How a CCA Sequence Protects Mature tRNAs and tRNA Precursors from Action of the Processing Enzyme RNase BN/RNase Z

Background: 3′-Terminal CCA-containing tRNAs and precursors are resistant to action of RNase BN/RNase Z.

Results: Arg²⁷⁴ and the two C residues are required for protection by the CCA sequence.

Conclusion: Presence of Arg²⁷⁴ and CC sequence prevents RNA substrate from moving into the RNase catalytic site.

Significance: This mechanism explains how mature tRNAs are protected from removal of the CCA sequence by a processing RNase.

Abstract

In many organisms, 3′ maturation of tRNAs is catalyzed by the endoribonuclease, RNase BN/RNase Z, which cleaves after the discriminator nucleotide to generate a substrate for addition of the universal CCA sequence. However, tRNAs or tRNA precursors that already contain a CCA sequence are not cleaved, thereby avoiding a futile cycle of removal and readdition of these essential residues. We show here that the adjacent C residues of the CCA sequence and an Arg residue within a highly conserved sequence motif in the channel leading to the RNase catalytic site are both required for the protective effect of the CCA sequence. When both of these determinants are present, CCA-containing RNAs in the channel are unable to move into the catalytic site; however, substitution of either of the C residues by A or U or mutation of Arg²⁷⁴ to Ala allows RNA movement and catalysis to proceed. These data define a novel mechanism for how tRNAs are protected against the promiscuous action of a processing enzyme.

Introduction

Cells contain a large repertoire of RNases that are involved in either maturation or targeted degradation of cellular RNAs, a process collectively known as RNA metabolism (1–4). Inasmuch as many of these RNases are also destructive enzymes with the ability to degrade or damage important functional RNAs (5, 6), various regulatory mechanisms must have evolved to protect cellular RNAs from unwanted degradation by these enzymes, although very little information is currently available in this area. Some RNAs such as tRNA and rRNA are stable in cells because they have extensive secondary structure or contain protective nucleotide sequences or because they are sequestered in RNP particles. In addition, certain RNases may also have built-in specificities that limit their action on functional RNAs (7). In Escherichia coli, which contains more than 20 RNases with different characteristics, it is likely that multiple mechanisms operate to protect cellular RNAs.

One clear example of such a protective mechanism involves the enzyme, RNase BN (8–11). RNase BN is the E. coli member of the RNase Z family of enzymes that participate in 3′ maturation of tRNA precursors in many organisms (12). Functional tRNAs require the universal CCA sequence at their 3′-terminus for amino acid attachment and for aminoacyl-tRNA action on the ribosome. In some bacteria such as E. coli, this sequence is already present in all the tRNA precursors, and 3′ maturation involves removal of extra residues following the CCA sequence, a process catalyzed by any one of a number of exoribonucleases (13–15). In organisms lacking an encoded -CCA sequence in all or some of their tRNA precursors, RNase Z makes an endonucleolytic cleavage following the discriminator nucleotide, and CCA is added by tRNA nucleotidyltransferase (16, 17). However, in no case does RNase Z remove the CCA sequence, resulting in protection of mature tRNAs (10, 11, 18).

In contrast to other members of the RNase Z family, which act solely as endoribonucleases (16, 17), RNase BN also displays a 3′-5′ exoribonuclease activity on model RNAs and tRNA precursor substrates in vitro (9, 10), and it acts as a dual function nuclease on tRNA precursors in vivo (11). RNase BN action on E. coli tRNA precursors that already contain a CCA sequence involves either a single endonucleolytic cleavage after the CCA sequence or exonucleolytic trimming to remove extra 3′ residues up to the CCA sequence, depending on the metal ion cofactor present (10). Like other RNase Z family members, it also does not remove the CCA sequence from mature tRNAs (10, 11, 18). Thus, RNase BN and other members of the RNase Z family have built-in specificity to avoid action on already CCA-containing precursors and on 3′ mature tRNAs, thereby avoiding a futile cycle of removal and resynthesis of CCA sequences. The mechanism of how this is accomplished is of considerable interest because it results in protection of mature tRNA molecules from unwanted damage.

In this study, we examined the mechanism of how the CCA sequence protects against RNase BN activity. Based on sequence alignments and structural model building, we identified a conserved amino acid residue, Arg²⁷⁴, located in the channel leading to the catalytic site of RNase BN that is required for discrimination against mature tRNAs. In addition, we determined the contribution of the CCA sequence itself to the protective effect. We found that two adjacent C residues completely block the action of RNase BN. Furthermore, we were able to show that the presence of both Arg²⁷⁴ and the two C residues prevents CCA-containing RNAs in the channel from moving to the catalytic site. Replacement of Arg²⁷⁴ or either of the two C residues in the CCA sequence allows RNA to proceed to the catalytic site. These data indicate that Arg²⁷⁴ interacts with the C residues of the CCA sequence leading to inhibition of RNase BN action. Based on these findings, we conclude that RNases can have built-in specificities that prevent their indiscriminate action on RNA substrates, thus explaining one mechanism that maintains the stability of mature tRNAs.

EXPERIMENTAL PROCEDURES

Materials

RNA oligonucleotides were synthesized by Dharmacon, Inc. ExpressHyb hybridization solution was purchased from Clontech. DpnI was from Fermentas. [γ-³²P]ATP, 5′-[³²P]pCp, and GeneScreen Plus hybridization transfer membrane were obtained from PerkinElmer Life Sciences. T4 polynucleotide kinase was obtained from New England Biolabs. T4 RNA ligase, calf intestine alkaline phosphatase, Nucway^TM spin columns, and the MEGAshortscript^TM kit were purchased from Ambion Inc. The GeneElute^TM PCR clean-up kit and bis(p-nitrophenyl) phosphate were from Sigma. The KOD Hot Start DNA polymerase was purchased from Novagen. Sequagel for denaturing urea-polyacrylamide gels was obtained from National Diagnostics. The His-Trap HP column was obtained from GE Healthcare. All other chemicals were reagent grade.

Site-directed Mutagenesis, Overexpression, and Purification of the Mutant Protein

The wild type rbn gene was cloned into plasmid pET15b, and its overexpression and purification were described previously (12). Site-directed mutagenesis with primers R1 (CGTCAGCTCGGCATATGATGACAAAGGTTGTCA) and R2 (TGTCATCATATGCCGAGCTGACGTGGGTAATGATTAG) was used to introduce the mutation R274A into a plasmid-encoded rbn gene. KOD Hot Start DNA polymerase was used for PCR. Wild type rbn-containing template plasmid was then digested by DpnI treatment at 37 °C for 2 h. Plasmids containing the mutant rbn gene were purified using a gel extraction kit (Qiagen) and transformed into E. coli BL21(DE3)I⁻II⁻. Mutant protein was overexpressed in E. coli strain BL21(DE3)I⁻II⁻/pLys and purified using the same procedure used previously for purification of wild type His-tagged RNase BN (9, 12). The purity of wild type and mutant RNase BN proteins was determined on an overloaded SDS-polyacrylamide gel (∼3.0 μg of the purified protein). For all of the proteins, a single band at ∼35 kDa was observed without any detectable contaminating bands.

Synthesis and 3′-End Labeling of Mature tRNA

E. coli genomic DNA was used as the template for PCR using KOD Hot Start DNA polymerase to synthesize full-length tRNA^Phe in an in vitro transcription reaction using the MEGAshortscript^TM transcription kit as described (10). The forward primer in the PCR contained the T7 RNA polymerase promoter sequence. PCR products were purified using the GeneElute^TM PCR clean-up kit. Mature tRNA was purified by phenol/chloroform/isoamylalcohol (25:24:1) extraction followed by ethanol precipitation as described (10).

Transfer-RNA^Phe was labeled at its 3′-end with 5′-[³²P]pCp using T4 RNA ligase in the presence of unlabeled ATP at 4 °C for 16 h as described previously (10). Nucway^TM spin column was used to remove unincorporated 5′-[³²P]pCp, and the 3′-terminal phosphate from 3′-[³²P]pCp-labeled tRNA^PheV was removed by calf intestine alkaline phosphatase. After dephosphorylation, 3′-end-labeled tRNA^PheV was purified as described (10).

RNase BN/Z Assay

A typical 30-μl reaction mixture contained 10 mm Tris-HCl, pH 7.5, 200 mm potassium acetate, 5 mm MgCl₂, and 3′-[³²P]pC-labeled tRNAs (∼0.05 μm) or 5′-³²P-labeled model RNA substrates (10 μm) and 0.14 μm purified wild type or mutant RNase BN, except as otherwise stated in the figure legends. Reaction mixtures were incubated at 37 °C. Portions were taken at the indicated times, and the reaction was terminated by the addition of 2 volumes of gel loading buffer (90% formamide, 20 mm EDTA, 0.05% SDS, 0.025% bromphenol blue, and 0.025% xylene cyanol). Reaction products were resolved on 20% denaturing 7.5 m urea polyacrylamide gels and visualized using a STORM 840 phosphorimaging device (GE Healthcare). ImageQuant (GE Healthcare) was used to quantitate the bands.

Phosphodiesterase Assay

Standard reaction conditions for the determination of the phosphodiesterase activity of RNase BN were 20 mm Tris-HCl (pH 7.4), 2 mm bis(p-nitrophenyl) phosphate, 0.6 μm of purified His-tagged wild type or mutant RNase BN, and 5 mm MgCl₂. Release of p-nitrophenol (ϵ = 11,500 m⁻¹ cm⁻¹ at pH 7.4) was continuously monitored for 2 min at 405 nm. One unit of activity corresponds to 1 μmol of p-nitrophenol liberated/min at 37 °C.

Northern Blot Analysis

tRNA^PheV samples digested by either wild type or mutant RNase BN were resolved on a 6% denaturing polyacrylamide gel in 0.5× Tris borate/EDTA buffer and transferred to a nitrocellulose membrane by horizontal transfer for 1.5 h at 150 mA using 0.5× Tris borate/EDTA as the transfer solution. DNA oligonucleotide probes complementary to the 5′-end of the tRNA were ³²P-labeled at their 5′-ends with T4 polynucleotide kinase. Probes were allowed to anneal to the transferred RNA by overnight incubation in ExpressHyb hybridization solution (Clontech Laboratories Inc.), and the detected bands were visualized by PhosphorImager analysis.

RESULTS

In early work from our laboratory, it was found that RNase BN could remove the 3′-terminal residue from tRNA-CU or tRNA-CA in vitro, whereas tRNA-CC and tRNA-CCA were essentially inactive as substrates (8). Subsequent work showed that RNase BN can remove residues following a CCA sequence but that it stops at the CCA sequence in a model RNA substrate (9). We also found that in vivo RNase BN can process the 3′-end of E. coli tRNA precursors without removing their encoded CCA sequence (11). Likewise, the CCA sequence has a protective effect on other members of the RNase Z family (17, 19). All of these data raise the interesting question of what structural features of RNase BN/RNase Z and of the CCA sequence make mature tRNA and CCA-containing precursors resistant to the actions of this RNase, thereby avoiding a futile cycle of removal and resynthesis of the 3′-CCA terminus.

Each C Residue of the CCA Sequence Affects Resistance to RNase BN Action

Because tRNA-CCA and tRNA-CC were essentially resistant to RNase BN, whereas tRNA-CU and tRNA-CA were substrates (8), it suggested that two consecutive C residues are sufficient to prevent RNase BN action. To examine this point in more detail and to evaluate the role of each C residue, we constructed a series of model oligonucleotide RNA substrates with the sequence G₅A₁₂NNA, where N represents C, U, or A. Each of the substrates was ³²P-labeled at its 5′-end. The action of RNase BN on these substrates was assessed by acrylamide gel analysis, and the disappearance of the starting material was quantified (Table 1). In agreement with earlier work (9), the substrate terminating with a CCA sequence was largely resistant to RNase BN under these assay conditions. In contrast, replacement of either C residue with a U or A residue rendered the oligonucleotide sensitive to RNase BN. Substitution of the C residue closer to the 3′-end had a greater effect than replacing the more 5′ C residue, but it is clear that each provides a major protective effect, indicating that two adjacent C residues are required to afford essentially complete protection.

TABLE 1.

Contribution of C residues to protection against the action of RNase BN

Model RNAs (10 μm) with the sequence G₅A₁₂X were labeled with ³²P at their 5′ end as described under “Experimental Procedures” and treated with RNase BN (1.7 μm) for 20 and 40 min. Digestion products were analyzed by 20% denaturing PAGE. The amount of substrate remaining after 20 and 40 min of digestion was quantified using ImageQuant (GE Healthcare). The data are expressed as the percentages of radioactivity present at 0 min for each model RNA. The values shown are the averages of three independent experiments.

Substrate (G5A12NNN)		Substrate remaining
		20 min	40 min
		%
NNN =	CCA	96 ± 2	94 ± 3
	UCA	25 ± 3	6 ± 1
	ACA	21 ± 3	7 ± 2
	CUA	7 ± 1	5 ± 1
	CAA	8 ± 2	4 ± 1

Although the first series of model substrates was designed to mimic mature tRNAs in which the CCA sequence is at the 3′-end of the RNA, we also examined a second group of substrates designed to mimic tRNA precursors. In these model substrates, the CCA sequence or its derivatives are followed by five additional A residues. As shown in Fig. 1, RNase BN removes the A residues following the CCA sequence, but then stops (lanes 7–9). In contrast, when the CCA sequence is replaced by CAA (lanes 2–4), ACA (lanes 10–12), or UUA (lanes 14–16), RNase BN action continues through the residues. Although there is a slight pause in the CAA- and ACA-containing substrates that is not seen with the UUA substrate, it is clear that only the presence of two adjacent C residues as is found in the CCA sequence results in almost a complete stop. These data indicate that RNase BN is essentially blocked by a CCA sequence when it resides in either a mature or a tRNA precursor-like substrate. Based on these findings, we conclude that RNase BN specifically recognizes the adjacent C residues in the CCA sequence, resulting in its inability to remove these nucleotides from an RNA molecule.

FIGURE 1.

Action of RNase BN on CCA-containing and various CCA-less model RNA substrates. Reactions were carried out as described under “Experimental Procedures” using 5′-³²P-labeled single-stranded model RNA substrates with the sequence G₅A₁₂XA₅ (10 μm). Digestion was carried out with purified RNase BN (1.7 μm) in the presence of Mg²⁺, and portions were withdrawn at 0, 20, and 40 min. Digestion products were analyzed by 20% denaturing PAGE. The numbers at the bottom of the gel are the lane numbers, and the numbers on the side are the chain lengths of standards (25 or 18) or of a product (6).

Identification of Amino Acid Residues in RNase BN That May Interact with the CCA Sequence

It is known from the crystal structures of E. coli RNase BN (20) and Bacillus subtilis RNase Z (21–23) that each member of the RNase Z family is a dimer with a core zinc-dependent β-lactamase domain that contains a HXHXDH metal binding motif as well as additional His and Asp residues that also coordinate to the two metal ions. Each subunit possesses a protruding flexible arm that is believed to have a role in tRNA binding. The metal-binding site is thought to be the catalytic site of the protein. A channel leading to the metal-binding site, which we term the catalytic channel, is responsible for the stable binding of the RNA substrate to facilitate the catalytic process. A channel leading from the metal-binding site, called the exit channel, binds the 3′-extra residues of tRNA precursors that are removed by endonucleolytic cleavage.

The fact that RNase BN exoribonucleolytically removes residues following the CCA sequence (Fig. 1) but then stops suggests that the terminal A residue of the CCA sequence is unable to move into the catalytic site. Taken together with the conclusion that the enzyme specifically recognizes the adjacent C residues of the CCA sequence, these data imply that interaction of one or more amino acids within the catalytic channel with the C residues is responsible for the inability of RNase BN to act on CCA-containing RNA substrates. This idea was re-enforced by our findings that mutations of amino acids in the exit channel had no effect on the protective function of the CCA sequence (data not shown). Thus, we focused our attention on amino acids in the catalytic channel that were positioned such that they could interact with the two C residues of the CCA sequence when the RNA chain was bound just short of the catalytic site.

To carry out this analysis, we made use of the known cocrystal structure of B. subtilis RNase Z and tRNA (21–23) to build a model of a tRNA clamped between the binding domain and the catalytic site of E. coli RNase BN (20). The E. coli RNase BN structure (Protein Data Bank entry 2CBN) was aligned with the B. subtilis RNase Z protein (Protein Data Bank entry 4GCW). The tRNA (Protein Data Bank entry 4GCW) was then superimposed on the aligned structure of E. coli RNase BN. Fig. 2 shows residues 69–74 of this tRNA and nearby amino acids. Based on this model building, we initially identified six amino acids in the catalytic channel that were close to the RNA substrate and had the potential to interact with RNA residues. These amino acids, Arg²⁴⁶, Lys²⁴², Met⁸², Ser⁸¹, Ser⁸³, and Trp¹¹¹, were each mutated to alanine to determine whether any might interact with the C residues of the CCA sequence. The resulting proteins were purified as described under “Experimental Procedures.” Measurement of the activity of the mutant proteins revealed that they retain almost full activity on bis(p-nitrophenyl) phosphate. They also could remove the five A residues following the CCA sequence of the model RNA substrate, G₅A₁₂CCAA₅. However, each of the mutant proteins retained the specificity of wild type RNase BN, stopping at the CCA sequence. Thus, none of these six amino acid residues appeared to participate in the anti-determinant function of the CCA sequence, and they were not studied further.

FIGURE 2.

Structure of the RNase BN catalytic channel. The top panel shows the structure of E. coli RNase BN (Protein Data Bank entry 2CBN) in a ribbon representation with the two subunits colored pale green and pale orange. To model the path of an RNA substrate downstream of the RNase BN catalytic site (indicated by the two zinc ions coordinated at the active site shown as blue spheres), we used the structure of tRNA bound to B. subtilis RNase Z (23) (Protein Data Bank entry 4GCW). Residues 69–74 of this tRNA^Thr precursor, positioned on RNase BN by superposition of the 4GCW structure on 2CBN, are shown in a stick representation. The bottom panel shows an expanded view of the catalytic channel leading into the catalytic site. Amino acids in the channel that interact with the RNA are shown. Note that RNA residues Cys72 and Gly71, two and three nucleotides downstream of the catalytic site, are located within H-bonding distance of Arg²⁷⁴ when positioned as a single strand in a syn conformation (base rotated from a glycosyl bond torsion angle of 158.6° in the 4GCW structure to 18°).

We next moved our search even more downstream along the catalytic channel. However, this dramatically increases the number of residues to consider because the channel is not well defined, and residues can also adopt different conformations upon RNA binding. To get around this, we also utilized a structure based sequence alignment. We reasoned that because mature tRNAs are resistant to RNase Z digestion in all organisms, members of the RNase Z family may contain a common motif responsible for the lack of activity on CCA-containing RNAs. To test this idea, a structure based sequence alignment study was performed using T-Coffee (24, 25). The sequences of RNase BN and RNase Z homologues from 22 other species were aligned (Fig. 3). Particular attention was paid to amino acids lining the RNase BN catalytic channel. We found a highly conserved motif, ²⁷⁰HXSXR²⁷⁴, very near the catalytic site of RNase BN that has the potential to interact with residues of tRNA located in the catalytic channel. In fact, His²⁷⁰ in this conserved motif is a known essential amino acid that coordinates to one of the two metal ions. His²⁷⁰ is present in a loop that extends away from the metal-binding site toward the catalytic channel. The two other conserved residues in the motif, Ser²⁷² and Arg²⁷⁴, are also present in this loop. The extensive sequence conservation of this motif, as well as that of nearby amino acids, made it an interesting candidate for additional mutational analysis.

FIGURE 3.

Members of RNase Z family contain a conserved HXSXR motif located near the catalytic site. A structure-based sequence alignment was generated by T-Coffee using sequences of RNase Z proteins from 22 different species. ESPript was used to present the sequence alignment (26). The sequences are from the Uniprot database. The sequences included are: RBN_ECOLI, E. coli RNase BN (entry no. P0A8V0); UniRef90_C4BUZ1, Enterobacteriaceae RNase BN; F7RBB1_SHIFL, Shigella flexneri RNase BN (F7RBB1); B7LM62_ESCF3, Escherichia fergusonii RNase BN (B7LM62); RBN_SALA4, Salmonella agona RNase BN (B5EZJ2); C1M711_9ENTR, Citrobacter sp. RNase BN (C1M711); D6DRB4_ENTCL, Enterobacter cloacae RNase BN (D6DRB4); RBN_KLEP3, Klebsiella pneumonia RNase BN (B5XNW7); H5UXS8_ESCHE, Escherichia hermannii RNase BN (H5UXS8); RBN_ERWT9, Erwinia tasmaniensis RNase BN (B2VHF1); RNZ_BACSU, B. subtilis RNase Z (P54548); NZ_BACAN, Bacillus anthracis RNase Z (Q81M88); NZ_BACC2, Bacillus cereus RNase Z (B71WQ5); RNZ_ENTFA, Enterococcus faecalis RNase Z (Q834G2); RNZ_LACLA, Lactococcus lactis RNase Z (Q9CHT8); RNZ_STRZJ, Streptococcus pneumonia RNase Z (C1CD41); RNZ_SYNY3, Synechocystis sp. RNase Z (Q55132); NZ_MICAN, Microcystis aeruginosa RNase Z (B0JGG3); RNZ_STAEQ, Staphylococcus epidermidis RNase Z (Q5HP47); RNZ_LACAC, Lactobacillus acidophilus RNase Z (Q5FKH3); RNZ2_MOUSE, Mus musculus ElaC2 protein (Q80Y81); and RNZ2_MOUSE, Homo sapiens ElaC2 (Q9BQ52). The stars indicate a conserved motif present in the catalytic channel. Other conserved residues are highlighted including the catalytic site residues Asp²¹² and His²⁷⁰.

Effect of Mutations in the Conserved HXSXR Motif on RNase BN Activity

To examine the role of amino acids in the conserved HXSXR motif on RNase BN activity and specificity, several mutations were made within the motif, and the resulting proteins were purified. These included conversion of His²⁷⁰, Ser²⁷², or Arg²⁷⁴ to alanine. The activity of these mutant proteins on both bis(p-nitrophenyl) phosphate and on model RNA substrates was measured. His²⁷⁰ is part of the active site, and as might be expected, its conversion to alanine caused a loss of RNase BN activity on both bis(p-nitrophenyl) phosphate, a small molecule that is the chromogenic substrate for the phosphodiesterase activity of RNase BN, and model RNAs (data not shown). Mutant S272A was also inactive on model RNAs, and it showed greatly reduced activity on the chromogenic substrate (data not shown). Consequently, we were unable to study either of these mutant proteins. In contrast, mutation of Arg²⁷⁴ to alanine only reduced phosphodiesterase activity ∼25% compared with the wild type protein, and activity against model RNAs was reduced to ∼30% of wild type RNase BN. Most importantly, this mutation changed the specificity of RNase BN.

To explore how the R274A mutation affected the specificity of RNase BN, we made use of the model substrates G₅A₁₂CCA and G₅A₁₂CCAA₅. As shown above and in Fig. 4A, wild type RNase BN is unable to act on G₅A₁₂CCA. (The small amount of product generated is due to contamination of the substrate by shorter molecules lacking a CCA sequence because the full-length substrate is essentially unaltered over a period of 2 h.) On the other hand, the R274A mutant protein digests G₅A₁₂CCA effectively such that >90% is gone in 2 h, generating a range of shorter products. (RNase BN is known to slow down considerably when the product length is reduced to ∼10 nucleotides (9).) On G₅A₁₂CCAA₅ (Fig. 4B), wild type RNase BN removes the five A residues but stops at the CCA sequence. The R274A mutant protein also removes the A residues, but it continues through the CCA sequence, generating the same shorter products as in Fig. 4A. These data show that the R274A mutant RNase BN can act on model RNA substrates that mimic both mature and precursor tRNAs. The findings also indicate that the Arg²⁷⁴ residue in the catalytic channel of RNase BN plays a critical role in the CCA inhibition of RNase BN action, suggesting that Arg²⁷⁴ interacts with the C residues of the CCA sequence.

FIGURE 4.

Activity of wild type and R274A mutant RNase BN on CCA-containing mature and precursor model tRNA substrates. 5′-³²P-labeled single-stranded model RNA substrates (10 μm) with the sequence G₅A₁₂CCA (A) or G₅A₁₂CCAA₅ (B) were digested with wild type or mutant RNase BN (13.6 μm) as described under “Experimental Procedures.” Portions were withdrawn at the indicated times and analyzed by denaturing PAGE. The values at the bottom of each lane indicate the percentage of the initial substrate remaining. Note that the band labeled 6 is the limit product generated by the WT enzyme.

Action of R274A Mutant RNase BN on tRNA

RNase BN does not act on mature tRNA either in vitro or in vivo because of the presence of a CCA sequence (10, 11). From the above experiments, it is evident that the R274A mutant protein is not affected by a CCA sequence on model RNA substrates. Thus, it was of considerable interest to determine whether the mutant RNase BN protein can also act on tRNA substrates particularly because RNase BN functions primarily as an exoribonuclease on model RNA substrates, but as an endoribonuclease on tRNA precursors in the presence of Mg²⁺, cleaving after the CCA sequence (10, 11). It does not work at all on mature tRNAs (10, 11). To examine how the R274A mutant protein acts on a tRNA substrate, full-length tRNA^Phe from E. coli (Fig. 5A) was synthesized as described under “Experimental Procedures” and was digested with either wild type or mutant RNase BN. The original substrate and cleavage products were detected by Northern blot analysis using a labeled DNA oligonucleotide complementary to the 5′-end of tRNA^Phe. As shown in Fig. 5B, full-length tRNA was resistant to wild type RNase BN; however, mutant RNase BN removes the CCA sequence from the 3′-end of the tRNA to generate a three-nucleotide shorter product. These data indicate that the inhibitory effect of the CCA sequence of tRNA substrates on RNase BN activity is relieved by mutation of Arg²⁷⁴.

FIGURE 5.

Action of wild type and R274A mutant RNase BN on tRNA^Phe. A, structure of tRNA^PheV is shown with the 3′-terminal CCA sequence in bold type. The 3′-terminal [³²P]C residue used in C is in parentheses. B, wild type and mutant RNase BN (1.12 μm) were used to digest full-length tRNA^Phe (∼0.05 μm). Portions were withdrawn at 0, 1.5, and 3 h. Cleavage products were analyzed by 6% denaturing PAGE, followed by Northern blotting using a 5′-probe. M, full-length tRNA. C, full-length tRNA (∼0.05 μm), labeled with [³²P]pC at its 3′-end, was treated with wild type and mutant RNase BN (1.12 μm) for 0, 1.5, and 3 h. Digestion products were analyzed by 20% denaturing PAGE. P, precursor tRNA.

To obtain additional information on how the Arg²⁷⁴ mutant protein removes the CCA sequence, tRNA^Phe was labeled at its 3′-end with 5′-[³²P]pCp and T4 RNA ligase. Calf intestine alkaline phosphate was then used to remove the 3′-terminal phosphate, and the resulting tRNA-CCA-[³²P]C was used as the substrate in reactions with wild type and mutant RNase BN. The gel in Fig. 5C shows that the wild type protein generates CMP, because it can remove the single extra nucleotide but is unable to remove any part of the CCA sequence. In contrast, a tetranucleotide product was observed using mutant RNase BN, and no CMP was generated. The additional bands seen with the R274A mutant have not been identified but may represent some nonspecific cleavages. These data demonstrate that the R274A mutant protein removes the CCA sequence from full-length tRNA endonucleolytically, cleaving after the discriminator nucleotide as is usual for members of the RNase Z family when the CCA sequence is absent. Thus, the CCA sequence does not display its usual inhibitory properties in the presence of the mutant RNase BN.

Use of Phosphodiesterase Assay to Explain How the CCA Sequence Inhibits RNase BN

The data presented to this point indicate that both the substrate (CC sequence) and the enzyme (Arg²⁷⁴) contribute to the inability of RNase BN to act on CCA-containing RNA molecules. To explain the mechanism of this inhibition, we hypothesized that Arg²⁷⁴, located in the catalytic channel, interacts with the two adjacent C residues of the CCA sequence and prevents the 3′ terminus of the RNA molecule from moving into the catalytic site (Fig. 6). This hypothesis predicts that when both Arg²⁷⁴ and the CC sequence are present, the catalytic site will remain empty, whereas when either determinant is absent, the RNA substrate will move into the catalytic site, and RNase action will proceed (Fig. 6). To test this prediction, we developed an assay using bis(p-nitrophenyl) phosphate. When the catalytic site is unoccupied because the RNA substrate is prevented from entering, bis(p-nitrophenyl) phosphate is able to enter, and phosphodiesterase activity will be observed. However, when the RNA substrate can move into the catalytic site, phosphodiesterase activity will be inhibited.

FIGURE 6.

Schematic representation of the cleavage of CCA-containing and CCA-less tRNAs at the catalytic site of RNase BN. cat site is the catalytic site of RNase BN, and N represents either A or U residues replacing C residues in the CCA sequence. The arrows indicate the sites of RNA cleavage after the discriminator nucleotide (D) of tRNA. 1, Arg²⁷⁴ of wild type RNase BN interacts with adjacent C residues at the 3′-end of tRNA and prevents tRNA from moving into the catalytic site. 2 and 3, mutation of Arg²⁷⁴ to Ala (step 2) or substitution of adjacent C residues with A or U residues (step 3) allows the 3′-end of tRNA to move into the catalytic site enabling cleavage to occur after the discriminator nucleotide. Bis(p-nitrophenyl) phosphate, the phosphodiesterase (PDase) substrate, binds to the catalytic site when Arg²⁷⁴ and CCA are present (1) but is unable to bind when either one of these two determinants is absent (steps 2 and 3).

This assay can distinguish between an empty and an occupied catalytic site and can be used to examine the mechanism of inhibition by CCA-containing RNAs, as is shown in Fig. 6. Wild type RNase BN phosphodiesterase activity is essentially unaffected by increasing concentrations of G₅A₁₂CCA, whereas the R274A mutant protein is dramatically inhibited (Fig. 7A). However, when a different RNA oligonucleotide is used, G₅A₁₂C, which lacks two adjacent C residues, the phosphodiesterase activities of both the wild type and the mutant RNase BN are inhibited (Fig. 7A). These data support our conclusion that when Arg²⁷⁴ and a CCA sequence on the RNA substrate are both present, the RNA cannot move into the catalytic site, allowing bis(p-nitrophenyl) phosphate to enter and phosphodiesterase activity to proceed. However, the absence of either Arg²⁷⁴ or the CCA sequence allows the substrate RNA to enter the catalytic site, resulting in inhibition of phosphodiesterase activity. Based on this information, we conclude that interaction of Arg²⁷⁴ with the adjacent C residues prevents RNA in the channel from moving into the catalytic site, thus providing an explanation for why CCA-containing RNAs are resistant to RNase BN action.

FIGURE 7.

Phosphodiesterase activity of wild type and R274A mutant RNase BN in the presence of RNA oligonucleotides. A, phosphodiesterase activity of wild type and R274A mutant RNase BN was measured as described under “Experimental Procedures” in the presence of the indicated concentrations of model RNAs. B, phosphodiesterase activity of wild type RNase BN in the presence of model RNAs with different 3′-ends, used at the indicated concentrations. Phosphodiesterase activity in the absence of RNA was set at 100. The values shown are the averages of three independent experiments.

The phosphodiesterase assay also allowed us to examine a variety of other nucleotide sequences for their ability to inhibit RNase BN action (Fig. 7B). Each of the model RNAs was based on the sequence G₅A₁₂ and was followed by a two- or three-nucleotide additional sequence. Compared with the CCA sequence, which displayed little inhibition of phosphodiesterase activity, all the other sequences are strongly inhibitory. However, within this grouping, UCA was least inhibitory, followed by ACA. These data indicate the more 3′ terminal C residue can, at least partly, retard movement of the RNA chain into the catalytic site, a result in complete agreement with the data in Table 1, which showed protection against RNase BN by different nucleotide sequences. A C residue in the more 5′ position, as in CUA or CAA, was more inhibitory of phosphodiesterase activity, and an AA sequence was the most inhibitory (Fig. 7B). These data provide a measure of the ability of different sequences to interact with Arg²⁷⁴ and to thereby inhibit movement into the catalytic site. Their close agreement with the protection data shown in Table 1 increases our confidence in these comparative values.

DISCUSSION

The data presented in this study define the parameters important for inhibition of RNase BN activity by CCA-containing RNAs, and they provide a mechanism that explains this inhibition. Thus, we showed that the two adjacent C residues of the CCA sequence and Arg²⁷⁴ in the catalytic channel are both necessary for effective inhibition. Substitution of either of the two C residues with an A or U residue or mutation of Arg²⁷⁴ to alanine abolishes the discrimination of RNase BN against CCA-containing RNAs. Although substrate RNAs with a single C residue slightly retard the action of RNase BN, the overall effect on RNase BN is minimal. Based on this information, we propose that there is a strong interaction between Arg²⁷⁴ in the catalytic channel of RNase BN and the two C residues of the CCA sequence. This interaction prevents the movement of CCA-containing RNA chains into the catalytic site (Fig. 6), and this was shown directly by the finding that the phosphodiesterase substrate, bis(p-nitrophenyl) phosphate was able to enter the catalytic site and be hydrolyzed even when such an RNA was present in the catalytic channel.

Interestingly, in earlier work (21), it was suggested that a loop structure located between strands β1 and β2 in the B. subtilis RNase Z was responsible for the protective effect of the CCA sequence. A highly conserved amino acid, Pro¹³, in that loop was also suggested to play an important role in the discrimination of cytosine residues from other bases. However, we could not obtain any evidence for involvement of this structure in the case of E. coli RNase BN. In fact, we have found that removal of amino acid residues Ser⁹ to Arg¹⁷ in this loop has no effect on the inability of RNase BN to remove a CCA sequence from a CCA-containing model RNA. Likewise, the loop mutant acts similarly to wild type RNase BN on another model RNA substrate in removing five A residues following a CCA sequence but then stopping (unpublished observations). Therefore, it is unlikely that either residues in the loop or Pro¹³ are involved in the CCA inhibition of RNase BN. However, further work is needed to clarify the apparent differences between the E. coli and B. subtilis enzymes.

Arginine is one of the most prevalent amino acids present in nucleic acid-binding domains of proteins. The reasons behind this fact include the length of its side chain, its capacity to interact in different conformations, and its ability to generate favorable hydrogen-bonding geometries (27). We attempted to examine whether Lys could replace Arg in position 274. Unfortunately, the derivative was inactive (unpublished observation). Other amino acids that are not as efficient as arginine in interacting with nucleic acids either lack long side chains or can only interact in one configuration (27). The importance of Arg for interaction with C residues in RNA molecules has been observed in several systems. In one example, arginine is part of the RNA-binding motif of an evolved PUF protein and is specifically involved in recognition of cytosine residues (28, 29). Arginine also forms direct hydrogen bonds to cytosine during the selection of adenine over cytosine by CCA-adding enzyme at position 76 of tRNA (30, 31).

One point that remains to be resolved is whether Arg²⁷⁴ is the sole amino acid interacting with the two C residues of the CCA sequence or whether other amino acid residues might be involved. Clearly, removal of Arg²⁷⁴ is sufficient for eliminating protection by the CCA sequence, and the bidentate structure of Arg may enable it to interact with both C residues. In fact, such bidentate interactions are seen in DNA-protein interaction (27). On the other hand, because mutations of other possible interacting amino acids, such as His²⁷⁰ and Ser²⁷², led to loss of activity, it was not possible to assess their roles. Also some other interacting amino acid might not have been examined. Cocrystal structures of RNase BN and CCA-containing RNA will be necessary to settle this question.

Mature tRNAs are resistant to digestion by RNases both in vitro and in vivo. Their complex secondary structures, the presence of an amino acid that protects the 3′ terminus, and, as shown here, the universal 3′-CCA sequence of tRNA, all contribute to the stability of tRNAs. Members of the RNase Z family, which are responsible for 3′-end maturation of tRNA precursors in many organisms, are important for survival in most of these organisms. However, these enzymes had to evolve a mechanism to distinguish mature, CCA-containing tRNAs from tRNA precursors that lack this sequence. As we have shown, members of the RNase Z family contain a strictly conserved motif, HXSXR, near the catalytic site whose Arg residue in RNase BN is responsible for the lack of action on mature tRNAs. We presume that a similar mechanism operates in other members of the RNase Z family. In the absence of Arg²⁷⁴, RNase BN removes residues after the discriminator nucleotide in a process identical to the removal of 3′-extra residues from CCA-less tRNA precursors by RNase Z. Thus, this family of RNases has a built-in specificity that avoids indiscriminate action on CCA-containing RNAs.

However, this is not the only mechanism for avoiding action on CCA-containing RNAs. Another RNase, termed RNase T, which is involved in the exonucleolytic 3′-end maturation of E. coli tRNA precursors, is also inhibited by the presence of a CC sequence (32). In this case, the mechanism of inhibition is completely different. Four phenylalanine and one glutamic acid residue of RNase T form a “C-filter” before the catalytic site of RNase T that screens out C residues (7). This leads to inhibition of RNase T on RNA substrates with C residues at their 3′-end, although the 3′ terminal A residue can be removed (33). It would not be surprising if additional mechanisms have evolved among the many RNases present in cells to avoid action on mature tRNAs as distinguished from tRNA precursors. The study presented here adds considerably to our knowledge of the RNase Z family of enzymes and of how they avoid acting on CCA-containing RNAs.