JOURNAL OF BACTERIOLOGY, OCt. 1993, p. 6212-6219
Copyright © 1993, American Society for Microbiology
Vol. 175, No. 19
Characterization of the Ferrous Iron Uptake System of
MEIKE KAMMLER, CLAUDIA SCHON, AND KLAUS HANTKE*
Lehrstuhlfiir Mikrobiologie II, Universitat Tubingen, Aufder Morgenstelle 28,
Received 12 April 1993/Accepted 26 July 1993
Escherichia coli has an iron(II) transport system (feo) which may make an important contribution to the iron
supply of the cell under anaerobic conditions. Cloning and sequencing of the iron(II) transport genes revealed
an open reading frame (feoA) possibly coding for a small protein with 75 amino acids and a membrane protein
with 773 amino acids (feoB). The upstream region offeoAB contained a binding site for the regulatory protein
Fur, which acts with iron(ll) as a corepressor in all known iron transport systems ofE. coli. In addition, a Fnr
binding site was identified in the promoter region. The FeoB protein had an apparent molecular mass of70 kDa
in sodium dodecyl sulfate-polyacrylamide gel electrophoresis and was localized in the cytoplasmic membrane.
The sequence revealed regions ofhomologyto ATPases, which indicates that ferrous iron uptake may be ATP
driven. FeoA or FeoB mutants could be complemented by clones with thefeoA orfeoB gene, respectively.
Since iron(III) is practically insoluble at neutral pH, many
aerobic microorganisms secrete siderophores, iron(III) che-
lating compounds, for their iron supply. Six different sidero-
phore-iron(III) transport systems in Escherichia coli have
been sequenced and analyzed, and many more in other
gram-negative bacteria have been characterized (4). These
transport systems share a common structure. A ferric sidero-
phore-specific receptor in the outer membrane delivers its
substrate in an energy-dependent mechanism to the peri-
plasm. The energy is provided by the TonB-ExbB-ExbD
complex (4). Even heme (33) and transferrin iron (6) are
taken up in a TonB-dependent manner. With the help of a
binding protein-dependent transport system, the ferric sid-
erophore crosses the cytoplasmic membrane (4).
Much less is known about the uptake of ferrous iron,
although there are some microorganisms which mainly or
exclusively use ferrous iron for their iron supply. The
best-studied examples are Bifidobacterium bifidum
which is one of the early colonizers of the intestine in
breast-fed infants. For the intracellular pathogen Legionella
pneumophila, no siderophores have been found, and it is
assumed that iron(III) reduced by the bacterium to iron(II) is
the main source of iron (14). The odontopathogen of human
dental caries, Streptococcus mutans, has been postulated to
use only ferrous iron furnished by reductants at the cell
surface (10). In yeast cells, an iron-regulated ferric reductase
which is assumed to supply ferrous iron for an uncharacter-
ized iron(II) uptake system has been cloned and sequenced
(7). All these organisms live in oxygen-restricted environ-
ments where ferrous iron may be available.
A major habitat of the facultative anaerobe E. coli is the
gut, where it helps to maintain anaerobic conditions. From
this point of view, it is not astonishing that E. coli also
possesses a ferrous iron transport system. Mutants ofE. coli
in the ferrous iron uptake system (feo) have been isolated
(12). Recently, it was shown that they are severely impeded
in their ability to colonize the mouse intestine (31).
Here we present the first characterization of an iron(II)
transport system in E. coli.
MATERIALS AND METHODS
Strains and plasmids, constructions, and growth conditions.
Strains are listed in Table 1. The media used were TY (8 g of
tryptone, 5 g of yeast extract, 5 g of NaCl per liter) and
nutrient broth (8 g ofnutrient broth and 5 g ofNaCl per liter).
P1 transductions have been described by Miller (17).
lactosidase was determined by the method of Miller (17).
One unit was defined as 1 nmol of ortho-nitrophenol pro-
duced per min per mg (dry weight). Representative values
from three experiments are given.
Standard methods for plasmid DNA isolation, restriction
endonuclease analyses, and ligations were carried out as
described by Sambrook et al. (25) or according to the
instructions of the suppliers. Phage and plasmid inserts are
shown in Fig. 1. In addition, the vectors pWKS30 (37) and
pBSK (Stratagene) were used for cloning or sequencing.
Plasmids not shown in Fig. 1 are pUH30, which contained
the 2.1-kb XhoI-SalI fragment with the end ofTnS andfeoA'
from phage lambda 4, and pUH20, in which the 4.2-kb
PstI-Kp7nIfragment ofpUH18 was deleted, which resulted in
a truncatedfeoB' gene.
A chromosomalfeoB-lacZ operon fusion was constructed
by inserting the 1-kb KpnI-EcoRV fragment of feoB into
pGP704 (18). lacZ was moved with NotI from pUJ8 (8) into
pBSK. With BamHI, the lacZ-containing fragment was
cloned into the BglII site behind 'feoB'. Orientation of lacZ
was checked by restriction analysis. This plasmid was
moved from strain SM10 lambda pir into MC4100 under
selection for ampicillin and streptomycin resistance, and
strain H5107 was obtained. P1 transduction of feoB::TnS
into H5107 showed removal of the ampicillin resistance and
the lac marker, indicating that the insertion site was infeoB
as expected. fnr-250 was introduced by cotransduction with
zcj-637::TnlO from strain RK5288 (28). H5108 was one of the
transductants which proved to be an fnr mutant on KNO3-
MacConkey plates (29). Both strains were always kept in the
presence of ampicillin to prevent loss of the insert.
Expression and localization of plasmid-encoded proteins.
Expression of proteins has been described (34). DNA frag-
ments were cloned into plasmid pT7-5, pT7-6, or pBSK+
and transcribed by the phage T7 RNA polymerase encoded
by the plasmid pGP1-2. The expression of the T7 RNA
FERROUS IRON UPTAKE SYSTEM OF E. COLI
TABLE 1. E. coli K-12 strains used
araD139 AlacU169 rpsL150 reLA1
flbBS301 deoC1 ptsF25 rbsR
As MC4100 but aroB
As H1443 butfhuF::Aplac Mu
As H1717 butfeoB7
As MC4100 butfhuF::MudX
As H1858 butfeoA::TnS
As H1858 butfeoB::TnS
As H1858 butfeoA::TnS
As MC4100 butfeoB-lacZ
As H5107 butfnr-250
As H1443 but Afur
As H2331 butfeoB::TnS
As H2331 butfeoA::TnS
As H1443 butfur-31 fhuF::Aplac
As MC4100 butfnr-250
aroB tsx malT tonB
tsx malT tonBfeoA::TnS
tsx malT tonBfeoB::TnS
aroB tsx malT tonBfeoA::TnS
aroB tsx malT tonBfeoB::TnS
thi thr leufhuA lacYsupE
polymerase is under the control of the lambdaPL promoter
and the gene for the heat-sensitive lambda repressor c1857.
For localizing FeoB, the procedure as described by Stauden-
maier et al. (27) was used, but without radioactive labeling,
since FeoB was so strongly expressed that it could be
identified by Coomassie blue staining of the gel. Because of
the differential centrifugation used to separate outer and
inner membranes, the cytoplasmic membrane was contami-
nated by outer membrane proteins.
Iron uptake experiments. Strains were grown overnight in
TY medium under anaerobic conditions and inoculated in
nutrient broth medium to about 2 x 108 cells per ml and
incubated at 37°C under aerobic conditions and shaking or
under anaerobic conditions in a filled 100-ml Erlenmeyer
flask without shaking. At about 8 x 108 cells per ml, the
culture was harvested, washed once in M9 medium (17) at
4°C, and kept on ice in M9 medium (0.2% glucose, 0.1 mM
nitrilotriacetate) at 1 x 109 cells per ml. The transport was
started after 5 min for warming up at 37°C by addition of 3
,uM ferrous iron labeled with
contained 300 ,uM 55FeC13 (12MBqq/,mol)and 100 mM
ascorbate to reduce the iron. At appropriate times, samples
filtered on 0.45-p,m-pore-size
washed two times with 2 ml of 0.1 mM LiCl. Incorporated
iron was determined by liquid scintillation counting.
Nucleotide sequence accession number. The accession num-
ber X71063 ECFEOAB was given to the feoAB gene se-
quence by the EMBL data library at Heidelberg, Germany.
sFe. The stock solution
Isolation offeo::Tn5 mutants. Strain H1771feoB7 (12) was
used to isolate feo complementing clones with high- and
low-copy-number vectors. However, the rare complement-
ing clones turned out to be highly unstable. For this reason,
we tried to clone a transposon-inactivatedfeo gene.feo::TnS
Lambda E3C10 (619)
FIG. 1. In the upper part the restriction map of the E. coli
chromosome covered by phage lambda 619 (16) is shown.feoAB is
located at 74.9 min between ompB andmaU. The restriction map of
the insert of phage lambda 4 and the inserts of the derived plasmids
are given. The coding region offeoB on phage lambda 4 is indicated
by an arrow, and the C-terminally truncatedfeoA is indicated by an
arrow under pUH31, which is a subclone derived from phage
lambda 4 (HindIII-AccI fragment). Phage lambda 4 was derived
from EMBL3. Vector of pUH16 was pACYC184 (25), vector for
pUH18 and pUH18E was pT7-6 and for pUH31 was pT7-5 (34), and
vector for pHSG924 was pHSG575 (35).
mutants were isolated by the following procedure. Strain
H1858 contained the iron-regulated reporter gene fluF::
MudlX (operon fusion of lac to fhuF). On MacConkey
lactose plates at high iron concentrations (40 p,M iron
added), this strain gives white colonies because ofrepression
by Fur-Fe2" of the lac fusion. When the free iron in the
medium was complexed by 50 p.M 2,2'-bipyridine, red
colonies appeared, showing derepression of the fhuF-lacZ
fusion. Cotransduction of zge-53::TnlO feoB7 into strain
H1858 led to red colonies at 40 p.M iron added. This
indicated derepression of the fusion gene because of the
missing iron supply caused by the defective ferrous iron
transport system. This Feo phenotype was used to screen a
pool of H1858 TnS mutants for red colonies on MacConkey
lactose medium. However, not only insertions infeo but also
VOL. 175, 1993
KAMMLER ET AL.
TTATTCCACA GCCLACTCA TAATATATTC CGGCAATATT TATCATTTCA TTAACAACTG 60
AAACCTTAAT TAAACATTAG CCAGTClGG TAATTC&CTA TTCGAATTAT ATITCGCTG 120
CGATATUCC TTGlGCCACA TCUCATTC1 GTCAGlTTAT TATTCAAZC AACATTCGCA 180
CACAT$TTAA GTATTGCTGA TG
ATCTCTCCTT TGTTGGCMA TClTCTGGTC TClTGTCGCT GTCUACGCC CC1TGAGGTA 300
GTTATCCAGT TAUTGGAAA CAAlC& CCYATGC&AT ACACTCCGA TACTGCGTGG 360
AAUTCACTG GCT?TTCCG TGAUATCAGC CCGGC&TATC GCCAAU
CT GCTTTCTCTT 420
G F SI
R Q K
GGCATGTTAC CTGGCTCCTC TTTTAATGTG GTGCGCGTCG CTCCACTCGG CGACCOCCTT 480
CATATCG&A C;CGTGTT GAGCT
1A AGATCGC C^TITTAGU540
K DL A
AT1CGGAT AACAACAATG AAAAATTA CCLTT 1C 600
AATTGGTAUT CCAUTTCTG GCUG1CAC GTTA1mUC CAGCTC&CTG GCTCACGTC& 660
I G NP N S
G K T TL F N
Q L T
G S R Q 28
GCGTGTAGGT AACTGGGCTG GCGTT1COGT CGAACGT1A GUAGGGCAUT TCTCCCAC 720
R V G NWA
G V T VE R KE G Q F S T T48
CGATCATCAG GTC&CGCTGG TGG&CCTGCC CGGC&rACTT TCTCTGA6CA CrClTCTC 780
D B QV T LV D L PG T Y
S L T
I S S 68
GCUGACCTCG CTCG&TGAGC MATCGGC
Q T S
TCACT1CATT TTGAGTGGCG ACGCCOGACT 840
L D I
QI A C
L S G
D A D L8S
GCTGATT1AC GTGMGATG CTCTIACCT TG&GOTUC CTGTCCTGA CGCT1CUaC 900
NV V D
A S N
Q L 108
GCTGGAACTC GGCATTCOCT GCATTGTGGC ACTGAACATG CTCGACATTG CCG1GAC
I PI V A
L N K
L DIA E
K Q 128
AUTATTCGT ATTGAA1TTG ATGTC?C GGCGCGTG GGC?GTCCGG TGATCS
D A L S
A R L
G C P
GCTUTGGC GCTCAGCIG GOCOTGITC GCTAT-AGC 1080
R G R
L K L
I D R Y K A 168
TAUCGAGAUT GTGGAICTGG TGCrTACGC AC&G(tGC
CTCUaGAAG CAG1TTC1CT 1140
GGCUUAAGTG ATGCCTTCOG ACATCOCGCT G&AACUCGT CGCTGGCTGG GCCGCAAT 1200
A K V N P
GCTGGAUGGC GATlTCFACA GO4GMXTA CGCMGTGA& GcTCGC ATCTGG&TGC 1260
D A 228
CGCCCTCGCC CGTCTGCTA ATGAGATGGl CGATCCGGCG CTGCACATTG CCGATCG 1320
FIG. 2. Nucleotide sequence of thefeo region and the amino acid sequence of the open reading frame FeoA and the protein FeoB. The
arrowhead indicates the position of TnS.
TCTGTGlTGT GGTAAGCAAC ACGACOGG CAGAACOCAG 1380
CCGTTTCACC ACTGQGGT1G ATATCGT GCTCAAaGT TTCCfCT
ATGTACCTG& TGTTCCGCT GGCTTCAC ATCGGOGG CGTTAC&GcC 1500
V N Y
L N F
GCTGTTTGAC GTCGGCTCCG TGGCGCTUT TGTGCATGGT ATTCUTGG& TTGGCTAC&C 1560
V L GI Q
GCTCC&CTTC CCGACTGGC TGACTlTCTT CCTCGCIG GGCTGGG GCGGCATTIA 1620
L A Q
G L G
ClCOGTGCTG CCACTGGGC CGCWATTGG ClTGA!C CTGTCCTCT CCTTCTTGA 1680
P L V
TATAYOCO GTGCGCGT TGIATMG C CGTCTGAC AGGWCTGGG 1740
CTTWXGGGG AAATCCFITG TTtCMCTGAT CGTCOGTC GGOGT
L P G K S
V P L
N G A
CGTACGC TGATCCC=CO TG&AkCgG TG&CAC %TT
D A P R
N N A
GTTT1T&TCC ZTCGGCGCGC GTCYGGCTT CTTCG
C G A
F A V
GUAGU;COGT GCOGGCOG TTCTCOCT GT
Q I GA
GTOC GGTkTfGTG O
V F S
I VN A
G1CTG G ATGCTCAAT CCTCUT GCGCO&A GMOLG TTCTOGA 2040
A T P
F V N
GTTO!XGGTC TATCATGTAC C1CACGTT11 AAGCCGTT ATCC&GACCT GGCAGOGTCT 2100
LP VY I V
P I V K
Q T N
TC GTCAGCT TMGL =
TTTCACAGC TTCTCG& C4IGGWA CGTCO&TkkC ATCAACT CGGOC
SF S L
S G K I
L A 5
V D I
S V S
T P V F K
T C&AGCCATT GGMCTOCTAUGATllCTG 2280
GrAGGCAACG GTTGGCC%T TTIC&GGC CATGOC
Q A TV G L F
GLGGTlG TgOGGACGCT 2340
E V VV
CAACAlCCC TAClOOGCAG kASATAIT
NTLY T A
GGACAGAG TFC&A= C&GUTTIAA 2400
E NI Q
CCTCGGTGAA GAGCTCA GTGCTAGA TGATOCCTGG CLGAQCCTG& ACACOCT
K D TF
insertions in tonBorfifrcould have led to red colonies. For
this reason red colonies were pooled, and phage P1 was
grown on these cells and used to transduce strain H1756
zge-53::TnlO to neomycin resistance. zge-53::TnlO was
known to be 90% cotransducible withfeoB7. The neomycin-
resistant transductants were tested for tetracycline resis-
tance, and those whichwere sensitive were assumed to carry
a feo::TnS mutation. By retransduction into strain H1858,
the mutant phenotype (derepression offhuF) was confirmed.
The mutants S067feoA::TnS and S074feoB::TnS were used
for further studies (for the distinction between feoA and
feoB, see below).
Cloning and sequencing offeo. In many attempts we were
unable to clone thefeo genes by complementation in strain
H1771. Very few white colonies were found on MacConkey
plates after introduction of different E. coli gene banks. The
clones isolated were unstable. Therefore, the gene region
was cloned into a lambda vector. Chromosomal DNA of
S067feoA::TnS was partially digested by Sau3A, and frag-
ments of about 20 kb were ligated into the vector lambda
EMBL3. The clone lambda 4 containing TnS was identified
by hybridization to a TnS DNA probe. DNA from both sides
ofTnS was cloned into the plasmids pUH31 and pUH16 (Fig.
1). Further subcloning into the vector pBSK and exonucle-
ase III digestion generated an ordered set of deletions which
was used to sequence thefeo gene region (Fig. 2). The open
reading frame feoA had been interrupted by the inserted
TnS. The restriction map of the phage lambda 4 insert
indicated thatfeo should be on phage 619 (E3C10) from the
Kohara collection (16). Hybridization with the 7.7-kb PstI
fragment of pUH18 confirmed the location of feo on this
phage. This result was in contrast to the previous mapping
locatingfeo near 38 min on the genetic map of E. coli (12).
Another trial to clone the whole feo gene region with
BamHI-EcoRI from the phage 619 into the vector pBSK
failed. Also, it was not possible to clone the KpnI-BamHI
fragment with a shortened feoB gene into a high-copy-
FERROUS IRON UPTAKE SYSTEM OF E. COLI 6215
CAGCCTTAGC GT1CTGATGA ACCQCATTGA AG0CAGCAA
GGCGACGGCG AATGGGTIC 2520
E M GT 648
CGGGGCGATG GGCGTGATGG ATCAGAAATT CGGTAGCGCA GC&GCAGCTT ACAGCTACCT 2580
G A N G V X
D Q KFGS AA A AYSYL 668
GATTTTCGTC CTGCTGTATG TACCATGTAT CTCGGTGATG GGGGCT1TCG CCCGTGAATC 2640
LY V P C
G AIA RE S 688
AAGCCGTGGC TGGATGGGCT TCTCC&TCCT GTGGGGGCTG AATATCGCFT ACTCACTGGC 2700
S R GS
W G LN
I AY SL A 708
AACATTGTTC TATCAGGTCG CC1GCT1CAG TCAGCATCCA ACTT1C1GQC TGGTGTGCU T 2760
TY Q VA S Y S
Q E PT Y SL V CI
TCEGGCGGT? ATCCTGTTTA ACTCGGG TATCGG
L A V
CTGCGCCG CGCGTAG00G 2820
L R R
FNI V VI G L
A R S R 748
GGTGGATATC GAACTGCTGG CACCCGCA GTCGGTAGC AGTTGCFCG
A T R K
CACCGGTAT TGQCCTTAAT GGCTTCLCTT ATTCAGGTGC GCG1mTGCT GG0GMTfCG92940
A A S T 768
E L L
S V S
S C C
T1GTTGGAAGNI&GATIAGCCAG ACUTTGMAC& CTCAGCC AATGTTIAC 3000
GOCCTGCTGC AACAACTGGA AAGTATGGGC AAUCCG
GGAGTCrGGA AG&&CCTCGC 3060
CTGGCAGTTG TAAAUGCTGC CCGGlAGGUA AGCCTGTCT GCGCGAGTGG 3120
TIGGOCIC GTTACCTTI CTCCATCM TGO =
TYGCGTGST G CTCC&TT T
AATGCI1CT C1ZGUC GG0GT11T 3240
CCTTTGUAG CUTCTCGCG GGATGAAUC TCGCTALTAC ACAGGTGTGG AGTGGCGCGT 3300
AGAGTCGCGG CAT?C&ACA ACAGGGAAG GAACG0CCTG AGC&
number vector or in the low-copy-number vector pWKS30
The 924-bp SspI fragment containing thefeoA gene region
was cloned from phage 619 into plasmid pHSG924 (Fig. 1).
The DNA fragment overlapping the TnS insertion site was
sequenced and showed complete agreement with the previ-
ously determined sequences at the transposon insertion site,
indicating that no deletions or rearrangements had occurred.
No obvious promoter region with a -10 and -35 motif
was found. However, a site with homology to Fnr binding
sites was detected (127 to 148) in that region. Fnr is an
activator of anaerobic energy generating systems like fuma-
rate and nitrate reductases. In addition, further downstream
a binding site for the regulatory protein Fur (Fur box 202 to
220) was detected.
An open reading frame of 75 amino acids, named FeoA,
and a calculated molecular mass of 8,371 Da started at bp
334. feoB started at bp 578, 16 bp after the stop codon of
feoA. feoB coded for a protein of 773 amino acids with a
calculated molecular mass of 84,473 Da. By the method of
Klein et al. (15), FeoB was predicted to be a membrane
protein with eight hydrophobic transmembrane helices. Two
sequences (1 to 21 and 79 to 100) showed homologies to
nucleotide binding sites of ATPases (see Fig. 7). Their
distance of 58 amino acid residues was in the range found in
other ATPases. It was likely that ferrous iron transport was
driven by ATP hydrolysis.
Two possible termination loops are indicated after the stop
A comparison of bp 2515 to 3349 revealed identity to the
sequence of bioH (bp 1742 to 2576 as deposited at EMBL
). This placesfeoAB between bioH and the ompB operon
at 74.9 min on the genetic map ofE. coli. The reasons for the
wrong mapping are not clear. We were unable in some
strains to cotransduce zge-53::TnlO with aroB, which points
FIG. 3. Expression of FeoB. Plasmids w
strain WM1576, and the proteins encoded were expressedwith the
helpof the T7 RNApolymerase-promoter system (34). (A) Autora-
diogram of rifampin-treated cells labeled with methionine.
Proteins of whole cells were separated by sodium dodecyl sulfate-
polyacrylamide gel electrophoresis (SDS-PAGE). Lane 1, WM1576
(pUH18); lane 2, WM1576(pUH2O)with a truncatedfeoB gene. (B)
Proteins of whole cells were separated by SDS-PAGE and stained
with Coomassie blue. ThepositionofFeoB at anapparentmolecular
mass of about 70 kDa is indicated by an arrow. Lane 1, molecular
mass standards(phosphorylase b, 97 kDa; bovine serum albumin, 67
kDa; ovalbumin, 45 kDa; carbonic anhydrase, 29 kDa); lane 2,
pUH18 feoB; lane 3, pUH18E feoB; lane 4, pBSK 18-18 feoB
mutant; lane 5, pBSK18E (insert frompUH18E, T7 promoter at 5'
end offeoB) feoB'; lane 6, pBKS E18 (T7 promoter at 3' end of
feoB) phenotype FeoB', but no expression offeoB from the T7
promoter; lane 7, pBSK vector. (C) Localization of FeoB in
WM1576. Shown are outer membranes of cells with pUH3O (lane 1)
and pUHl8 (lane 2); cytoplasmic membranes, contaminated with
outer membrane proteins, of cells with pUH3O (lane 3) and pUHl8
(lane 4); and soluble proteins of cells with pUH3O (lane 5) and
pUH18 (lane 6). FeoB (arrow) is visible only in the cytoplasmic
membrane fraction (lane 4).
torearrangements in the chromosome of the mutant studied
as a possible reason for this mistake.
Identification of the gene products. For the expression of
the encoded proteins, the system of Tabor and Richardson
(34)was used where the genes of interest are cloned behind
thegene 4)10 promoter ofphageT7 and efficienttranscription
is obtained with the Ti RNA polymerase.WM1576(pUH18)
was labeled with [35S]methionine, and FeoB was identified
as a protein with an apparent molecular mass of about 70
kDa(Fig. 3A). FeoB ishighly expressedwith the Ti system,
and even in whole cells without radioactive labeling the
proteincan be identified (Fig. 3B).Theproteinwas localized
in thecytoplasmic membrane (Fig. 3C).
With the plasmid pUH31 containing the truncated open
reading frame of
fe7 A, aprotein of about 9 kDa was found.
However, since we did not observe this protein with the
plasmid pHSG924 up to now, further experiments are nec-
essary to see whether the open reading frame is translated
Phenotype offeo mutants. To test the importanceoffeoA
andfeoB for the growth of cells, the mutationsfeoA::TnS
andfeoB::TnS were introduced into the mutant BR158 aroB
tonB. This strain is unable to produce its own siderophore
enterochelin (enterobactin) because of the arol mutation,
and it is unable totransport iron(III) siderophoresbecause of
VOL. 175, 1993
KAMMLER ET AL.
G A T A A T G A G A A
C T T T A T A A T A A
A T T A T T G A T A A
T A C A A A C A A A A
T G T A A G G A A A A
G A A T A T G A T T G
G A A A A T G A G A A
T A T A A T G A T A C
G A T A A T C A T T T
T C A T T A T T
T C A T T C T C
C T A T T T G C
T T A T T C G C
T A A T T C T T
C T A T T T G C
G C A T T A T T
G C A T T A T C
T C A A T A T C
G A T A A T G A T A A T C A T T A T C
A G A A A C C A T T C T C A T T A T C
FIG. 4. Growth of H5125 tonB feoA (A), H5126 tonBfeoB (B),
H5127 tonB aroBfeoA (C), and H5128 tonB aroB feoB (D) on TY
medium plates under aerobic conditions.
the tonB mutation. P1 transduction from strains S067
feoA::TnS and SO74feoB::TnS into strain BR158 and selec-
tion for neomycin resistance led to many unexpectedly well
growing transductants. The transductants were then further
cotransduced with feo. No feo aroB transductants were
detected. Cotransduction rate of both feo loci with aroB+
into strain H1443 was only 13 to 18%. The reason for the
good growth of the aroB+ tonB feo strains was not the
secretion of enterochelin into the medium, since it has been
shown that iron(III)-enterochelin uptake is TonB dependent.
dent uptake of iron has been described (11). This seems to be
the main source of iron for thefeo tonB double mutants.
To construct a feo tonB aroB triple mutant, the double
mutantfeo aroB was chosen as a donor for the transduction
into BR158. The transductants were streaked for single
colonies on nutrient broth. The tonBfeo aroB mutant formed
microcolonies on this medium, while the tonB feo aroB+
mutant was able to grow to large colonies. On TY medium,
the tonB feo aroB strains were able to form colonies, but
they were much smaller than the tonB feo strain colonies
(Fig. 4). There was no major difference in the phenotype
between thefeoA::TnS and thefeoB::Tn5 mutants. Growth
offeoB tonB aroB mutants on TY and nutrient broth could
be stimulated by the addition of 10 mM sodium citrate or by
high concentrations of iron-loaded 2,3-dihydroxybenzoate.
Regulation offeo by iron. Ferrous iron uptake was shown
to be repressible by iron. In addition, ferrous iron uptake
was derepressed in afur mutant (12). A Fur binding site was
found in the upstream region of feo by comparison to the
consensus sequence (4). Of the 19 bp, 12 were identical to
the consensus sequence (Fig. 5). The nonidentical base pairs
were also found in other well-established Fur binding sites,
except for a C instead of a T in position 6.
The presence of a Fur binding site was also demonstrated
in vivo. Strain H1717 with fhuF::lambda plac Mu as a
reporter gene is repressed by Fur on MacConkey agar with
40 ,uM iron added. By introducing a Fur binding site on a
high-copy-number plasmid, the low level of Fur protein is
titrated out, the reporter genefhuF is derepressed (30), and
it was found that aroB+ had been
a dihydroxybenzoate-promoted TonB-indepen-
A A A A A T C G A T C T C G T C A A A T T T
A A C C A T C A T T C A T A T C A A A T T T
A A T G A C G C A T G A A A T C A C G T T T
C C C T T T G A T A C C G A A C A A T A A T
A C T C T T G A T C G T T A T C A A T T C C
G A A T T T G A T T T A C A T C A A T A A G
FIG. 5. Sequence comparisons with the Fur box and the Fnr box
in front offeo. The sources of the Fur box sequences are given by
Braun and Hantke (4), and the sources of the Fnr boxes are given by
Eiglmeier et al. (9).
A A A * T T G A T *
A A C C T T G A G C C A C A T C A A C A T T
* * * A T C A A * T T T
the transformants grow as red colonies on the MacConkey
agar plates with iron. Strain H1717 was transformed with the
plasmid pBKS31 carrying the insert of pUH31 or with
pHSG924. Transformants with the high-copy-number plas-
mid pBKS31 yielded red colonies, while transformants with
the low-copy-number plasmid pHSG924 remained white.
This, together with the known sequence, indicated that there
is a Fur box on plasmid pBKS31 able to titrate the Fur
receptor in strain H1717 (30).
Influence of Fnr on feo expression. Fnr is the transcrip-
tional activator of anaerobic respiratory genes. Fe(II) has
been shown to influence the activity of Fnr. Severe iron
limitation led to a reduced activity of Fnr (21). A homology
to Fnr binding sites was found in the upstream region offeo
(Fig. 2 and 5). However, no canonical -10 region was
detected in a distance of 18 to 24 bp as was described for six
Fnr-activated genes by Eiglmeier et al. (9).
To gain more insight into the regulation offeo by Fnr, an
operon fusion feo-lacZ was constructed. The 1.3-kb KpnI-
EcoRV fragment offeoB was cloned into pGP704 (18). lacZ
from plasmid pJU8 (8) was cloned behind the 'feoB' frag-
ment. pGP704 is not able to replicate in a cell without the
presence of the pir gene in trans and can be used for the
generation of insertion mutants (18). The 'feoB'-lacZ-con-
taining plasmid was crossed into MC4100. Those cells in
which the homology of the 'feoB' fragment led to integration
of the pGP704 derivative into the chromosome were selected
with ampicillin. Although lacY is missing in this strain, low
iron-regulated P-galactosidase activity was observed with
strain H5107feoB-lacZ on MacConkey agar plates. ,B-Galac-
tosidase activity was tested from cells grown under anaero-
bic conditions in TY medium with 1% KNO3 added. In the
presence of iron, 72 U of ,3-galactosidase was produced in
strain H5107feoB-lacZ, while in strain H5108feoB-lacZfnr,
14 U were found. Inactivation of fnr led to a fivefold
3-galactosidase activity in the presence of iron,
indicating that Fnr is acting as an activator forfeo. With 50
,uM of dipyridyl, 85 U ofP-galactosidasewas observed in
strain H5108 feoB-lacZ fnr, demonstrating a sixfold dere-
pression probably mediated by Fur.
Complementation and iron transport. Plasmids pUH18,
pUH31, and pHSG924 were transformed into different feo
mutants. Strain H1771 feoB7 and strain S074 feoB::TnS
were complemented by pUH18, growing as white colonies
FERROUS IRON UPTAKE SYSTEM OF E. COLI
pmolFe/mg dry weight
FIG. 6. Ferrous iron uptake. Cells were grown under aerobic conditions (A) or anaerobic conditions (B and C) in nutrient broth medium,
and uptake of"Fe2" was measured. (A and B) Strain H1717 (V), strain H1771feoB7 (e), strain H1771feoB7(pBSK18EfeoB+) (e), and strain
SO67feoA::TnS(+)results are shown. (C) Iron uptake by thefollowingfiurmutants: strain H2331Afitr (V), strain H5101 AfurfeoB::Tn5 (U),
and H5102 AfurfeoA::TnS (*).
on MacConkey lactose agar plates with 40 ,uM iron. Plasmid
pUH31 was unable to complement anyfeo mutant. Plasmid
pHSG924 was able to complement the strains S067, S068,
S069, S070, S071, S072, and S073. However, this com-
plementation was only partially effective, since single colo-
nies were red to weakly red, while in more crowded regions
of the plate the colonies were white. The reason may be a
polar effect of the Tn5 insertion on the expression offeoB.
Ferrous iron uptake was significantly lower in H1771feoB
than in the parent strain H1717 (Fig. 6). Complementation
with pBSK18E led to a very high uptake of iron and was not
very much influenced by the aerobic or anaerobic growth
conditions. However, in feoA::TnS mutants, ferrous iron
uptake was only partly reduced. Since these results were
highly variable, the double mutants H5102AfitrfeoA::TnS
and H5101 AfiirfeoB::TnS were constructed in strain H2331
Aftir. Thefur mutation was introduced to obtain a constitu-
tive high expression of thefeo genes and to avoid variations
due to differing amounts of iron in the growth medium. The
iron uptake in thefeoA mutant was about 60% as high as in
the parent strain (Fig. 6), confirming the results obtained
with the fur' strains.
We had great difficulties in cloning the feo gene region.
There are two possible reasons for this failure. High expres-
sion of an integral membrane protein has often been shown
to be deleterious for the cells. The other, less likely, possi-
bility is that an uncontrolled high expression of the FeoB
ATPase activity leads to an energy deprivation.
This is the first ferrous iron transport system studied at the
molecular level. At first glance, FeoA (75 amino acid resi-
dues) and FeoB (773 amino acid residues) seemed to be
comparable to the cadmium export system (cadAC), in
which one relatively large membrane protein, CadA (727
amino acid residues), and CadC (122 amino acid residues)
catalyze the ATP-driven export of Cd2" (22). This transport
system is a member of a family of cation-translocating E1E2
ATPases, including K+- and Ca2+-ATPases (26). However,
no sequence homologies to these and any other proteins
were detected in the data base. A search for specific sites
indicated that there are two domains infeoB with homology
to nucleotide binding sites (Fig. 7) as described by Walker et
al. (36). This type of structure is found in eucaryotic and
bacterial F-type ATPases and in many transport-related
ATPases (traffic ATPases), as has been analyzed by Mimura
et al. (19). It is interesting to note that the binding protein-
dependent transport systems, including the iron(III) sidero-
phore transport systems, also belong to this superfamily (5).
However, obvious homologies between FeoB and the traffic
ATPases were observed only in the phosphate binding
domain and not in the nucleotide binding fold, which is more
similar to the domains of the E. coli ATPase and phospho-
fructokinase (36). How the transport of ferrous iron via Feo
is energized has not been tested, but the observed similari-
ties lead to the prediction that the uptake is driven by ATP.
The iron regulation of this system by Fur was observed
during the first characterization of the system (12) and was
confirmed by further experiments. The suggested Fur bind-
ing site deviates in position 6 from the consensus, where the
highly conserved T is replaced by a C (Fig. 5). Further
experiments are necessary to prove this suggestion and to
see why the C is tolerated. Similarly, for the suggested Fnr
box no convincing -10 region has been found, since the
TATIAT in position 158 (Fig. 2) seems to have an unallow-
able short spacing of 9 bp to the Fnr box. There is no doubt
that Fnr is activating the expression offeo, however. Fer-
rous iron uptake is certainly important for the iron supply of
the cells. This was demonstrated in a tonB mutant which is
unable to use iron(III) siderophores. However, the depen-
dence on feo could be demonstrated on nutrient broth and
TY medium plates only in an aroB background. Enteroche-
lin, the siderophore of E. coli, can be secreted by the aro+
strains, but they should not be able to use it, because of the
tonB mutation. This contradiction can be explained by a
TonB-independent 2,3-dihydroxybenzoate iron uptake sys-
tem (11). The aro+ strains are able to produce 2,3-dihydroxy-
benzoate as a precursor of enterochelin.
VOL. 175, 1993
KAMMLER ET AL.
G K I
G K V
Q R E
S E I
G B I
N Q V
G E F
G D V
G S R
G K V
K L T
GLF * G
GLF * G
I FVV G
TA I V G
G A G V G K T V F I M
G A G V G K T V N M M
D R G T G KTLA L A I
G P G S G K G I Q C E
P E S S G K T T L T L
P S G C G K S T L L R
P S G C G K S T-L L R
S S G S G K S T F L R
P S G S G K T T L_L R
P N G C G K S T L L N
H N G S G K S T-L L K
N P N S G K T TL F N
Adenylate kinase 106
FB D Q E § Q D V TTL.L F L
F B D * E g R D V
D * RE D A L. T T
K * I g Q P TLLL
L E K * H
H Y I L SD A DL
N I F R F T Q A
. F V D N I Y R Y T L A
D LS K Q A V A
V D AG P E T M T
I Q G LLV V I G G a G S Y Q G A
A S N L E R
FIG. 7. ATP binding site homologies to ATPases and traffic ATPases. The upper part shows the alignment to the phosphate binding
domain (5, 36). The lower part shows the alignment to part of the nucleotide binding fold (36). No obvious similarities of FeoB to the traffic
ATPases (19) were seen in this region.
The triple mutant forms only microcolonies on nutrient
broth; however, growth could be obtained after citrate was
added. Citrate may also be the reason for the slight growth
observed on TY medium plates, since it is known that the
ferric-citrate uptake system is induced above 0.1 mM citrate
(13), and induction is observed in TY medium. The low
molecular mass of the ferric iron dicitrate complex of 443 Da
may allow diffusion of this complex through the outer
membrane, independent of FecA, as has also been observed
in E. coli for the cloned Sfu iron uptake system of Serratia
marcescens (1). This iron transport system has been shown
to use ferric iron, but the way that iron(III) reaches the
periplasm is not known. Proteins homologous to SfuA have
been detected in Neisseria spp., and it is tempting to
speculate that in this organism the ferric iron comes from
transferrin, which is bound by a TonB-dependent receptor to
the cell surface (2, 6).
Ferrous iron uptake under oxygen-limited conditions in
the intestine seems to be more important for E. coli than
ferric iron transport. Recently, it was shown that feo mu-
tants were less efficient in the colonization of the mouse
intestine than feo+ strains (31). Iron(III) uptake by entero-
chelin, citrate, or aerobactin did not influence the colonizing
abilities of the strains tested (32). Similarly, it has been
shown for aerobactin-defective mutants ofShigella spp. that
they are not altered in their virulence (20), and it has been
speculated that heme may be a source of iron for these
intracellular pathogens (24). We think that also, in this case,
ferrous iron has to be considered as an important additional
The technical assistance ofM. D6ring is gratefully acknowledged.
We thank I. Stojiljkovic for help in construction ofthe lac fusion and
discussion, R. Koebnik for advice at the computer, and G. Wein-
stock, E. Bremer, and J. Kohara for sending strains and phages. We
are also grateful to J. Crosa and V. Braun for critical reading of the
The work was supported by the Deutsche Forschungsgemein-
1. Angerer, A., B. Klupp, and V. Braun. 1992. Iron transport
systems of Serratia marcescens. J. Bacteriol. 174:1378-1387.
2. Berish, S. A., C.-Y. Chen, T. A. Mietzner, and S. A. Morse.
1992. Expression of a functional neisserial fbp gene in Esche-
richia coli. Mol. Microbiol. 6:2607-2615.
3. Bezkorovainy, A., L. Soberg, M. Poch, and R. Miller-Catchpole.
1987. Ferrous iron uptake by Bifidobacterium bifidum var.
pennsylvanicum: the effect of metals and metabolic inhibitors.
Int. J. Biochem. 19:517-522.
4. Braun, V., and K. Hantke. 1991. Genetics of bacterial iron
transport, p. 107-138. In G. Winkelmann (ed.), Handbook of
microbial iron chelates. CRC Press, Inc., Boca Raton, Fla.
5. Burkhardt, R., and V. Braun. 1987. Nucleotide sequence of the
fhuC andfhuD genes involved in iron(III) hydroxamate trans-
port: domains in FhuC homologous to ATP-binding proteins.
Mol. Gen. Genet. 209:49-55.
6. Cornelissen, C. N., G. D. Biswas, J. Tsai, D. K. Paruchuri, S. A.
Thompson, and P. F. Sparling. 1992. Gonococcal transferrin-
binding protein 1 is required for transferrin utilization and is
homologous to TonB-dependent outer membrane receptors. J.
7. Dancis, A., D. G. Roman, G. J. Anderson, A. G. Hinnebusch,
and R. D. Klausner. 1992. Ferric reductase of Saccharomyces
cerevisiae: molecular characterization, role of iron uptake, and
transcriptional control by iron. Proc. Natl. Acad. Sci. USA
8. deLorenzo, V., M. Herrero, U. Jakubzik, and K. N. Timmis.
1990. Mini-TnS transposon derivatives for insertion mutagene-
sis, promoter probing, and chromosomal insertion of cloned
DNA in gram-negative eubacteria. J. Bacteriol. 172:6568-6572.
9. Eiglmeler, K., N. Honore, S. Iuchi, E. C. C. [in, and S. T. Cole.
1989. Molecular genetic analysis of FNR-dependent promoters.
Mol. Microbiol. 3:869-878.
10. Evans, S. L., J. E. L. Arcenaux, B. R. Byers, M. E. Martin, and
N. Aranha. 1986. Ferrous iron transport in Streptococcus mu-
tans. J. Bacteriol. 168:1096-1099.
11. Hancock, R. E. W., K. Hantke, and V. Braun. 1977. Iron
transport in Escherichia coli K-12: 2,3-dihydroxybenzoate-pro-
moted iron uptake. Arch. Microbiol. 114:231-239.
12. Hantke, K. 1987. Ferrous iron transport mutants in Escherichia
coli K-12. FEMS Microbiol. Lett. 44:53-57.
13. Hussein, S., K. Hantke, and V. Braun. 1981. Citrate dependent
iron transport system in Escherichia coli. Eur. J. Biochem.
FERROUS IRON UPTAKE SYSTEM OF E. COLI6219 Download full-text
14. Johnson, W., L. Varner, and M. Poch. 1991. Acquisition of iron
by Legionella pneumophila: role of iron reductase. Infect.
15. Klein, P., M. Kanehisa, and C. DeLisi. 1985. The detection and
classification of membrane-spanning proteins. Biochim. Bio-
phys. Acta 815:468-476.
16. Kohara, Y., K. Aidyama, and K. Isono. 1987. The physical map
of the whole E. coli chromosome: application of a new strategy
for rapid analysis and sorting of a large genomic library. Cell
17. Miller, J. H. 1972. Experiments in molecular genetics. Cold
Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
18. Miller, V. L., and J. J. Mekalanos. 1988. A novel suicide vector
and its use in construction of insertion mutations: osmoregula-
tion of outer membrane proteins and virulence determinants in
Vibrio cholerae requires toxR. J. Bacteriol. 170:2575-2583.
19. Mimura, C. S., S. R. Holbrook, and G. F.-L. Ames. 1991.
Structural model of the nucleotide-binding conserved compo-
nent of periplasmic permeases. Proc. Natl. Acad. Sci. USA
20. Nassif, X., M.-C. Mazert, J. Mounier, and P. J. Sansonetti. 1987.
Evaluation with an iuc::TnlO mutant of the role of aerobactin
production in the virulence of Shigella flexneri. Infect. Immun.
21. Niehaus, F., K. Hantke, and G. Unden. 1991. Iron content and
FNR-dependent gene regulation in Escherichia coli. FEMS
22. Nucifora, G., L. Chu, T. K. Misra, and S. Silver. 1989. Cadmium
resistance from Staphylococcus aureus plasmid pI258 cadA4
gene results from a cadmium-efflux ATPase. Proc. Natl. Acad.
Sci. USA 86:3544-3548.
23. O'Reagen, M., R. Gloeckler, C. Ledoux, I. Ohsawa, and Y.
Lemoine. 1989. Nucleotide sequence of the bioH gene of Esch-
enichia coli. Nucleic Acids Res. 17:8004.
24. Payne, S. M. 1989. Iron and virulence in Shigella. Mol. Micro-
25. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular
cloning: a laboratory manual, 2nd ed. Cold Spring Harbor
Laboratory, Cold Spring Harbor, N.Y.
26. Silver, S., G. Nucifora, L. Chu, and T. K. Misra. 1989. Bacterial
resistance ATPases: primary pumps for exporting toxic cations
and anions. Trends Biochem. Sci. 14:76-80.
27. Staudenmaier, H., B. Van Hove, Z. Yaraghi, and V. Braun.
1989. Nucleotide sequences ofthefecBCDE genes and locations
of the proteins suggest a periplasmic-binding-protein-dependent
transport mechanism for iron(III) dicitrate in Escherichia coli. J.
28. Stewart, V. 1982. Requirement of Fnr and NarL functions for
nitrate reductase expression in Escherichia coli. J. Bacteriol.
29. Stewart, V., and C. M. MacGregor. 1982. Nitrate reductase in
Eschenichia coli K-12: involvement ofchiC, chlE, and chiG loci.
J. Bacteriol. 151:788-799.
30. StojiUkovic, I., A. Baumler, and K. Hantke. Fur regulon in
gram-negative bacteria: identification and characterization of
new iron-regulated Escherichia coli genes by a Fur titration
assay. Submitted for publication.
31. Stojilikovic, I., M. CobeUic, and K. Hantke. 1993. Eschenichia
coli K-12 ferrous iron uptake mutants are impaired in their
ability to colonize the mouse intestine. FEMS Microbiol. Lett.
32. Stojiljkovic, I., M. Cobeljic, Z. Trgovcevic, and E. Salaj-Smic.
1991. The ability of rifampin-resistant Escherichia coli to colo-
nize the mouse intestine is enhanced by the presence of a
plasmid-encoded aerobactin-iron(III) uptake system. FEMS Mi-
crobiol. Lett. 90:89-94.
33. Stojiljkovic, I., and K. Hantke. 1992. Hemin uptake system of
Yersinia enterocolitica: similarities with other TonB-dependent
systems in Gram-negative bacteria. EMBO J. 11:4359-4367.
34. Tabor, S., and C. C. Richardson. 1985. A bacteriophage T7
RNA polymerase/promoter system for controlled exclusive ex-
pression of specific genes. Proc. Natl. Acad. Sci. USA 82:1074-
35. Takeshita, S., M. Sato, M. Toba, W. Masahashi, and T. Hashi-
moto-Goto.1987. High-copy-number and low-copy-number
plasmid vectors for lacZa-complementation and chlorampheni-
col- or kanamycin-resistance selection. Gene 61:63-74.
36. Walker, J. E., M. Saraste, M. J. Runswick, and J. N. Gay. 1982.
Distantly related sequences in the a- and a-subunits of ATP
synthase, myosin, kinases and other ATP-requiring enzymes
and a common nucleotide binding fold. EMBO J. 8:945-951.
37. Wang, R. F., and S. R. Kushner. 1991. Construction ofversatile
low-copy-number vectors for cloning, sequencing, and gene
expression in Escherichia coli. Gene 100:195-199.
VOL. 175, 1993