Use of photoaffinity crosslinking and molecular modeling to analyze the global architecture of ribonuclease P RNA.

Abstract

Bacterial ribonuclease P (RNase P), an endonuclease involved in tRNA maturation, is a ribonucleoprotein containing a catalytic RNA. The secondary structure of this ribozyme is well established, but comparatively little is understood about its 3‐D structure. In this analysis, orientation and distance constraints between elements within the Escherichia coli RNase P RNA‐pre‐tRNA complex were determined by intra‐ and intermolecular crosslinking experiments. A molecular mechanics‐based RNA structure refinement protocol was used to incorporate the distance constraints indicated by crosslinking, along with the known secondary structure of RNase P RNA and the tertiary structure of tRNA, into molecular models. Seven different structures that satisfy the constraints equally well were generated and compared by superposition to estimate helix positions and orientations. Manual refinement within the range of conformations indicated by the molecular mechanics analysis was used to derive a model of RNase P RNA with bound substrate pre‐tRNA that is consistent with the crosslinking results and the available phylogenetic comparisons.

Guanosine 2-NH2 groups of Escherichia coli RNase P RNA involved in intramolecular tertiary contacts and direct interactions with tRNA

“…We have observed a variety of nucleotides in the L18 and P8 regions in this and our previous studies (Hardt et al+, 1995b at which Rp-phosphorothioate-, 29-deoxy-, and/or inosine modifications interfered with tRNA binding (summarized in Fig+ 4A)+ Interference effects may either reflect a perturbation of direct contacts to the tRNA or may affect tRNA binding indirectly by disrupting RNase P RNA structure+ Based on the phylogenetic evidence mentioned above (Brown et al+, 1996), interference effects observed in the L18 and P8 regions are likely to reflect an indirect effect on tRNA binding by means of perturbing the L18-P8 tetraloophelix interaction+ This is also in line with the finding that helix P8 is part of the four-helix junction that is thought to be the main interaction site for the T arm of tRNA molecules (Nolan et al+, 1993;Harris et al+, 1994;Pan et al+, 1995;Loria & Pan, 1997)+ Figure 4B,C summarizes a model of the L18-P8 interaction based on the proposed nucleoside triple structures of Jaeger et al+ (1994), but with the modifications outlined below+ First, we incorporated an additional interaction involving the tetraloop nucleotides G314 and A317 (Fig+ 4B) for the following reasons: (a) this type of interaction between the first-and last-loop nucleotide has recently been observed in two GNRA tetraloophelix contacts in crystal structures of the hammerhead ribozyme and the P4/P6 domain of the Tetrahymena ribozyme (Pley et al+, 1994;Cate et al+, 1996), and (b) this type of interaction is fully supported by our interference data (Fig+ 4A)+ Second, we propose hydrogen bonding between the 2-NH 2 group of G316 and O2 of U104 to account for the observed inosine interference effect at G316 (Fig+ 4A,C)+ The nucleotide quadruple and triple shown in Figure 4B,C would be supported by inosine interference effects at G95, 314, and 316, as well as by Rp-phosphorothioate interference at A317 Illustration of other sites of inosine interference+ A sequence ladder generated by iodine hydrolysis of Rp-AMPaS-modified RNase P RNA was included for the analysis of the 39-P4 region (G356)+ Nucleotide positions indicated by circles are sites of tRNA binding interference effects attributable to the inosine modification alone (black circles) or to both the Rp-phosphorothioate and the inosine modification (open circles); the triangle (G68) indicates a 100% Rp-phosphorothioate interference effect+ Interference effects were verified in three to six independent experiments+ Also, to compensate for fluctuations in the amount of material loaded on gels, corresponding hydrolysis bands in the tRNA-binding (complex) and nonbinding RNase P RNA lanes were normalized to the average intensity differences of three to four reference bands for which an interference effect was absent (see Materials and Methods)+ Gel resolution and thus assignment of interference effects was poor for G105-109 (not shown), G250/251 and G259-262+ However, in the latter two cases, interference effects appeared to be predominantly at G250 and G259/260+ A second example of the G95 region is included to illustrate that the site of interference is at...…”

“…Ribonuclease P (RNase P) is the ubiquitous processing enzyme that generates the mature 59-termini of tRNAs+ It is a ribonucleoprotein in vivo, with the exception of RNase P from spinach chloroplasts, whose composition was reported to be purely proteinaceous (Wang et al+, 1988)+ In vitro, RNA subunits of RNase P enzymes from Bacteria are catalytically active in the absence of the protein component (Guerrier-Takada et al+, 1983)+ They are the only known RNA catalysts naturally devoted to act in trans+ RNase P enzymes recognize the acceptor stem/ T arm modules of tRNA molecules (e+g+, McClain et al+, 1987;Forster & Altman, 1990a;Kahle et al+, 1990;Thurlow et al+, 1991;Schlegl et al+, 1992;Hardt et al+, 1993a;Carrara et al+, 1995;Yuan & Altman, 1995)+ Substrate recognition has been studied most extensively for bacterial RNase P enzymes+ Binding of tRNA to the catalytic RNA mainly relies on tertiary contacts between functional groups of the two RNA molecules+ One region of Watson-Crick-type base-pairing, involving the two consecutive cytosines of tRNA CCA-termini and two guanosines in the internal L15/16 loop of Escherichia coli-like "type A" (Haas et al+, 1996) RNase P RNAs, has been indicated by genetic studies (Kirsebom & Svärd, 1994)+ Five functional groups in the T stem-loop of yeast tRNA Phe , including four ribose 29-hydroxyls and the 4-amino group of the conserved C56, were shown to be crucial for binding to Bacillus subtilis RNase P RNA (Loria & Pan, 1997)+ Biochemical, mutational, and photoaffinity crosslinking data indicate that the P7-P11 region of bacterial RNase P RNA is the main interaction site for the T stem-loop of tRNAs (Nolan et al+, 1993;Harris et al+, 1994;Pan et al+, 1995;Loria & Pan, 1997)+ The 29-hydroxyl of tRNA nucleotide 62 was inferred to interact with A230 of B. subtilis RNase P RNA (Pan et al+, 1995)+ Binding interference studies have identified pro-Rp oxygens and 29-OH groups in E. coli RNase P RNA whose modification impairs tRNA binding (Hardt et al+, 1995b+ A subset of these functional groups are potential candidates for direct contacts to the tRNA moiety, particularly those residing in regions reported to form photoreactive crosslinks with tRNA molecules+ Another subset of interfering modifications is anticipated to affect tRNA binding indirectly by perturbing crucial tertiary interactions or by favoring alternative folding states+ Since the gel retardation assay applied in this and previous studies (Hardt et al+, 1995b selects for thermodynamically stable tRNA binding to E. coli RNase P RNA under moderat...…”

Derivation of the three-dimensional architecture of bacterial ribonuclease P RNAs from comparative sequence analysis

“…At the other end of P2, a single, invariant nucleotide (G19) is inserted between P2 and P3 (see Figure 1). In the absence of any other feasible alternative stacking, we conserved in our models the P2/P3 stack already present in previous models (Harris et al, 1994;Westhof & Altman, 1994). Therefore, the assembly of P19, P2 and P3 forms the ®rst continuous helical stack in domain II (see Figure 1 and 2).The choice of the P19/P2 stack in our models prevents the once proposed stacking of P1 on P2.…”

“…Domain II includes as well stems P15, P19, and extensions P16/P17/P6 and P18 (in type A) or P5.1, P15.1/P15.2 (in type B). In spite of the early recognition of both structural types, most attention has been focused on the type A sequence from E. coli, for which two structural models for the RNase P RNA-pre-tRNA complex were independently proposed (Harris et al, 1994;Westhof & Altman, 1994). The Harris & Pace model was primarily designed as a frame for inter-and intramolecular crosslinking data (Burgin & Pace, 1990;Nolan et al, 1993;Oh & Pace, 1994), whereas the Westhof & Altman model was based on alternative cross-linking data (Guerrier-Takada et al, 1989) and chemical modi®cation data (Lumelsky & Altman, 1988;Shiraishi & Shimura, 1988;Kirsebom & Altman, 1989;Knap et al, 1990;Talbot & Altman, 1994b;Westhof et al, 1996b).…”

RNA structure prediction: Progress and perspective

“…For initial 3D structure modeling, the 3D structures of some common RNA molecules such as tRNAs , [9] group I introns [10] and RNase P [11] have been successfully built by RNA structure experts. In recent years, a variety of computational models have been developed for predicting RNA 3D structures, [13,14,[49][50][51] which have been tabulated in Table 2.…”