Metformin alters DNA methylation genome-wide via the H19/SAHH axis.

The molecular mechanisms underlying the antineoplastic properties of metformin, a first-line drug for type 2 diabetes, remain elusive. Here we report that metformin induces genome-wide alterations in DNA methylation by modulating the activity of S-adenosylhomocysteine hydrolase (SAHH). Exposing cancer cells to metformin leads to hypermethylation of tumor-promoting pathway genes and concomitant inhibition of cell proliferation. Metformin acts by upregulating microRNA let-7 through AMPK activation, leading to degradation of H19 long noncoding RNA, which normally binds to and inactivates SAHH. H19 knockdown activates SAHH, enabling DNA methyltransferase 3B to methylate a subset of genes. This metformin-induced H19 repression and alteration of gene methylation are recapitulated in endometrial cancer tissue samples obtained from patients treated with antidiabetic doses of metformin. Our findings unveil a novel mechanism of action for the drug metformin with implications for the molecular basis of epigenetic dysregulation in cancer. This novel mechanism of action also may be occurring in normal cells.

How a small DNA virus uses dsRNA but not RNAi to regulate its life cycle.

Mouse polyomavirus contains a circular DNA genome, with early and late genes transcribed from opposite strands. At early times after infection, genes encoded from the early transcription unit are predominantly expressed. After the onset of viral DNA replication, expression of genes encoded from the late transcription unit increases dramatically. At late times, late primary transcripts are inefficiently polyadenylated, leading to the generation of multigenomic RNAs that are precursors to mature mRNAs. These transcripts contain sequences complementary to the early RNAs and downregulate early-strand gene expression by inducing RNA editing. Our recent work leads to a model where the production of the multigenomic late RNAs is also controlled by the editing of poly(A) signals, directed by overlapping primary transcripts.

The fate of dsRNA in the nucleus: a p54(nrb)-containing complex mediates the nuclear retention of promiscuously A-to-I edited RNAs.

How do cells discriminate between selectively edited mRNAs that encode new protein isoforms, and dsRNA-induced, promiscuously edited RNAs that encode nonfunctional, mutant proteins? We have developed a Xenopus oocyte model system which shows that a variety of hyperedited, inosine-containing RNAs are specifically retained in the nucleus. To uncover the mechanism of inosine-induced retention, HeLa cell nuclear extracts were used to isolate a multiprotein complex that binds specifically and cooperatively to inosine-containing RNAs. This complex contains the inosine-specific RNA binding protein p54(nrb), the splicing factor PSF, and the inner nuclear matrix structural protein matrin 3. We provide evidence that one function of the complex identified here is to anchor hyperedited RNAs to the nuclear matrix, while allowing selectively edited mRNAs to be exported.

Intronless mRNA transport elements may affect multiple steps of pre-mRNA processing.

We have reported recently that a small element within the mouse histone H2a-coding region permits efficient cytoplasmic accumulation of intronless beta-globin cDNA transcripts. This sequence lowers the levels of spliced products from intron-containing constructs and can functionally replace Rev and the Rev-responsive element (RRE) in the nuclear export of unspliced HIV-1-related mRNAs. In work reported here, we further investigate the molecular mechanisms by which this element might work. We demonstrate here through both in vivo and in vitro assays that, in addition to promoting mRNA nuclear export, this element acts as a polyadenylation enhancer and as a potent inhibitor of splicing. Surprisingly, two other described intronless mRNA transport elements (from the herpes simplex virus thymidine kinase gene and hepatitis B virus) appear to function in a similar manner. These findings prompt us to suggest that a general feature of intronless mRNA transport elements might be a collection of phenotypes, including the inhibition of splicing and the enhancement of both polyadenylation and mRNA export.

Antisense RNA: function and fate of duplex RNA in cells of higher eukaryotes.

There is ample evidence that cells of higher eukaryotes express double-stranded RNA molecules (dsRNAs) either naturally or as the result of viral infection or aberrant, bidirectional transcriptional readthrough. These duplex molecules can exist in either the cytoplasmic or nuclear compartments. Cells have evolved distinct ways of responding to dsRNAs, depending on the nature and location of the duplexes. Since dsRNA molecules are not thought to exist naturally within the cytoplasm, dsRNA in this compartment is most often associated with viral infections. Cells have evolved defensive strategies against such molecules, primarily involving the interferon response pathway. Nuclear dsRNA, however, does not induce interferons and may play an important posttranscriptional regulatory role. Nuclear dsRNA appears to be the substrate for enzymes which deaminate adenosine residues to inosine residues within the polynucleotide structure, resulting in partial or full unwinding. Extensively modified RNAs are either rapidly degraded or retained within the nucleus, whereas transcripts with few modifications may be transported to the cytoplasm, where they serve to produce altered proteins. This review summarizes our current knowledge about the function and fate of dsRNA in cells of higher eukaryotes and its potential manipulation as a research and therapeutic tool.

The mouse histone H2a gene contains a small element that facilitates cytoplasmic accumulation of intronless gene transcripts and of unspliced HIV-1-related mRNAs.

Histone mRNAs are naturally intronless and accumulate efficiently in the cytoplasm. To learn whether there are cis-acting sequences within histone genes that allow efficient cytoplasmic accumulation of RNAs, we made recombinant constructs in which sequences from the mouse H2a gene were cloned into a human beta-globin cDNA. By using transient transfection and RNase protection analysis, we demonstrate here that a 100-bp sequence within the H2a coding region permits efficient cytoplasmic accumulation of the globin cDNA transcripts. We also show that this sequence appears to suppress splicing and can functionally replace Rev and the Rev-responsive element in the cytoplasmic accumulation of unspliced HIV-1-related mRNAs. Like the Rev-responsive element, this sequence acts in an orientation-dependent manner. We thus propose that the sequence identified here may be a member of the cis-acting elements that facilitate the cytoplasmic accumulation of naturally intronless gene transcripts.

Nuclear antisense RNA induces extensive adenosine modifications and nuclear retention of target transcripts.

Antisense RNA may regulate the expression of a number of eukaryotic genes, but little is known about its prevalence or mechanism of action. We have used a model system in which antisense control can be studied both genetically and biochemically. Late in polyoma virus infection, early-strand mRNA levels are down-regulated by nuclear antisense RNA from the late strand. Analysis of early-strand transcripts isolated late in infection revealed extensive base modifications. In many transcripts almost half of the adenosines were altered to inosines or guanosines. These results suggest modification of RNA duplexes by double-stranded RNA adenosine deaminase or a related enzyme. Probes that detect only modified RNAs revealed that these molecules are not highly unstable, but accumulate within the nucleus and are thus inert for gene expression. Antisense-induced modifications can account for most or all of the observed regulation, with the lowered levels of early-strand RNAs commonly observed late in infection resulting from the fact that many transcripts are invisible to standard hybridization probes. This work suggests that similar antisense-mediated control mechanisms may also operate under physiological conditions in uninfected eukaryotic cells, and leads to the proposal that there is a novel pool of nuclear RNAs that cannot be seen with many molecular probes heretofore used.

Transcription from plasmid expression vectors is increased up to 14-fold when plasmids are transfected as concatemers.

A protocol for increasing transcription from plasmid expression vectors is presented. A vector containing chloramphenicol acetyltransferase (CAT) gene was digested leaving the transcription cassette intact. Heat inactivation of restriction enzymes followed by ligation of the digestion products yielded concatemers which migrated as a single band in agarose gel electrophoresis. Mouse fibroblasts transfected with the concatemers gave a CAT activity that was 14-fold greater than that of cells transfected with a similar mass (equimolar gene number) of the native plasmid. The effect was independent of promoter type, restriction enzyme, number of restriction sites and with a noted exception, cell line.

A suboptimal 5' splice site is a cis-acting determinant of nuclear export of polyomavirus late mRNAs.

Mouse polyomavirus has been used as a model system to study nucleocytoplasmic transport of mRNA. Three late mRNAs encoding the viral capsid proteins are generated by alternative splicing from common pre-mRNA molecules. mRNAs encoding the virion protein VP2 (mVP2) harbor an unused 5' splice site, and more than half of them remain fully unspliced yet are able to enter the cytoplasm for translation. Examination of the intracellular distribution of late viral mRNAs revealed, however, that mVP2 molecules are exported less efficiently than are mVP1 and mVP3, in which the 5' splice site has been removed by splicing. Point mutations and deletion analyses demonstrated that the efficiency of mVP2 export is inversely correlated with the strength of the 5' splice site and that unused 3' splice sites present in the mRNA have little or no effect on export. These results suggest that the unused 5' splice site is a key player in mVP2 export. Interestingly, mRNAs carrying large deletions but retaining the 5' splice site exhibited a wild-type mVP2 export phenotype, suggesting that there are no other constitutive cis-acting sequences involved in mVP2 export. RNA stability measurements confirmed that the subcellular distribution differences between these mRNAs were not due to differential half-lives between the two cellular compartments. We therefore conclude that the nuclear export of mVP2 is strongly influenced by a suboptimal 5' splice site. Furthermore, results comparing spliced and unspliced forms of mVP2 molecules indicated that the process of splicing does not enhance nuclear export. Since mVP2 and some of its mutant forms can accumulate in the cytoplasm in the absence of splicing, we propose that splicing is not a prerequisite for mRNA export in the polyomavirus system; rather, removal of splicing machinery from mRNAs may be required. The possibility that export of other viral mRNAs can be influenced by suboptimal splicing signals is also discussed.

Novel biotinylated plasmid expression vectors retain biological function and can bind streptavidin.

A new method for coupling proteins to plasmid expression vectors is presented. Biotin was covalently attached to a plasmid expression vector containing a chloramphenicol acetyltransferase (CAT) gene. The specific label was one biotin per 100 bp. An electrophoretic mobility shift assay showed that the plasmid was capable of binding multiple streptavidin molecules. When transfected into mouse fibroblasts, the biotinylated plasmid retained 40% of the native plasmid's biological activity, as determined by CAT assay, and was not affected by the binding of streptavidin. The method allows for attachment of any protein to plasmid DNA expression vector while retaining biological function. Hybrid plasmids in which the transcription cassettes were kept free of biotin label were constructed by digesting biotinylated and unbiotinylated plasmids at sites outside the transcription cassette and re-ligating the digestion products. Electron microscopy studies show that the ligation products formed large tangled assemblages of plasmid DNA. When equimolar (with respect to gene number) amounts of these large hybrid biotinylated plasmids were transfected into mouse fibroblasts by means of calcium phosphate precipitation, an increase in CAT expression 25-fold greater than that of original biotinylated plasmid was observed. Slot-blot analysis of total DNA extracted from transfected cells shows that this enhanced activity was not due to increased transfection efficiency. Receptor-mediated delivery could not be shown when a complex comprising biotinylated asialoglycoprotein/streptavidin/biotinylated CAT expression vector was placed in media containing Hep G2 cells.

Role of polyadenylation in nucleocytoplasmic transport of mRNA.

To examine the role of polyadenylation in the nuclear export of mRNA, we have replaced the poly(A) signal in a Rev-responsive human immunodeficiency virus type 1-based reporter gene with a cis-acting hammerhead ribozyme. Transcripts from this gene thus acquire a 3' terminus by cis-ribozyme cleavage rather than by polyadenylation. The nuclear and cytoplasmic distribution of transcripts was investigated using transient gene expression and quantitative RNase protection assays. In the absence of Rev, a basal level of polyadenylated unspliced mRNA transcribed from a poly(A) signal-containing control reporter gene was detected in the cytoplasm of transfected COS7 cells. However, cytoplasmic ribozyme-cleaved unspliced RNA was only barely detectable. The nuclear/cytoplasmic (n/c) ratio of polyadenylated RNAs was 3.8, while the n/c ratio for ribozyme cis-cleaved RNAs was 33. The cytoplasmic localization of the polyadenylated unspliced mRNA was enhanced about 10-fold in the presence of Rev and the Rev-responsive element. In marked contrast to this, ribozyme cleaved RNA accumulated almost exclusively (n/c ratio of 28) in the nucleus in the presence of Rev. Actinomycin D time course analysis suggested that the low levels of the cytoplasmic ribozyme-cleaved RNAs in both the presence and absence of Rev were due to serve export deficiency of ribozyme-cleaved RNA. Finally, by inserting a 90-nucleotide poly(A) stretch directly upstream of the ribozyme cassette, we have demonstrated that a long stretch of poly(A) near the 3' end of a ribozyme-cleaved transcript is not sufficient for directing mRNA export. Taken together, these results suggest that polyadenylation is required for the nucleocytoplasmic transport of mRNA and that Rev interaction with the Rev-responsive element cannot bypass this requirement.

The sequence and context of the 5' splice site govern the nuclear stability of polyoma virus late RNAs.

We have examined the influence of splicing signals on the stability of polyoma virus late RNAs in the nucleus. Late primary transcripts contain a single 5' splice site and three alternative 3' splice sites. In earlier work we showed that the presence of introns was not required for late RNA accumulation, however, the 5' splice site was essential, as removal of only the 5' splice site was sufficient to destabilize late RNAs up to 100-fold when compared with early RNAs. A complementary clone which retained the 5' splice site but which carried small deletions of all late region 3' splice sites produced wild-type levels of unspliced late RNA. In order to extend this work we have constructed additional types of mutants. Point mutations in the 5' splice site confirmed its importance for RNA stability. Other mutants included constructs in which the spacing between the 5' splice site and the late promoter was altered and 5' splice site insertion mutants where a 58 bp fragment containing the 5' splice site sequence was inserted separately at various restriction sites in the late region. Both types of mutants lacked all of the late 3' splice sites and had only a single 5' splice site. RNase protection analyses of late and early RNAs from these constructs revealed that moving the 5' splice site away from the late promoter (or from its normal context) destabilized late RNAs > 10-fold relative to the wild-type. We conclude that both 5' splice site integrity and its proximity to the late promoter play important roles in the nuclear stability of polyoma virus late RNAs.

Characterization of the polyomavirus late polyadenylation signal.

The polyomavirus late polyadenylation signal is used inefficiently during the late phase of a productive viral infection. Inefficient polyadenylation serves an important purpose for viral propagation, as it allows a splicing event that stabilizes late transcripts (G. R. Adami, C. W. Marlor, N. L. Barrett, and G. G. Carmichale, J. Virol. 63:85-93, 1989; R. P. Hyde-DeRuyscher and G. G. Carmichael, J. Virol. 64:5823-5832, 1990). We have recently shown that late-strand readthrough transcripts serve as natural antisense molecules to downregulate early-strand RNA levels at late times in infection (Z. Liu, D. B. Batt, and G. G. Carmichael, Proc. Natl. Acad. Sci. USA 91:4258-4262, 1994). Thus, poor polyadenylation contributes to the early-late switch by allowing the formation of more stable late RNAs and by forming antisense RNA to early RNAs. The importance of late poly(A) site inefficiency in the viral life cycle has prompted us to map the cis elements of this site. Since the polyomavirus late site proved a poor substrate for in vitro polyadenylation, we used an in vivo assay which allowed us to map the cis sequences required for its function. In this assay, various fragments containing the AAUAAA and different surrounding sequences were placed 1.4 kb upstream of a second, wild-type signal. The second signal served to stabilize transcripts that are not processed at the upstream site, allowing accurate quantitation of relative poly(A) site use by an RNase protection assay. Processing was primary at the upstream site when a large fragment surrounding the poly(A) signal (50 nucleotides [nt] upstream and 90 nt downstream) was tested in this assay, demonstrating that this fragment contains the essential cis elements. Deletion analysis of this fragment revealed that most but not all upstream sequences can be removed with little effect on polyadenylation efficiency, indicating the absence of a strong stimulatory upstream element. Deletion of all but 25 nt downstream of the AAUAAA reduced polyadenylation activity only by half, demonstrating that processing can occur at this site despite the lack of downstream sequences. Thus, the core cis element for polyadenylation is quite small, with most important cis-acting elements lying within 19 nt upstream and 25 nt downstream of the AAUAAA sequence. This core contains the AAUAAA hexanucleotide, an upstream A/U-rich element, and three identical repeats of a 6-nt sequence, UAUUCA. Polyadenylation was eliminated or greatly reduced when either the AAUAAA or the three repeats were mutated.

Nuclear antisense RNA. An efficient new method to inhibit gene expression.

We describe an efficient new antisense RNA method to inhibit gene expression. Antisense RNAs that are retained in the nucleus bind to target transcripts and appear to lead to the degradation of their targets. Antisense RNAs can be expressed and accumulated specifically in the nucleus if they are not polyadenylated at their 3' ends. In antisense expression vectors we use a cis-acting ribozyme to generate 3'-ends independently of the polyadenylation machinery and thereby inhibit transport of RNA molecules from the nucleus to the cytoplasm. We have evaluated this method in the mouse polyoma virus model system, where nuclear antisense transcripts to the viral early transcription region efficiently reduced the level of viral early-strand RNAs. Nonspecific antisense RNA had no effects on viral gene expression. In comparative studies, nuclear antisense RNAs were significantly more effective in downregulating polyoma virus early RNAs than were conventional antisense molecules, which were processed by polyadenylation.

Polyadenylation and transcription termination in gene constructs containing multiple tandem polyadenylation signals.

The processes of pre-mRNA 3'-end cleavage and polyadenylation have been closely linked to transcription termination by RNA polymerase II. We have studied the relationship between polyadenylation and transcription termination in gene constructs containing tandem poly(A) signals, at least one of which is the inefficient polyomavirus late poly(A) site. When identical tandem viral signals were separated by fewer than 400 bp, they competed for polyadenylation. The upstream site was always chosen preferentially, but relative site choice was influenced by the distance between the signals. All of these constructs showed the same low level of transcription termination as wild type polyomavirus, which contains a single late poly(A) site. When tandem poly(A) signals were not identical, a stronger downstream signal could outcompete a weaker upstream signal for polyadenylation without altering the efficiency of transcription termination characteristic for use of the upstream signal. Thus, if a weak polyoma virus late poly(A) signal (associated with inefficient transcription termination) preceded a strong rabbit beta-globin signal (associated with efficient transcription termination), termination remained inefficient, but the distal signal was most often chosen for polyadenylation. These results are consistent with independent regulation of polyadenylation and transcription termination in this system and are discussed in light of current models for the dependence of transcription termination on a functional poly(A) site.

Targeted nuclear antisense RNA mimics natural antisense-induced degradation of polyoma virus early RNA.

We describe a general antisense strategy to inhibit target gene expression. The substitution of a cis-acting ribozyme for a polyadenylylation signal in an antisense expression vector results in the nuclear retention of RNAs and the efficient degradation of their targets. We demonstrate the utility of this system in polyoma virus, where early-strand RNA levels are downregulated in the nucleus by antisense late-strand counterparts. We show that mutations destabilizing these naturally occurring antisense transcripts lead to increased levels of early-strand RNAs. Furthermore, expression in trans of nuclear antisense transcripts lowers early-strand RNA levels and quantitatively mimics the natural regulation.