help button home button Genetics AJP: Heart and Circ
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Originally published as Genetics Published Articles Ahead of Print on October 16, 2004.

Genetics, Vol. 169, 107-122, January 2005, Copyright © 2005
doi:10.1534/genetics.104.035873

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
genetics.104.035873v1
169/1/107    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Lesuisse, E.
Right arrow Articles by Dancis, A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Lesuisse, E.
Right arrow Articles by Dancis, A.

Genome-Wide Screen for Genes With Effects on Distinct Iron Uptake Activities in Saccharomyces cerevisiae

Emmanuel Lesuisse*,1, Simon A. B. Knight{dagger}, Maïté Courel*, Renata Santos*, Jean-Michel Camadro* and Andrew Dancis{dagger}

* Laboratoire d'Ingéniérie des Protéines et Contrôle Métabolique, Département de Biologie des Génomes, Institut Jacques Monod, Unité Mixte de Recherche 7592 CNRS-Universités Paris 6 and 7, France
{dagger} Department of Medicine, Division of Hematology-Oncology, University of Pennsylvania, Philadelphia, Pennsylvania 19104

1 Corresponding author: Laboratoire d'Ingéniérie des Protéines et Contrôle Métabolique, Institut Jacques Monod, Tour 43, Université Paris 6 and Paris 7, 2 place Jussieu, F-75251 Paris Cedex 05, France.
E-mail: lesuisse{at}ijm.jussieu.fr

Manuscript received September 3, 2004. Accepted for publication October 8, 2004.


   ABSTRACT
TOP
 ABSTRACT
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 ACKNOWLEDGEMENTS
 LITERATURE CITED
 
We screened a collection of 4847 haploid knockout strains (EUROSCARF collection) of Saccharomyces cerevisiae for iron uptake from the siderophore ferrioxamine B (FOB). A large number of mutants showed altered uptake activities, and a few turned yellow when grown on agar plates with added FOB, indicating increased intracellular accumulation of undissociated siderophores. A subset consisting of 197 knockouts with altered uptake was examined further for regulated activities that mediate cellular uptake of iron from other siderophores or from iron salts. Hierarchical clustering analysis grouped the data according to iron sources and according to mutant categories. In the first analysis, siderophores grouped together with the exception of enterobactin, which grouped with iron salts, suggesting a reductive pathway of iron uptake for this siderophore. Mutant groupings included three categories: (i) high-FOB uptake, high reductase, low-ferrous transport; (ii) isolated high- or low-FOB transport; and (iii) induction of all activities. Mutants with statistically altered uptake activities included genes encoding proteins with predominant localization in the secretory pathway, nucleus, and mitochondria. Measurements of different iron-uptake activities in the yeast knockout collection make possible distinctions between genes with general effects on iron metabolism and those with pathway-specific effects.

RECENT advances have made possible genome-wide approaches to the biology of eukaryotes. Many genes and proteins implicated in iron metabolism have been identified in the past 10 years using classical genetic approaches of mutagenesis, phenotype screening, complementation, and homology searching. More recently, global transcriptome analysis has become possible, and the effects of iron manipulations on gene expression have been studied. Using these methods, a number of iron transporters and regulators have been discovered (KOSMAN 2003). However, many aspects are still missing from a complete picture of cellular iron metabolism. Mediators of intracellular iron trafficking, distribution, organelle transport, and processing must exist, although few have been identified. Regulators coordinating use of iron in cofactors such as heme and Fe-S also must exist and the cofactors must be made, trafficked, and delivered to target apoproteins. Most of these key components are yet to be found. Finally, essential processes such as transcription, translation, protein trafficking, organelle biogenesis, and secretion are likely to influence iron metabolism to different degrees and in different ways. A new tool, the haploid knockout collection of Saccharomyces cerevisiae, has recently become available. Each strain carries a complete deletion of a single ORF, and the collection can be manipulated and examined in microtiter well format. We wondered whether we could use this collection to discover novel genes involved in iron metabolism.

The idea underlying this screen is that yeast expresses two independent pathways by which iron can enter cells: siderophore and reductive (LESUISSE and LABBE 1989; DANCIS et al. 1990). Both pathways are regulated and responsive to iron availability (YUN et al. 2000a). Therefore, random or targeted mutations in the genome that alter basic processes involved in cellular iron metabolism will affect activities for both pathways. By contrast, mutations selectively impacting one or the other system will be due to altered expression, localization, or function of pathway-specific components.

The siderophore pathway mediates iron uptake from siderophores (YUN et al. 2000b; LESUISSE et al. 2001). Siderophores are small molecules that bind, solubilize, and chelate ferric iron in the environment with tremendous affinity. They are synthesized by a nonribosomal enzymatic process and secreted by bacteria and fungi, although not by S. cerevisiae. This yeast, however, has evolved a means for acquiring iron from siderophores made by other organisms. Plasma membrane transporters belonging to the major facilitator superfamily (MFS) are required for internalization of ferrisiderophore complexes (LESUISSE et al. 1998; HEYMANN et al. 1999; YUN et al. 2000b). Trafficking into a vesicular compartment and release of iron occurs subsequently (KIM et al. 2002). Intracellular release of iron from siderophores is probably mediated by special reductases and/or hydrolases. The details of intracellular trafficking of ferrisiderophores, iron release from ferrisiderophores, and iron distribution following release are still undefined (reviewed by HAAS 2003).

The reductive pathway mediates iron uptake from ferric chelates (reviewed by KOSMAN 2003). An externally directed plasma membrane ferric reductase activity dependent on Fre1 (DANCIS et al. 1992) and Fre2 (GEORGATSOU and ALEXANDRAKI 1994) is able to reduce ferric chelates of varying composition (including siderophores), thereby releasing ferrous iron that can be accessed by the high-affinity ferrous transport complex. The latter consists of two components, a multi-copper oxidase (Fet3) with an externally directed oxidase domain and a polytopic permease protein (Ftr1; STEARMAN et al. 1996). The two components together mediate high-affinity transport of iron into the cell, and both are required for this activity. Copper plays a special role in the reductive pathway, because copper is an obligate cofactor for the multi-copper oxidase (DANCIS et al. 1994). Copper delivery to the oxidase requires cellular copper uptake (mediated by Ctr1 and in some strains also by Ctr3) and copper delivery into the secretory pathway (mediated by the Atx1 metallochaperone and the Ccc2 P-type ATPase). Thus mutations that interfere with these steps or with the integrity of the secretory pathway result in defects of reductive iron uptake. Notably, copper proteins are not involved in iron uptake from siderophores (KNIGHT et al. 2002; KOSMAN 2003).

Both siderophore and reductive pathways for iron uptake are regulated in response to iron availability. Induction occurs under conditions of low iron and is mediated by binding of the Aft1 or Aft2 proteins to consensus sequence sites in the promoters of target genes (YAMAGUCHI-IWAI et al. 1996; BLAISEAU et al. 2001; RUTHERFORD et al. 2001). Expression of genes for both reductive and siderophore iron uptake are under control of Aft1/2, although their regulatory controls are not perfectly coordinated; modifying effects of other factors are likely to occur on the individual promoters. For example, siderophore uptake seems to turn on in response to more severe iron deprivation, and reductive iron uptake responds to milder iron deprivation. A class of mutants with defects in Fe-S cluster assembly exhibits a complex iron regulatory phenotype characterized by global effects on basic aspects of iron metabolism. Impaired iron-sulfur cluster assembly leads to constitutive Aft1/2-dependent gene expression and pleiomorphic phenotypes, including activation of both reductive and siderophore uptake systems. Within the cells of these mutants, iron accumulates within mitochondria and iron-sulfur cluster protein activities are deficient (JENSEN et al. 2004).

Here we screened a collection of 4847 haploid knockout strains [collection from European Saccharomyces cerevisiae Archive for Functional Analysis (EUROSCARF)] of S. cerevisiae for regulated activities that mediate cellular uptake of iron from siderophores or from ferric/ferrous salts. We identified hundreds of genes with effects on one or more of these activities. A subset of these was evaluated as homozygous diploid knockouts to decrease problems with suppressor mutations. Some of these genes were previously known to be involved in iron handling, although most were not previously connected to iron metabolism. An advantage of the present strategy is that it did not depend on evaluation of growth, which can be difficult to score and insensitive to subtle changes in iron homeostasis.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 ACKNOWLEDGEMENTS
 LITERATURE CITED
 
Yeast strains and growth conditions:
The entire collection of haploid knockout strains and a subset of diploid knockout strains of S. cerevisiae (constructed in BY4741 and BY4743, respectively) were purchased from EUROSCARF (Frankfurt, Germany). The genotype of the wild-type strain is: MATa his3{Delta}1 leu2{Delta}0 met15{Delta}0 ura3{Delta}0 (BY4741); MATa/MAT{alpha} his3{Delta}1/his3{Delta}1 leu2{Delta}0/leu2{Delta}0 met15{Delta}0/MET15 LYS2/lys2{Delta}0 ura3{Delta}0/ura3{Delta}0 (BY4743). The cells were grown in the wells of microtiter plates filled with complete YPD medium (100 µl medium/well). The cells were precultured for 48 hr at 30° without agitation and then diluted 10-fold in medium of the same composition and cultured for 5 hr under the same conditions before being used for experiments. For quality control, one mutant strain was randomly selected from each of six plates (YLR363w, YNR031c, YML033w, YLR278c, YOR041c, and YOR356w) and subjected to two PCR reactions for strain verification. The primers used were strain-specific primers A and KanB to detect the deleted gene and A and B to detect the wild-type gene. Primers were listed on the website for the Saccharomyces Genome Deletion Project. In each case, the correct deleted gene was detected and the wild-type allele was not detected, indicating that the well contained the correct strain deletion and furthermore that there was no evidence of contamination with other yeast strains.

Screens for siderophore uptake:
The 55Fe-labeled siderophore [ferrioxamine B (FOB), ferrichrome (FCH), triacetylfusarinine C (TAF), or enterobactin (ENB)] was added to the wells of the culture plates at a final concentration of 1 µM, and the cells were grown for 2 hr at 30°. The cells were then collected with a cell harvester (Brandel) and washed with water on the filter (Wallac, Gaithersburg, MD) before scintillation counting (Wallac Trilux MicroBeta).

Ferric reductase and ferrous uptake assays:
Cells were inoculated and grown as described above. After 5 hr of growth at 30°, the microtiter plates were centrifuged at 1500 x g for 5 min. The medium was aspirated, and the cells were washed twice in ice-cold 50 mM citrate, pH 6.6, 5% D-glucose and resuspended in 100 µl of the same. Optical density of the cell suspension was measured. For measurement of ferric reductase activity, 80 µl of the cell suspension was aliquoted to a fresh 96-well plate, and the assay solution of citrate buffer containing final concentrations of 1 mM ferric ammonium sulfate and 1 mM bathophenathrolinedisulfonate (BPS) was added. After 2 hr of incubation, absorbance of the ferrous-BPS complex was measured at 515 nm. The remaining cells were resuspended in 1 µM 55Fe citrate and incubated at 30°. After 1 hr, the cells were harvested (Tomtec Mach III M) onto fiber filters (Wallac). The filters were allowed to dry, and they were then soaked in scintillation fluid (Wallac BetaPlate Scint) before being counted using a 96-well liquid scintillation counter (Wallac Microbeta LS).

Clustering of data:
Mutant strains with similar patterns of activities for iron reductase, ferrous iron uptake, and siderophore iron uptake were identified by clustering analysis of the numerical data. Activities in the mutant strains were normalized by reference to the wild-type values, and these ratios were expressed as log base 2. We used the Cluster 3.0 software developed by EISEN et al. (1998) and modified by DE HOON et al. (2004) to perform k-mean clustering (k = 10) of the data, with uncentered correlation for similarity [metric for both strains (row) and activities (columns)]. Cluster 3.0 was retrieved at http://bonsai.ims.u-tokyo.ac.jp/~mdehoon/software/cluster. The output was visualized using the JavaTreeView software written by A. Saldanha (available at http://genome-www.stanford.edu/alok/TreeView).

Statistical analysis:
The diploid strains assayed for ferric reductase and ferrous uptake activities were assayed a total of four times. A regression analysis in the context of ANOVA was used to compare each mutant strain to the wild type (BY4743) to test for statistical differences. The regression is equivalent to multiple paired tests of each mutant compared, in turn, with the wild type. The underlying model for the ANOVA assumes that the population variances in each gene group are equal. A comparison of variances of the wild-type and the pooled mutants showed no significant difference in variances between the two groups. Significance level was set at P-value < 0.05.

RNA analysis:
For experiments requiring RNA isolation, cells were grown in complete YPD medium. After overnight preculture, the cells were diluted 10-fold and grown for 5 hr at 30° before RNA isolation. RNA was extracted as described previously (KOHRER and DOMDEY 1991). Northern blotting (using 20 µg of total RNA) and hybridization [at 42° in 50% (vol/vol) formamide] were done essentially as described (SAMBROOK et al. 1989). The 32P-labeled DNA fragments used as probes came from PCRs with the following primer sets: FET3 (623 bp), 5'-TTCTTGGACGATTTCTACTT-3' and 5'-GCAACTCTGGCAAACTTCTA-3'; SIT1 (872 bp), 5'-ACGCTAACCACATCTTCTCC-3' and 5'-TAACACTACAACCCAACCAA-3'. A 1.2-kb BamHI-HindIII fragment was used for the ACT1 gene.

Other:
The desferri-siderophores coprogen (CG) and ENB were purchased from the EMC microcollection GmbH (Germany). Desferri-FCH was purchased from Sigma (St. Louis). FOB refers to the commercially available mesylate derivative, Desferal (Novartis AG). TAF was a gift from H. Haas (Department of Molecular Biology, Medical University of Innsbruck, Austria).

Protoplasts were lysed and fractionated to purify isolated mitochondria as previously described (RAGUZZI et al. 1988).


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
ACKNOWLEDGEMENTS
 LITERATURE CITED
 
Initial screening of the haploid knockout collection for FOB uptake rates:
The collection of haploid knockout strains was evaluated using a quantitative assay for radioactive iron uptake from FOB. Results of the screen are shown as a scattergram (Figure 1), and they show a wide distribution of activities, with ~10% of clones exhibiting increased activity of >1.8-fold that of wild type. The number of clones exhibiting decreased activity (<0.6-fold for wild type) was significantly less, amounting to ~1%, but this may be due to lowered detection sensitivity for a decreased FOB uptake phenotype. The assays were repeated three times and a subset of strains with reproducibly abnormal uptake was selected for further study.



View larger version (17K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 1.— The complete haploid knockout collection from EUROSCARF was assayed for FOB uptake. Data for strains in sequentially numbered wells in microtiter plates (x-axis) were plotted as a scattergram against uptake data in counts per minute (y-axis). Due to the presence of numerous empty wells on the assay plates, the number of assays (>7000) exceeded the number of knockout strains (4847). The value for the wild type was 1009 cpm.

 
The FOB ferrisiderophore complex is colored yellow brown, whereas the iron-free siderophore is uncolored. Yeast strains such as AFT1-1up that accumulate FOB, yet are unable to dissociate the iron from the siderophore, form distinctive yellow colonies (LESUISSE et al. 2001). We screened the whole collection of haploid mutants for this phenotype. Seven mutants turned yellow due to accumulation of undissociated FOB (Figure 2). SPT4 (YGR063c, involved in RNA elongation) and YGR064w (unknown function) are overlapping genes, and thus further study will be required to establish which of these two genes is responsible for the observed phenotype. Other strains showing the same phenotype were mutated in CIK1 (which is involved in the organization of microtubule arrays), TIM18 (a translocase of the inner mitochondrial membrane), GGA2 (adaptor protein involved in Golgi-to-vacuole transport of special protein substrates), GGC1 [which we previously described as YHM1 (LESUISSE et al. 2004), a GTP/GDP exchanger of the mitochondrial inner membrane (VOZZA et al. 2004) that is involved in mitochondrial iron homeostasis]. Finally, YLL029w encodes a gene with unknown function (Figure 2). The color of colonies grown on FOB-containing agar plates was not directly related to the rate of FOB uptake by the cells, as illustrated by the ARF1 mutant, which showed high-FOB uptake and no color change indicative of ferrisiderophore accumulation (Figure 2). The colored colony phenotype therefore must reflect a specific defect in intracellular FOB dissociation that is separate from induction of the rate of cellular FOB uptake. Little is known about the mechanisms of intracellular iron release from internalized siderophores in S. cerevisiae. Theoretically, iron could be released from siderophores within cells by reduction of the chelated iron or by hydrolysis of the siderophore itself. The process could take place in a vesicular compartment or in the cytoplasm and might involve regulatory effects, depending on mitochondrial function. The effects on this process observed in the mutants identified here could be direct or indirect. For example, Tim18, a chaperone protein of the intermembrane space, is involved in translocation and membrane insertion of carrier proteins. Ggc1 is such a carrier protein and therefore is likely to depend on Tim18 function for correct localization and folding. Guanine nucleotides in turn are the substrate for Ggc1. Levels of guanine nucleotides apparently affect iron homeostasis and might influence siderophore iron release as well, although how these effects occur is unknown. The FOB accumulator phenotype of AFT1-1up mutants could indicate that Aft1 target genes mediate the effect, although none of the genes identified here are Aft1 regulated.



View larger version (24K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 2.— Colonies of haploid mutant strains that accumulated undissociated FOB. The indicated strains were spotted on a YPD-agar plate containing 50 µM FOB and allowed to grow for 3 days at 30°. The appearance of the colonies was recorded (left column). The rate of FOB uptake by each mutant was measured independently as described in MATERIALS AND METHODS. Values for each mutant were reported as n-fold the value of the wild-type FOB uptake rate (right column).

 
Candidate gene(s) for proteins directly participating in iron dissociation from FOB (some reductase or hydrolase, for example) were not identified here. Perhaps this is due to the nature of the screen itself. Mutant cells unable to dissociate siderophores at the intracellular level are expected to show increased siderophore uptake compared to wild-type cells when grown on siderophore-containing medium (because undissociated iron will be unavailable for Aft-mediated sensing). In the present screen, we selected strains grown on siderophore-free medium that showed high-FOB uptake, thus excluding putative mutants that would show unrepressed siderophore uptake because of a defect in siderophore iron sensing. In summary, the data of Figure 2 show that YGR063c/YGR064w, YLL029w, CIK1, TIM18, GGA2, and GGC1 are involved, maybe indirectly but specifically, in iron removal from intracellular FOB.

Our screen revealed that a special category of mutants was deregulated for FOB uptake: mutants that were affected in mitochondrial functions frequently showed high (1.5- to 2-fold that of the wild-type value) or very high (>2-fold that of the wild-type value) FOB uptake rates. This was the case for mutants deleted for genes encoding mitochondrial ribosomal proteins [MRP(L) genes] for {Delta}cox mutants and for {Delta}atp mutants (data not shown). We are currently studying these categories of mutants further to discriminate between specific effects and effects related to the rho– status of mutants. Results of this study will be presented elsewhere.

During the course of working with the haploid mutants, we noted that numerous strains showed changes in their uptake rates from one assay to the next. The changes were invariably from high uptake to lower uptake, and the changes became more pronounced over time and with repeated assaying (Table 1). We previously found that ssq1 and yfh1 mutant strains developed extragenic suppressor mutations in the genes mediating cellular iron uptake (STEARMAN et al. 1996; LESUISSE et al. 2003), and iron accumulation has been shown to produce a mutator phenotype (KARTHIKEYAN et al. 2002). We therefore became concerned that extragenic suppressor mutations arising in high-uptake mutant strains of the haploid knockout collection might be obscuring the true phenotypes. At this point, the haploid knockouts identified by the primary screen were reacquired as diploids carrying deletions of the corresponding ORFs. We reasoned that these would maintain more stable phenotypes, since recessive suppressor mutations would not be expressed. We also acquired the wild type, BY4743, and seven additional homozygous diploid mutants of genes previously characterized in iron uptake and homeostasis (FTR1, FET3, CTR1, ATX1, CCC2, SSQ1, and YHM1) to serve as controls.


View this table:
[in this window]
[in a new window]

  		TABLE 1 Progressive loss, in successive screens, of the high-FOB uptake phenotype in some haploid strains

 
Analysis of uptake from various siderophores:
A subset of the homozygous diploid mutants that showed a clear FOB uptake phenotype was analyzed for iron uptake from four different siderophores (FOB, FCH, TAF, and ENB). To further parse phenotypic categories, iron and copper content of the growth medium was manipulated, and reductase and ferrous-iron-uptake activities were measured. This first analysis sheds light on the overlap or differences among iron-uptake pathways. The data are shown in Figure 3. Each vertical column depicts data for a particular type of iron-uptake assay: ferric reductase, ferrous iron uptake, or siderophore iron uptake. Each horizontal row depicts the results for a single knockout clone, constructed by replacement of the open reading frame for a single gene, indicated by ORF number and name. The data represent mean values of triplicate measurements and are color coded, with red representing relative increased activity and green representing relative decreased activity compared to the parental control strain. The clustering software grouped the iron assays into three main nodes. The left-hand node consists of ferric reductase activity, measured following growth under low-copper (–Cu), copper-replete (+Cu), or low-iron (BPS) conditions. These correspond to moderately inducing, repressing, and maximally inducing conditions. The middle node groups ENB uptake with ferrous uptake, following low-copper (–Cu) or copper-replete (+Cu) growth. The right-hand node groups siderophore uptake from FOB, FCH, or TAF. What is immediately clear is that although these activities involved in iron acquisition may sometimes be coordinately expressed, they more often show divergent levels of expression. The divergences are particularly informative here, and we will discuss several examples shown in Figure 3.
  1. Horizontal nodes 2, 3. In these groups of mutants, ferrous uptake was low (green) and associated with increased ferric reductase activity (red). Uptake activities from siderophores FOB, FCH, and TAF increased (red). We can understand these phenotypes as consequences of a primary problem with ferrous iron transport and a secondary regulatory response. Iron in the growth medium was present as ferric iron chelates, and so an interruption of ferrous transport will result in cellular iron depletion and generation of an iron starvation signal. Thus when uptake activities were measured, ferrous transport was low, and ferric reductase was high due to iron starvation. Similarly, siderophore transport activities were induced by iron starvation experienced during growth of the cells.
  2. Horizontal nodes 1 and 9. In these groups of mutants, the most striking finding is a singular increase in uptake activity for the siderophore FOB. The effect is unlikely to represent a regulatory effect from iron starvation occurring during growth, because, as stated above, FOB iron was not present in the growth media. Furthermore, other iron-regulated activities, e.g., reductase, were not induced. We are left with the conclusion that FOB uptake was specifically altered, implying a regulatory signal specific for FOB uptake or a more direct gain of function effect on components of the uptake pathway.
  3. Horizontal node 5. In node 5, the FOB and FCH uptakes were increased, while other activities remained unperturbed. Again, these effects are unlikely to be the result of general misregulation of iron metabolism, because other reductase and ferrous transport was unaffected. The effects could be due to regulatory signals or upregulation of components shared by FOB and FCH uptake pathways.
  4. Horizontal node 6. In two mutant strains, {Delta}sit1 and {Delta}deg1, iron-uptake activities from FOB and FCH decreased, whereas other activities were minimally altered. SIT1 encodes a member of the MFS implicated in uptake of siderophores at the plasma membrane. The substrate specificity for Sit1p shows preference for FOB but also includes FCH (LESUISSE et al. 2001). The deg1 deletion strain phenocopies {Delta}sit1, raising the possibility that Deg1p modulates SIT1 expression or acts downstream of Sit1p in siderophore uptake. However, the decreased FOB and FCH uptake in these knockout mutants mirrors the increased FOB and FCH uptake in other mutants (horizontal blocks 5), indicating the existence of a specific trafficking pathway for handling these two siderophores.
  5. Horizontal nodes 3, 5, 6, 10 (and, to a lesser extent, 7). These groups show unique behavior for the TAF uptake activities, which remained unperturbed in the face of increased or decreased activity for transport of the other siderophores. However, in some cases TAF uptake seemed to move with the other transport activities and was clearly iron responsive. A possible explanation for these data is the existence of a threshold effect; for example, a lower-affinity Aft1-binding site on a MFS member promoter mediating TAF uptake could be envisaged. Another possibility, which does not exclude the previous one, is that TAF uptake and use involves different pathways of regulation and/or trafficking. This hypothesis is strengthened by our previous observations that the presence of TAF did not interfere with the uptake of other siderophores (FCH, FOB, or FC), which themselves showed complex relationships of competitive inhibition for uptake (LESUISSE et al. 1998, 2001). In addition, we observed that the Cyc8-Tup1 repressor had opposite effects on the uptake of FOB/FCH and on the uptake of TAF (LESUISSE et al. 2001).
  6. Horizontal node 8. In this node, the mutant strains showed increased cell reductase activity whatever the growth conditions and increased uptake activity whatever the iron source. This suggests that the strains grouped in this node suffer from a general misregulation of iron metabolism, as shown for mutants of mitochondrial iron homeostasis (yfh1, nfs1, isu1/2, etc.; LI et al. 1999; FOURY and TALIBI 2001; GERBER et al. 2004). Interestingly, four of six strains of this node were deleted for a gene encoding a mitochondrial protein (IMG2, YME1, MTM1, and YHM1/GGR1), highlighting once more the importance of the mitochondrial compartment for cellular iron homeostasis.
  7. Vertical column for ENB uptake. ENB uptake was found to group with ferrous transport activities rather than with other siderophore transport activities. Siderophores bind ferric iron with high affinity but ferrous iron with low affinity. Following reduction of the ferrisiderophore complex, ferrous iron could be released and permeated into the cell via the Fet3/Ftr1 ferrous transport complex (LESUISSE and LABBE 1989; YUN et al. 2000a). Grouping of the ENB data with the ferrous transport data suggests that iron bound to ENB was taken up via a reductive pathway and that a specific MFS member was not involved. We sought direct confirmation for this by measuring iron uptake from various iron sources in a {Delta}fre1{Delta}fre2 mutant strain, which lacks cell-surface ferrireductase activity (GEORGATSOU and ALEXANDRAKI 1994). Results are shown in Figure 4. As expected, uptake of ferric citrate decreased in the reductase-negative mutant compared to wild-type cells (Figure 4). This was also the case for CG, a siderophore for which no specific receptor has been found in S. cerevisiae (LESUISSE et al. 1998). Uptake of hydroxamate-type siderophores for which a specific nonreductive uptake pathway has been shown unambiguously in S. cerevisiae (FOB, FCH, TAF) was unaffected—or even increased—by the loss of cell-surface reductase activity (Figure 4). However, uptake of the catecholate-type siderophore ENB was strongly decreased in reductase-deficient cells (Figure 4). These results suggest that iron uptake from ENB follows a reductive pathway and that the putative siderophore transported via the ENB1 gene product remains to be identified. Another possibility is that a specific transporter for enterobactin (HEYMANN et al. 2000) could still exist and might be completely repressed in complete medium.



View larger version (31K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 3.— Clustering of data for ferric reductase, iron uptake from ferrous salts, and iron uptake from various siderophores by diploid knockout strains. FeRed, ferric reductase activity [under copper-deficient (–Cu), copper-replete (+Cu), or iron-deficient (BPS) conditions]; Fe(II), ferrous uptake activity. Results are expressed (according to the color scale) in log base 2 as the ratio of mutant to wild type. Means are from four experiments.

 


View larger version (21K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 4.— Iron-uptake rates from various iron sources by wild-type (solid columns) or {Delta}fre1{Delta}fre2 (open columns) cells of S. cerevisiae (S150-2B). Cells were incubated for 30 min in complete (YPD) medium containing 1 µM of the indicated iron source (as 55Fe) before washing and determining intracellular iron. Fe(III), ferric citrate. Mean ± SD is from four experiments.
 
Analysis of gene groups with characteristic phenotypes for reductive or nonreductive uptake:
In independent experiments, a set of 110 homozygyous diploid mutants, seven control strains (FTR1, FET3, CTR1, ATX1, CCC2, SSQ1, and GGC1), and wild type (BY4743) were analyzed for iron uptake from FOB (siderophore pathway) and ferrous ascorbate (reductive pathway). Ferric reductase was also measured, and iron and copper content of the growth media was manipulated. The clustering analysis grouped the data to identify 10 phenotypically similar groups of mutants. These phenotypic groups could be clustered into three more general categories. The defining characteristic of the first category was low-ferrous iron transport activity (Figure 5A). The hallmark of the second category was specific alteration (high or low) of FOB uptake activity (Figure 5B). Finally, mutants in the third category exhibited generalized induction of all iron transport activities compared with wild-type controls (Figure 5C). In each category, a few strains were assayed for their level of FET3 and SIT1 RNAs by Northern blotting (Figure 5).



View larger version (44K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 5.— Clustering of homozygous mutants with similar phenotypes for FOB iron uptake, ferric reductase, and ferrous iron uptake (left). Nodes of similar mutants were further grouped into three categories with different uptake signatures: (A) low-ferrous uptake, (B) high- or low-FOB uptake, and (C) high-FOB uptake, ferric reductase, and ferrous uptake. For each node, representative mutants were selected, and mRNA was isolated and tested by Northern blotting for SIT1 and FET3 mRNA levels (right). FeRed, ferric reductase activity [under copper-deficient (–Cu), copper-replete (+Cu), or iron-deficient (BPS) conditions]; Fe(II), ferrous uptake. Results are expressed (according to the color scale) in log base 2 as the ratio of mutant to wild type. Means are from four experiments.

 
Low-ferrous uptake:
The first category, characterized by low-ferrous iron transport, includes knockouts of genes with known functions in iron metabolism (Figure 5A). Within this category, the aft1 deletion mutant exhibited a unique phenotypic profile of decreased FOB transport activity and decreased ferrous transport activity (node 5A-1 in Figure 5A). Basal and induced ferric reductase levels were maintained, perhaps because of the ability of alternate transcription factors such as the copper-sensing Mac1p (or Aft2p) to mediate expression in the absence of Aft1p (GEORGATSOU et al. 1997; BLAISEAU et al. 2001). UME6 (node 5A-1) knockouts exhibited some aspects of the {Delta}aft1 phenotype, and the encoded protein (a zinc-finger transcription factor) may therefore play a role in iron sensing. FET3 and FTR1 are known Aft1p target genes. FET3 encodes a multi-copper oxidase, and the encoded protein is associated with Ftr1p to form the plasma membrane transporter complex for high-affinity ferrous uptake (STEARMAN et al. 1996). Thus, as expected, FET3 and FTR1 knockouts showed negligible high-affinity ferrous transport activity (node 5A-1). These mutants clustered here, although they differed from the AFT1 deletion phenotype in that FOB uptake was not decreased in mutants of the ferrous transporter. Surprisingly, FOB uptake and ferric reductase were not induced in these iron transport mutants, implying that inactivation of the ferrous transporter does not by itself produce severe iron starvation. Other iron acquisition pathways such as low-affinity iron transporters (DIX et al. 1997) likely can function in rich media to supply adequate iron to prevent starvation. This was confirmed by Northern blotting FET3 and SIT1 mRNAs from the {Delta}ftr1 mutant and from wild-type cells: transcription of these genes was only slightly increased in the mutant (Figure 5).

Mutants blocked in riboflavin biosynthesis ({Delta}rib4, node 5A-1) also showed deregulation of iron metabolism. However, the phenotype of the haploid {Delta}rib4 strain differed from that of the diploid {Delta}rib4/{Delta} strain, and both the haploid and diploid strains tended to accumulate extragenic suppressor mutations (data not shown). The effect of riboflavin synthesis on iron metabolism will be described elsewhere.

Node 5A-2 clusters data for mutants of CTR1, CCC2, and ATX1. These mutants exhibited low-ferrous transport activity, high ferric reductase, and high-FOB uptake. The corresponding genes have established roles in copper transport, and therefore the iron regulatory phenotypes of the mutants are likely secondary to effects on copper. CTR1 encodes a plasma membrane copper transport protein. The function of CTR1 is redundant with CTR3. However, the latter is inactivated by transposon insertion in some yeast genetic backgrounds (including BY4741/2), leading to dependence on CTR1 expression for high-affinity copper uptake and high-affinity iron uptake (KNIGHT et al. 1996). ATX1 encodes a copper chaperone, dedicated to delivering copper to the P-type ATPase Ccc2p, which in turn transports copper into the secretory pathway for Fet3p activation. The genes CTR1, CCC2, and ATX1 are involved in copper handling and delivery of copper to apo-Fet3, and thus the corresponding mutants have deficient Fet3 activity, accounting for the low-ferrous transport activities. Another characteristic of node 5A-2 is greater induction of FOB uptake and ferric reductase compared with node 5A-1 (FET3 and FTR1 mutants, for example). Similarly, transcription of FET3 and SIT1 was increased compared to wild-type levels in the {Delta}ccc2/{Delta} strain representative of this node (Figure 5). A possible explanation for these differences is provided by the role of Fet5, a multi-copper oxidase associated with the Fth1 permease in the vacuole membrane and dedicated to pumping iron outward to the cytoplasm (URBANOWSKI and PIPER 1999). Thus copper deficiency or impaired copper trafficking will lead to defects in both Fet3 and Fet5 oxidases, with consequently more severe cellular iron starvation than defects in either oxidase separately. BUD32, PKR1, and RAV1 knockouts also showed similar phenotypes to the copper trafficking mutants, with low-ferrous uptake partly compensated by copper addition to the media and constitutively high-FOB uptake (node 5A-2). The functions of Bud32 and of Pkr1 are unclear, but Rav1 is part of a complex named "regulator of the (H+)-ATPase of the vacuolar and endosomal membranes" (SEOL et al. 2001), which could be involved in copper trafficking in a way that remains to be determined.

Node 5A-3 includes protein sorting mutants localized to later stages in the secretory pathway. Also included are VMA mutants lacking constituents or regulators of the hydrogen transporting vacuolar ATPase. The iron phenotype of these mutants depended on transcriptional effects on iron transporter expression and also probably on post-transcriptional effects. Representative mutants of node 5A-3 ({Delta}pep12 and {Delta}vma6) showed a significant increase in the levels of FET3 and SIT1 transcription (Figure 5). Several mutants ({Delta}pep12, {Delta}vps15, {Delta}arf1, etc.) with impaired endosome fusion with late endosome and vacuolar compartments are expected to exhibit decreased recycling of Sit1 (KIM et al. 2002). Increased Sit1 at the cell surface in these mutants could then enhance cellular uptake of FOB by a post-transcriptional mechanism.

The signature iron phenotype of these mutants overlaps that produced by copper deficiency and could be explained by missorting of copper transporters and/or copper delivery components. Alternatively, disruption of acidification of the secretory pathway might interfere with efficient copper loading of apoproteins. Consistent with this idea, some ferrous transport defects in some of the mutants were corrected by growth in high concentrations of copper (50 µM added to YPD). The degree of correction was variable. No correction was observed for FET3 and FTR1 mutants as expected, since the copper-binding target Fet3p is absent or retained in the earliest part of the secretory pathway. Partial correction was observed for CTR1, ATX1, and CCC2 mutants. Some VPS mutants, such as {Delta}vps4 and {Delta}vps9, showed partial correction. The phenotypes of these diploid mutants in the BY4743 background were slightly different from previously published experience with haploid mutants in a different genetic background (YUAN et al. 1997). However, these differences are difficult to interpret, because the ability of copper to correct ferrous transport defects in various mutants is poorly understood. The effects are presumed to be mediated by an uncharacterized low-affinity copper transport pathway able to bypass the usual copper delivery pathway to Fet oxidases in the secretory pathway.

Altered FOB transport:
The second phenotypic category is characterized by altered FOB transport activity (Figure 5B). Iron use from siderophores involves a complex machinery regulated at transcriptional and post-transcriptional levels. The presence of siderophores in the medium has to be "sensed" by the cells prior to the siderophore receptors being exported at the plasma membrane (KIM et al. 2002). Transcription of genes for these receptors is itself subject to complex regulation, depending on Aft1/2 and Tup1/Cyc8 (LESUISSE et al. 2001). Iron-dependent and iron-independent signals are integrated in control of these genes. Endocytosis and recycling of the siderophore receptors depend on the machinery for intracellular protein trafficking, involving many regulatory steps at various levels. This may explain the high number of genes implicated in FOB uptake and regulation.

In node 5B-1 in Figure 5B, FOB uptake was increased but reductase and ferrous transport were not altered. In node 5B-2, FOB uptake was increased and ferric reductase was decreased under some conditions, while ferrous transport was variable. The phenotype of these mutants was probably not due to direct transcriptional effects on the "iron regulon" genes, as the levels of FET3 and SIT1 transcription were not significantly altered in representative strains ({Delta}gga2, {Delta}vam6, {Delta}ssn8; Figure 5). Many genes of these first two nodes (5B-1, 5B-2) are involved in transcriptional regulation and RNA processing (LSM1, LSM6, LSM7, PAT1, SSN3, CYC8, SSN8, MED1, etc.). Thus, effects of these mutants on FOB uptake are likely to be indirect (in view of the lack of change in SIT1 message level in the {Delta}ssn8 strain). Other genes in this group are involved in intracellular protein trafficking (GGA2, VAM6, APS3, etc.), and these might influence trafficking of siderophores or siderophore receptors. One of these genes, GGA2 (block 5B-1), is particularly interesting, since the corresponding mutant showed strongly increased FOB uptake with unaltered SIT1 transcription, while other iron-dependent activities (reductase and ferrous transport) were unchanged. Moreover, TAF uptake by the {Delta}gga2 mutant was unaltered (see Figure 3, node 5), and undissociated FOB accumulated in this mutant (see Figure 2). Thus, the gga2 mutation creates a specific interruption upstream of the dissociation step for FOB and FCH. Gga proteins are known to contribute to protein sorting by functioning as adaptors between cargo proteins and clathrin coats (KATZMANN et al. 2002). We are currently trying to use the {Delta}gga2 mutant as a tool to determine which cellular compartment(s) is involved in iron dissociation from FOB.

The node 5B-3 is very interesting. It includes two strains mutated in genes that are antisense to one another, BUD25 and HEM14. Both strains showed markedly increased FOB uptake, strongly decreased reductase, and unaltered ferrous transport. Both strains showed heme deficiency, although the Hem14 block in heme biosynthesis is a little leaky, because protoporphyrinogen can be oxidized into protoporphyrin nonenzymatically to some extent (CAMADRO et al. 1994). Both strains were able to grow slowly on YPD medium without supplementation with hemin or tween/ergosterol. Reductase activity of these strains was virtually absent, probably due to the lack of heme, which is a cofactor for the Fre protein reductases (SHATWELL et al. 1996). The observation that these strains showed very high uptake of FOB raises the question of the role of heme synthesis in iron homeostasis. In a recent study, CRISP et al. (2003) showed that heme-deficient mutants downregulated genes of the "iron regulon." However, regulation of genes involved in siderophore uptake by these mutants was not examined. The present study shows that SIT1 transcription was tremendously increased in the {Delta}hem14 mutant compared to wild-type cells (Figure 5), which could account for the observed phenotype of high-FOB uptake. In a previous study, we showed that some heme mutants showed constitutively high-FOB uptake (LESUISSE and LABBE 1989). We therefore studied this question further and analyzed the influence of global heme deficiency and of specific genes (the HEM gene family and the BUD25 gene) on siderophore and ferrous transport. Results of this study will be published elsewhere.

Node 5B-4 groups strains with low-FOB uptake activities. As expected, the {Delta}sit1 mutant was found in this group. The influence of HSP12 disruption on iron metabolism should be studied further, because the haploid and diploid mutants behaved differently. The haploid {Delta}hsp12 strain was initially selected for high-FOB uptake, but tended to loose this phenotype in the successive screens (see Table 1), while the diploid {Delta}hsp12/{Delta} strain showed a stable low-FOB uptake phenotype. The reason for such a discrepancy is unknown. The phenotype of the {Delta}deg1 strain resembles the {Delta}sit1 strain. DEG1 encodes a tRNA pseudouridine synthase and may play a role in regulating SIT1 expression or activity. Notably, the level of SIT1 transcript was very low in this strain (Figure 5).

Generalized induction of all activities:
The third phenotypic category (Figure 5C) groups together mutants that showed highly induced reductase, FOB transport, and ferrous transport. The corresponding ORFs are candidates for genes that function in general iron metabolism or iron sensing. Many categories of genes are included here, including genes for nuclear-encoded mitochondrial functions, genes implicated in endocytosis, vacuolar protein sorting, and transcriptional regulators. Transcriptional and post-transcriptional effects were involved in producing the mutant phenotypes. One strain taken as representative of node 5C-1 (YLL029w) showed a moderate increase (about twofold) in the level of FET3 and SIT1 transcripts (Figure 5). However, iron-uptake activities were more severely affected (about fivefold increased), suggesting that post-transcriptional regulation was also involved. In node 5C-2, a number of genes implicated in endocytosis and intracellular protein trafficking are grouped together. In contrast to what was observed in other endocytosis mutants (see strains of node 5A-3), no direct effect on FET3 and/or SIT1 transcription was observed here (Figure 5). Note that the biological significance of low-ferrous transport following growth in the presence of copper is unclear, but nonetheless, this characteristic seems to enable phenotypic grouping of endocytosis mutants. Several of the mutants in this node ({Delta}snf7, {Delta}vps20, {Delta}vps25, {Delta}vps28, {Delta}vps36, and {Delta}bro1) are expected to be defective in fusion of late endosome vesicles to the vacuole (KATZMANN et al. 2002). The consequences of this organellar trafficking abnormality may be to increase Sit1p concentration at the cell surface, leading to increased FOB uptake. In addition, there may be enhanced surface expression of reductase and ferrous transport proteins, leading to increased reductase activity and ferrous transport activity.

Statistical analysis of mutants with effects on iron uptake grouped according to proposed cellular localizations of the corresponding gene products:
The data up to this point have been analyzed using k-means clustering of uptake and reductase activities to group mutants according to phenotypic similarities. The multiple biological replicates (n = 4) of the data for ferrireductase and ferrous uptake allowed us to perform a more standard statistical analysis (regression in the context of ANOVA), comparing each mutant with the wild type. For this analysis only diploid mutants that had a FOB uptake rate of ≤0.6 times wild type (low) and of ≥1.8 times wild type (high) were included. The data from this alternative statistical analysis are presented according to the proposed cellular localization of the corresponding proteins (from Saccharomyces Genome Database Gene Ontology (GO) annotations) in Table 2. For ferric reductase activity and ferrous uptake, mutants are described with high or low activity if there was a statistical difference compared to wild type at a level of P < 0.05. An overview of Table 2 indicates that the majority of mutants that had altered FOB uptake also had altered ferrireductase activity. However, of the 81 mutants that had high-FOB uptake, only 28 had significantly altered ferrous uptake (7 had high and 21 had low-ferrous uptake—not including the mutant control strains). Likewise, of the five mutants that had low-FOB uptake, only 1 mutant also had abnormal (low) ferrous uptake. Thus, comparison of FOB uptake with ferrous uptake is likely to be more informative than comparison with ferrireductase. The lack of correlation between FOB and ferrous uptake in these mutants is probably reflective of the functional separation of these two pathways in bringing iron into the cell. However, organelles in which the FOB and ferrous uptake phenotypes converge might be key components of intracellular iron transport or iron regulation systems.


View this table:
[in this window]
[in a new window]

  		TABLE 2 GO proposed location of proteins corresponding to homozygous diploid mutants with high- or low-FOB uptake, ferrireductase activity, and ferrous uptake

 
The convergence is particularly apparent in the mitochondria where four mutants ({Delta}img2, {Delta}mtm1, {Delta}tom5, and {Delta}YCR024c) as well as the control mutant {Delta}ggc1 (formerly {Delta}yhm1) exhibit high-FOB and high-ferrous uptake (Table 2). Three of the corresponding proteins are involved in mitochondrial protein biosynthesis or protein import. Img2p is a constituent of mitochondrial ribosome complex. YCR024c encodes a mitochondrial protein with asparaginyl-tRNA synthetase activity (LANDRIEU et al. 1997). Disruption of either gene generates a petite phenotype, although why deletion also generates this particular iron phenotype is not clear. Tom5p is part of the mitochondria outer membrane translocase complex and may serve as a functional link between the outer membrane receptors and the general import pore of the mitochondria (DIETMEIER et al. 1997). A role for this protein in the import of proteins involved in Fe-S cluster and/or heme cannot be excluded. The effect of Tom5p on cellular iron uptake is likely to be post-transcriptional, since disruption of the corresponding genes did not significantly alter the level of FET3 and SIT1 transcripts (data not shown). The corresponding protein for the other mutant in this group, Mtm1p, is a mitochondrial carrier protein involved in activation of Sod2 (LUK et al. 2003). The {Delta}mtm1 mutant was shown to accumulate iron in the mitochondria (LUK et al. 2003). Overall, the phenotype of {Delta}mtm1 is very similar to that of strains defective in mitochondrial Fe-S cluster assembly—ssq1, nfs1, and the control mutant strain ggc1 (KNIGHT et al. 1998; LI et al. 1999; LESUISSE et al. 2004). This convergence of high-FOB uptake and high-ferrous uptake phenotypes in strains with mutations of genes encoding mitochondrial proteins underscores the importance of this organelle in cellular iron homeostasis and lends weight to our hypothesis (SANTOS et al. 2003) that a mitochondrial Fe-S cluster (that could be sensitive to oxidative stress especially by superoxides) performs a regulatory function in a signal transduction chain from the mitochondria to the cytosol to inform the cell of iron status. The majority (21) of the high-FOB mutants with a defect in ferrous uptake had low-ferrous uptake. This reciprocal phenotype convergence (high-FOB, low-ferrous uptake) was particularly notable in mutants whose corresponding proteins are located in or involved with the secretory/vacuole system and the nucleus. The secretory/vacuolar mutants include {Delta}pep7, {Delta}rav1, {Delta}vps9, {Delta}vma6, {Delta}vma10, and {Delta}vma13. The low-ferrous uptake of {Delta}rav1 and {Delta}vps9 mutants could be suppressed by the addition of copper to the growth media in a manner identical to the control mutants {Delta}ccc2 and {Delta}atx1. Rav1p is involved in the regulation of vacuolar acidification and Vps9 is a guanine nucleotide exchange factor involved in vacuolar protein transport (HAMA et al. 1999). Although the low-ferrous-uptake phenotype of {Delta}vam6, {Delta}vma10, and {Delta}vma13 cannot be corrected by the addition of copper to the growth media, all three corresponding proteins are essential in vacuolar acidification. Vma6p and Vma10p are subunits of the H(+)-ATPase (BAUERLE et al. 1993; SUPEKOVA et al. 1995), and Vma13p is required for H(+)-ATPase assembly. These data support the idea that a functional vacuole is required for normal iron and copper cellular homeostasis (SZCZYPKA et al. 1997; DAVIS-KAPLAN et al. 2004). When the vacuole is dysfunctional, ferrous uptake is diminished and the increase in FOB uptake is likely the result of a compensatory regulatory system.

The nucleus was also a site for reciprocal convergence of high-FOB and low-ferrous uptake for the mutants {Delta}tup1, {Delta}cyc8, {Delta}med1, {Delta}ssn3, and {Delta}taf1 (Table 2). The corresponding proteins all mediate transcription. We previously showed that the Tup1/Cyc8 general repressor complex was involved in iron-uptake regulation, with preferential effects on derepressing the siderophore uptake pathway (LESUISSE et al. 2001). At this stage, it is not possible to discriminate between factors that directly mediate transcription of the "iron regulon" genes and those that could act more indirectly. However, the copper correction of the low-ferrous uptake of the {Delta}med1 mutant might suggest that the corresponding protein mediates its effect on the genes encoding copper transport/distribution proteins.

Finally in this statistical analysis, one mutant had a distinct phenotype of low-FOB uptake and low-ferrous uptake. That mutant was {Delta}aft1, the well-characterized iron regulatory transcription factor (YAMAGUCHI-IWAI et al. 1995). The uniqueness of this mutant phenotype from a primary screen of >4847 mutants underscores the critical role that Aft1p plays in yeast iron homeostasis.

Phenotypic categories and iron distribution:
The initial screen identified many mutant strains of yeast with increased activity for iron uptake from FOB. The increased FOB uptake could be a regulatory consequence of cellular iron starvation occurring during growth (phenotypic category I). Alternatively, it could result from specific effects that increase siderophore transport into cells (phenotypic category II). Finally, increased FOB uptake activity could result from a general perturbation of iron sensing and iron regulation (phenotypic category III). We next evaluated examples of knockout mutants from each phenotypic category in terms of their intracellular iron distribution following iron loading with FOB as an iron source (Table 3). The results show that iron distribution was abnormal for mutants from each category, with proportionately decreased mitochondrial accumulation of iron. In the {Delta}vma10 mutant (category I), biogenesis of the vacuolar ATPase is disrupted. The iron phenotype was characterized by impaired ferrous transport and increased FOB uptake activity. When the cells were loaded with iron from FOB, cell-associated iron was increased as expected. However, mitochondrial iron was not proportionately increased. In the {Delta}gga2 mutant (category II), the FOB uptake was specifically induced with no alteration of reductase or ferrous transport. The {Delta}gga2 mutant had the additional phenotype of accumulating colored iron siderophore complexes, indicating a problem in dissociating them or distributing them intracellularly (see above). Following iron loading from FOB, cellular iron was increased, and again, mitochondrial iron was not proportionately increased. In mutants of TOM5, POR1, FMC1, and IMG2 (category II), mitochondrial functions are disrupted to varying degrees and in a pleiotropic manner. Reductive and FOB transport were induced, but again, following loading with iron from FOB, mitochondrial iron accumulation was not proportionately increased. The implication of all these results is that cellular iron uptake and mitochondrial iron accumulation from FOB are subject to separate controls. Many of the mutations that impacted on cellular iron uptake, inducing FOB transport activity, did not induce mitochondrial deposition of the iron. This is in contrast to the phenotype of Fe-S cluster assembly mutants in which induction of FOB uptake activity and mitochondrial iron accumulation are salient features. Two mutants (category III) that did not adhere to the pattern described above were examined. Instead, they exhibited increased iron transport activities (reductase, ferrous transport, and FOB transport), but following loading with iron via FOB, they accumulated mitochondrial iron in proportion to cellular iron, and so the ratio of these compartmental accumulations was not altered (Table 3). YLL029w and YGL220w encode uncharacterized ORFs. The corresponding proteins are reported to be cytoplasmic, and it is interesting to speculate that they may coordinately control cytoplasmic and mitochondrial iron distribution. YLL029w may be especially interesting to study further, since the corresponding mutant accumulated FOB as its undissociated form (see above). One important question that arises from the analysis of iron distribution in strains showing increased FOB uptake is indeed related to the cellular location where iron dissociation from the siderophore takes place. Is iron necessarily released from the siderophore before entering the mitochondria, or is there some pathway for siderophore uptake and use by the mitochondria itself? Our iron distribution studies do not answer this question. Mutants known to accumulate iron by the reductive pathway followed by iron deposition in the mitochondria ({Delta}yfh1, {Delta}atm1, etc.) also accumulate iron from siderophores (LESUISSE et al. 2001), but the intracellular distribution of this siderophore iron was never investigated. The cells deleted for YLL029w accumulated FOB in an undissociated form (whole cells were colored after growth with FOB) and showed increased iron in their mitochondria compared to wild-type cells (Table 3), but further work would be required to determine if part of the intracellular FOB could be found undissociated inside the mitochondria. Mitochondria are intrinsically more pigmented than whole cells, and thus accumulation of the siderophore pigment into this organelle is less easily determined than in whole cells. We are currently working on these questions related to the intracellular dissociation of siderophores.


View this table:
[in this window]
[in a new window]

  		TABLE 3 Intracellular iron distribution in selected strains belonging to different phenotypic categories

 
General conclusion and caveats:
This study of the entire set of viable haploid deletion strains of S. cerevisiae from the EUROSCARF collection revealed a huge number of genes with involvement in cellular iron homeostasis. In the initial screen for FOB uptake, >10% of the mutants (570 strains) showed altered activity compared to the wild-type strain. A subset with reproducibly abnormal phenotypes was selected and analyzed as homozygous diploids (197 strains including 79 with mutations in COX, ATP, and MRP(L) genes, which will be presented separately). Some strains affected in iron metabolism probably passed through the net of our successive screens for several reasons. First, all the strains that did not show altered iron uptake in the first screen were not studied further; thus, errors in measuring uptake values in the first screen may have led to discarding some potentially interesting strains. Second, our screen was more sensitive at detecting high-FOB uptake strains than low-FOB uptake strains (for technical reasons related to the specific activity of 55Fe that we used). Finally, a large number of mutants showed an unstable phenotype, with progressive attenuation of the high-uptake phenotype. Probably this was due to accumulation of extragenic suppressor mutations. In addition, it is important to note that only the strains considered as viable in the EUROSCARF collection were screened. Curiously, mutants with defective Fe-S cluster assembly were not selected here. Some were classified as essential genes (e.g., YFH1, ATM1) and others were missed, perhaps due to masking of the phenotype by suppressors (e.g., SSQ1). Despite all these limitations, in this study a large number of genes were implicated in iron metabolism for the first time. These included genes encoding predicted proteins located in the plasma membrane, cytoplasm, secretory pathway, nucleus, and mitochondria. Future work will endeavor to determine which of these affect iron metabolism directly via transport or sequestration of iron and which affect iron metabolism indirectly via signals or protein trafficking.

A special feature of the screen is the measurement of distinct uptake systems. One of the uptake systems (reductive) is active during growth in the usual media and its activity reflects iron regulation and the history of iron exposure during growth of the cells. The other system (siderophore) is generally not used during growth in customary media, because these media do not contain siderophores. However, activity for siderophore uptake is also responsive to iron homeostasis. Clustering of the uptake data makes possible a distinction between mutants with altered trafficking of specific components and mutants with generally altered iron metabolism.


   ACKNOWLEDGEMENTS
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 ACKNOWLEDGEMENTS
LITERATURE CITED
 
We thank David Dancis for technical support and Rachel Weinstein of the University of Pennsylvania Biostatistics Analysis Center for assistance with the statistical analysis. This work was supported by grants from the French Ministère de la Recherche (Programme de Recherches Fondamentales en Microbiologie, Maladies Infectieuses et Parasitaires and Réseau Infection Fongique) and from the Association pour la Recherche sur le Cancer (ARC 5439 and 4396). This work was supported by National Institutes of Health Grant DK-53953 (to A.D.).


   LITERATURE CITED
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 ACKNOWLEDGEMENTS
 LITERATURE CITED
 

BAUERLE, C., M. N. HO, M. A. LINDORFER and T. H. STEVENS, 1993 The Saccharomyces cerevisiae VMA6 gene encodes the 36-kDa subunit of the vacuolar H(+)-ATPase membrane sector. J. Biol. Chem. 268: 12749–12757.[Abstract/Free Full Text]

BLAISEAU, P. L., E. LESUISSE and J. M. CAMADRO, 2001 Aft2p, a novel iron-regulated transcription activator that modulates, with Aft1p, intracellular iron use and resistance to oxidative stress in yeast. J. Biol. Chem. 276: 34221–34226.[Abstract/Free Full Text]

CAMADRO, J. M., F. THOME, N. BROUILLET and P. LABBE, 1994 Purification and properties of protoporphyrinogen oxidase from the yeast Saccharomyces cerevisiae. Mitochondrial location and evidence for a precursor form of the protein. J. Biol. Chem. 269: 32085–32091.[Abstract/Free Full Text]

CRISP, R. J., A. POLLINGTON, C. GALEA, S. JARON, Y. YAMAGUCHI-IWAI et al., 2003 Inhibition of heme biosynthesis prevents transcription of iron uptake genes in yeast. J. Biol. Chem. 278: 45499–45506.[Abstract/Free Full Text]

DANCIS, A., R. D. KLAUSNER, A. G. HINNEBUSCH and J. G. BARRIOCANAL, 1990 Genetic evidence that ferric reductase is required for iron uptake in Saccharomyces cerevisiae. Mol. Cell. Biol. 10: 2294–2301.[Abstract/Free Full Text]

DANCIS, A., D. G. ROMAN, G. J. ANDERSON, A. G. HINNEBUSCH and R. D. KLAUSNER, 1992 Ferric reductase of Saccharomyces cerevisiae: molecular characterization, role in iron uptake, and transcriptional control by iron. Proc. Natl. Acad. Sci. USA 89: 3869–3873.[Abstract/Free Full Text]

DANCIS, A., D. S. YUAN, D. HAILE, C. ASKWITH, D. EIDE et al., 1994 Molecular characterization of a copper transport protein in S. cerevisiae: an unexpected role for copper in iron transport. Cell 76: 393–402.[CrossRef][Medline]

DAVIS-KAPLAN, S. R., D. M. WARD, S. L. SHIFLETT and J. KAPLAN, 2004 Genome-wide analysis of iron-dependent growth reveals a novel yeast gene required for vacuolar acidification. J. Biol. Chem. 279: 4322–4329.[Abstract/Free Full Text]

DE HOON, M. J., S. IMOTO, J. NOLAN and S. MIYANO, 2004 Open source clustering software. Bioinformatics 20: 1453–1454.[Abstract/Free Full Text]

DIETMEIER, K., A. HONLINGER, U. BOMER, P. J. DEKKER, C. ECKERSKORN et al., 1997 Tom5 functionally links mitochondrial preprotein receptors to the general import pore. Nature 388: 195–200.[CrossRef][Medline]

DIX, D., J. BRIDGHAM, M. BRODERIUS and D. EIDE, 1997 Characterization of the FET4 protein of yeast. Evidence for a direct role in the transport of iron. J. Biol. Chem. 272: 11770–11777.[Abstract/Free Full Text]

EISEN, M. B., P. T. SPELLMAN, P. O. BROWN and D. BOTSTEIN, 1998 Cluster analysis and display of genome-wide expression patterns. Proc. Natl. Acad. Sci. USA 95: 14863–14868.[Abstract/Free Full Text]

FOURY, F., and D. TALIBI, 2001 Mitochondrial control of iron homeostasis. A genome wide analysis of gene expression in a yeast frataxin-deficient strain. J. Biol. Chem. 276: 7762–7768.[Abstract/Free Full Text]

GEORGATSOU, E., and D. ALEXANDRAKI, 1994 Two distinctly regulated genes are required for ferric reduction, the first step of iron uptake in Saccharomyces cerevisiae. Mol. Cell. Biol. 14: 3065–3073.[Abstract/Free Full Text]

GEORGATSOU, E., L. A. MAVROGIANNIS, G. S. FRAGIADAKIS and D. ALEXANDRAKI, 1997 The yeast Fre1p/Fre2p cupric reductases facilitate copper uptake and are regulated by the copper-modulated Mac1p activator. J. Biol. Chem. 272: 13786–13792.[Abstract/Free Full Text]

GERBER, J., K. NEUMANN, C. PROHL, U. MUHLENHOFF and R. LILL, 2004 The yeast scaffold proteins Isu1p and Isu2p are required inside mitochondria for maturation of cytosolic Fe/S proteins. Mol. Cell. Biol. 24: 4848–4857.[Abstract/Free Full Text]

HAAS, H., 2003 Molecular genetics of fungal siderophore biosynthesis and uptake: the role of siderophores in iron uptake and storage. Appl. Microbiol. Biotechnol. 62: 316–330.[CrossRef][Medline]

HAMA, H., E. A. SCHNIEDERS, J. THORNER, J. Y. TAKEMOTO and D. B. DEWALD, 1999 Direct involvement of phosphatidylinositol 4-phosphate in secretion in the yeast Saccharomyces cerevisiae. J. Biol. Chem. 274: 34294–34300.[Abstract/Free Full Text]

HEYMANN, P., J. F. ERNST and G. WINKELMANN, 1999 Identification of a fungal triacetylfusarinine C siderophore transport gene (TAF1) in Saccharomyces cerevisiae as a member of the major facilitator superfamily. Biometals 12: 301–306.[CrossRef][Medline]

HEYMANN, P., J. F. ERNST and G. WINKELMANN, 2000 A gene of the major facilitator superfamily encodes a transporter for enterobactin (Enb1p) in Saccharomyces cerevisiae. Biometals 13: 65–72.[CrossRef][Medline]

JENSEN, L. T., R. J. SANCHEZ, C. SRINIVASAN, J. S. VALENTINE and V. C. CULOTTA, 2004 Mutations in Saccharomyces cerevisiae iron-sulfur cluster assembly genes and oxidative stress relevant to Cu,Zn superoxide dismutase. J. Biol. Chem. 279: 29938–29943.[Abstract/Free Full Text]

KARTHIKEYAN, G., L. K. LEWIS and M. A. RESNICK, 2002 The mitochondrial protein frataxin prevents nuclear damage. Hum. Mol. Genet. 11: 1351–1362.[Abstract/Free Full Text]

KATZMANN, D. J., G. ODORIZZI and S. D. EMR, 2002 Receptor downregulation and multivesicular-body sorting. Nat. Rev. Mol. Cell Biol. 3: 893–905.[CrossRef][Medline]

KIM, Y., C. W. YUN and C. C. PHILPOTT, 2002 Ferrichrome induces endosome to plasma membrane cycling of the ferrichrome transporter, Arn1p, in Saccharomyces cerevisiae. EMBO J. 21: 3632–3642.[CrossRef][Medline]

KNIGHT, S. A., S. LABBE, L. F. KWON, D. J. KOSMAN and D. J. THIELE, 1996 A widespread transposable element masks expression of a yeast copper transport gene. Genes Dev. 10: 1917–1929.[Abstract/Free Full Text]

KNIGHT, S. A., N. B. SEPURI, D. PAIN and A. DANCIS, 1998 Mt-Hsp70 homolog, Ssc2p, required for maturation of yeast frataxin and mitochondrial iron homeostasis. J. Biol. Chem. 273: 18389–18393.[Abstract/Free Full Text]

KNIGHT, S., E. LESUISSE, R. STEARMAN, R. D. KLAUSNER and A. DANCIS, 2002 Reductiv