The Ubp3/Bre5 deubiquitylation complex modulates COPII vesicle formation

Abstract The appropriate delivery of secretory proteins to the correct subcellular destination is an essential cellular process. In the endoplasmic reticulum (ER), secretory proteins are captured into COPII vesicles that generally exclude ER resident proteins and misfolded proteins. We previously characterized a collection of yeast mutants that fail to enforce this sorting stringency and improperly secrete the ER chaperone, Kar2 (Copic et al., Genetics 2009). Here, we used the emp24Δ mutant strain that secretes Kar2 to identify candidate proteins that might regulate ER export, reasoning that loss of regulatory proteins would restore sorting stringency. We find that loss of the deubiquitylation complex Ubp3/Bre5 reverses all of the known phenotypes of the emp24Δ mutant, and similarly reverses Kar2 secretion of many other ER retention mutants. Based on a combination of genetic interactions and live cell imaging, we conclude that Ubp3 and Bre5 modulate COPII coat assembly at ER exit sites. Therefore, we propose that Ubp3/Bre5 influences the rate of vesicle formation from the ER that in turn can impact ER quality control events.


| INTRODUCTION
Vesicle formation from the endoplasmic reticulum (ER) represents a central cellular quality control checkpoint. As COPII vesicles form at the ER membrane, they select cargo for enrichment into the emerging bud. For the most part, these cargo proteins are mature, folded proteins that have been released from the abundant ER chaperones that promote their folding and assembly. Such selection for folded cargo ensures that aberrant and misfolded proteins are not released to downstream compartments, where they may trigger aggregation and interrupt cell function. However, non-selective capture of nascent proteins can also occur, in a process known as bulk flow. 1,2 Indeed, non-specific leakage of ER resident proteins necessitates an ER retrieval pathway that recognizes escaped ER residents and returns them to the ER via COPI-coated vesicles.
We previously investigated the mechanisms that contribute to ER retention using a genetic screen that identified yeast mutants that have lax ER retention. 3 The basis for the screen was detection of the abundant ER lumenal chaperone, Kar2, at the cell surface. Kar2 is normally efficiently retained within the cell by the combination of ER sorting stringency preventing capture into COPII vesicles and efficient ER retrieval via the KDEL-receptor, Erd2, to collect any escaped protein. We reasoned that if ER leakage increases, then the retrieval pathway may become overwhelmed, resulting in release of Kar2 to the cell surface, which we detected using colony immunoblotting methods. This screen resulted in the identification of 87 mutants that impact various secretory pathway functions to yield an increase in Kar2 secretion. 3 One of the top hits in our screen was the p24 family of proteins, which function as export receptors for GPI-anchored proteins (GPI-APs). Absence of p24 proteins and the resulting reduction in GPI-AP packaging into COPII vesicles has multiple effects. 4,5 First, the local membrane bending energy of the ER seems to be reduced such that the bilayer can be deformed without structural rigidity conferred by the outer COPII coat scaffolding protein, Sec13. 6 Second, the space created in the vesicle by the lack of GPI-AP capture causes an increase in the non-specific bulk flow leakage of secretory proteins, misfolded proteins, and ER residents. 7 Increased bulk flow is the likely cause of Kar2 secretion; the retrieval pathway is still intact in p24 mutants, but retrieval capacity is outweighed by ER leakage. Adding a second mutation that decreases the size of COPII vesicles to the p24 mutants reversed both Kar2 secretion and increased bulk flow phenotypes, confirming the model that steric crowding plays a central role in ER sorting stringency. 7 Finally, cells lacking p24 proteins show constitutive activation of the unfolded protein response (UPR), likely as a result of chaperone depletion. 5 Here, we sought to use the Kar2 secretion phenotype as a tool to identify novel factors that impact COPII vesicle formation and/or architecture. We performed a genome-wide screen for mutants that reverse the Kar2 secretion associated with loss of the p24 protein, Emp24. We identify the ubiquitin protease complex, Ubp3/Bre5, as factors that impact vesicle formation from the ER. Loss of either Ubp3 or Bre5 reverses Kar2 secretion in many different mutant backgrounds, and also alleviates other phenotypes associated with loss of the p24 proteins. We propose that by modulating the ubiquitination state of the COPII subunit, Sec23, Ubp3, and Bre5 influence the rate of vesicle formation at the ER, and thereby contribute to sorting stringency.

| A high-throughput screen shows multiple pathways that reverse aberrant Kar2 secretion
In order to identify factors that influence ER sorting, we used synthetic genetic array (SGA) methodology to survey the yeast genome for mutants that reverse the Kar2 secretion phenotype of a p24 mutant. An emp24Δ query strain was crossed to the yeast deletion collection, and haploid double mutants containing the emp24Δ deletion plus deletion of a second gene were recovered ( Figure 1A).
Among the 4854 deletion strains in the collection, 72 strains did not yield haploid double mutants after the SGA procedure, likely because of synthetic sick/lethal interactions. Colony size and Kar2 secretion were measured in the resulting library of 4782 haploid double mutant strains, and normalized to that of an emp24Δ trp1Δ double mutant.
This control strain had undergone the SGA procedure, so had the same resultant genetic background as the library, and showed Kar2 secretion levels similar to the emp24Δ parental strain (Table S1).
Deletion of 110 genes resulted in impaired growth of cells in the emp24Δ background ( Figure 1B). These double mutants were excluded from the Kar2 secretion analysis because growth defects confound analysis of Kar2 levels from colony overlay immunoblots.
We calculated a Kar2 reversion index, which is the log2 ratio of the Kar2 secretion signal in a haploid double mutant relative to the emp24Δ trp1Δ mutant ( Figure 1B). Considering only double mutants with normal growth, the cohort of mutants with reduced Kar2 secretion included the emp24Δ lst1Δ mutant that we previously demonstrated reversed Kar2 secretion to wild-type levels ( Figure 1B).
Analysis of the Kar2 reversion indices across the double mutant library showed a mean reversion of −0.2 ( Figure 1B, inset) with a SD of 0.2. Using 2 standard deviations (ie, -0.6) as a cutoff threshold we identified 118 mutants with Kar2 secretion levels reduced relative to that of emp24Δ cells ( Figure 1C, Table S1). In order to classify the Kar2 secretion revertants we analyzed the biological processes in which they are involved. We observed GO-term enrichment for proteins involved in ribosome function/biogenesis and translation, and in endosomal trafficking. Additional categories that were not necessarily GO-enriched included proteins involved in ER-Golgi function, cell wall function and cell polarity, and transcription ( Figure 1D, Table S1).
Because of our interest in ER quality control, we focused on the ER-Golgi function hits, and specifically on two members of a known complex, Ubp3 and Bre5.
2.2 | The absence of the Ubp3/Bre5 complex reverts Kar2 secretion as a result of a reduction in ER-to-Golgi transport Ubp3 and Bre5 form a deubiquitylation complex that acts directly on Sec23 to regulate its turnover. 8,9 To confirm the results from our SGA screening, ubp3Δ and bre5Δ deletions were introduced by PCRmediated integration into an emp24Δ strain, and Kar2 secretion measured by colony immunoblot. Indeed, Kar2 secretion was reversed by both the UBP3 and BRE5 deletions ( Figure 2A). We next asked if the absence of the Ubp3/Bre5 complex could also rescue other phenotypes associated with Emp24 loss, including induction of the unfolded protein response (UPR), and bypass-of-Sec13. 4,5,10 To measure UPR activation, we introduced a UPRE-LacZ reporter plasmid 11 into wildtype and mutant strains, and quantified β-galactosidase activity. As previously reported, the emp24Δ mutant had high levels of constitutive UPR, which was reversed by additional deletion of either UBP3 or BRE5 ( Figure 2B). This reversal suggests that UPR activation in emp24Δ cells is primarily caused by increased export of ER chaperones, such that reversal of leakage conferred by loss of Ubp3/Bre5 alleviates ER stress, resulting in reduced activation of UPR. In contrast, UPR activation in the emp24Δ background was unaffected by deletion of two other reversion mutants, vps10Δ, and lst1Δ ( Figure 2B), although both the vps10Δ and lst1Δ single mutants also have constitutive UPR (Supplemental Figure 1).
Emp24 loss creates permissive conditions that result in a bypassof-sec-thirteen (bst) phenotype, in which the normally essential COPII subunit Sec13 can be deleted. 4,6,10 We therefore tested whether Sec13 bypass was reversed by the absence of the Ubp3-Bre5 complex. We combined a sec13Δ emp24Δ strain with deletions in UBP3 or BRE5 and tested for viability on media containing 5-FOA, which counter-selects for SEC13 on a URA3-marked plasmid. The bst phenotype of the emp24Δ strain was indeed reversed by UBP3 and BRE5 deletions ( Figure 2C). This observation supports the idea that the Ubp3/Bre5 complex promotes COPII vesicle formation, such that the absence of cargo during vesicle formation is no longer sufficient to provide viability when Ubp3/Bre5 function is lacking.
One trivial explanation for the multiple phenotypic reversions we observed upon loss of Ubp3/Bre5 is that defective ER export of GPI-APs in the emp24Δ background is restored. However, when we examined the maturation of the GPI-AP, Gas1, in emp24Δ and double mutant strains, we observed that GPI-AP transport is not rescued ( Figure 2D). Finally, we asked whether the reversion effects of Ubp3/ Bre5 loss were unique to the emp24Δ condition or whether their deletion would similarly revert other Kar2 secretion mutants. We used the SGA method to introduce ubp3Δ and bre5Δ deletions into a set of 49 Kar2 secretion mutants that we previously identified, 3 The absence of the Ubp3/Bre5 complex reverts Kar2 secretion as a result of a reduction in ER-to-Golgi transport. A, Extracellular Kar2 was detected by colony overlay immunoblot in the indicated yeast strains. Serial dilutions of cells at stationary phase were spotted onto YPD plates. After 5 hours growth, colonies were overlaid with nitrocellulose and secreted Kar2 detected with Kar2-specific antibodies. B, Exponentially growing wild-type and mutant cells carrying a UPRE-LacZ reporter gene were lysed, and ß-galactosidase activity measured using ONPG as a substrate. Error bars depict mean and SD of at least three independent experiments. Statistical test was a one-way ANOVA with Dunnett's multiple comparisons test. C, The indicated strains were grown to exponential phase, serially diluted and spotted on media containing 5-fluoroorotic acid (5-FOA) to counter-select for the wild-type pSEC13::URA3 plasmid. D, Wild type, emp24Δ, emp24Δ ubp3Δ and emp24Δ bre5Δ cells were subjected to pulse-chase analysis with [ 35 S] methionine. Gas1 was immunoprecipitated from cell lysates at the indicated times. F I G U R E 4 Ubp3/Bre5 mutants have secretory defects but normal COPII levels. A, COPII protein levels were analyzed in cell lysates prepared from equal numbers of exponentially growing cells of the indicated strains using immunoblot. Asterisk indicates unspecific band. B, Pulse-chase analysis was used to determine the half-lives of Sec23 and Sec24 in the indicated strains. After starvation cells were labeled with [ 35 S] methionine and chased for the indicated times. Sec23 and Sec24 were precipitated from cell lysates using specific antibodies. Immunoprecipitated proteins were detected by SDS-PAGE and autoradiography. C, Gas1 and CPY maturation were examined in wild-type, ubp3Δ and bre5Δ strains by pulse-chase with [ 35 S] methionine. Gas1 and CPY were immunoprecipitated from lysates at the indicated times and were detected by SDS-PAGE and autoradiography. The percent maturation from precursor (p) to mature (m) forms was quantified and plotted using Prism software. Error bars represent SD; n = 3 coat. Deletion of UBP3 in the context of various SEC24 alleles showed distinct effects. Cargo binding mutants had variable phenotypes in the absence of Ubp3, with the A-and D-site mutants reduced in viability but the C-site mutant unaffected ( Figure 3B). The sec24-m11 mutant, which impacts coat assembly and turnover 14 was also inviable when UBP3 was deleted ( Figure 3B). Our interpretation of these genetic interactions is that Ubp3/Bre5 are required to maintain vesicle formation in the context of partially dysfunctional COPII coat proteins.

| Loss of Ubp3/Bre5 affects COPII coat dynamics
Because the Ubp3/Bre5 complex has previously been defined as a modulator of Sec23 stability, we measured Sec23 steady state levels ( Figure 4A), and Sec23 turnover ( Figure 4B) in the presence and absence of Ubp3 and Bre5. In our hands, loss of Ubp3/Bre5 did not affect the stability of Sec23 or other COPII proteins. We also did not detect the ubiquitinated form of Sec23, which was previously observed. 9 These differences may reflect the different antibodies used in the different studies. In pulse-chase experiments to measure secretion rates for two proteins, Gas1 and CPY, we recapitulated previous observations that secretion was modestly impaired in both the ubp3Δ and bre5Δ mutants, consistent with reduced COPII abundance ( Figure 4B).
Previous morphological analysis of ubp3Δ and bre5Δ mutant strains showed accumulation of ER membranes consistent with reduced COPII vesicle formation. 9 We sought to more directly interrogate COPII function by visualizing COPII marked ER exit sites. Deletion of UBP3 or BRE5 had no obvious effect on the localization of Sec13-sfGFP at ER exit sites ( Figure 5A), although the number of ERES was reduced in both the ubp3Δ and bre5Δ strains ( Figure 5B).
Moreover, the total fluorescent signal for Sec13-sfGFP was reduced in the ubp3Δ and bre5Δ strains relative to wild-type ( Figure 5C), suggesting that coat stability is reduced in these cells. Similarly, Sec13-sfGFP pixel fluorescence intensity variance was reduced in the ubp3Δ and bre5Δ mutants ( Figure 5D), which indicates a slighly more homogeneous distribution between ERES and a cytoplasmic pool. 14,15 Together, these fluorescence measurements suggest that abundance and intracellular distribution of the COPII coat is subtly altered in the absence of Ubp3/Bre5. Increased lifetime of COPII proteins at ER exit sites is consistent with a reduction in the number of vesicles released from the ER. 14  In focusing on likely regulators of COPII function, we further characterized the Ubp3/Bre5 deubiquitination complex, which acts directly on Sec23 to manage its turnover. 9 In the context of lax protein quality control driven by loss of p24 proteins, additional loss of Ubp3/Bre5 most likely reduces Kar2 secretion simply by reducing the frequency of vesicle formation. The slow-down in Gas1 and CPY maturation we observed in the ubp3Δ and bre5Δ strains is suggestive of a general reduction in ER export efficiency. If fewer vesicles are made, COPII turnover at ER exit sites will be reduced, again consistent with the phenotype we observed in the absence of Ubp3/Bre5. Presumably the rate of vesicle formation in the ubp3Δ/bre5Δ mutants is not reduced to such an extent that sufficient cargo accumulates to trigger the UPR, which is not constitutively active in these strains. We propose that a reduction in vesicle formation associated with loss of Ubp3/Bre5 can explain the reversal of all phenotypes associated with p24 loss. With the number of vesicles reduced, cargo will out-compete ER residents for access to those vesicles, thereby reducing bulk flow and restoring ER sorting stringency. With vesicles again full of cargo proteins, the membrane bending energy at ER exit sites will again be high, requiring rigidity conferred by Sec13 and thus reversing the bst phenotype. Finally, if Kar2 leakage is prevented, then the likelihood of a constitutive ER stress condition is reduced, again consistent with the observed reduction in UPR. Therefore, we propose that Ubp3/ Bre5 play an important role in modulating the rate of vesicle formation from the ER that in turn can impact ER quality control events.

| DISCUSSION
Reduction in vesicle formation would explain the genetic interactions we and others have observed for the ubp3Δ and bre5Δ strains.
Bre5 was first linked to ER export via a genetic interaction with the COPII subunit Lst1/Sfb3. 9 We report multiple synthetic interactions with different alleles of Sec23 and Sec24 that impact distinct functions within the COPII coat. Moreover, high throughput analyses have showed numerous negative genetic interactions between ubp3Δ and bre5Δ and proteins that function in a variety of secretory events (thebiogrid.org 17,18

| Strains and plasmids
Yeast cultures were grown at 30 C in standard rich medium (YPD: 1% yeast extract, 2% peptone, and 2% glucose) or synthetic complete medium (SC: 0.67% yeast nitrogen base and 2% glucose supplemented with amino acids as needed). Strains (Table S2) were made by PCRbased integration of auxotrophic or drug-resistance markers. The plasmids (Table S3) were generated using standard cloning techniques.

| Synthetic genetic array
SGA was performed as previously described 22

| Serial dilutions
To observe growth in the absence of Sec13 and growth of strains carrying Sec23 or Sec24 mutations, cells were grown to saturation at 30 C in media lacking uracil. Serial dilutions (1:10) were made from the saturated cultures and spotted onto solid media without (control) or with 0.1% 5-Fluoroorotic acid (5-FOA, BioVision) to counter-select for the corresponding URA3 plasmid (pSEC13-URA3, pSEC23-URA3 or pSEC24-URA3). Plates were scanned at day 2 or day 3 after spotting and growth at 30 C.

| Western blot analysis
Protein extracts prepared by alkaline lysis of exponentially growing yeast were separated by SDS-PAGE. Sec23, Sec24, Sec13, and Sec31 proteins were detected with the corresponding specific antibodies (provided by Randy Schekman, UC Berkeley). Steady state levels of proteins were then detected with HRP-conjugated goat-anti-rabbit antibodies followed by ECL detection.

| Pulse chase analysis
Pulse-chase experiments were used to monitor intracellular transport of Gas1 and CPY and to monitor Sec23 and Sec24 degradation. These experiments were performed as previously described. 23 Briefly, strains were grown to mid-log phase at 30 C, starved for 15 minutes, and

| Fluorescence microscopy and FLIP assays
Sec13-sfGFP imaging was performed in cells grown to mid-log phase at 30 C in minimal media lacking tryptophan. Images were taken on a Nikon Ti2 with a 100×/1.49 NA Oil (TIRF) objective and a scientific complementary metal-oxide-semiconductor (sCMOS) camera.
FLIP was used to measure coat dynamics of Sec13-sfGFP. Imaging was performed in cells grown to mid-log phase at 30 C in minimal media lacking tryptophan. Images were taken on an Andor revolution spinning disk microscope with a 40×/1.3NA oil immersion objective and a EMCCD camera. Images were collected using the Andor iQ3 software. A small region of interest was repeatedly photobleached, and the fluorescence intensity of the whole cell was measured for a loss of signal, representing protein that had diffused into the bleaching area. Fluorescence intensity was measured using Fiji and statistical analysis was performed with Prism 7.0 (GraphPad Software).