(Received for publication, December 13, 1996, and in revised form, February 14, 1997)
From the Departments of Molecular Biology and Cell Biology, The Scripps Research Institute, La Jolla, California 92037
Wee1 tyrosine kinase regulates mitosis by carrying out the inhibitory tyrosine 15 phosphorylation of Cdc2 M-phase inducing kinase. Schizosaccharomyces pombe Wee1 is a large protein, consisting of a C-terminal catalytic domain of ~350 amino acids preceded by a N-terminal domain of ~550 residues. The functional properties of the Wee1 N-terminal domain were investigated by expressing truncated forms of Wee1 in S. pombe. Both positive and negative regulatory domains were identified. Sequences important for Wee1 function were mapped to a central region (residues 363-408). This region is not required for kinase activity or nuclear localization, suggesting it may be involved in substrate recognition. The negative regulatory domain resides in the N-terminal third of Wee1, Wee1 constructs lacking this domain are more effective at delaying mitosis than wild-type Wee1. The negative regulatory domain contains clusters of potential Cdc2 phosphorylation sites. Investigations to monitor the abundance of Wee1 mRNA and protein during the cell cycle were also carried out.
A key aim of cell cycle investigations has been to fully describe the control mechanism that regulates the onset of mitosis, which is often referred to as the mitotic control. In recent years major advances have been made in understanding the mitotic control, largely as a result of the confluence of several experimental approaches, in particular genetic and in vivo biochemical studies carried out with the fission yeast Schizosaccharomyces pombe and in vitro biochemical studies carried out with extracts made from oocytes of the frog Xenopus laevis (1). These studies have led to the following understanding of the mitotic control. The initiation of mitosis is brought about by Cdc2 serine/threonine kinase, acting in obligatory association with one or more B-type cyclins. B-type cyclins that are involved in promoting mitosis are periodically destroyed upon exit from M-phase, thus progression through S and G2 is accompanied by a steady increase in the abundance of Cdc2-cyclin B complex. Activation of Cdc2 also requires phosphorylation of a threonine residue in the T-loop region (threonine 167 in S. pombe Cdc2), although this phosphorylation does not appear to have an important role in regulating the periodic activation of Cdc2 (2). Cdc2-cyclin B is maintained in a repressed state during interphase due to phosphorylation of a tyrosine residue in the N-terminal lobe of Cdc2 (3). In S. pombe phosphorylation of Cdc2 on tyrosine 15 is performed by Wee1 and Mik1 tyrosine kinases, with Wee1 having the dominant role (4-11). In animal cells this phosphorylation is carried out by Wee1 and Myt1 kinases, with the latter enzyme also phosphorylating the preceding threonine residue (7, 12-17). Dephosphorylation of these residues and consequent activation of Cdc2 is largely carried out by Cdc25 dual specificity phosphatase (18-24), although other phosphatases may weakly contribute to the activation of Cdc2-cyclin B (25, 26).
Wee1 and Cdc25, two of the major regulators of Cdc2-cyclin B kinase, are themselves regulated by phosphorylation (1). In S. pombe, Nim1 serine/threonine protein kinase contributes to the induction of mitosis by carrying out inhibitory phosphorylation of Wee1 (27-31). In Xenopus oocytes and human tissue culture cells, Wee1 is the subject of a separate form of inhibitory phosphorylation that occurs during M-phase (15, 32-34). This second mechanism of inhibitory regulation of Wee1 is either directly or indirectly controlled by Cdc2-cyclin B kinase. Cdc25 phosphatase undergoes activating phosphorylation during M-phase, also by a mechanism that is either directly or indirectly controlled by Cdc2 (35-43). This type of Cdc25 regulation has been demonstrated in mammalian tissue culture cell, Xenopus oocyte extracts, and S. pombe.
It has been proposed that the activating phosphorylation of Cdc25 and
inhibitory phosphorylation of Wee1 that occurs during M phase may play
an important part in promoting the onset of mitosis (1). In this
positive feedback model, activation of a small fraction of Cdc2-cyclin
B can be rapidly amplified into total activation by a process in which
Cdc2-Cyclin B catalyzes stimulatory phosphorylation of Cdc25 and
inhibitory phosphorylation of Wee1 (1). The rod-shaped fission yeast
presents one of the best experimental systems to test this positive
feedback model of mitotic control. In S. pombe the size at
which cells initiate mitosis and undergo cell division is quite
sensitive to the gene dose of wee1+,
cdc25+, and nim1+, thus
determination of cell length at division provides an easy measurement
of mitotic timing (4, 5, 23, 30). In rich growth medium wild-type cells
divide at ~15 µm in length, wee1 cells
undergo division at approximately half the length of wild-type and
exhibit a wee phenotype, whereas cells that have extra copies of
wee1+ undergo division at longer cell lengths
that are directly related to wee1+ gene dose
(5).
This report deals with the function and regulation of Wee1 in S. pombe. Fission yeast Wee1 is a large protein, having a molecular mass of ~107 kDa (5). The protein kinase catalytic domain is confined to the C-terminal ~35 kDa of Wee1. Interestingly, the protein sequences of the large N-terminal domains of Wee1 homologs from S. pombe, the budding yeast Saccharomyces cerevisiae, the fruit fly Drosophila melanogaster, Xenopus, and humans are highly divergent (5, 15, 34, 44, 45). For this reason very little is known about which regions of Wee1 other than the catalytic domain are required for function in vivo. In vitro studies have shown that a truncated form of Wee1 containing only the ~35-kDa catalytic domain, produced in insect cells, is unable to phosphorylate Cdc2, although it is able to phosphorylate enolase, suggesting that some part of the N-terminal domain is important for substrate interaction (8). One feature that appears to be shared among Wee1 proteins from different species is a preponderance of serine-proline and threonine-proline dipeptides. SP and TP are the minimal consensus sequences for phosphorylation catalyzed by Cdc2 (46), thus one attractive model is that Cdc2 or another proline-directed protein kinase promotes the initiation of mitosis by carrying out inhibitory phosphorylation of Wee1. A key prediction of this model is that elimination of phosphorylation sites, either by truncation or mutation, should make Wee1 resistant to inhibition by phosphorylation and as a consequence the initiation of mitosis should be delayed or prevented.
S. pombe strains are all derived from
972h and 975h+ (47).
Procedures for genetic studies in S. pombe have been
described (48). YES and synthetic EMM2 media were used to grow S. pombe cells (48). Cell size measurements were determined using an eyepiece micrometer attached to a Zeiss Axioskop 20 microscope with a
100 X objective.
The
wee1+ open reading frame was amplified by
PCR1 from pWee1-1 (5) using the 5
oligonucleotide
5
-CTCCATATGAGATCTTATGGCTTACGGCGGTCC-3
(NdeI
site underlined) and the 3
oligonucleotide
5
-CAGCGGCCGCCAACATTCACCTGCCAATC-3
(NotI site
underlined). The wee1+
NdeI-NotI fragment was inserted downstream of the
thiamine-repressible nmt1 promoter in pREP1-Ha6H (28, 49).
To create wee1-
1 and wee1-
3, 5
primers
5
-CTCCATATGAGATCTAATGCATCTACTGGTGTA-3
and 5
-CTCCATATGAGATCTTCCATGGACTTTTTGAGG-3
(NdeI
sites underlined) were used separately with the 3
primer
5
-CAGCGGCCGCCAACATTCACCTGCCAATC-3
(NotI site
underlined) to amplify fragments of
wee1+. The NdeI-NotI
fragments were cloned into pREP1-Ha6H. pREP1wee1-
4, pREP1wee1-
5,
and pREP1wee1-
11 plasmids were obtained by deletion of the region
between two internal NcoI sites from pREP1wee1-
3, pREP1wee1-
1, and pREP1wee1+ plasmids,
respectively. Plasmids pREP1wee1-
2, pREP1wee1-
6, and
pREP1wee1-
10 plasmids were obtained by deleting the region between two internal NsiI sites from pREP1wee1-
1,
pREP1wee1-
5, and pREP1wee1+ plasmids, respectively. To
create pREP1wee1-
7, the wee1 933-1224 region from
pREP1wee1-
2 was obtained by PCR using the 5
oligonucleotide 5
-TAATGGACCTGTTAATCGAA-3
which hybridized to the nmt1
promoter and the 3
oligonucleotide: 5
-CACCCATGGGTTTTGAAGGAGTACT-3
(NcoI site underlined). The
NdeI-NcoI fragment from this PCR reaction was
cloned into pREP1wee1-
4 that had been digested with NdeI and NcoI. To create pREPwee1-
8, the wee1
1089-1224 fragment was obtained by PCR using the 5
primer
5
-CTCAGATCTGAGTTGTTAACTACTCCC (BglII site
underlined) and the 3
oligonucleotide:
5
-CACCCATGGGTTTTGAAGGAGTACT-3
(NcoI site
underlined). pREPwee1-
8 was constructed by inserting the
wee1 1089-1224 BglII-NcoI fragment
into pREP1wee1-
4. To create pREPwee1-
9, the wee1
1-933 fragment was obtained by PCR using the nmt1 promoter
oligonucleotide 5
-TAATGGACCTGTTAATCGAA-3
and the 3
oligonucleotide:
5
-CAAAAAGTCCATGGATGCATCCTTA-3
(NsiI and
NcoI sites underlined). pREP1wee1-
9 was constructed by
inserting the wee1 1-933 NdeI-NcoI
fragment into pREP1wee1+. Plasmid pREP1wee1-
12 was
constructed by blunt end ligation of pREP1wee1-
2 that had been
digested with NcoI and NotI. All nmt1:wee1 constructs were shuttled into pUR19-ars as
PstI-EcoRI fragments. Plasmid pUR19-ars was
created by removing the ClaI-ClaI ars fragment
from pUR19. To generate single chromosomal copies of wee1
constructs under the control of wee1 promoter,
BglII-KpnI fragments from pREPwee1-
1, -
2,
-
4, and -
7 were mobilized into pWee1-10 (5). Then pWee1-10,
-
1, -
2, -
4, and -
7 were digested with KpnI and
integrated into a wee1-50/mik1::LEU2
strain (strain PR265). The resulting strains are RA1264, RA1265,
RA1263, RA1266, and RA1262, respectively.
To create a strain in which genomic wee1+
encoded a protein having three copies of the HA epitope at the C
terminus, we first created a plasmid encoding
wee1+-3HA. This plasmid, pSP72, has a
5.7-kb EcoRI/XhoI fragment containing wee1+-3HA. At the 3 end of the
wee1 open reading frame is the sequence, GTGATTGTTGGCGGCCGCATCTTTTACCCATACGATGTTCCTGACTATGCGGGCTATCCCTATGACGTCCCGGACTATGCAGGATCCTATCCATATGACGTTCCAGATTACGCTGCTGTCGACTAAACCTTT. The underlined sequence are NotI and SalI
sites, respectively, which were created for the purpose of directional
cloning of the 3HA sequence. The bold sequence encodes three
consecutive copies of the HA epitope YPYDVPDYA (50), which was derived
from pGTEP1 by PCR amplification. Sequence from wee1
immediately flanks the NotI and SalI sites; the
TAA termination codon is italicized. The 5.7-kb
EcoRI/XhoI fragment from pSP72 was used to
transform a wee1::ura4+ strain (PR47)
along with the leu1-32 complementing co-transformation plasmid pART1. Ura
Leu+ transformants were
identified and confirmed to have replaced wee1::ura4+ with
wee1+-3HA by Southern hybridization
and immunoblotting. Strain LW1585 (wee1-3Ha leu1-32
ura4-D18) was used in this study.
Frozen cell pellets were thawed and resuspended in ice-cold LYSIS-1 buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 5 mM EDTA, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 5 µg/ml each of leupeptin, pepstatin, and aprotinin). Chilled glass beads were added to the meniscus and cells were broken by vigorous shaking in a vortexer for 5 min at 4 °C. The cell extract was collected and then centrifuged at 14,000 rpm in a Eppendorf Microfuge at 4 °C for 15 min. The concentration of cell extracts was estimated by determining OD280 and then normalized by adding LYSIS-1 buffer. Immunoblots were probed with anti-HA (12CA5) monoclonal antibodies, anti-Wee1 (number 7451), or anti-Cdc2 (number 1267) rabbit polyclonal antibodies (51). Immunoblots were subsequently incubated with anti-rabbit or anti-mouse IgG antibodies and signals were visualized using the ECL detection system (Amersham). Wee1-Ha6H protein was purified in denaturing conditions using Ni-NTA-agarose beads (Qiagen) according to the manufacturer's protocol. This involved breakage of cells in highly denaturing buffer G (6 M guanidine hydrochloride, 0.1 M sodium phosphate, 50 mM Tris-HCl, pH 8.0) and therefore no phosphatase or protease inhibitors were added.
Northern Hybridization AnalysisCell pellets were resuspended in HE buffer (50 mM Hepes-NaOH, pH 7.9, 5 mM Na2EDTA, 100 mM NaCl) and cells were lysed by vortexing at 4 °C in the presence of glass beads. Supernatants were transferred to tubes containing an equal volume of HES buffer (200 mM Hepes-NaOH, pH 7.9, 10 mM Na2EDTA, 200 mM NaCl, 2% SDS) and incubated at 37 °C for 1 h. RNA was purified by phenol/chloroform extraction and ethanol precipitation. Pellets were resuspended in water and RNA concentration was determined by measuring absorbance at 260 nm. Equivalent amounts of total RNA were loaded onto agarose/formaldehyde gels run and transferred to nylon membranes. Probe was made from the 8-kb genomic fragment containing wee1 and flanking sequences in the plasmid pWEE1-1 (5).
Immunofluorescence MicroscopyA 30% (w/v) solution of paraformaldehyde was dissolved into 20 ml of PEM buffer (100 mM Pipes, pH 6.9, 1 mM EGTA, 1 mM MgSO4) by incubation at 65 °C for 30 min. This solution was diluted 10-fold into a growing culture to fix cells. After 1 h, cells were collected and washed in PEM. Cells were resuspended into PEMS (PEM + 1 M sorbitol) with 0.625 mg/ml Zymolyase-20T (20,000 units/g, Seikagaku Corp.), 0.1 mg/ml NovoZymTM (Novo), and incubated for 0.5 to 1 h at 37 °C. Cells were collected and resuspended into PEMS supplemented with 1% Triton X-100 for 1 min. Cells were then washed three times with PEM and incubated in PEMBAL (1% bovine serum albumin, 0.1% sodium azide, 100 mM L-lysine monohydrochloride, pH 6.9) for 1 h at room temperature. Affinity purified anti-Wee1 (number 7451) antibodies were added in PEMBAL and the samples were incubated overnight at room temperature. After washing three times with PEMBAL, cells were incubated for 1 h at room temperature in PEMBAL containing fluorescein isothiocyanate-conjugated goat anti-rabbit IgG antibody (1/50 dilution) and RNase A (200 µg/ml). After the final wash with PEMBAL, samples were resuspended in phosphate-buffered saline containing 0.1% sodium azide and 4 µg/ml propidium iodide. SlowFadeTM Antifade Kit (Molecular Probes, Inc.) was used for mounting.
The major aim of this study was to evaluate
the functional and regulatory roles of the N-terminal domain of Wee1.
As shown in Fig. 1, the protein kinase catalytic domain
of Wee1 occupies the C-terminal third of the 881-amino acid protein.
There are 15 serine-proline or threonine-proline dipeptides located in
the N-terminal 300 amino acids of Wee1; these are potential sites of
inhibitory phosphorylation carried out by Cdc2-Cyclin B or other
proline-directed kinases. A series of wee1 constructs
encoding proteins having N-terminal truncations or internal deletions
were produced (see "Experimental Procedures") and placed under the control of the thiamine (vitamin B1)-repressible
nmt1 promoter (49). Plasmids containing these constructs
were integrated at the wee1 locus in wild-type and
wee1-50 mik1::ura4+ strains, in each
case leaving the genomic copy of wee1+ or
wee1-50 intact. Note that wee1-50 is a
temperature-sensitive mutation, thus wee1-50
mik1::ura4+ cells are viable and only
slightly smaller than wild-type at the permissive temperature of
25 °C, but undergo mitotic catastrophe when incubated at
temperatures above 30 °C (11). Southern hybridization analysis
confirmed that these strains contained a single integrated copy of the
plasmids. At the outset we noticed that the full-length version of
nmt1:wee1+ caused a moderate cell elongation
phenotype in the wee1-50 mik1 background when these
cells grown in medium that represses the nmt1 promoter (Fig.
1). Septated cells were ~21 µm in length, versus ~15
µm for wild-type cells. This suggests that the repressed nmt1+ promoter is severalfold more active than
the wee1+ promoter, underscoring the fact that
wee1+ expression is normally quite low.
Most of the wee1 constructs caused cell cycle arrest when
expressed from the induced nmt1 promoter in the wild-type
and wee1-50 mik1 backgrounds. These included the
wee1-
2 construct, which lacked the N-terminal 310 amino
acids, as well as the wee1-
7 construct, which lacked the
N-terminal 310 and internal 408-520 amino acids (Fig. 1). Constructs
which failed to cause a cell cycle arrest when highly overexpressed
included wee1-
3, which lacked the N-terminal 460 amino
acids, wee1-
4, which lacked the N-terminal 520 amino
acids, and wee1-
9, which lacked the internal 310-460
amino acids. The failure of constructs such as wee1-
3 to
cause cell cycle arrest could not be attributed to a problem with
protein expression; in fact, the amount of Wee1 protein detected in
cells expressing inactive constructs such as wee1-
3 and
wee1-
4 was much higher than in cells expressing active
constructs such as wee1-
1 and wee1-
2 (Fig.
2), which in turn expressed higher levels of Wee1
protein than cells expressing full-length wee1+
from the nmt1 promoter.2 It appears that
cells expressing active Wee1 constructs experience growth defects
caused by a massive increase in size long before the nmt1
promoter is fully derepressed, this may explain why inactive Wee1
constructs appear to be more abundant that active Wee1 constructs. It
is also possible the shorter Wee1 constructs are more stable proteins.
Nor could the properties of inactive constructs be attributed to loss
of kinase activity per se, as the minimal catalytic domain of Wee1
(residues 520-877, equivalent to wee1-
4), when produced in an insect cell expression system, retains vigorous
autophosphorylation activity (8). This finding was confirmed for the
wee1-
3 and wee1-
4 constructs expressed in
S. pombe, which also retained vigorous autophosphorylation
activity.2 It should be noted that high overexpression of
even the wee1-
3, wee1-
4, and
wee1-
9 constructs caused a moderate elongation of cell
length at division in wild-type cells and these constructs were able to
rescue wee1-50
mik1 (Fig. 1), indicating that these truncated forms of Wee1 were not completely inactive in
vivo.
The inability of the wee1-9 construct to cause cell cycle
arrest when expressed from the nmt1 promoter strongly
suggested that amino acids 310-460 are critical for Wee1 function
in vivo. Indeed, construct wee1-
6, which
consists of the 310-460 region fused to the catalytic domain (residues
520-881), was able to cause cell cycle arrest when overexpressed from
the nmt1 promoter (Fig. 1). The important sequences in the
310-460 region were further delineated by testing the activity of the
wee1-
7 construct, which contained amino acids 310-408
fused to the catalytic domain (Fig. 1). This construct also was active,
causing cell cycle arrest when overexpressed from the nmt1
promoter and rescuing wee1-50
mik1 when cells were grown
in nmt1 repressing medium. We finally tested the activity of
construct wee1-
8, which contained amino acids 363-408
fused to C-terminal protein kinase domain. This construct also caused
cell cycle arrest when expressed from the nmt1 promoter and
rescued wee1-50
mik1 when cells were grown in
nmt1 repressing medium (Fig. 1). These findings indicate
that N-terminal sequences that are critical for Wee1 function are
located in the 363-408 amino acid region.
Cdc2-Cdc13 complex is predominantly
and perhaps exclusively localized in the nucleus during G2
(52-54). We therefore expected that Wee1 would also be a nuclear
protein. This was investigated by carrying out immunolocalization
studies of Wee1. Anti-Wee1 antibodies produced no signal in wild-type
cells, presumably because of the very low abundance of Wee1 protein. To
circumvent this problem we overexpressed wee1-K596L, which
encodes a catalytically inactive form of Wee1 (6, 55). Staining with
anti-Wee1 antibodies produced a strong signal that closely coincided
with the nuclear DNA signal produced by staining with DAPI (Fig.
3, top panels). The nmt1:wee1K596L
construct was expressed from an episomal plasmid that was frequently
lost during mitotic divisions, therefore only a subset of cells stained
with the anti-Wee1 antibody. These observations indicate that Wee1
protein is localized in the nucleus, a finding consistent with
immunolocalization studies of overexpressed Wee1 protein encoded by
wee1-50 (56).
The critical central region of Wee1 contains a stretch of basic amino
acids 387-LSKQHRPRKNT-397 (basic
residues underlined) that potentially could be involved in directing
the nuclear localization of Wee1 (57). This possibility was initially
supported by the observation that Wee1-4 protein, which consists
only of the catalytic domain of Wee1, was predominantly localized in
the cytoplasm (Fig. 3, middle panels). Immunolocalization
analysis of Wee1-
9 protein, which specifically lacks the critical
central region together with some flanking sequence (Fig. 3,
bottom panels), provided a more direct test of this
hypothesis. This analysis showed that Wee1-
9 protein was
predominantly localized to the nucleus, producing a signal that was
very similar to that observed with Wee1-K596L. These findings suggest
that the 363-408 amino acid region is not essential for nuclear
localization, although it may be capable of promoting nuclear
localization in vivo.
The
nmt1:wee1 truncation and deletion constructs divided into
four classes when assayed for their ability to rescue wee1-50 mik1 in nmt1-repressing medium (Fig. 1). The
wee1-
3 and wee1-
4 constructs failed to
rescue wee1-50
mik1, constructs wee1-
6, -
7, and -
8 rescued but did not cause an
elongated phenotype; full-length wee1+ caused
cells to divide at ~21 µm, whereas three of the constructs having
an N-terminal deletion (i.e. wee1-
1, -
2,
and -
5) caused cell elongation phenotype that was more
severe than that caused by full-length wee1+.
The phenotypes caused by the wee1-
1, -
2,
and -
5 constructs suggested that the N-terminal domain of
Wee1 might have an inhibitory effect on Wee1 activity. This was more
carefully investigated by integrating a subset of wee1
truncation constructs, expressed from the wee1+
promoter, into a wee1-50
mik1 background. Integrants
were analyzed by Southern hybridization to confirm that they has a
single copy of the wee1 truncation constructs. The activity
of the constructs was first assayed by determining whether they rescued
the mitotic catastrophe phenotype observed in wee1-50
mik1 cells incubated at the restrictive temperature of
35.5 °C. The wee1-
4 construct failed to rescue the
mitotic catastrophe phenotype, whereas the wee1-
1,
wee1-
2, and wee1-
7 constructs complemented
the defect (Table I). These findings are consistent with
the behavior of the nmt1:wee1 constructs as described in
Fig. 1. Cell size measurements were then carried out with cultures
grown at 25 °C. This analysis showed that cells having single
integrated copies of wee1-
1, wee1-
2, or
wee1-
7 were significantly longer at division (22.4 ± 1.0, 23.9 ± 1.2, and 21.6 ± 0.9 µm, respectively),
than were cells having a single integrated copy of
wee1+ (16.8 ± 2.4 µm). On average the
cells expressing the activated truncation alleles of wee1
were 35% larger than cells expressing wee1+.
These findings show that truncation of the N-terminal ~300 amino acids of Wee1 results in a greater amount of Wee1 activity in vivo. This may be due either to increased specific activity of Wee1 or to greater stability of Wee1 protein.
|
The timing of mitosis is very
sensitive to the gene dose of wee1+, therefore
periodic changes in the cellular concentration of Wee1 would be
predicted to affect mitotic timing. For this reason a secondary aim of
our studies was to determine whether the amount of wee1
mRNA or Wee1 protein change during the cell cycle. We first
measured wee1 mRNA expression during the cell cycle. A
synchronous culture of wild-type cells was prepared by centrifugal
elutriation. This method separates cells on the basis of size,
generating a pure population of small cells that are in early
G2 (58). These cells were cultured for two cell cycles. At
regular intervals samples were collected for RNA preparation and
Northern blot analysis. Cell cycle periodicity was monitored by
counting the septation index. In agreement with previous studies (5),
two wee1 specific mRNA transcripts were detected, a more
prominent ~4.0-kb species and a less abundant ~3.4-kb form (Fig.
4). As expected, these mRNA species were absent in a
strain in which the wee1+ open reading frame was
replaced with the ura4+ gene
(wee1::ura4+). As shown in Fig. 4, the
level of wee1+ mRNA did not appear to
fluctuate during two synchronous rounds of division.
Experiments designed to monitor the abundance of Wee1 protein during
the cell cycle were problematic because of the extreme paucity of Wee1
protein. Wild-type levels of Wee1 could not be reliably detected by
standard immunoblotting methods. Wee1 is a dose-dependent
inhibitor of mitosis, therefore the Wee1 detection problem could not be
solved by increasing wee1+ copy number. We
initially solved this problem by creating a strain in which the genomic
copy of wee1+ was replaced with a version of
wee1+ that encoded Wee1 protein having an
C-terminal extension consisting of three copies of the Ha epitope,
which is recognized by 12CA5 monoclonal antibodies (anti-Ha). Cell size
was not noticeably affected by the replacement of wild-type
wee1+ with the
wee1+-3HA allele. A synchronous
culture of the wee1+-3HA strain was
made by centrifugal elutriation. Immunoblot analysis showed that a Wee1
signal was present throughout the cell cycle, but it appeared to
undergo a moderate oscillation (Fig. 5). The Wee1 signal
was highest in samples 1-3 and 7-11, these samples correspond to S
and G2 phases. In the first cell cycle the initial decrease
in the Wee1 signal, occurring in samples 3-4, immediately preceded the
rise in septation index. Septation follows the onset of mitosis in
S. pombe, thus the decrease in the Wee1 signal closely corresponds to M and G1 phases.
In this study we have focused on several key issues relevant to
the structure and function of S. pombe Wee1 tyrosine kinase. One aim of our studies was to identify regions of the Wee1 N-terminal noncatalytic domain that are important for Wee1 function in
vivo. As noted above, there are no obvious sequence homologies
between N-terminal domains of Wee1-like proteins isolated from
divergent species, therefore it was unclear which sequences of the
N-terminal domain would have functional importance. In fact, our
studies indicate that only a small region of the N-terminal domain is of critical importance. By testing various N-terminal truncations and
internal deletions the critical region was initially localized to amino
acids 310-460. We proceeded to show that the wee1-8 construct, containing only residues 363-408 fused to the catalytic domain (residues 520-881), retained the ability to cause cdc arrest when overexpressed from the derepressed nmt1 promoter.
Moreover, wee1-
8 also rescued wee1-50
mik1
mitotic catastrophe when expressed from the repressed nmt1
promoter.
What function is provided by the 363-408 region of Wee1? Previous
studies showed that a purified ~37-kDa C-terminal construct of Wee1
was able to phosphorylate enolase but not Cdc2-cyclin B in
vitro (8). In contrast, purified full-length Wee1 phosphorylated Cdc2 (when complexed to cyclin B) but did not phosphorylate enolase (8). These results strongly suggest that the Wee1 N-terminal regulatory
sequences are not required for Wee1 protein kinase activity per se. Our
in vivo studies extend these findings by showing that the
wee1-4 construct, which encodes the same protein as the
aforementioned ~37-kDa Wee1 truncation, is extremely defective at
inhibiting mitosis in vivo. Fusion of the 363-408 region to ~37 kDa Wee1 restores in vivo mitotic inhibition activity
to a level that is near wild-type levels. It is likely that the
363-408 sequence is important for substrate recognition. In this
regard it is worth recalling that previous in vitro studies
showed that monomeric Cdc2 is a very poor substrate for Wee1, whereas
Wee1 readily phosphorylates Cdc2 that is bound to cyclin B (8). This
raises the question of whether cyclin-B forms all or part of the
structure recognized by Wee1, or whether the primary affect of cyclin B
is to induce a conformational change in the structure of Cdc2 such that
the region encompassing tyrosine 15 is exposed to Wee1 (and Cdc25).
A second major aim of our studies was to determine whether the
N-terminal region of Wee1 contains domains of inhibitory regulation and
to evaluate the affect of deleting those domains on the mitotic control. We found that truncation of amino acids 1-153 or 1-310 (wee1-1 and wee1-
2) led to a delay of the
onset of mitosis, as shown by an increase of the length of cells at
division. This was observed both for cells having nmt1:wee1
constructs that were grown in nmt1 repressing conditions, as
well as for cells having single integrated copies of the
wee1 constructs expressed from the wee1 promoter.
The observed cell elongation is roughly equivalent to that observed in
cells carrying 2-3 extra copies of wee1+
integrated at the wee1+ locus (5). Thus we may
surmise that by truncating Wee1 we have interfered with the mechanism
of mitotic timing. None of the constructs caused a cell cycle arrest
when expressed from the wee1+ promoter at single
copy. This may indicate that negative regulation of Wee1 mediated
through the N-terminal domain has a modulatory effect but is not
essential for the induction of mitosis. As noted above, our findings do
not necessarily imply that the specific activity of Wee1 is increased
by truncation of the N-terminal region of Wee1, since it is also
possible that the truncation stabilizes Wee1 protein, leading to an
increased level of Wee1 kinase in vivo.
A third aim of our studies was to provide some basic information about the regulation of wee1+ mRNA and protein expression during the cell cycle. We found that the wee1+ mRNA Northern blot signal did not fluctuate during the cell cycle, whereas the Wee1 protein immunoblot signal did undergo a moderate oscillation. The decrease in the Wee1 immunoblot signal appeared to coincide with M-phase, thus our data suggest that Wee1 might be degraded during this phase of the cell cycle. Experiments are currently underway to determine whether Wee1 is destabilized in cells that are arrested in M-phase. Our data do not suggest that Wee1 abundance changes prior to the onset of mitosis, thus we do not believe that oscillation of Wee1 protein plays a role in determining the timing of mitosis.
We thank Clare McGowan, Odile Mondesert, and Kazuhiro Shiozaki and all members of the Scripps Cell Cycle Groups for their advice and encouragement. We thank G. Feiser and Ian Wilson for antibody reagents, George Tokiwa and Bruce Futcher for pGTEP1, and Kinsey Maundrell for pREP and pRIP plasmids.