Volume 15 Number 10 1987
Nucleic Acids Research
Competition studies with repressors and activators of viral enhancer function in F9 mouse
embryonal carcinoma cells
M.J.Sleigh, T.J.Lockett, J.Kelly and D.Lewy
CSIRO Division of Molecular Biology, PO Box 184, North Ryde, NSW 2113, Australia
Received January 21, 1987; Revised and Accepted April 27, 1987
ABSTRACT
DNA competition studies have been used to Investigate the presence of
a repressor of viral enhancer function In F9 mouse embryonal carcinoma
cells. The complete polyoma virus enhancer region, cotransfected Into F9
cells with the SV40 promoter/enhancer attached to a chloramphenlcol acetyl
transferase marker gene, Induced a small Increase In pSV^CAT expression.
This can be explained by preferential but weak binding by polyoma
sequences of a molecule repressing pSV^CAT transcription. Repressor
activity substantially disappeared when the cells were Induced to
differentiate by retlnolc acid. Repressor binding was localised to one
half of the polyona enhancer, but was lost on further fragmentation of this
region. It appears that multiple sequence elements nay be required for
repressor binding and that these are at least partially separable from the
complement of elements binding enhancer activating molecules.
INTRODUCTION
Cells of the F9 mouse embryonal carcinoma (EC) line differentiate In
response to treatment with retlnolc acid, with or without additional cyclic
AMP (1). In nonolayer culture, product cells resemble extraembryonlc
parietal endodera, a tissue produced from multlpotent Inner cell mass cells
at about the time of Implantation during mouse embryo development In vivo
(2).
A characteristic of this differentiation step, both _ln vivo and In F9
cells ^n_ vitro, is the new synthesis of extracellular matrix and other
proteins.
At about the same time, the cells becone permissive for
expression of a range of vtrel genes.
This activation of expression of
genes from viruses Including SV40, polyoma, Holoney murlne leukemia virus,
and cytomegalovlrus has been used In several laboratories to Investigate
gene regulation mechanisms In differentiating F9 cells (3-10).
It Is now evident that the block to viral gene expression In F9 and
other EC stem cells Is at the transcription step, and Is due to partial or
complete lack of function of viral transcrlptlonal enhancer sequences
© IR L Press Limited, Oxford, England.
4307
Nucleic Acids Research
(11-15).
Some evidence suggests the presence In EC cells of one or more
molecules which repress enhancer activity.
An adenovirus Ela-llke activity
has been reported In undlfferentlated but not differentiated EC cells
(16).
A product of the adenovlrus Ela region Is known to repress
transcription from the SV40 early gene promoter, apparently via an
Interaction with the enhancer sequences (17,18).
More recently, Gorman eit
al., (14) reported that In undlfferentlated F° cells, the association of
enhancers with viral gene promoters caused a decrease In transcription
compared with that from the sane promoters lacking enhancers.
Tn addition,
they found stimulation of transcription from the murlne sarcoma virus
promoter/enhancer when the Rous sarcoma virus (RSV) long terminal repeat
sequence was present In the same cells, suggesting that the RSV sequence
was preferentially sequestering a factor In P) cells which was acting to
repress enhancer function.
We have previously reported that the SV40 enhancer retains a small
amount of activity In undlfferentlated EC cells.
Competition studies
suggest that In both differentiated and undlfferentlated F9 cells, the
supply of enhancer activating molecules limits the amount of transcription
seen fron the SV40 early promoter/enhancer In transient assays (15). Thus
a repressor molecule, If present, possibly acts to block the consequences
of activator nolecule binding, rather than the binding Itself.
Tills nay
suggest that activating and repressing molecules are not In direct
competition, but bind to different DNA sequences, as described for the
lnterferon gene enhancer (19). Alternatively, the repressor may represent
a modified but Inactive fora of the activator, In which case binding of
activator and repressor to the sane DNA sequences might be expected.
One approach to Investigating such questions has been through
competition studies carried out _in vitro (20) or _ln vivo (21,22).
Previous
results (14) have suggested that competing DNA sequences could be used to
derepress viral enhancer function In F9 cells, presumably via sequestration
of a DNA-blndlng repressor nolecule.
We have used a range of DNA sequences
competing In transfected F9 cells with a marker gene driven by the SV40
pronoter/enhancer region.
These studies have provided further evidence for
repression of viral enhancer function In undlfferentlated (but not
differentiated) F9 EC cells and have suggested that different complements
of DNA sequence elements may be involved In mediating the repression and
enhancement of viral gene transcription.
4308
Nucleic Acids Research
MATERIALS AND METHODS
Cell lines and culture methods
The F9 cell line used for this study (0TF963) (23) was obtained from
Dr. A. Levlne, and was maintained on gelatin-coated dishes at 37° In a
humidified atmosphere of 7.5Z OO2.
Differentiation of the cells was
Initiated by addition to the isedlum of 5 x 10~7M all-trans retlnoic acid
(Sigma Chemical Company).
The flbroblast line L~a~ which Is thymidine
klnase and aprt negative, was obtained from Columbia University via Dr. K.
Raphael.
This was maintained in Dulbecco's modified Eagle's aedium (DMEM)
supplemented with 10Z fetal calf serum and dlamlnopimellc acid (50|ig/ml).
All other lines were maintained In DHEM + 101 heat Inactivated fetal calf
serum.
Tranafectlon procedure
Cells were plated to reach a density of approximately 106 per 10cm
dish at the time of transfectlon.
Transfection of DNA Into cells was
carried out by the calcium phosphate method (24). Sixteen hours after
addition of the DNA precipitate to the plates, cells were washed twice in
PBS containing 0.5mM EGTA (to remove loosely attached DNA) and then once in
PBS before addition of fresh medium containing mycostatin.
Twenty eight
hours after the medium change, cells were washed twice In PBS, harvested by
scraping and collected by centrifugatlon.
The cell pellet was washed In
PBS and then resuspended In 0.2 ml of 0.25M Trls-HCl buffer, pH 7.8.
Cell
extracts were prepared by son lea t Ion and then centrlfuged for 15 tnin in the
cold In an Eppendorf microcentrifuge to remove cell debris (25).
Enzyme Assays
(5-galactosldase activity was measured in cell extracts as described
(26).
CAT activity was assayed In cell extracts that had been heated at
65°C for 10 mln and then centrlfuged for a further 5 min.
This treatment
removed an acetyl coenzyme A-degrading activity but did not affect the CAT
enryme.
Assays were carried out using a rapid, non-chromatographlc
procedure which measures transfer of
C-labelled acetyl groups from acetyl
coenzyme A to unlabelled chloramphenlcol (27).
DNA used for transfectlon
Plasmids were propagated in E.coll RR1.
DNA for transfectlon was
Isolated by alkaline lysis of cells, with separation of supercolled forms
by two rounds of cesium chlorlde/ethldium bromide centrifugatlon (28).
Enzymes used In constructions were obtained from New England Blolabs, and
were used according to the suppliers' recommended conditlona.
4309
Nucleic Acids Research
PSV2CAT (25) was obtained from Dr. E. Dennis.
The construction and
use of ptkgal (B-galactosldsse gene coding sequences transcribed from the
herpes virus thynldlne klnase promoter) has been described previously (15).
pSVEN (15) contained one complete and one partial copy of the SV40
enhancer cloned into pBR322.
pttoE contained a synthetic oligonucleotide
equivalent of one of the 72 base pair repeats constituting the enhancer of
Moloney murlne sarcoma virus (29). The fragment (the promoter proximal
repeat) was synthesiser) as two 45-oers with complementary overlapping
sequences at the centre of the fragment.
The two oligomers (kindly
syntheslsed by Dr. G. Both) were annealed and extended to form a complete
double stranded molecule (28). This was cleaved at an EcoRl site Included
in the aolecule next to MoMSV-rtertved sequences, and cloned into FcoRI and
Smal-cut pUC18.
plgE contained the lmmunoglohulIn heavy chain gene enhancer (30)
cloned Into Sell- and Hindlll-cleaved pUC18.
The 313bp Insert was derived
by Hindlll and Xhol digestion of pPylgE, kindly supplied by I>r. J. de
Vllllers.
Complete and partial polyoma enhancer sequences for cloning into pIIC18
were derived from a series of deletion mutants described elsewhere (31) and
kindly supplied by Dr. G. Veldtnan.
In Table 1 and Figure 3.
The conttructs prepared are described
An additional construct (pPyF12) containing three
copies of the insert present in pPyF2 was obtained by EcoRI-PvuII digestion
of pPyF7, blunt ending of the 495bp fragment using Klenow DNa polymerase
(28) and llgatlon into Snal-cut pUClR.
RESULTS
Design of competition experiments
The a in of this series of experiments was to determine whether
competing DNA sequences present In the same cell as a transcribing gene
would preferentially sequester activating or repressing molecules binding
to enhancer elements.
The effect observed would be a net decrease or
Increase, respectively, In expression of the transcribing gene.
The transcribing gene used Is that coding for the bacterial enzyme
chloranphenlcol acetyl transferase (CAT), with transcription driven by the
SV40 early promoter and enhancer region (pSV2CAT; 25). In addition to
PSV2CAT, tranafected cells received 5(ig ptkgal to act as an internal
control for efficiency of transfection.
Cells receiving large amounts of
DNA during transfection showed some depression In endogenous
4310
Nucleic Acids Research
Table 1
Derivation of cloned polyoaa enhancer fragments
Cloned fragment
la pOC18
Parent plasmlda
Digestion to
generate fragment
pPyFl
B5137
EcoRI + PvuII
PPyF2
pPyCAT
EcoRI + PvuII
PPyP3
B5183
EcoRI + PvuII
pPyFA
P5179dB
EcoRI + PvuII
pPyF6
pPyCAT
EcoRI + Haelll
PPyF7
pPyCAT
EcoRI + partial PvuII
PPyF9
PPyF2
Sau3A( blunt ended )b
+ EcoRI
pPyFlO
P5179dB
pPyFll
pPyF2
HpalK blunt ended)b
+ EcoRI
NaellKblunt ended)c
+ EcoRI
a
parent plasmlds (other than pPyF2 which was generated as
described above) had structures as outlined previously (31).
Sau3A and Hpall-cut ends were converted to blunt end cuts using
Klenow DNA polymerase I (28).
Haelll - cut ends were converted to blunt end cuts using anmg
bean nuclease (Boehrlnger Mannheim) according to the suppliers'
instructions.
b
c
p-galactosldase expression but this effect appeared to be Independent of
the nature of DNA Introduced (Henderson & Sleigh, unpublished results) .
Thus "mock transfected" samples for determination of background levels of
p-galactosldase received pUC18 DNA (23.3ug per plate) but not ptkgal or
pSV2CAT.
The third conponent of transfectlon was the competing enhancer
sequence cloned into pBR322 or pUC18.
In practice, DNA precipitates were
prepared containing Increasing amounts of the competing molecule,
progressively replacing an equivalent anount of pUC18 DNA.
Under the conditions of cell growth and transfectlon used In these
studies, maximal transient expression Is reached when approx. 10-20ug of
pSVjCAT DNA Is used to transfect differentiated or undlfferentlated F9
cells.
Availability of enhancer-binding factors Is a limiting factor for
transcription in both cell types.
At these levels, transcription factors
binding to the SV40 early promoter are still in excess In F9 cells but
approaching full utilisation In differentiated F9 cells (15). The
4311
Nucleic Acids Research
C o n p e t l t l o n between pSVjCAT and pPyF7
T»bU 2
Cell
line
DNA f o r *
transfectlon
pOC 18
L~a~
F9
F9+RA
a
b
c
d
0
CAT*
fealc
CAT/d
assays
assays
gal
266
PSV2CAT
pSV £AT+F7,1:1
pSV2CAT*-F7,l:2
pSViCA 1 W-F7,l:4
15,221
13,261
10,654
10,604
pUC 18
pSV2CAT
pSV£AT+F7,l:l
pSV2CATfr7,l:2
pSV2CAT4-F7,l:4
302
930
850
1,894
1,810
pOC 18
pSVjCAT
pSV^CAT+F7,l:l
pSV2CATfF7,l:2
pSV£AT+F7,l:4
35,226
23,654
31,512
32,265
298
-
0.227
0.440
0.580
0.510
0.740
71,460
35,552
37,646
20,670
0.138
0.158
0.167
0.175
0.165
31,400
18,896
51,189
67,037
0.240
0.344
0.296
0.352
0.368
338,711
422,392
281,357
252,070
-
_
Relative
expression 6
100
50
53
29
100
60
163
213
_
100
124
83
74
23.3 Mg DNA was used for transf act Ions, consisting of 5 \% pSVjCAT and
5pg ptlegal where appropriate, pPyF7 In the ratio shown to pSVjCAT, and
the balance pUC18 DNA.
cpo In u-acetylchloramphenlcol after 60 mln. assay of approximately
30|Jl cell lysate prepared as described In Methods. 0.01 units of
purified CAT enxyme (P.L.Blochenlcals) yields approx. 20,000 cpa under
these conditions.
P-galactosldase assays were carried out on O.lal allquots of cell
lysate, using overnight assays as described In Methods.
Calculated after subtraction of aock-transfected blanks.
Results In the presence of competitor calculated relative to those for
pSV2CAT alone.
conpetltlon experlnents described here were carried out at 5nft PSV2CAT per
10 cm plate of cells.
Tills was the minimal level that could be used to
provide CAT expression that could be measured reliably In F9 cell extracts
(15).
However It did Impose sone limitations on the ratio of competing to
transcribing molecules that could be achieved.
Using the methods of cell
culture and transfectlon described here, DNA uptake capacity appeared to be
United to a maximum of around 25(ig DNA per 10 cm dish of cells (Results
not shown).
For this reason maximum competitor to gene molecular ratios
used In Initial experiments have been 4:1. An additional construct
containing multiple copies of the most effective competitor fragment has
also been prepared, providing a competitor:gene ratio of 12:1. The
transcribing gene, PSV2CAT, contained multiple regulatory elenents within
two 72bp repeats in the enhancer region (32, 33) while the competing
4312
Nucleic Acids Research
wo»
7~-—.
'
• • ^
so-
o ~^~
^
wo-
o
"
*
~~- -
9O-
mo90MO
O
wo-
o
o
—*--•-,
80-
ig
1
2
3
4
C0MWIII0H B*TIO
Flg.l
Effect of coapetlag enhancer sequences on pSV jCAT expression. F9
cells (•), F9 cells treated with RA for 4 days at the tine of transfectlon
(o) or L~a" cells (•) were transfected with 5 « pSV *AT, 5 H? ptkgal and
13.3Mg pOC18 DNA, progressively replaced by pSVEN (SV), pPy77 (Py), pHoE
(Ho) or plgE (Ig). Two days later CAT and p^galactosldase activities In
the cells were measured as described In Methods. Results were calculated
as described for Table 2, and are expressed for each point relative to the
normalised CAT expression obtained froa cells receiving pSVjCAT and pOC18
DNA but no conpetlng enhancer.
The coapetltor ratio Is the number of conpetlng enhancer molecules
Introduced to the cells relative to each pair of 72bp repeat sequences on
PSV2CAT. Results shown are averaged from three Independent experiments.
molecules also contained several regulatory sequence elements (Veldoan et
al., 1985).
Thus the exact ratio of activator or repressor binding
sequences In competing compared with transcribing DNA In each competition
experiment Is difficult to define any more exactly.
Characteristics of the CAT Assay method
The assay method used here for CAT enzyme activity In cell extracts
(27) differs In several respects from the chromatographlc assay used
commonly (25). Because enzyme activity Inactivating acetylCoA Is removed
by heating cell extracts at 65° , the assay remains linear over several
4313
Nucleic Acids Research
hours (Henderson and Sleigh, unpublished results).
The specific activity
of the acetylCoA can be varied to Increase the sensitivity of assay of low
activity samples.
Thirdly, the assay quantltates reaction product rapidly
and directly, Increasing both the consistency of results snd the ease with
which multiple cell samples can be assayed and duplicate assays carried
out.
These characteristics have been very Important In this study, where
activities have often been low, and It has been necessary to carry out
repeat experiments and duplicate cell treatments within experiments to
confirm the significance of quite small variations In enzyme activity.
Effect of various competing enhancer sequences on transient expression of
pSV?CAT
Cloned enhancer seouences from the lmmunoglobulln heavy chain gene
(plgE), Moloney murlne sarcona virus, (pMoE) polyoma virus (pPyF7) and SVAO
(pSVE») were tested In competition experiments with pSV^CAT to determine
whether they would Influence Its transcription rate by net removal of
either enhancer-activating or enhancer-repressing molecules.
The results
from one such experiment, carried out In LtK", F9 and differentiated F9
cells, are shown In Table 2. Relative PSV2CAT expression In the three
Hoes before and after correction for p galactosldase expression remained
approximately constant from one experiment to another although absolute
values obtained varied 8 one what.
Each experiment reported was carried out
at least three tines to confirm the consistency of results obtained.
Figure 1 summarises the results from these multiple experiments for
the four different coupeting enhancers listed above.
Inhibition of CAT
expression was produced In all cell types by the enhancer from Moloney
sarcoma virus (MoSV) and by the SVAO enhancer Itself.
The lumunoglobulln
enhancer, which would not be expected to be active In any of these cell
types, reduced CAT expression In L cells, hut not In the other cell types.
The complete polyoma enhancer region depressed pSV2 CAT expression In L and
differentiated F9 cells, but In F9 stem cells, a small but reproducible
stimulation of PSV2CAT expression was seen (Figure 1). This result Is
consistent with preferential binding by the polyooa enhancer of a repressor
molecule present In undlfferentlated F9 cells, but not In their
differentiated progeny or In L cells.
Localisation of the repreasor-blndlng sequences within the polyoma virus
enhancer
The enhancer sequences required for polyoma virus replication and
transcription have been characterised In several laboratories (11,31,34).
4314
Nucleic Acids Research
1
2
COMPETITOR
3
*
RATIO
Fig. 2
Effect of conpetlng polyoma enhancer fragments on pSV JCAT expression.
Experiments were carried out and results calculated and expressed exactly
as described for Figure 1. Competing polyoma enhancer sequences were the
entire enhancer region (pPyF7) or the A and B regions (11; pPyF2 and pPyFl
respectively). (•) F9 cells, (O) F9 + RA cells, (•) ITf cells.
Results shown are the average of those obtained In three Independent
experlnents for each conpetlng fragaent.
Sequences contributing to enhancer function are spread over the region frota
nucleotldes 5020-5270 with contributions from several different sequence
elements within this region.
In flbroblasts the "A" region (bases
5020-5130) Is the most active (11). In PCC3 embryonal carcinoma cells, the
"A" region Is relatively Inactive, with a snail amount of residual enhancer
function In the "B" region (bases 5130-5270).
This suggests that a polyoma
enhancer repressing nolecule present In F9 cells nay bind predominantly to
the A region, with activating molecules binding to both A and B.
When the
cloned A and B regions were Introduced Into cells In conpetltlon with
PSV2CAT, all of the ability of the complete polyona enhancer to stimulate
PSV2CAT expression sppeared to reside In the A region (pPyF2 - Figure 2).
This Is consistent with binding by the polyoma A region of a molecule
repressing pSV^CAT expression In F9 but not In other cells.
To further localise activator- and repressor-blndlng sequences, other
fractions of the polyoma enhancer region (Figure 3) were tested as
4315
Nucleic Acids Research
8000
MOO
8330
8300
S2B2
Fig. 3
Structure of polyoaa virus enhancer fragments used In competition
studies. The derivation of the fragsents is described In Table I. All
fragments ended at an EcoRI site corresponding to nucleotlde A635 of
polyooa (a BanHI site in the original virus). Teralnl of other sequences
contained In the fragmenta are shown. Unnumbered termini are at the Bell
site (nucleotlde 5021) or the PvuII sites at 5130 or 5267. Nucleotlde
nunberlng Is as In reference 31.
Also marked on the parent polyoaa sequence Is the region for the start
of late transcription (bar and arrow) and the A and B regions of the
polyoma enhancer (11).
coupetltors with PSV2CAT.
These results (Figure 4) demonstrate that no
fragment smaller than the complete A region (pPyF2) stimulated pSVjCAT
expression.
However, a comparison of the results for undlfferentiated and
differentiated F9 cells suggests that pPyF6 and pPyFll may retain some
ability to bind repressor molecules.
Several subsections of the A region
were effective In binding activator molecules, as evidenced by reduced
PSV2CAT expression.
cAHP requirement for disappearance of the repressor effect
Figure 5 summarises the competition results obtained for each fragment
tested, at the msxlmun level of competitor used.
This demonstrates the
relative effect of competing molecules in the 3 different cell types.
Interestingly, the reductions in PSV2CAT expression induced by molecules
which have net repressor-bindlng activity in F9 stem cells (pPyF7 and
pPyF2) were not as great In differentiated F9 cells as those seen In L
4316
Nucleic Acids Research
TOO
g
SO
-—o
— — —o
80
•
PyFtl
PyfO
r
\u
ESS ON -
it!
f
X
f.
100'
too
so-
SO
6O
60-
100
100-
so
80
6O
6O
= zz—•
4O
PyfW
4
1
COMPETITOfl RATIO
ffect of subtractions of the polyona virus enhancer on tranalent
expression of PSV2CAT. Experiaenta were carried out and results calculated
and expressed exactly as described for Figure 1. Sequences contained
within competing molecules are described In Figure 3. Results shown are
the average of those obtained In 2-4 Independent experiments for each
competing fragment. (•) F9 cells, (O) F9 + RA cella, (•) L"a" celle.
cells.
However some subsections of the A region which showed no net
repressor binding In F9 stem cells produced a greater reduction In pSV^CAT
expression In the differentiated cells than that seen In L cells.
One
explanation for this Is that the disappearance of the enhancer-binding
repressor is not complete even after 6 days exposure of F9 cells to
retlnolc acid.
It has been suggested (2, 35, 36) that treatment of F9 cells with cAMP
In addition to RA Is required for full differentiation to parietal
endodern-type cells.
In previous experiments on SV40 enhancer function, we
found that enhancer activation was similar in F9 cella treated with RA
alone (at 5xlO"7M) , or RA + lmM cAHP (15).
In competition experiments using pPyF2 and pPyF7, Inhibition of
PSV2CAT expression was similar In F9 cells treated only with RA (5xlO~7M),
or with RA (5x10-7M) + dibutyryl cAMP (lmM) (results not shown).
Thus
additional treatment with cAMP does not appear to lead to more complete
disappearance of the repressor effect than In cells treated with 5x10" H RA
alone.
4317
Nucleic Acids Research
Bdl
PvuP
PWJD
F9
I1
•26
-25
26
26
•26
I 1 11
•
-26
26
L
1"
I
26
•28-25-
28
25
l1 I
11 1
•2*
-25
26
1 11
25
•28-25-
1
1
Pig. 5
Effects of competing molecules on transient expression of pSVjCAT.
The figure compare* polyoma sequences contained In coapetlng molecules on
the left hand side with the percent change In pSV£AT expression seen at
the maxlnal level of competitor tested (4 competing molecules to each
pSVgCAT solecule). For each competitor, bars above the baseline represent
stimulation of pSVjCAT expression. Those below the base line show
Inhibition. For each competing fragment, effectiveness as an enhancer In a
DNA replication assay (31) It shown, relative to wild type (pPyF7 - 1001).
Effect of higher competitor:gene ratios on pSV^CAT expression
To determine whether higher levels of the polyoma "A" region could
further stimulate PSV2CAT expression In F9 cells, pPyF12, containing three
tandem copies of this region was prepared and tested In competition
experiments.
The results are shown In Figure 6.
Tills experiment confirmed
stimulation of PSV2CAT expression In F9 but not F9+RA cells at low (3-4:1)
competitor:gene ratios.
However, as the amount of competitor was
Increased, suppression of pSV2CAT expression was observed.
This can be
explained by simultaneous removal of enhancer activating molecules by the
4318
Nucleic Acids Research
F9+RA
PyF12
3
6
9
12
COMPETITOR RATIO
Fig. 6
Competition between high levels of polyoma enhancer region "A" and
PSV2CAT In transient assays.
Fxperiaents were carried out aDd results calculated as described for
Figure I. The competing plasmld, pPyF12, consisted of pUC18 containing
three head-to-tall copies of the "A" region Insert from pPyF2. ( ) F9
cells, ( ) F9+BA cells. Results shown are the average of those from two
transfectlon experiments.
competing plasmld.
This results In a rather narrow "window" In which
removal of repressor predominates In determining levels of PSV2CAT
expression.
DISCUSSION
Enhancer regions fron both viral and cellular genes have been shown to
consist of multiple sequence elements which may have Independent,
overlapping or equivalent functions In regulation of gene activity In
different cell types (32, 33; reviewed In 37). Specificity of enhancer
action may he nodulated by changing complements of sequence elements In
different enhancers (33). In addition, different cell types appear to vary
In the range and/or activity of enhancer-binding molecules (presumably
proteins) that they possess.
Such molecules nay exert a positive or
negative effect on expression of associated genes (for example, reference
22).
Results reported here, and previously (14, 38) suggest that In F9
embryonal carcinoma cells, negative regulatory factors binding to specific,
enhancer-associated DNA sequence elements Influence the level at which
viral genes are expressed.
After cell differentiation, this repressive
effect Is substantially reduced.
If disappearance of repressor molecules were the principal cause of
enhancer activation during cell differentiation, then removal of the
4319
Nucleic Acids Research
repressor by binding to competing DNA in undlfferentlated cells should
increase the enhancer effect to a level approaching that seen in
differentiated cells.
As seen In Figure 5, no such Increase was achieved.
The maxlnum stimulation of PSV2CAT expression by coapetlng DNA was 50-801,
while full restoration of enhancer sctlvlty should Increase expression to
at least ten times the level seen in undifferentlated cells (15).
In part this could be due to limitations Imposed by the experimental
design.
As demonstrated previously (15), enhancer activating nolecules
become Uniting for transcription when 10-15ug PSV2CAT Is transfected into
differentiated or undifferentiated F9 cells.
Since the competing polyona
DNA molecules that bound repressor also bound activators, as seen for the
results for Va~
and differentiated F9 cells, then the msxleum stimulation
of transcription could be Halted to two- to three-fold.
That this level was not always achieved may reflect the relative
affinities of enhancer elements and regions for DNA binding molecules.
It
has previously been reported that the SVAO and polyoma enhancer regions do
not compete strongly with each other for enhancer activating factors (17,
20, 39) and this Is confirmed by the relatively poor competitive effects
reported here.
SVAO, polyona and other viral enhancer regions may also
sequester different subsets of the total coaplement of factors (activators
or repressers) available in a particular cell type (20, 32, 33, 40). Thus,
despite the strong suppression of the HoMSV enhancer In F9 cells (13, 14),
the MoMSV enhancer segment described here did not show net repressor
binding In competition with pSVjCAT.
On the other hand, up to 75-fold
stimulation of expression from the MoMSV promoter/enhancer by RSV sequences
has been described (14), suggesting that these seauences have a very high
affinity for the repressor molecule(s) blocVlng MoMSV expression in F9
cells.
Since the polyoma sequences described in this report are themselves
strongly suppressed in F9 cells, but appear to have low affinity for the
SVAO enhancer repressor, the possibility arises that Bore than one
repressing factor Is present.
If so, then the polyoma sequences shown here
to be Important In binding the SV40 repressor may be different from those
of primary importance for repression of the function of the polyoma
enhancer Itself In EC cells.
An Important result froa the competition experiments of Figures 2, 4
and 5 Is that binding sequences for repressing and activating molecules are
partially separable (compare results for pPyF2 and pPyF4).
4320
Deletion of
Nucleic Acids Research
nucleotldes 5113-5130 from pPyF2, the most effective repressor binder ID F9
cells, produced pPyFll, which showed less net repressor binding In F9 cells
as well as a small decrease ID net activator binding In differentiated
cells.
This suggests that not only Is this sequence element very Important
for polyoma enhancer function, as suggested previously (31), but It also
contributes to the binding of the SV40 repressor.
However Its Interaction
with other sequence elements Is clearly different for the two functions.
In the absence of other "A" region sequences this element failed to bind
repressor nolecules but was an effective binder of activator molecules and
produced high enhancer activity (pPyF4 and pPyFlO of Figure 5 and reference
31).
Evidently the 5113-5130 element In coablnatlon with other elements of
the "A" region Is required for binding of the SV40 repressor, while the
same sequences In combination with the "B" region are optimal for activator
binding and enhancer activity.
Two sequence elements homologous to the SV40 enhancer are found In the
"A" region.
These lie between nucleotldes 5110-5120 and 5030-5040 of
polyoma (31), the elements being related to GT element II and Sph element
II (33) of the SV40 enhancer respectively.
A requirement for both of these
elements to effect repressor binding by polyoma sequences would be
consistent with the results described here.
Differential Interactions of repressors and activators with DNA
elements may have significant Implications for attempts to Isolate
activating and repressing molecules using In vitro DNA binding methods.
However they do not clarify the relationship between repressors and
activators, and how this changes as F9 cells differentiate.
The results
may be explained either by distinct molecules Interacting either directly
or Indirectly with different sequences In the enhancer, or by related
molecules binding to the same sequence but Interacting differently with
neighbouring sequence elements.
Studies on binding of nuclear proteins to
polyoroa enhancer sequences have not yet clarified the situation.
No
differences In binding patterns have been reported when extracts from
differentiated and undlfferentlated F9 cells are compared (41).
Several laboratories have Isolated mutants of polyoma virus which are
able to grow In undlfferentlated EC cells (reviewed In 42). The
outstanding feature of these mutants, and of revertants of Ineffective SV40
enhancer mutants assayed In other cell types (32), was that enhancer
activity was restored either by duplication of existing sequence elements
or by rearrangement (SV40) or mutation (polyoma) to create new ones.
An
4321
Nucleic Acids Research
additional feature of the polyoma nutants, as pointed out previously (43),
Is that the different types of sequence elements making up the enhancer
region have a different relative arrangement In the mutants compared with
wild type virus.
A particularly Interesting example la the F9.1 mutant,
where a point nutation creates a new enhancer core sequence in the B region
of the polyoma enhancer.
This mutant enhancer Is more effective in all
cell types and shows no evidence of repression In EC cells (12). It is
also resistant to repression by the adenovlrus ELa gene product (38). A
staple explanation Is that for polyoma virus the sequences around this
point mutation are required for repressor binding, but that alone would not
explain the more general Increase in effectiveness of this enhancer.
An
alternative possibility la that not only has a new element important for
binding activator ooleculea been created (explaining the general Increase
in effectiveness) but also that this la now differently placed with respect
to sequences Involved In enhancer repression In EC cells (nucleotides
5020-5130 from our experiments) thus reducing the effectiveness of bound
repressor.
The results reported here suggest some additional strategies for
generating enhancers that would escape binding by the SV40 repreBSor In EC
cells.
In particular, removal of nucleotldes 5040-5100 of polyoma should
remove repressor binding.
Such a deleted enhancer retains activator
binding activity (see pPyF4 and pPyFlO in Figure 5) and had undlmlnished
enhancer activity in a DNA replication assay (31). One report describes
the creation of a hybrid enhancer from the orlgln-dlstal half of the SV40
enhancer and nucleotldes 5090-5137 of polyoma (33). This enhancer, which
our results would predict should largely lacV repressor binding sequences
In the polyoma segaent, was functional In both EC and HeLa cells.
The
naturally occurring polyoma nutant PyFLC78 similarly lacks sequences from
the Bell site to nucleotlde 5096 (see Figure 3) in the A region of the
polyona enhancer.
This mutant had high enhancer activity in 3T3 cells and
was less suppressed than the wild type virus In PCC3 EC cells (11).
However in this virus there were also other rearrangements, deletions and
mutations, presumably to restore signals needed for Initiation of late
transcription, which normally occurs within the deleted region.
If slmilsr
repressor molecules act on both SV40 and polyoma virus enhancers in F9
cells, then an unrearranged equivalent of this enhancer region, i.e.
containing nucleotldes 5110-5270, should have equal activity in
4322
Nucleic Acids Research
differentiated and undlfferentlated F9 c e l l s .
Experiments to test this
hypothesis are in progress.
ACKNOWLEDGEMENTS
The authors a r e g r a t e f u l t o D r s . G. Veldroan, J . de V i l l l e r s and P.
Kamen f o r h e l p f u l d i s c u s s i o n s and supply of polyoma DNA, t o Dr. G. Both f o r
s y n t h e s i s of o l i g o n u c l e o t i d e a and t o D r s . P. Jennings and P. Molloy for
c r i t i c a l review o f the manuscript.
The work has been supported in part by
a grant t o MJS from the Australian Meat Research Committee.
REFERENCES
1.
S t r i c k l a n d , S . , and Mahdavl, V. (1978) C e l l , J5_, 393-403.
2.
Hogan, B.L.M., Barlow, D.P. and T i l l y , R. ( 1 9 8 3 ) . Cancer S u r v . , 2,
115-140.
3.
K e l l y , F. and Boccara, M. (1976) Nature, 262, 4 0 9 - 4 1 1 .
4.
T e l c h , N.H., Weiss, R . A . , Martin, G.R. andTowy, D.R. (1977) C e l l , 1 2 ,
973-982.
5.
Swartzenruber, D. F . , F r i e d r l c h , T.D. and Lehnan, J . H . (1977) J. C e l l
Physlol., 9 3 , 25-30.
6.
Oldstone, M.B.A., Tlshon, A . , Dutko, F . J . , Kennedy, S . I . T . , Holland,
J . J . and Lanpert, P.W. (1980) J . V i r o l . , J34, 256-265.
7.
Fujimura, F . K . , S l l b e r t , P . E . , Eckhart, W. and Llnney, E. (1981)
J . V i r o l . , 39, 306-312.
8.
Gautsch, JTO. and Wilson, M. (1983) Nature, 301_, 32-37.
9.
Huebner, K., Llnnenbach, A . , Ghosh, P . K . , ar-Rushdl, A . , Romanczuk,
H., Tsuchlda, N. and Croce, CM. (1983) In S i l v e r , L . H . , Martin,
G.R. and S t r i c k l a n d , S. ( e d s . ) , Teratocarcinoma Stea C e l l s , Cold
Spring Harbor Laboratory P r e s s , NY, pp.343-361.
10.
Niwa, 0 . , Yokota, Y . , I s h i d a , H. and Sugahara, T. (1983) C e l l , 22_,
1105-1113.
11.
Herboael, P . , Bourachot, B, and Yanlv, M. (1984) C e l l , 2 £ . 6 5 3 - 6 6 2 .
12.
Linney, E . , and Donerly, S. (1983) C e l l , ^ 5 , 693-699.
13.
Llnney, E . , D a v i s , B . , Overhauser, J . , Chao, E. and Fan, H. ( 1 9 8 4 ) .
Nature, 3 0 8 , 470-472.
14.
Gorman, C M . , Rlgby, P.W.J. and Lane, D.P. (1985) C e l l , 4 2 , 5 1 9 - 5 2 6 .
15.
S l e i g h , M.J. and L o c k e t t , T.J. ( 1 9 8 5 ) . Embo J . , 4, 3831Z3~837.
16.
I n p e r l a l e , M . J . , Kao, H . T . , Feldman, L . T . , Nevlns, J . R . and
S t r i c k l a n d , S. (1984) Mol. C e l l . B l o l . , 4 , 867-874.
17.
B o r r e l l l , E . , Hen, R. and Chambon, P. (19^4) Nature, 2 H t 6 0 8 - 6 1 2 .
18.
V e l c i c h , A. and Z i f f , E. (1985) C e l l , 4 0 , 705-716.
19.
Goodbourn, S . , B u r s t e l n , H. and M a n l a t l s , T. (1986) C e l l , £ 5 ,
601-610.
20.
Sassone-Corsl, P . , Wlldeman, A . , and Chambon, P. (1985) Nature, 3 1 3 ,
458-463.
21.
S c h o l e r , H.R. and Gruss, P. (1984) C e l l , 2 i » 4 0 3 - 4 1 1 .
22.
S c h o l e r , H.R. and Gruss, P. (1985) Embo J. , i, 3005-3013.
23.
Rosenstraus, M.J. and Levlne, A. ( 1 9 7 9 ) . C e l l , l]_, 3 3 7 - 3 4 6 .
24.
Wlgler, M., S l l v e r s t e l n , S . , L e e , L . S . , P e l l i c e r , A . , Cheng, Y . C and
Axel, R. ( 1 9 7 7 ) . C e l l , VI, 223-232.
25.
Gorman, C M . , Moffat, L . F . and Howard, B.H. (1982) Mol.Cell B l o l . , 2,
1044-1051.
4323
Nucleic Acids Research
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
4324
M i l l e r , J.H. (1972) Experiments In Molecular G e n e t i c s , published by
Cold Spring Harbor Laboratory P r e s s , NY.
S l e i g h , M.J. ( 1 9 8 6 ) . Anal. Blochetn. , 156, 251-256.
M a n l a t l s , T . , F t l t s c h , F.F. and SanbrooT7 J. (1982)
Molecular
(Honing. A Laboratory Manual, published by Cold Spring Harbor
Laboratory P r e s s , NY.
L a l n l n s , L . A . , Gruss, P . , P o z z a t t l , R. and Khoury, G. ( 1 9 8 4 ) .
J . V i r o l . , 4 9 , 183-18<».
B a n e r j l , J. , Olson, L. and Schaffner, V. (1983) C e l l , 33^, 729-740.
Veldnan, G., Lupton, S. and Pamen, R. ( 1 9 8 5 ) . Mol. Cell B l o l . , 5,
649-658.
Herr, V. and Clarke, J. ( 1 9 8 6 ) . C e l l , 45_, 4 6 1 - 4 7 0 .
Zenke, M., Grundstron, T., Hatthes, H., W l n t z e r l t h , M., S c h a t i , C.,
Wlldeman, A. and Chambon, P. ( 1 9 8 6 ) . Enbo J . , 2» 387-397.
Muller, W . J . , Mueller, C.R., Mes, A.M. and H a z z e l l . J.A. ( 1 9 8 3 ) . J.
V i r o l . , 4 7 , 586-599
Wang, S . - Y . , LaRosa, G.J. and Gudas, L.J. (1984) Dev. B l o l . , 107,
75-86.
Marottl, K.R., Brown, G.D. and S t r i c k l a n d , S. (1985) D e v . B l o l . , 108,
26-31.
S e r f l l n g , E . , J a s l n , M. and Schaffner, W. (1985) Trends Genet., 1,
224-230.
Hen, R. , B o r r e l l l , E . , Fromental, C. S a s s o n e - C o r s l , P. and Chambon,
P. ( 1 9 8 6 ) . Nature, 321, 2 4 9 - 2 5 1 .
P l e t t e , J . , K r y s z t e , M.H. and Yanlv, M. ( 1 9 8 5 ) . Embo J . , 4 ,
2675-2685.
Wlldeman, A.G., Sassone-Corsl, P . , Grundstron, T. , Zenke, M. and
Chambon, P. ( 1 9 8 4 ) . Fmbo J . , ^ . 3129-3133.
Ostapchuk, P . , D l f f l e y , J. F.X., Bruder, JK.T., S t l l l m a n , B . , Levlne,
A.J. and Hearing, P. (1986) Proc. N a t l . Acad. S c l . USA 8 3 , 8550-8554.
Amatl, P. (1985) C e l l , 43, 561-562.
Melln, P . , Plnon, H., R e l s s , C , Kress, C., Montreau, N. and Blangy,
D. (1985) Enbo J. , ^ , 1799-1803.