Mutational analysis of the nuclease domain of Escherichia coli ribonuclease III. Identification of conserved acidic residues that are important for catalytic function in vitro

The ribonuclease III superfamily represents a structurally distinct group of double-strand-specific endonucleases with essential roles in RNA maturation, RNA decay, and gene silencing. Bacterial RNase III orthologs exhibit the simplest structures, with an N-terminal nuclease domain and a C-terminal double-stranded RNA-binding domain (dsRBD), and are active as homodimers. The nuclease domain contains conserved acidic amino acids, which in Escherichia coli RNase III are E38, E41, D45, E65, E100, D114, and E117. On the basis of a previously reported crystal structure of the nuclease domain of Aquifex aeolicus RNase III, the E41, D114, and E117 side chains of E. coli RNase III are expected to be coordinated to a divalent metal ion (Mg(2+) or Mn(2+)). It is shown here that the RNase III[E41A] and RNase III[D114A] mutants exhibit catalytic activities in vitro in 10 mM Mg(2+) buffer that are comparable to that of the wild-type enzyme. However, at 1 mM Mg(2+), the activities are significantly lower, which suggests a weakened affinity for metal. While RNase III[E41A] and RNase III[D114A] have K(Mg) values that are approximately 2.8-fold larger than the K(Mg) of RNase III (0.46 mM), the RNase III[E41A/D114A] double mutant has a K(Mg) of 39 mM, suggesting a redundant function for the two side chains. RNase III[E38A], RNase III[E65A], and RNase III[E100A] also require higher Mg(2+) concentrations for optimal activity, with RNase III[E100A] exhibiting the largest K(Mg). RNase III[D45A], RNase III[D45E], and RNase III[D45N] exhibit negligible activities, regardless of the Mg(2+) concentration, indicating a stringent functional requirement for an aspartate side chain. RNase III[D45E] activity is partially rescued by Mn(2+). The potential functions of the conserved acidic residues are discussed in the context of the crystallographic data and proposed catalytic mechanisms.     

Mechanism of action of Escherichia coli ribonuclease III. Stringent chemical requirement for the glutamic acid 117 side chain and Mn2+ rescue of the Glu117Asp mutant

Escherichia coli ribonuclease III (EC 3.1.24) is a double-strand- (ds-) specific endoribonuclease involved in the maturation and decay of cellular, phage, and plasmid RNAs. RNase III is a homodimer and requires Mg(2+) to hydrolyze phosphodiesters. The RNase III polypeptide contains an N-terminal catalytic (nuclease) domain which exhibits eight highly conserved acidic residues, at least one of which (Glu117) is important for phosphodiester hydrolysis but not for substrate binding [Li and Nicholson (1996) EMBO J. 15, 1421-1433]. To determine the side chain requirements for activity, Glu117 was changed to glutamine or aspartic acid. The mutant proteins were purified as (His)(6)-tagged species, and both exhibited normal homodimeric behavior as shown by chemical cross-linking. The Glu117Gln mutant is unable to cleave substrate in vitro under all tested conditions but can bind substrate. The Glu117Asp mutant also is defective in cleavage while able to bind substrate. However, low level activity is observed at extended reaction times and high enzyme concentrations, with an estimated catalytic efficiency approximately 15 000-fold lower than that of RNase III. The activity of the Glu117Asp mutant but not that of the Glu117Gln mutant can be greatly enhanced by substituting Mn(2+) for Mg(2+), with the catalytic efficiency of the Glu117Asp-Mn(2+) holoenzyme approximately 400-fold lower than that of the RNase III-Mn(2+) holoenzyme. For RNase III, a Mn(2+) concentration of 1 mM provides optimal activity, while concentrations >5 mM are inhibitory. In contrast, the Glu117Asp mutant is not inhibited by high concentrations of Mn(2+). Finally, high concentrations of Mg(2+) do not inhibit RNase III nor relieve the Mn(2+)-dependent inhibition. In summary, these experiments establish the stringent functional requirement for a precisely positioned carboxylic acid group at position 117 and reveal two classes of divalent metal ion binding sites on RNase III. One site binds either Mg(2+) or Mn(2+) and supports catalysis, while the other site is specific for Mn(2+) and confers inhibition. Glu117 is important for the function of both sites. The implications of these findings on the RNase III catalytic mechanism are discussed.     

Crystallographic and modeling studies of RNase III suggest a mechanism for double-stranded RNA cleavage.

BACKGROUND: Aquifex aeolicus Ribonuclease III (Aa-RNase III) belongs to the family of Mg(2+)-dependent endonucleases that show specificity for double-stranded RNA (dsRNA). RNase III is conserved in all known bacteria and eukaryotes and has 1-2 copies of a 9-residue consensus sequence, known as the RNase III signature motif. The bacterial RNase III proteins are the simplest, consisting of two domains: an N-terminal endonuclease domain, followed by a double-stranded RNA binding domain (dsRBD). The three-dimensional structure of the dsRBD in Escherichia coli RNase III has been elucidated; no structural information is available for the endonuclease domain of any RNase III. RESULTS: We present the crystal structures of the Aa-RNase III endonuclease domain in its ligand-free form and in complex with Mn(2+). The structures reveal a novel protein fold and suggest a mechanism for dsRNA cleavage. On the basis of structural, genetic, and biological data, we have constructed a hypothetical model of Aa-RNase III in complex with dsRNA and Mg(2+) ion, which provides the first glimpse of RNase III in action. CONCLUSIONS: The functional Aa-RNase III dimer is formed via mainly hydrophobic interactions, including a "ball-and-socket" junction that ensures accurate alignment of the two monomers. The fold of the polypeptide chain and its dimerization create a valley with two compound active centers at each end of the valley. The valley can accommodate a dsRNA substrate. Mn(2+) binding has significant impact on crystal packing, intermolecular interactions, thermal stability, and the formation of two RNA-cutting sites within each compound active center.     

Heterodimer-based analysis of subunit and domain contributions to double-stranded RNA processing by Escherichia coli RNase III in vitro

Members of the RNase III family are the primary cellular agents of dsRNA (double-stranded RNA) processing. Bacterial RNases III function as homodimers and contain two dsRBDs (dsRNA-binding domains) and two catalytic sites. The potential for functional cross-talk between the catalytic sites and the requirement for both dsRBDs for processing activity are not known. It is shown that an Escherichia coli RNase III heterodimer that contains a single functional wt (wild-type) catalytic site and an inactive catalytic site (RNase III[E117A/wt]) cleaves a substrate with a single scissile bond with a k(cat) value that is one-half that of wt RNase III, but exhibits an unaltered K(m). Moreover, RNase III[E117A/wt] cleavage of a substrate containing two scissile bonds generates singly cleaved intermediates that are only slowly cleaved at the remaining phosphodiester linkage, and in a manner that is sensitive to excess unlabelled substrate. These results demonstrate the equal probability, during a single binding event, of placement of a scissile bond in a functional or nonfunctional catalytic site of the heterodimer and reveal a requirement for substrate dissociation and rebinding for cleavage of both phosphodiester linkages by the mutant heterodimer. The rate of phosphodiester hydrolysis by RNase III[E117A/wt] has the same dependence on Mg(2+) ion concentration as that of the wt enzyme, and exhibits a Hill coefficient (h) of 2.0+/-0.1, indicating that the metal ion dependence essentially reflects a single catalytic site that employs a two-Mg(2+)-ion mechanism. Whereas an E. coli RNase III mutant that lacks both dsRBDs is inactive, a heterodimer that contains a single dsRBD exhibits significant catalytic activity. These findings support a reaction pathway involving the largely independent action of the dsRBDs and the catalytic sites in substrate recognition and cleavage respectively.     

Ribonuclease III cleavage of a bacteriophage T7 processing signal. Divalent cation specificity, and specific anion effects.

Escherichia coli ribonuclease III, purified to homogeneity from an overexpressing bacterial strain, exhibits a high catalytic efficiency and thermostable processing activity in vitro. The RNase III-catalyzed cleavage of a 47 nucleotide substrate (R1.1 RNA), based on the bacteriophage T7 R1.1 processing signal, follows substrate saturation kinetics, with a Km of 0.26 microM, and kcat of 7.7 min.-1 (37 degrees C, in buffer containing 250 mM potassium glutamate and 10 mM MgCl2). Mn2+ and Co2+ can support the enzymatic cleavage of the R1.1 RNA canonical site, and both metal ions exhibit concentration dependences similar to that of Mg2+. Mn2+ and Co2+ in addition promote enzymatic cleavage of a secondary site in R1.1 RNA, which is proposed to result from the altered hydrolytic activity of the metalloenzyme (RNase III 'star' activity), exhibiting a broadened cleavage specificity. Neither Ca2+ nor Zn2+ support RNase III processing, and Zn2+ moreover inhibits the Mg(2+)-dependent enzymatic reaction without blocking substrate binding. RNase III does not require monovalent salt for processing activity; however, the in vitro reactivity pattern is influenced by the monovalent salt concentration, as well as type of anion. First, R1.1 RNA secondary site cleavage increases as the salt concentration is lowered, perhaps reflecting enhanced enzyme binding to substrate. Second, the substitution of glutamate anion for chloride anion extends the salt concentration range within which efficient processing occurs. Third, fluoride anion inhibits RNase III-catalyzed cleavage, by a mechanism which does not involve inhibition of substrate binding.     

Ethidium-dependent uncoupling of substrate binding and cleavage by Escherichia coli ribonuclease III

Ethidium bromide (EB) is known to inhibit cleavage of bacterial rRNA precursors by Escherichia coli ribonuclease III, a dsRNA-specific nuclease. The mechanism of EB inhibition of RNase III is not known nor is there information on EB-binding sites in RNase III substrates. We show here that EB is a reversible, apparently competitive inhibitor of RNase III cleavage of small model substrates in vitro. Inhibition is due to intercalation, since (i) the inhibitory concentrations of EB are similar to measured EB intercalation affinities; (ii) substrate cleavage is not affected by actinomycin D, an intercalating agent that does not bind dsRNA; (iii) the EB concentration dependence of inhibition is a function of substrate structure. In contrast, EB does not strongly inhibit the ability of RNase III to bind substrate. EB also does not block substrate binding by the C-terminal dsRNA-binding domain (dsRBD) of RNase III, indicating that EB perturbs substrate recognition by the N-terminal catalytic domain. Laser photocleavage experiments revealed two ethidium-binding sites in the substrate R1.1 RNA. One site is in the internal loop, adjacent to the scissile bond, while the second site is in the lower stem. Both sites consist of an A-A pair stacked on a CG pair, a motif which apparently provides a particularly favorable environment for intercalation. These results indicate an inhibitory mechanism in which EB site-specifically binds substrate, creating a cleavage-resistant complex that can compete with free substrate for RNase III. This study also shows that RNase III recognition and cleavage of substrate can be uncoupled and supports an enzymatic mechanism of dsRNA cleavage involving cooperative but not obligatorily linked actions of the dsRBD and the catalytic domain.     

Noncatalytic assembly of ribonuclease III with double-stranded RNA

Ribonuclease III (RNase III) represents a family of double-stranded RNA (dsRNA) endonucleases. The simplest bacterial enzyme contains an endonuclease domain (endoND) and a dsRNA binding domain (dsRBD). RNase III can affect RNA structure and gene expression in either of two ways: as a dsRNA-processing enzyme that cleaves dsRNA, or as a dsRNA binding protein that binds but does not cleave dsRNA. We previously determined the endoND structure of Aquifex aeolicus RNase III (Aa-RNase III) and modeled a catalytic complex of full-length Aa-RNase III with dsRNA. Here, we present the crystal structure of Aa-RNase III in complex with dsRNA, revealing a noncatalytic assembly. The major differences between the two functional forms of RNase III.dsRNA are the conformation of the protein and the orientation and location of dsRNA. The flexibility of a 7 residue linker between the endoND and dsRBD enables the transition between these two forms.     

RNase HIII from Chlamydophila pneumoniae can efficiently cleave double-stranded DNA carrying a chimeric ribonucleotide in the presence of manganese

Two ribonuclease Hs (RNase Hs) have been found in Chlamydophila pneumoniae, CpRNase HII and CpRNase HIII. This work is the first report that CpRNase HIII can efficiently cleave DNA-rN(1) -DNA/DNA (rN(1) , monoribonucleotide) in vitro in the presence of Mn(2+) , whereas the enzymatic activity of CpRNase HII on the same substrate was inhibited by Mn(2+) and dependent on Mg(2+) . However, the ability of both CpRNase Hs to cleave other alternative substrates (RNA/DNA hybrids and Okazaki-like substrates), was insensitive to the divalent ions changes, suggesting that high concentrations of Mn(2+) specifically repressed the ability of CpRNase HII to cleave DNA-rN(1) -DNA/DNA but activated this function in CpRNase HIII. Further in vivo experiments showed that the CpRNase HII complementation of Escherichia coli rnh(-) mutations in an Mg(2+) environment was suppressed by Mn(2+) . In contrast, Mn(2+) was indispensable for CpRNase HIII to complement the same mutations. Further, the cell growth inhibition and the genomic DNA sensitivity to alkali in the bacterial strain lacking RNase HII activity could be relieved by functional CpRNase HII or HIII with its compatible ion. Therefore, CpRNase HIII can execute cleavage activity on DNA-rN(1) -DNA/DNA under a Mn(2+) -rich environment and may function as a substitute for CpRNase HII under special physiological states.     

Monitoring the structure of Escherichia coli RNase P RNA in the presence of various divalent metal ions

Lead(II)-induced cleavage can be used as a tool to probe conformational changes in RNA. In this report, we have investigated the conformation of M1 RNA, the catalytic subunit of Escherichia coli RNase P, by studying the lead(II)-induced cleavage pattern in the presence of various divalent metal ions. Our data suggest that the overall conformation of M1 RNA is very similar in the presence of Mg(2+), Mn(2+), Ca(2+), Sr(2+) and Ba(2+), while it is changed compared to the Mg(2+)-induced conformation in the presence of other divalent metal ions, Cd(2+) for example. We also observed that correct folding of some M1 RNA domains is promoted by Pb(2+), while folding of other domain(s) requires the additional presence of other divalent metal ions, cobalt(III) hexamine or spermidine. Based on the suppression of Pb(2+) cleavage at increasing concentrations of various divalent metal ions, our findings suggest that different divalent metal ions bind with different affinities to M1 RNA as well as to an RNase P hairpin-loop substrate and yeast tRNA(Phe). We suggest that this approach can be used to obtain information about the relative binding strength for different divalent metal ions to RNA in general, as well as to specific RNA divalent metal ion binding sites. Of those studied in this report, Mn(2+) is generally among the strongest RNA binders.     

Characterization of Aquifex aeolicus ribonuclease III and the reactivity epitopes of its pre-ribosomal RNA substrates

Ribonuclease III cleaves double-stranded (ds) structures in bacterial RNAs and participates in diverse RNA maturation and decay pathways. Essential insight on the RNase III mechanism of dsRNA cleavage has been provided by crystallographic studies of the enzyme from the hyperthermophilic bacterium, Aquifex aeolicus. However, the biochemical properties of A. aeolicus (Aa)-RNase III and the reactivity epitopes of its substrates are not known. The catalytic activity of purified recombinant Aa-RNase III exhibits a temperature optimum of ～70-85°C, with either Mg2+ or Mn2+ supporting efficient catalysis. Small hairpins based on the stem structures associated with the Aquifex 16S and 23S rRNA precursors are cleaved at sites that are consistent with production of the immediate precursors to the mature rRNAs. Substrate reactivity is independent of the distal box sequence, but is strongly dependent on the proximal box sequence. Structural studies have shown that a conserved glutamine (Q157) in the Aa-RNase III dsRNA-binding domain (dsRBD) directly interacts with a proximal box base pair. Aa-RNase III cleavage of the pre-16S substrate is blocked by the Q157A mutation, which reflects a loss of substrate binding affinity. Thus, a highly conserved dsRBD-substrate interaction plays an important role in substrate recognition by bacterial RNase III.     

Activation and inhibition of rubber transferases by metal cofactors and pyrophosphate substrates

Metal cofactors are necessary for the activity of alkylation by prenyl transfer in enzyme-catalyzed reactions. Rubber transferase (RuT, a cis-prenyl transferase) associated with purified rubber particles from Hevea brasiliensis, Parthenium argentatum and Ficus elastica can use magnesium and manganese interchangably to achieve maximum velocity. We define the concentration of activator required for maximum velocity as [A](max). The [A](max)(Mg2+) in F. elastica (100 mM) is 10 times the [A](max)(Mg2+) for either H. brasiliensis (10 mM) or P. argentatum (8 mM). The [A](max)(Mn2+) in F. elastica (11 mM), H. brasiliensis (3.8 mM) and P. argentatum (6.8 mM) and the [A](max)(Mg2+) in H. brasiliensis (10 mM) and P. argentatum (8 mM) are similar. The differences in [A](max)(Mg2+) correlate with the actual endogenous Mg(2+) concentrations in the latex of living plants. Extremely low Mn(2+) levels in vivo indicate that Mg(2+) is the RuT cofactor in living H. brasiliensis and F. elastica trees. Kinetic analyses demonstrate that FPP-Mg(2+) and FPP-Mn(2+) are active substrates for rubber molecule initiation, although free FPP and metal cations, Mg(2+) and Mn(2+), can interact independently at the active site with the following relative dissociation constants K(d)(FPP) 相似文献   

Structural and mechanistic insight into stem‐loop RNA processing by yeast Pichia stipitis Dicer

Dicer is a member of the ribonuclease III enzyme family and processes double‐stranded RNA into small functional RNAs. The variation in the domain architecture of Dicer among different species whilst preserving its biological dicing function is intriguing. Here, we describe the structure and function of a novel catalytically active RNase III protein, a non‐canonical Dicer (PsDCR1), found in budding yeast Pichia stipitis. The structure of the catalytically active region (the catalytic RNase III domain and double‐stranded RNA‐binding domain 1 [dsRBD1]) of DCR1 showed that RNaseIII domain is structurally similar to yeast RNase III (Rnt1p) but uniquely presents dsRBD1 in a diagonal orientation, forming a catalytic core made of homodimer and large RNA‐binding surface. The second dsRNA binding domain at C‐terminus, which is absent in Rnt1, enhances the RNA cleavage activity. Although the cleavage pattern of PsDCR1 anchors an apical loop similar to Rnt1, the cleavage activity depended on the sequence motif at the lower stem, not the apical loop, of hairpin RNA. Through RNA sequencing and RNA mutations, we showed that RNA cleavage by PsDCR1 is determined by the stem‐loop structure of the RNA substrate, suggesting the possibility that stem‐loop RNA‐guided gene silencing pathway exists in budding yeast.     

Properties of cloned and expressed human RNase H1.

We have characterized cloned His-tag human RNase H1. The activity of the enzyme exhibited a bell-shaped response to divalent cations and pH. The optimum conditions for catalysis consisted of 1 mM Mg(2+) and pH 7-8. In the presence of Mg(2+), Mn(2+) was inhibitory. Human RNase H1 shares many enzymatic properties with Escherichia coli RNase H1. The human enzyme cleaves RNA in a DNA-RNA duplex resulting in products with 5'-phosphate and 3'-hydroxy termini, can cleave overhanging single strand RNA adjacent to a DNA-RNA duplex, and is unable to cleave substrates in which either the RNA or DNA strand has 2' modifications at the cleavage site. Human RNase H1 binds selectively to "A-form"-type duplexes with approximately 10-20-fold greater affinity than that observed for E. coli RNase H1. The human enzyme displays a greater initial rate of cleavage of a heteroduplex-containing RNA-phosphorothioate DNA than an RNA-DNA duplex. Unlike the E. coli enzyme, human RNase H1 displays a strong positional preference for cleavage, i.e. it cleaves between 8 and 12 nucleotides from the 5'-RNA-3'-DNA terminus of the duplex. Within the preferred cleavage site, the enzyme displays modest sequence preference with GU being a preferred dinucleotide. The enzyme is inhibited by single-strand phosphorothioate oligonucleotides and displays no evidence of processivity. The minimum RNA-DNA duplex length that supports cleavage is 6 base pairs, and the minimum RNA-DNA "gap size" that supports cleavage is 5 base pairs.     

The N-terminal domain that distinguishes yeast from bacterial RNase III contains a dimerization signal required for efficient double-stranded RNA cleavage

Yeast Rnt1 is a member of the double-stranded RNA (dsRNA)-specific RNase III family identified by conserved dsRNA binding (dsRBD) and nuclease domains. Comparative sequence analyses have revealed an additional N-terminal domain unique to the eukaryotic homologues of RNase III. The deletion of this domain from Rnt1 slowed growth and led to mild accumulation of unprocessed 25S pre-rRNA. In vitro, deletion of the N-terminal domain reduced the rate of RNA cleavage under physiological salt concentration. Size exclusion chromatography and cross-linking assays indicated that the N-terminal domain and the dsRBD self-interact to stabilize the Rnt1 homodimer. In addition, an interaction between the N-terminal domain and the dsRBD was identified by a two-hybrid assay. The results suggest that the eukaryotic N-terminal domain of Rnt1 ensures efficient dsRNA cleavage by mediating the assembly of optimum Rnt1-RNA ribonucleoprotein complex.     

Diffusely bound Mg2+ ions slightly reorient stems I and II of the hammerhead ribozyme to increase the probability of formation of the catalytic core

The hammerhead ribozyme is one of the best-studied small RNA enzymes, yet is mechanistically still poorly understood. We measured the Mg(2+) dependencies of folding and catalysis for two distinct hammerhead ribozymes, HHL and HH alpha. HHL has three long helical stems and was previously used to characterize Mg(2+)-induced folding. HH alpha has shorter stems and an A.U tandem next to the cleavage site that increases activity approximately 10-fold at 10 mM Mg(2+). We find that both ribozymes cleave with fast rates (5-10 min(-1), at pH 8 and 25 degrees C) at nonphysiologically high Mg(2+) concentrations, but with distinct Mg(2+) dissociation constants for catalysis: 90 mM for HHL and 10 mM for HH alpha. Using time-resolved fluorescence resonance energy transfer, we measured the stem I-stem II distance distribution as a function of Mg(2+) concentration, in the presence and absence of 100 mM Na(+), at 4 and 25 degrees C. Our data show two structural transitions. The larger transition (with Mg(2+) dissociation constants in the physiological range of approximately 1 mM, below the catalytic dissociation constants) brings stems I and II close together and is hindered by Na(+). The second, globally minor, rearrangement coincides with catalytic activation and is not hindered by Na(+). Additionally, the more active HH alpha exhibits a higher flexibility than HHL under all conditions. Finally, both ribozyme-product complexes have a bimodal stem I-stem II distance distribution, suggesting a fast equilibrium between distinct conformers. We propose that the role of diffusely bound Mg(2+) is to increase the probability of formation of a properly aligned catalytic core, thus compensating for the absence of naturally occurring kissing-loop interactions.     

Biochemical and genomic analysis of substrate recognition by the double-stranded RNA binding domain of yeast RNase III

Members of the RNase III family of double-stranded RNA (dsRNA) endonucleases are important enzymes of RNA metabolism in eukaryotic cells. Rnt1p is the only known member of the RNase III family of endonucleases in Saccharomyces cerevisiae. Previous studies have shown that Rnt1p cleaves dsRNA capped by a conserved AGNN tetraloop motif, which is a major determinant for Rnt1p binding and cleavage. The solution structure of the dsRNA-binding domain (dsRBD) of Rnt1p bound to a cognate RNA substrate revealed the structural basis for binding of the conserved tetraloop motif by alpha-helix 1 of the dsRBD. In this study, we have analyzed extensively the effects of mutations of helix 1 residues that contact the RNA. We show, using microarray analysis, that mutations of these amino acids induce substrate-specific processing defects in vivo. Cleavage kinetics and binding studies show that these mutations affect RNA cleavage and binding in vitro to different extents and suggest a function for some specific amino acids of the dsRBD in the catalytic positioning of the enzyme. Moreover, we show that 2'-hydroxyl groups of nucleotides of the tetraloop or adjacent base pairs predicted to interact with residues of alpha-helix 1 are important for Rnt1p cleavage in vitro. This study underscores the importance of a few amino acid contacts for positioning of a dsRBD onto its RNA target, and implicates the specific orientation of helix 1 on the RNA for proper positioning of the catalytic domain.     

Both N-terminal catalytic and C-terminal RNA binding domain contribute to substrate specificity and cleavage site selection of RNase III.

The double-stranded RNA-specific endoribonuclease III (RNase III) of bacteria consists of an N-terminal nuclease domain and a double-stranded RNA binding domain (dsRBD) at the C-terminus. Analysis of two hybrid proteins consisting of the N-terminal half of Escherichia coli RNase III fused to the dsRBD of the Rhodobacter capsulatus enzyme and vice versa reveals that both domains in combination with the particular substrate determine substrate specificity and cleavage site selection. Extension of the spacer between the two domains of the E. coli enzyme from nine to 20 amino acids did not affect cleavage site selection.