Pun1p is a metal ion-inducible, calcineurin/Crz1p-regulated plasma membrane protein required for cell wall integrity.

Under conditions of environmental stress, the plasma membrane is involved in several regulatory processes to promote cell survival, like maintenance of signaling pathways, cell wall organization and intracellular ion homeostasis. PUN1 encodes a plasma membrane protein localizing to the ergosterol-rich membrane compartment occupied also by the arginine permease Can1. We found that the PUN1 (YLR414c) gene is transcriptionally induced upon metal ion stress. Northern blot analysis of the transcriptional regulation of PUN1 showed that the calcium dependent transcription factor Crz1p is required for PUN1 induction upon heavy metal stress. Here we report that mutants deleted for PUN1 exhibit increased metal ion sensitivity and morphological abnormalities. Microscopical and ultrastructural observations revealed a severe cell wall defect of pun1∆ mutants. By using chemical cross-linking, Blue native electrophoresis, and co-immunoprecipitation we found that Pun1p forms homo-oligomeric protein complexes. We propose that Pun1p is a stress-regulated factor required for cell wall integrity, thereby expanding the functional significance of lateral plasma membrane compartments.

Introduction

The plasma membrane is involved in many cellular processes ensuring adaptation to changed environmental conditions. These include maintenance of intracellular ion homeostasis, regulation of signaling pathways, morphogenesis and cell wall biogenesis.

Studies on the plasma membrane provide strong evidence for a highly regulated structure and presence of discrete domains, like lipid rafts in higher eukaryotes as well as eisosomes and the membrane compartment of the arginine permease Can1 (MCC), being ergosterol-rich plasma membrane invaginations, in yeast [1]. Although their actual function still remains to be determined, these domains were shown to be involved in cell signaling [2] and endocytic protein turnover [3,4].

An important aspect associated with the survival of environmental stress situations, like osmotic or heat stress [5], is the adjustment of ion fluxes across the plasma membrane. Transporters, exchangers and channels, embedded in a highly flexible membrane system, allow an accurate adaptation of the cellular ion homeostasis to changing environmental conditions [6]. Consequently, tightly regulated processes for ion uptake through the plasma membrane, distribution to the appropriate subcellular compartments, as well as detoxification mechanisms are essential for cell survival.

In spite of their potential toxicity at higher concentrations, a number of transition metals e.g. manganese, cadmium, cobalt, zinc, and iron are important trace elements in humans and other organisms. The transport mechanisms and the regulation and function of involved proteins, the biological effects of some of these metals, their toxicity, as well as the respective regulatory mechanisms are not yet completely understood [7-10].

In search of genes involved in cellular adaptation to metal ion stress conditions, we found the PUN1 (YLR414c) gene as being significantly induced upon metal ion stress. PUN1 encodes a plasma membrane MCC protein which was recently shown to be induced during nitrogen starvation stress and involved in filamentous growth [11].

Surprisingly, Pun1p shares structural similarities with mammalian claudins, the major constituents of plasma membrane tight junctions [12], in particular a tetraspan topology and the highly conserved claudin family signature G-L-W-x-x-C-x(8-10)-C within the first extracellular loop (Interpro IPR017974).

In addition to Can1p and Pun1p, seven other proteins are known to be integral components of the MCC. Among them, the uracil permease Fur4p, the tryptophan/tyrosin permease Tat2p, and the tetraspan proteins Sur7p, Nce102p, Fmp45p, Fhn1p, and Ynl194cp, all five of unknown function [4,13]. Besides Fmp45p and Ynl194cp [14], Pun1p is a third S. cerevisiae paralog of Sur7p. Interestingly, all four proteins contain a cysteine motif similar to the claudin signature in their first extracellular loop, but only for Pun1p it is perfectly conserved [15].

Thus, the clustering of tetraspan proteins in MCC compartments raises the challenging question if these proteins and their common domains share evolutionary conserved structural and functional properties.

Materials and methods

Yeast strains, media and culture conditions

Saccharomyces cerevisiae strains were grown at 28 °C in liquid YPD medium (yeast extract/peptone/dextrose), on solid YPD in serial 10-fold dilution steps or in standard SD medium (0.67% yeast nitrogen base, 2% glucose, and amino acids as required). Escherichia coli strain DH10b (Invitrogen, Lofer, Austria) was cultivated at 37 °C in Luria–Bertani (LB) medium supplemented with 100 μg/ml ampicillin when appropriate. Media were supplemented with various metal ion salts when indicated. Yeast strains BY4741 (acc. no. Y00000), W303 (acc. no. 20000A) and FY1679 (acc. no. 10000 M) provided by Euroscarf (Frankfurt, Germany) were used as wild-type. Mutant derivatives of BY4741 used in this study were also provided by Euroscarf. These were BYhog1∆, BYcrz1∆, BYrlm1∆, BYste12∆, BYsko1∆, and BYdig1∆. Corresponding genes were completely deleted and replaced by the geneticin resistance module KanMX4. Strain W303msn2/4∆ was described previously [16]. Strain BY4741 bearing the chromosomal PUN1-GFP fusion was obtained from the yeast GFP-tagged collection [17]. The haploid deletion strain MApun1∆ was constructed from parental strain FY1679. According to the one step replacement protocol [18] the PUN1 ORF was deleted by homologous recombination with a HIS5 disruption cassette amplified from plasmid pSG634 with the primers PUN1-delfor: 5′-TTCGCTAGGAGCACTTATATTAGCCATTGTTGCATGCGCAGGATCCGTACGCTGCAGGTCGAC-3′ and PUN1-delrev: 5′-GACGGTGGGCCTCGATCTCACGCTACACCATCGAATGAAATTCAGTAGGCCACTAGTGGATCTG-3′. Verification of correct gene replacement was performed by analytical PCR using primers PUN1-up: 5′-AAATCGGGCGTACTATCAGCCAAGC-3′, PUN1-in: 5′-AAACACGAACCTATACTGACTAATAGG-3′, and HIS-in: 5′-TCTACAAAAGCCCTCCTACCCATG-3′.

Plasmid constructs

To express PUN1 from its endogenous promoter, the entire ORF and its flanking regions including 450 nt upstream of the ATG and 175 nt downstream of the stop codon, were PCR-amplified from FY1679 genomic DNA using the forward primer PUN1-fwd: 5′-TAGAGCTCAACTGTCACAGCCTCCCACTTGACC-3′ and reverse primer PUN1-rev: 5′-ATGTCGACAAGAAGATCATGCAACATCACC-3′. The PCR product was cloned via its introduced restriction sites SacI and SalI (underlined) into the centromeric vector YCplac22 [19], thereby creating the vector YCp-PUN1. To generate C-terminally triple HA-tagged Pun1p expressed via the strong MET25 promoter, the PUN1 coding sequence was PCR-amplified using the oligonucleotide primers PUN1-SpeI: 5′-AAACACTAGTTCGAAGGACGCTATAAGCATGAGG-3′ and PUN1-SalI: 5′-TTTCGTCGACAAATCAATGGTTTTTCCTCAATTGG-3′, and cloned via the SpeI and SalI restriction sites (underlined) into the multicopy vector YEpM351HA [20]. For expression of C-terminally triple HA-tagged Pun1p under the control of its endogenous promoter, PUN1 was PCR-amplified from FY1679 genomic DNA using the primer pair PUN1-fwd and PUN1-SalI. The PCR product was cloned via the SacI and SalI restriction sites into plasmid YEp351HA [21] and then subcloned together with the HA tag into plasmid YCplac33 [19], thereby creating the construct YCp-PUN1-HA. To generate the centromeric plasmid pUG35-PUN1-GFP expressing C-terminally GFP-tagged Pun1p under the control of the MET25 promoter, the PUN1 coding sequence was PCR-amplified using the oligonucleotide primers PUN1-SpeI (see above) and PUN1-EcoRI: 5′-TTTCAGAATTCAATCAATGGTTTTTCCTCAATTGG-3′ and cloned into the SpeI- and EcoRI-digested vector pUG35 [22].

Intracellular AsCl3 measurement

Strains FY1679 and MApun1∆ carrying the empty plasmid YEpM351HA and MApun1∆ overexpressing C-terminally HA-tagged PUN1 from plasmid YEpM-PUN1-HA (MApun1∆(PUN1)) were adjusted to an OD600 = 0.5 and grown in SD medium and SD medium supplemented with 250 μM AsCl3 for 6 h at 28 °C. Cells were collected and washed twice with Fluka highly pure deionized water (Sigma-Aldrich), OD600 and the dry weight were defined and the total intracellular As3+ concentration of whole cells (μg As3+/g cells) was measured by inductively coupled plasma mass spectrometry (ICP-MS, ARC Seibersdorf Research GmbH, Austria).

Fluorescence microscopy

Yeast cells used for analysis of GFP fusion proteins and for cell wall staining were grown to early log phase. Fluorescence microscopy was used to detect GFP and fluorescent dyes and differential interference contrast (DIC) optics were used to observe cell morphology. For Aniline Blue staining cells were washed with distilled water and incubated in 0.05% Aniline Blue for 5 min. For chitin staining washed cells were incubated in 1 mg/ml Calcofluor White for 5 min, washed twice and then observed. Concanavalin A–FITC staining was carried out by incubating washed cells in 0.1 mg/ml concanavalin A–FITC in 10 mM sodium phosphate buffer, pH 7.2, 150 mM NaCl for 10 min. For Filipin staining of plasma membrane ergosterol cells were washed with 50 mM potassium phosphate buffer, pH 5.5, stained with 5 μg/ml Filipin for 5 min and washed twice in the same buffer. Aniline Blue, Calcofluor White, concanavalin A–FITC, and Filipin were purchased from Sigma-Aldrich (Schnelldorf, Germany).

Fluorescence was visualized in living cells without fixation. Images were captured with equivalent exposures using a Zeiss Axioplan 2 fluorescence microscope with an AxioCam MRc5 CCD camera using AxioVision 4.8.1 software (Carl Zeiss, Oberkochen, Germany). Grayscale images were processed with Photoshop CS3 (Adobe, San Jose, CA).

Electron microscopy

FY1679 wild-type and MApun1∆ cells bearing the empty plasmid YCplac22, as well as MApun1∆ cells expressing Pun1p from plasmid YCp-PUN1 were grown in YPD to log phase (OD600 = 1). For cryofixation, cell pellets were introduced to flat sample holders of an EMPACT high-pressure freezer (LEICA Microsystems, Austria). Flat specimen holders were placed in a sample holder pod and tightly sealed. The samples were frozen as described previously for specimens of mammalian tissues [23]. Following freezing, the flat specimen holders were transferred under liquid nitrogen to an automatic freeze substitution unit (AFS; LEICA Microsystems). Freeze-substitution was performed as described elsewhere [24]. Thin sections were cut with an Ultracut S ultramicrotome (LEICA Microsystems), mounted on copper grids with Formvar support film, counterstained with uranyl acetate and lead citrate and examined at 80 kV in a JEOL JEM-1210 electron microscope. Images were acquired using a digital camera (Morada) for the wide-angle port of the TEM and analySIS FIVE software (Soft Image System).

Zymolyase sensitivity assay

FY1679 wild-type and MApun1∆ cells were grown to early log phase in YPD at 28 °C. Cells were harvested, washed, and diluted to an equal OD600 value in 10 mM Tris/Cl pH 7.5 supplemented with 0, 30 or 100 μg/ml zymolyase 20T (Seikagaku Corporation, Tokyo, Japan). Cells were incubated at 28 °C and the optical density was monitored photometrically at indicated time points by the use of a HITACHI U-2000 photometer.

For immunoblotting experiments, early log phase cells were washed, concentrated by centrifugation (OD600 = 10), resuspended in 10 mM Tris/Cl pH 7.5 and treated for 30 min with identical zymolyase concentrations as described above. Equal amounts of supernatants were subjected to SDS-PAGE and immunodetected with an antibody directed against cytoplasmic hexokinase 1 (Hxk1), which was released due to cell lysis.

Yeast cell extracts and immunoblotting

Cells were harvested and washed with distilled water. For lysis, equal cell amounts were incubated in 2 N NaOH and 1.25% β-mercaptoethanol for 10 min on ice. Proteins were precipitated with TCA (28% final concentration) for 15 min. Subsequent washings of the precipitate were performed using 90% acetone. Protein extracts were dissolved by boiling in SDS-loading buffer, separated on a 12% SDS-polyacrylamide gel, transferred to a PVDF membrane, and immunodetected.

The antibodies used in this study were mouse anti-HA (laboratory stock), mouse anti-GFP (Roche, Vienna, Austria), rabbit anti-Hxk1p (Biotrend, Cologne, Germany), rabbit anti-Pdr5p (generous gift of Karl Kuchler), mouse anti-Prc1p (Invitrogen), mouse anti-Por1p (Molecular Probes, Eugene, Oregon) and horseradish peroxidase-conjugated goat anti-mouse IgG and goat anti-rabbit IgG (Promega, Mannheim, Germany). Immunoblotting was performed in TBS-TWEEN plus 2.5% dry milk with the desired antibodies. The proteins were visualized by use of the SuperSignal West Pico system (Pierce, Rockford, Illinois).

Northern blot analysis

Overnight cultures (BY4741 wild-type, BYhog1∆, BYcrz1∆, BYrlm1∆, BYste12∆, BYsko1∆, BYdig1∆, W303 wild-type, and W303msn2/4∆) were grown in fresh YPD from OD600 = 0.15 to OD600 = 1.0 and then incubated with metal ions (MnCl2, AgNO3, HgCl2, CdCl2, AsCl3, NaCl; concentrations as indicated in the figures) for 30 min at 28 °C. Cells were harvested and total RNA was extracted as described previously [25]. Total RNA (20 μg) was separated on a formaldehyde gel; RNA was blotted onto a hybond membrane over night and UV-crosslinked. Probes were generated by PCR from chromosomal DNA using oligonucleotide primers PUN1-Nfwd: 5′-GGAGCACTTATATTAGCCATTGTTGCATGCGCGTCGACAACC-3′ and PUN1-Nrev: 5′-GATGACGGTGGGCCTCGATCTCACGCTACACCCCACTAGTGGATC-3′, as well as ACT1-fwd: 5′-ACCAAGAGAGGTATCTTGACTTTACG-3′ and ACT1-rev: 5′-GACATCGACATCACACTTCATGATGG-3′. Radioactive labeling was performed with (α-32P) dATP by random primed DNA labeling in the course of PCR. Blots were analyzed with an Amersham Biosciences Typhoon 8600 phosphor imaging system. Quantification of PUN1 levels relative to ACT1 was performed with ImageQuant 5.1 software (Molecular Dynamics).

Membrane fractionation

Yeast strains FY1679 expressing C-terminally HA-tagged Pun1p from plasmid YEpM-PUN1-HA and BY4741 expressing chromosomally GFP-tagged Pun1p either bearing the empty plasmid YCplac33 or coexpressing C-terminally HA-tagged Pun1p from plasmid YCp-PUN1-HA were grown in SD medium at 28 °C to mid-log phase. The washed cell pellet was resuspended in solution A (0.1 M Tris/SO4 pH 9.4; 10 mM DTT) and incubated for 10 min at 28 °C with shaking. Cells were spheroplasted and homogenized as described [21]. The membrane pellet was resuspended in TEDG buffer (10 mM Tris/Cl pH 7.5; 0.2 mM EDTA; 0.2 mM DTT; 20% glycerol) and gently layered on top of discontinuous sucrose gradients (25/30/35/40/45/50/60% and 30/43/53% sucrose in TED buffer (10 mM Tris/Cl pH 7.5; 0.2 mM EDTA; 0.2 mM DTT)). The gradients were centrifuged in a SW40Ti rotor for 2 h at 100,000g. Interphase fractions were collected, diluted fourfold with ice cold water, and again centrifuged at 30,000g for 20 min. The pellet was resuspended in buffer T (10 mM Tris/Cl pH 7.4), and the protein concentration was determined using Bio-Rad protein assay reagent according to the manufacturer's instructions.

Chemical cross-linking

Equal protein amounts (10 μg) of the 43% sucrose membrane fraction were supplemented with 20 mM Hepes buffer (pH 7.5) and the cysteine-specific cross-linking reagent ortho-phenyldimaleimide (o-PDM, Sigma-Aldrich) at increasing concentrations (0, 0.3, 3, 30, and 300 μM) in a total reaction volume of 50 μl. Probes were incubated for 15 min on ice, and the reaction was quenched by adding 5 μl of 100 mM N-ethylmaleimic acid (10 min on ice). Samples were supplemented with SDS-loading buffer, heated (5 min, 65 °C) and loaded on a 12% SDS-polyacrylamide gel, followed by immunodetection against HA-tagged Pun1p.

Blue native electrophoresis

Equal protein amounts (30 μg) of the same sucrose membrane fraction used for chemical cross-linking were supplemented with 50 μl extraction buffer (750 mM aminocaproic acid; 50 mM Bis–tris/Cl pH 7.0) and with digitonin (D), Triton X-100 (TX) or n-Dodecyl-β-d-maltoside (DM) at final concentrations indicated in the figure. After incubation on ice for 30 min, the samples were centrifuged at 45,000g for 30 min, and the supernatant was supplemented with a 0.25 volume of sample buffer (500 mM aminocaproic acid, 5% Serva blue G). Solubilized proteins were analyzed by BN-PAGE on a 5–18% linear polyacrylamide gradient [26]. Proteins were transferred to a PVDF membrane followed by immunodetection against HA-tagged Pun1p. The calibration standards (Amersham) used in the BN-PAGE were bovine thyroglobulin (669 kDa), horse spleen apoferritin (440 kDa), bovine liver catalase (232 kDa), bovine heart lactate dehydrogenase (140 kDa) and bovine serum albumin monomer (67 kDa).

Co-immunoprecipitation

Membrane fractions (43% sucrose, 100 μg total protein) of strain BY4741 expressing chromosomally GFP-tagged Pun1p, either coexpressing HA-tagged Pun1p from plasmid YCp-PUN1-HA or bearing the empty plasmid YCplac33, were resuspended in 10 mM Tris⁄Cl pH 7.4 and membrane proteins were solubilized by addition of TX to a final concentration of 0.1% and incubation for 30 min on ice. Samples were centrifuged at 43,000g for 30 min at 4 °C to remove nonsolubilized membrane debris. One hundred microliters of Protein A Dynabeads (Invitrogen) was washed with 10 mM Tris/Cl, pH 7.4 containing 0.1% TX. Coating of the beads was performed with a mouse anti-HA (laboratory stock) or mouse anti-GFP (Roche) antibody in the same buffer for 1 h at 4 °C under rotation. Antibody coated beads were washed twice and incubated with the clarified supernatant for 1 h at 4 °C under gentle rotation. After the binding reaction, the supernatant was removed, and the beads were washed three times with the same buffer mentioned above. Proteins were eluted from the beads by heating for 5 min at 80 °C in SDS sample buffer. The supernatants and the elution fractions were analyzed by SDS-PAGE and Western blotting was performed as described above.

Results

PUN1 is induced under heavy metal treatment of yeast cells

The transcriptional regulation of PUN1 in response to heavy metals was investigated by Northern blot analysis. To induce metal ion specific expression of PUN1, cells were treated with different concentrations of MnCl2, AgNO3, HgCl2, CdCl2, AsCl3, and NaCl prior to total RNA extraction. Compared to unstressed control cells, treatment with 1 or 2 mM MnCl2 increased PUN1 transcript level by a factor of 1.7 and 5.0, respectively. Treatment with 100 μM AgNO3 raised the PUN1 mRNA about 3-fold whereas 50 μM HgCl2 and 10 μM CdCl2 doubled the expression relative to control cells (Fig. 1, left panel). In contrast, treatment with 500 μM AsCl3 or 500 mM NaCl did not significantly change PUN1 transcript level (Fig. 1, right panel).

Fig. 1

Metal ion stress-dependent expression of PUN1. Induction of PUN1 by metal ion stress. Logarithmically grown BY4741 wild-type cells were induced by different concentrations of metal ions for 30 min as indicated in the figure. RNA samples were separated on a formaldehyde gel, blotted and incubated with radio labeled probes against PUN1 and ACT1, used as a loading control. Intensities of Northern blot signals were quantified by ImageQuant 5.1 software and normalized to untreated cells and the actin control.

Deletion of PUN1 changes sensitivity to metal ions

For investigation of PUN1 mediated phenotypic effects, the gene was deleted in strain FY1679 resulting in strain MApun1∆ (see Materials and methods). Both mutant strain MApun1∆ and its isogenic wild-type FY1679 were grown in YPD and YPD containing 1 mM MnCl2, 1.5 mM AsCl3, 1 mM NiSO4, and 25 mM CaCl2 at 28 °C. The growth rate of strain MApun1∆ was slightly decreased even without metal ions (Fig. 2A). The addition of manganese, arsenic, nickel, and calcium significantly reduced growth of the MApun1∆ mutant. Notably, in the presence of calcium pun1∆ cells reached a higher density than wild-type cells after 24 h (Fig. 2A). Other metal ions tested, like Al3+, Hg2+, Na+, or Cd2+ similarly reduced growth of the mutant strain (data not shown). To confirm that the mutant phenotype was caused by deletion of PUN1, the gene was reintroduced into the pun1∆ mutant strain. Expressing PUN1 from the centromeric plasmid YCplac22 in MApun1∆ cells restored growth on YPD plates supplemented with metal ions (Fig. 2B).

Fig. 2

Complementation and phenotyping of the pun1 deletion. A. Strains FY1679 wild-type and MApun1∆ were grown in YPD and YPD supplemented with 1 mM MnCl2, 1.5 mM AsCl3, 1 mM NiSO4, and 25 mM CaCl2 at 28 °C for 24 h under shaking. Growth was followed by measuring OD600. Growth curves shown represent an average of three individual measurements. MApun1∆ cells exhibited increased metal ion sensitivity compared to their isogenic wild-type. B. Complementation of the pun1∆ growth phenotype. Serial 10-fold dilutions of strains FY1679, carrying the empty plasmid YCplac22 and MApun1∆ carrying the empty plasmid or expressing PUN1 under the control of its endogenous promoter (YCp-PUN1), were spotted on YPD plates with and without metal ions (concentrations as indicated in the figure), and incubated at 28 °C for 48 h. Plasmid-based expression of PUN1 restored metal ion sensitivity of pun1∆ mutant cells. C. Measurement of intracellular AsCl3 concentration. Strains FY1679 and MApun1∆ carrying the empty plasmid YEpM351HA and MApun1∆ overexpressing C-terminally HA-tagged PUN1 from plasmid YEpM-PUN1-HA (MApun1∆(PUN1)) were grown in SD medium containing 250 μM AsCl3 and SD medium (w/o AsCl3) for 6 h at 28 °C. Cells were washed, OD600 and dry weight were defined and the total intracellular As3+ concentration of whole cells (μg As3+/g cells) was determined by ICP-MS (ARC Seibersdorf research GmbH, Austria). Values shown represent an average of three individual measurements. Absence of PUN1 caused a highly increased intracellular As3+ concentration.

Because pun1∆ mutant cells exhibited increased sensitivity to several metal ions and not only to those causing transcriptional induction of PUN1, these data rather suggest a more general defect in maintaining ion homeostasis and not a specific transport defect. Therefore, we determined the total intracellular As3+ concentration in wild-type, pun1∆, and YEpM351-PUN1 overexpressing cells after treatment of cells in liquid medium containing AsCl3, one of the ions which did not cause transcriptional induction of PUN1. The cells were treated for 6 h with 250 μM AsCl3 and prepared for ICP-MS analysis (see Materials and methods). Cells were washed and OD600 and the dry weight were used as a standard to calculate the total intracellular As3+ concentration. As shown in Fig. 2C, pun1∆ cells incorporated 5.7 times more As3+ than wild-type cells and about 3.5 times more than PUN1 overexpressing cells, supporting our idea of an increased ion permeability of pun1∆ cells.

Subcellular localization of Pun1p

To further characterize the mode of function of Pun1p, the subcellular localization of the protein was determined by fluorescence microscopy, using C-terminally GFP-tagged Pun1p. Fluorescence was visualized in living cells without fixation in wild-type cells harboring pUG35-PUN1-GFP growing in early exponential phase. Fluorescence of GFP-tagged Pun1p appeared in punctate patches at the cell periphery (Fig. 3). Previously Pun1p was found to colocalize with Sur7p [15], a plasma membrane protein localized to so called MCC compartments, ergosterol-rich domains involved in turnover regulation of transport proteins [4]. To confirm MCC localization of Pun1p, the same strain was treated with Filipin, a fluorescent polyene antibiotic which stains ergosterol (Fig. 3, false colored red). Indeed, fluorescence microscopy clearly revealed colocalization of Pun1p-GFP and MCC patches.

Fig. 3

Subcellular localization of Pun1p. Plasma membrane ergosterol patches colocalize with Pun1p-GFP. Logarithmically grown FY1679 cells expressing C-terminally GFP-tagged Pun1p from plasmid pUG35-PUN1-GFP were treated with Filipin, which stains plasma membrane ergosterol (false colored red). The merged images demonstrate the high degree of overlap of the ergosterol patches and Pun1p-GFP. DIC, differential interference contrast.

Stress induced accumulation of Pun1p at the plasma membrane

To determine the subcellular localization of Pun1p upon metal or osmotic stress, we used a strain with the chromosomally GFP-tagged gene. Interestingly, fluorescence of the chromosomally tagged version was hardly detectable when cells were grown under standard conditions (Fig. 4A, panel a and f). The signal significantly increased when cells were stressed for 60 and 120 min with 200 mM CaCl2 or 2 mM MnCl2 (Fig. 4A, panel b, c, g, and h). This corresponds to PUN1 transcriptional activation shown above (Fig. 1). Incubation of cells with 0.8 M NaCl did not significantly increase the fluorescence signal (panel d and i), thus excluding osmotic stress as a trigger. Furthermore, treatment of cells with 1 mM AsCl3 also failed to intensify fluorescence (panel e and j). These observations were also confirmed by detection of Pun1p-GFP in crude extracts after treatment of the same strain with metal ions for 2 h (Fig. 4B).

Fig. 4

Induction of PUN1 upon metal stress. A. Yeast strain BY4741, bearing a chromosomal PUN1-GFP fusion, was stressed with various metal ions as indicated in liquid YDP for 1 h (a–e) and 2 h (f–j) at 28 °C. Pun1p-GFP was highly induced upon exposure to 200 mM CaCl2 and 2 mM MnCl2. GFP-fluorescence and corresponding DIC images were obtained from living cells without fixation. B. The identical strain as shown in Fig. 4A was grown in YPD or treated with various metal ions as indicated for 2 h at 28 °C. Cells were collected and washed and total protein of whole cells was precipitated with TCA. Proteins were separated by SDS-PAGE and immunoblotted against GFP and Hexokinase-1 (Hxk1), used as a loading control. Western blot results clearly confirmed metal ion-induced expression of Pun1p-GFP. C. Influence of upstream regulatory elements on metal ion-induced transcription. Wild-type cells and mutants deleted for HOG1, CRZ1, RLM1, STE12, SKO1, DIG1, and MSN2/4 were analyzed by Northern blot for MnCl2 induced expression of PUN1. Red boxes highlight absent induction of PUN1 in strain BYcrz1∆ compared to the wild-type.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Transcriptional regulation of PUN1 upon metal ion stress

PUN1 expression has been reported to be dependent on the high osmotic response and cell wall integrity pathways in Congo red and zymolyase treated cells [27,28]. In these studies putative interaction motifs upstream of the PUN1 coding sequence were predicted, indicating the probability for participation of other signaling pathways in gene regulation. As has been shown previously, PUN1 is also regulated via the calcineurin dependent signaling pathway [29] and therefore might be dependent on the transcriptional activator Crz1p.

We investigated the influence of candidate factors on ion-mediated induction of PUN1 by Northern blot analysis (Fig. 4C). Transcript levels of PUN1 of cells treated with 2 mM MnCl2 were compared to unstressed cells. We examined PUN1 levels in cells deleted for HOG1, a protein kinase involved in osmoregulation [30] and in mutants lacking the stress-responsive transcription factors CRZ1, RLM1, SKO1, and MSN2/4 [31], as well as the mating/filamentous growth regulating factors STE12 and DIG1 [32]. Deletion of the calcineurin dependent transcription factor CRZ1 significantly impeded activation of PUN1 transcription (Fig. 4C, red boxes), indicating that manganese stress causes CRZ1 activation. However, the absence of CRZ1 did not interfere with the basal expression level of the gene (Fig. 4C, lanes 1 and 7). Cells depleted for HOG1 showed a reduced activation of PUN1 expression by a factor of 2 when stressed with MnCl2 (lanes 5 and 6), whereas the absence of RLM1 caused an increased induction of PUN1 by the same rate (lane 10). None of the other factors provided evidence for a direct influence on PUN1 gene transcription when the cells were stressed with MnCl2. Similar expression data were obtained with HgCl2 and AgNO3 treated cells (data not shown).

Deletion of PUN1 affects yeast cell wall composition

Reduced growth of pun1∆ cells even in rich medium and the microscopic observation that mutant cells tended to be smaller and more rounded shaped than wild-type cells pointed to a modified cell wall structure. The severe sensitivity exhibited by the MApun1∆ strain to metal ion treatment could be directly connected to ion transport or might be caused from changes of the cell wall composition. In addition, PUN1 was shown to be induced upon cell wall damage [33] and has promoter elements involved in regulation of cell wall integrity [27]. Therefore, we examined the cell wall morphology of the MApun1∆ strain in more detail. Principal cell wall constituents were stained by treatment with different dyes (Fig. 5A). Examination of cells stained with Aniline Blue and concanavalin A–FITC, fluorescent dyes binding to cell wall 1,3-β-d-glucan and mannoproteins, respectively, revealed that pun1∆ cells exhibited significantly weaker fluorescence compared to wild-type cells, indicating a reduced amount of these cell wall components (Fig. 5A, upper and middle panel). Analysis of cells treated with Calcofluor White, a fluorochromic dye that binds to chitin, showed normal chitin localization in the neck region of budding cells, although fluorescence again was less intense in pun1∆ cells (Fig. 5A, lower panel), probably due to a generally reduced cell wall layer.

Fig. 5

Modified cell wall composition of pun1∆ mutant cells. A. Staining of cell wall components. FY1679 wild-type and MApun1∆ cells were treated as indicated with Aniline Blue, concanavalin A–FITC, and Calcofluor White, staining cell wall 1,3-β-d-glucan, mannoproteins, and chitin, respectively. Fluorescence microscopy revealed reduced fluorescence of pun1∆ cells. B. Treatment with zymolyase. FY1679 wild-type and MApun1∆ cells were grown logarithmically in YPD medium, diluted to an equal OD600 value and treated with 0, 30, or 100 μg/ml zymolyase 20T. At time points indicated in the figure, OD600 was determined (upper panel). Equal amounts of FY1679 wild-type and MApun1∆ cells were treated for 30 min with identical zymolyase concentrations as used before. After centrifugation, equal amounts of the supernatant were immunodetected for cytoplasmic hexokinase 1 (Hxk1p), representing Hxk1p release caused by cell lysis (middle panel). Intracellular Hxk1p levels of both strains were compared by immunodetection after TCA precipitation of equal cell amounts (lower panel). C. Electron microscopic images of strains FY1679 wild-type (a-b), MApun1∆ (c and d), and MApun1∆ expressing PUN1 from plasmid YCp-PUN1. Principal cell wall constituents (GL and MP) and the plasma membrane (PM) are indicated. Images b, d, and f correspond to white frames in images a, c, and e, respectively. Plasmid-based expression of PUN1 restored the defective cell wall composition of MApun1∆ cells. PM, plasma membrane; GL, β-glucan layer; MP, mannoprotein layer. Black bars, 100 nm.

Furthermore, we monitored the sensitivity to zymolyase, an enzyme which specifically hydrolyzes 1,3-β-d-glucosidic linkages in yeast cell wall 1,3-β-d-glucans. Cells were identically grown, washed in Tris-buffer and incubated with 0, 30 and 100 μg/ml zymolyase T-20. Measurement of the optical density clearly showed rapid lysis of cells deleted for PUN1, whereas wild-type cells underwent only a slight reduction in their optical density (Fig. 5B, upper panel), indicating an abnormal cell wall 1,3-β-d-glucan content of pun1∆ cells. Overexpression of Pun1p resulted in lysis rates comparable to wild-type (data not shown). To confirm rapid lysis of pun1∆ cells, we also determined the amount of released cytoplasmic hexokinase 1 (Hxk1p) after zymolyase treatment. Equal cell amounts of both strains were treated for 30 min with identical zymolyase concentrations as used for optical density measurements. After a clarifying spin, the supernatants were subjected to SDS-PAGE and immunodetected with an antibody directed against Hxk1p. Indeed, pun1∆ cells released much more Hxk1p (Fig. 5B, middle panel), whereas both strains exhibited comparable intracellular Hxk1p levels, as shown by immunodetection of TCA precipitates of whole cells (Fig. 5B, lower panel).

Additionally, high resolution electron microscopic analysis of the cell wall ultrastructure of wild-type and pun1∆ cells indicated a modified cell wall composition of the mutant strain. MApun1∆ cells exhibited a dramatically reduced β-glucan layer (Fig. 5C, panel c and d). Furthermore, also the outer mannoprotein layer showed an aberrant composition and appeared to be less compact than in the wild-type. These changes likely caused the increased sensitivity to zymolyase (Fig. 5B). Importantly, complementation of pun1∆ with plasmid YCp-PUN1 completely restored the cell wall defect (Fig. 5C, panel e and f). Taken together, these findings clearly indicate an involvement of Pun1p in correct cell wall composition.

Pun1p builds high molecular weight complexes

Pun1p and claudins of higher eukaryotes share structural similarities, especially four predicted transmembrane domains and the conserved cysteine motif G-L-W-x-x-C-x(8-10)-C within the first extracellular loop. Although their molecular architecture is currently unknown, mammalian claudins were shown to form homo-oligomers [34].

To test the oligomerization capacity of Pun1p, sucrose membrane fractions of HA-tagged Pun1p were prepared and analyzed by chemical cross-linking, SDS-PAGE, and immunoblotting against the HA epitope of Pun1p. The addition of increasing concentrations of the cysteine-specific cross-linking reagent o-phenylenedimaleimide (o-PDM) produced complexes corresponding to putative dimers, tetramers, and hexamers of the protein (Fig. 6A, black arrows). Furthermore, small amounts of dimers and tetramers were also observed in the absence of o-PDM (first lane), implying a robust protein interaction. Interestingly, signals appeared as double bands, suggesting possible protein modifications.

Fig. 6

Oligomerization of Pun1p. A. Chemical cross-linking of Pun1p-HA. Membranes of FY1679 cells expressing plasmid YEpM-PUN1-HA were separated by ultracentrifugation on a discontinuous sucrose gradient (see Materials and methods). Equal amounts of the membrane extract were treated with increasing concentrations (0–300 μM) of the cross-linking reagent o-PDM. Proteins were separated by SDS-PAGE and immunoblotted against the HA epitope. Black arrows point to the Pun1p-HA monomer and higher-order oligomers (dimer to hexamer from bottom to top). o-PDM, o-phenylenedimaleimide. B. Blue native electrophoresis of Pun1p-HA. Equal amounts of the same membrane fraction used for chemical cross-linking were solubilized with the indicated detergents, applied to a native gradient gel and immunoblotted against the HA epitope. Arrowheads indicate putative oligomeric states of Pun1p-HA (dimer to dodecamer from bottom to top). Solid asterisk indicates an undefined protein complex containing Pun1p-HA. D, digitonin; TX, Triton X-100; DM, n-Dodecyl-β-d-maltoside; M, protein ladder. C. Co-immunoprecipitation (Co-IP) of Pun1p. Membrane fractions of strain BY4741 expressing chromosomally GFP-tagged Pun1p, either coexpressing HA-tagged Pun1p from plasmid YCp-PUN1-HA (lanes in blot area 1 and 3) or bearing the empty plasmid YCplac33 (lanes in blot area 2 and 4), were solubilized using 0.1% TX and incubated with anti-HA serum-coated (HA Co-IP) or anti-GFP serum-coated (GFP Co-IP) beads for 1 h at 4 °C. Unbound (supernatant, SN) and bound (elution, E) fractions were separated by SDS-PAGE and analyzed by immunoblotting with antibodies directed against HA (upper panel) or GFP (lower panel).

Additionally, identical membrane fractions were analyzed by Blue native electrophoresis for the formation of high molecular weight complexes. Pun1p-HA was solubilized with different concentrations of the detergents digitonin (D), Triton X-100 (TX), and n-Dodecyl-β-d-maltoside (DM) and detected by immunoblotting against HA. As shown in Fig. 6B, higher-order oligomers appeared, corresponding to di-, tetra-, hexa-, octa-, deca-, and dodecamers of the 35 kDa protein (indicated by arrowheads). Interestingly, we also detected a possible Pun1p-HA containing super-complex of unknown composition (solid asterisk). To our knowledge, these results demonstrate for the first time oligomeric complexes of a MCC protein.

To test for self-interaction of Pun1p, we performed Co-Immunoprecipitation (Co-IP) experiments. Sucrose membrane fractions of strain BY4741 expressing chromosomally GFP-tagged Pun1p, either coexpressing HA-tagged Pun1p from plasmid YCp-PUN1-HA or bearing the empty plasmid YCplac33, were solubilized and incubated with anti-HA serum-coated or anti-GFP serum-coated beads. Upon coexpression of GFP- and HA-tagged Pun1p, both proteins were detected in the anti-HA and anti-GFP immunoprecipitate (Fig. 6C, elution fractions, blot area 1 and 3). In the control experiments using membranes from cells expressing only GFP-tagged Pun1p, the protein was found exclusively unbound after incubation with HA-coated beads (Fig. 6C, supernatant fraction, blot area 2) as well as bound to GFP-coated beads (Fig. 6C, elution fraction, blot area 4). These results confirm self-interaction of Pun1p and suggest the protein complexes detected after cross-linking and Blue native electrophoresis to represent homo-oligomers of Pun1p.

Discussion

In this study we show that Pun1p is a calcineurin/Crz1p-regulated, metal ion stress-induced protein involved in correct cell wall organization, and a component of high molecular weight protein complexes. Pun1p has a tetraspan topology and contains the highly conserved claudin family signature G-L-W-x-x-C-x(8-10)-C within the first extracellular loop (www.uniprot.org/uniprot/Q06991). Furthermore, Pun1p belongs to a group of nine integral membrane proteins known to be localized to the ergosterol-rich MCC, a plasma membrane compartment found to be involved in endocytic regulation of transport proteins [4].

Deletion of PUN1 causes metal ion sensitivity

We found a strong induction of PUN1 transcription upon heavy metal treatment, whereas the pun1∆ strain exhibited a basic growth reduction and a pronounced sensitivity to several metal ions, not only to those causing transcriptional induction of PUN1. This led us to the suggestion that the growth defect in response to metal stress was caused by a general rather than a specific defect in ion homeostasis and that Pun1p may be directly involved in metal detoxification. To support this idea, we determined the total intracellular As3+ concentration of mutant cells compared to the wild-type and cells overexpressing PUN1 after AsCl3 treatment. Although the cellular concentration of As3+ was 5.7 times increased in pun1∆ cells compared to the wild-type, the multicopy expression of PUN1 did not confer As3+-resistance but only restored the As3+ content of the mutant to a level slightly increased compared to wild-type cells. Therefore, Pun1p is probably not directly involved in metal ion detoxification and the metal ion sensitivity of pun1∆ cells might rather be a consequence of a primary defect in maintaining intracellular ion homeostasis.

Crz1p mediates PUN1 expression upon metal ion stress

In a genome wide study, regulation of PUN1 expression was found to be dependent on the calmodulin/calcineurin signaling pathway [29]. To follow induction of PUN1 in vivo, we microscopically investigated chromosomally GFP-tagged Pun1p and found fluorescence to be significantly induced upon Ca2+ and Mn2+ exposure. This gave us the first hint that Crz1p, the major effector of calmodulin/calcineurin-regulated gene expression in yeast [29], might be involved in the PUN1 regulation.

Computational promoter analyses of genes induced upon cell wall stress revealed enrichment of certain motifs some of which also residing in the upstream region of PUN1 [27,33]. Here we show that Crz1p is crucial for transcriptional activation in response to Mn2+, Ag3+, and Hg2+ but not for basal gene expression. The metal ion-induced expression of PUN1 was also slightly decreased in a hog1∆ strain, but the level of reduction suggests that the HOG-pathway might be not directly linked to PUN1 gene expression, when metal ions are the cause of cellular stress. Additionally, deletion of RLM1 slightly increased PUN1 expression.

Based on these findings, induction of PUN1 upon ion stress can be partly explained. Ca2+ activates the calcineurin/Crz1p pathway, thereby directing Crz1p to the CDRE motif in the promoter region of PUN1. This seems to be a plausible step since this pathway was shown to confer high tolerance to ion stress [35]. Nevertheless, a direct regulatory interaction between Crz1p and PUN1 still has to be proven. Mn2+ has been known to be able to replace Ca2+ in the activation of calmodulin [36,37], thereby likely inducing the expression of PUN1 by mimicking the function of Ca2+. Induction of PUN1 by other metal ions can be explained by raising the intracellular Ca2+ level. This was previously shown for mammalian cells in presence of Cd2+ [38] and Hg2+ [39], and in yeast under conditions of Mg2+ deprivation [40]. Although we do not have evidence of a direct interaction between Crz1p and the PUN1 promoter, we suggest that the calcineurin/Crz1p pathway is the major inducer of PUN1 upon metal ion stress.

Pun1p is required for a normal cell wall composition

Fluorescence microscopy presented here revealed a subcellular localization of Pun1p in the ergosterol-rich plasma membrane MCC, consistent with previous results showing colocalization of Pun1p with Sur7p, which also localizes to the MCC [15]. Besides reduced growth in metal ion containing media, deletion of PUN1 also caused a morphological phenotype. Mutant cells appeared smaller than wild-type cells and were more rounded shaped. This was the first evidence that pun1∆ cells may suffer from cell wall damage. To further investigate this aspect, principal cell wall components, namely 1,3-β-d-glucan, mannoproteins, and chitin, were stained with specific dyes and analyzed by fluorescence microscopy. Indeed, MApun1∆ cells displayed weaker fluorescence compared to the wild-type, suggesting a reduced amount of these cell wall components. Furthermore, zymolyase treatment caused much faster lysis of pun1∆ cells compared to the wild-type, thereby confirming microscopical observations.

Finally, an aberrant cell wall structure was also observed by ultrastructural comparison of wild-type and pun1∆ cells. Electron microscopy revealed a drastically reduced β-glucan layer within pun1∆ cell walls and an outer mannoprotein layer of a less compact composition compared to the wild-type, together explaining the zymolyase sensitivity of mutant cells. This cell wall insufficiency is probably also the primary cause of the increased metal ion sensitivity of strain MApun1∆, although changed plasma membrane permeability cannot be excluded.

By comparison of these results with previous findings, evidence for involvement of Pun1p in cell wall assembly is growing. PUN1 was shown to be induced in five individual cell wall mutants (gas1∆, mnn9∆, fks1∆, knr4∆, and kre6∆) together with 78 other genes, combined in a so called “cell wall compensatory-dependent gene cluster” [33], and we found induced expression of PUN1 in absence of the cell wall integrity transcription factor RLM1, which regulates Hog1p-dependent PUN1 expression in response to cell wall damage [27,28]. Furthermore, the calmodulin/calcineurin pathway was shown to be responsible to induce genes involved in cell wall integrity [41]. Interestingly, a recent study identified PUN1 to be induced upon nitrogen starvation by the filamentous growth factor Mss11p and to be involved in pseudohyphal growth [11]. Facing the fact that filamentous growth is regulated by the cell wall integrity pathway and the cell wall itself [42], this suggests that absent filamentous growth in pun1∆ cells is primarily caused by a cell wall defect.

Pun1p forms higher-order oligomeric protein complexes

Since claudins of higher eukaryotes were found to oligomerize [34] and share structural similarities with Pun1p, the possibility of oligomerization of Pun1p was investigated. By using Blue native electrophoresis, chemical cross-linking with the cysteine-specific crosslinker o-PDM, and Co-IP experiments, we could demonstrate the presence of different oligomeric Pun1p complexes and confirmed self-interaction of Pun1p. Since the claudin domain within the first extracellular loop contains two cysteine residues and was shown to be involved in cell–cell contact formation in higher eukaryotes [43], we hypothesize that this conserved domain is involved in protein oligomerization by forming disulfide bonds. Interestingly, Pun1p appeared as a double band pattern, suggesting possible protein modifications. Sequence analysis using the ScanProsite program (http://ca.expasy.org/prosite/) indeed identified potential sites of protein modification, among them possible myristoylation, glycosylation, and phosphorylation sites of protein kinase C, the first component of the cell wall integrity signaling pathway discovered [6].

Conclusions

Data presented here support a model of Pun1p as being a stress-responsive protein involved in cell wall integrity, regulated by different signaling pathways dependent on the cause of cellular stress (Fig. 7). This idea is supported by a number of predicted regulatory motifs present within its promoter region [27]. Interestingly, also other proteins involved in cell wall formation were shown to be coordinately regulated, like the 1,3-β-d-glucan synthase subunit Fks2p, which is regulated by the cell wall integrity and calcineurin/Crz1p pathway [33].

Fig. 7

Comprehensive model of transcriptional induction of PUN1 upon different stress conditions.

Besides Sur7p [15], Pun1p is the second tetraspan protein localized in the MCC responsible for correct cell wall organization. So far, it is not clear how Pun1p contributes to cell wall integrity, especially under conditions of environmental stress. By oligomerization Pun1p may function as a scaffold for cell wall assembly or for recruitment of cell wall building proteins. Interestingly, in a recent genome wide study the 1,3-β-d-glucan synthase subunit Fks1p was found in a complex with Pil1p [44], a component of eisosomes [3], being sub-cortical protein assemblies which show overlapping localization with MCCs. Alternatively, absence of Pun1p may interfere with stress signaling and thereby cause disturbed cell wall integrity and adaptation to stress conditions.

Finally, mammalian claudins regulate paracellular ion transport between epithelial cells by the formation of charge-selective channels, and the charged residues in their first extracellular loop are believed to represent a selectivity filter [45]. Interestingly, Pun1p also contains a series of charged residues in its first loop. However, yeast is a single cellular organism without paracellular transport and currently we do not have evidence that Pun1p is involved in any form of ion transport. Nonetheless, results presented in this study relate the plasma membrane MCC with cell wall integrity and metal ion stress.