This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Wicky, S. Articles by Singer-Krüger, B. Search for Related Content PubMed PubMed Citation Articles by Wicky, S. Articles by Singer-Krüger, B.

Molecular Interactions of Yeast Neo1p, an Essential Member of the Drs2 Family of Aminophospholipid Translocases, and Its Role in Membrane Trafficking within the Endomembrane System

Institute for Biochemistry, University of Stuttgart, Stuttgart,¹ Max Planck Institute for Developmental Biology, Tübingen, Germany²

Received 19 December 2003/ Returned for modification 2 February 2004/ Accepted 3 June 2004

	ABSTRACT

Top Abstract Introduction Materials and Methods Results Discussion References

 
Neo1p is an essential yeast member of the highly conserved Drs2 family of P-type ATPases with proposed aminophospholipid translocase activity. Here we present evidence that Neo1p localizes to endosomes and Golgi elements. In agreement with that finding, the temperature-sensitive neo1-37 and neo1-69 mutants exhibit defects in receptor-mediated endocytosis, vacuole biogenesis, and vacuolar protein sorting. Furthermore, neo1 mutants accumulate aberrantly shaped membranous structures most likely derived from vacuoles and the endosomal/Golgi system. At permissive temperatures, HA-Neo1-69p, like wild-type Neo1p, is stable and associates with endosomes. In contrast, HA-Neo1-37p is rapidly degraded and is predominantly retained within the endoplasmic reticulum (ER). Thus, the two neo1 alleles affect the stability and localization of the mutant polypeptides in different ways. A C-terminally truncated and a C-terminally epitope-tagged version of Neo1p are nonfunctional and also mislocalize to the ER. In agreement with a role within the endomembrane system, Neo1p exhibits genetic and physical interactions with Ysl2p, a potential guanine nucleotide exchange factor for Arl1p. Interestingly, deletion of ARL1 rescues the temperature sensitivity of neo1-37 and neo1-69. We demonstrate that Arl1p in its myristoylated and GTP-bound form is responsible for the inhibitory effect. Thus, Neo1p, Ysl2p, and Arl1p represent three proteins that collaborate in membrane trafficking within the endosomal/Golgi system.

	INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

 
Arf proteins are important regulators of vesicular membrane traffic. They help bind coat proteins to membranes and act on lipid-modifying enzymes (13). They are the founding members of an expanding family of homologous proteins including Arl, Arp, and the remotely related Sar proteins, whose functions are much less defined (29). Like other small GTPases, Arf proteins cycle between an active, GTP-bound form, which is membrane associated, and an inactive, GDP-bound form, which is soluble. Enzymes that regulate this cycle are GTPase-activating proteins (GAPs) and guanine nucleotide exchange factors (GEFs). The best characterized GEFs for Arf proteins share a conserved domain termed the Sec7 domain. In Saccharomyces cerevisiae, four members of this family have been shown to contain GEF activity toward Arf1p and Arf2p (for a review, see reference 25). However, little is currently known about the GEFs that activate other Arf family proteins. Members of a growing family of ß-propeller proteins, termed the HERC family, are new candidate GEFs for such Arf family members (37). Recently, Ysl2p/Mon2p, a protein with remote homology to the Sec7 family, was identified (26). Ysl2p is peripherally associated with endosomal membranes and is required for endocytosis and vacuole biogenesis. On the basis of the physical interaction between Ysl2p and Arl1p and the specific suppression of ysl2 defects by Arl1p, it was proposed that Ysl2p could represent a GEF for Arl1p (26).

In this study, we identified Neo1p as new binding partner of Ysl2p. This protein belongs to the Drs2 family of P-type ATPases. Its members are missing negatively charged amino acids in transmembrane domains four and six and contain hydrophobic residues in their place, arguing against their activity as cation transporters (6, 46). In fact, a substantial amount of work suggests that these P-type ATPases may act as aminophospholipid (APL) translocases (17, 32, 46, 50), although direct evidence for this activity remains to be provided with a purified enzyme. The finding that Drs2p exhibits strong genetic interactions with clathrin heavy chain and Arf1p (9) led to the proposal that the generation of lipid asymmetry could be intimately connected to the process of coat assembly and budding. In fact, recent work showing that a complex composed of Drs2p and the Sec7 family Arf GEF, Gea2p, is functionally implicated in membrane transformation events (7) provided further evidence for this idea.

Among the five Drs2 family members present in S. cerevisiae, Neo1p is the only essential one (34). In contrast, deletion of DNF1, DNF2, and DNF3, either individually or in combination, does not affect growth, and drs2 cells fail to grow only at 23°C or below (9, 22). However, the lethality of the quadruple drs2 dnf1 dnf2 dnf3 mutant indicated that DRS2 and the DNF genes constitute an essential subfamily with substantial functional overlap (22). Drs2p resides at the late Golgi complex, where it is implicated in the formation of a specific class of clathrin-coated vesicles (9, 16). While Dnf3p was also localized to the late Golgi complex (22, 32), Dnf1p and Dnf2p were shown to reside primarily at the plasma membrane. There, they were found to mediate an energy-dependent influx of nitrobenz-2-oxa-1,3-diazole-labeled phosphatidylethanolamine,-serine, and -choline (32). Furthermore, drs2 dnf1 dnf2 cells, but not dnf1 dnf2 cells, exhibited a defect in the internalization of endocytic markers (32), demonstrating that Drs2p contributes to the generation of lipid asymmetry at the plasma membrane.

In this work we characterized in great detail the subcellular localizations of wild-type Neo1p and several Neo1p variants and analyzed their role in membrane trafficking. We found that only an N-terminally tagged hemagglutinin (HA)-Neo1p complemented neo1 defects. This HA-Neo1p localized primarily to endosomes and Golgi elements. In contrast, C-terminally modified Neo1p was nonfunctional and exhibited an aberrant localization in reticular structures similar to those described in a recent study by Hua and Graham (23). We demonstrate that temperature-sensitive Neo1 proteins are unstable and accumulate massively in the endoplasmic reticulum (ER). Consistent with a role of wild-type Neo1p within the endomembrane system, temperature-sensitive neo1 mutants exhibited defects in endocytosis, vacuolar protein sorting, and vacuole biogenesis. Significantly, these defects were already evident at permissive conditions. We show that NEO1 is a suppressor of ysl2 defects, and we provide biochemical and genetic evidence for an interaction between Neo1p, Ysl2p, and Arl1p. These results suggest a role of Neo1p in membrane trafficking within the endosomal/Golgi system closely linked to that of Ysl2p and Arl1p.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

 
Strains, media, and plasmids. Unless otherwise indicated, strains (Table 1) were grown in complete medium (yeast extract-peptone-dextrose [YPD]) or synthetic growth medium (SD medium) to early-logarithmic phase at 25°C in a rotary shaker. DNA manipulations were carried out by standard techniques. The plasmids used are listed in Table 2.

neo1

A YSL2-TAP-specific PCR fragment was generated by amplification from plasmid pBS1539 with the oligonucleotides described below and was inserted into the genome downstream of, and in frame with, the YSL2 open reading frame by homologous recombination (36). Similarly, ARL1-3-HA-, NEO1-3-HA-, and YSL2-3-HA-specific fragments were generated by PCR using p3xHA-HIS5 as a template and were inserted at the chromosomal ARL1, NEO1, and YSL2 loci, respectively. Diploid transformants were purified. and correct integration was verified by PCR. After sporulation and tetrad dissection of five individual NEO1 NEO1-3-HA::HIS5 diploids, His⁺ haploid progeny expressing Neo1p-C-HA was never obtained, suggesting that this modified version is nonfunctional.

Oligonucleotides used for PCR were as follows: S1,NEO1 (GGTATCAGTGTGCACGGAAAACTTCAGAGACAAATGCCTAACCCTCCTTCGTACGCTGCAGGTCGAC) and S2,NEO1 (CCAAACCAATTGAATTATGGAGTGGCAAACTCTTGCACTTTTGCATAACTCGATGAATTCGAGCTCG), 5'TAP,YSL2(GATGGCCTCCAAGATAAAGTTTTGGAATTATCACTTGGATTTACGAAACTAGACTCCATGGAAAAGAGAAG) and 3'TAP,YSL2 (CACACAACTATTTCTATAAGCATATCATACATACTACAATCTTATGTATTACGACTCACTATAGGG), 5'HA,ARL1 (GGTATTACCGAAGGTTTAGATTGGTTGATTGATGTTATAAAAGAGGAACAGTTAGGAGCAGGGGCGGGTGC) and 3' HA, ARL1 (GAGAAACATGTATACACTTACTAACTCTATTATGTTTGGATAGAGCTCCTTGAGGTCGACGGTATCGATAAG), 5'HA,NEO1 (ATAGAAGGCTACATCCTCCAAGTTATGCAAAAGTGCAAGAGTTTGCCACTCCAGGAGCAGGGGCGGGTGCA) and 3'HA,NEO1 (TGTAACAAAATAATATTATGCATAATCTATATCTTCTTTGTAAAAATAAGGAGGTCGACGGTATCGATAAG), and 5'HA,YSL2 (GATGGCCTCCAAGATAAAGTTTTGGAATTATCACTTGGATTTACGAAACTAGACGGAGCAGGGGCGGGTGC) and 3'HA,YSL2 (TCATACATACTACAATCTTATGTATTTCTTTTCTTTTAGCTATGCATACCAATGGCAGAGGTCGACGGTATCGATAAG).

Generation of DNA constructs encoding N-terminally HA epitope-tagged Neo1p, Neo1p^D503N, Neo1p^Ctail, and Arl1p mutants. To obtain the triple HA epitope tag at the N terminus of Neo1p, a NotI site was generated immediately after the start codon of NEO1 by using recombinant PCR. Two overlapping PCR products were generated by using pRS425-NEO1 as a template. One was generated with the mutant primer 5'-GAAGGAGGGTTAGGGCGGCCGCTCATTTGTCTCTGAAG-3' and oligonucleotide A (5'-ACCTAACACTAGTTTGGGT-3'), complementary to sequence 5' of the SpeI site in the NEO1 promoter region. The other was generated with the mutant primer 5'-CTTCAGAGACAAATGAGCGGCCGCCCTAACCCTCCTTC-3' and oligonucleotide B (5'-TTCCAGTCAGTCTCCCCAT-3'), homologous to sequence 3' of the BstXI site in the NEO1 coding region. The two PCR products served as templates in a second PCR, in which the flanking primers A and B were used. The PCR product was digested with SpeI and BstXI, and the fragment was subcloned into pBSK. Subsequently, the 111-bp NotI fragment encoding the triple HA epitope tag was introduced at the NotI site. After the presence of the correct nucleotide sequence was confirmed, the SpeI/NcoI 3-HA-NEO1 fragment was subcloned into pRS425-NEO1 to generate pRS425-HA-NEO1. The SpeI/SalI fragment encoding 3-HA-Neo1p was subcloned into the SpeI/SalI restriction sites of pRS315 to generate pRS315-HA-NEO1.

Plasmid pRS315-HA-NEO1 was transformed into BS811. After sporulation and tetrad dissection, neo1 haploids containing the plasmid-borne HA-Neo1p grew similarly to the wild type. They were propagated under nonselective conditions (YPD).

DNA encoding the Neo1p^D503N point mutation was generated by directed mutagenesis using a PCR-based protocol (4) and the mutant primer 5'-CCCTGTTTTGTTGCTTAGAAG-3'. As flanking primers, 5'-CTGACTGGAAACTACGTG-3' (5'NEO1,Pac1) and 5'-TACGACCGATAACCAAAGTAC-3' (3'NEO1,SnaB1) were used with pRS425-NEO1 as the template. The PCR product was digested with SnaBI and PacI and then subcloned into pRS425-NEO1 opened with SnaBI/PacI.

To generate NEO1^Ctail, a stop codon was introduced by using the GeneEditor mutagenesis kit (Promega). The HindIII/SalI NEO1 fragment, subcloned into pBSK, was mutagenized according to the manufacturer's instructions by using the mutagenic oligonucleotide 5'-CTATAAATTGCCTAGGCCGTCCAGACAG-3'. This created a stop codon at positions 3391 to 3393 of the NEO1 coding sequence and a StyI restriction site. The SalI/StuI subfragment of the mutated HindIII/SalI fragment was excised from pBSK and used to replace the wild-type SalI/StuI fragment in pRS315-HA-NEO1.

DNAs encoding the Arl1G2A, Arl1T32N, and Arl1Q72L point mutants were generated by directed mutagenesis as described previously (26), except that the 5'-flanking primer was replaced by 5'-ACCAGGATCCTTGTTAAAGAGTAAGCC-3'. The mutant primer used to generate Arl1G2A was 5'-GAACTAAAAATGTTAGCCATCTTGATCTATAACTCAC-3'. The PCR products were digested with BamHI and XhoI and subcloned into pRS316.

All PCR-amplified regions were sequenced to verify the mutations and to exclude the presence of PCR errors.

Coimmunoprecipitation using TAP-tagged Ysl2p and HA-tagged Neo1p. Coimmunoprecipitation was basically carried out according to the method of Wicky et al. (49) with minor modifications. BS912 and BS862 cell lysates were extracted with either 0, 0.01, 0.02, or 0.04% NP-40 (final concentration) for 20 min at 0°C and were centrifuged at 16,000 x g for 20 min. The supernatants were affinity purified with immunoglobulin G (IgG)-Sepharose, and after extensive washing, bound proteins were released by the TEV protease. Alternatively, after extraction with 0.01% NP-40, the low-speed supernatant was centrifuged at 100,000 x g for 1 h at 4°C in a TLA 120.2 rotor (Beckman Instruments). The 100,000 x g supernatant was subsequently immunoprecipitated with IgG-Sepharose as described above. Proteins released by the TEV protease were either directly analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) or subjected to a second immunoprecipitation via the residual calmodulin binding peptide of the tandem affinity purification (TAP) epitope (36). Ysl2p and HA-Neo1p were detected by immunoblotting using rabbit anti-Ysl2p (26) and mouse anti-HA (-HA; 16B12; Covance) as primary antibodies.

Immunofluorescence microscopy. Indirect immunofluorescence microscopy was performed as described previously (44). Single stainings were performed with the mouse monoclonal -HA antibody (16B12; Covance) at 1:1,000 (HA-Neo1p, HA-Ysl2p) or 1:8,000 (HA-Arl1p) and a Cy³-conjugated goat anti-mouse Fab fragment (Jackson ImmunoResearch) at 1:1,000. For double stainings, dilutions for the primary antibodies were as follows: 1:1,000 for mouse monoclonal -HA (16B12; Covance), 1:100 for rat monoclonal -HA (clone 3F10; Roche), 1:100 for mouse monoclonal -Pep12p (Molecular Probes), 1:2,000 for mouse monoclonal -60-kDa v-ATPase (Molecular Probes), 1:5 for mouse monoclonal -Ypt1p ascites fluid (a gift from D. Gallwitz), 1:100 for affinity-purified rabbit -Ypt51p (44), 1:300 for rabbit -c-Myc (sc-789; Santa Cruz Biotechnology), and 1:1,000 for affinity-purified mouse -GFP (clone 3E6; QBIOgene). All secondary antibodies were affinity purified and used at a dilution of 1:1,000. These included Alexa⁴⁸⁸- and Alexa⁵⁹⁴-conjugated goat anti-mouse IgG (1:250 for Ypt1p), Alexa⁴⁸⁸- and Alexa⁵⁹⁴-conjugated goat anti-rat IgG, and Alexa⁵⁹⁴- and Alexa⁴⁸⁸-conjugated goat anti-rabbit IgG (Molecular Probes). Green fluorescent protein (GFP)-Sec63p was directly observed with appropriate filter sets.

Endocytosis of LY and Ste2p, and Rer1p localization. Lucifer Yellow CH (LY) internalization experiments with cells grown at 25°C were performed as described previously (43).

Pheromone-induced endocytosis of Ste2p was carried out with MATa cells as described previously (26). Extracts of BS188 cells (MAT), treated for 10 min with cycloheximide, served as a control for the specificity of the -Ste2p antiserum.

The localization of GFP-Rer1p in BS64, BS915, and BS917 transformants carrying pSKY5RER1-0 was analyzed. After growth at 25°C, cells were shifted to 30°C for 1 h or to 37°C for 2 h, collected by centrifugation, and viewed on concanavalin A-coated coverslips with a fluorescence microscope by using the appropriate filter for GFP.

Generation of temperature-sensitive neo1 alleles by PCR mutagenesis. PCR mutagenesis was performed as described elsewhere (26) by using the 2.9-kb HindIII NEO1 fragment as a template and oligonucleotides 5'NEO,NcoI (5'-GTAACCATGGCAAAGGAAGC-3') and 3'NEO,ts-mut (5'-CCAAATCTTGATACCTGC-3'), annealing upstream and downstream of the internal NcoI and SnaBI restriction sites of NEO1, respectively. In the reaction, 4.8 mM MgCl₂, 0.2 to 0.4 mM MnCl₂, 0.2 mM dATP, 1 mM dCTP, 1 mM dGTP, and 1 mM dTTP were used. The purified PCR products were cotransformed with PacI/SnaBI-digested pRS315-NEO1 into neo1 cells carrying pRS316-NEO1 (BS845). After growth at 25°C on SD plates lacking uracil and leucine, approximately 4,600 transformants were replica plated onto SD-Leu plates containing 5-fluoroorotic acid and screened for growth at 25 and 37°C. Mutant plasmids were rescued from colonies that grew at 25°C, but not at 37°C, and were retransformed into BS845 to confirm the plasmid dependence of temperature-sensitive growth after loss of pRS316-NEO1. The mutations found in neo1-37 result in the following amino acid exchanges: F301S, D305G, V373A, S539P, L540P, N631S, Q655R, and V748A; those in neo1-69 result in I284T, I449M, E605G, L680P, and Y714H. The conservation of these amino acids was determined by sequence alignment of Neo1p orthologs found in Saccharomyces paradoxus, Saccharomyces mikatae, Saccharomyces bayanus, Schizosaccharomyces pombe, and Candida albicans (unpublished data). The sequence encoding the triple HA epitope was introduced into the neo1-37 and neo1-69 alleles by replacing the SpeI/NcoI fragment with that of pRS315-HA-NEO1.

Pulse-chase labeling and immunoprecipitation. Pulse-chase labeling and immunoprecipitation of carboxypeptidase Y (CPY) were performed as described previously (45). Pulse labeling to monitor HA-Neo1p was carried out at 25°C for 30 min. The chase was performed at either 25 or 37°C for 30, 60, 120, and 180 min. Cells collected at the different time points were lysed in 100 µl of ice-cold buffer (50 mM Tris-HCl [pH 7.5], 1 mM EDTA, 1x chymostatin-leupeptin-antipain-pepstatin cocktail [CLAP]) by using glass beads and were solubilized in the presence of 1% SDS and 150 mM NaCl. After a 10-fold dilution with immunoprecipitation buffer (1% deoxycholic acid, 1% Triton X-100, 50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1x CLAP) immunoprecipitations were performed with mouse -HA (12CA5; Roche) as described above.

Invertase secretion. Invertase secretion was analyzed with cells grown at 25°C. Invertase was induced at 30°C as described by Jochum et al. (26), and samples were removed after 0, 45, 60, and 75 min of incubation in YPD (0.1% glucose). The glycosylation state of invertase was determined after a 2-h induction by staining for invertase activity after native gel electrophoresis basically as described previously (3). The cell lysis and electrophoresis buffers were replaced by 100 mM Tris-HCl (pH 6.8)-10% glycerol-1 mM phenylmethylsulfonyl fluoride and 25 mM Tris base-200 mM glycine, respectively. sec18 cells, shifted to 37°C during the induction period, served as a control for the production of core-glycosylated invertase.

Transmission electron microscopy. Yeast cells were cryoimmobilized by high-pressure freezing according to the method of Hohenberg et al. (20) and were prepared for ultrastructural analysis as described previously (26). Ultrathin sections stained with aqueous uranyl acetate and lead citrate were viewed under a Philips CM10 electron microscope at 60 kV.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

 
NEO1, an essential member of the Drs2 family of P-type ATPases, is a suppressor of ysl2 defects. To identify proteins that function together with Ysl2p or bypass the requirement for Ysl2p, a screen for high-copy-number suppressors was performed in ysl2 cells, which are temperature sensitive at 37°C (26). This led to the isolation of a clone that contained a fragment of chromosome IX including the NEO1 gene. Further subcloning analysis revealed that NEO1 retained the suppressing activity (Fig. 1A). Based on the available sequence information, Neo1p is most homologous to the ATPase C6C3 from Schizosaccharomyces pombe and to ATPases IIa and IIb from humans and mice. The other four S. cerevisiae Drs2 family members are more divergent from Neo1p. Consistent with that, overexpression of Drs2p, the closest homolog of Neo1p (30% identity; 48%homology), did not suppress the growth defect of the ysl2 strain (Fig. 1A).

View larger version (86K): [in this window] [in a new window]   FIG. 1. NEO1 is a suppressor of ysl2 defects. (A) Growth of wild-type (wt) (RH1201) and ysl2 (BS747) cells, transformed with the indicated plasmids, on YPD at 37°C. (B) Pheromone-stimulated internalization of Ste2p was analyzed in cells (BS64, BS694, and BS694 transformants) grown at 25°C. After a 10-min preincubation at 30°C in the presence of 10 µg of cycloheximide/ml, -factor was added, and at the indicated times, samples were withdrawn and processed for immunoblotting with an anti-Ste2p antiserum. An extract of cells lacking Ste2p revealed a band (indicated by stars) nonspecifically labeled by anti-Ste2p. The positions of the 45- to 47-kDa and lower-mobility (hp+u) forms of Ste2p are indicated. (C) LY internalization in wt cells (RH1201), ysl2 cells (BS747), and ysl2 cells transformed with either pRS315-NEO1 or pRS425-NEO1 was analyzed. After a 1-h incubation with LY at 30°C, the cells were washed, mounted in low-melting-point agarose, and visualized by fluorescence (left panels) and Nomarski optics (right panels). Bar, 10 µm.

ysl2

ysl2

ysl2

Ctail

neo1

ysl2

Due to the strong delay in the pheromone-induced degradation of the -factor receptor Ste2p exhibited by ysl2 cells (26) (Fig. 1B), we analyzed whether expression of Neo1p from a CEN- or a 2 µm-based vector affected recovery from this endocytic transport defect. In the wild type, before treatment with -factor, Ste2p migrates principally as a band of 45 to 47 kDa. Soon after the addition of pheromone, a series of discrete high-molecular-mass bands are observed due to the hyperphosphorylation and ubiquitination (hp+u) of Ste2p. Ste2p is then completely degraded within 30 to 60 min (Fig. 1B). In ysl2 cells, overexpression of Neo1p from a 2 µm plasmid resulted in a clear recovery from the endocytic transport block (Fig. 1B). A similar rescue was observed upon expression of NEO1 from a single-copy vector. Under these conditions, however, the hp+u form of Ste2p persisted longer than in the wild type (Fig. 1B).

The vacuole biogenesis defect exhibited by ysl2 cells (26) was also suppressed by NEO1. While ysl2 cells exhibited highly fragmented vacuoles that were weakly stained by the fluid-phase marker LY, after 1 h of incubation at 30°C the major fraction of ysl2 transformants carrying NEO1 on either CEN- or 2 µm-based plasmids contained vacuoles of wild-type structure that were strongly stained by LY (Fig. 1C).

Neo1p interacts with Ysl2p in vivo. To further establish a relationship between Ysl2p and Neo1p, we checked for an in vivo interaction between the two proteins by coimmunoprecipitation. Functional versions of Ysl2p modified C-terminally with the TAP epitope tag and of Neo1p modified N-terminally with three copies of the influenza virus HA epitope were coexpressed (see Materials and Methods). TAP-Ysl2p assemblies were purified from total cellular extracts by using IgG-Sepharose followed by a mild elution from the affinity matrix with the site-specific TEV protease (36). The released protein assemblies were separated by SDS-PAGE and analyzed by immunoblotting. As shown in Fig. 2A, HA-Neo1p copurified specifically with TAP-Ysl2p. The coimmunoprecipitation of Neo1p was most efficient in the presence of 0.01% NP-40 during the isolation. This detergent concentration resulted in approximately 25 and 20% solubilization of Ysl2p and Neo1p, respectively, as determined by a 100,000 x g centrifugation. Precipitation of TAP-Ysl2p present in the 100,000 x g supernatant after treatment with 0.01% NP-40 still allowed the coisolation of HA-Neo1p (Fig. 2B, lane 6) but not that of Ypt51p or Pep12p (data not shown), indicating that the interaction was not mediated through membranes including Ysl2p and Neo1p. Finally, the interaction between TAP-Ysl2p and HA-Neo1p was nicely preserved when the two-step TAP (36) was performed (Fig. 2B, lane 4), supporting the high specificity of this interaction.

View larger version (14K): [in this window] [in a new window]   FIG. 2. Neo1p interacts with Ysl2p in vivo. (A) Coimmunoprecipitations were performed with cell extracts expressing the indicated epitope-tagged proteins in the absence (lanes 1 and 5) or presence (lanes 2 to 4 and 6 to 8) of NP-40 (concentrations as indicated). Proteins bound to IgG-Sepharose were released by the TEV protease and detected as described in Materials and Methods. (B) Coimmunoprecipitations were performed in the presence of 0.01% NP-40 by using either IgG-Sepharose alone (lanes 1 and 2) or a stepwise combination of IgG-Sepharose and calmodulin resin (TAP protocol) (lanes 3 and 4). Results of coimmunoprecipitation with a 100,000 x g supernatant after extraction with 0.01% NP-40 are shown in lanes 5 and 6.

View larger version (92K): [in this window] [in a new window]   FIG. 3. Neo1p localizes to punctate structures that are sensitive to a mutation in VPS27 and that colocalize with structures containing Pep12p, Ypt51p, and Tlg1p. (A, B, and C) Wild-type cells lacking HA-Neo1p (BS64) (A) or expressing HA-Neo1p (BS862) (B), or vps27 cells expressing HA-Neo1p (BS1088) (C), were grown at 25°C, fixed, and stained by indirect immunofluorescence with a monoclonal mouse anti-HA antibody as described in Materials and Methods. (D to I and L) Double immunofluorescence with BS862 cells and BS862 transformants expressing Myc-Tlg1p (F), Ypt51Q66L (G), or GFP-Rer1p (I). Cells were fixed and stained first with a rat (D, H, I, and L) or mouse (E, F, and G) monoclonal anti-HA antibody to detect HA-Neo1p and then with either -Pep12p (D), -Ypt51p (E and G), -Myc (F), -Ypt1p (H), -AFP (I), or -60-kDa v-ATPase (L). Secondary antibodies to the various IgGs are described in Materials and Methods. (K) Double immunofluorescence, as for panels D to J and L, with sec7 cells expressing 3-HA-Neo1p (BS1382) after a 1-h shift to 37°C in YPD medium containing 0.1% glucose. Bar, 5 µm.

As shown in Fig. 3D to G, the staining patterns of HA-Neo1p and all three endosomal proteins were very similar. Counting of the Neo1p-positive spots showed that at least 35% coincided with Ypt51p (n = 559), 48% coincided with Pep12p (n = 1,109), and 54% coincided with Myc-Tlg1p (n = 611), suggesting that Neo1p is spread throughout the endosomal system. As with vps27 cells, expression of Ypt51Q66L, a point mutant with a higher affinity for GTP, which causes the formation of enlarged endosomal structures (44), resulted in the generation of larger and less numerous Neo1p-positive structures. Under these conditions, the extent of colocalization of Neo1p and Ypt51p was greatly increased (Fig. 3G) (at least 69% of Neo1p spots coincided with Ypt51p [n = 77]). Together, these results clearly demonstrate that Neo1p resides on endosomal structures.

In spite of the considerable overlap between Neo1p and endosomal structures, some spots did not coincide. This may reflect the mechanism of protein sorting within the highly dynamic tubulovesicular endosomal system (30). Alternatively, Neo1p may partially localize to another subcellular compartment, most likely the Golgi complex. To address this possibility, the staining pattern of HA-Neo1p was compared to those of two Golgi proteins, the small GTPase Ypt1p and Rer1p, a retrieval receptor for ER membrane proteins (41). Although in wild-type cells the majority of Neo1p-positive spots did not coincide with those containing Ypt1p (Fig. 3H) (10% colocalization [n = 990]) or GFP-Rer1p (Fig. 3I) (8% [n = 538]), in temperature-sensitive sec7 cells incubated at 37°C the Neo1p staining pattern was affected like that of Ypt1p (Fig. 3K). As has been observed for other Golgi-localized proteins (42), the Neo1p structures collapsed into large clumps.

Finally, double labeling of Neo1p and the 60-kDa subunit of the vacuolar v-ATPase revealed that many of the Neo1p-positive spots were located in proximity to, but not within, the vacuolar membrane (Fig. 3L), consistent with the localization of Ypt51p-positive endosomes in the vicinity of the vacuole at the ultrastructural level (28).

Taken together, our results provide strong evidence that a major fraction of Neo1p is localized to early and late endocytic structures. Most likely, another but smaller portion is associated with the late Golgi complex.

Localization, stability, and identity of temperature-sensitive Neo1p mutants. To further investigate the function of Neo1p, temperature-sensitive neo1 alleles were generated by low-fidelity PCR and expressed in the neo1 strain after plasmid shuffling. This led to the isolation of mutants (among which neo1-37 and neo1-69 were studied in more detail) that grew like the wild type at 25 and 30°C but were temperature sensitive at 37°C (Fig. 4A). Similar to what has been described recently (23), the temperature sensitivity of neo1-37 and neo1-69 was suppressed by the presence of 1.5 M sorbitol in the growth medium (data not shown).

View larger version (58K): [in this window] [in a new window]   FIG. 4. Characterization of temperature-sensitive neo1 mutants: localization and stability of mutant polypeptides. (A) Dilution series of wild-type (wt) (BS64), neo1-37 (BS915), and neo1-69 (BS917) cells after growth at 37 and 25°C. (B) Cells expressing HA-Neo1-37p or HA-Neo1-69p (BS1001 and BS993, respectively) were grown at 25°C to early-log phase and either directly fixed and processed for indirect immunofluorescence as described in the legend to Fig. 3A through C or shifted to 37°C for the indicated times, fixed, and stained. (C) Double staining in cells expressing HA-Neo1-37p/GFP-Sec63p or HA-Neo1-69p/GFP-Sec63p (BS1285 and BS1286, respectively) was performed after growth at 25°C with a mouse monoclonal anti-HA antibody followed by Alexa594-conjugated anti-mouse IgG. The GFP-Sec63p signal was directly observed. (D) Staining of HA-Neo1-69p in vps27 (BS1383) cells was performed as described in the legend to Fig. 3A through C. (E) Cells expressing HA-Neo1p, HA-Neo1-37p, or HA-Neo1-69p (BS862, BS1001, or BS993, respectively) were labeled for 30 min at 25°C with [35S]methionine and chased for 0, 30, 60, 120, or 180 min at 25 or 37°C. At each time point, wt and mutant HA-Neo1p were immunoprecipitated as described in Materials and Methods. HA-Neo1p levels were quantified by phosphorimaging and expressed as percentages of the level of induced protein at the 0-min chase point. (F) Simplistic model of Neo1p. Localization of the putative membrane-spanning domains is based on the PredictProtein program (39). Positions of amino acid exchanges in the mutant Neo1-69p (hexagons) and Neo1-37p (stars) proteins are indicated (for details, see Materials and Methods).

In contrast, at 25°C the Neo1-69p staining pattern was punctate and similar to that of wild-type Neo1p. Since it was still sensitive to the vps27 mutation (Fig. 4D), Neo1-69p appears to localize to endosomes at permissive temperatures. After a 1-h shift to 37°C, Neo1-69p was found in brightly stained spots of increased size (Fig. 4B), and after 3 h it was detected in tubular structures similar to those labeled by Neo1-37p (Fig. 4B). However, a weak residual punctate staining of Neo1-69p was still observed.

Since the tubular staining pattern was reminiscent of the appearance of the ER, indirect immunofluorescence was conducted in mutant strains transformed with GFP-Sec63p, a component of the ER translocon. Since the autofluorescence of GFP was still preserved after fixation of cells grown at 25°C, labeling to detect the mutated Neo1p versions was performed only with the anti-HA antibody. While Neo1-37p clearly colocalized with GFP-Sec63p, Neo1-69p did not (Fig. 4C). Thus, at 25°C the majority of Neo1-37p resides within the ER, while Neo1-69p is still primarily associated with vps27-sensitive endosomes (Fig. 4D). Subcellular fractionation of a cell extract using sucrose density gradient centrifugation allowed us to exclude the idea of an association of some peripherally located Neo1-37p with regions of the plasma membrane and confirmed the localization of Neo1-37p to the ER (data not shown). Shifting the cells to 37°C resulted in the loss of the auto- and immunofluorescence signal of GFP-Sec63p, thereby preventing the simultaneous detection of Neo1p mutants and GFP-Sec63p. However, Neo1-37p most likely remained in the ER, where it seemed to aggregate. A substantial fraction of Neo1-69p also appeared to accumulate in the ER after 3 h of incubation at 37°C.

We also analyzed the fate of wild-type and mutant Neo1 proteins by pulse-chase labeling and subsequent immunoprecipitation. At 25°C, both wild-type Neo1p and Neo1-69p were long-lived proteins with estimated half-lives (t_1/2) of 257 and 226 min, respectively (Fig. 4E). In contrast, Neo1-37p was highly unstable, with a t_1/2 of approximately 36 min (Fig. 4E). When the chase was performed at 37°C, the stabilities of Neo1p and Neo1-69p were still comparable and were equally reduced to t_1/2 of 90 and 100 min, respectively. The stability of Neo1-37p was further decreased (t_1/2, 26 min) (Fig. 4E). In summary, under permissive conditions the mutations in neo1-37 cause the protein to be retained within the ER, where it is rapidly degraded. In contrast, Neo1-69p reveals a stability and localization similar to those of wild-type Neo1p.

Sequencing of the neo1 alleles revealed eight amino acid exchanges in Neo1-37p, of which three are present in the cytoplasmic loop between transmembrane helices 2 and 3 and five are in the large cytoplasmic loop between transmembrane helices 4 and 5 (Fig. 4F) (see Materials and Methods). Remarkably, the phenylalanine at position 301, conserved within APL family members, was changed to serine. In Neo1-69p, five amino acid changes were identified, one in transmembrane helix 4, one in the small cytoplasmic loop, and three in the large cytoplasmic loop (Fig. 4F). The L680P mutation could affect the helix structure within the long helix identified in the crystal structure of the Ca²⁺-ATPase (47). Aside from that, the mutations are not located in the known ATP-binding sites present in all P-type ATPases or in more specific sequence elements of the S. cerevisiae APL subgroup (6). A comparison of the Neo1p orthologs from S. cerevisiae, S. mikatae, S. bayanus, S. paradoxus, Schizosaccharomyces pombe, and C. albicans revealed that 7 of the 13 amino acid exchanges were at positions that are completely conserved among these yeast species. While the Neo1p orthologs of the four Saccharomyces species were identical at all positions mutated in S. cerevisiae Neo1p, both the Schizosaccharomyces pombe and the C. albicans Neo1p orthologs revealed differences from S. cerevisiae Neo1p at S539, N631, and Q655 (unpublished data).

Two C-terminal modifications of Neo1p result in loss of Neo1p function and cause mislocalization to the ER. In a recent study, the C-terminally 13-Myc epitope-tagged Neo1p was localized to the ER and to punctate structures, which partially overlapped with the cis and medial Golgi markers Och1p and Mnn1p (23). We also generated a C-terminally 3-HA epitope-tagged Neo1p (Neo1p-C-HA) in diploid cells by replacing one wild-type copy of the NEO1 gene with the construct. In sharp contrast to the functional N-terminally tagged HA-Neo1p, upon sporulation and tetrad dissection of five independent heterozygous diploids expressing Neo1p-C-HA, no haploid cells containing the NEO1-C-HA::HIS5 copy were obtained (Fig. 5A), implying that this form of Neo1p is nonfunctional. Indirect immunofluorescence of Neo1p-C-HA in heterozygous diploids containing one wild-type NEO1 copy revealed localization within continuous reticular and faint punctate structures (Fig. 5B), similar to that shown by Hua and Graham (23). Qualitatively, the staining differed from the more punctate pattern of N-terminal HA-Neo1p (Fig. 3 and 5C). Moreover, Neo1p-C-HA produced a much weaker signal overall than HA-Neo1p, suggesting that the C-terminally modified form, like other Neo1p mutants, may be unstable (see Fig. 4 and below).

View larger version (126K): [in this window] [in a new window]   FIG. 5. Two C-terminally modified versions of Neo1p are nonfunctional and mislocalize to the ER. (A) The heterozygous NEO1/NEO1::3-HA-HIS5 strain (BS757) was subjected to sporulation, tetrad dissection, and growth at 25°C. Shown are five tetrads (numbered 1 to 5; each in rows A to D) with a 2:0 segregation of the His–/His+ phenotype. (B to D) Staining by indirect immunofluorescence of Neo1p-C-HA (BS757) (B), HA-Neo1p (BS64 carrying pRS315-HA-NEO1) (C), or HA-Neo1pCtail (BS64 carrying pRS315-HA-NEO1Ctail) (D). Note that the exposure time for panel B was threefold longer than that for panels C and D. Bar, 5 µm.

Ctail

Ctail

neo1

Ctail

Ctail

ysl2

Ctail

The temperature-sensitive neo1-37 and neo1-69 mutants are defective in endocytosis, vacuolar protein sorting, and vacuole biogenesis. The localization of Neo1p within the endosomal system prompted us to investigate whether endocytosis was affected in the neo1 mutants. The effects on fluid-phase endocytosis and vacuole morphology were studied by endocytic internalization of LY under permissive conditions. In contrast to the wild type, both mutants exhibited fragmented vacuoles, which were only weakly labeled with LY (Fig. 6A), indicating reduced endocytic transport toward the vacuole.

View larger version (75K): [in this window] [in a new window]   FIG. 6. The temperature-sensitive neo1-37 and neo1-69 mutants reveal defects in vacuole biogenesis, endocytosis, and vacuolar protein sorting but not in general secretion. (A) LY internalization assays were performed with wild-type (wt), neo1-37, and neo1-69 cells (strains BS64, BS915, and BS917, respectively) grown at 25°C as described in the legend to Fig. 1C. (Left panels) Fluorescence; (right panels) Nomarski optics. Bar, 5 µm. (B) Pheromone-stimulated Ste2p turnover was analyzed in wt, neo1-37, and neo1-69 cells (BS64, BS915, and BS917, respectively) as describedin the legend to Fig. 1B. (C) CPY processing and sorting were analyzed in cells (see Fig. 6A) pulse labeled with [35S]methionine for 5 min at 30°C (lanes 1 and 5). Methionine was then added to initiate the chase for 5, 10, or 25 min (lanes 2 to 4 and 6 to 8, respectively). At each time point, aliquots were removed and converted to spheroplasts after the medium was separated from the cells. The spheroplast (lanes 1 to 4) and medium (lanes 5 to 8) fractions were processed for immunoprecipitation with anti-CPY antibodies and were analyzed by SDS-PAGE. p1, ER form; p2, Golgi form; m, mature form. (D) Invertase secretion was analyzed with cells (see panel A) grown in YPD medium (5% glucose) at 25°C. Invertase derepression was induced in YPD medium (0.1% glucose) at 30°C for 30 min. Samples were analyzed for external and total invertase activity after 0, 45, 60, and 75 min of invertase derepression. Values are means from three (neo1-37) or five (wt and neo1-69) different experiments. (E) Invertase glycosylation was analyzed for wt, neo1-37, neo1-69, and sec18 cells (see panel A and RH1737 in panel E, lane sec18) after derepression in YPD medium (0.1% glucose) at 30°C (wt, neo1-37, and neo1-69) or 37°C (sec18) for 2 h. Cell lysates were separated by native PAGE and stained for invertase activity as described in Materials and Methods. (F) Cells (see panel A) transformed with the GFP-Rer1p plasmid were grown at 25°C, shifted to 30°C for 1 h or to 37°C for 2 h, and observed with a fluorescence microscope.

Due to the intersection of endocytic and vacuolar sorting pathways and the fact that the sorting of vacuolar hydrolases is also impaired in many endocytic mutants, the transport and processing of CPY were analyzed. After a 5-min pulse with [³⁵S]methionine, wild-type cells display two precursor forms of ³⁵S-labeled CPY, representing an ER-modified (p1) and a Golgi-modified (p2) form. Upon delivery to the vacuole, p2 CPY is proteolytically processed, resulting in an active, lower-molecular-weight, mature form (Fig. 6C). At the permissive temperature, CPY transport and processing appeared to be somewhat delayed in both temperature-sensitive neo1 mutants. In the neo1-69 strain, maturation of the p1 to the p2 form was delayed by approximately 5 min. This delay appeared slightly prolonged in neo1-37 cells. Furthermore, the conversion to the mature form was significantly delayed in neo1-37/neo1-69 cells. In comparison to the wild type, in which processing was complete after 25 min of chase, only 52 and 45% of total CPY were found in the mature form in neo1-37 and neo1-69 cells, respectively, at that time point. Similar to what has been determined for arl1 and ysl2 cells (26), 8 and 15% of total CPY were secreted in the p2 form in neo1-37 and neo1-69 cells, respectively. These data are consistent with a role of Neo1p in protein trafficking within the late Golgi/endosomal system.

We also monitored the trafficking of the well-characterized enzyme invertase. After derepression of invertase by incubation in 0.1% glucose at 30°C, neo1-37 and neo1-69 cells exhibited a ratio of external to total invertase very similar to that in wild-type cells (Fig. 6D). Unlike sec18 (N-ethyl maleimide-sensitive factor allele) cells, which accumulated highly underglycosylated invertase due to a transport block between the ER and the Golgi complex, none of the neo1 mutants showed impaired glycosylation of invertase (Fig. 6E). This was in contrast to results obtained after a 1-h derepression at 37°C, which revealed fractions of incompletely glycosylated invertase in both neo1 mutants (data not shown).

To test if Neo1p is involved in anterograde or retrograde transport between the ER and the Golgi complex as previously suggested (23), the localization of the Golgi receptor protein Rer1p was analyzed in neo1-37 and neo1-69 cells under permissive conditions. Mislocalization of GFP-Rer1p to the vacuole is indicative of defective retrograde transport between the ER and the Golgi complex (COPI), while retention within the ER is caused by defective anterograde transport (41). As shown in Fig. 6F, GFP-Rer1p was properly localized to punctate structures typical for the Golgi complex in both neo1 mutants. However, shifting of the cells to 37°C for 2 h resulted in mislocalization of GFP-Rer1p to the fragmented vacuole compartment (Fig. 6F) (23).

In summary, our results suggest that Neo1p activity is required for endocytosis and vacuole biogenesis, but not for normal protein secretion or for retrograde transport between the ER and the Golgi complex. Nevertheless, aggregation of mutant Neo1 proteins within the ER under nonpermissive conditions appears to indirectly impair membrane trafficking along the secretory pathway.

Accumulation of aberrant tubular membrane protrusions in temperature-sensitive neo1 mutants. To study the neo1 mutants at the ultrastructural level, cells were quick-frozen after growth at 25°C or after a shift to 37°C and were subsequently analyzed under the electron microscope. Unlike wild-type cells, which usually contained a few (one to three) large vacuoles per central section (Fig. 7A), both neo1-69 (Fig. 7B and D) and neo1-37 (Fig. 7C) mutants were characterized by fragmentation and branching of the vacuolar compartment. The neo1 mutants accumulated extremely aberrantly shaped, flattened membrane elongations, which often extended from larger compartments with a greatly increased electron density (Fig. 7D, E, F, G, and K). These divergent structures exhibited an increased surface-to-volume ratio, reflected also in the higher matrix density of the vacuoles and endosomes, and frequently formed flat processes that ended in sharp edges. In the wild type, such tapered membrane structures were never observed. We also noticed the frequent accumulation of other types of membrane structures, such as multivesicular and ringlike tubular structures (Fig. 7E, F, H, and K). These prominent alterations in membrane-bound structures were observed both at 25 and at 37°C but were more frequently found under nonpermissive conditions. This phenotype hints at an important role of Neo1p in membrane trafficking within the endomembrane/late Golgi system.

View larger version (164K): [in this window] [in a new window]   FIG. 7. Electron microscopy of neo1 mutants. Wild-type (BS862) (A), neo1-37 (BS1001) (C), or neo1-69 (BS993) (B and D to K) cells were grown in YPD medium at 25°C to early-logarithmic phase and were either directly cryoimmobilized and processed for electron microscopy (A to D, G, I, and K) or shifted to 37°C for 3 h (E, F, and H), quick-frozen, and processed for electron microscopic analysis as described in Materials and Methods. Thin sections were viewed with an electron microscope. N, nucleus; M, multivesicular structure; arrowhead, sharp-edged membrane structure; asterisk, elongated process, originating from vacuole. Bar for panels A to G, 1 µm; bar for panels H to K, 0.5 µm.

ysl2 arl1

ysl2 arl1

ysl2 arl1

ysl2

ysl2

View larger version (82K): [in this window] [in a new window]   FIG. 8. Genetic interactions between NEO1 and ARL1. (A) A heterozygous ysl2::kanr arl1::URA3 strain (generated by mating BS694 with BS1105) transformed with pRS425-NEO1 was subjected to sporulation, tetrad dissection, and growth at 25°C. Shown are five tetrads (numbered 1 to 5, each in rows A to D) that were classified as either tetratype (T), parental ditype (PD), or nonparental ditype (NPD) based on uracil auxotrophy and resistance to G-418. The genotype of the haploid progeny is indicated; in slots 3D and 4C it is deduced. Shaded areas indicate ysl2 arl1 transformants. (B) Dilution series of arl1 neo1-69 (BS1323) and ARL1 neo1-69 (BS917) cells after growth at 30 or 37°C. (C) Growth of arl1 neo1-69 cells (BS1323) transformed with pRS316, pRS316-ARL1, pRS316-ARL1G2A, pRS316-ARL1Q72L, or pRS316-ARL1T32N after growth at 30 or 35°C.

neo1

arl1 neo1

arl1 neo1

ARL1 neo1

neo1

The temperature-sensitive growth of the neo1 mutants was, at least in part, directly dependent on the presence of ARL1, since transformation of ts arl1 neo1 cells with ARL1 recreated temperature sensitivity (Fig. 8C). To further resolve which form of Arl1p is inhibitory, different point mutants of Arl1p were transformed back into arl1 neo1-69 cells. Significantly, like the empty vector, Arl1pG2A, which cannot be myristoylated (27), and Arl1pT32N, presumably restricted to the GDP-bound form, did not cause temperature sensitivity of neo1-69 cells (Fig. 8C). In contrast, expression of Arl1pQ72L, which is assumed to exist mainly in the GTP-bound active form due to impaired GTP hydrolysis, did inhibit growth of the neo1-69 strain at 35°C. These results clearly suggest that both myristoylation and the GTP-bound status of Arl1p are responsible for the temperature sensitivity associated with the neo1-37 and neo1-69 mutations.

Since the prominent morphological alterations of the vacuolar system were one of the most striking phenotypes of the neo1 mutants, we asked whether deletion of ARL1 could rescue this defect. As determined by electron microscopic analysis of ultrathin sections, vacuolar fragmentation as such was not suppressed by arl1. However, deletion of ARL1 seemed to diminish the buildup of disk-shaped vacuolar membrane deformations (as shown in Fig. 7D to K) accumulating after incubation at 37°C for 3 h. While 79% of the ARL1 neo1-69 cells (n = 286) exhibited flattened membrane processes, this fraction was reduced to 60% in arl1 neo1-69 cells (n = 350). Concomitantly, the portion of cells with the spherical type of fragmented vacuolar structures increased from 21% in ARL1 neo1-69 cells to 40% in arl1 neo1-69 cells. In neo1-37 cells, we could not detect a significant change in vacuolar morphology after loss of ARL1. However, since the aberrantly shaped vacuole phenotype is not as pronounced in the neo1-37 strain as in the neo1-69 strain (15% of neo1-37 cells exhibit flattened membrane protrusions), a partial recovery is more difficult to detect. In any event, suppression of the temperature sensitivity of neo1-37 and neo1-69 cells by arl1 most likely does not rely solely on the rescue of the membrane transformation defects caused by neo1 mutants.

Localization by indirect immunofluorescence of HA-tagged Arl1p revealed in most cells a diffuse cytoplasmic and a very faint punctate staining pattern, probably representing the Golgi complex and/or endosomal elements (27) (Fig. 9A). In the neo1-69 strain, the punctate HA-Arl1p structures appeared brighter and larger, and they were clearly separated from the diffuse background in a significantly greater number of cells (Fig. 9A). Quantitation of this effect revealed that 25% of the neo1-69 cells (n = 318) exhibited this clear punctate Arl1p pattern, whereas only 8% of the wild-type cells (n = 283) showed the faint punctate structures. It is noteworthy that this effect on the Arl1p staining pattern was more obvious in neo1-69 cells incubated at 25°C than after a 3-h shift to 37°C. In neo1-37 cells, at both 25 and 37°C, the HA-Arl1p pattern was not significantly affected. This could imply that mutant Neo1-69p ought to be localized properly (see Fig. 4B) to elicit this effect on Arl1p.

View larger version (97K): [in this window] [in a new window]   FIG. 9. Neo1-69p affects the subcellular distribution of HA-Arl1p and HA-Ysl2p. (A) Indirect immunofluorescence with cells lacking HA-Arl1p (BS64) (control), wild-type (wt) cells expressing HA-Arl1p (BS1279), and neo1-69 cells expressing HA-Arl1p (BS1283) was performed as described in the legend to Fig. 3A. (B) Indirect immunofluorescence was performed with cells lacking HA-Ysl2p (BS64) (control) and with wt (BS1121) and neo1-69 (BS1402) cells expressing HA-Ysl2p as described in the legend to Fig. 3A. Asterisk indicates cells staining like control cells.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

 
Neo1p functions within the endomembrane/Golgi system. In this study, we have characterized Neo1p, the only essential member of the proposed APL translocases present in yeast. Extensive localization studies and phenotypic analysis of temperature-sensitive neo1 mutants allowed us to place Neo1p in the endosomal system. (i) Double labeling with established endosomal marker proteins (Ypt51p, Pep12p, and Tlg1p) clearly revealed an extended colocalization of Neo1p with all three proteins. (ii) Mutation of VPS27 and expression of the fusion-active Ypt51Q66L protein resulted in a collapse of the Neo1p-positive spots into a few large aggregates reminiscent of the class E compartment. (iii) At permissive temperatures, neo1 mutants displayed a delay in the endocytic transport of Ste2p and LY toward the vacuole as well as defective biogenesis of the vacuole. Significantly, the general secretory pathway, as shown by invertase glycosylation and secretion and Rer1p trafficking, was unaffected under these conditions. (iv) Neo1p revealed an intimate connection to Ysl2p and Arl1p, both of which act within the endosomal/late Golgi system (1, 26, 38).

In addition to its role within the endosomal system, Neo1p may operate in the Golgi complex. Although in wild-type cells the HA-Neo1p labeling pattern differed from that of the early Golgi markers Ypt1p and Rer1p, in sec7 cells the Neo1p structures collapsed into large clusters containing Ypt1p. Therefore, Neo1p most probably is associated with the late Golgi compartment. This would be consistent with the large extent of colocalization between Neo1p and Tlg1p and the weak colocalization with Ypt1p and Rer1p. Such a localization of Neo1p would also account for the defect in the processing and sorting of Golgi-modified p2 CPY in neo1 mutants.

In a recent study by Hua and Graham (23), Neo1p was proposed to reside within the ER and the early Golgi compartment and was implicated in retrograde membrane traffic between these organelles. The ER/Golgi localization was based on staining of C-terminally Myc-tagged Neo1p. The authors themselves acknowledged that the expression of their Neo1p-13Myc construct, in comparison to a third study (32), possibly "induced an artificially high concentration in the ER." In good agreement with this, our studies showed that two distinct C-terminal modifications of Neo1p gave rise to unstable proteins, which were either fully (HA-Neo1p^Ctail) or partially (Neo1p-C-HA) retained in the ER, like Neo1p-13Myc. Thus, the assignment of Neo1p to the ER solely based on the localization of the C-terminally tagged Neo1p (23) must be considered with caution. Adding further to this confusion, the C-terminally HA epitope-tagged Neo1p described by Pomorski et al. (32) colocalized with Pep12p and was functional. Hence, small variations in the sequence composition of the Neo1p C-terminal region, in combination with differences in strain background, might affect the proper biogenesis of Neo1p and thus could account for the observed discrepancies.

Instability and retention of mutant Neo1p within the ER. A key observation was that the mutant Neo1p versions accumulated in the ER. While Neo1-37p was already retained in the ER under permissive conditions, Neo1-69p still revealed the characteristic endosomal localization. However, as judged by the dramatic increase in the signal intensity, incubation at the nonpermissive temperature (37°C) led to massive accumulation of both mutant proteins within the ER. Not surprisingly, the buildup of this multispanning transmembrane protein may then affect anterograde and retrograde membrane trafficking within the early secretory pathway. Thus, it is conceivable that the protein transport defects observed in temperature-sensitive neo1 mutants after several hours of incubation at 37°C (23) are the indirect consequence of the retention and accumulation of mutant Neo1p proteins within the ER. In fact, like our neo1 alleles, the mutants described in the study by Hua and Graham (23) did not show impaired processing and transport of proteins along the secretory pathway under permissive conditions. Furthermore, the ER-to-Golgi transport defects observed after a 24-h Neo1p depletion using a Gal turnoff approach (23) could also be accounted for by indirect effects, which may have arisen after the very long time (10 generations) required for cell growth to slow down.

The finding that mutated Neo1 polypeptides are less stable and that their ability to exit from the ER is blocked is not without precedent. In fact, several mutants of the plasma membrane H⁺-ATPase Pma1p, another P-type ATPase, were characterized by disrupted folding, an accelerated rate of degradation, and accumulation within prominent punctate bodies positive for the molecular chaperone Kar2p (12, 15, 19). Thus, the various extents of ER mislocalization and the distinct stabilities of Neo1-37p and Neo1-69p could reflect a correlation between poor protein folding and ER retention similar to that established for Pma1p mutants. Like our neo1 mutants, cells accumulating Pma1p^D378N exhibited a slight delay in the maturation of p1 to p2 CPY (48). Since this delay was dependent on the accumulation of mutant Pma1p within the ER, it is possible that the defect observed in neo1-37 cells is the consequence of perturbed ER export of mutant Neo1 polypeptides. However, the fact that p1-to-p2 CPY maturation was also slightly slowed in neo1-69 cells at permissive temperatures in spite of the proper endosomal localization of the majority of Neo1-69p suggests that this delay may be caused by another, yet unidentified deficiency.

Another reason for the failure of the Neo1p mutants to exit the ER that may also be linked to improper folding could be a defect in association with a component required for quality control and efficient sorting of Neo1p to endosomes. It is noteworthy that two other APL family members (Drs2p and Dnf1p) were recently shown to form complexes with integral membrane proteins (Cdc50p and Lem3p, respectively). The assemblies were found to be essential for exit of the two P-type ATPases from the ER and for proper transport to their target compartment (40). It would not be surprising if Neo1p sorting were found to rely on a similar mechanism.

Essential role of Neo1p and the question of redundancy with other Drs2 family members. Neo1p is the only essential APL family member in S. cerevisiae. The lethality caused by NEO1 deletion cannot be rescued by overexpression of Drs2p, which acts at the late Golgi complex (9) (see below). Hints as to why Neo1p could be essential come from the morphology of the endosomal network and the finding that Neo1p spreads throughout this system. In all eukaryotes, endocytic organelles are characterized by a highly complex and pleiomorphic organization consisting of cisternal regions from which thin tubules and large vesicles emanate. The vesicles contain membrane invaginations and are therefore described as multivesicular (33; for a review, see reference 18). Due to the substantial amount of work suggesting a function for Drs2 family members as APL translocases implicated in the regulation of membrane shape (11, 14), it is not too far-fetched to propose a similar function for Neo1p. Thus, this protein could be essential for the generation of the complex structure of the endosomal/late Golgi membrane system. Support for this idea is provided by the morphological defects associated with the neo1 mutants at the ultrastructural level (see below).

The fact that neo1-37 cells grow like wild-type cells at 25°C, in spite of the mislocalization of Neo1-37p to the ER, suggests that either small quantities of Neo1p are properly localized or another Drs2 family member can replace Neo1p under permissive conditions, or both. The best candidate for such a substitute is Drs2p, which also resides in the late Golgi complex (9) and late endosomes (40). Genetic evidence for a functional overlap between these two APL family members is provided by the synthetic lethality displayed between a conditional neo1 mutant and drs2 (23) and the suppression of cdc50 defects by overexpression of NEO1 (40). Thus, Neo1p and Drs2p may indeed share an overlapping function in the late Golgi complex. Nevertheless, the fact that overexpression of Drs2p neither rescues the lethality caused by neo1 (22; our unpublished results) nor suppresses the temperature sensitivity of ysl2 (Fig. 1A) implies that Neo1p fulfills unique cellular functions.

Functional links among Neo1p, Ysl2p, and Arl1p. In this study, we reveal genetic and physical interactions between Neo1p and Ysl2p, and we provide multiple evidence for a genetic interaction between Neo1p and Arl1p. Furthermore, we show that the suppression of ysl2 cells by NEO1 is dependent on ARL1. These interactions, the similar localization of the three proteins, and the resemblance of a number of phenotypes shown by mutants of the three genes oblige us to conclude that Neo1p, Ysl2p, and Arl1p collaborate in membrane trafficking. Significantly, analogous interactions between another yeast APL family member (Drs2p), a Sec7 family Arf GEF (Gea2p), and Arf have recently been identified. While Drs2p and Gea2p have recently been shown to bind to each other physically via a nonconserved, membrane-proximal region within the Drs2p C-terminal tail and parts of the Gea2p Sec7 domain (7), the idea of an interaction between DRS2 and ARF1 was suggested on the basis of genetic evidence (9). The finding that Gea2p interacts directly and functionally with Drs2p is particularly intriguing, because it reveals another similarity between Gea2p, an established high-molecular-mass Arf GEF, and Ysl2p, a potential GEF for Arl1p (26) (see also below). Thus, the strikingly similar links between two potential APL translocases and Arf proteins and their respective regulators strongly support the idea that the generation of lipid bilayer asymmetry could be intimately connected to Arf activity and coat assembly during budding reactions (9, 24). Based on sequence and structural homology, Arl1p could possibly be grouped with the Arf family (29), whose members are implicated in the coat assembly of COPI coatomer and various clathrin adaptors. Therefore, Arl1p, Ysl2p, and Neo1p could, like Arf1p, Gea2p, and Drs2p, be involved in the assembly of coat components.

Since ultrastructural analysis of the neo1 mutants revealed a strong increase in the surface-to-volume ratio of the electron-dense vacuolar structures, it is possible that there is a lack in the proper coordination between translocase activity and vesicle scission. Similar deficiencies in membrane transformation events that lead to the segregation of secretory granules have been observed upon expression of a Gea2p mutant, which does not interact with Drs2p any more (7). The observation that mutated neo1 alleles, rather than neo1, were suppressed by arl1 could mean that mutant Neo1 proteins (either directly or indirectly) engage in unproductive interactions with Arl1p (possibly at the endosome/Golgi complex or at the ER). These assemblies may impair the process of membrane trafficking, thereby causing the growth defect at elevated temperatures. In fact, our results implicating the myristoylated and GTP-bound form of Arl1p as being detrimental in the neo1 ts background are consistent with this idea. The additional finding that the neo1-69 mutation indeed affects the subcellular distributions of Arl1p and Ysl2p (Fig. 9) further supports this idea.

Another strong argument for the concerted action of Neo1p, Ysl2p, and Arl1p is based on our initial findings that both NEO1 and ARL1 are suppressors of ysl2 defects. Multiple evidence indicates a role of Ysl2p in guanine nucleotide exchange for Arl1p (26), which could promote a tight membrane association of Arl1p^GTP. However, since at present we cannot exclude the possibility that the exchange on Arl1p requires another component (x) (Fig. 10), it is possible that Ysl2p plays a partial or different function in the activation of Arl1p (Fig. 10A). In the absence of Ysl2p (Fig. 10B, top), Arl1p may not be activated and thus may not bind efficiently to its target membrane, because the interaction between Arl1p^GDP and membranes is weak (in analogy to Arf1p^GDP [2]). However, an increase in Neo1p levels may help to recruit more Arl1p^GDP to the target membrane, either directly by Neo1p itself or through an increase in the concentration of specific lipids on the cytoplasmic leaflet, or both (Fig. 10B, bottom left). Alternatively, elevated amounts of Arl1p^GDP generated by ARL1 overexpression would also increase the chance of the transient association with Neo1p containing membranes and/or lipid microdomains generated by Neo1p (Fig. 10B, bottom right). In both cases, this may result in the activation of Arl1p to its GTP-bound form, bypassing the need for Ysl2p. In fact, evidence for the involvement of the anionic phosphatidylserine, one of the major substrates of APL translocases (10), in the membrane recruitment of both Arf1p^GDP and the Sec7 family GEF ARNO and in the efficient activation of Arf1p^GDP to Arf1p^GTP has been provided recently (2, 8). Clearly, further studies will be required to corroborate such a model and to determine the precise mechanism of Neo1p, Ysl2p, Arl1p and other components in endosome dynamics.

View larger version (28K): [in this window] [in a new window]   FIG. 10. Model for the concerted functioning of Ysl2p, Neo1p, and Arl1p in wild-type cells (A) and for suppression by NEO1 and ARL1 of ysl2 defects (B). Details are given in the Discussion. x, putative component that may act in concert with Ysl2p in the activation of Arl1p.

	ACKNOWLEDGMENTS

 
We are very grateful to D. Gallwitz, J. Holthuis, S. Munro, H. Riezman, H. Rudolph, K. Sato, and B. Séraphin for providing plasmids, strains, and reagents used in this study. We acknowledge R. Krause for initial homology searches and A. Ziegler for assistance in the screen for temperature-sensitive mutants of neo1. We thank H. Rudolph for critical reading of the manuscript, the reviewers for helpful comments, and D. H. Wolf for general advice.

This work was supported by a grant from the DFG to B.S.-K. (Si 635).

	FOOTNOTES

 
* Corresponding author. Mailing address: Institut für Biochemie, Universität Stuttgart, Pfaffenwaldring 55, D-70569 Stuttgart, Germany. Phone: 0049-711-685-4387. Fax: 0049-711-685-4392. E-mail: Singer-Krueger{at}po.uni-stuttgart.de