Efficient trans-cleavage by the Schistosoma mansoni SMα1 hammerhead ribozyme in the extreme thermophile Thermus thermophilus

Abstract

The catalytic hammerhead structure has been found in association with repetitive DNA from several animals, including salamanders, crickets and schistosomes, and functions to process in cis the long multimer transcripts into monomer RNA in vivo. The cellular role of these repetitive elements and their transcripts is unknown. Moreover, none of these natural hammerheads have been shown to trans-cleave a host mRNA in vivo. We analyzed the cis- and trans-cleavage properties of the hammerhead ribozyme associated with the SMα DNA family from the human parasite Schistosoma mansoni. The efficiency of trans-cleavage of a target RNA in vitro was affected mainly by both the temperature-dependent chemical step and the ribozyme–product dissociation step. The optimal temperature for trans-cleavage was 70°C. This result was confirmed when both the SMα1 ribozyme and the target RNA were expressed in the extreme thermophile Thermus thermophilus. Moreover, SMα1 RNA showed a remarkable thermostability, equal or superior to that of the most stable RNAs in this species, suggesting that SMα1 RNA has been selected for stability. Computer analysis predicts that the monomer and multimer transcripts fold into highly compact secondary structures, which may explain their exceptional stability in vivo.

INTRODUCTION

The hammerhead ribozyme is a type of natural catalytic RNA structure that was originally discovered within the satellite RNA of a plant pathogenic virus (1). Since then, its distribution has been seen to extend to other satellite RNAs, viroids and virusoids, where it participates in the processing of multimers produced during rolling circle replication (2). Surprisingly, this catalytic structure has been found associated with actively transcribed repetitive DNA from distantly related animals such as salamanders, schistosomes and crickets (3–5). Apart from the highly conserved catalytic core domain, these three repetitive DNA families are unrelated to each other and also differ in their transcription mechanisms. The Sat2 repetitive DNA from salamanders and the SINE SMα from schistosomes are transcribed by RNA polymerases II and III, respectively, but little is known about the satellite pDo500 family from cave crickets (4–6). Intriguingly, in the three known animal cases in which hammerhead structures are associated with transcribed repetitive DNA, the conservation of each DNA family in closely related species is notable (4–8), suggesting that each of these known genetic elements was derived vertically rather than by horizontal transfer.

The possible role or function of repetitive DNA in cells remains largely speculative. Active transcription of these repetitive elements and the accumulation of monomer-sized RNA resulting from hammerhead-mediated processing of multimer transcripts appears to be important evidence that the hammerhead ribozyme–repetitive DNA association is under strong selective pressure, leading to the speculation that these elements may play a cellular role or function (5) or that they may confer an evolutionary advantage to these species. It is thought that all natural hammerhead structures serve only to process long multimer transcripts in cis into monomer RNA, whether they arise from rolling circle replication or from transcribed tandemly repeated DNA (5). However, it is still unclear whether these catalytic structures can trans-cleave other host RNAs in vivo.

For that reason, we considered it pertinent to analyze the cis- and trans-cleavage properties of a natural hammerhead ribozyme associated with transcribed repetitive SMα DNA from the human parasite Schistosoma mansoni (4), because it has evolved within a cellular context, exposed to the constraints that therapeutic ribozymes may encounter. Our aim, therefore, was to determine whether the SMα1 hammerhead is able to trans-cleave an RNA target in vivo and to determine which rate limiting factors have the most important effect on the efficiency of the trans-cleavage reaction. We demonstrate here that the SMα1 hammerhead catalyzes chemical transesterification at exceptionally high temperatures, higher than those so far reported for any hammerhead ribozyme. Moreover, this is the first report of ribozyme activity in the extreme thermophile Thermus thermophilus and we provide evidence of efficient in vivo trans-cleavage of an RNA target.

MATERIALS AND METHODS

In vitro transcription and trans-cleavage kinetics

A synaptobrevin genomic fragment (GenBank accession no. U30291), containing a cleavable sequence within its intron, and the SMα1 monomer DNA (GenBank accession no. AF036739) were used as the target and hammerhead ribozyme, respectively (4). Plasmids containing the synaptobrevin fragment (651 bp) or the SMα1 monomer DNA (316 bp) were linearized with appropriate restriction enzymes prior to in vitro transcription as described elsewhere (5). The synaptobrevin target RNA (986 nt) included a 5′ leader (73 nt) and a 3′ tail (262 nt) from vector sequences (pBluescript KS+; Stratagene). The SMα1 RNA self-cleaves, producing the catalytically active 5′ product Pr1 (218 nt) and the inactive 3′ product Pr2 (127 nt). The [α-³²P]UTP-radiolabeled synaptobrevin RNA target and the SMα1 product Pr1 were gel purified as described (5). The rate of trans-cleavage activity was measured under single turnover conditions (excess ribozyme) at temperatures ranging from 0 to 50°C. Above 50°C the cleavage rate became too fast to be measured precisely by manual pipetting, thus data were obtained under multiple turnover conditions (excess target). All incubations were done on a thermal cycler (Perkin-Elmer) using thin walled PCR quality tubes. Typically, the gel-purified SMα1 Pr1 RNA and radiolabeled synaptobrevin RNA were mixed in 10 mM Tris–HCl pH 8.0, denatured at 95°C for 2 min, then cooled to 70°C over 2 min followed by a slow cooling step (3°C min^–1) to 30°C and kept at this temperature for 10 min. The mixture was then equilibrated at the temperature of interest for 5 min and trans-cleavage reactions were initiated by adding 10 mM MgCl₂ at the same temperature. Aliquots were taken at appropriate intervals and the reaction stopped immediately with 4 vol of stop solution (90% formamide, 30 mM EDTA, 1 M urea, 0.1% xylene cyanol and 0.1% bromophenol blue). After separation on 6% polyacrylamide–TBE gels containing 8 M urea, the intact target RNA and cleavage products were cut out, counted by liquid scintillation and k₂ or k_obs values were calculated as described elsewhere (5).

Plasmid constructs and in vivo trans-cleavage

The P_nar promoter, inducible by nitrate under anaerobic conditions, was chosen to control at our will expression of the SMα1 hammerhead ribozyme in T.thermophilus HB27::nar, a plasmid-free strain carrying the nitrate reductase-encoding cluster that allows its anaerobic growth in the presence of nitrate (9). Synaptobrevin expression was driven by the constitutive slpA gene promoter (10). Briefly, the P_nar promoter was ligated either to full-length SMα1 to generate P_nar–SMα1(WT) or to a shorter version in which 94 bp from the leading 5′ end were removed by EcoRV digestion (see Fig. 1A) yielding P_nar–SMα1(ΔEcoRV). The synaptobrevin fragment was ligated to each of two versions of the slpA promoter: the P_m promoter version encompassing the region from –74 to +2 and the longer P_s version (–74 to +90) (11). Four combinations of the transcription units with the promoters ligated in opposite orientation were cloned into the Escherichia coli–Thermus shuttle vector pMK18 (12): pMK18-1 [P_nar–SMα1(WT) + P_m–synaptobrevin], pMK18-2 [P_nar–SMα1(WT) + P_s–synaptobrevin], pMK18-3 [P_nar–SMα1(ΔEcoRV) + P_m–synaptobrevin] and pMK18-4 [P_nar–SMα1(ΔEcoRV) + P_s–synaptobrevin]. Recombinant T.thermophilus HB27::nar colonies harboring plasmids were grown aerobically at 70°C on rich medium (12) with kanamycin (30 µg/ml) to an OD₅₅₀ of 0.5. To induce P_nar-driven expression of the SMα1 RNA, 40 mM KNO₃ was added and shaking was stopped to maintain microaerophilic conditions, keeping the incubation at 70°C for 2 h. Aliquots of 40 ml of the NO₃-induced culture were transferred to flasks pre-warmed at 50, 60, 70 and 80°C, with rifampicin (200 µg/ml) to inhibit the bacterial RNA polymerase. Ten milliliter aliquots from these flasks were taken at 0, 5, 10 and 15 min incubation and frozen immediately to stop ribozyme activity. Total RNA was isolated under conditions that inhibit ribozyme activity using the FastRNA Kit–Blue (BIO101, CA). Both the intact synaptobrevin RNA target and the 5′ cleavage product were detected by northern hybridization with oligonucleotide probe O-SYN1 (5′-GTGAAATTAATAATGG-3′). The oligonucleotide probes O-RIB1 (5′-ATGTACCTGCATCTCA-3′) and O-RIB2 (5′-CTCGATATAGCCTTGA-3′) were used to detect intact SMα1 RNA and self-cleavage products. Labeled oligonucleotides (Gene Images 3′-Oligolabeling Module Kit) were hybridized and detected with an ECL detection kit (Amersham). Adult S.mansoni worms were collected from infected mice and total RNA was extracted as described elsewhere (13). Typically, 20–40 µg total RNA were separated on agarose–formaldehyde gels (14) or on 7% polyacrylamide–TBE gels containing 8 M urea, transferred to nylon membranes (Hybond-N⁺; Amersham Pharmacia Biotech) and hybridized with the [α-³²P]dCTP-labeled SMα1 DNA probe in Ultrahyb hybridization buffer (Ambion) following standard protocols (14).

Figure 1.

SMα1 and the substrate RNAs predictably fold into two alternative hammerhead conformations. (A) Hammerhead in the I/III format and predicted secondary structure of the product Pr1 resulting from SMα1 self-cleavage. The open triangle indicates the point at which 94 nt from the 5′-leader are deleted in ΔEcoRV. (B) Hammerhead in the I/II format. The cleavage site (CS) on the RNA target is indicated by an arrow.

RESULTS

Efficient trans-cleavage by the hammerhead ribozyme SMα1 at high temperatures in vitro

The effect of incubation temperature on trans-cleavage of a synaptobrevin genomic sequence containing a cleavable site within an intron was analyzed. Two alternative hammerhead structures are predicted to form when SMα1 RNA hybridizes to the synaptobrevin RNA target, corresponding to hammerheads in the I/III and in I/II formats (Fig. 1A and B, respectively). Under single turnover conditions (excess ribozyme), increasing the temperature from 0 to 50°C dramatically stimulated the rate of cleavage (k₂) (∼400-fold) (Fig. 2, circles). The calculated energy of activation for cleavage, 26 kcal mol^–1, in the temperature range 20–50°C is very much in agreement with that of well-behaved hammerhead duplex systems consisting of short ribonucleotides and suggests that ligation of products is essentially negligible (15–17). The k₂ data in the Arrhenius plot fit a straight line well (r² = 0.992), even though a slight curvature at 40–50°C is noticed (Fig. 2, circles). Curvatures like these have been reported in several cases and interpreted to be the result of conformational changes in the ground state structure by temperature (15,17,18). Apart from this consideration, the results clearly indicate that the chemical step is rate limiting over the temperature range tested. Due to the rapidity of the reaction above 50°C, excess target RNA (multiple turnover) was used to measure reaction rates (k_obs). Under these conditions the stimulation of cleavage by temperature was confirmed, with the highest k_obs value obtained at 70°C (1.2 min^–1) (Fig. 2, squares). The multiple turnover experiment established that the ribozyme–product dissociation step is also rate limiting, since the observed cleavage rate under multiple turnover conditions at 50°C (0.03 min^–1) was much lower (14-fold) than that under single turnover conditions (0.42 min^–1). Catalytically inactive structures probably coexist in the mixture and multiple turnover rates may be affected through parasite intra- and/or intermolecular competition, especially at lower temperatures. The structure in the I/III format (Fig. 1A) is thermodynamically more stable at all incubation temperatures (e.g. slower ribozyme–product dissociation) than that in the I/II format (Fig. 1B) because of the greater number of complementary base pairs in the duplex RNA forming stem III. For example, the calculated free energy (19) for helix III (Fig. 1A) is –10.3 kcal mol^–1 at temperatures below 50°C and shifts to –6.3 and –0.3 kcal mol^–1 at 70 and 100°C, respectively. In contrast, the short hairpin III (Fig. 1B) is highly unstable, with +0.5 kcal mol^–1 at 37°C and +1.2 kcal mol^–1 at 50°C. Most probably, both conformers coexist in the mixture, but it is reasonable to expect that at lower temperatures tight binding of the ribozyme and target in the I/III format can reduce multiple turnover. Therefore, the robust increase in k_obs with temperature increase is interpreted as the result of both the temperature-dependent chemical step and the ribozyme–product dissociation step being accelerated (20,21). Finally, the sharp change in slope above 70°C (Fig. 2) could be attributed to any of several rate limiting steps, such as ribozyme denaturation, slower ribozyme–substrate association or faster dissociation of the ribozyme–substrate duplex than the chemical step (17,21).

Figure 2.

Arrhenius plot of the rate of cleavage (k₂) under single turnover (excess ribozyme, circles) and multiple turnover (k_obs) (excess substrate, squares) conditions at different incubation temperatures in vitro. Both radiolabeled synaptobrevin RNA (986 nt) and unlabeled SMα1 Pr1 RNA (the 155 nt sequence is shown in Fig. 1A) from self-cleavage were gel purified and mixed at 1:40 and 40:1 molar ratios for single and multiple turnover experiments, respectively.

Efficient SMα1 hammerhead ribozyme cis- and trans-cleavage of a target RNA in T.thermophilus

To confirm the above observation that the hammerhead can self-cleave and trans-cleave the RNA target at temperatures unusually high for catalytic RNAs, and non-physiological for schistosomes, we chose to express both the ribozyme (inducible by nitrate and anoxia) and the target (constitutive) in the extreme thermophile T.thermophilus. Bacteria harboring each of the four constructs (see Materials and Methods) were grown at the optimal temperature for this strain (70°C) under inducible and non-inducible conditions for the P_nar promoter, and trans-cleavage was analyzed by northern hybridization. When the ribozyme was not induced, the target RNA remained intact (Fig. 3A, lanes 1, 3, 5 and 7). Upon induction of the P_nar promoter to express the wild-type (WT) (Fig. 3A, lanes 2 and 4) or the 5′ deletion ribozyme (ΔEcoRV) (Fig. 3A, lanes 6 and 8), the target RNA was cleaved and both the intact and 5′ cleavage product were detected. Trans-cleavage was achieved with similar efficiency by both the WT (Fig. 3A, lanes 2 and 4) and ΔEcoRV ribozymes (Fig. 3A, lanes 6 and 8). To estimate the efficiency of trans-cleavage in vivo, both ribozyme (WT) and target RNAs were expressed at 70°C either for 2 h followed by arrest of transcription with rifampicin or, alternatively, under continuous transcription. Under continuous expression of both ribozyme and target RNA, similar and constant amounts of both intact substrate and its 5′ cleavage product were detected (Fig. 3B, lanes 5–7; the ratio of intact substrate to product was 1.2 in lanes 5 and 6 and 1.0 in lane 7), because an equilibrium is established between ribozyme activity and RNA degradation. In contrast, when transcription was arrested with rifampicin, the 5′ product accumulated whereas the intact substrate decreased to much lower levels (Fig. 3B, lanes 1–4). Therefore, disappearance of the intact RNA target is better explained by trans-cleavage rather than faster degradation than its 5′ product. This conclusion is supported by the observation that under continuous expression both intact and cleavage product RNAs showed similar turnovers and their levels remained almost constant (Fig. 3B, lanes 5–7).

Figure 3.

In vivo trans-cleavage of the synaptobrevin RNA target in T.thermophilus at 70°C. (A) Northern blot showing trans-cleavage of synaptobrevin RNA (Syn) and its 5′ product (Ps1) by the wild-type (WT) and the 5′ deletion (ΔEcoRV) ribozymes. Transcription of the ribozyme from the P_nar promoter was induced upon addition of nitrate and reduced aeration (lanes 2, 4, 6 and 8). Lanes 1, 3, 5 and 7 correspond to bacterial cultures in which the ribozyme was not induced. Transcription of the synaptobrevin target was driven by the P_m (lanes 1, 2, 5 and 6) or the P_s (lanes 3, 4, 7 and 8) promoter. (B) Time course of trans-cleavage by the WT SMα1 ribozyme under continuous induction of the P_nar promoter. Rifampicin was added to bacterial cultures at 70°C and samples were analyzed by northern blot after 0, 5, 10 and 15 min incubation (+Rif, lanes 1–4). Parallel cultures were maintained under continuous transcription (–Rif, lanes 5–7). Lane 8 corresponds to a culture in which the ribozyme was not induced.

The effect of incubation temperature on cis- and trans-cleavage was analyzed in vivo. Upon arrest of transcription with rifampicin, trans-cleavage of the RNA target was observed at 50, 60 and 70°C. No cleavage was observed in the absence of induction of the WT or ΔEcoRV ribozyme (Fig. 4A and B, respectively, lanes ni), as expected. Both target and ribozyme RNAs were not detected after 5 min at 80°C (Fig. 4A and B, panels Osyn) because of their very fast degradation at this temperature. Although very little difference was observed at 50 and 60°C (Fig. 4A), trans-cleavage seems to be more efficient at 70°C. The optimal temperature for the ΔEcoRV ribozyme is also 70°C (Fig. 4B), at which temperature 5′ product accumulated concomitant with complete disappearance of intact substrate. The slight difference in cleavage efficiency suggests that the 5′-leading sequence of the WT ribozyme may somehow interfere in the interaction of the hammerhead ribozyme with the target RNA. Moreover, other factors, such as proteins or RNA diffusion, may have slowed down trans-cleavage by the WT ribozyme at 50–60°C. Nevertheless, these results confirm those of the in vitro experiments, in which the highest k_obs was obtained at 70°C.

Figure 4.

Effect of incubation temperature on self- and trans-cleavage by the SMα1 ribozyme in T.thermophilus. Bacterial cultures carrying the pMK18-2 (A) or pMK18-4 plasmid (B) were grown in parallel for 2 h at 70°C expressing the WT (A) or the ΔEcoRV (B) SMα1 ribozyme, before being transferred to pre-warmed flasks with rifampicin and kept at the indicated temperatures for 0, 5, 10 and 15 min. Total RNA samples from these cultures were analyzed by northern blot andhybridized either sequentially or simultaneously with oligonucleotide probes Osyn, Orib1 and Orib2 (see Materials and Methods). Lanes ni correspond to a bacterial culture in which the ribozyme was not induced. When a bacterial culture was incubated for 5 min at 80°C with both the ribozyme and target RNAs under continuous transcription (lane c), the intact RNAs from WT (Rib) and deletion ribozymes (ΔRib) were detected but not the self-processing 5′ (Pr1) and 3′ (Pr2) products or the intact synaptobrevin RNA (Syn).

The use of ribozyme-specific probes on the same northern blots showed that self-cleavage and concomitant production of both 5′ and 3′ fragments occurred at 50–70°C, but not at 80°C (Fig. 4A and B, panels Orib). At 80°C and under continuous transcription the apparent absence of self-processing by both the WT and ΔEcoRV ribozymes suggests that the RNA cannot fold into the active conformation or that the cleavage products are degraded too rapidly to be detected (Fig. 4A and B, middle and lower panels, lanes c). The WT ribozyme self-cleaves with highest efficiency at 70°C (Fig. 4A). At this temperature the intact RNA became undetectable 10 min after transcription arrest, while both the 5′ and 3′ products (Pr1 and Pr2, respectively) accumulated concomitantly. Since the intact RNA is intrinsically more stable than the 3′ product, disappearance of intact RNA is explained by self-processing rather than by degradation. On the other hand, the ΔEcoRV ribozyme showed similar self-cleavage efficiency at 50, 60 and 70°C (Fig. 4B). This confirms our observation of trans-cleavage (Fig. 4A), suggesting that the 5′-leader sequence in the WT ribozyme may interfere either in correct folding of the hammerhead domain or in the interaction with the RNA substrate. This effect is more obvious at 50–60°C than at 70°C, presumably because the higher temperature disrupts interfering interactions. Interestingly, the 5′ RNA product from self-cleaved ΔEcoRV RNA (Fig. 4B, Pr1) has a lower stability and degrades faster than the Pr1 product from the WT ribozyme (Fig. 4A, Pr1). Possibly, its compact secondary structure at the 5′ end may contribute to RNA stability (Fig. 1A). In bacteria the rate limiting step in mRNA degradation is usually cleavage at the 5′ end by an endoribonuclease, followed by exonuclease degradation (22). Moreover, stable secondary structures at the 5′ and 3′ ends block or retard 5′→3′ and 3′→5′ exoribonucleases in both eukaryotes and bacteria (22,23). Indeed, both 5′ (Pr1) and 3′ (Pr2) fragments from self-cleaved WT RNA (Fig. 4A) are equally stable and remain detectable at 70°C for at least 15 min, hinting that this repetitive sequence has evolved for high stability. This notable stability is superior to that of slpA mRNA, which is among the most stable RNAs in T.thermophilus (11).

Can the S.mansoni SMα1 hammerhead ribozyme trans-cleave a host RNA?

The above results suggest that the catalytic SMα1 RNA may not trans-cleave a host mRNA at physiological temperatures for S.mansoni. Indeed, none of the naturally found hammerhead ribozymes have been demonstrated to participate in trans-processing of a host mRNA in vivo (3–6). Since the limited availability of S.mansoni genomic sequences obstructs a search for additional potential RNA targets in vivo, we decided to address this question by first determining the relative abundance of SMα1 RNA in total RNA from adult schistosomes. Northern blot analysis (Fig. 5A) indicates that the SMα1 monomer (330 nt in Fig. 5B) is a very low abundance transcript, representing ∼0.005% of total RNA. This finding, although surprising, is similar to the observation reported for the pDo500 hammerhead ribozyme from Dolichopoda cave crickets (5). It is known that the SMα family is scattered throughout the schistosome genome and represents a significant proportion of the genomic DNA (24,25). One possible explanation is that only a small fraction of the repetitive sequences is actively transcribed or, alternatively, that most of the multimer transcript population is partially processed into monomers. Consistent with the latter possibility, the northern blot showed that SMα multimer transcripts are clearly predominant over the monomer (Fig. 5A). It is also possible that the monomer RNA is rapidly degraded in vivo. Though in T.thermophilus both the WT SMα1 RNA and its self-cleavage fragments showed considerable stability (Fig. 4A and B), in eukaryote cells the RNA may be degraded by one of several alternative mechanisms (26). Although the northern blot in Figure 5B shows some putative degradation products, the monomer RNA is visibly more abundant. To assess the trans-cleavage activity of the SMα1 hammerhead in total RNA from S.mansoni, we incubated radiolabeled synaptobrevin target RNA with total RNA from S.mansoni under optimal in vitro conditions. No trans-cleavage was detected even when increasing the amount of total RNA (Fig. 5C, lanes 2 and 3). Thus, the result confirms the above observation that active SMα1 ribozyme in total RNA samples is present at very low levels and/or the RNA is folded in an inactive conformation.

Figure 5.

SMα1 relative abundance in total RNA from S.mansoni and in vitro trans-cleavage. (A) Total RNA (20 µg) was separated on an agarose–formaldehyde gel along with known amounts of in vitro transcribed SMα1 RNA. Lane 1, S.mansoni adult worms collected from untreated mice; lanes 2 and 3, worms collected from mice treated with PZQ at 40 and 160 mg/kg, respectively. The autoradiograph was deliberately overexposed in order to detect the low abundance monomer RNA (M). The predominant transcripts of variable size above the monomer represent putative multimers transcribed from tandemly repeated SMα DNA with variable numbers of copies or partially processed. (B) Live worms were heat-shocked at 37 (lanes 1 and 3) or 42°C (lanes 2 and 4) for 1 h. Total RNA was isolated and 20 (lanes 1 and 2) or 40 µg (lanes 3 and 4) were separated on 7% polyacrylamide–urea gels along with molecular weight markers and analyzed by northern hybridization. The monomer RNA (M) and four other secondary bands (180, 160, 145 and 130 nt) were detected. These low molecular weight bands may be partially degraded products from monomers or multimers. Although speculative, the band of ∼1.3 kb (undetected in the agarose northern blot) may represent circular SMα1 monomers produced by a host RNA ligase and deserves further study in the future. (C) In vitro trans-cleavage of synaptobrevin RNA. Lane 1, radiolabeled target RNA and SMα1 RNA (40 ng each) were mixed and incubated under optimal cleavage conditions (see Materials and Methods). Lanes 2 and 3, the target RNA was mixed with S.mansoni total RNA (1 and 5 µg, respectively) and incubated as above; lanes 4 and 5, ribozyme and target RNAs were mixed with total RNA (0.15 and 1.5 µg, respectively) and incubated as above.

DISCUSSION

This work reports the functional analysis, both in vitro and in vivo, of a natural hammerhead ribozyme embedded within the repetitive SMα DNA family from the human parasite S.mansoni. To our knowledge, this is the first report demonstrating efficient trans-cleavage of an RNA target by a natural hammerhead ribozyme in vivo in an extreme thermophile. Moreover, the hammerhead structure remains catalytically active, both in cis and in trans, showing highest efficiency at 70°C in vivo and in vitro. Some residual activity at 80°C in vitro was also detected. This contrasts with the reported weak or undetectable activity of catalytic RNAs from extremophiles at physiological temperatures in vitro (27,28). A class of introns unique to Archaea has also been described, but they do not self-splice in vitro (29). Thus, one may speculate that the extreme environmental conditions at which these organisms thrive may not be favorable to sustain catalytic RNAs. However, this work indicates that high temperatures may favor catalysis.

A major unresolved question concerning the hammerhead motif associated with repetitive DNA is whether the catalytic hammerhead has a trans-cleavage role in vivo. If self-cleavage is the only function of the hammerhead in these repetitive DNA transcripts, as suggested (5), then the chemical step should become the crucial limiting step, with product dissociation being unimportant, and so the hammerhead kinetics would correspond to single turnover conditions. On the other hand, if the hammerhead has a trans-cleavage role, it should have evolved for efficient ribozyme–product dissociation at physiological temperatures for schistosomes. Our results point more towards the first possibility. First, the in vitro experiments demonstrate that two major steps in the reaction are rate limiting, namely ribozyme–product dissociation and the temperature-dependent chemical step. By extrapolating from the multiple turnover experiments in vitro, we conclude that in S.mansoni cells multiple turnover is limited or non-existent. This is principally because the ribozyme–product bi-molecule is highly stable at physiological temperatures (e.g. 37°C). This ribozyme–substrate system can fold in either the I/II or I/III conformation (Fig. 1A and B). Although the annealing protocol used for the in vitro experiments allows folding of the ribozyme and target RNA in both conformations, the I/III format may be favored because of its lower free energy, determined by the long hybridizing arms forming helices I and III. Therefore, if this repetitive family has been under selection pressure, one may expect that the present sequence has evolved towards high stability, with self-cleavage being more important than catalytic turnover (e.g. trans-cleavage). Another important argument favoring the hypothesis that these satellite sequences have been selected for stability is the presence of catalytically inactive members of the SMα family bearing mutations within the hammerhead domain, as found in other repetitive satellite DNA families containing this catalytic motif (4,5,7,8).

The in vivo experiments in T.thermophilus also lead to similar conclusions. The removal of 5′ sequences in the ribozyme had an important effect, improving self-cleavage at temperatures <70°C but decreasing the half-life of both intact and product RNAs, thus indicating once again that stability is more important than cleavage. Although the experiments are done at the permissive temperature for this species, cis- and trans-cleavage by the WT ribozyme are less efficient at 50°C than at 60–70°C. This can be explained both by the temperature dependence of the chemical step and the ribozyme–product dissociation step (15–18,20). At the highest incubation temperature tested (80°C), hammerhead denaturation, slower ribozyme–target association and/or faster RNA degradation are assumed to rate limit the reaction.

Nevertheless, it may still be possible that the SMα1 hammerhead could act as a trans-cleaving ribozyme in S.mansoni, provided that the highly stable ribozyme–substrate complex is efficiently dissociated. It is worth noting that RNA helicases can indeed unwind the ribozyme–product hybrids and thus modulate ribozyme activity (30). However, the low abundance of the monomer transcripts and the undetectable ribozyme activity in vitro in total RNA from S.mansoni are strong arguments against the possibility that they may trans-cleave a host RNA. Under in vivo conditions trans-cleavage would approach multiple turnover, but the above discussion predicts limited trans-cleavage, if any.

Self-processing of the long primary transcripts is also partial in S.mansoni cells (Fig. 5A). One explanation is that some members of this repetitive DNA family are inactivated by mutations occurring within the hammerhead core domain (4,24). Computer analysis with Mfold (19) suggests that two adjacent copies of SMα1 RNA within the long primary transcript can hybridize to each other in the antiparallel orientation due to extensive base pair complementation, forming either a double hammerhead structure cleaving each strand in trans (Fig. 6A) or single hammerheads cleaving in cis (Fig. 6B). Similarly, the monomer RNA produced after self-processing of a multimeric transcript predictably folds to itself and as a result the hybridizing arms remain locked, unable to trans-cleave (Fig. 6C). These structures are highly compact, composed of double-stranded stretches interrupted by small loops, and show remarkable symmetry. Interestingly, both these structural features and processing by a double-hammerhead structure are common to satellite RNA, viroids, virusoids and to the transcripts from repetitive pDo500 DNA from Dolichopoda crickets (2,5). These duplex molecules are thermodynamically very stable and thus the hybridizing ‘arms’ of the hammerhead may not be readily accessible for a target RNA at physiological temperatures in S.mansoni.

Figure 6.

Predicted folding and self-processing mechanism of a multimer SMα1 transcript. (A) A double hammerhead structure (indicated by arrows) consisting of two copies hybridized in the antiparallel orientation due to extensive base complementation, so that in this case each strand is trans-cleaved. (B) Another possible conformation with similar compactness but differing in that each SMα1 copy forms a single hammerhead, so that self-cleavage occurs. (C) Predicted folding of a single SMα1 monomer RNA produced by processing of the multimer transcripts in vivo.