Noncanonical coproporphyrin-dependent bacterial heme biosynthesis pathway that does not use protoporphyrin

Article abstract of DOI:10.1073/pnas.1416285112

It has been generally accepted that biosynthesis of protoheme (heme) uses a common set of core metabolic intermediates that includes protoporphyrin. Herein, we show that the Actinobacteria and Firmicutes (high-GC and low-GC Gram-positive bacteria) are unable to synthesize protoporphyrin. Instead, they oxidize coproporphyrinogen to coproporphyrin, insert ferrous iron to make Fecoproporphyrin (coproheme), and then decarboxylate coproheme to generate protoheme. This pathway is specified by three genes named hemY, hemH, and hemQ. The analysis of 982 representative prokaryotic genomes is consistent with this pathway being the most ancient heme synthesis pathway in the Eubacteria. Our results identifying a previously unknown branch of tetrapyrrole synthesis support a significant shift from current models for the evolution of bacterial heme and chlorophyll synthesis. Because some organisms that possess this coproporphyrin-dependent branch are major causes of human disease, HemQ is a novel pharmacological target of significant therapeutic relevance, particularly given high rates of antimicrobial resistance among these pathogens.

Full text of DOI:10.1073/pnas.1416285112

Noncanonical coproporphyrin-dependent bacterial
heme biosynthesis pathway that does not
use protoporphyrin
Harry A. Daileya,b,c,1, Svetlana Gerdesd, Tamara A. Daileya,b,c, Joseph S. Burcha, and John D. Phillipse
aBiomedical and Health Sciences Institute and Departments of bMicrobiology and cBiochemistry and Molecular Biology, University of Georgia, Athens,
GA 30602; dMathematics and Computer Science Division, Argonne National Laboratory, Argonne, IL 60439; and eDivision of Hematology, Department
of Medicine, University of Utah School of Medicine, Salt Lake City, UT 84132
Edited by J. Clark Lagarias, University of California, Davis, CA, and approved January 12, 2015 (received for review August 25, 2014)
It has been generally accepted that biosynthesis of protoheme
(heme) uses a common set of core metabolic intermediates that
includes protoporphyrin. Herein, we show that the Actinobacteria
and Firmicutes (high-GC and low-GC Gram-positive bacteria) are
unable to synthesize protoporphyrin. Instead, they oxidize copro-
porphyrinogen to coproporphyrin, insert ferrous iron to make Fe-
coproporphyrin (coproheme), and then decarboxylate coproheme
to generate protoheme. This pathway is specified by three genes
named hemY, hemH, and hemQ. The analysis of 982 representa-
tive prokaryotic genomes is consistent with this pathway being
the most ancient heme synthesis pathway in the Eubacteria. Our
results identifying a previously unknown branch of tetrapyrrole
synthesis support a significant shift from current models for the
evolution of bacterial heme and chlorophyll synthesis. Because some
organisms that possess this coproporphyrin-dependent branch are
major causes of human disease, HemQ is a novel pharmacological
target of significant therapeutic relevance, particularly given high
rates of antimicrobial resistance among these pathogens.
of a “primitive” pathway in Desulfovibrio vulgaris (13). This path-
way, named the “alternative heme biosynthesis” path (or ahb), has
now been characterized by Warren and coworkers (15) in sulfate-
reducing bacteria. In the ahb pathway, siroheme, synthesized
from uroporphyrinogen III, can be further metabolized by suc-
cessive demethylation and decarboxylation to yield protoheme (14,
15) (Fig. 1 and Fig. S1). A similar pathway exists for protoheme-
containing archaea (15, 16).
Current gene annotations suggest that all enzymes for pro-
karyotic heme synthetic pathways are now identified. However,
confounding issues exist. First, the validity of most hemN annota-
tions for oxygen-independent coproporphyrinogen oxidases (also
known as coproporphyrinogen dehydrogenase or decarboxylase)
is questionable, and there is no biochemical or genetic evidence
supporting the existence of a coproporphyrinogen oxidase en-
zyme in any of the Firmicutes or Actinobacteria (herein referred
to as Gram-positive bacteria). Second, in Gram-positive bacteria,
the protein HemQ, annotated as a member of the chlorite dis-
mutase (Cld) family, has been shown to be necessary for heme
synthesis (17), although no function has been ascribed to this pro-
tein (17, 18). Finally, protoporphyrin has never been identified as
a pathway intermediate in Gram-positive bacteria (19, 20).
Herein, we show that Firmicutes, Actinobacteria, and several
other bacterial taxa synthesize protoheme via a noncanonical
pathway that is different from the pathways found in eukaryotes,
Archaea, Proteobacteria, or sulfate-reducing bacteria (15, 16).
heme synthesis Gram-positive bacteria HemQ coproporphyrin HemN
|
|
|
|
any biological processes depend upon heme, an iron-con-
taining porphyrin (1, 2). This prosthetic group is essential
M
for the function of diverse proteins, including cytochromes, glo-
bins, peroxidases, catalases, and sensors that bind diatomic gases.
Moreover, heme affects multiple aspects of bacterial pathogen-
esis, such as the ability of Mycobacterium and Campylobacter to
scavenge reactive nitrogen species produced by host immune
systems (3, 4) and the ability of Staphylococcus to modulate
virulence (5, 6). Although the ability to synthesize heme is not
ubiquitous in distribution, there are few organisms that do not
synthesize heme and even fewer that lack heme altogether.
For over 50 y, it has been commonly accepted that the met-
abolic intermediates in the heme biosynthetic pathway are con-
served among all organisms, with the one variation being the
manner by which the first committed intermediate, 5-amino-
levulinate (ALA), is synthesized (1, 7). Whereas the route for
ALA synthesis that employs a single enzyme (HemA) using Gly
and succinyl-CoA as substrates was the first to be described (often
called the four-carbon or Shemin path), it is the glutamyl-tRNA–
based synthesis described in the 1970s (often called the five-car-
bon path), which relies upon two enzymes, glutamyl-tRNA re-
ductase and Glu-1-semialdehyde-2,1-aminomutase, that is most
prevalent across all forms of life (7, 8). Alternative enzymes to
accommodate aerobic vs. anaerobic lifestyles have been described
for the oxidative decarboxylation of coproporphyrinogen III to
protoporphyrinogen IX (aerobic: HemF, anaerobic: HemN/CpdH)
(7, 9, 10) and the oxidation of protoporphyrinogen IX to pro-
toporphyrin IX (aerobic: HemY, mixed: HemG and HemJ) (7, 11,
12) (Fig. S1). However, in each of these instances, the metabolic
intermediates in the pathway are invariable.
Significance
It has been accepted dogma that eukaryotes and heme-syn-
thesizing bacteria use the same metabolic intermediates in their
heme synthesis pathways, where protoporphyrin is the final
intermediate into which iron is inserted to make protoheme.
Herein, we present data demonstrating that Gram-positive
bacteria do not use protoporphyrin as an intermediate but, in-
stead, have an altered set of terminal reactions that oxidize
coproporphyrinogen to coproporphyrin and insert ferrous iron
into coproporphyrin, resulting in the formation of coproheme.
A newly characterized enzyme, HemQ, decarboxylates coproheme
to generate protoheme. Because some organisms that possess
this coproporphyrin-dependent branch are major causes of
human disease, HemQ is a novel pharmacological target of sig-
nificant therapeutic relevance, particularly given high rates of
antimicrobial resistance among these pathogens.
Author contributions: H.A.D., S.G., and J.D.P. designed research; H.A.D., S.G., T.A.D., J.S.B.,
and J.D.P. performed research; T.A.D. contributed new reagents/analytic tools; H.A.D.,
S.G., J.S.B., and J.D.P. analyzed data; and H.A.D., S.G., and T.A.D. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1To whom correspondence should be addressed. Email: hdailey@uga.edu.
The first experimental hint that the “classic” protoheme syn-
thetic pathway was not universal came in 1998 with the postulation
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1416285112/-/DCSupplemental.
www.pnas.org/cgi/doi/10.1073/pnas.1416285112
PNAS Early Edition
|
1 of 6

Fig. 1. Currently characterized tetrapyrrole biosynthesis pathways. The core enzymes present in all tetrapyrrole-synthesizing organisms are shown in the
blue box. The protoporphyrin-independent, ancestral heme biosynthesis pathway of the Firmicutes and Actinobacteria described in the current study is shown
in orange. The primitive heme pathway of the archaea and sulfate-reducing bacteria is shown in the violet box, and the modern heme pathway of Pro-
teobacteria and eukaryotes is shown in the red box. Enzyme abbreviations are those abbreviations currently in acceptance and are shown with single solid
lines. Pathways involving more than a single enzyme are shown as double-line arrows. The dashed line represents a proposed pathway for chlorophyll
synthesis in bacteria that do not use protoporphyrin. ALA, 5-aminolevulinic acid; COPRO, coproporphyrin; COPRO’GEN, coproporphyrinogen; GltR, glutamyl-
tRNA reductase; Glut-tRNA, glutamyl-tRNA; GSAMS, Glu-1-semialdehyde-2,1-aminomutases; HMB, hydroxymethylbilane; PBG, porphobilinogen; PROTO,
protoporphyrin; PROTO’GEN, protoporphyrinogen; succCoA, succinyl-coenzymeA; URO’GEN, uroporphyrinogen.
Given that some of the Gram-positive bacteria that possess this
Streptococcus species that do not synthesize heme. Thus, there
pathway are pathogens and contributors to human disease, HemQ exist no identifiable enzymes to produce protoporphyrinogen
represents a novel pharmacological target of significant thera-
peutic relevance.
in many bacterial genomes (otherwise encoding a full complement
of heme biosynthetic genes), especially in Firmicutes, Actino-
bacteria, Planctinomyces, and several other taxa (23) (Dataset S1).
Results
Identification of the Terminal Enzymes of Gram-Positive Heme
Synthesis. Our current bioinformatics analyses indicate that bona
fide CpdH (HemN) is found in Proteobacteria, cyanobacteria, the
Bacteroidetes/Chlorobi group, the Chlamydiae Chlamydiae/
Verrucomicrobia group, and several other phyla, but it does not
occur in any sequenced genomes of Firmicutes or Actinobacteria.
Likewise, the aerobic form of coproporphyrinogen oxidase,
HemF, is found in the same major taxa as HemN but is not
present in sequenced genomes of Firmicutes or Actinobacteria
(with very few exceptions). Extensive biochemical and bio-
informatics approaches in our laboratory failed to identify a
gene/protein in Gram-positive organisms that catalyzed the con-
version of coproporphyrinogen to protoporphyrinogen. However,
our screens did identify a protein of unknown function (HemQ)
that is essential for heme synthesis in Firmicutes and Actinobacteria
(17). These data and experiences led us to reevaluate the state of
knowledge about heme synthesis in bacteria.
Analysis of Current hemN Annotations. In the face of rapidly
expanding genomic data, annotation has had to rely on an in-
sufficiently small number of experimentally validated functional
assignments (21). Unfortunately, this intrinsically error-prone
process created substantial problems in the annotation of the
hemN/Z branch of the radical S-adenosyl methionine (SAM)
superfamily as all being oxygen-independent coproporphyrinogen
oxidase. Using a comparative genomics approach to identify genes
that encode coproporphyrinogen dehydrogenases (CpdHs), all
genes annotated as oxygen-independent coproporphyrinogen oxi-
dases (hemN and hemZ) were reexamined. This analysis was per-
formed in the SEED database (22), and the results are presented
in the SEED subsystem “Heme biosynthesis: Protoporphyrin- and
coproporphyrin-dependent pathways” (23).
Based on genomic context and amino acid sequence com-
parisons, gene products fit into four distinct groups within the
radical SAM enzyme superfamily (Table S1) and subsystem in
SEED (23). The first of these groups, group A, represents the bona
fide CpdH and contains the experimentally validated Escherichia
coli HemN. Multiple sequence alignment using ClustalW (24) (Fig.
S2) revealed that conserved regions correspond to those regions
present in all radical SAM enzymes [e.g., the S-adenosyl methi-
onine–binding sites and the Fe-S cluster motif (CXXXCXXC)].
However, only the group A family contains the motifs of E. coli
HemN that are essential for CpdH activity: 20GPRYTSYPTA29
and 308KNFQGYTT315 (7, 10). Thus, these data suggest that
only group A members possess the necessary structural attributes
to be classified as CpdH. Group B members are exemplified by
Vibrio cholerae HutW, a protein that does not possess CpdH/
CpoX activity but appears to be involved in heme uptake and/or
The hemY gene was originally identified as part of the Ba-
cillus subtilis gene cluster hemEHY (26), and thereafter was
shown to encode a protein that was capable of oxidizing pro-
toporphyrinogen IX to protoporphyrin IX (27, 28). A crystal
structure exists for this protein (29), which is of a similar size
and structure as the eukaryotic PpoX (30, 31) and Gram-neg-
ative HemY (32). However, it possesses four distinctions from
these enzymes: (i) it is a soluble monomer, (ii) it is relatively
insensitive to the herbicide acifluorfen, (iii) it is unable to com-
plement a ΔppoX (hemG) mutant of E. coli, and (iv) it is able to
oxidize coproporphyrinogen to coproporphyrin. Indeed, it oxidizes
coproporphyrinogen at a rate almost ninefold faster than it oxi-
dizes protoporphyrinogen (7.0 nmol·min−1 vs. 0.85 nmol·min−1
,
utilization (25). Group C members are found in a wide variety respectively) and has a lower apparent Km (0.56 μM vs. 0.95 μM,
of organisms, including many incapable of heme biosynthesis.
respectively) (28). Additionally, when B. subtilis HemY is over-
These group C genes are often clustered with those genes related expressed in E. coli, it causes the accumulation of coproporphyrin
to nucleotide metabolism. Finally, group D contains hemNs in the (27), not protoporphyrin. To justify the assignment of this HemY
Firmicutes and Actinobacteria, including those annotated in
as a protoporphyrinogen (rather than a coproporphyrinogen)
2 of 6
|
www.pnas.org/cgi/doi/10.1073/pnas.1416285112
Dailey et al.

oxidase, it has been proposed that, in vivo, it exists in a complex
with other proteins where HemY would be prevented from
encountering coproporphyrinogen. The possibility that HemY
functions in vivo in Gram-positive bacteria to oxidize cop-
roporphyrinogen to coproporphyrin had not been examined
experimentally.
HemY + HemH + HemQ of P. acnes (or M. tuberculosis) were
assayed in vitro with coproporphyrinogen and ferrous iron, the
products observed were protoheme and an intermediate, which
was identified by mass spectrophotometry as monovinyl, monop-
ropionyl heme, resulting from the decarboxylation of a single
propionate side chain (Fig. 2). The stereochemistry of the single
decarboxylation (i.e., vinyl at the A or B ring) was not determined.
When HemH and HemQ, together with coproporphyrin and
ferrous iron, or HemQ alone with coproheme was assayed,
Ferrochelatase (HemH) from B. subtilis has been well char-
ꢀ
acterized, including elucidation of its crystal structure at 1 Å (33,
34). HemHs from the Actinobacteria have also been identified
and characterized, and they have been found to have a [2Fe-2S]
cluster similar to what is found in metazoan ferrochelatases (34–
36). Ferrochelatases are structurally well conserved, with the ex-
ception that the enzymes from Gram-positive organisms lack a
small loop that forms one lip of the active site mouth (34) (Fig.
S3). Interestingly this lip, in ferrochelatases that possess it, en-
closes the active site during catalysis, thereby forming a snug pocket
that encloses the A and B rings (which possess vinyl groups at the
2, 4 position) of the macrocycle (37) (Fig. S1).
We examined the substrate specificity of ferrochelatase from
B. subtilis (a Firmicute) and Mycobacterium tuberculosis (an Acti-
nobacteria), along with human ferrochelatase, to determine if they
could use the coproporphyrin that would be produced in vivo by
HemY. All proteobacterial and eukaryotic ferrochelatases exam-
ined to date do not catalyze the insertion of iron into cop-
roporphyrin at measurable rates. We find that the tested purified
ferrochelatases from these representatives of both phyla use
coproporphyrin at a rate that is greater than is found with pro-
toporphyrin (Fig. S4 and Tables S2 and S3). Indeed, we determined
that the Km of M. tuberculosis ferrochelatase for coproporphyrin
III (10.5 μM) is over 60-fold lower than its Km for protoporphyrin
IX (720 μM) and that the kcat is 1.8 min−1 vs. 0.8 min−1 (Fig. S5).
For B. subtilis ferrochelatase, we determined that the Km for
coproporphyrin III is 7.8 μM and that the kcat is 0.11 min−1. In
our hands, the activity of B. subtilis ferrochelatase for pro-
toporphyrin is too low to measure, although it has previously been
reported that the Km for protoporphyrin is 8.0 μM (38). Of note is
that both I and III isomers of coproporphyrin serve as substrates
for these enzymes (Fig. S4). The substrate specificity with regard to
metals is not distinctly different from eukaryotic ferrochelatases
(Tables S2 and S3).
Previously, we identified HemQ (COG3253) as an essential
component of heme synthesis in Firmicutes and Actinobacteria
(17). Our initial genomic analysis identified HemQ as being the
best candidate for the missing coproporphyrinogen oxidase (17).
However, we were unable to detect any coproporphyrinogen
oxidase activity. Because Propionibacterium acnes HemH and
HemQ exist as a fusion protein, one possibility was that HemQ
served as a scaffold upon which HemY and HemH may assem-
ble. However, the lack of an identifiable stable complex in vitro
did not support this model. Since then, others have reexamined
HemQ, corroborating our finding that it lacks coproporphyri-
nogen oxidase activity, and have also demonstrated that KO of
hemQ in Staphylococcus aureus results in the accumulation of
coproporphyrin (18).
The structure of three HemQs (annotated as Clds) has been
determined and reveals that HemQ is a homopentamer (Protein
Data Bank ID codes 1T0T, 1VDH, and 3DTZ). Unfortunately,
the available protein structures of HemQs do not have bound
product or substrate. For the HemQs we studied, the purified
protein contains no metals or cofactors, but one can titrate pro-
toheme into the apoprotein to produce a stable hemoprotein (17).
With the realization that HemY + HemH of Gram-positive bac-
teria catalyze the formation of coproheme in vitro, we examined
HemQ to determine if it converts coproheme to protoheme.
Previously, we have shown that concurrent expression of HemY +
HemH + HemQ in E. coli lacking either HemH or HemG
resulted in rescue of these cells and that this occurs without
the accumulation of coproporphyrin (17). When recombinant
Fig. 2. HPLC chromatograms of reaction products of HemY/HC/Q reactions.
Assays were run as described in the main text. Elution time is shown on the x
axis, and absorbance is shown on the y axis. (A) Reaction products from
P. acnes HemY + HemH + HemQ with coproporphyrinogen and iron as sub-
strates are shown in the first line (cyan). Reaction products from assay with
coproporphyrin and iron as substrates with 50 μM FMN present are shown in
the second line (green). Protohemin (blue) and coprohemin (black) standards
are shown in the third and fourth lines, respectively. (B) Assay products from
M. tuberculosis HemQ assayed with coprohemin as a substrate in the presence
of 50 μM H2O2 (blue) or 50 μM FMN (black). (C) Assay products from
M. tuberculosis HemQ that was not heme-loaded (blue) or was loaded with
heme and then assayed with coprohemin as a substrate in the presence of
50 μM FMN. The mass spectrum results yielded 752.2 (coproheme + formate)
for peak 1, 706.2 (monovinyl, monopropionyl heme intermediate + formate)
for peak 2, and 660.2 (protoheme + formate) for peak 3.
Dailey et al.
PNAS Early Edition
|
3 of 6

protoheme was identified as the product, although at low con- existence of all known bacterial enzymes that catalyze the oxi-
dative decarboxylation of the 2, 4-position propionates of the
coproporphyrin(ogen) tetrapyrrole-to-vinyl groups. This list in-
cludes HemF, HemN, HemQ, and AhbD. The data are pre-
sented in tabular (Dataset S1) and diagrammatic (Fig. 4) forms.
HemF and HemN are present in Proteobacteria, cyanobacteria,
the Bacteroidetes/Chlorobi group, the Chlamydiae/Verrucomicro-
bia group, Aquificae, Gemmatimonadetes, and several other taxa,
and they overlap significantly with each other. HemQ is distinct
from HemF/N, and it is found in nearly every Firmicute and
Actinobacteria with currently sequenced genomes, and also in
evolutionarily early-branching (Acidobacteria and Planctomyces)
and transitional (Deinococcus-Thermus group) diderm phyla
(39, 40). HemQ is the sole coproheme decarboxylase in almost
70% of genomes of Gram-positive bacteria (Fig. 4, Fig. S6, and
Dataset S1). However, in ∼30% of the genomes possessing
HemQ, it coexists with a clear homolog of AhbD in the absence
of other enzymes of the siroheme-to-protoheme pathway (AhbA,
AhbB, and AhbC). Indeed, AhbD is found to co-occur more
frequently with HemQ (along with HemY and HemH) than it
does with AhbA, AhbB, and AhbC. Among genomes that possess
AhbD, ∼60% contain HemQ and do not contain AhbA, AhbB,
and AhbC. This phylogenetic occurrence profile strongly indi-
cates that coproheme decarboxylase AhbD, discovered origi-
nally as part of the siroheme-to-protoheme pathway (14, 16),
can also function in the HemQ-based coproporphyrin-depen-
dent heme biosynthesis route described in this work. Given that
AhbD is an anaerobic, radical SAM enzyme and that HemQ
functions in the presence of oxygen, the presence of both cop-
roheme decarboxylases in a single organism mimics the presence
of an anaerobic coproporphyrinogen decarboxylase (HemN) and
aerobic (HemF) coproporphyrinogen decarboxylase in many Gram-
negative bacteria.
HemY most frequently coexists with HemQ and AhbD (where
it would function as a coproporphyrinogen oxidase), but mini-
mally with HemF and HemN (where it would function as a pro-
toporphyrinogen oxidase) (Fig. 4). This strongly suggests that the
annotation for HemY should be coproporphyrinogen oxidase,
secondarily protoporphyrinogen oxidase. This finding is consis-
tent with previous data showing that HemJ is the most common
form of protoporphyrinogen oxidase in Proteobacteria (12, 41),
whereas HemY is found infrequently among Proteobacteria that
possess the classic pathway (Dataset S1). Attempts to delineate
HemYs into coproporphyrinogen vs. protoporphyrinogen oxidases
based upon sequence alignments proved futile (Fig. S6). Unlike
ferrochelatases, where a clear sequence/structural distinction can
be drawn due to the presence/absence of one active site lip, suffi-
cient differences/commonality could not be found among available
sequence data for HemYs. Additionally, no studies exist that char-
acterize the range of substrate specificity (i.e., coproporphyrinogen
vs. protoporphyrinogen) for HemYs to address the possibility that
some HemYs may function in vivo with both substrates.
centrations. Under identical conditions, however, HemQ did not
convert Ni-coproporphyrin into Ni-protoporphyrin.
The oxidative decarboxylation of coproheme to protoheme
generates four protons, along with two CO2s , and requires an
electron acceptor. Because HemQ has been noted to have low
peroxidase activity (17, 18), the possibility that HemQ was acting
as a decarboxylase/peroxidase was examined. Inclusion of H2O2
in the reaction of HemQ with coproheme resulted in produc-
tion of protoheme (Fig. 2B). However, the overall amount of
coproheme + protoheme present after the assay period was less than
expected, suggesting that the product heme was being degraded
in the presence of peroxide. This finding is consistent with pre-
vious observations that “holo-HemQ” degrades its heme in vitro
(18) and suggests that in the absence of a heme acceptor, HemQ-
bound heme may be degraded. Some common electron accept-
ors were examined for their ability to stimulate the reaction.
NAD, NADP, or FAD at a final concentration of 50 μM had no
impact on protoheme production by HemQ, but inclusion of
FMN resulted in conversion of coproheme into protoheme with
HemQ alone (Fig. 2B). Of note is that inclusion of FMN had no
measurable impact on protoheme formation by HemY + HemH +
HemQ from coproporphyrinogen + iron. Because HemQ is known
to bind both ferrous and ferric iron-protoporphyrin, the possibility
that a protoheme HemQ “holoenzyme” is the actual decarboxylase
rather than a product-bound form of the enzyme was examined.
Coproheme decarboxylase assays with either heme-free or heme-
bound HemQ were carried out, and there was no distinguishable
difference between them (Fig. 2C).
Genomic Analysis of Heme Synthesis in Firmicutes and Actinobacteria.
Our findings demonstrate that HemY, HemH, and HemQ in
these organisms effectively work to oxidize coproporphyrinogen
to coproporphyrin, insert ferrous iron to make coproheme, and
then decarboxylate the 2, 4-position propionates of coproheme
in a stepwise fashion to form the vinyl groups of protoheme
(Fig. 3). Thus, for the Gram-positive organisms selected for bio-
chemical characterization in this study, we have defined a new
pathway for protoheme synthesis that does not involve pro-
toporphyrin as an intermediate.
To avoid unwarranted generalization of our data, we have
used a genomic analysis approach to identify and compare the
Discussion
The first significant study of porphyrin metabolism in Gram-
positive bacteria was published in 1957 by Townsley and Neilands
(20), who examined iron and porphyrin metabolism in the Gram-
positive bacterium Micrococcus lysodeikticus. They reported that
lysed cell preparations were capable of synthesizing coproheme,
but not protoheme, and that no protoporphyrin was detectable.
In the 1970s, the Jacobs’ group (19) examined the terminal steps of
heme synthesis in a variety of bacteria. They found that although
the Proteobacteria E. coli and Pseudomonas denitrificans readily ac-
cumulated protoporphyrin when coproporphyrinogen was added
to cell-free extracts, the Gram-positive bacteria M. lysodeikticus
and S. aureus did not (19). Additionally, although E. coli and
P. denitrificans accumulated protoporphyrin when the heme pre-
cursor ALA was provided, the Gram-positive bacteria accumulated
Fig. 3. Reaction catalyzed by the enzymes HemY, HemH, and HemQ (details
are provided in main text). Coproporphyrinogen III (A), coproporphyrin III (B),
coproheme III (C), and protoheme IX (D) are shown. Pyrrole ring lettering and
side-chain numbering are shown on protoheme IX.
4 of 6
|
www.pnas.org/cgi/doi/10.1073/pnas.1416285112
Dailey et al.

Frequency of co-occurrence with:
HemF HemN HemQ AhbD
Occurs Total
% as only
with
genomes
decarboxylase HemY
(%)
in
selection
(%)
(%)
(%)
(%)
HemF
X
76.1
0.8
0.6
22.5
22.0
67.1
31.3
11
15
93
94
386
393
152
67
HemN 76.2
HemQ 2.0
X
1.3
X
0.5
27.6
X
3.3
3.0
AhbD
3.0
62.7
Fig. 4. Co-occurrence of HemF, HemN, HemQ, AhbD, and HemY. A Venn diagram illustrates the overlap of the four decarboxylases and HemY. The values
shown in the table represent the percentage of each of the four decarboxylases as it is found to occur within the same genome with any of the other car-
boxylases. Also listed are the occurrence of each decarboxylase with HemY and the total number of genomes in the selection containing each decarboxylase. It
should be noted that all heme-synthesizing Gram-positive bacteria possess HemY but few Gram-negative (Proteobacteria) bacteria have a hemY gene.
coproporphyrin and not protoporphyrin. Interestingly, neither
group interpreted their observations as the basis for coproheme
as a protoheme intermediate in Gram-positive bacteria but, in-
stead, explained the lack of protoporphyrin and the ability to syn-
specific acceptors. In vitro assays demonstrate that these acceptors
may include free FMN or hydrogen peroxide, but not molecular
oxygen, FAD, NAD, or NADP. Utilization of hydrogen peroxide
is an attractive possibility because the antepenultimate enzyme
thesize coproheme as interesting anomalies or artifacts of their in HemY generates six molecules of peroxide that would provide for
vitro systems.
more than the four required for HemQ. Having multiple pos-
sible acceptors would provide for the metabolic diversity that
many of the HemQ-containing bacteria possess.
We have demonstrated that within the Firmicutes, Actino-
bacteria, and evolutionary early-branching diderm taxa (39, 40),
the terminal steps for the synthesis of protoheme do not go
A few oddities exist that do not clearly fit into either the
through protoporphyrinogen and protoporphyrin; instead, classic protoporphyrin-dependent alternative (AhbA-, AhbB-,
HemY oxidizes coproporphyrinogen to coproporphyrin, HemH
inserts iron to form coproheme, and HemQ then decarboxylates
AhbC-, or AhbD-based) or coproporphyrin-dependent (HemQ-
or AhbD-based) pathway. Such organisms are flagged in the
the coproheme to form protoheme (Figs. 1 and 3). The manner in associated SEED subsystem (23) with the variant codes “hybrid-
which Gram-positive HemY and HemH function should be highly 1” or “hybrid-2” and are beyond the scope of this study. One
similar to the action of the canonical HemY and HemH, respec-
group of note, however, is the photosynthetic organisms that
tively, because they differ only in that they use coproporphyri- produce heme via the coproporphyrin-dependent pathway. How
nogen and coproporphyrin, rather than protoporphyrinogen and they synthesize chlorophyll remains a mystery, given that all known
protoporphyrin, as substrates. However, the mechanism by which chlorophyll-synthesizing pathways use protoporphyrin as an in-
HemQ functions is unclear at the present time. HemF and HemN, termediate. Such cases are exceedingly rare: Of 72 photosyn-
which catalyze the oxidative decarboxylation of coproporphyri- thetic organisms included in this study (23) (Dataset S1), only
nogen to protoporphyrinogen, function in a stepwise fashion with three species from two major taxonomic groups fall into this
the production of monovinyl, monopropionyl porphyrinogen, fol- category, namely, Heliobacterium modesticaldum of the Gram-
lowed by decarboxylation of a second ring propionate to make positive photosynthetic heliobacteria (44, 45) and Roseiflexus sp.
protoporphyrinogen and the overall production of two molecules
of CO2. HemF uses molecular oxygen as an electron acceptor,
and the enzyme from E. coli has been shown to be stimulated
by Mn (42). A model for catalysis by the HemF-type copro-
porphyrinogen oxidase proposed by Lash (43) has current ac-
RS-1 and Roseiflexus castenholzi of green nonsulfur bacteria (46,
47). These species are analyzed in more detail in SI Methods.
The inability of Firmicutes, Actinobacteria, and some evolu-
tionarily early-branching diderm bacteria (39, 40) to synthesize
protoporphyrin raises something of an evolutionary conundrum.
ceptance. For the oxygen-independent HemN, Jahn and coworkers If the general assumption is that aerobic metabolism arose after
(2, 7) have presented evidence that this radical SAM enzyme forms
photosynthetic organisms provided atmospheric oxygen, how does
one account for the presence of robust oxygen-using pathways
a substrate radical by hydrogen abstraction at the β-carbon of the
propionate side chain as the first step in the reaction sequence. For among the Gram-positive bacteria whose tetrapyrrole biosynthetic
the iron-sulfur cluster containing the radical SAM enzyme AhbD, it
has been proposed that it uses the same reaction mechanism as
HemN, which may be reasonable, given that both are radical SAM
enzymes that function in the absence of oxygen (14, 16).
HemQ differs from both HemF and HemN in that its sub-
strate has a fully unsaturated macrocyclic tetrapyrrole with a
coordinated iron. HemQ uses iron-coproporphyrin as a sub-
strate, and the conversion by HemQ of coproheme to protoheme
occurs for Fe-coproporphyrin but not for Ni-coproporphyrin,
suggesting a possible role for iron in the reaction. The fate of the
abstracted electrons appears not to be random but to involve
pathway seems to predate the advent of photosynthesis?
The demonstration that Gram-positive bacteria use a different
heme biosynthetic pathway that is specific for them, along with
the observation that KO of hemQ in S. aureus results in small
colony variants (18), suggests that this enzyme may represent a
valid selective antimicrobial target for pathogenic Gram-positive
bacteria. This finding could be of considerable value, given the
advent of drug-resistant strains of Staphylococcus, Listeria, and
Mycobacteria species. Antimicrobial resistance in Gram-positive
pathogens has been recognized as a public health threat by both
the WHO (48) and the US Centers for Disease Control and
Dailey et al.
PNAS Early Edition
|
5 of 6

Ultra Performance Liquid Chromatography Chromatograms of Reaction Pro-
ducts of HemY/H/Q Reactions. Analysis of products was performed by stan-
dard ultra performance liquid chromatography porphyrin/hemin analysis.
Specific details are described in SI Methods.
Prevention (49). There are also identified links between the
abundance of Firmicutes in the human gut microbiome and
obesity and potentially cancer (50). HemQ, identified herein, is
an essential enzyme for heme synthesis specific for Gram-posi-
tive bacteria capable of heme biosynthesis and is a novel phar-
macological target with significant therapeutic relevance.
Genomic Analysis. Many of these analyses are described in the text and/or in
the Dataset S1. Additional details are presented in SI Methods.
Methods
ACKNOWLEDGMENTS. We thank W. B. Whitman, I. Hamza, A. Medlock, and
J. DuBois for critical review of the manuscript. This work was supported by
NIH Grants DK096051 (to H.A.D.) and DK020503 and DK083909 (to J.D.P.).
Expression, Purification, and Assay of HemY, HemH, and HemQ. Many of these pro-
cedures are previously published, and specific details are included in SI Methods.
1. Hamza I, Dailey HA (2012) One ring to rule them all: Trafficking of heme and heme
synthesis intermediates in the metazoans. Biochim Biophys Acta 1823(9):1617–1632.
2. Munro AW, et al. (2009) Heme and hemoproteins. Tetrapyrroles: Birth, Life and
Death, eds Warren MJ, Smith AG (Landes Bioscience, Austin, TX), pp 160–183.
3. Ascenzi P, Bolognesi M, Milani M, Guertin M, Visca P (2007) Mycobacterial truncated
hemoglobins: From genes to functions. Gene 398(1-2):42–51.
26. Hansson M, Hederstedt L (1992) Cloning and characterization of the Bacillus subtilis
hemEHY gene cluster, which encodes protoheme IX biosynthetic enzymes. J Bacteriol
174(24):8081–8093.
27. Dailey TA, Meissner P, Dailey HA (1994) Expression of a cloned protoporphyrinogen
oxidase. J Biol Chem 269(2):813–815.
28. Hansson M, Hederstedt L (1994) Bacillus subtilis HemY is a peripheral membrane
protein essential for protoheme IX synthesis which can oxidize coproporphyrinogen
III and protoporphyrinogen IX. J Bacteriol 176(19):5962–5970.
4. Ascenzi P, Visca P (2008) Scavenging of reactive nitrogen species by mycobacterial
truncated hemoglobins. Methods Enzymol 436:317–337.
29. Qin X, et al. (2010) Structural insight into unique properties of protoporphyrinogen
oxidase from Bacillus subtilis. J Struct Biol 170(1):76–82.
5. Stauff DL, Skaar EP (2009) The heme sensor system of Staphylococcus aureus. Contrib
Microbiol 16:120–135.
30. Koch M, et al. (2004) Crystal structure of protoporphyrinogen IX oxidase: A key en-
zyme in haem and chlorophyll biosynthesis. EMBO J 23(8):1720–1728.
31. Qin X, et al. (2011) Structural insight into human variegate porphyria disease. FASEB J
25(2):653–664.
32. Corradi HR, et al. (2006) Crystal structure of protoporphyrinogen oxidase from Myx-
ococcus xanthus and its complex with the inhibitor acifluorfen. J Biol Chem 281(50):
38625–38633.
33. Al-Karadaghi S, Hansson M, Nikonov S, Jönsson B, Hederstedt L (1997) Crystal struc-
ture of ferrochelatase: The terminal enzyme in heme biosynthesis. Structure 5(11):
1501–1510.
34. Dailey HA, Dailey TA (2003) Ferrochelatase. The Porphyrin Handbook, eds Kadish KM,
Smith KM, Guilard R (Academic, San Diego), Vol 12, pp 93–122.
35. Dailey TA, Dailey HA (2002) Identification of [2Fe-2S] clusters in microbial ferroche-
latases. J Bacteriol 184(9):2460–2464.
6. Torres VJ, et al. (2007) A Staphylococcus aureus regulatory system that responds to
host heme and modulates virulence. Cell Host Microbe 1(2):109–119.
7. Layer G, Reichelt J, Jahn D, Heinz DW (2010) Structure and function of enzymes in
heme biosynthesis. Protein Sci 19(6):1137–1161.
8. Kannangara CG, et al. (1988) tRNA(Glu) as a cofactor in delta-aminolevulinate bio-
synthesis: Steps that regulate chlorophyll synthesis. Trends Biochem Sci 13(4):139–143.
9. Coomber SA, Jones RM, Jordan PM, Hunter CN (1992)
A putative anaerobic cop-
roporphyrinogen III oxidase in Rhodobacter sphaeroides. I. Molecular cloning, trans-
poson mutagenesis and sequence analysis of the gene. Mol Microbiol 6(21):3159–3169.
10. Layer G, Moser J, Heinz DW, Jahn D, Schubert WD (2003) Crystal structure of cop-
roporphyrinogen III oxidase reveals cofactor geometry of Radical SAM enzymes.
EMBO J 22(23):6214–6224.
11. Boynton TO, Daugherty LE, Dailey TA, Dailey HA (2009) Identification of Escherichia
coli HemG as a novel, menadione-dependent flavodoxin with protoporphyrinogen
oxidase activity. Biochemistry 48(29):6705–6711.
36. Shepherd M, Dailey TA, Dailey HA (2006) A new class of [2Fe-2S]-cluster-containing
protoporphyrin (IX) ferrochelatases. Biochem J 397(1):47–52.
12. Boynton TO, et al. (2011) Discovery of a gene involved in a third bacterial proto-
porphyrinogen oxidase activity through comparative genomic analysis and functional
complementation. Appl Environ Microbiol 77(14):4795–4801.
13. Ishida T, et al. (1998) A primitive pathway of porphyrin biosynthesis and enzymology
in Desulfovibrio vulgaris. Proc Natl Acad Sci USA 95(9):4853–4858.
14. Bali S, et al. (2011) Molecular hijacking of siroheme for the synthesis of heme and d1
heme. Proc Natl Acad Sci USA 108(45):18260–18265.
37. Medlock A, Swartz L, Dailey TA, Dailey HA, Lanzilotta WN (2007) Substrate inter-
actions with human ferrochelatase. Proc Natl Acad Sci USA 104(6):1789–1793.
38. Hansson M, Hederstedt L (1994) Purification and characterisation of a water-soluble
ferrochelatase from Bacillus subtilis. Eur J Biochem 220(1):201–208.
39. Gupta RS (1998) Protein phylogenies and signature sequences:
A reappraisal of
evolutionary relationships among archaebacteria, eubacteria, and eukaryotes. Mi-
crobiol Mol Biol Rev 62(4):1435–1491.
15. Bali S, Palmer DJ, Schroeder S, Ferguson SJ, Warren MJ (2014) Recent advances in the
biosynthesis of modified tetrapyrroles: The discovery of an alternative pathway for
the formation of heme and heme d 1. Cell Mol Life Sci 71(15):2837–2863.
16. Kühner M, et al. (2014) The alternative route to heme in the methanogenic archaeon
Methanosarcina barkeri. Archaea 2014:327637.
17. Dailey TA, et al. (2010) Discovery and Characterization of HemQ: An essential heme
biosynthetic pathway component. J Biol Chem 285(34):25978–25986.
18. Mayfield JA, et al. (2013) The chlorite dismutase (HemQ) from Staphylococcus aureus
has a redox-sensitive heme and is associated with the small colony variant phenotype.
J Biol Chem 288(32):23488–23504.
19. Jacobs NJ, Jacobs JM, Brent P (1971) Characterization of the late steps of microbial
heme synthesis: Conversion of coproporphyrinogen to protoporphyrin. J Bacteriol
107(1):203–209.
20. Townsley PM, Neilands JB (1957) The iron and porphyrin metabolism of Micrococcus
lysodeikticus. J Biol Chem 224(2):695–705.
21. Schnoes AM, Brown SD, Dodevski I, Babbitt PC (2009) Annotation error in public
databases: Misannotation of molecular function in enzyme superfamilies. PLOS
Comput Biol 5(12):e1000605.
22. Overbeek R, et al. (2005) The subsystems approach to genome annotation and its use
in the project to annotate 1000 genomes. Nucleic Acids Res 33(17):5691–5702.
23. SEED, Heme biosynthesis: Protoporphyrin- and coproporphyrin-dependent pathways.
Available at pubseed.theseed.org//SubsysEditor.cgi?page=ShowSubsystem&subsystem=
Heme_Biosynthesis%3A_protoporphyrin-_and_coproporphyrin-dependent_pathways.
Accessed January 5, 2015.
40. Gupta RS (2011) Origin of diderm (Gram-negative) bacteria: Antibiotic selection
pressure rather than endosymbiosis likely led to the evolution of bacterial cells with
two membranes. Antonie van Leeuwenhoek 100(2):171–182.
41. Kobayashi K, Masuda T, Tajima N, Wada H, Sato N (2014) Molecular phylogeny and
intricate evolutionary history of the three isofunctional enzymes involved in the ox-
idation of protoporphyrinogen IX. Genome Biol Evol 6(8):2141–2155.
42. Breckau D, Mahlitz E, Sauerwald A, Layer G, Jahn D (2003) Oxygen-dependent cop-
roporphyrinogen III oxidase (HemF) from Escherichia coli is stimulated by manganese.
J Biol Chem 278(47):46625–46631.
43. Lash TD (2005) The enigma of coproporphyrinogen oxidase: How does this unusual
enzyme carry out oxidative decarboxylations to afford vinyl groups? Bioorg Med
Chem Lett 15(20):4506–4509.
44. Sousa FL, Shavit-Grievink L, Allen JF, Martin WF (2013) Chlorophyll biosynthesis gene
evolution indicates photosystem gene duplication, not photosystem merger, at the
origin of oxygenic photosynthesis. Genome Biol Evol 5(1):200–216.
45. Xiong J, Bauer CE, Pancholy A (2007) Insight into the haem d1 biosynthesis pathway in
heliobacteria through bioinformatics analysis. Microbiology 153(Pt 10):3548–3562.
46. Hugenholtz P, Stackebrandt E (2004) Reclassification of Sphaerobacter thermophilus
from the subclass Sphaerobacteridae in the phylum Actinobacteria to the class
Thermomicrobia (emended description) in the phylum Chloroflexi (emended de-
scription). Int J Syst Evol Microbiol 54(Pt 6):2049–2051.
47. Kunisawa T (2011) The phylogenetic placement of the non-phototrophic, Gram-pos-
itive thermophile ‘Thermobaculum terrenum’ and branching orders within the phy-
lum ‘Chloroflexi’ inferred from gene order comparisons. Int J Syst Evol Microbiol 61(Pt
8):1944–1953.
24. Larkin MA, et al. (2007) Clustal W and Clustal X version 2.0. Bioinformatics 23(21):
2947–2948.
25. Abicht HK, Martinez J, Layer G, Jahn D, Solioz M (2012) Lactococcus lactis HemW
(HemN) is a haem-binding protein with a putative role in haem trafficking. Biochem J
442(2):335–343.
48. WHO (2014) Antimicrobial Resistance. Global Report on Surveillance 2014 (WHO, Geneva).
49. CDC (2013) Antibiotic Resistance Threats in the United States, 2013 (US Centers for
Disease Control and Prevention, Atlanta).
50. Yoshimoto S, et al. (2013) Obesity-induced gut microbial metabolite promotes liver
cancer through senescence secretome. Nature 499(7456):97–101.
6 of 6
|
www.pnas.org/cgi/doi/10.1073/pnas.1416285112
Dailey et al.