Unstable RNA Research Articles

Multiple hok genes on the chromosome of Escherichia coli.

The hok/sok locus of plasmid R1 mediates plasmid stabilization by the killing of plasmid-free cells. Many bacterial plasmids carry similar loci. For example, the F plasmid carries two hok homologues, flm and srnB, that mediate plasmid stabilization by this specialized type of programmed cell death. Here, we show that the chromosome of E. coli K-12 codes for five hok homologous loci, all of which specify Hok-like toxins. Three of the loci appear to be inactivated by the insertion elements IS150 or IS186 located close to but not in the toxin-encoding reading frames (i.e. hokA, hokC and hokE), one system is probably inactivated by point mutation (hokB), whereas the fifth system is inactivated by a major genetic rearrangement (hokD). In the ECOR collection of wild-type E. coli strains, we identified hokA and hokC loci without IS elements. A molecular and a genetic analysis show that the hokA and hokC loci specify unstable antisense RNAs and stable toxin-encoding mRNAs that are processed at their 3' ends. An alignment of the mRNA sequences reveals all the regulatory elements known to be required for correct folding and refolding of the plasmid-encoded mRNAs. The conserved elements include fbi that ensure a long-range interaction in the full-length mRNAs, and tac and antisense RNA target stem-loops that are required for translation and rapid antisense RNA binding of the processed mRNAs. Consistently, we find that the chromosome-encoded mRNAs are processed at their 3' ends, resulting in the presumed translationally active mRNAs. Despite the presence of all of the regulatory elements, the chromosome-encoded loci do not mediate plasmid stabilization by killing of plasmid-free cells. The chromosome-encoded mRNAs are poorly translated in vitro, thus yielding an explanation for the lacking phenotype. These observations suggest that the chromosomal hok-like genes may be induced by an as yet unknown signal.

Tsix, a gene antisense to Xist at the X-inactivation centre

In mammals, dosage compensation is achieved by X inactivation and is regulated in cis by the X-inactivation centre (Xic) and Xist. The Xic controls X-chromosome counting, choice of X to inactivate and initiation of silencing. Xic action culminates in a change in Xist RNA property from a scarce, unstable RNA to highly expressed Xist RNA that coats the future inactive X. Deleting a 65-kb region downstream of Xist results in constitutive Xist expression and X inactivation, implying the presence of a cis-regulatory element. In this region, we now report the discovery of a gene antisense to Xist. Tsix is a 40-kb RNA originating 15 kb downstream of Xist and transcribed across the Xist locus. Tsix sequence is conserved at the human XIC. Tsix RNA has no conserved ORFs, is seen exclusively in the nucleus and is localized at Xic. Before the onset of X inactivation, Tsix is expressed from both X chromosomes. At the onset of X inactivation, Tsix expression becomes monoallelic, is associated with the future active X and persists until Xist is turned off. Tsix is not found on the inactive X once cells enter the X-inactivation pathway. Tsix has features suggesting a role in regulating the early steps of X inactivation, but not the silencing step.

Kissing and RNA stability in antisense control of plasmid replication

Plasmids have been invaluable models for studies of sophisticated gene-regulation circuits. These studies show that plasmids have different mechanisms for measuring their intracellular concentrations and controlling their copy numbers[ 1 Wagner E.G.H Simons R.W Annu. Rev. Microbiol. 1994; 48: 713-742 Crossref Scopus (364) Google Scholar ]. Many plasmids use small, unstable antisense RNAs as principal regulators[ 1 Wagner E.G.H Simons R.W Annu. Rev. Microbiol. 1994; 48: 713-742 Crossref Scopus (364) Google Scholar , 2 Eguchi Y Itoh T Tomizawa J Annu. Rev. Biochem. 1991; 60: 631-652 Crossref PubMed Scopus (273) Google Scholar ](iteron-controlled plasmids[ 3 Nordström K Cell. 1990; 63: 1121-1124 Abstract Full Text PDF Scopus (80) Google Scholar ]are not considered here). These inhibitors bind to their cognate target RNAs and, through a variety of mechanisms[ 1 Wagner E.G.H Simons R.W Annu. Rev. Microbiol. 1994; 48: 713-742 Crossref Scopus (364) Google Scholar ], prevent initiation of plasmid replication. The complementarity between antisense and target RNAs suggested that RNA duplexes are the inhibitory structures.

Polyadenylation of stable RNA precursors in vivo.

Polyadenylation at the 3' terminus has long been considered a specific feature of mRNA and a few other unstable RNA species. Here we show that stable RNAs in Escherichia coli can be polyadenylated as well. RNA molecules with poly(A) tails are the major products that accumulate for essentially all stable RNA precursors when RNA maturation is slowed because of the absence of processing exoribonucleases; poly(A) tails vary from one to seven residues in length. The polyadenylation process depends on the presence of poly(A) polymerase I. A stochastic competition between the exoribonucleases and poly(A) polymerase is proposed to explain the accumulation of polyadenylated RNAs. These data indicate that polyadenylation is not unique to mRNA, and its widespread occurrence suggests that it serves a more general function in RNA metabolism.

Developmentally Regulated Xist Promoter Switch Mediates Initiation of X Inactivation

Developmental regulation of the mouse Xist gene at the onset of X chromosome inactivation is mediated by RNA stabilization. Here, we show that alternate promoter usage gives rise to distinct stable and unstable RNA isoforms. Unstable Xist transcript initiates at a novel upstream promoter, whereas stable Xist RNA is transcribed from the previously identified promoter and from a novel downstream promoter. Analysis of cells undergoing X inactivation indicates that a developmentally regulated promoter switch mediates stabilization and accumulation of Xist RNA on the inactive X chromosome.

Role of the nonstructural polyproteins in alphavirus RNA synthesis.

Alphaviruses are of general interest because they replicate in such a broad range of animals, including invertebrates as well as mammals and birds. This broad replicative potential is an essential part of their transmission strategy. The successful synthesis of many infectious progeny starts with the translation of the genome into the viral nonstructural proteins (nsPs) that form an unstable viral RNA-dependent RNA polymerase activity, which copies the infecting genome into a complementary, or minus-strand, RNA template of genome-length. Complexes composed of the minus-strand templates and long-lived viral replicases or transcriptases then produce throughout the rest of the infectious cycle large amounts of genomes or via internal initiation of a 3’-co-terminal, subgenomic mRNA. The 26S mRNA serves as a template for translation of the viral capsid and envelope proteins. Thus, the crucial step after an alphavirus enters the cell is the formation of the initial replication complex that produces minus-strand RNA since the number of minus strand templates that form during the early part of the infectious cycle determines the amount and rate of viral plus-strand RNA synthesis. This review will highlight research efforts directed toward understanding the synthesis of alphavirus RNA and its regulation.KeywordsNonstructural ProteinStrand SynthesisSemliki Forest VirusSindbis VirusAlfalfa Mosaic VirusThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Cyclooxygenase 2 RNA message abundance, stability, and hypervariability in sporadic Alzheimer neocortex.

Long-term treatment by nonsteroidal anti-inflammatory drugs has been shown to decrease the incidence of Alzheimer's disease (AD). Both platelet-activating factor and interleukin-1beta, potent mediators of the inflammatory and immune response, strongly induce transcription of the cyclooxygenase-2 (COX-2) gene in brain cells. Using Northern and RT-PCR analysis, we have determined in 15 control and 10 sporadic AD human neocortical samples (age range, 60-82 yr; postmortem interval [PMI] range, 0.7-16.0 hr) the levels of COX-2 RNA in relation to the constitutively expressed COX-1 and beta-actin RNA message levels. Our results indicate that in short PMI brain, COX-1 and COX-2 transcripts are relatively low abundance RNA messages, ranging from a mean of 6.8% of the beta-actin signal in controls to 8.5% of the beta-actin signal in AD-affected brain. A large variation in the signal intensity for COX-2 RNA was noted in both control and AD; although there was a trend for higher COX-2 RNA message abundance in AD neocortex to +11.5% of that of controls, it did not reach statistical significance (ANOVA = 0.45). Several human tissues, including heart, skeletal muscle, lung, kidney, and spinal cord, displayed 4.6- and 2.8-kb COX-2 RNA message isoforms; however, the 4.6-kb COX-2 RNA predominated in the hippocampus and association neocortex. COX-2 RNA message was found to be degraded at similar rates in both control and AD tissues, and a strong positive correlation between the PMI and the intensity of the COX-2 RNA signal was noted (ANOVA = 0.006). Linear regression analysis indicated that the 4.6-kb COX-2 RNA is an unstable short-lived RNA species with a half-life of not more than 3.5 hr, a feature characteristic of immediate early gene transcripts. Individual hypervariability in COX-2 RNA message abundance may reflect various degrees of expression of AD-related inflammatory processes.

Regulation of plasmid R1 replication: PcnB and RNase E expedite the decay of the antisense RNA, CopA.

The replication frequency of plasmid R1 is controlled by an unstable antisense RNA, CopA, which, by binding to its complementary target, blocks translation of the replication rate-limiting protein RepA. Since the degree of inhibition is directly correlated with the intracellular concentration of CopA, factors affecting CopA turnover can also alter plasmid copy number. We show here that PcnB (PAPl-a poly(A)polymerase of Escherichia coli) is such a factor. Previous studies have shown that the copy number of ColE1 is decreased in pcnB mutant strains because the stability of the RNase E processed form of RNAI, the antisense RNA regulator of ColE1 replication, is increased. We find that, analogously, the twofold reduction in R1 copy number caused by a pcnB lesion is associated with a corresponding increase in the stability of the RNase E-generated 3' cleavage product of CopA. These results suggest that CopA decay is initiated by RNase E cleavage and that PcnB is involved in the subsequent rapid decay of the 3' CopA stem-loop segment. We also find that, as predicted, under conditions in which CopA synthesis is unaffected, pcnB mutation reduces RepA translation and increases CopA stability to the same extent.

Programmed cell death by hok/sok of plasmid R1: Coupled nucleotide covariations reveal a phylogenetically conserved folding pathway in the hok family of mRNAs

The hok/sok system of plasmid R1 mediates plasmid maintenance by killing of plasmid-free cells. Translation of the stable toxin-encoding hok mRNA is repressed by the unstable Sok antisense RNA. Using genetic algorithm simulations and phylogenetic comparisons, we analyse five plasmid-encoded and two chromosome-encoded hok-homologous mRNAs. A similar folding pathway was found for all mRNAs. Metastable hairpins at the very 5′-ends of the mRNAs were predicted to prevent the formation of structures required for translation and antisense RNA binding. Thus the folding of the mRNA 5′-ends appears to explain the apparent inactivity of the nascent transcripts. In the full-length mRNAs, long-range 5′ to 3′ interactions were predicted in all cases. The 5′ to 3′ interactions lock the mRNAs in inactive configurations. Translation of the mRNAs is activated by 3′ exonucleolytic processing. Simulation of the 3′ processing predicted that it triggers rearrangements of the mRNA 5′-ends with the formation of translational activator and antisense RNA target hairpins. Alignment of the mRNA sequences revealed a large number of nucleotide covariations that support the existence of the proposed secondary structures. Furthermore, coupled covariations support the folding pathway and provide evidence that the mRNA 5′-ends pair with three different partners during the proposed series of dynamic RNA rearrangements.

Stabilization of Xist RNA Mediates Initiation of X Chromosome Inactivation

The onset of X inactivation is preceded by a marked increase in the level of Xist RNA. Here we demonstrate that increased stability of Xist RNA is the primary determinant of developmental up-regulation. Unstable transcript is produced by both alleles in XX ES cells and in XX embryos prior to the onset of random X inactivation. Following differentiation, transcription of unstable RNA from the active X chromosome allele continues for a period following stabilization and accumulation of transcript on the inactive X allele. We discuss the implications of these findings in terms of models for the initiation of random and imprinted X inactivation.

Programmed cell death in bacteria: translational repression by mRNA end-pairing.

The hok/sok and pnd systems of plasmids R1 and R483 mediate plasmid maintenance by killing plasmid-free cells. Translation of the exceptionally stable hok and pnd mRNAs is repressed by unstable antisense RNAs. The different stabilities of the killer mRNAs and their cognate repressors explain the onset of translation in plasmid-free cells. The full-length hok and pnd mRNAs are inert with respect to translation and antisense RNA binding. We have previously shown that the mRNAs contain two negative translational control elements. Thus, the mRNAs contain upstream anti-Shine-Dalgarno elements that repress translation by shielding the Shine-Dalgarno elements. The mRNAs also contain fold-back-inhibition elements (fbi) at their 3' ends that are required to maintain the inert mRNA configuration. Using genetic complementation, we show that the 3' fbi elements pair with the very 5' ends of the mRNAs. This pairing sets the low rate of 3' exonucleolytical processing, which is required for the accumulation of an activatable pool of mRNA. Unexpectedly, the hok and pnd mRNAs were found to contain translational activators at their 5' ends (termed tac). Thus, the fbi elements inhibit translation of the full-length mRNAs by sequestration of the tac elements. The fbi elements are removed by 3' exonucleolytical processing. Mutational analyses indicate that the 3' processing triggers refolding of the mRNA 5' ends into translatable configurations in which the 5' tac elements base pair with the anti-Shine-Dalgarno sequences.

Mechanism of Post-segregational Killing: Secondary Structure Analysis of the Entire Hok mRNA from Plasmid R1 Suggests a Fold-back Structure that Prevents Translation and Antisense RNA Binding

Thehok/soksystem of plasmid R1 mediates plasmid stabilization by killing of plasmid-free cells. The Hok mRNA is very stable and can be translated into Hok killer protein. Translation of the Hok mRNA is inhibited by the small unstable Sok antisense RNA. Translation ofhokis coupled to an overlapping reading frame termedmok. Translation ofmokis tightly regulated by Sok RNA, and Sok RNA thus regulateshoktranslation indirectly throughmok. The rapid decay of Sok RNA explains the onset of Hok synthesis in newborn plasmid-free segregants. However, a second control level is superimposed on this simple induction scheme, since the full-length Hok mRNA was found to be translationally inactive whereas a 3′-end truncated version of it was active. We have therefore previously suggested, that the 3′-terminal region of the full-length Hok mRNA encodes an element which prevents its translation. This element was termedfbi(fold-back inhibition). Here we describe thein vitrosecondary structure of the entire Hok mRNA. Our results suggest a closed structure in which the 3′-end of the full-length Hok mRNA folds back onto the translational initiation region ofmok. This structure explains why full-length Hok mRNA is translationally silent. The proposed structure was further supported by results obtained using mutations in the 3′-endfbielement. These “structure closing” mutations affected the structure much further upstream in themoktranslational initiation region and concomitantly prevented antisense RNA binding to the same region of the mRNA. These results lend further support to the induction model that explains onset of Hok mRNA translation in plasmid-free segregants. The most important regulatory element in this model is the FBI structure formed between the 3′-end and themoktranslational initiation region. This structure renders Hok mRNA translationally inactive and prevents antisense RNA binding, thus allowing the accumulation of a pool of mRNA which, by slow 3′-end processing, is activated in plasmid-free segregants, eventually leading to the elimination of these cells.

Mechanism of post-segregational killing: translation of Hok, SrnB and Pnd mRNAs of plasmids R1, F and R483 is activated by 3'-end processing.

The gene systems hok/sok of R1, srnB of F and pnd of R483 mediate plasmid maintenance by killing of plasmid-free segregants. Translation of the very stable mRNAs encoding the killer proteins is regulated by small unstable antisense RNAs. The differential decay rates of the inhibitory antisense RNAs and the mRNAs encoding the killer proteins is the basis for the onset of killer mRNA translation in newborn plasmid-free segregants and the killing of these cells. We have suggested previously that this requires that the killer mRNAs occur in two forms. A translationally inactive form was proposed to be converted into a 3'-truncated, translationally active mRNA. In the presence of the antisense RNA, translation from this killer mRNA should be inhibited. In this communication we present in vivo and in vitro evidence that support this model. The requirement for 3'-processing for killer gene expression is demonstrated. By using in vitro techniques it is shown that full-length Hok mRNA is translationally inactive, whereas a 3'-end truncated version of the Hok mRNA is translationally active. In vitro secondary structure probing suggests that the 3'-end of the full-length Hok mRNA folds back onto the translational initiation region of the mok gene and thereby inhibits translation of the mRNA. By inference we conclude that the Pnd and SrnB mRNAs are regulated by a similar mechanism.

PcnB is required for the rapid degradation of RNAI, the antisense RNA that controls the copy number of ColE1-related plasmids.

The replication of ColE1-related plasmids is controlled by an unstable antisense RNA, RNAI, which can interfere with the successful processing of the RNAII primer of replication. We show here that a host protein, PcnB, supports replication by promoting the decay of RNAI. In bacterial strains deleted for PcnB a stable, active form of RNAI, RNAI*, which appears to be identical to the product of 5'-end processing by RNAase E, accumulates. This leads to a reduction in plasmid copy number. We show, using a GST-PcnB fusion protein, that PcnB does not interfere with RNAI/RNAII binding in vitro. The fusion protein, like PcnB, has polyadenylating activity and is able to polyadenylate RNAI (and also another antisense RNA, CopA) in vitro.

Quantitative Model of ColE1 Plasmid Copy Number Control

Initiation of replication of the Escherichia coli plasmid ColE1 is inhibited by formation of a complex between a small RNA (RNA I) and the pre-primer for DNA synthesis (RNA II). Complex formation (and inhibition of replication) is enhanced by the plasmid-encode Rom protein. The in vitro kinetics of complex formation were previously studied both experimentally and theoretically. The in vivo concentrations and half-lives of RNA I, RNA II and Rom protein have been measured recently. We present a dynamic model for the in vivo replication control mechanism that accounts for the measured concentration values. From the model we deduce a simple formula for the steady-state plasmid concentration. Our result agree with a previous simple steady-state analysis done by Brenner and Tomizawa, in that plasmid copy number is most strongly dependent on the per plasmid rate of RNA I synthesis. However, our model predicts other parameter dependencies that are not evident from or at variance with the previous analysis. Accordingly, we predict that plasmid copy number is greatly influenced by changes in the rate constant describing the formation of an initial unstable RNA I-RNA II complex, but is only slightly influenced by changes in the dissociation rate of this complex. Plasmid copy number per average cell volume is predicted to increase linearly with increases in the RNA II synthesis rate and with increases in the generation time of the host culture. Rom protein, which promotes conversion of the unstable RNA I-RNA II complex to a stable complex, serves to decrease copy number, however, its presence or absence does not seem to qualitatively alter the copy number control mechanism. Our model predicts the quantitative increase of plasmid copy number in rom - mutants. Several experiments are suggested to investigate the predications of the model.

Evidence for instability of mRNAs containing AUUUA motifs mediated through translation-dependent assembly of a > 20S degradation complex.

Many short-lived mRNAs, including those encoding lymphokines, cytokines, and proto-oncogenes, contain an AU-rich sequence in their 3'-untranslated regions. These AU domains and, more specifically, AUUUA motifs within them, are widely thought to mediate the extreme instability of the corresponding mRNAs. This is most clearly true for granulocyte monocyte colony stimulating factor (GM-CSF) mRNA whose AUUUA motifs are conserved phylogenetically and whose presence in an otherwise stable beta-globin mRNA results in a 50-fold decrease in accumulated mRNA level. We show that RNA instability conferred by the GM-CSF AU motif requires the mRNA to be actively translated and the AU motif to be within the 3'-untranslated region. By analysis of the sedimentation characteristics, we identify a large (> 20S), divalent cation-independent complex found only on unstable RNAs. Like instability, complex formation (1) is dependent on translation of the RNA, (2) requires intact AUUUA motifs, and (3) is blocked by ribosome translocation across the AU-rich motif. We propose that RNA instability mediated by the AU motif is achieved through translation-dependent assembly of this large mRNA-destabilizing complex.

Mechanism of killer gene activation. Antisense RNA-dependent RNase III cleavage ensures rapid turn-over of the stable Hok, SrnB and PndA effector messenger RNAs

The hok/sok, srnB and pnd systems of plasmids RI, F and R438 mediate plasmid maintenance by killing plasmid-free segregants. The systems encode exceptionally stable full-length mRNAs that code for potent cell toxins that kill the cells from within. The systems also produce truncated mRNAs whose appearance is correlated with killing activity. The truncated mRNAs are shortened by 35 to 70 nucleotides in the 3′ ends, but have the same 5′ ends as the full-length transcripts. Translation of the stable killer mRNAs is regulated by unstable antisense RNAs that are complementary to the leader regions of the full-length and truncated mRNAs. We show here, that both the presence of the antisense RNA and of the host enzyme RNase III is required for rapid cleavage of the truncated mRNAs, and we map the cleavage point in the Hok mRNA in vitro and in vivo to be located between nucleotides +245 and +246. The RNase III cleavage products of the Hok mRNA were found to be very unstable in vivo. Thus, RNase III cleavage seems to be the initial event leading to decay of the killer mRNAs. In an rnc − strain, the truncated mRNA species were found in steady-state cells. This observation indicates that the truncated mRNAs are formed constitutively and independently of the presence of the antisense RNAs. Thus, the antisense RNAs prevent the accumulation of the truncated mRNAs solely by mediating their rapid hydrolysis by RNase III. Furthermore, the generation of the truncated killer mRNAs in the rnc − host indicate that RNase III is dispensable for induction of the killer gene systems. Based on these and on observations obtained previously, we present a molecular model that explains the activation of the killer mRNAs in plasmid-free segregants and after addition of rifampicin.

Control of 4-8S RNA transcription at the midblastula transition in Xenopus laevis embryos.

Transcription of Xenopus laevis U1 snRNA genes is subject to a precise program with respect both to the timing of activation at the midblastula transition (MBT) and to the relative levels of the two embryonic U1 RNAs (xU1b1 and b2) that are made. Here, we demonstrate that exogenous xU1b genes injected into developing X. laevis embryos come under the same controls as the endogenous genes. Injected U1 genes, unlike exogenous RNA polymerase III genes, remain quiescent until MBT and their activation at MBT requires protein synthesis during the early cleavage stages. Significantly, the onset of 4-8S RNA transcription occurs at the normal time, even when the DNA content of the embryo has been increased by injection of exogenous DNA or reduced through cleavage arrest, indicating that transcriptional activation at MBT is independent of the ratio of DNA (nucleus) to cytoplasm. In cleavage-arrested (coenocytic) embryos, the reduced level of DNA at MBT results both in a decrease in snRNA and tRNA synthesis (reflecting the lower gene dosage) and in a prolonged synthesis of large amounts of unusual RNA polymerase III transcripts, OAX RNAs. In normally cleaving embryos, small amounts of these unstable OAX RNAs (encoded by satellite I DNA) are synthesized only briefly at MBT. Our demonstration that RNA and DNA metabolism is aberrant in cleavage-arrested embryos requires reevaluation of previous experiments on transcriptional activation that utilized such coenocytic embryos.

The rifampicin‐inducible genes srn6 from F and pnd from R483 are regulated by antisense RNAs and mediate plasmid maintenance by kiiling of plasmid‐free segregants

The gene systems srnB of plasmid F and pnd of plasmid R483 were discovered because of their induction by rifampicin. Induction caused membrane damage, RNase I influx, degradation of stable RNA and, consequently, cell killing. We show here that the srnB and pnd systems mediate efficient stabilization of a mini-R1 test-plasmid. We also show that the killer genes srnB' and pndA are regulated by antisense RNAs, and that the srnC- and pndB-encoded antisense RNAs, denoted SrnC- and PndB-RNAs, are unstable molecules of approximately 60 nucleotides. The srnB and pndA mRNAs were found to be very stable. The differential decay rates of the inhibitory antisense RNAs and the killer-gene-encoding mRNAs explain the induction of these gene systems by rifampicin. Furthermore, the observed plasmid-stabilization phenotype associated with the srnB and pnd systems is a consequence of this differential RNA decay: the newborn plasmid-free cells inherit the stable mRNAs, which, after decay of the unstable antisense RNAs, are translated into killer proteins, thus leading to selective killing of the plasmid-free segregants. Thus our observations lead us to conclude that the F srnB and R483 pnd systems are phenotypically indistinguishable from the R1 hok/sok system, despite a 50% dissimilarity at the level of DNA sequence.

Analysis of unstable RNA transcripts of insecticidal crystal protein genes of Bacillus thuringiensis in transgenic plants and electroporated protoplasts

We have examined expression of several insecticidal crystal protein (ICP) genes of Bacillus thuringiensis in transgenic tobacco plants and electroporated carrot protoplasts. We determined that low levels of lepidopteran toxin cryIA(b) ICP gene expression in plants and electroporated carrot cells is due to RNA instability. We used a series of 3' deleted by cryIA(b) constructs directed by the cauliflower mosaic virus 35S promoter to demonstrate that this instability is minimally contained in the first 579 bases of the gene in both systems. This instability may result from 5'----3' as well as 3'----5' RNA metabolism. The coleopteran toxic cryIIIA gene was also examined in electroporated carrot cells, and found to be poorly expressed. A model for improvement of ICP RNA stability in plants is presented.