Max and Inhibitory C-Myc Mutants Induce Erythroid DiErentiation and Resistance to Apoptosis in Human Myeloid Leukemia Cells

Max and inhibitory c-Myc mutants induce erythroid dierentiation and resistance to apoptosis in human myeloid leukemia cells

Matilde CanÄ elles1, M Dolores Delgado1, Kathy M Hyland2, Ana Lerga1, Carlos Richard3, Chi V Dang2 and Javier LeoÂ n1

1Departamento de BiologõÂa Molecular, Facultad de Medicina, 39011 Santander, Spain; 2Division of Hematology, Department of Medicine, The John Hopkins University School of Medicine, Baltimore, Maryland 21205; 3Servicio de HematologõÂa, Hospital Universitario MarqueÂs de Valdecilla, 39011 Santander, Spain

We have used the human leukemia cell line K562 as a site (Ems) (Blackwood and Eisenman, 1991; Prender- model to study the role of c-myc in dierentiation and gast et al., 1991). Myc/Max activates transcription of apoptosis. We have generated stable transfectants of reporter genes carrying Ems in their promoters, while K562 constitutively expressing two c-Myc inhibitory the homodimer Max/Max is inactive as a transcrip- mutants: D106-143, that carries a deletion in the tional activator (Kretzner et al., 1992; Kato et al., transactivation domain of the protein, and In373, that 1992; Amin et al., 1993; Gu et al., 1993). Also, the carries an insertion in the DNA-interacting region. We overexpression of Max results in suppression of cell show here that In373 is able to compete with c-Myc for transformation mediated by c-Myc (Makela et al., Max binding and to inhibit the transformation activity of 1992; Prendergast et al., 1991; Mukherjee et al., 1992; c-Myc. K562 cells can dierentiate towards erythroid or Amati et al., 1993a; Gu et al., 1993; Koskinen et al., myelomonocytic lineages. K562 transfected with c-myc 1994; Lindeman et al., 1995). These data are consistent mutants showed a higher expression of erythroid with the idea that inactive Max/Max dimers compete in dierentiation markers, without any detectable eects vivo with Myc/Max for common DNA binding sites. in the myelomonocytic dierentiation. We also trans- Expression of c-myc is induced during mitogenic fected K562 cells with a zinc-inducible max gene. Ectopic stimulus and required for cell growth, while terminal Max overexpression resulted in an increased erythroid dierentiation of many cell types is accompanied by dierentiation, thus reproducing the eects of c-myc down-regulation of c-myc expression (Blackwood et al., inhibitory mutants. We also studied the role of c-myc 1992; Marcu et al., 1992; Kato and Dang, 1992; mutants and max in apoptosis of K562 induced by Meichle et al., 1992; Evan and Littlewood, 1993). okadaic acid, a protein phosphatases inhibitor. The Myeloid leukemia cell lines have been broadly used as expression of D106-143 and In373 c-myc mutants and models to study the molecular basis of the prolifera- the overexpression of max reduced the apoptosis tion-dierentiation switch. Following induction of mediated by okadaic acid. The common biochemical dierentiation in myeloid cell lines, c-myc is down- activity of D106-143 and In373 is to bind Max and regulated. This has been found in murine (M1, MEL, hence to titrate out c-Myc to form non-functional Myc/ WEHI3B) and human cell lines (HL60, U937) Max dimers. Similarly, Max overexpression would (reviewed in Marcu et al., 1992). Consistently, decrease the relative levels of c-Myc/Max with respect constitutive expression of c-myc inhibits the chemi- to Max/Max. The results support a model where a cally-induced dierentiation of some of these cell lines threshold of functional c-Myc/Max is required to (Chisholm et al., 1992; Coppola and Cole, 1986; maintain K562 cells in an undierentiated state and to Dmitrovski et al., 1986; Larsson et al., 1988) and undergo drug-mediated apoptosis. inhibition of c-myc expression induce dierentiation (Holt et al., 1988; Prochownik et al., 1988; Nguyen et Keywords: Max; c-Myc mutants; K562; erythroid al., 1995). Paradoxically, it has been shown that c-Myc dierentiation; apoptosis; okadaic acid overexpression induces apoptosis in dierent cell systems under conditions where c-Myc is normally down-regulated (reviewed in Harrington et al., 1994; Packham and Cleveland, 1995). The apoptosis-promot- Introduction ing activity of c-myc was originally observed in rodent ®broblasts but also takes place in murine myeloid cells c-Myc contains a transcriptional activation domain deprived of IL3 or exposed to antiproliferative and a basic/helix ± loop ± helix/leucine zipper (bHLH- cytokines (Askew et al., 1991; Lotem and Sachs, LZ) domain that mediates sequence-speci®c DNA 1993; Selvakumaran et al., 1994). binding and heterodimerization with Max, itself Human K562 is a bipotential cell line that expresses another bHLH-LZ protein. All the known biological erythroid markers and can be further dierentiated in functions of c-Myc depend on its dimerization with vitro towards erythroid lineage with 1-a-D-arabinofur- Max (reviewed in Amati and Land, 1994; Vastrik et al., anosylcytosine (ara-C) and other inducers (Rowley et 1994). Myc/Max heterodimer binds to a DNA al., 1981), or towards myelomonocytic lineage (with consensus sequence CACGTG, termed E Box Myc expression of monocytic and megakaryocytic markers) with phorbol esters (Kamano et al., 1990; Shen et al., Correspondence: J LeoÂ n 1992). We have previously shown that c-myc is down- Received 18 July 1996; revised 8 November 1996; accepted 11 regulated during the erythroid and myelomonocytic November 1996 dierentiation of K562 cells (Gomez-Casares et al., Max and Myc in differentiation and apoptosis of K562 cells MCanÄelles et al 1316 1993) and that c-myc overexpression leads to a partial we have produced and puri®ed these hexahistidine inhibition of erythroid dierentiation in K562 cells tagged proteins from bacteria (Figure 1a). Although (Delgado et al., 1995). On the other hand, although In373 contains four extra amino acids, the polylinker K562 cells are relatively resistant to apoptosis the sequence of pDS-Myc(342 ± 439) contributes to amino protein phosphatases inhibitor okadaic acid (OA) acids that exceed the In373 polypeptide by four amino readily induces apoptosis in K562 (Zheng et al., 1994; acids. Hence the In373 polypeptide showed a faster Lerga et al., 1995). Therefore K562 cell line provides a mobility on SDS ± PAGE. useful model system to study the involvement of c-Myc In373 is unable to bind DNA (Figure 1b, lane 4), in growth, dierentiation and apoptosis. In the present whereas tMyc forms a distinct homodimeric complex work we have investigated the eects of dominant- (Figure 1b, lane 3). In this experimental design, Max negative c-myc mutants and max on dierentiation and (6 ng) was mixed with an excess of tMyc (600 ng) so that apoptosis of K562. Two mutant c-myc genes were tMyc-tMyc homodimers and tMyc-Max heterodimers used: D106-143 and In373. These mutants were chosen are readily detected on EMSA (Figure 1b, lane 5). The because: (i) they are completely inactive for Rat-1A Max-Max homodimer existed at low levels under these cells transformation and rat embryo cells (REC) Myc- conditions. Addition of increasing amounts In373 Ras cotransformation (Stone et al., 1987); (ii) they are (Figure 1b, lanes 6 to 9) resulted in diminishing intensity two of the most potent dominant negative c-myc of shifted bands corresponding to tMyc/Max, Max/Max mutants inhibiting the transformation of Rat-1 cells and tMyc/tMyc. These results indicate that recombinant by v-ABL or BCR/ABL, and in the case of D106-143, In373 protein is able to disrupt DNA binding by tMyc inhibiting the cotransformation of REC by Myc-Ras and Max homodimers and heterodimers. (Dang et al., 1989; Sawyers et al., 1992), and (iii) they To determine whether In373 is able to heterodimer- carry alterations in two dierent domains of c-Myc ize with Max in vivo, we tested the abilities of these protein: the transactivation domain and the speci®c proteins to interact using a mammalian two-hybrid DNA binding domain. The Max-dimerizing region is assay that had been reported previously (Dang et al., not aected in the mutants used. D106-143 lacks 1991). The dimerization domains of wild-type c-Myc or amino acids 106 to 143 of wild-type c-Myc. This region of In373 (including the insertion mutation) were fused is able to act as an independent transcriptional to the Ga14 DNA binding domains. Constructs activation domain and is responsible for the transfor- encoding these proteins were cotransfected with either mation and transrepression activities of c-Myc (Kato et the control pNLVP plasmid, which encode the VP16 al., 1990; Li et al., 1994; Brough et al., 1995; Lee et al., activation domain, or with pVPMax that encodes Max 1996). The Myc mutant In373 carries an insertion in fused to the VP16 activation domain. The interaction the DNA binding region (Stone et al., 1987) and we of these hybrid proteins would reconstitute the show now that it binds to Max, but the dimer is unable function of GAL4 to transactivate a GAL4 driven to bind DNA. We present evidence suggesting that the reporter, G5-E1B-CAT (Figure 1c). When compared to expression of either c-myc inhibitory mutants or max the controls with pNLVP, both wild-type c-Myc and enhance the erythroid dierentiation of K562 cells In373 sequences were able to interact with Max in without any detectable eect on myelomonocytic transfected cells. The protein levels of GAL4-Myc and dierentiation. Furthermore, we have found that GAL4-In373 are comparable in transfected CHO cells expression of max and c-myc mutants signi®cantly as determined by immunoblotting (data not shown). reduces the drug-mediated apoptosis of K562. These results indicate that the insertion at position 373 did not disrupt Myc ability to interact with Max. Although the In373 mutant was shown to be non- transforming (Stone et al., 1987) and to interrupt Results oncogenic Abl-mediated transformation (Sawyers et al., 1992), it is unknown whether it is able to inhibit c- Characterization of the In373 mutant Myc transforming activity. Therefore we analysed its To study c-Myc role in dierentiation of K562 cells, we eect on a cotransformation assay of REC mediated set out to obtain cell lines constitutively expressing c- by ras and myc. Similar to the D106-143 mutant (Dang myc inhibitory mutants. We used for transfections two et al., 1989), In373 is able to dramatically inhibit potent dominant negative c-myc mutants: D106-143 and transformation of REC by wild-type c-Myc and In373. As the biochemical activities of In373 are activated Ras. As a control, we also cotransfected unknown, we ®rst analysed its ability to bind DNA, In6, another insertional mutant which does not act as dimerize with Max and inhibit c-Myc cotransformation an inhibitory mutant (Stone et al., 1987). Cotransfec- activity. Previous studies have demonstrated that c-Myc tion of In6 increased the number of foci in the presence alone is unable to bind the E-box sequence, CACGTG of wild-type c-Myc and activated Ras (Figure 1d). This unless at very high concentrations. However, a was the expected result as In6 is able to cotransform truncated form of c-Myc (tMyc), which contains amino REC. Together with previous results in stably acids 342 ± 439, is able to bind the E-box sequence as a transfected cells (Sawyers et al., 1992), our results homodimer (Kato et al., 1992). When mixed with Max indicate that In373 is a potent dominant negative protein, tMyc is able to form heterodimers with Max inhibitor of c-Myc transforming activity. that bind DNA resulting in three resolvable protein- DNA complexes: tMyc-tMyc, Max-Max and tMyc- Generation of K562 cell lines stably transfected with c- Max. To determine if the dominant negative tIn373 Myc mutants protein, which also contains Myc amino acids 343 ± 439 with an insertion of four serins at residue 373, is able to The mutants D106-143 and In373 were transfected into aect the DNA binding properties of these complexes, K562 cells by electroporation and several G418- Max and Myc in differentiation and apoptosis of K562 cells MCanÄelles et al 1317

ab In373 –––+– ++– 123456789101112

Std tMyc In373 Max

33 Max-Max — tMyc-Max — 22

14 tMyc-tMyc —

free — probe

Max –+– –+++++–+ – tMyc ––+–++++++––

c d

Figure 1 Characterization of In373. (a) Puri®ed recombinant hexahistidine fusion protein of tMyc, In373 or Max were resolved on SDS ± PAGE and stained with Coomassie blue. The left lane (Std) shows prestained standards and the molecular masses (kDa) are indicated on the left margin. For Myc and In373, a contaminating 30 kDa protein is noted. (b) The Myc mutant In373 inhibits DNA binding by tMyc and Max in electrophoretic mobility shift assay. The probe is a 260 bp CACGTG containing oligonucleotide (pDW14) that resolves the dierent DNA bound dimers (indicated at the left margin). The amounts of recombinant proteins in each reaction were: Max, 6 ng; tMyc, 0.6 mg; tIn373, 0.6 to 3.3 mg (lanes 5 to 9). (c) Mammalian two-hybrid assay demonstrates an interaction between Max and either wild-type Myc or In373. Gal-Myc is the GAL4 DNA binding domain fused to c-Myc amino acids 262 ± 439, Gal-In373 contains c-Myc amino acids 262 ± 439 with an insertion of four serines at residue 373. pNLVP encodes the VP16 transactivation domain. VP-Max is Max fused at its N-terminus to the VP16 transactivation domain. The CAT activities were derived from duplicated experiments and normalized to the controls with pNLVP. (d) In373 inhibits Myc cotransformation of rat embryo cells in the presence of activated H-Ras (EJ-ras). Each 100 mm plate (in quadruplicates) of rat embryo cells were lipofected with 5 mg of Ras plasmid alone or with 5 mg of Myc expression plasmid. Cotransfections were with 5 mg of either In373 or In6 genes. Transformed foci per plate are indicated on the ordinate

resistant clones were selected. In order to assess the fragment and of extrabands for the ®ve analysed stable integration of the transfected genes into the cell clones. The intensity of the band of the transfectants as genome, three K562 sublines transfected with D106- compared with the corresponding of parental cells 143 (KD2, KD4 and KD11) and two transfected with indicates that transfected cells carried multiple copies In373 (KIA and KIB) were subjected to Southern of the c-myc gene. Northern blotting analysis showed analysis. The results (Figure 2a) indicate that the in all ®ve cases an augmented expression of steady- transfected c-myc genes are integrated in the genome of state c-myc mRNA with respect to non-transfected the ®ve analysed clones. The hybridization revealed the K562 cells (Figure 2b). The increase ranged from three- presence of multiple copies of the c-myc EcoRI to 10-fold, as assessed by ®lm densitometry (not Max and Myc in differentiation and apoptosis of K562 cells MCanÄelles et al 1318 shown). The K562 clones transfected with D106-143 same size that the endogenous c-Myc. To be used as (KD2, KD4 and KD11) were analysed for mutant Myc controls, several K562 lines transfected with a plasmid protein expression by Western blotting, and for all the carrying a G418 resistance gene were generated cases a c-Myc doublet was observed, corresponding to (Kneo2, Kneo3, Kneo5, Kneo6, Kneo8). the endogenous and exogenous c-Myc proteins (Figure 2c). As reported (Stone et al., 1987), the mobility of c-Myc mutants enhance erythroid dierentiation of K562 D106-143 protein was smaller than that of the endogenous protein. In sharp contrast with the Growth of K562 cells transfected with mutant c-myc mRNA expression, the levels of D106-143 protein in genes was reduced by about 25% with respect to the transfected cells were 4 ± 6-fold lower (by ®lm parental cells and vector-transfected cells, as assessed densitometry) than those of the wild-type c-Myc. No by [3H]thymidine incorporation and cell counts. Also, information could be obtained on the amount of In373 both mutants drastically reduced K562 clonogenicity in in the transfectants as the mutant protein has about the agar after transfection (not shown). In contrast to the cells expressing mutant c-myc genes, cells overexpres- sing wild-type c-myc (KmycJ cells in the presence of zinc) or Kneo cells did not show signi®cant dierences in their growth rate with respect to uninduced or a K562 KD2 KD4 KD11 KIA KIB M parental cells (not shown). We studied the eects of c- — 21,2 Myc mutants in myelomonocytic and erythroid dierentiation. Transfectant cells did not show any increase of myelomonocytic/megakaryocytic markers as — 5,1 expression of GpIIb ± IIIa, cluster formation, adher- — 4,2 ence to plastic and fraction of cells positive for the — 3,5 nitroblue tetrazolium reduction test. Furthermore, when myelomonocytic dierentiation was induced by treatment with 10 nM TPA for 3 days, the extent of —2 dierentiation of KD2, KD4, KIA and KIB cell lines was similar to that of parental K562 cells or Kneo cell lines, as assayed by cluster formation and nitroblue tetrazolium reduction (about 30% of positive cells in b all lines) (not shown). Therefore, myelomonocytic K562 KD2 KD4 KD11 KIA KIB dierentiation of K562 seemed to be unaected by expression of Myc mutants. We then analysed the erythroid dierentiation in c-myc K562 cells expressing inhibitory c-myc genes, as compared with the parental cells or neo expressing cells. Cell lines KD2, KD4, KIA and KIB showed a signi®cant increase in the erythroid dierentiation as assessed by the expression of e-globin mRNA, an embryonic globin expressed in K562 (Charnay and rRNAs Maniatis, 1983) (Figure 3a). The amounts of globin mRNAs are compared in Figure 3b to show that the basal expression of e-globin is 2 ± 3-fold higher than in parental cells. This result was con®rmed by the cytochemical reaction of benzidine, which detects hemoglobinized cells. The fraction of benzidine- c positive cells in KD2, KD4, KIA and KIB cells was K562 KD2 KD4 KD11 M signi®catively higher than in parental K562 cells and —97 neo transfectants (Figure 3c). No major dierences in the extent of erythroid dierentiation were found MM Myc between the D106-143 and In373 transfectants. The D106-143 percentage of benzidine-positive cells in the transfec- —50 tants was similar to that of parental K562 cells treated with ara-C. This drug induces erythroid dierentiation Figure 2 Analysis of K562 cell lines stably transfected with c- of K562 accompanied by irreversible growth arrest and myc mutants. (a) Presence of c-myc exogenous sequences in K562 sublines tranfected with D106-143 (KD2, KD4 and KD11) and more pronounced erythroblastoid morphology (Rowley with In373 (KIA, KIB). High-molecular weight DNAs from the et al., 1981). Ara-C induces an 8 ± 10-fold increase in cell lines were subjected to Southern analysis as described in the the fraction of hemoglobinized cells in K562 and Kneo text, and hybridized to a human c-myc probe. The size of markers clones, but only 2 ± 3-fold in the D106-143 and In373 in kb are indicated at the right. (b)c-myc mRNA expression in transfected K562 cell lines. Total RNAs were prepared from the transfectants (Figure 3c). indicated cell lines and analysed by Northern analysis using a human c-myc probe. The lower panel shows the ®lter stained with ethidium bromide to assess RNA integrity and loading. (c) Generation of cell lines transfected with inducible max Immunoblot of c-Myc proteins in K562 and KD2, KD4 and gene KD11 cell lines. The position of endogenous normal c-Myc and D106-143 mutant are indicated. The position of the molecular Increased levels of Max would result in the higher weight markers (kDa) are indicated at the right relative levels of Max/Max inactive homodimers that Max and Myc in differentiation and apoptosis of K562 cells MCanÄelles et al 1319 a c K562 KD2 KD4 KIA KIB

CACACACACA ∋ -globin

rRNAs

Figure 3 K562 sublines expresssing c-myc mutants show enhanced erythroid dierentiation. (a) e-globin mRNA expression of transfected cells. RNAs from the indicated cell lines were prepared from control cells (lanes `C') and cells treated for 3 days with 1 mM ara-C (lanes `A'). (b) Quanti®cation of e-globin mRNA signals of the above Northern blot with respect to the 18s rRNA. The radioactivity of the signals was determined as described in the text and represented with respect to the maximum value. (c) Fraction of benzidine-positive cells in transfectant cell lines. Benzidine test was carried out in untreated growing cells and cells treated for 3 days with 1 mM ara-C. The maximum score (KD2 cells treated with ara-C, 58%) was set at 100% and the other data were normalized to this value. Bars indicate standard deviations from three separate experiments

compete with c-Myc/Max for common DNA binding dierences were observed in the growth rate of sites. To investigate the eects of Max overexpression transfected cells upon addition of 50 mM ZnSO4 (not in K562 we transfected an expression vector where max shown). The addition of ara-C to Kmax12 and cDNA is under the control of murine metallothionein Kmax16 resulted in a further increase in benzidine- promoter and can be induced by the addition of zinc to positive cells (about 2-fold). This increase, however, the media. The plasmid was electroporated into K562 was much smaller than that observed for parental and transfectant cells were selected with hygromycin B K562 cells (8 ± 10-fold) (Figure 5). In contrast, we did as described in Materials and methods. Two cell lines, not detect any increase in the expression of markers of Kmax12 and Kmax16 were selected which showed a myelomonocytic dierentiation (reduction of nitroblue marked increase in max mRNA (Figure 4a) and tetrazolium, adherence, expression of GpIIb-IIIa) upon protein (Figure 4b) upon addition of 75 mM ZnSO4. induction of Max expression by zinc. Also, the The increase in max expression with zinc was dose- induction of Max expression did not alter the growth dependent. Both cell lines express higher amounts of rate of Kmax12 and Kmax16 (not shown). Addition of max than parental cells in the absence of zinc addition, zinc did not induce any erythroid dierentiation in due to the transcriptional leakiness of the metallothio- KMMT cells, transfected with the empty vector nein promoter. (pHEBoMT) (Figure 5).

Max overexpression enhances erythroid dierentiation of Expression of c-Myc mutants impairs apoptosis of K562 K562 As c-Myc mediates apoptosis in dierent cell types To monitor changes in the extent of erythroid deprived of growth factors, we ®rst investigated dierentiation, we determined the fraction of benzi- whether this eect is reproduced in K562 cells. Cells dine positive cells in Kmax12 and Kmax16 cells in the were shifted to medium containing 2% or no serum for presence and absence of zinc. When the expression of 24 or 48 h and apoptosis was assessed by DNA

Max was induced by 50 mM ZnSO4, a signi®cant fragmentation and scoring of apoptotic cells after increase in the fraction of hemoglobinized cells was Giemsa staining. As shown in Figure 6a, K562 cells observed for both cell lines, being higher for Kmax12 showed some DNA degradation after 2 days of culture (Figure 5). The fraction of benzidine-positive cells after in the absence of serum. The fraction of apoptotic cells zinc treatment was close to that observed for parental at this point was only 6% as determined by cells treated with ara-C. An increased number of cells morphological analysis of Giemsa-stained prepara- with morphology compatible with that of basophilic tions. However, cells grew in the presence of 2% of erythroblasts was also observed (not shown). No serum without any sign of apoptosis. We analysed Max and Myc in differentiation and apoptosis of K562 cells MCanÄelles et al 1320

a Kmax12 Kmax16

0 25 50 100 200 0 25 50 100 200 µM Zn

max

rRNAs

b Kmax12 Kmax16 0 2 6 12 24 0 2 6 12 24 hours

max

rRNAs

Figure 5 Induction of max expression results in erythroid dierentiation. K562, KMMT, Kmax12 and Kmax16 cells were exposed to either 50 mM ZnSO4,1mM ara-C or to 50 mM ZnSO4 for 4 h prior ara-C induction and then to 1 mM ara-C for three c days. Untreated cells were used as control. The maximum score Kmax12 Kmax16 (Kmax12 cells treated with Zn+ara-C, 65%) was set at 100% M and the other data were normalized to this value. Bars indicate K562 0 12 24 48 0 12 24 48 hours standard deviations from three separate experiments 27,5 —

L p21/22 Max 16,5 — Figure 4 Induction of max expression in transfected K562 cell lines. Northern analysis of max expression in Kmax12 and a K562 KmycJ Kmax16 sublines exposed to the indicated concentrations of ZnSO4 for 4 h (a)orto75mMZnSO4 for the indicated times (b). 0% 2% 0% 2% FCS Lower panels show the ®lters stained with ethidium bromide ––++––++ ––++––++ Zn showing the rRNAs. (c) Immunoblot of Max in K562, Kmax12 M 12121212 12121212 days and Kmax16 cells exposed to 75 mM ZnSO4 for the indicated times. Molecular mass markers (kDa) are indicated at the left

internucleosomal DNA fragmentation and the fraction of apoptotic cells in KmycJ cells deprived of fetal calf serum in the presence and absence of zinc. We found that the overexpression of c-myc induced by 75 mM b 0% FCS 2% FCS ZnSO4 only resulted in a modest increase of apoptosis as assessed by internucleosomal DNA fragmentation K562KD2KD4 KIAKIB K562 KD2 KD4 KIA KIB (Figure 6a). The fraction of apoptotic cells after 2 days M2323232323 232323232 3 of serum deprivation was about 4% and rose to 8% in days the presence of zinc (mean values from three independent experiments). The same experiment was performed with cells expressing c-Myc inhibitory mutants, and we observed no signi®cant changes in the extent of DNA fragmentation in KD2, KD4, KIA and KIB with respect to parental cells after 3 days of incubation in low-serum medium (Figure 6b). There- Figure 6 (a) Apoptosis of K562 and KmycJ cells incubated in fore, the extent apoptosis induced by serum depriva- the absence of serum. K562 and KmycJ cells were grown for one tion in K562 is low and shows little change by the or two days in the absence or presence of 2% fetal calf serum (FCS) and in the absence or presence of 75 mM ZnSO4 as expression of c-Myc or inhibitory c-Myc mutant indicated at the top of the ®gure. Apoptosis was determined by proteins. internucleosomal DNA fragmentation as described in the text. (b) In view of the low level of apoptosis induced by Apoptosis of cells expressing dominant negative c-myc mutants. serum deprivation in K562 cells, we set out to study a The indicated cell lines were treated for two or three days in the possible role of c-Myc in drug-induced apoptosis. K562 absence or presence of 2% serum as indicated. Apoptosis was determined by detection of internucleosomal DNA fragmentation are remarkably resistant to apoptosis mediated by as above. Lane M refers to DNA molecular weight markers many drugs, but okadaic acid (OA), a serine/threonine (lambda DNA digested with EcoRI and HindIII) Max and Myc in differentiation and apoptosis of K562 cells MCanÄelles et al 1321 a a ––––++++++ Zn K562KD2KIA KIB K562 KD4 –++++++––– OA MMCCCOACOAC OAC OA OA OA 01231231/4 13days

c-myc

12345678 910 11 12

rRNAs b 120

100 b ––––++++++ Zn –++++++––– OA 80 M0123123123 days

40 APOPTOTIC CELLS (%) Figure 7 Apoptosis induced by OA during enforced expression of c-myc. KmycJ cells were treated with 15 nM OA or 75 mM 20 ZnSO4 as indicated on the top of each lane. (a) Northern blot showing c-myc expression in KmycJ cells. Filter was hybridized to c-myc probe as indicated. A picture of the ®lter after transfer 0 shows the rRNAs stained with ethidium bromide. The size of the KD2 KD4 KIA KIB exogenous c-myc mRNA is smaller than the endogenous one due KD2 KD4 KIA KIB K562 Kneo2 Kneo3 Kneo6 Kneo8 to the lack of the ®rst untranslated exon in the transfected K562 Kneo2 Kneo3 Kneo6 Kneo8 construct (Delgado et al., 1995). (b) Apoptosis detected by DNA UNTREATED OA fragmentation assay. KmycJ cells were treated with 15 nM OA and/or 75 mM ZnSO as indicated on the top of the picture and 4 c the DNA was processed as indicated in Materials and methods. 120 Lane M shows DNA molecular weights markers (lambda DNA digested with HindIII)

100 protein phosphatases inhibitor, can induce apoptosis in K562 at nanomolar concentratioans (Zheng et al., 80 1994; Lerga et al., 1995). We ®rst investigated whether c-myc overexpression modi®ed the apoptosis induced by OA in KmycJ cells. The results (Figure 7a) 60 con®rmed that the induction of c-myc mRNA expression by 75 mM ZnSO4 was not impaired by 15 nM OA. The Northern blot of Figure 7a also shows 40 that the expression of endogenous c-myc mRNA was slowly down-regulated following the OA treatment, as

previously described (Lerga et al., 1995). The extent of NUCLEOSOMS IN CYTOPLASM (%) 20 apoptosis was determined by oligonucleosomal DNA fragmentation and the results (Figure 7b) showed that the ectopic expression of c-myc did not signi®cantly 0 modify the apoptosis provoked by OA. The same result K562 KD2 KD4 KIA KIB was found by scoring the fraction of morphologically Figure 8 Apoptosis induced by OA is reduced in cell sublines apoptotic cells, which was about 70% after 48 h of OA expressing c-myc mutants. (a) Internucleosomal DNA fragmenta- treatment in the absence and presence of zinc (not tion assay from control cells (lanes `C') and cells treated with shown). 15 nM OA (lanes `OA'). Cells were treated for 48 h (lanes 1 to 8) However, c-Myc could be required for OA to trigger or for 36 h (lanes 9 to 12). Lanes M show DNA molecular weights markers (lambda DNA digested with HindIII). (b) the apoptosis response in the cells. To investigate this Fraction of apoptotic cells, determined by morphological possibility we analysed the eect of c-Myc mutants on analysis after Giemsa staining. Cells were treated for 48 h with OA-mediated apoptosis of K562 cells. KD2, KD4, 15 nM OA and the fraction of apoptotic cells was determined. The KIA and KIB cells were treated with 15 nM OA for fraction of apoptotic cells in the absence of OA was below 3% for all cell lines. The maximum score (Kneo6 cells treated with OA, 48 h and the level of apoptosis was determined by 80% apoptotic cells) was set at 100% and the other data were genomic DNA fragmentation. A reduction was normalized to this value. Bars indicate standard deviations from observed in the apoptosis in the transfectants, as three (Kneo, KD2, KD4) or four (K562, KIA and KIB) separate compared to K562 cells (Figure 8a). This partial experiments. (c) Apoptosis quanti®cation by the level of inhibition of apoptosis was con®rmed by the fraction cytoplasmic nucleosomes. Cells were treated for 48 h with 15 nM OA, lysed and the amounts of DNA-histone complexes of apoptotic cells after Giemsa staining (Figure 8b) and were determined as described in Materials and methods. The cell the level of cytoplasmic nucleosomes, determined by lines tested are indicated Max and Myc in differentiation and apoptosis of K562 cells MCanÄelles et al 1322 a Kmax12 and Kmax16 were treated with 15 nM OA in

the presence and absence of 50 mM ZnSO4 for 48 h and the fragmentation of DNA was analysed. A represen- tative experiment is shown in Figure 9a. The DNA K562 KMMT Kmax16 Kmax12 K562 KMMT Kmax16 Kmax12 K562 KMMT Kmax16 Kmax12 K562 KMMT Kmax16 Kmax12 –––– ++++ – – – – +++ + Zn fragmentation induced by OA was reduced upon zinc –––– ––––++++++++OA treatment in the Kmax16 and to a larger extent in Kmax12 cells, as compared to K562 (parental cells) and KMMT (cells transfected with the empty vector). This result was con®rmed by determining the fraction of apoptotic cells after OA treatment in the presence of zinc. As shown in Figure 9b, a decrease in the percentage of apoptotic cells was observed in Kmax12 and Kmax16 cells treated with 50 mM

ZnSO4, with respect to K562 or KMMT cells. Again, b apoptosis supression was small for Kmax16 and very signi®cant for Kmax12 cells, thus reproducing the result observed in the internucleosomal DNA fragmentation assay. This result was parallel to that found for erythroid dierentiation, where Kmax12 was also the most dierentiated cell line. The fraction of apoptotic cells was already reduced in transfected cells even in the absence of zinc (Figure 9b), a result consistent with the high basal expression of Max in these cells (Figure 4). However, some further reduction in apoptosis was observed in the presence of zinc in max-transfected cells. This reduction was small, but taking together the results from the DNA fragmentation assay (Figure 9a) and the fraction of apoptotic cells (Figure 9b), we conclude that OA-mediated apoptosis decreased for Kmax12 and Kmax16 cells in the presence of zinc. In contrast, the extent of apoptosis was similar for K562 cells and KMMT cells in the absence or presence of zinc, indicating that the decrease in apoptosis in Kmax12 and Kmax16 was a consequence of the induction of Max expression.

Figure 9 Apoptosis induced by OA is reduced in max expressing cell lines. (a) Internucleosomal DNA fragmentation assay for Discussion K562, KMMT, Kmax 12 and Kmax16 cells. Cells were treated for 48 h with 50 mM ZnSO4 (Zn) and 15 nM OA as indicated. (b) Fraction of apoptotic cells, determined by morphological analysis The human myeloid leukemia K562 cell line provides a after Giemsa staining. Cells were treated for 48 h with 50 mM unique model system where the biological functions of ZnSO4 (Zn) in the presence or absence of 15 nM OA as indicated. c-Myc can be studied during dierentiation towards The fraction of apoptotic cells was determined after Giemsa two separate lineages (erythroid and myelomonocytic) staining as indicated in Materials and methods. The maximum and during drug-induced apoptosis. The eect of c- score (KMMT treated with zinc and OA, 84% apoptotic cells) was set at 100% and the other data were normalized to this value. Myc dominant negative mutants has been studied in Bars indicate standard deviations from six separate experiments dierentiation in murine preadipocytic and erythroleukemia cells (Freytag et al., 1990; Ohmori et al., 1993) and apoptosis of murine ®broblasts (Evan et al., 1992). ELISA (Figure 8c). The inhibition of apoptosis shown However, to our knowledge the eects of c-Myc by transfectant cells is only partial, as can be expected mutants and Max have not been studied in differentia- from the low expression of mutant proteins (Figure 2c). tion and apoptosis in the same cell line. We have The average apoptotic cells for K562, Kneo2, Kneo3, carried out such study on human K562 myeloid cells. Kneo6 and Kneo8 (control cell lines) is 74.3+9.4% We have found that the expression of either two (+s.d.), while the average for KD2, KD4, KIA and dierent c-Myc mutants or ectopic Max overexpression KIB is 41.0+14.3%. Statistical analysis of the data have an inhibitory eect on drug-mediated apoptosis of con®rms that the dierence between control and K562 cells, and enhance erythroid dierentiation, while mutant c-myc transfected cells is signi®cative at a myelomonocytic dierentiation was not modi®ed. 97.3% con®dence level. We also assessed by Northern Among the c-Myc mutant proteins tested, D106-143 analysis that the expression of D106-143 and In373 c- and In373 have been shown to be completely inactive myc alleles was not modi®ed by the treatment with for all the biological activities of c-Myc as transforma- 15 nM OA (not shown). tion of Rat-1a ®broblasts (Stone et al., 1987), apoptosis of Rat-1 cells (Evan et al., 1992), DNA synthesis of NIH3T3 cells (Goruppi et al., 1994), and REC Max overexpression impairs apoptosis of K562 cotransformation with Ras (Stone et al., 1987). Both We next studied whether Max overexpression modi®ed D106-143 and In373 act as dominant negative c-Myc the extent of apoptosis induced by OA in K562 cells. mutants in transformation of Rat-1 cells mediated by Max and Myc in differentiation and apoptosis of K562 cells MCanÄelles et al 1323 Bcr/Abl or v-Abl (Sawyers et al., 1992) and D106-143 1990; Sawyers et al., 1992), despite phenotypic changes in the Myc-Ras cotransformation of REC (Dang et al., described for the transfectants. To our knowledge, the 1989). Conversely to wild-type c-myc, these mutants K562 transfectants reported here and the Friend are unable to inhibit dierentiation of murine erythroleukemia cells (Bar-Ner et al., 1992) constitute erythroleukemia cells (Bar-Ner et al., 1992; Ohmori et the only examples where the constitutive expression of al., 1993) and preadipocytic cells (Freytag et al., 1990). a dominant negative c-Myc mutant is documented in The deletion of the D106-143 protein includes the Myc stably transfected cells. This is probably due to the box II (amino acids 122 ± 143), a region conserved high basal expression of normal c-myc in these cell amongst Myc family members. This box is required for lines, which allows the expression of limited amounts cell transformation and transcriptional repression of inhibitory c-Myc mutants. dependent on initiator (Inr) element of promoters, We took advantage of the K562 model to but seems to be dispensable for transactivation (Kato investigate c-Myc involvement in dierentiation and et al., 1990; Amin et al., 1993; Li et al., 1994; Brough apoptosis. We have previously shown that K562 cells et al., 1995; Lee et al., 1996). So it is likely that D106- show a high level of basal c-myc expression and this 143 inhibits Myc function by interfering with not only expression is down-regulated when the cells are transactivation but also with the transcriptional induced to dierentiate towards erythroid and repression activity of c-Myc. In contrast, the potent myelomonocytic lineages (Gomez-Casares et al., dominant negative mutant In373 carries an insertion in 1993). Therefore we asked whether the expression of the ®rst helix of the HLH domain in the C-terminal c-myc inhibitory mutants and the overexpression of part of c-Myc. We have shown that In373 is able to max interferes with dierentiation of K562 cells. We dimerize with Max in vivo, interrupts DNA binding by show in this paper that the expression of D106-143 Myc/Max heterodimers and inhibits Myc cotransform- mutant, of In373 mutant and the overexpression of ing ability in rat embryo cells (Figure 1) thus acting as Max conferred signs of erythroid dierentiation in dominant negative c-Myc protein in this assay. These human myeloid cells. The three proteins antagonize data con®rm a recent report showing that In373 the eects of c-Myc in proliferation assays. Now we inhibits Myc/Max binding to DNA in reticulocytes extend this fact to a dierentiation model system. lysates (Katzav et al., 1995). Therefore, our results Both c-Myc mutants sequester Max so as to form strongly indicate that the biochemical basis of In373 heterodimers that are expected to be biologically function is disruption of Myc/Max DNA binding. The inactive. Therefore, it is conceivable that Max levels common biochemical activity of D106-143 and In373 is are limiting in K562, and the dimerization of Max to bind Max and hence to titrate out c-Myc to form with D106-143 or In373 results in a decrease of wild- non-functional Myc/Max dimers. type Myc/Max heterodimers. The induction of Max in cells transfected with the inducible gene would result in increased levels of Max/Max dimers, competing Eects on growth and dierentiation with Myc/Max heterodimers for their molecular We generated K562 stably transfected with the D106- targets. In agreement with the results presented here, 143 and In373 mutant c-myc genes. The expression of it has been previously observed that ectopic c-myc both mutants impaired K562 growth, as expected from overexpression partially inhibits erythroid differentia- the requirement of c-Myc in cell proliferation (Heikkila tion of K562 cells (Baker et al., 1994; Delgado et al., et al., 1987; Prochownik et al., 1988; Eilers et al., 1991; 1995). In murine erythroleukemia cells the ectopic Evan et al., 1992; Nguyen et al., 1995). However, we expression of c-Myc inhibits dierentiation, while wanted to isolate stable transfectants constitutively overexpression of Max did not increase the differ- expressing the c-Myc inhibitory mutants so as to study entiation (Dunn et al., 1994). This discrepancy with their eects on dierentiation and apoptosis of K562 our results could be explained again if basal Max cells. Cell lines transfected with D106-143 contained levels relative to c-Myc were lower in K562 cells than multiple copies of the gene and expressed high levels of in murine erythroleukemia cells. c-myc mRNA, but the immunoblots detected lower We cannot rule out the possibility that the biological levels of D106-143 with respect to endogenous wild- responses described in the transfectant K562 cells are type c-Myc. The levels of In373 in the transfectants due to mechanisms independent of c-Myc. However, could not be determined but we believe they are also the common molecular consequence of the expression low as no signi®cant increase in the total c-Myc protein of D106-143, In373 and Max would be to decrease the band was detected in Western blots (not shown). This amount of active c-Myc/Max dimers bound to their result could be expected from the growth inhibitory DNA targets. The mechanism by which lowering the eect of the D106-143 and In373 c-Myc mutants. levels of c-Myc/Max would enhance erythroid differ- Therefore, only those clones expressing low amounts of entiation of K562 is unknown. One possibility is that mutant c-Myc proteins will be selected after long-term the primary eect is to impair the growth of K562 and culture. Consistently, the stable transfectants obtained the dierentiation would come as a consequence of showed a slightly reduced growth rate with respect to this. However, the fact that growth arrest by low parental cells. It has been described the presence of serum, high cell density (our unpublished observations) transcripts of inhibitory mutant c-myc genes in stable or treatment by interferons (GoÂ mez-Casares et al., transfectants (Stone et al., 1987; Mukherjee et al., 1993) do not result in dierentiation argues against the 1992) and the expression of mutant c-Myc proteins idea that growth arrest per se is sucient for K562 after transient transfection (Stone et al., 1987). erythroid dierentiation. Also, Kmax12 and Kmax16 However, it is reportedly dicult to document the cells grow at similar rates in the absence and presence presence of these mutant c-Myc proteins in stably of zinc, while erythroid dierentiation is enhanced in transfected cells (Stone et al., 1987; Freytag et al., the latter case. Max and Myc in differentiation and apoptosis of K562 cells MCanÄelles et al 1324 K562 cells are bipotential, as can dierentiate into work, Figure 7). Therefore, c-Myc could be required at erythroid or myelomonocytic lineages. Thus this the earlier stages of the process (the `condemnation' system could allow the detection of a dierential phase) (Earnshaw, 1995) and the expression of mutant role of c-Myc depending on the dierentiation c-myc genes will impair the apoptosis mediated by OA. pathway. It is noteworthy that we did not detect The mechanism by which OA induces apoptosis is any signi®cant change in the expression of myelomo- unknown, although this eect is common to dierent nocytic markers in cells transfected with c-myc cell types (Boe et al., 1991; Ishida et al., 1992; Kiguchi mutants or with inducible max after zinc addition. et al., 1994). It has been reported recently that cdc25A Also, the extent of myelomonocytic dierentiation is a target gene for c-Myc and that cdc25A can induce induced by TPA was similar for the transfectants and apoptosis in ®broblasts (Galaktionov et al., 1996). On the parental cells. Consistently with this result, we the other hand, OA can activate cdc25A through previously found that ectopic c-myc overexpression protein phosphatase 2A inhibition (reviewed in Wera does not modify TPA-mediated myelomonocytic and Hemmings, 1995). These results suggest that dierentiation of K562 (Delgado et al., 1995). To cdc25A could be one of the mediators of the apoptotic our knowledge this is the ®rst example of a eect of OA and that Max and inhibitory c-Myc dierential involvement of c-Myc in two differentia- mutants may counteract the up-regulation of cdc25A tion pathways of the same cell type. mediated by c-Myc. Further work is required to test this hypothesis. Taken together, our results support a model where a threshold level of functional c-Myc/ Eect on apoptosis Max dimers is required to maintain the cells in an Myc-mediated apoptosis has been explained by undierentiated state and to trigger an apoptotic proposing that apoptosis is a physiological function response to drug treatment. of c-Myc but normally inhibited by growth factors (Harrington et al., 1994; Packham and Cleveland, 1995). The apoptotic function of c-Myc requires dimerization with Max (Amati et al., 1993b; Bissonn- ette et al., 1994). We have found that K562 cells are Materials and methods relatively resistant to apoptosis by serum deprivation and the overexpression of c-Myc only slightly increased Bacterially synthesized proteins this apoptosis (Figure 6). This is in sharp contrast with Truncated c-Myc (tMyc; amino acids 342 ± 439) and Max the dramatic apoptotic response to c-Myc overexpres- proteins were produced as previously described (Kato et sion of ®broblasts and myeloid cells deprived from al., 1992). The truncated mutant Myc protein, tIn373, was growth factors (Askew et al., 1991; Evan et al., 1992). produced from the expression vector pDS-In373 that Consistently, expression of inhibitory c-Myc mutants encodes amino acids 343 ± 439 of c-Myc with an insert of did not decrease the apoptosis mediated by serum four serines at position 373. pDS-In373 was generated by deprivation. We conclude that the cell death induced subcloning a blunt HindIII fragment from a polymerase by growth factors deprivation in K562 is to a large chain reaction ampli®ed fragment using Sp65MycA2-In373 (Stone et al., 1987) as the template. The primers were T3 extent Myc-independent. K562 are also remarkably and 5'-ACCAGCCCAGG TCCTCGG-3' (human c-myc resistant to apoptosis induced by drugs as compared to sequence starting at codon 343). The polylinker sequence other human myeloid cell lines (Martin et al., 1990; of pDS-Myc(342 ± 439) contributes to amino acids that Ritke et al., 1994; McGahon et al., 1994; Benito et al., exceed the tIn373 polypeptide encoded by pDS-In373 by 1995). However, apoptosis is readily induced in K562 two amino acids. All proteins are tagged with a by OA (Zheng et al., 1994; Lerga et al., 1995). OA hexahistidine tail allowing for puri®cation over nickel inhibits protein serine/threonine phosphatases type 1 agarose as described previously (Kato et al., 1992). and 2A, being more active against type 2A (reviewed in Cohen et al., 1990; Wera and Hemmings, 1995). We Electrophoretic mobility shift assay (EMSA) have previously shown that OA is capable of inducing apoptosis in K562 at concentrations (15 n , used in EMSAs were performed as previously described with a M CACGTG containing radiolabeled 260 bp probe pDW14 the present work) that inhibit phosphatase type 2A but (Wechsler and Dang, 1992). This probe, in contrast to not type 1 (Lerga et al., 1995). The apoptosis mediated shorter oligonucleotides, allows resolution of the mobilities by OA in K562 is not further augmented by c-Myc of the dierent tMyc and Mac dimers. overexpression, which is in agreement with the high basal level of c-Myc expression in K562 cells. However, by all criteria analysed, the extent of OA-mediated Mammalian two-hybrid assay apoptosis was reduced in cells expressing the mutant c- The two-hybrid assay was performed as described (Dang Myc proteins D106-143 and In373 as well as in cells et al., 1991). The constructs pGal-Myc or GM(262 ± 439), where Max is overexpressed. The result suggests that a pNLVP, pVPMax(8 ± 112) and the reporter G5-E1b-CAT reduction in active c-Myc/Max complexes is respon- were described previously (Kato et al., 1990). pGal-In373 sible for the decreased apoptotic response to OA. We was constructed from the mutant In373 (Stone et al., conclude that c-Myc is involved in this apoptotic 1987) by ligating a Myc-In373 ClaI ± NsiI fragment into process triggered by protein phosphatase inhibition. the GAL4 vector pGALO as described for GM(262 ± 439) (Kato et al., 1990). Chinese hamster ovary (CHO) cells We have found that OA-mediated apoptosis is were transfected with 2 or 4 mg of activator DNA and accompanied by slow down-regulation of c-myc and 2 mg of reporter using DEAE dextran as described (Kato max, but their mRNA and protein are still detectable et al., 1990). Assays for chloramphenicol acetyltransferase after 24 h of treatment with OA, when cells are (CAT) activities were described previously (Kato et al., committed to apoptosis (Lerga et al., 1995; and this 1990). Max and Myc in differentiation and apoptosis of K562 cells MCanÄelles et al 1325 Mammalian expression c-myc and max plasmids with addition of 0.5% SDS and 0.4 mg/ml proteinase K. After an overnight incubation at 378C, DNA was extracted The plasmids pMLV-D106-143, pMLV-In373 and pMLV- by precipitation with 0.4 M NaCl and isopropanol and In6 carrying mutant c-myc genes have been previously dissolved in TE. 15 mg of DNA were treated with EcoRI, described (Stone et al., 1987). D106-143 carries a deletion separated in 0.7% agarose gel and transferred to between the 106 and 143 amino acids of c-Myc and In373 nitrocellulose membranes (Millipore) essentially as de- gene carries an insertion of four serines between amino scribed (Sambrook et al., 1989). Total RNA was isolated acids 373 and 374 of human c-Myc. In6 carries an insertion from cells by the acid guanidine thiocyanate method between amino acid 6 and 7, and is used as an inactive (Chomczynski and Sacchi, 1988). RNA samples (25 mg insertion mutant (Stone et al., 1987). pM21 encodes the per lane) were electrophoresed on 1% agarose-formalde- normal human c-myc gene. All these genes are under the hyde gels and transferred to nitrocellulose membranes control of the Moloney murine leukemia virus LTR (Stone (Millipore) as described (Sambrook et al., 1989) except that et al., 1987). The vector pHEBoMT (kindly provided by F 1 mg of ethidium bromide was added to each RNA sample Grignani) carries the mouse metallothionein I promoter prior to gel loading. A picture of the ®lter under u.v. light and the hygromycin resistance gene (Grignani et al., 1990). was obtained after transfer to assess the amount and Plasmid pHEBoMT-max was constructed by ligation of the integrity of the rRNAs. The Southern and Northern blots 1.9 kb EcoRI insert of plasmid pBSmax (kindly provided were hybridized at 428C in 40% formamide, 5% dextran by R Dalla-Favera) into the BamHI site of the pHEBoMT sulphate, 0.8 M NaCl, 50 mM sodium phosphate (pH 7), plasmid after blunt-ending the EcoRI and BamHI sites 0.2% SDS, 150 mg of denatured salmon sperm DNA per ml with the Klenow fragment. pHEBoMT-max encodes the and DNA probe labeled with [32P]dCTP by random primed human p21 Max under the control of the metallothionein labeling (Pharmacia kit). The ®lters were washed to a ®nal promoter. stringency of 0.56SSC and 0.1% SDS at 658Cand autoradiographed at 7708C. The signals were quanti®ed Rat embryo cell cotransformation assay with a Bio-Rad Molecular Imager apparatus. Probes for human c-myc, max and e-globin genes were as described The myc ± ras cotransformation assays were performed as (Delgado et al., 1995). described previously (Dang et al., 1989). Transformed foci were determined from quadruplicate experiments at 3 weeks after lipofection of rat embryo ®broblasts. Immunoblots Cell pellets were lysed in a solution containing K562 cell lines and transfection 100 mM Tris (pH 6,8), 8% b-mercaptoethanol, 4% SDS and 20% glycerine. Protein content was measured using K562 cells were obtained from ATCC. KmycJ are K562 Bio-Rad Protein Assay. 40 mg of protein per lane were cells stably transfected with a zinc-inducible c-myc gene, separated in 10% acrylamide gels and transferred to PVDF under the control of the mouse metallothionein I promoter membranes (Immobilon, Millipore), using a semi-dry (Delgado et al., 1995). KMMT are K562 cells stably electroblotter (Millipore) and a buer containing 10 mM transfected with the vector pHEBoMT (Delgado et al., Trizma, 96 mM glycine and 10% methanol. c-Myc and 1995). Cells were grown in RPMI-1640 medium (Whit- Max proteins were detected by the anti-Myc monoclonal taker) supplemented with 8% fetal calf serum (Biochrom) antibody 9E10 (Oncogene Science) or the anti-Max and gentamicin (80 mg/ml). KmycJ and KMMT were polyclonal antibody (UBI). Membranes were then devel- grown in the presence of 0.4 mg/ml of G418. 50 mgof oped using chemiluminescent detection (Tropix). pMLV-In373 and pMLV-D106-143 plasmids (Stone et al., 1987) were transfected along with 5 mg of pSV2neo into K562 cells by electroporation (400 v, 500 mF) using a Bio- Apoptosis determinations Rad Gene Pulser apparatus. After electroporation the cells For analysis of internucleosomal DNA fragmentation were incubated for 48 h and 0.6 mg/ml of G418 was added. (Duke and Cohen, 1986), cells were lysed in a solution 40 mg of pHEBoMT-max plasmid were also transfected by containing 10 mM Tris, 1 mM EDTA and 0.2% Triton X- electroporation and selected in 0.2 mg/ml of hygromycin B. 100. Cytoplasmic fraction of the lysates was obtained by Clones were obtained from the G418-resistant and centrifugation and further adjusted to 150 mM NaCl, hygromycin-resistant pools by limiting dilution in micro- 40 mM EDTA, 1% SDS and treated with 200 mg/ml of titer wells and were expanded and analysed. To be used as proteinase K. DNA fragments were precipitated with controls, K562 cells were also transfected with a plasmid ethanol and separated in 1.5% agarose gels containing expressing the G418 resistance gene (pMAMneo, Clontech) 0.1 mg/ml ethidium bromide. Quanti®cation of apoptosis and selected as indicated above. by determination of cytoplasmic nucleosomes was carried out by the Cell Death Detection ELISA assay (Boehringer Assessment of cell growth rate and dierentiation Mannheim). Cytospin preparations were stained by Giemsa and the fraction of apoptotic cells was scored under the Cell growth and viability were assayed by hemocytometer microscope. Apoptotic cells were identi®ed by chromatin and the trypan blue exclusion test. Erythroid differentia- condensation and marginalization, nuclear fragmentation tion was induced by adding to growing cells 1 mM ara-C. and cytoplasmic shrinkage. At least 200 cells were analysed The fraction of hemoglobin-producing cells was scored by for each point by observers without knowledge of the the benzidine assay essentially as described (Rowley et al., sample analysed. 1981). Myelomonocytic dierentiation was induced by treating the cells with 10 nM TPA. The dierentiation was determined by cell clustering activity, the ability to reduce nitroblue tetrazolium and the expression of Acknowledgements glycoprotein IIb-IIIa, determined by ¯ow cytometry as We thank Pilar Frade for tissue culture and photographic described (Delgado et al., 1992). work, Robert Eisenman and Steven Collins for critical reading of the manuscript and helpful advice, Franco Grignani, Riccardo Dalla-Favera, Robert Eisenman and DNA and RNA analysis Waturu Shoji for clones, Narciso Benitez for statistical Cellular DNA was isolated by lysing cells in a solution analysis and illustrations design and M Luisa GarcõÂ afor containing 10 mM Tris (pH 8), 150 mM NaCl, 3 mM EDTA technical assistance. This work was supported by grants Max and Myc in differentiation and apoptosis of K562 cells MCanÄelles et al 1326 DGICYT PB92-506CO-2 and CICYT SAF96-0083 from fellowship from Gobierno Vasco, KMH by NIH training Spanish Government, a grant from FundacioÂ nRamoÂn grant T32GM07814 and AL by a FPI fellowship from Areces and NIH grant CA51497. MC was supported by a Ministerio de EducacioÂ n y Ciencia, Spain.

References

Amati B and Land H. (1994). Curr. Op. Genet. Develop., 4, Goruppi S, Gustincich S, Brancolini C, Lee WMF and 102 ± 108. Schneider C. (1994). Oncogene, 9, 1537 ± 1544. Amati B, Brooks MW, Levy N, Littlewood TD, Evan GI and Grignani F, Lombardi L, Inghirami G, Sternas L, Cechova Land H. (1993a). Cell, 72, 233 ± 245. K and Dalla-Favera R. (1990). EMBO J., 9, 3913 ± 3922. Amati B, Littlewood TD, Evan GI and Land H. (1993b). Gu W, Cechova K, Tassi V and Dalla-Favera R. (1993). EMBO J., 12, 5083 ± 5087. Proc.Natl.Acad.Sci.USA,90, 2935 ± 2939. Amin C, Wagner AJ and Hay N. (1993). Mol. Cell. Biol., 13, Harrington EA, Fanidi A and Evan GI. (1994). Curr. Op. 383 ± 390. Genet. Develop., 4, 120 ± 129. Askew DS, Ashmun RA, Simmons BC and Cleveland JL. Heikkila R, Schwab G, Wickstrom E, Loke SL, Pluznik DH, (1991). Oncogene, 6, 1915 ± 1922. Watt R and Neckers LM. (1987). Nature, 328, 445 ± 449. Baker SJ, Pawlita M, Leutz A and Hoelzer D. (1994). Holt JT, Redner RL and Nienhuis AW. (1988). Mol. Cell. Leukemia, 8, 1309 ± 1317. Biol., 8, 963 ± 973. Bar-Ner M, Messing LT, Cultraro CM, Birrer MJ and Segal Ishida Y, Furukawa Y, Decaprio JA, Saito M and Grin S. (1992). Cell Growth Di., 3, 183 ± 190. JD. (1992). J. Cell. Physiol., 150, 484 ± 492. Benito A, Grillot D, NunÄ ez G and FernaÂ ndez-Luna JL. KamanoH,OnnishiH,TanakaT,IkedaK,OkabeAand (1995). Am. J. Pathol., 146, 481 ± 490. Irino S. (1990). Leuk. Res., 14, 831 ± 839. Bissonnette RP, McGahon A, Mahboubi A and Green DR. Kato GJ, Barret J, Villa-GarcõÂ a M and Dang CV. (1990). (1994). J. Exp. Med., 180, 2413 ± 2418. Mol. Cell. Biol., 10, 5914 ± 5920. Blackwood EM and Eisenman RN. (1991). Science, 251, Kato GJ, Lee WMF, Chen L and Dang CV. (1992). Genes 1211 ± 1217. Develop., 6, 81 ± 92. Blackwood EM, Kretzner L and Eisenman RN. (1992). Curr. Kato GJ and Dang CV. (1992). FASEB J., 6, 3065 ± 3072. Op. Genet. Develop., 2, 227 ± 235. Katzav S, Pakham G, Sutherland M, Aroca P, Santos E and Boe R, Gjertsen BT, Vintermyr OK, Houge G, Lanotte M Cleveland JL. (1995). Oncogene, 11, 1079 ± 1088. and Doskeland SO. (1991). Exp. Cell Res., 195, 237 ± 246. KiguchiK,GlesneD,ChubbCH,FujikiHandHuberman Brough DE, Hofmann TJ, Ellwood KB, Townley RA and E. (1994). Cell Growth Dier., 5, 995 ± 1004. Cole MD. (1995). Mol. Cell. Biol., 15, 1536 ± 1544. Koskinen PJ, VaÈ strik I, MaÈ kelaÈ TP, Eisenman RN and Charnay P and Maniatis T. (1983). Science, 220, 1281 ± 1283. Alitalo K. (1994). Cell Growth Dier., 5, 313 ± 320. Chisholm O, Stapleton P and Symonds G. (1992). Oncogene, Kretzner L, Blackwood EM and Eisenman RN. (1992). 7, 1827 ± 1836. Nature, 359, 426 ± 429. Chomczynski P and Sacchi N. (1988). Anal Biochem., 162, Larsson LG, Ivhed I, Gidlund M, Pettersson U, VennstroÈ m 156 ± 159. B and Nilsson K. (1988). Proc. Natl. Acad. Sci. USA, 85, Cohen P, Holmes CFB and Tsukitani Y. (1990). Trends 2638 ± 2642. Biochem. Sci., 15, 98 ± 102. Lee LL, Dolde C, Barrett J, Wu CS and Chi D. (1996). J. Coppola JA and Cole MD. (1986). Nature, 320, 760 ± 763. Clin. Invest., 97, 1687 ± 1695. Dang CV, Barrett J, Villa-GarcõÂ a M, Resar LMS, Kato GJ Lerga A, Belandia B, Delgado MD, Cuadrado MA, Richard and Fearon ER. (1991). Mol. Cell. Biol., 11, 954 ± 962. C, Ortiz JM, MartõÂ n-Perez J and LeoÂ n J. (1995). Biochem. Dang CV, McGuire M, Buckmire M and Lee WMF. (1989). Biophys. Res. Commun., 215, 889 ± 895. Nature, 337, 664 ± 666. Li L, Nerlov C, Prendergast G, MacGregor D and Zi EB. Delgado MD, Quincoces AF, GoÂ mez-Casares MT, MartõÂ nez (1994). EMBO J., 13, 4070 ± 4079. CA, Cuadrado MA, Richard C and LeoÂ n J. (1992). Cancer Lindeman GJ, Harris AW, Bath ML, Eisenman RN and Res., 52, 5948 ± 5979. Adams JM. (1995). Oncogene, 10, 1013 ± 1017. Delgado MD, Lerga A, CanÄ elles M, GoÂ mez-Casares MT and Lotem J and Sachs L. (1993). Cell Growth Dier., 4, 41 ± 47. LeoÂ n J. (1995). Oncogene, 10, 1659 ± 1665. MaÈ kelaÈ TP, Koskinen PJ, VaÈ strik I and Alitalo K. (1992). Dmitrovski E, Kuelk WM, Hollis GF, Kirsch IR, Bender TP Science, 256, 373 ± 377. and Segal S. (1986). Nature, 322, 748 ± 750. McGahon A, Bissonnette RP, Schmitt M, Cotter KM, Green Duke RC and Cohen JJ. (1986). Lymphokine Res., 5, 289 ± DR and Cotter TG. (1994). Blood, 83, 1179 ± 1187. 299. Marcu KB, Bossone SA and Patel AJ. (1992). Annu. Rev. Dunn BK, Cogliati T, Cultraro CM, Bar-Ner M and Segal S. Biochem., 61, 809 ± 860. (1994). Cell Growth Dier., 5, 847 ± 854. Martin S, Lennon S, Bonham A and Cotter T. (1990). J. Earnshaw WC. (1995). Curr. Op. Cell Biol., 7, 337 ± 343. Immunol., 145, 1859 ± 1867. Eilers M, Schirm S and Bishop JM. (1991). EMBO J., 10, Meichle A, Philipp A and Eilers M. (1992). Biochim. Biophys. 133 ± 141. Acta., 1114, 129 ± 146. Evan GI, Wyllie AH, Gilbert CS, Littlewood TD, Land H, Mukherjee B, Morgenbesser SD and DePinho R. (1992). Brooks M, Waters CM, Penn LZ and Hancock DC. Genes Dev., 6, 1480 ± 1492. (1992). Cell, 69, 119 ± 128. Nguyen HQ, Selvakumaran M, Liebermann DA and Evan GI and Littlewood TD. (1993). Curr. Op. Genet. Homan B. (1995). Oncogene, 11, 2439 ± 2444. Develop., 3, 44 ± 49. Ohmori Y, Jun-ichi T, Takada S, Lee WMF and Obinata M. Freytag SO, Dang CV and Lee WMF. (1990). Cell Growth (1993). Oncogene, 8, 379 ± 386. Dier., 1, 339 ± 343. Packham G and Cleveland JL. (1995). Biochim. Biophys. Galaktionov K, Chen X and Beach D. (1996). Nature, 382, Acta., 1242, 11 ± 28. 511 ± 517. Prendergast GC, Lawe D and Zi EB. (1991). Cell, 65, 395 ± GoÂ mez-Casares MT, Delgado MD, Lerga A, Crespo P, 407. Quincoces AF, Richard C and LeoÂ n J. (1993). Leukemia, Prochownik EV, Kukowska J and Rodgers C. (1988). Mol. 7, 1824 ± 1833. Cell. Biol., 8, 3683 ± 3695. Max and Myc in differentiation and apoptosis of K562 cells MCanÄelles et al 1327 Ritke MK, Bergoltz VV, Allan WP and Yalowich JC. (1994). Stone J, De Lange T, Ramsay G, Jakobovits E, Bishop JM, Biochem. Pharmacol., 48, 525 ± 533. Varmus H and Lee W. (1987). Mol. Cell. Biol., 7, 1697 ± Rowley PT, Ohlsson-Wilhelm BM, Farley BA and LaBella S. 1709. (1981). Exp. Hematol., 9, 32 ± 37. VaÈ strik I, MaÈ kelaÈ TM, Koskinen PJ, Klefstrom J and Alitalo Sambrook J, Fritsch EF and Maniatis T. (1989). Molecular K. (1994). Crit. Rev. Oncogenesis, 5, 59 ± 68. Cloning: A Laboratory Manual, 2nd edition. Cold Spring Wechsler DS and Dang CV. (1992). Proc. Natl. Acad. Sci. Harbor Laboratory. Cold Spring Harbor, New York. USA, 89, 7635 ± 7639. Sawyers CL, Callahan W and Witte ON. (1992). Cell, 70, Wera S and Hemmings BA. (1995). Biochem. J., 311, 17 ± 29. 901 ± 910. Zheng B, Chambers TC, Raynor RL, Markham PN, Gebel Selvakumaran M, Lin HK, Sjin RTT, Reed JC, Liebermann HM, Vogler WR and Kuo JF. (1994). J. Biol. Chem., 269, DA and Homan B. (1994). Mol. Cell. Biol., 14, 2352 ± 12332 ± 12338. 2360. Shen W-F, Detmer K, Mayhews CHE, Hack FM, Morgan DA, Largman C and Lawrence HJ. (1992). EMBO J., 11, 983 ± 989.