Differentiation lineage-specific expression of human

 
CONTINUE READING
Differentiation lineage-specific expression of human
  heat shock transcription factor 2
           LILA PIRKKALA,*,† TERO-PEKKA ALASTALO,*,‡ PÄIVI NYKÄNEN,*,§ LAURA SEPPÄ,*,§
           AND LEA SISTONEN*,1
           *Turku Centre for Biotechnology, †Department of Biology, ‡Department of Anatomy, and
           §
             Department of Biochemistry and Pharmacy, Åbo Akademi University, University of Turku,
           FIN-20521 Turku, Finland

ABSTRACT         Differentiation of multipotential hema-       elements of differentiation lineage-specific gene loci
topoietic cells into lineage-committed precursors in-          (for review, see refs 1, 2). Differentiation lineage-
volves the selection and maintenance of appropriate            restricted transcription factors are thus good candi-
programs of gene expression, regulated by specific             dates for mediating the various signals in immature
transcription factors. Using human K562 erythroleuke-          cells to establish or maintain differentiation lineage-
mia cells capable of differentiating along erythroid and       specific gene expression. Therefore, it is essential to
megakaryocytic lineages, we explore the differentia-           study the expression, complex formation, and chroma-
tion-related role of heat shock transcription factor 2         tin accessibility of these specific transcription factors to
(HSF2), which belongs to a family of transcription             understand the mechanisms underlying the complex
factors generally known to regulate heat shock gene            processes of cell differentiation.
expression. We demonstrate that enhanced HSF2 ex-                 The human K562 erythroleukemia cell line is a
pression and the acquisition of HSF2 DNA binding               multipotent hematopoietic precursor cell line derived
activity are strictly specific for erythroid characteristics   from a patient with chronic myeloid leukemia in blast
of K562 cells. Our results reveal a multistep regulatory       crisis (3). K562 cells can be induced to differentiate
process of HSF2 gene expression. In K562 cells under-          along several lineages, thus providing a model system
going hemin-mediated erythroid differentiation, the            to study gene expression during hematopoiesis. The
increase in HSF2 protein levels is preceded by tran-           K562 cell line has been used extensively in studies of
scriptional induction of the HSF2 gene, accompanied            hemin-induced erythroid differentiation (4 –6). Treat-
by increased HSF2 mRNA stability. In contrast, during          ment of K562 cells with hemin, the synthetic form of
megakaryocytic differentiation induced by the phorbol          heme, induces the synthesis of red cell-specific pro-
ester TPA, expression of HSF2 is rapidly down-regu-            teins, such as globin polypeptides, as a result of tran-
lated, leading to a complete loss of the HSF2 protein.         scriptional activation of the embryonic a- and b-like
These results indicate that the determination of HSF2          globin genes, z and e, respectively, as well as the fetal
expression occurs at the early stages of lineage com-          g-globin and adult a-globin genes (6). Hemin treat-
mitment. Taken together, our data suggest that HSF2            ment does not, however, lead to terminal differentia-
could function as a lineage-restricted transcription fac-      tion of K562 cells, thereby dissociating increases in
tor during differentiation of K562 cells along either the      intracellular hemoglobin content from other events
erythroid or the megakaryocytic pathway.—Pirkkala,             considered central to erythroid differentiation (5).
L., Alastalo, T.-P., Nykänen, P., Seppä, L., Sistonen, L.    On the other hand, treatment of K562 cells with the
Differentiation lineage-specific expression of human           tumor promoter 12-O-tetradecanoyl-phorbol 13-ace-
heat shock transcription factor 2. FASEB J. 13,                tate (TPA)2 shifts these cells toward megakaryocytic
1089 –1098 (1999)

Key Words: HSF2 z K562 cells z erythroid differentiation
 z megakaryocytic differentiation
                                                                 1
                                                                   Correspondence: Turku Centre for Biotechnology, Bio-
                                                               City, 5th Floor, Tykistökatu 6, FIN-20521 Turku, Finland.
Differentiation of lineage-committed precursors                E-mail: lea.sistonen@btk.utu.fi
                                                                 2
from multipotential hematopoietic cells involves the               Abbreviations: DMEM, Dulbeccos’s modified Eagle’s me-
selection and maintenance of appropriate programs of           dium; FCS, fetal calf serum; HSE, heat shock element; HSF,
gene expression. These programs presumably arise               heat shock factors; Hsp, heat shock protein; K562, erythro-
                                                               leukemia; Molt-4, T-lymphoblastic leukemia; PDGF, platelet-
from changes in the functional balance of cellular             derived growth factor; SDS-PAGE, sodium dodecyl sulfate-
transcription factors, resulting in the formation of           polyacrylamide gel electrophoresis; TPA, 12-O-tetradecanoyl-
stable active transcription complexes at the regulatory        phorbol 13-acetate; TRX, thioredoxin.

0892-6638/99/0013-1089/$02.25 © FASEB                                                                                 1089
pathway of differentiation, leading to loss of their        ity, hemin-induced transcription of heat shock
erythroid properties and to acquisition of several          genes, and erythroid differentiation of K562 cells
megakaryoblastoid characteristics, including synthe-        (27). In this study, we have analyzed the expression
sis and secretion of platelet-derived growth factor         of human HSF2 in K562 cells induced to differenti-
(PDGF) polypeptides as well as synthesis and surface        ate by hemin or TPA along the erythroid or the
expression of glycoprotein IIIa (for review, see ref 7).    megakaryocytic lineage, respectively. Our results re-
TPA is known to exert its effects on various cellular       veal that the expression of HSF2 is strictly and
processes, such as growth and differentiation,              specifically regulated in a lineage-restricted manner,
through activation of protein kinase C (8). For             i.e., hemin enhances and TPA down-regulates HSF2
example, TPA-induced megakaryocytic differentia-            expression in K562 cells, suggesting that HSF2 might
tion of the human HEL erythroleukemia cells has             be an important transcriptional regulator involved in
been shown to be mediated by protein kinase C (9).          erythroid differentiation.
   Transcriptional activation of heat shock genes, which
ultimately leads to increased synthesis of heat shock
proteins (Hsp’s), is regulated by a family of transcrip-    MATERIALS AND METHODS
tion factors called heat shock factors (HSFs) that          Cell culture and experimental treatments
respond to external stimuli, such as elevated tempera-
tures and diverse physiological and environmental           K562 (erythroleukemia) and Molt-4 (T-lymphoblastic leuke-
stressors (for review, see refs 10, 11). In yeast and       mia) cells were maintained in RPMI 1640 medium supple-
Drosophila, only one HSF-encoding gene has been             mented with 10% fetal calf serum (FCS) and antibiotics
                                                            (penicillin and streptomycin) in a humidified 5% CO2 atmo-
identified (12, 13), whereas in vertebrates several mem-    sphere at 37°C. Raji (Burkitt’s lymphoma) cells were main-
bers of the HSF family (HSF1–4) have been cloned            tained in Dulbecco’s modified Eagle’s medium (DMEM)
(14 –18). In mammalian cells, HSF1 mediates the ubiq-       containing 10% FCS and antibiotics. HeLa (cervical carci-
uitous response to stress stimuli, whereas HSF2 is          noma) cells were grown as a monolayer in DMEM containing
regulated by distinct signaling mechanisms. HSF2 is         5% FCS and antibiotics. K562 cells stably overexpressing
                                                            HSF2-a and HSF2-b isoforms (2a-C7 and 2b-D5, respectively;
abundantly expressed and constitutively active in           ref 27) were maintained in RPMI 1640 medium containing
mouse embryonal carcinoma cells, at the blastocyst          G418 (500 mg/ml; Life Technologies, Inc., Paisley, U.K.). For
stage during mouse embryogenesis, and during sper-          experimental treatments, cells were seeded at 5 3 106 cells
matogenesis, suggesting a role for HSF2 as a develop-       per 10 cm-diameter plate (HSF2-a- and HSF2-b-overexpress-
                                                            ing cells were plated in RPMI 1640 medium without G418).
mental regulator (19–23). In addition, HSF2 binds to a      Hemin (Aldrich, Milwaukee, Wis.) was added to a final
specific DNA binding sequence (heat shock element,          concentration of 30 mM, TPA (Sigma, St. Louis, Mo.) to 10
HSE) in the hsp70 gene promoter, leading to abun-           nM, and actinomycin D (Sigma) to 6.4 mg/ml; cells were
dant expression of Hsp70 protein during hemin-medi-         incubated at 37°C for the time periods indicated. Heat shock
ated erythroid differentiation of K562 cells (24 –27).      was performed at 42°C in a waterbath.
During embryogenesis, however, the pattern of HSF2          Gel mobility shift analysis
DNA binding activity does not coincide with the expres-
sion profile of any of the known Hsp’s (22). The            Whole-cell extracts were prepared from experimentally
expression of thioredoxin (TRX) is induced in K562          treated cells, as described previously (33), and incubated (12
cells in response to hemin in an HSF2-dependent             mg protein) with a 32P-labeled oligonucleotide representing
                                                            the proximal HSE of the human hsp70 promoter. The
manner (28), providing evidence that HSF2 might             protein–DNA complexes were analyzed on a native 4% poly-
regulate genes other than the known heat shock genes.       acrylamide gel as described previously (33). The signal inten-
Recently, the roles of distinct HSFs have been proposed     sities of the protein–DNA complexes were quantitated using a
to overlap depending on stimulatory signals. For exam-      phosphorimaging scanner (Bio-Rad, Hercules, Calif.). For
ple, HSF2 activation and consequent transcriptional         antibody supershift experiments to analyze HSF1 and HSF2
                                                            composition in the HSE binding complex in K562 cells
induction of heat shock genes have been indicated in        treated with 30 mM hemin for 18 h or 10 nM TPA for 24 h,
cells where the ubiquitin-proteasome pathway is inhib-      dilutions (1:10, 1:50, and 1:100) of antisera against mouse
ited (29). Moreover, mutated yeast cells carrying a         HSF1 and mouse HSF2 (amHSF1 and amHSF2, respectively;
lethal HSF deletion can be rescued by human HSF2,           a kind gift from Dr. Richard Morimoto; ref 34) were added to
                                                            whole-cell extracts and incubated at 25°C for 15 min prior to
but not by HSF1 (30).
                                                            gel mobility shift analysis. For the competition experiment,
   More complexity to the regulatory functions of           the binding reaction mixture contained 0.1 ng of the labeled
HSF2 is added by the finding that HSF2 exists as two        HSE oligonucleotide and a 50-, 100-, or 200-fold molar excess
alternatively spliced isoforms, HSF2-a and HSF2-b           of the unlabeled HSE oligonucleotide or a 100-fold molar
(31, 32). According to our recent results, HSF2-a is        excess of an unspecific oligonucleotide.
the predominantly expressed isoform in K562 cells
                                                            SDS-PAGE and Western blot analysis
(27). A molar excess of HSF2-a is required for the
hemin-mediated activation of HSF2 since overex-             Whole-cell extracts (12 mg protein) were subjected to 8%
pression of the HSF2-b isoform inhibits HSF2 activ-         sodium dodecyl sulfate-polyacrylamide gel electrophoresis

1090   Vol. 13   June 1999                        The FASEB Journal                                      PIRKKALA ET AL.
(SDS-PAGE) and transferred to nitrocellulose filter (Protran      K562 cells toward the megakaryocytic pathway. First,
Nitrocellulose; Schleicher & Schuell, Keene, N.H.) using a        we confirmed the ability of K562 cells to differentiate
Bio-Rad semidry transfer apparatus. HSF2 was detected by a
polyclonal antibody specific to mouse HSF2 (34), the induc-       along the erythroid and megakaryocytic lineages by
ible form of Hsp70 by 4g4 (Affinity Bioreagents, Inc., Nes-       hemin and TPA by analyzing the expression of
hanic Station, N.J.), the constitutively expressed Hsc70 by       specific markers, i.e., the fetal g-globin and PDGF
SPA-815 (StressGen, Victoria, B.C., Canada), and fetal g-glo-     polypeptide, respectively. In accordance with earlier
bin by PBF-R (Isolab, Akron, Ohio). Horseradish peroxidase-
conjugated secondary antibodies were purchased from Pro-
                                                                  studies (5, 40), accumulation of the fetal g-globin
mega (Madison, Wis.) and Amersham (Little Chalfont, U.K.).        polypeptide was markedly increased after 48 h of
The blots were developed with an enhanced chemilumines-           hemin treatment as compared with untreated cells,
cence method (Amersham).                                          whereas during the corresponding treatment with
                                                                  TPA, the g-globin polypeptide could not be detected
Northern blot analysis                                            (Fig. 1A, left panel). Abundant expression of
                                                                  PDGF-B mRNA was observed by 24 – 48 h of TPA
Poly(A) mRNA was isolated from the treated cells using a
poly(A) mRNA purification kit (Pharmacia, Piscataway, N.J.).      incubation, but not on hemin treatment (Fig. 1A,
RNA was separated on a 1% agarose-formaldehyde gel, trans-        right panel).
ferred to nylon filter (Hybond-N; Amersham), and hybridized          As shown in the left panel of Fig. 1B, HSF–HSE
at 65°C with a [a-32P]dCTP (50 mCi, 3000 Ci/mmol; ICN,            complex formation was induced by 16 h treatment
Irvine, Calif.) -labeled 931 bp HindIII/PstI cDNA insert
                                                                  with hemin, but not by TPA. Consistent with our
coding for human HSF2 (hHSF2 cDNA was a kind gift from
Dr. Robert Kingston; ref 16), a 500 bp cDNA insert coding for     earlier results, HSF2 was primarily activated in he-
human TRX (ref 28), a 1510 bp cDNA insert coding for              min-treated K562 cells, as shown by antibody super-
human PDGF-B (a kind gift from Dr. Kari Alitalo), and             shift assay using specific antisera against mouse HSF1
[a-32P]dCTP-labeled plasmids for the following genes: hu-         and HSF2 (Fig. 1B, middle panel; ref 26). TPA did
man hsp70 (pH2.3; ref 35), rat GAPDH (pGAPDH; ref 36),
and human b-actin (pHFbA-1; ref 37). After hybridization,         not abolish the hemin-induced HSF2 DNA binding
filters were washed with high stringency conditions (0.1X         activity (he1TPA; Fig. 1B). Likewise, HSF2 DNA
SSC-0.1% SDS at 65°C; 1X SSC is 0.15 M sodium chloride and        binding could not be activated by hemin after a TPA
0.015 M sodium citrate), and visualized by autoradiography.       pretreatment (TPA1he; Fig. 1B, left panel), suggest-
The intensities of radioactive signals were quantitated using a   ing that commitment of K562 cells to an erythroid
computerized image analysis (Microcomputer Imaging De-
vice version M4, Imaging Research, Inc.) or a phosphorimag-       differentiation pathway requires activation of HSF2.
ing scanner (Bio-Rad).                                            Furthermore, HSF2 activation appears to be irrevers-
                                                                  ible, as the already activated HSF2 could not be
Nuclear run-on analysis                                           inactivated by TPA. After several hours of exposure
                                                                  to TPA, a faster migrating, smaller molecular weight
Nuclear run-on transcription reactions were performed with        DNA binding complex of unknown origin was ob-
nuclei isolated from hemin- or heat shock-treated cells in the
presence of 100 mCi of [a-32P]dUTP (3000 Ci/mmol; Amer-           served (asterisks in Fig. 1B). It appears, however, that
sham) as described previously (38). Radiolabeled RNA was          this DNA binding complex did not contain HSF2,
hybridized to nitrocellulose immobilized 931 bp HindIII/PstI      since neither the specific antisera against mouse
cDNA insert coding for human HSF2 (16) and plasmids for           HSF1 or HSF2 nor an excess of nonradiolabeled
the following genes: human hsp70 (pH2.3; ref 35), human           HSE oligonucleotide displaced the corresponding
hsp90/89a (pUCHS801; ref 39), human b-actin (pHFbA-1;
ref 37), and a Bluescript vector (Stratagene, San Diego,          band (Fig. 1B, right panel).
Calif.). The hybridizations were carried out in 50% form-            To analyze the effects of hemin and TPA on HSF2
amide-6X SSC-10X Denhardt’s-0.2% SDS at 42°C for 72 h.            expression, the levels of HSF2 protein were analyzed
Filters were washed with high stringency conditions (0.2X         upon exposure to hemin or TPA or to combined
SSC-0.2% SDS at 65°C) and visualized by autoradiography.
The intensities of radioactive signals were quantitated using a
                                                                  hemin and TPA treatment. In contrast to the in-
phosphorimaging scanner (Bio-Rad).                                creased accumulation of HSF2 protein in K562 cells
                                                                  exposed to hemin for 16 h, treatment with TPA
                                                                  resulted in a dramatic reduction of HSF2 protein by
RESULTS                                                           4 h (41) and a complete loss by 16 –24 h (Fig. 1C).
                                                                  However, a similar decrease in the levels of HSF2
Activation and expression of HSF2 are specific for                protein was not detected in cells exposed to TPA for
the erythroid properties of K562 cells                            1– 6 h after a 16 h pretreatment with hemin
                                                                  (he1TPA; Fig. 1C). When K562 cells were preincu-
HSF2 DNA binding is induced in K562 cells commit-                 bated with TPA for 1 to 4 h prior to addition of
ted to differentiate along the erythroid lineage by               hemin for 16 h, the TPA-induced loss of HSF2
hemin (26). The differentiation lineage specificity of            protein could not be reversed with hemin (TPA1he;
HSF2 expression has, however, remained unknown.                   Fig. 1C). Together with previous studies (26, 27), this
Therefore, we examined the activation and expres-                 provides evidence that the HSE binding activity
sion of HSF2 during TPA-induced differentiation of                induced by hemin (Fig. 1B) requires presence of

DIFFERENTIAL EXPRESSION OF HSF2                                                                                      1091
Figure 1. Inhibition of HSF2 activa-
tion during TPA-mediated megakaryo-
cytic differentiation of K562 cells. A)
Analysis of erythroid and megakaryo-
cytic markers of K562 cells treated with
hemin and TPA, respectively. Whole-
cell extracts (12 mg) isolated from
control (C), hemin-treated (he; 30 mM
for 48 and 96 h), and TPA-treated
(TPA; 10 nM for 48 and 96 h) K562 cells were analyzed on a 12% SDS-PAGE and immunoblotted using antibodies against fetal
g-globin and Hsc70 (left panel). Right panel: after treatment of K562 cells with hemin (he; 30 mM for 48 h) or TPA (10 nM for 3,
8, 24, and 48 h), poly(A) mRNA was isolated and analyzed by Northern blotting using 32P-labeled cDNA probes for PDGF-B and
GAPDH. GAPDH was used as a control for equal loading of samples. The mRNA sizes are indicated on the right. B) Analysis of HSF2
DNA binding activity. Left panel: whole-cell extracts from control (C), hemin-treated (he; 30 mM), TPA-treated (TPA; 10 nM) K562
cells, and K562 cells subjected to combined hemin and TPA treatment were analyzed by gel mobility shift assay. Extracts (12 mg) were
incubated with a 32P-labeled oligonucleotide representing the proximal HSE of the human hsp70 promoter. Protein–DNA complexes
were resolved on a 4% nondenaturing polyacrylamide gel. C and he indicate untreated and hemin-treated (16 h) cells, respectively.
Numbers above the lanes indicate the duration of TPA treatment either after (he1TPA) or before (TPA1he) a 16 h hemin
treatment. HSF indicates the specific inducible HSF2–HSE complex, CHBA indicates the constitutive HSE binding activity reported
previously (33), NS denotes nonspecific protein–DNA interaction, and Free indicates free probe. Asterisk marks an unknown, faster
migrating DNA binding complex. Middle panel: extracts (12 mg) from K562 cells treated with 30 mM hemin for 18 h were incubated
in the presence of 1:10, 1:50, and 1:100 dilutions of either the antiserum specific to mouse HSF2 (amHSF2) or mouse HSF1
(amHSF1), as indicated, prior to the gel mobility shift assay. Right panel: extracts (12 mg) from K562 cells treated with 10 nM TPA
for 24 h were incubated in the presence of the antiserum specific to mouse HSF2 (amHSF2) or mouse HSF1 (a1) or in the presence
of a 50-, 100-, or 200-fold molar excess of the unlabeled HSE oligonucleotide or a 100-fold molar excess of an unspecific
oligonucleotide (non-self) prior to the gel mobility shift assay. C) Prolonged TPA treatment leads to loss of HSF2 protein. The same
samples (12 mg) used in the gel mobility shift assay in panel B were analyzed on an 8% SDS-PAGE and immunoblotted using
antibodies against HSF2 and Hsc70. The double band in the HSF2 blot corresponds to the alternatively spliced human HSF2 isoforms.
The slower migrating band indicates HSF2-a, and the faster migrating band HSF2-b.

1092   Vol. 13   June 1999                               The FASEB Journal                                         PIRKKALA ET AL.
Figure 2. HSF2 steady-state mRNA levels on
hemin or TPA treatment in K562 cells. A)
Poly(A) mRNA isolated from control (C) and
hemin-treated (he; 30 mM for 6, 16, and 24 h)
K562 cells was analyzed by Northern blotting
using 32P-labeled cDNA probes for HSF2,
hsp70, thioredoxin (TRX), and b-actin.
Hsp70 and TRX were used as positive controls
for hemin inducibility and b-actin was used as
a control for equal loading of samples. The
mRNA sizes are indicated on the right. B)
Poly(A) mRNA isolated from control (C) and
TPA-treated (TPA; 10 nM for 1, 2, 3, 4, 6, 24,
and 48 h) K562 cells was analyzed by Northern
blotting using 32P-labeled cDNA probes for
HSF2 and GAPDH.

HSF2. The amounts of Hsc70 remained constant on      HSF2 gene expression is regulated both at the
both hemin and TPA treatment.                        transcriptional level and by mRNA stabilization in
   To examine whether the hemin-induced in-          hemin-treated K562 cells
crease and the TPA-induced decrease in HSF2
protein were due to changes in HSF2 mRNA             Next we wanted to establish whether the induction of
expression, K562 cells were exposed to hemin or      HSF2 expression on hemin treatment was regulated
TPA for various time periods and poly(A) mRNA        on the transcriptional level; nuclear run-on assay was
samples were analyzed by Northern blotting. In       performed with nuclei isolated from untreated, heat-
                                                     shocked, and hemin-treated K562 cells. Transcrip-
untreated cells, a basal HSF2 mRNA expression
                                                     tion analysis revealed a modest but consistent 1.5- to
was detected, and the levels of HSF2 mRNA grad-
                                                     2-fold increase in HSF2 gene transcription upon
ually increased up to sixfold by 24 h of hemin
                                                     exposure to hemin for 16 h (Fig. 3A), as normalized
treatment (Fig. 2A). The amounts of hsp70, TRX,
                                                     against b-actin transcription. As expected, in con-
and b-actin mRNAs were analyzed in comparison
                                                     trast to transcriptional induction of the classical heat
with HSF2 mRNA in the same samples. Consistent       shock genes hsp70 and hsp90 in response to hemin
with our previous results (28), hsp70 mRNA was       and heat shock, transcription of the HSF2 gene was
also induced by hemin treatment, but, in contrast    not induced by heat shock (Fig. 3A). This is consis-
to HSF2 mRNA, hsp70 mRNA expression reached          tent with the finding that HSF2 DNA binding is
the maximum level already at 16 h and started to     activated in hemin-treated but not in heat-shocked
decrease thereafter (Fig. 2A). mRNA expression of    K562 cells (26). It is worth noting that the transcrip-
human TRX, another gene that has been shown to       tional induction of HSF2 gene was not detected in
be regulated in concert with HSF2 activation (28),   the earlier study (42). This discrepancy may be due
was increased by hemin with kinetics similar to      to the differences in experimental conditions, be-
HSF2 mRNA (Fig. 2A). b-Actin was used to con-        cause a whole plasmid containing the mouse HSF2
firm the equal loading of mRNA samples. In           cDNA was used earlier as a probe instead of the
contrast to the results obtained with hemin, upon    more specific human HSF2 cDNA insert used in the
exposure to TPA, HSF2 mRNA expression was            present study (for details, see Materials and
gradually down-regulated after 6 h of TPA treat-     Methods).
ment to the extent that the HSF2 mRNA levels            Because the prominently elevated HSF2 steady-
were decreased to 40% from the levels detected in    state mRNA levels on hemin treatment (Fig. 2A)
untreated cells, as normalized against GAPDH         were unlikely due to the modest, at most twofold
mRNA levels; HSF2 mRNA levels were barely de-        transcriptional induction (Fig. 3A), we wanted to
tectable at 24 – 48 h of TPA treatment (Fig. 2B).    determine whether the half-life of HSF2 mRNA

DIFFERENTIAL EXPRESSION OF HSF2                                                                         1093
Figure 3. Induction of HSF2 gene transcription and stabilization of HSF2 mRNA in
                                      K562 cells undergoing erythroid differentiation. A) Transcription rates of HSF2,
                                      hsp70, hsp90, and b-actin genes were analyzed by nuclear run-on assay. Equal
                                      number of nuclei from control (C), heat-shocked (HS; 1 h at 42°C), and hemin-
                                      treated (he; 30 mM for 6 and 16 h) K562 cells were used for in vitro labeling of newly
                                      synthesized transcripts that were hybridized to immobilized DNA probes. Hsp70 and
                                      hsp90 were used as positive controls for heat shock and hemin inducibility, and
                                      b-actin was used as an internal control for equal loading of samples. BS indicates
                                      plasmid Bluescript. Note the longer exposure time for the HSF2 blot. B) After
                                      treatment of K562 cells with hemin (30 mM for 16 h) or heat shock (HS; 1 h at
                                      42°C), treated and control cells were incubated with actinomycin D (actD; 6.4
                                      mg/ml) for 1, 2, 3, 4, and 6 h. Poly(A) mRNA was isolated and analyzed by Northern
                                      blotting using 32P-labeled cDNA probes for HSF2, hsp70, and TRX. The mRNA sizes
                                      are indicated on the right. C) The intensities of radioactive signals were quantitated
                                      using a phosphorimaging scanner and the values obtained for HSF2 mRNA were
                                      normalized against the respective values for TRX mRNA, the half-life of which was
                                      not affected by hemin treatment.

was affected by hemin. To prevent de novo gene            cells when the values were normalized against TRX
transcription, actinomycin D was added to control,        mRNA (Fig. 3B, C). The half-life of HSF2 mRNA
hemin-treated, or heat-shocked cells for different        was less than 1 h in heat-shocked cells, whereas it
time periods, and the HSF2 mRNA levels were               was ;2 h in control cells. We also analyzed the
monitored over the ensuing 4 h period. TRX                half-life of hsp70 mRNA in control, hemin-treated,
mRNA was used as a normalization control in the           and heat-shocked cells. As shown in Fig. 3B, the
quantitative analysis, since the half-life of TRX         half-life of hsp70 mRNA was ;2 h in control cells
mRNA was not affected by hemin (Fig. 3B, C).              and 4 h in hemin-treated and heat-shocked cells,
Exposure to hemin led to a marked stabilization of        which is in agreement with earlier studies (43, 44).
HSF2 mRNA. After 4 h of actinomycin D treat-              To examine whether the stabilization of HSF2
ment, the relative level of HSF2 mRNA in hemin-           mRNA required novel protein synthesis, K562 cells
treated cells was clearly higher than in control          were incubated in the presence of cycloheximide,

1094   Vol. 13   June 1999                    The FASEB Journal                                           PIRKKALA ET AL.
Figure 4. Down-regulation of HSF2
                                                                                mRNA expression on TPA treatment is
                                                                                mediated through the HSF2 promoter.
                                                                                Whole-cell extracts (12 mg) from control
                                                                                (C), hemin-treated (he; 30 mM for 24 h),
                                                                                and TPA-treated (T; 10 nM for 24 h)
                                                                                K562 cells as well as K562 cells stably
                                                                                overexpressing either mouse HSF2-a (2a-
                                                                                C7) or mouse HSF2-b (2b-D5) isoforms
                                                                                under the control of human b-actin pro-
                                                                                moter (27) were analyzed on an 8%
                                                                                SDS-PAGE and immunoblotted using an-
tibodies against HSF2, Hsp70, and Hsc70. Note that the exogenous mouse HSF2 isoforms (mHSF2-a and mHSF2-b) migrate
slightly faster on an SDS-PAGE than the endogenous human HSF2 isoforms (the upper double band, see legend to Fig. 1C), and
can therefore be separated from the slower migrating human HSF2 counterparts.

an inhibitor of protein synthesis, either alone or in          The differentiation lineage-dependent expression
combination with hemin, and poly(A) mRNA lev-                  of HSF2 upon treatment with hemin or TPA is
els were analyzed. However, no marked changes in               specific for K562 cells
the HSF2 mRNA levels were observed during cy-
cloheximide treatment (data not shown), suggest-               Finally, we wanted to determine the specificity of the
ing that the stabilization of HSF2 mRNA is not                 differentiation pathway-dependent expression of
dependent on de novo protein synthesis. However,               HSF2. To this end, in addition to K562 cells, various
the amounts of HSF2 protein gradually decreased                human cell lines, such as Raji (Burkitt’s lymphoma),
on cycloheximide treatment (data not shown),                   Molt-4 (T-lymphoblastic leukemia), and HeLa (cer-
which is consistent with the study recently re-                vical carcinoma) cells, were treated with either he-
ported by Mathew and co-workers (29).                          min or TPA for 24 h and whole-cell extracts were
                                                               analyzed by Western blotting. As shown in Fig. 5, in
Down-regulation of HSF2 mRNA by TPA requires                   all cell lines tested, HSF2 protein was readily detect-
the presence of HSF2 promoter                                  able also in samples isolated from untreated cells. Yet
                                                               the amounts of HSF2 varied considerably between
                                                               different cell lines, Raji cells containing the highest
The decrease in HSF2 mRNA expression by TPA
                                                               and HeLa cells the lowest levels. A prominent hemin-
(Fig. 2B) prompted us to investigate whether this
                                                               induced increase in HSF2 protein was observed only
down-regulation occurred at the promoter level. For
                                                               in K562 cells. Similarly, the TPA-mediated loss of
this purpose, we made use of stably transfected K562
                                                               HSF2 protein was specific for K562 cells induced to
cell clones overexpressing either mouse HSF2-a (2a-
                                                               differentiate along the megakaryocytic lineage, since
C7) or HSF2-b (2b-D5) isoforms under the control
                                                               in the other cell lines there was essentially no change
of human b-actin promoter (27). Consistent with
                                                               in the amounts of HSF2 protein (Fig. 5). Hsc70
earlier results, hemin-induced accumulation of both
                                                               protein levels are shown as a control for equal
the endogenous and exogenous HSF2 protein was
                                                               loading of samples.
observed in 2a-C7 cells, but not in 2b-D5 cells (Fig. 4;
ref 27). As expected, in K562 cells as well as in the
transfected cell clones, the endogenous human
HSF2 protein decreased during TPA treatment (Fig.              DISCUSSION
4). In contrast, the exogenous mouse HSF2 isoforms
expressed under the control of human b-actin pro-              We have shown that HSF2 protein levels increase in
moter were not affected by TPA treatment (Fig. 4),             concert with acquisition of DNA binding activity
indicating that the inhibiting effect of TPA on HSF2           during hemin-induced erythroid differentiation of
expression is likely to be mediated through the                human K562 cells (27, 42). In this study, we have
endogenous HSF2 promoter.                                      analyzed the molecular levels at which the regulation

                                                                               Figure 5. The differential expression of
                                                                               HSF2 upon treatment with hemin or TPA is
                                                                               specific for K562 cells. Whole-cell extracts
                                                                               (12 mg) from control (C), hemin-treated
                                                                               (he; 30 mM for 24 h), and TPA-treated (T;
                                                                               10 nM for 24 h) K562, Raji, Molt-4, and
                                                                               HeLa cells were analyzed on an 8% SDS-
                                                                               PAGE and immunoblotted using antibodies
                                                                               against HSF2 and Hsc70.

DIFFERENTIAL EXPRESSION OF HSF2                                                                                       1095
of HSF2 expression occurs in K562 cells induced to          of HSF2 for maintaining and promoting erythroid
differentiate along the erythroid or megakaryocytic         properties is further emphasized by the absence of this
lineage. Our results reveal that in hemin-treated           factor in K562 cells undergoing megakaryocytic differ-
K562 cells, HSF2 expression is up-regulated both at         entiation. It is well established that certain hematopoi-
the transcriptional level and by mRNA stabilization.        etic-restricted transcription factors, such as GATA-1
The enhanced expression of HSF2 is strictly specific        and NF-E2, are coexpressed within the erythroid and
for the erythroid properties of K562 cells; during          megakaryocytic differentiation lineages, raising the
TPA-induced megakaryocytic differentiation, HSF2            possibility of common programs or mechanisms of
expression is efficiently and rapidly down-regulated,       gene activation in the various differentiation pathways
leading to a complete loss of HSF2 protein.                 (51–53). GATA-1 was originally identified through its
   The K562 erythroleukemia cell line provides a            interaction within the b-globin locus control region
unique human cell model with which to study gene            (54, 55). Subsequently, a consensus DNA binding motif
expression during hematopoiesis. In vivo, the prolif-       for GATA-1 has been found in the cis-regulatory ele-
eration and maturation of megakaryocyte and ery-            ments of virtually all known erythroid-expressed gene
throid precursors are regulated by two structurally         promoters (for review, see ref 56). In K562 cells, the
related growth factors, thrombopoietin and erythro-         hemin-induced in vivo binding of HSF2 to the HSEs
poietin, respectively (for review, see ref 45). K562        within the human hsp70 promoter results in transcrip-
cells possess several erythroid properties that are         tional induction of the hsp70 gene (26, 42). However,
enhanced upon treatment with hemin (3, 46 – 48),            it is evident that HSF2 might have other target genes
whereas treatment with TPA leads to loss of their           apart from the known heat shock genes (22, 28). We
erythroid characteristics, directing these cells toward     speculate that during erythroid differentiation of K562
megakaryocytic lineage (49). We show here that the          cells, in addition to or instead of activating the classical
expression of HSF2 is differentially regulated in           heat shock genes, HSF2 might be a potential candidate
K562 cells undergoing either erythroid or                   either alone or in combination with certain erythroid-
megakaryocytic differentiation. HSF2 seems to be            specific transcription factor(s) to regulate the expres-
needed for maintaining and enhancing the ery-               sion of erythroid-specific genes. In contrast, during
throid properties of K562 cells, as characterized by        differentiation along the megakaryocytic lineage, HSF2
measuring globin expression (27), whereas HSF2 is           seems to be dispensable, and its expression is down-
rapidly and efficiently down-regulated during TPA-          regulated.
mediated differentiation along the megakaryocytic              The inducible expression of HSF2 indicates a differ-
lineage. Identification and characterization of spe-        ent regulatory mechanism as compared with the stress-
cific markers for the distinct hematopoietic lineages       responsive transcription factor HSF1. Upon activation,
are important for understanding the molecular               expression of the HSF1 gene appears to remain unal-
mechanisms that underlie the differentiation pro-           tered but the HSF1 protein undergoes posttransla-
cesses and malignant transformation. Recently, dif-         tional modifications such as oligomerization (i.e., trim-
ferential expression of the Kell blood group and            erization) and hyperphosphorylation (34, 57).
CD10 antigens, two related membrane metallopep-             Although HSF2 is known to be converted from an inert
tidases, was reported in K562 cells undergoing              dimeric state to an active trimer, the activation process
megakaryocytic and erythroid differentiation (50).          is accompanied by enhanced accumulation of HSF2
Expression of Kell and CD10 antigens represent              protein (42). In this study, the transcriptional induc-
relatively late differentiation markers, whereas ex-        tion of HSF2 in hemin-treated K562 cells was found to
pression of HSF2 is up-regulated and down-regu-             be 1.5- to 2-fold, and the stability of HSF2 mRNA was
lated within a few hours in the presence of hemin           markedly increased in the presence of hemin, in con-
and TPA, respectively. Furthermore, the differential        trast to the rapid decay of HSF2 mRNA in untreated
regulation of HSF2 seems to be specific for the             and heat-shocked cells. In general, although the mech-
progenitor cell-like ability of K562 cells to differen-     anisms underlying mRNA stabilization are not yet well
tiate along several lineages, since in Raji, Molt-4, and    understood, at least certain conserved features affect-
HeLa cells the levels of HSF2 protein did not essen-        ing the mRNA half-lives, such as sequence determi-
tially change when treated with hemin or TPA.               nants and trans-acting regulatory factors, have been
Therefore, HSF2 expression could provide an early           characterized (for review, see ref 58). By using cyclo-
hallmark for lineage-specific differentiation path-         heximide to inhibit protein synthesis, we show that the
ways of K562 cells.                                         stabilization of HSF2 mRNA by hemin does not involve
   Although HSF2 is present in various cell types and       novel synthesis of an HSF2 mRNA binding protein.
tissues (22, 31), it is to our knowledge the first tran-    Whether the half-life of HSF2 mRNA is regulated by
scription factor described whose expression is strictly     stable RNA-interacting protein(s) remains to be eluci-
regulated in K562 cells differentiating along the ery-      dated. However, it was found that in the presence of
throid and megakaryocytic lineages. The importance          cycloheximide, the levels of HSF2 protein rapidly de-

1096   Vol. 13   June 1999                        The FASEB Journal                                    PIRKKALA ET AL.
creased, suggesting that the hemin-mediated increase                      12.   Wiederrecht, G., Seto, D., and Parker, C. S. (1988) Isolation of
                                                                                the gene encoding the S. cerevisiae heat shock transcription
in HSF2 protein during erythroid differentiation could                          factor. Cell 54, 841– 853
be due to a stabilizing effect of hemin on some yet                       13.   Clos, J., Westwood, J. T., Becker, P. B., Wilson, S., Lambert, K.,
unknown HSF2-interacting protein(s). The short half-                            and Wu, C. (1990) Molecular cloning and expression of a
                                                                                hexameric Drosophila heat shock factor subject to negative
life of HSF2 protein might serve an important regula-                           regulation. Cell 63, 1085–1097
tory function considering the need for rapid down-                        14.   Rabindran, S. K., Giorgi, G., Clos, J., and Wu, C. (1991)
regulation of HSF2 during the megakaryocytic                                    Molecular cloning and expression of a human heat shock factor,
                                                                                HSF1. Proc. Natl. Acad. Sci. U. S. A. 88, 6906 – 6910
differentiation.                                                          15.   Sarge, K. D., Zimarino, V., Holm, K., Wu, C., and Morimoto,
   In conclusion, HSF2 provides an example of tran-                             R. I. (1991) Cloning and characterization of two mouse heat
scription factors, the expression of which is strictly                          shock factors with distinct inducible and constitutive DNA-
                                                                                binding ability. Genes Dev. 5, 1902–1911
regulated at multiple levels in a differentiation lin-                    16.   Schuetz, T. J., Gallo, G. J., Sheldon, L., Tempst, P., and
eage-specific manner. In light of our study, regula-                            Kingston, R. E. (1991) Isolation of a cDNA for HSF2: evidence
tion of HSF2 expression could be one of the key                                 for two heat shock factor genes in humans. Proc. Natl. Acad. Sci.
                                                                                U. S. A. 88, 6911– 6915
determinants in the commitment of K562 cells to                           17.   Nakai, A., and Morimoto, R. I. (1993) Characterization of a
either erythroid or megakaryocytic pathway of differ-                           novel chicken heat shock transcription factor, HSF3, suggests a
entiation. The processes actually governing cell dif-                           new regulatory pathway. Mol. Cell. Biol. 13, 1983–1997
                                                                          18.   Nakai, A., Tanabe, M., Kawazoe, Y., Inazawa, J., Morimoto, R. I.,
ferentiation in vivo are certainly more complex, and                            and Nagata, K. (1997) HSF4, a new member of the human heat
only spatially and temporarily organized combina-                               shock family which lacks properties of a transcriptional activa-
tions of various components can ensure normal                                   tor. Mol. Cell. Biol. 17, 469 – 481
                                                                          19.   Mezger, V., Rallu, M., Morimoto, R. I., Morange, M., and
hematopoietic development.                                                      Renard, J.-P. (1994) Heat shock factor 2-like activity in mouse
                                                                                blastocysts. Dev. Biol. 166, 819 – 822
   We thank Olli Ritvos for valuable suggestions concerning               20.   Murphy, S. P., Gorzowski, J. J., Sarge, K. D., and Phillips, B.
HSF2 expression during TPA-mediated megakaryocytic dif-                         (1994) Characterization of constitutive HSF2 DNA-binding ac-
ferentiation of K562 cells. We are also grateful to John E.                     tivity in mouse embryonal carcinoma cells. Mol. Cell. Biol. 14,
Eriksson, Carina I. Holmberg, Panu Jaakkola, and Päivi J.                      5309 –5317
Koskinen for discussions and critical comments on the manu-               21.   Sarge, K. D., Park-Sarge, O.-K., Kirby, J. D., Mayo, K. E., and
                                                                                Morimoto, R. I. (1994) Expression of heat shock factor 2 in
script. This work was supported by the Academy of Finland,                      mouse testis: potential role as a regulator of heat-shock protein
the Sigrid Jusélius Foundation, and the Finnish Cancer                         gene expression during spermatogenesis. Biol. Reprod. 50, 1334 –
Organization (L.S.), and by Turku Graduate School of Bio-                       1343
medical Sciences (L.P. and T.-P.A.).                                      22.   Rallu, M., Loones, M. T., Lallemand, Y., Morimoto, R. I.,
                                                                                Morange, M., and Mezger, V. (1997) Function and regulation of
                                                                                heat shock factor 2 during mouse embryogenesis. Proc. Natl.
                                                                                Acad. Sci. U. S. A. 94, 2392–2397
REFERENCES                                                                23.   Alastalo, T.-P., Lönnström, M., Leppä, S., Kaarniranta, K.,
                                                                                Pelto-Huikko, M., Sistonen, L., and Parvinen, M. (1998) Stage-
 1.   Felsenfeld, G. (1992) Chromatin as an essential part of the               specific expression and cellular localization of the heat shock
      transcriptional mechanism. Nature (London) 355, 219 –224                  factor 2 isoforms in the rat seminiferous epithelium. Exp. Cell
 2.   Enver, T., and Greaves, M. (1998) Loops, lineage, and leukemia.           Res. 240, 16 –27
      Cell 94, 9 –12                                                      24.   Singh, M. K., and Yu, J. (1984) Accumulation of a heat shock-
 3.   Lozzio, C. B., and Lozzio, B. B. (1975) Human chronic myelog-             like protein during differentiation of human erythroid cell line
      enous leukemia cell line with positive Philadelphia chromo-               K562. Nature (London) 309, 631– 633
      some. Blood 45, 321–334                                             25.   Theodorakis, N. G., Zand, D. J., Kotzbauer, P. T., Williams,
 4.   Cioe, L., McNab, A., Hubbel, H. R., Meo, P., Curtis, P., and              G. T., and Morimoto, R. I. (1989) Hemin-induced transcrip-
      Rovera, G. (1981) Differential expression of the globin genes in          tional activation of the hsp70 gene during erythroid maturation
      human leukemia K562(S) cells induced to differentiate by                  in K562 cells is due to a heat shock factor-mediated stress
      hemin or butyric acid. Cancer Res. 41, 237–243                            response. Mol. Cell. Biol. 9, 3166 –3173
 5.   Dean, A., Erard, F., Schneider, A. B., and Schechter, A. N.         26.   Sistonen, L., Sarge, K. D., Phillips, B., Abravaya, K., and Mori-
      (1981) Induction of hemoglobin accumulation in human K562                 moto, R. I. (1992) Activation of heat shock factor 2 during
      cells by hemin is reversible. Science 212, 459 – 461                      hemin-induced differentiation of human erythroleukemia cells.
 6.   Charnay, P., and Maniatis, T. (1983) Transcriptional regulation           Mol. Cell. Biol. 12, 4104 – 4111
      of globin gene expression in the human erythroid cell line          27.   Leppä, S., Pirkkala, L., Saarento, H., Sarge, K. D., and Sistonen,
      K562. Science 220, 1281–1283                                              L. (1997) Overexpression of HSF2-b inhibits hemin-induced
 7.   Alitalo, R. (1990) Induced differentiation of K562 leukemia               heat shock gene expression and erythroid differentiation in
      cells: a model for studies of gene expression in early megakaryo-         K562 cells. J. Biol. Chem. 272, 15293–15298
      blasts. Leukemia Res. 14, 501–514                                   28.   Leppä, S., Pirkkala, L., Chow, S. C., Eriksson, J. E., and Sistonen,
 8.   Glazer, R. I. (1994) Protein kinase C in multidrug resistance,            L. (1997) Thioredoxin is transcriptionally induced upon activa-
      neoplastic transformation, and differentiation. In Protein Kinase         tion of heat shock factor 2. J. Biol. Chem. 272, 30400 –30404
      C (Kuo, J. F., ed) pp. 171–198, Oxford University Press, New        29.   Mathew, A., Mathur, S. K., and Morimoto, R. I. (1998) The heat
      York                                                                      shock response and protein degradation: regulation of HSF2 by
 9.   Hong, Y., Martin, J. F., Vainchenker, W., and Erusalimsky, J. D.          the ubiquitin-proteasome pathway. Mol. Cell. Biol. 18, 5091–5098
      (1996) Inhibition of protein kinase C suppresses megakaryo-         30.   Liu, X.-D., Liu, P. C. C., Santoro, N., and Thiele, D. J. (1997)
      cytic differentiation and stimulates erythroid differentiation in         Conservation of a stress response: human heat shock transcrip-
      HEL cells. Blood 87, 123–131                                              tion factors functionally substitute for yeast HSF. EMBO J. 16,
10.   Wu, C. (1995) Heat shock transcription factors: structure and             6466 – 6477
      regulation. Annu. Rev. Cell Dev. Biol. 11, 441– 469                 31.   Fiorenza, M. T., Farkas, T., Dissing, M., Kolding, D., and
11.   Morimoto, R. I. (1998) Regulation of the heat shock transcrip-            Zimarino, V. (1995) Complex expression of murine heat shock
      tional response: cross talk between a family of heat shock                transcription factors. Nucleic Acids Res. 23, 467– 474
      factors, molecular chaperones, and negative regulators. Genes       32.   Goodson, M. L., Park-Sarge, O.-K., and Sarge, K. D. (1995)
      Dev. 12, 3788 –3796                                                       Tissue-dependent expression of heat shock factor 2 isoforms

DIFFERENTIAL EXPRESSION OF HSF2                                                                                                                1097
with distinct transcriptional activities. Mol. Cell. Biol. 15, 5288 –         tion rather than transcriptional activation. Proc. Natl. Acad. Sci.
       5293                                                                          U. S. A. 95, 2319 –2324
33.    Mosser, D. D., Theodorakis, N. G., and Morimoto, R. I. (1988)           45.   Metcalf, D. (1994) Thrombopoietin—at last. Nature (London)
       Coordinate changes in heat shock element-binding activity and                 369, 519 –520
       HSP70 gene transcription rates in human cells. Mol. Cell. Biol. 8,      46.   Andersson, L. C., Nilsson, K., and Gahmberg, C. G. (1979)
       4736 – 4744                                                                   K562—a human erythroleukemia cell line. Int. J. Cancer 23,
34.    Sarge, K. D., Murphy, S. P., and Morimoto, R. I. (1993)                       143–147
       Activation of heat shock gene transcription by HSF1 involves            47.   Andersson, L. C., Jokinen, M., and Gahmberg, C. G. (1979)
       oligomerization, acquisition of DNA binding activity, and nu-                 Induction of erythroid differentiation in the human leukaemia
       clear localization and can occur in the absence of stress. Mol.               cell line K562. Nature (London) 278, 364 –365
       Cell. Biol. 13, 1392–1407                                               48.   Rutherford, T. R., Clegg, J. B., and Weatherall, D. J. (1979) K562
35.    Wu, B., Hunt, C., and Morimoto, R. I. (1985) Structure and                    human leukaemic cells synthesize embryonic haemoglobin in
       expression of the human gene encoding major heat shock                        response to haemin. Nature (London) 280, 164 –165
       protein HSP70. Mol. Cell. Biol. 5, 330 –341                             49.   Tetteroo, P. A., Massaro, F., Mulder, A., Schreuder-van Gelder,
36.    Fort, P., Marty, L., Piechaczyk, M., El Sabrouty, S., Dani, C.,               R., and von dem Borne, A. E. (1984) Megakaryoblastic differ-
       Jeanteur, P., and Blanchard, J. M. (1985) Various rat adult                   entiation of proerythroblastic K562 cell-line cells. Leuk. Res. 8,
       tissues express only one major mRNA species from the glycer-                  197–206
       aldehyde-3-phosphate dehydrogenase multigenic family. Nucleic           50.   Belhacène, N., Maulon, L., Guérin, S., Ricci, J. E., Mari, B.,
       Acids Res. 13, 1431–1442                                                      Colin, Y., Cartron, J. P., and Auberger, P. (1998) Differential
37.    Gunning, P., Ponte, P., Okayama, H., Engel, J., Blau, H., and                 expression of the Kell blood group and CD10 antigens: two
       Kedes, L. (1983) Isolation and characterization of full-length                related membrane metallopeptidases during differentiation of
       cDNA clones for human a-, b-, and g-actin mRNAs: skeletal but                 K562 cells by phorbol ester and hemin. FASEB J. 12, 531–539
       not cytoplasmic actins have an amino-terminal cysteine that is          51.   Martin, D. I. K., Zon, L. I., Mutter, G., and Orkin, S. H. (1990)
       subsequently removed. Mol. Cell. Biol. 3, 787–795                             Expression of an erythroid transcription factor in megakaryo-
38.    Banerji, S. S., Theodorakis, N. G., and Morimoto, R. I. (1984)                cytic and mast cell lineages. Nature (London) 344, 444 – 447
       Heat shock-induced translational control of HSP70 and globin            52.   Romeo, P.-H., Prandini, M.-H., Joulin, V., Mignotte, V., Prenant,
       synthesis in chicken reticulocytes. Mol. Cell. Biol. 4, 2437–2448             M., Vainchenker, W., Marguerie, G., and Uzan, G. (1990)
39.    Hickey, E., Brandon, S. E., Smale, G., Lloyd, D., and Weber,                  Megakaryocytic and erythrocytic lineages share specific tran-
       L. A. (1989) Sequence and regulation of a gene encoding a                     scription factors. Nature (London) 344, 447– 449
       human 89-kilodalton heat shock protein. Mol. Cell. Biol. 9,
                                                                               53.   Andrews, N. C., Erjument-Bromage, H., Davidson, M. B.,
       2615–2626
                                                                                     Tempst, P., and Orkin, S. H. (1993) Erythroid transcription
40.    Mäkelä, T. P., Alitalo, R., Paulsson, Y., Westermark, B., Heldin,
                                                                                     factor NF-E2 is a haematopoietic-specific basic-leucine zipper
       C. H., and Alitalo, K. (1987) Regulation of platelet-derived
                                                                                     protein. Nature (London) 363, 722–728
       growth factor gene expression by transforming growth factor
                                                                               54.   Ney, P. A., Sorrentino, B. P., McDonagh, K. T., and Nienhuis,
       beta and phorbol ester in human leukemia cell lines. Mol. Cell.
                                                                                     A. W. (1990) Tandem AP-1-binding sites within the human
       Biol. 7, 3656 –3662
                                                                                     beta-globin dominant control region function as an inducible
41.    Holmberg, C. I., Leppä, S., Eriksson, J. E., and Sistonen, L.
                                                                                     enhancer in erythroid cells. Genes Dev. 4, 993–1006
       (1997) The phorbol ester 12-O-tetradecanoylphorbol 13-acetate
       enhances the heat-induced stress response. J. Biol. Chem. 272,          55.   Talbot, D., and Grosveld, F. (1991) The 59HS2 of the globin
       6792– 6798                                                                    locus control region enhances transcription through the inter-
42.    Sistonen, L., Sarge, K. D., and Morimoto, R. I. (1994) Human                  action of a multimeric complex binding at two functionally
       heat shock factors 1 and 2 are differentially activated and can               distinct NF-E2 binding sites. EMBO J. 10, 1391–1398
       synergistically induce hsp70 gene transcription. Mol. Cell. Biol.       56.   Orkin, S. H. (1995) Transcription factors and hematopoietic
       14, 2087–2099                                                                 development. J. Biol. Chem. 270, 4955– 4958
43.    Theodorakis, N. G., and Morimoto, R. I. (1987) Posttranscrip-           57.   Baler, R., Dahl, G., and Voellmy, R. (1993) Activation of human
       tional regulation of hsp70 expression in human cells: effects of              heat shock genes is accompanied by oligomerization, modifica-
       heat shock, inhibition of protein synthesis, and adenovirus                   tion, and rapid translocation of heat shock transcription factor
       infection on translation and mRNA stability. Mol. Cell. Biol. 7,              HSF1. Mol. Cell. Biol. 13, 2486 –2496
       4357– 4368                                                              58.   Ross, J. (1995) mRNA stability in mammalian cells. Microbiol.
44.    Kaarniranta, K., Elo, M., Sironen, R., Lammi, M. J., Goldring,                Rev. 59, 423– 450
       M. B., Eriksson, J. E., Sistonen, L., and Helminen, H. J. (1998)
       Hsp70 accumulation in chondrocytic cells exposed to high                                                Received for publication October 14, 1998.
       continuous hydrostatic pressure coincides with mRNA stabiliza-                                          Revised for publication January 18, 1999.

1098      Vol. 13    June 1999                                      The FASEB Journal                                                PIRKKALA ET AL.
You can also read