A Plasma-Membrane E-MAP Reveals Links of the Eisosome With Sphingolipid Metabolism and Endosomal Trafficking

A Plasma-Membrane E-MAP Reveals Links of the Eisosome With Sphingolipid Metabolism and Endosomal Trafficking
of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  nature structural & molecular biology   VOLUME 17 NUMBER 7 JULY 2010 901 resource The pasma membrane is the defining feature of the ce, separatingits interior from the exterior space. It contros exchange and commu-nication processes between the ce and its environment. The deivery of ceuar materia to the pasma membrane or ce exterior is medi-ated by exocytosis. Conversey, endocytosis is used to take up pasmamembrane and externa components. In addition, many signaingprocesses occur at the pasma membrane simutaneousy and are oftenreguated by the endocytosis of receptors or deivery of messengermoecues. To coordinate these processes and maintain ce integ-rity under changing conditions, both pasma-membrane protein andipid composition are reguated and adjusted to externa conditions.Despite impressive advances in our understanding of these individuaprocesses, it is not we understood how they are coordinated.To accommodate its many functions, the pasma membrane ishighy organized, both spatiay and temporay. In Saccharomycescerevisiae , severa pasma-membrane domains of different compo-sition are distinguishabe by ight microscopy. This organization ismediated, at east in part, by eisosomes, arge protein compexes thatunderie one of the domains, named MCC after the marker proteinCan1 found there. When PIL1 , encoding a major eisosome compo-nent, is deeted, ces have abnorma pasma-membrane structurewith arge invaginations and oss of MCC protein organization 1,2 .In addition, the endocytosis of severa pasma-membrane proteinsis either acceerated or deayed 2,3 . The moecuar function of eiso-somes is sti unknown, but recent data show that they interact withsphingoipid-reguated Pkh-kinases, which phosphoryate their corecomponents and are required for efficient endocytosis 4–6 . In addi-tion to Pkh-kinases, Tor kinase compex 2 (TORC2) is impicated insphingoipid metaboism reguation 7 . However, it is uncear how thesedifferent signaing pathways are controed and coordinated as we aswhat their downstream effects are. Experimenta evidence supportsa mode in which reguation of sphingoipid, stero and gycerophos-phoipid eves in the pasma membrane are coordinated, but mecha-nistic insights as to how this is achieved are currenty acking 8,9 . Torevea functiona inks between the different processes, we generateda quantitative genetic-interaction map targeting a arge set of genesimpicated in pasma-membrane function.Genetic interactions have ong been used to dissect functionareationships between genes. Cassicay, researchers have ookedfor quaitative differences between observed phenotypes of doubemutants and the phenotypes of the two reated singe mutants. Morerecenty, we empoyed the epistatic miniarray profie (E-MAP)approach, a variation on synthetic genetic arrays 10 . This aows forthe quantitative anaysis of genetic interactions, incuding negative(for exampe, synthetic sick or etha) as we as positive ones (forexampe, suppression) 11 . For this approach, a comprehensive setof doube mutants is generated and their growth is measured. Todetermine individua genetic interactions, deviations of growth ratesfrom the medians of a combinations with one particuar gene arecacuated for each combination as a quantitative interaction score 1 Institut Pasteur de Montevideo, Montevideo, Uruguay. 2 Max Planck Institute o Biochemistry, Organelle Architecture and Dynamics, Martinsried, Germany. 3 Department o Cellular and Molecular Pharmacology, University o Caliornia, San Francisco, Caliornia, USA. 4 The Blavatnik School o Computer Science,Tel Aviv University, Tel Aviv, Israel. 5 Department o Biochemistry and Biophysics, University o Caliornia and Howard Hughes Medical Institute, San Francisco,Caliornia, USA. 6 Max Planck Institute o Biochemistry, Proteomics and Signal Transduction, Martinsried, Germany. 7 University o Southern Denmark, Departmento Biochemistry and Molecular Biology, Odense, Denmark. 8 Present address: Whitehead Institute or Biomedical Research, Cambridge, Massachusetts, USA. 9 Theseauthors contributed equally to this work. Correspondence should be addressed to N.J.K. (krogan@cmp.ucs.edu) or T.C.W. (twalther@biochem.mpg.de).Received 10 December 2009; accepted 9 April 2010; published online 6 June 2010;doi:10.1038/nsmb.1829 A pasma-membrane E-MAP reveas inks of the eisosomewith sphingoipid metaboism and endosoma trafficking Pabo S Auiar 1,9 , Forian Fröhich 2,9 , Michae Rehman 2,9 , Mike Shaes 3,9 , Ior Uitsky  4,8 , Austina Oivera-Couto 1 ,Hannes Braber 3 , Ron Shamir 4 , Peter Water 5 , Matthias Mann 6 , Christer S Ejsin 7 , Nevan J Kroan 3 & Tobias C Wather 2 The plasma membrane delimits the cell and controls material and information exchange between itself and the environment.How different plasma-membrane processes are coordinated and how the relative abundance of plasma-membrane lipids andproteins is homeostatically maintained are not yet understood. Here, we used a quantitative genetic interaction map, or E-MAP,to functionally interrogate a set of ~400 genes involved in various aspects of plasma-membrane biology, including endocytosis,signaling, lipid metabolism and eisosome function. From this E-MAP, we derived a set of 57,799 individual interactions betweengenes functioning in these various processes. Using triplet genetic motif analysis, we identified a new component of the eisosome,Eis1, and linked the poorly characterized gene EMP70 to endocytic and eisosome function. Finally, we implicated Rom2,a GDP/GTP exchange factor for Rho1 and Rho2, in the regulation of sphingolipid metabolism.  902 VOLUME 17 NUMBER 7 JULY 2010 nature structural & molecular biology resource (or S-score) 12,13 . Each mutation has a genetic-interaction profie, orphenotypic signature, consisting of a its S-scores with a other genesin the E-MAP. A particuary usefu parameter to judge the simiaritiesof profies is to compare correations of two genes’ interactions with aother genes in the set. In addition, bioinformatic extraction based onmathematica modes can be appied to yied functiona modues in anunbiased fashion from E-MAP datasets, and correations and S-scorescan be used to revea their connections 14,15 . The E-MAP approach hasbeen previousy used to functionay interrogate severa processes, andthe dissection of genetic interactions from these E-MAPs has ed to adeuge of bioogica insights in a variety of processes 11,16–18 .Here we report an E-MAP targeting pasma-membrane functions to gene-rate previousy unknown bioogica insight reating to pasma-membranefunctions. Using this E-MAP, we have inked two new genes ( EMP70 and EIS1 ) to eisosome function and uncovered a ink between GDP/GTPexchange protein Rom2 signaing and sphingoipid metaboism. RESULTSOverview of the plasma-membrane E-MAP To address functiona reationships between pasma-membrane processes,we systematicay determined the genetic interactions among a set of 374 genes invoved in pasma-membrane bioogy. We seected candidategenes encoding proteins functioning in membrane transport and organi-zation, especiay eisosomes, actin patches, endocytosis and exocytosis.In addition, we picked genes invoved in ergostero and sphingoipidmetaboism, as these ipids are impicated in many pasma-membraneprocesses. Our seection criteria were based on avaiabe functiona anno-tation (gene ontoogy terms) and a iterature survey. We aso incudeda diverse set of genes whose products ocaize to the pasma membraneand/or interact geneticay or physicay with previousy characterizedpasma-membrane genes/proteins. The seected genes were categorizedinto the functiona groups presented in Figure 1a and Supplementary Table 1 . We incuded a number of genes anayzed in previous systematicgenetic studies to faciitate comparison between datasets 11,16,17 . Fromthis set, we quantitated a tota of 57,799 genetic interactions using theE-MAP approach (~83% of the possibe interactions).Previousy, we found that gene pairs encoding physicay interact-ing proteins are enriched for positive genetic interactions and showa higher propensity for having highy correated genetic-interactionprofies 11,16,17 . To assess the richness and quaity of the genetic- interaction data of the pasma-membrane E-MAP, we compared theousy deveoped agorithm that defines functiona modues from quan-titative genetic and PPI data 14 ( Supplementary Fig. 2 ). This methodidentified 18 modues encompassing 53 genes ( Supplementary Fig. 2  and Supplementary Table 3 ). Genes in each modue have simiargenetic-interaction profies and form a connected subnetwork inthe PPI network. These modues corresponded to known proteincompexes, such as the F-actin capping protein compex and theAP-3 adaptor, or to known pathways, such as sphingoipid metabo-ism, the HOG osmosensory pathway and ergostero biosynthesis( Supplementary Fig. 2 ). To identify modues for which PPI data isnot avaiabe, we performed the moduar anaysis without requiringPPI connectivity ( Supplementary Fig. 3 ). This identified 29 mod-ues encompassing 190 genes ( Supplementary Table 4 andhttp://acgt.cs.tau.ac.i/pmemap). This anaysis yieded simiar amountsof modues for the pasma membrane and the previousy reportedE-MAP on the eary secretory pathway  11 ( Supplementary Table 5 ).Additiona information can be extracted by considering interactionsof singe genes with modues (data not shown). Insights from hierarchical clustering of the genetic-interaction data Each mutant engenders a genetic-interaction profie, or phenotypic sig-nature, representing how it geneticay interacts with a other mutantstested. Comparison of these profies using hierarchica custering( Fig. 2 , Supplementary Data andhttp://interactome-cmp.ucsf.edu/pasma_membrane/) is a powerfu and unbiased approach to identify genes of the same pathway. In the foowing, we provide a brief summary of severa functiona connections reveaed by such gene custering. RVS161 and RVS167  encode proteins that operate together in mem-brane remodeing during endocytosis 22 . As expected from their over-apping functions, rvs161 Δ and rvs167  Δ custered together with highcorreation (correation = 0.54; Fig. 2 , inserts 2 ). Consistent with previ-ous reports, both share positive genetic interactions with a number of genes invoved in fatty-acid eongation for sphingoipid synthesis, suchas FEN1 and SUR4 (ref. 23) ( Fig. 2 , insert 2d). Notaby, we observedpositive interactions with genes encoding components of the Hog1MAP-kinase cascade and the ergostero biosynthesis pathway ( erg3 Δ , erg5 Δ , erg6  Δ , Fig. 2 , inserts 2). In additions to changes in their steros,these erg  mutants have atered sphingoipid composition 8 . Thus,defects resuting from deetion of  RVS genes coud be compensated by  erg  mutants via changes in sphingoipids. Aso in ine with previouswork, both rvs161 Δ and rvs167  Δ show negative interactions with actin Correlation between interaction partners        F     r     e     q     u     e     n     c     y interaction score        F     r     e     q     u     e     n     c     y 0–5   –   1 .   0  –   0 .   8  –   0 .   6  –   0 .  4  –   0 .   20   0 .   2   0 .  4   0 .   6   0 .   8   1 .   0 –4–3–2–10123450.ckingOtherUnknownEisosomesCell wallSignalingLipid metabolismMembranetransportersVacuolarfunctionExocytosisGene expressionPeroxisomes cba Figure 1 Composition o the plasma membrane E-MAP.( a ) Genes selected or the plasma membrane E-MAP areclassiied according to their biological unction. ( b , c )Genes encoding proteins interacting with each other aremore likely to show positive genetic interactions ( b ) andcorrelated genetic interaction proiles ( c ). Green, interactionand correlation scores o gene pairs known to encodeinteracting proteins; black, the remainder o gene pairs. pairwise correation of genetic-interactionprofies to a high-quaity set of protein-proteininteractions (PPIs) 19 and found that the powerof the genetic map to predict PPIs is compar-abe to that of previousy pubished E-MAPs( Supplementary Fig. 1 ). Furthermore, com-parison of interaction scores or correationcoefficients of gene pairs encoding physicay interacting proteins 19–21 (see Supplementary Table 2 ) among a pasma-membraneE-MAP gene pairs reveaed that they have ahigher ikeihood to interact positivey and tohave correated genetic-interaction profies( Fig. 1b , c , yeow area under the green graph).Conversey, gene pairs with highy correatedinteraction profies and positive interactionsare ikey to physicay interact.To better visuaize groups of interactinggenes and their reationships, we used a previ-  nature structural & molecular biology   VOLUME 17 NUMBER 7 JULY 2010 903 resource cytoskeeton genes, such as BBC1 ,  JSN1 and BZZ1 (refs. 10,24–26) ( Fig. 2 ,insert 2a). In addition, we found severa previousy unrecognizedreationships, incuding negative interactions between the RVS genesand ire1 Δ and hac1 Δ  , two mediators of the unfoded protein response(UPR) contro system for endopasmic reticuum function. Possiby, cesreact to Rvs deficiency by atering ipid synthesis or transport, whichin turn activates the UPR. Ces acking the UPR in addition to the Rvsproteins coud have decreased fitness. Consistent with this notion, arecent genome-wide study found the UPR activated in rvs Δ ces 27 .We aso detected many genetic interactions and highy corre-ated profies between genes encoding actin-patch components. Forexampe, sla1 Δ and ede1 Δ , which function in endocytosis, are highy correated (correation = 0.64, Fig. 2 , insert 1) and show a nega-tive genetic interaction (interaction score = −7.7). Unexpectedy,given its function in exocytosis rather than endocytosis, we asofound chs6  Δ to be highy correated with sla1 Δ and ede1 Δ (correa-tions ede1 Δ - chs6  Δ = 0.53 and sla1 Δ - chs6  Δ = 0.43; Fig. 2 , insert 1).Furthermore, these three genes a resut in negative genetic interac-tions when any two of them are combined. Coectivey, this indicatesthat Chs6 might function in coordinating exo- and endocytosis,perhaps by deivering a subset of cargos to the pasma membrane 28 .In this scenario, chs6  Δ woud ead to the depetion of an endocyticfactor from the pasma membrane and, as a consequence, a decreasein endocytosis efficiency. Combination with mutants defective inunderstand eisosome function in vivo , we geneticay anayzed its corecomponents, PIL1 and LSP1 . As the encoded proteins are >70% identi-ca and are stoichiometric components of the eisosome, we expected very simiar genetic profies for them. Unexpectedy, PIL1 and LSP1  showed very different genetic interactions and, accordingy, custer indifferent regions of the E-MAP (correation = 0.038; Fig. 2 , insert 4).This paraes the ce-bioogica observation that deetion of  PIL1 butnot LSP1 resuts in strong effects on pasma-membrane organizationand protein turnover.To gain further insight into eisosome function, we anayzed the tripetgenetic motifs (TGMs) in which  pil1 Δ participates 17 . TGMs are thesimpest motifs apart from binary interactions and can exist in fourforms: type I (a three genes showing positive genetic interactions),type II (two positive and one negative), type III (two negative and onepositive) and type IV (three negative interactions) ( Fig. 3a ). We havepreviousy shown that genes with a positive genetic interactions (typeI TGM) are enriched for functioning in the same pathway  17 . We there-fore assembed a compete map of type I TGMs found in the pasma-membrane E-MAP ( Supplementary Fig. 4 ). Because Pi1 has a moreprominent roe than Lsp1 in eisosome and pasma-membrane function,we extracted a type I TGMs invoving  pil1 Δ ( Fig. 3b ). In this represen-tation, we highighted genes that are important for eisosome ocaizationor are cosey reated to such genes ( YMR031c and EMP70 , respec-tivey  31 ; green nodes in  Fig. 3b ) and characterized them further. 2c43b3a2d2b2a1 LSP1PIL1 2a2b42c2d    I   P   T   1   S   K   N   1   P   B   S   2   H   O   G   1   S   S   K   1   E   N   T   5   I   N   P   5   3 RVS161RVS167    L   C   B   4   F   E   N   1   S   U   R   4   S   U   R   2   L   C   B   3   R   O   M   2   P   K   H   3   F   K   S   1   L   E   M   3   S   C   S   2   D   N   F   1   A   R   L   1   S   Y   S   1   A   R   L   3   T   L   G   2   G   E   T   3   G   E   T   1   A   P   M   3   A   P   L   5   V   P   S   2   9   V   P   S   3   5   V   P   S   1   7   P   E   P   8   M   O   N   2   V   P   S   1   3   P   E   R   1   A   R   F   1   S   W   A   2   C   O   G   7   C   O   G   6   C   O   G   5   C   O   G   8   R   I   C   1 ARL1SYS1ARL3TLG2GET3GET1APM3APL5VPS29VPS35VPS17PEP8MON2VPS13PER1ARF1SWA2COG7COG6COG5COG8RIC1    A   B   P   1   G   I   M   4   G   G   A   1   T   W   F   1   Y   L   R   4   0   8   C   M   S   B   3   B   Z   Z   1   J   S   N   1   S   U   R   7   B   B   C   1   A   I   P   1   Y   G   L   0   7   9   W   H   A   C   1   I   R   E   1 ARF-related proteinsARF-related proteinsReceptor for Arl1 and Arl3Endosomal t-SNAREGet complexAP3 complexRetromere complexCOPI retrograde trafficking(Synthetic lethal with Arf1) AuxillinCOG complex–3–2–10123Actin patchActin patchActin patchActin cytoskeletonActin cytoskeleton    S   L   A   1   E   D   E   1   C   H   S   6   C   L   A   4   T   P   M   1   A   K   R   1   B   E   M   3   R   G   D   1   S   L   T   2   B   C   K   1   T   A   T   1   S   M   Y   1   C   A   P   2   B   N   R   1   C   A   P   1   A   R   C   1   8 SLA1EDE1CHS6CLA4TPM1AKR1BEM3RGD1SLT2BCK1TAT1SMY1CAP2BNR1CAP1ARC18    L   A   G   1   T   S   C   1   0   D   P   L   1   L   A   C   1   L   C   B   4   F   E   N   1   S   U   R   4   S   U   R   2   L   C   B   3 CSG2SCS7    E   R   G   3   E   M   P   7   0   E   R   G   6   E   R   G   5   L   A   G   1   T   S   C   1   0   D   P   L   1   L   A   C   1   L   C   B   4   F   E   N   1   S   U   R   4   S   U   R   2   L   C   B   3   R   O   M   2 LAG1TSC10DPL1LAC1LCB4FEN1SUR4SUR2LCB3ROM2    R   O   M   2 Negativeinteractions(syntheticallysick/lethal)Positiveinteractions(suppressive/ epistatic) MKS1    R   T   G   1 –1.0–0.8–0.6–0.4– –1.0 –0.8 –0.6 –0.4 –0.2 0 0.2 0.4 0.6 0.8 1.0 Figure 2 Overview o the clustergram o theplasma membrane E-MAP. Top, selected areasare marked in the overview and highlightedas inserts 1–4. Yellow, positive geneticinteractions; blue, negative genetic interactions.Bottom, genes with correlating genetic proilesare shared between RTG1 and MKS1. Pairwisecorrelations between RTG1 and MKS1 and allother genes in the plasma membrane E-MAPwere calculated and plotted against each other. this process woud further decrease thefitness of the resuting strains.We aso observed many strong geneticinteractions between trafficking compexes.Genes encoding the retromer compex( VPS17  , VPS29 , VPS35 , PEP8 ), the COGcompex ( COG5 , COG6  , COG7  , COG8 ) orthe AP3 compex (  APM3 ,  APL5 ) a formedhighy correated custers in the pasma- membrane E-MAP ( Fig. 2 , insert 4). In addi-tion, potentia new connections between thesecompexes and heretofore poory character-ized components of the endocytic machinery are apparent in these custers. As an exampe,the retromer compex cocusters with deetionof   MON2 (correation = 0.48), a gene encod-ing an evoutionariy conserved scaffodingprotein functioning in endosome-to-Gogitrafficking 29 . Our data suggest that Mon2 actstogether with the retromer in this process.Many genes encoding members of signaingcascades showed strong genetic reationships.For exampe, two kinases of the ce integrity MAP kinase signaing modue, St2 (the MAPkinase) and Bck1 (the MAP kinase kinasekinase) 30 , showed one of the highest correations(0.75). Simiary, genes encoding components of retrograde signaing ( RTG1 , RTG2 , RTG3 and  MKS1 ) a custer together (correation = 0.44)indicating that a pairs have high correationcoefficients (for exampe,  MKS1 / RTG1 correa-tion coefficient = 0.59; Fig. 2 , bottom). Functional links involving eisosomes Athough the eisosome has been inked toendocytosis reguation, detais regardingits bioogica roes remain unresoved. To  904 VOLUME 17 NUMBER 7 JULY 2010 nature structural & molecular biology resource EIS1 / YMR031c  encodes a novel eisosome component Because  ymr031c Δ and  pil1 Δ have a positive genetic interactionand a correated interaction profie ( Fig. 3b ), we tested whether thecorresponding proteins physicay associate. To this end, we fused thesequence encoding the green fuorescent protein (GFP) tag to PIL1  at its endogenous ocation in the yeast genome and immune-purifiedthe expressed Pi1-GFP from a yeast cuture that was metaboicay abeed with heavy, nonradioactive ysine (SILAC) 32 . In parae, weperformed a mock purification from contro, ight-abeed wid-typeces. We identified 533 proteins present over a 10,000-fod dynamicrange in the mixed euates from both purifications. As expected, wefound Pi1 and Lsp1 as we as the recenty identified eisosomes bind-ing protein Mrp8 to be significant outiers, with a high ratio of abeedto nonabeed protein, indicating that they are specific interactors 2,33  ( P  < 0.0001;  Fig. 4a ). In addition, we found a number of other specificinteractors, incuding Ymr031c, which is consistent with a recentreport 34 . To independenty confirm this observation, we performedimmunoprecipitations of TAP-tagged Ymr031c and, as a contro, Lsp1,and we found that both specificay precipitated Pi1 ( Fig. 4b ). To testwhether Ymr031c coocaizes with Pi1, we fuorescenty tagged bothproteins. The signa from Pi1 and Ymr031c perfecty overapped ateisosomes ( Fig. 4c , upper pane; Pearson correation = 0.81 ± 0.06).Consistent with these data, Ymr031c was recenty detected at MCCs 3 .One prediction for a genuine eisosome component is that it reoca-izes to eisosome remnants in a PIL1 deetion strain 2 . We thereforeinvestigated Ymr031c-GFP ocaization in  pil1 Δ ces and found thatboth Ymr031c and the eisosome component Lsp1 ocaized to one ora few eisosome remnants in the ce periphery ( Fig. 4d ). To investigatewhether YMR031c has a roe in eisosome architecture or assemby, wedeeted it and anayzed the ocaization of eisosome core componentsin the resuting strain. For both Pi1 and Lsp1-GFP, we observed sub-stantiay increased cytosoic fuorescence in  ymr031c Δ ces ( Fig. 4e , f  ).Coectivey, these data show that Ymr031c is physicay associatedwith eisosomes and is required for their norma formation. We havetherefore named this gene EIS1 . EMP70 is an early endosomal and vacuolar protein In the genetic network of the pasma-membrane E-MAP, EMP70 is thestrongest candidate for a functiona reationship with PIL1 because (i)the two genes have highy correated genetic profies (correation of  PIL1  and EMP70 = 0.37 ( EMP70 has the most simiar profie to PIL1 of a theE-MAP genes);  Fig. 5a ); (ii) the two genes participate in two type I TGMs( Fig. 3b ); and (iii) the Emp70 homoog Tmn2 is required for norma Pi1-GFP ocaization 31 . In addition, our moduar anaysis identified EMP70  and PIL1 as part of the same six-gene modue ( Supplementary Fig. 3 ;S-score between PIL1 and EMP70 = 1.78; Supplementary Table 4 ). b EMP70 ISC1RIC1ERS1WSC4 YMR031C CHS1ARL1 RVS161 PER1PIL1SKM1 SPS22  ARN1STV1 +++Type I–Type IVType II++–Type III+–––– a Negative genetic interactionPositive genetic interaction MergeDistance along plasma membrane    I  n   t  e  n  s   i   t  y   (   A   U   ) Lsp1-GFP    W   T      p       i       l       1       ∆ Pil1-GFPPil1-GFP  ymr031c ∆    E   i  s  o  s  o  m  e  n  u  m   b  e  r   E   i  s  o  s  o  m  e   G   F   P   f   l  u  o  r  e  s  c  e  n  c  e   (   A   U   ) 0102030405060012345600.    C  y   t  o  s  o   l   i  c   G   F   P   f   l  u  o  r  e  s  c  e  n  c  e   (   A   U   )    W   T   y  m  r  0  3  1  c   ∆    W   T   y  m  r  0  3  1  c   ∆    W   T   y  m  r  0  3  1  c   ∆    W   T   y  m  r  0  3  1  c   ∆    W   T   y  m  r  0  3  1  c   ∆    W   T   y  m  r  0  3  1  c   ∆ ba dce Pil1    Y  m  r   0   3   1  c  -   T   A   P   i  n  p  u   t   L  s  p   1  -   T   A   P   i  n  p  u   t   C  o  n   t  r  o   l   i  n  p  u   t   Y  m  r   0   3   1  c  -   T   A   P  e   l  u  a   t  e   L  s  p   1  -   T   A   P  e   l  u  a   t  e   C  o  n   t  r  o   l  e   l  u  a   t  e log 2 (ratio heavy/light) Pil1Ygr130cYkl105cMrp8Ymr031cYmr086wLsp1Pkh1 Lsp1-GFP00.    E   i  s  o  s  o  m  e  n  u  m   b  e  r   E   i  s  o  s  o  m  e   G   F   P   f   l  u  o  r  e  s  c  e  n  c  e   (   A   U   )   C  y   t  o  s  o   l   i  c   G   F   P   f   l  u  o  r  e  s  c  e  n  c  e   (   A   U   )    l  o  g    2    (   i  n   t  e  n  s   i   t  y   ) f 302520–4 –2 0 2 4Ymr031c-GFP Pil1-RFPmarsYmr031c-GFPLsp1-GFP  ymr031c ∆ in  ymr031c  Δ or control cells. Representative midsections are shown. For each experiment, the number o eisosomes per cell, the GFP luorescence pereisosome and the cytosolic GFP luorescence were quantiied rom at least 100 cells and are shown below the images. Scale bars, 2.5 μ m. Figure 4   YMR031C   / EIS1 encodes an eisosome component. ( a ) Ainitypuriication and MS analysis o heavy labeled cells expressing GFP-tagged Pil1and untagged control cells. Averaged peptide intensities are plotted againstheavy/light SILAC ratios. Signiicant outliers ( P  < 0.0001) are colored in orangeor light blue ( P  < 0.05); other identiied proteins are shown in dark blue.( b ) Pulldown puriication rom cells expressing tandem ainity-tagged Lsp1,Ymr031c or untagged control cells. Inputs and eluates rom the pulldown wereblotted and probed with antibodies against Pil1. ( c ) Colocalization o GFP-taggedYmr031c with RFPmars-tagged Pil1. Representative conocal midsectionsare shown. The graph shows the intensity proiles or both channels along theperimeter o the cell. ( d ) PIL1 is required or normal localization o Ymr031c.Ymr031c-GFP or Lsp1-GFP was expressed and imaged either in WT or pil1 Δ  cells. Representative conocal midsections are shown. ( e , f ) Ymr031c is requiredor normal eisosome ormation. Pil1-GFP ( e ) or Lsp1-GFP ( f ) was expressed Figure 3 TGMs o the plasma membrane E-MAP. ( a ) All our potential TGMsare shown. Nodes in vertical order represent involvement in the same pathway;horizontal orientation indicates possible parallel pathways. ( b ) Type I TGMsthat have PIL1 as a node. Nodes in green represent a gene important or Pil1-GFP localization ( YMR031C  ) or a homolog o such a gene ( EMP70  ) 31 .  nature structural & molecular biology   VOLUME 17 NUMBER 7 JULY 2010 905 resource These genetic inks prompted us to investigate EMP70 in more detai( Fig. 5 ). We fuorescenty tagged Emp70 with GFP and found that itocaizes in a compex pattern consisting of a centra ring reminis-cent of vacuoes and severa bright foci in the cytopasm that oftenseem connected to the vacuoe ( Fig. 5b and Supplementary Video 1 ).Emp70 was previousy found in an endosoma membrane fraction 35 .We therefore tested whether cytosoic Emp70 foci represent endosomes.We used a number of endosoma markers and found Emp70-GFP focito coocaize with Kex2, marking the eary endosome, which in yeastis functionay continuous with the trans-Gogi network. In contrast,Emp70 ocaization did not overap with the ate endosoma/prevacuoarmarker Vps5 ( Fig. 5b and Supplementary Fig. 5a ).To test whether the Emp70-abeed compartments are part of theendocytic route, we used the endocytosis tracer FM4-64. This ipiddye is incorporated in the pasma membrane, taken up by endocytosisand trafficked through the endosoma system to the vacuoe 36 . Wefound in puse-chase experiments that eary FM4-64 intermediatescoocaize with Emp70 foci ( Fig. 5c , 0 min). As the dye migratedthrough the endocytic system, it aso coocaized with a subset of Emp70-positive foci toward the end of the reaction but markedy essat intermediate time points ( Fig. 5c , 30 min). At the fina time point,FM4-64 ceary abeed the vacuoe deimiting membrane where itcoocaized with the Emp70-GFP ring staining. Trafficking from eary endosomes can be bocked by incubation of ces at 16 °C, which eadsto the accumuation of FM4-64 (ref. 37). Emp70-GFP amost perfecty coocaized with FM4-64 when the atter was accumuated in such a‘16 °C compartment’, further arguing that Emp70 ocaizes to eary endosomes ( Supplementary Fig. 5b ). Strains harboring a deetedor C-terminay tagged SNF7  (an ESCRT-III gene) show a ‘cass E’ vacuoar protein sorting defect. This is characterized by coapse of endosomes to one or a few arge cass E compartments 38,39 . Underthese conditions, Emp70-GFP formed fewer, very arge custers thatcoocaized with Snf7-RFPmars marked cass E compartments andshowed reduced vacuoar membrane staining ( Fig. 5d ). From thesedata, we concude that Emp70 ocaizes to eary endosomes and the vacuoe. To better characterize the ocaization of Emp70 in thesetwo poos, we quantitated the reative amount of Emp70 coocaizingwith markers for each organee and found 48% of Emp70 to oca-ize in the TGN/endosoma compartment and 41% at the vacuoarmembrane ( Fig. 5g  ). Figure 5 The eisosome-linked Emp70 isan early endosomal protein. ( a ) Genes withcorrelating genetic proiles are shared between PIL1 and EMP70  but not PIL1 and LSP1 .Correlation coeicients between the geneticproile o PIL1 and each o the other 373proiles in the E-MAP are plotted on the x  axisagainst, on the  y  axis, either the similar seto values or the LSP1 proile with all otherproiles (blue) or those or EMP70  with all otherproiles (red). Labeled points indicate somegenes with proiles that are positively correlatedwith both the proile o PIL1 and that o EMP70  . CC values in blue and red indicate thecorrelation coeicients or the ull set o blueor red points plotted. ( b ) Emp70 colocalizeswith Kex2. Emp70-GFP and Kex2-RFPmarswere coexpressed and imaged. Representativeconocal midsections are shown. ( c ) Emp70localizes to an FM4-64 marked endocyticcompartment. Cells expressing Emp70-GFP(green) were pulse labeled with FM4-64 (red)and imaged or 1 h. Images o midsections ocells at selected time are shown as indicated.( d ) Emp70 localizes to the class E compartmentin SNF7  mutants. GFP-tagged Emp70 wasexpressed in cells harboring nonunctionalSn7-RFPmars, resulting in the clustering oendosomal proteins in the class E compartment.Representative conocal midsections areshown. ( e ) Emp70-GFP oci localize to the cellperiphery. Emp70-GFP (green) was expressedin cells harboring the luorescent eisosomesmarker Lsp1-MARS. Representative mid- (let)and top sections (right) are shown. Boxeshighlight selected areas o colocalization.( f ) PIL1 is required or normal Emp70 localizationto the cell periphery. Emp70-GFP was expressedin cells expressing the plasma membrane markerYlr413w-RFPmars, and oci overlaying thismarker were counted in more than 100 WT and pil1 Δ cells. Results are shown as a histogram onumber o spots opposed to the plasma membrane in each cell. ( g ) Quantitation o the organelle distribution o Emp70. Emp70-GFP was imaged in live cellsand analyzed or colocalization with Kex2-RFPmars ( n = 100), vacuolar FM4-64 ( n = 91), Sn7-RFPmars ( n = 93, diploid strain expressing one tagged Sn7allele) and Lsp1-Cherry ( n = 107). The relative area o overlap between signals was quantiied as a percentage o total area occupied by Emp70 signal. Boxplots representing maxima, 75th percentile, median, 25th percentile and minima are shown or the colocalization with each marker. Scale bars, 2.5 μ m. a CC = 0.038CC = 0.471PIL1–1.0    L   S   P   1   (   b   l  u  e   )   E   M   P   7   0   (  r  e   d   ) –1.0–0.8–0.6–0.4– LSP1PIL1PIL1ERG6COS10EMP70ERG3SUR2INP51YSR3Emp70-GFP b Emp70-GFP d    F   M   4  -   6   4   M  e  r  g  e 0 15 3060Time (min) ce WT pil1 ∆    P  e  r  c  e  n   t  a  g  e  o   f  c  e   l   l  s No. spots near the plasma membrane0 2 4 6 8 10 0 2 4 6 8 100515102025303545405005151020253035454050 f    E  m  p   7   0  -   G   F   P   L  s  p   1  -   R   F   P  m  a  r  s   M  e  r  g  e Midsection020406080100    C  o   l  o  c  a   l   i  z  a   t   i  o  n  w   i   t   h   E  m  p   7   0   (   %   )    K  e  x  2   F   M  4  -  6  4  S  n  f   7    L  s  p  1  g –0.8 –0.6 –0.4 –0.2 0 0.2 0.4 0.6 0.8 1.0Kex2-RFPmars Merge    E  m  p   7   0 Snf7-RFPmars MergeTop section
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks