JBC Avanti Polar Lipids

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Hastings, M. L.
Right arrow Articles by Munroe, S. H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hastings, M. L.
Right arrow Articles by Munroe, S. H.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

J Biol Chem, Vol. 275, Issue 15, 11507-11513, April 14, 2000


Post-transcriptional Regulation of Thyroid Hormone Receptor Expression by cis-Acting Sequences and a Naturally Occurring Antisense RNA*

Michelle L. HastingsDagger §, Hema A. IngleDagger , Mitchell A. Lazar, and Stephen H. MunroeDagger ||

From the Dagger  Department of Biology, Marquette University, Milwaukee, Wisconsin 53201 and the  Division of Endocrinology, Diabetes, and Metabolism, Departments of Medicine and Genetics, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania 19104



    ABSTRACT
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

Thyroid hormone (T3) coordinates growth, differentiation, and metabolism by binding to nuclear thyroid hormone receptors (TRs). The TRalpha gene encodes T3-activated TRalpha 1 (NR1A1a) as well as an antagonistic, non-T3-binding alternatively spliced product, TRalpha 2 (NR1A1b). Thus, the TRalpha 1/TRalpha 2 ratio is a critical determinant of T3 action. However, the mechanisms underlying this post-transcriptional regulation are unknown. We have identified a non-consensus, TRalpha 2-specific 5' splice site and conserved intronic sequences as key determinants of TRalpha mRNA processing. In addition to these cis-acting elements, a novel regulatory feature is the orphan receptor RevErbAalpha (NR1D1) gene, which is transcribed from the opposite direction at the same locus and overlaps the TRalpha 2 coding region. RevErbAalpha gene expression correlates with a high TRalpha 1/TRalpha 2 ratio in a number of tissues. Here we demonstrate that coexpression of RevErbAalpha and TRalpha regulates the TRalpha 1/TRalpha 2 ratio in intact cells. Thus, both cis- and trans-regulatory mechanisms contribute to cell-specific post-transcriptional regulation of TR gene expression and T3 action.



    INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

Thyroid hormone receptors (TRs)1 mediate the diverse physiological effects of thyroid hormone (T3) (1-3). TRs are encoded by two closely related genes, erbAalpha and erbAbeta , in all vertebrates (4). Additional receptor diversity is generated by alternative processing of the TRalpha and TRbeta pre-mRNAs. The alternatively spliced isoforms of TRalpha , TRalpha 1 (NR1A1a) (5) and TRalpha 2 (NR1A1b), are of particular interest. Although both are widely expressed, they are functionally antagonistic. Due to its variant C terminus, TRalpha 2, unlike TRalpha 1, does not bind T3 and lacks the major activation function present in other TRs (6, 7). The dominant negative activity of TRalpha 2 appears to reflect both competition for binding to TR target genes as well as altered protein-protein interactions (8-11).

Expression of the two TRalpha isoforms is highly regulated, with each mRNA expressed in a tissue-specific and developmentally regulated fashion. In some tissues TRalpha 2 represents a relatively minor fraction of the TR-related isoforms, while in other tissues such as brain, kidney, and testes, TRalpha 2 is the most abundant isoform (12, 13). A developmental increase in the T3 responsiveness of some cells correlates with a decrease in the relative expression of TRalpha 2 mRNA, suggesting that TRalpha 2 expression modulates T3 action (13). Thus, TRalpha 2 may act as a tissue-specific antagonist of T3-responsive gene activation.

Interestingly, the alternative post-transcriptional processing of the TRalpha transcript that gives rise to TRalpha 2 mRNA occurs exclusively in mammals (14-17). The mammalian TRalpha gene is also remarkable in that it partially overlaps the gene for another nuclear receptor that is encoded on the opposite DNA strand. This receptor gene, RevErbAalpha (RevErb, NR1D1), is convergently transcribed with respect to the TRalpha gene such that its 3' end overlaps sequences coding for TRalpha 2, but not TRalpha 1 (18, 19). The unusual organization of these two genes, which code for structurally related transcription factors, may have important implications for regulation. In many cells and tissues, high levels of RevErb mRNA correlate with low levels of TRalpha 2 mRNA relative to TRalpha 1 (18, 20). Furthermore, conditions that alter RevErb expression levels often result in a reciprocal change in the ratio of TRalpha 2 to TRalpha 1 (21, 22).

One possible explanation for these observations is that the complementary 3' ends of the RevErb and TRalpha 2 transcripts form duplexes that negatively regulate TRalpha 2 mRNA. In this case specific base pairing interactions between RevErb and TRalpha 2 RNA may provide a transcript-specific mechanism for regulating the level of TRalpha 1 and TRalpha 2 mRNA levels. Although antisense RNA provides a well established mechanism for regulation of gene expression in prokaryotic systems, it is poorly characterized in eukaryotes (23). Thus, the molecular mechanisms regulating alternative processing of the TRalpha mRNAs are of interest from both a physiological and mechanistic perspective.

In this study, we examine the requirements for alternative processing of TRalpha 1 and TRalpha 2 mRNA using a model erbAalpha minigene. Our results demonstrate that the activity of the TRalpha 2-specific 5' splice site (ss) is dependent on sequence elements located within the TRalpha 2-specific intron. We also find that RevErb expression results in a decrease in TRalpha 2 mRNA relative to TRalpha 1, consistent with a role for RevErb mRNA in blocking TRalpha 2 expression. These findings have important implications for the regulation and evolution of the thyroid hormone response in mammals.


    EXPERIMENTAL PROCEDURES
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

Plasmids-- TRalpha and RevErb transcripts were expressed from a 9.1-kb MscI/BamHI fragment that spans exons 7-10 of the rat TRalpha gene and exons 2-8 of the overlapping RevErb gene (18). This fragment was cloned into the HindIII and NotI sites of the pRcCMV vector (InVitrogen) to create pCMV-erbAm. A plasmid expressing TRalpha and RevErb was constructed by reversing the orientation of this insert relative to the CMV promoter and inserting the SV40 early promoter adjacent to the TRalpha gene. In this plasmid, pSVerbA/Rev(CMV), the 9.1-kb insert was cloned between the NotI and ApaI sites of pRcCMV and a 344-bp PvuII/HindII SV40 fragment replaced downstream vector sequences extending through another SV40 promoter to a BclI site. A control plasmid, pSVerbA(Delta CMV), lacking the CMV promoter, was made by deleting the NruI/NotI fragment upstream of the insert in pSVerbA/Rev(SV). The pSVerbA/Rev plasmids, with and without the CMV promoter, were modified by insertion of the complete green fluorescent protein (GFP) coding sequence as a 0.75-kb KpnI/NotI fragment upstream of RevErb to produce pSVerbA/GFP-Rev(CMV) and pSVerbA/GFP-Rev(Delta CMV). pTKerbA/Rev(RSV) was constructed by replacing the CMV promoter in pSVerbA/Rev(CMV) with the RSV promoter and the SV40 promoter with the HSV1 thymidine kinase (TK) promoter. The RSV and TK promoters in this construct were obtained as a NdeI/HindIII fragment of pRSV2 (the gift of M. T. McNally, Medical College of Wisconsin) (24), and a BglII/HindIII fragment from pXK70R (the gift of K. W. Wilcox, Medical College of Wisconsin) (25), respectively. A control plasmid, pTKerbA(Delta RSV), was made by omitting the RSV promoter.

The minigenes with TRalpha 2-specific 5'ss mutations were constructed using polymerase chain reaction. In pCMV-erbAm(Delta 5'ss), the wild-type sequence (AG/GUGACU) at the 5'ss (/) is replaced with the sequence CC/CACACA. pCMV-erbAm(5'ss+5G) and pCMV-erbAm (5'ss+6G) contain a single G residue substitution at positions 5 and 6 of the TRalpha 2-specific 5'ss, respectively. pCMV-erbAm(Delta BS) and pCMV-erbAm(Delta XS) were made by deleting a 3.5-kb Bsu36I/SspI and a 3.2-kb XbaI/SspI from the TRalpha 2 intron, respectively. pCMV-erbAm(Delta BS+344H) was constructed by inserting a 344-bp SphI/PstI fragment of lambda  phage into the deletion in pCMV-erbAm(Delta BS).

For further analysis of the splicing enhancer, plasmid pCMV-erbAm was modified by deleting sequences extending from a BstEII site 260 nt to a XbaI site in the vector. A Bsu36I site within exon 10 was eliminated by end-filling to produce pErbA, which contains a unique Bsu36I site immediately downstream of the TRalpha 1 stop codon. Expression and processing of the TRalpha mRNA from pErbA is comparable to that of pCMV-erbAm. pErbA(Delta BS) and pErbA(Delta XS) were derived from pErbA and include the same deletions as pCMV-erbAm(Delta BS) and pCMV-erbAm(Delta XS), respectively. pErbA(SE30) and pErbA(SE80) were constructed by polymerase chain reaction amplification of the pErbA minigene using primers for sequences extending exactly 160 and 210 nt downstream, respectively, of the TRalpha 2 5'ss. Sequences of plasmids and primers are available upon request.

Cell Transfection and RNA Isolation-- COS-1 and 293 cells were grown in minimal essential medium supplemented with fetal bovine serum. Cells were replated at a density of 5 × 105 cells/plate 24 h prior to transfection. Transfection of cells and isolation of RNA were carried out as described previously (20). Skeletal muscle RNA was prepared as a control from adult Harlan Sprague-Dawley rats (26).

RNA Assays-- Northern analysis and RNase protection assays were carried out as described previously (20). The TRalpha 2 poly(A) site probe was prepared from a cloned PstI/DdeI fragment yielding an RNA probe that overlaps the final 323 nt of TRalpha 2. The TRalpha 1 poly(A) site probe was made from a cloned StyI/HincII fragment, which overlapped the final 140 nt of TRalpha 1. The TRalpha 2 cDNA probe was prepared from a cloned 135-bp StuI/FseI fragment of TRalpha 2 cDNA plasmid palpha 2Delta C-Sac/stop (27) that produced a probe that overlapped 80 nt of TRalpha 1 and 134 nt of TRalpha 2 mRNA. The exon 9A probe was prepared from a cloned DdeI/EcoO109I fragment that overlapped 97 nt of intron 8 and 154 nt of exon 9. RevErb poly(A) site probes were prepared from a cloned 516-bp StyI/EcoRI fragment. All other probes have been described (20). Detailed information on each probe is available upon request. RNA assays were quantitated as by direct radioanalytic scanning of 32P-labeled blots or gels, following appropriate corrections for background and probe length and composition as described previously (20).


    RESULTS
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

Expression of Accurately Processed TRalpha Minigene Transcripts-- Formation of TRalpha 1 and TRalpha 2 reflects competition between two mutually exclusive alternative processing reactions: polyadenylation at the proximal TRalpha 1-specific poly(A) site or splicing of the TRalpha 2-specific 5'ss to exon 10 (Fig. 1A). To facilitate analysis of the requirements for alternative processing of TRalpha mRNAs, we constructed a minigene expression plasmid that includes the alternatively spliced exons and the entire region of overlap between TRalpha 2 and RevErb in addition to other TRalpha and RevErb sequences (Fig. 1A). When this minigene was transiently expressed in COS-1 cells, the resulting transcripts were alternatively spliced and polyadenylated at exactly the same sites used in the endogenous transcripts. RNase protection assays (Fig. 1B) using probes complementary to the TRalpha 2-specific 5'ss (lane 2), spliced TRalpha 2 mRNA (lane 6), and to both polyadenylation sites (lanes 8-10) confirmed that the minigene transcripts were accurately processed. The sizes of the protected fragments are identical to those observed for endogenous mRNAs isolated from skeletal muscle (Fig. 1B, lanes 1, 5, and 11). Similarly, Northern blotting revealed two transcripts (3.9 and 1.4 kb; Fig. 2B, lane 1) corresponding exactly to the expected sizes of the alternatively processed minigene transcripts. Only very low levels of endogenous transcripts were present in these cells, which were seen upon a 10-fold overexposure of lanes containing RNA from untransfected COS-1 cells (cf. Fig. 1B (lanes 3 and 7) and Fig. 2B (lane 5)). These results validate the use of the minigene as a model system for alternative processing of the TRalpha mRNAs.



View larger version (54K):
[in this window]
[in a new window]
  		Fig. 1.   Structure and expression of the endogenous TRalpha gene locus and TRalpha minigenes. A, diagram of the TRalpha /RevErb genes (top) and the pCMV-erbAm minigene (bottom). Boxes represent exons; lines, introns; hatched boxes, untranslated regions; block arrows, direction of transcription and poly(A) sites (pA); horizontal arrow, hCMV promoter; shaded box, TRalpha 1-specific sequence; stippled box, TRalpha 2-specific sequence; angled line, TRalpha 2 splicing. Dotted lines indicate the RevErb and TRalpha 2 overlap. Probes used in RNase protection assays are labeled and presented as bars beneath protected regions. Sizes of protected fragments are noted for each probe. B, RNase protection assays of endogenous and minigene TRalpha RNA. Assays were carried out with total RNA from COS cells transfected with pCMV-erbAm (T, lanes 2, 6, and 8-10), rat skeletal muscle (S, lanes 1, 5, and 11), untransfected cells (C, lanes 3 and 7) or with tRNA alone (t, lanes 4 and 12). Marker lanes (M) are MspI-cut pBR322. Probes used to assay erbAalpha RNA processing were the TRalpha 2-specific 5'ss (lanes 1-4), a TRalpha 2 cDNA probe spanning the TRalpha 2-specific 5'ss/3'ss junction (lanes 5-7) or the TRalpha 1 and TRalpha 2 poly(A) probes separate or in a mixture (lanes 8, 9, and 10-12, respectively). Protected fragments corresponding to TRalpha 1 mRNA (alpha 1) and TRalpha 2 mRNA (alpha 2) and RNAs reading through the poly(A) sites (open arrows) are indicated.



View larger version (32K):
[in this window]
[in a new window]
  		Fig. 2.   TRalpha 2 5' splice site sequence is critical for balance between TRalpha 1 and TRalpha 2 mRNA processing. A, metazoan consensus 5'ss (28), the TRalpha 1 coding sequence for mammals and non-mammals, and three separate 5'ss mutations made in pCMV-erbAm (+5G, +6G, Delta ss) are shown. Mutations are in lowercase, and non-consensus nucleotides are underlined. TRalpha consensus sequences correspond to the following data base files (GenBankTM accession nos.): human (X55068), rat (M18028), mouse (X51983), sheep (Z68308), pig (AJ005797), frog (Xenopus laevis) (M35343, M35344), chicken (Y00987), penguin (AJ002363), flounder (D16461, D16462), and zebrafish (U54796). B, Northern blot analysis of 2 µg of poly(A)+ RNA from COS cells transfected with indicated plasmids. RNA from untransfected cells was included as a control (C, lane 5). Blots were hybridized with a probe from the region common to TRalpha 1 and TRalpha 2.

Site-specific Substitutions within the TRalpha 2 5'ss Dramatically Affect Alternative Processing-- The sequence of the TRalpha 2 5'ss differs from the metazoan 5'ss consensus at the highly conserved +5 position (Fig. 2A) (28). This sequence is also conserved in non-mammalian TRalpha genes, except for the presence of a purine at the position corresponding to +6 of the TRalpha 2 5'ss (Fig. 2A). This difference may account for the lack of splicing in the non-mammalian TRalpha gene. To test this hypothesis, we constructed three 5'ss mutations in the context of the rat TRalpha minigene (Fig. 2A). The first mutation (+5G) changes the +5C residue to G, thereby creating a 5'ss that matches the consensus. A second mutation (+6G) converts position +6 of the 5'ss from a U to a G as found in non-mammalian TRalpha 1 genes. The third mutation (Delta ss) eliminates this 5'ss by changing six residues of the consensus sequence. Northern blot analysis of poly(A) mRNA from the different constructs demonstrates that the wild-type TRalpha transcripts were alternatively processed with a TRalpha 1/TRalpha 2 ratio of 3.3 (Fig. 2B, lane 1). Strikingly, the 5'ss consensus mutation (+5G) results in a complete switch to TRalpha 2 splicing, with no detectable use of the TRalpha 1 polyadenylation site (Fig. 2B, lane 2). The apparent increased expression of TRalpha 2 (and some endogenously expressed RNAs) in lane 2 resulted from sample loading variation and is not reproducible. However, this overloading serves to demonstrate the complete absence of the TRalpha 1 minigene transcript in the +5G mutation. The single-nucleotide difference between the mammalian and non-mammalian sequences is also important for splice site activity, as was apparent from the lack of splicing of the +6G mutant, which increased the ratio of TRalpha 1/TRalpha 2 mRNA about 7-fold (Fig. 2B, lane 3). Remarkably, this single base change reduced splicing nearly to the level of the Delta ss mutant, which has a completely crippled splice site and expresses only TRalpha 1 (Fig. 2B, lane 4). The switch to TRalpha 1 processing and absence of cryptic splicing with the +6G and Delta ss mutations, and the switch to TRalpha 2 expression with the +5G mutation, demonstrate that the suboptimal TRalpha 2 5'ss allows both TRalpha 1 and TRalpha 2 to be derived from the primary transcript. Taken together, these results demonstrate that the strength of the TRalpha 2 5'ss is critical for determining the balance between TRalpha 1 and TRalpha 2 mRNA processing.

Identification of a Splicing Enhancer within the TRalpha 2 Intron-- The suboptimal 5'ss in TRalpha 2 mRNA presumably allows other factors to influence the balance between TRalpha 2 alternative splicing and TRalpha 1 polyadenylation. We considered that the activity of the TRalpha 1 poly(A) site may be important in determining TRalpha alternative processing. However, deletion of the TRalpha 1 poly(A) site produced only a modest enhancement of TRalpha 2 mRNA splicing (data not shown). In addition, decreasing the length of the TRalpha 2-specific intron by removing more than 60% of the central portion of this intron had little effect on TRalpha alternative processing (data not shown).

We hypothesized that intronic sequences may regulate TRalpha 2 mRNA processing. To test this, a deletion beginning 130 nt downstream of the TRalpha 2 5'ss and ending 60 nt upstream of the TRalpha 2 3'ss was constructed. The resulting minigene has a drastically shortened TRalpha 2 intron of about 200 nt (Delta BS, Fig. 3A). By shortening the intron and removing the competing TRalpha 1 polyadenylation site, this mutation was expected to enhance TRalpha 2 splicing efficiency. Surprisingly, the deletion caused a striking decrease in splicing (Fig. 3B, cf. lanes 2 and 4). Northern blot analysis and RNase protection assays confirm that the major product of this deletion is a transcript that is unspliced at the TRalpha 2-specific 5'ss and extending to the TRalpha 2 poly(A) site (results not shown).



View larger version (34K):
[in this window]
[in a new window]
  		Fig. 3.   Sequences within the TRalpha 2 intron enhance splicing. A, diagram of minigenes with intron deletions. Dotted lines indicate deletions starting at XbaI (X), Bsu36I (B), and SspI (S) sites. Numbers show position of these sites relative to the alpha 2 5'ss. Black box represents Bsu36I/XbaI fragment. Hatched box is heterologous sequence. Other symbols are as detailed in Fig. 1. B, RNase protection assays of total RNA from transfected COS cells using the 5'ss probe. Fragments corresponding to mRNAs unspliced and spliced at the TRalpha 2-specific 5'ss are labeled u and s, respectively, since deleted minigenes lack the TRalpha 1 poly(A) site. C, quantitation of RNase protection assays is shown as the percentage of the total TRalpha RNA that is spliced. D, RNase protection assays with the alpha 2 5'ss probe of RNA from COS cells transfected with minigenes containing the indicated deletions (lanes 1-4) and the 5'ss consensus mutation +5G (Fig. 2) alone (lanes 5 and 6) or in combination with Delta BS (lanes 7 and 8).

The loss of splicing observed with Delta BS indicates either that specific sequences within the deleted intron are required for TRalpha 2 splicing or that the intron is shorter than a nonspecific minimal length. To test these possibilities, 344 nt of heterologous sequence was inserted into the Delta BS deletion (Delta BS+344H, Fig. 3A). This nonspecific lengthening of the intron did not improve the efficiency of splicing (Fig. 3B, lanes 7 and 8). In contrast, adding back 250 nt of the TRalpha 2 intron to the 5' end of the Delta BS deletion in the Delta XS deletion (Fig. 3B, lanes 5 and 6) restored splicing efficiency to wild-type levels as shown quantitatively in Fig. 3C. These results provide evidence for a splicing enhancer sequence within a 250-nt region located 130-380 nt downstream of the TRalpha 2 5'ss.

The TRalpha 2 splicing enhancer sequence was required for splicing at the inherently weak TRalpha 2-specific 5'ss. However, the enhancer was not required when the TRalpha 2 5'ss was mutated to the consensus (+5G) 5'ss sequence (Fig. 3D, cf. lanes 4 and 8). This suggests that the function of the TRalpha 2 intronic splicing enhancer is to regulate the suboptimal splicing characteristic of the wild-type TRalpha 2 5'ss.

Mapping of a Highly Conserved Intronic Splicing Enhancer-- As a first step toward more precisely mapping sequences required for enhancer activity, a phylogenetic comparison of sequences from this region of the TRalpha 2 gene in mammals and non-mammalian vertebrates was carried out. The region analyzed includes coding and non-coding sequences from the first 500 nt of exon 9. The coding regions common to TRalpha 1 and TRalpha 2 and specific to TRalpha 1 are very highly conserved among mammals with 96-98% identity (Fig. 4A). These sequences also are highly conserved among other vertebrates, with 80-88% identity between rat and frog. The conservation of these coding sequences reflects the highly conserved structure of the C-terminal portion of TRalpha 1. It is noteworthy that a short, 80-nt segment (dark shading, Fig. 4A) immediately downstream of the TRalpha 1 stop codon, and within the 250-nt Bsu36I/XbaI splicing enhancer fragment, is also conserved among mammals at a level (95%) comparable to that of the coding sequences immediately upstream. This high level of similarity drops off sharply (70-72% identity) further downstream of the stop codon, as seen in a comparison of three additional segments of identical length. Thus, the first 80 nt of the TRalpha 1 3'-untranslated region are conserved at a level similar to the region coding for the highly conserved C-terminal portion of TRalpha 1.



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 4.   Mapping of the intronic splicing enhancer for TRalpha 2 mRNA. A, representation of sequences within the 5' end of exon 9. The TRalpha 2 5'ss and TRalpha 1 stop codon, UAA, are noted. The region of exon 9 common to TRalpha 1 and TRalpha 2 along with downstream sequences specific to the TRalpha 1 exon and TRalpha 2 intron were compared in human, rat, sheep, and frog. Sequences were compared using GCG programs PILEUP and GAP within segments of indicated sizes. A highly conserved 80-nt region is darkly shaded, TRalpha 1-specific exonic sequences are small boxes, and untranslated regions are hatched. The percentage identity of each segment is shown in bold beneath the region compared. GenBankTM accession nos. are as follows: human, X55068; rat, M18028; sheep, Z68308; frog (X. laevis) X17385. B, schematic representation of TRalpha minigenes derived from pErbA. Sequences downstream of the Bsu36I site that were added back to the Delta BS intron deletion are indicated by +80 and +30. Positions of sequences relative to the alpha 2-specific 5'ss are noted. C, RNase protection assay (top) using the 5'ss probe shown in Fig. 1. Quantitation of splicing as determined by RNase protection assays is shown graphically (bottom) and represents the average of four to eight experiments with the exception of Delta BS+30, which represents a single experiment. D, RNase protection assay using the exon 9A probe, which extends into intron 8. The top band (unmarked) represents undigested probe. I8/I9 indicates unspliced transcripts containing both intron 8 and intron 9. Other protected fragments are labeled as in Fig. 3B.

In contrast to the conservation of coding sequence among all vertebrates, the non-coding sequences in non-mammalian vertebrates have essentially no similarity to the mammalian genes: only 29-39% identity in gapped alignments of 80-nt segments (Fig. 4A). The very high conservation observed among mammalian sequence within the first 80 nt of the 3'-untranslated region of TRalpha 1 mRNA strongly suggests that this region has a specific function, such as splicing, that is conserved in mammals but not in non-mammalian vertebrates. To test the role of the conserved 80-nt region in splicing, this sequence was inserted into the Delta BS deletion and assayed for its ability to promote splicing of TRalpha 2 (Fig. 4B). Splicing of this transcript, SE80, was compared with that of identical constructs that contain the Delta BS and Delta XS sequences previously described. The addition of 80 nt to the 5' end of Delta BS (positions 130-210 in the TRalpha 2-specific intron) results in splicing levels slightly higher than that seen in Delta XS or the wild-type TRalpha minigenes (Fig. 4B, lanes 5 and 6). This result indicates that SE80 includes all of the sequences specific to Delta XS that are needed to enhance TRalpha 2 splicing. We refer to this 80-nt sequence as the TRalpha 2 splicing enhancer, SEalpha 2.

To precisely map SEalpha 2, a 30-nt segment at the 5' end of the conserved 80 nt region was also tested for enhancer activity (Fig. 4B). This transcript, SE30, displays nearly the same reduced level of splicing seen in Delta BS (Fig. 4C, lanes 7 and 8). These results demonstrate that 50 nt located between positions 160 and 210 of the TRalpha 2-specific intron are required for efficient splicing. To determine if SEalpha 2 affects splicing of intron 8, as well as the TRalpha 2-specific intron 9, splicing was assayed with a probe, spanning both the 3'ss and 5'ss of exon 9A. The results show that intron 8 was spliced out of approximately 95% of the transcripts from Delta BS, Delta XS, and SE80, a level similar to that observed for the wild-type transcript (Fig. 4D). Thus, the presence or absence of SEalpha 2 downstream of exon 9A has no effect on the activity of the upstream 3'ss, while dramatically altering the activity of the 5'ss for exon 9A.

RevErb Gene Expression Regulates the Processing of TRalpha Pre-mRNA-- The structure of the TRalpha locus is remarkable in that the final exon of TRalpha 2 overlaps the 3' end of RevErb transcripts encoded on the opposite DNA strand (Fig. 1A). We have shown previously that antisense transcripts derived from RevErb sequences regulate TRalpha splicing in vitro (29). However, such regulation has not been shown in vivo. To investigate the effect that the naturally occurring RevErb antisense RNA has on alternative processing of TRalpha mRNAs, we expressed RevErb from the opposite strand of the same insert used to express TRalpha mRNAs. RevErb minigene mRNA, beginning with exon 2, was expressed at high levels (Fig. 5, A and B, lane 4). This mRNA is correctly spliced and polyadenylated, as indicated by Northern blotting and RNase protection assays carried out at selected splice junctions (Fig. 5, B and C, and data not shown). Importantly, the 3' end of the transcript mapped to a site that precisely matched that of endogenous RevErb mRNA from skeletal muscle (data not shown). Thus, the 3' end of the rat RevErb minigene transcript, like the endogenous RNA, overlaps only TRalpha 2 and not TRalpha 1 (18, 30).



View larger version (43K):
[in this window]
[in a new window]
  		Fig. 5.   RevErb expression results in lower level of TRalpha 2 mRNA relative to TRalpha 1. A, diagram of TRalpha /RevErb minigene expression system. Filled arrows indicate convergent promoters, and black bars show probes for Northern blots. B, Northern analysis (left) of cytoplasmic poly(A)+ mRNA from 293 cells transfected with pSVerbA/Rev(CMV) (+R, lanes 1 and 4) or pSVerbA(Delta CMV) (-R, lanes 2 and 5), or untransfected cells (C, lanes 3 and 6) using probes specific for TRalpha (lanes 1-3) or RevErb mRNA (lanes 4-6). Quantitation of results (right) from two separate experiments is expressed as the ratio of alpha 1/alpha 2 in the presence (+R) and absence (-R) of RevErb. C, RNase protection assays of total RNA from 293 cells transfected with pSVerbA/Rev(CMV) (e/R, +R, lanes 1 and 2), pSVerbA(Delta CMV) (e/R, -R, lanes 3 and 4), pSVerbA/GFP-Rev(CMV) (e/R(G), +R, lanes 5 and 6), pSVerbA/GFP-Rev(Delta CMV) (e/R(G), -R, lanes 7 and 8) or untransfected cells (C, lane 9). TRalpha and RevErb mRNA was detected with the alpha 2 cDNA probe (upper) and RevErb probe (lower), respectively. Graphical representation of data at right represents results of assays of two experiments.

TRalpha mRNA levels in 293 cells transfected with the minigene expressing both the TRalpha and RevErb genes (+R) were compared with those in cells expressing only TRalpha (-R, Fig. 5A). A probe hybridizing to the common regions of TRalpha 1 and TRalpha 2 mRNA was used for Northern analysis of poly(A)+ RNA (Fig. 5, A and B) to allow the TRalpha 1/TRalpha 2 ratio to be determined accurately. A comparison of TRalpha expression with and without RevErb coexpression showed a striking 2.1-fold increase in the ratio of TRalpha 1/TRalpha 2 in the presence of RevErb relative to mRNA levels in its absence (Fig. 5B). These results were confirmed by RNase protection assays of total RNA, using the TRalpha 2 cDNA probe shown in Fig. 1 (Fig. 5C). Using this probe, coexpression of RevErb resulted in 2.8 ± 0.4-fold increase in the TRalpha 1/TRalpha 2 ratio (Fig. 5C). The ratio of RevErb/TRalpha 2 in these experiments was 34 ± 5 (Fig. 5C), which is similar to ratios observed in cells expressing high levels of endogenous RevErb mRNA (20). The results were not due to promoter-related artifacts because similar results were obtained with a minigene employing the RSV early promoter to express RevErb and the herpes simplex virus TK promoter to drive TRalpha expression (data not shown).

We also considered the possibility that the shift in ratio of TRalpha 1/TRalpha 2 observed upon expression of RevErb reflects direct or indirect effects of expressing the RevErb receptor protein, rather than the overlapping RevErb transcript. Although RevErb translation normally initiates at a start codon in exon 1, which is missing in the minigene, it is possible that the processed RevErb transcript is translated from an in-frame start codon in exon 2 (18). To eliminate possible translation of the RevErb reading frame, cells were transfected with plasmids that have the complete coding sequence for GFP terminated by a stop codon placed upstream of the RevErb minigene. This construct, pSVerbA/GFP-Rev(CMV), produced a GFP-RevErb mRNA transcript that efficiently expressed GFP (results not shown). Expression of this plasmid also showed a very similar shift in TRalpha 1/TRalpha 2 ratio as compared with the same construct lacking the CMV promoter (Fig. 5C). Since translation of any portion of the RevErb coding sequence downstream of the stop codon following GFP is unlikely (31), transcription but not translation of the RevErb sequences is required for an increase in the TRalpha 1/TRalpha 2 ratio. Taken together, these results demonstrate that overexpression of RevErb antisense RNA in vivo is sufficient to alter expression of the alternatively processed TRalpha 1 and TRalpha 2 mRNAs.


    DISCUSSION
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

We have identified several features of the TRalpha gene that contribute to the tissue-specific and developmentally regulated alternative mRNA processing required for expression of functionally antagonistic TRalpha proteins in mammals. These features, including a suboptimal 5'ss, a cis-acting splicing enhancer, and a naturally occurring antisense transcript, are absent in non-mammalian vertebrates, which do not express TRalpha 2. The suboptimal nature of the splice site allows for the regulation of TRalpha 1 and TRalpha 2 mRNA expression by both cis and trans mechanisms. Results of our analysis provide insight into the evolution of this complex locus as well as requirements for its regulation.

We have identified a sequence-specific TRalpha 2 intronic splicing enhancer, SEalpha 2, located in a region 130-210 nt from the TRalpha 2 5'ss, and immediately downstream of the TRalpha 1 stop codon (Fig. 4). A sequence of just 80 nt, but not 30 nt, within this region stimulates TRalpha 2 splicing 4-fold. SEalpha 2 is also required for efficient splicing of pre-mRNA in vitro (29).2 SEalpha 2 is unusual in that it represents an intronic splicing enhancer that influences differential splicing and polyadenylation. Sequence analysis of SEalpha 2 reveals that the element is very purine-rich (73% A+G). In this respect it resembles purine-rich exonic splicing enhancers (ESEs) (32-34). ESEs generally function by binding members of the arginine-serine-rich family of splicing factors (SR proteins) and activating a weak upstream 3'ss (34). SEalpha 2 differs from most other purine-rich splicing enhancers in that it is located directly downstream of an alternative 5'ss rather than a 3'ss. In two other instances where purine-rich splicing regulatory elements have been identified within introns, the sequences inhibit rather than enhance splicing (24, 35). These elements appear to act as a decoy by directing the partial assembly of competing but non-functional spliceosomes. However, examples of purine-rich intronic splicing enhancers have also been identified, suggesting that purine-rich sequences within an intron can regulate splicing in a positive as well as negative manner (36, 37).

The location of SEalpha 2 between two alternative processing sites also resembles in some respects a purine-rich splicing enhancer located between two alternative 5'ss of exon 5 in the caldesmon gene (38, 39). However, the caldesmon splicing enhancer appears to function as an ESE, whereas several results show that SEalpha 2 functions as an intronic element. Unlike ESEs, SEalpha 2 does not affect the activity of the upstream 3'ss of intron 8 (Fig. 4D). SEalpha 2 is also active in vitro in transcripts lacking an upstream 3'ss (results not shown). Finally, SEalpha 2 functions independently of a competing TRalpha 1 polyadenylation signal, thus indicating that its location within an exon is not a requirement for activity. These results suggest that SEalpha 2 represents an unusual type of purine-rich enhancer, which promotes splicing from an intronic location. Further analysis of such elements is needed to determine whether they promote splicing by a mechanism distinct from that of exonic elements.

The evolution of the TRalpha 2 5'ss within a highly conserved region of the TRalpha gene provides a unique opportunity to assess the process by which a novel regulatory function evolves by establishment of an alternative processing pathway. The co-evolution of the overlapping TRalpha and RevErb genes defines a unique series of events. Whereas a RevErb homolog has been identified in insects (40), the expression of TRalpha 1 is restricted to vertebrates, and TRalpha 2 to mammals. This evolutionary evidence suggests that RevErb arose first, followed by TRalpha (41). Since the C-terminal end of RevErb is homologous to that of many other nuclear receptor proteins, whereas that of the overlapping TRalpha 2 is unique, it seems clear that the evolution of TRalpha 2 alternative splicing was the final step in the evolution of this locus (41). Our analysis extends the insight into this evolutionary process. Once the RevErb and the TRalpha genes were juxtaposed, formation of an alternative 5'ss required only a single base change (+6T) within TRalpha . Although this change is sufficient for a low level of splicing (Fig. 2), the activity of this non-consensus site may have been augmented through the gradual evolution of SEalpha 2. The intronic region of TRalpha 2 may provide the only available site for positioning a splicing enhancer (Fig. 4A), as the flanking coding regions in exons 9A and 10 are very highly conserved (98-99% identity in humans and rats).

Both TRalpha and RevErb mRNAs encode nuclear receptor proteins of the NR1 type (5). The similar structure of these closely linked genes raises the possibility that the genetic organization reflects requirements for coordinate regulation such as that observed for many other loci where tight linkage between similar, but functionally distinct genes is highly conserved. Interestingly, both RevErb and TRalpha 2 are inhibitors of gene transcription, RevErb as an active repressor and TRalpha 2 primarily as an inhibitor of thyroid hormone-stimulated transcription (42-44).

The present study provides a direct demonstration that RevErb expression alters the balance between TRalpha 1 and TRalpha 2 mRNA processing in vivo. The 2-3-fold effect of RevErb expression on TRalpha mRNA levels with the minigene system (Fig. 5) agrees with what we have observed upon increasing endogenous RevErb expression by pharmacological means (22, 30). The increase in the ratio of TRalpha 1/TRalpha 2 mRNAs with increasing levels of RevErb relative to TRalpha 2 mRNA is also similar to that observed in a series of B lymphocyte-derived cell lines (20) as well as other tissues (18). However, in the minigene system, the change in TRalpha 1 and TRalpha 2 mRNA with increasing RevErb seems restricted to a range that is somewhat narrower than that observed in vivo. The identification of cis-acting elements that influence TRalpha 1 and TRalpha 2 processing may indicate that balanced expression is also determined by regulatory proteins binding to SEalpha 2. Regulation of TRalpha 1 and TRalpha 2 at multiple levels likely explains why the relationship between RevErb and TRalpha mRNAs does not hold in all instances (45).

RevErb expression could increase the TRalpha 1/TRalpha 2 ratio by a number of mechanisms, including selective interference with exon 10 transcription, inhibition of TRalpha 2 processing, or destabilization of TRalpha 2 mRNA. However, the kinetics of TRalpha transcription and mRNA turnover upon induction of RevErb expression are consistent with a block in TRalpha 2 processing as opposed to a change in TRalpha 2 turnover or transcription (21, 22). Furthermore, RevErb antisense RNAs complementary to exon 10 efficiently and specifically block alpha 2 splicing in vitro (29). The simplest explanation for these results and those described here is that RevErb interferes selectively with the splicing process that generates TRalpha 2.

The selective decrease in TRalpha 2 mRNA levels in response to RevErb expression has implications for regulation of T3 action. Because TRalpha 2 inhibits transcriptional activation by TRalpha 1 in a dose-dependent manner (15, 16), any change in the ratio of these regulatory proteins is expected to alter T3 action. The physiological significance of a 2-3-fold change in the expression of these receptors is supported by studies demonstrating that tissues with a higher TRalpha 1/TRalpha 2 mRNA ratio are more responsive to T3 than tissues with a 2-fold lower ratio (46). Such an effect would be especially important in tissues such as the heart that depend specifically on TRalpha for the T3 response (47).


    ACKNOWLEDGEMENTS

We thank Lori Vigue, Leslie Snell, Richard Baker, Catherine Wickland, and Heather Shoven for skillful assistance; Kent Wilcox and Mark McNally for providing recombinant plasmids; and Mark McNally for helpful comments on this manuscript.


    FOOTNOTES

* This work was supported in part by National Institutes of Health Grants DK 48034 and GM 55922 (to S. H. M.) and DK45586 (to M. A. L.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

§ Supported by a predoctoral fellowship from the Arthur J. Schmitt Foundation. Current address: Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 11724.

|| To whom correspondence should be addressed: Dept. of Biology, Wehr Life Sciences Bldg., Marquette University, P. O. Box 1881, Milwaukee, WI 53201. Tel.: 414-288-1485; Fax: 414-288-7357; E-mail: stephen.munroe@marquette.edu.

2 M. L. Hastings and S. H. Munroe, unpublished results.


    ABBREVIATIONS

The abbreviations used are: TR, thyroid hormone receptor; T3, thyroid hormone; RevErb, RevErbAalpha ; ss, splice site; nt, nucleotide(s); ESE, exonic splicing enhancer; TK, thymidine kinase; GFP, green fluorescent protein; bp, base pair(s); kb, kilobase pair(s); CMV, cytomegalovirus; RSV, Rous sarcoma virus.


    REFERENCES
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES


1. Brent, G. A., Williams, G. R., Harney, J. W., Forman, B. M., Samuels, H. H., Moore, D. D., and Larsen, P. R. (1991) Mol. Endocrinol. 5, 542-548[Abstract]
2. Lazar, M. A. (1993) Endocr. Rev. 14, 184-194[CrossRef][Medline] [Order article via Infotrieve]
3. Oppenheimer, J. H., Schwartz, H. L., and Strait, K. A. (1994) Eur. J. Endocrinol. 130, 15-24[Medline] [Order article via Infotrieve]
4. Evans, R. M. (1988) Science 240, 889-895[Abstract/Free Full Text]
5. Nuclear Receptors Committee. (1999) Cell 97, 161-163[CrossRef][Medline] [Order article via Infotrieve]
6. Koenig, R. J., Lazar, M. A., Hodin, R. A., Brent, G. A., Larsen, P. R., Chin, W. W., and Moore, D. D. (1989) Nature 337, 659-661[CrossRef][Medline] [Order article via Infotrieve]
7. Lazar, M. A., Hodin, R. A., and Chin, W. W. (1989) Proc. Natl. Acad. Sci. U. S. A. 86, 7771-7774[Abstract/Free Full Text]
8. Katz, D., and Lazar, M. A. (1993) J. Biol. Chem. 268, 20904-20910[Abstract/Free Full Text]
9. Liu, R. T., Suzuki, S., Miyamoto, T., Takeda, T., Ozata, M., and DeGroot, L. J. (1995) Mol. Endocrinol. 9, 86-95[Abstract]
10. Yang, Y. Z., Burgos-Trinidad, M., Wu, Y., and Koenig, R. J. (1996) J. Biol. Chem. 271, 28235-28242[Abstract/Free Full Text]
11. Tagami, T., Kopp, P., Johnson, W., Arseven, O. K., and Jameson, J. L. (1998) Endocrinology 139, 2535-2544[Abstract/Free Full Text]
12. Strait, K. A., Schwartz, H. L., Perez-Castillo, A., and Oppenheimer, J. H. (1990) J. Biol. Chem. 265, 10514-10521[Abstract/Free Full Text]
13. Hodin, R. A., Meng, S., and Chamberlain, S. M. (1994) Endocrinology 135, 564-568[Abstract]
14. Izumo, S., and Mahdavi, V. (1988) Nature 334, 539-542[CrossRef][Medline] [Order article via Infotrieve]
15. Koenig, R., Warne, R. I., Brent, G. A., Harney, J. W., Larsen, P. R., and Moore, D. D. (1988) Proc. Natl. Acad. Sci. U. S. A. 85, 5031-5035[Abstract/Free Full Text]
16. Lazar, M. A., Hodin, R. A., Darling, D. S., and Chin, W. W. (1988) Mol. Endocrinol. 2, 893-901[Abstract]
17. Mitsuhashi, T. G., Tennyson, G. E., and Nikodem, V. M. (1988) Proc. Natl. Acad. Sci. U. S. A. 85, 5804-5808[Abstract/Free Full Text]
18. Lazar, M. A., Hodin, R. A., Darling, D. S., and Chin, W. W. (1989) Mol. Cell. Biol. 9, 1128-1136[Abstract/Free Full Text]
19. Miyajima, N., Horiuchi, R., Shibuya, Y., Fukushige, S., Matsubara, K., Toyoshima, K., and Yamamoto, T. (1989) Cell 57, 31-39[CrossRef][Medline] [Order article via Infotrieve]
20. Hastings, M. L., Milcarek, C., Martincic, K., Peterson, M. L., and Munroe, S. H. (1997) Nucleic Acids Res. 25, 4296-4300[Abstract/Free Full Text]
21. Lazar, M. A., Hodin, R. A., Cardona, G., and Chin, W. W. (1990) J. Biol. Chem. 265, 12859-12863[Abstract/Free Full Text]
22. Chawla, A., and Lazar, M. A. (1993) J. Biol. Chem. 268, 16265-16269[Abstract/Free Full Text]
23. Kumar, M., and Carmichael, G. G. (1998) Microbiol. Mol. Biol. Rev. 62, 1415-1434[Abstract/Free Full Text]
24. McNally, M. T., Gontarek, R. R., and Beemon, K. (1991) Virology 185, 99-108[CrossRef][Medline] [Order article via Infotrieve]
25. McKnight, S. L., Gavis, E. R., and Kingsbury, R. (1981) Cell 25, 385-98[CrossRef][Medline] [Order article via Infotrieve]
26. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A., and Struhl, K. (1987) Current Protocols in Molecular Biology , John Wiley & Sons, New York
27. Katz, D., Berrodin, T. J., and Lazar, M. A. (1992) Mol. Endocrinol. 6, 805-814[Abstract]
28. Burge, C. B., Tuschl, T., and Sharp, P. A. (1999) The RNA World , 2nd Ed. , pp. 525-560, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY
29. Munroe, S. H., and Lazar, M. A. (1991) J. Biol. Chem. 266, 22083-22086[Abstract/Free Full Text]
30. Lazar, M. A., Jones, K. E., and Chin, W. W. (1990) DNA Cell Biol. 9, 77-83[Medline] [Order article via Infotrieve]
31. Kozak, M. (1992) Annu. Rev. Cell Biol. 8, 197-225[CrossRef]
32. Xu, R. J., Teng, J., and Cooper, T. A. (1993) Mol. Cell. Biol. 13, 3660-3674[Abstract/Free Full Text]
33. Tanaka, K., Watakabe, A., and Shimura, Y. (1994) Mol. Cell. Biol. 14, 1347-1354[Abstract/Free Full Text]
34. Hertel, K. J., Lynch, K. W., and Maniatis, T. (1997) Curr. Opin. Cell Biol. 9, 350-357[CrossRef][Medline] [Order article via Infotrieve]
35. Kanopka, A., Mühlemann, O., and Akusjärvi, G. (1996) Nature 381, 535-538[CrossRef][Medline] [Order article via Infotrieve]
36. Kosaki, A., Nelson, J., and Webster, N. J. G. (1998) J. Biol. Chem. 273, 10331-10337[Abstract/Free Full Text]
37. McCullough, A. J., and Berget, S. M. (1997) Mol. Cell. Biol. 17, 4562-4571[Abstract]
38. Humphrey, M. B., Bryan, J., Cooper, T. A., and Berget, S. M. (1995) Mol. Cell. Biol. 15, 3979-3988[Abstract]
39. Elrick, L. L., Humphrey, M. B., Cooper, T. A., and Berget, S. M. (1998) Mol. Cell. Biol. 15, 343-352
40. Segraves, W. A., and Hogness, D. S. (1990) Genes Dev. 4, 204-219[Abstract/Free Full Text]
41. Laudet, V., Hanni, C., Coll, J., Catzeflis, F., and Stehelin, D. (1992) EMBO J. 11, 1003-1013[Medline] [Order article via Infotrieve]