(Received for publication, November 22, 1996)
From the Department of Pharmaceutical Sciences,
University of Colorado School of Pharmacy and the University of
Colorado Cancer Center, Denver, Colorado 80262 and the
DNA
Technology Group, Amgen, Inc., Boulder, Colorado 80301
DNA topoisomerase II (topo II
) is an
essential proliferation-dependent nuclear enzyme which has been
exploited as an anti-tumor drug target. Since the proliferative status
of human leukemia cells is associated with expression of the
c-myb proto-oncogene, c-Myb was investigated as a
trans-activator of the topo II
gene. Using topo II
promoter-luciferase reporter plasmids, c-myb expression caused trans-activation of the topo II
promoter a
maximum of ~4.5-fold over basal levels in HL-60 human promyelocytic
leukemia cells. Trans-activation was submaximal with higher
levels of c-myb expression plasmid but a Myb protein
lacking its negative regulatory domain resulted in ~19-fold
trans-activation. Mutagenesis and 5
-deletion studies
revealed that Myb trans-activation was mediated via a
Myb-binding site at positions
16 to
11 and that this region governed the bulk of basal topo II
promoter activity in human leukemia cells. Trans-activation of topo II
by c-Myb was
lymphoid- or myeloid-dependent. However, B-Myb, a more
widely-expressed Myb family member, caused topo II
trans-activation in both HL-60 cells and HeLa epithelial
cervical carcinoma cells. These data provide evidence for a new
Myb-responsive gene which is directly linked to and required for
cellular proliferation.
DNA topoisomerase II (topo II)1 is a nuclear enzyme whose catalytic activity is absolutely required for cellular proliferation. The enzyme catalyzes the relaxation and decatention of DNA as a result of its unique coordination of double-strand DNA breakage, strand passage, and religation activities (reviewed in Refs. 1 and 2) via an ATP-modulated, protein-clamp intermediate (3, 4). Roles for topo II catalytic activity have been demonstrated or proposed in DNA replication, transcription, and chromosomal segregation (5-7). Its role in decatenating newly replicated DNA has been shown to be essential for cellular survival in yeast (5, 8, 9). In addition, nuclear protein-protein interactions with topo II may have implications for modulating its catalytic activity (10) and in maintaining stability of the genome (11).
Two forms of mammalian topo II have been identified ( and
), each
encoded by distinct genes located on separate chromosomes (12). The
form of the enzyme has been the most intensively studied since it is
usually the more abundant form in proliferating cells (13). Not
surprisingly, topo II
expression has been linked to cellular
proliferative status in nearly every system studied (14-18). However,
little is known regarding the transcriptional mechanisms underlying
this proliferation-dependent expression of topo II
.
Recent data suggest that the NF-Y transcription factor family may be
responsible for loss of topo II
expression in confluence-arrested cells (19).
In previous investigations, HL-60 leukemia cells have proven an
excellent model in which to study topo II transcriptional regulation
during the transition from growth to terminal differentiation (20). In
these studies, promoter-reporter deletion analysis of the topo II
5
-flanking region revealed the presence of a consensus c-Myb-binding
site within the proximal promoter (20). c-Myb has therefore emerged as
the first candidate factor for regulating topo II
gene expression in
proliferating hematopoietic cells.
The viral myb oncogene was described independently in two avian retroviruses: avian myeloblastosis virus and the E26 virus (21). The v-Myb protein was originally shown to be a sequence-specific DNA-binding transcriptional activating protein (22) transduced from a normal cellular progenitor, c-Myb (23). Cellular and viral Myb proteins can bind a variety of sequences but the consensus that has emerged is TAAC(G/T)G or TAACNG (22, 24, 25). Disruption of DNA binding activity is most commonly accomplished experimentally by mutating the TAAC core (26). Viral Myb proteins were found to be truncated at both the C and N termini, implying that normal regulation of this protein in non-transformed cells occurred via these regions and that transformation was a result of their truncation (27). As might be predicted, the cellular form of Myb is also implicated in normal cellular proliferation, albeit in a more regulated fashion. Specifically, the expression of c-Myb is closely linked to proliferative status in normal hematopoiesis; c-Myb is down-regulated following terminal differentiation to monocytes, neutrophils, or erythrocytes (28, 29). Antisense oligonucleotides to c-myb inhibit myeloid cell proliferation (30) while overexpression of c-Myb protein is known to inhibit cellular differentiation in erythroid and myeloid cell lines (31, 32). In vivo, transgenic mice homozygous for a null c-myb allele do not survive past gestation day 15 due to their lack of hepatic hematopoiesis (33). More directed, tissue-specific inactivation of c-Myb in transgenic mice expressing dominant negative c-Myb peptides in immature T cells blocked thymopoiesis and prevented proliferation of mature T cells (34).
Functionally, the c-Myb protein can be separated into three distinct domains: a DNA-binding domain, a trans-activating domain, and a negative regulatory domain (NRD). The DNA-binding domain consists of three 51-52 amino acid imperfect repeats, each containing three conserved tryptophan residues evenly spaced (35, 36). Protein NMR studies have recently confirmed that the second and third of these repeats are required for contact with the essential components of the Myb consensus binding site: the first A, C, and G nucleotides of the TAACNG sequence (37). The 50-amino acid trans-activation domain lies C-terminal to the DNA-binding domain and is hydrophilic and acidic (23), consistent with trans-activating domains of other factors. Much attention has been devoted to the negative regulatory domain present at the C terminus of the c-Myb protein. This region contains a classical leucine zipper and evidence currently suggests that when c-Myb homodimerizes via this domain, its DNA-binding and trans-activating capacity is compromised (38). At high c-Myb concentrations, dimerization is favored and DNA-binding and trans-activation are ameliorated (38); Myb is therefore described as being self-limiting or self-squelching. In fact, the transforming capacity of C-terminally truncated viral Myb proteins has been shown to result from increased trans-activating capacity due to loss of this negative regulatory domain (39). The C terminus also contains a serine phosphorylation site that negatively regulates trans-activating activity. Mutation of this phosphorylation site to alanine results in a 2-7-fold increase in the trans-activating capacity of c-Myb (40). The C-terminal domain appears to serve other functions as well. Favier and Gonda (41) have shown that this domain can bind other proteins which may act either as co-activators or suppressors (38). In fact, Ness' (42) group has shown that the N and C termini of c-Myb can associate and preclude binding of the co-activator, p100. Hence, there are multiple levels at which c-Myb can be regulated via its C-terminal negative regulatory domain.
In this report, we demonstrate that c-Myb activates expression of a
luciferase reporter gene under control of the human topo II
promoter. A role for c-Myb in topo II
expression appears, however,
to be restricted to cells of hematopoietic origin. Use of dominant
negative c-Myb inhibitors indicates that c-Myb is likely to play a
major role in basal topo II
expression in HL-60 leukemia cells.
Evidence is also presented that activation of topo II
expression may
extend to other more widely expressed Myb family members such as B-Myb,
a factor whose own expression is linked to the G1/S
boundary of the cell cycle. This report establishes an association
between topo II
, an enzyme essential to the completion of cell
division, and a transcription factor family linked intimately to tumor
cell proliferation.
HL-60 human promyelocytic leukemia cells (ATCC CCL 240), U937 histiocytic leukemia cells (ATCC CRL 1593), CCRF-CEM lymphoblastic leukemia cells (ATCC 119), and HeLa S3 human cervical carcinoma cells (ATCC CCL 2.2) were obtained from the American Type Culture Collection (Rockville, MD). All cells were maintained at 37° C in a humidified atmosphere containing 5% CO2. Exponentially growing suspension cultures of leukemia cells were propagated by subculturing at 5 × 105 cells/ml every 2-3 days in RPMI 1640 medium supplemented with 10% fetal bovine serum, 50 units/ml penicillin G, and 50 µg/ml streptomycin sulfate. HeLa cells were maintained in monolayer culture in Dulbecco's modified essential medium containing the same serum supplements and were subcultured at 5 × 105 cells per 100-mm dish every 2-3 days. For all gene transfer experiments, all cells were freshly subcultured at the indicated cell densities 48 h prior to transfection.
Reporter and Expression PlasmidsAll topo
II-promoter-reporter vectors were constructed in the background of
pA3LUC (43) as described previously (20). CMV-driven c-Myb
and c-Myb
NRD plasmids (44) were a gift from Dr. Prem Reddy (Fels
Institute, Temple University, Philadelphia, PA). The SV40-driven c-Myb
expression plasmid pMbm1 (31) was kindly provided by Dr. Edward
Prochownik (Children's Hospital, Pittsburgh, PA). The dominant
negative c-Myb expression plasmid (34) (pSCDMS/MenT) was a gift of Dr.
Kathy Weston (Institute for Cancer Research: Royal Cancer Hospital,
London). The human B-Myb expression plasmid (pKCB-myb) was generously
provided by Dr. Roger Watson (Ludwig Institute for Cancer Research,
London). Internal control
-galactosidase expression plasmids were
obtained from either Dr. David Gordon (RSV-
-gal; University of
Colorado Health Sciences Center, Denver, CO) or Clontech (CMV-
-gal).
A passive, dominant negative Myb plasmid (CMV-mybDBD) was constructed by polymerase chain reaction amplification of the DNA-binding domain of
c-Myb (codons 84-212) followed by subcloning into the HindIII/BamHI site of pcDNA3 (Invitrogen, La
Jolla, CA). The topo II promoter-reporter vector (
562TOP2LUC/MBSmut)
was constructed by polymerase chain reaction mutagenesis of the
putative Myb binding site at
16 to
11 from TAACCG to
T
C
using the 4-oligonucleotide method of
Higuchi et al. (45). The fidelity of all constructs generated by polymerase chain reaction was confirmed by dideoxy DNA
sequencing using Sequenase 2.0 (U. S. Biochemical Corp.). All plasmid
DNA used for transfections was propagated in Escherichia coli DH5
, then isolated by a standard alkaline-SDS lysis
protocol followed by double purification on isopycnic CsCl
gradients.
Exponentially growing
HL-60, U937, or CCRF-CEM cells were seeded as described above, then
collected by centrifugation at 700 × g for 5 min. The
resulting cell pellet was resuspended in fresh RPMI 1640, with all
supplements, at a cell density of at least 5 × 107
cells/ml. Two-hundred microliters of this cell suspension (~1 × 107 cells) was combined with 20 µg of reporter plasmid
and the indicated amounts of effector plasmid, then transferred to a
0.4-cm gap width metal-lined electroporation cuvette (Invitrogen). DNA
was introduced to the HL-60 cells by delivering a charge to the cell suspension using an IBI geneZAPPER electroporator set at 250 V and 950 microfarads. The observed time constant was usually within the
range of 48-56 ms. The cell suspension was then removed to a 60-mm
culture dish containing 3 ml of supplemented RPMI 1640. Harvest times
varied from 4 to 24 h as indicated in each figure. In general,
each group contained three samples and experiments were performed 3-5
times.
For the HeLa cell experiments, the procedure was essentially the same as described above except that Dulbecco's modified essential medium was used for resuspension and plating, 1 × 106 cells were used in each transfection, and the electroporation conditions were 180 V and 500 microfarads. Time constants generally ranged from 21 to 26 ms.
Experiments with CMV-driven expression vectors included an appropriate
amount of the empty CMV vector, pcDNA3, such that each group had an
equivalent total amount of CMV vectors. Controlling each group with
empty vector was required since the CMV promoter alone appears to usurp
general transcription factors; therefore, basal topo II promoter
activity varies between experiments as a result of the fact that
different amounts of total CMV vectors were used in each experiment.
However, the absolute magnitude of Myb-dependent
trans-activation remained relatively constant regardless of
the amount of CMV vectors employed.
Cell pellets collected by centrifugation at 1000 × g for 5 min were washed once with phosphate-buffered saline, then resuspended in 100 µl of 100 mM potassium phosphate, pH 7.8, containing 1 mM dithiothreitol. Cells were lysed by three cycles of freezing in dry ice, thawing at 37° C for 30-45 s, and vortexing for 15 s in order to liberate luciferase enzyme. Following the third freeze-thaw cycle lysates were centrifuged at 10,000 × g for 10 min, then supernatants were removed, quantitated, and immediately assayed for luciferase activity. Luciferase activity was determined by combining 25-40 µl of cell supernatant with 160-175 µl of luciferase assay buffer (100 mM potassium phosphate, pH 7.8, 1 mM dithiothreitol, 15 mM MgSO4, 5 mM ATP). The luciferin substrate (Analytical Bioluminescence, San Diego, CA) was prepared at a final concentration of 1 mM in 100 mM potassium phosphate, pH 7.8, 1 mM dithiothreitol, and 100 µl was used to initiate each reaction. Luminescence of each reaction was indicated by arbitrary light units as quantitated over 15 s by a Los Alamos Diagnostics 535 Luminometer (Turner Designs, Mountain View, CA). Each sample was assayed twice, background was subtracted, and individual data expressed as total light units backcalculated for the entire cell sample. Group data was expressed as mean ± S.E.
In most cases a plasmid containing the cytomegalovirus immediate early
promoter/enhancer upstream of the E. coli lacZ gene (encoding -galactosidase), denoted pCMV-
-gal, was co-transfected (0.5-1 µg) as an internal control. Preliminary experiments with B-myb expression plasmid revealed that the CMV-
-gal
construct was responsive to this factor. Therefore, experiments
employing the B-myb expression plasmid used an RSV-driven
-gal internal control to alleviate this problem. After luciferase
assay, an aliquot of cell supernatant was used for assay of
-gal
activity using 2-nitrophenyl-
-D-galactopyranoside as the
substrate according to the method detailed by Sambrook et
al. (46). E. coli
-galactosidase (Boehringer
Mannheim) was used as the standard. Luciferase values were corrected
for
-gal internal control activity by dividing total light units by
total milliunits of
-gal activity.
The common link between c-Myb, topo II, and
hematopoietic cell proliferation led to questioning whether c-Myb could
activate the topo II
promoter. To test this hypothesis, a CMV-driven
c-myb expression plasmid was co-transfected into HL-60
leukemia cells along with the topo II
promoter-luciferase reporter
construct,
562TOP2LUC. As illustrated in Fig. 1, c-Myb
was both a potent and efficacious activator of the topo II
promoter.
As little as 0.2 µg of c-myb expression plasmid activated
reporter expression reproducibly and this effect achieved a maximum of
4.6-fold over basal levels (Fig. 1). The expression of c-myb
under control of the less robust SV40 promoter (pMbm1) also activated
reporter expression more than 3-fold. Interestingly, the dose-response relationship with CMV-myb exhibited an inverted U-shape, as
has been described previously for other c-Myb-responsive genes (38). This diminution of trans-activation at high levels of
c-myb expression is believed to occur due to the
self-squelching nature of its NRD through which inhibitory protein
interactions and phosphorylation can occur (38, 40, 42).
It would then follow that a c-Myb molecule lacking the NRD but still
containing its DNA-binding and trans-activating domains would produce a more classical dose-response curve, that is, without the self-squelching plateau. To confirm this hypothesis with the topo
II promoter, HL-60 cells were co-transfected with
562TOP2LUC and
varying amounts of plasmids driving production of either full-length c-Myb or c-Myb lacking the NRD (
NRD). The latter truncated c-Myb protein proved to be a far more effective trans-activator of
the the topo II
promoter than wild-type c-Myb. As illustrated in Fig. 2, c-Myb produced a dose-response profile similar
to that in Fig. 1 with a maximum of 4.7-fold activation over basal
levels. In contrast, c-Myb
NRD resulted in 10.5-fold activation at an amount of expression plasmid where the wild-type c-Myb effect had
plateaued, achieving a peak of 18.6-fold stimulation of basal topo
II
promoter activity (Fig. 2). Specificity of both c-Myb proteins
for trans-activating the topo II
promoter was
demonstrated by the fact that neither construct activated reporter
activity when under control of a heterologous herpes simplex virus
thymidine kinase promoter (pT81LUC (47)) or with the promoterless
luciferase parent vector, pA3LUC (data not shown). Taken
together, these data support the suggestion that c-Myb
trans-activates the topo II
promoter and that this
trans-activation function is autoregulated via its
C-terminal NRD.
Experiments presented thus far employed a 20-24 h post-transfection
harvest time. It was reasoned that a temporal assessment of
c-Myb-mediated trans-activation of the topo II promoter
might support a direct role for c-Myb in this response, as opposed to c-Myb acting as an initiator of a cascade culminating in topo II
promoter activation. HL-60 cells were therefore co-transfected with
562TOP2LUC and either the control CMV empty vector or the CMV-myb
effector plasmid and luciferase activity quantitated at times ranging
from 4 to 18 h post-transfection. Strikingly, robust (3.2-fold)
topo II
promoter trans-activation was observed as quickly
as 4 h post-transfection (Fig. 3). Although basal
luciferase expression expectedly decreases at later time points, the
magnitude of trans-activation remained 3-5-fold at later
time points. If Myb were acting indirectly through activation of a
second factor, a greater delay in appearance of
trans-activation would have been expected. For example,
increased expression of topo II
or other unrelated genes does not
occur for 12-72 h after treatment with a pleiotropic activator of gene
expression, the histone deacetylase inhibitor sodium butyrate (20,
48-50). The observed rapidity of this c-Myb-mediated
trans-activation is therefore consistent with a direct role
for this factor in topo II
gene regulation.
c-Myb trans-Activation of the Topo II
The consensus c-Myb-binding site that
has been described is TAACNG, although flanking bases also appear to
contribute to binding affinity (22, 24, 25). A match for this site
exists in the human topo II promoter at
16 to
11 (5
-TAACCG-3
)
and this site is conserved exactly in the corresponding region of the
Chinese hamster topo II
promoter described recently (51). A series of 5
-promoter deletions were constructed in the context of the luciferase reporter vector and then independently co-transfected with
CMV-myb into HL-60 cells. As shown in Fig. 4,
c-Myb trans-activated both a larger promoter construct which
extended 5
from
562 (
1200) as well as minimal promoter-reporter
constructs (
90 and
51) by approximately the same magnitude
(3-4-fold) over basal levels. These data suggest that an element which
was positively responsive to c-Myb was located within
51 of the human
topo II
promoter, consistent with the location of the putative
Myb-binding site at
16 to
11.
To directly ascertain the functionality of this putative
c-Myb-binding site, the sequence TAACCG was altered by
sitedirected mutagenesis to TC
in the
context of
562TOP2LUC. As shown in Fig. 5, mutation of
this site completely abrogated c-Myb-dependent trans-activation in HL-60 cells. In addition, mutation
of the c-Myb site resulted in a dramatic 97% reduction in basal
activity of this construct (Fig. 5). This observation appears to be
unique to the topo II
gene since c-Myb-binding site mutations in
other gene promoters does not greatly affect basal transcription levels (52, 53). It is therefore possible that in the TATA-less topo II
promoter, the c-Myb site may overlap with transcriptional initiation
sequences. Consistent with this suggestion, the Myb site lies
immediately between two palandromic promoter sequences which have been
suggested previously to be involved in transcriptional initiation of
this gene (54). In fact, c-Myb may form a bridge with the basal
transcriptional apparatus through its recently demonstrated interaction
with p100, a factor known to bind TFIIE (42).
c-Myb as an Endogenous Regulator of the Topo II
Although mutation of the topo II promoter
c-Myb-binding site greatly reduced basal reporter expression from this
construct, its proximity to the transcriptional initiation site
confounds our understanding of the relative contribution of c-Myb to
basal topo II
expression in human leukemia cells. A question which therefore remained was whether c-Myb is an endogenous modulator of topo
II
gene expression in myeloid cells. The antisense oligonucleotide approach has been taken previously to ameliorate candidate
c-Myb-dependent responses (30). In preliminary experiments,
an antisense c-Myb expression vector co-transfected into HL-60 cells
did indeed cause a dose-dependent attenuation of topo II
promoter-driven luciferase activity (data not shown). However, a
scrambled control expression vector had a similar, but
non-sequence-specific effect. This data may be reconciled by the recent
demonstration of a non-antisense mechanism for antisense c-Myb
oligonucleotides (55). These sequences appear to exert an
antiproliferative effect via a "G quartet" sequence contained
within, possibly by sequestering basic fibroblast growth factor when
administered extracellularly (56). Therefore, these results indicated
an alternative approach would be necessary to inactivate endogenous
c-Myb.
Weston's group has recently constructed an elegant chimeric expression
vector which encodes the c-Myb DNA-binding domain fused with the
Drosophila engrailed transcriptional repressor domain (34).
When the plasmid is introduced into cells the resulting protein
competes for binding of Myb consensus sites with endogenous c-Myb,
repressing rather than activating transcription. This chimera was
demonstrated to be effective both in transfection assays as well as in
T-cell specific gene expression in transgenic mice (34). We therefore
obtained this dominant negative expression construct (pSCDMS/MenT) for
co-transfection of HL-60 cells with 562TOP2LUC. A second, passive,
dominant negative competitor of c-Myb binding was also employed which
consisted only of the c-Myb DNA-binding domain
(pCMV-mybDBD). To first demonstrate the effectiveness of
this approach, the dominant negative constructs were tested for their
ability to block both c-Myb or c-Myb
NRD-mediated
trans-activation of the topo II
promoter. As illustrated
in Fig. 6A, the passive and active c-Myb
dominant negative peptides effectively reduced c-Myb-stimulated
promoter activity to ~120 and 30% of basal levels, respectively,
whether wild-type or truncated c-Myb was used as the activator. This
finding confirmed that the dominant-negative Myb proteins could
attenutate the effects of exogenously-derived Myb. But most
importantly, the dominant negative peptides also reduced basal topo
II
promoter activity to 60% (for mybDBD) and 38%
(pSCDMS/MenT) of control (CMV only control group, Fig. 6A). This effect was specific for the topo II
promoter since the dominant negative Myb constructs had no effect on the internal control
-galactosidase expression plasmid. This finding suggests that c-Myb,
or perhaps a related family member, plays a central role in topo II
expression in proliferating HL-60 leukemia cells.
The specificity of the more efficacious dominant negative construct,
pSCDMS/MenT, was addressed by comparing its dose-response relationship
in HL-60 cells versus HeLa cells, a human cervical carcinoma
line known to lack c-Myb. In Fig. 6B, the dominant negative construct elicited a reproducible suppression of topo II promoter activity in HL-60 cells with as little as 0.2 µg of plasmid, and further suppression was dose-dependent. In contrast, up to
2 µg of pSCDMS/MenT had no suppressive effect and even slightly
stimulated the topo II
promoter-reporter when co-transfected into
HeLa cells. Therefore, the specificity of this dominant negative Myb
construct for endogenous c-Myb in HL-60 cells was confirmed by its lack of suppressive effect in a c-Myb-negative cell line (Fig.
6B). Furthermore, this lack of effect in HeLa cells also
suggests that the Myb dominant negative protein was not acting in HL-60
cells simply by physical blockade of transcriptional initiation.
We have also begun to investigate the cell type-specific
trans-activation by c-Myb of topo II promoter-reporter
constructs in other hematopoietic or lymphoid cells. While c-Myb is
normally expressed in immature cells of these lineages, only a subset
of these cells are capable of supporting trans-activation in
the case of one c-Myb target gene, mim-1 (57). Ness et
al. (58) demonstrated that cell-specific
trans-activation by c-Myb was dependent on at least one
differentially expressed factor, NF-M, which acts in a bipartite
fashion with c-Myb. It is clear that c-Myb alone is not capable of
substantial trans-activation in either epithelial cells or
fibroblasts (53, 58), presumably due to lack of a combinatorial
activator like NF-M or other recently characterized Myb co-activators
(42, 59). Therefore, co-transfections with CMV-myb and
562TOP2LUC were performed with three other human tumor cell lines:
U937 histiocytic leukemia cells, CCRF-CEM lymphoblastic pre-T cell
leukemia, and HeLa epithelial cervical carcinoma cells. As depicted in
Fig. 7, both the U937 and CCRF-CEM cells could support
substantial c-Myb-mediated trans-activation of the topo II
promoter-reporter construct. While CCRF-CEM cells were somewhat less inducible than U937 cells, it should be recognized that CCRF-CEM cells already possess the highest basal topo II
promoter activity of
any cell line tested to date; therefore, other cooperating factors may
already be in near-limiting quantities. In contrast, c-Myb was less
effective in stimulating topo II
transcription in HeLa cells,
especially with high amounts of c-Myb expression plasmid. These results
suggest that c-Myb regulation of topo II
is cell-type specific,
consistent with other previously described Myb target genes.
B-Myb Also trans-Activates the Human Topo II
In
almost every cell type tested, topo II abundance is directly
proportional to proliferation rate or growth fraction. Since c-Myb
production is restricted to hematopoietic cells with a few exceptions
(60), it seems unlikely that this factor is responsible for the
proliferation-dependent regulation of topo II
observed in numerous other cell types. Other Myb-related factors, especially B-Myb, appear to be more widely expressed in other tissues (61), including human myeloid cells (62). B-Myb is also intimately associated
with cell cycle progression. Of note, B-myb transcriptional repression in G0 and early G1 phase and
activation at the G1/S boundary are associated with two
distinct classes of E2F-containing DNA binding complexes (63). Recent
work also suggests the p107 retinoblastoma family member may exert its
growth-suppressive activity by repressing B-myb
transcription, perhaps in a complex with an E2F protein (64). With this
link to cellular growth control, B-Myb was therefore evaluated as a
potential activator of the topo II
promoter in both a myeloid and
non-myeloid cell line.
When a CMV-driven B-myb expression vector (pKCB-myb; gift of
Dr. R. J. Watson) was co-transfected with 562TOPLUC into HL-60 cells,
robust increases in reporter activity were observed (Fig. 8). While this activation required greater amounts of
B-myb expression plasmid than with c-myb plasmids
(Fig. 1), maximal promoter activation achieved a far greater magnitude
(15-fold at 15 µg) and a plateau was not observed (Fig. 8). This
observation suggests that putative B-Myb co-activators are not limiting
in HL-60 cells under the conditions employed. But in contrast with
c-Myb, B-Myb also trans-activated the topo II
promoter
construct in HeLa cervical carcinoma cells (Fig. 8). This observation
is consistent with the fact that in HeLa cells, B-Myb has been shown to
be an effective trans-activator of other c-Myb-responsive
genes (65). It is therefore conceivable that other more widely
expressed Myb family members like B-Myb serve to activate topo II
expression in non-hematopoietic cell types.
This report provides evidence for trans-activation of
the topo II promoter by a transcription factor family well known to play a role in cellular proliferation. Based on its catalytic role
essential for chromosomal segregation, topo II
is a logical downstream effector for the growth-stimulatory effects of the Myb
transcription factor family. Activation of the topo II
promoter by
c-Myb was comparable to that described for other c-Myb-responsive genes
in that, 1) an inverted U-shaped dose-response relationship was
observed for wild-type c-Myb; 2) a C-terminal-truncated c-Myb protein
lacking the NRD led to more robust trans-activation; 3) c-Myb-mediated trans-activation occurred rapidly following
transfection; 4) c-Myb-mediated trans-activation was
selective for hematopoietic cells; 5) mutation of the c-Myb-binding
site abrogated c-Myb-mediated trans-activation; and 6) a
dominant negative c-Myb protein was capable of attentuating
trans-activation by either c-Myb or c-Myb
NRD.
Of greatest relevance is the fact that the dominant negative c-Myb
proteins were able to block 40-60% of the basal activity of a topo
II promoter-reporter construct in HL-60 cells. Specifically, the
fact that the passive c-Myb DBD competitor reduced basal promoter activity supports an endogenous role for c-Myb in topo II
expression in HL-60 cells. The possibility has been considered that the dominant negative proteins simply prevented assembly of the basal transcription apparatus since an intact Myb site also appeared to be essential for
the majority of basal topo II
expression (Fig. 5). However, the more
active of the two c-Myb dominant negative proteins had no effect or
slightly stimulated topo II
promoter activity in HeLa
cells (Fig. 6B). This finding suggests that the dominant negative proteins acted in HL-60 cells by precluding c-Myb binding and
not via general blockade of transcriptional initiation. Taken together,
the mutagenesis and dominant negative experiments support a role for
endogenous c-Myb in topo II
expression in HL-60 cells.
The significance of this finding extends beyond basic studies of
transcriptional regulation. The transcriptional basis of topo II
regulation has also been investigated as a mechanism of tumor cell
resistance to drugs which trap topo II
in covalent complexes with
DNA (66). For example Sp3, an Sp1 family member which acts as a
transcriptional repressor, is believed to bind an Sp1 site in the human
topo II
promoter (67). In topo II-drug-resistant carcinoma cell
lines which underexpress topo II
, resistance appears to be due to
overexpression of the Sp3 repressor (67). With Myb proteins appearing
responsible for the bulk of basal topo II
promoter activity in HL-60
cells, altered Myb regulation should also be investigated as another
basis for topo II-drug resistance in leukemias.
While dominant negative Myb proteins also clearly abrogate
trans-activation by ectopically expressed c-Myb proteins, it
is not yet known whether the function of other Myb family members is
also inactivated by these inhibitors. This point will be particularly important to investigate further since B-Myb, which is also produced by
HL-60 cells (62), does not always bind to and trans-activate through the same element as c-Myb (24, 68). In fact, the lack of effect
of the Myb/engrailed chimera in HeLa cells (Fig.
6B), which are known to produce B-Myb, suggests that B-Myb
may not be acting through the c-Myb site at 16 to
11 of this
promoter.
This report also carries implications for understanding the role of
c-Myb in cell proliferation. While a pivotal role for c-Myb in
hematopoietic cell growth has been appreciated, few c-Myb target genes
are known which trigger or facilitate proliferation. For example, c-Myb
is known to trans-activate the cdc2 kinase gene
(52) but its role in trans-activating the DNA polymerase gene remains equivocal (69). The present report therefore implicates
topo II
, an enzyme essential for chromosomal segregation prior to
mitosis, as another gene target for the proliferative effects of c-Myb
at least in hematopoietic cells.
Conversely, the self-squelching property of c-Myb via its NRD presents
this factor as a rather attractive candidate for regulation of topo
II. Topo II
levels rarely vary by more than 2- or 3-fold in
proliferating cells in culture (70-72). In fact, attempts at overexpression of recombinant topo II
in a variety of hosts have been difficult and often require conditional expression systems (73).
Stringent control of normal, cellular topo II
expression therefore
logically requires a proliferation-dependent
transcriptional program that itself is either tightly regulated, or is
intrinsically self-regulating. The outcome of our experiments with
c-Myb begins to establish a causal relationship for this logic. How
this hypothesis might extend to topo II
regulation in
non-hematopoietic tissues is somewhat less clear. We have begun to
investigate the role of other Myb family members, like
B-myb, which are more widely expressed in different tissues
(61). But unlike c-Myb, the C terminus of B-Myb does not act in a
negative regulatory fashion (74). Instead, the B-Myb C terminus is
believed to bind co-activating proteins whose abundance appears to
limit its efficacy as a trans-activator (65). Based on this
finding, it is surprising that the trans-activating effect
of B-Myb did not appear to be limiting in transfection studies (Fig.
7). Nonetheless, B-myb expression is tightly regulated at
the G1/S boundary (63) and precedes the known increase in topo II
expression in G2/M (71, 72). The results in Fig. 8 make it tempting to speculate that the Myb family of transcription factors might play a global role in topo II
regulation, but
investigation of B-Myb effects in a more extensive panel of
non-hematopoietic cells is warranted.
We gratefully acknowledge the generous plasmid gifts of Drs. Edward Prochownik, Prem Reddy, Roger Watson, and Kathy Weston which enabled these studies. We also thank Drs. Tim Bender, Carlos Catalano, Scott Ness, and Linda Shapiro for their helpful discussions throughout the course of this work.