Eur. J. Biochem. 161, 103-109 (1986) 0 FEBS 1986

A (re)initiation-dependent cell-free protein-synthesis system from mouse erythroleukemia cells Michael BADER and Thomas F. SARRE Institut fur Biologie 111, Universitat Freiburg i. Br.

(Received June Z/August 14, 1986) - EJB 86 0522

Cultured mouse erythroleukemia cells (MEL cells) can be induced in vivo to erythroid differentiation which is marked by the onset of globin mRNA and haemoglobin synthesis. When these cells are briefly exposed to hypertonic growth medium prior to lysis, the resulting post-mitochondria1 supernatants show a high in vitro protein-synthesis activity. Amino acid incorporation is linear up to 60 min; more than 80% of this is due to (re)initiation, as shown by the inhibition with edeine. Extracts from induced cells reach only a third of overall incorporation as compared to extracts from uninduced cells. This reduction of the protein-synthesizing capacity is also observed in v i v a Polyacrylamide gel electrophoresis shows that extracts from uninduced cells faithfully translate their endogenous mRNA, whereas in extracts from induced cells, non-globin protein synthesis is reduced and globin is preferentially synthesized. Haemin (40 pM) as well as purified eukaryotic initiation factor 2 (eIF-2) from rabbit reticulocytes enhance amino acid incorporation in both kinds of extracts, which suggests that both uninduced and induced MEL cells contain a haemin-controlled eIF-2a kinase. This system should be useful for studying the mechanisms controlling protein synthesis in a nucleated differentiating cell.

Regulation of gene expression in eukaryotic cells is not confined to transcription, but takes place on the translational level as well. Cellular differentiation, viral infection, various stress conditions etc. lead to changes in the degree and pattern of translational activity of the cell [l -31, but little is known about the control mechanisms involved.

The regulation of the rate of initiation of protein synthesis seems to play a major role, and great progress has been made in determining the mechanisms that regulate initiation in the rabbit reticulocyte lysate [l, 41 and in extracts from interferon- treated cells [ l , 51. However, it is not known whether these mechanisms represent general regulatory mechanisms for a variety of cells and physiological conditions.

In recent years, many attempts have been made to prepare in vitro protein-synthesis systems from nucleated cells, but most of these systems proved to be rather inefficient in trans- lational activity and in their capacity to (re)initiate poly- peptide synthesis [6 - 121. For investigations on the mechanisms controlling differential mRNA translation, we have chosen cultured mouse erythroleukemia cells (MEL cells) which can be induced to erythropoietic differentiation in culture by a series of chemical agents [13, 141. This differentiation process includes, within three to four days post-induction, the onset of globin mRNA and haemoglobin synthesis and the reduction of cell growth, cell size and overall RNA content [13 - 151. As described for other systems (see

Correspondence to T. F. Sarre, Institut fur Biologie 111, Schanzle- stral3e 1, D-7800 Freiburg, Federal Republic of Germany

Abbreviations. MEL, mouse erythroleukemia; NP-40, Nonidet P-40; SDS, sodium dodecyl sulfate; eIF-2, eukaryotic initiation fac- tor 2.

Enzyme. Creatine phosphokinase (EC 2.7.3.2).

above), cell-free extracts from MEL cells exhibit a rather poor translational activity consisting mainly of the elongation of pre-existing nascent peptide chains.

Based on the observation that a brief exposure of cells to hypertonic growth medium leads to a rapid and reversible block of polypeptide chain initiation in vivo [16, 171, we have treated MEL cells with 0.2 M KC1 prior to lysis. The resulting cell-free extracts show a high in vitro protein synthesis activity and faithfully translate their endogenous mRNA. More than 80% of [3 ?3]methionine incorporation is due to (re)initiation. This paper describes the preparation and properties of the cell-free extracts from uninduced and induced MEL cells and the effect of haemin and purified eIF-2 from rabbit re- ticulocytes on the (re)initiation rate. This system may offer the possibility of investigating the mechanisms controlling translation in a nucleated differentiating cell.

MATERIALS AND METHODS Materials

Dulbecco’s modified eagle medium (MEM) with or without L-methionine, 100 x non-essential amino acids and foetal calf serum were from Gibco (Wiesbaden). Hexa- methylenebisacetamide, Mops, NP-40 and sodium deoxy- cholate were from Sigma, penicillin from Serva (Heidelberg) and benzidine from Fluka (Buchs). Streptomycin and L-amino acids were purchased from Merck (Darmstadt). Dithiothreitol and haemin were from Roth (Karlsruhe). Edeine was ob- tained from Calbiochem. Creatine phosphate, creatine phosphokinase (EC 2.7.3.2). ATP and GTP were from Boehringer, Mannheim. ~-[~~S]methionine (900 - 1300 Ci/ mmol) was obtained from Amersham Buchler (Braun- schweig).

104

Culture and induction of MEL cells

MEL cells, line B8 (equivalent to the original Friend cell line 745), were a generous gift from Dr W. Ostertag (Heinrich Pette Institut fur Experimentelle Virologie und Immunologie, Hamburg). Cells were grown in our laboratory in Dulbecco's MEM containing 2 x non-essential amino acids, 1 x penicillin/ streptomycin and 10% foetal calf serum. Cells were cultured with initial cell densities of lo4- lo5 cells/ml at 37°C and 3% C 0 2 and passaged or harvested at a final cell density of lo6 cells/ml. Induction to erythrodifferentiation was achieved by addition of sterile 0.25 M hexamethylenebisacetamide to the medium to a final concentration of 4.5 mM. The amount of induced, i.e. haemoglobin-synthesizing cells, was determined by benzidine staining as described [18].

Labelling of cells in vivo After determining the viability of the cells by staining with

trypan blue, 5 x lo6 viable cells were removed from a culture flask under sterile conditions, pelleted by centrifugation at 1000 x g for 4 rnin and washed twice with 1 ml of ice-cold phosphate-buffered saline. Subsequently, cells were gently re- suspended in 1 ml of pre-warmed culture medium containing a fifth of the original methionine and 100 pCi [35S]methionine, and incubated at 37°C and 3% COz under occasional shaking. After 1 h cells were repelleted, washed twice with 1 ml phosphate-buffered saline and lyzed by the addition of 100 pl lysis buffer (20 mM Tris/HCl pH 7.5, 100 mM NaCI, 2 mM MgC12, 1 mM dithiothreitol, 1 % NP-40,0.2% sodium deoxy- cholate). The cell debris was removed by centrifugation at 10000 x g for 20 min at 4°C and the resulting supernatant was stored at - 80 "C. The protein concentration was determined by the method of Bradford [19] and appropriate amounts of supernatant were either precipitated with trichloroacetic acid for liquid scintillation counting or mixed with electrophoresis sample buffer for gel electrophoresis (see below).

Preparation of cell-jiree extracts

Uninduced or induced MEL cells were harvested by cen- trifugation at 1000 x g for 4 rnin at 4°C and washed once in cold isotonic buffer (20 mM Mops/KOH pH 7.2, 130 mM NaCl, 5 mM KCl, 7.5 mM magnesium acetate, 5 m M glucose). The packed cells were resuspended in 1.5 vol. cold hypotonic buffer (10 mM Mops/KOH pH 7.2, 10 mM KC1, 1.5 mM magnesium acetate, 2 mM dithiothreitol) and kept on ice for 10 min. Cell lysis was achieved by 15 strokes in a Dounce homogenizer using a tight-fitting pestle. The cell ex- tract was cleared by centrifugation at 10000 x g for 4 rnin at 4"C, and the cytoplasmic supernatant was prepared by a subsequent centrifugation at 10000 x g for 20 min at 4°C. 5O-pl aliquots of this cell-free extract were immediately frozen and stored in liquid nitrogen.

For the KCl treatment prior to lysis, cells were harvested by cenrifugation at IOOOxg for 4 min at 20°C and re- suspended in 1/10 of the original volume of pre-warmed culture medium without methionine, containing additional 0.2 M KCl. Under Occasional shaking, cells were incubated for 10 min at 37°C. Thereafter, the preparation procedure followed the protocol described above.

Protein synthesis in vitro Standard reaction assays (20 pl) contained 5 p1 of MEL

cell-free extract and the following components in final concen-

trations: 20 mM Mops/KOH pH 7.0,0.2 mM ATP, 0.05 mM GTP, 15 mM creatine phosphate, 0.1 mg/ml creatine phos- phokinase, 5 pM of all 20 L-amino acids, 0.2 pM [35S]me- thionine and 1 mM dithiothreitol. The optimal concentrations of Mg2+ and K + ions had been determined to be 1 mM and 100 mM, respectively, and were adjusted in the assay. Incubation was carried out for 60 rnin at 30°C.

A 0.4 mM haemin solution was prepared freshly for each experiment by dissolving 1 mg haemin successively in 0.2 ml 1 M KOH, 1 mlO.2 M Tris/HCl pH 7.5 and 2.5 ml H20 . The pH of this solution was adjusted to 7.5 by addition of 1 M HCl. [35S]Methionine incorporation was measured by re- moval of 5-p1 aliquots and subsequent precipitation with trichloroacetic acid as described [20]. Filters were counted in a Beckman LS 7000 liquid scintillation counter at an efficiency of approximately 90%. For gel electrophoresis, the aliquots were mixed with electrophoresis sample buffer (see below).

Polyacrylamide gel electrophoresis and autorudiogruphy

SDS/polyacrylamide gel electrophoresis was performed following the method of Laemmli [21] with the modifications of Studier [22]. Gradient gels were used composed of 10 - 20% (w/v) acrylamide, 0.27 -0.54% (w/v) bisacrylamide, 375 mM Tris/HCl pH 8.8, 0.1% (w/v) SDS, 2.5% (v/v) glycerol. Stacking gels had the composition of 3Yo (w/v) acrylamide, 0.08% (w/v) bisacrylamide, 125 mM Tris/HCl pH 6.8, 0.1% (w/v) SDS. Gels had the dimension 17 x 9.5 cm and 1.2 mm in thickness. The running buffer contained 192 mM glycine, 25 mM Tris base and 0.1 YO (w/v) SDS. Gels were usually run with 150-V constant voltage (approx. 50 mA) for 2.5 h.

Samples were mixed with 0.5 vol. 3 x electrophoresis sample buffer containing 0.24 M Tris/HCl pH 6.8, 6% (w/v) SDS, 30% (v/v) glycerol, 0.3 M dithiothreitol, 0.002% (w/v) bromophenol blue, and heated for 5 rnin at 95'C. The following proteins were used in a molecular mass marker mix: P-galactosidase (1 30 kDa), phosphorylase b (94 kDa). bovine serum albumin (68 kDa), ovalbumin (45 kDa), carboanhy- drase (30 kDa) and myoglobin (18 kDa). Gels were stained with Coomassie blue, destained and dried. For autorddiog- raphy, dried gels were exposed for I - 3 days at - 8O"C, using Fuji RX X-ray films.

Sucrose density gradient cen trijiugation analysis

3 A260 units of MEL cell-free extract were applied onto a 5-ml linear 15 - 40% (w/w) sucrose gradient containing 10mM Mops/KOH pH 7.2, 25mM KCI, 10mM NaCl, 1 mM MgCl2, 0.5 mM dithiothreitol, and centrifuged at 53000 rev./min for 45 min at 4OC (Beckman rotor SW 65). The contents of the tube were withdrawn from the bottom with a piercing unit and pumped through the flow cell of a Perkin Elmer recording spectrophotometer. Absorbance was measured at 260 nm.

Preparation of eIF-2 from rabbit reticulocytes

The procedure followed the method of Staehelin et al. [23] with some modifications, which are outlined briefly. The rabbit reticulocyte ribosomal salt wash was precipitated with crystalline ammonium sulfate in four steps (0 - 25%, 25 - 40%, 40 - 50% and 50 - 70%). After dialysis against IF buffer (20 mM Tris/HCl pH 7.6, 14 mM 2-mercaptoethanol, 0.1 mM EDTA, 10% glycerol) containing 100 mM KCl, the 40-50% ammonium sulfate cut was applied to a DEAE-

105

Table 1. [35S]Methionine incorporation in vitro with extracts from uninduced M E L cells prepared under different conditions In vitro protein synthesis was carried out as described in Materials and Methods. 5-pl aliquots were precipitated by trichloroacetic acid after a 60-min incubation at 30°C. Where indicated, edeine was present during incubation at a concentration of 5 pM. Typical cxperiments are shown

Extracts from x [35S]Methionine incorporation Inhibition by uninduced MEL cells edeine

- edeine + edeine difference

CPm Yo

Prepared without preincubation 81 51 30 37 Preincubated in medium without

Preincubated in medium without methionine 181

methionine plus additional 0.2 M KCl 81

139 42 23

13 68 84

cellulose (Whatman DE-52) column, equilibrated with the same buffer. After extensive washing at 100 mM KCl, protein was eluted with IF buffer containing 125 mM, 150 mM, 175 mM and 200 mM KC1 in a stepwise manner. Fractions of 0.5 ml were collected. The peak fractions of the 125 mM KC1 step contained highly (> SO?’”) purified eIF-2; 5O-pl aliquots were frozen and stored in liquid nitrogen.

RESULTS

EfSeect of KCI, prior to the preparation of ce1l:free extracts, on in vitro protein synthesis

When cell-free extracts for in vitro protein synthesis are prepared from mammalian tissue or cultured cells, the proce- dure generally includes three steps : (a) the resuspension of cells in a hypotonic lysis buffer, (b) lysis by homogenization and (c) the removal of nuclei and mitochondria by a 10000 x g centrifugation [6 - 121. MEL cell extracts, prepared by this method and incubated under the appropriate conditions for in vitro protein synthesis, show incorporation of [35S]methio- nine, but translation ceases after 20 min. If initiation inhibi- tors (e.g. edeine, aurin tricarboxylic acid) are added to the assay, a 20 - 40% inhibition of [35S]methionine incorporation is observed (Table I), indicating that most of the incorpora- tion is due to elongation of pre-existing peptide chains. Modifications of the incubation conditions (e.g. changes of temperature or pH, changes of ATP, GTP or creatine phosphate concentrations, potassium acetate instead of KC1 etc.) did not substantially influence this result (data not shown).

As has been reported previously [16, 171, high concentra- tions of potassium ions lead to a reversible in vivo block of peptide chain initiation without affecting the elongation of nascent chains. Therefore, the protocol for the preparation of cells was changed. Before cells were washed and lyzed, they were incubated for 10min at 37°C in a medium without methionine containing additional 0.2 M KCI. The incubation without methionine merely serves for the depletion of endogenous methionine and thus enhances [35S]methionine incorporation. The percentage of edeine-sensitive incorpora- tion, however, is not significantly altered (Table 1). The pre- treatment of cells with 0.2 M KC1 results in cell-free extracts, in which the (edeine-insensitive) elongation of pre-existing peptide chains is almost completely abolished. The edeine- sensitive amino acid incorporation, due to (re)initiation events, seems to be unimpaired or slightly enhanced. Thus,

the portion of initiation-dependent [35S]methionine incor- poration is increased from 23% to 84% (Table 1).

Sucrose density gradient centrifugation analysis of M EL cell extracts shows the effect of hypertonicity on the polyribosomal profile (Fig. 1). Extracts prepared according to the conventional protocol show the expected profile of intact polyribosomes (Fig. 1 A) as do extracts from cells in- cubated in medium without methionine. In cell extracts pre- pared from KC1-treated cells, polyribosomes disappeared almost completely and 80s particles accumulated (Fig. 1 B). We assume that the latter represent loose couples of ribosomal subunits and not 80s initiation complexes, since initiation in vitro is sensitive to edeine (Table 1) which is reported to block the joining of the 60s ribosomal subunit to the preinitiation complex 11 [24].

Characteristics of protein synthesis in vitro and in vivo

The time course of [35S]methionine incorporation by MEL cell extracts prepared after high KCI treatment is outlined in Fig. 2 (A: uninduced cells; B: induced cells). The induced cells were harvested three days post-induction, consisting of approximately 80% benzidine-positive, i.e. haemoglobin- synthesizing, cells.

In both extracts, protein synthesis is nearly linear up to 60 min after an initial lag phase of 10 min (Fig. 2); more than 80% of [35S]methionine incorporation into protein is due to (re)initiation as demonstrated by the addition of edeine. However, with extracts from induced MEL cells, the incorpo- ration rate is lower and overall protein synthesis is only a third of that obtained with extracts from uninduced MEL cells (Fig. 2). Since MEL cells show a marked decrease in RNA content during the course of induction [25], i.e. from approximately 500 pg/108 cells to 200 - 250 pg on day 3 post- induction (our data), we tested for the possibility that this reduction affects poly(A)-rich RNA and thus accounts for the diminished amino acid incorporation observed in vitro. However, extracts from uninduced and induced MEL cells contain the same amount of poly(A)-rich RNA, i.e. ap- proximately 1.5 - 2 pmol/Az6,, unit of extract as measured by hybridization to [3H]poly(U) following the method of Safer et al. [26].

Consequently, cells were pulse-labelled with [ 35S]methio- nine in vivo during the time course of induction to see whether the reduction of protein synthesis can be observed also in vivo. In these experiments, care was taken that the same number of cells were used and that the cells were viable. As Fig. 3A

106

A

0.65

0.2

* C - P

4" 0.1

B 1.10

0.65

0.2

u

c 3 - 0 B

0.1

Fig. 1. Sucrose gradient analysis of MEL cell extracts. 3 AZ60 units of extract from uninduced MEL cells, without preincubation of the cells (A) or with preincubation in the presence of 0.2 M KCl prior to lysis (B), were analyzed. Sedimentation was from left to right, the position of 80s particles is marked

L

27500 - 25000 -

3 22500

!? B

LL

: 20000 - v

c)

17500

8 l5000 - E .-

12500 ... B 9 ... 10000 -

5 7500 1 - ul $ 5000

2500 -

0 -

o 10 zo 30 a 50 no o LO 20 30 +o M no TIME (min) TIME (min)

Fig. 2. Time course of (3SS]methionine incorporation in MEL ceN extracts. Extracts from uninduced (A) and induced (B) MEL cclls. preincubated in the presence of 0.2 M KCI prior to lysis, were assayed for in vitro protein synthesis. Induced cells had been harvested at day 3 post-induction. 5O-pl assays contained 2 A260 units of extract and included no addition ( O ) , 5 pM edeine (O), 40 p M haemin (A) or 4 pg purified eIF-2 from rabbit reticulocytes (+). 5 4 samples were taken at the times indicated and precipitated with trichloroacetic acid as described in Methods

shows, overall [35S]methionine incorporation into protein capacity of cell extracts from induced MEL cells merely re- decreases after 2 days post-induction and represents only one flects a status occurring already in vivo. third of the initial value on the fourth day. Globin synthesis The product analysis of in vitro protein synthesis and the becomes visible 22 h post-induction and seems to be corresponding pattern of proteins synthesized in vivo is shown unaffected by the process which reduces the synthesis of other in Fig. 4. In the case of uninduced cells and the corresponding proteins (Fig. 3 B). Thus, the diminished protein-synthesizing extracts, no significant difference, either in the pattern or in

107

I 0 22 48 70 93

h post-induction

Fig. 3. [35S]Methionine incorporation in vivo into protein by MEL cells during the time course of induction. Cells were induced with 4.5 mM hexamethylenebisacetamide. At the times indicated, 5 x lo6 viable cells were labelled in vivo as described in Materials and Methods. Equal amounts of protein (30 pg), as determined by the method of Bradford [19], were either precipitated with trichloroacetic acid (A) or analyzed by polyacrylamide gel electrophoresis (B). In the latter case, the autoradiogram is shown. Arrows indicate the position of the protein markers on the stained gel; the arrowhead marks the position of globin

the relative amount of proteins synthesized, can be detected (Fig. 4; lanes 1 and 2). Only high-molecular-mass proteins ( M , >130000) seem to be translated less efficiently in vitro (Fig. 4, lane 2). Thus, the protein synthesis system described here allows a faithful translation of endogenous mRNA. Ex- tracts from induced cells, however, show a pattern of transla- tion (Fig. 4, lane 6 ) which is different from that obtained in vivo (Fig. 4, lane 5): with a few exceptions where the synthesis of polypeptides is either similar or slightly enhanced, overall protein synthesis is clearly reduced and globin is preferentially synthesized. The reason(s) for this result will be discussed below. In both extracts, the initiation inhibitor edeine virtually abolishes [35S]methionine incorporation into protein (Fig. 4; lanes 4 and 8) which indicates that more than 90% of in vitro protein synthesis is due to (re)initiation events.

Effect of haemin and eIF-2 on protein synthesis

Since induced MEL cells at day 3-4 post-induction synthesize haemoglobin preferentially (Fig. 3 B) and thus be- come comparable to reticulocytes, we investigated whether or not haemin exhibits the same stimulating effect on in vitro protein synthesis as is found in the rabbit reticulocyte lysate [l, 41. Protein synthesis in cell-free extracts from induced MEL cells does respond to the addition of haemin (Fig. 2B), but the effect is far less than expected and, furthermore, is observed with extracts from uninduced MEL cells as well (Fig. 2A), though they do not synthesize haemoglobin (Fig. 3B). This non-globin-specific effect of haemin has also been observed by others previously [8, 271. Fig. 4 (lane 3) shows that the stimulation of protein synthesis by haemin is

not at all restricted to globin but affects all proteins synthe- sized. In the case of extracts from induced cells, however, a preferential stimulation of globin synthesis becomes visible (Fig. 4, lane 7).

Purified eIF-2 from rabbit reticulocytes exhibits a com- parable stimulation of protein synthesis to haemin (Fig. 2A, B), as has been described in haemin-depleted rabbit re- ticulocyte lysates [l, 41. This observation, together with the data obtained by the addition of haemin to the system, suggests that MEL cell extracts may contain (a) haemin- controlled eIF-2a kinase(s). Preliminary data indicate that this is the case for both uninduced and induced MEL cells (Th. Sarre, unpublished work).

DISCUSSION This paper describes the preparation and characteristics

of a cell-free protein-synthesis system from normal and differ- entiating MEL cells. A short-time incubation of uninduced or induced MEL cells in the presence of 0.2 M KCl prior to lysis results in cell-free extracts in which more than 80% of in vitro protein synthesis is due to (re)initiation (Table 1 ; Figs 2 and 4). The molecular basis for the effect of KC1 on translation in vivo is unclear [16, 171, but our data are in agreement with the suggestion of Yates and Nuss [17] that hypertonic treatment inhibits polypeptide chain initiation at a step prior to the mRNA-dependent joining of the 60 S ribosomal subunits to the preinitiation complex I.

Extracts from induced MEL cells show a distinct reduction of their protein synthesizing capacity (Fig. 2 B) compared to extracts from uninduced cells (Fig. 2A). This finding, howev-

Fig. 4. Analysis of translation products synthesized by cell-free extracts. After preincubation in the presence of 0.2 M KCl prior to lysis, extracts were prepared from uninduced and induced MEL cells (3 days post-induction) as described in Materials and Methods. For comparison, 5 x lo6 viable cells from the respective cell culture were labelled in vivo. Samples were subjected to SDS/polyacryamide gel electrophoresis with subsequent autoradiography. The autoradio- gram shows: the in vivo translation products from uninduced (lane 1 ) and induced MEL cells (lane 5), 5-pl samples taken from a 60-min incubation with extracts from uninduced (lanes 2-4) and induced MEL cells (lanes 6-8). Incubation was carried out as described in Methods and the assays contained no addition (lanes 2 and 6), 40 p M haemin (lanes 3 and 7) or 5 pM edeine (lanes 4 and 8). Arrows indicate the position of the protein markers on the stained gel; the arrowhead marks the position of globin

er, reflects a situation already present in vivo with intact cells (Fig. 3 A) and leads us to the assumption that overall protein synthesis is reduced in favour of globin synthesis during the time course of induction (Fig. 3B; Fig. 4, lanes 1 and 5).

Extracts from uninduced MEL cells faithfully translate their endogenous mRNA (Fig. 4, lanes 1 and 2), whereas extracts from induced cells show a pattern of translation sig- nificantly different from that observed in vivo (Fig. 4, lanes 5 and 6): Globin is preferentially synthesized and a major por- tion of non-globin protein synthesis is clearly reduced.

As to the in vivo data (Fig. 3B; Fig. 4, lane 9, one has to take into account that induced cells at day 3 -4 post-induction actually represent a mixture of 10-20% uninduced and 80- 90% induced cells, hence the pattern of proteins labelled in vivo represents a superposition of normal protein synthesis promoted by uninduced cells and the translation occurring in the differentiating cells. In cell-free extracts, the postmito- chondrial cytoplasms from (20%) uninduced and (80%) in- duced cells are mixed. Thus, influences on the translational activity of one by the other cytoplasm may become visible.

Based on these results, we suggest that the differentiation process involves the synthesis or activation of (a) translational inhibitor(s) specific for non-globin protein synthesis. Whether or not the observed enhancement of globin synthesis is just a mere consequence of this or requires the concomitant

appearance of a globin-specific activator of translation, has to be elucidated. Since the overall poly(A)-rich RNA content of MEL cells is not reduced during induction, the observed inhibition of non-globin protein synthesis implies that non- globin mRNA is, at least in part, shifted into the pool of untranslated free cytoplasmic messenger ribonucleoprotein. Such changes in mRNA utilization during ertyhrodifferentia- tion of MEL cells are currently being investigated in our laboratory (G. Schulze, unpublished results) and have been reported for a few distinct mRNA species recently [28, 291.

The existence of translational inhibitors in either uninduced or induced MEL cells has been reported before [30 - 331. Hardesty and coworkers have purified an inhibitor from uninduced MEL cells which shows eIF-212 kinase activity and inhibits protein synthesis in the rabbit reticulocyte system [32]. This is in agreement with our finding of a haemin- controlled eIF-2a kinase (see below), but the translational inhibitor(s) postulated above should have different characteristics. The nature of the translational inhibitor dc- scribed by Stringer et al. [33] is unclear, since it was tested in a cell-free system from MEL cells that was merely elongating pre-existing nascent peptide chains. Thus, further ex- perimental evidence is necessary to corroborate the postulate of inhibitors of non-globin protein synthesis in induced MEL cells.

Since induction of MEL cells leads to the accumulation of haemoglobin-synthesizing erythroblasts, i.e. the precursor cells of reticulocytes, we investigated whether these cells develop the haemin dependence of protein synthesis described in detail for rabbit reticulocytes [l, 41. In fact, protein synthesis in extracts from induced MEL cells can be stimulated by haemin (Fig. 2B) and the effect may even be specific for globin synthesis (Fig. 4; lanes 6 and 7). However, the effect of haemin is less distinct as expected, and can also be observed in extracts from uninduced MEL cells where no globin synthesis occurs (Fig. 2A; Fig. 4, lanes 2 and 3). The latter finding is in agree- ment with data from our laboratory that postribosomal supernatants from both uninduced and induced MEL cells contain a haemin-controlled eIF-2a kinase.

The reasons why haemin and eIF-2 do not show a more pronounced effect on translation in our system are obscure. We assume that cell-free extracts from KC1-treated MEL cells are enriched in preinitiation complexes prior to the joining of mRNA and/or the 60s ribosomal subunit. Thus, endogenous eIF-2 might not be accessible to the inherent kinase(s) during the first round of initiation. Only during reinitiation should phosphorylation of eIF-2 lead to an inhibition of protein synthesis, which, in turn, can be prevented by haemin or additional eIF-2 (Fig. 2A and B).

The cell-free protein synthesis system from uninduced and induced MEL cells described here should allow more detailed studies on the following mechanisms controlling protein syn- thesis in a nucleated cell on its way to erythroid differentia- tion: (a) the translational inhibitor(s) that promote a shut-off of non-globin protein synthesis during differentiation, and (b) the haemin-controlled eIF-2cc kinase(s) and their role for globin and non-globin protein synthesis.

This work was supported by a grant from the Deurschr For- schungsgemeinschuft to Dr K. Hike (Hi 188/3-3), whose advice and comments we gratefully acknowledge. We thank M. Gorlach, G. Schulze, Drs R. Hertel and W. Michalke for helpful criticism during the work and the preparation of the manuscript.

REFERENCES 1.

2. 3.

4.

5. 6. 7. 8.

9. 10.

11.

12.

13.

14.

15. 16.

17.

Jagus, R., Anderson, W. F. &Safer, B. (1981) Progr. Nucleic Acid

Austin, S. A. & Clemens, M. J. (1981) Biosci. Rep. 1, 35-44. Bablanian, R. (1984) in Comprehensive virology (Fraenkel-

Conrat, H. &Wagner, R. R., eds), pp. 391 -429, Plenum Press, New York.

Jackson, R. J. (1982) in Protein biosynthesis in eukaryotes (Perez- Bercoff, R., ed.) pp. 363-418, Plenum Press, New York.

Lengyel, P. (1982) Annu. Rev. Biochem. 51,251 -282. Falvey, A. K. & Staehelin, T. (1970) J. Mol. Biol. 53, 21 -34. Skup, D. & Millward, S. (1977) Nucleic Acids Res. 4,3581 -3587. Dabney, B. J. & Beaudet, A. L. (1978) J. Biol. Chem. 253,7124-

Fischer, I. & Moldave, K. (1981) Anal. Biochem. 113, 13-26. Person, A., Nielsen, P., Beaud, G. & Trachsel, H. (1984) Biochim.

Morley, S. J. & Jackson, R. J. (1985) Biochim. Biophys. Acta 825,

Morley, S. J., Buhl, W.-J. &Jackson, R. J . (1985) Biochim. Bio-

Marks, P. A. & Rifkind, R. A. (1978) Annu. Rev. Biochem. 47,

Reuben, R. C., Rifkind, R. A. & Marks, P. A. (1980) Biochim.

Curtis, P. J. (1980) Biochim. Biophys. Acta 605,347 - 364. Saborio, J. L., Pong, S . 3 . & Koch, G. (1974) J . Mol. Biol. 85,

Yates, J. R. & Nuss, D. L. (1982) J. Biol. Chem. 257, 15030-

Res. Mol. Biol. 25, 127 - 185.

7126.

Biophys. Acta 783, 152 - 157.

45 - 56.

phys. Acta 825, 57 - 69.

41 9 - 448.

Biuphys. Acta 605, 325 - 346.

195-21 1.

15034.

109

18. Orkin, S. H., Harosi, F. I. & Leder, P. (1975) Proc. Natl Acad.

19. Bradford, M. M. (1976) Anal. Biochem. 72,248-254. 20. Jackson, R. J. & Hunt, T. (1983) Methods Enzymol. 96,50-74. 21. Laemmli, U. K. (1970) Nature (Lond.) 227,680-685. 22. Studier, F. W. (1973) J. Mol. Biol. 79, 231-248. 23. Staehelin, T., Erni, B. & Schreier, M. H. (1979) Methods Enzjmol.

24. Santon, J. B. & Stanley, W. M. Jr (1978) Biochem. Biophys. Res.

25. Shcrton, C. C. & Kabat, D. (1976)Dev. Biol. 48, 118-131. 26. Safer, B., Jagus, R. & Kemper, W. M. (1979) Methods Enzymol.

27. Weber, L. A., Feman, E. R. & Baglioni, C. (1975) Biochemistry

28. Yenofsky, R., Cereghini, S., Krowczynska, A. & Brawerman, G. (1983) Mol. Cell Biol. 3, 1197-1203.

29. Krowczynska, A., Yenofsky, R. & Brawerman, G. (1985) J. Mol. B id . 181, 231 -239.

30. Cimadevilla, J. M. & Hardesty, B. (1975) Biochem. Biophys. Res Commun. 63,931 -937.

31. Cimadevilla, J. M., Kramer, G., Pinphanichakarn, P., Konecki, D. & Hardesty, B. (1975) Arch. Biochem. Biophys. 171, 145- 153.

32. Pinphanichakarn, P., Kramer, G. & Hardesty, B. (1977) J . Biol. Chem. 252,2106-2112.

33. Stringer, E. A. & Friend, C. (1982) Proc. Natl Acad. Sci. USA 79,

Sci. USA 72,98 - 102.

60,136-165.

Commun. 84,985 - 992.

60,61-87.

14, 5315-5321.

1839 - 1843.