Cell, Vol. 6. 495-503,

August

1976, Copyright0

1976

by MIT

Biosynthesis and Stability of Globin mRNA in Friend Cells Cultured Erythroleukemic Haim Aviv, Zeev Voloch, Roberto Bastos, and Shoshana Levy Department of Virology The Weizmann Institute Rehovot, Israel

of Science

Summary Biosynthesis and stability of the mRNA population In DMSO-induced Frlend erythroleukemlc cells were studied after labeling the RNA with 3H-uridine and then chasing it with nonlabeled urldine. Globln RNA metabolism was studied by hybridization to excess complementary DNA covalently coupled to oligo(dT)-cellulose. After a labeling period of 120 min, 2-4% of the poly(A)-containing labeled RNA was In globin RNA; it decayed with a half-life of 16-17 hr. The rest of the poly(A)-containing RNA was composed of two kinetic populations: 65-90% decayed with a half-life of about 3 hr, while 10% decayed with a half-life of about 37 hr. The portion of globln RNA in labeled poly(A)-containing RNA behaved in an unexpected fashlon during the chase period. During the initial chase period, the percentage of globin RNA increased rapidly, reaching a maximum of about 15% at 20 hr, but it subsequently declined gradually. Based on these findings, a model was built that describes the changes In the propotilon of globln mRNA In poly(A)-containing RNA during contlnuous synthesis and after chase of the labeled RNA. It appears that if the parameters described remaln constant during the maturatlon of erythroblasts, then this model would not account for the almost exclusive presence of globln RNA In the reticulocyte. By far the most effective way to achieve this hlgh level of globin RNA is the destabllizatlon of the mRNA population which is more stable than globln RNA, and not the stabilization of globln RNA itself. Introduction Cellular differentiation in eucaryotic tissues is often characterized by the production of a relatively high abundance of a specific protein (or group of proteins). Although the factors determining this differentiation process are not known, there appears to be an enrichment of the messenger RNAs which code for the differentiation-specific proteins. Thus in reticulocytes, over 90% of the protein synthesized is hemoglobin with globin mRNA comprising a similar proportion of total cellular mRNA (Kruh and Borsook, 1956; Schweet, Lamfrom, and Allen, 1956; Lockard and Lingrel, 1969; Aviv and Leder, 1972).

Such enrichment of specific mRNA in the differentiated cells can result from either transcriptional or post-transcriptional controls. Transcriptional regulation can express itself positively by acceleration of the transcription of specific genes, or negatively by selective suppression of all genes unrelated to differentiation. Post-transcriptional regulation can be carried out either by degradation in the nucleus of the mRNA species not required for differentiation or by differential stability of messenger RNAs in the cytoplasm, the net result being a progressive accumulation of one specific class of mRNA with time (Kafatos, 1972). It is obvious that the proportion of a particular mRNA in the total mRNA population is not only a function of its own half-life, but also of the half-lives of all the other mRNA species. Indeed, the possibility that control of differentiation may be related to the relative stabilities of mRNA was suggested long ago (Marks, Burka, and Schlesinger, 1962; Rifkind, Danon, and Marks, 1964; Stewart and Papaconstantinou, 1967; Fantoni et al., 1968; Kafatos, 1972; Marks and Rifkind, 1973), but a kinetic model of this process was only recently described by Kafatos (1972). Until recently, however, direct investigation of the metabolism of specific mRNAs was hampered by the lack of suitable techniques directly measuring specific mRNA synthesis and degradation. An important step toward solving this problem has been the purification of several eucaryotic mRNA species and the synthesis of their complementary DNAs (cDNA) (Ross et al., 1972a; Kacian et al., 1972; Verma et al., 1972). These techniques enabled measurements of mRNA accumulation (Ross, Ikawa, and Leder, 197213; Ross et al., 1974; Leder et al., 1973; Preisler et al., 1973; Gilmour et al., 1974). Nevertheless, the details of the more dynamic aspect of mRNA metabolism (that is, the rates of mRNA synthesis and degradation) could not be studied for several reasons. First, quantitative measurements of mRNA synthesis and degradation require conditions of excess DNA hybridization, which require large quantities of cDNA. Second, the background levels of contaminating labeled RNA are too high to allow the direct measurement of the synthesis and degradation of an mRNA coded by a unique gene. We have previously described the use of cDNA covalently bound to cellulose; in the globin system, the background level was 0.02% (Levy and Aviv, 1976). This procedure thus makes it possible to overcome the difficulties involved in measuring directly the synthesis and degradation of globin mRNA. In the present study, we have measured directly the synthesis and degradation of globin mRNA as well as that of total poly(A)-containing RNA in

Cell 496

DMSO-treated Friend erythroleukemic cells. The half-life of globin mRNA is about 16-17 hr, while the half-life of about 90% of poly(A)+ RNA is as short as 3 hr. The remaining poly(A)+ RNA, however, shows greater stability than globin mRNA. Here we discuss the contribution of this stable fraction in limiting the relative accumulation of globin RNA in Friend cells, and propose a model which quantitatively describes the kinetics of globin mRNA accumulation in proerythroblasts and reticulocytes.

Results Synthesis of Globin RNA The “induction” of globin synthesis and the accumulation of globin RNA observed in DMSO-treated Friend cells (Friend et al., 1971; Ross et al., 1972b, 1974) need not necessarily result from the transcriptional activation of the globin gene. Posttranscriptional stabilization of globin RNA can also play a role in its accumulation. In theory, the contribution from each of these mechanisms can be evaluated experimentally by direct measurements of globin RNA synthesis and degradation. The effect of l-4 days of treatment with DMSO on the rate of globin RNA synthesis was studied in Friend cells after labeling with 3H-uridine for 120 min. RNA was extracted and hybridized to excess globin cDNA covalently bound to cellulose (Venetianer and Leder, 1974; Levy and Aviv, 1976). Even using this technique, we could not measure the synthesis of globin RNA in the nucleus. 3H-RNA extracted from the nuclei of DMSO-treated cells labeled for 120 min does not show any measurable level of globin RNA synthesis above the background of 0.02%. Labeled globin RNA can be detected in the cytoplasm of FLC on the second day after DMSO treatment. As shown in Figure 1, levels of 0.06% were found at that time, 0.10% on the third day, and 0.20% on the fourth day (Figure 1). When newly synthesized cytoplasmic RNA was fractionated on oligo(dT)-cellulose, 6-10% of the labeled RNA was shown to contain poly(A). The amount of newly synthesized globin RNA in total labeled poly(A)+ RNA increased after DMSO treatment in a pattern similar to that of globin RNA in total RNA, except that the proportion of labeled globin RNA was about 10 fold higher in the former (Figure 1). At least 90% of globin RNA synthesized during the 2 hr labeling period contained poly(A). In the experiment represented in Figure 1, the percentage of labeled globin RNA in poly(A)+ RNA was 1.4% after 4 days in DMSO; however, in many other experiments, globin RNA in poly(A)+ RNA reached 3.54% after 4 or 5 days in DMSO. It should be stressed that the appearance of newly synthesized

globin RNA in the cytoplasm of DMSO-treated cells does not necessarily imply that DMSO induces the transcription of the globin gene. DMSO may merely stabilize globin RNA which is being continuously synthesized and subsequently broken down either in the nucleus or in the cytoplasm.

Stability of Globln RNA The fact that the net accumulation of globin RNA is the result of simultaneous synthesis and degradation required the study of the stability of globin RNA. Cells exposed to DMSO for 4 or 5 days were labeled with 3H-uridine for 2 hr and then chased by transferring them to a medium containing 10 mM uridine and 5 mM cytidine. The insert in Figure 2 shows I

I

1.4

I

Poly(A)

I

I

I

containing RNA

t

1.21.00.8 0.6 -

3

Cytoplasmic

Time Figure

1. Biosynthesis

of Globin

1

RNA

in hours RNA upon

DMSO

Treatment

Lower panel: cells cultured in the presence of DMSO (O-4 days) were labeled for 120 min with ‘H-uridine (50 pc/ml). Radioactivity incorporated varied between 5 and 7 x 105 cpm/lOb cells. About 1 O6 cpm (1.3 AzbO units) cytoplasmic RNA were hybridized to cDNAceltutose (estimated 1 pg of cDNA) as described in Experimental Procedures. Upper panel: poly(A)+ RNA was separated from poly(A)RNA on oligo(dT)-cellulose. About &IO% of the RNA labeled in 120 min containedpoly(A), wlthlittledifferencefound betweenDMSO-treated and control cells. Nonspecific binding of poly(A)RNA to oligo(dT)cellulose was about 0.1%. l-2 x lo5 cpm (0.05 A260 units) were hybridized to cDNA-cellulose as in the lower panel. The percentage of labeled RNA extracted from DMSO-treated cells which hybridized with cDNA is plotted.

Messenger 497

RNA

Metabolism

and Differentiation

the efficacy of the chase conditions used. RNA extracted from the cytoplasm was hybridized to excess cDNA-cellulose as described in Experimental Procedures. In Figure 2, the relative proportion of labeled globin RNA in total labeled cytoplasmic RNA is shown at different times after chase. Since most of these labeled cytoplasmic RNAs are rRNA and tRNA, which are rather stable molecules (Loeb, Howell, and Tompkins, 1965; Emerson, 1971; Abelson et al., 1974), we find it convenient to express the decay of poly(A)+ RNA as a ratio of this species to total labeled RNA. The pattern is very similar if we plot the amount of labeled poly(A)+ RNA per cell. The decay of globin RNA is exponential with a half-life (tl/,) of 17&l hr (Figure 2, closed circles). The proportion of globin mRNA in the messenger population-which is ultimately the crucial factor in the expression of the globin gene in erythroid dif-

I

I IO

I

1

1

Figure 2. Decay Kinetics (Labeling Period 2 Hr)

of Labeled

Globin

I

I

a:

RNA and Poly(A)+

RNA

20 30 40 Time after chase(hours)

50

Cells seeded at 4 X lOd/ml were treated for 4.5 days with 1.7% DMSO with one change of medium 14 hr before labeling. The culture reached 2 x lo6 cells per ml and about 90% of the cells were benzidine-positive. Cells were labeled for 120 min and chased as described in Experimental Procedures. The efficiency of the chase treatment is shown in the insert. About 106 cpm (1.2 Azao units) of total cytoplasmic RNA were hybridized with cDNA-cellulose under the standard conditions. At zero time &), 0.4% of the counts hybridized to globin cDNA. This value (t. of chase) is taken as 100% for calculation of the relative proportion of globin IH-RNA. A background level of 0.02%, which was obtained by using XH-RNA from mouse lymphoma cells, was subtracted. The percentage of poly(A)+ RNA to t, in this experiment was 9.5%, which served as 100% for calculations of relative ratio of poly(A)+ to poly(A)RNA. (M) decay kinetics of globin RNA; (O-O) decay kinetics of poly(A)+ RNA.

ferentiation-is determined also by the rate of synthesis and stability of nonglobin mRNA as is obvious from the analysis of Kafatos (1972). The best experimental measurement, although not ideal (Milcarek, Price, and Penman, 1974), for total mRNA sequences makes use of their poly(A) content. The variation in the amount of labeled poly(A)+ RNA in DMSO-treated Friend cells with time after chase is shown in Figure 2 (open circles). Two major components of poly(A)+ RNA with different half-lives exist. The major component (Ml), which constitutes about 65-90% of the population, is degraded very rapidly (ts = 3 hr). However, a component (M2), constituting about lo-15% of the population, is much more stable than globin RNA, and has a half-life of about 35-40 hr. Poly(A)+ RNA was further characterized using sucrose density gradients and appears to be in the range of 6S-16s (Figure 3). The size distribution of the labeled poly(A)+ RNA did not change even after a chase of 50 hr. It also should be pointed out that the preparation of poly(A)+ RNA using the method described in Experimental Procedures yields poly(A)+ RNA which is very lightly contaminated by tRNA and rRNA (Figure 3). That poly(A)+ RNA is not heavily contaminated with rRNA and tRNA was also shown by recycling of poly(A)+ RNA on oligo(dT)-cellulose. While only about 0.1% of rRNA binds to oligo(dT)-cellulose under the conditions used here, about 80-90% of poly(A)+ RNA binds (data not shown). The synthesis and degradation of globin RNA and poly(A)+ RNA can be formulated using simple kinetic equations (see Discussion), which describe the proportion of globin RNA in poly(A)+ RNA as a function of the duration of chase. In the initial period, since Ml decays faster than globin RNA, a rather rapid increase in the proportion of globin RNA will be observed; subsequently, since globin RNA is degraded faster than M2, a gradual decrease in the percentage of globin RNA will be observed. This pattern was indeed observed when the percentage of labeled globin RNA in poly(A)+ RNA was measured experimentally (Figure 4). Moreover, a mathematical analysis of the kinetics using computer simulation shows that the experimental data in Figure 4 agreed closely with the theoretical curve (not shown) calculated using the parameters obtained from Figure 2. This analysis predicts that in cells labeled for 40 hr (rather than 2 hr), however, the proportion of M2 in the poly(A)+ RNA population should reach a theoretical value of 36%. This proportion was indeed found when cells were labeled for 40 hr with 3Huridine and subsequently chased (Figure 5). It can also be seen in Figure 5 that the half-lives of Ml and M2 were about 3 and 35 hr, respectively, while

Cell 498

I-

4s

18S

28s

20 N ‘0; a v

16 12 8

E

the half-life of globin RNA was 15-16 hr, in good agreement with the estimates from Figure 2. It should be stressed that these parameters of globin and total mRNA synthesis and degradation are not unique to DMSO-treated Friend cells in culture, which may have peculiar properties since they are actively producing C type particles (Friend, Preisler, and Scher, 1974). Recent studies in our laboratory indicate that mRNA metabolism in spleen cells from anemic mice has very similar parameters (R. Bastos, Z. Voloch, and H. Aviv, manuscript submitted for publication).

Dlscussion In the process of erythroid differentiation, ic series of changes take place, starting

I Fraction number Figure

3. Size Distribution

of Polv(A)+

RNA

“H-RNA was prepared as dePoly(A)-containing cytoplasmic scribed in the legend to Figure 2. RNA was fractionated by a 5-20% sucrose gradient (4 hr at 50,000 rpm in a Beckman SW-50 rotor at 4°C). (A) 104 cpm poly(A)+ ‘H-RNA from cells labeled for 120 min. (B) 104 cpm of 3H-RNA labeled as in (A), but chased for 23 hr as described in Experimental Procedures. (C) 5 x 103 cpm of SHRNA labeled as in (A), but chased for 49 hr. In every gradient, an equal amount of total cytoplasmic RNA labeled with 32P for a long time (24 hr) was included. (U) 3H cpm per fraction; (o--O) QP cpm per fraction.

I

I

a dramatfrom the

I

IO 20 30 40 Time after chase{ hours) Figure 4. Kinetics of Globin ‘H-RNA after Chase

IH-RNA

Accumulation

in Poly(A)+

l-2 x 106 cpm of total cytoplasmic RNA obtained as in 2 were separated into poly(A)+ and poly(A)RNA. Input for hybridization ranged between 4000 to 20,000 cpm. The at time zero (t) of chase was hybridized using 100 and of cDNA-cellulose suspension (Figure 7) and the value of ization was 3.8% of the input RNA.

Figure counts sample 200 pl hybrid-

Messenger 499

RNA

Metabolism

and

Differentiation

stem cells-where globin mRNA is absent-through the erythroblastic stages-where globin mRNA is present but constitutes only a rather small fraction of a heterogeneous population of messengers-and ending in the reticulocyte, whose mRNA is composed almost exclusively by that coding for the (Y and ,l? chains of hemoglobin (see Introduction). Due to experimental limitations, it is as yet impossible to define the initiation of globin RNA synthesis in terms of either transcriptional activation or stabilization of globin RNA sequences in the nucleus. Studies are in progress to reduce further the background of the present hybridization technique. In this study, we have used pulse and pulse-chase labeling techniques to investigate the metabolism of globin mRNA as well as nonglobin poly(A)+ RNA in DMSO-treated Friend cells. 2-4% of the cytoplasmic poly(A)+ RNA labeled in 2 hr was globin mRNA,

$ .06-

oz

.04-

I

Figure 5. Decay Period 40 Hr)

I IO Kinetics

I

I I 20 30 40 50 Time after chase (hours) of Globin

I

RNA and Poly(A)+

-,04 B

RNA (Labeling

Cells at a concentration of 4 x lOa/ml were seeded in the presence of 1.6% DMSO. After 3 days (106 cells per ml, 37% S+ cells), 40 pc/ml of IH-uridine (2.2 Ci/mmole) were added. 40 hr later (2.106 cells per ml, 90% Bf). the cells were washed and chased by further incubation in fresh medium containing 10 mM uridine and 5 mM cytidine. Incorporation was linear up to 30 hr (6 x 105 cpm/lOb ceils). 4 x 106 cpm of cytoplasmic RNA were hybridized to cDNA in the standard conditions. At zero time of chase, 0.09% of the RNA was hybridized, and this value was used as 100% in the calculation of the relative proportion of globin in cytoplasmic RNA. The background (Figure 2, legend) value of 0.02% was subtracted from all experimental results. The decay of poly(A)+ RNA labeled for 40 hr is also shown, about 3.2% of cytoplasmic XH-RNA containing poly(A)+ RNA in this experiment. (M) kinetics of decay of globin RNA; (0-O) kinetics of decay of poly(A)+ RNA.

which decayed with a half-life of about 17 hr. 8090% of the labeled poly(A)+ RNA (Ml) had a much shorter half-life-t,/, being as short as 3 hr (Figures 2 and 5). Surprisingly, however, another fraction of poly(A)+ RNA exists (M2) which is twice as stable as globin RNA-its tl/, being about 35 hr. As discussed below, we believe that this fraction of nonglobin poly(A)+ RNA is responsible for the low percentage of globin RNA measured in erythroblastic cells. The parameters of degradation, after chase, can be represented by simple kinetic equations, as follows: Ml

= Ml.e-kit;

M2 = M2,e-W;

G = GOe-W

where Ml, M2, and G are the concentrations of the three classes of mRNA; Ml,, M2,, and G, are their initial concentrations at the zero time of chase; and k,, kZ, k1 are the degradation rate constants of the three classes of RNA. The relationship of k to halflife is given by k = In 2/t%. The percentage of globin RNA (%G) in the total mRNA population at .. any time IS by defmmon %G = M,GxlOO + M2+G. A detailed analysis of the time function %G will be published later (R. Bastos and H. Aviv, manuscript submitted for publication). However, we wish to stress several points. First, when both synthesis and degradation of RNA proceed simultaneously, globin RNA will reach a steady state level of only lo-2096 of poly(A)+ RNA. This concentration was indeed observed experimentally using labeled cDNA and excess RNA hybridization (Ross et al., 1974; Friend et al., 1974; Gilmour et al., 1974; our laboratory, unpublished data). Second, when synthesis of total RNA stops, during maturation of erythroblasts-or when labeled RNA is chased, as in the experiments described here-if k, > kj > kz, under most conditions of Ml,,, M2,, and G,, a plot of %G as a function of time goes through a maximum, the maximal value being much lower than 50%. The significance of this lies in the fact that if one assumes that the kinetic parameters of mRNA metabolism operating in Friend cells as described above and in spleen cells from anemic mice (R. Bastos et al., manuscript submitted for publication) remain constant during reticulocytosis, the proportion of globin RNA in the cell can never reach the exceedingly high values of 95%. We therefore propose that during reticulocytos/s a drastic change in mRNA metabolism needs to take place which enables the cell to synthesize hemoglobin almost exclusively. It is interesting to speculate on possible changes in mRNA metabolism which may bring about such a drastic change within a limited span of time (the estimated time interval between orthochromatic proerythroblast and reticulocyte stages is believed to be in the range of 20-40 hr; Stohlman, 1970).

Cell 500

The effect of changing the kinetic parameters tl/, (Ml), tx (M2), and t% (G) on the accumulation of globin RNA after chase (or after RNA synthesis stops) was carried out with the aid of a computer (Figure 6). It is clear from Figure 6A that the level of globin RNA in poly(A)+ RNA changes very little after chase, even if the half-life of Ml (the unstable fraction of mRNA) is as short as 0.5 hr. Furthermore, stabilization of globin mRNA is insufficient to raise its concentration significantly (Figure 6C). If the t, of globin RNA were to increase to 80 hr, even 100 hr after chase, globin RNA concentration would reach only 50%. In contrast, when the tl/, of M2 (the stable fraction of mRNA) is scanned, the I

11 lnrtial rrlativs

proportion { $:O:y

t n’l7hr

,c

$zoi!f!+= : c .

fixed tl,2(MI)=3hr

60

t ,/2(M2)=37h t&G)scanning

25

50

75

Ti me of chase( hours) Figure 6. Computer Analysis of the Accumulation of Globin in Poly(A)+ RNA as a Function of Time after Chase

RNA

Initial proportions of Ml:M2 and G (0.5:0.4:0.1) were calculated from steady state conditions using the following parameters: rates of relative synthesis of Ml = 0.66, M2 = 0.10, G = 0.04; half-life of the corresponding RNAs in this calculation were Ml = 3 hr, M2 = 37 hr, and G = 17 hr. The effect of the following changes of t1,2 on the percentage of globin RNA accumulated in poly(A)+ RNA after chase or RNA synthesis shut-off were calculated: (A) t,,z of Ml ranged from 0.5 to 3.5 hr; (6) t,,> of M2 ranged from 4 to 36 hr; (C) t,,x of globin RNA ranged from 16 to 60 hr.

effect on the percentage of globin RNA is dramatic (Figure 6B). Thus when tK (M2) is reduced to 4 hr, the concentration of globin RNA will very rapidly approach 100%. Note that the initial concentrations of Ml, M2, and G are assumed to be their steady state values, 0.5, 0.4, and 0.1, respectively. These values were calculated from the synthesis and degradation parameters measured in Friend cells and anemic spleen cells. But our essential conclusion, which emphasizes the importance of M2 in preventing the accumulation of globin RNA, is not significantly altered if somewhat different initial concentrations are assumed. Thus even if M2 is only lo%, Ml 80%, and G 10% at the time when RNA synthesis stops, the destabilization of M2 would still be by far the most efficient way to concentrate globin RNA in the messenger RNA fraction of the cell. In short, we propose that during the maturation of erythroblasts to form reticulocytes, a change in the stability of mRNA should take place. It appears that the most efficient way to yield a high proportion of globin RNA in the mRNA population of reticulocytes is to destabilize that fraction of mRNA (M2) which is more stable than globin RNA. This model could be criticized for several reasons. First, it is not clear whether the three different populations of mRNA are present in every cell in the culture. Although 80-90% of the cells studied were benzidine-positive after DMSO treatment (and thus probably synthesizing hemoglobin), the M2 population may be present in the remaining cells only. This possibility is improbable because M2 comprises only 10% of poly(A)+ RNA in noninduced cells (unpublished data). Another criticism that may be raised is that our measurements focus on poly(A)-containing RNA, while as much as 30% of mRNA may lack poly(A) (Milcarek et al., 1974). We do not believe this will seriously affect our proposed model. During the course of erythroid cell maturation from proerythroblasts to reticulocytes, the percentage of globin mRNA in poly(A)-containing RNA changes from about 10% to over 95%; it therefore appears that the most important changes in the parameters of mRNA metabolism should be those connected with the poly(A)+ fraction of RNA. Changes in the kinetics of poly(A)RNA may be interesting to analyze; however, their relevance to our model should be minimal. The following questions relating to the importance of mRNA stability as expressed during erythropoiesis can be raised. What determines the stability of mRNA? In particular, what is the reason for the existence of a variety of classes of mRNA having different half-lives? A similar two-class distribution of mRNA population was found in a variety of cells (Singer and Penman, 1970; Pucket, Chambers, and Darnell, 1975; R. H. Singer, G.

Messenger 501

RNA

Metabolism

and Differentiation

Kessler, and D. Yaffe, personal communication). The two classes of mRNA may represent different mRNA sequences or similar RNA sequences having a different half-life. Furthermore, is the half-life a function of “environment” (that is, cytoplasmic factors which are mRNA-specific) or, alternatively, is stability determined by the nature of mRNA itself? If a selective destabilization of poly(A)+ mRNA does take place during reticulocytosis, what is its mechanism? Does the stability of globin mRNA change during differentiation? From a theoretical point of view (Kafatos, 1972) as well as from our own studies, it is obvious that the stability of other mRNAs is at least as important in determining the effective concentration of globin mRNA as is the stability of globin mRNA itself. It remains to be seen whether or not similar mRNA classes are present and have a role in controlling differentiation-specific mRNA in other systems. Furthermore, erythropoietin is believed to act at two levels, accelerating the replication of erythropoietin-sensitive cells and raising the level of globin RNA in proerythroblasts (Marks et al., 1962; Cantor et al., 1972; Terada et al., 1972; Marks and Rifkind, 1973; Conkie et al., 1975). The effect of erythropoietin on transcriptional activation of the globin genes or on the relative stabilities of mRNAs is not known. The approach described here may be useful in answering these questions. Expertmental

Procedures

Cell Growth and lnductlon by DMSO Friend erythroleukemic cells (745) were obtained from M. Adesnik and subcloned in our laboratory. The cells were inoculated in petri dishes (10 cm) at 105 cells per ml, or as specified in the legends, in Dulbecco’s modified Eagle’s medium (Bio Lab, Jerusalem), with a lower concentration of CaCI, (0.4 mM) and MgSOd (0.15 mM), supplemented with 15% (heated at 56°C for 30 min) fetal calf serum (GIBCO), and incubated in a humid CO? incubator (5-6% COZ) at 37°C. Induction was carried out by the addition of dimethylsulfoxide (DMSO) (Fluka) to a final concentration of 1.5-1.7%. The growth of the cells was only slightly inhibited by DMSO and reached a final concentration of 2 x 106 cells per ml. Benzidine tests were performed on intact cell suspensions as described by Orkin, Harosi. and Leder (1975). After 4-6 days in DMSO, 80-90% of the cells are B+. Pulse Labellng of RNA and Chase Approximately 1 O* cells were collected by centrifugation (1000 x g. 10 min) and resuspended in 20 ml of prewarmed fresh medium: 1 mCi of 5, 6-‘H-uridine (45 Ci/mmole; Amersham) was added. Incorporation was linear for 7-8 hr. After 120 min, cells were collected and washed twice with 20 ml of medium. For chase experiments, the cells were then resuspended in fresh medium containing 10 mM uridine and 5 mM cytidine. As seen in the insert to Figure 2, incorporation in cytoplasmic RNA continued for a short time after the start of the chase, then leveled off and slowly declined. If actinomycin D was used to chase the SH-uridine, incorporation into RNA ceased rapidly. After about 10 hr, however, the amount of labeled RNA dropped sharply, probably as a result of cell death and RNA decomposition. Actinomycin D could therefore not be used in experiments when long chase times were required.

Preparation of RNA At the end of the labeling period, the cells were collected by centrifugation as above and washed once in 0.01 M Tris-HCI (pH 7.4) containing 0.15 M NaCI. Nuclei were separated from cytoplasm by gentle mixing of the cells in a solution containing 0.25 M sucrose, 0.15 M NaCI. 3 mM MgCb, 10 mM Hepes (pH 8.0) and 0.5% Triton X-100, and centrifuged for 20 min. The nuclear pellet was washed once more with the same buffer without Triton X-100. The combined supernatants of the two washes contain the cytoplasm of the cells. Nuclear RNA was processed according to Penman (1966) with some modifications. After DNAase digestion, the buffer was adjusted to 0.25 M NaCI, 25 mM MgCb, 10 mM Tris (pH 7.4) 0.5% SDS, and 50 Fg/gl proteinase K were added. Proteolysis proceeded for 15 min at 37°C. The solution was extracted once with phenol at 60°C and the aqueous phase was extracted 3 more times with chloroform containing 1% isoamyl alcohol at room temperature. Ethanol precipitation and a second DNAase digestion were performed as described by Penman (1966). Cytoplasm/c RNA was extracted several times with phenol:chloroform (1:l) in 0.1 M Tris-Cl (pH 9.0) at room temperature (Aviv and Leder, 1972). PO/y(A)-contahlng RNA was purified as described (Aviv and Leder, 1972), except for the following modifications (Singer and Penman, 1973): 0.1% SDS was present in all solutions, and 100 pg of tRNA (yeast; Sigma) were added to the sample [0.3-0.5 ml packed oligo(dT)-cellulose]. Mouse-glob/n IW-RNA was prepared by the Commerford procedure (Commerford, 1971) as described elsewhere (Levy and Aviv, 1976). Preparallon of cDNA-cellulose was essentially by the method of Venetianer and Leder (1974). Hybridization of Labeled RNA wlth cDNA-Cellulose The detailed conditions for hybridization and elution of labeled RNA were described elsewhere (Levy and Aviv, 1976). Briefly, the reaction (0.5 ml) contained 50% deionized formamide (Fluka), 0.6 M NaCI, 0.01 M Tris-HCI (pH 7.5), 0.1% SDS, 1 mM EDTA, 50 pg poly(A), 100 pg yeast tRNA, and labeled RNA and cDNA-cellulose I

I

50

1

I

I

I

100 150 Globin cDNA-cellulose(@)

Figure 7. Hybridization of Total Cytoplasmic Concentrations of Globin cDNA-Cellulose

SH-RNA

I

I

200 to Different

7 X 105 cpm )H-RNA from 5 day induced cells were hybridized to increasing amounts of globin cDNA-cellulose at the standard conditions. The value of cDNA-cellulose is given in gl of cellulose suspension (1:3) because the exact amount of cDNA is difficult to determine. Our estimate is that 100 pl of this cDNA suspension are equivalent to about 1 pg of cDNA.

Cell 502

Received

January

23, 1976;

revised

April 29, 1976

References Abelson, H. T., Johnson, Cell 1, 161-165. Aviv, H.. and 1408-1412.

Leder.

L. F., Penman, P. (1972).

S., and Green,

Proc.

Nat.

Acad.

Sci.

Cantor, L. N., Morris, A. J., Marks, P. A., and Rifkind. Proc. Nat. Acad. Sci. USA 69, 1337-1341. Commerford,

S. L. (1971).

Biochemistry

Conkie, D., Kleiman, L., Harrison, Cell Res. 93, 315-324. Emerson, RNA

added (pg)

Figure 8. Competition Hybridization of Mouse Globin and Cytoplasmic ‘H-RNA by Nonlabeled Mouse and Globin RNA

rz+RNA Chicken

68,000 cpm (0.1 1.19) 9s mouse globin ‘ZSI-RNA (u) purified on oligo(dT)-cellulose or 250,000 cpm of cytoplasmic ‘H-RNA (0-O) from 5 day DMSO-induced cells were hybridized to globin cDNA-cellulose in the presence of increasing concentrations of unlabeled 9s mouse globin RNA. 0.21% of the “H-RNA and 73% of rz5l-RNA were hybridized to cDNA without competitor. These values are taken as 100% on the ordinate. As a control, chicken 9s globin RNA was used as a competitor (A-A). A background value of 0.02% was subtracted from IH-labeled, hybridized RNA. as specified. Hybridization was carried out in scintillation vials shaking in a water bath at 40°C. To ensure an excess of cDNA-cellulose, we determined a concentration curve for each preparation of cDNA with in vivo labeled ‘H-RNA as shown in Figure 7. In every experiment, the validity of the saturation curve was tested by using 2 vol (100 and 200 pl) of a 1:3 cDNA-cellulose:water suspension. cDNA and globin RNA concentrations were determined by a competition hybridization technique, which will be published later (R. Bastos et al., manuscript in preparation). Specificity of the hybridization assay was determined by the following criteria. Purified mouse globin 9s r25l-labeled RNA was hybridized to cDNA-cellulose in the presence of increasing amounts of nonlabeled RNA (Figure 8). In another series of experiments, in vivo labeled cytoplasmic ‘H-RNA was hybridized. The competition curve of both the purified globin 1251and of ‘H-RNA by unlabeled 9s RNA were identical to each other (Figure 8). On the other hand, competition by 9s globin RNA from chicken reticulocytes was negligible (Figure 8). If we assume that poly(A)-containing 9s RNA from reticulocytes is pure globin mRNA, it follows that the “H-RNA hybridized to globin cDNA is globin RNA. In addition, when poly(A)-containing ‘H-RNA was melted from cDNA and rehybridized to cDNA-cellulose, the percentage of hybridized, labeled RNA was increased from 3% on the first cycle to about 70% on the second cycle (not shown). The T, of the hybrid between XH-RNA and cDNA-cellulose was 66°C under the condition tested (0.1 M NaCI), and is very similar to that of purified mouse globin 1251-RNA (64°C) indicating that the hybrids of IH-labeled RNA with cDNA-cellulose are composed of long nucleotide sequences. Acknowledgments We thank Drs. M. Revel, U. Littauer, C. Prives, and H. Ozer for helpful suggestions; Mrs. Y. Bernstein for technical assistance; and Mr. E. Tepper for editing the manuscript. We thank Drs. J. Gruber and J. Beard for AMV reverse transcriptase. This work is supported in part by a contract from the National Cancer Institute.

C. P., Jr. (1971).

H. (1974). USA

R. A. (1972).

70, 1993-1999.

P. R., and Paul, J. (1975).

Nature

69,

New Biol. 232,

Fantoni, A., De La Chapell, A., Rifkind, (1968). J. Mol. Biol. 33, 79-91.

Exp.

101-106.

R. A., and Marks,

Friend, C., Scher. W., Holland, J. G., and Sato, Nat. Acad. Sci. USA 68, 378-382.

P. A.

T. (1971).

Proc.

Friend, C., Preisler, H. D., and Scher, W. (1974). In Current Topics in Developmental Biology, 8, A. A. Moscona and A. Monroy, eds. (New York: Academic Press), p, 81. Gilmour, R. S., Harrison, P. R.. Windass, J. D.. Affara, Paul, J. (1974). Cell Differentiation 3, 9-22. Kacian, D. L., Spiegelman, J. A., Dow, L., and Marks, 167-169. Kafatos. scription Kruh,

S., Bank, A., Terada, M., Metafora, P. A. (1972). Nature New Biol. 235,

F. C. (1972). Fifth Karolinska Symposium in Reproductive Tissue, p. 319. J., and Borsook,

H. (1956).

S., and Aviv,

Lockard, Commun.

H. (1976).

R. E., and Lingrel, 37, 204-212.

Loeb, J. N., Howell, 1093-l 095. Marks,

Biochemistry,

in press.

J. B. (1969).

Biochem.

Milcarek,

C., Price,

R. A. (1973).

R., and Penman,

Science

Penman,

S. (1966).

Science

749,

175. 955-961. Proc.

Nat.

Cell 3, l-10.

P. (1975).

Proc.

Nat. Acad.

Friend,

C. (1973).

J. Mol. Biol. 77, 117-130.

Preisler, H. D., Housman, D., Scher, W., and Proc. Nat. Acad. Sci. USA 70, 2956-2959. Puckett. L., Chambers, S., and Darnell, Acad. Sci. USA 72, 389-393. Rifkind, R. A., Danon, 22, 599-611. Ross. Acad.

Res.

D. (1962).

S. (1974).

F. I., and Leder,

Biophys.

G. (1965).

Marks, P. A., Burka, E. R., and Schlesinger, Acad. Sci. USA 48. 2163-2171. Orkin. S. H., Harosi, Sci. USA 75, 98-102.

Tran-

S., Ikawa, Y., Aviv, H., Symp. Quant. Biol. 38,

R. R., and Tomkins,

P. A., and Rifkind,

on Gene

J. Biol. Sci. 220, 905-915.

Leder, P., Ross, J., Gielen, J., Packman, and Swan, D. (1973). Cold Spring Harbor 753-761. Levy,

N. A., and

D., and

Marks,

J., Aviv, H., Scolnik, E., and Sci. USA 69, 264-268.

Ross, J., Ikawa, Y.. and Leder, USA 69, 3620-3623. ROSS, J., Gielen, J., Packman, J. Mol. Biol. 87, 697-714.

J. E. (1975). P. A. (1964).

Leder,

P. (1972a).

P. (1972b). S., Ikawa,

Proc.

Proc.

Nat.

J. Cell Biol. Proc.

Nat.

Nat. Acad.

Sci.

Y., and Leder,

Schweet, R., Lamfrom, H., and Allen, Sci. USA 44, 1029-1035.

E. (1958).

Singer,

J. Mol. Biol. 78, 321-334.

Stewart, 370.

R. H., and Penman,

S. (1973).

J., and Papaconstantinou.

J. (1967).

Stohlman, F.. Jr. (1970). In Regulation of Gordon, A. S., ed. (New York: Appleton-Century

Proc.

P. (1974). Nat.

Acad.

J. Mol. Biol. 29, 357Hematopoiesis, I, Crofts), p, 317.

Messenger 503

RNA

Metabolism

Terada, M., Cantor, Marks, P. A. (1972).

L.. Metafora, S.. Rifkind, R. A., Bank, A., and Proc. Nat. Acad. Sci. USA 69, 3575-3579.

Venetianer. P., and Leder, 77, 3892-3895. Verma, Nature

and Differentiation

P. (1974).

I. M., Temple, G. F., Fan, New Biol. 235, 163-167.

Note Added

Proc. H., and

Nat.

Acad.

Baltimore,

Sci.

USA

D. (1972).

In Proof

In recent experiments, a background of O.OOl-0.003% was obtained. It appears that in treated cells (labeled for IO min with SH-uridine), globin RNA synthesis was accelerated as compared to control cells, thus indicating that the globin gene is transcriptionally regulated in these cells.

Biosynthesis and stability of globin mRNA in cultured erythroleukemic Friend cells.

Cell, Vol. 6. 495-503, August 1976, Copyright0 1976 by MIT Biosynthesis and Stability of Globin mRNA in Friend Cells Cultured Erythroleukemic Hai...
837KB Sizes 0 Downloads 0 Views