Purification and Properties of a Specific Escherichia coli Ribonuclease which Cleaves a Tyrosine Transfer Ribonucleic Acid Precursor

SUMMARY Precursor molecules of Escherichia coli wild type and mutant tyrosine tRNA’s contain at both their 5’ and 3’ termini extra nucleotides in addition to those of the mature tRNA molecule. The early steps of processing these precursor molecules must involve specific ribonuclease cleavage. We report the isolation from E. coli extracts of the specific endonucleolytic RNase which cleaves only a single phosphodiester bond of the 129 nucleotide tyrosine tRNA precursor molecule. This cleavage removes all extra nucleotides present at the 5’ terminus of the precursor as a 41 nucleotide fragment, exposing the 5’ end of the mature tRNA. After sufficient purification, this activity has no effect upon the extra nucleotides at the 3’ end of the tRNA precursor. Therefore processing of the two ends of this molecule must be carried out by different enzymatic activities. This novel RNase activity, which we have called RNase P, has been purified by washing ribosomes with 0.2 M NH&l, followed by ammonium sulfate fractionation and chromatography on DEAE-Sephadex and phosphocellulose. At this stage it shows no evidence of other E. coli RNase activities. RNase P requires both monovalent and divalent


molecules
of Escherichia coli wild type and mutant tyrosine tRNA's contain at both their 5' and 3' termini extra nucleotides in addition to those of the mature tRNA molecule.
The early steps of processing these precursor molecules must involve specific ribonuclease cleavage. We report the isolation from E. coli extracts of the specific endonucleolytic RNase which cleaves only a single phosphodiester bond of the 129 nucleotide tyrosine tRNA precursor molecule.
This cleavage removes all extra nucleotides present at the 5' terminus of the precursor as a 41 nucleotide fragment, exposing the 5' end of the mature tRNA.
After sufficient purification, this activity has no effect upon the extra nucleotides at the 3' end of the tRNA precursor.
Therefore processing of the two ends of this molecule must be carried out by different enzymatic activities. This novel RNase activity, which we have called RNase P, has been purified by washing ribosomes with 0.2 M NH&l, followed by ammonium sulfate fractionation and chromatography on DEAE-Sephadex and phosphocellulose. At this stage it shows no evidence of other E. coli RNase activities. RNase P requires both monovalent and divalent cations for optimal activity, and has a pH optimum of 8 It is evident that this precursor molecule must be cleaved in viva in order to give rise to the functional tRNA. Such processing of the tRNA New Haven, Connecticut.
precursor by crude S 30 extracts of E. coli has recently been reported by 9ltman and Smith (2), who showed that the extra segments at both the 5' and 3' ends of the precursor could be removed under proper conditions. One hypothesis to explain these results is that specific endonucleases exist in I?. coli including one which could cleave the phosphodiester bond immediately adjacent to the normal 5' end of the tyrosine tRNA sequence, yielding the mature 5' terminus and a 41 nucleotide fragment. However, such a hypothetical 41 nucleotide fragment was not recovered after digestion with crude extracts, although variable amounts of a smaller fragment containing the 22 to 23 bases nearest the 5' end of the precursor were observed (2).
In order to see whether any of these previously described nucleases are involved in processing the tRNA precursor, as well as t.o characterize further the activity or activities responsible for this processing, we have purified this activity.
We have found that more than one ribonuclease activity is involved in the processing, and we describe the extensive purification of one of these activities.
This enzyme turns out to be a new ribonuclease specific for the cleavage of only a single phosphodiester bond in the entire tRNA precursor molecule. The precursor, and other RNA bands on polyacrylamide gels, were eluted as follows.
The region depicting the desired area was cut from an exposed x-ray film, which was used as a template for excising t,he desired regions of the 20 x 40 cm slab gel. The gel band was mechanically homogenized aud eluted in 0.1 M Tris-TICl, pH 9.1, 0.5 RI NaCl, 0.01 M EDTA. Equal volumes (1 to 5 ml) of disrupted polyacrylamide suspended in the above buffer and water-saturated phenol were homogenized mechanically, centrifuged for 10 min at 10,000 x g, and the aqueous layer taken; the other phase was re-extracted with an equal volume of the above buffer by st,irring, and the pooled aqueous phases were filtered through a 0.45.nm Millipore filter. One ABe unit of E. coli tRNA was added per ml of aqueous phase, and the RNA was precipitated with 2.5 volumes of absolute ethanol at -2O", centrifuged, and the pellet collected and resuspended in 1 ml of 0.2 M sodium acetate, pH 5.5. Rn'A was precipitated again with 2.5 volumes of absolute ethanol at -20". After centrifugation, the resulting precipitate was lyophilized and resuspended in 0.1 ml of distilled water. Specific activity of the precursor was 0.5 to 2 X lo6 cpm per pg.
Other Polynucleotides-32P-labeled mixed tRNA was prepared from the same cultures as the tRN.1 precursor and purified in a similar fashion. f2 phage RNA was grown and purified according to Dahlberg (13).
[3H]poly(AIY) copolymer, specific activity 24.3 PC1 per pmole, was that used by Robertson et al. (11). Poly(G) Poly ( were added, and 10 pg per ml of pancreatic DNaae were added to the resultant slurry. After 30 min at 4", the mixture was centrifuged for 10 min at 8,000 rpm in 12-ml glass centrifuge tubes in the SS34 rotor of the Sorvall RCS-B centrifuge. The supernatant was then centrifuged for 40 min at 15,500 rpm as above; the resulting 30,000 x g supernatant is called S 30 and was that in which precursor cleaving activity was previously detected (2). R.ibosomes were prepared from this S 30 suprrnatant by centrifugation for 4 hours at 45,000 rpm in the type 65 rotor of the Reckman model L ultracentrifuge.
The upper two-thirds of the resulting S 100 supernatant was removed with a Pasteur pipette, and the rest was discarded.
The ribosomal pellet was rinsed with 2 ml of l3uffer A, which was discarded, and the pellet was then redissolved in 2 ml of fresh Isuffer A. After removal of an aliquot of resuspended ribosomes for assays, the ribosomes were washed with the desired concentration of NH,Cl as follows. An appropriate arnount of 4 M NH&I in Buffer A was added, aiid the mixture was transferred to a Beckman cellulose nitrate ceiitrifuge tube (3 X 2 inch).
This tube was attached to the cup of a varinble-speed Vortex mixer (Lab-Line Instruments) with vinyl tape and allowed t,o agitate gently overnight at 4". The volume leas increased to 5 ml with Buffer A containing the appropriate NH,CI concentration, and the mixture was centrifuged for 4 hours in the SW39 or SW50 rotor of the Beckman model L ultracentrifuge at 37,000 rpm.
The upper two-thirds of the resulting superna tant was removed and retained, while the rest was discarded. The ribosomes were again resuspended in Buffer A, an aliquot, removed, and a further washing st,ep initiated. Protein con centrations were determined by the procedure of Lowry et al. (18).
Determination of Radioactivity--Samples dried on glass fiber filters or on paper were assayed for radioactivity using the to- tive analysis of the kinetics of tRKA precursor reactions \~a:: performed after cutting out the appropriate bands from the gel as described above.
The radioactivity in each sample was assayed by placing it in an empty vial and measuring the Cerenkov radiation in a scintillation spectrometer. The efficiency of detectiou of z21' under optimal settings was 950/, using scintillation fluid and 2Ocl, for Cerenkov radiation.
Centrifugation was carried out for 10 hours in the SW50 rotor of the Beckman L265B ultracentrifuge at a t)emperature of 5".

Partial
Purification of a Specific Precursor-cleaving Activity-We atternpted to devise a purification procedure which avoided harsh treatment of subcellular components for as long as possible. In particular we attempted not to disrupt ribosomes. This ap-Irroach was used successfully by Robertson et al. in their purification of E. coli RNase III (11) and was also an alternative purification of JG:. coli RKase IT suggested by Spahr (19).
An S 30 extract from 5 g of E. coli lUREGO was prepared as described under "Experimental Procedure," and the ribosomes were isolated and washed as described under "Experimental Procedure." The resuspended ribosornes and the supernatants, after washing at various NH&l coucentrations, were assayed for their ability to cleave t,he tRNh precursor. Fig. 1 shows the effect of various subcellular fractions upon the tyrosine tRNA precursor. The large change in mobility of the precursor upon specific cleavage is almost entirely due to the removal of the 41 rrucleotides located to the 5' side of the tRNA sequence in the precursor, since the 5'-terminal region accounts for 41 of the 44 extra uucleotides in the precursor. A comparison of Lanes I and 14 shows that the 8 30 extract contains an activity which cleaves the precursor to yield a major product with the mobility of tRNA, as noted by Altman and Srnith (2) In light of the results shown in Fig. 1, it is likely that the activity eluted in 0.1 M NH&l corresponds to the latent nonspecific RiYase cornponents, the rnajority of which were not removed from ribosomes until they were washed in 0.5 M NH4C1. Activity present in fract,ions eluted from DEAE-Sephades in 0.5 RI NIIdCl could be conceiitrated by addition of 0.6 g of ammonium sulfate per ml of enzyme solution, followed by centrifugation as described in the legend to Fig. 2, and  A l-ml aliquot was further analyzed on a phosphocellulose column 1 cm in height prepared in a Pasteur pipette.
Stepwise elution of this column was carried out exactly as described for the DEAE-Sephadex step, using Buffer B containing 5% sucrose and the NH&l concentrations indicated in Fig. 3. We find that all of the activity is recovered in the flowthrough of the column (Buffer B containing 0.02 M NH&l). At this point the amount of protein in the active fractions is not detectable.
However, an estimate of the minimum extent of purification through the DEAE-Sephadex step is presented below.
Specijicity of Precursor-cleaving Activity-We have previously shown that crude E. coli extracts split the precursor to give the tRNA sequence (with partial loss of the extra nucleotides at the 3' end) a fragment comprising the first 22 to 23 nucleotides from the precursor 5' end, and mono-and dinucleotides (2). These products could have resulted from a specific single cleavage splitting off the 41 nucleotide fragment which was subsequently partly degraded by other enzymes in the crude extracts. This interpretation has now been shown to be correct. Fig. 4a shows that the more highly purified enzyme gives only two major RNA products on polyacrylamide gel electrophoresis. One ("tRNA") migrates slightly behind mature tRNA, while the second (5' fragment) moves more slowly than did the 22 nucleo-nucleotides from the precursor segment are: 1, Gp; 2, ApGp; 3, CPAPGP; 4, CPCPAPGP; 5, APUPAPAPGP; 6, UPAPAPAP-APGP; 7, CPUPUPCPCPCPGP; 8, CPAPUPUPAPCPCPCPGP. Separation is by electrophoresis on cellulose acetate in pyridine acetate, 7 M urea, pH 3.5, from right to left; and on DEAE paper in 7y0 formic acid (v/v) from top to bottom. tide fragment described before (2). The products of digestion of these two bands with Tr and pancreatic ribonucleases were examined; Fig. 4, b and c, shows the separation of the Tr ribonuclease products.
The "tRNA" band contains the entire tRNA sequence from the terminal pGpG. . . , and includes the extra nucleotides at the 3' end of the precursor. Partial loss of these nucleotides results in an additional 3'Qerminal T1 ribonuclease product which is present when the crude extract's are used for processing but which is absent in Fig. 4b.
The digestion products of the 5' fragment ( Fig. 4c) are those expected from the complete 41 nucleotide 5' segment. GpUp is absent from the pancreatic ribonuclease products obtained with this fragment; instead an additional nucleotide migrating in the position of GpU is found.
An alkaline hydrolysis of this gave Gp, and it is tentatively identified as GpU. This pancreatic ribonuclease product would be expected if the specific RNase splits the 3' phosphodiester bond of the last nucleotide in the precursor 5' segment.
These results indicate that the enzymatic activity which we are studying here has a single simple mode of action on the precursor; we designate this activity ribonuclea,se P. We can further conclude that RNase P purified to this stage is already free of the activity or activities which degrade the 5'-terminal fragment and remove the extra nucleotides from the 3' end of the precursor. The sequence of tyrosine tRNA precursor and its cleavage point by RNase P are shown in Fig. 5 Assay conditions were those described under "Experimental Procedure." All substrates were used at the same specific radioactivity and were prepared from pulse-labeled cells as described previously (I, 2) and under "Experimental Procedure." The assays depicted in Lanes 1 to 7 contained the 0.5 M ribosomal wash activity shown in Fig. 1, Lane 6, and contained 0.025 mg of protein per ml. Reactions shown in Lanes 8 to 10 contained RNase P activity in the 0.2 M ribosomal wash (Fig. 1, Lane Q), and the final protein concentration was 0.08 mg per ml.
No enzyme was added to the reaction shown in Lane 11. Potential substrates and other additions to the assays were as follows. Properties of RNase P--We have studied the kinetics of the cleavage reaction by RNase P, both in crude extracts and after purification through the DEAE-Sephadex step, by quantifying the release of specific RNA products with time.
In this way, we hope not only to find whether the properties of the reaction are altered by the removal of components during the purification, but also to estimate the minimum extent of purification. Results obtained with the more highly purified enzyme fraction, shown in Fig. 64, indicate that the recovery of products accounts quantitatively for the loss of tRNA precursor.
In addition, the ratio of radioactivity recovered in tRNA to that in the 5' fragment at any given time is close to that expected for two RNA's of identical specific activity which have their size ratio (88 to 41 nucleotides). This result is one indication that the properties of the RNase P cleavage reaction have not been grossly altered by the purification steps. Furthermore, 154 pg of S 30 protein were added to one reaction shown in Fig. 6, while less than 1 pg of protein from the RNase P DEAE-fraction was added to the other.
Since the initial rate of tRNA precursor cleavage by S 30 is 2.8 times that of the DEAE-fraction as determined from the data on which this comparison is based, we conclude that the enzyme in this fraction has been purified more than 50.fold, but the real figure is undoubtedly higher. The effect of ionic conditions on the RNase P reaction has been surveyed using the same assay system already described for the purification.
We have tested various salt and pH conditions with the intention of comparing the RNase P cleavage reaction with those carried out by other E. coli RNases.
For this reason, and because of the cumbersome nature of the standard assay, we have only screened certain carefully chosen sets of conditions. The results which we have obtained for a purified RNase P preparation, summarized in for RNase P at various stages of the purification is in the neighborhood of 8.0, with lower activity observed both at pH 7.0 and pH 8.5.
The size of RNase P was estimated by a velocity sedimentation through a 7 to 257, sucrose density gradient in the absence of NH&!1 as described under "Experimental Procedure." Purified catalase and hemoglobin were centrifuged in a separate tube as markers.
In some experiments low levels of activity were recovered moving slightly ahead of catalase, which has a reported  (12) also had no RNase Plike activit,y.
The specific endonucleolytic mode of action of RNase I' rules out its identity with the major exonucleolytic activity associated with F. coli RNase II (4,19,21) or with polynucleotide phosphorylase (8). In order to test whether RNase III of E. coli is also responsible for the RNase P activity, we incubated an excess of highly purified RNase III (11) with the tyrosine tRNA precursor mlder optimal conditions for both enzymes (which are almost exactly the same).
We found that RNase III has no effect on the electrophoretic mobility of the tRNA precursor.
An RNase activity associated with ribosomes which has been called RNase V (7) has been identified with an activity which degrades messenger RKA's in an exonucleolytic fashion from their 5' ends. This process apparently depends upon the state of activity of the ribosomes, and our RNase P preparations do not' appear to have such an acbivity. We conclude that RNase P is a novel RNase activity of E. coli. This enzyme may cleave all l?. coli tRNA precursors or only the one for tyrosine tRSA.
The choice of one of these alternat'ives will await the isolation of additional E. coli tRNA precursors. The subcellular fractionation of RNase P shown in Fig. 1 dem and also renders it soluble in .5c/; trichloroacetic acid as deter-onstrates that processing of the two ends of the tRNA precursor is carried out by different enzymatic activities. This finding is confirmed by the fingerprinting data shown in Fig. 4. The results shown in Fig. 1 also show that RNase P shares with several other important E. coli proteins the property of loose association with ribosomes within crude extracts capable of biologically important reactions. For example, RNase III (11) and the various factors required for the initiation of protein synthesis (24) show behavior very similar to that of RNase P with regard to their removal from ribosomes by NH&l. While association with ribosomes in estracts does not necessarily mean that these enzymes and factors are located upon these particles in viuo, analysis of the behavior of the other E. coli RNases during such gentle subcellular fractionation as that described here might reveal other imeresting functional associations within the cell. The behavior of RNase P upon ammonium sulfate fractionation (Fig. 2) and chromatography upon DEAE-Sephadex (Fig.  3) and phosphocellulose is different from that of the bulk of E. coli proteins and has greatly facilitated our purification of this enzyme.
In light of these properties, it is possible that the active form of RNase P, which must have a strong negative charge, could be associated with some nucleic acid.
In light of the specific cleavage of a single phosphodiester bond within the 129 nucleotide tRNA precursor, in contrast to the behavior of other RNases, we conclude that RXase P is the first, ribonuclease to be described which has such a high degree of specificity.
Several examples of specific DNases have been reported (25)(26)(27). The fact that RNase P creates a 5' phosphate end group also makes it unique among E. coli endonucleolytic RNases so far characterized.
Altman and Smith (2) suggested on the basis of studying various mutated tyrosine tRKA precursors that baxe changes both near and far away from the point of cleavage may affect the rate of cleavage.
Those studies as well as the ones reported here suggest that RNA secondary and tertiary structure may be as important as sequence in the action of RNase P. -4n assessment of the relative importance of these substrate properties in determining RNase P specificity could also be important in a more general study of RNA to protein interactions.
Although we have only performed a few experiments on the additional RXase activities present in 0.5 M NH&l ribosomal washes (Figs. 7 and 8), their results are worthy of some comment. From their studies of mutant tRNA's, Altman and Smith (2) proposed the existence of a degradative pathway for tRNA precursor other than the one leading to mature tRNA. RNase activity present in 0.5 M NH&l ribosomal wash fractions which can degrade tyrosine tRNA precursor but not mature tRNA (Figs. 7 and 8) may correspond to this proposed scavenger enzyme.
In addition, the activity or activities responsible for degrading the 5'.terminal fragment of the tyrosine tRNA precursor in less pure RNase P fractions, as well as that which proc 5251 esses the 3' end of the precursor molecule (see Figs. 1 and 2), may also be present in this 0.5 M NH&l wash. An interesting feature of this RNase activity is its latent nature.
The existence of such latent RNases associated with ribosomes might help to account for reports of R1Jase activities which are dependent upon ribosomal configuration or protein synthetic activity (7). Finally, we conclude that the use of natural substrates such as tRNA precursors as one aspect of the rigorous characterization of E. coli ribonucleases should be instrumental in revealing new aspects of the regulation of RNA metabolism.