Using DDRT-PCR, we compared the mRNA content of untreated and TNF-treated mouse embryonic fibroblasts (MEFs). Among differentially represented fragments, we identified and cloned a novel TNF-stimulated gene named Tsg-5. This gene, mapped to mouse chromosome 14, has three exons that can be alternatively spliced giving rise to two mRNA species, one spanning three exons and another that skips the second exon. Analysis of full-length Tsg-5 cDNA revealed a potential start codon within exon 2 encoding an ORF of 40 amino-acids. No homology with known mouse or human sequences, neither at the nucleotide nor at the amino-acid level could be found in public databases. In MEFs, Tsg-5 is induced by tumor necrossis factor-α (TNF) and IL-1β, albeit with distinct kinetics. TNF-induced Tsg-5 expression is NF-κB-dependent as it was inhibited by MG132, lactacystin, Bay 11–7083, and Bay 11–7085. Analysis of Tsg-5 expression in vivo revealed that the gene and its encoded polypeptide are constitutively expressed in the thymus and ovary, whereas, in LPS-treated mice, Tsg-5 mRNA can be detected in the spleen, lung, and brain. Our data suggest that Tsg-5 encodes a new, rare transcript, with a very tight regulation of expression and differential splicing.
Tumor necrosis factor α(TNF) is a pleiotropic, proinflammatory cytokine that plays an important role in host defenses and in the pathophysiology of chronic inflammatory disorders.1 TNF elicits a particularly broad spectrum of organismal and cellular responses, including lymphocyte and leukocyte activation, and migration, fever, acute-phase response, cell proliferation, differentiation, and apoptosis.2
In order to exert its many biological functions, TNF binds to two distinct cell surface receptors, TNF receptor superfamily 1A (TNFRSF1A) and TNF receptor superfamily 1B (TNFRSF1B)3 that are present on virtually all cell types.4 Binding of TNF trimers to these receptors leads to aggregation and trimerization of the receptor,5 followed by the activation of a signaling cascade and modulation of gene expression, mainly via the transcription factor NF-κB.6
TNF-stimulated genes are, in turn, responsible for the biological effects attributed to TNF. These genes include, for example, different cytokines and chemokines, cell adhesion molecules, acute phase proteins, and many other proteins participating in host defenses.7 In addition, TNF also activates the expression of genes encoding transcription factors, for example, interferon regulatory factor-1 (IRF-1).8 The role of IRF-1 in the regulation of TNF-induced genes remains unclear, but the fact that TNF and interferons (IFNs) can modulate the expression of a common set of genes might be related to TNF's ability to induce IRF-1.9
In order to further clarify the molecular mode of action of TNF, as well as the involvement of IRF-1 in the regulation of TNF-induced genes, we decided to search for new TNF-stimulated genes comparing, by differential display, the mRNA content of TNF-treated and control, untreated, mouse embryonic fibroblasts (MEFs) derived from wild-type (WT) or IRF-1-deficient mice (IRF-1−/−). Here we describe the identification, cloning, structural, and initial functional characterization of a new TNF-responsive gene, which we named TNF-stimulated gene-5 (Tsg-5).
Results and discussion
Identification of a new TNF-regulated gene by DDRT-PCR
With the original goal of identifying IRF-1-dependent TNF-stimulated genes, we compared, by DDRT-PCR, the mRNA content of WT and IRF-1−/− MEFs stimulated with TNF. In the course of these studies, we diverged from our original aim and concentrated instead on the characterization of a 253-bp cDNA fragment (GenBank accession number AF004564) that was more pronounced in lanes corresponding to TNF-treated, IRF-1−/− MEFs than in TNF-treated WT cells. In view of the fact that IRF-1 is known as a positively acting transcription factor,10 it appeared that the corresponding cDNA could be an interesting example of a gene whose expression is negatively regulated by IRF-1. In fact, we could find few reports in the literature that would describe IRF-1 as a negative regulator of gene expression.11,12 Moreover, this 253-bp fragment showed no homology with known mouse sequences deposited in public databases and thus could represent a new TNF-responsive gene.
To confirm the differential expression of the mRNA associated with this 253-bp cDNA fragment, we performed a series of Northern blot analyses using total RNA from either WT or IRF-1−/− MEFs and the 253-bp fragment as a probe. These experiments revealed that this cDNA fragment hybridized with an mRNA of about 570 bases, expressed preferentially in immortalized IRF-1−/− cells treated with TNF. No detectable levels of Tsg-5 mRNA could be observed in untreated IRF-1−/− cells or in WT cells that were either treated with TNF or left untreated (Figure 1, left panel). In contrast, KC mRNA, known to be upregulated by TNF,13 could be readily detected in both WT and IRF-1−/− cells stimulated with TNF, suggesting that the differential expression of Tsg-5 was not because of a lack of responsiveness to TNF in WT cells. When we repeated the Northern blot analysis using total RNA from primary fibroblast cultures of the same genotypes, both WT and mutant cells expressed comparable levels of Tsg-5 mRNA in response to TNF (Figure 1, right panel), suggesting that the observed differences in the original comparison between immortalized WT and IRF-1−/− MEFs might be a consequence of the immortalization procedure rather than the absence of IRF-1. It is noteworthy that two other immortalized or transformed cell lines, NIH3T3 and L929, did not express Tsg-5 in response to TNF, but again, KC mRNA was augmented in response to TNF (data not shown).
Cloning of full-length Tsg-5 cDNA
Since we found no homology between the 253-bp of the Tsg-5 cDNA fragment cloned by DDRT-PCR with other sequences deposited in public databases, we decided to further characterize Tsg-5 cDNA. We screened a mouse cDNA library prepared from mRNA isolated from TNF-treated, immortalized IRF-1−/− MEFs, and identified and sequenced 10 independent cDNA clones that specifically hybridized with the original Tsg-5 cDNA fragment. Two major cDNA species could be characterized, one having three exons (Figure 2, clones A and D) and one in which the second exon was alternatively spliced out (Figure 2, clones B and C). In a single clone, a cryptic donor splice site, located 12-bp upstream from the more frequently used donor splice site in exon 1, was identified (Figure 2, clone C and Figure 3). Also, a single clone with a much larger first exon, extending towards its 5′ end was identified (Figure 2, clone D). The identified exon 3 (270-bp) contains the entire 253-bp fragment originally identified by DDRT-PCR.
In order to confirm the expression of the two Tsg-5 splice variants, we designed a pair of primers that are complementary to sequences within exons 1 and 3 of the Tsg-5 gene (P1 and P2, Figure 2, upper panel), and performed RT-PCR using total RNA from TNF-treated or control, untreated IRF-1−/− MEFs (Figure 2, lower panel). Indeed, both mRNA species can be detected in TNF-treated cells after 30 cycles of amplification and, after 35 cycles, even the untreated cells have detectable levels of both forms of Tsg-5 mRNA. This event of alternative splicing was further confirmed by the sequencing of these two PCR products.
Sequence analysis of the 3-exon Tsg-5 mRNA revealed the presence of a strong Kozak consensus sequence14 surrounding an AUG codon located within exon 2. This AUG codon would give rise to an open reading frame (ORF) encoding a polypeptide of 40 amino acids (Figure 3, upper panel). The predicted polypeptide does not contain a putative signal sequence. Despite this observation, there are examples of poly-peptide molecules that, even without a signal peptide, are secreted into the extracellular space via alternative pathways such as translocation (of the plasma membrane or of intracellular membranes) or pinching off from the plasma membrane of vesicles enriched in a given leaderless secretory protein.15 Further experiments are needed to determine whether the Tsg-5 polypeptide is secreted to the extracellular space.
The alternative spliced form of Tsg-5 mRNA, lacking exon 2, has also a potential AUG start codon located in exon 1. However, this codon is embedded in a less conserved Kozak consensus sequence and it would give rise to an ORF encoding for a 34 amino-acid long polypeptide that also lacks a signal peptide. The two isoforms would share identity in the 31 amino acids encoded by exon 3 (amino acids, underlined in Figure 3). Neither of the two polypeptide isoforms share homology with protein sequences deposited in available public databases.
At the 3′ UTR, Tsg-5 gene has a 174-bp long se-quence within which we identified a polyadenylation signal sequence (AUUAAA) 147-bp downstream from the UGA terminal codon (Figure 3, underlined in nucleotide sequence). All 10 cDNA clones sequenced showed no variation at the 3′-UTR, neither in length nor in the site of cleavage and addition of the Poly(A) tail.
Characterization of Tsg-5 genomic structure and chromosomal localization
In order to complete the characterization of the Tsg-5 gene, we next determined the size of introns 1 and 2, and for this we used long-range PCR with primers complementary to the boundaries of the three Tsg-5 exons. For intron 1, we amplified a DNA fragment of about 2.3-kb whereas, for intron 2, we amplified a DNA fragment of about 9-kb. These fragments were subcloned and their extremities were sequenced in order to confirm donor and acceptor splicing sites. Acceptor and donor sites found in the exon/intron boundaries of Tsg-5 have the consensus GT-AG found in the majority of intron-containing genes.16 With the availability of the mouse genome sequence (Celera Discovery System, www.celera.com), we could confirm the exact size and sequence of the entire introns 1 and 2, as well as the sequences of the three exons. Hence, the entire Tsg-5 gene extends over a region of 11.2-kb of the mouse genome and is located on chromosome 14, cytogenetic band 14D2, between the RB1 and the serotonin receptor 2A (Htr2) genes. Tsg-5 gene is approximately 980-kb from RB1 gene and approximately 280-kb from Htr2 gene. This region is syntenic to human chromosome 13p1417 but, nevertheless, we were unable to find a sequence homologue to the mouse Tsg-5 gene. Tsg-5 sequences can be accessed at http://www.ncbi.nlm.nih.gov and the accession numbers for Tsg-5 gene and mRNAs are AY139113, AY139114, and AY139115.
Mapping the Tsg-5 transcription start site
Comparison of the cDNA sequences obtained from 10 independent clones revealed that the first exon in clone D (upper panel in Figure 2) is 234-bp long in contrast with clone A in which the first exon is only 96-bp long. The entire cDNA sequence of clone A is 425-bp long, being compatible to the Tsg-5 mRNA size estimated by Northern blot analyses (about 570 bases). In contrast, clone D cDNA is 563-nucleotide long and, considering that the average size of poly(A) tail is about 150–200 nucleotides,18 this clone represents an mRNA that exceeds the detected size for the predominant mRNA. A possible explanation for the extended 5′ end observed in exon 1 of clone D would be the usage of an alternative minimal promoter and transcriptional start site, upstream from the commonly used transcriptional start site that gave rise to the remaining nine clones. Taking this possibility into account, we hypothesized that the bona fide Tsg-5 promoter and its transcriptional start site could be located within the extended exon 1 sequence found in clone D.
Indeed, sequence analysis of the extended exon 1 of clone D revealed the presence of a potential core promoter with a TATA-Box element and one cis-element with the consensus binding site for the transcription factor NF-κB, 40-bp upstream to this putative TATA Box (Figure 4a). Owing to the very low levels of Tsg-5 mRNA, we used a RT-PCR-based approach to investigate whether the main Tsg-5 transcription start would indeed be located downstream from the TATA-Box motif. Three forward primers were designed on the basis of the sequence of exon 1 from clone D, as depicted in Figure 4b (upper panel), and were used in conjunction with the P2 antisense primer, complementary to Tsg-5 exon 3. As template, we used total RNA derived from TNF-treated IRF-1−/− MEFs reverse transcribed with oligo(dT). Analysis of the RT-PCR products revealed the amplification of a prominent band when P1 and P2 primers were used. In contrast, a faint band could be observed with primers P3 and P4. Thus, we suggest that the main transcription start site is located downstream from the indicated TATA motif.
A functional repetitive element in Tsg-5 gene
Sequence analysis of the exon 1 and the 5′ flanking genomic DNA sequence of Tsg-5 gene, obtained from Celera database (www.celera.com), revealed a strong homology with the repetitive element of the mouse transcript (MT) family (Figure 4a, underlined sequence). MT elements are 400-bp dispersed repetitive element represented between 4 × 104 and 9 × 104 times per haploid mouse genome and have the hallmarks of solitary long terminal repeats (LTRs).19,20 It was recently demonstrated that MT elements belong to a subclass of the mammalian LTR retrotransposon superfamily (MaLR). The majority of MaLR sequences in the GenBank database are found as solitary LTRs, presumably resulting from a high frequency of LTR–LTR recombination and retrotransposon-like genome excision.21
As shown in Figure 4a (underlined sequence), this repetitive element contains the entire Tsg-5 exon 1, the predicted core promoter, and the main Tsg-5 transcriptional start site plus one consensus NF-κB binding site. Additionally, within this repetitive element, we could identify three putative AP-1 binding sites, two of them located within a palindromic sequence. The immediate 5′ upstream region from this repetitive element contains another three consensus NF-κB binding sites. Together, the NF-κB and AP-1 binding sites might be involved in the transcription activation of Tsg-5 in response to TNF.
It has been suggested that repetitive sequences could move within the genome as a way of providing regulatory sequences for gene expression and, indeed, a number of repetitive sequences are frequently inserted into promoter regions and can influence gene expression.22 Hence, exon 1 and a portion of the Tsg-5 promoter is another example of a repetitive sequence that, during evolution, acquired a functional role within the mouse genome.
Tsg-5 is transcriptionally activated by TNF
Having characterized the structure of the Tsg-5 gene, we next determined the kinetics of Tsg-5 induction by TNF. Northern blot analyses using total RNA from immortalized IRF-1−/− MEFs exposed to TNF for various periods of time revealed that Tsg-5 mRNA can be detected after 2 h of TNF treatment (Figure 5a). We observed a slow accumulation of Tsg-5 mRNA and maximum induction occurred at 6 h post-treatment. Also, high levels of mRNA were maintained for at least 16 h after TNF treatment.
To determine whether Tsg-5 is transcriptionally activated in response to TNF and to see whether its delayed kinetics is because of a delayed transcription or accumulation of stable mRNA transcribed at very low levels, we measured, by RT-PCR, the levels of unprocessed Tsg-5 pre-mRNA, which can be considered as an indicator of ongoing gene transcription.23,24 Since intron sequences are rapidly spliced from the primary transcript,25,26 assessment of the relative steady state of primary transcripts provides a reliable estimate of the rate of active transcription. As can be observed in Figure 5b, RT-PCR analysis revealed that, 40 min after treatment with TNF, a discrete increase in the levels of Tsg-5 mature mRNA can be detected. Importantly, at this time point, we can detect maximum levels of the primary transcript, which is sustained for almost 16 h. Moreover, the fact that we can detect an increase in Tsg-5 pre-mRNA upon TNF treatment confirms that the Tsg-5 gene is transcriptionally activated by TNF. These data also indicate that the delayed kinetic observed by Northern blot analysis (Figure 5a) reflects the slow accumulation of Tsg-5 mRNA that is transcribed, at low levels, throughout the period of TNF treatment.
This conclusion would imply that Tsg-5 mRNA is rather stable. In order to confirm that this is indeed the case, we determined the half-life of Tsg-5 mRNA in IRF-1−/− MEFs stimulated with TNF for 7 h following blockage of transcription by actinomycin D (ActD). As shown in Figure 5c, in samples exposed to ActD, Tsg-5 pre-mRNA starts to decay 30 min after cessation of transcription and was not detectable after 60 min of treatment with ActD, confirming the efficacy of transcription blockage. In contrast, treatment with ActD had no influence in the levels of mature Tsg-5 mRNA, even after 4 h of transcription blockage. Hence, we can conclude that, whereas Tsg-5 pre-mRNA is rapidly processed, mature Tsg-5 mRNA has a rather long half-life.
Tsg-5 induction by TNF is dependent of NF-κB activation
The Rel/NF-κB transcription factor family plays an important role in TNF-induced gene activation.6 The presence of three κB sites in the 5′ region of Tsg-5 gene suggests that NF-κB may be involved in Tsg-5 regulation by TNF.
In order to investigate the role of NF-κB in Tsg-5 gene expression, we determined the levels of Tsg-5 mRNA in cells treated with TNF in the presence of various inhibitors of NF-κB activation. We used two distinct classes of inhibitors that are known to interfere with different steps of NF-κB activation. Bay 11-7083 and Bay 11-7085 are structurally related compounds that act by inhibiting TNF- induced phosphorylation of IκBα, although their precise target(s) are yet unidentified.27 The other two inhibitors, MG132 and lactacystin, prevent NF-κB activation by inhibiting proteasome activity and, consequently, blocking the degradation of ubiquitinated IκBα by the 26S proteasome.
As demonstrated in Figure 6a, pretreatment of immortalized IRF-1−/− MEFs with Bay 11-7085 led to a significant inhibition of TNF-induced Tsg-5 gene expression in a dose-dependent manner. The same results were observed using Bay 11-7083 (data not shown). These results suggest that Tsg-5 induction by TNF is dependent on TNF-induced IκBα phsphorylation. In addition, treatment with MG132 and lactacystin also led to a complete inhibition of Tsg-5 gene induction by TNF, indicating that such induction is dependent on 26S proteasome activity (Figure 6b and c).
Poppers et al28 demonstrated that, in FS-4 cells maintained in the continuous presence of TNF, IκBα is rapidly degraded but IκBα reappearance is incomplete since cells do not become desensitized to TNF signaling, and newly synthesized IκBα continues to be phosphorylated and degraded. Continued signaling by TNF in these cells, even after 15 h of continuous treatment, was demonstrated by persistent activation of the IKK complex, ongoing proteasome-mediated IκBα degradation, continued nuclear localization of p65/RelA, and the persistent presence of IκBα This observation appears to be cell specific since, in other experimental systems, restoration of normal levels of IκBα protein can be observed even in the continuous presence of TNF.29,30 In order to determine the role of NF-κB in the prolonged induction of Tsg-5 gene, we treated IRF-1−/− MEFs with TNF for 15 h and lactacystin for another 2 extra hours. As it can be observed in Figure 6d, Tsg-5 gene transcription was impaired in lactacystin-treated cells, but not in control cells treated with TNF alone. Together, data from Figure 6 indicate that, indeed, NF-κB activity is necessary for both the basal and TNF-induced Tsg-5 gene transcription.
In contrast to what was observed for TNF, Poppers et al28 demonstrated that, in FS-4 cells treated with IL-1β, IκBα returns to normal basal level after 1 h of treatment. In agreement with the notion that, in IL-1β-treated cells, NF-κB activity is rather transient, we observed that, in IRF-1−/− MEFs, induction of Tsg-5 gene expression by TNF was also transient with higher levels of mRNA being detected between 1 and 4 h after IL-1β-treatment, decaying thereafter (data not shown).
Differential expression of Tsg-5 in immortalized WT and IRF-1−/− cells
The differences in the induced expression of Tsg-5 in WT and IRF-1−/− immortalized MEFs prompted us to investigate whether such difference was because of loss of the gene during immortalization or whether it would be owing to epigenetic modifications of Tsg-5 gene such as differential cytosine methylation in Tsg-5 promoter. As demonstrated in Figure 7a, the immortalized WT cell line does have a functional Tsg-5 gene that also responds to TNF, but with significant reduction in transcription as well as in the steady-state levels of Tsg-5 mRNA, confirming the original data from the DD-RTPCR. The observation that IκBα mRNA is modulated by TNF in a comparable manner in both cell types indicates that the observed differences are not because of altered TNF signaling. It is noteworthy that, 19 h after TNF treatment, the steady-state level of Tsg-5 and IκBα mRNAs are dramatically reduced in immortalized WT MEFs, whereas it remained augmented in IRF-1−/− cells.
In vertebrates, DNA methylation, either reversibly or irreversibly, regulates genome functions affecting gene transcription and chromatin structure.31 Several lines of evidence suggest that a significant fraction of all CpG islands is prone to progressive methylation in certain abnormal cells such as cancers32 and permanent cell lines.33,34 Therefore, we considered the possibility that the differences in Tsg-5 gene expression between immortalized WT and IRF-1−/− MEFs could be a consequence of differential cytosine methylation status of Tsg-5 promoter.
In silico inspection of the Tsg-5 promoter region did not reveal the presence of CpG island, thus reducing the chance of methylation interference in Tsg-5 expression. However, several transcription factors, including AP-2, c-Myc/Myn, the AMP-dependent activator CREB, E2F, and NF-κB, recognize sequences that contain CpG residues, and binding to each has been shown to be inhibited by methylation.35 In Tsg-5 promoter, at least one CpG dinucleotide is adjacent to a potential NF-κB-binding site.
To investigate the potential impact of DNA methylation in differential expression of Tsg-5 gene observed in immortalize WT or IRF-1−/− MEFs, we used the cytosine analogue 5-Aza-2′-deoxycytidine (5-Aza-CdR), an inhibitor of DNA methyltransferase enzymes. DNA substituted with 5-Aza-CdR covalently binds DNA methyltransferase enzymes, leading to loss of activity and extensive DNA hypomethylation.36 As shown in Figure 7b, treatment of WT MEFs with 5-Aza-CdR did not lead to an augmented expression of Tsg-5 gene in response to TNF as one would expect if the observed difference between the two cells would be a consequence of hypermethylation of the Tsg-5 gene promoter in WT cells. On the contrary, mRNA levels of Tsg-5 in WT cells were further reduced in 5-Aza-CdR-tretead WT cells. In IRF-1−/− cells, treatment with 5-Aza-CdR had no detectable effect neither in the constitutive nor in the TNF-induced expression. This observation clearly rules out differential methylation of the Tsg-5 promoter as the mechanism responsible for its reduced expression in immortalized WT MEFs. Furthermore, these data could indicate that a potential repressor that would be involved in the silencing of Tsg-5 is in turn the target for hypermethylation and its function could be diminished in IRF-1−/− cells.
Tissue-specific expression of Tsg-5 gene
Finally, we determined tissue distribution of Tsg-5 mRNA in vivo. Mice were left untreated or were treated with LPS (30 μg/animal for 8 h) and the levels of Tsg-5 mRNA in various organs were determined by RT-PCR. As shown in Figure 8, Tsg-5 gene is expressed constitutively in ovary and thymus and both forms of Tsg-5 mRNA transcripts are detectable.
Induced expression of Tsg-5 gene was detected in lungs, spleen, and in the brain, but not in heart, liver, kidney, testis, and muscle. Importantly, only the three-exon form of Tsg-5 is detectable in these organs. Amplification of the HGRPT cDNA was used to monitor RNA quantity and quality in all samples. The lack of Tsg-5 gene expression in other organs cannot be attributed to an inefficient LPS induction, since Northern blot analysis of the same RNA samples revealed a clear induction of the KC gene in all organs surveyed. It is noteworthy that Tsg-5 gene expression could not be detected in several organs that were efficiently stimulated by LPS, even using an extremely sensitive methodology (Southern blot analysis of PCR products). This observation suggests that, in vivo, Tsg-5 gene expression is controlled by a very tight mechanism. Not only the expression of Tsg-5 gene appears to be tightly regulated, but the splicing event also seems to be under very restricted control. Whereas both species of mRNA were detected in a constitutive manner, LPS induced the exclusive accumulation of the 3-exon containing mRNA. Further investigation of the Tsg-5 promoter region as well as its splicing events would be necessary to clarify these mechanisms at the molecular level.
Several studies have demonstrated the constitutive expression of TNF, IL-1β, and their receptors in the thymus and ovary37,38,39,40,41,42 and we could speculate a role of these cytokines in maintaining the constitutive expression of Tsg-5 observed in these organs. Therefore, we used TNF-43 and TNFRSF1A- deficient44 mice to investigate their role in regulation of constitutive expression of Tsg-5 gene. TNF-deficient mice showed a significant reduction of Tsg-5 expression both in thymus and in ovary (Figure 9a). For TNFRSF1A-deficient mice, Tsg-5 expression was severely impaired in the ovary whereas, in thymus, we only observed a discrete, but consistent, reduction of Tsg-5 mRNA (Figure 9b). Therefore, we can speculate that in thymus, TNF-induced Tsg-5 expression could be mediated also by TNFRSF1B. In ovary, TNFRSF1A-mediated signaling appears to be critical for the constitutive expression of Tsg-5 gene. Interestingly, Tsg-5 expression in ovary of WT animals varies between different animals (Figure 9b) and such variation is not associated with a particular phase of the estrous cycle (data not shown).
In normal ovary, TNF and IL-1β were implicated in ovulation, a process that has been characterized as an inflammatory reaction.45 TNF and IL-1β have also been implicated in ovarian follicular development and atresia, steroidogenesis, and corpus luteum function (including formation, development, and regression).46 Moreover, some alterations related to ovarian functions have been described in TNFRSF1A knockout mice,47 and impaired expression of TNF-regulated genes in these animals may contribute to development of the observed phenotypes. Thus, further characterization of the biological function of Tsg-5 gene might contribute to elucidate the role of TNF system in the physiology of the ovary.
To confirm that the Tsg-5-encoded mRNA is indeed translated and gives rise to a polypeptide, we used immunohistochemistry to detect Tsg-5 protein in thymus of untreated mice and in cultures of freshly isolated TNCs. Figure 10 (panel a, upper half) shows a specific staining in thymus, which is more prominent in septum and in the medulla, with very punctual cells stained within the cortex. Specific staining was also observed in freshly isolated thymic nurse cells (TNCs) (Figure 10, panel a, lower half). This pattern of expression is compatible with the data presented in Figure 10, panel b, showing that Tsg-5 mRNA can be detected primarily in TNCs and in phagocytic cells of the thymic reticulum, but not in thymocytes. The faint band observed in the lane corresponding to thymocytes is likely because of the contamination of this cell prep since they are freshly prepared.
In conclusion, we have identified and characterized the structure of a new, not yet annotated, mouse gene, which is expressed in vivo in a very restricted and tissue-specific manner. Tsg-5 gene is expressed both constitutively as well as in response to inflammatory stimuli and, in both instances, its expression is tightly regulated both at the level of transcription and splicing. These characteristics would suggest that Tsg-5 protein has a rather restricted biological activity and further experiments could help us in deciphering its role in thymus and ovary development as well as in the inflammatory response. It is also noteworthy that, even after having the complete mouse genome sequenced, experiments in wet labs will be the only way to identify genes that are expressed at low levels or under specific conditions.
Material and methods
Cell culture and cytokine treatment
Primary MEFs were prepared from E14 embryos obtained by mating a pair of 129/Sv mice, deficient for the IRF-1 gene.9 Spontaneously immortalized MEFs prepared from WT or IRF-1−/− embryos were kindly provided by Drs Charles Weissmann and Yi-Li Yang, formerly at the University of Zurich. Cells were grown in Dullbecco's modified Eagle medium (DMEM) supplemented with antibiotics and 10% fetal bovine serum (FCS). Cells were treated with recombinant mouse TNF (R&D, Minneapolis, USA) or IL-1β (R&D, Minneapolis, USA) as described in the figure legends. For inhibition of transcription, cells were treated with actinomycin D (5 μg/ml, Sigma). The NF-κB inhibitors lactacystin, MG-132, Bay 11-7082 and Bay 11-7085 (Calbiochem), and the DNA methyltransferase inhibitor 5-Aza-2′-deoxycytidine (Sigma) were used at concentrations indicated in the figure legends. TNCs and cultures of nonepithelial phagocytic cells of the thymic reticulum (PTR) were obtained as described before.48,49,50
For in vivo induction of Tsg-5 gene, 5-week-old mice (129/Sv) were injected intraperitoneally with LPS (E. coli 0111: B4, Sigma, 30 μg/animal) in a final volume of 0.2 ml of sterile saline. Control mice received only the vehicle. Mice were killed 8 h after treatment, and indicated organs were frozen immediately in liquid nitrogen.
Total RNA from cells or frozen organs was isolated using TRIzol reagent (Invitrogen Life Technologies) according to the manufacturer's recommendations.
Differential display was performed following the protocol of Liang and Pardee51 with some modifications.52,53 Briefly, total RNA was treated with DNaseI and 200 ng of total RNA were mixed with one of the four anchored oligoT11VN primers at 2.5 μM final concentration. Total RNA and anchored oligo(dT) were heated to 70°C for 10 min followed by an ice-incubation for 2 min. Reverse transcription was carried out using SUPERSCRIPT II (Invitrogen Life Technologies) according to the manufacturer's recommendations. One-tenth (2.0 μl) of cDNA first strand reaction was used as template for PCR amplification in the presence of T11VN plus a 10-bp primer, 20 μM of each dNTP, 2.5 units of Sequencing Grade® Taq DNA polymerase (Promega), 10 μCi of Redivue [α-32P]dCTP (Amersham Pharmacia Biotech). Three random primers were used throughout this work: P13, 5′-IndexTermCTGATCCATG-3′; P14, 5′-IndexTermCTGCTCTCAA-3′; and P15, 5′-IndexTermCTTGATTGCC-3′. PCR amplification was carried out for 40 cycles (94°C for 30 s, 40°C for 2 min, and 72°C for 30 s, followed by 10 min extension at 72°C). PCR products were fractionated through a denaturing 6% polyacrylamide, 8 M urea gel. Bands of interest were excised from the gel, eluted in deionized water, and reamplified by PCR using the same conditions described for differential display except that [α-32P]dCTP was replaced by nonradioactive dCTP. Reamplified cDNA fragments were purified by Wizard PCR Preps (Promega), cloned into pUC18 (SureClone System, Pharmacia), and sequenced (ABI PrismTM, PE Applied Biosystems, USA).
A genomic DNA library from 129/Sv-Ev mouse strain54 was kindly provided by Drs U Muller and Charles Weissmann, formerly at the University of Zurich. A unidirectional cDNA library in ZAP Express vector® was prepared by Stratagene, using total RNA derived from IRF-1−/− MEFs treated with TNF for 8 h. Screening and purification of positive clones were performed as described by Ansubel et al.55 Sequence analysis of Tsg-5 gene and cDNAs clones were done in an ABI PrismTM. Introns were originally identified and sequenced by long-range PCR, using eLONGase Amplification System (Invitrogen Life Technologies) and primers complementary to the extremity of Tsg-5 exons. The following primer pairs were used: P1 (5′-IndexTermtgagttccagtcctgacttccttg-3′, specific for exon 1) and P5 (5′-IndexTermctcacggtttgaggtacagtcca-3′, specific for exon 2) for intron 1, and P6 (5′-IndexTermccactctcagagcacacctcctca-3′, specific for exon 2) and P2 (5′-IndexTermagattccctggtccttcgtt-3′, specific for exon 3) for intron 2.
After availability of the mouse genome sequence (Celera Discovery System, www.celera.com), the entire Tsg-5 gene sequence including intron/exon boundaries and the 5′UTR was confirmed in silico.
Analysis of steady-state levels of mRNA and pre-mRNA
For Northern blot analysis, 20 μg of total RNA was fractionated through a 1% denaturing agarose gel and transferred by capillarity onto Hybond N filters (Amersham Pharmacia Biotech). Prehybridization, hybridization, and washes were performed as described by Church and Gilbert.56 The Tsg-5 cDNA probe corresponds to the third exon plus 20-bp of the second intron. The IκBα cDNA probe (positions 177–625, accession number gb/U36277) was cloned by RT-PCR in our lab using the following primer pairs: 5′-IndexTermaaggacgaggagtacgagca-3′ (sense) and 5′-IndexTermagacacgtgtggccattgta-3′ (antisense). The KC chemokine probe corresponds to the full-length KC cDNA (accession number J04596) subcloned into pUC18 vector. The mouse GAPDH cDNA (positions 952–1226, accession number gb/M32599) was cloned by RT-PCR in our lab and used as control for ensuring equal RNA loading. Probes were labeled by random priming, using Redivue [α-32P]dCTP (3000 Ci/mmol). Nylon filters were exposed to Kodak Hyperfilm (Amersham Pharmacia Biotech) at −70°C, with intensifying screen.
For RT-PCR, 1 μg of total RNA was reverse transcribed with SUPERSCRIPT II and oligo(dT) (Invitrogen Life Technologies) following the protocol supplied with the reagents. One-tenth of the first-strand reaction was used as template for PCR. For Tsg-5 mRNA analysis, the following primer pairs were used: P1 exon-specific primer (5′-IndexTermtgagttccagtcctgacttccttg-3, sense) and P2 exon 3-specific primer (5′-IndexTermagattccctggtccttcgtt-3′, antisense). Two additional oligonucleotide primers specific for exon 1 were used to map the main transcriptional start site, P4 (5′-IndexTermtggtgccatctctggactggta-3′, sense), and P3 (5′-IndexTermggttctataagagagcaggctga-3′, sense). The underlined nucleotides in P3 primer correspond to the putative TATA-Box element in the 5′ flanking region in Tsg-5 gene. The P7 intron 2-specific primer (5′-IndexTermatctaaatatgtattttacctgcag-3′) was designed for Tsg-5 pre-mRNA amplification. This oligonucleotide is specific for the 3′ end of Tsg-5 intron 2, adjacent to intron 2 and exon 3 boundaries. The underlined ‘ag’ dinucleotide in primer P7 corresponds to the acceptor splice site of Tsg-5 intron 2. When indicated, hypoxanthine-guanine phosphoribosyltransferase (HGPRT) cDNA was also amplified from the same samples in order to ensure equal amounts of RNA (5′-IndexTermatcagtcaacgggggacata-3′, sense; 5′-IndexTermttgcgctcatcttaggcttt-3′, antisense). For the analyses of Tsg-5 pre-mRNA, 50 μg of total RNA was treated with 10 units of RQ1 RNase-free DNase I (Promega) in order to remove contaminating genomic DNA. PCR products were fractionated through a 1.2% agarose gel and stained by ethidum bromide or blotted onto nylon membrane for Southern blot analysis as indicated.
Generation of anti-Tsg-5 antibodies
Rabbit polyclonal antibodies were raised by Bethyl Laboratories (Montgomery, USA), by immunization with a cocktail of two polypeptides, one corresponding to positions 4–17 (IndexTermCTSNREITIALNQS) and a second corresponding to positions 23–40 (IndexTermLQQMETIHENYSLAKCRE) of the predicted polypeptide encodes by Tsg-5 gene. The peptides were separately conjugated to KLH as a carrier, using maleimide chemistry, linking sulfhydryl of peptide to the carrier and injected as immunogen.
Cryostatic sections from thymus of 3 to 4-week-old mice were fixed with acetone and used for immunohistochemistry as described elsewhere.57 In order to avoid nonspecific staining, specimens were first incubated with 10 μg/ml normal mouse Ig diluted in 1% BSA/PBS for 30 min. Specific staining with appropriate dilutions (1 : 100) of the anti-Tsg-5 antibodies was carried out for 1 h. Specimens were then washed three times in PBS and subjected to the respective peroxidase-conjugated second antibody (Amersham, Buckinghamshire, UK) for 1 h. For negative controls, we substitute specific antibodies by preimmune sera, applied at the same concentration.
Kollias G, Douni E, Kassiotis G, Kontoyiannis D . On the role of tumor necrosis factor and receptors in models of multiorgan failure, rheumatoid arthritis, multiple sclerosis and inflammatory bowel disease. Immunol Rev 1999; 169: 175–194.
Tracey KJ, Cerami A . Tumor necrosis factor, other cytokines and disease. Annu Rev Cell Biol 1993; 9: 317–343.
Lewis M, Tartaglia LA, Lee A et al. Cloning and expression of cDNAs for two distinct murine tumor necrosis factor receptors demonstrate one receptor is species specific. Proc Natl Acad Sci USA 1991; 88: 2830–2834.
Hohmann HP, Remy R, Brockhaus M, van Loon AP . Two different cell types have different major receptors for human tumor necrosis factor (TNF alpha). J Biol Chem 1989; 264: 14927–14934.
Banner DW, D'Arcy A, Janes W et al. Crystal structure of the soluble human 55 kd TNF receptor–human TNF beta complex: implications for TNF receptor activation. Cell 1993; 73: 431–445.
Baud V, Karin M . Signal transduction by tumor necrosis factor and its relatives. Trends Cell Biol 2001; 11: 372–377.
Agaarwall BB, Samanta A, Feldmann M . TNF-alfa. In: Oppenheim JD, Feldmann M (eds). Cytokine Reference. Academic Press: London, 2000, pp 413–434.
Fujita T, Reis LF, Watanabe N, Kimura Y, Taniguchi T, Vilcek J . Induction of the transcription factor IRF-1 and interferon-beta mRNAs by cytokines and activators of second-messenger pathways. Proc Natl Acad Sci USA 1989; 86: 9936–9940.
Reis LF, Ruffner H, Stark G, Aguet M, Weissmann C . Mice devoid of interferon regulatory factor 1 (IRF-1) show normal expression of type I interferon genes. EMBO J 1994; 13: 4798–4806.
Taniguchi T, Ogasawara K, Takaoka A, Tanaka N . IRF family of transcription factors as regulators of host defense. Annu Rev Immunol 2001; 19: 623–655.
Kirchhoff S, Oumard A, Nourbakhsh M, Levi BZ, Hauser H . Interplay between repressing and activating domains defines the transcriptional activity of IRF-1. Eur J Biochem 2000; 267: 6753–6761.
Manzella L, Gualdi R, Perrotti D et al. The interferon regulatory factors 1 and 2 bind to a segment of the human c-myb first intron: possible role in the regulation of c-myb expression. Exp Cell Res 2000; 256: 248–256.
Ohmori Y, Fukumoto S, Hamilton TA . Two structurally distinct kappa B sequence motifs cooperatively control LPS-induced KC gene transcription in mouse macrophages. J Immunol 1995; 155: 3593–3600.
Kozak M . An analysis of 5′-noncoding sequences from 699 vertebrate messenger RNAs. Nucleic Acids Res 1987; 15: 8125–8148.
Rubartelli A, Sitia R . Secretion of mammalian proteins that lack a signal sequence. In: Kuchler K, Rubartelli A, Holland BI (eds). Unusual Secretory Pathways: From Bacteria to Man. R.G. Landes: Austin, 1997, pp 87–104.
Mount SM . A catalogue of splice junction sequences. Nucleic Acids Res 1982; 10: 459–472.
Hsieh CL, Bowcock AM, Farrer LA et al. The serotonin receptor subtype 2 locus HTR2 is on human chromosome 13 near genes for esterase D and retinoblastoma-1 and on mouse chromosome 14. Somat Cell Mol Genet 1990; 16: 567–574.
Sheiness D, Puckett L, Darnell JE . Possible relationship of poly(A) shortening to mRNA turnover. Proc Natl Acad Sci USA 1975; 72: 1077–1081.
Heinlein UA, Lange-Sablitzky R, Schaal H, Wille W . Molecular characterization of the MT-family of dispersed middle-repetitive DNA in rodent genomes. Nucleic Acids Res 1986; 14: 6403–6416.
Bastien L, Bourgaux P . The MT family of mouse DNA is made of short interspersed repeated elements. Gene 1987; 57: 81–88.
Smit AF . Identification of a new, abundant superfamily of mammalian LTR-transposons. Nucleic Acids Res 1993; 21: 1863–1872.
Britten RJ . Cases of ancient mobile element DNA insertions that now affect gene regulation. Mol Phylogenet Evol 1996; 5: 13–17.
Swee MH, Parks WC, Pierce RA . Developmental regulation of elastin production. Expression of tropoelastin pre-mRNA persists after down-regulation of steady-state mRNA levels. J Biol Chem 1995; 270: 14899–14906.
Zhang M, Pierce RA, Wachi H, Mecham RP, Parks WC . An open reading frame element mediates posttranscriptional regulation of tropoelastin and responsiveness to transforming growth factor beta1. Mol Cell Biol 1999; 19: 7314–7326.
Bauren G, Wieslander L . Splicing of Balbiani ring 1 gene pre-mRNA occurs simultaneously with transcription. Cell 1994; 76: 183–192.
Zhang G, Taneja KL, Singer RH, Green MR . Localization of pre-mRNA splicing in mammalian nuclei. Nature 1994; 372: 809–812.
Pierce JW, Schoenleber R, Jesmok G et al. Novel inhibitors of cytokine-induced IkappaBalpha phosphorylation and endothelial cell adhesion molecule expression show anti-inflammatory effects in vivo. J Biol Chem 1997; 272: 21096–21103.
Poppers DM, Schwenger P, Vilcek J . Persistent tumor necrosis factor signaling in normal human fibroblasts prevents the complete resynthesis of I kappa B-alpha. J Biol Chem 2000; 275: 29587–29593.
Sun SC, Ganchi PA, Ballard DW, Greene WC . NF-kappa B controls expression of inhibitor I kappa B alpha: evidence for an inducible autoregulatory pathway. Science 1993; 259: 1912–1915.
de Martin R, Vanhove B, Cheng Q et al. Cytokine-inducible expression in endothelial cells of an I kappa B alpha-like gene is regulated by NF kappa B. EMBO J 1993; 12: 2773–2779.
Attwood JT, Yung RL, Richardson BC . DNA methylation and the regulation of gene transcription. Cell Mol Life Sci 2002; 59: 241–257.
Baylin SB, Herman JG . DNA hypermethylation in tumorigenesis: epigenetics joins genetics. Trends Genet 2000; 16: 168–174.
Antequera F, Boyes J, Bird A . High levels of de novo methylation and altered chromatin structure at CpG islands in cell lines. Cell 1990; 62: 503–514.
Jones PA, Wolkowicz MJ, Rideout III WM et al. De novo methylation of the MyoD1 CpG island during the establishment of immortal cell lines. Proc Natl Acad Sci USA 1990; 87: 6117–6121.
Tate PH, Bird AP . Effects of DNA methylation on DNA-binding proteins and gene expression. Curr Opin Genet Dev 1993; 3: 226–231.
Santi DV, Norment A, Garrett CE . Covalent bond formation between a DNA-cytosine methyltransferase and DNA containing 5-azacytosine. Proc Natl Acad Sci USA 1984; 81: 6993–6997.
Giroir BP, Brown T, Beutler B . Constitutive synthesis of tumor necrosis factor in the thymus. Proc Natl Acad Sci USA 1992; 89: 4864–4868.
Wolf SS, Cohen A . Expression of cytokines and their receptors by human thymocytes and thymic stromal cells. Immunology 1992; 77: 362–368.
Deman J, Van Meurs M, Claassen E, Humblet C, Boniver J, Defresne MP . In vivo expression of interleukin-1 beta (IL-1 beta), IL-2, IL-4, IL-6, tumour necrosis factor-alpha and interferon-gamma in the fetal murine thymus. Immunology 1996; 89: 152–157.
Ryffel B, Brockhaus M, Greiner B, Mihatsch MJ, Gudat F . Tumour necrosis factor receptor distribution in human lymphoid tissue. Immunology 1991; 74: 446–452.
Chen HL, Marcinkiewicz JL, Sancho-Tello M, Hunt JS, Terranova PF . Tumor necrosis factor -alpha gene expression in mouse oocytes and follicular cells. Biol Reprod 1993; 48: 707–714.
Simon C, Frances A, Piquette G, Polan ML . Immunohistochemical localization of the interleukin-1 system in the mouse ovary during follicular growth, ovulation, and luteinization. Biol Reprod 1994; 50: 449–457.
Marino MW, Dunn A, Grail D et al. Characterization of tumor necrosis factor-deficient mice. Proc Natl Acad Sci USA 1997; 94: 8093–8098.
Rothe J, Lesslauer W, Lotscher H et al. Mice lacking the tumour necrosis factor receptor 1 are resistant to TNF-mediated toxicity but highly susceptible to infection by Listeria monocytogenes. Nature 1993; 364: 798–802.
Espey LL . Current status of the hypothesis that mammalian ovulation is comparable to an inflammatory reaction. Biol Reprod 1994; 50: 233–238.
Terranova PF, Rice VM . Review: cytokine involvement in ovarian processes. Am J Reprod Immunol 1997; 37: 50–63.
Roby KF, Son DS, Terranova PF . Alterations of events related to ovarian function in tumor necrosis factor receptor type I knockout mice. Biol Reprod 1999; 61: 1616–1621.
Wekerle H, Ketelsen UP . Thymic nurse cells—Ia-bearing epithelium involved in T-lymphocyte differentiation? Nature 1980; 283: 402–404.
Houben-Defresne MP, Varlet A, Goffinet G, Boniver J . Thymic nurse cells are the first site of virus replication after inoculation of the radiation leukemia virus. Leuk Res 1982; 6: 231–241.
Papiernik M, Nabarra B, Savino W, Pontoux C, Barbey S . Thymic reticulum in mice. II. Culture and characterization of nonepithelial phagocytic cells of the thymic reticulum: their role in the syngeneic stimulation of thymic medullary lymphocytes. Eur J Immunol 1983; 13: 147–155.
Liang P, Pardee AB . Differential display of eukaryotic messenger RNA by means of the polymerase chain reaction. Science 1992; 257: 967–971.
Silva AM, Pires EG, Abrantes EF, Ferreira LR, Gazzinelli RT, Reis LF . Application of the differential display RT-PCR strategy for the identification of inflammation-related mouse genes. Braz J Med Biol Res 1999; 32: 845–852.
Silva AM, Reis LF . Sodium salicylate induces the expression of the immunophilin FKBP51 and biglycan genes and inhibits p34cdc2 mRNA both in vitro and in vivo. J Biol Chem 2000; 275: 36388–36393.
Muller U, Steinhoff U, Reis LF et al. Functional role of type I and type II interferons in antiviral defense. Science 1994; 264: 1918–1921.
Ausubel FM, Brent R, Kingston RE et al. Screening of Recombinant DNA Libraries. John Wiley & Sons, Inc.: New York, 1996.
Church GM, Gilbert W . Genomic sequencing. Proc Natl Acad Sci USA 1984; 81: 1991–1995.
Villa-Verde DM, Lagrota-Candido JM, Vannier-Santos MA, Chammas R, Brentani RR, Savino W . Extracellular matrix components of the mouse thymus microenvironment. IV. Modulation of thymic nurse cells by extracellular matrix ligands and receptors. Eur J Immunol 1994; 24: 659–664.
We thank Anna Christina de Matos Salim and Elizangela Monteiro for performing all sequences reported in this work and Drs Mike Marino and João Santana da Silva for TNF−/− and TNFRSF1A−/− mice, respectively. We also thank Dr Patricia Bozza for her criticisms during an early phase of these experiments, Carlos Ferreira, Miyuki F da Silva, and Suely Nonogaki for helping with tissue sections, Drs Ana Paula Lepique, Anamaria A Camargo, and Jan Vilcek for critically reading the manuscript, and all members of our labs for their suggestions and discussions. EFA was supported by a predoctoral fellowship from CNPq/Ministry of Science and Technology and is now a postdoctoral fellow from FAPESP. FFC is a predoctoral fellow from FAPESP. This work was supported in part by grants from FAPESP (LFLR) and CNPq, Ministry of Science and Technology.
About this article
Cite this article
Abrantes, E., Pires, E., Carvalho, A. et al. Identification, structural characterization, and tissue distribution of Tsg-5: a new TNF-stimulated gene. Genes Immun 4, 298–311 (2003). https://doi.org/10.1038/sj.gene.6363949
- differential display