BY0074-0012.indd Page 129
12/18/06
4:33:07 PM elhi /Volumes/ju108/POIN002/symbosia_indd%0
12
Biochem. Soc. Symp. 74, 129–139
(Printed in Great Britain)
© 2007 The Biochemical Society
Our FABulous VACation: a
decade of phosphatidylinositol
3,5-bisphosphate
Stephen K. Dove1 and Zoë E. Johnson
School of Biosciences, University of Birmingham, Birmingham B15 2TT, U.K.
Abstract
PtdIns(3,5)P2 was discovered about a decade ago and much of the machinery
that makes, degrades and senses it has been uncovered. Despite this, we still
lack a complete understanding of how the pieces fit together but some patterns
are beginning to emerge. Molecular functions for PtdIns(3,5)P2 are also elusive,
but the identification of effectors offers a way into some of these processes.
An examination of the defects associated with loss of synthesis of PtdIns(3,5)P2
in lower and higher eukaryotes begins to suggest a unifying theme; this lipid
regulates membrane retrieval via retrograde trafficking from distal compartments
to organelles that are more proximal in the endocytic/lysosomal system. Another
unifying theme is stress signalling to organelles, possibly both to change their
morphology in response to external insults and to maintain the lumenal pH or
membrane potential of organelles. The next few years seem likely to uncover
details of the molecular mechanisms underlying the biology of this fascinating
lipid. This review also highlights some areas where further research is needed.
Introduction
Ten years of research on the phosphoinositide PtdIns(3,5)P2 have yielded
broad biochemical understanding of how it is made, degraded and perceived
and has offered insights into its cellular roles [1–3]. Most of these data have
come from studies in the yeast Saccharomyces cerevisiae, so the main focus of
this review will be a summary of recent research on PtdIns(3,5)P2 in yeast; the
reader is directed to two recent reviews for a more exhaustive treatment of the
functions of PtdIns(3,5)P2 in multicellular organisms [2,3]. We also highlight
studies in Caenorhabditis elegans and mammalian cells that were reported after
the publication of those reviews [4,5].
1
To whom correspondence should be addressed (email [email protected]).
129
BY0074-0012.indd Page 130 12/5/06 2:10:27 PM elhi /Users/elhi/Desktop/5 Dec_Bijender/indd
130
S.K. Dove and Z.E. Johnson
Figure 1 The Fab1p pathway. Despite a decade of research, the physical
and functional relationships between the various proteins in this pathway
remain ill-defined. It is known that Vac14p and Fig4p physically associate but the
mechanism by which Vac7p and Vac14p control Fab1p is still unknown. Some
effectors of PtdIns(3,5)P2 are indicated but their molecular functions remain to
be explained in detail.
Like all phosphoinositides, PtdIns(3,5)P2 is rapidly derived from the ubiquitous phospholipid PtdIns by sequential phosphorylation [1,6]. The phosphorylations occur in a fixed sequence (Figure 1) and are catalysed by two
distinct families of highly regulated phosphoinositide kinases. The first step in
PtdIns(3,5)P2 synthesis is 3-phosphorylation of PtdIns by Vps34p-type PtdIns
3-kinases to form PtdIns3P (see Figure 1) [1,6]. Vps34p was first identified as an
enzyme required for correct sorting of vacuolar (lysosomal) hydrolytic enzymes
at the TGN (trans-Golgi network), implying that PtdIns(3,5)P2 would have
roles in the endosomal system [7,8]. This expectation was fulfilled when Fab1p
was identified as the yeast PtdIns3P5-kinase; FAB1 mutants had already
been shown to have multiple defects in vacuole (lysosome) function [9–11].
Subsequent studies identified Vac7p and Vac14p, two additional proteins that
are required for full Fab1p activity [12–14]. Yeast knockouts of these genes
synthesize very little PtdIns(3,5)P2 and display many characteristic defects of
fab1 mutants.
The defects in mutants deficient in PtdIns(3,5)P2 synthesis include swelling
and hypertrophy of the vacuole [9], caused by a failure of membrane retrieval
from this compartment [15], a defect in the sorting of ubiquitinated proteins into
MVBs (multivesicular bodies) [16], sensitivity of the cells to multiple stresses
[9,11] and failure to acidify the vacuole [2,9]. These phenotypes are partially
separable, so they appear to be consequences of faults in independent functions
[14,16]. All depend on the lipid kinase function of Fab1p as they are displayed
by fab1 knockout mutants that express a kinase-dead Fab1p (fab1K) [3].
Phosphoinositides mediate their effects on cellular processes primarily
through binding to specific effector proteins, and this multiplicity of functions
suggests that cells may contain multiple pools of PtdIns(3,5)P2 that influence
multiple PtdIns(3,5)P2 effectors [3]. These phosphoinositide–protein interactions
usually serve to regulate the activity, localization or stability of the protein
effectors and translate into cellular responses via the altered properties of the
effector molecules. Proposed effectors for PtdIns(3,5)P2 include the PROPPIN
family of seven-bladed ß-propellers (Svp1p/Atg18p, Hsv1p/Atg21p and Hsv2p
© 2007 The Biochemical Society
BY0074-0012.indd Page 131 12/5/06 2:10:27 PM elhi /Users/elhi/Desktop/5 Dec_Bijender/indd
Molecular functions for PtdIns(3,5)P2
131
in yeast). Svp1p/Atg18p is implicated in regulating PtdIns(3,5)P2-dependent
retrieval of membrane and proteins from the vacuole to the late endosome
[15]. The PtdIns(3,5)P2-dependent functions of Hsv1p/Atg21p and Hsv2p
remain to be defined in detail. Proteins implicated in PtdIns(3,5)P2-dependent
MVB sorting include the Epsin-like proteins Ent3p and Ent5p [17] and the
CHMP family protein Vps24p [18]. No protein effectors of PtdIns(3,5)P2
have yet been described that are required for vacuole acidification [2,3]. We have
recently identified a protein, Svp3p, that appears to lie downstream of Fab1p
on a PtdIns(3,5)P2-dependent stress response pathway (S.K. Dove, unpublished
work).
Phosphoinositide signals are terminated by phosphoinositide phosphomonoesterases. PtdIns(3,5)P2 can be 5-dephosphorylated by SAC (suppressor of
actin) domain-containing phosphatases such as Fig4p [19,20], or 3-dephosphorylated by myotubularin-type phosphatases [21,22].
PtdIns(3,5)P2 in organisms other than yeasts.
PtdIns(3,5)P2 is present in all major groups of eukaryotes, including
plants, animals and fungi. It is therefore no surprise to find that much of the
cellular machinery for making, degrading and sensing PtdIns(3,5)P2 is likewise
conserved throughout the Eukarya [3]. Most eukaryotes seem to encode a
single FAB1-like PtdIns3P 5-kinase, and also VAC14 and FIG4 genes. There
are often multiple PROPPIN genes in a single organism, but whether the
Epsin-like proteins present in most eukaryotes correspond to Ent3p and Ent5p
awaits further investigation. VAC7 seems to be present only in fungi.
The limited number of studies in which functions for PtdIns(3,5)P2 have
been investigated in multicellular organisms have all confirmed that Fab1p-type
enzymes represent the principle route for PtdIns(3,5)P2 biosynthesis. They also
confirm that a major function of PtdIns(3,5)P2 is in control of lysosome or
late endosome size and/or function. Expression of a dominant negative and
kinase-dead forms of PIKfyve, the mammalian Fab1p homologue, produces
a phenotype that includes extensive vacuolation of mammalian cells, as does
treatment with a cell-permeant PIKfyve inhibitor [23,24].
The identity of the swollen compartment in animal cells has not been
unequivocally determined. This is an important issue, since the animal equivalent
of the yeast vacuole is a matter of some dispute. Many researchers assume the
yeast vacuole is a lysosome-like organelle, but recent definition of the BLOC
(Biogenesis-of-Lysosome-related-organelles-Complex) protein complexes, which
are required for the formation of mammalian lysosomes, and the absence of BLOC
proteins from yeast have called this into question [24]. The vacuole (pH 6.0) is less
acidic than a classical lysosome (pH 5.0–4.0), also suggesting important differences.
Could it be that the vacuole is actually more like a modified late endosome?
However, the identity of the swollen compartment in studies of the
C. elegans PtdIns3P 5-kinase mutant ppk-3 seems in no doubt [4]. In this
organism, the huge vacuolar organelles that accumulate in many cell types
are characterized by the presence of lysosomal glycoproteins, indicating that
© 2007 The Biochemical Society
BY0074-0012.indd Page 132 12/5/06 2:10:27 PM elhi /Users/elhi/Desktop/5 Dec_Bijender/indd
132
S.K. Dove and Z.E. Johnson
these organelles have a lysosomal or post-lysosomal origin. The authors of this
study also contend that PtdIns(3,5)P2-dependent membrane retrieval may occur
only from post-lysosomal compartments in higher organisms, since lysosomes
themselves form normally in these mutants [4]. The lysosomes in ppk-3 mutants
were not normally acidified, but this did not impair their ability to degrade
proteins; these properties phenocopy the yeast fab1Δ mutant.
These results suggest that PtdIns(3,5)P2 is relevant to the function of true
lysosomes and that the defects associated with the yeast fab1Δ mutants can be
generalized to include at least some metazoans. However, a deeper understanding
of the roles of PtdIns3P 5-kinases and PtdIns(3,5)P2 in higher organisms is
hindered by the fact that the PtdIns3P 5-kinases so far examined seem to localize
to different compartments in each study: e.g. the human kinase appears on early
endosomes [25], while the mouse kinase appears mainly on late endosomes
[26]. Whether this represents cell type-specific rather than organism-specific
differences awaits further investigation. Moreover, many of the studies have been
undertaken with epitope-tagged and over-expressed proteins, the localization of
which may be perturbed. It is possible that Fab1p-like enzymes normally cycle
between several different compartments but can accumulate in one of these
when trafficking is blocked or slowed following over-expression or tagging.
Another recent study suggesting strong parallels with the yeast system
identified PIKfyve, the mammalian Fab1p homologue, as a factor required for
replication of enveloped viruses, such as Ebola and HIV [5]. This requirement
was not for sorting of viral proteins into MVBs, as might have been expected,
many RNA viruses are known to assemble the MVB sorting machinery at the
cell surface or induce exocytosis of MVBs thereby gaining their envelopes
and effecting escape from the cell. Instead, it appears that PIKfyve is needed
for a membrane retrieval step involved in retrograde late endosome-to-Golgi
trafficking (Step 2 in Figure 2). PIKfyve mutants, and also Rab9, TIP47 and p40
mutants, showed a block in viral assembly because the membrane-localized viral
Gag protein could not get from the CD63-positive late endosome to the TGN
(Step 2, Figure 2) and hence to the cell surface (Step 3, Figure 2) [5]. Caution
should be exercized when interpreting these results, as this might represent a
PtdIns(3,5)P2-independent function of PIKfyve; it has been demonstrated that
PIKfyve can also phosphorylate p40, a kelch domain-containing effector of the
Rab9 GTPase, and so localize p40 to membranes [27].
PtdIns(3,5)P2 might also be required for this retromer-mediated late
endosome-to-Golgi trafficking step: one homologue of the PtdIns(3,5)P2 effector
protein Svp1p/Atg18p (known as WIPI-49) is implicated in this pathway in
mammalian cells (but it is reported to bind to PtdIns3P) [28]. Further studies are
required to clarify the role of PIKfyve and PtdIns(3,5)P2 in the retromer pathway.
PtdIns(3,5)P2 and stress
A dramatic accumulation of PtdIns(3,5)P2 occurs in yeast cells subjected
to a hyperosmotic stress, beginning within seconds of challenge [1]. This can
boost PtdIns(3,5)P2 concentrations up to 30 times higher than those present in
© 2007 The Biochemical Society
BY0074-0012.indd Page 133 12/5/06 2:10:27 PM elhi /Users/elhi/Desktop/5 Dec_Bijender/indd
Molecular functions for PtdIns(3,5)P2
133
Figure 2 PtdIns(3,5)P2- and/or Fab1p-dependent membrane trafficking.
In a generalized cell, PtdIns(3,5)P2 or the enzyme that synthesizes it (a Type III
PtdInsP kinase; Fab1p in yeast or PIKfyve in mammals) is required for two
retrograde membrane trafficking steps. In yeast, Step 1 is PtdIns(3,5)P2dependent, and the same could be true in C. elegans, while in mammals Step 2
appears to depend on the protein kinase activity of PIKfyve. Step 2 is required
for the life cycle of many enveloped viruses because their Gag protein (or
equivalent) must traffic via Step 2 if they are to be delivered to the cell surface
via Step 3. Is it possible that phosphoinositides specify the direction of trafficking
in the endocytic system, with PtdIns3P required for anterograde trafficking and
PtdIns(3,5)P2 for retrograde trafficking? Or is there some mechanism based on
the inter-conversion of PtdIns3P of PtdIns(3,5)P2 that balances the amount of
membrane fluxing through the endocytic system in the two directions?
unstimulated cells. This discovery was unexpected; fab1K mutants display no
gross defect in growth under hyperosmotic conditions, but the response displays a number of features that suggest that it is a tightly regulated event. First,
the PtdIns(3,5)P2 accumulation is transient, and the magnitude and duration of
the response are proportional to the applied osmotic stress; 0.4 M NaCl produces a much lesser accumulation of PtdIns(3,5)P2 than 0.9 M, peaking at approx.
5 min, and PtdIns(3,5)P2 levels return to unstimulated levels within approx.
10 min (S.K. Dove, unpublished work). In contrast, it can take an hour for the
PtdIns(3,5)P2 concentration to return to normal after a 1.1 M NaCl treatment
and the peak does not occur until approx. 15 min [1].
Similar, though less dramatic, PtdIns(3,5)P2 accumulation has been reported
in algal cells [29] and one type of mammalian cell [30] in response to osmotic
stress; most mammalian cells do not respond thus. Similarly, if the mouse Fab1p
homologue, PIKfyve, is transfected into fab1Δ yeast and the yeast are subjected
to hyperosmotic shock, then a comparable accumulation of PtdIns(3,5)P2 occurs
to that occurring in wild-type cells [3,31]; the machinery seems to be sufficiently
conserved to be essentially interchangeable.
© 2007 The Biochemical Society
BY0074-0012.indd Page 134 12/5/06 2:10:28 PM elhi /Users/elhi/Desktop/5 Dec_Bijender/indd
134
S.K. Dove and Z.E. Johnson
The mechanism of PtdIns(3,5)P2 accumulation has not been studied
extensively in any of these systems but could result from increased synthesis,
decreased degradation or some combination of the two. Our initial
data suggested that increases in PtdIns3P 5-kinase activity are primarily
responsible, at least in yeast at early time-points (S.K. Dove, unpublished
work). This is borne out by the finding that dominant mutants of Fab1p can
be isolated that produce much higher basal levels of PtdIns(3,5)P2 but also
accumulate PtdIns(3,5)P2 only to the normal stressed level when subjected
to a hyperosmotic stress [20]. One way of looking at this is to suggest that
increases in kinase activity are the main determinant of stress-stimulated
accumulation of PtdIns(3,5)P2; if decreased degradation was the cause,
these hyper-active kinase mutants would be predicted to accumulate more
PtdIns(3,5)P2 after stress. This is based upon a model in which these mutants
are more active because they are partially released from auto-inhibition, but
still have the same maximum intrinsic lipid kinase activity as wild-type Fab1p
(see below).
The stress-provoked accumulation of PtdIns(3,5)P2 is reversed by 5dephosphorylation by Fig4p-type phosphoinositide 5-phosphomonoesterases
[20]. Unexpectedly, however, a recent report suggests that Fig4p is also required for hyperosmotic shock-mediated accumulation of PtdIns(3,5)P2;
the authors suggest that Fig4p acts both as a protein phosphatase and as a
Fab1p activator [20]. An alternative possibility is that some proportion of the
PtdIns(3,5)P2 accumulation is a result of inhibition of the lipid phosphatase
activity of Fig4p. If so, it is probable that Vac14p controls this process; Vac14p
is required for the some of the PtdIns(3,5)P2 accumulation during hyperosmotic stress and is known to associate with Fig4p, localize this protein at the
vacuole and stabilize it [20,32]. Vac14p certainly seems to be required for
normal dephosphorylation of PtdIns(3,5)P2 after any stress-mediated accumulation; some workers find far less Fig4p in vac14Δ cells, while others have
reported no difference.
Deleting VAC14 has a more profound effect on PtdIns(3,5)P2 synthesis,
under both basal and stress stimulated conditions, than deletion of FIG4, so
Vac14p probably also activates Fab1p directly [13,14]. Vac14p is a large 880
amino acid peripheral membrane protein whose association with the vacuole
membrane appears to be partially dependent on the presence of Fab1p and/
or PtdIns(3,5)P2. There is little other information on this protein, except that
two-hybrid studies suggest a direct interaction between Fig4p and Vac14p [13],
whereas we and others have so far failed to find a direct interaction between
Vac14p and Fab1p.
Available evidence strongly suggests that Vac7p, confined to fungi, is the
major mediator of PtdIns(3,5)P2 accumulation in stressed yeast [20]; Vac7p is
a 1166 amino acid residue integral membrane protein that has no recognized
domains [12]. It is not currently known how Vac7p makes its way to the vacuole
(e.g. via either the CPY or AP-3 pathways). Recent work suggests that Vac7p
acts independently of the Vac14p-Fig4p complex, in accord with the observation
that this protein is absent from animals and plants [20].
© 2007 The Biochemical Society
BY0074-0012.indd Page 135 12/5/06 2:10:28 PM elhi /Users/elhi/Desktop/5 Dec_Bijender/indd
Molecular functions for PtdIns(3,5)P2
135
Fab1p: normally an auto-inhibited kinase?
The isolation of constitutively active mutants of Fab1p (such as FAB1-14,
FAB1-6 and FAB1-5) that are dominant over wild-type Fab1p [19,20] lends
weight to the suggestion that Vac14p–Fig4p and Vac7p facilitate PtdIns(3,5)P2
synthesis by relieving an auto-inhibition of Fab1p. Essentially, the hypothesis
is that these Fab1p mutants are ‘locked’ in a partially activated conformation,
possibly as a result of disruption of auto-inhibitory interactions between the
domains of Fab1p. We made this suggestion because even massive over-expression
of Fab1p, to levels that exceed 100 times the normal levels of this protein, does
not cause major increases in PtdIns(3,5)P2 levels in cells. This can hardly be
explained by the presence of cytoplasmic inhibitors as these could never be
abundant enough to hold all of the extra Fab1p in check under such conditions.
Verification of this hypothesis must await expression and in vitro measurements
of the kinase activities of the Fab1p variants. Our predictions would be that
these enzymes will be more intrinsically active in isolation than wild-type Fab1p
and that these hyperactive Fab1p mutants will be less susceptible to activation
by Vac7p and the Vac14p–Fig4p complex; it would be expected that wild-type
and hyperactive Fab1p variants will all achieve approximately the same maximal
kinase activity in the presence of these activators. This appears to be the situation
in vivo, [20] but a test of this model will have to await an in vitro kinase assay
that is sufficiently physiological to allow Vac14p and/or Vac7p activation of
Fab1p to be measured directly.
Stress sensors in the Fab1p pathway?
One question that still remains to be answered is how stress signals are
transmitted to Fab1p and by which protein(s)? It is known that the Hog1p
MAPK (mitogen-activated protein kinase) pathway does not communicate
with Fab1p, and so there must exist an osmo-sensor or membrane tension
sensor in cells that in some way transmits this signal to Fab1p [1]. PtdIns(3,5)P2
accumulation begins very rapidly in response to stress, so the activation cascade
is probably short, and is unlikely to involve gene activation. Indeed, the presence
of a CCT chaperonin-like domain in the Fab1p protein has led to the suggestion
that Fab1p may itself sense stress by detecting changes in the conformation of
a partner protein(s) (Vac7p and Vac14p). This is possible since the CCT-like
domain of Fab1p corresponds to the substrate-binding region of a chaperone
and the fab1-1 allele has a mutation in this CCT domain that prevents Fab1p
activation in response to hyper-osmotic stress [3].
PtdIns(3,5)P2 in regulation of vacuole pH and connections with stress
Osmotic stress is not the only way to cause PtdIns(3,5)P2 accumulation; both
heating and alkalinization of the extracellular medium to pH 7.5 provoke a similar
response [33], albeit much less potently and with slower kinetics of induction,
than hyperosmotic stress. The fact that these two stresses cause PtdIns(3,5)P2
© 2007 The Biochemical Society
BY0074-0012.indd Page 136 12/5/06 2:10:28 PM elhi /Users/elhi/Desktop/5 Dec_Bijender/indd
136
S.K. Dove and Z.E. Johnson
accumulation is particularly satisfying, because the growth of fab1Δ mutants is
compromised under the same two conditions [9,33]. Both heat and alkalinization
produce more sustained responses, with the peak of PtdIns(3,5)P2 maintained
for at least an hour. It will be very interesting to examine the regulation of the
Fab1p pathway in response to these disparate stimuli to discover to what extent
these responses use the same sensory elements.
The authors of the alkali stress study suggest that their response is due
to buffer affecting the vacuole pH; Tris is taken up by endocytosis [33]. They
suggest that Fab1p is being activated to re-establish the acidic pH of the vacuole
via activation of the vacuolar H+-ATPase and that the Fab1p pathway is
activated by alkalinization of the vacuole relative to the cytosol. There appears
to be some merit to this suggestion given that when they buffered the medium
to more acidic pHs (at least in the presence of low concentrations of weak
acids) they reduced basal levels of PtdIns(3,5)P2 to below those normally seen
in resting cells. This strongly suggests that the Fab1p pathway can sense and is
able to respond to pH, becoming progressively more active as the extracellular
fluid becomes more alkaline. Whether this is detected through the vacuole or
some other compartment downstream of endocytosis of extracellular medium
will remain unresolved until this response is examined in mutants that cannot
endocytose extracellular fluid (e.g. end4 mutants).
Whatever the case, this response seems to fit nicely with the observation
that PtdIns(3,5)P2 is required for proper vacuole acidification; it suggests that
the PtdIns(3,5)P2 response is dynamic and adapts to the pH across the vacuolar
membrane. A possible prediction of this model might be that a regulator or subunit
of the vacuolar H+-ATPase (or an anion channel required for vacuole acidification;
e.g. a Cl− channel) might serve as a yet-to-be-identified PtdIns(3,5)P2 effector.
Functions for stress-stimulated PtdIns(3,5)P2 accumulation
The function of stress-stimulated PtdIns(3,5)P2 accumulation is not entirely
clear. One plausible hypothesis is that some stresses increase PtdIns(3,5)P2,
bringing about activation of the H+-ATPase (as discussed above). This is very
likely for alkali stress, but not for hyperosmotic stress. One study has carefully
examined the effects of acute hyperosmotic stress on cell and vacuole volume
and pH in S. cerevisiae [34]. It reports a very rapid (approx. 50%) decrease in the
volume of both the cell and the vacuole within 1 min of applying hyperosmotic
stress. This is a surprising result as it has been assumed that the vacuole acts as
a water reservoir for the cell during hyperosmotic stress. This study however,
rules out such a mechanism, as the cell and the vacuole shrink at the same
time and rate, suggesting that they are osmotically continuous. This work also
shows that the pH of the vacuole and the cytosol of S. cerevisiae decreases by
0.5 pH units during the passive flow of water out of the cell/vacuole during
hyperosmotic stress. This occurs because the H+ permeability of cell membranes
is known to decrease during hyperosmotic stress and so retention of protons in a
lowered cell volume is sufficient to explain this pH shift [34]. This suggests that
the function of PtdIns(3,5)P2 during hyperosmotic stress is not connected with
the H+-ATPase, since lowering internal pH (via weak acids) of both cytosol and
vacuole normally reduces PtdIns(3,5)P2.
© 2007 The Biochemical Society
BY0074-0012.indd Page 137 12/5/06 2:10:28 PM elhi /Users/elhi/Desktop/5 Dec_Bijender/indd
Molecular functions for PtdIns(3,5)P2
137
So if PtdIns(3,5)P2 does not bring about activation of the H+-ATPase in
hyper-osmotically stressed cells, what else might it be doing? Elegant studies
from Lois Weisman’s laboratory have begun to shed light on another possible
function for the hyperosmotically provoked PtdIns(3,5)P2 accumulation; this was
the first laboratory to show vacuole fragmentation in response to hyper-osmotic
stress, and that this fragmentation requires PtdIns(3,5)P2 (see Figure 3) [14].
Moreover, the vacuoles of unstressed cells fragment in other situations in which
PtdIns(3,5)P2 accumulates (e.g. over-expression of Fab1p or expression of one of
the hyperactive Fab1p mutants). This suggests that Fab1p and/or PtdIns(3,5)P2
is required for regulating vacuole size, number and surface area-to-volume ratio
[20]. It is also consistent with the observation that fab1Δ cells have a defect in
vacuole inheritance, as this process requires regulated changes in vacuole shape
to form the vesicles and tubules of the so-called segregation structure that is
required for inheritance.
One obvious way that PtdIns(3,5)P2 could bring about changes in
vacuole size, shape and number is by an increase in membrane budding from
the surface of the vacuole, which is the process thought to be controlled
by the PtdIns(3,5)P2 effector Svp1p/Atg18p; svp1Δ cells accumulate 6-fold
more PtdIns(3,5)P2 than wild-type cells when stressed, but display a defect
in vacuole fragmentation [15]. However, svp1Δ cells do not have an obvious
defect in vacuole inheritance (at least in the BY4742 strain background);
segregation structures appear very commonly in svp1Δ mutants, but not in
fab1Δ mutants.
Studies on Yck3p, one of three yeast isoforms of casein kinase I, have
offered another possible insight into the above. Yck3p was recently identified
Figure 3 Effect of hyperosmotic stress on vacuole morphology in fab1
and yck3 mutants. Idealized yeast cells are shown with a representation
of the size and morphology of their vacuoles. After 10 min of hyperosmotic
stress, wild-type vacuoles fragment and remain in this state until the stress is
relived. In fab1 mutants this vacuole fragmentation never happens while in yck3
mutants, fragmentation occurs but is not sustained and reverses by 60 min.
© 2007 The Biochemical Society
BY0074-0012.indd Page 138 12/5/06 2:10:29 PM elhi /Users/elhi/Desktop/5 Dec_Bijender/indd
138
S.K. Dove and Z.E. Johnson
in a screen for mutants that show the initial hyperosmotically driven and
PtdIns(3,5)P2-dependent vacuole fragmentation, but which subsequently show
re-fusion of the fragmented vacuoles (see Figure 3) [35]. Yck3p is therefore
required to maintain vacuoles in the fragmented state after PtdIns(3,5)P2
levels have returned to basal. This study suggests that one target of Yck3pmediated phosphorylation might be Vps41p, an important part of the HOPS
tethering complex that is required for homotypic vacuole–vacuole fusion [35].
Phosphorylation of Vps41p is proposed to block tethering and so ‘freeze’
vacuoles in the fragmented state after hyperosmotic stress.
These data suggest the possibility of some relationship between PtdIns(3,5)P2
accumulation and Yck3p activation as these processes happen sequentially in
response to the same stimuli. It seems probable that Yck3p acts after PtdIns(3,5)P2,
and is therefore downstream of Fab1p on an additional effector arm of this
pathway. At the very least, a common sensor probably lies upstream of both
Fab1p and Yck3p. When considering other possible connections, it is notable
that the two Svp1p-related PROPPIN proteins, Hsv1p/Atg21p and Hsv2p, can
both localize to the vertex ring of fusing vacuoles, exactly where the HOPS
complex localizes [36]. It is tempting to speculate that these two PtdIns(3,5)P2
effectors may be involved in PtdIns(3,5)P2-dependent inhibition of homotypic
vacuole fusion.
Studies on the Fab1p pathway and the functions of PtdIns(3,5)P2 just keep
on yielding surprises, and this looks set to continue for the foreseeable future.
Might the next ten years yield a more complete understanding of the functions
of PtdIns(3,5)P2 than for any other inositol lipid?
We would like to thank Professor Bob Michell, Dr Frank Cooke, Dr Mike Clague,
Professor Harald Stenmark, Dr Paul Whitley and Dr Sylvie Urbe for stimulating
discussions over the years. We would also like to thanks Dr Nevin Perera, Dr Victoria
Heath and Dr Robert McEwen for past contributions to research in this area and
present members of the Phosphoinositide Laboratory in Birmingham for sharing
their ideas and views. The BBSRC, The Wellcome Trust and the Royal Society fund
research in our laboratory. S.K.D. is a Royal Society University Research Fellow.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
Dove, S.K., Cooke, F.T., Douglas, M.R., Sayers, L.G., Parker, P.J. and Michell, R.H. (1997)
Nature 390, 187–192
Efe, J.A., Botelho, R.J. and Emr, S.D. (2005) Curr. Opin. Cell Biol. 17, 402–408
Michell, R.H., Heath, V.L., Lemmon, M.A. and Dove, S.K. (2006) Trends Biochem. Sci. 31,
52–63
Nicot, A.S., Fares, H., Payrastre, B., Chisholm, A.D., Labouesse, M. and Laporte, J. (2006)
Mol. Biol. Cell 17, 3062–3074
Murray, J.L., Mavrakis, M., McDonald, N.J., Yilla, M., Sheng, J., Bellini, W.J., Zhao, L., Le
Doux, J.M., Shaw, M.W., Luo, C.C. et al. (2005) J. Virol. 79, 11742–11751
Whiteford, C.C., Brearley, C.A. and Ulug, E.T. (1997) Biochem. J. 323, 597–601
Herman, P.K. and Emr, S.D. (1990) Mol. Cell Biol. 10, 6742–6754
Schu, P.V., Takegawa, K., Fry, M.J., Stack, J.H., Waterfield, M.D. and Emr, S.D. (1993)
Science 260, 88–91
Yamamoto, A., DeWald, D.B., Boronenkov, I.V., Anderson, R.A., Emr, S.D. and Koshland,
D. (1995) Mol. Biol. Cell 6, 525–539
© 2007 The Biochemical Society
BY0074-0012.indd Page 139 12/5/06 2:10:29 PM elhi /Users/elhi/Desktop/5 Dec_Bijender/indd
Molecular functions for PtdIns(3,5)P2
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
139
Cooke, F.T., Dove, S.K., McEwen, R.K., Painter, G., Holmes, A.B., Hall, M.N., Michell,
R.H. and Parker, P.J. (1998) Curr. Biol. 8, 1219–1222
Gary, J.D., Wurmser, A.E., Bonangelino, C.J., Weisman, L.S. and Emr, S.D. (1998) J. Cell
Biol. 143, 65–79
Bonangelino, C.J., Catlett, N.L. and Weisman, L.S. (1997) Mol. Cell Biol. 17, 6847–6858
Dove, S.K., McEwen, R.K., Mayes, A., Hughes, D.C., Beggs, J.D. and Michell, R.H. (2002)
Curr. Biol. 12, 885–893
Bonangelino, C.J., Nau, J.J., Duex, J.E., Brinkman, M., Wurmser, A.E., Gary, J.D., Emr,
S.D. and Weisman, L.S. (2002) J. Cell Biol. 156, 1015–1028
Dove, S.K., Piper, R.C., McEwen, R.K., Yu, J.W., King, M.C., Hughes, D.C., Thuring, J.,
Holmes, A.B., Cooke, F.T., Michell, R.H. et al. (2004) EMBO J. 23, 1922–1933
Odorizzi, G., Babst, M. and Emr, S.D. (1998) Cell 95, 847–858
Friant, S., Pecheur, E.I., Eugster, A., Michel, F., Lefkir, Y., Nourrisson, D. and Letourneur,
F. (2003) Dev. Cell 5, 499–511
Whitley, P., Reaves, B.J., Hashimoto, M., Riley, A.M., Potter, B.V. and Holman, G.D. (2003)
J. Biol. Chem. 278, 38786–38795
Gary, J.D., Sato, T.K., Stefan, C.J., Bonangelino, C.J., Weisman, L.S. and Emr, S.D. (2002)
Mol. Biol. Cell 13, 1238–1251
Duex, J.E., Tang, F. and Weisman, L.S. (2006) J. Cell Biol. 172, 693–704
Walker, D.M., Urbe, S., Dove, S.K., Tenza, D., Raposo, G. and Clague, M.J. (2001) Curr.
Biol. 11, 1600–1605
Schaletzky, J., Dove, S.K., Short, B., Lorenzo, O., Clague, M.J. and Barr, F.A. (2003) Curr.
Biol. 13, 504–509
Ikonomov, O.C., Sbrissa, D. and Shisheva, A. (2001) J. Biol. Chem. 276, 26141–26147
Gautam, R., Novak, E.K., Tan, J., Wakamatsu, K., Ito, S. and Swank, R.T. (2006) Traffic 7,
779–792
Cabezas, A., Pattni, K. and Stenmark, H. (2006) Gene 371, 34–41
Ikonomov, O.C., Sbrissa, D., Yoshimori, T., Cover, T.L. and Shisheva, A. (2002) J. Biol.
Chem. 277, 46785–46790
Sbrissa, D., Ikonomov, O.C. and Shisheva, A. (2005) Methods Enzymol. 403, 696–705
Jeffries, T.R., Dove, S.K., Michell, R.H. and Parker, P.J. (2004) Mol. Biol. Cell 15, 2652–2663
Meijer, H.J. and Munnik, T. (2003) Annu. Rev. Plant Biol. 54, 265–306
Sbrissa, D. and Shisheva, A. (2005) J. Biol. Chem. 280, 7883–7889
McEwen, R.K., Dove, S.K., Cooke, F.T., Painter, G.F., Holmes, A.B., Shisheva, A., Ohya,
Y., Parker, P.J. and Michell, R.H. (1999) J. Biol. Chem. 274, 33905–33912
Rudge, S.A., Anderson, D.M. and Emr, S.D. (2004) Mol. Biol. Cell 15, 24–36
Mollapour, M., Phelan, J.P., Millson, S.H., Piper, P.W. and Cooke, F.T. (2006) Biochem.
J. 395, 73–80
Vindelov, J. and Arneborg, N. (2002) Yeast 19, 429–439
LaGrassa, T.J. and Ungermann, C. (2005) J. Cell Biol. 168, 401–414
Barth, H., Meiling-Wesse, K., Epp0le, U.D. and Thumm, M. (2002) FEBS Lett. 512,
173–179
© 2007 The Biochemical Society