(Received for publication, February 27, 1997)
From the Ludwig Institute for Cancer Research, 91 Riding House Street, London W1P 8BT, United Kingdom and the
§ MRC Graduate Programme and the ¶ Department of
Biochemistry and Molecular Biology, University College London, Gower
Street, London WC1E 6BT, United Kingdom
The mammalian phosphoinositide 3-kinases (PI3Ks)
p110,
, and
form heterodimers with Src homology 2 (SH2)
domain-containing adaptors such as p85
or p55PIK.
The two SH2 domains of these adaptors bind to phosphotyrosine residues
(pY) found within the consensus sequence pYXXM. Here we
show that a heterodimer of the Drosophila PI3K, Dp110, with an adaptor, p60, can be purified from S2 cells with a pYXXM
phosphopeptide affinity matrix. Using amino acid sequence from the
gel-purified protein, the gene encoding p60 was cloned and mapped to
the genomic region 21B8-C1, and the exon/intron structure was
determined. p60 contains two SH2 domains and an inter-SH2 domain but
lacks the SH3 and breakpoint cluster region homology (BH) domains found in mammalian p85
and
. Analysis of the sequence of p60 shows that
the amino acids responsible for the SH2 domain binding specificity in
mammalian p85
are conserved and predicts that the inter-SH2 domain
has a coiled-coil structure. The Dp110·p60 complex was immunoprecipitated with p60-specific antisera and shown to possess both
lipid and protein kinase activity. The complex was found in larvae,
pupae, and adults, consistent with p60 functioning as the adaptor for
Dp110 throughout the Drosophila life cycle.
Studies in vertebrates, Drosophila melanogaster and
Caenorhabditis elegans, suggest that both the structure and
the function of receptor tyrosine kinases
(RTKs)1 are highly conserved across
metazoan organisms (1). Upon stimulation with an extracellular ligand,
RTKs dimerize and transphosphorylate (2). This tyrosine
phosphorylation enables the recruitment of signaling molecules
containing SH2 or phosphotyrosine binding (PTB) domains that recognize
phosphotyrosines within specific amino acid motifs (3). In this way,
the SH2 domain-containing adaptors for Class IA PI3Ks (43)
are recruited to tyrosine-phosphorylated RTKs and associated substrates
containing the pYXXM motif (4). The recruitment of Class
IA PI3Ks to activated RTKs coincides with a dramatic
increase in the production of 3 phosphorylated phosphoinositides.
These 3
phosphorylated phosphoinositides are thought to act as second
messengers that affect cell growth, differentiation, membrane
trafficking, and cytoskeletal organization (4).
In mammals, there are at least three Class IA PI3Ks,
p110,
, and
(5, 6, 44) that can associate with a number of adaptors. Three distinct genes encode p85
, p85
, and
p55PIK, and additional adaptors are generated from
alternatively spliced p85
transcripts (7-13). Each of these
adaptors contains two SH2 domains, both of which selectively bind
peptides containing phosphotyrosine with a methionine at the +3
position (pYXXM) (14, 15), and an inter-SH2 domain, which
mediates binding to Class IA PI3Ks (16). The p85
and
p85
adaptors also contain an SH3 domain and a BH domain at the N
terminus, whereas p55PIK and two splice variants of p85
have short N-terminal extensions. Despite this structural diversity,
there has been no reported selectivity of binding between different
adaptors and p110
,
, or
(44).
We are using Drosophila to examine the role of the Class IA PI3Ks genetically and to provide an in vivo system to identify downstream targets. Many molecules downstream of RTKs in Drosophila are structurally and functionally homologous to their mammalian counterparts. The best characterized example of this is the Ras/MAPK pathway downstream of the Sevenless, Drosophila EGF receptor and Torso RTKs (17). Drosophila possesses a Class IA PI3K, Dp110 (also known as PI3K_92E), which is homologous to mammalian Class IA PI3Ks (18, 19). Previously, we have shown that the ectopic expression of Dp110 in larval imaginal discs affects cell growth but not cell differentiation (19).
Here we present the affinity purification of p60, the adaptor for Dp110, using immobilized phosphopeptides containing the pYXXM motif. Peptides derived from the purified protein were sequenced, and degenerate PCR and cDNA cloning were used to isolate the p60 cDNA. The structural and functional conservation of p60 with the mammalian adaptors is discussed. The Dp110·p60 complex possessed lipid and protein kinase activity and was found in Drosophila larvae, pupae, and adults and the S2 cell line. The genomic structure and location of the p60 gene has been determined to facilitate the identification and generation of genetic reagents to study the function of this PI3K adaptor in vivo.
S2 cell
lysates were prepared essentially as described (19) by lysis in buffer
containing 1% Triton X-100 and the protease inhibitors, 1 mM phenylmethylsulfonyl fluoride, 18 µg/ml aprotinin, 2 µg/ml leupeptin, 1 µg/ml pepstatin, 5 mM benzamidine.
Detergent lysates of Oregon R third instar larvae, pupae, and adult
flies were prepared in the same buffer, but which additionally
contained 5 µM diisopropylfluorophosphate and 15 µM
N--p-tosyl-L-lysine chloromethyl
ketone, by homogenization with a polytron (2 × 10 s, on
ice). The homogenate was clarified by centrifugation at 20,000 × g for 1 h and then at 20,000 × g for
30 min. Affinity purification using phosphopeptide coupled to Actigel
(Sterogene) beads was performed as described (19).
Proteins
affinity purified from approximately 1010 S2 cells were
resolved by SDS-PAGE and stained with Coomassie Blue R250. Each band
was excised and digested with lysyl endopeptidase C (Wako Chemicals),
and the resulting peptides were resolved by passage through an AX-300
pre-column and then a C18 (150 × 1 mm) column (Relasil). The
peptides were sequenced at the 500 fmol level using an ABI Procise
system. Peptides 6 and 9 (see Fig. 2A) were used to design
the degenerate primers, CA(A/G)GA(A/G)(C/T)TITT(C/T)CA(C/T)TA(C/T)ATGGA and (C/G)(A/T)IACIGT(C/T)TC(A/G)TAIA(A/G)(C/T)TG(C/T)TG, respectively. Poly(A)+ mRNA was prepared from S2 cells using
oligo(dT)-cellulose (Stratagene) and used to synthesize first strand
cDNA with Moloney murine leukemia virus reverse transcriptase
(Pharmacia Biotech Inc.). PCR reactions were performed with cDNA as
a template and 0.025 units/µl Taq, 1.5 mM
MgCl2, 0.2 mM dNTPs, and 1 µM
each primer in a total volume of 50 µl. 40 cycles of amplification
(94 °C for 30 s, 50 °C for 30 s, and 72 °C for
60 s) were performed. The 200-base pair product obtained was
cloned into pGEM-T (Promega), sequenced, and found to encode peptides 7 and 8 (see Fig. 2A). This fragment was used to screen a
gt10 eye imaginal disc cDNA library (A. Cowan), and four
positive clones were isolated that appeared to be identical by
restriction mapping and by sequencing of the 5
and 3
ends. One clone
was digested with EcoRI, and the two resulting fragments were subcloned into pBluescript SKII (Stratagene). PCR and direct sequencing of the original lambda clone showed the two fragments to be
adjacent. Each fragment was sequenced in both directions with T3, T7,
and p60-specific primers on an ABI 373 automated DNA sequencer.
Immunological Methods
Peptides corresponding to the N
terminus (CGGMQPSPLHYSTMRPQ, CGGSLVDPNEDELRMA) and the C terminus
(CGGLYWKNNPLQVQMIQLQE, CGGSLEAEAAPASISPSNFSTSQ) of p60, including a CGG
coupling linker at the N terminus, were coupled to maleimide-activated
keyhole limpet hemocyanin as directed (Pierce). Antibodies were raised
in rabbits against pools of N-terminal (p60N) and C-terminal
antigens (
p60C). Immunoblotting was performed using
Dp110
(1:1000) (19),
p60N (1:1000), or
p60C (1:2000) and developed with
enhanced chemiluminescence as directed (Amersham Life Science).
Immunoprecipitation was performed by incubating the lysate with a 1:200
dilution of antisera (6 µl) for 1 h at 4 °C and then adding
protein A-Sepharose (Pharmacia) beads and incubating for an additional
30 min. The beads were washed in the same manner as the peptide-coupled
beads (19).
For the lipid and protein kinase assays, the
beads were washed three times in lysis buffer and twice in 20 mM Tris-HCl (pH 7.4), 100 mM NaCl, 0.1 mM EGTA. Lipid kinase assays were performed essentially as
described (20) in 60 µl of 20 mM Tris-HCl (pH 7.4), 100 mM NaCl, 0.1 mM EGTA, 2.5 mM
MgCl2, 100 µM ATP containing 2.5 µCi
[-32P]ATP and 200 µM sonicated
phosphatidylinositol (Sigma). The reaction was incubated for 30 min at
room temperature and terminated with acidified chloroform, and the
lipid was extracted and resolved by thin layer chromatography with
chloroform/methanol/4 M ammonium hydroxide (45:35:10).
Protein kinase assays were performed in 30 µl of 20 mM
Tris-HCl (pH 7.4), 100 mM NaCl, 0.1 mM EGTA,
2.5 mM MgCl2, 100 µM ATP
containing 2.5 µCi [
-32P]ATP for 30 min at room
temperature and resolved by SDS-PAGE on 7.5% polyacrylamide gels.
Phosphopeptides, containing one or both of the
pYXXM motifs found at positions 740 and 751 of the human
PDGF receptor, bind selectively to the adaptors for Class
IA PI3Ks (14). When coupled to agarose beads, these
phosphopeptides can be used to affinity purify heterodimeric complexes
containing the adaptors bound to Class IA PI3Ks (9, 21). We
investigated whether this approach could be used to identify an adaptor
for Dp110, the Drosophila Class IA PI3K. Three
proteins of approximately 145, 120, and 60 kDa were affinity purified
from Drosophila S2 cells, using the tyrosine phosphorylated
peptide GGYMDMSKDESVDpYVPML (pY751) coupled to agarose beads (Fig.
1). The same proteins were purified when the peptide was
phosphorylated on tyrosine 740 or on both tyrosine 740 and tyrosine
751, but they were not recovered with beads lacking peptide (data not
shown). The affinity purified complex possessed lipid and protein
kinase activities, and immunoblotting with
Dp110 antisera showed
that the 120-kDa protein was Dp110 (see below). We washed the complex
at high stringency to determine which of the remaining bands was the
adaptor for Dp110. Washing with lysis buffer containing increasing
concentrations of sodium chloride removed the majority of the 145-kDa
band, suggesting that p60 was the adaptor (Fig. 1). A large scale
affinity purification was performed, and peptides derived from p60,
p120, and p145 were sequenced (see "Experimental Procedures").
Three of the peptide sequences obtained from p120 confirmed that it was
Dp110 (LMANYTGL, EYQVYGISTFN, and LHVLE). Two peptides sequenced from
p145, FMEXIYTDVR and FXNNXXCGYIL, revealed that this protein was
Drosophila phospholipase C
(PLC
D) (22). Interestingly,
human PLC
can be affinity purified from human cell lines using the
same pYXXM phosphopeptide2, and
the C-terminal SH2 domain of mammalian PLC
can interact with
pYXXM motifs (14) though it binds preferentially to
phosphotyrosines in other sequence contexts (15). Nine peptide
sequences were obtained from p60 and used to design degenerate PCR
primers. PCR amplification from first strand cDNA derived from S2
cell mRNA generated a 200-base pair fragment that was used to
isolate p60 cDNAs from an eye imaginal disc cDNA library (see
"Experimental Procedures"). These cDNAs contained an open
reading frame that could encode a protein with a predicted size of 57.5 kDa, which contains all nine of the peptide sequences recovered from
p60 (Fig. 2A).
Sequence Analysis of p60
Like the other identified adaptors
for Class IA PI3Ks, the predicted amino acid sequence of
p60 includes two SH2 domains and an inter-SH2 domain. However, the SH3
and BH domains in p85 and
, the N-terminal extensions in
p55PIK and the p55
splice variants, and the proline-rich
SH3 domain-binding motifs (23) found in all mammalian adaptors are
absent in p60 (Fig. 2B). p60 has a short N terminus (similar
in size to the N terminus of p50
, a recently isolated splice variant
of p85
(12)), and a unique C terminus of 70 amino acids that shows no significant similarity to other proteins. When the amino acid sequences of the core SH2-inter-SH2-SH2 region of p60, p85
,
p85
, and p55PIK are compared, p60 shows an equal degree
of similarity to all three mammalian adaptors (Fig. 2A, data
not shown).
The N-terminal and C-terminal SH2 domains are the most conserved
regions of p60 and are 58% and 48% identical to the respective domains of bovine p85. The three-dimensional structures of the N-terminal and C-terminal SH2 domains of p85
in complex with pYXXM phosphopeptides have been determined by x-ray
crystallography and nuclear magnetic resonance (24, 25). These SH2
domain structures identify the amino acids responsible for the
pYXXM binding specificity. These amino acids, including the
phenylalanine of the beta strand E and the leucines of the loop between
the alpha helix B and the beta strand G, are conserved in p60.
Together, these three amino acids, shown in Fig. 2A, define
the hydrophobic pocket that allows the specific binding of methionine
three residues C-terminal to the phosphotyrosine (26).
The inter-SH2 domain of mammalian p85 mediates binding to the Class
IA PI3K, p110
. Modelling studies of this inter-SH2
domain predict a two- or four-helix antiparallel coiled-coil structure similar to the solved crystal structure of the inter-SH2 domain of
ZAP-70 (16, 27).3 The inter-SH2 domain of
p60 is approximately 20% identical to the corresponding region of the
mammalian adaptors. Despite this low homology, this region of p60 is
likely to form a similar structure since it contains the leucine-rich
heptad repeats characteristic of coiled-coil alpha helical bundles (see
Fig. 2). Furthermore, BLITZ and BLASTP data base searches (28, 29) show
that this region of p60 is significantly homologous to coiled-coil
regions of proteins that form stable heterodimers (data not shown).
A BLASTN (28)
data base search with the p60 nucleotide sequence identified a sequence
tagged site (STS, Dm0574) from the Berkeley Drosophila
Genome Project that is identical to the 3-untranslated region of the
p60 cDNA and has been mapped to the genomic region, 21B6-C2 (30).
To further characterize the genomic structure of the p60 gene, we
performed Southern analysis of a
-phage contig of the region (kindly
provided by M. Noll). This analysis determined the position and
orientation of the p60 gene with respect to the published
EcoRI restriction map (31). Furthermore, the exon/intron structure of the gene was determined by PCR, subcloning, and sequence analysis of the genomic clones (Fig. 2C). The gene encoding
p60 has three exons and probably overlaps the breakpoint of the
deficiency Df(2L)al at 21B8-C1.
We initially purified p60 from S2 cells, an embryonically
derived cell line. We next sought to characterize the expression of p60
at different stages of the Drosophila life cycle. The
Dp110·p60 complex was affinity purified from Triton X-100 lysates of
third instar larvae, pupae, and adult flies using pYXXM
phosphopeptide beads (Fig. 3A). The complex
is present in all stages examined although we consistently recovered
lower levels from larvae than from the other stages. Immunoblotting
with antisera against N- and C-terminal sequences of p60 and against
Dp110 confirmed the presence of these proteins (Fig. 3B). An
additional 55-kDa band can be seen that is immunoreactive with p60N
but not
p60C (Fig. 3, p60*). This protein might be a form
of p60 that has been degraded from the C terminus, a splice variant of
p60 lacking the C terminus or the product of another gene that
cross-reacts with
p60N. We believe that this protein is a degraded
form of p60 since its appearance coincides with lower levels of
full-length p60 and because lysate preparation in the absence of
certain protease inhibitors or following the freezing and thawing of
samples resulted in increased levels of this smaller band (data not
shown). In addition, the exon/intron structure of the p60 gene
indicates that the 55-kDa band is unlikely to be encoded by an
alternatively spliced transcript (Fig. 2C).
Enzymatic Activity of the Dp110·p60 Complex
The Dp110·p60
complex possesses lipid and protein kinase activity when affinity
purified from S2 cells with phosphopeptide beads. We confirmed these
results using immunoprecipitation as an independent method of purifying
Dp110·p60. Silver staining and immunoblots show that both p60N and
p60C can immunoprecipitate the Dp110·p60 complex though
p60C is
more efficient than
p60N (data not shown). Dp110, whether
immunoprecipitated or affinity purified, is able to autophosphorylate
and, to a lesser degree, to transphosphorylate p60 (Fig.
4A). The autophosphorylation is comparable to
that shown by human p110
(44), whereas the phosphorylation of p60 is
reminiscent of the phosphorylation of p85
by mammalian p110
(32).
However, p60 does not contain a phosphorylation site homologous to
serine 608 in p85
(32) (Fig. 2A). Consistent with other
Class IA PI3Ks in complex with their adaptors, both the
immunoprecipitated Dp110·p60 complex and the affinity purified complex possessed lipid kinase activity as assessed by the conversion of phosphatidylinositol to phosphatidylinositol 3-phosphate (Fig. 4B).
p60 has been affinity purified from Drosophila with a pYXXM phosphopeptide previously used to purify mammalian adaptors for Class IA PI3Ks. Analysis of the p60 amino acid sequence and the lipid and protein kinase activity of the Dp110·p60 complex indicates that p60 is both structurally and functionally homologous to the mammalian adaptors for Class IA PI3Ks. Since p60 is the most divergent member of the family identified to date, its sequence provides an insight into the evolution of the structure and function of these molecules. Notably, the residues responsible for the SH2 domain binding specificity are conserved, and the prediction of a coiled-coil structure for the inter-SH2 domains of the adaptor subunits for Class IA PI3Ks is supported.
Since we have shown that the pYXXM phosphopeptide can be used to purify adaptors in complex with Class IA PI3Ks from mammals and Drosophila, affinity purification with this phosphopeptide might also be used to isolate homologous PI3K complexes from many species. Interestingly, the SH2 domain of the Drosophila signaling molecule Drk also binds to the same phosphotyrosine motif recognized by its mammalian homologue Grb2 (33), Thus, affinity purification with peptides containing specific phosphotyrosine motifs might be used to isolate SH2 domain-containing proteins from various organisms.
Putative YXXM docking sites for the Dp110·p60 complex are
found in the RTK Dret, the RTK substrate, Dos, and the
Drosophila homologue of the insulin receptor, INR (34-36).
However, it remains to be shown, either biochemically or genetically,
whether these motifs are used in vivo. INR contains three
YXXM motifs and is able to bind the N-terminal SH2 domain of
mammalian p85 when phosphorylated (36). Class IA PI3Ks
associate with the mammalian insulin receptor via multiple
pYXXM motifs in its substrate IRS-1 and are thought to
mediate many of the effects of insulin stimulation (37). Consistent
with an analogous role for Class IA PI3Ks downstream of the
Drosophila INR, certain mutations in inr and the
ectopic expression of dominant negative Dp110 both affect imaginal disc cell growth (19, 38). It must be noted that mammalian adaptors for
Class IA PI3K can also bind to the pYVXV motif
in the Met receptor though with a lower affinity than for
pYXXM motifs (39). Therefore, it is possible that p60 might
recognize phosphotyrosine binding sites other than
pYXXM.
It is likely that p60 is the only adaptor for Class IA
PI3Ks present in Drosophila. However, we cannot rule out the
possibility that additional adaptors exist. Together with Dp110, p60
was the predominant protein that was affinity purified from
Drosophila with the pYXXM phosphopeptide (Fig.
3A). Immunoblotting of the affinity purified material with
both p60N and
p60C did not detect any larger splice variants of
p60 that might contain SH3 or BH domains. Furthermore, we have not
found exons encoding these domains when sequencing genomic DNA 9 kilobases upstream of the most 5
-exon of p60. Similarly, probing
Northern blots with the p60 cDNA revealed only one band (data not
shown). Thus, we conclude that any additional adaptors for Class
IA PI3Ks that exist in Drosophila must be
present at very low levels in the tissues that we have examined and/or other have a highly restricted expression pattern. Degenerate PCR and
extensive cDNA library screening (18, 19, 40) have revealed only
one member of each class of PI3K in Drosophila, suggesting
that, in common with other gene families, there are less PI3K isoforms
in Drosophila than in mammals.
If p60 is the only adaptor for Class IA PI3Ks in
Drosophila, this implies that the adaptor of the common
ancestor of vertebrates and flies consisted of the core
SH2-inter-SH2-SH2 region. Even though SH3 domains are found in other
Drosophila signaling molecules, for example, Drk (41), their
absence in p60 suggests that the SH3 and BH domains found in mammalian
p85 and
are the result of more recent evolution. p60 also lacks
the SH3 domain-binding, proline-rich sequences found in all mammalian
adaptors for Class IA PI3Ks. This again suggests that these
motifs are involved in a more recently evolved mode of regulation that
might be related to the presence of the SH3 domain found in p85
and
.
We thank Nick Totty, Alistair Sterling, Hans Hansen, and Justin Hsuan for peptide microsequencing and Marina Cotsiki and Christopher Odell for DNA sequencing. We are grateful to Yamanouchi Pharmaceutical Co. Ltd for phosphopeptides. We thank Joy Alcedo and Marcus Noll for providing genomic clones and restriction maps. We thank Thomas Twardzik, Thomas Raabe, and Martin Heisenberg for helpful discussions, and we are grateful to Ivan Gout, Markéta Zvelebil, Bart Vanhaesebroeck, Kyoichiro Higashi, and Ernst Hafen for useful discussions.