Ethanol Destabilizes Liver Galβl, 4GlcNAc α2,6-Sialyltransferase, mRNA by Depleting a 3′-Untranslated Region-Specific Binding Protein

Among the biomarkers that are specific for heavy alcohol consumption, plasma carbohydrate-deficient transferrin and sialic acid-deficient apolipoprotein J have been established as viable markers (Stibler and Borg, 1981; Tsutsumi et al., 1994; Lakshman et al., 1999). We have shown previously that hepatic sialylations of transferrin and apolipoprotein J, the two N-glycosylated proteins, and apolipoprotein E, an O-glycosylated protein, are markedly impaired in chronic ethanol-fed rats compared with the pair-fed controls (Lakshman et al., 1993, 2001; Ghosh et al., 2001). Furthermore, we showed a parallel retention of newly synthesized glycoproteins such as apolipoprotein E and transferrin within the liver cell in chronic ethanol-fed rats (Marmillot et al., 2001).

To define the significance and molecular mechanisms of aberrant sialylation in alcoholics, we focused attention on sialyltransferases, the key enzymes involved in the metabolism of glycoproteins and glycolipids. Sialyltransferases are a family of glycosyltransferases that catalyze the transfer of sialic acid to terminal positions on the carbohydrate groups of glycoproteins and glycolipids in the Golgi compartment. Our previous studies (Ghosh and Lakshman, 1997; Rao and Lakshman, 1997, 1999) have shown that chronic ethanol feeding in rats caused a marked decrease (59% reduction compared with controls) of the Galβl, 4GlcNAc α2,6-sialyltransferase (ST6Gal l) activity and its synthetic rate, as well as its mRNA level in liver. We also showed that the decrease ST6Gal l mRNA level was due to its decreased stability. Because of this observation, it was necessary to investigate whether the regulation was at the posttranscriptional level.

The complete cDNA for rat ST6Gal l is 4.2 kb, with a 2.7-kb 3′-untranslated region (UTR) region. Sequence analysis by computer software revealed several conserved sequences among different species within ST6Gal l 3′-UTR. Therefore, it is reasonable that this extremely long 3′-UTR may play important roles in the posttranscriptional regulation of rat ST6Gal l. The 3′-UTR of many eukaryotic mRNAs has been implicated in a variety of cellular processes, such as mRNA stability, processing, polyadenylation, localization, and translational regulation. In each case, functional activity appears to be mediated, in part, by a specific interaction of RNA-binding proteins, which target cellular RNAs to form RNA-protein complexes (Jain et al., 1995; Zaidi and Malter, 1995; Wang et al., 1996; Joseph et al., 1998; Ostareck-Lederer et al., 1998; Sakai et al., 1999; Gilmore et al., 2001; Kufel et al., 2003). More and more reports have suggested that mRNA stability is frequently determined by RNA-protein interactions, and those interactions frequently occur within the 3′-UTR. Thus, RNA-protein interactions are crucial for maintaining proper RNA metabolism. We have undertaken to analyze the RNA-protein interactions within the 3′-UTR of ST6Gal l to characterize the binding protein with which this RNA sequence interacts and to investigate whether chronic ethanol feeding affects the status of this binding protein and consequently that of the ST6Gal l mRNA transcripts. It will be unequivocally demonstrated in the present investigation that a 41-kDa binding protein specifically binds to the 3′-UTR of ST6Gal l mRNA and stabilizes the mRNA. Significantly, chronic alcohol feeding decreases this liver cytosol binding protein, leading to impaired binding and destabilization of ST6Gal l mRNA.

Materials and Methods

Animal Feeding. Male Wistar rats (Charles River, Wilmington, MA) were maintained on a standard laboratory chow until they reached a body weight of around 150 g. They were then divided into four groups and pair-fed the nutritionally adequate control or indicated alcohol diets for 8 weeks. The diets were isoenergetic and were formulated according to the modified method reported by Lieber and DeCarli (1982). The ethanol calories in the experimental groups were 10.8, 21.6, and 36% of the total dietary calories, respectively. The control diet had an equicaloric amount of dextrin-maltose in place of ethanol. All of the animals in various experimental groups gained their normal weights, and there were no statistically significant differences in body weight gains in various groups (data not shown).

Sialic Acid Index of Plasma Apolipoprotein J. At the end of the experimental period, all of the animals from various groups were euthanatized by exsanguination and pentobarbital anesthesia (50 mg/kg i.p.), and the blood plasma was collected from control and 36% ethanol groups and stored at –80°C. One milliliter of plasma from each animal was subjected to affinity column purification of apolipoprotein J and subsequent determination of sialic acid index of plasma apolipoprotein J (SIJ) as described by us previously (Ghosh et al., 2001).

Preparation of Liver Cytosol Extracts. At the end of the experimental period, all of the animals from various groups were euthanatized by exsanguination and pentobarbital anesthesia (50 mg/kg i.p.), and the livers were harvested and washed three times with ice-cold phosphate-buffered saline (PBS). The liver cytosol fraction was prepared as described previously (Marmillot et al., 2001) and was either used directly or stored at –80°C. The total protein concentration of the cytosol fraction was determined by the Bradford method (Bio-Rad, Hercules, CA) and generally diluted to be 5 μg/μl.

Preparation of the Riboprobe Templates. DNA templates for use in the synthesis of mRNA probe for in vitro transcription were prepared by amplifying a 2029-bp fragment from 3′-UTR immediately downstream of the stop codon of rat ST6Gal l by RT-PCR and from gel-purified templates sequentially truncated from its 3′ end (Fig. 4). Total RNA (5 μg) extracted from rat liver were used in 20 μl of reverse transcription reaction containing 10 pM oligo(dT), 10 μM dNTPs, and 1 unit of reverse transcriptase. Typical PCR reaction mixture included 2 μl of cDNA templates from RT, 10 pM each primer, 10 μM dNTPs, 3 mM MgCl₂, 10× buffer, and 2 units of high-fidelity TaqDNA polymerase in a reaction volume of 50 μl. The PCR conditions were 2 min at 94°C followed by 35 cycles at 94°C for 30 s, 54°C for 30 s, and 72°C for 3 min. The primer pairs used were: forward primer, 5′-TAATACGACTCACTATAGACCTAGCCAGGCACCCTTA-3′, and reverse primer, 5′-TGCCTTATAATGAGCGTGTGAC-3′. The forward primer included a T7 RNA polymerase promoter sequence, which is underlined.

Preparation of ³²P-Labeled RNA Transcripts.³²P-Labeled and unlabeled mRNA fragments were transcribed in vitro with T7 RNA polymerase from PCR-amplified 3′-UTR of ST6Gal l and truncated templates according to the method described by Milligan et al. (1987). In brief, the reactions were performed in the presence of 1 μg of template in a buffer containing 2 nM [α-³²P]UTP (800 Ci/mmol), 2.5 mM of each unlabeled rGTP, rCTP, and rATP, 40 units of RNasin, and 2 U of T7 polymerase. Following incubation for 1 h at 37°C, template DNA was digested with 1 unit of RNase-free DNase, and unincorporated nucleotides were removed by Centricon-30 membrane filtration (Millipore, Billerica, MA). ³²P-Labeled RNA transcripts were quantified by liquid scintillation counting.

RNA-Protein Electrophoretic Gel Mobility Shift Assays.³²P-Labeled RNA at a concentration of 2 nM was incubated with different amounts of cytosol proteins (1-200 μg) in a 10-μl solution containing 10 mM HEPES, pH 7.5, 25 mM KCl, 10% glycerol, and 1 mM dithiothreitol at 30°C for 30 min. For competition experiments, the cytosol fraction was incubated for 10 min with 100-fold molar excess of unlabeled RNA transcripts before incubation with the labeled RNA. In some experiments, the cytosol fraction was treated with proteinase K at a final concentration of 2 mg/ml for 30 min at 37°C, or the cytosol fraction was heated at 56°C for 10 min before conducting the standard binding assay as above. Then, 2 μl of 6× native gel loading buffer (30% glycerol, 0.025% bromphenol blue, and xylene cyanol) was added, and the RNA-protein complexes were resolved on an 8% native polyacrylamide gel in 0.5× Tris borate/EDTA buffer (45 mM Tris-HCl, pH 8.3, 45 mM borate, and 2.5 mM EDTA). Gels were pre-electrophoresed for 30 min at 20 mA followed by electrophoresis at 30 mA for 2 to 4 h at 4°C. Gels were dried and exposed to XAR film (PerkinElmer, Wellesley, MA) with an intensifying screen at –70°C overnight.

Ribonuclease Digestion of RNA-Protein Complex. Equal amounts of labeled RNA transcripts were used in the standard binding reaction as described above either before or after treatment with 1 ng of RNase A (Ambion, Austin, TX), 0.1 U of RNase V1 (Ambion), 20 U of RNase T1 (Ambion), and 50 U of S1 Ribonuclease (Promega, Madison, WI) for 30 min at 30°C. The reaction mixtures were then resolved by 8% native PAGE gel and exposed to XAR film at –80°C for an appropriate time.

UV Cross-Linking of Protein and RNA. The RNA-protein binding reaction was set up with ³²P-labeled RNA and 50 μg of cytosol protein as described above. After incubation for 30 min at 30°C, 20 U of RNase T1 was added, and the reaction mixture was incubated for 30 min at 30°C to digest unprotected RNA. The binding reaction mixture was then exposed to UV light in Stratalinker UV light box, model 1800 (Stratagene, La Jolla, CA) at automatic setting. After irradiation, the resulting ³²P-labeled proteins were resolved by 10% SDS-PAGE in the presence of high-resolution molecular mass markers. The gel was dried and exposed to XAR film with an intensifying screen at –70°C for 3 days.

Mapping the Specific Protein Binding Sequence on ST6Gal l mRNA 3′-UTR. RT-PCR-amplified 2029-bp ST6Gal l 3′-UTR template was sequentially digested with AvrII, BglII, SacI, and KpnI. The resulting fragments of 304, 831, 1425, and 1742 bp were gel-purified and were used as templates to generate RNA probes designated as probe 5, probe 4, probe 3, and probe 2, respectively. All of these RNA probes were used in binding assay with cytosol proteins as described above. Four 80-bp RNA probes that cover probe 5 (304 bp), were synthesized by Operon Technologies, Inc. (Alameda, CA). Later, RNA probes corresponding to a 13-bp consensus sequence, and the rest of the 67-bp nonconsensus sequences on the 80-bp region were also synthesized. These probes were 5′-end labeled with biotin so that they can bind tightly with streptavidin-coated paramagnetic particles (SA-PMPs) (Promega). Binding was performed by adding 1 μg of biotin-labeled RNA probe to 50 μl of liver cytosol extract for 30 min at room temperature in binding buffer (10 mM HEPES, pH 7.5, 25 mM KCl, 10% glycerol, and 1 mM dithiothreitol). Streptavidin magnetic beads were washed in binding buffer (lacking glycerol), and resuspended beads were added to the binding reaction for 25 min at room temperature. The magnetic beads with immobilized RNA and its bound proteins were separated from crude cell extract on a magnetic stand. The RNA-protein complex bound to the SA-PMP was washed twice with the binding buffer, and proteins were eluted by boiling for 5 min in sample buffer (62.5 mM Tris-HCl, pH 6.8, 25% glycerol, 2% SDS, 0.01% bromphenol blue, and 1% 2-mercaptoethanol). SDS-PAGE gel was run, and the gel was stained with Coomassie Brilliant Blue stain.

Mutagenesis Analysis of the 13-bp Consensus Sequence. Four RNA probes with mutations corresponding to the following sequences were synthesized with 5′-end biotin label: consensus, 5′-CAGCCTCCTCCCT-3′; Mprobe 1, 5′-CCTCCTCCTCCCT-3′; Mprobe 2, 5′-CAGCCTCCCACCT-3′; Mprobe 3, 5′-CAG TTTCCTCCCT-3′; and Mprobe 4, 5′-CAGCCTCCGCCCT-3′. Each of these probes was tested for its efficiency in binding with liver cytosol fraction under standard conditions as described above.

Specificity Analysis of Protein-ST6Gal I mRNA Interaction. The ST8Sia 1 gene was selected as a control to investigate the specificity of this protein-ST6Gal I mRNA interaction. RNA probes for both ST8Sia 1 and ST6Gal I were synthesized according to their respective 3′-UTRs, and the probes were 5′-end-labeled with biotin. These probes were then used in the binding assay as described above.

Results

Partial Identification of the Liver Cytosol Protein That Forms Complex with the 3′-UTR of Rat ST6Gal l mRNA. The interaction of the cytosol fraction from control rat liver with ³²P-labeled ST6Gal l mRNA probe of 2029-bp length covering nucleotides immediately downstream from the stop codon was analyzed by electrophoretic gel mobility shift assays (EMSA) on 8% native polyacrylamide gels. As can be seen in Fig. 1 A, incubation of the ³²P-labeled ST6Gal l mRNA probe with as little as 1 μg of cytosol protein retarded the migration of the probe, leading to the appearance of a RNA-protein complex radioactive band (Fig. 1A, lane 1). There was an increase in the intensity of this band with increasing amount of the cytosol protein up to 50 μg, beyond which no further increase occurred (Fig. 1A, lanes 1–7).

Prior incubation of the cytosol fraction with 100-fold molar excess of unlabeled cold RNA probe completely quenched the binding of ³²P-labeled RNA probe as evidenced by the loss of the intensity of the RNA-protein complex band (Fig. 1B, lane 4). Likewise, RNA-protein complex formation was abolished by treating the cytosol fraction with proteinase K or preheating at 56°C for 10 min (Fig. 1B, lanes 5 and 6).

Importance of Secondary Structure of RNA in RNA-Protein Complex Formation.Figure 1B (lanes 1–3) shows the interaction of the cytosol protein with the RNA probe that was first heat-denatured at 90°C for 10 min and either cooled down rapidly on ice to maintain the denatured structure (Fig. 1B, lane 2) or cooled down gradually to room temperature to allow the stable RNA secondary structure to form (Fig. 1B, lane 3), compared with normal binding reaction with unheated RNA probe (Fig. 1B, lane 1). It is obvious from the figure that the same strong signal for the RNA-protein complex band could be observed with both denatured and renatured RNA probes.

The RNA-Protein Interaction Protects 3′-UTR from RNase Digestion. To investigate whether the formation of protein-ST6Gal l complex plays any functional role in ST6Gal l mRNA stability, RNase digestion assays were performed with the longest RNA probe (probe 1) either before or after the binding reactions. As can been seen from Fig. 2, the specific protein binding to the 3′-UTR of ST6Gal l completely protected it from digestion by RNases S, V1, and A and partially protected it from RNase T1. It is worthy to note that RNase S is nonspecific for digesting RNA; RNase V1 cleaves only base-paired nucleotide, whereas RNase A cleaves 3′-U and C residues. On the other hand, RNase T1 cleaves only at the 3′-G residue; therefore, it is possible that some RNA sequence may remain intact after RNase T1 digestion.

Partial Characterization of Protein Interacting with 3′-UTR of ST6Gal l mRNA by UV Cross-Linking. As a direct method to determine the molecular mass of this protein under denatured condition, the products of the binding reaction with the ³²P-labeled ST6Gal l RNA probe were UV cross-linked, and the noncross-linked RNA was removed by RNase T1 digestion. The resulting ³²P-labeled RNA-protein complex was separated by SDS-PAGE along with a set of molecular mass standards on a separate lane. As shown in Fig. 3, a single band corresponding to 41 kDa was seen.

Thirteen-Base Pair Motif of ST6Gal l mRNA Is Sufficient for the Specific Protein to Bind. To identify the shortest possible nucleotide sequence of 3′-UTR of ST6Gal l with which the binding protein interacts, the 2029 bp of the RT-PCR template was digested with restriction enzymes, and the sequentially truncated templates were used for RNA probe synthesis and EMSA (Fig. 4A). As shown in Fig. 4B, all of the probes listed in Fig. 4A interacted with the protein, indicating that probe 5, which is only 304 bp immediate downstream from the stop codon, was clearly sufficient to facilitate the binding interaction. However, it was difficult to visualize the bands by EMSA, even with prolonged exposure with RNA probes that were shorter than probe 5, due to the limited incorporation of [³²P]UTP. Therefore, after narrowing the binding region to 304 bp downstream from the stop codon (probe 5), 5′-biotin-labeled RNA probes were made within this region and were used to further map the binding region. Therefore, four 5′-end biotin labeled 80-bp RNA probes covering the entire length of the 304-bp RNA probe 5 were synthesized and tested in the standard binding reaction. Surprisingly, in contrast to other nonspecific faint bands that were observed when the cytosol fraction was incubated with the magnetic beads alone (Fig. 5A, lane 6), a strong band corresponding to 41 kDa was observed with equal binding intensity exclusively for each of these 80-bp probes when they were incubated with the cytosol fraction and the magnetic beads (Fig. 5A, lanes 2–5). Further analysis revealed that all of the four 80-bp probes shared the 13-bp consensus sequence as shown in Fig. 5B. Further mapping of the 80-bp probe showed that only the RNA probe corresponding to the 13-bp consensus sequence showed binding (Fig. 6A, lane 1) but not to the probe corresponding to the remaining nonconsensus sequence (Fig. 6A, lane 3). Significantly, when the nucleotides AG and TC were mutated to CT and CA, respectively, in the 13-bp consensus sequence, the corresponding RNA probes completely lost their binding ability (Fig. 6B, lanes 1 and 2). However, when only the nucleotide T was mutated to nucleotide G (Fig. 6B, lane 5), the corresponding RNA probe still showed partial binding ability. The mutation of other nucleotides, like CC to TT, did not affect the binding (Fig. 6B, lane 4).

The Specificity of the Cytosolic Binding Protein That Interacts with ST6Gal I mRNA. ST8Sia 1 is another member of sialyltransferase family, which links the sialic acid residues to another sialic acid residue of brain glycoconjugates through an α2,8-bond. In an attempt to test the specificity of this binding protein to ST6Gal I mRNA, it was reasonable to test whether this binding protein interacted with ST8Sia 1 mRNA probe. Biotin-labeled ST8Sia 1 RNA probe was used along with the labeled ST6Gal I probe to study the specificity of their binding to the liver cytosolic fraction. As shown in Fig. 7, a band corresponding to 41 kDa was found only with ST6Gal I probe (lane 1) but not with ST8Sia 1 probe (lane 2) under identical conditions.

Effects of Chronic Ethanol Exposure in Vivo on the Status of the Binding Protein and Plasma SIJ. As shown in Fig. 8A, there was a dietary ethanol concentration-dependent decrease in the binding intensity of the RNA-protein complex band that virtually disappeared when the cytosol fraction from the livers of rats fed with 36% of their dietary calories because ethanol was tested in the standard binding reaction. The intensity of the UV cross-linked ³²P-labeled RNA protein complex band also progressively decreased with increasing concentration of dietary ethanol (Fig. 8B). There was also a 45% (p < 0.05) decrease in plasma SIJ (14.2 ± 1.7 in the 36% ethanol group versus 25.8 ± 3.7 in the control group).

Discussion

ST6Gal l mediates the transfer of α2,6-linked sialic acid to glycoproteins. Down-regulation of the expression of this gene and consequent impaired activity of ST6Gal l has been associated with the increase of asialoconjugates in the blood of chronic alcoholics and retention of glycoproteins in the liver. In our previous studies (Ghosh and Lakshman, 1997; Rao and Lakshman, 1997, 1999), we showed that chronic ethanol feeding in rats caused a marked decrease (59% reduction compared with controls) of the ST6Gal l activity as well as its mRNA level in liver that was due to decreased stability of ST6Gal l mRNA. In our present work, we have further characterized the mechanism of action of ethanol in the destabilization of ST6Gal l mRNA. We report here for the first time the identification and partial characterization of a liver cytosol protein that specifically binds to the 3′-UTR of the ST6Gal l mRNA and plays a role in its stability. Significantly, chronic ethanol feeding decreases the intracellular concentration of this binding protein leading to the destabilization of ST6Gal l mRNA and its down-regulation by rapid degradation.

The observed binding intensity of RNA-protein complex as a function of protein concentration (Fig. 1A) reveals that the binding reaction seems to be saturated at 50 μg of cytosol protein when the RNA concentration is kept at 2.0 nM. Unfortunately, it is premature to carry out Scatchard analysis using the crude cytosol fraction to determine the dissociation constant for this binding protein. This has to wait until the purification of this protein to homogeneity is achieved. Nonetheless, if the stoichiometry of interaction of the RNA with the binding protein is mole per mole, our data imply that the relative abundance of this binding protein in normal rat liver cytosol fraction is approximately 0.0035% of total cytosol protein (based on our finding that the molecular mass of the binding protein is 41 kDa). The specificity of this binding is demonstrated by virtual absence of the binding in the presence of 100-fold excess of the unlabeled probe (Fig. 1B, lane 4). The polypeptide nature of this binding protein is confirmed by its sensitivity to proteinase K and heat treatment (Fig. 1B, lanes 5 and 6). Based on the fact that the binding protein interacts equally well with both native and denatured RNA probes (Fig. 1B, lanes 2 and 3), it is clear that the proper secondary structure of RNA is not an absolute prerequisite for this interaction to occur. The fact that the RNA-protein complex was resistant to digestion by various types of RNases, whereas the naked RNA was totally destroyed on treatment by the same RNases, clearly shows that the binding protein specifically protects the RNA from degradation (Fig. 2). UV cross-linking experiments revealed a molecular mass of the binding protein to be around 41 kDa (Fig. 3). The fact that the 3′-UTR of ST8Sia 1, another related sialyltransferase, failed to show any interaction with this binding protein (Fig. 7) strongly supports our finding this binding protein is highly specific for interaction with the ST6Gal l 3′-UTR.

Based on the fact that the secondary structure of ST6Gal l 3′-UTR is not critical for this specific protein-ST6Gal l interaction (Fig. 1B, lanes 2 and 3), it is logical that this protein should interact with the ST6Gal l 3′-UTR in a sequence specific manner. Our data shows that the RNA probe that covers the 304 bp immediate downstream from the stop code of ST6Gal l 3′-UTR shows tight binding to this binding protein (Fig. 4). Surprisingly, all four RNA probes (5-1, 5-2, 5-3, and 5-4) of 80-bp length spanning the entire 304-bp probe showed equal binding intensity to the binding protein (Fig. 5A). Sequence alignment analysis of these four probes revealed that the corresponding cDNA sequences for the four 80-bp RNA probes share 13-bp consensus sequences (Fig. 5B). Further analysis confirmed that this consensus sequence is not present in other areas of the ST6Gal l 3′-UTR and coding region. The fact that only the 13-bp consensus sequence but not the remaining nonconsensus sequence on the 80-bp region showed binding activity (Fig. 6A) clearly establishes that this 13-bp consensus sequence serves as the specific binding site for this protein. Mutagenesis analysis conclusively proved that the conserved nucleotides AG and TC are critical for the RNA-protein interaction (Fig. 6B). Thus, a mutation of even a single nucleotide seems to affect the binding.

Although the identity of the 41-kDa protein is yet to be determined, its specificity and high-affinity interaction with the narrow UTR region of ST6Gal l mRNA clearly indicates its critical role in regulating the ST6Gal l mRNA metabolism as evidenced by the influence of chronic ethanol exposure that leads to its destabilization (Ghosh and Lakshman, 1997; Rao and Lakshman, 1997, 1999). The formation of the RNA-protein complex progressively decreased with increasing dietary ethanol concentration leading to its virtual disappearance in the livers of rats fed with 36% of the total calories as ethanol (Fig. 8, A and B). A parallel 45% (p < 0.05) decrease in plasma SIJ in the 36% ethanol group compared with the control group fully confirms our concept that chronic ethanol feeding destabilizes ST6Gal l mRNA level by decreasing the amount of this specific binding protein that interacts with the 3′-UTR of ST6Gal l mRNA and leads to the generation of asialoconjugates in the blood of alcoholics.

Our present study demonstrates for the first time another unique interaction of a liver cytosolic-specific binding protein with the 3′-UTR region of ST6Gal l mRNA that protects it from degradation in normal rat liver. We further show that chronic ethanol exposure down-regulates this mRNA by depleting this specific binding protein leading to ethanol-mediated destabilization of ST6Gal l mRNA. It is significant to point out that the amount of ethanol consumption, when it is fed at 36% of the total dietary calories, amounts to 12 to 14 g/kg body weight per day, a value that is comparable with ethanol drinking by human heavy alcoholics (>100 g/day that is equivalent to ≥6 drinks/day). Thus, this specific defect caused by chronic ethanol exposure in alcoholics is the most likely cause for the blood appearance of asialoconjugates such as carbohydrate-deficient transferrin (Stibler and Borg, 1987), sialic acid-deficient apolipoprotein J (Ghosh et al., 2001; present study), and asialo-α1-acid glycoprotein (Tsutsumi, 1994) that serve as excellent biomarkers for chronic alcohol consumption.