Mutational Analysis of the Joining Regions Flanking Helix P18 inE. coliRNase P RNA

J. Mol. Biol. (1996) 259, 422–433

Mutational Analysis of the Joining Regions Flanking Helix P18 in E. coli RNase P RNA Wolf-Dietrich Hardt and Roland K. Hartmann* Institut fu¨r Biochemie Freie Universita¨t Berlin Thielallee 63, 14195 Berlin FRG

We have studied variants of Escherichia coli RNase P RNA with base exchanges in the joining regions flanking helix P18, which form part of the ribozyme core structure. Mutant RNase P RNAs were analyzed for: (1) specific tRNA binding by gel retardation; (2) catalytic performance in single turnover reactions; (3) structural perturbations utilizing Pb2+-induced hydrolysis; and (4) in vivo function by complementation analysis in E. coli RNase P mutant strains. Our in vitro experiments revealed that the base moieties of nucleotides (nt) 303 and 331 to 333 neither significantly contribute to tRNA binding or structural stabilization of RNase P RNA nor to active site chemistry. Single base exchanges at nt 300, 301 and 330 reduced tRNA binding, while having little effect on the catalytic rate, which demonstrates that these nucleotides are involved in forming the high affinity (pre-)tRNA binding site. In contrast, point mutations at the strictly conserved positions nt 328, 329, 334 and 335 reduced tRNA binding affinity as well as the catalytic rate, suggesting that these mutations additionally disrupted important interactions in the catalytic center. Probing by Pb2+ revealed that particularly the mutations that affected catalytic function most strongly perturbed a more extended region (nt 248 to 335) known to be involved in tRNA binding. Under high salt conditions (r0.8 M NH4 +), catalytic defects of the mutant RNase P RNAs were much less pronounced, suggesting that structural perturbations leading to increased electrostatic repulsion between phosphate groups were the main cause for observed functional defects. Only mutant C334 retained a largely increased pre-steady-state Km(pss) under high salt conditions. We conclude that the base at position 334 is directly involved in a contact crucial to pre-tRNA binding. A complementation analysis demonstrated the important role in vivo of the joining regions flanking helix P18. None of the bases could be mutated without affecting bacterial viability. 7 1996 Academic Press Limited

*Corresponding author

Keywords: E. coli ribonuclease P RNA; mutational analysis; gel retardation; Pb2+-induced hydrolysis; in vivo complementation studies

Present addresses: W.-D. Hardt, School of Medicine, Department of Microbiology, SUNY at Stony Brook, Stony Brook, NY 11794-5222, USA. R. K. Hartmann, Medizinische Universita¨t zu Lu¨beck, Institut fu¨r Biochemie, Ratzeburger Allee 160, 23538 Lu¨beck, FRG Abbreviations used: PCR, polymerase chain reaction; tRNAGly, glycine-specific transfer RNA; rnpB, rnpA, genes encoding the RNA (rnpB) and protein (rnpA) subunits of E. coli RNase P; pre-tRNA, precursor tRNA; wt, wild-type; kreact , Km(pss), single turnover (pre-steady-state) kinetic constants (ES); kreact , maximal cleavage rate for the single turnover reaction at saturating enzyme concentrations; Km(pss) , enzyme concentration at the half-maximal rate of cleavage; appKd = apparent equilibrium dissociation constant; OAc, acetate; DEPC, diethylpyrocarbonate; nt, nucleotide(s). 0022–2836/96/230422–12 $18.00/0

Introduction Ribonuclease P (RNase P) is a ubiquitous ribonucleoprotein particle, which cleaves tRNA precursors to generate their mature 5' termini. In bacteria it is composed of a large RNA subunit, approximately 400 nt in size, and a small basic protein of approximately 120 amino acid residues (Brown & Pace, 1992). RNA subunits of bacterial RNase P enzymes were shown to be catalytically active in the absence of protein components (Guerrier-Takada et al., 1983). So far, only little is known about functional groups of RNase P RNA that interact with substrate RNAs or constitute the active site. Phylogenetic analyses and several lines of biochemical evidence have indicated that a sub7 1996 Academic Press Limited

RNase P RNA Mutations

region of bacterial RNase RNA, corresponding to nt 300 to 335 in Escherichia coli RNase P RNA, is critical for enzyme function: (1) nt G300, A328, G329, A330, A334 and U335 (Figure 1) are strictly conserved among bacterial RNase P RNAs (Haas et al., 1994). Except for G300 and A330, these nucleotides are also conserved in eukaryotic RNase P RNA subunits (Tranguch & Engelke, 1993). (2) Simultaneous mutation of nt 328 to 330, 334 and 335 or deletions in the region of nt 300 to 335 strongly impaired activity of Thermus thermophilus RNase P RNA (Schlegl et al., 1994). (3) Based on photocrosslinking experiments (Burgin & Pace, 1990; Oh & Pace, 1994), the 5' and 3' ends of the tRNA moiety were inferred to be close to nt 330 to 333 of RNase P RNA (Harris et al., 1994; Westhof & Altman, 1994). Likewise, a 4-thioU residue at position −1 of pre-tRNA pSu3-3XGC (Sva¨rd & Kirsebom, 1993) yielded crosslinks with G332 and A333 of E. coli RNase P RNA (Joanna Kufel & Leif A. Kirsebom, personal communication). (4) A chemical footprinting study revealed protection of E. coli RNase P RNA from modification at the same sites in the

423 presence of mature and precursor tRNA, except for a precursor-specific protection at G332 and a simultaneous enhancement of reactivity at A333 (LaGrandeur et al., 1994). (5) Rp-phosphorothioate modifications 5' of G300, A317 and A330 were recently shown to interfere with tRNA binding to E. coli RNase P RNA (Hardt et al., 1995a). (6) In earlier investigations, point mutations G to A329 (Shiraishi & Shimura, 1986), A to U334 and the double mutation C333/U334 (Baer et al., 1988) of E. coli RNase P RNA were reported to affect catalytic performance. (7) Based on the bulk of biochemical and crosslinking data, Westhof & Altman (1994) proposed in their three-dimensional model of the E. coli RNase P RNA-pre-tRNA complex that nt 328 to 331 may be part of the catalytic center, and suggested U331 as a candidate for the coordination of a catalytic magnesium ion. To study the contribution of base identities in single-stranded segments flanking helix P18 to enzyme function in more detail, we have constructed a series of E. coli RNase P RNAs with nucleotide exchanges in these regions. Point

Figure 1. Secondary structures of E. coli RNase P RNA according to Haas et al. (1994) and Thermus thermophilus pre-tRNAGly (Schlegl et al., 1992). Sites of specific Pb2+-induced hydrolysis in RNase P RNA are indicated by Roman numerals according to Ciesiolka et al. (1994). Grey-shaded circles indicate sites of hydrolysis whose intensities decrease in the presence of tRNA; site IVb (grey-shaded squares) is not seen in the absence of tRNA. The broken line marks a novel tertiary interaction identified recently (Mattsson et al., 1994). nt G292 and G293, which base-pair with the 3'-CCA end of tRNAs (Kirsebom & Sva¨rd, 1994), have been highlighted; nt 304/327 and 305/326 are shown unpaired since base-pairing at these positions is disfavored by phylogenetic evidence (James W. Brown & Norman R. Pace, personal communication) and our lead probing data (see the text). The arrow above nucleotide + l of the pre-tRNAGly indicates the canonical RNase P cleavage site.

424

RNase P RNA Mutations

Table 1. Pre-steady-state kinetic parameters for pre-tRNA processing and apparent equilibrium dissociation constants (appKd ) for tRNA binding by mutant RNase P RNAs Standard assay conditions (0.1 M NH4 OAc) Mutant wt C300 G301 C303 C304/G327 U328 C329 U330 A331 C332 U333 C334 A335 U321/U328 D309/ U330/G363

High salt conditions (e0.8 M NH4 OAc)

kreact (min−1 )

Km(pss) (mM)

kreacta/ Km(pss)

appKd (mM)

DGapp (kcal/mol)

kreact (min−1 )

Km(pss) (mM)

kreacta/ Km(pss )

NH4 OAc (M)

21 9 18 23 0.08 0.3 1.4 17 23 19 21 5 5 2 × 10−4

0.2 1.5 3.8 0.4 3.3 4.4 5.6 2.7 0.7 0.4 0.3 7.4 4.2 1.5

105 6 4.7 57.5 2.4 × 10−2 6.8 × 10−2 0.25 6.3 32.9 47.5 70 0.7 1.2 1.3 × 10−4

0.02 5 1.5 0.02 7.2 6.3 4.7 0.22 0.02 0.02 0.02 4.0 1.7 2

−10.9 −7.5 −8.2 −10.9 −7.3 −7.4 −7.5 −9.4 −10.9 −10.9 −10.9 −7.6 −8.2 −8.1

15 27 — — 7 9 14 34 — — — 24 11 0.06

0.14 0.2 — — 0.4 0.4 0.3 0.2 — — — 2.5 0.4 0.3

107 135 — — 17.5 22.5 46.7 170 — — — 9.6 27.5 0.2

1.5 1.5 — — 1.5 1.5 1.5 1.5 — — — 0.8 0.8 2.1

3 × 10−4

0.2

1.5 × 10−3

1.8

−8.1

0.03

0.07

0.43

2.1

appKd values were determined by gel retardation; T. thermophilus pre-tRNAGly (Figure 1) was used as the substrate for single turnover (E*S) cleavage assays performed at 37°C either in standard buffer A (0.1 M NH4 OAc, 0.1 M Mg(OAc)2 (pH 6.6); see Materials and Methods) or in the same buffer containing elevated NH4 OAc concentrations (high salt conditions); optimal NH4 OAc concentrations (indicated in the last column on the right side) were determined in the presence of 5 mM RNase P RNA, while keeping the Mg2+ concentration at 0.1 M. Kinetic and thermodynamic constants were derived from three independent experiments. Deviations found in individual experiments indicate that errors reach up to 230%; DGapp = −RT ln(1/appKd ), with R = 0.00198 kcal mol−1 K−1 and T in Kelvin. a kreact /Km(pss) is given in mM−1 min−1.

mutations were designed to introduce base identities rarely or not at all found at corresponding positions in known bacterial RNase P RNAs (RNase P RNA sequence compilation, kindly provided by James W. Brown and Norman R. Pace). Additional RNase P RNA variants U321/U328 and D309/ U330/G363 (Table 1) were isolated fortuitously owing to PCR errors.

Results A small number of mutant E. coli RNase P RNAs with nucleotide exchanges in single-stranded regions flanking helix P18 have been characterized in previous studies (Shiraishi & Shimura, 1986; Baer et al., 1988; Kirsebom et al., 1988). We were interested in a more systematic mutational analysis and in a comparison of the different mutant RNAs under uniform conditions. Moreover, we have applied a variety of assays to gain a deeper insight into the nature of functional defects. This included (1) gel retardation experiments to monitor specific tRNA binding, (2) single turnover kinetics, (3) structural probing by Pb2+-induced hydrolysis, and (4) complementation studies in E. coli RNase P mutant strains. tRNA binding to mutant RNase P RNAs Gel retardation was employed to measure tRNAGly binding to mutant RNase P RNAs apart from the overall catalytic reaction. Apparent equilibrium dissociation constants (appKd ) were determined by analyzing formation of the specific

tRNA–RNase P RNA complex in the presence of trace amounts of 32P-labeled pre-tRNA and increasing amounts of RNase P RNA (Hardt et al., 1995b). Note that excess amounts of wild-type and mutant RNase P RNA (except for low activity variants U321/U328 and D309/U330/G363) led to maturation of the pre-tRNA during the ten minute incubation step preceding gel loading. Thus, we measured binding of mature tRNAGly to RNase P in the case of most mutant RNase P RNAs, which was supported by the appearance of the 5' cleavage product on the non-denaturing polyacrylamide gels (data not shown). However, precursor and mature tRNAGly bind with very similar affinity to E. coli RNase P RNA (appKd E 20 nM, data not shown), as shown by gel retardation analyses performed in the presence of Ca2+ as the only divalent cation, which prevents significant maturation of pre-tRNA during such experiments. Thus, in the light of appKd values in the mM range measured for many mutant RNAs (Table 1), possible minor differences in binding affinity of mature and precursor tRNAGly are unlikely to affect the basic conclusions drawn in this study. Effects of RNase P RNA mutations on tRNA binding under standard assay conditions mainly fall into two categories (summarized in Table 1). One group of mutations had no effect on tRNA binding (C303, A331, C332, U333), while the other mutations, except for U330, affected tRNA binding dramatically, showing e75-fold reductions in binding affinity. Mutant U330 had an intermediate phenotype, resulting in a tenfold reduced binding affinity. These results show that ground state tRNA

425

RNase P RNA Mutations

binding is severely impaired in the case of mutants C300, G301, U328, C329, C334, A335 and the double and triple mutants.

however, retained largely decreased kreact values, while their Km(pss) values were well in the range of Km(pss) determined for wild-type E. coli RNase P RNA (Table 1, right part).

Kinetics of pre-tRNA cleavage by mutant RNase P RNAs

Structural probing with Pb2+

In single turnover reactions (pre-steady-state, ES), each enzyme molecule cleaves at most one substrate molecule. Release of mature tRNA has no effect on the processing reaction. This simplifies the interpretation of kinetic data, since only the chemical and preceding steps are considered. Single turnover kinetic constants were determined using trace amounts of 32P-labeled pre-tRNAGly (Figure 1) as the substrate; kreact (min−1 ) is the rate constant for the single turnover reaction at saturating [E] and Km(pss) (pss = pre-steady-state) corresponds to the enzyme concentration at the half-maximal rate of cleavage. Base exchanges at positions 303 and 331 to 333 had no or only minor effects on catalytic performance under standard assay conditions (Table 1, left part). Mutants G301 and U330 had a 10 to 20-fold increased Km(pss) , while kreact was unaffected. For mutant C300, a moderate reduction (about two fold) of the catalytic rate and a more severe increase in Km(pss) (about eight fold) was observed. Mutants C334, A335, and especially U328, C329 as well as the double and triple mutants C304/G327, U321/U328 and D309/U330/G363 showed the most pronounced changes in both kinetic parameters. It has been shown in previous studies that increased concentrations of monovalent salt can partially compensate for structural perturbations of RNase P RNA (Darr et al., 1992; Haas et al., 1994) or self-splicing group II introns (Chanfreau & Jacquier, 1994). At low monovalent salt concentrations, defects of mutant RNAs are thought to mainly result from electrostatic repulsion of phosphate groups due to perturbations of the phosphate-backbone (Smith et al., 1992). In order to determine whether the effects of mutations on kinetic constants were due to structural perturbations of this kind, we determined for each mutant RNA the optimal NH4 OAc concentrations at which highest pre-steady-state rate constants were obtained. When kreact and Km(pss) were measured under these high salt conditions (Table 1, right part), effects of mutations at positions nt 300 and nt 330 were completely suppressed. An almost complete rescue was also observed for mutants U328, C329, A335 and, quite remarkably, for the double mutant C304/G327. These results indicate that the observed catalytic defects were indeed mainly due to structural perturbations of the ribose-phosphate backbone rather than being a direct consequence of the loss of specific interactions involving the functional groups of bases at these locations. Only the single mutant C334 retained an about 20-fold increased Km(pss) at its optimal NH4 OAc concentration. Mutants U321/U328 and D309/U330/G363,

Lead ions are excellent probes to sense changes in the overall structure of RNase P RNA due to base exchanges or tRNA binding (Ciesiolka et al., 1994; Tallsjo¨ et al., 1993; Zito et al., 1993). In general, two categories of Pb2+-induced cleavage sites are observed in RNA molecules: first, highly specific sites due to tight metal ion binding sites, and, second, cleavages of lower intensity and specificity that are restricted to single-stranded RNA regions. Specific sites of the first category found in E. coli RNase P RNA are marked by Roman numerals in Figure 1. Mutant RNase P RNAs were labeled at their 3' or 5' ends, and were subjected to Pb2+-induced hydrolysis, as illustrated for several mutant RNAs in Figure 2. Assays employing 2 mM end-labeled RNase P RNAs were either performed in the absence or presence of excess amounts of pre-tRNAGly (10 mM). Point mutations at positions 303, 331, 332 and 333 had no effect on the Pb2+ hydrolysis pattern, either in the presence or absence of tRNA (Figure 2 and Table 2). In good correlation, these four mutant RNAs showed kreact , Km(pss) and appKd values very similar to the wild-type ribozyme (Table 1). In contrast, mutants with strong defects in catalysis and tRNA binding showed marked changes of their Pb2+ hydrolysis pattern in the region comprising nt 248 to 335 (including the specific sites IIc, III, IVa, V and VI; Figure 2, lanes 1 to 4; summarized in Table 2). In the presence of tRNAGly, the majority of mutant RNase P RNAs, like the wild-type RNA, showed the changes in the Pb2+ hydrolysis pattern attributed to specific tRNA binding (Ciesiolka et al., 1994; Figure 2, lanes 5 to 8; Table 2). However, tRNA-dependent reductions of hydrolysis at sites III and V and appearance of hydrolysis at site IVb were less pronounced for many mutant RNAs, which correlated with their increased Km(pss) and appKd values (Table 1). Hydrolysis at site IVb was completely absent in the most inactive mutant RNAs U321/U328 and D309/U330/G363. Only mutant RNAs C303, U330, A331, C332 and U333 showed tRNA-dependent intensity changes at sites III, IVb and V very similar to the wild-type RNA. No significant deviations from the wild-type Pb2+ hydrolysis pattern were detected outside the region comprising nt 248 to 335 in any of the mutant RNAs, with the exception of the helix P4 region (Figure 1). Here, we observed subtle tRNA-induced changes in the hydrolysis pattern around G72, which slightly differed between wild-type RNase P RNA and several mutant RNAs (Figure 3). The central core helix P4 includes most of the Rp-phosphorothioate modifications in RNase P RNA that were found to interfere with tRNA binding (Hardt et al., 1995a) or

Figure 2. Pb2+-induced hydrolysis patterns of (3'-32P)-labeled mutant E. coli RNase P RNAs at 50 mM Tris-HCl (pH 7.1 at 25°C), 0.1 M Mg(OAc)2 , 0.1 M NH4 OAc; 1.3 mg (10 pmol) of RNase P RNAs were incubated with either 2 mg (65 pmol) of pre-tRNAGly (lanes 5 to 8) or 2 mg of E. coli 50 S ribosomal RNA (lanes 1 to 4) at different Pb(OAc)2 concentrations for 15 minutes at 37°C. Lanes 1 and 5, control lanes in the absence of Pb2+; 2 and 6, 18 mM Pb(OAc)2 ; 3 and 7, 35 mM Pb(OAc)2 ; 4 and 8, 70 mM Pb(OAc)2 ; lanes T1 , limited digestion with RNase T1 ; L, limited alkaline hydrolysis. Main sites of Pb2+-induced hydrolysis are marked by Roman numerals (see also Figure 1) and brackets mark the region of nt 326 to 335.

427

RNase P RNA Mutations

Table 2. Changes in Pb2+-induced hydrolysis patterns of mutant RNase P RNAs Site/region − tRNA

C300 G301 C303 C304/G327 U328 C329 U330 A331 C332 U333 C334 A335 U321/U328 D309/U330/G363

+ tRNA

nt 326-335

nt 307-322

VI

V

wt wt wt oo oo o o wt wt wt o o oo oo

wt wt wt wt + wt wt wt wt wt wt wt ++ ++

wt + wt wt ++ ++ + wt wt wt wt wt ++ ++

wt wt wt wt + wt wt wt wt wt wt wt +b − −b

IIc/III/IVaa

III/IVb/Va

o o wt oo o oo o wt wt wt oo oo oo o

o o wt oo oo o wt wt wt wt o o ooo ooo

Based on Pb2+ hydrolysis experiments as shown in Figure 2; wt, wild-type pattern; moderate (o), strong (oo) or very strong (ooo) deviations from the wild-type pattern; moderate (+) or large (+ +) increases in susceptibility to Pb2+ hydrolysis; (− −) decreased susceptibility to Pb2+ hydrolysis. a Relative intensities of cleavage products derived from Pb2+ hydrolysis at indicated sites. b More diffuse hydrolysis pattern. For Pb2+-induced sites of hydrolysis indicated by Roman numerals, see Figure 1.

pre-tRNA cleavage (Harris & Pace, 1995). In addition, manganese rescue experiments suggested direct metal ion coordination to the pro-Rp oxygen atoms at A67 (Harris & Pace, 1995), U69 and C70 (Hardt et al., 1995a).

Complementation studies Plasmids carrying the wild-type or mutant RNase P RNA genes under control of the native E. coli rnpB promoter were introduced into temperature-

Figure 3. Pb2+-induced hydrolysis patterns of (5'-32P)-labeled wild-type and several mutant RNase P RNAs in the absence (lanes 1 to 4) or presence (lanes 5 to 8) of pre-tRNAGly. For experimental details, see the legend to Figure 2; tRNA-induced changes in the lead hydrolysis pattern in the region of G72 are marked by bars next to lanes 4 and 8.

428

RNase P RNA Mutations

Table 3. Complementation of E. coli RNase P mutant strains by rnpB mutant alleles rnpA49 Mutant RNase P RNA Wild-type C300 G301 C303 C304/G327 U328 C329 U330 A331 C332 U333 C334 A335 U321/U328 D309/U330/G363 pSP64

DW2

30°C

42°C

30°C

42°C

+ + + + + + + + + + + + + + + +

+ − − (+) − −− −− − − − −− −− −− −− −− −−

+ + + + + + + + + + + + + + + +

+ − − (+) − − −− − − − (+) −− −− −− − −−

Growth of mutant strains transformed with the different rnpB alleles was analyzed on LB plates at permissive (30°C) and non-permissive (42°C) temperatures. With both strains, none of the mutant alleles showed the same complementation efficiency as the wild-type allele at 42°C; + > (+) > − > − −, decreasing order of complementation efficiency based on numbers and sizes of colonies.

sensitive E. coli RNase P mutant strains (rnpA49 and DW2/pDW160). The rnpA49 strain carries an A to G transition in the rnpA gene encoding the protein subunit (C5) of E. coli RNase P, which results in an arginine to histidine exchange at position 46 of C5 (Kirsebom et al., 1988). The mutation leads to temperature-sensitive bacterial growth (Schedl et al., 1974). In vitro studies have indicated that the mutant protein is defective in forming active RNase P (Baer et al., 1989). It has been shown, however, that overproduction of functional RNase P RNA can suppress the mutant phenotype in vivo (Kirsebom et al., 1988, and references cited therein), which is paralleled by stimulation of RNase P activity in vitro in the presence of excess RNase P RNA over C5A49 (Baer et al., 1989). In strain DW2/pDW160, the chromosomal gene encoding the RNase P RNA subunit (rnpB) has been replaced with a chloramphenicol acetyltransferase gene, and a complementing rnpB gene is provided on a plasmid temperature-sensitive for replication (Waugh & Pace, 1990). Suppression of the conditionally lethal phenotype at 42°C is achieved by expression of functional RNase P RNA from a second compatible plasmid (Waugh & Pace, 1990). Thus, cell growth relies entirely on the plasmid-encoded rnpB gene under investigation. Only the wild-type gene, but none of the mutant alleles, was able to fully restore growth at the non-permissive temperature in rnpA49 and DW2/ pDW160 bacteria (Table 3). Among the mutant rnpB genes, C303 was most proficient in partially suppressing the mutant phenotypes, which correlated with its almost wild-type catalytic performance in vitro (Table 1). Variant U333 was ambiguous, displaying no complementation effect in rnpA49 bacteria, but partial restoration of growth at the

non-permissive temperature in DW2/pDW160 bacteria. Very small but detectable complementation effects were observed for mutant alleles C300, G301, C304/G327, U330, A331 and C332 in rnpA49 as well as in DW2/pDW160 bacteria at 42°C; rnpB alleles carrying the U328, C329, C334, A335, U321/U328 or D309/U330/G363 mutations were most inefficient in supporting growth of the mutant strains at 42°C.

Discussion As outlined in the Introduction, several lines of evidence have indicated that the single-stranded regions flanking helix P18 of RNase P RNA play an important role in ribozyme function. Using several different assay systems, we have analyzed a series of RNase P RNAs with point mutations in these regions in order to determine the functional contributions of individual nucleotides. Specific effects of single base exchanges Single base exchanges at positions 303, 331, 332 and 333 had no significant effect on tRNA binding and only marginal effects on pre-steady-state kinetic parameters (Table 1, left part). In addition, Pb2+ hydrolysis patterns of these mutants were identical to that of the wild-type RNase P RNA (Figure 2 and Table 2). We conclude that, under in vitro assay conditions, functional groups of the bases at these positions neither significantly contribute to structural stabilization of RNase P RNA nor to (pre-)tRNA binding or formation of the active site. This is in line with an earlier report showing that mutant C333 had wild-type catalytic properties in vitro (Baer et al., 1988). Point mutations at positions 300, 301 and 330 had strong effects on Km(pss) , while showing no (or in the case of mutant C300 only a small) effect on the catalytic rate (Table 1). Likewise, mutants C300, G301 and U330 showed decreases in the binding affinity for mature tRNA in gel retardation assays (>tenfold increases in appKd ; Table 1). Our results indicate that the loss of binding energy due to mutations at nt 300, 301 or 330 affects substrate binding in the ground state and transition state to a very similar extent (Fersht, 1985). Thus, bases at these positions contribute to formation or stabilization of the specific (pre-)tRNA–RNase P RNA complex, but do not affect active site geometry. Effects of mutations C300 and U330 on Km(pss) were suppressed under high salt conditions (Table 1, right part). It is therefore likely that distortion of the ribozyme structure due to base exchanges at these positions impaired (pre-)tRNA binding by virtue of increased electrostatic repulsion. Among the mutant RNase P RNAs with single base exchanges analyzed in this study, U328 and C329 showed the strongest reductions (e15-fold) in kreact . Higher, but still significantly reduced catalytic rates were observed for mutants C334 and A335. This is in line with previous reports, which revealed reduced catalytic rates for E. coli mutants A329 and

429

RNase P RNA Mutations

U334 (Baer et al., 1988; Shiraishi & Shimura, 1986). Increased Km(pss) and appKd values (Table 1) and reduced catalytic rates of mutant RNAs U328, C329, C334 and A335 suggest that the loss of binding energy preferentially destabilizes the enzyme–substrate transition state complex. Thus, nt 328, 329, 334 and 335 play a role in both ground state binding of (pre-)tRNA and formation of the active site. The 1500-fold reduction in catalytic efficiency (kreact / Km(pss) ) observed for mutant U328 reveals that, in wild-type RNase P RNA, nt A328 plays a key role in the tertiary structural organization of the ribozyme core. The catalytic defects of mutants U328, C329 and A335 could be largely compensated by increased salt concentrations, suggesting that electrostatic repulsion of phosphate groups was the main cause for malfunction. Significant structural distortion is corroborated by the strongly deviating Pb2+ hydrolysis patterns observed for these mutants (Figure 2 and Table 2). C334 was the only mutant RNase P RNA among those with single base exchanges that retained an about 20-fold increased Km(pss) at the optimal salt concentration (0.8 M NH4+ ; Table 1, right part). Thus, in wild-type RNase P RNA, A334 seems to support an intra- or intermolecular base-specific interaction crucial to tRNA binding. Specific interaction of RNase P RNA and (pre-)tRNA is largely mediated through hydrogen bonding (Smith et al., 1992). Calculation of binding energies from appKd values (DGapp , Table 1) indicates that mutations C300, G301, U328, C329, C334 and A335 lead to the disruption of at least one hydrogen bond in the RNase P RNA–tRNA complex. However, at present it is impossible to assign observed effects to the disruption of specific intra- or intermolecular hydrogen bonds due to the low resolution of current structural models of the pre-tRNA–RNase P RNA complex (Westhof & Altman, 1994; Harris et al., 1994). In addition, base substitutions caused structural changes in the more extended region comprising nt 248 to 335 of RNase P RNA (Figure 2 and Table 2), which may suggest that observed losses of binding energy reflect the sum of several minor distortions. The role of the nt 248 to 335 region for interaction with the tRNA Based on intra- and intermolecular crosslinking data, nt 247 to 249 and the nt 328 to 335 region have been positioned close to each other in the tertiary structure, and A248 and G332 have been placed in the vicinity of the pre-tRNA cleavage site (Harris et al., 1994). Likewise, Westhof & Altman (1994) positioned nt + 1 of the tRNA domain close to A249 and A330. Furthermore, the tRNA CCA 3' end interacts with the internal bulge region of nt 254 to 259/291 to 295 (Oh & Pace, 1994; LaGrandeur et al., 1994; Kirsebom & Sva¨rd, 1994) and general metal ion cleavage sites (Kazakov & Altman, 1991) are clustered in the region of nt 248 to 300 (Figure 1), suggesting the binding of structurally important

metal ions within this part of RNase P RNA. Moreover, intensities of Pb2+ induced hydrolysis at sites III, IVb and V change upon tRNA binding (Figure 2; Ciesiolka et al., 1994), and Rp-phosporothioate modifications in E. coli RNase P RNA 5' of A248, A249, G300, A317 and A330 were shown to interfere with tRNA binding (Hardt et al., 1995a). In this study, we have identified seven base moieties (nt 300, 301, 328-330, 334 and 335) among single-stranded nucleotides flanking helix P18, which strongly affect enzyme function. Interestingly, each of these single base exchanges resulted in a decreased tRNA affinity as well as an altered pattern of susceptibility to Pb2+ induced hydrolysis in the region comprising nt 248 to 335. This string of experimental evidence suggests that nt 248 to 335 may form a subdomain of RNase P RNA and supports the role of this region for docking the substrate, and the terminal part of the acceptor stem in particular, into the active site. Structure and function of P18 Evidence for a direct or indirect role of P18 in (pre-)tRNA binding has been reported. Deletion variants of RNase P RNAs from E. coli (Haas et al., 1994) and T. thermophilus (Schlegl et al., 1994) lacking P18 showed largely increased Km values and an Rp-phosphorothioate modification at A317 interfered with tRNA binding (Hardt et al., 1995a). In this study we have included two RNase P RNAs (U321/U328 and D309/U330/G363) with mutations in helix P18. Both mutant RNAs showed very strong defects in catalytic function and tRNA binding, which were only partially suppressed under high salt conditions (Table 1). Nucleotides in helix P18 became largely susceptible to Pb2+ hydrolysis in these two RNAs, suggesting that functional defects may at least partially be attributed to the disruption of helix P18. The apical tetraloop of P18 was insusceptible to Pb2+ hydrolysis in wild-type RNase P RNA. Surprisingly, the U328 mutation, although located outside of P18, seems to unfold the apical part of P18, as inferred from increased susceptibility to Pb2+hydrolysis (Figure 2 and Table 2). Also, nt 300 to 305 became more accessible in mutant RNAs U328, C329, U321/U328 and D309/U330/G363. This suggests that P18 and its flanking sequences (nt 300 to 305 and nt 326 to 335) are part of a higher-order structural context. It may also explain why some mutant RNAs with structural perturbations involving P18, such as variant U321/U328, have more defective phenotypes than previously characterized deletion variants in which P18 has been replaced by a single C residue (Haas et al., 1994; Schlegl et al., 1994). Two non-canonical base-pairs have been proposed to form at the base of helix P18 (Figure 1; Haas et al., 1994). However, nt 326 and 327 were susceptible to Pb2+-induced hydrolysis in the wild-type RNA and all variants (except for mutant U328/U321; Figure 2). This may be attributable

430 either to the absence of base-pairing at these positions or to partial opening of the two non-canonical terminal base-pairs. To analyze the role of these putative base-pairs in more detail, we have designed the double mutation C304/G327 which would have the potential to replace the putative G304-U327 pair, thereby stabilizing the helix end. However, susceptibility of nt 326 and 327 to Pb2+ hydrolysis was even enhanced in mutant RNA C304/G327 (Figure 2). In addition, the double mutation caused a 250-fold reduction of the catalytic rate as well as large increases in Km(pss) and appKd under standard assay conditions (Table 1). These observations argue against base-pairing of nt 304/327 and 305/326 (which is also indicated by phylogenetic evidence: James W. Brown & Norman R. Pace, personal communication) and suggest that structural constraints other than helicity are crucial at these positions.

Complementation analysis None of the mutant RNase P RNAs analyzed in this study was able to fully complement RNase P function in vivo (Table 3). This is in line with an earlier study which showed that neither the C333 nor the U334 mutant allele was able to complement the mutant phenotype of rnpA49 bacteria (Baer et al., 1988). The failure of RNAs with mutations at positions 303 and 331 to 333, which behaved very similar to the wild-type RNase P RNA in vitro, to fully support bacterial survival may be related to sub-optimal folding of mutant RNAs under conditions of the cellular ionic milieu or to impaired binding of the protein subunit in vivo. In the presence of the protein subunit C5, E. coli RNase P RNA is protected from enzymatic cleavage and chemical modification at nt 266 to 287 (Talbot & Altman, 1994; Vioque et al., 1988), suggesting that RNA–protein contacts occur in this part of the RNA. Since many of the mutations analyzed here affected the structure of this region, which harbors the Pb2+ hydrolysis sites IVa and IVb (Figure 2 and Table 2), impaired C5 binding may well have contributed to the malfunction of these mutant RNAs in vivo. Also, some mutations might lead to rapid RNA degradation or inefficient processing of RNase P RNA primary transcripts. However, this seems less likely since it has been demonstrated for a variety of rnpB alleles carrying mutations (Lumelsky & Altman, 1988) or short deletions (Lawrence & Altman, 1986) that the mutant RNase P RNAs are stably produced in vivo. Likewise, mutant RNA with a G to A substitution at nt 329 was shown to be stable in vitro and in vivo (Shiraishi & Shimura, 1986). The only reported example of a mutation that resulted in enhanced degradation of E. coli RNase P RNA has been the replacement of G89 with A (Shiraishi & Shimura, 1986). Although this mutation locates to a region different from the one analyzed here, it suggests that instability of some mutant RNase P RNAs cannot be completely ruled out.

RNase P RNA Mutations

Concluding remarks We have studied a sub-region of the core structure of RNase P RNA by site-directed mutagenesis. Although severe functional defects were observed in vitro and none of the mutant RNAs was able to fully replace the wild-type RNA in vivo, evidence is lacking that any of the nucleotides exchanged here may be directly involved in transition state stabilization by means of functional groups of bases. However, strictly conserved nt 328 in particular but also nt 329, 334 and 335 were revealed to play an indirect though important role in facilitating the chemical step by stabilizing the catalytic core structure. In addition, our study yields strong evidence that the adenine at position 334 provides a major direct contribution to the energy of substrate binding. Severe functional defects due to base exchanges in the region of nt 300 to 335 correlated with structural changes in the more extended region of nt 248 to 335. This points to a high degree of structural organization within this part of the ribozyme, which may be considered a subdomain of RNase P RNA.

Materials and Methods Bacteria E. coli strains NHY322 (rnpA49) (Kirsebom et al., 1988) and DW2/pDW160 (Waugh & Pace, 1990) were used for complementation studies. Mutagenesis Mutations were introduced into the E. coli RNase P RNA gene by a three-step PCR protocol. (1) In the first step, a shortened RNase P RNA gene fragment carrying a truncation at the 3' end was obtained by using the following primers: primer 1 carried a 5' overhang which introduced a terminal EcoRI site and the T7 promoter at the 5' end of the coding sequence; primer 2 covered the site of mutation and contained a single mismatch specific for the individual mutant gene. (2) In the second step, the 5' strand of the PCR product from the first step was annealed to the partially overlapping primer 3, which encoded the remaining 3' portion of the gene and a terminal BamHI site; this partially overlapping hybrid was converted to a double-stranded product by extending the 3' ends of both strands (‘‘fill in’’ reaction). (3) In the third step, this double-stranded product, representing the full-length mutant gene, was amplified by use of primers l and 3. PCR products were then digested with EcoRI and BamHI and cloned into the plasmid pSP64 (Promega) by standard techniques (Maniatis et al., 1982). Altered RNase P RNA genes were verified by DNA dideoxy sequencing using double-stranded plasmid DNA and T7 DNA polymerase according to the protocol provided by the manufacturer (Pharmacia). Preparation of RNAs RNase P RNAs and pre-tRNAGly (carrying the sequence CCAAUA at the 3' terminus) were synthesized by runoff transcription with bacteriophage T7 RNA polymerase; 1 ml assays (in DEPC-treated water) contained 80 mM

431

RNase P RNA Mutations

Hepes-HCl (pH 7.5), 22 mM MgCl2 , 1 mM spermidine, 3.75 mM each ATP, CTP, GTP and UTP, 120 mg of bovine serum albumin, 5 units of pyrophosphatase (Sigma), about 15 pmol (linearized plasmid DNA) or 300 pmol (PCR-amplified DNA) of template DNA and 2000 units of T7 RNA polymerase (MBI Fermentas, Lithuania). After six hours at 37°C, 10 ml of DNase I (Boehringer Mannheim, RNase-free) was added, followed by another incubation for 30 minutes at 37°C and phenol/chloroform (1:1, v/v) extraction. RNAs were purified on 10% (for pre-tRNA) or 5% (for RNase P RNAs) polyacrylamide/ 8 M urea gels as described (Ciesiolka et al., 1994). RNA concentrations were determined by absorption at 260 nm (1 A260 unit = 37 mg). Pre-tRNAGly was internally labeled with [a-32P]CMP as described (Hardt et al., 1993). 3' End-labeling of RNase P RNAs was performed using [5'-32P]pCp and T4 RNA ligase (England & Uhlenbeck, 1978). For 5' end-labeling, RNase P RNAs were synthesized in the presence of ApG as initiator of transcription (Hardt et al., 1995a). Processing assays Single turnover (pre-steady-state) experiments were performed with 1 to 10 nM 32P-labeled pre-tRNAGly ˇ erenkov cpm per data point) and excess (about 20,000 C amounts of E. coli RNase P RNAs (0.1 to 10 × Km(pss) ; E*S) as described (Hardt et al., 1995b). Assays were performed in buffer A (0.05 M TrisOAc, 0.1 M Mg(OAc)2 , 0.1 M NH4 OAc and 2 mM EDTA (pH 6.6 at 37°C)) at 37°C, unless stated otherwise. For mutant RNase P RNAs we could show that variations of the amount of 32P-labeled substrate did not affect the observed reaction rates (kobs ) at a fixed excess enzyme concentration. This largely excluded that mutant RNase P RNAs represented a heterogeneous pool of different conformers including a low proportion of wild-type-like folded molecules. Gel retardation analyses Assays were performed as described (Hardt et al., 1995b). RNase P RNAs were pre-incubated separately for 60 minutes at 37°C in buffer A (see above) including 5% glycerol in a total volume of 9 ml; 1 ml of 32P-labeled ˇ erenkov cpm) in buffer A was pre-tRNAGly (about 20,000 C added, followed by another ten minutes of incubation at 37°C. Electrophoresis, quantification of complex formation and determination of apparent equilibrium dissociation constants (appKd ) for the tRNA–RNase P RNA complex were performed as described (Hardt et al., 1995b). The equation Kd = [RNase P RNA] − 0.5[tRNA] was used for appKd determinations (Pyle et al., 1990). In our recent study (Hardt et al., 1995b) we used a modified equation Kd = 0.7[RNase P RNA] − 0.5[tRNA] for appKd determinations. The factor of 0.7 considered that only about 70% of the wild-type E. coli RNase P RNA molecules were found to be capable of gel-resolvable tRNA binding after a one to two hour preincubation step at 37°C. We have not analyzed here the proportion of mutant RNase P RNAs that were able to bind tRNA under these conditions. Therefore, a correcting factor was omitted in the above equation, which, however, had no influence on the basic conclusions drawn in this study. It should be noted that in those cases where we varied the amount of 32P-labeled tRNA in the presence of excess amounts of mutant RNase P RNAs in gel retardation experiments, no significant changes in appKd were observed (data not shown). It is thus unlikely (as also inferred from the kinetic experiments, see above) that

increases in appKd observed for mutant RNase P RNAs are due to the presence of very low proportions of RNA conformers binding tRNA with wild-type affinity. Hydrolysis by Pb2+ A portion (1.3 mg; 10 pmol) of the (3'-32P)- or (5'-32P)labeled RNase P RNAs was pre-incubated with either 2 mg (65 pmol) of pre-tRNAGly or 2 mg of E. coli 50 S ribosomal RNA for one hour at 37°C in 1.25 × buffer B (buffer B: 0.05 M Tris-HCl (pH 7.1 at 25°C), 0.1 M Mg(OAc)2 , 0.1 M NH4 OAc). In the case of RNase P RNAs with single nucleotide exchanges, the pre-tRNA substrate was entirely or almost completely converted to mature tRNA during the one hour pre-incubation step. Pb2+ hydrolysis reactions were started by adding 1 ml of lead acetate solution (0.09 to 0.35 M) to 4 ml of RNA solution. In control lanes 1 ml of water was added. After a 15 minute incubation at 37°C, hydrolysis reactions were stopped by addition of 7 ml of loading buffer (67% (v/v) formamide, 0.3 × buffer B, 2.7 M urea, 100 mM EDTA) and shockfreezing in liquid nitrogen. Complementation studies Mutant RNase P RNA genes were put under control of the native E. coli RNase P RNA gene promoter. This was accomplished as follows: (1) an NheI/SmaI fragment (0.8 kb), harboring the 5' portion of the wild-type rnpB gene (and its natural promoter and upstream sequences; the SmaI site corresponds to nt 288 to 293, Figure 1), was excised from plasmid pDW160 (Waugh & Pace, 1990) and cloned into pSP64 (Promega) digested with NheI and SmaI; (2) a 0.8 kb fragment, which was excised from the recombinant pSP64 construct utilizing NaeI (yielding blunt ends, adjacent to the NheI site in pSP64) and SmaI, was cloned into the SmaI site of pUC18; one clone with the desired orientation (direction of transcription of rnpB from the EcoRI to the SmaI site within the pUC18 construct) was selected; (3) a 0.8 kb EcoRI/SmaI fragment, harboring the 5' portion of rnpB and its natural upstream sequences, was excised from the recombinant pUC18 plasmid and substituted for the 0.3 kb EcoRI/SmaI fragments of pSP64 mutant RNase P RNA constructs (see above, mutagenesis), thereby eliminating the T7 promoter and introducing the natural rnpB promoter. Recombinant plasmids were introduced into E. coli strains NHY322 (rnpA49) and DW2/pDW160. Both strains were grown at 30°C in LB medium; 30 mg/ml chloramphenicol was added to cultures of DW2/pDW160. Competent cells were prepared as described (Mandel & Higa, 1970). For transformation, 2 ml (about 2 mg of plasmid DNA) of pSP64 constructs, harboring the mutant rnpB alleles under control of the natural rnpB promoter, were mixed with 80 ml of competent cells, followed by incubation for 15 minutes on ice, three minutes at 30°C, 15 minutes on ice, three minutes at 30°C and 15 minutes on ice; tenfold serial dilutions (100 to 10−3 ) of cell suspensions (30 ml) were plated directly in duplicate on LB agar plates containing 100 mg/ml ampicillin (NHY322) or 100 mg/ml ampicillin and 30 mg/ml chloramphenicol (DW2/ pDW160). The two sets of plates were incubated in parallel at 30°C or 42°C, respectively, for 18 to 42 hours.

Acknowledgements We thank Volker A. Erdmann for continuous support, Werner Schro¨der for the synthesis of DNA oligonucleo-

432 tides, Leif A. Kirsebom and Norman R. Pace for kindly providing E. coli strains NHY322 and DW2/pDW160, and James W. Brown, Norman R. Pace, Joanna Kufel and Leif A. Kirsebom for providing unpublished results. Financial support for these studies from the Deutsche Forschungsgemeinschaft (Ha 1672/4-1 and SFB 344/C2) and the Fonds der Chemischen Industrie is acknowledged. W.-D.H. received a stipend from the Boehringer Ingelheim Fonds.

References Baer, M. F., Reilly, R. M., McCorkle, G. M., Hai, T.-Y., Altman, S. & RajBhandary, U. L. (1988). The recognition by RNase P of precursor tRNAs. J. Biol. Chem. 263, 2344–2351. Baer, M. F., Wesolowski, D. & Altman, S. (1989). Characterization in vitro of the defect in a temperature-sensitive mutant of the protein subunit of RNase P from Escherichia coli. J. Bacteriol. 171, 6862–6866. Brown, J. W. & Pace, N. R. (1992). Ribonuclease P RNA and protein subunits from bacteria. Nucl. Acids Res. 20, 1451–1456. Burgin, A. B. & Pace, N. R. (1990). Mapping the active site of ribonuclease P RNA using a substrate containing a photoaffinity agent. EMBO J. 9, 4111–4118. Chanfreau, G. & Jacquier, A. (1994). Catalytic site components common to both splicing steps of a group II intron. Science, 266, 1383–1387. Ciesiolka, J., Hardt, W.-D., Schlegl, J., Erdmann, V. A. & Hartmann, R. K. (1994). Lead-ion induced cleavage of RNase P RNA. Eur. J. Biochem. 219, 49–56. Darr, S. C., Zito, K., Smith, D. & Pace, N. R. (1992). Contributions of phylogenetically variable structural elements to the function of the ribozyme ribonuclease P. Biochemistry, 31, 328–333. England, T. E. & Uhlenbeck, O. C. (1978). 3'-Terminal labelling of RNA with T4 RNA ligase. Nature, 275, 560–561. Fersht, A. (1985). Enzyme Structure and Mechanism, W. H. Freeman and Company, New York. Guerrier-Takada, C., Gardiner, K., Marsh, T., Pace, N. & Altman, S. (1983). The RNA moiety of ribonuclease P is the catalytic subunit of the enzyme. Cell, 35, 849–857. Haas, E. S., Brown, J. W., Pitulle, C. & Pace, N. R. (1994). Further perspective on the catalytic core and secondary structure of ribonuclease P RNA. Proc. Natl Acad. Sci. USA, 91, 2527–2531. Hardt, W.-D., Schlegl, J., Erdmann, V. A. & Hartmann, R. K. (1993). Gel retardation analysis of E. coli M1 RNA-tRNA complexes. Nucl. Acids Res. 21, 3521– 3527. Hardt, W.-D., Warnecke, J. M., Erdmann, V. A. & Hartmann, R. K. (1995a). Rp-phosphorothioate modifications in RNase P RNA that interfere with tRNA binding. EMBO J. 14, 2935–2944. Hardt, W.-D., Schlegl, J., Erdmann, V. A. & Hartmann, R. K. (1995b). Kinetics and thermodynamics of the RNase P RNA cleavage reaction: analysis of tRNA 3'-end variants. J. Mol. Biol. 247, 161-172. Harris, M. E. & Pace, N. R. (1995). Identification of phosphates involved in catalysis by the ribozyme RNase P RNA. RNA, 1, 210–218. Harris, M. E., Nolan, J. M., Malhotra, A., Brown, J. W., Harvey, S. C. & Pace, N. R. (1994). Use of photoaffinity crosslinking and molecular modeling to analyze the global architecture of ribonuclease P RNA. EMBO J. 13, 3953–3963.

RNase P RNA Mutations

Kazakov, S. & Altman, S. (1991). Site-specific cleavage by metal ion cofactors and inhibitors of M1 RNA, the catalytic subunit of RNase P from Escherichia coli. Proc. Natl Acad. Sci. USA, 88, 9193–9197. Kirsebom, L. A. & Sva¨rd, S. G. (1994). Base-pairing between Escherichia coli RNase P RNA and its substrate. EMBO J. 13, 4870–4876. Kirsebom, L. A., Baer, M. F. & Altman, S. (1988). Differential effects of mutations in the protein and RNA moieties of RNase P on the efficiency of suppression by various tRNA suppressors. J. Mol. Biol. 204, 879–888. LaGrandeur, T. E., Hu¨ttenhofer, A., Noller, H. F. & Pace, N. R. (1994). Phylogenetic comparative chemical footprint analysis of the interaction between ribonuclease P RNA and tRNA. EMBO J. 13, 3945–3952. Lawrence, N. P. & Altman, S. (1986). Site-directed mutagenesis of M1 RNA, the RNA subunit of Escherichia coli ribonuclease P. J. Mol. Biol. 191, 163–175. Lumelsky, N. & Altman, S. (1988). Selection and characterization of randomly produced mutants in the gene coding for M1 RNA. J. Mol. Biol. 202, 443–454. Mandel, M. & Higa, A. (1970). Calcium dependent bacteriophage DNA infection. J. Mol. Biol. 53, 154–162. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982). Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Mattsson, J. G., Sva¨rd, S. G. & Kirsebom, L. A. (1994). Characterization of the Borrelia burgdorferi RNase P RNA gene reveals a novel tertiary interaction. J. Mol. Biol. 241, 1–6. Oh, B.-K. & Pace, N. R. (1994). Interaction of the 3'-end of tRNA with ribonuclease P RNA. Nucl. Acids Res. 22, 4087–4094. Pyle, A. M., McSwiggen, J. A. & Cech, T. R. (1990). Direct measurement of oligonucleotide substrate binding to wild-type and mutant ribozymes from Tetrahymena. Proc. Natl Acad. Sci. USA, 87, 8187–8191. Schedl, P., Primakoff, P. & Roberts, J. (1974). Processing of E. coli tRNA precursors. Brookhaven Symp. Biol. 26, 53–76. Schlegl, J., Fu¨rste, J. P., Bald, R., Erdmann, V. A. & Hartmann, R. K. (1992). Cleavage efficiencies of model substrates for ribonuclease P from Escherichia coli and Thermus thermophilus. Nucl. Acids Res. 20, 5963–5970. Schlegl, J., Hardt, W.-D., Erdmann, V. A. & Hartmann, R. K. (1994). Contribution of structural elements to Thermus thermophilus ribonuclease P RNA function. EMBO J. 13, 4863–4869. Shiraishi, H. & Shimura, Y. (1986). Mutations affecting two distinct functions of the RNA component of RNase P. EMBO J. 5, 3673–3679. Smith, D., Burgin, A., Haas, E. S. & Pace, N. R. (1992). Influence of metal ions on the ribonuclease P reaction. J. Biol. Chem. 267, 2429–2436. Sva¨rd, S. G. & Kirsebom, L. A. (1993). Determinants of Escherichia coli RNase P cleavage site selection: a detailed in vitro and in vivo analysis. Nucl. Acids Res. 21, 427–434. Talbot, S. J. & Altman, S. (1994). Gel retardation analysis of the interaction between C5 protein and M1 RNA in the formation of the ribonuclease P holoenzyme from Escherichia coli. Biochemistry, 33, 1399–1405. Tallsjo¨, A., Sva¨rd, S. G., Kufel, J. & Kirsebom, L. A. (1993). A novel tertiary interaction in M1 RNA, the catalytic

RNase P RNA Mutations

subunit of Escherichia coli RNase P. Nucl. Acids Res. 21, 3927–3933. Tranguch, A. J. & Engelke, D. R. (1993). Comparative structural analysis of nuclear RNase P RNAs from yeast. J. Biol. Chem. 268, 14045–14053. Vioque, A., Arnez, J. & Altman, S. (1988). Protein-RNA interactions in the RNase P holoenzyme from Escherichia coli. J. Mol. Biol. 202, 835–848. Waugh, D. S. & Pace, N. R. (1990). Complementation of an RNase P RNA (rnpB) gene deletion in Escherichia coli

433 by homologous genes from distantly related eubacteria. J. Bacteriol. 172, 6316–6321. Westhof, E. & Altman, S. (1994). Three-dimensional working model of M1 RNA, the catalytic RNA subunit of ribonuclease P from Escherichia coli. Proc. Natl Acad. Sci. USA, 91, 5133–5137. Zito, K., Hu¨ttenhofer, A. & Pace, N. R. (1993). Leadcatalyzed cleavage of ribonuclease P RNA as a probe for integrity of tertiary structure. Nucl. Acids Res. 21, 5916–5920.

Edited by J. Karn (Received 10 January 1996; received in revised form 15 March 1996; accepted 19 March 1996)