Characterization of ribonuclease H activity associated yeast RNA polymerase A.

The ribonuclease H activity associated with yeast RNA polymerase A degrades a variety of RNA-DNA hybrids with apparently no base specificity. The nuclease activity was characterized using (r-A), . (dT), and (r-G), . (dC), hybrids. It shows an absolute requirement for a divalent cation, Mg’+ or Mn2+. The pH curve is bimodal, with optima at pH 6.5 and 8. The optimal temperature depends strongly on the nature of the divalent cation. The activity is inhibited by low salt concentrations, EDTA, N-ethylmaleimide, and (ri),. The nuclease activity is also drastically reduced under conditions in which polymerization can proceed, even with limiting concentrations of ribonucleoside triphosphates. RNA polymerase A executes an exonucleolytic attack on the hybridized RNA, producing a mixture of a mononucleotide 5’-phosphate and of a dinucleotide with a 5’-phosphate end. The pattern of degradation products varies strongly with the incubation conditions. Depending on pH and divalent cation used, the dinucleotidelmononucleotide ratio can vary by a factor of 20 or more. During the course of these experiments it was found that Mn*+ ions, in absence of enzyme, catalyze the hydrolysis of (r-A), to a mixture of acid-soluble oligonucleotides of decreasing length, terminated with a 3’-phosphate group. After mild dissociation of RNA polymerase A with urea, the RNase H activity was recovered associated with the fractions containing both dissociated polypeptides, A48 and AZ4,: subunits, and with the RNA polymerase A* which is lacking these subunits. The RNase H activity again comigrated with A* enzyme upon polyacrylamide gel electrophoresis. Enzyme A* produced a mixture of dinucleotides and mononucleotides like Enzyme A, although in a significantly different ratio; and this ratio varied with the incubation conditions. The possibility of two distinct RNase H associated with the polymerase is discussed. The role of such ribonuclease H activity is still unclear, however, its presence in yeast RNA polymerase A supports the hypothesis that eukaryotic RNA polymerases are part of a multienzyme system involved in chromatin activity.

The ribonuclease H activity associated with yeast RNA polymerase A degrades a variety of RNA-DNA hybrids with apparently no base specificity. The nuclease activity was characterized using (r-A), . (dT), and (r-G), . (dC), hybrids. It shows an absolute requirement for a divalent cation, Mg'+ or Mn2+. The pH curve is bimodal, with optima at pH 6.5 and 8. The optimal temperature depends strongly on the nature of the divalent cation. The activity is inhibited by low salt concentrations, EDTA, N-ethylmaleimide, and (ri),. The nuclease activity is also drastically reduced under conditions in which polymerization can proceed, even with limiting concentrations of ribonucleoside triphosphates.

RNA polymerase
A executes an exonucleolytic attack on the hybridized RNA, producing a mixture of a mononucleotide 5'-phosphate and of a dinucleotide with a 5'-phosphate end. The pattern of degradation products varies strongly with the incubation conditions. Depending on pH and divalent cation used, the dinucleotidelmononucleotide ratio can vary by a factor of 20 or more. During the course of these experiments it was found that Mn*+ ions, in absence of enzyme, catalyze the hydrolysis of (r-A), to a mixture of acid-soluble oligonucleotides of decreasing length, terminated with a 3'-phosphate group. After mild dissociation of RNA polymerase A with urea, the RNase H activity was recovered associated with the fractions containing both dissociated polypeptides, A48 and AZ4,: subunits, and with the RNA polymerase A* which is lacking these subunits. The RNase H activity again comigrated with A* enzyme upon polyacrylamide gel electrophoresis.
Enzyme A* produced a mixture of dinucleotides and mononucleotides like Enzyme A, although in a significantly different ratio; and this ratio varied with the incubation conditions.
The possibility of two distinct RNase H associated with the polymerase is discussed. The role of such ribonuclease H activity is still unclear, however, its presence in yeast RNA polymerase A supports the hypothesis that eukaryotic RNA polymerases are part of a multienzyme system involved in chromatin activity.
Two distinct enzymes which hydrolyze ribonucleic acid hybridized to DNA have been described in yeast cells (1,2).
$ The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Materials
Enzymes -Yeast RNA polymerase A was prepared as previously described (7). The most purified fraction (phosphocellulose fraction) was used throughout this study. Homogeneous yeast RNA polymerase B (8) and Escherichia coli RNA polymerase (9) were prepared as previously described.

Nucleic
Acids and Nucleotides -Synthetic polymers (rfl,, (dTl,, and (dC) The unlabeled nucleotides used as markers and nucleoside triphosphates were obtained from P-L Biochemicals except for pAp, which was kindly synthesized for us by M. Blandin (Saclay). The 32P-labeled nucleoside triphosphates were purchased from Amersham; the 3H-and '%-labeled nucleoside triphosphates were obtained from C. E. N. Saclay. The melting curve of (rA), (dT), hybrid was carried out in a Cary 15 spectrophotometer equipped with a thermostatically controlled cell-housing. polymerized 13HlAMP; 60 to 80 cpm/pmol), 0.025 M Tris/HCl, pH 8, 5 mM MgCl, and 3 pg of RNA polymerase A. After 30 min at 37", 80 pg of serum albumin (40 ~1) was added as carrier, followed by 15 ~1 of 50% trichloroacetic acid; the precipitate was discarded in the cold and the radioactivity of acid-soluble material was measured by liquid scintillation in a toluene/Triton X-loo-based fluid. RNA polymerase was assayed in 0.1-ml mixtures containing 0.07 M Tris/HCl, pH 8, 1 rnM dithiothreitol, 5 mM MgCl,, 1 mM concentration each of ATP and [3H]UTP (25 cpmlpmol), 3 pg of d(A-T),, and enzyme. After 30 min at 30", radioactivity of acid-precipitable material was measured.

Analysis of Digestion Products by Thin Layer Chromatography
-Enzymatic hydrolysis was stopped by addition of EDTA and aliquots of the mixture were subjected to thin layer chromatography using the following systems. pppA, pAp, and pA (or Ap) were separated on PEI-cellulose (Merck; Catalogue No. F-5725) with 0.4 M potassium phosphate buffer, pH 4.5; 0.6 M phosphate was used to resolve pppG, pGp, and pG (or Gp). Potassium borate buffer (0.7 M), pH 9.2, separated pA from Ap and pAp on PEI-cellulose; similarly pG, Gp, and pGp were separated with 0.8 M borate. The borate buffer also allowed a good resolution of acid-soluble oligonucleotides of increasing length. When the samples contained glycerol or salts which disturb the migration, a preliminary washing of the plate was carried out with water. Then the thin layer plate was dried and the buffer was applied. After 2 h at room temperature it had migrated almost to the top of the plate. The system for two-dimensional chromatography was adapted to thin layer cellulose (Merck, Catalogue No. 5716) from the method of Pataki (11). The first dimension was developed with 1-propanol/NH,OH/H,O (6:3:1) for 5 h, the second dimension chromatography was for 2 h with 2-propanol/ saturated ammonium sulfate/H,0 (2:79:19). Two successive chromatography stages in the first direction with the same solvent improve the separation.
Polyacrylamide Gel Electrophoresis -The method of Laemmli (2) was used to analyze the proteins by dodecyl sulfate gel electrophoresis in a 12.5% acrylamide slab gel. Electrophoresis under nondenaturing conditions was performed similarly in a slab gel but without dodecyl sulfate. The RNase H activity was located in the gel by incubating l-mm slices in 0.15 ml of RNase H incubation mixture for 2 h at 37". The undegraded [3Hl(rA), (dT), hybrid was removed by acid precipitation and the acid-solubilized radioactivity determined as indicated above. In the latter case, the labeled DNA strand was found to remain undigested.

Some
The fact that all these substrates were used with similar efficiency suggested that the RNase H had a rather broad base specificity.

Effect of pH-The
RNase H activity was measured with (rA), . (dT), as substrate at different pH values, in Trislmaleate (pH 5 to 7) and Tris/HCl (above pH 7). The activity occurred within a very broad range of pH. Two pH optima were clearly found, near pH 6.5 and pH 8 (Fig. 2). As already shown in Fig. 1, the RNase H activity at pH 6.5 was about 50% lower than at pH 8. We checked that the single-stranded homopolymer (rA), resisted degradation under all these conditions (Fig. 2 Hybrid degradation is given in picomoles of total polymerized nucleotides rendered acid-soluble in O.l-ml sample. The asterisk indicates the radioactive polymer. MgCl, and 10 mM Tris buffer to minimize the variation in ionic strength due to the buffer. Effect of Ions -The divalent cation requirement for the RNase H activity could be fulfilled either by Mgz+ (5 mM) or Mn2+ (1 mM). The concentration response curves were essentially the same at the two pH optima, 6.5 or 8 (Fig. 3). In contrast with these results, which were obtained with (rA), . (dT), as substrate, the hybrid (rG), (dC), was degraded well only in the presence of Mn2+ ions, at a rather sharp optimal concentration around 1 mM. In the presence of Mg2+ Right panels, the effect of increasing concentration of salts, NaCl, or ammonium sulfate, was followed using the standard incubation mixture described under "Methods," at pH 8.1 (0-O) or 6.56 (0-O) as above. Hybrid degradation was measured as indicated under "Methods." presented in Fig. 3. The nuclease activity was practically maximal in the absence of salts and was inhibited by an increase in salt concentration above 0.01 M KC1 or ammonium sulfate. Again, this applied at pH 6.5 or pH 8.
Influence of Temperature -In view of the requirement of the enzyme for the hybrid structure (6), we investigated the effect of temperature on degradation of synthetic hybrids having very different thermal stability. Hydrolysis of (rG); (dC),, in the presence of Mn2+ ions, was optimal at 50" (Fig. 4). With (rA), * (dT),, the optimum temperature was dependent upon the nature of the divalent cation. With Mg2+, the rate of degradation was optimal at 37", whereas with Mn*+ ions the rate of reaction increased up to 50" and was still quite high at 60". This was surprising, since under these precise ionic conditions (with 1 mM Mn2+) the melting temperature of (rA), * (dT), was 53" (see Fig. 4); (with 5 mM Mg2+ the melting temperature of this polymer was 63" which was considerably higher). The explanation for this unexpected finding is that, above 50", Mn2+ ions catalyze an extensive degradation of (rA), polymer to acid-soluble products (Fig. 4). No such degradation of (rG), was observed up to 60" nor was any degradation of (rA), seen with Mg*+ ions. Similar observations had been previously reported (13). Inhibition -Concentrations of EDTA capable of binding all of the divalent cations present inhibited the activity completely. RNase H was also very sensitive to moderate ionic strength levels, as shown above (Fig. 3). Addition of a I-fold excess of (rA), strands over (dT), reduced the activity 70%. A I-fold excess of (dT), strand over (rA), inhibited the rate of reaction 85%. This inhibition was likely due to unproductive binding of the enzyme to the single-stranded polymer.
Addition of a lo-fold excess of (rf), during the course of the reaction immediately A control without enzyme was also included (m-m).
hybrid. This supports a nonprocessive mechanism of degradation. Nuclease activity was also completely inhibited by the -SH reagent N-ethylmaleimide at 10m3 M. The exact sensitivity of the enzyme to the inhibitor was not determined, because the storage buffer for the enzyme contained an -SH compound. The inhibitory effect of ribonucleoside triphosphates, which is related to the tight association of the nuclease activity with RNA polymerase, is described separately below. Inhibitory Effect ofRibonucleoside Triphosphates -We previously reported that addition of ATP prevented the degradation of the (rA), moiety of the (rA), . (dT), hybrid, whereas CTP, UTP, or GTP were ineffective (6). This observation was investigated in more detail, using two different hybrids, (rA), * (dT), or (rG), . (dC),, and the corresponding complementary ribonucleoside triphosphates, ATP or GTP, at varying concentrations.
In the experiment reported in Fig. 5, both the polymerization and the degradation reaction caused by RNA polymerase occurred simultaneously. As shown in Fig. 5 the presence of relatively low concentrations of ATP, in the range of 5.10m5 to 10m4 M, drastically reduced the rate of hydrolysis of hybridized (rA), strands. In this range of ATP concentration, the rate of the (dT),-directed polymerization reaction remained very low. The rate of polymerization increased steeply at high concentrations of ATP, up to 1O-3 M, but this was not followed by a further reduction in the degradation reaction which occurred at a significant background level. Basically, the same observations were made with the (rG), 1 (dC), polymer (Fig. 5), showing again that the polym- Synthesis of 13Hl(rG), in the presence of lSHIGTP (A----A). erization reaction was preponderant.
As expected in this case, only the complementary nucleotide GTP was inhibitory, but not ATP.

Mode of Action of Ribonuclease H Partial
Digest of (rG), . (dC), -In the above experiments, the degradation reaction was followed by monitoring the release of acid-soluble radioactive products. In the following, breakdown products were isolated by chromatography on polyethyleneimine cellulose sheets. Aliquots of the reaction mixture containing 32P-labeled (rG),* (dC), were withdrawn at various times and subjected to chromatography with phosphate buffer (Fig. 6). Three spots were found on the autoradiogram. The material remaining at the origin, corresponding to the undegraded polymer, progressively disappeared, and gave rise to two new components (Spots 1 and 2). From their R, values, these spots could correspond, respectively, to a mononucleotide and dinucleotide. There was no evidence of trailing of the radioactive material at the origin. To quantify these observations the compounds were eluted from the chromatogram with 0.1 N HCl and their radioactivity determined. The percentage of total radioactivity recovered in Spots 1 and 2 as a function of time is shown in Fig. 6. These breakdown products accounted for more than 80% of the hybrid degrada- Top, incubation mixture (0.5 ml) contained Ia2Pl(rG),.(dC), (80 pmol of CMP; 230 cpm/pmol), 21 pg of RNA polymerase A, 0.06 M Tris/HCl, pH 8, and 1 rnM MnCIZ. Incubation was at 37". At various times, 0, 2, 5, 10, 20, 30, and 40 min, two 20-~1 aliquots were withdrawn, one was used to measure acid-soluble radioactivity, the other was supplemented with 5 ~1 of 0.1 M EDTA and directly applied to a polyethyleneimine cellulose plats, which was developed with 0.6 M phosphate buffer as described under "Methods" and then subjected to radioautography.
The control (called C) on the left, corresponds to a sample of IS2Pl(rG),. (dC), incubated for 60 min at 37" in absence of enzyme. Bottom, the three spots, revealed by autoradiography, were scraped off, eluted with 0.1 M HCl, and their radioactivity determined by liquid scintillation. Results are given as per cent of total radioactivity recovered at the level of the different spots; percentage of total input radioactivity solubilized in acid (A-A); per cent radioactivity present in Spot 1 (O-O) and in spot 2 (0-O).
tion, as measured in parallel by acid precipitation. The ratio of radioactivity present in Spot 1 versus that in Spot 2 was 1:6. (A minor spot, faster than Spot 1, possibly free phosphate, was not taken into account.) Nature of Degradation Products -The degradation products from [32P](rG), . (dC), were identified by two-dimensional chromatography on thin layer cellulose plates (Fig. 7). The technique allows the resolution of a complex mixture of ribomononucleotides and related compounds (11). Co-chromatography with several unlabeled markers, pG, Gp, ppG, pGp, allowed the identification of Spot 1 as 5'-GMP. Left panels, [32Pl(rG), . (dC), was used as substrate. A, a sample of incubation mixture was directly spotted and analyzed in the presence of pG, Gp, and pGp as markers. B, Spot 2 from the above experiment (called pGpG in A) was scraped off, hydrolyzed with alkali, and subjected to the same analysis in the presence of pGp isomers as markers. Right panels, [32Pl(rA), . (dT), was used as substrate.
C, a sample of incubation mixture, was directly spotted on the thin layer plate together with pA, Ap, and pAp as markers. In D, enzymatic hydrolysis was followed by alkaline hydrolysis and chromatography with the above markers. With the exception of the spots called pGpG and pApA, for which no markers were available, all of the other radioactive spots revealed by autoradiography comigrated with the unlabeled markers which were located under UV light.
Spot 2 was inferred after alkaline hydrolysis.
The compound was eluted with 0.3 M KOH and treated with alkali for 16 h at 37". Upon alkaline hydrolysis Spot 2 disappeared and gave rise to two new spots which co-migrated with the 2'(3')guanosine diphosphate isomers of pGp. This indicated that Spot 2 corresponded to the dinucleotide pGpG, hydrolyzed to pGp with alkali.
Similarly, hydrolysis of [32Pl(rA), 9 (dT), by RNA polymerase A released two compounds corresponding to Y-AMP (PA) and to the dinucleotide pApA (Fig. 7). A minor spot with the migration behavior of adenosine was also found with 14Clabeled substrate (result not shown). Further proof for the identification of the products was provided after alkaline hydrolysis by chromatography of the digest on polyethyleneimine cellulose with 0.8 M borate buffer. The compounds cochromatographed perfectly with authentic pA and pAp. The rate of release of the two products was compared by chromatography.
The compounds were eluted and their radioactivity determined.
The molar ratio of pA versus pApA was found in the range of 1:3, which corresponded well to the molar ratio of pG versus pGpG found previously (once the appropriate corrections were made for their phosphate content) (see Fig. 6).
While in the course of these experiments we were interested to analyze the products of the hydrolysis of [32Pl(rA), . (dT), TABLE I Factors affecting pattern of degradation of (rA), (dT), ' Degradation of l'%l(rA),.(dT). by RNA polymerase A or A* (3 pg) was carried out as described in Fig. 3 either with Mg*+ (5 mM) at pH 8 or 6.56 or with Mn2+ (1 mm) at pH 8. After 30 min incubation at 37", aliquots of the mixture were subjected to chromatography on polyethyleneimine cellulose as described in Fig. 6. The spots of pA, of pApA, and of the undegraded material at the origin were located by autoradiography, and their radioactivity determined as indicated in Fig. 6. The results are given as percentage of total radioactivity recovered at the level of the two compounds. The molar ratio of pApA to pA was computed from the above data. The influence of the nature of the activator cation was most striking in this respect. With Mg2+ ions, at pH 8, the major degradation product of (rA), . (dT), was the dinucleotide pApA, whereas, with MnZ+ ions, the mononucleotide pA predominated slightly (Table I). A strong variation in the ratio of these two products was also found as a function of pH. At pH 6.56 and with Mg"+, only trace amounts of mononucleotide was formed, and the dinucleotidelmononucleotide ratio increased by a factor of 5 compared to pH 8.1 (Table I) In the initial report on the association of RNase H activity with RNA polymerase A, it was noted that some activity was also present at the level of RNA polymerase A* (6). However, RNA polymerase A* had a specific activity, which was lower than that of RNA polymerase A in terms of RNase H. This suggested that one or both polypeptides dissociated from RNA polymerase A (Subunits Adl( and A,,,s, of 48,000 and 34,500 daltons) were somehow involved in the nuclease activity.
To investigate this point further, RNA polymerase A was subjected to a mild dissociation treatment with 1.5 M urea in order to favor the formation of A* enzyme and accumulate the two polypeptides.
After the urea treatment the enzyme was applied to a phosphocellulose column; the proteins were eluted by a gradient of ammonium sulfate in absence of urea, and the fractions were assayed for RNA polymerase and RNase H activity (Fig. 8) (7)) was diluted to 0.17 mg/ml with a buffer containing 20 mM TrislHCl, pH 8, 0.01 M 2-mercaptoethanol, 0.5 mM EDTA, 0.05 M ammonium sulfate, 10% glycerol, and 1.5 M urea, at 0". The sample wss loaded onto a phosphocellulose column (10 cm x 1.25 cm*) equilibrated with the above buffer. The column was washed with 30 ml of buffer without urea and the proteins (A 280) (--) were eluted with 100 ml of a linear gradient from 0.05 to 0.4 M ammonium sulfate in the same buffer without urea. Fractions of 1.2 ml were collected and 20 ~1 was assayed for RNA polymerase (O-O) and RNase H (0-O) activities using the standard assays described under "Methods." Fractions 90 to 100 were first concentrated about 3-fold by dialysis at -20" against the same buffer as above containing 70% glycerol then assayed for RNase H activity (---1. Conductivity measurements indicated that Ehzyme A* was eluted at 0.19 M ammonium sulfate, Enzyme A at 0.25 M, and the Subunits & and A345 at 0.36 M ammonium sulfate. The fractions were pooled as indicated in the figure. Aliquots of the proteins pooled were precipitated with 15% trichloroacetic acid and analyzed by dodecyl sulfate electrophoresis on slab gel (inset); the numbers on the gel correspond to the number of the fractions pooled. elution was characteristic of these two forms of enzyme which were also identified by their subunit content (Fig. 8, inset).
The dissociated subunits were eluted tandemly at a higher salt concentration, as evidenced by sodium dodecyl sulfate gel electrophoresis.
RNase H activity co-chromatographed with Enzyme A, but also, despite the dissociation treatment, with Enzyme A*. Interestingly, the fractions containing the dissociated polypeptides were also found, after concentration, to have RNase H activity (Fig. 8) (7). One part of the gel was stained to locate the protein bands, and the other was sliced and used to assay RNase H activity (Fig. 9). The main band of protein migrated as RNA polymerase A*, ahead of a minor band of Enzyme A. After slicing the gel, RNase H activity was again clearly found at the level of A*. The dissociated polypeptides being basic proteins (14) do not enter the gel under these conditions.
RNA polymerase A* also produced a mixture of mononucleotides and dinucleotides, like Enzyme A, although in a significantly different ratio (see Table I). Less mononucleotide was produced compared to Enzyme A. Sometimes the mononucleotide spot was barely detectable. It can be seen in Table  I that the pattern of degradation  products varied with the  incubation conditions, as with A enzyme. 9. RNase H activity at the level of A* enzyme upon polyacrylamide gel electrophoresis. RNA polymerase A (15 pg; glycerol gradient step 7)) was subjected to electrophoresis on 5% polyacrylamide slab gel prepared essentially according to Laemmli (12) excepted that sodium dodecyl sulfate was omitted (nondenaturing conditions). Electrophoresis was run at 4", under constant current (20 mA/gel, 1 mm thick) for 3.5 h. One strip of gel was stained and used to locate the protein bands with Coomassie blue: under these conditions most of the RNA polymerase was recovered as A* enzyme. The Position of the protein bands is indicated in the figure. One adjacent strip was cut into l-mm slices which were incubated into 0.15 ml of RNase H assay mixture containing [3Hl(rA), . (dT), (330 pmol of AMP residues; 36 cpm/pmol) for 2 h at 37". Acid-solubilized radioactivity was determined as usual.

DISCUSSION
The persistence of an RNase H activity in homogeneous preparations of yeast RNA polymerase A is most interesting. This nuclease activity behaves as an exonuclease which produces a mixture of two different products, a 5'-phosphate mononucleotide and a dinucleotide with a 5'-phosphate end. The exonucleolytic mechanism of degradation is inferred from the fact that these products accounted for all of the acidsoluble radioactivity recovered. Oligonucleotides of increasing length were not observed. If the RNase H activity was that of an endonuclease, short oligonucleotides which cannot form stable hybrids at 50" or even at 37" would accumulate and never yield mononucleotides as final degradation products (15). The drastic and immediate inhibitory effect of a competitor nucleic acid, such as (ri),, when added during the course of degradation, was suggestive of a nonprocessive mechanism of degradation.
The nuclease appears to have a broad base specificity because it degraded the (r-G), strand of (r-G), . (dC), hybrids as well as the (rA), strand of (rA); (dT), polymers, provided that the appropriate activator cation was added. The RNA. DNA hybrid made with denatured T, DNA as template by yeast RNA polymerase B was degraded as well.