Negative Control of the Poly(A)-binding Protein mRNA Translation Is Mediated by the Adenine-rich Region of Its 5 * -Untranslated Region*

Translation of the mRNA for the poly(A)-binding protein (PABP) may be autoregulated by the binding of PABP to the A-rich segment of its 5 * -untranslated region (UTR). To test this hypothesis, we examined the effect of different fragments of the 5 * -UTR from human PABP cDNA on the translation of the b -galactosidase ( b -Gal) reporter gene. Presence of the A-rich sequence from the 5 * -UTR of PABP mRNA inhibited expression of the chimeric b -Gal gene in transfected HeLa cells. The differences in expression of b -Gal polypeptide was due to the translational repression of b -Gal mRNA containing the A-rich 5 * -UTR of PABP mRNA. The A-rich region of the 5 * -UTR located within nucleotides 58–146 of PABP mRNA was sufficient to mediate translational control of this mRNA expression. We also examined the effect of overexpression of PABP mRNA in HeLa cells. The ectopic PABP mRNA without the A-rich 5 * -UTR region was translated efficiently, whereas the translation of the endogenous PABP mRNA was substantially inhibited in the transfected cells. In contrast, the ectopic PABP mRNA containing the A-rich 5 * -UTR region did not show similar effect on the translation of the endogenous PABP mRNA in these cells. These results suggest that feedback control of mRNA translation is involved in regulating PABP expression in HeLa cells.

Cellular mRNAs are often complexed with a group of RNAbinding proteins. One of the best studied RNA-binding proteins is the 72-kDa poly(A)-binding protein (PABP), 1 that shows specific interaction with the 3Ј poly(A) region of all eukaryotic mRNAs (1)(2)(3). PABP is ubiquitous and highly conserved in eukaryotic cells. The poly(A)-binding region contains 90 amino acid domains (1)(2)(3). Plants and mammals both contain several related polypeptides with specific affinity toward the poly(A) region of the mRNA. These PABP-related polypeptides show tissue-specific expression and may be regulated by growth conditions (4,5). PABP is essential for viability of eukaryotic cells and is involved in regulating mRNA translation and stability. Results of both in vitro (6) and in vivo (7) studies suggest that the complex of PABP and poly(A) tail plays an important role in mRNA translation. It is believed that the poly(A)-PABP complex is involved in joining the 60 S subunit to the 48 S preinitiation complex (7). PABP also stimulates the binding of 40 S subunits to the mRNA (8). The ability of PABP to interact with the eIF4G and eIF4B is important for its role in the initiation of mRNA translation (9 -13). The interactions between PABP and these initiation factors also take place in absence of poly(A) (11). Therefore, it is not clear whether the free PABP or PABP-poly(A) complex is involved in mRNA translation in vivo. In addition, PABP is also involved in poly(A) tail shortening probably through its interaction with a poly(A)-specific 3Ј-exonuclease (12,14).
Several studies indicated that PABP expression is regulated at the post-transcriptional level during the cell cycle (15)(16)(17)(18)(19)(20). In resting cells, the majority of cellular PABP mRNA was found in the non-translated state. Following stimulation of growth by serum, a rapid transition of the mRNA to the translated state occurred in mouse fibroblasts (19,20). In contrast, during differentiation of murine myoblasts growth arrest resulted in the migration of repressed PABP mRNA to the translated state. This process was also accompanied by reduced stability of the PABP mRNA in differentiated myotubes resulting in little change in PABP synthesis (18).
Translation of a specific mRNA is often regulated by the interaction of regulatory proteins with defined sequences of the mRNA. For instance, feedback control of ribosomal protein L10 synthesis by the level of free L10 protein is mediated by its binding to the 5Ј-UTR of the L10 mRNA (21,22). Similarly, synthesis of ferritin is regulated in vertebrates by the iron level (23). The inhibition of ferritin mRNA translation requires the binding of a 98-kDa translation repressor polypeptide to the conserved 25-nucleotide-long sequence of its 5Ј-UTR (24,25). In addition the 3Ј-UTR of some mRNAs is involved in regulating translation. The translations of protamine (26) and myocyte enhancer factor-2a mRNAs are controlled by RNA-binding proteins that interact with sequences located within the 3Ј-UTR of these mRNA (27). How the 3Ј-end of these mRNAs influence ribosome binding at the 5Ј-end is unknown. It is possible that 3Ј-UTR binding protein(s) may interfere with the function of poly(A)-PABP complex.
In contrast to the examples cited above, regulation of PABP mRNA translation appears to be more subtle. Instead of completely blocking PABP mRNA translation, the distribution of this mRNA between translationally active and inactive populations is modulated. Interestingly, analysis of PABP mRNA sequences in various species showed the presence of conserved highly A-rich sequence in its 5Ј-UTR. This raises the possibility of negative feedback control of PABP mRNA translation. Previous studies from this and other laboratories have shown that PABP is able to bind to the 5Ј-UTR A-rich regions and inhibit PABP mRNA translation in a rabbit reticulocyte lysate cellfree system (28, 29), and removal of this sequence from the PABP mRNA enhances its translation in vitro. We have also shown that the translation of full-length PABP mRNA, but not the truncated mRNA from which the A-rich region of the 5Ј-UTR was removed, can be inhibited in a cell-free system by adding purified human PABP. Furthermore, by using UV-mediated cross-linking of RNA and proteins, we have shown that PABP binds to the first 224 nucleotides of the 5Ј-UTR of PABP mRNA, and this may repress its own expression (28). Therefore in the present study, we extended these investigations to determine whether similar inhibition of mRNA translation occurs in vivo by the A-rich 5Ј-UTR of PABP mRNA. We have also examined if PABP synthesis was regulated in HeLa cells when the attempt was made to increase PABP synthesis by ectopic expression of this mRNA.

EXPERIMENTAL PROCEDURES
Plasmid Construction-The parent plasmid pHu73 is a full-length cDNA clone of human PABP mRNA (2). The 487-bp EcoRI/BglI fragment of the 5Ј-UTR of this clone was used in PCR reactions with appropriate primers to produce amplified DNA consisting of different regions of the PABP mRNA 5Ј-UTR as shown in Fig. 1 (panel A). The PCR primers used in our studies also contained additional sequences to provide the NcoI and EcoRI sites at 5Ј-and 3Ј-ends, respectively, of the amplified products. The PCR reactions were carried out in the buffer containing 10 mM Tris-HCl, pH 9.0, 1.5 mM MgCl 2 , 50 mM KCl, 0.1% Triton X-100, and 200 M dNTPs for 35 cycles with 2 units of Taq DNA polymerase (Life Technologies, Inc.). Each cycle was at 94°C 15 s, 50°C 30 s, and 72°C 1 min, and the final cycle was at 72°C for 5 min. The amplified product of the correct size was purified from an agarose gel using the GeneClean kit (BIO/CAN, Montreal, Canada) and was digested with NcoI and EcoRI before cloning. The PCMV-SPORT-␤-Gal (Life Technologies, Inc.) vector was also digested with NcoI and EcoRI, and the 368-bp fragment was removed. The 5Ј-UTR of PABP mRNA was then ligated to the 7485-bp-long vector fragment that contained all the essential features for the expression of the reporter ␤-Gal gene. Additional cloning was performed by digesting the vector with SunI to generate a 285-bp deletion in the 3Ј-coding region to express a truncated ␤-Gal polypeptide as internal controls. The sequences of the 5Ј-UTR regions of all clones were examined by automated DNA sequencing.
For ectopic expression of human PABP cDNA, two plasmids were constructed. One of these plasmids (pCMV⌬PABP) was designed to lack the adenine-rich region from the 5Ј-UTR of human PABP cDNA clone pHU73. The second plasmid (pCMV89⌬PABP) was constructed by adding an 89-bp PCR-generated fragment ( Fig. 1, panel A) containing the A-rich 5Ј-UTR region to the first construct, the pCMV⌬PABP. The pCMV-SPORT-␤-Gal (Life Technologies, Inc.) was digested with BamHI and NcoI, and the plasmid backbone without the entire ␤-Galtranscribed region was isolated and ligated to the SacII and BamHI fragment of pHU73 (Fig. 1, panel B).
Transfection of Cells-Approximately 1 ϫ 10 6 HeLa W5 cells (ATCC) grown on a 12-well dish in Dulbecco's modified EagleЈs medium (Life Technologies, Inc.) supplemented with 10% fetal calf serum (Life Technologies, Inc.) were used for transfection. Subconfluent cultures were transfected in Opti-MEM medium (Life Technologies, Inc.) without serum and antibiotics. For transfection, 1.5 g of the reporter plasmid DNA was incubated with 10 g of LipofectAMINE in 400 l of Opti-MEM medium at 20°C for 30 min before being added to the cells. Cells were incubated with the DNA/liposome mixture for 5 h at 37°C, and then the DNA-containing medium was removed, and transfected cells were allowed to grow for 24 h in Dulbecco's modified EagleЈs medium with 10% fetal calf serum. In some experiments, 0.5 g of a second plasmid DNA coding for a truncated ␤-Gal polypeptide ( Fig. 1) was used as an internal control to monitor the transfection efficiencies between experiments.
Measurement of mRNA Levels-The transfected cells were washed three times with PBS and then lysed in 4 M guanidine thiocyanate (Fluka, Switzerland). Total RNA from the cells was isolated by phenol/ chloroform extraction of the lysate as described previously (31). The level of specific mRNAs was measured by reverse transcription (RT) and PCR. The RT step was performed with 200 units of M-MLV reverse transcriptase (Life Technologies, Inc.) in 50 mM Tris-HCl, pH 8.3, 75 mM KCl, 10 mM dithiothreitol, 3 mM MgCl 2 , containing 0.5 mM each of dNTPs. The reaction was performed at 37°C for 1 h. The cDNA was amplified by the PCR technique with Taq DNA polymerase for 25 cycles using mRNA specific primers (Fig. 1, panel A) as described earlier. The conditions used for RT-PCR gave linear increase in the amplified products with 0.25 to 2 g input of total cellular RNA. In some reactions additional bands were seen; however, their presence did not affect the linear response.
The level of ␤-Gal mRNA was also measured in some experiments by RNase protection (32) using the Hybspeed RPA Kit (Ambion; Austin, TX). The PCMV-SPORT-␤-Gal vector was linearized by digesting with Bpu11021 enzyme, and the 381 nucleotides long antisense RNA complementary to the 3Ј-UTR of the ␤-Gal mRNA was transcribed in vitro by using T7 polymerase. Similarly, a 374-nucleotide-long antisense RNA to the human GAPDH mRNA was produced by the T7 polymerase transcription of a commercially available clone (Ambion, TX). This antisense RNA was used in RNase protection assay of GAPDH mRNA, and due to the presence of genetically engineered additional sequences, produced a 316-nucleotide-long RNase protected fragment. Approximately 5 g of RNA was hybridized with 5 ϫ 10 5 cpm (approximately 5 ng) of the antisense RNA, and RNase-protected samples were analyzed by 2% agarose gel electrophoresis (33). The levels of ectopic and endogenous PABP mRNA were measured by S1 nuclease protection of specific oligodeoxynucleotide probes using the multi NPA kit from Ambion. For ectopic PABP mRNA detection a 36-mer oligonucleotide complementary to the vector-derived unique 3Ј-UTR of the ectopic PABP mRNA with 5 nucleotides mismatched at its 5Ј-end was used. The sequence of this oligomer is (5Ј-cgcatGCACGCGTAAGCTTGGGCCCCTCGAGACTCT-3Ј), complementary to nucleotides 680 -710 of the vector (Life Technologies, Inc.). For detecting endogenous PABP mRNA, a 30-mer oligonucleotide complementary to the 5Ј-UTR region of the endogenous PABP mRNA with 5 nucleotides mismatched at its 5Ј-ends was used. The selected 5Ј-UTR was deleted from the ectopic mRNA and therefore was specific for the endogenous mRNA. The sequence of this oligomer (5Ј-ggcggGCGGGTCGGTCTCGGCTGCCAGTAA-3Ј) is complementary to nucleotides 196 -220 of human PABP mRNA. The lowercase letters of both oligomers represent the mismatched bases. The levels of human GAPDH mRNA were measured in these studies using a commercially available 44-mer (Ambion) which gives a 30-mer S1 nuclease-protected fragment. These probes were 5Ј-end-labeled using [␥-32 P]ATP and T4 polynucleotide kinase (33). The radiolabeled probe (5 ng Ϸ5 ϫ 10 5 cpm) was hybridized to 5 g of cellular RNA and subjected to S1 nuclease treatment (31). The S1 nuclease protected oligomers were analyzed by 12% polyacrylamide, 8 M urea gel electrophoresis.
Subcellular Fractionation-Following transfection, the cells were lysed in 200 l of polysomal buffer (10 mM MOPS, pH 7.2, 250 mM NaCl, 2.5 mM MgOAc, 0.5% Nonidet P-40, 0.1 mM phenylmethylsulfonyl fluoride, 200 g/ml heparin, 50 g/ml cycloheximide). After removing the nuclei and cell debris by centrifugation at 12,000 ϫ g for 10 min, the polysomes were pelleted by ultracentrifugation at 100,000 ϫ g for 1 h in a 75 Ti rotor (Beckman) at 4°C and resuspended in 200 l of polysomal buffer (34). The supernatant was used as the source for post-polysomal material. The RNA from both fractions was isolated in the presence of 4 M guanidine thiocyanate (31). Samples corresponding to equal numbers of cells were used for analysis. In some experiments, the 12,000 ϫ g supernatant of the cytoplasmic faction was centrifuged in a 10-ml 10 -50% sucrose gradient, containing 25 mM HEPES, pH 7.0, 50 mM KCl, 2 mM MgOAc, 50 g/ml cycloheximide, 15 mM 2-mercaptoethanol at 40,000 rpm in a Beckman SW 41Ti rotor for 2.5 h (35). Gradient fractions were collected using a Buchler Auto Densi-Flow IIC apparatus (Buchler Instruments). Total RNA from each fraction was isolated as described before and precipitated with ethanol using 5 g of yeast tRNA as the carrier.

Effect of PABP 5Ј-UTR on ␤-Gal
Expression-Earlier studies (28, 29) using in vitro transcribed mRNAs have shown that PABP could bind to the 5Ј-UTR of its mRNA and inhibits its translation in vitro. However, the mRNA transcribed in vitro differed from its in vivo counterpart in terms of capping and methylation of the 5Ј-end and polyadenylation of the 3Ј-end, which may be important for RNA-protein interactions. Presence of additional intracellular RNA-binding proteins may also influence the binding of PABP to the 5Ј-UTR of its mRNA. Therefore, it was necessary to test the autoregulation of PABP mRNA translation in a cellular environment. This was performed by expressing chimeric ␤-Gal mRNAs consisting of different lengths of the PABP mRNA 5Ј-UTR upstream of the ␤-Gal open reading frame in HeLa cells.
The constructs used in these studies are shown in Fig. 1, panel B. Results of transient expression of ␤-Gal polypeptide by these various plasmids are shown in Fig. 2. Western blot analyses of the transfected cell extracts show that among the different chimeric constructs used for transfection, maximum level of ␤-Gal polypeptide was present in the p264-Gal-transfected cells (lane b). This construct lacked the putative PABPbinding site at the 5Ј-UTR. In contrast, presence of most of either the 5Ј-UTR (p487-Gal, lane a), or oligo(A)-rich half of the 5Ј-UTR (p223-Gal, lane c), or as little as the predominantly A-rich region between nucleotides 58 and 146 (p89-Gal, lanes d and e) caused significant reduction of the ␤-Gal level. Maximum (80%) reduction in ␤-Gal level was observed with the shortest PABP 5Ј-UTR construct p89-Gal (lanes d and e). The p223-Gal construct also showed a similar 70 -80% reduction of the ␤-Gal level (lane c). These results suggest that the distance between the A-rich region and the initiation codon per se did not have a significant influence on the reduction in the ␤-Gal polypeptide level. But when the A-rich region was moved further away from the initiation codon in the P487-Gal construct, the effect of this region was somewhat muted. The reason for this is unclear but suggests a less than all and none influence of the relative proximity between initiation codon and the regulatory sequence.
It should be noted that the level of ␤-Gal polypeptide was significantly lower in cells transfected with the chimeric plasmids than that with the parent plasmid (pCMV-SPORT-␤-Gal) (Fig. 2, compare lanes f and b). Measurement of ␤-Gal mRNA levels in these cells showed that this was primarily due to a higher ectopic mRNA level in cells transfected with pCMV-SPORT-␤-Gal (Fig. 3). The reason for these differences is not clear but may be due to the presence of additional transcriptional modulating sequences in the parent plasmid which was removed during the subcloning procedures. In all experiments presented here the difference in transfection efficiencies between experiments was monitored by co-transfecting cells with a second plasmid containing a truncated ␤-Gal gene. The results show that there were no significant differences in the levels of the truncated ␤-Gal polypeptide between different transfection experiments (Fig. 2, ⌬␤-Gal). Similar results were observed in three independent experiments.
In order to determine whether translational control of the chimeric ␤-Gal mRNA was responsible for the differential expression of the ␤-Gal polypeptide in transfected cells, we measured ␤-Gal mRNA steady-state levels ( Fig. 3) by RNase protection analyses. The mRNAs produced by all constructs used in these studies had common 3Ј-ends; therefore, we used antisense RNA transcribed from Bpu11021;digested pCMV-SPORT ␤-Gal plasmid to measure mRNA levels. As internal controls we also measured the level of GAPDH mRNA using the same RNA samples. The antisense GAPDH RNA was transcribed in vitro from commercially available clone from Ambion and produced a slightly shorter RNase-protected fragment due to engineered mismatches in the RNA sequence. The chimeric ␤-Gal mRNA levels were normalized, using the GAPDH mRNA levels, and the results (Fig. 3) show that there was little difference in the ␤-Gal mRNA levels in cells transfected with p89-Gal p223-Gal, p264-Gal, and p487-Gal. However, as discussed earlier compared with these constructs there were approximately 3 times more ␤-Gal mRNA present in cells transfected with the parent CMV-SPORT-␤-Gal plasmid. Comparison of the ␤-Gal polypeptide (Fig. 2) and mRNA levels (Fig. 3) therefore suggests that the p89-Gal, p223-Gal, and p487-Gal mRNAs were less efficiently translated than the p264-Gal and pCMV-SPORT-␤-Gal-derived mRNAs. These results suggest that the presence of the short, A-rich regions of the PABP mRNA 5Ј-UTR inhibited mRNA translation. Furthermore, the polypeptide levels in p487-Gal and PCMV-SPORT-␤-Gal-transfected cells were proportional to their mRNA levels. These results suggest that apart from the A-rich segment between nucleotides 58 and 146 of the PABP mRNA 5Ј-UTR other regions of this UTR had a minimal effect on mRNA translation. Influence of the 5Ј-UTR on mRNA Stability-Complexes involving PABP and 3Ј poly(A) of mRNAs are known to play a role in regulating mRNA stability. Therefore, the presence of a potential PABP-binding site at the 5Ј-UTR of the chimeric ␤-Gal mRNA could influence the interaction of PABP with the 3Ј poly(A) region of these mRNAs and affect mRNA stability. We therefore investigated the half-life of different chimeric ␤-Gal mRNAs in transfected cells. This was carried out by treating the cells 24 h after transfection with actinomycin D to block RNA synthesis. The steady-state levels of the mRNA was measured by RT-PCR at different times following inhibition of RNA synthesis. For each set of experiments, we also measured the level of GAPDH mRNA as internal controls. The results indicate that the half-lives of chimeric ␤-Gal mRNAs containing different regions of PABP mRNA 5Ј-UTR were similar (Fig.  4). The p89-Gal mRNA containing the shortest 5Ј-UTR decayed at approximately the same rate as the p264-Gal mRNA containing the 5Ј-UTR region proximal to the initiation site (Fig. 4,  panel B). In all transfections, no detectable difference was observed in the stability of GAPDH mRNA (Fig. 4, panel A). These results suggest that presence of a potential PABP-binding site at the 5Ј-UTR had no detectable effect on mRNA stability.
Presence of mRNA in the Translationally Repressed Fraction-Previous studies from several laboratories showed the efficiency of translation of a specific mRNA can be monitored by examining the distribution of the mRNA between the polysomal and post-polysomal fractions (36). A simple way to test this is to fractionate the cytoplasmic extract by centrifugation at 100,000 ϫ g for 1 h to remove the polysomes from the extract. The pellet recovered after centrifugation consists of materials with sedimentation coefficients Ն120 S. This represents translationally active mRNAs that are bound to more than one 80 S ribosome. The supernatant fraction therefore contains most of the non-ribosome and monosome-bound mRNAs (36). To examine whether fusion of different regions of the PABP mRNA 5Ј-UTR affects the subcellular distribution of the chimeric ␤-Gal mRNA, we measured the steady-state levels of these mRNAs in subcellular fractions. As internal controls we also analyzed the distribution of GAPDH mRNA between polysomal and post-polysomal fractions. For these analyses we have used RT-PCR reactions which showed values comparable to that of nuclease protection assays (see Fig. 6). This offered a quick comparison of the translational status of mRNAs from all constructs. The results (Fig. 5) clearly showed a preferential distribution of the adenine-rich 5Ј-UTR containing ␤-Gal mRNA (p89-Gal, compares lanes d and i) in the post-polysomal fraction. Nearly 90% of cytoplasmic p89-Gal mRNA was present in the non-translated population. This distribution pattern was not markedly different when the adenine-rich sequence was located further downstream from the initiation codon of ␤-Gal mRNA as was the case with the p223-Gal and p487-Gal mRNAs (lanes f and h). In these chimeric transcripts, the putative PABP-binding site of the 5Ј-UTR of PABP mRNA was located at different distances from the initiation codon ranging from approximately 100 in p89-Gal to approximately 500 in p487-Gal. The results of mRNA distribution studies therefore confirm that the distance between AUG and the translational regulatory region had no significant effect on the ability of this adenine-rich sequence to inhibit mRNA translation. In contrast to those mRNAs containing the A-rich domain, p264-␤-Gal was found predominantly in the translated polysomal fraction (Fig.  5, lane b). Simultaneous analyses of the GAPDH mRNA distribution between translationally active and inactive populations showed no variation in cells transfected with different constructs. In all cases more than 90% of GAPDH mRNA was present in the translationally active polysomal fraction (Fig. 5,  lanes a-e). The cytoplasmic distribution of GAPDH mRNA also suggested that the observed translational repression was not due to unexpected loss of mRNA from the polysomal population during lysis and subcellular fractionation of the cells. The distribution of chimeric mRNAs between the polysome and post-polysome fractions was in general agreement with the levels of ␤-Gal polypeptide (Figs. 2 and 5). However some discrepancy was noticed because despite any detectable difference in the distribution profile between p487 and p89-Gal mRNAs there was almost 2-fold difference in ␤-Gal polypeptide levels in cells transfected with these constructs. The reason for the discrepancy is not clear but is most likely due to the difference in the amount of monosome-bound slowly translated mRNAs in the post-polysomal fractions. This could also explain FIG. 4. Determination of mRNA stability. HeLa cells were transfected with different constructs to express the chimeric PABP-␤-Gal mRNA as described previously. Following transfection, the cells were allowed to recover in the normal growth medium for 24 h, then 5 g/ml actinomycin D was added to the culture to block RNA synthesis. At different times the cellular RNA was extracted and used for RT-PCR to measure the steady-state levels of chimeric ␤-Gal mRNAs and the internal control GAPDH mRNA. Reverse transcriptions (RT) of all RNA samples were performed with a primer (5Ј-CCA GAC CAA TGC CTC CCA GAC CGG C-3Ј) complementary to the ␤-Gal mRNA sequence adjacent to the fused PABP mRNA 5Ј-UTR. The second primer (5Ј-CAT GCC ATG GGG CCC CGC AGC AGC T-3Ј) used for PCR amplification of the p264-Gal cDNA was different from that used for p89-Gal, p223-Gal, and p487-Gal chimeric mRNAs (5Ј-CAT GCC ATG GCA CAT TTA TTA TTA AA-3Ј). The RT-PCR products were approximately 170, 250, 350, and 520 bp for p89-Gal, p223-Gal, 264-Gal, and p487-Gal mRNAs, respectively. The level of GAPDH mRNA was also measured as internal control using the following primers (sense 222-236, 5Ј-CCA CCC ATG GCA AAT TCC ATG GCA-3Ј; antisense 787-811, 5Ј-TCT AGA CGG CAG GTC AGG TCC ACC-3Ј). The numbers indicate the nucleotide number of the mRNA. A 589-bp PCR amplification product represents the GAPDH mRNA control. Amplifications without primers and the RT step were performed to check the absence of PCR products as negative controls but are not shown here. Panel A, levels of different RT-PCR products; panel B, the half-lives of p89-Gal, p264-Gal, and p487-Gal mRNAs were calculated by scanning the intensities of bands corresponding to different mRNAs as described previously (18). The average of the experiments are shown in the bar graph. why some ␤-Gal polypeptide (Fig. 2) was observed in p89-Galtransfected cells.
Further studies were performed to more accurately examine the cytoplasmic distribution of chimeric mRNAs. This was done by fractionating the cytoplasmic extract in a sucrose gradient. Previous experiments showed that p89-Gal mRNA was poorly translated, whereas the p264-Gal mRNA was efficiently translated (Fig. 5). Therefore, the distribution of these two mRNAs in the sucrose gradient fractions was monitored by RNase protection analysis. The results show that the p264-Gal mRNA was mostly present in the polysomal fractions of the gradient (fractions 6 -11). Only a small amount of this mRNA was found in the monosome-disome region (fraction 5). In contrast the majority of p89-Gal mRNA was detected in the slower sedimenting region (fractions 1-5) of the gradient. This region consisted of free mRNPs, 40 S ribosomal subunit, and 80 S monosome-bound mRNAs (Fig. 6D). In addition disomes may also be present in fraction 5. To determine that absence of p89-Gal mRNA in the polysome fractions was not due to ad-ventitious dissociation of polyribosomes or simply the consequence of false negatives in RNase protection reactions, the same samples were analyzed for the distribution of GAPDH mRNA. The results (Fig. 6, panel C) show that the polysomal distribution profile of the GAPDH mRNA was similar to that of the p264-Gal mRNA. Thus our results suggest that formation of pre-initiation complex between mRNA and 40 S ribosomal subunits was not completely inhibited by the presence of PABP at the 5Ј-UTR. It is however likely that scanning of the mRNA by the ribosomal subunit was hindered by 5Ј-UTR-bound PABP resulting in much slower joining of the 60 S ribosomal subunit to form the initiation complex.
Although the A-rich 5Ј-UTR region of PABP mRNA could confer repression of translation of the reporter ␤-Gal mRNA, it was not evident whether overexpression of PABP in HeLa cells is prevented through this sequence. To investigate this we reasoned that if overexpression of PABP beyond a threshold level is detrimental to cells, attempts to overexpress it in cells by transfection with PABP cDNA driven by a strong promoter would result in silencing of the endogenous PABP expression. This will be particularly evident if the construct for ectopic PABP expression lacks the presumptive autoregulatory A-rich 5Ј-UTR. We therefore developed two constructs. The first one lacks the 1-223-nucleotide-long region containing the A-rich sequence of the 5Ј-UTR. The second plasmid was created by joining a short fragment containing the A-rich segment of the 5Ј-UTR with this construct (Fig. 1, panel B). Transfections of cells with these constructs were carried out, and then cytoplasmic distribution of endogenous and ectopic PABP mRNAs was measured. The mRNA levels were measured by S1 nuclease protection of oligodeoxynucleotides specific for either the ectopic PABP or the endogenous PABP mRNA. We took advantage of the 5Ј-UTR deletion and the presence of unique 3Ј-UTR in the ectopic PABP mRNA in designing specific oligomers. Results (Fig. 7) show that there was no detectable nucleaseprotected oligomer in non-transfected cells when the primers specific for ectopic PABP mRNA were used (Fig. 7, panel A,  lanes c and a). Analysis of the subcellular distribution of these mRNAs shows that the ectopic PCMV⌬PABP mRNA was efficiently translated, and nearly 90% of this mRNA was present in the polysomal fraction (Fig. 7, panel A, lanes e and f). In contrast 90% of the PCMV89⌬PABP mRNA was found in the repressed post-polysomal fraction (lanes g and h). Most interesting, however, was the effect on the translation of the endogenous PABP mRNA. In untreated control cells, 70% of the endogenous mRNA was found in the translated polysomal fraction (Fig. 7, panel B, lanes c and d). A dramatic change in its translation was observed in PCMV⌬PABP-transfected cells (lanes e and f). The majority of the endogenous PABP mRNA was found in the translationally repressed state following expression of the ⌬PABP mRNA. These results show that efficient translation of the truncated unregulated PABP mRNA could force the PABP mRNA with the autoregulatory sequence to the non-translated state, thus preventing overexpression of the polypeptide. This shift to the non-translated state was unique to PABP mRNA translation and was not observed for the GAPDH mRNA translation. In mock-treated control, PCMV⌬PABP-, and PCMV89⌬PABP-transfected cells a similar distribution of GAPDH mRNA between polysomal and nonpolysomal fractions was observed. In all circumstances more than 80 -90% of GAPDH mRNA was present in the polysomal fraction (Fig. 7, panel C, lanes c, e, and f). It should also be noted that more endogenous PABP mRNA was present in polysomes than the chimeric ␤-Gal mRNAs and the ectopic PABP mRNA with the regulatory sequence at their 5Ј-UTR (Figs. 5,  panel B, and 7, panel B). The reason for this is not known but FIG. 6. Analysis of the cytoplasmic distribution of mRNA by sucrose gradients. HeLa cells were transfected with either p89-Gal or p264-Gal plasmids, and cytoplasmic extracts were prepared as described previously. The cytoplasmic fractions were centrifuged in a 10-ml gradient of 10 -50% sucrose containing 25 mM HEPES, pH 7.0, 50 mM KCl, 2 mM MgOAc, 50 g/ml cycloheximide, and 15 mM 2-mercaptoethanol at 40,000 rpm in a Beckman SW 41Ti rotor for 2.5 h. Gradient fractions were collected using a Buchler Auto Densi-Flow IIC apparatus. Total RNA from each fraction was isolated using phenol/chloroform extraction and precipitated with ethanol using 5 g of yeast tRNA as carrier. Samples were used to measure p89-Gal (panel A), p264-Gal (panel B), and GAPDH mRNA (panel C) levels by RNase protection assay as described in Fig. 3. Complete digestion of probe in absence of RNA (ϪRNA) and the undigested probe (ϪRNase) are shown as controls. The absorption profile of a representative gradient is shown in panel D. The arrows indicate the direction of migration and the positions of migrations of ribosomal subunits and monosomes. could be due to the ectopic expression of these mRNAs in the background of cellular PABP mRNA. Presence of additional sequences near the AUG codon in the endogenous mRNA could also influence the extent of its translation.
When ectopic PABP mRNA containing the putative autoregulatory region was expressed in PCMV89⌬PABP-transfected cells, only a small decrease in the translation of endogenous PABPmRNA was noticed (panel C, lanes g and h). Under these conditions approximately 55% of the endogenous PABP mRNA was present in polysomes. To test further whether the level of PABP was altered under these conditions, its cellular level was measured by Western blotting. The results (panel D) show that although in the transfected cells the combined cellular level of endogenous and ectopic PABP mRNA was almost twice that in non-transfected control cells (panels A and B), it did not produce a corresponding increase in the cellular polypeptide level. Compared with the control cells in the PCMV⌬PABP-transfected cells, there was only approximately 25% increase in the PABP level. In contrast the PABP level was reduced by approximately 50% in the PCMV89⌬PABP-transfected cells. This reduction, although slightly more than what was expected from the mRNA distribution profile, was reproducible and within the range of experimental variation and sensitivity of the detection techniques. The distribution profile of mRNAs and the level of PABP were measured several times with good reproducibility.

DISCUSSION
The presence of approximately 60% adenine base between nucleotides 71 and 131 of the PABP mRNA 5Ј-UTR offers potential PABP-binding site(s) for autoregulation of its translation. In support of this autoregulation model it was shown that removal of the first 223 nucleotides of its 5Ј-UTR improves PABP synthesis in vitro (28,29). In the studies reported here further examination of the influence of this oligo(A)-rich 5Ј-UTR sequence in cultured HeLa cells also demonstrated the inhibitory effect of PABP on mRNA translation. The presence of the 58 -146 nucleotides from PABP mRNA 5Ј-UTR resulted in inhibition of translation of the reporter ␤-Gal mRNA in HeLa cells. Although the adenine-rich region was located between nucleotides 71 and 131, we used a slightly larger region (nucleotides 58 -146) of the 5Ј-UTR in our analyses. Due to the predominance of a single base the A-rich region was not clonable without some flanking sequences. The results of our studies did not rule out the possibility of other regulatory sequences existing within the 5Ј-UTR of PABP mRNA. PABP mRNA translation appeared to be differently regulated in different cells (5, 16 -18). It is therefore possible that its translation can be regulated by sequence(s) other than the putative PABPbinding sequence in other cell lines or in a tissue-specific manner. It is conceivable that the presence of cell-specific RNAbinding proteins may be involved in differential regulation of translation of PABP mRNA. However, our major concern was to test the possibility of autoregulation and the potential role of the oligo(A)-rich sequence in mRNA translation in vivo. To test further this model, we examined the effect of ectopic expression of this mRNA on the translation of endogenous mRNA. Our results showed that translation of the endogenous PABP mRNA was repressed when attempts were made to overexpress PABP. These results suggest that the level of PABP and its mRNA may determine how much of the cytoplasmic pool of the PABP mRNA is translated. This checkpoint in PABP expression depends on the presence of the A-rich region of its 5Ј-UTR.
In cells transfected with the PABP expression vectors, the mRNA levels were increased by more than 2-fold. But only a modest increase (25%) in the PABP level was found (Fig. 7). In the cells expressing the unregulated ectopic mRNA, this was achieved by almost completely blocking the translation of endogenous mRNA and translating the ectopic mRNA. On the other hand when the ectopic mRNA had the auto-regulatory sequence (ARS) both endogenous and ectopic mRNAs translation were controlled. However, the translation of the ectopic PABP mRNA with a shorter 5Ј-UTR and the ARS positioned closer to the initiation codon than in the endogenous mRNA was repressed more than that of the endogenous mRNA. The reason for this difference is not clear but may reflect the influence of other regulatory sequences. As a result of nearly complete inhibition of the ectopic mRNA translation and a small shift of the endogenous PABP mRNA to the non-translated state, the levels of PABP in these cells were reduced by 50%. Thus, there was over-compensation of PABP expression.
The precise mechanism of PABP-mediated feedback control of translation is not known. However, the results suggest that 48 S preinitiation complexes were formed but were not efficiently converted to the initiation complex. It is therefore possible that binding of PABP at the ARS may decrease the rate of the 40 S ribosomal subunit movement along the mRNA. According to this model the distance between the initiation codon and the ARS is unlikely to have a major influence on mRNA translation. The distance of the ARS from the 5Ј-cap region, however, might have stronger influence than that of the ARS and the AUG codon. This is conceivable since the 3Ј poly(A)-PABP complex may be involved in the translation initiation through its interaction with the eIF4G. Therefore, formation of a similar complex close to the 5Ј-capped end may mute their functional interaction. Further studies are required to understand how ARS-PABP complex works.
Feedback control of PABP mRNA translation may serve an important purpose in regulating the cellular level of free PABP. Under normal circumstances, cytoplasmic PABP is present in excess over the availability of the poly(A)-binding sites (37). The reason for the presence of free PABP is not clear. However, some of this may result from degradation or shortening of the poly(A) tail. Presence of multiple poly(A)-binding domains in PABP (38) potentially contributes to the dynamic nature of its interaction with the 3Ј poly(A) tail (39). This could allow movement of PABP between the 3Ј poly(A) tail and other lower affinity PABP binding sequences (37) including the ARS at the 5Ј UTR of PABP mRNA. How efficiently PABP would interact with the lower affinity binding sequences may depend on its cellular level.
Presumably an optimal level of PABP is required for maintaining an equilibrium between its various functions. Our studies have shown that the presence of the putative PABP-binding site at the 5Ј-UTR of PABP mRNA reduces by approximately 3-fold the translation of this mRNA both in vitro (28) and in vivo (Fig. 3). Although the presence of a certain excess amount of PABP is tolerated and may even be required for normal cellular function, its expression beyond a threshold could be controlled at the mRNA translation level. Excess PABP beyond the threshold level may detrimentally alter the pattern of protein synthesis in the cell, for example by increasing translation of inherently inefficient mRNAs. In addition stability of some unstable mRNAs may also increase if excess PABP is made. Together, these events could have a dramatic effect on the growth regulation and cellular differentiation. This possibility is supported by the previous observation that overexpression of eIF4E results in cellular transformation (41). Since the role of PABP in translation is believed to be mediated through its interaction with the cap structure and eIF4G (of which eIF4E is a component), it will be important to control PABP expression through feedback mechanisms. This is further supported by the observation that overexpression of PABP in Xenopus oocytes influenced the deadenylation and translation of maternal mRNAs (42).