Search

The C4 repressor of the temperate bacteriophages P1 and P7 inhibits antirepressor (Ant) synthesis and is essential for establishment and maintenance of lysogeny. C4 is an antisense RNA acting on a target, Ant mRNA, which is transcribed from the same promoter. The antisense-target RNA interaction requires processing of C4 RNA from a precursor RNA. Here we show that 5' maturation of C4 RNA in vivo depends on RNase P. In vitro, Escherichia coli RNase P and its catalytic RNA subunit (M1 RNA) can generate the mature 5' end of C4 RNA from P1 by a single endonucleolytic cut, whereas RNase P from the E. coli rnpA49 mutant, carrying a missense mutation in the RNase P protein subunit, is defective in the 5' maturation of C4 RNA. Primer extension analysis of RNA transcribed in vivo from a plasmid carrying the P1 c4 gene revealed that 5'-mature C4 RNA was the predominant species in rnpA+ bacteria, whereas virtually no mature C4 RNA was found in the temperature-sensitive rnpA49 strain at the restrictive temperature. Instead, C4 RNA molecules carrying up to five extra nucleotides beyond the 5' end accumulated. The same phenotype was observed in rnpA+ bacteria which harbored a plasmid carrying a P7 c4 mutant gene with a single C-->G base substitution in the structural homologue to the CCA 3' end of tRNAs. Implications of C4 RNA processing for the lysis/lysogeny decision process of bacteriophages P1 and P7 are discussed.

Oligonucleotides that can hybridize to single-stranded complementary polypurine nucleic acid targets by Watson-Crick base pairing as well as by Hoogsteen base pairing, referred to here as foldback triplex-forming oligonucleotides (FTFOs), have been designed. These oligonucleotides hybridize with target nucleic acid sequences with greater affinity than antisense oligonucleotides, which hybridize to the target sequence only by Watson-Crick hydrogen bonding [Kandimalla, E. R. and Agrawal, S. Gene(1994) 149, 115-121 and references cited therein]. FTFOs have been studied for their ability to destabilize quadruplexes formation by RNA or DNA target sequences. The influence of various DNA/RNA compositions of FTFOs on their ability to destabilize RNA and DNA quadruplexes has been examined. The ability of the FTFOs to destabilize quadruplex structures is related to the structurally and thermodynamically stable foldback triplex formed between the FTFO and its target sequence. Antisense oligonucleotides (DNA or RNA) that can form only a Watson-Crick double helix with the target sequence are unable to destabilize quadruplex structures of RNA and DNA target sequences and are therefore limited in their repertoire of target sequences. The quadruplex destabilization ability of FTFOs is dependent on the nature of the cation present in solution. The RNA quadruplex destabilization ability of FTFOs is -20% higher in the presence of sodium ion than potassium ion. The use of FTFOs, which can destabilize quadruplex structure, opens up new areas for development of oligonucleotide-based therapeutics, specifically, targeting guanine-rich sequences that exist at the ends of pro- and eukaryotic chromosomes and dimerization regions of retroviral RNA.

The Escherichia coli micF RNA is a prototype for a class of antisense RNAs encoded by genes at different loci from those that code for their target RNAs. RNAs in this class exhibit only partial complementarity to their targets. micF RNA binds to and regulates the stability of ompF mRNA in response to a variety of environmental stimuli. The secondary structures of micF RNA, ompF-213 mRNA (a segment containing the 213 nucleotides at the 5'-terminus of the target message), and the micF RNA/ompF-213 mRNA duplex were analyzed in vitro by partial digestion with structure-specific ribonucleases and chemical modification. Both micF RNA and ompF mRNA have single-stranded 5'-ends and contain stable stem-loop structures. Strong phylogenetic support for the proposed secondary structure for E. coli micF RNA is provided by a comparison of structural models derived from micF sequences from related bacteria. The micF RNA/ompF-213 mRNA duplex interaction appears to involve only a short segment of micF RNA. Unfolding of only one stem-loop of micF RNA and a minor stem-loop of ompF-213 mRNA appears to be necessary to form the duplex. The probing data suggest that the Shine-Dalgarno sequence and AUG start codon of ompF mRNA, found in single-stranded regions in the free message, are base-paired to micF RNA in the RNA/RNA duplex.

Lysogenic Acetobacter methanolicus strains carrying the prophage Acm1 were found to be unable to synthesize both the capsular polysaccharide (CPS) and the O-specific side-chain of lipopolysaccharide (LPS) and to represent rough variants of the host bacterium. A 262 bp DNA fragment of phage Acm1, obviously required for interference with LPS biosynthesis, was cloned and expressed in Escherichia coli. Independently of the O-type, transformation of various E. coli strains with the recombinant DNA resulted in a suppression of biosynthesis of the O-specific chains. The DNA fragment of phage Acm1 contained three very short open reading frames of 21, 24, and 36 bp. However, attempts to express phage-encoded peptides were not successful. Instead, the Acm1-derived DNA fragment was shown to code for the synthesis of a trans-acting RNA molecule of 97 nucleotides, designated lbi (LPS biosynthesis-interfering) RNA. This RNA contains sequence complementarity to E. coli target RNA sequences and appears to have the ability to form intracellularly RNA hybrid duplexes with mRNA. The data presented in this study support the hypothesis that the phenotypic effect of conversion to rough-type LPS is accompanied by the expression of an antisense RNA of phage Acm1.

Naturally occurring antisense RNAs in prokaryotes are generally short, highly structured and untranslated. Stem-loops are always present, and loop regions serve as primary recognition structures in most cases. Single-stranded tails or internal unstructured regions are required for initiation of stable pairing between antisense and target RNA. Most antisense RNAs contain bulged-out nucleotides or small internal loops in upper stem regions. Here we investigated the role of the bulged-out nucleotides of CopA (the copy number regulator of plasmid R1) in determining the binding properties of this antisense RNA to its target in vitro and the efficiency of a translational inhibition in vivo. The introduction of perfect helicity in the region of the two bulges in CopA decreased pairing rate constants by up to 180-fold, increased equilibrium dissociation constants of the 'kissing intermediate' up to 14-fold, and severely impaired inhibition of repA expression. A previously described loop size mutant of CopA showed decreased pairing rates, but, in contrast to the bulge-less mutant CopAs, shows a decreased dissociation constant of the 'kissing complex'. We conclude that removal of the specific bulges/internal loops within the stem-loop II of CopA impairs the inhibitor, and that creation of an internal loop at a different position does not restore activity, emphasizing the optimal folding of wild-type CopA. The accompanying paper shows that an additional function of bulges can be protection from RNase III cleavage.

Bulged-out nucleotides or internal loops are present in the stem-loop structures of several antisense RNAs. We have used the antisense/target RNA system (CopA/CopT) that controls the copy number of plasmid R1 to examine the possible biological function of bulged-out nucleotides. Two regions within the major stem-loop of the antisense RNA, CopA, carry bulged-out nucleotides. Base pairing in either one or both of these regions of the stem was restored by site-specific mutagenesis and in one case a new internal loop was introduced. The set of mutant and wild-type CopA variants was characterized structurally in vitro. The results reported here indicate a possible function of the bulges: their presence protects CopA RNA from being a substrate for the double-strand-specific enzyme RNase III. In vitro cleavage rates were drastically increased when either the lower or both bulges were absent. This is paralleled by a similar, but not identical, effect of the bulges on metabolic stability of the CopA RNAs in vivo. The degradation pathways of wild-type and mutant CopA in various strain backgrounds are discussed. In the accompanying paper, we address the significance of bulges in CopA for binding to the target RNA in vitro and for its inhibitory efficiency in vivo.

Retinoic acid (RA) in the form of isotretinoin (Accutane) and tretinoin (Retin-A) is a clinically important compound in the treatment of dermatologic disorders. However, it is also a potent teratogen associated with a number of serious congenital malformations. Generally, these malformations involve the craniofacial structures derived from the first and second branchial arches. To determine how altered gene expression may contribute to the observed RA-induced defects, pregnant LM/Bc mice were administered (5 mg/kg) all-trans RA on gestational day (GD) 8:12. First and second branchial arches were removed from control and teratogen-treated embryos on GD 10:00 10:12, or 12:00, processed by in situ transcription/aRNA techniques, and analyzed for alterations in gene expression. In these studies, a panel of 40 candidate genes that are known to be important in mammalian craniofacial development were examined. This analysis revealed significant differences in the expression level of the nicotinic acetylcholine receptor subunit alpha (NAChR), transforming growth factor beta 2 (TGF beta 2), type 1 cellular retinoid binding protein (CRBP-1), retinoic acid receptor gamma (RAR gamma), and cAMP response element binding protein (CREB). The alterations observed in the expression of these genes following RA exposure may prohibit normal morphogenetic processes within the second branchial arch and lead to the observed malformations.

Several catalytic antisense RNAs directed against different regions of the genomic or antigenomic RNA of Sendai virus were constructed. All RNAs contained the same catalytic domain based on hammerhead ribozymes but some had deletions or mutations resulting in imperfect helices I and III. Pre-annealed substrate/ribozyme complexes were used to determine the rates of the cleavage process for the different ribozymes under single-turnover conditions. It was found that the sequence context surrounding the cleavable motif influenced the cleavage efficiencies. Deletions or mutations of nucleotides 2.1 or 15.1 and 15.2 according to the numbering system for hammerhead ribozymes of Hertel et al. destroyed catalytic activity. Deletions of nucleotide 2.2 or additional nucleotides in the helix I-forming region of the ribozyme did not destruct, but only reduced the cleavage efficiencies. Similar results were observed for a deletion of nucleotide 15.3. Simultaneous deletions within helices I and III resulted in alternative cleavage sites. The potential consequences for the specificity of the ribozyme reaction are discussed.

When designed to cleave a target RNA in trans, the hammerhead ribozyme contains two antisense flanks which form helix I and helix III by pairing with the complementary target RNA. The sequences forming helix II are contained on the ribozyme strand and represent a major structural component of the hammerhead structure. In the case of an inhibitory 429 nucleotides long trans-ribozyme (2as-Rz12) which was directed against the 5'-leader/gag region of the human immunodeficiency virus type 1 (HIV-1), helix II was not pre-formed in the single-stranded molecule. Thus, major structural changes are necessary before cleavage can occur. To study whether pre-formation of helix II in the non-paired 2as-Rz12 RNA could influence the observed cleavage rate in vitro and its inhibitory activity on HIV-1 replication, we extended the 4 base pair helix II of 2as-Rz12 to 6, 10, 21, and 22 base pairs respectively. Limited RNase cleavage reactions performed in vitro at 37 degrees C and at physiological ion strength indicated that a helix II of the hammerhead domain was pre-formed when its length was at least six base pairs. This modification neither affected the association rate with target RNA nor the cleavage rate in vitro. In contrast to this, extension of helix II led to a significantly decreased inhibition of HIV-1 replication in human cells. Together with the finding of others that shortening of helix II to less than two base pairs reduces the catalytic activity in vitro, this observation indicates that the length of helix II in the naturally occurring RNAs with a hammerhead domain is already close or identical to the optimal length for catalytic activity in vitro and in vivo.

Trans-cleaving hammerhead ribozymes with long target-specific antisense sequences flanking the catalytic domain share some features with conventional antisense RNA and are therefore termed 'catalytic antisense RNAs'. Sequences 5' to the catalytic domain form helix I and sequences 3' to it form helix III when complexed with the target RNA. A catalytic antisense RNA of more than 400 nucleotides, and specific for the human immunodeficiency virus type 1 (HIV-1), was systematically truncated within the arm that constituted originally a helix I of 128 base pairs. The resulting ribozymes formed helices I of 13, 8, 5, 3, 2, 1 and 0 nucleotides, respectively, and a helix III of about 280 nucleotides. When their in vitro cleavage activity was compared with the original catalytic antisense RNA, it was found that a helix I of as little as three nucleotides was sufficient for full endonucleolytic activity. The catalytically active constructs inhibited HIV-1 replication about four-fold more effectively than the inactive ones when tested in human cells. A conventional hammerhead ribozyme having helices of just 8 nucleotides on either side failed to cleave the target RNA in vitro when tested under the conditions for catalytic antisense RNA. Cleavage activity could only be detected after heat-treatment of the ribozyme substrate mixture which indicates that hammerhead ribozymes with short arms do not associate as efficiently to the target RNA as catalytic antisense RNA. The requirement of just a three-nucleotide helix I allows simple PCR-based generation strategies for asymmetric hammerhead ribozymes. Advantages of an asymmetric design will be discussed.

Using an in vitro ribonuclease protection assay, it was shown that synthetic antisense transcripts from the 5'-upstream region of the beta-tubulin gene are efficiently imported into isolated Leishmania mitochondria. Import occurred after a lag of about 30 min at 25 degrees C and was dependent on ATP. Preincubation experiments suggested that import consists of a slow interaction of mitochondria with RNA, followed by rapid ATP-dependent uptake. Import was saturable with antisense RNA at about 1 nM concentration, and sequence-specific, as shown by lack of import of other labelled transcripts. Deletion analysis demonstrated a correlation between efficiency of import and the number of oligopurine motifs on the antisense RNA. Several small ribosomal RNAs (srRNAs) and Leishmania tRNA competed with antisense RNA for import. Incubation of mitochondria with srRNAs and tRNA in the presence of radiolabelled UTP resulted in the ribonuclease-resistant labelling of these RNAs by the mitochondrial terminal uridylyl transferase. Extracts of isolated mitochondria contain a factor binding to antisense RNA, as shown by gel retardation assay. These observations indicate the presence of a receptor-mediated import pathway for srRNAs and tRNA in Leishmania mitochondria.

Antisense RNA-mediated transcriptional attenuation is the mode of replication control of several plasmids, among them pIP501. This mechanism implies that the repR mRNAs can fold into two mutually exclusive structures. The formation of one of these structures is induced by binding of the antisense RNA and results in premature termination. Since the fate of the nascent mRNA transcripts depends on the binding rate of the antisense RNA to its target, the control is kinetic. We have studied the antisense RNA, RNAIII, and target RNA, RNAII, whose interaction determines the replication frequency of plasmid pIP501. RNA secondary structures were analyzed using structure-specific RNases. RNA binding was studied in vitro with normal size and truncated RNAIII species. An in vitro single-round attenuation assay was developed that permits qualitative and quantitative assessment of inhibition by RNAIII. The effect of varying concentrations of RNAIII species on attenuation was tested and inhibition rate constants were calculated. The inhibition rate constants were at least 10 times higher than the pairing rate constants. Thus, steps preceding stable RNA duplex formation are sufficient to induce RNAIII-dependent termination of nascent RNAII transcripts.

We isolated and characterized mutants of ColE2 with increased copy number (cop) and those with reduced sensitivity to the wild-type incA gene (inc). Both types of mutations were single-base substitutions in the incA region and simultaneously increased the plasmid copy number and reduced the inhibitory activity of the incA gene on ColE2 DNA replication. Most of the cop mutations also reduced sensitivity to the wild-type incA gene. These mutations were located in the region specifying the large stem-and-loop structures of RNA I and the 5' portion of the Rep mRNA. All these results indicate that RNA I interacts with the Rep mRNA and thereby inhibits expression of the Rep protein at a post-transcriptional step and that this is probably the only mechanism that controls the ColE2 Rep protein expression. It is suggested that only portions of the nucleotides in the loop region are involved in initial (kissing) interaction of these RNAs. The total level of rep gene expression in the host cells appears to be kept constant (at a level characteristic for each cop allele) irrespective of the actual plasmid copy number above a certain level, when rep gene expression is regulated by the incA gene on the same plasmid. These seem to be the basic mechanisms for the replication control of ColE2.

The replication frequency of plasmid R1 is controlled by an antisense RNA, CopA, that inhibits the synthesis of the replication initiator protein, RepA, at the post-transcriptional level. This inhibition is indirect and affects translation of a leader peptide reading frame (tap). Translation of tap is required for repA translation (Blomberg et al., 1992). Here we asked whether an RNA stem-loop sequestering the repA ribosome-binding site blocks tap translation-independent repA expression. Destabilization of this structure resulted in tap-independent RepA synthesis, concomitant with a loss of CopA-mediated inhibition; thus, CopA acts at the level of tap translation. Structure probing of RepA mRNAs confirmed that the introduced mutations induced a local destabilization in the repA ribosome-binding site stem-loop. An increased spacing between the repA Shine-Dalgarno region and the start codon permitted even higher repA expression. In Incl alpha/IncB plasmids, an RNA pseudoknot acts as an activator for rep translation. We suggest that the regulatory pathway in plasmid R1 does not involve an activator RNA pseudoknot.

Expression of the Rep protein of colicin E2 plasmid is negatively controlled at a post-transcriptional step by a plasmid-coded RNA (RNAI). RNAI is complementary to the 5' nontranslated region of the Rep mRNA and can be folded into two stem-loop structures. Several cop mutations that increased the copy number of the plasmid have been mapped in the RNAI coding region. Here we demonstrated that RNAI and the Rep mRNA rapidly form a stable complex under near physiological conditions. The cop mutations drastically reduced the binding rate. The reduction correlated to the in vivo phenotypes of the cop mutant plasmids, indicating that binding of RNAI to the Rep mRNA is involved in the regulation of the Rep protein expression. Binding properties of RNAI and the Rep mRNA with the cop mutations and those with various extents of deletion suggested that the initial interaction of the two RNAs occurs between the loop portions of the larger of the two stem-loop structures. RNAI does not completely sequester the putative S/D region, even when the two RNAs form a complete duplex. Possible mechanisms of the inhibition of the rep gene expression by binding of RNAI to the Rep mRNA under such a situation were discussed.

The efficiency of conjugation of F-like plasmids is regulated by the FinOP fertility inhibition system. The transfer (tra) operon is under the direct control of the TraJ transcriptional activator which, in turn, is negatively regulated by FinP, an antisense RNA, and FinO, a 22 kDa protein. Recently, FinO has been shown to extend the chemical stability of FinP in vivo in the absence of traJ mRNA. The in vitro secondary structures of both the FinP and TraJ RNAs were determined by the use of single- and double-strand-specific nucleases; both RNAs were found to have double stem-loop structures that are complementary to each other and, therefore, FinP RNA and TraJ RNA have the potential to form a duplex with each other. This was verified by in vitro binding experiments. The reaction was shown to be biomolecular with an apparent rate constant (kapp) of 5 x 10(5)M-1s-1, a value that is similar to those found for other natural antisense RNA systems. Preliminary evidence for the in vivo formation of the FinP-TraJ RNA duplex was obtained by primer extension of the traJ mRNA; the presence of both FinO and FinP was required to cause a dramatic reduction in the steady-state level of traJ mRNA, perhaps as a result of RNase III degradation of the resulting RNA duplex.

Peptide nucleic acids (PNAs) are novel DNA mimics in which the sugar-phosphate backbone has been replaced with a backbone based on amino acids. PNAs exhibit sequence-specific binding to DNA and RNA with higher affinities and specificities than unmodified DNA. They are resistant to nuclease and protease attack in serum and cellular extracts and, thus, appear very promising as diagnostic and biomolecular probes, and possibly as antisense and antigene drugs.

We examined, on the basis of the fact that sense and antisense peptides have affinity for each other, whether any relationship exists between homologous restriction factor (HRF20), a membrane inhibitor of the terminal stage of homologous complement attack, and the antisense sequence of the terminal complement components C8 and C9. In this article, we demonstrate that there are two regions of C9 that contain antisense sequences to one continuous region of HRF20 and that this relationship exists between human HRF20 and human C9, but not mouse C9. We also found one region of the C8 beta chain that contains an antisense sequence to HRF20.

The catalytic domain of a hammerhead ribozyme was incorporated into a 413 nucleotides long antisense RNA directed against the 5'-leader/gag region of the human immunodeficiency virus type 1 (HIV-1) (pos. +222 to +634). The resulting catalytic antisense RNA was shown to cleave its target RNA in vitro specifically at physiological ion strength and temperature. We compared the antiviral effectiveness of this catalytic antisense RNA with that of the corresponding unmodified antisense RNA and with a mutated catalytic antisense RNA, which did not cleave the substrate RNA in vitro. Each of these RNAs was co-transfected into human SW480 cells together with infectious complete proviral HIV-1 DNA, followed by analysis of HIV-1 replication. The presence of the catalytically active domain resulted in 4 to 7 fold stronger inhibition of HIV-1 replication as compared to the parental antisense RNA and the inactive mutant. Kinetic and structural studies performed in vitro indicated that the ability for double strand formation was not changed in catalytic antisense RNA versus parental antisense RNA. Together, these data suggest that the ability to cleave target RNA is a crucial prerequisite for the observed increase of inhibition of the replication of HIV-1.

The rate of double strand formation between procaryotic antisense RNA and complementary RNA in vitro is known to correlate with the effectiveness of antisense RNA in vivo. In this work, an in vitro assay for determining the hybridization rates of a large number of antisense RNA species was developed. A set of HIV-1-directed antisense RNAs with the same 5'-end but successively shortened 3'-ends was produced by alkaline hydrolysis of a 150 nt HIV-1-directed antisense transcript. This mixture was used to determine hybridization rates for individual chain lengths with a complementary HIV-1-derived RNA in vitro. The second order binding rate constants of individual antisense RNA species differed by more than a factor of 100, although in some cases, slow-hybridizing and fast-hybridizing antisense RNAs differed by only two or three 3'-terminally-located nucleotides. The results indicated that there was not a trivial dependence of binding rates on the chain length of antisense RNAs. Further, the binding rate constants determined in vitro for individual antisense RNA species correlated with the extent of inhibition of HIV-1 replication in vivo.

Specific features of a model for the secondary structure of the self-cleaving RNA sequences (ribozymes) of hepatitis delta virus were rigorously tested. Using a self-cleaving form of the antigenomic sequence, mutations were made in the 5' and 3' sequences of each of four duplex regions within the proposed ribozyme structure. Precursor RNA from each variant sequence was prepared and the kinetics of cleavage in 10 mM Mg2+ at 37 degrees was examined. The data was quantified to determine an end point and a first-order rate constant for cleavage with each mutant by fitting the data to the exponential form of the first-order rate equation. With regard to the final extent of cleavage, most mutations in these regions appeared to have little effect, however, the kinetics indicated that disruption of the potential for basepairing resulted in dramatic decreases in the rate constant for cleavage. These results are consistent with the idea that most of the mutations affected ribozyme activity rather than an equilibrium between precursor and cleavage products. Mutations that reduced rates were compensated by changes that restored the potential for Watson-Crick pairing. Ribonuclease probing of ribozyme variants containing mismatches and compensatory changes allowed direct correlation of structural changes with the mutations. This provided an independent validation of the functional kinetic assay. Thus, site-directed mutagenesis was consistent with a proposed ribozyme secondary structure containing 4 distinct base-paired regions.

Most natural antisense RNAs display a high degree of secondary structure with stem-loops as their most prominent feature. Mutations affecting the inhibitory activity of these RNAs most often map in or close to loop regions in both the antisense and target RNAs. The primary recognition loops often contain 5-7 unpaired nucleotides. Nucleotide changes in the loops affect the binding rate and, hence, the inhibitory effect on the activity of the target RNA. Here we address the question whether loop sizes affect binding rates between antisense and target RNAs, using the replication control system of plasmid R1 as a model system. By creating a series of loop size mutants we show that loop size alterations have strong effects on the binding rates between the two reactant RNAs in vitro, and that most of the mutations analyzed display corresponding effects on antisense RNA control in vivo. Our data suggest that the three-dimensional structures of antisense and target RNA stem-loops are crucial for determining binding rates. The implications of these results for the design of efficient artificial antisense RNA control systems are discussed.

RNAs derived from the genomic and antigenomic hepatitis delta virus are capable of self-cleavage, and thus have the potential for serving as ribozymes in a trans-cleaving reaction. Because the catalytic core of such an enzymatic RNA was not evident from phylogenetic data, we took a step-wise approach to identifying the core, reducing the RNA in size, and characterizing various properties for each size class. Thus, a 186-nucleotide antigenomic RNA (termed Ag180) was found to be capable of cleaving well in 20 M formamide (Smith and Dinter-Gottlieb, 1991), and this unusual stability in formamide was lost by reducing the 3' end of the molecule, leaving a 140-nucleotide RNA (Ag 140). Both RNAs showed only intramolecular cleavage at a wide range of concentrations, and a number of conformers could be seen in the Ag140 RNA, some of which were resistant to cleavage at 37 degrees C. Since Ag140 could not cleave in 20 M formamide, the 5' and 3' termini of Ag180 were truncated and produced Ag5-84, which cleaved to 100% at 37 degrees C in less than 0.25 min. Internal deletions of the Stem IV region resulted in Ag5-73, still capable of efficient cleavage, although with a lessened stability in formamide. A trans-cleaving enzyme-substrate pair was finally derived from this RNA, and it consisted of a 67-nucleotide enzyme that cleaved a 13-nucleotide RNA substrate.

The human nucleolar p120 protein is a proliferation-associated antigen which is expressed in G1 and peaks during the early S phase of the cell cycle. Overexpression of the human p120 protein caused the transformation of NIH 3T3 cells and expression of an antisense p120 construct inhibited the growth of NIH 3T3 cells (Perlaky et al., Cancer Res., 52:428-436, 1992). The middle region of the antisense p120 RNA was found to be almost as inhibitory as the full length antisense construct but the 5' and 3' antisense portions did not affect NIH 3T3 cell proliferation. After the mouse p120 complementary DNA was cloned and sequenced, comparison with the human p120 complementary DNA showed a striking conservation of 85% of the nucleotide sequence and 96% of the amino acid sequence. The two ends of the p120 molecule had less homology in their nucleotide and amino acid sequences. Based on this homology, the observed inhibitory effects of the middle portion of antisense human p120 RNA may be related to suppression of mouse p120 expression by RNA:RNA duplex formation. The high evolutionary conservation of the middle region suggests it has a critical role for the function of this protein.

Antisense oligonucleotides are an attractive potential alternative to conventional drugs as antiviral agents. A major advantage is the relatively simple rational design of oligonucleotides which should bind only to specific nucleic acid sequences, compared with conventional drugs which are frequently targeted against sites of unknown structure in proteins. Progress to date provides hope for the development of a new class of antiviral chemotherapeutics based on antisense oligonucleotides.

RNA enzymes or ribozymes are receiving considerable attention for their potential use as highly specific inhibitors of gene expression. From the basic science perspective, the mechanisms by which ribozymes catalyze site-specific cleavage (and in some cases ligation) reactions provide exciting and active areas of scientific investigation. The most recent developments in our understanding of the molecular mechanisms of catalysis, as well as in vivo applications of ribozymes, are highlighted.

To understand the role of individual genes in regulating biological processes, one must be able to interfere specifically with either their expression or function. While monoclonal antibodies have proven very useful in studying cell surface proteins, the specific inhibition of intracellular proteins in viable cells is a much more difficult problem. The goal of antisense technology is to develop small oligonucleotides, plasmids, or retroviral vectors which can be readily introduced into living cells in order to inhibit specific gene expression. In this review, we briefly describe the principles of antisense usage, including problems of cellular uptake and intracellular distribution, mechanism of antisense action, and the properties of various oligonucleotide derivatives. In addition we present several examples of the biological effects of antisense administration used to study the role of specific genes in the regulation of cell growth and differentiation.

Efforts have been made to improve the biological stability of phosphodiester (PO) oligonucleotides by the addition of various modifications to either the 3', 5' or both the 3' and 5' ends of an oligonucleotide. ISIS 1080, a phosphorothioate (PS) 21-mer oligonucleotide complementary to the internal AUG codon of UL13 mRNA in HSV-1, reduces the infectious yield of HSV-1 in HeLa cells to 9.0% +/- 11%. PO analogs of ISIS 1080 containing three PS linkages placed on the 3' (ISIS 1365), 5' (ISIS 1370), both the 3' and 5' (ISIS 1364) ends or with four linkages in the middle (ISIS 1400) demonstrated reduced antiviral efficacy compared to fully PS ISIS 1080. Thermal denaturation profiles demonstrated that these oligonucleotides hybridized to complementary DNA or RNA with equivalent binding affinities. All were able to support E. coli RNAse H cleavage of the HSV mRNA to which they were targeted. The stability of the congeners in cell culture medium containing 10% fetal calf serum (FCS), HeLa cytosolic extract, HeLa nuclear extract and in intact HeLa cells revealed that ISIS 1080 was most resistant to nucleolytic digestion through 48 hours. Partial PS oligonucleotides exhibited increased degradation compared to the fully thioated oligonucleotide by exonuclease activity in FCS and endonuclease activity in cell extracts or intact cells. Thus, the reduced efficacy of partial compared to fully PS oligonucleotides against HSV-1 in HeLa cells may result from increased degradation of the mixed PO/PS oligonucleotides.

RNA duplexes containing the modified base 2-amino-adenine in place of adenine are stabilized through the formation of three hydrogen bonds in 2-amino A.U base pairs. Antisense 2'-O-alkyloligoribonucleotide probes incorporating 2-aminoadenosine are thus able to efficiently affinity select RNP particles which are otherwise inaccessible. This has allowed the efficient and specific depletion of U5 snRNP from HeLa cell nuclear splicing extracts. U5 snRNP is shown to be essential for spliceosome assembly and for both steps of pre-mRNA splicing. The absence of U5 snRNP prevents the stable association of U4/U6 but not U1 and U2 snRNPs with pre-mRNA.

Natural antisense RNAs have stem-loop (hairpin) secondary structures that are important for their function. The sar antisense RNA of phage P22 is unusual: the 3' half of the molecule forms an extensive stem-loop, but potential structures for the 5' half are not predicted to be thermodynamically stable. We devised a novel method to determine the secondary structure of sar RNA by examining the electrophoretic mobility on non-denaturing gels of numerous sar mutants. The results show that the wild-type RNA forms a 5' stem-loop that enhances electrophoretic mobility. All mutations that disrupt the stem of this hairpin decrease mobility of the RNA. In contrast, mutations that change the sequence of the stem without disrupting it (e.g. change G.U to A.U) do not affect mobility. Nearly all mutations in single-stranded regions of the structure also have no effect on mobility. Confirmation of the proposed 5' stem-loop was obtained by constructing and analyzing compensatory double mutants. Combinations of mutations that restore a base-pair of the stem also restore mobility. The genetic phenotypes of sar mutants confirm that the proposed secondary structure is correct and is essential for optimal activity of the antisense RNA in vivo.