Báo cáo khoa học: GHP, a new c-type green heme protein from Halochromatium salexigens and other proteobacteria potx

GHP, a new c-type green heme protein from
Halochromatium salexigens and other proteobacteria
Gonzalez Van Driessche
1
, Bart Devreese
1
, John C. Fitch
2
, Terrance E. Meyer
2
,
Michael A. Cusanovich
2
and Jozef J. Van Beeumen
1
1 Department of Biochemistry, Microbiology and Physiology, Laboratory for Protein Biochemistry and Protein Engineering, Ghent University,
Belgium
2 Department of Biochemistry and Molecular Biophysics, University of Arizona, Tucson, AZ, USA
Purple phototrophic bacteria produce a large variety
of soluble electron-transfer proteins, the most wide-
spread and abundant of which are cytochromes c
2
, c
4
,
c¢, reaction center tetra-heme cytochrome c, flavocyto-
chrome c–sulfide dehydrogenase, bacterial ferredoxin,
and high-potential iron-sulfur protein [1,2]. Minor
components are often difficult to purify and generally
have not been characterized to any extent. Some are
truly soluble, but others exhibit low solubility, or they

are proteolytic fragments of more abundant mem-
brane-bound proteins. Cytochrome c
4
, for example, is
normally membrane-bound, but is shown to be present
at low concentrations in the soluble fraction of many
species of bacteria. Others, such as Sphaeroides heme
protein (SHP) and di-heme cytochrome c, are rarely
observed, although they are also widespread. In Rho-
dobacter sphaeroides, the SHP gene is associated with
those for a membrane-spanning cytochrome b and the
Keywords
cysteic acid; cytochrome c; green heme
protein; Halochromatium salexigens; mass
spectrometry
Correspondence
J. Van Beeumen, Universiteit Gent,
Vakgroep Biochemie, Fysiologie en
Microbiologie, Laboratorium voor
Eiwitbiochemie en Eiwitengineering,
Ledeganckstraat 35, B-9000 Gent, Belgium
Tel: +32 9264 5109
Fax: +32 9264 5338
E-mail:
Database
The protein sequence data reported in this
paper will appear in the UniProt Knowledge-
base under the accession number P84848
(Received 21 February 2006, revised 14
April 2006, accepted 27 April 2006)

doi:10.1111/j.1742-4658.2006.05296.x
We have isolated a minor soluble green-colored heme protein (GHP) from
the purple sulfur bacterium, Halochromatium salexigens, which contains
a c-type heme. A similar protein has also been observed in the purple
bacteria Allochromatium vinosum and Rhodopseudomonas cryptolactis. This
protein has wavelength maxima at 355, 420, and 540 nm and remains
unchanged upon addition of sodium dithionite or potassium ferricyanide,
indicating either an unusually low or high redox potential, respectively.
The amino-acid sequence indicates one heme per peptide chain of 72 resi-
dues and reveals weak similarity to the class I cytochromes. The usual sixth
heme ligand methionine in these proteins appears to be replaced by a cys-
teine in GHP. Only one known cytochrome has a cysteine sixth ligand,
SoxA (cytochrome c-551) from thiosulfate-oxidizing bacteria, which is low-
spin and has a high redox potential because of an un-ionized ligand. The
native size of GHP is 34 kDa, its subunit size is 11 kDa, and the net charge
is )12, accounting for its very acidic nature. A database search of complete
genome sequences reveals six homologs, all hypothetical proteins, from
Oceanospirillum sp., Magnetococcus sp., Thiobacillus denitrificans, Dechloro-
monas aromatica, Thiomicrospira crunogena and Methylobium petroleophi-
lum, with sequence identities of 35–64%. The genetic context is different
for each species, although the gene for GHP is transcriptionally linked to
several other genes in three out of the six species. These genes, coding for
an RNAse, a protease ⁄ chaperone, a GTPase, and pterin-4a-carbinolamine
dehydratase, appear to be functionally related to stress response and are
linked in at least 10 species.
Abbreviations
CID, collision-induced dissociation; GHP, green heme protein; PCD, pterin-4a-carbinolamine dehydratase.
FEBS Journal 273 (2006) 2801–2811 ª 2006 The Authors Journal compilation ª 2006 FEBS 2801
di-heme cytochrome c, and is probably regulated dif-
ferently from the major cytochromes [3]. One of the

best characterized species in this context is the purple
bacterium Allochromatium vinosum, from which six
minor cytochromes have been reported, for example
cytochrome c-553(550), which is a c
4
, and cyto-
chrome c-551, which is a c
8
[4,5]. In this organism, we
have observed a green heme protein (GHP) which we
have now succeeded in purifying from the related pur-
ple bacterium Halochromatium salexigens, where it is
somewhat more abundant, and we have characterized
it sufficiently to identify it in other genomes, either
from its gene sequence or from its spectral characteris-
tics. H. salexigens is interesting because it has a 10%
salt optimum and has a higher sulfide tolerance than
other purple sulfur bacteria studied [6]. As expected,
the same major electron-transfer proteins were isolated
from the soluble fraction of H. salexigens, including
cytochromes c
4
, c¢, reaction center tetra-heme cyto-
chrome c, flavocytochrome c–sulfide dehydrogenase,
bacterial ferredoxin, and high-potential iron-sulfur
protein [7], but there are some interesting proteins not
previously characterized from A. vinosum or other
members of the Chromatiaceae, such as the photo-
active yellow protein [8].
We here present the complete amino-acid sequence

and some specific characteristics of the unusual bacter-
ial GHP, and show that it may be a member of a so
far unknown group of class I c-type cytochromes.
Results
Isolation of GHP
The soluble proteins from the H. salexigens superna-
tant were purified as described in the Experimental
procedures section. The ratio A
280
⁄ A
420
was 0.38 for
the best green-colored protein fractions. The pyridine
hemochrome spectrum displayed a peak at 550 nm,
indicating that GHP contains a c-type heme with an
absorption coefficient of 68 mm
)1
Æcm
)1
Æheme
)1
at
420 nm. The absorption spectrum has wavelength
maxima at 420, 355, and 540 nm, as shown in Fig. 1.
GHP was unaffected by the reducing agent sodium
dithionite or by the oxidant potassium ferricyanide,
indicating that it has either a very low or high redox
potential, respectively. The usual wavelength maximum
for the Soret peak of oxidized low-spin c-type cyto-
chromes is 410–415 nm and that of high-spin heme is

near 400 nm. In the visible region, the corresponding
maxima are at 530 and 500 nm, which is unlike GHP
in either case. The Soret peak of reduced low-spin
cytochromes is near 416–419 nm, and there are two
peaks in the visible region near 525 and 550 nm, which
is also unlike GHP. Reduced high-spin proteins have
Soret peaks near 420–430 nm and a single broad peak
in the visible region in the vicinity of 540–550 nm,
characteristics that are most similar to those of GHP.
However, we were left with insufficient quantities to
characterize the spectral properties further, which must
await cloning and overproduction. Additional data on
heme ligation will be presented below. The native size
of GHP was estimated by gel filtration to be 34 kDa.
SDS ⁄ PAGE, however, indicates a mass of 11 kDa.
Amino-acid sequence of GHP and molecular
mass
The complete amino-acid sequence (Fig. 2) of GHP
was determined by a combination of automated
Edman degradation and tandem MS of peptides gener-
ated by different digestions on apoprotein and native
protein (Table 1). GHP contains 72 amino-acid resi-
dues, and the measured mass (Fig. 3A) of component
A (8970.3 Da) is in good agreement with the theoret-
ical mass of the complete sequence containing one
heme group (8968.3 Da). Component B (8840.9 Da)
lacks the N-terminal glutamic acid (8839.2 Da), as
pointed out in Table 1. Isoform I, represented by pep-
tides S5 and S6, has the sequence ALHGQVRM.
Although expressed in low amount, the presence of a

second isoform (II) was found by MS analysis of pep-
tide S3, originating from the nonspecific cleavage of
Glu-C endoproteinase at the aspartic acid residues 34
and 42, of which the N-terminal sequence was deter-
mined to be: ALHGQVYD (Table 1).
There are three cysteines in the GHP. Non-modified
and reduced ⁄ aminopropylated peptides generated from
250
600
Absorbance
Wavelength (nm)
420
540
355
280
0.0
1.0
Fig. 1. UV-visible spectrum of green colored heme protein (GHP) in
20 m
M Tris ⁄ HCl, pH 7.5. The spectrum was unaffected by the addi-
tion of sodium dithionite or potassium ferricyanide. The protein con-
centration was  11 l
M.
GHP, a new cytochrome c from H. salexigens G. Van Driessche et al.
2802 FEBS Journal 273 (2006) 2801–2811 ª 2006 The Authors Journal compilation ª 2006 FEBS
the Glu-C endoproteinase digestion on the native pro-
tein were deposited on a MALDI-MS target plate to
determine which peptides contain a free cysteine. Pep-
tides covering the sequence regions 32–44 (1457.7 Da)
and 32–48 (1941.9 Da) were shown to contain a modi-

fied cysteine, by a mass shift of 57 Da to 1514.7 Da
and 1998.9 Da, respectively (data not shown). As
expected from the conventional CXXCH heme-binding
motif, the two other cysteines are involved in covalent
heme binding [9].
Heme iron oxidation state
As mentioned above, the UV spectrum (Fig. 1) of the
GHP shows a 5–10 nm red-shift of the Soret peak,
suggesting unusual heme ligation if the protein is oxid-
ized low-spin. We therefore aimed to examine the nat-
ure of this ligation by MS techniques. The analyses
showed that heme-containing peptides were present in
the chromatographic fractions S7 and S8 (Table 1)
obtained after digestion of the native protein with
Glu-C endoproteinase. The heme is thus covalently
bound to peptides covering the N-terminal regions
1–21 and 1–31, the former originating by incomplete
cleavage of a Glu-Val bond (Fig. 2). No evidence for
cross-linking of the heme with other peptides of the
protein has been found. The heme-containing peptides
were also subjected to MALDI MS fragmentation ana-
lysis, and one example of such an MS ⁄ MS spectrum is
presented in Fig. 3B. For comparison, Fig. 3C shows
the same type of spectrum of the heme-containing pep-
tides from horse cytochrome c, Paracoccus sp. cyto-
chrome c
4
and Rhodoferax fermentans cytochrome c¢,
the spectra of which are identical. In general, the pro-
toporphyrin group bears two negative charges, result-

ing from two ionized pyrrole rings. As the central iron
atom has either two or three positive charges, the
heme group as a whole is either neutral or singly posi-
tively charged if the heme propionates are excluded
from consideration [10,11]. On the basis of the chem-
ical formula C
34
H
32
N
4
O
4
Fe, a mass of 616.2 Da
should be measured during mass analysis, and that is
exactly what we observe in Fig. 3C. GHP, however,
releases a singly charged heme ion of 617.2 Da
(Fig. 3B), which suggests that the heme as a whole is
neutral, because of Fe(II) and considering a proton to
be responsible as a charge carrier. From these data, it
appears that the heme iron from the GHP is in the
Fe(II) oxidation state.
Evidence for oxidized cysteine in the native
protein
In the MALDI mass spectrum of the non-modified
Glu-C peptide 32–48, peaks appeared at differences
of 16, 32, 48 and 64 mass units from the major
peak at 1943.0 Da (Fig. 4A). The peak at m ⁄ z
1959.0 Da can theoretically be interpreted as the oxi-
dation product of either a methionine or a cysteine

to Met-SO or Cys-SOH, respectively, and the peak
Fig. 2. Sequence alignment of H. salexigens GHP (HS) with homologs ( and from
Oceanospirillum sp. MED92 (OS), Thiomicrospira crunogena (TC), Magnetococcus sp. Mc-1 (MC), Dechloromonas aromatica (DA), Thiobacil-
lus denitrificans (TD) and Methylobium petroleophilum (MP). Identical residues are in bold. Percentage identities are presented at the right of
the alignment.
G. Van Driessche et al. GHP, a new cytochrome c from H. salexigens
FEBS Journal 273 (2006) 2801–2811 ª 2006 The Authors Journal compilation ª 2006 FEBS 2803
at 1975.0 Da may represent Cys-SO
2
H or MetSO
2
.
Proof that the mass of 1959.0 Da refers to the pep-
tide containing Cys-SOH is found from the spectrum
in Fig. 4B, which shows the MALDI MS ⁄ MS data
for the precursor ion at 1975.0 Da. There is clear
evidence for the presence of the further oxidation
Table 1. Mass and sequence analysis of selected peptides obtained after digestion of purified apoprotein with trypsin (T) and Asp-N endo-
proteinase (D), and of native protein with Glu-C endoproteinase (S). Amino acids identified by Edman degradation and MS tandem fragmen-
tation are presented in normal and bold letters, respectively. Selected peptides are numbered following their elution during chromatography.
-, Amino acid not identified during sequence analysis but already determined from other peptides; NS, no signal during MS analysis.
Peptide Sequence
Measured
mass
Calculated
mass
Sequence
position
T1 RVESLDALHGQVR 1478.8
a

1478.8 29–41
T2 AIDAQALVDQNC 2850.8
b,c
2850.3 3–28
T3 EAIDAQALVDQN 2979.9
b,c
2979.3 2–28
T4 3109.9
b,c
3110.3 1–28
T5 EYYNFEP 960.4
a
960.4 66–72
T6 M-EQNLELT-FDDQVDAVTTLLNR NS (2855.2) 42–65
D1 DQNCT- -HGSEYTR 1668.7
a
1668.7 11–25
D2 DER- - -SL 1002.5
a
1002.5 26–33
D3 DALHGQVRM 1025.5
a
1025.5 34–42
D4 DREYYNFEP 1231.8
a
1231.5 64–72
d
D5 DALHGQVRMCE 1612.8
a
1612.7 34–47

D6 TTLLNREYYN- - - 1944.1
a
1943.9 57–72
D7 DDQVDAVTTLLNREYYN 2402.3
a
2401.1 53–72
D8 A GQVRMCEQN 2289.2
a
2289.1 34–52
B GQVRMCE 2746.5
a
2746.3 34–56
S1 QNLE NS (502.2) 45–48
S2 VYTRDERRVE 1321.9
a
1321.7 22–31
S3 ALHGQVYD 901.6
a
901.4 35–42
e
S4 YYNFEP 831.4
b
831.3 67–72
S5 SLDALHGQVRM(Cys-SO
3
)EQNLE 1990.0
a
1989.8 32–48
S6 SLDALHGQVRMCEQNLE 1942.0
a

1941.9 32–48
S7 A AID-QALVDQ- - -G 4110.1
b
4110.1 1–31
B EAIDAQALVDQN-TG-HGSEVYT-D 3981.1
b
3981.0 2–31
C 3852.1
b
3852.0 3–31
S8 A AID-Q-LV-QN-TG 2805.8
b
2805.8 1–21
B EAIDAQ-LVDQN-TG-HGS 2676.8
b
2676.7 2–21
C 2547.8
b
2547.7 3–21
S9 LTWFDDQVDAVTTLLNRE 2134.8
a
2135.0 49–66
a
MALDI.
b
ESI.
c
For complete removal of the covalently bound mercury molecule, the peptide was reduced with dithiothreitol.
d
Nonspecific

cleavage originated by deamidation of asparagine to aspartic acid.
e
Peptide from GHP isoform II.
001
0
%
0
578
0029
mass (Da) mass (Da)
9.0488
3
.0798
*
*
*
*
)27-2(B
)27-1(A
001
0
%
006
03
6
2.81
6
2.716
2.616
001

0
%
006
036
2.716
2.816
2.916
*
*
*
mass (Da)
A
B
C
Fig. 3. (A) MALDI mass spectrum of native GHP. The asterisks represent an additional mass of 16 Da compared with the mass of the peaks
preceding it. (B) MS ⁄ MS fragmentation of the heme-containing peptide of the GHP showing a detailed spectrum of the heme. (C) The same
fragmentation analysis as in (B), but now for horse cytochrome c, as well as for Paracoccus diheme cytochrome c and Rhodoferax fermen-
tans cytochrome c¢.
GHP, a new cytochrome c from H. salexigens G. Van Driessche et al.
2804 FEBS Journal 273 (2006) 2801–2811 ª 2006 The Authors Journal compilation ª 2006 FEBS
product Cys-SO
2
H(b
12
-b
11
) and no evidence for the
oxidation of Met-SO to methionine sulfone (Met-
SO
2

). Moreover, the small peak at 1277.7 Da refers
to the species b
12
-H
2
SO
2
, due to the phenomenon of
b-elimination under collision-induced dissociation
(CID) conditions [12,13]. Concerning the mass peak
at 1991.0 Da (Fig. 4A), the MALDI MS ⁄ MS spec-
trum of the precursor ion at 1991.0 Da (Fig. 4C)
shows a distance of 151 Da between the b
12
–b
11
ions, which fits perfectly the residual mass of cysteic
acid (Cys-SO
3
H), an interpretation that is further
supported by the peak at 1277.7 Da for the species
b
12
–H
2
SO
3
. When analyzed by fragmentation analysis
after electrospray ionization (ESI MS ⁄ MS) (Fig. 4D),
the same precursor ion does not show a clear b

12
species, but the mass difference between b
13
–b
11
ions
(280.2 Da) fits exactly the sequence cysteic acid–
glutamic acid. This assignment is further supported
by the loss of 82 Da for b
13
–H
2
SO
3
(at 1407.1 Da)
and of 82 Da from the undetected b
12
species (expec-
ted at 1360.1 Da), giving rise to b
12
–H
2
SO
3
at
1278.0 Da. Finally, the peak at 2007.0 Da in Fig. 4A
can be explained by the presence of one cysteic acid
and one methionine sulfoxide in the same peptide.
Although there is evidence for the presence of singly
oxidized methionine in the protein, it occurs only

after complete oxidation of the cysteine, which is
remarkable because methionines are known to be the
most susceptible residues to oxidation by almost all
forms of reactive oxygen species [14]. This indicates
that the modification of cysteine to cysteic acid at
position 43 may have occurred in vivo and that it
did not originate by nonspecific oxidation during
isolation or storage of the GHP.
* - H SO (82 Da) from b12 and/or b13
23
100
%
0
1943.0
1959.0 1975.0
1991.0
2007.0
1937
2011
Cys-SOH
Cys-SO H
3
Cys-SO H
2
Met-SO
and Cys-SO H
3
100
%
0

1056
1532
1208.7
1077.6
1251.7
1472.7
1343.7
Met
Glu
Mass (Da)
Mass
A
B
C
D
(Da)
1277.7
Cys-SO H
2
b10
b11
b12
b13
-HSO
22
100
%
0
800 1800
822.5

1208.6
1077.6
1251.7
1488.7
Mass (Da)
1277.7
921.5
1616.8
1730.8
1359.6
Val
Asn
Arg
Glu
Met
Gln
b8
b9
b11
b13
b14
b10
b12
b15
Cys-
SO H
3
1489.1
1407.1
1077.9

921.7
1844.3
1731.2
1617.2
1278.0
1208.9
1252.0
822.6
b8
b9
b10
b11
b13
b15
b14
b16
Val
Asn
Arg
LeuMet
Gln
100
%
0
800
1800
*
*
Mass (Da)
*

Cys-SO H
+Glu
3
Fig. 4. (A) Detailed MALDI mass spectrum of a Glu-C peptide from sequence region 32–48 (theoretical mass 1941.9 Da); the interpretation
of the molecular species follows from the spectra B–D. (B) MALDI MS ⁄ MS characterization of the precursor ion of 1975.0 Da. (C) MALDI
MS ⁄ MS characterization of a precursor ion of 1991.0 Da. (D) ESI-MS ⁄ MS of the ion at m ⁄ z 997.5 Da (z ¼ 2). Ion fragments are indicated
using the international convention for the nomenclature of MS fragmentation. All values represent the singly protonated species.
G. Van Driessche et al. GHP, a new cytochrome c from H. salexigens
FEBS Journal 273 (2006) 2801–2811 ª 2006 The Authors Journal compilation ª 2006 FEBS 2805
Discussion
Sequence identification
Searches using the blast engine returned six homologs,
annotated as hypothetical proteins, with identities ran-
ging from 38% to 44% with H. salexigens GHP, as
shown in the alignment of Fig. 2. The Magnetococcus
sp., Dechloromonas aromatica, Thiomicrospira crunogena
and Thiobacillus denitrificans GHP homologs are more
closely related to one another, at 46–64% identity. The
Methylobium petroleophilum and Oceanospirillum GHP
homologs, at 31–46% identity, are the most divergent
on a percentage basis. The sizes of translated genes and
the location of the signal peptides for the homologs
indicate that the H. salexigens GHP sequence represents
the whole protein and that it is not a proteolytic
fragment of a larger membrane-bound protein. The
sequence alignment furthermore suggests that GHP may
be a class I cytochrome, as there is a consensus
GxxxHxxxCxxCH heme-binding motif near the N-ter-
minus and a YY motif near the C-terminus. Secondary-
structure prediction (using predator) indicates that

GHP from all seven species may be composed of three
helices, an N-terminal helix that binds the heme (resi-
dues 4–18), a central helix (residues 28–43), and a C-ter-
minal helix (residues 54–67), which is consistent with the
identification of GHP as a class I heme protein [1]. The
usual distance of the YY motif from the methionine resi-
due as the heme sixth ligand in class I cytochromes is 17
residues, although there are exceptions. This distance
refers to L49 of GHP, a position that is not conserved in
the GHP homologs. There is a methionine residue in
H. salexigens GHP at position 42, but it is not con-
served in the homologs either. Instead, they contain a
conserved cysteine at position 43, which suggests that
this residue may have functional relevance, either as a
heme ligand or because of its presumed location near
the face of the heme.
Oxidation of a cysteine residue to cysteine sulfonic
acid and ⁄ or sulfinic acid in nonheme iron metalloen-
zymes has previously been observed in a serine ⁄ threon-
ine and a tyrosine phosphatase [15,16], a nitrile
hydratase [17], a cysteine dioxygenase [18] and an inacti-
vated peptide deformylase [19]. The last of these is very
labile under oxygen stress, resulting in the oxidation of
the catalytic Fe
2+
center of the deformylase into the
inactive Fe
3+
ion, accompanied by the conversion of
Cys90, a ligand of the catalytic iron center, into a cys-

teine sulfonic acid. On the other hand, the light-sensitive
nitrile hydratase from Rhodococcus sp. N-771 is a bac-
terial metalloenzyme, catalyzing the hydration of nitriles
to the corresponding amides, of which Cys112 is
post-translationally modified to a cysteine sulfinic acid
in the native form [17]. The biological role of Cys-SO
2
H
in the photoresponse and ⁄ or catalysis remains unclear.
In addition to the five enzymes in which oxidized cys-
teine residues appear to play a role in the catalytic activ-
ity, the peroxiredoxins should also be mentioned with
respect to the present discussion on GHP. Peroxiredox-
ins are a family of peroxidases that reduce hydrogen
peroxide and alkyl hydroperoxides to water and corres-
ponding alcohols, respectively, with the use of reducing
equivalents provided by thiol-containing proteins such
as thioredoxin or glutathione [20]. The active-site cys-
teine is selectively oxidized to cysteine sulfenic acid dur-
ing catalysis. The sulfenic group is usually unstable and
can be further oxidized to sulfinic acid and ⁄ or cysteic
acid, leading to inactivation of peroxidase [21,22]. The
presence of sulfinic acid in peroxiredoxin was originally
thought to be irreversible, but is now known to be
repaired by reaction with one of the two cysteines of the
other subunit of the homodimer, forming a disulfide,
which is subsequently reduced by thioredoxin ⁄ thi-
oredoxin reductase [23,24]. Whether the presence of a
cysteic acid in GHP originates by post-translational
modification and ⁄ or is the consequence of the need of

a higher oxidation state in the catalytic function of
the protein, as it is in peroxiredoxin, will remain
unanswered until evidence is found that points to its bio-
logical role.
Origin of the H. salexigens GHP green color and
nature of heme ligation
As the GHP from H. salexigens has a conventional
c-type heme (616.2 Da), as determined by tandem MS
analysis, and is covalently linked to the protein
through the classical CXXCH sequence, we conclude
that the color of the GHP is not a consequence of any
covalent heme modification, but is rather the result of
an unusual heme ligation or heme environment.
Low-spin heme proteins are usually colored red, but
high-spin proteins are either olive green or brown. We
presented evidence from homology to class I cyto-
chromes that the heme ligand in GHP may be a cys-
teine, and if so, it should be low-spin and red in color.
The absorption spectrum is most similar to reduced
high-spin proteins, and MS data indicate that the heme
is present in the reduced state. However, neither result
is conclusive as very few proteins with cysteine ligands
have been spectrally characterized and the heme could
have become reduced during the isolation of the pep-
tides. Furthermore, the heme should have a low redox
potential if the cysteine is ionized, as in the M80C
mutant of cytochrome c [25], or a high-potential if
GHP, a new cytochrome c from H. salexigens G. Van Driessche et al.
2806 FEBS Journal 273 (2006) 2801–2811 ª 2006 The Authors Journal compilation ª 2006 FEBS
un-ionized, as in SoxA [26]. In fact, SoxA contains

two His ⁄ Cys-ligated low-spin hemes, one of which can-
not be reduced by dithionite and is presumably the
catalytic heme with Cys ionized, whereas the other
becomes protonated upon reduction. Furthermore, a
post-translational modification of the catalytic heme
ligand converts it into a persulfide, which is presuma-
bly related to its enzymatic role in sulfur oxidation.
Whether such a modification occurs in GHP remains
to be determined. It is remarkable that the M80C
mutant of cytochrome c has an unusually low redox
potential, a red-shifted Soret peak with lower absorp-
tion coefficient than the wild-type, and a prominent
delta peak, at 355 nm, similar to GHP [25]. At present,
the heme environment in GHP and its sixth ligand
remain speculative, although we favor a low-spin heme
with an ionized cysteine ligand.
Genetic organization
At present, we do not have experimental evidence for
the biological function of the GHP. Its identification
as a heme protein, however, is important in that it
elevates the homologs from being hypothetical to genes
with a clear structural feature, thus improving annota-
tions in future genome sequences. Furthermore, the
arrangement of a gene in an operon may provide clues
to the functional role of the protein under study. In
Fig. 5, we can see that the context of each of the
homologous GHP genes is different, although related.
The Dechloromonas GHP gene overlaps with that for a
GTPase (which contains a nucleotide-binding zinc fin-
ger domain) by four bases on the downstream side,

and overlaps a gene for a membrane-bound prote-
ase ⁄ chaperone by four bases on the upstream side as
well. Furthermore, the protease is in the opposite ori-
entation to that of an mRNA-degrading RNAse. For
most species that have these three associated genes,
such as Nitrosomonas europaea and Chromobacterium
violaceum, a pterin-4a-carbinolamine dehydratase
(PCD) gene coding for the enzyme PCD, which regen-
erates the phenylalanine hydroxylase cofactor and also
serves a regulatory role, takes the place of GHP. Thio-
bacillus is an exception in that the GHP gene is 11
muecaloivmuiret
ca
bomo
rh
C
snacifir
t
inedsullicaboihT
e
sA
NR
enorepahc/esaetorP
DCP
esaPTG
nieto
r
PemeHneerG
e
sa

lor
dyH
esaremosiediflusid-loihT
mu
lihpoe
lort
ep
mu
i
bo
l
yht
e
M
aeaporuesanomosortiN
67EN57
EN
77EN
87EN
4-
4-
563
2VC6632V
C7632V
C8632
VC
851
11
4-
4-

821
27
511
9381dbT
0
48
1dbT
1481dbT
2481dbT
aci
tam
or
asano
mo
rolhce
D
75
4-
4-
10
03o
raD
2003oraD
3003oraD
4003oraD
31-
8-
5242AepM
6242AepM
7242AepM

4-
4
-
1192AepM
4192AepM2192AepM
55
3192AepM
69
424
2AepM
rotalugerlanoitpircsnarT
Fig. 5. Genetic organization of the GHP homologs from N. europeae, C. violaceum, T. denitrificans, D. aromatica and M. petroleophilum. The
numbers above the central lines indicate the separation of genes; a minus sign refers to an overlap with an adjacent gene, and a number
without a sign refers to the stretch of bases between adjacent genes. Numbers below indicate the position on the chromosomes, following
the annotations from the Joint Genome Institute ( />G. Van Driessche et al. GHP, a new cytochrome c from H. salexigens
FEBS Journal 273 (2006) 2801–2811 ª 2006 The Authors Journal compilation ª 2006 FEBS 2807
bases upstream of the PCD, which in turn is 158 bases
upstream of the homologous GTPase gene. However,
the Thiobacillus genome appears to lack the prote-
ase ⁄ chaperone. The Methylobium GHP gene overlaps a
thiol disulfide isomerase gene, which overlaps a hydrol-
ase gene by eight bases. At another locality on the gen-
ome, the homologous protease ⁄ chaperone, the PCD
and the GTPase form an operon opposite the RNAse,
similar to the arrangement in Dechloromonas but with
the PCD replacing the GHP. Thiomicrospira crunogena
GHP is not part of an operon, but is somewhat loosely
associated with a protease ⁄ chaperone and DNA heli-
case. The above four enzymes are clustered on another
part of the chromosome (data not shown).

N. europeae, Polaromonas sp., C. violaceum, and
Rhodoferax ferrireducans contain the same operon
structure as Methylobium, but they lack a correspond-
ing gene for GHP. The PCD gene is located elsewhere
on the chromosome in Ralstonia eutropha, Azoarcus
sp., Methylococcus capsulatus, Bordetella pertussis,
Methylobacillus flagellatus, and Burkholderia cepacia,
but the RNAse, protease, and GTPase remain clus-
tered. There are many other species that have these
genes, although there are cases in which only two of
them are organized in a cluster, such as the RNAse
and GTPase in Pseudomonas aeruginosa or the RNAse
and the protease in Escherichia coli.
There are thus sufficient examples of these gene
arrangements to indicate that their occurrence is not
random. Furthermore, the GHP does not appear to
have been randomly inserted in the cluster, as it displa-
ces the PCD in Dechloromonas and the protease in
Thiobacillus, indicating that these insertions were inde-
pendent evolutionary events. The RNAse, called ‘orn’
in E. coli, is an essential gene that carries out the final
step in mRNA degradation from small oligoribonucle-
otides to mononucleotides [27]. The GTPase, called
YjeQ in E. coli, is also an essential gene that binds to
ribosomes to regulate translation [28]. The GTPase has
an OB domain at the N-terminus and a zinc-finger
domain at the C-terminus which are involved in bind-
ing. PCD is known to have two roles, as an enzyme
and as a regulator. It normally functions as a homo-
tetramer, but in its regulatory role it forms heterotetra-

mers to activate partner proteins [29]. In this case, the
likely binding partner is the GTPase. The prote-
ase ⁄ chaperone, which is called HtpX in E. coli [30], is
membrane bound and helps to refold denatured pro-
teins or to degrade those that cannot be refolded. It is
a heat shock protein that is under the control of factor
sigma 32 and a histidine kinase–response regulator
pair, called CpxAR, which is only present in close rela-
tives of the enteric bacteria [31]. The protease ⁄ chaper-
one is generally activated when denatured proteins
appear in the periplasm.
From all these considerations, it appears that GHP
may be part of a heat shock response, with respect to
the particular biochemical function of H. salexigens
perhaps being related to its capacity as a sulfur carrier.
That is, denatured iron-sulfur proteins would release
sulfide, which may react with GHP and be oxidized.
PCD, RNAse and GTPase are known to be cytoplas-
mic enzymes and, furthermore, the enzymatic site of
the membrane-bound protease ⁄ chaperone is on the
cytoplasmic side. Thus, GHP is the only protein in this
group that is periplasmic. In the particular case of
Methylobium, the thiol disulfide isomerase is periplas-
mic, like GHP. The heat shock response includes both
cytoplasmic and periplasmic components. In addition
to thiol disulfide isomerases, proline cis–trans isom-
erases, which are often periplasmic, are involved in
protein renaturation and are generally included in the
heat shock response. We aim to investigate the exact
functional role of GHP as soon as more material

becomes available after cloning, expression, genetic
and physiological characterization, complemented by
structural studies using X-ray crystallography.
Experimental procedures
Bacterial culture and protein isolation
H. salexigens, strain SG 3201 (gift from P. Caumette, Uni-
versite
´
de Pau, France), was grown on the recommended
medium [6] at pH 7.4, with the addition of 100 gÆL
)1
NaCl,
0.5 gÆL
)1
sodium acetate, and 0.5 gÆL
)1
sodium thiosulfate.
Cells were grown phototrophically for 5 days in 1-liter pre-
scription bottles to stationary phase, which resulted in a
density of 2.5 g per liter culture. Some 400 g wet cell paste
was suspended in 1.5 L 0.1 m Tris ⁄ HCl, pH 8.0, and soni-
cated to fractionate the cells in 200-mL aliquots for
3 · 4 min at 4 °C. The suspension was centrifuged at
38 000 g for 20 min, and the supernatant was then centri-
fuged at 245 000 g for 3 h. The cell-free extract from H. sa-
lexigens was adsorbed on to DEAE-cellulose, and the
column developed with a stepwise salt gradient. Electron-
transfer proteins were eluted in the order: high-potential
iron-sulfur protein (80 mm NaCl), cytochrome c¢ (180 mm
NaCl), flavocytochrome c–sulfide dehydrogenase (250 mm

NaCl), GHP (300 m m NaCl), bacterial ferredoxin (poorly
resolved at 350 mm NaCl), and cytochrome c553 (500 mm
NaCl). The GHP fraction was further purified by gel filtra-
tion on Sephadex G75, by ammonium sulfate precipitation
(50–70% saturation), and by chromatography on hydroxy-
apatite, where it was eluted at 25 mm phosphate from a
0–100 mm gradient. Final purification was achieved by
FPLC on a TSK column.
GHP, a new cytochrome c from H. salexigens G. Van Driessche et al.
2808 FEBS Journal 273 (2006) 2801–2811 ª 2006 The Authors Journal compilation ª 2006 FEBS
Heme removal and protein modification
The covalently bound heme was removed by treatment of
90 lg native protein with mercuric (II) chloride in 6 m urea
containing 0.1 m HCl, as described by Ambler & Wynn [32].
After overnight incubation at 37 °C, the reaction mixture
was desalted by ultrafiltration through a Centricon-3 Protein
Concentrator (Amicon, Danvers, MA, USA) with 100 mm
ammonium bicarbonate, pH 8.0, as final buffer solution.
Enzymatic digestions
The first half of the concentrated apoprotein solution in
100 mm ammonium bicarbonate, pH 8.0, was used for
digestion with trypsin (Boehringer, Mannheim, Germany)
at an enzyme to substrate ratio (mass ⁄ mass) of 1 : 55 for
20 h at 37 °C. The second half of the concentrated apopro-
tein was digested with 2 lg Asp-N endoproteinase (Boeh-
ringer) in 100 mm ammonium bicarbonate, pH 8.0, and
incubated overnight at 37 °C. Glu-C endoproteinase (Boeh-
ringer) digestion was performed on 100 lg native protein at
an enzyme to substrate ratio of 1 : 40 in 100 mm ammo-
nium bicarbonate, pH 8.0, at 37 °C for 4 h.

Peptide purification
Peptides from the tryptic and Glu-C endoproteinase diges-
tion on apoprotein and native protein, respectively, were
separated on a Pep-S II C
2
⁄ C
18
column (4.6 · 250 mm;
Amersham Biosciences, Uppsala, Sweden) using a model 870
three-headed plunger pump, a model 8800 system controller,
and a UV monitor set at 220 nm (all parts from Dupont,
Wilmington, DE, USA). The generated peptides from the
Asp-N endoproteinase digestion on the apoprotein were
purified on a Brownlee PTC-C
18
column (2.1 · 25 mm;
PerkinElmer, Boston, MA, USA) using a SMART chroma-
tographic system (Amersham Biosciences). In all cases, the
peptides were eluted with a gradient of 0.1% trifluoroacetic
acid in water (solvent A) to 0.08% trifluoroacetic acid in
80% acetonitrile (solvent B), at an appropriate flow rate, as
recommended by the manufacturer of the column.
Sequence and molecular mass analyses
Sequence analyses were carried out on a 477A pulsed liquid
sequenator with on-line detection of the phenylthiohydanto-
in-amino acids on a 120A separation system (Applied Bio-
systems, Foster City, CA, USA).
MS analyses were initially carried out on an ESI hybrid
quadrupole-time-of-flight (Q-TOF) mass spectrometer
(Micromass, Manchester, UK), equipped with a nanoflow

Z-spray ionization system. MS ⁄ MS of peptides using CID
was performed with argon as the collision gas, at a collision
energy of 20–40 eV, depending on the mass and charge
state of the precursor ion. The MS ⁄ MS spectra were trans-
formed using the MaxEnt Sequence Software supplied with
the mass spectrometer. Before mass analysis, the HPLC
fractions were dried and redissolved in 40 lL 50% acetonit-
rile ⁄ water, containing 0.1% formic acid;  10 lL were used
for the ESI and MS ⁄ MS analyses. The instrument was cal-
ibrated with myoglobin and trypsinogen in the mass range
500–2500 m ⁄ z.
The collected peptides were also spotted on to a MALDI
sample target plate with a matrix consisting of a saturated
solution of a-cyano-4-hydroxycinnamic acid (Sigma, St
Louis, Missouri) in 50% acetonitrile ⁄ 0.1% trifluoroacetic
acid. MS and MS ⁄ MS were performed in the positive re-
flectron mode using a MALDI-TOF mass spectrometer
(model 4700 Proteomic Analyzer; Applied Biosystems).
Peptide mass spectra were obtained in the mass range 700–
4000 Da, with calibration in the default mode, and major
or interesting peaks were selected for MS determination of
their amino-acid sequences. Tandem mass spectra were
acquired by accelerating the precursor ions to 8 keV, select-
ing them with the timed gate set to a window of 6 Da, and
performing CID at 1 keV. Gas pressure (air) in the CID
cell was switched off, or at medium pressure, depending on
the obtained fragmentation or sequence information. The
mass of the native GHP was determined in the linear mode
with horse cytochrome c as internal calibrant.
Determination of free cysteines

For the determination of free cysteines, about 500 pmol
native protein was digested with 0.2 lg Glu-C endoprotein-
ase in 25 mm ammonium bicarbonate, pH 7.8, at 37 °C for
1 h. Half of the digestion mixture was then reduced with
dithiothreitol and treated with 3-bromopropylamine follow-
ing the procedure of Jue & Hale [33]. After incubation at
room temperature for 1 h, 1 lL of the reaction mixture was
mixed with an equal volume of saturated a-cyano-4-
hydroxycinnamic acid in 50% acetonitrile ⁄ 0.1% trifluoro-
acetic acid and directly analyzed using MALDI MS
techniques. The untreated half of the digestion mixture was
analyzed by MS in a comparative way.
Acknowledgements
This work was supported in part by a grant to M.A.C.
from the National Institutes of Health (GM21277),
and by a research grant from the Bijzonder Onder-
zoeksfonds of Ghent University (B ⁄ 06840 ⁄ 01).
References
1 Meyer TE (1996) Evolution and classification of c-type
cytochromes. In Cytochrome c, a Multidisciplinary
G. Van Driessche et al. GHP, a new cytochrome c from H. salexigens
FEBS Journal 273 (2006) 2801–2811 ª 2006 The Authors Journal compilation ª 2006 FEBS 2809
Approach (Scott RA & Mauk AG, eds), pp. 33–99. Uni-
versity Science Books, Sausalito, CA.
2 Van Driessche G, Vandenberghe I, Devreese B, Samyn
B, Meyer TE, Leigh R, Cusanovich MA, Bartsch RG,
Fischer U & Van Beeumen JJ (2003) Amino acid
sequences and distribution of high-potential iron-sulfur
proteins that donate electrons to the photosynthetic
reaction center in phototropic proteobacteria. J Mol

Evol 57, 181–199.
3 Leys D, Backers K, Meyer TE, Hagen WR, Cusanovich
MA & Van Beeumen JJ (2000) Crystal structures of an
oxygen-binding cytochrome c from Rhodobacter sphaer-
oides. J Biol Chem 275, 16050–16056.
4 Gaul DF, Gray GO & Knaff DB (1983) Isolation and
characterization of two soluble heme c-containing pro-
teins from Chromatium vinosum. Biochim Biophys Acta
723, 333–339.
5 Samyn B, De Smet L, Van Driessche G, Meyer TE,
Bartsch RG, Cusanovich MA & Van Beeumen JJ
(1996) A high-potential soluble cytochrome c-551 from
the purple phototrophic bacterium Chromatium vinosum
is homologous to cytochrome c
8
from denitrifying pseu-
domonads. Eur J Biochem 236, 689–696.
6 Caumette P, Baulaigue R & Matheron R (1988) Charac-
terization of Chromatium salexigens Sp-Nov, a halophi-
lic chromatiaceae isolated from mediterranean salinas.
Syst Appl Microbiol 10, 284–292.
7 Meyer TE & Cusanovich MA (2003) Discovery and
characterization of electron transfer proteins in the
photosynthetic bacteria. Photosynth Res 76, 111–126.
8 Koh M, Van Driessche G, Samyn B, Hoff WD, Meyer
TE, Cusanovich MA & Van Beeumen JJ (1996)
Sequence evidence for strong conservation of the photo-
active yellow proteins from the halophilic phototrophic
bacteria Chromatium salexigens and Rhodospirillum sale-
xigens. Biochemistry 35, 2526–2534.

9 Allen JW, Daltrop O, Stevens JM & Ferguson SJ (2003)
C-type cytochromes: diverse structures and biogenesis
systems pose evolutionary problems. Philos Trans R Soc
Lond B 358, 255–266.
10 He F, Hendrickson CL & Marshall AG (2000) Unequi-
vocal determination of metal atom oxidation state in
naked heme proteins: Fe (III) myoglobin, Fe (III) cyto-
chrome c, Fe (III) cytochrome b5, and Fe (III) cyto-
chrome b5 L47R. J Am Soc Mass Spectrom 11, 120–
126.
11 Yang F, Bogdanov B, Strittmatter EF, Vilkov AN,
Gritsenko M, Shi L, Elias DA, Ni S, Romine M, Pasa-
Tolic L, et al. (2005) Characterization of purified c-type
heme-containing peptides and identification of c-type
heme-attachment sites in Shewanella oneidenis cyto-
chromes using mass spectrometry. J Proteome Res 4,
846–854.
12 Wang Y, Vivekananda S, Men L & Zhang Q (2004)
Fragmentation of protonated ions of peptides contain-
ing cysteine, cysteine sulfinic acid, and cysteine sulfonic
acid. J Am Soc Mass Spectrom 15, 697–702.
13 Dai J, Zhang Y, Wang J, Li X, Lu Z, Cai Y & Qian X
(2005) Identification of degradation products formed
during performic oxidation of peptides and proteins by
high-performance liquid chromatography with matrix-
assisted laser desorption ⁄ ionization and tandem mass
spectrometry. Rapid Commun Mass Spectrom 19,
1130–1138.
14 Vogt W (1995) Oxidation of methionyl residues in pro-
teins: tools, targets, and reversal. Free Radic Biol Med

18, 93–105.
15 Goldberg J, Huang HB, Kwon YG, Greengard P, Nairn
AC & Kuriyan J (1995) Three-dimensional structure of
the catalytic subunit of protein serine ⁄ threonine phos-
phatase-1. Nature 376, 745–753.
16 Persson C, Sjoblom T, Groen A, Kappert K, Engstrom
U, Hellman U, Heldin CH, den Hertog J & Ostman A
(2004) Preferential oxidation of the second phosphatase
domain of receptor-like PTP-alpha revealed by an anti-
body against oxidized protein tyrosine phosphatases.
Proc Natl Acad Sci USA 101, 1886–1891.
17 Tsujimura M, Dohmae N, Odaka M, Chijimatsu M,
Takio K, Yohda M, Hoshino M, Nagashima S &
Endo I (1997) Structure of the photoreactive iron
center of the nitrile hydratase from Rhodococcus sp.
N-771. Evidence of a novel post-translational
modification in the cysteine ligand. J Biol Chem 272,
29454–29459.
18 Yamaguchi K & Hosokawa Y (1987) Cysteine dioxy-
genase. Methods Enzymol 143, 395–403.
19 Rajagopalan PT & Pei D (1998) Oxygen-mediated inac-
tivation of peptide deformylase. J Biol Chem 273,
22305–22310.
20 Wood ZA, Schroder E, Harris JR & Poole LB (2003)
Structure, mechanism and regulation of peroxiredoxins.
Trends Biochem Sci 28, 32–40.
21 Wagner E, Luche S, Penna L, Chevallet M, Van Dors-
selaer A, Leize-Wagner E & Rabilloud T (2002) A
method for detection of overoxidation of cysteines: per-
oxiredoxins are oxidized in vivo at the active-site

cysteine during oxidative stress. Biochem J 366, 777–
785.
22 Rabilloud T, Heller M, Gasnier F, Luche S, Rey C,
Aebersold R, Benahmed M, Louisot P & Lunardi J
(2002) Proteomics analysis of cellular response to oxida-
tive stress. Evidence for in vivo overoxidation of peroxi-
redoxins at their active site. J Biol Chem 277, 19396–
19401.
23 Woo HA, Chae HZ, Hwang SC, Yang KS, Kang SW,
Kim K & Rhee SG (2003) Reversing the inactivation of
peroxiredoxins caused by cysteine sulfinic acid forma-
tion. Science 300, 653–656.
24 Chevallet M, Wagner E, Luche S, Van Dorsselaer A,
Leize-Wagner E & Rabilloud T (2003) Regeneration of
GHP, a new cytochrome c from H. salexigens G. Van Driessche et al.
2810 FEBS Journal 273 (2006) 2801–2811 ª 2006 The Authors Journal compilation ª 2006 FEBS
peroxiredoxins during recovery after oxidative stress:
only some overoxidized peroxiredoxins can be reduced
during recovery after oxidative stress. J Biol Chem 278,
37146–37153.
25 Raphael AL & Gray HB (1991) Semisynthesis of axial-
ligand (position-80) mutants of cytochrome c. JAm
Chem Soc 113, 1038–1040.
26 Bamford VA, Bruno S, Rasmussen T, Appia-Ayme C,
Cheesman MR, Berks BC & Hemmings AM (2002)
Structural basis for the oxidation of thiosulfate by a
sulfur cycle enzyme. EMBO J 21, 5599–5610.
27 Ghosh S & Deutscher MP (1999) Oligoribonuclease is
an essential component of the mRNA decay pathway.
Proc Natl Acad Sci USA 96, 4372–4377.

28 Daigle D, Campbell T & Brown E (2004) Impact of
YjeQ, an essential sub-stoichiometric ribosome-asso-
ciated GTPase from Escherchia coli, on translation
in vitro. FASEB J 18, C17–C18.
29 Beckett D (2004) Functional switches in transcription
regulation: molecular mimicry and plasticity in protein–
protein interactions. Biochemistry 43, 7983–7991.
30 Kornitzer D, Teff D, Altuvia S & Oppenheim AB
(1991) Isolation, characterization, and sequence of an
Escherichia coli heat-shock gene, Htpx. J Bacteriol 173,
2944–2953.
31 Shimohata N, Chiba S, Saikawa N, Ito K & Akiyama
Y (2002) The Cpx stress response system of Escherichia
coli senses plasma membrane proteins and controls
HtpX, a membrane protease with a cytosolic active site.
Genes Cells 7, 653–662.
32 Ambler RP & Wynn M (1973) The amino acid
sequences of cytochromes c-551 from three species of
Pseudomonas. Biochem J 131, 485–498.
33 Jue RA & Hale JE (1993) Identification of cysteine resi-
dues alkylated with 3-bromopropylamine by protein
sequence analysis. Anal Biochem 210, 39–44.
G. Van Driessche et al. GHP, a new cytochrome c from H. salexigens
FEBS Journal 273 (2006) 2801–2811 ª 2006 The Authors Journal compilation ª 2006 FEBS 2811