a 40-residue region that is conserved among a subset of site specific recombinase (integrase family).

Deletion of human JK segments by site-specific recombination recognizing the conserved nonamer and heptamer sequences

Mapping and partial sequencing of the productive K chain genomic DNA of FK-001 demonstrated a 1.8-kb deletion including the JK2, JK3, JK4, and JK5 segments. This deletion occurred between the heptamer recombination signal sequence of the JK2 segment and the heptamer-like sequence located 1.8 kb downstream of the JK2 segment. The recombination reaction kept the reciprocally joined signal sequences on the chromosome and deleted the intervening DNA segment. The cloned FK-001 K chain gene was expressed efficiently in mouse myeloma cells, demonstrating that the 1.8-kb deleted region conferred no functions for gene expression.

Identification of a recombinational signal sequence-specific DNA-binding protein(s) of Mr 115,000 in the nuclear extracts from immature lymphoid cell lines

Rearrangements of immunoglobulin genes are mediated by highly conserved heptamer and nonamer recombinational signal sequence. Using a protein-blotting procedure, a heptamer and nonamer recombinational signal sequence-specific DNA-binding protein(s) was examined in the nuclear extracts from lymphoid and nonlymphoid cell lines. Nuclear extracts were subjected to SDS-polyacrylamide gel electrophoresis, and transferred by electroblotting to nitrocellulose filters. Then the filters were hybridized to 32P-labelled synthetic double-stranded heptamer-23bp-nonamer or nonamer-12bp-heptamer recombinational signal sequence probes. A relatively large amount of a DNA-binding protein(s) of Mr 115,000 for both probes was detected in the nuclear extracts from immature B and immature T cell lines. No DNA-binding proteins were detected in a myeloma cell line. Interestingly, this DNA-binding protein(s) might be able to recognize both heptamer and nonamer. Recombinational signal sequence-specific DNA-binding activity of the protein(s) and the presence of the protein(s) in a stage-specific manner strongly suggest that the protein(s) of Mr 115,000 detected here may play an important role in the recombination of Ig and TCR genes.

Purification and characterization of a protein that binds to the recombination signal sequence of the immunoglobulin J kappa segment

A protein that binds to the recombination signal sequence (RS) of the immunoglobulin J kappa segment was purified almost to homogeneity from the nuclear extract of a murine pre-B cell line 38B9. A similar binding protein was found in lymphoid cell lines but not in non-lymphoid cell lines. The binding activity was associated with a polypeptide with a molecular weight of 60,000. DNase I footprinting analysis demonstrated that this binding protein interacted with the heptamer and several 3' bases close to the heptamer. The Kd value of the J kappa RS binding protein to the J kappa RS was 1 nM. One base substitution in the heptamer of the J kappa RS greatly reduced the affinity of the J kappa RS binding protein. The high specificity of the binding site of the J kappa RS binding protein suggests that this protein may be involved in V-J recombination.

NBP, a protein that specifically binds an enhancer of immunoglobulin gene rearrangement: purification and characterization

Immunoglobulin and T-cell receptor (TCR) genes are encoded in discrete germ line DNA segments that are joined by site-specific recombination during lymphocyte development. These DNA rearrangements are mediated by conserved heptamer and nonamer DNA sequence elements that lie near the sites of recombination. In this paper we show that the nonamer element coincides with the recognition site for a specific DNA-binding protein: mutations within the nonamer sequence, but not outside of it, decrease affinity for the binding protein by 300- to 1000-fold. Deletion of the binding site for the protein results in at least a 50-fold decrease in recombination frequency in vivo. By a combination of conventional and recognition site affinity chromatography, we have achieved greater than 20,000-fold purification of the protein from calf thymus, with an overall yield of 22%. The purified protein, which we now call nonamer-binding protein (NBP), has an apparent molecular weight of 63,000 and a frictional ratio of 1.27, suggesting that it exists as a globular monomer in 0.5 M NaCl. Our observations suggest that NBP is a component of the recombinational apparatus.

Methylated DNA-binding protein is present in various mammalian cell types

A DNA-binding protein from human placenta, methylated DNA-binding protein (MDBP), binds to certain DNA sequences only when they contain 5-methylcytosine (m5C) residues at specific positions. We found a very similar DNA-binding activity in nuclear extracts of rat tissues, calf thymus, human embryonal carcinoma cells, HeLa cells, and mouse LTK cells. Like human placental MDBP, the analogous DNA-binding proteins from the above mammalian cell lines formed a number of different low-electrophoretic-mobility complexes with a 14-bp MDBP-specific oligonucleotide duplex. All of these complexes exhibited the same DNA methylation specificity and DNA sequence specificity. From the extracts of rat and calf tissues, oligonucleotide protein complexes formed that also had the same specificity as human placental MDBP although they had a higher electrophoretic mobility probably due to digestion by proteases in the nuclear extracts. Although MDBP activity was found in various mammalian cell types, it was not detected in extracts of cultured mosquito cells and so may be associated only with cells with vertebrate-type DNA methylation.