Download 2-Oxoacid dehydrogenase multienzyme complexes

Microbiology (2000), 146, 1061–1069
Printed in Great Britain
2-Oxoacid dehydrogenase multienzyme
complexes in the halophilic Archaea ? Gene
sequences and protein structural predictions
Keith A. Jolley,1 Deborah G. Maddocks,1 Shan L. Gyles,1 Zoe$ Mullan,1
Sen-Lin Tang,2 Michael L. Dyall-Smith,2 David W. Hough1
and Michael J. Danson1
Author for correspondence : Michael J. Danson. Tel : j44 1225 826509. Fax : j44 1225 826779.
e-mail : M.J.Danson!bath.ac.uk
1
Centre for Extremophile
Research, Department of
Biology and Biochemistry,
University of Bath,
Bath BA2 7AY, UK
2
Department of
Microbiology &
Immunology, University of
Melbourne, Parkville,
Australia
All Archaea catalyse the conversion of pyruvate to acetyl-CoA via a simple
pyruvate oxidoreductase. This is in contrast to the Eukarya and most aerobic
bacteria, which use the pyruvate dehydrogenase multienzyme complex [PDHC],
consisting of multiple copies of three component enzymes : E1 (pyruvate
decarboxylase), E2 (lipoate acetyl-transferase) and E3 (dihydrolipoamide
dehydrogenase, DHLipDH). Until now no PDHC activity has been found in the
Archaea, although DHLipDH has been discovered in the extremely halophilic
Archaea and its gene sequence has been determined. In this paper, the
discovery and sequencing of an operon containing the DHLipDH gene in the
halophilic archaeon Haloferax volcanii are reported. Upstream of the DHLipDH
gene are 3 ORFs which show highest sequence identities with the E1α, E1β and
E2 genes of the PDHC from Gram-positive organisms. Structural predictions of
the proposed protein product of the E2 gene show a domain structure
characteristic of the E2 component in PDHCs, and catalytically important
residues, including the lysine to which the lipoic acid cofactor is covalently
bound, are conserved. Northern analyses indicate the transcription of the
whole operon, but no PDHC enzymic activity could be detected in cell extracts.
The presence in the E2 gene of an insertion (equivalent to approximately 100
aa) not found in bacterial or eukaryal E2 proteins, might be predicted to
prevent multienzyme complex assembly. This is the first detailed report of the
genes for a putative 2-oxoacid dehydrogenase complex in the Archaea, and the
evolutionary and metabolic consequences of these findings are discussed.
Keywords : halophile, central metabolism, pyruvate dehydrogenase complex,
dihydrolipoamide dehydrogenase
INTRODUCTION
plexes (BCODHC)] are multienzyme systems catalysing
the general reaction :
The 2-oxoacid dehydrogenase complexes [ODHCs : the
pyruvate dehydrogenase complex (PDHC), the 2oxoglutarate dehydrogenase complex (OGDHC) and
the branched-chain 2-oxoacid dehydrogenase com-
2-oxoacidjCoAjNAD+
.................................................................................................................................................
Abbreviations : BCODHC, branched-chain 2-oxoacid dehydrogenase complex ; DHLipDH, dihydrolipoamide dehydrogenase ; ODHC, 2-oxoacid dehydrogenase complex ; OGDHC, 2-oxoglutarate dehydrogenase complex ;
PDHC, pyruvate dehydrogenase complex ; TPP, thiamine pyrophosphate.
The GenBank accession number for the sequence reported in this paper is
AF068743.
acyl-CoAjCO j
#
NADHjH+
They are all three component systems consisting of
multiple copies of enzymes E1 (2-oxoacid decarboxylase), E2 (lipoate acyl-transferase) and E3 (dihydrolipoamide dehydrogenase, DHLipDH). Whilst E1 and
E2 are specific to the individual complexes, the E3
component is the same gene product in each complex
and catalyses an identical reaction (Perham, 1991, 1996).
Central to the catalytic mechanism of these complexes
0002-3889 # 2000 SGM
1061
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 06:07:27
K. A. J O L L E Y a n d O T H E R S
(a)
O
R–C–COOH
TPP–H
E1
CO2
R–CH–TPP
OH
CoASH O
R–C~SCoA
O
R–C~S–Lip–SH
SH
Lip SH
E2
S
Lip S
E3
FAD
–SH S– –
+
HB–
FAD
–S–S–
NADH + H+
NAD
+
Net reaction:
O
R–C–COOH + CoASH + NAD+
(b)
O
R–C–COOH
O
+
R–C–SCoA + CO2 + NADH + H
TPP–H
O
R–C~SCoA
CoASH
CO2
R–CH–TPP
OH
FeS
ox
Fd
red
R–C–TPP˙
O
FeS
red
Fd
ox
FeS
ox
Fd
red
FeS
red
Fd
ox
.................................................................................................................................................
Fig. 1. The proposed enzymic mechanisms for the oxidative decarboxylation of 2-oxoacids. (a) The 2-oxoacid
dehydrogenase multienzyme complexes of the Bacteria and
Eukarya. E1, 2-oxoacid decarboxylase ; E2, dihydrolipoyl
acyltransferase ; E3, DHLipDH. (b) 2-Oxoacid : ferredoxin oxidoreductase of the halophilic Archaea. [It should be noted
that in the methanogenic archaeon, Methanosarcina barkerii,
the iron–sulphur centres of both the enzyme and the ferredoxin
are only reduced in the presence of both substrates, pyruvate
and Coenzyme-A (Bock et al., 1997), and that a similar
mechanism may occur in the halophilic enzyme.] Symbols : B, a
histidine base on DHLipDH ; Fd, ferredoxin ; FeS, an enzymebound iron–sulphur cluster ; Lip, enzyme-bound lipoic acid ; TPPH, thiamine pyrophosphate.
(Fig. 1a) is the acyl-carrying cofactor, lipoic acid, which
is covalently attached to E2 and serves to connect the
three active sites and channel substrate through the
complex. In addition to forming the catalytic core of
PDHC, the E2 component also comprises the structural
core, to which the E1 and E3 components are noncovalently bound.
These complexes are found in the Eukarya and most
aerobic bacteria, whereas the Archaea possess simpler 2oxoacid oxidoreductases [reviewed by Danson (1993)
and references therein]. In the halophilic Archaea, the
pyruvate and 2-oxoglutarate oxidoreductases have an
α β structure (Plaga et al., 1992) and the catalytic
# #
mechanism
does not involve the participation of a lipoic
acid moiety (Fig. 1b). Instead, the hydroxyethyl group is
transferred directly from thiamine pyrophosphate (TPP)
to Coenzyme-A and the reducing equivalents are passed
via an iron–sulphur centre to ferredoxin (Kerscher et al.,
1981).
No 2-oxoacid dehydrogenase activity has ever been
found in the Archaea. It was therefore a surprise to
discover the presence of DHLipDH and its substrate,
lipoic acid, in the halophilic Archaea (Danson et al.,
1984 ; Pratt et al., 1989). The only known function of
DHLipDH is as the third enzyme component of the
ODHCs along with the glycine cleavage system, the
activity of which has also not been reported in the
Archaea (reviewed by Danson, 1988, 1993). The gene
encoding DHLipDH from Haloferax volcanii has been
cloned and sequenced (Vettakkorumakankav &
Stevenson, 1992), and from sequence alignments it is
clearly related to the DHLipDHs from bacterial and
eukaryal ODHCs. However, consistent with these
complex activities not being present in this organism, a
DHLipDH-negative mutant of H. volcanii was
unaffected in its ability to grow on a variety of carbon
sources that would require metabolism of pyruvate, 2oxoglutarate and glycine (Jolley et al., 1996). Neither
wild-type nor mutant strains grew on the branchedchain amino acids.
The function of DHLipDH in the halophilic Archaea is
therefore still a mystery. However, in this paper we
report the sequencing of the region upstream of the
DHLipDH gene in H. volcanii and find that the gene
appears to be the fourth ORF in an operon containing
other genes whose putative protein products are homologous to the E1α, E1β and E2 components of the
PDHC from Gram-positive bacteria. The domain structure of these proteins is predicted and the conservation
of functional amino acids is identified. Finally, we show
that the operon is transcribed and discuss possible
reasons why this does not lead to a detectable 2-oxoacid
dehydrogenase activity.
METHODS
Materials. Restriction endonucleases, Vent DNA polymerase
and bovine serum albumin were obtained from New England
Biolabs. [γ$#-P]dCTP was supplied by Amersham. The λ DNA
preparation kit was from Qiagen, DNA primers were
synthesized by PE-Applied Biosystems and the Random
Primed DNA Labelling Kit was from Boehringer Mannheim.
Strains and plasmids. H. volcanii (WFD 11) was grown at
37 mC in 18 % salt-water modified growth medium (SWMGM), consisting of 14n4 % NaCl, 1n8 % MgCl ;6H O,
#
#
2n1 % MgSO ;7H O, 0n42 % KCl, 0n5 % peptone and 0n1 %
%
#
yeast extract, pH 7n2 (Nuttall & Dyall-Smith, 1993).
Escherichia coli XL-1 Blue MRA(P2) and the λEMBL3
replacement vector were obtained from Stratagene. Plasmid
pNAT82, which contains the cloned DHLipDH gene from
H. volcanii, was a kind gift of Professor K. J. Stevenson
(University of Calgary, Alberta, Canada).
Construction of a genomic DNA library. H. volcanii genomic
DNA was prepared as described by Jolley et al. (1996) and a
genomic DNA library was then constructed in λEMBL3.
Partially digested chromosomal DNA (200 µg), after Sau3A1
treatment at 37 mC for 30 min, was fractionated by density
gradient centrifugation using 5–25 % NaCl and inserts of
approximately 15 kb were pooled and ligated to 1 µg BamHI
arms of λEMBL3. The mixture was packaged according to the
manufacturer’s instructions and E. coli XL-1 Blue MRA(P2)
was infected with the packaging mix.
1062
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 06:07:27
Archaeal 2-oxoacid dehydrogenase complexes
DNA sequencing. Direct sequencing of pNAT82 and selected
λ DNA clones was performed on a Perkin Elmer ABI Prism 377
DNA Sequencer using gene-specific primers.
Sequence alignments and secondary structural predictions.
Sequence alignments were performed using the program
- at the European Bioinformatics Server (http:\\
www2.ebi.ac.uk\blast2). Secondary structural predictions
were made using the Predict-Protein Server at EMBL (http:\\
www.embl-heidelberg.de\predictprotein), using the program
PHDsec (Rost, 1996).
Total RNA isolation and Northern blotting. H. volcanii was
cultured aerobically at 37 mC in 18 % SW-MGM, and anaerobically under N in the same medium but including 10 mM
#
NaNO . At mid-exponential
phase (A l 0n5–0n8), a 1 ml
$
&&!
sample from each culture was quickly removed
to a pre-chilled
1n5 ml microcentrifuge tube and placed on ice. The cells were
pelleted by centrifugation (11 000 g, 1 min, 4 mC), the supernatant carefully removed using a micropipette and the cells
resuspended and lysed in 80 µl lysis buffer (25 mM NaOH,
0n5 % SDS, 5 mM EDTA, 8 % sucrose, 5 µM 1,2-cyclohexanediamine N,N,Nh,Nh-tetraacetic acid). After incubation
at 37 mC for 15 min, the tubes were placed on ice for 2 min and
then 30 µl pre-cooled sodium acetate (3 M, pH 5n6) was added
and the solution vortexed for a few seconds. The lysate was
centrifuged (11 000 g, 30 min, 4 mC) to remove precipitated
proteins and the supernatant removed to a fresh tube. The
RNA was precipitated from the supernatant by adding 2 vols
ethanol and the subsequent RNA pellet was washed twice
with 70 % ethanol before being air-dried (30 min, 25 mC,
under a slight vacuum) and finally dissolved in diethylpyrocarbonate-treated water.
Agarose gel electrophoresis was performed using a modified
version of the method described by Goda & Minton (1995).
RNA was denatured by adding one-fifth volume of 60 mM
sodium phosphate buffer, pH 6n8, containing 0n25 % bromophenol blue, 0n25 % xylene cyanol, 30 % glycerol, 1n2 %
SDS, and heating at 75 mC for 5 min. Samples were immediately loaded into the wells of a 10 cm long 1 % agarose gel
and electrophoresed until the bromophenol blue dye had
reached within 2 cm of the bottom. The gel was prepared by
adding guanidine thiocyanate (final concentration 20 mM)
and 0n6 µl ethidium bromide (10 mg ml−") to 35 ml molten
(cooled to 60 mC) agarose immediately prior to pouring
into a casting tray pre-soaked for 2 h in 1 % SDS.
After electrophoresis, the RNA was transferred to a nylon
membrane by capillary-elution. The membrane was then
prehybridized (30 min, 65 mC) in hybridization solution
(0n25M Na HPO , pH 7n2, 7 % SDS) before the radiolabelled
DNA probe# (the %4n3 kbp insert from pNAT82 ; see Fig. 2) was
added and hybridization allowed to proceed at 68 mC overnight. The membrane was then rinsed in 0n2iSSC\0n1 % SDS
at room temperature (Sambrook et al., 1989) before washing
(with agitation) at 65 mC in 0n2iSSC\0n1 % SDS. Membranes
were then briefly air-dried at room temperature, covered in
plastic wrap and exposed to X-ray film (12–24 h, k70 mC,
with an intensifier screen).
RESULTS
Sequence of plasmid pNAT82
The gene encoding H. volcanii DHLipDH has previously
been cloned as part of a 4n3 kbp genomic MboI
restriction fragment and incorporated into pBluescript
KSj to give plasmid pNAT82. The DHLipDH gene
(1425 bp) was sequenced from this plasmid and was
shown to lie at one end of the cloned fragment in
pNAT82 (Vettakkorumakankav & Stevenson, 1992).
PCR amplification, subcloning and homologous expression of this gene in the parent organism required the
use of a halophilic archaeal rRNA promoter to achieve
the production of active enzyme, suggesting that the
DHLipDH gene does not have its own promoter
immediately upstream of the coding region (Jolley et al.,
1996). Moreover, N. N. Vettakkorumakankav and K. J.
Stevenson (personal communication) noticed that the 76
upstream bases of their published sequence contained
part of an ORF that showed high similarity to the E2
component of bacterial and eukaryal ODHCs.
Therefore, we decided to sequence the whole of the
DHLipDH upstream region in pNAT82 and in doing so
found two further complete ORFs (1563 and 984 bp,
respectively), plus 207 bp at the 3h end of a third, but
incomplete ORF (Fig. 2). The complete nucleotide
sequence of pNAT82 is thus 4n25 kbp. The proximity of
the 4 ORFs in pNAT82 (Fig. 2), with no recognizable
promoters preceding ORFs 2, 3 and 4, suggest an operon
structure with the DHLipDH gene being the final
component. It was therefore necessary to return to
genomic DNA to obtain the chromosomal sequence
upstream of that cloned into pNAT82.
The complete operon sequence
A genomic DNA library of H. volcanii was constructed
in λEMBL3, as described in Methods. To probe this
library, a 600 bp region of pNAT82, corresponding to
the 3h end of ORF1 and the 5h end of ORF2, was PCRamplified and the product, after purification, was
radiolabelled by random priming with [α-$#P]dCTP. In
all, 28 positive colonies were identified in the primary
screen, two of which (clones 8 and 24) gave the strongest
hybridization signals on the secondary screens. Using
sequencing primers corresponding to that part of ORF1
identified in pNAT82, it was found that clone 24
contained this ORF, and the complete sequence of this
gene, plus the upstream region, was subsequently
obtained.
As summarized in Fig. 2, the four ORFs are all in the
same direction and are tightly spaced : ORF1 and ORF2
coding regions overlap by 4 nt, with the ATG start
codon of ORF2 sharing bases with the ORF1 TGA stop
codon ; ORF2 and ORF3 are in-frame with each other,
separated by three bases ; ORF3 and ORF4 (the
DHLipDH gene) share a single nucleotide, with the
TTA stop codon of ORF3 overlapping the ATG start
codon of the DHLipDH gene.
The start codon of ORF1 appears to be GTG. This
prediction is based on sequence alignments with homologous genes in the Bacteria and Eukarya (see below),
the lack of any nearby in-phase ATG and the presence of
a good potential Shine–Dalgarno sequence 8 bp upstream of the GTG codon (i.e. CAGGAGG, which is
complementary to the first 7 nt of the 3h end of H.
volcanii 16S rRNA). Furthermore, approximately 30–
1063
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 06:07:27
K. A. J O L L E Y a n d O T H E R S
Total sequence obtained: 5385 bp
pNAT82 (4255 bp)
.....................................................................................................
Inter-gene distance (bp)
–4
+3
–1
Direction of transcription
No. of aa
ORF 1
ORF 2
ORF 3
373
328
521
ORF 4 (DHLipDH)
475
TGA-26 bp–tttttt
ORF 1 protein sequence
M S V L
40 bp further upstream of the proposed Shine–Dalgarno
sequence are two AT-rich sequences, either of which
may act as a promoter (Palmer & Daniels, 1995).
Interestingly, these two sequences are inverted repeats of
each other. Downstream of ORF4 is a good potential
transcriptional stop signal (poly-dT tract) (Fig. 2).
The upstream regions of ORFs 2, 3 and 4 do not appear
to have any clear promoter motifs, suggesting that the
four ORFs may function as an operon. Also, there are no
obvious Shine–Dalgarno sites upstream of these ORFs
and the tight (mainly overlapping) spacing of start and
stop codons could well indicate that the translation of
these genes is coupled (Normark et al., 1983). These
features are consistent with our previous attempts to
express the subcloned DHLipDH gene (ORF4), where
expression was only successful when a halophilic rRNA
promoter was added upstream (Jolley et al., 1996).
Identification of homologues in the sequence
databases
A  search of the sequence databases showed that
the proteins encoded by the four ORFs from H. volcanii
were most similar to the components of the ODHCs
from Bacillus stearothermophilus. The five highest
identity scores, with their corresponding P-values, are
given in Table 1 and it can be seen that the pyruvate
dehydrogenase components were most consistently
identified, although the list also included the homologous enzymes from the OGDHC and BCODHC.
Significantly, the Haloferax proteins from ORFs 1, 2, 3
and 4 corresponded to the Bacillus E1α, E1β, E2 and E3
(DHLipDH) components, and the gene order is the same
in the two genomes.
Predicted structural domains and motifs of the E2
component of PDHC
The E2 component of the ODHCs of the Bacteria and
Eukarya serves as both the catalytic centre and the
structural core of these multienzyme systems. Moreover,
it has a number of sequentially arranged structural
Fig. 2. The arrangement, inter-gene
distances and proposed direction of
transcription of the four ORFs constituting
the proposed H. volcanii operon. The corresponding number of amino acids of the
protein products are given. The sequence
contained within clone pNAT82 is indicated.
The DNA sequence at the 5’ end of ORF1,
plus the upstream region, is shown to
illustrate the proposed promoter sequences
(underlined), the Shine–Dalgarno sequence
(double-underlined) and the GTG start
codon. The proposed transcriptional stop
signal is also shown.
domains associated with these various functions
(Perham, 1991, 1996). Thus, in PDHC, the E2 polypeptide consists of one or more N-terminal lipoyl
domains followed by a peripheral E1 and E3 subunitbinding domain and then the C-terminal structural core
and acetyltransferase domain. Each of these motifs is
connected by extended, flexible linker regions. The
number of lipoyl domains depends on the species of
origin.
From the sequence alignments, it is predicted that H.
volcanii ORF3 encodes an E2-like protein. If this is
indeed the case, then the boundaries of typical E2
structural domains and motifs should be identifiable
within the predicted amino acid sequence. However,
before this can be investigated, account must be taken of
the sequence alignments which show that the H. volcanii
gene contains an extra 306 bp within the middle of the
sequence that are not present in any other of the E2
sequences in the databases. This ‘ insert’ is in-frame with
the rest of the sequence and would give an extra 102 aa
(residues 155–256 in the 521 aa complete sequence) to
the polypeptide. The possible significance and origin of
this sequence is considered in the Discussion.
With respect to the protein (minus the insert) that would
be produced from H. volcanii ORF3, secondary structural predictions were made using the PHDsec program
on the EMBL Predict-Protein Server (see Methods). This
program constructs a multiple sequence alignment from
the databases (in the current case, 36 sequences were
selected by the program) and this is used as input for
secondary structural predictions by a system of neural
networks. Three structured domains are predicted,
separated by two regions of 36 and 37 aa that have mean
probabilities of loop assignment (on a scale of 0–9) of 8n1
and 7n8, respectively. This is the same arrangement of
domains as a typical E2 polypeptide of an ODHC and
therefore the domains of the halophilic protein have
tentatively been assigned the same names and functions
(Fig. 3).
Analysing known E2 sequences (e.g. from B. stearothermophilus) with the PHSsec program gives almost
1064
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 06:07:27
Archaeal 2-oxoacid dehydrogenase complexes
Table 1. Sequence identities to proteins in databases
.................................................................................................................................................................................................................................................................................................................
Identification of database homologues was performed using  . P-score, probability.
H. volcanii ORF
Database match
1 (373 aa)
2 (328 aa)
3 (419 aa)
4 (475 aa)
K44
Species
Function
Length
(aa)
Accession
no.
Bacillus stearothermophilus
Bacillus subtilis
Acholeplasma laidlawii
Mycobacterium tuberculosis
Streptomyces avermitilis
Bacillus stearothermophilus
Bacillus subtilis
Acholeplasma laidlawii
Aeropyrum pernix
Deinococcus radiodurans
Bacillus stearothermophilus
Staphylococcus aureus
Myxococcus xanthus
Aeropyrum pernix
Acholeplasma laidlawii
Bacillus stearothermophilus
Bacillus subtilis
Staphylococcus aureus
Deinococcus radiodurans
Escherichia coli
E1α PDHC
E1α PDHC
E1α PDHC
E1α PDHC
E1α BCODHC
E1β PDHC
E1β PDHC
E1β PDHC
E1β PDHC
E1β ODHC
E2 PDHC
E2 PDHC
E2 BCODHC
E2 PDHC
E2 PDHC
E3 PDHC
E3 PDHC
E3 PDHC
E3 PDHC
E3 PDHC\OGDHC
368
370
345
367
381
324
324
327
325
344
427
430
416
412
544
470
470
468
467
473
P21873
P21881
P35485
O06161
Q53592
P21874
P21882
P35488
Q9YBC3
AAF09622
P11961
Q59821
Q9X6X2
Q9YBC6
P35489
P11959
P21880
Q59822
AAF11916
P00391
102 aa insert
82
35
Lipoyl E1 & E3 binding
domain
domain
228
Catalytic & structural
core domain
.................................................................................................................................................
Fig. 3. Predicted structural domains of the E2 protein that
would be produced from H. volcanii ORF3. The domain
boundaries were deduced from secondary structural predictions
made using the PHDsec program of the EMBL Predict-Protein
Server, as described in Methods. The three domains are shown
as open boxes (with the corresponding number of amino acids
contained within), and the inter-domain segments and
proposed insert are shown as filled boxes. The position of the
conserved lysine residue (K44) that is thought to be the residue
to which a lipoic acid is attached, is indicated by an arrow.
identical secondary structural predictions. Therefore,
the domains of the halophilic protein have been compared with those E2 proteins that have been structurally
analysed by NMR and X-ray crystallography and a
number of features have been identified.
Lipoyl domains consist of approximately 80 residues arranged in two β-sheets forming a
flattened β-barrel, with a core of hydrophobic residues
(Green et al., 1995). The lipoyl domain of the E2 from H.
volcanii, extending from residues 1 to 82, is evident by
The lipoyl domain.
Identity
(%)
44
43
45
45
43
56
54
50
52
48
48
47
48
45
40
49
48
47
43
44
P-score Overlap
(aa)
k75
k73
k72
k72
k68
k94
k92
k85
k84
k84
k93
k88
k86
k79
k73
k116
k115
k108
k98
k97
347
347
343
344
329
322
322
318
322
324
390
390
395
390
390
474
470
471
465
454
the conserved DKA motif around the presumed lipoyl
lysine, K44 (Dardel et al., 1993), as well as high
homology throughout the region. The secondary structural predictions indicate the presence of two β-sheets
and there is also a large proportion of hydrophobic
residues in the core area. The structure is similar to the
B. stearothermophilus PDHC E2 in having a single
lipoyl domain.
Peripheral subunit-binding domains of E2 chains consist of
approximately 35 residues and form two parallel αhelices, separated by a short strand, a helix-like turn and
an irregular loop, with a hydrophobic core (Robien
et al., 1992 ; Kalia et al., 1993). From the sequence
alignments, the domain in H. volcanii consists of
residues 119–154 and the secondary structural predictions indicate the α-helical and β-strand regions in the
expected positions. Many of the core hydrophobic
residues are conserved, whereas others are exchanged
for different hydrophobic groups. The ‘ insert ’ referred
to above (155–256) is thus immediately adjacent to the
C-terminal side of this domain.
The peripheral E1 and E3 subunit-binding domain.
It has been suggested that an aspartate and a threonine
residue in the PDHC E2 of B. stearothermophilus form
a hydrogen bond crucial to domain stability (Kalia et al.,
1993). The aspartate is conserved throughout all known
E2 sequences, including the H. volcanii E2 (D152),
whereas the threonine is replaced by a serine in many
1065
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 06:07:27
K. A. J O L L E Y a n d O T H E R S
(a)
H. volcanii E2 (123)
(153)
B. stearothermophilus E2 (132)
(162)
(b)
Insert (207)
(227)
H. volcanii E2 (120)
(140)
.................................................................................................................................................................................................................................................................................................................
Fig. 4. Partial amino acid sequence alignments of the proposed E2 protein from H. volcanii. Positions in the primary
sequences are numbered and identities are indicated with an asterisk. (a) Alignment of aa 123–153 with the peripheral
E1 and E3 subunit-binding domain of the B. stearothermophilus E2 component of the PDHC. Residues implicated in the
structure and function of the B. stearothermophilus protein are indicated in bold type. (b) Alignment of aa 207–227 of H.
volcanii E2 (within the insert sequence) with amino acids 120–140 of the same polypeptide (outside the insert sequence).
structures. In H. volcanii E2 it appears as S142 (Fig. 4a).
Since the two residues are similar in structure, it is
predicted that the stabilizing hydrogen bond is present
in the H. volcanii structure.
With respect to the domain’s possible subunit-binding
function, most of the residues implicated in binding the
E3 dimer in the B. stearothermophilus structure are
conserved in H. volcanii (detailed in Mande et al., 1996).
Thus, S124, R126, R130 and R147 are all conserved,
whereas R127 is a similarly charged lysine residue in the
B. stearothermophilus structure. G141 is conserved and
has been implicated in the binding of E1 (Kalia et al.,
1993), although E150 is either a lysine or an arginine in
most other sequences and has been implicated in
electrostatic interactions with E1. These sequence alignments are illustrated in Fig. 4(a).
The catalytic and structural core domain. The catalytic and
structural core domain of the H. volcanii E2, from
residues 294 to 521, is highly conserved with the B.
stearothermophilus E2 protein, especially around the
proposed active site histidine (H493) and aspartate
(D497) (Hendle et al., 1995).
The inter-domain regions in
the E. coli PDHC are dominated by Ala and Pro residues,
interspersed with charged and hydrophilic amino
acids ; the conformation of one of these linkers has been
studied by NMR spectroscopy and shown to be conformationally flexible, although not a random coil (Perham,
1991, and references therein). The proposed interdomains segments of the H. volcanii protein (residues
83–118 and 257–293) are predicted not to have a defined
secondary structure (see above for loop prediction
probabilities). They possess, respectively, 10 Ala plus 16
Asp\Glu, and 11 Ala, 4 Pro plus 6 Asp\Glu residues,
and thus fit with their being flexible linker regions as
found in other E2 enzymes from non-archaeal 2-oxoacid
complexes. The proposed insert (residues 155–256),
which is situated between the peripheral subunit binding
domain and its adjacent loop region (Fig. 3), also appears
not to possess a well-defined secondary structure.
The inter-domain segments.
When the insert sequence is compared with the amino
acid sequence databases, the best matches are with E2
components of the ODHCs, even though they do not
themselves contain this extra sequence. However, the
matches are partial and of a low identity. For example,
there is partial alignment of 20 residues of the insert with
amino acids 120–139 of its own E2 sequence (within
the peripheral subunit binding domain), giving a 60 %
identity score (Fig. 4b).
The E1α and E1β components
These components of the ODHCs do not contain a
domain structure that can be used to characterize the
homologous H. volcanii proteins. However, amino acid
sequence comparisons of TPP-binding enzymes have
identified a common sequence motif of approximately
30 residues in length, beginning with the highly conserved sequence GDG and concluding with a second
highly conserved NN (Hawkins et al., 1989). Approximately 10 aa to the C-terminal side of the GDG sequence
there is usually an E or D residue, followed about 5 and
11 residues further on by a generally conserved A and P
residue, respectively. Immediately preceding the NN
sequence is a cluster of hydrophobic amino acids.
In PDHCs, the binding site for TPP is thought to reside
in the E1α and not the E1β subunits (Stepp & Reed,
1985) and it is in the E1α enzyme that the conserved
motif is found. The proposed E1α protein of H. volcanii
contains this motif, the level of identity (Fig. 5) arguing
that it has retained the functional property of TPP
binding thought to be associated with this structural
element.
Enzyme activities
In the context of our discovering genes for a putative
ODHC, it is worth reiterating that we and others have
never detected enzymic activities of PDHC, OGDHC or
the BCODHCs in H. volcanii or any other halophilic
archaeon. However, DHLipDH activity is present in all
strains tested. Assays for the ODHCs have been repeated
with the strain of H. volcanii used in this work and the
absence of enzymic activity has been reconfirmed.
Positive controls for these activities have been performed
with cell extracts of E. coli.
1066
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 06:07:27
Archaeal 2-oxoacid dehydrogenase complexes
.....................................................................................................
Fig. 5. Alignment of aa 164–196 of the
proposed E1α protein from H. volcanii with
the putative TPP-binding motif from the E1α
protein of the B. stearothermophilus PDHC.
Identities are indicated with an asterisk.
Residues of the B. stearothermophilus
sequence in bold type are those that are
most highly conserved throughout TPPbinding enzymes (Hawkins et al., 1989).
H. volcanii E1α
B. stearothermophilus E1α
Northern blots
Total RNA prepared from H. volcanii was subjected to
agarose gel electrophoresis and transferred to nylon
membranes as described in Methods. Northern analysis
using the H. volcanii insert from pNAT82 (Fig. 2)
highlighted a single RNA band at 5n2 kb from both the
aerobically and anaerobically grown cells.
DISCUSSION
The sequence data presented in this paper, the structural
predictions of the protein products of the four ORFs
discovered and the previous identification of DHLipDH
as the enzymically active product of what is now ORF4,
all strongly suggest that H. volcanii possesses an operon
encoding an ODHC. Moreover, sequence identities to
proteins in the databases support the tentative conclusion that this system is a PDHC.
Northern analysis of H. volcanii RNA strongly indicates
that the whole operon is transcribed as a single message
in both aerobically and anaerobically grown cells ; that
is, a 5n2 kb RNA species hybridizing to the cloned
insert in pNAT82 has been detected, and this is very
close to the predicted length (5n4 kb) from the combined E1α, E1β, E2 and E3 gene sequences. The known
translation of the fourth ORF (DHLipDH), the proximity of all four genes in the operon and the occurrence
of overlapping start and stop codons (between ORFs 1
and 2, and between ORFs 3 and 4) suggest translational
coupling and therefore the production of enzymes E1α,
E1β and E2.
However, neither PDHC nor any ODHC activity has
been detected in H. volcanii, nor have they been found in
any archaeon (Danson, 1988, 1993). Assuming that all
four protein products are produced, a number of
explanations are possible. First, a functional ODHC
may be produced but the substrate is not pyruvate, 2oxoglutarate nor any of the branched-chain 2-oxoacids.
Consistent with a functional role, either in the current
organism or at least in its recent evolutionary history, is
the adaptation of all four proteins to an extremely
halophilic environment. Proteins from the halophilic
Archaea have highly negatively charged surfaces to
attract water in the form of hydrated potassium ions and
thereby to remain soluble in high salt concentrations
(Danson & Hough, 1997). As expected, H. volcanii
DHLipDH has a characteristically high content of acidic
amino acids (18 %) and a molecular model of the enzyme
shows a typical halophilic protein surface (Jolley et al.,
1997). The other three components that would be
produced from the operon also show a high content of
acidic amino acids : E1α, 19 % ; E1β, 16 % ; E2, 21 % ;
compared with a mean of 13n0p1n6 % across the PDHC
components from Gram-positive and Gram-negative
bacteria. However, it should be noted that the creation
of a DHLipDH-negative strain of H. volcanii failed to
identify any metabolic defect in the organism (Jolley et
al., 1996) and so any such complex is not essential to the
organism on the growth substrates tested.
Second, ORFs 1, 2 and 3 may produce non-functional
proteins for an ODHC even though an active DHLipDH
(ORF 4) is synthesized. Consistent with this, halophilic
and other archaea possess 2-oxoacid oxidoreductases as
alternative enzymes for the required catabolic functions.
However, many essential features of the proposed E1α
and E2 proteins have been conserved (E1β structural and
functional data, apart from primary sequences, are not
available in non-archaeal organisms for comparison),
suggesting retention of function. These include the E2
lysine residue that is lipoylated in non-archaeal E2
components, and we have previously shown the presence
of lipoic acid in the halophilic Archaea (Pratt et al.,
1989), although its cellular location was not established.
An added complicating factor to the question of
functional proteins being produced from the operon is
the nature of the ‘ insert ’ thought to be present in the E2
component. It is not clear if this sequence is of
importance to the function in H. volcanii, or if it has
arisen through aberrant crossing-over at the gene level
(see Fig. 4b), resulting in a non-functional E2 protein
that either is inactive or cannot assemble into the normal
core structure that binds E1α, E1β and DHLipDH.
Clearly, our findings suggest a number of experiments to
probe further the possibilities, although none are
technically easy due to the requirement of halophilic
enzymes for high ( 2 M) concentrations of salt ; therefore they are outside the scope of the current paper.
However, it is interesting to note that a DHLipDH gene
is present in the genomes of Methanococcus janaschii
and Methanobacterium thermoautotrophicum. In contrast to Mc. janaschii, which has only the DHLipDH
gene, Mb. thermoautotrophicum also has genes that
show homology with the proposed E1α, E1β and E2
genes from H. volcanii, although the identities are low
(17, 23 and 18 %, respectively) and they are not in an
operon structure (intergenic distances of 105, 25 and
145 kbp, respectively). The proposed E1α of Mb.
thermoautotrophicum does show the TPP-binding mo1067
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 06:07:27
K. A. J O L L E Y a n d O T H E R S
tif. However, prediction of structural motifs of the E2
protein failed to demonstrate the characteristic E2
domain structure and sequence alignments show that
very few of the functional residues conserved in the
H. volcanii E2 are present in the protein from the
methanogen. Interestingly, no homologues to E1α, E1β,
E2 or DHLipDH could be identified in the genomes of
Archaeoglobus fulgidus and Pyrococcus horikoshii.
Since the completion of the work reported in this paper,
four genes have been identified within the genome of
Aeropyrum pernix that show homology to the components of PDHC. In turn, they are 40–50 % identical at
the amino acid level with the H. volcanii sequences given
here, but there are no reported enzymological studies
from this particular thermophilic archaeon. It is hoped
that our analyses will point the way to future studies
in A. pernix and a wide range of other Archaea to
investigate the origin and evolution of the ODHCs in
particular and of the pathways of central metabolism in
general. The presence of ODHC genes in the halophilic
and thermophilic Archaea adds further weight to the
idea that archaeal metabolism shares many common
features with that of the Bacteria and the Eukarya, and
that the pathways of central metabolism may have been
laid down before the divergence of the three domains
(Danson et al., 1998).
stearothermophilus pyruvate dehydrogenase multienzyme complex. J Mol Biol 229, 1037–1048.
Goda, S. K. & Minton, N. P. (1995). A simple procedure for gel
electrophoresis and Northern blotting of RNA. Nucleic Acids Res
23, 3357–3358.
Green, J. D. F., Laue, E. D., Perham, R. N., Ali, S. T. & Guest, J. R.
(1995). Three-dimensional structure of a lipoyl domain from the
dihydrolipoyl acetyltransferase component of the pyruvate dehydrogenase multienzyme complex of Escherichia coli. J Mol Biol
248, 328–343.
Hawkins, C. F., Borges, A. & Perham, R. N. (1989). A common
structural motif in thiamin pyrophosphate-binding enzymes.
FEBS Lett 255, 77–82.
Hendle, J., Mattevi, A., Westphal, A. H., Spee, J., de Kok, A.,
Teplyakov, A. & Hol, W. G. J. (1995). Crystallographic and
enzymic investigations on the role of Ser558, His610, and Asn614
in the catalytic mechanism of Azotobacter vinelandii dihydrolipoamide acetyltransferase. Biochemistry 34, 4287–4298.
Jolley, K. A., Rapaport, E., Hough, D. W., Danson, M. J., Woods,
W. G. & Dyall-Smith, M. L. (1996). Dihydrolipoamide dehydro-
genase from the halophilic archaeon Haloferax volcanii –
homologous overexpression of the cloned gene. J Bacteriol 178,
3044–3048.
Jolley, K. A., Russell, R. J. M., Hough, D. W. & Danson, M. J.
(1997). Site-directed mutagenesis and halophilicity of dihydro-
lipoamide dehydrogenase from the halophilic archaeon, Haloferax volcanii. Eur J Biochem 248, 362–369.
Kalia, Y. N., Brocklehurst, S. M., Hipps, D. S., Appella, E.,
Sakaguchi, K. & Perham, R. N. (1993). The high resolution
ACKNOWLEDGEMENTS
We thank the Biotechnological and Biological Research
Council, UK (research grants to M. J. D., D. W. H. and Garry
Taylor, and to Julian Chaudhuri, D. W. H. and M. J. D., and
Earmarked Studentships to K. A. J. and D. G. M.), the Royal
Society and NATO (travel grants to M. J. D. and D. W. H.)
and the Australian Research Council (research grant to
M. D. S.) for generous financial support. We are also indebted
to K. J. Stevenson and N. N. Vettakkorumakankav (University of Calgary, Alberta, Canada) for the supply of plasmid
pNAT82 and for helpful discussions concerning their initial
sequence data.
structure of the peripheral subunit-binding domain of dihydrolipoamide acetyltransferase from the pyruvate dehydrogenase
multienzyme complex of Bacillus stearothermophilus. J Mol Biol
230, 323–341.
Kerscher, L. & Oesterhelt, D. (1981). The catalytic mechanism of
2-oxoacid : ferredoxin oxidoreductases from Halobacterium
halobium. Eur J Biochem 116, 595–600.
Mande, S. S., Sarfaty, S., Allen, M. D., Perham, R. N. & Hol,
W. G. J. (1996). Protein-protein interactions in the pyruvate
dehydrogenase multienzyme complex : dihydrolipoamide dehydrogenase complexed with the binding domain of dihydrolipoamide acetyltransferase. Structure 4, 277–286.
Normark, S., Bergstron, S., Edlund, T., Grundstrom, T., Jaurin, B.,
Lindberg, F. P. L. & Olsson, O. (1983). Overlapping genes. Annu
REFERENCES
Bock, A.-K., Scho$ nheit, P. & Teixeira, M. (1997). The iron-sulfur
centres of the pyruvate : ferredoxin oxidoreductase from
Methanosarcina barkeri (Fusaro). FEBS Lett 414, 209–212.
Danson, M. J. (1988). Dihydrolipoamide dehydrogenase : a ‘ new’
function for an old enzyme? Biochem Soc Trans 16, 87–89.
Danson, M. J. (1993). Central metabolism of the Archaea. New
Compr Biochem 26, 1–24.
Danson, M. J. & Hough, D. W. (1997). The structural basis of
protein halophilicity. Comp Biochem Physiol 117A, 307–312.
Danson, M. J., Eisenthal, R., Hall, S., Kessell, S. R. & Williams,
D. L. (1984). Dihydrolipoamide dehydrogenase from halophilic
archaebacteria. Biochem J 218, 811–818.
Danson, M. J., Russell, R. J. M., Hough, D. W. & Taylor, G. L.
(1998). Comparative enzymology as an aid to understanding
evolution. In Thermophiles : the Keys to Molecular Evolution and
the Origin of Life ?, pp. 255–267. Edited by M. W. W. Adams
& J. Wiegel. London : Taylor & Francis.
Dardel, F., Davis, A. L., Laue, E. D. & Perham, R. N. (1993). The
three-dimensional structure of the lipoyl domain from Bacillus
Rev Genet 17, 499–525.
Nuttall, S. D. & Dyall-Smith, M. L. (1993). HF1 and HF2 : novel
bacteriophages of halophilic Archaea. Virology 197, 678–684.
Palmer, J. R. & Daniels, C. J. (1995). In vivo definition of an
archaeal promoter. J Bacteriol 177, 1844–1849.
Perham, R. N. (1991). Domains, motifs and linkers in 2-oxoacid
dehydrogenase multienzyme complexes : a paradigm in the design
of a multifunctional protein. Biochemistry 30, 8501–8512.
Perham, R. N. (1996). Interaction of protein domains in the
assembly and mechanism of 2-oxo acid dehydrogenase multienzyme complexes. In Alpha-Keto Acid Dehydrogenase Complexes, pp. 1–15. Edited by M. S. Patel, T. E. Roche & R. A.
Harris. Basel : Birkha$ user.
Plaga, W., Lottspeich, F. & Oesterhelt, D. (1992). Improved
purification, crystallization and primary structure of
pyruvate : ferredoxin oxidoreductase from Halobacterium halobium. Eur J Biochem 205, 391–397.
Pratt, K. J., Carles, C., Carne, T. J., Danson, M. J. & Stevenson,
K. J. (1989). Detection of bacterial lipoic acid : a modified gas
1068
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 06:07:27
Archaeal 2-oxoacid dehydrogenase complexes
chromatographic-mass spectrometric procedure. Biochem J 258,
749–754.
Robien, M. A., Clore, G. M., Omichinski, J. G., Perham, R. N.,
Appella, E., Sakaguchi, K. & Gronenborn, A. M. (1992). Three-
dimensional solution structure of the E3-binding domain of the
dihydrolipoamide succinyltransferase core from the 2oxoglutarate dehydrogenase multienzyme complex of Escherichia
coli. Biochemistry 31, 3463–3471.
Rost, B. (1996). PHD : predicting one-dimensional protein structure by profile based neural networks. Methods Enzymol 266,
525–539.
Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular
Cloning : a Laboratory Manual, 2nd edn. Cold Spring Harbor,
NY : Cold Spring Harbor Laboratory.
Stepp, L. R. & Reed, L. J. (1985). Active-site modification of
mammalian pyruvate dehydrogenase by pyridoxal 5h-phosphate.
Biochemistry 24, 7187–7191.
Vettakkorumakankav, N. N. & Stevenson, K. J. (1992). Dihydrolipoamide dehydrogenase from Haloferax volcanii : gene cloning,
complete primary sequence and comparison to other dihydrolipoamide dehydrogenases. Biochem Cell Biol 70, 656–663.
.................................................................................................................................................
Received 24 November 1999 ; revised 14 February 2000 ; accepted
25 February 2000.
1069
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 06:07:27