RNA-Induced Changes in the Activity of the Endonuclease Encoded by the R2 Retrotransposable Element

Abstract

R2 is a non-long terminal repeat retrotransposable element that inserts itself site specifically in the 28S rRNA genes of arthropods. The 120-kDa protein encoded by R2 has been shown to cleave one strand of the 28S gene at the target site and to use the 3′ hydroxyl group generated from this nick to prime reverse transcription of its own RNA. This reaction has been termed target-primed reverse transcription (TPRT). Cleavage of the second DNA strand can occur in the presence or absence of reverse transcription but requires RNA. In this study, more sensitive in vitro assays have enabled further characterization of these reactions. R2 protein is capable of only a single round of TPRT because, once bound to the target DNA, it does not dissociate at physiological ionic strengths. Analysis of the role of RNA in the DNA cleavage reaction has revealed that the binding of RNA induces the R2 protein to form a multimeric complex. While larger complexes may form, the active component appears to be a dimer based on sedimentation studies and the change in stoichiometry of the cleavage reaction from a 1:1 ratio of protein subunit to target DNA in the absence of RNA to a 2:1 ratio of subunit to DNA target in the presence of RNA. Nonspecific RNA can also induce formation of this RNA-protein (RNP) complex, but the association of the protein with R2 RNA is stronger as revealed by its stability in 0.4 M NaCl. Finally, formation of the RNP complex gives rise to a 150-fold increase in the ability of the R2 endonuclease to find the target site. The specificity of this RNP complex is sufficiently great that it can find the 28S gene target site and conduct the TPRT reaction with total genomic DNA.

Retrotransposable elements can be divided into two distinct classes based on the sequences of their reverse transcriptase (RT) domains (5, 36, 37). The most obvious structural difference between these two classes of elements is the presence or absence of long terminal repeats; thus, they have been called the LTR and non-LTR classes. The structure and retrotransposition mechanism of the LTR class closely resemble those of retroviruses (1, 29, 32). Indeed, certain LTR-retrotransposable elements have transmission properties that qualify them as retroviruses (17, 31). In addition to lacking terminal repeats, members of the non-LTR class of retrotransposable elements do not encode an integrase domain and have no sequences complementary to tRNA for the priming of reverse transcription. The absence of these critical components of the retroviral integration machinery suggests that the mechanism of non-LTR retrotransposition is very different from that of retroviruses (1, 7).

The mechanism of non-LTR retrotransposition has been best characterized for the R2 element of arthropods. This element specifically integrates into a unique site in the 28S rRNA genes of its arthropod host (2, 3, 15). The single 120-kDa protein encoded by the R2 element from Bombyx mori (R2Bm) has been shown to contain endonuclease and RT activities (22, 35). The RT activity differs from that of LTR retrotransposons and retroviruses in that it is capable of using the 3′ hydroxyl group generated by the endonuclease at the DNA target site to prime reverse transcription of the R2 RNA. This critical initial step in the retrotransposition of R2 elements has been termed target primed reverse transcription, or TPRT. Purified R2 protein and RNA templates are capable of conducting the initial steps of the R2 integration reaction in vitro. The R2 protein first cleaves the noncoding strand of the 28S rRNA gene target site (the strand which serves as the template for RNA synthesis). We will hereafter refer to this DNA strand as the primer strand. The RT activity then polymerizes the cDNA directly onto the new DNA 3′ end generated by primer strand cleavage. Cleavage of the second DNA strand (hereafter referred to as the non-primer strand) occurs after reverse transcription. The R2 TPRT reaction is highly specific for the R2 RNA and the 28S target site and does not require complementarity between the RNA template and the DNA target (22). The R2 RNA sequences recognized by the R2 protein correspond to the 250-nucleotide (nt) 3′ untranslated region of the R2 element (23). Similar TPRT-like reactions are believed to be used by most other non-LTR retrotransposons (9, 20, 26, 27, 30).

Based on the sequences of their RT domains, the non-LTR retrotransposable elements have been shown to be more closely related to mitochondrial group II introns than to LTR-retrotransposable elements (7, 37). Strengthening this phylogenetic placement, yeast mitochondrion group II introns, aI1 and aI2, have been shown to insert themselves into unoccupied target DNA sites by a TPRT mechanism. An RNA-protein (RNP) complex containing the intron-encoded protein and intron RNA cleaves the target DNA site and uses the 3′ hydroxyl group of one strand to prime cDNA synthesis (39). While the catalytic subunit for cleavage of the DNA strand used for priming reverse transcription is the intron-encoded protein, the catalyst for cleavage of the non-primer strand is the intron RNA itself. Cleavage of this non-primer strand occurs before primer strand cleavage and is catalyzed by the reverse splicing of the intron RNA into the DNA site (12, 38, 40).

A requirement for RNA in DNA cleavage has also been noted in the case of the R2 TPRT reaction (22). Cleavage of the primer (first) strand occurs in the absence of RNA, but cleavage of the non-primer (second) DNA strand requires RNA. In this study, we have further characterized these cleavage reactions, including studies of the role played by RNA. We show that binding of the RNA by the R2 protein induces the protein to form a higher-ordered complex, probably a dimer, which is required for non-primer strand DNA cleavage. We further show that this RNP complex is considerably more efficient at finding the DNA target site for the TPRT reaction than is the protein alone.

MATERIALS AND METHODS

Protein purification.

R2 protein was purified from Escherichia coli JM109/pR260 as previously described (22), except that the starting volume of the JM109/pR260 culture was increased to 1.2 liters. Before protein elution, the Q-Sepharose column was washed with 8 column volumes and the DNA-cellulose column was washed with 10 column volumes. R2 protein eluted from the DNA cellulose column was concentrated approximately fivefold on a Centricon-50 column (Amico) and dialyzed against a solution of 50% glycerol, 0.4 M NaCl, 25 mM Tris-HCl (pH 7.5) and 1 mM dithiothreitol (DTT) at 4°C. The concentrated R2 protein was stored after dialysis at −20°C. Protein stored in this manner could be used for over 1 month after isolation, with only minor decreases in the stoichiometry of the primer strand cleavage reaction. More significant losses of activity with time were detected in the TPRT reaction and non-primer strand cleavage. R2 protein concentrations were determined by comparing the intensity of Coomassie-brilliant-blue-stained bands with that of known concentrations of bovine serum albumin after sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

Preparation of DNA substrates.

The DNA substrate (target DNA) for the in vitro R2 cleavage and TPRT reactions was a 110-bp fragment of the 28S rRNA gene. This DNA substrate was generated by PCR amplification from clone pB109 (22) with one primer (AATTCAAGCAAGCGCGG) complementary to the 28S sequence 50 bp upstream of the R2 site and a second primer (CTAAGGATCCCGTTAATCCATTCATGCGCG) complementary to the region 60 bp downstream of the R2 site. The PCR was carried out in 50 μl of buffer containing 1 ng of pB109 plasmid, 0.25 μM (each) primer, 50 μCi of [α-³²P]dATP (3,000 Ci/mmol; Dupont-New England Nuclear), 80 μM dATP, 200 μM (each) dCTP, dGTP, and dTTP, and 2.5 U of Taq DNA polymerase (Life Technologies). The PCR products were separated on an 8% native polyacrylamide gel, and the region corresponding to 110 bp was excised from the gel, ground into small pieces, and eluted at room temperature in 10 mM Tris–1 mM EDTA (pH 8) (TE buffer). DNA eluant was separated from the polyacrylamide gel pieces on a Quik-Sep polypropylene column (Isolab), ethanol precipitated, and dissolved in TE buffer.

In vitro synthesis of RNAs.

Construct pBmR2-249A4 was used as template for in vitro transcription of the R2-specific RNA (23). The construct was linearized with XmnI (New England Biolabs), digested with proteinase K, extracted with phenol-chloroform, and ethanol precipitated. Each in vitro transcription reaction mixture contained 1 μg of the linearized template in 40 μl of transcription buffer (Promega) containing 1 mM DTT, 0.4 mM (each) nucleoside triphosphate (NTP), 20 U of RNAguard ribonuclease inhibitor (Pharmacia) and 10 U of T7 RNA polymerase (Promega). Incubations were at 37°C for 1.5 h. The products of the reaction were incubated with 15 U of RNase-free DNase I (Pharmacia) for 40 min at 37°C. The RNA was recovered after phenol extraction by ethanol precipitation and resuspended in 10 μl of water. This synthesized R2 RNA was 283 nt in length and contained the 249-nt 3′ untranslated region of the R2 element and 34 nt of pBSK(+) vector sequence at its 5′ end. Synthesis of a control RNA was identical to that described above, with the uninserted pBS(+) plasmid digested with PvuII. The RNA synthesized from this template was 277 nt in length.

In vitro R2 reactions.

Unless otherwise specified, all cleavage and TPRT reactions were performed in 20-μl volumes containing 50 mM Tris-HCl (pH 8)–200 mM NaCl–10 mM MgCl₂–1 mM DTT. The concentration of R2 protein and the DNA substrate for each reaction is given in the legend for each figure. Approximately 10 pmol (1.0 μg) of in vitro-synthesized R2 RNA or vector RNA was added to those assay mixtures containing RNA, which is 30- to 300-fold in excess of that of the R2 protein in the reaction. The deoxynucleoside triphosphate (dNTP) concentration was 25 μM for all TPRT reactions. All cleavage and TPRT reaction mixtures were incubated at 37°C, and the reactions stopped by the addition of 3 μl of 0.5 M EDTA. The reaction products were extracted with phenol-chloroform or chloroform alone, precipitated by 95% ethanol, and dissolved in 10 μl of TE buffer. For electrophoresis on 8% denaturing polyacrylamide gels, 2-μl aliquots of the reaction products were mixed with 10 μl of formamide buffer (95% formamide–20 mM EDTA), boiled for 5 minutes, and chilled on ice before loading. All electrophoresis was done at room temperature at 500 V in 90 mM Tris–90 mM borate–1 mM EDTA (pH 8). Quantitation of the reaction products was conducted with a PhosphorImager (Molecular Dynamics) after the gel was dried.

Genomic DNA digestion assay.

Genomic DNA was isolated from Drosophila melanogaster W1118 (6). About 1.2 μg of genomic DNA was digested by ClaI at 37°C for 2 h in a total reaction volume of 20 μl. This predigested genomic DNA was then incubated with the R2 protein (with or without RNA) at 37°C for 1 h under the conditions used in the in vitro assays described above. After R2 protein digestion, the DNA samples were fractionated on a 1.5% agarose gel, transferred to nitrocellulose membranes, and probed with a 280-bp labeled segment of the 28S gene starting 74 bp downstream of the R2 insertion site as previously described (15).

Glycerol gradients.

Approximately 240 ng of R2 protein either alone or with 3 μg of RNA in 200 mM NaCl–50 mM Tris-HCl (pH 8)–10 mM MgCl₂–1 mM DTT was loaded onto 5 ml of 10 to 30% glycerol gradients containing either 0.2 or 0.4 M NaCl in 50 mM Tris-HCl (pH 8). Centrifugation with the SW50.1 rotor was for 17 h at 30,000 rpm at 2°C. Fractions were collected from the bottom of the tube, and the presence of R2 protein was detected by assaying 0.1-ml aliquots of each fraction in a 0.4-ml DNA cleavage reaction mixture with the 110-bp labeled DNA substrate. After incubation, the products were phenol-chloroform extracted, ethanol precipitated, resuspended in 3 μl of water plus 10 μl of formamide buffer, and fractionated on 8% denaturing polyacrylamide gels, and the level of cleavage was quantitated with a PhosphorImager. l-Lactic dehydrogenase (123,000 to 131,000 Da), catalase (250,000 Da), and apoferritin (440,000 to 460,000 Da) (Sigma) were used as protein molecular weight standards. These protein standards were individually sedimented on gradients identical to those used for the R2 protein. Protein in each fraction was detected by trichloroacetic acid precipitation, followed by electrophoresis on a sodium dodecyl sulfate–10% polyacrylamide gel and Coomassie brilliant blue staining.

RESULTS

In vitro assays for DNA cleavage and reverse transcription.

Our previous assay for R2 endonuclease and TPRT activities utilized an entire pUC plasmid, containing a 1.1-kb 28S rDNA insertion bearing the R2 target site (22–25). Cleavage and reverse transcription were followed by direct staining of the reaction products with ethidium bromide or by the addition of labeled dNTPs. In these previous studies, the R2 protein was used immediately after purification, as the protein rapidly lost activity upon storage. We have now identified a method that enables the storage of R2 protein in 50% glycerol at −20°C for at least 1 month (see Materials and Methods). The use of this stored protein allows greater uniformity and control over the stoichiometry of the reaction. We have also utilized in this report a new substrate for the in vitro assays. As diagrammed in Fig. 1A, the target DNA substrate is a uniformly labeled 110-bp DNA fragment generated by the addition of [α-³²P]dATP in a PCR amplification (see Materials and Methods). The R2 target site is located 5 bp from the middle of this 110-bp fragment. Because R2 endonuclease cleavage generates a 2-bp 5′ overhang at the target site, cleavage of this DNA gives rise to fragments of lengths 48, 50, 60, and 62 nt that can be resolved in an 8% denaturing polyacrylamide gel. An additional advantage to the use of this ³²P-labeled target DNA is that the increased sensitivity allows a 25-fold reduction in the amount of R2 protein used in each reaction (100 ng reduced to 4 ng). The R2 RNA template used in the assay is the shortest R2 RNA that has been shown to maximally support the TPRT reaction (23). This RNA, referred to as “R2 RNA,” has a 3′ end which corresponds to the precise boundary of the R2 element with the 28S gene, the complete 249-nt 3′ untranslated region of the R2 element from B. mori, and 34 nt of the pBS vector at its 5′ end. Control RNA was a 277-nt RNA derived entirely from the pBS vector, referred to as vector RNA.

FIG. 1.

R2 DNA cleavage and reverse transcription assays. (A) Diagram of the nucleic acid substrates and products of the reactions conducted in panel B. The RNA used as template corresponded to the 3′ untranslated region of the R2 element plus an additional 34 nt of vector sequence. The DNA substrate is a uniformly labeled 110-bp fragment generated by PCR amplification (see Materials and Methods). (B) Autoradiograph of the reaction products separated on a 17-cm 8% denaturing polyacrylamide gel. Reactions were conducted in 50 mM Tris-HCl (pH 8)–10 mM MgCl₂–0.2 M NaCl–1 mM DTT in 20-μl volumes for 1 h at 37°C. Each reaction mixture contained 20 ng (250 fmol) of internally labeled target DNA. Other components in the incubation mixtures are indicated by the plus and minus signs at the bottom of the figure. The mixtures contained 40 ng of R2 protein (300 fmol), 1 μg of RNA (10 pmol), and 25 μM dNTPs. Lane 1, no protein; lane 2, R2 protein alone; lane 3, R2 protein and vector RNA; lane 4, R2 protein, vector RNA, and dNTPs; lane 5, R2 protein and R2 RNA; and lane 6, R2 protein, R2 RNA, and dNTPs. Lane M, DNA size standards (end-labeled HaeIII fragments of φX174). The sizes of these fragments (in nucleotides), starting at the bottom, are: 72, 118, 194, 234, 271, 281, 310, 603, 872, 1,078, and 1,352.

Shown in Fig. 1B are examples of the DNA cleavage and reverse transcription reaction products with this 110-bp target DNA. When target DNA was incubated with R2 protein alone (lane 2), cleavage of only the primer (first) strand occurred, resulting in two bands of 50 and 60 nt. When vector RNA (lane 3) or R2 RNA (lane 5) was added to the cleavage reaction mixtures, cleavage of both strands occurred, with all four cleavage products visible on the gel. The level of non-primer strand cleavage is considerably higher with vector RNA than with R2 RNA. When dNTPs were added to enable reverse transcription, the products of the assay were not changed in the case of the vector RNA template (lane 4), but in the case of the R2 template (lane 6), the 60-nt fragment was reduced and the TPRT band of approximately 340 nt was generated. An additional faint band of 140 nt in lane 6 originated from the initiation of reverse transcription at an internal site within the R2 RNA (23). Note that the level of non-primer strand cleavage in lane 6 was increased to a level similar to that of vector RNA (lanes 3 and 4). This increase is consistent with our previous findings that non-primer strand cleavage is inhibited when the R2 RNA template is present but that the absence of nucleotides does not allow reverse transcription (22). To more clearly resolve the four DNA strands after cleavage, the reaction products were also separated on higher-resolution DNA sequencing gels. As shown in Fig. 2, the location of the primer strand cleavage is identical in the presence or absence of RNA. The location of non-primer strand cleavage in the presence of RNA is also identical whether or not that RNA was used in the reverse transcription reaction.

FIG. 2.

DNA products of the R2 cleavage reaction separated on a high-resolution sequencing gel. All reactions were conducted under the same conditions as in Fig. 1, but the products were separated on a 43-cm 8% DNA sequencing gel. Lane 1, no protein; lane 2, R2 protein; lane 3, R2 protein with vector RNA and dNTPs; and lane 4, R2 protein, R2 RNA, and dNTPs. The sizes of the cleavage products (in nucleotides) are shown at the right. The low levels of 62- and 48-nt fragments generated in lane 2 are a result of a small amount of bacterial RNA contaminating the R2 protein preparation used in this experiment. The weak band at 51 nt in all cleavage reactions is due to the addition of one extra nontemplated terminal nucleotide in the PCR used to generate the target DNA.

The ability to detect all four DNA products of the cleavage reaction is lacking for the reaction with the group II introns, in which one or both products of non-primer strand cleavage are detected only after treatment with RNase (38, 40). In the group II reaction, non-primer strand cleavage is a reverse splicing reaction that results in the covalent attachment of the intron RNA to the DNA strand. The cleavage products generated by the R2 protein are of the expected size (Fig. 1 and 2) and do not change if the reaction mixtures are treated with RNase A (data not shown), indicating that there is no covalent attachment of RNA to the DNA target.

Kinetics of the TPRT reaction.

Fig. 3 shows the products of a TPRT reaction as a function of time. Primer strand cleavage is detected in 1 min, the TPRT product is detected in 5 min, and non-primer strand cleavage is detected in 15 min. The generation of primer strand cleavage and TPRT products is essentially complete after 15 min, while non-primer strand cleavage products accumulate slowly for the first hour. Kinetics are similar if the reactions are conducted in the absence of dNTPs to allow only DNA cleavage or in the absence of RNA to allow only primer strand cleavage (data not shown). The rapid cessation of these reactions could be the result of the loss of protein activity after 15 min or the inability of the R2 protein to dissociate from the first DNA substrate and initiate cleavage of a second DNA substrate.

FIG. 3.

Time course of a complete TPRT reaction. Each 20-μl reaction mixture contained 200 fmol of internally labeled target DNA, 30 fmol of R2 protein, 10 pmol of R2 RNA, and 25 μM dNTPs. Incubation times (in minutes) at 37°C are indicated at the bottom. The reaction products were extracted, precipitated, and heat denatured before separation on an 8% denaturing polyacrylamide gel. The 48-nt band is poorly separated from the 50-nt band on these gels (Fig. 1). Lane M, DNA size standards identical to those shown in Fig. 1.

To determine if the R2 protein rapidly lost activity at 37°C, R2 protein in the absence of RNA was preincubated in the reaction buffer for 20 min before the addition of target DNA for a second 20-min incubation. As shown in Fig. 4, lane 1, this preincubated protein retained its ability to conduct the primer strand cleavage reaction. In lane 2, the R2 protein was preincubated for 20 min with a molar excess of unlabeled target DNA under our standard incubation conditions (0.2 M NaCl). The salt concentration was then raised to 0.8 M NaCl in order to dissociate any R2 protein that might be bound to target DNA. After the addition of salt, ³²P-labeled target DNA was added, and the reaction mixture was diluted with 3 volumes of buffer to reduce the salt concentration to 0.2 M NaCl again for a second 20-min incubation. The R2 protein retained cleavage activity after the first 20-min incubation and was able to cleave the primer strand of the labeled DNA in the second 20-min incubation. The levels of these cleavage products are approximately equal in lanes 1 and 2, suggesting that R2 protein activity was fully retained after one cycle of DNA cleavage and dissociation. As a control, in lane 3, the initial incubation with unlabeled DNA, the dilutions with buffer, and the addition of labeled DNA for a second incubation were performed as in lane 2 except that the salt concentration was maintained at 0.2 M at all times. Only a very low level of cleavage was detected in lane 3, confirming that the R2 protein needed to be physically dissociated from the unlabeled DNA in the first incubation to enable cleavage of the labeled DNA in the second incubation. Similar results to those shown in Fig. 4 have been obtained when R2 RNA and dNTPs are present in the reaction mixture (data not shown). We conclude that the in vitro R2 assays presented in this report represent single-cycle reactions because the R2 protein is unable to dissociate from the first target DNA it binds. The protein remains fully functional, however, because dissociating the protein from the cleavage site by the addition of salt restores its ability to cleave a second DNA target.

FIG. 4.

Demonstration that the R2 protein can conduct a second cleavage reaction if dissociated from a first DNA target. In lane 2, 30 fmol of R2 protein was preincubated with 300 fmol of unlabeled target DNA at 37°C for 20 min under the same conditions as described for the experiment shown in Fig. 1. To this 20-μl reaction mixture was then added 4 μl of 3.8 M NaCl–50 mM Tris-HCl (pH 8)–10 mM MgCl₂ to raise the NaCl concentration to 0.8 M. Following the salt increase, 200 fmol of radioactively labeled target DNA and sufficient buffer to reduce the NaCl concentration to 0.2 M were added. The mixture was then incubated at 37°C for another 20 min. In lane 1, the R2 protein was preincubated without the target DNA in the nicking buffer at 37°C for 20 min before a mixture of labeled and unlabeled DNA (to duplicate the concentration and specific activity of that in the second incubation mixture in lane 2) was added to the tube for the second 20-min incubation. In lane 3, the reaction was performed in the same fashion as in lane 2, except that the reaction volume was raised with buffers containing 0.2 M NaCl to allow the R2 protein to remain bound to the target DNA in all steps. Lane M, DNA size standards identical to those shown in Fig. 1.

RNA association changes the protein-DNA stoichiometry of the DNA nicking reaction.

Because the R2 protein is only able to conduct one cycle of the DNA cleavage reaction, we can readily estimate the reaction stoichiometry by determining the total number of R2 protein and DNA target molecules present in each reaction mixture and the fraction of that DNA which is cleaved (see Materials and Methods). In the absence of RNA, we have calculated that within our experimental error (10%), essentially all 120-kDa R2 protein molecules in our assays are capable of cleaving the primer strand of a target DNA molecule. Presumably, the reason for such a high level of active enzyme is that the purification procedure used to isolate the R2 protein depends upon the ability of the protein to bind both RNA and DNA substrates at high ionic strengths (22). However, this 1:1 stoichiometry of protein to target DNA in the cleavage reaction changes upon the addition of RNA. Figure 5A shows a time course for a DNA cleavage reaction with a constant amount of R2 protein in the presence or absence of R2 RNA. Unlike the experiments shown in Fig. 1 and 2, the DNA substrate is approximately sevenfold in excess of the level of R2 protein. While both reactions are complete after 20 min, the presence of R2 RNA results in a twofold reduction in the total level of cleaved DNA relative to the level in the absence of RNA. This twofold reduction was uniformly detected with different preparations of target DNA and R2 protein and was observed with the substitution of vector RNA for that of R2 RNA (data not shown). To eliminate the possibility that RNA was simply serving as a nucleic acid competitor for the target DNA, the experiment shown in Fig. 5B was conducted. In this experiment, the total level of DNA cleavage after 15 min was determined with increasing concentrations of R2 RNA. The twofold reduction in cleavage caused by the RNA was detected at the lowest RNA concentration tested (which is still in molar excess relative to the protein) and remained constant over a 16-fold increase in RNA.

FIG. 5.

Effect of R2 RNA on the number of DNA molecules cleaved by the R2 protein. All incubations were performed at 37°C with 200 fmol of target DNA and 30 fmol of R2 protein (i.e., a six- to sevenfold molar excess of target DNA over R2 protein). (A) Time course of the cleavage reaction with and without R2 RNA. The RNA concentration was 50 μg/ml in those reaction mixtures with RNA (10 pmol in each 20-μl reaction mixture). Numbers on the y axis represent the averaged values and the standard errors from phosphorimage quantitation for three independent experiments. All experiments are normalized to the level of DNA cleavage after 30 min in the absence of R2 RNA. (B) Level of DNA cleavage by the R2 protein as a function of increasing concentrations of R2 RNA. All incubations were performed for 30 min at 37°C. The twofold reduction in the level of cleavage at the lowest RNA concentration and the lack of a further reduction suggest that the RNA does not act as a competitive inhibitor in the cleavage reaction.

The simplest model to explain the twofold reduction in the cleavage reaction is that RNA induces the R2 protein to form a dimer and only one of the two subunits of the RNP complex remains capable of cleaving the target DNA. These results are also consistent with the formation of a higher multimeric complex as long as only one-half of the protein subunits in the complex retain the ability to nick DNA. Because the monomer R2 protein cannot dissociate from the target site after cleavage (Fig. 4), one prediction of this multimer model is that if the RNA is added after protein has already bound the target site, TPRT and non-primer strand cleavage will not be stimulated. We have tested this prediction by preincubating R2 protein with excess substrate DNA for 20 min before the addition of R2 RNA. A subsequent 20-min incubation gave less than 10% of the TPRT product and non-primer strand cleavage obtained in a parallel experiment in which the R2 protein and RNA were mixed before the introduction of the target DNA (data not shown). This result supports the model that R2 RNA induces the R2 protein to form a multimer rather than simply inducing a conformational change in the R2 protein.

Sedimentation analysis of the R2 protein-RNA complexes.

To more directly demonstrate that RNA can induce the R2 protein to form a multimeric complex, a series of sedimentation experiments were performed. As shown in Fig. 6A, in 0.2 M NaCl (the ionic conditions of our DNA cleavage and TPRT assays), R2 protein alone sediments at a rate similar to that of l-lactic dehydrogenase, a globular protein standard of approximately 125 kDa. This sedimentation rate is consistent with the 120-kDa R2 protein existing as a monomer in the absence of RNA. In the presence of excess levels of the 283-nt R2 RNA, the R2 protein sediments significantly faster, with a broad peak centered near the position of apoferritin, a protein standard of approximately 450 kDa. A similar shift to a faster-sedimenting form was induced by the presence of the 277-nt vector RNA. While this experiment suggests that vector RNA is equally able to form a multimeric complex with the R2 protein, this complex must differ from that of the R2 RNA because it cannot support the TPRT reaction (Fig. 1) (22). To directly demonstrate a different interaction of the R2 protein with R2 RNA versus vector RNA, we conducted a series of sedimentation experiments at higher salt concentrations. Figure 6B shows the results of one such sedimentation experiment conducted in 0.4 M NaCl. In the absence of RNA or the presence of vector RNA, the R2 protein sedimented at a rate consistent with that of a monomer. In the presence of R2 RNA, the protein sedimented faster as a broad peak and somewhat more slowly than the 450-kDa apoferritin peak. Thus the RNP complex formed with R2 RNA is significantly more stable at higher ionic strengths than the complex formed with vector RNA.

FIG. 6.

Sedimentation of the R2 protein with and without RNA on 10 to 30% glycerol gradients. Two-hundred-microliter samples containing 240 ng of R2 protein alone, R2 protein with 3 μg of R2 RNA, or R2 protein with 3 μg of vector RNA were loaded onto gradients. Samples contained 50 mM Tris-HCl (pH 8), 10 mM MgCl₂, and 1 mM DTT. The 5-ml 10 to 30% glycerol gradient itself contained 50 mM Tris-HCl (pH 8) and 0.2 or 0.4 M NaCl. Fractions were collected from the bottom of the tube. R2 protein was detected by assaying primer strand cleavage activity of each fraction in 0.2 M NaCl. Protein standards were l-lactic dehydrogenase (Lac), catalase (Cat), and apoferritin (Apo) and were individually fractionated on gradients identical to those used for the R2 protein (see Materials and Methods). The location and molecular weight of each protein standard are indicated at the top of panel A. (A) Sedimentation of the R2 protein in 200 mM NaCl. (B) Sedimentation of the R2 protein in 400 mM NaCl.

As described in the previous section, the twofold change in the stoichiometry of protein to DNA target upon RNA binding is consistent with the R2 protein forming a multimeric complex in which only half of the protein subunits are capable of cleaving the primer DNA strand. Unfortunately, the very-broad sedimenting peak formed by the protein-RNA complex around the position of a 450-kDa globular protein does not indicate with any certainty whether this complex is a dimer or a tetramer. Indeed, the trailing of the RNA-protein peak to faster-sedimenting forms, especially in 0.2 M NaCl, is even more pronounced at higher protein concentrations, indicating that different multimeric complexes are being formed in a concentration-dependent fashion (data not shown). Unfortunately, because we do not know how the binding of RNA by the protein affects the overall conformation or density of the resulting complex, further sedimentation experiments to resolve the sizes of the complexes will be difficult. For the remainder of this article, we will refer to the RNA-protein complex as a dimer because that represents the simplest form which is consistent with the stoichiometry of the cleavage reaction. However, it is clearly possible that in our in vitro assay, tetramers or even larger complexes are present with each complex capable of binding and cleaving multiple target DNA molecules at a ratio of two protein subunits per DNA target.

RNA binding increases the specificity of the R2 protein for the DNA target.

We have conducted a number of studies to detect differences in activity between the R2 protein monomer and RNP complex. Gel shift assays revealed that both the monomer and complex could quantitatively bind the target site at picomolar concentrations (data not shown). More revealing were experiments conducted to monitor the specificity of the cleavage reaction. Examples of the DNA nicking assays conducted at NaCl concentrations from 20 to 400 mM are shown in Fig. 7. As shown in panel A, in the absence of RNA, maximum levels of specific target site cleavage can be found in 100 and 200 mM NaCl. When the salt concentration was higher than 0.2 M NaCl, the level of DNA cleavage was greatly reduced. At salt concentrations below 100 mM NaCl, the level of specific cleavage was reduced, with a corresponding increase in cleavage products of lengths other than 50 and 60 nt. Shown in Fig. 7B is the cleavage reaction conducted in the presence of R2 RNA. Again, maximum levels of cleavage, which can be seen from 100 to 200 mM NaCl, decreased when the salt concentration was increased. At lower salt concentrations, however, specificity for the target site remains high, as only the 50- and 60-nt fragments could be detected. (Non-primer strand cleavage can also be detected in these cleavage reactions conducted with RNA.) These results indicate that under low-ionic-strength conditions, the R2 protein monomer loses its specificity for recognizing and cleaving the target site, while the R2 protein-RNA complex retains its specificity. Vector RNA sequences appeared to increase the target specificity of the R2 protein at low ionic strength to about the same degree as R2 RNA (data not shown).

FIG. 7.

Effect of NaCl concentration on the R2 cleavage reaction without RNA (A) or with 10 pmol of R2 RNA (B). All assays were performed under the same cleavage conditions as those for the experiment shown in Fig. 5A, except that the NaCl concentrations was varied from 20 to 400 mM as indicated at the bottom. All incubations were conducted at 37°C for 30 min.

The increased specificity of the R2 protein for the target sequence induced by the binding of RNA was also demonstrated in DNA competition experiments. Shown in Fig. 8 are experiments in which the R2 protein was incubated with the 110-bp labeled DNA target and increasing amounts of pUC18 vector DNA as a competitor. The NaCl concentration in these assays was 0.2 M. In the absence of R2 RNA, pUC18 DNA effectively competed with the 110-bp target molecule. At the highest level of competitor DNA tested, 38 μg/ml, target site cleavage was only 5% of the level in the absence of competitor. In the presence of R2 RNA, the R2 protein was capable of efficient cleavage of the target site at all concentrations of competitor DNA tested. Based on the estimated concentration of competitor DNA needed to reduce the level of cleavage by 50% (2 μg/μl without RNA and 300 μg/μl with RNA), the specificity of the R2 protein for the target site was increased 150-fold by the addition of RNA. It should be noted that to optimize the effect of the competing DNA, the incubations shown in Fig. 8 were conducted for only 5 min, when the kinetics of the reaction were determined to still be in the linear range (see also Fig. 3).

FIG. 8.

DNA competition assays with and without R2 RNA. First-strand cleavage assays were performed under the same cleavage conditions as those described for the experiment shown in Fig. 5A except that various amounts of pUC18 DNA were added to each reaction mixture. Reaction mixtures were incubated at 37°C for 5 min, and reactions were stopped by the addition of 3 μl of 0.5 M EDTA. The tubes were then placed on dry ice. The level of DNA cleavage at each pUC18 concentration was normalized to the level of DNA cleaved in the absence of DNA competitor. Each value represents the average and the standard error obtained from three independent experiments. In tubes without RNA, 1 μg of RNase A was included to ensure that no RNA was present.

The R2 protein-RNA complex can find the 28S target site in total genomic DNA.

To determine if the R2 protein-RNA complex had the ability to find the 28S gene target site in the presence of the entire host genome, we conducted the experiment shown in Fig. 9. Total genomic DNA was first digested with the restriction enzyme ClaI which, as diagrammed in panel A, cleaves the 28S gene 1.55 kb upstream and 0.65 kb downstream of the R2 target site. Thus, the R2 target site for those 28S genes that do not contain an R1 or R2 insertion (defined here as uninserted ribosomal DNA [rDNA]units) are located on a 2.2-kb fragment (the band labeled U). Because no ClaI sites are present within either the R1 or R2 elements of this species, 28S genes containing either a full-length R1 or R2 insertion give rise to restriction fragments of lengths of 5.8 and 7.2 kb, respectively (the bands labeled R1 and R2) (16). The ClaI-digested genomic DNA was then treated with R2 protein alone or with R2 protein and either R2 RNA or vector RNA in the presence of dNTPs. After incubation, the DNA was fractionated on an agarose gel, blotted onto a nitrocellulose membrane, and hybridized with a 28S gene probe corresponding to the 300-bp region downstream of the R1 and R2 target sites (Fig. 9A). As shown in Fig. 9B, incubation of the genomic DNA with increasing amounts of R2 protein without RNA gave rise to only low levels of a 650-bp band, the size expected for the cleavage of the 28S gene target site (the band labeled Cleavage). R2 protein plus vector RNA resulted in a significant increase of the 650-bp cleavage band. R2 protein plus R2 RNA gave rise to a band of approximately 920 bp (labeled TPRT), indicating that the R2 protein had found the target site of the uninserted rDNA units and conducted the TPRT reaction with the 283-nt R2 RNA. The intensity of the TPRT band and the corresponding reduction in intensity of the original 2.2-kb 28S band indicated that the R2 protein was best able to find the 28S target site when bound to its own R2 RNA. The R2 protein in the presence of RNA was also able to cleave and conduct TPRT with those 28S genes containing R1 insertions [labeled TPRT (R1-inserted)], although they are not easily seen in Fig. 9 because they are obscured by the 28S genes with R2 insertions. Those 28S genes already containing an R2 insertion (labeled R2) were not cleaved by the R2 protein.

FIG. 9.

Total genomic DNA digested with the R2 endonuclease. (A) Diagram of the D. melanogaster rDNA unit. The sizes of the ClaI restriction fragments for uninserted 28S genes, genes with R1 insertions, and genes with R2 insertions are indicated. The fractions of the total number of 28S genes that are represented by these three conditions are indicated next to the diagrams. The location of the radiolabeled 28S gene probe used in the genomic blot in panel B is shown by the diagonally shaded boxes. (B) Genomic blot of DNA digested with ClaI and the R2 endonuclease. In each lane, 1.2 μg of D. melanogaster DNA was initially digested with ClaI, the NaCl concentration was then adjusted to 0.2 M, dNTPs were added to a final concentration of 25 μM, and either 4, 12, or 20 ng of R2 protein (30, 90, or 150 fmol) and 1 μg of RNA (10 pmol) were added for a 1-h incubation. The amount of R2 protein added to each tube (in nanograms) is indicated at the bottom of each lane. Lane M, DNA size standards. The sizes of the fragments (in kilobases), starting at the bottom, are 0.60, 0.87, 1.08, 1.35, 2.0, 2.3, 4.4, 6.7, 9.9, and 23.

DISCUSSION

In this study, we have extended our characterization of the protein-DNA and protein-RNA interactions involved in the TPRT reaction of R2 elements. These studies were made possible by more-sensitive in vitro assays utilizing labeled 110-bp DNA targets and our ability to store active R2 protein for extended periods. With this assay, the R2 protein was shown to be capable of only one round of the cleavage and reverse transcription reaction, because the protein remained bound to the first target DNA molecule. Given the ability of the R2 protein to bind quantitatively to this site in the picomolar range, it was not surprising that once the enzyme is bound to the target site, it cannot dissociate. The finding that the R2 protein remained bound even after DNA cleavage is also consistent with the location of the R2 protein recognition sequences on the target DNA. Bal 31 deletions of the 28S gene target site indicated that the recognition sequences of the R2 protein extended from approximately 25 bp upstream of the insertion site to approximately 5 bp downstream of this site (21, 35). Therefore, even after double-stranded DNA cleavage, most of the recognition sequences are intact, and the protein presumably remains bound to the upstream sequences. While not conducted by our in vitro assays, completion of a full R2 integration reaction in vivo requires the attachment of the newly synthesized cDNA to the upstream 28S gene sequences. We have postulated several possible mechanisms for this attachment (11). The dissociation of the R2 protein from the upstream DNA sequences after second-strand cleavage is likely to be a key factor in this attachment reaction.

One useful advantage of the R2 protein’s ability to initiate only one round of DNA cleavage is that it has allowed a determination of the stoichiometry of this reaction. Purified R2 protein was found to be capable of binding the DNA target site and conducting primer strand cleavage at a 1:1 ratio of protein monomer to DNA target. However, in the presence of RNA, the R2 protein cleaved DNA at a ratio of two protein subunits per target site. As diagrammed in Fig. 10, this finding suggests that the RNA induces the formation of a higher-ordered complex in which only half of the subunits can cleave the DNA. While the sedimentation and cleavage properties of the RNP complex are consistent with the formation of a protein dimer, larger complexes can still be formed.

FIG. 10.

Model for the activity and conformational changes of R2 protein upon association with RNA. The 120-kDa R2 protein is shown in this diagram as an oval. R2 monomers can bind the 28S target site but are able to nick only the first DNA strand. In the presence of R2 RNA, R2 protein monomers associate to form a higher-order complex. The exact nature of this complex is not known, but the active unit appears to be a dimer, based on the changes in stoichiometry of the cleavage reaction from a 1:1 ratio of protein subunit to target DNA in the absence of RNA to a 2:1 ratio of subunit to DNA target in the presence of RNA. It is possible that under our in vitro conditions, tetramers or even larger R2 complexes are formed, with each complex capable of cleaving multiple-target DNA molecules at a ratio of two protein subunits per DNA target. The R2 RNP complex is capable of first- and second-strand cleavage and reverse transcription of R2 RNA. The R2 RNA in this diagram is drawn to imply that secondary structure of the RNA is important in the R2 protein-RNA interaction (see reference 25). The R2 protein dimer is drawn bound by one RNA molecule, because each dimer appears able to conduct only one reverse transcription reaction. The actual number of RNA molecules bound per protein monomer is not known.

Demonstration that the R2 protein forms a higher-order complex with RNA explains why cleavage of the primer DNA strand can occur with protein alone whereas cleavage of the non-primer DNA strand requires RNA (22). The RNA requirement for double-stranded cleavage of the target site was in principle similar to the RNA requirement for the DNA cleavage in the TPRT reaction catalyzed by group II introns (reviewed in reference 4). In the case of the group II introns, the intron RNA aids in target site recognition by annealing to the target DNA and is the catalyst for non-primer strand cleavage by a reverse splicing reaction (8, 12, 38, 40). The result of this reverse splicing is the covalent attachment of the intron RNA to one or both ends of the coding strand of the DNA target. In the case of the R2 TPRT reaction, it seemed unlikely that RNA would be serving such a catalytic role in DNA cleavage, because even vector RNA sequences are able to promote non-primer strand cleavage. It was still possible, however, that the 5′ or 3′ end of the RNA was serving as a recipient of a transesterification reaction catalyzed by the R2 protein. In this study, we have demonstrated that there is no covalent attachment of the RNA to the target site. All four DNA strands of the predicted size are readily detected after cleavage in the absence of RNase. Thus, RNA appears to be necessary for non-primer strand cleavage only because it is a protein multimer that is required to cleave both DNA strands.

The degree to which an RNA template promotes reverse transcription and non-primer strand cleavage varies greatly. Vector RNA promotes the highest levels of non-primer strand cleavage yet is not utilized as a template in the TPRT reaction. R2 RNA is utilized as a template in the TPRT reaction but does not promote high levels of non-primer strand cleavage unless dNTPs are added to permit reverse transcription (Fig. 1) (22). Thus, the binding of these two RNAs appears to induce alternative conformational changes in the R2 protein. The sequences of the R2 RNA that are needed to induce the conformation capable of TPRT are distributed throughout the 250-nt 3′ untranslated region (23). This recognition appears to depend upon the secondary and tertiary structure of the RNA rather than simply on its primary sequence. RNA derived from the 3′ untranslated region of both the B. mori and D. melanogaster R2 elements is utilized by the B. mori protein with similar efficiencies in the TPRT reaction; yet, there is little similarity of nucleotide sequence between the RNAs of these two R2 elements (25). Consistent with this suggestion, R2 RNA can assume a secondary structure which is conserved between elements from different insect species (25).

The efficiency with which the R2 RNA is used as a template in the TPRT reaction is also highly dependent upon the exact sequence at its 3′ end. Elimination of just a few nucleotides from the 3′ end of the RNA or, conversely, the addition of a few nucleotides of 28S rRNA sequences significantly changes the ability of the R2 protein to position the RNA for TPRT (23, 24). Because we have been unable to identify a promoter at the 5′ end of R2 elements (10), one likely model of R2 expression requires cotranscription with the 28S gene. Based on the secondary structure of 28S rRNA (13) in such a cotranscript, the 28S rRNA sequence downstream of the R2 element would form helices with 28S rRNA sequences upstream of R2. We are currently testing whether the addition of rRNA sequences at both ends of the R2 sequences increases the efficiency of the reverse transcription and non-primer strand cleavage reactions.

We have also shown in this study that the formation of an R2 RNP complex not only allows cleavage of both strands but also results in more specific binding of the endonuclease to the DNA target site. The R2 protein monomer cleaves nontarget DNA sequences at low ionic strength and is readily competed from the target site by other DNA sequences at physiological ionic strengths. The RNP complex, on the other hand, remains highly specific for the 28S target site at all salt concentrations and cannot be easily competed from this site by other DNA sequences. Many proteins are known to bind to DNA as dimers or as tetramers (reviewed in reference 28). The recognition sequences for these proteins usually have dyad symmetry. The R2 protein recognition site also has symmetry around a sequence 14 bp upstream of the insertion site (CTCA/GAGT 7 to 10 bp upstream and its inverse sequence 18 to 21 bp upstream). While this region of symmetry is centered in the region known to be involved in R2 protein binding, it should be noted that this symmetry is a direct result of a highly conserved secondary structure of the encoded 28S rRNA (13). Thus, it is possible that this region of dyad symmetry is unrelated to R2 protein binding. Indeed, the fact that the R2 protein monomer alone does not dissociate at physiological ionic strength and can nick the target site at NaCl concentrations as high as those of the RNP complex (Fig. 7) suggests a model in which only one of the subunits of a dimer is predominantly responsible for DNA binding. In this model, the association of the second subunit alters the conformation of the first subunit only in a manner that increases its specificity for the target site. It should be possible to directly address the question of whether only one subunit or both subunits make contact with the DNA target by comparing DNase I and methylation footprints of the R2 monomer and multimer bound to the target site.

While a single subunit of the R2 protein dimer is capable of specific DNA binding and primer strand cleavage, there are no data to indicate whether or not this same subunit cleaves the non-primer DNA strand or contains the active site for the RT. The most-detailed structural studies of an RT have been conducted with the enzyme encoded by the human immunodeficiency virus (HIV) (18, 33). The HIV RT forms a heterotypic dimer, with a p66 subunit serving as the catalyst in the reaction and a p51 subunit serving to bind the tRNA primer for the reverse transcription. Remarkably, while both the p66 and p51 peptides contain the same N-terminal amino acid sequences, they assume different conformations and functions in the complex (18, 33). Given the conservation of critical residues and segments associated with the catalytic subunit between non-LTR retrotransposons and retroviruses (19, 26), we suggest that one of the subunits of the R2 dimer assumes a conformation similar to that of the p66 catalytic subunit of HIV. If it is eventually shown that the second R2 protein subunit is responsible for binding the primer in the reaction (i.e., the nicked DNA target), the similarity between the R2 and HIV reactions may be high.

Further characterization of the R2 RNP complex would be aided by a determination of those segments of the 120-kDa R2 protein that are responsible for the protein-protein, protein-RNA, and protein-DNA interactions. Based on sequence homology, the central one-third of the R2 protein is the RT domain. Located at the amino-terminal and carboxyl-terminal ends of the protein are domains that contain conserved cysteine-histidine motifs that could be involved in the binding to nucleic acid (14). Two-dimensional nuclear magnetic resonance studies of a peptide containing the amino-terminal cysteine-histidine domain have revealed that it binds a zinc ion and can form a tertiary structure similar to that of TFIIIa zinc fingers (34). Mutagenesis of the amino-terminal, central, and carboxyl-terminal domains, followed by attempts to complement their activities as heterodimers should reveal the activities associated with each domain and the enzymatic functions of each subunit in the R2 protein complex.