Probing phosphoinositide functions in signaling and membrane trafficking.

Abstract

The inositol phospholipids (PIs) comprise a family of eight species with different combinations of phosphate groups arranged around the inositol ring. PIs are among the most versatile signaling molecules known, with key roles in receptor-mediated signal transduction, actin remodeling and membrane trafficking. Recent studies have identified effector proteins and specific lipid-binding domains through which PIs signal. These lipid-binding domains can be used as probes to further our understanding of the spatial and temporal control of individual PI species. New layers of complexity revealed by the use of such probes include the occurrence of PIs at intracellular locations, the identification of phosphatidylinositol signaling hotspots and the presence of non-membrane pools of PIs in cell nuclei.

Full text

Page 1

Probing phosphoinositide functions in
signaling and membrane trafficking
C. Peter Downes1, Alex Gray1and John M. Lucocq2
1Division of Cell Signalling, School of Life Sciences, MSI/WTB Complex, University of Dundee, Dundee, UK DD1 5EH
2Division of Cell Biology and Immunology, School of Life Sciences, MSI/WTB Complex, University of Dundee, Dundee, UK DD1 5EH
The inositol phospholipids (PIs) comprise a family of
eight species with different combinations of phosphate
groups arranged around the inositol ring. PIs are among
the most versatile signaling molecules known, with key
roles in receptor-mediated signal transduction, actin
remodeling and membrane trafficking. Recent studies
have identified effector proteins and specific lipid-
binding domains through which PIs signal. These lipid-
binding domains can be used as probes to further our
understanding of the spatial and temporal control of
individual PI species. New layers of complexity revealed
by the use of such probes include the occurrence of PIs
at intracellular locations, the identification of phospha-
tidylinositol signaling hotspots and the presence of non-
membrane pools of PIs in cell nuclei.
Introduction
The inositol-containing glycerophospholipids, collectively
known as phosphoinositides (inositol phospholipids; PIs),
are among the most versatile ofregulatory molecules,with
strikingly diverse roles in cell signaling and vesicle-based
transport mechanisms. This versatility arises from the
chemistry of the myo-inositol moiety, which is attached to
diacylglycerol (DAG) via a di-ester phosphate at the D-1
position, leaving five free hydroxyls, three of which are
phosphorylated in different combinations by lipid kinases.
PIs that have been identified in eukaryotic cells, their
biosynthesis and metabolic interconversions are illus-
trated in Figure 1. The enzymes involved in these
pathways have been reviewed extensively elsewhere
[1–6]. With the recent discovery of phosphatidylinositol
(3,4,5)-trisphosphate [PtdIns(3,4,5)P3] in fission yeast [7],
it now appears that most of the PIs in Figure 1 are
conserved from yeast to man. Here, we focus on PI-binding
proteins and, in particular, on the use of specific lipid-
binding domains as probes for the quantitative temporal
and spatial analysis of PIs in cells (Box 1).
The DAG moiety accounts for the fact that PIs are
predominantly if not exclusively associated with cell
membranes. By contrast, it is the exposed headgroups of
PIs that bind to effector proteins and through which their
signaling functions are realized. The structural diversity of
these headgroups, the existence of effector domains with
highaffinityandselectivityforparticularPIspeciesandthe
non-uniform distribution of PIs among subcellular mem-
branes are crucial for the fidelity of PI-dependent signaling
mechanisms.IfthenumberofPIspeciesseemscomplexand
confusing, the range of effector domains is even more
dramatically diverse. They include pleckstrin homology
(PH), phagocyte oxidase (PX), epsin N-terminal homology
(ENTH) and Fab1p, YOTB, Vac1p and EEA1 (FYVE)
domains, as well as an assortment of proteins that bind to
PIs with varying degrees of specificity through small
patches of basic amino acids [4,8]. These domains serve
primarily to target their host proteins to specific cellular
locations and, in some cases, perhaps, might more directly
regulateproteinfunction.Theiruseasprobesdependsupon
ahighdegreeofspecificityforthetargetlipid,affinitiesthat
allow detection of the target at the levels that occur
naturally in cell membranes and an understanding of
secondary interactions that might restrict the distribution
oftheprobetoaparticularcompartment,limitingitsusefor
detection of the target at other sites. With the possible
exceptionofPtdIns5PandPtdInsitself,suitableprobesnow
existforallofthePIsinFigure1,implyingthateachofthese
lipids has one or more signaling roles. What follows
considers how we can use and avoid abusing these probes
to study the functions of individual PI species.
PtdIns(4,5)P2: multiple functions and locations
The lipid phosphatidylinositol
[PtdIns(4,5)P2] undoubtedly represents a focal point in
PI-dependent signaling in both metabolic and functional
terms. It is synthesized mainly by a diverse family of
PtdIns4P 5-kinases and serves as the substrate for two
powerful receptor-regulated signal-generating enzymes.
PI-phospholipase C (PI-PLC) cleaves PtdIns(4,5)P2,
simultaneously producing two second messengers, DAG
and inositol (1,4,5)-trisphosphate [15]. Type I PI 3-kinases
[1], on the other hand, convert PtdIns(4,5)P2 to
PtdIns(3,4,5)P3, of which more later. If this wasn’t enough,
PtdIns(4,5)P2itself can bind to an increasing number of
effector domains, through which it appears to be a crucial
regulator of: actin polymerization and anchorage to
plasma membranes and vesicular structures; assembly/
disassembly of vesicular coats; invagination and scission
of endocytic vesicles; regulated secretion; the turnover of
focal adhesion complexes and several types of plasma
membrane KCchannel [2,3,16–21].
Many of the proteins that bind to PtdIns(4,5)P2
physiologically appear to do so through clusters of basic
(4,5)-bisphosphate
Corresponding author: Downes, C.P. (c.p.downes@dundee.ac.uk).
Available online 8 April 2005
Review
TRENDS in Cell BiologyVol.15 No.5 May 2005
www.sciencedirect.com0962-8924/$ - see front matter Q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.tcb.2005.03.008

Page 2

residues against a relatively hydrophobic background but
have little else in common and might not share a common
structural fold. By contrast, three distinct domains have
been shown to bind to this lipid, with varying degrees of
specificity. These include proteins containing ENTH and
AP180 N-terminal homology (ANTH) and PH domains
[3,22]. Of these modules, the PH domain of PLCd1has
been extensively characterized and utilized as a probe to
study PtdIns(4,5)P2dynamics in cells.
PH domains are protein modules of w120 amino acids,
and there are w250 representatives within the human
genome. While many PH domains bind to PIs, they often
do so with relatively low affinity and/or specificity. PH
PLCd1is an exception in that it binds to PtdIns(4,5)P2and
its headgroup, Ins(1,4,5)P3, sufficiently selectively to be
used as a cellular probe for these compounds [23–25]. The
function of these interactions is to target PLCd1to plasma
membranes enriched in the substrate lipid, with
Ins(1,4,5)P3competing for this binding and hence acting
as a feedback inhibitor of PLCd1when Ins(1,4,5)P3levels
are raised. Recently, it has been shown that, while the
interaction with PtdIns(4,5)P2 appears essential for
plasma membrane binding of PH PLCd1, this might also
require additional interactions involving the C-terminal a
helix and a short loop between the b6and b7sheets [26].
Green fluorescent protein (GFP)-tagged PH-PLCd1
labelspredominantlytheplasmamembranewhenanalysed
by confocal microscopy. The stimulation of PLC-coupled
receptors is usually accompanied by a pronounced trans-
location of this membrane fluorescence to the cytosol.
Several studies have concluded that the agonist-depen-
dent translocation of PH-PLCd1 is due to increased
cytosolic Ins(1,4,5)P3 and not the decline of plasma
membrane PtdIns(4,5)P2[24,25]. However, under circum-
stances where the intensity of stimulation leads to a
significant fall in the concentration of PtdIns(4,5)P2, it
seems likely that this must also contribute to the cytosolic
redistribution of the probe.
A potential limitation of studies with live cells is that
overexpressed PI-binding proteins could sequester
PtdIns(4,5)P2 from its endogenous binding partners.
This limitation can be overcome in experiments using
the tagged PH domain on fixed cells. In this approach,
cells are either permeabilized or sectioned to allow access
of the immunological detection system. Using the on-
section method adapted for electron microscopy, it has
been possible to study the distribution of PtdIns(4,5)P2at
high resolution (see Box 2). These studies revealed a major
pool of PtdIns(4,5)P2in plasma membranes, and smaller
pools in intracellular membranes, including Golgi, endo-
somes and endoplasmic reticulum [27]. PtdIns(4,5)P2was
also found in the nuclear matrix, where it was associated
not with membranes but with dense regions of hetero-
chromatin, in agreement with studies using antibodies
against PtdIns(4,5)P2[28] (see Box 3 for a discussion of
nuclear inositol lipids). An important advantage of the
TRENDS in Cell Biology
OH
HO
OO
2
OH
OH
P
O
O
O–
O
H
O
O
O
H
1
3
45
6
DAG
Phosphatidylinositol
Myo-inositol
headgroup
PI5P
ING2-PHD
PI(3,4)P2
TAPP1-PH
PI3P
FYVE
PX
PI(3,5)P2
Svp1p
DAG
C1
PI(4,5)P2
PLCδ1-PH
ENTH
PI4P
FAPP1-PH
PI(3,4,5)P3
GRP1-PH
Btk-PH
ARNO-PH
IP3
Figure 1. Biosynthesis and metabolism of phosphoinositide signals. Phosphatidylinositol (PtdIns) is the precursor of all the phosphoinositides (PIs) shown. Most of the
naturally occurring phosphatidylinositol has a saturated fatty acid esterified at the 1-position of glycerol, with a polyunsaturated fatty acid at the 2-position (typically stearic
acid and arachidonic acid, respectively, as illustrated). It is generally assumed that all PtdIns species share this composition, which is likely to have a bearing on whether PIs
are components of lipid rafts. The PIs shown differ with respect to the number and distribution of mono-ester phosphate groups around the inositol ring and are
interconverted by kinase and phosphatase enzymes that comprise 14 distinct enzyme families. Protein domains that have been shown to interact with a particular PtdIns
species are indicated in red, while specific PI-binding proteins and the relevant domains they harbour are indicated in black.
Review
TRENDS in Cell BiologyVol.15 No.5 May 2005 260
www.sciencedirect.com

Page 3

Box 1. Methods for determining PtdIns/protein binding affinities and specificities
Ligand binding specificity and affinity are key determinants of the suitability of any probe for analysing the cellular content, spatial distribution and
dynamics of an individual phosphoinositide species (Table I).
Table I
Lipid/protein overlay (‘Fat blot’)
A range of lipids is spotted onto nitrocellulose, soaked in buffer containing the protein of interest, washed and absorbed protein detected by
enhanced chemiluminescence [9].
Advantages
Quick, easy technique familiar for non-specialist laboratories
Many lipid species analysed in parallel
Disadvantages
Purely qualitative
Prone to both false positives and negative
Must be used in conjunction with a quantitative approach
Pull down assays
Several variations on these, e.g. sucrose-loaded unilamellar lipid vesicles can be rapidly centrifuged for analysis of bound proteins by western
blot. Alternatively, tagged proteins can be pulled down on agarose beads and absorbed lipid vesicles detected by using radiolabeled carrier
lipids [10–12]
Advantages
Quantitative
Disadvantages
Slow separation of bound and free ligand means that low-affinity
interactions will be underestimated
Vesicle composition can be varied to resemble cellular membranes
Surface plasmon resonance (BIACore)
Several BIACore chips can be employed with protein attached when the adsorption of large unilamellar vesicles containing the lipid of interest
can be detected. Alternatively, with lipids adsorbed to a hydrophobic chip, the binding of protein in the mobile phase can be detected [9,13].
Advantages
Potentially quantitative
Disadvantages
The nature of adsorption of lipids to hydrophobic chips remains ill-
characterized
Instrumentation is expensive and binding kinetics often prove
complex
Lipid composition can be varied
Binding toequilibrium makes it suitable for detecting low-as well as
high-affinity interactions
Ligand displacement
Adaptation of methods introduced for high-throughput screening of PI 3-kinases. Uses a sensor complex in which the binding of a pleckstrin-
homology (PH) domain to its cognate lipid immobilized on allophycocyanin generates a time-resolved fluorescence resonance energy transfer
(TR-FRET) signal. Exogenous lipids with high affinity for the PH domain dissociate the sensor complex and reduce the TR-FRET signal [13,14].
Advantages
Very quantitative
Large number of assays run in parallel
Suitable for both soluble (e.g. inositol phosphates) and lipid/
vesicular ligands
Disadvantages
Optimal signals require the use of low levels of detergent
Box 2. Ultrastructural localization of phosphoinositides on cryosections
A traditional approach to localize lipids is to express GFP-tagged
constructs in living cells, but these may have dominant interference
effects on labeling. In on-section labeling, the cells are fixed and
sectioned without prior transfection. The sections are then decorated
using a tagged version of the lipid-binding module, which is then
immunologically detected for microscopy. In the EM version illustrated
here (Figure I), localization is performed on thin cryosections, using
electron-dense marker colloidal gold as the readout. The affinity
labeling is performed close to zero degrees C to prevent extraction of
phosphoinositides (PIs) from the sectioned membranes [27]. When
combined with unbiased stereological quantitation of the gold signal
[30], this approach can be used to localize PIs in cellular compartments
[27,29]. Controls for specificity include: (i) competing-out binding
using soluble phosphoinositide headgroups [27], (ii) physiological
modulation of endogenous PIs before fixation [27,29] and (iii) point
mutations that reduce or prevent PI binding [29]. On-section labeling
can also be applied to fluorescence microscopy for quantitative
detection of PIs as a biomarker in clinical samples [31].
TRENDS in Cell Biology
Localization
Slicing
Living cell
PhosphoinositideGST-tagged PH domainAnti-GST antibody
Fixation
EM marker
Figure I.
Review
TRENDS in Cell BiologyVol.15 No.5 May 2005261
www.sciencedirect.com

Page 4

on-section method is that it opens-up compartments and
allows more even access of the labeling reagents to
cellular structures. Furthermore, the quantitative
evaluation of the particulate gold readout at the EM
level allows dispersed, but weak-labeled, compartments to
be detected (27,29,30).
A crucial question that can now be addressed is
whether PtdIns(4,5)P2content is polarized in particular
regions of membranes or uniformly distributed. The idea
that lipid second messengers might be concentrated in
membrane microdomains such as lipid rafts has arisen
using mainly indirect approaches. In fact, as PIs commonly
possess polyunsaturated fatty acids at the 2-position
(see Figure 1), they seem unlikely raft components, a
conclusion supported by ultrastructural analyses [32,33].
In a recent study, Huang et al. [34] applied ultrafast
acquisition and super-resolution deconvolution micro-
scopy to adipocytes expressing GFP-tagged PH-PLCd1.
This approach revealed large-scale patches of plasma
membrane that were enriched in PtdIns(4,5)P2beneath
which were present clathrin-containing vesicular struc-
tures and dense concentrations of polymerized actin.
These results provide evidence for the involvement of
PtdIns(4,5)P2as a major determinant coordinately regu-
lating actin polymerization and endocytic vesicle traffick-
ing. Importantly, the PtdIns(4,5)P2-rich patches observed
in these studies were probably too large to correspond to
rafts (with lateral dimensions of greater than several
micrometers [32]). Ultrastuctural studies by on-section
labeling (Box 2) also showed enrichment of PtdIns(4,5)P2
in actin-rich lamellipodia-like structures [27]. It is
important to consider how such apparently stable
PtdIns(4,5)P2-enriched loci might be maintained. As the
rate of lateral diffusion of lipids within the bilayer is likely
to be very fast compared with the turnover numbers of
lipid-metabolizing enzymes. We suggest that barriers to
diffusion would be needed to reduce the rate at which
PtdIns(4,5)P2could escape from such ‘hotspots’ in addition
to the specific targeting of synthetic enzymes to and/or
degradative enzymes away from these zones. Such a
diffusion barrier could be provided by the high concen-
tration of PtdIns(4,5)P2-binding proteins, associated with
the regulation of actin turnover, surrounding such sites.
These binding proteins would act like a fine sieve so that,
at any one time, most of the ligand would be free (and
hence detectable by the probe), but its progress through
the sieve would be retarded. The combination of high-
resolution microscopic techniques and the availability of
selective binding probes can now begin to address whether
PtdIns(4,5)P2-enriched microdomains occur commonly
and the mechanisms that support the polarized distri-
bution of signaling lipids in cells.
PtdIns(3,4,5)P3and PtdIns(3,4)P2: lipids mediating PI
3-kinase-dependent signaling pathways
PtdIns(3,4,5)P3 is synthesized by Type I PI 3-kinases
(PI3Ks), a reaction that is reversed by the tumor-
suppressor lipid phosphatase, PTEN (for ‘phosphatase
and tensin homolog deleted on chromosome ten’; see
Figure 2). In quiescent cells, PtdIns(3,4,5)P3is present at
low levels [w0.1% of the level of its precursor,
PtdIns(4,5)P2], but, upon stimulation of tyrosine kinase
and some G-protein-coupled receptors, its concentration
can increase by factors ranging from 2- to 100-fold. The
sources of cellular phosphatidylinositol 3,4-bisphosphate
[PtdIns(3,4)P2] are not completely established, but it
probably arises mainly through the dephosphorylation of
PtdIns(3,4,5)P3by 5-phosphatases such as SHIP (for ‘SH2
domain containing inositol phosphatase’) and SHIP2 [35]
and it occurs in amounts similar to its precursor. Together,
PtdIns(3,4,5)P3 and PtdIns(3,4)P2 appear to account
for the many and varied cellular functions of Type I PI
3-kinases, which include cell growth and proliferation,
resistance to apoptosis, regulation of cytoskeleton
dynamics, membrane trafficking and many of the meta-
bolic responses to insulin [1,36]. These pervasive roles
explain the intense interest in PI 3-kinase signaling
pathways as a potentially rich source of targets for
therapeutic intervention in cancer, type II diabetes,
cardiac failure, inflammatory diseases and problems
involving auto/hyper-immune responses.
PtdIns(3,4,5)P3 and PtdIns(3,4)P2 regulate a wide
range of effector proteins primarily by targeting them to
specific membrane locations via high affinity binding to
one or more PH domains [9]. Because of the relatively low
concentrations of PtdIns(3,4,5)P3and/or PtdIns(3,4)P2in
cells, PH domains that bind to these lipids physiologically
must display a high degree of specificity, especially by
comparison with the much more abundant PtdIns(4,5)P2.
Several PH domains that display the requisite selectivity
Box 3. Phosphoinositides in the nucleus: phospholipids, but
not as we know them
Phosphoinositides (PIs) were first found to be components of highly
purified nuclei more than 20 years ago [48]. It came as a big surprise,
however, when it was discovered that PIs were present in nuclei that
had been stripped of membrane structures [49]. Recently, probes for
PtdIns(4,5)P2(see Box 2 and Ref. [48]) identified this lipid in the
nuclear matrix within electron-dense structures, nuclear lamina and
nuclear speckles. This raises important questions about the
physicochemical form of these lipids in the absence of obvious
membrane structures. As it can be detected with probes that interact
primarily with the headgroup, it seems most likely that nuclear
PtdIns(4,5)P2exists either in micellar or proteolipid complexes that
envelop the hydrophobic fatty acids, leaving the headgroups
exposed.
Functions of nuclear PIs
Nuclear PtdIns(4,5)P2, as elsewhere in the cell, is likely to have
multiple functions. It appears to be a substrate for PI-phospholipase
C enzymes (PLCs; especially PLCb1) and hence might regulate
DAG-sensitive protein kinase Cs. Although its role in intranuclear
Ca2Csignaling is not clear, Ins(1,4,5)P3, as it does in yeast, might
instead serve as a substrate in the synthesis of several inositol
polyphosphates implicated in the control of mRNA export and
chromatin remodeling [50–52]. Nuclear PIs are also likely to be
substrates for PI 3-kinases, but so far PtdIns3P is the only 3-PtdIns
that has been detected directly [53]. The presence of PtdIns(3,4,5)P3,
however, is inferred by the occurrence of Type I PI 3-kinases and a
novel regulatory GTPase (PIKE, PtdIns 3-kinase enhancer), one form
of which is localized to nuclei [54]. As for nuclear PI-binding proteins,
chromatin-associated ING2, through its plant homeodomain (PHD)
finger, might be the first known functional binding partner for
PtdIns5P, through which it might regulate p53-dependent responses
to DNA damage [55].
Review
TRENDS in Cell BiologyVol.15 No.5 May 2005 262
www.sciencedirect.com

Page 5

have now been characterized and provide extremely
valuable tools for studying the temporal and spatial
distribution of PI 3-kinase signaling in cells. A problem
with many of the PH domains that are in use for this
purpose, for example the PH domain of protein kinase
B/Akt, is that they bind to both PtdIns(3,4,5)P3 and
PtdIns(3,4)P2[37]. The PH domains of Bruton’s tyrosine
kinase and several members of the general receptor for
phosphoinositides 1 (Grp1) family (e.g. Grp1 itself, Arf
nucleotide binding site opener, ARNO, and cytohesin 1),
however, are highly selective for PtdIns(3,4,5)P3. Recent
crystallographic studies of Grp1 and ARNO PH domains
have established that the insertion of a single glycine
residue in the b1/b2loops of these domains is sufficient to
convert them from being monospecific to being dual
binders of both PtdIns(3,4,5)P3and PtdIns(3,4)P2[38].
When expressed in quiescent cells, GFP-tagged
PtdIns(3,4,5)P3-specific PH domains are mainly cytosolic,
but they translocate primarily to plasma membranes in
response to stimuli that activate PI 3-kinases [3,39,40].
This type of approach has been used to show the highly
polarized accumulation of PtdIns(3,4,5)P3at the leading
edge of Dictyostelium during chemotaxis towards a source
of cyclic AMP [41]. A more sophisticated approach was
used by Sato and colleagues [42] who developed a
targetable probe in which Grp1–PH was fused between
cyan and yellow fluorescent proteins. PtdIns(3,4,5)P3
binding caused a conformational rearrangement that
enhanced the fluorescence resonance energy transfer
(FRET) signal between CFP and YFP. This probe was
targeted to plasma membranes by inserting a CAAX motif
and to endomembranes by mutation of Cys181 of the
CAAX motif to a serine. In response to platelet-derived
growth factor (PDGF), these probes detected rapid
production of PtdIns(3,4,5)P3 in plasma membranes,
followed by its accumulation in endomembranes within
w1 min. Furtherexperiments
PtdIns(3,4,5)P3was produced in situ in endomembranes
and required clathrin-mediated endocytosis of the PDGF
receptor. Hence, there now exist probes that allow a high
degree of temporal resolution of PI 3-kinase signaling in
distinct subcellular compartments.
In addition to PH domains, monoclonal antibodies are
available that interact with PtdIns(3,4,5)P3, but which
appear to detect principally cytosolic components in
labeling experiments with fixed cells [43]. As the anti-
bodies were raised against protein conjugates of the lipid
headgroup, it might be that they are detecting either
protein-bound lipid or indeed the headgroup itself,
Ins(1,3,4,5)P4.
Currently, the only well-characterized PtdIns(3,4)P2-
specific binding protein is the C-terminal PH domain of
TAPP1 (for: ‘tandem PH-domain-containing protein 1)
[9,44]. TAPP1 might function as part of a feedback loop to
downregulate tyrosine kinase signaling. It binds to the
first of five PDZ domains of protein tyrosine phosphatase
establishedthat
TRENDS in Cell Biology
PI(4,5)P2
PLC
PI(3,4,5)P3
IP3
DAG
BTKs
PDK1/PKB
RHO
Ca2+
GEFs
PI(4,5)P2
PI 3-kinase
PTEN
ABPs
PI(4,5)P2 Hot spots
RAC
Actin
Integrins
Cytokines
Growth factors
Insulin
PH, ENTH, ANTH
domains
Basic patches
PH domains
PKC
C1 domains
Figure 2. Lipid-binding domains mediating phosphoinositide functions at the plasma membrane. PtdIns(4,5)P2 [PI(4,5)P2] is the substrate for stimulus-dependent
phospholipase C enzymes (PLCs) and type I phosphoinositide 3-kinases (PI 3-kinases) that generate three second messengers: soluble Ins(1,4,5)P3(IP3) and membrane-
associated diacylglycerol (DAG), and PtdIns(3,4,5)P3[PI(3,4,5)P3]. PtdIns(4,5)P2[PI(4,5)P2) also has a plethora of signaling roles. These include the regulation of actin
polymerization and membrane anchorage, and endocytic vesicle trafficking. These functions are mediated through diverse protein domains (ENTH, ANTH and PH) and
positively charged patches on structurally unrelated proteins. The diversity and relative abundance of actin-binding proteins capable of binding to PtdIns(4,5)P2might create
an effective barrier to diffusion of this lipid away from F-actin-rich structures, possibly accounting for the occurrence of PtdIns(4,5)P2hotspots in such regions of the plasma
membrane. PtdIns(3,4,5)P3binds to and regulates a wide range of proteins mediated through their PH domains, which account for their binding specificity towards lipids.
Exemplified by Btk family members, serine/threonine kinases, such as PDK1 and PKB and GDP–GTP exchange factors (GEFs) for Rho-family G-proteins, which lead to
regulation of actin dynamics. Agonist-stimulated endocytosis of growth-factor receptors leads to the internalization of signaling complexes, which retain activated
PI 3-kinases and synthesize PtdIns(3,4,5)P3that accumulates in components of the endomembrane system.
Review
TRENDS in Cell Biology Vol.15 No.5 May 2005 263
www.sciencedirect.com

Page 6

L1 (PTPL1) through its C-terminal tail and maintains the
cytosolic pool of this phosphatase in quiescent cells. Upon
stimulation of PI 3-kinase signaling, the increase in
PtdIns(3,4)P2causes translocation of the TAPP1–PTPL1
complex to plasma membranes, where presumably it
functions to reverse the actions of activated tyrosine
kinases [45]. GFP-tagged TAPP1–PH is therefore a
suitable probe for the detection of PtdIns(3,4)P2within
cells. Recently, a GST-tagged TAPP1 PH has been utilized
to map the ultrastructural distribution of its cognate
lipid ligand by immunoelectron microscopy, as described
in Box 2. These studies revealed the presence of
PtdIns(3,4)P2 in the endoplasmic reticulum and multi-
vesicular endosomes, as well as in the plasma membrane
[29]. Only the plasma membrane and endosomal pools
were sensitive to the expression of PTEN, implying
segregated metabolism and functions for these distinct
pools of this signaling lipid.
The high degree of selectivity of certain PH domains for
PtdIns(3,4,5)P3 and PtdIns(3,4)P2 has prompted their
development as versatile tools in a range of assay
procedures. These assays rely on the generation of a
sensor complex involving the PH domain and a synthetic
biotinylated version of the target lipid. The presence of
exogenous lipid, produced for example in a PI 3-kinase
assay, or present as a component of a tissue extract, can be
quantitatively detected as it competes with the biotinyl-
ated lipid within the sensor complex [46]. These methods
are proving extremely valuable in high-throughput
screening of PI 3-kinases and as nonradioactive assays
of signaling lipids in the tissues of transgenic animals
carrying mutations of PI 3-kinase signaling components
[47]. With the range of specific lipid-binding domains
currently available, these methods should be adaptable
specifically to the detection of most of the lipids depicted in
Figure 1.
PtdIns4P, PtdIns3P and PtdIns(3,5)P2: phosphoinositides
associated predominantly with intracellular membranes
Research into the roles of PIs on intracellular membranes
has lagged well behind that of PIs found at the cell surface.
PtdIns(4,5)P2, PtdIns(3,4)P2 and PtdIns(3,4,5)P3 are
rarely found intracellularly, but a large body of evidence
now implicates monophosphorylated PIs in the functions
of intracellular organelles, in particular, membrane traffic
[3,56]. The two principal lipids of interest are phospha-
tidylinositol 3-phosphate (PtdIns3P) and phosphatidyl-
inositol 4-phosphate (PtdIns4P), of which the latter is the
more abundant in cells and present at levels similar to
those of PtdIns(4,5)P2. These lipids target cytoplasmic
TRENDS in Cell Biology
PI
PI3P
PI4P
PI(4,5)P2
PI
PI3P
PI(3,5)P2
Golgi
Endosomes
AP1
Spectrin (PH)
Epsin R
(ENTH)
EEA1
(FYVE)
Retromer
(PX)
Ent3p
(ENTH)
Hrs
(FYVE)
FAPP1/OSBP
(PH)
Key:
ClathrinARF1Cargo
Figure 3. Subcellular distribution of phosphoinositides and localization of trafficking machinery in the Golgi complex and endosomal system. PtdIns4P (PI4P) and PtdIns3P
(PI3P) are characteristic phosphoinositides of the Golgi and endosomal system, respectively. At these sites, accurate targeting of trafficking machinery is dependent on both
lipid and protein–protein interactions (for example with small GTPases, cargo or coat proteins). In each case, the protein/protein complex is followed by the relevant lipid-
binding domain in parenthesis. At the Golgi, PI4P is generated from phosphatidylinositol (PtdIns, PI) by PtdIns 4-kinases and acts as both a substrate for PtdIns(4,5)P2
[PI(4,5)P2] and a localizor of the budding machinery. For trans-Golgi localization, PtdIns4P-binding proteins (FAPP1, OSBP and AP1) engage with at least two components,
PtdIns4P and the small GTPase ARF, and, in the case of AP1, also with cargo and clathrin. AP1 might act sequentially with GGA proteins in trafficking between the TGN and
endosome, but the precise details of the trafficking steps are unclear. FAPP1 regulates Golgi export en route to the plasma membrane. The Golgi-localized lipid phosphatases
driving hydrolysis of PtdIns(4,5)P2and PtdIns4P are not shown. In endosomes, PtdIns3P is synthesized through the action of the Vps34p and PtdIns(3,5)P2on late endosomes
by mammalian PIKfyve/yeast Fab1p. PtdIns3P-binding domains targeting trafficking machinery to endosomes include FYVE, PX and ENTH (epsin N-terminal homology
domain) domains. Initial weak membrane association by binding to PtdIns3P is strengthened by membrane insertion (FYVE, PX domains, ENTH) or by association with
coat/tether proteins [retromer (PX); EEA1/rab5 (FYVE)]. The PtdIns3P-binding FYVE domain of Hrs sorts ubiquitinated cargo into clathrin-coated regions before uptake into
multivesicular bodies, while the ENTH-domain-containing Ent3p is a clathrin binding effector of PtdIns(3,5)P2needed for a similar internalization step [together with another
PtdIns(3,5)P2-specific protein Ent5, Ent3 and the ubiquitin-binding protein Vps27p are required for sorting into the multivesicular body]. Another ENTH-domain-containing
protein, Epsin R, mediates retrograde endosome–TGN traffic. It binds to PtdIns4P through its ENTH domain and its hydrophobic a-helix might be used to insert between lipid
headgroups and induce membrane curvature. Epsin R can also interact with AP1 (PtdIns4P binder), clathrin and the SNARE protein Vti1b.
Review
TRENDS in Cell Biology Vol.15 No.5 May 2005264
www.sciencedirect.com

Page 7

protein complexes to specific membranes, with PtdIns4P
concentrated in the Golgi and PtdIns3P in the endosome
system (Figure 3). Four principle PI-binding domains
have been implicated in localizing functional proteins to
these compartments, including PH domains (mainly Golgi
localized), ENTH domains, found in Golgi and endosomes,
and FYVE and PX domains involved in targeting to
endosomes [8]. A major feature is their involvement in
membrane trafficking and in the recruitment and turn-
over of cytoplasmic coats required for tubule and vesicle
production [3,56]. Interestingly, a lipid interaction on the
Golgi or endosomes is often not the sole determinant of
membrane association, and other factors, usually pro-
teins, are required [3]. This has two advantages for vesicle
coat assembly/disassembly. First multiple low-affinity
interactions facilitate dynamic interactions over the very
small time scalesneeded forrapid coated-vesicle assembly,
budding and coat disassembly reactions and, second, they
allow lipid signals to be proof-read by a second or even
third protein signal to target specifically to a small focus
on a budding membrane [57]. Recruitment to the trans-
Golgi network (TGN) is a good example.
The clathrin adaptor protein AP1 binds to PtdIns4P
(although not through one of the recognized phospholipid-
binding domains), but association with the membrane
requires the small GTPase ARF [58]. AP1 also has to bind
to cargo destined for transport in addition to its role in
recruiting the clathrin coat. It is likely that AP1
dissociates from the membrane after budding, but pre-
sumably, as each one of these binding sites is removed, the
binding becomes weaker and the complex falls apart. A
second adaptor for clathrin-coated structures in the TGN
region has been identified called Epsin R, and this
contains an ENTH domain that is again selective for
PtdIns4P [59]. Multiple interactions include an expected
ARF1 interaction [60] and association with a SNARE
protein cargo [61], and Epsin R appears to cooperate with
AP1 to form clathrin-coated structures. Epsin R might
function in retrograde trafficking from endosomes to the
TGN [62]. Another multiple interaction mechanism has
been observed for two proteins, FAPP1 and oxysterol-
binding protein (OSBP), known to bind to PtdIns4P
through their PH domains. FAPP1 was identified as a
PtdIns4P-binding protein that controls traffic from the
TGN to the cell surface and also binds to ARF1, which
itself can promote recruitment of a PtdIns 4-kinase that
could generate the FAPP1 ligand, PtdIns4P [63]. The
interaction with ARF is mediated through the PH domain
(63), and several other reports suggest that PH domains
themselves can interact with proteins as well as with
phosphoinositides [64–66]. Coat and cargo interactions
have not so far been described for FAPP1. The functions
and binding partners of PtdIns4P and PtdIns3P are
illustrated in Figure 3.
So, are any of these lipid-binding domains useful as
probes forPtdIns4P?When expressed in cells,GFP-tagged
AP1 and FAPP1 show localization that is dependent on
PtdIns4P, but the protein–protein interactions confound
the unbiased detection of PtdIns4P. One solution might be
to use on-section labeling because protein–protein inter-
actions have a good chance of being disrupted by extensive
aldehyde crosslinking, although careful controls would be
required. One such control is to employ a mutant lipid-
binding domain that no longer binds to the target
phosphoinositide [27] or to make mutants that abrogate
specifically the protein–protein interaction (Box 2).
Targeting to the endosomal system, rich in PtdIns3P
generated from PtdIns by type III PI 3-kinase (yeast
Vps34p and its mammalian homolog), has a decidedly
different flavor. Here, two additional lipid-binding
domains play a role in targeting both coat and tethering
proteins. FYVE domains bind to PtdIns3P specifically and
target several proteins to endosomal locations. Interest-
ingly, the FYVE domain has rather limited affinity,
explained by a relative lack of high-affinity hydrogen
bonds being made with the PtdIns headgroup [8]. Affinity
can be increased by nonpolar side-chains of a membrane
insertion loop inside the membrane and/or by dimeriza-
tion of FYVE domains. The latter is reflected in the need
for using dimers of FYVE domains to label PtdIns3P in
cells [53,67], although dimerization might not be required
for FYVE domain proteins with higher affinity for
membranes [68]. Additional interactions with coat, cargo
and accessory proteins again ensure the precise localiz-
ation, as in EEA-1 with Rab5 and Hrs with clathrin and
ubiquitinated cargo destined for uptake into multivesicu-
lar endosomes.
Some PX domains also interact with PtdIns3P and are
involved in protein sorting and vesicle coat assembly but,
again, despite high selectivity, have a rather limited
affinity for the lipid alone [3,8]. The affinity can be
increased by insertion of a membrane association loop or
by multiple interactions within a cytoplasmic coat such as
the retromer complex, which mediates endosome-to-Golgi
retrieval [69,70] and increases affinity by multiple PX
interactions combined with cargo binding [71].
So, can the endosomal binding domains be used as
probes for PI localization? This application has been
pioneered by Gillooly et al., who have used an FYVE
domain dimer to localize PtdIns3P both in cells and on EM
sections [53,67]. An FYVE domain construct that allows
dimerization to be induced by exogenous addition of a cell-
penetrant crosslinker has also been used to good effect in
the detection of endosomal PtdIns3P [67]. Combining
these approaches with the on-section labeling methods
illustrated in Box 2 might be necessary to complete the
picture in terms of the overall PtdIns3P content and
distribution within cells.
Phosphatidylinositol 3,5-bisphosphate [PtdIns(3,5)P2]
is the most recently identified phosphoinositide and is
synthesized from PtdIns3P by Fab1p in S. cerevisiae and
by an FYVE-domain-containing PtdIns kinase (PIKfyve)
in mammalian cells [72,73]. In the latter case, the FYVE
domain appears to target the enzyme to PtdIns3P-rich
membranes (see above). The phenotype of fab1-deficient
yeast is complex, implicating PtdIns(3,5)P2in vacuole-to-
lysosome membrane trafficking, vacuole acidification,
packaging of proteins in multivesicular bodies (MVBs)
and growth at elevated temperatures, suggesting the prob-
able involvement of functionally distinct PtdIns(3,5)P2-
binding proteins. This has now been confirmed with the
identification of several effector proteins that appear to
Review
TRENDS in Cell BiologyVol.15 No.5 May 2005265
www.sciencedirect.com

Page 8

bind to PtdIns(3,5)P2 with the anticipated degree of
specificity. These comprise Ent3p, Ent5p and Vps24p,
which link Vps7p-mediated ubiquitin sorting into MVBs
[74,75], Svp1p, which is required for membrane recycling
from the vacuole [76], and sorting nexins that mediate
transport from endosomes to the trans-Golgi [77]. As the
yeast proteins do not account fully for the fab1D
phenotype, it seems that additional PtdIns(3,5)P2effector
proteins await discovery.
Surprisingly, these effector proteins interact with
PtdIns(3,5)P2through a diverse set of PI-binding modules.
For example, Vps24p and its mammalian homolog possess
polybasic N-termini that do not resemble other known
PI-binding domains, whereas Ent3p and Ent5p bind to
PtdIns(3,5)P2 through their ENTH domains [74,78].
Centaurin b2, by contrast, has a PH domain that
selectively binds to PtdIns(3,5)P2[9], while sorting nexins,
such as SNX-1, utilize a PX domain for this purpose [77].
The most recently described PtdIns(3,5)P2effector pro-
tein, Svp1p, required for membrane recycling from the
vacuole in yeast, is the first protein identified that binds to
phosphoinositides through basic patches on a b-propeller
structure [76].
Whether any of these modules represent suitable
probes for the unbiased detection of PtdIns(3,5)P2,
however, is open to question. In the case of SNX-1, it is
not clear whether PtdIns(3,5)P2 or PtdIns3P is the
physiological ligand. This protein also possesses a BAR
(Bin/Amphiphysin/Rvs) domain that mediates dimeriza-
tion and the detection of highly curved membrane surfaces
[77]. So its cognate lipid ligands might only be detected
with high affinity in the correct biophysical context. The
full-length mammalian homolog of Vps24p binds to
PtdIns(3,4)P2as well as to PtdIns(3,5)P2and, although
it localizes to endosomes under certain conditions, this
might involve additional hydrophobic interactions and/or
oligomerization. Expression of the N-terminal half of this
protein, which is responsible for phosphoinositide binding,
exerts dominant-negative effects that clearly distort the
labeled compartment, telling us much about the function
of Vps24p but not about the global distribution of
PtdIns(3,5)P2. It might be that suitable probes could be
developed as for PtdIns3P by enhancing the affinity of
probes for PtdIns(3,5)P2by engineering suitable dimers or
chimeras and combining live-cell and on-section labeling
approaches.
Concluding remarks
The unique versatility of phosphoinositides as intracellu-
lar signals arises from four distinct aspects:
† The combinatorial actions of lipid kinases and
phosphatases generates the range of molecules depicted
in Figure 1, where many of the components are both
precursors and products of signaling enzymes.
† A temporal aspect in which individual signals can be
produced transiently and metabolized rapidly.
† A spatial aspect in which phosphoinositides are
produced and maintained in distinct cellular locations
and organelles.
† Finally, each PI can interact with a range of effector
www.sciencedirect.com
molecules, each of which must possess a PI-binding
domain with an appropriate degree of specificity.
Future research will make increasing use of PI-binding
domains and probes to study the spatial aspects of lipid
signaling in increasing detail. ‘Pure’ PI-specific binding
domains might, however, prove to be quite rare, while
‘coincidence detection’, in which the fidelity of protein
targeting requires multiple, relatively low-affinity, inter-
actions, appears especially important for many of the
intracellular functions of phosphoinositides. The latter
means that individual phosphoinositides can participate
in the assembly of different signaling complexes in
different locations. The usefulness of PI-binding domains
as probes depends on understanding their lipid binding
characteristics and any secondary interactions that might
contribute to localization within cells. Even where
secondary interactions are important, useful probes can
be engineered, for example by making dimers of the lipid-
binding moiety, and dominant interference effects can be
avoided by on-section labeling techniques. With these
points in mind, you might never again need to reach for
the organic solvents or have a grasp of thin-layer
chromatography to bring inositol phospholipids into your
experimental repertoire.
Acknowledgements
We thank Yvonne Lyndsay and Terry Smith for help with Figure 1.
Research in the C.P.D. laboratory is supported by the Medical Research
Council. J.M.L. is supported by the Wellcome Trust and the Lowe Trust.
References
1 Cantley, L.C. (2002) The phosphoinositide 3-kinase pathway. Science
296, 1655–1657
2 Giudici, M.L. et al. (2004) Phosphatidylinositol phosphate kinases.
J. Endocrinol. Invest. 27, 137–142
3 Roth, M.G. (2004) Phosphoinositides in constitutive membrane traffic.
Physiol. Rev. 84, 699–730
4 Wishart, M.J. and Dixon, J.E. (2002) PTEN and myotubularin
phosphatases: from 3-phosphoinositide dephosphorylation to disease.
Trends Cell Biol. 12, 579–585
5 Backers, K. et al. (2003) The termination of PI3K signaling by SHIP1
and SHIP2 inositol 5-phosphatases. Adv. Enzyme Regul. 43, 15–28
6 Simonsen, A. et al. (2001) The role of phosphoinositides in membrane
transport. Curr. Opin. Cell Biol. 13, 485–492
7 Mitra, P. et al. (2004) A novel phosphatidylinositol (3,4,5)P3pathway
in fission yeast. J. Cell Biol. 166, 205–211
8 Lemmon, M.A. (2003) Phosphoinositide recognition domains. Traffic 4,
201–213
9 Dowler, S. et al. (2000) Identification of pleckstrin-homology-domain-
containing proteins with novel phosphoinositide-binding specificities.
Biochem. J. 351, 19–31
10 James, S.R. et al. (1996) Specific binding of the Akt-1 protein kinase to
phosphatidylinositol 3,4,5-trisphosphate without subsequent acti-
vation. Biochem. J. 315, 709–713
11 Gaullier, J-M. et al. (1998) FYVE fingers bind PtdIns3P. Nature 394,
432–433
12 Levine, T.P. and Munro, S. (2002) Targeting of Golgi-specific
pleckstrin homology domains involves both PtdIns4P-kinase-depen-
dent and -independent components. Curr. Biol. 12, 695–704
13 Yu, J.N. and Lemmon, M.A. (2001) All PHOX homology domains from
Saccharomyces cerevisiae specifically recognize phosphatidylinositol
3-phosphate. J. Biol. Chem. 276, 44179–44184
14 Komander, D. et al. (2004) Structural insights into the regulation of
PDK1 by phosphoinositides and inositol phosphates. EMBO J. 23,
3918–3928
15 Rhee, S.G. et al. (2001) Regulation of phosphoinositide-specific
phospholipase C. Annu. Rev. Biochem. 70, 281–312
Review
TRENDS in Cell BiologyVol.15 No.5 May 2005266

Page 9

16 Cremona, O. and de-Camilli, P. (2001) Phosphoinositides in membrane
traffic at the synapse. J. Cell Sci. 114, 1041–1052
17 Czech, M.P. (2003) Dynamics of phosphoinositides in membrane
retrieval and insertion. Annu. Rev. Physiol. 65, 791–815
18 Takenawa, T. and Itoh, T. (2001) Phosphoinositides, key molecules for
regulation of actin cytoskeletal organization and membrane traffic
from the plasma membrane. Biochim. Biophys. Acta 1533, 190–206
19 Jonmey, P.A. and Lindberg, U. (2004) Cytoskeletal regulation: rich in
lipids. Nat. Rev. Mol. Cell Biol. 5, 658–666
20 Toker, A. (1998) The synthesis and cellular roles of phosphatidyl-
inositol 4,5-bisphosphate. Curr. Opin. Cell Biol. 10, 254–261
21 Ling, K. et al. (2002) Type Ig phosphatidylinositol phosphate kinase
targets and regulates focal adhesions. Nature 420, 89–93
22 Lemmon, M.A. and Ferguson, K.M. (2000) Signal-dependent mem-
brane targeting by pleckstrin homology (PH) domains. Biochem. J.
350, 1–18
23 Cifuentes, M.E. et al. (1994) D-myo-inositol 1,4,5-trisphosphate
inhibits binding of phospholipase C-delta 1 to bilayer membranes.
J. Biol. Chem. 269, 1945–1948
24 Hirose, K. et al. (1999) Spatiotemporal dynamics of inositol 1,4,5-
trisphosphate that underlies complex Ca2Cmobilization patterns.
Science 284, 1527–1530
25 Young, K.W. et al. (2004) NMDA-receptor regulation of muscarinic-
receptor-stimulated inositol 1,4,5-trisphosphate production and pro-
tein kinase C activation in single cerebellar granule neurons.
J. Neurochem. 89, 1537–1546
26 Varnai, P. et al. (2002) Inositol lipid binding and membrane
localization of isolated pleckstrin homology (PH) domains. Studies
on the PH domains of phospholipase C delta 1 and p130. J. Biol. Chem.
277, 27412–27422
27 Watt, S.A. et al. (2002) Subcellular localization of phosphatidylinositol
4,5-bisphosphate using the pleckstrin homology domain of phospho-
lipase C d1. Biochem. J. 363, 657–666
28 Schiavo, G. et al. (2004) Curr. Top. Microbiol. Immunol. 282, 177–206
29 Watt, S.A. et al. (2004) Detection of novel intracellular agonist-
responsive pools of phosphatidylinositol 3,4-bisphosphate using the
TAPP1 pleckstrin homology domain in immunoelectron microscopy.
Biochem. J. 377, 653–663
30 Lucocq,J.M.etal.(2004)Arapidmethodforassessingthedistributionof
gold labelling on thin sections. J. Histochem. Cytochem. 52, 991–1000
31 Evans, C. et al. (2004) Comparative proteomics of primitive haema-
topoietic cell populations reveals differences in expression of proteins
regulating motility. Blood 103, 3751–3759
32 van Rheenan, J. and Jalink, K. (2002) Agonist-induced PIP2
hydrolysis inhibits cortical actin dynamics: regulation at a global
but not at a micrometer scale. Mol. Biol. Cell 13, 3257–3267
33 London, E. (2002) Insights into lipid raft structure and formation
from experiments in model membranes. Curr. Opin. Struct. Biol.
12, 480–486
34 Huang, S. et al. (2004) Phosphatidylinositol-4,5-bisphosphate-rich
plasma membrane patches organize active zones of endocytosis and
ruffling in cultured adipocytes. Mol. Cell. Biol. 24, 9102–9123
35 Mitchell, C.A. et al. (2002) Inositol polyphosphate 5-phosphatases:
lipid phosphatases with flair. IUBMB Life 53, 25–36
36 Vanhaesebroeck, B. and Alessi, D.R. (2000) The PI3K-PDK1 connec-
tion: more than just a road to PKB. Biochem. J. 346, 561–576
37 Gray, A. et al. (1999) The pleckstrin homology domains of protein
kinase B and GRP1 (general receptor for phosphoinositides-1) are
sensitive and selective probes for the cellular detection of phospha-
tidylinositol 3,4-bisphosphate and/or phosphatidylinositol 3,4,5-
trisphosphate in vivo. Biochem. J. 344, 929–936
38 Cronin, T.C. et al. (2004) Structural determinants of phosphoinositide
selectivity in splice variants of Grp1 family PH domains. EMBO J. 23,
3711–3720
39 Oatey, P.B. et al. (1999) Confocal imaging of the subcellular
distribution of phosphatidylinositol 3,4,5-trisphosphate in insulin
and PDGF-stimulated 3T3-L1 adipocytes. Biochem. J. 344, 511–518
40 Walker, S.A. et al. (2004) GFP fusion proteins to study signaling in live
cells. Methods Mol. Biol. 273, 407–420
41 Postma, M. et al. (2004) Sensitisation of Dictyostelium chemotaxis by
phosphoinositide-3-kinase-mediated
patches. J. Cell Sci. 117, 2925–2935
self-organising signaling
42 Sato, M. et al. (2003) Production of PtdInsP3at endomembranes is
triggered by receptor endocytosis. Nat. Cell Biol. 5, 1016–1022
43 Niswender, K.D. et al. (2003) Immunocytochemical detection of
phosphatidylinositol 3-kinase activation by insulin and leptin.
J. Histochem. Cytochem. 51, 275–283
44 Thomas, C.C. et al. (2001) Crystal structure of the phosphatidyl-
inositol 3,4-bisphosphate-binding pleckstrin-homology domain-con-
taining protein 1 (TAPP1): molecular basis of lipid specificity.
Biochem. J. 358, 287–294
45 Kimber, W.A. et al. (2003) Interaction of the protein tyrosine
phosphatase PTPL1 with the PtdIns(3,4)P2adaptor protein TAPP1.
Biochem. J. 376, 525–535
46 Gray, A. et al. (2003) Nonradioactive methods for the assay of
phosphoinositide 3-kinases and phosphoinositide phosphatases and
selective detection of signaling lipids in cell and tissue extracts. Anal.
Biochem. 313, 234–245
47 Ali, K. et al. (2004) Essential role for the p110 delta phosphoinositide
3-kinase in the allergic response. Nature 431, 1007–1011
48 Martelli, A.M. et al. (2004) Nuclear inositides: facts and perspectives.
Pharmacol. Ther. 101, 47–64
49 Vann, L.R. et al. (1997) Metabolism and possible compartmental-
ization of inositol lipids in isolated rat-liver nuclei. Biochem. J.
327, 569–576
50 Steger, D.J. et al. (2003) Regulation of chromatin remodeling by
inositol polyphosphates. Science 299, 114–116
51 York, J.D. et al. (1999) A phospholipase C-dependent inositol poly-
phosphate kinase pathway required for efficient messenger RNA
export. Science 285, 96–100
52 Osborne, S.L. et al. (2001) Nuclear PtdIns(4,5)P2 assembles in a
mitotically regulated particle involved in pre-mRNA splicing. J. Cell
Sci. 114, 2501–2511
53 Gillooly, D.J. et al. (2000) Localisation of phosphatidylinositol
3-phosphate in yeast and mammalian cells. EMBO J. 19, 4577–4588
54 Ye, K. and Snyder, S.H. (2004) PIKE GTPase: a novel mediator of
phosphoinositide signaling. J. Cell Sci. 117, 155–161
55 Gozani, D. et al. (2003) The PHD finger of the chromatin-associated
protein ING2 functions as a nuclear phosphoinositide receptor. Cell
114, 99–111
56 De Matteis, M.A. and Godi, A. (2004) PI-loting membrane traffic. Nat.
Cell Biol. 6, 487–492
57 Owen, D.J. (2004) Linking endocytic cargo to clathrin: structural and
functional insights into coated vesicle formation. Biochem. Soc. Trans.
32, 1–14
58 Wang, Y.J. et al. (2003) Phosphatidylinositol 4 phosphate regulates
targeting of clathrin adaptor AP-1 complexes to the Golgi. Cell 114,
299–310
59 Mills, I.G. et al. (2003) EpsinR: an AP1/clathrin interacting protein
involved in vesicle trafficking. J. Cell Biol. 160, 213–222
60 Hirst, J. et al. (2003) EpsinR: an ENTH domain-containing protein
that interacts with AP-1. Mol. Biol. Cell 14, 625–641
61 Hirst, J. et al. (2004) EpsinR is an adaptor for the SNARE protein
Vti1b. Mol. Cell Biol. Sep 15 [Epub ahead of print].
62 Saint-Pol, A. et al. (2004) Clathrin adaptor epsinR is required for
retrograde sorting on early endosomes. Dev. Cell 6, 525–538
63 Godi, A. et al. (2004) FAPPs control Golgi-to-cell-surface membrane
traffic by binding to ARF and PtdIns(4). Nat. Cell Biol. 6, 393–404
64 Snyder, J.T. et al. (2003) The pleckstrin homology domain of
phospholipase C-beta2 is an effector site for Rac. J. Biol. Chem. 278,
21099–21104
65 Touhara, K. et al. (1994) Binding of G-protein beta gamma-subunits to
pleckstrin homology domains. J. Biol. Chem. 269, 10217–10220
66 Yao, L. et al. (1994) The pleckstrin homology domain of Bruton
tyrosine kinase interacts with protein kinase C. Proc. Natl. Acad. Sci.
U. S. A. 91, 9175–9179
67 Gillooly, D.J. et al. (2003) Phosphatidylinositol 3-phosphate is
found in microdomains of early endosomes. Histochem. Cell Biol.
120, 445–453
68 Hayakawa, A. et al. (2004) Structural basis for endosomal targeting by
FYVE domains. J. Biol. Chem. 279, 5958–5966
69 Arighi, C.N. et al. (2004) Role of the mammalian retromer in sorting of
the cation-independent mannose 6-phosphate receptor. J. Cell Biol.
165, 123–133
Review
TRENDS in Cell BiologyVol.15 No.5 May 2005 267
www.sciencedirect.com

Page 10

70 Seaman, M.N. (2004) Cargo-selective endosomal sorting for retrieval
to the Golgi requires retromer. J. Cell Biol. 165, 111–122
71 Burda, P. et al. (2002) Retromer function in endosome-to-Golgi
retrograde transport is regulated by the yeast Vps34 PtdIns 3-kinase.
J. Cell Sci. 115, 3889–3900
72 Dove, S.K. et al. (1997) Osmotic stress activates phosphatidylinositol-
3,5-bisphosphate synthesis. Nature 390, 187–192
73 Gray, J.D. et al. (1998) Fab1p is essential for PtdINs3P 5-kinase
activity and the maintenance of vacuolar size and membrane
homeostasis. J. Cell Biol. 143, 65–79
74 Friant, S. et al. (2003) Ent3P is a PtdIns(3,5)P2effector required for
protein sorting to the multivesicular body. Dev. Cell 5, 499–511
75 Eugster, A. et al. (2004) Ent5p is required with Ent3p and Vps27p for
ubiquitin-dependent protein sorting into the multivesicular body. Mol.
Biol. Cell 15, 3031–3041
76 Dove, S.K. et al. (2004) Svp1p defines a family of phosphatidylinositol
3,5-bisphosphate effectors. EMBO J. 23, 1922–1933
77 Carlton, J. et al. (2004) Sorting nexin-1 mediates tubular endosome-
to-TGN transport through coincidence sensing of high curvature
membranes and 3-phosphoinositides. Curr. Biol. 14, 1791–1800
78 Whitley, P. et al. (2003) Identification of mammalian Vps24p as an
effector of phosphatidylinositol 3,5-bisphosphate-dependent endo-
some compartmentation. J. Biol. Chem. 278, 38786–38795
Endeavour
the quarterly magazine for the history
and philosophy of science
You can access Endeavour online via
ScienceDirect, where you’ll find a
collection of beautifully illustrated
articles on the history of science, book
reviews and editorial comment.
Featuring
Sverre Petterssen and the Contentious (and Momentous) Weather Forecasts for D-Day, 6 June 1944 by J.R. Fleming
Food of Paradise: Tahitian breadfruit and the Autocritique of European Consumption by P. White and E.C. Spary
Two Approaches to Etiology: The Debate Over Smoking and Lung Cancer in the 1950s by M. Parascandola
Sicily, or sea of tranquility? Mapping and naming the moon by J. Vertesi
The Prehistory of the Periodic Table by D. Rouvray
Two portraits of Edmond Halley by P. Fara
and coming soon
Fighting the ‘microbe of sporting mania’: Australian science and Antarctic exploration in the early twentieth century
by P. Roberts
Learning from Education to Communicate Science as a Good Story by A. Negrete and C. Lartigue
The Traffic and Display of Body Parts in the Early-19th Century by S. Alberti and S. Chaplin
The Rise, Fall and Resurrection of Group Selection by M. Borrello
Pomet’s great ‘‘Compleat History of Drugs’’ by S. Sherman
Sherlock Holmes: scientific detective by L. Snyder
The Future of Electricity in 1892 by G.J.N. Gooday
The First Personal Computer by J. November
Baloonmania: news in the air by M.G. Kim
and much, much more . . .
Locate Endeavour on ScienceDirect (http://www.sciencedirect.com)
Review
TRENDS in Cell BiologyVol.15 No.5 May 2005268
www.sciencedirect.com