Interaction of the Endocytic Scaffold Protein Pan1 with the Type I Myosins Contributes to the Late Stages of Endocytosis

Abstract

The yeast endocytic scaffold Pan1 contains an uncharacterized proline-rich domain (PRD) at its carboxy (C)-terminus. We report that the pan1-20 temperature-sensitive allele has a disrupted PRD due to a frame-shift mutation in the open reading frame of the domain. To reveal redundantly masked functions of the PRD, synthetic genetic array screens with a pan1ΔPRD strain found genetic interactions with alleles of ACT1, LAS17 and a deletion of SLA1. Through a yeast two-hybrid screen, the Src homology 3 domains of the type I myosins, Myo3 and Myo5, were identified as binding partners for the C-terminus of Pan1. In vitro and in vivo assays validated this interaction. The relative timing of recruitment of Pan1-green fluorescent protein (GFP) and Myo3/5-red fluorescent protein (RFP) at nascent endocytic sites was revealed by two-color real-time fluorescence microscopy; the type I myosins join Pan1 at cortical patches at a late stage of internalization, preceding the inward movement of Pan1 and its disassembly. In cells lacking the Pan1 PRD, we observed an increased lifetime of Myo5-GFP at the cortex. Finally, Pan1 PRD enhanced the actin polymerization activity of Myo5–Vrp1 complexes in vitro. We propose that Pan1 and the type I myosins interactions promote an actin activity important at a late stage in endocytic internalization.

INTRODUCTION

Endocytosis is a coordinated series of molecular events that includes the identification and selection of cargo, coated pit formation, coated pit invagination, and vesicle scission that releases the vesicle into the interior of the cell. Recent studies in both mammalian cells and yeast have begun to order the recruitment of endocytic proteins (Kaksonen et al., 2003, 2005; Merrifield et al., 2004). These spatiotemporal colocalization studies correlate the function of a protein with the timing of its appearance in the process. In Saccharomyces cerevisiae, these endocytic events occur at dynamic cortical patches that contain actin and other proteins (Engqvist-Goldstein and Drubin, 2003); the proteins seen early at patches (Las17, Sla1, Sla2, and Pan1) bind to cargo, membrane phospholipids, and clathrin and function in the initial cargo selection and vesicle formation steps. The later arriving proteins (Myo5, Abp1, and Arp2/3 complex) interact with early proteins and function in the later steps of endocytosis, such as actin dynamics and vesicle scission (Kaksonen et al., 2003; Jonsdottir and Li, 2004).

Pan1 is an endocytic scaffold protein that is found at early patches, remains throughout patch maturation, and moves inward from the cell cortex as the new vesicle is formed and released (Kaksonen et al., 2003). PAN1 is essential in yeast, and Pan1 interacts with many different endocytic proteins (Figure 1A) (Sachs and Deardorff, 1992; Wendland et al., 1996). The functions of its binding partners suggest that Pan1 is a central factor in regulating transitions from early-to-late events in endocytic internalization. For example, the amino (N)-terminal region of Pan1 binds to several different proteins that are known or suspected cargo adaptor proteins: Sla1 (Tang et al., 2000), the yeast epsins Ent1 and Ent2 (Wendland and Emr, 1998), and the yeast CALM homologues Yap1801 and Yap1802 (Wendland et al., 1999), which are all predicted to act in early stages of internalization. The carboxy (C)-terminal region of Pan1 contains a region that binds to filamentous actin and an acidic (A) domain that binds to the Arp2/3 complex. Together, these domains cooperate to stimulate Arp2/3 complex activity and to promote actin polymerization (Duncan et al., 2001; Toshima et al., 2005). The later events in internalization, possibly including vesicle scission, involve actin polymerization (Yarar et al., 2005) and Arp2/3 complex recruitment (Merrifield et al., 2005). The central region of Pan1 mediates the formation of Pan1–Pan1 complexes (Miliaras et al., 2004), and also seems to link the N-terminal and C-terminal functions of Pan1 into a single protein complex (Miliaras et al., 2004; Toshima et al., 2005). Thus, Pan1 is poised to coordinate and regulate the sequence of early-to-late events necessary for efficient endocytic internalization.

Figure 1.

Pan1 and its C-terminal PRD. (A) Schematic representation of the Pan1 protein illustrating its major protein interaction domains. Under each region is a list of identified protein interactions. The black circles represent Prk1 phosphorylation sites. The C-terminal fragment used in the two-hybrid screen (aa1232–1480) is indicated by the black line. (B) Protein sequence alignment of the PRD of Pan1 and mutant pan1-20. The sequence starts at amino acid 1310 and ends at 1480 for the wild-type and 1488 for the mutant protein. The black line highlights the type I myosin SH3 domain ligand consensus sequence, PXXXPPXXP.

One region of Pan1 that has been almost completely uncharacterized is a proline-rich domain (PRD) at its extreme C-terminus (Figure 1, A and B). The PRD contains multiple candidate PXXP ligand motifs for binding Src homology (SH)3 or WW domains, suggesting that perhaps additional interactions exist for Pan1 that need to be accounted for in the modeling of Pan1 functions.

Type I myosins are highly conserved actin-based motors that participate in polarized morphogenesis, cell migration, and endocytosis (Geli and Riezman, 1996; Brown, 1997). Typically, myosins have N-terminal motor domains that bind actin in an ATP-dependent cycle and C-terminal tails that bind cargo through protein interaction domains. Type I myosins have an SH3 domain in their C-terminal tails, and fungal type I myosins also have an A domain at the extreme C-terminus that binds and activates the Arp2/3 complex (Geli and Riezman, 1996; Goodson et al., 1996; Evangelista et al., 2000; Lechler et al., 2000). MYO3 and MYO5 are the only genes encoding type I myosins in the Saccharomyces cerevisiae genome. Deletion of either gene does not affect growth; however, a double mutant is severely compromised (Goodson et al., 1996; Lechler et al., 2000). myo5Δ cells, but not myo3Δ cells, have a temperature-sensitive endocytosis defect, indicating that Myo5 may be more competent in endocytosis than Myo3 (Geli and Riezman, 1996). Myo3 and Myo5 both localize to cortical actin patches (Goodson et al., 1996), and a recent study reports that Myo5 appears transiently at cortical endocytic patches, immediately preceding the fast movement of vesicles away from the membrane, suggesting a function in vesicle scission (Jonsdottir and Li, 2004). Both the motor and Arp2/3 activities of Myo5 have been implicated in completing the late stages of endocytosis (Sun et al., 2006). In support of their endocytic roles, the C-terminal tails of Myo3 and Myo5 interact with multiple endocytic proteins: Las17, a yeast homologue of Wiskcott Aldrich Syndrome protein (WASP) (Evangelista et al., 2000); Vrp1, a yeast homologue of human WASP-interacting protein (Anderson et al., 1998); and Arc19 and Arc40, components of the Arp2/3 actin nucleating complex (Evangelista et al., 2000; Lechler et al., 2000). Vrp1 is an activator of the Arp2/3 actin polymerization activity of type I myosins in both S. cerevisiae and Schizosaccharomyces pombe (Sirotkin et al., 2005; Sun et al., 2006).

As more proteins are temporally placed along the timeline of endocytic events, a mechanistic understanding of the functions of each protein is needed to develop a complete model of endocytosis. This study describes our investigation into the function of the Pan1 PRD.

MATERIALS AND METHODS

Media and Materials

Yeast cells were grown in either standard rich medium (yeast extract-peptone) or synthetic medium (containing yeast nitrogenous base supplemented with amino acids for plasmid selection). Dextrose (glucose) as the standard carbon sugar source was added to a final concentration of 2%. 5-Fluoroorotic acid was used at 750 μg/ml in synthetic medium. Geneticin (G-418; Invitrogen, Carlsbad, CA) was used at 250 μg/ml in rich medium. Bacterial cells were grown on standard Luria broth media containing 50 μg/ml carbenicillin (US Biologicals, Swampscott, MA), 100 μg/ml ampicillin, 30 μg/ml kanamycin, and/or 34 μg/ml chloramphenicol as appropriate to maintain plasmids. Materials were obtained from Fisher Scientific (Pittsburgh, PA) or Sigma-Aldrich (St. Louis, MO) unless otherwise stated.

Yeast Strains and Plasmids

Standard methods were used for growth, sporulation, tetrad dissections, DNA manipulations, and transformations of yeast (Burke and Stearns, 2000). The yeast strains used in this study are listed in Table 1. BWY2043 (Pan1-GFP), BWY2319 (Myo3-RFP), and BWY2405 (Myo5-RFP) strains were generated by C-terminal chromosomal integration of polymerase chain reaction (PCR) products as described in Longtine et al. (1998). The monomeric red fluorescent protein (RFP) plasmid used for the RFP tagging was a gift from W. Huh (UCSF) (Huh et al., 2003). BWY2510, contains pan1ΔPRD-NAT in a synthetic genetic array (SGA) competent strain, was constructed by PCR-based double integration of a premature stop codon, with endogenous 3′ untranslated sequence by using (5′-gaagaagcaaagattggtcatcctgatcatgcacgtgct TGA agaatttaatttgctttctaagatacg-3′) and (5′-GTGACCCGGCGGGGACGAGGCAAGCTAAACAGATCTtgaagctgtaaaggaagaattggatcg-3′) along with a NAT resistance (Goldstein and McCusker, 1999) cassette by using (5′-AGATCTGTTTAGCTTGCCTCGTCC-3′) and (5′-cagaaattagtatacatacgtatctatagaaagcaaattaaatctGAATTCGAGCTCGTTTTCGACACTG-3′). The plasmids used in this study are listed in Table 2. DNA manipulations were performed using standard techniques. All restriction enzymes were purchased from New England Biolabs (Beverly, MA).

Table 1.

Yeast strains used in this study

Strain	Genotype	Source
AH109	Yeast two-hybrid screen stain	Clontech
SFY526	lacZ reporter strain	Clontech
DDY1810	MATa leu2-3,112 trp1-Δ901 ura3-52 prb1-1122 pep4-3 prc1-407 gal2	Drubin laboratory stain
BWY2043	MATα ADE his3-11,15 leu2-3,112 ura3-1 trp1-1 can1-100 PAN1-GFP-KAN	Miliaras and Wendland (2004)
BWY2319	MATa ade2-1 his3-11,15 leu2-3,112 ura3-1 trp1-1 can1-100 MYO3-RFP-KAN	This study
BWY2405	MATa ade2-1 his3-11,15 leu2-3,112 ura3-1 trp1-1 can1-100 MYO5-RFP-KAN	This study
BWY2409	MATa ade2-1 his3-11,15 leu2-3,112 ura3-1 trp1-1 can1-100 MYO3-RFP-KAN	This study
	MATα ADE his3-11,15 leu2-3,112 ura3-1 trp1-1 can1-100 PAN1-GFP-KAN
BWY2411	MATa ade2-1 his3-11,15 leu2-3,112 ura3-1 trp1-1 can1-100 MYO5-RFP-KAN	This study
	MATα ADE his3-11,15 leu2-3,112 ura3-1 trp1-1 can1-100 PAN1-GFP-KAN
SEY6210	MATa leu2-3 ura3-52 his3-Δ200 trp1-Δ901 lys2-801 suc2-Δ9	Laboratory stain
BWY1041	MATa leu2-3 ura3-52 his3-Δ200 trp1-Δ901 lys2-801 suc2-Δ9 pan1-20	Wendland et al. (1996)
BWY2469	MATα ade2-1 his3-11,15 leu2-3,112 ura3-1 trp1-1 can1-100 pan1::HIS myo3::KAN [p416::PAN1]	This study
BWY2471	MATα ade2-1 his3-11,15 leu2-3,112 ura3-1 trp1-1 can1-100 pan1::HIS MYO3-AΔW-NAT myo5::KAN [p416::PAN1]	This study
BWY2474	MATα ade2-1 his3-11,15 leu2-3,112 ura3-1 trp1-1 can1-100 pan1::HIS myo3::KAN MYO5-AΔW-NAT [p416::PAN1]	This study
BWY2510	MATa can1Δ::STE2pr-Sp_his5 lyp1Δ::STE3pr-LEU2 his3Δ1 leu2Δ0 ura3Δ0 met15Δ0 LYS2 pan1ΔPRD::NAT	This study
BWY2514	MATα can1Δ::STE2pr-Sp_his5 lyp1Δ::STE3pr-LEU2 his3Δ1 leu2Δ0 ura3Δ0 met15Δ0 LYS2 pan1ΔPRD::NAT	This study
BWY2515	MATα can1Δ::STE2pr-Sp_his5 lyp1Δ::STE3pr-LEU2 his3Δ1 leu2Δ0 ura3Δ0 met15Δ0 LYS2 pan1ΔPRD::NAT act1-121::KAN	This study
BWY2519	MATα can1Δ::STE2pr-Sp_his5 lyp1Δ::STE3pr-LEU2 his3Δ1 leu2Δ0 ura3Δ0 met15Δ0 LYS2 act1-121::KAN	This study
BWY2520	MATα can1Δ::STE2pr-Sp_his5 lyp1Δ::STE3pr-LEU2 his3Δ1 leu2Δ0 ura3Δ0 met15Δ0 LYS2	This study
BWY2521	MATα can1Δ::STE2pr-Sp_his5 lyp1Δ::STE3pr-LEU2 his3Δ1 leu2Δ0 ura3Δ0 met15Δ0 LYS2 pan1ΔPRD::NAT las17-14::KAN	This study
BWY2523	MATα can1Δ::STE2pr-Sp_his5 lyp1Δ::STE3pr-LEU2 his3Δ1 leu2Δ0 ura3Δ0 met15Δ0 LYS2 las17-14::KAN	This study
DDY2926	MATα ARP3::LEU2 ARC18-HA::HIS3 leu2-3,112 ura3-52 his3Δ200 lys2-801	Martin et al. (2005)

Table 2.

Plasmids used in this study

Plasmid	Description	Details	Source
pBW530	pDBD-Pan1-AP	pGBT9::PAN1 (aa1232–1480)	Miliaras and Wendland (2004)
pBW1175	pAD-Myo3	pACT2::MYO3 (aa973–1271)	This study
pBW1181	pAD-Myo3	pACT2::MYO3 (aa1037–1271)	This study
pBW1185	pAD-Myo5	pACT2::MYO5 (aa933–1215)	This study
pBW1209	pAD-myo3 SH3W1157S	pACT2::myo3 (aa973–1271)	This study
pBW1210	pAD-myo5 SH3W1123S	pACT2::myo5 (aa933–1169)	This study
pBW1204	pHIS6-Pan1-PRD	pET28c::PAN1 (aa1310–1480)	This study
pBW1206	pHIS6-pan1-20-PRD	pET28c::pan1-20 (aa1310–1488)	This study
pBW1226	pHIS6-Pan1-PRDΔ12	pET28c::PAN1 (aa1310–1468)	This study
	pGST-Myo3 SH3	pGEX::MYO3 (aa1120–1189)	Evangelista et al. (2000)
pBW1214	pGST-Myo5 SH3	pGEX::MYO5 (aa1088–1149)	This study
	pMyo3-myc	pRS315::MYO3-myc	Evangelista et al. (2000)
	pMyo5-myc	pRS416::MYO5-myc	Evangelista et al. (2000)
pBW940	p415-PAN1	pRS415::PAN1	This study
pBW941	p415-pan1-20	pRS415::pan1-20	This study
	pSTE6-GFP	pRS426::STE6-GFP	Huyer et al. (2004)
pDD1213	pVrp1-myc	pRS426::Gal-Vrp1-TEV-myc	Sun et al. (2006)
pDD1705	pMyo5-myc	pRS426::Gal-Myo5-TEV-myc	Sun et al. (2006)

Morphological Analysis of Cells by Electron Microscopy

Cells were grown overnight to an early log phase in selective media and prepared by glutaraldehyde fixation as described previously (Rieder et al., 1996). Sections were cut on a Reichert Ultracut T ultramicrotome, poststained with uranyl acetate and lead citrate, and observed on a Philips TEM 420 at 80 kV. Images were captured with an Olympus Soft Imaging Solutions (Münster, Germany) Megaview III digital camera.

Genetic Analysis

SGA analysis was performed as described previously (Tong and Boone, 2006). When using the temperature sensitive allele library, the cells were grown at room temperature (22°C) and for 1 d longer at each step. For the final double mutant selection, three sets of plates were pinned and grown at 22, 26, and 30°C.

Yeast Two-Hybrid Screening and Quantitative β-Galactosidase Analysis

A two-hybrid screen (Phizicky and Fields, 1995) was performed by transforming the FRYL library (Fromont-Racine et al., 1997) into AH109 cells (Clontech, Mountain View, CA) containing pBW530, a plasmid of Gal4-DNA-binding domain fusion of Pan1p-APRD (amino acid [aa] 1232–1480), and grown on YNB-TRP-LEU-HIS-ADE media. From ∼750,000 transformants, 25 positive clones were obtained, and the activation domain library plasmid was rescued and confirmed by retransforming into AH109 with pBW530 before being sequenced. For quantitative analysis, SFY526 cells transformed with a Gal-DNA–binding domain plasmid (pBW530) and transcriptional-activation domain plasmid were grown in selective liquid culture to mid-log phase and assayed for the expression of Galp-LacZ as described previously (Jarvis et al., 1988). Each value represents the average and SD for three independent quantifications.

Recombinant Pull-Down Experiments

Bacterial lysates were prepared from Rosetta cells containing either a pGEX-based glutathione S-transferase (GST)-fusion plasmid or a pET28c-based His₆-fusion plasmid. Cells were grown to 0.6 ODs in SB media containing 34 μg/ml chloramphenicol and 50 μg/ml carbenicillin or 30 μg/ml kanamycin, respectively, and then they were induced with 0.1 mM isopropyl β-d-thiogalactoside for 3 h at 37°C. Cells were harvested by centrifugation, washed, and resuspended in lysis buffer [20 mM Tris, pH 7.5, 1 mM EDTA, and 4-(2-aminoethyl)benzenesulfonyl fluoride protease inhibitors]. Lysis was achieved by a slow freeze at −20°C and a gentle thaw to room temperature. Then, lysozyme was added to 400 μg/ml and the lysates were incubated and mixed on a Nutator at room temperature for 20 min, finished with a final step of sonication on ice. The lysates were cleared with a 15,000 rpm in an SS-34 rotor (Sorvall, Newton, CT) at 4°C, and KCl was added to 150 nM and 0.2% Tween 20. Lysates were stored at −80°C. Thawed His₆ or GST construct lysates were incubated with nickel agarose beads (QIAGEN, Valencia, CA) or glutathione agarose beads, respectively, for 1 h at room temperature to allow for binding. Protein-bound beads were washed three times with phosphate-buffered saline (PBS), and then they were tested for approximately equal concentrations of protein-bound by SDS-polyacrylamide gel electrophoresis (PAGE), standard Coomassie staining, and immunoblot analysis. To test for binding between the PRD and SH3 domains, protein-bound beads were incubated with lysates of the target for 1 h at room temperature. Nickel beads with bound His₆-PRD proteins were incubated with GST lysates and glutathione beast with GST-bound SH3 domains were incubated with His₆-PRD lysates. Beads were subjected to gentle centrifugation and washed three times with PBS. The initial supernatants and bound protein pellets were denatured in Laemmli buffer and run on SDS-PAGE gels for immunoblotting.

Coimmunoprecipitation Experiments

DDY1810 cells transformed with a myosin-myc plasmid were grown to mid-log phase and dounced in lysis buffer (20 mM HEPES, pH 6.8, 0.2 M sorbitol, 50 mM KOAc, and 2 mM EDTA) with a protease inhibitor cocktail (Roche Diagnostics, Indianapolis, IN). Extracts were prepared from the P3 pellet to a concentration of 20 ODs/ml in lysis buffer with 1% Triton X-100. Immunoprecipitations were performed with 500 μl of extracts, containing 10 ODs, polyclonal antibody against Pan1 (gift from A. Sachs, UC Berkeley), and protein A-Sepharose beads (experiments were repeated 5 times for Myo5 and 2 times for Myo3). For immunoblot analysis, 0.5% of the extracts and 10% of the immunoprecipitations were loaded for detection by using a monoclonal mouse Myc antibody (Santa Cruz Biotechnology, Santa Cruz, CA).

Protein Purification

Rabbit skeletal muscle actin, hemagglutinin (HA)-tagged Arp2/3 complex, Myo5, and Vrp1 were purified as described previously (Martin et al., 2005; Toshima et al., 2005; Sun et al., 2006).

GST-Pan1 PRD was expressed in and purified from the protease-deficient DDY1810 yeast strain. Cells were washed, and cell pellets frozen in liquid N₂ were resuspended in PBS-KCl buffer (20 mM phosphate, 150 mM NaCl, 100 mM KCl, and 2 mM dithiothreitol). The cells were lysed using an EmulsiFlex C-3 (Avestin, Ottawa, Ontario, Canada) for 5 min at pulses of 21,000 psi. The lysate was clarified at 30,000 × g for 10 min at 4°C. The lysate was incubated with glutathione agarose beads (Invitrogen) for 3 h at 4°C, washed with the PBS-KCl buffer, and the GST fusion protein was eluted with a 50 mM glutathione, 50 mM Tris, pH 7.4, solution. Protein concentrations were determined using SYPRO dye (Invitrogen) with bovine serum albumin as a standard. GST and GST-Pan1 PRD were dialyzed in 20 mM HEPES, pH 7.5, 1 mM EDTA, 50 mM KCl, and 5% (vol/vol) glycerol.

SDS-PAGE and Immunoblotting

Proteins were separated on polyacrylamide mini gels (7.5–15%) at 18–25 mA in SDS running buffer (3 mM SDS, 25 mM Tris base, and 192 mM glycine), and then they were transferred onto nitrocellulose membranes at 80 V for 90 min in cold transfer buffer (20% methanol, 0.0375% SDS, 48 mM Tris base, and 39 mM glycine). The membranes are blocked in 5% milk in Tris-buffered saline/Tween 20 (TBST) (10 mM Tris, pH 7.5, 0.25 M NaCl, and 0.025% Tween 20). Blots were incubated in the specified primary antibody, washed three times in TBST, incubated with secondary antibodies conjugated to horseradish peroxidase (Pierce Chemical, Rockford, IL), and diluted 1:5000 in milk solution for 45–60 min. Blots were washed again three times in TBST, and then they were developed with SuperSignal West Pico Chemiluminescent Substrate (Pierce Chemical) for 5 min at room temperature. The chemiluminescence was visualized on a Fluorochem 8000 chemiluminescence system (Alpha Innotech, San Leandro, CA).

Dual Fluorescence Microscopy

Fluorescent images were captured using a Sensicam QE charge-coupled device camera (Cooke, Romulus, MI) on an Axiovert 135 TV inverted microscope (Carl Zeiss, Jena, Germany) equipped with a 100×/1.4 numerical aperture objective, Ludl motorized filter wheels (Ludl Electronic Products, Hawthorne, NY), fluorescein isothiocyanate (FITC) and Texas Red filter sets (Semrock, Rochester, NY), and IPLab software (Scanalytics, Fairfax, VA). In the capture two-color movies, IPLab software was used to drive the Sensicam QE camera and motorized filter wheels (Ludl Electronic Products) with FITC and Texas Red filter sets paired with a 4,6-diamidino-2-phenylindole/FITC/Texas Red multiband dichroic (Semrock). For the live cell imaging, cells were grown to early log phase on rich medium plates containing excess adenine at 26 or 30°C. Cells were placed in 2 μl of complete minimal media on the surface of an uncoated glass coverslip, and then they were inverted onto a glass slide. All imaging was performed at room temperature. Image analysis was performed using National Institutes of Health ImageJ (http://rsb.info.nih.gov/ij/) or SlideBook software (Intelligent Imaging Innovations, Denver, CO).

Actin Nucleation Assays

Actin assembly was performed using 2 μM rabbit skeletal muscle actin (5% pyrene labeled) as described previously (Sun et al., 2006), and assembly was monitored using a FluoroMax 3 fluorometer (Jobin-Yvon Horiba, Edison, NJ) with an excitation and emission of 365 and 407 nm, respectively. Barbed end concentrations were calculated using the rate of actin polymerization at 50% polymerization and a rate constant of 8.7 μM⁻¹ s⁻¹, as described previously (Higgs and Pollard, 1999).

RESULTS

The Temperature-sensitive Allele pan1-20 Has a Defect in the PRD

pan1-20 is a temperature-sensitive allele generated by ethyl methane sulfonate treatment that was identified in a screen for endocytic mutants a decade ago (Wendland et al., 1996). Strains carrying pan1-20 have been broadly used as a canonical endocytic mutant; however, the mechanistic nature of the mutation has not been characterized. To better understand this allele and to place the experimental findings of pan1-20 cells in a molecular context, we sequenced the PAN1 open reading frame from pan1-20 cells, and we identified a single nucleotide deletion of guanine 4285. This deletion causes a shift in the open reading frame of the protein starting at amino acid 1429, which occurs in the middle of the PRD. The mutant protein is eight amino acids longer than wild type; however, the last 60 residues differ from the wild-type protein (Figure 1B).

As reported previously, the pan1-20 allele causes lethality, endocytic defects, actin cytoskeletal abnormalities, and extended membrane invaginations at high temperatures (Wendland et al., 1996). However, the functional consequences of a significant frameshift mutation should be apparent at all temperatures, although higher temperatures may exacerbate certain effects or phenotypes. To more rigorously assess the status and stability of the pan1-20 mutant protein, we performed quantitative Western blotting of cell extracts prepared from wild-type and pan1-20 cells grown at either 26°C or shifted to 37°C for 3 h. This experiment showed that full-length, intact, pan1-20 protein was present at roughly 70% of wild-type levels at 26°C and at one third of wild-type levels at 37°C (Figure 2A). When we examined the phenotypes of pan1-20 cells at the permissive temperature of 26°C, we observed endocytic defects as seen by the stabilization of the endocytic cargo Ste6-GFP at the plasma membrane (Figure 2B; Liu et al., 2007) and mild actin phenotypes (Wendland et al., 1996). Morphological analysis by electron microscopy of pan1-20 cells grown at permissive temperature did not show extended invaginations. However, there were approximately twofold more membrane invaginations in pan1-20 cells (42 invaginations in 24 cells) than in wild-type cells (23 invaginations in 26 cells) (Figure 2C). These data show that the phenotypes of pan1-20 cells are indeed present at permissive temperatures, although in milder form than at the nonpermissive temperature, and that they are consistent with a defect in endocytic internalization at the point of vesicle scission. Thus, the C-terminus of Pan1 may have a distinct role in endocytosis.

Figure 2.

pan1-20 cells have endocytic defects at permissive temperature. (A) Analysis of Pan1 protein in wild-type and pan1-20 cells. Trichloroacetic acid (TCA) extracts of cells, at 26°C or grown at 26°C and shifted to 37°C for 3 h, were prepared and analyzed by immunoblot analysis with anti-Pan1 antibodies. Each band was quantified and the value relative to wild type grown at 26°C (100%) is reported below each lane. (B) Ste6-GFP localization in mid-log phase wild-type and pan1-20 cells. Bar, 5 μm. (C) Conventional electron microscopy of wild-type (WT) and pan1-20 cells. Arrows indicate invaginations, putative sites of vesicle scission. n, nucleus; v, vacuole; m, mitochondria. Bar, 0.5 μm.

A Genetic Analysis of the PRD of PAN1

The identification of an altered PRD corresponding to the pan1-20 allele is the first report of the importance of this region within the protein. However, the PRD of PAN1 is not normally essential to cells, because previous studies have failed to observe any fitness defect or temperature sensitivity when the PRD was deleted (Sachs and Deardorff, 1992; Duncan et al., 2001; Miliaras et al., 2004). It is possible that the temperature-sensitive and endocytic defects of the pan1-20 mutation might be due to more than the loss of the C-terminal sequence within the PRD, perhaps through a subtle effect on protein stability or the additional nonnative C-terminal residues (Figure 2A). Nonetheless, our findings highlight a potential importance for the PRD. Many endocytic proteins are redundant or share overlapping functions with other proteins that can mask their individual contributions (Engqvist-Goldstein and Drubin, 2003). Therefore, to determine whether there is an in vivo role for the Pan1 PRD, we sought to identify genetic backgrounds that caused specific sensitivity to loss of the Pan1 PRD. A pan1ΔPRD strain was engineered by introducing a stop codon at aa1310 and removing the coding sequence 3′ of this site while maintaining 200 base pairs of the 3′ untranslated region, followed by an NAT resistance marker. Western blots confirmed that the pan1ΔPRD strain produced a truncated protein at levels similar to endogenous protein (Figure 3A).

Figure 3.

Synthetic genetic interactions of pan1ΔPRD. (A) The pan1ΔPRD strain was confirmed by immunoblot analysis. TCA lysates from a wild-type and the pan1ΔPRD cells were prepared, and an immunoblot analysis was performed with anti-Pan1 antibodies. (B) Synthetic sick interactions with las17-14 and act1-121. Tetrads resulting from crosses with pan1ΔPRD. (C) Synthetic lethal interaction with sla1Δ. The circles highlight the double mutants. P, parental ditype; NPD, nonparental ditype; T, tetratype. (D) Ste3 stabilization/endocytosis was determined by treating each strain with 5 μg/ml cycloheximide, and removing aliquots at each time. Lysates were probed with rabbit anti-Ste3 antibodies and quantified by chemiluminescent detection.

We first screened the pan1ΔPRD strain against an unbiased library of temperature-sensitive KanMX-marked alleles. This collection was constructed by integrating previously characterized temperature-sensitive alleles into the S288C genetic background (Li and Boone, personal communication). The collection contains alleles of 348 genes that are either essential or severely sick as deletions, covering 25 of the 33 biological process designations curated in the gene ontology terms (see Supplemental Material). To identify synthetic lethal or synthetic sick double mutant genetic interactions, we performed an SGA analysis (Tong et al., 2001; Tong and Boone, 2006) against the library of temperature-sensitive alleles by using the pan1ΔPRD query strain. Synthetic sick interactions were identified with las17-14 and act1-121, and they were subsequently confirmed by tetrad analysis (Figure 3B). Thus, this broad screen identified interactions with ACT1, which encodes actin, and with LAS17, the yeast WASP homologue, which controls the actin nucleating activity of the Arp2/3 complex and functions at endocytic sites. In cycloheximide chase experiments monitoring the internalization and degradation of endogenous Ste3 at the permissive temperature of 26°C, the double mutants had greater endocytic defects than the single mutants (Figure 3D).

We also used a focused set of haploid deletion strains, corresponding to 59 genes with known endocytic, actin, and membrane functions (see Supplemental Material). This second round of SGA analysis with pan1ΔPRD revealed a synthetic lethal interaction with sla1Δ (Figure 3C). Sla1 binds to the C-terminus of End3 and to the N-terminus of Pan1; this complex performs unknown functions in endocytic internalization (Tang et al., 2000; Zeng et al., 2001). The interactions identified in these genetic studies highlight a role for the PRD of Pan1 in actin dynamics and endocytic internalization, which is consistent with the phenotypes observed in pan1-20 cells.

The Pan1 PRD Directly Interacts with the SH3 Domains of the Type I Myosins Myo3 and Myo5 In Vitro and In Vivo

To identify protein interaction partners for the PRD of Pan1, we performed a yeast two-hybrid screen with a fragment of Pan1 (aa1232–1480), containing both the A domain and the PRD. The A domain was included because constructs of the PRD alone proved to be self-activating in the two-hybrid system. From a genomic yeast two-hybrid library (Fromont-Racine et al., 1997), multiple clones of the type I myosins Myo3 and Myo5 were identified (Table 3). All of the clones obtained in the screen contained the C-terminal SH3 domain of the myosins (Figure 4A). To determine whether the interaction with Pan1 was through the SH3 domain, we made single point mutations at a conserved tryptophan residue in the myosin SH3 domain: W1157S in Myo3 and W1123S in Myo5. This mutation causes Myo3 mislocalization in vivo and reduces SH3 domain-dependent physical interactions of Myo3 with its other known partners in yeast (Evangelista et al., 2000; Geli et al., 2000). We have observed a similar mislocalization with the corresponding mutation in Myo5 (Barker and Wendland, unpublished data). The SH3 domain mutants did not interact with the C-terminus of Pan1 in the two hybrid assay (Table 4), demonstrating that the interaction with Pan1 is dependent on the SH3 domain of the type I myosins.

Table 3.

Two-hybrid interactions of Pan1 C-terminal tail (aa1232–1480) with the C-terminal tails of the type I myosins Myo3 and Myo5

DNA binding domain fusion	Activation domain fusion	Lac Z expression (Miller units)
Pan1 AP (aa1250–1480)	Vector	0.18 ± 0.03
	Myo3 (aa973–1271)	6.48 ± 0.62
	Myo3 (aa1037–1271)	3.89 ± 0.27
	Myo5 (aa933–1169)	17.21 ± 1.22

Figure 4.

The type I myosins Myo3 and Myo3 interact with the PRD of Pan1. (A) Schematic representation of the general domain structure of the type I myosins: with the N-terminal myosin motor domain, three IQ repeats where myosin light chains bind, tail homology (TH) domain 1, TH2, an SH3 domain, and a C-terminal acidic (A) domain. The sequences identified in the yeast two-hybrid clones are represented as bars underneath. (B) His₆ pull-downs. Nickel beads loaded with His₆ or His₆-Pan1 PRD were incubated with lysates of GST, GST-Myo3 SH3, or GST-Myo5 SH3. Bound proteins were detected by immunoblot analysis with antibodies to GST. (C) Reciprocal GST pull-downs. Glutathione beads loaded with GST, GST-Myo3 SH3, or GST-Myo5 SH3 were incubated with lysates of His₆, His₆-Pan1 PRD, or His₆-Pan1 PRDΔ12. Bound proteins were detected by immunoblot analysis with antibodies to His₆. (D) Coimmunoprecipitation of Myo3 and Myo5. Extracts prepared from cells expressing Myo3-myc (left) or Myo5-myc (right) were immunoprecipitated with anti-Pan1. Immunoprecipitated proteins were detected by immunoblot analysis with anti-Myc.

Table 4.

Two hybrid interactions of Pan1 (aa1250–1480) with type I myosin C-terminal tails are abolished with single point mutations (W->S) in their SH3 domains

DNA binding domain fusion	Activation domain fusion	Lac Z expression (Miller units)
Pan1 AP (aa1250–1480)	Vector	0.10 ± 0.00
	Myo3 (aa973–1271)	6.02 ± 0.48
	Myo3W1157S (aa973–1271)	0.12 ± 0.02
	Myo5 (aa933–1169)	16.59 ± 1.54
	Myo5W1123S (aa933–1169)	0.13 ± 0.00

To determine whether the interaction of both Myo3 and Myo5 with Pan1 is direct or mediated through interactions with other endocytic proteins, we recombinantly expressed the PRD of Pan1 and the SH3 domains of the myosins, and we performed pull-down experiments from bacterial lysates. Experiments with His₆-Pan1 PRD bound to nickel beads and incubated with GST-SH3 lysates demonstrated a direct interaction between the PRD of Pan1 and the SH3 domains of both type I myosins (Figure 4B). Reciprocal experiments performed with GST-SH3 domains bound to glutathione beads incubated with His₆-Pan1 PRD lysates confirmed the same interactions (Figure 4C). The recombinant His₆-Pan1 PRD fragment migrated at a molecular mass larger than predicted (20.5 kDa predicted vs. 38 kDa observed); however, this is not uncommon for proteins rich in proline residues. Recombinant SH3 domain mutants (myo3^W1157S and myo5^W1123S) failed to interact with the PRD of Pan1 in similar pull-down experiments (data not shown). Thus, the interaction between Pan1 and the type I myosins is mediated through a classical proline-rich motif binding to an SH3 domain, respectively.

A consensus sequence PxxxPPxxP has been reported as the ligand for the type I myosin SH3 domains, and a single predicted ligand is present within the last 12 residues of the PRD of Pan1 (Tong et al., 2002). To test whether this C-terminal ligand is important for the interaction between Pan1 and the myosins, we made a truncated construct, His₆-Pan1 PRDΔ12 (aa1310–1468), in which the last 12 amino acids were removed. The GST-SH3 domains did not bind this truncated His₆-Pan1 PRDΔ12 from bacterial lysates (Figure 4C). Consistent with this, the smaller proteolytic fragments of the affinity-purified N-terminally tagged His₆-Pan1 PRD, which would be expected to lack the extreme C-terminal residues, also did not bind to the GST-SH3 domain proteins (Figure 4C). Together, these data demonstrate that the last 12 residues of Pan1 are required for the binding of Pan1 to the type I myosins.

The Pan1–myosin interaction was confirmed as a bona fide in vivo interaction through coimmunoprecipitation of both Myo3 and Myo5 with Pan1 (Figure 4D). Endogenous Pan1 was immunoprecipitated from cells containing either Myo3-myc or Myo5-myc encoded by a plasmid, and the bound myosins were detected using an anti-Myc antibody. This confirms that the interaction can occur in the context of the full-length proteins. Although the degree of coimmunoprecipitation was somewhat weak, these results are reproducible and consistent with a transient interaction in vivo (see below). Together, these data indicate that the Pan1 and the type I myosins interact directly in vivo.

Spatiotemporal Colocalization Analysis of Pan1 and the Type I Myosins

With the biochemical confirmation of an interaction between Pan1 and the type I myosins, we proceeded to examine the spatial and temporal nature of the interaction in living cells. Pan1 was tagged with GFP at its C-terminus by integration at the PAN1 locus. The dynamics of Pan1-GFP were consistent with previous reports, having a lifetime of ∼30–40 s and culminating with movement toward the interior of the cell, which is interpreted as invagination or vesicle scission (Kaksonen et al., 2003). Myo3 and Myo5 were tagged with monomeric RFP at their C-termini by integration at their respective chromosomal loci. The Myo5-RFP had a lifetime similar to that in previous reports (Jonsdottir and Li, 2004; Sun et al., 2006), and unlike Pan1-GFP, the Myo5-RFP remained at the cell cortex for its duration. We observed Myo3-RFP to behave similarly to Myo5-RFP, as expected, also showing no internal movement. Statistically, we found no difference between the behavior of the myosins; however, the Myo5-RFP signal had a greater intensity.

To characterize the spatial and temporal features of the Pan1–myosin interaction, we mated Pan1-GFP cells with cells expressing a myosin-RFP and observed the colocalization of green and red fluorescence in the resultant diploid cells. Experiments were performed with both Myo3-RFP and Myo5-RFP; however, we only show Myo5 because the behavior of the myosins was indistinguishable. In single, static images, Pan1-GFP and the myosin-RFPs have a partial colocalization (Figure 5A), with some patches showing colocalization and some patches containing only Pan1-GFP or Myo-RFP. However, when analyzed temporally by two-color time-lapse movies, complete colocalization was ultimately seen for all patches: each Pan1-GFP patch was joined by myosin-RFP in its lifetime (Figure 5, B and C) (n = 31 for Myo5-RFP; n = 9 for Myo3-RFP). These results are consistent with the relative timing of protein behavior reported by Sun et al. (2006). Interestingly, the myosin always colocalized with Pan1 just before its internal movement, and then the myosin stayed at the cell cortex, whereas Pan1 moved inward (Figure 5, B and C). The intensities of both Pan1 and the myosin dissipated around the same time (Figure 5C). The precise timing of colocalization behavior suggests that the Pan1–myosin complex is poised to play a role in vesicle invagination and/or scission.

Figure 5.

Spatial and temporal colocalization of Pan1 and Myo5. (A) Single frames from the GFP and RFP channels of a movie and a merged image show the colocalization of Pan1-GFP and Myo5-RFP. Time to acquire one pair was 1.5 s. Bar, 5 μm. (B) Kymograph representation of Pan1-GFP (top) and Myo5-RFP (middle) in a single patch over time. The curvature of Pan1p indicates the patch is moving off the cortex, toward the interior of the cell. (C) Time series showing composition of a single patch over 45 s. The time between each frame is 1.5 s. Top and middle, two separate channels, GFP and RFP respectively; bottom, merged image. (D) Kymographs of Myo5-GFP in wild-type, pan1Δprd, act1-121, pan1Δprd act1-121, las17-14, and pan1Δprd las17-14 cells. Images were collected every second for 120 s. (E) A bar graph indicating the average Myo5-GFP lifetimes from the cells shown in D; n = 20 patches for wild-type and pan1Δprd, n = 10 patches for all others. Asterisk (*) indicates p < 0.0001 in paired Student's t tests for the two strains.

It has been previously shown that type I myosins localize to the membrane in an SH3-dependent manner (Anderson et al., 1998; Evangelista et al., 2000). To determine whether the PRD of Pan1 plays a role in the localization of the myosins, a Myo5-GFP plasmid was transformed into wild-type and pan1ΔPRD cells. A fluorescent GFP signal was observed at the membrane in both cells, indicating that the PRD of Pan1 does not influence the recruitment of Myo5 to the plasma membrane. However, we did observe an alteration in the lifetime of Myo5-GFP at the cortex; 13 ± 1 s in wild-type cells versus 20 ± 2 s in pan1ΔPRD cells (n = 20 spots for each cell type; paired Student's t test, p < 0.0001) (Figure 5, D and E). An increase in Myo5-GFP lifetime was also observed in pan1-20 cells at permissive temperature (data not shown). Having observed a physiological effect of the lack of the Pan1 PRD, we went on to measure the lifetime of Myo5-GFP in our other genetic backgrounds: act1-121, act1-121 pan1ΔPRD, las17-14, and las17-14 pan1ΔPRD. In the single act1-121 and las17-14 mutants, the lifetime of Myo5-GFP was longer than in wild-type cells; however, in double mutants combined with pan1ΔPRD, the lifetimes were significantly increased (Figure 5E). This increase in the lifetime of Myo5 at the membrane suggests that some function of the type 1 myosin is perturbed or delayed as a result of the absence of the PRD of Pan1.

Pan1 PRD Enhances the Actin Activity of the Type I Myosins

The type I myosins are one of four Arp2/3-stimulating actin nucleation promoting factors (NPFs). In fluorescent actin polymerization assays, the myosins on their own are weak activators, and they require the presence of verprolin, Vrp1, for more potent activity (Sirotkin et al., 2005; Sun et al., 2006). Because Vrp1 interacts with the SH3 domains of the myosins, we wanted to test whether the SH3 domain interaction with Pan1 had any effect on the NPF activity of the myosins. However, Pan1 itself is an NPF, with its own A domain. Therefore, we purified GST-PRD lacking the A domain to assess the effect of the PRD of Pan1 in isolation. GST-Pan1 PRD affected neither Arp2/3 nucleation activity nor Myo5 weak NPF activity (Figure 6 and data not shown). The addition of Vrp1 increased Myo5 NPF activity approximately threefold. GST-Pan1 PRD further enhanced the Arp2/3 complex activation by the Myo5p–Vrp1p complex: 75 nM GST-Pan1 PRD consistently increased the Myo5-Vrp1 NPF activity by 1.4-fold and 100 nM increased the activity by 1.8-fold, over that of Myo5 with Vrp1 and of Myo5 with Vrp1 and GST. Thus, the enhancement of Myo-Vrp1 NPF activity increased with increasing amounts of GST-Pan1 PRD.

Figure 6.

Pan1 PRD enhances Arp2/3 complex activation by the Myo5p–Vrp1p complex. Representative actin nucleation reactions. Rabbit skeletal muscle actin (2 μM; 5% pyrene labeled) was polymerized with 10 nM Arp2/3 complex, 15 nM Myo5p, 50 nM Vrp1, and 75 nM or 100 nM GST or GST-Pan1 PRD (PRD), when indicated.

The effect of Pan1 PRD is specific to the Myo–Vrp1 complex. The combination of GST-Pan1 PRD and Vrp1, in the absence of Myo5, had no NPF activity (data not shown). Similarly, when Myo5 concentration was increased to 50 nM, without Vrp1, PRD did not increase NPF activity (data not shown). In reciprocal experiments, the SH3 domains of the myosins had no effect on the NPF activity of Pan1 (data not shown). Together, results of these actin polymerization assays illustrate a direct effect for the PRD of Pan1 on the NPF activity of the Myo5–Vrp1 complex.

DISCUSSION

More than 60 proteins have been implicated in budding yeast endocytosis, but how their activities are coordinated, and how they influence each other, are not known. The type I myosins Myo3 and Myo5 are the first proteins shown to interact with the PRD of Pan1. Our biochemical studies confirm that this is a direct physical interaction, which represents a classical binding of an SH3 domain to a proline-rich motif (Figure 4). We have characterized the spatiotemporal dynamics of this interaction in vivo and shown that the myosins colocalize with Pan1 in the late stages of endocytosis (Figure 5). We also observed that in the absence of the Pan1 PRD, there is an extended lifetime of Myo5-GFP at cell surface patches (Figure 5D). In actin polymerization assays, the PRD of Pan1 enhanced the actin nucleating activity of Myo5 in a Vrp1-dependent manner. The effect of Pan1 PRD in the actin assays offers evidence that it might influence the activity of the myosins in vivo. In the absence of the Pan1 PRD, the activity of Myo5-Vrp1 is not maximal, which could explain the increased lifetimes of Myo5 at the plasma membrane in pan1 PRD mutants, pan1-20 and pan1ΔPRD. Together, these findings support a model in which the interaction of Pan1 PRD with the myosins contributes actin activity that is required for efficient endocytosis.

The most striking feature from the spatiotemporal studies of the Pan1–myosin interaction is the intriguing timing of the interaction: the myosin occurs late in the Pan1 patch lifetime, just before the inward motion of the Pan1 patch (Figure 5). The interaction is brief (∼6–8 s) and terminates with the inward movement of Pan1, whereas the myosin remains at the membrane cortex. The inward movement of Pan1 and other proteins is interpreted as corresponding to an invagination event followed by vesicle movement just after scission (Kaksonen et al., 2003; Kaksonen et al., 2005). The type I myosins remain at the cell surface during these events, and they are thought to be key factors in generating and moving actin filaments to produce invagination forces (Sun et al., 2006). Pan1 and the type I myosins are each actin NPFs for the Arp2/3 actin nucleation complex, although a recent study suggests that Pan1 has a low specific activity (Sun et al., 2006). Both of these actin NPFs also bind filamentous actin (Albanesi et al., 1985; Zot et al., 1992; Toshima et al., 2005), and together are predicted to act late during the endocytic process. Our data are consistent with a model in which Pan1 triggers positive feedback on the Myo5–Vrp1 complex to facilitate the formation of force-generating actin filaments that are properly organized and linked to the invaginating vesicle coat.

Pan1 is not the first, nor only, protein that interacts with the SH3 domains of the type I myosins. Many models of myosin's actin Arp2/3 activity depict the heterogeneity of ligand binding to the SH3 domain of the type I myosins as a means by which multiple copies of the Arp2/3 complex are assembled on the actin. The SH3 domain binding partner proteins typically have other interaction domains that could cluster additional actin activators; this may be important for the actin requirements in endocytosis. Pan1 is a good candidate for this role in vivo, because it binds to and activates the Arp2/3 complex in addition to binding Myo3/5. We do not presently know whether this could explain the mechanism by which the Pan1 PRD stimulates Myo5-Vrp1 activity, especially because in the in vitro actin polymerization assays only the PRD was present. It is possible that the GST–PRD proteins used for the in vitro studies dimerized via the GST moieties, thereby contributing to a local concentrating effect of actin binding and polymerizing functions.

Could there be a more specific role for a Pan1–myosin complex in invagination and scission? Actin NPF activity is unique to fungal type I myosins, as type I myosins from other species lack an A domain (Evangelista et al., 2000). In fungi, this actin activity of the myosins could supplant the requirement of a dynamin-like protein in neck elongation/vesicle scission, and simultaneously provide an actin component to the process as well. Consistent with this idea, it has been suggested recently that a role for dynamin in animal cells is to help create elongated tubules, whereas tension mediated by the actin cytoskeleton is required for the scission of these tubules (Itoh et al., 2005). Additionally, a recent report shows that Myosin IE interacts with dynamin and the 5′ inositol phosphatase synaptojanin, further suggesting a physical and potential regulatory link between type I myosins and scission machinery (Krendel et al., 2007).

In mammalian cells, vesicle scission is known to involve the GTPase dynamin. However, yeast has no obvious dynamin-like protein implicated in endocytosis. Scission defects in yeast have thus far been monitored through morphological analysis of electron micrographs to observe accumulated plasma membrane invaginations. We found this phenotype in cells carrying the pan1-20 allele, which has a frameshift mutation within the PRD, and it should not interact with the type I myosins in vivo. A similar phenotype has been reported for a type I myosin allele, myo5^E901A (Jonsdottir and Li, 2004), suggesting that both Pan1 and the type I myosins act at the same functional point in the endocytosis pathway.

Our SGA experiments provide further evidence consistent with an in vivo role for the Pan1 PRD in an actin-related late endocytic process. First, the two temperature-sensitive alleles we uncovered as having synthetic sick phenotypes, act1-121 and las17-14, are both intimately linked to actin polymerization functions. These are allele-specific interactions, because we recovered only one of 17 act1 alleles and one of three las17 alleles in the collection tested in our screen (Figure 3 and Supplemental Material). Phenotypes of the act1-121 allele are not well described in the literature (Wertman et al., 1992), but there are a couple of features of interest. First, mutations in this region of Act1 (E83A, K84A) generally have the most deleterious effects on endocytosis (Whitacre et al., 2001). Second, purified act1-121 protein does not efficiently polymerize into filaments in vitro (Miller and Reisler, 1995). Either hyperstable or slowly polymerizing F-actin filaments might be expected to alter the dynamics and regulation of the late stages of endocytosis. Third, act1-121 mutant diploid cells have the relatively unique phenotype compared with many other act1 alleles of hyperpolarized and elongated buds, a phenotype also associated with mutations in septins and the p21-activated kinase homologue Cla4 (Drubin et al., 1993; Richman et al., 1999; Versele and Thorner, 2004). Cla4 also regulates Myo3 and Myo5 activity (Wu et al., 1996). Therefore, through act1-121, our genetic findings for the pan1ΔPRD tie into the genetics of the activity of type I myosins.

The las17-14 temperature-sensitive allele (W41E, L133S) also exhibited synthetic sickness when combined with pan1ΔPRD. This LAS17 allele is in the collection of temperature-sensitive alleles used in the SGA analysis, but it has not been characterized in the literature. Comparisons of Las17 with WASP homologues suggest that one of the mutated residues, W41E, may correspond to the invariant tryptophan residue positioned from 23 to 25 residues away from the start of WH1/EVH1 domains; this is one of the aromatic residues implicated in the WH1/EVH1 domain binding to peptide ligands (Rong and Vihinen, 2000). WH1/EVH1 domains typically bind proline-rich peptides, and the N-terminal region of Las17 that contains the WH1 domain has been reported to bind to Vrp1/End5 (Naqvi et al., 1998). It will be interesting to know whether the las17-14 mutant protein is disrupted for Vrp1 binding or Arp2/3 activation; if so, this might explain the synthetic sickness we observe with our pan1ΔPRD allele.

We also confirmed a synthetic lethality between sla1Δ and pan1ΔPRD. Sla1 is a well-established Pan1 binding partner that is an endocytic adaptor and also is an inhibitor of Las17-dependent Arp2/3 stimulation (Tang et al., 2000; Howard et al., 2002; Rodal et al., 2003). It has been shown previously that the pan1-4 allele exhibits synthetic lethality when combined with sla1Δ (Tang et al., 2000); however, this allele corresponds to an extensive C-terminal truncation lacking hundreds of amino acids (Zeng et al., 2001), whereas our truncation is missing just the final 170 amino acids and preserves the F-actin and acidic Arp2/3 binding sites. Perhaps the combination of losing regulation of both Las17 and the type I myosins is particularly deleterious; for example, the Arp2/3 stimulatory activity of Las17 is required in cells lacking the Arp2/3 activity of the type I myosins (Evangelista et al., 2000; D'Agostino and Goode, 2005). Likewise, loss of Las17 regulation by deleting SLA1, along with altered Myo3 or Myo5 regulation by deleting the PRD of Pan1 could lead to irrevocably perturbed actin structures. Consistent with this idea, pan1-20 cells that also lack the key PRD region show exaggerated actin structures that seem to be either very thick cables or dramatically elongated cortical patches when grown at the nonpermissive temperature (Wendland et al., 1996). Alternatively, the pan1-20 and pan1ΔPRD proteins may have other activities affected in addition to their altered myosin SH3 ligand, such as putative regulated interactions between the N-termini and C-termini of Pan1 (Miliaras et al., 2004; Toshima et al., 2005).

Interestingly, the original screen that identified the pan1-20 allele also identified a mutation in SHE4 (Wendland et al., 1996). She4 is a chaperone for type I and type V myosins and is essential for myosin function under stress conditions (Toi et al., 2003; Wesche et al., 2003). This screen did not aim to identify mutants affecting one specific stage of endocytosis; however, the coincidence of the same screen identifying both the pan1-20 allele and a factor important for the type I myosins could suggest that these two mutants in fact act at the same substage of endocytic internalization.