TFG-1 function in protein secretion and oncogenesis

Abstract

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

Figure 1. C. elegans TFG-1 interacts with the ER exit site component SEC-16

(a) Schematic representation of human and C. elegans Sec16 isoforms. The central conserved domain (CCD) is highlighted in each protein. (b) SEC-16 was immunoprecipitated from C. elegans embryo extract and blotted with α-TFG-1 antibodies (n=3). A mock IP was conducted in parallel using rabbit IgG. (c) TFG-1 was immunoprecipitated from C. elegans embryo extract and blotted with α-SEC-16 antibodies (n=3). A mock IP was conducted in parallel using rabbit IgG. (d) GST alone and GST-tagged full length SEC-16 were immobilized on glutathione agarose beads, which were incubated with an extract generated from E. coli expressing recombinant TFG-1. Following a series of washes, proteins were eluted using reduced glutathione, separated by SDS-PAGE, and either stained using Coomassie (top) or immunoblotted using TFG-1 antibodies (bottom). (e) Polyhistidine-tagged full length and truncated forms of TFG-1, either encoding amino acids 1–195 (TFG-1^(N)) or 196–486 (TFG-1^(C)), were purified from E. coli onto nickel affinity resin and incubated with freshly prepared whole worm extract (n=3). Imidazole eluted proteins were separated by SDS-PAGE, stained with Coomassie (top), and blotted with α-SEC-16 antibodies (bottom). For each figure, uncropped scans of all gels and immunoblots are provided in Supplementary Fig. S6.

Figure 2. TFG-1 localizes to ER exit sites that are juxtaposed to the Golgi

(a) Dissected C. elegans gonads were fixed and stained using Cy2-labeled α-TFG-1 and Cy3-labeled α-SEC-16 antibodies (n=8). Both individual and merged images of proximal oocytes with TFG-1 in green and SEC-16 in red are shown (Bar, 10 μm). The upper right image is the boxed area in the panel below magnified 6× (Bar, 2 μm). Also shown is a schematic of the C. elegans reproductive system, which includes a syncytial stem cell niche in the distal gonad (boxed region in light green) and proximal oocytes that have undergone cellularization (boxed region in dark green). (b and c) Lowicryl sections of C. elegans oocytes were stained with antibodies against TFG-1 or a combination of TFG-1 and SEC-13 antibodies. Arrows highlight Golgi cisternae. Large arrowheads point out 15 nm gold particles associated with immunoreactive TFG-1, and small arrowheads highlight 5 nm gold particles associated with SEC-13. Bars, 100 nm. An inset is provided in panel c to clearly show the distribution of 5 nm particles at higher magnification (Inset bar, 15 nm). In addition, a three-dimensional reconstruction of TFG-1 immunolocalization is shown. The image was generated using the software Reconstruct from serial 50 nm thin sections. Vesicles were reconstructed using the sphere setting, and all other components (ER, ERGIC, coats, Golgi stacks) were generated using the Boissonnat surface setting. Light grey: ER; dark grey: COPII coat; orange: ER-derived transport vesicles and ERGIC; red, green and blue: Golgi cisternae; from cis to trans, respectively. (d) An electron micrograph illustrating two ER exit sites and adjacent Golgi complexes in the proximal most oocyte of an animal following high pressure freezing and freeze substitution (Bar, 500 nm). On the right is a three-dimensional reconstruction of the same pair of Golgi complexes and associated ER exit sites. The ER exit sites are surrounded by vesicles that fuse to form the ERGIC. Light grey: ER; dark grey: COPII coat; orange: ER-derived transport vesicles and ERGIC; yellow, red and blue: Golgi cisternae; from cis to trans, respectively.

Figure 3. TFG-1 regulates SEC-16 levels on ER exit sites

(a) In the proximal gonad, a 300 nm section of the early secretory pathway (ER exit sites, ERGIC, and Golgi) was analyzed by electron tomography. ER exit sites are highlighted by arrowheads. On the left are individual sections from the tomographic stack. On the right are two orthogonal views of the tomogram following three-dimensional reconstruction. Light grey: ER; black: COPII coat; orange and yellow: ER-derived transport vesicles and ERGIC; green, red, blue: Golgi cisternae; diffuse grey: not further resolvable matrix. Bar, 100 nm. (b) An electron micrograph illustrating ER exit sites and adjacent Golgi complexes in the distal gonad following high pressure freezing and freeze substitution. An arrowhead highlights the presence of budding vesicle from smooth ER (Bar, 100 nm). Below is a three-dimensional construction of the same Golgi complexes and associated ER exit sites. (c) Dissected gonads from control, TFG-1 depleted, and SEC-16 depleted animals were fixed and stained using Cy2-labeled α-TFG-1 and Cy3-labeled α-SEC-16 antibodies. Individual and merged images of the distal gonad with TFG-1 in green and SEC-16 in red are shown (Bar, 10 μm). (d) Fluorescence intensity of SEC-16 in the distal gonad was measured in control and TFG-1 depleted animals, and intensities were segregated into low, medium and high thresholds. To establish individual thresholds, a histogram of fluorescence intensities was equally divided into three regions, and the number of ER exit sites within each area was calculated. The bar graph indicates the percentage of all ER exit sites that fall into a specific threshold. For each condition, at least 1000 unique ER exit sites were examined. Error bars represent mean +/− SEM; 10 different animals. **p < 0.01 compared with control, calculated using a paired t test. (e) Western blots of extracts prepared from animals depleted of TFG-1 by RNAi (n=3). Serial dilutions of extracts prepared from control animals were loaded to quantify depletion levels. Blotting with α-CAR-1 antibodies was performed to control loading.

Figure 4. The amino-terminus of TFG-1 mediates it oligomerization

The results presented in each panel are representative of at least three individual experiments performed. In all cases, the intensities of each band were measured to identify the peak elution fraction, which was used to calculate either a Stokes radius or sedimentation value, depending on the experiment. (a) Western blots using SEC-16 antibodies (top) or TFG-1 antibodies (bottom) of C. elegans embryo extract fractionated on a Superose 6 gel filtration column. The peaks corresponding to SEC-16 and TFG-1 partially overlap. A Stokes radius was calculated for each protein based on comparison with the elution profiles of known standards. (b-d) Recombinant polyhistidine-tagged TFG-1 or fragments of TFG-1 described in Figure 1e were expressed and purified from E. coli extracts using nickel resin. A Coomassie stained gel of the peak elution fractions after fractionation of the recombinant proteins on a Superose 6 gel filtration column are shown (top). Proteins were fractionated on a 10–30% glycerol gradient (bottom), and S-values were calculated based on the location of characterized standards run on a parallel gradient. (e) Western blots of control and TFG-1 depleted C. elegans whole worm extracts fractionated on a Superose 6 gel filtration column and probed with SEC-16 antibodies (top) or SEC-13 antibodies (middle panels). Fractionation of HGRS-1, a component of the ESCRT-0 complex, was examined in both control and TFG-1 depleted conditions (bottom panels), to ensure gel filtration profiles were directly comparable. Stokes radii were calculated for each protein based on comparison with the elution profiles of known standards.

Figure 5. TFG-1 is required for COPII recruitment and protein secretion

(a) Dissected C. elegans gonads were fixed and stained using Cy2-labeled α-SEC-13 and Cy3-labeled α-SEC-16 antibodies (n=15). Merged images of the distal gonad with SEC-13 in green and SEC-16 in red are shown on the left (Bar, 10 μm). Panels to the right are magnified 5× views of the boxed area in the adjacent panel (Bar, 2 μm). (b) Bar graph showing the average ratio of SEC-13 to SEC-16 fluorescence intensities in control and TFG-1 depleted animals. For each condition, at least 250 unique ER exit sites in the distal gonad were examined. Error bars represent mean +/− SEM; 6 different animals. No statistically significant difference was observed, based on a calculation using a paired t test. (c) Swept field confocal optics were used to image anesthetized control (n=15) and TFG-1 depleted (n=15) adult animals expressing GFP:SNB-1 and mCherry:PH. Scale bar, 10 μm. (d) Electron micrographs illustrating the early secretory pathway in the proximal most oocyte of control (left) and TFG-1 depleted (right) animals following high pressure freezing and freeze substitution (Bar, 100 nm). Arrowheads highlight ER exit sites. Below each micrograph is a three-dimensional reconstruction of the same regions. Light grey: ER; dark grey: COPII coat; orange: ER-derived transport vesicles and ERGIC; green, red and blue: Golgi cisternae; from cis to trans, respectively.

Figure 6. Human TFG functions at ER exit sites and binds to Sec16

(a) Swept field confocal optics were used to image HeLa cells that had been transiently transfected with GFP:TFG and mCherry:Sec16B (n=42). Representative color overlays of mCherry:Sec16B (red) and GFP:TFG (green) are shown. Scale bar, 10 μm. (b) Swept field confocal optics were used to monitor the recovery of GFP:TFG after photobleaching (n=15). A 3× magnified view of the boxed region where GFP:TFG was 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 average percentage of GFP:TFG and mCherry:Sec16B fluorescence recovered as a function of time in seconds relative to the bleach (error bars represent means +/− SEM for each time point; n=15 different cells for each fluorescent fusion protein). (d) Western blots of HeLa cell extract fractionated on a Superose 6 gel filtration column (n=3). A Stokes radius was calculated for human TFG based on comparison with the elution profiles of known standards. (e, f) A GST-tagged, truncated form of human TFG, amino acids 1–193 was expressed and purified from E. coli extracts using glutathione agarose (n=3), and the GST tag was subsequently cleaved using Prescission Protease prior to loading onto a gel filtration column or glycerol gradient. A Coomassie stained gel of the peak elution fractions after fractionation of the recombinant protein, referred to as TFG^(N), on a Superose 6 gel filtration column are shown (e). The protein was also fractionated on a 10–30% glycerol gradient (f), and an S-value was calculated based on the location of characterized standards run on a parallel gradient (n=3). (g) Antibodies directed against mCherry were used to immunoprecipitate mCherry:Sec16B from HeLa cells transiently transfected with GFP:TFG or a GFP fusion to the amino-terminus of TFG referred to as GFP:TFG^(N) (n=3). Isolated proteins were separated by SDS-PAGE and blotted with α-TFG (left) and α-GFP antibodies (right).

Figure 7. Targeting of the NTRK1 kinase domain to ER exit sites is sufficient to activate NTRK1-mediated downstream signaling

(a–c) Swept field confocal optics were used to image HeLa cells that had been transiently transfected with mCherry:Sec16B and GFP fusions to either the amino terminus of TFG referred to as GFP:TFG^(N) (n=18), the transmembrane 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 to the oncogene characterized previously (21). Representative color overlays of mCherry:Sec16B (red) and GFP fusions (green) are shown. Scale bar, 10 μm. (d) Bar graph showing the percent co-localization between the GFP fusions described above and mCherry:Sec16B (error bars represent means +/− SEM for each condition; n=15 different cells for each condition and at least 800 unique ER 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 kinase domains (as indicated) were separated by SDS-PAGE and blotted using a phospho-specific ERK1-ERK2 antibody (top) and a pan-ERK1-ERK2 antibody (bottom).