Poly(A)-binding Protein Positively Affects YB-1 mRNA Translation through Specific Interaction with YB-1 mRNA*

The major protein of cytoplasmic mRNPs from rabbit reticulocytes, YB-1, is a member of an ancient family of proteins containing a common structural feature, cold-shock domain. In eukaryotes, this family is represented by multifunctional mRNA/Y-box DNA-binding proteins that control gene expression at different stages. To address possible post-transcriptional regulation of YB-1 gene expression, we examined effects of exogenous 5′- and 3′-untranslatable region-containing fragments of YB-1 mRNA on its translation and stability in a cell-free system. The addition of the 3′ mRNA fragment as well as its subfragment I shut off protein synthesis at the initiation stage without affecting mRNA stability. UV cross-linking revealed four proteins (69, 50, 46, and 44 kDa) that specifically interacted with the 3′ mRNA fragment; the inhibitory subfragment I bound two of them, 69- and 50-kDa proteins. We have identified these proteins as PABP (poly(A)-binding protein) (69 kDa) and YB-1 (50 kDa) and demonstrated that titrating out of PABP by poly(A) strongly and specifically inhibits YB-1 mRNA cap+poly(A)− translation in a cell-free system. Thus, PABP is capable of positively affecting YB-1 mRNA translation in a poly(A) tail-independent manner.

In bacteria, accumulation of major cold-shock proteins at low temperatures was shown to result mainly from strong and selective stabilization of their mRNAs (30,31). The crucial role of eukaryotic CSD proteins in major cellular events suggests that there is precise regulation of their expression. At present the post-transcriptional control of eukaryotic gene expression is to a large extent attributed to the presence of specific sequences within mRNA 5Ј-and 3Ј-untranslatable regions (UTRs), which serve as targets for binding of certain proteins and complementary RNAs (32,33). Here, we studied this type of regulation during in vitro synthesis of rabbit p50, a member of the eukaryotic CSD protein family, which is the major protein of reticulocyte mRNPs and is virtually identical to the human Y-box binding protein YB-1. To determine a possible role of YB-1 mRNA UTRs in YB-1 post-transcriptional regulation, we examined effects of exogenous mRNA fragments containing different UTR parts on YB-1 mRNA translation and stability in a cell-free system from rabbit reticulocytes.
We found that addition of the 3Ј mRNA subfragment I, including the last 42 nucleotides of YB-1 mRNA coding region and the first 107 nucleotides of its 3Ј-UTR, shut off YB-1 synthesis mainly at the initiation stage without affecting mRNA stability. The inhibition by this subfragment was also observed in cell-free systems translating luciferase or endogenous globin mRNAs. In contrast, subfragment II containing the rest of YB-1 mRNA 3Ј-UTR or the 5Ј fragment did not affect translation. Two proteins were found to specifically bind the inhibitory subfragment I, leading us to believe that they play an important role in mRNA translation. We have identified these proteins as poly(A)-binding protein (PABP) (69 kDa) and YB-1 (50 kDa) and demonstrated that titrating out of PABP by poly(A) strongly and specifically inhibited YB-1 both cap ϩ poly(A) Ϫ and cap ϩ poly(A) ϩ mRNA translation in a cellfree system. Thus, PABP is capable of positively affecting YB-1 mRNA translation in a poly(A) tail-independent manner.

EXPERIMENTAL PROCEDURES
Plasmid Construction-pBluescript II SK YB-1 construct containing YB-1 cDNA, pSP36T-LucA50 construct containing 5Ј-UTR of ␤-globin mRNA fused with the luciferase coding sequence, and pSP64/CAT were described previously (34 -36) pBluescript II SK YB-1 A50 was made as follows. The fragment containing 50-nucleotide-long poly(A) was cut from pSP36T-LucA50 construct by digesting with HindIII and PstI and ligated between the same sites of pBluescript II SK (Stratagene). The obtained construction was named pBluescript II SK A50. Then YB-1 gene was cut from pBluescript II SK YB-1 by digesting with HindIII and SmaI and ligated between HindIII and BglII of pBluescript II SK A50 (blunt ends after BglII restriction were made using the Klenow fragment). pBluescript II SK 3Ј-UTR YB-1 construct was made as follows. The 3Ј-UTR fragment of YB-1 cDNA was digested with PvuII and BamHI from pBluescript II SK YB-1 and ligated between the EcoRV and BamHI sites of pBluescript II SK (Fig. 1).
pBluescript II SK subfragment II 3Ј-UTR YB-1 construct was made as follows. 5Ј-UTR, the coding sequence, and 3Ј-UTR subfragment I of YB-1 cDNA were removed from pBluescript II SK YB-1 construct by digesting with EcoRV and BspTI, followed by fill-in reaction with Klenow fragment and blunt end religation (Fig. 1).
The transcription was performed as described previously (37). Capped transcripts were obtained by a reaction where instead of 5 mM GTP a mixture of 0.2 mM GTP and 0.8 mM m7GpppG (Amersham Biosciences) was used. To generate 32 P-labeled fragments of YB-1 mRNA, [ 32 P]UTP (2000 Ci/mM; Radioisotop) was added to the reaction, and the concentration of unlabeled UTP was reduced to 0.05 mM.
In Vitro Translation Assays-Translation of endogenous or exogenous mRNA in a rabbit reticulocyte cell-free system was performed as described previously (38). The incubation mixture (15 l) contained reticulocyte lysate (7.5 l), 10 mM Hepes-KOH, pH 7.6, 100 mM potassium acetate, 1 mM magnesium acetate, 8 mM creatine phosphate, 0.5 mM spermidine, 0.2 mM GTP, 0.8 mM ATP, 1 mM dithiothreitol, 25 M of each of 20 amino acids except for the labeled ones (Leu or Met), and 6 Ci of [ 14 C]Leu (Ͼ300 mCi/mmol; Radioisotop) or 1 Ci of [ 35 S]Met (Ͼ700 Ci/mmol; Radioisotop). In the case of micrococcal nuclease (MN)treated lysate, the mixture was supplemented with appropriate mRNA as indicated in the figure legends. Translation was carried out at 30°C for 2 h. The incorporation of [ 14 C]Leu into the protein was measured in 5-l aliquots as described previously (34). [ 35 S]Met-incorporated proteins were analyzed by 10 -22% SDS-PAGE followed by autoradiography. The relative amount of radioactivity in the bands was determined using a Packard Cyclone Storage Phosphor System (Packard Instrument Co., Inc.).
Sucrose Gradient Analysis-Cell-free translation mixtures (30 l) with endogenous mRNA were incubated at 30°C for 5 min, cooled, and immediately layered onto 15-40% (w/v) linear sucrose gradients in buffer containing 10 mM Tris-HCl, pH 7.6, 100 mM NaCl, 5 mM MgCl 2 . Centrifugation was carried out at 45,000 rpm in an SW-60 rotor (Beckman Instruments) for 50 min. All gradients were monitored for absorbance at 254 nm during their collection from the bottom.
Northern Blot Analysis-Total RNA was extracted from translation mixtures by the phenol/chloroform method. 7 g of total RNA from each reaction (measured by A 260 ) was separated by electrophoresis on a denaturing 1.5% agarose gel containing 2.2 M formaldehyde and 1ϫ MOPS buffer (40 mM MOPS, 10 mM sodium acetate, 1 mM EDTA, pH 7.0). RNA was transferred onto a nylon membrane (Hybond-N, Amersham Biosciences) and cross-linked using transilluminator cross-linker (Vilber-Lourmat) at 0.15 J/cm 2 . Membrane-bound RNA was hybridized to a 1054-nt fragment of YB-1 cDNA probe or a 600-nt fragment of ␣-globin cDNA probe labeled with [ 32 P]dATP (Multiprime DNA labeling system, Amersham Biosciences) in 50% formamide, 5ϫ SSC, 5x Denhard's solution, 1% SDS, 100 g/ml denatured salmon sperm DNA at 65°C for 12-16 h. The membrane was washed twice with 2ϫ SSC, 0.1% SDS for 5 min at room temperature, twice with 0.2ϫ SSC, 0.1% SDS for 5 min at room temperature, twice with 0.2ϫ SSC, 0.1% SDS for 15 min at 42°C, and twice with 0.1ϫ SSC, 0.1% SDS for 15 min at 68°C and analyzed by autoradiography.
Electrophoretic Mobility Shift Assay-0.5 l of MN-treated rabbit reticulocyte lysate was incubated with radiolabeled RNA (10,000 cpm) in a final volume of 10 l (10 mM Hepes-KOH, pH 7.6, 100 mM KCl, 0.5 mM dithiothreitol, 1 mg/ml bovine serum albumin) at 30°C for 15 min. In the case of proteinase K treatment, 10 g of proteinase K and SDS up to 0.1% were added and followed by incubation for 5 min. After incu-bation, the protein-RNA complexes were analyzed by 6% PAGE in 0.5ϫ TBE buffer (1ϫ: 89 mM Tris, 89 mM boric acid, 2 mM EDTA), followed by autoradiography.
UV Cross-linking Assays-5 l MN-treated rabbit reticulocyte lysate was incubated with radiolabeled RNA (100,000 cpm) in a final volume of 18 l (10 mM Hepes-KOH, pH 7.6, 100 mM KCl) at 30°C for 15 min. The reactions were irradiated at 1.5 J/cm 2 in a transilluminator crosslinker (Vilber-Lourmat), incubated for 45 min at 37°C with 0.05 unit/l MN and 0.5 g/l RNase A, and analyzed by 10 -22% SDS-PAGE followed by autoradiography.
Depletion of YB-1 from Reticulocyte Lysate-For immunodepletion experiments, nuclease-treated reticulocyte lysate (30 l) was diluted 3-fold with buffer (10 mM Hepes-KOH, pH 7.6, 100 mM KCl) and incubated with 20 l of anti-YB-1 or preimmune antibodies fixed on protein A-Sepharose beads for 2 h at room temperature with gentle agitation. Then the protein A-Sepharose beads were removed by centrifugation. The efficiency of depletion of YB-1 was controlled by immunoblotting.
Depletion of PABP from Reticulocyte Lysate-Nuclease-treated reticulocyte lysate (300 l) was diluted 3-fold with buffer (10 mM Hepes-KOH, pH 7.6, 200 mM KCl) and incubated with 200 l of poly(A)agarose beads (Amersham Biosciences) for 2 h at 0°C with gentle agitation. Then the poly(A)-agarose beads were removed by centrifugation. The efficiency of depletion of PABP was controlled by immunoblotting.
Immunoblotting-For Western blot analysis, proteins were separated by SDS-PAGE and transferred to a nitrocellulose membrane. The membrane was blocked overnight at 20°C with 1% bovine serum albumin, 1% polyvinylpyrrolidone, and 0.05% Tween 20 in buffer (10 mM Tris-HCl, pH 7.6, and 150 mM NaCl) and probed with appropriate antibodies. Immunocomplexes were detected with alkaline phosphatase-coupled secondary antibodies (Promega) according to the manufacturer's recommendation.
YB-1 and PABP Purification-YB-1 and PABP from mRNPs of rabbit reticulocytes were obtained as described earlier (39). After dissociation of ribosomes using 33 mM EDTA polyribosomal mRNPs were adsorbed to the oligo(dT)-cellulose (Amersham Biosciences) in buffer (10 mM Tris-HCl, pH 7.6, 250 mM NaCl, 2 mM EDTA, and 0.5 mg/ml heparin) at 4°C and then the column was washed with the same buffer solution. mRNPs were eluted with buffer (10 mM Tris-HCl, pH 7.6, 2 mM EDTA) at 37°C, pelleted by centrifugation at 100,000 rpm for 2 h in the type TLA-100.3 rotor (Beckman) at 4°C and dissolved in buffer (10 mM Tris-HCl, pH 7.6, 100 mM NaCl). mRNP proteins dissociated from mRNA using 3 M LiCl, and the mRNA was precipitated overnight at Ϫ20°C. The RNA precipitate was removed by centrifugation at 12,000 rpm for 15 min at 4°C. Proteins from the supernatant fraction were dialyzed against buffer (10 mM Hepes-KOH, pH 7.6, 200 mM KCl) and separated from aggregates by centrifugation at 12,000 rpm for 15 min at 4°C. About 500 g of mRNP proteins were applied to a Superose 6 HR/10/30 column (Amersham Biosciences) equilibrated with buffer (10 mM Hepes-KOH, pH 7.6, 200 mM KCl). Chromatography was done at a flow rate of 0.4 ml/min using a fast protein liquid chromatography system. The fraction volume was 1.5 ml.

RESULTS
3Ј Fragment of YB-1 mRNA Inhibits YB-1 in Vitro Synthesis-As we mentioned above, in eukaryotes, 5Ј-and 3Ј-UTRs of mRNA are involved in regulation of activity, stability, and intracellular distribution of mRNA through interaction with specific regulatory proteins or RNAs (32, 33, 40 -42). Here we studied the effect of exogenous 5Ј and 3Ј fragments of rabbit YB-1 mRNA (Fig. 1) on YB-1 synthesis in a rabbit reticulocyte cell-free translation system. If these mRNA fragments compete with YB-1 mRNA for binding to specific regulatory proteins or RNAs, we would expect to detect a change in YB-1 mRNA activity and/or stability.
The addition of the 3Ј fragment containing the 231-nt UTR and the last 42 nt of YB-1 coding sequence to the cell-free system translating YB-1 cap ϩ poly(A) Ϫ mRNA strongly inhibited YB-1 synthesis ( Fig. 2A, lanes 3 and 4). In contrast, the 132-nt 5Ј fragment containing 5Ј-UTR did not affect YB-1 mRNA translation ( Fig. 2A, lanes 5 and 6). To better localize the inhibitory signal within the 3Ј fragment, two its subfragments were tested in a cell-free translation system. The upstream 149-nt-long subfragment I caused strong inhibition of YB-1 synthesis (Fig. 2B, lanes 6 and 8), while the downstream 124-nt-long subfragment II did not (lanes 9 -11).
The addition of subfragment I also inhibited translation of luciferase cap ϩ poly(A) Ϫ mRNA (Fig. 3A) and natural globin mRNA (Fig. 3B), demonstrating that this inhibition is not specific for YB-1 mRNA translation.
Subfragment I of YB-1 mRNA Inhibits Translation Initiation without Affecting mRNA Stability-The stage of protein synthesis affected by subfragment I may be identified by studying its effect on the polysomal profile in a rabbit reticulocyte cellfree system. Inhibition of translation at the stage of initiation is known to be accompanied by a decay of polysomes, whereas inhibition of elongation/termination results in larger polysomes. Subfragment I or II was added to the cell-free translation system with endogenous mRNA, and the mixture was incubated for 5 min and analyzed by centrifugation in sucrose gradient. The polysomal profiles are shown in Fig. 4. The addition of subfragment I resulted in a complete decay of polysomes (cf. Fig. 4, B with A, presenting control with no subfragment), while with subfragment II added, the profile was similar to that of the control (cf. Fig. 4, C with A). The subfragment I-induced decay of polysomes may result from inhibition of translation at the initiation stage. However, it cannot be ruled out that subfragment I may stimulate mRNA degradation, which would also result in a decay of polysomes.
To verify the latter suggestion, we studied the effect of subfragments I and II on YB-1 and ␣-globin mRNA stability in the translation reaction. The total RNA was isolated from cell-free translation systems described above after 5-min incubation, subjected to gel electrophoresis (Fig. 4F), and analyzed (along with control YB-1 or ␣-globin mRNA) by Northern blot hybridization to 32 P-labeled YB-1 (Fig. 4D) or ␣-globin cDNA (Fig.  4E). No notable difference in mRNA content and size was observed between experiment and control. This allowed concluding that polysome decay caused by subfragment I did not result from mRNA degradation but was caused by inhibition of translation itself mostly or exclusively at the stage of initiation.
Four Proteins Specifically Interact with the 3Ј Fragment of YB-1 mRNA-To ascertain the possible proteins interacting with the 3Ј fragment and affecting mRNA activity, we performed an electrophoretic mobility shift experiment using reticulocyte lysate and 32 P-labeled 3Ј fragment. Addition of the lysate to this labeled fragment resulted in retardation of the labeled band in the gel (Fig. 5A, compare lanes 1 and 2). This retardation effect was avoided by proteinase K treatment (Fig.  5A, lane 3) and was sensitive to the amount of non-labeled 3Ј fragment added (lanes 4 -6), but not to addition of other RNAs (lanes 7-12), revealing specific binding of some lysate protein(s) to the labeled fragment. To characterize these proteins, we UV-irradiated the mixture containing reticulocyte lysate and the labeled 3Ј fragment. Four proteins with molecular weights of about 69, 50, 46, and 44 kDa were found to cross-link to the labeled fragment (Fig. 5B, lane 2). The same proteins were cross-linked to the labeled 3Ј fragment under conditions of an active translation system with exogenous or endogenous  mRNA (2-and 15-min incubation). These proteins, with the exception for 50-kDa protein, are efficiently competed out by the non-labeled 3Ј fragment (Fig. 5B, lanes 3-5). Other RNAs used in this experiment demonstrated less competition (Fig.  5B, lanes 6 -14) revealing that these three proteins, 69, 46, and 44 kDa, bound the 3Ј fragment specifically. The inability of the non-labeled 3Ј fragment to efficiently substitute for the labeled fragment in retarded complexes (Fig. 5A, lanes 4 -6) and in UV cross-linking to 50-kDa protein (Fig. 5B, lanes 3-5) indicated that there was an excess of 50 kDa protein in the lysate. For this reason we could not yet conclude that the interaction of the 3Ј fragment of YB-1 mRNA with 50-kDa protein was specific.
Next, we studied the binding specificity of 69-, 50-, 46-, and 44-kDa proteins to subfragments I and II of the 3Ј fragment to determine which of these proteins may affect mRNA translational activity. Using UV cross-linking, we analyzed the binding of lysate proteins to the labeled 3Ј fragment in the presence of increasing amounts of non-labeled competitors: 3Ј fragment, subfragment I, and subfragment II. Similarly to what was shown above, the non-labeled 3Ј fragment efficiently competed out 69-, 46-, and 44-kDa proteins from the complex containing the labeled fragment (Fig. 6A, lanes 3-5). Moreover, we found that subfragment I competed with the labeled fragment only for 69-kDa protein (Fig. 6A, lanes 6 -8), while subfragment II competed only for 46-and 44-kDa proteins (lanes 9 -11).
Then, the 3Ј fragment and subfragments I and II were radioactively labeled and incubated with the lysate separately. Analysis of proteins UV cross-linked to each of these RNAs revealed that 69-and 50-kDa proteins specifically bound subfragment I (Fig. 6B, lane 5), while 46-and 44-kDa proteins bound subfragment II (lane 6). This result suggests that 50and 69-kDa proteins may be involved in protein synthesis and promote translation, and their specific binding by exogenous subfragment I may cause inhibition of translation. However, we detected much higher amounts of 50-kDa protein compared with 69 kDa in reticulocyte lysate, which makes the latter a more likely candidate.
50-kDa Protein Specifically Interacting with Subfragment I Is YB-1-Subfragment I of YB-1 mRNA was found to have the motif AACAUC in its UTR portion (Fig. 1). According to the literature, this motif specifically interacts with Xenopus FRGY2 that displays 64% homology to YB-1 (43). This led us to believe that 50-kDa protein specifically cross-linked to subfragment I in rabbit reticulocyte lysate is YB-1. To verify this suggestion, YB-1 isolated from mRNPs of rabbit reticulocytes was cross-linked to 32 P-labeled subfragment I in the presence of increasing amounts of unlabeled subfragment I or II as a competitor. Subfragment I substituted for its labeled form in the complex with YB-1 (Fig. 7A, lanes 4 -6) much more efficiently than subfragment II (lanes 7-9), thereby showing YB-1 specificity for subfragment I. To demonstrate that 50-kDa protein specifically cross-linked to subfragment I in rabbit reticulocyte lysate is YB-1 as well, 32 P-labeled subfragment I was added to the lysate, UV crosslinked, treated with a nuclease mixture, incubated with anti-YB-1 antibodies, and adsorbed onto protein A-Sepharose. The adsorbed 32 P-labeled proteins were analyzed by SDS-PAGE followed by autoradiography. As seen from Fig. 7B, lane 2, only the 50-kDa protein cross-linked to subfragment I precipitated with anti-YB-1 antibodies, whereas 69-kDa protein did not adsorb onto protein A-Sepharose beads (Fig. 7B, lane 3). None of them interacted with preimmune antibodies and adsorbed onto protein A-Sepharose (Fig. 7B, lanes 4 and 5). In another experiment, MN-treated rabbit reticulocyte lysate was incubated with anti-YB-1 antibodies, and YB-1 was precipitated with protein A-Sepharose beads. Preimmune antibodies were used as a control. After the precipitation, equal aliquots of all lysates were subjected to immunoblotting and UV cross-linking to 32 P-labeled subfragment I (Fig. 7C). The 32 P-band corresponding to 50-kDa protein was absent only from the YB-1depleted lysate (Fig. 7C, lane 2). Hence, the 50-kDa protein specifically interacting with subfragment I in rabbit reticulocyte lysate is YB-1.
69-kDa Protein Specifically Interacting with Subfragment I Is PABP-The analysis of the subfragment I primary structure showed that it is A-rich (34%). Some A residues are assembled in 3-5-nt clusters (Fig. 1). Besides, the molecular mass of 69-kDa protein is close to that of PABP. To check whether 69-kDa protein specifically interacting with subfragment I has a preferential affinity for poly(A), we studied by UV crosslinking the effect of homoribopolynucleotides on 69-kDa protein binding to subfragment I (Fig. 8A). As seen, poly(A), unlike poly(U), poly(C), and poly(G), strongly suppressed binding of 69 kDa protein to subfragment I, thereby demonstrating a preferential affinity of this protein for the poly(A) sequence. Therefore, it was suggested that 69-kDa protein is PABP. To verify this suggestion, MN-treated rabbit reticulocyte lysate was incubated with poly(A)-agarose beads with subsequent precipitation of PABP. The PABP depletion was controlled by immunoblotting. As seen from Fig. 8B, lane 2, the incubation resulted in complete depletion of PABP. Then the PABP-depleted lysate was subjected to UV-crosslinking to 32 P-labeled subfragment I (Fig. 8C, lane 3). The absence of the 32 P-band indicated that the 69-kDa protein was PABP indeed. Addition of purified PABP to the PABP-depleted lysate resulted in re-appearance of the 32 Pband (Fig. 8C, lane 4) that disappeared again upon addition of increasing amounts of unlabeled subfragment I (Fig. 8C, lanes  5-7) but not II (Fig. 8C, lanes 8 -10). This confirms PABP  6. Interaction of subfragments I and II of YB-1 mRNA with proteins in rabbit reticulocyte lysate. A, aliquots of rabbit reticulocyte lysate (5 l) were incubated for 15 min at 30°C with 32 P-labeled 3Ј fragment (0.33 pmol) in the presence of increasing amounts of unlabeled 3Ј fragment or each of its two subfragments (0.33, 0.65, and 1.3 pmol), UV cross-linked, treated with RNase A and MN, analyzed by SDS-PAGE, and visualized by autoradiography. B, aliquots of rabbit reticulocyte lysate (5 l) were incubated for 15 min at 30°C with 32 P-labeled 3Ј fragment or each of its two subfragments (0.2 pmol), UV cross-linked, treated with RNase A and MN, analyzed by SDS-PAGE, and visualized by autoradiography. specificity to subfragment I of YB-1 mRNA.
Exogenous Poly(A) Specifically Inhibits Translation of Nonpolyadenilated YB-1 mRNA-Specific binding of YB-1 and PABP to YB-1 mRNA suggested that these proteins could selectively affect translation of YB-1 mRNA. Here we studied the effect of PABP titrating out on YB-1 cap ϩ poly(A) Ϫ and cap ϩ poly(A) ϩ mRNA translation. As controls, we used capped luciferase, CAT, and globin poly(A) ϩ or poly(A) Ϫ mRNAs. mRNAs were translated in a cell-free system with [ 35 S]Met (Fig. 9, A-E) or [ 14 C]Leu (Fig. 9F) in the presence of increasing amounts of poly(A). The proteins were separated by SDS-PAGE, and [ 35 S]Met translation products were detected by autoradiography (Fig. 9, A-E). The incorporation of [ 14 C]Leu into the protein was measured in 5-l aliquots as described under "Experimental Procedures" (Fig. 9F). Addition of poly(A) was shown to inhibit YB-1 mRNA translation much stronger (Fig. 9, A and D) than translation of any other mRNA (Fig. 9, B-C and E-G). This effect did not depend on whether or not the YB-1 mRNA had a poly(A) tail. Hence, the PABP titrating out by its binding to poly(A) inhibited predominantly YB-1 mRNA translation. This allowed us to conclude that PABP positively affects YB-1 mRNA translation in a poly(A) tail-independent manner. DISCUSSION Our experiments demonstrated that synthesis of the major mRNP protein YB-1 in the rabbit reticulocyte cell-free translation system may be strongly suppressed by exogenous 3Ј fragment of YB-1 mRNA. Moreover, this RNA fragment inhibited synthesis of other proteins tested. Northern blot analysis of mRNA in a translation reaction inhibited by addition of the 3Ј fragment showed mRNA intactness. The protein synthesis inhibition was accompanied by a rapid decay of polysomes, which shows that inhibition occurs mostly or exclusively at the initiation stage of translation. Our idea is that 3Ј-UTR of YB-1 mRNA specifically interacts with protein(s) necessary to initiate translation of all cellular mRNAs. Through this interaction, these proteins produce an additional positive effect on YB-1 translation, then in trans titrating out of these proteins by 3Ј fragments would be expected to cause inhibition of synthesis of YB-1, as well as other cellular proteins.
According to the literature, the majority of proteins specifically associated with 3Ј-UTRs of various mRNAs cause inhibition of mRNA translation (44,45). However, there are recent data demonstrating the presence of enhancing sequences within 3Ј-UTRs, as well as a possibility of mRNA translation activation through interaction of specific proteins with these sequences. These observations were made for mouse protamine 1 mRNA (46,47), fibronectin mRNA (48), manganese superoxide dismutase mRNA (49 -51), mRNA for the ␤ subunit of mitochondrial H ϩ -ATP synthase (52), human fibroblast growth factor 2 mRNA (53), and ornithine decarboxylase mRNA (54).
UV cross-linking revealed four proteins that specifically interacted with the 3Ј fragment of YB-1 mRNA in rabbit reticulocyte lysate. Their molecular masses are ϳ69, 50, 46, and 44 kDa. Two of them, 69-and 50-kDa proteins, specifically bind to the inhibitory subfragment I, while 46-and 44-kDa proteins specifically interact with the rest of YB-1 mRNA 3Ј-UTR. As suggested above, the association with 69-and/or 50-kDa protein is required for mRNA translation. Quantitative analysis of subfragment I required to bind all 50-and 69-kDa proteins in the lysate revealed higher amounts of 50-kDa as compared with 69-kDa protein.
Subfragment I was shown to contain a hexamer (AACAUC) that, according to the literature, specifically binds FRGY2, a YB-1 homolog from Xenopus laevis (43). A reporter mRNA with this motif is more sensitive to the inhibitory effect of FRGY2 (24). In immunoprecipitation experiments, we have identified 50-kDa protein as YB-1 and showed that isolated YB-1 specifically binds to subfragment I of YB-1 mRNA. Evidently, a specific heteropolymeric sequence is implicated in this binding, since none of the homoribopolynucleotides (including poly(G) displaying high affinity for YB-1 (55)) can compete with subfragment I for binding to YB-1 (Fig. 8A). The nucleotide sequence of subfragment I specifically binding YB-1 is currently under study, and soon it will be clear whether the predicted hexameric motif AACAUC belongs to the binding site.
The analysis of nucleotide sequence of the 3Ј fragment also showed that subfragment I is A-rich (34%), with As arranged in 3-5-nt clusters. This allowed us to suggest that 69-kDa protein specifically binding to this fragment is PABP. Experiments on competition between subfragment I and homoribopolynucleotides for 69-kDa protein proved that this protein really had a higher affinity for poly(A). Selective removal of PABP from rabbit reticulocyte lysates by its adsorption onto poly(A)-agarose beads led to a complete disappearance of the protein specifically cross-linked to subfragment I of YB-1 mRNA, which proved that 69-kDa protein was PABP indeed. Addition of purified PABP to the 69-kDa protein-depleted lysate led to re-appearance of the 69 kDa band. Experiments with competitive unlabeled subfragments I and II confirmed a higher affinity of the added PABP for subfragment I.
Since both identified proteins, YB-1 and PABP, are universal major mRNP proteins and both are required for mRNA translation initiation (PABP stimulates initiation of cap ϩ poly(A) ϩ mRNA (56), while YB-1 stimulates both cap ϩ poly(A) ϩ and cap Ϫ poly(A) Ϫ mRNAs (39, 57)), it becomes clear why subfrag- ment I inhibited translation of all mRNAs tested. Most likely, subfragment I has a dual effect on protein synthesis: it inhibits translation initiation due to PABP titrating out on the one hand and due to YB-1 titrating out on the other.
Finally, we demonstrated that PABP titrating out by poly(A) strongly and specifically inhibited translation in a cell-free system of both poly(A) ϩ and poly(A) Ϫ YB-1 mRNA. Hence, PABP specifically and positively affects YB-1 mRNA translation in a poly(A) tail-independent manner. Inhibition of translation of other tested mRNAs by exogenous poly(A) was far less strong and with only a minor difference between poly(A) ϩ and poly(A) Ϫ , the latter being in a good agreement with the literature data (35). A stronger poly(A) inhibitory effect on translation of YB-1 mRNA probably resulted from a lower PABP affinity for a specific PABP-binding sequence of YB-1 mRNA as compared with its affinity for the poly(A) tail.
The translational status of mRNA in the cell is determined by the ratio between two major mRNA-associated proteins: YB-1 and PABP (58). PABP is known to be capable of inhibiting its own synthesis through interaction with a 61-nt A-rich element of 5Ј-UTR of its mRNA (59,60). Also, as reported here, PABP can be involved in positive regulation of YB-1 synthesis, thereby maintaining the PABP to YB-1 ratio optimal for translation. PABP and YB-1 specifically interact with the same fragment of YB-1 mRNA, and therefore, overlapping of their binding sites cannot be ruled out; then excess YB-1 might displace PABP from its complex with YB-1 mRNA, thus negatively regulating its own synthesis. This suggestion is currently under verification.