ABSTRACT
In
Saccharomyces cerevisiae
commitment to cell division occurs late in the G1 phase of the cell cycle at a
point called Start and requires the activity of the Cdc28 protein kinase and
its associated G1 cyclins. The Swi4,6-dependent cell cycle box binding factor, SBF, is important for maximal
expression of the G1 cyclin and
HO
endonuclease genes at Start. The cell cycle regulation of these genes is
modulated through an upstream regulatory element termed the SCB (Swi4,6-dependent cell cycle box, CACGAAA), which is dependent on both
SWI4
and
SWI6
. Although binding of Swi4 and Swi6 to SCB sequences has been well characterized
in
vitro
, the binding of SBF
in vivo
has not been examined. We used
in vivo
dimethyl sulfate footprinting to examine the occupancy of SCB sequences
throughout the cell cycle. We found that binding to SCB sequences occurred in
the G1 phase of the cell cycle and was greatly reduced in G2. In the absence of either Swi4 or Swi6, SCB sequences were not occupied at any cell
cycle stage. These results suggest that the G1-specific expression of SCB-dependent genes is regulated at the level of DNA binding
in vivo
.
Passage through Start in
Saccharomyces cerevisiae
requires the activity of the Cdc28 kinase and the G1 cyclins, Cln1, Cln2 and
Cln3 (reviewed in
1
).
CLN1
and
CLN2
are periodically expressed and their protein levels also peak at Start,
concurrent with maximal Cdc28 kinase activity (
2
,
3
).
CLN3
, however, is expressed at low levels throughout the cell cycle and Cln3-Cdc28 kinase activity is not significantly periodic (
2
). Two additional yeast G1 cyclins,
PCL1
and
PCL2
(formerly called
HCS26
and
ORFD
respectively) are also maximally expressed in G1 (
2
,
4
-
6
). Recent biochemical and genetic data suggest that Pcl1 and Pcl2 complex with
another cyclin-dependent kinase, Pho85, to promote cell cycle progression (
6
,
7
).
Maximal expression of the G1 cyclins,
CLN1
,
CLN2
,
PCL1
and
PCL2
, at Start requires the activity of a transcription factor, SBF (SCB binding
factor), which binds the repeated upstream regulatory sequence CACGAAA (SCB,
Swi4,6-dependent cell cycle box) (reviewed in
8
). Multiple copies of this sequence are sufficient to confer cell cycle-regulated transcription upon a reporter gene (
9
,
10
). SBF is composed of at least two proteins, Swi4 and Swi6 (
4
,
11
-
13
). Swi4 specifically binds the SCB sequence, whereas Swi6 binds indirectly via
its interaction with the C-terminus of Swi4 (
14
-
16
). In the absence of Swi4 or Swi6 or if the SCB sequences are deleted,
transcription of
HO
is abolished (
11
,
17
,
18
) and that of the G1 cyclins is greatly reduced (
4
,
19
-
22
). More recently, detailed analysis of
CLN2
transcriptional activation has suggested that other non-SCB consensus elements, which may also depend on
SWI4
, are involved in the periodic transcriptional regulation of
CLN2
(
21
,
22
).
The binding of SBF to SCB sequences has been well characterized
in vitro
(
4
,
11
,
13
,
15
). The upstream regulatory sequence, URS2, of the
HO
endonuclease gene contains 10 SCB elements (
18
). In gel mobility shift assays, promoter sequences from URS2 support SBF
complex formation in extracts prepared from cells throughout the cell cycle
(the `L' or `lower complex') (
13
; B.J.Andrews, unpublished results). In addition, a minor cell cycle-regulated complex of slightly slower mobility forms on URS2 probes and is
Cdc28 dependent (the `U' or `upper complex') (
13
). Since the appearance of this upper complex occurs slightly after the
transcriptional induction of
HO
and persists following
HO
repression after Start, its significance with respect to
HO
transcriptional activation is not clear (
13
).
A related transcription factor, the
Mlu
I cell cycle box binding factor (MBF, also called DSC1), is composed of Swi6 and
the Swi4 homolog Mbp1 (
20
,
23
,
24
). MBF binds an upstream regulatory element (MCB, consensus ACGCGTNA) which is
present in the promoters of several genes whose expression is induced at Start,
including
SWI4
, the S phase cyclins
CLB5
and
CLB6
and other genes involved in DNA synthesis (
23
-
28
). The MCB element can also direct Start-specific transcription of a heterologous gene and periodic binding of MBF
to MCB elements has been observed
in vitro
(
25
). In
Schizosaccharomyces pombe
the MBF-like complex, DSC1, also binds its target sequence in a cell cycle-regulated manner
in vitro
(
29
,
30
). So far the only
in vivo
observation to suggest that MBF and SBF DNA binding may be cell cycle-regulated is the subcellular localization of the Swi6 protein; Swi6 is
nuclear throughout G1 phase and is predominantly cytoplasmic from late G1 until
late M phase (
13
,
48
).
One hypothesis consistent with these data is that the SBF and MBF transcription
factors modulate the cell cycle-specific expression of their target genes through periodic changes in
their binding activity (
8
,
31
,
32
). In order to examine this question directly we used
in vivo
dimethyl sulfate (DMS) footprinting to assay for SBF binding throughout the
cell cycle. Our
in vivo
data support the notion that temporal regulation of SBF binding is involved in the G1-specific transcriptional induction of SCB-containing genes.
The following yeast strains were used for footprinting experiments: KN699
(=W303a;
2
),
MAT
[alpha]
ade2-1 trp1-1 leu2-3,112 his3-11,15 ura3 can1-100
[
psi
+
] (Figs
2
and
6
); BY105,
MAT
[alpha]
trp
[Delta]
63 ura3-52 lys2-801
a
ade2-107
o
his3
[Delta]
200 leu2-
[Delta]
1
(isogenic to JO14;
4
; Fig.
3
); BY210,
MAT
[alpha]
swi4
[Delta]
HIS3
, otherwise isogenic to BY105 (Figs
4
and
6
); BY211,
MAT
[alpha]
swi6
::
HIS3
, otherwise isogenic to BY105 (Figs
5
and
6
). The SCB::
lacZ
plasmid, pBA259, was constructed using synthetic SCB oligonucleotides as
described (
11
). The plasmid carries five copies of the SCB consensus sequence upstream of a
CYC1
::
lacZ
reporter gene in vector YEp24 (Fig.
1
). Yeast strains were transformed using standard techniques (
33
).
Cells were grown to a density of ~1 * 10
7
cells/ml in complete synthetic medium lacking uracil (SC-URA) (
34
). Except where noted, galactose was used as a carbon source and cells were
grown and elutriated at room temperature. Elutriations were performed as
previously described, with minor modifications (
2
). For loading onto the centrifugal elutriator ~2.0-4.0 * 10
10
cells were pelleted and resuspended in ~200 ml medium (for flocculant strains, such as
swi6
deletion strains, 50 mM EDTA was included in the loading buffer). Cells were
loaded at 18 ml/min into a Beckman JC-MI centrifugal elutriator (rotor type JE-5.0) at 2400 r.p.m., pre-equilibrated with the appropriate medium (for most experiments
the `conditioned' medium from the supernatant taken from the pelleted cells was
used). Once cells had equilibrated within the rotor the pump speed was
gradually increased by 2 ml/min every 3-5 min. The eluate was collected in 200-400 ml increments and cell size and budding were monitored
microscopically and using a Coulter channelizer. Typically the first wild-type daughter cells eluted at 30-36 ml/min, whereas for the
swi4
and
swi6
deleted cells (unbudded cell size greater) the first elution peak occured at >42 ml/min. In particular,
for the
swi6
[Delta] cells elution required both an increase in pump speed to 46 ml/min and a
decrease in rotor speed to 1800 r.p.m. A maximum of 10% of the load volume
could be recovered as a homogeneous population of unbudded cells. For the wild-type elutriations a total of five elutriation experiments were performed,
using two different strain backgrounds (Figs
2
and
3
and data not shown); for the
swi4
[Delta] and
swi6
[Delta] strains a total of two elutriations for each strain were performed,
using similar genetic backgrounds (Figs
4
and
5
and data not shown).
In vivo
DMS footprinting was performed using a modification of published protocols (
37
,
38
). For each sample ~1 * 10
9
cells were divided in half and pelleted. Half of the sample was rinsed once in
1 M sorbitol and frozen at -20oC for subsequent DNA extraction and
in vitro
DMS modification. The remaining pellet was resuspended in 1 ml medium (SC-URA with either galactose or raffinose as appropriate). Ten microliters of
DMS were added, mixed well and incubated for 5 min at room temperature. To stop
the reaction 50 ml cold distilled water was added and the cells were
immediately pelleted at 4oC, rinsed with 5 ml 1 M sorbitol and frozen at -20oC.
DNA was prepared from the frozen cell pellet by DTAB extraction as previously
described (
39
). Following resuspension of the purified DNA the unmodified DNA samples were subjected to
in vitro
DMS modification as described (
40
).
For PCR amplification of the DMS-modified genomic DNA, a primer corresponding to the 3'-end of the
URA3
gene in the SCB::
lacZ
plasmid was used: URA3PCR, 5'-ATTTGAGAAGATGCGGCCAGC-3'. DNA was first quantitated by measuring A
260
and ~20 [mu]g/sample were digested with
Apa
I (to facilitate primer annealing). To improve loading differences between samples, DNA digestion
and quantitation were checked by agarose gel electrophoresis prior to PCR amplification. PCR amplification was
carried out under the following conditions: 50 mM KCl, 10 mM Tris-HCl, pH 9.0 (at 25oC), 1% w/v Triton X-100, 0.2 mM each of dATP, TTP, dGTP, dCTP, 0.5 U Taq
polymerase (Promega), ~1 pmol
32
P-5'-end-labeled primer (
37
,
38
). Amplification was carried out for 18 cycles (1 min at 94oC, 2 min at 56oC, 2 min at 76oC per cycle). The products were phenol/chloroform extracted,
ethanol precipitated and resolved on an 8% w/v denaturing acrylamide gel.
Purified SCB::
lacZ
plasmid was also sequenced using the
32
P-end-labeled URA3PCR primer and was run alongside the genomic DNA samples
as a marker. The gel was dried and exposed to Kodak XAR-5 film at -70oC or exposed to a Molecular Dynamics PhosphorImager screen.
Gels were exposed on a Molecular Dynamics screen and scanned using a Molecular
Dynamics PhosphorImager and Imagequant (version 3.1) software. To obtain the
graphical representation of the footprints shown in Figures
2
,
4
,
4
and
6
a line was drawn vertically through the center of the lane and the intensity at
each point on the line was integrated by percent area to generate a linear plot
of band intensities. To compare the pattern obtained by
in vitro
DMS modification with the pattern obtained by
in vivo
DMS modification samples were normalized for slight differences in DNA
quantitation, PCR amplification and gel loading. First, a `background' line was
drawn between the sample lanes (where no DNA was loaded) and the values at each
point on this line were subtracted from the values at each equivalent point on
the sample line. Second, a band was chosen near the SCB sequences that did not
appear to be protected by DMS modification in any experiments and was assigned
an arbitrary value of 1000. For example, the raw phosphorimager values for each
peak normalized to 1000 in the experiment shown in Figure
2
D are as follows (each pair listed as
in vitro
and
in vivo
DMS-modified respectively): log sample, 577.66, 618.74; sample 1, 367.39,
236.98; sample 2, 432.29, 349.95; sample 3, 208.44, 291.14; sample 4, 251.15,
564.88. The phosphorimager plots shown in Figure
6
are taken from the same data as shown in Figures
2
,
4
and
5
. Only the bottom two SCB sequences are shown for clarity. No consistent
protection of 5' or 3' flanking sequences was apparent in the phosphorimager plots (data
not shown). Several other bands, outside the SCB sequences, were also tested as
normalization standards and gave similar results (data not shown). The band
closest to the SCB sequences was chosen for the final presentation of the data
for the following reason. Since Taq polymerase stops at modified purines, there
will be a gradient of more intense radioactive bands from the bottom to the top
of the gel (the more modified the DNA, the shorter the products). We found that
this `gradient' differed slightly between the
in vitro
and
in vivo
DMS-modified samples (data not shown). Therefore, the farther the
normalization band is from the region of interest, the greater the skewing of
results toward either the
in vitro
or
in vivo
pattern. This sample-to-sample variation can occur between any modified sample, thus we also
chose the SCB-proximal band to compare the
in vivo
DMS patterns shown in Figure
6
.
In order to examine binding to SCB sequences
in vivo
, a high copy number plasmid containing five SCB consensus sites inserted
upstream of
lacZ
was introduced into yeast cells (Fig.
1
). Like chromosomal SBF targets, the tandem SCB elements serve as an orientation-independent enhancer to confer cell cycle regulation of
lacZ
transcription, with maximal expression at Start (
9
,
10
). Expression of
lacZ
absolutely depends on the SCB promoter element and is not transcribed in the
absence of Swi4 or Swi6 (
11
,
41
). Thus high copy expression of the SCB::
lacZ
gene parallels expression of a genomic SCB-dependent promoter, such as
HO
, in both its regulation and transcriptional induction.
Dimethyl sulfate was used for
in vivo
footprinting in order to minimize the perturbation of cells, since treatment
times are short (4-5 min) and can be carried out on intact cells at room temperature.
Methylated purines were visualized after one-sided PCR amplification using Taq polymerase (
37
). As a control for the DMS modification pattern in the absence of bound
proteins, DNA was isolated from each sample and then subjected to DMS modification
in vitro
, followed by PCR amplification (see Materials and Methods).
A wild-type yeast strain carrying the SCB::
lacZ
plasmid was examined for SCB footprinting during exponential growth. DMS
modification of an asynchronous population showed weak protection of the SCB
sequences compared with the same DNA sample after deproteinization and
in vitro
DMS modification (Fig.
2
A, lanes 1 and 2, and data not shown). In particular, slight protection was
observed over the dA residues on the purine-rich strand (5'-CACGAAA-3') of each SCB sequence (Fig.
2
A). There was no significant protection of the pyrimidine-rich strand
in vivo
(5'-GTGCTTT-3'; Fig.
2
A, lane 2, and data not shown). The weak protection over the dA residues in the
SCB sequences was confirmed by phosphorimager analysis of the
in vivo
DMS-treated sample compared with
in vitro
modified DNA from the same sample (Fig.
2
D, `log').
Since asynchronous cultures contain cells in both G1 and G2 (by FACS analysis;
Fig.
2
C), we sought to enrich the SCB footprint by analysis of a homogeneous G1 cell
population. G1 daughter cells were obtained by centrifugal elutriation and re-inoculated into conditioned medium. This method minimizes any perturbation
of physiological state, since it does not involve temperature shift, media shift or drug addition. Progression of the re-inoculated culture was monitored throughout the cell cycle by cell size, FACS,
percentage budded cells and Northern analysis (Fig.
2
B and C).
In contrast to the asynchronous population, the homogeneous G1 daughter culture
showed significant protection over the SCB sequences in the
lacZ
promoter (Fig.
2
A, lanes 4 and 5). Protection of the dA residues within each SCB was more
pronounced than in asynchronous cultures (Fig.
2
A, compare lanes 2, 4 and 5). In addition, slight protection over the dG residue
on the purine-rich strand was also observed (Fig.
2
A, lanes 4 and 5). Some residues between the SCB sequences were also slightly
protected in G1 cells on the purine-rich strand (Fig.
2
A). The dG residues on the pyrimidine-rich strand of the SCB were only weakly footprinted (Fig.
2
A, lane 3, and data not shown). As cells progressed into G2 the SCB footprint
over these sequences was diminished to a level similar to that observed in the
asynchronous population (Fig.
2
A, lanes 6 and 7).
Due to technical limitations on the DNA yields of elutriated samples, some of
the lanes within one experiment were underloaded relative to others (Fig.
2
A, compare lanes 2 and 4). For this reason, DMS modification was also carried
out on deproteinized DNA for each sample and normalized to the
in vivo
DMS modification pattern using a phosphorimager (see Materials and Methods).
For example, phosphorimager analysis of the weak protection apparent in the
residues below the SCB sequences (close to
URA3
; Fig.
2
A, lanes 4 and 5) showed that this protection was not cell cycle specific
(Materials and Methods and data not shown). However, the phosphorimager plots
did confirm the weak protection over the dA residues in the SCB for the
asynchronous cell population (Fig.
2
D, `log'). In the G1 samples, strong protection over dA and dG residues of the
SCB was also observed after phosphorimager normalization, including weak
protection over two adjacent purine residues between the SCBs (Fig.
2
D, panels 1 and 2). As cells progressed into G2 this protection declined
significantly (Fig.
2
D, panel 4).
Northern analysis showed that the peak of
lacZ
transcription in G1 coincided with or followed closely after footprinting over
the SCB region (Fig.
2
B and C). The induction of
lacZ
transcription showed parallel kinetics to another SCB-dependent gene,
CLN2
(Fig.
2
C). The
lacZ
transcript peaked slightly later than
CLN2
and its level persisted compared with
CLN2
mRNA (Fig.
2
C), possibly due to differences in mRNA half-life. The large samples required within each experiment did not allow us
to follow the transcripts further into the cell cycle. However, similar
kinetics of
lacZ
and
CLN2
induction were also observed in another elutriation (data not shown).
A similar pattern of DMS protection was observed in all G1 cell populations,
whether in early or late G1 (Fig.
3
). For example, the footprinting pattern in small daughter cells (11 fl),
although of weak intensity due to the small sample size, did not significantly
differ from the pattern in larger G1 cells (Fig.
3
, lanes 2 and 4-7). One minor difference observed was that protection of the dG residue
of the SCB was more pronounced in cells in mid-G1 compared with early G1 or asynchronous cultures (Fig.
3
). As cells entered G2, protection over the SCB sequences was diminished (Fig.
3
, lanes 9 and 10). For example, cells that were 75% budded showed almost no
footprint as compared with an
in vitro
DMS-modified sample (Fig.
3
, lanes 10 and 11).
To determine whether the
in vivo
SCB footprint was due to SBF or some other DNA binding factor we examined
protection of SCB sequences throughout the cell cycle in strains that were
deleted of
SWI4
or
SWI6
.
SWI4
and
SWI6
disruption strains were each elutriated and examined for DMS footprinting
in vivo
. In contrast to wild-type cells, no significant footprint over the SCB sequences was observed
in either deletion strain at any stage in the cell cycle (Figs
4
A and
5
A). Northern analysis of elutriated cells from both
SWI4
and
SWI6
deletion strains showed that
lacZ
was not transcribed at any stage of the cell cycle (Figs
4
B and C and
5
B and C). However, as has been seen previously, low levels of
CLN2
were still expressed in the absence of
SWI4
or
SWI6
(Figs
4
B and C and
5
B and C;
21
,
22
).
Phosphorimager analysis was performed on the
in vivo
footprinting patterns for both the
swi4
[Delta] and
swi6
[Delta] elutriation experiments (Figs
4
D and 5D). In
swi4
[Delta] cells no detectable footprint was observed in either the asynchronous
population, G1 or G2 cells (Fig.
4
D). In
swi6
[Delta] cells no detectable footprint was seen in any elutriated sample (Fig.
5
D). The very weak footprint observed over the dA residues in the
swi6
[Delta] asynchronous cell population (Fig.
2
D, panel 1) was not observed in other experiments or in the purified G1 or G2
samples from this experiment (Fig.
5
D, panels 2-4, and data not shown).
We also directly compared the
in vivo
DMS modification patterns for wild-type,
swi4
[Delta] and
swi6
[Delta] cells between G1 cells and cells that had entered G2 (Fig.
6
). In wild-type cells there was a clear protection of SCB sequences in the G1 cells
compared with cells that had progressed through Start, both visually and by
phosphorimager analysis (Fig.
6
A). In
swi4
or
swi6
deletion strains, however, there was no detectable difference between the
in vivo
DMS modification patterns in G1 cells compared with cells later in the cell
cycle (Fig.
6
B and C).
We chose centrifugal elutriation and DMS footprinting to analyze the cell cycle
regulation of SCB binding
in vivo
. We found that footprinting over SCB sequences was specifically detected in G1
daughter cells purified by centrifugal elutriation and was not detected in
cells that had progressed into G2. Although the least physiologically
perturbing, cell sampling by elutriation usually precludes analysis beyond one
cell cycle. In contrast, other methods involving cell cycle arrest and release
can typically retain synchrony for two cell cycles. However, we found only a
weak footprint at SCB sequences in wild-type cells using several methods of cell cycle arrest, including [alpha]-factor, nocodazole and temperature-sensitive
cdc28
alleles (data not shown).
The technical difficulty of obtaining enough DNA from each cell cycle sample for
DMS footprinting also necessitated the use of a high copy number plasmid. For example, in these experiments single copy genomic footprinting would require ~1 l of purified, elutriated cells for each cell cycle time point. While
ligation-mediated PCR (LM-PCR) analysis has been successfully used in yeast for
in vivo
footprinting of single copy genomic sequences (
40
), we were unable to detect an SCB footprint with LM-PCR (data not shown). It is possible that the number of protected
sequences was too low to be detected after ligation and PCR amplification. High copy number plasmids were also used in initial studies of
cell cycle-regulated binding of the origin recognition complex, ORC, at ARS1 (
42
,
43
).
Using
in vivo
DMS modification we observed protection of several purines in the SCB promoter
during G1. Most pronounced was protection of the dA residues of the SCB, C
The absence of SCB binding in a
swi4
deletion strain supports the hypothesis that it is SBF that is bound to the
promoter and not some other DNA binding factor, such as MBF. MBF complexes
contain Swi6 and the Swi4 homolog Mbp1 (reviewed in
8
).
In vitro
Mbp1 and Swi4 can bind to the same sequences and binding of SBF to the SCB can be specifically competed by MCB sequences (
24
,
25
). Since Mbp1 is still present in the
SWI4
deletion strain, we presume that the specificity of binding of MBF and SBF is more
tightly regulated
in vivo
and that MBF cannot substitute for binding to the SCB sequences in the absence of Swi4. A recently identified
factor, SCELA (
S.cerevisiae
E2F-like activity), can efficiently bind SCB sites
in vitro
(
44
). Our results suggest that SCELA does not detectably bind SCB sequences
in vivo
.
We have found that Swi4 protein alone cannot bind SCB sequences
in vivo
.
In vitro
the N-terminal DNA binding domain of Swi4 is sufficient to footprint an SCB
consensus sequence in the
CLN2
promoter (
15
). However, in our experiments protection of SCB sequences was not detected in
swi6
[Delta] cells.
A regulatory mechanism may prevent binding of Swi4 alone to the SCB sequences in
G1. Since Swi6 protein is found predominantly in the nucleus only in G1, it is
possible that Swi6 nuclear localization is important for SBF complex formation
(
13
,
48
).
Although the SCB sequences upstream of the G1 cyclins
CLN1
and
CLN2
are required for their maximum expression at Start, recent studies have
indicated that cell cycle regulation of these genes is not strictly dependent
on upstream SCB sites (
21
,
22
). In particular, deletion of the SCB and MCB elements upstream of
CLN2
does not completely abolish its periodic expression (
21
,
22
). Similarly, deletion of three upstream MCB elements does not abolish the G1
periodicity of
SWI4
mRNA (
26
). Thus SBF and MBF may be only partially responsible for the amplification of
G1 cyclin levels at Start. We have shown a correlation between the timing of
SCB binding
in vivo
and transcriptional activation of a reporter gene that is completely SBF
dependent. Whether the role of SBF is to help initiate transcription or amplify
it, binding of SCB sequences at the time of
CLN
transcriptional induction is consistent with a role for SBF binding in the
initial induction of SCB target genes in early G1. However, since SCB sequences
were also occupied in early G1, prior to the peak in SCB-driven transcription, it is possible that a second event after SCB binding
may be required for transcriptional induction.
The fact that SCB binding is G1 specific raises the interesting question as to
what activates the binding of SBF in early G1 and what events must occur, if
any, after SBF binding to activate transcription. Several studies have shown
that transcriptional activation of the G1 cyclins and
HO
are dependent on active Cdc28 kinase (reviewed in
8
). In particular, recent experiments show that SBF- and MBF-driven transcription of the G1 cyclins is activated by the Cln3-Cdc28 kinase (
45
,
46
). Our finding that SCB binding occurs specifically in G1 suggests that the DNA
binding activity of SBF may be a direct target for such regulation. Other
studies have suggested an additional role for Cdc28 in regulating SBF activity.
In particular, it has previously been shown that Clb2-Cdc28 kinase is required for repression of cyclin expression in G2 and
that Clb2 can bind Swi4 (
47
). These results suggest that Clb2-Cdc28 may negatively regulate G1 cyclin expression by direct interaction
with Swi4. Our finding that SCB sites are not significantly occupied in G2 is
consistent with the possibility that Swi4 regulation by Clb2-Cdc28 kinase may occur at the level of SCB binding
in vivo
.
A second study has revealed cell cycle-regulated binding to SCB sequences
in vivo
. Koch
et al
. find G1-specific binding of SCB sequences within the
CLN2
promoter
in vivo
[Koch,C., Schleiffer,A., Ammerer,G. and Nasmyth,K. (1995)
Genes Dev
. in press].
We are indebted to M. Tyers for suggestions and help with initial elutriations.
We are grateful to C. Koch and K. Nasmyth for helpful discussions and for
communicating results prior to publication. For critical comments we thank C.
Greider, A. Spence, P. Sadowski, M. Tyers and members of the laboratory. We
thank C. Smith for assistance with FACS analysis, R. Collins and T. Beattie for advice regarding phosphorimager quantitation, J. Wei for
computer assistance and B. Bolychuk and P. Schwartz for photography. LH was a
post-doctoral fellow of the National Cancer Institute of Canada. This work was
supported by a grant from the Medical Research Council of Canada to BJA. BJA is
a Medical Research Council of Canada Scholar.
+
Present address: Amgen Institute, 620 University Avenue, Toronto, Ontario M5G
2C1, Canada
REFERENCES
Return