Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Alu transcripts: cytoplasmic localisation and regulation by DNA methylation

Alu transcripts: cytoplasmic localisation and regulation by DNA methylation Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 Nucleic Acids Research, 1994, Vol. 22, No. 6 1087-1095 © 1994 Oxford University Press Alu transcripts: cytoplasmic localisation and regulation by DNA methylation Wen-Man Liu1, Richard J.Maraia3, Carol M.Rubin1 and Carl W.Schmid1'2* 1Section of Molecular and Cellular Biology and department of Chemistry, University of California, Davis, CA 95616 and 3National Institute of Child Health and Human Development, NIH, Bethesda, MD 20892, USA Received September 30, 1993; Revised and Accepted February 7, 1994 EMBL accession nos U02044 - U02068 (incl.) ABSTRACT Full length Alu transcripts in HeLa cells are detected sequences and some Alu sequences persist in mature mRNA (8, by primer extension using reverse transcriptase and are 9). We refer to Pol IE-directed Alu transcripts as being authentic transcripts, as distinct from Alus that are transcribed by virtue also analyzed as cloned cDNA sequences. The 5' end of these transcripts corresponds to the transcriptional of their interspersion within other transcription units. The start site for RNA polymerase III indicating that these interspersion of Alu sequences in Pol U-directed transcripts RNAs are transcribed from their internal polymerase III precludes the use of S1 mapping and nuclear run on assays to promoters. The Alu transcripts found in cytoplasmic detect Pol HI Alu transcripts confidently. Many Alu hybridization probes cross hybridize with abundant 7SL RNA complicating poly A + RNAs appear to be organized into RNPs as assayed by sucrose gradient sedimentation. Present at the use of Northern blot hybridization analysis (5, 6, 7, 10). about one hundred to one thousand copies per cell, the Primer extension by reverse transcriptase tests whether the 5' Alu transcripts are rare as compared to 7SL RNA. In end of an Alu transcript corresponds to the initiation site expected agreement with previous reports that methylation for Pol Hi-directed transcription. However many Alu primer extension products are prematurely truncated, presumably due inhibits Pol Ill-directed transcription of Alu in vitro, treatment of HeLa cells with 5-azacytldine results in Alu to secondary structures in the GC-rich Alu sequence. This DNA hypomethylation and an increase in the problem is further compounded by the abundance of Alu abundance of the Alu transcript. Sequence analysis homologs in both 7SL RNA and Pol H-directed transcripts, which shows that many different Alu repeats including can introduce a high background of prematurely truncated primer extension products as compared to primer extension products members of all subfamilies are transcribed by Pol III In vivo. cDNA sequences of the Pol Ill-directed transcripts resulting from relatively rare Alu transcripts (5, 6, 7). exactly match the A box of the Pol III promoter element Given these technical challenges, little is known about the whereas in other Alu transcripts this element is not regulation and fate of Alu transcripts. Full-length Alu transcripts, faithfully conserved. which can vary in size from 300 to 500 nt, have been detected in total RNA and in the polyadenylated (poly A+ ) cytoplasmic fractions (5, 6, 7). This size heterogeneity is presumably due INTRODUCTION to variable 3' ends as the Alu Pol III promoter initiates transcription at the first base of the Alu repeat (11, 12). Alu Human Alu repeats are thought to retrotranspose through an RNA repeats belong to subfamilies of different evolutionary ages and intermediate (1, 2). Also, Alu repeats contain an internal RNA while observed Alu transcripts represent all subfamilies, they are polymerase III (Pol III) promoter and most are transcribed in enriched in young subfamily members (6, 7). Results from in vitro by Pol III (1, 2, 3). Despite the abundance of Alus and vitro transcription assays indicate that the A box element of the their transcriptional potential, corresponding Pol HI transcripts internal Pol HI promoter sequence might be a critical determinant are very rare in tissues and cell lines examined to date and have of the relative transcriptional activity of Alu repeats (13). Full been extremely difficult to demonstrate convincingly (4). length Alu cDNA sequences available through one study do not However, identification of Alu subfamily consensus sequences test this prediction, as the A box sequences were replaced during (1,2) has allowed the use of specific oligonucleotide probes to the PCR cloning step which generated the cDNA clones (6). demonstrate that Alu transcripts do indeed accumulate in vivo Although the Alu consensus sequence is a dimeric structure, (5, 6, 7). shorter transcripts corresponding to Pol Ill-initiated transcripts Unusual problems that prevent or complicate the use of many of the left Alu monomer have also been observed (5, 7, 10). conventional techniques to study Alu transcription deserve These transcripts accumulate in the nonpolyadenylated (poly A—) attention. Because Alus are ubiquitously interspersed in other fraction of the cytoplasm and are designated small cytoplasmic transcription units, about 10% of pre mRNA consists of Alu *To whom correspondence should be addressed Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 1088 Nucleic Acids Research, 1994, Vol. 22, No. 6 Alu (scAlu) RNA. Some full length dimeric Alu transcripts may 81 -10 3 of the PV Alu subfamily consensus sequence (2). Oligo be precursors to monomeric scAlu RNAs (7). 783, 5'-ATGCCGAACTTAGTGCGG-3', is complementary to positions 112-129 (S-domain) of 7SL RNA. Oligo 2334. Factors provided or induced by adenovirus and herpes virus 5'-GCCGGGACAACACATT-3', is complementary to positions increase the steady state abundance of Alu containing transcripts 82-9 7 of Clone 350- 3 RNA. Adapt-TI7, 5'-GACTCGAGT- (14, 15). Those studies do not distinguish between scAlu and CGACATCGA(T)17-3\ Adapt-1, 5'-GACTCGAGTCGAC- full length Alu transcripts and the mechanism of this stimulation ATCGA-3', and 18mer, 5'-GAGAATTCTCGGCTCACTGC- is unknown. Alu repeats are heavily methylated in DNAs isolated A-3' were used for PCR amplification of Alu cDNA as shown from various somatic tissues and certain Alu subgroups are on the diagram below. The 18mer nests one nucleotide upstream undermethylated or almost completely unmethylated in sperm to the 17mer. DNA (16, 17, 18). In vitro transcription of Alu templates is inhibited by a repressor which is specific toward methylated DNA (Adapt-1) (13). <-EcoRI (18mer) > (Adapt-T17) Here we: i) examine the subcellular localization of Alu > (PV51) <— (17mer) transcripts; ii) test the effect of DNA methylation on the A L U- > A A A A A I— • expression of Alu repeats in vivo; and iii) examine the Pol III 0 240nt intragenic promoter structure of Alus that generate full length transcripts. Primer extension assays RNA in 20 /*l of annealing buffer [10 mM Tris-HCl pH 7.5, MATERIALS AND METHODS 1 mM EDTA, 0.3 M KC1, 1 unit Inhibit-ACE (5prime-3prime)] RNA preparation and 5-1 0 pmole (about 105 cpm) of a primer were incubated HeLa cells were grown to 4-6 x 105 cells per ml in spinner in boiling water for 2 minutes and then allowed to anneal at 42°C flasks at 37°C in SMEM (Minimum Essential Medium, GIBCO) for 1 hour. Primers were labeled by T4 polynucleotide kinase containing 5 % calf serum. Cells were also grown for either eight with (7-32P)-ATP and were ethanol precipitated. Following or twenty days in the presence of 8/iM 5-azacytidine (19, 20). addition of 80 /tl of reverse transcriptase solution (10 mM Cells were washed once in 10 volumes of 4°C PBS (phosphate Tris-HCl pH 8.6, 5 mM MgCl , 5 mM DTT, 1 mM dNTP, buffered saline), collected by centrifugation, resuspended in five and 1 unit Inhibit-ACE), the reaction mixture was incubated with volumes of cytoplasmic RNA extraction buffer [140 mM NaCl, 200 units of MMLV reverse transcriptase (USB) for 3 hours at 35 °C. The reaction was stopped by the addition of 100 /xl phenol/ 1.5 mM MgCl , 10 mM Tris-HCl pH 8.4, 1 mM DTT, 0.5% chloroform mixture. The extracted aqueous phase was ethanol NP-40, 0.5% Triton X-100, 20 mM VRCs (vanadyl precipitated, and the DNA pellet was resuspended in 6.5 /*l ribonucleoside complexes)], and lysed by 20 — 30 strokes of a denaturing loading dye mix, denatured at 90°C for 5 min, and Thomas tissue grinder at 4°C. Nuclei were pelleted by electrophoresed in 5 — 8% polyacrylamide-8M urea gel. centrifugation at 2,000 rpm (Sorvall SS-34 rotor) for ten minutes. M13mpl8 ssDNA sequencing provided size markers. The gel The cytoplasmic supernatant was removed, and intact nuclei were was exposed to XAR X-ray film with a Quanta III intensifying resuspended in GnSCN buffer (4M guanidinium thiocyanate, 25 screen at — 70°C. mM Na-citrate pH 7.0, 0.5% sarkosyl, 0.1 M mercaptoethanol) with a Pyrex tissue grinder. Both nuclear and cytoplasmic Northern blot results fractions were extracted several times with equal volumes of phenol/chloroform and chloroform and were ethanol precipitated. RNA blotting was done from 5.5% polyacrylamide gel (40:1 The nuclear RNA was purified by CsCl centrifugation to remove acrylamide:bis) as previously described (10, 22). Cytoplasmic genomic DNA. Poly A + and poly A - RNAs were selected as RNA (65 ng) and poly A + RNA (19 /tg) were electrophoresed with 40 ng of sonicated salmon sperm DNA per sample. After described (21). soaking out urea with transfer buffer, the gel was transferred RNP isolations and sucrose gradients to Gene Screen Plus (NEN) at 30 volts overnight using a Bio Rad transblot cell with plate electrodes. The membrane was UV- PBS-washed cell pellets were suspended in sucrose gradient buffer crosslinked and dried throughly. The membrane was incubated (100 mM KC1, 10 mM Hepes pH 7.8, 1 mM DTT, 10 /ig/ml with hybridization solution (6xSSC, 2xDenhardt's, 0.5% SDS, cycloheximide), 0.5 mM PMSF (phenylmethylsulfonyl fluoride), 100 /ig/ml yeast RNA) at the hybridization temperature (58°C) 2 /tg/ml TLCK (N-a-p-tosyl-L-lysine chloromethyl ketone), 20 before hybridization for 18 hours. The end-labeled hybridization mM VRCs, and either 3 mM EDTA (for ribosomal subunit probe designated Alu-24 has been previously described (7, 22). fractionation) or 3 mM MgCl (to maintain poly somes). The membrane was washed twice with 2xSSC at room Cytoplasmic RNAs were fractionated by 5-3 5 % (3 mM EDTA) temperature and once at the hybridization temperature. or 18—42 % (3 mM MgCl ) sucrose density gradient Quantitation of the bands was done with the ImageQuant program centrifugation in a Beckman SW 40.1 Ti rotor at 4°C. of a phosphorimager (Molecular Dynamics). Denatured 0X174 Centrifugation was at 35,000 rpm for 7 hours for 5-35 % sucrose DNA was used as the size marker in this experiment. gradients and at 30,000 rpm for 2 hours for 18-42% sucrose gradients. Twelve equal volume fractions were collected. cDNA cloning and sequence analysis Oligodeoxyribonucleotides The 240 nt and 350 nt bands from primer extension products All oligonucleotides were custom synthesized. Oligo 17mer, were excised from the gel and electroeluted. Following poly (dA) 5'-GCGATCTCGGCTCACTG-3', corresponds to positions tailing, PCR amplification and EcoRJ/SaJI directional cloning into 238—221 of the Alu consensus sequence. Oligo PV51, 5'-A- pUC19 were performed as previously described (22) except that CCATCCCGGCTAAAACGGTGA-3', corresponds to positions the primers were Adapt-T17 and Adapt-1 (5'). and 18mer (3'). Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 Nucleic Acids Research, 1994, Vol. 22, No. 6 1089 Double strand DNA cycle sequencing (BRL) was employed for Cyto. sequence determinations. I I A + A' T N Assaying DNA niethylation For blot hybridization, 50 /*g of genomic DNA (from cells treated for twenty days with 5-azacytidine and untreated control cells) was digested overnight with 50 units of Tthl 111 (ca. ten fold overdigestion), phenol extracted, and similarly redigested with Tthl 111. After phenol extraction and ethanol precipitation the sample was divided into two equal aliquots for overnight digestion with either BstUI (50 units) or TaqI (27 units). After agarose gel (1.3%) electrophoresis, the samples were blotted onto 0.2 micron nitrocellulose and hybridized with oligo PV51 (16, 18). At the final washing temperature employed in this experiment (63°C), this oligonucleotide selectively hybridizes to the PV Alu subfamily (5, 18). Alu repeats were isolated by agarose gel electrophoresis as a visible 190 nt restriction fragment in TaqI —Aspl digests of genomic DNAs. Aspl is a Tthl 1II isoschizomer and the digestion conditions were similar to those described above. Equal aliquots of these preparations were subjected to HpaTJ and Mspl digestion and were terminally labeled with y-32P ATP using T4 polynucleotide kinase. The products were compared by acrylamide gel electrophoresis (18, 23). RESULTS Primer extension of Alu transcripts Initially, we used oligonucleotide primers situated exactly on the 3' end of the Alu consensus sequence, hoping to obtain full length primer extension products. However, this region includes an almost exact match to 7SL RNA and a close match to 28S rRNA. Figure 1. Primer Extension of control and HeLa cell RNAs. 60ng and 6ng of a control Alu template (7) were used as templates for primer extension with oligo [The 3 ' Alu consensus sequence is GAGCGAGACTCCGTCT- I7mer giving as expected a 240 nt cDNA as the principal product. 3 ^g poly C which closely matches the sequence GAGCGAGACCCGT- A + and 3 fig poly A— cytoplasmic RNAs, and cellular equivalents of total CGC beginning at position 943 in human 28S rRNA (see cytoplasmic RNA (60 /ig), and nuclear RNA (6 y.%) were each used as templates Discussion)]. Truncated primer extension products resulting from for primer extension reactions with oligo 17mer Two products of 240 and 350 these very abundant RNAs interfered with the detection of the nt were observed with cytoplasmic RNAs (poly A+ and total fractions). much rarer Alu transcript (data not shown). To avoid these complications, an oligonucleotide complementing positions 221 to 238 of the Alu consensus sequence which is totally dissimilar RNA by comparing the intensities of their respective primer to 7SL and ribosomal RNA sequences is used in the primer extension products (see below). The 7SL RNA primer extension extension assays reported here. Using a control template this product is about four orders of magnitude more abundant then primer, produces the expected, 238 nt, full length extension the corresponding Alu product. Assuming that there are one product (Figure 1, Lanes 1 and 2). million 7SL RNAs per cell (24), this result suggests that there are 100 Alu transcripts per cell. While this comparison is not A 240 nt primer extension product is observed in total likely to be an exact measure of the relative amount of two RNAs cytoplasmic HeLa RNA (Figure 1, Lane 5). Comparing which differ so gready in abundance, it does indicate the relative equivalent masses of the cytoplasmic poly A + and A— fractions paucity of the Alu transcripts. As an alternative measure of the (Lanes 3 and 4), the product results from a polyadenylated RNA. abundance of the Alu transcript, we find that the primer extension Comparing equivalent cellular amounts of nuclear and poly A + product resulting from 12 ng of an in vitro Alu transcript is cytoplasmic RNA, the Alu product appears to be more abundant equivalent to the product resulting from 5 x 107 cell equivalents in the cytoplasmic fraction (Figure 1). Replications of this of poly A + RNA. This comparison suggests that there are one experiment, comparing either equvalent cellular amounts or equal thousand copies of Alu 240 nt RNA per cell. masses of nuclear and cytoplasmic RNA agree with a cytoplasmic poly A+ identification of this transcript. Sequence studies In addition to the 240 nt product, a longer 350 nt primer reported below show that this primer extension product includes extension product is routinely observed in the poly A + Alu sequences having 5' ends corresponding to the known cytoplasmic fraction (Figure 1). Corresponding products have initiation site for Pol Ill-directed transcription, so that the 240 been observed in primer extension experiments using other nt primer extension product includes authentic Alu transcripts. primers at other positions within the Alu consensus sequence These findings confirm previous reports that Alu transcripts are thereby mapping the start site of this transcript 110 nt upstream cytoplasmic (5, 6). from the interspersed Alu element (6). As described below, this product results from an Alu repeat positioned near the 5' end The scarcity of this transcript is noteworthy. We attempted to of what is probably an mRNA. compare the relative abundance of the Alu transcript and 7SL Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 1090 Nucleic Acids Research, 1994, Vol. 22, No. 6 A 123456789 10 11 12 6 7 8 9 10 11 28s • * 18s . ft tRNA- :• * | / f C 45678 9 10 11 « • I § $ i • -7SL(128) « ft ft I Figure 2. Sucrose gradient fraction of cytoplasmic RNA in the presence of EDTA to dissociate nbosomal subunits. A shows ethidium bromide stained agarose gel electrophoresis of the reulting sedimentation fractions after protein extraction. The RNA fractions were next separated into poly A+ and poly A - fractions (B). Primer extension of the poly A+ RNA was performed with the 17mer primer for AJu repeats (B) and pnmer extension of the poly A - RNA was performed with ohgo 783 for 7SL RNA (C). Sucrose gradient fractionation of cytoplasmic components these alternatives. However, as shown below by Northern blot hybridization, the Alu transcripts are 300 to 500 nt in length, corresponding to 8.5S. We therefore believe that the Alu To determine the subcellular localization of the Alu transcript, cytoplasmic RNPs were fractionated by sucrose gradient transcripts are probably organized into RNPs. sedimentation in the presence of EDTA which dissociates the Polyribosomes were fractionated from monoribosomes and ribosomal subunits (Figure 2A). As assayed by the 240 nt primer other RNPs by sedimentation in the presence of magnesium ions extension product, the sedimentation velocity of Alu transcripts (Figure 3A). Again as assayed by the 240 nt primer extension (Figure 2B,ca. Fractions 6 to 8) is intermediate between that of product, most Alu transcripts sediment at about the same velocity the 40S ribosomal subunit (18S RNA, Figure 2A, Fractions 4 as 7SL RNA-containing SRPs (Figures 3B, 3C, Fractions 8-10). and 5) and free tRNAs (Figure 2A, Fractions 10 and 11) and These findings do not preclude a translational role for the Alu- is approximately coincident with the sedimentation velocity of RNP but merely show that it. like SRP, is not an obligatory 7SL RNA (Figure 2C, Fractions 7 to 9), which is quantitatively ribosomal component (Discussion). contained in the signal recognition particle (SRP) sedimenting Interestingly, the sedimentation velocities of the 240 nt (Figure at 1 IS. These findings suggest that the authentic Alu transcripts 3B, Fractions 8-10) and 350 nt (Figure 3B, Fractions 3-8) are associated with RNPs sedimenting at about 1 IS or are high primer extension products are fully resolved when the ribosomes molecular weight RNAs. The paucity of the Alu transcript are stabilized by magnesium but are not resolved in the presence frustrated obvious experiments designed to distinguish between of EDTA (Figure 2B). This observation is evidence that the 350 Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 Nucleic Acids Research, 1994, Vol. 22, No. 6 1091 B345678 9 10 1 2345678 9 10 28s - 18s- tRNA- C 345678 9 10 7SL (128) *t! Figure 3. Sucrose gradient fractionation of cytoplasmic RNA in the presence of MgCl to stabilize ribosomes. A shows ethidium bromide stained gel electophoresis of the resulting sedimentation fractions after protein extraction. The RNA fractions were next separated into poly A+ and poly A - fractions. Primer extension of the poly A+ RNA was performed with the 17mer primer for Alu repeats (B) and primer extension of the poly A - RNA was performed with oligo 783 for 7SL RNA (C). nt product results from a ribosome-associated mRNA; its labeling of restriction digests of Alu sequence preparations. A enrichment in the poly A+ fraction supports this interpretation. BstUI site, CpGpCpG, is fortuitously located in the Alu A box and can be cleaved when it is unmethylated (18). Double digestion Also, the base sequence reported below shows that the 350 nt at the BstUI site and the consensus Tthllll site 260 nt primer extension product includes transcripts from unique loci and that the Alu elements within these sequences are not correctly downstream releases the predicted band as detected by blot positioned to direct Pol III transcription of the corresponding hybridization if the BstUI site is unmethylated (Figure 4A, Lanes RNA. This interpretation is supported by previous primer 1 and 3). As a control on the total amount of DNA, double digestion at the consensus TaqI and Tthl 1II sites releases a 189 extension studies (6) which map the start site of the corresponding nt band (Figure 4A. Lanes 2 and 4). BstUI cleaves a minor but transcript to a position 110 nt upstream from the Alu element. significant fraction of the Alu sequences in untreated HeLa cells Effect of DNA methylation on transcript abundance (Figure 4A, Lane 3). Treatment of HeLa cells with 5-azacytidine substantially increases the susceptibility of the Alu BstUI site to Depending on the tissue type, as much as 30% of the cleavage, showing the demethylation of this site in some Alu 5-methylcytosine content of human DNA resides within Alu members (Figure 4A, Lane 1). Methylation of the BstUI site repeats (16, 18). HeLa cells were treated with 5-azacytidine to within the A box is sufficient to repress Pol El-directed reduce the level of DNA methylation (Materials and Methods). transcription of AJu templates (13). To test the methylation status The efficacy of this treatment was demonstrated by two different of other sites, the Alu repeats from 5-azacytidine treated and experiments: blot hybridization of genomic DNA digests and end Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 1092 Nucleic Acids Research, 1994, Vol. 22, No. 6 effect of methylation on the abundance of the Alu transcript A 12 3 4 (Figure 4B). 396 - Northern blot analysis confirms and extends the finding that 5-azacytidine treatment has a pronounced effect on the abundance of the Alu transcript (Figure 4C). In agreement with previous 214 - ** observations, two principal length transcripts are observed by blot hybridization. The shorter transcript in the poly A - fraction (Lanes 1 and 2) corresponds to 118nt scAlu RNA (5. 7). The longer transcripts in the poly A + fraction (Lanes 3 and 4), ca. 300 to 450 nt, are the sizes expected for Pol Ill-directed Alu transcripts (Introduction; 5, 6, 7). The high molecular weight smear of hybridization possibly results from a combination of Control 5'AzaC I II I longer Pol Ill-directed Alu transcripts and the presence of Alu 1 2 3 4 1 6 64 1 6 6.4 repeats in other transcription units. Comparing Lanes 3 and 4, treatment with 5-azacytidine increases the abundance of the presumptive Pol Ill-directed Alu transcripts by about eight-fold, in satisfactory agreement with the results of primer extension discussed above. Interestingly, 603 - the effect of 5-azacytidine treatment on the abundance of the 118 •350 nt scAlu transcript is much less pronounced (Figure 4C; Discussion). Reprobing the same Northern blot shows that the 310 - the steady state level of 5S rRNA is not affected by 5-azacytidine 281 - treatment (data not shown). This finding also provides an internal control for the amount of RNA loaded in Lanes 1 and 2 so that the small increase in the abundance of scAlu is significant. 240 Authenticity and identity of the Alu transcripts Coincidence of the 3' end of a primer extension product and the 118 - initiation site for Pol Hi-directed Alu transcription would provide the most direct evidence available that the primer extension products result from authentic Alu transcripts. Also, the A box of the RNA Pol III promoter element is located very near the 5' end of the Alu sequence. For these reasons, a cloning strategy Figure 4. Effects of 5-azacytidine treatment on HeLa DNA and Alu RNA. which maps and preserves the 3' end of the primer extension A: Blot hybridization of genomic DNA extracted from 5-azacytidine treated (Lanes product (i.e., the 5' end of the corresponding Alu transcript) was I and 2) and untreated cells (Lanes 3 and 4) was performed on DNAs digested with BstUI and Tthlll l (Lanes 1 and 3) and Taql and Tthllll (Lanes 2 and adopted (22). The primer extension product was first tailed with 4). Probe was the Alu specific oligonucleotide PV 51. Marker lengths in nt are oligo dA and then linearly amplified using an oligonucleotide that shown on the left-hand side. B: Pnmer extension of poly A + cytoplasmic RNA consists of an adapter primer linked to a short run of Ts, thereby extracted as indicated from 5-azacytidine treated cells (eight days) and untreated 'capping' the 5' end of the resulting cDNA sequence. By using control cells using the 17mer Alu specific oligonucleotide primer. Either 1.6 /ig a non-Alu homologous primer for PCR, this method decreases or 6.4 /ig of RNA was used as indicated. Similar results are obtained upon treating cells for twenty days with 5-azacytidine. C: Northern blot analysis of poly A — the likelihood of inadvertant amplification of Alu family members (Lanes 1 and 2) and poly A + (Lanes 3 and 4) cytoplasmic RNA isolated from interspersered in high molecular weight nucleic acids (7, 22; HeLa cells grown for eight days in the presence (Lanes 2 and 4) and absence Discussion). The resulting product was PCR amplified with the (Lanes 1 and 3) of 5-azacytidine. adapter primer and an additional Alu primer that is nested with respect to the oligonucleotide originally used for reverse transcription. The resulting clones were subsequently screened untreated cells are first isolated as a 189 bp restriction fragment by gel electrophoresis to discard cDNAs (approximately half) resulting from a Taql and Aspl (Tthl 1II isoschizomer) digest which were obviously truncated with respect to the full length of the DNAs (18). There are three consensus sites for HpaTI and Alu consensus sequence; about one-half of the remaining clones its isoschizomer within this Alu fragment. The resulting were sequenced. preparation is next subjected to digestion with either Hpall, which does not cleave its methylated site, or Mspl, which is indifferent With respect to their 5' ends as defined by the poly T tail, to methylation, followed by direct end labeling (18, 23). A 67 three groups of cDNAs were obtained by this procedure: four nt Hpall indicator band released from Alus in DNA from cDNAs are truncated before the 5' end of the Alu consensus 5-azacytidine treated cells is about twice as intense as the same sequence; eight begin precisely at the Pol III transcription start band from untreated cells (data not shown). In summary, site; and eleven sequences have one to seven additional 5' 5-azacytidine treatment results in hypomethylation of several nucleotides, usually A residues (Figure 5). The truncated cDNAs consensus Alu CpGs. are presumably products of incomplete reverse transcription and could result from either authentic Pol III Alu transcripts or Alus As assayed by the 240 nt primer extension product, the Alu that were transcribed within longer transcription units from other transcript is more abundant (ca. five-fold) in 5-azacytidine treated promoters. The tendency of reverse transcriptase to terminate cells than in the control cells. The level of the 350 nt product prematurely within the G-rich 5' end of Alu sequences is as well as other bands of unknown identity are not altered by illustrated by truncated primer extension products of 7SL RNA 5-azacytidine treatment, providing an internal control for the Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 Nucleic Acids Research, 1994, Vol. 22, No. 6 1093 GGCC*GQqCQCQGTQqCT — GAQATCQAQACCAT — (Alu consensus) (Promoter consensus) RRYNNRRYGG GATCRANNCC Sub.Fa» . # Hit . fCpQ CLONES (T) PV 1 19 4 *..A — Precise 28 8 6 * A — Precise 9 15 12 *...T — Precise 13 12 A5 A...* A — Major 23 7 5 A A — ...T. J G Major 20 8 8 ....C — ... T G Major 15 7 9 ....*....C — ... T G Major 15 10 A4 *...T — ... T G 17 CT * — PV 10 15 7 AA * — Precise 10 17 A7 CTGTT...T* AT — I... Precise 17 8 Precise 11 13 A6 ACAGTT...** — . . .fl-A 1 A *.A£...A — ... T G Major 24 5 A3 A * A A... — ...T.. A G Major 29 5 A2 CT. . .T . . .T.T A — ...T.. A G Major 28 4 1 5 GAA *A CA..C.A. — ...T.. A A Major 31 7 Major 19 5 1 0 AAAAA *...T...S — ...T. J A--G 18 6 1 6 AAAAA * — ...T.. A A Major 24 8 Major 3 GTTCCCT *...£...! — ...T G .T. I G Major 350-2 ...T*A. . .AT . . 27 4 350-3 GT..*C. ..A.A... .T . .C G Major 33 7 ******* Al . . CG 9 Precise 21 *****, 13 .T.T .T 6 T G Major 26 ******* 2 A. . Major 25 11 T G **************. 11 Major 15 8 Figure 5. Sequence analysis of Alu cDNA clones Comparison of Alu cDNA sequences to the Alu consensus sequence for positions 1 — 17 and positions 73 — 86 (2 and references therein) and the A and B boxes of the Pol III promoter consensus (33). Clones 1 - 17 and clones Al -A 7 (5-azacytidine treated cells) are derived from the 240 nt pnmer extension product. Clones 350-2 and 350-3 are from the 350 nt band. The subfamily identity, the number of mismatches with respect to the subfamily consensus sequence, and the number of CpG residues are listed for each clone. - indicates uncompared sequence at a position. • indicates base identity to the consensus. * indicates a base deletion. Base substitutions are shown by the appropriate letter. Bold letters indicate a substitution with respect to the Pol in promoter consensus. These sequences have been assigned GenBank accession numbers U02044-U02068. reported in Figures 2 and 3. Consequently, incomplete Alu cDNA CT or from an authentic Alu transciprt which artifactually sequences are not informative about the source of their parent acquired a single C residue at position - 2 (Figure 5). The transcripts and will not be examined further. We interpret the assignment of Clone A2 is similarly ambiguous. Both are formally full length cDNAs as primer extension products resulting from included in the class as each has flanking non-T residues. Pol III Alu transcripts (Figure 5). For the same reason, we In addition to cDNAs corresponding to the 240 nt primer conclude that the 240 nt primer extension products examined in extension product, two Alu-containing cDNAs (Clones 350-2 and the previous experiments (Figures 1 through 4) are a faithful 350-3) were derived from the 350 nt primer extension product. indication of authentic Alu transcripts. Regarding the eleven These two cDNAs certainly result from longer primary transcripts cDNAs which have additional residues on their 5' ends, many and are therefore pooled with the eleven Alu cDNAs, discussed of these cannot be attributed to Pol in transcription from the above, giving a total of thirteen cDNAs in this class (Figure 5). promoter of the interspersered Alu repeat (Discussion). Most of In agreement with previous reports, each full length Alu cDNA these eleven cDNAs probably result from longer transcripts which corresponds to a distinct Alu repeat, showing that many different happen to have contained an interspersed Alu element Alu loci must be transcribed (Figure 5) (6, 7). Included among (Discussion). For simplicity, we refer to these as 'interspersed' the full length cDNA sequences are four representatives of the Alu transcripts. However, the assignment of Clone 17, and Major subfamily, three representatives of the Precise subfamily perhaps also of Clone A2, as transcripts may result from a cloning and one representative of the PV subfamily (1, 2). The PV artifact which deserves special comment. Clone 17 has a single subfamily Alu member was obtained by screening cDNA clones C residue in what would otherwise be a continuous perfect poly using oligonucleotide hybridization and is not statistically T sequence generated by the tailing strategy described above. represented in this small sampling of cDNA sequences. However, In another cDNA, a single C residue was observed in the middle relative to their genomic abundance, the three Precise subfamily of its, long poly T run and was undoubtedly artifactually members are overrepresented compared to the four Major introduced during PCR amplification (unpublished). Clone 17 subfamily members (Figure 5). This finding agrees with previous either results from a Alu transcript flanked by the dinucleotide observations that representatives of younger Alu subfamilies are Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 1094 Nucleic Acids Research, 1994, Vol. 22, No. 6 transcriptionally more active than older subfamily members (5, RNAs detected here by many orders of magnitude and 6,7) . contamination resulting from these transcripts is probably There is a difference in the conservation of Pol III promoter unavoidable. Sinnett et al use an oligonucleotide primer that elements (32) between cDNAs corresponding to authentic Alu corresponds to the 5' end of the Alu consensus sequence for transcripts and cDNAs corresponding to Alu transcripts. All eight second strand cDNA synthesis (6). As they discuss, that cloning full-length Alus have perfect consensus A boxes and only one strategy does not preserve the 5' end of the parent RNA and cDNA (Clone 5) has a nonconsensus B box substitution. Four therefore cannot unambiguously discriminate betwen Alu RNAs of the thirteen Alu cDNAs have nonconsensus consensus A boxes which result from Pol II- and Pol Ill-directed transcription. In this study, less than half of the observed cDNAs result from and four have nonconsensus B boxes. authentic Alu transcripts as judged by their 5' ends; the majority Individual cDNAs differ from their respective consensus result from Alus interspersed within other transcription units. sequences at various base positions (Figure 5). The authentic Alu transcript cDNAs have an average of 15.5 (7.0%) differences Determination and regulation of Alu template activities from their consensus sequences; for the Alu cDNAs the average is 20.8 (9.5%) differences. The slight difference between these The steady state abundance of Alu RNA observed here is very two averages mostly reflects the different subfamily composition low. The expression of Alu repeats may be determined by several of the two classes of cDNAs; the average divergence is 8% for factors, including their DNA methylation, structural variation of the Precise subfamily and 13.5 % for the Major subfamily (25). their internal Pol III promoters, sequence context, and lifetime of the resulting transcripts. Remarkably, the sequence of Clone 4 differs by only a single nucleotide from the PV Alu consensus. The average CpG content Methylation represses transcription of several different Pol Hi- of the two classes is also not very different. The authentic Alu directed templates in vivo, and in particular represses the transcript cDNAs have an average of 11 CpGs; the Alu cDNAs transcription of Alu repeats in vitro (13, 17). Stimulation of Alu have an average of 9.4 CpGs (Figure 5). In summary, the single expression in HeLa cells by 5-azacytidine treatment and its distinguishing feature for these two classes of cDNAs is the concomitant effect on DNA methylation provide evidence that fidelity of their A box elements. This element is more faithfully the transcription of Alu repeats, like that of other Pol Ill-directed represented in younger Alu members, which in rum are somewhat templates, is repressed by DNA methylation. Alu repeats are almost completely methylated in certain somatic tissues and enriched in the authentic Alu transcript cDNAs. The A box is interestingly, certain Alu subgroups are almost quantitatively especially important in determining the strength of a Pol III unmethylated in the male germ line (16, 17, 18). Thus it is likely promoter (13 and references therein). that the low level of Alu expression is attributable, at in least The Alu element in Clone 350-2 is preceded by an 85 nt part, to transcriptional repression caused by extensive sequence and similarly the Alu element in Clone 350-3 is methylation. A small number of Alu repeats in HeLa cells are preceded by a different 90 nt sequence. Neither of these two hypomethylated (18) and plausibly this subgroup includes more flanking sequences is present in sequence data banks and, in active Alu templates. Methylation of sequences is often particular, neither is a known repetitive sequence. Primer determined by their sequence context (27), and more active extension of poly A-I- cytoplasmic RNA using an oligonucleotide hypomethylated Alu templates might reside in a favorable derived from Clone 350-2 did not yield product, but primer sequence context. Chromatin structure and DNA methylation extension using an oligonucleotide from Clone 350-3 yielded synergistically repress Alu transcription in vitro (28). several discrete-length primer extension products, including one that is consistent with the structure of the 350 nt primer extension Sequences flanking certain Pol Ill-directed templates, such as product described above: an Alu positioned 110 nt downstream the 7SL RNA gene, significantly stimulate their transcriptional from a transcription start site. We therefore tentatively assign activity (29). In the specific case of an Alu source gene (30), cDNA Clone 350-3 to the mRNA associated with the 350 nt cis acting elements stimulate its transcription in vitro by about primer extension product. ten-fold (Chesnokov and Schmid, unpublished). Alu transcriptional activity is probably subject to both positive and negative regulation. DISCUSSION More transpositionally active source genes closely match their Technical solution to a long-standing problem consensus sequences, raising the question of why these sequences Authentic but relatively rare Alu transcripts have been difficult have a transpositional advantage over those which have suffered to identify unambiguously against the very high background of a higher number of base substitutions (1,2). The relative template activity of Alus in vitro is in part determined by the fidelity of Alu sequences interspersed within hnRNA and even some mature mRNAs (Introduction). Many Alu cDNA sequences observed the internal Pol III promoter sequences. The A box is especially here begin precisely at the 5' start site employed by Pol III important in determining the strength of a Pol III promoter (13). transcription in vitro and very likely represent authentic Alu The A box sequence in the Pol Ul-directed Alu transcripts exactly transcripts. Similar evidence shows that scAlu RNAs result from matches the consensus sequence for this element, which is less faithfully preserved in the interspersed transcripts. Pol IE-directed Pol III transcription (7). Alu cDNAs with extended 5' ends most likely result from transcripts that contain an interspersed Alu Alu transcripts reported here would likely have a transcriptional element although in certain cases, reverse transcriptase can add advantage relative to Alus having nonconsensus base substitutions one or two untemplated nucleotides to a primer extension product in their Pol III promoter elements. (26). The additional bases are mostly As. The direct repeats This finding may partially explain the selective advantage that flanking Alu elements are usually A-rich, suggesting the some potential Alu source genes enjoy over others. Younger Alu probability that these additional nucleotides were templated by repeats closely match their respective subfamily consensus the parent RNA. Polymerase II-directed Alu transcripts comprise sequences, including the Pol III promoter consensus sequence. 10% of hnRNA, exceeding the small number of cytoplasmic Conservation of the promoter elements may therefore provide Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 Nucleic Acids Research, 1994, Vol. 22, No. 6 1095 6 Sinnett. D.. Richer. C . Deragon. J M. and Laboda, D. (1992) J. Mol. Bid.. younger Alu sequences with a relative transcriptional advantage 226. 689-706. and a consequent retrotranspositional advantage (2, 13). Other 7. Maraia. R.J.. Driscoll. C . Bilyeu. T.. Hsu. K.. and Darlington, G.J. (1993) factors, such as the structure and lifetime of the RNA, in addition Mol. Cell Biol.. 13, 4233-4241 to methylation and positive cis acting elements discussed above, 8. Weiner, A.M.. Deimnger. P.L. and Efstratiudis. A. (1986) Ann. Rev. would all select among the many potential source genes (6). Biochcm., 55, 631-661. 9. Schmid. C.W., Deka. N. and Matera. A.G. (1990) In Adolpf, K.W.(ed) Chromosomes: Eukaryotic, Prokaryotic and Viral. CRC Press. Boca Raton. Possible functions of Alu transcripts Honda, pp 323-358. Considering only authentic (i.e., Pol Ill-directed) transcripts, two 10. Chang. D.-Y and Maraia. R.J. (1993) J. Biol. Chem.. 268, 6423-6428. different forms of Alu RNA are present in the cytoplasm: left 11 Elder. J. T.. Pan. J., Duncan. C.H. and Weissman. S. M. (1981) Nucleic Acids Res.. 9. 1171-1189 monomer scAlu RNA and full length Alu RNA (5, 6, 7). ScAlu 12 Fuhrman. S.A.. Deininger. P.L.. LaPorte, P . Friedmann. T. and Geiduschek. RNA is thought to result from the processing of full length Alu E.P (1981) Nucleic Acids Res.. 9, 6439-6456. transcripts, presumably in the nucleus. The cytoplasmic 13. Liu. W. M. and Schmid. C W. (1993) Nucleic Adds Res.. 21, 1351 - 1359. localization of the full length Alu transcripts and their association 14. Jang K. L. and Latchman. D.S. (1989) FEBS Len.. 258, 255-258. with RNPs suggest that the full length and scAlu RNAs may have 15. Panning. B. and Smiley. J.R. (1993) Molec. and Cell Biol., 13. 3231-3244. distinct cytoplasmic functions or activities. Also, 5-azacytidine 16. Schmid, C.W. (1991) Nucleic Acids Res.. 19, 5613-5617. 17. Kochanek. S.. Renz. D and Doerfler. W. (1993) EMBO Journal 12. treatment increases relative abundances of the full-length and left 1141-1151. monomer transcripts by different degrees. Among other 18. Hellmann-Blumberg. U.. Hintz. M.F. and Schmid. C.W. (1993) Mol. Cell possibilities, this difference might result from limiting factors Biol.. 13. 4523-4530. necessary for the accumulation of scAlu RNA. 19 Snyder, R and Lachmann. P. (1989) Mutation Res.. 226. 185-190. 20. Foster. R.. Jahroudi. N. and Gedamu. L. (1991) Biochim. Biophvs. Ada. The presence of 100 to 1000 cytoplasmic Alu transcripts per 1088. 373-379 cell, possibly organized as RNPs, suggests that the transcripts 21. Maniatis. T.. Fntsch. E F. and Sambrook. J. (1989) Molecular Cloning: have a function. The most obvious possibility is a role in some A Laboratory Manual Cold Spring Harbor University Press. Cold Spring aspect of translation. The Alu homolog, 7SL RNA, is an integral Harbor. 22. Maraia, R. (1991) Nucleic Acid Res., 19. 5695-5702. component of SRP and one function of SRP is translation arrest 23. Monk. M.. Boubelik. M. and Lchnert. S. (1987) Development. 99. 71-382. (31). SRP proteins bind the Alu-like region of 7SL RNA and 24. Baserga, S. J. and Steitz. J.A (1993) The RNA World. Cold Spring Harbor the 68 kD subunit of SRP forms complexes with Alu transcripts Press. Cold Spring Harbor, pp. 359-381. in vitro (31,32). Interestingly the major families of short 25. Willard. C . Nguyen. H. T . and Schmid. C.W. (1987) J. Mol. Evol., 26, interspersed repeats in mammals can be classified into two 180-186. 26 Swanstrom. R., Varmus. H.E.. and Bishop. J.M. (1981) J. Biol. Chem., sequence superfamilies. Primate Alu repeats and rodent Bl 256. 1115-1121. sequences are homologous to 7SL RNA (1, 2). Rodent B2 27 Bird. A. P. (1992) Cell. 70. 5-8 . sequences, rabbit C repeats, and various other repeat sequence 28 Englander, E.W..Wolffe. A.P. and Howard. B.H. (1993) J. Biol. Chem., families in various mammalian species are homologous to tRNA 268. 19565-19573 sequences. Both superfamilies are ancestrally related to sequences 29. Bredow. S.. Sung, D.. Muller, J.. Kleinert. H. and Benecke. B.-J. (1990) Nucleic Acids Res.. 18, 6779-6784 which have a translational role. Thus there are several reasons 30 Leeflang, E.P.. Liu. W-M . Chesnokov, I. and Schmid. C.W. (1993) J. to guess that an Alu RNP might have a translational role. The Mol. Ewl.. .37. 559-565 sequence similarity between a short region of 28S rRNA and the 31 Strub. K.. Moss, J.. and Walter. P (1991) Mol. Cell. Biol., 11. 3949-3959. human Alu sequence (Results) might merely be an accidental 32. Andrews. P G. and Kole. R. (1987) J. Biol Chem.. 262, 2908-2912. match or it might result from such a functional relationship. Alternatively, the presence of the full-length transcript in HeLa cells might result from the leaky repression of Alu templates, and while these transcripts may not have a necessary function in HeLa cells they might have such a function in other tissues and cell types in which Alu repeats are appropriately expressed. ACKNOWLEDGEMENTS This work was supported in part by USPHS grant GM 21346 (C.S.), the Agricultural Experiment Station at the University of California, Davis. (C.S.), and a Jastro Shields Scholarship Award (W.M.L.). We thank Ms Sandra Malakauskas for her invaluble technical assistance. REFERENCES 1 Deininger. P L.. Batzer. M.A.. Hutchison III. C.A and Edgell. M H (1992) Trends in Genet . 8. 307-311. 2 Schmid. C W and Maraia. R (1992) Current Opinion in Genetics ami Development• Genomes and Emlution, 2. 874 — 882. 3 Liu. W-M . Leeflang. E.P. and Schmid. C W (1992) Biochim Biophvs Actu. 1132. 306-308. 4. Paulson. E. K. and Schmid. C.W. (1986)NucleicAddsRcs., 14,6145-6158 5 Matera. A.G.. Hellmann-Blumberg, U and Schmid. C W. (1990) Mol. Cell. Biol , 10. 5424-5432. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nucleic Acids Research Oxford University Press

Alu transcripts: cytoplasmic localisation and regulation by DNA methylation

Download PDF
9 pages

Loading next page...
 
/lp/oxford-university-press/alu-transcripts-cytoplasmic-localisation-and-regulation-by-dna-n1svmvJt9c

References (10)

Publisher
Oxford University Press
Copyright
© Oxford University Press
ISSN
0305-1048
eISSN
1362-4962
DOI
10.1093/nar/22.6.1087
Publisher site
See Article on Publisher Site

Abstract

Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 Nucleic Acids Research, 1994, Vol. 22, No. 6 1087-1095 © 1994 Oxford University Press Alu transcripts: cytoplasmic localisation and regulation by DNA methylation Wen-Man Liu1, Richard J.Maraia3, Carol M.Rubin1 and Carl W.Schmid1'2* 1Section of Molecular and Cellular Biology and department of Chemistry, University of California, Davis, CA 95616 and 3National Institute of Child Health and Human Development, NIH, Bethesda, MD 20892, USA Received September 30, 1993; Revised and Accepted February 7, 1994 EMBL accession nos U02044 - U02068 (incl.) ABSTRACT Full length Alu transcripts in HeLa cells are detected sequences and some Alu sequences persist in mature mRNA (8, by primer extension using reverse transcriptase and are 9). We refer to Pol IE-directed Alu transcripts as being authentic transcripts, as distinct from Alus that are transcribed by virtue also analyzed as cloned cDNA sequences. The 5' end of these transcripts corresponds to the transcriptional of their interspersion within other transcription units. The start site for RNA polymerase III indicating that these interspersion of Alu sequences in Pol U-directed transcripts RNAs are transcribed from their internal polymerase III precludes the use of S1 mapping and nuclear run on assays to promoters. The Alu transcripts found in cytoplasmic detect Pol HI Alu transcripts confidently. Many Alu hybridization probes cross hybridize with abundant 7SL RNA complicating poly A + RNAs appear to be organized into RNPs as assayed by sucrose gradient sedimentation. Present at the use of Northern blot hybridization analysis (5, 6, 7, 10). about one hundred to one thousand copies per cell, the Primer extension by reverse transcriptase tests whether the 5' Alu transcripts are rare as compared to 7SL RNA. In end of an Alu transcript corresponds to the initiation site expected agreement with previous reports that methylation for Pol Hi-directed transcription. However many Alu primer extension products are prematurely truncated, presumably due inhibits Pol Ill-directed transcription of Alu in vitro, treatment of HeLa cells with 5-azacytldine results in Alu to secondary structures in the GC-rich Alu sequence. This DNA hypomethylation and an increase in the problem is further compounded by the abundance of Alu abundance of the Alu transcript. Sequence analysis homologs in both 7SL RNA and Pol H-directed transcripts, which shows that many different Alu repeats including can introduce a high background of prematurely truncated primer extension products as compared to primer extension products members of all subfamilies are transcribed by Pol III In vivo. cDNA sequences of the Pol Ill-directed transcripts resulting from relatively rare Alu transcripts (5, 6, 7). exactly match the A box of the Pol III promoter element Given these technical challenges, little is known about the whereas in other Alu transcripts this element is not regulation and fate of Alu transcripts. Full-length Alu transcripts, faithfully conserved. which can vary in size from 300 to 500 nt, have been detected in total RNA and in the polyadenylated (poly A+ ) cytoplasmic fractions (5, 6, 7). This size heterogeneity is presumably due INTRODUCTION to variable 3' ends as the Alu Pol III promoter initiates transcription at the first base of the Alu repeat (11, 12). Alu Human Alu repeats are thought to retrotranspose through an RNA repeats belong to subfamilies of different evolutionary ages and intermediate (1, 2). Also, Alu repeats contain an internal RNA while observed Alu transcripts represent all subfamilies, they are polymerase III (Pol III) promoter and most are transcribed in enriched in young subfamily members (6, 7). Results from in vitro by Pol III (1, 2, 3). Despite the abundance of Alus and vitro transcription assays indicate that the A box element of the their transcriptional potential, corresponding Pol HI transcripts internal Pol HI promoter sequence might be a critical determinant are very rare in tissues and cell lines examined to date and have of the relative transcriptional activity of Alu repeats (13). Full been extremely difficult to demonstrate convincingly (4). length Alu cDNA sequences available through one study do not However, identification of Alu subfamily consensus sequences test this prediction, as the A box sequences were replaced during (1,2) has allowed the use of specific oligonucleotide probes to the PCR cloning step which generated the cDNA clones (6). demonstrate that Alu transcripts do indeed accumulate in vivo Although the Alu consensus sequence is a dimeric structure, (5, 6, 7). shorter transcripts corresponding to Pol Ill-initiated transcripts Unusual problems that prevent or complicate the use of many of the left Alu monomer have also been observed (5, 7, 10). conventional techniques to study Alu transcription deserve These transcripts accumulate in the nonpolyadenylated (poly A—) attention. Because Alus are ubiquitously interspersed in other fraction of the cytoplasm and are designated small cytoplasmic transcription units, about 10% of pre mRNA consists of Alu *To whom correspondence should be addressed Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 1088 Nucleic Acids Research, 1994, Vol. 22, No. 6 Alu (scAlu) RNA. Some full length dimeric Alu transcripts may 81 -10 3 of the PV Alu subfamily consensus sequence (2). Oligo be precursors to monomeric scAlu RNAs (7). 783, 5'-ATGCCGAACTTAGTGCGG-3', is complementary to positions 112-129 (S-domain) of 7SL RNA. Oligo 2334. Factors provided or induced by adenovirus and herpes virus 5'-GCCGGGACAACACATT-3', is complementary to positions increase the steady state abundance of Alu containing transcripts 82-9 7 of Clone 350- 3 RNA. Adapt-TI7, 5'-GACTCGAGT- (14, 15). Those studies do not distinguish between scAlu and CGACATCGA(T)17-3\ Adapt-1, 5'-GACTCGAGTCGAC- full length Alu transcripts and the mechanism of this stimulation ATCGA-3', and 18mer, 5'-GAGAATTCTCGGCTCACTGC- is unknown. Alu repeats are heavily methylated in DNAs isolated A-3' were used for PCR amplification of Alu cDNA as shown from various somatic tissues and certain Alu subgroups are on the diagram below. The 18mer nests one nucleotide upstream undermethylated or almost completely unmethylated in sperm to the 17mer. DNA (16, 17, 18). In vitro transcription of Alu templates is inhibited by a repressor which is specific toward methylated DNA (Adapt-1) (13). <-EcoRI (18mer) > (Adapt-T17) Here we: i) examine the subcellular localization of Alu > (PV51) <— (17mer) transcripts; ii) test the effect of DNA methylation on the A L U- > A A A A A I— • expression of Alu repeats in vivo; and iii) examine the Pol III 0 240nt intragenic promoter structure of Alus that generate full length transcripts. Primer extension assays RNA in 20 /*l of annealing buffer [10 mM Tris-HCl pH 7.5, MATERIALS AND METHODS 1 mM EDTA, 0.3 M KC1, 1 unit Inhibit-ACE (5prime-3prime)] RNA preparation and 5-1 0 pmole (about 105 cpm) of a primer were incubated HeLa cells were grown to 4-6 x 105 cells per ml in spinner in boiling water for 2 minutes and then allowed to anneal at 42°C flasks at 37°C in SMEM (Minimum Essential Medium, GIBCO) for 1 hour. Primers were labeled by T4 polynucleotide kinase containing 5 % calf serum. Cells were also grown for either eight with (7-32P)-ATP and were ethanol precipitated. Following or twenty days in the presence of 8/iM 5-azacytidine (19, 20). addition of 80 /tl of reverse transcriptase solution (10 mM Cells were washed once in 10 volumes of 4°C PBS (phosphate Tris-HCl pH 8.6, 5 mM MgCl , 5 mM DTT, 1 mM dNTP, buffered saline), collected by centrifugation, resuspended in five and 1 unit Inhibit-ACE), the reaction mixture was incubated with volumes of cytoplasmic RNA extraction buffer [140 mM NaCl, 200 units of MMLV reverse transcriptase (USB) for 3 hours at 35 °C. The reaction was stopped by the addition of 100 /xl phenol/ 1.5 mM MgCl , 10 mM Tris-HCl pH 8.4, 1 mM DTT, 0.5% chloroform mixture. The extracted aqueous phase was ethanol NP-40, 0.5% Triton X-100, 20 mM VRCs (vanadyl precipitated, and the DNA pellet was resuspended in 6.5 /*l ribonucleoside complexes)], and lysed by 20 — 30 strokes of a denaturing loading dye mix, denatured at 90°C for 5 min, and Thomas tissue grinder at 4°C. Nuclei were pelleted by electrophoresed in 5 — 8% polyacrylamide-8M urea gel. centrifugation at 2,000 rpm (Sorvall SS-34 rotor) for ten minutes. M13mpl8 ssDNA sequencing provided size markers. The gel The cytoplasmic supernatant was removed, and intact nuclei were was exposed to XAR X-ray film with a Quanta III intensifying resuspended in GnSCN buffer (4M guanidinium thiocyanate, 25 screen at — 70°C. mM Na-citrate pH 7.0, 0.5% sarkosyl, 0.1 M mercaptoethanol) with a Pyrex tissue grinder. Both nuclear and cytoplasmic Northern blot results fractions were extracted several times with equal volumes of phenol/chloroform and chloroform and were ethanol precipitated. RNA blotting was done from 5.5% polyacrylamide gel (40:1 The nuclear RNA was purified by CsCl centrifugation to remove acrylamide:bis) as previously described (10, 22). Cytoplasmic genomic DNA. Poly A + and poly A - RNAs were selected as RNA (65 ng) and poly A + RNA (19 /tg) were electrophoresed with 40 ng of sonicated salmon sperm DNA per sample. After described (21). soaking out urea with transfer buffer, the gel was transferred RNP isolations and sucrose gradients to Gene Screen Plus (NEN) at 30 volts overnight using a Bio Rad transblot cell with plate electrodes. The membrane was UV- PBS-washed cell pellets were suspended in sucrose gradient buffer crosslinked and dried throughly. The membrane was incubated (100 mM KC1, 10 mM Hepes pH 7.8, 1 mM DTT, 10 /ig/ml with hybridization solution (6xSSC, 2xDenhardt's, 0.5% SDS, cycloheximide), 0.5 mM PMSF (phenylmethylsulfonyl fluoride), 100 /ig/ml yeast RNA) at the hybridization temperature (58°C) 2 /tg/ml TLCK (N-a-p-tosyl-L-lysine chloromethyl ketone), 20 before hybridization for 18 hours. The end-labeled hybridization mM VRCs, and either 3 mM EDTA (for ribosomal subunit probe designated Alu-24 has been previously described (7, 22). fractionation) or 3 mM MgCl (to maintain poly somes). The membrane was washed twice with 2xSSC at room Cytoplasmic RNAs were fractionated by 5-3 5 % (3 mM EDTA) temperature and once at the hybridization temperature. or 18—42 % (3 mM MgCl ) sucrose density gradient Quantitation of the bands was done with the ImageQuant program centrifugation in a Beckman SW 40.1 Ti rotor at 4°C. of a phosphorimager (Molecular Dynamics). Denatured 0X174 Centrifugation was at 35,000 rpm for 7 hours for 5-35 % sucrose DNA was used as the size marker in this experiment. gradients and at 30,000 rpm for 2 hours for 18-42% sucrose gradients. Twelve equal volume fractions were collected. cDNA cloning and sequence analysis Oligodeoxyribonucleotides The 240 nt and 350 nt bands from primer extension products All oligonucleotides were custom synthesized. Oligo 17mer, were excised from the gel and electroeluted. Following poly (dA) 5'-GCGATCTCGGCTCACTG-3', corresponds to positions tailing, PCR amplification and EcoRJ/SaJI directional cloning into 238—221 of the Alu consensus sequence. Oligo PV51, 5'-A- pUC19 were performed as previously described (22) except that CCATCCCGGCTAAAACGGTGA-3', corresponds to positions the primers were Adapt-T17 and Adapt-1 (5'). and 18mer (3'). Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 Nucleic Acids Research, 1994, Vol. 22, No. 6 1089 Double strand DNA cycle sequencing (BRL) was employed for Cyto. sequence determinations. I I A + A' T N Assaying DNA niethylation For blot hybridization, 50 /*g of genomic DNA (from cells treated for twenty days with 5-azacytidine and untreated control cells) was digested overnight with 50 units of Tthl 111 (ca. ten fold overdigestion), phenol extracted, and similarly redigested with Tthl 111. After phenol extraction and ethanol precipitation the sample was divided into two equal aliquots for overnight digestion with either BstUI (50 units) or TaqI (27 units). After agarose gel (1.3%) electrophoresis, the samples were blotted onto 0.2 micron nitrocellulose and hybridized with oligo PV51 (16, 18). At the final washing temperature employed in this experiment (63°C), this oligonucleotide selectively hybridizes to the PV Alu subfamily (5, 18). Alu repeats were isolated by agarose gel electrophoresis as a visible 190 nt restriction fragment in TaqI —Aspl digests of genomic DNAs. Aspl is a Tthl 1II isoschizomer and the digestion conditions were similar to those described above. Equal aliquots of these preparations were subjected to HpaTJ and Mspl digestion and were terminally labeled with y-32P ATP using T4 polynucleotide kinase. The products were compared by acrylamide gel electrophoresis (18, 23). RESULTS Primer extension of Alu transcripts Initially, we used oligonucleotide primers situated exactly on the 3' end of the Alu consensus sequence, hoping to obtain full length primer extension products. However, this region includes an almost exact match to 7SL RNA and a close match to 28S rRNA. Figure 1. Primer Extension of control and HeLa cell RNAs. 60ng and 6ng of a control Alu template (7) were used as templates for primer extension with oligo [The 3 ' Alu consensus sequence is GAGCGAGACTCCGTCT- I7mer giving as expected a 240 nt cDNA as the principal product. 3 ^g poly C which closely matches the sequence GAGCGAGACCCGT- A + and 3 fig poly A— cytoplasmic RNAs, and cellular equivalents of total CGC beginning at position 943 in human 28S rRNA (see cytoplasmic RNA (60 /ig), and nuclear RNA (6 y.%) were each used as templates Discussion)]. Truncated primer extension products resulting from for primer extension reactions with oligo 17mer Two products of 240 and 350 these very abundant RNAs interfered with the detection of the nt were observed with cytoplasmic RNAs (poly A+ and total fractions). much rarer Alu transcript (data not shown). To avoid these complications, an oligonucleotide complementing positions 221 to 238 of the Alu consensus sequence which is totally dissimilar RNA by comparing the intensities of their respective primer to 7SL and ribosomal RNA sequences is used in the primer extension products (see below). The 7SL RNA primer extension extension assays reported here. Using a control template this product is about four orders of magnitude more abundant then primer, produces the expected, 238 nt, full length extension the corresponding Alu product. Assuming that there are one product (Figure 1, Lanes 1 and 2). million 7SL RNAs per cell (24), this result suggests that there are 100 Alu transcripts per cell. While this comparison is not A 240 nt primer extension product is observed in total likely to be an exact measure of the relative amount of two RNAs cytoplasmic HeLa RNA (Figure 1, Lane 5). Comparing which differ so gready in abundance, it does indicate the relative equivalent masses of the cytoplasmic poly A + and A— fractions paucity of the Alu transcripts. As an alternative measure of the (Lanes 3 and 4), the product results from a polyadenylated RNA. abundance of the Alu transcript, we find that the primer extension Comparing equivalent cellular amounts of nuclear and poly A + product resulting from 12 ng of an in vitro Alu transcript is cytoplasmic RNA, the Alu product appears to be more abundant equivalent to the product resulting from 5 x 107 cell equivalents in the cytoplasmic fraction (Figure 1). Replications of this of poly A + RNA. This comparison suggests that there are one experiment, comparing either equvalent cellular amounts or equal thousand copies of Alu 240 nt RNA per cell. masses of nuclear and cytoplasmic RNA agree with a cytoplasmic poly A+ identification of this transcript. Sequence studies In addition to the 240 nt product, a longer 350 nt primer reported below show that this primer extension product includes extension product is routinely observed in the poly A + Alu sequences having 5' ends corresponding to the known cytoplasmic fraction (Figure 1). Corresponding products have initiation site for Pol Ill-directed transcription, so that the 240 been observed in primer extension experiments using other nt primer extension product includes authentic Alu transcripts. primers at other positions within the Alu consensus sequence These findings confirm previous reports that Alu transcripts are thereby mapping the start site of this transcript 110 nt upstream cytoplasmic (5, 6). from the interspersed Alu element (6). As described below, this product results from an Alu repeat positioned near the 5' end The scarcity of this transcript is noteworthy. We attempted to of what is probably an mRNA. compare the relative abundance of the Alu transcript and 7SL Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 1090 Nucleic Acids Research, 1994, Vol. 22, No. 6 A 123456789 10 11 12 6 7 8 9 10 11 28s • * 18s . ft tRNA- :• * | / f C 45678 9 10 11 « • I § $ i • -7SL(128) « ft ft I Figure 2. Sucrose gradient fraction of cytoplasmic RNA in the presence of EDTA to dissociate nbosomal subunits. A shows ethidium bromide stained agarose gel electrophoresis of the reulting sedimentation fractions after protein extraction. The RNA fractions were next separated into poly A+ and poly A - fractions (B). Primer extension of the poly A+ RNA was performed with the 17mer primer for AJu repeats (B) and pnmer extension of the poly A - RNA was performed with ohgo 783 for 7SL RNA (C). Sucrose gradient fractionation of cytoplasmic components these alternatives. However, as shown below by Northern blot hybridization, the Alu transcripts are 300 to 500 nt in length, corresponding to 8.5S. We therefore believe that the Alu To determine the subcellular localization of the Alu transcript, cytoplasmic RNPs were fractionated by sucrose gradient transcripts are probably organized into RNPs. sedimentation in the presence of EDTA which dissociates the Polyribosomes were fractionated from monoribosomes and ribosomal subunits (Figure 2A). As assayed by the 240 nt primer other RNPs by sedimentation in the presence of magnesium ions extension product, the sedimentation velocity of Alu transcripts (Figure 3A). Again as assayed by the 240 nt primer extension (Figure 2B,ca. Fractions 6 to 8) is intermediate between that of product, most Alu transcripts sediment at about the same velocity the 40S ribosomal subunit (18S RNA, Figure 2A, Fractions 4 as 7SL RNA-containing SRPs (Figures 3B, 3C, Fractions 8-10). and 5) and free tRNAs (Figure 2A, Fractions 10 and 11) and These findings do not preclude a translational role for the Alu- is approximately coincident with the sedimentation velocity of RNP but merely show that it. like SRP, is not an obligatory 7SL RNA (Figure 2C, Fractions 7 to 9), which is quantitatively ribosomal component (Discussion). contained in the signal recognition particle (SRP) sedimenting Interestingly, the sedimentation velocities of the 240 nt (Figure at 1 IS. These findings suggest that the authentic Alu transcripts 3B, Fractions 8-10) and 350 nt (Figure 3B, Fractions 3-8) are associated with RNPs sedimenting at about 1 IS or are high primer extension products are fully resolved when the ribosomes molecular weight RNAs. The paucity of the Alu transcript are stabilized by magnesium but are not resolved in the presence frustrated obvious experiments designed to distinguish between of EDTA (Figure 2B). This observation is evidence that the 350 Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 Nucleic Acids Research, 1994, Vol. 22, No. 6 1091 B345678 9 10 1 2345678 9 10 28s - 18s- tRNA- C 345678 9 10 7SL (128) *t! Figure 3. Sucrose gradient fractionation of cytoplasmic RNA in the presence of MgCl to stabilize ribosomes. A shows ethidium bromide stained gel electophoresis of the resulting sedimentation fractions after protein extraction. The RNA fractions were next separated into poly A+ and poly A - fractions. Primer extension of the poly A+ RNA was performed with the 17mer primer for Alu repeats (B) and primer extension of the poly A - RNA was performed with oligo 783 for 7SL RNA (C). nt product results from a ribosome-associated mRNA; its labeling of restriction digests of Alu sequence preparations. A enrichment in the poly A+ fraction supports this interpretation. BstUI site, CpGpCpG, is fortuitously located in the Alu A box and can be cleaved when it is unmethylated (18). Double digestion Also, the base sequence reported below shows that the 350 nt at the BstUI site and the consensus Tthllll site 260 nt primer extension product includes transcripts from unique loci and that the Alu elements within these sequences are not correctly downstream releases the predicted band as detected by blot positioned to direct Pol III transcription of the corresponding hybridization if the BstUI site is unmethylated (Figure 4A, Lanes RNA. This interpretation is supported by previous primer 1 and 3). As a control on the total amount of DNA, double digestion at the consensus TaqI and Tthl 1II sites releases a 189 extension studies (6) which map the start site of the corresponding nt band (Figure 4A. Lanes 2 and 4). BstUI cleaves a minor but transcript to a position 110 nt upstream from the Alu element. significant fraction of the Alu sequences in untreated HeLa cells Effect of DNA methylation on transcript abundance (Figure 4A, Lane 3). Treatment of HeLa cells with 5-azacytidine substantially increases the susceptibility of the Alu BstUI site to Depending on the tissue type, as much as 30% of the cleavage, showing the demethylation of this site in some Alu 5-methylcytosine content of human DNA resides within Alu members (Figure 4A, Lane 1). Methylation of the BstUI site repeats (16, 18). HeLa cells were treated with 5-azacytidine to within the A box is sufficient to repress Pol El-directed reduce the level of DNA methylation (Materials and Methods). transcription of AJu templates (13). To test the methylation status The efficacy of this treatment was demonstrated by two different of other sites, the Alu repeats from 5-azacytidine treated and experiments: blot hybridization of genomic DNA digests and end Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 1092 Nucleic Acids Research, 1994, Vol. 22, No. 6 effect of methylation on the abundance of the Alu transcript A 12 3 4 (Figure 4B). 396 - Northern blot analysis confirms and extends the finding that 5-azacytidine treatment has a pronounced effect on the abundance of the Alu transcript (Figure 4C). In agreement with previous 214 - ** observations, two principal length transcripts are observed by blot hybridization. The shorter transcript in the poly A - fraction (Lanes 1 and 2) corresponds to 118nt scAlu RNA (5. 7). The longer transcripts in the poly A + fraction (Lanes 3 and 4), ca. 300 to 450 nt, are the sizes expected for Pol Ill-directed Alu transcripts (Introduction; 5, 6, 7). The high molecular weight smear of hybridization possibly results from a combination of Control 5'AzaC I II I longer Pol Ill-directed Alu transcripts and the presence of Alu 1 2 3 4 1 6 64 1 6 6.4 repeats in other transcription units. Comparing Lanes 3 and 4, treatment with 5-azacytidine increases the abundance of the presumptive Pol Ill-directed Alu transcripts by about eight-fold, in satisfactory agreement with the results of primer extension discussed above. Interestingly, 603 - the effect of 5-azacytidine treatment on the abundance of the 118 •350 nt scAlu transcript is much less pronounced (Figure 4C; Discussion). Reprobing the same Northern blot shows that the 310 - the steady state level of 5S rRNA is not affected by 5-azacytidine 281 - treatment (data not shown). This finding also provides an internal control for the amount of RNA loaded in Lanes 1 and 2 so that the small increase in the abundance of scAlu is significant. 240 Authenticity and identity of the Alu transcripts Coincidence of the 3' end of a primer extension product and the 118 - initiation site for Pol Hi-directed Alu transcription would provide the most direct evidence available that the primer extension products result from authentic Alu transcripts. Also, the A box of the RNA Pol III promoter element is located very near the 5' end of the Alu sequence. For these reasons, a cloning strategy Figure 4. Effects of 5-azacytidine treatment on HeLa DNA and Alu RNA. which maps and preserves the 3' end of the primer extension A: Blot hybridization of genomic DNA extracted from 5-azacytidine treated (Lanes product (i.e., the 5' end of the corresponding Alu transcript) was I and 2) and untreated cells (Lanes 3 and 4) was performed on DNAs digested with BstUI and Tthlll l (Lanes 1 and 3) and Taql and Tthllll (Lanes 2 and adopted (22). The primer extension product was first tailed with 4). Probe was the Alu specific oligonucleotide PV 51. Marker lengths in nt are oligo dA and then linearly amplified using an oligonucleotide that shown on the left-hand side. B: Pnmer extension of poly A + cytoplasmic RNA consists of an adapter primer linked to a short run of Ts, thereby extracted as indicated from 5-azacytidine treated cells (eight days) and untreated 'capping' the 5' end of the resulting cDNA sequence. By using control cells using the 17mer Alu specific oligonucleotide primer. Either 1.6 /ig a non-Alu homologous primer for PCR, this method decreases or 6.4 /ig of RNA was used as indicated. Similar results are obtained upon treating cells for twenty days with 5-azacytidine. C: Northern blot analysis of poly A — the likelihood of inadvertant amplification of Alu family members (Lanes 1 and 2) and poly A + (Lanes 3 and 4) cytoplasmic RNA isolated from interspersered in high molecular weight nucleic acids (7, 22; HeLa cells grown for eight days in the presence (Lanes 2 and 4) and absence Discussion). The resulting product was PCR amplified with the (Lanes 1 and 3) of 5-azacytidine. adapter primer and an additional Alu primer that is nested with respect to the oligonucleotide originally used for reverse transcription. The resulting clones were subsequently screened untreated cells are first isolated as a 189 bp restriction fragment by gel electrophoresis to discard cDNAs (approximately half) resulting from a Taql and Aspl (Tthl 1II isoschizomer) digest which were obviously truncated with respect to the full length of the DNAs (18). There are three consensus sites for HpaTI and Alu consensus sequence; about one-half of the remaining clones its isoschizomer within this Alu fragment. The resulting were sequenced. preparation is next subjected to digestion with either Hpall, which does not cleave its methylated site, or Mspl, which is indifferent With respect to their 5' ends as defined by the poly T tail, to methylation, followed by direct end labeling (18, 23). A 67 three groups of cDNAs were obtained by this procedure: four nt Hpall indicator band released from Alus in DNA from cDNAs are truncated before the 5' end of the Alu consensus 5-azacytidine treated cells is about twice as intense as the same sequence; eight begin precisely at the Pol III transcription start band from untreated cells (data not shown). In summary, site; and eleven sequences have one to seven additional 5' 5-azacytidine treatment results in hypomethylation of several nucleotides, usually A residues (Figure 5). The truncated cDNAs consensus Alu CpGs. are presumably products of incomplete reverse transcription and could result from either authentic Pol III Alu transcripts or Alus As assayed by the 240 nt primer extension product, the Alu that were transcribed within longer transcription units from other transcript is more abundant (ca. five-fold) in 5-azacytidine treated promoters. The tendency of reverse transcriptase to terminate cells than in the control cells. The level of the 350 nt product prematurely within the G-rich 5' end of Alu sequences is as well as other bands of unknown identity are not altered by illustrated by truncated primer extension products of 7SL RNA 5-azacytidine treatment, providing an internal control for the Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 Nucleic Acids Research, 1994, Vol. 22, No. 6 1093 GGCC*GQqCQCQGTQqCT — GAQATCQAQACCAT — (Alu consensus) (Promoter consensus) RRYNNRRYGG GATCRANNCC Sub.Fa» . # Hit . fCpQ CLONES (T) PV 1 19 4 *..A — Precise 28 8 6 * A — Precise 9 15 12 *...T — Precise 13 12 A5 A...* A — Major 23 7 5 A A — ...T. J G Major 20 8 8 ....C — ... T G Major 15 7 9 ....*....C — ... T G Major 15 10 A4 *...T — ... T G 17 CT * — PV 10 15 7 AA * — Precise 10 17 A7 CTGTT...T* AT — I... Precise 17 8 Precise 11 13 A6 ACAGTT...** — . . .fl-A 1 A *.A£...A — ... T G Major 24 5 A3 A * A A... — ...T.. A G Major 29 5 A2 CT. . .T . . .T.T A — ...T.. A G Major 28 4 1 5 GAA *A CA..C.A. — ...T.. A A Major 31 7 Major 19 5 1 0 AAAAA *...T...S — ...T. J A--G 18 6 1 6 AAAAA * — ...T.. A A Major 24 8 Major 3 GTTCCCT *...£...! — ...T G .T. I G Major 350-2 ...T*A. . .AT . . 27 4 350-3 GT..*C. ..A.A... .T . .C G Major 33 7 ******* Al . . CG 9 Precise 21 *****, 13 .T.T .T 6 T G Major 26 ******* 2 A. . Major 25 11 T G **************. 11 Major 15 8 Figure 5. Sequence analysis of Alu cDNA clones Comparison of Alu cDNA sequences to the Alu consensus sequence for positions 1 — 17 and positions 73 — 86 (2 and references therein) and the A and B boxes of the Pol III promoter consensus (33). Clones 1 - 17 and clones Al -A 7 (5-azacytidine treated cells) are derived from the 240 nt pnmer extension product. Clones 350-2 and 350-3 are from the 350 nt band. The subfamily identity, the number of mismatches with respect to the subfamily consensus sequence, and the number of CpG residues are listed for each clone. - indicates uncompared sequence at a position. • indicates base identity to the consensus. * indicates a base deletion. Base substitutions are shown by the appropriate letter. Bold letters indicate a substitution with respect to the Pol in promoter consensus. These sequences have been assigned GenBank accession numbers U02044-U02068. reported in Figures 2 and 3. Consequently, incomplete Alu cDNA CT or from an authentic Alu transciprt which artifactually sequences are not informative about the source of their parent acquired a single C residue at position - 2 (Figure 5). The transcripts and will not be examined further. We interpret the assignment of Clone A2 is similarly ambiguous. Both are formally full length cDNAs as primer extension products resulting from included in the class as each has flanking non-T residues. Pol III Alu transcripts (Figure 5). For the same reason, we In addition to cDNAs corresponding to the 240 nt primer conclude that the 240 nt primer extension products examined in extension product, two Alu-containing cDNAs (Clones 350-2 and the previous experiments (Figures 1 through 4) are a faithful 350-3) were derived from the 350 nt primer extension product. indication of authentic Alu transcripts. Regarding the eleven These two cDNAs certainly result from longer primary transcripts cDNAs which have additional residues on their 5' ends, many and are therefore pooled with the eleven Alu cDNAs, discussed of these cannot be attributed to Pol in transcription from the above, giving a total of thirteen cDNAs in this class (Figure 5). promoter of the interspersered Alu repeat (Discussion). Most of In agreement with previous reports, each full length Alu cDNA these eleven cDNAs probably result from longer transcripts which corresponds to a distinct Alu repeat, showing that many different happen to have contained an interspersed Alu element Alu loci must be transcribed (Figure 5) (6, 7). Included among (Discussion). For simplicity, we refer to these as 'interspersed' the full length cDNA sequences are four representatives of the Alu transcripts. However, the assignment of Clone 17, and Major subfamily, three representatives of the Precise subfamily perhaps also of Clone A2, as transcripts may result from a cloning and one representative of the PV subfamily (1, 2). The PV artifact which deserves special comment. Clone 17 has a single subfamily Alu member was obtained by screening cDNA clones C residue in what would otherwise be a continuous perfect poly using oligonucleotide hybridization and is not statistically T sequence generated by the tailing strategy described above. represented in this small sampling of cDNA sequences. However, In another cDNA, a single C residue was observed in the middle relative to their genomic abundance, the three Precise subfamily of its, long poly T run and was undoubtedly artifactually members are overrepresented compared to the four Major introduced during PCR amplification (unpublished). Clone 17 subfamily members (Figure 5). This finding agrees with previous either results from a Alu transcript flanked by the dinucleotide observations that representatives of younger Alu subfamilies are Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 1094 Nucleic Acids Research, 1994, Vol. 22, No. 6 transcriptionally more active than older subfamily members (5, RNAs detected here by many orders of magnitude and 6,7) . contamination resulting from these transcripts is probably There is a difference in the conservation of Pol III promoter unavoidable. Sinnett et al use an oligonucleotide primer that elements (32) between cDNAs corresponding to authentic Alu corresponds to the 5' end of the Alu consensus sequence for transcripts and cDNAs corresponding to Alu transcripts. All eight second strand cDNA synthesis (6). As they discuss, that cloning full-length Alus have perfect consensus A boxes and only one strategy does not preserve the 5' end of the parent RNA and cDNA (Clone 5) has a nonconsensus B box substitution. Four therefore cannot unambiguously discriminate betwen Alu RNAs of the thirteen Alu cDNAs have nonconsensus consensus A boxes which result from Pol II- and Pol Ill-directed transcription. In this study, less than half of the observed cDNAs result from and four have nonconsensus B boxes. authentic Alu transcripts as judged by their 5' ends; the majority Individual cDNAs differ from their respective consensus result from Alus interspersed within other transcription units. sequences at various base positions (Figure 5). The authentic Alu transcript cDNAs have an average of 15.5 (7.0%) differences Determination and regulation of Alu template activities from their consensus sequences; for the Alu cDNAs the average is 20.8 (9.5%) differences. The slight difference between these The steady state abundance of Alu RNA observed here is very two averages mostly reflects the different subfamily composition low. The expression of Alu repeats may be determined by several of the two classes of cDNAs; the average divergence is 8% for factors, including their DNA methylation, structural variation of the Precise subfamily and 13.5 % for the Major subfamily (25). their internal Pol III promoters, sequence context, and lifetime of the resulting transcripts. Remarkably, the sequence of Clone 4 differs by only a single nucleotide from the PV Alu consensus. The average CpG content Methylation represses transcription of several different Pol Hi- of the two classes is also not very different. The authentic Alu directed templates in vivo, and in particular represses the transcript cDNAs have an average of 11 CpGs; the Alu cDNAs transcription of Alu repeats in vitro (13, 17). Stimulation of Alu have an average of 9.4 CpGs (Figure 5). In summary, the single expression in HeLa cells by 5-azacytidine treatment and its distinguishing feature for these two classes of cDNAs is the concomitant effect on DNA methylation provide evidence that fidelity of their A box elements. This element is more faithfully the transcription of Alu repeats, like that of other Pol Ill-directed represented in younger Alu members, which in rum are somewhat templates, is repressed by DNA methylation. Alu repeats are almost completely methylated in certain somatic tissues and enriched in the authentic Alu transcript cDNAs. The A box is interestingly, certain Alu subgroups are almost quantitatively especially important in determining the strength of a Pol III unmethylated in the male germ line (16, 17, 18). Thus it is likely promoter (13 and references therein). that the low level of Alu expression is attributable, at in least The Alu element in Clone 350-2 is preceded by an 85 nt part, to transcriptional repression caused by extensive sequence and similarly the Alu element in Clone 350-3 is methylation. A small number of Alu repeats in HeLa cells are preceded by a different 90 nt sequence. Neither of these two hypomethylated (18) and plausibly this subgroup includes more flanking sequences is present in sequence data banks and, in active Alu templates. Methylation of sequences is often particular, neither is a known repetitive sequence. Primer determined by their sequence context (27), and more active extension of poly A-I- cytoplasmic RNA using an oligonucleotide hypomethylated Alu templates might reside in a favorable derived from Clone 350-2 did not yield product, but primer sequence context. Chromatin structure and DNA methylation extension using an oligonucleotide from Clone 350-3 yielded synergistically repress Alu transcription in vitro (28). several discrete-length primer extension products, including one that is consistent with the structure of the 350 nt primer extension Sequences flanking certain Pol Ill-directed templates, such as product described above: an Alu positioned 110 nt downstream the 7SL RNA gene, significantly stimulate their transcriptional from a transcription start site. We therefore tentatively assign activity (29). In the specific case of an Alu source gene (30), cDNA Clone 350-3 to the mRNA associated with the 350 nt cis acting elements stimulate its transcription in vitro by about primer extension product. ten-fold (Chesnokov and Schmid, unpublished). Alu transcriptional activity is probably subject to both positive and negative regulation. DISCUSSION More transpositionally active source genes closely match their Technical solution to a long-standing problem consensus sequences, raising the question of why these sequences Authentic but relatively rare Alu transcripts have been difficult have a transpositional advantage over those which have suffered to identify unambiguously against the very high background of a higher number of base substitutions (1,2). The relative template activity of Alus in vitro is in part determined by the fidelity of Alu sequences interspersed within hnRNA and even some mature mRNAs (Introduction). Many Alu cDNA sequences observed the internal Pol III promoter sequences. The A box is especially here begin precisely at the 5' start site employed by Pol III important in determining the strength of a Pol III promoter (13). transcription in vitro and very likely represent authentic Alu The A box sequence in the Pol Ul-directed Alu transcripts exactly transcripts. Similar evidence shows that scAlu RNAs result from matches the consensus sequence for this element, which is less faithfully preserved in the interspersed transcripts. Pol IE-directed Pol III transcription (7). Alu cDNAs with extended 5' ends most likely result from transcripts that contain an interspersed Alu Alu transcripts reported here would likely have a transcriptional element although in certain cases, reverse transcriptase can add advantage relative to Alus having nonconsensus base substitutions one or two untemplated nucleotides to a primer extension product in their Pol III promoter elements. (26). The additional bases are mostly As. The direct repeats This finding may partially explain the selective advantage that flanking Alu elements are usually A-rich, suggesting the some potential Alu source genes enjoy over others. Younger Alu probability that these additional nucleotides were templated by repeats closely match their respective subfamily consensus the parent RNA. Polymerase II-directed Alu transcripts comprise sequences, including the Pol III promoter consensus sequence. 10% of hnRNA, exceeding the small number of cytoplasmic Conservation of the promoter elements may therefore provide Downloaded from https://academic.oup.com/nar/article-abstract/22/6/1087/1032924 by DeepDyve user on 05 August 2020 Nucleic Acids Research, 1994, Vol. 22, No. 6 1095 6 Sinnett. D.. Richer. C . Deragon. J M. and Laboda, D. (1992) J. Mol. Bid.. younger Alu sequences with a relative transcriptional advantage 226. 689-706. and a consequent retrotranspositional advantage (2, 13). Other 7. Maraia. R.J.. Driscoll. C . Bilyeu. T.. Hsu. K.. and Darlington, G.J. (1993) factors, such as the structure and lifetime of the RNA, in addition Mol. Cell Biol.. 13, 4233-4241 to methylation and positive cis acting elements discussed above, 8. Weiner, A.M.. Deimnger. P.L. and Efstratiudis. A. (1986) Ann. Rev. would all select among the many potential source genes (6). Biochcm., 55, 631-661. 9. Schmid. C.W., Deka. N. and Matera. A.G. (1990) In Adolpf, K.W.(ed) Chromosomes: Eukaryotic, Prokaryotic and Viral. CRC Press. Boca Raton. Possible functions of Alu transcripts Honda, pp 323-358. Considering only authentic (i.e., Pol Ill-directed) transcripts, two 10. Chang. D.-Y and Maraia. R.J. (1993) J. Biol. Chem.. 268, 6423-6428. different forms of Alu RNA are present in the cytoplasm: left 11 Elder. J. T.. Pan. J., Duncan. C.H. and Weissman. S. M. (1981) Nucleic Acids Res.. 9. 1171-1189 monomer scAlu RNA and full length Alu RNA (5, 6, 7). ScAlu 12 Fuhrman. S.A.. Deininger. P.L.. LaPorte, P . Friedmann. T. and Geiduschek. RNA is thought to result from the processing of full length Alu E.P (1981) Nucleic Acids Res.. 9, 6439-6456. transcripts, presumably in the nucleus. The cytoplasmic 13. Liu. W. M. and Schmid. C W. (1993) Nucleic Adds Res.. 21, 1351 - 1359. localization of the full length Alu transcripts and their association 14. Jang K. L. and Latchman. D.S. (1989) FEBS Len.. 258, 255-258. with RNPs suggest that the full length and scAlu RNAs may have 15. Panning. B. and Smiley. J.R. (1993) Molec. and Cell Biol., 13. 3231-3244. distinct cytoplasmic functions or activities. Also, 5-azacytidine 16. Schmid, C.W. (1991) Nucleic Acids Res.. 19, 5613-5617. 17. Kochanek. S.. Renz. D and Doerfler. W. (1993) EMBO Journal 12. treatment increases relative abundances of the full-length and left 1141-1151. monomer transcripts by different degrees. Among other 18. Hellmann-Blumberg. U.. Hintz. M.F. and Schmid. C.W. (1993) Mol. Cell possibilities, this difference might result from limiting factors Biol.. 13. 4523-4530. necessary for the accumulation of scAlu RNA. 19 Snyder, R and Lachmann. P. (1989) Mutation Res.. 226. 185-190. 20. Foster. R.. Jahroudi. N. and Gedamu. L. (1991) Biochim. Biophvs. Ada. The presence of 100 to 1000 cytoplasmic Alu transcripts per 1088. 373-379 cell, possibly organized as RNPs, suggests that the transcripts 21. Maniatis. T.. Fntsch. E F. and Sambrook. J. (1989) Molecular Cloning: have a function. The most obvious possibility is a role in some A Laboratory Manual Cold Spring Harbor University Press. Cold Spring aspect of translation. The Alu homolog, 7SL RNA, is an integral Harbor. 22. Maraia, R. (1991) Nucleic Acid Res., 19. 5695-5702. component of SRP and one function of SRP is translation arrest 23. Monk. M.. Boubelik. M. and Lchnert. S. (1987) Development. 99. 71-382. (31). SRP proteins bind the Alu-like region of 7SL RNA and 24. Baserga, S. J. and Steitz. J.A (1993) The RNA World. Cold Spring Harbor the 68 kD subunit of SRP forms complexes with Alu transcripts Press. Cold Spring Harbor, pp. 359-381. in vitro (31,32). Interestingly the major families of short 25. Willard. C . Nguyen. H. T . and Schmid. C.W. (1987) J. Mol. Evol., 26, interspersed repeats in mammals can be classified into two 180-186. 26 Swanstrom. R., Varmus. H.E.. and Bishop. J.M. (1981) J. Biol. Chem., sequence superfamilies. Primate Alu repeats and rodent Bl 256. 1115-1121. sequences are homologous to 7SL RNA (1, 2). Rodent B2 27 Bird. A. P. (1992) Cell. 70. 5-8 . sequences, rabbit C repeats, and various other repeat sequence 28 Englander, E.W..Wolffe. A.P. and Howard. B.H. (1993) J. Biol. Chem., families in various mammalian species are homologous to tRNA 268. 19565-19573 sequences. Both superfamilies are ancestrally related to sequences 29. Bredow. S.. Sung, D.. Muller, J.. Kleinert. H. and Benecke. B.-J. (1990) Nucleic Acids Res.. 18, 6779-6784 which have a translational role. Thus there are several reasons 30 Leeflang, E.P.. Liu. W-M . Chesnokov, I. and Schmid. C.W. (1993) J. to guess that an Alu RNP might have a translational role. The Mol. Ewl.. .37. 559-565 sequence similarity between a short region of 28S rRNA and the 31 Strub. K.. Moss, J.. and Walter. P (1991) Mol. Cell. Biol., 11. 3949-3959. human Alu sequence (Results) might merely be an accidental 32. Andrews. P G. and Kole. R. (1987) J. Biol Chem.. 262, 2908-2912. match or it might result from such a functional relationship. Alternatively, the presence of the full-length transcript in HeLa cells might result from the leaky repression of Alu templates, and while these transcripts may not have a necessary function in HeLa cells they might have such a function in other tissues and cell types in which Alu repeats are appropriately expressed. ACKNOWLEDGEMENTS This work was supported in part by USPHS grant GM 21346 (C.S.), the Agricultural Experiment Station at the University of California, Davis. (C.S.), and a Jastro Shields Scholarship Award (W.M.L.). We thank Ms Sandra Malakauskas for her invaluble technical assistance. REFERENCES 1 Deininger. P L.. Batzer. M.A.. Hutchison III. C.A and Edgell. M H (1992) Trends in Genet . 8. 307-311. 2 Schmid. C W and Maraia. R (1992) Current Opinion in Genetics ami Development• Genomes and Emlution, 2. 874 — 882. 3 Liu. W-M . Leeflang. E.P. and Schmid. C W (1992) Biochim Biophvs Actu. 1132. 306-308. 4. Paulson. E. K. and Schmid. C.W. (1986)NucleicAddsRcs., 14,6145-6158 5 Matera. A.G.. Hellmann-Blumberg, U and Schmid. C W. (1990) Mol. Cell. Biol , 10. 5424-5432.

Journal

Nucleic Acids Research – Oxford University Press

Published: Mar 25, 1994

There are no references for this article.