Purification and Characterization of an Endoribonuclease from Nucleoplasm and Nucleoli of HeLa Cells*

An endoribonuclease has been isolated from HeLa cell nuclei. Approximately 70% of the enzyme appears to be nucleolar bound; 30% is in the nucleoplasm. Studies of the purified enzyme reveal that the enzyme is an endonuclease of estimated molecular weight 16,000. It produces oligonucleotides bearing 5’-phosphate end groups. The enzyme degrades poly(C) and poly(U), as well as rRNA and heterogeneous nuclear RNA. Poly(A), double-stranded RNA, and DNA are not cleaved. The enzyme is heat-labile and is inhibited by 10 mM Mg*+ and 50 mM NaCl. The enzyme is probably distinct from previously described nuclear endonucleases. The ribosomal RNA, messenger RNA, and transfer RNA of eukaryotic cells are transcribed as larger precursor molecules which are then cleaved by ribonucleases into smaller mature products (l-4). A number of ribonucleases have been found in the nuclei of animal cells. Winicov enzymological rRNA’

The ribosomal RNA, messenger RNA, and transfer RNA of eukaryotic cells are transcribed as larger precursor molecules which are then cleaved by ribonucleases into smaller mature products (l-4). A number of ribonucleases have been found in the nuclei of animal cells. Winicov and Perry (5) have reviewed recent work on the enzymological aspects of rRNA' processing while I have summarized in Table I the known ribnnucleases from nuclei of animal cells. Although the physiological roles of these enzymes are unknown, some of them have been implicated in processing of rRNA (8,(11)(12)(13).
In earlier work concerning enzymatic processing of rRNA, the fractionation of nucleoplasmic and nucleolar ribonuclease activities on DEAE-cellulose columns was described (13). These fractionations showed the presence of at least three degradative activities. I have continued these studies and report here the purification of one of the ribonucleases. MATERIALS AND METHODS Cell Fractionation -HeLa cells (S,) were grown in Joklik modified Minimal Essential Medium plus 10% fetal calf serum (Grand Island Biological Co.) in spinner bottles. Cells were collected at a density of 4 x lo5 cells/ml.
For each cell fractionation, 3 x lo9 cells were used. All steps were carried out at O-4". Cells were swollen in 60 ml of low salt buffer (0.01 M Tris, pH 7.4, 0.01 M NaCl, 0.0015 M MgCl,) for 10 min; 2 M sucrose and 10% Triton X-100 were then added to final concentrations of 0.25 M and 0.25%, respectively.
The mixture was given 10 strokes in a Dounce homogenizer and was then centrifuged at 200 x g for 5 min to yield a pellet of nuclei. The nuclear pellet was washed by resuspension in 10 ml of RSB plus 1 ml of 10% Tween 40 and 0. 5  SDS was added to 0.1% to stop the reaction. The reaction mixture was fractionated on a P-30 Bio-Gel column (0.4 x 20 cm), previously washed with low salt buffer containing 0.1% SDS, and eluted with the same buffer; fractions (150 ~1) were collected. NaOH was added to each fraction to 0.3 N and the mixtures digested overnight at 37". HCl was then added to adjust the pH to between 6 and 7. The samples were applied to DEAE-Sephadex A-25 column (0.5 x 4 cm). Nucleosides were eluted with two l-ml portions of low salt buffer and nucleotides were eluted with two l-ml portions of low salt buffer containing 0.2 M NaCl, followed by two l-ml portions of low salt buffer containing 0.3 M NaCl. The ratio of total counts to the nucleoside counts was used to calculate the average chain length in each Bio-Gel fraction.

DEAE-cellulose
Fractionation- Fig.  1 shows the DEAEfractionation of crude nucleoplasmic (a) and nucleolar (b) extracts. For this, 50 ml (50 to 100 mg of protein) of nucleoplasmic or 10 ml (10 to 20 mg of protein) of nucleolar extract were placed on a DEAE-cellulose column (2 x 10 cm or 0.9 x 10 cm) previously washed with low salt buffer. After the samples had passed through, the columns were eluted with a linear gradient of 50 ml of low salt buffer and 50 ml of low salt buffer containing the indicated concentrations of NaCl. Each fraction was assayed using [3H]poly(C) as substrate. Multiple peaks of enzymatic activity were found both in the nucleoplasmic and nucleolar extracts, including a substantial fraction that was not retained by the column. The active fractions that did not bind to DEAE were pooled to give Fraction I (Table II). Approximately 20% of the input protein was found in Fraction I, although these fractions contained almost no was placed on a column (2 x 10 cm) and eluted with a linear gradient of 50 ml of low salt buffer and 50 ml of low salt buffer containing 0.2 M NaCl; 3.3-ml fractions were collected. The A,,, was determined (---) and 30 ~1 of each fraction was removed to assay for enzymatic activity (0) using poly(C) as substrate. b, crude nucleolar extract (13 mg of protein) was placed on a column (0.9 x 10 cm) and eluted with a linear gradient of 50 ml of low salt buffer and 50 ml of low salt buffer containing 0.4 M NaCl; 2.5-ml fractions were collected. Fractions between the arrows were pooled to yield Fraction I. nucleic acid as indicated by A280/A26,, ratios, which were in the range of 1.0 to 1.5 for both nucleoplasmic or nucleolar extracts. Phosphocellulose Fractionation -The pooled fraction from the DEAE-cellulose column was dialyzed against 0.01 M Tris (pH 7.4), 0.01 M NaCl, and 1 mM dithiothreitol.
A portion of the nucleolar material, containing 107 units of enzyme in 4 ml, was placed on a phosphocellulose column (0.9 x 4 cm). After the samples had passed through, the column was eluted batchwise, first with 5 ml of the dialysis buffer and then 5 ml each of potassium phosphate buffer (pH 7.2) at concentrations of 0.1, 0.2, 0.3, 0.5, and 1.0 M, each containing 1 mM dithiothreitol. Fractions of 5 ml were collected. After dialyzing all fractions against low salt buffer, the fractions containing enzymatic activity were pooled (Fraction II).
The enzyme was eluted from phosphocellulose with 0.2 M potassium phosphate buffer, although occasionally some activity was found in the 0.3 M phosphate buffer eluate. Larger volumes of sample (up to 25 ml of Fraction I) were also used on identical columns with similar results.  Nucleoplasmic Fraction I enzyme showed an identical fractionation on phosphocellulose.
Gel Filtration -The phosphocellulose fraction (Fraction II) was further purified by Sephadex G-100 gel filtration.
The material in 7.5-ml portions was layered onto a column (2 x 45 cm) previously equilibrated with low salt buffer. Elution was carried out with the same buffer. Fractions of 3.3 ml were collected at 15-min intervals. At this stage, the enzyme eluted as a single symmetrical peak of activity.
Calibration of the column with human y-globulin, ovalbumin, and cytochrome c gave an estimated molecular weight of the enzyme of 16,000. To stabilize the activity, 1 mg/ml of albumin was added to each of the active fractions. Enzyme preparations from nucleoplasmic and nucleolar extracts showed identical elution patterns on Sephadex. Although it is not possible to precisely determine the distribution of the enzyme between nucleoplasm and nucleoli because of the presence of other activities in the crude fractions, we have made an estimate based on the recovery of the purified enzymes (Table  II). This indicates that at least 70% of the enzyme is associated with the nucleolus.

Enzymatic Properties
Purified enzyme fractions from nucleoplasm and nucleoli have identical properties in all cases.
Endonucleolytic Cleavage-When using [3H]poly(C) as substrate, the enzyme gave increasing amounts of alcohol-or acidsoluble radioactivity with increasing times of incubation (Fig.  2). However, the alcohol-soluble counts were considerably lower. This is because only oligonucleotides of length 6 or smaller are soluble in alcohol while oligonucleotides up to 20 nucleotides long are soluble in trichloroacetic acid (18). No mononucleotides were detectable even when nearly 100% of the substrate was made acid-soluble. These results show that initially the enzyme cleaves poly(C) into oligonucleotides of relatively large size. This is shown directly in Fig. 3, where the poly(C) was examined directly by gel electrophoresis before and after enzymatic cleavage. After 5 min of incubation with a low level of enzyme, when no alcohol-soluble counts can be detected, the treated substrate was fragmented as indicated by its increased mobility. In addition, the size of the oligonucleotides after enzymatic digestion was examined by fractiona- tion on a P-30 Bio-Gel column. Individual fractions were hydrolyzed with NaOH. The average size of each fraction was determined by measurement of the nucleotide to nucleoside ratio (see "Materials and Methods"). The substrate, [3H1poly(C), had an average size of 40 nucleotides, while after endonuclease digestion the oligonucleotides were from 25 to 11 nucleotides in length. The presence of nucleosides in the NaOH digests of the product indicates that the oligonucleotides are terminated with 3'-OH groups. The validity of this conclusion was confirmed by demonstrating the absence of phosphatase in the purified enzyme (see below).
Substrate Specificity- Fig.  4 shows the treatment of poly(0, poly(U), poly(A), 28 Sand 18 S rRNA, and hnRNA by the purified enzyme. Poly(C) and poly(U) are the best substrates, although rRNA and hnRNA are also cleaved. Poly(A) is not degraded at all. Double-stranded poly(G):poly(C) was also tested as substrate and no degradation occurred in a 2-h incubation. No DNase activity could be detected in a similar assay with purified enzyme.
Effect ofpH-The enzyme shows a broad pH profile with an optimum at approximately pH 7.2. The enzymatic activity is identical in potassium phosphate buffer and in Tris buffer.
Effect of Mg2+ Concentration -Dialysis against 1 mM EDTA has almost no effect on enzymatic activity. When dialyzed enzyme is assayed in the presence of increasing amounts of W+, enzymatic activity is reduced: 10 mM Mg2+ inhibits by 50%.
Effect of Zonic Strength-The purified enzyme was sensitive to increasing concentrations of NaCl. NaCl at 0.04 M gives 50% inhibition.
NaF at 10 mM had no effect on enzymatic activity although 3 mM NaF has been shown to be inhibitory to a nuclear exonuclease (10). Potassium ion at similar concentrations (10 mM) had no effect on the enzyme, nor did phosphate buffer.

(right).
Hydrolysis of poly(C) by venom phosphodiesterase and purified endonuclease.
Temperature Lability of Enzyme-The purified enzyme is heat-labile, although high enzyme concentrations or the presence of albumin give considerable protection. Heating the enzyme at 58" in the presence of albumin (1 mglml) for 10 min results in 10% loss of activity, while in the absence of albumin all activity is lost in 5 min. Albumin is necessary for storage of the enzyme at -20". Enzyme stored without albumin loses 50% of its activity in 72 h, whereas all activity is retained in its presence.
Product Analysis -Several lines of evidence show that the enzyme yields oligonucleotides bearing 5'-phosphate end groups. In the first series of experiments, snake venom phosphodiesterase was used. This phosphodiesterase catalyzes the hydrolysis of oligonucleotides with 3'-hydroxyl end groups to yield mononucleoside 5'-phosphates (19). Fig. 5 shows that when poly(C) is incubated first with the purified endonuclease and then with snake venom phosphodiesterase it is hydrolyzed much more rapidly than without prior endonuclease treatment. This indicates that the cleavage by the endonuclease is such as to produce 3'-hydroxyl and 5'-phosphate end groups.
Experiments were also performed using spleen phosphodiesterase. This enzyme's mode of action is complementary to that of the venom enzyme; a 5'-hydroxyl group is necessary (19). The presence of increasing amounts of purified endonuclease in reaction mixtures containing poly(C) and spleen phosphodiesterase give inhibition of mononucleotide released by the spleen enzyme. By contrast, comparable amounts of pancreatic ribonuclease A had a slight stimulatory effect (Fig. 6). This result is consistent with the cleavage of the poly(C) to yield fragments bearing 5'-phosphate end groups.
Long term incubation ( ase gave only 5'-CMP on thin layer chromatography plates (Fig. 7). No cytidine can be detected even after 18 h of incubation indicating the absence of any contaminating phosphatase.

DISCUSSION
Multiple peaks of enzymatic activities were found by DEAEcellulose fractionation of either the nucleoplasmic or nucleolar extracts (Fig. 1). Because of this, the extent of purification (Table II) is difficult to quantitate. Moreover, there seemed to be an inhibitor of endonuclease activity in nucleoli. I detected only very low poly(C) degradative activity in crude nucleoli, while the activity of enzyme increased greatly after the DEAE-cellulose chromatography.
No comparable increase of activity occurred with nucleoplasmic fractions.
We have begun the purification of nuclear ribonucleases because of this potential role in RNA processing (13). Based on the character of various intermediate rRNAs produced by in uiuo maturation of ribosomal RNA and on data determined in the electron microscope, Winicov and Perry (5) have postulated four distinct endonucleolytic cleavage sites in the rRNA precursor molecules. These sites are common in several animal species. HeLa cells and mouse L-cells share the same cleavage sites, although the order of cleavage is different. Therefore, since processing endonucleases may be universal, it is of interest to compare the HeLa enzyme to that of other enzymes already implicated in processing.
The endonuclease described here is very similar to one described by Winicov and Perry (8) from mouse L-cells. Both enzymes cleave poly(C) but not poly(A), are inhibited by Mg*+, and are found bound to nucleoli. However, two important differences exist. The HeLa enzyme produces oligonucleotides having a 5'-phosphate end group, while the enzyme previously described produced oligonucleotides with 3'-phosphate end groups. It is of course possible that since the mouse enzyme was not highly purified, the termini observed were due to a contaminating activity. However, the latter enzyme does not cleave poly(U) while that from HeLa does. The HeLa endonuclease can also be compared to the enzyme isolated from pre-ribosomal particles by Prestayko et al. (9). The cellular location and substrate specificities are similar, however, this enzyme has been shown to be extremely stable to heating while our enzyme is heat-labile.
Cordis et al. (10) have isolated an endonuclease from rat liver nuclei. This enzyme, like the HeLa enzyme reported here, produces oligonucleotides with 5'-phosphate end groups. However, the HeLa enzyme does not cleave poly(A) as does that from rat liver. Heppel isolated an endonuclease from pig liver (71, similar to the enzyme of Cordis et al. (10). It also attacks poly(A) to give oligonucleotides bearing 5'-phosphate end groups. It appears possible based on the above discussion as well as our experimental results (Fig. 1) that a number of nuclear endonucleases exist. However, some of the apparent differences may be due to the method of purification used and the purity obtained, since the presence of inhibitors or other nucleases may influence the observed characteristics.
The properties of the endonuclease described here make it a good candidate for a processing enzyme. The enzyme produces oligonucleotides bearing 5'-phosphate termini which are good substrates for the 3'-OH specific exonuclease of Sporn et al. (6). The endonuclease by itself or in combination with the exonuclease could cleave and trim the precursor to create the proper intermediates and the mature rRNA. Moreover, the fact that the enzyme is inhibited by NaCl and Mg*+ is consistent with the stability of precursor 45 S RNA in the high salt preparative method of Penman (15,20).