Determination of the Catalytic Parameters of the N-terminal Half of Escherichia coli Ribonuclease E and the Identification of Critical Functional Groups in RNA Substrates*

Ribonuclease E is required for the rapid decay and correct processing of RNA in Escherichia coli. A detailed understanding of the hydrolysis of RNA by this and related enzymes will require the integration of structural and molecular data with quantitative measurements of RNA hydrolysis. Therefore, an assay for RNaseE that can be set up to have relatively high throughput while being sensitive and quantitative will be advantageous. Here we describe such an assay, which is based on the automated high pressure liquid chromatography analysis of fluorescently labeled RNA samples. We have used this assay to optimize reaction conditions, to determine for the first time the catalytic parameters for a polypeptide of RNaseE, and to investigate the RNaseE-catalyzed reaction through the modification of functional groups within an RNA substrate. We find that catalysis is dependent on both protonated and unprotonated functional groups and that the recognition of a guanosine sequence determinant that is upstream of the scissile bond appears to consist of interactions with the exocyclic 2-amino group, the 7N of the nucleobase and the imino proton or 6-keto group. Additionally, we find that a ribose-like sugar conformation is preferred in the 5′-nucleotide of the scissile phosphodiester bond and that a 2′-hydroxyl group proton is not essential. Steric bulk at the 2′ position in the 5′-nucleotide appears to be inhibitory to the reaction. Combined, these observations establish a foundation for the functional interpretation of a three-dimensional structure of the catalytic domain of RNaseE when solved.

RNaseE is a single-stranded-specific endoribonuclease (1)(2)(3) that is required for the rapid decay of many if not most transcripts in Escherichia coli (for reviews, see Refs. 4 -6) and the correct processing of precursors of, for example, ribosomal and transfer RNAs (7,8). RNaseE generates products terminating in a 3Ј-hydroxyl and a 5Ј-phosphate (9) indicating that the reaction involves the attack of water on the susceptible phosphodiester bond followed by scission of 3Ј-oxygen-phosphorus bond. The RNaseE polypeptide consists of 1061 residues (Fig. 1,   panel A). The N-terminal half (NTH) 1 of RNaseE is sufficient for basal catalytic activity (10) that is able to support cell viability (11). The RNaseE-catalyzed hydrolysis of RNA in vivo is influenced by the C-terminal half (12)(13)(14), which contains an arginine-rich RNA-binding domain (10, 14 -17) and several sites (18,19) that serve in the assembly of the E. coli RNA degradosome (for reviews, see Refs. 20 and 21). The 3Ј-exonuclease polynucleotide phosphorylase, the RhlB helicase, and the glycolytic enzyme enolase are the other major components of this macromolecular complex (22,23).
In general, it appears that endonucleolytic hydrolysis by RNaseE initiates the decay of transcripts in E. coli (for reviews, see Refs. 4 -6). Moreover, it has been shown in vitro that RNaseE preferentially cleaves RNAs that have a 5Ј-monophosphate group (24). As primary transcripts in E. coli have a triphosphate group at their 5Ј end and the RNaseE-catalyzed hydrolysis generates downstream intermediates that have a 5Ј-monophosphate group (9), the ability of RNaseE to assess the phosphorylation status of 5Ј ends may ensure that after this enzyme has made the first cut in a primary transcript, it will tend to cut any sites downstream before starting on another intact transcript. The N-terminal half of RNaseE is sufficient to discriminate substrates that have a monophosphate group at the 5Ј end (25,26). There is evidence for "5Ј end dependence" in vivo and it has been suggested that this facet of RNaseE-catalyzed cleavage may contribute to the finding that RNA decay occurs in general without the production of significant levels of intermediates in E. coli (27).
Polypeptides that have sequences similar to E. coli RNaseE have been predicted from approximately half of the eubacterial genomes that have been sequenced and in the plastids of plants (for review, see Ref. 28). In all of the cases that we know of, the sequence similarity covers at least part of the N-terminal half of E. coli RNaseE. Only some members of the RNaseE family are elaborated with N-and C-terminal extensions, and these vary greatly in length and have little sequence conservation (28). In the case of E. coli, it is known that the C-terminal domain is required for the normal rapid decay of mRNA in vivo (12)(13)(14) and it has been suggested that extensions serve to adapt the activity of the catalytic domains of RNaseE-related enzymes to mediate patterns of RNA processing and decay that are appropriate for the corresponding bacteria or plastid (14).
Oligonucleotide substrates synthesized using phosphorami-dite chemistry are being used increasingly to study hydrolysis of RNA by RNaseE in vitro. In the first study of this enzyme that used oligonucleotide substrates, it was found that a decaribonucleotide (ACAGU2AUUUG) corresponding in sequence to the single-stranded region at the 5Ј end of RNAI, the antisense regulator of ColE1-type plasmid replication, is cut at least as efficiently as the corresponding full-length transcript by RNaseE (29). It was subsequently found for the 9 S precursor of 5 S rRNA that the "a" site (ACAGA2AUUUG) but not the "b" site (AUCAA2AUAAA) is cut efficiently (17). The b site, in contrast to the other two, lacks a G nucleotide upstream within 1-2 nucleotides of the phosphodiester bond, which within the context of 9 S RNA is hydrolyzed. Moreover, the swapping of a G for the A at position 4 in the b site oligonucleotide was sufficient to confer efficient hydrolysis, whereas the swapping of a U for the G at position 4 in the a site blocked hydrolysis of the oligonucleotide substrate (17). The alignment of known RNaseE sites (e.g. Ref. 30) has revealed that many have a G nucleotide that is 5Ј and is in close proximity to the susceptible bond. In light of the swapping experiments described above, this suggests that such positioned G nucleotides are of general importance as determinants of RNA hydrolysis in E. coli. The mechanism by which the b site within the context of a 9 S transcript is cleaved by RNaseE despite the absence of an upstream G remains to be determined. It is possible that the recruitment of RNaseE is regulated by some secondary structural element(s) (17) as suggested relatively early in the study of RNaseE (1,30) or the distance between susceptible sequences and a monophosphate at the 5Ј end (25). From the above, it can be seen that an understanding of the recognition and hydrolysis of RNA by RNaseE will require a systematic analysis of the contribution of RNA sequence and structure, including the phosphorylation status at the 5Ј end, the multiple domains within the C-terminal half of RNaseE, and the effects of other degradosome component. With this in mind, we have developed an assay system described here that can have relatively high throughput while being sensitive and quantitative. We also describe the optimization of reaction conditions for the N-terminal catalytic half of RNaseE and the determination of Michaelis-Menten parameters for this enzyme. In addition, we have identified functional groups in the G determinant that are required for efficient hydrolysis and the importance of substrates being ribose. This work establishes a foundation for investigating the recognition and hydrolysis of RNA by the RNaseE family of enzymes.

EXPERIMENTAL PROCEDURES
Synthesis of Oligonucleotide Substrates-The oligoribonucleotides used in this study are shown in Table I. They were synthesized using ABI reagents on an ABI 391 DNA synthesizer with a standard 1-mol DNA assembly cycle with the exception that the coupling time was extended to 15 min with a 1-s acetonitrile delivery at 7.5 min. The standard phosphoramidites were purchased from ChemGenes, Inc. The exocyclic amino groups of adenine and cytosine were benzoylprotected, whereas guanine was isobutyryl-protected. Inosine- Deprotected oligonucleotides were added to a DNAPac PA-100 (4 ϫ 250 mm) anion-exchange column (Dionex) that had been equilibrated with 50 mM NaCl, 20% (v/v) acetonitrile and were eluted with a gradient of 10 -70% 1 M NaCl containing 20% (v/v) acetonitrile over 20 min. The peak fractions were further purified using a Jupiter C 18 reverse-phase (5 m, 300 Å, 4 ϫ 250 mm) column (Phenomenex) equilibrated with 100 mM NH 4 OAc and eluted with a gradient of 0 -35% 100 mM NH 4 OAc/acetonitrile (1:1) using a DX500 high pressure liquid chromatography system (Dionex). Q-Tof (quadrupole time-of-flight) mass spectrometry analysis in conjunction with maximum entropy analysis of the m/z spectrum revealed that the molecular mass of each of the oligoribonucleotides was within 0.7 Da of those calculated.
Production and Purification of the N-terminal Half of RNaseE-This was done using the pET expression and purification system (31)(32)(33). A DNA fragment encoding the first 529 amino acid residues of RNaseE with a decahistidine tag at the C terminus was generated using a pair of primers with the following sequences: 5Ј-CATGCCATGGAAAGAAT-GTTAATCAACG-3Ј (start codon is in italics, and NcoI site is underlined) and 5Ј-CGGGATCCTTAATGATGATGATGATG-ATGATGAT-GATGATGCAGCGCAGGTTGTTCCG-3Ј (anti-stop codon is in boldface, and BamHI site is underlined). The PCR product was cut with NcoI and BamHI and inserted between the corresponding sites in the expression vector pET16b (Novagen). The resulting construct, called pRne529c, was then sequenced from both strands. BL21(DE3) cells containing this plasmid were grown at 37°C in 750 ml of LB medium containing 50 g/ml carbenicillin. When the optical density of the culture at 600 nm was 0.5-0.6, cells were induced by the addition of isopropyl-␤-D-thiogalactoside to 1 mM and harvested by centrifugation (5000 ϫ g for 15 min at 4°C) after an additional 2 h of incubation. The cell pellet was resuspended in lysis buffer containing 20 mM Tris-HCl, pH 7.6, 50 mM imidazole, 0.5 M NaCl, and 400 l of a mixture of protease inhibitors (Sigma). The cells were lysed by mechanical disruption with a French pressure cell press (SLM Instruments, Inc.) at 25,000 p.s.i., and the lysate was cleared by centrifugation (100,000 ϫ g for 30 min at 4°C) following by filtration through a 0.45-m Filtopur S TM filter (Sarstedt). The soluble extract was applied to a 5-ml nickel-chelating column (Amersham Biosciences) that had been charged with Ni 2ϩ and equilibrated with lysis buffer as per the vendor's instructions. The column was washed with five volumes of lysis buffer, and then the tagged protein was eluted with an 80-ml gradient of imidazole (0 -1000 mM) in lysis buffer. The elution of protein peaked at 400 mM imidazole. The protein in the peak was concentrated using MacroSep ultrafiltration device with a 10-kDa cut-off membrane as described by the vendor (Flowgen). The concentrate was applied to a Sephacryl S-200 HR column (1.6 ϫ 60 cm, Amersham Biosciences) that had been equilibrated with 20 mM Tris-HCl, pH 7.6, and 0.5 M NaCl and protein eluted with the same buffer. Fractions corresponding to the first of two peaks that eluted after the void volume were pooled, and the protein was concentrated as described above. The protein in the first peak corresponds to a multimeric predominantly tetrameric form of RNaseE. 2 The final protein concentration was determined by measuring the absorbance of samples at 280 nm. The extinction coefficient used for NTH-RNaseE tagged at the C terminus was 26,920 M Ϫ1 cm Ϫ1 as calculated using the ExPASy Molecular Biology Server tool (ca. expasy.org/). The purity of the preparation was greater than 95% as judged by SDS-PAGE analysis. 2 The Dependence of the Hydrolysis of RNA by RNaseE on pH-The effects of pH on RNaseE hydrolysis was studied using a substrate concentration of 6 M and an enzyme concentration of 1 nM. Under these conditions, the initial reaction rate approaches its maximum. The reactions contained 10 mM MgCl 2 , 100 mM NaCl, 0.1% (v/v) Triton X-100, 1 mM dithiothreitol, and 25 mM of a buffering agent. The following pH values were analyzed: 5.5, 6.0, 6.5 (MES-HCl), 7.0, 7.3, 7.6, 8.0 (Tris-HCl), 8.5, 9.0, 9.5 (Bis-Tris-propane-HCl), 10.0, and 10.5 (CAPS-HCl). The reactions were initiated by the addition of an enzyme that had been diluted in the reaction buffer. Reaction mixtures were incubated at 37°C, and aliquots were taken at appropriate time points. Samples were quenched by mixing aliquots with an equal volume of 20 mM EDTA and analyzed by ion-pair reverse phase chromatography using a Dionex high pressure liquid chromatography system with an in-line fluorescence detector (Model RF-2000) and automated sample injector (Model ASI-100). Chromatograms were analyzed with the use of Chromeleon SP1 software, version 6.4. A Transgenomic DNA Sep® cartridge was equilibrated with 70:30 mixture of 2.5 mM tetrabutyl ammonium bromide, 1 mM EDTA, and 0.1% (v/v) acetonitrile (buffer A) and 2.5 mM tetrabutyl ammonium bromide, 1 mM EDTA, and 70% (v/v) acetonitrile (buffer B). RNA oligonucleotides were eluted with a gradient of 30 -60% buffer B over 10 min. The system was calibrated using a 5-mer, p-BR5-Fl, which corresponded to the 3Ј product of RNaseE hydrolysis but was synthesized chemically. Initial rates of reaction at the various substrate concentrations were determined by plotting the amount of product against time. Data from triplicate experiments were averaged and fitted to a double ionization equation (35) using Kaleida-Graph software (Synergy Software, Reading, PA) as shown in Equation 1, where K E1 and K E2 are acid dissociation constants of the free enzyme, k obs is the rate constant at a given pH, and k 0 is the maximal rate constant of the catalytically competent form of the enzyme.
where v is the initial rate, [S] is the substrate concentration, and [E] is the total enzyme concentration.

Assay of RNaseE under Pseudo-first Order Conditions-
The reaction conditions and assay procedures were as described above. The final substrate and enzyme concentrations were 10 and 100 nM, respectively. The experiments were done at least three times, and average data were plotted and fitted to the pseudo-first order equation as shown in Equation 3, where P is percentage of the product, S 0 is the initial percentage of substrate (100%), and [E] is the total enzyme concentration.

RESULTS
The Assay of RNaseE by High Pressure Liquid Chromatography Analysis of Fluorescein-labeled RNA-As an alternative to incorporating 32 P at the 5Ј end of oligoribonucleotides postsynthesis, we added fluorescein (Fl) to their 3Ј end during chemical synthesis (for details of the RNAs used in this study, see Fig. 1). When required, a monophosphate (p) was added chemically to the 5Ј end during synthesis to avoid the significant variation in efficiency that can occur when T4 polynucleotide kinase is used. In contrast to 32 P-labeled substrates, we find that those with fluorescein can be kept for many months without detectable loss of label. Using a fluorescence imaging system, we have been able to detect bands in polyacrylamide gels that correspond to Ͻ10 femtomoles of fluorescein-labeled RNA. 4 This is comparable to the sensitivity of detecting 32 Plabeled RNA by phosphorimaging (36). We have now automated the resolution and detection of RNA in reaction mixtures using ion-pair reverse phase chromatography coupled with an in-line fluorescence detector (for details, see "Experimental Procedures"). As shown in Fig. 2, panel A, a 5-nucleotide 3Ј-(Fl)-labeled fragment produced by hydrolysis of p-BR10-Fl, a 10-nucleotide substrate corresponding in sequence to the RNaseE site at the 5Ј end of RNAI from pBR322 (29), can be separated readily. Moreover, this system can separate oligonucleotide substrates that differ only in the presence or absence of a monophosphate at the 5Ј end (data not shown). Quantitation was calibrated using p-BR5-Fl, a synthetic mimic of the 5-nucleotide hydrolysis product. The fluorescence response was linear over our working range of 40 fmol-10 pmol (Fig. 2, panel B). Furthermore, the amount of fluorescence from p-BR5-Fl was the same as that from equal amounts of p-BR10-Fl or a derivative p-BR13-Fl, which has three extra Gs at the very 5Ј end (data not shown).
Optimization of Conditions for RNaseE Hydrolysis of Oligonucleotide Substrates-The reaction conditions for the NTH of RNaseE were optimized using a stepwise approach. First, the pH of the buffer was optimized and then the magnesium concentration and finally the concentration of NaCl (Fig. 3) 1. RNaseE and oligoribonucleotide substrates. A, schematic representation of the full-length enzyme. The catalytic domain resides within the N-terminal half, which is the most evolutionarily conserved. The C-terminal half contains arginine-rich regions that serve as an RNA-binding domain and sites that serve as a scaffold for the assembly of the RNA degradosome. The positions of regions rich in prolines are also shown along with a region rich in proline and acidic residues. The latter moderates RNaseE activity by a mechanism that has yet to be established (15). B, the extent of the RNaseE polypeptide used in this study. It has an oligohistidine tag at its C terminus. C, sequence and secondary structure of RNAI from plasmid pBR322. The structure is drawn as reported previously (34,56). The positions of phosphate (p) and fluorescein (Fl) groups on the oligonucleotide substrates are shown. The major site of RNaseE hydrolysis is indicated by scissors. The sequences of the BR10 and BR13 substrates are as described previously (29). was p-BR13-Fl. Preliminary experiments that used published conditions (29) to compare the hydrolysis of BR13 (radiolabeled at 5Ј end) with and without fluorescein at the 3Ј end found that the rates of the reactions were comparable, suggesting that this label at the 3Ј end does not have a significant effect on hydrolysis by the NTH of RNaseE (data not shown). The plot of the reaction rate against pH was bell-shaped (panel A), indicating that catalysis requires both unprotonated and protonated functional groups. By fitting the pH rate profile to a double ionization equation (see Equation 1) using non-linear regression analysis, the values of pK a of these groups were estimated to be 8.16 and 8.62. The optimum pH was found to be 8.3.
The optimum concentration of Mg 2ϩ in a buffer containing 100 mM NaCl was found to be between 15 and 25 mM (panel B). Although the highest "average" rate value was obtained for 20 mM Mg 2ϩ , there is no statistically significant difference in the values between 15 and 25 mM Mg 2ϩ as indicated by the overlap in the error bars. The effect of salt on the rate of RNA hydrolysis by RNaseE was also investigated in a buffer containing 15 mM Mg 2ϩ , and the optimum concentration was found to be between 100 and 200 mM (panel C). Monovalent ions such as Na ϩ tend to affect the rate of enzyme-catalyzed reactions by serving as bulk electrolytes that interact with charged surface residues rather than serving specific structural or catalytic roles, whereas divalent ions can serve both types of role. The role of Mg 2ϩ divalent metals ions in the reaction catalyzed by RNaseE is being investigated further and will be reported elsewhere.
Measurement of Michaelis-Menten Parameters-We next de-

FIG. 2. Separation and quantitation of the products of hydrolysis of BR10 by RNaseE.
A, elution profiles of hydrolysis products of BR10. The RNaseE hydrolysis reaction mixture contained 400 nM substrate and 1 nM enzyme and was incubated for 30 min as described under "Experimental Procedures." The 1-nucleotide ladder was generated by incubating 1 g of substrate and 1 unit of S1 nuclease (Invitrogen) at 37°C for 1 min. The positions of the p-BR10-Fl substrate and the p-BR5-Fl hydrolysis product are indicated by single and double asterisks, respectively. Numbers above each peak indicate the size of the 3Ј-labeled products. B, calibration of the fluorescence response. Different amounts of p-BR5-Fl from 40 fmol to 10 pmol were loaded onto the column, and the area under the corresponding peaks of fluorescence was calculated. Each value represents the average of three independent measurements. The bars represent the mean Ϯ S.D. of each value. The inset shows the values for peak areas corresponding to amounts lower than 1 pmol of p-BR5-Fl. . We also determined Michaelis-Menten parameters for the hydrolysis of p-BR10-Fl (panel B), which has only 5 nucleotides between the susceptible bond and the 5Ј-monophosphate, whereas p-BR13-Fl has 8 nucleotides. The objective of this experiment was to further investigate an observation made using an oligoadenosine substrate that phosphodiester bonds within a distance of 8 nucleotides from a monophosphate group are less susceptible to hydrolysis (36). The K M and k cat values obtained for p-BR10-Fl were 9.70 M and 1.1 s Ϫ1 , respectively. To quantitate the relative susceptibility of p-BR10-Fl and p-BR13-Fl to hydrolysis by RNaseE, we compared their reaction rates, which were derived from their k cat /K M ratios. Consistent with the oligoadenosine result (36), p-BR10-Fl is less susceptible than p-BR13-Fl to hydrolysis by RNaseE (see Table I).
Identification  (Table I) were obtained by fitting the amount of product formed against time to the pseudofirst order equation (see Equation 3). As reported previously for the two major sites of RNaseE hydrolysis in 9 S RNA (17), we found that the G one-nucleotide upstream of the susceptible phosphodiester bond in the pBR322 site is an important sequence determinant. The replacement of the G at position 4 with an A reduced the reaction rate by 12.0-fold. Compared with guanosine (Fig. 5), adenosine has an amino group in place of a keto group at the 6-position and lacks a proton at the N1-position and an amino group at the 2-position. To determine specifically whether the 2-amino group is important for hydrolysis by RNaseE, we next incorporated inosine at nucleotide 4. The reaction rate was reduced by only 4.7-fold, suggesting that a group(s) in addition to the 2-amino group has a role. This was confirmed by assaying the hydrolysis of an oligonucleotide substituted at nucleotide 4 with 2-aminopurine, which lacks the N1-proton and 6-keto group. The rate of hydrolysis of this substrate was reduced by 10.3-fold, indicating that in a guanosine upstream of a susceptible bond, the N1-proton or 6-keto group or both along with the 2-amino group contributes to the hydrolysis of RNA by RNaseE. We also found that the reaction rate of an oligonucleotide substituted with 7-deaza-G, which has a carbon in place of a nitrogen at the 7-position in the nucleobase, was reduced by 6.0-fold. The requirement for a particular group on the imidazole ring of guanosine provides an explanation for the finding that oligonucleotides substituted with uracil, which lacks an imidazole ring, in place of guanosine are cleaved by RNaseE at a reduced rate (17). The lack of a 2-amino group on uracil may also contribute to the greatly reduced rates of hydrolysis that have been observed. The Importance of Being Ribose-Hydrolysis of p-BR10-Fl substituted with 2Ј-deoxyuridine at nucleotide position 5 was not detected between positions 5 and 6, indicating that the 2Ј-OH in nucleotide 5 is required for hydrolysis by RNaseE. One possible role for the 2Ј-hydroxyl group in the upstream nucleotide was that it participated in the hydrolysis reaction by forming a hydrogen bond to the 3Ј-leaving group oxygen. To investigate this possibility, we replaced the hydroxyl at the 2Ј position in nucleotide 5 with fluorine, which is unable to share a proton. This caused only a 2.6-fold reduction in the reaction  3.49 ϫ 10 7 BR10 G4(A) 1.25 ϫ 10 6 BR10 G4(inosine) 3.18 ϫ 10 6 BR10 G4(2-amino purine) 1.46 ϫ 10 6 BR10 G4(7-deaza G) 2.51 ϫ 10 6 BR10 U5(dU) ND* BR10 U5(2Ј-OMe U) ND** BR10 U5(2Ј-NH 2 U) 2.97 ϫ 10 5 BR10 U5(2Ј-F U) 5.71 ϫ 10 6 rate ( Table I), indicating that proton sharing by the 2Ј-OH group is not critical for the RNaseE-catalyzed reaction, although it may make a modest contribution. As the sugars of 2Ј-fluorine-substituted nucleosides and ribonucleosides have a preference for the C3Ј-endo conformation whereas 2Ј-deoxynucleosides are predominantly in the C2Ј-endo sugar conformation (40), we next investigated whether the pucker of the sugar in the upstream nucleotide is important (Table I). This was done by synthesizing an oligonucleotide substituted at nucleotide position 5 with 2Ј-amino-2Ј-deoxyuridine, which has a preference for the C2Ј-endo sugar conformation (41) and is able to share a proton at the 2Ј position. Hydrolysis of this oligonucleotide was barely detected. The reaction rate was reduced around 50-fold. Combined, the above results indicated that the efficient hydrolysis of RNA by RNaseE is dependent on the pucker of the ribose sugar. Hydrolysis of an oligonucleotide substituted at nucleotide 5 with 2Ј-O-methyl, which blocks proton sharing while maintaining a ribose-like sugar pucker, was not detected. In light of the result with the deoxy-2Ј-fluorouridine-substituted substrate, this finding suggests that RNaseE hydrolysis is hampered by a bulky group at this particular position. Although 2Ј-O-methylcontaining oligonucleotides have been shown previously not to be cleaved by RNaseE, the cause could not be linked to a specific position because more than one substitution had been incorporated (42). DISCUSSION A number of factors have been shown to influence hydrolysis of RNA by E. coli RNaseE such as the sequence of hydrolysis sites, the secondary structure and 5Ј phosphorylation status of the substrate, and domains within the C-terminal half of the enzyme. However, there is as yet no detailed understanding of the relative influence of these factors and how they combine mechanistically to regulate the recognition and hydrolysis of RNA by RNaseE. Although numerous RNaseE polypeptides and RNA substrates have been constructed that differ in do-main composition and structure, respectively, many if not most comparisons of RNA hydrolysis in vitro have been based on limited kinetic data under conditions that have not been rigorously defined. Although this has been sufficient to identify factors that influence hydrolysis, it has not been possible to thoroughly test the correctness of models that have been proposed. A significant barrier to the generation of sufficient kinetic data to do this, at least in our case, was the labor and time required to prepare and purify radiolabeled substrates, run polyacrylamide gels, and measure band intensities from autoradiographs or phosphorimages. The new assay approach described here provides much higher throughput while being sensitive and quantitative.
The reactions were done as described previously (29) with samples being withdrawn at different time points and quenched. However, ion-pair reverse phase chromatography was used in place of polyacrylamide gel electrophoresis to separate substrate and products and an automated injection system was used to load samples. The resolution provided by ion-pair reverse phase chromatography is more than comparable with that of polyacrylamide gel electrophoresis (Fig. 2) and has also been used, for example, for footprinting studies (43). The quantity of each RNA species as it was eluted was determined by an in-line instrument that measured the amount of fluorescence from fluorescein that was attached to substrates during their chemical synthesis. The fluorescent label was placed at the 3Ј end to allow RNaseE to sense the phosphorylation status at the 5Ј end of substrates. This positioning also blocks 3Ј-exonucleolytic attack, and we have been able for the first time to use small defined substrates to assay RNaseE activity within the degradosome and crude cell extracts. 5 The first polypeptide of RNaseE we have chosen to study is the N-terminal catalytic half of RNaseE. In doing so, we aim to provide a platform for investigating in future studies the role of ancillary domains within the C-terminal half of the E. coli enzyme and a reference for the many RNaseE homologues in eubacteria and plant plastids that only have sequence similarity to the NTH of E. coli RNaseE. As in previous studies (10,17,29,36), we expressed NTH-RNaseE using the pET system (31)(32)(33), which generates an oligohistidine tag that facilitates purification by immobilized metal affinity chromatography (44,45). In this study, however, we used a polypeptide that was tagged at the C instead of the N terminus as extension of the latter may reduce the activity of RNaseE. In experiments not described here, we found that the amount of RNaseE produced by the pET system is inversely related to the activity of the cloned RNaseE gene and that the levels of NTH-RNaseE produced were higher when the oligohistidine tag was positioned at the N terminus. Consistent with our notion that the activity of RNaseE is affected by extension of the N terminus, there is evidence that the ability of E. coli RNaseG, an orthologue of RNaseE, when overproduced to compensate for the loss of RNaseE is affected by extension of its N terminus (compare Refs. 46 and 47). However, altering the position of the tag does not appear to affect the specificity of the enzyme (compare Fig.  2 with Ref. 29).
The assay was used first to optimize the reaction conditions for the N-terminal domain of RNaseE (Fig. 3). The substrate for these experiments was BR13, which has three guanosines followed by the sequence of the RNaseE site at the 5Ј end of pBR322 RNAI (Fig. 1). Previous studies have shown that this substrate is sufficient to direct hydrolysis by purified RNaseE of the phosphodiester bond that appears in the context of RNAI to be the most susceptible to hydrolysis in vivo. A substrate 5 L. Tong and K. J. McDowall, unpublished data.

FIG. 5. The bases introduced at nucleotide position 4 in BR10.
The structure of uracil is provided for the purpose of comparison. The effect of introducing uracil in place of guanine has been described previously (17). concentration of 6 M was used. As a K M value of 0.7 M was obtained in initial experiments (data not shown) using the starting conditions (36), the concentration of substrate was sufficiently higher than the K M at the start of optimization for the reaction rates to approach their maximum. We found that catalysis is dependent on unprotonated and protonated functional groups (panel A). The values of pK a were estimated to be 8. 16 and 8.62. The environment of an amino acid side chain within a protein has been shown to affect the value of pK a by as much as 3 pH units (e.g. Refs. 35 and 48). By taking this into consideration, comparisons with the values of pK a calculated for free amino acids (49) have been used to implicate particular side chains in the reactions of certain enzymes (50,51). In the case of RNaseE, the ionizable residues required for catalysis could correspond to lysine, tyrosine, cysteine, or histidine residues, which in the free form have pK a values of 10.4, 9.6, 8.3, and 6.3, respectively. There are a number of lysine, tyrosine, cysteine, and histidine residues within the NTH of E. coli RNaseE. Of these, three lysines at positions 103, 109, and 149 and three tyrosines at positions 42, 123, and 256 are highly conserved across the RNaseE family 6 and located within the smallest RNaseE polypeptide that has been found to have hydrolytic activity (52). Hydrated metal ions such as Mg(H 2 0) 6 2ϩ could also be one of the ionizable groups that are required for the hydrolysis of RNA by RNaseE. Irrespective of the actual identity of the proton donor and acceptor, their detection is suggestive of general acid-general base catalysis, which is used for example by the RNaseH1 family of ribonucleases (53).
By increasing the pH from 7.6 to 8.3 and the MgCl 2 concentration from 10 to 15 mM (Fig. 3, panels A and B), we estimate that we have been able to increase the rate of RNA hydrolysis by at least 4-fold. This means that previous biochemical studies that investigated the relative roles of ribonucleases in various steps in RNA decay and processing, e.g. poly(A) removal (36), may have underestimated the capabilities of RNaseE. After optimizing conditions for the hydrolysis of RNA by the Nterminal half of RNaseE, we determined Michaelis-Menten parameters for the hydrolysis of p-BR13-Fl and p-BR10-Fl (Fig.  4). The latter lacks the three extra Gs at the 5Ј end of BR13. The values of k cat we determined (1.4 and 1.1 s Ϫ1 , respectively) are similar to those of other phosphotransferases, RNaseHI (1.5 s Ϫ1 , see Ref. 39) and EcoRV (0.70 s Ϫ1 , see Refs. 37 and 38), and may reflect similarities in the chemistries used to cleave phosphodiester bonds following the formation of an enzymesubstrate complex. The K M values calculated for p-BR13-Fl and p-BR10-Fl using the optimized conditions were 2.07 and 9.70 M, respectively. Interestingly, the K M value calculated for p-BR13-Fl using the optimized conditions appears to be 2-3fold higher than that calculated using the previous reaction conditions (36). This possible effect on K M is being investigated further by studying the effect of pH and magnesium on the hydrolysis of RNA under pseudo-first order conditions as well as steady-state conditions. The higher K M value for p-BR10-Fl is consistent with the finding that phosphodiester bonds Ͻ8 nucleotides from a monophosphate group at the 5Ј end are less susceptible to hydrolysis (36). By now determining kinetic parameters for reactions between RNaseE polypeptides that contain additional domains and substrates that are more complex in structure, it will be possible to address the correctness of models that have been proposed for RNaseE, such as a role for the arginine-rich RNA domain in the recruitment of structured RNaseE to the catalytic site (17) and the effect of distance from the 5Ј end (36).
While we were undertaking our kinetic analysis of the N-terminal half of RNaseE, collaborators were successful in obtaining crystals of this part of RNaseE. 2 We therefore decided to generate biochemical data that in hand with a 3-dimensional structure should lead to the identification of enzyme-substrate interactions that are candidates for having important roles in the hydrolysis of RNA. Our approach was to modify the RNA because, unlike the amino acid residues involved in catalysis, the targets of catalysis, i.e. susceptible bonds, had been identified. Previous studies had shown that a guanosine positioned within 1-2 nucleotides upstream of susceptible bonds is an important determinant of hydrolysis by E. coli RNaseE. Therefore, we synthesized a number of oligonucleotide substrates that had functional groups substituted in the guanosine at nucleotide 4 in p-BR10-Fl. We found under pseudo-first order conditions that at least three functional groups in this sequence determinant are required for efficient RNaseE hydrolysis of RNA (Table I). The lack of the 2-amino group along with the lack of the N1 proton or 6-keto group is sufficient to explain the reduced rate of hydrolysis of oligonucleotides containing G to A substitutions, whereas the lack of the imidazole ring along with the lack of the 2-amino group can explain the reduced rates resulting from G to U substitutions (17). The functional groups in additional sequence determinants that have recently been shown to affect the efficiency of cleavage by E. coli RNaseE (54) can be characterized using the approach described here. It is now clearly established experimentally that RNaseE is sensitive to the presence of specific nucleotides at multiple positions relative to the site of cleavage even though it can hydrolyze substrates with low sequence complexity such as oligomeric adenosine and uracil (36,55). In addition to sequence, our analysis of substitutions of the 2Ј-OH on the ribose sugar directly 5Ј to the phosphodiester bond that is normally cleaved by RNaseE suggests that a ribose-like sugar pucker is important for efficient hydrolysis.