Isolation of the ecm33⁺, aah3⁺, and gaz2⁺ genes as multicopy suppressors of zinc transporter cis4-1 mutant

We have previously demonstrated that Cis4 is a zinc transporter belonging to the CDF protein family, and plays a role in Golgi membrane trafficking in fission yeast [11]. To better understand the function of Cis4, we screened for genes that when overexpressed could suppress the Cl⁻ hypersensitivity of cis4-1 mutant. The cis4-1 mutant cells grew well in rich YPD medium, however, in the presence of 0.15 M MgCl₂, the cis4-1 cells failed to grow, whereas wild-type cells grew well (Figure 1A). Notably, overexpression of the ecm33⁺ gene partially suppressed the MgCl₂ sensitivity of cis4-1 mutant, and overexpression of the aah3⁺ and gaz2⁺ genes more strongly suppressed the MgCl₂ sensitivity of the cis4-1 mutant (Figure 1A). Then we further determined the growth rates of ecm33⁺ overexpression in liquid media to assess the level of the suppression of the phenotype. Consistently, results showed that cis4-1 mutant cells harboring the multicopy vector grew almost normally but stopped growing 1 h after shift to the YPD media containing 0.15 M MgCl₂. However, the cis4-1 mutant cells expressing ecm33⁺ gene could grew in the presence of 0.15 M MgCl₂, although the growth was slower than that of the cis4-1 mutant cells harboring cis4⁺ gene (Figure 1B). Likewise, these three genes complemented the FK506-sensitive phenotype of the cis4-1 mutant (Figure 1A). Then we examined in Δcis4 deletion mutants the effects of the overexpression of ecm33⁺, aah3⁺, and gaz2⁺ genes, respectively, and results showed that these genes also suppressed the MgCl₂-sensitive and FK506-sensitive growth defect of the Δcis4 cells (our unpublished data).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Isolation of Ecm33, Aah3, and Gaz2 as multicopy suppressors of the cis4-1 mutant cells.

(A) The cis4-1 mutant cells were transformed with either the pDB248 multicopy vector or the vector containing ecm33⁺, aah3⁺, or gaz2⁺. Cells were then streaked onto plates containing YPD, YPD plus 0.15 M MgCl₂, or YPD plus 0.5 µg/ml FK506 and then incubated for 4 days at 30°C. (B) The cis4-1 mutant cells were transformed with either the pDB248 multicopy vector (closed triangles), or the vector containing cis4⁺ (open circles) or ecm33⁺ (open squares). Cells were then diluted with fresh EMM or EMM plus 0.15 M MgCl₂ and incubated at 30°C. Growth was recorded by measurement of the absorbance at 660 nm. Data were averaged from three independent experiments, each sample done in duplicate.

https://doi.org/10.1371/journal.pone.0041946.g001

The ecm33⁺ gene encodes a 43.3 kDa protein (Ecm33) comprising 421 amino acids and containing a signal peptide for GPI anchor in its N-terminus. The aah3⁺ gene encodes a cell surface GPI-anchored protein (Aah3) consisting of 564 amino acids (63.2 kDa) and containing an alpha-amylase domain as well as a DUF1966 domain of unknown function. The gaz2⁺ gene, based on the nucleotide sequence determination, encodes a conserved fungal protein of 317 amino acids. Notably, the N-terminal portion of Gaz2 contains an amino acid signal sequence, and in addition, the gaz2⁺ gene product (Gaz2) has a serine-rich region. These three proteins are conserved in fungi, and Ecm33 is structurally similar to the budding yeast Pst1p and Ecm33p, while Aah3 and Gaz2 have no apparent S. cerevisiae ortholog. The amino acid sequence similarity among Ecm33, Gaz2, and Aah3 are considerably low, and the domain structure is distinct from each other. Ecm33 is a member of the Ecm33/Sps2 family, Aah3 is an alpha-amylase protein, and Gaz2 seems to be a non-enzymatic serine-rich cell wall protein. The only structural thing they have in common is that they contain signal peptides for ER entry and GPI anchoring. Probably, a common feature of these three proteins is that they are highly glycosylated, and the suppression is related to their glycosylation onto the proteins. As the feature of these three proteins is their high glycosylation, the suppression might be due to an indirect effect of overexpressing the GPI proteins.

Phenotypes of ecm33⁺, gaz2⁺, and aah3⁺ deletion mutants

We constructed a null mutation in the ecm33⁺ and gaz2⁺ genes, respectively (see Materials and Methods) and found that the gaz2 deletion mutant was also viable (Figure 2A, upper panel), indicating that Gaz2 is not essential for cell viability. Then we compared the phenotypes of ecm33⁺, gaz2⁺, and aah3⁺ gene deletion mutants. With regard to the cis phenotypes including FK506 sensitivity and MgCl₂ sensitivity [11], Δecm33 cells exhibited sensitivity to both FK506 and MgCl₂, whereas the Δgaz2 and Δaah3 cells were not sensitive to FK506 or MgCl₂. With regard to CaCl₂ sensitivity, Δaah3 cells failed to grow on YPD plate containing 0.15 M CaCl₂, whereas Δecm33 and Δgaz2 cells grew well on the same plate. With regard to temperature sensitivity, Δaah3 cells were very sensitive to cold temperature while the others were not sensitive, and all were not sensitive to high temperature. With regard to the altered sensitivity to the plasma membrane perturbing agent, sodium dodecyl sulfate (SDS), Δecm33 and Δgaz2 but not Δaah3 cells were significantly more resistant to SDS as compared with that of the wild-type cells (Figure 2A, upper panel).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. The Δecm33, Δaah3, and Δgaz2 mutants displayed distinct phenotypes.

(A) Phenotypes of the Δecm33, Δaah3, and Δgaz2 mutants. Upper panel, Cells were streaked onto each plate as indicated, and then incubated at 30°C for 4 days, at 36°C for 3 days or at 17°C for 7 days. Lower panel, MgCl₂-sensitive and FK506-sensitive phenotypes of Δecm33 and cold temperature-sensitive phenotype of Δaah3 were osmoremedial, whereas CaCl₂-sensitive phenotype of Δaah3 was not. Cells were streaked onto each plate as indicated, and then incubated at 30°C for 4 days or at 17°C for 7 days. (B) Effects of overexpression of the gaz2⁺, aah3⁺, and cis4⁺ genes on the phenotypes of Δecm33, and effects of overexpression of the ecm33⁺, gaz2⁺, and cis4⁺ genes on the phenotypes of Δaah3. Wild-type cells, Δecm33 or Δaah3 cells transformed with a control vector or the vector containing ecm33⁺, gaz2⁺, aah3⁺, and cis4⁺ were spotted onto YPD plates or YPD plus 0.15 M MgCl₂ and incubated at 30°C for 4 days.

https://doi.org/10.1371/journal.pone.0041946.g002

Because some of the GPI-anchored proteins were found to be involved in cell wall integrity [15], we then examined whether the phenotypes of these three GPI-anchored protein mutants were suppressible by osmotic stabilization of the medium with sorbitol. Our results showed that in Δecm33 cells, sorbitol suppressed the FK506 sensitivity and MgCl₂ sensitivity of the cells. In Δaah3 cells, sorbitol suppressed the cold temperature sensitivity of the cells, whereas sorbitol failed to suppress the CaCl₂ sensitivity of the cells (Figure 2A, lower panel). Consistent with these results, Morita et al showed that the morphological defect of Δaah3 cells were not rescued in the presence of 1.2 M sorbitol-YES medium [13].

Analysis of the overlapping functions among the three GPI-anchored proteins

As described above, the domain structures of Ecm33, Aah3, and Gaz2 are distinct, therefore it seems likely that ecm33⁺, aah3⁺ and gaz2⁺ are not functionally redundant. To test this possibility, we examined the effects of the overexpression of gaz2⁺ and aah3⁺ on the phenotypes of the ecm33⁺ deletion mutants, as well as the overexpression of ecm33⁺ and gaz2⁺ on the phenotypes of the aah3⁺ deletion mutants. As shown in Figure 2B, the results showed that overexpression of gaz2⁺, but not aah3⁺, suppressed the MgCl₂-sensitive growth defect of the ecm33⁺ gene deletion mutants. On the other hand, overexpression of the ecm33⁺ or gaz2⁺ genes failed to suppress the phenotypes of the aah3⁺ deletion mutants. Thus, these findings suggest that the structures of the three GPI-anchored proteins are distinct from each other, and that these three proteins have only partial overlapping functions. We also examined the effects of the overexpression of cis4⁺ on the phenotypes of the Δecm33 and Δaah3 mutants. The results showed that overexpression of cis4⁺ suppressed the MgCl₂-sensitive growth defect of the Δecm33 mutants, but failed to suppress the CaCl₂-sensitive phenotype of the Δaah3 mutants (Figure 2B).

Deletion analysis of the ecm33⁺ gene

To determine the functional region of Ecm33, we prepared a series of truncated forms of Ecm33. Structural feature of the deletion mutants employed in this study is illustrated in Figure 3A. Results showed that in Δecm33 mutants, the overexpression of the full-length Ecm33 as well as Ecm33 fragment A, fragment B, fragment C, fragment F, and fragment G suppressed the phenotypes of the mutants (Figure 3B). However, overexpression of Ecm33 fragment D, fragment E, fragment H, fragment I, fragment J, and fragment K failed to suppress the phenotypes of Δecm33 mutants (Figure 3B). Overexpression of these truncated versions of Ecm33 showed similar genetic suppression profile of the cis4-1 mutant as compared with that of the Δecm33 cells (our unpublished data).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Deletion analysis of the ecm33⁺.

(A) Structural features of the truncated mutants of Ecm33. (B) The Δecm33 mutant phenotype suppression by Ecm33 truncated mutants. Cells transformed with the multicopy vector or vector containing various truncated genes were spotted onto each plate as indicated and incubated for 4 days at 30°C. (C) Protein levels of Ecm33 examined by immunoblot analysis. The ecm33 deletion cells were transformed with the ecm33⁺, or truncated ecm33⁺ gene were grown to mid-log phase in EMM at 30°C. Cells were washed and incubated for 24 h and then analyzed by immunoblotting using anti-Ecm33 monoclonal antibody as described under “Materials and Methods.” The panel indicated as CBB stain shows the CBB staining of the same gel to show the presence of equal amount of proteins in each lane.

https://doi.org/10.1371/journal.pone.0041946.g003

We also examined the protein levels of truncated mutants of Ecm33 in the Δecm33 cells by immunoblotting with anti-Ecm33 monoclonal antibody [12] (Figure 3C). The immunoblot analysis detected an appreciable amount of Ecm33 fragments A, C, F, G, and I, but failed to detect fragment B, D, E, H, J or K. These results are consistent with the above results that overexpression of Ecm33 fragments A, C, F, G except for fragment I suppressed the phenotypes of the Δecm33 cells. In addition, the Ecm33 fragment B was not detected in the cells by immunoblotting, although overexpression of the fragment B suppressed the phenotypes of the Δecm33 cells. It is possible that the Ecm33 fragment B contains the epitope for the monoclonal antibody. The reasons for the inability of the antibody to detect other Ecm33 fragments as well as the functional importance of these fragments are unknown.

Subcellular localization of Ecm33 and Gaz2

In order to investigate the subcellular localization and membrane trafficking of the Ecm33 and Gaz2 protein, plasmids carrying GFP-Ecm33 and GFP-Gaz2 fusions, respectively, were constructed. On Ecm33, a GFP carrying the S65T mutation was inserted into 60 bp from the N-terminus of Ecm33. Fujita et al demonstrated that HA- and mRFP-tagged versions of Gas1, a well-characterized GPI-anchored protein in Saccharomyces cerevisiae, were generated by inserting the tags in Gas1 immediately following the N-terminal signal sequence and both the tagged proteins were functional [16]. Therefore, we constructed GFP-Ecm33 fusion by inserting the GFP tag immediately following the N-terminal signal sequence. Results showed that the construct of GFP-Ecm33 was functional as cells expressing GFP-Ecm33 suppressed the phenotypes of Δecm33 mutants (Figure 4A). Then, we examined the localization of GFP-Ecm33 expressed from its own promoter in wild-type cells, and results showed that GFP-Ecm33 localized to the cell surface or the medial regions. This observation was similar to the sterol localization of filipin fluorescence that was enriched in the plasma membrane at the growing cell tips and at the site of cytokinesis (Figure 4B). Also, this finding was consistent with the data obtained using anti-Ecm33 antibody by immunofluorescence microscopy [12]. On Gaz2, a GFP was inserted into 600 bp from the N-terminus of Gaz2 because deletion of 500–600 bp of Gaz2 gene did not affect its suppression ability on the phenotype of cis4-1 mutant. However, the construct was not functional and cells expressing GFP-Gaz2 failed to suppress the phenotypes of the cis4-1 mutant (our unpublished data). Presumably, this is because the GFP tag is inserted somewhere in the middle of the Gaz2 protein.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Subcellular localization of GFP-Ecm33.

(A) GFP-tagged Ecm33 was functionally similar to the non-tagged protein. Cells transformed with the multicopy vector harboring gene encoding GFP-Ecm33 or the empty vector was streaked onto each plate containing YPD or YPD plus 0.12 M MgCl₂, and then incubated for 4 days at 30°C. (B) GFP-Ecm33 localized to the cell surface or medial regions in wild-type cells. Wild-type cells expressing chromosome-borne GFP-Ecm33 were cultured in EMM medium at 27°C, and were examined by fluorescence microscopy; Sterol localization detected by filipin staining (see Materials and Methods) in wild-type cells. Bar: 10 µm. (C) GFP-Ecm33 localized to the structure surrounding the nucleus in wild-type cells when cells were treated with BE49385A. Wild-type cells expressing chromosome-borne GFP-Ecm33 were cultured in EMM medium and incubated at 27°C, and were examined by fluorescence microscopy at 0 hour, 1 hour, 2 hour, and 4 hour after 1 µg/ml BE49385A was added to the medium. Bar: 10 µm. (D) GFP-Ecm33 primarily localized to the structure surrounding the nucleus and to the cell surface in its8-1 mutant, while the localization of GFP-Ecm33 was normal in Δcis4 mutant. The its8-1 mutant and Δcis4 mutant expressing chromosome-borne GFP-Ecm33 were cultured in EMM medium and incubated at 27°C, and were examined by fluorescence microscopy. Bar: 10 µm. (E) In the zinc-deficient medium GFP-Ecm33 primarily localized to the intracellular dots and to the ER in Δcis4 cells. The wild-type cells and Δcis4 mutant expressing chromosome-borne GFP-Ecm33 were cultured in the zinc deficient EMM medium at 27°C for 48 hours, and were examined by fluorescence microscopy. Bar: 10 µm.

https://doi.org/10.1371/journal.pone.0041946.g004

Next, we examined the effect of BE49385A, an inhibitor of Its8, on the subcellular localization of GFP-Ecm33 in wild-type cells. We previously identified a mutation in the its8⁺ gene that are involved in GPI anchor synthesis, and showed that Its8 is a molecular target of BE49385A [10]. In wild-type cells, before the addition of BE49385A as shown in Figure 4C, GFP-Ecm33 localized to the cell surface or the medial regions. Then, 1 hour after the addition of 1 µg/ml BE49385A to the medium, GFP-Ecm33 mostly localized at the cell surface (Figure 4C, arrowheads) and also localized at the structure surrounding the nuclei that are considered to be the endoplasmic reticulum (ER) (Figure 4C, arrows). Then, 2 hours or 4 hours after treatment with BE49385A, GFP-Ecm33 mostly localized to the nuclear envelope and the peripheral ER rather than the cell surface (Figure 4C, arrows). In its8-1 mutant cells, the subcellular localization of GFP-Ecm33 was also examined. As expected, GFP-Ecm33 primarily localized to the ER and to the cell surface in the its8-1 mutant cells (Figure 4D, arrows), suggesting that the impairment of GPI anchor synthesis caused the defective attachment of GPI-anchor to the Ecm33 protein thereby resulting in the abnormal GFP-Ecm33 localization in the ER. Then we also examined the subcellular localization of GFP-Ecm33 in the Δcis4 cells, and results showed that GFP-Ecm33 was observed at the cell surface or the medial region in the Δcis4 cells (Figure 4D) similar to that observed in the wild-type cells (Figure 4B). We further examined the effect of zinc deficiency on the subcellular localization of GFP-Ecm33 in the Δcis4 cells by removing zinc from the EMM medium. As shown in Figure 4E, the fluorescence of GFP-Ecm33 in wild-type cells was enriched in the cell surface and the medial region in the zinc-deficient medium. In contrast, in Δcis4 cells GFP-Ecm33 primarily localized to the intracellular dots and to the ER in the zinc-deficient medium (Figure 4E).

Genetic interaction between cis4⁺ and its8⁺ genes

Then we examined the effect of Zn²⁺ on the phenotypes of the its8-1 mutant cells. The results showed that the addition of Zn²⁺ to the medium significantly rescued the high temperature-sensitive and FK506-sensitive phenotypes of the its8-1 mutant (Figure 5A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 5. Genetic interaction between cis4⁺ and its8⁺ genes.

(A) Effect of the addition of extracellular Zn²⁺ on the phenotypes of its8-1 mutant. Wild-type or its8-1 mutant cells were spotted onto each plate as indicated and then incubated for 4 days at 27°C or at 35°C. (B) The its8-1Δcis4 double mutants showed more marked temperature sensitivity than the single mutants, while the double mutants showed similar BE49385A-sensitivity as compared with that of the its8-1 mutant. Wild-type, its8-1, Δcis4, and its8-1Δcis4 cells were spotted onto each plate as indicated and then incubated for 4 days at 27°C or at 33°C.

https://doi.org/10.1371/journal.pone.0041946.g005

Our previous study suggested that Cis4 localizes to the cis-Golgi and was involved in Golgi membrane trafficking through regulating the zinc homeostasis [11]. In order to investigate the functional relationship between Cis4 and Its8, we constructed its8-1Δcis4 double mutants, and examined the effect of temperature, BE49385A, and zinc deficiency respectively on these cells. On the effect of temperature, in the its8-1Δcis4 double mutants, these cells exhibited more marked temperature sensitivity than that of the its8-1 single mutants (Figure 5B), suggesting that there is a genetic interaction between Its8 and Cis4. On the effect of BE49385A, in the Δcis4 cells interestingly, the growth of these single deletion cells was significantly inhibited by BE49385A as compared with that of the wild-type cells, although the sensitivity of the Δcis4 cells was not as severe as that of the its8-1 mutant (Figure 5B). In the its8-1Δcis4 double mutants, notably, these cells exhibited the BE49385A-sensitive growth defects similar to that of the its8-1 single mutant. On the effect of zinc deficiency, in the its8-1Δcis4 double mutants interestingly, these cells were observed to have a very small colony size similar to that of the Δcis4 single mutant in the zinc-deficient medium (Figure 5B). These results suggest that the impairment of GPI-anchor synthesis and zinc-ion homeostasis in the double mutant is similar to that of its8-1 and Δcis4 single mutants, respectively.

Furthermore, we examined the effect of overexpressed multicopy suppressors of the cis4-1 mutant on the MgCl₂-sensitive phenotypes of the apm1-1 mutant, an allele of apm1⁺ gene that encodes μ1A subunit of the clathrin-associated adaptor protein complex 1(AP-1) implicated in the Golgi/endosome function [17]. The multicopy suppressors of the cis4-1 mutant identified here were the three genes encoding GPI-anchored protein namely ecm33⁺, aah3⁺ and gaz2⁺, and in addition, two other multicopy suppressor genes including pmp1⁺, and SPCC1322.03 (trp1322⁺) that also suppressed the MgCl₂-sensitive phenotype of cis4-1 mutant (Materials and Methods). The pmp1⁺ gene encodes a dual-specificity MAPK phosphatase that negatively regulates the Pmk1 MAPK signaling [18]. The trp1322⁺ gene encodes transient receptor potential (TRP)-like ion channel that mediates the cytoplasmic Ca²⁺ rise caused by the extracellularly added CaCl₂ [19]. As shown in Table 1, all the multicopy suppressors with the exception of the aah3⁺ gene significantly suppressed the MgCl₂-sensitive phenotype of the apm1-1 mutant.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Complementation of the MgCl₂-sensitive phenotype of the cis4-1 mutant and apm1-1 mutant.

https://doi.org/10.1371/journal.pone.0041946.t001

Localization of GFP-Ecm33 in various membrane trafficking mutants

As shown above, the overexpression of the GPI-anchored proteins suppressed the MgCl₂ sensitivity of the mutant allele of the apm1⁺ gene. This prompted us to hypothesize that Apm1 may play roles in membrane trafficking of the GPI-anchored proteins. Then, we examined the localization of GFP-Ecm33 in Δapm1 cells. In wild-type cells, GFP-Ecm33 clearly localized at the cell surface and the medial regions as shown in Figure 4B. In Δapm1 cells, in contrast, GFP-Ecm33 primarily localized as dot-like structures that were observed in the cytoplasm (Figure 6A, arrows) as well as at the cell surface and the division site (Figure 6A, arrowheads). Next, we examined the localization of GFP-Ecm33 in other membrane trafficking mutants namely, ypt3-i5 mutant and chc1-1 mutant. The Rab/Ypt GTPase Ypt3 has been implicated in the membrane trafficking associated with the Golgi complex, and its mutation confers sensitivity to FK506 and defects in cell wall integrity [20]. The chc1⁺ gene encodes clathrin heavy chain Chc1 involved in intracellular protein transport. Results showed that the chc1-1 mutant exhibited its (its for immunosuppressant- and temperature-sensitive) phenotype [21] (Figure S1A). Sequence analysis of the genomic DNA from the chc1-1 mutant revealed that arginine at 1615 was mutated to a termination codon by a C-to-T transition (CGA to TGA), and resulted in a truncated protein product lacking 51 amino acids downstream of the mutation (Figure S1B). In the membrane trafficking mutants including chc1-1 mutant and ypt3-i5 mutant, results showed that GFP-Ecm33 also localizes as intracellular dot-like structures (Figure 6A, arrows) in addition to the cell surface and the division site (Figure 6A, arrowheads). So, we examined whether the dot-like fluorescence of GFP-Ecm33 co-localized with the endocytic tracer dye FM4-64 during an early stage of endocytosis in Δapm1, chc1-1 mutant and ypt3-i5 mutants. After 5 min of dye uptake, most of the GFP-Ecm33 dot-like structures co-localized with FM4-64-positive structures in Δapm1, chc1-1 mutant and ypt3-i5 mutants (Figure 6A, Merge). This strongly suggests that the intracellular dot-like fluorescence of GFP-Ecm33 represents Golgi/endosome compartments. Then we further examined the co-localization of GFP-Ecm33 with Krp1 fused to monomeric red fluorescent protein (RFP) at its C-terminus. Krp1 is a furin/Kex2 homolog that resides in the Golgi/endosome [22], [23]. As shown in Figure 6B, intracellular GFP-Ecm33 mostly co-localized with Krp1-RFP (Figure 6B). Thus, GFP-Ecm33 localized at Golgi/endosome structures in addition to the cell surface and the division site in these mutants, suggesting that GPI-anchored proteins were not correctly transported and were retained at the Golgi/endosome structures in these membrane trafficking mutants. Similarly, in the wild-type cells, GFP-Gaz2 also clearly localized at the cell surface and medial regions (Figure 6C), while in Δapm1 cells, GFP-Gaz2 localized as intracellular dot-like structures (Figure 6C). Interestingly, all of the membrane trafficking mutants that were tested showed hypersensitivity to BE49385A (Figure 6D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 6. Subcellular localization of GFP-Ecm33 in membrane trafficking mutants.

(A) The Apm1-deletion (Δapm1), chc1-1 mutant and ypt3-i5 mutant expressing chromosome-borne GFP-Ecm33 cultured in EMM medium were incubated with FM4-64 fluorescent dye for 5 min at 27°C to visualize Golgi/endosomes. GFP-Ecm33 localization and FM4-64 fluorescence were examined under a fluorescence microscope. Bar: 10 µm. (B) Localization of GFP-Ecm33 and Krp1-RFP in the membrane trafficking mutants. The Apm1 deletion (Δapm1), chc1-1 mutant and ypt3-i5 mutant expressing chromosome-borne GFP-Ecm33 were transformed with the pREP1 vector expressing Krp1 fused to RFP at its C-terminus. Cells were grown to early log phase in EMM containing 4 µM thiamine at 27°C, and were examined by fluorescence microscopy. Bar: 10 µm. (C) GFP-Gaz2 localized to the cell surface or medial regions in wild-type cells, while GFP-Gaz2 localized to the intracellular dots in the Δapm1 cells. Cells transformed with the multicopy vector GFP-Gaz2 were cultured in EMM medium at 27°C, and were examined by fluorescence microscopy. Bar: 10 µm. (D) All the membrane trafficking mutants tested namely ypt3-i5, chc1-1, and Δapm1 mutants showed hypersensitivity to BE49385A. Wild-type, ypt3-i5, chc1-1, and Δapm1 cells were spotted onto each plate containing YPD or YPD plus 30 ng/ml BE49385A as indicated, and then incubated at 27°C for 4 days.

https://doi.org/10.1371/journal.pone.0041946.g006

GPI-anchored proteins and clathrin-mediated post-Golgi membrane trafficking

Important finding of this study is the role of clathrin-mediated post-Golgi membrane trafficking pathway in the transport of GPI-anchored proteins in fission yeast. In budding yeast, GPI-anchored proteins are transported from the ER to the Golgi apparatus in vesicles distinct from those containing non-GPI-anchored proteins such as the general amino acid permease Gap1p and pro-alpha factors [24], [25], [26], and the transport of GPI-anchored proteins and non-GPI-anchored proteins from the trans-Golgi network (TGN) to the plasma membrane is also regulated by different sorting and packaging machinery [27]. Consistent with these, Castillon et al observed that GPI-anchored proteins accumulate in ER exit sites (ERES) that are distinct from those in which other secretory proteins accumulate [28]. More recently, Rivier et al reported that in mammalian cells, GPI-anchored and other secretory proteins are not segregated upon exit from the ER, in contrast to the remarkable segregation seen in budding yeast [29].

In this study in fission yeast, we observed the subcellular localization of a GPI-anchored protein Ecm33 using GFP fusion proteins in vivo. Results showed the abnormal localization of GFP-Ecm33 in all of the membrane trafficking mutants tested including ypt3-i5, chc1-1, and Δapm1 mutants. The AP-1 complex that links a clathrin to the membrane plays a role in the post-Golgi membrane trafficking including exit transport from the TGN to endosomes, endosomes to the TGN, and TGN or endosomes to the plasma membrane [23]. In the Δapm1 and chc1-1 mutants, GFP-Ecm33 was primarily seen as dot-like structures presumed to be the Golgi/endosomes, suggesting the delay in the clathrin-mediated post-Golgi membrane trafficking of GPI-anchored protein Ecm33 to the cell surface in these membrane trafficking mutants. Moreover, GFP-Ecm33 also localized at Golgi/endosome structures in addition to the cell surface and the division site in the ypt3-i5 mutant. Ypt3 is involved at multiple steps of the fission yeast membrane trafficking events, namely, at the exit from the trans-Golgi as well as at the later step of the exocytic pathway [20]. Instead, Ypt31p and Ypt32p, the homolog of Ypt3 in S.cerevisiae, have been reported that implicated in the exocytic pathway and mediates intra-Golgi traffic or the budding of post-Golgi vesicles from the trans-Golgi [30], [31], [32]. Thus, our results suggest that Ypt3 plays roles in transport of GPI-anchored proteins at multiple steps including the exit from the trans-Golgi as well as at the later step of the exocytic pathway. Furthermore, the localization of GFP-Gaz2 in the Δapm1 cells was similar to that of GFP-Ecm33, strongly suggesting that GPI-anchored proteins are transported through a clathrin-mediated post-Golgi membrane trafficking pathway that is required for the efficient transport of other secretory proteins in fission yeast. Of course, our results do not rule out the possibility that some GPI-anchored proteins might still be separately sorted from the secretory proteins in fission yeast. Further studies will be required to clarify the molecular mechanisms of membrane trafficking of GPI-anchored proteins. Given the high similarity between the fission yeast and the mammalian cells, this study may provide a basis for understanding the precise mechanism of membrane trafficking of GPI-anchored proteins in higher eukaryotes.

Cis4 is involved in the membrane trafficking of GPI-anchored proteins

In the present study, we present several lines of evidence that suggests a role of Cis4 in membrane trafficking of GPI-anchored proteins. In our previous study, we established that zinc transporter Cis4 is implicated in Golgi membrane trafficking through the regulation of zinc homeostasis in fission yeast [11]. In this study, we first showed that the overexpression of several GPI-anchored proteins that have distinct structures, namely Ecm33, Aah3, and Gaz2, suppressed the phenotypes of the Δcis4 mutants. Second, there is a genetic interaction between the genes encoding Cis4 and Its8, because the its8-1Δcis4 double mutant cells were more sensitive to high temperature than that of the single mutants. Third, the Δcis4 mutants were sensitive to BE49385A, an inhibitor of Its8. In particular, the its8-1Δcis4 double mutants exhibited the same BE49385A-sensitive growth defects as that of the its8-1 single mutants, while the double mutant showed defective growth similar to that of the Δcis4 single mutants in the zinc-deficient EMM medium. Fourth, overexpression of a majority of the multicopy suppressors of the cis4-1 mutant complemented the MgCl₂-sensitive phenotype of the apm1-1 mutant. The biosynthesis of GPI anchors and its attachment to the target protein are carried out on the ER membrane, and then transported to the plasma membrane by vesicular trafficking [25]. Previous study reported that the synthesis of GPI anchors is zinc dependent both in vitro and in vivo [33], [34]. It is consistent with our results that the addition of Zn²⁺ to the medium significantly rescued the high temperature-sensitive and FK506-sensitive phenotypes of the its8-1 mutant (Figure 5A). However, the lines of evidence as presented above suggest that the zinc transporter Cis4 may be indirectly involved in the membrane trafficking of GPI-anchored proteins, rather than involved in the synthesis of GPI anchored proteins. Consistent with this hypothesis, the subcellular localization of GFP-Ecm33 in the Δcis4 cells was normal and did not exhibit an abnormal accumulation in the ER in the normal medium (Figure 4D). Probably, there exists in the Golgi as yet unknown zinc-requiring components that are involved in the membrane trafficking of GPI-anchored protein, and that Cis4 delivers Zn²⁺ to the Golgi to regulate membrane trafficking of GPI-anchored proteins.

Strains, media, genetic and molecular biology techniques

S. pombe strains used in this study are listed in Table S1. The complete medium, YPD, and the minimal medium, EMM, have been described previously [35]. Standard S. pombe genetic and recombinant-DNA methods were performed as described previously except where noted [36]. Gene disruptions are denoted by lowercase letters representing the disrupted gene followed by two colons and the wild-type gene marker used for disruption (for example, gaz2::ura4⁺). Gene disruptions are abbreviated by the gene preceded by Δ (for example, Δgaz2). Proteins are denoted by roman letters and only the first letter is capitalized (for example, Gaz2). Tacrolimus (FK506) was obtained from Astellas Pharma (Tokyo, Japan). All other chemicals and reagents were purchased from commercial sources.

Multicopy suppressor screen

To identify multicopy suppressors of the high MgCl₂ sensitivity of cis4-1 mutant, a genomic library cloned into the vector pDB248 [37] was transformed into the cis4-1 mutant cells. The Leu⁺ transformants were replica-plated onto YPD plates containing 0.15 M MgCl₂ and the plasmid DNA was recovered from transformants that showed a plasmid-dependent rescue. These plasmids complemented MgCl₂ sensitivity of the cis4-1 mutant. By DNA sequencing, the suppressing plasmids were found to belong to six classes, with one class containing the cis4⁺ gene [11], and other classes containing the ecm33⁺ gene, SPBC1E8.05, aah3⁺ gene, pmp1⁺ gene, and trp1322⁺ gene (SPCC1322.03). Here we focus on the ecm33⁺ gene, SPBC1E8.05, and the aah3⁺ gene that encodes GPI-anchored proteins, and renamed SPBC1E8.05 gene as gaz2⁺ (gaz for GPI-anchored protein that suppress the zinc transporter deletion).

Knockout of the ecm33⁺, gaz2⁺, and aah3⁺ genes

To knockout the ecm33⁺ gene, a PCR-based targeted gene deletion method was prepared by the Cre-loxP-mediated marker removal procedure as described previously [38] using the sense primer 5′-CAT AGC AAG AGC AGC AAC CAA AAG AGA TCC CAA AAC TAA AGC ACC AGC AGT GAA GCC GTT AGA AGC GGC TGA GCC CAA TAG GCC GAA ATC GGC AAA ATC CC-3′, and the antisense primer 5′-GTT GTT CAA ATC ATT CGC TCT CAC TCT TCT TTT CGC CGC AGC TCG CGT ACA AGC TGC TTC CAA CTG CTC CAG CGG CCC GGT GAT GGT TCA CGT AGT GGG CC-3′. The disruption of the gene was checked using genomic Southern hybridization (our unpublished data).

To knockout the gaz2⁺ gene, a one-step gene disruption by homologous recombination was performed [39]. The gaz2::ura4⁺ disruption was constructed as follows. The open reading frame (ORF) of gaz2⁺ was amplified by PCR (forward primer2232, 5′-CCG CTC GAG CAC CAT GAA GTT GTC TTT CAT TTT ATC TAC TCT CG-3′; reverse primer 2233, 5′-ATA GTT TAG CGG CCG CCA AGA AAC AAG GCA ATA GCA GAA ACA ACA CC-3′) from the genomic DNA and was subcloned into the XhoI/NotI site of BlueScriptSK (+). Then a HindIII fragment containing the ura4⁺ gene was inserted into the HindIII site of the previous construct. The construct containing the disrupted gaz2⁺ gene was digested with XhoI/NotI, and the gaz2::ura4⁺ fragment was used to transform the diploids (5A/1D, Table S1). Stable integrants were subsequently cloned on plates containing the medium lacking uracil, and gene disruption by the gaz2⁺ derivative containing the ura4⁺ insertion was verified by genomic Southern blotting (our unpublished data).

The aah3⁺ gene deletion mutant (h⁻ leu1-32 ura4-D18 ade6-M210 aah3::KanMX4) was purchased from BioNEER (South Korea) [40]. We constructed the aah3⁺ gene deletion cells that are not auxotrophic for uracil or adenine by the genetic cross between wild-type cells HM123 and the above strain to make KP5075 (h⁻ leu1 aah3::KanMX4) (Table S1).

Plasmids constructions

The ecm33⁺, gaz2⁺, and aah3⁺ genes, respectively, together with their promoter regions were amplified by PCR using the genomic DNA of wild-type cells as a template. The primers used were summarized in Table S3. The amplified products containing the ecm33⁺ or gaz2⁺ genes were digested with PstI, while the amplified products containing the aah3⁺ gene was digested with HindIII. All the resulting fragments were subcloned into BlueScriptSK (+), to give pKB8044, pKB7850, and pKB8147 respectively.

A series of truncated Ecm33 mutants were constructed as follows. The ecm33⁺ gene lacking about 100 base pairs from the predicted region was amplified by PCR using the plasmid pKB8044 as a template, to yield fragment A, fragment B, fragment C, fragment D, fragment E, fragment F, fragment G, fragment H, fragment I, fragment J, and fragment K (Figure 3A). The primers in each mutant were derived from the upstream and the downstream regions to be deleted, as shown in Table S2.

To study the subcellular localization of Ecm33 and Gaz2, plasmids carrying GFP-Ecm33 and GFP-Gaz2 fusions, respectively, were constructed as follows. First, the sequence for GFP lacking the start codon was amplified by PCR (primers shown in Table S3) from the plasmid containing GFP carrying S65T mutation, and was subcloned into BamHI site of BlueScript SK(+). Next, BamHI site was constructed at 60 bp of Ecm33 and at 600 bp of Gaz2, respectively, by PCR technique using the primers shown in Table S3. Then, a BamHI fragment containing GFP S65T mutation was inserted into the BamHI site of the Ecm33 construct and the Gaz2 construct as described above. To obtain the chromosome-borne GFP, the fused gene GFP-Ecm33 was subcloned into the vector containing the ura4⁺ marker under the control of its own promoter, and was integrated into the chromosome at the ura4⁺ gene locus of KP1248 as described [17], [20]. A successful integration was confirmed by PCR and Southern blot (our unpublished data).

Bioinformatics

Database searches were performed using the National Center for Biotechnology Information BLAST network service (www.ncbi.nlm.nih.gov) and the Sanger Center S. pombe database search service (www.sanger.ac.uk). Sequence alignment was performed using protein BLAST and the ClustalW program.

Microscopy and miscellaneous methods

Methods in light microscopy, such as fluorescence microscopy that was used to observe the localization of GFP-tagged proteins and FM4-64 labeling, were performed as described [17], [20]. Krp1-RFP was expressed as described previously [41]. A fluorescent cholesterol probe, filipin was used to stain sterol as previously described [42].

Cell extract preparation and immunoblot Analysis

Cell extract preparation was performed as described previously [43]. Protein extracts (10–20 µg/5 µl) were subjected to immunoblot analysis with anti-Ecm33 monoclonal antibody [12]. Obtained gel profiles were also visualized by CBB staining as a loading control.