Upon facing stress conditions
the budding yeast Saccharomyces
cerevisiae is able to choose among
different cellular fates such as unicellular or multicellular organization,
with each organism being able to undertake growth modes accordingly to the
different needs as differential growth modes can confer the cell protection and
enhance its dissemination and substrate colonization. When either
haploid or diploid Saccharomyces cerevisiae cells are
subject of such stress like during starvation conditions, they will develop two
different types of growth: adhesive/invasive for the first and pseudohyphae for
the latter. For both of these phenomena,
yeasts will experience an increment in cell length, change of polarity and an
augmentation of cell to cell adhesion; in pathogens, this kind of development
is fundamental for host-cell attachment, virulence and tissue invasiveness.
It has been identified that Flo11
have a central role in pseudohyphal (filamentous growth) growth: when FLO11 is deleted, diploids of Saccharomyces
cerevisiae strain Σ1278b do not form pseudohyphae and there is
no agar invasion in haploids. In nitrogen starvation conditions, wild type
diploids but not haploids will accumulate FLO11 transcripts. (Lo and Dranginis, 1998) FLO11also has a particularly long
promoter expanding for about 3 kilo bases (kb) and intensive studies by Baumgarner
et
al. have unveiled the presence of
a paradigmatic epigenetic toggle switch controlled by two cis-interfering long noncoding
RNAs. The region upstream the FLO11
ORF
encodes ICR1, an ncRNA of about 3.2 kb
long whose transcription inhibits that of FLO11 trough the mechanism of promoter
occlusion. The second
ncRNA, PWR1
is transcribed from the complementary, Watson strand, blocking ICR1 transcription and
in consequence, promoting that of FLO11. These two
transcripts, IRC1
and PWR1, are at the
same time subjected to the control of two transcription factors, Sfl1 and Flo8
and chromatin remodeler Rpd3L, an HDAC with the paradoxical effect of promoting
FLO11
transcription. In the model proposed by the authors, Rpd3L localizes to the FLO11 promoter and
contributes to the repression of ICR1 transcription
by facilitating Flo8 binding and PWR1 transcription
together with FLO11.
In
addition to nitrogen depletion, amino acid starvation has been
reported to induce an increase of FLO11 expression. When
transcriptional coactivators Rsc1 and Gcn5 are deleted, FLO11
transcription is impaired together with the concomitant changes on invasive
growth, but in the case of amino acid starvation such deletion have a different
effect. FLO11
transcription is very low, but rsc1Δ
and gcn5Δ
still show invasive growth and the FLO11prom-lacZ-encoded
β-galactosidase activities increase importantly. (
rewrite to point out the function of Gcn5)
It has been demonstrated that Gcn5, a histone acetyltransferase
(HAT) that has functions as coactivator in transcriptional regulation,
contribute largely to transcriptional activation and particularly to enhance
substrate specificity in combination with other histone modifications. (Clements et al., 2003) Gcn5p is the
catalytic subunit for three chromatin modifying complexes of the ADA, SAGA and
SILK/SALSA that act regulating a wide arrange of genes either in a positive and
negative fashion.
Additionally, it was shown by
Fischer et al. that amino acid starvation
induced FLO11-dependent adhesive growth of
the rsc1Δ and gcn5Δ strains although FLO11transcription remained very
low. The double deletion strain rsc1Δflo11Δ, however, does not grow
adhesively, suggesting that the adhesion of the rsc1Δ strain at amino acid
starvation is FLO11-dependent. Hence, it was resolved that very low
FLO11 transcripts are essential and sufficient for derepression of FLO11 expression and adhesive growth
during amino acid starvation. (Fischer et al., 2008)
Results
The histone
acetyl transferase (HAT) Gcn5
is involved in pseudohyphal development.
We
tested the haploid and diploid strains for phenotype changes caused by the
deletion of GCN5 (Figure 1A). In the case of
the haploid, it was shown through wash tests that GCN5 is required to develop
invasive growth; the same was true for snf1 and FLO11. In the case of the diploid,
the mutant strain didn’t develop pseudohyphae similarly to the pseudo-diploid;
an α haploid containing a single MATa1 plasmid whose WT presents the
same behavior than natural diploids. FLO11 mRNA levels have different
pattern of expression in WT diploids than that in WT haploids during amino acid
starvation (Figure 1B); the first have an striking increase peaking at around
1.5h, but clearly decreasing until getting back to basal levels after two more
hours. Haploids, on the other hand, have mRNA levels that increase gradually
and only reach that equal half the maximum detected for the diploid strain,
however such expression is stably maintained for about four hours, until a
final sharp drop. Nevertheless,
FLO11
expression is heavily impaired in both haploid and diploid gcn5
mutants
(Figure 1C) under amino acid starvation conditions, with just very low levels
still remaining. To study if this effect is due to the histone acetyl
transferase (HAT) activity of GCN5, we also tested a strain that expresses full
length GCN5 carrying a catalytic impairing mutation, E173Q. The level of expression are
similar to those of the deletion mutant, so we can infer that the HAT activity
is required for FLO11 expression during amino acid starvation.
Invasive
and filamentous
growth phenotypes
are
restored in deletion mutant strains carrying an
episomal GCN5.
FLO11 expression is required for
haploid invasive growth, and flo11 strains are totally washed
away when tested (Figure 2A). The lower levels of FLO11 expression on gcn5
strains have a repercussion in that phenotype and
less than 20% of the total growth remains after the wash test. However, when
the yeast are transformed with a plasmid containing GCN5, the invasive growth
increases up to the double than the deletion mutant on 50% the level of the WT
strain, but when the transformation was done with a plasmid carrying the
non-catalytic, E173Q GCN5, the invasive growth levels are
similar to the gcn5 strain. A similar outcome can
be seen in diploids, where filamentous development in the gcn5 and the E173Q strains is
heavily impaired, but is significantly rescued with the GCN5 carrying plasmid
(Figure 2B).
The transcription factor
Gcn4 binds to FLO11
promoter in response to 3-AT
induction.
Amino
acid starvation is sensed into the cell trough diverse pathways, and microarray
data characterized GCN4 as a positive regulator
during that phenomenon. Being a broad gene activator, it’s well arguable that
an interaction may exist between the FLO11 promoter and this transcription
factor, so we analyze that possibility using chromatin immunoprecipitation with
8 primers located at different regions along the extensive upstream control
region (Figure 3A). For haploid strains, we measured the interaction at 0, 2
and 4 hours (Figure 3B) and time 0 and 1 hour for diploid (Figure 3C). Our
results shows that the regions covered by primers NR2 and NR5 exhibit a notable
enrichment for both kind of strains, with a striking multifold increase for the
diploid. To our understanding this is the first report concerning such direct
interaction between Gcn4 and the FLO11 promoter.
Gcn5 regulate the
transcription of an
ncRNA (ICR1) and a flocculin gene (FLO11)
in response to 3-AT induction.
In
the extensive promoter region upstream of FLO11 two long non-coding RNAs are
found controlling an aforementioned epigenetic toggle switch (Figure 4A) where FLO11 expression is halted via the
promoter occlusion mechanism when ICR1 transcription takes place.
For the haploid strain, the pattern of IRC1
RNA expression turned out remarkably similar among
the WT and gcn5
strains when they were subject of 3-AT induction and
also when they were not starved and nevertheless FLO11 RNA levels are significantly
different (Figure 4B). The low levels of FLO11 RNA expressed are consistent
and correlate to the phenotype shown in Figure 1A, were a big percentage, but
not all, of the invasive growth is lost. Noticeably the only difference in the
pattern can be found at about -1 Kb from the FLO11 transcription start site, and
while small it seems to be significant.
We
found a radically different story told by the pattern of expression from the
diploid strains; even in normal nutritional status the gcn5 strain has a significantly
higher level of ICR1
transcriptional activity around -2 and -1 Kb upstream
the FLO11 transcriptional start. Under
amino acid starvation conditions, this difference is magnified several fold and
the FLO11 profile shows a silencing of
the FLO11 transcription that supports
the observations annotated in Figure 1B, no pseudohyphal development is
observed for gcn5. This results indicate that
GCN5 exerts a different control, at least in magnitude, among haploid and
diploids. It should also be noted that the regions with a higher ICR1 expression correspond to
those two regions enriched by Gcn4 binding.
Gcn5 is a negative regulator
of ICR1 in diploid strains.
ICR1 is a long non coding
transcript whose expression inhibits that of FLO11,
