Discovery and characterization of a unique
mycobacterial heme acquisition system
Michael V. Tulliusa,1, Christine A. Harmstonb,1, Cedric P. Owensb, Nicholas Chimb, Robert P. Morseb, Lisa M. McMathb,
Angelina Iniguezb, Jacqueline M. Kimmeya, Michael R. Sawayac, Julian P. Whitelegged, Marcus A. Horwitza,
and Celia W. Gouldingb,e,2
Departments ofbMolecular Biology and Biochemistry andePharmaceutical Sciences, University of California, Irvine, CA 92697;
Department of Medicine, School of Medicine, University of California, Los Angeles, CA 90095;cUniversity of California–Department of Energy, Institute of
Genomics and Proteomics, Los Angeles, CA 90095-1570; anddThe Pasarow Mass Spectrometry Laboratory, Semel Institute for Neuroscience and Human
Behavior, David Geffen School of Medicine, University of California, Los Angeles, CA 90024
aDivision of Infectious Diseases,
Edited* by David S. Eisenberg, University of California, Los Angeles, CA, and approved February 11, 2011 (received for review July 7, 2010)
Mycobacterium tuberculosis must import iron from its host for
survival, and its siderophore-dependent iron acquisition pathways
are well established. Here we demonstrate a newly characterized
pathway, whereby M. tuberculosis can use free heme and heme
from hemoglobin as an iron source. Significantly, we identified the
genomic region, Rv0202c–Rv0207c, responsible for the passage of
heme iron across the mycobacterial membrane. Key players of this
heme uptake system were characterized including a secreted pro-
tein and two transmembrane proteins, all three specific to myco-
bacteria. Furthermore, the crystal structure of the key heme carrier
protein Rv0203 was found to have a unique fold. The discovery of
a unique mycobacterial heme acquisition pathway opens new ave-
nues of exploration into mycobacterial therapeutics.
X-ray crystallography|iron uptake
multidrug resistance and the AIDS pandemic, which has resulted
in a large increase in persons highly susceptible to tuberculosis.
There is an urgent need to understand the pathogen’s survival
tactics. Iron, essential for pathogen survival, is sequestered from
its host by well-characterized siderophore-mediated iron acqui-
sition pathways. Pathogenic mycobacteria including Mtb synthe-
size two types of siderophores—the cell-bound water-insoluble
mycobactins and the secreted water-soluble exochelins, which
have a mycobactin-like backbone (1, 2), and are referred to as
exomycobactins hereafter. Exomycobactins, the most abundant
molecules secreted by Mtb on a molar basis, can remove iron from
human transferrin and lactoferrin and transport it to either
mycobactins in the Mtb cell wall (3) or the iron transport system,
IrtA/B (4). However, there are disadvantages to using side-
rophore-mediated iron uptake pathways alone. First, mycobactin
and exomycobactin production requires at least 11 different
enzymes; hence mycobacteria must devote considerable meta-
bolic energy to synthesize them (5). Second, transferrin iron
accounts for <1% of the body’s total iron whereas heme iron
accounts for >80% (6). A recent study suggests that there is an
alternative mycobacterial iron uptake pathway. A recombinant
bacillus Calmette-Guérin (rBCG) mutant with an interrupted
mycobactin/exomycobactin biosynthetic pathway surprisingly mul-
tiplied, albeit at a rate considerably less than the parental strain, in
SCID mice (7). This result led us to postulate that another type of
iron uptake pathway exists. Given that a number of pathogenic
bacteria acquire iron from the largest source, heme iron (8), it is
highly plausible that mycobacteria may use a heme uptake system.
Over the past two decades, it has become evident that heme is
a major source of iron for both Gram-negative and Gram-positive
bacteria (9, 10). An extensive bioinformatics search with known
membrane-bound and cell wall-associated heme/hemoglobin re-
ceptors, heme transporter, or secreted heme scavenging proteins
(hemophores) found no homologs in mycobacterial proteomes.
However, past studies suggest the presence of an uncharacterized
he human pathogen, Mycobacterium tuberculosis (Mtb), cur-
heme acquisition system in mycobacteria. First, the addition of
exogenous heme to an Mtb mutant with an interrupted heme
mutant growth (11). Second, the addition of hemoglobin can in-
crease growth of mycobacteria (12). Third, a gallium substituted
heme derivative is toxic to Mycobacterium smegmatis cells, sug-
gesting that gallium-metalloporphyrin is acquired by the myco-
bacteria and is either used in the cell wall environment, thus
preventing electron transfer, or broken down in the cytoplasm to
release toxic metal (13).
Here we show that Mtb has a mechanism to use exogenous
heme as its iron source. Identification of the genomic region re-
sponsible for heme uptake reveals several mycobacterial-specific
proteins including a secreted heme-binding protein with a unique
fold. Mycobacterial iron acquisition is potentially a therapeutic
target (14); the identification of all possible avenues of iron up-
take, and thus the discovery of a unique mycobacterial heme up-
take system, is central to the development of therapeutic strat-
egies targeting iron acquisition.
Mtb Uses Heme Iron. To determine whether Mtb has a functional
heme uptake system, we constructed an Mtb mutant deficient in
iron uptake, MtbΔmbtB, by disrupting mbtB (thereby disrupting
mycobactin/exomycobactin biosynthesis) (Fig. S1A). Similar to
BCGΔmbtB (7), MtbΔmbtB does not grow in 7H9 media con-
taining ∼130 μM Fe3+(as ferric ammonium citrate), unless sup-
plemented with exogenous mycobactin (Fig. 1 A and B). However,
in the absence of mycobactin, MtbΔmbtB displays near maximal
growth in the presence of submicromolar concentrations of heme
or human hemoglobin (Fig. 1 A and B), confirming that Mtb has
a heme-iron uptake system, with the ability to scavenge both free
heme and heme bound to hemoglobin as an iron source. This
system also functions in wild-type Mtb that produces mycobactins
and exomycobactins, as evidenced by its growth in iron-depleted
media in the presence of heme or hemoglobin, but not in their
absence (Fig. S2).
Identification and Characterization of a Putative Hemophore. In an
attempt to identify proteins involved in mycobacterial heme
uptake, we used heme–agarose affinity chromatography to pull
down secreted heme-binding proteins that may be potential
Author contributions: M.V.T., M.A.H. and C.W.G. designed research; M.V.T., C.A.H., C.P.O.,
N.C., R.P.M., L.M.M., A.I., J.M.K., and C.W.G. performed research; M.V.T, N.C., M.R.S., J.P.W.,
M.A.H, and C.W.G. analyzed data; and M.V.T., N.C., M.A.H., and C.W.G. wrote the paper.
The authors declare no conflict of interest.
*This Direct Submission article had a prearranged editor.
Data deposition: The atomic coordinates have been deposited in the Protein Data Bank,
www.rcsb.org/pdb (PDB ID code 3MAY).
1M.V.T. and C.A.H. contributed equally to this work.
2To whom correspondence should be addressed. E-mail: firstname.lastname@example.org.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
| March 22, 2011
| vol. 108
| no. 12
hemophores. We incubated Mtb culture filtrate containing se-
creted proteins with heme–agarose beads, washed the beads,
and eluted bound proteins under denaturing conditions. SDS/
PAGE showed only one eluted protein band, which we analyzed
correspond to Rv0203, an uncharacterized mycobacteria-specific
protein (Fig. 1C). The gene Rv0203 encodes for a protein with
a signal peptide (15). To investigate the effect of Rv0203 on
the ability of Mtb to scavenge heme, we deleted Rv0203 from
MtbΔmbtB to create a double-deletion mutant, which was con-
and its own promoter (Fig. S1C). Growth of MtbΔmbtBΔRv0203
in 7H9 medium supplemented with heme was reduced 83% com-
pared with MtbΔmbtB and reduced 61% in 7H9 medium supple-
of heme or hemoglobin, the growth rate of MtbΔmbtBΔRv0203
was ∼50% slower than that of MtbΔmbtB (Table S1). Analysis of
mRNA expression of genes immediately flanking deleted Rv0203
demonstrated that the deletion had no negative polar effect, as
the genes were expressed and even up-regulated compared with
MtbΔmbtB (Table S2). Furthermore, expression of Rv0203 from
a plasmid fully restored growth of MtbΔmbtBΔRv0203 in the pres-
ence of heme and hemoglobin (Fig. 2B, Fig. S3, and Table S3) and
restored secretion of Rv0203 protein into the culture filtrate (Fig.
S3C). Together, these results demonstrate that Rv0203 plays an
responsible for the attenuated phenotype.
To confirm that Rv0203 binds heme, we purified the re-
combinant mature protein (lacking the signal peptide) from
Escherichia coli and determined its heme binding capabilities.
Apo-Rv0203 was incubated with 1.5-M equivalents of heme, and
excess heme was removed by gel filtration chromatography. The
protein eluted in one brownish-yellow colored peak. The heme–
Rv0203 complex was confirmed by absorption spectroscopy,
which revealed a maximum Soret band (γ-peak) at 407 nm that
upon reduction with sodium hydrosulfite shifted to 421 nm, with
the emergence of distinct β- and α-peaks at 526 and 557 nm, re-
spectively, indicative of a heme binding protein (Fig. 3A). Titra-
tion of heme into Rv0203 resulted in a 1:1 stoichiometry of heme
toRv0203 (Fig.3B), andits heme-binding affinity was determined
to have a Kdof 5.4 × 10−6M by isothermal calorimetry. In com-
parison, hemophores from Bacillus anthracis IsdX1 and Porphyr-
omonas gingivalis HmuY have Kds of ∼10−6M (16, 17), whereas
the hemophore from Serratia marcescens, HasA, has a heme af-
finity Kdof ∼10−10M (18). Thus, the heme binding affinity ob-
served for Rv0203 is consistent with that of other known bacterial
hemophores. Also, circular dichroism (CD) analysis shows no ap-
upon the addition of heme (Fig. 3C). Thus, we have shown that
Rv0203 is a heme-binding protein.
Structure of the Putative Hemophore. To further characterize
Rv0203,wesolved the crystal structure by multiwavelength anom-
alous diffraction (MAD) phasing to 2.5 Å, yielding an Rfree/Rwork
of 19.5/25.6 (Table S4). A structural homology search using the
Dali server indicates that Rv0203 has a unique fold and is highly
atypical among heme transport proteins (19). The predominately
α-helical structure consists of a dimer of dimers, where each
monomer consists of five α-helices (Fig. 3D). Within each mono-
mer, the two longest α-helices (α1 and α4) are tethered at the N
terminus of α1 and the C terminus of α4 by a disulfide bond be-
tween Cys40 and Cys114 and capped by two short helices (α2 and
α3) with the α5 helix (α5) lying perpendicularly across α1 and α4
helices. The dimer is formed by antiparallel interactions between
the α1 helices, and the four helices of the dimer form an anti-
parallel helical sheet, with the two shorter helices capping each
end of the dimer, forming a catcher’s mitt-like structure. Two
dimers associate to produce an off-tilt cage-like structure, where
the α5 helices form a weak hydrophobic core in the center of the
cage (Fig. 3D and Fig. S4A). The monomers in the dimer su-
perimpose on each other to give a root mean square of 2.0 Å. The
most obvious differences between these two polypeptide chains
(0μM Hm orHb).Growth was determinedbymeasuring absorbance at 750 nm. Three independent experiments were performed with similar results. (C) SDS/
PAGE of Mtb culture filtrate fractions eluted from a hemin–agarose column. Lane 1, molecular weight marker; lane 2, purified recombinant Rv0203; lanes
3 and 4, eluted fractions.
Mtb uses heme as an iron source and secretes a heme-binding protein. (A and B) MtbΔmbtB was grown in 7H9–0.01% tyloxapol media containing
Rv0203, and membrane protein,
MmpL11, facilitate heme uptake.
grown in 7H9–0.01% tyloxapol me-
dia containing mycobactin J (myc)
(10 ng/mL), 0.4 μM heme (0.4 μM
Hm), 0.2 μM human hemoglobin
(0.2 μM Hb), or no added supple-
ment (No Sup.). Growth was de-
at 750 nm. Experiments were per-
formed three to five times with similar results. (B) Complementation analysis of MtbΔmbtBΔRv0203 and MtbΔmbtBΔmmpL11 in studies in which the non-
complemented strains contain the control plasmid pNBV1. Growth and measurement conditions were the same as in A although the medium contained
50 μg/mL hygromycin. Three independent experiments were performed, and growth was compared as described in Table 1. The data are displayed as
percentage of growth (mean ± SE) relative to MtbΔmbtB pNBV1.
| www.pnas.org/cgi/doi/10.1073/pnas.1009516108Tullius et al.
are (i) the loop region connecting α2 and α3 and (ii) the loop
region connecting α4 to α5, causing the α5 helices not to super-
impose upon one another (Fig. S4B).
The Mtb hemophore, Rv0203, is a mycobacteria-specific protein
and has a unique fold with an unusual self-association. However,
comparison with the structurally divergent S. marcescens hemo-
phore, HasA (20) (Fig. 3D and Fig. S4C), reveals a similar heme-
binding motif in Rv0203 (Fig. 3E and Fig. S4E), which is in-
dicative of convergent evolution akin to the catalytic triads found
in various, otherwise unrelated, serine proteases (21). The
ligands that coordinate heme iron in HasA are His32 and Tyr75
(4.8 Å apart), which is hydrogen-bonded to His83 (Fig. S4E).
Within the structure of Rv0203, there is a similar potential heme-
binding motif, where Tyr59 and His89 are 7.8 Å apart and Tyr59
ishydrogenbonded toHis63(Fig.3E).To investigate theeffectof
potential heme-binding residues on Rv0203, we introduced point
mutations Tyr59Ala, His63Ala, and His89Ala (and Met56Ala
and Met127Ala as controls) into Rv0203. Each Rv0203 mutant
protein was purified,bound with heme, and further purified by gel
filtration chromatography. Compared with Rv0203 and other
mutants, Rv0203-Tyr59Ala does not bind heme (Fig. 3F). Addi-
tionally, absorbance spectra of Rv0203 and mutants revealed
Soret bands of 407 nm for all but Rv0203-Tyr59Ala (Fig. S4F).
CD analysis confirmed that Rv0203-Tyr59Ala is mainly α-helical
and folded as observed for Rv0203 (Fig. S4G). Thus, Tyr59
appears to play an important role in heme binding in Rv0203, and
binding of heme predominately through a tyrosine axial ligand is
not unprecedented. Biochemical and structural analysis of a per-
iplasmic heme-binding protein, ShuT, indicates that one tyrosine
residue is the axial heme-iron ligand (22, 23), and heme iron is
five-coordinated by ShuT. In addition, the heme iron in two other
heme-binding proteins, IsdA and IsdC, involved in Staphylococ-
residue (24, 25). These results infer that Tyr59 is critical for heme
binding within the mycobacterial potential hemophore, Rv0203.
reduced heme-Rv0203 (blue line). The Inset shows the magnified Q-band region between 450 and 650 nm. (B) Titration of heme to apo-Rv0203 (10 μM) is
observed by the spectral change at 405 nm between protein plus heme and heme alone. The protein saturates at 10 μM heme. (C) Circular dichroism spectrum
of apo-Rv0203 (black line) and heme-Rv0203 (orange line). (D) Tetrameric structure of Rv0203. Each polypeptide chain is colored differently and the disulfide
bond is in stick representation, where the carbon and sulfur atoms are colored orange and yellow, respectively. (E) The putative heme-binding site with Tyr59,
His63, His89 in stick representation, where carbon, nitrogen, and oxygen atoms are colored white, blue, and red, respectively. (F) Wild type and mutants of
heme-bound Rv0203 after gel-filtration chromatography. Rv0203-Y59A is colorless, indicating that Y59 is critical for heme binding.
Mycobacteria-specific Rv0203 binds heme and has a unique structural fold. (A) The UV/vis absorbance spectra of heme-Rv0203 (orange line) and
attenuated for growth in the presence of heme and hemoglobin
MtbΔmbtBΔRv0203 and MtbΔmbtBΔmmpL11 are
% growth compared with
MtbΔmbtB, mean ± SE*
17 ± 6 (P = 0.002)§
39 ± 8 (P = 0.01)
13 ± 3 (P = 0.0001)
30 ± 8 (P = 0.03)
19 ± 7 (P = 0.002)
40 ± 9 (P = 0.04)
13 ± 3 (P = 0.0001)
33 ± 12 (P = 0.06)
*Mean of four or five experiments (heme) or three experiments (hemoglo-
†Growth of MtbΔmbtBΔRv0203 or MtbΔmbtBΔmmpL11 in the presence of
heme or hemoglobin (0.2 or 0.4 μM) compared with growth of MtbΔmbtB
under the same conditions measured at maximal difference, with the data
fit to a standard sigmoid growth model (SI Methods).
‡Growth of each strain in the presence of heme or hemoglobin after nor-
malization to growth in the presence of mycobactin J.
§Statistical analysis performed using a paired, two-tailed Student’s t test.
Tullius et al. PNAS
| March 22, 2011
| vol. 108
| no. 12
Identification of Transmembrane Heme Transporters. Unlike other
organisms, such as Yersinia pestis, S. aureus, B. anthracis, and
Pseudomonas aeruginosa, where most of the genes encoding pro-
(16, 26–28), in Mtb, Rv0203 is not located within an operon.
However, Rv0203 is surrounded by genes that encode for pre-
dicted transmembrane proteins from Rv0201c to Rv0207c (Fig.
4A). Of particular interest are Rv0202c (MmpL11) and Rv0206c
(MmpL3), which both belong to a family of 13 mycobacterial
transmembrane proteins (MmpLs) and are proposed to be in-
volved in transport of molecules across the mycobacterial mem-
brane (29–31). Thus, we postulate that MmpL11 and MmpL3 are
both heme transporters. Previously, it was shown that mmpL3 is
likely to be essential (32), and our attempts to delete the potential
heme uptake region (Rv0201c–Rv0207c) or mmpL3 alone from
MmpL11 in heme uptake, a double mutant with deleted Rv0202c
(mmpL11) was constructed by replacing most of the mmpL11
coding region with an apramycin resistance cassette and its pro-
of MtbΔmbtBΔmmpL11 in 7H9 media supplemented with heme
or hemoglobin was reduced compared with MtbΔmbtB (heme,
87% reduction; hemoglobin, 70% reduction; Fig. 2A and Table 1)
and the growth rate was ∼50% slower than that of MtbΔmbtB
of mRNA expression of the genes immediately flanking deleted
mmpL11 revealed that they were expressed and up-regulated com-
a plasmid partially restored growth of the MtbΔmbtBΔmmpL11
Table S3), which may be a consequence of the non-iron–regulated
promoter on the plasmid controlling the expression of mmpL11.
Taken together, these results suggest that MmpL11 plays a signifi-
cant role in heme uptake. MmpL3 has 25% sequence identity to
MmpL11. Given the sequence homology and their close genomic
in the transport of heme across the mycobacterial membrane.
Genomic Region Essential for Heme Uptake. To further investigate
the potential heme uptake region, we turned to the model or-
ganism,M.smegmatis,which isasoil-dwelling Mycobacterium(33)
with alternate iron pathways compared with intracellular patho-
genic mycobacteria. M. smegmatis has a similar genomic organi-
zation to the corresponding proposed heme uptake region in Mtb
(Fig. 4A). To determine whether these genes are essential for
heme uptake in M. smegmatis, we constructed a knockout mutant
to tolerate deletion of MmpL3 and thus the entire proposed
were passaged in iron-deplete 7H9 medium to remove intra-
cellular iron and growth under iron-deplete, 1-μM iron or heme
conditions was monitored (Fig. 4B). No growth was observed
for either WT or MsmegΔhemeuptake under iron-deplete con-
of either iron or heme. However, MsmegΔhemeuptake grew sub-
stantially in media supplemented with iron but not heme. Impor-
tantly, complementation of MsmegΔhemeuptake with a plasmid
containing the Mtb genes inclusive of Rv0202c–Rv0207c fully re-
stored growth in the presence of heme to that of WT (Fig.
4B). These results suggest that the Mtb genomic region between
Rv0202c and Rv0207c encodes for key proteins involved in myco-
bacterial heme uptake.
To confirm the hypothesis that MmpL11 and MmpL3 physi-
cally bind heme for transportation across the membrane, we
purified the soluble domains of each potential heme transporter
and determined their heme binding capabilities. We predicted
the domain boundaries of the two extracellular (E1 and E2) and
cytoplasmic (C1) soluble domains (Fig. 5A and Fig. S4H) and
cloned, expressed, and purified them (Fig. 5B). Each domain was
incubated with heme, subjected to gel filtration chromatography,
smegmatis homologs, Msmeg0242 and Msmeg0243, are depicted with solid arrows. The dashed arrows depict genes with high sequence homology between
the two genomes. Note that the corresponding genomic region in M. smegmatis contains all of the genes in the genomic region of Mtb surrounding
Rv0203. (B) Growth curves of M. smegmatis mc2155 (WT), MsmegΔhemeuptake (genes inclusive of Msmeg0240–Msmeg0251, designated KO) and
MsmegΔhemeuptake::Rv0202c-Rv0207c (complemented with Mtb genes inclusive of Rv0202c–Rv0207c, designated KO+Mtb complement) in 7H9–0.01%
tyloxapol-NoFe medium supplemented with 1 μM FeCl3, 1 μM heme, or no supplement (No Sup).
Use of M. smegmatis to identify Mtb genes involved in heme uptake. (A) Mtb and M. smegmatis genomic regions surrounding Rv0203. Rv0203 and M.
| www.pnas.org/cgi/doi/10.1073/pnas.1009516108 Tullius et al.
and then analyzed by UV/vis spectroscopy. For both MmpL11
andMmpL3,thespectra ofextracellular domainshadSoretpeaks
between 399 and 410 nm, indicative of heme-binding protein
domains. In contrast, the cytoplasmic domains showed no evi-
dence of heme binding (Fig. 5 C and D). These results imply
extracellular to intracellular heme transport and offer concrete
support of the previous proposal that several MmpL proteins may
be involved in mycobacterial metabolite import (32). The high
sequence homology and close genomic proximity of MmpL3 and
MmpL11 (and also the hemophore Rv0203), and the similar
heme binding properties of their soluble domains, support roles
of both MmpL3 and MmpL11 in heme uptake.
Our findings offer strong support that there is a previously un-
characterized mycobacterial heme uptake system. We identified a
unique secreted heme-binding protein and several membrane
proteins encoded by the region inclusive of Rv0202c–Rv0207c that
are implicated in sequestering heme iron from the host and trans-
portation of heme across the cell wall and membrane (Fig. 6). One
potential energy source for heme transport is ATP (as seen with
other heme uptake systems) (34), as several membrane proteins
in a previous study of MmpL proteins, MtbΔmmpL11 established
infection normally in a murine low-dose aerosol model of tuber-
culosis but was significantly less lethal; median survival of C57BL/6
mice infected with wild-type Mtb was 265 d vs. 398 d for mice in-
fected with MtbΔmmpL11 (32). Importantly, this result suggests
that heme uptake may contribute to the pathogenesis of Mtb.
Mtb may encounter heme or hemoglobin at both intracellular
and extracellular sites of infection. Macrophages degrade senescent
red blood cells at a rate of 2 million/s (36). Thus, intracellular Mtb
in macrophage phagosomes may interact with heme thatdiffusesto
its phagosome from the phagolysosome, where the erythrocyte is
degraded, within the same host cell; alternatively, as some meta-
bolically active Mtb resides in phagolysosomes (37), the bacteria
may interact directly with heme or hemoglobin released from
erythrocytes.ExtracellularMtbmay encounter heme orhemoglobin
in the bloodstream during dissemination from primary to secondary
sites of infection, as well as in tuberculous lung cavities into which
bleeding has occurred, a phenomenon that may result in hemop-
tysis, a symptom of tuberculosis; in these settings, utilization of
heme may be facilitated by hemolysins expressed by Mtb (38).
The discovery of a mycobacterial heme uptake system that has
the ability to use heme from human hemoglobin challenges
a long existing paradigm that Mtb obtains iron solely via exo-
mycobactins and mycobactins scavenging iron from iron-con-
taining proteins, especially transferrin and lactoferrin. To further
support the hypothesis that mycobacteria may use heme as an
iron source, a cytoplasmic mycobacterial heme-degrading protein,
MhuD, has recently been characterized, suggesting that heme iron
traversing the membrane may be degraded to provide iron under
iron-deplete conditions (Fig. 6) (39). Moreover, the likely essenti-
which is very similar to that of MmpL3 (Rv0206c), Fig. S4H. There are 11 transmembrane helices, two extracellular domains (E1 and E2), and one cytoplasmic
domain (C1). (B) SDS/PAGE of purified domains of both MmpL11 and MmpL3. Domains of MmpL11 are to the left of the lane with molecular weight markers
and domains of MmpL3 are to the right of the lane with molecular weight markers (kDa), as indicated. (C and D) UV/vis absorbance spectra of domains E1, E2,
and C1 from MmpL11 (C) and MmpL3 (D) after each domain was incubated with heme and repurified by gel filtration. For both MmpL11 and MmpL3, both
E1 and E2 bind heme whereas C1 does not.
Examination of heme binding capabilities of the predicted soluble domains of MmpL11 and MmpL3. (A) Predicted topology of MmpL11 (Rv0202c),
host hemoglobin (Hb, brown tetramer). Rv0203 delivers heme to membrane
proteins MmpL11 (pink) or MmpL3 (light blue), where heme is shuttled through
breaks down heme to release iron. Black dashed lines indicate the direction of
heme uptake from host to cytoplasm, although the exact sequence of events is
still to be determined along with other protein players within this pathway.
Proposed model of mycobacterial heme uptake. Depicted is the struc-
Tullius et al. PNAS
| March 22, 2011
| vol. 108
| no. 12
ality of mmpL3 (Rv0206c), which encodes a probable heme trans-
port protein, suggests that heme uptake may play a pivotal role in
mycobacterial pathogenicity in vivo. In conclusion, we confirmed
that a heme uptake pathway exists for Mtb and identified the key-
of iron-sequestering pathways in Mtb and opens new avenues in
mycobacterial biology for the development of therapeutics.
All bacterial strains, plasmids, and primers used in this work are listed in
An MtbΔmbtB mutant (Erdman strain) defective in mycobactin biosyn-
thesis was constructed via an allelic exchange method, which uses a tem-
perature-sensitive sacB plasmid, as previously described for a BCGΔmbtB
strain (7). The double mutants, MtbΔmbtBΔRv0203 and MtbΔmbtBΔmmpL11,
were constructed via specialized transduction (40), and details are outlined
in SI Methods. Mtb wild type and mutants were cultured in 7H9–0.01%
tyloxapol containing no supplement, various concentrations of heme or
human hemoglobin, or mycobactin J and grown for 20–28 d, until the
cultures reached stationary phase. Growth was monitored by removing
aliquots for absorbance measurements at 750 nm. The cloning, protein
purification (Table S8), and crystallization protocols have been described for
other Mtb proteins (41), and crystals of Rv0203 were grown by a hanging-
drop, vapor-diffusion method against a reservoir containing 1.7 M LiSO4,
0.1 M Hepes, pH 7.8, at room temperature. Methodology for structure
determination has been previously described (42). Heme binding charac-
terization is outlined in SI Methods (43). An M. smegmatis heme uptake
knockout mutant was created as described previously (44) and the com-
plement plasmid with Mtb genes was constructed as described in SI Meth-
ods (7). M. smegmatis wild type and mutants were passaged in 7H9–0.01%
tyloxapol-NoFe and then cultured in media containing no supplement, 1 μM
heme, or 1 μM FeCl3. Cell density was measured at 750 nm at time points
over a 72-h period.
ACKNOWLEDGMENTS. The authors thank Dr. John T. Belisle, Colorado State
University, National Institutes of Health, National Institute of Allergy and
Infectious Diseases Contract N01 AI-75320, for the generous supply of Mtb
H37Rv genomic DNA. We also thank all of the staff at Stanford Synchrotron
Radiation Lightsource for their invaluable help in data collection. We thank
Laleh Rezaei-Homami and Saša Masleša-Galić for technical assistance and
Jeffrey Gornbein for assistance with statistical analysis. Finally, we thank
Drs. Tom Poulos, Sheryl Tsai, Duilio Cascio, and Morgan Beeby for invaluable
discussions and the TB structural genomics consortium for their support. This
work was supported by a grant from the National and California American
Lung Association RG-78755-N (to C.W.G.) and by National Institutes of
Health Grants AI081161 (to C.W.G.), AI068135 (contract to C.W.G.),
A1078691 (to L.M.M.), AI068413 (to M.A.H.), AI078691 (to L.M.M.), and
HL077000 (to M.A.H.).
1. Gobin J, et al. (1995) Iron acquisition by Mycobacterium tuberculosis: Isolation and
characterization of a family of iron-binding exochelins. Proc Natl Acad Sci USA 92:
2. Wong DK, Gobin J, Horwitz MA, Gibson BW (1996) Characterization of exochelins of
Mycobacterium avium: Evidence for saturated and unsaturated and for acid and ester
forms. J Bacteriol 178:6394–6398.
3. Gobin J, Horwitz MA (1996) Exochelins of Mycobacterium tuberculosis remove iron
from human iron-binding proteins and donate iron to mycobactins in the M.
tuberculosis cell wall. J Exp Med 183:1527–1532.
4. Rodriguez GM, Smith I (2006) Identification of an ABC transporter required for iron
acquisition and virulence in Mycobacterium tuberculosis. J Bacteriol 188:424–430.
5. Rodriguez GM (2006) Control of iron metabolism in Mycobacterium tuberculosis.
Trends Microbiol 14:320–327.
6. Crichton R (2001) Inorganic Biochemistry of Iron Metabolism: From Molecular
Mechansims to Clinical Consequences (Wiley, West Sussex, UK).
7. Tullius MV, Harth G, Maslesa-Galic S, Dillon BJ, Horwitz MA (2008) A replication-
limited recombinant Mycobacterium bovis BCG vaccine against tuberculosis designed
for human immunodeficiency virus-positive persons is safer and more efficacious than
BCG. Infect Immun 76:5200–5214.
8. Wilks A, Burkhard KA (2007) Heme and virulence: How bacterial pathogens regulate,
transport and utilize heme. Nat Prod Rep 24:511–522.
9. Cescau S, et al. (2007) Heme acquisition by hemophores. Biometals 20:603–613.
10. Wandersman C, Stojiljkovic I (2000) Bacterial heme sources: The role of heme,
hemoprotein receptors and hemophores. Curr Opin Microbiol 3:215–220.
11. Parish T, Schaeffer M, Roberts G, Duncan K (2005) HemZ is essential for heme
biosynthesis in Mycobacterium tuberculosis. Tuberculosis (Edinb) 85:197–204.
12. Raghu B, Sarma CR, Venkatesan P (1993) Effect of hemoglobin on the growth of
mycobacteria and the production of siderophores. Indian J Pathol Microbiol 36:376–382.
13. Stojiljkovic I, Kumar V, Srinivasan N (1999) Non-iron metalloporphyrins: Potent
antibacterial compounds that exploit haem/Hb uptake systems of pathogenic
bacteria. Mol Microbiol 31:429–442.
14. Monfeli RR, Beeson C (2007) Targeting iron acquisition by Mycobacterium
tuberculosis. Infect Disord Drug Targets 7:213–220.
15. Ben Amor Y, et al. (2005) Immunological characterization of novel secreted antigens
of Mycobacterium tuberculosis. Scand J Immunol 61:139–146.
16. Maresso AW, Garufi G, Schneewind O (2008) Bacillus anthracis secretes proteins that
mediate heme acquisition from hemoglobin. PLoS Pathog 4:e1000132.
17. Wójtowicz H, et al. (2009) Unique structure and stability of HmuY, a novel heme-
binding protein of Porphyromonas gingivalis. PLoS Pathog 5:e1000419.
18. Deniau C, et al. (2003) Thermodynamics of heme binding to the HasA(SM)
hemophore: Effect of mutations at three key residues for heme uptake. Biochemistry
19. Holm L, Kääriäinen S, Rosenström P, Schenkel A (2008) Searching protein structure
databases with DaliLite v.3. Bioinformatics 24:2780–2781.
20. Arnoux P, et al. (1999) The crystal structure of HasA, a hemophore secreted by
Serratia marcescens. Nat Struct Biol 6:516–520.
21. Tang J, James MN, Hsu IN, Jenkins JA, Blundell TL (1978) Structural evidence for gene
duplication in the evolution of the acid proteases. Nature 271:618–621.
22. Eakanunkul S, et al. (2005) Characterization of the periplasmic heme-binding protein shut
from the heme uptake system of Shigella dysenteriae. Biochemistry 44:13179–13191.
23. Ho WW, et al. (2007) Holo- and apo-bound structures of bacterial periplasmic heme-
binding proteins. J Biol Chem 282:35796–35802.
24. Grigg JC, Vermeiren CL, Heinrichs DE, Murphy ME (2007) Haem recognition by
a Staphylococcus aureus NEAT domain. Mol Microbiol 63:139–149.
25. Sharp KH, Schneider S, Cockayne A, Paoli M (2007) Crystal structure of the heme-IsdC
complex, the central conduit of the Isd iron/heme uptake system in Staphylococcus
aureus. J Biol Chem 282:10625–10631.
26. Létoffé S, Redeker V, Wandersman C (1998) Isolation and characterization of an
extracellular haem-binding protein from Pseudomonas aeruginosa that shares
function and sequence similarities with the Serratia marcescens HasA haemophore.
Mol Microbiol 28:1223–1234.
27. Mazmanian SK, et al.(2003) Passage ofheme-ironacrossthe envelope of Staphylococcus
aureus. Science 299:906–909.
28. Rossi MS, et al. (2001) Identification and characterization of the hemophore-
dependent heme acquisition system of Yersinia pestis. Infect Immun 69:6707–6717.
29. Converse SE, et al. (2003) MmpL8 is required for sulfolipid-1 biosynthesis and
Mycobacterium tuberculosis virulence. Proc Natl Acad Sci USA 100:6121–6126.
30. Domenech P, et al. (2004) The role of MmpL8 in sulfatide biogenesis and virulence of
Mycobacterium tuberculosis. J Biol Chem 279:21257–21265.
31. Jain M, Cox JS (2005) Interaction between polyketide synthase and transporter
suggests coupled synthesis and export of virulence lipid in M. tuberculosis. PLoS
32. Domenech P,Reed MB,BarryCE,3rd (2005)Contributionofthe Mycobacteriumtuberculosis
MmpL protein family to virulence and drug resistance. Infect Immun 73:3492–3501.
33. Ratledge C, Ewing M (1996) The occurrence of carboxymycobactin, the siderophore of
pathogenic mycobacteria, as a second extracellular siderophore in Mycobacterium
smegmatis. Microbiology 142:2207–2212.
34. Köster W (2001) ABC transporter-mediated uptake of iron, siderophores, heme and
vitamin B12. Res Microbiol 152:291–301.
35. Pal D, Eisenberg D (2005) Inference of protein function from protein structure.
36. Soe-Lin S, et al. (2009) Nramp1 promotes efficient macrophage recycling of iron
following erythrophagocytosis in vivo. Proc Natl Acad Sci USA 106:5960–5965.
37. Lee BY, Clemens DL, Horwitz MA (2008) The metabolic activity of Mycobacterium
tuberculosis, assessed by use of a novel inducible GFP expression system, correlates
with its capacity to inhibit phagosomal maturation and acidification in human
macrophages. Mol Microbiol 68:1047–1060.
38. Wren BW, et al. (1998) Characterization of a haemolysin from Mycobacterium
tuberculosis with homology to a virulence factor of Serpulina hyodysenteriae.
39. Chim N, Iniguez A, Nguyen TQ, Goulding CW (2010) Unusual diheme conformation of
the heme-degrading protein from Mycobacterium tuberculosis. J Mol Biol 395:595–608.
40. Bardarov S, et al. (2002) Specialized transduction: An efficient method for generating
marked and unmarked targeted gene disruptions in Mycobacterium tuberculosis, M.
bovis BCG and M. smegmatis. Microbiology 148:3007–3017.
41. Goulding CW, Perry LJ (2003) Protein production in Escherichia coli for structural
studies by X-ray crystallography. J Struct Biol 142:133–143.
42. Goulding CW, et al. (2002) Crystal structure of a major secreted protein of
Mycobacterium tuberculosis-MPT63 at 1.5-A resolution. Protein Sci 11:2887–2893.
43. Izadi N, et al. (1997) Purification and characterization of an extracellular heme-
binding protein, HasA, involved in heme iron acquisition. Biochemistry 36:7050–
44. Converse SE, Cox JS (2005) A protein secretion pathway critical for Mycobacterium
tuberculosis virulence is conserved and functional in Mycobacterium smegmatis. J
| www.pnas.org/cgi/doi/10.1073/pnas.1009516108Tullius et al.