The LIM-only protein Lmo2 is a bridging molecule assembling an erythroid, DNA-binding complex which includes the TAL1, E47, GATA-1 and Ldb1/NLI proteins.

The LIM-only protein Lmo2, activated by chromosomal translocations in T-cell leukaemias, is normally expressed in haematopoiesis. It interacts with TAL1 and GATA-1 proteins, but the function of the interaction is unexplained. We now show that in erythroid cells Lmo2 forms a novel DNA-binding complex, with GATA-1, TAL1 and E2A, and the recently identified LIM-binding protein Ldb1/NLI. This oligomeric complex binds to a unique, bipartite DNA motif comprising an E-box, CAGGTG, followed approximately 9 bp downstream by a GATA site. In vivo assembly of the DNA-binding complex requires interaction of all five proteins and establishes a transcriptional transactivating complex. These data demonstrate one function for the LIM-binding protein Ldb1 and establish a function for the LIM-only protein Lmo2 as an obligatory component of an oligomeric, DNA-binding complex which may play a role in haematopoiesis.

Chromosomal translocations and leukaemia: a role for LMO2 in T cell acute leukaemia, in transcription and in erythropoiesis.

The LMO2 gene associated with T cell acute leukaemia has been used as an example of a gene activated by association with the T cell receptor genes after chromosomal translocations. The gene is shown to encode a LIM protein which is involved in protein interactions and during normal haematopoiesis is necessary for erythroid development. LMO2 has been shown to cause tumours when aberrantly expressed and to be able to heterodimerise with TAL1 to facilitate tumour development.

Protein dimerization between Lmo2 (Rbtn2) and Tal1 alters thymocyte development and potentiates T cell tumorigenesis in transgenic mice.

The LMO2 and TAL1 genes were first identified via chromosomal translocations and later found to encode proteins that interact during normal erythroid development. Some T cell leukaemia patients have chromosomal abnormalities involving both genes, implying that LMO2 and TAL1 act synergistically to promote tumorigenesis after their inappropriate co-expression. To test this hypothesis, transgenic mice were made which co-express Lmo2 and Tal1 genes in T cells. Dimers of Lmo2 and Tal1 proteins were formed in thymocytes of double but not single transgenic mice. Furthermore, thymuses of double transgenic mice were almost completely populated by immature T cells from birth, and these mice develop T cell tumours approximately 3 months earlier than those with only the Lmo2 transgene. Thus interaction between these two proteins can alter T cell development and potentiate tumorigenesis. The data also provide formal proof that TAL1 is an oncogene, apparently acting as a tumour promoter in this system.

The MAP kinase phosphorylation site of TAL1 occurs within a transcriptional activation domain.

Alteration of the TAL1 gene is the most common genetic lesion found in patients with T cell acute lymphoblastic leukemia. TAL1 encodes a basic helix-loop-helix transcription factor that is phosphorylated on serine residue 122 by the mitogen-activated protein (MAP) kinase ERK1. Here we show that the amino-terminal sequences of TAL1 (residues 1-166) function in vivo as a transcriptional activation domain. Mutation of serine residue 122 reduces the potency of the transactivation domain by more than half. The data suggest that the amino-terminal transactivation domain of TAL1 is positively regulated by S122 phosphorylation and that the functional properties of TAL1 can be influenced by signal transduction pathways that involve the MAP kinases.

Specific in vivo association between the bHLH and LIM proteins implicated in human T cell leukemia.

The protein products of proto-oncogenes implicated in T cell acute lymphoblastic leukemia include two distinct families of presumptive transcription factors. RBTN1 and RBTN2 encode highly related proteins that possess cysteine-rich LIM motifs. TAL1, TAL2 and LYL1 encode a unique subgroup of basic helix-loop-helix (bHLH) proteins that share exceptional homology in their bHLH sequences. We have found that RBTN1 and RBTN2 have the ability to interact with each of the leukemogenic bHLH proteins (TAL1, TAL2 and LYL1). These interactions occur in vivo and appear to be mediated by sequences within the LIM and bHLH domains. The LIM-bHLH interactions are highly specific in that RBTN1 and RBTN2 will associate with TAL1, TAL2 and LYL1, but not with other bHLH proteins, including E12, E47, Id1, NHLH1, AP4, MAX, MYC and MyoD1. Moreover, RBTN1 and RBTN2 can interact with TAL1 polypeptides that exist in assembled bHLH heterodimers (e.g. TAL1-E47), suggesting that the RBTN proteins can influence the functional properties of TAL1. Finally, we have identified a subset of leukemia patients that harbor tumor-specific rearrangements of both their RBTN2 and TAL1 genes. Thus, the activated alleles of these genes may promote leukemia cooperatively, perhaps as a result of bHLH-LIM interactions between their protein products.

Positive and negative transcriptional control by the TAL1 helix-loop-helix protein.

Tumor-specific activation of the TAL1 gene is the most common genetic defect associated with T-cell acute lymphoblastic leukemia. The TAL1 gene products possess a basic helix-loop-helix (bHLH) motif, a protein-dimerization and DNA-binding domain found in several transcription factors. TAL1 polypeptides interact, in vitro and in vivo, with class A bHLH proteins (e.g., E47) to form heterodimers with sequence-specific DNA-binding activity. In this study, we show that TAL1 can regulate the transcription of an artificial reporter gene that contains binding sites for bHLH heterodimers involving TAL1. Transcription of the reporter is strongly induced by E47-E47 homodimers and moderately induced by TAL1-E47 heterodimers. Thus, in a cellular environment that allows formation of E47-E47 homodimers (e.g., in the absence of Id regulatory proteins) TAL1 can repress transcription by recruiting E47 into bHLH complexes with less transcriptional activity (i.e., TAL1-E47 heterodimers). However, in other settings TAL1 can activate transcription because TAL1-E47 heterodimers are more resistant to negative regulation by Id proteins. Hence, TAL1 can potentially regulate transcription in either a positive or negative fashion.

Formation of in vivo complexes between the TAL1 and E2A polypeptides of leukemic T cells.

Tumor-specific activation of the TAL1 gene occurs in approximately 25% of patients with T-cell acute lymphoblastic leukemia (T-ALL). The TAL1 gene products possess a basic helix-loop-helix (bHLH) domain that interacts in vitro with the bHLH proteins (E12 and E47) encoded by the E2A locus. We have now applied two independent methods, the two-hybrid procedure and co-immunoprecipitation analysis, to demonstrate that TAL1 and E2A polypeptides also associate in vivo. These studies show that the bHLH domain of TAL1 selectively interacts with the bHLH domains of E12 and E47, but not with the Id1 helix-loop-helix protein. TAL1 does not self-associate to form homodimeric complexes, implying that the in vivo functions of TAL1 depend on heterologous interaction with other bHLH proteins such as E12 and E47. Co-immunoprecipitation analysis revealed the presence of endogenous TAL1/E2A complexes in Jurkat cells, a leukemic line derived from a T-ALL patient. Thus, the malignant properties of TAL1 may be due to obligate interaction with the E2A polypeptides.

Evidence for regulation of human platelet adenylate cyclase by phosphorylation. Inhibition by ATP and guanosine 5'-[beta-thio]diphosphate occur by distinct mechanisms.

1. Incubation of human platelet membranes with guanosine 5'-[beta gamma-imido]triphosphate (p[NH]ppG) causes a time-dependent increase in the activation of adenylate cyclase due to Gs (the stimulatory GTP-binding protein). Forskolin enhances adenylate cyclase activity but does not interfere with the process of activation. The activation follows first-order kinetics in both the presence and the absence of the assay components. 2. ATP in the presence or the absence of an ATP-regenerating system of phosphocreatine and creatine kinase inhibits activation. 3. Hydrolysis of ATP to ADP does not lead to receptor-mediated inhibition of adenylate cyclase acting via Gi (the inhibitory GTP-binding protein). The ADP analogue adenosine 5'-[beta-thio]diphosphate (ADP[S]) does not inhibit the activation process. 4. Phosphocreatine alone inhibits adenylate cyclase activation at concentrations above 1 mM. 5. Inhibition by phosphocreatine is not due to the chelation of free Mg2+ ions. 6. Inhibition by ATP and the other assay components occurs throughout the activation process, decreasing both the rate of activation and the maximum activity obtained. 7. Maximal activation of adenylate cyclase after prolonged incubation with p[NH]ppG slowly reverses in the presence of the assay components. 8. A 10-fold excess of the GDP analogue guanosine 5'-[beta-thio]diphosphate (GDP[S]) over p[NH]ppG inhibits the activation process completely, at all stages of the time course. 9. Preincubations in the presence and absence of ATP, cyclic AMP, phosphocreatine and creatine kinase show equal sensitivity to increasing GDP[S] concentration. These data show that the inhibition observed in the presence of ATP is not due to endogenous or contaminating guanine nucleotides, and suggest that phosphoryl transfer may regulate adenylate cyclase activity.