Cordycepin AN INHIBITOR OF NEWLY SYNTHESIZED GLOBIN MESSENGER RNA*

The effect of cordycepin (3’-deoxyadenosine) on newly synthesized globin mRNA in cultured mouse fetal liver erythroid cells is investigated. At cordycepin concentrations that do not inhibit amino acid incorporation into acid-precipitable material, the quantity of pulse-labeled (radio-active) globin mRNA nucleotide sequences is reduced by 90%, as compared to adenosine-treated controls. The reduction of radioactivity in globin-specific RNA sequences is greater than the inhibition of total RNA synthesis in experiments in which the labeling times range from 6 to 60 min. Control experiments demonstrate that cordycepin does not reduce the recovery of total cell RNA or steady state (unlabeled) globin mRNA. The hybridization assay used to detect radioactive globin mRNA sequences is independent of the cellular location or the number of 3’4erminal adenylate residues in the mRNA-containing molecules. These data thus indicate that cordycepin inhibits newly synthesized mRNA as effectively as it inhibits ribosomal and transfer RNA synthesis. Of the several classes of eukaryotic RNA (rRNA, hnRNA,’ mRNA, and tRNA), only hnRNA and mRNA contain 3’-terminal poly(A) sequences Some hnRNA and mRNA

The effect of cordycepin (3'-deoxyadenosine) on newly synthesized globin mRNA in cultured mouse fetal liver erythroid cells is investigated. At cordycepin concentrations that do not inhibit amino acid incorporation into acidprecipitable material, the quantity of pulse-labeled (radioactive) globin mRNA nucleotide sequences is reduced by 90%, as compared to adenosine-treated controls. The reduction of radioactivity in globin-specific RNA sequences is greater than the inhibition of total RNA synthesis in experiments in which the labeling times range from 6 to 60 min. Control experiments demonstrate that cordycepin does not reduce the recovery of total cell RNA or steady state (unlabeled) globin mRNA. The hybridization assay used to detect radioactive globin mRNA sequences is independent of the cellular location or the number of 3'4erminal adenylate residues in the mRNA-containing molecules. These data thus indicate that cordycepin inhibits newly synthesized mRNA as effectively as it inhibits ribosomal and transfer RNA synthesis.
The role that poly(A) plays in mRNA function or metabolism has not been clearly defined. It may function in the cytoplasm since the mRNA of some viruses that replicate exclusively therein is polyadenylated (Johnston and Bose, 1972;Ehrenfeld, 19741, and adenylate residues are added onto maternal mRNA in the cytoplasm of sea urchin embryos @later Wilt, 1973). It is unlikely that poly(A) is required for translation since the initial rates of translation of adenylated and nonadenylated mRNAs are similar (Fromson and Verma, 1976;Gedamu and Dixon, 1976;Sippel et aZ., 1974; * This work was supported by Grant CA-07175 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Supported in part by Predoctoral Training Grant CA-5002. To whom corresoondence should be addressed. ' The abbreviations used are: hnRNA, heterogeneous nuclear RNA. defined here as that Dortion of nuclear RNA that is distinct from rRNA and tRNA; MFL, mouse fetal liver. Soreq et al., 1974;Nude1 et al., 1976). The poly(A) might retard the degradation of polysomal mRNA (Sippel, et al ., 1974;Soreq et al., 1974;Nude1 et al., 1976).
It has been suggested that poly (A) functions in the processing and/or transport of mRNA from the nucleus. Evidence supporting this hypothesis has come, in large part, from experiments with the adenosine analog cordycepin (3'-deoxyadenosine). Cordycepin was originally isolated from the liquid growth medium of the mold Cordyceps militaris. It is triphosphorylated in animal cells (Klenow, 1963) and, at sufficient concentrations, inhibits rRNA, tRNA (Siev et al., 1969), and protein synthesis (LaTorre and Perry, 1973) in intact cells. The following data support the hypothesis that post-transcriptional polyadenylation is essential for complete processing and/or transport of nuclear mRNA-containing molecules. 1) Cordycepin blocks post-transcriptional polyadenylation in the nucleus (Darnell et al., 1971a;Mendecki et al., 1972;Nakazato et al., 1974). 2) Under conditions in which nuclear polyadenylation is inhibited 75% or more, synthesis of hnRNA is either unaffected or is inhibited to a lesser extent (Siev et al ., 1969;Penman et al., 1970;Darnell et al., 1971a;Mendecki et al., 1972;Nakazato et al., 1974;Desrosiers et al., 1976). 3) The percentage of inhibition of newly synthesized (nucleus-derived) poly(A)-containing mRNA in polysomes is approximately the same as the inhibition of nuclear poly(A) (Mendecki et al., 1972;Nakazato et al., 1974;Milcarek et al., 1974); the accumulation of polysomal mRNA lacking poly(A) is also reduced, but to a lesser extent than is the poly(A)-containing mRNA (Milcarek et al., 1974). One interpretation of these data is that cordycepin inhibits the accumulation of newly synthesized polysomal mRNA by blocking polyadenylation of nuclear mRNA molecules or their precursors. The nonpolyadenylated molecules are then unable to be processed or transported.
This interpretation is supported by the fact that hnRNA synthesis is relatively resistant to cordycepin. However, this hypothesis makes the important, but unproved, assumption that cordycepin does not inhibit mRNA transcription, i.e. that the cordycepin-resistant hnRNA contains unprocessed mRNA molecules. Another interpretation is that total hnRNA consists of two distinct classes of RNA that respond differently to cordycepin: 1) mRNA molecules and their putative precursors; and 2) nonmessenger RNA molecules. The nonmessenger fraction, representing the vast majority of the hnRNA, might be resistant to cordycepin, while only the minor portion of hnRNA that is related to mRNA is inhibited.
This interpretation would Inhibition of Globin mRNA by Cordycepin 2629 account for the fact that cordycepin inhibits the accumulation of polysomal mRNA more so than hnRNA. Globin mRNA is transcribed initially into a precursor molecule that is 2-to 3-fold larger than globin mRNA and is cleaved to generate the mature molecule (Curtis and Weissmann, 1976;Ross, 1976;Kwan et al., 1977). This precursor is polyadenylated.2 Since cordycepin inhibits polyadenylation, it seemed feasible to exploit this analog to investigate the role of poly (A) in the cleavage, processing, and transport of a specific mRNA precursor.

MATERIALS AND METHODS
Details of the isolation and culture conditions for mouse fetal liver cells and the procedure for RNA isolation were as described (Ross, 1976). The liver of the 14-day-old mouse embrvo consists urimarilv of erythroid precursor cells (Paul et al., 1969): As primary cultures, these cells mature and synthesize appreciable quantities of adult mouse hemoglobin and globin mRNA (Chui et al., 1971;Ramirez et al., 1975;Ross, 1976). For isolation of total cell RNA, whole cells were lysed and extracted with phenol plus chloroformlisoamvl alcohol. The nucleic acids were then centrifuged in an isopycnic'cesium chloride gradient; the RNA pelleted to the bottom of the tube and the DNA remained in the gradient. The pellet was resuspended in buffer and concentrated by ethanol or trichloroacetic acid precipitation (see Ross, 1976 for details).
The hybridization methods for detecting labeled and unlabeled globin mRNA nucleotide sequences were as described (Ross, 1976  is that cordycepin inhibits newly synthesized globin mRNA nucleotide sequences even with relatively short labeling times. To determine the effect of different cordycepin concentrations, MFL cells were preincubated with adenosine or cordycepin for 30 min and were then labeled with 13Hluridine for 60 min. Total RNA and radioactive globin mRNA nucleotide sequences were quantitated. As expected, increasing concentrations of cordycepin reduced incorporation into total RNA (Table III). At concentrations of 1 pg/ml or greater, cordycepin also reduced the percentage of newly synthesized globin mRNA sequences (Table III). With as little as 3 yg/ml, the per cent hybridization was reduced approximately 2-fold, from 0.3 to 0.15%. Therefore, even at low concentrations, the MFL cells (lo? were incubated in 1 ml of growth medium for 75 min, at which time cordycepin or adenosine was added. The cells were incubated an additional 30 min, and then [5,6JH]uridine (200 &i/ml) and '"C-amino-acids (2 &i/ml) were added. After 60 min of labeling, 0.1 ml of cells was removed and trichloroacetic acidprecipitated to determine the uridine and amino acid incorporation. The remainder of the cultures was used to isolate total cell RNA, which was hybridized to unlabeled globin cDNA as described (Ross, 1976 synthesis (or accumulation) of globin mRNA nucleotide sequences is more sensitive to cordycepin than is total RNA synthesis.
One trivial explanation for these results was that cordycepin preferentially inhibited a specific size class of nuclear RNA. This seemed unlikely, based on sedimentation analysis of hnRNA from control uersus cordycepin-treated mammalian tissue culture cells (e.g. see Penman et al., 1970). Nevertheless, to determine the effect of cordycepin with MFL cells, total cell pulse-labeled RNA from treated or untreated cultures was sedimented in a sucrose gradient. Although incorporation was reduced in all size classes of RNA from cordycepin-treated cells, there was no selective decrease in the 4 to 28 S RNA (Fig. 2). There was slightly greater inhibition of 30 to 50 S RNA. This result indicates that cordycepin does not selectively inhibit the 10 to 18 S portion of the hnRNA, which contains globin mRNA and its precursor.  (Ross, 1976) at a concentration of 1.0 x 101/ml. After 75 min in culture, adenosine or cordycepin was added to a final concentration of 20 pg/ ml. After an additional 50 min, PH]uridine was added to a final concentration of 200 pCi/ml. After 10 min of labeling. the cells were harvested, and total cell RNA was extracted and-concentrated by ethanol precipitation. The RNA was pelleted out of ethanol and resuspended in 0.01 M Tris-Cl, pH 7.5, and 0.002 M EDTA, heated to 45°C ibr 2 min, rapidly chilled, and applied to a 15 to 30% linear sucrose gradient prepared in the same buffer. The gradients were centrifuged in a SW56 rotor, 35,000 rpm, 6 h, 5°C. Fractions of approximately 0.2 ml were collected dropwise by gravity. The absorbance at 260 nm of each fraction was measured to determine the positions of the internal ribosomal RNA markers (arrows), and the material was then counted directly after addition of 5 ml of RIA solution (Research Products, Incorporated). 0---0, adenosine; O-0, cordycepin.

DISCUSSION
Previous investigations revealed that cordycepin inhibited the appearance of newly synthesized, polyadenylated RNA molecules. However, these studies did not exclude the possibility that the analog was blocking both polyadenylation and structural gene transcription because the cordycepin effect was evaluated in terms of the quantity of total polyadenylated RNA being made. The inhibitory effect might have resulted from the following: 1) inhibition of polyadenylation of mRNA; 2) inhibition of transcription of structural genes; or 3) a combination of the above. The novel aspect of our study is that we have directly determined the effect of cordycepin on the appearance of the heteropolymeric portions of specific mRNA's. This approach permitted us to monitor more directly the effect of cordycepin on gene transcription.
These experiments demonstrate that cordycepin inhibits the transcription and/or accumulation of globin mRNA nucleotide sequences under conditions in which amino acid incorporation is not appreciably reduced. This effect is observed with labeling times as short as 6 min, and the per cent inhibition of total RNA synthesis is less than that of the globin mRNA sequences. Although these results do not unequivocally distinguish between reduced synthesis or accelerated degradation of the mRNA sequences, the data strongly support the interpretation that cordycepin inhibits mRNA transcription. This conclusion is consistent with the fact that cordycepin triphosphate inhibits transcription in vitro by prokaryotic RNA polymerases (Shigeura and Boxer, 1964;Maale et al., 1975) and by mammalian cell RNA polymerase II, either with purified DNA as a template or in isolated nuclei (Maale et al., 1975;Desrosiers et al ., 1976). Cordycepin also inhibits transcription from mitochondrial DNA in intact cells (Hirsch and Penman, 1974).
A major conclusion from these studies is that experiments with cordycepin in intact cells must be interpreted cautiously. Previous studies demonstrated that cordycepin inhibited both polyadenylation and the appearance of newly synthesized mRNA in polysomes, and it was concluded that polyadenylation was very likely required for mRNA processing and/or transport. This conclusion was strengthened by the observation that hnRNA synthesis was either unaffected or only slightly reduced by cordycepin. The data presented here demonstrate that cordycepin inhibits globin mRNA transcription and/or accumulation when the labeling times are as brief as or briefer than those in the above cited experiments. Therefore, the inhibition of polysomal mRNA might be due to inhibition of mRNA transcription, rather than inhibition of polyadenylation.
Three additional factors relevant to this point should be mentioned. 1) Since it exerts multiple effects in mammalian cells, cordycepin might inhibit the accumulation of polysomal mRNA by a combination of effects. For example, when HeLa cells were pulse-labeled for 7.5 min and then incubated (chased) with either actinomycin or actinomycin plus cordycepin, the quantity of radioactive polysomal mRNA was reduced by 70% in the actinomycin plus cordycepin cells, as compared with actinomycin alone . When the labeling time was 20 min, so that a greater proportion of the nuclear mRNA sequences was more completely processed, there was little, if any, reduction of labeled polysomal mRNA with cordycepin (Penman et al., 1970). These experiments provide good evidence that cordycepin blocks transport of preformed (and, presumably, incompletely processed) nuclear mRNA sequences to the cytoplasm. 2) Our conclusions are based on the assumption that the effect of cordycepin on mRNA synthesis is a general one, i.e. that it inhibits most, if not all, mRNA's. A major asset of the cDNA hybridization assay is the ability to detect newly synthesized (radioactive) mRNA sequences by a specific method that is independent of general structural features, such as the poly (A) or the cap. However, only two mRNA's (for a-and /3-globin) have been assayed so there is no proof that total mRNA transcription is inhibited by cordycepin. Nevertheless, there are no known structural features of globin mRNA that would promote selective inhibition by cordycepin. In most respects the properties of globin mRNA are similar to those of other mRNA's [Gould and Hamlyn, 19'73;Lim and Canellakis, 1970;Burr and Lingrel, 1971;Perry and Scherrer, 1975;see Dayhoff, 1972, for the primary structures of adult mouse globins). Therefore, there is no reason to suspect that the inhibition of globin mRNA synthesis by cordycepin is unique. The in vitro experiments support this assumption since cordycepin triphosphate is a potent inhibitor of transcription of total DNA by RNA polymerase II (Maale et al., 1975;Desrosiers et al., 1976). 3) We have not investigated the number of 3'-terminal adenylate residues or the binding to affinity columns (e.g. oligo(dT)cellulose) of RNA from cordycepin-treated cells. Our experiments were concerned solely with the effect of cordycepin on the heteropolymeric portion of globin mRNA molecules and their precursors. Experiments are in progress to determine whether the globin-specific RNA transcribed in the presence of cordycepin contains poly (A) or oligo(A) (Mendecki et al., 1972).