Structural basis of a ribozyme's thermostability: P1–L9 interdomain interaction in RNase P RNA
Abstract
For stability, many catalytic RNAs rely on long-range tertiary interactions, the precise role of each often being unclear. Here we demonstrate that one of the three interdomain architectural struts of RNase P RNA (P RNA) is the key to activity at higher temperatures: disrupting the P1–L9 helix–tetraloop interaction in P RNA of the thermophile Thermus thermophilus decreased activity at high temperatures in the RNA-alone reaction and at low Mg2+ concentrations in the holoenzyme reaction. Conversely, implanting the P1–P9 module of T. thermophilus in the P RNA from the mesophile Escherichia coli converted the latter RNA into a thermostable one. Moreover, replacing the E. coli P1–P9 elements with a pseudoknot module that mediates the homologous interaction in Mycoplasma P RNAs not only conferred thermostability upon E. coli P RNA but also increased its maximum turnover rate at 55°C to the highest yet described for a P RNA ribozyme.
Keywords
INTRODUCTION
The ribonucleoprotein enzyme RNase P catalyzes tRNA 5′-end maturation in all organisms and organelles (Schön 1999). In bacteria, the RNA subunit (P RNA), ∼380 nucleotides (nt) in size and encoded by the rnpB gene, forms a specific complex with a single protein cofactor of ∼13 kDa. In vitro, bacterial P RNA has robust catalytic activity in the absence of the protein (Guerrier-Takada et al. 1983), but the latter is essential for function in vivo (Schedl et al. 1974; Gößringer et al. 2006). P RNAs from bacteria are subdivided into two distinct structural groups, termed type A (for “ancestral”) and type B (for “Bacillus”; Hall and Brown 2001). P RNAs of both types consist of two independent folding domains, the specificity (S-) and the catalytic (C-) domains (Loria and Pan 1996, 1997), which are oriented toward each other by three interdomain loop–helix contacts (Fig. 1A, L18–P8, L8–P4, P1–L9) in P RNAs of type A architecture. These long-range interactions, considered as architectural struts that stabilize the RNA's overall conformation, were originally identified by phylogenetic covariation analyses (Brown et al. 1996; Massire et al. 1997, 1998) and later confirmed in the crystal structure of P RNA from Thermotoga maritima (Torres-Larios et al. 2005). The importance of the aforementioned as well as other (intradomain) long-range tertiary interactions for P RNA function were also demonstrated experimentally (e.g., Darr et al. 1992; Pomeranz Krummel and Altman 1999). However, the specific functional contributions of the L18–P8, L8–P4, and P1–L9 contacts are poorly understood.
(A) Secondary structure presentation of T. thermophilus RNase P RNA according to Massire et al. (1998), with the three interdomain tetraloop–helix contacts L18–P8, L8–P4, and P1–L9 indicated by dashed lines. (B) P1–P9 region of the T. thermophilus wild-type P RNA (wt, left) and with the nucleotide exchanges in loop L9 that disrupt the P1–L9 interaction (mL9, right). (C) Processing assays with the two T. thermophilus P RNA variants. Assay conditions: 0.1–3 μM P RNA, <1 nM 5′-end labeled ptRNA, 100 mM Mg(OAc)2, 100 mM NH4OAc, 50 mM MES, and 2 mM EDTA (pH 6.0), and 37°C or 55°C as indicated. Error bars indicate standard deviations.
We recently reported that P RNAs from thermophilic bacteria share a 5′-GYAA L9 tetraloop and a P1 receptor site consisting of a G-C base-pair (bp) tandem, a combination not present in other bacteria (Marszalkowski et al. 2006). Also, helices P1 and P9 are stabilized in P RNAs from thermophiles by helix extension and/or deletion of nucleotide bulges. These observations prompted us to scrutinize the importance of the P1–L9 interaction in P RNA of the thermophile Thermus thermophilus, an A-type P RNA with robust in vitro activity at temperatures of up to 75°C (Hartmann and Erdmann 1991) and to compare it to that of the mesophile Escherichia coli.
RESULTS AND DISCUSSION
Activity of T. thermophilus P RNA with a disrupted P1–L9 interaction
Initially, we disrupted the P1–L9 interaction in T. thermophilus P RNA by a sequence change of the L9 loop (Fig. 1B, mL9, 5′-UUCG). At 100 mM Mg2+ in RNA-alone single turnover reactions, the mutant showed a defect at 55°C, but not at 37°C (Fig. 1C). This was caused mainly by a two- to threefold increase in K m(sto) (K m(sto), the single turnover K m, describes the enzyme concentration at which the half maximum rate under conditions of [E] >> [S] is achieved) (Table 1). An even more pronounced defect (up to 24-fold decrease in cleavage rate) was seen in activity assays with reconstituted RNase P holoenzymes at 37°C and 2.0 or 4.5 mM Mg2+ (Table 2).
Summary of kinetic data of RNA-alone activity assays
k obs [min−1] in RNase P holoenzyme assays of T. thermophilus and E. coli wild-type (wt) and mutant P RNAs
Role of the P1–L9 interaction in folding of T. thermophilus P RNA
Effects of the L9 mutation on P RNA folding were analyzed by native PAGE (Buck et al. 2005). While two main conformers were observed for T. thermophilus wild-type (wt) P RNA (Fig. 2A, bands 1,2), the L9 mutation indeed impeded formation of the faster migrating conformer 2, regardless of the presence or absence of P protein. To further characterize the two conformers of T. thermophilus wt P RNA, we eluted them separately from a native polyacrylamide gel at 4°C and reloaded aliquots of each eluate onto another native gel. Under these conditions, with the temperature not exceeding 15°C during elution and reelectrophoresis, both conformers did not reequilibrate to a substantial extent (Fig. 2B). In parallel, equal amounts of the two eluted conformers were tested in the holoenzyme reaction, performed at 22°C to mimize conformational reequilibration while maintaining high rates of substrate conversion. Approximately 15-fold higher activity (data not shown) was observed with the eluted conformer 2 (Fig. 2B, lane b) relative to eluted conformer 1 (Fig. 2B, lane a).
(A) Native PAGE analysis of T. thermophilus P RNA and its L9 mutant in the presence (+) or absence (−) of P protein. Trace amounts (≤50 fmol) of 3′-end labeled P RNA were incubated in buffer KN4.5 for 5 min at 55°C followed by 50 min at 37°C, and either with our without the addition of 40 nM RNase P protein for another 15 min at 37°C. Samples were loaded and run on 11.25% polyacrylamide gels in 1× THE buffer supplemented with 100 mM NH4OAc and 4.5 mM MgCl2 (for further details, see Materials and Methods). (B) Stability of T. thermophilus wt P RNA conformers studied by reelectrophoresis after elution. T. thermophilus wt P RNA (10 nmol including 3′-end labeled P RNA) was preincubated (see Materials and Methods) and then loaded on a native 11.25% polyacrylamide gel as in A. The two conformers (bands 1,2) were then eluted separately at 4°C, and aliquots of each band were loaded onto another native 11.25% polyacrylamide gel. Electrophoreses were performed at ≤15°C. Under these conditions, with the temperature not exceeding 15°C during elution and reelectrophoresis, both conformers did not reequilibrate to a substantial extent.
The catalytic deficits of the L9 mutant RNA observed at low metal ion concentration in the holoenzyme reaction thus appear to be directly correlated to a folding defect. Accordingly, the L9–P1 interaction is crucial for folding into a conformation that either stabilizes the catalytic core of the enzyme and its metal ion binding sites or has increased affinity for the P protein and/or the substrate.
Role of the P1–L9 interaction in E. coli P RNA
We then investigated the role of the P1–L9 interaction in P RNA from the mesophile E. coli (Fig. 3A) by disrupting the P1–L9 interaction (Fig. 3B, E. coli mL9, 5′-UCCG mutation) as done previously (Pomeranz Krummel and Altman 1999). However, this neither substantially impaired the RNA-alone reaction at 37°C (in accordance with Pomeranz Krummel and Altman 1999) or 55°C nor the holoenzyme reaction (Fig. 3C; Tables 1, 2), thus contrasting our results for T. thermophilus P RNA. The mutation was further neutral to viability (Table 3): The E. coli L9 mutant rnpB gene provided on a low copy plasmid fully rescued growth of the E. coli P RNA mutant strain BW (Wegscheid and Hartmann 2006). Our data thus indicate that the P1–L9 interaction does not make a significant contribution to overall stabilization of E. coli P RNA. We like to note that an E. coli rnpB gene with the same L9 mutation failed to rescue the phenotype of E. coli strain NHY322 rnpA49, for which the temperature sensitivity of a mutant P protein at 43°C is compensated by P RNA overexpression (Pomeranz Krummel and Altman 1999). We do not know the reasons for this discrepancy, but the different nature of the test strains precludes a direct comparison.
In vivo complementation screen in E. coli rnpB mutant strain BW
Secondary structure presentation of (A) E. coli wild-type (wt) P RNA and (B) alterations in the P1–P9 mutants thereof. mL9: Nucleotide changes in loop L9 that disrupt the P1–L9 interaction; P1P9 Tth: E. coli P1P9 replaced with the corresponding elements of T. thermophilus P RNA; P1P9 Mgen: E. coli P1P9 replaced with the pseudoknot from M. genitalium; P1P9 MgenA117C: variant of P1P9 Mgen, with a single nucleotide (A117C) exchanged. (C) Processing assays with the E. coli P RNA variants performed as described in the legend to Figure 1C.
Introduction of T. thermophilus and Mycoplasma genitalium P1–P9 elements into E. coli P RNA
We next replaced the P1–P9 elements of E. coli with those of T. thermophilus P RNA (Fig. 3B, E. coli P1P9 Tth). Compared to the E. coli wild type (wt), ribozyme activity of this hybrid P RNA at 37°C was also essentially unchanged, but relatively increased to a substantial extent at 55°C (Fig. 3C; Table 1). Again the mutant P RNA was functional in vivo (Table 3). Finally, an exceptional kind of P1–P9 interaction exists in some P RNAs of the genus Mycoplasma, which possess a P RNA of type B and where P9 and P1 juxtapose via a pseudoknot instead of the tetraloop–helix contact apparent in many type A P RNAs (Massire et al. 1997). Built into E. coli P RNA as an interaction module expected to be of particularly high stability (Fig. 3B, E. coli P1P9 Mgen), the pseudoknot of M. genitalium had little effect on the RNA-alone reaction at 37°C but substantially increased the cleavage rate at 55°C (Fig. 3C), the maximum cleavage rate obtained with this variant even exceeding that of E. coli P1P9 Tth (Table 1). In the holoenzyme reaction, cleavage rates were roughly twice that of E. coli wt P RNA (Table 2), and again, the hybrid P RNA was fully functional in vivo (Table 3). Finally, we intended to stabilize the interacting module even further by an A117 to C mutation in Mgen P9 to introduce an additional Watson–Crick base pair (Fig. 3B). This mutation generated a variant with the highest maximum turnover rate at 55°C yet described for a P RNA ribozyme (Table 1).
The fast substrate turnover catalyzed by the chimeric RNAs E. coli P1P9 Mgen and MgenA117C in the RNA-alone reaction at 55°C was associated with an increase in the single turnover K m(sto) compared with cleavage at 37°C, in particular for variant E. coli P1P9 MgenA117C (Table 1). This effect may be due to a loss in substrate binding affinity or to a reduction in the fraction of P RNAs that are competent to bind the substrate at 55°C. While this K m effect may have various reasons, the stronger K m effect for variant E. coli P1P9 MgenA117C (Fig. 3B) may be explained by the mutant C117–G108 bp constraining the orientation of A118, such that stacking of A118 on another bulged A residue in P11 (A233) (Fig. 3A; Lescoute and Westhof 2006) is weakened and becomes unstable at the higher temperature of 55°C.
CONCLUSIONS
From the presented data we conclude that the P1–L9 tertiary contact is crucial for folding and activity of P RNA from the thermophile T. thermophilus at high temperatures and low magnesium ion concentrations, conditions that require high intrinsic stability of the RNA. Our native PAGE analysis suggests that this interacting module represents one of the key anchoring points toward folding into the most active RNA conformer. These results extend present knowledge on the thermostability of P RNA from T. thermophilus: While Pan and coworkers identified 12 key base identities in the T. thermophilus S-domain that explain its stability increase over the corresponding E. coli domain (Baird et al. 2006), we were able to pinpoint an interdomain strut crucial for function of the entire RNA.
In contrast, the P1–L9 interaction appears degenerate in E. coli-like mesophilic P RNAs and could be completely disrupted without compromising function in vitro as well as in vivo. Yet, though not made use of in E. coli, in this P RNA the location of the P1–L9 interaction module has nonetheless remained a key position for placing an architecturally highly effective interdomain strut: If building in a strongly interacting P1–L9 module, E. coli P RNA gains substantially in activity at high temperatures. The architectural framework governing P RNA overall statics has thus been preserved between the thermophilic and mesophilic P RNA.
Finally, our findings identify the P1–L9 receptor/tetraloop of T. thermophilus and the corresponding pseudoknot from M. genitalium as (thermo)stabilization modules for complex RNAs that comply with the concept of modular “plug and play” interchangeability (Costa and Michel 1995, 1997; Qin et al. 2001). They may thus expand the spectrum of modules that can be exploited to stabilize larger RNAs for structural analysis (Ferré-D'Amaré et al. 1998).
MATERIALS AND METHODS
Cloning of transcription templates and complementation plasmids
All plasmids were constructed by standard PCR and cloning techniques.
Preparation and labeling of substrate and P RNA
ptRNA substrate and P RNAs were obtained by run-off transcription with recombinant bacteriophage T7 RNA polymerase and were 5′- and 3′-end labeling essentially as described (Heide et al. 1999; Busch et al. 2000). The T. thermophilus ptRNAGly substrate was transcribed from plasmid pSBpt3′HH linearized with BamHI (Busch et al. 2000); E. coli wild-type P RNA was transcribed from plasmid pJA2′ linearized with FokI (Vioque et al. 1988; Busch et al. 2000), T. thermophilus P RNA variants from plasmid pUC19TthrnpBwt and pUC19TthrnpBmL9 linearized with EheI; all E. coli mutant P RNAs were transcribed from respective pUC19-derivatives linearized with EcoRI (E. coli mL9, E. coli P1P9 Tth, E. coli P1P9 Mgen, and E. coli P1P9 MgenA117C). The gift of pUC19TthrnpBwt from Andreas Werner, group of Eric Westhof, IBMC Strasbourg, is gratefully acknowledged.
RNase P RNA and holoenzyme activity assays
RNase P RNA-alone activity assays were performed in 100 mM Mg(OAc)2, 100 mM NH4OAc, 50 mM MES, and 2 mM EDTA (pH 6.0), with trace amounts of 5′-end labeled substrate and P RNA concentrations in the range of 0.1–3 μM. Before starting the reaction, substrate and P RNA were preincubated separately in the assay buffer (ptRNAGly substrate for 5 min at 55°C and 20 min at 37°C, P RNAs for 5 min at 55°C and 55 min at 37°C). Reactions were run for 5–30 min at 37°C or 55°C as indicated. Activity of RNase P holoenzymes was measured in buffer KN (20 mM HEPES, 150 mM NH4OAc, 2 mM spermidine, 0.05 mM spermine, 4 mM β-mercaptoethanol at pH 7.4 at 37°C) containing 2 or 4.5 mM Mg(OAc)2 (buffer KN2 and KN4.5; Wegscheid and Hartmann 2006), at concentrations of 10 nM P RNA, 100 nM substrate, and 40 nM P protein (recombinantly expressed and purified exactly as in Marszalkowski et al. 2006). Prior to the reaction, substrate and P RNA were preincubated in the reaction buffer (substrate as above, P RNA for 5 min at 55°C and 50 min at 37°C); P protein was then added to the P RNA and preincubation continued for another 5 min at 37°C, after which substrate was added. Analysis of cleavage reactions and data evaluation were carried out as described (Busch et al. 2000).
Folding analysis by native PAGE
Folding analyses were performed essentially as described (Wegscheid and Hartmann 2006). In Figure 2A, trace amounts (≤50 fmol) of 3′-end labeled P RNA were incubated in buffer KN4.5 (see above) for 5 min at 55°C followed by 50 min at 37°C and either with or without the addition of 40 nM RNase P protein for another 15 min at 37°C. After addition of an equal volume of loading buffer (10% [v/v] glycerol, 4.5 mM MgCl2, 0.025% [w/v] each bromophenol blue and xylene cyanol), samples were run on an 11.25% (v/v) polyacrylamide gel containing 66 mM HEPES, 33 mM Tris, 0.1 mM EDTA, 100 mM NH4OAc, and 4.5 mM Mg(OAc)2 (pH 7.4) (buffer system adapted from Buck et al. 2005), with the gel temperature not exceeding 15°C. RNA bands were visualized with a Bio-Imaging Analyzer (FLA 3000-2R, Fujifilm). For reelectrophoresis and activity testing of T. thermophilus wt P RNA conformers (Fig. 2B), 20 nM of unlabeled P RNA plus 20 × 103 Cerenkov cpm of 3′-end labeled P RNA per gel lane were preincubated in buffer KN4.5 for 10 min at 55°C in a volume of 7 μL; after addition of 7 μL loading buffer (see above), samples were subjected to native PAGE as described above. After electrophoresis and 1 h of phosphorimager plate exposition, the two conformers were excised separately from the gel and eluted in 50 μL of buffer KN4.5 (see above) without Mg2+ by shaking overnight at 4°C. To analyze stability of the conformers, eluted material from each band was subsequently supplemented with loading buffer and run on a second native polyacrylamide gel (Fig. 2B). To compare the conformers for differences in catalytic activity, ∼3.5 nM of each eluted conformer were used in holoenzyme cleavage assays performed under multiple turnover conditions (reaction buffer KN4.5, 100 nM ptRNA, 40 nM B. subtilis RNase P protein). Assays were carried out at 22°C without any preincubation steps to eliminate potential heat-induced conformational changes.
In vivo complementation assays in E. coli strain BW
In vivo complementation assays were performed in the E. coli rnpB mutant strain BW with derivatives of the low copy plasmid pACYC177 precisely as described (Wegscheid and Hartmann 2006).
ACKNOWLEDGMENTS
This work was supported by the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie.
Footnotes
-
Reprint requests to: Dagmar K. Willkomm, Institut für Pharmazeutische Chemie, Philipps-Universität Marburg, Marbacher Weg 6, D-35037 Marburg, Germany; e-mail: willkomm{at}staff.uni-marburg.de; or Roland K. Hartmann, Institut für Pharmazeutische Chemie, Philipps-Universität Marburg, Marbacher Weg 6, D-35037 Marburg, Germany; e-mail: roland.hartmann{at}staff.uni-marburg.de; fax +49-6421-2825854.
-
Article published online ahead of print. Article and publication date are at http://www.rnajournal.org/cgi/doi/10.1261/rna.762508.
-
- Received August 2, 2007.
- Accepted September 10, 2007.
- Copyright © 2008 RNA Society
