Gs and the Ras binding domain of p110 are both important regulators of PI3K signalling in neutrophils

Sabine Suire^1, 4, Alison M. Condliffe^1, 2, 4, G. John Ferguson¹, Chris D. Ellson^1, 3, Hervé Guillou¹, Keith Davidson¹, Heidi Welch¹, John Coadwell¹, Martin Turner¹, Edwin R. Chilvers², Phillip T. Hawkins^1, 5 & Len Stephens^1, 5

¹ The Babraham Institute, Babraham, Cambs CB22 3AT, UK.

² Respiratory Medicine Division, University of Cambridge School of Clinical Medicine, Addenbrooke's and Papworth Hospitals, Cambridge, UK.

³ Current address: Centre for Cancer Research, 40 Ames Street, Cambridge, MA 02139, USA.

⁴ These authors contributed equally to this work.

⁵ These authors contributed equally to this work.

Through their ability to regulate production of the key lipid messenger PtdIns(3,4,5)P₃, the class I phosphatidylinositol-3-OH kinases (PI(3)Ks) support many critical cell responses^{1, 2}. They, in turn, can be regulated by cell-surface receptors through signals acting on either their adaptor subunits (for example, through phosphotyrosine or Gs) or their catalytic subunits (for example, through GTP-Ras). The relative significance of these controlling inputs is undefined in vivo. Here, we have studied the roles of Gs, the adaptor p101, Ras and the Ras binding domain (RBD) in the control of the class I PI(3)K, PI(3)K, in mouse neutrophils. Loss of p101 leads to major reductions in the accumulation of PtdIns(3,4,5)P₃, activation of protein kinase B (PKB) and in migration towards G-protein activating ligands in vitro, and to an aseptically inflamed peritoneum in vivo. Loss of sensitivity of PI(3)K to Ras unexpectedly caused similar reductions, but additionally caused a substantial loss in production of reactive oxygen species (ROS). We conclude that Gs, p101 and the Ras–RBD interaction all have important roles in the regulation of PI(3)K in vivo and that they can simultaneously, but differentially, control distinct PI(3)K effectors.

Class I PI(3)Ks are heterodimers made up of regulatory and p110 catalytic subunits. The mammalian class I PI(3)Ks are divided into two subfamilies: the class IA PI(3)Ks (named after their catalytic subunits, p110, and ), which are associated with SH2 domain-containing adaptor subunits and can be activated by tyrosine kinase-based mechanisms; and the class IB PI(3)K (p110), which associates with either p84 (ref. 3; also called p87^PIKAP, ref. 4) or p101 adaptors⁵, through which it can be activated by Gs. All class I PI(3)K catalytic subunits can bind, and be activated by, GTP-Ras proteins^{6, 7} (with the possible exception of PI(3)K).

PI(3)K is most highly expressed in neutrophils, where it is predominantly associated with p101 (approximately 95%)³. Biochemical and cellular studies have shown that p101/p110 is substantially activated by Gs and in a manner that is dependent on p101 (refs 5, 9, 10, 11). Recombinant PI(3)K has also been shown to be activated by GTP-Ras proteins (N, K, H and R) directly and independently of other regulators (such as Gs or phosphotyrosine residues^{7, 12}). The crystal structures of p110, both free and bound to active Ras, have shown that five residues in p110 are critical for binding GTP-Ras; substitutions at these sites create a mutant protein (DASAA; T232D, K251A, K254S, K255A and K256A) that can neither bind or be activated by GTP-Ras in vitro or in transfected cells⁷.

Neutrophils are circulating, myeloid-derived cells. Adjacent to sites of injury or infection, neutrophils rapidly exit the circulation and migrate through the surrounding basement membrane towards ongoing damage¹³. This migration process is underpinned by the chemotactic ability of neutrophils to move towards sources of a variety of inflammatory mediators (for example, fMLP, IL-8 MIP-2 and Kc, the murine analogue of Gro-). As neutrophils approach sites of inflammation they become primed through exposure to a number of cytokines released in this context, such as lipopolysaccharide (LPS) and tumour necrosis factor (TNF). This sensitises the neutrophils and dramatically increases their ability to both phagocytose apoptotic cells or pathogens and release ROS that are involved in killing¹⁴. These functional responses enable neutrophils to drive the resolution of inflammatory responses.

Mice survive genetic ablation of PI(3)K; however, PI(3)K- deficient neutrophils fail to respond appropriately to various inflammatory mediators^{15, 16, 17}. Accumulation of PtdIns(3,4,5)P₃, phosphorylation of PKB, production of ROS and chemotaxis in response to ligands such as fMLP and C5a are all severely inhibited. Use of isoform-selective PI(3)K inhibitors has shown that PI(3)K can act as a master kinase that controls subsequent activation of PI(3)Ks and ¹⁸. Detailed study of PI(3)K- deficient mice has shown that this protein has important functions in the heart, lymphocytes, endothelium, platelets and mast cells, indicating the broad significance of these signalling cascades^{19, 20, 21}.

The accepted dogma is that PI(3)K is rapidly activated by Gs that are released by dissociation of G_i-containing G-proteins, because PtdIns(3,4,5)P₃ signals in these responses peak within 12 s and typically decline within 1 min^{22, 23, 24}. The role of Ras is likely to be relatively minor because of these rapid kinetics, but little work has yet addressed the role of Ras in regulation of PI(3)Ks in vivo. We have asked whether there is a role for endogenous Ras in regulating PI(3)K in mouse neutrophils by creating a PI3K^DASAA knock-in.

Published work in the field is consistent with the hypothesis that p101 is needed for G activation of PI(3)K, that it only functions in a complex with PI(3)K and that it is not essential for Ras activation of p110^{5, 7, 11}. These observations suggest loss of p101 would serve as a specific means of assessing the role of Gs, as distinct from Ras, in the activation of PI(3)K. We have created mice lacking p101 in an attempt to address these issues.

Hence, the combined analysis of p101^-/- and PI(3)K^DASAA/DASAA mice also enables a currently unique window into the relative regulatory power of adaptor subunit-mediated versus Ras-mediated signals on class I PI(3)K activity and, furthermore, the potential for an analysis of the ability of a single PI(3)K to differentially control effectors as a function of its mode of activation.

The mouse p101 locus was disrupted in E14 stem cells by the in-frame insertion of a -Gal gene at codon 11. p101^-/- mice were viable, fertile, of normal size and with normal blood counts (Methods and see Supplementary Information, Fig. S1). Western blotting showed that neutrophils from p101^-/- mice contained no detectable p101 and unchanged expression of p110, , or (Fig. 1). Significantly, because the p84 adaptor for PI(3)K is the neighbouring gene to p101 on mouse chromosome 11, expression of p84 message and protein was not reduced; in fact, it was increased by approximately 50% compared with wild type (this should be viewed in the context that approximately 95% of p110 in mouse neutrophils is bound to p101 and 5% to p84).

Figure 1. Expression and distributions of PI(3)K subunits in neutrophils isolated from p101-/- and p110DASAA/DASAA mice. (a) Western immunoblots demonstrating the loss in expression of p101, unchanged expression of PI(3)K catalytic subunits, appearance of -galactosidase and increased expression of p84 in neutrophils from p101-/- mice, compared with p101+/+ mice. (b) Quantification of immunoblots (similar to a) comparing expression of PI(3)K subunits in neutrophils from p101-/- compared with p101+/+ mice. Endogenous GAPDH was quantified in parallel to demonstrate equivalent loading in experiments analysing p110 expression. Data are means s.e.m. (n values are as indicated). (c) Expression of PI(3)K in neutrophils from p110+/+, p110+/- and p110DASAA/DASAA (D/D) mice. Western immunoblots were probed in parallel with anti-p110 and anti-p67phox antibodies and the ratio of p110:p67phox expression was quantified for each individual sample and compared with those from p110+/+ mice (scaled as 1.0). Very similar ratios were obtained if p110 expression was normalised through -COP (coat-protein complex; data not shown). Data are means s.e.m. (n = 3, from 3 independent experiments). (d) Distribution of p110 in neutrophils from p101-/-, p110DASAA/DASAA and wild-type mice. Neutrophils were sonicated to prepare an homogenate (H) and ultracentrifuged to derive cytosolic (C) and membrane fractions (M), which were immunoblotted (C and H, 3.5 105; and M, 3.5 106 cell equivalents per lane, respectively) to quantify p110, G and GAPDH in each of the fractions relative to the starting homogenate (100%). Data are from two experiments and are means s.e.m. (n = 4; see also Supplementary Information, Fig. S3). Full Figure and legend (25K)

A p110^DASAA (p110^DASAA/DASAA) knock-in mouse was created using standard approaches (Methods and see Supplementary Information, Fig. S2). Pure, recombinant 6His–p110^DASAA has the same basal catalytic activity, ability to bind p101 and activation by Gs as 6His–p110, prepared in parallel (see Supplementary Information, Fig. S3). In the context of published data⁷ this suggests that p110^DASAA should be bound by p101 (or p84), and be fully sensitive to Gs but insensitive to Ras. p110^DASAA/DASAA mice were viable, fertile, of normal size and with generally normal blood counts (see Supplementary Information, Fig. S1), and quantitative western blotting indicated that expression of other class I PI(3)K catalytic subunits was normal (data not shown), but expression of p110^DASAA/DASAA was reduced (Fig. 1). This reduction in expression of p110^DASAA/DASAA could be because of minor changes in the stability of the p110^DASAA/DASAA protein that are secondary to the amino-acid substitutions, or the levels of p110 message, caused by the small quantity of foreign, intronic sequence remaining after the gene targeting. To control for any potentially confounding effects that these small reductions in PI(3)K-subunit expression may have had on our analysis of p110^DASAA/DASAA mice, we also compared them with p110^+/- mice.

Analysis of the subcellular distribution of endogenous p110 in p101^-/- or p110^DASAA/DASAA-expressing versus wild-type neutrophils showed, however, that these genetic modifications or minor changes in expression had no effect on the ability of PI(3)K to associate with cell membranes in vivo (Fig. 1 and see Supplementary Information, Fig. S3).

A variety of cell-based assays were used to assess the overall integrity of G-protein dependent signalling in neutrophils isolated from p101^-/- and p110^DASAA/DASAA mice. fMLP and C5a-stimulated phosphorylation of p42/44 ERKs, p38 and JNK with the same dose-dependency and maximal effects as in wild-type cells, and activation of Ras proteins by fMLP, as assessed by Raf–RBD pulldown, was also unchanged in both genetically manipulated lines (see Supplementary Information, Fig. S4).

To establish the degree to which PI(3)K could be activated in response to G-protein-coupled agonists in neutrophils from p101^-/- and p110^DASAA/DASAA mice, the accumulation of its catalytic products, PtdIns(3,4,5)P₃ and PtdIns(3,4)P₂, was measured. Substantial reductions in the accumulation of these two lipids was found in response to both fMLP and C5a, in neutrophils from both the p101^-/- and p110^DASAA/DASAA mice (Fig. 2 and data not shown). The reductions in PtdIns(3,4,5)P₃ accumulation in neutrophils from p101^-/- mice were particularly marked at sub-maximal doses of chemoattractant (Fig. 2e). The decrease in PtdIns(3,4,5)P₃ accumulation in response to fMLP observed in neutrophils from p110^-/- mice and reported by our group¹⁸ and others^{15, 16, 17}, was only slightly greater than those we report here (approximately 90% and 75% reductions at 10 s and 1 min, respectively).

Figure 2. Accumulation of PtdIns(3,4,5)P3 is reduced in neutrophils isolated from p110DASAA/DASAA (p110D/D) and p101-/- mice. (a–e) 32P-Pi labelled neutrophils from either p101+/+ (solid line) versus p101-/- (short-dashed line) mice (a, c, e) and p110+/+ (solid line) versus p110DASAA/DASAA (long-dashed line) mice (b, d) were stimulated for various times with fMLP (10 M; a and b) or C5a (100 nM; c and d), or for 10 s with various doses of C5a (e). PtdIns(3,4,5)P3 (and other phosphoinositide) levels were quantified by deacylation, HPLC and scintillation counting. Data presented are one experiment respresentative of four in a; means s.e.m. (n = 3, except 60 s p110DASAA/DASAA, which was a range of n = 2) in b; means range (n = 2, except p101-/- 10 s and p101+/+ 2 min, which were n = 1) in c; means s.e.m. (n = 3, except p110DASAA/DASAA 1 min, which was a range of n = 2) in d; and means range (n = 2, except 0.6 nM C5a, which were s.e.m., n = 3) in e. Full Figure and legend (31K)

In keeping with the reduced accumulation of PtdIns(3,4,5)P₃ and PtdIns(3,4,)P₂, activation of PKB, as measured by the phosphorylation status of Ser 473 in response to fMLP and C5a, was reduced in neutrophils from both p101^-/- and p110^DASAA/DASAA but not p110^+/- mice (Fig. 3).

Figure 3. Phosphorylation of Ser 473-PKB in response to chemoattractants is reduced in neutrophils from p101-/- and p110DASAA/DASAA mice. (a) Immunoblots showing that phosphorylation of Ser 473-PKB, in response to various doses of C5a, is reduced in neutrophils derived from p101-/- mice. (b) Quantification of data from experiments similar to that shown in a, Data presented are means s.e.m. (n = 4, from four experiments). (c) Quantification of immunoblots comparing the phosphorylation of Ser 473-PKB, in response to 10 M fMLP, in neutrophils from p110+/+, p110+/- and p110DASAA/DASAA (p110D/D) mice. Data are means s.e.m. (n = 3, p110+/-; n = 4, p110+/+; n = 4, p110DASAA/DASAA). Full Figure and legend (23K)

The ability of neutrophils from p110^-/-, p101^-/-, p110^+/-, p110^DASAA/DASAA and wild-type control mice to respond chemotactically or chemokinetically to fMLP, C5a, Kc and MIP2 was assessed. The chemotactic and chemokinetic responses of neutrophils from p110^-/- (as reported previously^{15, 16, 17, 25}), p101^-/- and p110^DASAA/DASAA, but not p110^+/- mice, were significantly and similarly reduced for all four inflammatory mediators (Fig. 4). Furthermore, in a standard in vivo model (thioglycollate-induced aseptic peritonitis) neutrophils from both p101^-/- and p110^DASAA/DASAA mice, similarly to those from p110^-/- mice^{15, 16, 17}, exhibited substantially reduced recruitment into the peritoneum (Fig. 4).

Figure 4. The migration of neutrophils from mice lacking p110 or p101 or expressing p110DASAA/DASAA is substantially reduced. (a) Cells (1 106) from mice of the indicated genotype were loaded into the upper chamber of a transwell insert. In gradient mode, the chemoattractants were only placed in the lower chamber (MIP-2, 1 nM; Kc, 3 nM; fMLP, 1 M; C5a, 2 nM; all sub-maximal concentrations in this assay). us, unstimulated. Migration was assessed by counting the neutrophils that had appeared in the lower chamber after 1 h. Data are presented as percentage of wild-type cells that had migrated in gradient mode towards MIP-2 (called 100%). Means s.e.m. (n values are as indicated). (b) The migration of neutrophils in the presence of uniform concentrations of chemoattractant was measured, as described in a. Results are presented as percentage of migration to MIP-2 measured in gradient mode shown in a. (c) The migration of neutrophils into an aseptically inflamed peritoneum depends on p101 and the RBD of p110. Thioglycollate was injected into the peritoneum of various mice (genotypes indicated). After 4.5 h the number of neutrophils in the peritoneum of both the injected and un-injected (ui) control mice was quantified. The data presented are means s.e.m. (n values are as indicated). The numbers of neutrophils in the un-injected controls of the various genotypes were not significantly different and hence were pooled for presentation. Full Figure and legend (59K)

The ability of neutrophils from p101^-/-, p110^+/- and p110^DASAA/DASAA mice, to produce ROS in response to fMLP or C5a was assayed and compared with neutrophils from p110^-/- mice and wild-type controls. As anticipated, there was a substantial and equivalent reduction in ROS generation by both primed and unprimed neutrophils from p110^DASAA/DASAA and p110^-/- mice, compared with controls. Surprisingly, there was no significant difference in ROS formation between neutrophils from p101^-/- and p101^+/+ mice, in response to a wide range of sub-maximal (in the context of our observation that PtdIns(3,4,5)P₃ production is most reduced in p101^-/- neutrophils at lower agonist concentrations) and maximal doses of fMLP and C5a, and in both primed and unprimed neutrophils (Fig. 5 and data not shown).

Figure 5.  Production of ROS in response to chemoattractants is reduced in neutrophils from p110DASAA/DASAA but not p101-/- mice. (a) Production of ROS by unprimed mouse neutrophils is dependent on the RBD of p110 but not p101. Formation of ROS in response to fMLP, C5a and phorbol myristate acetate (PMA; 1 M, 100 nM and 300 nM final concentrations; which were all sub-maximal) was quantified by chemiluminescense assay. Data are presented as percentages of the response of neutrophils isolated from wild-type mice s.e.m. (n values are as indicated). (b) Production of ROS in response to fMLP by isolated, primed (TNF-) murine neutrophils (genotypes as indicated; the single, open-squared symbol shows parallel data obtained with neutrophils from p110-/- mice), data are presented in relative response units as means s.e.m. (n = 8, p110+/+; n = 6, p110D/D; n = 3, p110+/-;). Similar experiments with TNF- primed, p101-/- neutrophils showed no defect in ROS accumulation in response to neither fMLP nor C5a (data not shown). (c) Production of ROS in response to sub-maximal stimulation with C5a in unprimed neutrophils from p101+/+ (solid lines) and p101-/- mice. Data are presented as percentage of maximum response of that genotype s.e.m. (n = 4, from one experiment typical of four). No p101-dependent defect in ROS accumulation is revealed at a range of sub-maximal concentrations of C5a. (d) Data as in c, except the cells were stimulated with sub-maximal concentrations of fMLP. Full Figure and legend (59K)

These results demonstrate an essential role for p101 in signalling through PI(3)K to PtdIns(3,4,5)P₃ and PtdIns(3,4)P₂, activation of PKB and migration both in transwell assays and in vivo, indicating that the roles of p101 in mouse neutrophils overlap with those of p110. The extent of the roles for p101 specifically, and PI(3)K adaptors generally, in these PI(3)K-dependent responses are probably slightly underestimated through limited redundancy between p84 and p101 and the 50% increase in p84 expression in the absence of p101.

The above results and argument suggest the scale of the role of Gs in activating PI(3)K will be approximated by the loss of function resulting from p101 deficiency. Therefore, we conclude that our results also demonstrate that, as expected, Gs have substantial roles in the regulation of PI(3)K activity in mouse neutrophils.

Currently, the identity of the signalling pathway regulating the RBD of PI(3)K in these contexts is not clear. An extensive screen of the ability of Ras-related GTPases to regulate class I PI(3)Ks, in both transfected mammalian cells and in vitro, has found that a variety of GTP-bound Ras proteins (for example, N-, K-, R- and H-Ras) could activate PI(3)K, but Rap and other closely related proteins (Rap 1 and 2, Rheb or Rin) could not^{8, 12}. Therefore, it is very likely that a Ras family GTPase regulates this domain in vivo. A variety of G-protein-activating ligands can stimulate rapid and substantial increases in the amount of total GTP-Ras in both mouse and human neutrophils²⁸ (for example, 20–40-fold increase in 10 s; see Supplementary Information, Fig. S4). The species of Ras protein involved has not been characterised, but neutrophils probably contain K- and N-, but not H-, Ras. Similarly, the identity of the Ras-GEF (guanine nucleotide exchange factor), and how it is activated in these contexts, is uncertain, although a variety of results with various pharmacological inhibitors indicate no role for Src-type tyrosine kinases or phospholipase C activity²⁹.

Several further important conclusions can be drawn from a comparison of the phenotypes resulting from loss of p101 and knock-in of a Ras-insensitive form of PI(3)K. First, we can show that in a signalling cascade engaged by a single receptor, signals through both G–p101 and Ras–p110 are essential for normal activation. Second, a quantitative analysis of the scale of loss of PtdIns(3,4,5)P₃ accumulation in neutrophils from p101^-/- and p110^DASAA/DASAA mice, mapped against our (and others) work previously defining the reduction in PtdIns(3,4,5)P₃, suggests that Ras and G signals combine synergistically during physiological engagement of this pathway. This property is clearly translated into assays of migration in vivo and is entirely consistent with previous work indicating Ras and G can synergistically activate PI(3)K 'in transfecto'⁷. Third, different regulators of PI(3)Ks (here, Gs and Ras) can simultaneously but differentially regulate distinct PI(3)K effectors.Very similar extents of loss and kinetics of PtdIns(3,4,5)P₃ accumulation were observed in response to fMLP in neutrophils from p101^-/- and p110^DASAA/DASAA mice, making it difficult to explain the differential regulation of ROS formation on the basis of temporal specificity or intensity thresholds for responses. Furthermore, previous use of PI(3)K catalytic-site inhibitors would suggest that the reduction in PtdIns(3,4,5)P₃ production observed in p101^-/- neutrophils should be associated with significant reductions in ROS formation⁷. The likely explanation is that the Ras and p101 signals are spatially distinct and that this translates into differences in the sites of activation of PI(3)K and accumulation of PtdIns(3,4,5)P₃, and ultimately to qualitative differences in activation of effectors. This suggests that the class I PI(3)K cassette can use a combination of multiple PI(3)K isoforms and has the ability to decode different regulators independently providing flexibility and specificity to its signalling.

Generation of p110^DASAA/DASAA and p101^-/- mice.

Several clones encoding p110 genomic sequence were isolated from RPCI mouse PAC library 21 (gift from P. de Jong, UK HGMP Resource Centre, Hinxton, UK) by Southern screens. An 11.4 kb fragment, bracketed by XmnI sites, encompassing exons 1–4 was isolated from clone RP21-531 A23 and inserted into pGEM-T-easy to form the basis of a p110 gene-targeting vector. A smaller fragment containing exon 2 was subcloned into pBS; standard site directed mutagenesis was used to alter five amino acids in exon 2 (T232D, K251A, K254S, K255A and K256A) and create an additional NheI site. The added NheI site was used to track the presence of the mutated sequence. This modified segment was sequenced and then re-introduced into the targeting vector. A loxP-flanked cassette containing a neomycin gene was inserted in the upstream intron (see Supplementary Information, Fig. S2).

The final targeting vector was digested by NotI to remove excess pGEM-T easy vector sequence and used to transfect E14 129sv ES cells (Babraham Institute Gene Targeting Facility, Babraham Institute, Babraham, UK). Southern blots of 122 clones were screened for homologous recombination using a ³²P-oligonucleotide probe, diagnostic for correct insertion at the 3' end of the vector. Positive clones were re-screened with a probe diagnostic for correct insertion at the 5' end of the targeting vector. Two clones were sequence validated and used for blastocyst injection. Male chimeras from one of these clones were bred with female C57/bl6 animals to generate p110^{DASAA/wild type} mice on a mixed 129sv x C57/bl6 background. Progeny were crossed with an X-linked Cre-deletor strain (on a mixed 129sv x C57/bl6 background³⁰). Deletion of the neomycin-resistance cassette was confirmed by appropriate Southern blot analysis. Genotyping of the mice was routinely performed by PCR amplification of an approximate 850 base-pair region through exon 2 and subsequent diagnosis by susceptibility of the PCR product to cleavage by NheI to generate 350 and 500 base pair fragments from p110^DASAA alleles.

Two overlapping mouse genomic clones containing approximately 30 kb of p101 genomic sequence were isolated from a mouse 129sv library in FIXII (gift from N. Allen, Babraham Institute). A 4.7 kb KpnI fragment encompassing exons 1–3 was cloned into pBluescript and exon 1 was mutated by standard site-directed mutagenesis to include a BamHI restriction site at codon 11. This allowed in-frame insertion of a -galactosidase–neomycin resistance cassette from pSA-gal (gift from N. Allen) at this site (see Supplementary Information, Fig. S1). This construct formed the basis of the amino-terminal arm of the targeting vector, which was assembled in pBluescript together with a 7.5 kb carboxy-terminal XhoI fragment, resulting in the deletion of most of exon 1 and all of exons 2 and 3. The final targeting vector was linearised using a unique SalI site at the end of the C-terminal arm and transfected into E14 129sv stem cells. Probes, indicative of appropriate insertion at either the 3' or 5' ends of the targeting vector, were used to screen Southern blots of approximately 500 neomycin-resistant clones. Two positive clones were expanded and used in blastocyst-injections to derive two lines of p101^-/- mice on mixed 129sv C57/bl6 backgrounds. Mice were genotyped by Southern analysis. The results in this manuscript are derived with mice of one strain; key results (Figs 2, 3) were reproduced with the second strain.

p110 mice were obtained from M. Wymann (Basel, Switzerland) and maintained on a mixed 129sv C57/bl6 background. All mice were kept in isolators under specific pathogen-free conditions (Home Office Project licence PPL 80/1875).

The presence and quantity of p101, p84, p110, p110, p110 and p110 were determined by immunoblotting. The anti-p101 antibody was raised in sheep against GST-tagged p101 (amino acids 1–1031; porcine) and affinity purified using 6His-tagged p101 (amino acids 560–1031; porcine) covalently crosslinked to HiTrap Affinity (Pharmacia/LKB, Uppsala, Sweden). The affinity-purified antibody was used in immunoblots in TBS (pH 8.0; 0.1% Tween 20 and 4% skimmed milk powder) at 1 g ml^-1, detected with a rabbit anti-sheep HRP-antibody (Santa Cruz Biotechnology, Santa Cruz, CA) and quantified using ECL-plus reagent (Amersham, Amersham, UK). Rabbit, anti-p84 and anti-p110 antibodies were used as previously described³. The anti-p110 antibody detects recombinant p110 and p110^DASAA with equivalent sensitivity in immunoblots (data not shown). Anti-p110 and antibodies were from Santa Cruz (sc 7174 and sc 602, respectively) and anti-p110 antibodies (monoclonal L125) were a gift from M.Turner (Cambridge, UK). Anti-GAPDH and p67^phox monoclonal and anti-G polyclonal antibodies were from Biogenesis (Oxford, UK) and Transduction laboratories (Franklin Lakes, NJ) and S. Mumby (Dallas, TX).

To measure protein kinase activation, neutrophils (below) were stimulated (5 10⁶ ml^-1 in HBSS, at 37 °C for 180 s, unless otherwise noted), rapidly diluted with cold PBS, sedimented by centrifugation, aspirated and solubilised into ice-cold lysis buffer (as recommended), centrifuged (12,000g for 10 min, 4 °C) and the supernatants mixed with 4 SDS–PAGE sample buffer.

Mouse neutrophils were fractionated as follows: after stimulation with fMLP or its vehicle cells, cells were resuspended in ice-cold buffer (20 mM HEPES–NaOH at pH 7.2, 4 °C, 0.2 M sucrose, 0.13 M NaCl, 5 mM EGTA, 1 mM MgCl₂, 20 g ml^-1 each of leupeptin, pepstatin, antipain and aprotinin and 0.1 mM PMSF at 1.5 10⁷ ml^-1) and probe sonicated to yield a homogenate. Nuclei were sedimented by centrifugation in a microfuge (5,000g for 10 min at 4 °C) and the supernatant was ultracentrifuged (100,000g for 30 min at 4 °C) to yield a cytosolic fraction and light membrane pellet. The membranes were washed, resuspended (50 l) and aliquots of the homogenate, cytosolic (both 3.5 10⁵ cell equivalents) and membrane fractions (3.5 10⁶ cell equivalents) were immunoblotted with antibodies against p110, G (membrane marker) and GAPDH (cytosolic marker).

Bone marrow cells were collected into HBSS (with Ca²⁺and Mg²⁺; Sigma, Poole, UK) containing 0.25% fatty acid free, low endotoxin BSA and 14 mM HEPES–NaOH at pH 7.2 (37 °C), at 5 10⁶ ml^-1. These cell suspensions (200 l) were added to the top of a transwell filter (polycarbonate, 3 m pore; Millipore, Millerica, MA) inserted into a 24-well plate (Ultra-low attachment; Costar, Corning, Acton, MA) containing 300 l of the above buffer with chemoattractants or vehicle. The plates were transferred to an incubator (37 °C, atmospheric CO₂, humidified). After 1 h, the filter inserts were removed, the medium in the lower chamber was removed and retained, and 300 l of HBSS (Ca²⁺ and Mg²⁺ free), containing 5 mM EDTA was added. After 30 mins on ice, the buffer in each of the wells was removed and pooled. A further 300 l of HBSS (as above) was rinsed through the wells and added to the pools which were made up to 1 ml with HBSS. The cells recovered from each of the wells were analysed by FACS — counting neutrophils through a gate defined by Gr-1-positive cells in the total bone marrow. Total bone marrow preparation was stained for Gr-1 (after incubation in Fc-block) with 0.125 mg ml^-1 PE-conjugated Ly-6G (Gr-1) and Ly-6C (RB6-8CS; Becton Dickenson, Franklin Lakes, NJ) before being washed and analysed. This analysis was used to define a gate through which the concentration of neutrophils in unstained samples of the bone marrow cell suspension, and hence the number of neutrophils that went into each upper well, could be estimated and also to quantify the number of neutrophils that had migrated into the lower chambers.

Production of ROS was quantified as previously described¹⁸ using luminol-dependent chemiluminescence. Neutrophils in HBSS were primed with TNF- (human) for 25 mins (5 10⁶ ml^-1). Light emmision was recorded in a Berthold MicroLumat Plus luminometer (data output is in relative response units s-¹). Chemoattractants were added through the injection port. Data were integrated over the first 3 min of stimulation to derive the total response units in that time.

	Top

1. Vanhaesebroeck, B. et al. Synthesis and function of 3-phosphorylated inositol lipids. Annu. Rev. Biochem. 70, 535–602 (2001). | Article | PubMed | ISI | ChemPort |
2. Cantley, L. C. The phosphoinositide 3-kinase pathway. Science 296, 1655–1657 (2002). | Article | PubMed | ISI | ChemPort |
3. Suire, S. et al. p84, a new G-activated regulatory subunit of the type IB phosphoinositide 3-kinase p110. Curr. Biol. 15, 566–570 (2005). | Article | PubMed | ChemPort |
4. Voigt, P., Dorner, M. B. & Schaefer, M. Characterization of P87^Pikap, a novel regulatory subunit of phosphoinositide 3-kinase that is highly expressed in heart and interacts with PDE3B. J. Biol. Chem. 231, 9977–9986 (2006). | Article |
5. Stephens, L. R. et al. The G sensitivity of a PI3K is dependent upon a tightly associated adaptor, p101. Cell 89, 105–114 (1997). | Article | PubMed | ISI | ChemPort |
6. Rodriguez-Viciana, P. et al. Phosphatidylinositol-3-OH kinase as a direct target of Ras. Nature 370, 527–532 (1994). | Article | PubMed | ISI | ChemPort |
7. Pacold, M. E. et al. Crystal structure and functional analysis of Ras binding to its effector phosphoinositide 3-kinase . Cell 103, 931–943 (2000). | Article | PubMed | ISI | ChemPort |
8. Rodriguez-Viciana, P., Sabatier, C. & McCormick, F. Signaling specificity by Ras family GTPases is determined by the full spectrum of effectors they regulate. Mol. Cell Biol. 24, 4943–4954 (2004). | Article | PubMed | ISI | ChemPort |
9. Krugmann, S., Cooper, M. A., Williams, D. H., Hawkins, P. T. & Stephens, L. R. Mechanism of the regulation of type IB phosphoinositide 3OH-kinase by G-protein subunits. Biochem. J. 362, 725–731 (2002). | Article | PubMed | ChemPort |
10. Krugmann, S., Eguinoa, A., McGregor, A. H., Hawkins, P. T. & Stephens, L. R. Structural analysis of a novel isoform of phosphoinositide 3OH-kinase. Biochem. Soc. Trans. 25, S604 (1997). | PubMed | ChemPort |
11. Brock, C. et al. Roles of G in membrane recruitment and activation of p110/p101 phosphoinositide 3-kinase . J. Cell Biol. 160, 89–99 (2003). | Article | PubMed | ISI | ChemPort |
12. Suire, S., Hawkins, P. & Stephens, L. Activation of phosphoinositide 3-kinase by Ras. Curr. Biol. 12, 1068–1075 (2002). | Article | PubMed | ISI | ChemPort |
13. Simon, S. I. & Green, C. E. Molecular mechanics and dynamics of leukocyte recruitment during inflammation. Annu. Rev. Biomed. Eng. 7, 151–185 (2005). | Article | PubMed | ChemPort |
14. Sheppard, F. R. et al. Structural organization of the neutrophil NADPH oxidase: phosphorylation and translocation during priming and activation. J. Leukoc. Biol. 78, 1025–1042 (2005). | Article | PubMed | ChemPort |
15. Sasaki, T. et al. Function of PI3K in thymocyte development, T cell activation, and neutrophil migration. Science 287, 1040–1046 (2000). | Article | PubMed | ISI | ChemPort |
16. Hirsch, E. et al. Central role for G protein-coupled phosphoinositide 3-kinase in inflammation. Science 287, 1049–1053 (2000). | Article | PubMed | ISI | ChemPort |
17. Li, Z. et al. Roles of PLC-2 and -3 and PI3K in chemoattractant-mediated signal transduction. Science 287, 1046–1049 (2000). | Article | PubMed | ISI | ChemPort |
18. Condliffe, A. M. et al. Sequential activation of class IB and class IA PI3K is important for the primed respiratory burst of human but not murine neutrophils. Blood 106, 1432–1440 (2005). | Article | PubMed | ISI | ChemPort |
19. Crackower, M. A. et al. Regulation of myocardial contractility and cell size by distinct PI3K-PTEN signaling pathways. Cell 110, 737–749 (2002). | Article | PubMed | ISI | ChemPort |
20. Hirsch, E. et al. Resistance to thromboembolism in PI3K-deficient mice. FASEB J. 15, 2019–2021 (2001). | PubMed | ChemPort |
21. Puri, K. D. et al. The role of endothelial PI3K activity in neutrophil trafficking. Blood 106, 150–157 (2005). | Article | PubMed | ChemPort |
22. Rickert, P., Weiner, O. D., Wang, F., Bourne, H. R. & Servant, G. Leukocytes navigate by compass: roles of PI3K and its lipid products. Trends Cell Biol. 10, 466–473 (2000). | Article | PubMed | ISI | ChemPort |
23. Wymann, M. P. & Marone, R. Phosphoinositide 3-kinase in disease: timing, location, and scaffolding. Curr. Opin. Cell Biol. 17, 141–149 (2005). | Article | PubMed | ISI | ChemPort |
24. Weiner, O. D. Regulation of cell polarity during eukaryotic chemotaxis: the chemotactic compass. Curr. Opin. Cell Biol. 14, 196–202 (2002). | Article | PubMed | ISI | ChemPort |
25. Thomas, M. J. et al. Airway inflammation: chemokine-induced neutrophilia and the class I phosphoinositide 3-kinases. Eur. J. Immunol. 35, 1283–1291 (2005). | Article | PubMed | ChemPort |
26. Sasaki, A. T., Chun, C., Takeda, K. & Firtel, R. A. Localized Ras signaling at the leading edge regulates PI3K, cell polarity, and directional cell movement. J. Cell Biol. 167, 505–518 (2004). | Article | PubMed | ChemPort |
27. Funamoto, S., Meili, R., Lee, S., Parry, L. & Firtel, R. A. Spatial and temporal regulation of 3-phosphoinositides by PI 3-kinase and PTEN mediates chemotaxis. Cell 109, 611–623 (2002). | Article | PubMed | ISI | ChemPort |
28. Coffer, P. J. et al. Comparison of the roles of mitogen-activated protein kinase kinase and phosphatidylinositol 3-kinase signal transduction in neutrophil effector function. Biochem. J. 329, 121–130 (1998). | PubMed | ISI | ChemPort |
29. Zheng, L., Eckerdal, J., Dimitrijevic, I. & Andersson, T. Chemotactic peptide-induced activation of Ras in human neutrophils is associated with inhibition of p120-GAP activity. J. Biol. Chem. 272, 23448–23454 (1997). | Article | PubMed | ChemPort |
30. Schwenk, F., Baron, U. & Rajewsky, K. A cre-transgenic mouse strain for the ubiquitous deletion of loxP-flanked gene segments including deletion in germ cells. Nucleic Acids Res. 23, 5080–5081 (1995). | PubMed | ISI | ChemPort |
31. de Rooij, J. & Bos, J. L. Minimal Ras-binding domain of Raf1 can be used as an activation-specific probe for Ras. Oncogene 14, 623–625 (1997). | Article | PubMed | ChemPort |
32. Welch, H. C. et al. P-Rex1 regulates neutrophil function. Curr. Biol. 15, 1867–1873 (2005). | Article | PubMed | ChemPort |

	Top

PI(3)K mice were provided by M. Wymann. Thanks to M. Skynner for help in the initial isolation of p101 genomic clones and thanks also to T. Green for use of a blood-cell counter. Thanks to P. Arnaud for design of PCR primers and A. Segonds-Pichon for help with statistics. C.E. was a Beit fellow. A.M.C. was supported by a Wellcome Trust Intermediate fellowship. Different parts of this work were funded by grants from the Biotechnology and Biological Sciences Research council (BBSRC) and Cancer Research UK (CRUK).

Competing interests statement:  The authors declare that they have no competing financial interests.