| Nucleic Acids Research | Pages |
The double-stranded RNA-binding protein Xlrbpa promotes RNA strand annealing
Introduction
Materials And Methods
Expression and purification of recombinant Xlrbpa
Expression and purification of recombinant dsRBD Xl-2
Expression of His-tagged dsRBD constructs
Northwestern assays
Preparation of RNA substrates
RNA-annealing assay
Results
Xlrbpa exhibits RNA strand-annealing activity
An isolated dsRBD is sufficient to promote RNA strand annealing
Heparin inhibits RNA strand annealing
RNA strand annealing does not depend on RNA binding
Discussion
Acknowledgements
References
The double-stranded RNA-binding protein Xlrbpa promotes RNA strand annealing
ABSTRACT
INTRODUCTION
The double-stranded RNA-binding domain (dsRBD) is a conserved protein motif of ~70 amino acids that binds exclusively to double-stranded (ds)RNAs and RNA-DNA hybrids but fails to bind dsDNA (1-3). The reason for this binding specificity can most likely be explained by the unique A-form helix that is assumed by both RNA-RNA and RNA-DNA hybrids but not by dsDNA, which normally assumes a B-form helix (4). dsRBDs can be found in many RNA-binding proteins from Escherichia coli to man (5,6). The functions of dsRBD-containing proteins vary considerably and range from an RNase to an RNA localizing protein (5,6). Also, the number of dsRBDs found in any given protein can vary and range from a single dsRBD in E.coli RNase III to up to five dsRBDs found in the Drosophila Staufen protein. Not all of the multiple dsRBDs found in some proteins show RNA-binding activity as an isolated domain in vitro, but seem to contribute to a cooperative binding effect when in context with other dsRBDs (6).
The NMR structure of two dsRBDs has been solved and shows a conserved folding pattern for these two domains, suggesting a common structure for all dsRBDs (5,7). Conserved amino acids are located on one surface of the dsRBD and could make direct contact with the backbone of RNA, which is confirmed by site-directed mutagenesis and footprint experiments (3,6,7).
Consistent with the various functions of dsRBD-containing proteins, many of them contain other functional or catalytic domains. Drosophila maleless protein, for instance, carries an RNA helicase domain while the dsRNA-dependent, interferon-activated kinase PKR contains a conserved protein kinase domain at its C-terminal end (8,9). dsRNA adenosine deaminase, dsRAD or ADAR, on the other hand, contains an adenosine deaminating domain required for conversion of adenosines to inosines (10).
In contrast, the Xenopus RNA-binding protein A (Xlrbpa) and its human homologue TAR-RNA-binding protein (hsTRBP) are composed of three dsRBDs and contain no other obvious catalytic domains. Both proteins are relatively small, with 299 and 346 amino acids for Xlrbpa and hsTRBP, respectively (11,12). The three dsRBDs found in both proteins comprise >63 (Xlrbpa) and 53% (hsTRBP) of the entire protein. The dsRBDs are evenly distributed with linker regions of 30-60 amino acids between them (11).
hsTRBP has been isolated by its ability to bind HIV TAR-RNA but has since been implicated as a modulator or antagonist of dsRNA-dependent, interferon-activated protein kinase PKR (13,14). The antagonistic effect of hsTRBP has been explained by an RNA sequestering activity of hsTRBP or by direct interaction of hsTRBP with PKR (14,15).
Xlrbpa, in contrast, has been shown to interact with most structured cellular RNAs, leading to its association with ribosomes and hnRNPs (11). Despite an intensive search, no Xenopus homologue of PKR has been identified (G.Schabbauer and M.Jantsch, unpublished observation). Therefore, a role for Xlrbpa as a regulator of PKR can neither be confirmed nor excluded at this point.
However, just like other hnRNP proteins, Xlrbpa might simply act as a general hnRNP helping to complex structured RNAs. Alternatively, Xlrbpa might help to fold or stabilize double-stranded regions in RNA. RNA chaperonin-like activity has previously been shown for other RNA-binding proteins (16,17).
To test whether Xlrbpa or an isolated dsRBD could act as an RNA chaperone we have purified full-length Xlrbpa and individual dsRBD of this protein and tested them for their ability to promote RNA strand annealing. Here we can show that an isolated dsRBD and also full-length Xlrbpa can promote RNA strand annealing in vitro.
MATERIALS AND METHODS
Expression and purification of recombinant Xlrbpa
The entire open reading frame (ORF) encoding Xlrbpa was amplified by PCR using primers carrying restriction sites. The 5[prime] primer carried a BglII and the 3[prime] primer contained a KpnI site. The PCR product was gel-purified, cut with BglII and KpnI and ligated into a modified pET 3 vector carrying BamHI and KpnI sites (18).
For protein expression the resulting plasmid, pET 3.5, was transformed into E.coli BL21(DE3). Cultures (300 ml) were grown in 2× TY medium containing 50 mg/l ampicillin. Protein expression was induced by addition of IPTG to 1 mM final concentration when cultures reached an OD600 of 0.7. After 3 h of induction cells were harvested by centrifugation and frozen at -70°C. Inclusion bodies were prepared as described (19). Inclusion bodies were subsequently dissolved in buffer A (6 M urea, 10 mM NaCl, 20 mM HEPES, pH 7.4) and loaded on a MacroprepS column (BioRad, CA). Proteins were fractionated in a linear 150 ml gradient of buffer A and buffer B (6 M urea, 1 M NaCl, 20 mM HEPES, pH 7.4). Most Xlrbpa protein eluted at 300 mM NaCl. The fraction containing the protein was diluted 1:2 with dilution buffer (10 mM NaCl, 20 mM HEPES, pH 7.4), giving rise to a final buffer composition of 155 mM NaCl, 3 M Urea, 20 mM HEPES, pH 7.4). The diluted fraction was subsequently loaded on a S-Sepharose column (Pharmacia, Uppsala, Sweden) and eluted in a continuous 100 ml gradient of buffers C (3 M urea,10 mM NaCl, 20 mM HEPES, pH 7.4) and D (3 M urea,1 M NaCl, 20 mM HEPES, pH 7.4). The protein eluted at 300 mM NaCl and was >90% pure as judged by staining of an SDS-PAGE gel. Subsequently, purified protein was fractionated on a size exclusion column, dialysed against several changes of dH2O containing 0.05% TFA and lyophilized.
Expression and purification of recombinant dsRBD Xl-2
The second dsRBD of Xlrbpa was amplified by PCR using primers carrying a BamHI site (5[prime] primer) and a KpnI site (3[prime] primer). The PCR product was cut with these two enzymes and cloned into a modified pET3 vector cut with the same enzymes.
The resulting plasmid, designated pET 3.2, was used for protein expression. Expression and purification of recombinant pET3.2 was identical to that of full-length Xlrbpa (see above).
Expression of His-tagged dsRBD constructs
All three dsRBDs of Xlrbpa, chimeras between the first and second dsRBDs and mutant versions of the second dsRBD were cloned in the His fusion vector pRSET as described (6). His fusion proteins were produced in E.coli BL21(DE3) by induction with IPTG. Recombinant proteins were purified by Ni-chelating chromatography, following the manufacturer's protocol (Qiagen, Germany).
Northwestern assays
Northwestern assays of recombinant Xlrbpa or Xl2 were performed as described (6).
Preparation of RNA substrates
For the annealing assay a pBluescript plasmid containing a 290 bp long region of the Xlrbpa cDNA was used. To generate the cold RNA, the plasmid was linearized with KpnI. A 346 nt long run-off transcript was generated from the T7 promoter. The complementary, 105 bp long, hot RNA was transcribed from the T3 promoter after linearization of the plasmid with PstI. To label the RNA, 30 µCi [[alpha]-32P]UTP was added to the transcription reaction. Both hot and cold RNA were purified on a denaturing acrylamide gel. The two RNAs are complementary in a region of 87 nt.
RNA-annealing assay
For RNA-annealing assays the protein of interest was brought to 5 µl in 2× RNA annealing buffer (200 mM KCl, 1 mM magnesium acetate, 1 mM DTT, 0.1 mg/ml BSA, 40 mM HEPES, pH 7.6) and preincubated for 10 min at 30°C. The RNA solution was prepared by heating 4 and 2 nM solutions of cold and hot RNAs, respectively, to 65°C for 5 min. After quick chilling on ice equal volumes of each RNA were mixed on ice. An equal volume of this solution containing 2 nM cold RNA and 1 nM hot RNA was added to the protein and the annealing reaction was allowed to take place for 10 min at 30°C. RNase T1 (Boehringer-Mannheim) was then added to a final concentration of 1 U/µl and the reaction was incubated for 15 min at 37°C. An aliquot of 100 µl TE buffer was added and the RNA was extracted with phenol:chloroform:isoamyl alcohol (25:24:1) and precipitated by addition of 2.5 vol ethanol. RNA pellets were recovered by centrifugation, washed with 70% cold ethanol and resuspended in RNA urea buffer. RNA fragments were resolved by electrophoresis on a 10% (29:1) acrylamide:bisacrylamide-8 M urea gel. The gel was dried and exposed.
RESULTS
Xlrbpa exhibits RNA strand-annealing activity
To test whether Xlrbpa might exhibit strand-annealing activity the cDNA encoding the protein was cloned in a pET vector which is suitable for protein expression in E.coli (18). Full-length Xlrbpa could be purified from E.coli expressing the protein by two rounds of ion exchange chromatography. The purified protein exhibited dsRNA-binding activity when tested in northwestern assays (Fig.
Figure 1. Northwestern RNA-binding analysis of purified Xlrbpa and the isolated second dsRBD, Xl2. (a) Coomassie stained gel on which 4 µg purified Xlrbpa and Xl2 were loaded. (b) Northwestern RNA-binding assay of Xlrbpa and Xl2. After electrophoresis, proteins were transferred to a PVDF membrane, refolded by urea treatment and incubated with radiolabelled rI/rC. Both proteins show RNA binding. Since RNA binding of full-length Xlrbpa is ~10-fold stronger than that of the isolated Xl2 domain, a longer exposure of the Xl2 lane is shown. Figure 2. Full-length Xlrbpa promotes RNA strand annealing. (a) RNA strand annealing was assayed using two partially complementary RNAs. A 340 nt long cold RNA (broken line) and a 105 nt long hot RNA (triangles) that could base pair over an 87 bp long complementary region were mixed in the presence or absence of purified Xlrbpa. Upon digestion with RNase T1 an ~90 bp long fragment will be protected if annealing takes place. (b) Cold top strand (TS) and radiolabelled bottom strand (BS) were mixed with various concentrations of Xlrbpa. After incubation for 10 min at 30°C, RNase T1 was added to determine the amount of annealed and thus protected fragment. Final concentrations of TS and BS were 2 and 1 nM, respectively. Lanes 1-3, annealing assay in the presence of 3 or 2 µM or no Xlrbpa. Lanes 4-9, controls. Lanes 4 and 6 check the integrity of undigested input RNAs. Lane 5 controls RNase T1 digestion in the absence of annealing. Lane 7, TS and BS RNAs annealed in the presence of 1 M KCl. In lanes 8 and 9 both strands or only the radiolabelled bottom strand, respectively, were incubated with Xlrbpa to check for RNase contamination of the protein preparation. Figure 3. Schematic representation of dsRBD constructs used in this study. The sequences of Xl1, Xl2, Xl3 and construct 2/1 are aligned with each other. Conserved amino acids found in the dsRBD consensus sequence are in bold. Positions of the F/L and R/P mutations in the Xl2 domain are indicated. This experiment showed that full-length Xlrbpa can promote strand annealing in a concentration-dependent manner. Maximum annealing was obtained at a 1000-fold molar excess of protein over RNA, which amounts to a nucleotide:protein ratio of ~1:2.
An isolated dsRBD is sufficient to promote RNA strand annealing
We have shown previously that the second dsRBD found in Xlrbpa (Xl2) can bind dsRNA as an isolated domain. In contrast, the other two dsRBDs, termed Xl1 and Xl3, show no RNA-binding activity by themselves but contribute to a cooperative binding effect (6). We therefore wanted to know whether RNA strand annealing could also be exhibited by Xl2, the isolated second dsRBD of Xlrbpa (Fig.
The single Xl2 domain was thus expressed and purified from E.coli. Like full-length Xlrbpa, the purified Xl2 domain shows RNA binding in a northwestern assay, however, at a much reduced level when compared with Xlrbpa (Fig.
Annealing experiments performed with the isolated dsRBD show that Xl2 is sufficient to promote RNA strand annealing in a concentration-dependent manner (Fig.
Figure 4. The isolated dsRBD Xl2 is sufficient to promote RNA strand annealing. RNA-annealing assay performed with various concentrations of purified Xl2. Final concentrations of TS and BS were 2 and 1 nM, respectively. In lanes 1-6 the final concentration of Xl2 is given in µM. Lanes 7-12, controls as described in Figure 2. Xlrbpa and Xl2 were purified from E.coli using ion exchange chromatography. It is thus possible that trace amounts of E.coli proteins were co-purified which could lead to the observed annealing effect. We therefore decided to repeat the experiments with proteins purified by other means. To do this, the Xl2 domain was expressed as a 6×His fusion protein and purified by Ni-chelating chromatography. Also, to test whether contaminating E.coli protein could lead to the annealing effect, a protein purification with empty plasmid was performed in parallel. To test for contaminating RNase activity the purified proteins were incubated with radioactive single-stranded RNA alone or with a mixture of complementary hot and cold RNA. After incubation at 37°C for 15 min the RNAs were recovered by phenol extraction, precipitated and analysed on a denaturing acrylamide gel (Fig. RNA annealing was observed with the Xl2 domain purified by either method while the mock protein preparation showed no annealing effect. This suggests that the observed annealing effect is an intrinsic property of the dsRBD and not caused by co-purifying E.coli proteins. Again, maximum annealing was observed at a nucleotide:protein ratio of 1:2 for FPLC-purified Xl2 while His-tagged Xl2 showed maximum annealing at a nucleotide:protein ratio of 1:4. Interestingly, more input RNA was converted to the double-stranded form by His-tagged protein than by FPLC-purified Xl2, indicating that the former preparation was slightly more active. FPLC purification of Xl2 involves denaturation of proteins while the His-tagged proteins were purified under native conditions. We have recently shown that refolded Xl2 is slightly less active in RNA binding than Xl2 purified under native conditions (23). It is thus possible that the observed differences in annealing activity of the two Xl2 preparations reflect differences in their folding status.
Heparin inhibits RNA strand annealing
Xlrbpa is a basic protein with a calculated pI of 8.24. The isolated Xl2 domain is even more basic, with a calculated pI of 9.99. It is therefore conceivable that the strand-annealing activity exhibited by those two proteins was due to their basic nature, which would attract negatively charged RNA molecules and bring them into physical proximity, thus facilitating the initial step of finding the right partner for annealing. We therefore tested whether the strand-annealing activities of Xlrbpa and Xl2 could be competed by the polyanion heparin. As shown in Figure
RNA strand annealing does not depend on RNA binding
To further investigate whether the charge of the dsRBD is important for strand annealing we compared the first and third dsRBD of Xlrbpa and mutant versions of the second dsRBD for their RNA-annealing activity. Xl1, the first dsRBD in Xlrbpa, has a calculated pI of 7.82 when expressed as a His-tagged fusion protein, while His-tagged Xl3, the third dsRBD of Xlrbpa, has a calculated pI of 5.98. Thus, Xl1 and Xl3 are neutral or even acidic when compared with Xl2. Mutation F/L, in contrast, exchanges a central conserved phenylalanine for a leucine in the Xl2 domain. The resulting pI of 9.99 is identical to that of the wild-type Xl2 domain, yet the F/L mutation is unable to bind dsRNA (6). Another mutation, R/P, replaces a basic arginine at the C-terminal end of the Xl2 domain by a proline, thereby reducing the charge of the domain to a pI of 9.31 and disrupting a predicted [alpha]-helical structure in the region of the mutation (6,7). Finally, mutation 2/1 is a chimera that fuses the N-terminal half of the Xl2 domain to the C-terminal half of the Xl1 domain (6). His-tagged 2/1 construct has a calculated pI of 9.24. The position of the various mutations is depicted in Figure
It should also be noted that only Xl2 can efficiently bind dsRNA as an isolated domain (6). However, Xl1 and Xl3 can contribute to a cooperative binding effect when expressed in combination with other dsRBDs. Similarly, mutations F/L and R/P and the 2/1 construct show greatly reduced RNA-binding activity (6). Therefore, comparison of the annealing activities exhibited by those constructs not only tells whether the net charge of the dsRBD is important for RNA annealing but also indicates whether RNA binding is a prerequisite for RNA strand annealing in vitro.
Thus, all the mentioned constructs were expressed as His-tagged fusion proteins and purified from the soluble protein fraction by Ni-chelating chromatography. The purified proteins were tested for the absence of contaminating RNase activities and used in RNA annealing assays (Fig.
Interestingly, the almost neutral Xl1 domain showed an annealing activity comparable with that of the Xl2 domain. The F/L and R/P mutations showed even stronger RNA-annealing activity than the wild-type Xl2 domain. Maximum strand-annealing activity was reached at a protein:nucleotide ratio of 1:1 and 1:2 for the F/L and R/P mutations, respectively. The third dsRBD of Xlrbpa, Xl3, and the chimeric 2/1 construct, in contrast, showed barely detectable annealing activities. Only at protein:nucleotide ratios of 8:1 was a small percentage of the total RNA input annealed by those two constructs (Fig.
Taken together, the data indicate that the net charge of the dsRBD is not a major determinant of RNA annealing. However, since heparin inhibits the strand-annealing effect of Xl2 we cannot exclude the possibility that individual basic amino acids that might be exposed on a certain substructure of the dsRBD are important for strand annealing. In fact, the Xl3 domain and the 2/1 construct are almost completely inactive in our annealing assays. The Xl3 domain fits the dsRBD consensus sequence only at its C-terminal end, while the N-terminal region of this domain appears less conserved (Fig.
Our data also indicate that RNA binding is not a prerequisite for RNA strand annealing and that the two functions are obviously separable.
Figure 5. Annealing activity is specific for the dsRBD. FPLC-purified Xl2 (Xl2, lanes 1-5), His-tagged Xl2 purified by Ni-chelation chromatography (His-Xl2, lanes 6-10) and a mock protein preparation (lanes 11-13) were used for annealing assays. Final protein concentrations in each reaction are indicated in µM. For the mock reaction a protein preparation from empty plasmid-transformed E.coli was used, in parallel with the His-Xl2 preparation. After Ni-chelating chromatography the sample was dialysed, lyophilized and resuspended in the same volume as His-Xl2. Aliquots of 2, 1 and 0.5 µl of this preparation were added to the three control reactions. Annealing activity is observed with Xl2 and His-Xl2, while the mock preparation showed no RNA annealing. Lanes 14-20, controls.
DISCUSSION
The dsRNA-binding protein Xlrbpa is composed of three tandemly arranged dsRBDs. Many other dsRBD-containing proteins have additional catalytic or functional domains. In contrast, Xlrbpa seemingly lacks such additional regions, raising questions about the cellular function of the protein.
We have previously shown that Xlrbpa is associated with the majority of cellular RNAs, rRNAs and hnRNAs. Association of Xlrbpa with hnRNAs is reflected in a tight association of the protein with hnRNPs (11), suggesting that Xlrbpa might function as a general RNP. Additionally, Xlrbpa might play a role in facilitating RNA folding and stabilization of double-stranded regions. We could show here that Xlrbpa or even isolated dsRBDs can promote RNA strand annealing.
RNA-annealing activity has already been suggested for the dsRBDs of human dsRNA-activated, interferon-dependent kinase PKR. It was shown that binding of one dsRBD-containing molecule to a TAR-RNA region containing bulges facilitates binding of a second dsRBD-containing molecule, suggesting that binding of the first peptide induces a more uniform double-stranded conformation of the substrate molecule, thus facilitating binding of a second dsRBD (3). Our data are consistent with this interpretation, as we could show that a dsRBD facilitates RNA annealing and might therefore help to form and stabilize imperfectly paired, double-stranded regions.
Figure 6. RNA annealing of Xl2 can be competed by heparin. RNA-annealing assay performed in the presence of 20 µM Xl2 and various amounts of heparin. Lanes 1-8, decreasing amounts of heparin (amounts indicated in ng) were added to the annealing assay. Lanes 9-15, controls as in Figure 2. As little as 100 ng heparin can efficiently inhibit RNA strand annealing. Strand-annealing activity has also been shown for other RNA-binding proteins, like Tif3, eIF4B, YRA1 and a number of hnRNP proteins from HeLa cells (20-22). However, all these proteins contain different RNA-binding domains of the RRM or RGG type. In fact, fractionation of HeLa nuclei revealed nine major fractions with RNA-annealing activity, eight of which most likely contained previously known hnRNP proteins containing RRM or RGG RNA-binding domains (21). One fraction, however, contained an unidentified basic protein of 33 kDa, which was suggested to represent the splicing factor SF2. In the light of the fact that Xlrbpa shows RNA-annealing activity it seems more likely to us that this basic, 33 kDa protein previously found in HeLa cells is in fact the human homologue of Xlrbpa, rather than SF2. TAR-RNA binding protein (hsTRBP), the human homologue of Xlrbpa, is also basic in nature and has a mobility of exactly 33 kDa on SDS-PAGE (11). Our assays demonstrated a concentration dependence for RNA annealing of Xlrbpa or individual dsRBDs. Maximum annealing was achieved at a 200-1000-fold molar excess of protein over dsRNA substrate or at protein:nucleotide ratios ranging from 8:1 to 1:2, depending on the construct used. This suggests that dsRBDs work in a stoichiometric rather than catalytic fashion. However, it is also possible that our E.coli-produced proteins are not fully active, possibly due to misfolding of the proteins. In fact, RNA-binding assays performed with E.coli-produced Xlrbpa or Xl2 already indicated that only a small fraction of proteins, typically <1/106, are biologically active (6). It thus seems possible that much higher annealing activities could be achieved in vivo when Xlrbpa is properly folded. This view is also supported by our finding that Xl2 purified under native conditions by Ni-chelating chromatography shows stronger RNA-annealing activity than Xl2 protein purified by FPLC under denaturing conditions. Alternatively, it should be considered that our artificial substrates might not properly mimic the in vivo situation and that natural substrates might have a much higher affinity for Xlrbpa in vivo. Figure 7. RNA annealing does not depend on RNA binding. RNA-annealing assays were performed with serial (1:2) dilutions of His-Xl2 (lanes 1-5), His-Xl1 (lanes 6-10), His-F/L (lanes 11-15), His-R/P (lanes 16-20), His-Xl3 (lanes 21-25), His-2/1 (lanes 26-30) or the `mock' protein preparation (lanes 31-35). Lanes 36-40 are controls as in Figure 2. Protein concentrations ranged from 20 (highest concentration) to 1.2 µM (lowest concentration) for all proteins. The mock protein preparation was added from 5 to 0.3 µl. His-Xl2 and His-Xl1 showed comparable annealing activities, while the His-F/L and His-R/P mutants showed stronger annealing activity than the wild-type His-Xl2 construct. His-tagged Xl3 and chimeric construct His-2/1, in contrast, showed only minor annealing activity, while the control showed no annealing at all. The fact that the polyanion heparin can efficiently inhibit strand annealing by the dsRBD suggests that the basic nature of this motif is responsible for strand annealing to a large extent. This finding is also in good agreement with the fact that dsRBDs fail to bind single-stranded nucleic acids (1,3), which excludes the possibility that a dsRBD binds to single-stranded RNA thereby helping to expose bases required for binding, a mechanism, termed `matchmaker', which has been proposed for the annealing activities exhibited by several RRM-containing proteins. It should also be mentioned in this context that RNA-annealing assays performed in the presence of polylysine or the basic region of the transcription factor GCN4 also lead to strand annealing which can be easily competed by heparin, indicating that the presence of a basic protein is sufficient for RNA annealing in vitro (data not shown). Interestingly, RNA-annealing activity was also exhibited by the Xl1 domain. This dsRBD fails to bind dsRNA as an isolated domain (6) and has a relatively neutral pI of 7.8. Nonetheless, Xl1 contains a similar number of basic residues to Xl2. It is therefore possible that the basic charge of individual amino acids within dsRBDs rather than the net charge of the domain is crucial for RNA annealing. Alternatively, it is possible that the polyanion heparin negatively influences folding of the dsRBD or physically blocks the interaction of dsRBDs with RNA, thereby inhibiting strand annealing of the dsRBD. The fact that the chimeric 2/1 construct and the Xl3 domain fail to exhibit strand-annealing activity also indicates that not all dsRBDs promote strand annealing simply by virtue of their basic charge. At least the 2/1 construct has a basic pI of 9.24. Failure of this construct to promote RNA strand annealing therefore indicates that other factors, such as proper folding or the entire structure of a dsRBD, are important for this activity. Our finding that domains unable to bind dsRNA can promote RNA annealing indicates that RNA binding by dsRBDs is not a prerequisite for RNA annealing. In fact, it even appears that mutants F/L and R/P that fail to bind or show reduced RNA binding can exhibit a slightly stronger RNA-annealing effect than wild-type Xl2. It could be, for instance, that once annealing has taken place some dsRBDs bind to the duplex RNA molecule. A dsRBD that is bound to dsRNA might no longer promote annealing of other RNA molecules. dsRBDs unable to bind dsRNA, in contrast, could promote many annealing reactions, thereby leading to a stronger annealing effect. Taken together, we could show that full-length Xlrbpa and also an isolated dsRBD can promote RNA folding. This indicates that Xlrbpa which lacks any obvious catalytic or functional domains besides its three dsRBDs might function as an RNA chaperone in vivo by helping to fold and stabilize otherwise unfavourable or instable RNA structures. This idea is in good agreement with our finding that Xlrbpa is very abundant and can be found associated with almost all cellular RNAs (11). RNA chaperone-like activities have recently also been discussed for other RNA-binding proteins and might thus be a common feature of this class of proteins (17). Our finding that dsRBDs unable to bind dsRNA by themselves are able to promote RNA strand annealing also sheds new light on the function of `inactive' dsRBDs found in several proteins. Inactive domains might facilitate strand annealing, thereby helping to create the substrates required for RNA binding. Amongst several proteins, dsRBDs can also be found in ADAR1 and ADAR2, two RNA-editing enzymes that act exclusively on dsRNAs (24). It has been suggested that the site and extent of RNA editing mediated by these enzymes could be controlled by the stability of the double-stranded region defining the site of editing (24). It will be interesting to determine whether the annealing activity of dsRBDs described here also influences the extent of ADAR-mediated RNA editing by stabilizing substrate RNAs. It should be noted that strand annealing is probably not the only function of Xlrbpa. hsTRBP, the human homologue of Xlrbpa, has been shown to interact with and to modulate human PKR, a kinase activated by dsRNAs (13-15). Nonetheless, the mechanism of PKR-TRBP interaction is somewhat obscure at this point. While one set of experiments indicates that TRBP can only interact with PKR via an RNA linker, possibly competing for RNA binding, other experiments suggest a physical protein-protein interaction of the two dsRBD-containing proteins (14,15). Given that Xlrbpa and hsTRBP are almost exclusively composed of RNA-binding domains, it will be interesting to see whether the proposed protein-protein interaction can be attributed to a single region in TRBP. However, we favour the first hypothesis, in which hsTRBP (or Xlrbpa) directly competes with PKR for binding sites on dsRNA, thereby preventing activation of the kinase by sequestering its substrate RNAs.
ACKNOWLEDGEMENTS
The authors would like to thank members of the laboratory for helpful discussion of the manuscript. This work was supported by the Austrian Science Foundation with grants nos P9665 and P11444.
REFERENCES
This article has been cited by other articles:
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 19 Sep 1998
Copyright©Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
J. C. Fisk, M. L. Ammerman, V. Presnyak, and L. K. Read
TbRGG2, an Essential RNA Editing Accessory Factor in Two Trypanosoma brucei Life Cycle Stages
J. Biol. Chem.,
August 22, 2008;
283(34):
23016 - 23025.
[Abstract]
[Full Text]
[PDF]
M. L. Ammerman, J. C. Fisk, and L. K. Read
gRNA/pre-mRNA annealing and RNA chaperone activities of RBP16
RNA,
June 1, 2008;
14(6):
1069 - 1080.
[Abstract]
[Full Text]
[PDF]
M. Hallegger, A. Taschner, and M. F. Jantsch
RNA aptamers binding the double-stranded RNA-binding domain
RNA,
November 1, 2006;
12(11):
1993 - 2004.
[Abstract]
[Full Text]
[PDF]
C. Lu and N. Fedoroff
A Mutation in the Arabidopsis HYL1 Gene Encoding a dsRNA Binding Protein Affects Responses to Abscisic Acid, Auxin, and Cytokinin
PLANT CELL,
December 1, 2000;
12(12):
2351 - 2366.
[Abstract]
[Full Text]
B. Tian and M. B. Mathews
Functional Characterization of and Cooperation between the Double-stranded RNA-binding Motifs of the Protein Kinase PKR
J. Biol. Chem.,
March 23, 2001;
276(13):
9936 - 9944.
[Abstract]
[Full Text]
[PDF]
S. Li and A. E. Koromilas
Dominant Negative Function by an Alternatively Spliced Form of the Interferon-inducible Protein Kinase PKR
J. Biol. Chem.,
April 20, 2001;
276(17):
13881 - 13890.
[Abstract]
[Full Text]
[PDF]
M. A. Skabkin, V. Evdokimova, A. A. M. Thomas, and L. P. Ovchinnikov
The Major Messenger Ribonucleoprotein Particle Protein p50 (YB-1) Promotes Nucleic Acid Strand Annealing
J. Biol. Chem.,
November 21, 2001;
276(48):
44841 - 44847.
[Abstract]
[Full Text]
[PDF]
This Article 
Abstract
Print PDF (306K)
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (12)
Request Permissions
Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Articles by Hitti, E.
Articles by Jantsch, M. F.
Search for Related Content
PubMed
PubMed Citation
Articles by Hitti, E.
Articles by Jantsch, M. F.
Social Bookmarking
What's this?