(Received for publication, January 23, 1997, and in revised form, May 20, 1997)
From ICOS Corporation, Bothell, Washington 98021, § Department of Biological Chemistry, University of
Michigan, Ann Arbor, Michigan 48109, ¶ Fred Hutchinson Cancer
Research Center, Seattle, Washington 98104
We have identified a novel p110 isoform of
phosphatidylinositol 3-kinase from human leukocytes that we have termed
p110. In addition, we have independently isolated p110
from a
mouse embryo library on the basis of its ability to interact with
Ha-RasV12 in the yeast two-hybrid system. This unique
isoform contains all of the conserved structural features
characteristic of the p110 family. Recombinant p110
phosphorylates
phosphatidylinositol and coimmunoprecipitates with p85. However, in
contrast to previously described p110 subunits, p110
is expressed in
a tissue-restricted fashion; it is expressed at high levels in
lymphocytes and lymphoid tissues and may therefore play a role in
phosphatidylinositol 3-kinase-mediated signaling in the immune
system.
Phosphatidylinositol (PI)1 3-kinase
was originally identified as an activity associated with viral
oncoproteins and growth factor receptor tyrosine kinases that
phosphorylates PI and its phosphorylated derivatives at the 3-hydroxyl
of the inositol ring (1). The purification and subsequent molecular
cloning of PI 3-kinase revealed that it is a heterodimer consisting of p85 and p110 subunits (2-5).
The p85 subunit acts to localize PI 3-kinase activity to the plasma
membrane by virtue of the interaction of its SH2 domain with
phosphorylated tyrosine residues (present in an appropriate local
sequence context) in target proteins (6). Two isoforms of p85 have been
identified: p85, which is ubiquitously expressed, and p85
, which
is primarily found in brain and lymphoid tissues (7). The p110 subunit
contains the catalytic domain of PI 3-kinase, and three isoforms of
p110 have thus far been reported (
,
, and
) (3, 8, 9). The
identification of p110
revealed additional complexity within this
family of enzymes. p110
is most closely related to p110
and
(45-48% identity in the catalytic domain) but does not make use of
p85 as a targeting subunit. p110
contains an additional domain
termed a pleckstrin homology domain near the amino terminus. The
pleckstrin homology domain allows interaction with the
subunits
of heterotrimeric G proteins that appears to regulate its activity and
subcellular localization (9).
Additional members of this growing gene family include more distantly related lipid and protein kinases such as Vps34, TOR1, and TOR2 of Saccharomyces cerevisiae (and their mammalian homologues such as FRAP and mTOR), the human ataxia telangiectasia gene product, and the catalytic subunit of DNA-dependent protein kinase (10).
The levels of phosphatidylinositol-3,4,5-triphosphate, the primary product of PI 3-kinase activation, are elevated upon treatment of cells with a wide variety of agonists (11). This observation has implicated PI 3-kinase activation in a diverse range of cellular responses including cell growth, differentiation, and apoptosis (1, 12, 13). The downstream targets of the phosphorylated lipids generated following PI 3-kinase activation have not been well characterized. However some isoforms of protein kinase C are directly activated by phosphatidylinositol-3,4,5-triphosphate in vitro. The protein kinase C-related protein kinase AKT has also been shown to be activated by PI 3-kinase, although the mechanism of this has yet to be determined (14).
PI 3-kinase also appears to be involved in a number of aspects of leukocyte activation. PI 3-kinase physically associates with the cytoplasmic domain of CD28, which is an important co-stimulatory molecule for the activation of T cells in response to antigen (15, 16). Activation of T cells through CD28 lowers the threshold for activation by antigen and increases the magnitude and duration of the proliferative response. These effects are linked to increases in the transcription of a number of genes including the T cell growth factor interleukin 2 (17). Mutation of CD28 such that it can no longer interact with PI 3-kinase leads to a failure to initiate interleukin-2 production, suggesting a critical role for PI 3-kinase in T cell activation (15). Based on studies using the PI 3-kinase inhibitor wortmannin, there is evidence that PI 3-kinase(s) is also required for some aspects of leukocyte signaling through G protein-coupled receptors (18).
We report here the cloning of novel human and murine p110 isoforms
(p110) with a highly restricted pattern of expression. p110
is
expressed predominantly in leukocytes and may therefore play a role in
PI 3-kinase-mediated signaling in the immune system.
Degenerate
oligonucleotide primers were designed for use in the PCR reaction based
on sequences conserved in the catalytic domain of known PI 3-kinases.
The sense primer was 5-GCAGACGGATCCGGIGAYGAYHKIAGRCARGA-3
encoding the sequence GDDLRQD, and the antisense primer was
5
-GCAGACGAATTCRWRICCRAARTCIRYRTG-3
(where I is inosine, R
is A or G, Y is C or T, H is A, C, or T, W is A or T, and K is G or T)
encoding the amino acid sequence HIDFGH. BamHI and
EcoRI restriction sites are underlined. PCR reactions
consisted of 100 ng of cDNA template (from human peripheral blood
mononuclear cells (PBMC) activated for 18 h with 10 ng/ml phorbol
myristate acetate and 250 ng/ml calcium ionophore (Sigma)), 10 µg/ml
oligonucleotide primers, 50 mM KCl, 10 mM
Tris-HCl (pH 8.4), 1.5 mM MgCl2, 200 mM dNTPs, and 1 unit of Taq polymerase in a
final volume of 100 µl. Reactions were performed using denaturation for 1 min at 94 °C, annealing at 60 °C for 2 min, and 3 cycles of
extension for 1 min at 72 °C. The procedure was then repeated using
a 56 °C annealing temperature for 3 cycles, 52 °C annealing temperature for 3 cycles, and 50 °C annealing temperature for 30 cycles. Amplified products were gel-purified, digested with BamHI and EcoRI, and subcloned into the vector
pBluescript SKII+ (Stratagene) for sequencing. All DNA for sequencing
was prepared using the Wizard Miniprep DNA purification system
(Promega, Madison, WI). Sequencing was performed on the Applied
Biosystems model 373 automated sequencer. Data bank searches were
performed using the BLAST program, and protein and DNA alignments were
made using the Geneworks program (Intelligenetics Inc., Mountain View,
CA). One clone contained a 399-base pair insert that encoded a
133-amino acid open reading frame showing ~80% identity with
p110
.
To identify a full-length cDNA (which we subsequently
termed p110), specific oligonucleotide primers were designed based on the sequence of the PCR product. The forward primer was
5
-CATGCTGACCCTGCAGATGAT-3
, and the reverse primer was
5
-AACAGCTGCCCACTCTCTCGG-3
. These were used to screen the
aforementioned human PBMC cDNA library. Successive rounds of PCR
were first performed on pools of 100,000 clones and subsequently on
smaller pools until a single clone (termed PBMC 249) was isolated by
colony hybridization using the PCR product (labeled by random priming)
as a probe. This cDNA was not full-length. Therefore, to identify
longer cDNA clones, the same approach was used to screen a cDNA
library from human macrophages (19). This led to the isolation of an
additional cDNA (M928) that extended the cDNA sequence by 1302 base pairs at the 5
end. The remaining 5
end of the cDNA was
obtained by 5
RACE PCR (CLONTECH, Palo Alto, CA.)
Two antisense cDNA-specific oligonucleotide primers were designed
at the 5
end of cDNA M928 for RACE PCR reactions. The primary RACE
primer was 5
-GGGCCACATGTAGAGGCAGCGTTCCC-3
. The nested RACE primer was
5
-GGCCCAGGCAATGGGGCAGTCCGCC-3
. Marathon RACE reactions were set up
using a human leukocyte Marathon-ready cDNA template and the
Advantage core PCR reaction kit (CLONTECH) following manufacturer protocol. Touchdown PCR cycling conditions were
modified to improve the specificity of the Marathon RACE PCR primary
reaction as follows: denaturation at 94 °C for 2 min followed by 5 cycles of 94 °C for 30 s, annealing and extension at 72 °C
for 3 min, 5 cycles of 94 °C for 30 s, annealing and extension
at 70 °C for 3 min, and 25 cycles of 94 °C for 30 s, annealing and extension at 68 °C for 3 min.
The 5 RACE PCR products were gel-purified and subcloned into the TA
vector PCRII (Invitrogen, San Diego, CA) according to manufacturer
instructions. Three independent clones were sequenced.
A full-length cDNA for p110 was assembled
from clones 249, 928, and the 5
RACE PCR products. The 5
RACE product
ODH18 was used as a template in the PCR using the primers ODH5
FLAG,
5
-AGTTACGGATCCGGCACCATG(GACTACAAGGACGACGATGACAAG)CCCCCTGGGGTGGACTGCCC-3
and ODH3
Paste 5
-CCACATGTAGAGGCAGCGTTCC-3
. The 5
primer
includes a BamHI site (underlined) and encodes the peptide
sequence DYKDDDDK (shown in parentheses), which is recognized by the M2
anti-FLAG monoclonal antibody (Kodak Scientific Imaging Systems). The
resulting PCR product was digested with BamHI and
AflII and was ligated along with an
AflII/PvuI fragment derived from the clone M928 and a PvuII/XbaI fragment derived from PBMC clone
249 into the BamHI/XbaI sites of the mammalian
expression vector pcDNA3 (Invitrogen). The resulting plasmid,
(pcDNA3:FLAG/p110
), uses the cytomegalovirus promoter to drive
the expression of full-length p110
preceded by an initiating
methionine residue and the FLAG epitope.
cDNA encoding a fragment (amino acids
141-291) of mouse p110 termed Rip36 (Ras-interacting protein) was
identified by screening a mouse embryo library using
Ha-RasV12 as bait in the yeast two-hybrid system. The yeast
two-hybrid system and the plasmids have been described (20). The 5
end of the cDNA was obtained by RACE PCR using the reaction conditions described in Vojtek et al. (21) and the following
specific oligonucleotides: 1) for the reverse transcription
reaction, 5
-CGCGGATCCTGCTGACACGCAATAAGCCG-3
; 2)
cDNA-specific primer 1, 5
-TAGGCACCTGCAGATGTACTG-3
;
and 3) cDNA-specific primer 2, 5
-CGCGGATCCTGCTGACACGCAATAAGCCG-3
.
The template for the reverse transcription reaction was spleen
RNA from adult mouse, prepared as described (21). Sequence analysis of
cDNAs present after the RT-PCR reaction identified 558 nucleotides of p110
coding sequence, corresponding to amino acids 1-186 and 36 nucleotides of 5
-untranslated region. An in-frame stop is present 12 nucleotides upstream of the predicted initiating ATG.
The 3 end of murine p110
was obtained by RACE PCR using the Expand
PCR system (Boehringer Mannheim) according to manufacturer instructions. The cDNA-specific oligonucleotides used for
amplification are 5
-GCCAGTTTTGTGAAGAGGCTG-3
and
5
-AGCGGATCCCAGTACATCTGCAGGTGCCTAC-3
. A product of 3.6 kilobases was
subcloned, restriction mapped, and sequenced. This cDNA contains
2,415 nucleotides of p110
coding sequence followed by
3
-untranslated region.
A full-length cDNA for the mouse p110 was obtained by reverse
transcription of spleen RNA from adult mice followed by 35 cycles of
PCR using the Expand system (Boehringer Mannheim) according to
manufacturer instructions and the following two
cDNA-specific oligonucleotides:
5
-GGAAGATCTTGGCGATGCCCCCTGGGGTGGACTGC-3
and 5
-GGAAGATCTGCGGCCGCCTACTGTCGGTTATCCTTG-3
. The latter
oligonucleotide was also used in the reverse transcription reaction.
The reverse transcription reaction contained the following
components/10 µl reaction: 2 µg of total spleen RNA from adult
mice, 4 pmol of the gene-specific primer, 1 µl of 10 mM
dNTPs (U. S. Biochemical Corp.), 1 unit of RNasin (Promega), 200 units
of Superscript reverse transcriptase (Life Technologies, Inc.), 2 µl
of 5 × RT buffer supplied by the manufacturer (Life
Technologies), 5 mM dithiothreitol. The RT reaction was
incubated at 44 °C for 2 h. Five µl of the RT reaction was
amplified by PCR in a reaction volume of 50 µl using the Expand
system (Boehringer Mannheim). The reaction conditions were denaturation
at 94 °C for 2 min followed by 35 cycles of denaturation at 94 °C
for 20 s, annealing at 57 °C for 30 s, and extension at
68 °C for 3.5 min. The cDNA was subcloned directly to the pT7
blue T vector (Novagen) and subsequently transferred to the mammalian
expression vector pEBG, which drives the expression of a glutathione
S-transferase fusion protein, including the complete coding
sequence of mouse p110
(22).
Radiolabeled
[32P]cDNA probes were prepared by PCR using 10 ng of
plasmid DNA template encoding p110. The forward primer was 5
-CTGCCATGTTGCTCTTGTTGA-3
, and the reverse primer was
5
-GAGTTCGACATCAACATC-3
. PCR reactions contained 50 mM
KCl, 10 mM Tris-HCl (pH 8.4), 1.5 mM
MgCl2, 200 µM dATP, dTTP, dGTP, and 1 µM dCTP, 50 µCi of [
-32P]dCTP (DuPont
NEN), 10 µg/ml cDNA-specific primers, and 2 units of
Taq DNA polymerase. Reactions were heated for 4 min at
94 °C followed by 15 cycles of denaturation for 1 min at 94 °C,
annealing for 1 min at 55 °C, and extension for 2 min at 72 °C.
Unincorporated nucleotides were removed by passing the reaction over a
Sephadex G-50 column (Boehringer Mannheim). A multiple tissue northern blot (CLONTECH) was probed and washed under
stringent conditions according to manufacturer recommendations. The
autoradiograph was exposed for 1-4 days at
80 °C with
intensifying screens.
The epitope-tagged p110 was transfected into COS cells
using DEAE dextran (23). Three days after transfection, the cells were
serum-starved overnight in Dulbecco's modified Eagle's medium plus
0.1% fetal bovine serum. The plates were rinsed once with calcium- and
magnesium-free phosphate-buffered saline and lysed in 3 ml of buffer R
per confluent 150-mm plate (buffer R is 1% Triton X-100, 150 mM NaCl, 10 mM Tris (pH 7.4), 1 mM
EDTA, 0.5% Nonidet P-40, 0.2 mM phenylmethylsulfonyl
fluoride, 1% aprotinin). p110
was immunoprecipitated using the
monoclonal antibody anti-FLAG M2 (Kodak Scientific Imaging) according
to manufacturer recommendations. The immunoprecipitates were washed
three times with buffer R and twice with PAN buffer (100 mM
NaCl, 10 mM PIPES, 20 µg/ml aprotinin) and resuspended in
PAN buffer. One-tenth of the immunoprecipitates were incubated for 15 min at 30 °C in a total volume of 10 µl containing 0.2 mg/ml
phosphatidylinositol (Sigma), 20 mM HEPES (pH 7.4), 5 mM MnCl2, 0.45 mM EGTA, 10 µCi of
[
-32P]ATP, 10 µM ATP. The reactions were
terminated by the addition of 100 µl of 1 M HCl. The
phospholipids were extracted once with 200 µl of chloroform/methanol
(1:1 v/v) and once with 80 µl of HCl/methanol (1:1), lyophilized to
dryness, resuspended in 10 µl of chloroform/methanol, and spotted
onto a 1% potassium oxalate-impregnated silica 60 thin layer
chromatography plate (V. W. R. Scientific). The phospholipids were
resolved by ascending chromatography in chloroform, methanol, 4 M NH4OH (9:7:2) and visualized by
autoradiography. Crude phospholipid standards (Sigma) were run in
parallel with the radiolabeled samples and visualized by exposing the
plate to iodine vapor.
Using a strategy based on amplification of conserved PI 3-kinase
sequences, we have identified a novel human member of this family that
we have termed p110. A combination of cDNA library screening and
5
RACE PCR has led to the identification of cDNAs encompassing the
complete coding region of p110
. The deduced amino acid sequence of
p110
is shown in Fig. 1.
In an independent search for mouse Ha-RasV12-interacting
proteins (Rips) using the yeast two-hybrid system, we identified two clones related in sequence to p110: Rip31 and Rip36. Using a PCR-based strategy and gene-specific oligonucleotide primers derived from the Rip36 sequence, a full-length cDNA was isolated (see "Materials and Methods"). Sequence analysis suggests that this clone is the murine p110
, since it shares 94% amino acid sequence identity with human p110
(compared with 56% identity between human
p110
and
; an alignment of human and mouse p110
is shown in
Fig. 1) and has a similar pattern of expression in vivo (see below).
The sequences of human and murine p110 include open reading frames
predicted to encode for proteins of 1044 and 1043 amino acids,
respectively, with an expected molecular mass of 119,505 Da for the
human clone (~120 kDa). The sequences around the predicted initiating
methionines are in good agreement with that required for optimal
translational initiation (24). The presence of stop codons in the
5
-untranslated sequence is consistent with isolation of the complete
coding region of p110
(data not shown). Consistent with the
predicted size of the encoded translation product, Western blotting of
immunoprecipitates from COS cells transiently transfected with an
epitope-tagged form of human p110
detected a protein of ~110 kDa
(see below).
Comparison of the sequence of the carboxyl-terminal catalytic domain of
p110 with those of other PI 3-kinases reveals that it is most
closely related to p110
. p110
is 72% identical to p110
in
this region and is less closely related to p110
(49%) or p110
(45%), whereas cpk/p170 and the yeast Vps34 protein show the lowest
homology (31 and 32%, respectively). This is confirmed by dendrogram
analysis; p110
and p110
form a distinct sub-branch of the PI
3-kinase family (Fig. 2). The distantly related ataxia telangiectasia gene product and the catalytic subunit of
DNA-dependent protein kinase have been included for
comparison (Fig. 2). These proteins are structurally related to PI
3-kinases and have protein kinase activity (25) but have not yet been
shown to possess lipid kinase activity (26, 27).
p110 and p110
form heterodimers with a common p85 subunit (3-5,
8). We examined the association of recombinant p110
with p85. When
either epitope-tagged human (Fig. 3, panel B)
or mouse (Fig. 4, lane 4) p110
was
expressed in COS or 293T cells and recovered by immunoprecipiation, a
85-kDa protein was detected with p85-specific antiserum in the
immunoprecipitates. Thus, both human and murine p110
associate with
endogenous p85 after transfection into COS or 293T cells, respectively.
Association of human p110
with p85 could also be detected following
expression of epitope-tagged p110
in the lymphoid cell line Jurkat
(data not shown). The association of p110
, p110
, and p110
with
p85 is consistent with the presence of a conserved p85 interaction
domain at the amino terminus of these isoforms. This region is lacking
in p110
, which is targeted to the plasma membrane via its
interaction with the
/
subunits of heterotrimeric G proteins.
This interaction is dependent on a p110
-specific adaptor protein,
p101 (28).
It has been demonstrated that PI 3-kinase is an important
intermediate in the Ras pathway (29, 30). A specific region at the
amino terminus of the p110 subunit, residues 133-314, termed the
Ras regulatory domain (underlined in Fig. 1), is responsible for this interaction (30). Comparison of the sequence of both human and
murine p110
with other p110 subunits indicates that this region is
also conserved in p110
, including a lysine residue (residues 227 of
p110
and 223 of p110
), which has been shown to be essential for
physical association with Ras (30). Moreover, a relatively short amino
acid domain of p110
(amino acids 141-291, the amino acids of
p110
encoded by the Rip36 clone) is sufficient to interact with
Ha-RasV12 in vivo in the yeast two-hybrid system
and further delineates the Ras regulatory domain to a 151-amino acid
region. The interaction of Ha-RasV12 with this domain of
p110
requires active Ras and an intact Ras effector domain (data not
shown). Thus, p110
is likely to mediate some of the effects of Ras,
although the p110
Ras-interacting region is less conserved than the
putative p85 binding site or catalytic domain.
Whereas the activation of PI 3-kinase in a wide range of biological
systems has been extensively studied, less is known concerning the cell
type-specific expression of particular p110 isoforms. Northern blot
analysis of the expression of p110 in human and murine tissues
reveals a single transcript of ~5.4 kilobases (consistent with the
size of the composite cDNAs). In humans, the highest levels of
expression are seen in PBMC, spleen, and thymus (Fig. 5). After prolonged exposure of the autoradiograph, low
levels of p110
expression could be detected in testes, uterus,
colon, and small intestine but not in other tissues examined including prostate, heart, brain, and liver (data not shown.) p110
is also abundantly expressed in adult mouse spleen as well as in testis (Fig.
5B). The elevated expression of p110
mRNA in mouse
but not human testes is noteworthy and may reflect a true difference in
expression between species. Alternatively, a number of genes is
expressed specifically at elevated levels in postmeiotic haploid cells
in the testis. Abnormally sized transcripts that may be more stable
than the transcripts found in diploid cells are commonly found.
However, it is not clear whether these transcripts are translated (Ref.
31 and references cited therein). In the case of p110
, the RNA
expressed in the testis may or may not be translated. If p110
protein is expressed, then it is possible that the protein has a
specific role in development of the male germ line. The restricted
expression of p110
is in contrast to p110
and p110
, which
appear to be widely expressed (3, 8).
To test whether p110 has PI 3-kinase activity,
immunoprecipitates from COS cells transfected with epitope-tagged
p110
were incubated with [32P]ATP and
phosphatidylinositol, and the radiolabeled phospholipids were resolved
by chromatography. A product was detected that migrates slightly slower
than the PI 4-phosphate (PIP) standard, consistent with the
generation of PI 3-phosphate (32) (Fig. 6). This enzyme activity was sensitive to the PI 3-kinase inhibitor wortmannin (data
not shown). Similar results were obtained when purified phosphatidylinositol was used as a substrate (data not shown). Whereas
these results demonstrate that the cDNA clone that we isolated
encodes a functional PI 3-kinase, it cannot be assumed that the
in vitro substrate specificity of a particular isoform reflects its activity on membrane lipids in intact cells (11).
The interaction of multiple p110 catalytic subunit isoforms, p110,
p110
, and p110
, with a common adaptor protein, p85, suggests that
the nature of the phosphorylated lipids generated in response to a
particular agonist may be regulated at least in part by the
cell/tissue-specific expression of the different isoforms of the
catalytic subunit. In cells such as leukocytes, where it is likely that
multiple p110 isoforms are expressed, it will be of interest to
determine the relative contribution made by these multiple isoforms to
processes such as cell activation and migration.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) U86587 and U86453.
We thank David Turner, Bart Vanhaesebroeck, and Pablo Rodriguez-Viciana for discussions and Johnny Stine for help in preparing the figures.