 |
INTRODUCTION |
Among the multiple families of related protein kinases that have
been described, the mammalian NIMA related kinases, or Neks, have
proven to be elusive in their functional characterization. For the most
part, the relatedness of these kinases is restricted to sequence
homology within the catalytic domain, as none of the Neks appear to be
functionally related to NIMA. The prototype for this family, NIMA
(never in mitosis, gene
A), was originally identified in Aspergillus
nidulans as a serine/threonine kinase critical for cell cycle
progression. NIMA is specifically required to initiate the cytological
aspects of mitosis. Temperature-sensitive mutants of NIMA or
overexpression of dominant negative forms of NIMA cause cells to arrest
in G2 with uncondensed DNA and interphase microtubules (1).
In addition, overexpression of NIMA in fungus as well as in mammalian
cells results in the early onset of mitotic events, including chromatin
condensation and depolymerization of microtubules (2-4). The ability
of NIMA to functionally regulate mitosis in higher organisms has
suggested the existence of a conserved NIMA-like pathway in eukaryotes.
However, only in the filamentous ascomycete, Neurospora
crassa, and the fission yeast Schizosaccharomyces pombe
have functional homologs been identified (5, 6).
In recent years, several mammalian Neks have been identified. These
typically contain 40-50% sequence identity, which is confined to the
catalytic domain. Nek1, the first mammalian Nek to be described, is
highly expressed in meiotic germ cells, and was proposed to play a role
in meiotic events (7). More recently, positional cloning studies
revealed Nek1 as the gene that is altered in polycystic kidney disease, although its precise function remains unknown (8). Nek2
represents the best characterized mammalian Nek. Nek2 displays
cell-cycle dependent expression similar to NIMA, both being most
abundant at the onset of mitosis (9). Endogenous Nek2 associates with
centrosomes, and overexpression of active Nek2 in cells causes a
pronounced splitting of centrosomes, required for G2/M
transition (10). Nek2 phosphorylates a centrosomal coiled-coil protein,
c-Nap1, and also associates with protein phosphatase 1 (10, 11). These
findings suggest that Nek2 contributes to proper centrosomal function.
Little is known about the function of other mammalian Neks. Nek3
displays a marginal variation through the cell cycle, with highest
expression observed in G0-arrested cells. However, Nek3
overexpression studies have provided no evidence for a cell-cycle
regulated function (12, 13). Like Nek1, Nek4/STK2 is also highly
expressed in germ cells, but shows no cell-cycle regulated expression
(13, 14). Unlike all other NIMA-related kinases, Nek6 and Nek7 encode
their catalytic domains within the C terminus, possibly representing a
novel subfamily of Neks (15). Recently, Nek6 and Nek7 were shown to be
capable of phosphorylating and activating the p70 ribosomal S6 kinase
(16).
Here we report the identification of Nek8 and its candidate substrate,
Bicd2. Nek8 contains an N-terminal Nek-like catalytic domain, a
C-terminal coiled-coil domain and a central region with homology to
guanine nucleotide exchange factor
(GEF)1 domains. Nek8 is
biochemically related to Neks, sharing substrate specificities and
biochemical properties required for maximal kinase activity. In
addition, multimerization and autophosphorylation of Nek8 are important
for its activation. Bicd2 associates tightly with Nek8 and is
phosphorylated by Nek8 in vitro. Bicd2 is a human homolog of
the Drosophila coiled-coil protein Bicaudal D (BicD). In
flies, BicD associates with microtubules and functions to properly localize developmental factors. Bicd2 displays a filamentous cellular staining pattern that is dependent on proper microtubule morphology. Overexpressed Nek8 co-localizes with Bicd2 in cells. Similar to Nek3,
expression and activity of Nek8 exhibits only slight variation across
the cell cycle, with highest activity observed in serum deprived
G0-arrested cells. In nocodazole-treated cells, Nek8 undergoes a mobility shift indicative of phosphorylation, which correlates with a dramatic reorganization of Bicd2 localization.
| |
EXPERIMENTAL PROCEDURES |
Purification of Nek8 and Bicd2--
Lungs were isolated from New
Zealand White rabbits injected with 100 µg/kg huIL-1
15 min prior
to sacrifice. Lungs were washed in cold PBS, freeze-thawed, and
homogenized in 20 mM Tris-HCl, pH 8.5, 30 mM
-glycerophosphate, 10 mM NaF, 25 mM
p-nitrophenyl phosphate, 1 mM DTT, 1 mM EDTA, 1 mM EGTA, 1 mM PMSF, 0.1 mM leupeptin, 10% glycerol, and 0.1% Nonidet P-40. After
centrifugation and ultrafiltration, the resultant supernatant was made
25% in ammonium sulfate. Proteins precipitated by this treatment were
collected by centrifugation. Pelleted precipitates were resolubilized
in buffer A (20 mM Tris-HCl, pH 8.5, 50 mM
-glycerophosphate, 1 mM DTT, 1 mM EDTA, 1 mM EGTA, 1 mM PMSF, 0.1 mM
leupeptin) and dialyzed. After centrifugation the resultant supernatant
was sequentially filtered and the homogenate was applied to a column of
Source15Q (Amersham Bioscience). Nek8 eluted from the Source15Q runs
was diluted 1:2 with buffer A containing 10% glycerol (buffer B) and applied to a column of Reactive Green 19 (Sigma). Fractions containing Nek8 activity were pooled and concentrated to a final volume of 5 ml,
and the Nek8 concentrate was loaded onto a HiLoad Superdex 200 (Amersham Bioscience) column. Fractions containing Nek8 activity were
combined, made 0.1% in Nonidet P-40, incubated at 37 °C for 5 min,
and applied to a column of heparin-Sepharose (Amersham Bioscience).
Fractions containing Nek8 activity were combined, diluted with buffer C
(buffer B containing 0.1% Nonidet P-40), incubated at 37 °C for 5 min, and applied to a HR5/5 MonoQ (Amersham Bioscience) column.
Fractions containing Nek8 activity were pooled, pH adjusted to 7.0, and
applied to a Bio-Sil SEC-400 (Bio-Rad) column equilibrated with buffer
D (20 mM Tris-HCl, pH 7.0, 10 mM
-glycerophosphate, 1 mM DTT, 1 mM EDTA, 1 mM EGTA, 1 mM PMSF, 0.1 mM
leupeptin, 10% glycerol, 0.1% Nonidet P-40). Proteins were eluted
from the column, 0.5-ml fractions were collected and 0.5 µl from each
fraction was assayed for Nek8 activity. An additional 0.5 µl was
removed for SDS-PAGE following incubation with 10 mM Hepes,
pH 7.4, 10 mM MnCl2, 25 µM ATP,
and 20 µCi of [
32P]ATP for 30 min at 30 °C.
Nek8 containing fractions were combined and applied to a 35-µl
microbore MonoQ (Amersham Bioscience) column equilibrated with buffer E
(20 mM Tris-HCl, pH 8.5, 10 mM
-glycerophosphate, 1 mM DTT, 1 mM EDTA, 1 mM EGTA, 1 mM PMSF, 0.1 mM
leupeptin, 10% glycerol, 0.1% Nonidet P-40). Nek8 was applied to the
column and eluted after washing with buffer E. Fractions were
collected, 0.25 µl was removed to assay for Nek8 activity, and 0.25 µl was removed for radiolabeling in a modified kinase assay and
SDS-PAGE analysis as detailed above. Virtually all (~90%) of the
Nek8 activity eluted in fractions 12 and 13.
Sequence Analysis of Nek8 and Bicd2 Peptides--
A portion of
the Nek8 containing fraction was incubated for 90 min at 30 °C with
10 mM MnCl2, 25 µM ATP, and 20 µCi of [32P]ATP, and analyzed by SDS-PAGE and
PhosphorImager analysis (Molecular Dynamics). Nek8 and Bicd2 bands were
excised and subjected to in-gel trypsin digestion (17) performed in 20 mM NH4HCO3, pH 8.0, at 37 °C for
16 h. Peptides were extracted, concentrated, and loaded onto a
capillary C18 column (Vydac) equilibrated with 0.1%
trifluoroacetic acid. Peptides were eluted with an acetonitrile gradient (1%/min) over 90 min. Eluted peptides were monitored spectrophotometrically at 214 nm and peaks collected. Peak fractions were analyzed by matrix-assisted laser desorption ionization using a
Lasermat Mass Analyzer (Finnigan MAT, Bremen, Germany) and/or by triple
quadrapole mass spectroscopy (Finnigan MAT TSQ 700 with electrospray
ionization). The remainder was sequenced by Edman degradation using an
ABI automated protein sequencer.
Cloning of Human Nek8--
During a round of high throughput
sequencing of a human dendritic cell subtracted cDNA library (18),
sequence corresponding to one of the rabbit-derived Nek8 peptides was
identified. Additional sequencing of a particular clone (HH0381)
revealed that this cDNA contained identical sequences identified in
several rabbit Nek8 peptides. HH0381 was used as a probe to screen a
second human dendritic cell cDNA library. One positive clone was
isolated and sequenced, containing an open reading frame of 640 amino
acids, including an initiator methionine. To identify the remainder of the Nek8 open reading frame, a new DNA probe was designed and used to
probe a human dermal fibroblast library in 
gt10. This screen
identified overlapping clones spanning the entire open reading frame of
Nek8. The full Nek8 coding region was amplified and sequenced from both
human dendritic cell and dermal fibroblast cDNA libraries by PCR.
All sequencing was performed on ABI/PE 373 and 377 automated
sequencers. Accession number for Nek8 is AY048580.
Cloning of Human Bicd2--
Three peptide sequences identified
from the rabbit Bicd2 in-gel tryptic digestion analyzed by mass
spectroscopy matched a human 71-amino acid sequence derived from
chromosome 9 (accession number T11454). This was used to probe several
human cDNA libraries. A partial cDNA was found in a human KB
cell library, and used as a probe to screen Raji and human dermal
fibroblast libraries. The entire open reading frame was present in a
composite of four Raji clones and three dermal fibroblast library
clones. Confirmation that the full Bicd2 sequence was present was
determined by PCR and sequencing of a full-length clone from the Raji
library (accession number for Bicd2 is AY052562).
Detection of Nek8 Activity and Mapping Phosphorylation Sites on
-Casein--
Activity of partially purified Nek8 from rabbit lung
and His-Nek8 purified from COS7 cells was measured using
dephosphorylated -casein (Sigma) as a substrate. Kinase was
incubated in 20 mM Hepes, pH 7.4, 10 mM
MnCl2, 10 µg/ml heparin, 25 µM ATP, 1 µCi of [32P]ATP, and 5 µg of dephosphorylated
-casein at 30 °C for 20 min. Reactions were stopped by addition
of sample buffer, analyzed by SDS-PAGE, and quantitated using a
PhsophorImager. One unit of -casein kinase activity was defined as
the amount of Nek8 needed to incorporate 1 pmol of phosphate into
-casein in 1 min under standard assay conditions. To map the
-casein phosphorylation sites, 1 mg of dephosphorylated -casein
in 20 mM Tris-HCl, pH 7.4, 10 mM
MnCl2, 30 µM ATP, 1 mM PMSF, 0.1 mM leupeptin, 0.05 mM E-64, and 100 µCi of
[-32P]ATP was incubated with partially purified Nek8
at 30 °C for 3 h. The reaction was stopped by addition of 100%
trichloroacetic acid on ice. Trichloroacetic acid-precipitated proteins
were centrifuged, washed, air dried, and resolubilized in 100 mM ammonium bicarbonate, pH 8.6, containing 10 µg of
sequencing grade Glu-C (Roche Molecular Biochemicals) and incubated
16 h at 37 °C. The reaction was stopped by addition of
trifluoroacetic acid to a final concentration of 1% and the resultant
peptides separated by reverse phase high performance liquid
chromatorgraphy on a Vydac C18 column developed with an acetonitrile
gradient. Column fractions were analyzed by Cerenkov counting and Edman
degradation using an ABI automated protein sequencer. Peptide
substrates composed of identified sequences were synthesized on an ABI
Model 430 Peptide Synthesizer. Kinetic analysis of potential peptide
substrates resulted in the selection of: RRR-HLPPLLLQSWMHQPHQ for our
standard Nek8 peptide assay.
In Vitro Kinase Assays--
For Nek8 P81 peptide filter binding
assays, 10 µl of Nek8 and 2 mM peptide substrate were
incubated in assay buffer as previously described for 20 min at
30 °C. Reactions were stopped by addition of 100% formic acid.
Assay mixtures were spotted onto Whatman P81 filters, washed with 75 mM H3PO4, and analyzed by Cerenkov counting. One unit of Nek8 peptide activity is defined as that amount
of Nek8 needed to incorporate 1 pmol of phosphate into the peptide
substrate, RRR-HLPPLLLQSWMHQPHQ, in 1 min under standard assay
conditions. Specific activity is defined as units of Nek8 activity/mg
of protein. For kinase assays using non-peptide substrates (myelin
basic protein, histone H1 or H3, -casein, and -casein) 2-5 µg
of exogenous substrate was added to the above reaction mixture and
incubated as above. Reactions were stopped by addition of sample
buffer, analyzed by SDS-PAGE, autoradiography, and PhosphorImager analysis.
Cell Culture and Transfections--
COS7 cells were maintained
in Dulbecco's modified Eagle's medium containing F-12 nutrient
mixture (Invitrogen), 100 µg/ml penicillin, 100 µg/ml
streptomycin, 100 µg/ml glutamine (PSG), and
5% fetal bovine serum. HT29 and HeLa cells were grown in Dulbecco's modified Eagle's medium supplemented with 20% fetal bovine serum and
PSG. Human dermal microvascular endothelial cells (MVEC, Clonetics) were maintained using the manufacturer's recommended medium
(Clonetics). Mouse embryo fibroblasts were maintained in Dulbecco's
modified Eagle's medium containing PSG, 15% fetal bovine serum, 1 mM Na pyruvate (Invitrogen), 100 µM
non-essential amino acids (Invitrogen), and 0.2% -mercaptoethanol.
HeLa cells were transfected with Flag-Nek8 and HA-Bicd2 using FuGENE6
(Roche Molecular Biochemicals) according to the manufacturers
instructions and harvested 48 h after transfection. Transiently
expressed proteins were detected by Western blotting with Flag M2 and
anti-HA (12CA5) monoclonal antibodies.
Antibodies--
The Nek8 rat IgM monoclonal was generated
against His-purified Nek8 using standard techniques. The Nek8 rabbit
polyclonal antibody was raised against a pool of four KLH-coupled
peptides derived from the human Nek8 sequence (amino acids: 70-96,
133-150, 407-430, and 776-796). The Bicd2 rabbit polyclonal antibody
was raised against a pool of two KLH-coupled peptides
corresponding to the N and C termini of human Bicd2 (peptide
9.11, GCARLVMEAQPEWLRAEVKRLSHELAET; peptide 9.792, LTQRLELLELDHEQTRRGRAKACG). For anti-Bicd2 affinity purification, 50 mg
of pooled peptides were coupled to Affi-10 resin (Bio-Rad). After
washing in salt buffer and blocking nonspecific sites on resin, Bicd2
antiserum was passed over the column and eluted with 100 mM
glycine, pH 2.5, into 1 M Tris, pH 9.5. Fractions were
analyzed by Bradford protein assay, SDS-PAGE, and Western blotting. The
unbound flow-through was used as a control for determining specificity
of the Bicd2 antibody.
Purification of His-Nek8 from COS7 Cells--
COS7 cells were
transfected using DEAE/dextran as described (19). Seven days after
transfection, cells were washed in cold PBS and lysed in buffer
containing 50 mM Hepes, pH 7.4, 0.1 mM EDTA,
0.1 mM EGTA, 30 mM -glycerophosphate, 25 mM p-nitrophenyl phosphate, 1 mM
DTT, 0.2 mM Na3VO4, 10% glycerol,
1 mM PMSF, and 0.1 mM leupeptin. After
centrifugation, cleared lysates were loaded onto a Ni-NTA (Qiagen)
column equilibrated with 20 mM NaPO4, pH 7.4, 300 mM NaCl, and 5 mM imidazole. His-tagged
protein was eluted with an imidazole gradient (20-250 mM).
Fractions were collected and analyzed by silver stain SDS-PAGE, Western
blotting, and Nek8 in vitro kinase assays.
Immunoprecipitations and in Vitro Kinase Assays--
HT29 cells
were washed and counted for cell cycle analysis, and the remaining
cells were lysed on ice in a buffer containing 50 mM Hepes,
pH 7.4, 5 mM MnCl2, 10 mM
MgCl2, 5 mM EGTA, 2 mM EDTA, 100 mM NaCl, 5 mM KCl, 0.1% Nonidet P-40, 1 mM PMSF, 1 µg/ml leupeptin, 20 mM
-glycerophosphate, 20 mM NaF, 0.3 mM
Na3VO4, 1% aprotinin, 1 mM DTT,
and 10% glycerol. Similar conditions were used for Nek8
immunoprecipitations from MVEC and HeLa cells. Lysates were normalized
for total protein content and Nek8 was immunoprecipitated using either
Flag-M2 or an anti-Nek8 rat IgM monoclonal antibody, or rat IgM + IgG
as a control. Immune complexes were recovered using Protein G for
Flag-M2, or UltraLink Immobilized Neutravidin beads coated with
biotinylated goat anti-rat IgM + IgG (Jackson IR). Complexes were
washed three times in lysis buffer, once with 20 mM Tris,
pH 7.5, 25 mM -glycerophosphate, 2 mM EGTA,
2 mM DTT, and 1 mM
Na3VO4, and once in buffer containing 20 mM Hepes, pH 7.4, and 10 mM MnCl2.
Peptide and non-peptide substrate kinase assays were performed as
described above.
Immunofluorescence Microscopy--
Mouse embryo
fibroblasts and HeLa cells were grown on glass chamber slides
(Lab Tek). After stimulation, cells were washed once in PBS, fixed with
ice-cold methanol, permeabilized with 0.2% Triton X-100 in PBS, and
blocked in PBS containing 5% bovine serum albumin. Mouse embryo
fibroblasts were incubated with anti-Bicd2 and immune complexes were
detected with Texas Red-conjugated anti-rabbit IgG (Jackson
Immunoresearch). HeLa cells were incubated with anti-Bicd2 and/or Flag
M2 followed by Alexa Fluor 488 anti-rabbit IgG and/or Alexa Fluor 546 anti-mouse IgG (Molecular Probes), respectively. Slides were mounted
with coverslips using Anti-Fade (Molecular Probes) and viewed on a
laser scanning confocal microscope (Zeiss).
Cell Cycle Synchronization--
HT29 cells were collected at
roughly 80% confluency for exponentially growing cells. For
G0 arrest, cells were grown in serum-free medium for
48 h. Cells were synchronized at G1/S by addition of medium containing 20% fetal bovine serum and 5 µg/ml aphidicolin to
G0-arrested cells for 24 h. For collection of S-phase
cells, aphidicolin-arrested cells were washed and released into-drug free medium for 6.5 h. G2/M phase arrested cells were
prepared by incubation in full media supplemented with 0.4 µg/ml
nocodazole for 20 h. Nocodazole-arrested cells were washed and
released for another 20 h for G2 release cells. Cell
cycle profiles were determined by flow cytometric analysis of DNA
content using a FACScan (BD PharMingen) as described (20).
| |
RESULTS |
Isolation, Purification, and Cloning of Human Nek8 and
Bicd2--
The use of exogenous substrates to identify novel enzymatic
activities in response to various stimuli has been a successful method
in the identification of many protein kinases. Previous studies had
reported the identification of a specific -casein phosphorylating
activity induced by IL-1, named -casein kinase (21-23). This novel
activity was shown to be specific for phosphorylating -casein and
not -casein, and was distinct from any known mitogen-activated protein kinase or hsp27 kinase. To identify and characterize this putative protein kinase, we looked for the presence of an
IL-1-activated -casein kinase in extracts derived from the lungs of
IL-1-treated rabbits.
Lung homogenates were generated from naive rabbits or rabbits treated
with IL-1. Extracts were collected and chromatographed sequentially
over multiple columns as described in Table
I. Column fractions were collected and
assayed for -casein kinase activity, and active fractions were
pooled and prepared as loads for subsequent chromatographic steps. Fig.
1 indicates that following the final purification (microMonoQ), virtually all of the activity eluted in 2 fractions, with >75% found in fraction 12 (Fig. 1A). This was in contrast to fractions obtained from the lungs of naive rabbits,
in which no major peak of activity was observed (data not shown).
Analysis of fraction 12 on a silver-stained SDS-PAGE revealed the
presence of several bands, with two prominent bands migrating at 123 and 108 kDa (Fig. 1B). To determine which one of these bands
corresponded to a -casein kinase, fraction 12 was incubated in the
presence of [-32P]ATP in the absence of exogenous
substrate. Fig. 1C indicates that the 123-kDa band was
phosphorylated, either by autophosphorylation or by a co-purifying
kinase. No significant phosphate incorporation was observed in any of
the other bands, either because they were already phosphorylated
in vivo, or were not substrates for the kinase. As many
kinases are known to exhibit autophosphorylating activity, we predicted
that the phosphorylated 123-kDa band corresponded to the putative
-casein kinase. We also focused on the 108-kDa band, as this was the
second most abundant band on the silver-stained gel.
View this table:
[in this window]
[in a new window]
|
Table I
Nek8 purification scheme
Rabbits were treated with IL-1 for 15 min, lungs were collected and
homogenates were sequentially purified using the indicated columns to
obtain the final microMono Q fraction (fraction 12) exhibiting
-casein kinase activity (see "Materials and Methods" for
details). Activity was measured as the ability of a partially purified
fraction to phosphorylate -casein. Specific activity is defined as
units of Nek8 activity/mg of protein. Fold purification is defined as
the ratio of specific activity to the initial specific activity (0.43 units/mg).
|
|
View larger version (28K):
[in this window]
[in a new window]
|
Fig. 1.
Purification of Nek8. A,
activity profile from microbore MonoQ column chromatography
representing the final Nek8 purification step. Fractions were assayed
for their ability to phosphorylate -casein. Activity is defined
under "Experimental Procedures." The majority of -casein kinase
activity eluted in fractions 12 and 13. B, analysis of
fraction 12 by silver staining and autoradiography. Left,
silver-stained gel of fraction 12, indicating the presence of multiple
bands. The most prominent bands migrated at 123 and 108 kDa.
Right, fraction 12 was incubated with
[-32P]ATP for 20 min and analyzed by SDS-PAGE and
autoradiography. Only the 123 kDa becomes phosphorylated.
|
|
To obtain amino acid sequence of the putative autophosphorylating
kinase and the 108-kDa co-purifying protein, in-gel trypsin digestions
were performed and the resultant peptides were analyzed by mass
spectrometry. One of the peptides derived from the 123-kDa phosphorylated band, GAFGEATLYR, was recognized as a conserved portion
from subdomain I of protein kinases (24). This and other peptide
sequences were compared with sequences of cDNA clones derived from
human dendritic cell and dermal fibroblast cDNA libraries. Five
rabbit peptide sequences encompassing 61 amino acids were identified,
and displayed 100% amino acid identity with the human sequence (Fig.
2A). Additional clones were
isolated and used to identify an open reading frame coding for a
979-amino acid protein kinase (Fig. 2A). Similarly, tryptic
peptide sequences derived from the 108-kDa protein were used in the
cloning of a human 824-amino acid molecule. Analysis of the 108-kDa
protein sequence indicated that this molecule was a coiled-coil
protein.
View larger version (54K):
[in this window]
[in a new window]
|
Fig. 2.
Nek8 sequence and structure.
A, amino acid sequence of human Nek8. Catalytic domain
is shown in italics. RCC1-like domain is shown in
bold. Coiled-coil domain is denoted by the boxed
residues. Underlined residues represent the conserved rabbit
peptide sequences obtained from the in-gel trypsin digestion. Residues
331-352 comprised two peptides. Black diamonds correspond
to intron-exon boundaries. Bottom, schematic of Nek8 and Bicd2, indicating the locations of various
domains within the sequences. CC indicates coiled-coil
domains. Nek8, GenBankTM accession number AY048580. Bicd2,
GenBankTM accession number AY052562 (not shown).
B, comparison of the catalytic domains of NIMA-related
kinases, created by the PILEUP program (WGCG) using a pairwise
alignment. Percent identity of the catalytic domains was calculated
using BESTFIT (WGCG). Nek1, Nek4/STK2, and Nek7 are mouse sequences,
Nek2, Nek3, and Nek6 are human sequences. CG10951 is the
Drosophila gene product. C, alignment of the
RCC1-like domain of Nek8 with human RCC1, human RPGR, and UVR8 from
Arabidopsis thaliana. Position of glycine and proline
residues critical for proper arrangement of the seven repeats are shown
in dark gray. Conserved hydrophobic residues also important
for structure determination are indicated in light gray.
Location of the seven blades based on the RCC1 structure are indicated
above the sequences. The second half of blade 7 and the
first half of blade 1 are not shown. To maintain the circular structure
of the propeller, half of the first blade is encoded in the N-terminal
end of the protein, and the other half is on the C-terminal end of the
protein. Residues important for RCC1 exchange factor activity on Ran
are indicated by black boxes on the RCC1 sequence. Note that
these are poorly conserved in Nek8. Amino acid numbers correspond to
Nek8.
|
|
A database search of the 123-kDa protein sequence revealed that the
N-terminal catalytic domain was most similar to NIMA-related kinases
(Fig. 2B). Therefore, we named this protein Nek8. Among the
NIMA-related kinases, the sequence of Nek8 is more closely related to
Nek1, Nek3, and Nek4/Stk2, sharing 41-42% amino acid identity over
the catalytic domain with each (Fig. 2B). An open reading
frame found on human chromosome 14q24.3 (accession number AC007055)
corresponds to the full Nek8 sequence, comprised of 22 exons over
roughly 42,000 nucleotides. A gene product encoding a related protein
kinase with a GEF domain was identified through the
Drosophila genome project (CG10951), and has 32% identity spanning the entire molecule, and 43% identity over the catalytic domain. This may correspond to a Drosophila homolog of Nek8.
The only other known protein with both a kinase domain and a putative GEF domain is the multidomain protein Trio, containing a catalytic domain and two separate Rac- and Rho-specific GEF domains (25). Nek8
also contains a C-terminal coiled-coil domain, similar to Nek1, Nek2,
and NIMA.
The central portion of Nek8 is comprised of a roughly 300-amino acid
region with homology to RCC1. RCC1 functions as a GEF for the
GTP-binding protein Ran and is required for chromosome condensation
(26). The structure of RCC1 is a seven-bladed propeller made up of
-strands (27). Each blade is made up of a repeat of 50-60 residues,
consisting of four glycines, one proline, and hydrophobic residues
which form a structural consensus. RCC1-like domains have been found in
a growing number of proteins, including RPGR (retinitis
pigmentosa GTPase regulator), the
gene responsible for X-linked retinitis pigmentosa (28). However, the
role of RCC1-like domains is not clear (29, 30). Only some appear to
possess GEF activity on Ras superfamily members, others have no
intrinsic GEF activity but appear to be involved in mediating protein-protein interactions (28, 31). An alignment of RCC1-like domains from various proteins indicates that Nek8 possesses the residues required to form a propeller, although structural residues for
the 7th blade are poorly conserved (Fig. 2C).
However, residues identified to be important for exchange factor
activity by alanine scanning and mutagenesis appear not to be well
conserved in Nek8 (32). Within its C terminus Nek8 contains a
coiled-coil domain, as predicted by CoilScan (GCG). Coiled-coil domains
are thought to be involved in mediating protein-protein interactions by
generating either hetero- or homodimers. Indeed, dimerization of Nek2
via one of its coiled-coil domains that resembles a leucine zipper
motif is required for activation (33).
Tryptic peptide sequences derived from the 108-kDa protein were used to
probe cDNA libraries and revealed that this was a human homolog of
the Drosophila coiled-coil protein, BicaudalD (BicD). BicD
is a 90-kDa coiled-coil protein made up of five heptad repeats (34,
35). Studies from dominant and recessive BicD mutations indicate that
BicD associates with cytoskeletal components and functions to properly
transport and/or localize developmental factors important for oogenesis
and abdominal patterning. In addition, the phosphorylation state of
BicD is also important for its function (36). Two independent human
BicD homologs have been described (37, 38). Both human homologs encode
roughly 840 amino acid proteins with a predicted coiled-coil structure,
appear to be ubiquitously expressed and map to different chromosomes
(37, 38). The clone we obtained has 99% identity with the partial cDNA described by Ishikawa et al. (37) known as
KIAA0699. Our clone contains 8 additional residues at the N terminus,
including an initiator methionine, and a stop codon 32 residues before
the predicted stop for KIAA0699, resulting in an open reading frame of
824 amino acids. The other human BicD homolog, Bicd1, is 63% identical
to KIAA0699. Because the 108-kDa protein that co-purified with Nek8
corresponded to a full-length cDNA of KIAA0699, we named this
protein Bicd2.
Biochemical Properties of Nek8--
As Nek8 activity from rabbit
lung displayed the properties of a -casein kinase as well as
sequence homology to Neks, we examined the biochemical characteristics
of Nek8. NIMA-related kinases have also been demonstrated to
preferentially phosphorylate -casein, distinguishing them from
casein kinase II, which phosphorylates -casein. To determine which
sites on -casein were phosphorylated by Nek8, we generated
-casein tryptic peptides following incubation of -casein with
His-Nek8 in the presence of [-32P]ATP. Peptide mapping
indicated that Nek8 preferentially phosphorylated peptides containing
serine residues at positions 57, 124, and 143 of -casein (Fig.
3). In each case the target serine was
positioned immediately downstream from a glutamine residue and upstream
from a large hydrophobic residue. In addition, a proline residue was located six positions upstream from the serine in all cases (Fig. 3A). Based on this mapping data, a series of basic
residue-tagged peptides were synthesized and analyzed as substrates.
Fig. 3B shows that the phosphorylation of one of the
-casein peptides (residues 48-64) revealed a similar pattern to the
phosphorylation of full-length -casein when analyzed by
two-dimensional thin layer chromatography. When the two phosphorylated
substrates were mixed and analyzed the pattern was completely
overlapping, indicating that the sites phosphorylated in the
full-length molecule correspond to those in the peptide (Fig.
3B). We were able to demonstrate phosphorylation of the
peptides containing each of the mapped target sites with
Km values in the 300-500 µM range, as
compared with a Km of 24 µM for
full-length -casein (data not shown). For comparison, purified NIMA
phosphorylates -casein with a Km of 38 µM (39). The observation that all three phosphorylation
sites in -casein had conserved sequences surrounding them suggested
that this might represent a minimal consensus recognition site for
substrate phosphorylation. However, generation of peptides in which the
conserved glutamine residue, the upstream proline residue, or the
adjacent hydrophobic residue was altered, had no effect on the peptide
substrate phosphorylation by Nek8 (data not shown). In subsequent
peptide kinase assays, the Ser-143 containing peptide
(RRR-HLPPLLLQSWMHQPHQ), referred to as the QSW peptide, was used.
View larger version (50K):
[in this window]
[in a new window]
|
Fig. 3.
Nek8 phosphorylates
-casein at three similar sites.
A, alignment of three -casein derived peptides that were
phosphorylated by Nek8. Numbering is based on a -casein variant
(GenBankTM accession number A59068). Residues in
italics were not part of the synthesized peptides, but are
shown to illustrate the presence of similar residues downstream of each
phosphorylation site. Asterisk indicates the serine
phosphorylated in all three peptides, as determined by P81 peptide
filter binding kinase assays. Residues identical in all three are in
gray boxes. Conserved residues or identical residues in two
of the three peptides are indicated by the open boxes. B,
phosphorylation of the -casein peptide overlaps phosphorylation
of full-length -casein by Nek8. Purified His-Nek8 was incubated with
either the -casein peptide 48-64 or full-length -casein in the
presence of [-32P]ATP. Substrates were isolated and
subjected to partial acid hydrolysis before analysis by two-dimensional
thin layer chromatography as described (48). O indicates
origin. Numbers were arbitrarily assigned to phosphorylated
spots. Left panel, migration pattern of partially hydrolyzed
phosphorylated -casein peptide. Middle panel, migration
pattern of partially hydrolyzed phosphorylated full-length -casein.
Right panel, migration pattern of a mixture of the 48-64
peptide with full-length -casein.
|
|
As other -casein kinases, including Nek1 and Nek2, have displayed a
preference for Mn2+ over Mg2+ to satisfy their
divalent cation requirements, we tested whether Nek8 might also have
such a preference (9). In conditions when either -casein or the QSW
peptide was used as a substrate, Nek8 activity was stimulated further
with Mn2+ over Mg2+. In addition, decreasing
concentrations of Mn2+ (1-5 mM
versus 10 mM) resulted in an increase of Nek8
activity (data not shown). Similar results have been observed for Nek2 (9, 40). Additionally, Nek8 was unable to utilize
[-32P]GTP as a phosphate donor, was not inhibited by
heparin, and its activity was sensitive to increasing concentrations of
detergent (data not shown).
A range of protein and peptide substrates were assayed for their
ability to be phosphorylated in vitro by His-Nek8. Nek8 also phosphorylated myelin basic protein, histone H1, and histone H3, but to
a lesser degree than -casein (data not shown). Collectively, the
overlapping biochemical properties, substrate specificities, and
sequence similarity of Nek8 with other NIMA-related kinases is
consistent with its role as a newly identified member of this group of kinases.
Nek8 Has Autophosphorylating Activity--
The initial analysis of
purified Nek8 from rabbit lung suggested that it had
autophosphorylating activity, as incubation of the microMono Q fraction
containing -casein kinase activity with [-32P]ATP
resulted in its phosphorylation (Fig. 1B). Similarly,
incubation of His-Nek8 with [-32P]ATP also results in
autophosphorylation. We performed phosphoamino acid analysis on
His-Nek8 purified from COS7 cells and found that >90% of the
phosphate was incorporated on serine (Fig.
4A). This has also been
demonstrated for Nek2 (9). For NIMA, autophosphorylation has been
proposed to occur on a conserved threonine residue (Thr-199) located
within the activation loop of the kinase, and mutation of this site
renders the kinase functionally inactive (5). We tested to see if Nek8
might also phosphorylate its own activation loop by synthesizing an
activation loop peptide, spanning residues 199-222 of Nek8. Fig.
4B shows that Nek8 was able to phosphorylate the activation
loop peptide, and this corresponded to about half the phosphate
incorporation observed with the QSW peptide. To determine whether the
phosphorylation on the activation loop peptide occurred on threonine,
activation loop peptide mutants were synthesized in which one or both
threonine residues (Thr-210, Thr-214) were mutated to alanine.
Alteration of Thr-214 on the T214A peptide had no effect on its
phosphorylation by Nek8, however, mutation of Thr-210 significantly
reduced the ability of the T210A peptide mutant to be phosphorylated
(Fig. 4B). A T210A,T214A double mutant peptide was
phosphorylated to the extent of the T210A single mutant. Based on
sequence alignments, NIMA and all mammalian Neks except for Nek6 and
Nek7, which have a serine substitution, contain a threonine at this
position, suggesting that this residue is an important regulatory site.
Our data suggests that one site for Nek8 autophosphorylation may be
Thr-210 within the activation loop. It will be important to test
whether mutation of this site within the context of full-length Nek8
has any effects on autophosphorylation and/or autoactivation.
View larger version (53K):
[in this window]
[in a new window]
|
Fig. 4.
Nek8 has autophosphorylating activity.
A, Nek8 is predominantly phosphorylated on serine.
Phosphorylated His-Nek8 purified from COS7 cells was cut out of a dried
SDS-PAGE and subjected to phosphoamino acid analysis as described under
"Experimental Procedures." An autoradiograph of the cellulose plate
is shown. Arrow denotes origin, and P denotes
partially hydrolyzed peptides. The positions of phosphoserine
(S), phosphothreonine (T), and
phosphotyrosine (Y) are indicated. B, Nek8
phosphorylation of activation loop peptides. Peptide filter binding
kinase assays were performed in duplicate with purified His-Nek8 and
the indicated peptide substrates. Wild-type activation loop peptide
sequence was as follows:
199KKLNSEYMAETLVGTPYYMSPE222. C,
autophosphorylation induces a reduction-insensitive alteration in the
electrophoretic mobility of Nek8. His-Nek8 purified from COS7 cells was
incubated in the absence or presence of [-32P]ATP for
the indicated times, then analyzed on a silver stained reducing
SDS-PAGE and by autoradiography. Lower panel is a
silver-stained gel indicating the mobility of Nek8 in the absence of
ATP. Upper panel is a silver-stained gel indicating the
mobility of Nek8 in the presence of ATP. An autoradiograph of the
upper panel shows the same pattern as the silver-stained gel
(not shown). At increasing times, monomeric Nek8 undergoes a minor
shift indicative of phosphorylation. Nek8 also forms high-molecular
weight multimers in the presence of ATP which are stable in the
presence of reducing agents and SDS.
|
|
Although Nek8 phosphorylation on Thr-210 of the activation loop peptide
may represent an autophosphorylation site, the bulk of Nek8 phosphate
incorporation occurs on serine. One possibility is that Nek8
autophosphorylation on Thr-210 activates the kinase such that it may
then autophosphorylate on serine residues within the molecule,
resulting in full activation.
While analyzing Nek8 autophosphorylation activity in vitro,
we noticed that incubation of Nek8 with ATP for increasing times resulted in the formation of reduction-insensitive complexes migrating at roughly twice the molecular weight or higher (Fig. 4C).
This change in mobility was ATP dependent, as incubation of Nek8 in the
absence of ATP had no effect on its mobility. This result suggests that
Nek8 autophosphorylation leads to dimerization. Dimerization of Nek2
has also been observed, and has been shown to mediate
trans-autophosphorylation of Nek2, resulting in increased kinase
activity (33). This raises the possibility that autophosphorylation and
higher order complex formation may also be involved in regulating Nek8
kinase activity in vivo.
Nek8 Is Not Regulated by IL-1--
The isolation and purification
of Nek8 was originally undertaken to identify a previously described
novel -casein kinase activity induced specifically by IL-1 (21, 22).
Nek8 was subsequently cloned based on tryptic peptide sequence obtained
from purified fraction 12 containing IL-1 induced -casein kinase
activity. To determine whether the Nek8 molecule that was cloned
corresponded to the -casein kinase activity identified in IL-1
stimulated rabbit lung, we tested to see if human Nek8 could be
activated by IL-1. Primary human dermal MVEC were treated with PMA,
IL-1, or tumor necrosis factor for 15 min or left untreated. Endogenous Nek8 was immunoprecipitated from cells and assayed for in
vitro kinase activity using the -casein-derived QSW peptide as
a substrate. Fig. 5 indicates that IL-1
was only a mild activator of Nek8, increasing Nek8 activity 1.4-fold
when compared with unstimulated cells. PMA and tumor necrosis factor
treatment had virtually no effect on Nek8 activity. Activation of the
ERK and JNK mitogen-activated protein kinases in response to PMA or
cytokines was as expected, indicating that these cells were responsive
to these stimuli (Fig. 5). This result suggests that although Nek8 is
indeed a potent -casein kinase, its activity is unlikely to be
strongly regulated by IL-1. Other assays to demonstrate IL-1 mediated
Nek8 activation have shown similar results (data not shown). Fig.
1B indicated that several proteins were present in
microMonoQ fraction 12, of which the 123-kDa protein, or Nek8, was most
abundant. This protein was also the only one to display any autokinase
activity, supporting the idea that this band was indeed the
unidentified kinase. From our current results, however, we cannot rule
out the possibility that one of the less abundant proteins with lower molecular weight visible by silver stain corresponded to the IL-1 regulated activity. To date, we have been unsuccessful in identifying any Nek8 specific activators, and overexpression of Nek8 in cells results in its constitutive
activation.2 One possibility
is that Nek8 activity is regulated by its association with other
proteins or by dimerization. This could affect its localization in
cells and thereby influence its ability to become activated. This is
consistent with the idea that a binding partner may regulate Nek8
activity either by masking activation domains or by altering
localization of Nek8 within the cell.
View larger version (41K):
[in this window]
[in a new window]
|
Fig. 5.
Regulation of Nek8 activity by IL-1.
Primary dermal MVEC were stimulated with 1 mg/ml PMA, 100 ng/ml huTNF,
or 50 ng/ml huIL-1 for 15 min. Endogenous Nek8 was
immunoprecipitated with a rat anti-Nek8 monoclonal antibody, or rat
anti-IgM as a mock immunoprecipitation. To measure Nek8 kinase
activity, immunoprecipitated Nek8 was incubated with QSW peptide and
[-32P]ATP, and analyzed by SDS-PAGE and
autoradiography. Phosphorylated QSW peptide is indicated. Expression of
endogenous Nek8 (anti-Nek8 rabbit polyclonal antibody), phospho-JNK,
JNK, phospho-ERK2, and ERK2 (Cell Signaling Technology) in the total
lysates are indicated. Similar results were obtained in three
independent experiments.
|
|
Bicd2 Is an in Vitro Substrate for Nek8--
Studies from
Drosophila have indicated that BicD exists in multiple
phosphorylation states, and that phosphorylation is required for proper
localization of BicD during oogenesis (36). To see if Bicd2 could be
phosphorylated by Nek8, purified His-Nek8 and His-Bicd2 made separately
in COS7 cells were incubated in the presence of
[-32P]ATP, either alone or together, and then analyzed
by SDS-PAGE and autoradiography. Fig.
6A indicates that as
previously noted, Nek8 was capable of autophosphorylation. Incubation
of Bicd2 with ATP alone had no effect, however, when Nek8 and Bicd2
were incubated together, both proteins became phosphorylated,
demonstrating that Nek8 can directly phosphorylate Bicd2 in
vitro.
View larger version (37K):
[in this window]
[in a new window]
|
Fig. 6.
Nek8 phosphorylates Bicd2 in
vitro, and Nek8 and Bicd2 associate in
vivo. A, His-Nek8 and His-Bicd2 were
independently purified from COS7 cells with Ni-NTA, eluted with
imidazole and dialyzed. Purified lysates were mixed together in
the presence of [-32P]ATP and
MnCl2 and incubated at 30 °C for 20 min. Lysates were
analyzed by SDS-PAGE and autoradiography. B, HeLa cells were
transiently transfected with Flag-Nek8 and HA-Bicd2 and harvested after
48 h. Left panel, HA Western blot of HA-Bicd2 and
Flag-Nek8 immunoprecipitated from cells using anti-HA (12CA5) or
anti-Flag antibodies, respectively, or mouse IgG as a mock IP.
Middle panel, anti-Flag Western blot indicating expression
of Flag-Nek8 in HeLa lysate. Right panels, Western blots of
untransfected HeLa cell extracts with either anti-Bicd2 polyclonal sera
or unbound flow-through fraction from the affinity column.
C, endogenous Nek8 and Bicd2 associate in
vivo. Nek8 was immunoprecipitated from primary dermal MVEC with
rat anti-Nek8 monoclonal antibody, or rat anti-IgM as a mock IP.
Left panel, Bicd2 Western blot with anti-Bicd2 polyclonal
antibody, indicating the presence of Bicd2 in the Nek8
immunoprecipitation, but not the mock immunoprecipitation. Right
panel, Nek8 Western blot with anti-Nek8 polyclonal antibody,
indicating the presence of Nek8 in the Nek8 IP, but not the mock
immunoprecipitation.
|
|
Nek8 and Bicd2 Associate in Vivo--
The co-purification of Bicd2
with Nek8 from rabbit lung suggested that these proteins could
associate in vivo. To further test this, we transiently
expressed Flag-Nek8 and HA-Bicd2 in HeLa cells. Flag-Nek8 was
immunoprecipitated and analyzed by Western blotting for the presence of
associated HA-Bicd2. Fig. 6B shows that Bicd2 was associated
with Nek8, and was not detectable in a mock immunoprecipitation as a
control. We also compared the signal specificity of anti-Bicd2 with the
unbound flow-through fraction generated from an anti-Bicd2 affinity
purification. The band corresponding to Bicd2 was no longer detectable
in a Western blot using the unbound flow-through, indicating that our
antibody was specific (Fig. 6B). In addition, we were also
able to detect association of endogenous Nek8 and Bicd2 when Nek8 was
immunoprecipitated from primary MVEC (Fig. 6C). These data
support the findings that the co-purification of Bicd2 with Nek8
activity from rabbit lung represented a specific interaction between
these proteins.
Tissue Distribution and Localization of Nek8 and Bicd2--
To
examine the tissue distribution of Nek8, we used Northern blot analysis
with a probe spanning amino acids 195-569 of human Nek8. A major
transcript of ~6.5 kb was identified in all mouse tissues examined,
but was most abundant in heart, liver, kidney and testis (Fig.
7A). We also observed varying
levels of expression in multiple cell lines, including Raji B cells,
smooth muscle cells, and fibroblasts (data not shown). Based on reverse
transcriptase-PCR analysis, the tissue distribution has been shown to
be ubiquitous for KIAA0699, the partial cDNA corresponding to Bicd2
(37). Using human-rodent hybrid panels, the chromosomal location of Bicd2 was determined to be chromosome 9 (37).
View larger version (33K):
[in this window]
[in a new window]
|
Fig. 7.
Distribution and localization of Nek8 and
Bicd2. A, expression of Nek8 in adult mouse
tissues. A murine multiple tissue Northern blot
(CLONTECH) was probed with a 375-amino acid
fragment of human Nek8, spanning part of the kinase and RCC1-like
domain (amino acids 194-569). B, subcellular distribution
of Bicd2. a-d, overexpressed Flag-Nek8 co-localizes
with endogenous Bicd2. HeLa cells were transfected with Flag-Nek8,
fixed after 48 h, and stained with anti-Bicd2 to reveal Bicd2
(a), anti-Flag to reveal Nek8 (b), or both in an
overlay (c). Staining with Bicd2 unbound flow-through shows
a diffuse predominantly nuclear pattern (d).
e-g, microtubule depolymerizing agents alter the
localization of Bicd2. Mouse embryo fibroblasts were left untreated
(e), or treated for 2 h with 10 µM
colchicine (f) or 10 µM nocodazole
(g) prior to fixation. Cells were stained with
anti-Bicd2.
|
|
We also examined the distribution of Bicd2 and Nek8 in cells by
indirect immunofluorescence. Endogenous Bicd2 in HeLa cells and mouse
embryo fibroblasts is localized in a dispersed filamentous pattern,
reminiscent of a cytoskeletal structure (Fig. 7B). Labeling of cells with the anti-Bicd2 depleted unbound flow-through eliminates this staining, indicating that the filamentous pattern is specific for
Bicd2. Although we were unable to detect endogenous Nek8 by immunofluorescence, Flag-Nek8 expressed in HeLa cells exhibits a
cytoplasmic staining pattern that overlaps with endogenous Bicd2 (Fig.
7B). It will be important to determine whether endogenous Nek8 also co-localizes with Bicd2, and whether this association is regulated.
In Drosophila, BicD has been reported to associate with
microtubules, and treatment of flies with the microtubule disrupting agent colchicine mimics a BicD loss of function phenotype (41, 42).
Co-immunofluorescence with antibodies against Bicd2 and tubulin
indicates that, unlike Drosophila, the distribution of Bicd2
and microtubules in murine fibroblasts is similar but not overlapping
(data not shown). Despite this, treatment of fibroblasts with agents
known to disrupt microtubules can dramatically alter the subcellular
distribution of Bicd2 (Fig. 7B). When fibroblasts were
treated with either nocodazole or colchicine, Bicd2 distribution was no
longer filamentous, but in a discrete perinuclear pattern. This
suggests that although Bicd2 may associate indirectly with the
microtubule cytoskeleton, microtubule morphology is important for
proper Bicd2 localization.
Expression of Nek8 during the Cell Cycle--
As NIMA and Nek2
have been demonstrated to be regulated during cell cycle progression,
we also examined the expression and activity of Nek8 through the cell
cycle. Extracts were prepared from exponentially growing HT29
epithelial cells or cells arrested in distinct cell cycle phases as
described under "Experimental Procedures." Fig.
8A illustrates the profiles of
each cell cycle stage analyzed as determined by flow cytometry.
Expression of endogenous Nek8 in lysates from each stage was determined
by immunoblotting with an anti-Nek8 polyclonal antibody. Fig.
8B indicates that Nek8 expression varied only mildly across
the cell cycle, with highest expression observed in G1 and
S-phase cells. Interestingly, in G2/M nocodazole-arrested
cells, Nek8 was seen to migrate as a doublet, suggesting that some
fraction of Nek8 was phosphorylated. The total amount of Nek8 did not
appear to fluctuate, however. In cells that were released from the
nocodazole arrest, the slower migrating form of Nek8 was no longer
visible, and total Nek8 expression was slightly reduced. Expression of
Bicd2 across the cell cycle was constant (Fig. 8B). As
expected, cdc2 was expressed and phosphorylated in all stages except in
G0-arrested cells (Fig. 8B). The amount of
immunoprecipitated Nek8 reflected that which was present in the total
lysates, with the highest amounts observed in G1 and S-phase cells (Fig. 8C). When we looked for the presence of
associated Bicd2 on immunoprecipitated Nek8, we found that the amount
of Bicd2 bound to Nek8 also reflected total Nek8 expression. That is,
more Bicd2 was bound to Nek8 in G1 and S-phase cells, where more Nek8 was immunoprecipitated (Fig. 8C). To determine
Nek8 activity across the cell cycle, immunoprecipitated Nek8 was
assayed for activity using the QSW peptide as a substrate in P81 filter binding kinase assays. Fig. 8D shows that there was little
variation in Nek8 activity across the cell cycle, although Nek8
activity was consistently higher in G0-arrested cells. This
profile does not correlate with the observed expression of Nek8 in
total lysates (Fig. 8B), where expression was slightly
increased in G1 and S-phase cells. Moreover, the presence
of a phosphorylated form of Nek8 in G2/M-phase cells showed
no measurable effects on Nek8 activity, suggesting that factors other
than phosphorylation may be important in Nek8 activation. Slightly
increased protein and activity levels in G0-arrested cells
have been reported for Nek3 (12). The modest differences observed in
Nek8 expression and activity across the cell cycle suggest that the
primary method of Nek8 regulation may be by other means.
View larger version (40K):
[in this window]
[in a new window]
|
Fig. 8.
Expression and kinase activity of Nek8 across
the cell cycle. A, flow cytometric profiles of cell
cycle progression. Cell lysates were prepared from exponentially
growing HT29 cells, or from cells arrested in distinct cell-cycle
phases as described under "Experimental Procedures." A portion of
each population was analyzed for DNA content by propidium iodide
staining and FACS analysis. B, expression of Nek8 across the
cell cycle in HT29 cells. Lysates were prepared from samples
illustrated in panel A, normalized by Bradford analysis
using bovine serum albumin as a standard and analyzed by Western
blotting with anti-Nek8 polyclonal antibody. Note that in
G2/M-arrested cells Nek8 migrates as a doublet. Lysates
were also analyzed by immunoblotting with anti-Bicd2 polyclonal
antibody and anti-phospho-cdc2 (Cell Signaling Technology).
C, expression of Nek8 and Bicd2 in Nek8 immunoprecipitates.
The amount of immunoprecipitated Nek8 correlates with expression in the
total lysate. Nek8 was immunoprecipitated with rat anti-Nek8 monoclonal
antibody and analyzed by Western blotting with Nek8 and Bicd2
polyclonal antibodies. D, Nek8 activity across the cell
cycle. Immunoprecipitated Nek8 was assayed for activity using the QSW
peptide as a substrate in P81 filter binding kinase assays. Activity is
shown as picomole of phosphate incorporated/assay. Each bar
corresponds to an average of picomole of phosphate incorporated from
three independent experiments with similar results.
|
|
| |
DISCUSSION |
In a search for an unidentified -casein kinase activity
reported to be stimulated exclusively by IL-1, we identified a novel protein kinase whose catalytic domain shares sequence homology with
NIMA-related kinases, termed Nek8. Nek8 contains an N-terminal kinase
domain and a C-terminal coiled-coil domain, resembling Nek1, Nek2, and
NIMA. Nek8 is unique in that it contains a central domain with homology
to RCC1, a chromatin bound GEF for the nuclear GTPase Ran. Although the
residues predicted to be important in maintaining the propeller
structure of RCC1 are conserved in Nek8, many of the residues important
for exchange factor activity appear not to be conserved. We have been
unable to detect association of Nek8 with a number of GTPases,
including Ran, Rac, Ras, Ral, and Rab11, and it is not known whether
Nek8 has any intrinsic exchange factor
activity.3 We cannot rule out
the possibility that Nek8 may associate with or promote GDP exchange on
other small GTPases.
The purification of Nek8 was directed by isolating a -casein kinase
activity. Three serine residues in -casein were readily phosphorylated by Nek8. Although the sequences surrounding each of
these residues corresponded to a consensus phosphorylation recognition
site, alteration of the conserved residues in each peptide substrate
had no effect on their ability to be phosphorylated by Nek8. In all
cases the phosphorylated serine was adjacent to a glutamine residue in
a QS orientation. It is worth noting that Bicd2 also contains a QS
motif, and Nek8 itself contains two QS and one QT motif. Whether or not
any of these sites are phosphorylated by Nek8, or whether this
represents a minimal recognition site, has not been determined.
Although mutation of the glutamine residue in the peptide substrates
has no effect on phosphorylation, this may not adequately reflect
effects that might be observed in the context of the full-length
molecule. The sites on -casein which are phosphorylated by other
Neks have not been identified.
Nek8 also shares biochemical characteristics with Nek1 and Nek2. Like
Nek1 and Nek2, Nek8 activity is stimulated by Mn2+ over
Mg2+ as a divalent cation requirement (7, 9). Nek8 activity is also sensitive to detergent, is insensitive to heparin, and cannot
utilize GTP as a phosphate donor, properties that have also been
described for Nek2 (40).
The regulation of Nek8 activity appears to be mediated by its
autophosphorylation, multimerization, and possibly by its association with Bicd2. Based on the ability of Nek8 to phosphorylate a peptide derived from its own activation loop, we predict that Nek8
autophosphorylates on Thr-210. The analogous site is highly conserved
among the mammalian Neks, and is a known autophosphorylation site in
NIMA (5). Although these data suggest that Thr-210 may serve as an
important regulatory site, mutation of Thr-210 within full-length Nek8
will be required to confirm these observations. Phosphoamino acid
analysis of full-length Nek8 revealed that the majority of
phosphorylation was on serine. Therefore, autophosphorylation of
Thr-210 must represent only a fraction of total phosphate
incorporation. Since phosphorylation of residues within the activation
loop of most kinases results in their activation, we would predict that
autophosphorylation of Thr-210 occurs first and thereby allows
subsequent phosphorylation of other sites within the molecule. We
cannot rule out the possibility that other kinases are also required
for Nek8 activation. We also observed that Nek8 autophosphorylation
leads to complex formation. The molecular weight of some of these
multimers is consistent with Nek8 dimerization. Nek8 multimers are
reduction insensitive and occur in the presence of purified protein,
suggesting that these are highly stable structures, and require no
additional proteins to form. However, Nek8 complexes do require ATP, as
incubation of Nek8 in the absence of ATP has no effect on its mobility
on SDS-PAGE. Our data is consistent with other reports demonstrating that Nek2 exists as a dimer, and dimerization results in
autophosphorylation and kinase activation (33). Dimerization of Nek2 is
dependent on a C-terminal coiled-coil domain which contains a unique
leucine zipper motif. Nek8 also contains a C-terminal coiled-coil
domain, which may be important for its dimerization. For both Nek8 and Nek2, it has not been possible to distinguish whether
autophosphorylation leads to complex formation, or vice versa.
The identification of Nek8 was based on a search for an IL-1 mediated
-casein kinase activity reported by Guesdon et al. (21,
22). Although we identified Nek8 activity in a homogenate from
IL-1-treated rabbits that was absent in homogenates from naïve
rabbits, we do not believe that Nek8 corresponds to the previously
described activity. First, Nek8 is a 123-kDa protein, whereas the
activity reported by Guesdon et al. (21) behaved as a 65-kDa
protein. Second, and perhaps more importantly, we have been unable to
detect significant Nek8 kinase activation in response to IL-1.
Moreover, perturbation of Nek8 in cells has no effect on any well
characterized IL-1-mediated responses, such as secretion of IL-6 or
granulocyte macrophage-colony stimulating factor, or activation of
molecules in IL-1 signaling
pathways.4 Our data suggest
that the isolation of Nek8 as an IL-1 regulated activity was
fortuitous, and the biological function of Nek8 is unlikely to be
involved in IL-1 mediated events.
The purification of Nek8 from rabbit lung revealed that a second
protein, Bicd2 co-chromatographed with Nek8 activity. Bicd2 is a human
homolog of the Drosophila coiled-coil protein, BicD. Nek8
can phosphorylate Bicd2 in vitro, and the endogenous
proteins associate in vivo, raising the possibility
that Bicd2 may be an in vivo substrate for Nek8.
Interestingly, the substrate identified for Nek2 is structurally
related to Bicd2. c-Nap1 is a centrosomal coiled-coil protein that
associates with and is phosphorylated by Nek2 and is important for
centrosome cohesion (43, 44). We have found that the distribution of
Bicd2 is in a filamentous pattern, resembling a cytoskeletal structure,
and transiently expressed Nek8 co-localizes with Bicd2. Recessive
mutations in Drosophila at the BicD locus disrupt
the formation and maintenance of the microtubule cytoskeleton and block
oocyte differentiation (42). Other mutations demonstrated to interfere
with phosphorylation of BicD also disrupt oocyte differentiation by
preventing its accumulation in the pro-oocyte (36). Similarly, BicD
gain of function mutants have also established the importance of BicD in localizing or transporting morphogenetic factors. In mammals, the
BicD homolog Bicd2 may play a similar role. Its filamentous distribution in cells suggests it may associate with cytoskeletal structures. Consistent with our findings, Hoogenraad et al.
(45) recently reported that mammalian Bicd2 associates with dynein, a
microtubule based motor. The localization of Bicd2 is dependent on
microtubule morphology, as treatment of cells with agents that disrupt
microtubules can alter Bicd2 distribution. Microtubule disruption
induced by nocodazole also induced phosphorylation of Nek8, as
determined by mobility shift on SDS-PAGE. It will be important to
determine whether nocodazole treatment also affects Bicd2
phosphorylation, and what the subcellular localization of endogenous
Nek8 might be under these conditions.
Finally, we examined Nek8 expression and activity through the cell
cycle. Our data suggest that, like Nek3, Nek8 activity was not
significantly changed throughout the cell cycle, yet we repeatedly
observed somewhat elevated levels in serum-starved G0-arrested cells. We did observe a shift in Nek8 mobility
in G2/M-arrested cells and this correlated with an
increased amount of Bicd2 associated to immunoprecipitated Nek8. It is
possible that the effects seen at G2/M are a consequence of
the drugs used for cell cycle synchronization, which lead to
microtubule destabilization, as we also saw changes in Bicd2
localization in response to nocodazole. These observations may reflect
changes in Nek8 and Bicd2 in response to microtubule integrity in a
cell cycle independent manner. Other circumstances affecting
microtubule stability include activation of signaling pathways,
initiation of apoptosis, or cell motility (46). Discoveries of
microtubule motors in complexes with signaling components suggest that
there is likely to be cross-communication between signaling molecules
and regulation of microtubule dynamics (47). We have identified Nek8 as
a novel NIMA-related kinase and an associated coiled-coil protein,
Bicd2. Nek8 displays many of the characteristics found in other
mammalian Neks, including structure, substrate specificity, and
regulation of kinase activity by autophosphorylation and complex
formation. It is tempting to speculate that Nek8 may either regulate or
be regulated by cytoskeletal dynamics. More studies are required to
determine whether Nek8 and Bicd2 may function in cellular processes
that are mediated by microtubule integrity.
,
TOP
ABSTRACT
INTRODUCTION
To whom correspondence should be addressed. Tel.: 206-587-0430;
Fax: 206-624-7496; E-mail: hollandp@immunex.com.
