Supplementary Information for 2005-08-09311C
SI Files
The following figures are in Photoshop jpeg format:
Figure S1.jpg 1.2 MB
Figure S2a.jpg 1.1 MB
Figure S2b.jpg 1.1 MB
Figure S2c.jpg 2.5 MB
Figure S3.jpg 336 KB
Figure S4.jpg 332 KB
Figure S5.jpg 1.1 MB
Figure S6.jpg 892 KB
The following movies are in QuickTime format:
Movie 1a (Sec7).mov 19.9 MB
Movie 1b (Sec7).mov 16 MB
Movie 2a (Sys1-Sec7).mov 9.3 MB
Movie 2b (Sys1-Sec7).mov 7.4 MB
Movie 3a (2-color).mov 14.3 MB
Movie 3b (2-color).mov 9.2 MB
Movie S1a (CM).mov 2 MB
Movie S1b (SC).mov 424 KB
Movie S2 (z-Stacks).mov 13.3 MB
Movie S3a (Vrg4).mov 14.7 MB
Movie S3b (Vrg4).mov 11.3 MB
The following tables are in pdf format:
Table S1.pdf 36 KB
Table S2.pdf 28 KB
Titles and descriptions are given below.
SUPPLEMENTARY METHODS
Yeast strains and plasmids. All experiments were performed with the haploid S.cerevisiae strain JK9-3d, which has the genotype leu2-3,112, ura3-52, rme1, trp1, his41. Yeast were grown in rich glucose medium (YPD) or minimal glucose medium (SD)2.
A strain that overexpresses SEC7GFP mildly (~3fold) has the chromosomal SEC7 gene replaced with a chimeric gene encoding Sec7p tagged at its Cterminus with three tandem copies of GFP, and also has an identical SEC7GFP fusion gene expressed from the ACT1 promoter in an integrating plasmid. Replacement of the endogenous SEC7 gene with the SEC7GFP fusion gene was done using plasmid pSSEC7-mGFPx3, which was constructed as follows. The pmGFP13 plasmid, which encodes three tandem copies of a monomeric GFP, was constructed in the same way as pEGFP133,4 except that the monomerizing A206K mutation5 was introduced into the EGFP gene. pUSEURA36 was digested with BamHI and EagI, and a BamHI-NotI fragment from pmGFP-13 was inserted to create pSSEC7mGFPx3. This plasmid was linearized with SpeI, and pop-in/pop-out gene replacement was performed as described3. Integration of a second SEC7GFP gene under control of the ACT1 promoter was done using plasmid YIplac204-A/C-SEC7-mGFPx3, which was constructed as follows. YIplac2047 was digested with HincII and HindIII, and the CYC1 transcription terminator8 was inserted as an EcoRV-HindIII fragment. The ACT1 promoter was then amplified by PCR and inserted between the NdeI and EcoRI sites, creating YIplac204A/C. Into this plasmid, a full-length SEC7GFP fusion gene containing the triple-monomeric GFP cassette was inserted between the Asp718I and XbaI sites. The resulting plasmid was linearized with Bsu36I for integration at the TRP1 locus.
A strain that overexpresses GFP-VRG4 mildly (~2fold) has the chromosomal VRG4 gene replaced with a chimeric gene encoding Vrg4p tagged at its Nterminus with GFP, and also has a CEN plasmid containing an identical GFPVRG4 fusion gene expressed from the VRG4 promoter. Replacement of the endogenous VRG4 gene with the GFPVRG4 fusion gene was done using plasmid YIplac211-sGFP-VRG4, which was constructed as follows. YIplac2117 was digested with EcoRI and HindIII, and an EcoRI-HindIII fragment from YCplac33sGFP-VRG4 (see below) was inserted to create YIplac211sGFPVRG4. This plasmid was linearized with HpaI for pop-in/pop-out gene replacement. Expression of a second copy of GFPVRG4 was done using plasmid YCplac33-sGFP-VRG4, which was constructed as follows. The full-length VRG4 gene, including 5’ and 3’ UTR’s, was amplified by PCR and inserted between the EcoRI and HindIII sites of YCplac337. A gene (kindly provided by Pam Silver) encoding the sGFP variant9 was inserted after the second codon of VRG4 to create YCplac33-sGFP-VRG4.
The GFP-VRG4/SEC7-DsRed strain was derived from the GFP-VRG4 strain. The chromosomal SEC7 gene was replaced with a chimeric gene encoding Sec7p tagged at its Cterminus with six tandem copies of DsRed-Monomer (also known as DsRed.M1), which has a low tendency to alter the localization of fusion partners (D.E.S., unpublished observations). In addition, the same SEC7DsRed gene was expressed from the TPI1 promoter in an integrating plasmid, resulting in strong (~12fold) overexpression. The plasmids for gene replacement (pSSEC7DsRed.M1x6) and overexpression (YIplac204T/CSEC7DsRed.M1x6) were made in the same way as pSSEC7mGFPx3 and YIplac204A/CSEC7mGFPx3, except that the triple-GFP cassette was replaced with a cassette containing six tandem copes of the DsRed.M1 gene, and the ACT1 promoter was replaced with the stronger TPI1 promoter.
The SYS1GFP/SEC7-DsRed strain was derived from a strain constructed as above to overexpress SEC7-DsRed. The chromosomal SYS1 gene was then replaced with a chimeric gene encoding Sys1p tagged at its Cterminus with six tandem copies of monomeric GFP. In addition, the SYS1-GFP gene was expressed from the ACT1 promoter in an integrating plasmid, resulting in strong (~15-fold) overexpression. Replacement of the endogenous SYS1 gene with a SYS1GFP fusion gene was done using plasmid YIplac211-Sys1-GFPx6, which was constructed as follows. The full-length SYS1 gene, including 5’ and 3’ UTR’s, was amplified by PCR and inserted between the EcoRI and HindIII sites of YIplac2117. A gene encoding six tandem copies of monomeric GFP4 was then inserted in place of the stop codon. The resulting plasmid was linearized with BglII for pop-in/pop-out gene replacement. Overexpression of SYS1-GFP was done using plasmid YIplac128-A/C-Sys1-GFPx6, which was constructed as follows. YIplac128-A/C was constructed from YIplac1287 as described above for YIplac204-A/C. Into this plasmid, a full-length SYS1-GFP fusion gene containing the hexa-monomeric GFP cassette was inserted between the EcoRI and XbaI sites. The resulting plasmid was linearized with EcoRV for integration at the LEU2 locus.
Immunofluorescence microscopy and immunoblotting. Immunofluorescence microscopy was performed as previously described3. Rabbit polyclonal anti-Sec7p (a gift of Alex Franzusoff) was affinity purified and used at 1:100 dilution. Mouse monoclonal anti-GFP antibody (a mixture of clones 7.1 and 13.1; Roche) was used at 5µg/ml.
The levels of Sec7p overexpression were estimated by immunoblotting10 using an anti-GFP polyclonal antibody (Abcam; #6556). A strain expressing SEC7GFP as a gene replacement was assumed to contain the wild-type level of Sec7p, and was compared to strains that also expressed a second copy of SEC7GFP from either the ACT1 promoter (~3fold overexpression) or the TPI1 promoter (~12-fold overexpression). The level of Sys1p overexpression (~15fold) was estimated using a similar approach. The level of Vrg4p overexpression was assumed to be ~2fold because the second copy of GFPVRG4 was expressed from the VRG4 promoter on a low-copy CEN plasmid.
SUPPLEMENTARY TABLES AND FIGURES
Table S1. Duration of labeling of cisternae with Sec7p-GFP or GFPVrg4p.
We analyzed ten Sec7p-GFP movies, labeled A–J, where D corresponds to Movies 1a and 1b. (Six cisternae from Movies 1a and 1b were used for quantitation, but for clarity only four of them are marked in the movies and displayed in Figure 1.) We also analyzed seven GFP-Vrg4p movies, labeled A–G, where B corresponds to Movies S3a and S3b. For each movie, we tracked 1–6 cisternae that could be unambiguously resolved from all of the other cisternae in each zstack of the 4D dataset. The duration of labeling with the GFP-tagged marker protein was defined as the interval between the first and last time points at which the cisterna was visible. The average duration of labeling for each marker was estimated by summing all 29 values for the Sec7pGFP-labeled cisternae, or all 15 values for the GFPVrg4p-labeled cisternae. SEM, standard error of the mean.
Table S2. Early Golgi cisternae consistently become late Golgi cisternae.
We made four separate 10-min 4D movies of cells from the strain containing both GFPVrg4p and Sec7pDsRed, and analyzed the 2D projections and 4D datasets to record all instances in which a green fluorescent early Golgi cisterna unambiguously changed into a red fluorescent late Golgi cisterna (“Observed”). The expected number of such transitions during a 10-min period (“Expected”) was then calculated as follows. The average number of early Golgi cisternae in each cell was estimated by counting the number of GFPVrg4p-labeled cisternae at the beginning of the movie and at each 1-min interval thereafter. Based on the expectation that an early cisterna should mature after 2–3 min, we multiplied the average number of early cisternae per cell by 3.3–5 to yield the expected number of transitions in 10 min. The Observed values ranged from 44% to 91% of the Expected values. This discrepancy could indicate that some cisternae fail to mature. However, we consider it more likely that some transitions were overlooked either because the fluorescence signals during the transition phase were unusually dim, or because the relevant cisternae were too close to other cisternae to be clearly resolved.
Figure S1. Experimental strategy for distinguishing between the cisternal maturation and stable compartments models.
For simplicity, a Golgi apparatus is depicted as consisting of cis, medial, and trans cisternae. Green indicates the presence of a GFP-tagged early Golgi resident protein, and red indicates the presence of a DsRed-tagged late Golgi resident protein. The black dots represent secretory cargo proteins. With S.cerevisiae, individual cisternae are seen by fluorescence microscopy as spots. Each diagram of a Golgi apparatus with its accompanying yeast cell represents a single time point in an experiment.
a, Prediction of the cisternal maturation model. Individual cisternae should change in color from green to yellow to red. At the first time point (top), a trans cisterna (arrow) and a cis cisterna (arrowhead) are visible. At the second time point (middle), the trans cisterna is no longer visible because it has matured by exporting its late Golgi marker, while the cis cisterna has matured into a medial cisterna, which is yellow because it contains both early and late Golgi markers. At the third time point (bottom), the medial cisterna has matured into a trans cisterna, which is red because it has exported the early Golgi marker while acquiring a full complement of the late Golgi marker. During this entire process, secretory cargo proteins remain within the maturing cisternae. The recycling of Golgi resident proteins is shown as being mediated by retrograde vesicles, but other recycling mechanisms are possible. See Movie S1a for an animation.
b, Prediction of the stable compartments model. Individual cisternae should remain the same color indefinitely. At the first time point (top), a trans cisterna (arrow) and a cis cisterna (arrowhead) are visible. This pattern is unchanged at the second (middle) and third (bottom) time points because each cisterna retains its characteristic identity. Meanwhile, secretory cargo proteins move forward from one cisterna to the next. The transport of secretory cargo proteins is shown as being mediated by anterograde vesicles, but other transport mechanisms are possible. See Movie S1b for an animation.
Figure S2. Analysis method for the 4D datasets.
a, Schematic diagram of the 4D analysis. At each time point, we capture a zstack of optical sections by moving the focal plane along the zaxis from the bottom to the top of the cell. The individual optical sections can then be displayed on a computer screen. Each cisterna typically appears in two or more optical sections. When the zstack is collapsed to generate a 2D projection, cisternae that were separate may appear to be connected.
b, Representative zstack of optical sections from Movie 1a, showing cisternae labeled with Sec7pGFP. The cisterna marked with the arrow was used for quantitation. (This cisterna was not marked in Movie 1a and was not included in the edited Movie 1b, but corresponds to cisterna D1 in Table S1.) At the time point shown, the 2D projection (bottom panel) appears to show this cisterna contacting another cisterna, which is marked with an arrowhead. However, the individual optical sections reveal that these two cisternae were separate along the zaxis.
c, Representative zstack of optical sections from Movie 3a, showing cisternae labeled with GFPVrg4p and Sec7pDsRed. The Sec7pDsRed-labeled cisterna marked with the arrow is the same one that is marked with an arrow in Fig. 3 and Movie 3a. At the time point shown, the 2D projection (bottom panel) appears to show this cisterna contacting a GFPVrg4p-labeled cisterna. However, the individual optical sections reveal that these two cisternae were separate along the zaxis.
Figure S3. Overexpression of Sec7p-DsRed does not change the relative localizations of GFPVrg4p and Sec7p.
The localizations of GFPVrg4p and Sec7p-DsRed were compared using immunofluorescence with antibodies against GFP (green) and Sec7p (red). Three representative cells are shown for each strain. Scale bars, 2 µm.
a, The strain contained only GFP-Vrg4p, which was overexpressed ~2fold.
b, The strain contained GFPVrg4p, which was overexpressed ~2fold, as well as Sec7pDsRed, which was overexpressed ~12fold.
Figure S4. Strains containing tagged Vrg4p and/or tagged Sec7p exhibit normal trafficking of CPY.
Yeast strains were pulse-labeled for 10 min with 35S-Met, chased for 0–30 min as indicated at the top of the figure, and subjected to immunoprecipitation with anti-CPY antibody followed by SDS-PAGE and autoradiography. Marked are the ER-localized p1 form, the Golgi-localized p2 form, and the vacuole-localized mature form11. This analysis was done in parallel for the untagged parental wild-type strain (WT), the strain expresssing SEC7-GFP, the strain expressing GFP-VRG4, and the strain expressing both SEC7DsRed and GFPVRG4.
Figure S5. The labeling of cisternae with GFPVrg4p is similar in duration to the labeling with Sec7pGFP.
Cropped frames from Movies S3a and S3b of cells mildly overexpressing GFP-Vrg4p. Numbers indicate the time (min:s) after the movie was initiated. The arrowhead and arrows point to three GFP-Vrg4p-labeled Golgi cisternae that acquired fluorescence, increased in brightness, then progressively lost fluorescence. The bottom panels show the same images edited to display only the marked cisternae. Cisternae labeled with GFP-Vrg4p often showed more complex shape changes than those labeled with Sec7pGFP. Scale bar, 2 µm.
Figure S6. SDS-PAGE and autoradiography data for the pulse-chase analyses of afactor and carboxypeptidase Y.
These data were used to generate the graphs shown in Fig. 4. MATa cells expressing Sec7p-GFP were pulsed for 1 min with 35SMet for 1min (afactor) or 2min (CPY). At the indicated times, the chase was terminated by treating aliquots of the cells or culture medium with trichloroacetic acid. Samples were subjected to immunoprecipitation with anti-a-factor antibody or anti-CPY antibody, followed by SDS-PAGE and autoradiography.
a, Transport kinetics of afactor. In the top panel, aliquots of the cells were analyzed to measure the ER export kinetics of pro-a-factor. The prominent band is the ER-localized core pro-a-factor, which disappears upon transport to the Golgi and consequent elaboration of the oligosaccharide side chains12. In the bottom panel, aliquots of the culture medium were analyzed. The appearance of mature afactor in the medium reflects the entire process of transport from the ER to the extracellular space. Mature afactor appears as a diffuse band because of its small size. This effect, combined with the very brief pulse time, resulted in a low signal-to-noise ratio. To verify that the diffuse band was indeed mature afactor, we confirmed that it did not appear in a pulse-chase experiment with an isogenic MATa strain (data not shown).