PtdIns4P is a key regulator of the secretory pathway and plays an essential role in trafficking from the Golgi. Our recent work demonstrated that spatial control of PtdIns4P at the ER (endoplasmic reticulum) and Golgi co-ordinates secretion with cell growth. The central elements of this regulation are specific phosphoinositide 4-kinases and the phosphoinositide phosphatase Sac1. Growth-dependent translocation of Sac1 between the ER and Golgi modulates the levels of PtdIns4P and anterograde traffic at the Golgi. In yeast, this mechanism is largely dependent on the availability of glucose, but our recent results in mammalian cells suggest that Sac1 phosphatases play evolutionarily conserved roles in the growth control of secretion. Sac1 lipid phosphatase plays also an essential role in the spatial control of PtdIns4P at the Golgi complex. A restricted pool of PtdIns4P at the TGN (trans-Golgi network) is required for Golgi integrity and for proper lipid and protein sorting. In mammalian cells, the stress-activated MAPK (mitogen-activated protein kinase) p38 appears to play a critical role in transmitting nutrient signals to the phosphoinositide signalling machinery at the ER and Golgi. These results suggest that temporal and spatial integration of metabolic and lipid signalling networks at the Golgi is required for controlling the secretory pathway.
Stimulation of cell growth in eukaryotes requires both increased biosynthetic capacity and enhanced membrane-based transport of macromolecules. It is well established that ribosome biogenesis and protein translation are rapidly up-regulated upon growth stimulation, and regulatory mechanisms have been characterized . Analysis of yeast mutants showed that a functional secretory pathway is also critical for cell-surface growth . Yet how anterograde trafficking of proteins and lipids is controlled by growth and nutrient-based signals is largely unknown. The importance of secretion for cell growth has been highlighted in several studies, suggesting that growth of cancer cells requires stimulated membrane traffic. For example, expression of a subset of trafficking regulators belonging to the family of Rab and Arf GTPases are up-regulated in liver, breast and ovarian cancer [3,4]. Among the different classes of membrane transport regulators, phosphoinositide lipids play a particularly prominent role. The different phosphoinsositide species display unique compartment-specific distributions within the cell and provide spatial cues for membrane trafficking . Within the secretory pathway, PtdIns4P appears to be the key phosphoinositide regulator . PtdIns4P is concentrated at Golgi membranes and has essential functions in Golgi integrity, regulation of Golgi enzyme localization and anterograde trafficking from the Golgi . The present review summarizes our current understanding of how Golgi PtdIns4P levels are regulated by metabolic and growth signals. The central elements of this control mechanisms are evolutionarily conserved and consist of specific lipid signalling enzymes that associate with the Golgi under certain growth conditions. This growth-specific regulation of Golgi PtdIns4P is crucial for co-ordinating the capacity of the secretory pathway with cell proliferation.
Metabolic control of Golgi phosphoinsoitides in yeast
A genetic screen in yeast for novel regulators of secretion discovered a critical role for Golgi PtdIns4P in controlling anterograde trafficking from the Golgi to the plasma membrane . Yeast cells have three PI4Ks (phosphoinositide 4-kinases), namely Pik1p, Stt4p and Lsb6p [8–10]. Both Stt4p and Pik1p are essential for cell viability, but regulate different cellular processes [11–13]. Pik1p localizes at the Golgi and is essential for anterograde trafficking . In contrast, Stt4p resides at the plasma membrane and synthesizes a pool of PtdIns4P that is converted into PtdIns(4,5)P2 and plays a role in actin cytoskeletal organization and protein kinase C signalling . Both Pik1p and Stt4p play also distinct roles in regulating MAPK (mitogen-activated protein kinase) signalling [14,15].
The most important lipid phosphatase that controls intracellular levels of PtdIns4P is the polyphosphoinositide phosphatase Sac1p . The yeast SAC1 gene was originally identified in a screen for suppressors of actin mutations, but the functional basis for this genetic interaction remains unclear . A conserved N-terminal lipid phosphatase domain in yeast Sac1p (also termed the Sac domain) is the defining property of a specific class of lipid phosphatases . Sac phosphatase domain-containing proteins are able to dephosphorylate a variety of phosphoinositide species including PtdIns3P, PtdIns4P and PtdIns(3,5)P2 . The recently resolved crystal structure of the Sac phosphatase domain of yeast Sac1p shows that the Sac domain comprises two closely packed subdomains, a unique N-terminal domain and a catalytic lipid phosphatase subdomain, which contains a conserved catalytic CX5R(T/S) motif [18,19]. Topology analysis showed that Sac1p is a type II membrane protein containing two transmembrane domains near the C-terminus . Among the Sac domain-containing lipid phophatases, Sac1p is the only integral membrane protein, and the majority of Sac1p localizes at the Golgi complex and at the ER (endoplasmic reticulum) [20,21]. Deletion of yeast SAC1 results in a dramatic elevation of cellular PtdIns4P levels and leads to an increase in the forward transport of Chs3p chitin synthase from its intracellular stores to the cell periphery and thus leads to specific cell wall defects [22,23]. Elimination of yeast Sac1p additionally causes impaired ATP transport into the ER .
In contrast with its mammalian orthologues, yeast Sac1p lacks any consensus sequence motifs that regulate retrieval to the ER . Previous studies have shown that Dol-P-Man (dolichol phosphate mannose) synthase Dpm1p interacts with Sac1p and is responsible for ER localization of Sac1p during exponential cell growth  (Figure 1). Dpm1p synthesizes Dol-P-Man, which is an essential mannosyl donor for luminal steps in oligosaccharide biosynthesis at the ER . The interaction between Sac1p and Dpm1p persists only during exponential growth and is rapidly abolished when cell growth slows due to restricted nutrient availability . Under glucose-starved conditions, Sac1p quickly translocates to the Golgi, which requires the adaptor protein Rer1p and the COPII (coat protein complex II) [25,27] (Figure 1). The glucose-starvation-induced ER–Golgi shuttling of Sac1p is promptly reversed when glucose is added back  (Figure 1). Golgi association of the PI4K Pik1 is also regulated by glucose. Under high glucose conditions, Pik1p localizes at the Golgi by binding to its non-catalytic subunit Frq1p and produces a Golgi-specific pool of PtdIns4P [27,28] (Figure 1). During nutrient deprivation, the Pik1p–Frq1 complex is rapidly released from the Golgi to the cytoplasm [27,28] (Figure 1). A portion of Pik1 also translocates into the nucleus  (Figure 1). This reaction is rapidly reversed upon restoration of nutrient supply [27,28].
Glucose regulation of Golgi phosphoinositides in yeast
Nutrient-dependent Sac1p localization controls distinct pools of PtdIns4P, generated by different lipid kinases [27,29,30]. When nutrients are plentiful, Sac1p resides at the ER, but controls a specific pool of PtdIns4P at the plasma membrane synthesized by Stt4p lipid kinase [29,30]. This reaction occurs at ER–plasma membrane contact sites and requires the small ER adaptor proteins Scs2p/Scs22p and Osh (oxysterol-binding homology) proteins that act as PtdIns4P sensors and activate Sac1 phosphatase activity  (Figure 1). The physiological significance of this interesting enzymatic reaction is unclear, but, given that individual yeast Osh proteins associate with different intracellular membranes , this type of regulation may represent a general mechanism for controlling phosphoinositide signalling at organellar membrane interfaces. During glucose starvation, Sac1p quickly translocates to the Golgi and turns over a pool of PtdIns4P that is generated by Pik1p, which in turn down-regulates anterograde traffic . Glucose availability therefore regulates the Golgi association of Sac1p lipid phosphatase and the Pik1p–Frq1p lipid kinase complex in an opposite manner and thus aligns the trafficking capacity of the Golgi with cell growth conditions.
Phosphoinositide signalling at the Golgi and growth control of secretion in mammalian cells
The mechanism for regulating Golgi PtdIns4P is evolutionarily conserved and present in organisms from yeast to mammals. As in yeast, the mammalian Sac1 lipid phosphatase orthologues appear to be central elements in this regulation . However, Golgi PtdIns4P in mammals is synthesized by several PI4Ks . Previous studies have shown that both PI4KIIα and PI4KIIIβ are involved in generating PtdIns4P at the Golgi [33,34]. These kinases may control functionally distinct Golgi PtdIns4P pools and perhaps regulate different lipid and protein trafficking pathways [34–37]. PI4KIIIβ is the closest homologue of yeast Pik1p and forms a complex with the N-myristoylated Ca2+-binding NCS1 (neuronal calcium sensor 1), which is homologous with yeast Frq1p . Whether the PI4KIIIβ–NCS1 complex dissociates from the Golgi upon serum or nutrient starvation remains unknown, but the activity of PI4KIIIβ was shown to be regulated by PKD (protein kinase D) and 14-3-3 proteins .
Like yeast Sac1p, mammalian Sac1 proteins translocate between the ER and Golgi in response to growth conditions, which is regulated by growth factors rather than glucose  (Figure 2). Serum starvation induces almost quantitative accumulation of Sac1 at the Golgi . This trafficking reaction is functionally equivalent to the glucose starvation-induced Golgi accumulation of Sac1p in yeast and also results in specific down-regulation of Golgi PtdIns4P, which dampens anterograde traffic . Differently from yeast Sac1p, all mammalian Sac1 orthologues contain a canonical dilysine COPI (coat protein complex I) recognition motif . Consequently, trafficking of mammalian Sac1 proteins is regulated differently from yeast Sac1p. Under normal growth conditions, the dilysine motif is sufficient for continuous retrieval of Sac1 from the Golgi to the ER, where the majority of the enzyme is localized in steady state [40,41]. Starvation-induced accumulation of mammalian Sac1 at the Golgi requires oligomerization of the protein, which depends on a leucine repeat motif present at the N-terminal cytoplasmic domain [40,41]. The exact mechanism by which oligomerization facilitates anterograde trafficking of Sac1 is unclear, but it is possible that active recruitment into COPII vesicles plays a role. Previous studies have shown that oligomerization of membrane cargos such as plasma membrane receptors can be a requirement for ER exit . Alternatively, oligomerization could mask the COPI-interacting motif, which would disable retrograde trafficking of Sac1.
Mitogen-dependent regulation of Golgi phosphoinositides in mammals
The shift in localization of Sac1 from the ER to the Golgi occurs within 24–48 h of serum starvation . It therefore appears likely that the actual ER–Golgi redistribution of Sac1 is initiated after cells have exited the cell cycle. In contrast, when quiescent cells are stimulated by growth factors, Sac1 rapidly returns to the ER with a half-time of approximately 15 min, and this reaction appears to be one of the earliest events triggered by growth signalling  (Figure 2). The growth-factor-induced stimulation of retrograde trafficking of Sac1 coincides with the dissociation of Sac1 oligomers and requires p38 MAPK activity  (Figure 2). Whether Sac1 itself is a p38 MAPK target is unknown. It is clear, however, that activation of p38 MAPK by expression of activated alleles of the upstream MKK (MAPK kinase) 3 or MKK6 kinases is sufficient to trigger Sac1 oligomer dissociation and stimulate retrograde traffic, even in the absence of growth factor signalling . These results indicate that trafficking of Sac1 is regulated not only during proliferation and during quiescence, but also probably when cells are exposed to stressors that activate p38 MAPK. Interestingly, inhibition of p38 MAPK by specific drugs eliminates the rapid up-regulation of secretion that is normally observed when cells are stimulated with growth factors . These results provide direct evidence that both p38 MAPK signalling and Sac1 lipid phosphatase activity play central roles in regulating constitutive secretion in response to starvation and proliferation signals.
Role for Sac1 lipid phosphatase in spatial regulation of Golgi PtdIns4P
Mammalian Sac1 has an additional important housekeeping function at the Golgi during cell proliferation that is necessary for the spatial regulation of PtdIns4P. In rapidly dividing cells, the majority of Sac1 resides at the ER [40,41]. Yet a certain proportion of Sac1 is also located at the Golgi, even when cells are stimulated with growth factors [40,43]. The Golgi-resident portion of SAC1 is restricted to cisternal Golgi regions and absent from the TGN (trans-Golgi network) . This distribution of Sac1 results in a steep cis–trans PtdIns4P gradient, with the majority of PtdIns4P remaining confined to the TGN. The spatial segregation of PtdIns4P and Sac1 is dependent on Sac1 phosphatase activity . These results therefore provide a mechanistic explanation for the polarized pools of Golgi PtdIns4P reported previously .
What is the physiological significance of spatially restricted Golgi PtdIns4P pools? Many of the effectors of Golgi PtdIns4P also bind to activated ARF1 (ADP-ribosylation factor 1) and play a role in anterograde trafficking from the TGN . The concentrated pool of PtdIns4P at the TGN provides directionality for these processes. Golgi PtdIns4P is also important for the targeting of several lipid-transfer proteins . In addition, PtdIns4P is required for the regulation of the Golgi phospholipid flippase Drs2 . These factors appear to play a role in controlling the lipid composition at the TGN, which may facilitate formation of vesicular and tubular transport carriers . This idea is supported by the fact that the PtdIns4P effector FAPP2 (four-phosphate adaptor protein 2) not only functions as a glucosylceramide-transfer protein, but also has an independent activity in membrane tubule formation [37,46–48]. Finally, PtdIns4P interacts with proteins that play a role in Golgi morphology. For example the PtdIns4P effector GOLPH3 (Golgi phosphoprotein 3) binds to unconventional myosins and seems to participate in shaping the Golgi . Vps74 (vacuolar protein sorting 74), the yeast homologue of GOLPH3, also binds to PtdIns4P, but, in addition, interacts with several Golgi glycosylation enzymes and regulates the proper Golgi localization of these proteins . A role for spatially controlled Golgi PtdIns4P pools in Golgi morphology and retrograde trafficking of glycosylation enzymes is supported by Sac1 siRNA (small interfering RNA) knockdown experiments. Elimination of Sac1 lipid phosphatase induces delocalization of Golgi PtdIns4P and causes plasma membrane accumulation of glycosylation enzymes . In addition, Golgi morphology and mitotic spindle organization is significantly impaired . PtdIns4P therefore plays multiple spatially defined roles at the Golgi. Additional regulatory factors, such as small GTPases from the Arf and Rab family, are required in a coincidence detection mechanism to selectively recruit effector proteins to the appropriate sites.
Among the seven described phosphoinositide lipids, PtdIns4P is the most important regulator of the secretory pathway. PtdIns4P plays critical roles in regulating Golgi morphology and stability, but is particularly important for facilitating lipid and protein sorting at the Golgi. Metabolic and mitogenic control of PtdIns4P links the capacity of the secretory pathway to cell growth and proliferation. Although lipid phosphatases and kinases that regulate Golgi PtdIns4P levels have been identified, it remains unresolved how growth and nutrient signals are transmitted to the lipid signalling machinery at the ER and Golgi. Further characterization of the control of Golgi phosphoinositides will be important to gain new insights into the cross-talk between growth signalling and the secretory pathway, which may help to identify novel drug targets against the uncontrolled cell growth in cancer and other cell proliferative diseases.
Signalling 2011: a Biochemical Society Centenary Celebration: A Biochemical Society Focused Meeting held at the University of Edinburgh, U.K., 8–10 June 2011. Organized and Edited by Nicholas Brindle (Leicester, U.K.), Simon Cook (The Babraham Institute, U.K.), Jeff McIlhinney (Oxford, U.K.), Simon Morley (University of Sussex, U.K.), Sandip Patel (University College London, U.K.), Susan Pyne (University of Strathclyde, U.K.), Colin Taylor (Cambridge, U.K.), Alan Wallace (AstraZeneca, U.K.) and Stephen Yarwood (Glasgow, U.K.).
coat protein complex I
coat protein complex II
dolichol phosphate mannose
Golgi phosphoprotein 3
mitogen-activated protein kinase
neuronal calcium sensor 1
protein kinase D
This work was supported in part by the National Institutes of Health [grant numbers GM71569 and GM84088].