The role of non-catalytic binding subsites in the endonuclease activity of bovine pancreatic ribonuclease A.

Bovine pancreatic ribonuclease A catalyzes the depolymerization of RNA. There is much evidence that several subsites, in addition to the main catalytic site, are involved in the formation of the enzyme-substrate complex. This work analyzes the pattern of oligonucleotide formation by ribonuclease A using poly(C) as substrate. The poly(C) cleavage shows that the enzyme does not act in a random fashion but rather prefers the binding and cleavage of the longer substrate molecules and that the phosphodiester bond broken should be 6-7 residues apart from the end of the chain to be preferentially cleaved by ribonuclease A. The results demonstrate the model of the cleavage of an RNA chain based on the cooperative binding between the multisubsite binding structure of ribonuclease A and the phosphates of the polynucleotide (Parés, X., Nogués, M. V., de Llorens, R., and Cuchillo, C. M. (1991) in Essays in Biochemistry (Tipton, K. F., ed) Vol. 26, pp. 89-103, Portland Press Ltd., London). The contribution to the enzymatic process of the non-catalytic phosphate-binding subsite (p2) adjacent to the catalytic center has been analyzed in p2 chemically modified ribonuclease A or by means of site-directed mutagenesis. In both cases deletion of p2 abolishes the endonuclease activity of ribonuclease A, which is substituted by an exonuclease activity. All these results support the role of the multisubsite structure of the enzyme in the endonuclease activity and in the catalytic mechanism.

Bovine pancreatic ribonuclease A catalyzes the depolymerization of RNA. There is much evidence that several subsites, in addition to the main catalytic site, are involved in the formation of the enzyme-substrate complex. This work analyzes the pattern of oligonucleotide formation by ribonuclease A using poly(C) as substrate. The poly(C) cleavage shows that the enzyme does not act in a random fashion but rather prefers the binding and cleavage of the longer substrate molecules and that the phosphodiester bond broken should be 6 -7 residues apart from the end of the chain to be preferentially cleaved by ribonuclease A. The results demonstrate the model of the cleavage of an RNA chain based on the cooperative binding between the multisubsite binding structure of ribonuclease A and the phosphates of the polynucleotide (Paré s, X., Nogué s, M. V., de Llorens, R., and Cuchillo, C. M. (1991) in Essays in Biochemistry (Tipton, K. F., ed) Vol. 26, pp. 89 -103, Portland Press Ltd., London). The contribution to the enzymatic process of the non-catalytic phosphate-binding subsite (p 2 ) adjacent to the catalytic center has been analyzed in p 2 chemically modified ribonuclease A or by means of sitedirected mutagenesis. In both cases deletion of p 2 abolishes the endonuclease activity of ribonuclease A, which is substituted by an exonuclease activity. All these results support the role of the multisubsite structure of the enzyme in the endonuclease activity and in the catalytic mechanism.
Bovine pancreatic ribonuclease A (RNase A) 1 (EC 3.1.27.5) is a well known enzyme whose main physical, chemical, and enzymatic properties have been the subject of extensive reviews (Richards and Wyckoff, 1971;Blackburn and Moore, 1982;Eftink and Biltonen, 1987). RNase A is an endonuclease that cleaves 3Ј,5Ј-phosphodiester linkages of single-stranded RNA when the base of the nucleotide in the 3Ј position is a pyrimidine. The use of well defined low molecular mass substrates such as pyrimidine 2Ј,3Ј-cyclic mononucleotides and dinucleoside monophosphates provided most of the kinetic data on which all mechanistic studies on RNase A have relied. Only a few kinetic studies have been carried out with longer oligonucleotides and homopolynucleotides such as poly(U) (Irie et al., 1984a(Irie et al., , 1984b, but no detailed studies with RNA as substrate have been carried out. The reasons for this are (i) the difficulties in the kinetic analysis derived from the complex structural features of the RNA molecule, (ii) the difficulty of monitoring a very fast reaction in a reliable fashion, and (iii) the spectrophotometric methods used can only give an average measure of all the species produced during the reaction, without giving any idea about either the size distribution of the products or the characteristics of the bond broken. The use of HPLC techniques has circumvented some of the problems concerning the analysis of the products of the reaction. Using an anion exchange column it was possible to measure the products of the reaction at high concentrations of the low molecular mass substrate cytidine 2Ј,3Ј-cyclic phosphate (CϾp). This method allowed the analysis of the partition of the reaction between the synthesis of the dinucleotide CpCϾp and the hydrolysis to 3Ј-CMP (Guasch et al., 1989;Cuchillo et al., 1991) and the separation between transphosphorylation and hydrolysis reactions using CpC as substrate (Cuchillo et al., 1993). In addition to the catalytic center, several phosphate-binding subsites that recognize the negatively charged phosphates of RNA have been described (Parés et al., 1980;Parés et al., 1991;Fontecilla-Camps et al., 1994;Nogués et al., 1995) (Fig. 1). A non-catalytic phosphate-binding subsite (p 2 ) adjacent to the catalytic center (p 1 ) was postulated from the chemical modification of RNase A with the halogenated nucleotide 6-chloropurine-9-␤-D-ribofuranosyl 5Ј-monophosphate (cl 6 RMP) (Parés et al., 1980). The reaction yielded a major derivative (derivative II) with the nucleotide label attached to the ␣-amino group. Different studies suggested that the phosphate group was bound in a specific phosphate-binding subsite, p 2 , and that Lys-7 and Arg-10 were involved in that subsite (Arú s et al., 1981;Irie et al., 1986;de Llorens et al., 1989;Richardson et al., 1990). The structure of derivative II was recently solved by x-ray crystallography (Boqué et al., 1994), and the structure found is in accordance with the location of p 2 , B 3 , and R 3 2 subsites at the N-terminal region of the protein. By means of * This work was supported by Grants PB91-0480 and PB93-0872 from the Dirección General de Investigación Científica y Técnica of the Ministerio de Educación y Ciencia (Spain) and GRQ93-2093 from Comissió Interdepartamental de Recerca i Tecnologia of the Generalitat de Catalunya (Spain). Equipment purchase was made possible by the Fundació M. F. de Roviralta (Barcelona, Spain). 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.
1 The abbreviations used are: RNase A, bovine pancreatic ribonuclease A; derivative II, a covalent derivative obtained by reaction of RNase A with 6-chloropurine riboside 5Ј-monophosphate in which the label is attached to the ␣-amino group of Lys-1; K7Q and K7Q plus R10Q, ribonuclease A in which Lys-7 and Lys-7 and Arg-10 are replaced by Gln; cl 6 RMP, 6-chloropurine-9-␤-D-ribofuranosyl 5Ј-monophosphate; CϾp, cytidine 2Ј,3Ј-cyclic phosphate; CpC, cytidylyl (3Ј-5Ј)cytidine; (Cp) n CϾp, an oligocytidylic acid of n ϩ 1 residues ending in a 2Ј,3Ј-cyclic phosphate; HPLC, high pressure liquid chromatography. kinetic analysis of RNase A derivatives obtained by site-directed mutagenesis, an indirect role for the p 2 subsite in the enzyme's catalytic mechanism was demonstrated (Boix et al., 1994). In the present work the pattern of oligonucleotide formation by RNase A using poly(C) as substrate was analyzed by means of reversed-phase HPLC, and the contribution of the binding subsite p 2 to the product distribution was studied in both native RNase A and in p 2 -modified molecules. The results indicate that in the poly(C) cleavage RNase A does not act in a random fashion but rather prefers the binding and cleavage of the longer substrate molecules. In addition the phosphodiester bond broken should be 6 -7 nucleotides apart from at least one of the ends to be preferentially cleaved by RNase A. Finally, deletion of p 2 abolishes the endonuclease activity of RNase A, which is substituted by an exonuclease activity. All these facts can be explained in terms of the multisubsite structure of the enzyme.

Materials
Bovine pancreatic RNase (5 ϫ crystallized), cl 6 RMP, poly(C), CϾp and 3Ј-CMP were from Sigma. Acetonitrile HPLC grade was obtained from Farmitalia Carlo Erba (Milan, Italy) and ammonium acetate HPLC grade was from Scharlau (Ferosa, Barcelona, Spain). Distilled water treated with a Milli Q water purification system (Millipore Corp., Milford, MA) was used throughout. All other reagents were of analytical grade. The reversed-phase HPLC column, Nova-Pak C 18 (2 ϫ 150 mm), was purchased from Waters Corp. (Milford, MA), and the anion exchange column Nucleosil 10SB was obtained from Macherey, Nagel and Co. (Dü ren, Germany).

Apparatus
All HPLC experiments were carried out with a modular Waters system consisting of two pumps (model 510), a liquid chromatography injector (U6K), and an absorbance detector (model 490E). The system was controlled by the Maxima 820 program (Dynamic Solution, Division of Millipore Corp.). An Eppendorf centrifuge was used routinely to remove any particulate material before injecting the samples into the HPLC system.

Protein Purification
The RNase A fraction was obtained by a modification of the method of Taborsky (Alonso et al., 1986). Derivative II was obtained by reaction between cl 6 RMP and RNase A according to Parés et al. (1980) and Alonso et al. (1986). K7Q and K7Q plus R10Q RNase A mutants were obtained and characterized according to Boix et al. (1994).

Analysis of the Digestion Products of Poly(C)
Digestion of Poly(C) and Separation of Oligocytidylic Acids-50 l of a 5 mg/ml poly(C) solution in 10 mM Hepes-KOH pH 7.5 was digested with 10 l of 30 nM RNase A at 25°C. At different time intervals, between 0 and 45 min, the products of the reaction were analyzed according to the method of McFarland and Borer (1979) by injecting 20 l of the reaction mixture on a reversed-phase HPLC column (Nova-Pak C 18 ) equilibrated with solvent A (10% ammonium acetate (w/v) and 1% acetonitrile (v/v) in water). The elution was carried out at a flow rate of 1 ml/min with an initial 10-min wash and a 50-min linear gradient from 100% solvent A to 10% solvent A plus 90% solvent B (10% ammonium acetate (w/v) and 11% acetonitrile (v/v) in water). After each run the system was washed for 10 min with 100% acetonitrile and reequilibrated with solvent A. Product elution was monitored and quantified from the absorbance at 260 nm. The analysis at reaction time t ϭ 0 was used to determine the presence of oligonucleotides in the poly(C) substrate. The elution position of small oligonucleotides was deduced from the pattern obtained after the reaction mixture was incubated for a long time (100 min) when no poly(C) was left. The separation pattern found by McFarland and Borer (1979) for the chemical hydrolysis of poly(C) was also taken into account for the identification of the different peaks.
A similar process was used to analyze the cleavage of poly(C) by derivative II and K7Q and K7Q plus R10Q RNase A mutants. In these cases the initial enzyme concentrations were 50, 35, and 62 nM, FIG. 1. Schematic diagram of the RNase A active site and adjacent binding subsites. B, R, and p refer to the binding subsites for base, ribose, and phosphate, respectively. B 1 is specific for pyrimidines and B 2 "prefers" purines. The phosphate group of the phosphodiester bond hydrolyzed by the enzyme binds to p 1 . Amino acids located near each site are indicated .
FIG. 2. Analysis by reversed-phase HPLC column (Nova-Pak C 18 column) of products obtained from poly(C) digestion by RNase A at time intervals from 0 to 45 min. See "Methods" for digestion and separation conditions. Note that in each chromatogram the best scale on the ordinate is used.

respectively.
Separation of CϾp and 3Ј-CMP-CϾp and 3Ј-CMP produced in the poly(C) digestion by RNase A eluted from the Nova-Pak C 18 column near the injection peak but so close to one another that no direct quantification was possible. The separation of these mononucleotides was achieved by concentrating the fractions eluted during the first 2 min in a Speed-Vac system. 200 l of the concentrated sample was injected on an anion exchange column (Nucleosil 10SB) equilibrated with 0.2 M ammonium acetate, pH 7.5. Sample separation was carried out with the starting eluent at a flow rate of 1 ml/min according to Alonso et al. (1985). Commercial CϾp and 3Ј-CMP were used as standards.

Separation of the Poly(C) Digestion Products-
The poly(C) used as substrate is a high molecular mass polymer. Fig. 2 (t ϭ 0 min) shows that poly(C) is eluted as a single fraction by reversed-phase HPLC (Nova-Pak C 18 ), indicating that although the sample is not electrophoretically homogeneous (information provided by Sigma) all high molecular mass components are eluted as a single peak and that oligonucleotides are not present in the sample at the initial conditions. Fig. 2 also shows the elution profile by reversed-phase HPLC of the (Cp) n CϾp oligonucleotides obtained from the poly(C) digestion with a low concentration of RNase A at different times within the range 0 -45 min. The elution position of the small oligonucleotides obtained from poly(C) digestion by RNase A was deduced from the pattern obtained after the reaction mixture was incubated for a long time (100 min) when no high molecular mass poly(C) was left (Fig. 3). The separation pattern found by McFarland and Borer (1979) for the chemical hydrolysis of poly(C) was also taken into account for the identification of the different peaks. Previous studies about the degradation of poly(C), poly(U), and poly(A) by RNase A had shown that the oligoribonucleotidic acids have a general structure of (Cp) n CϾp containing a 2Ј,3Ј-cyclic phosphate terminus, except for the fragment arising from the 3Ј terminus of the initial molecules of substrate (Imura et al., 1965;Irie et al., 1984b). The two mononucleotide products 3Ј-CMP and CϾp elute very near the injection peak and thus they cannot be directly measured. The smallest structure that can be clearly separated is the dinucleotide CpCϾp with an elution time of around 4 min. Oligomers of increasing size elute sequentially as a function of the amount of organic phase in the eluent.
The analysis of the oligonucleotide size distribution (Figs. 2, 4, and 5) shows that under the conditions used only polynucleotide fragments are formed during the early stages of incubation. However, shortly thereafter (5 min) a clear trend toward the formation of oligonucleotides with a size of about 6 -7 residues is observed. As expected, at the end of the process there is a clear increase in the number of small size oligonucleotides. These results suggest that the enzyme prefers binding and cleavage of long substrates and that to be preferentially cleaved by RNase A the phosphodiester bond has to be some six-seven nucleotides apart from at least one of the ends of the molecule.
The rate of appearance of mononucleotides CϾp and 3Ј-CMP was followed by rechromatography on an anion exchange HPLC column (Nucleosil 10SB) of the fraction eluted during the first 2 min of chromatography of poly(C) digestion products on the Nova-Pak C 18 column ( Table I). The formation of mononucleotides is a very slow process. For example, after 45 min of poly(C) digestion when no poly(C) is left but oligonucleotides are present in the medium as RNase A substrates (Fig. 2), the absorbance area at 260 nm of mononucleotide fraction accounts for only 14% of the total, 92% of which is CϾp and the remain- ing 8% of which is 3Ј-CMP. In agreement with previous results on CpC digestion by RNase A (Guasch et al., 1989;Cuchillo et al., 1993) there is an accumulation of CϾp in the reaction medium before its transformation to 3Ј-CMP.
Sequential Cleavage of Poly(C) Products-The action of RNase A on poly(C) gives rise to products which, in turn, are also substrates of the enzyme. This peculiarity results in a competition of the different substrate species for the enzyme leading to a partition of the catalyst between them. As shown by Fersht (1985) this partition is determined by the k cat /K m ratio of each competing substrate. Although in such a complex system as the present one it is very difficult to evaluate the k cat /K m ratios of the high molecular mass substrates, it is clear from Fig. 2 that there is a substrate size preference which, overall, can be viewed as a sequential process. The enzyme first acts on the longer substrates and then on the intermediate size substrates, and eventually all the CϾp is converted to 3Ј-CMP. This was checked by analyzing the evolution of some selected species. The original poly(C) substrate, the hexanucleotide (Cp) 5 CϾp, and dinucleotide CpCϾp intermediates were chosen. Fig. 4 shows the experimental values for the evolution of the three species. The progress curves for the three species are very similar to what would be expected from two consecutive and irreversible reactions (Fersht, 1985) according to Scheme I.

Poly(C)
RNase A 3 ͑Cp͒ 5 CϾp RNase A 3 CpCϾp SCHEME I The hexanucleotide species (Cp) 5 CϾp behaves as a transient intermediate, whereas CpCϾp accumulates during the observed reaction times. This product distribution suggested that RNase A shows a clear preference for polynucleotide or oligonucleotide substrates of high molecular mass rather than for low molecular mass substrates. These results can be explained according to the multisubsite structure of RNase A Nogués et al., 1995). The formation of the RNA⅐RNase A complex is mainly driven by interactions between the phosphate groups of the substrate and the active site (p 1 ) and the main phosphate-binding subsites (p 0 and p 2 ) of the enzyme (Fig. 1). However, other electrostatic interactions between phosphate groups of RNA and basic amino acid residues located at the surface of the protein are also involved in leading to optimal catalytic efficiency. The total occupancy of these binding subsites gives the best conformation for activity, and the additional binding energy clearly favors the action on the higher substrates.
Effect of the p 2 Phosphate-binding Subsite on the RNase A Substrate Preference-The additional p 2 phosphate-binding subsite adjacent to the 3Ј-side of the active site was postulated from the specific reaction of the halogenated nucleotide cl 6 RMP with RNase A (Parés et al., 1980). The major derivative, derivative II, is the result of an affinity-labeling reaction in which the phosphate group of the nucleotide binds to the specific p 2 phosphate-binding subsite before the formation of the covalent

FIG. 5. Comparison of (Cp) n C>p (n ‫؍‬ 0 -8) formation from poly(C) cleavage by RNase A (Ⅺ) and RNase A-modified forms (derivative II, K7Q plus R10Q RNase A mutant, and K7Q RNase A mutant as marked on the panels) ([o]).
Area percent in each case has been determined from the area of the corresponding peak eluted from the Nova-Pak C 18 column. Due to the different enzyme activities comparisons have been established using as reference the same undigested poly(C) fraction. Note that in each graphic the best scale on the ordinate is used. a Area percent of the fraction eluted during the two initial minutes of the Nova-Pak C 18 column with respect to the total area. b Distribution between CϾp and 3Ј-CMP (as percent area) found in the initial fraction eluted from the Nova-Pak C 18 column.
bond (Parés et al., 1980;Boqué et al., 1994). From chemical modification studies of RNase A Lys-7 and Arg-10 were postulated as constituent amino acid residues of p 2 (Richardson et al., 1990). Crystallographic studies of derivative II have confirmed the presence of Lys-7 in p 2 (Boqué et al., 1994), and several studies have also suggested a role for Lys-7 in the catalytic mechanism of RNase A (Brü nger et al., 1985;Filippi et al., 1987;Boix et al., 1994). The role of Arg-10 is less clear because structural studies of RNase A and RNase A⅐ligand complexes have indicated that this residue establishes a salt linkage with Glu-2, an interaction that contributes to the stability of the N-terminal ␣-helix (Boqué et al., 1994;Fontecilla-Camps et al., 1994). However, site-directed mutagenesis studies have pointed out that both Lys-7 and Arg-10 are constituents of a cationic cluster that forms the p 2 phosphatebinding subsite. In the absence of Lys-7, the nearby positive charge of Arg-10 can partially fulfill the phosphate binding role of p 2 . However, complete deletion of p 2 decreases the catalytic efficiency in p 1 even for small substrates such as CϾp, which only interacts in p 1 (Boix et al., 1994).
For these reasons we chose derivative II and K7Q and K7Q plus R10Q RNase A mutants to check the role of the p 2 phosphate-binding subsite in the poly(C) cleavage pattern by RNase A. In all three derivatives substrate binding in the p 2 subsite is greatly decreased due to either the occupancy of this subsite by the phosphate group of the marker nucleotide (derivative II) or to the partial (K7Q RNase A mutant) or total absence (K7Q plus R10Q RNase A mutant) of positive charges necessary for the enzyme-substrate electrostatic interactions in this region. Fig. 5 clearly shows that differences in oligonucleotide formation pattern take place as a consequence of a non-functional p 2 binding subsite. Due to the lower enzyme activity of RNase A mutants with respect to the native enzyme (Boix et al., 1994) the comparison has been established between digestion products containing the same percentage of undigested poly(C) frac-tion that is at equivalent stages of the digestion process. In contrast with the native enzyme, the cleavage of poly(C) by the RNases modified at p 2 shows a pattern in which, from the initial reaction time, the major product formed corresponds to the mononucleotide fraction and other small products, without accumulation of the hexanucleotide. This behavior is more apparent in both the derivative II and the K7Q plus R10Q with a totally non-functional p 2 binding site than with K7Q in which there is still a positive charge near p 2 . These findings indicate that the modified RNase cleaves the substrate in an exonucleolytic fashion in contrast with native RNase A, which acts preferentially as an endonuclease.

DISCUSSION
The results shown in Fig. 2 indicate that the breakdown of poly(C) catalyzed by RNase A is not a random process even though all internucleotide bonds in the substrate are susceptible to attack by the enzyme. Instead, it can be considered as taking place roughly in consecutive steps; during the early part of the reaction longer fragments are expected for a random endolytic reaction, but as the reaction proceeds there is a significant accumulation of oligocytidylic acids of 6 -7 residues, which in the final stages of the reaction are transformed into the mononucleotide CϾp and eventually into 3Ј-CMP ( Fig. 4 and Table I). These results suggest that the enzyme prefers binding and cleavage of long substrates and that the phosphodiester bond broken should be some 6 -7 residues apart from the end of the chain to be preferentially cleaved by RNase A. It should be noted that although this treatment is indicative of the behavior of RNase A on poly(C) it cannot be used to calculate accurately the kinetic parameters. This is because (i) although the hexanucleotide product has been taken as the main intermediate one should consider a broader population of products and (ii) the same enzyme concentration has been used for the two consecutive steps, i.e. the initial concentration, but FIG. 6. Model of the cleavage of an RNA chain by RNase A that explains the preference of the enzyme for long polynucleotide substrates. The model is based on the cooperative binding between the multiple protein subsites and the phosphates of the polynucleotide.
Step 1, a long RNA chain binds to RNase A.
Step 2, cleavage occurs in the active site resulting in the formation of two shorter oligonucleotide fragments, one of them ending with a 2Ј,3Ј-cyclic phosphate (2Ј,3Јc).
Step 3, the cooperative binding is weakened in the oligonucleotide fragments and this favors their replacement by a longer chain. In subsequent reactions the shorter fragments will also be broken, and eventually the hydrolytic step (formation of a 3Ј-phosphate from the 2Ј,3Ј-cyclic phosphodiesters) occurs when most of the substrate has already been cleaved by transphosphorylation. Adapted from Parés et al. (1991). the relative concentration of enzyme in each step is smaller as they are catalyzed by the same enzyme which, therefore, is partitioned between the two steps. An accurate knowledge of the k cat /K m for each substrate would allow knowledge of the proportion of enzyme acting in each reaction.
These results are supported by kinetic studies that suggest the contribution of phosphate-binding subsites to the catalysis. Steady-state kinetic studies of RNase A with oligonucleotides as substrates indicate that the k cat values increase with the substrate size, the k cat value for UpApA and UpApG being 3-5 times higher than those of UpA (Irie et al., 1984a). The use of oligouridylic acids has shown similar results (Irie et al., 1984b). From these findings it is demonstrated that the p 0 , p 1 , and p 2 binding sites play an important role in catalysis. Moreover, when poly(U) is used as substrate the V max value is 3-20 times higher than the values with oligouridylic acids depending on the assay conditions (Irie et al., 1984b). These kinetic results suggest that additional binding subsites must contribute to the catalytic efficiency. Crystallographic analysis of complexes between the protein and the deoxyadenylic acid tetramer supports the existence of multiple subsites in RNase A. A virtual DNA strand composed of 12 nucleotides traces out a nearly continuous path. The binding between protein and nucleic acid takes place through salt bridges between phosphate groups and nine charged side chains (McPherson et al., 1986).
Our results support the model of the cleavage of an RNA chain based on cooperative binding between the multisubsite structure of RNase A and the phosphates of the polynucleotide as proposed by Parés et al. (1991) (Fig. 6). This model proposes that the strong binding between RNA and RNase A takes place primarily through electrostatic interactions between the phosphate groups of the substrate and basic amino acid residues located on the surface of the protein. After the cleavage of the phosphodiester bond at the active site the cooperative binding is weakened in the fragments, and the displacement by a new long chain of RNA occurs. Subsequently, shorter fragments will be progressively cleaved in the order of maximum to minimum occupancy of the RNase A subsites. Finally, the hydrolytic step takes place.
The contribution of phosphate-binding subsites to the enzymatic process has been analyzed in p 2 chemically modified RNase A (derivative II) or by means of site-directed mutagenesis (K7Q and K7Q plus R10Q mutants). In both cases the product distribution pattern from poly(C) cleavage is altered (Fig. 5), and the significant increase in the mononucleotide fraction with respect to the production of oligonucleotides indicates that in these modified RNase A forms an exonuclease activity is favored. The RNase specificity has also been altered by site-directed mutagenesis of amino acid residues of B 1 , the pyrimidine binding site (delCardayré and Raines, 1994); the replacement of Thr-45 by amino acids with smaller side chains (i.e. Ala or Gly) increases the efficiency for poly(A) transphosphorylation up to the level of poly(C), and at the same time a processive poly(A) cleavage is seen in contrast with the distributive cleavage characteristic of wild-type RNase A. Thus, these substitutions not only modify the specificity of substrate binding, but they also alter the cleavage pattern. In addition, an indirect role for the amino acid residues of the p 2 binding subsite in the RNase A catalytic process has been demonstrated, since the catalytic efficiency is strongly affected not only in the case of an RNA substrate, which occupies all the binding subsites, but also in the case of the CϾp substrate, which only interacts in the catalytic center (B 1 R 1 p 1 ) (Boix et al., 1994). In conclusion, the best catalytic efficiency of RNase A is produced with long substrates as a consequence of the occupation of all the phosphate-binding sites. The p 2 phosphate-binding subsite also plays an important role in the catalytic mechanism and contributes significantly to the endonuclease activity of RNase A.