Induction of globin mRNA accumulation by hemin in cultured erythroleukemic cells

Cell, Vol. 8, 513420,

August

1976,

Copyright

0 1976

by MIT

Induction of Globin mRNA Accumulation in Cultured Erythroleukemic Cells Jeffrey Ross and Deborah Sautner McArdle Laboratory for Cancer Research University of Wisconsin Madison, Wisconsin 53706

Summary The role of heme in erythrold development is investigated in erythroleukemic (Frfend) cells. Exogenous hemin induces the accumulation of globln mRNA and globin protein in T&Cl2 erythroleukemia cells to levels comparable to those induced by polar solvents, such as dlmethylsulfoxide (DMSO). The hemin concentration required for maximal induction (lo-4 M) is the same as that which stimulates globln message translation In retlculocytes or cell-free retlculocyte lysates. Hemin and DMSO together cause T3-Cl2 cells to accumulate 8-9 fold more globin mRNA than either inducer indlvldually. The kinetics of globin mRNA induction in hemin as compared to DMSO are very different: globin message accumulation begins 4 hr after hemin addition, but not until 30-40 hr after DMSO addition. Blllverdin induces 20-40 fold less hemoglobin than hemin; delta-amlnolevulinic acid and porphobilinogen do not induce. Introduction Cultured murine erythroid cells (Friend leukemia cells) that can synthesize hemoglobin have been used as a model system to study aspects of red cell development. These cells accumulate heme and globin (Friend et al., 1971; Boyer et al., 1972; Kabat et al., 1975) globin mRNA (Ross, Ikawa, and Leder, 1972; Preisler et al., 1973), and erythrocytespecific membrane antigens (Ikawa, Furusawa, and Sugano, 1973) in the presence of appropriate inducers. The most frequently utilized inducer has been dimethylsulfoxide (DMSO), although other polar solvents (Tanaka et al., 1975) and short chain organic acids (Ikawa, Aida, and Saito, 1975; Leder and Leder, 1975) are potent inducers as well. The mechanism of action of these compounds is unknown. A major question concerns whether these cells will respond also to stimuli that function in vivo. Erythropoietin, a hormone required for red cell maturation, does stimulate erythroleukemic cells to accumulate heme and to form colonies in plasma clots, but the degree of stimulation is small relative to DMSO and requires the simultaneous presence of DMSO (Preisler and Giladi, 1974; Goldstein et al., 1974). In contrast, primary cultured erythroid cells derived directly from an animal source respond to exogenous erythropoietin alone

by Hemin

(Gallien-Lartigue and Goldwasser, 1964; Terada et al., 1972). One interpretation of these results is that the erythroleukemic Friend cells are analogous to erythroblasts whose development has been blocked subsequent to the erythropoietin responsive stage. If so, regulatory molecules that function only in the later stages of erythroid development, such as heme, may reactivate the developmental process in these cells. Exogenous hemin stimulates the translation of globin mRNA in intact reticulocytes (Bruns and London, 1965) and in reticulocyte ceil-free translation systems (Adamson, Herbert, and Kemp, 1969); it has a similar effect with other mRNAs in other cell-free systems (Beuzard, Rodvien, and London, 1973). Moreover, the accumulation of heme and globin in intact cells is apparently coordinated, since factors that affect the synthesis of one usually affect the other to a similar extent (Kruh and Borsook, 1956; Morell, Savoie, and London, 1958). Since erythroleukemic cells represent nucleated erythroid precursors that are more immature than reticulocytes (Friend et al., 1971), it was possible to ask if hemin has a regulatory role in these cells, as well as in reticulocytes. (Heme is protoporphyrin IX containing reduced iron; hemin contains oxidized iron. Otherwise, they are identical. Where applicable, heme and hemin are distinguished in this paper. For example, heme is present in intracellular hemoglobin, while the material added to the culture fluid is hemin.) The effect of hemin has been investigated in two different erythroleukemic cell lines: clone 745 and T3-Cl2. Clone 745 cells contain significant constitutive levels of hemoglobin and globin mRNA, both of which are increased following DMSO treatment (Preisler et al., 1973). The T3-Cl2 line (Furusawa, Ikawa, and Sugano, 1971) provides a particularly useful model system for these experiments, since these cells contain low constitutive levels of globin mRNA and no detectable hemoglobin when cultured without inducers (Ross et al., 1974). In these experiments, we have attempted to answer two questions concerning the potential role of hemin in erythroid maturation: first, will hemin induce detectable quantities of hemoglobin, and second, will hemin increase the quantity of globin mRNA? Results Induction of Globin Synthesis by Hemin To determine whether hemin induced globin accumulation, an immunodiffusion technique was devised. Globin was prepared from highly purified mouse hemoglobin and was used to immunize sheep. Immune serum formed a precipitin band against mouse hemoglobin, but not against duck

Cell 514

or human hemoglobin and not against a solution of 10-4 M hemin. The precipitin band stained dark blue with benzidine, which demonstrated the presence of large amounts of hemin in the precipitated protein. Serum obtained from the same animal prior to immunization failed to react with mouse hemoglobin; this result indicated that the antibody appeared in response to the mouse globin immunogen. Initial experiments showed that extracts from T3-Cl2 cells treated with hemin (10-4 M), DMSO (l%, v/v), or a combination thereof, formed precipitin bands, while no bands were observed with extracts from untreated controls. To estimate the average quantity of hemoglobin per cell, an endpoint dilution experiment was performed. Lysates from a specified number of cells were serially diluted to determine the highest dilution at which a precipitin band would form. By comparison with a mouse hemoglobin standard of known concentration, the hemoglobin content per cell could be determined. The results, summarized in Table 1 (column 4) indicate that heminor DMSO-treated T3-Cl2 cells contained approximately 2.2 pg of hemoglobin on theaverage.The minimum detectable hemoglobin concentration was 0.1 pg per cell with this assay (Experimental Procedures). Therefore, untreated cells contained less than 0.1 pg hemoglobin. Cells treated simultaneously with DMSO and hemin accumulated 10 fold more hemoglobin (23 pg) than cells treated with either inducer alone, and contained approximately 7 pg more hemoglobin than the adult mouse erythrocyte (Altman, 1961). A summary of several properties of hemin-treated T3-Cl2 cells is shown in Table 1. The doubling time and saturation density in treated and untreated cultures were similar (columns 2 and 3). Hemin was effective between 5 x 10-5 and 10-4 M; 10-5 M was ineffective, and concentrations >lO-4 M have not been used (see Table 3). The concentration of hemin that maximally stimulates globin translation in vitro is also in the range of 10-4 M (Bruns and Table

1. Properties

of Erythroleukemic

T3-Cl2 Cell Doubling Time

Cells Cultured

with Hemin,

Inducer

0-W

Saturation Density (Cell Number per ml x 10-S)

None

12

5.9

13

0.8

DMSO

(l%,

v/v)

Hemin

(lo-4

M)

DMSO

(1%) plus

Hemin

(10-a

M)

14

4.6

14

7.3

London, 1965; Adamson et al., 1969). Maximal induction with DMSO occurs at a concentration of 1.5% in T3-Cf2 cells, but the growth rate is less than that of untreated controls (Ross et al., 1974). In contrast, the growth rate in 1% DMSO is the same as that in controls or hemin-treated cells (Table 1). 1% DMSO was therefore chosen as a standard inducer in these studies. To determine the specificity of hemin induction and the effect of hemin on cells that produce hemoglobin without inducers, a similar experiment was performed with three other cell lines. The SCRF-145 and SCRF-179 lines are lymphoblastoid cells, one of which (-145) produces an atropic C type virus (Dr. F. C. Jensen, personal communication). The clone 745 line is an erythroleukemia cell that is constitutive for detectable hemoglobin production but can be stimulated by DMSO to accumulate still greater quantities of hemoglobin (Preisler et al., 1973). As shown in Table 2, hemoglobin was not detected in the lymphoblastoid cells under any conditions. Globin mRNA was also not detected in these cells with hemin or DMSO, under conditions where >5 x 10-4 pg per cell would have been detected. In clone 745 cells, hemin induced a 10 fold increase in hemoglobin content. DMSO, at a concentration that did not decrease the growth rate relative to control or hemin-treated cells, induced hemoglobin 20 fold over the control levels, while DMSO and hemin together acted synergistically to elevate hemoglobin levels 60 fold. These data indicate that hemin does not induce hemoglobin accumulation in lymphoblastoid cells. It does induce hemoglobin accumulation in erythroleukemic cells constitutive for hemoglobin production, but it is not as active as DMSO in these cells. Globin mRNA Accumulation in the Presence of Hemin The finding that hemin induced hemoglobin accumulation suggested that hemin, like DMSO, may have induced globin message accumulation. The DMSO,

or DMSO

Hemoglobin on Day 5

plus

per Cell

Hemin

(PS)

(PS)

Globin mRNA per Cell on Day 5 (Pg x 10’)

to.1

12

0.01

Total RNA per Cell on Day 5

1

2.2 (2-2.5)

6

2.2 (2-2.5)

12

0.6

11

9.1

23

(20-31)

T3-Cl2 Cells were cultured as described in Experimental Procedures. For flasks receiving both DMSO and hemin, the DMSO was added first, and hemin was added soon thereafter (l-10 min). Hemoglobin determinations were made by immunodiffusion, and the results represent the average of four determinations in four separate experiments. Numbers in parentheses represent the range of values in these experiments. The immunodiffusion assay as described here cannot detect hemoglobin at a concentration lower than 0.1 pg per cell. The pg of globin message per cell were calculated from the percentage of globin message obtained from hybridization analyses shown in Figure 1.

Globin 515

mRNA

Induction

with Hemin

Table 2. The Effect of Hemin, DMSO. SCRF-179 (Lymphoblastoid) Cells

or DMSO

plus Hemin

Clone

on Clone

745 (Erythroleukemic)

or on SCRF-145

and

745

Hemoglobin on Day 5

SCRF-145

Inducer

(Pg)

Globin mRNA per Cell on Day 5 (PS x 10’)

None

0.1

0.04

to.1

2

0.91

to.1

1

0.18

to.1

8

0.91

to.1

DMSO

(196, v/v)

Hemin

(IO-4

M)

DMSO

(1%)

plus

Hemin

(IO-4

M)

per Cell

Hemoglobin Day 5

Kinetics of Globin mRNA Accumulation in T3-Cl2 Cells One question raised by these studies is whether the inductive pathways of hemin and DMSO differ. If so, unique characteristics of the hemin induction mechanism in erythroleukemic cells could reflect events that occur in the development of nucleated erythroid precursor cells in the intact animal. As an initial approach to this question, the kinetics of globin message accumulation in hemin as opposed to DMSO have been compared. A characteristic feature of DMSO induction in some erythroleukemic lines is a lag period between introduction of the inducer and the earliest detectable erythroid

per Cell on

(pg)

Logarithmically growing cells were subcultured to fresh growth medium at 5 x 103 cells per ml. Inducers were 5 days, cells were harvested, and cytoplasmic lysates and RNA were prepared. All cells treated with inducers the same rate as untreated controls. Hemoglobin determinations represent the average of three separate experiments. was determined by hybridization Cot analysis, exactly as described for T3-Cl2 cells (Figure 1).

alternative possibility was that hemin stimulated the translation of constitutive globin message in uninduced cells (Ross et al., 1974). T3-Cl2 cells were cultured with hemin (lo-4 M), DMSO (l%, v/v), or hemin plus DMSO for 5 days, at which time total cellular RNA was prepared and hybridized to 3Hglobin cDNA. The results (Figure 1 and Table 1, column 6) show that hemin induced a 60 fold increase in globin mRNA as compared to untreated controls. The percentage of globin mRNA in hemin-treated cells was one third that in DMSO-treated cells (Figure l), but there was only a 2 fold difference in the actual quantity of globin message per cell (Table 1, column 6). These data demonstrate that hemin, like DMSO, induced globin mRNA accumulation in T3-Cl2 cells. It also induced globin mRNA accumulation in clone 745 erythroleukemia cells, which are constitutive for hemoglobin production (Table 2). In this case, however, hemin was not as effective as 1% DMSO in stimulating globin message accumulation. Moreover, whereas hemin and DMSO together stimulated hemoglobin synergistically in T3Cl2 cells, the globin mRNA levels in clone 745 cells treated simultaneously with both inducers was not greater than in cells treated with DMSO alone.

or SCRF-179

IO'

added 4 hr later, After grew at approximately Globin mRNA content

102

103

Cot (moles set/l) Figure

1. Hybridization

to T3-Cl2

Total Cell RNA to “H-Globin

cDNA

Logarithmically growing T3-Cl2 cells were passed to fresh growth medium at a concentration of 5 x 10’ cells per ml. 4 hr later, cells received hemin (final concentration IO-4 M). DMSO (l%, v/v). hemin plus DMSO (10-a M plus l%), or pH 7.8 buffer used to dissolve the hemin (see Experimental Procedures). After 4.5 days in culture, cells were harvested and total cellular RNA was extracted. Different amounts of RNA were then incubated with IH-globin cDNA (500 cpm) for 18 hr at 65°C. Purified 9s mouse globin mRNA was obtained as previously described (Ross et al., 1974) and was annealed under identical conditions. The Cot, for this RNA was 2.4 x 10-3 (data not shown). The percentage of the cDNA that hybridized was determined by the Sl nuclease assay, 2% of the cDNA being nuclease-resistant in the absence of RNA. Appropriate controls were performed to exclude the possibility that nuclease resistance was due to enzyme inhibition rather than hybridization (Ross et al., 1974). (0-O) untreated control; Cot% = 3.2 x 103 (by extrapolation); (D--B) 10-4 M hemin; Cot, = 5 x 101; (U) 1% DMSO; Cotw = 1.4 x 10’; (&A) 10-4 M hemin plus 1% DMSO; Cot, = 2.6.

response of the cells (McClintock and Papaconstantinou, 1974; Levy et al., 1975). To determine the early globin message kinetics in hemin or DMSO or a combination thereof, T3-Cl2 cells were passed to fresh medium, and inducer was added 4 hr later. At intervals thereafter, cells were harvested, and total cell RNA was prepared and incubated with 3Hglobin cDNA. With DMSO there was a 30-40 hr lag

Cell 516

before the globin message content began to increase (Figure 2, circles; Ross et al., 1974). In contrast, globin message accumulation was observed as early as 4 hr after hemin addition (squares). With a combination of inducers, there was a lag period, as with DMSO alone, but its duration was 14-16 hr shorter (triangles). These experiments indicated that the kinetics of globin message accumulation in hemin and DMSO were different. By 24 hr after addition of inducers, the globin mRNA concentration of hemin-treated cells had risen 4 fold above the basal level, while that of DMSO-treated cells had remained unchanged. To confirm this observation and to quantitate the difference more precisely, the globin mFiNA level in cells cultured for 24 hr with or without inducers was determined by Cot analysis. These data (Figure 3) indicate that the globin mRNA level in DMSO-treated cells (closed circles) was identical with that in untreated controls (open circles). Hemin-treated cells contained 4 fold more globin message (squares), while cells cultured in hemin

plus DMSO (triangles) contained twice as much globin mRNA as did DMSO-treated cells. These data confirm the fact that in T3-Cl2 cells, globin mRNA induction with hemin begins 28-32 hr earlier than with DMSO.

Effects of Heme Pathway Intermediates T3-Cl2 Cells

on

Heme is synthesized through a series of essentially irreversible reactions, starting with the condensation of glycine and succinyl coenzyme A to form delta-aminolevulinate. Some of these reactions occur within, others outside of, mitochondria. To determine whether several heme synthetic intermediates induce globin accumulation, hemoglobin assays were performed on lysates of T3-Cl2 cells grown for 5 days with delta-aminolevulinate or porphobilinogen. Table 3 shows that neither compound induced hemoglobin accumulation to detectable levels. However, biliverdin, the product of the first of a series of reactions involved in heme catabolism, did induce low levels of hemoglobin. The primary structure of biliverdin is similar to heme, but one of the methylene bridges is broken, resulting in a linear rather than closed tetrapyrrole (Figure 4). The reaction whereby heme is converted to biliverdin occurs in phagocytic reticuloendothelial cells, not in erythroid cells, and biliverdin is not a z .p! 0 ‘C

g

I

O-

20-

ii!! 402 n 600 I m= 60.-c

20

40

60

Time 1hr) Figure 2. Kinetics of Globin mRNA Accumulation of Hemin or DMSO or Hemin plus DMSO

s I

I

I

IO2

IO3

C,t (moles se& I) in the Presence

Logarithmically growing T3-Cl2 cells were passed to fresh growth medium at 2 x 104 cells per ml. 4 hr later, freshly prepared hemin (final concentration 1 O-4 M) or DMSO (1%) or hemin (1 O-4 M) plus DMSO (1%) was added. At intervals thereafter cells were counted and harvested, and total cellular RNA was isolated. Zero time in the figure is the time at which inducers were added. The growth rate of the cells was virtually identical during the 70 hr experimental period (data not shown). 10 pg of RNA were incubated with ‘Hglobin cDNA for 18 hr, and the percentage hybridized was determined. 2% of the cDNA was nuclease-resistant in the absence of RNA, and this background has been subtracted from each point. (M) DMSO; ( M) hemin; (&A) hemin plus DMSO.

Figure 3. Hybridization Kinetic Analysis with 3H-Globin RNA from T3-Cl2 Cells Cultured 24 Hr with or without

cDNA and Inducers

Logarithmically growing T3-Cl2 cells were passed to fresh growth medium at 2 x 104 cells per ml. 4 hr later, cells received hemin (10-d M), DMSO (l%, v/v), hemin plus DMSO (lo-4 M plus 1%) or pH 7.6 buffer used to dissolve the hemin. 24 hr after the addition of inducers, cells were harvested, and total cellular RNA was isolated and hybridized with 31-l-globin cDNA. as described in the legend to Figure 1. (o--o) untreated control; Cot% = 103; (U) 10-d M hemin; Cot, = 2.5 x 102; (O-O) 1% DMSO; CotH = 103; (CA) 10-a M hemin plus 1% DMSO; CotK = 4.7 x 102.

I

Globin 517

mRNA

Table

3. Effect

Induction

with

of Hemin

Hemin

and Heme-Related

Compounds

on Hemoglobin

Accumulation

in T3-Cl2

Concentrations Tested Compound Delta-Aminolevulinic Porphobilinogen

Hemoglobin on Day 5 (pg)

UW Acid

Cells

10-2,

lo-‘,

10-4,

10-5,

10-J,

4 x 10-4,

10-4, 10-S

Hemin

10-4,

5 x 10-5,

Biliverdin

10-4,

5 x 10--5,10-s,

10-6, 10-5,

IO-’

to.1

10-b


10-b

Range of Effective Concentration

per cell

04

2.2 (at 10-d M)

5 x 10-5-10-4

0.1-0.2

1 O-4 only

Compounds were added 4 hr after the cells had been subcultured to fresh growth medium. 5 days later, lysates were prepared and hemoglobin assays were performed by immunodiffusion. The growth rate and saturation density were similar to that for 10-a M hemtn with each compound at each concentration, with the exception of aminolevulinate, in which case cells failed to divide at 10-Z or 10-a M.

heme precursor (Falk, 1964), suggesting that the T3-Cl2 cells do not synthesize heme from biliverdin. Experiments with other heme-related compounds, such as bilirubin, were hindered by their low solubility in aqueous solvents (unpublished observations). Discussion These data demonstrate that hemin induces hemoglobin accumulation in two lines of erythroleukemic Friend cells. In the T3-Cl2 line, hemin is as effective an inducer as 1% DMSO (Table 1). Hemin represents a new class of erythroleukemic cell inducer, which differs both structurally and functionally from the polar solvent inducers. Its effect may be restricted to erythroid cells, since control SCRF lines, which are lymphoblastoid, fail to accumulate detectable quantities of hemoglobin when cultured with hemin (Table 2). It is the only known inducer that is active without a co-inducer and that has a regulatory role in primary erythroid cells derived directly from an animal source (Levere and Granick, 1965; Schulman, 1975). The major and somewhat unexpected finding is that growth in hemin, like DMSO, results in the accumulation of globin mRNA. These data are consistent with, but by no means prove, that the inducers affect globin gene transcription. Another possibility is that they somehow increase the stability of globin message. These experiments were not designed to determine the rate of synthesis of globin protein, since the immunodiffusion assay detects only the steady state levels of hemoglobin. For this reason, the data provide no conclusive information concerning potential effects of hemin on translation. Nevertheless, it is of interest that hemin induces globin mRNA and hemoglobin to approximately the same extent as does DMSO after 5 days of treatment of T3-Cl2 cells (Table 1, columns 4 and 6). In contrast, there are differences in the relative extents of induction of hemoglobin and globin mRNA with the two in-

&I,

tH CH.

BILIVERDIN

HEME Figure

4. The

Structures

of Heme

and Biliverdin

ducers in clone 745 cells (Table 2). In view of the limitations of the assays, however, the basis for these differences is not understood at this time. The mechanism of hemin induction could involve one or all of the following. First, hemin could function in a manner analogous to DMSO and the other polar solvents. It would then be mere coincidence that the inducer (hemin) was also closely linked to the developmental process itself. Most of the polar solvents are low molecular weight, aprotic compounds with planar configuration (Tanaka et al., 1975; Preisler and Lyman, 1975). Hemin is not a polar solvent, nor is it strictly planar, since the iron atom and the four pyrrole rings are tilted slightly out of the plane formed by the four methene bridge carbon atoms (Falk, 1964). Moreover, globin message induction kinetics in hemin differ from those in DMSO (Figures 2 and 3). These data indicate that hemin and DMSO differ functionally. Second, a contaminant in the hemin preparation could function as the inducing agent. The (Y and p absorbance maxima of hemin dissolved in pyridine as described by Falk (1964) were 558 and 524 rnp (unpublished observation), compared to the published values of 557 and 526 (Falk, 1964). Furthermore, a single spot with the expected Rr (0.34) was observed when a solution of 10-X M hemin was chromatographed in water, n-propyl alcohol, and pyridine, as described by Chu and Chu (1955; and unpublished data). Thus the hemin used in these experiments is very

Cell 518

pure, although a minor contaminant responsible for induction cannot be completely excluded. Third, hemin itself or a metabolite or breakdown product of hemin directly affects globin gene transcription or globin message stability. Hemin solutions for these experiments were prepared exactly as described for cell-free translation studies (Freedman, Geraghty, and Rosman, 1974), but the hemin was then incubated in aqueous growth medium, which could alter the structure of hemin (Falk, 1964). Since the mechanism of action of DMSO or hemin alone is not understood, the synergism between them (Figure 1, Tables 1 and 2) cannot be explained at this point. It seems improbable that DMSO facilitates the entry of hemin into cells. If it had, the onset of globin mRNA accumulation with both inducers together should have occurred as early as, or earlier than, with hemin alone, which was not the case (Figure 2). It may be of some interest to determine whether exogenous hemin at levels below 10-5 M stimulates the response of cells to DMSO. If so, this observation may in part explain why the extent of DMSO-induced hemoglobin accumulation varies with different types of sera (Paul and Hickey, 1974) or with different lots of the same serum (unpublished observations)-that is, in view of the synergism between hemin and DMSO, the variable response may be related to the amount of hemin in the serum. There are many examples of genetic control in procaryotes, in which regulatory molecules induce operon expression. In eucaryotes, specific receptor proteins for hormones (O’Malley and Means, 1974; Samuels and Tsai, 1973; Koerner et al., 1975) may induce the expression of one or more genes in differentiated cells. In this context, a direct analogy seems to exist between hemin and estrogen induction kinetics. The ovalbumin mRNA content of the chick oviduct increases as early as 3 hr after administration of estrogen to chicks previously treated with estrogen and then withdrawn (Cox, Haines, and Emtage, 1974). The rapidity of the responses with hemin or estrogen implies, but does not prove, a direct effect on gene expression. A related question concerns whether the hemin effect, like that of DMSO, is pleiotropic or is restricted to globin. DMSO induces accumulation of several proteins including delta-aminolevulinate synthetase (Ebert and Ikawa, 1974) and carbonic anhydrase (Kabat et al., 1975). DMSO, but not hemin, induces a 2-3 fold increase in acetylcholine esterase activity in T3Cl2 cells (P. Barald and J. Ross, manuscript in preparation), which suggests that the response to hemin is more restricted than to DMSO. The major question raised by these experiments, of course, concerns the role of hemin in globin gene expres-

sion in nucleated intact animal. Experimental

erythroid

precursor

cells

of the

Procedures

Cells and Culture Conditions T3-Cl2 and clone 745 cells were cultured in HAM F-12 medium containing 10% fetal calf serum (GIBCO, unheated) at 37°C in a humidified 5% CO2 atmosphere. They were subcultured every 3-5 days to maintain them in logarithmic growth. SCRF-145 is a cultured, X-ray-induced lymphoma cell from a BloAzR mouse. This cell produces an atropic C type virus that is noninfectious for N or B mouse, rabbit, rat, and human cells. SCRF-179 cells are virusnegative lymphoblastoid cells from the spleen of (NZC x NXB)F,. Both SCRF lines, which were provided by Dr. Fred C. Jensen (Scripps Clinic and Research Foundation, La Jolla, California), were cultured under the same conditions as were the erythroleukemia lines. Inducers (DMSO or hemin) were routinely added 4 hr after subculturing cells to fresh medium at a concentration of 0.5-2 x lo4 cells per ml. Hemin (bovine; Sigma) or biliverdin (Sigma) were dissolved in 0.2 M KOH; 1 ml of 0.2 M Tris-Cl (pH 7.8) and 2.6 ml Hz0 were added, and the pH was adjusted to 7.8 with 1 N HCI, the final concentration being IO-> M (Freedman et al., 1974). The solutions were then filtered through a 0.2 p membrane (Nalgene). Only freshly prepared solutions were used for these experiments, although frozen solutions provided identical results (unpublished observations). Control cell cultures routinely received pH 7.8 buffer, prepared as described above but lacking inducer. When DMSO and hemin were introduced into the same flask, DMSO was added first. Hemin was then added, usually l-10 min later, depending upon the number of flasks involved. Flasks were then gently shaken to mix the inducers. Delta-aminolevulinate and porphobilinogen (Sigma) were dissolved directly in growth medium and filtered as described above. Immunological Studles Hemoglobin was purified from membrane-free lysates of washed mouse erythrocytes on CM-Sephadex with a linear gradient [O.Ol M phosphate buffer (pH 6.8) to 0.02 M phosphate buffer (pH 9.1)]. Globin was prepared from this purified hemoglobin by acid acetone precipitation. Antibody to mouse globin was raised in sheep injected once with 14 mg of globin in complete Freund’s adjuvant and then twice more with 3.5 mg of globin in incomplete Freund’s adjuvant. Serum obtained from the same animal before the first (preimmune) and after the third (immune) injection was heated to 56°C for 30 min prior to use. Assays were performed at room temperature by adding 10 pl of cell lysate (see below) and immune serum to separate wells on an immunodiffusion plate (Hyland) and scoring for precipitin bands 18 hr later. Preimmune serum failed to react with authentic mouse hemoglobin, while immune serum reacted with mouse but not with human or duck hemoglobin (unpublished observations). A single precipitin band was observed with freshly prepared mouse hemoglobin as antigen. The band reacted rapidly with benzidine, forming a dark blue color; large amounts of heme were therefore present in the precipitated protein. The precipitin band with lysates from inducer-treated T3-Cl2 cells formed a line of identity with authentic adult mouse hemoglobin. Cell lysates were prepared by resuspending washed cells at 4°C in lysis buffer [O.Ol M Tris-Cl (pH 7.2) 0.1 M NaCI, 0.0015 M MgCl?, 0.5% (v/v) Triton X-l 001 at a concentration of l-5 x 105 cells per pl. After vigorous shaking for 15 set, nuclei were pelleted at 4°C for 10 min at 1500 rpm in the PR-J centrifuge. The supernatant was removed and IO ~1 were used in the immunodiffusion assay. Where noted, lysates were diluted in water before application to the immunodiffusion plate. The minimum detectable hemoglobin concentration in all experiments reported here was 0.1 pg per cell, as determined by reconstruction experiments using known quantities of purified hemoglobin as a standard (data not shown).

Globin 519

mRNA

Induction

with

Hemin

RNA Isolation and Hybridization Assays Total cell RNA was isolated by a modification of the cesium chloride technique described by Glisin, Crkvenjakov, and Byus (1974). Cells were harvested by centrifugation and washed once in cold F-12 medium. The cell pellet was then lysed at room temperature by adding 2.7 ml of lysis buffer [4% (w/v) sarkosyl, 0.05 M Tris-Cl (pH 7.2) 0.05 M EDTA] per 5 x 106 cells. The lysate was mixed vigorously for 4 min and then 1 .O g of cesium chloride (Schwarz/ Mann, optical grade) was added. After shaking vigorously for 2-3 min more, the lysate was layered over a 1.5 ml cushion of 5.7 M cesium chloride in a nitrocellulose SW 56 tube. The cesium chloride cushion solution had been filtered prior to use as described by Glisin et al. (1974). After centrifugation in the SW 56 rotor at 32,000 rpm for 15-18 hr in the Beckman L2-65B ultracentrifuge at 25”C, the tubes were drained and cut exactly as described by Glisin et al. (1974). Control experiments demonstrated that recovery of total RNA and of globin mRNA sequences by this method was the same and often 5-15% greater than standard methods using phenol extraction and DNAase (J. Ross, unpublished observations). Hybridization assays for globin mRNA were performed as previously described (Ross et al., 1974) with the following modifications. Reactions were carried out in a final volume of 0.03 ml in 6 x 50 mm glass tubes containing approximately 100 pg (500 cpm) of ‘H-cDNA, 0.6 M NaCI, 0.13 mM EDTA, 0.001 M Tris-Cl (pH 7.2), 0.015% (w/v) SDS, 30 pg tRNA, and 4 pg calf thymus DNA. After addition of all components, reactions were covered with paraffin oil and incubated at 65°C. Sl nuclease assays were performed as described (Ross et al., 1974). Acknowledgments We are grateful to Dr. Fred Jensen for SCRF cells; to Dr. Barry Whitney for assistance with the benzidine staining technique; and to Drs. Gerald Mueller and David Shemin for critical reviews of the manuscript. This investigation was supported by a grant from the NIH. Received

February

19, 1976;

revised

May 5. 1976

References Adamson, S. D., Herbert, E., and Kemp, S. F. (1989). Effects of hemin and other porphyrins on protein synthesis in a reticulocyte lysate cell-free system. J. Mol. Biol. 42, 247-258. Altman, P. L. (1981). Blood and Other D.C.: Federation of American Societies and Medicine), p. 116.

Body Fluids Washington. for Experimental Biology

Beuzard. Y., Rodvien, R., and London, I. M. (1973). on the synthesis of hemoglobin and other proteins cells. Proc. Nat. Acad. Sci. USA 70, 1022-1026.

Effect of hemin in mammalian

Boyer, S. H., Wuu, K. D.. Noyes, A. N., Young, R., Scher, W., Friend, C., Preisler, H. D., and Bank, A. (1972). Hemoglobin biosynthesis in murine virus-induced leukemic cells in vitro: structure and amounts of globin chains produced. Blood 40, 823-835. Bruns, G. P., and London, I. M. (1965). The effect of hemin on the synthesisof globin. Biochem. Biophys. Res. Commun. 18,236-242. Chu, T. C., and Chu, E. J. (1955). Paper chromatography complexes of porphyrins. J. Biol. Chem. 212, 1-7.

of iron

Cox, R. F.. Haines. M. E., and Emtage, J. S. (1974). Ouantitation of ovalbumin mRNA in hen and chick oviducts by hybridization to complementary DNA. Eur. J. Biochem. 49, 225-238. Ebert. P.. and Ikawa, Y. (1974). Induction of I-aminolevulinic acid synthetase during erythroid differentiation of cultured leukemia cells. Proc. Sot. Exp. Biol. Med. 746. 601-604. Falk. J. E. (1964). Elsevier).

Porphyrins

and

Metalloporphyrins

(New

York:

Freedman, M. L., Geraghty, M.. and Rosman, trol of globin synthesis. J. Biol. Chem. 249,

J. (1974). Hemin 7290-7294.

con-

Friend, C., Scher, W., Holland, J. G., and Sato, T. (1971). Hemoglobin synthesis in murine virus-induced leukemic cells in vitro: stimulation of erythroid differentiation by dimethysulfoxide. Proc. Nat. Acad. Sci. USA 68, 378-382. Furusawa, erythrocyte of Friend

M., Ikawa, Y., and Sugano, H. (1971). Development of membrane-specific antigen(s) in clonal cultured cells virus-induced tumor. Proc. Jap. Acad. 47, 220-224.

Gallien-Lartigue. O., and Goldwasser, E. (1964). Hemoglobin synthesis in marrow cell culture: the effect of rat plasma on rat cells. Science 145. 277-279. Glisin, V., Crkvenjakov, isolated by cesium 2633-2637.

R.. and Byus, C. (1974). Ribonucleic chloride centrifugation. Biochemistry

acid 13,

Goldstein, K., Preisler, H. D., Lutton, J. D., and Zanjani, E. D. (1974). Erythroid colony formation in vitro by dimethylsulfoxidetreated erythroleukemic cells. Blood 44, 831-836. Ikawa, Y., Furusawa, M., and Sugano, H. (1973). Erythrocyte membrane antigen in Friend leukemia cells, Bibl. Haematol. 39, 955-967. Ikawa, Y., Aida, M., and Saito, M. (1975). in cultured Friend leukemia cells induced extract preparation. Gann 66. 583-584.

Hemoglobin production by a human placental

Kabat, D., Sherton, C. C.. Evans, L. M., Bigley, R., and Koler, R. D. (1975). Synthesis of erythrocyte-specific proteins in cultured Friend leukemia cells. Cell 5, 331-338. Koerner, D., Schwartz, H. L.. Surks. M. I,, Oppenheimer, J. H., and Jorgensen, E. C. (1975). Binding of selected iodothyronine analogues to receptor sites of isolated rat hepatic nuclei: high correlation between structural requirements for nuclear binding and biological activity. J. Biol. Chem. 250, 6417-6423. Kruh, J.. and Borsook, H. (1956). Hemoglobin synthesis reticulocytes in vitro. J. Biol. Chem. 220, 905-915.

in rabbit

Leder, A., and Leder, P. (1975). Butyric acid, a potent inducer of erythroid differentiation in cultured erythroleukemic cells. Cell 5, 319-322. Levere, R. D., and Granick, S. (1965). Control of hemoglobin synthesis in the cultured chick blastoderm by {-aminolevulinic acid synthetase: increase in the rate of hemoglobin formation with {aminolevulinic acid. Proc. Nat. Acad. Sci. USA 54, 134-137. Levy, J., Terada. M., Rifkind. R. A., and Marks, P. A. (1975). Induction of erythroid differentiation by dimethylsulfoxide in cells infected with Friend virus: relationship to the cell cycle. Proc. Nat. Acad. Sci. USA 72, 28-32. McClintock, P. R., and Papaconstantinou, J. (1974). Regulation of hemoglobin synthesis in a murine erythroblastic leukemic cell: the requirement for replication to induce hemoglobin synthesis. Proc. Nat. Acad. Sci. USA 71, 4551-4555. Morell, H., Savoie, J. C., and London, I. M. (1958). The biosynthesis of heme and the incorporation of glycine into globin in rabbit bone marrow in vitro. J. Biol. Chem. 233, 923-929. O’Malley, B. W.. and Means, A. R. (1974). Female and target cell nuclei. Science 183, 61 O-620.

steroid

hormones

Paul, J., and Hickey, I. (1974). Haemoglobin synthesis in inducible, uninducible and hybrid Friend cell clones, Exp. Cell Res. 87, 20-30. Preisler, H. D., and Giladi, M. (1974). Erythropoietin responsiveness of differentiating Friend leukemia cells. Nature 257, 645-646. Preisler. H. D., and Lyman, G. (1975). Differentiation kemia cells in vitro: properties of chemical inducers. tiation 4, 179-l 85.

of erythroleuCell differen-

Preisler, H. D., Housman, D., Scher, W., and Friend, C. (1973). Effects of 5-bromo-2’-deoxyuridine on production of globin messenger RNA in dimethylsulfoxide-stimulated Friend leukemia cells. Proc. Nat. Acad. Sci. USA 70, 2956-2959.

Cell 520

Ross, J., Ikawa, Y., and Leder, P. (1972). Globin messenger induction during erythroid differentiation of cultured leukemic Proc. Nat. Acad. Sci. USA 69, 3620-3623.

RNA cells.

Ross, J., Gielen, J., Packman, S.. Ikawa, Y., and Leder. P. (1974). Globin gene expression in cultured erythroleukemic cells. J. Mol. Biol. 87, 697-714. Samuels, H. H., and Tsai, J. S. (1973). Thyroid hormone action in cell culture: demonstration of nuclear receptors in intact cells and isolated nuclei. Proc. Nat. Acad. Sci. USA 70, 3468-3492. Schulman, H. M (1975). Evidence supporting for the hemin-controlled translational repressor in reticulocytes. Biochim. Biophys. Acta 414,

a physiological role of globin synthesis 161-166.

Tanaka, M., Levy, J., Terada, M., Breslow, R., Rifkind, R. A., and Marks, P. A. (1975). Induction of erythroid differentiation in murine virus infected erythroleukemia cells by highly polar compounds. Proc. Nat. Acad. Sci. USA 72, 1003-1006. Terada, M., Cantor, L., Metafora, S., Rifkind, R. A., Bank, A., and Marks, P. A. (1972). Globin messenger RNA activity in erythroid precursor cells and the effects of erythropoietin. Proc. Nat. Acad. Sci. USA 69, 3575-3579.