Download Sec14p-like proteins regulate phosphoinositide homoeostasis and

346
Biochemical Society Transactions (2006) Volume 34, part 3
Sec14p-like proteins regulate phosphoinositide
homoeostasis and intracellular protein and lipid
trafficking in yeast
C.J. Mousley, K.R. Tyeryar, M.M. Ryan and V.A. Bankaitis1
Department of Cell and Developmental Biology, Lineberger Comprehensive Cancer Center, School of Medicine, University of North Carolina at Chapel Hill,
Chapel Hill, NC 27599-7090, U.S.A.
Abstract
The major PI (phosphatidylinositol)/PC (phosphatidylcholine)-transfer protein in yeast, Sec14p, co-ordinates
lipid metabolism with protein transport from the Golgi complex. Yeast also express five additional gene
products that share 24–65% primary sequence identity with Sec14p. These Sec14p-like proteins are termed
SFH (Sec Fourteen Homologue) proteins, and overexpression of certain individual SFH gene products rescues
sec14-1ts -associated growth and secretory defects. SFH proteins are atypical in that these stimulate the
transfer of PI, but not PC, between distinct membrane bilayer systems in vitro. Further analysis reveals
that SFH proteins functionally interact with the Stt4p phosphoinositide 4-kinase to stimulate PtdIns(4,5)P2
synthesis which in turn activates phospholipase D. Finally, genetic analyses indicate that Sfh5p interfaces
with the function of specific subunits of the exocyst complex as well as the yeast SNAP-25 (25 kDa
synaptosome-associated protein) homologue, Sec9p. Our current view is that Sfh5p regulates PtdIns(4,5)P2
homoeostasis at the plasma membrane, and that Sec9p responds to that regulation. Thus SFH proteins
individually regulate specific aspects of lipid metabolism that couple, with exquisite specificity, with key
cellular functions.
Introduction
PITPs (phosphatidylinositol-transfer proteins) are characterized by their ability to mediate the transfer of phosphatidylinositol or PC (phosphatidylcholine) monomers between membrane bilayers in vitro [1,2]. The major PITP in
the yeast Saccharomyces cerevisiae, encoded by the essential
SEC14 gene, localizes to the Golgi apparatus where its function is necessary for vesicle-mediated protein transport from
this organelle [3,4]. Observations regarding the function of
Sec14p in vivo derive from extensive characterization of the
conditional sec14-1ts allele and analysis of extragenic loss of
function mutations that permit cell viability in the complete
absence of essential Sec14p function. This ‘bypass Sec14p’
condition has proven invaluable in increasing our understanding of how Sec14p regulates secretory vesicle biogenesis
from the trans-Golgi network. We now know it does so
by generating an appropriate pro-secretory lipid environment in those membranes [2,3]. Accordingly, Sec14p dysfunction leads to a significant derangement of several aspects
of lipid metabolism which subsequently impose a defect
in secretory vesicle biogenesis from trans-Golgi membranes
Key words: lipid trafficking, phosphatidylinositol-transfer protein (PITP), phospholipase D,
Sec14p, target membrane soluble N-ethylmaleimide-sensitive fusion protein attachment protein
receptor (t-SNARE), yeast.
Abbreviations used: DAG, diacylglycerol; ER, endoplasmic reticulum; GFP, green fluorescent
protein; PA, phosphatidic acid; PI, phosphatidylinositol; PITP, PI-transfer protein; PC,
phosphatidylcholine; PLD, phospholipase D; PS, phosphatidylserine; SFH, Sec Fourteen
Homologue; (t-)SNARE, (target membrane) soluble N-ethylmaleimide-sensitive fusion protein
attachment protein receptor.
1
To whom correspondence should be addressed (email [email protected]).
C 2006
Biochemical Society
[4–9]. This is accompanied by a dramatic disorganization of
the actin cytoskeleton in Sec14p-deficient cells [10].
Interestingly, Sec14p-independent cell growth in all
‘bypass Sec14p’ mutants requires the constitutive activation
of PLD (phospholipase D) [11,12], encoded by the nonessential SPO14 gene, which catalyses the hydrolysis of PC
to choline and PA (phosphatidic acid). It is postulated that the
hydrolysis of PC by PLD generates a critical pool of PA that
is subsequently converted into DAG (diacylglycerol), and
that it is this generation of DAG at the expense of PC that
helps compensate for the loss of Sec14p function in sec14-1ts
cells [7,12–14].
The essential function of Sec14p notwithstanding, it is now
clear the PITP repertoire of even simple yeast cells is more
complex than expected. The yeast genome encodes five additional genes whose translational products are homologous
with Sec14p and are required for PLD activation in bypass
sec14 mutants [12]. Herein, we summarize our current understanding of the function of these Sec14p-like yeast PITPs, and
discuss roles for such proteins in regulating membrane trafficking and dynamics of the actin cytoskeleton.
The yeast Sec14p-like family
The S. cerevisiae genome encodes five genes whose protein
products share at least 25% identity and 45% similarity to
Sec14p, and are termed SFH1–SFH5 (Sec Fourteen Homologue 1–5) (Figure 1) [12]. Soluble fractions prepared
by salt-stripping of membranes harvested from yeast
cells overproducing Sfh2p, Sfh3p, Sfh4p or Sfh5p exhibit
Non-Vesicular Intracellular Traffic
Figure 1 The S. cerevisiae Sec14p-like gene family
A schematic comparison of Sec14p with the five yeast Sec14p-like proteins is shown. The gene names are given on the left and the percentage
primary sequence identity/similarity to Sec14p is indicated within the
grey area that represents the region of homology with Sec14p. The number of residues in each protein is given at the upper right of each corresponding box depiction.
Sec14p-independent PI (phosphatidylinositol)-transfer activity in vitro, indicating that the corresponding SFH proteins
are also PITPs. These SFH proteins are biochemically distinguished from Sec14p in their inability to transfer PC in vitro,
however [12]. That Sfh2p, Sfh4p and Sfh5p are functionally
related to Sec14p is indicated by two additional lines of
evidence. Firstly, increased gene dosage of SFH2, SFH4 or
SFH5 is sufficient to rescue the conditional lethality of
sec14-1ts yeast. Secondly, this phenotypic rescue is accompanied by improved Golgi secretory capacity, as demonstrated
by the increased capacity of sec14-1ts cells to secrete invertase
[12]. It is also interesting that the SFH protein most related to
Sec14p by homology is not a functional Sec14p on the basis
of several important criteria. Yeast cells overexpressing Sfh1p
have at best a very low grade Sec14p-independent PI-transfer
activity in vitro, and Sfh1p overexpression does not effect an
efficient rescue of sec14 defects in vivo [12,15].
Their biochemical similarities notwithstanding, we now
know that SFH proteins represent a novel class of nonclassical fungal PITPs whose individual functions are substantially unique and non-redundant. The nature of these
individual functions is only now coming to light, and recent
progress on this subject is reviewed below.
SFH PITPs and Stt4p stimulate PLD activity
The SFH proteins do not individually, or collectively, execute
essential cellular functions. Also, en bloc deletion of the SFH
genes in the sec14-1ts genetic background does not compromise cell viability, implying that SFH proteins do not define a
minor family of PITPs which share functional redundancy
with Sec14p. However, SFH proteins are required for
Sec14p-independent cell growth in ‘bypass Sec14p’ mutants,
as collective deletion of all SFH genes abolishes Sec14pindependent growth [12]. Intriguingly, this phenotype is reminiscent of that associated with PLD deficiency in sec14-1ts
cells, as constitutive PLD activity is required to support
Sec14p-independent cell growth in all ‘bypass Sec14p’
mutants [11].
The SFH/PLD phenotypic similarity in the context of
‘bypass Sec14p’ suggested a first clue regarding a biological
function for the SFH PITPs, namely, a role for SFH proteins
in optimal stimulation of PLD activity. Experimental evidence supports this concept. Collective ablation of the SFH2–
SFH5 genes in yeast significantly compromises the PLD
activity in vivo evoked by Sec14p inactivation, as measured
both by PLD-dependent release of choline and generation
of PA [12]. Thus SFH proteins support the course of PLD
activation that occurs in Sec14p-deficient cells. This finding
enlightens our understanding of a conundrum presented by
the demonstration that PLD activity is strongly elevated upon
inactivation of Sec14p. Liscovitch and Cantley [16] have previously synthesized the known lipid activation properties of
PLD and phosphoinositide kinases into an attractive proposal that PITPs drive a positive feedback loop that stimulates PLD enzymatic activity. While the activation of PLD
by inactivation of Sec14p is spectacularly counter to this
model, the potential employment of the SFH PITPs in an
SFH/phosphoinositide kinase/PLD regulatory circuit provides an alternative PITP/PLD circuit that satisfies the basic
tenets of the same model.
SFH proteins regulate PI metabolism
in vivo
The biochemical properties of SFH proteins suggest a simple
mechanism for how these PITPs activate PLD. Given that
PtdIns(4,5)P2 is an obligate cofactor for PLD enzymatic activity [17], we anticipated that SFH proteins stimulate
the synthesis of 4-OH phosphorylated phosphoinositides
in vivo. Indeed, SFH2 overexpression increases both
PtdIns(4)P and PtdIns(4,5)P2 in yeast strains with baseline
phosphoinositide levels, yet PtdIns(3)P levels are not affected
[10]. Overproduction of Sfh4p and Sfh5p, unlike overproduction of Sfh2p, does not increase PtdIns(4)P, but does result
in increased PtdIns(4,5)P2 . Taken together, these findings
imply that Sfh2p, Sfh4p and Sfh5p modulate PtdIns(4)P
and/or PtdIns(4,5)P2 metabolism in vivo. Similarly, collective
ablation of the SFH2–SFH5 genes evokes a 40% reduction
in bulk PtdIns(4,5)P2 relative to isogenic wild-type controls
[10].
How do these non-conventional PITPs stimulate phosphoinositide synthesis? Sfh2p, Sfh4p and Sfh5p do not appear
to directly stimulate the activity of any of the three yeast
phosphoinositide 4-kinases (e.g. Pik1p, Stt4p and Lsb6p),
nor do they stimulate the activity of the single yeast phosphoinositide 4-phosphate 5-kinase in vitro (Mss4p; [10]).
Rather, SFH proteins probably stimulate phosphoinositide
production by regulating the delivery of PI substrate to the
appropriate phosphoinositide kinase(s).
Do SFH proteins channel PI to a specific phosphoinositide
kinase or are these proteins more promiscuous in this regard
in the usual physiological setting? The results indicate that
the functional SFH/phosphoinositide kinase interface is a
C 2006
Biochemical Society
347
348
Biochemical Society Transactions (2006) Volume 34, part 3
Table 1 Increased phosphatidylinositol 4-phosphate in
phosphoinositide phosphatase-deficient yeast with elevated
Sfh2p or Sfh5p
Yeast sec14∆ cki1∆ double mutants exhibit baseline levels of all phosphoinositide species and these mutants are specifically sensitized to
increased Stt4p activity by inactivation of the Sac1p phosphoinositide
phosphatase [10]. Steady-state phosphatidylinositol 4-phosphate levels
for the parental triple mutant strain carrying a control YEp(URA3) plasmid
or Sfh2p and Sfh5p overexpression plasmids (YEpSFH2 and YEpSFH5
respectively) are given. Values are expressed as percentage of total
deacylatable inositol phospholipids.
Phosphatidylinositol 4-phosphate
Yeast strain genotype
(% total deacylatable inositol
phospholipid)
sec14∆ cki1 sac1∆ YEp(URA3)
sec14∆ cki1 sac1∆ YEp(SFH2)
sec14∆ cki1 sac1∆ YEp(SFH5)
5.7 ± 0.4
7.6 ± 0.9
9.3 ± 0.5
specific one. Disruption of either SFH2 or SFH5 exacerbates
the growth defects associated with stt4-4ts , but not pik1-83ts
growth defects, implying that both Sfh2p and Sfh5p functionally interact with Stt4p [10]. In agreement with the genetic
interaction data, Stt4p-dependent PtdIns(4)P synthesis is
stimulated by the overexpression of either SFH2 or SFH5
when the experiment is performed in a phosphoinositide
phosphatase-deficient genetic background (Table 1) [10].
A functional interface between the Sfh proteins and Stt4p
strongly implies that these molecules are important in synthesizing extra-Golgi pools of phosphoinositides.
As specific SFH proteins functionally interact with Stt4p,
and SFH proteins are required to activate PLD, it is expected
that PLD activity should also be sensitive to robustness of
Stt4p function. This expectation is also realized as thermal
inactivation of the Stt4-4ts protein evokes reductions in PLD
activation that are comparable in magnitude with those recorded upon wholesale deletion of the SFH genes [10]. This
is fully consistent with the idea that PLD activation in vegetative cells requires the concerted activity of SFH proteins
and Stt4p.
Sfh2p and Sfh5p restore polarized actin
organization to Sec14p-deficient cells
Are there other downstream cellular effects mediated by
SFH regulation of the available PtdIns(4,5)P2 pool? In this
regard, Sec14p orthologues in Schizosaccharomyces pombe
and Arabidopsis thaliana are implicated in modulating postGolgi membrane trafficking as well as in the organization
of the actin cytoskeleton [18,19]. Actin polarization is
responsive to exocytic membrane flux and phosphoinositides
[20,21]. In yeast, the distribution of actin is tightly polarized
whereby actin filaments traverse the entire length of the
mother cell, enter the bud, and concentrate as cortical patches
within the bud. However, this distinct organization is dramatically compromised in Sec14p-deficient yeast as the cortical
C 2006
Biochemical Society
actin patches become distributed throughout the mother and
daughter cells as puncta and the actin cables are diminished
[10]. Consistent with SFH proteins displaying functional
homology with Sec14p, overexpression of SFH2 or SFH5
corrects the actin cytoskeletal disorganization that stems
from Sec14p insufficiency. The effects of SFH proteins on
regulation of actin dynamics are also discernible in SEC14
yeast. The actin cytoskeleton in sfh∆ mutants is unable to
undergo the rapid remodelling normally recorded when yeast
cells are challenged with stresses such as heat shock [10].
Sfh5p-mediated control of plasma
membrane PtdIns(4,5)P2 and efficient
Sec9p t-SNARE (target membrane soluble
N-ethylmaleimide-sensitive fusion protein
attachment protein receptor) function
Stt4p and Mss4p reside in the yeast plasma membrane [22].
This raises the possibility that SFH proteins modulate exocytic and/or endocytic events at the plasma membrane. To
investigate whether the yeast SFH proteins modulate exocytosis, each SFH gene product was individually overexpressed in a number of sec ts mutants that block membrane
trafficking at various points along the secretory pathway.
Overexpression of SFH1–SFH4 fails to restore growth of
any sects strain (other than sec14ts ) at restrictive temperatures,
but SFH5 overexpression reverses the temperature-sensitive
growth and secretory defects associated with each of the
sec8-9ts , sec10-2ts and sec15-1ts mutations [10]. These mutations all affect subunits of the exocyst complex that itself
plays an essential role in docking the 90 nm Golgi-derived
secretory vesicles to the plasma membrane [23]. The mechanism by which overexpression of SFH5 suppresses defects
in components of the exocyst is PLD-independent.
How does Sfh5p levy these effects on exocyst function?
The answer seems to lie in the interface between regulation
of plasma membrane PtdIns(4,5)P2 levels and the activity of
Sec9p, the yeast version of the neuronal SNAP-25 (25 kDa
synaptosome-associated protein) [24]. In this regard, overexpression of the Mss4p phosphoinositide 4-phosphate 5kinase resembles Sfh5p overexpression in its partial rescue
of the temperature-sensitive growth defects associated with
sec8-9ts , sec10-2ts and sec15-1ts mutant alleles, as well as
growth defects associated with sec9ts , sec1ts and sec2ts mutants.
By the same token, overexpression of the Sec9p t-SNARE
elicits partial rescue of a spectrum of sects mutations that very
closely resembles that recorded for Mss4p overexpression
[10]. Both the Mss4p- and the Sec9p-dependent rescue
is Sfh5p-dependent and PLD-independent. Moreover, the
Mss4p-dependent rescue is specific with respect to the phosphoinositide-4-phosphate 5-kinase activity, as elevated dosage of either the PIK1 or LSB6 phosphoinositide 4-kinase
structural genes fails to evoke the same effects [10].
The Sfh5p-dependent rescue of post-Golgi sects secretory
defects levied by overexpression of Mss4p implies a mechanism involving elevation of a plasma membrane pool of
PtdIns(4,5)P2 , and evidence in favour has been obtained
Non-Vesicular Intracellular Traffic
Figure 2 Sfh5p localizes to peripheral ER
Figure 3 An intermembrane contact site model for SFH protein
Fluorescence imaging experiments show that a Myc epitope-tagged
Sfh5p (in red; upper right) substantially co-localizes with the peripheral,
function
(A) An SFH protein (grey) may itself be an integral component of an
but not the nuclear envelope-associated, aspect of the ER, as marked by
a Sec61p–GFP chimaera (in green; upper left). The merge panel (lower
right) shows that Sfh5p is restricted to discrete subdomains of peripheral
intermembrane contact site and employ its PI-binding capacity (in black)
to impose specific PI trafficking through that site. (B) An SFH protein may
employ its ability to channel PI to a phosphoinositide kinase to generate
ER (arrows). That Sfh5p is not co-localized with Stt4p on the yeast
plasma membrane was demonstrated previously [10]. The lower left
panel shows a Nomarski image of the cells examined.
a phosphoinositide platform that recruits intrinsic components (Factor X)
of an intermembrane contact site. In this scenario, the SFH protein does
not directly utilize its PI-binding properties to impose PI-trafficking specificity through such a site. Other lipid species could pass through such a
site.
from PtdIns(4,5)P2 imaging experiments. A PLD–GFP (green
fluorescent protein) chimaera is a reliable PtdIns(4,5)P2 sensor that translocates to the plasma membrane upon Mss4p
overexpression [25]. This relocalization of PLD–GFP is compromised in sfh5∆ cells, indicating that Sfh5p contributes to
the synthesis and/or maintenance of the Mss4p-dependent
pool of plasma membrane PtdIns(4,5)P2 [10].
How does PtdIns(4,5)P2 regulate Sec9p t-SNARE function? While the details are unclear, phosphoinositidemediated regulation of SNARE function may well emerge
as a common theme. In this regard, PtdIns(4,5)P2 reduces the
diffusion rate of t-SNAREs in membranes [26]. Also, phosphoinositides co-operate with DAG and ergosterol in controlling the formation of SNARE complexes required for the
homotypic fusion of vacuolar membranes [27].
Discussion
Herein, we discuss the body of biochemical and genetic evidence to indicate that SFH proteins are non-classical PITPs
that channel PI to the Stt4p phosphoinositide 4-kinase. This
metabolic channelling influences a number of cellular functions that include PLD activation, organization of the cytoskeleton and, in the case of Sfh5p, plasma membrane
phosphoinositide homoeostasis. In the last case, the Sfh5presponsive plasma membrane PtdIns(4,5)P2 pool couples
with the activity of the post-Golgi secretory vesicle docking/
fusion machinery at the plasma membrane. Our interpreta-
tion of the various results presented here (and elsewhere) is
that SFH proteins, and PITPs in general, are individually
dedicated to the modulation of specific lipid metabolic
events that themselves couple with very specific biological
functions. This exquisite functional specificity is at odds
with the startling functional promiscuity exhibited by PITPs
in cell-free or permeabilized membrane systems that purportedly reconstitute PITP-dependent processes. We find it
an attractive notion that SFH proteins, and perhaps PITPs
in general, represent specificity factors that help define the
identity of distinct phosphoinositide/lipid pools.
How SFH proteins may play a role in PI channelling and
pool specification is unclear. With regard to channelling, SFH
proteins are peripheral membrane proteins that need to be
salt-stripped from membranes [10]. Thus these proteins may
not mobilize PI between membranes via a soluble PI/SFH
protein transport intermediate. The case of Sfh5p suggests
other interesting possibilities. Sfh5p, while functionally engaged with Stt4p, nonetheless localizes to subdomains of
the peripheral ER (endoplasmic reticulum) (Figure 2), and
not to the plasma membrane where Stt4p resides. The
localization patterns of Sfh5p and Stt4p are not coincident
[10]. Taken together, the results are consistent with a model
whereby Sfh5p acts as a component of a machinery that
forms intermembrane contact sites, thereby providing a nonvesicular and non-carrier mechanism for transfer of PI from
one membrane to another (Figure 3). The lipid-binding
specificity of Sfh5p could either gate the contact site with
regard to which lipids are competent to pass through it, or
Sfh5p may function in regulating the assembly of such a
site.
C 2006
Biochemical Society
349
350
Biochemical Society Transactions (2006) Volume 34, part 3
The contact site model may be generally applicable to
the SFH proteins. In this regard, Sfh4p is implicated in
an intermembrane contact site-dependent transport of PS
(phosphatidylserine) from the ER to the extramitochondrial
PS decarboxylase [28] where it is subsequently converted
into phosphatidylethanolamine. Interestingly, this metabolic
involvement is specific to Sfh4p, as neither Sec14p, Sfh2p,
Sfh3p nor Sfh5p can substitute for Sfh4p in this pathway
[10]. This finding argues that SFH proteins do not define
general components of a putative contact site machinery. It
leaves open the possibility that SFH proteins define lipid pool
identity by regulating the function of specific intermembrane
contact sites that control the metabolic flux into and out of
specific lipid pools.
This work was supported by grant RO1-GM3370 from the National
Institutes of Health to V.A.B.
References
1 Bankaitis, V.A., Malehorn, D.E., Emr, S.D. and Greene, R. (1989)
J. Cell Biol. 108, 1271–1281
2 Bankaitis, V.A., Aitken, J.R., Cleves, A.E. and Dowhan, W. (1990)
Nature (London) 347, 561–562
3 Cleves, A.E., McGee, T.P. and Bankaitis, V.A. (1991) Trends Cell Biol. 1,
30–34
4 Cleves, A.E., McGee, T.P., Whitters, E.A., Champion, K.M., Aitken, J.R.,
Dowhan, W., Goebl, M. and Bankaitis, V.A. (1991) Cell 64, 789–800
5 McGee, T.P., Skinner, H.B., Whitters, E.A., Henry, S.A. and Bankaitis, V.A.
(1994) J. Cell Biol. 124, 273–287
6 Skinner, H.B., McGee, T.P., McMaster, C.R., Fry, M.R., Bell, R.M. and
Bankaitis, V.A. (1995) Proc. Natl. Acad. Sci. U.S.A. 92, 112–116
7 Kearns, B.G., McGee, T.P., Mayinger, P., Gedvilaite, A., Phillips, S.E.,
Kagiwada, S. and Bankaitis, V.A. (1997) Nature (London) 387, 101–105
8 Bankaitis, V.A., Phillips, S.E., Yanagisawa, L., Li, X., Routt, S. and Xie, Z.
(2005) Adv. Enzyme Regul. 45, 155–170
C 2006
Biochemical Society
9 Hama, H., Schnieders, E.A., Thorner, J., Takemoto, J.Y. and DeWald, D.
(1999) J. Biol. Chem. 274, 34294–34301
10 Routt, S.M., Ryan, M.M., Tyeryar, K., Rizzieri, K.E., Mousley, C.,
Roumanie, O., Brennwald, P.J. and Bankaitis, V.A. (2005) Traffic 12,
1157–1172
11 Xie, Z., Fang, M., Rivas, M.P., Faulkner, A., Sternweis, P.C., Carman, G.C.,
Engebrecht, J. and Bankaitis, V.A. (1998) Proc. Natl. Acad. Sci. U.S.A. 95,
1246–1251
12 Li, X., Routt, S., Xie, Z., Cui, X., Fang, M., Kearns, M.A., Bard, M., Kirsch, D.
and Bankaitis, V.A. (2000) Mol. Biol. Cell 11, 1989–2005
13 Yanagisawa, L., Marchena, J., Xie, Z., Li, X., Poon, P.P., Singer, R.,
Johnston, G., Randazzo, P.A. and Bankaitis, V.A. (2002) Mol. Biol. Cell 13,
2193–2206
14 Phillips, S., Sha, B., Topalof, L., Xie, Z., Alb, J., Clenchin, V., Swigart, P.,
Cockcroft, S., Luo, M., Martin, T. et al. (1999) Mol. Cell 4, 187–197
15 Schnabl, M., Oskolkova, O.V., Holic, R., Brezna, B., Pichler, H.,
Zagorsek, M., Kohlwein, S.D., Paltauf, F., Daum, G. and Griac, P. (2003)
Eur. J. Biochem. 270, 3133–3145
16 Liscovitch, M. and Cantley, L.C. (1995) Cell 81, 659–662
17 Sciorra, V.A., Rudge, S.A., Wang, J., McLaughlin, S., Engebrecht, J. and
Morris, A.J. (2002) J. Cell Biol. 159, 1039–1049
18 Vincent, P., Chua, M., Nogue, F., Fairbrother, A., Mekeel, H., Xu, Y.,
Allen, N., Bibikova, T.N., Gilroy, S. and Bankaitis, V.A. (2005) J. Cell Biol.
168, 801–812
19 Nakase, Y., Nakamura, T., Hirata, A., Routt, S.M., Skinner, H.B., Bankaitis,
V.A. and Shimoda, C. (2001) Mol. Biol. Cell 4, 901–917
20 Goldschmidt-Clermont, P.J., Machesky, L.M., Baldasarre, J.J. and Pollard,
T.D. (1990) Science 247, 1575–1578
21 Roumanie, O., Wu, H., Molk, J.N., Rossi, G., Bloom, K. and Brennwald, P.
(2005) J. Cell Biol. 170, 583–594
22 Audhya, A. and Emr, S.D. (2002) Dev. Cell 2, 593–605
23 TerBush, D.R., Maurice, T., Roth, D. and Novick, P. (1996) EMBO J. 15,
6483–6494
24 Brennwald, P., Kearns, B., Champion, K., Keranen, S., Bankaitis, V. and
Novick, P. (1994) Cell 79, 245–258
25 Coluccio, A., Malzone, M. and Neiman, A.M. (2004) Genetics 166, 89–97
26 Wagner, M.L. and Tamm, L.K. (2001) Biophys. J. 81, 266–275
27 Fratti, R.A., Jun, Y., Merz, A.J., Margolis, N. and Wickner, W. (2004)
J. Cell Biol. 167, 1087–1098
28 Wu, W.-I., Routt, S., Bankaitis, V.A. and Voelker, D. (2000) J. Biol. Chem.
275, 14446–14456
Received 12 December 2005