Purification and Characterization of the tRNA-processing Enzyme RNase BN*

RNase BN, a tRNA-processing enzyme previously shown to be required for the 3′-maturation of certain bacteriophage T4-encoded tRNAs, was overexpressed and purified to near homogeneity from Escherichia coli. The purified enzyme, which is free of nucleic acid, is an α2-dimer with a molecular mass of ∼65 kDa. RNase BN displays a number of unusual catalytic properties compared with the other exoribonucleases of E. coli. The enzyme is most active at pH 6.5 in the presence of Co2+ and high concentrations of monovalent salts. It is highly specific for tRNA substrates containing an incorrect residue within the universal 3′-CCA sequence. Thus, tRNA-CU and tRNA-CA are effective substrates, whereas intact tRNA-CCA, elongated tRNA-CCA-Cn, phosphodiesterase-treated tRNA, and the closely related tRNA-CC are essentially inactive as substrates. RNA or DNA oligonucleotides also are not substrates. These data indicate that RNase BN has an extremely narrow substrate specificity. However, since tRNA molecules with incorrect residues within the -CCA sequence are not normally produced in E. coli, the role of RNase BN in uninfected cells remains to be determined.

tRNA genes in Escherichia coli are normally transcribed as precursor molecules that require processing at both their 5Јand 3Ј-ends to generate the mature functional forms (1). Maturation of the 5Ј-end of tRNA precursors is carried out by a single endoribonuclease, RNase P, whereas 3Ј-processing generally is a multistep process requiring the action of both endoand exoribonucleases to remove the extra nucleotides following the encoded 3Ј-terminal -CCA sequence (1)(2)(3). A single pathway for 3Ј-maturation does not exist. Rather, it appears to proceed in a stochastic manner such that any one of five exoribonucleases (RNase II, D, BN, T, or PH) may act to complete the 3Ј-maturation process (2)(3)(4).
The process is more complicated in bacteriophage T4-infected Escherichia coli. In this situation, eight new tRNAs encoded by the T4 genome are synthesized (6). Four of these tRNA precursors lack the usually encoded -CCA sequence, and during their 3Ј-maturation, the incorrect residues must be removed and be replaced by the -CCA sequence through the action of tRNA nucleotidyltransferase (5,6). In certain mutant E. coli strains, BN and CAN, this process does not occur (7,8).
Thus, these strains do not support the growth of a mutant T4 phage (BU33) because a phage-encoded suppressor tRNA Ser , required for the translation of an amber mutation in a BU33 head protein, cannot be processed to its mature form due to a defect in 3Ј-maturation (9). The molecular basis for this phenotype in strains BN and CAN is the deficiency of the exoribonuclease RNase BN (10,11). Thus, in contrast to uninfected cells, in which at least five exoribonucleases can remove the extra residues following the -CCA sequence during 3Ј-maturation of tRNA precursors, only RNase BN appears able to remove incorrect residues within the -CCA sequence during phage T4 tRNA maturation.
To learn more about the structure, function, and specificity of RNase BN, we first identified and cloned the rbn gene encoding the enzyme (12). We discovered that rbn is nonessential in E. coli and that it encodes a polypeptide of 32.8 kDa (12). In this paper, we describe the overexpression, purification, and characterization of RNase BN.
Culture Conditions-Cells were routinely grown at 37°C in YT medium or on YT plates (14). Ampicillin, when added, was present at a concentration of 200 or 300 g/ml. Growth in liquid medium was followed by absorbance measurements at 600 nm. Materials-[ 3 H]Poly(A), blue dextran 2000, Ultrogel-AcA44, and DEAE-Sephadex A-50 were obtained from Amersham Pharmacia Biotech. [ 14 C]ATP and [␥-32 P]ATP were obtained from NEN Life Science Products. Unlabeled poly(A) used to dilute the radioactive material was obtained from Sigma. Bacterial alkaline phosphatase was purchased from Worthington. N-Ethylmaleimide, 5,5Ј-dithiobis(2-nitrobenzoic acid), p-hydroxymercuribenzoate, and protein size markers were obtained from Sigma. Mercuric chloride was obtained from J. T. Baker Inc. Hydroxylapatite HT, Affi-Gel blue, and cellulose were purchased from Bio-Rad. All other chemicals were reagent-grade.
RNase BN Assay-RNase BN removes the 3Ј-terminal mononucleotide AMP from the phage tRNA precursor analogue tRNA-C[ 14 C]A. This substrate was prepared as described previously (15). Reaction mixtures of 100 l contained 20 mM Hepes, pH 6.5, 0.2 mM CoCl 2 , 200 mM KCl, 8 or 16 g of tRNA-C[ 14 C]A (10 4 cpm/nmol), and cell extract or purified enzyme. Samples were incubated for 30 min at 37°C, and acid-soluble radioactivity was determined. One unit of RNase BN is the amount of enzyme that releases 1 nmol of AMP in 1 h.
Overexpression of RNase BN-Plasmid pBSrbn ϩ carrying the rbn gene cloned under control of a T 7 promoter was transformed into strain BL21(DE3) pLys. Bacteriophage DE3 contains the gene for T 7 RNA polymerase controlled by the lacUV5 promoter, which is inducible by isopropyl-␤-D-thiogalactopyranoside. The pLys plasmid carries the gene for T 7 lysozyme, which inhibits basal levels of T 7 transcription in the uninduced cell. Cells were grown to an absorbance of 0.5 at 600 nm in YT medium containing 200 g/ml ampicillin. To maintain the rbn ϩ plasmid, cells were then harvested by centrifugation at 5000 rpm for 5 min, washed with YT medium to remove ␤-lactamase, and then resuspended in fresh YT medium containing 300 g/ml ampicillin. The cul-* This work was supported by Grant GM16317 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ʈ To whom correspondence should be addressed. Tel.: 305-243-3150; Fax: 305-243-3955; E-mail: mdeutsch@med.miami.edu. ture was grown to an absorbance of just under 1 at 600 nm. The cells were spun down, washed with YT medium, and resuspended in YT medium containing 300 g/ml ampicillin and 0.4 mM isopropyl-␤-Dthiogalactopyranoside to induce T 7 RNA polymerase. Cells were then grown for an additional 4 h before harvesting.
Preparation of Extracts-All steps of the purification procedure were carried out at 4°C in the cold room. Ten grams of induced cells (wet weight) were suspended in 2.5 volumes of Buffer A (10 mM Tris-HCl, pH 7.5, 1 mM dithiothreitol, 0.1 mM EDTA, and 0.1 mM phenylmethylsulfonyl fluoride) and passed twice through an Amicon French press at 10,000 p.s.i. DNase I was added to the resulting extract at a concentration of 10 g/ml, and the sample was incubated at 0°C for 45 min. The extract was then centrifuged at 20,000 rpm for 40 min, and the soluble fraction obtained was centrifuged at 45,000 rpm for an additional 120 min in a Ti-70.1 rotor in a Beckman L8 -70M ultracentrifuge.
DEAE-Sephadex Chromatography-The supernatant fraction was diluted 2-fold with Buffer A containing 10% glycerol to lower the ionic strength and passed through a 0.22-syringe filter. The resulting 12-ml sample was applied to a DEAE-Sephadex A-50 column (2.8 ϫ 13 cm) and washed with Buffer A containing 50 mM KCl until the A 280 was Ͻ1. The column was eluted using a linear gradient of 50 -700 mM KCl in Buffer A. Eighty 1-ml fractions were collected. Fractions 51-73, with the highest RNase BN activity, were combined. This material was concentrated 5-fold with an Amicon Diaflo YM-10 membrane and dialyzed against 2 liters of Buffer B (5 mM potassium phosphate, pH 7.5, 1 mM dithiothreitol, 0.1 mM EDTA, and 0.1 mM phenylmethylsulfonyl fluoride).
Hydroxylapatite Chromatography-The dialyzed 6.7-ml sample was applied to a hydroxylapatite column (2.8 ϫ 13 cm) that contained 10% cellulose to improve flow rates and that had been equilibrated with Buffer B. The column was eluted with a linear gradient of 5-50 mM phosphate in Buffer B. Eighty 1-ml fractions were collected. Fractions 35-55, with the highest activity, were combined and concentrated ϳ10fold to a final volume of 2 ml by ultrafiltration as described above.
Ultrogel-AcA44 Chromatography-The concentrated sample was applied to an Ultrogel-AcA44 column (1.8 ϫ 78 cm) equilibrated with Buffer A containing 1 M KCl. The column was eluted with the same buffer. Eighty 2-ml fractions were collected. Fractions 42-74, containing RNase BN activity, were combined, dialyzed against 2 liters of Buffer A, and concentrated ϳ30-fold by ultrafiltration to a final volume of 2 ml.
Affi-Gel Blue Chromatography-The sample was applied to an Affi-Gel blue column (0.8 ϫ 10 cm) and washed with Buffer A containing 50 mM KCl until the A 280 was Ͻ1. The column was eluted with a linear gradient of 0 -1 M KCl in Buffer A. Fifty 75-l fractions were collected, and the active fractions, in tubes 25-27 (see Fig. 1), were combined and frozen at Ϫ80°C. Side fractions, 19 -24 and 28 -33, were also combined and concentrated 2-fold by ultrafiltration. This latter sample was split into three portions and stored at Ϫ80, Ϫ20, and Ϫ20°C with 50% glycerol to ascertain the best storage conditions.
Other Assays-Protein was determined by the method of Bradford (16) with bovine serum albumin as a standard or by absorbance at 280 nm.

RESULTS
Overexpression and Purification of RNase BN-Overexpression of RNase BN is deleterious to cell growth and generally leads to loss of the expression plasmid. 1 However, use of the inducible T 7 expression system (with changing of the medium to remove ␤-lactamase and to restore ampicillin levels) resulted in a 25-fold overproduction of RNase BN during the 4-h induction period. In several experiments, it was found that this amount of time gave maximum expression of RNase BN.
As shown in Table I, RNase BN was purified from a highspeed supernatant fraction using a series of four chromatographic steps. The overall purification from the overexpressed extract was 46-fold, equivalent to ϳ1000-fold from a normal extract. The apparent recovery was 50%; however, as is clear from the increased activity in steps 2 and 3, RNase BN activity in the supernatant fraction actually is inhibited 2-3-fold, presumably by the large amount of nucleic acid present. This inhibition was verified directly by mixing of purified RNase BN with the initial supernatant fraction (data not shown).
Purified RNase BN was stored for 1 month under a variety of conditions and then re-assayed to determine optimal conditions for storage. Although the enzyme was relatively stable under all the conditions tested, it was most stable at Ϫ20°C in the presence of 50% glycerol or at Ϫ80°C after rapid freezing in a dry ice/ethanol bath. Under these conditions essentially no RNase BN activity was lost during the period of storage.
Purity of RNase BN-Based on the close correspondence between RNase BN activity and absorbance at 280 nm upon elution from Affi-Gel blue (Fig. 1), it was apparent that RNase BN was quite pure. This was confirmed by SDS-polyacrylamide gel electrophoresis (Fig. 2). Overexpression of RNase BN is evident by comparison of lanes 6 and 7. The progressive purification of RNase BN can be seen in lanes 6 to 2. We estimate that the most purified material (lane 2) was at least 95% pure.
Spectral analysis of RNase BN, carried out in the range of 200 -350 nm, revealed no unusual peaks (data not shown). The A 280 /A 260 ratio was 1.89, indicating that purified RNase BN is free of nucleic acid. Thus, RNase BN does not require nucleic acid for activity.
Molecular Mass and Subunit Structure-SDS-polyacrylamide gel electrophoresis (Fig. 2) indicated that purified RNase BN contains only a single species of polypeptide chain with a molecular mass of ϳ37 kDa. This is in good agreement with the predicted size of RNase BN of 32.8 kDa based on the sequence of the rbn gene (12).
Based on gel filtration on Ultrogel-AcA44 (step 4 of the purification procedure) and assuming that it is a globular protein, the native molecular mass of RNase BN is ϳ65 kDa. These data strongly support the conclusion that RNase BN functions as an ␣ 2 -dimer, as previously suggested from earlier genetic studies in which an interruption-deletion mutant of RNase BN displayed a strong dominant-negative effect (12).
Effect of Sulfhydryl Reagents-Based on the nucleotide sequence of the rbn gene, RNase BN contains 2 cysteine residues/ polypeptide chain (12). To ascertain whether these residues are present in their reduced form and whether they might play a role in enzyme activity, purified RNase BN was incubated with the sulfhydryl reagents N-ethylmaleimide, 5,5Ј-dithiobis(2-nitrobenzoic acid), p-hydroxymercuribenzoate, and mercuric chloride. As presented in Table II, each of these reagents was 1 C. Callahan and M. P. Deutscher, unpublished observation. found to be a potent inhibitor of RNase BN, decreasing RNase BN activity ϳ90% upon preincubation for 30 min at 37°C. These data show that at least 2 of the cysteine residues in native RNase BN are in the reduced form and that cysteine residues are required for enzyme activity. Whether these residues are directly required for catalysis or indirectly by maintaining structure, such as the dimeric form, remains to be determined. Temperature Sensitivity of RNase BN-Samples of RNase BN were incubated at various temperatures for 10 min prior to determining the amount of activity remaining. The data in Fig.  3 show that RNase BN remained stable for 10 min at temperatures up to 45°C, and at 50°C, still retained 66% of its activity. However, at 60°C, 90% of the activity was lost in 10 min. Additional experiments revealed that RNase BN retained ϳ25% of its activity upon incubation for 40 min at 50°C (data not shown). These data indicate that RNase BN is moderately stable at elevated temperatures.
Optimal Conditions for RNase BN Activity-RNase BN was assayed under a variety of conditions to assess the requirements for optimal activity. Assay at pH values between 6.0 and 8.0 in Mes 2 and Hepes buffers indicated that RNase BN is most active at pH 6.5. Activity at pH 8.0 was 70% of that at pH 6.5, whereas that at pH 6.0 was 95% of that at pH 6.5. The relatively low pH optimum for RNase BN is unusual among the exoribonucleases of E. coli, which generally function optimally at more alkaline pH values.
Among the divalent cations tested, significant activity was observed only with Co 2ϩ and Mg 2ϩ , with Co 2ϩ being considerably more effective at 0.2 mM (Table III). A more detailed examination of these two cations (Fig. 4) revealed that at its optimum (0.2 mM), Co 2ϩ was approximately twice as effective as Mg 2ϩ at its optimum (1 mM). Both cations displayed quite sharp optima. These findings are quite unusual as Co 2ϩ has not been observed previously to function as such an effective cofactor for RNases.
RNase BN also requires a monovalent cation for activity. With K ϩ as the cation, the optimal concentration was 200 -400 mM (Fig. 5). At 200 mM, the monovalent cation requirement could be satisfied by Li ϩ , which was as effective as K ϩ ; Na ϩ and Rb ϩ were ϳ70% as effective, and Cs ϩ was ϳ30% as effective. A monovalent cation is also required when Mg 2ϩ is the divalent cation. Although a number of other E. coli exoribonucleases are stimulated by monovalent cations, the absolute requirement of RNase BN for a monovalent cation is unusual.
Substrate Specificity of RNase BN-RNase BN is highly spe- 2 The abbreviation used is: Mes, 4-morpholineethanesulfonic acid.   cific in its mode of action (Table IV). tRNA molecules in which a nucleotide residue within the normal 3Ј-terminal -CCA sequence has been replaced such as tRNA-CA and tRNA-CU were highly active as substrates, whereas all the other RNA molecules tested, even those with very similar structures, were essentially inactive. Thus, intact tRNA-CCA, diesterasetreated tRNA (which lacks 3Ј-terminal residues), and tRNA-CCA-Cn (which contains residues following the normal 3Ј terminus) were all extremely poor substrates. Even the closely related tRNA-CC, which differs from the active substrates by a single terminal residue, was almost inactive. Given this high degree of specificity against such similar molecules, it is not surprising that rRNA and poly(A) also were poor substrates (Table IV).
A second group of substrates consisting of RNA and DNA oligonucleotides 11-17 residues in length were also tested as substrates (Table V). The sequences of the RNA oligonucleotides are those of the 3Ј terminus of E. coli tRNA 1 Tyr and its precursor molecules with 3 or 6 extra 3Ј-residues. These molecules were labeled at their 5Ј termini with 32 P, and RNase BN action was assessed on a 20% acrylamide gel, which can detect shortening by even a single nucleotide residue. Based on this analysis, none of these molecules were substrates for RNase BN under standard assay conditions.
Products of RNase BN Action-Based on earlier work in which relatively crude preparations of RNase BN were found to release UMP from tRNA-CU, it was suggested that this enzyme is an exoribonuclease (10). To confirm this conclusion and to ensure that the release of mononucleotide was not due to a secondary reaction resulting from the presence of a contaminating activity, the products of the reaction catalyzed by highly purified RNase BN were examined. Thus, using tRNA-C[ 14 C]A as a substrate, both the acid-soluble and tRNA products were determined (Table VI).
If RNase BN were an exoribonuclease, the 14 C-labeled acidsoluble product would be expected to be AMP, which, upon treatment with alkaline phosphatase, would be converted to the uncharged molecule [ 14 C]adenosine. If, on the other hand, RNase BN were an endoribonuclease, any acid-soluble oligonucleotides produced would remain charged after treatment with the phosphatase and would elute with the "nucleotide" fraction. Columns of the anion exchanger Dowex AG 1-X2 were used to separate radioactive nucleoside from nucleotide species. As shown in Table VI, after phosphatase treatment, 93% of the acid-soluble radioactivity was eluted from the ion-exchange column with water, consistent with its being the nucleoside adenosine.
To further verify the exoribonucleolytic action of RNase BN, the tRNA product was treated with tRNA nucleotidyltransferase in the presence of [ 3 H]CTP. Removal of AMP from tRNA-CA would generate tRNA-C, a substrate for the incorporation of CMP by tRNA nucleotidyltransferase. As also shown in Table VI

DISCUSSION
In this paper, the purification of RNase BN in its native untagged form has been described. RNase BN is now the seventh of the eight known E. coli exoribonucleases (all except RNase R) to have been purified to near homogeneity. Interestingly, RNase BN is the sixth of the eight enzymes to exist as a multimer and the fifth to be an ␣ 2 -dimer (polynucleotide phosphorylase is a trimer). Inasmuch as there is relatively little overall sequence similarity or similarity in size among the exoribonucleases as a group, it is intriguing that there is commonality in their quaternary structure. Perhaps, this represents some consistency among these enzymes in their interactions with substrate. As yet, there is very little detailed structural information about this class of enzymes, so the significance of their multimeric forms will have to await further work.
In contrast to its shared structural similarity, RNase BN is unique among the E. coli exoribonucleases in a number of its catalytic properties. These include a low pH optimum of 6.5, a strong preference for Co 2ϩ as the required divalent cation, and a requirement for a high concentration of monovalent ions or for elevated ionic strength. In addition, RNase BN displays a highly unusual and strict substrate specificity. Of the molecules tested, only those containing alterations within the 3Јterminal -CCA sequence serve as effective substrates, resulting in removal of the incorrect nucleotide. Other closely related molecules, even tRNA-CC, are inactive or very poorly active as substrates. This unusual substrate specificity makes RNase BN ideally suited for its, so far, only known biological role: the maturation of a subset of bacteriophage T4 tRNA precursors that lack the -CCA sequence (5). RNase BN is essential for the 3Ј-processing of at least some of this group of tRNAs (6,9). This, of course, raises the interesting question of what the role of RNase BN is in the uninfected E. coli cell. Clearly, this enzyme has not been maintained solely for its role during phage infection. It is known that RNase BN can contribute to the maturation of cellular tRNAs when other processing exoribonucleases are absent (3,4). However, it is the poorest of all the exoribonucleases in this regard. RNase BN is not an essential enzyme in E. coli, and its absence has no effect on cell growth under usual laboratory conditions (12). It is possible that RNase BN is important during a stress response. In fact, RNase BN expression is affected by manipulation of certain proteins altered during heat shock, 1 but further work is necessary to clarify this effect. One might expect that the unusual substrate specificity of RNase BN would be a clue to its cellular function. Yet in E. coli, in contrast to higher organisms, all tRNA genes encode the universal -CCA sequence (17). Thus, there are no known conditions in which incorrect residues would be present in this sequence other than in the case of errors during transcription.
From the foregoing discussion, it is evident that there is much remaining to be learned about the structure and function of RNase BN. Now, with the availability of purified protein and of mutant strains lacking RNase BN (12), continued progress can be expected in studies of this interesting enzyme.