This Article Abstract Full Text (PDF) Other Versions of this Article: MCB.02289-05v1 27/9/3290    most recent Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Ji, X. Articles by Liebhaber, S. A. Search for Related Content PubMed PubMed Citation Articles by Ji, X. Articles by Liebhaber, S. A.

The 3' Untranslated Region Complex Involved in Stabilization of Human -globin mRNA Assembles in the Nucleus and Serves an Independent Role as a Splice Enhancer

Departments of Genetics and Medicine,¹ Department of Medicine, Renal-Electrolyte and Hypertension Division, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania 19104²

Received 30 November 2005/ Returned for modification 27 December 2005/ Accepted 8 February 2007

	ABSTRACT

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Posttranscriptional controls, mediated primarily by RNA-protein complexes, have the potential to alter multiple steps in RNA processing and function. Human -globin mRNA is bound at a C-rich motif in the 3' untranslated region (3'UTR) by the KH domain protein -globin poly(C)-binding protein (CP). This "-complex" is essential to cytoplasmic stability of -globin mRNA in erythroid cells. Here we report that the 3'UTR -complex also serves an independent nuclear role as a splice enhancer. Consistent with this role, we find that CP binds -globin transcripts prior to splicing. Surprisingly, this binding occurs at C-rich sites within intron I as well as at the 3'UTR C-rich determinant. The intronic and 3'UTR CP complexes appear to have distinct effects on splicing. While intron I complexes repress intron I excision, the 3'UTR complex enhances splicing of the full-length transcript both in vivo and in vitro. In addition to its importance to splicing, nuclear assembly of the 3'UTR CP complex may serve to "prepackage" -globin mRNA with its stabilizing complex prior to cytoplasmic export. Linking nuclear and cytoplasmic controls by the action of a particular RNA-binding protein, as reported here, may represent a modality of general importance in eukaryotic gene regulation.

	INTRODUCTION

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Posttranscriptional controls are central to the establishment and regulation of eukaryotic gene expression (15, 35). Recent data highlight critical links between nuclear and cytoplasmic processes involved in these regulatory pathways. Three prominent examples illustrate how the "nuclear history" of a transcript can affect its fate in the cytoplasm. (i) Translational activity of maternal mRNAs in Xenopus oocytes reflects whether precursor transcripts have transited the nuclear splicing pathway (32). More recent studies demonstrate the same relationship in mammalian cells (4, 36, 37, 52). The mechanism(s) that links nuclear splicing and cytoplasmic translational activity remains poorly defined. (ii) Accelerated cytoplasmic destruction of mRNAs containing premature termination codon-containing mRNAs (nonsense-mediated mRNA decay) is dependent upon the position of the premature termination codon relative to exon-exon junctions (16, 28). This nuclear "history" (i.e., where introns were located) is recorded by deposition of a multiprotein exon-junction complex at the site of each splice (16, 26, 53). (iii) A specific subset of cytokine mRNAs are stabilized in the cytoplasm by binding of hnRNP D p37 to a 3' untranslated region (3'UTR) AU-rich motif. Remarkably, the formation of this 3'UTR RNP occurs exclusively in the nucleus prior to mRNA export (7). Thus, nuclear import of a shuttling RNA-binding protein can be a prerequisite for its deposition on target transcripts and subsequent stabilization of an mRNA(s) in the cytoplasm (7). While the overall impact of "nuclear history" on the cytoplasmic fate of specific mRNAs is evident based on these and other examples (27, 43), the underlying mechanisms involved in integrating these controls and the degree to which they can be generalized remain to be determined.

The erythrocyte is perhaps the most specialized cell in mammalian organisms. Its role in oxygen transport is intimately linked to high-level expression of globin proteins. Defects in globin synthesis result in an array of inherited anemias and underlie the most common of human genetic disorders (42). Erythroblasts undergo global and irreversible transcriptional arrest during terminal differentiation, and their mRNA population shifts from high complexity to greater than 95% globin mRNAs. This shift reflects selective destabilization of most nonglobin mRNAs and reciprocal stabilization of globin mRNAs (47). Loss of -globin mRNA stability due to mutations in the -globin gene results in a common form of -thalassemia (Constant Spring) (10, 24). Thus, robust expression of globins is dependent on gene transcription early in the differentiation process and selective stabilization of globin mRNAs late in the process.

High-level stability of human -globin mRNA (h-globin mRNA) in the erythroblast is dependent on a pyrimidine-pure and C-rich sequence in the 3'UTR (51). This sequence is bound by a 39-kDa poly(C)-binding protein. This -globin poly(C) binding protein (CP) (50) (also referred to as poly(C)-binding protein and hnRNP E) (12, 25, 38) comprises a set of isoforms sharing characteristic triple repeats of the 65-amino-acid KH domain RNA binding motif (14, 30). Each CP isoform contains two KH domains grouped near the N terminus and a third KH domain located at the C terminus. This overall structure is shared by the nuclear hnRNP K protein involved in general transcript packaging (44) and by the mammalian Nova proteins that are implicated in alternative splicing control in the central nervous system (19).

Recent studies suggest that CPs have a nuclear, as well as cytoplasmic, function(s). The major CP isoforms, CP1, CP2, and CP2-KL, are present in the nucleus as well as the cytoplasm (9, 13) and shuttle between the two compartments (9). This shuttling is mediated by a novel set of nuclear localization signals (NLS) and an N-terminal leucine-rich nuclear export signal (NES). These data suggest a model, as yet untested, in which CPs bind h-globin transcripts in the nucleus and then travel with the processed mRNA to the cytoplasm. The possibility that association of CPs with h-globin transcripts may have a specific nuclear function(s) is supported by the observations that CP1 is concentrated in nuclear speckles (34) and that CP2-KL interacts in vivo with the splicing cofactor 9G8 (12). Nuclear roles for the CPs, while suggested by these lines of evidence, remain to be identified.

In the current study we explored the potential nuclear role of CP in h-globin gene expression. We began with the observation that the 3'UTR -complex affects the relative levels of unspliced and spliced -globin mRNA in transfected cells. Subsequent studies revealed that CPs are loaded on -globin transcripts in the erythroblast nucleus and that CP targets C-rich sites in intron I as well as the previously defined 3'UTR stability determinant. In vitro splicing studies support a role for CP in -globin transcript splicing, suggest that the two sets of CP complexes have distinct functions in this pathway, and strengthen the conclusion that the 3'UTR -complex acts as a splice enhancer. Thus, assembly of CP complexes on the h-globin transcript may link nuclear and cytoplasmic controls by regulating transcript processing and prepackaging the mature -globin mRNA for maximized cytoplasmic stabilization.

	MATERIALS AND METHODS

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Cell culture. Human erythroleukemia (K562) cells were grown in RPMI 1640 medium, mouse erythroleukemia (MEL) cells were cultured in minimal essential medium, and HeLa cells were grown in Dulbecco modified Eagle medium supplemented with 10% fetal bovine serum. All culture media contained 100 U of penicillin/ml and 100 µg of streptomycin sulfate/ml, and conditions were maintained at 37°C in a 5% CO₂ incubator. MEL cells and HeLa cells with a stably expressed transfected "Tet-off" transactivator (MEL/tTA and HeLa/tTA cells) were used for conditional expression of the h-globin mRNA; these cells have been previously described in detail (22).

Cell transfection. MEL/tTA cells were transfected with the indicated pTet plasmid DNA by electroporation (20, 22). Two micrograms of the pTet plasmid DNA and 18 µg of carrier DNA were added to a MEL/tTA cell suspension (3 x 10⁷ to 5 x 10⁷per 0.5 ml). Electroporation was performed in a BRL Cell-Porator system under the following settings: 250 V, 1,180 µF, and low resistance. Cells were then cultured in medium without tetracycline for 24 h to induce expression from the transfected gene. The HeLa/tTA cells, grown as an adherent culture on routine tissue culture plastic ware, were split 1 day prior to transfection so that they were at 70% confluence at the time of transfection. The HeLa/tTA cell transfections were carried out using the liposomal reagent Trans-IT (Mirus). pTet plasmid DNA (0.2 µg) and carrier DNA (5.8 µg) were mixed and coated with 12 µl of Trans-IT before they were added to cells (22). Cells were then cultured in medium without tetracycline for 24 h to induce expression from the transfected pTet plasmid.

Cell extract preparation and immunoprecipitation. K562 cells or MEL/tTA cells transfected with pTet-^WT or pTet-^Neut DNA were washed with ice-cold phosphate-buffered saline twice and resuspended in 1,000 µl of ice-cold RSB100 buffer (10 mM Tris-HCl [pH 7.4], 100 mM NaCl, 2.5 mM MgCl₂) containing 0.5% Triton X-100. Total cell extracts were prepared by performing three 10-second sonication pulses on ice (Sonic Dismembrator [Fisher Scientific] set to a scale of 10). The sonicated material was layered onto a 30% (wt/vol) sucrose cushion in RSB100 and centrifuged at 4,000 x g for 15 min (33), and the supernatant was collected and used for immunoprecipitation experiments. Immunoprecipitation was carried out as described previously (20). Immunoprecipitated pellets were extracted and ethanol precipitated prior to RNase protection assay (RPA) or reverse transcription-PCR (RT-PCR) analysis.

SDS-PAGE. Cell extracts were separated by sodium dodecyl sulfate (SDS)-12.5% polyacrylamide gel electrophoresis (PAGE) and electroblotted to nitrocellulose membranes (Protran BA 85; Schleicher & Schuell) for 1 h at 150 mA in transfer buffer (20 mM Tris, 150 mM glycine, 20% methanol), using a Semi-phor transfer apparatus (Hoefer). The membranes were blocked in 3% nonfat milk in 1x phosphate-buffered saline for 1 h at room temperature, followed by an additional hour with primary antisera. Primary rabbit antibodies to the CP isoforms have been previously detailed and characterized (8). Anti-hnRNP L antibody, a mouse monoclonal antibody, was a gift from Gideon Dreyfuss (University of Pennsylvania). Horseradish peroxidase-labeled secondary antibodies (Amersham) were used as detailed by the supplier. Donkey anti-rabbit immunoglobulin G-horseradish peroxidase and sheep anti-mouse immunoglobulin G-horseradish peroxidase secondary antibodies were used at a 1:5,000 dilution (Amersham), and signals were developed by ECL (Boehringer Mannheim).

EMSA. RNA oligonucleotides were synthesized by the University of Pennsylvania Nucleic Acid Core Facility and were 5' end labeled using T4 polynucleotide kinase (New England Biolabs, Beverly, MA) and [-³²P]ATP (Amersham). All labeled oligonucleotides were gel purified on 12% denaturing gels prior to use (17). A series of in vitro transcription templates were generated by PCR. An Sp6 promoter sequence was incorporated at the 5' ends of 5' primers to facilitate in vitro transcription. The PCR products include the following: 1 to 575 (full-length -globin mRNA), 1 to 238, 217 to 474, 370 to 474, 370 to 575, and 466 to 575 (full-length 3'UTR). The full-length intron I and intron II templates and the two overlapping intron I probes were prepared the same manner. These RNA probes for the electrophoretic mobility shift assay (EMSA) studies were generated using the Maxiscript SP6 kit (Ambion). EMSAs were carried out as described previously (8) with minor modifications. To perform RNA EMSA, each in vitro-transcribed RNA probe (20,000 cpm) was incubated with 10 to 15 µg of S100 extract from K562 cells. The incubation was in 20 µl of binding buffer (10 mM Tris-HCl [pH 7.4], 150 mM KCl, 1.5 mM MgCl₂, and 0.5 mM dithiothreitol) at room temperature for 20 min. The binding samples were subsequently incubated with RNase T₁ (20 U; Roche) at room temperature for 10 min. Addition of RNase T₁ (RNase T₁ cleaves 3' of single-stranded G residues) to the binding reaction mixture after the complex has formed will degrade any unbound RNA probe and will enhance the resolution of the RNP complex on the gel. One microliter of heparin (50 mg/ml) was added to each reaction mixture 10 min prior to loading. Samples were resolved on a 5% native polyacrylamide gel. For EMSA with synthetic RNA oligonucleotides, 5 ng of each oligonucleotide (approximately 20,000 cpm) was mixed with 30 µg of K562 S100 extract and then incubated and gel analyzed as detailed above, with the exception that the RNase T₁ step was omitted.

Nuclear extract preparation and in vitro splicing assays. HeLa cell nuclear extract was prepared as described previously (11), and in vitro splicing assays followed a published procedure (18). Capped pre-mRNA substrates were synthesized by in vitro transcription using Sp6 or T7 RNA polymerase (Ambion). In vitro splicing was performed with 8.0 µl of HeLa cell nuclear extract in 25-µl reaction mixtures containing 21 fmol (100,000 cpm) of substrate in the presence of 2.8 mM ATP, 14 mM creatine phosphate (Sigma), 4.5 mM MgCl₂, and 85 mM KCl. The mixtures were incubated at 30°C for indicated times, followed by addition of 125 µl of stop buffer (100 mM Tris-HCl [pH 7.5], 10 mM EDTA-Na₂ [pH 8.0], 1% SDS, 150 mM NaCl, 300 mM sodium acetate [pH 5.2]). After phenol-chloroform extraction and ethanol precipitation, the pellet was resuspended in loading buffer or diethyl pyrocarbonate-treated water and used for RT-PCR assay.

Depletion of CP proteins from HeLa extracts. "Poly(C)-depleted" nuclear extracts were generated using poly(C)-agarose beads (Sigma). An equal amount of unconjugated protein A-Sepharose CL-4B beads was used in a parallel "mock" depletion procedure. Two milligrams of nuclear extract was bound to 0.2 mg of agarose beads for 20 min at 4°C in nuclear preparation buffer D (20 mM HEPES, pH 7.9; 100 mM KCl; 0.2 mM EDTA; 20% glycerol). Following a brief spin to pellet the beads, the supernatant containing unbound protein was incubated with an additional 0.2 mg of beads. This was repeated once, and the final supernatant was recovered and concentrated to 6 mg/ml with a Centricon 10 spin column.

For immunodepletion studies, 2 mg of nuclear extracts was incubated for 1 h at 4°C with anti-CP2/KL antibody (FF3) conjugated to protein A-Sepharose CL-4B beads or with an equal amount of rabbit preimmune serum conjugated and used in parallel. The unbound fraction was recovered following a brief spin to pellet the beads. This was repeated two more times, and the final supernatant was diluted with nuclear preparation buffer D and concentrated to 6 mg/ml with a Centricon 10 spin column.

RPA. Internally ³²P-labeled probes used for the RPA were generated by in vitro transcription of plasmids containing cDNA inserts for h-globin (29), human glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and mouse GAPDH (Ambion, Austin, TX). RPA was carried out as described previously (20). Radioactivity in bands of interest was quantified by PhosphorImager analysis (Storm 840; Molecular Dynamics).

RT-PCR. RNA (2.5 µl) was incubated with 1 pmol of reverse primer, 1 mM each deoxynucleoside triphosphate, 2.5 U of anti-RNase (Ambion), 50 U of Moloney murine leukemia virus reverse transcriptase (Promega), and 1x Moloney murine leukemia virus RT buffer (Promega) in a volume of 12.5 µl. After incubation at 37°C for 1 h, the samples were used as a template for PCR. The forward primer (20 pmol) was end labeled by incubation with 5 µl of [³²-P]ATP (6,000 Ci/mmol), 1x reaction buffer (New England Biolabs), and 20 U of T4 polynucleotide kinase (New England Biolabs) for 60 min at 37°C, with termination at 70°C for 10 min. The PCR mixtures included 5 µl of the RT product, 0.2 mM deoxynucleoside triphosphates, 1.5 mM MgCl₂, 2.5 µl of the labeled primer, 2.5 µg of each primer, 0.25 U of AmpliTaq (Perkin-Elmer), and 1x PCR buffer II (Perkin-Elmer). The PCRs were performed for various numbers of cycles depending on the primers used. Samples were visualized by 6% denaturing PAGE and quantified with the PhosphorImager (ImageQuant; Molecular Dynamics). Primers used were as follows: h-globin between exon 2 and exon 3 (315 bp), forward 5'-GTGGACGACATGCCCAACGC-3' and reverse 5'-CCCACTCAGACTTTATTCAA-3'; h-globin between exon 1 and exon 2 (238 bp), forward 5'-ACTCTTCTGGTCCCCACAGACTCA-3' and reverse 5'-CAGGGCGTCGGCCACCTTCTTG-3'; human GAPDH (180 bp), forward 5'-CAACTACATGGTTTACATGTTC-3' and reverse 5'-GCCAGTGGACTCCACGAC-3'; mouse GAPDH (236 bp), forward 5'-TTCACCACCATGGAGAAGGC-3' and reverse 5'-GGCATGGACTGTGGTCATGA-3'; and unspliced h-globin transcript between intron I and exon 2 (142 bp), forward 5'-CCAAACCCCACCCCTCACTCT-3' and reverse 5'-CAGGGCGTCGGCCACCTTCTTG-3'.

	RESULTS

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
The PR stability determinant in the h-globin 3'UTR plays a role in nuclear processing. In prior studies we established a "Tet-off" transcriptional control system in MEL cells that was used to monitor mRNA half-lives of -globin mRNAs in a variety of structural settings (22). These studies confirmed an essential role for the 3'UTR pyrimidine-rich (PR) motif and its bound CP in h-globin mRNA stabilization (22). It was also possible to visualize unspliced -globin transcripts in this system by assaying the cells immediately following the Tet-controlled transcriptional pulse. Using this approach, and comparing unspliced and spliced -globin mRNAs either containing (^WT) or lacking (PR) the PR motif, we observed that deletion of the 3'UTR PR stability determinant results in a decrease in the proportion of spliced transcripts from 74% to 46% of total transcripts (Fig. 1A and B). This apparent decrease in splicing activity exceeds what could be reasonably attributed to the differences in the relative stabilities of the cytoplasmic ^WT and PR mRNAs (10.5 versus 7.5 h, respectively [22]). Thus, the 3'UTR PR determinant, in addition to its role in stabilization of cytoplasmic -globin mRNA, appears to enhance nuclear processing of -globin transcripts.

View larger version (15K): [in this window] [in a new window]   FIG. 1. The 3'UTR PR determinant enhances h-globin transcript splicing efficiency in transfected erythroid and nonerythroid cells. (A) Diagrams of the human -globin (WT) transcript and derivative transcripts in which a 42-base segment encompassing the 3'UTR PR stability determinant is either deleted (PR) or replaced by an identically sized unrelated ("neutral") sequence (Neut). A schematic of the RPA using an internally 32P-labeled RNA probe to quantify unspliced (intron I-containing) and spliced -globin transcripts is shown at the bottom. (B) Impact of the PR determinant on -globin transcript splicing in erythroid cells. The WT and PR genes were expressed for a 4-h window in MEL cells stably transfected with a tetracycline transactivator gene (MEL/tTA cells) (see Materials and Methods). RNAs were extracted immediately following the transcriptional pulse and were analyzed by RPA. The bands corresponding to unspliced and spliced -globin transcripts are indicated to the left of the analytic gel. The level of splicing of intron I of the -globin transcripts is calculated as a ratio of spliced to total -globin transcripts. These data (means and standard deviations) are indicated on the histogram (n = 3). (C) Impact of the PR determinant on -globin transcript splicing in nonerythroid cells. HeLa cells stably transfected with the tTA plasmid (HeLa/tTA) were pulsed with WT and Neut mRNAs (as described for panel B). RNA harvested after a 24-hour transcriptional pulse was analyzed by RPA. Ratios of spliced to total transcripts were determined and plotted on the histogram (n = 3).

CP2 binds to unspliced -globin transcripts. Analysis of h-globin expression in MEL and HeLa cells suggests that the 3'UTR CP complex enhances splicing (Fig. 1). This predicts that CP interacts with the -globin transcript prior to splicing. The human erythroblastoid K562 cell line expresses abundant levels of wild-type h-globin mRNA (1). To test for an interaction between CP and unprocessed -globin transcripts, CP-containing RNA-protein (RNP) complexes were immunoprecipitated from K562 cell extract, and the mRNA content was analyzed by RT-PCR. Immunoprecipitation was done with an affinity-purified antiserum that specifically recognizes the two major CP isoforms, CP2 and CP2-KL (lab antiserum designation FF3 [8]). For convenience these two isoforms, both of which are encoded at the CP2 (PCBP2) locus, are referred to collectively as CP2. The immunoprecipitation was optimized to maximize efficiency and selectivity and to preserve mRNP integrity (see Materials and Methods). RNA extracted from the immunoprecipitated pellets was amplified with primers designed to detect intron I-containing -globin transcripts (Fig. 2A). The controls included a set of primers that detect GAPDH mRNA, an mRNA that is not bound by CP (48) and a parallel immunoprecipitation using preimmune serum. These studies reveal robust and specific enrichment for intron I-containing -globin transcripts in CP complexes. In contrast, GAPDH mRNA is present only at trace levels in the same immunoprecipitated sample, and neither -globin nor GAPDH RNA is detected in the immunoprecipitation carried out with preimmune serum. These data indicate that CP binds to -globin transcripts prior to splicing.

View larger version (27K): [in this window] [in a new window]   FIG. 2. CP2 is associated with unspliced -globin transcripts. (A) CP2 binds to unspliced -globin transcripts. K562 cell extract was immunoprecipitated (IP) with anti-CP2 (FF3) or preimmune (P.I.) serum. -globin transcripts containing intron I and human GAPDH mRNA (G) (negative control) were detected by RT-PCR in the input sample. The RNAs isolated from the RNP pellets generated from the immunoprecipitation reactions were also assayed. The -globin RT-PCR primer set shown below the gel specifically detects transcripts that retain intron I. (B) CP2 binds at one or more sites in the unspliced -globin transcript in addition to the 3'UTR PR binding site. WT and Neut genes were transfected into MEL cell line MEL/tTA (22). Cellular extracts prepared after a 24-hour transcriptional pulse were immunoprecipitated with FF3 or preimmune antiserum. -globin transcripts with and without intron II were assayed by RT-PCR in the starting extracts (input) and in the FF3 and preimmune RNP pellets. The RT-PCR primer set is shown below the gel.

CP2 binds unspliced -globin transcripts at one or more intronic sites.

CP2 binds to two C-rich regions within -globin intron I. The preceding experiments reveal that CP2 binds unspliced -globin transcripts at one or more sites unassociated with the 3'UTR PR (Fig. 2B). Additional CP binding sites in the -globin transcript were sought by RNA EMSA. CP2 RNP complexes were identified on the basis of poly(C) sensitivity and supershifting with the anti-CP2 antiserum. An EMSA-based survey of the mature -globin mRNA confirmed that CP binding is restricted to the 3'UTR (Fig. 3). Introns I and II were then tested for CP binding. The EMSA reveals strong and selective binding of CP to intron I (Fig. 4A). Intron I is 54% C, and it contains three prominent C-rich patches. These C-rich patches 1, 2, and 3, are located adjacent to the 5' (patch 1) and 3' (patches 2 and 3) ends of the intron (Fig. 4B and C). EMSA of overlapping segments encompassing either the 5' or 3' terminus of intron I revealed that both regions bind CP2, although binding to the 3' half of the intron was by far the most robust (Fig. 4B).

View larger version (55K): [in this window] [in a new window]   FIG. 3. Fully processed -globin mRNA contains a single CP binding site that is located in the 3'UTR. (A) Diagram of the mature -globin mRNA and EMSA probes. The PR determinant is highlighted as a shaded rectangle. Each of the probes used for EMSA and their coordinates are indicated below the diagram. The termini of each probe are numbered, with the base 1 corresponding to the capped 5' terminus. (B) EMSA scan of the spliced -globin mRNA to detect CP binding sites. Each 32P-labeled RNA probe was incubated with K562 S100 extract, and RNase T1-resistant complexes were visualized on a nondenaturing 6% polyacrylamide gel. In the indicated lanes (+) a poly(C) competitor or an antibody to CP2/CP2KL (FF3 antiserum) was added to the binding reaction mixture (see Materials and Methods). The positions of the -complex (bracket) and of the supershifted complex (arrow) are indicated.

View larger version (54K): [in this window] [in a new window]   FIG. 4. CP2 binds to intron I of the -globin transcript. (A) CP2 binds to intron I but not intron II sequences. In vitro-transcribed intron I and intron II probes were individually incubated with K562 S100 extract. Reactions were analyzed on nondenaturing gels as described for Fig. 3. A diagram of the unspliced -globin transcript is shown below the two gel panels; the three exons are represented by the rectangles and the two introns by the connecting lines. The positions of the intron I and intron II probes are indicated below the transcript diagram. (B) CP2 binds to sites adjacent to the 5' and 3' splice sites in intron I. Designations are as in panel A. The two partially overlapping probes used in the assay are indicated below their respective gel panels. The C-rich regions adjacent to the 5' and 3' splice sites of intron I are indicated by the shaded ovals (for the sequence, see panel C). (C) Sequence of h-globin intron I. C-rich regions (patches 1, 2, and 3) with the potential to serve as individual CP binding sites adjacent to the 5' and 3' splice sites are boxed and shaded.

View larger version (50K): [in this window] [in a new window]   FIG. 5. Mapping of the CP binding site(s) in intron I. (A) Diagram of probes used for binding site mapping. Four parental RNA oligonucleotides (I-1 to I-4) covering the entire intron I and I-1/mut with two C-to-G substitutions were chemically synthesized and in vitro kinase labeled for the EMSA. The three C-rich patches are highlighted in shaded ovals. (B) Probes I-1 through I-4 and I-1/mut were each incubated with K562 S100 extract. Complex formation was analyzed on nondenaturing gels as described above. In the indicated lanes, a poly(C) competitor or an antibody to CP2/CP2KL (FF3 antiserum) was added to the reaction mixture. The positions of the -complex (bracket) and of the supershifted complex (arrow) are indicated.

View larger version (57K): [in this window] [in a new window]   FIG. 6. Determination of the minimal binding site within the I-4 region of intron I of the -globin transcript. (A) Diagram of probes used for minimal binding site mapping. The I-4 parental RNA oligonucleotide and the derivative I-4/mut2, I-4/mut3, and I-4/mut2/3 RNA oligonucleotides with C-to-T substitutions within either or both of the two C-rich patches were chemically synthesized and in vitro kinase labeled for the EMSA. (B) In vitro-transcribed -globin 3'UTR RNA, I-4, I-4/mut2, I-4/mut3, and I-4/mut2/3 were each incubated with K562 S100 extract. In the indicated lanes, a poly(C) competitor, an antibody to CP2/CP2KL (FF3 antiserum), or an antibody to CP1 (FF1 antiserum) was added to the reaction mixture. The positions of the -complex (bracket) and of the supershifted complex (arrow) are indicated.

The CP2 complex within intron I acts as a local repressor of intron I splicing.

View larger version (40K): [in this window] [in a new window]   FIG. 7. CP2 inhibits in vitro excision of intron I. (A) Depletion of CPs from HeLa cell nuclear extract by adsorption to poly(C) beads. Western analyses of extracts treated with unconjugated beads (mock) or with beads conjugated with poly(C) are shown. The panels were individually probed with antisera specific to CP2, CP1, and hnRNP L (loading control). (B) In vitro splicing of intron I is enhanced by CP depletion. Left, the internally labeled (asterisks) splicing substrate extending from exon 1 through exon 2 of the -globin transcript incubated with no extract, mock-depleted extracts, or poly(C)-depleted extracts. Right, splicing of a control transcript lacking CP binding sites (PIP 11 mRNA). The diagram at the bottom shows the internally 32P-labeled RNA splicing substrate that encompasses exon 1 and intron I and extends into exon 2 of the h-globin transcript. (C) CP2 immunodepletion of HeLa cell nuclear extract. The immunodepletion using anti-CP2 (FF3) or preimmune serum (mock) was monitored by Western blotting. Primary antibodies used in the Western blots recognize CP2 (FF3 antiserum), CP1 (FF1 antiserum), or hnRNP L (a kind gift of G. Dreyfuss, University of Pennsylvania). (D) In vitro splicing of intron I from a substrate extending from exon 1 through exon 2 is sensitive to CP2. Splicing was carried out in extract immunodepleted with preimmune serum or anti-CP2 antiserum (FF3). The splicing reaction was activated by the addition of ATP with or without the addition of recombinant CP2 or CP1. Splicing efficiency (spliced transcripts as a percentage of total transcripts) was assessed by RT-PCR using the indicated primer set. (E) Semiquantitative analysis of pre-RNA and mRNA levels by RT-PCR. Left panel, in vitro-transcribed full-length unspliced -globin transcript and mature -globin mRNA were mixed at different concentrations (shown above the respective lanes). The mixtures were reverse transcribed and then PCR amplified using labeled primer (as for panel D). Right panel, the PCR output (ratio of pre-RNA to mRNA) was plotted against the input (ratio of pre-RNA to mRNA).

The structural basis for CP2 RNP functions in intron I splicing was explored by analyzing the splicing of -globin transcripts containing the base substitutions in C-rich patches 2 and 3 in intron I (Fig. 6A). Each splicing substrate, extending from exon 1 through exon 2, was incubated in native HeLa cell nuclear extract, and the splice products were quantified. Inactivation of the CP2 complex at C-rich patch 2 (mut2) doubled the splicing activity (24% to 47%) (Fig. 8). A similar enhancement of splicing was seen when the CP2 complex at C-rich patch 2 was disrupted by a parallel set of C-to-G substitutions (24% to 40%) (data not shown). The base substitutions in patch 3 (mut3) had only a minimal effect on the splicing reaction. A combination of mut2 and mut3 retained the enhancement of splicing activity seen with mut2 alone, doubling splicing to 46%. We conclude from these results that CP2 RNP complex at C-rich patch 2 exerts a local repressive effect on splicing of intron I.

View larger version (20K): [in this window] [in a new window]   FIG. 8. CP complexes formed at C-rich patch 2 of intron I of the -globin transcript represses splicing. The internally labeled WT or its derivative mut2 and mut3 transcripts (as in Fig. 6A) extending from exon 1 through exon 2 of the -globin transcript were incubated with native HeLa cell nuclear extracts for 2 h. After incubation, RNA was extracted and separated by denaturing PAGE. The splicing efficiency was calculated as a percentage of spliced to total -globin transcripts.

The 3'UTR -complex acts as an enhancer of intron I splicing.

View larger version (48K): [in this window] [in a new window]   FIG. 9. In vitro splicing of full-length -globin transcript is enhanced by the 3'UTR PR determinant. (A) Unspliced WT and Neut transcripts have equivalent stabilities in the in vitro splicing system. In vitro-transcribed full-length WT and Neut pre-RNAs were incubated with HeLa cell extract in the absence of ATP for the indicated times (hours). RNAs were recovered and quantified by RPA. The signal intensities at each time point are indicated as a percentage of the starting level. (B and C) In vitro-transcribed sets of full-length -globin transcript (WT) and its derivative mutants (as in Fig. 8) or the -globin transcript lacking the PR determinant (Neut) and its corresponding mutants were incubated with HeLa cell nuclear extracts for 1 h. The splicing reaction was activated by the addition of ATP. Splicing products were detected by RT-PCR (as in Fig. 7D). The splicing efficiency was calculated as a percentage of spliced to total -globin transcripts.

	DISCUSSION

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Previous studies have demonstrated that the 3'UTR -complex is a major determinant of a-globin mRNA stability. This high-affinity interaction is based not only on a C-rich determinant but also on the presentation of this determinant in a single-strand configuration within the mRNA (17). Base substitutions within the -globin 3'UTR that block CP binding result in a marked decrease in mRNA accumulation in vivo. Using in vitro approaches and literature searches, it was possible to identify -complexes on the 3'UTRs of additional stable mRNAs (17, 23). In each case the complex was linked to an mRNA with unusual stability (38, 40, 45). These findings suggested that the -complex might constitute a general determinant of high-level mRNA stability (17).

The mechanism by which the -complex stabilizes cytoplasmic -globin mRNA is currently under study. It has been demonstrated that CP protein, and not its C-rich binding site, is the critical determinant of mRNA stabilization (22). For example, an -globin mRNA that had been destabilized by deletion of the native C-rich binding site (PR region) in the 3'UTR can be fully restabilized by artificial "tethering" of a CP fusion protein to the 3'UTR (22). Thus, the role of the C-rich determinant appears to be primarily one of targeting CP binding.

While cytoplasmic roles of CP have been addressed in multiple systems (31, 39), the idea that CPs have a nuclear function(s) has received less attention. A nuclear role for CPs is, however, suggested by several observations. A set of novel NLS and an N-terminal leucine-rich NES have been mapped and characterized in these proteins (9). The combination of NLS and NES predicts that CP1, CP2, and CP2-KL shuttle between the nuclear and the cytoplasmic compartments (9). In addition, CP protein is concentrated in nuclear speckles and interacts with a number of nuclear proteins, including hnRNP L, hnRNP K, hnRNP I (polypyrimidine-tract binding protein [PTB]), Y-box-binding protein, and splicing factor 9G8 (reference 12 and our unpublished data). It remains to be determined how these interactions factor into CP function.

In the current study, we provide evidence that CPs are involved in the nuclear processing of -globin transcripts. CPs bind to unspliced a-globin transcripts within intron I as well as in the 3'UTR. Binding of CPs within intron I appears to directly repress intron I splicing when assessed in vitro. In contrast, the 3'UTR -complex enhances -globin splicing in transfected cells and increases the efficiency of splicing in vitro. Thus, the impact of CP on -globin splicing is complex; while the intron I complex may be having a local repressive effect on intron excision, the 3'PR complex is able to enhance splicing of the full-length transcript. How these two sets of CP complexes interact in cis to control overall mRNP structure and processing can now be addressed.

CP binds to several sites on the unspliced -globin transcripts. The major CP binding site (patch 2) in is in close proximity to the predicted branch point site and the polypyrimidine tract, two regions that are critical to intron excision. Binding of CP in this region represses intron I splicing, and the C-to-T substitutions (or corresponding C-to-G substitutions [data not shown]) at those sites (I-4/mut2) which disrupt the formation of -complex increased the splicing of intron I (Fig. 6, 8, and 9). This effect could be mediated through the same mechanism as proposed for the binding of Drosophila Sex lethal (Sxl) to the 3' splice site, which represses splicing of Transformer (Tra) exon 2. In this case, the bound protein is postulated to interfere with U2AF binding (2). Similarly, binding sites for PTB have been shown to reside in close proximity to some 3' splice sites, and PTB binding at these sites may block U2AF binding and inhibit splicing (49). Along the same mechanistic lines, hnRNP A1 binding to sites near the branch point of a human immunodeficiency virus type 1 tat intron have been shown to prevent U2 binding (46). Taken together, our depletion and mutagenesis studies indicate that the disruption of the -complex at the splice acceptor site releases the inhibitory effect of CP binding and results in an increased splicing efficiency.

The basis for splicing enhancement by the 3'UTR CP complex is less clear. It is possible that the 3'UTR complex enters into a long-range interaction with the intron I complex to generate a higher-order RNP substrate that relieves the local splicing blockade and favors intron excision. Such long-range interactions may rely on the documented ability of CPs to homo- and heterodimerize (reference 21 and our unpublished observations). The ability of the 3'UTR complex to counter the repressive effect of intron I-bound CP2 and to promote splicing is likely to involve additional interacting proteins and formation of complex RNP assemblies. Further structural and biochemical studies will be needed to identify CP-associated proteins involved in these functions and to test for long-range interactions and cooperativity between the two sets of CP binding sites.

It is of interest that the role of CPs in -globin transcript splicing may be isoform specific. While CP1 and CP2 both bind effectively to the 3'UTR PR motif (Fig. 6B and our prior studies), the binding within intron I is specific to CP2 (Fig. 6B). In addition, the splice site repression at intron I appears to be specific to CP2; selective immunodepletion of CP2 isoforms (CP2 and/or CP2-KL) from the splicing extract results in a significant increase in intron I excision, and this is selectively reversed by readdition of CP2 (Fig. 7C and 7D). CP1, which remains in the CP2-depleted extract, is not able to compensate for the loss of CP2 isoforms. While CP2 shares more than 80% sequence identity with CP1 (31), studies in a number of systems have supported distinct functions (31, 39). For example, BCR-ABL expression in chronic myelocytic leukemia cells selectively induces cellular CP2 levels that in turn inhibit C/EBP mRNA translation via a putative blockade of 40S scanning. In contrast, CP1 has no similar capacity for inhibition of C/EBP expression in these cells (41). Likewise, binding of CP2 to stem-loop IV of the poliovirus mRNA internal ribosome entry site is essential for efficient poliovirus translation in HeLa cell extracts. Although recombinant CP1 and CP2 can both bind to stem-loop IV in vitro (13), only recombinant CP2 is able to restore poliovirus internal ribosome entry site activity in CP-depleted HeLa cell lysate (3). Thus, the CP isoforms can mediate specific biologic functions that may not be mirrored by in vitro binding activities. One potential explanation for these differences is that CP isoforms may associate with distinct sets of protein partners that are required for their respective activities.

The results presented here allow us to add CP to a rather limited list of regulatory factors that can influence the efficiency of splicing positively or negatively depending on the site of binding (2). While CP has been recently shown to affect splicing efficiency of two alternative exons, the actual sites of binding were not demonstrated in these studies (5). We now show that the positioning of CP complexes within the transcript dictates opposing effects on the splicing reaction. Such behavior has been shown for several splicing regulatory proteins and may reflect a general role of certain RNA-binding proteins to maintain the fidelity of splicing for constitutively spliced exons and introns as well as combinatorial control involved in alternatively spliced genes. In constitutive splicing, as occurs for globin transcripts, there may be a requirement for RNP complexes that suppress "cryptic" 5' or 3' splice sites embedded in the nascent transcript. Various hnRNP proteins have been shown to bind such sequences and ensure proper splice site pairing (2). Factors that enhance the recognition of appropriate splice sites must work within this background of negative regulation. Thus, the positions of the regulatory sequences relative to different candidate splice sites affect splicing outcome. Balancing positive and negative influences may offer an additional level of posttranscriptional control in splicing. Seen in this context, our findings that CP proteins display position-dependent splicing enhancement or repression may well be indicative of a broader role in regulating splicing efficiency and in the maintenance of splicing fidelity.

	ACKNOWLEDGMENTS

 
This work was supported by grants from the National Institutes of Health (R37 HL 65449 and P01-CA72765 to S.A.L. and CA093769 to R.P.C.), a Cooley's Anemia Foundation Fellowship (to X.J.), and the generosity of the Doris Duke Foundation.

	FOOTNOTES

 
* Corresponding author. Mailing address: Departments of Genetics and Medicine, University of Pennsylvania School of Medicine, Room 428 CRB, 415 Curie Blvd., Philadelphia, PA 19104. Phone: (215) 898-7834. Fax: (215) 573-5157. E-mail: liebhabe{at}mail.med.upenn.edu

Published ahead of print on 26 February 2007.

	REFERENCES

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
1. Andersson, L. C., K. Nilsson, and C. G. Gahmberg. 1979. K562—a human erythroleukemic cell line. Int. J. Cancer 23:143-147.[Medline]

2. Black, D. L. 2003. Mechanisms of alternative pre-messenger RNA splicing. Annu. Rev. Biochem. 72:291-336.[CrossRef][Medline]

3. Blyn, L. B., J. S. Towner, B. L. Semler, and E. Ehrenfeld. 1997. Requirement of poly(rC) binding protein 2 for translation of poliovirus RNA. J. Virol. 71:6243-6246.[Abstract]

4. Bourdon, V., A. Harvey, and D. M. Lonsdale. 2001. Introns and their positions affect the translational activity of mRNA in plant cells. EMBO Rep. 2:394-398.[Medline]

5. Broderick, J., J. Wang, and A. Andreadis. 2004. Heterogeneous nuclear ribonucleoprotein E2 binds to tau exon 10 and moderately activates its splicing. Gene 331:107-114.[CrossRef][Medline]

6. Carstens, R. P., W. L. McKeehan, and M. A. Garcia-Blanco. 1998. An intronic sequence element mediates both activation and repression of rat fibroblast growth factor receptor 2 pre-mRNA splicing. Mol. Cell. Biol. 18:2205-2217.[Abstract/Free Full Text]

7. Chen, C. Y., N. Xu, W. Zhu, and A. B. Shyu. 2004. Functional dissection of hnRNP D suggests that nuclear import is required before hnRNP D can modulate mRNA turnover in the cytoplasm. RNA 10:669-680.[Abstract/Free Full Text]

8. Chkheidze, A. N., D. L. Lyakhov, A. V. Makeyev, J. Morales, J. Kong, and S. A. Liebhaber. 1999. Assembly of the alpha-globin mRNA stability complex reflects binary interaction between the pyrimidine-rich 3' untranslated region determinant and poly(C) binding protein CP. Mol. Cell. Biol. 19:4572-4581.[Abstract/Free Full Text]

9. Chkheidze, A. N., and S. A. Liebhaber. 2003. A novel set of nuclear localization signals determine distributions of the CP RNA-binding proteins. Mol. Cell. Biol. 23:8405-8415.[Abstract/Free Full Text]

10. Clegg, J. B., D. J. Weatherall, and P. F. Milner. 1971. Haemoglobin Constant Spring—a chain termination mutant? Nature 234:337-340.[CrossRef][Medline]

11. Dignam, J. D., R. M. Lebovitz, and R. G. Roeder. 1983. Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res. 11:1475-1489.[Abstract/Free Full Text]

12. Funke, B., B. Zuleger, R. Benavente, T. Schuster, M. Goller, J. Stevenin, and I. Horak. 1996. The mouse poly(C)-binding protein exists in multiple isoforms and interacts with several RNA-binding proteins. Nucleic Acids Res. 24:3821-3828.[Abstract/Free Full Text]

13. Gamarnik, A., and R. Andino. 1997. Two functional complexes formed by KH domain containing proteins with the 5' noncoding region of poliovirus RNA. RNA 3:882-892.[Abstract]

14. Gibson, T. J., J. T. Thompson, and J. Heringa. 1993. The KH domain occurs in a diverse set of RNA-binding proteins that include the antiterminator NusA and is probably involved in binding to nucleic acid. FEBS Lett. 324:361-366.[CrossRef][Medline]

15. Hazelrigg, T. 1998. The destinies and destinations of RNAs. Cell 95:451-460.[CrossRef][Medline]

16. Hentze, M. W., and A. E. Kulozik. 1999. A perfect message: RNA surveillance and nonsense-mediated decay. Cell 96:307-310.[CrossRef][Medline]

17. Holcik, M., and S. A. Liebhaber. 1997. Four highly stable eukaryotic mRNAs assemble 3' untranslated region RNA-protein complexes sharing cis and trans components. Proc. Natl. Acad. Sci. USA 94:2410-2414.[Abstract/Free Full Text]

18. Hovhannisyan, R. H., and R. P. Carstens. 2005. A novel intronic cis element, ISE/ISS-3, regulates rat fibroblast growth factor receptor 2 splicing through activation of an upstream exon and repression of a downstream exon containing a noncanonical branch point sequence. Mol. Cell. Biol. 25:250-263.[Abstract/Free Full Text]

19. Jensen, K. B., B. K. Dredge, G. Stefani, R. Zhong, R. J. Buckanovich, H. J. Okano, Y. Y. Yang, and R. B. Darnell. 2000. Nova-1 regulates neuron-specific alternative splicing and is essential for neuronal viability. Neuron 25:359-371.[CrossRef]