(Received for publication, February 12, 1997)
From the Agonist-sensitive phosphoinositide pools are
maintained by recently-identified wortmannin (WT)-sensitive
phosphatidylinositol (PI) 4-kinase(s) (Nakanishi, S., Catt, K. J., and
Balla, T. (1995) Proc. Natl. Acad. Sci. U.S.A. 92, 5317-5321). Two loosely membrane-associated WT-sensitive type III PI
4-kinases were isolated from bovine adrenal cortex as
[3H]WT-labeled 110- and 210-kDa proteins. Based on
peptide sequences from the smaller enzyme, a 3.9-kilobase pair (kb)
cDNA with an open reading frame encoding a 90-kDa protein was
isolated from a bovine brain cDNA library. Expression of this
cDNA in COS-7 cells yielded a 110-kDa protein with WT-sensitive PI
4-kinase activity. Northern blot analysis of a human mRNA panel
showed a single ~3.8-kb transcript. Peptide sequences obtained from
the 210-kDa enzyme corresponded to those of a recently described rat 230-kDa PI 4-kinase. A 6.5-kb cDNA containing an open reading frame
of 6129 nucleotides that encoded a 230-kDa protein, was isolated from
brain cDNA. Northern blot analysis of human mRNA revealed a major
7.5-kb transcript. The molecular cloning of these novel WT-sensitive
type III PI 4-kinases will allow detailed analysis of their signaling
and other regulatory functions in mammalian cells.
Phosphatidylinositol (PI)1 4-kinases
were first recognized as the enzymes that provide PI 4-phosphate for
the synthesis of PI(4,5)P2, the precursor of two important
intracellular messengers, inositol 1,4,5-trisphosphate and
diacylglycerol (1). Early studies indicated that multiple PI kinases
are present in cellular homogenates, based on their individual
sensitivities to detergents (2, 3). It has been widely accepted that
the major PI 4-kinase activity is present in the plasma membrane and
regulates the synthesis of phosphoinositides destined for
hormone-regulated hydrolysis (3). This tightly membrane-bound activity
has been purified from detergent extracts of various membrane sources
including the red blood cell membrane (4-7), which contains a
~56-kDa enzyme termed type II PI 4-kinase. Another form of PI
4-kinase, the type III enzyme, was subsequently described in detergent
extracts of bovine brain membranes and differed from the type II enzyme
by its larger size (over 200 kDa based on sedimentation
characteristics), and lower affinity for both ATP and PI (8).
Receptor-mediated regulation of PI 4-kinase activity has been indicated
by the rapid increases of PI(4)P levels observed in agonist-stimulated
cells (9-11), but few data are available to support direct regulation
of this enzymatic activity by either G protein-coupled or growth factor
receptors (12-14). Increased PI kinase activity is associated with
activated receptor and non-receptor tyrosine kinases and viral
oncoproteins (15-17), but this activity phosphorylates PI on the 3- rather than the 4-position of its inositol ring (18). This enzyme,
termed type-I PI kinase or PI 3-kinase, contains a 110-kDa catalytic
and an 85-kDa regulatory subunit. It produces 3-phosphorylated
phosphoinositides by utilizing PI, as well as PI(4)P and
PI(4,5)P2, as substrates (see Ref. 19 for a review).
Several PI 3-kinases have been purified and cloned (20-24), and some
of these, such as the yeast Vps34p, only phosphorylate PI and interact
with a 150-kDa subunit. A 210-kDa form of PI 3-kinase has also been
reported in Drosophila (24).
Recently, the fungal metabolite wortmannin (WT), a potent inhibitor of
PI 3-kinases, has been used to define cellular functions that are
regulated by PI 3-kinases. Although WT was not believed to inhibit PI
4-kinases (25-27), we observed that micromolar concentrations of the
compound abolish the sustained formation of inositol
1,4,5-trisphosphate in agonist-stimulated cells by inhibiting a PI
4-kinase enzyme (28). Further characterization of this soluble (loosely
membrane-bound) WT-sensitive PI 4-kinase activity demonstrated its
similarity to the type III PI 4-kinase and showed that the bovine brain
type III enzyme (as originally described) displays similar WT
sensitivity (29). In contrast, the major cellular PI 4-kinase activity, the type II enzyme, was insensitive to WT and thus unlikely to participate in the synthesis of hormone-sensitive phosphoinositide pools (29). The sensitivity of the type III PI 4-kinase(s) to WT
(although significantly less than that of PI 3-kinases) raised the
possibility that they have structural similarities to PI 3-kinases.
In the present study we report the purification of 110- and 210-kDa
WT-sensitive, type III PI 4-kinase enzymes from the bovine adrenal
cortex, and the molecular cloning of cDNAs encoding these novel
enzymes.
DEAE-Sepharose, SP-Sepharose (bulk media), and
heparin-Sepharose, butyl-Sepharose, MonoQ, and MonoS columns were from
Pharmacia Biotech (Uppsala, Sweden). Phosphatidylinositol and ATP were
from Fluka (Ronkonkoma, NY) and Sigma, respectively.
[ In the final
purification, 60 bovine adrenal cortices were homogenized in three
batches (20 each) in Buffer A (20 mM Tris/HCl, pH 7.3, 1 mM EDTA, 1 mM dithiothreitol, 100 µM AEBSF, and 10 µg/ml leupeptin) containing 1 M NaCl as described previously (29), except that the
volumes were upscaled proportionally. All manipulations were performed
at 4 °C unless otherwise indicated. After centrifugation (100,000 × g, 90 min), the supernatant was taken to
40% ammonium sulfate (40% saturation) and the precipitated material
collected by centrifugation (10,000 × g, 20 min) and
the pellets stored at The active fractions from three such purifications were combined and
re-chromatographed on the Heparin column to concentrate samples and
remove glycerol. Ammonium sulfate was then added to the active
fractions (80 mg/ml), and the material was loaded onto a 10-ml
butyl-Sepharose minicolumn that was pre-equilibrated with Buffer A/80
mg/ml ammonium sulfate at room temperature. After loading, the column
was washed with the same buffer until the absorption of the eluent
returned to base line. At this point the column was eluted with Buffer
A at 1 ml/min, and 1.5-min fractions were collected. Active fractions
were pooled, diluted with three volumes of Buffer A, and loaded onto a
Mono Q column (HR 5/5). The column was washed with Buffer A, 50 mM NaCl for 10 min and eluted with a linear gradient of
0.05-0.5 M NaCl in Buffer A over 50 min at a flow rate of
1 ml/min. Two peaks of activity were eluted from this column and were
collected and pooled separately. The pooled fractions were diluted with
two volumes of Buffer B and loaded onto a Mono S columns (HR 5/5),
followed by a 10-min wash with 50 mM KCl in Buffer B and
elution with a linear gradient of 0.05-0.5 M KCl in Buffer
B over 50 min at 1 ml/min. Fractions that contained the activity were
combined, precipitated with 5% (w/v) trichloroacetic acid, and
analyzed on an 8% SDS-polyacrylamide gel. After Coomassie staining,
the bands that corresponded to the respective enzymes previously
identified with [3H]WT binding (see below) were cut out
for subsequent peptide sequencing.
Coomassie-stained bands at
210 and 110 kDa were excised and subjected to in situ
proteolytic digestion with modified trypsin (sequencing grade, Promega,
Madison, WI) essentially according to the method of Moritz et
al. (30). Washing steps were performed at 50 °C. The resulting
digest was separated at 0.25 ml/min with a gradient described by
Fernandez et al. (31) on a narrow bore (2.1 × 250 mm)
Vydac 218TP52 column and guard column (Separations Group, Hesperia, CA)
at 35 °C using a System Gold HPLC equipped with a model 507 autosampler, model 126 programmable solvent module, and model 168 diode
array detector (Beckman, Fullerton, CA). The column effluent was
monitored at 215 and 280 nm, and fractions collected at 30-s intervals
were stored at Aliquots
of the fractions eluted from the various columns were incubated for 20 min at room temperature in a total volume of 100 µl of PBS containing
0.4 µCi of [3H]WT-17-ol. This corresponds to 200 nM WT, a concentration that is sufficient to label proteins
with lower affinity to WT than PI 3-kinase. In some cases the samples
were preincubated with 10 nM unlabeled WT for 10 min prior
to labeling to occupy high-affinity binding sites such as those of the
PI 3-kinase(s). Proteins were precipitated with trichloroacetic acid
(5% final) and subjected to SDS-PAGE. When ethanol precipitation was
used instead of trichloroacetic acid (since WT binding has been
reported to be acid-labile), there was no difference in the protein
labeling. After fixation and Coomassie staining, the gels were
impregnated with EN3HANCE (NEN Life Science Products)
solution, and after drying were exposed at -70 °C for 1-2 weeks
with Hyperfilm (Amersham).
A size-enriched cDNA
library was created in the pSPORT1 plasmid from bovine brain cortex
using sucrose-fractionated (>2.5 kb) poly(A)+-selected RNA
and the SuperscriptTM plasmid system for cDNA cloning (Life
Technologies, Inc.) according to the manufacturer's instruction. This
library contained 1.25 million primary clones and after amplification was stored in aliquots of glycerol stock at An oligonucleotide primer was designed from the peptide sequence,
QLQSIWEQE, obtained from tryptic fragments of the 110-kDa PI 4-kinase,
and an antisense primer based on the conserved KDRHNGN sequence that is
common to all known PI 4-kinases. PCR amplification using this primer
pair (5 The isolation of the cDNA encoding the 210-kDa enzyme was initially
attempted by colony hybridization of the brain cDNA library with a
PCR-amplified fragment of the human PI4K Northern blot analysis was performed
utilizing an mRNA panel of several human tissues
(CLONTECH, Palo Alto, CA) and the random primer-labeled 1.8-kb fragment of the EcoRI digest of one of
the clones (c365) (corresponding to nucleotides 1960-3750 of
PI4KIII COS-7 cells were
grown to about 70% confluence in Dulbecco's modified Eagle's medium,
10% fetal bovine serum on 10 cm culture dishes. Cells were transfected
with 5 ml of Opti-MEM medium containing 10 µg/ml LipofectAMINE (Life
Technologies, Inc.) and 5 µg plasmid DNA (pcDNA3.1(+), containing
the PCR-amplified clone of PI4KIII As reported previously (29), two PI
4-kinase activities with native molecular sizes of ~110 and ~200
kDa were identified and partially purified from the bovine adrenal
cortex. The catalytic properties of the two components were
indistinguishable, raising the possibility that the two peaks represent
monomeric and dimeric forms of the same enzyme or that the larger is a
heterodimer associated with another protein subunit. Based on the
presumed similarity with PI 3-kinases, a purification procedure similar
to that used for preparation of PI 3-kinases (33, 34) was employed by
sequential chromatographies on DEAE-Sepharose, SP-Sepharose,
heparin-Sepharose, butyl-Sepharose, MonoQ, and MonoS columns (Fig.
1). Two activities were clearly separated on MonoQ
chromatography, and [3H]WT-17-ol-binding (35) was used to
correlate the PI 4-kinase activity of the effluent fractions with the
[3H]WT labeling of the proteins. As shown in Fig.
2, SDS-PAGE analysis of the WT-labeled proteins revealed
that the two peaks of PI 4-kinase activity correlated with two labeled
proteins of ~210 and ~110 kDa. The radioactivity bound to these
proteins was only slightly reduced by preincubation of the fractions
with 10 nM unlabeled WT (Fig. 3). Such
treatment greatly reduced the labeling of PI 3-kinase(s) (data not
shown), consistent with the higher affinity of these enzymes for WT
(29).
Due to their relatively low abundance, these proteins were not purified
to homogeneity even after passage through multiple chromatography
steps. However, WT labeling allowed their clear identification on SDS
gels after Coomassie staining. The two protein bands of interest were
then cut out from the gels and digested with trypsin. Only one
unequivocal peptide sequence was obtained from the larger enzyme:
EFDFFNK, which showed homology to a recently cloned human PI 4-kinase,
PI4K Tryptic digestion of the smaller protein yielded seven unequivocal
peptide sequences (LSEQLAHTPTAFK, QLQSIWEQE, VENEDEPVR, LATLPTK, EFIK,
EPVFIAAGDIR, and EPGVQA). None of these exhibited homology to any known
protein sequence available in the data base. However, a computer
alignment of the three human PI 3-kinases (
A primer was
designed based on the comparison of the nucleotide sequences encoding
QLQSIWEQE, and the corresponding nucleotide sequences of PIK1 and the
three PI 3-kinases based on the putative amino acid alignment described
above. This primer was used in combination with another primer designed
on a conserved sequence (KDRHNGN) that is present in all of the
then-known PI 4-kinases (PIK1, STT4, and PI4K
The peptide sequence (EFDFFNK) that was obtained from the larger enzyme
indicated its similarity to the 200-kDa yeast PI 4-kinase, STT4 (37)
and to a smaller human PI 4-kinase, PI4K
Because
of the similarity of the larger bovine PI 4-kinase to the recently
characterized 230-kDa rat enzyme, (32), our expression studies were
focused on the smaller enzyme. A full-length clone of the 110-kDa
enzyme was obtained using RT-PCR of mRNA isolated from bovine
adrenal glomerulosa cells with long range PCR and the mammalian
expression plasmid pcDNA3.1(+) (see "Experimental Procedures"
for details). Plasmid DNA was expressed in COS-7 cells, and kinase
activity was measured in the soluble fractions after homogenization and
sonication of the transfected cells. As shown in Fig. 7,
the transfected cells contained increased PI 4-kinase activity, which
after chromatography on DEAE Sepharose minicolumns was completely
inhibited by 1 µM WT. Control COS-7 cells that were
treated only with LipofectAMINE also contained some endogenous WT-sensitive PI 4-kinase. [3H]WT binding to the extracts
obtained from transfected COS-7 cell and subsequent SDS-PAGE showed
prominent labeling of a 110-kDa band that was identical to the 110-kDa
enzyme purified from bovine adrenal cortex (Fig. 7). Preincubation of
the extracts with increasing concentrations of radioinert WT before
[3H]WT labeling indicated an affinity that was consistent
with the lower sensitivity of this kinase than PI 3-kinase to
inhibition by WT (29). Several [3H]WT-labeled bands were
present in mock-transfected COS-7 cells and showed no change upon
transfection. The most prominent of these was a large protein of >250
kDa that is probably a member of the family of PI kinase-related
enzymes (38). The endogenous 210-kDa kinase of COS-7 cells showed only
very faint labeling, and the endogenous 110-kDa labeled band is
probably a mixture of PI 3- and 4-kinases.
Phosphorylation of phosphatidylinositol by PI 4-kinases has long
been considered as the initial reaction for the synthesis of membrane
phosphoinositides that serve as precursors for the agonist-stimulated
formation of inositol 1,4,5-trisphosphate and diacylglycerol. While the
most abundant cellular PI 4-kinase, the tightly membrane-bound type II
enzyme, was believed to be the most likely candidate to perform this
function, our recent studies revealed that the sustained formation of
hormone-sensitive inositide pools requires the participation of a
WT-sensitive PI 4-kinase enzyme (28). Our analysis of the sensitivity
of the various cellular PI 4-kinases to WT also showed that only the type III and not the type II enzyme(s) show such sensitivity (29). Purification of the WT-sensitive, loosely membrane-associated type III
PI 4-kinase from the bovine adrenal cortex revealed it to be a mixture
of two major activities with molecular sizes of ~200 and 110 kDa
(29). [3H]WT labeling and subsequent SDS-PAGE analysis
confirmed the existence of two separate proteins of 210 and 110 kDa,
both of which showed catalytic properties characteristic of type III PI
4-kinases and similar WT sensitivities (29).
Based on the expected similarity between these enzymes and PI
3-kinases, we employed a purification scheme that was successful for
the isolation of PI 3-kinases (33, 34). Although these separation
methods were also applicable to the PI 4-kinases, the stronger
interaction of these enzymes with the ion exchangers did not allow as
efficient separation from the majority of proteins as in the case of PI
3-kinases. Nevertheless, a sufficient amount of protein was obtained to
permit petide sequences to be obtained from direct digests of the SDS
gel slices containing the two proteins of interest.
Isolation and analysis of cDNA clones encoding these two
enzymes confirmed their identity with the purified proteins and
revealed that they were mammalian homologs of the yeast STT4 and PIK1
enzymes. Although the overall homology compared with the respective
yeast enzymes is low (24% and 16% on the protein level for the bovine 110-kDa enzyme versus PIK1 and the 210-kDa enzyme
versus STT4, respectively), these enzymes show a large
degree of conservation within their lipid kinase/protein kinase and
lipid kinase unique domains. However, sequence comparison within the
kinase domain clearly defined two groups of the type III PI 4-kinases
(Fig. 8 and Ref. 36). The marked homology of the 210-kDa
enzyme with the much shorter (97 kDa) human PI4K The cDNA sequence encoding the smaller WT-sensitive enzyme,
PI4KIII Both the 110- and 210-kDa enzymes were found to be sensitive to WT,
although at higher concentrations than those needed to inhibit PI
3-kinases (29). A recent study identified Lys-802 of PI3K The physiological roles and modes of regulation of the multiple PI
4-kinases are not clear at present. However, our data on the WT
sensitivity of the maintenance of agonist-sensitive
PI(4,5)P2 pools indicate that one or both of the currently
described enzymes participates in this process. In the yeast, deletion
of PIK1 but not STT4 is lethal. The latter mutant has an
osmolarity-dependent phenotype (37, 43) and increased
staurosporine sensitivity, indicating its possible connection with
PKC-dependent pathways. PIK1 has also been cloned from
Saccharomyces cerevisiae with the aid of antibodies raised
against the nuclear pore complex and is presumably also present in the
nucleus (44), but its function (if any) there is not known. The
localization of the epitope-tagged 230-kDa rat PI 4-kinase was largely
Golgi-associated in overexpressing COS cells (32), but this could
reflect the artificially high production of the protein in such cells.
The role of the very abundant WT-resistant type II PI 4-kinase(s) is
even more enigmatic, as well as its relationship to the WT-sensitive PI
4-kinase enzymes or to the yeast enzymes. The increasing number of
enzymes that are known to be regulated by various phosphorylated
inositides, mostly via their plekstrin homology domains (45-48),
indicates that regulation of inositide synthesis at several subcellular locations might be an important means of controlling the assembly and
function of active signaling complexes.
In summary, the present results describe the identification,
isolation, and molecular cloning of two WT-sensitive PI 4-kinase enzymes from the bovine brain and adrenal cortex. These enzymes are
mammalian homologs of two yeast PI 4-kinases, PIK1 and STT4, that
appear to have clearly distinct functions in yeast. The cloning of
these enzymes should facilitate the clarification of their role(s) and
regulation in mammalian cells, and the understanding of the multiple
roles of phosphoinositides in cell regulation.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) U88531 and U88532. We are grateful to Dr. Larry Mertz (Life
Technologies, Inc., Gaithersburg, MD) for helpful advice and to Dr. Y. Yamada (NIDR, National Institutes of Health) for providing the rat EST
clone. The skillful technical work of Yue Zhang and the assistance of Dr. Hong Ji are also greatly appreciated.
Shortly after the completion of this study,
Meyers and Cantley (51) and Nakagawa et al. (52) reported
the cloning of a WT-sensitive human and rat PI 4-kinase, respectively,
(40) that are homologs of the bovine PI4KIII
Endocrinology and Reproduction Research
Branch,
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
Addendum
REFERENCES
Materials
-32P]ATP (6000 Ci/mmol) and
[
-32P]dCTP (3000 Ci/mmol) were from Amersham, and
[3H]wortmannin-17-ol (19.7 Ci/mmol) was from NEN Life
Science Products. Reagents for SDS-PAGE were obtained from Bio-Rad,
CA), and restriction enzymes were purchased from Life Technologies,
Inc. or New England Biolabs (Beverly, MA). The T/A cloning kit was
obtained from Promega (Madison, WI), and the sequencing kits from
Amersham. All other chemicals were of HPLC or analytical grade.
70 °C until further processing. Ammonium
sulfate precipitates were dissolved in 100 ml of Buffer A and dialyzed
overnight at 4 °C against 10 volumes of Buffer A containing 30 mM NaCl. The dialyzed protein solution was then diluted to
500 ml with Buffer A/30 mM NaCl and applied to a 5 × 70-cm DEAE-Sepharose column pre-equilibrated with the same buffer.
After loading, the column was washed with 2 liters of Buffer A, 30 mM NaCl and eluted with a linear gradient of 0.03-0.5
M NaCl in Buffer A with a flow rate of 10 ml/min. The PI
4-kinase activity of the fractions was measured as described previously
(28). Active fractions were combined and diluted with Buffer B (20 mM MES, pH 6.8, 1 mM EGTA, 0.5 mM dithiothreitol, 100 µM AEBSF, and 1 µg/ml leupeptin) to
a conductance of 5 millisiemens/cm. This material was loaded onto an
SP-Sepharose column (1.5 × 50 cm) that was pre-equilibrated with
Buffer B. After loading, the column was washed with 300 ml Buffer B and eluted with a linear gradient of 0-1 M KCl in Buffer B at
a flow rate of 1 ml/min. Active fractions were combined again and
diluted with Buffer A to a conductance of 22 millisiemens/cm and
applied to a 5-ml heparin-Sepharose minicolumn, which was equilibrated with 0.2 M NaCl in Buffer A. The column was washed with 20 ml of Buffer A, 0.2 M NaCl and then fitted to an HPLC
system (Gilson, Middleton, WI) for elution with a gradient of 0.2-1
M NaCl in Buffer A at a flow rate of 1 ml/min at room
temperature. The UV absorption (280 nM) of the eluent was
monitored by a Stratos UV-detector (Thomson Instruments, Vienna, VA),
and 1-ml fractions were collected on ice. Active fractions were saved
and stored at
20 °C with 10% glycerol.
70 °C. Fractions (125 µl) containing tryptic
peptides were applied in 30-µl aliquots to a Biobrene (Applied
Biosystems, Foster City, CA)-treated glass fiber filter and dried prior
to amino acid sequencing on a model 477A pulsed-liquid protein
sequencer equipped with a model 120A PTH analyzer (Applied Biosystems)
using methods and cycles supplied by the manufacturer. Data were
collected and processed by a model 610A data analysis system (Applied
Biosystems). Amino acid sequences were searched in the GCG-Swiss
Protein Data base (University of Wisconsin Genetics Computer
Group).
70 °C.
-ctgcaRtctatttgggaRcaag-3
and 5
-attgttaccgttgtgtctgtcctt-3
) yielded a 300-bp product, which was ligated into the PGEM-Easy plasmid
(Promega) and subjected to dideoxy sequencing. This DNA fragment, which
encoded an amino acid sequence with high homology to PIK1, was random
primer-labeled and used to screen the size-enriched bovine brain
cDNA library. Homology search of the data base with the 300-bp
sequence also revealed high homology to a rat EST (R46930) (Y-162) that
was subsequently provided by Dr. Y. Yamada (NIDR, National Institutes
of Health) and on sequencing showed extensive homology with PIK1. The
1.3-kb EcoRI/XhoI insert of this clone was also
used for screening the bovine brain library in subsequent experiments.
Positive colonies were isolated and the plasmids cut with
EcoRI/NotI restriction enzymes to determine their
insert-size. Several clones were isolated and sequenced, the longest of
which (c354) contained a 3-kb insert. The missing 5
end of the
mRNA was obtained by 5
-RACE (version 2.0, Life Technologies, Inc.) following the manufacturer's instructions. The full-length clone used
for transfection studies was created by long-range PCR amplification (Elongase, Life Technologies, Inc.) from cDNA prepared by reverse transcriptase (Superscript, Life Technologies, Inc.; Rethrotherm, Epicentre Technologies) from mRNA isolated from cultured bovine adrenal glomerulosa cells. The primers used for this amplification were
(5
-aggatccgagaaatggcacacctcag-3
and
5
-cggcaagctctagagttaccacatgatc-3
), and the 3.3-kb product was ligated
into the pcDNA3.1(+) plasmid (Invitrogen, Carlsbad, CA) after
digestion with XbaI/BamHI.
, due to the homology indicated by the peptide sequences obtained from the 210-kDa bovine protein. The longest insert isolated was a 3.3-kb product that lacked a
poly(A)+ tail due to the use of NotI enzyme
during creation of the cDNA library and the presence of an internal
NotI site after the stop codon in this sequence. The
3
-untranslated region was then obtained by 3
-RACE, and the 5
end of
the transcript was amplified with PCR using primers based on the rat
sequence (32) that became available during the course of these studies.
Completion of the 5
sequence information was achieved by amplification
from the cDNA library with 5
primers designed on the flanking
plasmid (pSPORT1) sequence. Full-length expressable clones were created by long-range PCR amplification from the cDNA of bovine adrenal glomerulosa cells (primers: 5
cgcggatcctgtgca-gagaccggcatgtgtggag-3
and 5
-cggaattccacacagagaccggctctgattgtc-3
) and ligation of the 6.3-kb
product into the pcDNA3.1(+) plasmid after digestion with BamHI and EcoRI.
) or the full 2.9-kb insert of a partial clone of PI4KIII
(c2D5) (3286-6200). After prehybridization at 42 °C in Hybrisol I
(Oncor, Gaithersburg, MD), hybridization was performed overnight at
42 °C. The blot was washed several times with increased stringency
with a final wash of 0.2 × SSC, 0.1% SDS at 55 °C for 10 min.
The blots were analyzed both by autoradiography and by a PhosphorImager (Molecular Dynamics).
). After 8 h the medium was
replaced with Dulbecco's modified Eagle's medium/10% fetal bovine
serum and culture was continued for selected periods. PI 4-kinase
activity was then measured in the soluble fractions after lysing the
cells in 500 µl of ice-cold Buffer A containing 150 mM
NaCl, followed by sonication and centrifugation at 14,000 × g for 30 min at 4 °C. For [3H]WT binding,
the supernatants obtained from two such plates were combined, diluted,
and applied to 1-ml DEAE-Sepharose columns, and after washing, eluted
with 500 mM NaCl in Buffer A. This eluent was then
concentrated on Amicon filters and subjected to PI 4-kinase activity
measurement and [3H]WT binding followed by SDS-PAGE.
Soluble Extracts of Bovine Adrenal Cortex Contain Two Distinct
WT-sensitive PI 4-Kinase
Fig. 1.
Fractionation of WT-sensitive PI 4-kinase
activity during sequential chromatographic steps. Ammonium sulfate
precipitates from the soluble fraction of adrenocortical homogenates
were dissolved and chromatographed on the indicated ion exchange
columns as described under "Experimental Procedures." The PI
4-kinase activities of the effluent fractions were assayed in the
presence of Triton X-100 (to inhibit PI 3-kinase) and were found to be
inhibited by WT (data not shown).
[View Larger Version of this Image (17K GIF file)]
Fig. 2.
Separation of WT-sensitive PI 4-kinase
activities by MonoQ ion exchange chromatography. Enzyme
preparations enriched by sequential chromatographies (Fig. 1) were
loaded on a MonoQ HR 5/5 column and eluted with a gradient of NaCl as
detailed under "Experimental Procedures." PI 4-kinase activity
measurements and [3H]WT-binding followed by SDS-PAGE
analysis was performed on the effluent fractions. The two peaks of
enzyme activity correlated with the labeling of 210-kDa and 110-kDa
proteins.
[View Larger Version of this Image (34K GIF file)]
Fig. 3.
Correlation between PI 4-kinase activity and
[3H]WT labeling of the effluent fractions of MonoS
chromatography of MonoQ-purified 210-kDa (left) and 110-kDa
(right) bovine adrenal PI 4-kinase. Active fractions
eluted from the MonoQ column (Fig. 2) containing the 210-kDa and
110-kDa PI 4-kinase, respectively, were loaded on MonoS HR 5/5 columns
and eluted with a KCl gradient as described under
"Experimental Procedures." PI 4-kinase activity
measurements and [3H]WT-binding followed by SDS-PAGE
analysis were performed on the effluent fractions.
[3H]WT-binding was only slightly reduced by 10 min
preincubation of the samples with 10 nM unlabeled WT
(lower strip of bands) consistent with the lower affinity of
these enzymes for WT. Similar treatment greatly reduced the labeling of
PI 3-kinase (data not shown).
[View Larger Version of this Image (28K GIF file)]
(36), and to the yeast PI 4-kinase, STT4 (37). Other peptide
peaks from the digest were mixtures of peptides that yielded multiple
sequences that were homologous to PI4K
(data not shown). Further
purification of the peptides was not attempted because of the minute
amounts available. Although the predicted molecular size of PI4K
is
only 97 kDa, we concluded that it probably represents a smaller splice
variant of the 210-kDa bovine enzyme, which would be the mammalian
homolog of the yeast 200-kDa STT4.
,
, and
), and the
125-kDa yeast PI 4-kinase, PIK1, together with one of the sequences,
QLQSIWEQE, revealed a possible alignment for the latter (see Fig. 8).
This served as the basis for designing primers so that further sequence
information could be obtained from this enzyme.
Fig. 8.
Comparison of the amino acid sequences of
identified PI 3- and PI 4-kinases within their catalytic lipid
kinase/protein kinase domain (A) or lipid-kinase unique
domain (B). Conserved residues that are present in all
or most of these enzymes are indicated by the dark areas,
and those that are conserved within groups are shown on a light
gray background. The Lys residue to which WT binds covalently in
PI 3-kinases (42) is labeled with an asterisk.
PI3K, -
, and -
as well as huVps34p and
PI4K
are human sequences (Refs. 20, 36, 21, 49, and 22,
respectively), PI3K_68D is a sequence from
Drosophila (24), while Vps34p, PIK1, and STT4 are yeast sequences (Refs. 50, 43, and 37,
respectively). Dic-3K1 and Dic-4K are the DdPIK1
and DdPIK4 sequences described in Dictyostelium (23).
PI4Kp230 is the sequence from the 230-kDa rat PI 4-kinase
(32), and Y-162 is the sequence of the rat EST provided by
Dr. Y. Yamada.
[View Larger Version of this Image (46K GIF file)]
) but differs slightly
in all PI 3-kinases (GDRHNXN). Amplification of a
size-enriched bovine brain cDNA library with these primers yielded
a 300-bp product with a nucleotide sequence showing extensive homology
with PIK1. Screening of the size-enriched bovine brain cDNA library
with this product yielded several clones, one of which contained a
3.0-kb insert (c354). Sequencing of this clone confirmed its homology
with PIK1, especially within its catalytic domain, and revealed an open
reading frame that encoded all seven peptide sequences obtained from
the purified protein. Northern blot analysis showed that a single
transcript of ~3.8 kb was present in several tissues (Fig.
6A), and 5
-RACE was employed to capture the missing 5
region of the mRNA. The full-length transcript was found to be 3859 bp with an open reading frame of 2403 nucleotides that encodes a
protein of 90 kDa (Fig. 4). Interestingly, one of the
isolated clones contained an additional stretch of 48 nucleotides
encoding a 16-amino acid serine-rich sequence that is intercalated
within one of the peptide sequences obtained from the purified protein,
and must represent a splice variant of the enzyme. This sequence was
not present in clones that were obtained by PCR for mammalian
expression studies and probably represents a minor variant of the
enzyme (Fig. 4).
Fig. 6.
Northern hybridization of human mRNA from
several tissues with 32P-labeled probes for the 110-kDa
(PI4KIII, upper panels) and the 210-kDa (PI4KIII
,
lower panels) bovine adrenal PI 4-kinase. Membranes
containing 2 µg of poly(A)+-selected human RNA
(CLONTECH) were hybridized with
random-prime-labeled cDNA probes for the respective enzymes as
described under "Experimental Procedures." After washing (0.2 SSC,
0.1% SDS, 55 °C), the membranes were subjected to analysis by a
PhosphorImager after 24 h of exposure.
[View Larger Version of this Image (93K GIF file)]
Fig. 4.
Nucleotide and predicted amino acid sequence
of the 110-kDa type III bovine PI 4-kinase. The sequence within
the box shows a minor splice variant of the enzyme as
indicated by one of the isolated clones. The peptide sequences obtained
from the purified 110-kDa enzyme are underlined.
[View Larger Version of this Image (60K GIF file)]
(36). This and other less
certain peptide sequences that were obtained from peptide mixtures of
the enzyme digests of the 210-kDa bovine PI 4-kinase suggested that
this enzyme is a mammalian homolog of the yeast enzyme, STT4, and a
larger form of PI4K
. Screening of the bovine cDNA library with a
cloned fragment (1130-2908) of PI4K
yielded several clones, one of
which contained a 3.3-kb insert. Sequencing of these clones showed
their homology with PI4K
, but revealed a difference in the 5
end.
While this work was in progress, the cloning of a rat 230-kDa PI
4-kinase that showed strong sequence homology with the sequenced bovine
clones was reported (32). The 5
end of the bovine sequence was
obtained by PCR, using primers based on the 5
sequence of the rat
enzyme and the longest bovine cDNA clone, and 5
-RACE was used to
determine the bovine sequence that corresponded to the 5
primer (rat)
and 5
-flanking sequences. In addition, the 3
-untranslated region of
the mRNA was obtained with 3
-RACE, since all the clones were truncated after the stop codon due to a NotI site in the
sequence and the use of NotI during construction of the
cDNA library. The full-length sequence of 6520 nucleotides that was
reconstructed in this way encoded a 230-kDa protein that showed 92%
homology at the protein level with the rat PI 4-kinase enzyme (32)
(Fig. 5). Northern blot analysis showed the presence of
a prominent ~7.5-kb transcript and minor amounts of smaller
transcripts in some tissues, notably the placenta (Fig.
6B).
Fig. 5.
Nucleotide and predicted amino acid sequence
of the 210-kDa type III bovine PI 4-kinase. Only the unequivocal
peptide sequence obtained from the purified 210-kDa protein is shown
underlined.
[View Larger Version of this Image (108K GIF file)]
Fig. 7.
Expression of the 110-kDa bovine PI 4-kinase
in COS-7 cells. COS-7 cells were grown to 70% confluence and
transfected with LipofectAMINE reagent and the pcDNA3.1(+) plasmid
containing the PCR-amplified insert encoding PI4KIII. After 48 h, cells were harvested and sonicated as detailed under "Experimental
Procedures." The soluble fractions from control and transfected COS-7
cells were applied to DEAE-Sepharose minicolumns (for separation from WT-insensitive PI 4-kinases) and eluted with 0.5 M NaCl in
Buffer A. After concentration, samples were assayed for PI 4-kinase
activity in the presence and absence of 1 µM WT (after 10 min of preincubation) or were subjected to [3H]WT-binding
after a 10-min preincubation in the presence of increasing concentrations of unlabeled WT. Partially purified 110-kDa PI 4-kinase
was also labeled with [3H]WT. Samples were resolved on
10% SDS-polyacrylamide gels and after impregnation with
EN3HANCE solution were exposed to x-ray films for 2 weeks
at
70 °C.
[View Larger Version of this Image (39K GIF file)]
enzyme described by
Wong and Cantley (36) is consistent with the possibility that the latter might represent a splice variant of the human homolog of the
bovine enzyme. The reason for the difference between the catalytic properties of the 210-kDa enzyme (type III) and those reported for the
expressed PI4K
(type II) is not clear at present. However, while
these studies were in progress, a novel cDNA that encodes a 230-kDa
rat PI 4-kinase was isolated by homology cloning (32) and the
purification and partial cloning of a 200-kDa bovine brain PI 4-kinase
was reported (39). Although their WT-sensitivities were not examined,
both of these larger enzymes were identified as type III PI 4-kinases
that are very similar or identical to the 210-kDa enzyme reported in
the present study. Based on these findings, we propose the term
PI4KIII
to denote the identity of this larger WT-sensitive enzyme as
a type III form with homology to human PI4K
, and PI4KIII
for the
110-kDa bovine PI 4-kinase.
, contained all seven peptide sequences in an open reading frame for a 90-kDa protein. Although this enzyme is a homolog of the
yeast PIK1, its sequence similarity is confined largely to the
C-terminal third of the molecule that contains the catalytic domain
(Fig. 8A). Unlike the yeast enzyme, in which the lipid kinase unique domain is located near the N-terminal end, in the bovine
enzyme this domain is more distantly positioned from the N terminus
(Fig. 8B). Interestingly, in all three PI 3-kinases (
,
, and
), as well as in the large PI 4-kinases, this domain is
even further away from the N-terminal part of the molecule (36).
PI4KIII
also contains a proline-rich sequence in its N-terminal
region (Fig. 8). This sequence may promote the interaction of this
enzyme with SH3 domains (40), but may also serve as an N-terminal
processing signal as suggested in the case of cytochrome P450 enzymes
(41). This kinase could also interact with membranes through the
putative myristoylation site at its N-terminal end. In contrast,
interaction of the larger enzyme, PI4KIII
, with membranes may also
be aided by its putative plekstrin homology domain located between its
lipid kinase unique and lipid kinase/protein kinase catalytic domains
as described for PI4K
and STT4 (36). The shorter bovine PI 4-kinase,
PI4KIII
, showed closest homology to PI3K
of the PI 3-kinase
family members. We found no sequence homology with PI 3-kinases that
would suggest the interaction of this protein with the p85 regulatory
protein of PI 3-kinase, and no indication was found during purification
for the existence of heterodimers of the PI 4-kinases. The relatively
large difference between the calculated molecular size of the protein
and its apparent size on SDS-PAGE raises the possibility that the
enzyme undergoes posttranslational modification. Similarly, the yeast
enzyme PIK1 shows a larger apparent size on SDS gels (125 kDa) than its
calculated molecular size (119 kDa).
as the
site to which WT binds covalently within the putative ATP-binding
domain (42), probably with the help of additional interactions with
other residues, including Glu-821, Ser-919, and His-936 of PI3K
(42). Although the residue corresponding to Lys-802 is highly conserved
among all PI kinases, including the two enzymes described in this
study, WT sensitivities show great variations even within PI 3-kinases.
This suggests that the stabilization of WT binding by other residues is
an important determinant of the inhibitory potency of this compound.
Interestingly, none of the additional residues named above are
conserved between PI 3- and PI 4-kinases. It is important to note that
the WT sensitivity of these enzymes is dependent on the experimental
conditions, in particular on ATP concentration, incubation time, and
pH. Since both of the enzymes have a high Km for
ATP, it is quite likely that at the prevailing ATP concentrations in
intact cells, their WT sensitivity is even lower than under in
vitro conditions. These factors must be considered when
interpreting data on the WT sensitivity of cellular responses.
*
The costs of publication of this
article were defrayed in part by the
payment of page charges. The article
must therefore be hereby marked
"advertisement" in
accordance with 18 U.S.C. Section
1734 solely to indicate this fact.
§
To whom correspondence should be addressed: Endocrinology and
Reproduction Research Branch, NICHD, National Institutes of Health,
Bldg. 49, Rm. 6A36, 49 Convent Dr., Bethesda, MD 20892-4510. Tel.:
301-496-2136; Fax: 301-480-8010; E-mail: tambal{at}box-t.nih.gov.
1
The abbreviations used are: PI,
phosphatidylinositol; PI(4)P, phosphatidylinositol 4-phosphate;
PI(4,5)P2, phosphatidylinositol 4,5-bisphosphate; WT,
wortmannin; bp, base pair(s); kb, kilobase pair(s); RACE, rapid
amplification of cDNA ends; HPLC, high performance liquid
chromatography; PAGE, polyacrylamide gel electrophoresis; AEBSF,
4-(2-aminoethyl)-benzenesulfonyl fluoride hydrochloride; MES,
2-(N-morpholino)ethanesulfonic acid.
described in this
study.
©1997 by The American Society for Biochemistry and Molecular Biology, Inc.