TFG-1 function in protein secretion and oncogenesis et al., 2011.pdf · 2015-01-15 · ARTICLES TFG-1 function in protein secretion and oncogenesis KristenWitte1,4,AmberL.Schuh1,4,JanHegermann2,AliSarkeshik3,JonathanR.Mayers1,KatrinSchwarze2,

ART I C L E S

TFG-1 function in protein secretion and oncogenesis

Kristen Witte1,4, Amber L. Schuh1,4, Jan Hegermann2, Ali Sarkeshik3, Jonathan R. Mayers1, Katrin Schwarze2,John R. Yates III3, Stefan Eimer2 and Anjon Audhya1,5

Export of proteins from the endoplasmic reticulum in COPII-coated vesicles occurs at defined sites that contain the scaffoldingprotein Sec16. We identify TFG-1, a new conserved regulator of protein secretion that interacts directly with SEC-16 and controlsthe export of cargoes from the endoplasmic reticulum in Caenorhabditis elegans. Hydrodynamic studies indicate that TFG-1 formshexamers that facilitate the co-assembly of SEC-16 with COPII subunits. Consistent with these findings, TFG-1 depletion leads toa marked decline in both SEC-16 and COPII levels at endoplasmic reticulum exit sites. The sequence encoding the aminoterminus of human TFG has been previously identified in chromosome translocation events involving two protein kinases, whichcreated a pair of oncogenes. We propose that fusion of these kinases to TFG relocalizes their activities to endoplasmic reticulumexit sites, where they prematurely phosphorylate substrates during endoplasmic reticulum export. Our findings provide amechanism by which translocations involving TFG can result in cellular transformation and oncogenesis.

The trafficking of most secretory cargoes begins with their export fromthe endoplasmic reticulum1,2. In metazoans, cargoes are packagedinto vesicles that emerge at defined sites on the endoplasmicreticulum and ultimately fuse with the endoplasmic reticulum–Golgiintermediate compartment (ERGIC). This process relies on theefficient recruitment of a set of soluble factors known as the COPIIcoat, which is composed of two multimeric protein complexes(Sec23–Sec24 and Sec13–Sec31; refs 3,4). A number of rare diseaseshave been directly linked to perturbations in COPII function, includingcraniolenticulosutural dysplasia, in which amutation in Sec23A leads toimpaired collagen export from the endoplasmic reticulum5,6. AlthoughCOPII vesicle formation has been reconstituted with purified coatproteins on synthetic membranes7, regulators of COPII assemblyremain largely unexplored.Of the known COPII-interacting proteins, the small GTPase Sar1

and the putative scaffolding protein Sec16 are among the bestcharacterized8. When activated by the Sec12 exchange factor, Sar1recruits the Sec23–Sec24 complex, which forms an adaptor layer forSec13–Sec31 lattice assembly, completing the COPII coat9,10. Sec23also functions as a Sar1-activating protein, which is stimulated bySec31, leading to coat disassembly following vesicle budding11. Incontrast to Sar1, which has been shown to bind only Sec23 directly,Sec16 interacts with all COPII coat components, potentially servingas a scaffold for their recruitment12–15. Furthermore, Sec16 maystabilize the COPII coat to prevent premature disassembly followingactivation of the Sec23 GTPase-activating protein16. Surprisingly,

1Department of Biomolecular Chemistry, University of Wisconsin–Madison Medical School, 1300 University Avenue, Madison, Wisconsin 53706, USA. 2EuropeanNeuroscience Institute and Center for Molecular Physiology of the Brain (CMPB), 37077 Goettingen, Germany. 3Department of Cell Biology, The Scripps ResearchInstitute, La Jolla, California 92037, USA. 4These authors contributed equally to this work.5Correspondence should be addressed to A.A. (e-mail: [email protected])

Received 23 August 2010; accepted 7 February 2011; published online 10 April 2011; DOI: 10.1038/ncb2225

few other proteins have been implicated in the regulation of COPIIrecruitment. Considering the necessity for controlling secretory fluxduring cell differentiation and development, additional factors thatgovern this process probably exist, and some may function throughdirect regulation of Sar1 or Sec16. Here, we report the identificationof a new regulator of endoplasmic reticulum export, which interactsdirectly with Sec16 and controls COPII subunit accumulation atendoplasmic reticulum exit sites.

RESULTSIdentification of SEC-16-interacting proteins in C. elegansOn the basis of the sequence of human Sec16A, we identifieda Sec16 homologue in C. elegans encoded by ZK512.5, whichwe named sec-16 (Fig. 1a). Antibodies directed against SEC-16recognized a band at the appropriate relative molecular massof 135,000 (∼Mr 135K) following immunoblot analysis of aC. elegans lysate and were used to immunoprecipitate SEC-16 and associated proteins from detergent-solubilized embryoextracts (Supplementary Fig. S1a–c). We conducted solution massspectrometry to identify 19 proteins that were reproducibly isolatedby SEC-16 antibodies, which were individually depleted usingRNA interference (RNAi). We found that 7 out of 19 geneproducts were necessary for embryo production: SEC-16, fourCOPII subunits (SEC-23, SEC-24.2, SEC-13 and SEC-31), the Sar1-specific guanine nucleotide exchange factor SEC-12 and TFG-1(Supplementary Table S1).

550 NATURE CELL BIOLOGY VOLUME 13 | NUMBER 5 | MAY 2011

© 2011 Macmillan Publishers Limited. All rights reserved.

ART I C L E S

Total (5%)

cMock IP (100%)

TFG-1 IP (100%)

SEC-16

Total (5%)

Mock IP (100%)

SEC-16 IP (100%)

Tfg-1

a

2,357Hs Sec16A

1,060Hs Sec16B

1,232Ce SEC-16

CCD

CCD

CCD

TFG-1

TFG-1(C)

TFG-1(N)

75Mr (K)

50

25

SEC-16

TFG

-1

TFG

-1(C

)

TFG

-1(N

)

37

SEC-16

GST

GST

–SEC

-16

GST

TFG-1

75

50

25

37

20

100150

Mr (K)

b

d e

Figure 1 C. elegans TFG-1 interacts with the endoplasmic reticulum exit-sitecomponent SEC-16. (a) Schematic representation of human (Hs) andC. elegans (Ce) Sec16 isoforms. The central conserved domain (CCD) ishighlighted in each protein. (b) SEC-16 was immunoprecipitated from C.elegans embryo extract and blotted with anti-TFG-1 antibodies (n = 3).A mock immunoprecipitation (IP) assay was conducted in parallel usingrabbit immunoglobulin G. (c) TFG-1 was immunoprecipitated from C.elegans embryo extract and blotted with anti-SEC-16 antibodies (n = 3).A mock immunoprecipitation assay was conducted in parallel using rabbitimmunoglobulin G. (d) GST alone and GST-tagged full-length SEC-16 wereimmobilized on glutathione agarose beads, which were incubated with anextract generated from Escherichia coli expressing recombinant TFG-1.Following a series of washes, proteins were eluted using reduced glutathione,separated by SDS–PAGE and either Coomassie blue stained (top) orimmunoblotted using TFG-1 antibodies (bottom). (e) Polyhistidine-taggedfull-length and truncated forms of TFG-1, either encoding amino acids1–195 (TFG-1(N)) or 196–486 (TFG-1(C)), were purified from E. coli ontonickel affinity resin and incubated with freshly prepared whole-worm extract(n=3). Imidazole-eluted proteins were separated by SDS–PAGE, Coomassieblue stained (top) and blotted with anti-SEC-16 antibodies (bottom).Uncropped scans of gels and immunoblots are provided in SupplementaryFig. S6.

We focused our attention on the interaction between SEC-16 andTFG-1. Immunoblot analysis of a C. elegans lysate using TFG-1antibodies revealed at least two closely migrating bands immediatelybelow the Mr 75K marker, significantly larger than the predictedsize of Mr 49.8K (Supplementary Fig. S1d). Similarly, recombinantTFG-1 also exhibited a slow migration on SDS–polyacrylamidegel electrophoresis (SDS–PAGE; see Fig. 1e). Analysis of a SEC-16immunoprecipitate using TFG-1 antibodies confirmed their interaction(Fig. 1b), and TFG-1 immunoprecipitation led to an enrichmentof SEC-16 (Fig. 1c). We also generated recombinant forms of both

TFG-1 and SEC-16 and measured their association in vitro. Wefound that a glutathione S-transferase (GST) fusion to SEC-16,but not GST alone, was capable of binding to recombinant TFG-1(Fig. 1d). In addition, we generated polyhistidine-tagged recombinantforms of full-length TFG-1 and two truncations of TFG-1 encodingthe amino terminus (amino acids 1–195) and carboxy terminus(amino acids 196–486), and measured their ability to interactwith endogenous SEC-16. In a similar fashion to the full-lengthrecombinant protein, the N-terminal portion of TFG-1 was ableto bind to SEC-16, whereas the C terminus did not (Fig. 1e).These data confirm a direct biochemical association between SEC-16 and TFG-1.

TFG-1 localizes to endoplasmic reticulum exit sites with SEC-16To determine whether SEC-16 interacts with TFG-1 at endoplasmicreticulum exit sites, we examined the C. elegans reproductive system,which is enriched in TFG-1 (Supplementary Fig. S1e). TFG-1 antibodiesstained punctate structures throughout the cytoplasm of oocytes,85% of which co-localized with SEC-16 on endoplasmic reticulumexit sites (Fig. 2a). In many cases, TFG-1 staining extended beyondthe puncta labelled with SEC-16 (Fig. 2a, see ×6 magnification).Using immunogold electron microscopy, we further defined thelocalization of TFG-1 in proximal oocytes to a cloud-like regionat endoplasmic reticulum exit sites that spread to the ERGIC(Fig. 2b). The high concentration of labelling observed indicatedthat TFG-1 may form a matrix in this region, which wouldcorrespond well to the elevated electron density seen there byelectron microscopy. Dual immunogold labelling with TFG-1 andSEC-13 antibodies indicated that both proteins localize to an identicalarea next to endoplasmic reticulum exit sites, although the SEC-13labelling was more discrete (Fig. 2c). We conclude that TFG-1localizes to endoplasmic reticulum exit sites with both SEC-16and COPII machinery.The architecture of endoplasmic reticulum exit sites was also

examined using electron microscopy. We found that endoplasmicreticulum exit sites range in size from 70 to 150 nm, and alwayscontain coated buds emerging from smooth endoplasmic reticulumthat are directed towards ERGIC and Golgi membranes (Fig. 2dand Supplementary Fig. S1f,g and Video S1). Electron tomographywas used to confirm these findings and provide a high-resolutiondepiction of COPII vesicle biogenesis (Fig. 3a and SupplementaryVideo S2). Consistent with these observations and our data showingthat TFG-1 localizes specifically to endoplasmic reticulum exit sites,we failed to observe co-localization between TFG-1 and the glucuronyltransferase SQV-8, a marker of the cis/medial Golgi (SupplementaryFig. S1h). Instead, TFG-1 appeared directly adjacent to SQV-8-labelled structures, further illustrating that in C. elegans endoplasmicreticulum exit sites are closely juxtaposed to the Golgi, forming anintegrated secretion unit.We next examined the distal region of the C. elegans germline, which

constitutes a stem-cell niche that constitutively requires membranebiosynthesis and secretion17. Endoplasmic reticulum exit sites in thisregion were morphologically similar to those seen in the proximalgonad (Fig. 3b). Immunofluorescence analysis of exit sites in the distalgermline revealed a distribution of intensities for both SEC-16 andTFG-1 (Fig. 3c,d). However, the ratio of the fluorescence intensities

NATURE CELL BIOLOGY VOLUME 13 | NUMBER 5 | MAY 2011 551


ART I C L E S

TFG-1 (10 nm gold)b

TFG-1 (15 nm)/SEC-13 (5 nm)c

Pro

xim

al o

ocyt

ea

Proximal oocytes

Distal gonad

SEC-16/TFG-1TFG-1SEC-16 (endoplasmic reticulum exit site)

SEC-16/TFG-1 (×6 magnification)

Endoplasmic reticulum exit sites

d

Figure 2 TFG-1 localizes to endoplasmic reticulum exit sites that arejuxtaposed to the Golgi. (a) Dissected C. elegans gonads were fixed andstained using Cy2-labelled anti-TFG-1 and Cy3-labelled anti-SEC-16antibodies (n = 8). Both individual and merged images of proximaloocytes with TFG-1 in green and SEC-16 in red are shown (scalebar, 10 µm). The upper right image is the area outlined in the panelbelow magnified sixfold (scale bar, 2 µm). Also shown is a schematicrepresentation of the C. elegans reproductive system, which includes asyncytial stem-cell niche in the distal gonad (light-green highlighted area)and proximal oocytes that have undergone cellularization (dark-greenhighlighted area). (b,c) Lowicryl sections of C. elegans oocytes werestained with antibodies against TFG-1 or a combination of TFG-1 andSEC-13 antibodies. Arrows highlight Golgi cisternae. Large arrowheadspoint out 15nm gold particles associated with immunoreactive TFG-1,and small arrowheads highlight 5 nm gold particles associated withSEC-13. Scale bars, 100nm. An inset is provided in c to clearly showthe distribution of 5 nm particles at higher magnification (inset scale

bar, 15 nm). In addition, a three-dimensional reconstruction of TFG-1immunolocalization is shown. The image was generated using the softwareReconstruct from serial 50 nm thin sections. Vesicles were reconstructedusing the sphere setting, and all other components (endoplasmic reticulum,ERGIC, coats, Golgi stacks) were generated using the Boissonnat surfacesetting. Light grey, endoplasmic reticulum; dark grey, COPII coat; orange,endoplasmic-reticulum-derived transport vesicles and ERGIC; red, greenand blue, Golgi cisternae, from cis to trans, respectively. (d) An electronmicrograph illustrating two endoplasmic reticulum exit sites and adjacentGolgi complexes in the most proximal oocyte of an animal followinghigh-pressure freezing and freeze substitution (scale bar, 500nm). Onthe right is a three-dimensional reconstruction of the same pair ofGolgi complexes and associated endoplasmic reticulum exit sites. Theendoplasmic reticulum exit sites are surrounded by vesicles that fuse toform the ERGIC. Light grey, endoplasmic reticulum; dark grey, COPII coat;orange, endoplasmic-reticulum-derived transport vesicles and ERGIC;yellow, red and blue, Golgi cisternae, from cis to trans, respectively.

between SEC-16 and TFG-1 largely did not vary among differentendoplasmic reticulum exit sites, indicating that in wild-type animalsthe levels of SEC-16 correlated with those of TFG-1, regardless ofwhether they were high or low (Supplementary Fig. S2a).To determine localization dependencies between TFG-1 and

SEC-16, we depleted each protein using RNAi. Depletion ofTFG-1 did not result in a significant decline in the number ofSEC-16-labelled endoplasmic reticulum exit sites (SupplementaryFig. S2b). However, the distribution of SEC-16 intensities wasmarkedly altered under these conditions (Fig. 3d). The number of

exit sites harbouring a high level of SEC-16 declined significantly(approximately 30-fold), whereas the number of low-intensitysites increased approximately twofold. As the stability of SEC-16 was not diminished following depletion of TFG-1 (Fig. 3e),these data indicate that TFG-1 controls the abundance of SEC-16at endoplasmic reticulum exit sites. Depletion of SEC-16 led tonearly complete loss of puncta labelled by TFG-1, and the fewstructures that were observed seemed abnormally large and didnot co-stain for the COPII component SEC-13 (Fig. 3c and ourunpublished data). We conclude that TFG-1 requires SEC-16 for its



ART I C L E S

SEC-16/TFG-1TFG-1SEC-16

tfg-1(RNAi) sec-16 (RNAi)

Dis

tal g

onad

Control Control

Distal gonadProximal gonada

c

Low-intensity Endoplasmic

reticulum exit sites

Medium-intensity Endoplasmic


Per

cent

age

of e

ndop

lasm

ic

retic

ulum

exi

t si

tes

High-intensity Endoplasmic


60

20

40

80 Control

tfg-1(RNAi)

∗∗

∗∗

∗∗

d

e

b

Control

tfg-

1(R

NA

i)

100%50% 25% 10% 100%

SEC-16

TFG-1

CAR-1

Figure 3 TFG-1 regulates SEC-16 levels on endoplasmic reticulum exitsites. (a) In the proximal gonad, a 300nm section of the early secretorypathway (endoplasmic reticulum exit sites, ERGIC and Golgi) was analysedby electron tomography. Endoplasmic reticulum exit sites are highlightedby arrowheads. On the left are individual sections from the tomographicstack. On the right are two orthogonal views of the tomogram followingthree-dimensional reconstruction. Light grey, endoplasmic reticulum; black,COPII coat; orange and yellow, endoplasmic-reticulum-derived transportvesicles and ERGIC; green, red, blue, Golgi cisternae; diffuse grey, notfurther resolvable matrix. Scale bar, 100nm. (b) An electron micrographillustrating endoplasmic reticulum exit sites and adjacent Golgi complexesin the distal gonad following high-pressure freezing and freeze substitution.An arrowhead highlights the presence of a budding vesicle from smoothendoplasmic reticulum (scale bar, 100nm). Below is a three-dimensionalreconstruction of the same Golgi complexes and associated endoplasmicreticulum exit sites. (c) Dissected gonads from control, TFG-1-depletedand SEC-16-depleted animals were fixed and stained using Cy2-labelled

anti-TFG-1 and Cy3-labelled anti-SEC-16 antibodies. Individual and mergedimages of the distal gonad with TFG-1 in green and SEC-16 in red areshown (scale bar, 10 µm). (d) The fluorescence intensity of SEC-16 inthe distal gonad was measured in control and TFG-1-depleted animals,and intensities were segregated into low, medium and high thresholds. Toestablish individual thresholds, a histogram of fluorescence intensities wasequally divided into three regions, and the number of endoplasmic reticulumexit sites within each area was calculated. The histogram indicates thepercentages of all endoplasmic reticulum exit sites that fall within specificthresholds. For each condition, at least 1,000 unique endoplasmic reticulumexit sites were examined. Error bars represent mean±s.e.m.; 10 differentanimals. ∗∗P <0.01 when compared with control, calculated using a pairedStudent t -test. (e) Western blots of extracts prepared from animals depletedof TFG-1 by RNAi (n=3). Serial dilutions of extracts prepared from controlanimals were loaded to quantify depletion levels. Blotting with anti-CAR-1antibodies was carried out to control loading. Uncropped scans of theimmunoblots are provided in Supplementary Fig. S6.

localization, whereas SEC-16 requires TFG-1 to accumulate normallyon endoplasmic reticulum exit sites.

TFG-1 self-associates to form hexamers that regulate properSEC-16 complex assemblyOn the basis of size-exclusion chromatography, native SEC-16 andTFG-1 exhibit similar elution profiles, which correspond to globularcomplexes larger thanMr 800K (Fig. 4a). As TFG-1 was amenable to re-combinant expression, wemeasured its Stokes radius (107.8 Å) and sed-imentation value (7.0 S) in the absence of SEC-16 (Fig. 4b). On the basisof these data, we calculated the native∼Mr of TFG-1 to be 318K, nearly

identical to the predicted mass of a TFG-1 hexamer. Furthermore, wefound that the N terminus assembled into an octamer, whereas theC terminus formed a dimer (Fig. 4c,d). These data indicate that theN terminus of TFG-1mediates its oligomerization, whereas the C termi-nusmay restrict the full-length protein to formhexamers in solution. Aslevels of TFG-1 and SEC-16 at endoplasmic reticulum exit sites co-varyin control animals and loss of TFG-1 leads to diminished accumulationof SEC-16 on endoplasmic reticulum exit sites, such hexamers of TFG-1probably play a role in proper SEC-16 complex assembly. Consistentwith this idea, depletion of TFG-1 caused a reduction in the Stokesradius of SEC-16 (by ∼70Å) isolated from whole animals (Fig. 4e).



ART I C L E S

SEC-16

TFG-1

Superose 6 gel filtration

1098 11 12 13765

Stokes radius = 125.3 Å



TFG-1: 6xHis


1098 11Glycerol gradient (10–30%)

7621 3 4 5Load

MW

S value = 7.0 S

TFG-1: 6xHis

Load

MW

1098 11 12 1376100

Mr (K)

75

75100

Mr (K)

b

131211Glycerol gradient (10-30%)

10965 7 8Load

MW

S value = 6.7 S



Load

MW

TFG-1(N): 6xHis

109 11 12 13 14 15

25

37Mr (K)

TFG-1(N): 6xHis25

37Mr (K)

1098 11 12 13Superose 6 gel filtration

765

tfg-1(RNAi); Stokes radius = 55.8 Å

HGRS-1

HGRS-1

Wild type

tfg-1 (RNAi)


14

11109 121314Superose 6 gel filtration876 15161718

tfg-1(RNAi); Stokes radius = 24.2 Å

Control; Stokes radii = 128.1, 54.3, 23.9 Å

11109 121314Superose 6 gel filtration876 15161718



SEC-16

SEC-13

SEC-13

e


12Load

MW

14 15

37

50Mr (K)

TFG-1(C): 6xHis

Glycerol gradient (10-30%)

621 3 4 5MW

37

50Mr (K)

S value = 3.1 S


TFG-1(C): 6xHis

d

a

c

Figure 4 The N terminus of TFG-1 mediates its oligomerization. Theresults presented in each panel are representative of at least threeindividual experiments. In all cases, the intensities of each band weremeasured to identify the peak elution fraction, which was used tocalculate either a Stokes radius or sedimentation value, depending onthe experiment. (a) Western blots using SEC-16 or TFG-1 antibodies ofC. elegans embryo extract fractionated on a Superose 6 gel filtrationchromatography column. The peaks corresponding to SEC-16 and TFG-1partially overlap. A Stokes radius was calculated for each protein onthe basis of comparison with the elution profiles of known standards.(b–d) Recombinant polyhistidine-tagged TFG-1 or fragments of TFG-1described in Fig. 1e were expressed and purified from E. coli extractsusing nickel resin. A Coomassie-blue-stained gel of the peak elution

fractions after fractionation of the recombinant proteins on a Superose 6gel filtration chromatography column is shown. Proteins were fractionatedon a 10–30% glycerol gradient, and S values were calculated on thebasis of the location of characterized standards run on a parallel gradient.(e) Western blots of control and TFG-1-depleted C. elegans whole-wormextracts fractionated on a Superose 6 gel filtration column and probed withSEC-16 or SEC-13 antibodies. Fractionation of HGRS-1, a componentof ESCRT-0 (endosomal sorting complex required for transport-0), wasexamined in both control and TFG-1-depleted conditions to ensure thatgel filtration profiles were directly comparable. Stokes radii were calculatedfor each protein on the basis of comparison with the elution profiles ofknown standards. Uncropped scans of gels and immunoblots are providedin Supplementary Fig. S6.

Concomitantly, loss of TFG-1 also affected the assembly of complexescontaining COPII proteins. Analysis of native SEC-13 by size-exclusionchromatography demonstrated that the COPII subunit is a componentof a large complex (Stokes radius of 128.1 Å), similar in size to SEC-16and TFG-1 (Fig. 4e). In addition, SEC-13 was found in a 54.3 Å com-plex, which probably corresponds to a SEC-13–SEC-31 heterotetramerthat was described previously10, and as a globular monomer (Stokes ra-dius of 23.9 Å). However, following TFG-1 depletion, most SEC-13 wasfound in a monomeric state (Stokes radius of 24.2 Å), although we con-tinued to detect the COPII subunit in high-molecular-weight fractionsas well. We conclude that the normal assembly of large complexes con-taining SEC-16 andCOPII subunits is facilitated by TFG-1 hexamers.

TFG-1 is required for protein secretionWe next examined the fate of COPII localization in animals with andwithout TFG-1. In control animals, the COPII subunit SEC-13 is foundin close apposition to SEC-16-labelled exit sites and on the nuclearenvelopes of proliferating germ nuclei (Fig. 5a). Following depletionof TFG-1, the intensities of both SEC-13 and SEC-16 on punctatestructures within the germline declined markedly, whereas localizationof SEC-13 to the nuclear envelope was largely unaffected (Fig. 5a). Onmeasuring the fluorescence intensities of SEC-13 and SEC-16 in controland TFG-1-depleted animals, we found that their ratio was not alteredsignificantly, indicating that the remaining SEC-16 on endoplasmicreticulum exit sites can continue to recruit COPII components in theabsence of TFG-1, albeit to lower levels (Fig. 5b).

To determine whether the substantially reduced levels of SEC-16 andSEC-13 were sufficient to support protein secretion following TFG-1depletion, we examined several secreted integral membrane proteins.We expressed a green fluorescent protein fusion to the v-SNAREsynaptobrevin (GFP–SNB-1) and anmCherry fusion to the PH domainof rat phospholipase C PLC11 (mCherry-PH), which binds specificallyto the plasma membrane18. In the distal region of the germlineunder steady-state conditions the majority of GFP–SNB-1 co-localizeswith mCherry-PH (Fig. 5c). However, in TFG-1-depleted animals,GFP–SNB-1 accumulated in the endoplasmic reticulum (Fig. 5c).Similarly, we found that several other transmembrane proteins,including theC. elegans caveolin-like protein CAV-1 and the glucuronyltransferase SQV-8, which normally traffic to other organelles followingsynthesis, accumulate in the endoplasmic reticulum in the absence ofTFG-1 (Supplementary Fig. S2c,d).To further examine the effect of TFG-1 depletion, we used

serial-section electron microscopy. In control animals, endoplasmicreticulum exit sites were juxtaposed to well-organized stacks ofmembranes corresponding to the ERGIC and Golgi cisternae (Fig. 5d,left). Vesicles that were 48.2± 1.2 nm (n = 15) in diameter wereobserved throughout the region between the endoplasmic reticulumand ERGIC. In contrast, subsequent to TFG-1 depletion, ERGIC andGolgi membranes seemed smaller and more poorly stacked, and theendoplasmic reticulum became fragmented (Fig. 5d, right). In addition,fewer Golgi networks were observed overall. However, consistentwith our findings that reduced levels of SEC-16 at endoplasmic



ART I C L E S

SEC-16/SEC-13 ×5 magnification

Con

trol

Inte

nsity

rat

io

(SE

C-1

3/S

EC

-16)

tfg-

1(R

NA

i)

Control

tfg-1(RNAi)

mCherry–PH/GFP–SNB-1mCherry–PH GFP–SNB-1

Distal gonad

Distal gonad

0.75

0.25

0.50

1.00

Control

tfg-1(RNAi)

a

c

Control tfg-1(RNAi)db

Figure 5 TFG-1 is required for COPII recruitment and protein secretion.(a) Dissected C. elegans gonads were fixed and stained using Cy2-labelledanti-SEC-13 and Cy3-labelled anti-SEC-16 antibodies (n = 15). Mergedimages of the distal gonad with SEC-13 in green and SEC-16 in red areshown on the left (scale bar, 10 µm). Panels to the right are magnified×5 views of the outlined area in the adjacent panel (scale bar, 2 µm).(b) Histogram showing the average ratio of SEC-13 to SEC-16 fluorescenceintensities in control and TFG-1-depleted animals. For each condition,at least 250 unique endoplasmic reticulum exit sites in the distal gonadwere examined. Error bars represent mean±s.e.m.; six different animals.No statistically significant difference was observed, on the basis of a

calculation using a paired Student t -test. (c) Swept-field confocal opticswere used to image anaesthetized control (n = 15) and TFG-1-depleted(n = 15) adult animals expressing GFP–SNB-1 and mCherry–PH. Scalebar, 10 µm. (d) Electron micrographs illustrating the early secretorypathway in the most proximal oocyte of control and TFG-1-depletedanimals following high-pressure freezing and freeze substitution (scalebar, 100nm). Arrowheads highlight endoplasmic reticulum exit sites.Below each micrograph is a three-dimensional reconstruction of the sameregions. Light grey, endoplasmic reticulum; dark grey, COPII coat; orange,endoplasmic-reticulum-derived transport vesicles and ERGIC; green, red andblue, Golgi cisternae, from cis to trans, respectively.

reticulum exit sites can continue to recruit COPII following depletionof TFG-1, we identified vesicles that were 48.8± 0.7 nm (n = 39)in diameter between the endoplasmic reticulum and the ERGIC.Although the sizes of vesicles were not affected by TFG-1 depletion,their distribution was altered.Most vesicles observed accumulated closeto the endoplasmic reticulum and failed to migrate towards the ERGICand Golgi membranes. These data indicate that TFG-1 is requiredfor the efficient, directed movement of COPII vesicles away from theendoplasmic reticulum, but not for their initial formation.We also generated animals that stably express a GFP fusion to the

Golgi Rab-type GTPase RAB-6, which mediates retrograde transportfrom the Golgi to the endoplasmic reticulum, to study perturbationsresulting from TFG-1 depletion. As RAB-6 also cycles between theGolgi and early endosomes19, we co-expressed an mCherry fusion tothe early endosome marker RAB-5, to enable us to distinguish betweenendosome- and Golgi-associated RAB-6. In control animals, RAB-5and RAB-6 exhibited distinct localizations in the distal region of thegermline (Supplementary Fig. S2e). However, following depletionof either SEC-16 or TFG-1, RAB-6 accumulated on fewer, largepunctate structures in the germline that also harboured RAB-5,indicating that RAB-6 localization was shifted from the Golgi toenlarged early endosomes (Supplementary Fig. S2e). These datafurther support the idea that TFG-1 depletion leads to a defect innormal Golgi assembly.

Human TFG localizes to endoplasmic reticulum exit sites andinteracts with Sec16C. elegans TFG-1 has a single human homologue, previously annotatedTFG (for TRK-fused gene; Supplementary Fig. S3a). We generated aGFP fusion to full-length TFG and co-transfected it into HeLa cellswith an mCherry fusion to human Sec16B to measure co-localization.

Analysis of cells expressing both markers indicated that more than 90%of GFP–TFG-labelled puncta also containedmCherry–Sec16B (Fig. 6a).These data demonstrate that, analogously to C. elegans TFG-1, humanTFG functions at endoplasmic reticulum exit sites.We next examined the dynamics of Sec16B relative to TFG using

fluorescence recovery after photobleaching. Similarly to GFP–Sec16A(ref. 20), we found that only a fraction (56.8 ± 6.3%) of thefluorescence of mCherry–Sec16B recovered following photobleaching,with a half-time of 5.68± 0.34 s. In contrast, a smaller proportionof GFP–TFG recovered (33.6± 5.1%), which exhibited a longerhalf-time of 8.22±0.64 s (Fig. 6b,c). These data indicate that Sec16Band TFG are not loaded onto exit sites simultaneously and insteadthat they assemble onto endoplasmic reticulum exit sites throughdistinct mechanisms.Similarly to our findings with C. elegans TFG-1, endogenous human

TFG exhibited a large Stokes radius (Fig. 6d), and a recombinant formof its N terminus (amino acids 1–193) assembled as an octamer with anativeMr of 172K (Fig. 6e,f). Thus, human TFG and C. elegans TFG-1share multiple physical properties, which indicate that they may share acommon function in vivo. Consistent with this hypothesis, depletion ofTFG caused a substantial delay in the secretion of VSVG(ts045)–GFP,a type I transmembrane protein used widely to study the mammaliansecretory pathway (Supplementary Fig. S3b).To determine whether human TFG binds to Sec16, we transfected

cells stably expressingmCherry–Sec16B with a GFP fusion to either full-length TFG or its N terminus, and conducted immunoprecipitationsusing mCherry antibodies. We found that endogenous TFG wasrecovered under these conditions (Fig. 6g). In addition, we foundthat both GFP-tagged isoforms were able to bind to mCherry–Sec16B(Fig. 6g). These data support the hypothesis that the N terminus of TFGinteracts with Sec16 in an evolutionarily conservedmanner.



ART I C L E S

mCherry–Sec16B/GFP–TFGmCherry–Sec16B GFP–TFG


TFG


10 11 12 13


14 15Load

MW

TFG(N)25Mr (K)


11Glycerol gradient (10–30%)

10965 7 8MW

S value = 6.42 S

25 Mr (K)

20TFG(N)

anti-mCherry IP+ + + +– –

anti-TFG IB anti-GFP IB

97 10 11 12 138

a GFP–TFG

Prebleach Bleach Postbleach–10 s 0 s 30 s

×3 magnification–10 s 0 s 30 s

bP

erce

ntag

e of

initi

al e

ndop

lasm

ic

retic

ulum

exi

t-si

te fl

uore

scen

ce

reco

vere

d

Time (seconds relative to bleach)

mCherry–Sec16BGFP–TFG

100

80

60

20

40

–10 0 10 20 30

c d e

f g

TFG

GFP–TFG

GFP–TFG(N)

mCherry–Sec16B

Figure 6 Human TFG functions at endoplasmic reticulum exit sites and bindsto Sec16. (a) Swept-field confocal optics were used to image HeLa cellsthat had been transiently transfected with GFP–TFG and mCherry–Sec16B(n = 42). Representative colour overlays of mCherry–Sec16B (red) andGFP–TFG (green) are shown. Scale bar, 10 µm. (b) Swept-field confocaloptics were used to monitor the recovery of GFP–TFG after photobleaching(n=15). A threefold-magnified view of the outlined region where GFP–TFGwas bleached is shown below. Times are in seconds relative to the bleach.Scale bars, 10 µm (top) and 1 µm (bottom). (c) Graph showing the averagepercentage of GFP–TFG and mCherry–Sec16B fluorescence recovered asa function of time in seconds relative to the bleach (error bars representmeans±s.e.m. for each time; n = 15 different cells for each fluorescentfusion protein). (d) Western blots of HeLa cell extract fractionated on aSuperose 6 gel filtration chromatography column (n =3). A Stokes radiuswas calculated for human TFG on the basis of comparison with the elution

profiles of known standards. (e,f) A GST-tagged, truncated form of humanTFG, amino acids 1–193, was expressed and purified from E. coli extractsusing glutathione agarose (n=3), and the GST tag was subsequently cleavedusing PreScission Protease before loading onto a gel filtration column orglycerol gradient. A Coomassie-blue-stained gel of the peak elution fractionsafter fractionation of the recombinant protein, referred to as TFG(N), on aSuperose 6 gel filtration chromatography column is shown (e). The proteinwas also fractionated on a 10–30% glycerol gradient (f), and an S valuewas calculated on the basis of the location of characterized standards runon a parallel gradient (n=3). (g) Antibodies directed against mCherry wereused to immunoprecipitate mCherry–Sec16B from HeLa cells transientlytransfected with GFP–TFG or a GFP fusion to the N terminus of TFG referredto as GFP–TFG(N) (n =3). Isolated proteins were separated by SDS–PAGEand immunoblotted (IB) with anti-TFG and anti-GFP antibodies. Uncroppedscans of gels and immunoblots are provided in Supplementary Fig. S6.

Mis-targeting of the NTRK1 kinase domain to endoplasmicreticulum exit sites causes the hyperactivation of downstreamNTRK1 effectors and cell transformationTFG was initially identified as part of a gene fusion with theneurotrophic tyrosine kinase receptor type 1 (NTRK1, also called TrkA),resulting in the formation of an oncogene21. Subsequent studies haveshown that TFG is also a fusion partner of the anaplastic lymphomareceptor tyrosine kinase ALK in some anaplastic large cell lymphomas22.In both instances, the sequences encoding the N terminus of TFG andthe C terminus of the other gene are fused, resulting in an oncogenicproduct. We found that the N-terminal domain of TFG localizes toendoplasmic reticulum exit sites and is capable of redirecting theC-terminal domain of NTRK1 there as well (Fig. 7a–d). These data areindicative of a role for TFG in targeting the kinase activities of NTRK1and ALK to endoplasmic reticulum exit sites.NTRK1 is a high-affinity receptor for nerve growth factor23

(NGF). Binding of NGF to NTRK1 leads to its dimerization andautophosphorylation, ultimately causing the activation of severaldownstream signalling cascades, including the Ras–Raf–MEK (MAP-kinase kinase)–ERK (extracellular signal-regulated kinase) pathway, topromote cell survival and growth24. The oncogenic TFG–NTRK1 fusionprotein is constitutively active25, and expression of GFP–TFG–NTRK1

in hTERT-RPE1 cells caused an increase in the levels of phospho-ERK1and -ERK2 when compared with control cells (Fig. 7e). To determinewhether dimerization of the kinase domain is sufficient to initiatesignalling downstream of NTRK1, we linked the constitutive dimerGST to NTRK1 and expressed it as a GFP fusion in hTERT-RPE1 cells.Under these conditions, levels of phospho-ERK1 and -ERK2 increasedonly twofold when compared with control cells (Fig. 7e). These dataindicate that dimerization of the NTRK1 kinase domain alone is notadequate to cause the hyperactivation of its downstream effectors as isobserved following TFG–NTRK1 expression.As suggested earlier, an alternative mechanism by which fusion to

TFGmay cause oncogenic NTRK1 activity is through itsmislocalizationto endoplasmic reticulum exit sites. Therefore, we tested the effect oftargeting the NTRK1 kinase domain to exit sites independently of TFG.We fused a region of Sec16B (amino acids 1–791), which localizes toendoplasmic reticulum exit sites similarly to the full-length protein,to NTRK1 and expressed it as a GFP fusion in hTERT-RPE1 cells(GFP–Sec16B–NTRK). Hydrodynamic analysis of a portion of thisregion in Sec16B (amino acids 34–234) indicated that it oligomerizes invitro (Supplementary Fig. S4a). GFP–Sec16B–NTRK1 localized to endo-plasmic reticulum exit sites (Supplementary Fig. S4b), and expressionof this fusion protein resulted in hyperactivation of ERK1 and ERK2,



ART I C L E S

a mCherry–Sec16B/GFP–TFG(N)mCherry–Sec16B GFP–TFG(N)

TFG

Per

cent

age

ofS

ec16

B c

o-lo

caliz

atio

n

TFG–

NTRK

TFG(N) NTRK(C)

GST–NTR

K

Sec16

B–NTR

K

TFG–N

TRK

Contro

l

60

20

40

80

ERK2ERK1

phospho-ERK2phospho-ERK1

GFP–NTRK(C)mCherry–Sec16BmCherry–Sec16B/

GFP–NTRK(C)

GFP–TFG–NTRKmCherry–Sec16BmCherry–Sec16B/GFP–TFG–NTRK

b

cd e

Figure 7 Targeting of the NTRK1 kinase domain to endoplasmic reticulumexit sites is sufficient to activate NTRK1-mediated downstream signalling.(a–c) Swept-field confocal optics were used to image HeLa cells that hadbeen transiently transfected with mCherry–Sec16B and GFP fusions toeither the N terminus of TFG, referred to as GFP–TFG(N) (n = 18), thetransmembrane and kinase domains of NTRK1, referred to as GFP–NTRK(C)

(n = 15), or a TFG(N)–NTRK1(C) fusion (n = 28), which is equivalent tothe oncogene characterized previously21. Representative colour overlaysof mCherry–Sec16B (red) and GFP fusions (green) are shown. Scale bar,

10 µm. (d) Histogram showing the percentage of co-localization between theGFP fusions described above and mCherry–Sec16B (error bars representmeans±s.e.m. for each condition; n =15 different cells for each conditionand at least 800 unique endoplasmic reticulum exit sites were examined).(e) Extracts from hTERT-RPE1 cells stably transfected with GFP alone(control) or various GFP fusions to the NTRK1 transmembrane and kinasedomains (as indicated) were separated by SDS–PAGE and blotted using aphospho-specific ERK1–ERK2 antibody and a pan-ERK1–ERK2 antibody.Uncropped scans of immunoblots are provided in Supplementary Fig. S6.

similar to the effect following GFP–TFG–NTRK1 expression (Fig. 7e).We conclude that targeting of the NTRK1 kinase domain to endoplas-mic reticulum exit sites is necessary for its constitutive activity in cells.To test whether the endoplasmic reticulum exit-site targeting of

NTRK1 was sufficient to transform cells, we generated NIH3T3 celllines expressing either GFP, GFP–GST–NTRK1, GFP–TFG–NTRK1or GFP–Sec16B–NTRK1. Cells expressing the endoplasmic reticulumexit-site-targeted forms of NTRK1 produced at least eight times morecolonies than the dimerized form of NTRK1 (Supplementary Fig. S4c).These data indicate that localization of NTRK1 to endoplasmicreticulum exit sites strongly contributes to its ability to transform cells.

DISCUSSIONWe have demonstrated that TFG-1 localizes to endoplasmic reticulumexit sites and is required for protein secretion in the C. elegans germline.In addition, we have found that TFG-1 accumulates at endoplasmicreticulum exit sites in several other tissues, including the intestine, hy-podermis andmuscle (Supplementary Fig. S5), indicative of a commonrole in multiple cell types. We further demonstrate that human TFGfunctions at endoplasmic reticulum exit sites and probably regulates thesecretion of multiple cargoes. Consistent with this idea, overexpressionof human TFG has been shown to partially rescue the traffickingdefect of a mutant isoform of the cystic fibrosis transmembraneconductance regulator (CFTR-F508del), which aberrantly accumulatesin the endoplasmic reticulum in cystic fibrosis patients26,27. As levels ofTFG-1 at endoplasmic reticulum exit sites seem to correlate with levelsof SEC-16 and COPII, extra TFG-1 may increase COPII recruitmentand stimulate secretion from the endoplasmic reticulum28–30.On the basis of our data, we speculate that TFG family members

generate a matrix at endoplasmic reticulum exit sites, which mayserve as a molecular sink to help retain COPII components locallyand facilitate efficient vesicle formation and egress. The increased

membrane flux mediated by TFG-1 may also function to maintainproper Golgi organization in tissues under a high secretory demand,such as the C. elegans germline. Together, our findings reveal a newcomponent of the early secretory pathway that regulates anterogradetrafficking from the endoplasmic reticulum.Furthermore, our studies also provide a key insight into the

mechanism by which TFG functions as a proto-oncogene. In the caseof the TFG–NTRK1 fusion generated by a chromosomal translocation,concentrating constitutive NTRK1 kinase activity at endoplasmicreticulum exit sites may cause the premature stimulation of multipleeffectors, including components of the ERK1–ERK2 kinase cascade,which leads to cell transformation. Notably, stimulation of ERK2 atendoplasmic reticulum exit sites may cause hyper-phosphorylation ofSec16, resulting in the formation of new exit sites that would recruitextra TFG–NTRK1 (ref. 31). Such a feed-forward mechanismmay playan important role during oncogenesis. �

METHODSMethods and any associated references are available in the onlineversion of the paper at http://www.nature.com/naturecellbiology/

Note: Supplementary Information is available on the Nature Cell Biology website

ACKNOWLEDGEMENTSThis work was supported in part by grants from the NIH (1R01GM088151-01A1 toA.A. and P41RR011823 to J.R.Y.).We thank B.Weaver and P. Bertics for use of tissueculture facilities, R. Landick for use of a Gradient Master, K. Oegema for markerstrains and antibodies, E. Chapman for use of Metamorph software and S. Koenigfor help with RNAi.We also thankD. Stephens, P. Kiley andmembers of the Audhyalaboratory for critically reading this manuscript.

AUTHOR CONTRIBUTIONSK.W., A.L.S., S.E. and A.A. conceived and designed experiments. K.W., A.L.S., A.S.,J.H., J.R.M., K.S., S.E. andA.A. carried out experiments and analysed data. S.E., J.R.Y.and A.A. contributed reagents, materials and analysis tools. A.A. wrote the paper.



http://www.nature.com/naturecellbiology/

ART I C L E S

COMPETING FINANCIAL INTERESTSThe authors declare no competing financial interests.

Published online at http://www.nature.com/naturecellbiologyReprints and permissions information is available online at http://npg.nature.com/reprintsandpermissions/

1. Lee, M. C., Miller, E. A., Goldberg, J., Orci, L. & Schekman, R. Bi-directional proteintransport between the ER and Golgi. Annu. Rev. Cell Dev. Biol. 20, 87–123 (2004).

2. Bonifacino, J. S. & Glick, B. S. The mechanisms of vesicle budding and fusion. Cell116, 153–166 (2004).

3. Bickford, L. C., Mossessova, E. & Goldberg, J. A structural view of the COPII vesiclecoat. Curr. Opin. Struct. Biol. 14, 147–153 (2004).

4. Dancourt, J. & Barlowe, C. Protein sorting receptors in the early secretory pathway.Annu. Rev. Biochem. 79, 777–802 (2010).

5. Boyadjiev, S. A. et al. Cranio-lenticulo-sutural dysplasia is caused by a SEC23Amutation leading to abnormal endoplasmic-reticulum-to-Golgi trafficking. Nat.Genet. 38, 1192–1197 (2006).

6. Lang, M. R., Lapierre, L. A., Frotscher, M., Goldenring, J. R. & Knapik, E. W.Secretory COPII coat component Sec23a is essential for craniofacial chondrocytematuration. Nat. Genet. 38, 1198–1203 (2006).

7. Matsuoka, K. et al. COPII-coated vesicle formation reconstituted with purified coatproteins and chemically defined liposomes. Cell 93, 263–275 (1998).

8. Miller, E. A. & Barlowe, C. Regulation of coat assembly—sorting things out at the ER.Curr. Opin. Cell Biol. 22, 447–453 (2010).

9. Yoshihisa, T., Barlowe, C. & Schekman, R. Requirement for a GTPase-activatingprotein in vesicle budding from the endoplasmic reticulum. Science 259,1466–1468 (1993).

10. Stagg, S. M. et al. Structure of the Sec13/31 COPII coat cage. Nature 439,234–238 (2006).

11. Bi, X., Mancias, J. D. & Goldberg, J. Insights into COPII coat nucleation from thestructure of Sec23.Sar1 complexed with the active fragment of Sec31. Dev. Cell 13,635–645 (2007).

12. Espenshade, P., Gimeno, R. E., Holzmacher, E., Teung, P. & Kaiser, C. A. YeastSEC16 gene encodes a multidomain vesicle coat protein that interacts with Sec23p.J. Cell Biol. 131, 311–323 (1995).

13. Gimeno, R. E., Espenshade, P. & Kaiser, C. A. COPII coat subunit interactions:Sec24p and Sec23p bind to adjacent regions of Sec16p. Mol. Biol. Cell 7,1815–1823 (1996).

14. Shaywitz, D. A., Espenshade, P. J., Gimeno, R. E. & Kaiser, C. A. COPIIsubunit interactions in the assembly of the vesicle coat. J. Biol. Chem. 272,25413–25416 (1997).

15. Whittle, J. R. & Schwartz, T. U. Structure of the Sec13–Sec16 edge element,a template for assembly of the COPII vesicle coat. J. Cell Biol. 190,347–361 (2010).

16. Supek, F., Madden, D. T., Hamamoto, S., Orci, L. & Schekman, R. Sec16ppotentiates the action of COPII proteins to bud transport vesicles. J. Cell Biol. 158,1029–1038 (2002).

17. Li, L. & Xie, T. Stem cell niche: structure and function. Annu. Rev. Cell Dev. Biol.21, 605–631 (2005).

18. Shi, A. et al. EHBP-1 functions with RAB-10 during endocytic recycling inCaenorhabditis elegans. Mol. Biol. Cell 21, 2930–2943 (2010).

19. Mallard, F. et al. Early/recycling endosomes-to-TGN transport involves two SNAREcomplexes and a Rab6 isoform. J. Cell Biol. 156, 653–664 (2002).

20. Hughes, H. et al. Organisation of human ER-exit sites: requirements for thelocalisation of Sec16 to transitional ER. J. Cell Sci. 122, 2924–2934 (2009).

21. Greco, A. et al. The DNA rearrangement that generates the TRK-T3 oncogene involvesa novel gene on chromosome 3 whose product has a potential coiled-coil domain.Mol. Cell Biol. 15, 6118–6127 (1995).

22. Hernandez, L. et al. TRK-fused gene (TFG) is a new partner of ALK in anaplastic largecell lymphoma producing two structurally different TFG-ALK translocations. Blood94, 3265–3268 (1999).

23. Schecterson, L. C. & Bothwell, M. Neurotrophin receptors: old friends with newpartners. Dev. Neurobiol. 70, 332–338 (2010).

24. Klesse, L. J. & Parada, L. F. Trks: signal transduction and intracellular pathways.Microsc. Res. Tech. 45, 210–216 (1999).

25. Greco, A. et al. Role of the TFG N-terminus and coiled-coil domain inthe transforming activity of the thyroid TRK-T3 oncogene. Oncogene 16,809–816 (1998).

26. Zhang, F., Kartner, N. & Lukacs, G. L. Limited proteolysis as a probe forarrested conformational maturation of delta F508 CFTR. Nat. Struct. Biol. 5,180–183 (1998).

27. Trzcinska-Daneluti, A. M. et al. High-content functional screen to identifyproteins that correct F508del-CFTR function. Mol. Cell Proteomics. 8,780–790 (2009).

28. Heinzer, S., Worz, S., Kalla, C., Rohr, K. & Weiss, M. A model for the self-organizationof exit sites in the endoplasmic reticulum. J. Cell Sci. 121, 55–64 (2008).

29. Farhan, H., Weiss, M., Tani, K., Kaufman, R. J. & Hauri, H. P. Adaptation ofendoplasmic reticulum exit sites to acute and chronic increases in cargo load. EMBOJ. 27, 2043–2054 (2008).

30. Hughes, H. & Stephens, D. J. Assembly, organization, and function of the COPII coat.Histochem. Cell Biol. 129, 129–151 (2008).

31. Farhan, H. et al. MAPK signalling to the early secretory pathway revealed bykinase/phosphatase functional screening. J. Cell Biol. 189, 997–1011 (2010).



http://www.nature.com/naturecellbiology

http://npg.nature.com/reprintsandpermissions/

http://npg.nature.com/reprintsandpermissions/

DOI: 10.1038/ncb2225 METHODS

METHODSAntibodies. C. elegans SEC-16, TFG-1 and CYP-33E1 antibodies were raised inrabbits by immunization with GST fusions to either a fragment of SEC-16 (aminoacids 813–1,014), full-length TFG-1 or a fragment of CYP-33E1 (amino acids244–392) produced in E. coli. The antibodies were affinity purified from serumby binding to columns of the same antigens following removal of the GST tagsby cleavage with PreScission Protease. To produce SEC-13-specific monoclonalantibodies, ten mice were each immunized with 50 µg of purified GST–SEC-13fusion protein in PBS using complete Freund’s adjuvant. Hybridoma fusions weregenerated following booster injections and subsequently screened by enzyme-linkedimmunosorbent assay according to standard protocols. Mouse monoclonal SEC-13antibodies were purified through Protein A Sepharose beads. Antibodies directedagainst human TFG were purchased from Bethyl Laboratories (A302-341A andA302-343A). Antibodies against ERK1 and ERK2 and phospho-ERK1 and -ERK2were obtained from Millipore (06–182) and Invitrogen (44680G), respectively.CAR-1, SQV-8, GFP and mCherry antibodies have been described elsewhere32–34.All antibodies were used at 1 µgml−1.

Immunofluorescence, live-cell imaging and electron microscopy. Imageswere acquired on a swept-field confocal microscope (Nikon Ti–E) equipped witha Roper CoolSnap HQ2 CCD (charge-coupled device) camera using a Nikon×60, 1.4 numerical aperture Planapo oil objective lens. Acquisition parameterswere controlled by Nikon Elements software, and image analysis was conductedusing Metamorph software. Immunofluorescence of fixed gonads was observed asdescribed previously33 using directly labelled rabbit antibodies at a concentrationof 1 µgml−1. Briefly, 60–120 Z sections at 0.2 µm steps were acquired (dependingon sample thickness). The fluorescence intensity of each endoplasmic reticulumexit site was always confined to a maximum of seven Z planes, which were usedto generate a maximum-intensity projection. To calculate the fluorescence intensityof proteins that localize to endoplasmic reticulum exit sites, the total intensity ina box containing the endoplasmic reticulum exit site (from a maximum-intensityprojection) was measured and the camera background was subtracted. For live-cellimaging of C. elegans gonads, animals were anaesthetized and mounted on anagarose pad. HeLa cells were grown on 35mm glass-bottom dishes maintainedat 37 ◦C for time-lapse imaging. Photobleaching was carried out using a 405 nmlaser, coupled into a Nikon Photo Activation Illuminator Unit to create a singlediffraction-limited spot. The signal at the first post-bleach time was subtractedfrom all post-bleach measurements, and the percentage of fluorescence recoveredat each time was calculated by dividing by the difference between the pre-bleachand first post-bleach measurements. Kaleidagraph software was used to fit the dataand calculate the maximal fractional recovery and the half-time for recovery. Todetermine the percentage of co-localization between two proteins, endoplasmicreticulum exit sites were analysed individually using Nikon Elements software. TheVSVG(ts045)–GFP trafficking assay was conducted as described previously35, withmodifications as noted in the legend of Supplementary Fig. S3b.

For high-pressure freezing, a 100-µm-deep aluminium platelet (MicroscopyServices) was filled with 20 adult worms in a suspension of E. coli and frozen usinga BalTec HPM 10. Freeze substitution was carried out in a Leica EM AFS at−90 ◦Cas described previously36. For electron microscopy studies, 50 nm longitudinalsections of C. elegans were cut using a Leica UC6 ultramicrotome. Ribbons ofsections were transferred onto Formvar-coated copper slot grids. The grids wereplaced in drops of 4% uranyl acetate, washed with water and dried. They werethen transferred onto lead citrate37 and rinsed again with water. Micrographs weretaken with a Proscan CCD HSS in a Zeiss EM 902A electron microscope operatedin bright-field mode. For three-dimensional reconstructions, serial sections ofendoplasmic reticulum exit sites were imaged, and files were aligned linearly usingthe software Reconstruct38. For electron tomography, 250–300 nm EPON sectionswere transferred onto Formvar-coated copper slot grids and stained similarly asdescribed for thin sections. Ten-nanometre gold beads were applied to both sides.An orthogonal tilt series was acquired on a JEOL JEM 2100 at 200 kV from−55◦ to 55◦ (1◦ increments). Reconstruction was done using IMOD software(http://bio3d.colourado.edu/imod). Immuno-electron microscopy was conductedas described previously36.

Worm strains, RNAi and cell culture. All C. elegans strains were derived fromthe Bristol strain N2. The generation of animals expressing fluorescent fusions withCAV-1, SNB-1 and the PH domain of rat PLC1δ has been described previously18,34.Double-stranded RNA (dsRNA) was synthesized as described previously33 fromtemplates prepared by using primers listed in the Methods section to amplify N2genomic DNA. For most RNAi experiments, early-L4-stage hermaphrodites weresoaked in dsRNA for 24 h at 20 ◦C. Animals were allowed to recover for 48 hbefore analysis. For partial depletions, late-L4-stage worms were soaked in dsRNAfor 24 h, followed by a 22 h recovery period, and then analysed. For large-scaledepletion of TFG-1 necessary for gel filtration studies, L4 animals were fed bacteriaexpressing dsRNA directed against TFG-1 for 72 h, before collecting and extractpreparation. HeLa and NIH3T3 cells were maintained in DMEM supplemented

with 10% FBS (HeLa) or 10% FCS (NIH3T3), penicillin, streptomycin andl-glutamine at 37 ◦C in the presence of 5% CO2. hTERT-RPE1 cells were grownsimilarly, with the exception that DMEM/F-12 media was used. Transfections wereconducted using Lipofectamine 2000 (Invitrogen), and cells were selected usingpuromycin (1 µgml−1 for HeLa cells and 12.5 µgml−1 for hTERT-RPE1 cells).Colony-formation assays on monolayers of contact-inhibited NIH3T3 cells werecarried out as described previously39.

Primers, dsRNAs and plasmids used in this study. To generatedsRNAs directed against TFG-1, the following primer sets were used: (5′-AATTAACCCTCACTAAAGGCTGCTGTGGTGGAGCATATC-3′ and 5′-TAATA-CGACTCACTATAGGATCTCTCGGCTCCAAAACAA-3′), (5′-AATTAACCCTC-ACTAAAGGCTTAATCTGCTCGACTTGCT-3′ and 5′-TAATACGACTCACTATA-GGATGGTGCATTCAAACGGAGC-3′) and (5′-AATTAACCCTCACTAAAGG-TTACTGCTGATACGGCGACT-3′ and 5′-TAATACGACTCACTATAGGCAGCA-GCAGCAATTCGGAGC-3′). Each dsRNA produced using these primer setsyielded similar results. To generate dsRNAs directed against SEC-16, the followingprimer sets were used: (5′-AATTAACCCTCACTAAAGGGCGGTCTGCGAGTTT-AGATT-3′ and 5′-TAATACGACTCACTATAGGCAAGCGGATGCAAGAAGAAT-3′), (5′-AATTAACCCTCACTAAAGGGCAAACCTGTAATTTAAAAT-3′ and 5′-TAATACGACTCACTATAGGCTTCTGGTTTCGATATGAGT-3′) and (5′-AA-TTAACCCTCACTAAAGGCTCCACTTCTAGCACTTCGC-3′ and 5′-TAATACGA-CTCACTATAGGGTAACGATCAACAACATAAC-3′). Each dsRNA producedusing these primer sets yielded similar results. For short interfering RNAexperiments, the following oligonucleotide was used to efficiently deplete TFG:5′-CUUCUCAGCCUACUAAUUA-3′. The control short interfering RNA usedwas directed against the endoplasmic reticulum exit-site component TANGO1(5′-GCAAUAACCUCAACUCUAUUU-3′), which had been shown previously notto affect the trafficking of VSVG–GFP (ref. 40). To generate polyhistidine fusionconstructs, DNAs were cloned by restriction digest into pRSETA, which encodesa 6×histidine tag on the N terminus. To generate GST fusion constructs, DNAswere cloned by restriction digest into pGEX6P-1, which encodes a GST tag onthe N terminus, followed by a PreScission Protease cleavage site. A C. eleganscomplementary DNA library was used as a template for cloning recombinantexpression constructs by PCR, and cDNAs for human TFG and NTRK1 werepurchased from Open Biosystems.

Mass spectrometry analysis and biochemistry. Gravid adult hermaphroditeswere grown synchronously in liquid culture, and embryos were isolated in buffercontaining 0.6N NaOH and 20% bleach. Extracts were generated in lysis buffer(50mM HEPES, 1mM EDTA, 1mM MgCl2, 100mM KCl and 10% glycerol) andused for immunoprecipitations as described previously32. For mass spectrometry,proteins were precipitated using trichloroacetic acid. The trichloroacetic acid pelletswere solubilized and treated with Endoproteinase Lys-C (0.1 µg µl−1), followed bytrypsin (0.5 µg µl−1). Following digestion, the proteins were pressure-loaded onto afused-silica capillary desalting column, placed in line with aHewlett Packard Agilent1100 Quaternary Pump (Version 1.4) and analysed using a customized four-stepseparationmethod (90, 120, 120 and 150min respectively; ref. 41). For each step, onefull-scanmass spectrum (400–2,000m/z) occurred followed by five data-dependentMS/MS spectra at a 35%normalized collision energy. The spectrawere searchedwiththe SEQUEST algorithm42 against theWormBaseC. elegans (Version 2.11 created on10 January 2010) database.

For immunoprecipitations, HeLa or hTERT-RPE1 cells were grown in 15 cmplates, and extracts were generated in lysis buffer, which were subjected tocentrifugation at 100,000g before use. Recombinant protein expression was carriedout using BL21 (DE3) E. coli, and purifications were conducted using eitherglutathione agarose beads (forGST fusions) or nickel affinity resin (for polyhistidine-tagged proteins). For samples applied to a Superose 6 gel filtration column, theStokes radius of each protein or protein complex was calculated from its elutionvolume on the basis of the elution profiles of characterized standards. Glycerolgradients (10–30%) were poured using a Gradient Master and fractionated fromthe top by hand. Sedimentation values were calculated by comparing the positionof the peak with that of characterized standards run on a separate gradient inparallel. To determine the native molecular mass of proteins, the following equationwas used: M = 6πηNas/(1− υρ), where M is the native molecular mass, η isthe viscosity of the medium, N is Avogadro’s number, a is the Stokes radius, s isthe sedimentation value, υ is the partial specific volume and ρ is the density ofthe medium43. Immunoblotting of extracts and immunoprecipitates were carriedout as described previously32. To determine whether endogenous SEC-16 couldinteract with various fragments of recombinant TFG-1, extracts were generated fromanimals freshly collected from 15 plates measuring 10 cm that were subsequentlysubjected to sonication in lysis buffer containing 1% Triton X-100. Extracts wereclarified by centrifugation (100,000g ) before incubation with proteins bound tonickel affinity resin. To determine the level of TFG-1 depletion following RNAi,both control and dsRNA-treated animals (60 each) were hand-picked and placedinto Eppendorf tubes containing 100 µl of lysis buffer. Samples were sonicated

NATURE CELL BIOLOGY


http://bio3d.colourado.edu/imod

METHODS DOI: 10.1038/ncb2225

in a water bath sonicator set to 80 ◦C in the presence of 1× sample bufferfor 10min, after which they were boiled for 10min at 100 ◦C before separationby SDS–PAGE.

Statistical analysis. Statistical significance was evaluated by carrying out a two-tailed Student t -test.

32. Cheeseman, I. M. et al. A conserved protein network controls assembly of the outerkinetochore and its ability to sustain tension. Genes Dev. 18, 2255–2268 (2004).

33. Audhya, A. et al. A complex containing the Sm protein CAR-1 and the RNA helicaseCGH-1 is required for embryonic cytokinesis in Caenorhabditis elegans. J. Cell Biol.171, 267–279 (2005).

34. Sato, K. et al. Dynamic regulation of caveolin-1 trafficking in the germ line andembryo of Caenorhabditis elegans. Mol. Biol. Cell 17, 3085–3094 (2006).

35. Presley, J. F. et al. ER-to-Golgi transport visualized in living cells. Nature 389,81–85 (1997).

36. Rostaing, P. et al. Analysis of synaptic ultrastructure without fixative usinghigh-pressure freezing and tomography. Eur. J. Neurosci. 24, 3463–3474 (2006).

37. Reynolds, E. S. The use of lead citrate at high pH as an electron-opaque stain inelectron microscopy. J. Cell Biol. 17, 208–223 (1963).

38. Fiala, J. C. Reconstruct: a free editor for serial section microscopy. J. Microsc. 218,52–61 (2005).

39. Perlaky, L. et al. Increased growth of NIH/3T3 cells by transfection with human p120complementary DNA and inhibition by a p120 antisense construct. Cancer Res. 52,428–436 (1992).

40. Saito, K. et al. TANGO1 facilitates cargo loading at endoplasmic reticulum exit sites.Cell 136, 891–902 (2009).

41. Diop, S. B. et al. Reptin and pontin function antagonistically with PcG and TrxGcomplexes to mediate Hox gene control. EMBO Rep. 9, 260–266 (2006).

42. Eng, J., McCormack, A. & Yates, J. R. III An approach to correlate tandem massspectral data of peptides with amino acid sequences in a protein database. J. Am.Soc. Mass Spectrom. 5, 979–989 (1994).

43. Siegel, L. M. & Monty, K. J. Determination of molecular weights and frictionalrations of proteins in impure systems by use of gel filtration and density gradientcentrifugation. Application to crude preparations of sulfite and hydroxylaminereductases. Biochim. Biophys. Acta 112, 346–362 (1966).

NATURE CELL BIOLOGY


S U P P L E M E N TA RY I N F O R M AT I O N

WWW.NATURE.COM/NATURECELLBIOLOGY 1

DOI: 10.1038/ncb2225

Figure S1 TFG-1 functions with SEC-16 at ER exit sites. (a) C. elegans embryo extract was separated by SDS-PAGE and immunoblotted using SEC-16 antibodies. (b) C. elegans embryo extract was subjected to differential centrifugation in the presence or absence of 1% Tx-100. The 13,000 x g pellet (P13), 100,000 x g pellet (P100), and 100,000 x g supernatant (S100) were recovered and immunoblotted using a-SEC-16 (top) or a-TFG-1 (bottom) antibodies. The P13 fraction is enriched with ER and plasma membrane proteins, while the P100 fraction contains Golgi and endosome membranes (our unpublished data). (c) SEC-16 was immunoprecipitated from detergent-solubilized C. elegans embryo extract, and purified proteins were resolved by SDS-PAGE, followed by silver staining. Bands corresponding to the molecular weights of SEC-16 and TFG-1 are highlighted. (d) C. elegans embryo extract was separated by SDS-PAGE and immunoblotted using TFG-1 antibodies. (e) L1 stage larvae harboring the temperature sensitive glp-1(q224) mutation were grown at 15ºC or 25ºC immediately after hatching. At 25ºC, in the absence of GLP-1

function, larvae fail to generate a germline (1). Extracts were made from 30 animals grown at each temperature and immunoblotted using a-TFG-1 (top), a-CYP-33E1 (middle; not expressed in the germline), or a-CAR-1 (bottom; expressed only in the germline). Serial dilutions of extracts prepared from animals grown at 15ºC were loaded to quantify levels of TFG-1 remaining in animals without a germline. (f) Higher magnifications of the ER exit sites shown in panel g with arrowheads marking ER exit sites. On the right are sketches illustrating the ER exit sites (drawn to scale), ribosomes (black dots) and COPII coats (thick black lines). Bar, 50 nm. (g) Higher magnifications of two EM sections from the reconstructed series shown in Figure 2d. Arrowheads highlight ER exit sites. Bars, 200 nm. (h) Dissected gonads were fixed and stained using Cy2-labeled a-TFG-1 and Cy3-labeled a-SQV-8 (Golgi) antibodies (n=3). Both individual and merged images of proximal oocytes with TFG-1 in green and SQV-8 in red are shown (Bar, 10 mm). The panel on the right is the boxed region magnified 6x relative to the adjacent images (Bar, 2 mm).

b

SEC-16

TFG-1

1% Tx-100

P13 P100

S100

Lysis Buffer

P13 P100

S100

e

TFG-1

100%

50%

20%

10%

100%

15 Co 25 Co

glp-1(q224)

CYP-33E1

CAR-1

g

f

Witte et al., Supplemental Figure S1

Control

SEC-16

sec-16

(RNAi)

75kD

50kD

25kD

37kD

20kD

100kD150kD

a

Control

tfg-1

(RNAi)

TFG-175kD

50kD

25kD

37kD

20kD

100kD150kD

d

75kD

50kD

25kD

37kD

20kD

100kD150kD

10kD

SEC-16

250kD

TFG-1

MW

SEC-16 I

Pc

SQV-8 (Golgi) TFG-1 SQV-8 / TFG-1 SQV-8 / TFG-1 (6x)h

Pro

xim

al O

ocy

te



2 WWW.NATURE.COM/NATURECELLBIOLOGY

Figure S2 TFG-1 is required for secretion and normal localization of RAB-6. (a) Fluorescence intensities of SEC-16 and TFG-1 in the distal gonad were measured in control animals, and at each ER exit site, a ratio was calculated. Each point in the scatter plot indicates a ratio between the intensities. Although more than 600 ER exit sites were analyzed, only a small (random) selection of ratios is shown due to space limitations. A line indicating the average intensity ratio for all analyzed ER exit sites is shown. (b) Dissected C. elegans gonads from control and TFG-1 depleted animals were fixed and stained using Cy3-labeled a-SEC-16 antibodies (n=15 animals each). The number of ER exit sites (puncta harboring SEC-16) was counted for a 50 mm section from the distal-most region of the germline (in the medial confocal plane of each gonad). Error bars represent mean +/- SEM; 15 different animals for each condition. No statistically significant difference was observed, based on a calculation using a paired t test. (c) Dissected C. elegans gonads from control and TFG-1 depleted animals

expressing GFP:CAV-1 were fixed and stained using Cy2-labeled a-GFP and Cy3-labeled a-SQV-8 antibodies (n=12). Merged images of the proximal oocytes with GFP:CAV-1 in green and SQV-8 in red are shown. Scale bar, 10 mm. (d) Dissected C. elegans gonads from TFG-1 depleted animals expressing GFP:CAV-1 were fixed and stained using Cy2-labeled a-GFP and Cy3-labeled a-RET-1 (reticulon) antibodies, which label ER tubules (n=8). Merged images of the proximal oocytes with GFP:CAV-1 in green and RET-1 in red are shown. Scale bar, 10 mm. (e) Live control (top), TFG-1 depleted (middle), and SEC-16 depleted (bottom) animals stably expressing GFP:RAB-6 and mCherry:RAB-5 were anesthetized and mounted onto agarose pads for imaging (n=15 each). Both the distal gonad and uterus are indicated in each series of panels. In the absence of TFG-1 or SEC-16, animals become sterile, as indicated by the lack of embryos in the uterus. Bar, 10 mm. On the right are 4x zoomed images of the boxed regions shown in the merged panels to the left. Bar, 2.5 mm.

sec-16

(RNAi)

Co

ntro

l

mCherry:RAB-5 / GFP:RAB-6mCherry:RAB-5 GFP:RAB-6

tfg-1

(RNAi)

Uterus

DistalGonad

Uterus

DistalGonad

Uterus

DistalGonad

SQV-8 / CAV-1CAV-1 SQV-8

Co

ntro

ltfg-1

(RNAi)

Pro

ximal O

ocyte

c

e


Inte

nsit

y R

atio

.75

.25

.50

1.0

SEC-16TFG-1

a

Co

ntro

lNum

ber

of

ER

ES

(per

50 μ

m s

ecti

on)

150

50

100tfg-1

(RNAi)

b

d CAV-1 RET-1 (ER marker) RET-1 / CAV-1

tfg-1

(RNAi)

Co

rtical View

4x ZoomRAB-5 / RAB-6




Figure S3 Human TFG is a conserved protein required for VSVG(ts045)-GFP secretion. (a) Cartoon illustrating the conserved domains found in TFG isoforms. Based on previous studies, the PB1 domain may function in protein-protein interactions (2). The specific roles of the conserved coiled-coil domain (CC) and the proline/glutamine enriched domain (PQ) remain largely undefined. (b) HeLa cells were co-transfected with a plasmid encoding VSVG(ts045)-GFP and either control or TFG siRNA (see Methods

for more information). After 36 hours, cells were shifted to 40°C for 16 hours, and then shifted to 32°C for imaging (3 confocal Z sections acquired at 6 min intervals). Fluorescence intensity (total integrated intensity minus camera background) of VSVG(ts045)-GFP in the ER was measured at every time point (n=10 for each condition). The percentage of the initial fluorescence intensity is shown over time. Error bars represent mean +/- SEM; 10 different cells.

100 200 300

20

40

60

80

100

Time (minutes relative to temp shift)

Per

cent

age

of

init

ial E

RFl

uore

scen

ce (V

SV

G-G

FP)

TFG siRNAControl

a

b


C. elegans TFG-1PB1 CC

486

H. sapiens TFGPB1 CC

400

PQ rich

PQ rich




Figure S4 The region of Sec16B necessary for its targeting to ER exit sites forms a homo-oligomer and targeting of NTRK1 to ER exit sites is necessary for cell transformation. (a) The results presented in both panels are representative of at least three individual experiments performed. Recombinant polyhistidine-tagged Sec16B (amino acids 34-234) was expressed and purified from E. coli extracts using nickel resin. The protein was fractionated on a 10-30% glycerol gradient (top), and an S-value was calculated based on the location of characterized standards run on a parallel gradient. A Coomassie stained gel of the peak elution fractions after fractionation of the protein on a Superose 6 gel filtration column is shown (bottom). In both cases, the intensities of the bands corresponding to the amino-terminus of Sec16B (amino acids 34-234) were measured to identify the peak elution fraction, which was used to calculate

either a Stokes radius or sedimentation value, depending on the experiment. (b) Swept field confocal optics was used to image hTERT-RPE1 cells that had been stably transfected with either GFP:TFG(N)-NTRK(C) or GFP:Sec16B(1-

791)-NTRK(C) (n=13 each). Scale bar, 10 mm. (c) The bar graph shows the number of colonies (at least 0.01 cm2 in diameter) that grew in a 6 cm2 region of individual tissue culture plates seeded with NIH3T3 cells transfected with either GFP, GFP:GST-NTRK1, GFP:TFG-NTRK1, or GFP:Sec16B-NTRK1. In each case, 500 cells, as calculated using a hemocytometer, were seeded onto a monolayer of control NIH3T3 cells. Colony formation was scored after 14 days. Error bars represent mean +/- SEM; 3 different experiments. **p < 0.01 compared with results obtained using cells expressing GFP:GST-NTRK1, calculated using a paired t test.

87Glycerol Gradient (10-30%)

6521 3 4LoadMW

S-Value = 4.04 S

25 kD

37kD

Sec16B

Stokes Radius = 80.3 Å

96 10 11 12Superose 6 Gel Filtration

13 14MW

20 kD

37kD

20 kD

(34-234)

157 8

25 kD Sec16B(34-234)

a b

GFP:TFG-NTRK

hTERT-RPE1

GFP:Sec16B-NTRK

c

20

10

15

25

5Co

loni

es (p

er 6

cm

)2

GFP:

TFG-NTRK

GFP:

Sec16

B-NTRK

GFP:

GST-NTRK

GFP

**

**





Figure S5 TFG-1 localizes to ER exit sites in several distinct tissues. (a-c) Immunogold-electron microscopy was used to define the localization of C. elegans TFG-1 in the intestine (a), hypodermis (b) and muscle (c) tissues.

Immunoreactive gold particles bound to Tfg-1 are highlighted by arrowheads. In each case, Tfg-1 is found in a cloud-like region that spreads from ER exit sites to the ERGIC/Golgi membranes. Bars, 100 nm.

TFG-1 (intestinal cell)a b TFG-1 (hypodermal cell)

c TFG-1 (muscle cell)





Figure S6 Full scan data of immunoblots and Coomassie stained gels. In several cases, the nitrocellulose used for immunoblot was cut into strips to minimize the amount of antibody necessary for analysis. Scans of entire nitrocellulose strips are shown.

Fig. 1c

IB: SEC-1675kD

100kD150kD

Fig. 1b

IB: TFG-1

50kD75kD

100kD

IB: TFG-1

75kD100kD150kD

50kD

25kD

37kD

Fig. 1d

75kD100kD150kD

Fig. 1e

IB: SEC-16

Fig. 1e

Coomassie

75kD100kD150kD

50kD

25kD

37kD

20kD

250kDFig. 3e

IB: CAR-1

50kD

37kD

25kD

100kD150kD

Fig. 3e

IB: SEC-16

Fig. 3e

75kD

50kD

IB: TFG-1

Fig. 4a

IB: SEC-16

100kD150kD250kD 75kD

50kD

IB: TFG-1

Fig. 4a

Fig. 4b

75kD100kD150kD

50kD

25kD

37kD

20kD

Coomassie

75kD100kD150kD

50kD

25kD

37kD

20kD

Fig. 4b

Coomassie

75kD100kD150kD

50kD

25kD

37kD

20kD

Fig. 4c

Coomassie

75kD100kD150kD

50kD

25kD

37kD

20kD

Fig. 4c

Coomassie

75kD100kD150kD

50kD

25kD

37kD

20kD

Fig. 4d

Coomassie

75kD100kD150kD

50kD

25kD

37kD

20kD

Fig. 4d

Coomassie

250kD

Fig. 4e

IB: SEC-16

100kD150kD

IB: SEC-13

Fig. 4e50kD

25kD

37kD50kD

25kD

37kD

IB: SEC-13

Fig. 4e

IB: HGRS-1

75kD100kD150kD

Fig. 4e

IB: HGRS-175kD

100kD150kD

Fig. 4e250kD

50kD37kD

75kD100kD150kD

Fig. 6d

IB: TFG

75kD100kD150kD

50kD

25kD

37kD

20kD

250kDFig. 6f

Coomassie

75kD100kD150kD

50kD

25kD

37kD

20kD

250kDFig. 6e

Coomassie

75kD100kD150kD

50kD

25kD

37kD

250kD

Fig. 6g

IB: TFG

75kD100kD

150kD

50kD

37kD

250kD

Fig. 6g

IB: GFP

50kD

37kD

50kD

37kD

Fig. 7e Fig. 7e

IB: ERK1/2 IB: phos-ERK1/2





Supplemental Table

Table S1 Essential SEC-16 interacting proteins identified by mass spectrometry. Table showing the seven proteins identified by solution mass spectrometry following immunoprecipitation of SEC-16 from C. elegans embryo extract (n=3), which also result in sterility following their depletion. The percent sequence coverage and molecular weight of each protein are also shown.




Supplemental Movie Legends

Movie S1 Three-dimensional reconstruction of the early secretory pathway in C. elegans oocytes. The portion of the ER harboring the ER exit site is shown in grey. Both ER exit sites are covered with a COPII coat depicted in dark grey. ER derived vesicles and the ERGIC are shown in orange. The Golgi cisternae composed of cis, medial and trans-Golgi compartments are highlighted in yellow, red, and blue, respectively. Transport vesicles surrounding the trans-Golgi cisternae are also colored in blue. In addition, at the rim of the lower cis-Golgi cisternae, a protein coat, which is most likely COPI, is also visible and shown in dark grey. The movie illustrates the assembly of the early secretory pathway and provides multiple views of its organization.

Movie S2 Tomogram of the early secretory pathway observed in the proximal gonad. Each frame of the video corresponds to a digital section of the tomogram moving along the tomographic Z-axis.

Supplemental References

1. Maine, E.M., and Kimble, J. (1989). Identification of genes that interact with glp-1, a gene required for inductive cell interactions in Caenorhabditis elegans. Development 105, 133-43.

2. Moscat, J., Diaz-Meco, M.T., Albert, A., and Campuzano, S. (2006). Cell signaling and function organized by PB1 domain interactions. Mol. Cell 23, 631-640.


Documents

TFG-1 function in protein secretion and oncogenesis et al., 2011.pdf · 2015-01-15 · ARTICLES TFG-1 function in protein secretion and oncogenesis KristenWitte1,4,AmberL.Schuh1,4,JanHegermann2,AliSarkeshik3,JonathanR.Mayers1,KatrinSchwarze2,