The Effects of Polyamines on a Residue-specific Human Plasma Ribonuclease

A ribonuclease, purified some 2700-fold from human plasma, exhibited a strong predilection for the hydrolysis of intemucleotide bonds containing cytidylic acid. Analysis of [3’-32P]- and [5’-32P]phosphoryl-terminal fragments obtained after enzymic digestion of rabbit liver and yeast RNA indicated that the nucleotide found at the 3’ terminus of the fragments was invariably cytidylic acid. The nucleotide at the 5’ terminus varied between cytidylic and uridylic acids in a ratio of 9:l. When characterized by DEAE-cellulose chromatography, approximately 70% of the digest consisted of oligonucleotides from 4 to 8 nucleotides in length. Enzyme activity, when measured in low ionic strength buffer, could be increased severalfold above control levels the addition of either of the polyamines, spermidine or spermine. These substances also restored nucleolytic activity to preparations inhibited by such ordered synthetic polyribonucleotides as polyguanylic acid. Estimations of the molecular weight of the enzyme, both by Sephadex gel filtration and sucrose density centrifugation, indicate that the weight may vary, depending on the presence or absence of certain cations. Of the cations examined, spermidine and spermine appear to have the greatest effect, causing an alteration in molecular weight from greater than 150,000 to approximately 32,000.

It has been shown that the polyamines can exert considerable effects on the activity of a ribonuclease isolated from a soil-living organism, Citrobacter sp. (l-3).
The polyamine, spermidine, was found to enhance enzyme activity, to alter the residue specificity exhibited by the enzyme in its attack on RNA, and to reverse the competitive inhibition of enzyme activity caused by highly ordered synthetic polyribonucleotides. Inasmuch as these observations suggested that the polyamines may be playing a significantly different role in RNA metabolism than had hitherto been suspected (4,5), it was of interest to determine whether these effects were limited to the Citrobacter system or were a reflection of a more general phenomenon in which the polyamines regulated RNA concentration within the cell by controlling RNase activity.
To explore this latter view, the effects of polyamines on a number of ribonucleases were examined.
In the earlier paper of this series, the enzymes chosen for study were all highly specific, but were of microbial origin.
In the present paper, the ribonuclease from human plasma was selected because of a report suggesting specificity for cytidylic acid (6). The enzyme, purified approximately 2700-fold, exhibited a strong predilection for internucleotide bonds of cytidylic acid. Spermidine and spermine were found to cause as much as a 7-fold increase in enzyme activity.

Materials
Human plasma was purchased from a commercial blood bank or was obtained from volunteers.
Yeast RNA, purified by the method of Crestfield et al. (7) was prepared in the laboratory. Rabbit liver RNA was purified by the method of Delihas and Staehelin (8).

Assay of Human Plasma
Ribonuclease-The standard assay system contained 1.5 pmol of polymer or 1.0 mg of RNA, 100 perno of buffer (Tris-HCl, pH 7.5, ior (C), or sodium phosphate, pH 8.0, for RNA), and 0.5 mg of bovine serum albumin and enzyme in 1 ml. After incubation for 15 min at 25", the reaction was stopped by the addition of 1 ml of 2 N perchloric acid and the reaction vessel chilled for 10 min in an ice bath. The cloudy reaction mixture was clarified by centrifugation and the absorb-ante of the acid-soluble nucleotides was measured at 260 nm (13). An enzyme unit is defined as that amount of enzyme needed to cause an increase in absorbance of 1.0 under conditions of the assay, when (C), was used as the substrate. Molecular Weioht of Human Plasma Ribonuclease-The molecular weight of human ilasma ribonuclease was estimated by the gel filtration method of' Andrews (14), using as protein standards bovine serum albumin (67.000) (15 After 16 hours of incubation at 37", the digest was subjected to ultrafiltration through a Centriflo membrane cone (CF50, Amicon Corp., Lexington, Mass.) to remove all fragments of molecular weight above 50,000. An aliquot (1.0 ml) then was treated with 0.1 N HCl at 30" for 1 hour to convert the cyclic phosphodiester bonds to the monoester form followed by dialysis for-5 hours against five changes (15 liters) of 0.1 M Tris-HCl buffer. DH 7.6. Terminal nhosnhate nrouos were removed from the dialyzed oligonucleotfdes by treatment with alkaline phosphatase (9) for 2 hours at 37". Mechanical shaking of the reaction mixture with a chloroform-octanol (9:l) solvent was used to remove all protein from the solution.
The reaction mixture was incubated for 2 hours at 37" and extracted with the chloroform-octanol solvent followed by ether as described above. The [3'-32P]phosphoryl-terminal oligonucleotide fraction was dialyzed for i ho&s against 10 liters-of water and digested with 1.0 N NaOH for 16 hours at 37". Ten micromoles of each of the four 3'-mononucleotides were added to the digest which was added to a column (0.9 X 6 cm) of Dowex l-X8 formate.
The mononucleotides were eluted with increasing normality of formic acid as described by Hurlbert et al. (22). Each of the four mononucleotide peaks was identified by absorption spectrum.
Each peak was collected, reduced in volume by reduced pressure distillation, and brought to constant specific activity by paper electrophoresis (Whathan No. 3MM paper at 3000 volts. 190 ma for 2 hours at 24" in formic acid) followed bv thin layer' chromatography on ECTEOLA-cellulose using 0.1 h~ NaCl as developing solvent.
The amount of mononucleotide isolated was estimated from its molar extinction coefficient (23). Radioactivity was measured in a Beckman L-150 liquid scintillation spectrometer.

Preparation of [5'-32P]PhosphoryLterminal
Oligonucleotides-Another aliquot (1 ml) was taken from the rabbit liver RNA digest described above and was treated with alkaline phosphatase for 2 hours at 37". All protein was removed from the incubation mixture by mechanical shaking with the chloroform-octanol solvent as described above. After removing residual solvent with ether, the oligonucleotide-containing fraction (aqueous phase) was dialyzed for 2 hours against two changes (8 liters) of 0.1 M Tris-HCl, pH 7.6. The dialyzed fraction was added to a reaction mixture-consisting of 1 mmol of [Y-~*P]ATP (specific activitv 5 X 10" cnm aer mmol). 0.05 ml of 0.6 M M&l%. 0.05 ml of 0.3 M dithiothreitbl, &nd 5 units of polynucleoti& kinase (11). After 3 hours at 37", the solution was extracted with the chloroform-octanol solvent followed by ether and was dialyzed for 2 hours against four changes (16 liters) of 0.1 h% Tris-HCl buffer, pH 7.6. The [5'-32P]phosphoryl-terminal oligonucleotide solution was treated with 0.3 ml of purified snake venom phosphodiesterase (10) to liberate the 5'-mononucleotides.
After 20 hours at 37", 10 pmol of each of the four 5'-mononucleotides were added to the reaction mixture, which was made alkaline with NHdOH. The nucleotide peaks were separated and brought to constant specific activity by chromatographic procedties as above, except that an additional system, ascending paper chromatography on Whatman No. 3MM paper with isobutyric acid, NHrOH, and water (66:1:33) as developing solvent was included as the final step.

Purijication of Enzyme
The following operations were performed at O-5". The purification data are summarized in Table I. Step 1: Ammonium Sulfate Fractionation-With the pH maintained at 7.0 by the dropwise addition of 1 N NaOH, solid ammonium sulfate (28 g) was added wit,h stirring to 100 ml of human plasma over a 20-min period.
The enzyme solution was clarified by centrifugation and an additional 28 g of ammonium sulfate were added to the supernatant solution in the same manner. The precipitate which formed between 40 and 80% saturation was collected by centrifugation and was dissolved in 0.01 M phosphate buffer, pH 6.8, to a final volume of 36 ml. The yellow, somewhat turbid, solution was dialyzed for 3 hours against 100 volumes of the same low molarity phosphate buffer.
Step 2: Cellulose Phosphate Chromatography-To a column (2 X 40 cm) of cellulose phosphate, which had been equilibrated previously with 0.01 M phosphate buffer, pH 6.8, 36 ml of the dialyzed enzyme were added. After the column was washed with 150 ml of eqiilibrating buffer 250 ml of a linear gradient of elution from 0.5 to 1.5 M KC1 in eauiiibratine buffer were annlied. Fractions, 4 ml in volume, were collectedY ( Fig. 1) and i<ose fractions with the highest activity (63 to 94) were combined and were dialyzed for 3 hours against 40 volumes of 0.05 M phosphate buffer, pH 8.
Step 3: Afinity Chromatography-The enzyme solution (110 ml) obtained in the previous step was applied to a column containing 2 g of activated Sepharose 4B (0.9 X 7.0 cm) to which 10 pmol of (G), were covalently bound by the method of Poonian et al. (26). After addition of the entire enzyme solution, the column was washed with 30 ml of 0.05 M phosphate buffer, pH 8.0, to remove completely an initial protein peak which contained asmall amount of r;bonuElease activity.
The ribonuclease was eluted with 0.05 M nhosnhate buffer. DH 8. containing 1 M KC1 (Fig. 2). Fractions, l-ml in volume, werk collected and those (39 to 46) having most of the enzyme activity were combined and used for all subsequent studies.

pH Optima
When the production of acid-soluble nucleotides was used as an index of enzyme activity, considerable variation in pH optima was found, not only between individual substrates, but also between different buffers used in the measurement of hydrolysis of the same substrate. Thus (Fig. 3), at the optimum pH (7.5) in Tris-HCl buffer, hydrolysis of (C), was approximately 1.5 times as great as it was at the optimum pH (6.5) in phosphate buffer. The hydrolysis of RNA showed the same optimum pH in either Tris-HCl or phosphate buffer (pH 8), but the enzyme 600/, of the enzyme was adsorbed to the column but could be eluted by buffer solution containing 5 mM spermine or spermidine. These latter fractions, or fresh enzyme, when applied to a column equilibrated with 0.5 mM spermine or spermidine in the buffer system were eluted at a position corresponding to a molecular weight of 32,000 (Fig. 4). The aggregation phenomenon was reversible by the addition or removal of polyamine from the equilibrating buffer. Neither Ca2+ nor Mg2f at the same concentrations as the polyamines was able to produce these effects. P 1.2 Sucrose Density Centrifugation-The molecular weight of the plasma ribonuclease also was determined by ultracentrifugation in a sucrose density gradient, both in the presence and absence of spermidine.
As seen in Fig. 5, the peak of enzyme activity, in the absence of spermidine, was found at the bottom of the gradient with a gradual reduction in activity as the less dense sucrose concentrations were approached. The pattern seemed to indicate a spectrum of molecular species ranging in weight from well over 150,000 downward.
In the presence of the polyamine two sharp peaks of enzyme activity, with estimated molecular weights of 45,120 and 28,050, respectively, were readily differentiated. Centrifugation in the presence of spermine produced similar results, except considerably greater recovery of enzyme was attained.

Although
Ca2f and Mg2+ at the same concentration produced an apparent shift in the In each case, 700 units of enzyme were applied to a column.
In the first experiment (m--m), the column was equilibrated and eluted with 0.05 M Tris-HCl buffer, pH 7.5. The other curves show the elution patterns when the column was equilibrated and eluted with the same buffer containing 0.5 mM spermine (O---0) or spermidine (n--A).
Enzyme activity was measured against (C), in the standard system and is shown in terms of increase in absorption at 260 nm. Enzyme activity was measured as described in Fig. 4. Effect of cations OTL hydrolysis of RNA and (C), by human plasma ribonuclease Enzyme activity against RNA was measured in the standard assay system, except that the reaction mixture contained 10 pm01 of sodium phosphate buffer, pH 8. When the effect of cation on hydrolytic activity was -examined, 1 pmol was added, as indicated, to the reaction mixture just before the addition of 13 units of dialyzed enzyme. With (C), as substrate, the reaction mixture contained 10 pmol of Tris-HCl buffer, pH 7.5, and incubation time was 738 min. When the effect of cation on enzyme activity was studied, 0.5 wmol was added, as indicated, just before the addition of 4 units of dialyzed enzyme. Ribonuclease activit.y was assayed as described under "Methods." Addition I Eflect of Cations on Enzyme Activity-The hydrolytic activity of human plasma ribonuclease is decreased significantly when the buffer concentration in the reaction medium is reduced from 0.1 M to 0.01 M. Activity against RNA could be restored by the addition of various cations to the dilute buffer medium (Table  II).
The greatest stimulations occurred when either of the polyamines, spermidine or spermine, was added at an optimal concentration of 1 m&I. At the latter concentration, enhancement of activity by Ca*f, i\ilg*f, or putrescine was considerably less. Raising the concentration of Mg2+ or putrescinc to 5 mM resulted in an enhancement of activity which approached that seen with spermidine and spermine.
Other cations did not affect activity, with the exception of Zn2+ and Ni2+ (not shown in Table), which acted as strong inhibitors.
The diminished hydrolysis of (C), in the dilute buffer system was similarly increased by various cations (Table II).
At a concentration of 0.5 mM, spermine and spermidine caused 2-and a-fold stimulations of activity against (C),, respectively, whereas C&2+, Mgz+, and putrescine had lesser or no effect. At a IO-fold greater concentration, all of the cations save spermine produced further enhancement of hydrolysis.
Spermine inhibited above 0.5 mM, possibly because of precipitation of the substrate. It should be noted that the conditions under which the hydrolysis of the synthetic polymer and that of RNA were studied are quite different.
More enzyme was required, as well as a longer incubation time for the digestion of the latter substance than for polycytidylic acid. The enzyme has a considerably greater affinity for the synthetic polymer than for RNA. Inhibition of Nuclease Activity by Synthetic Polynucleotides-A number of synthetic polyribonucleotides having ordered secondary structure have been reported to inhibit microbial nucleolytic enzyme activity (1,27). Although some of these same compounds also inhibit human plasma ribonuclease, the degree of inhibition appears to be dependent on the substrate under study (Table III).
The amount of ordered polynucleotide needed to produce 500/, inhibition of RNA hydrolysis can be as plasma ribonuclease Activity was measured in the standard assay system using either 1.5 pmol of (C), or 1.0 mg of yeast RNA as substrate.
The inhibitor concentrations shown are those which caused 5O'j'& inhibition of hydrolytic activity.
Enzyme activity was assayed as described under "Methods." Four units of enzyme were added to a reaction mixture containing 1.5 pmol of substrate and 100 pm01 of Tris-HCl buffer, pH 7.5. After incubation at 25" for 736 min, the reaction was stopped by the addition of 1 ml of 20 mM lanthanum nitrate in 2 N perchloric acid. Enzyme activity was measured as described under "Methods." Reaction conditions were as described in Table II, except that the reaction mixtures were incubated for 71% min at 25". When the effects of inhibiting agents and the cations on the reaction were studied, 0.5 pmol of cation and inhibitor (at the concentration indicated below) was added. Hydrolytic activity, measured as described under "Methods, " is expressed as a percentage of control activity.  Although the synthetic polynucleotides are not as strong inhibitors of RNA hydrolysis, reversal of this inhibition is less readily accomplished (Table V). This may be related to the fact that much higher inhibitor levels are used. Only spermine appeared to be consistently effective, and it induced only a partial restoration of activity.

Specijicity of Enzyme jar Synthetic
Polyribonucleotides-The purified enzyme has a preference for cytidylic acid residues in RNA.
Of the homopolymers examined, only polycytidylic acid was attacked (Table VI).

Hydrolysis of Nucleic Acids
Analysis of Digestion Products of RNA-No evidence of hydrolysis of either single-or double-stranded DNA was found. When yeast RNA (2.5 mg) was incubated in 0.1 M phosphate buffer, pH 8, with 20 units of enzyme at 37", the analysis of the digestion products (28) after 16 hours revealed the presence of only one mononucleotide, cyclic 2':3'-CMP. Analysis of the 3' and 5' termini of the digestion products, large and small, was carried out. Rabbit liver RNA was chosen as a substrate, primarily because it is mammalian in origin and, secondly, because it can be purified in relatively large quantities.
Following a 16-hour hydrolysis of this RNA by human plasma ribonuclease, 321' was introduced enzymatically into the 3' and 5' termini of the digest fragments, and their terminal (i.e. radioactive) nucleotides were isolated and characterized.
Approximately 90% of the total radioactivity at the 5' terminus was in cytidylic acid, whereas 10% was found in uridylic acid (Table  VII).
Almost 100 '% of the radioactivity at the 3' terminus was in cytidylic acid (Table VII).
Clearly the predominant cleavage of phosphodiester bonds in rabbit liver RNA was between cytidylic acid moieties.
Identical results were obtained in the digestion of yeast RNA.
In studies in which undigested rabbit liver RNA was used, incorporation of 32P was less than 2yo of that occurring in the enzyme-degraded RNA. Thus, corrections for pre-existing end groups would not significantly change the data presented.
Size of Oligonucleotide Prwments-To establish the average much as 200 times greater than the amount needed to produce the same degree of inhibition of the hydrolysis of polycytidylic acid (Table III).
An earlier study (1,2) showed that this type of inhibition often can be completely reversed by polyamines.
Complete restoration of the enzyme activity against (C), can be induced by 0.5 mM spermine or spermidine, except in the case of (G, U), inhibition (Table IV).
The same concentration of Ca2f or Mg2+ is significantly less effective in overcoming inhibition. See "Methods" section for reaction conditions. Each combined nucleotide fraction (designated by Roman numerals) was obtained from a digest of rabbit liver RNA that had been subjected to DEAE-cellulose chromatography (29) (see Fig. 6). The combined fraction, adjusted to pH 8, was freed of urea and salt by rechromatography on separate columns of DEAE-cellulose (carbonate) as described by Tomlinson and Tener (29). Upon completion of this second chromatography the nucleotidecontaining fractions were combined, brought to dryness by lyophilization, and the residue redissolved in a small volume of water. When the process had been repeated, the residue was dissolved in 2 ml of water and an aliquot used to determine the average size of the nucleotides by the total phosphorus to terminal phosphorus ratio (30 6. Chromatography of human plasma ribonuclease digest of rabbit liver RNA. To a solution (5 ml) containing 25 mg (420 absorbance units) of rabbit liver RNA and 5OO~mol of PO, buffer, pH 8, 30 units of human plasma ribonuclease were added. The reaction mixture was incubated at 37" and at the end of 16 hours was placed in a boiling water bath for 5 min. After centrifugation the supernatant solution (220 absorbance units) was applied to a column (2.5 X 50 cm) of DEAF-cellulose (chloride form) that had been prepared as described by Tomlinson and Tener (29). The column was washed with 500 ml of water and then 3 liters of a linear gradient of elution from 0 to 0.3 M NaCl in 7 M urea and 0.0025 M sodium acetate buffer, pH 4.7, was applied.
Fractions B ml in volume were collected at a flow rate of approximately 20 ml per hour. The bars at the base of the elution diagram represent the number of fractions which, when combined, were used for phosphorus analysis. The Roman numerals at the top of each peak are used to designate the combined fractions. length of oligonucleotides produced within a given period, use was made of the procedure of Tomlinson and Tener (29) in which the products of hydrolysis were separated on a DEAE-cellulose column using 7 M urea as the eluting solvent. The hydrolytic products were separated on the basis of the length of the nucleotide chain (Fig. 6). The degree of polymerization within each combined fraction, designated by Roman numerals, was established by determining the ratio of total phosphorus to terminal phosphorus in each fraction.
The results of this analysis, as well as the percentage of the distribution of the oligonucleotides, are summarized in Table VIII.
From these data approximately 70% of the recovered products of rabbit RNA hydrolysis consisted of oligonucleotides equal to or greater than 5 nucleotides in length.
Essentially the same type of results was obtained with yeast RNA. The liberation of C-cyclic-l' during the course of hydrolysis of polycytidylic acid suggests that the human plasma ribonuclease functions as other ribonucleases in that it cleaves the phosphodiester bond between a nucleoside 3'.phosphate and the 5'-hydroxyl group of the adjacent nucleotide (31). That this method of hydrolytic attack is probably the same when RNA is the substrate is indicated by the finding in RNA digests of Ccyclic-l' and oligonucleotides terminating in cyclic phosphates as digestion products.
Aside from revealing that the human plasma ribonuclease possesses endonucleolytic activity, analysis of the fragments present in the RNA digest is of considerable interest in another sense as well, primarily because of the apparent high specificity exhibited by the enzyme for ribonucleotide bonds containing cytidylic acid residues.
Initially, a strong suggestion of this specificity came from the rather poor hydrolytic activity against synthetic polyribonucleotides that did not contain cytidylic acid. After the introduction, by enzymic means, of 321' into the 3' and 5' termini of oligonucleotide fragments present in the RNA digest, it was possible to show that the enzyme cleaved exclusively between cytidylic acid residues at the 3' terminus and exhibited a preferential specificity for cytidylic over uridylic acid residues in an approximate ratio of 9: 1 at the 5' terminus.
The high specificity for cytidylic acid residues is of particular moment in view of the recent report of a large tract of cytidylic acid residues within the RNA of an encephalomyocarditis virus (32). This latter observation suggests the possibility that other RNAs with long tracts of cytidylate residues may exist and that it is these substances which are the natural substrates of the human enzyme.
Another aspect of the enzyme's characteristics which is particularly striking is its response to the polyamines, especially spermidine and spermine.
The effects of these compounds on enzyme activity are in many respects similar to those seen with a microbial enzyme, Citrobacter nuclease (1,2). Enzyme activity, for example, in the presence of these compounds is enhanced severalfold over control levels. The inhibition of hydrolytic activity observed with the ordered polynucleotides can, with some exceptions, be reversed by spermidine and spermine. Other cations which can enhance enzyme activity weren't as effective in this area.
The most dramatic effect of the polyamines on the human enzyme is their apparent ability to control the degree of aggregation of the enzyme molecule.
It was found, for example, that the apparent molecular weight of the enzyme, when determined by gel filtration 011 Sephadex columns, is decreased from approximately 150,000 to 32,000 in the presence of the polyamines, spermidine and spermine.
These compounds similarly cause a change in the position of the enzyme in a sucrose gradient after centrifugation, indicating a shift in molecular weight from well over 150,000 to approximately 29,000. These studies suggest that the enzyme undergoes aggregation-disaggregation reactions under the influence of the polyamines.
The previously expressed view (1, 2) that polyamines may represent control factors in the metabolism of nucleic acids thus is further supported by the effects noted upon human plasma ribonuclease.