|Home | About | Journals | Submit | Contact Us | Français|
To satisfy the high demand for ribosome synthesis in rapidly growing eukaryotic cells, short duplexes between the U3 small nucleolar RNA (snoRNA) and the precursor ribosomal RNA (pre-rRNA) must form quickly and with high yield. These interactions, designated the U3-ETS and U3-18S duplexes, are essential to initiate the processing of small subunit rRNA. Previously, we showed in vitro that duplexes corresponding to those in Saccharomyces cerevisiae are only observed after addition of one of two proteins: Imp3p or Imp4p. Here, we used fluorescence-based and other in vitro assays to determine whether these proteins possess RNA chaperone activities and to assess whether these activities are sufficient to satisfy the duplex yield and rate requirements expected in vivo. Assembly of both proteins with the U3 snoRNA into a chaperone complex destabilizes a U3-stem structure, apparently to expose its 18S base-pairing site. As a result, the chaperone complex accelerates formation of the U3-18S duplex from an undetectable rate to one comparable to the intrinsic rate observed for hybridizing short duplexes. The chaperone complex also stabilizes the U3-ETS duplex by 2.7 kcal/mol. These chaperone activities provide high U3-ETS duplex yield and rapid U3-18S duplex formation over a broad concentration range to help ensure that the U3-pre-rRNA interactions limit neither ribosome biogenesis nor rapid cell growth. The thermodynamic and kinetic framework used is general and thus suitable to investigate the mechanism of action of other RNA chaperones.
RNA chaperones have long been recognized as proteins that help RNA trapped in a nonfunctional conformation to adopt its functional form by using activities such as RNA annealing, strand-exchange and duplex destabilization1. Activity is not limited to aiding transitions only from the “misfolded” to the “folded” form but also from one functional form to another. Alternate RNA conformations often represent subsequent steps along a reaction pathway in processes like pre-mRNA splicing or ribosome biogenesis. RNA chaperones are often obligatory when cellular demands dictate that these steps along the reaction pathway occur quickly and efficiently. To assess how these demands are met we investigate how RNA chaperones mediate interactions between U3 small nucleolar RNA (snoRNA) and the precursor ribosomal RNA (pre-rRNA) in eukaryotic ribosome biogenesis (reviewed in refs. 2–4), a process essential to cellular growth and linked to cancer (reviewed in ref. 5).
Fast and efficient initiation of small ribosomal subunit (SSU) biogenesis is needed to supply the hundreds to thousands of ribosomes per minute required by rapidly growing eukaryotic cells. Formation of two short duplexes between the U3 snoRNA and the pre-rRNA, designated U3-ETS and U3-18S, is a prerequisite for the endonucleolytic cleavages that initiate SSU biogenesis6–11. These cleavages liberate the 18S precursor from the transcribed pre-rRNA, which embeds the 5.8S, 18S and 25S–28S rRNAs between internal and external transcribed spacers, ITSs and ETSs, respectively (Fig. 1).
Formation of both the U3-ETS and U3-18S duplexes docks the U3 snoRNA and its associated proteins, designated the SSU processome12–14, onto the pre-rRNA in a manner expected to recruit the as-yet unidentified U3-dependent endoribonuclease(s) for cleavage at A0, A1 and A2 (Fig. 1). Cleavage at A2 releases the 18S precursor from the pre-rRNA and is observed by electron microscopy to occur during pre-rRNA transcription with an estimated half-life of ~85 s in vivo15. As a prerequisite for cleavage, the U3-pre-rRNA duplexes are expected to form even faster. We also expect a high duplex yield (> 90%) because duplex formation is essential for pre-rRNA processing and growth6–11.
To achieve sufficient duplex yield and formation rates, RNA chaperones are needed to overcome two limitations (Fig. 1b). First, yield is limited by thermodynamic instability of the short U3-ETS duplex, made of only 10 base pairs. Second, a kinetic barrier limits formation of the other hybrid: the U3-18S duplex. Before the U3-18S duplex can form, the conserved box A/A’-stem structure must unfold to expose its base-pairing site. Formation of the U3-18S duplex thus occurs via two steps: unfolding and hybridization.
Previously, we showed with qualitative in vitro assays using minimal substrates16 that the U3-18S and U3-ETS duplexes corresponding to those in Saccharomyces cerevisiae are only observed after addition of one of two proteins, Imp3p or Imp4p, presumably by overcoming these two limitations. Both proteins are part of the SSU processome, required for the U3-dependent cleavages and thus essential12,17. Our findings on these S. cerevisiae proteins are expected to apply to other eukaryotes because the pre-rRNA processing pathways, including the U3-pre-rRNA base pairing potential18–20, and the associated trans acting factors, including Imp3p and Imp4p, have counterparts in higher eukaryotes21.
Imp3p and Imp4p share the same minimal U3 binding site16, U3 MINI, raising the possibility that they assemble into a ternary complex. Does this ternary complex form and if so does it possess chaperone activities sufficient to satisfy the in vivo requirements for rapid formation and high yield of the U3-pre-rRNA duplexes?
In this report, we address these questions by developing fluorescence based and other assays to ascertain the magnitude of the limits to U3-pre-rRNA yield and formation rate and the extent to which Imp3p and Imp4p overcome these limitations using minimal substrates. We demonstrate assembly of Imp3p, Imp4p and U3 MINI into a ternary complex and show that it does not alter the association rate constant of either the U3-18S or U3-ETS duplex. Rather, assembly of the complex removes the kinetic unfolding barrier to expose the U3 MINI bases to permit apparently spontaneous formation of the U3-18S duplex. The assembled complex increases the stability of the U3-ETS duplex by 2.7 kcal/mol (20 °C), thereby increasing the yield of this short duplex. Estimates based on our findings show that activities of this assembly, designated the chaperone complex, are needed to satisfy the in vivo demands for rapid formation and high yield of the U3-pre-rRNA duplexes.
The distance dependent nature of fluorescence resonance energy transfer (FRET) is ideally suited to monitor assembly of the chaperone complex and formation of the U3-18S and U3-ETS duplexes. In our steady-state FRET (ssFRET, i.e. with continuous illumination and observation) assays one molecule is labeled with the donor fluorescein (Fl) and its potential partner with the acceptor tetramethylrhodamine (Rh). When partner macromolecules interact they produce a ssFRET value above background if the fluorophore pairs are sufficiently close (between ~15 and 80 Å). To confirm that the fluorescent labels do not interfere with binding activity, we determined that the RNA-protein Kd values using fluorescently labeled molecules (data not shown) were within a factor of two of those measured previously with radiolabeled RNA and unlabeled protein16. To determine the duplex association (kon) and dissociation (koff) rate constants we monitored the signal change associated with the donor emission because it is larger than that of the acceptor emission. This phenomenon is due in part to more FRET-independent crosstalk from the donor to the acceptor than vice versa because of the asymmetry of their emission peaks. Three lines of evidence provide confidence that the Fl signal monitors duplex formation in accord with FRET. First and foremost, addition of acceptor containing RNA molecules results in a decrease in the Fl peak emission with concomitant increase in Rh peak emission for each case studied, whereas addition of an unlabeled partner to the Fl-labeled U3 MINI does not quench donor emission (Supplementary Fig. S1). Second, duplex kon values were within a factor of two of those determined with fluorophores attached to different sites on the RNA substrates (data not shown). Third, the Kd calculated by dividing U3-18S duplex koff by its corresponding kon is within a factor of two of the mean Kd value measured by electrophoretic mobility shift assays (Table 1; Supplementary Fig. S2).
To distinguish the U3-ETS duplex parameters from those of the U3-18S duplex, the former are designated henceforth kon (ETS), koff (ETS) and Kd (ETS) and the latter kon (18S), koff (18S) and Kd (18S).
Our previous findings16 showed that Imp3p and Imp4p share the same minimal RNA binding site, U3 MINI. To test whether these proteins assemble with U3 MINI into a ternary chaperone complex, we used ssFRET assays in which Fl-labeled Imp3p (Fl-Imp3p) contained the donor, Rh-labeled Imp4p (Rh-Imp4p) contained the acceptor, and U3 MINI was unlabeled (Fig. 2). Addition of Rh-Imp4p to a pre-formed binary complex of Fl-Imp3p and U3 MINI resulted in a FRET efficiency (EFRET, calculated as described in Materials and Methods) value of 0.26 ± 0.01, significantly above background (0.03), consistent with assembly (Fig. 2a, open bar). In contrast, only background EFRET values were observed when Fl-Imp3p was added to either Rh-Imp4p (Fig. 2a, hashed bar) or a pre-formed binary complex of Rh-Imp4p and U3 MINI (Fig. 2a, black bar). Addition of unlabeled Imp4p to a preformed binary complex of Fl-Imp3p and U3 MINI showed no signal change and thus confirmed that the observed FRET signal results from the proximity of the Rh and Fl labels and not protein binding (Fig. 2b). These findings support the notion that an RNA-dependent chaperone complex assembles from U3 MINI, Imp3p and Imp4p and lead to the hypothesis that Imp3p binds to U3 MINI before Imp4p.
To verify assembly of the chaperone complex with full length U3 snoRNA, metal affinity chromatography was used to capture N-terminal His6–tagged Imp3p (His6-Imp3p) in the presence of untagged Imp4p and full-length U3 snoRNA (Fig. 2c). Denaturing PAGE analysis of the loaded mixtures and eluted fractions after washes that include high salt (1 M NaCl) shows that His6-Imp3p and Imp4p associate with each other in the presence of U3 snoRNA (Fig. 2d, lane 2). To confirm that retention on the column arises from interaction with the tagged protein, we showed that neither the unlabeled Imp4p nor U3 snoRNA remain bound after the washes (Fig. 2d, lane 4).
Our in vitro assembly findings are consistent with previous immunoprecipitation studies using S. cerevisiae cell extracts, which showed that prior binding of Imp3p is needed to incorporate Imp4p into the SSU processome22. Such correlation between our in vitro studies and immunoprecipitation assays of others helps validate our in vitro system as biologically relevant.
To ascertain the limits to duplex yield and hybridization rate constants we investigated how addition of protein affects the duplex stability (Kd) and duplex kinetics (kon and koff). Evaluation of these effects will also discriminate between the six mechanistic models that are envisioned to overcome these limitations (Fig. 3). Given that Kd = koff/kon, we determined any two of these values, which are sufficient to calculate the third value for a single step model of reversible duplex formation as illustrated by the U3-ETS duplex (Fig. 1b). In contrast, formation of the U3-18S duplex is most readily modeled with two discernable steps: unfolding of the box A/A’ stem structure and subsequent hybridization (Fig. 1b). For the first (unfolding) step, we estimated the Keq equilibrium constant between U3 MINI and its unfolded form, designated U3 MINI*, and how protein binding affects this constant. For the second (hybridization) step in U3-18S duplex formation, appropriate conditions were used to determine kon (18S) and koff (18S) directly rather than Kd (18S). The binding affinity is expected to include contributions from both the unfolding and hybridization steps resulting in an apparent Kd (18S) (Fig. 1b).
Chaperone activity can mediate U3-18S duplex formation by affecting the unfolding step, the hybridization step or both. We begin by investigating the unfolding step (Fig. 4). To place an upper estimate on the free energy of unfolding the U3-stem structure we obtained reversible UV melting data for U3 MINI from S. cerevisiae (Fig. 4a). The melting temperature of 54 °C corresponds to a free energy of 4 kcal/mol with a Keq of 10−3 at 20 °C. As this stem structure is conserved among eukaryotes, it is expected to remain folded even at the growth temperature of vertebrates (37 – 42 °C) with only trace quantities (~0.1%) of U3 MINI*, the unfolded form of U3 MINI. To ensure rapid formation of the U3-18S duplex, helix destabilization activity is thus anticipated.
To test whether assembly of the chaperone complex opens up the U3-stem structure and thereby changes Keq, we used time-resolved FRET (trFRET). Unlike ssFRET, trFRET23,24 compares the nanosecond-scale donor fluorescence decay in the presence and absence of the acceptor to determine with high precision the distribution of distances separating the fluorophore pair. We measured trFRET of a doubly labeled U3 MINI with Fl at the 5’ end and Rh on the opposite side of the box A/A’ stem (Fl-U3 MINI-Rh, Fig. 4b) in the presence and absence of proteins and 18S (decay curves shown in Supplementary Fig. S3). Determination of the Fl-Rh distance distributions in U3 MINI alone showed that 93% of the RNA molecules yield a short (~19 Å) mean Fl-Rh distance (Fig. 4c, grey line), as expected from a donor and acceptor on opposite sides of an A-form helical RNA (Fig. 4d). The remaining 7% of RNA molecules reside in a conformation of larger and more broadly distributed Fl-Rh distances (Fig. 4c, grey dashed line), consistent with the presence of a small fraction of U3 MINI dimer (Fig. 4c, inset).
Upon addition of Imp3p to Fl-U3 MINI-Rh (Fig. 4c, dashed black line), the mean Fl-Rh distance increases by 13 Å from 19 Å to 32 Å with a concomitant sharpening of the distance distribution (Fig. 4c, compare the solid grey and dashed black lines). Subsequent addition of Imp4p and 18S results in only minor changes (Fig. 4c, compare the black line with the dashed and dotted black lines) supporting the view that Imp3p is primarily responsible for unfolding U3 MINI to U3 MINI*.
The 13 Å increase readily accommodates an open box A/A’ stem structure but not a fully extended U3 MINI* that could separate the Fl-Rh pair by as much as 100 Å, and thus abolish FRET. To account for the lack of change in distance distributions upon addition of Imp4p and 18S, our data are most consistent with a model in which the 3’ segment of U3 MINI loops back as shown in Fig. 4d, as a result of Imp3p binding. This arrangement remains unchanged upon addition of Imp4p and 18S. Our trFRET data suggest that assembly of the chaperone complex mediates the first unfolding step by opening up the box A/A’ stem structure.
To test whether the chaperone complex affects the second U3-18S hybridization step, we investigated how the kon (18S) and koff (18S) change upon addition of Imp3p and Imp4p (Fig. 5; Table 1). We determined the kon (18S) by monitoring the time-dependent donor quenching of 5’-Fl-labeled U3 MINI (Fl-U3 MINI) upon addition of equimolar amount of 3’-Rh labeled 18S (18S-Rh) in the presence of saturating amounts of protein. This stoichiometry was used to ensure a 1:1 donor-to-acceptor ratio for a maximum ssFRET signal change. Duplex formation with a kon (18S) of (7 ± 1) × 105 M−1s−1 was observed only after assembly of the chaperone complex (Fig. 5b, black squares, and c). To ensure that this rate directly monitors the bimolecular hybridization step we verified that this kon (18S) was the same, within error, as that determined using more conventional pseudo first-order conditions (excess 18S-Rh, Fig 5d).
In sharp contrast to rapid U3-18S hybridization in the presence of protein, duplex formation was not detectable in the absence of protein even when up to 1 µM concentrations of 18S-Rh were used (Fig. 5b, compare the traces with open and grey circles). Likewise, no shift was detected with electrophoretic mobility shift assays using up to 200 µM U3 MINI with trace amounts of 32P-18S (data not shown).
To place an upper estimate on kon (18S) after the box A/A’ stem has opened up, we used a fragment of U3 MINI, designated MINI-17, which retains only the 17 nucleotides involved in the U3-18S duplex, including the mismatches (the dashed box in Fig. 5a shows the MINI-17-18S duplex). The deleted flanking nucleotides of U3 MINI remove the 3’ half of the box A/A’ stem structure and thus eliminate the need to unfold this structure prior to U3-18S hybridization (Fig. 1b). In the absence of protein, the kon (18S) for MINI-17 hybridizing with 18S is (7 ± 1) × 105 M−1s−1 (Supplementary Fig. S4; Table 1), identical to the kon (18S) observed for U3 MINI in the presence of protein (Fig. 5c). The equivalence of these rate constants supports the view that protein binding has removed the barrier to U3-18S duplex formation by unfolding the box A/A’ stem structure to expose the base-pairing site.
To determine the duplex koff (18S), we chased the pre-formed fluorescently labeled U3-18S duplex with at least 100-fold excess of unlabeled 18S. The time-dependent exponential increase in Fl emission was used to determine that koff (18S) is (2 ± 1) × 10−3 s−1 in the presence of protein (Fig. 6a). In the absence of protein, koff (18S) was not measured because formation of this duplex was not observed. To estimate koff (18S) in the absence of protein, we therefore determined koff (18S) for the MINI-17 – 18S duplex; the observed rate constant of (1.0 ± 0.1) × 10−4 s−1 is 20-fold slower than the U3-18S duplex dissociation rate constant in the presence of protein (Fig. 6b).
Before comparing these different substrates (U3 MINI and MINI-17), it is useful to consider the step that limits formation of other short duplexes. Classic kinetic studies have shown that hybridization of complementary nucleic acid strands proceeds via two steps: nucleation and elongation25. Once diffusion juxtaposes bases from two complementary strands, formation of 3 to 4 contiguous base pairs is sufficiently long-lived to nucleate the process. Elongation completes hybridization of the remaining base pairs that flank the nucleation site. Nucleation, not elongation, limits hybridization of two complementary and unstructured RNA strands to form duplexes from 8 to ~20 base pairs in length26. Consequently, they share the same duplex kon of ~106 M−1s−1, independent of their sequence27,28. Equivalent kon values are observed for formation of two short duplexes studied herein: the U3-18S duplex in the presence of Imp3p and Imp4 and the MINI-17–18S duplex in the absence of protein. Given the common kon value it is reasonable to expect that formation of these duplexes is also limited by nucleation.
As a result, comparing how kon (18S) and koff (18S) differ for the U3-18S duplex in the presence of Imp3p and Imp4p and for the MINI-17–18S duplex in the absence of protein offers insight into the mechanism of the hybridization step. Addition of Imp3p and Imp4p does not change the duplex kon (18S) whereas koff (18S) increases by 20-fold, corresponding to a 1.7 kcal/mol destabilization of the U3-18S duplex product (Fig. 7, compare superimposed dotted green (MINI-17-18S duplex) and black (U3-18S duplex) lines). Of the six possible mechanisms, only product destabilization increases koff (18S) and Kd (18S) without changing kon (18S) (Fig. 3; Table 1). The kinetic findings provide evidence that Imp3p and Imp4p do not affect the forward hybridization barrier because kon (18S) remains unchanged but they do destabilize the product duplex after it is formed.
The findings from trFRET, UV melting and kinetic studies suggest that protein binding accelerates formation of the U3-18S duplex by unfolding U3 MINI to U3 MINI* (the first step) instead of stimulating annealing activity (the second step). In the absence of protein, two factors limit the amount of U3 MINI* and the subsequent U3-18S duplex (Fig. 7, grey dashed line). First, the 4 kcal/mol stability of the box A/A’ stem structure limits the percentage of U3 MINI* to ~0.1 % (Keq = 10−3; Fig. 4a). Second, entropy favors U3 MINI* refolding to U3 MINI rather than bimolecular hybridization. As a result, the protein free reaction is unfavorable. In contrast, trFRET data show that U3 MINI* is the only species detected upon addition of Imp3p (Keq 1), with negligible differences observed upon subsequent addition of Imp4p and 18S (Fig. 4c and d). By increasing Keq from 10−3 to 1 assembly of the chaperone complex unfolds U3 MINI into a stable U3 MINI* to accommodate annealing with 18S and ensures that the reaction proceeds energetically downhill from U3 MINI to U3 MINI* to the U3-18S duplex, in contrast to the protein free reaction (Fig. 7, compare grey dashed (no protein) and black lines (protein added)).
Unlike the kinetic barrier that prevents detectable U3-18S duplex formation in the protein free reaction (Fig. 5b), the yield of the other hybrid, the U3-ETS duplex, is limited by thermodynamic instability (Fig. 1b). Our previous qualitative findings showed that Imp3p and Imp4p increase the yield of the U3-ETS duplex16. To quantify the magnitude of this increase we determined the Kd (ETS) and kon (ETS) values by using electrophoretic mobility shift assays16 and ssFRET assays, respectively, in the presence and absence of Imp3p and Imp4p (Materials and Methods). Assembly of the chaperone complex decreases the Kd (ETS) by 100-fold from (7 ± 2) × 10−7 M to (7 ± 3) × 10−9 M (Fig. 8a and b; Table 2), which corresponds to an increase of 2.7 kcal/mol (20 °C) in duplex stability. We determined kon (ETS) by monitoring the time-dependent donor quenching of the 3’-Fl-labeled U3 MINI (U3 MINI-Fl) upon addition of an equimolar amount of 5’-Rh labeled ETS (Rh-ETS) in the presence and absence of saturating amounts of protein (Fig. 8c; Table 2). In contrast to changes in duplex affinity, kon (ETS) is the same in the presence ((5 ± 1) × 105 M−1s−1) and absence of protein ((6 ± 2) × 105 M−1s−1), within experimental error (P > 0.05) (a representative trace of the no protein reaction is shown in Fig. 8d). The equivalence of these kon (ETS) values to the intrinsic rate constant for formation of short duplexes27,28 supports the view that hybridization is unhindered even in the presence of Imp3p and Imp4p.
Upon assembly of the chaperone complex, the Kd (ETS) decreases by 100-fold and kon (ETS) remains unchanged, favoring a product stabilization mechanism over the competing alternatives (Fig. 3). A change in Kd rules out transition state stabilization, whereas no change to kon rules out substrate stabilization and destabilization as well as a combined mechanism. By multiplying Kd (ETS) and kon (ETS), koff (ETS) is predicted to increase, which rules out product destabilization. Product stabilization is the only model in which protein decreases Kd without changing kon. The absence of a change to kon (ETS) upon addition of protein reflects an unchanged hybridization barrier (Fig. 9). After hybridization the protein stabilizes this duplex by 2.7 kcal/mol to ensure high yield.
Our findings support a model in which the product U3-ETS duplex is stabilized by docking into a binding pocket created by Imp3p and Imp4p (Fig. 9). Duplex docking is expected to occur only after the duplex forms because the kon (ETS) is unaffected by the presence of protein (Table 2) and the U3 nucleotides involved in hybridization are accessible to ribonuclease digestion16. A concave binding pocket is an attractive possibility because it most readily accommodates the cylindrical shape of the A-form duplex product.
By extrapolating our findings using minimal substrates in vitro to the corresponding reactions occurring with full-length pre-rRNA and U3 snoRNA we can estimate whether the chaperone complex satisfies the in vivo demands for rapid formation and high yield of the U3-pre-rRNA duplexes. Addition of the extra pre-rRNA sequences and the numerous trans acting factors found in vivo will undoubtedly affect these results. However, given the many potential complications arising from misfolding of larger RNA substrates, it is important to first determine how RNA chaperones alter duplex formation rates and the yields of minimal RNA substrates.
To calculate duplex rates and yields that simulate in vivo conditions, the nucleolar concentrations of U3 snoRNA and the pre-rRNA are required. Even though these values are unknown, estimates are possible. High-resolution mapping of rDNA and U3 snoRNA territories in the nucleolus of S. cerevisiae using optical microscopy indicates volumes of 0.5 × 10−15 L and 1.5 × 10−15 L, respectively29. Given that about 4,000 copies of pre-rRNA30 and between 400 and 1,000 copies of U3 snoRNA31 are expected for rapidly growing cells, the concentration of the U3 snoRNA is between 0.4 and 1 µM and that of the pre-rRNA is ~13 µM. Undoubtedly, the concentration will not be homogeneous throughout the nucleolus; hence our calculations use a broad concentration range from 0.01 to 10 µM. It is also possible to approximate the yield of the U3-pre-rRNA duplexes in vivo. About 1 in 10 pre-rRNA transcripts are cleaved at A3 before A2 (and A0 and A1; Fig. 1a) resulting in a 23S intermediate rather than the standard 20S precursor (personal communication K. Karbstein, University of Michigan). Because U3-pre-rRNA hybridization is a prerequisite for the A0–A2 cleavages, it is reasonable to assume that these duplexes have not yet formed in the 23S intermediates. Given these considerations, we estimate that 90% of the pre-rRNA forms a duplex with the U3 snoRNA in vivo.
To assess whether the chaperone complex sufficiently accelerates the rate of U3-18S duplex formation, we calculated half-lives for the reaction as a function of substrate concentration (Fig. 10a is based on values in Table 1). As described in the introduction, formation of the U3-18S duplex is a prerequisite for the U3-dependent cleavages that release the 18S precursor from the pre-rRNA. In rapidly growing cells these cleavage events have an estimated half-life of ~85 s in vivo15. The prerequisite formation of the U3-18S duplex is thus expected to be even faster. In the absence of Imp3p and Imp4p, the formation of the U3-18S duplex is not observed. In sharp contrast, in the presence of protein, the half-life for duplex formation is less than 85 s when the pre-rRNA concentration exceeds 7 nM (based on the kinetic parameters in Table 1). This analysis supports the argument that Imp3p and Imp4p are necessary and sufficient to fulfill the need for rapid formation of this duplex in vivo.
Consistent with in vivo expectations, the presence of Imp3p and Imp4p ensures a high U3-ETS duplex yield over a broad concentration range of both substrates (U3 snoRNA and pre-rRNA) based on calculated percent yield (Fig. 10b and c). In the absence of protein, pre-rRNA and U3 snoRNA (assuming equimolar amounts) must exceed estimates of their nucleolar concentrations (>63 µM) to achieve high duplex yield (>90%). In contrast, lower substrate concentrations (> 0.63 µM), in line with in vivo estimates, are sufficient to ensure high U3-ETS duplex yield in the presence of Imp3p and Imp4p.
The U3-ETS and U3-18S hybridizations were modeled as separate bimolecular reactions because the pre-rRNA was divided into two minimal substrates (ETS and 18S); however, intramolecular reactions may also occur in vivo with full-length pre-rRNA (Fig. 1a). During pre-rRNA transcription, the U3-ETS duplex may hybridize first as a bimolecular reaction because the ETS site is transcribed before the 18S site. A stable U3-ETS duplex is needed for subsequent intramolecular U3-18S hybridization. The half-life for this intramolecular reaction may occur even faster than those in Fig. 10a due to higher effective concentration (lower entropic barrier). It is reasonable to assume that unfolding of the box A/A’ stem structure will still limit the U3-18S reaction in the absence of protein. Our in vitro studies provide evidence that the presence of Imp3p and Imp4p will alleviate this kinetic unfolding barrier to accelerate U3-18S hybridization and enhance the stability of the U3-ETS duplex.
In ribosome biogenesis, the U3/Imp3p/Imp4p chaperone complex is expected to position the SSU processome for the early pre-rRNA cleavage events that release the SSU precursor by stimulating docking between the U3 snoRNA and two complementary sites on the pre-rRNA: the U3-ETS and U3-18S duplexes (Fig. 1). To keep up with the high demand that rapidly growing cells have for producing ribosomes, formation of these duplexes has to be fast15 (half-life < ~85 s) and efficient6–11 (duplex yield > ~90%). In this study, we developed in vitro ssFRET and trFRET-based assays to demonstrate assembly of the chaperone complex (Fig. 2) and show that it possesses the RNA chaperone activities (Figs. 4– 6 and and8)8) necessary to satisfy these in vivo demands (Fig. 10).
In the absence of protein, formation of the U3-18S duplex is not observed in vitro but is expected to occur in two steps: unfolding of U3 MINI to U3 MINI* and hybridization (Figs. 1b and and7).7). Assembly of the chaperone complex, and particularly binding of Imp3p, destabilizes the conserved box A/A’ stem structure to expose its 18S base-pairing site by unfolding U3 MINI to U3 MINI* (Figs. 1b and and4).4). Unfolding this stem structure is sufficient to accelerate this reaction from an undetectable rate to the intrinsic hybridization rate for short duplexes (~106 M−1s−1).
The U3-18S chaperone activity is expected to be needed throughout the eukaryotic kingdom of life because the box A/A’ structure, the pre-rRNA base-pairing potential18–20 and the sequences of Imp3p and Imp4p are conserved17,21. In the absence of protein, formation of the U3-18S duplex is not observed. In contrast, in the presence of Imp3p and Imp4p, rapid formation of this duplex occurs over a wide range of substrate concentrations (Fig. 10a).
To ensure high yield of the U3-ETS duplex over a physiologically relevant range of pre-rRNA concentrations the chaperone complex stabilizes the duplex by decreasing the Kd (ETS) by ~100-fold (Figs. 8 – 10; Table 2). The chaperone complex binds to the U3-ETS duplex after it is formed to increase the duplex stability by 2.7 kcal/mol. Even though high nucleolar concentrations of pre-rRNA and U3 snoRNA are expected, the chaperone complex is needed to ensure high duplex yield when the concentration of both substrates is less than 630 nM (with equimolar substrate concentrations).
Product stabilization of the U3-ETS duplex may be needed throughout the eukaryotic kingdom. Previous studies of the chaperone activities of Imp3p and Imp4p indicate that they are not very sensitive to sequence variation of the U3-ETS duplex as long as hybridization potential is maintained9,16. Thus, in single-cell eukaryotes, where short U3-ETS base pairing is conserved, the chaperone complex ensures that high U3-ETS duplex yield limits neither ribosome biogenesis nor rapid cell growth. In the frog Xenopus laevis and other higher eukaryotes the one short U3-ETS duplex is replaced by two short duplexes separated by a number of nucleotides32. More study is needed to establish whether a chaperone complex stabilizes one or both of these duplexes, which are expected to be unstable due to their short lengths.
With quantitative data in hand it is possible to estimate the need for trans acting factors to release the SSU precursor from the SSU processome (U3-18S dissociation) or to recycle the SSU processome for another round of pre-rRNA processing (U3-ETS dissociation) (Fig. 1). Earlier qualitative analysis showed that removal of Imp3p and Imp4p from the U3-ETS duplex leads to duplex dissociation16. To estimate the need for “release factors” for dissociation of the ETS portion of the pre-rRNA, we calculated koff (ETS) values and corresponding dissociation half lives (Table 2). Addition of Imp3p and Imp4p increases the dissociation half-lives from ~ 2 s to ~200 s. Given the ~85 s half-life for the U3-dependent cleavages, protein is required to ensure that the U3-ETS duplexes remain intact long enough for cleavage to occur. For release of the 18S portion of the pre-rRNA, dissociation half-lives were calculated from observed koff (18S) values (Table 1). Addition of Imp3p and Imp4p decreases the dissociation half-life from ~2 hr to ~350 s. Protein addition thus reduces the need for helicase activity while ensuring that the U3-18S duplex remains intact long enough to release the SSU precursor. Possibly, proteins are used to temporally regulate the U3-18S duplex dissociation that releases the 18S nucleotides, which are part of a universal pseudoknot structure of mature ribosomes10. Releasing these 18S nucleotides at the proper time chaperones 18S folding by ensuring that this centrally located pseudoknot does not form prematurely and that these nucleotides are not trapped in an incorrect structure.
Proteins and RNA chaperones often do not possess specific activity; rather they target a large number of substrates. To avoid targeting correctly folded substrates, which may be harmful, protein chaperones preferentially bind to misfolded proteins by recognizing exposed hydrophobic patches that serve as distinctive features. How RNA chaperones avoid this problem is less clear because misfolded and correctly folded RNAs offer few if any distinguishing recognition features. A recent study by Russell and coworkers33 showed that for one substrate the increased stability of the correctly folded RNA compared to its misfolded counterpart protects it from the unwanted attention of RNA chaperones. Our studies illustrate another strategy, whereby the RNA chaperone acts site-specifically. Imp3p and Imp4p preferentially bind to the 5’ portion of the U3 snoRNA16, and because assembly is RNA dependent (Fig. 2), the chaperone complex targets a specific site.
Our kinetic and thermodynamic framework predicts that the chaperone complex is needed to accelerate U3-18S duplex formation from an undetectable rate to ~106 M−1s−1 and thereby ensure that this process limits neither ribosome biogenesis nor rapid cell growth. In contrast, the 2.7 kcal/mol of U3-ETS duplex stabilization provided by the chaperone complex will help, but may not be essential, to achieve the needed high duplex yield given the high concentrations of U3 snoRNA and pre-rRNA expected in the nucleolus. Determining how proteins change duplex kon, koff and Kd of RNA duplexes is a general strategy to investigate the mechanism by which other RNA chaperones satisfy the cellular demands for fast, efficient and site-specific structural rearrangements.
Details of the trFRET assays and derivation of equations for determining kon under non-pseudo first order conditions are found in Supplementary Materials and Methods.
All reactions were carried out in reaction buffer (20 mM Tris (pH 8.0), 100 mM KCl, 30 mM NH4Cl and 0.5 mM MgCl2) at 20 °C, unless otherwise specified. Prior to use, U3 MINI was refolded in a manner to maximize formation of the box A/A’ stem relative to dimer formation: samples were heated to > 90 °C for 2 min, followed by 10 min incubation on ice. The chaperone complex was assembled by incubating heat annealed U3 MINI with saturating amounts of Imp3p (≥ 1.5 µM) for 60 minutes. Subsequently, Imp4p was added at saturating amounts (≥ 0.5 µM) and allowed to reach equilibrium for at least 30 minutes.
In FRET based assays, we used Fl as the donor (excitation peak at 493 nm and emission peak at 520 nm) and Rh as the acceptor (excitation peak at 550 nm and emission peak at 580 nm). EFRET values were calculated as before34 by determining the Fl and Rh emission peaks heights and correcting them for direct acceptor excitation.
Imp3p and Imp4p from S. cerevisiae were recombinantly expressed and purified as before16. Imp3p and Imp4p were labeled with fluorescein (Fl) and tetramethyl rhodamine (Rh). Fl-5-maleimide was reacted with Imp3p to make Fl-Imp3p by forming thioether linkages with cysteine residues according to the recommendations of the manufacturer (Invitrogen). Likewise Rh-5-maleimide was reacted with Imp4p to make Rh-Imp4p.
The pHis6-Imp3p expression vector contained the Imp3 ORF and the N-terminal MGSSHHHHHHSSGLVPRGSH tag cloned into pET21d using XbaI and NotI restriction sites (New England Biolabs). His6-Imp3p was expressed at 16 °C overnight (16 – 17 hr) in E. coli BL21(DE3) cells supplemented with a vector coding for rare tRNAArg condons as before35. After cell breakage in 20 mM Tris (pH 8.0) and 200 mM NaCl, inclusion bodies were solubilized in 20 mM Tris (pH 8.0), 600 mM NH4Cl, 50 mM MES (pH 6.5) and 1 M Urea. The solubilized protein was purified via Talon resin (Clontech) by following the recommendations of the manufacturer and stored at a concentration of ≥70 µM in 20 mM Tris (pH 8.0) and 50 mM MgCl2.
All modified RNA molecules represent S. cerevisiae sequences and were synthesized and PAGE purified by Dharmacon to ensure complete label incorporation. To ensure that EFRET values reflect the distance between the donor and acceptor, both fluorophores must be freely rotating. To enable this mobility we attached each fluorophore via a six-carbon linker to its RNA oligomer and verified that the fluorescence polarization values were < 0.3. To label with Rh, RNA oligomers were synthesized with an amino group at the 3’ terminus, 5’ terminus or internally at C5 of one uracil residue and then reacted with Rh-succinimide (5-TAMRA, Invitrogen) according to the recommendations of the manufacturer. 5’-Fl was attached during synthesis. To label Fl at the 3’ terminus, RNA oligomers were synthesized with an amino group and then reacted with Fl-succinimide (6-FAM, Invitrogen) according to the recommendations of the manufacturer.
U3 MINI (5’-GGA CGU ACU UCA UAG GAU CAU UUC UAU AGG AAU CGU CAC UCU UUG ACU) represents nucleotides 4–50 of the U3 snoRNA with an additional 5’ terminal G, added originally to enable T7 in vitro transcription16. MINI-17 (5’ UAC UUC AUA GGA UCA UU) includes only the U3 nucleotides involved in 18S hybridization (the dashed box in Fig. 5a represents the MINI-17–18S duplex). The ETS site 5’-UCA AAG AGU G reflects nucleotides 470 – 479 of the pre-rRNA and the 18S site 5’-GGU UGA UCC UGC CAG UA reflects nucleotides 6 – 22 of the mature 18S rRNA.
The chaperone complex was made using His6-Imp3p, Imp4p and U3 snoRNA in a 50 µl volume and complexes were separated via metal affinity chromatography. The U3 snoRNA was produced by run-off transcription using a linearized plasmid DNA template16. The template was digested with DNase I (Promega) and nucleotides less than ~200 nucleotides, including abortive transcripts, were removed by RNeasy MINI Kit (Qiagen). Before using, the U3 snoRNA was refolded: the RNA was incubated at 100 °C (3 min), cooled at room temperature (3 min), then incubated with 10 mM MgCl2 at 55 °C (3 min). Prior to loading, 50 µl Co2+ Talon affinity resin (Clontech) was pre-equilibrated with ten column-volumes (cv) of reaction buffer. Protein and RNA complexes containing His6-Imp3p bound to the metal affinity resin, whereas excess untagged RNA and protein were eluted with a series of washes: 5 cv of reaction buffer, 2 cv of reaction buffer with increased ionic strength (1M NaCl added) and 3 cv of reaction buffer to restore the ionic strength of the column. The 1 M salt wash was necessary to eliminate nonspecific Imp4p-resin interaction. Complexes captured by the affinity resin were eluted with 4 cv of 300 mM imidazole in reaction buffer and resolved on a 15% denaturing SDS PAGE. Protein and RNA molecules were visualized by silver staining and ethidium bromide, respectively.
UV absorbance melting curves were collected at 260 nm from 10.5 to 84.5 °C in one-degree increments on a Varian Cary 1E UV-Visible spectrophotometer using unlabeled U3 MINI in reaction buffer. The absorption and temperature were processed using the program PRISM. The slope at the mid point of the transition in a plot of normalized A260 vs T (K) was used to estimate ΔH (at Tm) and ΔG° (at 20 °C) as previously described for a single transition36.
To monitor U3-18S duplex association rates, two methods were used. Stopped-flow kinetics were used to measure kon (18S) under pseudo first-order conditions, where [18S-Rh] greatly exceeds [Fl-U3 MINI] throughout the titration. Rapid mixing techniques were used to measure kon (18S) under nonpseudo first-order conditions, where [18S-Rh] equals [Fl-U3 MINI]. The latter condition ensures a 1:1 donor-to-acceptor ratio for a maximum ssFRET signal change. In the presence of protein 2, 5, 10 and 15 nM equimolar substrate concentrations were used. No reaction was detected in the absence of protein. For assays monitoring MINI-17–18S duplex formation 2, 5 and 10 nM equimolar substrate concentrations were used.
A SLM 8000C Spectrofluorimeter was used to collect data under nonpseudo first-order conditions in which the concentrations of the FRET donor and acceptor were equal. Excitation and emission slits were set to 2 mm and 4 mm, respectively. Integration time was set at 1 s and emission counts were recorded as a function of time to provide an optimal signal of at least 4,000 counts. The fractional decrease in the fluorescence of Fl-U3 MINI is proportional to the fractional increase in the amount of U3-18S duplex ([AB]apparent, M) formed and thus can be calculated using:
where f0 is the fluorescence at time t = 0, ft is the fluorescence at any time t, and A0 is the concentration of limiting substrate, either Fl-U3 MINI or 18S-Rh. In this case, since Fl-U3 MINI and 18S-Rh are equimolar, A0 can be either one. The rate of duplex hybridization is similar to a bi-molecular reversible reaction, which under equimolar concentrations has a form of the first term in equation 2a below (see Supplementary Material and Methods for derivation of equation 2a – 2e).
To account for photobleaching a second exponential parameter with kpb as the rate constant and d as the amplitude is added to Supplementary equation S9. The region of the data that defines koff (18S) overlaps the contribution from photobleaching and thus limits the ability to calculate koff (18S) from this data.
To determine the half-life t1/2 for the U3-18S duplex was set to 0.5 and solved for t
To verify that the fluorophores did not bias our measurements, hybridization rate constants were also determined for substrates labeled at different sites: unlabeled U3 MINI and doubly labeled 18S (Fl-18S-Rh). Reactions were initiated by adding a preformed ternary complex to Fl-18S-Rh.
For pseudo first-order conditions, increasing concentrations of 18S-Rh (150 nM to 3000 nM) were mixed in a Stopped-Flow Reactor (SLM Aminco FP-120) with a preformed ternary complex with a final concentration of 38 nM Fl-U3 MINI and saturating amounts of Imp3p and Imp4p. FRET dependent fluorescein quenching was recorded with minimum 1 ms resolution, at 520 nm as a function of time with AB2 Luminescence Spectrophotometer version 5.31 using a slit width of 2 nm for excitation (at 490 nm) and 16 nm for emission (at 520 nm). Data were averaged for a minimum of 8 shots per concentration of 18S-Rh. A representative averaged decay trace for 150 nM 18S-Rh is shown in the inset of Fig. 5d. The rate constant kobs was calculated by fitting decay traces to
where fis the fluorescence, A is f at infinite time, and B and C reflect the amplitude of each exponential with rate constants kobs and kobs2, respectively. The fast phase (kobs) increased with increasing concentration of 18S-Rh. The slow phase (kobs2) has a rate consistent with that of the photobleaching from the fluorophore and shows negligible dependence on concentration of 18S-Rh and was thus assigned kobs2 as the photobleaching rate constant.
The koff (18S) was determined by chasing a preformed duplex of Fl-U3 MINI and 18S-Rh with a large excess of unlabeled 18S (> 100-fold) in the presence of saturating amounts of protein. The chase resulted in an exponential growth of the fluorescein emission as RNA molecules with no FRET signal (Fl-U3 MINI–18S duplex and liberated 18S-Rh) replaced those with a FRET signal (Fl-U3 MINI–18S-Rh duplex). To determine koff (18S), the time-dependent increase in the Fl fluorescence was fit to
where f is the fluorescence at any time t, f0 is the fluorescence before chase is initiated and fmax + f0 is f at infinite time. The rate constant koff (18S) is determined under conditions where [18S] is high enough to ensure that this rate constant is independent of [18S]. The same method was used to determine koff (18S) for Fl-MINI-17 and 18S-Rh.
To determine Kd (ETS) values for the U3-pre-rRNA duplexes, U3 MINI was titrated with trace amounts of 32P 5’ end-labeled ETS either in the presence or absence of saturating amounts of Imp3p and Imp4p in reaction buffer. The RNA complexes were allowed to incubate for ~45 min and were resolved on 12 % (absence of proteins) and 6 % (presence of protein) non-denaturing PAGE gel (50:1 cross-linking ratio) for 45 min at 125 V at 4 °C. The gel was polymerized with 40 mM Tris-acetate (pH 8.0), 1 mM EDTA, 50 mM KCl and 2.5 % (v/v) glycerol and the same was used as the running buffer. Equal volume of 40 % (w/v) sucrose, 50 mM KCl and 80 mM Tris-acetate (pH 8.0) was added to the samples just before loading, to allow the sample to sink in the well. The bound and free 32P-ETS species were visualized by autoradiography using a Fuji Imaging plate (Bas 2024) and Typhoon 9400 (Amershan Biosciences, GE) and quantified using Image Quant 5.0. Kd values were determined by fitting the fraction of 32P-ETS bound as a function of [U3 MINI] in the absence or presence of protein using
where Ymax and Ymin are the fraction bound values at saturating and limiting [U3 MINI], respectively.
The quadratic solution to binding equations was used to calculate the duplex yield from Kd (ETS) values using
where Kd is Kd (ETS), concentrations of U3 MINI (M) and ETS (M) are represented as A and B, respectively, and min(A,B) is the lowest value of either A or B.
Hybridization rate constants of the U3-ETS duplex were measured under equimolar concentrations of donor and acceptor (i.e. equimolar substrate concentrations). In the presence of protein 10, 25 and 50 nM equimolar substrate concentrations were used for the assay; in the absence of proteins 1, 2, 5, 25 and 50 nM equimolar substrate concentrations were used. Ten fold concentrated Rh-ETS was added to Fl-U3 MINI with either saturating amounts of Imp3p and Imp4p or their buffers with no protein such that the final concentrations of substrates are equal. As with the determination of kon (18S) (above), the ssFRET dependent fluorescein quenching was converted into the appearance of [AB]apparent using equation (1) and fit to equation (2a) to determine kon (ETS).
To ascertain that the fluorophores did not interfere with activity, we determined kon (ETS) using 10 nM Fl-ETS-Rh and using substrates in which the donor and acceptor labels on the two RNA substrates were exchanged.
All equations were fit using least squares by PRISM5. The mean and standard deviation reported for the kinetic and thermodynamic values are calculated from at least three measurements.
We are grateful to D.H. Harrison and K.E. Neet for valuable advice and discussion, J.A. Piccirilli for use of his UV spectrophotometer and M.J. Plantinga, M. Tata and M. Havens for technical assistance and discussion. This work was supported by grants from the US National Institutes of Health (GM070491) to C.C.C and (GM062357) to N.G.W.