GW body disassembly triggered by siRNAs independently of their silencing activity.

GW bodies (or P-bodies) are cytoplasmic granules containing proteins involved in both mRNA degradation and storage, including the RNA interference machinery. Their mechanism of assembly and function are still poorly known although their number depends upon the flux of mRNA to be stored or degraded. We show here that silencing of the translational regulator CPEB1 leads to their disappearance, as reported for other GW body components. Surprisingly, the same results were obtained with several siRNAs targeting genes encoding proteins unrelated to mRNA metabolism. The disappearance of GW bodies did not correlate with the silencing activity of the siRNA and did not inhibit further silencing by siRNA. Importantly, in most cases, GW bodies were rapidly reinduced by arsenite, indicating that their assembly was not prevented by the inhibition of the targeted or off-target genes. We therefore propose that some siRNA sequences affect mRNA metabolism so as to diminish the amount of mRNA directed to the GW bodies. As an exception, GW bodies were not reinduced following Rck/p54 depletion by interference, indicating that this component is truly required for the GW body assembly. Noteworthy, Rck/p54 was dispensable for the assembly of stress granules, in spite of their close relationship with the GW bodies.

The translational regulator CPEB1 provides a link between dcp1 bodies and stress granules.

The cytoplasmic polyadenylation element-binding protein (CPEB) has been characterized in Xenopus laevis as a translational regulator. During the early development, it behaves first as an inhibitor and later as an activator of translation. In mammals, its closest homologue is CPEB1 for which two isoforms, short and long, have been described. Here we describe an additional isoform with a different RNA recognition motif, which is differentially expressed in the brain and ovary. We show that all CPEB1 isoforms are found associated with two previously described cytoplasmic structures, stress granules and dcp1 bodies. This association requires the RNA binding ability of the protein, whereas the Aurora A phosphorylation site is dispensable. Interestingly, the rck/p54 DEAD box protein, which is known as a CPEB partner in Xenopus and clam, and as a component of dcp1 bodies in mammals, is also present in stress granules. Both stress granules and dcp1 bodies are involved in mRNA storage and/or degradation, although so far no link has been made between the two, in terms of neither morphology nor protein content. Here we show that transient CPEB1 expression induces the assembly of stress granules, which in turn recruit dcp1 bodies. This dynamic connection between the two structures sheds new light on the compartmentalization of mRNA metabolism in the cytoplasm.

Targeting the kinesin Eg5 to monitor siRNA transfection in mammalian cells.

RNA interference, the inhibition of gene expression by double-stranded RNA, provides a powerful tool for functional studies once the sequence of a gene is known. In most mammalian cells, only short molecules can be used because long ones induce the interferon pathway. With the identification of a proper target sequence, the penetration of the oligonucleotides constitutes the most serious limitation in the application of this technique. Here we show that a small interfering RNA (siRNA) targeting the mRNA of the kinesin Eg5 induces a rapid mitotic arrest and provides a convenient assay for the optimization of siRNA transfection. Thus, dose responses can be established for different transfection techniques, highlighting the great differences in response to transfection techniques of various cell types. We report that the calcium phosphate precipitation technique can be an efficient and cost-effective alternative to Oligofectamine in some adherent cells, while electroporation can be efficient for some cells growing in suspension such as hematopoietic cells and some adherent cells. Significantly, the optimal parameters for the electroporation of siRNA differ from those for plasmids, allowing the use of milder conditions that induce less cell toxicity. In summary, a single siRNA leading to an easily assayed phenotype can be used to monitor the transfection of siRNA into any type of proliferating cells of both human and murine origin.

In vivo kinetics of mRNA splicing and transport in mammalian cells.

The kinetics of pre-mRNA processing in living cells is poorly known, preventing a detailed analysis of the regulation of these reactions. Using tetracycline-regulated promoters we performed, during a transcriptional induction, a complete analysis of the maturation of two cellular mRNAs, those for LT-alpha and beta-globin. In both cases, splicing was appropriately described by first-order reactions with corresponding half-lives ranging between 0.4 and 7.5 min, depending on the intron. Transport also behaved as a first-order reaction during the early phase of beta-globin expression, with a nuclear dwelling time of 4 min. At a later time, analysis was prevented by the progressive accumulation within the nucleus of mature mRNA not directly involved in export. Our results further establish for these genes that (i) splicing components are never limiting, even when expression is induced in naive cells, (ii) there is no significant RNA degradation during splicing and transport, and (iii) precursor-to-product ratios at steady state can be used for the determination of splicing rates. Finally, the comparison between the kinetics of splicing during transcriptional induction and during transcriptional shutoff reveals a novel coupling between transcription and splicing.

Mature mRNAs accumulated in the nucleus are neither the molecules in transit to the cytoplasm nor constitute a stockpile for gene expression.

In higher eukaryotes, the regulation of pre-mRNA processing is still poorly known. The accumulation of various mature mRNAs, which can be observed in the nuclei of mammalian cells, is suggestive of a regulatory role of transport. However, the significance of these nuclear mRNA is presently unknown. We have used a tetracycline-regulated promoter to investigate the dynamics of these pools of mRNAs upon arrest of transcription. We observed, for beta-globin and LT-alpha genes, a slow disappearance of these mRNA from the nucleus, with an apparent half-life that is similar to their cytoplasmic half-life. In view of these dynamics, these mRNA cannot simply be mature mRNAs in transit to the cytoplasm. They could be mRNAs retained in the nucleus, provided that the regulation of mRNA stability is comparable in the nucleus and the cytoplasm. But, because of their limited stability, these nuclear mRNAs cannot constitute a significant stock for gene expression. Alternatively, they could reflect a bidirectional transport of mRNA, that is, to and from the cytoplasm, which would provide a direct explanation for the similarity in both compartments of their half-life and poly(A) tail shortening over time.

Secondary alveolar echinococcosis in lymphotoxin-alpha and tumour necrosis factor-alpha deficient mice: exacerbation of Echinococcus multilocularis larval growth is associated with cellular changes in the periparasitic granuloma.

The availability of mice carrying a deletion of LT-alpha and tumour necrosis factor (TNF)-alpha genes enabled us to investigate the role of the TNF during alveolar echinococcosis. We compared the growth rate of Echinococcus multilocularis in LT-alphaTNF-alpha +/+ mice to that of mice having either no or only one LT-alphaTNF-alpha functionnal allele. LT-alphaTNF-alpha -/- mice harboured a significantly higher parasite burden than did the other two populations at 5, 10, and 15 weeks of infection, and they did not survive thereafter. Liver metacestodes removed from these mice were alive and the dehydrogenase activities of peritoneal metacestodes were decreased. Liver lesions regressed in most wild-type mice. Indeed, dead parasites were cordoned by granulomas containing numerous macrophages and lymphocytes leading to focal liver fibrosis at an early stage of infection. In contrast, most of LT-alphaTNF-alpha -/- mice harboured metacestodes interspersed with leucocytes, realising purulent abscesses with secondary extensive irregular fibrosis at a late stage of infection. Heterozygous mice had behavioural characteristics intermediate between homozygous mutants and wild-type mice. Levels of E. multilocularis-specific delayed-type hypersensitivity and serum antibodies were slightly decreased in LT-alphaTNF-alpha -/- mice. This study shows that TNF-alpha and/or LT-alpha genes play an essential role in the immune protection mechanisms against E. multilocularis at the site of infection.

Regulation of mRNA splicing and transport by the tyrosine kinase activity of src.

The regulation of transcription by signal transduction pathways is well documented. In addition, we have previously shown that src can regulate pre-mRNA processing. To investigate which functional domains of src are involved in the regulation of splicing and transport of Lymphotoxin alpha (LTalpha) transcripts, we have used src mutants in the catalytic, SH2 and SH3 domains in association with the Y527F or the E378G activating mutation. Our results establish that the regulation of pre-mRNA processing and transcription can occur independently of each other. The splicing and transport phenotypes require an intact tyrosine kinase domain and both are insensitive to the deletion of the SH3 domain. Therefore these phenotypes do not depend upon the recruitment through the SH3 domain of src of RNA binding proteins (Sam 68, hnRNP K). By contrast, deletions in the SH2 domain have no effect on splicing but either abolish or exacerbate the transport phenotype depending upon the activating mutation (Y527F or E378G). These divergent responses are associated with specific changes in the pattern of tyrosine phosphorylated proteins. Thus, the regulation of transcription, splicing and mRNA transport implicate different effector pathways of src. Furthermore, analysis of the transport phenotype reveals the interplay between the SH2 and catalytic domain of the protein.

Efficient, repeated adenovirus-mediated gene transfer in mice lacking both tumor necrosis factor alpha and lymphotoxin alpha.

The efficiency of adenovirus-mediated gene transfer is now well established. However, the cellular and the humoral immune responses triggered by vector injection lead to the rapid elimination of the transduced cells and preclude any efficient readministration. The present investigation focuses on the role of tumor necrosis factor alpha (TNF-alpha), a proinflammatory cytokine, and the related cytokine lymphotoxin alpha (LTalpha), in mounting an immune reaction against recombinant adenovirus vectors. After gene transfer in the liver, mice genetically deficient for both cytokines (TNF-alpha/LTalpha-/-), in comparison with normal mice, presented a weak acute-phase inflammatory reaction, a reduction in cellular infiltrates in the liver, and a severely impaired T-cell proliferative response to both Adenoviral and transgene product antigens. Moreover, we observed a strong reduction in the humoral response to the vector and the transgene product, with a drastic reduction of anti-adenovirus immunoglobulin A and G antibody isotypes. In addition, the reduction in antibody response observed in TNF-alpha/LTalpha-/- and TNF-alpha/LTalpha+/- mice versus TNF-alpha/LTalpha+/+ mice links antibody levels to TNF-alpha/LTalpha gene dosage. Due to the absence of neutralizing antibodies, the TNF-alpha/LTalpha knockout mice successfully express a second gene transduced by a second vector injection. The discovery of the pivotal role played by TNF-alpha in controlling the antibody response against adenovirus will allow more efficient adenovirus-based strategies for gene therapy to be proposed.

Control of nitric oxide production by endogenous TNF-alpha in mouse retinal pigmented epithelial and Muller glial cells.

Since the induction of nitric oxide synthase (NOS) by lipopolysaccharide (LPS) has been suggested to be partially dependent of the synthesis of tumor necrosis factor alpha (TNF alpha), we have investigated in vitro the production of NO in retinal cells from mice deficient in Lymphotoxin alpha (LT alpha)/TNF alpha. Treatment of retinal Müller glial (RMG) and retinal pigmented epithelial (RPE) cells from both wild-type and knockout mice with LPS and interferon gamma (IFN gamma) induced NO synthesis as determined by nitrite release into the media and was correlated to an increase in NOS-2 mRNA levels, evaluated by RT-PCR. However, the level of nitrite and the accumulation of mRNA was always less in cells from LT alpha/TNF alpha knockout mice than in wild type mice. Simultaneous addition of TNF alpha restored the level of NO synthesis by RMG and RPE cells from LT alpha/TNF alpha knockout mice stimulated with LPS and IFN gamma to wild type levels. Transforming growth factor beta (TGF beta) blocked LPS/IFN gamma-induced NO production is RMG and RPE cells from wild-type and LT alpha/TNF alpha knockout mice. Our results demonstrate that induction of NO synthesis in RMG and RPE cells by LPS and IFN gamma is dependent in part on endogenous TNF alpha while inhibition of NO production by TGF beta does not require a modulation of TNF alpha synthesis.

Mice heterozygous for a deletion of the tumor necrosis factor-alpha and lymphotoxin-alpha genes: biological importance of a nonlinear response of tumor necrosis factor-alpha to gene dosage.

The tumor necrosis factors (TNF-alpha and lymphotoxin, or LT-alpha) are important mediators of the immune and inflammatory responses, and it has been proposed that a positive feedback loop could boost the expression of the TNF to sufficiently high levels to fend off infections. To investigate this phenomenon and its biological consequences, we have generated LT-alpha/TNF-alpha knockout mice and compared mice having one or two functional LT-alpha/TNF-alpha alleles. In response to lipopolysaccharide (LPS) stimulation, TNF-alpha levels in the circulation or in the supernatant of macrophage cultures were 20- to 100-fold lower in heterozygous samples than in their wild-type counterparts. This differential increased with the intensity of stimulation and throughout the response, supporting the involvement of a positive feedback loop. Moreover, the heterozygous mice had an increased bacterial load following Listeria monocytogenes infection and exhibited a bimodal response to the association of D-galactosamine and LPS which was similar to that of wild-type mice at low doses of LPS and more like that of homozygous mutants at high doses. These results therefore establish the biological importance of the nonlinear response of TNF-alpha levels to gene dosage, and these mice provide a unique tool to study how the propensity to produce TNF can determine the immunological fitness of individuals.

Induction of pgp3 expression and reversion of the multidrug resistance phenotype in 9-OH-ellipticine-resistant Chinese hamster lung fibroblasts transfected with the MYC oncogene.

Chinese hamster lung cells resistant to the DNA topoisomerase II inhibitor 9-OH-ellipticine (DC-3F/9-OH-E) are cross resistant to various drugs through the expression of the MDR phenotype. The myc oncogene was approximately 10-fold amplified and 20-fold overexpressed in parental DC-3F cells as compared with DC-3F/9-HO-E cells. Transfection of the resistant cells with a mouse c-myc gene did not alter the resistance to topoisomerase II inhibitors and, in cells with a low multidrug (MDR) expression, reversed this phenotype. Northern and Western blot analyses revealed an increased expression of pgp1 in the DC-3F/9-OH-E cells, which was not modified in the myc-transfected clones. However, myc expression in these clones resulted in an increased expression of pgp3, roughly in proportion to the level of myc expression. Transfection of the DC-3F/9-OH-E cells with the human MDR3 gene, homologous to pgp3, also resulted in the reversion of the MDR phenotype. These results show that (1) expression of the transfected myc gene positively regulates pgp3 expression but has no effect on pgp1; (2) when observed, reversion of the MDR phenotype is proportional to the levels of myc and pgp3 expression; and (3) this reversion, resulting from pgp3 expression, is associated with a decreased functional activity of the pgp1 protein and might require an appropriate balance of pgp1 and pgp3 expression.

Lipopolysaccharide-induced cytokine cascade and lethality in LT alpha/TNF alpha-deficient mice.

BACKGROUND: Tumor necrosis factor alpha (TNF-alpha) is often considered the main proinflammatory cytokine induced by lipopolysaccharide (LPS) and consequently the critical mediator of the lethality associated with septic shock. MATERIALS AND METHODS: We used mice carrying a deletion of both the lymphotoxin alpha (LT-alpha) and TNF-alpha genes to assess the role of TNF in the cytokine cascade and lethality induced by LPS. RESULTS: Initial production of IL-1 alpha, IL-1 beta, IL-6, and IL-10 is comparable in wild-type and mutant mice. However, at later times, expression of IL-1 alpha, IL-1 beta, and IL-10 is prolonged, whereas that of IL-6 decreases in mutant mice. Expression of IFN-gamma is almost completely abrogated in mutants, which is in agreement with a more significant alteration of the late phase of the cytokine cascade. We measured similar LD50 (600 micrograms) for the intravenous injection of LPS in mice of the three genotypes (+/+, +/-, -/-), demonstrating that the absence of TNF does not confer long-term protection from lethality. However, death occurred much more slowly in mutant mice, who were protected more efficiently from death by CNI 1493, an inhibitor of proinflammatory cytokine production, than were wild-type mice. DISCUSSION: Thus, while TNF-alpha is not required for the induction of these cytokines by LPS, it modulates the kinetics of their expression. The lethality studies simultaneously confirm a role for TNF as a mediator of early lethality and establish that, in the absence of these cytokines, other mediators take over, resulting in the absence of long-term protection from LPS toxicity.

Transcription of unr (upstream of N-ras) down-modulates N-ras expression in vivo.

The ras proteins (Harvey, Kirsten and N-ras) are key regulators of signal transduction and a perturbation of their GDP/GTP cycle is frequently observed in tumors. In mammals, N-ras constitutes with unr (upstream of N-ras) a tightly linked tandem of ubiquitously expressed genes. Although unr and N-ras appear to be involved in distinct functions, this unusual genetic organization could be important for the regulation of N-ras expression. Specifically, transcription of unr could negatively regulate that of N-ras by transcriptional interference. To investigate this possibility, we have deleted the unr promoter by homologous recombination in murine embryonic stem cells. Analysis of tissues of heterozygous mice revealed an increase in N-ras mRNA accumulation ranging between 20 and 65%, in agreement with the suppression of a transcriptional interference.

Abnormal organisation of the splenic marginal zone and the correlated leukocytosis in lymphotoxin-alpha and tumor necrosis factor alpha double deficient mice.

In this study we further characterized the phenotype at the homeostasis of mice genetically deficient in Tumor Necrosis Factor-alpha and Lymphotoxin-alpha (LT-alpha TNF-alpha -/-). As initially observed in LT-alpha -/- mice, these mice are devoid of lymph nodes and Peyer's patches, while in their spleen the white and red pulp domains are no more detectable. In the blood the leukocytosis dominated by lymphocytosis is not solely due to the absence of lymph nodes. Indeed, this abnormality was shown to be correctable by the transfer of wild type bone marrow in the absence of lymph node. We now report that the metallophilic macrophages of the marginal zone are no more detectable with an antibody reactive to sialoadhesin, a macrophage restricted transmembrane molecule known to bind myeloid and lymphoid cells. The absence of sialoadhesin within the marginal zone, a critical domain for lymphocyte trafficking towards the white pulp suggests a possible cellular basis for the observed blood leukocytosis. In addition, in the peritoneal cavity of LT-alpha TNF-alpha-/- mice, the size of the resident leukocyte population is increased. By their amplitudes these leukocytosis are similar within the blood and the peritoneal compartments.

ALL-1 interacts with unr, a protein containing multiple cold shock domains.

The ALL-1 gene is involved in human acute leukemia through chromosome translocations and fusion to partner genes, or through partial tandem duplications. ALL-1 is the human homologue of Drosophila trithorax which transregulates the homeotic genes of the Antennapedia and bithorax complexes controlling body segment identity. ALL-1 encodes a very large protein of 3968 amino acids which presumably interacts with many proteins. Here we applied yeast two hybrid screening to identify proteins interacting with the N-terminal segment of ALL-1. One protein obtained in this way was the product of the unr gene. This protein consists of multiple repeats homologous to the cold shock domain (CSD), a motif common to some bacterial and eukaryotic nucleic acids-binding proteins. The minimal region on unr required for the interaction with ALL-1 included two CSD and two intervening polypeptides. The interaction was confirmed by in vitro binding studies, and by coimmunoprecipitation from COS cells overexpressing the relevant segments of the two proteins. These results suggest that unr is involved in an interaction of ALL-1 with DNA or RNA.

Regulation of pre-mRNA processing by src.

BACKGROUND: Changes in gene expression in response to external signals provide a key mechanisms for the regulation of higher eukaryotic cell functions. The importance of transcriptional control in the response of cells to growth factors and cytokines has been extensively documented, but gene expression has also been shown to be controlled at other levels, such as the stability of mRNA in the cytoplasm, its localization and translation. By contrast to transcriptional control, little is known of the contribution of pre-mRNA nuclear processing to the regulation of gene expression, as most of our knowledge of pre-mRNA processing in vivo is indirect, being inferred from comparisons of transcription rates and levels of mRNA accumulation. RESULTS: In this study, we have used as a model the well-characterized maturation pathway of transcripts of the cytokine, tumour necrosis factor beta (TNF beta). We have used the murine TNF beta gene as a reporter for pre-mRNA processing, using a co-transfection approach to investigate whether overproduction of proteins involved in signal transduction influences the processing of TNF beta transcripts. Although transfection of both activated ras and src genes led to an increase in RNA accumulation in the nuclear and cytoplasmic compartments, as expected from their transactivation of the TNF beta expression vector, only src induced a modification of RNA processing. Comparison of several modes of src activation indicated that two distinct effects of src on pre-mRNA processing can be coupled: one involves slowing down splicing and the other allows the export of partially spliced transcripts. These effects can be observed not only on the three introns of TNF beta but also on transcripts from a beta globin expression vector. DISCUSSION: We have characterized how the processing of transcripts of TNF beta and beta globin is regulated by the signal transduction pathway that includes the Src protein, establishing that external signals have the capacity to regulate gene expression at a post-transcriptional level within the nucleus. Src seems to act on a general mechanism of splicing and/or mRNA transport, but its biologically relevant targets are likely to be restricted to genes for which either alternative processing pathways are in competition, or the kinetics of splicing is critical. This regulation could reflect a modulation by Src of the activity of components of the splicing and transport machineries, but could also involve RNA-binding proteins, which have been shown to interact with Src.

In vitro inhibition of the pim-1 protooncogene by chimeric oligodeoxyribonucleotides composed of alpha- and beta-anomeric fragments.

We show that oligodeoxyribonucleotides (oligos) composed of alpha- and beta-anomeric sections can be used as antisense compounds. An octamer has been chosen as an effector domain to form a substrate for RNaseH. This octamer is complementary to the translation start site of the pim-1 protooncogene mRNA. Chimeric alpha-beta oligos and their beta-analogs have a similar binding affinity for their target. These oligos also direct efficient RNaseH-mediated cleavage of target mRNA. Among all alpha-beta oligos studied, one with an alpha-fragment bound by its 3'-end to the 3'-end of the beta-octamer is the most resistant to nucleolytic digestion in biological media. The alpha-beta oligos have been found to inhibit in vitro translation of pim-1 RNA with specificity.

Anti-tumor necrosis factor modulates anti-CD3-triggered T cell cytokine gene expression in vivo.

De novo expression of TNF, IFN gamma, IL-3, IL-4, and IL-6 genes was initiated rapidly by treatment of mice with anti-CD3. A specific feature of this reaction was that TNF was derived exclusively from T cells. TNF was produced both as a mature soluble trimeric protein and as a 26-kD anti-TNF-reactive protein compatible with membrane-anchored TNF. Pretreatment with anti-TNF did not affect anti-CD3-triggered TNF mRNA expression in T cells. In contrast, in vivo and in vitro anti-TNF treatment upregulated anti-CD3-induced IFN gamma mRNA expression and inhibited IL-4 mRNA expression. These latter effects were not dependent on TNF neutralization: pretreatment with soluble recombinant 55-kD TNF receptor (TBPI) as an alternative TNF-neutralizing agent did not modify the anti-CD3-induced cytokine profile. These results suggest that a direct interaction between anti-TNF and T cell membrane-anchored TNF could account for the observed modulation of cytokine gene expression. The increased expression of INF gamma mRNA observed in anti-TNF-treated animals correlated with a decrease in IL-3 and IL-6 mRNA expression. Conversely, IFN gamma blockade by a neutralizing anti-IFN gamma mAb led to a substantial increase in both IL-3 and IL-6 gene expression induced by anti-CD3. Taken together, these results strongly argue for the existence, in the anti-CD3-induced cytokine cascade, of IFN gamma-dependent regulation of IL-3 production, which in turn modulates IL-6 production.

Infection with a Kirsten-retrovirus can induce a multiplicity of tumorigenic phenotypes in the interleukin-3-dependent FDC-P1 cells.

The identification of ras oncogenes in both human and animal tumors as well as in preleukemic and precancerous lesions suggests that activated ras genes participate in neoplastic development, yet the precise role of ras oncogenes in leukemogenesis is not clear. To assess the functional role of ras genes in tumorigenesis, we introduced with a retroviral vector either a wild-type (Gly-12) or a mutant (Val-12) Kirsten ras cDNA into the cells of a factor-dependent myeloid cell line, FDC-P1. FDC-P1 cells are nontumorigenic and their proliferation is dependent on either interleukin-3 (IL-3) or granulocyte-macrophage colony-stimulating factor (GM-CSF). The Ki-Val 12-infected FDC-P1 cell population is still strictly IL-3-dependent but has acquired the ability to survive up to 72 hours in the absence of growth factor and to form tumors in nude mice. These tumors are easily established into cell lines that are clonal and show a multiplicity of phenotypes with respect to their growth factor dependence. These results suggest that, in contrast with the overexpression of a normal Ki-ras, Ki-ras oncogene can efficiently promote the tumorigenic conversion of FDC-P1 cells. However, the clonality of the tumors as well as the distinct phenotypes indicates that other genetic events are required for tumorigenicity. Therefore, in FDC-P1 cells, an activated ras gene acts as a dominant oncogene through the induction of tumor progression. Finally, in this simple experimental system we observed a multiplicity of tumorigenic phenotypes which are reminiscent of those observed in patients with acute myeloid leukemia.

In vivo cooperation between introns during pre-mRNA processing.

In higher eukaryotes the large number of introns present in most genes implies that the pre-mRNA processing machinery should be efficient and accurate. Although this could be achieved at the level of each intron, an attractive alternative would be that interactions between introns improve the performance of this machinery. In this study we tested this hypothesis by comparing the processing of transcripts of the tumor necrosis factor beta gene, which differ only by their number of introns. We took advantage of the ordered splicing of the three introns present in this gene to design constructs that should generate, as primary transcripts, molecules that are normally produced by splicing. We established that the apparent splicing rate of intron 3 is increased 2.5- and 3.5-fold by the presence of one or two other introns on the primary transcript, respectively. Similarly, the apparent splicing rate of intron 2 is increased by the presence of intron 1. As these effects involve the splice sites of the upstream intron, these observations support the existence of cooperative interactions between introns during pre-mRNA processing.