(Received for publication, April 3, 1997, and in revised form, June 3, 1997)
From Onyx Pharmaceuticals, Richmond, California 94806-5206
Activation of the Cdc2·cyclin B kinase is a pivotal step of mitotic initiation. This step is mediated principally by the dephosphorylation of residues threonine 14 (Thr14) and tyrosine 15 (Tyr15) on the Cdc2 catalytic subunit. In several organisms homologs of the Wee1 kinase have been shown to be the major activity responsible for phosphorylating the Tyr15 inhibitory site. A membrane-bound kinase capable of phosphorylating residue Thr14, the Myt1 kinase, has been identified in the frog Xenopus laevis and more recently in human. In this study, we have examined the substrate specificity and cell cycle regulation of the human Myt1 kinase. We find that human Myt1 phosphorylates and inactivates Cdc2-containing cyclin complexes but not complexes containing Cdk2 or Cdk4. Analysis of endogenous Myt1 demonstrates that it remains membrane-bound throughout the cell cycle, but its kinase activity decreased during M phase arrest, when Myt1 became hyperphosphorylated. Further, Cdc2·cyclin B1 was capable of phosphorylating Myt1 in vitro, but this phosphorylation did not affect Myt1 kinase activity. These findings suggest that human Myt1 is negatively regulated by an M phase-activated kinase and that Myt1 inhibits mitosis due to its specificity for Cdc2·cyclin complexes.
The cyclin-dependent kinases (Cdks)1 are a family of highly conserved serine/threonine kinases that mediate many of the cell cycle transitions that occur during duplication. Each of these Cdk catalytic subunits associates with a specific subset of regulatory subunits, termed cyclins, to produce a distinct Cdk·cyclin kinase complex that, in general, functions to execute a specific cell cycle event (1). The placement of a specific Cdk·cyclin complex function with a particular cell cycle transition has led to the identification of key downstream substrates as well as upstream regulatory mechanisms. Well characterized Cdk·cyclin-regulated cell cycle transitions that have been identified thus far include Cdk4·cyclin D and Cdk6·cyclin D complexes with the pRB-gated G1/S transition, Cdk2·cyclin E and Cdk2·cyclin A complexes with the initiation and progression of DNA replication and Cdc2·cyclin A and Cdc2·cyclin B complexes with the initiation of mitosis (reviewed in Refs. 2-4).
Activation of the Cdk·cyclin kinases during these transitions is controlled by a variety of regulatory mechanisms. For the Cdc2·cyclin B complex, inhibition of kinase activity during S phase and G2 is accomplished by phosphorylation of Cdc2 residues Thr14 and Tyr15, which are positioned within the ATP-binding cleft (5-7). Phosphorylation of Thr14 and/or Tyr15 is believed to suppress catalytic activity by disrupting the orientation of the ATP molecule present within this cleft (8, 9). The abrupt dephosphorylation of these residues by the Cdc25 phosphatase results in the rapid activation of Cdc2·cyclin B kinase activity and downstream mitotic events (10). While phosphorylation/dephosphorylation of the conserved Tyr15 site in Cdk2 likely plays an important role in regulating the G1/S transition, the role that Cdk2 Thr14 phosphorylation plays is less clear (11, 12). Phosphorylation of the corresponding inhibitory tyrosine residue in Cdk4 has also been observed (13, 14). It has been proposed that Thr14/Tyr15 phosphorylation functions to permit a cell to attain a critical concentration of inactive Cdk·cyclin complexes, which, upon activation, induces a rapid and complete cell cycle transition (15). There is evidence in mammalian cells that Thr14/Tyr15 phosphorylation also functions, in part, to delay Cdk activation after DNA damage (7, 13, 16-18).
The Schizosaccharomyces pombe wee1 gene product was the first kinase identified that is capable of phosphorylating Tyr15 in Cdc2 (19). Homologs of the Wee1 kinase have been subsequently identified and biochemically characterized from a wide range of species including human, mouse, frog, Saccharomyces cerevisiae, and Drosophila (20-26). In vertebrate systems, where Thr14 in Cdc2 is also phosphorylated, the Wee1 kinase was capable of phosphorylating Cdc2 on Tyr15, but not Thr14, indicating that another kinase was responsible for Thr14 phosphorylation (21, 27). Direct evidence for the existence of a kinase activity that phosphorylated Cdc2 on Thr14 in the membrane fractions of Xenopus egg extracts has been reported (28). The Xenopus gene encoding this membrane-associated kinase, the Myt1 kinase, has been isolated, and its gene product was shown to be capable of phosphorylating Thr14 and, to a lessor extent, Tyr15 in Cdc2 (29). A human Myt1 homolog displaying similar properties has been recently identified (30) (see "Results"). An apparently unrelated, non-membrane-associated Thr14 kinase activity has also been identified in bovine thymus cytosol (31).
While Xenopus Myt1 clearly plays an important role in regulating the rapid early embryonic cell cycles, its role in cells that have distinct G1 and G2 phases as well as growth and checkpoint controls is not known. To investigate the role of Cdk Thr14 phosphorylation in somatic cell cycle control, we have initiated a biochemical analysis of the human Myt1 kinase. The results presented reveal that Myt1, unlike Wee1, exhibits a restricted substrate specificity; it was capable of phosphorylating Thr14 only on Cdc2·cyclin and not Cdk2·cyclin complexes. We also found that endogenous Myt1 kinase activity was reduced during a drug-induced M-phase arrest, concomitant with hyperphosphorylation of the Myt1 protein, implying that Myt1 may be negatively regulated by phosphorylation.
A 325-base pair DNA fragment,
corresponding to an internal region of EST clones 335241 and 336264, was polymerase chain reaction-generated using oligos
5-AGCAGCCTCTCCAGCAACTGG-3
and 5
-CAGAGAAGACCATGGAGTTCC-3
(5
and 3
primers, respectively) and the first strand cDNA synthesis of fetal
brain RNA (Clonetech) as a template. This DNA fragment was
32P-labeled by random priming (Pharmacia) and used as a
probe to screen approximately 7 × 105 clones of a
human placental Lambda ZAPII cDNA library (Stratagene). Of nine
positives isolated, clone 16 was found to contain the longest cDNA
insert. The complete DNA sequence of this 1.98-kilobase pair insert was
determined using an ABI sequencer.
Frozen pellets of Sf9 cells containing
baculovirus-expressed proteins were thawed and lysed by Dounce
homogenization in 10 mM Hepes, pH 7.5, 1 mM
EDTA, 1% Nonidet P-40, 1 mM DTT, and protease inhibitors
(1 mM Pefabloc (Boehringer), 5 µg/ml aprotinin, 5 µg/ml leupeptin) at 4 °C, followed by ultracentrifugation for 30 min at
42,000 rpm in a TLA45 rotor at 4 °C. The clarified lysates were
aliquoted and stored at 80 °C. Combinations of various Cdk, cyclin, Wee1, and Myt1 lysates were mixed together with an ATP regeneration system (5 × stock: 0.5 mg/ml creatine phosphokinase, 10 mM ATP, 350 mM phosphocreatine, 20 mM Hepes, pH 7.5, 10 mM MgCl2) for
25 min at 25 °C. To determine the Cdk·cyclin kinase activity in
these extracts, 1 µl of the lysate was incubated with 9 µl of a
histone H1 kinase mix (50 mM Tris, pH 7.5, 10 mM MgCl2, 1 mM DTT, 5 µg histone
H1, 50 µM ATP, 1 µCi of [
-32P]ATP
(3000 Ci/mmol) for 10-15 min at 25 °C. This reaction was stopped
with SDS sample buffer, resolved by SDS-PAGE, and analyzed by
autoradiography.
The coding regions of Myt1, full-length Wee1 (clone 1E-12
encoding p98/100Wee1 (22), provided by Dr. Tony Hunter, Salk
Institute), and cyclins were cloned into the baculovirus transfer
vector pAcoG such that the expressed protein contained an N-terminal
EE-epitope (32). The monomeric Cdc2 in Fig. 3C contained a
C-terminal EE-tag and the cyclin B1 complexed with Cdc2-AF in Fig.
2A contained an N-terminal glutathione
S-transferase fusion. All other Cdc2 and Cdk2 subunits used
in this study contained a C-terminal HA tag and have been described
previously (11, 33). The coding region of Wee1, which corresponds to
the p49Wee1 truncated protein (20), was polymerase chain
reaction-generated using the full-length Wee1 clone 1E-12 as a
template. Cdk·cyclin complexes were obtained by mixing and incubating
individual Cdk and cyclin lysates together prior to purification,
except in the case of Cdk4·cyclin D1, which was obtained by
co-infecting Sf9 cells with Cdk4 and cyclin D1-expressing baculoviruses. Additionally, Sf9 lysates containing Cdk-activating kinase (Cdk7 and cyclin H) (34) had been added to Cdc2-wild type and
cyclin B1 lysates during the activation incubation.
Baculovirus-expressed proteins were immunoaffinity purified from Sf9
cell lysates using anti-Glu-Glu (EE) antibody-coupled agarose beads
(32) or affinity-purified using glutathione-agarose beads (Sigma).
In vitro 32P-labeling reactions were performed
by preincubating 1 µl (1-7 µg) of purified Cdk·cyclin complexes
in 4 µl of KAB (50 mM Tris, pH 7.5, 10 mM
MgCl2, 10 mM DTT) containing 25 µM ATP for 20 min at 25 °C, followed by addition of
0.5 µl p49Wee1 (0.78 mg/ml) or Myt1 (2.2 mg/ml) in 4.5 µl of KAB
containing 5 µCi of [-32P]ATP (3000 Ci/mmol) and
incubating at room temperature for 20 min. The reaction was stopped
with sample buffer and resolved by 12% SDS-PAGE.
Mammalian cell lines CEM, U2-0S, and C-33A used in this
study were obtained from the American Type Culture Collection. Various Myt1-containing DNA fragments were cloned into a derivative of mammalian expression vector pcDNA3 (Invitrogen) that contained a
5-HA epitope-coding sequence such that the expressed Myt1 proteins contained an N-terminal HA epitope tag. The expressed Myt1 proteins are
designated as follows: Myt1-1b, residues 10-499; Myt1-1A, residues
75-499; Myt1-K
, residues 10-499 containing a
Lys139 to Ala139 change that was created by
site-directed mutagenesis; Myt1-
T.M., residues 10-499 containing a
deletion of the transmembrane domain (residues
379-397)2 (30). The S100 and
P100 fractions were prepared from Myt1-expressing cell lines that had
been lysed by Dounce homogenization in an ice-chilled hypotonic buffer
(10 mM Hepes, pH 7.5, 5 mM KCl, 1.5 mM MgCl2, protease inhibitors). This lysate was
ultracentrifuged for 30 min at 40,000 rpm in a TLA45 rotor at 4 °C
to obtain the S100 supernatant. The membrane-bound proteins in the P100
fractions were solubilized by extracting this fraction in the same
buffer containing 1% Nonidet P-40 for 30 min at 4 °C. The remaining
insoluble material was removed by ultracentrifugation for 30 min at
40,000 rpm in a TLA45 rotor at 4 °C.
Synchronization of CEM cells by drug treatments was performed by culturing cells for 16-23 h in the presence of either 5 mM hydroxyurea (Calbiochem), 1 mM mimosine (Sigma), or 50 ng/ml nocodazole (Calbiochem). U2-OS cells were treated with 125 ng/ml nocodazole.
AntibodiesMyt1 peptides N2 (MPMPTEGTPPPLSGC, residues 10-23), C1 (CNSEPPRGSFPSFEPRN, residues 472-487), and C2 (CRNLLSMFEDTLDPT, residues 486-499) were coupled to activated keyhole limpet hemocyanin (Pierce) (via the included cysteine residues) and injected into rabbits (two per peptide), resulting in the respective sera N2-1, N2-2, C1-1, C1-2, C2-1, and C2-2. Mouse antibodies against Myt1 protein were raised by injecting three mice with purified Myt1 (residues 9-499). Monoclonal antibody 12CA5 was used to detect HA-tagged proteins.
Myt1 Immunoprecipitation and in Vitro Kinase AssaysMammalian cells were lysed in ELB (50 mM
Hepes, pH 7.4, 250 mM NaCl, 0.1% Nonidet-P40, 1 mM DTT) with protease inhibitors and clarified by
microcentifugation at 8200 × g for 15 min. Myt1 was
immunoprecipitated by incubating 2 µl of anti-Myt1 antibody serum in
0.2-0.4 ml of ELB containing 1 mg of lysate for 30 min at 4 °C,
followed by addition of 25 µl of protein A-Sepharose (1:1 slurry,
Pierce). After 1 h at 4 °C, the beads were pelleted, washed
three times with ELB, and once with KAB. To assess kinase activity, 10 µl of Cdc2·cyclin B1 kinase reaction (1.5 µl of Cdc2·cyclin B1
Sf9 lysates mix, see above, in 8.5 µl of KAB containing 7.5 µM ATP) was added to the beads and incubated for 45 min
at room temperature with periodic mixing. Additionally, 10 µM final concentration of the corresponding peptide was
included during kinase assays of C-terminal antibody
immunoprecipitates. This reaction was then assayed for histone H1
kinase activity by adding 5 µl of a histone H1 kinase mix (50 mM Tris, pH 7.5, 10 mM MgCl2, 1 mM DTT, 5 µg of histone H1, 1 µCi of
[-32P]ATP (3000 Ci/mmol)) and incubating for 15 min at
30 °C with agitation. This reaction was stopped with SDS sample
buffer, resolved by SDS-PAGE, and analyzed by autoradiography. For
quantitation, the labeled bands were excised, and
32P-incorporation was determined by Cerenkov counting. The
mean of individual kinase assays performed on two or three aliquots of
the same immunoprecipitation was determined.
Clarified mammalian cell lysates, immunoprecipitates, and Sf9 cell lysates were fractionated by SDS-PAGE (Novex) and transferred to Immobilon polyvinylidene difluoride membranes (Millipore) using a Mini Trans-Blot cell (Bio-Rad). The filters were blocked with Blotto (15% nonfat dry milk in TBST (0.15 M NaCl, 0.1% Tween 20, and 10 mM Tris, pH 8.0)), washed with TBST, and incubated with primary antibodies (1:1000) and secondary antibodies, horseradish peroxidase-conjugated goat anti-mouse or horseradish peroxidase-conjugated goat anti-rabbit (1:10,000 dilution, Bio-Rad), in blocking solutions. Immunodetection was performed using an enhanced chemiluminescence system (Amersham Corp.).
Human Myt1 Analysis in Xenopus Egg ExtractsAnalysis of Myt1 proteins in Xenopus interphase and mitotic egg extracts was performed essentially as described by Stukenberg et al. (35) except that the Myt1 proteins were produced from pcDNA3-containing Myt1 plasmids in a coupled transcription/translation system (TNT, Promega) in the absence of radiolabel.
To identify a human homolog of the Xenopus MYT1 gene, we performed a TBLASTN search of the EST data base using the amino acid sequence encoded by Xenopus MYT1 as the query (29). As expected, the majority of high scoring matches were protein kinases, owing to the extensive conservation of residues within the catalytic domain of protein kinases (36). However, two nearly identical human ESTs (dbEST Id: 336264 and 335241) were found to match significantly amino acid tracts within the C-terminal (noncatalytic domain) 130 residues of the Xenopus Myt1 kinase. A full-length cDNA clone was isolated and found to encode the recently described human Myt1 kinase (30).
Myt1 inactivates Cdc2 but Not Cdk2 Kinase ActivityTo analyze
human Myt1 activity, we examined its ability to inactivate Cdk·cyclin
complexes in a cell-free extract system. We used the baculovirus
expression system to produce insect cell extracts that individually
contained high levels of either human Cdk, cyclin, Wee1, or Myt1
proteins. It has been shown previously that Cdk·cyclin kinase
activation can be reconstituted upon mixing individual Cdk and
cyclin-containing lysates (33); thus, addition of an inhibitory
activity would be expected to block this activation. We first assessed
the ability of this assay system to detect Wee1 activity. Lysates
containing wild-type Cdc2 were mixed with cyclin B1 lysates and with
either mock-infected or Wee1-containing lysates. After a brief
incubation, an aliquot of this lysate mixture was removed and
Cdc2·cyclin B1 kinase activity was determined by performing a histone
H1 kinase reaction. As shown in Fig.
1A, addition of Wee1 lysates
effectively inhibited Cdc2·cyclin B1 kinase activation. The Wee1
inhibitory activity was dependent on the presence of the
Tyr15 residue in Cdc2, since both Cdc2-F15
(Tyr15 replaced with phenylalanine) and Cdc2-AF
(Thr14 and Tyr15 replaced with alanine and
phenylalanine) proteins complexed to cyclin B1 retained full activity
in the presence of Wee1 lysates.
When Myt1 lysates were combined together with Cdc2 and cyclin B1-containing lysates, a dramatic reduction in Cdc2 kinase activation was also observed (Fig. 1B). Furthermore, Myt1 lysates inhibited both Cdc2-A14 and Cdc2-F15 activation, indicating that Myt1 can phosphorylate either Thr14 or Tyr15 inhibitory sites in Cdc2; although it appears that Myt1 preferentially phosphorylates Thr14 since Cdc2-A14 lysates consistently retained some histone H1 kinase activity.
To determine whether Myt1 can inactivate Cdk2·cyclin complexes, we measured the kinase activity of Cdk2 that had been activated in the presence of Myt1-containing lysates. We used the Cdk2-AF mutant as a reference to control for any non-Thr14/Tyr15 phosphorylation inhibitory activity that may be present in the Myt1 lysate. As shown in Fig. 1C, an equivalent amount of histone H1 kinase activity was present in wild-type Cdk2 and Cdk2-AF lysates that had been mixed with either cyclin A or cyclin E lysates in the presence of Myt1, indicating that neither Cdk2·cyclin A nor Cdk2·cyclin E complexes were inactivated by Myt1. However, when Cdc2 was substituted for Cdk2 in the cyclin A complex, Myt1 significantly reduced histone H1 kinase activity, as compared with mock-treated or the Cdc2-AF control lysate (Fig. 1D). Thus, the Cdk subunit, not the cyclin subunit, is a key determinant for whether a particular Cdk·cyclin complex is recognized by the Myt1 kinase.
Human Myt1 Phosphorylates Cdc2 but Not Cdk2 in VitroTo
analyze Myt1 substrate specificity further, various combinations of
affinity purified Myt1, Wee1, and Cdk·cyclin kinases were incubated
together in an in vitro kinase reaction containing [-32P]ATP followed by SDS-PAGE analysis of the
32P-labeled proteins. As shown in Fig.
2A, both purified Myt1 and Wee1 were capable of phosphorylating wild-type Cdc2 as well as a
kinase-deficient Cdc2 mutant (K
), each of which was
complexed with cyclin B1. The Cdc2-AF mutant was not
32P-labeled by Myt1 or Wee1, indicating that these kinases
were phosphorylating Cdc2 on residue Thr14 and/or
Tyr15. Direct phosphoamino acid analysis and immunoblot
analysis using anti-Cdc2 phospho-specific antibodies showed that human
Myt1, while having some Tyr15 activity, preferentially
phosphorylated Cdc2 on Thr14 (data not shown), consistent
with previous studies of human and Xenopus Myt1 (29,
30).
When Cdk2·cyclin complexes were used as substrates in these kinase reactions, we found that Myt1 failed to phosphorylate Cdk2 that was complexed with cyclins A or E, as well as Cdk4 complexed with cyclin D1 (Fig. 2B). In contrast, Wee1 readily phosphorylated Cdk2 complexed with cyclins A or E, but not Cdk4 complexed with cyclin D1, as has been previously demonstrated (22). Thus, using either a crude lysate system or an in vitro kinase reaction with purified components, we consistently observed that Myt1 only phosphorylated and inactivated Cdk·cyclin complexes in which Cdc2 was the catalytic subunit.
To assess whether Myt1 phosphorylation of Cdc2 requires cyclin association, we performed a Myt1 kinase reaction using monomeric Cdc2 as substrate. As shown in Fig. 2C, Myt1 did not phosphorylate monomeric Cdc2. However, reconstitution of the Cdc2·cyclin B1 complex, by preincubating monomeric Cdc2 with monomeric cyclin B1, enabled Myt1 to phosphorylate Cdc2, indicating that Myt1 can only phosphorylate Cdc2 that is complexed with a cyclin subunit. The truncated Wee1 kinase was capable of 32P-labeling monomeric Cdc2, but this phosphorylation was enhanced when the cyclin-bound form was reconstituted, which is consistent with previous studies (37).
Human Myt1 Is Hyperphosphorylated during MitosisTo examine the regulation of endogenous Myt1, rabbit polyclonal antibodies were raised against three different peptides corresponding to Myt1 residues 10-23, 472-487, and 486-499. These antibodies are designated N2-1, C1-1, and C2-1, respectively. When used for immunoblotting of total cell lysates, cytosolic fractions, and detergent-solubilized membrane fractions prepared from asynchronously growing C-33A human cells, each of the anti-Myt1 antibodies recognized a 66-kDa membrane-associated protein (Fig. 3). This 66-kDa protein comigrated with baculovirus-expressed human Myt1 as well as Myt1 that was produced in an in vitro transcription/translation-coupled reaction (data not shown).
In Xenopus mitotic extracts, the electrophoretic mobility of
Xe-Myt1 is drastically decreased, concomitant with Myt1
hyperphosphorylation and decreased kinase activity (29). To determine
whether human Myt1 exhibits similar cell cycle regulation, we examined
Myt1 in lysates prepared from unsynchronized human CEM cells as well as
cells that had been arrested at G1, S, and M phases of the cell cycle by treatment with mimosine, hydroxyurea, and nocodazole, respectively. Immunoblot analysis using anti-Myt1 C-terminal peptide antibodies (C1-1 and C2-1) showed that Myt1 shifted to a slower migrating form in mitotic lysates (Fig.
4A). The main contribution to
this modification is phosphorylation, since phosphatase treatment of
the mitotic lysate increased the migration of Myt1 to nearly that
observed in G1- and S-phase extracts (Fig. 4B).
At least five laddering Myt1 forms were detectable in the
phosphatase-treated mitotic lysate, indicating that multiple residues
are phosphorylated during mitosis. Further, immunoblot analysis of
cytosolic and detergent-extracted membrane fractions prepared from
nocodazole-arrested CEM cells revealed that human Myt1 remained
membrane-associated during M-phase arrest (Fig. 4C).
Interestingly, anti-Myt1 antibody N2-1 failed to recognize the slower
migrating Myt1 protein in mitotic lysates but was capable of detecting
Myt1 in lysates prepared from unsynchronized as well as mimosine, or
hydroxyurea-arrested CEM cells (Fig. 4A). Phosphatase treatment of the mitotic lysate alleviated this Myt1 masking effect, suggesting that this antibody recognized only unphosphorylated Myt1
(data not shown). Consistent with the immunoblot results, this
N-terminal antibody was incapable of immunoprecipitating mitotic Myt1
protein (Fig. 5A).
Since Myt1 likely spans the endoplasmic reticulum membrane2
(30), it is possible that membrane association is required for Myt1
hyperphosphorylation observed during mitosis. To investigate this
question, we established a U2-OS cell line that stably expressed a
non-membrane-associated Myt1 mutant, Myt1T.M., which lacks the
19-residue transmembrane domain (see "Experimental Procedures"). This Myt1
T.M.-expressing cell line was arrested in mitosis with nocodazole, and the soluble (S100) and detergent-solubilized
particulate (P100) fractions were prepared. Immunoblot analysis of
these fractions revealed that the S100-containing Myt1
T.M. protein
exhibited a reduced mobility shift similar to that seen for the
endogenous Myt1 protein (Fig. 4D) present in mitotic
extracts. This result indicates that a kinase(s) exists in the cytosol
that is capable of phosphorylating Myt1 during mitosis, independent of
Myt1 association with endoplasmic reticulum membranes. We cannot,
however, rule out the possibility that this unidentified kinase is also
present in the endoplasmic reticulum lumen.
To characterize the kinase activity of endogenous human Myt1, anti-Myt1 antibody immunoprecipitates from CEM cell extracts were tested for their ability to inactivate various Cdk·cyclin complexes. Fig. 5A shows that the C-terminal anti-Myt1 antibodies were capable of immunoprecipitating Myt1 from lysates prepared from nonsynchronized and mitotic cells, whereas, as described above, the N-terminal antibody failed to recognize mitotic Myt1. Consistent with our analysis of recombinant Myt1, anti-Myt1 immunoprecipitates dramatically reduced the histone H1 kinase activity of Cdc2·cyclin B1 and Cdc2-F15·cyclin B1 complexes, partially inhibited Cdc2-A14·cyclin B1 complexes, and had no effect on the Cdc2-AF·cyclin B1 complex (Fig. 5B). Likewise, anti-Myt1 immunoprecipitates failed to inactivate Cdk2·cyclin A and Cdk2·cyclin E complexes (Fig. 5C). These results demonstrate that the endogenous Myt1 kinase phosphorylates preferentially Cdc2 on residue Thr14 and does not phosphorylate Cdk2.
To determine if the kinase activity of human Myt1 was cell
cycle-dependent, we compared the kinase activity of Myt1
immunoprecipitates isolated from nonsynchronized and
nocodazole-arrested CEM cells. Using two different C-terminal Myt1
peptide antibodies, Myt1 immunoprecipitated from mitotic extracts had
approximately 40-50% reduced kinase activity (Fig.
6A). Analysis of Myt1
immunoprecipitates using an N-terminal Myt1 peptide antibody showed
that Myt1 was active in asynchronous, as well as mimosine- and
hydroxyurea-arrested cells (Fig. 6C). As discussed above,
this N-terminal peptide antibody failed to immunoprecipitate Myt1 from
nocodazole-arrested cells. These results strongly suggest that Myt1 is
active during G1, S, and G2 but becomes
inactive during M phase, concomitant with Myt1
hyperphosphorylation.
Phosphorylation of Human Myt1 in Xenopus Extracts and by Cdc2·cyclin B1
The similarity of cell
cycle-dependent kinase activity and phosphorylation state
for both human and Xenopus Myt1 suggested that Myt1 kinases
may be regulated by a conserved mechanism. As an initial test of this
possibility we compared the electrophoretic mobility of human Myt1 that
had been incubated in Xenopus interphase and mitotic egg
extracts. As shown in Fig. 7A,
Xenopus mitotic extracts contained an activity that
drastically reduced the electrophoretic mobility of each of the four
human Myt1 proteins tested. The reduced mobility observed for Myt1
truncation mutants lacking the N-terminal 57 residues (-1A,
lane 4) and C-terminal 206 residues2 indicates
that phosphorylated residues responsible for the Myt1 mobility shift
are not contained exclusively within either of these regions.
Since the Xenopus mitotic extracts were derived from
interphase egg extracts that had been induced into an M-phase state by addition of exogenous cyclin B1, it is possible that Cdc2·cyclin B1
may phosphorylate Myt1 directly, perhaps acting as an autocatalytic feedback mechanism such as has been proposed for Xenopus
Cdc25 and Wee1 (24, 38). To address this possibility, we first
determined whether Cdc2·cyclin B1 could affect the electrophoretic
mobility of a kinase-defective Myt1 (Myt1-K). Sf9 lysates
containing Cdc2·cyclin B1 were incubated together with a
Myt1-K
lysate in the presence of an ATP regeneration
system. Immunoblot analysis of this reaction showed that the
Myt1-K
protein migrated with a reduced mobility after
Cdc2·cyclin B1 treatment (Fig. 7B), albeit to a lesser
degree than observed for Myt1 in mitotic extracts derived from human
CEM cells (data not shown). Interestingly, Fig. 7B also
shows that the Myt1 N2-1 antibody exhibited a reduced ability to
recognize Myt1 that had been phosphorylated by the Cdc2·cyclin B1
kinase. The direct phosphorylation of Myt1 by Cdc2·cyclin B1 was
confirmed by including [
-32P]ATP in an in
vitro kinase reaction that contained immunoprecipitated Myt1 and
purified Cdc2·cyclin B1 (Fig. 7C). Phosphoamino acid analysis showed that Cdc2·cyclin B1 phosphorylated Myt1 on both threonine and serine residues (data not shown).
Using an Sf9 lysate system, we examined whether the activity of Myt1 was affected by Cdc2·cyclin B1 phosphorylation. A Myt1-containing lysate was first mixed with a Cdc2-AF·cyclin B1 lysate, followed by subsequent addition of wild-type Cdc2·cyclin B1. The latter Cdc2·cyclin B1 complex was isolated by immunoprecipitation and its kinase activity was determined by performing a histone H1 kinase assay. We observed that phosphorylation of Myt1 by Cdc2-AF·cyclin B1 had no affect on its ability to inactivate Cdc2 kinase activity (data not shown). Hence, phosphorylation by Cdc2 does not appear to directly block Myt1 catalytic activity.
In this study we report several biochemical properties of the human Myt1 kinase. We found that human Myt1 was capable of phosphorylating and inactivating Cdc2 associated with cyclin A or cyclin B1. In contrast, however, Myt1 was unable to phosphorylate or inactivate Cdk·cyclin complexes that function earlier in the cell cycle. In particular, Myt1 did not phosphorylate Cdk4·cyclin D1, Cdk2·cyclin E, or Cdk2·cyclin A complexes. The inability of Myt1 to phosphorylate Cdk4 was not unexpected since Cdk4 contains an alanine residue at the corresponding Thr14 position. Since Myt1 readily phosphorylated Cdc2·cyclin A complexes, Myt1 specificity appears to be determined principally by the Cdk subunit. In vivo, Cdk2 Tyr15 is phosphorylated to much greater extent than Thr14 (11, 12), consistent with our in vitro results that Myt1 does not phosphorylate Cdk2 on Thr14. It is possible, however, that Myt1 can phosphorylate Cdk2 that has been initially phosphorylated on Tyr15, as has been previously suggested (11). Alternatively, additional kinases may be responsible for phosphorylating Thr14 in Cdk2, such as a non-membrane-associated Thr14 kinase activity that is present in bovine thymus cytosol (31). In either case, the significantly greater extent of Tyr15 phosphorylation on endogenous Cdk2 suggests that Wee1, or another Wee1-like kinase, is the major kinase activity that suppresses Cdk2·cyclin kinase activity during the G1/S transition.
The kinase activity of human Myt1 was decreased during M-phase arrest. This decreased activity correlated with hyperphosphorylated and electrophoretically slower forms of Myt1. Indeed, phosphorylation of the high number of TP and SP amino acid doublets in both human and Xenopus Myt1, 10 and 12, respectively, could account for the dramatic Myt1 hyperphosphorylation observed during M phase. Many of these TP/SP doublets conform to the Cdc2·cyclin B1 phosphorylation consensus site (39), consistent with our observation that Cdc2·cyclin B1 readily phosphorylated Myt1 in vitro. Our finding that an anti-Myt1 N-terminal peptide antibody recognized interphase but not M-phase Myt1, or Myt1 phosphorylated by Cdc2·cyclin B1 in vitro, strongly suggests that a residue near the N terminus is phosphorylated during mitosis.
Which kinase(s) phosphorylates Myt1 during mitosis? Our in vitro data demonstrate that Cdc2·cyclin B1 can phosphorylate Myt1 on sites that reduce its mobility on SDS-polyacrylamide gels. While this phosphorylation did not affect Myt1 kinase activity, we cannot rule out the possibility that, in vivo, Cdc2 phosphorylation of Myt1 may be a requisite for further modifications that directly block Myt1 catalytic activity. Additionally, the partial reduction in Myt1 gel mobility induced by Cdc2·cyclin B1 phosphorylation suggests that an additional kinase(s) may phosphorylate Myt1. Further experiments will be required to discern the mechanism of Myt1 regulation during the cell cycle.
We thank Brian Karlak and Eric Vermaas for assistance in identifying the initial human Myt1 clones. We are also very grateful to the following: Dr. Janice Williams for generously providing purified Cdk2·cyclin A and Cdk2·cyclin E complexes; Dr. David Morgan (University of California, San Francisco) for providing numerous Cdc2 and Cdk2 clones and recombinant baculoviruses; Dr. Emma Lees (DNAX) for providing cyclin clones; Dr. Todd Stukenberg (HMS) for providing Xenopus extracts and cyclins; Drs. Michelle Garrett, Nancy Pryer, and Abdallah Fanidi for helpful discussions; David Lowe and Drs. Jim Litts and Barbara Belisle for expressing and purifying the recombinant baculovirus proteins used throughout this study.