Purification and Properties of a Novel Pyrimidine-specific Endoribonuclease Termed Endoribonuclease VI1 from Calf Thymus That Is Modulated by Polyadenylate”

Endoribonuclease VII, a novel endoribonuclease from calf thymus, was identified and purified by us. The purified enzyme has M, = 74,000; its homogeneity was checked by analysis in polyacrylamide gels (both in the presence and in the absence of sodium dodecyl sulfate). The nuclease cleaves poly(U) and poly(C) while other single-stranded homopolyribo- as well as polydeoxyribonucleotides are not degraded; poly(A,C) is hydrolyzed to a smaller extent, while poiy(U) ‘ poly(A) is not degraded at all. Poly(A) modulates the poly(U)-degrading activity; at a molar ratio of approx- imately 1 [poly(A)]:lO [poly(U)], a more than 100% stimulation of the enzyme activity was achieved, while at lower ratios an almost complete inhibition of the enzyme activity resulted. Binding studies revealed that endoribonuclease VI1 has a marked affinity for poly(A) and poly(U). During hydrolysis, oligo(U)lz fragments with 3”OH and 5’-P termini are formed. The basic enzyme (PI = 8.5) has its activity optimum at pH 7.2, requiring neither monovalent nor divalent cations; the enzyme is not inhibited by thiol group reagents. Several lines of evidence suggesting a role of endoribonu- clease VI1 in mRNA processing are presented.

). However, little is known about the regulatory mechanisms which must exist to control the activities of the poly(A) anabolic and poly(A) catabolic enzymes during the polyadenylation of hnRNA. Evidence is presented that this control may be performed first by poly(A)-binding proteins (10, l l ) , second by tubulin and G-actin (12), third by modified nucleotides in the poly(A) tract (13), and fourth *This investigation was supported by a grant awarded by the Deutsche Forschungsgemeinschaft. 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.
$ To whom correspondence should be addressed. Polynucleotides and nucleotides are abbreviated according to the recommendations of the Commission on Biochemical Nomenclature (57). It has been proposed that the poly(A) segment of mRNA governs the metabolic stability of mRNA, facilitates the transport of mRNA through the nuclear membrane, and orients the splicing of mRNA precursors (surveyed in Refs. 19 and 20). The appealing model, in which poly(A) aligns the splicing sites of mRNA precursors (201, requires the following: first, a strict conservation of sequences present in the triple-stranded structures formed between poly(A) and the oligo(U) and oligo(A) stretches within the intervening sequences; second, stabilizing elements for these triple-stranded structures which could be small nuclear RNA species (21); and third, recognition sites for the hitherto unknown splicing enzymes. In the present paper, we describe a novel ribonuclease termed endoribonuclease VII. This enzyme shows properties that characterize the nuclease as a possible candidate for a splicing enzyme in the hypothesized poly(A)-mediated splicing process: binding to poly(A) and poly(U), degradation specificity toward poly(U) but not toward poly(A), and modulation of poly(U)-hydrolyzing activity by defined poly(A) concentrations of a distinct chain length.

VI1
Endoribonuckase Vff Assay-Unless stated otherwise, all reactions were carried out in a final volume of 100 pl at 37 "C for 10 min. Each reaction mixture contained 100 mM Tris-HC1 (pH 7.2), 2.5 mM MgC12, 20 mM EDTA, 400 mM urea, 50 pg of bovine serum albumin, and 8 nmol of [3H]p~ly(U) (specific radioactivity, 40,000 dpm/nmol)/90 pl, and 10 pl of the enzyme preparation (containing 3 enzyme units in maximum). For determination of enzymic activity, 40-pl aliquots were routinely taken and acid-insoluble material was collected, applying the GF/C filter technique (25). The details of these procedures were given earlier (26). 40% of [3H]poly(U) was found to be bound at time 0 of incubation to the filters. Determinations revealed that oligo(U)lo fragments were not bound to GF/C filters. Under these assay conditions, the reaction velocity increased linearly during the first 15 min. One unit of enzyme is defined as the amount which converts 1 nmol of [3H]poly(U) into an acid-soluble form during a 10-min period under the standard conditions. Phosphodiesterase Assays-Phosphodiesterase assays were composed basically as described (27). 0.4 nmol of [3H]oligo(U) (1.4 X 10' dpm) were incubated (37 "C, 10 min) in a reaction volume of 200 pl with 5 pg of phosphodiesterase I in a 200 mM Tris-HC1 buffer (pH 8.8, 10 mM MgCl,). 40-pl samples were taken at 0 and 10 min, and analyzed by gel chromatography on Sephadex G-50 (9). The monoribonucleotides appeared at a V,lV, of 4.0-4.4, while the oligoribonucleotides eluted at a VJVO of lower than 3.5.
The phosphodiesterase I1 assay contained 0.4 nmol of [3H]oligo(U), 5 pg of phosphodiesterase I1 in a 2.2 mM sodium succinate HCI buffer (pH 6.5). Incubation was performed at 37 "C and the product was analyzed with respect to the amount of oligoribonucleotides and monoribonucleotides (9).
Binding Assay of Nucleic Acids to Endoribonuckase VZZ-Complex formation between nucleic acids and endoribonuclease VI1 was determined applying the nitrocellulose filter binding technique of Jones and Berg (28). The binding mixture (total volume, 100 pl) contained 20 mM Tris-HC1 (pH 7.5), 300 mM KCl, 2 nmol of labeled RNA or DNA, and 1.2 pg of enzyme protein. After incubation (5 min, 37 "C), the samples were placed onto the filters and washed with 30 ml of the binding buffer supplemented with 1% dimethyl sulfoxide. The radioactivity retained on the filters was determined. In the control assays, the enzyme was replaced by 3 pg of bovine serum albumin.
The sedimentation coefficient of endorihonuclease VI1 complex was determined by velocity gradient centrifugation as described (30). The enzyme samples were applied to 10-30% sucrose gradients (in a 10 mM Tris-HC1 buffer, pH 8.0, supplemented with 10 mM NaCl and 400 mM urea) and centrifuged for 4 h at 4 "C and 40,000 rpm in an SW 41 rotor.
Analytical isoelectric focusing was carried out on a 15-ml column. After dialysis against 0.9% NaCI, 4 M urea, 2 mM dithiothreitol, and 5% (w/v) sucrose, samples of 200-400 pg of protein were added in the middle of a linear 5-50% sucrose gradient (9 ml) containing 5% carrier ampholytes (pH 3.5-lo), 1 mM dithiothreitol, 4 M urea, and 100 pg of ferritin as marker. Focusing was performed at 4 "C for 19 h with 300-V constant voltage and for 1 additional h with 400 V. As buffer systems, 10 mM phosphoric acid (anode) and 20 mM NaOH (cathode) were used. Fractions of 0.4 ml were collected and the pH of each fraction was measured at 4 "C. Endoribonuclease VI1 activity of each fraction was determined in the standard assay.
Preparative isoelectric focusing was carried out on a 60-ml column and a 50-ml linear sucrose gradient under otherwise identical conditions as described for the analytical procedure. After focusing, fractions of 2 ml were collected.
The chain length of the reaction product was determined by gel chromatography using Sephadex G-50 (9). A column (1.1 X 34.5 cm) was equilibrated with 100 mM Tris-HC1 (pH 7.5) and 2 M urea. After application of 1-ml samples, fractions of 0.45 ml were collected and counted in 10 ml of Aquasol each.
The nature of the monoribonucleosides and -nucleotides was determined by ion exchange chromatography using DEAE-Sephadex A-25 (9, 31). Applying a bicarbonate buffer, the monomers appeared at the following molarity of the buffer: Urd, 0.05 M; P-2':3'-Urd, 0.18 M; 5'-UMP, 0.26 M; and 3'-UMP, 0.32 M. Electrophoresis in the presence of sodium dodecyl sulfate was carried out according to the method of Weber and Osborn (32). Protein samples were heated for 5 min at 100 "C in the presence of sodium dodecyl sulfate and mercaptoethanol. A 30-pg sample of protein was analyzed on 10% gels. The gels were stained with Coomassie brilliant blue.
Polyacrylamide gel electrophoresis (10% separation gel, 6% spacer gel) was performed in tubes in the presence of 6 M urea (33), using a 120 mM acetic acid, 300 mM &alanine (pH 4.3) buffer system. Prior to electrophoresis, the enzyme samples were dialyzed (12 h, 2 "C) against the electrophoresis buffer (pH 4.3, supplemented with 6 M urea). The gels were stained with Coomassie brilliant blue. In order to detect endoribonuclease VI1 activity in situ, the 7.4-cm gels were cut into 2-mm slices. Each slice was incubated in 200 pl of endoribonuclease VI1 incubation mixture for 30 min at 37 "C under shaking; after incubation, acid-insoluble material was collected using the GF/ C-filter technique (7).

Purification of the Enzyme
All procedures were carried out at 0-4 "C unless stated otherwise; a summary of the steps is given in Table I.
Step 1: Crude Extract-200 g of frozen calf thymus from 2month-old animals were cut into pieces, thawed, and suspended in 1 liter of 0.2 M NaC1, 5 mM MgC12, 10 mM Tris-HC1, pH 8.0. The suspension was mixed in a Waring Blendor for 10 min and subsequently homogenized in an Ultra Turrax (Janke-Kunkel, Staufen) for 15 min. After standing for 30 min, the extract was centrifuged for 15 min at 12,000 X g. About 950 ml of supernatant fraction (fraction I) with 13.3 mg/ml of protein were obtained. The ratio A Z m &~ was 0.77.
Step 2: Ammonium Sulfate Precipitation-Fraction I (950 ml) was brought to 0.45% saturation with solid ammonium sulfate under stirring and allowed to stand for 60 min. The supernatant was collected by centrifugation (15 min; 12,000 X g) and ammonium sulfate was added again to bring the solution to 0.75% saturation. The precipitate formed was obtained by centrifugation (15 min; 12,000 X g), then dissolved in 50 ml of 100 mM Tris-HC1 (pH 8.0; 5 mM MgC12, 10 mM D-galactose), and dialyzed for 24 h against 100 mM Tris-HC1

TABLE I
Purification of endoribonuclease VU The enzyme Was purified as described in the text. One unit of activity is defined as the amount of enzyme producing 1 nmol of acid-soluble material in 10 min at 37 "C from poly(U). The calculation of the specific activity as well as of the degree of purification is based first on that enzyme activity that was determined in nonisoelectrically focused preparations (= overall activity), and second on that portion of enzyme activity in the crude extract that was measured after isoelectric focusing around the typical PI for endoribonuclease VI1 (pH 8.2-9.2).
Step 3: Sephadex G-150 Gel Filtration-Fraction I1 was passed in three successive runs (40-ml aliquots each) through a Sephadex G-150 column (4.5 x 26.5 cm) equilibrated with 100 mM Tris-HC1 (pH 8.0), 5 mM MgCI2, 10 mM D-galactose; 5-ml fractions were collected. The elution of protein was followed by absorbance at 280 nm and poly(U)-degrading enzyme activity was determined as described under "Experimental Procedures." The high molecular weight enzyme fractions between VJV0 values (34)  Step 4: DEAE-cellulose Fractionation-Fraction 111 was applied to a DEAE-cellulose column (2.2 X 7 cm) equilibrated with 100 mM Tris-HC1 (pH 8.0; 5 mM MgCl,, 10 mM Dgalactose). After washing with 3 column volumes of the equilibration buffer, the poly(U)-degrading enzyme was eluted from the column with the equilibration buffer, supplemented with 100 mM NaCI. Fractions of 2 ml were collected during elution. Each fraction was assayed for protein (A2m and A230; Ref. 29) and poly(U)-degrading enzyme activity. Active fractions were pooled for fraction IV. Total protein at this stage averaged about 200 mg in 310 ml. The A z a o A 2~ ratio changed to greater than 1.5, indicating that most nucleic acids have been removed (35). The specific poly(U)-degrading enzyme activity was determined to be 6930 units/mg.
Step 5: Phosphocellulose Adsorption-Fraction IV was dialyzed against a 50 mM Tris-acetate buffer (pH 4.2; 5 mM MgC12, 15% (v/v) glycerol). After centrifugation (15 min; 12,000 x g), it was applied to a phosphocellulose column, equilibrated with the same Tris-acetate buffer. After washing with 2 column volumes of the equilibration buffer, the poly(U)-degrading enzyme was recovered by changing the pH of the buffer to 8.0 (200 mM Tris-HC1) ( Fig. 2 A ) . Fractions 28-34 were pooled and contained in 42 m1/36.8 mg of protein (fraction V). The specific activity of the poly(U)-degrading enzyme activity was determined to be 11,480 units/mg. Fraction V contained two poly(U)-degrading enzyme activities which can be distinguished by their susceptibility to EDTA. In the absence of EDTA and in the presence of 2.5 mM MgC12, the enzyme preparation degraded poly(U) up to 3'-UMP ( Fig.  1). However, in the presence of 20 mM EDTA, only oligo(U) fragments (5"phosphate termini) with a chain length of around 10 UMP units were formed (Fig. 1).
Step 6: Preparative Isoelectric Focusing-Fraction V was subjected to isoelectric focusing in order to separate the two poly(U)-degrading enzyme activities (Fig. 2B). Two peaks of activity were obtained one peak at pH 8.5 (designated endoribonuclease VII) and a second one at pH 4.1. Fractions 28-30 (endoribonuclease VII) and 16-21 were pooled. The PI 8.5 enzyme accounted for 35% and the PI 4.1 enzyme for 65% of the enzyme activity recovered after focusing. Ampholine, which was present in the fraction, was removed by Sephadex G-150 gel filtration (column size, 2.2 X 10 cm; further details are given with the description of purification step 3). The specific activities of the final preparations were determined to be: endoribonuclease VII, 1.2 X lo5 units/mg; p14.1 enzyme, 0.3 X lo6 units/mg. The protein concentration of this endoribonuclease VI1 fraction was 62 Fg/ml.  and endoribonuclease V, we calculated the degree of purification in a second way as follows. The specific activity of the purified enzyme was related to that portion of enzyme activity in the crude extract, which was measured after electric focusing around the characteristic pl range (pH 8.2-9.2) of the endoribonuclease VI1 (in Table I termed PI 8.5 activity). In the crude extract, 6.1% of the total poly(U)-degrading enzyme activity was localized within a PI range of 8.2-8.8. A calculation on this basis revealed a 6340-fold purification of endoribonuclease VI1 ( Table I). The calculation of the yield of the purification of the enzyme, which amounts to 1.2%, is based on the overall poly(U)-degrading activity and the total protein content, present in the crude extract. Stored in 50% (v/v) glycerol, Fraction VI was stable for more than 5 months without measurable change of its enzyme activity. Freezing of the enzyme in the absence of glycerol at -20 "C resulted in a 85% loss of activity.

Enzymic Contamination
The purified endoribonuclease VI1 was free of nucleotidase and exoribonuclease activity. This conclusion must be drawn from experiments in which the nature of the oligo(U) products formed by this enzyme was determined. As summarized later, only oligo(U)g_14 fragments with 3"OH and 5"phosphate termini were detected as end products during the reaction with step VI enzyme in the absence of EDTA. The possible existence of a phosphodiesterase in the preparation can also be excluded, because no DNA-degrading activity was detectable (Table 11).

Properties
If not stated otherwise, the experiments were performed with Fraction VI. Sedimentation Property and Molecular Weight-Fraction VI of endoribonuclease VI1 was analyzed by velocity sucrose gradient centrifugation (Fig. 3). The enzyme was recovered in the top fractions (58%) and in the bottom fractions (42%); the latter had a sedimentation coefficient of approximately 45 s.
The two activity fractions (bottom fraction and top fraction) were analyzed by gel electrophoresis both under denatured and native conditions. Analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed (Fig. 44) the bottom fraction (lane b) to be composed of six major proteins (one with an M, = 74,000, which comprised about 5% of the total protein and five species with M, < 50,000) while the top fraction (lane a) contained only one protein species with M , = 74,000. That protein component in the preparations that was associated with the endoribonuclease VI1 activity was successfully identified after separation in urea-polyacrylamide gels (Fig. 4B). After separation of the bottom fraction (lane b), two main bands were detected in the separation gel ( R F 0.05 and 0.85); approximately 40% of the protein migrated into the separation gel. In contrast, in the top fraction, only one band had been visualized after electrophoresis (lane a), which had the same relative electrophoretic mobility as the fast migrating band from the bottom fraction. The in situ determination of nuclease activity in urea gels revealed (Fig. 4C) that the two protein bands in the separating  gel from the bottom fraction and the single band from the top fraction coincide with the nuclease activity.
From these findings, the hint is taken that the enzyme which was present in the top fractions after sucrose gradient centrifugation was purified to homogeneity. Its apparent M, = 74,000. It is presently assumed that the second nucleaseassociated band identified in urea gels ( R F 0.05) represents a relatively stable protein complex.
Isoelectric Point-The PI of fraction VI of endoribonuclease VI1 was determined in a 15-ml sucrose density gradient, containing 4 M urea. The enzyme activity was recovered in a single sharp peak at a pH of 8.5. The enzyme activity coincides with the protein band (not shown).
Substrate Specificity-The specificity of endoribonuclease VI1 was tested in assays using different natural and synthetic polyribonucleotides as substrates ( Table 11). The enzyme was found to hydrolyze preferentially poly(U) and poly(C) and, with a slightly lower affinity, poly(A,C) as well. No other activity. In parallel unstained urea gels, the nuclease activity was determined after slicing the gel as described under "Experimental Procedures." Nuclease activity is shown in the separated top ( l a n e B, a; 0 ) and bottom fractions ( l a n e B, b; 0). single-stranded deoxyribo-and ribopolymers were degraded. Poly(U) hybridized to poly(A) was not a substrate for the enzyme. In a 1:1 molar mixture between p~l y ( [~H ] U ) and poly(A), the poly(U) substrate was hydrolyzed only to 13% (Table 11). This result already indicates that poly(A) inhibits the endoribonuclease VI1 activity. A more detailed description of this outcome is given later. DNA was not hydrolyzed, while tRNA was degraded with an efficiency of 44%, if compared with poly(U) (100%).
Kinetics of Poly(U) Degradation-Under the experimental conditions used (addition of 1.3 enzyme units), the kinetics of poly(U) degradation during the first 15 min was linear and without a lag phase (Fig. 5A); then the curve leveled off when high molecular weight substrate was exhausted.
The Michaelis constant with poly(U) as substrate (2 to 10 p~ phosphate in the reaction mixture) was found to be 11.3 PM with respect to phosphate. The maximal velocity was 13.2 nmol of acid-soluble materiallpg protein X min.
The molecular activity of the enzyme estimated from the results described in Fig. 1 shows that 1.4 enzyme units (= 12 ng of enzyme complex, corresponding to 1 ng of the 74,000-Da enzyme protein) degrades 1 nmol of po1y(U)135 (with respect to phosphate) or 7.5 pmol of po1y(U)135 (with respect to entire polymer) during an incubation period of 10 min to oligo(U)12 fragments. This means that 8.4 X lo9 enzyme molecules have cleaved approximately 4.9 X l O I 3 phospho-  (Fig. 1). The average chain length of the fragments was estimated graphically, using the semilogarithmic plot procedure, in which the chain lengths of the oligoribonucleotides are correlated with their VJVo values. Calibration was performed with marker oligoribonucleotides (9). By this approach, the oligo(U) fragments were determined to consist of 11.5 k 2.3 UMP residues. The chemical nature of the oligo(U) fragments at their 3' and 5' termini was determined enzymatically using phosphodiesterases I and 11. Oligo(U) fragments were formed during incubation of ['H]poly(U) with 12 units of endoribonuclease VI1 for 60 min. Subsequently the samples were heated (2 min, 95 "C) to destroy the nuclease activity and the reaction products were separated on Sephadex G-50 (Fig. 1). Fractions 14-19 ( V,/Vo, 1.84-2.5) were taken, pooled, and concentrated in a rotary evaporator (at 40 "C) to 100 pl. The residue was desalted by a Sephadex G-15 gel filtration column (0.8 X 12 cm) and the oligoribonucleotides were collected. 0.4 nmol of oligo(U) was analyzed in both the phosphodiesterase I and I1 assays (see "Experimental Procedures"). After incubation, 94% of the oligomers were converted to monoribonucleotides after incubation with phosphodiesterase I, while only 2% were degraded in the presence of phosphodiesterase I1 (Table 111).
The nature of the monoribonucleotides, released after incubation with phosphodiesterase I, was analyzed by ion exchange chromatography (see "Experimental Procedures") and determined to be to 96% 5'-UMP. These data prove that the oligo(U)II products, formed by endoribonuclease VI1 are 3'hydroxy-terminated and 5"phosphate-terminated oligonucleotides (271, and indicate that this enzyme is a 3"endoribonuclease. Effect of Enzyme Concentration-Poly(U) degradation was proportional to the enzyme concentration within a 0.005-0.02-pg range in the standard assay containing 8 nmol of poly(U). Again, at higher concentrations, the rate decreased due to the limitation of the substrate (Fig. 523).
Effect of pH and Selected Compounds-The pH optimum was determined in a 50 mM Tris-acetate buffer system between pH 4.0 and 8.0. The maximal activity was measured at pH 7.2. At pH 6.5, the enzyme activity was determined to be 68% of the optimal activity and at pH 8.0 only 36%. Addition of 2.5 mM MgCL or MnClz had no influence on enzyme activity. At higher ionic strength, present in the reaction TABLE I11

Analysis of oligo(U) product formed by endoribonuclease VII
Poly(U) was exhaustively hydrolyzed by the nuclease and the oligo(U) fragments were collected by gel filtration (details are given in Fig. 1). After desalting, phosphodiesterase I or I1 was added to oligo(U) and incubated as described under "Experimental Procedures." The amount of oligoribonucleotides left was determined by Sephadex G-50 gel filtration. The monoribonucleotide product, formed during the phosphodiesterase I reaction, was analyzed on ion exchange chromatography. Details are given in the text. mixture, the enzyme activity decreased; addition of 100 mM NaCl reduced the activity to 62% and of 300 nM to less than 5%.
For optimal enzyme activity, urea a t a concentration between 300 and 500 mM had to be added to the reaction mixture. In the absence of urea, only 65% of the activity could be detected (control, 100%). Concentrations above 0.8 M urea are inhibitory; a t 1 M, the enzyme activity decreased by 10% and at 3 M by 41%. The enzyme was not inhibited by the following reagents for thiol groups at a concentration of 5 mM: N-ethylmaleimide, o-iodosobenzoate, iodoacetamide and HgC12. K-phosphate (pH 7.2), however, was determined to be a strong inhibitor; 10 mM reduced the activity by 27% and 100 mM by almost 100%.
Effect of Temperature-The temperature optimum was determined to be 37 "C; the temperature coefficient Qlo (36) 37/ 27 "C was 1.84 and the 37/47 "C was 1.87. Endoribonuclease VI1 is a thermally unstable enzyme. More than 95% of its activity had been destroyed irreversibly after heating to 56 "C for 7 min.
Modulation of the Enzyme Activity by Homopolymers-As summarized above (Table 11), endoribonuclease VI1 does not use poly(A), poly(G), and poly(&) as substrate. However, these homopolymers affected the activity of this enzyme in assays with its specific substrate poly(U) (Fig. 6). For these experiments, we chose poly(A), poly(G), and poly(dA) fractions of comparable size (85-120 nucleotide units). Poly(dA) and poly(G) were found to cause only a strong inhibitory effect on endoribonuclease VI1 activity (fraction VI), while poly(A)9s stimulated the enzyme by more than 100% a t a limited concentration range; maximal stimulation was obtained at a concentration of 0.5 nmol/assay. At concentrations higher than 2 nmol/assay, poly(A) caused an inhibition of 50%. These results indicate that at a molar ratio (based on phosphate content) of approximately 1 [poly(A)]:lO [poly(U)] maximal stimulation of the enzyme activity is achieved. At lower ratios, 1:2 (Fig. 6) or 1:l (Table 11), the poly(U) hydrolysis is inhibited by 52 or 87%, respectively. The extent of stimulation is dependent on the chain length of the polymer. Studying the same concentration range, oligo(A)4 was found to have only a little modulating effect on the enzyme activity; a maximal stimulation of 15% and a maximal inhibition of 12% was achieved. The described effects of homopolymers on endoribonuclease VI1 activity are not restricted to the enzyme, present in the purification state of step VI, but was observed also with the purified enzyme (top fraction after sucrose gradient centrifugation; Fig. 3) as well as with the particleassociated enzyme in the bottom fraction (Fig. 3).
In a comparative series of experiments, we have determined the influence of po1y(A)95 on the activity of ribonuclease A by adjusting its enzyme concentration so that almost the same amount of poly(U) is converted to an acid-soluble form as in the studies with endoribonuclease VII. Under these conditions, poly(A) inhibits ribonuclease A even at very low concentrations, down to 0.5-1 nmol/assay (Fig. 6).
In a further approach, the high affinity of poly(A) for endoribonuclease VI1 can be demonstrated. Applying the nitrocellulose filter technique, we studied the binding capacity of endoribonuclease VI1 (fraction VI) for different nucleic acids. The KC1 concentration in the assay was adjusted to 300 mM in order to suppress the enzymic activity. Under this condition, endoribonuclease VI1 was determined to bind not only to the poly(U) substrate, but also to the poly(A) modulator, while in the presence of radioactively labeled poly(C), poly(G), and DNA, less than 8% of the input radioactivity is retained on the filters (Fig. 7). In control experiments, with 3 Fg of bovine serum albumin/assay, a nonspecific binding of poly( t3H]A) andpoly( [3H]U) around 3-5% of the radioactivity is found. The radioactively labeled nucleic acid species (amount of radioactivity per assay: poly(A), 2310 dpm; poly(U), 1800 dpm; poly(C), 250 dpm; and DNA, 560 dpm) were incubated in the absence or presence of endoribonuclease VII. After incubation, the amount of radioactivity retained on nitrocellulose filters was determined as described under "Experimental Procedures." The percentage of radioactivity retained on filters in the presence of endoribonuclease VI1 (fraction VI) is shown. In control assays using bovine serum albumin instead of endoribonuclease VII, less than 5% of the labeled nucleic acids were bound.
Endoribonuclease VII, after purification by ammonium sulfate fractionation, gel chromatography, ion exchange chromatography, and isoelectric focusing, has been analyzed by velocity sucrose gradient centrifugation. During this procedure, the enzyme was separated into two fractions, a top ( t 5 S) and a bottom fraction (45 S). The top fraction containing only one protein species of M , = 74,000 is associated with the activity. Using this criterion, the enzyme, present in the top fraction, was purified to homogeneity. The bottom fraction contained this enzyme species as well, yet together with four major (Mr <50,000) and five minor proteins. The enzyme in this complex is associated not only with protein but also with RNA (not shown), suggesting that endoribonuclease VI1 is cific Endoribonuclease VII 7039 bound to RNP particles. Since the enzyme activities present in the two fractions responded to substrates in identical fashion (not shown) and to poly(A) with the same typical dose-dependent stimulation or inhibition and since they can be characterized by the same electrophoretic mobility, we assume that the activity in the top fraction ((5 S) had been liberated from the 45 S particles.
From densitometer tracings of Coomassie-stained urea gels, we conclude that approximately 8% of the total protein of the 45 S particle is endoribonuclease VII. With this figure, the relative molecular mass, the amount of pure enzyme which had been obtained from 200 g of thymus gland (yield 1.2%), and the fact that approximately 4.6 X lo9 cells are present in 1 g of thymus (45), the concentration of endoribonuclease VI1 can be estimated as 2.2 x lo4 molecules/cell. In comparison, endoribonuclease VI1 is present in a concentration of 4 X lo4 molecules/oviduct cell (42) and endoribonuclease V in a concentration of 1.6 X IO5 molecules/thymus cell (9). The molecular activity of the enzyme was determined to be 180 cleavages performed/min/molecule of enzyme (see "Results" Endoribonuclease VI1 is a basic protein (isoelectric point, 8.5) which shows an activity optimum at pH 7.2 without requirement for divalent cations for its activity and is insensitive towards thiol group reagents (e.g. N-ethylmaleimide), very much in contrast to other cellular RNases which degrade double-and single-stranded RNAs and which are inhibited already by low concentrations of N-ethylmaleimide (46).
One interesting feature of endoribonuclease VI1 is its affinity for poly(A). Dose-response experiments revealed that poly(A) causes an over 100% stimulation of the enzymic poly(U) degradation, if a concentration ratio (based on moles of phosphate) of approximately 1 [poly(A)]:lO [poly(U)j was adjusted in the assay. This stimulation is observed with poly(A) only, while oligo(A), is nearly ineffective. At higher ratios, poly(A) inhibits poly(U)-hydrolyzing activity up to 50%. In previous experiments, it was established (47) that poly(A) is inhibitory for pancreatic ribonuclease A. Therefore, we performed comparative studies which revealed that poly(A) is not stimulatory for this enzyme under such concentration conditions at which this polymer positively affected endoribonuclease VII. The stimulation of endoribonuclease VI1 activity was specific for poly(A); poly(dA) and poly(G) caused an inhibitory effect only. Binding experiments confirmed the evidence that endoribonuclease VI1 interacts not only with poly(U) but also with poly(A). Although endoribonuclease VI1 utilizes both poly(U) and poly(C) as substrate, the filter binding studies revealed only an affinity of the enzyme for poly(U) but not for poly(C). We therefore suppose that the enzyme has the property of discrimination between these two substrates due to different binding affinities. As resported under "Results," endoribonuclease VI1 showed substantial activity with a tRNA substrate; the hydrolysis products are oligoribonucleotides in nature. No stimulation of the enzyme had been observed with tRNA (data not given), which might be due to the lack of oligo(U) sequences in this polymer.
The functional significance of poly(U)-andpoly(C)-specific endoribonuclease VII, which seems to be a constituent of an RNP complex and whose activity is modulated by poly(A), is not known. However, based on the protein pattern of the endoribonuclease VI1 complex and the probable existence of RNA in this particle, a homology with the small nuclear RNPs (snRNP) might be assumed (48). 7-11 proteins have been identified in snRNPs (21,49); most of them have M , = 10,000-50,000, while one species has been identified as a M, by guest on March 24, 2020 http://www.jbc.org/ Downloaded from = 70,000 protein (50). An almost identical composition has been established for the endoribonuclease VI1 complex in the present study. A discrepancy between the two complexes exists with respect to the sedimentation coefficient; while snRNPs sediment with 11-12 S (51), the endoribonuclease VI1 complex has a sedimentation coefficient of 45 S. Taking the known high affinity of snRNA (or snRNP) to 30 S nuclear particles (52, 53) and to hnRNP monomers (51) into consideration, a hypercomplex is presently assumed. It could explain the high Svedberg value for the endoribonuclease VII-containing complex. This assumption is supported by the observation of Howard (54) showing that 30-60 S RNP complexes from mouse erythroleukemia cell nuclei are composed of snRNP and hnRNP. We are exploring these and other possible RNP interactions to locate more precisely endoribonuclease VI1 in its functional position in the cell.
Studies from Jelinek and Leinwand (55) revealed a complementary pairing of snRNA in the snRNP complex with nuclear and cytoplasmic poly(A)-containing RNA. The recognition signal for this complex formation is not known, but it may constitute a transitional conformation of hnRNP, aligned by the poIy(A) terminus (20). This putative complex could provide a splicing matrix for these enzyme(s). Thus far, only RNase 111, P, or Q are thought to be candidates for enzymes that cleave RNA precursors during their processing (56). We propose endoribonuclease VI1 as a further potential splicing enzyme, because it comprises relevant properties which are thought to be prerequisites for a poly(a)-facilitated splicing event: (a) reversible storing in a protein/RNA complex; (b) base specificity and structure specificity of the substrate to be hydrolyzed; and (c) a unique dependence from poly(A), which modulates its activity in a stoichiometric fashion.