close

Вход

Забыли?

вход по аккаунту

?

346

код для вставкиСкачать

. 14: 813–825 (1998)
Quantitative Imaging of TATA-Binding Protein in
Living Yeast Cells
GEORGE H. PATTERSON, STEPHANIE C. SCHROEDER, YU BAI, P. ANTHONY WEIL AND
DAVID W. PISTON*
Department of Molecular Physiology and Biophysics, Vanderbilt University, Nashville, TN 37232, U.S.A.
Received 23 October 1997; accepted 16 January 1998
We describe the quantitative monitoring of TATA-binding protein (TBP) localization and expression in living
Saccharomyces cerevisiae cells. We replaced the endogenous TBP with a green fluorescent protein (GFP) · TBP
fusion, which was imaged quantitatively by laser scanning confocal microscopy (LSCM). When GFP · TBP
expression was altered by using various promoters, the levels measured by LSCM correlated well with the levels
determined by immunoblot of whole cell extract protein. These results show that GFP · TBP imaging not only offers
a method of measurement equivalent to a more conventional technique but also provides real-time quantitation in
living cells and subcellular localization information. Time-lapse confocal imaging of GFP · TBP in mitotic yeast cells
revealed that it remains localized to the nucleus and displays an asymmetric distribution (1:0·7) between mother and
daughter cells. Based on this and data from a mutant which underexpresses GFP · TBP, we suggest that intracellular
levels of TBP are near rate-limiting for growth and viability. 1998 John Wiley & Sons, Ltd.
Yeast 14: 813–825, 1998.
  — Green Fluorescent Protein; GFP; confocal microscopy; mitosis
INTRODUCTION
Transcription in eukaryotic cells is catalysed by
three distinct DNA-dependent RNA polymerases
(RNAPI, RNAPII and RNAPIII), each of which
transcribes one class of genes (class I, II or III;
Hernandez, 1993; Zawel and Reinberg, 1995).
Eukaryotic RNA polymerases cannot specifically
initiate transcription, and each utilizes a distinct
set of general transcription factors (GTFs) that
facilitate accurate initiation and elongation of
transcription. The TATA-binding protein (TBP),
however, is required for transcription by all three
RNA polymerases since TBP is a subunit of SL1,
TFIID and TFIIIB (Hernandez, 1993). Consequently, TBP is the most central transcription
factor in a eukaryotic cell (Hernandez, 1993).
Numerous biochemical and genetic studies have
shown the importance of TBP binding to DNA
*Correspondence to: D. W. Piston, Department of Molecular
Physiology and Biophysics, 702 Light Hall, Vanderbilt University, Nashville, TN 37232, U.S.A. Tel.: (+1) 615 322 7030; fax:
(+1) 615 322 7236.
CCC 0749–503X/98/090813–13 $17.50
1998 John Wiley & Sons, Ltd.
and to other proteins, but changes in TBP concentration and its subcellular localization in vivo have
been more difficult to assay.
Changes in expression level and/or subcellular
distribution of TBP within a living yeast cell could
have significant consequences upon cell physiology, presumably by up- or down-regulating
transcription globally or by modulating expression
of subsets of yeast genes. The consequences of
altered TBP levels have previously been examined
by transient transfection of TBP-encoding genes in
metazoan cells where it was found that TBP
appeared limiting for transcription. For example,
overexpression of TBP in Drosophila Schneider
cells resulted in increased activity from all subclasses of RNAPIII promoters (Trivedi et al.,
1996) as well as from RNAPII TATA-containing
promoters but not from TATA-less promoters
(Colgan and Manley, 1992). Likewise, overexpression of TBP in HeLa and CHO cells was found to
potentiate the response of some transactivators but
not others (Sadovsky et al., 1995). Conversely,
overexpression of a TBP-sequestering factor
814
known as Dr1 caused a decrease in both RNAPIIand RNAPIII-mediated transcription (Kim et al.,
1997). Whether TBP and/or other GTFs are ratelimiting in yeast is less clear. In general, overexpression of TBP in normal cells does not
significantly affect cell growth rates (Poon et al.,
1991), although under certain conditions, overexpression of TBP can increase yeast growth rates
(Auble et al., 1994; Bai et al., 1997; Kim et al.,
1997). In none of these cases, however, was the
effect of TBP overexpression examined for a large
number of specific mRNA-encoding genes. Moreover, the subcellular localization of overexpressed
TBP was not determined. Thus, it is possible that
the reason TBP overexpression showed such a
small effect on overall cell physiology is the failure
of the overexpressed TBP to localize to the
nucleus.
Cell fractionation studies performed in mammalian cells indicate that TBP is normally localized
to the nucleus and is presumably bound to chromatin during interphase, though Segil et al. (1996)
showed that a significant fraction of TBP is cytoplasmic during mitosis. However even during
mitosis, 10–20% of the TFIID remains associated with the DNA. Similarly, the RNAPI-specific
SL1 TBP–TAFI complex localizes with the rRNA
genes throughout the cell cycle (Jordan et al., 1996;
Roussel et al., 1996). Presumably, during yeast cell
division, the mother cell transcription machinery is
distributed equally between daughter cells during
the cell division, although this idea has never been
explicitly tested. Likewise, both cytoplasmic and
chromosome-associated proteins are likely to partition equally during telophase and be equally
shared between the dividing cells.
Unlike mammalian cells, Saccharomyces
cerevisiae exhibits a closed mitosis. Moreover, the
daughter cell is substantially smaller than the
mother cell, so that the two cells may contain
unequal allotments of the mother cell’s machinery.
Further, the yeast cell nuclear membrane remains
intact during mitosis (Robinow and Johnson,
1987; Wheals, 1987) and may therefore present
a barrier to diffusion and/or redistribution of
nuclear proteins throughout the mitotic cycle.
Despite the fact that the daughter cell is smaller
than the mother cell, one could reasonably assume
that a key transcription factor such as TBP is
shared equally between the two cells, particularly if
TBP remains bound to the equally distributed
DNA. However, this aspect of TBP biology has
heretofore not been examined.
1998 John Wiley & Sons, Ltd.
. .   .
In this study, we used the green fluorescent
protein (GFP) to both quantitatively monitor
intracellular TBP levels and investigate its subcellular localization and dynamics in living S.
cerevisiae cells. Originally isolated and cloned from
the jellyfish Aequorea victoria (Chalfie et al., 1994),
GFP has been used to study protein localization
and gene expression in many heterologous systems. Using the S65T derivative of GFP (Heim
et al., 1995) fused to TBP, we have replaced the
endogenous TBP-encoding gene with a gene which
encodes a GFP · TBP fusion protein. Although
this fusion protein doubles the effective mass of
TBP, it is fully functional and cells containing the
‘green’ TBP grow at the same rate as the wild-type
parental yeast cell line. We have validated the
quantitative imaging experiments by showing that
microscopy results with four different GFP · TBP
expression levels closely correlate with results from
protein immunoblot analysis. Finally, we have
monitored GFP · TBP localization throughout the
cell cycle and have found an asymmetric distribution of this transcription factor between mother
and daughter cells. Although the significance of
this particular finding is currently unknown, we
discuss possible explanations and roles that this
asymmetry may have in yeast cell physiology.
MATERIALS AND METHODS
Plasmids and yeast strains
All plasmids were propagated in Escherichia coli
XL1Blue in LB containing 100 ìg/ml ampicillin.
The S65T derivative of GFP was made with the
Kunkel method of site-directed mutagenesis
(Sambrook et al., 1989) using the TU#65 plasmid
(Chalfie et al., 1994) as a template and the mutagenic oligonucleotide, 5-TGAACACCATAAGT
GAAAGTAGTGACA-3 as the annealing primer
to give TU#65-S65T. Using the TU#65-S65T as a
template, PCR amplification products of S65T
were generated using the N-terminal oligonucleotide primer 5-GCGCGTCGACATGAGTAAAG
GAGAAGAACT-3 containing a SalI restriction
site (underlined) and the C-terminal primer 5-GC
GCCTCGAGTTTGTATAGTTCATCCATGC-3
containing an XhoI restriction site (underlined).
Since this GFP variant was used for most of the
work in this paper, we refer to the S65T variant
simply as GFP hereafter in this report. In the one
experiment where the wild-type GFP is used, it
is denoted as such. The amplification products

. 14: 813–825 (1998)
   
containing GFP encoding sequences were gel
purified and digested with the restriction endonucleases, SalI and XhoI, and ligated into a SalIdigested pTFIID plasmid, a yeast expression
plasmid which uses the promoter and regulatory
sequences of the TBP-encoding gene to drive low
to moderate levels of inserted ORFs (Poon et al.,
1991). SalI cloning of the GFP into pTFIID
generated pTFIID-GFP, which expresses low
levels of GFP. The GFP was also ligated in-frame
to the SalI site at the N-terminus of the TBP ORF
in pTFIID-WT (Poon et al., 1991) to generate
pTFIID-GFP · TBP, which expresses low levels of
the GFP · TBP fusion protein. Proper insertion of
the GFP ORF in all expression plasmid constructs was determined by restriction endonuclease
digestion and DNA sequence analysis.
Three other expression vectors, pPGK, pGAL
and pTvnPED, were also used to express GFP,
TBP and GFP · TBP at various levels. We
previously used pPGK-WT and pGAL-WT to
overexpress TBP (Poon et al., 1991). Expression
plasmid pTvnPED expresses the inserted ORFs at
sub-wild-type levels, relative to pTFIID because
the TBP-encoding gene regulatory sequences
which drive expression in this plasmid carry a
transversion in one of the TBP-encoding gene
positive cis elements present in the pTFIID
(Schroeder et al., 1994). To generate the various
overexpression (i.e. pPGK and pGAL) and underexpression (i.e. pTvnPED) constructs, the GFP
and GFP · TBP ORFs were removed from
pTFIID-GFP and pTFIID-GFP · TBP, respectively, by digestion with the restriction endonucleases, SalI and BamHI, gel purified, and
ligated into similarly digested pPGK, pGAL and
pTvnPED plasmids.
GFP plasmids were transformed into the
S. cerevisiae strain, YTW22 (Poon et al., 1991),
originally derived from the parental strain,
YPH252 (MATa ura3-52 trp1-Ä1 his3-Ä200
leu2-Ä1 lys2-801amber ade2-101ochre). YTW22 has
a deletion of the chromosomal copy of the
TBP-encoding gene and a covering plasmid,
pURA-TFIID. The pURA-TFIID plasmid was
replaced with the pTvnPED, pTFIID, pPGK and
pGAL expression vectors using the plasmid shuffle
method. For this experiment, growth of the transformed YTW22 cells on 5-fluoroorotic acid
selected for the loss of the pURA-TFIID plasmid,
leaving only the transformed copy of TBP or
GFP · TBP (Boeke et al., 1987). Strains carrying
the TBP expression plasmids pTvnPED-WT,
1998 John Wiley & Sons, Ltd.
815
pTFIID-WT, pPGK-WT and pGAL-WT are
referred to as YTvnPED-TBP, YTFIID-TBP,
YPGK-TBP and YGAL-TBP, respectively. Strains
carrying the GFP · TBP expression plasmids
pTvnPED-GFP · TBP,
pTFIID-GFP · TBP,
pPGK-GFP · TBP and pGAL-GFP · TBP are
referred to here as YTvnPED-GFP · TBP,
YTFIID-GFP · TBP, YPGK-GFP · TBP and
YGAL-GFP · TBP, respectively. YTW22 was
grown in rich medium and the above-described
strains were grown in synthetic complete (SC)
medium, supplemented with the necessary nutrients for selection and the requisite primary carbon
source, depending on the strain. Growth curves for
each strain were generated by monitoring both
culture optical density at 600 nm and cell number
by direct microscopic cell counting while growing
in SC media at 30C.
Expression and purification of protein standards
His6-tagged GFP protein was expressed in E.
coli and purified as described in Patterson et al.
(1997). Recombinant TBP was prepared, as
described in Poon et al. (1993) from isopropylthioâ--galactopyranoside (IPTG)-induced E. coli
BL21pLysS cells containing a pET11D vector
expressing TBP. To further purify the TBP, the
peak protein fractions from the heparin-sepharose
column were pooled and precipitated by dialysis in
Spectro/Por Molecular porous membrane tubing
(MWCO 12–14,000) against 200 m-Tris pH 7·9,
3·78 -(NH4)2SO4, 1 m-EDTA, and 1 m-âmercaptoethanol (BME) for 22 h at 4C. The
precipitated TBP was pelleted by centrifugation and resuspended in BB500 (20 m-HEPES,
10% glycerol, 0·2 m-EDTA, 5 m-MgCl2,
10 m-BME,
1 m-benzamidine,
0·1 mphenylmethylsulfonyl fluoride and 500 m-NaCl)
and then applied to a sephacryl S-300 gel filtration column (2100 cm) pre-equilibrated with
BB500. The protein concentration of the fractions
was determined by the absorbance at 280 nm using
15,000  1 cm 1 as the extinction coefficient
for TBP. The TBP concentration was verified
with known amounts of bovine serum albumin.
Purity of the resulting TBP was determined to be
>95% by densitometry of a Coomassie brilliant
blue-stained 12% sodium dodecyl sulfate (SDS)–
polyacrylamide gel (PAGE). GST-TBP was
affinity purified using glutathione-agarose resin
(Sigma, St Louis, MO) from IPTG-induced
E. coli BL21pLysS containing a pGEX

. 14: 813–825 (1998)
. .   .
816
plasmid (Stratagene, La Jolla, CA) with TBPencoding cDNA subcloned in-frame with the GST.
The cells were lysed in LB200 (20 m-Tris pH 7·9,
200 m-NaCl) by sonication and the insoluble
debris was removed by centrifugation. The supernatant was incubated with resin prepared according to the manufacturer’s specifications for 2 h at
4C on a tiltboard. The resin was washed with
LB200 and eluted with 20 m reduced glutathione
in LB200. Purification efficiency was determined as
described above and concentration was determined
by bicinchoninic acid assay (Pierce, Rockford, IL).
Protein blot analysis
Protein blots were performed as described by
Bai et al. (1997) with the exception that the transfer buffer was 48 m-Tris, 39 m-glycine, and 20%
methanol. After protein transfer, the membrane
was probed with affinity-purified anti-TBP polyclonal rabbit IgG (Poon et al., 1991), affinitypurified anti-TAFII30 polyclonal rabbit IgG, or
anti-GFP monoclonal antibody (Clontech Laboratories, Inc., Palo Alto, CA) in TBST (100 mTris–Cl pH 7·5, 150 m-NaCl, and 0·1% (v/v)
Tween 20) at room temperature for 2 h. The
membrane was incubated at room temperature
for 1 h with a 1:30,000 dilution of horseradish
peroxidase (HRP)-conjugated anti-rabbit IgG for
the TBP and TAFII30 antibodies or a 1:3000
dilution of HRP-conjugated anti-mouse IgG in
TBST for the GFP antibody and detected by
enhanced chemiluminescence (ECL) and a BioRad (Hercules, CA) ECL Phosphorimager screen.
The relative amount of TBP and GFP · TBP for
each strain was determined using the mean pixel
values of the appropriate bands. Recombinant
TBP and GST-TBP were used as standards in
order to compare the protein blot data with the
imaging data.
Microscopy and analysis
Laser scanning confocal microscopy (LSCM)
was performed using a Zeiss LSM410 microscope
with a 40 Plan Neofluar 1·3 NA oil immersion
objective (Carl Zeiss, Thornwood, NY). The
488 nm line of an argon-krypton ion laser was used
for excitation. A Q498LP (Chroma Technology
Corp., Brattleboro, VT) dichroic mirror and a
HQ512/27 emission filter designed specifically
for GFP was used in the imaging experiments
(Niswender et al., 1995).
1998 John Wiley & Sons, Ltd.
Yeast cells were imaged in Mat-Tek culture
dishes (Mat-Tek Corp., Ashland, MA) in SC
media or on slides coated with 1% agarose containing SC media. When imaging in liquid media, the
cells were allowed to settle to the coverslip covering the bottom of the dish before imaging. Steadystate measurements were made on a single optical
section through the middle of the cells. For the
objective lens and confocal pinhole used here, the
optical sections were 1 ìm thick (Sandison et al.,
1995). Time-lapse image sets were acquired in
three dimensions (10 focal planes spaced 1 ìm
apart) every 10 min for 12 h.
Two channel (fluorescence and Nomarski DIC)
images were analysed using Adobe Photoshop 3.0
(Adobe Systems Incorporated, Mountain View,
CA), NIH Image 1.61 (National Institutes of
Health, Bethesda, MD), and Alice 2.4 (Hayden
Image Processing Group, Boulder, CO). Mean
pixel values from background regions were subtracted from regions of interest for normalization.
Intracellular GFP concentrations were determined
using mean pixel values from images of known
concentrations of purified His6-GFP (S65T) in
deep well slides as standards.
RESULTS
GFP · TBP replaces TBP with no effect on cell
viability and growth
GFP · TBP experiments were performed in yeast
strains where a plasmid carrying the gene encoding
normal TBP was replaced, by plasmid shuffle,
with a plasmid-borne gene encoding a GFP · TBP
fusion protein. Although GFP has been successfully fused to a number of proteins without affecting their function (Cubitt et al., 1995; Gerdes and
Kaether, 1996), we were concerned that tagging
such a central component of the transcription
machinery with such a large probe might interfere
with its ability to function properly. We assessed
the ability of GFP · TBP to support yeast cell
viability by performing growth curves (Figure 1A),
which showed that the YTFIID-GFP · TBP yeast
strain grows at the same rate (doubling time of
2·5 h) as the YTFIID-TBP strain. This similarity
in the growth rates was also found for two of
the other yeast strains that we generated (YPGKTBP and YPGK-GFP · TBP; YGAL-TBP and
YGAL-GFP · TBP; data not shown). As discussed
further below, the two TBP-encoding gene
promoter mutant strains (YTvnPED-TBP and

. 14: 813–825 (1998)
   
817
expressed GFP · TBP, we performed protein
immunoblot experiments (Figure 1B). Yeast whole
cell extracts (WCE) were prepared from equivalent
cell numbers of the YTFIID-TBP and YTFIIDGFP · TBP strains, which yielded roughly equal
amounts of protein. Equivalent protein amounts
of these extracts were fractionated by SDS–PAGE,
blotted to a filter, and TBP was detected with
anti-TBP IgG. The yeast strains, YTFIID-TBP
and YTFIID-GFP · TBP, expressing TBP (lane 1)
and GFP · TBP (lane 2), respectively, express
roughly equivalent amounts of the expected size
proteins (TBP 27 kDa and GFP · TBP
54 kDa. The slight difference in signal intensities
between these two forms of TBP is attributable to
the decreased electrotransfer efficiency of the larger
GFP · TBP. Importantly, no 27 kDa TBP is
present in the YTFIID-GFP · TBP extract, indicating that the intact fusion protein alone is
providing TBP functionality in these cells. The
data in Figure 1 indicate that the 27 kDa GFP
protein fused to the N-terminus of TBP does
not hamper yeast vegetative cell growth and
that GFP · TBP is an adequate replacement for
wild-type TBP.
Figure 1. GFP · TBP replaces TBP with no detriment to yeast
cell growth. The plasmid shuffle method was used to replace the
TBP-encoding gene with the GFP · TBP-encoding gene. (A)
Growth curve comparisons were performed with the yeast
strains YTFIID-TBP and YTFIID-GFP · TBP, which express
TBP and GFP · TBP respectively from the TBP-encoding gene
promoter. (B) A protein immunoblot for TBP was performed
on WCE from each of the strains. The strain from which the
protein was derived is listed above each lane. The mobility
of molecular weight standards run in an adjacent lane are
indicated to the left of the gel.
YTvnPED-GFP · TBP) grew 30% and 80%
more slowly than the YTFIID strains, exhibiting doubling times of 3·25 h and 4·5 h,
respectively (data not shown).
To examine the relative extents of TBP
production/accumulation in these yeast strains
and, more importantly, prove that the GFP moiety
was not simply proteolysed from the GFP · TBP
fusion protein, thereby generating ‘TBP’ from the
1998 John Wiley & Sons, Ltd.
Use of different promoters allows controlled
variation of TBP and GFP · TBP expression levels
We cloned TBP, GFP and GFP · TBP into a
family of expression plasmids which should yield
50-fold variation of protein expression levels
(Poon et al., 1991; Schroeder et al., 1994). To
assure that we achieved this variation of expression
levels, protein immunoblot experiments were
performed. Shown in Figure 2 are protein levels
for the YTFIID-TBP and YTFIID-GFP · TBP
strains (lanes 1 and 2) in comparison with levels
from the YTvnPED yeast strains that underexpress TBP and GFP · TBP (lanes 3 and 4) and the
YPGK (lanes 5 and 6) and YGAL (lanes 7 and 8)
strains that both overexpress TBP and GFP · TBP.
Quantitation of five replicates of this experiment
shows that in relation to pTFIID, the pTvnPED,
pPGK and pGAL expression plasmids drive 0·6fold, 19-fold and 28-fold amounts of TBP protein
production and 0·5-fold, 13-fold and 14-fold
amounts of GFP · TBP. Probing for TAFII30
showed that an equal amount of whole cell protein
was loaded into lanes 1–4 and that lanes 5–8
received 1/10 of lanes 1–4, confirming that the
observed protein levels were not a loading artifact
(data not shown). The increased degradation of

. 14: 813–825 (1998)
. .   .
818
Figure 2. TBP and GFP · TBP protein levels can be systematically
varied in the yeast cell. The plasmid shuffle method was used to
replace the endogenous TBP-encoding gene with TBP- and
GFP · TBP-encoding genes expressed from four different promoters, the TBP promoter (lanes 1 and 2), TvnPED, a promoter
mutant (lanes 3 and 4), the PGK promoter and regulatory sequence
(lanes 5 and 6), and the GAL10 promoter and regulatory sequence
(lanes 7 and 8). Quantitative protein immunoblots were performed
on WCE derived from yeast strains expressing TBP and GFP · TBP
at these four different expression levels. The strain from which the
WCE was derived is listed above each lane. The proteins in lanes
5–8 were diluted 1/10 before gel electrophoresis. TAFII30 immunoblots were performed and quantitated as internal controls (data not
shown). The mobility of molecular weight standards run in an
adjacent lane are indicated to the left of the gel.
the fusion protein visible in lane 8 may account for
the discrepancy between the TBP and GFP · TBP
levels in the YGAL strain. Most importantly,
though, no 27 kDa ‘TBP’ appeared to be generated
from GFP · TBP, even when it was overexpressed.
The relative values of TBP production were all
expected from previous studies (Poon et al., 1991)
with the exception of the YTvnPED strains, which
have not been previously analysed for TBP production levels. Previous work on the cis elements
involved in TBP expression (Schroeder et al., 1994)
showed that transversion mutagenesis of the TBPencoding gene PED cis element (nucleotides 147
to 128 relative to the transcription start site)
dramatically reduced lacZ reporter gene expression levels to <1% of the wild-type promoter level.
In contrast, pTvnPED used here produced protein
levels that were 60% of wild-type levels (58%
for expression of TBP and 66% for GFP · TBP).
It must be noted, however, that unlike the previous
work, we are demanding that pTvnPED drive
sufficient expression of TBP and GFP · TBP to
support yeast cell growth, and the cells have
responded to this selection in two ways. First,
quantitative Southern blot analyses demonstrated
that the pTvnPED-TBP and pTvnPEDGFP · TBP copy number was amplified (four
1998 John Wiley & Sons, Ltd.
copies/cell). Secondly, the cells also utilized posttranscriptional, translational, and/or posttranslational mechanisms to increase TBP and
GFP · TBP expression to levels sufficient to
support life.
Quantitative image analysis yields protein values
that correlate well with protein blot analysis
We next calibrated our quantitative imaging
methodologies using the family of yeast strains
which yielded systematic variation of GFP · TBP
levels. A single 1 ìm thick optical section through
the middle of each ‘green’ yeast strain, YTvnPEDGFP · TBP,
YTFIID-GFP · TBP,
YPGKGFP · TBP and YGAL-GFP · TBP, is shown in
Figure 3. Quantitative image analysis (>20 cells
analysed for each strain) revealed 13-fold
and 24-fold increase in GFP · TBP levels (over
YTFIID) for the YPGK and YGAL strains,
respectively, while the YTvnPED strain showed
a 0·7-fold reduction in GFP · TBP. The
GFP · TBP was localized exclusively to the
nucleus in both the YTvnPED and the YTFIID
strains. However in both overexpressing strains,
only 12-fold wild-type levels of GFP · TBP
were located in the nucleus, while the rest was

. 14: 813–825 (1998)
   
819
Figure 3. Expression of GFP · TBP allows direct observation of TBP in
living yeast cells. Single 1 ìm-thick optical sections of GFP fluorescence
(green) from yeast expressing GFP · TBP are shown for each of the four
strains as indicated (see Figure 2). The fluorescent images are overlaid on
Nomarski DIC images of the same cells (red). Scale bar is 5 ìm.
distributed into the cytoplasm except for the vacuole (Figure 3). GFP expressed alone (which lacks
a nuclear localization signal) was distributed
uniformly throughout the yeast cell, excluding the
vacuole (data not shown). As shown in Figure 4,
the relative expression levels determined by LSCM
imaging and protein immunoblotting correlate
quite well.
By quantitation of the protein blots, we estimated that the average number of TBP molecules
present in each cell of the YTvnPED-TBP,
YTFIID-TBP, YPGK-TBP and YGAL-TBP
strains were 5000, 8500, 160,000 and
240,000, respectively. To perform a similar
analysis using the LSCM imaging data, we compared the total fluorescence intensity in the yeast to
the fluorescence of known amounts of purified
His6-tagged GFP (S65T). From these measurements, we estimated that 500, 700, 10,500
and 17,500 GFP · TBP molecules per cell were
present in the YTvnPED-GFP · TBP, YTFIIDGFP · TBP, YPGK-GFP · TBP and YGALGFP · TBP strains. The reason for the discrepancy
1998 John Wiley & Sons, Ltd.
between the two measurements is unknown,
but several possible causes are discussed below.
Regardless, since the data acquired from imaging
GFP · TBP expressed at various levels correlates
very well with the more conventional technology
of protein immunoblotting, we are confident that
the GFP imaging measurements offer accurate
representations of relative protein levels.
Time lapse imaging of dividing yeast cells shows
that TBP retains nuclear localization throughout
the cell cycle and reveals an asymmetric
distribution of TBP between mothers and
daughters
In addition to the steady-state measurements
described above, we also measured the localization
and distribution of GFP · TBP throughout the
cell cycle (Figure 5A). For these experiments, we
acquired ten 1 ìm thick optical sections, which
assured that each entire yeast cell would be imaged
completely at every time point. Imaging each preparation of cells for 12 h allowed us to follow cells

. 14: 813–825 (1998)
. .   .
820
Figure 4. Quantitative GFP imaging correlates with quantitative protein immunoblotting. Quantitative protein immunoblots were performed on WCE prepared from each of the
strains expressing TBP and GFP · TBP using purified recombinant TBP and purified recombinant GST · TBP as standards
(standards not shown). LSCM and quantitative image analysis
were performed on each of the strains expressing GFP · TBP
using purified recombinant His6-GFP (S65T) as a standard.
Each analysis of the TBP and GFP · TBP levels in the
YTvnPED, YPGK and YGAL strains is graphically represented relative to levels in the YTFIID strains.
through several cell cycles. In Figure 5A, the ten
fluorescence optical sections taken of a group of
living yeast cells were combined (green) and overlaid upon a single Nomarski DIC image taken
through the middle of the cells (red). The nuclear
dynamics and shape changes that we observed
appeared to be the same as those reported earlier
(Robinow and Johnson, 1987; Wheals, 1987; Jones
et al., 1993). During mitosis in the YTFIID and
YTvnPED strains, the GFP · TBP remains localized to the nucleus throughout the cell cycle. This
is not the subcellular localization pattern observed
in mammalian cells undergoing mitosis. One explanation for this data is that S. cerevisiae undergoes
a closed mitosis where the nuclear membrane does
not break down (Robinow and Johnson, 1987;
Wheals, 1987) presumably ‘trapping’ TBP in the
nucleus.
Analysis of three-dimensional image sets of the
YTFIID-GFP · TBP cells revealed an asymmetric
distribution of the GFP · TBP between the mother
and daughter cells after nuclear division. Image
quantitation showed that, on average, the daughter cells received 70% of the TBP of the mother
cells (Figure 5). Although variations in signal
1998 John Wiley & Sons, Ltd.
levels between yeast cells were observed in these
analyses, the temporal series of images in Figure
5A shows a typical pattern of GFP · TBP levels in
a mother cell and a daughter cell during one cell
cycle. Image quantitation is shown in Figure 5B.
The GFP · TBP level in the mother cell increases
before mitosis and then drops back to a ‘normal’
level after budding. The daughter cell clearly
receives less GFP · TBP than the mother, but
GFP · TBP content increases steadily to a ‘normal’
level over 1 h. As a control to determine if this
asymmetry is a GFP folding artifact, a time-lapse
experiment was performed on a similar yeast strain
expressing the wild-type GFP fused to TBP. Presumably, if the asymmetry is due to the kinetics of
GFP folding, the more slowly folding (Heim et al.,
1995) wild-type GFP would result in an even
greater asymmetry. However, quantitative analysis
of these images revealed an average ratio of
1:0·65 between mothers and daughters (data not
shown). A similar analysis of time-lapse images of
the YTvnPED-GFP · TBP strain showed no difference between the mother and daughter cell
GFP · TBP levels (Figure 5C).
DISCUSSION
Use and limitations of GFP in quantitative
imaging of living cells
After constructing a GFP · TBP fusion, we used
this intrinsic fluorescent probe to study both
cellular transcription factor content and dynamics.
By genetic manipulation of the yeast, S. cerevisiae,
we replaced the endogenous TBP-encoding gene
with various forms of a GFP · TBP-encoding gene
that allowed us to monitor GFP · TBP protein
expression levels and dynamics without potential
complications to interpretation due to the presence
of the wild-type untagged version of TBP. We were
initially concerned that the introduction of such a
large tag as GFP (27 kDa) on to such a central
component of the transcription apparatus would
hamper its activity. However, we were able to
demonstrate that there were no differences between the growth rates of yeast expressing TBP
and of yeast expressing GFP · TBP at normal
levels (Figure 1A). Thus, we could conclude that in
fact GFP does not interfere with TBP function.
This evidence coupled with the observation that
GFP · TBP was efficiently and stably produced as
an intact fusion protein (Figure 1B) allowed us to
readily monitor TBP in living yeast cells.

. 14: 813–825 (1998)
   
821
Figure 5. GFP · TBP is asymmetrically distributed between mother cells and daughter cells during
mitosis. (A) A series of time-lapse images of GFP · TBP distribution in the YTFIID-GFP · TBP strain
was acquired by LSCM at 5-min intervals corresponding to the 30–55 min time frame in (B). Scale bar
is 5 ìm. (B) Quantitative analysis of time-lapse images of YTFIID-GFP · TBP progressing through the
cell cycle was performed. These two traces represent the GFP · TBP levels in a typical mother–daughter
pair during one cell cycle. (C) Quantitative image analysis was performed on YTFIID-GFP · TBP and
YTvnPED-GFP · TBP mother–daughter pairs immediately after completion of nuclear division (corresponds to the 25 min time point in panel A). Student’s t tests confirmed a difference in GFP · TBP levels
between YTFIID-GFP · TBP mother and daughter cells (*P<0·01) but not between YTvnPEDGFP · TBP mother and daughter cells.
Qualitative observations of TBP localization can
be performed in fixed and immunostained cells.
However, using GFP genetically fused to TBP
opens up the possibility of performing quantitative
investigations of TBP in living cells in real time. A
potential problem with quantitation of GFP · TBP
(in particular, the S65T derivative) is the temperature dependence of chromophore formation. To
circumvent this caveat, our imaging experiments
were performed at 25C, where formation of the
GFP (S65T) chromophore was 100% complete
when expressed in bacteria (Patterson et al., 1997).
In the future, we could eliminate this problem by
using a temperature-stable GFP mutant, such as
the EGFP (Cormack et al., 1996; Siemering et al.,
1996; Patterson et al., 1997). To address this
complication and convince ourselves that our GFP
1998 John Wiley & Sons, Ltd.
measurements were indicative of normal intracellular TBP action, we compared our LSCM
imaging quantitation with measurements of TBP
concentration from protein immunoblotting.
By using protein immunoblotting methods,
we previously demonstrated that the expression
plasmids
pTFIID-WT,
pPGK-WT,
and
pGAL-WT resulted in expression of 1-fold,
10-fold and 30-fold amounts of TBP in yeast
cells. Our analysis of GFP · TBP expression from
this family of expression plasmids by LSCM indicated that quantitative imaging of GFP · TBP
gives a good relative representation of TBP concentrations inside cells. However, we do not at
present understand why the imaging measurements of the absolute intracellular TBP levels did
not agree with the biochemical measurements. A

. 14: 813–825 (1998)
822
combination of several factors could be responsible for the 10-fold difference between the two
sets of measurements. First, it is possible that not
all of the GFP is folding into the fluorescent state.
Our experiments were performed at a temperature
that should allow efficient chromophore formation
(Patterson et al., 1997), but if for some reason
GFP folds less efficiently in yeast than bacteria,
this could affect image quantitation. Secondly, the
yeast cell wall could be filtering or scattering some
GFP fluorescence resulting in a decrease in the
amount of detected fluorescent signal produced in
the cells. Finally, the internal environment of the
yeast cell may be suboptimal for GFP fluorescence
(Patterson et al., 1997). The intracellular pH of a
yeast cell has been estimated to be 6·6–7·0
(Garcı́a-Arranz et al., 1994). Determination of
GFP (S65T) fluorescence as a function of pH
indicates that its brightness is only slightly perturbed (10–15%) at this pH (Patterson et al.,
1997). However, a small loss of signal due to pH
coupled with the other factors could reduce the
apparent number of GFP molecules detected, and
thus account for the discrepancy between the
imaging and biochemical quantitation.
Overexpression of TBP in vivo
Since TBP is thought to be limiting for transcription in vivo (Colgan and Manley, 1992), overexpressing it might be expected to increase total
cellular transcription and possibly increase the
growth rate of the cell. Previous studies which
examined the effects of overexpressing TBP in
yeast found that their growth rate characteristics
did not change (Poon et al., 1991). However, it was
not known if the ‘extra’ TBP failed to localize to
the nucleus, which is a trivial explanation for this
result. The data presented in this report show that
in the YPGK-GFP · TBP and YGAL-GFP · TBP
yeast strains, 12-fold wild-type levels of
GFP · TBP is localized to the nucleus (Figure 3)
while the rest is cytoplasmic. Whether this 12-fold
value is a physical limit (i.e. the nucleus is full) or
if the extra GFP · TBP simply cannot be transported into or retained in the nucleus (i.e. all
transport machineries, binding sites, transcription
components, etc. are occupied) is unknown. It is
also possible that the fluorescence observed outside the nucleus in the YPGK and YGAL strains
derives from degraded GFP · TBP (Figure 2, lanes
6 and 8) and is not forced to the cytoplasm by any
nuclear localization limit. Regardless, by being
1998 John Wiley & Sons, Ltd.
. .   .
localized to the nucleus, at least 12 times the
normal amount of GFP · TBP has the potential
to participate in transcription, yet this ‘extra’
TBP does not result in an increase in yeast cell
growth. Presumably, another component of the
transcription machinery becomes limiting under
these conditions.
In contrast, the low expression yeast strains,
YTvnPED-GFP · TBP and YTvnPED-TBP, had
slower growth rates than normal and expressed
GFP · TBP and TBP to 66% and 58% of their
wild-type counterparts, YTFIID-GFP · TBP and
YTFIID-TBP, respectively (Figure 4). The two
strains also contained multiple TBP and
GFP · TBP gene copies. These data suggest that
there is, in fact, a lower limit of TBP protein
content below which yeast cells are non-viable. It is
interesting to note that the TBP steady-state levels
in cells expressing these two forms of TBP from the
pTvnPED expression plasmid are the same as that
found in newly generated daughter cells (60–70%
normal). The implications of this lower TBP level
on global gene transcription are currently being
investigated.
Asymmetric distribution of TBP between mother
and daughter cells
The asymmetric distribution of GFP · TBP
between the mother cell and the daughter cell
(Figure 5B) is a surprising observation, and several
possible reasons for this distribution are discussed
here. During mitosis in budding yeast, the nuclear
membrane does not disassemble, but instead the
nucleus migrates to the bud site and undergoes a
division much like a dividing cell (Robinow and
Johnson, 1987; Wheals, 1987). Since the nucleus
appears to divide equally between each mother and
daughter (Jones et al., 1993), a nuclear volume
difference is probably not the underlying reason
for the observed asymmetric distribution of
GFP · TBP. A second possibility is that the difference in TBP levels between the two nuclei is an
artifact produced by GFP · TBP degradation.
However, as shown in Figures 1B and 2, only a
very small amount of degradation of the fusion
protein can be seen in the TFIID-GFP · TBP
strain WCE, so this is probably not the cause
for the asymmetric GFP · TBP distribution. In
addition, the GFP · TBP in the daughter cell
increases to the mother cell concentration over
time, suggesting that this level is the ‘normal’
resting GFP · TBP level. A third possibility is that

. 14: 813–825 (1998)
   
this observation is an artifact due to the kinetics of
the GFP chromophore formation. This GFP variant (S65T) exhibits a chromophore formation time
constant of 0·45 h and could result in a ‘fluorescence lag’ of the GFP · TBP (Heim et al., 1995).
However, this possibility seems unlikely since a
similar time-lapse experiment performed on
yeast (YWT-wtGFP · TBP) expressing the more
slowly folding wild-type version of GFP fused to
TBP gave a similar asymmetry (1:0·65) between
the mothers and daughters (data not shown).
Another possibility is that the mRNAGFP · TBP is
asymmetrically distributed between mother and
daughter cells thus leading to the unequal levels of
GFP · TBP in the divided cells. Molecular genetic
analyses of Ash1p found that this protein was
asymmetrically localized to the daughter cells
during mitosis (Bobola et al., 1996; Sil and
Herskowitz, 1996) and that this asymmetry is
dictated by the localization of mRNAASH1 (Long
et al., 1997). Therefore, it is possible that the
mother cell contains more mRNAGFP · TBP which
leads to an asymmetric expression of GFP · TBP
protein; however, summation of the GFP · TBP
levels in the mother–daughter pair after nuclear
separation approximately equals the level of
GFP · TBP observed in the mother cell before
nuclear division (Figure 5B and data not shown).
In any case, it remains that the level of GFP · TBP
protein in the daughter cell immediately after
nuclear division is less than that in the mother cell.
The mechanism for the asymmetric GFP · TBP
distribution is most likely quite complicated, but
could involve partial association of GFP · TBP
with DNA, other transcription factors, or the
mother cell nuclear matrix during mitosis. As
mentioned earlier, 10–20% of the mammalian
TFIID complex (Segil et al., 1996) and all of the
mammalian SL1 complex (Jordan et al., 1996;
Roussel et al., 1996) remain associated with the
chromosomes throughout mitosis via mechanisms
which are currently not understood.
In mitotic cells during the M-phase, there is a
global inhibition of transcription by all three
classes of DNA-dependent RNA polymerases
(Prescott and Bender, 1962). Phosphorylation of
several of the GTFs and their components, including TBP, is thought to play a significant role in this
transcriptional repression (Gottesfeld et al., 1994;
White et al., 1995; Leresche et al., 1996; Segil et al.,
1996). One of the hypotheses for the mechanism
of the mitotic inhibition of transcription (Segil
et al., 1996) holds that the redistribution of some
1998 John Wiley & Sons, Ltd.
823
transcription factors may be a mechanism for
changing gene expression either during the cell
cycle or during development. The possible effects
upon gene expression that lower GFP · TBP concentrations could have in the daughter cells is
unknown, though it is not unreasonable to
postulate that this decreased concentration of TBP
could cause significant changes in specific gene
transcription.
Conclusions
A thorough understanding of TBP levels and
dynamics inside the living cell will ultimately be
crucial to unravelling the complexities of gene
regulation on a global, genome-wide scale. This
was one of the main reasons we attempted to
construct and utilize the GFP · TBP reporter. Our
detailed comparisons of quantitative GFP imaging with a more accepted biochemical method
validates the use of GFP as a quantitative tool for
biological research. Some of the applications of
GFP that we describe here, such as steady-state
comparisons of TBP levels in various strains, could
easily be made using protein immunoblotting.
However, kinetic monitoring of GFP · TBP
throughout the cell cycle is not possible using
biochemical methods. Notably, it was our kinetic
observations of GFP · TBP concentrations and
intracellular dynamics that led to the discovery of
asymmetric TBP distributions during yeast cell
division. How or why this distribution occurs is
presently unknown, but the implications for
transcription in a freshly budded cell containing
0·7-fold normal TBP levels may lead to a better
understanding of the physiological role of TBP.
ACKNOWLEDGEMENTS
We thank Kevin Gerrish and Steven Sanders for
assistance with protein immunoblotting. These
studies were supported by grants from the
Beckman Foundation Young Investigator Program and the Whitaker Foundation Biomedical
Engineering Research Program to D.W.P., and
from NIH to P.A.W. (GM52461). During part of
this work, G.H.P. was an NIH trainee (GM08320).
Confocal microscopy was performed at the Cell
Imaging Shared Resource, supported by the
Vanderbilt Cancer Center (CA68485) and
Diabetes Research and Training Center
(DK20593).

. 14: 813–825 (1998)
824
REFERENCES
Auble, D. T., Hansen, K. E., Mueller, C. G. F., Lane,
W. S., Thorner, J. and Hahn, S. (1994). Mot1, a
global repressor of RNA polymerase II transcription,
inhibits TBP binding to DNA by an ATP-dependent
mechanism. Genes Develop. 8, 1920–1934.
Bai, Y., Perez, G. M., Beechem, J. M. and Weil, P. A.
(1997). Structure-function analysis of TAF130:
identification and characterization of a high-affinity
TATA-binding protein interaction domain in the N
terminus of yeast TAFII130. Mol. Cell. Biol. 17,
3081–3093.
Bobola, N., Jansen, R.-P., Shin, T. H. and Nasmyth, K.
(1996). Asymmetric accumulation of Ash1p in postanaphase nuclei depends on a myosin and restricts
yeast mating-type switching to mother cells. Cell 84,
699–709.
Boeke, J. D., Trueheart, J., Natsoulis, G. and Fink,
G. R. (1987). 5-Fluoroorotic acid as a selective agent
in yeast molecular genetics. Methods Enzymol. 154,
164–175.
Chalfie, M., Tu, Y., Euskirchen, G., Ward, W. W. and
Prasher, D. C. (1994). Green fluorescent protein as a
marker for gene expression. Science 263, 802–805.
Colgan, J. and Manley, J. L. (1992). TFIID can be a
rate limiting in vivo for TATA-containing, but not
TATA-lacking, RNA polymerase II promoters. Genes
Develop. 6, 304–315.
Cormack, B. P., Valdivia, R. H. and Falkow, S. (1996).
FACS-optimized mutants of the green fluorescent
protein (GFP). Gene 173, 33–38.
Cubitt, A. B., Heim, R., Adams, S. R., Boyd, A. E.,
Gross, L. A. and Tsien, R. Y. (1995). Understanding,
improving and using green fluorescent proteins. TIBS
20, 448–455.
Garcı́a-Arranz, M., Maldonado, A. M., Mazón, M. J.
and Portillo, F. (1994). Transcriptional control of
yeast plasma membrane H + -ATPase by glucose. J.
Biol. Chem. 269, 18,076–18,082.
Gerdes, H.-H. and Kaether, C. (1996). Green fluorescent
protein: applications in biology. FEBS Letters 389,
44–47.
Gottesfeld, J. M., Wolf, V. J., Dang, T., Forbes, D. J.
and Hartl, P. (1994). Mitotic repression of RNA
polymerase III transcription in vitro mediated by
phosphorylation of a TFIIB component. Science 263,
81–84.
Heim, R., Cubitt, A. B. and Tsien, R. Y. (1995).
Improved green fluorescence. Nature 373, 663–664.
Hernandez, N. (1993). TBP, a universal eukaryotic
transcription factor? Genes Develop. 7, 1291–1308.
Jones, H. D., Schliwa, M. and Drubin, D. G. (1993).
Video microscopy of organelle inheritance and
motility in budding yeast. Cell Motility and the
Cytoskeleton 25, 129–142.
Jordan, P., Mannervik, M., Tora, L. and CarmoFonseca, M. (1996). In vivo evidence that TATA 1998 John Wiley & Sons, Ltd.
. .   .
binding protein/SL1 colocalizes with UBF and RNA
polymerase I when rRNA synthesis is either active or
inactive. J. Cell Biol. 133, 225–234.
Kim, S., Na, J. G., Hampsey, M. and Reinberg, D.
(1997). The Dr1/DRAP1 heterodimer is a global
repressor of transcription in vivo. Proc Natl. Acad. Sci.
USA 94, 820–825.
Leresche, A., Wolf, V. J. and Gottesfeld, J. M. (1996).
Repression of RNA polymerase II and III transcription during M phase of the cell cycle. Exp. Cell Res.
229, 282–288.
Long, R. M., Singer, R. H., Meng, X., Gonzalez, I.,
Nasmyth, K. and Jansen, R.-P. (1997). Mating type
switching in yeast controlled by asymmetric localization of ASH1 mRNA. Science 277, 383–387.
Niswender, K. D., Blackman, S. M., Rohde, L.,
Magnuson, M. A. and Piston, D. W. (1995). Quantitative imaging of green fluorescent protein in cultured
cells: comparison of microscopic techniques, use in
fusion proteins and detection limits. J. Microscopy
180, 109–116.
Patterson, G. H., Knobel, S. M., Sharif, W. D., Kain,
S. R. and Piston, D. W. (1997). Use of the greenfluorescent protein (GFP) and its mutants in
quantitative fluorescence microscopy. Biophys. J. 73,
2782–2790.
Poon, D., Schroeder, S. C., Wang, C. K., et al. (1991).
The conserved carboxy-terminal domain of Saccharomyces cerevisiae TFIID is sufficient to support normal
cell growth. Mol. Cell. Biol. 11, 4809–4821.
Prescott, D. M. and Bender, M. A. (1962). Synthesis of
RNA and protein during mitosis in mammalian tissue
culture cells. Exp. Cell Res. 26, 260–268.
Robinow, C. F. and Johnson, J. S. (1987).Yeast cytology: an overview. The Yeasts, vol. 4. Academic Press,
Inc., NY, pp. 7–120.
Roussel, P., André, C., Comai, L. and HernandezVerdun, D. (1996). The rRNA transcription machinery is assembled during mitosis in active NORs and
absent in inactive NORs. J. Cell Biol. 133, 235–246.
Sadovsky, Y., Webb, P., Lopez, G., et al. (1995). Transcriptional activators differ in their responses to overexpression of TATA-box-binding protein. Mol. Cell.
Biol. 15, 1554–1563.
Sambrook, J., Fritsch, E. F. and Maniatis, T. (1989).
Site-directed mutagenesis of cloned DNA. Molecular
Cloning: A Laboratory Manual, vol. 2. Cold Spring
Harbor Laboratory Press, NY, pp. 15.3–15.108.
Sandison, D. R., Piston, D. W., Williams, R. M. and
Webb, W. W. (1995). Resolution, background rejection, and signal-to-noise in widefield and confocal
microscopy. Applied Optics 34, 3576–3588.
Schroeder, S. C., Wang, C. K. and Weil, P. A. (1994).
Identification of the cis-acting DNA sequence
elements regulating the transcription of the Saccharomyces cerevisiae gene encoding TBP, the TATA
binding protein. J. Biol. Chem. 269, 28,335–28,346.

. 14: 813–825 (1998)
   
Segil, N., Guermah, M., Hoffmann, A., Roeder, R. G.
and Heintz, N. (1996). Mitotic regulation of TFIID:
inhibition of activator-dependent transcription and
changes in subcellular localization. Genes Develop. 10,
2389–2400.
Siemering, K. R., Golbik, R., Sever, R. and Haseloff, J.
(1996). Mutations that suppress the thermosensitivity
of green fluorescent protein. Curr. Biol. 6, 1653–1663.
Sil, A. and Herskowitz, I. (1996). Identification of
an asymmetrically localized determinant, Ash1p,
required for lineage-specific transcription of the yeast
HO gene. Cell 84, 711–722.
Trivedi, A., Vilalta, A., Gopalan, S. and Johnson,
D. L. (1996). TATA-binding protein is limiting for
1998 John Wiley & Sons, Ltd.
825
both TATA-containing and TATA-lacking RNA
polymerase III promoters in Drosophila cells. Mol.
Cell. Biol. 16, 6909–6919.
Wheals, A. E. (1987). Biology of the cell cycle in yeasts.
The Yeasts, vol. 1. Academic Press, Inc., NY,
pp. 283–390.
White, R. J., Gottlieb, T. M., Downes, C. S. and
Jackson, S. P. (1995). Mitotic regulation of a TATAbinding-containing complex. Mol. Cell. Biol. 15,
1983–1992.
Zawel, L. and Reinberg, D. (1995). Common themes in
assembly and function of eukaryotic transcription
complexes. Ann. Rev. Biochem. 64, 533–561.

. 14: 813–825 (1998)
Документ
Категория
Без категории
Просмотров
2
Размер файла
188 Кб
Теги
346
1/--страниц
Пожаловаться на содержимое документа