A guanosine 3':5'-monophosphate-sensitive nuclease from Bacillus brevis.

In toluene-treated cells of Bacillus brevis, newly synthesized RNA is rapidly degraded in a reaction that is inhibited by cyclic guanosine 3':5'-monophosphate (cGMP) and by 1,10-phenanthroline. This appears to be due to a ribonuclease found in cell-free extracts of B. brevis which is inhibited by cGMP and related compounds as well as by 1,10-phenanthroline. The cGMP-sensitive nuclease hydrolyzes synthetic polynucleotides, yielding nucleoside 5'-monophosphates as the sole products, even during the early stages of hydrolysis. Synthetic polynucleotides terminated by a 3'-phosphate are resistant to hydrolysis. While with 3'-hydrolysis of the polymer. The enzyme is therefore an exonuclease that degrades polynucleotides from the 3' end to product 5'-mononucleotides. It also acts on denatured but not on native DNA. Activity is greatest in the presence of Mn2+ and is not affected by the presence of monovalent cations. 1,10-Phenanthroline, but not 1,7-phenanthroline, inhibits the nuclease even when Mn2+ is present in excess. The inhibition of the enzyme by cGMP is noncompetitive, and cGMP itself is not hydrolyzed. The sensitivity of the nuclease to inhibition depends strikingly on the nature of the substrate and is lost when the enzyme is assayed at high pH. These observations suggest that cGMP inhibits the nuclease by combining with an allosteric site on the enzyme. Although cGMP was found to be the most effective inhibitor, other nucleoside 3':5'-monophosphates and derivatives of 5'-GMP can also inhibit the nuclease. Since measurements of cGMP in B. brevis have not revealed detectable amounts (less than 5 times 10-8 M), the substance that modulates the activity of the nuclease under physiological conditions remains to be identified.

newly synthesized RNA is rapidly degraded in a reaction that is inhibited by cyclic guanosine 3':5'-monophosphate (cGMP) and by l,lO-phenanthroline.
This appears to be due to a ribonuclease found in cell-free extracts of B. brevis which is inhibited by cGMP and related compounds as well as by 1, lo-phenanthroline.
The cGMP-sensitive nuclease hydrolyzes synthetic polynucleotides, yielding nucleoside 5'-monophosphates as the sole products, even during the early stages of hydrolysis. Synthetic polynucleotides terminated by a 3'-phosphate are resistant to hydrolysis, while with 3'-hydroxyl-terminated substrates the release of the 3'-terminal nucleotide occurs much more rapidly than the over-all hydrolysis of the polymer. The enzyme is therefore an exonuclease that degrades polynucleotides from the 3' end to produce 5'-mononucleotides.
It also acts on denatured but not on native DNA. Activity is greatest in the presence of Mn2+ and is not affected by the presence of monovalent cations. 1 ,lO-Phenanthroline, but not 1,7-phenanthroline, inhibits the nuclease even when Mn2+ is present in excess.
The inhibition of the enzyme by cGMP is noncompetitive, and cGMP itself is not hydrolyzed.
The sensitivity of the nuclease to inhibition depends strikingly on the nature of the substrate and is lost when the enzyme is assayed at high PH. These observations suggest that cGMP inhibits the nuclease by combining with an allosteric site on the enzyme. Although cGMP was found to be the most effective inhibitor, other nucleoside 3': 5'-monophosphates and derivatives of 5'-GMP can also inhibit the nuclease. Since measurements of cGMP in B. brevis have not revealed detectable amounts (less than 5 x 10e8~), the substance that modulates the activity of the nuclease under physiological conditions remains to be identified. and stored at -20". 7'dIrcorc 7'rcul;nc~,,t~Samples (j0 ml) were removed from growing cultures and rcntrifuged at 5000 X g for 5 min at 25". The cells were suspended in 0.5 ml of 0.1 i\~ triethanolamine hydrochloride, pH 7.5, and toluenc (5 ~1) was added. The mixtures were gently agitated at 25" for 10 min and then centrifuged at 5000 X g for 5 min at 25".
The sedimentcd material was suspended in 0.05 M triethanolamine hydrochloride, pH 7.5, containing 50% glycerol and stored at -20". RNA synthesis by cells treated with toluene was measured by the method described previously for purified RNA polymcrase (9). Degradation of newly synthesized RNA was estimated by arresting ItNA synthesis after 10 min by the addition of streptolydigin (2 pg per ml) and then continuing incubation at 37" for the appro-Driate time with or without cvclic nucleotides (2 mM). The rcac-&on was terminated by the a'bdition of 0.3 N tr'ichloioacetic acid, and the residual acid-insoluble radioactivity was compared with samples that had not been further incubated after the addition of streptolydigin.

RNA
Degradation in Toluene-treated Cells and in Cell-free Extracts-Cells of B. brevis, made permeable to small molecules by treatment with toluenc, were able to incorporate [W]UTP into RNA in a reaction that depended on all four nucleosidc triphosphates and that was inhibited by rifampicin and streptolydigin (1). When further RNA synthesis was prevented by the addition of streptolydigin, the radioactive product was found to be rapidly degraded.
As shown in Fig. 1, the half-life of newly synthesized RSA was 10 min. The addition of cG;\Il' reduced the rate of RXA degradat,ion to one-half, while cA1IJ' had a smaller effect. 1, lo-l'henanthrolinc (1.0 rnl\l) completely inhibited RNA degradation.
Cell-free extracts of B. brevis mediated the rapid hydrolysis of synthetic polyribonucleotides. Fig. 2 shows the effect of cGbl1' on the hydrolysis of [3B]poly(U) in the presence of Mn2+ and Mg2+ at pH 7.1. The reaction was inhibited by low concentrations of cGMl', with 85% inhibition at 1 mM. This indicated that most of the ribonuclease activity measured under these conditions was due to an enzyme which is sensitive to the cyclic nucleotide.

Bnzume
Purification-All ooerations were carried out at 4". 2. Inhibition of ribonuclease activity in crude extracts by cGMP.
Crude extract of B. brevis (16 pg of protein; 0.0 unit) was assayed under standard conditions at various concentrations of cGMP as described under "Experimental Procedure." were disrupted in a French pressure ccl1 at 8,000 p.s.i. The crude extract was centrifuged at 20,000 X g for 10 min and then at 150,000 x g for 4 hours:. The supernatant solution was subjected to gel filtration on a column (2.5 X 90 cm) containing 430 ml of Ijio-Gel A-0.5m equilibrated with UufTer A plus 1 M KH&l. .4 single symmetrical peak of ribonuclease activity emerged from the column after about 300 ml. The active fractions were pooled, dialyzed against IWfer A, and passed over a column containing 10 ml of I>I5AScellulose equilibrated with 13uffer A. Elution was carried out with a linear gradient generated by 200 ml each of Buffer A and Jluffer A plus 0.5 111 KCl. Ribonuclease activity appeared in the cfflurnt at about 0.12 1~ KCl. The pooled active fractions were dialyzed against Buffer A and applied to a column containing 10 ml of ECCTEOLA-cellulose equilibrated with Buffer A. The column was developed with a linear gradient produced by 200 ml each of ISuffer A and Buffer A plus 0.3 M KCl. The enzyme clutcd as a sharp peak at about 0.06 M KC1 (Fig. 3). One-third of the recovered ribonuclease activity (10 ml) was diluted with an equal volume of Buffer A and passed over a column of l)N'A-cellulose in this buffer at a flow rate of 2 ml per hour. The column was then washed with 5 ml of Buffer A containing 0.1 hi KC1 and the enzyme was eluted with Buffer A plus 0.6 M BCl. Protein (--) was determined spectrophotometrically (15). (i Only one-third of Fraction VI was processed. b Activity was low due to inhibition by DNA eluted from DNAcellulose. The results of the purification procedure are summarized in Table I. Since the I~~&xllulose cluatc (Fraction VII) contained some 1)X,4 which inhibits the hydrolysis of poly(U) (see below), its activity could not be measured accurately and may be considerably higher than indicated in Table I. '1 he total purification achieved was therefore at least 300.fold with a yield of 7 o/c or higher.
Fraction VII, when subjected to clcctrophoresis on polyacrylamide gels, revealed one major and a minor protein component (Fig. 4A). Upon polyacrylamide gel elcctrophoresis in the presence of sodium dodccyl sulfate, the major protein component appeared as a sharp doublet band whose molecular weight, estimated by comparison with the electrophorctic mobility of standard proteins, was 66,000 (Fig. 4B). '1 he purified enzyme was relatively unstable.
n'hen stored in Buffer A at -lo", one-half of the cnzymic activity was lost in about 2 weeks.
The ribonuclease was purified on the basis of its ability to hydrolyze poly(U).
However, as will be discussed below, the purified enzyme was also able to hydrolyze other polynucleotides such as poly(A) and denatured LIKA, albeit at a considerably

(right). Estimation of molecular weight of nuclease by gel filtration.
Crude extract of R. brevis (14 mg of protein), P-galactosidase (240 wg), and glycerol kinase (60 pg) in 0.5 ml of Buffer A plus 1 M NHaCl was subjected to gel filtration on a column (2.5 X 45 cm) of Bio-Gel A-0.5m as described in the text. Fractions of 5 ml were collected and assayed for enzyme activity.
lower rate. The ratio of the capacities to hydrolyze poly(U) and poly(A) was relatively constant throughout the purification procedure (Table 1). Furthermore, the activities which hydrolyzed polyU, both at pH 7.1 and 10.6, and denatured DNA eluted in a common peak from ECTEOLA-cellulose (Fig. 3), suggesting that a single enzyme acted on all these substrates.

Metal
Ion Requirement-As shown in Table II, the hydrolysis of synthetic polynucleotides required the presence of hIn2+, which could be only partially replaced by hIg2f.
Nuclease activity was inhibited by 1 , 10.phenanthroline at concent.rations considerably lower than that of ;\Inz+ present in the incubation mixture, while 1,7-phenanthroline was without effect. It should be noted that the hydrolysis of poly(U) was somewhat less sensitive to the chelator than the hydrolysis of other substrates. Thus, while 1 , 10.phenanthroline (0.5 mM) inhibited the hydrolysis of poly(A,C), poly(A), and denatured I)n'A each by 70%, the rate of poly(U) hydrolysis was reduced by only 35%.
The addition of KCl, NaCl, and NH&l (0.1 M) had no effect on the activity of the nuclease.
Mode of Action-When polyuridylic acid was hydrolyzed with an excess of enzyme, the sole reaction product was 5'.UMP (Fig.  6). Similarly, the only radioactive product obtained with poly-(A,C) lab&d with 14C in the adenine moiety was 5'.A1:II' (not shown).
Even under conditions of incomplete hydrolysis (5 to lo%), nucleosidc 5'.monophosphates represented more than 90y0 of the hydrolysis products of poly(U) and poly(A,C). Table III shows the action of the nucleasc on (Ap)zUp and  Fig. 7. The hydrolysis of the Wlabeled 3'.hydroxyl end was considerably more rapid than the over-all hydrolysis of the polynucleotide, measured by the solubilization of the %labeled residues.
Substrate Specificity--Resides poly(U), the purified nuclease preparation could hydrolyze a variety of polynucleotides (Table  IV).
The observed rates of hydrolysis of poly(A) and denatured DNA were considerably lower than those of poly(U) and poly-(A,C).
However, the concentrations of the polynucleotides used were probably not saturating; moreover, the rate of exonucleolytic degradation will be a function of the concentration of 3' termini rather than of nucleotide residues.
Consequently, since the average chain lengths of the various polynucleotides had not been determined, it is not possible to evaluate their relative effectiveness as substrates.
Nevertheless, the rates of hydrolysis of denatured and native DNA can be directly compared since they come from the same DNA preparation, and it is clear that denatured I>KA is hydrolyzed much more rapidly. Table V shows the effect of unlabeled polynucleotides on the hydrolysis of labclcd substrates.
An equivalent amount of poly(A) completely inhibited the hydrolysis of [3H]poly(U), suggesting that the poly(U)-poly(A) complex is resistant to hydrolysis. Denatured, but not native DiYA inhibited the hydrolysis of polyribonuclcotides while poly(U) inhibited the hydrolysis of denatured DNii, supporting the idea that ribonuclease and deoxyribonucleasc activities arc attributes of a single enzyme. It should be noted that ribonuclease activity was inhibited at a very low DNA : RNA ratio whereas a considerable excess of RNA was required to inhibit DNA hydrolysis, suggesting that the The hydrolysis by nuclease (Fraction VI) of the substrates indicated was assayed in the presence of unlabeled polynucleotides as described under "Experimental Procedure," except that 0.3 N trichloroacetic acid was used as the precipitant when DNA was the substrate.
Unlabeled DNA was derived from salmon sperm raphy on polyethyleneimine-impregnated cellulose in 0.5 hr LiCl, 99% of the recovered radioactivity was found to migrate with cGLI1' while less than 0.1 'i;o was found to be associated with the position of the 5'.GlIl' standard.
In contrast, the hydrolysis of other polynucleotides was much less sensitive to inhibition by cGRI1'. At 1 m&r, cGJ\II' inhibited the hydrolysis of poly(A,C) by only 507, and that of poly(A) and denatured l)NA hardly at all. cGXI1' reduced the V max of the nuclease without affecting the apparent K, for the polynucleotidc substrate. This is illustrated in Fig. 9 for poly(A,C); similar noncompctitivc kinetic data were also observed with poly(U). Table VI compares the effects of various nucleotides on the hydrolysis of poly(U).
Neither guanosine 3'.diphosphatc-5'.diphosphate nor the reaction product 5'-ULIl' (not shown) had an effect on poly(U) hydrolysis. 10. Effect of pH on ribonuclease activity in the absence and presence of cGMP. Nuclcase (1.4 kinits of Fraction VI) was assayed under standard conditions except that the pII was varied as indicated, without (closed symbols) and with (oprt~ symbols) 0.05 rnM cGMP. The buffer systems used wcrc 20 mM Tris maleate (0,0) or 20 mM ethanolamine hydrochloride (W, 0).
EJeecl of pll-The hydrolysis of poly(U) by the nuclease of B. brevis occurred over a broad range of pI1, with more than halfmaximal activity between pl1 7.1 and 10.6 ( Fig. 10). However, the sensitivity to inhibition by cG111' declined in the higher pH range, so that the pH optimum was shifted upward in the presence of the cyclic nucleotidc The nuclcase activity observed at high pH did not seem to be due to a different enzyme, since it showed the same substrate specificity as that observed at pH 7.1 and produced 5'.mononuclcotides as the only product (not shown).
Nevertheless, in order to assure maximum sensitivity to inhibition by cGRIP, most of our studies were done at pH 7.1, where the enzyme exhibited about half-maximal activity.
Levels of Cyclic Nucleotides i n B. brevis-We made attempts to measure the intracellular levels of cGhI1' and cAM1' in samples of early and late exponential as well as early and late sporulating cultures of B. brevis.
Although cyclic nucleotidcs added as internal standards were recovered in good yield, no cndogcnous cGhI1' and cA111' was found.
The sensitivity of the radioimmunoassay used was such that intracellular concentrations of 5 X lop8 nr cG111' and 5 x 1O-7 hr &fill' would have been detected.

DISCUSSION
Although the turnover of mRNA in bacteria has been widely studied, little is known about the factors that regulate this process. It was therefore of considerable intcrcst that the degradation of newly synthesized RNA in tolucne-treated cells of B. brcvis was markedly inhibited by the addition of cGilIl', suggesting the presence of a nucleasc whose activity is regulated by the cyclic nucleotidc.
This was confirmed by the observation that rnost of the ribonucleasc activity in cell-free extracts of B. brevis could be inhibited by cGL\il' (Fig. 2). Moreover, RNA degradat.ion in toluene-treated cells and the ribonucleasc in cell-free extracts were similarly inhibited by 1, IO-phenanthrolinc.
After partial purification, the cGXl'-scnsitivc enzyme could be characterized as an exonuclease because it produced ribonuclcosidc 5'-monophosphatcs as the sole products even during the very early stages of hydrolysis of synthetic polynucleotides (Fig. 4). Moreover, polynuclcotides bearing a 3'.phosphate were relatively resistant to the action of the enzyme (Table III). Their slow hydrolysis did not lead to the rclcase of the 3'.terminal nucleotide and may have been due to a small amount of endo-nuclease contaminating our partially purified enzyme preparation. Studies on the time course of hydrolysis of the double labeled polynuclcotide (Ap)fiU showed a much more rapid solubilization of the 3'-terminal uridylate than of the adenylate residues (Fig. 7)) a result consistent with random exonucleolytic degradation from the 3' end. The purified exonuclease preparation from B. brevis could hydrolyze a variety of single-stranded polynucleotides. Several lines of evidence suggest that the action on the various substrates is the manifestation of a single enzyme.
(a) The ratio of the capacities to hydrolyze poly(U) and poly(A) was relatively constant during the purification procedure (Table I) ; (b) the activities that hydrolyzed poly(U) and denatured DNA eluted in a common peak from IXTEOLA-cellulose (Fig. 3); (c) the ribonuclcase activity was adsorbed to denatured DNA bound to cellulose; (d) the hydrolysis of poly(U) and poly(A) was inhibited by denatured DNA, while the hydrolysis of DNA was inhibited by polyribonucleotides (Table V); and (e) all activities were inhibited to some extent by cG1IP and 1 , 10.phenanthroline, even though the relative sensitivities to these agents did depend on the nature of the substrate.
A striking property of the nuclease from B. brevis is its sensitivity to inhibition by cGI\IP. Two possible mechanisms could account for this phenomenon.
The first of these is the direct combination of the cyclic nucleotide with the active site of the enzyme, either as a substrate or as a substrate analog. This possibility is unlikely, since no hydrolysis of cGXP by the enzyme could be detected and since the inhibition was strictly noncompetitive (Fig. 9). hloreover, specificity studies showed that nuclcotides not containing a phosphodiester linkage, such as 5'-G,\Il' and guanosine 3', 5'.bisphosphate, were also somewhat inhibitory, and that cU1\11' and cC~IP were weaker inhibitors than cGXI', in spite of the fact that the enzyme preferred to hydrolyze polynucleotides containing pyrimidine residues. We must therefore consider a second possibility, namely that the inhibitors do not bind to the active site of the enzyme but to an allosteric site. Strictly speaking, the occurrence of noncompetitive inhibition is sufficient evidence for the existence of an allosteric site, but, ordinarily, the dissociation of catalytic activity from inhibition is regarded as a stronger criterion.
Such a phenomcnon was indeed observed at high pH, where the hydrolysis of poly(U) still proceeded at a relatively rapid rate but was insensitive to inhibition bycGM1' (Fig. 10). Since the ribonuclease activities assayed at pH 7.1 and 10.6 eluted in a single peak from ECTEOLA-cellulose (Fig. 3) and the catalytic properties at pH 7.1 and 10.6 were very similar, it is unlikely that these activities are due to two different enzymes. Rather, it appears that at high pH the enzyme has selectively lost the ability to respond to allosteric inhibition.
Rforeover, the differences in sensitivity to inhibition by cGX1 of the hydrolysis of different substrates (Fig.  8) can be more readily understood in terms of an allosteric mechanism than in terms of binding of the cyclic nucleotide to the active site. Finally, it should be noted that the molecular weight of the nuclease from B. brevis, 360,000, is very much higher than that of other ribonucleases known and suggests that the enzyme is composed of subunits, a property characteristic also of allosteric enzymes.
The physiological function of the exonuclease of B. brevis is still uncertain, but the experiments with toluene-treated cells suggest that the enzyme may be involved in the degradation of mRNA. An important question concerns therefore the nature of the substance that inhibits the ribonuclease under physiological conditions.
Our analyses revealed that the intracellular levels of cGllH' and CAMP in B. brevis are at least several orders of magnitude lower than the concentrations that produce 50% inhibition of the ribonuclease.
Other workers have also been unable to detect CAMP in either Bacillus lichenijormis (17) or Bacillus megaterium (18), while the level of cGM1' in B. lichenijormis has been found to be less than 8 x 10m9 M (17). Therefore, it is unlikely that cGhIP is the agent that modulates the activity of the ribonuclease in vivo, in spite of the fact that it was the most inhibitory of the substances that we have examined with the purified enzyme. A naturally occurring inhibitor of the nuclease of B. brevis, perhaps structurally related to cGMP, thus remains to be identified.