1. Introduction

Polarized growth is crucial for various biological processes across yeast and filamentous fungi, which is achieved through the cytoskeleton-based directional transport of cargo to polarized domains [1]. As a result of its asymmetric growth and the polar delivery of organelles, the budding yeast, 

Saccharomyces cerevisiae, is considered to be the preferred model system for studying the mechanisms and molecules of polarized growth and faithful organelle inheritance in eukaryotic cells ^[1][2]. Rho GTPase Cdc42 is essential for the control of polarized growth during bud emergence, by recruitment of a yeast-specific complex called the polarisome, which is comprised of formin Bni1, nucleation-promoting factor (NPF) Bud6, Pea2, scaffolding protein Spa2 and receptor protein Epo1 ^[2][3][4]. During budding, Cdc42 also initiates the formation of a physical diffusion barrier at the neck, comprising septins, which compartmentalizes the bud plasma membrane (PM) from the mother ^[5][6]. However, distinct and not fully characterized protein complexes organize the contact sites between the PM and the endoplasmic reticulum.

, is considered to be the preferred model system for studying the mechanisms and molecules of polarized growth and faithful organelle inheritance in eukaryotic cells [1,2]. Rho GTPase Cdc42 is essential for the control of polarized growth during bud emergence, by recruitment of a yeast-specific complex called the polarisome, which is comprised of formin Bni1, nucleation-promoting factor (NPF) Bud6, Pea2, scaffolding protein Spa2 and receptor protein Epo1 [2,3,4]. During budding, Cdc42 also initiates the formation of a physical diffusion barrier at the neck, comprising septins, which compartmentalizes the bud plasma membrane (PM) from the mother [5,6]. However, distinct and not fully characterized protein complexes organize the contact sites between the PM and the endoplasmic reticulum.

Bem3 localizes to the sites of polarisome growth through its C-terminal Rho GTPase-activating protein domain, which negatively regulates a Rho-type GTPase Cdc42 ^[7][8][9]. This domain is preceded by a lipid-binding pleckstrin homology (PH) domain, a PX (phox) domain and an N-terminal region that harbors a predicted coiled-coil domain ^[8][10][11]. A previous study showed that the N-terminal coiled-coil domain of Bem3 interacts directly with the C-terminal coiled-coil domains of Epo1, which is a new member of the polarisome ^[2][12], suggesting a novel role for the polarisome in linking Bem3 to its functional target, Cdc42, during the budding process.

Bem3 localizes to the sites of polarisome growth through its C-terminal Rho GTPase-activating protein domain, which negatively regulates a Rho-type GTPase Cdc42 [7,8,9]. This domain is preceded by a lipid-binding pleckstrin homology (PH) domain, a PX (phox) domain and an N-terminal region that harbors a predicted coiled-coil domain [8,10,11]. A previous study showed that the N-terminal coiled-coil domain of Bem3 interacts directly with the C-terminal coiled-coil domains of Epo1, which is a new member of the polarisome [2,12], suggesting a novel role for the polarisome in linking Bem3 to its functional target, Cdc42, during the budding process.

Scs2 is a homolog of mammalian synaptobrevin-associated protein, which is a conserved integral ER protein and a component of a lipid-sensing complex ^[12][13]. Scs2 serves as anchors to the ER for cytoplasmic proteins (including Opi1p), through a conserved motif known as FFAT, two phenylalanines (FF) in an acidic tract ^[14][15][16]. Scs2 also contributes to the tethering of the ER to the septins and to the robust inheritance of the cER, in that its single deletion already leads to a severe reduction in the number of cER-PM contact sites and an up-regulated unfolded protein response ^{[12][14][16][17]}. Epo1, which was founded to be the PM-located receptor for Scs2, can promote cER tethering at sites of polarized growth ^[2][12]. In budding yeast, there exists an Scs2–Epo1–Bem3 polarisome complex that is required to keep ER tubules or the PM-attached cER close to the tip of the bud during tip growth. The Epo1-Scs2 connection might pull the cER actively into the bud, and then the connection between Epo1 and Scs2 is dissolved during the M phase (Mitosis phase) of the cell cycle ^[2]. Whether and how Epo1 assists Scs2 in its newly discovered roles as an ER-septin tether and in spindle positioning remain open questions for future experiments.

Scs2 is a homolog of mammalian synaptobrevin-associated protein, which is a conserved integral ER protein and a component of a lipid-sensing complex [12,13]. Scs2 serves as anchors to the ER for cytoplasmic proteins (including Opi1p), through a conserved motif known as FFAT, two phenylalanines (FF) in an acidic tract [14,15,16]. Scs2 also contributes to the tethering of the ER to the septins and to the robust inheritance of the cER, in that its single deletion already leads to a severe reduction in the number of cER-PM contact sites and an up-regulated unfolded protein response [12,14,16,17]. Epo1, which was founded to be the PM-located receptor for Scs2, can promote cER tethering at sites of polarized growth [2,12]. In budding yeast, there exists an Scs2–Epo1–Bem3 polarisome complex that is required to keep ER tubules or the PM-attached cER close to the tip of the bud during tip growth. The Epo1-Scs2 connection might pull the cER actively into the bud, and then the connection between Epo1 and Scs2 is dissolved during the M phase (Mitosis phase) of the cell cycle [2]. Whether and how Epo1 assists Scs2 in its newly discovered roles as an ER-septin tether and in spindle positioning remain open questions for future experiments.

To investigate the mechanism by which Epo1 anchors cER to the bud tip in yeast, we determine the X-ray structure of the C-terminal coiled-coil domains 2, 3 and 4 of Epo1 (named Epo1

^CC2-CC4

) in complex with the N-terminal coiled-coil domain of Bem3 (named Bem3

^CC1

). The structure reveals that the Bem3

^CC1

 domain forms a homodimer to bind four Epo1

^CC2-CC4

 molecules, with two CC3 domains of Epo1 providing an interface for binding to each Bem3

^CC1

. Moreover, through H/D exchange assay, we determine that the N-terminus 12 residue of Scs2 is responsible for binding to Epo1. Thus, Epo1 serves as a key receptor link between Scs2 and Bem3.

2. Bem3-Epo1 Complex Is a Hexamer

The N-terminal coiled-coil domain of Bem3 was previously reported to interact directly with the C-terminal domain of Epo1 [2] (

Figure 1

A). We started with the co-expression of the N-terminal domain of Bem3 (residue 1–99, CC1 domain) and the C-terminal domain of Epo1 (residue 746–943, CC2-CC4 domain) in 

E. coli

 BL21(DE3) cells, and purified the Bem3-Epo1 complex using size-exclusion chromatography. The fractions were analyzed using SDS-PAGE, which showed that Bem3

^CC1

 and Epo1

^CC2-CC4

 can form a stable complex (

Figure 1

B). To determine the accurate mass of the Bem3

^CC1

 and Epo1

^CC2-CC4

 complex, we performed analytical gel filtration combined with MALS. Interestingly, while both Bem3

^CC1

 and Epo1

^CC2-CC4

 in isolation act as dimers in solution, the mixture of Bem3

^CC1

 and Epo1

^CC2-CC4

 eluted from MALS corresponded to a hexamer (about 100 kDa molecular mass) (

Figure 1

C).

Figure 1. Functional domains and structures of Epo1 and Bem3. (A) Schematic diagram of domain structure of Epo1 and Bem3. Epo1 contains an N-terminal of unknown functional domain (residues 1–494) followed by four CC domains: CC1 is a predicted domain; CC2 to CC4 are side-by-side, with numbers indicating the amino acid positions of the start and endpoints of each domain. Bem3 contains CC1, PX, PH and GAP domains. (B) An elution profile shows the separation of the Epo1^CC2-CC4 and Bem3^CC1 complex from excess Bem3 using size-exclusion chromatography (left); confirmation of the purified Epo1 and Bem3 complex by Coomassie Brilliant Blue (CBB)-stained SDS/PAGE (right). (C) The sizes of the Epo1^CC2-CC4 and Bem3^CC1 complex are determined using MALS coupled with gel filtration. The data are representative of at least three repetitions.

To determine the atomic structure of the protein complex between Bem3

To determine the atomic structure of the protein complex between Bem3

^CC1

 and Epo1

^CC2-CC4

, we crystallized the Bem3

^CC1

–Epo1

^CC2-CC4

 complex and determined its structure to 3.5 Å resolution by X-ray crystallography (

Figure 2

A and 

Table S1, supplementary could be found in https://www.mdpi.com/1422-0067/22/8/3812). The model of the Epo1

). The model of the Epo1

^CC2-CC4

–Bem3

^CC1

 complex was built manually via several rounds of restrained individual atomic displacement parameter refinement, which allowed us to visualize the structure of the Epo1–Bem3 complex in detail. The final complex structure contained the heterologous hexamer of two Bem3

^CC1

 molecules and four Epo1

^CC2-CC4

 molecules in one asymmetric unit (

Figure 2

A), which was consistent with the aforementioned data obtained from MALS (

Figure 1

C), but the superposition of two Epo1

^CC2-CC4

 had obvious conformation changes (

Figure S1

).

Figure 2.

 Structure of the Epo1 and Bem3 complex. (

A

) Epo1 and Bem3 form the hexamer structure in the asymmetric unit, one Bem3 in complex with an Epo1 homodimer and two trimers side-by-side to form a heterologous hexamer; two promoters of Epo1 are colored violet and cyan, respectively. The Bem3 dimers show in the colors salmon and light blue (Bem3 and Bem3′). (

B

) The CC2 dimer interface of Epo1 is shown as sticks with the key residues highlighted. (

C

) The interface of the Epo1 CC4 dimer. (

D

) The sequence alignment of scEpo1 and different species are conserved and similar residues are highlighted with blue and light blue; residues involved in the CC3 dimer interface are indicated by a red triangle. “sc”, “cg”, “dh”, “eg” and “tb” represent Saccharomyces cerevisiae, Candida glabrata, Debaryomyces hansenii, Eremothecium gossypii and Tetrapisispora blattae, respectively.

3. The Overall Structure of Epo1

^CC2-CC4

–Bem3

^CC1

The Epo1

^CC2-CC4

–Bem3

^CC1

 complex is a heterologous hexamer with four Epo1

^CC2-CC4

 molecules and a Bem3

^CC1

 homodimer. The structure of Epo1

^CC2-CC4

 comprises three α-helix, a CC2 helix (amino acids 821 to 851), a longer CC3 (amino acids 866 to 905) and a C-terminal CC4 (amino acids 912 to 939). Similar to Bem3

^CC1

, two Epo1 molecules are aligned in parallel and interact directly with each other to form homodimers; the observed buried area between the two promoters is 3522 Å

²

. Two Epo1 homodimers located at the two sides of the Bem3 homodimer finally assemble into a heterologous hexamer (

Figure 2

A).

The complex structure reveals that the two Epo1 homodimers interacted with each other via two interfaces. In the first interface, residue S837 from Epo1 α1 stabilized α1′ by forming hydrogen bonds with S837, E841 and α2 S879 formed a hydrogen bond with α2′ Q844, while α2 K877 and α1 E841 formed a salt bridge (

Figure 2

B). The second region of intermolecular interactions was that of L924, I928 and I935 forming several hydrophobic interactions, which were buttressed by a salt bridge between α3 D932 and α3′ R931 (

Figure 2

C). Moreover, Epo1

^CC2-CC4

 was a highly conserved cross-species (

Figure 2

D).

4. Interaction of Epo1

^CC2-CC4

–Bem3

^CC1

As mentioned above, Bem3 formed a homodimer mainly via its CC1 helices; CC1 and CC1′ form the parallel contents between the coiled-coil dimer (

Figure 2

A and 

Figure 3

A). The Bem3

^CC1

 folded into a dimeric parallel coiled-coil that was 61.8 Å long, with a buried solvent-assessable area of 914 Å

²

 (

Figure 3

A). Mutations of Bem3 N56 and Y66A disrupted the Bem3 dimer (

Figure 3

C) and these parts constituted the primary interface for Bem3 bound to Epo1 (

Figure 4

A). The Epo1–Bem3 interface regions could be subdivided into a central compartment and a side compartment. The major interface, a hydrophilic region, comprised Epo1 α2 and α2′ to form a zipper with Bem3, consisting mainly of hydrogen bonds, involving residues from Bem3 (K58, Q62 and E69), Epo1 α2 (E872 and R876) and α2′ (Q888) (

Figure 4

A,B), and a salt bridge between E65 of Bem3 and K881 of the Epo1 α2′ helix (

Figure 4

B). The minor interface contained a local intermolecular hydrogen bond network, involving the side chains of three critical residues (Q906 of Epo1 α2′, R84 and E85 of Bem3′) and the carbonyl oxygen of Epo1 A899 (

Figure 4

C). In order to investigate the role of the above-described key residues involved in interactions between Epo1

^CC2-CC4

 and Bem3

^CC1

, we performed site-directed mutagenesis and a subsequent SPR experiment.

Figure 3.

 The dimer interaction of Bem3

^CC1.

 (

A

) The interface residues between the Bem3

^CC1

 dimer. (

B

,

C

) The sizes of wild-type (wt) Bem3

^CC1

 (theoretical molecular mass 11.39 kDa) and some key residue mutants were determined by MALS coupled with gel filtration. The estimated molecular masses are shown on the right axis. (

D

) Sequence alignment of Bem3 from different species: Saccharomyces cerevisiae (sc), Candida glabrata (cg), Debaryomyces hansenii (dh), Eremothecium gossypii (eg) and Tetrapisispora blattae (tb). The scBem3 are numbered and aligned, while the secondary structures of Bem3 are labeled on top.

Figure 4.

 Heterologous interaction of Epo1 and Bem3. (

A

) Specific interaction between Epo1 and Bem3, with key regions boxed. (

B

,

C

) CC3 of two Epo1 present at the interaction surface with Bem3-CC; a stick representation of the key residue is shown. (

D

) The specific interaction between Bem3-CC and different Epo1-CC (WT and mutant) is characterized by SPR. Epo1-CC is seen binding to Bem3, Epo1-CC-K881A to Bem3, Epo1-CC Q88A to Bem3 and Epo1-CC R84A to Bem3.

Consistent with our structural observations, a mutation of K881 disrupted the interactions between Epo1 and Bem3, while Q884A still maintained the stable interaction (

Figure 4

D). In addition, we tested the binding of Bem3 mutants to Epo1 using SPR, and found that the single-substitution of R84A, as well as inter-domain hydrogen bond network disruption, severely attenuated the interactions between Bem3 and Epo1 (

Figure 4

D).

5. Discussion

Epo1 proteins contain four coiled-coil (CC) domains: CC1 located at its N-terminus and CC2 to CC4 at its C-terminus. In this study, we determined the CC2-CC4 to be in a complex with the Bem3 CC1 domain. Protein folding propensity analyses indicated that CC2-CC4 region formed a dimer with an “L” shaped structure, for novelty, in the polarisome complex. In order to identify the key Epo1 CC domain for interaction with Bem3, we isolated CC2, CC3, and CC2-CC3, respectively. While the CC2 and CC3 clones showed no expression, CC2-CC3 could be expressed and bind with Epo1. Though Bem3 CC1 binding was not expected to interfere with the CC4 domain of Epo1, a deletion of CC4 would cause the Epo1–Bem3 interaction ability to decrease. This result suggested that the interaction between Bem3 and Epo1 might have been dependent on their coiled-coil domains. Indeed, as the CC1 domain of Bem3 docked on the CC3 surface of the Epo1 dimer, the overall Epo1-Bem3 structure may be relatively rigid for a hexamer. We did not find a detectable electron density between residues 746 and 808, indicating structural flexibility in this region.

Epo1 is critical in yeast bud development due to its function as a receptor that recruits Scs2 ^[2][12]. This article reports a key complex structure of Epo1–Bem3, which forms a specific contact to meet the specific demands of rapid membrane and cell wall extension at the cell tip. Specifically, this complex provides a platform for Scs2 interaction and contributes to the establishment of cell polarity, which are a fundamental processes in the life of a yeast bud.

Epo1 is critical in yeast bud development due to its function as a receptor that recruits Scs2 [2,12]. This article reports a key complex structure of Epo1–Bem3, which forms a specific contact to meet the specific demands of rapid membrane and cell wall extension at the cell tip. Specifically, this complex provides a platform for Scs2 interaction and contributes to the establishment of cell polarity, which are a fundamental processes in the life of a yeast bud.

In conclusion, our structure-function studies on the Epo1–Bem3–Scs2 complex reveal several important features. First, the N-terminus of Scs2 bound with Epo1 acts as a receptor of Scs2, and ORP1 and Opi1 also bind to Scs2 via the FFAT motif, which plays an important role in maintaining ER morphology. While we can only speculate on the dual function of Scs2, this is crucial for various biological processes. Second, the Bem3

^CC1

 dimer is the major determinant that is crucial for its polarized localization, as this segment can recruit the Epo1 protein in the absence of other structural elements of Bem3.