Phytochelatin synthase is required for tolerating metal toxicity in a basidiomycete yeast and is a conserved factor involved in metal homeostasis in fungi

Phytochelatin synthase (PCS) is an enzyme that catalyzes the biosynthesis of phytochelatin from glutathione. Phytochelatins protect cells against the toxic effects of non-essential heavy metals, such as cadmium, and hence growth is restricted in the presence of these metals in mutants in PCS-encoding genes. PCS genes from fungi have been characterized in only two species in the Ascomycota, and these genes are considered sparsely distributed in the fungal kingdom. A gene encoding a putative PCS was identified in Sporobolomyces sp. strain IAM 13481, a fungus that is a member of the Pucciniomycotina subphylum of the Basidiomycota. The function of this PCS1 gene was assessed by heterologous expression in the Ascomycota yeasts Saccharomyces cerevisiae and Schizosaccharomyces pombe, and by mutating the gene in Sporobolomyces. The gene is required for tolerance to toxic concentrations of non-essential cadmium as well as the essential metal copper. Pcs1 homologs in fungi and other eukaryotes have putative targeting sequences for mitochondrial localization: the S. pombe homolog was fused to green fluorescent protein and it co-localized with a mitochondrial dye. Evaluation of the presence or absence of PCS and PCS-like homologs in the genome sequences of fungi indicates that they have a wide distribution, and the absence in most Ascomycota and Basidiomycota (the Dikarya) species can be explained by a small number of gene losses. The ecology of the species within the fungi carrying putative PCS genes, the phenotypes of phytochelatin synthase mutants in two major fungal lineages, and the presence of homologs in many non-Dikarya lineages parallel what is seen in the plant and animal kingdoms. That is, PCS is a protein present early during the evolution of the fungi and whose role is not solely dedicated to combating toxic concentrations of non-essential metals.


Background
Metal ions are cofactors that are required for many enzymes and transcriptional regulators, yet these same essential metals can become toxic at high concentrations, and non-essential metals are often toxic [1]. Organisms have different mechanisms to tolerate or combat exposure to high levels of metals. Understanding those mechanisms and then employing this information has many potential benefits. Amongst them, the knowledge can aid in bioremediation processes of environmental sites contaminated with metals, in modifying organism properties, or in using metals as antimicrobial agents.
One approach used by organisms to avoid metal toxicity is to chelate metal ions with another molecule to sequester the metal within the cell. Phytochelatins are small peptides that bind heavy metals. A gene encoding a phytochelatin synthase was identified in the late 1990s in three independent studies. In one, a wheat cDNA was cloned that conferred resistance to cadmium when expressed in budding yeast Saccharomyces cerevisiae [2]. Deletion of the homologous gene that was identified in the fission yeast Schizosaccharomyces pombe rendered the mutant sensitive to cadmium. In a second study, the phytochelatin synthase was identified by map-based cloning in Arabidopsis thaliana [3]. A deletion strain for the gene was also made in S. pombe, resulting in a strain that exhibited sensitivity to arsenic and cadmium. In the third study, the A. thaliana gene was identified by its expression in an S. cerevisiae yap1Δ strain through its ability to confer resistance to cadmium [4].
With a clear role in protecting against the harmful effects of metals, the original research and much subsequent research focused on the roles phytochelatin synthase have in non-essential metal toxicity, including the potential manipulation of the genes for bioremediation purposes. However, dealing with non-essential metals is unlikely to be the sole role of the phytochelatin synthases that have been identified, because they are found in organisms that are not exposed to those metal ions at high concentrations, a puzzle that has been noted previously [5,6]. More recently, a role in the homeostasis of essential metal ions has been defined because one phenotype of loss of PCS activity is zinc sensitivity in A. thaliana and in S. pombe in a Zn-homeostasis mutant background [7].
The fungi share many of the same habitats with plants, and, like plants, as individual organisms they have restricted capabilities to move location in response to environmental conditions. As addressed above, the analysis of the genome sequence of S. pombe revealed a homolog of phytochelatin synthase. Other homologs were not subsequently characterized in the fungi until recently when a PCS1 homolog was identified in the genome sequence of the truffle-forming species Tuber melanosporum and noted in two other species of Schizosaccharomyces [8]. The fungi are a group of more than a million species [9,10]. The sparse distribution of PCS reported in the fungi led to the hypothesis that the Schizosaccharomyces PCS genes may have been acquired in those species by a horizontal gene transfer event [5].
This current research identified a putative phytochelatin synthase gene in a basidiomycete and aimed to test if the gene encoded a functional phytochelatin synthase. Analysis of genome sequences of fungi reveals a presence of PCS in many other fungal lineages outside of the Ascomycota and Basidiomycota, indicating a far wider distribution of this gene in the fungi than previously appreciated.

Results
Sporobolomyces sp. encodes a putative phytochelatin synthase homolog that can protect S. cerevisiae from toxic levels of heavy metals A putative phytochelatin synthase (PCS) homolog was identified during the analysis and annotation of the genome sequence of a strain of Sporobolomyces. The rarity of this gene in reports from fungi triggered an investigation into the function of this gene in Sporobolomyces, and then the evolution of PCS homologs in the fungi.
One approach used to clone phytochelatin synthases from plants was to express plant cDNAs in a S. cerevisiae strain with a yap1 mutation and identify cadmiumresistant transformants. Yap1 is a transcription factor that controls the expression of genes required for protection against oxidative and other stresses, such as cadmium [11,12]. Expression of the PCS-encoding genes increased the resistance of this mutant to toxic metals. In an equivalent experiment, the Sporobolomyces PCS1 cDNA was cloned into an S. cerevisiae plasmid that enables transcript induction in response to galactose in the medium. Wild type and yap1 mutant strains containing this plasmid and the control empty plasmid were cultured overnight in media containing glucose or galactose, and plated onto medium supplemented with cadmium. The presence of the Sporobolomyces gene increased the resistance of both wild type and yap1 mutant strains to cadmium ( Figure 1). This also occurred under the noninducing conditions in which there is residual expression from the GAL1 promoter [13], indicating that even a small amount of PCS is able to combat cadmium toxicity. These findings are consistent with a role of the Sporobolomyces gene in biosynthesis of phytochelatin.
The Sporobolomyces pcs1 mutant is sensitive to high metal concentrations Heterologous expression of the PCS1 homolog of Sporobolomyces supported the hypothesis that this gene can synthesize protective molecules in fungi. In order to test the function of the gene directly in Sporobolomyces, a gene replacement strain was constructed ( Figure 2A). Regions flanking the PCS1 gene were amplified by PCR and used to generate an overlap product that was transformed into a Sporobolomyces uracil auxotroph. The mutant strain was identified by PCR analysis (Figure 2B), and gene replacement confirmed by Southern blot analysis ( Figure 2C). As a control, a complemented strain was generated to include with phenotypic comparisons. For this purpose, the URA5 gene was "recycled" for use as a transformation marker, by selection on 5-FOA medium for strains with spontaneous mutations. A wild type copy of PCS1 was fused adjacent to the wild type copy of the URA5 gene by overlap PCR, and this construct transformed into the pcs1Δ ura5 strain to produce the complemented control strain.
Different metals were tested for their effects on the wild type, pcs1 deletion mutant and complemented strains. Greatest sensitivity in the deletion mutant was found for cadmium and copper ( Figure 2D). A slight decrease in growth in high levels of zinc was also observed. In contrast, no major difference was observed for toxic levels of cobalt chloride, sodium arsenite, iron sulfate, manganese chloride, or reactive oxygen species H 2 O 2 and t-butyl-hydroperoxide.
The transcript levels of the PCS1 gene in Sporobolomyces were examined by northern blot for the wild type strain cultured overnight in YPD alone or in the presence of three metals: copper, cadmium and zinc. No difference in transcript level was noted between any of these treatments or with the strain grown in YPD medium (data not shown). The regulation of phytochelatin synthase activity levels varies between species, although in many the transcript and protein are expressed constitutively with an induction of enzyme activity by direct interaction with the metal [14,15], and presumably a similar mechanism occurs in Sporobolomyces.
Sporobolomyces sp. PCS1 complements the phenotypes associated with loss of phytochelatin synthase from the ascomycete S. pombe To investigate the functions of PCS1 more widely in the fungi and if the Sporobolomyces gene could complement the loss of PCS function, S. pombe was used. First, a gene mutant (pcs1Δ) was constructed in S. pombe that replaced the pcs1 gene with the KanMX construct. The cDNA of Sporobolomyces and the ORF of S. pombe were amplified and cloned into a plasmid that enables a low level of expression of these genes when thiamine is present in media as used here [16]. The plasmids were transformed into the S. pombe deletion mutant, and growth tested on media containing cadmium ( Figure 3A). The S. pombe pcs1 mutant was sensitive to cadmium, as reported previously. This sensitivity was complemented by addition of the wild type copy of the gene from either S. pombe or Sporobolomyces into the mutant strain.
Testing growth under other metal conditions revealed that the S. pombe pcs1 mutant also has phenotypes in the presence of copper. That is, there was increased sensitivity to growth on copper. In addition, at lower concentrations the wild type strain became pigmented, while the pcs1 mutant strain had less pigmentation, indicating a change in copper homeostasis ( Figure 3B).

Phytochelatin synthase is localized in the mitochondria in S. pombe
Where PCS functions within the cell is largely unexplored, but one prediction is that it can be localized to the mitochondria [17,18]. The amino acid sequences of the Sporobolomyces and S. pombe homologs were examined with two software programs (Psort II and MitoProt) that predict subcellular localization, and both predicted mitochondrial localizations because of the presence of a putative mitochondrial targeting sequence at the N-terminus of the proteins.
To test its localization, fusions were created to produce proteins with Pcs1 at the N-terminus and GFP at the C-terminus, using both the Sporobolomyces and S. pombe homologs. The Sporobolomyces Pcs1-GFP fusion construct used the native promoter, and was transformed into the pcs1Δ strain. While the construct could complement the cadmium-sensitivity of the mutation, no fluorescence was observed for these strains (data not shown). To date, achieving a GFP signal in Sporobolomyces has been elusive, even with the use of a modified version of GFP that contains a Sporobolomyces intron Figure 1 Heterologous expression of Sporobolomyces PCS1 in S. cerevisiae increases its resistance to cadmium. The PCS1 cDNA was expressed in a vector allowing induction in response to galactose. Wild type and the yap1 mutant strain were transformed with this plasmid or the empty control and cultured overnight in YNB with either glucose or galactose as the carbon source. Ten-fold serial dilutions were plated onto YPD medium with or without cadmium sulfate and cultured 2 d at 30°C.
that is functional in another basidiomycete yeast (i.e. Cryptococcus neoformans). Others working to express foreign genes in basidiomycetes have found that this may require optimizing constructs to alter the GC content of their DNA or include introns (see as examples [19][20][21][22]).
As an alternative strategy, the localization of the S. pombe PCS protein was explored. The S. pombe PCS1 homolog was fused to GFP by PCR, the construct cloned, and the plasmid was transformed into the S. pombe pcs1Δ strain. The fusion construct complements the cadmium sensitive phenotype ( Figure 4A). In the AS008-AS002 and AS003-AS009 were used to generate 5′ and 3′ fragments for targeted homologous replacement. Primers AS020-AS004 were used to amplify the probe used in the Southern blot, and primers AS005-AS006 used for the PCR in panel B. Restriction enzyme sites relevant to the Southern blot are provided along with the relative positions in parentheses. (B) PCR analysis using primers AS005-AS006 on the wild type and pcs1 deletion strains. (C) Southern blot analysis of gene replacement of PCS1 with the URA5 gene. Genomic DNA was digested with BamHI, EcoRI or HindIII restriction enzymes and resolved on a 1% agarose gel. 32 P-labelled DNA amplified with primers AS020-AS004 was hybridized to the blot. Note that the amplification of the probe yielded a second non-specific product, which hybridizes to the bands marked with *. (D) Phenotypes of the gene replacement strain compared to wild type and complemented (Δ + PCS1) strain. Ten-fold serial dilutions were plated onto YPD medium with or without supplements, and cultured for 3 d at 22°C. strain expressing this construct, GFP was found in punctate dots or filaments. The pattern of fluorescence co-localized with the mitochondrial dye MitoTracker ( Figure 4B), indicating that Pcs1 is targeted to the mitochondria.

Phytochelatin synthase is widely distributed in the fungi
The genome sequences of fungi were analyzed for the presence or absence of homologs of PCS. Many lineages in the ascomycetes and basidiomycetes did not encode a copy. However, homologs were found in the Pezizomycetes  and Taphriniomycotina (Ascomycota) and Pucciniomycotina (Basidiomycota) ( Table 1). For the lineages outside of the monophyletic Dikarya, the presence of a PCS was common as it was absent from only two clades examined (the Microsporidia, known for highly reduced genome sizes and gene content, and the one species of Entomophthoromycotina with an available genome sequence). The distribution is consistent with an origin of the gene early in the evolution of the fungi, and then loss on a number of occasions. Major losses in the Basidiomycota would have been in the lineages that gave rise to the Agaricomycotina and Ustilaginomycotina. The presence of homologs in Wallemia species, considered either as a sister clade or a basal member of the Agaricomycotina [23,24], could imply two losses. Another major loss was in the Ascomycota lineage that branched after the divergence of the class Pezizomycetes from the other species in the Pezizomycotina phylum. Other losses (e.g. in the Saccharomycotina or the obligate parasites) are consistent with gene fates in these lineages.
The amino acid sequences of candidate PCS proteins from fungi were downloaded and aligned prior to making phylogenetic trees. After examining alignments of the full amino acid sequences, the variable N and C terminal ends were removed due to their poor alignments.
However, both ends should be important for function. The N-terminus is likely involved in subcellular targeting in some species and/or could have a role in protein activity. The C terminus in many homologs is cysteine rich, but the positions of the residues do not align. For example, S. pombe has within its C terminus CCX 5 CCX 3 CC; Sporobolomyces has CCX 8 CXC; one Phycomyces blakesleeanus homolog has CX 2 CX 3 C; and Catenaria anguillulae has CCCX 11 CXCC (X indicates any amino acid residue).
To build additional support for an early origin for the PCS in fungi, phylogenies were constructed from the fungal homologs, and one maximum likelihood analysis is illustrated in Figure 5. The relationships between many of the homologs are poorly supported based on bootstrapping methods, which is to be expected because of the lack of resolution amongst the basal lineages of the fungi based on conserved housekeeping genes [25]. Nevertheless, distinct groups with high bootstrap support are detected, many of which arrange the PCS homologs within groups of species with monophyletic origins. Other cases were more complicated. For example, the Basidiomycota have two groups that are split by homologs from the Ascomycota, and may indicate horizontal transfer of a gene between lineages. The amino acid alignment also revealed that four candidate PCS proteins in the Mucoromycotina have amino acid substitutions in the cysteine residue within the active triad site of characterized phytochelatin synthases [26][27][28]. This draws into question whether these proteins function as enzymes or if they play some other role in these species. In the phylogeny, the four PCS-like proteins form a clade with 100% bootstrap support.
While most ascomycete and basidiomycete species have no PCS, other fungal species show an expansion in copy number. For example, two homologs are in Puccinia graminis (Basidiomycota) as a tandem gene duplication in the genome. Most of the Mucoromycotina species have more than one copy, which is similar to a general expansion in gene number through whole genome or segmental genome duplications [29,30]. The duplication dates prior to the separation of these species. Curiously Catenaria anguillulae has two putative PCS proteins that are positioned in distinct parts of the phylogeny, which would suggest an interspecies exchange event occurred during its history. The evidence for expansion in copy number would support a hypothesis that the PCS candidates provide a selective advantage to these species, rather than being a vestigial property found in the earlier lineages of the fungi.

Discussion
Balancing metal homeostasis is crucial for cells to survive and function properly. Organisms must adapt to the environmental availability of such elements, either to take up as much as needed under depleted conditions or to avoid the toxic effects of higher concentrations. Molecules that chelate or sequester metal ions play key roles in these processes.
The emphasis on the function of phytochelatin synthases is their role in protection against the toxic effects of heavy metals, especially cadmium [15,31,32]. However, the distribution of the genes encoding putative phytochelatin synthases in organisms, and the locations of these organisms in naturehere the discussion being on members of Figure 5 Phylogeny of PCS proteins from the fungi. The predicted amino acids were aligned, and a common region spanning 243 residues was compared by maximum likelihood analysis. Numbers above the nodes indicate percentage bootstrap support from 100 reiterations if 70% or higher. Species in the same lineages are color-coding where possible, and their classification is provided in Table 1. Fonticula alba is a cellular slime mold (non-fungal). the fungal kingdom, but similar points have been raised for plants and other organisms (e.g. [6,33])is not consistent with this being their primary role. Heavy metal concentrations can certainly be high in parts of the environment; for example, an estimated 140 million people drink water contaminated by detrimental concentrations of arsenic [34]. However, for many of the fungal species that encode potential phytochelatin synthases these have not been isolated from environments that have such levels of metals. For instance, the Sporobolomyces sp. strain used in this study was isolated from the leaf surface of a willow tree [35], although Pucciniomycotina yeasts have been isolated from natural and contaminated metal sites with high levels of cadmium or other metals (e.g. [36][37][38]). One hypothesis is that the phytochelatins may play a primary role in the homeostasis of essential metals, such as copper or zinc that, while essential, may reach toxic concentrations.
One example of an essential metal that has an interaction with PCS is copper. For instance, in S. pombe most focus has been on the cadmium responses mediated by PCS, but there are at least two copperdependent effects when the gene is mutated. One is on growth rate and the second is on colony color. The source of this pigmentation is unknown, but based on studies in S. cerevisiae it is likely due to precipitation of copper sulfite in the cell wall, as occurs in S. cerevisiae and where the pigmentation is influenced by cysteine and glutathione levels that are the precursors of phytochelatins [39,40].
A disadvantage of using phytochelatins to protect against metals compared to a mechanism of exclusion or secretion is that the metal remains within the cell. A cell would therefore have limited ways to avoid accumulating a potentially hazardous element within itself. An alternative hypothesis has been proposed in metal hyperaccumulating plants, that toxic metals may serve to protect those species against herbivory (for reviews, see [41,42]). A similar hypothesis is worth exploring for fungi, since certain species produce macroscopic sexual fruiting structures that are eaten and yeasts can be consumed by insects and nematodes. That said, counter to this hypothesis is that the truffle fungus Tuber melanosporum and other truffle genomes encode a phytochelatin synthase [8], yet this has not prevented them from become one of the world's most prized and expensive foods.
Phytochelatin synthase localization in the mitochondria has been proposed based on the presence of a putative targeting sequence at the N-terminal end of the homolog from helminth parasite Schistosoma mansoni [17,18]. Alternative splicing produces two isoforms that are predicted to be mitochondrial and cytoplasmic [17]. The full-length (putative mitochondrial) isoform is inactive when expressed in Escherichia coli, whereas removing the N-terminus yields an active enzyme [18].
The localization of PCS proteins in eukaryotic cells has not been explored extensively, so we examined the localization of Pcs1-GFP fusion proteins. In S. pombe, the Pcs1 protein localized to the mitochondria based on the co-localization of the Pcs1-GFP fluorescence with a standard mitochondrial dye. Analysis of other putative PCS proteins using prediction software also identifies candidate N-terminal mitochondrial targeting sequences in other homologs, from plants like A. thaliana to animals like Caenorhabditis elegans. However, computation predictions have limitations in their accuracy. As one example, the two PCS proteins in A. thaliana are distributed in the cytoplasm with no evidence of targeting to the mitochondria [43]. Alignment of the protein sequences from five Pcs1 homologs and previous characterization of the catalytic regions indicates that the predicted cleavage site to remove the putative mitochondrial targeting region lies within the conserved regions, whereas the fungal species and S. mansoni have an extension in the N-terminus that accommodates the predicted mitochondrial targeting sequences (Additional file 1: Figure S1). The presence of PCS in the mitochondria in S. pombe, and potentially other organisms, is interesting from the perspective of where in the cell phytochelatins would be synthesized, where they would be required to chelate metals, and to have an idea about the intracellular transport of the phytochelatin-metal complex into vacuoles [44].
Analysis of the distribution of candidate phytochelatin synthase in fungal genomes reveals a presence throughout many lineages. This parallels recent observations suggesting that putative phytochelatin synthases are present in several of the early land plants and in many animal lineages, and may therefore be an ancestral property in each of these eukaryote groups [45][46][47][48]. The most obvious losses in the fungi appear in the two most species-rich lineages, the ascomycetes and basidiomycetes. It is unclear why so many of these species have lost this gene, and what would have taken its place. Another protein class that is important for chelation of metals is the metallothioneins. One hypothesis is that during evolution other mechanisms have taken over the control of balancing metal concentrations in those fungi without the PCS homologs.

Conclusions
The experiments reported here and the analysis of a phytochelatin synthase gene from a basidiomycete species extends the distribution of these enzymes into a second major lineage of the fungi, and shows that they have a conserved function in protection against toxic concentrations of metals. The analysis of fungal genomes, especially those of species other than members of the ascomycetes and basidiomycetes and that are often considered "basal" in the kingdom, demonstrates the presence of homologs widely. The functions of these homologs, in addition to protection against heavy metals like cadmium, are likely in other aspects of metal homeostasis, and remain to be explored in these organisms, as does the potential application of PCS of species with the gene for instance in bioremediation of contaminated sites.

Strains
Sporobolomyces sp. strain IAM 13481 was used as the wild type. The ura5 auxotroph strain AIS2 was isolated previously [49], and was the recipient for gene replacement experiments. Saccharomyces cerevisiae strain BY4743 and the yap1::KanMX gene replacement strain in this genetic background were used as host strains for expression of the Sporobolomyces PCS1 homolog. The Schizosaccharomyces pombe strain MM72-4A is an ura4 mutant, derived from wild type L972. More details about the strains are provided in Table 2.

Expression of Sporobolomyces PCS1 in ascomycete yeasts
The cDNA of PCS1 from Sporobolomyces was amplified using primers AS005 and AS006 from cDNA that was reverse transcribed from total RNA using an oligo-dT primer with Superscript III (Invitrogen, Grand Island, NY, USA). Primers in this study were obtained from Integrated DNA Technologies (Coralville, IA, USA), and their sequences are provided in Table 3. The amplicon was digested with KpnI and XbaI restriction enzymes, and ligated into the same sites in plasmid pYES2 that enables expression in Saccharomyces cerevisiae under the control of a galactose-inducible promoter (Invitrogen). A plasmid, pAS1, was identified with no errors by DNA sequencing. The cDNA was amplified from this plasmid with primers AS016 and AS017 and cloned into pCR2.1 (Invitrogen). An insert without errors was identified by sequencing (plasmid pAS2), it was digested with NdeI, and the PCS1 fragment ligated into the NdeI site of plasmid pREP42 to form plasmid pAS3. The S. pombe PCS1 open reading frame was amplified from genomic DNA with primers AS018 and AS019 and cloned into pCR2.1. An error-free insert was identified (plasmid pAS4) and excised with BamHI and NdeI, and ligated into the BamHI-NdeI site of pREP42 to form plasmid pAS5. The pREP42 plasmid allows expression in S. pombe from a mid-strength nmt promoter [16]. The plasmids were transformed into S. cerevisiae or S. pombe uracil auxotrophic strains using the lithium acetate/PEG method, and selected on yeast nitrogen base (YNB) medium.
Targeted gene replacement of PCS1 homologs in Sporobolomyces and Schizosaccharomyces The PCS1 gene of Sporobolomyces was replaced with a URA5 selectable marker by homologous recombination. A construct with 1,334 and 1,232 bp on either side of the wild type URA5 gene was created by forming a three-part DNA fragment with overlap PCR. The flanks of the PCS1 gene were amplified with primers AS001-AS002 and AS003-AS004. The URA5 gene was amplified with primers ALID0562-ALID0564. The three fragments were mixed, and an overlap fragment amplified with primers AS001-AS004. The construct was transformed into cells of the AIS2 uracil auxotrophic strain (a ura5 mutant) plated on yeast nitrogen base agar + 1 M sorbitol, by biolistic delivery of the DNA coated onto gold beads, as described previously [49]. A strain with the targeted gene replacement was identified by PCR analysis. To complement the pcs1Δ mutation, first a spontaneous mutation in the URA5 gene of the pcs1::URA5 ura5 strain was isolated by plating on 5-fluoroorotic acid medium. The full-length PCS1 gene that includes the promoter and terminator regions was amplified from wild type DNA using primers AS004-AS020, and URA5 amplified with primers ALID0562-ALID0564. A two-fragment overlap PCR was performed using primers AS004-ALID0564. This construct was transformed into the pcs1 ura5 mutant strain by biolistic delivery of the DNA and transformants were selected on yeast nitrogen base medium. The pcs1 gene of S. pombe was also replaced by homologous recombination. A construct with 360 and 339 bp on either side of the KanMX cassette was created by overlap PCR. The 5′ flank was amplified with primers AS012 and AS013, the 3′ flank with AS014 and AS015, and the KanMX selectable marker with primers KanMX F and KanMX R. Equimolar amounts were mixed, and the overlap PCR performed with primers AS012-AS015. The product was transformed into strain MM72-4A using the lithium acetate/PEG method. Cells were recovered overnight in liquid medium (5 g/L yeast extract, 10 g/L peptone, 3% glucose), and plated as aliquots onto the same medium that included agar and containing 50 or 100 μg/L G-418. A gene replacement strain was identified by PCR analysis.

Nucleic acid manipulations
Genomic DNA was isolated using a CTAB buffer extraction protocol [50]. RNA was isolated using TRI Reagent (Sigma-Aldrich, St. Louis, MO, USA). Nucleic acids were blotted to Zeta-Probe membrane (Bio-Rad, Hercules, CA, USA), and probed with 32 P-dCTP labeled DNA fragments using standard methods. As a loading and transfer control for northern blots, a fragment of the constitutively-expressed GPD1 gene, encoding glyceraldehyde-3-phosphate dehydrogenase, was amplified with primers ALID1432-ALID1433 for use as a probe.

Subcellular localization using GFP fusions
The PCS1 gene from S. pombe was fused in frame at the C-terminal to the gene encoding green fluorescent protein. The PCS1 gene was amplified with primers AS018-ALID2092. GFP was amplified with primers AISV066-ALID2091. An overlap PCR product was amplified with primers AS018-AISV066, and this product cloned into plasmid pCR2.1. The construct was excised from the pCR2.1 plasmid with BamHI-NdeI and cloned into the BamHI-NdeI site of plasmid pREP42 to make plasmid pAS6. The plasmid was transformed into the S. pombe pcs1Δ strain AS8 with the lithium acetate/PEG method, and transformants were selected on YNB.
The yeast strains were cultured in Edinburgh Minimal Medium (MP Biomedicals, Solon, OH, USA) and stained with MitoTracker red (Invitrogen), at 3 nM, with the staining of the cells resuspended in water for 20 min. Cells were washed and then re-suspended in phosphate buffered saline. Fluorescence was examined using an Olympus Fluoview FV300 confocal microscope.

Phylogenetic, genome sequence, and other computational analyses
Putative phytochelatin synthases were sought by BLASTp queries against the MycoCosm (US Department of Energy), NCBI, and Broad Institute databases [51,52]. The last queries were made on 1 November 2014. Protein sequences were aligned with ClustalW and the alignment inspected by eye. Protein predictions were examined for unusual gaps or sequence divergences that may reflect annotation errors. Revised annotations were generated in six cases: for Pyronema confluens, Rhizophagus irregularis, Tuber melanosporum, Rhizopus oryzae, Batrachochytrium dendrobatidis, and an alternative model in the MycoCosm database selected for Phycomyces blakesleeanus. After constructing a new alignment, the N and C terminal variable ends were removed, yielding an alignment of 243 amino acids. The alignment is provided in the Additional file 2. Evolutionary relationships between PCS proteins were explored using MEGA6 [53]. The best-fit model for maximum likelihood was selected, and a phylogeny was constructed using maximum likelihood (LG + G; using all amino acid sites). 100 bootstraps were used to infer support for the nodes within the tree.
For predicting subcellular localization of Pcs1 homologs, the MitoProt and Psort II programs were used [54,55].