January N-terminal signals in the SNX-BAR paralogs Vps5 and Vin1 guide coat complex formation Author List: 1 2 Shawn P. Shortill 0 1 2 Mia S. Frier 0 1 2 Michael Davey 0 1 2 Elizabeth Conibear conibear@cmmt.ubc.ca 0 1 2 . Centre for Molecular Medicine and Therapeutics, British Columbia Children's Hospital Research Department of Medical Genetics, University of British Columbia , Vancouver, BC VH6 3N1 , Canada Institute, University of British Columbia , Vancouver, BC V5Z 4H4 , Canada 2024 25 2024 191 196

*equal contribution Running title: Coat assembly by SNX-BAR N-termini

 VINE retromer Vps29 Vrl1 VARP endosome sorting nexin -

33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

 Abstract

Endosomal coat complexes assemble by incorporating membrane-binding subunits such as those of the sorting nexin (SNX) family. The S. cerevisiae SNX-BAR paralogs Vin1 and Vps5 are respective subunits of the endosomal VINE and retromer complexes that arose from a fungal whole genome duplication. Interactions mediated by the Vin1 and Vps5 BAR domains are required for protein complex assembly and membrane association. However, a degree of promiscuity is predicted for yeast BAR-BAR pairings, suggesting that another mechanism guides the formation of specific endosomal coat complexes. Previous work by our group and others has implicated the unstructured N-terminal domains of Vin1 and Vps5 in complex assembly. Here, we map N-terminal signals in both SNX-BAR paralogs that contribute to the formation and function of two distinct endosomal coats in vivo. Whereas Vin1 leverages a polybasic region and adjacent hydrophobic motif to bind Vrl1 and form VINE, the N-terminus of Vps5 interacts with the retromer subunit Vps29 at two separate sites. We show that one of these Vps5 motifs binds to a conserved hydrophobic pocket in Vps29 that is shared with other accessory proteins and targeted by a bacterial virulence factor in humans. Lastly, we examined the sole isoform of Vps5 from the milk yeast K. lactis and found that ancestral yeasts may have used a nested N-terminal signal to form both VINE and retromer. Our results suggest that the specific assembly of Vps5-family SNX-BAR coats depends on inputs from unique N-terminal sequence features in addition to BAR domain coupling, expanding our understanding of endosomal coat assembly mechanisms.

 Introduction

SNX-BARs are a conserved subfamily of dimerizing sorting nexin proteins that deform membranes into cargo-enriched tubules to promote membrane transport (Frost et al., 2009; Shortill, Frier, & Conibear, 2022; van Weering et al., 2010). Many SNX-BARs contain an extended, unstructured N-terminus, a lipid-binding PX domain and a concave BAR domain that drives dimerization (Ma & Burd, 2020; Van Weering & Cullen, 2014). Partner selection is thought to be enforced by complementary sets of charged amino acids in the extended interfaces of BAR domains, establishing a pairing code that promotes the specific assembly and function of SNX-BARs (Van Weering et al., 2012). Mutation of either the internal matching charged residues or the distal ends of BAR domains that facilitate tip-to-tip SNX-BAR oligomerization and tubule extension disrupts sorting (Dislich et al., 2011; Lopez-Robles et al., 2023; Van Weering et al., 2012). The S. cerevisiae SNX-BAR paralogs Vps5 and Vin1 arose from a fungal whole genome duplication event (Byrne & Wolfe, 2005; Wolfe & Shields, 1997) and subsequently diverged to assume new roles. We previously determined that Vin1 specifically binds to the VPS9-domain GEF Vrl1 to form a novel endosomal coat complex that we named VINE (Shortill et al., 2022). In contrast, Vps5 is a well-studied 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 subunit of the ubiquitous retromer complex—a conserved endosomal assembly that promotes the retrograde transport of numerous protein cargoes (Bean et al., 2017; Burd & Cullen, 2014; Ma & Burd, 2020). Yeast retromer has been described as a heteropentamer composed of a Vps5-Vps17 SNX-BAR dimer and a Vps26-Vps35-Vps29 trimer (Seaman et al., 1998). In humans retromer refers strictly to the VPS26-VPS35VPS29 trimer, which can associate with SNX-BAR dimers or other membrane adaptors to regulate diverse sorting processes (Arighi et al., 2004; Wassmer et al., 2007; Cui et al., 2019; Seaman, 2021; Steinberg et al., 2013; Courtellemont et al., 2022; Kvainickas et al., 2017; Simonetti et al., 2019).

Vin1 associates with Vrl1 through its unstructured N-terminus (Shortill et al., 2022). Likewise, both Vps5 and its human ortholog SNX1 interact with the Vps35-Vps29-Vps26 trimer through their low complexity N-termini, although the motifs involved have yet to be described (Seaman & Williams, 2002; Gullapalli et al., 2004). Because Vin1 and Vps5 evolved from a common ancestor, they could have acquired divergent N-termini that dictate protein complex selectivity, and therefore functional specificity. Using a combination of structural predictions, live-cell imaging, and biochemical assays, we mapped the motifs in the Vps5 and Vin1 N-termini that facilitate binding to Vps29 and Vrl1, respectively. We identified a shared Leu-Phe motif as well as sequence features unique to either paralog. In the closely related yeast K. lactis, the sole Vps5 isoform has nested N-terminal motifs for binding to Vrl1 and Vps29, indicating a possible ancient regulatory relationship between VINE and retromer. Taken together, our findings help explain the functional divergence between Vin1 and Vps5 and demonstrate the contributions of unstructured regions in SNX-BAR proteins to the assembly and function of membrane sorting complexes.

 Results Vrl1 recognizes polybasic and Leu-Phe signals in the Vin1 N-terminus

Vin1 and its paralog Vps5 leverage N-terminal interactions to form VINE and retromer, respectively. Previously, we found that Vrl1 interacts with the N-terminus of Vin1, but not Vps5, and that this interaction depends on a 20 amino acid (aa) motif in the Vin1 N-terminus (aa 76-95) (Shortill et al., 2022). To identify the N-terminal sequence elements that allow Vrl1 to distinguish Vin1 from Vps5, we examined the predicted interface between Vin176-95 and Vrl1 that was obtained using AlphaFold2 through the ColabFold platform (Jumper et al., 2021; Mirdita et al., 2022); Figure 1A-C). The Vin1 peptide and Vrl1 make extensive contact, with numerous predicted sidechain interactions that span the full Vin176-95 region.

We systematically mutated Vin1-mScarletI (mScI) between aa 76-95 by substituting residues with alanine 100 in sets of two (Figure 1D) and determined if these mutants recognize Vrl1 in our previously established 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 chimeric subcellular recruitment assay (Shortill et al., 2022); Figure 1E, F). Mutation of consecutive residues within a Vin1 basic region, R88, R89, R90 and R91 (herein referred to as the “polybasic” site), completely blocked recruitment by the Vrl1 chimera (p<0.0001) while mutation of K82 and L83 or F84 and T85 had more intermediate disruptive effects (p<0.0001 and p<0.05, respectively). This Vin1 polybasic region is predicted to interact with Vrl1 at the conserved acidic patch that we previously demonstrated was necessary for recognition (Shortill et al., 2022); Figure 1B), suggesting that charged interactions at this site serve an important role in VINE assembly.

To further resolve the contribution of the Vin1 K82, L83, F84 and T85 residues to the interaction with Vrl1, we made an additional L83A F84A mutant (Figure 2A) and compared its recruitment by the Vrl1 chimera to that of the previously tested K82A L83A and F84A T85A mutants (Figure 2B, C). Mutation of this LeuPhe motif caused a severe loss of recruitment by the Vrl1( 1-703 )YPE chimera (p<0.0001; Figure 2C), suggesting that this pair of hydrophobic residues is critical for VINE formation. The Leu-Phe motif is predicted to interface with a hydrophobic patch in the Vrl1 AnkRD that is adjacent to the acidic site (Figure 1C). Introducing a charged residue to this hydrophobic patch (Vrl1L497D) strongly reduced recruitment of the Vin1 N-terminus (Vin11-116-mScI; p<0.0001; Figure 2D, E), suggesting that this is another site of interaction between Vrl1 and the Vin1 N-terminus.

We also assessed the ability of Vin1 polybasic and Leu-Phe mutants to recruit full-length Vrl1-Envy to puncta. Vrl1 localization in vin1∆ cells was restored by Vin1-mScI, and introducing either R90A R91A or L83A F84A mutations to Vin1 disrupted this localization (p<0.0001; Figure 2F, G), suggesting that VINE is only partially assembled and localized in these cells. Taken together, these findings indicate that Vrl1 recognizes two distinct features of the Vin1 N-terminus to form VINE—a polybasic stretch that interacts with an acidic site and a Leu-Phe motif that binds an adjacent hydrophobic patch.

 Vps5 binds to Vps29 through a bipartite motif in its N-terminus

In similar fashion to the interactions that promote formation of VINE, the N-terminus of Vps5 is critical for interaction with the Vps26-Vps35-Vps29 retromer trimer in yeast (Seaman & Williams, 2002). Mutation of a conserved hydrophobic pocket in Vps29 blocks association of the trimer with the Vps5-Vps17 dimer and results in a moderate defect in progression of the vacuolar hydrolase CPY through the retromer131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 dependent vacuolar protein sorting pathway (Collins et al., 2005). Despite this, how the Vps5 N-terminus associates with the trimer remains incompletely understood.

Using AlphaFold2 (Mirdita et al., 2022), we obtained a confident binding prediction between aa 155-198 of the unstructured Vps5 N-terminus and two distinct sites in Vps29 (Figure 3A, Figure S1A, B). A Vps5 FTDPL motif (herein referred to as “pocket-binding”) was predicted to interact with Vps29 at the previously identified conserved hydrophobic pocket, with Vps5 L196 projecting into the hydrophobic core (Figure 3B). We also identified a Vps5 β-strand composed of the sequence PRILFDS (herein referred to as “sheetbinding”), which is predicted to form a lateral extension of a Vps29 β-sheet found on the surface opposite to the conserved hydrophobic pocket and is separated from the pocket-binding motif by a ~30 aa unstructured loop region (Figure 3C). Like Vin1, the Vps5 sheet site possesses a Leu-Phe motif predicted to leverage multiple hydrophobic contacts with Vps29 (Figure 1C and 3C).

Substituting the pocket- and sheet-binding motifs with alanine in Vps5-mScI either separately or together (bipartite-Ala; Figure 3D) reduced its co-purification with triple hemagglutinin (3HA)-tagged Vps29 (p<0.0001; Figure 3E, F). Disruption of either pocket- or sheet-binding motifs caused intermediate binding defects, whereas combined disruption of both sites resulted in binding loss similar to a full N-terminal truncation mutant (Vps5275-end; p<0.0001; Figure 3E, F). Moreover, mutating either the sheet-binding L160 and F161 residues to alanine, or substituting the pocket-binding L196 to lysine, resulted in Vps29 binding defects comparable to complete alanine substitution of either motif (Figure 3D, E, F), suggesting that these residues make important contributions to the interaction with Vps29.

To test the impact of these mutations on retromer-mediated sorting we used a functional assay based on the sorting of newly synthesized CPY at the Golgi. The CPY receptor Vps10 is recognized by retromer at endosomes and recycled to perform additional rounds of CPY transport (Seaman et al., 1998). If recycling is disrupted, soluble CPY is secreted from the cell. Truncation of the Vps5 N-terminus causes CPY processing defects (Seaman & Williams, 2002), and we found that a full N-terminal deletion mutant (Vps5275-end) also displays a strong secretion phenotype (Figure 3G, H). The Vps5 N-terminal pocket- and sheet-binding mutants secreted CPY at levels roughly proportional to the severity of demonstrated Vps29 binding deficiency, although not to the same level as truncation of the N-terminus (p<0.0001; Figure 3E, 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191

F). Taken together, these data suggest that Vps5 interacts with Vps29 through an N-terminal bipartite pocket- and sheet-binding motif that is critical for both retromer assembly and efficient sorting. Because both Vps5 and Vin1 leverage their N-termini to form retromer and VINE, respectively, we wondered if these sites could act as sole determinants of coat formation. To test this, we generated a series of mScI-tagged truncation mutants of Vin1 and Vps5 containing either the N-terminus or PX-BAR regions alone, as well as chimeras with respective N-termini swapped to that of the paralog (Figure S2A). None of the Vin1-mScI fragments or Vin1Nt chimera were able to recover the punctate localization of Vrl1-Envy in vin1∆ cells (Figure S2B), suggesting that Vrl1 requires input from both the Vin1 N-terminus and Vin1 BAR domain to form VINE. Conversely, while both the Vps5 N-terminus alone and Vps5Nt-Vin1∆Nt chimera failed to recover the punctate localization of the retromer subunit Vps35-GFP in vps5∆ cells, the Vps5 PXBAR region alone was able to partially complement (Figure S2C), in line with a previous report (Seaman & Williams, 2002). Taken together, these findings indicate that, while the N-termini of these SNX-BAR paralogs are important for determining coat complex membership, they are not sufficient and input from the respective SNX-BAR domains also contributes to selectivity.

The sole Vps5 protein from milk yeast K. lactis forms both VINE and retromer In S. cerevisiae, protein products of the duplicated VIN1 and VPS5 genes have diverged to assume separate roles in forming the VINE and retromer complexes, respectively. Ancestors of the modern milk yeast K. lactis did not undergo the same whole genome duplication (WGD) as those of S. cerevisiae (Kellis et al., 2004; Keogh et al., 1998), and therefore its genome encodes a single ortholog of the VPS5 gene. Interestingly, the K. lactis genome also encodes orthologs of both the VRL1 and VPS17 genes (Byrne & Wolfe, 2005), suggesting that the single copy of Vps5 in this organism may form both VINE and retromer (Figure 4A).

To investigate if K. lactis Vps5 can form both VINE and retromer, we first generated an RFP-tagged K. lactis Vps5 construct for ectopic expression in S. cerevisiae (klVps5-mScI). We observed klVps5-mScI at small puncta in wild-type cells that were greatly enhanced in a vps5∆ strain and lost in a vps17∆ strain (p<0.0001 and p<0.05, respectively; Figure 4B, C), suggesting that klVps5 competes with S. cerevisiae Vps5 (scVps5) for scVps17 binding and retromer formation. Indeed, klVps5-mScI strongly co-purified with scVps29-3HA (Figure S3A) and partially rescued CPY sorting defects (Figure S3B). These results 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 demonstrate that klVps5 can cooperate with the S. cerevisiae machinery to form a hybrid retromer complex with limited cargo sorting capabilities.

We then examined the ability of klVps5 to form a hybrid VINE complex. Neither scVrl1-Envy nor the artificially recruited scVrl1 chimera could recover the loss of klVps5-mScI puncta in vin1∆vps17∆ cells (Figure S4A), but moderate recruitment of the klVps5 N-terminus by the artificially recruited scVrl1 chimera suggested that a weak interaction may occur despite species-specific differences (Figure S4B). We wondered if the K. lactis Vrl1 homolog (klVrl1) might more readily bind klVps5 to form VINE. Indeed, ectopic expression of klVrl1-Envy greatly increased bright klVps5-mScI puncta in wild-type cells and overcame the loss of puncta in vin1∆vps17∆ cells (p<0.01; Figure 4D, E), suggesting that klVrl1 and klVps5 can cooperate to form the K. lactis version of VINE in S. cerevisiae, herein referred to as klVINE. Taken together our results indicate that klVps5 can form both klVINE and retromer. klVps5 recognizes klVrl1 and scVps29 at an overlapping site in its N-terminus To understand how both VINE and retromer could assemble while sharing the single Vps5 isoform in K. lactis, we took a structural modeling approach. Using AlphaFold2, we first obtained confidently predicted interactions for the BAR domain of klVps5 with those of either klVrl1 or klVps17, consistent with the idea that klVps5 can participate in both complexes (Mirdita et al., 2022; Figure S5A, B). We then identified a region of the klVps5 N-terminus between aa 172-203 that was confidently predicted to associate with klVps29 (Figure 5A, Figure S6A). Like the interaction between S. cerevisiae Vps5 and Vps29 (Figure 3A), klVps5 was predicted to interact with klVps29 through bipartite pocket- and sheet-binding sites (Figure 5A). The pocket-binding sites of both klVps5 and scVps5 contain a common DPL motif, while the sheetbinding sites share the Leu-Phe motif which is predicted to interface with the same region of Vps29. Next, we used AlphaFold2 to predict binding between the entire klVps5 N-terminus and the VPS9 and AnkRD domains of klVrl1 (klVrl1183-717; Figure S6B). Strikingly, the ~16 aa loop region between the klVps5 pocket- and sheet-binding sites contains a RTRRHP motif that was confidently predicted to interact with the same region of klVrl1 that forms the VINE interface in S. cerevisiae (Figure 5B, Figure S6B). This RTRRHP motif is not predicted to form any contacts with klVps29 and has electrochemical similarity to the polybasic region that we identified in S. cerevisiae Vin1 (Figure 1E, F). Taken together, these results 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 suggest that the same ~32 aa region of the klVps5 N-terminus contains overlapping signals that contribute to the formation of both VINE and retromer in K. lactis.

To test if these predicted interactions are important in vivo, we introduced a series of mutations to the Nterminus of klVps5-mScI (Figure 5C). klVrl1-Envy was largely displaced from bright puncta in vin1∆vps17∆ cells when either the N-terminus of klVps5-mScI was truncated (klVps5204-end) or the basic RTRRHP motif was substituted to alanine (p<0.0001 and p<0.001, respectively; Figure 5D, E). In contrast, alanine substitution of either the predicted klVps29 sheet- and pocket-binding sites, or associated Leu-Phe motif, had no effect on VINE formation (Figure 5D, E), suggesting that the N-terminal polybasic site specifically contributes to ectopic VINE formation.

Despite only partial rescue of S. cerevisiae vps5Δ phenotypes by klVps5, mutations in the klVps5 Nterminus resulted in significant sorting defects. Truncation of the klVps5 N-terminus increased CPY secretion (Figure 5F, G), similar to the effect of deleting the S. cerevisiae Vps5 N-terminus (Figure 3G, H). Disruption of the pocket- and sheet-binding sites also caused a significant defect (p<0.01, Figure 5F, G). The effects of mutations in K. lactis Vps5 were in agreement with those in S. cerevisiae Vps5, with the bipartite mutations eliciting a more severe defect than mutations in either pocket- or sheet-binding sites alone, while not accounting for the full contribution of the N-terminal unstructured region (Figure 3G, H, Figure 5F, G). In contrast to the requirements for klVINE formation (Figure 5D, E), alanine substitution of the klVps5 N-terminal polybasic site did not result in obvious CPY sorting defects. Taken together, these data suggest that the N-terminus of K. lactis Vps5 may contribute to the formation of both VINE and retromer through interactions at a nested site (Figure 5H).

 Discussion

Here we revealed mechanisms underlying SNX-BAR assembly by studying the S. cerevisiae paralogs Vps5 and Vin1. We found that these proteins use shared and unique motifs in their N-termini to guide coat complex selection, with input from their respective BAR domains. Our results also suggest that nested motifs confer dual roles for the sole K. lactis Vps5 isoform as a subunit of both VINE and retromer. Taken together, our work delineates a mechanism by which endosomal coats select SNX-BAR proteins by recognizing unstructured N-terminal domains, which are common in this family of membrane adaptors.

The N-termini of Vps5-family SNX-BARs promote coat assembly

The unstructured N-terminal regions of both Vin1 and Vps5 contribute to the specificity of endosomal coat formation in S. cerevisiae. Both Vin1 and Vps5 leverage a Leu-Phe motif, yet the Vin1 Leu-Phe motif is 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 predicted to contact a Vrl1 hydrophobic patch, whereas the Vps5 Leu-Phe forms part of a short β-strand predicted to laterally extend a Vps29 β-sheet. Vin1 also interacts at a conserved surface on the Vrl1 AnkRD through a polybasic sequence, while a Vps5 FTDPL motif associates with Vps29 at a conserved hydrophobic pocket. Interestingly, the duplicated human orthologs of Vin1 and Vps5, SNX1/SNX2, also feature extended N-termini that interact with the retromer-associated sorting nexin SNX27 (Chandra et al. 2021; Simonetti et al., 2019; Yong et al., 2021). Yeast two-hybrid experiments have demonstrated that SNX1/SNX2 also bind to VPS29 and VPS35 (Haft et al., 2000; Rojas et al., 2007) and Gullapalli et al. (2004) mapped retromer binding to the SNX1 unstructured N-terminus.

Similarly to Vps5, SNX1 binds to the conserved hydrophobic pocket in VPS29 (Swarbrick et al., 2011). This pocket is of special interest since it is targeted by various effector proteins in humans (Baños-Mateos et al., 2019). The bacterial pathogen L. pneumophila interferes with retromer by binding the VPS29 pocket through its effector protein RidL (Finsel et al., 2013). Additionally, the Vrl1 homolog and RAB21 GEF VARP, the RAB7 GAP TBC1D5, the Commander complex subunit VPS35L, and the WASH complex subunit FAM21 also compete for this site (Crawley-Snowdon et al., 2020; Hesketh et al., 2014; Healy et al. 2023; Guo et al. 2023), suggesting that the VPS29 pocket is a multi-functional interface that could spatiotemporally control Rab dynamics and regulate interactions with sorting nexins and other accessory proteins. Further work is needed to determine the relevance of the interaction between SNX1 and the VPS29 pocket in vivo, and its role in regulating the incorporation of metazoan retromer into a variety of sorting complexes.

The unstructured N-termini of Vps5/Vin1 and SNX1/SNX2 are each at least one hundred aa long yet known interactions can be mapped to very short motifs (Simonetti et al., 2022). It is possible that remaining Nterminal sequences mediate additional regulatory interactions. SNX5/6 interact with p150Glued, a component of the dynactin machinery responsible for controlling dynein motor function, to drive cargo transport (Hong et al., 2009; Wassmer et al., 2009) and SNX2 binds the ER protein VAP to regulate lipid exchange at ERendosome contacts (Dong et al., 2016). Accessory interactions could occur simultaneously to capture auxiliary factors that enhance sorting, or in a stepwise manner to ensure specificity while adding a layer of temporal control.

 Assembly and function of VINE and retromer in K. lactis

The genome of milk yeast K. lactis encodes individual copies of the VPS5, VPS17 and VRL1 genes. Our experiments based on ectopic expression in S. cerevisiae suggest that klVps5 forms both VINE and retromer, at least in part through nested motifs in its N-terminus. The use of overlapping signals ensures that Vps5 can engage with only one complex at a time, and raises questions about how the relative levels of retromer and VINE are regulated. The partitioning of klVps5 between retromer and VINE complexes could be driven by its abundance and its relative affinity for klVrl1 or subunits of retromer. In theory, posttranslational modifications could alter the activity or availability of one motif. Because the relevant pathways are unlikely to be present in S. cerevisiae, further studies in K. lactis are needed to uncover the regulatory mechanisms that govern subunit sharing. Taken together, our work has clarified mechanisms of SNX-BAR complex assembly in yeast, which could provide insight into similar strategies used by human 290 291 292 293 294 295 296 297 298 299

SNX-BARs.

 Materials and Methods Key Resources Table Reagent type (species) or resource

Antibody Antibody Antibody Antibody Antibody Other Other Other Other Other

 Designation

Anti-HA (Mouse monoclonal) Anti-RFP (Mouse monoclonal) Anti-HA (Rabbit polyclonal) HRP-Anti-Mouse (Goat polyclonal) Anti-CPY (Mouse monoclonal) nProtein A Sepharose 4 Fast Flow Amersham ECL Amersham ECL Prime Yeast/Fungal ProteaseArrest Concanavalin A Sigma-Aldrich C2010 AbCam Jackson Thermo Fisher Scientific Cytiva Cytiva Cytiva GBiosciences 786-435

 Source or reference

Covance

 Identifiers

MMS-101R; HA.11 ChromoTek 6G6

 Additional information

WB (1:1,000) WB (1:2,000) Ab9110

CoIP 115-035-146

WB (1:20,000) A-6428

WB (1:500) 17528004

CoIP GERPN2209 GERPN2232

Chemiluminescent reagent Chemiluminescent reagent Yeast protease inhibitor; 100X Yeast live-cell imaging preparation

Software, algorithm Software, algorithm Software, algorithm

MetaMorph GraphPad Prism ColabFold AlphaFold2 Advanced

MDS Analytical Technologies GraphPad Software Jumper et al., 2021; Mirdita Yeast strains and plasmids used in this study are described in Supplementary Files 1 and 2, respectively. Yeast strains were built in the BY4741 strain background using homologous recombination-based integration unless otherwise indicated. Gene deletions, promoter exchanges and tags were confirmed by colony PCR and either western blot or fluorescence microscopy where possible. Plasmids were built by NEBuilder® HiFi DNA assembly or homologous recombination in yeast, recovered in Escherichia coli, and confirmed by sequencing. K. lactis DNA constructs were built using native sequences obtained from strain NRRL Y-1140 (ATCC 8585).

 Bioinformatic analysis of protein folding

Prediction of protein structure and binding interfaces was performed using the AlphaFold2 advanced server accessed through the ColabFold platform with default settings (Jumper et al., 2021; Mirdita et al., 2022). Structural models were processed for presentation using PyMOL2 (Schrödinger, LLC, New York, New York).

 Fluorescence microscopy and automated image analysis

Yeast strains were prepared for microscopy by overnight growth, followed by four hours of growth at 30 °C in fresh synthetic dextrose-based (SD) media and plating on 96-well glass-bottom plates (Cellvis, Mountain View, California) coated in concanavalin A. Images were collected using a DMi8 microscope (Leica Microsystems, Wetzlar, Germany) equipped with an ORCA-flash 4.0 digital camera (Hamamatsu Photonics, Shizuoka, Japan) and a high-contrast Plan-Apochromat 63x/1.30 Glyc CORR CS oil immersion lens (Leica Microsystems, Wetzlar, Germany). The MetaMorph 7.8 software package was used for image acquisition and processing (MDS Analytical Technologies, Sunnyvale, California). The intensity of a given fluorophore is scaled identically in all micrographs within an experiment. Images were resized using Photoshop CC 2020 (Adobe, San Jose, California) and arranged in Illustrator CC 2020 (Adobe, San Jose, California). MetaMorph 7.8 scripted journals were used to quantify raw images (MDS Analytical Technologies, Sunnyvale, California). The Count Nuclei feature utilizing intensity above local background 326 327 328 329 330 (IALB) was used to exclude dead cells and identify live cells. The Granularity feature was used to identify puncta in a dead cell-masked intermediate image based on IALB. Masking functions were performed using the Arithmetic function with Logical AND. RFP channel signal is represented as magenta in merged micrographs to improve accessibility.

 Coimmunoprecipitation and western blotting

For co-immunoprecipitation analyses, yeast cultures were grown to log phase at 30 °C in synthetic dextrosebased (SD) media and 75 OD600 of yeast were collected and incubated for 15 minutes in 50 mM Tris-Cl with 10 mM DTT (pH 9.5) at room temperature, followed by one hour digestion in spheroplasting buffer (1.2 M sorbitol, 50 mM KH2PO4, 1 mM MgCl2 and 250 µg/ml zymolase at pH 7.4) at 30 °C. After washing spheroplasts twice with 1.2 M sorbitol, they were frozen at -80°C then incubated for 10 minutes at room temperature in 500 µl lysis buffer (0.5% Triton X-100, 50 mM HEPES, 1 mM EDTA, 50 mM NaCl, 1 mM PMSF and 1x fungal Protease Arrest, pH 7.4). 50 µl of lysate was collected for each sample and mixed with 2x Laemmli buffer (4% SDS, 20% glycerol, 120 mM Tris-Cl (pH 6.8), 0.01g bromophenol blue and 10% beta-mercaptoethanol) for analysis by western blot. A polyclonal rabbit anti-HA antibody (ab9110, Abcam) was added to remaining lysates and incubated at 4 °C for 1 hour. Protein A Sepharose beads (Cytiva) were added and incubated at 4 °C for 1 hour, then beads were washed three times in lysis buffer and resuspended in 50 µl of Thorner buffer (8 M Urea, 5% SDS, 40mM Tris-Cl (pH 6.4), 1% beta-mercaptoethanol, 0.4 mg/mL bromophenol blue) and heated for five minutes at 80 °C. Proteins were separated on 8% SDSPAGE gels and detected by western blot using monoclonal mouse anti-HA (MMS-101R; Covance) or monoclonal mouse anti-RFP antibodies (6G6; ChromoTek) prior to secondary antibody treatment with polyclonal goat anti-mouse conjugated to horseradish peroxidase (115–035-146; Jackson ImmunoResearch Laboratories). Blots were developed with Amersham ECL (GERPN2209, Cytiva) or Amersham ECL Prime (GERPN2232, Cytiva) chemiluminescent western blot detection reagents and exposed using the Vilber Fusion FX with automatic settings (Vilber Smart Imaging, Collégien, France). Densitometry of scanned films was performed using ImageJ (Schneider et al., 2012).

For detection of secreted carboxypeptidase Y (CPY), yeast cultures were grown to log phase at 30 °C, serially diluted 1:1 and spotted to synthetic dextrose (SD) based media and incubated 12-15 hours at 30 °C under a nitrocellulose membrane. For test of CPY secretion in klVps5-expressing strains the vps5Δ control strain was spotted in duplicate to provide separate data points for normalization and statistical comparison. Nitrocellulose was washed with dH2O and probed using a monoclonal mouse anti-CPY antibody (A-6428, Thermo Fisher Scientific) followed by secondary antibody treatment with polyclonal goat anti-mouse conjugated to horseradish peroxidase (115–035-146; Jackson ImmunoResearch Laboratories) and development as described above. 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387

 Statistical analysis of quantitative data

Statistical analyses were performed using GraphPad Prism 9.1.0 (GraphPad Software, San Diego, California). A 95% confidence threshold (P < 0.05) was used for hypothesis testing. Microsoft Excel 2019 was used to graph data (Microsoft, Redmond, Washington). Bar graphs show mean value of all collected biological replicates, with data points from individual replicates represented as scatter plots coloured by replicate. Error bars represent standard error of the mean (SEM).

 Acknowledgements

We thank Dr. Luc Berthiaume (University of Alberta, Edmonton, Canada) for generously sharing rabbit anti-GFP serum. We gratefully acknowledge funding support from the Natural Sciences and Engineering Research Council of Canada (grant RGPIN-2022-04573 to EC and PGS-D3 Doctoral Scholarship to SPS); Canada Foundation for Innovation (Leading Edge Fund 30636); Canadian Institutes of Health Research (CGS-M Frederick Banting and Charles Best Canada Graduate Scholarship to SPS and MSF and CGS-D Frederick Banting and Charles Best Canada Graduate Scholarship to MSF); BC Children’s Hospital Research Institute Jan M. Friedman Graduate Studentship to SPS; University of British Columbia Catalyst Paper Corporation Affiliated Fellowship to SPS, Gertrude Langridge Graduate Scholarship and Elwyn Gregg Memorial Affiliated Fellowships to MSF; 4-Year Doctoral Fellowship to SPS and MSF; and University of British Columbia Medical Genetics Rotation Award to MSF.

 Conflict of Interest Statement

The authors declare that there are no conflicts of interest. mammalian retromer in sorting of the cation-independent mannose 6-phosphate receptor. Journal of Cell Biology, 165(1), 123–133. https://doi.org/10.1083/jcb.200312055 Baños-Mateos, S., Rojas, A. L., & Hierro, A. (2019). VPS29, a tweak tool of endosomal recycling.

Current Opinion in Cell Biology. https://doi.org/10.1016/j.ceb.2019.03.010 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411

Bean, B. D. M., Davey, M., & Conibear, E. (2017). Cargo selectivity of yeast sorting nexins. Traffic, 18(2), 110–122. https://doi.org/10.1111/tra.12459 Burd, C., & Cullen, P. J. (2014). Retromer: A master conductor of endosome sorting. Cold Spring Harbor

Perspectives in Biology, 6(2). https://doi.org/10.1101/cshperspect.a016774 Byrne, K., & Wolfe, K. (2005). The Yeast Gene Order Browser: Combining curated homology and syntenic context reveals gene fate in polyploid species. Genome Research, 15(10), 1456–1461. Chandra, M., Collins, B. M., & Jackson, L. P. (2021). Biochemical basis for an interaction between SNX27 and the flexible SNX1 N-terminus. Advances in Biological Regulation, 100842. Collins, B., Skinner, C., Watson, P., Seaman, M., & Owen, D. (2005). Vps29 has a phosphoesterase fold that acts as a protein interaction scaffold for retromer assembly. Nature Structural & Molecular Biology, 12(7), 594–602. https://doi.org/10.1038/NSMB954 Courtellemont, T., De Leo, M. G., Gopaldass, N., & Mayer, A. (2022). CROP: a retromer-PROPPIN complex mediating membrane fission in the endo-lysosomal system. EMBO J. 41(10): e109646. Crawley-Snowdon, H., Yang, J.-C., Zaccai, N. R., Davis, L. J., Wartosch, L., Herman, E. K., Bright, N.

A., Swarbrick, J. S., Collins, B. M., Jackson, L. P., Seaman, M. N. J., Luzio, J. P., Dacks, J. B., Neuhaus, D., & Owen, D. J. (2020). Mechanism and evolution of the Zn-fingernail required for interaction of VARP with VPS29. Nature Communications 2020 11:1, 11(1), 1–15. Cui, Y., Carosi, J. M., Yang, Z., Ariotti, N., Kerr, M. C., Parton, R. G., Sargeant, T. J., & Teasdale, R. D. (2019). Retromer has a selective function in cargo sorting via endosome transport carriers. Journal of Cell Biology. https://doi.org/10.1083/jcb.201806153 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437

Dislich, B., Than, M. E., & Lichtenthaler, S. F. (2011). Specific amino acids in the BAR domain allow homodimerization and prevent heterodimerization of sorting nexin 33. The Biochemical Journal, 433(1), 75–83. https://doi.org/10.1042/BJ20100709 Dong, R., Saheki, Y., Swarup, S., Lucast, L., Harper, J. W., & De Camilli, P. (2016). Endosome-ER Contacts Control Actin Nucleation and Retromer Function through VAP-Dependent Regulation Finsel, I., Ragaz, C., Hoffmann, C., Harrison, C. F., Weber, S., van Rahden, V. A., Johannes, L., & Hilbi, H. (2013). The Legionella Effector RidL Inhibits Retrograde Trafficking to Promote Intracellular Replication. Cell Host & Microbe, 14(1), 38–50. https://doi.org/10.1016/j.chom.2013.06.001 Frost, A., Unger, V., & De Camilli, P. (2009). The BAR domain superfamily: Membrane-molding Gullapalli, A., Garrett, T. A., Paing, M. M., Griffin, C. T., Yang, Y., & Trejo, J. (2004). A role for sorting nexin 2 in epidermal growth factor receptor down-regulation: Evidence for distinct functions of sorting nexin 1 and 2 in protein trafficking. Molecular Biology of the Cell, 15( 5 ), 2143–2155. Guo, Q., Chen, K., Gimenez-Andres, M., Jellett, A. P., Gao, Y., Simonetti, B., Liu, M., Danson, C. M., Heesom, K. J., Cullen, P. J., & Collins, B. M. (2023). Structural basis for coupling of the WASH subunit FAM21 with the endosomal SNX27-Retromer complex. bioRxiv 2023.08.15.553351 Haft, C., de la Luz Sierra, M., Bafford, R., Lesniak, M., Barr, V., & Taylor, S. (2000). Human orthologs of yeast vacuolar protein sorting proteins Vps26, 29, and 35: assembly into multimeric complexes. Molecular Biology of the Cell, 11(12), 4105-4116. Healy, M. D., McNally, K. E., Butkovic, R., Chilton, M., Kato, K., Sacharz, J., McConville, C., Moody, E. R. R., Shaw, S., Planelles-Herrero, V. J., Yadav, S. K. N., Ross, J., Borucu, U., Palmer, C. S., Chen, K., Croll, T. I., Hall, R. J., Caruana, N. J., Ghai, R., Nguyen, T. H. D., Heesom, K. J., Saitoh, S., Berger, I., Schaffitzel, C., Williams, T. A., Stroud, D. A., Derivery, E., Collins, B. M., 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 & Cullen, P. J. (2023). Structure of the endosomal Commander complex linked to RitscherSchinzel syndrome. Cell 186(10):2219-2237.e29.

Hesketh, G., Pérez-Dorado, I., Jackson, L. P., Wartosch, L., Schäfer, I. B., Gray, S. R., McCoy, A. J., Zeldin, O. B., Garman, E. F., Harbour, M. E., Evans, P. R., Seaman, M. N. J., Luzio, J. P., & Owen, D. J. (2014). VARP is recruited on to endosomes by direct interaction with retromer, where together they function in export to the cell surface. Developmental Cell, 29( 5 ), 591–606. Hong, Z., Yang, Y., Zhang, C., Niu, Y., Li, K., Zhao, X., & Liu, J.-J. (2009). The retromer component SNX6 interacts with dynactin p150Glued and mediates endosome-to-TGN transport. Cell Research, 19(12), Article 12. https://doi.org/10.1038/cr.2009.130 Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M., Ronneberger, O., Tunyasuvunakool, K., Bates, R., Žídek, A., Potapenko, A., Bridgland, A., Meyer, C., Kohl, S. A. A., Ballard, A. J., Cowie, A., Romera-Paredes, B., Nikolov, S., Jain, R., Adler, J., … Hassabis, D. (2021). Highly accurate protein structure prediction with AlphaFold. Nature 2021 596:7873, 596(7873), 583–589. Kellis, M., Birren, B. W., & Lander, E. S. (2004). Proof and evolutionary analysis of ancient genome duplication in the yeast Saccharomyces cerevisiae. Nature, 428(6983), 617–624. Keogh, R. S., Seoighe, C., & Wolfe, K. H. (1998). Evolution of gene order and chromosome number in Saccharomyces, Kluyveromyces and related fungi. Yeast (Chichester, England), 14( 5 ), 443–457. Kvainickas, A., Jimenez-Orgaz, A., Nägele, H., Hu, Z., Dengjel, J., & Steinberg, F. (2017). Cargoselective SNX-BAR proteins mediate retromer trimer independent retrograde transport. Journal of Cell Biology. https://doi.org/10.1083/jcb.201702137 Lopez-Robles, C., Scaramuzza, S., Astorga-Simon, E. N., Ishida, M., Williamson, C. D., Baños-Mateos, S., Gil-Carton, D., Romero-Durana, M., Vidaurrazaga, A., Fernandez-Recio, J., Rojas, A. L., 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488

Bonifacino, J. S., Castaño-Díez, D., & Hierro, A. (2023). Architecture of the ESCPE-1 membrane coat. Nature Structural & Molecular Biology. https://doi.org/10.1038/s41594-023-01014-7 Ma, M., & Burd, C. G. (2020). Retrograde trafficking and plasma membrane recycling pathways of the budding yeast Saccharomyces cerevisiae. Traffic, 21(1). https://doi.org/10.1111/tra.12693 Mirdita, M., Schütze, K., Moriwaki, Y., Heo, L., Ovchinnikov, S., & Steinegger, M. (2022). ColabFold: Making protein folding accessible to all. Nature Methods, 19(6), Article 6. Rojas, R., Kametaka, S., Haft, C., & Bonifacino, J. S. (2007). Interchangeable but essential functions of SNX1 and SNX2 in the association of retromer with endosomes and the trafficking of mannose6-phosphate receptors. Molecular and Cellular Biology, 27(3), 1112-1124. Seaman, M. N. J. (2021). The Retromer Complex: From Genesis to Revelations. Trends in Biochemical Seaman, M. N. J., McCaffery, J. M., & Emr, S. D. (1998). A membrane coat complex essential for endosome-to-Golgi retrograde transport in yeast. Journal of Cell Biology, 142(3), 665–681. Seaman, M., & Williams, H. (2002). Identification of the functional domains of yeast sorting nexins Vps5p and Vps17p. Molecular Biology of the Cell, 13(8), 2826–2840. direct traffic at late endosomes. Current Opinion in Cell Biology, 76, 102087. is an endosomal VPS9-domain GEF and SNX-BAR coat. ELife, 11, e77035. Shortill, S. P., Frier, M. S., & Conibear, E. (2022). You can go your own way: SNX-BAR coat complexes Shortill, S. P., Frier, M. S., Wongsangaroonsri, P., Davey, M., & Conibear, E. (2022). The VINE complex 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514

Simonetti, B., Guo, Q., Giménez-Andrés, M., Chen, K.-E., Moody, E. R. R., Evans, A. J., Chandra, M., binds ESCPE-1 to transfer cargo proteins during endosomal recycling. PLOS Biology, 20(4), Simonetti, B., Paul, B., Chaudhari, K., Weeratunga, S., Steinberg, F., Gorla, M., Heesom, K. J., Bashaw, G. J., Collins, B. M., & Cullen, P. J. (2019). Molecular identification of a BAR domaincontaining coat complex for endosomal recycling of transmembrane proteins. Nature Cell Biology, 21(10). https://doi.org/10.1038/s41556-019-0393-3 Steinberg, F., Gallon, M., Winfield, M., Thomas, E. C., Bell, A. J., Heesom, K. J., Tavaré, J. M., & Cullen, P. J. (2013). A global analysis of SNX27-retromer assembly and cargo specificity reveals a function in glucose and metal ion transport. Nature Cell Biology. Swarbrick, J. D., Shaw, D. J., Chhabra, S., Ghai, R., Valkov, E., Norwood, S. J., Seaman, M. N. J., & Collins, B. M. (2011). VPS29 is not an active metallo-phosphatase but is a rigid scaffold required for retromer interaction with accessory proteins. PLoS ONE. Van Weering, J. R. T., & Cullen, P. J. (2014). Membrane-associated cargo recycling by tubule-based endosomal sorting. Seminars in Cell and Developmental Biology, 31, 40–47. Van Weering, J. R. T., Sessions, R. B., Traer, C. J., Kloer, D. P., Bhatia, V. K., Stamou, D., Carlsson, S.

R., Hurley, J. H., & Cullen, P. J. (2012). Molecular basis for SNX-BAR-mediated assembly of distinct endosomal sorting tubules. EMBO Journal, 31(23), 4466–4483. van Weering, J. R. T., Verkade, P., & Cullen, P. J. (2010). SNX-BAR proteins in phosphoinositidemediated, tubular-based endosomal sorting. Seminars in Cell and Developmental Biology, 21(4), 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 reveals SNX5 and SNX6 as potential components of the mammalian retromer. Journal of Cell Science, 120(Pt 1), 45–54. https://doi.org/10.1242/JCS.03302 Wassmer, T., Attar, N., Harterink, M., van Weering, J. R. T., Traer, C. J., Oakley, J., Goud, B., Stephens, D. J., Verkade, P., Korswagen, H. C., & Cullen, P. J. (2009). The Retromer Coat Complex Coordinates Endosomal Sorting and Dynein-Mediated Transport, with Carrier Recognition by the trans-Golgi Network. Developmental Cell, 17(1), 110–122. Wolfe, K. H., & Shields, D. C. (1997). Molecular evidence for an ancient duplication of the entire yeast genome. Nature. https://doi.org/10.1038/42711 Yong, X., Zhao, L., Hu, W., Sun, Q., Ham, H., Liu, Z., Ren, J., Zhang, Z., Zhou, Y., Yang, Q., Mo, X., Hu, J., Billadeau, D. D., & Jia, D. (2021). SNX27-FERM-SNX1 complex structure rationalizes divergent trafficking pathways by SNX17 and SNX27. Proceedings of the National Academy of Sciences, 118(36). https://doi.org/10.1073/PNAS.2105510118 AlphaFold2-predicted interaction between Vrl1181-703 and the Vrl1-binding portion of the Vin1 N-terminus (Vin176-95). Interacting surface on Vrl1 shown in blue. (B) Basic residues within Vin176-95 are predicted to associate with Vrl1. (C) nonpolar residues in Vin176-95 are predicted to associate at an adjacent site on the Vrl1 AnkRD. (D) Diagram of assay used to test recruitment by chimeric Vrl1 construct (top) of the indicated pairwise alanine substitution mutants of the Vin1 N-terminus (bottom). (E) Quantification of RFP puncta per cell for Vin1 mutants. One-way ANOVA with Dunnett’s multiple comparison tests; n = 3, cells/strain/replicate ≥ 878; not significant, n.s. = p > 0.05, * = p < 0.05, *** = p < 0.001, **** = p < 0.0001. Blue statistical significance labels correspond to a Dunnett-corrected ANOVA performed against YPEcontaining bait while brown labels correspond to a Dunnett-corrected ANOVA performed against wildtype Vin1. (F) Differential effects of pairwise alanine substitution in the Vin1 N-terminus on recruitment by the AnkRD-containing Vrl1( 1-703 )YPE chimera. Representative images quantified in E. Scale bars, 2 543 544 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 way ANOVA with Dunnett’s multiple comparison tests; n = 3; not significant, n.s. = p > 0.05, * = p < 0.05, ** = p < 0.01, **** = p < 0.0001. Blue statistical significance labels correspond to a Dunnett-corrected ANOVA performed against untagged Vps29 while brown labels correspond to a Dunnett-corrected ANOVA performed against 3HA-tagged Vps29 with wild-type Vps5-mScI. (G) Disruption of the Vps5 Nterminal pocket- or sheet-binding sites cause CPY secretion. Paired disruption of the pocket- and sheetbinding sites (bipartite-ala) caused the most severe secretion phenotype. (H) Quantification of secreted CPY by densitometry and normalized to measured secreted CPY from the shown vps5Δ strain. One-way ANOVA with Dunnett’s multiple comparison tests; n=3; not significant, n.s. = p > 0.05, ** = p < 0.01, **** = p < 0.0001. Blue statistical significance labels correspond to a Dunnett-corrected ANOVA performed against a vps5Δ strain while brown labels correspond to a Dunnett-corrected ANOVA performed against a vps5Δ strain with plasmid-expressed wild-type Vps5-mScI. Error bars report SEM. aa, amino acid. mScI, mScarletI. Nt, N-terminus. 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640

Whereas deletion of the N-terminus (klVps5204-end) and alanine substitution of the polybasic predicted klVrl1-binding site causes loss of klVrl1 localization, disruption of either predicted klVps29 pocket- or sheet-binding sites have no effect. (E) Quantification of GFP puncta per cell in D. One-way ANOVA with Dunnett’s multiple comparison tests; n = 3, cells/strain/replicate ≥ 515; not significant, n.s. = p > 0.05, ** = p < 0.01, **** = p < 0.0001. Blue statistical significance labels correspond to a Dunnett-corrected ANOVA performed against the empty vector vin1∆vps17∆ strain while brown labels correspond to a Dunnett-corrected ANOVA performed against wild-type klVps5-mScI. (F) Disruption of the klVps5 Nterminus either by complete deletion or alanine substitution of the pocket- and sheet-binding sites causes CPY secretion. Alanine substitution of the predicted klVrl1-binding polybasic site in the klVps5 N-terminus does not create a significant CPY secretion phenotype. (G) Quantification of secreted CPY by densitometry and normalized to measured secreted CPY from the vps5Δ strain. One-way ANOVA with Dunnett’s multiple comparison tests; n=3; not significant, n.s. = p > 0.05, * = p < 0.05, ** = p < 0.01, **** = p < 0.0001. Blue statistical significance labels correspond to a Dunnett-corrected ANOVA performed against a vps5Δ strain while brown labels correspond to a Dunnett-corrected ANOVA performed against a vps5Δ strain with plasmid-expressed wild-type klVps5-mScI. (H) Model for VINE and retromer assembly with N-terminal binding sites highlighted for Vin1 and Vps5, respectively. A polybasic region unique to Vin1 drives interaction with the Vrl1 AnkRD, while a bipartite motif in the Vps5 N-terminus associates at two distinct sites in Vps29—a conserved hydrophobic pocket and a β-sheet on the opposite surface of the protein. The sole Vps5 isoform in K. lactis possesses all three of these N-terminal motifs and can form both VINE and retromer. Error bars report SEM. aa, amino acid. mScI, mScarletI. Nt, N-terminus. OE, overexpressed.

 Supplementary Figure and File Legends

Figure S1. Confidence measures of Vps29-Vps5 N-terminus binding prediction. (A) AlphaFold2predicted interaction between Vps29 and the N-terminus of Vps5 (Vps51-276) with pLDDT scores mapped to each residue of Vps29 (top) or Vps5 (bottom). In each case, the interacting partner is shown in gray. (B) AlphaFold2-generated predicted alignment error (PAE; Jumper et al., 2021; Mirdita et al., 2022) for Vps29 and Vps51-276 indicates an intermolecular interaction in the off-diagonal boxes. aa, amino acids. Figure S2. The SNX-BAR domains of Vin1 and Vps5 are indispensable for endosomal coat targeting. (A) Schematic of Vin1 and Vps5 truncations and chimeric fusion constructs used in B, C. (B) The fulllength Vin1-mScI protein is required to complement the loss of Vrl1-Envy puncta in vin1∆ cells. A chimeric protein with the N-terminus of Vin1 and SNX-BAR domains of Vps5 also fails to complement in both 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 vin1∆ and vin1∆vps17∆ cells. (C) The SNX-BAR domains of Vps5 are sufficient to partially complement the loss of Vps35-GFP puncta in vps5∆ cells. A chimeric protein with the N-terminus of Vps5 and SNXBAR domains of Vin1 fails to complement in both vps5∆ and vps5∆vps17∆ cells.

Figure S3. klVps5 binds Vps29 and partially complements CPY secretion in S. cerevisiae. (A) Both mScI-tagged S. cerevisiae and K. lactis isoforms of Vps5 co-purify with Vps29-3HA in S. cerevisiae. (B) Both mScI-tagged S. cerevisiae and K. lactis isoforms of Vps5 partially complement the CPY secretion in vps5∆ S. cerevisiae cells. For the S. cerevisiae Vps5 isoform, complementation depends on its unstructured N-terminus.

Envy.

Figure S4. klVps5 does not readily form VINE with scVrl1. (A) S. cerevisiae Vrl1-Envy and K. lactis Vps5-mScI both fail to localize in vin1∆vps17∆ cells. The artificially recruited Vrl1( 1-703 )YPE chimera does not recruit K. lactis Vps5-mScI to puncta. (B) The artificially recruited Vrl1( 1-703 )YPE chimera weakly recruits the K. lactis Vps5 N-terminus fused to mScI in vin1∆ cells. mScI, mScarletI. YPE, Ypt35(PX)Figure S5. klVps5 forms predicted BAR dimers with klVps17 and klVrl1. (A) AlphaFold2-predicted interaction between the K. lactis Vps5 and Vps17 BAR domains along the canonical BAR-BAR dimerization interface (top) and with pLDDT scores mapped to each residue (bottom). (B) AlphaFold2predicted interaction between the K. lactis Vps5 and Vrl1 BAR domains along the canonical BAR-BAR dimerization interface (top) and with pLDDT scores mapped to each residue (bottom). Figure S6. Confidence measures of klVps5 N-terminal binding predictions with klVrl1 or klVps29. (A) AlphaFold2-predicted interaction between K. lactis Vps29 and the N-terminus of K. lactis Vps5 (klVps51-265) with pLDDT scores mapped to each residue of Vps29 (bottom left) or Vps5 (bottom right). In each case, the interacting partner is shown in gray. AlphaFold2-generated predicted alignment error (PAE; Jumper et al., 2021; Mirdita et al., 2022) for klVps29 and klVps51-265 indicates an intermolecular interaction in the off-diagonal boxes (top right). (B) AlphaFold2-predicted interaction between the K. lactis Vrl1 VPS9 and AnkRD domains (klVrl1183-717) and the N-terminus of K. lactis Vps5 (klVps51-265) with pLDDT scores mapped to each residue of klVrl1 (bottom left) or Vps5 (bottom right). In each case, the interacting partner is shown in gray. AlphaFold2-generated predicted alignment error (PAE; Jumper et al., 2021; Mirdita et al., 2022) for klVrl1183-717 and klVps51-265 indicates an intermolecular interaction in the off-diagonal boxes (top right). aa, amino acids.

Supplementary File 1. List of Saccharomyces cerevisiae strains used in this study.

Supplementary File 2. List of plasmids used in this study.

A

VIN176-95 N

Length 1 100 200 300 400 500 600 700 800 900 1000 1100 (aa)

PX

BAR

C

VPS9 VPS9

AnkRD

PX-like BAR

C

R91 R90 R89 R92

R88 E86

K87

T85 S95

E94 F84

L83 F545

K541 n.s. **** *** * 100°

K79 Y80

V538

10°

D77 D498

L500 L497

L512 n.s. **** n.s.

****

AnkRD

E 0.6 ll e rC0.5 e P a tc0.4 n u P Ic0.3 S m in10.2 V e rag0.1 e v A 0

**** ** ** * *** D

GFP OE:

Vrl1X-D: pVIN11-116 -mScI

GFP

Merge

0.25 y v n 0.2 -E ll l1 e r tV rC0.15

e h P g i rB tca 0.1 ge un ra P0.05 e v A

0 Vin1XX-AA: **** **** **** **** ** ****

L497D ***** GFP OE:

YPE pVRL1( 1-703 )YPE vrl1 vin1∆ pVIN11-116-mScI

LF-AA pVIN1-mScI vrl1 vin1∆ pVRL1-Envy n.s.

**** RRb-AA

I32 L25 L266

D275

I264

V277 L196

L2 L252

P195 D194

Y265

Y263 V138

D162

A140 F161

L160

I159 E139

Pocket F192

Sheet F147

T142

R158 Y141 t N 5 s p V Nt

Vps5 aa155 Sheet Loop Pocket aa198

G K P R I L F D S A R A Q R N S K R N H S L K A K R T T A S D D T I K T P F T D P L K K s tn Vps5Pocket-Ala G K P R I L F D S A R A Q R N S K R N H S L K A K R T T A S D D T I K T P A A A A A K K ta Vps5Sheet-Ala G K A A A A A A A A R A Q R N S K R N H S L K A K R T T A S D D T I K T P F T D P L K K u MVps5Bipartite-Ala G K A A A A A A A A R A Q R N S K R N H S L K A K R T T A S D D T I K T P A A A A A K K Vps5LF-AA G K P R I A A D S A R A Q R N S K R N H S L K A K R T T A S D D T I K T P F T D P L K K Vps5L196K G K P R I L F D S A R A Q R N S K R N H S L K A K R T T A S D D T I K T P F T D P K K K vrl1 vps5∆ pVPS5-mScI WT WT275-ePnodckSeAthlaeeABtlaiparLtiFteA-laALA196K VPS29-3HA - + + + + + + + kDa 130 100 70 130 100 70 55 35 1

O D 6 10-1 00/

m 10-2 L

F

H vps5∆ W27T5-ePndockeAtSlaheBeiAptlaartiteAlaLF-AAL196K vrl1 120 FP iInP )T100

y toW80 -Rα its

ve 60 ge ten it rveaA Idnan lea 40

R20 y 120 its ) n 5∆100 e t s In vp 80 **** **** vrl1

n.s. n.s.

B (% 0 **** **** **** VPS29-3HA - + + + + + + + pVPS5-mScI WT W2T75-enPdockeAtlaSheBeAitplaartiteAla

LF-AAL196K ***** ****

VINE Retromer Both? Duplication

Pre-WGD yeast VRL1 VPS5 VPS17

WGD

S. cerevisiae VRL1 VPS5 VPS17

VIN1

K. lactis

VRL1 VPS5 VPS17 WT vps5∆ vps17∆

D OE pklVRL1-Envy * ****

E

WT vps5∆

vps17∆ vrl1 pklVPS5-mScI VIN1VPS17

vin1∆vps17∆ vrl1 OE pklVPS5-mScI A B C ge uP a r e v A

GFP Merge 100 200 300 400 500 600 700 200 400 600 800 1000

Loop klVps29 203

Pocket aa lkV uMklVps5Bipartite-Ala A A A A A A A A A R T R R H P I K P T Q E K A S I A A A A A A A

klVps5LF-AA K L N K A A S S A R T R R H P I K P T Q E K A S I S I H D P L V

WT

OE pklVPS5-mScI vin1∆vps17∆ 204-end BasicAla vrl1 OE pklVRL1 -Envy VINE AnkRD

A

WT

204-end BasicAla SheetAla PocketAla LF-AA

OE pklVRL1-Envy OE pklVPS5-mScI vrl1 vin1∆vps17∆ S. cerevisiae

Basic

K. lactis

AnkRD

Vrl1 B

Pocket

PX

BAR

C VPS9

AnkRD

PX

BAR

C T182

R183 α-CPY

R181

 AnkRD

Sheet

AnkRD 0.5 DO

6 0.25 0/m 0

L OE pklVPS5-mScI -

W20T4-enBdasicAlSaheeAPtlaockeAtlaLF-iApAartiteAla B 120 n**.*s*. n**.*s*. WT -end 204

Ala Basic klRetromer

Vps29

Vps17

 Arighi , C. N. , Harmell , L. M. , Aguilar , R. C. , Haft , C. R. , & Bonifacino , J. S. ( 2004 ). Role of the Danson , C. M. , Williams , T. A. , Collins, B. M. , & Cullen , P. J. ( 2022 ). SNX27 -Retromer directly Wassmer, T. , Attar , N. , Bujny , M. , Oakley , J. , Traer , C. , & Cullen , P. ( 2007 ). A loss-of-function screen 172 aa klVps5 klVps29 Sheet klVrl1 Site 0 . 125