Schistosome Satellite DNA Encodes Active Hammerhead Ribozymes

Abstract

Using a computer program designed to search for RNA structural motifs in sequence databases, we have found a hammerhead ribozyme domain encoded in the Smα repetitive DNA of Schistosoma mansoni. Transcripts of these repeats are expressed as long multimeric precursor RNAs that cleave in vitro and in vivo into unit-length fragments. This RNA domain is able to engage in both cis and trans cleavage typical of the hammerhead ribozyme. Further computer analysis of S. mansoni DNA identified a potential trans cleavage site in the gene coding for a synaptobrevin-like protein, and RNA transcribed from this gene was efficiently cleaved by the Smα ribozyme in vitro. Similar families of repeats containing the hammerhead domain were found in the closely related Schistosoma haematobium and Schistosomatium douthitti species but were not present in Schistosoma japonicum or Heterobilharzia americana, suggesting that the hammerhead domain was not acquired from a common schistosome ancestor.

Schistosomes are a family of digenetic trematodes that parasitize many animal species; three members of this group, Schistosoma mansoni, Schistosoma haematobium, and Schistosoma japonicum, infect over 200 million people worldwide. These blood flukes possess a complex life cycle initiated by the release of eggs from a human host. Fluke eggs produce larvae, called miracidia, which infect snails, a secondary host. Snails in turn shed cercariae, another larval form, which are able to penetrate human skin, transform into schistosomula, and, after a complex migration and differentiation process, develop into sexual adults. Adults produce eggs to complete the cycle. Hopefully, the study of the genomic structure of these species could provide key information for the more effective control of these devastating parasites.

Most eukaryotic genomes contain families of interspersed repetitive DNA called SINEs (37), the sequences of which are generally related to tRNAs or 7SL RNA (1, 3, 10, 41). Their repetitive nature is thought to be due to an amplification process involving reverse transcription of RNA transcripts which are subsequently integrated into host DNA (21, 40). The species S. mansoni contains a family of SINEs, the 335-bp Smα repeats, which occur over 10,000 times in the haploid genome. Many copies of this repeat are clustered on the W female chromosome, while others are dispersed throughout the genome (38). In spite of this ability to amplify themselves, no function has been ascribed to SINEs.

Transcripts of the highly conserved family of satellite DNAs (Sat2) found in the newt do, however, possess a self-processing activity typical of the hammerhead domain found in plant viroids and their satellite RNAs (5, 7, 13, 39). Tandem arrays of the Sat2 repeats are dispersed throughout the genome of Notophthalmus viridescens and other newt species. Their transcripts have tissue-specific 5′ ends, suggesting that transcription and/or self-cleavage are regulated in vivo (14). Although suggestive, no cellular role has yet been assigned to these self-cleaving transcripts (17). We show here that the Smα repeats from S. mansoni and their counterparts in S. haematobium and Schistosomatium douthitti contain a hammerhead catalytic domain having both cis and trans cleavage properties. These observations raise the possibility that the presence of self-cleaving domains in repetitive DNA is more than coincidental.

MATERIALS AND METHODS

Searching the database for hammerhead ribozymes.

The program RNAMOT was used to search for hammerhead ribozyme domains in GenBank, release April 1996 (23). The format of the descriptor used by the program to accomplish this task is illustrated in Fig. 1.

FIG. 1.

Searching the GenBank database for hammerhead ribozyme RNA domains. (Top) The secondary structure of the hammerhead domain with the conserved nucleotides indicated. The part for which the descriptor was written is shown in boldface. (Bottom) The descriptor used in the program RNAMOT (23) was composed of the following features: s1 H1 s2 H1 s3. The s1 feature scans for a 12-nucleotide (nt) sequence of the form NNNNNCUGANGA, the first 5 nucleotides of which are represented by N (any nucleotide) and which includes the required sequence CUGANGA, which is part of the conserved catalytic core. The feature H1, s2, H1 corresponds to the helix II region closed by a loop of 4 to 10 nucleotides (s2). The numbers 4:4 refer to a helix of 4 bp having zero mismatches but allowing the wobble GU. In this helix, a GC pair is required at the base. The s3 feature requires an 8-nucleotide sequence of the form GAAASNNN, where S represents C or G. This feature contains the remaining part of the conserved catalytic core and the variable nucleotides of the 3′ recognition helix. In order to catalyze a cleavage reaction, RNA defined by this descriptor must recognize a substrate RNA by base pairing. The sequence of a possible substrate is represented, and the arrow indicates the cleavage site. The numbering system is that of Hertel et al. (19).

Organisms.

The Puerto Rican strain of S. mansoni was obtained from the Institute of Parasitology of McGill University and was maintained by being cycled through Biomphalaria glabrata (Puerto Rican strain) and female, CD-1 outbred mice (Charles River, St. Constant, Québec, Canada). S. mansoni adult worms were obtained by perfusion of mice infected 7 weeks previously with 150 cercariae (29). They were washed in phosphate-buffered saline at pH 7.2, snap frozen in liquid nitrogen, and stored at −80°C. Schistosomula of S. mansoni were prepared by artificial transformation of cercariae (29), washed in phosphate-buffered saline, pelleted by centrifugation, snap frozen in liquid nitrogen, and stored at −80°C.

Frozen specimens of Schistosomatium douthitti and ethanol-preserved Heterobilharzia americana were kindly provided by Scott Snyder at the University of New Mexico. Frozen specimens of adult S. japonicum and S. haematobium were provided by a National Institute of Allergy and Infectious Diseases supply contract (AI 55270).

Identification and amplification of Smα repeats in different organisms.

DNA was purified from S. mansoni, S. haematobium, S. japonicum, Schistosomatium douthitti, and H. americana as follows. Frozen worms were suspended in 5 mM Tris-HCl (pH 8)–100 mM EDTA–0.5% sodium dodecyl sulfate (SDS). Proteinase K was added to a final concentration of 50 μg/ml, and the mixture was incubated for 3 h at 60°C. Worm lysates were then extracted once with phenol and then with phenol-chloroform (1:1). DNA in the aqueous phase was obtained by ethanol precipitation. To obtain DNA from B. glabrata, the same procedure was used, but in this case, the frozen snails were first ground in liquid nitrogen to destroy the shell. Smα repeats were amplified from the above DNA preparations with two sets of primers. Primers 1A and 1B (see Fig. 2A) are 5′CCCATCGCACAAGCAAGTGG3′ and 5′CACTTAGTATTGTTTGTTTGAATC3′, respectively. Primer 1C, used to amplify Schistosomatium douthitti satellite DNA, is 5′TATAGGTTTTAGTGTCATTG3′. Primers 2A and 2B are 5′GACGCGCGTTTCGTCCTATT3′ and 5′CTGGATTCCACTGCTATCCA3′, respectively. PCRs were carried out with either Vent DNA polymerase (New England Biolabs [NEB]) or Taq DNA polymerase (Pharmacia) with the buffers supplied by the manufacturers. PCR conditions were as follows: 50 pmol of the primers, 200 μM (each) deoxynucleoside triphosphates (dNTPs), and 1 U of polymerase in 1× Vent DNA polymerase or Taq DNA polymerase buffer. PCR cycles were for 30 or 60 s at 94°C; 30 or 60 s at 45, 50, 55, or 60°C; and 30 or 60 s at 72°C. Amplified bands were cloned into pBLUESCRIPT (Stratagene) and sequenced by existing procedures (35).

FIG. 2.

A family of repetitive sequences coding for self-cleaving transcripts in S. mansoni. (A) General organization of Smα repeats into three regions: the 5′ region, which has similarity to tRNA and contains boxes A and B, required for transcription by RNA polymerase III; the middle region, encompassing the hammerhead domain (HH); and the 3′ region. Primers used for the experiments reported here are indicated as 1A, 2A, 1B, and 2B, and their positions on the map correspond to the parts of the sequence to which they are complementary. (B) Alignment of the tRNA and hammerhead (HH) ribozyme domains of 18 different Smα clones. The tRNA sequence is that of the serine 3 tRNA from the rat (accession no. K00371). In the tRNA domain, box A and B nucleotides are indicated in boldface. In the hammerhead domain, nucleotides essential for catalysis are indicated in boldface. (C) Summary of the hammerhead domain sequences found in 18 different Smα clones. The cleavage site is denoted by cs. Substitutions found in different isolates are denoted by arrows directed out of the structure, insertions are denoted by arrows directed toward the structure, and deletions are denoted by Δ.

In vitro transcription and cleavage kinetics.

Individual clones were PCR amplified as described above with oligonucleotide 1B and an oligonucleotide containing the sequence of the T7 RNA polymerase promoter 5′ to the sequence of oligonucleotide 1A, 5′TAATACGACTCACTATAGGCCCATCGCACAAGCAAGTGG3′. Transcription reactions with T7 RNA polymerase (NEB) were as described previously (28). The reaction mixtures contained 40 mM Tris-HCl (pH 8.0) at 37°C; 12 mM MgCl₂; 5 mM dithiothreitol; 2 mM spermidine- (HCl)₃; 25 mM NaCl; 1 mM (each) ATP, CTP, and GTP and 0.5 mM UTP; 10 μCi of [α-³²P]UTP (3,000 Ci/mmol); 1 μM DNA template; 40 U of RNasin (Pharmacia); and 100 U of T7 RNA polymerase (NEB). Reaction mixtures were incubated for 2 to 4 h at 30 or 37°C. To determine the cleavage rate in cis, full-length transcripts from Smα or Sdα templates were gel purified, eluted, phenol extracted, and recovered by ethanol precipitation. Cleavage reactions were carried out in 40 mM Tris-HCl (pH 8.0)–10 mM MgCl₂–1 μM RNA at the times and temperatures indicated in the figures. Reactions were terminated by adding an equal volume of 95% (vol/vol) formamide–0.1% xylene cyanol–0.1% bromophenol blue–10 mM EDTA. Products were heated at 95°C for 1 min, analyzed by electrophoresis in 6% polyacrylamide–8 M urea gels, and quantified by densitometry of the corresponding autoradiogram.

The synaptobrevin-like gene DNA template was obtained from S. mansoni DNA by PCR amplification with Vent DNA polymerase (NEB) and the primers 5′CTGTCAGAAAACATAGATAG3′ and 5′AAGCTTCATAAAAATATTTA3′. PCR cycles were 1 min at 94°C, 1 min at 55°C, and 1 min at 72°C. The PCR products were cloned into pBLUESCRIPT and sequenced. The 5′ cleavage product of the Smα transcript was obtained after in vitro transcription and gel purification as described previously. The synaptobrevin-like protein RNA fragment was also obtained by in vitro transcription from the cloned PCR fragment described above, after insertion into pBLUESCRIPT and digestion with XbaI. The full-length products of both the ribozyme and the substrate were gel purified and quantified by UV spectroscopy. The rate of the trans reaction between the 5′ cleavage product of the cis reaction and the synaptobrevin-like protein RNA was determined under single-turnover conditions. Constant amounts of substrate (2 nM) were incubated with increasing amounts of ribozyme (from 2- to 64-fold molar excess) for 4 h. Ribozyme and substrate were mixed in 40 mM Tris-HCl (pH 8.0)–10 mM MgCl₂–1 mM EDTA, heated for 1 min at 95°C, and cooled on ice. Reactions were started by adding 10 mM MgCl₂ and terminated by adding an equal volume of 95% (vol/vol) formamide–0.1% xylene cyanol–0.1% bromophenol blue–10 mM EDTA. The k_cat/K_m values are derived from the equation ln F/t = k_cat/K_m, where F is the fraction of uncleaved substrate at the end of the reaction at time t.

RNA ligation-dependent PCR.

A sample of 1 μg of in vitro-transcribed RNA was incubated in 1× polynucleotide kinase buffer (NEB)–1 mM ATP–10 U of polynucleotide kinase for 30 min at 37°C. The phosphorylated RNA was then ligated to the 14-mer oligoribonucleotide 5′ACGGUCUCACGAGC3′, in 50 mM HEPES (pH 7.5)–10 mM MgCl₂–1 mM ATP–20 mM dithiothreitol–1 μg of RNase-free bovine serum albumin–10% dimethyl sulfoxide–6 U of T4 RNA ligase (NEB) in a final volume of 20 μl. The reaction mixture was incubated overnight at 15°C, the reaction was stopped by heating at 65°C and the products were recovered by ethanol precipitation. Ligated RNA was reverse transcribed in 1× Vent DNA polymerase buffer (NEB) supplemented with 1 mM (each) dNTPs in a volume of 20 μl with 50 pmol of the primer 1B and 50 U of Moloney murine leukemia virus reverse transcriptase (NEB). The reaction was carried out for 2 h at 37°C. The resulting cDNA was then amplified by using primer 1B and the deoxyribonucleotide version of the oligoribonucleotide 14-mer used in the ligation reaction. PCRs were carried out in a final volume of 100 μl with 1 U of Vent DNA polymerase for 30 cycles at 94, 50, and 72°C. The PCR products were cloned and sequenced.

RNA purification, Northern blotting, RT-PCR, and primer extension.

Total RNA was obtained by treatment of frozen worms previously powdered on dry ice with 4 M guanidinium isothiocyanate, followed by phenol chloroform extraction and ethanol precipitation. Then it was treated for 1 h with DNase I (Pharmacia), 1 U/μg of total RNA, in 40 mM Tris-HCl (pH 7.9)–6 mM MgCl₂, at 37°C. For Northern blots, the RNA was fractionated on 1.4% formaldehyde-agarose gels and transferred to Hybond-N nylon membranes (Amersham) in 20× SSC (1× SSC is 0.15 M NaCl, 0.015 M sodium citrate [pH 7.0], and 0.1% SDS). Hybridization was performed by incubation of the membranes with ³²P-labeled probes (T7 Quickprime; Pharmacia Biotech Inc.) at 65°C in 7% SDS–0.25 M Na₂HPO₄ (pH 7.4)–1% bovine serum albumin (Gibco). The membranes were washed twice at 65°C in 2× SSC for 10 min and once in 0.2× SSC for 30 min. For reverse transcriptase PCR (RT-PCR), 1 μg of total RNA was incubated at 30°C for 30 min with 50 pmol of primer 1B in 20 μl of 1× Vent DNA polymerase buffer (NEB) supplemented with 2 mM dNTPs. Then, 40 U of RNase-Guard (Pharmacia) and 54 U of Moloney murine leukemia virus reverse transcriptase was added and the reaction was continued for 1 h at 37°C. The cDNA was amplified by adding primers 1A and 1B (50 pmol of each), 2 U of Vent DNA polymerase, and 1× Vent DNA polymerase buffer to make 100 μl. The PCR was carried out for 30 cycles of 1 min each at 94, 55, and 72°C. The products were resolved by agarose gel electrophoresis. Primer extension with 10 μg of total RNA was under the conditions described elsewhere (35).

Nucleotide sequence accession number.

The nucleotide sequences reported here have been assigned the following GenBank accession numbers: for S. mansoni Smα, GenBank AF036739 to AF036756; for S. haematobium Shα, GenBank AF036389 to AF036398; and for Schistosomatium douthitti Sdα, GenBank AF036399 to AF036404.

RESULTS

Identification and cloning of hammerhead-containing repeats from S. mansoni.

We have searched the DNA sequence bank (GenBank, version April 1996) for putative hammerhead ribozyme domains with a search engine known as RNAMOT (23). This program screens sequences for potential secondary and tertiary structural elements to uncover cryptic RNA motifs undetected by primary sequence analysis. The descriptor used to search for the hammerhead ribozyme is shown in Fig. 1. The descriptor does not include the base-pairing requirements of helices I and III of the consensus hammerhead domain because we wanted to leave open the possibility of finding a domain that might act in trans; that is, the representation of the two helices as single stranded is equivalent to defining them as substrate recognition arms. This search not surprisingly found the known hammerhead domains in plant viroids and their satellite RNAs and in several newt species but unexpectedly found them among the satellite DNA sequences from the human blood fluke S. mansoni as well (accession no. SCMRSLA). The chance occurrence of this domain is one in 10¹³ nucleotides based on the descriptor used in the search. Moreover, immediately downstream of this domain is a region complementary to the substrate recognition arms, fitting the substrate requirements for this ribozyme.

To study this region, PCR primers were designed to bind to the 5′ and 3′ ends of the repetitive unit (primers 1A and 1B [Fig. 2A]), and amplifications with them were carried out on genomic DNA from S. mansoni. Several amplification products, consisting mainly of a 335-bp fragment and multiples thereof, were isolated. The 335-bp fragment band was cloned into pBLUESCRIPT, and 18 independent clones were sequenced. The 18 clones possessed very similar sequences composed of three regions (Fig. 2A): (i) a tRNA-like region at the 5′ terminus containing the RNA polymerase III promoter elements, boxes A and B (Fig. 2B); (ii) the hammerhead domain (Fig. 2B and C); and (iii) a 3′-terminal domain. Six of the 18 clones possessed a hammerhead domain containing all essential nucleotides and base pairing, including two adjacent GC base pairs and an internal loop in helix I known to enhance the catalytic activity of the newt hammerhead. These features fulfill all known criteria for hammerhead activity. The other 12 sequences exhibited a small number of nucleotide changes in both single-stranded and helical regions, likely rendering them catalytically inactive based on in vitro data.

Transcripts from the Smα family of satellite DNA self-cleave in vitro.

Two of the cloned repeat sequences were chosen to characterize the cleavage properties of the hammerhead domain: one corresponded to a canonical hammerhead (Sm1 [Fig. 2B]) and the other contained a G→C change at position 5 (Sm3). In vitro transcripts of the two clones were prepared from the PCR-generated templates containing a T7 promoter with T7 RNA polymerase. As demonstrated in Fig. 3A, the repeat containing the canonical hammerhead domain cleaved during transcription, while the repeat containing the mutated hammerhead did not cleave.

FIG. 3.

In vitro self-cleavage of the transcripts derived from Smα repeats. (A) Self-cleavage during in vitro transcription at 37°C of different Smα-derived templates. Lane 1, transcription products of the Sm1 template. Lane 2, product after shortening of the Sm1 template by the restriction endonuclease ClaI. Lane 3, product after shortening of the Sm1 template by restriction endonuclease NdeI. Lane 4, pattern obtained from the Sm3 template carrying a G5➛C base substitution. In vitro transcription was less efficient on the templates treated with restriction endonucleases. Numbers at right indicate sizes in base pairs. (B) Kinetics of self-cleavage at 30°C in 10 mM magnesium and at pH 8. Gel-purified full-length transcripts were incubated under the cleavage conditions described in Materials and Methods, and the reaction was stopped at the indicated time. The products were resolved on a 6% acrylamide–8 M urea gel. Numbers at right indicate size in base pairs. (C) The intensities of the bands in panel B were measured by densitometry, normalized to the background of degradation, and graphed on a semilog plot to calculate the rate of the reaction. S, concentration of substrate at time t; So, initial substrate concentration.

The fragment of DNA containing the catalytically active domain was then subjected to restriction enzyme digestion prior to transcription in order to help define the catalytic domain. Figure 3A shows that the length of the 3′ cleavage product varied directly with template length, as would be expected if the hammerhead domain was responsible for cleavage. The cleavage site itself was mapped by primer extension analysis (shown in Fig. 4C) and RNA ligation-dependent PCR, both with primer 1B (data not shown). Both techniques identify the C in the sequence ▾CUG at the 5′ end of the 3′ cleavage product. In addition, the 3′ cleavage product could be labeled with radioactive phosphate by T4 polynucleotide kinase in the presence of [γ-³²P]ATP, indicating the presence of a 5′ hydroxyl group. Under the conditions used during in vitro transcription, 58% of the transcripts were cleaved at 37°C and 37% were cleaved at 30°C. Cleavage required Mg²⁺, the optimal concentration of which was 10 mM at pH 8. The kinetics of cleavage, shown in Fig. 3B and C, were determined at 30°C. The k_cat of self-cleavage was 0.30 ± 0.05 min⁻¹. Transcripts of repeats 5, 7, 10, 12, and 20, which also contain consensus hammerhead domains, have cleavage rates between 0.22 and 0.36 min⁻¹.

FIG. 4.

Expression of the Smα family of repetitive DNA in vivo. (A) Northern blot analysis with total RNA from schistosomula (lane 1), adult males (lane 2), adult females (lane 3), and in vitro-transcribed Smα1 repeat (lane 4). The quantity of RNA used in each analysis was judged equivalent by ethidium bromide staining of the gel and by comparison of the intensity of the rRNA bands in each preparation. (B) Reverse transcription-PCR analysis using primers 2A and 2B from Fig. 2A. Total DNA-free RNAs from adult males (lane 1) or adult females (lane 2) were reverse transcribed with primer 2B. The cDNA was amplified by Vent DNA polymerase (see Materials and Methods). In lane 3, a mixture of both female and male RNA preparations was subjected to the same treatments as were the preparations in lanes 1 and 2, except that reverse transcriptase was omitted. Lane 4 is the 123-bp ladder from Gibco. (C) Primer extension analysis of in vivo transcripts from the Smα family of satellite DNA. In vitro-transcribed Sm1 repeat (lane 1) and total RNAs from adult males (lane 2) and adult females (lane 3) were annealed with primer 1B and reverse transcribed with Superscript Kit II from Gibco at 42°C. The products were resolved on a 6% acrylamide–8 M urea gel. Sequencing reactions performed with primer 1B on the Sm1 template were run in the same gel. Numbers to the right of panel A and the left of panels B and C indicate size in base pairs. The arrow on the right indicates the positions of the expected cleavage products.

Smα repeats are expressed in vivo.

The presence and high conservation of the RNA polymerase III promoter elements among different repeats of the Smα family suggested that the repeat region is transcribed in vivo; however, the absence of termination signals raised the issue of the length of such transcription products. We addressed these questions by the analysis of total cellular RNA by Northern blotting, RT-PCR, and primer extension experiments. The autoradiogram obtained from probing the Northern blot of total RNA from both female and male adult schistosomes and schistosomulas with an Sm1 PCR fragment produced the band pattern shown in Fig. 4A. Since RNA quantities are approximately equivalent in each lane, it would appear that these developmental stages express the Smα repeats at virtually the same level. Also, the size of the major band in all cases corresponds to that of the unit repeat. Bands of greater length, corresponding to multiples of unit-length Smα, are also present and are more evident after overexposure of the autoradiogram. These data suggest that the major band arises after hammerhead processing of long multimeric transcripts. However, the preponderance of single-unit lengths in light of the large number (approximately two-thirds) of putatively inactive forms requires that cleavage take place at sites which are in trans or at least distal to the catalytic domain unless only the active repeats are transcribed.

In the RT-PCR protocol, reverse transcription was performed with primer 2B (Fig. 2A), and primer 2A was added for PCR. These primers were chosen because they would amplify either multimeric transcripts or unit-length transcripts derived from hammerhead processing of multimeric transcripts; in contrast, they could not amplify single repeats of the unit as defined in Fig. 2A. RT-PCR results (Fig. 4B) confirmed that Smα repeats are expressed in both female and male adult schistosomes. Sequences of these products were the expected permutated versions of the sequences obtained from genomic DNA with primers 1A and 1B.

The 5′ end of Smα transcripts was also studied to determine whether unit-length transcripts were generated by transcription alone or by transcription followed by cleavage. Normally, the 5′ terminus of the RNA would contain the polymerase III promoter sequence; however, if the terminus of the unit-length transcript is produced by cleavage of long transcripts, then the 5′-terminal sequence should be the same as that produced by in vitro cleavage. By using primer extension with primer 1B, the two possibilities can be distinguished by the lengths of the extension products. The data presented in Fig. 4C confirm that the in vivo product is identical to the in vitro product and, therefore, results from hammerhead cleavage.

trans cleavage of transcripts from the Smα family.

The high proportion of unit-length Smα transcripts despite the preponderance of repeats containing inactive hammerhead domains indicates that simple self-cleavage is not the only processing mechanism to which the Sm transcripts are subjected. Active ribozymes from distant sites would have to recognize and cleave a target sequence from a repeat containing a disabled hammerhead in order to explain the data shown in Fig. 4A. Active transcripts could first self-cleave and then go on to cleave a repeat with an inactive ribozyme either in cis or in trans (Fig. 5A). Note that, because of the polarity of cleavage, a catalytic domain which had been involved in self-cleavage could cleave only upstream in cis, whereas trans cleavage could be accomplished at any site. Consider as well that trans cleavage implies the use of either the I/II format of the hammerhead (8) or the more familiar I/III format. In the I/II format, the catalytic domain would provide the essential CUGACGA sequence and the target would provide the conserved GAAA sequence and the cleavage site GUC (Fig. 5A).

FIG. 5.

The Smα self-cleavage products as a trans-acting ribozyme. (A) A model of trans-acting ribozyme catalysis in the I/II format. The catalytic fragment is produced from self-cleavage of an active Smα repeat, which is shown in boldface. The substrate is a transcript of the Smα family with a disabled hammerhead ribozyme or the precursor mRNA of the synaptobrevin-like protein in the I/II format. The I/III format is shown in Fig. 1. (B) Cleavage kinetics of the trans hammerhead reaction with synaptobrevin-like protein precursor mRNA, obtained by in vitro transcription from a cloned template. Different concentrations of gel-purified 5′ and 3′ cleavage products from the self-cleavage reaction of Smα 1 transcripts (see Materials and Methods) were incubated in the cleavage conditions described in Materials and Methods with in vitro-transcribed synaptobrevin-like protein RNA for 4 h at 37°C. The products were resolved on a 6% acrylamide–8 M urea gel. Numbers at right indicate size in base pairs. (C) The intensity of the bands in panel B was measured by densitometry, normalized to the background of degradation, and graphed on a semilog plot to calculate the rate of the reaction. S/t, ratio of substrate concentration per time unit.

trans cleavage also raises the issue of cleavage of other cellular targets. A BLAST search of the known S. mansoni sequences in GenBank for the predicted substrate sequence produced one such target, the gene coding for a synaptobrevin-like protein (accession no. SMU30291 [Fig. 5A]). The potential target sequence is 97% identical to that of the Smα family and is found in the only known intron for this gene whose entire sequence is not yet available. Since there is no catalytic domain in this target region, it would be provided presumably from a self-cleaved Smα transcript (Fig. 5A).

In order to demonstrate the feasibility of the trans reaction, we cloned a 650-bp fragment of the synaptobrevin-like protein gene from S. mansoni by PCR and prepared substrate RNA from it by in vitro transcription with T7 RNA polymerase. Figure 5B shows that the 650-ribonucleotide fragment was readily cleaved into fragments of 419 and 231 nucleotides in the presence of cleavage products of the Smα repeat. The 5′ product of cleavage alone also produces these bands (data not shown). The cleavage site was finely mapped by RNA ligation-dependent PCR, which permitted sequencing of the entire 3′ cleavage product. This sequence corresponds to that expected from the cleavage suggested in Fig. 5A. The cleavage rate of the synaptobrevin-like protein RNA fragment was measured under single-turnover conditions, where the ribozyme is in large excess over the substrate so that the observed rate of cleavage is independent of product release. The amount of cleaved substrate increased linearly with ribozyme concentration (Fig. 5B and C), reaching 70 to 85% after 4 h of incubation with a 64-fold molar excess of the ribozyme. The catalytic efficiency of this hammerhead reaction is k_cat/K_m = 500 M⁻¹ s⁻¹, which is comparable to the efficiency (10 to 500 M⁻¹ s⁻¹) of artificially engineered hammerhead ribozymes against substrates of similar length (18).

The distribution of hammerhead-containing satellite DNA in the Schistosomatidae family.

Schistosomes have adult forms in different vertebrates and larval stages in various molluscan hosts. The species that parasitize humans have been classically grouped with respect to egg morphology and snail host type into African (S. mansoni and S. haematobium) and Asian (S. japonicum) schistosomes, while the American species, Schistosomatium douthitti and H. americana, parasitize small mammals. The phylogeny of rRNA sequences confirms this morphogeographic classification separating African, American, and Asian species (2).

The presence of satellite DNA coding for the self-cleaving repeats in the human parasite S. mansoni suggested that the distribution of the Smα family among the Schistosomatidae as a function of host type could be of interest. This idea led to the PCR amplification of DNA from the schistosomes listed in Table 1 with two sets of primers: 1A and 1B and 2A and 2B (Fig. 2A). The latter set was useful because of its specificity for the hammerhead ribozyme domain. Results of these PCR amplifications are presented in Table 1. In agreement with the previous phylogenies, Smα repeats were readily amplified with both sets of primers in S. haematobium but not in S. japonicum or H. americana. In addition, the hammerhead-specific primers allowed the amplification of a family of repeats from Schistosomatium douthitti. Amplification of DNA from B. glabrata, the intermediate molluscan host of S. mansoni, as a control was negative.

TABLE 1.

PCR screening for satellite DNA of the α family in different species

Species	Result for primer combinationa
	1A-1B	2A-2B	1A-1C
S. mansoni	+	+	NA
S. haematobium	+	+	NA
Schistomatium douthitti	±	±	+
S. japonicum	−	−	−
H. americana	−	−	−
B. glabrata	−	−	−
Mus musculus	−	−	−

^a±, amplification products were seen with the primer combination but did not have the same size as the α repeats in S. mansoni. −, no amplification products were observed after several PCRs where the annealing temperatures varied from 45 to 55°C and template DNA concentrations varied from 0.05 to 1 μg. NA, not applicable. 

Several repeats from S. haematobium were cloned and sequenced. The Shα repeats (named in the same manner as the Smα repeats in S. mansoni) were 92% identical to the Smα family and had a similar three-domain organization including the polymerase III promoter region and the hammerhead domain; however, unusual variations were found in the hammerhead domain (Fig. 6A). Most of the clones contained a three-A insertion immediately after the putative self-cleavage site, like the Smα repeats Sm4, Sm14, and Sm18 presented in Fig. 2B. In contrast to the Smα repeats, for which most sequence variations were found in the conserved hammerhead core, all Shα sequences had an intact hammerhead ribozyme core; variations were often found in the adjacent helices.

FIG. 6.

Summary of sequences corresponding to the hammerhead domains found in S. haematobium (A) and Schistosomatium douthitti (B). Substitutions found in different isolates are denoted by arrows directed out of the structure; insertions are indicated by arrows directed toward the structure. CS, cleavage site.

The repeats amplified with primers 2A and 2B from Schistosomatium douthitti were also cloned, and the nucleotide sequences of 12 independent clones were 91% identical to the Smα repeats. Comparison of these sequences with those of S. mansoni showed a number of differences which were localized in the region where primer 1B should bind, thereby explaining why this primer in conjunction with primer 1A did not produce the corresponding band. This new sequence, however, was then used to design a new primer, 1C, which allowed amplification of the α satellite DNA from Schistosomatium douthitti without predetermination of the hammerhead sequence by the primers. Sequencing of the 340-bp amplification product revealed a novel hammerhead motif shown in Fig. 6B. Instead of the ubiquitous GUC▾ triplet 5′ to the cleavage site, Sdα repeats possess an AUC▾ triplet. Also, nucleotide substitutions that potentially disrupt the core of the hammerhead ribozyme were found in two of six sequenced clones. The rate of self-cleavage for the active Sdα transcripts was 0.02 min⁻¹, around 10 times lower than the rate calculated for the self-cleaving transcripts from S. mansoni.

DISCUSSION

Comparison of hammerhead domains from repetitive DNAs.

The occurrence of the hammerhead domain in the Smα family of repeated sequences of S. mansoni superficially resembles the case of a hammerhead domain in the Sat2 repeats from the newt (17). Both contain adjacent GC base pairs and a loop in helix 1, which are required for activity in the newt (14, 31). In addition, other characteristics of the newt hammerhead are found in that of the schistosome, such as the spacing between the internal loop in helix I and the cleavage site as well as the identity of several nucleotides in the external loop and the distal portion of helix II (43). Nevertheless, the two occurrences of the hammerhead domain are likely unrelated based on the facts that (i) there is little overall sequence similarity between the satellite DNAs from the two species, (ii) repeats in the newt are transcribed by RNA polymerase II with small nuclear RNA promoter elements (9) while in schistosomes polymerase III promoter elements seems to be implicated, and (iii) kinetic analysis of the hammerhead domain from schistosomes demonstrates its greater catalytic activity compared to that from the newt. Comparison of the secondary structure models for both domains shows that stem III in schistosomes could be more stable, because it contains 3 bp compared with 2 in the newt. Also, the two organisms, newts and schistosomes, are unrelated in terms of their phylogenetic position. We conclude that the hammerhead-containing satellite DNAs in these two species are not evolutionarily related. However, if ribozyme domains are present in many more repetitive sequences than is currently known, then the occurrence of the hammerhead domain in repetitive DNA may not be completely coincidental.

To date, all naturally occurring hammerheads have been found to cleave a GUC▾ site, with two exemptions: GUA▾ in the lucerne transient streak virus (15) and AUA▾ in the satellite RNA from barley yellow dwarf virus (27). Thus, the AUC▾ site found in Schistosomatium douthitti is unique and its presence contrasts with data from previous reports indicating that this site was not suitable for cleavage in the I/III trans format (32). Surprisingly, AUC▾ does seem to be a good cleavage site in the I/II trans format (34), which could be the format for the S. mansoni hammerhead (see below).

Functional implications of trans-acting self-cleaving transcripts.

The apparent lack of function for repeated sequences has given rise to the selfish DNA hypothesis to rationalize their existence and propagation. According to this model, repeated sequences are not useful to the host but are maintained because they have discovered sequence-specific replication and amplification strategies. Elimination of these sequences would thus require improbable multiple deletion events (11, 30). On the other hand, some repeated sequences contain motifs for transcriptional regulation (20, 33) and some regions of human Alu repeats are immutable, suggesting a role for these repeats in the evolution of primates (6). Copies of repetitive DNA, like the Alu family, are thought to arise via an endonuclease-dependent integration of reverse transcripts into genomic DNA (4). In some repeats of S. mansoni and most repeats of S. haematobium, the presence of three adenines after the cleavage site also suggests a retrotranscription origin, since these adenines could be derived from the polyadenylation of the cleaved transcript. We have formulated a model for Smα dispersion in schistosomes (Fig. 7). Transcription of tandem repeats would yield long multimeric transcripts which self-process. Reverse transcription of the processed transcripts permits dispersion of the unit to other sites in the genome. However, incorporation of the processed transcripts at isolated sites is a dead end for expansion, because polymerase III promoter elements are located at the 3′ end of the processed transcripts. There would be no such barrier to propagation of multimeric transcripts. This scenario provides schistosomes with a method of limiting the copy number of repetitive DNA. In the newt, a similar function is suggested by the fact that monomeric transcripts are found predominantly in the ovaries while transcripts in somatic tissues are mostly dimers and larger multimers (14). In this case, cycles of self-amplification would be limited in the germ line. Interestingly, when the Smα repeats were first cloned, Spotila et al. noticed that some members of the family lost the tRNA homology region as predicted in our model (38). They suggested that transcripts from the repeats underwent a processing reaction before reinsertion into the genome.

FIG. 7.

A model for the propagation and function of the α satellite DNA in schistosomes. Tandem repeats or monomeric α sequences are transcribed by RNA polymerase III. The long transcripts are processed by two mechanisms: (i) intrarepeat cleavage and (ii) interrepeat cleavage. The products of self-cleavage reactions then act in trans on other multimeric transcripts of the α family or in transcripts such as the one coding for the synaptobrevin-like protein gene or the OZ.A retroposon. Reintegration in the genome of reverse transcripts from cleaved repeats creates dead ends in the transposition process because these sequences possess the polymerase III promoter at their 3′ end. However, reintegration of nonprocessed multimeric transcripts creates new sites from which further transcription of Smα repetitive elements can occur.

As for the function of the schistosome repeats, we have proposed that hammerhead-containing transcripts could act in trans to cleave other RNAs. This possibility is supported by the fact that a synaptobrevin-like protein mRNA has the required elements of a substrate in either the I/II or the I/III configuration (Fig. 5A). In fact, this target sequence, which is 97% identical to Smα repeats, might have originated from retrotransposition of a Smα fragment. It is of note that several human genes (5% of cDNAs in the GenBank) also contain a segment of an Alu sequence, in their 5′ or 3′ noncoding regions or their coding region (26, 42). A role in RNA processing in newts has also been advanced because the hammerhead-containing transcripts were found in 12S riboprotein particles (25).

Sex chromosome-specific, Smα-related sequences have been isolated in schistosomes by representational difference analysis (12). One of the isolated clones (OZ.A) codes for a female-specific retroposon (accession no. SMU12442). This retroposon is 76% identical to members of the Smα family and therefore could be considered another target of the trans reaction. We have found that the OZ.A retroposon also has sequence similarity with the microcopia element dhMiF2 of Drosophila hydei. An interesting parallel between the Drosophila microcopia retrotransposon and the Smα family is that both are enriched in the heterogametic sexual chromosome (the male chromosome in the fruit fly). microcopia encodes a testis-specific antisense RNA complementary to the sequence of its own reverse transcriptase gene (24). This antisense RNA could be involved in controlling the germ line expression of transposon-encoded proteins much as the trans-acting ribozymes in schistosomes could control propagation of repetitive sequences and transposable elements.

The phylogenesis of self-cleaving repeats in schistosomes.

The distribution of hammerhead-containing repeats in members of the Schistosomatidae is limited and does not resemble the distribution produced via a common ancestor. Horizontal transmission between organisms in the same host, however, may be possible. In the laboratory, cross-mating can readily be accomplished between S. haematobium and S. mansoni when they share a hamster host (22). These two species also coparasitize their human hosts in Africa. In a locality of Bahia, Brazil, 47% of wild rodents are infected with S. mansoni (36), a favorable situation for interspecies crosses between S. mansoni and rodent schistosomes like Schistosomatium douthitti.

Concluding remarks.

The function of genes is usually predicted by first translating open reading frames in the DNA and then using the predicted protein sequence to find homologs in the protein database. This paradigm dominates the field of functional genomics emerging from genome sequencing efforts. One of the shortcomings of this strategy is that the functionality of RNA sequences is completely ignored. The fact that RNAs likely anteceded proteins in biological evolution (16) suggests that searching databases for RNA domains could provide novel insights into the biology of organisms. The unexpected finding of an active hammerhead domain in schistosome repetitive DNA reported here testifies to the potential of the genomic study of RNA, that is, ribonomics. Perhaps when a more detailed knowledge of functional RNA motifs and domains has been achieved, a wide variety of functions may be found encoded in what previously had been considered junk DNA.