A Ribosome-associated Inhibitor of the Digestion of Polyadenylate-containing Ribonucleic Acid*

Abstract An inhibitor associated with guinea pig adrenal ribosomes inhibits the digestion of polyribosome-associated poly(A)-rich messenger RNA purified from human peripheral lymphocytes by two ribonucleases: (a) a ribonuclease which is co-purified with the poly(A)-rich RNA and which initially cleaves the poly(A) tract and (b) an endoribonuclease (poly(A)ribonuclease) purified from bovine adrenal cortex cytoplasm which cleaves and digests the poly(A) tract of poly(A)-rich RNA synthetic poly(A). The inhibitor has been purified 150- to 200-fold with respect to specific activity. It is rapidly inactivated by heat, freeze-thawing, or by incubation with trypsin, but treatment with DNase or RNase has no effect on its activity. Glycerol gradient density centrifugation indicates an S of 1.5 to 2.0 for the inhibitor. Kinetic studies suggest a competitive mode of inhibition. The inhibitor fails to prevent digestion of synthetic poly(A) by poly(A) ribonuclease and appears to be ineffective in preventing digestion of poly(A)-rich RNA by several other ribonucleases.

$ Teaching and Research Scholar of the American College of Physicians. To whom all requests for reprints should be addressed.
It appears that mRNA consists of two portions: one which serves an information function, and the other, the poly(A) tract, a regulatory function. This regulatory function of the poly (A) tract remains undefined. Drug and hybridization studies have suggested a role in transport of mRNA from nucleus to cytoplasm (4) ; however, the presence of a poly (A) tract in viruses whose life cycle is confined to the cytoplasm (19,21) and the demonstration of cytoplasmic polyadenylylation of transcripts in sea urchin embryos (42) suggest other or additional roles.
We have previously reported that polyribosome-associated poly(A)-rich RNA isolated from cultured lymphocytes using the nitrocellulose filter technique (6) is cleaved by a putative ribonuclease co-purifying with the RNA, resulting in the removal of a segment which sediments at 4 to 6 S in linear SDS sucrose density gradients (43). The base composition of this 4 to 6 S endoribonuclease digestion product is identical to that of the isolated poly(A) tract (44). We have subsequently identified and partially purified a cytoplasmic ribonuclease from bovine adrenal cortex that digests synthetic poly (A) and cleaves the poly (A) tract of the poly(A)-rich RNA. Many properties of this enzyme are similar to those described for a ribonuclease isolated from pig liver nuclei (40). A ribosome-associated factor inhibits digestion of poly(A)-rich RNA by both the RNAassociated ribonuclease and the partially purified cytoplasmic ribonucleases. We report here the purification of this inhibitor and the initial definition of its properties. The identification of a specific inhibitor of ribonucleases which digest the poly (A) tract suggests that regulation of the digestion of the poly (A) tract can occur in the cytoplasm of the eukaryotic cells.

Identification and PuriJication
of Inhibitor-Addition of a 0.5 M KC1 extract of guinea pig adrenal ribosomes entirely prevented the digestion of poly(A)-rich RNA by its associated ribonuclease or by poly(A) ribonuclease as assayed by the retention on Millipore filters (Table I) or by retention on GF/A filters impregnated with poly(U) immobilized by ultraviolet radiation (48) (data not shown).
Addition of eluate following the digestion did not alter retention on the filter (Table I).
The inhibitory factor was purified from the 0.5 M KC1 eluate of guinea pig adrenal ribosomes.
Purification resulted in a 110. to 190-fold increase in specific activity of the inhibitor ( Table  II).
The increase in specific activity would be much greater if considered with respect to total cytoplasmic protein.
No inhibitor activity was found in the 100,000 x g supernatant; however, the presence of many other ribonuclease activities may have prevented detection of inhibitor activity in this and other crude fractions by the nitrocellulose binding assay. The inhibitor activity was extremely labile; in many preparations all activity was lost during the dialysis, during the phosphocellulose chromatography, or with a single freeze-thawing. Poor recovery of activity through the purification technique could, in part, reflect a loss of activity of the inhibitor during the many manipulations.
Inhibitor activity sedimented with peak activity at 1.5 to 2 S on linear glycerol density gradients (Fig. 1).
The inhibition of digestion of poly(A)-rich RNA by the associated ribonuclease or poly (A) ribonuclease was proportional to the amount of inhibitor added (Fig. 2).
Con$rmation of Inhibitor Activity by Rate-Zonal Analysis of mRNA---In order to confirm that this apparent inhibition of ribonuclease digestion was complete and could not represent an artifact of the method of assay, the poly(A)-rich RNA was sub- jetted to gradient analysis on linear SDS-MeSO sucrose density gradients following exposure to poly(A) ribonuclease in the presence or absence of inhibitor.
Lymphocyte poly(A)-rich RNA was pretreated to prevent self-digestion (see "Methods") and was employed as a substrate for poly (A) ribonuclease. When inhibitor purified through the ammonium sulfate fraction step was used, endonucleolytic digestion by poly(A) ribonuclease was inhibited, but new digestion products sedimenting at the top of the gradients were noted (Fig. 3A). When purified inhibitor was used, there was complete inhibition of the digestion produced by the poly(A) ribonuclease without appearance of the small digestion products (Fig. 3B). Similar results were obtained when the effect of inhibitor was examined with regard to the self-digestion of poly(A)-rich RNA by the associated ribonuclease.
Properties of Inhibitor-The inhibitor was inactivated by heating to 90" for 5 min (Table III).
It withstood freeze-thawing poorly with 50 to 100% loss of activity.
Storage in buffers with low salt concentration was associated with rapid loss of activity.
The inhibitor was inactivated by digestion for 15 min at 37" with pronase (50 I.cg per ml). Incubation of the purified inhibitor with trypsin (10 pg per ml) for 30 min at 37" in Buffer A, followed by addition of trypsin inhibitor (11 rg per ml) abolished the ability to inhibit self-digestion of poly(A)-rich RNA or digestion by poly(A) ribonuclease.
Similarly, preincubation for 20 min at 37" with trypsin immobilized on Sepharose beads abolished inhibitory activity (Table III).
Incubation with DNase or with pancreatic RNase A immobilized on Sepharose beads had no effect on inhibitor activity (Table III).
Kinetic Studies-The kinetics of the inhibition of poly(A)rich RNA digestion was studied by assaying digestion of varying amounts of lymphocyte poly(A)-rich RNA pretreated to prevent self-digestion in mixtures containing constant amounts of poly (A) ribonuclease and inhibitor. Two concentrations of inhibitor found to confer incomplete protection against digestion of poly(A)-rich RNA by poly(A) ribonuclease were selected. A Lineweaver-Burk analysis of the inhibitor suggests a competitive mode of inhibition (Fig. 4). Neither the enzyme nor inhibitor has been purified to homogeneity; therefore, the Ki cannot be accurately stated.  1 (left), Sedimentation profile of inhibitor. Inhibitor was prepared from 25 g of guinea pig adrenal cortex ( Table I). The fraction of eluate from the phosphocellulose column with peak inhibitory activity was dialyzed against Buffer C and sedimented through a linear 5 to 30% glycerol gradient (see "Methods"). Inhibitor activity was assayed using 20-~1 aliquots in duplicate determinations with 50 ng of [3H]uridylate lymphocyte poly(A)rich RNA as described in the legend to Table I

FIG. 3. Assessment of inhibition of ribonuclease digestion by sedimentation
analysis of poly(A)-rich RNA. Each assay contained in 50 11 of Buffer A, 400 ng of [3H]adenosine lymphocyte poly(A)-rich RNA (160 cpm per 10 ng) pretreated to prevent selfdigestion (see "Methods") and, where indicated, poly(A) ribonuclease (0.08 pg per ml) or inhibitor purified through ammonium sulfate fractionation (300 pg per ml) or sedimentation through the glycerol gradient (4 pg per ml) ( Table I). Following incubation at 37" for 15 min, the samples were adjusted to 0.3% SDS and an equal volume of dimethylsulfoxide was added. The samples were allowed to remain at room temperature for 10 min and a O-fold volume of 0.3oJo SDS was then added. The samples were then applied to linear 5 to 20% sucrose density gradients, centrifuged, and analyzed (see "Methods").
Additions to the incubation mix- Substrate SpecijEcity of Inhibitor-h order to clarify the substrate requirements for the inhibitory activity, the ability of the inhibitor to prevent digestion of synthetic polyriboadenylic acid by poly(A) ribonuclease was tested. As shown in Table IV, the inhibitor did not prevent digestion of synthetic poly(A) even RNA (108 cpm per 10 ng) pretreated to prevent selfdigestion and poly(A) RNase (0.08 rg per ml). Following incubation for 15 min at 37', samples were evaluated for digestion (see "Methods").
Results are the average of duplicates differing by less than 3y0. [3H]IJridylate-lymphocyte poly(A)-rich RNA pretreated to prevent self-digestion was used as substrate for digestion by poly-(A) ribonuclease.
In each assay, purified inhibitor (4 pg per ml) was first incubated with gentle agitation for 20 min at 37" in Buffer A alone, with DNase I (10 rg per ml), trypsin immobilized on Sepharose beads (0.5 unit), or pancreatic ribonuclease immobilized on Sepharose beads (0.07 unit) or for 5 min at 90" in Buffer A. The Sepharose beads were removed by centrifugation.
The radiolabeled poly(A)-rich RNA (350 ng) and poly(A) ribonuclease (0.08 pg per ml) were then added, the assay volume was adjusted to lOOJ, and incubations were continued for 15 min at 37'. Assay for digestion was done as described in "Methods." The preparation of DNase I used was demonstrated by gradient analysis of mRNA to contain endoribonuclease activity. DNase I (10 pg per ml) digested 0.7 ng of poly(A)-rich mRNA under the conditions of this assay. Results are the average of duplicates differing by less than 4%; similar results were obtained in an experiment of identical design using  Similar results were obtained in four separate experiments of identical design with concentrations of poly(A) RNase which varied from 0.08 to 0.8 rg per ml.  (27), it is possible that the presence of specific proteins associated with the mRNA was necessary for the action of the inhibitor. In order to evaluate this possibility, deproteinized lymphocyte poly(A)-rich RNA was prepared by pronase digestion of the mRNA as described under "Methods." When the deproteiniied mRNA was used as a substrate for poly(A) ribonuclease, the purified inhibitor was able to inhibit the digestion of the RNA at concentrations similar to those required for inhibition with nondeproteinized poly(A)rich RNA (Table V). Kinetic studies revealed quantitatively similar inhibition of poly(A) RNase digestion of deproteinized or nondeproteinized poly(A)-rich RNA by purified inhibitor (data not shown). Enzyme Speci$city of Inhibitor-Possible specificity for the inhibition of ribonuclease digestion of poly(A)-rich RNA was suggested by the failure of the purified inhibitor to prevent digestion of poly(A)-rich RNA by low concentrations of various other ribonucleases. Thus, digestion of poly(A)-rich RNA by Tr RNase, Tt RNase, or E. coli ribonuclease II was not inhibited by concentrations of inhibitor sufficient to quantitatively inhibit digestion by poly(A) ribonuclease (Table VI). The inhibitor similarly failed to prevent digestion of poly(A)-rich RNA by concentrations of pancreatic RNase sufficient to digest more than 5Oa/, of radiolabeled RNA (Table VI) ; at lower concentrations, however, there was partial inhibition (Table VI). In addition, the inhibitor did not prevent digestion of [*H]RNA 4 DNA hybrids by ribonuclease H (0.8 to 0.008 pg per ml) (data not shown).
Studies of Identity of Inhibitor-The experiments demonstrating lability of inhibitor suggested that it was not a polyamine or histone. Purified spermine, spermidine, or lysine-or argininerich histone were unable to inhibit digestion of mRNA by poly(A) RNase at concentrations effective for the inhibitor described above. At much higher concentrations, they partly inhibited digestion of synthetic poly(A) by poly(A) ribonuclease, another characteristic distinguishing them from the ribosomeassociated inhibitor. The kinetics of this inhibition revealed noncompetitive inhibition (data not shown). For these reasons, it is concluded that the ribosome-associated inhibitor described above is not a polyamine. Since the existence of a prokaryotic RNA ligase has recently been reported (49) and since such an enzymatic activity would explain much of the data presented here on "inhibition" of RNases, the inhibitor was tested for possible activity as an RNA l&se. Various concentrations of synthetic [r'C]polyriboadenylic acid (average size 4 to 5 S) were added to 100 ~1 of Buffer A containing [aH]poly(A)-rich RNA or 60 ng of [3H]poly(A)-rich RNA pretreated to prevent self-digestion, 0.08 pg of poly(A) ribonuclease, and various concentrations of inhibitor.
The molar concentration of the [14C]poly(A) was varied from 0.5-to 10-fold that of the [3H]poly(A)-rich RNA. Following incubation for 15 min at 37", the samples were adjusted to 0.3% SDS and analyzed by sedimentation through linear SDS sucrose density gradients.
No alteration of the sedimentation of W-labeled material from 4 to 5 S to heavier moieties characteristic of the lymphocyte poly(A)-rich RNA was observed. Addition of various concentrations of ATP (10m3 to 10d7 M), GTP (10m3 to lOA M), or MgCL (1 to 10 mM) had no effect on this result nor did use of poly(A) tracts prepared from lymphocyte poly(A)-rich RNA.

DISCUSSION
It appears that mRNA consists of two portions: one which serves an informational function, and the other, the poly (A) tract, a regulatory function.
If this is correct, the synthesis and cleavage of the poly(A) tract could regulate mRNA function at one or more of several post-transcriptional levels. The results presented above describe a factor dissociated from ribosomes by 0.5 M KC1 which inhibits digestion of poly(A)-rich RNA by two specific, recently described RNases (43).
Extensive purification of this inhibitor by salt fractionation, phosphocellulose chromatography, and rate-zonal centrifugation have permitted some definition of its properties.
The inhibitor was eluted from the polyribosome fraction of guinea pig adrenal cortex with no inhibitor activity detectable in the 100,000 X g supernatant fraction.
However, the large amount of other ribonuclease activities in the cytosol may well have obscured inhibitor activity.
Hence, the association of inhibitor with the ribosomal fraction does not prove conclusively that it is a true ribosome-associated protein. The inhibitor is a small (1.5 to 2 S) protein which specificallv inhibits two ribonuclease activities which digest the poly(Aj tract of mRNA. The first of these, isolated with lymphocyte poly(A)-rich RNA, is specific for the poly(A) tract (44); the second is a partially purified cytoplasmic endoribonuclease from bovine adrenal cortex which digests synthetic poly(A) as well as the poly (A) tract of lymphocyte mRNA. The latter ribonuclease shares many properties of an endoribonuclease identified in pig liver nuclei (40). Kinetic studies suggest a competitive mode of inhibition.
The inhibitor does not appear to represent an altered form or subunit of the poly(A) ribonuclease since it has extremely different physical properties and sediments at a markedly different s 20,W on linear glycerol density gradients. The purified inhibitor preparation does not contain RNA ligase activity nor can ribonuclease activity be demonstrated once the inhibitory activity is lost upon storage or freeze-thawing.
The inhibitor failed to prevent the digestion of synthetic poly(A) by poly(A) ribonuclease although it inhibited digestion of the poly(A) tract of poly(A)-rich RNA.
Since the method of isolation of poly(A) -rich RNA may fail to remove all associated proteins (27), one possible explanation is that these proteins are, in some way, required for inhibition.
However, deproteinization did not alter the ability of the inhibitor to prevent digestion by poly (A) ribonuclease.
This suggests that a portion of the mRNA other than the poly(A) tract may be required for expression of the inhibitor activity against poly (A) ribonuclease. Apparent enzyme specificity of the inhibitor is suggested by the failure of the purified inhibitor to prevent digestion of poly(A)rich RNA by a number of ribonucleases other than the associated RNase or poly(A) RNase. However, since (a) cleavage of the poly(A) tract is detected with an amplified sensitivity compared to digestion of other portions of the RNA because the assay method is based on retention of RNA on a nitrocellulose filter due to the properties of the poly(A) tract, and (b) since the poly-(A) ribonuclease has not been purified to homogeneity, the inhibitor may be present in relatively greater concentration with respect to actual poly (A) ribonuclease concentration compared to other ribonucleases tested.
At very low concentrations of pancreatic ribonuclease, producing less than 3% digestion of radiolabeled mRNA, addition of inhibitor did prevent release of radiolabeled RNA from the nitrocellulose filter.
Following the identification of cytoplasmic RNase inhibitors (50), several studies have been interpreted to suggest that these ribonuclease inhibitors are required to maintain the integrity of polyribosomes (51, 52) l The purification and characterization of these inhibitors have been quite difficult due to their characteristic lability, but partial purification has been achieved (53, 54). The partially purified inhibitor stabilized incubated polyribosomes suggesting that mRNA cleavage was entirely prevented (54). The possibility that induction of ribonuclease inhibitor activity is regulated was suggested by increased activity in the regenerating rat liver (55).
The existence of a protein capable of inhibiting the ribonuclease which digests the poly(A) tract suggests that eukaryotic cells could regulate removal of the poly(A) tract from mRNA in the cytoplasm.
Complexing of the ribonuclease with an inhibitor may be a mechanism commonly used in eukaryotic cells to control RNase activity; a previously described example is "alkaline RNase" activity (56-58).
It has been proposed that mRNA digestion is regulated by partition of the alkaline RNase between 7005 a soluble RNase inhibitor and a ribosomal site which could bind the enzymatic subunit (56, 58).
The demonstration of a poly(A) tract in viruses whose life cycle occurs entirely in the cytoplasm (19, 21) suggested a cytoplasmic role for the poly (A) tract. This possibility was given additional support by the demonstration that cytoplasmic polyadenylylation of pre-existing transcripts occurs in the sea urchin oocyte following fertilization and that this polyadenylylation of subribosomal RNA is associated with rapid translocation of the mRNA to the polyribosomal fraction (42, 57). Therefore, mRNA function or stability in the cytoplasm of eukaryotic cells might be affected by addition or removal of poly(A) tracts with regulation occurring at one or more of several levels.