SNARE chaperone Sly1 directly mediates close-range vesicle tethering

Abstract

The essential Golgi protein Sly1 is a member of the Sec1/mammalian Unc-18 (SM) family of SNARE chaperones. Sly1 was originally identified through remarkable gain-of-function alleles that bypass requirements for diverse vesicle tethering factors. Employing genetic analyses and chemically defined reconstitutions of ER-Golgi fusion, we discovered that a loop conserved among Sly1 family members is not only autoinhibitory but also acts as a positive effector. An amphipathic lipid packing sensor (ALPS)-like helix within the loop directly binds high-curvature membranes. Membrane binding is required for relief of Sly1 autoinhibition and also allows Sly1 to directly tether incoming vesicles to the Qa-SNARE on the target organelle. The SLY1-20 mutation bypasses requirements for diverse tethering factors but loses this ability if the tethering activity is impaired. We propose that long-range tethers, including Golgins and multisubunit tethering complexes, hand off vesicles to Sly1, which then tethers at close range to initiate trans-SNARE complex assembly and fusion in the early secretory pathway.

Figure 1.

New gain-of-function SLY1* alleles. (A) Selection used in this study. A library of SLY1* alleles was constructed by mutagenic PCR and cloned into a single-copy plasmid. The library was then transformed into a SLY1 uso1∆ strain, with USO1 provided on a balancer plasmid bearing the counterselectable URA3 marker. Ejection of pUSO1 was forced by 5-fluoroorotic acid (5-FOA). This strategy positively selects viable cells carrying dominant mutant SLY1* alleles that bypass the otherwise essential USO1 requirement. (B) Locations within Sly1 (PDB ID 1MQS) of single missense substitutions that suppress requirements for Ypt1, or for both Ypt1 and Uso1. The loop is indicated in purple, with the dashed line denoting the portion of the loop not resolved in the crystal structure. Yellow shading indicates the domain 3a helical hairpin which, by analogy to Vps33 and Munc18-1, is hypothesized to scaffold assembly of Qa- and R-SNARE trans-complexes.

Figure S1.

Some SLY1 alleles require multiple substitutions to suppress the lethal uso1∆ phenotype. (A–D) Locations of amino acid substitutions in four representative SLY1 alleles recovered in our screen. (E) Growth phenotypes show that most single substitutions are unable to suppress the loss of Uso1. Many of the same single mutants suppress the loss of Ypt1 (see Table S1). The multisite allele SLY1-5c, although retrieved in our primary screen, was unable to suppress the uso1∆ allele in secondary screening.

Figure 2.

Setup and characterization of the in vitro fusion system. (A) Reporter systems for lipid and content mixing. RPLs (reconstituted proteoliposomes) are prepared with encapsulated content mixing FRET pair, and with the membranes doped with an orthogonal FRET pair. (B) SNARE topology of the RPLs used in this study, and soluble components added to stimulate fusion. (C–E) Characterization of the system using content mixing readout. (C) Requirement for Sly1. Reactions were set up with Q- and R-SNARE RPLs, and 3% PEG. Fusion activity was monitored for 5 min, and then Sly1 was added at time = 0 (arrows) to the indicated final concentrations. Note that fusion activity is saturated at 100 nM (Sly1). (D) Requirement for tethering. Reactions were set up with Q- and R-SNARE RPLs, with the indicated final concentrations of PEG. Fusion activity was monitored for 5 min, and then Sly1 was added to a final concentration of 250 nM. Note that at 6% and 7% PEG, some Sly1-independent fusion occurs prior to Sly1 addition. (E) Effects of the SNARE disassembly machinery. Reactions were set up with Q- and R-SNARE RPLs, and with or without PEG, Sec17, Sec18, ATP, and Sly1, as indicated. Fusion was initiated by adding Sly1. For C–E, points show mean ± SEM of three independent experiments; in many cases the error bars are smaller than the symbols. Gray lines show least-squares nonlinear fits of a second-order kinetic model.

Figure 3.

Gain-of-function Sly1 mutants alleviate the tethering requirement in vitro. (A–D) Reactions were set up as in Fig. 2, with the initial mixture containing R-SNARE and Qabc-SNARE RPLs and, as indicated for each row of panels, 0 or 3% PEG, and in the absence or presence of Sec17, Sec18 (both 100 nM), and ATP (1 mM). After a 5-min incubation, wild-type Sly1 or the indicated mutants were added (arrows) at 0, 25, 100, or 400 nM to initiate fusion. Points show the mean ± SEM from three or more independent experiments; in many cases the error bars are smaller than the symbols. Gray lines show least-squares nonlinear fits of a second-order kinetic model.

Figure 4.

The Sly1 regulatory loop has a positive function in vivo. (A) AlphaFold2 rendering, showing the location of Sly1 loop replacement with engineered linkers (blue). Sequences of the linker insert designs, and growth phenotypes of the corresponding mutants are presented in Table S2. The domain 3a SNARE assembly template is shown in yellow. (B) The sly1∆loop mutant is temperature-sensitive for growth. Dilutions of liquid cultures were spotted as 10× serial dilutions onto YPD agar plates and incubated for 2 days at 30° or 37°C. These are knock-ins at the genomic SLY1 locus, in the Y8205 strain background used for SGA analysis. (C) Selected SGA results. Genes exhibiting synthetic interactions with sly1∆loop are shown. Scores indicate log_e synthetic growth defects (red) or intergenic suppression (blue). A score of −4.6 indicates a 100× synthetic growth defect. Complete SGA results are presented in Data S1. (D) Gene Ontology Overrepresentation Test of the sly1∆loop SGA dataset. Genes with log_e synthetic defect scores less than or equal to −0.5 were included in the analysis. Bars show all GO-Slim Biological Process categories with statistically significant enrichment scores (*P < 0.05; **P < 10⁻²; ***P < 10⁻⁶). P values were calculated using Fisher’s exact test and adjusted for multiple comparisons (Bonferroni’s correction; count = 732). Additional details are presented in Data S1. (E) SEC17 overproduction is toxic in cells expressing sly1∆loop. sly1∆ mutant cells were maintained with a counterselectable SLY1 balancer plasmid and transformed with single-copy plasmids bearing either SLY1 or sly1∆loop, as well as plasmids carrying SEC17 or sec17-FSMS (Schwartz and Merz, 2009). The balancer plasmid was ejected by plating dilutions on media with 5-FOA and growth was assayed after 2 days of growth at 30°C.

Figure 5.

The Sly1 regulatory loop has a positive function in vitro. (A–D) Fusion activity of Sly1∆loop versus wild-type Sly1. Master mixes were assembled as in Fig. 3 and incubated for 5 min at 30°C. Fusion was initiated by adding (arrows) Sly1 or Sly1∆loop at the concentrations indicated in the legend adjacent to panel B. Reactions were run in the absence (A and C) or presence (B and D) of 3% PEG; and in the absence (A and B) or presence (C and D) of Sec17, Sec18 (both 100 nM), and ATP. Points show mean ± SEM from three or more independent experiments; in some cases, the error bars are smaller than the symbols. Gray lines show least-squares nonlinear fits of a second-order kinetic model. (E) Purified Sly1∆loop protein is folded. Circular dichroism spectra of wild-type Sly1 and Sly1∆loop. The spectra are normalized to account for small differences in molecular mass and concentration. A comparison of lipid and content mixing signals for the experiments in panels B and D is presented in Fig. S3.

Figure 6.

Amphipathic helix α21 is indispensable for normal Sly1 function. (A) CONSURF analysis of evolutionary conservation within the Sly1 loop. Helix α21 is the most highly conserved portion of the loop. Locations of gain-of-function mutations, and hydrophobic residues within the loop are indicated, as are the five substitutions in the Sly1-pα21 mutant. (B) Position of helix α21 within Sly1. Note that no gain-of-function mutations within α21 have been identified. The loop is purple; the domain 3a templating domain is yellow. (C) Helix α21 and residues immediately upstream have the potential to fold into a strongly amphipathic α-helix. The helical wheel renderings comprise the region underlined in black and were produced using HELIQUEST; hydrophobic moment (µH) is indicated. (D) Growth phenotypes of cells carrying sly1-pα21, SLY1-20-pα21, and other alleles were assayed in a sly1∆ strain with a SLY1 balancer plasmid, which is ejected in the presence of 5-FOA. (E–J) RPL fusion with (E–G) Sly1-pα21 and (H–J) the compound mutant Sly1-20-pα21. For reference, fusion is also plotted for Sly1 and Sly1^∆loop. Reactions were set up with (E and H) 0% PEG, Sec17 and Sec18 (100 nM each), and ATP (1 mM); (F and I) 3% PEG and no Sec17, Sec18 (100 nM each), or ATP; or (G, J, and F) 0% PEG, Sec17 and Sec18 (100 nM each), and ATP (1 mM). Fusion was initiated at time = 0 by adding Sly1 or its mutants, at the concentrations indicated in the legends at the right side of the figure. Points show mean ± SEM from three or more independent experiments; in many cases the error bars are smaller than the symbols. Gray lines show least-squares nonlinear fits of a second-order kinetic model. Lipid mixing traces for panels G and J are presented in Fig. S4.

Figure S2.

In silico estimation of membrane binding by Sly1 helix α21 and its polar mutant pα21. (A and B) PMIpred results for helices α21 (A) and pα21 (B) were generated by the server at https://pmipred.fkt.physik.tu-dortmund.de/curvature-sensing/ (van Hilten et al., 2023a, , Preprint). Similar results for α21 and pα21 were obtained when the model was initialized with either negatively charged or neutral membranes. These estimated binding parameters for α21 predict stronger binding than for the well-characterized GMAP-210 ALPS domain, and similar binding to the engineered high-affinity ALPS_cond mutant (Magdeleine et al., 2016; van Hilten et al., 2023a).

Figure S3.

Lipid as well as content mixing is defective with Sly1∆loop. Parallel lipid (top) and content mixing (bottom) results are shown from the same sets of reactions. The content mixing traces are identical to those shown in Fig. 5, B and D, and are shown here to facilitate comparison with the lipid mixing data. All reactions contained 3% PEG. (A) The reactions in A omitted Sec17 and Sec18. (B) The reactions in B contained 100 nM each of Sec17 and Sec18, and ATP. Lipid mixing is reported as raw fluorescence counts in arbitrary units. As the membranes mix, FRET from Marina Blue DHPE to NBD-DHPE (initially in separate liposomes) quenches Marina Blue emission at 465 nm. Points and bars in all traces show means ± SEM of data from three separate experiments.

Figure S4.

Lipid, as well as content mixing, is defective with Sly1∆ polar loop mutants. Parallel lipid (top) and content mixing (bottom) results are shown from the same sets of reactions. (A and B) The content mixing data in A and B are identical to Fig. 6, G and J, and are shown here to facilitate comparison with the lipid mixing data. All reactions contained 3% PEG and 100 nM each of Sec17 and Sec18, and ATP. Lipid mixing is reported as raw fluorescence counts in arbitrary units. As the membranes mix, FRET from Marina Blue DHPE to NBD-DHPE (initially in separate liposomes) quenches Marina Blue emission at 465 nm. Points and bars in all traces show means ± SEM of data from three separate experiments.

Figure 7.

Appending the Sly1 loop to the amino terminus of Sly1∆loop partially restores function. (A) Chimeric constructs were prepared with different fragments of the Sly1 loop appended to the N-terminus of Sly1 via a short, flexible linker (see Table S3 for details). Mutants designated pα21 had the five polar substitutions in the appended loop as described in Fig. 6 C. (B) The loop-Sly1 mutants were expressed from the native SLY1 promoter on single-copy plasmids. Growth of a sly1∆ strain was assessed in the presence of the indicated constructs following ejection of a SLY1 balancer plasmid by plating on media containing 5-FOA. (C–F) Fusion driven by mutants with fragments of the loop (C and E) or polar derivatives of the same fragments (D and F). Points show mean ± SEM of three independent experiments; in many cases the error bars are smaller than the symbols. Gray lines show least-squares nonlinear fits of a second-order kinetic model.

Figure 8.

Sly1 helix α21 binds membranes, with a preference for higher curvature. TMR-α21 and TMR-pα21 peptides were added to liposomes of nominal diameter 30 and 200 nm, which contained 1% TRPE as a fluorescence acceptor. (A) Emission spectra of peptides or liposomes (30 nm diameter, 6.7% ergosterol) measured separately, and the sums of the peptide and liposome spectra. The sums represent the no-FRET condition. Both the TMR-α21 and TMR-pα21 spectra are plotted; they overlap almost exactly. Vertical dashed lines at 585 and 610 nm indicate emission peaks for labeled peptides and liposomes, respectively. (B) Example of FRET data. Spectra from binding reactions containing liposomes (30 nm diameter, 6.7% ergosterol, 500 µM total lipid) and 25 µM TMR-α21 or TMR-pα21 are shown. The no-FRET condition is shown for reference. (C) Normalized FRET ratios for binding reactions containing the indicated combinations of liposomes and peptides, as in panel B. Traces and bars in A–C show means and ±95% confidence bands from four independent experiments.

Figure 9.

Hyperactive forms of Sly1 tether high-curvature vesicles to immobilized Sed5. (A) The ability of Sed5-bound Sly1 to directly tether vesicles was tested using a bead-based assay system. GST-Sed5 was adsorbed to glutathione-sepharose (GSH) beads, and wild-type or mutant forms of Sly1 were added to the reaction mixture. After 5 min, Texas Red-DHPE labeled liposomes were added to the mixtures, incubated for 15–20 min, and imaged by confocal microscopy (10× objective). A false-color scale was chosen to emphasize small differences in contrast under conditions with less tethering. The micrographs are representative of at least two independent assays per condition. (B–D) To quantify tethering efficiency, we used a spin-down assay. Binding reactions set up as for microscopy were subjected to low-speed centrifugation to sediment the GSH beads and associated proteins and vesicles. The supernatant was discarded and detergent was added to the pellet to liberate bound fluorescent lipids; the resulting signal was quantified by fluorometry. In C, Sly1-20 was present for each condition. In D, Sly1* was preincubated with a 6:1 excess of soluble Sed5-Habc or Sed5-N-Habc, as indicated. Y-axes show bead-associated fluorescence (au, arbitrary units) after subtracting background from blanks containing only buffer. Bars indicate means ±95% confidence intervals for 4–10 independent experiments. Binding of the Sly1 variants to immobilized Sed5 was efficient and nearly stoichiometric (Fig. S5).

Figure S5.

Association of Sly1 and its variants with immobilized GST-Sed5. Binding reactions were set up as in Fig. 9 with the indicated Sly1 variants (80 pmol per 500 µl reaction) and GST-Sed5 cytoplasmic domain immobilized on glutathione-agarose (GST-Sed5Cyt; 150 pmol added to beads per reaction). The binding and wash buffer contained 10 mg/ml BSA. After sedimenting and washing the beads, proteins in the pellet were eluted with SDS-PAGE sample buffer, separated on 10% polyacrylamide gels, and stained with Coomassie blue. Source data are available for this figure: SourceData FS5.

Figure 10.

Working model. (A) Long-range tethering is mediated by coiled-coil Golgin family tethers and multisubunit tethering complexes (MTC’s). Flexibility or buckling of long-range tethers allows the vesicle to dwell in the region near Sly1 so that handoff can occur. (B) Mechanism of close-range tethering. Sly1 is anchored to the N-terminal domain of the Qa-SNARE on the target membrane. Note that in the closed ground state, the loop and helix α21 (magenta) occlude the sec22 binding site on the Sly1 SNARE templating domain (yellow). Binding of α21 to an incoming vesicle’s membrane pulls open the autoinhibitory loop and tethers the vesicle to Sly1, likely in a spatial orientation optimal for Sec22 binding to Sly1’s SNARE templating domain (yellow). In panels ii and iii, helix α21 is shown but the unstructured portion of the loop is omitted for clarity. (C) The Sly1 loop is conformationally heterogeneous in the crystal structure (PDB ID 1MQS). Temperature (B) factors in the Sly1 crystal are shown by color and backbone trace thickness. The highest disorder is in α20–α21, the thick red peptide segment.