Cdc28 kinase activity regulates the basal transcription machinery at a subset of genes.

Abstract

The cyclin-dependent kinase Cdc28 is the master regulator of the cell cycle in Saccharomyces cerevisiae. Cdc28 initiates the cell cycle by activating cell-cycle-specific transcription factors that switch on a transcriptional program during late G1 phase. Cdc28 also has a cell-cycle-independent, direct function in regulating basal transcription, which does not require its catalytic activity. However, the exact role of Cdc28 in basal transcription remains poorly understood, and a function for its kinase activity has not been fully explored. Here we show that the catalytic activity of Cdc28 is important for basal transcription. Using a chemical-genetic screen for mutants that specifically require the kinase activity of Cdc28 for viability, we identified a plethora of basal transcription factors. In particular, CDC28 interacts genetically with genes encoding kinases that phosphorylate the C-terminal domain of RNA polymerase II, such as KIN28. ChIP followed by high-throughput sequencing (ChIP-seq) revealed that Cdc28 localizes to at least 200 genes, primarily with functions in cellular homeostasis, such as the plasma membrane proton pump PMA1. Transcription of PMA1 peaks early in the cell cycle, even though the promoter sequences of PMA1 (as well as the other Cdc28-enriched ORFs) lack cell-cycle elements, and PMA1 does not recruit Swi4/6-dependent cell-cycle box-binding factor/MluI cell-cycle box binding factor complexes. Finally, we found that recruitment of Cdc28 and Kin28 to PMA1 is mutually dependent and that the activity of both kinases is required for full phosphorylation of C-terminal domain-Ser5, for efficient transcription, and for mRNA capping. Our results reveal a mechanism of cell-cycle-dependent regulation of basal transcription.

Full text

Page 1

Cdc28 kinase activity regulates the basal transcription
machinery at a subset of genes
Pierre Chymkowitcha,b, Vegard Eldholma,b,1, Susanne Lorenzc,d,1, Christine Zimmermanna,b, Jessica M. Lindvalle,
Magnar Bjøråsa,b, Leonardo A. Meza-Zepedac,d, and Jorrit M. Enserinka,b,2
aDepartment of Molecular Biology, Institute of Microbiology, andbCentre for Molecular Biology and Neuroscience, Oslo University Hospital, Sognsvannsveien
20, NO-0027 Oslo, Norway;cNorwegian Microarray Consortium, Department of Molecular Biosciences, University of Oslo, anddDepartment of Tumor Biology,
The Norwegian Radium Hospital, Montebello, NO-0310 Oslo, Norway; andeBioinformaticService, Vastertappsvagen 12, SE-132 35 Saltsjo-Boo, Sweden
Edited by Kevin Struhl, Harvard Medical School, Boston, MA, and approved May 21, 2012 (received for review January 4, 2012)
The cyclin-dependent kinase Cdc28 is the master regulator of the
cell cycle in Saccharomyces cerevisiae. Cdc28 initiates the cell cycle
by activating cell-cycle–specific transcription factors that switch on
a transcriptional program during late G1 phase. Cdc28 also has
a cell-cycle–independent, direct function in regulating basal tran-
scription, which does not require its catalytic activity. However,
the exact role of Cdc28 in basal transcription remains poorly un-
derstood, and a function for its kinase activity has not been fully
explored. Here we show that the catalytic activity of Cdc28 is
important for basal transcription. Using a chemical-genetic screen
for mutants that specifically require the kinase activity of Cdc28
for viability, we identified a plethora of basal transcription factors.
In particular, CDC28 interacts genetically with genes encoding
kinases that phosphorylate the C-terminal domain of RNA poly-
merase II, such as KIN28. ChIP followed by high-throughput se-
quencing (ChIP-seq) revealed that Cdc28 localizes to at least 200
genes, primarily with functions in cellular homeostasis, such as the
plasma membrane proton pump PMA1. Transcription of PMA1
peaks early in the cell cycle, even though the promoter sequences
of PMA1 (as well as the other Cdc28-enriched ORFs) lack cell-cycle
elements, and PMA1 does not recruit Swi4/6-dependent cell-cycle
box-binding factor/MluI cell-cycle box binding factor complexes.
Finally, we found that recruitment of Cdc28 and Kin28 to PMA1
is mutually dependent and that the activity of both kinases is re-
quired for full phosphorylation of C-terminal domain–Ser5, for effi-
cient transcription, and for mRNA capping. Our results reveal a
mechanism of cell-cycle–dependent regulation of basal transcription.
Cdk1|Cks1|Ctk1|Cdk7|Cdk9
C
sufficient for cell-cycle regulation in the budding yeast Saccharo-
myces cerevisiae, phosphorylating a large number of substrates to
coordinate the cell cycle (1). In late G1, Cln3-Cdc28 complexes
phosphorylate Whi5, leading to its dissociation from the tran-
scription factor complex Swi4/6-dependent cell-cycle box-binding
factor (SBF), a Swi4-Swi6 heterodimer. Dissociation of Whi5
activatesSBF,whichtheninducestranscriptionoftheG1program
that includes cyclins CLN1, CLN2, CLB5, and CLB6 (2). Cln1,2-
Cdc28 complexes can also phosphorylate Whi5, setting up a pos-
itive feedback loop that ensures coherent cell-cycle entry (3).
Transcriptional activationinvolves assembly of RNA polymerase
II (RNAPII) and general transcription factors at the promoter
region of genes. The C-terminal domain (CTD) of Rpb1, the
largest subunit RNAPII, consists of multiple repeats of the hepta-
peptide Y1S2P3T4S5P6S7, and residues within the CTD are differ-
entially phosphorylated during transcription (4). Early in the
transcription cycle, Kin28 phosphorylates the CTD on serine 5,
which serves as a mark for recruitment of the mRNA capping
machinery (5). As RNAPII elongates, phosphorylated S5 levels
decrease progressively because of the action of the CTD-S5P–
specificphosphatasesRtr1andSsu72,andserine2phosphorylation
increases toward the 3′ end of the ORF as the result of kinase
yclin-dependent kinases (CDKs) drive the cell cycle in eu-
karyotic cells. Cdc28, also known as “Cdk1,” is necessary and
activity of Bur1 and Ctk1 (4). Phosphorylated CTD-S2 serves as
a docking site for a multitude of protein complexes involved in
histone modification, chromatin remodeling, mRNA polyade-
nylation, and transcription termination (4). Recently, CTD-S7
also was shown to be phosphorylated (6). Phosphorylation of this
residue is carried out by Kin28 and Bur1, and although its func-
tion is obscure in budding yeast, it contributes to expression of
noncoding RNA and mRNA splicing in mammalian cells (7, 8).
Interestingly, recent studies have identified a kinase-indepen-
dent role for Cdc28 in basal transcription (9, 10). Through its
binding partner Cks1, Cdc28 is recruited to thepromoter region of
several genes, including GAL1 and the mitotic regulator CDC20,
whereitrecruitsthe19Sproteasome(9).The19Sproteasomethen
evicts nucleosomes to facilitate transcriptional activation (11).
However,theexactfunctionofCdk1inbasaltranscriptionremains
poorly understood. For example, it is not clear whether the cata-
lytic activity of Cdc28 also may be involved in direct regulation of
the basal transcription machinery, and the genes at which Cdc28
acts to control transcription remain unknown.
Here, we studied the function of Cdc28 kinase activity in basal
transcription. Genome-wide analysis using ChIP followed by high-
throughput sequencing (ChIP-seq) of the localization of Cdc28 to
chromatinrevealedthat itassociateswithhighly transcribedgenes,
including PMA1. Using PMA1 as a model, we found that Cdc28
has a kinase-dependent function in transcription that is partially
redundant with the CTD kinase Kin28. In particular, Cdc28 and
Kin28 cooperate to recruit and phosphorylate RNAPII on CTD-
S5 and to promote mRNA capping. These results identify a role
for Cdc28 kinase activity in regulation of basal transcription at
a subset of genes and reveal a mechanism by which the cell cycle
directly regulates the basal transcription machinery.
Results
CDC28 Interacts Genetically with Genes Involved in Basal Transcription.
We recently identified the genetic network of CDC28 and dis-
covered that genes involved in the regulation of basal transcrip-
tion are overrepresented (12). We screened for additional genetic
interactions between CDC28 and genes involved in transcription.
We made use of the cdc28-as1 allele, which encodes a form of
Cdc28 that is sensitive to the highly specific inhibitor 1-NM-PP1
(13); in spot assays the 1-NM-PP1 IC50is 200 nM for cdc28-as1
mutants, whereas WT cells are resistant to a concentration of at
least 10 μM (12). We crossed this allele into an array of selected
deletion mutants and screened for double mutants that failed to
Author contributions: P.C. and J.M.E. designed research; P.C., V.E., S.L., C.Z., and J.M.E.
performed research; P.C., M.B., and J.M.E. contributed new reagents/analytic tools; P.C.,
S.L., J.M.L., L.A.M.-Z., and J.M.E. analyzed data; and P.C. and J.M.E. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1V.E. and S.L. contributed equally to this work.
2To whom correspondence should be addressed. E-mail: jorrit.enserink@rr-research.no.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1200067109/-/DCSupplemental.
10450–10455
| PNAS
| June 26, 2012
| vol. 109
| no. 26www.pnas.org/cgi/doi/10.1073/pnas.1200067109

Page 2

grow in the presence of low doses of 1-NM-PP1. Interestingly,
cdc28-as1 strongly interacted genetically with genes encoding
components of the RAD6 pathway, the Paf1 complex, and the
Ccr4-NOT complex (Table S1, Fig. 1A, and ref. 12), which are
important for the elongation step of transcription (14–16). CDC28
mRNA and protein levels were not affected in these mutants
(Fig. S1 A–D), excluding the possibility that these genetic inter-
actions are caused by the reduced expression of CDC28. These
data indicate that Cdc28 kinase activity might have a function in
basal transcription.
Accordingly, we observed that CDC28 interacted genetically
with KIN28 and CTK1, which encode CTD kinases, and with
BUR2, which encodes a nonessential cyclin for the essential CTD
kinase Bur1 (Fig. 1B, Fig. S1E, and Table S1). Growth of these
double mutants was impaired even in the absence of 1-NM-PP1,
likely because of the slightly reduced kinase activity of Cdc28-as1
(13). cdc28-as1 mutants also were hypersensitive to overexpression
of FCP1 and RTR1, which encode CTD phosphatases (17, 18), as
well as ESS1, a peptidyl-prolyl isomerase that stimulates CTD-S5
dephosphorylation by Fcp1 (19) (Fig. S1F).
We have shown previously that the rad6Δ cdc28-as1 double
mutant has a cell-cycle defect caused by a defect in transcription
of cyclins (12). This defect could be largely rescued by additional
deletion of WHI5, indicating that Rad6 and Cdc28 function in
parallel pathways to regulate transcription of cyclins. Similarly
the genetic interactions between CDC28 and either KIN28,
CTK1, or BUR2 could be caused by cell-cycle defects. kin28-as,
ctk1Δ, and bur2Δ mutants indeed have a cell-cycle defect, in
particular when combined with the cdc28-as1 allele (Fig. S2 A–
C). However, the cell-cycle defects of these double mutants
could not be rescued by additional deletion of WHI5, indicating
that Cdc28 must have a function that is independent of Whi5 and
that overlaps with Kin28, Ctk1, and Bur2.
Cdc28 Localizes to a Subset of Genes. Cdc28 has been shown pre-
viously to localize to CDC20 and GAL1 (9, 10). To gain further
insight in the identity of the genes that may be regulated directly
by Cdc28, we mapped the genome-wide chromatin localization
of Cdc28 using ChIP-seq. Interestingly, Cdc28 could be detected
at more than 2,000 genes (Dataset S1). For our further studies
we used a stringent, arbitrary cutoff of log2 = 1.5-fold enrich-
ment over background, which included ∼200 genes (Dataset S1
and Fig. S2D), including PMA1 (Fig. 2A). The recruitment of
Cdc28 to PMA1 and several other genes was confirmed by con-
ventional ChIP assays (Fig. 2B). Further analysis of the dataset
showed that 85% of the 200 Cdc28-enriched ORFs fell within
the 10th percentile of the highest expressed genes (Fig. S2E),
indicating that Cdc28 preferentially localizes to highly tran-
scribed genes. Perhaps surprisingly, Gene Ontology (GO) anal-
ysis revealed an overrepresentation of genes with housekeeping
functions, such as carbohydrate metabolism and cell-wall main-
tenance, as well as ribosomal components, but few cell-cycle
genes (Fig. 2C and Dataset S2).
We hypothesized that localization of Cdc28 to these genes is
important for their transcription. Indeed, inhibition of Cdc28 ki-
nase activity significantly reduced PMA1, GLN1, and MDH2
mRNA levels; in contrast, transcription of ACT1, SSE1, and
TEC1, genes that do not recruit Cdc28, wasnot affected (Fig. 2D).
These data indicate that localization of Cdc28 to genes such as
PMA1 may serve to boost transcription during cell-cycle entry,
whenCdc28becomesactive.Indeed,whenWTcellswerereleased
from α factor-induced G1 phase arrest (low Cdc28 activity), the
subsequent reactivation of Cdc28 coincided with increased PMA1
mRNA levels (Fig. 2E). PMA1 mRNA levels peaked slightly later
than CLN2 mRNA and returned to basal levels when CLB2
mRNA peaked, indicating that Cdc28 is particularly important for
PMA1 transcription shortly after the G1–S transition. In contrast,
transcription of ACT1, which is not enriched for Cdc28, remained
unchanged (Fig. 2E). The effect of Cdc28 on PMA1 transcription
was not mediated by SBF or MluI cell-cycle box binding factor
(MBF), because the PMA1 promoter lacks SBF and MBF binding
sites; indeed, we could not detect the SBF/MBF complexes at
PMA1 (Fig. S3A). Moreover, Cdc28 also contributed to PMA1
transcription in cells that had been synchronized with nocodazole
inM phase(Fig.S3BandC),whenSBFandMBFareinactive(1).
These results also exclude the possibility that the effect of Cdc28
on PMA1 transcription is caused by an indirect, cell-cycle stage-
dependent effect. Rather, our data indicate that Cdc28 directly
controls transcription of PMA1, and possibly most of the other
genes identified by ChIP-seq, because promoter sequence analysis
revealed no enrichment for any cell-cycle element (Dataset S3).
Together, these results suggest that Cdc28 boosts transcription of
a subset of genes early in the cell cycle.
Cdc28 and Kin28 Cooperate to Regulate RNAPII at PMA1 The finding
that CDC28 strongly interacted genetically with KIN28 suggests
that Cdc28 might cooperate with Kin28 to regulate basal tran-
scription. Indeed, PMA1 mRNA levels were significantly reduced
in 1-NM-PP1–treated cdc28-as1 single mutants and kin28-as
single mutants (Fig. 3A). However, in the kin28-as cdc28-as1
INO80 complex
RAD6 epistasis group
NAD(+)-dependent deacetylases
Cdc73/Paf1 complex
Proteasome, ubiquitination
Elongator complex
HDACs
SAGA/SLIK complex
HIR complex
ISW1 complex
Ccr4/NOT complex
DSIF complex Other
TRAMP complex
TFIIs complex
Component of multiple complexes
COMPASS, H3-K4 methyltransferase
Histone H3-K79 mehtyltransferase
Histone H3-K56 acetyltransferase
Histone H3-K36 methyltransferase
Nucleosome assembly
Histones
CTD kinase complex; P-TEFb
CTD kinase complex; TFIIH
A
CDC28 cdc28-as1
WT
kin28as
ctk1Δ
bur2Δ
DMSO 100 nM 1-NM-PP1
CDC28
400 nM 1-NM-PP1
CDC28
B
GCN5
ELP4
BRE1
HDA1
SET1
UBP8
ARP8
HIR1
LEO1
SIR1
SDC1
NHP10
TAF14
UBR1
RTF1
LGE1
RAD6
DST1
HIR2
CHD1
RPD3
IKI3
ISW1
PAP2
RPN4
SPT4
KIN28
CTK1
BUR2
SIR4
SIR3
SIR2
CDC73
NOT4
RTT109
RFX1
ASF1
RAD5
CDC28
HOS1
NAS6
DEF1
HHO1
HST2
CCR4
DOT1
RAD18
MMS2
HTZ1
SET2
Strong
Intermediate
Moderate
No genetic interaction
Strength of genetic interaction:
RPN10
cdc28-as1
cdc28-as1
Fig. 1.
transcription. (A) Network diagram summarizing genetic interactions with
cdc28-as1. Genes are represented as nodes and interactions are shown as
edges. Nodes are coloredaccording to their GO biological process (some were
manually annotated from the literature). For the complete dataset see Table
S1. (B) Ten-fold dilutions of cultures were spotted on YPD with DMSO or in-
creasing doses of 1-NM-PP1. Mutants harboring the kin28-as allele can be
inhibited by 1-NM-PP1 (21).
Genetic interactions between CDC28 and genes involved in basal
Chymkowitch et al.PNAS
| June 26, 2012
| vol. 109
| no. 26
| 10451
GENETICS

Page 3

double mutant the mRNA levels already were decreased by 50%
in absence of 1-NM-PP1, and they were reduced further after
treatment with 1-NM-PP1 (Fig. 3A). 1-NM-PP1 treatment had
no additional effect on PMA1 mRNA levels in ctk1Δ cdc28-as1
or bur2Δ cdc28-as1 double mutants (Fig. S3D). These results
show that Cdc28 has a function in regulation of basal tran-
scription that is redundant with Kin28.
To gain further insight into the function of Cdc28 in basal
transcription, we first analyzed the levels of Rpb1 at PMA1 and
SSE1 using ChIP. Interestingly, Rpb1 levels at PMA1 were in-
creased moderately in untreated cdc28-as1 and kin28-as mutants
compared with WT cells (Fig. 3C), although this increase did not
result in higher mRNA levels (Fig. 3A), indicating a potential
defect in transcription elongation. This effect was highly re-
producible and was observed in multiple independently derived
isolates. More importantly, however, the levels of Rpb1 at PMA1
were strongly reduced after 1-NM-PP1 treatment of cdc28-as1
and kin28-as single mutants and kin28-as cdc28-as1 double
mutants (Fig. 3C), showing that the kinase activity of Cdc28 and
Kin28 is important for regulation of RNAPII. In contrast,
treatment of cdc28-as1 single mutants with 1-NM-PP1 had no
effect on Rpb1 levels at SSE1, a gene that does not recruit Cdc28
(Fig. 3D). Taken together, our data show that Cdc28 kinase
activity regulates RNAPII at genes enriched for Cdc28, such as
PMA1, but not at genes devoid of Cdc28, like SSE1.
Next, we tested whether Cdc28 cooperates with Kin28 in
phosphorylating the RNAPII-CTD at PMA1 by ChIP using
phospho-specific CTD-S5 antibodies. As expected, treatment of
WT cells with 1-NM-PP1 did not affect CTD-S5 phosphorylation
(Fig. 3E). Treatment of either the cdc28-as1 or the kin28-as single
mutant with 1-NM-PP1 resulted in a reduction of CTD-S5P at
PMA1(Fig. 3E),which wascaused mostly by the reduction in pan-
RNAPII (Fig. 3C). However, in absence of 1-NM-PP1, phos-
phorylation of CTD-S5 was very low in cdc28-as1 kin28-as double
mutants (Fig. 3E), consistent with the reduced PMA1 mRNA
levels in this mutant (Fig. 3A). Importantly, Cdc28 had no effect
on CTD-S5 phosphorylation at genes devoid of Cdc28, such as
SSE1 (Fig. 3F). We obtained similar results with the temperature-
sensitive kin28-16 allele [Fig. S3E; kinase activity of Kin28-16 is
inhibited at the restrictive temperature (37 °C), but its protein
levels remain unchanged, and promoter occupancy of several
transcription initiation factors remainsintact (16,20)]. Thestrong
defect in S5 phosphorylation in the double mutant even in the
absence of 1-NM-PP1 likely is because both cdc28-as1 and kin28-
as are slightly hypomorphic alleles (13, 21); the combination of
these alleles may have resulted in a much stronger RNAPII
phosphorylation defect than would be expected from either single
mutant alone.
Next, we tested how Kin28 and Cdc28 might affect each other’s
recruitment to ORFs. As expected, we found that Cks1 (a com-
ponent of the Cdc28 holoenzyme) was present at PMA1 but not
SSE1 in vivo (Fig. S3G). Interestingly, we observed that re-
cruitment of Kin28 and Cks1/Cdc28 to PMA1 was mutually de-
pendent (Fig. S3 G and H), and the levels of Cks1 at PMA1
strongly correlated with CTD-S5 phosphorylation (Fig. 3E). We
could not detect any Cks1 at PMA1 in untreated cdc28-as1 kin28-
as double mutants (Fig. S3G), a result that is consistent with the
idea that cdc28-as1 and kin28-as are hypomorphic, redundant
alleles. Taken together, these data indicate that the combined
activity of Kin28 and Cdc28 is required for regulation of RNAPII.
A
B
0
6000
4500
3000
1500
relative abundence (AU)
7500
9000
PMA1
1.84
GLN1
1.34
MDH2
1.28
ACT1
0.62
SSE1
0.4
TEC1
0.006
C
CDC28
cdc28-as1
DMSO
1-NM-PP1
D
E
PMA1 and ACT1 mRNAs (AU)
*
*
**
mRNA levels
PMA1
0
1500
CDC28-TAP
CDC28
0
1500
PMA1
200 nt
ChIP: CT
ChIP: TAP
CDC28 CDC28-TAP
0
.3
.2
.1
% of input
Ch.V
PMA1
Ch.V
PMA1
CDC28
cdc28-as1
CDC28
cdc28-as1
CDC28
cdc28-as1
CDC28
cdc28-as1
CDC28
cdc28-as1
Ribosome,
translation
Metabolism
Histones
Cell wall
integrity
Heat shock
proteins
Other
PMA1
ACT1
CLN2
CLB2
0 30 6012090
200
400
600
800
1000
1200
2000
6000
10000
14000
18000
Time (min)
CLN2 and CLB2 mRNAs (AU)
mRNA levels
Cdc28 ChIP-seq: PMA1 locus
Fig. 2.
antibodies on strains expressing untagged Cdc28 (CDC28) or TAP-tagged Cdc28 (CDC28-TAP). (B) ChIPs were performed using a TAP antibody (ChIP: TAP) or no
antibody (ChIP: CT) in CDC28 and CDC28-TAP strains. Samples were analyzed by qPCR using PMA1-specific primers or primers annealing in an untranscribed
region of chromosome V (Ch.V). Values are given as percentage of input. Error bars indicate SEM of three independent experiments. (C) Graphic repre-
sentation of the GO analysis performed on the ChIP-seq dataset (Dataset S2). (D) WT cells and cdc28-as1 mutants were treated with DMSO or 1 μM 1-NM-PP1
for 1 h, and PMA1, GLN1, MDH2, ACT1, SSE1, and TEC1 mRNA levels were analyzed by qPCR. The values beneath the gene names represent the relative
enrichment of Cdc28 at these genes (log2-fold over background; see Dataset S1). The dashed line separates genes enriched and not enriched for Cdc28. *P <
0.05; **P < 0.03. (E) Transcription of PMA1 but not ACT1 peaks during the early cell cycle. α factor-arrested cells were released in YPD, and mRNA levels were
determined at the indicated time points.
Cdc28 regulates transcription. (A) Graphic representation of ChIP-seq results showing the PMA1 locus. ChIP experiments were performed using TAP
10452
| www.pnas.org/cgi/doi/10.1073/pnas.1200067109 Chymkowitch et al.

Page 4

Combined Activity of Cdc28 and Kin28 Is Required for CTD-S5 Phos-
phorylation.Because Cdc28 is present at a range of highly expressed
genes (Dataset S1 and Fig. S2 D and E), we hypothesized that a
substantial fraction of cellular RNAPII might be susceptible to
regulation by Cdc28, thereby allowing RNAPII analysis by West-
ern blotting. We analyzed global RNAPII phosphorylation in cell
5000
4000
3000
2000
1000
0
PMA1 mRNA
WT
cdc28-as1 kin28-ascdc28-as1
kin28-as
A
B
C
D
E
relatibve abundance (AU)
DMSO
1-NM-PP1
0
.3
.2
.1
0
2
1
Rpb1 ChIP (PMA1)
CTD-S5P ChIP (PMA1)
% of input
% of input
DMSO
1-NM-PP1
DMSO
1-NM-PP1
PMA1
200 nt
SSE1
200 nt
0
.1
% of input
% of input
.2
Rpb1 ChIP (SSE1)
CTD-S5P ChIP (SSE1)
0
2
1
DMSO
1-NM-PP1
*
*
*
WT
cdc28-as1 kin28-as cdc28-as1
kin28-as
WT
cdc28-as1 kin28-as cdc28-as1
kin28-as
WT
cdc28-as1 kin28-as cdc28-as1
kin28-as
WT
cdc28-as1 kin28-as cdc28-as1
kin28-as
F
DMSO
1-NM-PP1
Fig. 3.
mRNA levels were analyzed by qPCR. *P < 0.05. (B) Location of the PMA1 and SSE1 regions amplified in ChIP assays. (C–F) Cells were grown to log phase
and were treated with DMSO or 1 μM 1-NM-PP1 for 1 h, and levels of Rpb1 (C and D) and CTD-S5P (E and F) were determined by ChIP using Rpb1 (4F8) and
CTD-PS5 (3E8) antibodies (6). Values are given as the percentage of inputs after subtracting the values obtained for an untranscribed region in ChrV and
no-antibody controls. Error bars indicate SEM of three independent experiments.
Cdc28 regulates RNAPII during PMA1 transcription. (A) WT cells or cdc28-as1 mutants were treated with DMSO or 1 μM 1-NM-PP1 for 1 h, and PMA1
32P-GST-CTD
75
GST-CTD
A
B
D
----
++++
Wild-type
cdc28-as1ctk1Δ
cdc28-as1
--
++
bur2Δ
cdc28-as1
- Rpb1
- CTD-S2P
- Cdc11
1-NM-PP1
250
1
0
50
250
C
GST-CTD
Cdc28
hCDC2
-
+++
-
+ + ++
+
+
------
----
----
++++
1NM-PP-1
250
- Rpb1
- CTD-S5P
Wild-type
cdc28-as1
kin28-askin28-as
- Cdc11
- CTD-S7P
1
0
50
250
250
1
0
p<0.018
Inputs (10%)
+
GST pull-down
+
-
Extract
250
+++--
+
-
GST-Cdc28
50
Rpb1
*
GST-Cdc28
75
CTD-S5P
CTD-S7P
CTD-S2P
cdc28-as1
ctk1Δ
bur2Δ
*
Fig. 4.
were treated with DMSO or 1 μM 1-NM-PP1 for 1 h, and cell lysates were analyzed by Western blotting with Rpb1, CTD-S5P, CTD-S2P, CTD-S7P, and Cdc11
(loading control) antibodies. Graph depicts quantification of the levels of CTD-S5P, CTD-S2P, and CTD-S7P in 1-NM-PP1–treated samples, first normalized to
pan-Rpb1 and then to DMSO treatment. Error bars indicate SEM of three independent experiments. (C) Immobilized, recombinant GST-Cdc28/Clb5/Cks1 was
incubated with cell lysate, and retained proteins were analyzed by Western blot using GST and Rpb1 antibodies. Asterisk indicates a nonspecific protein cross-
reacting with the Rpb1 antibody. (D) Recombinant GST-CTD was incubated with purified recombinant Cdc28 or hCDC2 in the presence of [γ-32P] ATP. After
SDS/PAGE, the gel was Coomassie stained (Lower), and GST-CTD phosphorylation was analyzed with a phosphorimager (Upper).
Cdc28 and Kin28 collaborate in RNAPII CTD-S5 phosphorylation. (A and B) Cdc28 and Kin28 cooperate to phosphorylate CTD-S5. Log-phase cultures
Chymkowitch et al.PNAS
| June 26, 2012
| vol. 109
| no. 26
| 10453
GENETICS

Page 5

lysates using phospho-specific Rpb1 CTD antibodies (6). Treat-
ment of the cdc28-as1 single mutant with increasing concen-
trations of 1-NM-PP1 slightly reduced global phosphorylation
levels of CTD-S5 (Fig. 4A), in particular at higher doses of 1-
NM-PP1 (or the related inhibitor 1-NA-PP1) (Fig. S4A), but had
no effect on CTD-S7 (Fig. 4A). Consistent with previous findings
(22), 1-NM-PP1 treatment of the kin28-as single mutant de-
creased CTD-S5 phosphorylation as well as phosphorylation of
CTD-S7. However, CTD-S5 phosphorylation was significantly
more reduced in the 1-NM-PP1–treated kin28-as cdc28-as1
double mutant than in either single mutant, whereas CTD-S7
phosphorylation was not affected (Fig. 4A and Fig. S4A). Similar
results were obtained with mutants harboring the kin28-16 allele
(Fig. S4B). These data show that the combined activity of Kin28
and Cdc28 is required for efficient phosphorylation of CTD-S5.
CTD-S2 is phosphorylated by Ctk1 and Bur1 (4). Indeed,
CTD-S2 phosphorylation was strongly decreased in ctk1Δ and
bur2Δ single mutants (Fig. 4B). Global CTD-S2 phosphorylation
in ctk1Δ cdc28-as1 and bur2Δ cdc28-as1 double mutants was
identical to that of the ctk1Δ and bur2Δ single mutants, indicating
that Cdc28 is not involved in CTD-S2 phosphorylation. Taken
together, these results strongly suggest that Cdc28 kinase activity
promotes phosphorylation of the CTD at S5 but not at S2 or S7.
We then tested whether Cdc28 can directly bind and phos-
phorylate the CTD in vitro using recombinant purified GST-
Cdc28/Cks1/Clb5. We found that Rpb1 is present in the pool of
proteins pulled down from cell extracts by GST-tagged Cdc28/
Cks1/Clb5 holoenzyme (Fig. 4C) and that immobilized Suc1 (the
Schizosaccharomyces pombe homolog of Cks1) efficiently pulled
down Rpb1 (Fig. S3F). Furthermore, Cdc28 and human CDC2
efficiently phosphorylated recombinant purified GST-CTD in
in vitro kinase assays (Fig. 4D). We also tested phosphorylation
of peptides in which S2, S5, and S7 were replaced by alanine.
Cdc28 and hCDC2 failed to phosphorylate peptides with S5A
substitutions, but these kinases phosphorylated S7A and S2A
peptides as efficiently as the WT peptide (Fig. S4 C and D). In
contrast, hCDK9 [a CTD kinase with preference for CTD-S2
in vivo but that can phosphorylate both S2 and S5 in vitro (23)]
phosphorylated S2A peptides less efficiently than the WT pep-
tide (Fig. S4F). hCDK9 also completely failed to phosphorylate
S5A peptides [but it should be noted that the general lack of
activity toward S5A peptides by Cdc28, hCDC2, and hCDK9
could be caused by the formation of a structure that precludes
phosphorylation of other residues (24, 25)]. Therefore, although
Cdc28 and hCDC2 phosphorylated S2A and S7A mutant pep-
tides efficiently in vitro, and an intact Ser5 residue appeared to
be required for these kinases to phosphorylate a CTD heptad
dimer peptide, we cannot exclude the possibility that these kinases
also have activity toward Ser2/7 in vitro.
Cdc28 Is Involved in mRNA Capping. CTD-S5 phosphorylation by
Kin28 serves as a mark for the recruitment of the mRNA-cap-
ping machinery (22, 26, 27). Because we found that the activity of
both Cdc28 and Kin28 is required for efficient CTD-S5 phos-
phorylation, we hypothesized that Cdc28 also is involved in
PMA1 mRNA capping. We first used ChIP to analyze the re-
cruitment of the capping enzyme Ceg1 to the promoter of PMA1
and SSE1 in WT and cdc28-as1 cells. Interestingly, treatment of
the cdc28-as1 mutant with 1-NM-PP1 resulted in reduced Ceg1
levels at PMA1 (Fig. 5A) but not at SSE1 (Fig. 5B), indicating
that Cdc28 indeed has a function in mRNA capping. Therefore,
we immunoprecipitated capped mRNA with an anti–7-methyl-
guanosine antibody (28), followed by reverse transcription and
quantitative PCR (qPCR) using primers specific for PMA1 and
SSE1. Strikingly, after 1-NM-PP1 treatment the level of capped
PMA1 mRNA was reduced by 60% in both cdc28-as1 and kin28-
as single mutants as compared with WT cells (Fig. 5C). Capped
PMA1 mRNA was nearly undetectable in 1-NM-PP1–treated
cdc28-as1kin28-asdouble-mutantcells,showingthatbothkinases
contributed to the capping process (Fig. 5C). In contrast, SSE1
mRNA capping was not dependent on Cdc28 activity (Fig. 5D).
Finally, overexpression of ABD1 and CEG1, which encode com-
ponents of the capping machinery, restored growth of cdc28-as1
kin28-as double mutants in the presence of 1-NM-PP1 (Fig. 5E),
adding genetic evidence to the finding that CDC28 and KIN28
have a redundant function in mRNA capping.
Discussion
In this study, we found that Cdc28 is enriched by at least log2 =
1.5-fold in ∼200 protein-encoding genes and that transcription of
at least several of these genes depends on Cdc28 kinase activity.
These findings extend previous studies reporting a kinase-in-
dependent role for Cdc28 in regulation of transcription (9, 10).
The genes enriched for Cdc28 primarily include housekeeping
genes important for cell-wall integrity, energy supply, translation,
A
Cdc28 Kin28
Y1S2P3T4S5P6S7
RNA pol II
Transcription
mRNA capping
P
E
F
C
% of input
KIN28
kin28-as
KIN28 [ABD1]
kin28-as [ABD1]
KIN28 [CEG1]
kin28-as [CEG1]
DMSO
cdc28-as1
CDC28
100 nM 1-NM-PP1
cdc28-as1
CDC28
% of input
(relative to DMSO)
B
D
CEG1CEG1-MycCEG1-Myc
cdc28-as1
.005
.01
0
.015
.02
Ceg1-Myc ChIP (PMA1)
DMSO
1-NM-PP1
CEG1CEG1-MycCEG1-Myc
cdc28-as1
.04
.06
0
.08
.1
% of input
Ceg1-Myc ChIP (SSE1)
DMSO
1-NM-PP1
.02
WT
cdc28-as1
kin28-as
cdc28-as1
kin28-as
0
1
1.2
.2
.4
.6
.8
PMA1 mRNA capping
WT
cdc28-as1
kin28-as
cdc28-as1
kin28-as
0
1
1.2
.2
.4
.6
.8
% of input
(relative to DMSO)
SSE1 mRNA capping
1.4
Fig. 5.
cells were treated with DMSO or 1 μM 1-NM-PP1 for 1 h, and Ceg1-9Myc
recruitment was analyzed by ChIP using an anti-Myc antibody. Values are
given as the percentage of inputs after subtracting the values obtained
for an untranscribed region in ChrV. Error bars indicate SEM of two inde-
pendent experiments. (C and D) Log-phase cells were treated with DMSO
or 1 μM 1-NM-PP1 for 1 h, and capped PMA1 or SSE1 mRNA was analyzed
as described in Experimental Methods. Values were normalized to total
PMA1 and SSE1 mRNA levels and are shown as the percentage of the
values obtained with DMSO-treated samples. Error bars indicate SEM of
three independent experiments. (E) Strains transformed with plasmids
overexpressing either ABD1 or CEG1 were spotted on plates supplemented
with DMSO or 1-NM-PP1 and were incubated at 30 °C until colonies ap-
peared. (F) Model for Cdc28 in regulation of transcription. See Discussion
for details.
Cdc28 and Kin28 function in mRNA capping. (A and B) Log-phase
10454
| www.pnas.org/cgi/doi/10.1073/pnas.1200067109Chymkowitch et al.

Page 6

andchromatinarchitecture(Fig.2CandDatasetS2).Wepropose
that upon cell-cycle entry, Cdc28 directly stimulates transcription
of these genes to maintain their protein levels as the bud grows
and total cell volume increases.
The function of Cdc28 in basal transcription at least partially
overlaps with that of Kin28, because the activity of both kinases
is required for efficient phosphorylation of CTD-S5 and mRNA
capping. We currently do not know how Cdc28 is recruited to
ORFs. Recruitment of Kin28 and the Cks1-Cdc28 holoenzyme
was mutually dependent, and recruitment of Cdc28 depended on
its own activity, raising the possibility that Cdc28 promotes its own
recruitment through a positive feedback loop. It is plausible that
Cdc28 promotes CTD-S5 phosphorylation by recruiting RNAPII
(Fig. 3C) as well as Kin28 (Fig. S3H), which subsequently phos-
phorylates CTD-S5. However, we cannot exclude the possibility
that Cdc28 directly phosphorylates CTD-S5; indeed, hCDC2
phosphorylates the CTD in vivo, and it can phosphorylate CTD-S2
and CTD-S5 in vitro (29, 30), although the consequence has re-
mained unclear.
In conclusion, Cdc28 promotes recruitment of RNAPII and
Kin28, which promotes phosphorylation of RNAPII on CTD-S5,
to regulate directly the expression of a subset of genes during the
early cell cycle (see Fig. 5F for a model).
Experimental Procedures
Yeast Strains and Plasmids. S. cerevisiae strains were grown in standard yeast
extract peptone dextrose (YPD) medium. Strains were derived directly from
the S288c strain RDKY3032 (31) using either standard gene-replacement
methods or intercrossing (see Table S2 for strains and plasmids).
ChIP Assays and ChIP-Seq. Experiments were performed as previously de-
scribed (12, 32), with minor modifications (SI Experimental Procedures).
In Vitro Kinase Assays. Kinase reactions were carried out as described (33),
with minor modifications (SI Experimental Procedures).
Cap-mRNA Immunoprecipitation. Cap-mRNA immunoprecipitations were per-
formedasdescribed(22),withminormodifications(SIExperimentalProcedures).
ACKNOWLEDGMENTS. We thank Drs. S. Hahn, D. L. Bentley, H. Araki,
S. Buratowski, D. Ostapenko, and M. A. Osley for generously providing strains
andreagents;Drs.Y.BarralandM.Timmersforsuggestions;andDrs.F.Coinand
C. Wittenberg for critically reading the manuscript. J.M.E. is supported by
Outstanding Young Investigator Award 180499 from the Norwegian Research
CouncilandbyGrant2010036fromtheNorwegianHealthAuthoritySouth-East.
1. Enserink JM, Kolodner RD (2010) An overview of Cdk1-controlled targets and pro-
cesses. Cell Div 5:11.
2. Wittenberg C, Reed SI (2005) Cell cycle-dependent transcription in yeast: Promoters,
transcription factors, and transcriptomes. Oncogene 24:2746–2755.
3. Skotheim JM, Di Talia S, Siggia ED, Cross FR (2008) Positive feedback of G1 cyclins
ensures coherent cell cycle entry. Nature 454:291–296.
4. Buratowski S (2009) Progression through the RNA polymerase II CTD cycle. Mol Cell
36:541–546.
5. Rodriguez CR, et al. (2000) Kin28, the TFIIH-associated carboxy-terminal domain ki-
nase, facilitates the recruitment of mRNA processing machinery to RNA polymerase II.
Mol Cell Biol 20:104–112.
6. Chapman RD, et al. (2007) Transcribing RNA polymerase II is phosphorylated at CTD
residue serine-7. Science 318:1780–1782.
7. Akhtar MS, et al. (2009) TFIIH kinase places bivalent marks on the carboxy-terminal
domain of RNA polymerase II. Mol Cell 34:387–393.
8. Egloff S, et al. (2007) Serine-7 of the RNA polymerase II CTD is specifically required for
snRNA gene expression. Science 318:1777–1779.
9. Morris MC, et al. (2003) Cks1-dependent proteasome recruitment and activation of
CDC20 transcription in budding yeast. Nature 423:1009–1013.
10. Yu VP, Baskerville C, Grünenfelder B, Reed SI (2005) A kinase-independent function of
Cks1 and Cdk1 in regulation of transcription. Mol Cell 17:145–151.
11. Chaves S, Baskerville C, Yu V, Reed SI (2010) Cks1, Cdk1, and the 19S proteasome
collaborate to regulate gene induction-dependent nucleosome eviction in yeast. Mol
Cell Biol 30:5284–5294.
12. Zimmermann C, et al. (2011) A chemical-genetic screen to unravel the genetic net-
work of CDC28/CDK1 links ubiquitin and Rad6-Bre1 to cell cycle progression. Proc Natl
Acad Sci USA 108:18748–18753.
13. Bishop AC, et al. (2000) A chemical switch for inhibitor-sensitive alleles of any protein
kinase. Nature 407:395–401.
14. Kruk JA, Dutta A, Fu J, Gilmour DS, Reese JC (2011) The multifunctional Ccr4-Not
complex directly promotes transcription elongation. Genes Dev 25:581–593.
15. Squazzo SL, et al. (2002) The Paf1 complex physically and functionally associates with
transcription elongation factors in vivo. EMBO J 21:1764–1774.
16. Xiao T, et al. (2005) Histone H2B ubiquitylation is associated with elongating RNA
polymerase II. Mol Cell Biol 25:637–651.
17. Mosley AL, et al. (2009) Rtr1 is a CTD phosphatase that regulates RNA polymerase II
during the transition from serine 5 to serine 2 phosphorylation. Mol Cell 34:168–178.
18. Kobor MS, et al. (1999) An unusual eukaryotic protein phosphatase required for
transcription by RNA polymerase II and CTD dephosphorylation in S. cerevisiae. Mol
Cell 4:55–62.
19. Kops O, Zhou XZ, Lu KP (2002) Pin1 modulates the dephosphorylation of the RNA
polymerase II C-terminal domain by yeast Fcp1. FEBS Lett 513:305–311.
20. Ng HH, Robert F, Young RA, Struhl K (2003) Targeted recruitment of Set1 histone
methylase by elongating Pol II provides a localized mark and memory of recent
transcriptional activity. Mol Cell 11:709–719.
21. Liu Y, et al. (2004) Two cyclin-dependent kinases promote RNA polymerase II tran-
scription and formation of the scaffold complex. Mol Cell Biol 24:1721–1735.
22. Kanin EI, et al. (2007) Chemical inhibition of the TFIIH-associated kinase Cdk7/Kin28
does not impair global mRNA synthesis. Proc Natl Acad Sci USA 104:5812–5817.
23. Glover-Cutter K, et al. (2009) TFIIH-associated Cdk7 kinase functions in phosphoryla-
tion of C-terminal domain Ser7 residues, promoter-proximal pausing, and termina-
tion by RNA polymerase II. Mol Cell Biol 29:5455–5464.
24. Bienkiewicz EA, Moon Woody A, Woody RW (2000) Conformation of the RNA poly-
merase II C-terminal domain: Circular dichroism of long and short fragments. J Mol
Biol 297:119–133.
25. Jones JC, et al. (2004) C-terminal repeat domain kinase I phosphorylates Ser2 and Ser5
of RNA polymerase II C-terminal domain repeats. J Biol Chem 279:24957–24964.
26. Schroeder SC, Schwer B, Shuman S, Bentley D (2000) Dynamic association of capping
enzymes with transcribing RNA polymerase II. Genes Dev 14:2435–2440.
27. Komarnitsky P, Cho EJ, Buratowski S (2000) Different phosphorylated forms of RNA
polymerase II and associated mRNA processing factors during transcription. Genes
Dev 14:2452–2460.
28. Luhrmann R, et al. (1982) Isolation and characterization of rabbit anti-m3 2,2,7G
antibodies. Nucleic Acids Res 10:7103–7113.
29. Cisek LJ, Corden JL (1989) Phosphorylation of RNA polymerase by the murine ho-
mologue of the cell-cycle control protein cdc2. Nature 339:679–684.
30. Zhang J, Corden JL (1991) Identification of phosphorylation sites in the repetitive
carboxyl-terminal domain of the mouse RNA polymerase II largest subunit. J Biol
Chem 266:2290–2296.
31. Flores-Rozas H, Kolodner RD (1998) The Saccharomyces cerevisiae MLH3 gene func-
tions in MSH3-dependent suppression of frameshift mutations. Proc Natl Acad Sci
USA 95:12404–12409.
32. Lefrançois P, Zheng W, Snyder M (2010) ChIP-Seq using high-throughput DNA se-
quencing for genome-wide identification of transcription factor binding sites.
Methods Enzymol 470:77–104.
33. Chymkowitch P, Le May N, Charneau P, Compe E, Egly JM (2011) The phosphorylation
of the androgen receptor by TFIIH directs the ubiquitin/proteasome process. EMBO J
30:468–479.
Chymkowitch et al.PNAS
| June 26, 2012
| vol. 109
| no. 26
| 10455
GENETICS