THE JOURNAL OF BIOLOGICAL CHEMISTRY
© 2003 by The American Society for Biochemistry and Molecular Biology, Inc.
Vol. 278, No. 42, Issue of October 17, pp. 40612–40620, 2003
Printed in U.S.A.
A Specific Role of the Yeast Mitochondrial Carriers Mrs3/4p in
Mitochondrial Iron Acquisition under Iron-limiting Conditions*
Received for publication, July 21, 2003, and in revised form, August 4, 2003
Published, JBC Papers in Press, August 5, 2003, DOI 10.1074/jbc.M307847200
Ulrich Mühlenhoff‡, Jochen A. Stadler§, Nadine Richhardt‡, Andreas Seubert¶,
Thomas Eickhorst¶, Rudolf J. Schweyen§, Roland Lill‡, and Gerlinde Wiesenberger§储
From the ‡Institut für Zytobiologie und Zytopathologie, Philipps-Universität Marburg, Robert-Koch Str. 6, 35033
Marburg, Germany, §Max F. Perutz Laboratories, Department of Microbiology and Genetics, University of Vienna,
Dr. Bohrgasse 9, 1030 Wien, Austria, and ¶Fachbereich Chemie, Philipps-Universität Marburg, Hans-Meerwein Strasse,
35032 Marburg, Germany
The yeast genes MRS3 and MRS4 encode two members
of the mitochondrial carrier family with high sequence
similarity. To elucidate their function we utilized genome-wide expression profiling and found that both deletion and overexpression of MRS3/4 lead to up-regulation of several genes of the “iron regulon.” We therefore
analyzed the two major iron-utilizing processes, heme
formation and Fe/S protein biosynthesis in vivo, in organello (intact mitochondria), and in vitro (mitochondrial extracts). Radiolabeling of yeast cells with 55Fe
revealed a clear correlation between MRS3/4 expression
levels and the efficiency of these biosynthetic reactions
indicating a role of the carriers in utilization and/or
transport of iron in vivo. Similar effects on both heme
formation and Fe/S protein biosynthesis were seen in
organello using mitochondria isolated from cells grown
under iron-limiting conditions. The correlation between
MRS3/4 expression levels and the efficiency of the two
iron-utilizing processes was lost upon detergent lysis of
mitochondria. As no significant changes in the mitochondrial membrane potential were observed upon
overexpression or deletion of MRS3/4, our results suggest that Mrs3/4p carriers are directly involved in mitochondrial iron uptake. Mrs3/4p function in mitochondrial iron transport becomes evident under ironlimiting conditions only, indicating that the two
carriers do not represent the sole system for mitochondrial iron acquisition.
Proteins belonging to the mitochondrial carrier family
(MCF)1 form a large group of structurally related proteins,
which exist exclusively in eukaryotes (for reviews, see Refs.
1–3). Typical mitochondrial carrier proteins have a molecular
mass of about 35 kDa, contain six membrane-spanning segments, and have a tripartite structure. Each of the three parts
* This work was supported by grants from the Deutsche Forschungsgemeinschaft, Sonderforschungsbereich 593, the European Commission
(Grant QLG1-CT-2001-00966), the Fonds der Chemischen Industrie,
and Project T105 by the Austrian Ministry of Education, Science and
Culture. The costs of publication of this article were defrayed in part by
the payment of page charges. This article must therefore be hereby
marked “advertisement” in accordance with 18 U.S.C. Section 1734
solely to indicate this fact.
储 Holds a Herta Firnberg position funded by the Austrian Ministry of
Education, Science and Culture. To whom correspondence should be
addressed. Tel.: 43-1-4277-54614; Fax: 43-1-4277-9546; E-mail:
[email protected].
1
The abbreviations used are: MCF, mitochondrial carrier family; HA,
hemagglutinin; TMRM, tetramethylrhodamine methyl ester; AES,
atomic emission spectrometry; MS, mass spectrometry; ICP, inductively coupled plasma.
is made up of about 100 amino acids and shares sequence
homology to the other modules of the proteins. Most members
of the MCF are integral proteins of the mitochondrial inner
membrane and function in the shuttling of various metabolites
and cofactors between the cytosol and mitochondria. The substrates of mitochondrial carrier proteins are rather diverse in
size and composition, ranging from protons transported by the
uncoupling protein up to large molecules such as ATP and ADP
exchanged by the ADP/ATP carrier. Other members of this
family are responsible for the transport of phosphate, citrate,
fumarate/succinate, carnitine/acylcarnitine, flavine adenine
dinucleotide (FAD), and other substrates. One member of the
MCF has been located in peroxisomes where it was shown to
transport ATP in exchange for AMP (4).
The genome of the yeast Saccharomyces cerevisiae contains
35 open reading frames that encode members of the MCF (5–7).
The genes can be divided into five subclasses. (i) The functions
of the encoded proteins are known, and their transport activities were investigated by expression in Escherichia coli and
reconstitution of the purified proteins in liposome vesicles (e.g.
Mir1p/PiC, phosphate (8); Arg11p/ORC, ornithine (9); Odc1p
and Odc2p, 2-oxoglutarate and 2-oxoadipate (10); Acr1p/SFC,
succinate/fumarate (11); Tpc1p, thiamine pyrophosphate (12);
etc.). (ii) The substrates of mitochondrial carrier proteins encoded by the second group are predicted from investigations of
mutant strains, but so far the carriers have not been reconstituted in liposome vesicles (e.g. Flx1p, FAD (13); Leu5p, coenzyme A (14)). (iii) Several genes encoding members of the MCF
were identified by genetic screens that did not immediately
allow any conclusions with respect to their substrates (MRS3
and MRS4 (15); MRS12/RIM2 (16); YHM1/SHM1 (17); YHM2
(18)). (iv) Limited information from systematic studies is available for a few genes encoding predicted mitochondrial carrier
proteins (19, 20), while (v) nothing is known about the remaining open reading frames.
The MRS3 and MRS4 genes were initially identified in a
screen for multicopy suppressors of a mitochondrial RNA splicing defect (15). Overexpression of either MRS3 or MRS4 suppresses the phenotype caused by depletion of Mrs2p (21), a
mitochondrial inner membrane protein involved in Mg2⫹ transport (22, 23). Mrs3p (314 residues) and Mrs4p (304 residues)
are closely related; they contain 76% identical and 87% similar
amino acids over a stretch of 290 residues. With the exception
of the mitochondrial ADP/ATP carriers (Aac1p, Aac2p, and
Aac3p), which show identities/similarities ranging from 74/88%
(Aac1p versus Aac3p) up to 90/97% (Aac2p versus Aac3p), they
are the pair with the highest sequence similarity within the 35
yeast MCF members. Genome-wide transcription profiling indicated that MRS4 is co-induced with several iron-uptake
40612
This paper is available on line at http://www.jbc.org
Import of Iron into Mitochondria
genes in an AFT2–1up mutant, suggesting a role of this carrier
in mitochondrial metal transport (24). Another hint toward the
function of Mrs3/4p comes from previous work by Foury and
Roganti (25) showing that (i) deletion of MRS3/4 in a frataxin
deletion mutant (⌬yfh1 cells) is able to suppress mitochondrial
iron accumulation, (ii) overexpression of MRS4 in ⌬yfh1 cells
leads to higher mitochondrial iron load, and (iii) overexpression
of MRS4 in wild-type cells results in increased heme formation,
pointing to a role of these carriers in mitochondrial iron
homeostasis.
To elucidate the function of Mrs3/4p we used genome-wide
expression analyses and found that both deletion and overexpression of MRS3/4 lead to induction of the so-called iron regulon in S. cerevisiae. Genes of this regulon are under the control
of the iron-sensing transcription activators Aft1p and Aft2p
that induce gene transcription under iron-limiting conditions
(24, 26 –29). These observations prompted us to investigate the
activities of the two major iron-utilizing processes, i.e. the
biosynthesis of heme from protoporphyrin IX and the formation
of cellular Fe/S proteins, in yeast cells either lacking or overexpressing both MRS3 and MRS4. Both biosynthetic reactions
are performed in the mitochondrial matrix, yet little is known
about how iron is imported into the organelles. This is mainly
due to the fact that no mitochondrial iron transporters have
been identified so far. Furthermore mechanistic studies on
mitochondrial iron transport have been hampered by the fact
that iron avidly associates with membranes (30). Nevertheless,
using heme synthesis and Fe/S protein biogenesis as an indirect, yet faithful measure for iron transport, it was shown
recently that import of ferrous iron was dependent on a membrane potential and occurs independently of ATP (30, 31).
Building on these studies, our investigations presented here
reveal a clear correlation between the cellular MRS3/4 levels
and the efficiency of both heme biosynthesis and Fe/S protein
biogenesis in iron-deprived cells. Hence our findings strongly
suggest a role of these carriers in mitochondrial iron uptake
under iron-limiting conditions.
EXPERIMENTAL PROCEDURES
Yeast Strains and Cell Growth—The following strains of S. cerevisiae
were used: W303-1A (MATa, ura3–1, ade2–1, trp1–1, his3–11,15, leu2–
3,112), which served as wild type, GW344-1b (W303-1A mrs3⌬::LEU2,
termed ⌬mrs3), GW344-3c (W303-1A mrs4⌬::URA3, termed ⌬mrs4),
GW344-6d (W303-1A mrs3⌬::LEU2 mrs4⌬::URA3, termed ⌬mrs3/4),
and ⌬cyt2 (W303-1A, cyt2::LEU2; Ref. 32). Disruption of both MRS3
and MRS4 was performed in the diploid strain W303 using the short
flanking homology gene disruption method (33) and confirmed by PCR
analysis. One diploid (GW344), which was heterozygous for both MRS3
and MRS4, was sporulated, and tetrads were dissected. MATa segregants containing a disrupted MRS3 and/or a disrupted MRS4 gene
were selected and used for further investigations. MRS3 and MRS4
were overexpressed in wild-type cells from plasmids pGW822 (MRS3;
LEU2) and pGW821 (MRS4; URA3) under the control of the ADH1
promoter (see below). Cells were grown in rich (YP (1% yeast extract,
2% peptone)) and minimal (SC) media containing the desired carbon
sources with or without FeCl3 (34) or in lactate medium (35).
Construction of Plasmids Overexpressing MRS3 or MRS4 —For
pGW821, a 1.4-kb SspI-XbaI fragment containing the MRS4 open reading frame and 464 bp of the 3⬘ flanking sequences (note that XbaI site
is not derived from MRS4 genomic sequence) was cloned into the
blunted BamHI and the XbaI site of vector pVT102-U (36). For
pGW822, a 2.5-kb PsiI-SmaI fragment containing the MRS3 open reading frame and 579 bp of the 3⬘ flanking region (note that SmaI site is not
derived from MRS3 genomic sequence) was ligated into the blunted
HindIII restriction site of vector pAAH5 (37).
RNA Isolation—For Northern blot analyses wild-type, ⌬mrs3/4, and
MRS3/4n cells were grown as for in vivo labeling experiments (see
below) and harvested at A600 ⫽ 1, and RNA was prepared using the hot
acidic phenol method (38).
For microarray analyses two sets of experiments were performed.
First, wild-type cells and ⌬mrs3/4 cells were grown in YPD (1% yeast
extract, 2% peptone, 2% glucose) to A600 ⫽ 0.5 and harvested, and RNA
40613
was prepared essentially according to the hot acidic phenol method (38)
with the only difference being that three chloroform extractions were
performed instead of one. Second, W303-1A transformed with pAAH5
and pVT102-U (wild type) and W303-1A transformed with pGW822 and
pGW821 (MRS3/4n) were cultured in SD-Leu⫺Ura⫺ and harvested at
A600 ⫽ 0.5; RNA was prepared as described above.
DNA Microarray Analysis—Yeast DNA arrays were obtained from
the Ontario Cancer Institute Microarray Centre. Reverse transcription,
probe cleanup, and microarray hybridization were performed according
to the manufacturer’s protocol. Microarrays were read using an Axon
GenePix 4000B laser scanner (Axon Instruments) and analyzed using
the GenePix Pro 3.0 software. Experiments were repeated three times
using independent cultures.
55
Fe Incorporation into Fe/S Cluster Apoproteins and Heme in
Vivo—In vivo labeling of yeast cells with radioactive iron (55FeCl3) and
measurement of 55Fe incorporation into Fe/S proteins by immunoprecipitation and liquid scintillation counting was carried out as described
previously (39, 40). The following Fe/S reporter proteins were used:
mitochondrial biotin synthase (Bio2p) and a hemagglutinin (HA)tagged version of cytosolic Rli1p (41, 42), both overexpressed from the
2␮ plasmid p424GPD under the control of the strong constitutive TDH3
promoter (44). For the determination of heme formation in vivo, cells
were radiolabeled with 55Fe for 2 h, cell lysates were extracted with
butyl acetate, and the amount of 55Fe-labeled heme was determined by
liquid scintillation counting as described previously (30).
55
Fe Incorporation into Fe/S Cluster Apoproteins and Heme in
Vitro—The determination of Fe/S protein formation in intact mitochondria and in mitochondrial extracts in vitro was carried out as described
previously using mitochondria isolated from Bio2p-overexpressing,
iron-starved yeast cells (31). Heme formation of isolated mitochondria
was determined as described previously (30). 25 ␮g of mitochondria
were resuspended in 500 ␮l of buffer A (20 mM HEPES, pH 7.4, 50 mM
KCl, 1 mM MgSO4, 1 mM sodium ascorbate, 0.6 M sorbitol) and incubated with 2 mM NADH at 25 °C for 3 min. Energized mitochondria
were subsequently incubated with 2.5 ␮M deuteroporphyrin and 1 ␮Ci
of 55Fe for 10 min at 25 °C. The reaction was stopped by addition of 5
mM FeCl3 and 0.25 M HCl, heme was extracted with butyl acetate, and
the amount of 55Fe-labeled heme was determined by liquid scintillation
counting (30). For each data set, a blank value measured in the presence of 10 ␮M carbonyl cyanide m-chlorophenylhydrazone and in the
absence of added NADH was subtracted (see Ref. 30). For analysis of
heme formation activities of ferrochelatase in mitochondrial extracts,
25 ␮g of mitochondria were incubated in 50 ␮l of buffer A containing
0.15% ␤-dodecylmaltoside for 2 min on ice. The lysed sample was
diluted to 500 ␮l with buffer A and analyzed for 55Fe-labeled heme
formation as described above.
Determination of Mitochondrial Membrane Potential—Relative
membrane potentials of isolated mitochondria were estimated by fluorescence spectroscopy in a Kontron SMF25 spectrometer essentially as
described previously (45). The spectrometer was calibrated by setting
the fluorescence emission at 590 nm of a 0.5 ␮M tetramethylrhodamine
methyl ester (TMRM) solution to 100%. 100 ␮g of iron-starved mitochondria were resuspended in buffer A containing 0.5 ␮M TMRM, and
the observed fluorescence emission at 590 nm was recorded. After 2
min, 1 mM NADH was added, and the fluorescence emission was recorded. These values were subtracted from the value obtained in the
absence of NADH to obtain the amount of fluorescence quenching.
Determination of Elemental Concentrations in Isolated Mitochondria
by ICP-MS and ICP-AES—Mitochondria (400 ␮g of protein) were
treated with 250 ␮l of concentrated nitric acid (69%, w/w, semiconductor quality; BASF, Ludwigshafen, Germany) at 70 °C for 1 h in an
ultrasonic bath. The solution turned brown, and all solid matter dissolved. After cooling to room temperature the dissolution liquid was
transferred into 10-ml polypropylene bottles and diluted with ultrapure
water (Millipore Gradient and Elix, Millipore GmbH, Eschborn, Germany) to a 5-ml total volume. The blank of the nitric acid, the containers, and the buffers was below the limits of detection. The quantification was made using aqueous standards acidified to the same amount of
nitric acid as that present in the dissolved mitochondria samples. The
samples were analyzed using an atomic emission spectrometer (ICPAES Spectroflame P, Spectro A.I., Kleve, Germany) equipped with a
N2-flushed UV monochromator for the measurement of very low wavelengths necessary for sensitive phosphorus determinations. The following lines were used for quantification by atomic emission spectrometry
(ICP-AES): phosphorus, 177.500 nm; phosphorus, 178.290 nm; magnesium, 279.553 nm; manganese, 257.610 nm; zinc, 213.856 nm. The
transition metals occurring at much lower concentrations were analyzed using an elemental mass spectrometer (ICP-MS VG Elemental
40614
Import of Iron into Mitochondria
TABLE I
Genes involved in iron homeostasis are induced in ⌬mrs3/4 cells
Induction/reduction of gene expression is presented as the relative signals of the DNA array analysis observed with RNA extracted from ⌬mrs3/4
and wild-type cells. Standard deviation of three independent experiments is given. Information about gene functions was extracted from MIPS
(Munich Information Center for Protein Sequences) and SGD (Saccharomyces Genome Database) data bases.
Name
Iron homeostasis
FIT1
ARN2
ARN1
TIS11
FET3
FRE3
FTR1
ARN3
Name
Fe/S cluster proteins
LEU1
GLT1
Others, known function
BAT1
RPN10
OAC1
GDH1
GAP1
Gene
Induction
Function
YDR534C
YHL047C
YHL040C
YLR136C
YMR058W
YOR381W
YER145C
YEL065W
10.652 ⫾ 4.016
4.657 ⫾ 2.672
3.279 ⫾ 1.054
2.326 ⫾ 0.518
2.153 ⫾ 0.235
2.032 ⫾ 0.435
1.905 ⫾ 0.173
1.771 ⫾ 0.121
Gene
Induction
YGL009C
YDL171C
0.195 ⫾ 0.069
0.442 ⫾ 0.026
Leucine biosynthesis, Fe/S protein
Glutamate synthase, Fe/S protein
YHR208W
YHR200W
YKL120W
YOR375C
YKR039W
0.325 ⫾ 0.044
0.421 ⫾ 0.062
0.510 ⫾ 0.104
0.583 ⫾ 0.065
0.584 ⫾ 0.051
Branched-chain amino acid aminotransferase, Fe/S cluster biosynthesis?
Ubiquitin-dependent protein degradation
Oxalacetate carrier, sulfate porter
Glutamate dehydrogenase
General amino acid permease
Iron transport
Siderophore transport
Siderophore transport
Putative transcription factor, Aft1p consensus motif
Iron transport
Siderophore transport
Iron transport
Siderophore transport
Function
Plasmaquad 2 Turbo⫹ STE, Thermo, Winsford, UK). The following
isotopes were used for quantification: 26Mg, 55Mn, 57Fe, 59Co, 60Ni, 63Cu,
65
Cu, 64Zn, and 66Zn. Typical concentrations of the elements in solution
were low to mid ng/ml for the metals and ␮g/ml for phosphorus.
Miscellaneous Methods—The following published methods were
used: manipulation of DNA and PCR (46); transformation of yeast cells
(47); isolation of yeast mitochondria (35); Northern blot analyses (46);
preparation of whole cell lysates by mechanical cell disruption with
glass beads (48); immunostaining and immunoprecipitation (49); and
enzyme activities of malate dehydrogenase, aconitase (50), citrate synthase (51), and succinate dehydrogenase (52, 53). The standard error of
the determination of enzyme activities was between 5 and 15%. The
iron content of mitochondria was determined as described previously
(54).
RESULTS
Deletion and Overexpression of Both MRS3 and MRS4
Causes Up-regulation of Genes Involved in Iron Homeostasis—We constructed a series of isogenic strains with identical
auxotrophic markers where MRS3 and MRS4 are either deleted (termed ⌬mrs3/4 cells) or overexpressed under the strong
constitutive ADH1 promoter (termed MRS3/4n cells). Overexpression of both MRS3 and MRS4 leads to a severe growth
defect on media containing ethanol and glycerol as the sole
carbon sources at 16, 28, and 36 °C and at 36 °C on glucosecontaining media. Deletion of either MRS3 or MRS4 caused no
significant phenotype, but simultaneous disruption of both
genes resulted in a temperature-sensitive growth defect on
fermentable and non-fermentable carbon sources (data not
shown).
We investigated the global transcriptional responses to
changes in the expression levels of MRS3 and MRS4 using
DNA microarray technology. Expression profiles of early logarithmic cultures of ⌬mrs3/4 and MRS3/4n mutant cells were
compared with those of isogenic wild-type cells. Genes whose
expression was significantly up- or down-regulated at least
1.5-fold are shown in Table I for the ⌬mrs3/4 mutant and in
Table II for MRS3/4 overexpression. The complete data sets are
available upon request. Interestingly both deletion and overexpression of MRS3 and MRS4 caused induction of genes of the
“iron regulon.” This regulon contains genes for the high affinity
iron transport system constituted by FET3 and FTR1 (55, 56),
genes encoding siderophore transporters (57–59) and facilitators of iron transport (60). In the case of ⌬mrs3/4 cells, no other
genes were induced, while MRS3/4 overexpression caused up-
regulation of several genes involved in, e.g. stress response and
copper homeostasis. This finding is consistent with the more
severe growth phenotype of MRS3/4n as compared with
⌬mrs3/4 cells. Genes down-regulated in ⌬mrs3/4 cells include
LEU1 and GLT1 coding for Fe/S proteins (Table I). In MRS3/4n
cells several genes involved in mitochondrial energy metabolism, as well as genes involved in transcription and translation,
were repressed (Table II). These data indicate that the mitochondrial carrier proteins Mrs3/4p affect cellular iron metabolism either indirectly by performing a regulatory function or
directly by participating in iron transport across the mitochondrial inner membrane.
The Efficiency of Heme Formation in Vivo Is Influenced by
Cellular Levels of MRS3/4 —The up-regulation of genes of the
iron regulon upon deletion or overexpression of MRS3 and
MRS4 prompted us to investigate the major iron-utilizing processes, i.e. the formation of heme from protoporphyrin IX and
the biosynthesis of Fe/S proteins. Both processes take place in
the mitochondrial matrix. First, we determined the effects of
MRS3/4 levels on the de novo formation of heme in vivo. To this
end, wild-type, ⌬mrs3/4, and MRS3/4n cells were cultivated in
iron-free minimal medium and radiolabeled with 55Fe. Cells
were lysed mechanically with glass beads, and newly synthesized radiolabeled heme was determined by extraction into
organic solvents (30). As shown in Fig. 1A, ⌬mrs3/4 cells displayed a small but reproducible decrease (1.4-fold) in the incorporation of 55Fe into heme as compared with wild-type cells.
On the contrary, overexpression of MRS3/4 induced a more
than 2-fold increase in heme formation. Fig. 1B shows a Northern blot of RNAs prepared from iron-starved cells confirming
the overexpression of MRS3 and MRS4 in MRS3/4n cells and
the lack of the respective transcripts in the ⌬mrs3/4 cells. The
correlation of 55Fe incorporation into heme in vivo with the
expression levels of MRS3 and MRS4 points to a potential role
of Mrs3/4p in cellular iron transport and/or iron utilization in
vivo.
The Efficiency of Fe/S Protein Biogenesis in Vivo Is Influenced by MRS3/4 Expression Levels—Next we examined Fe/S
protein biogenesis in wild-type, ⌬mrs3/4, and MRS3/4n cells by
analyzing the de novo incorporation of radioactive 55Fe into
cellular Fe/S proteins in vivo (39, 42). First, the mitochondrial
biotin synthase (Bio2p) was used as a reporter after overpro-
Import of Iron into Mitochondria
40615
TABLE II
Genes involved in iron homeostasis are induced in MRS3/4n cells
Induction/reduction of gene expression is presented as the relative signals of the DNA array analysis observed with RNA extracted from
MRS3/4n and wild-type cells. Standard deviation of three independent experiments is given. Information about gene functions was extracted from
MIPS (Munich Information Center for Protein Sequences) and SGD (Saccharomyces Genome Database) data bases.
Name
Iron homeostasis
FIT2
FIT1
FIT3
ARN2
ARN4
FET3
FTR1
ISU1
Metal ion transport
IST1
CUP1-1
CUP1-2
Stress response
SNZ1
GLK1
Others, known function
TMT1
IDH2
OAC1
BDH1
IDH1
RHR2
ARG1
ARG3
ICY2
BNA1
TDH2
Others, unknown function
Name
Mitochondrial metabolism
MIR1
ATP2
COX4
HSC82
Gene expression
EFT1
SIN3
YEF3
Others, known function
RNR4
Others, unknown function
Gene
Induction
YOR382W
YDR534C
YOR383C
YHL047C
YOL158C
YMR058W
YER145C
YPL135W
16.810 ⫾ 14.099
16.461 ⫾ 9.827
7.632 ⫾ 2.959
5.516 ⫾ 3.636
3.048 ⫾ 0.930
2.841 ⫾ 0.618
2.457 ⫾ 0.324
2.270 ⫾ 0.500
YBR086C
YHR053C
YHR055C
3.042 ⫾ 1.513
2.634 ⫾ 0.795
2.246 ⫾ 0.546
Similar to calcium and sodium channel proteins
Copper-binding metallothionein
Copper-binding metallothionein
YMR096W
YCL040W
3.715 ⫾ 0.839
1.887 ⫾ 0.367
Response to nutrient limitation
Stress response, glycolysis
YER175C
YOR136W
YKL120W
YAL060W
YNL037C
YIL053W
YOL058W
YJL088W
YPL250C
YJR025C
YJR009C
2.872 ⫾ 1.367
2.695 ⫾ 0.430
2.565 ⫾ 0.747
2.293 ⫾ 0.529
2.255 ⫾ 0.379
2.088 ⫾ 0.425
2.027 ⫾ 0.214
1.922 ⫾ 0.225
1.895 ⫾ 0.306
1.852 ⫾ 0.313
1.735 ⫾ 0.201
Trans-aconitate methyltransferase
Mitochondrial isocitrate dehydrogenase
Oxalacetate carrier, sulfate porter
Butanediol dehydrogenase
Mitochondrial isocitrate dehydrogenase
Glycerol metabolism
Arg biosynthesis
Arg biosynthesis
Unknown, Aft1p consensus motif
Nicotinic acid biosynthesis
Glyceraldehyde-3-phosphate dehydrogenase
YDR340W
YHR162W
YGL117W
YLR334C
2.477 ⫾ 0.400
2.322 ⫾ 0.253
1.913 ⫾ 0.213
1.705 ⫾ 0.143
Unknown
Unknown
Unknown
Unknown
Gene
Induction
Function
YJR077C
YJR121W
YGL187C
YMR186W
0.522 ⫾ 0.109
0.529 ⫾ 0.073
0.544 ⫾ 0.088
0.557 ⫾ 0.051
Mitochondrial phosphate transport
Mitochondrial ATPase complex ␤ subunit
Cytochrome c oxidase
Chaperone, plasma membrane cation-transporting ATPase
YOR133W
YOL004W
YLR249W
0.444 ⫾ 0.127
0.457 ⫾ 0.060
0.585 ⫾ 0.063
Translation elongation factor
Transcription factor
Translation elongation factor
YGR180C
0.581 ⫾ 0.072
DNA replication
YBL100C
0.460 ⫾ 0.121
Unknown
duction. Cells were cultivated in iron-poor minimal medium
(SD) and radiolabeled with 55Fe. Subsequently a cell lysate was
prepared, and the 55Fe-labeled Bio2p was immunoprecipitated
with Bio2p-specific antibodies as described previously (42). In
wild-type cells, a high amount of 55Fe could be co-immunoprecipitated using Bio2p-specific antibodies (Fig. 2A), whereas
only background signals (⬍2 pmol of 55Fe/g of cells) were found
in control experiments using nonspecific antisera or cells that
did not overexpress Bio2p (not shown, see Ref. 42). In ⌬mrs3/4
cells, a 2.5-fold reduction in the amount of 55Fe associated with
Bio2p was observed, indicating a defect in the biosynthetic
pathway for mitochondrial Fe/S proteins in vivo. Twice the
amount of 55Fe was co-immunoprecipitated with anti-Bio2p
antibodies from MRS3/4n cells as compared with wild-type
cells. Notably overproduction of proteins of the Fe/S cluster
assembly machinery did not cause any conspicuous alterations
in Fe/S protein biogenesis (see e.g. Refs. 40 and 42, not shown).
Thus, these data indicate that cellular MRS3/4 levels have a
direct influence on the de novo formation of Fe/S proteins in
mitochondria.
Function
Iron transport
Iron transport
Iron transport
Siderophore transport
Siderophore transport
Iron transport
iron transport
Fe/S cluster biogenesis
A similar result was obtained when a hemagglutinin-tagged
version of the cytosolic Fe/S protein Rli1p was used as a reporter (Fig. 2B). In ⌬mrs3/4 cells, the amount of 55Fe association with Rli1p-HA was reduced 3.5-fold as compared with the
wild-type; in MRS3/4n cells it was increased by a factor of 1.7.
Apparently the levels of the two mitochondrial carriers also
influence the maturation of Fe/S proteins in the cytosol. When
the de novo Fe/S cluster incorporation into Bio2p was investigated in the single deletion mutants ⌬mrs3 or ⌬mrs4, no significant differences were observed (Fig. 2C), indicating redundancy of the proteins. Taken together, the in vivo analyses of
heme formation and de novo Fe/S protein formation unequivocally demonstrate that both processes decline upon deletion of
MRS3/4 and significantly increase upon overproduction of
these carrier proteins. These results strongly argue in favor of
a role for Mrs3/4p in the utilization and/or transport of iron in
vivo.
No Influence of MRS3/4 Expression Levels on Iron Utilization
in Mitochondria Isolated from Cells Grown in Iron-replete Medium—To further elucidate the role of Mrs3p and Mrs4p in
40616
Import of Iron into Mitochondria
FIG. 1. Cellular MRS3/4 expression levels correlate with the de
novo formation of heme in vivo. A, wild-type, ⌬mrs3/4, and MRS3/4n
cells were grown in “iron-poor” minimal medium supplemented with
glucose (SD) for 16 h. Cells were radiolabeled with 55Fe for 90 min and
lysed, and heme was extracted by butyl acetate. The amount of 55Felabeled heme was determined by liquid scintillation counting of the
organic extract (30). Error bars indicate the standard deviation of the
measurements. B, wild-type, ⌬mrs3/4, and MRS3/4n cells were grown
in “iron-poor” minimal medium as above and harvested at A600 ⫽ 1.
Total RNA was isolated, and 6-␮g samples were analyzed by RNA blot
hybridization. Note that the weak band above the MRS4 transcript in
RNA extracted from ⌬mrs3/4 cells is caused by cross-hybridization with
RNA running ahead of the 18 S rRNA.
mitochondrial iron transport we determined the enzyme activities of Fe/S cluster-containing enzymes, heme formation activities, and iron content of mitochondria isolated from wild-type,
⌬mrs3/4, and MRS3/4n cells grown in lactate medium (35)
containing standard concentrations of iron. Surprisingly the
mitochondrial preparations of the different strains did not
show any significant differences in the activities of the Fe/S
proteins aconitase and succinate dehydrogenase (Fig. 3A).
Moreover as judged by the bathophenanthroline iron quantification assay, no substantial changes in the iron accumulation
were observed in all three types of mitochondria (not shown).
Finally heme formation, i.e. the insertion of 55Fe into added
deuteroporphyrin by ferrochelatase, was similar in mitochondria isolated from wild-type, ⌬mrs3/4, and MRS3/4n cells (Fig.
3B). Taken together, our analysis showed that the expression
levels of MRS3/4 did not significantly influence the efficiency of
iron-dependent processes in mitochondria isolated from cells
grown in lactate medium. These results are in sharp contrast to
the data reported above on the de novo formation of heme and
Fe/S proteins in vivo. However, those analyses in vivo were
carried out using cells cultivated under iron-limiting conditions, whereas for the in organello analyses, mitochondria were
isolated from cells cultivated in iron-rich medium. Hence we
reason that the influence of MRS3/4 levels on mitochondrial
iron-dependent processes might become apparent only under
iron-limiting conditions.
MRS3/4 Levels Affect Transport of Iron to Ferrochelatase in
Mitochondria Isolated from Iron-starved Cells—To verify a potential role of Mrs3/4p in mitochondrial iron transport, mitochondria were isolated from wild-type, ⌬mrs3/4, and MRS3/4n
cells after cultivation in iron-free medium and were analyzed in
vitro. First, heme formation by intact mitochondria was investigated by following the in organello incorporation of radioactive 55Fe into deuteroporphyrin (30). As shown in Fig. 4A, 55Fe
FIG. 2. Cellular MRS3/4 levels affect the de novo maturation of
mitochondrial and cytosolic Fe/S proteins in vivo. Wild-type,
⌬mrs3/4, and MRS3/4n cells overproducing either Bio2p or a HA-tagged
version of Rli1p were grown and radiolabeled as described in Fig. 1. Cell
extracts were subjected to immunoprecipitations with specific antibodies. The amount of 55Fe that was co-immunoprecipitated with Bio2p (A)
or Rli1p (B) was quantified by liquid scintillation counting. C, wild-type,
⌬mrs3, and ⌬mrs4 cells overproducing Bio2p were analyzed for maturation of Bio2p as above. Error bars indicate S.D. of the measurements.
incorporation into heme was 2-fold lower in mitochondria from
⌬mrs3/4 cells as compared with wild-type organelles, while
mitochondria from cells overexpressing MRS3/4 displayed an
increase by a factor of 1.5. Under our assay conditions, iron
import is the rate-limiting step of heme formation because the
ferrochelatase-catalyzed insertion of iron into deuteroporphyrin is a fast reaction, and ferrochelatase is present in high
excess in mitochondria (30, 61). Therefore, the observed differences in heme formation in the various mitochondrial preparations reflect differences in the rate of iron transport into the
matrix. Consequently we expected that heme formation measured after lysis of the mitochondria should be similar in all
three samples. To test this idea, mitochondria from wild-type,
⌬mrs3/4, and MRS3/4n cells were dissolved in detergent-containing buffer, and the ferrochelatase activities of the detergent
lysates were subsequently determined by our radioassay. The
amount of iron incorporation into deuteroporphyrin was not
significantly different in all three lysates (Fig. 4B). Notably
Import of Iron into Mitochondria
40617
FIG. 5. Mitochondrial membrane potential is not altered in
⌬mrs3/4 or MRS3/4n cells. Organelles were energized with 1 mM
NADH in the presence of 0.5 ␮M TMRM, and the NADH-inducible
fluorescence quenching of TMRM was recorded at 590 nm (45). As a
control, mitochondria from the respiration-deficient strain ⌬cyt2 lacking cytochrome c1 heme lyase were analyzed in parallel. Error bars
indicate the S.D. of the measurements.
FIG. 3. Deletion of MRS3 and MRS4 causes no obvious ironrelated mitochondrial defects in yeast cells cultivated in ironreplete media. Wild-type, ⌬mrs3/4, and MRS3/4n cells were grown in
lactate medium for 16 h. Mitochondria were isolated and analyzed for
the enzyme activities of the Fe/S proteins aconitase and succinate
dehydrogenase (SDH) relative to that of malate dehydrogenase (MDH)
(A) and the formation of radioactive [55Fe]heme (B). Error bars indicate
the standard deviation of the measurements.
FIG. 4. MRS3/4 levels influence the efficiency of iron transport
to ferrochelatase in iron-depleted mitochondria. Mitochondria
were isolated from wild-type, ⌬mrs3/4, and MRS3/4n cells cultivated in
iron-poor minimal medium supplemented with glucose (SD). A, mitochondria were energized with NADH and incubated with 1 ␮Ci of 55Fe
and 2.5 ␮M deuteroporphyrin for 10 min. Heme was extracted by butyl
acetate, and the amount of 55Fe-labeled heme was determined by liquid
scintillation counting. B, mitochondria were lysed in buffer containing
␤-dodecylmaltoside, and the lysate was analyzed for [55Fe]deuteroheme
formation activity. Error bars indicate the standard deviation of the
measurements. The absolute value of heme formation in intact wildtype mitochondria was 3-fold lower as compared with that observed
after detergent lysis. The values for the two other mitochondrial samples were normalized for their respective wild-type controls.
heme formation in lysates of wild-type mitochondria was approximately 3 times more active than that in intact wild-type
mitochondria, substantiating that iron transport is the ratelimiting step in heme formation (30). Moreover, as judged by
immunostaining, ferrochelatase (Hem15p) levels were similar
in mitochondria from the respective strains (see Fig. 6C). In
conclusion, these data unambiguously demonstrate that the
mitochondrial carriers Mrs3p and Mrs4p are involved in the
transport of iron across the mitochondrial inner membrane.
This involvement may either be direct in that Mrs3p and
Mrs4p actively transport ferrous iron themselves or indirect in
that they influence the mitochondrial membrane potential,
which is known to be required for efficient mitochondrial iron
uptake (30).
We therefore determined the relative membrane potential of
the different mitochondrial preparations using a fluorescence
spectrophotometric assay based on the indicator TMRM. This
dye is incorporated into the inner membrane of energized mitochondria, and its fluorescence characteristics can be used as
an absolute measure of the mitochondrial membrane potential
in intact cells and in isolated organelles (45). Since our heme
formation assays were carried out with intact mitochondria
energized by NADH, we determined the NADH-inducible membrane potential of the mitochondria in vitro. This was accomplished by measuring the relative differences of the fluorescence emission of TMRM at 590 nm in the presence and
absence of externally added NADH (45). As shown in Fig. 5, the
observed NADH-inducible fluorescence quenching showed no
significant differences in mitochondria isolated from wild-type,
⌬mrs3/4, and MRS3/4n cells. In contrast, when mitochondria of
the respiration-deficient mutant ⌬cyt2 (32) were analyzed, only
a small change in the fluorescence emission of TMRM was
observed upon addition of NADH. We therefore conclude that
the deletion or overexpression of MRS3/4 did not significantly
alter the mitochondrial membrane potential. Taken together,
these data suggest that Mrs3/4p may be directly involved in
ferrous iron transport. It should be noted, however, that the
iron transport into mitochondria as determined by our heme
formation assay was only reduced, but not completely blocked,
upon deletion of MRS3/4 (cf. Fig. 4). This clearly indicates that
these carriers either do not transport iron themselves and just
regulate the efficiency of transport or do not represent the only
proteins responsible for mitochondrial iron transport. Further,
since iron transport was measured through the formation of
heme, the conclusions of this analysis only apply to the transport of iron to ferrochelatase.
The Efficiency of Fe/S Protein Maturation in Mitochondria
Isolated from Iron-starved Cells Is Influenced by MRS3/4 Expression Levels—To make a more general statement, we studied the de novo incorporation of 55Fe into Fe/S proteins in
iron-depleted mitochondria in organello as an indirect, yet
40618
Import of Iron into Mitochondria
TABLE III
Elemental concentrations of mitochondria isolated
from iron-starved yeast cells
Phosphorus content and citrate synthase activities of the various
mitochondria preparations were measured and used as reference for the
total amount of mitochondria. All values are shown relative to these
numbers together with the standard deviation for four (wild-type cells)
or seven (⌬mrs3/4 and MRS3/4n cells) individual samples.
Ratio
Element to phosphorus
Manganese (mmol/mol)
Iron (mmol/mol)
Nickel (␮mol/mol)
Cobalt (␮mol/mol)
Copper (mmol/mol)
Zinc (mmol/mol)
Element to citrate
synthase activity
Manganese (␮mol/unit)
Iron (mmol/unit)
Nickel (␮mol/unit)
Cobalt (␮mol/unit)
Copper (mmol/unit)
Zinc (mmol/unit)
FIG. 6. MRS3/4 levels influence the efficiency of de novo maturation of mitochondrial Fe/S proteins in vitro. Mitochondria
overproducing Bio2p were isolated from iron-starved wild-type,
⌬mrs3/4, and MRS3/4n cells. A, organelles were radiolabeled with 55Fe
for 3 h under anaerobic conditions in the presence of 0.2 mM cysteine
and 2 mM NADH as described previously (31). Reactions were terminated by addition of EDTA, mitochondria were lysed, and the amount of
55
Fe assembled into Bio2p was determined by immunoprecipitation as
described in Fig. 2. B, mitochondria were lysed with Triton X-100 under
anaerobic conditions, the clarified extracts were radiolabeled with 55Fe
for 3 h, and the amounts of 55Fe assembled into Bio2p were determined
by immunoprecipitation. Error bars indicate the standard deviation of
the measurements. C, immunostaining of the indicated proteins in the
mitochondrial preparations investigated.
faithful measure of iron import into isolated organelles (31).
Intact mitochondria harboring overexpressed Bio2p were incubated with radioactive 55Fe and reisolated by centrifugation,
and radiolabeled Bio2p was subsequently immunoprecipitated
from a clarified detergent lysate. The amount of 55Fe incorporation into Bio2p was 1.7-fold lower in mitochondria from the
⌬mrs3/4 strain as compared with wild-type organelles (Fig.
6A). In mitochondria isolated from MRS3/4n cells, Bio2p was
assembled 5-fold more efficiently relative to the wild-type control. The levels of Bio2p protein were comparable in all three
preparations as judged by immunostaining (Fig. 6C). Hence,
similar to the study of heme formation, the results of the in
organello analysis of Fe/S protein formation are qualitatively
similar to those obtained in vivo (cf. Fig. 2). We therefore
conclude that MRS3/4 levels affect the efficiency of both major
iron-utilizing processes within intact mitochondria under ironlimiting conditions.
To investigate whether the influence of these carriers on
Fe/S protein maturation was due to effects on mitochondrial
iron import or due to modulations of the efficiency of the mitochondrial Fe/S cluster assembly machinery, the maturation of
Bio2p was measured in mitochondrial detergent extracts as
described earlier (31). A mitochondrial detergent lysate was
prepared from iron-starved cells overproducing Bio2p and incubated with 55Fe under anaerobic conditions. The radiolabeled Fe/S holoprotein was isolated by immunoprecipitation,
and Bio2p-associated radioactivity was quantified by scintillation counting. The amount of 55Fe incorporation into Bio2p in
Wild type
⌬mrs3/4
MRS3/4n
0.31 ⫾ 0.01
0.72 ⫾ 0.13
30 ⫾ 7
28 ⫾ 1
0.32 ⫾ 0.02
2.8 ⫾ 0.7
0.20 ⫾ 0.02
0.34 ⫾ 0.21
30 ⫾ 15
14 ⫾ 3
0.27 ⫾ 0.03
1.7 ⫾ 0.4
0.26 ⫾ 0.02
1.77 ⫾ 0.18
47 ⫾ 25
26 ⫾ 4
0.34 ⫾ 0.08
3.8 ⫾ 0.2
120 ⫾ 2
0.28 ⫾ 0.04
12 ⫾ 3
11 ⫾ 0.3
0.12 ⫾ 0.01
1.1 ⫾ 0.3
78 ⫾ 4
0.14 ⫾ 0.09
11 ⫾ 4
5.6 ⫾ 1.0
0.10 ⫾ 0.01
0.6 ⫾ 0.1
101 ⫾ 8
0.67 ⫾ 0.07
18 ⫾ 10
9.8 ⫾ 1.1
0.13 ⫾ 0.03
1.4 ⫾ 0.1
mitochondrial extracts derived from the ⌬mrs3/4 strain was
comparable to wild-type extracts (Fig. 6B). This suggests that
the reduction of Fe/S protein formation observed with intact,
Mrs3/4p-deficient mitochondria was likely caused by a decreased efficiency of iron transport into the organelle. Surprisingly, however, in extracts from MRS3/4n mitochondria, the
amount of 55Fe incorporation into Bio2p remained elevated
(2.5-fold) as compared with wild-type controls. The increased
efficiency of Fe/S cluster formation in intact mitochondria can
therefore only in part be explained by improved mitochondrial
iron uptake upon overexpression of MRS3/4. A likely explanation of this observation may be that Mrs3/4p transport a low
molecular mass compound that may stimulate the activity of
mitochondrial Fe/S cluster assembly. The biochemical nature
of the stimulating component, however, currently remains elusive yet may include chelators of iron or divalent ions other
than ferrous iron.
Determination of Metal Contents in Mitochondria Isolated
from Iron-starved Cells—Finally we determined the steadystate concentrations of biologically relevant divalent metals in
mitochondria isolated from iron-starved wild-type, ⌬mrs3/4,
and MRS3/4n cells. Preparations were mineralized with concentrated nitric acid and subjected to elemental analysis using
an elemental mass spectrometer. Metal contents of the preparations were normalized to either citrate synthase activities or
the amounts of phosphorus as an internal standard. This element proved to be a reliable measure of the amount of mitochondria (Table III). In mitochondria from the ⌬mrs3/4 strain,
the iron content was lower by a factor of 2 as compared with
wild-type cells, whereas in mitochondria from MRS3/4n cells
iron increased 2.5-fold. Apparently the amount of iron correlates well with the MRS3/4 levels of the individual preparations and the respective iron transport activities measured by
heme formation and Fe/S protein biogenesis (see above). These
findings further substantiate the direct involvement of Mrs3/4p
in mitochondrial iron acquisition. Moreover the mitochondrial
zinc content showed a quantitatively similar correlation with
MRS3/4 levels. Cobalt was reduced in mitochondria from
⌬mrs3/4, and nickel contents increased by a factor of 2 upon
overexpression of MRS3/4 (Table III). No significant variations
were observed for copper and manganese. We therefore conclude that Mrs3/4p carrier proteins may also be involved in the
mitochondrial uptake of Zn2⫹ and possibly Mn2⫹ and Co2⫹, i.e.
Import of Iron into Mitochondria
metals with similar ion radii as Fe2⫹. Finally we tested the
possibility whether the presence of Zn2⫹, Mn2⫹, and Co2⫹
might improve Fe/S cluster assembly in mitochondrial extracts.
However, these metals slightly inhibited Bio2p holoprotein assembly in our in vitro assay (not shown, Ref. 31) ruling out that
these metal ions were responsible for the improved Fe/S protein formation in vitro in MRS3/4n mitochondria.
DISCUSSION
Several lines of evidence have previously suggested a role of
the mitochondrial carrier proteins Mrs3p and Mrs4p in metal
ion homeostasis. First, high copy numbers of MRS3 and MRS4
are able to suppress the deletion of MRS2, a gene encoding a
mitochondrial Mg2⫹ channel (21, 23). Second, both deletion and
overexpression of MRS3/4 genes result in complex changes in
heavy metal tolerance or sensitivity of yeast cells (25).2 Third,
MRS4 expression is under control of the iron-dependent transcription factor Aft2p and was found to be induced in ⌬yfh1
cells lacking mitochondrial frataxin, a protein involved in mitochondrial Fe/S protein biogenesis (24, 25, 42, 62, 63). Fourth,
deletion of the two carrier genes abolishes the accumulation of
iron in mitochondria of ⌬yfh1 cells, a hallmark phenotype of
cells deficient in frataxin (25, 64). Finally, overexpression of
MRS4 stimulated heme production in isolated mitochondria
(25). Together these data point toward a possible role of these
carriers in mitochondrial iron metabolism. Moreover, our genome-wide DNA profiling study of cells lacking or overexpressing the MRS3/4 genes further adds to this notion by demonstrating that both conditions lead to the induction of genes of
the iron regulon. Although the sum of these observations
clearly suggests a role of Mrs3/4p in mitochondrial iron homeostasis and/or trafficking, no consistent biochemical characterization has been carried out so far.
In this communication, we have made an effort to substantiate the suspected role of Mrs3/4p in mitochondrial iron acquisition by biochemically evaluating the effects of altered
MRS3/4 levels on the two major iron-utilizing processes in vivo,
in organello (intact mitochondria), and in vitro (lysed mitochondria). Our analyses showed a stimulation of both heme formation and the maturation of cellular Fe/S proteins upon overexpression and partial inhibition upon double deletion of MRS3/4
in vivo. Remarkably MRS3/4 levels influenced the rate of biosynthesis of both mitochondrial and cytosolic Fe/S proteins.
This is a further indication that iron utilized for Fe/S protein
maturation in the cytosol is shuttled through mitochondria and
is not inserted directly into cytosolic Fe/S apoproteins (65).
A similar correlation between the MRS3/4 levels and the
formation of heme and Fe/S proteins was observed in organello
using intact mitochondria from iron-deprived cells, demonstrating that the influence of MRS3/4 levels on heme and Fe/S
protein biogenesis was the result of specific changes taking
place within mitochondria. In case of heme formation, this
correlation was completely lost upon lysis of the mitochondria
with detergent. This unambiguously demonstrates that the
altered heme formation activities in cells and intact mitochondria of ⌬mrs3/4 and MRS3/4n strains were directly caused by
alterations of mitochondrial iron uptake and were not the result of different ferrochelatase levels or activities. We further
show that these effects were not related to changes in the
mitochondrial membrane potential, which has been shown to
be essential for iron uptake (30). Taken together, these data
indicate that the carriers Mrs3/4p are involved in iron transport into mitochondria.
The situation is more complex for Fe/S protein biogenesis as
an indicator of iron import into mitochondria. Similar to the
2
G. Wiesenberger, unpublished observations.
40619
findings on heme formation, the deletion of MRS3/4 resulted in
a decrease in Fe/S protein formation in intact mitochondria,
and this effect disappeared upon detergent lysis of the organelles. However, intact mitochondria and mitochondrial detergent extracts carrying overproduced Mrs3/4p were 5- and
2-fold, respectively, more active in Fe/S protein biogenesis than
wild-type samples. While the results of the MRS3/4 deletion
are compatible with the idea of an iron transport function of the
encoded carriers, the data obtained upon overproduction, while
not being contradictory, are more complex to interpret. How
might these data be explained? The sustained increase in Fe/S
protein biogenesis activity in MRS3/4n mitochondrial extracts
may be accounted for by the 2-fold induction of ISU1, a central
component of Fe/S protein biogenesis, under these conditions
(Table II). However, overexpression of central genes of the Fe/S
cluster assembly machinery including ISU1, ARH1, YFH1,
SSQ1, and JAC1 do not significantly stimulate this process (42,
43).3 It should be noted in passing that Mrs3/4p are the only
proteins known to us that detectably stimulate Fe/S protein
biogenesis above wild-type levels. As an alternative explanation, the stimulatory effect of MRS3/4 overexpression might be
understood by assuming that these carriers facilitate mitochondrial accumulation of components other than iron that are
rate-limiting for Fe/S protein assembly and thus stimulate the
activity of the Fe/S cluster assembly machinery. While the
biochemical nature of such a hypothetical compound remains to
be identified, it is tempting to speculate that a likely candidate
may be an iron chelator. Complex formation with a chelator
may increase the efficiency of iron transport across the mitochondrial inner membrane.
Several indications allow the conclusion that Mrs3/4p might
be involved in transport of other cations. First, high copy numbers of MRS3 and MRS4 can substitute for the lack of Mrs2p,
a protein involved in mitochondrial Mg2⫹ uptake (21, 23). Isolated mitochondria from a mrs2⌬ mutant contain only about
50% of the Mg2⫹ found in wild-type mitochondria, yet overexpression of MRS3 or MRS4 restores the mitochondrial steadystate levels of Mg2⫹ to nearly wild-type levels (66). Second,
deletion of MRS3 and MRS4 leads to a complex metal resistance and hypersensitivity (25).2 These findings may in part be
explained by transport of these ions via Mrs3/4p. However,
since changes in MRS3/4 levels induce the expression of high
affinity iron transporters, these phenotypes may as well reflect
complex changes of the metal transport characteristics of the
plasma membrane. Third, determination of metal contents in
isolated mitochondria from iron-depleted cells using ICP-MS
showed that the mitochondrial iron and zinc contents varied
with the expression levels of MRS3/4 and that the mitochondrial cobalt and manganese contents declined upon deletion of
these genes. Previously other metal transporters such as
Smf2p have been shown to carry various divalent ions across a
membrane (67).
The fact that the iron regulon is induced upon both deletion
and overexpression of MRS3/4 may serve as a further indication for a role in mitochondrial iron transport. Upon deletion of
these transporters, insufficient amounts of iron are delivered to
mitochondria, causing reduced Fe/S protein biosynthesis,
which in turn causes the up-regulation of the iron regulon. In
this respect, the ⌬mrs3/4 strain behaves rather similarly to
mutants with defects in Fe/S cluster biosynthesis (65, 68, 69).
On the other hand, overproduction of Mrs3/4p may cause a
depletion of the cytosolic iron pool since a larger portion of the
cellular iron is transferred to mitochondria, a situation that
also might lead to induction of genes responsible for iron up-
3
U. Mühlenhoff and R. Lill, unpublished data.
40620
Import of Iron into Mitochondria
take (70, 71). Remarkably up-regulation of iron uptake in response to depletion of MRS3/4 is not specific to S. cerevisiae. In
the pathogenic basidiomycete Cryptococcus neoformans a mutation that induces constitutive ferric reductase and iron uptake was shown to be located in the homologue of yeast MRS3/4
(72).
The influence of MRS3/4 levels on mitochondrial heme and
Fe/S protein biosynthesis was observed under iron-limiting conditions only. In mitochondria isolated from cells cultivated in rich
(iron-replete) medium, no differences were found. Apparently
other transporters for iron must exist that fully suffice for mitochondrial iron supply under iron-replete conditions. The conclusion that Mrs3/4p do not represent the central mitochondrial iron
transporters is evident also under iron-limiting conditions where
a significant basal iron uptake was detectable even in the absence of these carrier proteins. It is well possible that Mrs3/4p
levels may also affect iron-related biosynthetic processes under
iron-replete conditions, but in our mutants they obviously are too
small to be detected by our assays. In this context it should be
mentioned, that a small stimulation of heme formation in mitochondria isolated from MRS4 overexpressing cells grown under
iron replete conditions has been reported previously for a different strain background (25).
The combination of in vivo, in organello, and in vitro studies
allowed us to characterize the carriers Mrs3/4p as the first
proteins with a direct function in mitochondrial iron transport.
Whether the proteins import the iron in the form of free ferrous
iron or together with an iron chelator cannot be concluded with
certainty from our studies. Nevertheless our in organello studies demonstrating an influence of Mrs3/4p levels on the efficiency of iron transport may be regarded in favor of direct iron
transport through these proteins since no further additions
were needed to support ferrous iron uptake into mitochondria
(30). Building on our findings presented here, the possibility to
purify and reconstitute carrier proteins into proteoliposomes
will open new avenues to define the exact nature of the transported substrate(s) and the mechanism of transport.
Acknowledgment—We thank Gerhard Adam for critically reading
the manuscript.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
Aquila, H., Link, T. A., and Klingenberg, M. (1985) EMBO J. 4, 2369 –2376
Walker, J. E. (1992) Curr. Opin. Struct. Biol. 2, 519 –526
Palmieri, F. (1994) FEBS Lett. 346, 48 –54
Palmieri, L., Rottensteiner, H., Girzalsky, W., Scarcia, P., Palmieri, F., and
Erdmann, R. (2001) EMBO J. 20, 5049 –5059
El Moualij, B., Duyckaerts, C., Lamotte-Brasseur, J., and Sluse, F. E. (1997)
Yeast 13, 573–581
Belenkiy, R., Haefele, A., Eisen, M. B., and Wohlrab, H. (2000) Biochim.
Biophys. Acta 1467, 207–218
Palmieri, L., Runswick, M. J., Fiermonte, G., Walker, J. E., and Palmieri, F.
(2000) J. Bioenerg. Biomembr. 32, 67–77
Wohlrab, H., and Briggs, C. (1994) Biochemistry 33, 9371–9375
Palmieri, L., De Marco, V., Iacobazzi, V., Palmieri, F., Runswick, M. J., and
Walker, J. E. (1997) FEBS Lett. 410, 447– 451
Palmieri, L., Agrimi, G., Runswick, M. J., Fearnley, I. M., Palmieri, F., and
Walker, J. E. (2001) J. Biol. Chem. 276, 1916 –1922
Palmieri, L., Lasorsa, F. M., De Palma, A., Palmieri, F., Runswick, M. J., and
Walker, J. E. (1997) FEBS Lett. 417, 114 –118
Marobbio, C. M., Vozza, A., Harding, M., Bisaccia, F., Palmieri, F., and Walker,
J. E. (2002) EMBO J. 21, 5653–5661
Tzagoloff, A., Jang, J., Glerum, D. M., and Wu, M. (1996) J. Biol. Chem. 271,
7392–7397
Prohl, C., Pelzer, W., Diekert, K., Kmita, H., Bedekovics, T., Kispal, G., and
Lill, R. (2001) Mol. Cell. Biol. 21, 1089 –1097
Wiesenberger, G., Link, T. A., von Ahsen, U., Waldherr, M., and Schweyen,
R. J. (1991) J. Mol. Biol. 217, 23–37
Van Dyck, E., Jank, B., Ragnini, A., Schweyen, R. J., Duyckaerts, C., Sluse, F.,
and Foury, F. (1995) Mol. Gen. Genet. 246, 426 – 436
Kao, L. R., Megraw, T. L., and Chae, C. B. (1996) Yeast 12, 1239 –1250
Cho, J. H., Ha, S. J., Kao, L. R., Megraw, T. L., and Chae, C. B. (1998) Mol.
Cell. Biol. 18, 5712–5723
19. Lalo, D., Stettler, S., Mariotte, S., Gendreau, E., and Thuriaux, P. (1994) Yeast
10, 523–533
20. Winzeler, E. A., Shoemaker, D. D., Astromoff, A., Liang, H., Anderson, K.,
Andre, B., Bangham, R., Benito, R., Boeke, J. D., Bussey, H., Chu, A. M.,
Connelly, C., Davis, K., Dietrich, F., Dow, S. W., El Bakkoury, M., Foury, F.,
Friend, S. H., Gentalen, E., Giaever, G., Hegemann, J. H., Jones, T., Laub,
M., Liao, H., Davis, R. W., et al. (1999) Science 285, 901–906
21. Waldherr, M., Ragnini, A., Jank, B., Teply, R., Wiesenberger, G., and
Schweyen, R. J. (1993) Curr. Genet. 24, 301–306
22. Bui, D. M., Gregan, J., Jarosch, E., Ragnini, A., and Schweyen, R. J. (1999)
J. Biol. Chem. 274, 20438 –20443
23. Kolisek, M., Zsurka, G., Samaj, J., Weghuber, J., Schweyen, R. J., and Schweigel, M. (2003) EMBO J. 22, 1235–1244
24. Rutherford, J. C., Jaron, S., Ray, E., Brown, P. O., and Winge, D. R. (2001)
Proc. Natl. Acad. Sci. U. S. A. 98, 14322–14327
25. Foury, F., and Roganti, T. (2002) J. Biol. Chem. 277, 24475–24483
26. Yamaguchi-Iwai, Y., Stearman, R., Dancis, A., and Klausner, R. D. (1996)
EMBO J. 15, 3377–3384
27. Casas, C., Aldea, M., Espinet, C., Gallego, C., Gil, R., and Herrero, E. (1997)
Yeast 13, 621– 637
28. Yamaguchi-Iwai, Y., Ueta, R., Fukunaka, A., and Sasaki, R. (2002) J. Biol.
Chem. 277, 18914 –18918
29. Stadler, J. A., and Schweyen, R. J. (2002) J. Biol. Chem. 277, 39649 –39654
30. Lange, H., Kispal, G., and Lill, R. (1999) J. Biol. Chem. 274, 18989 –18996
31. Mühlenhoff, U., Richhardt, N., Gerber, J., and Lill, R. (2002) J. Biol. Chem.
277, 29810 –29816
32. Zollner, A., Rödel, G., and Haid, A. (1992) Eur. J. Biochem. 207, 1093–1100
33. Wach, A., Brachat, A., Rebischung, C., Steiner, S., Pokorni, K., te Heesen, S.,
and Philippsen, P. (1998) in Methods in Microbiology (Brown, A., and Tuite,
M., eds) Vol. 26, pp. 67– 81, Academic Press, San Diego, CA
34. Sherman, F. (1991) Methods Enzymol. 194, 3–21
35. Daum, G., Böhni, P. C., and Schatz, G. (1982) J. Biol. Chem. 257, 13028 –13033
36. Vernet, T., Dignard, D., and Thomas, D. Y. (1987) Gene (Amst.) 52, 225–233
37. Ammerer, G. (1983) Methods Enzymol. 101, 192–201
38. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith,
J. A., and Struhl, K. (eds) (1993) Current Protocols in Molecular Biology,
John Wiley and Sons, Boston
39. Kispal, G., Csere, P., Prohl, C., and Lill, R. (1999) EMBO J. 18, 3981–3989
40. Lange, H., Kispal, G., Kaut, A., and Lill, R. (2000) Proc. Natl. Acad. Sci.
U. S. A. 97, 1050 –1055
41. Lange, H., Lisowsky, T., Gerber, J., Mühlenhoff, U., Kispal, G., and Lill, R.
(2001) EMBO Rep. 2, 715–720
42. Mühlenhoff, U., Richhardt, N., Ristow, M., Kispal, G., and Lill, R. (2002) Hum.
Mol. Genet. 11, 2025–2036
43. Mühlenhoff, U., Richhardt, N., Gerber, J., and Lill, R. (2003) EMBO J. 22, in
press
44. Mumberg, D., Müller, R., and Funk, M. (1995) Gene (Amst.) 156, 119 –122
45. Scaduto, R. C., and Grotyohann, L. W. (1999) Biophys. J. 76, 469 – 477
46. Sambrook, J., and Russell, D. W. (2001) Molecular Cloning: A Laboratory
Manual, 3rd Ed., Cold Spring Harbor Press, Cold Spring Harbor, NY
47. Gietz, D., St. Jean, A., Woods, R. A., and Schiestl, R. H. (1992) Nucleic Acids
Res. 20, 1425
48. Woontner, M., and Jaehning, J. A. (1990) J. Biol. Chem. 265, 8979 – 8982
49. Harlow, E., and Lane, D. (1988) Antibodies: A Laboratory Manual, Cold Spring
Harbor Laboratory, Cold Spring Harbor, NY
50. Kispal, G., Csere, P., Guiard, B., and Lill, R. (1997) FEBS Lett. 418, 346 –350
51. Kispal, G., Rosenkrantz, M., Guarente, L., and Srere, P. A. (1988) J. Biol.
Chem. 263, 11145–11149
52. Robinson, K. M., von Kieckebusch-Guck, A., and Lemire, B. D. (1991) J. Biol.
Chem. 266, 21347–21350
53. Robinson, K. M., and Lemire, B. D. (1995) Methods Enzymol. 260, 34 –51
54. Li, J., Kogan, M., Knight, S. A., Pain, D., and Dancis, A. (1999) J. Biol. Chem.
274, 33025–33034
55. Askwith, C., Eide, D., Van Ho, A., Bernard, P. S., Li, L., Davis-Kaplan, S., Sipe,
D. M., and Kaplan, J. (1994) Cell 76, 403– 410
56. Stearman, R., Yuan, D. S., Yamaguchi-Iwai, Y., Klausner, R. D., and Dancis,
A. (1996) Science 271, 1552–1557
57. Heymann, P., Ernst, J. F., and Winkelmann, G. (1999) Biometals 12, 301–306
58. Heymann, P., Ernst, J. F., and Winkelmann, G. (2000) Biometals 13, 65–72
59. Heymann, P., Ernst, J. F., and Winkelmann, G. (2000) FEMS Microbiol. Lett.
186, 221–227
60. Protchenko, O., Ferea, T., Rashford, J., Tiedeman, J., Brown, P. O., Botstein,
D., and Philpott, C. C. (2001) J. Biol. Chem. 276, 49244 – 49250
61. Labbe-Bois, R., and Camadro, J. M. (1994) in Metal Ions in Fungi (Winkelmann, G., and Winge, D. R., eds) pp. 413– 453, Marcel Dekker Inc., New
York
62. Gerber, J., Mühlenhoff, U., and Lill, R. (2003) EMBO Rep. 4, 906 –911
63. Rutherford, J. C., Jaron, S., and Winge, D. R. (2003) J. Biol. Chem. 278,
27636 –27643
64. Babcock, M., De Silva, D., Oaks, R., Davis-Kaplan, S., Jiralerspong, S.,
Montermini, L., Pandolfo, M., and Kaplan, J. (1997) Science 276,
1709 –1712
65. Lill, R., and Kispal, G. (2000) Trends Biochem. Sci. 25, 352–356
66. Gregan, J., Kolisek, M., and Schweyen, R. J. (2001) Genes Dev. 15, 2229 –2237
67. Luk, E. E., and Culotta, V. C. (2001) J. Biol. Chem. 276, 47556 – 47562
68. Gerber, J., and Lill, R. (2002) Mitochondrion 2, 71– 86
69. Foury, F., and Talibi, D. (2001) J. Biol. Chem. 276, 7762–7768
70. Eide, D. J. (1998) Annu. Rev. Nutr. 18, 441– 469
71. Van Ho, A., Ward, D., and Kaplan, J. (2002) Annu. Rev. Microbiol. 56, 237–261
72. Nyhus, K., Ozaki, L., and Jacobson, E. S. (2002) Med. Mycol. 40, 581–591