In vivo production of catalase containing haem analogues
Myriam Brugna*, Lena Tasse and Lars Hederstedt
Microbiology Group, Department of Biology, Lund University, Sweden
Introduction
Haem is present in most organisms as the prosthetic
group of proteins such as cytochromes, oxygenases,
haemoglobins, and catalases [1]. The versatile chemical
reactivity of the iron ion is expanded in the haem
molecule, giving the wide functional range of haem
proteins [2].
The physicochemical properties of the haem mole-
cule are, however, troublesome for cells [3]. At neutral
pH, haem is an amphiphilic, poorly water-soluble
molecule. In the reduced state, in the presence of
molecular oxygen, haem is also a potent catalyst for
reactive oxygen species formation. These toxic effects
must be avoided when haem is transported and stored
within cells. Haem, synthesized within the cell or
acquired from the environment, and destined to finally
be a prosthetic group of a haem protein, is therefore
Keywords
antibacterial agents; catalase;
Enterococcus faecalis; haem protein;
metal porphyrins
Correspondence
L. Hederstedt, Department of Biology,
Biology Building, Lund University,
So
¨
lvegatan 35, SE-22362 Lund, Sweden
Fax: +46 46 222 41 13

Haem (protohaem IX) analogues are toxic compounds and have been con-
sidered for use as antibacterial agents, but the primary mechanism behind
their toxicity has not been demonstrated. Using the haem protein catalase
in the Gram-positive bacterium Enterococcus faecalis as an experimental
system, we show that a variety of haem analogues can be taken up by bac-
terial cells and incorporated into haem-dependent enzymes. The resulting
cofactor-substituted proteins are dysfunctional, generally resulting in
arrested cell growth or death. This largely explains the cell toxicity of haem
analogues. In contrast to many other organisms, E. faecalis does not
depend on haem for growth, and therefore resists the toxicity of many
haem analogues. We have exploited this feature to establish a bacterial
in vivo system for the production of cofactor-substituted haem protein vari-
ants. As a pilot study, we produced, isolated and analysed novel catalase
variants in which the iron atom of the haem prosthetic group is replaced
by other metals, i.e. cobalt, gallium, tin, and zinc, and also variants con-
taining meso-protoheme IX, ruthenium meso-protoporphyrin IX and
(metal-free) protoporphyrin IX. Engineered haem proteins of this type are
of potential use within basic research and the biotechnical industry.
Structured digital abstract
l
MINT-7722358, MINT-7722368: katA (uniprotkb:Q834P5) and katA (uniprotkb:Q834P5)
physically interact (
MI:0915)bycopurification (MI:0025)
Abbreviations
Co-PP, cobalt protoporphyrin IX; Cu-PP, copper protoporphyrin IX; Fe-meso, iron meso-protoporphyrin; Fe-PP, protohaem IX; Ga-PP, gallium
protoporphyrin IX; Mg-PP, magnesium protoporphyrin IX; MIC, minimal inhibitory concentration; Ni-PP, nickel protoporphyrin IX; Pd-meso,
palladium meso-protoporphyrin IX; PP, protoporphyrin IX; Ru-meso, ruthenium meso-protoporphyrin IX; Sn-PP, tin protoporphyrin IX;
Zn-PP, zinc protoporphyrin IX.
FEBS Journal 277 (2010) 2663–2672 ª 2010 The Authors Journal compilation ª 2010 FEBS 2663
most likely transiently bound to specific protein factors

negative bacteria such as Escherichia coli, no outer
membrane constituting a possible barrier for the
uptake of porphyrin compounds; (c) E. faecalis does
not synthesize haem; (d) E. faecalis does not depend
on haem for growth, and grows well also without
haem; (e) if supplied with haem, the bacterium is
capable of aerobic respiration [7–10] and produces
two haem proteins – a membrane-bound cytochrome
bd respiratory quinol oxidase [7], and a cytoplasmic
monofunctional catalase [6]; and (f) the growth of
many bacteria is inhibited by noniron metalloporphy-
rins, but E. faecalis strains are generally resistant to
this type of compound ([11] and this work).
Catalase (EC 1.11.1.6) is present in most aerobic
organisms, and serves in part to protect the cells from
the toxic effects of hydrogen peroxide by catalysing its
decomposition into O
2
and H
2
O [12,13]. E. faecalis
catalase is a typical monofunctional catalase [14]. It is
a homotetrameric enzyme containing one proto-
haem IX (Fe-PP) molecule per KatA polypeptide of
478 amino acids. The structure of the E. faecalis
enzyme is rather complex (Fig. S1). Each monomer
has an extended arm-like N-terminal domain and a
major globular C-terminal domain, with the haem-con-
taining active site deeply buried in a b-barrel structure.
The four subunits are tightly associated with the N-ter-

magnesium protoporphyrin IX (Mg-PP), nickel proto-
porphyrin IX (Ni-PP), and zinc protoporphyrin IX
(Zn-PP), and for protoporphyrin IX (PP), was
> 150 lm, and for tin protoporphyrin IX (Sn-PP)
it was > 130 lm, i.e., these compounds are not
toxic to E. faecalis. Ruthenium meso-protoporphyrin
IX (Ru-meso) and iron meso-protoporphyrin IX
(Fe-meso) were found to be somewhat toxic, and palla-
dium meso-protoporphyrin IX (Pd-meso) very toxic
(Table 1).
In vivo synthesis of gallium-substituted catalase
Ga
3+
and Fe
3+
are similar in size and form a very
stable complex with PP. In contrast to iron, the gallium
ion is not oxidized or reduced under physiological
Cofactor-substituted catalase M. Brugna et al.
2664 FEBS Journal 277 (2010) 2663–2672 ª 2010 The Authors Journal compilation ª 2010 FEBS
conditions. Ga-PP is very toxic to Gram-negative
bacteria (MIC < 2 lm) and some Gram-positive bacte-
ria such as Staphylococcus species (MIC < 3 lm) [11].
Ga-PP seems to be transported into Gram-negative
bacteria via haem uptake systems [11]. The toxicity of
Ga-PP has been suggested to result from the incorpora-
tion of this haem analogue into haem proteins, which
thereby become nonfunctional, but this has not been
experimentally demonstrated [5]. Reasons for the
resistance of E. faecalis to Ga-PP (Table 1) could be

-tagged iron-containing (Fe-PP-KatA) and
Ga-PP-KatA catalases were purified from cell-free
extracts by using metal affinity chromatography. The
purity of preparations was evaluated by SDS ⁄ PAGE,
which, for both Fe-PP-KatA and Ga-PP-KatA,
showed one polypeptide band corresponding to KatA
with an apparent molecular mass of 54 kDa (Fig. 2).
In the case of Ga-PP-KatA, an additional protein
band of about 110 kDa was observed. This band cor-
responds to KatA dimer, as determined by immuno-
blot analysis (Fig. 1, lane I).
Amino acid analysis of preparations of isolated
Fe-PP-KatA and Ga-PP-KatA confirmed the composi-
tion of KatA polypeptide as deduced from the katA
Table 1. Toxicity of porphyrins, porphyrin concentrations used for
growth and presence of catalase protein in E. faecalis V583 ⁄
pLUF15. ND, not done.
Porphyrin added
to the growth
medium MIC (l
M)
Porphyrin
concentration
used in the
growth medium
(lM)
a
KatA
polypeptide
present

sequence, and was used for quantitative determination
of KatA protein. Metal analysis combined with the
protein analysis showed the presence of 1.12 mol of
iron atoms per mol of KatA polypeptide in Fe-PP-
KatA. Isolated Ga-PP-KatA contained 0.96 mol of
gallium per mol of KatA polypeptide, and only trace
amounts of iron (Table 2).
Porphyrin present in Fe-PP-KatA and Ga-PP-KatA
was extracted from the isolated catalase proteins by
using acid ⁄ acetone, and analysed by RP-HPLC. The
porphyrin of Ga-PP-KatA eluted as a single peak at
7.0 min. Reference Ga-PP had the same retention time.
Fe-PP and the porphyrin extracted from Fe-PP-KatA
both eluted at 8.5 min. This showed that Fe-PP and
Ga-PP are present in the isolated proteins, and
excluded the possibility that these metalloporphyrins
added to the culture are modified during transport into
E. faecalis or after being incorporated into KatA poly-
peptide, as is the case for catalase HPII of E. coli [16].
Enzymatic and spectroscopic properties of
Ga-PP-KatA
Isolated Ga-PP-KatA showed less than 1% catalase
activity as compared with Fe-PP-KatA (Table 2).
Enzyme activity measurement with the cytoplasmic cell
fraction from E. faecalis V583 ⁄ pLUF15 containing
Ga-PP-KatA as compared with that containing Fe-PP-
KatA showed the same relative results (data not
shown). This confirmed that Ga-PP-KatA is essentially
inactive, and that this is not due to inactivation during
isolation of the protein.

The resulting preparations were pure or contai-
ned some contaminating proteins (in the cases of
Co-PP-KatA, Sn-PP-KatA, and Fe-meso-KatA) as
evaluated by SDS ⁄ PAGE (gel not shown). The cata-
lase activities of purified proteins are presented in
Table 2. Fe-meso-KatA showed 35% activity as
Table 2. Properties of isolated normal and cofactor-substituted catalases. ND, not done.
Variant
Porphyrin
present
a
Metal content
(mol ⁄ mol KatA)
PP content
b
(mol ⁄ mol KatA)
Relative
activity
c
(%)
Co-PP-KatA ? Co, 0.70; Fe, 0.001 ND < 1
Cu-PP-KatA PP Cu, 0.04; Fe, < 0.01 1.1 2
Fe-PP-KatA Fe-PP Fe, 1.12 ND 100
Ga-PP-KatA Ga-PP Ga, 0.96; Fe, < 0.01 ND < 1
Mg-PP-KatA PP Mg, 0.02; Fe, 0.08 0.7 < 1
d
PP-KatA PP Fe, 0.02 1.2 2
Sn-PP-KatA Sn-PP Sn, 0.76; Fe, 0.001 ND < 1
Zn-PP-KatA ? and PP Zn, 0.58; Fe, 0.02 0.075 3
Fe-meso-KatA Fe-meso Fe, 0.65 ND 35

crosslinks are formed in porcine erythrocyte catalase
when the enzyme is stored at 4 °C for more than
1 week [19]. Disulfide bond formation can be excluded
in the case of E. faecalis catalase, because this protein
does not contain any cysteine, and proteins were
reduced before SDS ⁄ PAGE.
All isolated catalase variants were analysed for
cobalt, copper, iron, gallium, magnesium, ruthenium,
tin and zinc content. The major metal found in each
preparation was generally the same as that contained
in the metalloporphyrin added to the growth medium
(Table 2). Notable exceptions, however, were the prep-
arations of isolated Cu-PP-KatA and Mg-PP-KatA,
which contained only low amounts of metals; Cu-PP-
KatA contained 0.04 mol of copper per mol of KatA
polypeptide, and Mg-PP-KatA contained 0.02 mol of
magnesium per mol of KatA. Similarly, the prepara-
tion of isolated PP-KatA contained little metal,
< 0.08 mol of metal per mol of KatA polypeptide,
except for zinc, which was found at 0.13 mol per mol
of KatA. In the following, we deal separately with the
metal-substituted (Co-PP-KatA, Sn-PP-KatA, Zn-PP-
KatA, Fe-meso-KatA, and Ru-meso-KatA) and the
porphyrin-containing but metal-deprived (Cu-PP-
KatA, Mg-PP-KatA, and PP-KatA) catalases.
Characterization of metal-substituted catalases
Porphyrins present in the different preparations with
metal-substituted catalase were analysed by RP-HPLC.
For Fe-meso-KatA and Sn-PP-KatA, the chromato-
gram of the extracted porphyrin completely agreed

buffer (pH 8.0).
M. Brugna et al. Cofactor-substituted catalase
FEBS Journal 277 (2010) 2663–2672 ª 2010 The Authors Journal compilation ª 2010 FEBS 2667
maxima (396, 499, 531 and 619 nm as compared with
406, 504, 541 and 625 nm).
The fluorescence emission spectrum of the porphyrin
contained in Zn-PP-KatA, in a pyridine⁄ water ⁄
NaOH ⁄ Tween-80 mixture, was not similar to that of
the reference solution of Zn-PP in the same solvent,
indicating that the bound compound is not Zn-PP
(Fig. 4). Moreover, fluorescence spectral analysis
showed that isolated Zn-PP-KatA contained a small
amount of PP (0.075 mol of PP per mol of KatA poly-
peptide) (Fig. 4; Table 2). Zn-PP-KatA contained
0.58 mol of zinc per mol of KatA polypeptide
(Table 2). These results, taken together, indicate that
zinc in Zn-PP-KatA is bound to a nonfluorescent,
unidentified porphyrin compound. This conclusion is
consistent with the light absorption spectrum of
Zn-PP-KatA, which presents features reminiscent of
the spectrum of PP-KatA (maxima at 554, 574 and
629 nm) (see next section) and features of the unidenti-
fied zinc porphyrin (maxima at 421 and 670 nm)
(Fig. 3A–C).
Characterization of catalase obtained with PP in
the growth medium
The HPLC chromatogram of porphyrin extracted from
PP-KatA showed one major species, with the same
retention time (20 min) as PP (data not shown). To
further characterize and analyse the amount of

absorbance spectroscopy (Fig. 3C) and fluorescence
spectroscopy (Fig. 4) showed that PP was the major
porphyrin in the two proteins. Approximately 1 mol of
PP and 0.7 mol of PP per mol of KatA polypeptide
were found in Cu-PP-KatA and Mg-PP-KatA, respec-
tively (Table 2).
These findings suggested that the metal ion of the
porphyrin is removed during transport of the porphy-
rin from the medium to the catalase protein in the
cytoplasm. Alternatively, but less likely, the metal-
loporphyrin is incorporated into catalase and the metal
is subsequently lost from the protein, leaving the PP
bound to the protein.
Discussion
Our findings demonstrate, first, that the haem pros-
thetic group of catalase can be replaced by various
haem analogues. Moreover, they explain the general
cellular toxicity of noniron metalloporphyrins. They
also show the potential of the bacterium E. faecalis as
Fig. 4. Fluorescence emission spectra of PP, Mg-PP, Cu-PP, and
Zn-PP, and of the porphyrins contained in isolated PP-KatA, Mg-PP-
KatA, Cu-PP-KatA, and Zn-PP-KatA. The excitation wavelengths
used are indicated in Experimental procedures. Emission spectra
peak maxima are indicated by dotted lines. The solvent was pyri-
dine ⁄ water ⁄ NaOH ⁄ Tween-80 (see Experimental procedures for
details).
Cofactor-substituted catalase M. Brugna et al.
2668 FEBS Journal 277 (2010) 2663–2672 ª 2010 The Authors Journal compilation ª 2010 FEBS
an in vivo system for the production of cofactor-substi-
tuted haem proteins.

unable to biosynthesize haem. The usefulness of this
method is, however, restricted to compounds that are
not toxic to E. coli, thus excluding many metal-substi-
tuted porphyrins.
For the in vivo production of catalase containing
haem analogues, we grew E. faecalis strain V583 ⁄ -
pLUF15, which overproduces catalase polypeptide, in
the presence of the respective haem analogue in the
growth medium. Noniron metalloporphyrin com-
pounds are generally toxic to bacteria [11], but, with
the notable exception of Pd-meso, they do not inhibit
growth of E. faecalis (Table 1). This allowed us to pro-
duce and isolate catalases in which the normal iron
atom is replaced by other metals, i.e. cobalt, gallium,
tin, and zinc. We could also produce catalase proteins
containing Fe-meso and Ru-meso. Supplementation of
the haem-free growth medium with PP resulted, much
to our surprise, in PP-containing catalase, i.e. protein
containing a metal-free porphyrin group.
The hemH gene (Ef1989) of E. faecalis V583 appar-
ently encodes a ferrochelatase. Ferrochelatases catalyse
the last step in haem synthesis, i.e. the insertion of
ferrous iron into PP. As growth of bacteria in the
presence of PP did not result in catalase containing
Fe-PP, it appears that the hemH gene is not expressed,
or that the HemH protein lacks ferrochelatase activity
with PP. Lactococcus lactis and E. faecalis are closely
related bacteria. L. lactis contains a hemH gene (called
hemZ), and can apparently take up Fe-PP from the
growth medium. Indirect evidence with a HemZ-defi-

in bacteria, and the mechanisms by which haem is
incorporated into soluble and membrane-bound pro-
teins inside cells have not been determined [4]. It is evi-
dent from our results that these putative transport
components also work with haem analogues. The toxic-
ity of noniron metallo-PPs for bacteria has been sug-
gested, but not previously demonstrated, to result from
these compounds being taken up into the cell and
incorporated into vital haem proteins instead of haem
[5]. We show here that haem analogues are indeed
taken up and can be incorporated into haem proteins,
resulting in their inactivation. Pd-meso and, to a lesser
extent, Ru-meso were found to be toxic also for
E. faecalis (Table 1). This toxicity might be connected
to meso-protoporphyrin IX, as the bacterium tolerated
Fe-PP better than Fe-meso.
M. Brugna et al. Cofactor-substituted catalase
FEBS Journal 277 (2010) 2663–2672 ª 2010 The Authors Journal compilation ª 2010 FEBS 2669
The semisynthetic catalases that we have produced
in the present investigation are metalloproteins of a
type hitherto not described. Our results show that the
E. faecalis cell can be used for the in vivo production
of novel haem proteins. Following the same general
approach, one may be able to generate a great variety
of complex soluble and membrane-bound substituted
haem proteins.
Experimental procedures
Bacterial strains and growth conditions
E. faecalis V583 cells containing plasmid pLUF15 (a deriv-
ative of plasmid pAM401 containing the katA–his

Preparative cultures were grown in 1 L portions in 5 L baf-
fled Erlenmeyer flasks at 200 r.p.m. in a rotary incubator.
Cultures with porphyrins added were protected from light
to avoid possible photoeffects. The cells were harvested by
centrifugation, at 5000 g for 30 min, when the cultures
reached late exponential growth phase as determined from
the attenuance measured at 600 nm (D600 nm).
Isolation of cytoplasmic cell fraction and
purification of catalase
The cytoplasmic fraction of E. faecalis cells was prepared
essentially as described previously [6], except in the case of
large-scale preparations. The cells from 3 L cultures were
suspended in 50 mm potassium phosphate buffer (pH 8.0),
incubated with 1 mgÆmL
)1
lysozyme at 37 °C with shaking
for 1 h, and then broken in a French pressure cell operated
at 16 000 p.s.i. The His
6
-tagged catalases were isolated from
the cytoplasmic cell fraction, corresponding to a 3 L culture,
with a 1 mL HiTrap chelating column (Pharmacia Bio-
tech), loaded with nickel ions, according to the manufac-
turer’s instructions. The cytoplasmic fraction,
supplemented with 300 mm NaCl and 1 mm histidine, was
loaded onto the affinity column, equilibrated in 50 mm
potassium phosphate buffer (pH 8.0) containing 300 mm
NaCl and 1 mm histidine. This buffer was used to wash
the column, and His
6

spectrofluorophotometer, using excitation wavelengths of
406 nm for PP and Cu-PP, 424 nm for Zn-PP, and 418 nm
for Mg-PP. The excitation and emission slits were 3 and
10 nm, respectively. Known concentrations of PP dissolved
in the same solvent as the samples were used for calibration
of the fluorometer readings.
Other methods
Extraction of porphyrins from purified catalase protein and
analysis by RP-HPLC were performed as described by Sone
and Fujiwara [31]. Protein concentrations were determined
using the bicinchoninic acid assay (Pierce Chem Co.), with
BSA as the standard. Concentrations of KatA were
determined by the combined use of quantitative amino acid
analysis (performed at the Department of Laboratory Medi-
cine, Malmo
¨
University Hospital), pyridine haemochromogen
Cofactor-substituted catalase M. Brugna et al.
2670 FEBS Journal 277 (2010) 2663–2672 ª 2010 The Authors Journal compilation ª 2010 FEBS
analysis [32], and rocket immunoelectrophoresis. Immuno-
blot and rocket immunoelectrophoresis were performed
with rabbit anti-KatA serum as previously described [6].
Catalase activity was assayed as described previously [6].
Light absorption spectra were recorded with a Shima-
dzu UV-2101PC spectrophotometer. Metal content analysis
was performed by using inductively coupled plasma MS.
Acknowledgements
This work was supported by grant 621-2007-6094 from
the Swedish Research Council and a grant from the
Crafoord Foundation. M. Brugna was the recipient of

dependent oxidative phosphorylation in Streptococcus
faecalis. J Gen Microbiol 58, 247–260.
9 Pritchard GG & Wimpenny JW (1978) Cytochrome
formation, oxygen-induced proton extrusion and respi-
ratory activity in Streptococcus faecalis var. zymogenes
grown in the presence of haematin. J Gen Microbiol
104, 15–22.
10 Ritchey TW & Seeley HW (1974) Cytochromes in
Streptococcus faecalis var. zymogenes grown in a
haematin-containing medium. J Gen Microbiol 85,
220–228.
11 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.
12 Zamocky M, Furtmu
ˆ
ller PG & Obinger C (2008)
Evolution of catalases from bacteria to humans.
Antioxid Redox Signal 10, 1527–1547.
13 Chelikani P, Fita I & Loewen PC (2004) Diversity of
structures and properties among catalases. Cell Mol Life
Sci 61, 192–208.
14 Ha
˚
kansson KO, Brugna M & Tasse L (2004) The
three-dimensional structure of catalase from Enterococ-
cus faecalis. Acta Crystollogr D Biol Crystallogr 60,
1374–1380.
15 Prakash K, Prajapati S, Ahmad A, Jain SK & Bhakuni

23 Deniau C, Couprie J, Simenel C, Kumar V, Stojiljkovic
I, Wandersman C, Delepierre M & Lecroisey A (2001)
1
H,
15
N and
13
C resonance assignments for the gallium
protoporphyrin IX–HasA
sm
hemophore complex.
J Biomol NMR 21, 189–190.
24 Balog E, Galantai R, Ko
¨
hler M, Laberge M & Fidy J
(2000) Metal coordination influences substrate binding
in horseradish peroxidase. Eur Biophys J 29, 429–438.
25 Watanebe Y (2002) Construction of haem enzymes:
four approaches. Curr Opin Chem Biol 6, 208–216.
M. Brugna et al. Cofactor-substituted catalase
FEBS Journal 277 (2010) 2663–2672 ª 2010 The Authors Journal compilation ª 2010 FEBS 2671
26 Woodward JJ, Martin NI & Marletta MA (2007) An
Escherichia coli expression-based method for haem
substitution. Nat Methods 4, 43–45.
27 Duwat P, Sourice S, Cesselin B, Lamberet G, Vido K,
Gaudu P, Loir Yl, Violet F, Loubie
`
re P & Gruss A
(2001) Respiration capacity of the fermenting bacterium
Lactococcus lactis and its positive effects on growth and

copy-edited or typeset. Technical support issues arising
from supporting information (other than missing files)
should be addressed to the authors.
Cofactor-substituted catalase M. Brugna et al.
2672 FEBS Journal 277 (2010) 2663–2672 ª 2010 The Authors Journal compilation ª 2010 FEBS