| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Laboratory of Neurobiology2
Office of Clinical Research,3 National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina 27709
Departments of Medicine and Biochemistry,4 Duke University Medical Center, Durham, North Carolina 27710
Department of Pathology and Lab Medicine,5 Robert Wood Johnson Medical School, New Brunswick, New Jersey 08901
Molecular Histology Center,6 Environmental and Occupational Health Sciences Institute, University of Medicine and Dentistry, Piscataway, New Jersey 08854
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
AU-rich element, cytokines, deadenylation, gene regulation, mRNA turnover, placenta, rodent-specific genes, trophoblast
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Three important features of the TTP TZF domain distinguish TTP family members from more distantly related CCCH proteins. First, there is strictly defined intrafinger spacing of C-X8-C-X5-C-X3-H. Second, there is strict interfinger spacing of exactly 18 amino acids. Third, there is a conserved and characteristic amino acid lead-in sequence (R/K)YKTEL immediately preceding the first cysteine of each finger. Alteration of any of these features of the CCCH fingers can interfere with RNA binding activity [9].
TTP is highly conserved throughout evolution, having been identified in mammals [10], frogs [11], and zebrafish (conceptual translation of GenBank accession number AW4223344). More distantly related proteins, whose similarity to mammalian TTP is restricted largely to the TZF domain, are also found in flies [12], nematodes [13], and yeast [14]. It is not yet clear whether the TTP-like proteins found in nonvertebrate species represent functional TTP orthologues.
Other vertebrate proteins contain CCCH TZF domains similar in primary structure to that of TTP as well as other conserved domains. Two such proteins have been identified in mammals, amphibians, and fish as well as a fourth family member that has been identified only in frogs and fish [7, 11]; this last member may actually represent a cluster of closely related proteins [15]. The two additional mammalian members of this vertebrate TTP protein subfamily are referred to here as ZFP36L1 (also known as CMG1 [16], TIS11B [17], ERF1 [18], BRF1 [19], and Berg36 [20]), and ZFP36L2 (also known as TIS11D [17], ERF2 [21], and BRF2 [22]). This nomenclature is based on the approved gene symbols for TTP (Zfp36 in mouse and ZFP36 in human), with ZFP36L1 referring to ZFP36-like 1. ZFP36L1 and ZFP36L2 each contain a TTP-like CCCH TZF domain and can exhibit RNA binding, deadenylating, and degrading activities similar to those of TTP [7]. Although the genes encoding mouse ZFP36L1 and ZFP36L2 are regulated differently from Zfp36 and are expressed to a large extent in different tissues, their similar mRNA binding and destabilizing properties to TTP suggest that considerable functional overlap exists among the members of this protein family.
Using database searches with the TTP TZF domain, we identified a putative fourth family member in mouse and rat genomic sequences and expressed sequence tags (ESTs). This new protein has been designated ZFP36L3 and is encoded by Zfp36l3. This protein exhibits both similarities with and differences from the previously known family members in terms of biochemistry and patterns of expression; some of these are described in this paper.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A partial cDNA encompassing the open reading frame of Zfp36l3 was cloned using reverse-transcription-polymerase chain reaction (RT-PCR) from mouse placenta total RNA (Clontech, Palo Alto, CA). The primers for the PCR amplification using Platinum Pfx DNA polymerase (Invitrogen/Life Sciences) were 5'-ACGTagatctGCCAACAACAATCTGAACCGC-3' (forward) and 5'-GCGTGAgtcgacTCATTTTTCAGAGTCCGAGAAGCG-3' (reverse); the lowercase letters indicate the restriction sites for BglII and SalI, respectively. These primer sequences were based on mouse genomic DNA sequences surrounding the open reading frame between bp 195 110 and bp 192 936 (reverse complement) of GenBank accession number NT_039706.2. The resulting 2.18-kb amplification product was blunt cloned into the pCR-Script vector (Stratagene, La Jolla, CA), and the complete sequence was determined by dRhodamine dye terminator cycle sequencing. This plasmid was designated pZfp36l3-ORF.
Plasmid pL3-probe was created by PCR amplifying bp 194 350– 192 933 (from NT_039706.2) from plasmid pZfp36l3-ORF. The resulting 1.42-kb amplification product was blunt cloned into the pCR-Script vector, and the complete sequence was determined by dRhodamine dye terminator cycle sequencing.
Plasmid pl3-int was created by PCR amplifying bp 927 986–927 204 (from NT_039706.3) from plasmid green fluorescent protein (pGFP)-Zfp36l3 (see the following discussion), using the forward primer 5'-GTGCCTgctagcTATATGGTTGGTCCTGCCCTTACTCCTGGTG- 3' and the reverse primer 5'-ATGACGctcgagACCTGGAGCCAATATGCCGGTAGAG-3', where the lowercase bases were designed to introduce NheI and XhoI sites, respectively, into the resulting amplification product. The resulting 782-b amplification product was cloned into the modified pSK- vector, FactorXa/SK- (Stratagene); the correct sequence was confirmed by dRhodamine dye terminator cycle sequencing.
Plasmids pGFP-TTP, pGFP-TTP.C147R, and pGFP-Zfp36l1 have been described [23, 24]. Plasmid pGFP-Zfp36l3 was created by excising the 2.18-kb BglII-SalI fragment from pZfp36l3-ORF and ligating it into pEGFPC1 (Clontech), which had been digested with BglII and SalI. pGFP-Zfp36l3 encodes the full-length mouse ZFP36L3 protein, amino acids 1– 725, linked at the amino terminus in frame with GFP. A hybridization probe for blotting was created by digesting pGFP-Zfp36l3 with BglII and SalI to release the entire 2.18-kb coding insert, which was then labeled by random priming with 
32P-dCTP.
Plasmid pMLP.TNF-3'UTR contains the mouse MARCKS-like protein (MLP) promoter and coding region fused to the tumor necrosis factor -3'-untranslated region (TNF-3'UTR) (data not shown).
Cell Culture and Transfections
Human embryonic kidney 293 cells were maintained in Dulbecco modified Eagle medium (Life Technologies, Inc., Gaithersburg, MD), supplemented with 10% (v/v) fetal bovine serum, 100 U/ml of penicillin, and 100 µg/ml of streptomycin. Transient transfections were performed using the calcium phosphate method, exactly as described previously [24].
Northern Blot Analysis
Total cellular RNA was isolated from mouse tissues or transfected 293 cells using the RNEasy system (Qiagen, Valencia, CA), according to the manufacturer's instructions. All animal experiments conformed to the National Institute of Environmental Health Sciences guidelines for animal care. Northern blots of mouse placenta stages, mouse conceptus stages, and mouse conceptus tissues were purchased from SeeGene (Seoul, Korea). The remaining Northern blots were prepared as described elsewhere [24] and hybridized to random primed, 32P-labeled probes. The probes used for detecting Zfp36l3 transcripts consisted of the inserts released from the indicated plasmids. The probe used for detecting Zfp36l1 transcripts consisted of nucleotides 920–1320 of GenBank accession number X52590. The probe used for detecting GFP, alone or in fusion transcripts, was EGFP released from pEGFPC1. Blots were also hybridized with a random primed, 32P-labeled Gapd cDNA probe (described in Carballo et al. [25]) to monitor gel loading.
In Situ Hybridization Histochemistry
In situ hybridization was performed essentially as described [26]. Briefly, placentas from pregnant C57Bl6J mice were removed and frozen in isopentane at –30°C for 10–15 sec, followed by several minutes in powdered dry ice. Ten-micron sections were cut on a cryostat and mounted onto aminopropyltriethoxys (AAS) subbed slides. All sections for a given probe were cut at the same thickness. The sections were dehydrated at 25°C for several minutes and then frozen at –20°C until used for hybridization. Oligonucleotide DNA probes (48 bp) were synthesized (Integrated DNA Technologies, Coralville, IA) with the antisense probe designed to be complementary to sequences from the Zfp36l3 transcript; a sense probe was also synthesized that corresponded to the same portion of the mRNA. The antisense probe was complementary to bp 195 060–195 107 of GenBank accession number NT_039706.2; this sequence corresponds to the extreme amino terminus of mouse ZFP36L3. This oligonucleotide was 60% identical to the sequence of the Zfp36l2 transcript and 56% identical to the Zfp36l1 transcript, with no similarities to the sequence of Zfp36. A probe for mouse Gapd mRNA was used as a positive tissue quality control; this was complementary to bp 169–216 of GenBank accession number NM_008084.1.
Each DNA probe had a G:C ratio of approximately 50%. Probes were 3' end-labeled with 35S-dATP (specific activity >1000 Ci/mM, Dupont-NEN) using terminal transferase (Roche, Mannheim, Germany, or Fermentas, Hanover, MD). Incorporation of radioactivity for each probe was measured and was at least 50% of the total radioactivity. Specific activity was determined, and labeled probes were purified with phenol/chloroform extraction and precipitation. Probes were used within 24 h of preparation. The specific activities of the probes used in this experiment were 92 000– 118 000 cpm/pmol.
Hybridization conditions were identical for each antisense/sense probe pair and the housekeeping gene probe. Hybridization and wash temperatures were determined empirically by comparing the strength of the hybridization signal following three incubation and three wash temperatures for each probe. Optimal conditions resulted in the highest hybridization signal for the antisense probe with the least nonspecific signal for the sense probe. Optimal conditions for the Zfp36l3 antisense probe were 33°C for the incubation and 36°C for the first washing procedure.
Sections were fixed for 5 min in 4% (w/v) formaldehyde in phosphate-buffered saline (PBS) at room temperature, washed three times for 10 sec each in fresh PBS, acetylated in 0.1 M triethanolamine with 0.25% (w/v) acetic anhydride, and finally dehydrated through graded ethanol solutions. The sections were exposed to labeled probes in a solution containing 50% (w/v) formamide, 4x SSC, sheared single-stranded DNA (0.5 mg/ml), yeast tRNA (0.25 mg/ml), Denhardt solution (1x), dextran sulfate (10%, w/v), and DTT (100 mM). A total of 1 x 106 cpm was applied to each section. Sections were hybridized overnight at 33°C, washed in 2x SSC containing 50% (w/v) formamide for 1 h at 36°C and then washed in 1x SSC for 1 h at room temperature before dehydration and drying. The slides and a previously calibrated standard were exposed to a phosphor screen (Amersham Biosciences, Piscataway, NJ) for 5 d, followed by imaging on an Amersham Biosciences Typhoon 9400 variable mode imager. Scan parameters were set for best resolution and a pixel size of 25 µm. The image file was viewed using ImageQuant software (Amersham) and exported as a TIF file. After the exposure, a subset of dried slides of interest was stained with cresyl violet using standard histochemical techniques.
For emulsion autoradiography, the remaining slides previously hybridized with labeled oligonucleotide probes were dipped in a photographic emulsion consisting of 50% (v/v) Kodak NTB2 emulsion and a 50% (v/ v) solution of 0.1% (w/v) Dreft detergent. The emulsion was kept at 42– 44°C in a darkroom. After slides were dipped, they were allowed to dry 2 h in the dark at room temperature, then placed in a sealed slide box with a desiccant and protected from light. The slides were stored at 4°C until development, in this case after 10 wk. Slides were then developed in Kodak D-19 for 2 min, stopped in distilled water for 15 sec, and fixed in Kodak rapid fix without hardener for 2 min, all at 16°C. Slides were then washed in running tap water for 20 min and immediately counterstained with cresyl violet and coverslipped using standard techniques.
Duration of exposure was based on the level of expression as determined by autoradiography. Typical exposures ranged from 3 wk to several months; multiple exposures were done for each experiment.
Preparation of Protein Extracts for Deadenylation Assays and Immunoblotting
Total cellular protein was isolated from transfected 293 cells as follows: Cells were washed twice with ice-cold PBS and lysed for 5 min on ice in an ice-cold lysis buffer (10 mM Hepes [pH 7.6], 40 mM KCl, 100 mM NaCl, 50 mM NaF, 0.25% [v/v] Nonidet-P40, 5% [w/v] glycerol, 0.5 mM phenyl methylsulfoxyl fluoride, 2 µg/ml leupeptin, and 1 µg/ml pepstatin). Cell rupture and the release of nuclei were monitored microscopically. Cell nuclei and debris were removed by centrifugation at 12 000 x g for 15 min at 4°C. Fifty or 30 µg of total protein were subjected to SDS-polyacrylamide gel electrophoresis (PAGE) on 10% (w/v) acrylamide or 4–12% (w/v) acrylamide gradient gels and transferred to nitrocellulose. Western blotting was performed as described [24], using the anti-GFP polyclonal Living Color Antibody (CLONTECH) at a dilution of 1/5000 or a rabbit polyclonal antibody directed at a carboxyl terminal peptide of ZFP36L2 [27]. For Western blots of tissue extracts, four placentas at E14.5 were collected from pregnant female mice and kept in RNAlater (Ambion Inc., Austin, TX) at 4°C. Dissected spleens from adult mice were rapidly frozen and kept at –20°C. Each organ was individually pulverized in liquid nitrogen, and protein extracts were made as previously described [27]. Three hundred micrograms of total protein were loaded per lane in a 10% SDS-PAGE gel, and the gel was transferred to a nitrocellulose membrane. Western blots were performed using the ZFP36L2 carboxyl-terminal peptide antiserum [27], alone or in the presence of the competing peptide (10 µg/ml), or with the preimmune serum at the same dilution (1: 2500).
Fluorescence Microscopy
To observe cells expressing GFP-fusion proteins, 293 cells grown on six-well plates (Falcon; Becton-Dickinson) were transfected with 0.8 ng of plasmid encoding GFP-TTP or GFP-ZFP36L3 fusion proteins. In all cases, pPS+ vector DNA (Stratagene) was added to bring the total transfected DNA to 750 ng. At 18–24 h after transfection, living cells were observed and images captured with an Olympus IX70 fluorescence microscope in either the presence or the absence of leptomycin B (Sigma-Aldrich) at 10 ng/ml final concentration in 0.07% (final concentration; v/v) methanol. An equal amount of methanol was used for the controls.
Analysis of RNA Binding
RNA mobility gel shift assays were performed essentially as described previously [6, 7]. Briefly, 5 µg of 293 protein extract were prepared as described previously and incubated with 32P-labeled TNF ARE RNA probe, and reactions were analyzed on 4% (w/v) nondenaturing polyacrylamide gels followed by autoradiography. The probes were digested with RNAse T1 prior to the gel shift assay so that the poly(A) tail and vector sequences were cleaved from the probe before it was allowed to bind to proteins.
In Vitro Deadenylation Assay
Transfection of 293 cells, preparation of protein extracts, and performance of the assay were performed exactly as described previously [28]. Briefly, 10 µg of protein from each extract from cells transfected with plasmids encoding various GFP fusion proteins were incubated with different 32P-labeled TNF ARE RNA probes as indicated in the figure legend. As described in [28], probe ARE contained bases corresponding to bp 1309–1332 of the cDNA for TNF (GenBank accession number X02611), probe ARE-A50 contained bases corresponding to bp 1309–1332 of the cDNA for TNF (GenBank accession number X02611) followed by 50 As, probe V contained 58 b transcribed from the multiple cloning site of vector SK- (Stratagene), and probe A50 contained the same 58 b followed by 50 As. After 1 h at 37°C, the reactions were terminated by the addition of 0.2 M EDTA to a final concentration of 20 mM, phenol extracted and analyzed by electrophoresis using 6% (w/v) polyacrylamide gels containing 7 M urea, followed by autoradiography.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Searches of the NCBI mouse genome database with the TZF domain of mouse TTP (GenBank accession number NP_035886.1; amino acids 95–158) revealed the expected matches with the genes encoding mouse TTP (Zfp36; chromosome 7; 2e-35), Zfp36l1 (chromosome 12; 2e-23), and Zfp36l2 (chromosome 17; 1e-23) as well as with a novel TZF domain sequence on the mouse X chromosome (currently corresponding to GenBank accession number NT_039706.3; bp 928 712–928 521; 1e-21). This predicted amino acid sequence was then used to probe the nr database, which revealed a match to a hypothetical mouse cDNA and protein (GenBank accession numbers XM_285657 and XP_285657, respectively). Although the initial hypothetical versions of these sequences predicted an intron of varying sizes, the most recent version of this predicted mRNA sequence (XM_285657.3; released August 31, 2004) contained no intron, with a continuous open reading frame predicting a protein of 725 amino acids. We found that the mRNA size proposed by XM_285657.3 closely paralleled the observed mRNA size and found no evidence for an intron (see the following discussion). In addition, some of the putative intron sequences were also contained in mouse ESTs that contained identical predicted amino acid sequences to portions of the protein (see the following discussion). Therefore, our data support the most recent NCBI release; that is, the encoded protein corresponds to a single, continuous open reading frame between bp 929 074 and 926 901 of NT_039706.3. We have named this protein zinc finger protein 36-like 3 (ZFP36L3), with the mouse gene name Zfp36l3, to correspond to the existing nomenclature for the other Zfp36 family members. This gene symbol was approved by the Mouse Genomic Nomenclature Committee.
On November 2, 2004, the predicted ZFP36L3 protein sequence was reflected in the mouse GenBank EST database in 11 ESTs. Together, these spanned all the proposed protein sequence, and three of these ESTs (GenBank accession numbers CF551613, CF617717 and CK030738) contained sequences corresponding to the original proposed intron sequences now excluded from XM_285657.3. All the identified mouse ESTs were cloned from cDNA libraries that contained mouse placenta or extraembryonic membranes.
We also used the mouse ZFP36L3 protein sequence to screen the draft rat genome and existing rat ESTs. The mouse protein identified an apparently orthologous protein sequence in the database (XP_228661.2); the 722 amino acid hypothetical protein was 67% identical to the mouse protein. It corresponded to a sequence on the rat X chromosome (bp 485 913–488 078 of NW_048051.1; see the following discussion). Blasting the rat EST database with the proposed rat protein sequence (XP228661.2) yielded three ESTs; all these were also from placenta libraries.
To date, we have not identified orthologous predicted protein sequences in any other mammalian species. This includes both humans and chimpanzees, whose X chromosomes were searched carefully for orthologous genes and pseudogenes.
Cloning and Characterization of a Mouse Zfp36l3 cDNA
RT-PCR using primers flanking the predicted Zfp36l3 open reading frame, as discussed previously, yielded a single amplification product of 2.18 kb, appearing exclusively in RT-PCR reactions using placenta RNA (data not shown). The sequence of this PCR product was identical to that of the open reading frame formed between bp 929 074 and 926 901 in the genomic contig NT_039706.3. This cDNA sequence has been deposited in GenBank (accession number AY661338).
The Zfp36l3 cDNA encoded a 725 amino acid protein with predicted Mr of 72 346, pI 5.55. It contained a single TTP-like TZF domain spanning amino acids 121–184 and across this region shared 76% amino acid identity with mouse TTP. An alignment of the TZF domain of ZFP36L3 with mouse TTP, ZFP36L1, and ZFP36L2 is shown Figure 1A and demonstrates the great similarity among all four family members in this region, including the two arginines previously reported to be necessary for nuclear import of TTP (Fig. 1A, arrowheads [29]). A significant difference between ZFP36L3 and the other family members was the presence of a proline instead of a leucine in the lead-in sequence to the second zinc finger (Fig. 1A, arrow). Otherwise, other than the TZF domain, there was little sequence similarity between ZFP36L3 and the other family members, except for a small region at the extreme carboxyl-terminus (Fig. 1B). In this figure, the arrow points to an aspartate in ZFP36L3, instead of a branched chain amino acid, that may prevent this region from being an effective nuclear export sequence [24, 30]. The peptide in ZFP36L2 that served as the antigen for an antibody that cross-reacts with ZFP36L3 is underlined (see the following discussion).
|
Another remarkable feature of this protein sequence is the presence in the carboxyl-terminal half a repeated element, which in the mouse consists of 10 consecutive repeats of the sequence AAMAPGAALAPAAALTPA or a close variant and in the proposed rat sequence 11 repeats of the sequence AALVPGAAMAPGTALAPGAA or a variant. In both species, these regions of the proteins can also be viewed as sequences of 38 consecutive repeats of the six amino acid motif GAALAP or a related variant. These large regions of both proteins are predicted by various secondary structure algorithms to contain hydrophobic alpha helices, suggesting potential involvement in membrane association or other subcellular localization. For example, the program TMPred (http://www.ch.embnet.org/software/TMPRED_form.html) predicts seven strong transmembrane helices, with the amino terminus inside the cell.
Expression of Zfp36l3 Transcripts in Mouse Tissues
Northern blot analysis of the Zfp36l3 transcripts, using total RNA from a panel of mouse and human tissues, showed that a 
2.4-kb Zfp36l3 transcript was expressed solely in mouse placenta (lane 9 in Fig. 2A) and was not detected in any of the other mouse tissues tested or in human placenta. Northern blotting was also performed on blots containing RNA from various mouse conceptus stages and tissues. Zfp36l3 mRNA was barely detectable in RNA from the total mouse conceptus starting at Embryonic Day (E) 8.5, reached peak levels between E13.5 and E14.5, and remained readily detectable through E18.5 (Fig. 2B, top panel). In contrast, Zfp36l1 mRNA was detected in total conceptus RNA from all stages tested (Fig. 2B, middle panel). Analysis of various conceptus tissues at E18.5 revealed that Zfp36l3 expression was limited to extraembryonic structures; it was detectable in placenta and yolk sac/amnion but not in the fetus (Fig. 3A). In contrast, Zfp36l1 mRNA was detected in all conceptus tissues tested (Fig. 3A, middle panel).
|
|
As discussed previously, the sequence of the Zfp36l3 RT-PCR product and the corresponding genomic sequence include a region initially predicted to encode an intron of 332 bp (bp 927 658–927 326 of GenBank accession number NT_039706.3) as reflected in the "spliced" proposed cDNA sequence of accession numbers XM_285657.1 and XM_285657.2. To determine whether this region of the mRNA was expressed in placenta RNA, a probe corresponding to this putative intron only was used in Northern blotting of mouse placenta RNA at mixed gestational ages. As shown in Figure 3B, this probe hybridized to the same single transcript, of approximately 2.4 kb, as did the longer cDNA Zfp36l3 probe (Fig. 3B). In neither case was there evidence for the existence of a shorter transcript, even after longer exposure of the Northern blot.
Expression of Zfp36l3 Transcripts Within the Placenta
We used ISHH with 35S-labeled DNA oligonucleotide probes to determine the predominant sites of expression within the mouse placenta at E14.5, when placental expression appeared to be maximal. Sections were evaluated initially by phosphor screen at low magnification to confirm probe specificity. The antisense probe hybridized predominantly to the central portion of the placenta, corresponding to the trophoblastic placenta under light microscopy, whereas an outer rim of tissue that corresponded to maternal decidua was essentially negative (Fig. 4A). A sense probe, labeled to approximately the same specific activity, yielded minimal signal when hybridized to an adjacent placenta section (Fig. 4B), whereas an antisense Gapd probe showed a strong signal with all placental layers (Fig. 4C). A Zfp36l3 antisense probe showed no specific hybridization to a kidney section (Fig. 4D), whereas the kidney section was readily recognized by the Gapd probe (Fig. 4E). In Fig. 4, A–E, all sections were exposed to the phosphor screen for the same time in the same cassette, and image adjustments were identical for each section. Similar results were obtained in placentas from E10.5 and E12.5 (not shown). These studies confirmed the specificity of the antisense probes used; we therefore used emulsion autoradiography with the same probes to evaluate regional expression of Zfp36l3 transcripts within the placenta at E10.5, E12.5, and E14.5.
|
In a coronal section of placenta at E14.5 evaluated by emulsion autoradiography, there was minimal expression of the Zfp36l3 transcript in the maternal decidua (Fig. 4F) and the allantois (Fig. 4J). There was moderate expression in the trophoblast giant cells (Fig. 4G), higher levels of expression in the spongiotrophoblast cells (Fig. 4H), and highest expression in the cells of the labyrinthine layer (Fig. 4I). All cell types in the trophoblast layers, with the exception of fetal and maternal red cells, appeared to be labeled under these conditions. Similar results were obtained in sections from placentas at E10.5, E12.5, and E14.5 based on the evaluation of multiple sections at each time point.
Expression of ZFP36L3 Protein
Because of the similarity of the carboxyl-terminal sequences between ZFP36L2 and ZFP36l3 (see Fig. 1B), we tested an antibody raised to this peptide from ZFP36L2 for reactivity against the ZFP36L3 protein, assuming that their large difference in molecular weight would permit differentiation on Western blots. The antibody directed at the carboxyl-terminus of ZFP36L2 (anti-L2) recognized a major band of approximately Mr 150 000 in 293 cells expressing the GFP-ZFP36L3 fusion protein as well as a presumed fragment of lower apparent molecular weight (Fig. 5A). These bands were not seen in cells expressing GFP alone and were not recognized by the appropriate preimmune serum in a parallel blot (Fig. 5A). The same upper band was recognized by an anti-GFP antibody, although the lower band was not, suggesting that the smaller species was missing the amino-terminal site of fusion with GFP (Fig. 5B). Both antibodies also recognized expressed GFP-ZFP36L2, but only the GFP antibody recognized GFP alone (Fig. 5B).
|
We then tested whether this antibody could cross react with endogenous ZFP36L3 in normal mouse placenta. Under the same blotting conditions, the anti-L2 antibody recognized a prominent band of Mr 115 000 that was not seen in a spleen extract that was probed in parallel (Fig. 5C, left panel). This placenta band was not seen with the preimmune serum in a parallel blot (Fig. 5C, middle panel), and no signal was seen when the antibody incubation was performed in the presence of the immunizing ZFP36L2 peptide (Fig. 5C, right panel). Although the predicted size of the protein is Mr 72 000, other members of this protein family migrate at anomalously high apparent Mr on SDS polyacrylamide gels, and the value of Mr 115 000 corresponds well with the Mr 150 000 seen with the GFP-ZFP36L3 fusion protein. We speculate that the large, alanine-rich repeat domain described previously may contribute to the proposed anomalous migration. However, these data suggest strongly that the ZFP36L3 protein as well as its mRNA are expressed in normal placenta. Further experiments to confirm this identification are in progress.
Evidence Against Nuclear-Cytoplasmic Shuttling of ZFP36L3
TTP, ZFP36L1, and ZFP36L2 are all nuclear-cytoplasmic shuttling proteins that rely on the export receptor CRM1 for nuclear export [24]. To determine whether this is also true for ZFP36L3, 293 cells were transfected with plasmids encoding GFP fusions with TTP or ZFP36L3, treated with either carrier or the CRM1 inhibitor leptomycin B (LMB), and then living cells were examined by fluorescence microscopy (Fig. 6). In the absence of LMB, both GFP-TTP and GFP-ZFP36L3 appeared to be largely if not exclusively cytoplasmic (Fig. 6, A, C, E, and G). However, in the presence of LMB for either 2 or 4 h, GFP-TTP was almost completely localized to the nucleus (Fig. 6, F and H), whereas GFP-ZFP36L3 remained exclusively cytoplasmic (Fig. 6, B and D). This is evidence against ZFP36L3 shuttling between the nucleus and cytosol and is different from the behavior of all three other known family members in this assay [24].
|
Binding of ZFP36L3 to the TNF ARE
TTP, ZFP36L1, and ZFP36L2 can each bind to the ARE found in TNF, granulocyte-macrophage colony-stimulating factor (GM-CSF), and interleukin-3 (IL-3) mRNA 3'-UTRs [7], and their ability to do so requires an intact TZF domain. To test whether ZFP36L3 shares this RNA binding ability, 293 cells were transfected with plasmids encoding amino-terminal GFP fusions with mouse TTP and ZFP36L3. Cell extracts were used in RNA electrophoretic mobility shift assays as previously described [6, 7]. As shown in Figure 7, GFP-TTP and GFP-ZFP36L3 could each bind to the TNF ARE probe, whereas GFP alone did not alter the binding of endogenous 293 cell proteins to the probe.
|
Effect of ZFP36L3 on the Degradation of ARE-Containing mRNA
In cotransfection studies, reporter mRNA transcripts containing an ARE from TNF, GM-CSF, or IL-3 become deadenylated and degraded in the presence of TTP and related proteins [6, 7], and their ability to do so requires an intact TZF domain. To test whether ZFP36L3 shares this ability to promote degradation of target ARE-containing mRNA, a cotransfection system was used. Briefly, 293 cells were cotransfected with expression vectors encoding a reporter gene and GFP fusion constructs with either TTP or ZFP36L3. Levels of the reporter mRNA were then analyzed by Northern blotting. The reporter construct in this case was pMLP.TNF-3'UTR, in which the stable mouse MLP mRNA was modified by the presence of 517 b of the TNF mRNA ARE in the 3'-UTR. This fusion mRNA is degraded in the presence of TTP, ZFP36L1, and ZFP36L2 (data not shown). As shown in Figure 8A (upper panel), in the presence of GFP-TTP or GFP-ZFP36L3, the reporter mRNA was markedly degraded, with modest accumulation of the deadenylated probe fragment. Expression of the various GFP fusion transcripts from different amounts of transfected plasmid DNA was confirmed by Northern blotting using a probe for the GFP component of the fusion transcripts (Fig. 8A, lower panel).
|
We recently showed that TTP, ZFP36L1, and ZFP36L2 could promote the deadenylation of ARE-containing RNA probes in a cell-free system and that the ability to do so was dependent on an intact TZF domain [28]. To test whether ZFP36L3 shared this deadenylating ability, extracts from transfected 293 cells were incubated with 32P-labeled, single-stranded RNA probes that contained or lacked a portion of the TNF ARE, with (ARE-A50) or without (ARE) a 50-bp poly(A) tail. We also used probes containing vector sequences linked to a poly(A) tail without an ARE (A50) and vector sequences alone (V). Details of these probes have been described [28]. The amounts and sizes of the reaction products were analyzed by electrophoresis and autoradiography. As shown in Figure 8B (upper panel), in the presence of either GFP-TTP (lanes 7 and 8) or GFP-ZFP36L3 (lanes 9 and 10), there was a decrease in the amount of the ARE-poly(A)-containing probe and an increase in the accumulation of the ARE probe lacking its poly(A) tail (arrow) when compared to the presence of GFP alone (lane 6). An additional control for the GFP-alone sample was an identical incubation in the presence of EDTA (lane 5), which inhibits endogenous deadenylating enzymes such as the poly(A) exonuclease PARN. The ability of the protein-containing extracts to deadenylate the probes was dependent on the presence of the ARE since no such probe degradation occurred when probes containing vector sequences were used that were linked to an ARE alone (Fig. 8B, upper panel, lanes 1–4), linked to a poly(A) tail without an ARE (Fig. 8B, lower panel, lanes 15–20), or vector sequences alone (Fig. 8B, lower panel, lanes 11– 14). Expression of the various GFP fusion proteins was confirmed by immunoblotting using an antibody to GFP (data not shown).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The ZFP36L3 protein has a predicted Mr 72 345 and pI 5.39, in contrast to the smaller sizes and basic pI of the other mammalian family members. This protein could be readily expressed as an amino terminal GFP fusion protein in 293 cells, in which it appeared to be largely if not exclusively cytosolic. We also demonstrated by cross-reactivity with an antibody to the closely related carboxyl-terminal sequence of ZFP36L2 that the endogenous protein apparently can be detected in protein extracts from mouse placenta. This protein exhibits several differences in properties when compared to the other three TTP family members, in addition to its larger predicted size and its acidic pI. A striking difference is the presence in the carboxyl-terminal half of ZFP36L3 of a long repeated element predicted to represent hydrophobic helices. Another difference lies in the TZF domains, which in ZFP36L3 from both mouse and rat contain a proline in the lead-in sequence to the second zinc finger, that is, KYKTEP instead of KYKTEL in the other proteins, and RYKTEL in all four members leading into the first zinc finger. This proline would not be predicted to have major effects on the structure of the TZF domain when fit into the coordinates of the human ZFP36L2 (TIS11D) TZF domain [31]. In addition, a similar proline is present leading into the second zinc finger of the TZF domain in a Danio rerio protein (GenBank accession number NP_996938), whose closest mammalian orthologue is mouse ZFP36L2; similar proline-containing sequences are found in translations of ESTs from other fish (e.g., salmon, CK881669; Tetraodon nigroviridis, CAG10191.
Outside the TZF region there was little similarity to the other three family members, with the exception of the extreme carboxyl-terminus. Earlier work demonstrated that TTP, ZFP36L1, and ZFP36L2 can shuttle between the nucleus and cytoplasm and accumulate in the nucleus in cells treated with leptomycin B (LMB), an inhibitor of the export receptor CRM1 [24]. ZFP36L1 and ZFP36L2 each contain a functional carboxyl-terminal leucine-rich nuclear export sequence (NES), required for both binding to CRM1 and nuclear export [24]. This region is highly conserved in ZFP36L3 between amino acids 712 and 725. However, in ZFP36L3 the fourth hydrophobic residue of the NES, isoleucine in ZFP36L1 and ZFP36L2, is changed to aspartate in both the mouse and the rat sequences, which should preclude binding to CRM1 [30, 32].
The nuclear import of TTP, ZFP36L1, and ZFP36L2 may depend on conserved arginine residues in the TZF domain [24, 29], which are also conserved in the TZF domain of ZFP36L3; however, it is not yet known whether the ZFP36L3 TZF domain alone can mediate nuclear import. It is possible that nuclear import of ZFP36L3 is blocked by some unknown mechanism, such as membrane association, or even by the large size of the protein. Since the nucleocytoplasmic shuttling of TTP may represent a means of regulating TTP's activity [24, 33, 34], the apparent lack of ZFP36L3 shuttling may prove to be another difference from the other family members in the regulation of its cellular activity.
Another important difference between ZFP36L3 and the other TTP family members is its apparent species specificity. Whereas orthologues for TTP, ZFP36L1, and ZFP36L2 genes have been found in many mammals and in Xenopus [11], to date ZFP36L3 has been found only in the mouse and rat. Database searches of the human and chimpanzee genomes, as well as of EST collections from all available species, have not identified ZFP36L3 outside the murine lineage. In addition, we have not detected Zfp36l3 by Southern blotting of human genomic DNA or its transcript by Northern blotting of guinea pig, hamster, sheep, or human RNA (R.S. Phillips and P.J. Blackshear, unpublished data). According to a recent analysis of the rat genome [35], only 31 genes were identified that were expressed in both the mouse and the rat but not in humans, and only 10 of these lacked orthologues in other mammals. Mechanisms proposed for these rodent-specific genes included rapid evolution of the genes, so that human orthologues might not be apparent; genes arising de novo from noncoding DNA; or conversion to pseudogenes in the human genome. As stated by Gibbs et al. [35], "The paucity of rodent-specific genes indicates that de novo invention of complete genes in rodents is rare." The determination of whether Zfp36l3 is truly rodent specific or possibly even murine specific will require further study.
Tissue-specific expression of Zfp36l3 mRNA was also much more highly restricted than those of Zfp36, Zfp36l1, and Zfp36l2. TTP is expressed in most tissues, with highest levels in the spleen [1], whereas Zfp36l1 and Zfp36l2 are virtually ubiquitously expressed. In contrast, Zfp36l3 transcripts were detected almost exclusively in the placenta after approximately E9.5. The highest levels of transcript expression were seen in essentially all the cells of the labyrinthine layer, with slightly lower apparent expression in the spongiotrophoblast and giant cell layers and minimal detectable expression in the allantois and the maternal decidua. We predict that ZFP36L3 will be found to play a role in the destabilization of one or more transcripts in the trophoblastic layers of mouse and rat placenta; the identification of these putative mRNA targets will require future studies. Some of these are under way, including attempts to disrupt Zfp36l3 and to determine its putative mRNA targets by immunoprecipitation and identification of associated mRNAs.
Despite these many differences between ZFP36L3 and the other members of the TTP protein family, we found that ZFP36L3 behaved indistinguishably from TTP, ZFP36L1, and ZFP36L2 in assays of RNA binding in a cell-free system, promoting decay of an ARE-containing target RNA in intact cell transfection experiments [7, 9] and promoting ARE-dependent deadenylation in cell-free assays [28]. Since the physiological role of TTP in the mouse includes promoting the down-regulation of the ARE-containing mRNAs encoding TNF and GM-CSF [25, 36], we predict that ZFP36L3 will play a similar role as an mRNA-destabilizing protein, with mRNA targets in the placenta. Although TTP remains the only CCCH protein whose physiological target mRNAs have been identified, recent studies of ZFP36L1 and ZFP36L2 in mouse physiology suggest a reproductive role for these proteins. For example, disruption of Zfp36l1 resulted in embryonic lethality at approximately E9.5, due in most cases to a failure of chorioallantoic fusion to form a functional placenta [37]; in this case, expression of the transcript was highest in the allantois, in contrast to the apparent high-level expression of Zfp36l3 in the labyrinthine layer and chorionic plate. Disruption of Zfp36l2, leading to decreased expression of an amino-terminal truncated protein, led to complete female infertility [27], with the defect apparently in the further cell division of the two-cell stage embryo. The data presented here suggest that ZFP36L3 could function in an analogous manner to TTP, playing a role in the down-regulation of yet unidentified placenta-specific target mRNAs. If this is true, one could speculate that one of the other family members assumes this role in the placental physiology of nonmurine mammals.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received: 1 February 2005.
First decision: 18 February 2005.
Accepted: 21 March 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. J. Horner, W. S. Lai, D. J. Stumpo, and P. J. Blackshear Stimulation of Polo-Like Kinase 3 mRNA Decay by Tristetraprolin Mol. Cell. Biol., April 15, 2009; 29(8): 1999 - 2010. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. D. Frederick, S. B. V. Ramos, and P. J. Blackshear A Unique C-terminal Repeat Domain Maintains the Cytosolic Localization of the Placenta-specific Tristetraprolin Family Member ZFP36L3 J. Biol. Chem., May 23, 2008; 283(21): 14792 - 14800. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Cao, J. F. Urban Jr, and R. A. Anderson Cinnamon Polyphenol Extract Affects Immune Responses by Regulating Anti- and Proinflammatory and Glucose Transporter Gene Expression in Mouse Macrophages J. Nutr., May 1, 2008; 138(5): 833 - 840. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. S. Lai, J. S. Parker, S. F. Grissom, D. J. Stumpo, and P. J. Blackshear Novel mRNA Targets for Tristetraprolin (TTP) Identified by Global Analysis of Stabilized Transcripts in TTP-Deficient Fibroblasts Mol. Cell. Biol., December 15, 2006; 26(24): 9196 - 9208. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Benjamin, M. Schmidlin, L. Min, B. Gross, and C. Moroni BRF1 Protein Turnover and mRNA Decay Activity Are Regulated by Protein Kinase B at the Same Phosphorylation Sites Mol. Cell. Biol., December 15, 2006; 26(24): 9497 - 9507. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
