Available via license: CC BY 4.0
Content may be subject to copyright.
%E
JOURNAL
OF
BIOLOGICAL CHEMISTRY
0
1994 by
The
American Society for Biochemistry and Molecular Biology, Inc.
Val.
269,
No.
21, Issue
of May
27,
pp.
14939-14945,1994
Printed
in
U.S.A.
The Primary Structure
of
Halocyanin, an Archaeal Blue Copper
Protein, Predicts a Lipid Anchor
for
Membrane Fixation*
(Received
for
publication, January 13, 1994, and
in
revised
form,
February 24, 1994)
Stephan Mattarl, Birgit Scham, Stephen
B.
H.
KentO,
Karin Rodewaldn, Dieter Oesterheltn, and
Martin Engelhardlll
From the $Max-Planck-Institut fur molekulare Physiologie, Rheinlanddamm 201, 44139 Dortmund, Germany; the
llMax-Planck-Znstitut fur Biochemie, Am Klopferspitz 18a, 85152 Martinsried, Germany; and the PScripps Research
..
.
Institute,
La
Jolla,
California 92037
Halocyanin, a small blue copper protein, was isolated
from the haloalkaliphilic archaeon
Natronobacterium
phuruonis.
The
NH,
terminus was not accessible to
Edman degradation. About
7Wo
of the amino acid
se-
quence was determined by protein sequence analysis.
The sequence information of two peptides was used for
cloning and sequencing the halocyanin gene (hcy). The
open reading frame codes for 489 base pairs, which ac-
count for a protein with 163 amino acids and a molecular
mass of 17,223 Da. The discrepancy between this value
and the molecular mass of 15,456 1.5 Da for the copper-
free protein determined by electrospray mass spectrom-
etry can be explained by a post-translational processing
of the gene product. The NH,-terminal sequence of the
open reading frame contains a motif that is character-
istic for prokaryotic lipoproteins. Assuming a similar
processing for halocyanin, Cys at position 25 of the pri-
mary transcript would be modified by a diphytanyl
(g1ycerol)diether. Subsequently, the precursor is cleaved
by a signal peptidase II-like protease and then acety-
lated at its NH,-terminal a-amino group. These modifi-
cations would yield a protein with a calculated molecu-
lar mass of 15,456 Da.
A
comparison
of
the primary
structure of halocyanin with a number of other blue
copper proteins places it into the plastocyanin-related
group.
Small blue copper (type
I)
proteins are found in
a
variety of
organisms, like bacteria, as well as in plant and animal tissues.
The blue color of these proteins,
with
an absorption maximum
at
600 nm, and their unique magnetic and redox properties are
due to the special copper-ligand interaction (for a review, see
Sykes (1991)). The four copper ligands are generally Cys, Met,
and two His, with the exception of stellacyanin, where the Met
is replaced by Gln (Bergman
et al.,
1977). Ryden (1988) has
undertaken
a
classification
of
blue copper proteins and distin-
guishes in the class
of
type
I
copper proteins five distinct fami-
lies whose primary sequences can be fitted into the same eight-
stranded fold characteristic,
e.g.
for plastocyanin (Colman
et
al.,
1978).
To
the structurally and functionally best character-
ized group belong the plastocyanins, which are found in plants,
algae, and cyanobacteria. They participate
in
the photosyn-
*
This
research
was
supported by Grant
EN
87/9-1
from
the Deutsche
Forschungsgemeinschaft. The costs
of
publication
of
this article
were
therefore
be
hereby
marked
“aduertisement”
in
accordance with
18
defrayed
in
part
by
the payment
of
page
charges. This article
must
U.S.C.
Section
1734
solely
to indicate
this
fact.
to
the GenBankTMIEMBL Data Bank with accession number(s)
230236.
The nucleotide sequence(s) reported in this paper has been submitted
fur molekulare Physiologie, Postfach 102664,
44026
Dortmund,
Ger-
11
To
whom correspondence should
be
addressed: Max-Planck-Institut
many.
Tel.:
49-231-1206-372;
Fax:
49-231-1206-229.
thetic electron transfer chain, connecting the cytochrome
b,f
complex and photosystem
1.
Three further families, the ami-
cyanins, pseudoazurins, and azurins, were detected only in
bacteria. They are involved in electron transfer reactions, like
all
type
I
copper proteins, whose function is known.
For
ex-
ample, amicyanin serves
as
primary electron acceptor for the
methylamine dehydrogenase and passes those electrons onto
either
a
soluble cytochrome
c
or to pseudoazurin (Tobari and
Harada, 1981). Azurins are thought to transfer electrons from
cytochrome
chS1
to cytochrome oxidase in the bacterial respira-
tory chain. The functional role of phytocyanins (Ryden, 19881,
with stellacyanin and cusacyanin as members,
is
presently un-
known. A sixth group, auracyanins, has been proposed
that
might replace
a
soluble cytochrome
c554
in
photosynthetic elec-
tron
transfer of
Chloroflexus aurantiacus
(McManus
et al.,
1992).
The length
of
the peptide chain
of
small blue copper proteins
vanes between 100 and 150 amino acids. Post-translational
modifications with carbohydrates are reported for a number of
non-photosynthetic blue proteins, members of the group of phy-
tocyanins (Ryden, 19881, and auracyanins (McManus
et al.,
1992).
Two
unusual azurin-like bacterial outer membrane pro-
teins appear to be lipoproteins (Strittmatter and Hitchcock,
1986; Gotschlich and Seiff, 1987). Generally, these proteins are
soluble but differ in their subcellular location. Most of the bac-
terial representatives, like azurin and rusticyanin, are located
in the periplasmic space.
Plant
plastocyanins are loosely bound
to
the inner thylakoid membrane surface in chloroplasts.
Auracyanins, which are only released from isolated membranes
by salt washing,
also
seem
to
be localized in the periplasm
(McManus
et al.,
1992).
All small blue copper proteins discovered and studied
so
far
have been isolated from organisms belonging only to two do-
mains of life, eukaryotes and bacteria. Recently,
a
blue copper
protein, halocyanin, from
Natronobacterium pharaonis
(Scharf
and Engelhard, 1993) has been isolated and characterized.
N.
pharaonis
belongs to the third domain
of
life, the archaea.
Their natural habitats are the North African soda lakes of Wadi
Natrun. They require high salt conditions and grow in the pH
range from
8.5
to
11
(Soliman and Triiper, 1982). A blue copper
protein from
Halobacterium salinarium,
another archaeon,
has
also been described (Steiner, 1983).
Halocyanin is
a
typical blue copper protein (Scharf and
Engelhard, 1993) with a characteristic fine structure
of
the
EPR spectrum,
an
absorption maximum
at
600 nm, and
a
mo-
lecular mass around 15,000 Da. The amino acid composition
is
noticeable because of
its
high amount of the acidic amino acids
Asp and Glu. The COOH-terminal sequence shows significant
homologies to plastocyanin and other blue copper proteins, par-
ticularly by harboring three possible copper ligands, Cys, His,
and Met, which are also consistent with the circular dichroic
14939
14940
Amino Acid Sequence
of
Halocyanin
spectrum in the visible range. Halocyanin adopts mainly
@-structural elements as shown by circular dichroic spectrum of
halocyanin in the
W
region that are also found in other small
blue copper proteins. The
NH,
terminus
is
not accessible to
Edman degradation. Halocyanin
is
released from the mem-
brane by mild treatment with detergents that infers
it
to be a
peripheral membrane protein. The midpoint potential is pH-
dependent and at a pH
of
7.3
is
E,
=
183
mV (versus standard
hydrogen electrode) (Brischwein
et
al.,
1993).
The observation that blue copper proteins do occur in ar-
chaea bears phylogenetic significance. Ryden (1984) places the
origin
of
blue copper proteins in
a
Chlorobium-like bacterium
from which not only the blue copper proteins but also the blue
oxidases evolved. It is not clear
at
which stage the three do-
mains of life separated and where the root has to be placed
(Woese
et
al.,
1990; Zillig
et
al.,
1992). However, the analysis of
a
family of proteins
that
is
so
widely distributed among the
living systems might contribute
to
the open phylogenetic ques-
tions.
In this study the primary sequence of halocyanin is pre-
sented as deduced from the
DNA
sequence of the coding gene
and partial amino acid sequence determination. The experi-
mental data indicate
a
post-translational modification in which
a
lipid moiety is attached to the protein explaining the affinity
of halocyanin to membranes and hydrophobic surfaces. The
comparison of the primary sequence of halocyanin with those of
other small blue copper proteins reveals its close relationship to
the group of plastocyanins.
EXPERIMENTAL PROCEDURES
Materials-Restriction endonucleases and modifying enzymes were
obtained from Boehringer Mannheim (Mannheim, Germany). Nucleo-
tides, buffer, and
Taq
polymerase for PCR' were purchased from
Am-
ersham (Braunschweig, Germany). Digoxigenin-11-dUTP, blocking
reagent,
anti-digoxigenin-alkaline
phosphatase-Fab fragments, and
3-(2'-spiroadamantane)-4-methoxy-4-(3''-phosphory~oxy~-pheny~-1,2-
dioxetane (AhfPPD) for Southern blotting experiments were obtained
from Boehringer Mannheim. The pBluescript IIKS+ vector was from
Stratagene (La Jolla, CAI.
been described by Scharf and Engelhard (1993). 4-Vinylpyridine was
The sources of materials used for the isolation of halocyanin have
from Aldrich (Heidenheim, Germany). TPCK-treated trypsin was ob-
tained from Serva (Heidelberg, Germany). Endoproteinase Glu-C from
Staphylococcus aureus was purchased from Boehringer Mannheim.
Reversed-phase HPLC columns for peptide separation were from
Waters (Eschborn, Germany) (p-Bondapak
C,,,
3.9
x
300 mm, 10 pm).
An
Aquapore RP 300 C, column (2.1
x
100 mm,
10
pm) from Pierce
Chemical Co. was used for on-line HPLC electrospray mass spectrom-
etry.
Oligonucleotides-The oligonucleotides that were used for PCR and
for sequencing were synthesized on an automatic DNA synthesizer
(381A; Applied Biosystems).
Bacterial Strains-Bacterial strains were
as
follows: N. pharaonis
SP1/28 (hR+, sR-II+, Car"' (Scharf, 1992)); Escherichia coli XL1-Blue
(recAI, endA1, gyrA96, thi-1, hsdRl7, supE44, relAl, lac, [F'proAB,
laclq, ZAM15, TnlO(tetr)l).
Isolation
of
DNA-If not otherwise indicated, standard molecular
biological methods were used according to Sambrook et
al.
(1989). Chro-
mosomal DNAfrom
N.
pharaonis was isolated and purified according to
Vogelsang et
al.
(1983). Minipreparations of
E.
coli plasmids were per-
formed by a rapid boiling procedure (Holmes and Quigley, 1981). Plas-
mid preparation for subsequent sequencing was done with Quiagen
tip-20 columns (Diagen, Diisseldorf, Germany). DNA from agarose gels
was isolated using the biotrap electroelution system from Schleicher
&
Schuell (Dassel, Germany).
Polymerase Chain Reaction-PCR was performed using
50
PM
of each
dNTP, 10 pmol of each primer,
500
ng of genomic DNA from N. phara-
open reading frame; HPLC, high performance liquid chromatography;
The abbreviations used are: PCR, polymerase chain reaction; ORF,
TPCK,
L-1-tosylamido-2-phenylethyl
chloromethyl ketone; AMPPD,
3-(2'-spiroadamantane)-4-methoxy-4-(3"-phospho~~oxy)-pheny~-1,2-di-
oxetane.
onis
as
template, and 2 units
of
Taq
polymerase in a final volume of 100
p1 using the PCR buffer from Amersham. PCR was carried out in
a
thermal cycler (Perkin-Elmer) for 30 cycles with
1
min
of
denaturation
at
95 "C,
1
min of annealing at 60 "C, and
1
min at 72 "C for the
extension reaction. The primers were 16-fold degenerated 25-mer oli-
gonucleotides derived from partial sequence analysis of halocyanin
(Mattar, 1993) referred to
as
NPC lhin (5'-GCG/C GAC CAG TCG/C
GTG/C ACGIC GAC AAC G-3') and NPC 2rev (5°C GAT G/CAC GICGC
G/CCC GTA CAT GKCC CTG-3'). A nonradioactive-labeled DNA probe
for Southern blots was generated by PCR using the same conditions,
only with 32.5 p~ d7TP and 17.5
p~
digoxigenin-11-dUTP instead of
50
p~
dlTP.
Southern Blotting-Chromosomal DNA from N. pharaonis was re-
stricted with a number of endonucleases, and about 1.5 pg of each digest
was fractionated on 1.0% (w/v) agarose gels and transferred onto nylon
membranes by vacuum blotting. Subsequent hybridization and detec-
tion were done according to Holtke et
al.
(1992) using
a
digoxigenin-ll-
dUTP-labeled probe obtained by PCR as mentioned above and
a
chemiluminescent reaction with AMPPD. The stringency employed was
hybridization
at
42 "C in 40% formamide and washing
at
50
"C
with 0.2
x
ssc.
Cloning
of
hcy Gene-For cloning, a Sac1 digest of
N.
pharaonis
genomic DNA was separated on an agarose gel, and fragments of about
3
kilobases were recovered by electroelution, ligated into a compatible
prepared pBluescript vector, and transformed into
E.
coli XL1-Blue by
electroporation using
a
Gene Pulser apparatus (Bio-Rad). Around 900
clones containing recombinant DNA were screened by colony hybridiza-
tion according
to
Buluwela et al. (1988) with the above mentioned probe
and under the same conditions used for Southern hybridization.
DNA Sequence Analysis-The samples were sequenced using the
Taq
DyeDeoxy Terminator Cycle sequencing
kit
(Applied Biosystems), and
products were analyzed on an automated DNA sequenator (3738; Ap-
plied Biosystems). After sequencing of four clones with the two PCR
primers, further primers, based on the obtained sequence data, were
synthesized, and the sequence of 620 nucleotides was determined for
both strands with the exception of
18
nucleotides of the
5'
terminus and
105 nucleotides of the
3'
terminus.
Isolation
of
Halocyanin-Growth of
N.
pharaonis and protein puri-
fication were performed
as
described (Scharf and Engelhard, 1993).
Proteolysis
of
Halocyanin-The protein was reduced and alkylated
prior to enzymatic digestion. One mg of lyophilized halocyanin was
dissolved in 400
1.11
of
0.5
M
Tris,
pH 5.6, containing
5
M
guanidine
hydrochloride and reduced with 20 pmol
of
dithiothreitol under nitro-
gen at
50
"C
for 2 h, followed by carboxymethylation with
50
pmol of
iodoacetic acid at room temperature in the dark for 30 min. Both re-
duction and alkylation were repeated, and the alkylation was stopped
by the addition of 70 pmol of dithiothreitol. For pyridylethylation
the first reduction step was performed for 4 h with 30 pmol of
dithiothreitovmg
of
halocyanin
at
a protein concentration
of
2.5 mdml
in the above mentioned buffer. Total thiols were then alkylated with
0.5
eq of 4-vinylpyridine under nitrogen in the dark for 14 h according to
Fujita et al. (1984). After reduction and alkylation, the sample was
desalted by extensive dialysis against the buffer used for subsequent
digestion of the protein with endoproteinases. The carboxymethylated
halocyanin was digested with 2.5%
(w/w)
endoproteinase Glu-C in 900
pl
of
25 mM sodium phosphate, pH 8,
at
room temperature for 14 h. The
pyridylethylated protein was digested with 3% (w/w) TPCK-trypsin in
600
p1
of 0.1
M
Tris,
pH 8.5, 10 mM CaCI,
at
room temperature for 14 h.
Peptide Separation-Samples were lyophilized after digestion, dis-
solved in 0.1% trifluoroacetic acid, and subjected
to
reversed-phase
HPLC on a p-Bondapak C,, column with
a
flow rate of
1
mymin. For the
peptide map of Glu-C, solvent A was 0.1% trifluoroacetic acid in water,
and solvent B was 0.1% trifluoroacetic acid in MeCN.
For
the tryptic
map, solvent A was 0.01% trifluoroacetic acid in water, and solvent B
was 0.009% trifluoroacetic acid in 84% MeCN. The efiluent was moni-
tored by absorbance at 230 nm.
Automated NH2-terminal Amino Acid Sequence Analysis-Edman
degradation was performed on
a
gas phase sequenator with integrated
analysis of phenylthiohydantoin derivatives by an on-line HPLC system
(473A Applied Biosystems).
Mass
Spectrometry-For electrospray mass spectrometry, 7.5 pg of
halocyanin in
6
M
guanidine hydrochloride was applied to an Aquapore
RP 300 C, reversed-phase column using an ammonium bicarbonate
buffer system with solvent A being 10 mM NH,HCO,, pH 7.6, and sol-
vent B being 10% A
+
90% MeCN. The sample was eluted with
a
gradient from
0
to 70% B over 30 min
at
a flow rate of 150 pvmin. The
monitoring wavelengths were 600 and 230 nm. 30% of the effluent was
guided directly into a Sciex (Toronto,
Canadal API-111
electrospray mass
Amino Acid Sequence
of
Halocyanin
14941
FIG.
1.
Nucleotide sequence
of
the
hcy
gene with
5'-
and 3"flanking re-
gions and amino acid sequences from
tion.
The ORF begins at position
+I
with
peptides after endoproteinase diges-
an ATG codon and ends at position +492
with a TAA stop codon. Upstream of the
ATG two putative promoter elements, box
A
(-35
to -28) and box B
(-6
to -3), are
shown in italics and shaded. Arrows
above the nucleotide sequence indicate
the positions of synthetic oligonucleotides
used as sequencing primers. Primers
NPC lhin and NPC 2rev used for PCR
and generation of the DNA probe are also
assigned. The DNA-derived amino acid
sequence is compared with peptide se-
quences of halocyanin, marked as bars,
below the corresponding sequences. Nota-
tions
G
and
T
refer to peptides obtained
after cleavage of the protein with endo-
proteinase Glu-C (see also Fig. 2) and
trypsin (data not shown). Dotted lines in-
dicate termination of the peptide se-
quence analysis. Not clearly identified
residues are marked by
x.
-56
5'
GTTGGCACGGCTCAAAG~AATrrrArCrrTCCTCGGCACATACAGTAATA'IJCGAC
+1
ATGAAAGACATCAGTCGACGCCGCTTCGTACTCGGGACCGGTGCAACGGTCGCGGCGGCA
1MKDISRRRFVLGTGATVAAA
61
ACGCTCGCAGGCTGTAACGGCAATGGCAACGGCAACGGCAACGGTAACGGCAACGGAGAA
ZTLAGCNGNGNGNGNGNGNGE
-
NPClhin
-
4
121
CCGGACACGCCGGAAGGCCGTGCCGACCAGTTCCTGACCGACAACGACGCCCTCATGTAT
41PDTPEGRAOOFLTONOALMY
1
T13
181
GATGGCGACATCACCGACGAGACGGGTCAGGACGAGGTCGTCGTCGTGACCGGTGCCGGT
610G0IT0ETG00EVVVVTGAG
b
GI
.............
241
AACAACGGCTTCGCGTTCGACCCCGCCGCGATTCGTGTCGATGTCGGCACGACCGTCACG
81NNGFAFOPAAlRVOVGTTVT
G15
I
r5
............
301
TGGGAGTGGACCGGCGACGGTGGCGCACACAACGTCGTCTCCGAGCCCGAAAGCGACTTC
1OlWEWTGDGGAHNVVSEPESOF
"X
G7
x
I
,
C9
n
a
..........
-c
361
GAGTTCGAGAGCGACCGCGTCGATGAAGAAGGATTCACATTCGAGCAGACCTTCGACGAC
~~~EFESDRVOEEGFTFEOTFDI?
""
___(I
"
"
G3
I1
GI
0
G5
7.
c
NPCZrsv
421
GAAGGCGTTGCGCTCTACGTCTGTACGCCCCACCGTGCACAGGGCATGTACGGCGCTGTC
141EGVALYVCTPHRAOGMYGAV
11
GI2
.....
TI T2
481
A~TCGAGTAATCGCTCTCCGTACCACTCATCCCCGAGGGTCGGA~GC~CCACA~~TTT
-
161
I
V E
*
541
CGACTTTTATATTATCGCTCGCTA
spectrometer. The instrument was run in negative ion mode with a
spray needle potential of
-5000
V
and an orifice potential of
-65
V.
Data
were acquired continuously over the range 400-2400
mlz
with data
collection every
0.5
mass units and a dwell time of 2 ms
(1
scad8
s).
Stored data were plotted as total ion current
uersus
time to identify the
major components eluting from the column. Mass spectral data corre-
sponding to the major components were examined using the program
MacSpec 3.21bl (Sciex). Calculated masses were derived from the pre-
dicted amino acid sequence using the program MacProMass
(Terry
Lee,
City of Hope, Duarte, CA).
NH,-terminal Peptide Extraction-The purified halocyanin was freed
of detergents by organic fast protein liquid chromatography gel filtra-
tion (Superose 12TM prep grade, HR
16/50
column; Pharmacia LKB
Biotechnology Inc.) in H20:2-propanol:MeCN (4:3:3;
H,O
with 0.0025%
trimethylamine). The lyophilized protein was digested with 2.5% (w/w)
endoproteinase Glu-C in
50
m~
NH,HCO,, pH
8,
at room temperature
for 22 h. The digest was extracted two times with chloroform, and the
Aminex AG 50-W-X4 column
(0.5
x
2.5 cm; Bio-Rad) developed with
10%
organic phase was applied to cation-exchange chromatography
on
an
acetic acid in methanol. The unretarded fractions were analyzed by
amino acid analysis.
Amino Acid Analysis-For amino acid analysis, samples were hydro-
lyzed in
6
N
HC1 with 0.1% phenol for 24
h
at 110
"C.
The hydrolysate
(Biotronik, model
7000).
was freed of solvents and analyzed on an amino acid analyzer
RESULTS
Gene
Isolation
and
Sequencing-The identification of the
halocyanin gene
(hcy)
was carried out using PCR techniques.
-1
60
20
120
40
180
60
240
80
300
100
360
120
420
140
480
160
540
163
564
3'
PCR was performed using the genomic DNA from
N.
pharaonis
as
a
template and degenerated primers derived from the se-
quence of two peptides from halocyanin (peptide
1,
ADQ-
FLTDNDALMYDGDITDETGQ; peptide 2, AQGMYGAVIVE).
The oligonucleotide
NPC
lhin was deduced from peptide
1
and
NPC 2rev from peptide
2.
The degeneration of the primers was
reduced using
a
codon usage table derived from halobacterial
protein coding genes (Soppa, 1994). The second peptide was
identified as the COOH terminus of the protein (Scharf and
Engelhard, 1993); consequently the first peptide has to be lo-
cated toward the NH, terminus. Direct sequencing of both
strands of the obtained PCR product revealed
a
343-base pair
DNA fragment. It coded for about
80%
of the total protein
sequence estimated from molecular weight measurement of the
protein by mass spectrometry (see below).
For
the isolation
of
the
hcy
gene, a nonisotopic-labeled DNA
probe was generated by PCR using NPC lhin and NPC 2rev
as
primers and digoxigenin-11-dUTP
as
label. Subsequent South-
ern blot analysis
of
the genomic DNA digested with several
endonucleases resulted in
a
single band at
-3
kilobases of a
Sac1
digest, which hybridized with the DNA probe. The de-
tected 3-kilobase fragment was cloned, and positive clones were
identified by colony hybridization and analyzed by double
strand sequencing.
In Fig.
1
the complete nucleotide sequence of the
hcy
gene
is
14942
Amino Acid Sequence
of
Halocyanin
TABLE
I
Comparison of the putative hcy promoter elements with consensus sequences for archaeal promoters and promoters
of
several archaeal protein-coding genes
The start codon is underlined, and known transcription initiation is shown in italics and boldface type.
(I),
consensus sequence taken from
Thomm and Wich (1988); (2), consensus sequence taken from Reiter
et al.
(1988); (31, halocyanin;
(4),
N.
pharaonis
haloopsin (Lanyi
et al.,
1990);
(51,
bacterioopsin (DasSarma
et
aZ.,
1984);
(S),
bacterial opsin-related protein (Betlach
et
al.,
1984);
(71, haloopsin (Blanck and Oesterhelt, 1987).
nt. nucleotide.
Gene
Box A
Intervening sequence
Box B
(1)
CTTATGTA
...
18-22 nt
...
ATGC
A
(2)
TTTAXA
...
17-22 nt
...
(3)
hCY
(4)
PhOP
(5)
bop
(6)
brp
TTTTGATG
(7)
hop
GTTATTTA
...
22 nt
...
TTTATGTT
GTTTGATT
GTTACACA
...
21 nt
...
...
23 nt
...
...
20
nt
...
...
19
nt
...
T~~g
A
TTGCACAU
TGGAGCCU
TTGCU
TTCATU
TCCGAACACU
shown. On the 620-base pair segment, an open reading frame
(ORF) can be identified.
It
starts
at
position
+1
with an ATG
codon and ends at position 492 with
a
TAA stop codon. The GC
content of 63% for the coding region corresponds well with
average values of the
N.
pharaonis genome (Tindall et
al.,
1984;
Soliman and Triiper, 1982). The ORF codes for 163 amino acids
corresponding to
a
protein with
a
molecular mass of 17,223 Da.
The 5'-noncoding region of the hcy gene contains two puta-
tive promoter elements. These RNA polymerase binding sites
were defined by comparison with promoter structures of other
archaeal genes. They are located
at
conserved positions -35 to
-28 (box A) and -6 to
-3
(box B) and show good homologies to
the archaeal consensus sequences (Reiter et
al.,
1988; Thomm
and Wich, 1988). In Table
I
the potential promoter elements of
hcy are compared with several archaeal protein-coding genes,
including
N.
pharaonis halorhodopsin. No putative ribosomal
binding site upstream of the ORF complementary to the 3'-end
of the 16
S
rRNA of Natronobacterium magadii (Lodwick et al.,
1991j, which is closely related
to
N.
pharaonis, could conclu-
sively be identified. Furthermore, no potential terminator
structure, according to Brown et
al.
(19891, was found down-
stream of the coding region.
Peptide Isolation and Sequencing-To determine also the
protein sequence of halocyanin,
it
was purified as previously
described (Scharfand Engelhard, 1993). As already shown, the
NH, terminus was not accessible to Edman degradation (Scharf
and Engelhard, 1993). For the elucidation of the primary struc-
ture of the polypeptide chain, around
50
nmol of the reduced
and acylated protein was digested with the endoproteinases
S.
aureus Glu-C and trypsin.
15
nmol of the Glu-C (Fig. 2) and
13
nmol of the tryptic proteolysate (data not shown) were sepa-
rated on
a
reversed-phase CIS column. All major peaks were
from unblocked peptides and amenable
to
gas phase sequenc-
ing. The first unretarded peak of the HPLC trace was not of
peptidic origin
as
shown by amino acid analysis. Corresponding
amino acid sequences delineated in Fig.
1
(omitting the redun-
dant information of hydrolysis products of larger sequenced
peptides) are compared with those derived from the DNA se-
quence.
The comparison confirmed
that
the 489-base pair-long ORF
encodes halocyanin. No differences between the two sequences
could be detected. However, the peptides recovered from re-
versed-phase chromatography comprised only about 72% of the
ORF (with the NH,-terminal region missing).
Extraction
of
NH,-terminal Peptide-Due to the missing of
an NH,-terminal peptide or an HPLC fraction that was blocked
to Edman degradation, further attempts were made to resolve
the amino-terminal structure of halocyanin. Because halocya-
nin is
a
peripheral membrane protein (Scharf and Engelhard,
1993), the loss of the NH,-terminal peptide after reversed-
_"-
ol
'I
"'
I
I I
I1
0
10
20
30
40
50
t
(rnin)
FIG.
2.
€&versed-phase
HPLC
separation
of
halocyanin after
endoproteinase
Glu-C
digestion.
About
50
nmol of reduced and car-
boxymethylated protein
was
digested with endoproteinase Glu-C. The
digest was applied
to
a reversed-phase p-Bondapak
C,,
column (3.9
x
solvent B was
0.1%
trifluoroacetic acid in MeCN.
300 mm,
10
pm). Solvent A was
0.1%
trifluoroacetic acid in water, and
phase HPLC could be an effect of
its
extreme hydrophobicity.
This assumption could be confirmed by extraction of
a
Glu-C
digest of halocyanin that had been freed from detergent by
organic fast protein liquid chromatography gel filtration. After
extraction of the proteolysate with chloroform, the organic
phase was further purified by cation-exchange chromatogra-
phy. Assuming that the a-amino group of mature halocyanin is
blocked, the NH,-terminal peptide, which also lacks Lys or Arg,
should be the only one that does not bind to an anionic column
material. The amino acid composition of the fraction, which
eluted unretarded from the ion-exchange column, resembles
that of a peptide with the sequence (NG),EPDTPE. A possible
NH,-terminal Cys that precedes the (NG), motif
has
not been
identified. It might be possible that the hydrolysis product@) of
this
modified Cys escaped the detection in the amino acid anal-
ysis.
Electrospray Mass Spectrometry-A sample of halocyanin
was separated on
a
reversed-phase C, column using an ammo-
nium carbonate/acetonitrile buffer system. A stream splitter
was directly connected
to
the mass spectrometer. The major
peak contained the native protein, which could be shown by the
absorption at
600
nm, with a mass of 15,521
.c
2 Da. The same
peak also contained
a
minor fraction
with
a
molecular mass of
15,456
2
1.5
Da
(Fig.
3). The difference of 65
2
3.5 Da between
the two masses correlates quite well with one atom
of
copper
(63.6 Da)
so
that the second mass can be assigned to the copper-
free halocyanin.
Amino Acid Sequence
of
Halocyanin
14943
(-11'
tO6.
I
0
3
1600
1800
2000
2
ME
FIG.
3.
On-line reversed-phase
HPLC
electrospray mass spec-
trometry (negative ion mode)
of
halocyanin solubilized
in
6
M
guanidine hydrochloride.
The
miz
value
of
each
ion
is indicated
along with the number of charges.
Inset,
molecular mass spectrum.
DISCUSSION
Halocyanin had been isolated from the membrane fraction of
the haloalkaliphilic
N.
pharaonis
(Scharf and Engelhard,
1993), which belongs
to
the archaea (Woese
et
al.,
1990). The
observation that halocyanin
is
a
peripheral membrane protein
has now been substantiated by the sequence analysis of the
mature protein and the analysis of the coding gene
hey.
The
molecular mass calculated for the gene product
is
17,223 Da
and
is
much greater than that of the copper-free halocy~in,
which was determined to be 15,456 Da by mass spectrometry.
The COOH terminus has not been processed since both DNA
and amino acid sequences agree. On the other hand, the NH,
terminus
is
blocked (Scharf and Engelhard, 1993), indicating a
post-translational modification.
The 5'-end
of
the ORF codes
for
a 24-amino acid-long region
that exhibits all features characteristic for a signal peptide of
bacterial lipoproteins (Duffaud
et al.,
1985). In Fig. 4 the NH,-
terminal sequence of halocyanin
is
compared with signal se-
quences of five lipoproteins from various organisms. The pro-
totypical signal sequence contains a cluster of positive charges
in the NE,-terminal region, which
is
followed by hydrophobic
residues and a polar COOH-terminal processing site. This lat-
ter region comprises the positions
-3
to
+1
for which the con-
sensus sequence Leu-Ala-Gly-Cys was proposed by Hayashi
and Wu (1990). For the NH,-terminal sequence of halocyanin,
all these features are found with a positive net charge at the
NH, terminus, a hydrophobic stretch at the center, and Leu-
Ala-Gly-Cys at the processing site. Furthermore, a secondary
structure calculation (Chou and Fasman, 1978) of the halocya-
nin precursor predicts a p-turn structure immediately follow-
ing the cysteine residue. This was also noted as characteristic
for lipoprotein signal sequences (Giam
et al.,
1984).
Apparently, the gene product of
hcy
possesses attributes that
are amenable to the general rules for the structure of signal
sequences
of
bacterial l~poproteins, which are recognized by the
prokaryotic signal peptidase I1 (houye
et
at.,
1983; Pollitt
et al.,
1986). This peptidase cleaves the NH,-terminal peptide bond of
Cys after
it
is
modified by a diglyceride. In a subsequent step,
the NH, terminus
is
blocked by acylation (an exception
is
de-
scribed in the case of the cytochrome subunit of the photosyn-
thetic reaction center (Weyer
et al.,
1987)). Under the assump-
tion that
N.
pharaonis
also possesses a signal peptidase
11,
this
general scheme (Hantke and Braun, 1973; Tokunaga
et
al.,
1982) can be applied to the processing
of
the halocyanin pre-
cursor. According to this mechanism, one would expect a pro-
-
25
-20
-15
-10
-5
-1fl
+5
(1
1
MtATZLVLGAVILGSTLLAGCSSNA
(2)
MLgYTCNALFLGSLILLSGCDNSS
(3)
M&QLIVNSVATVALASLVAGCFEPo
(5)
MgAYLALI SAAVI GLAACSQEP
(4)
MtLWFSTL$L?$AAAVLLFSCVAtAGCANNO
(6)
MdEL
S$%EFVLGTGATVAAATLAGAkNGNG
FIG.
4.
Comparison
of
the signal
and
NH,-tenninal sequence
of
halocyanin with those from bacterial lipoproteins.
(I),
major
outer membrane lipoprotein,
E.
coli;
(2),
pullulanase,
KEebsieEla pneu-
moniae;
(3),
cytochrome subunit of the photosynthetic reaction center,
R.
viridis;
(4),
penicillinase,
BacilEus Eicheniformis;
(51,
H8
outer mem-
brane protein,
Neisseria gonorrhoeae;
(61,
halocyanin,
N.
pharaonis.
The sequence information was taken from Hayashi and Wu
(1990)
(14)
and Gotschlich and Seiff
11987)
(5).
The
cleavage sites are indicated
by
a
line,
and
Cys
at position
+1
is shaded. Charged amino acids
in
the
NH~-terminal region are marked by
a
+
or
-,
and the h~drophobic
residues are in
boldface.
tein carrying a lipid moiety at the NH,-terminal Cys via a
thioether linkage. Furthermore, because of the inaccessibility
of the NH, terminus
to
protein sequencing,
it
has to be modi-
fied. Calculation of the molecular mass
of
such a modified pro-
tein yields 15,470 Da if an ester bond of two C,, phytanyl
hydrocarbon chains
to
glycerin and a formylated a-amino group
is
assumed. This
is
not congruent with the value obtained by
mass spectrometry. On the other hand, one of the characteristic
properties
of
archaea
is
the occurrence of C,C,, and C,C,, core
diether lipids (reviewed by De Rosa
et
al.
(1986)), and these
have
also
been found in
N.
pharaonis
(Tindall
et at.,
1984).
Therefore,
it
seems likely that the anchor of the hydrocarbons
is
an ether and not an ester. Taking these considerations into
account and assuming an acetylation of the NH, terminus,
a
mass of 15,456 Da
is
calculated. This value
is
almost exactly
what was measured by mass spectrometry. The processing of a
halocyanin precursor would result in a mature protein of 139
amino acids with Cys as NH,-terminal residue. It would have
been covalently modified by an a-amino acetylation and a
C,,C,, diphytanyl diether lipid linked to Cys by
a
thioether.
The above mentioned modifications known for prokaryotic
lipoproteins would explain that the isolation of an NH,-termi-
nal peptide after enzymatic cleavage presumably failed due to
the high hydrophobicity of the
terminal
peptide, which
could not be recovered from reversed-phase HPLC column. The
same explanation was given by Weyer
et
al.
(1987) for the
cytochrome subunit of the photosynthetic reaction center from
Rhodopseudomonas viridis,
also a lipoprotein. Due to the ex-
pected hydrophobic character of the amino-terminal Glu-C pep-
tide,
it
should be possible to extract the peptide into organic
solvents.
An
amino acid analysis of the extraction mixture re-
vealed the amino acid composition of the expected NH,-termi-
nal peptide (C(NG1,EPDTPE). However, Cys could not be iden-
tified because
a
thioether modification would not completely be
removed during hydrolysis.
Summa~zing these results, the halocyanin precursor seems
to be processed in
a
similar manner
as
it
is
proven for bacterial
cells. If this observation can be substantiated by
e.g.
direct
analysis of the NH,-terminal peptide and/or amino acid, it
would provide an interesting new mechanistic and phyloge-
netic insight into the signal peptide processing of archaea,
which apparently can utilize strategies matching those ob-
served in bacteria.
The amino acid sequence of mature halocyanin displays close
similarities
to
other type I copper proteins. In Fig.
5
the se-
quence
of
halocyanin
(HC)
is
aligned with representative ex-
amples like the blue copper protein from cucumber peelings
14944
Amino Acid Sequence
of
Halocyanin
1
cc
PC
GGWT-
15
PA
O--Du
12
KGAEGAM
16
HCCNGNGNGNGNGNGNGEP5TPEGRADOFLTDNDhLMYDGDITDET
AC
TOPPAAOP~T~PATOAANA~GGSNVVNETPA
PDAL
43
AZ
AE
0-MQFNT
17
5
6
7
cc
""
PAGhKVYTS-
PC
---SMSEEOLLNAK
PA
---_
GAEKFKSKIN
HC
-------."
DRVDE
AC
GDTANALXWTAMLN
AZ
DDSRVIAHTKCIGS
0
G
A
A
D
S
P/
P
96
99
96
139
154
7
28
of
halocyanin
(HC),
cusacyanin
(CC;
Cucumzs sutiuus,
eucaryota), plastocyanin
(PC;
Populus
nigru,
from chloroplasts), pseudoazurin
(PA;
FIG.
5.
Alignment
of
the putative
mature halocyanin sequence
with
several small
blue
copper proteins.
Compared are
the
sequences
Alcaligenes faeculis
S-6,
procaryota), auracyanin
B-1
(AC;
C.
uuruntiucus,
procaryota), and azurin
(AZ;
Pseudomonas ueruginosu,
procaryota). The
alignment was based on the data
of
Rydh and Hunt
(1993).
@-Strands (data taken from known crysta~lo~aphic structure analysis) are
underlined
and
numbered
I-?.
Clusters of similar residues are indicated by
shaded ureas.
The
copper ligands are
framed.
(CC),
plastocyanin
(PC),
pseudoazurin
(PA),
auracyanin
(AC),
and azurin
(AZ).
The general structural elements of blue cop-
per proteins are common for all selected examples. In the re-
gion of the COOH-terminal copper ligands Met, His, and Cys,
as has already been described (Scharf and Engelhard,
19931,
one finds a high amount of homology. But also the fourth li-
gand, His,
is
located upstream in
a
cluster of similarities. There
are four further similar areas that, interestingly, are located
at
positions where @-strands have been identified by x-ray struc-
tural analysis (see Adman
(1991)).
In Fig.
5,
these regions are
u~der~~~ed
and
numbered.
The folding topology of halocyanin,
like other type
I
copper proteins, apparently adopts mainly
@-pleated sheets, which is cgngruent with
its
circular dichroism
spectrum (Scharf and Engelhard,
1993).
The NH, terminus of halocyanin
is
not common
to
other
copper proteins.
It
is
much longer comprising about
45
amino
acids with
a
very acidic stretch of
10
negative charges and only
one
Arg.
Interestingly, the NH,-terminal sequence of
Halobac-
terium halobium
ferredoxin contains also a high degree of
acidic residues
(40%),
although the similarity to chloroplast
ferredoxins in the COOH-terminal part
is
strikingly high (Hase
et
al.,
1977).
This sequence pattern found in halobacteria might
be caused by the adaptation to life in concentrated
salt
solu-
tions (Lanyi,
1974).
Upstream of this acidic region
in
halocya-
nin,
a
repeating sequence
is
found where the doublet Asn-Gly
is
recurring seven times. Since this motif
is
a direct neighbor to
the lipid-modified Cys, one might assume that this stretch pro-
vides
a
flexible anchor allowing halocyanin to interact easily
sorting of lipoproteins to either the periplasm or cytoplasm
(Bouvier
et al.,
1991).
An
Asp
at
this position directs the li-
poprotein to the cytoplasmic site, whereas the replacement of
this residue by Ser, Asn, or Glu causes these proteins to be
transported to the outer membrane. Halocyanin possesses an
Asn at this position that would indicate it to be periplasmatic.
It
would be interesting if these sorting signals found for bacte-
ria
are also preserved in archaea.
The phylogenetic relation of halocyanin has been provision-
ally analyzed. Ryd6n and Hunt
(1993)
propose five distinct
families of small blue proteins. Plastocyanin, pseudoazurin,
and amicyanin cluster together to form a
group
of plastocyanin-
related proteins. All of these proteins are of bacterial origin if
plastocyanin as chloroplast protein counted as bacterial pro-
tein.
It
is interesting to note that another example of the simi-
larity between ch~oropl~st- and archaeal-type proteins
is
ferre-
doxin (Hase
et
al.,
1977).
All calculations place halocyanin into
the plastocyanin-related superfamily with almost equal dis-
tance to pseudoazurin and plastocyanin. More data on other
archaeal proteins have to be available if conclusions about the
origin of small blue copper proteins and dates of events can be
drawn.
Halocyanin emerges as the
first
example of an archaeal
small blue copper protein whose amino acid sequence has been
determined. The protein displays structural properties that are
novel for archaea and also for small type
I
copper proteins. The
post-translational introduction of
a
lipid moiety has not previ-
ously been observed in archaea nor has it been proven directly
and without pe~urbation with
its
reaction partner. for blue copper proteins.
tion of halocyanin, that have to be addressed, but for which an
an^^^^^^^^^^^^^^^^^^^^.
~~~~~~~~i~,s~~~~~~
answer is not yet available. The physiological significance
of
ad+e the molecular genetic methods,
halocyanin is not
known.
Its
midpoint potential ofE,
=
183
mV
(uersus
standard hydrogen electrode) (Brischwein
et
al.,
1993)
REFE~NCES
would enable halocyanin
to
interact with a terminal oxidase
Adman,
E.
T.
(1991)
&,,.
protein
ckem,
42,
145-197
There are two questions, the physiological role and the loca-
adopting a function like azurin. This assumption would imply a
Bergman,
C.,
Gmdvik,
E.&
NP~,
P.
0.
Strid,
L.
(1977)
&&em.
BWw.
location
Of
halocyanin at
the
site
Of
the membrane'
Betla&,
EA.,
Friedman,
J.,
Boyer,
€$.
W.
&
Heifer,
F.
(1984)
NucZeie
Acids
Res.
12,
Res. Commrm.
77,1052-1059
The position
+2
of the signal sequence appears to be critical for
7949-7959
Amino Acid Sequence
of
Halocyanin
14945
Blanck,
A.
&
Oesterhelt, D.
(1987)
EMBO J.
6,265-273
Bouvier,
J.,
Pugsley,
A.
P.
&
Stragier,
P.
(1991)
J. Bacteriol.
173, 5523-5531
Brischwein, M., Scharf, B., Engelhard, M.
&
Mhtele, W.
(1993)
Biochemistry
32,
Brown,
J.
W., Daniels, C.
J.
&
Reeve,
J.
N.
(1989)
CRC Crit. Reu. Microbiol.
16,
Buluwela, L., Forster,
A,,
Boehm, T.
&
Rabitts, T.
H.
(1988)
Nucleic Acids Res.
17,
Chou,
P.
Y.
&
Fasman, G. D.
(1978)
Annu. Rev. Biochem.
47,251-276
Colman, P. M., Freeman, H. C., Guss,
J.
M., Murata, M., Noms, V.
A,,
Ramshaw,
DasSarma,
S.,
RajBhandary,
U.
L.
&
Khorana, H.
G.
(1984)
Proc. Natl. Acad. Sci.
De Rosa, M., Gambacorta,
A.
&
Gliozzi,
A.
(1986)
Microbiol. Rev.
60,
70-80
Duffaud, G. D., Lehnhardt,
S.
K., March,
P.
E.
&
Inouye,
M.
(1985)
Cur,: Top.
Fujita, V.
S.,
Black,
S.
D., Tan;
G.
E., Koop, D. R.
&
Coon, M.
J.
(1984)
Proc. Natl.
Giam, C.-Z., Chai, T., Hayashi,
S.
&
Wu, H. C.
(1984)
Eu,:
J.
Biochem.
141,331-337
Hantke,
K.
&
Braun, V.
(1973)
Eu,:
J.
Biochem.
34, 284-296
Gotschlich, E. C.
&
Seiff, M. E.
(1987)
FEMS Microbiol. Lett.
43, 253-255
Hase,
T.,
Wakabayashi,
S.,
Matsubara,
H.,
Kerscher, L., Oesterhelt. D., Rao,
K.
K.
Hayashi,
S.
&
Wu,
H.
C.
(1990)
J. Bioenerg. Biomembr.
22, 451-471
Holmes, D.
S.
&
Quigley, M.
(1981)AnaL
Biochem.
114, 193-197
Holtke, H. J., Sagner,
G.,
Kessler, C.
&
Schmitz,
G. (1992)
BioTechniques
12,
Inouye,
S.,
Franceschini,
T.,
Sato, M., Itakura,
K.
&
Inouye,
M.
(1983)
EMBO
J.
2,
Lanyi,
J.
K.
(1974)
Bacterid. Reu.
38, 272-290
Lanyi,
J.
K.,
Duschl,
A,,
Hatfield, G.
W.,
May,
K.
&
Oesterhelt, D.
(1990)
J.
Biol.
Lodwick,
D.,
Ross,
H.
N.
M., Walker,
J.
A.,
Almond,
J.
W.
&
Grant, W. D.
(1991)
Mattar,
S.
(1993)
Seguenzstudien
zum
Halocyanin uon Natronobacterium phara-
McManus,
J.
D., Brune, D. C., Han, J., Sanders-Loehr, J.. Meyer,
T.
E., Cusanovich,
13710-13717
287-337
452
J.
A.
M.
&
Venkatappa, M.
P.
(1978)
Nature
272,319-324
U.
S.
A.
81,
125-129
Membr. Zkansp.
24, 65-104
Acad.
Sci.
U.
S.
A.
81.42604264
&
Hall,
D.
0.
(1977)
FEES
Lett.
77,308-310
10P113
87-91
Chem.
265,1253-1260
Syst. Appl. Microbiol.
14, 352-357
onis. Dissertation, Ruhr-Universitat Bochum
Pollitt,
S.,
Inouye,
S.
&
Inouye, M.
(1986)
J.
Biol.
Chem.
261, 1835-1837
Reiter, W.-D., Palm, P.
&
Zillig, W.
(1988)
Nucleic Acids Res.
16, 1-19
Ryden, L.
(1984)
in Copper Proteins and Copper Enzymes (Lontie,
R.,
ed) Val.
1,
pp.
Ryden, L.
(1988)
in Oxidases and Related Redox Systems (King,
T.
P.
&
Manson,
H.,
Ryden,
L.
&
Hunt, L.
T.
(1993)
J.
Mol.
Evol.
36, 41-66
Sambrook,
J.,
Fritsch, E.
F.
&
Maniatis, T.
(1989)
Molecular Cloning:
A
Laboratory
Manual, 2nd Ed., Cold Spring Harbor Laboratory, Cold Spring Harbor,
NY
Scharf, B.
(1992)
Vergleichende Untersuchungen des sensorischen Rhodopsins
I1
in
Halobacterium halobium und Natronobacterium pharaonis und Charak-
terisierung des Halocyanins, des ersten blauen Kupferproteins aus Archaebak-
terien. Ph.D. thesis, Ruhr-Universitat Bochum
M.
A,,
Tollin,
G.
&
Blankenship, R. E.
(1992)
J.
Biol.
Chem.
267,653145540
157-182,
CRC Press, Inc., Boca Raton, FL
eds) pp.
349-366,
Liss, Inc., New York
Scharf,
B.
&
Engelhard, M.
(1993)
Biochemistry
32, 12894-12900
Soliman, G.
S.
H.
&
Triiper,
H.
G.
(1982)
Zentralbl. Bakteriol. Mikrobiol.
1
Abt.
Soppa,
J.
(1994)
System. Appl. Microbiol.
16, 725-733
Steiner, M.
(1983)
Isolierung und Charakterisierung des nativen Halorhodopsins
aus Halobacterium halobium.
Ph.D.
thesis,
Ludwig-Maximillian-Universitiit
Miinchen
Orig.
C
3, 318329
Strittmatter,
W.
&
Hitchcock, P.
J.
(1986)
J.
Exp. Med.
164,2038-2048
Sykes,
A.
G.
(1991)
Adu. Inorg. Chem.
36, 377-408
Thomm, M.
&
Wich, G.
(1988)
Nucleic Acids Res.
16, 151-163
Tindall, B.
J.,
Ross,
H. N. M.
&
Grant,
W.
D.
(1984)
Syst. Appl. Microbiol.
6,41-57
Tobari,
J.
&
Harada,
Y.
(1981)
Biochem. Biophys. Res. Commun.
101,502-508
Tokunaga,
M.,
Tokunaga, H.
&
Wu, H. C.
(1982)
Proc.
Natl. Acad. Sci.
U.
S.
A.
79,
Vogelsang, H., Oertel, W.
&
Oesterhelt, D.
(1983)
Methods Enzymol.
97,226241
Weyer, K.
A.,
Schafer, W., Lottspeich,
E
&
Michel, H.
(1987)
Biochemistry
26,
Woese, C. R., Kandler,
0.
&
Wheelis, M. L.
(1990)
Proc.
Natl. Acad.
Sci.
U.
S.
A.
87,
2909-2914
45764579
Zillig, W., Palm, P.
&
Klenk, H.-P.
(1992)
in The Origin and Evolution ofProkaryotic
and Eukaryotic
Cells
(Hartman, H.
&
Matsuno,
K.,
eds) pp.
163-182,
World
Scientific Publishing Co., River Edge, NJ
2255-2259
