A highly conserved RNA-binding protein for cytoplasmic mRNA localization in vertebrates.

BACKGROUND: Cytoplasmic mRNA localization is a widespread mechanism for restricting the translation of specific mRNAs to distinct regions of eucaryotic cells. This process involves specific interactions between cellular factors and localization signals in the 3' untranslated regions of the localized mRNA. Because only a few of these cellular factors have been identified, it is not known whether common factors are utilized for the localization of different mRNAs. We recently discovered Vera, a protein that binds specifically to the Vg1 localization element and is involved in the localization of Vg1 mRNA in Xenopus oocytes. RESULTS: To characterize further the role of Vera in the localization of Vg1 mRNA, we have purified the Vera protein and cloned its gene. Vera is homologous to chicken zip-code-binding protein (ZBP), which binds to a short RNA sequence required for localization of beta-actin mRNA in chick embryo fibroblasts. The predicted amino-acid sequences of Vera and ZBP contain five RNA-binding domains and putative signals for nuclear localization and export. Like the binding of ZBP to beta-actin mRNA, Vera specifically binds to a repeated sequence motif in the Vg1 localization element that is required for Vg1 mRNA localization in Xenopus oocytes. CONCLUSIONS: Vera, a highly conserved component of the mRNA localization machinery, participates in localizing different mRNAs in different cell types. Thus, Vera appears to be a general factor for mRNA localization, and additional factors may be required to specify diverse patterns of RNA localization.

Structural similarities between hammerhead ribozymes and the spliceosomal RNAs could be responsible for lack of ribozyme cleavage in yeast.

Hammerhead ribozymes have been proposed as potential therapeutic agents for the treatment of viral and other diseases. However, a clear understanding of the cleavage reaction in vivo is not available at present. In these studies, we chose the yeast Saccharomyces cerevisiae as a model system to study the effects of hammerhead cleavage on gene expression in vivo. Several reporter genes were employed to monitor the self-cleaving characteristics of three different ribozymes. We show that these ribozymes decrease expression of some reporter genes by interfering with splice site selection or translation initiation and not by in vivo cleavage of the RNA transcripts. In fact, it appears that although these ribozymes can efficiently self-cleave the RNA in vitro, they are not able to function in vivo. We have identified a yeast splicing protein that interacts in vivo with our cis-ribozyme by specifically recognizing the ribozyme structure (Lin and Rossi, 1996). This interaction does not occur if different secondary structures are used in place of the ribozyme. The binding of this protein to the ribozyme can account for the inability of ribozymes to efficiently cleave in yeast. Remarkably, when yeast extracts are added to in vitro trans-cleavage reactions, the cleavage ability of the ribozyme is hampered, whereas the addition of mammalian extracts yields an enhancement of the reactions. These results confirm the presence of factor(s) that can block ribozyme function in the yeast intracellular environment.

Localization of Xenopus Vg1 mRNA by Vera protein and the endoplasmic reticulum.

In many organisms, pattern formation in the embryo develops from the polarized distributions of messenger RNAs (mRNAs) in the egg. In Xenopus, the mRNA encoding Vg1, a growth factor involved in mesoderm induction, is localized to the vegetal cortex of oocytes. A protein named Vera was shown to be involved in Vg1 mRNA localization. Vera cofractionates with endoplasmic reticulum (ER) membranes, and endogenous Vg1 mRNA is associated with a subcompartment of the ER. Vera may promote mRNA localization in Xenopus oocytes by mediating an interaction between the Vg1 3' untranslated region and the ER subcompartment.

Ribozymes expressed within the loop of a natural antisense RNA form functional transcription terminators.

As an initial step towards developing a widely applicable system for expressing small ribozymes (Rz) in various cell types using T7 RNA polymerase, we have replaced the loop domain of a natural prokaryotic antisense RNA (RNAout from Tn10), with hammerhead (Hh) Rz. RNAout was chosen, because the stem of its secondary structure gives it an unusually long half life in Escherichia coli which should also confer in vivo stability to small RNA sequences expressed within its loop domain. In order to define the 3' end of the Rz-RNAout chimeric RNAs, a poly(U) tract was inserted just 3' of the RNAout stem. Molecular analysis indicates that these RNAs function both as transcription terminators and Rz. In addition, the RNAs are stable in E. coli and can be expressed in mammalian cells. These results show that certain characteristics of a naturally evolved RNA, RNAout in this case, can be used to provide additional functions to short RNAs containing Hh Rz without disrupting the enzymatic activity of the Rz.

A simple method for randomly mutating cloned DNA fragments by using chemical mutagens and the polymerase chain reaction.

Random mutagenesis of cloned DNA fragments can be an important tool for understanding genes and/or gene products. I report here a simple procedure for generating random mutations in any cloned DNA fragment in vitro. This method utilizes the polymerase chain reaction (PCR) to amplify chemically damaged single-stranded DNA containing the DNA fragment of interest. A mutagenized pool of plasmid DNA is produced by cloning the PCR product (containing chemically induced sequence heterogeneity) into a new vector. Therefore, extreme chemical treatments can be used that biologically inactivate the original vector. In this report, 10% of the plasmids contained a mutation in a 97-nucleotide DNA insert, giving a mutagenesis frequency of 10(-3)/base pair. Of 3000 yeast transformants screened, seven intron mutations were isolated that activate a cryptic 3' splice site, suggesting that 2% of the mutations could give rise to the desired phenotype. The seven mutations included transitions, transversions and single nucleotide deletions, demonstrating the well-balanced repertoire of nucleotide changes derived from this procedure.

Biological and functional aspects of catalytic RNAs.

Several naturally occurring ribozymes have now been well characterized with respect to their in vivo and in vitro activities. Through detailed biochemical and genetic analyses, it has become possible to alter the substrate specificity of each ribozyme using simple Watson-Crick base pairing. Several laboratories, therefore, have designed ribozymes to cleave viral or other cellular transcripts in vitro with the hope of developing these molecules as antiviral or therapeutic agents. In addition to Watson-Crick base pairing, however, other factors such as protein or RNA tertiary interactions are involved in the ribozyme cleavage activity. Although several engineered ribozymes have been used successfully to reduce gene expression in vivo, it is difficult to determine whether gene expression has been reduced by the cleaving activity of the ribozyme or by its inherent antisense activity. In order to discriminate between these two activities and optimize potentially therapeutic ribozymes, it is imperative to develop in vivo assays in which the antisense activity of ribozymes is negligible.

Unexpected point mutations activate cryptic 3' splice sites by perturbing a natural secondary structure within a yeast intron.

The 3' splice site of the budding yeast Kluyveromyces lactis actin gene (ACT) intron is distally spaced (122 nucleotides) from its branchpoint and is also preceded by a silent PyAG located 43 nucleotides upstream. We devised a genetic screen that resulted in the isolation of several randomly induced cis-acting mutations that activate the silent PyAG as a 3' splice site. These mutations fall within a region surrounding this PyAG, which can hypothetically fold into a higher-order structure. Site-directed mutational analyses demonstrate that a hairpin structure in this region is required for correct 3' splice-site selection. Analysis of the point mutations suggests that local breathing of the hairpin near the first PyAG can lead to its activation. These data demonstrate that 3' splice-site selection is not a consequence of a linear, directional scanning mechanism, but support the notion of a critical positioning requirement for 3' splice-site selection. We speculate on the possible origin of this intron-encoded structural motif, which has homology to a bacterial transposon and suggests one possible origin for alternative splicing mechanisms in higher eukaryotes.

Exploring the use of antisense, enzymatic RNA molecules (ribozymes) as therapeutic agents.

Antisense catalytic RNAs that specifically base-pair with and cleave target RNA sequences have potential for use as therapeutic agents against viral as well as endogenous gene expression. With the ultimate goal of developing anti-human immunodeficiency virus type 1 (HIV-1) ribozymes for therapeutic use, we have been exploring ways to improve upon the functional activity of ribozymes in living cells. This is being done by the systematic exploration of parameters that affect antisense, and hence ribozyme, function. These include target accessibility, stability of the catalyst, methods for delivery, and intracellular localization of the ribozyme. In addition, we have been examining the kinetic consequences of having extra, nontargeted sequences appended to the ribozyme flanking sequences. Perhaps the single most important consideration for ribozyme effectiveness in an intracellular environment is the accessibility of the target RNA for cleavage. By exploiting the mechanisms by which naturally occurring antisense RNAs interact with their target sequences, we hope to be able to address this problem of targeting and fully capitalize upon the potential of ribozymes as therapeutic agents.

Kluyveromyces lactis maintains Saccharomyces cerevisiae intron-encoded splicing signals.

The actin (ACT) gene from the budding yeast Kluyveromyces lactis was cloned, and the nucleotide sequence was determined. The gene had a single intron 778 nucleotides in length which possessed the highly conserved splicing signals found in Saccharomyces cerevisiae introns. We demonstrated splicing of heterologous ACT transcripts in both K. lactis and S. cerevisiae.

Reconstitution of quinone reduction and characterization of Escherichia coli fumarate reductase activity.

Resolution of the fumarate reductase complex (ABCD) of Escherichia coli into reconstitutively active enzyme (AB) and a detergent preparation containing peptides C and D resulted in loss of quinone reductase activity, but the phenazine methosulfate or fumarate reductase activity of the enzyme was unaffected. An essential role for peptides C and D in quinone reduction was confirmed by restoration of this activity on recombination of the respective preparations. Neither peptide C nor peptide D by itself proved capable of permitting quinone reduction and membrane binding by the enzyme when E. coli cells were transformed with plasmids coding for the enzyme and the particular peptides. Transformation of a plasmid coding for all subunits resulted in a 30-fold increase in membrane-bound complex, which exhibited, however, turnover numbers for succinate oxidation and fumarate reduction that were intermediate between the high values characteristic of chromosomally produced complex and the relatively low values found for the isolated complex. It is also shown that preparations of the isolated complex and membrane-bound form of the enzyme, as obtained from anaerobically grown cells, are in the deactivated state owing to the presence of tightly bound oxalacetate and thus must be activated prior to assay.