Microbiology (2000), 146, 487–495
Printed in Great Britain
The genes for erythritol catabolism are
organized as an inducible operon in
Brucella abortus
Fe! lix J. Sangari, Jesu! s Agu$ ero and Juan M. Garcı! a-Lobo
Author for correspondence : Juan M. Garcı! a-Lobo. Tel : j34 42 201948. Fax : j34 42 201945.
e-mail : jmglobo!medi.unican.es
Departamento de Biologı! a
Molecular, Facultad de
Medicina, Universidad de
Cantabria, Unidad
Asociada al Centro de
Investigaciones Biolo! gicas,
CSIC, Cardenal Herrera
Oria s/n, 39011 Santander,
Spain
Erythritol utilization is a characteristic of pathogenic Brucella abortus strains.
The attenuated vaccine strain B19 is the only Brucella strain that is inhibited by
erythritol, so a role for erythritol metabolism in virulence is suspected. A
chromosomal fragment from the pathogenic strain B. abortus 2308 containing
genes for the utilization of erythritol was cloned taking advantage of an
erythritol-sensitive Tn5 insertion mutant. The nucleotide sequence of the
complete 7714 bp fragment was determined. Four ORFs were identified in the
sequence. The four genes were closely spaced, suggesting that they were
organized as a single operon (the ery operon). The first gene (eryA) encoded a
519 aa putative erythritol kinase. The second gene (eryB) encoded an erythritol
phosphate dehydrogenase. The function of the third gene (eryC) product was
tentatively assigned as D- erythrulose-1-phosphate dehydrogenase and the
fourth gene (eryD) encoded a regulator of ery operon expression. The operon
promoter was located 5' to eryA, and contained an IHF (integration host factor)
binding site. Transcription from this promoter was repressed by EryD, and
stimulated by erythritol. Functional IHF was required for expression of the
operon in Escherichia coli, suggesting a role for IHF in its regulation in B.
abortus. The results obtained will be helpful in clarifying the role of erythritol
metabolism in the virulence of Brucella spp.
Keywords : erythritol operon, Brucella abortus
INTRODUCTION
Brucella abortus is a Gram-negative pathogenic bacterium which causes bovine brucellosis. B. abortus has a
tropism for the reproductive organs, colonizing the
placenta and fetus of pregnant cows and thus causing
abortions. The preferential use of erythritol is characteristic of the genus Brucella. Erythritol is used by
Brucella in preference to glucose (Anderson & Smith,
1965) and erythritol promotes the growth of some
Brucella strains (Meyer, 1967). The metabolic pathway
for degradation of erythritol in Brucella was described
by Sperry & Robertson (1975a). B. abortus strain B19 is
a spontaneous attenuated mutant widely used to vac.................................................................................................................................................
Abbreviation : IHF, integration host factor.
The GenBank accession number for the sequence reported in this paper is
U57100.
cinate cattle. Erythritol inhibits, rather than promoting,
the growth of B19. Sperry & Robertson (1975b) reported
the absence in strain B19 of -erythrulose-1-phosphate
dehydrogenase, a critical enzyme in the erythritol
catabolic pathway. This defect causes both the accumulation of the toxic intermediate -erythrulose 1phosphate and a depletion of ATP levels, leading to
inhibition of bacterial growth. Erythritol metabolism is
the best known and most important difference between
B19 and virulent B. abortus strains. Consequently, a
correlation has been proposed between the capacity of
B. abortus to metabolize erythritol and the virulence of
this species. Erythritol has been found in the bovine
placenta, and a chemotactic role has been attributed to
this polyalcohol that would explain the specific
localization of B. abortus in this organ (Smith et al.,
1962). Several reports have claimed a correlation between erythritol oxidation and virulence in different
animal models (Meyer, 1966).
0002-3571 # 1999 SGM
487
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 14:53:47
F. J. S A N G A RI, J. A G U$ E R O a n d J. M. G A R C I! A - L O B O
Table 1. Bacterial strains and plasmids
Strain/plasmid
E. coli
DH5α
S17.1
RU4404
MC294
MC296
B. abortus
2308
2308Nx
227
B19
B19Nx
FJS-2
Plasmids
pBluescript II SK(j)
pKT231
pUC4K
pKOK.4
pSU6004
pSU6010
pSU6104
pSU6099
pSU6100
pSU6111
Relevant characteristics
φ80dlacZ ∆M15 ∆(argF–lacZYA)U169
deoR endA1 gyrA96 hsdR17 recA1
relA1 supE44 thi-1
thi pro his recA RP4–2-Tc : : Mu
Km : : Tn7
MM294 [thi-1 endA1 hsdR17
supE44] : : Tn1725
ara argEam ∆(lac–pro) gyrA metB
∆(recA) rpoB srl : : Tn10 supF thi
MC294 himA∆82 : : Tn10 hip∆3 : cam
Virulent strain
Spontaneous nalidixic acid resistant
mutant
2308Nx [eryB : : Tn5]
Vaccine strain
Spontaneous nalidixic acid resistant
mutant
2308Nx [eryD : : Km]
ApR
KmR SmR
KmR
ApR CmR TcR
7n7 kb EcoRI ery fragment from 2308 in
pBluescript II SK(j)
Same insert cloned in pKT231
eryA in pBluescript II SK(j)
eryAB in pBluescript II SK(j)
eryABC in pBluescript II SK(j)
eryABCD in pBluescript II SK(j)
The role of erythritol metabolism as a virulence factor
remains obscure. This prompted us to undertake a
genetic characterization of the metabolism of erythritol
and its contribution to virulence in B. abortus. As a first
step we generated a library of transposon insertion
mutants of B. abortus strain 2308 as previously described
(Sangari & Agu$ ero, 1991). A mutant was isolated
(mutant 227) that did not metabolize erythritol and was
inhibited by this compound, mimicking the behaviour of
the vaccine strain B19. This insertion mutant was used
to clone the ery : : Tn5 region from its chromosome, and
then the corresponding regions from both strains 2308
and B19. By comparison of these regions a deletion was
found in the chromosome of the B19 strain, allowing the
development of a PCR assay for the rapid and unequivocal differentiation of the B19 vaccine strain from
other Brucella strains (Sangari & Agu$ ero, 1991 ; Sangari
et al., 1994). The transposon insertion mutants in the ery
genes were used to demonstrate that oxidation of
erythritol was not essential for virulence in a mouse
Reference
Life Technologies
Simon et al. (1983)
Ubben & Schmitt (1986)
Gamas et al. (1986)
Gamas et al. (1986)
Sangari & Agu$ ero (1991)
Sangari & Agu$ ero (1994)
Sangari et al. (1994)
This work
Stratagene
Bagdasarian et al. (1981)
Pharmacia
Kokotek & Lotz (1991)
Sangari & Agu$ ero (1994)
Sangari et al. (1994)
This work
This work
This work
This work
model of B. abortus infection (Sangari et al., 1998). The
aim of the work described here was to characterize the
complete ery region from B. abortus 2308 and gain some
insight into its regulation, thus enabling further analysis
of the association between erythritol catabolism and
virulence in Brucella.
METHODS
Bacterial strains, plasmids, and growth conditions. Bacterial
strains and plasmids used in this work are listed in Table 1. B.
abortus was grown on tryptose broth (TB) or tryptose agar
(TA) (Difco). Plates were usually incubated at 37 mC, under a
5 % CO atmosphere. Erythritol (Sigma) was added to the
medium #at 1 % as a nutritional supplement or at 1 mg ml−" for
induction of the ery operon. Escherichia coli strains were
usually grown on Luria–Bertani broth or plates. Concentrations of antibiotics in the agar plates were as follows :
ampicillin, 50 µg ml−" ; chloramphenicol, 25 µg ml−" ; kanamycin, 50 µg ml−" ; nalidixic acid, 25 µg ml−".
Genetic and DNA methods. E. coli transformation was
488
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 14:53:47
Erythritol utilization operon of Brucella abortus
Table 2. Synthetic oligonucleotides
Name
M2990
eryA.1
eryA.2
Tn1725.I
Tn1725.II
Sequence
CAGGTCGCTCGCAGCCCCTTCC
CGCATCACCGGTGGTGCT
CTTCCCGATAGGCAACGA
GAGCTGTCACGAGAACACCG
CCCTTACGGATGCCCGGAAA
performed either by the calcium chloride method (Sambrook
et al., 1989), or by electroporation (Dower et al., 1988) using
a Bio-Rad Gene-Pulser apparatus. Conjugation between E.
coli S17.1 containing mobilizable plasmids and B. abortus was
performed in solid medium on nitrocellulose filters for 6 h
(Sangari & Agu$ ero, 1991). Transconjugants were obtained on
selective plates. Restriction endonucleases, ligase and other
DNA enzymes were used according to the manufacturers’
directions. DNA sequencing was performed by the dideoxy
chain-termination method (Sanger et al., 1977) with plasmid
templates using a Taq DNA polymerase based kit (Promega).
Transposon mutagenesis. Plasmid pSU6010 (Table 1), con-
sisting of the complete ery region from B. abortus 2308 in a
7714 bp EcoRI fragment cloned in the broad-host-range,
mobilizable vector pKT231 (Bagdasarian et al., 1981), was
able to complement the erythritol susceptibility of both B.
abortus strain B19 and the ery : : Tn5 mutant 227 (Sangari et
al., 1994). This plasmid was subjected to mutagenesis with the
chloramphenicol-resistance transposon Tn1725 to map the
position of ery genes precisely. To do this, pSU6010 was
introduced into E. coli strain RU4404 (Ubben & Schmitt,
1986), carrying Tn1725 in the chromosome. After 2 d growth
at 30 mC, plasmid DNA was isolated and used to transform E.
coli DH5α to chloramphenicol resistance. Plasmid DNA was
purified from individual transformants and the location of the
transposon was mapped with EcoRI and HindIII.
Construction of a B. abortus eryD mutant. To construct a B.
abortus eryD mutant by allelic replacement, we cloned a 2 kb
NruI–HindIII fragment containing eryD from the chromosome of strain 2308 (Sangari et al., 1994) into pBluescript SK.
This plasmid was called pSU6101 (Fig. 1). The central 0n5 kb
BalI fragment from the eryD insert in pSU6101 was then
replaced with a 1n2 kb kanamycin-resistance cassette from
pUC4K. The resulting plasmid, pSU6103, contained a deleted
eryD allele interrupted by the kanamycin-resistance cassette.
The insert in pSU6103 was cloned into the plasmid pKOK.4,
which can be transferred by conjugation into B. abortus, but
which does not replicate in this species. In this way the
plasmid pSU6106 was obtained and transferred into B. abortus
2308Nx. Kanamycin-resistant transconjugants were analysed
by Southern blot hybridization with an eryD probe to check
for replacement of the locus. One colony with the correct
genomic structure (strain FJS-2) was selected for further work.
The construction procedure is shown diagrammatically in
Fig. 1.
Phenotypic expression of ery genes in E. coli. E. coli cells
containing ery recombinant plasmids were plated on
Use
eryA mRNA primer extension
Sequencing eryA
Sequencing eryA
Sequencing Tn1725 insertions
Sequencing Tn1725 insertions
MacConkey agar base supplemented with 1 % erythritol.
Colonies able to utilize erythritol were red on these plates.
Primer extension mapping of the 5h end of ery RNA. Total
RNA was prepared from 10 ml cultures as described previously (Garrido et al., 1993). For primer extension studies,
RNA was used as template for the synthesis of cDNA by AMV
reverse transcriptase (Boehringer Mannheim). Synthetic oligonucleotide primer M2990 (Table 2) was 5h end-labelled with
polynucleotide kinase and [γ-$#P]ATP. The products of the
extension reactions were analysed in 6 % urea-polyacrylamide
gels. A sequencing ladder using the same primer was run in the
gel beside the extension products to map the 5h end of the
mRNA.
Sequence analysis and comparison. The DNA sequence was
analysed with the GCG programs (Devereux et al., 1984) in
the EMBNET node at the Centro Nacional de Biotecnologı! a,
Madrid. Sequence comparisons against databases were done
with the  programs (Altschul et al., 1997) at the NCBI
 server. Analysis of DNA curvature was performed with
the program .
RESULTS
Tn1725 mutagenesis of the ery region
A total of 80 independent Tn1725 insertions in pSU6010
were analysed. The location of the transposon in each
individual mutant was determined by mapping with
EcoRI and HindIII (Fig. 2). In 38 of them (pSU6022 to
pSU6059) the transposon was found integrated into the
Brucella DNA insert. In order to evaluate the effect of
Tn1725 insertion on the Ery phenotype, each mutant
was introduced by conjugation into B. abortus B19 and
the transconjugants were checked for their ability to
grow on erythritol-containing plates. The results of this
complementation experiment are also summarized in
Fig. 2.
Nucleotide sequence of the ery region
The nucleotide sequence of the complete EcoRI insert in
pSU6010 was determined on both DNA strands. Three
approaches were used for sequencing : (i) sequencing of
Tn1725 insertions using specific primers at the trans489
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 14:53:47
F. J. S A N G A RI, J. A G U$ E R O a n d J. M. G A R C I! A - L O B O
.....................................................................................................
Fig. 1. Schematic diagram of the construction of an eryD mutant by allelic replacement in strain 2308Nx. A NruI–HindIII
fragment containing eryD was cloned in
pBluescript SK (pSU6101), and an internal
BalI fragment was replaced with a kanamycinresistance cassette, giving pSU6103. The
interrupted region was transferred to the
suicide plasmid pKOK.4 (pSU6106), and used
to produce the eryD mutant strain FJS-2 by
homologous recombination.
.................................................................................................................................................................................................................................................................................................................
Fig. 2. Physical map of the EcoRI fragment containing the ery region from the chromosome of B. abortus 2308 cloned in
pSU6004. The arrows under the map point to the positions of different Tn1725 insertions in that plasmid. These
insertions were named pSU6022 to pSU6059, and they are identified in the figure by the last two digits of their plasmid
numbers. A plus sign below the arrow indicates that the plasmid complemented the Ery− phenotype of strain B19 ; a
minus sign indicates lack of complementation. The position of the ery ORFs as well as the ery (P1) promoter and
terminator is shown in the lower part of the figure.
poson ends, (ii) subcloning of specific restriction fragments that were sequenced with universal primers, and
(iii) use of synthetic oligonucleotides as primers to
sequence the remaining gaps (Table 2). The main
features of the sequence are described below and shown
in Fig. 2.
The region between nucleotides 1 and 376 showed an
AjT content much higher than the 41 mol % mean of
the Brucella genome. A sequence identical to the
integration host factor (IHF) binding site (Craig &
Nash, 1984) was found in this region (116–129). The
sequence TTTACAN TGTTAT (P1) was similar
#&#
#)"
to the canonical
E. coli ")promoter. Furthermore,
some
stretches of alternating As and Ts were suggestive of the
existence of intrinsic curvature by analysis with the
program  (data not shown). All these data
strongly suggested that this region could be involved in
the transcriptional start and regulation of expression of
downstream genes.
490
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 14:53:47
Erythritol utilization operon of Brucella abortus
Table 3. Sequence homologies between Ery proteins and gene products in the databases
ORF
Size
(bp)
Closest homologous gene/gene product (accession no.)
EryA
EryB
EryC
EryD
519
502
309
316
E. coli xylB\xylulose kinase (X04691)
E. coli glpD\glycerol-3-phosphate dehydrogenase (M55989)
Alcaligenes hydrogenophilus hupL\hydrogenase (S56898)
Rhodobacter sphaeroides smoC\operon regulator (AF018073)
Klebsiella pneumoniae dalR\operon regulator (AF04525)
aa identity
(%)/similarity
(%)*
30n5\39n0
50n4\68n5
38n9\51n8†
36n0\48n7
30n7\42n0
* Percentage identity and similarity were determined using the GAP program from the Genetics
Computer Group, with gap weight and gap length values of 8 and 2 respectively.
† Comparison limited to a 57 aa region containing the best fit between the two proteins.
Four ORFs were identified in the region spanning
basepairs 377 to 5400. The eryA gene could start either
at the GTG codon at bp 377 or at the ATG at basepair
386. The presence of the sequence GAAAG, similar to
the E. coli ribosome-binding site, 3 bp upstream of the
GTG codon suggested that this was the initiation codon
of eryA. The stop codon for this gene was found at
position 1934. Translation of this sequence would result
in a 519 aa polypeptide with a calculated molecular
mass of 54n4 kDa. The next gene in the sequence (eryB)
would start at the ATG at basepair 1949 and its stop
codon was found at basepair 3455. This gene would
translate into a 502 aa polypeptide with a calculated
molecular mass of 56n2 kDa. The eryC gene started at
3465 and ended at 4392. Its product would be 309 aa
long with a molecular mass of 35 kDa. Finally, eryD was
found between positions 4422 and 5370 ; its translated
product would be a protein of 316 aa with a molecular
mass of 33n5 kDa. Genes eryA, eryB and eryC were
closely spaced with very short intergenic regions in
which sequences resembling the E. coli ribosomebinding site were found, and there were only 30 bp
between eryC and eryD. A sequence found 3h to eryD
between basepairs 5372 and 5412 was able to fold as a
stem–loop structure and had some analogies with
transcription terminators. These data suggested that the
four genes eryABCD could constitute a single transcriptional unit. Three additional genes, encoding a triose
phosphate isomerase, a subunit of the galactose-6phosphate isomerase and a regulatory protein belonging
to the DeoR family, were found in the sequence
downstream of the ery operon.
Analysis of the ery gene products by sequence
similarity
The relationship of the polypeptides deduced from the
nucleotide sequence of the ery region to known gene
products was determined using the program .
The results of these comparisons are summarized in
Table 3.
The  program identified homology between the
eryA gene product and several bacterial xylulose kinases.
The closest homologue to EryA was the product of the
E. coli gene xylB, encoding a xylulose kinase. Homology
of EryA with bacterial glycerol kinases was also
detected. A motif search in the EryA sequence against
PROSITE revealed the presence of signature 2 of the
FGGY family of carbohydrate kinases, PS00445 (Reizer
et al., 1991). This suggested that the eryA gene product
could be a kinase for erythritol.
EryB was found to be a homologue of aerobic glycerol3-phosphate dehydrogenases. The greatest identity
(50n4 %) was with the enzyme from E. coli. EryB had the
PROSITE signature of FAD-dependent glycerol-3-phosphate dehydrogenases, PS00977. These findings and the
structural similarity between glycerol phosphate and
erythritol phosphate suggested that EryB was the erythritol phosphate dehydrogenase needed for the catabolism of erythritol.
We failed to find any protein in the database with
extended similarity to the complete EryC. The most
significant match corresponded to a limited region
(57 aa) of the large subunit (HupL) of Alcaligenes
hydrogenophilus hydrogenase (accession number
S56898).
EryD showed similarity to several DNA-binding regulatory proteins. The closest homologue of EryD was
SmoC, the regulator of a Rhodobacter sphaeroides
operon involved in polyol transport and metabolism.
Residues 21–42 of EryD were identified with a 71 %
probability as a helix–turn–helix motif using the method
described by Dodd & Egan (1990). This finding
suggested that EryD was a DNA-binding protein with a
regulatory function.
Identification of the ery operon promoter
Functionality of the putative promoter identified by
sequence analysis was studied by primer extension.
491
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 14:53:47
F. J. S A N G A RI, J. A G U$ E R O a n d J. M. G A R C I! A - L O B O
1
2
3
4
5
A C G T
Synthetic oligonucleotide M2990 (Table 2), complementary to bases 314–335 of the ery sequence, was used to
identify the transcriptional start point. The results of the
primer extension study are shown in Fig. 3. A major
band was obtained whose size indicated that the putative
transcriptional start point was the residue G291, located
10 bp downstream of the promoter. This result confirmed that P1 was a functional promoter that directs the
transcription of the ery operon.
Regulation of the expression of the ery operon in
B. abortus
Primer extension of total RNA was used as a quantitative method to evaluate ery expression in B. abortus.
Addition of erythritol to the culture medium 30 min
before cell harvest resulted in a fourfold increase in ery
RNA expression in strain 2308. The amount of ery RNA
in B. abortus B19 was independent of the presence of
erythritol and was approximately the same as in B.
abortus 2308 growing in medium containing erythritol.
This effect could be due to the lack of EryD activity in
strain B19. To test this possibility we constructed an
eryD mutant from strain 2308 by gene replacement.
Strain FJS-2 is B. abortus 2308Nx with the chromosomal
eryD gene interrupted by a kanamycin-resistance gene
(Fig. 1). The level of transcription from the P1 promoter
was similar in the FJS-2 mutant and the B19 strain.
Transcription from P1 in FJS-2 was also unaffected by
the addition of erythritol to the culture medium. These
results confirmed the role of EryD in repression of
expression from P1 and suggested that erythritol acted
as an inducer, probably by binding to EryD.
T
A
C
A
A
T
A
G
A
G
G
T
T
G
T
A
C
G
C
G
G
5′ T
–10
+1
Fig. 3. Products obtained by reverse
transcriptase extension with the primer
M2990 (Table 2) on total mRNA obtained
from the following B. abortus strains and
culture conditions : 1, strain 2308 ; 2, strain
2308 grown in the presence of erythritol ; 3,
eryD mutant FJS-2 ; 4, FJS-2 grown in the
presence of erythritol ; 5, B19. A sequencing
ladder obtained with the same primer was
included as a size marker.
Expression of ery genes in E. coli
Since wild-type E. coli cells are unable to utilize
erythritol we introduced recombinant plasmids containing different genes from the ery operon into E. coli
HB101 and studied the ability of the recombinants to use
erythritol. Using this assay and the set of plasmids
described in Fig. 4, we determined that the minimal
region required for degradation of erythritol in E. coli
was eryABC (Fig. 4). In addition, plasmid pSU6004 was
introduced into several E. coli strains to investigate the
effects of various E. coli genes on ery expression. Strain
MC296, a himA hip (IHF−) double mutant did not
utilize erythritol upon introduction of pSU6004, whereas
the isogenic strain with the wild-type IHF alleles
(MC294) did. This finding correlated with the presence
of an IHF-binding site in the ery control region and
strongly suggested a role for IHF in the regulation of
expression of the ery operon. E. coli DH5α also produced white colonies on McConkey-erythritol plates
when transformed with the ery plasmids. The relA
mutation could be responsible for the poor expression of
the ery operon in this strain.
DISCUSSION
Erythritol metabolism in B. abortus has been studied
from a biochemical perspective. Degradation of the
polyalcohol starts with erythritol phosphorylation and
is followed by a dehydrogenation step that produces erythrulose 1-phosphate. Two more dehydrogenation
steps and a decarboxylation reaction are required
to finally produce dihydroxyacetone phosphate and
carbon dioxide (Sperry & Robertson, 1975a). We have
sequenced a genomic DNA fragment from B. abortus
containing four genes organized as an operon that we
492
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 14:53:47
Erythritol utilization operon of Brucella abortus
.................................................................................................................................................................................................................................................................................................................
Fig. 4. Diagram showing the plasmids constructed to determine the genes needed for erythritol utilization in E. coli. All
the constructs were based on the pBluescript (SK) vector. Plasmid pSU6104 contained eryA. It was constructed by cloning
an EcoRI fragment from the pSU6010 : : Tn1725 derivative pSU6054 (see Fig. 2). Plasmid pSU6099 contained eryAB. It was
constructed from the marked EcoRI–HincII fragment from pSU6004. Plasmid pSU6100 contained eryABC. The EcoRI–ClaI
fragment shown in the figure was used in its construction. Plasmid pSU6111 contained eryABCD and was constructed by
cloning the indicated EcoRI–NruI fragment. The genetic map of the ery operon is included below the pSU6004 map to
indicate the ery genes contained in each plasmid. The colour of the colonies formed by E. coli HB101 containing each
plasmid on MacConkey-erythritol plates (red or white) is indicated.
have called the ery operon. Introduction of different
combinations of these genes into E. coli indicated that
eryAB and C were sufficient for erythritol degradation.
The similarity of the EryA and EryB proteins to sugar
kinases and glycerol-3-phosphate dehydrogenase, respectively, fit the biochemical data. Accordingly we
assigned the function of erythritol kinase and erythritol
phosphate dehydrogenase to EryA and EryB, respectively. A possible discrepancy in this identification would
be that biochemical data have indicated the use of NAD
as the electron acceptor for the dehydrogenation step,
whereas the close identity of EryB with aerobic glycerol3-phosphate dehydrogenase, and its sequence signature,
suggested the use of FAD.
We have been unable to assign a function to EryC by
sequence comparison. A clue to EryC function was
obtained upon analysis of the ery operon in the vaccine
strain B. abortus B19, which is the only known B.
abortus isolate whose growth is inhibited by erythritol.
The metabolism of erythritol in B. abortus B19 was
studied biochemically by Sperry & Robertson (1975a),
who found that this strain lacks -erythrulose-1-phosphate dehydrogenase activity. We have previously described the presence of a deletion in the B19 chromosomal ery locus (Sangari et al., 1994). The results of the
present study show that the deletion in B19 is located
between positions 3975 and 4676. It affects two genes,
eryC and eryD. As result of the deletion, 137 aa at the
carboxyl end of EryC and 85 aa at the amino end of
EryD, including its putative DNA-binding domain, are
lost. A nonfunctional fused polypeptide containing the
amino half of EryC and the carboxyl end of EryD could
be produced in B19. Thus the B. abortus B19 strain is an
eryCD double mutant. The defect in B19 was complemented in trans by plasmids containing the complete ery
region and by plasmids with Tn1725 insertions in eryA,
eryB and eryD. Plasmids with Tn1725 insertions in eryC
were the only ones that failed to complement the Ery−
phenotype of B19 (Fig. 2). This indicated that the
product of the eryC gene could be the enzyme erythrulose-1-phosphate dehydrogenase.
Sequence analysis and comparison showed that EryD
was a protein with a helix–turn–helix motif and
probably possessed some regulatory function. To obtain
further evidence for this, we constructed an eryD mutant
(FJS-2) and compared the levels of ery transcription in
this mutant and a wild-type strain. Transcription from
the ery promoter (P1) was more active in both the
mutant FJS-2 and strain B19 than in strain 2308. This
result strongly suggested a repressor function for EryD.
The increase in ery mRNA level induced by erythritol in
strain 2308 indicated an inducing function for erythritol,
probably through binding to EryD. Many transcription
repressors with a helix–turn–helix motif bind sugars
that function as inducers. In addition to control by
EryD, we have found an IHF-binding site upstream of
the P1 promoter and results suggested that IHF was
required for ery expression in E. coli. Although IHF has
not yet been described in Brucella, it is generally accepted
that all bacteria must contain a protein of this type. The
role of IHF as a transcriptional regulator in other species
has been well documented (Perez-Martin et al., 1994).
493
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 14:53:47
F. J. S A N G A RI, J. A G U$ E R O a n d J. M. G A R C I! A - L O B O
The genes in the ery operon described here encode three
enzymes that are able to transform erythritol into 3keto--erythrose 4-phosphate according to the pathway
described by Sperry & Robertson (1975a). E. coli cells
provided with these three enzymes were capable of
erythritol utilization. This observation suggests that the
enzymes for degradation steps beyond 3-keto--erythrose 4-phosphate may be organized in other pathways,
and their genes found at other chromosomal locations.
Since this is the first description of an operon for
erythritol utilization, comparisons are not possible.
However, the analysis of operons for polyol utilization
(Heuel et al., 1998) highlights the lack of a gene for
erythritol transport. Erythritol was shown to be a
substrate for the E. coli glycerol facilitator GlpF (Heller
et al., 1980). Furthermore it is known that B. abortus
B19 mutants tolerant to erythritol arise frequently, and
that those mutants show impaired ability to grow on
glycerol culture media (Sangari et al., 1996). Thus it is
possible that erythritol is transported into Brucella
through a GlpF analogue, and later phosphorylated by
erythritol kinase, trapping the polyol in the bacterial
cytoplasm. The results reported here explain the
utilization of erythritol by B. abortus, as well as the
inhibition of growth in strain B19 by overproduction of
the kinase EryA and the dehydrogenase EryB. The
growth-promoting effect of erythritol has not been
addressed in this work but there could be a link between
erythritol and the aromatic pathway. Erythrose 4phosphate, the precursor for the biosynthetic pathway
for aromatic compounds in bacteria, may be easily
obtained from the intermediates of erythritol catabolism. Thus erythritol may be crucial to obtain
molecules such as aromatic amino acids and catechols,
which play important physiological roles in Brucella,
providing a connection between erythritol metabolism
and virulence in B. abortus.
ACKNOWLEDGEMENTS
This work was supported by grants BIO96-1398-C02-02 and
BIO98-0674 from the Spanish ‘ Plan Nacional de IjD ’. We
thank Jose! Pe! rez-Martı! n for the computer analysis of DNA
bending.
REFERENCES
Altschul, S. F., Madden, T. L., Schaffer, A. A., Zhang, J., Zhang, Z.,
Miller, W. & Lipman, D. J. (1997). Gapped  and - : a
new generation of protein database search programs. Nucleic
Acids Res 25, 3389–3402.
Anderson, J. D. & Smith, H. (1965). The metabolism of erythritol
in Brucella abortus. J Gen Microbiol 38, 109–124.
Bagdasarian, M., Lurz, R., Ruckert, B., Franklin, F. C., Bagdasarian,
M. M., Frey, J. & Timmis, K. N. (1981). Specific-purpose plasmid
cloning vectors. II. Broad host range, high copy number, RSF
1010-derived vectors, and a host–vector system for gene cloning
in Pseudomonas. Gene 16, 237–247.
Craig, N. L. & Nash, H. A. (1984). E. coli integration host factor
binds to specific sites in DNA. Cell 39, 707–716.
Devereux, J., Haeberli, P. & Smithies, O. (1984). A comprehensive
set of sequence analysis programs for the VAX. Nucleic Acids Res
12, 387–395.
Dodd, I. B. & Egan, J. B. (1990). Improved detection of helix–
turn–helix DNA-binding motifs in protein sequences. Nucleic
Acids Res 18, 5019–5026.
Dower, W. J., Miller, J. F. & Ragsdale, C. W. (1988). High efficiency
transformation of E. coli by high voltage electroporation. Nucleic
Acids Res 16, 6127–6145.
Gamas, P., Burger, A. C., Churchward, G., Caro, L., Galas, D. &
Chandler, M. (1986). Replication of pSC101 : effects of mutations
in the E. coli DNA binding protein IHF. Mol Gen Genet 204,
85–89.
Garrido, T., Sanchez, M., Palacios, P., Aldea, M. & Vicente, M.
(1993). Transcription of ftsZ oscillates during the cell cycle of
Escherichia coli. EMBO J 12, 3957–3965.
Heller, K. B., Lin, E. C. & Wilson, T. H. (1980). Substrate specificity
and transport properties of the glycerol facilitator of Escherichia
coli. J Bacteriol 144, 274–278.
Heuel, H., Shakeri-Garakani, A., Turgut, S. & Lengeler, J. W.
(1998). Genes for -arabinitol and ribitol catabolism from
Klebsiella pneumoniae. Microbiology 144, 1631–1639.
Kokotek, W. & Lotz, W. (1991). Construction of a mobilizable
cloning vector for site-directed mutagenesis of gram-negative
bacteria : application to Rhizobium leguminosarum. Gene 98,
7–13.
Meyer, M. E. (1966). Metabolic characterisation of the genus
Brucella. V. Relationship of strain oxidation rate of i-erythritol to
strain virulence for guinea pigs. J Bacteriol 92, 584–588.
Meyer, M. E. (1967). Metabolic characterisation of the genus
Brucella. VI. Growth stimulation by i-erythritol compared with
strain virulence for guinea pigs. J Bacteriol 93, 996–1000.
Perez-Martin, J., Rojo, F. & de Lorenzo, V. (1994). Promoters
responsive to DNA bending : a common theme in prokaryotic
gene expression. Microbiol Rev 58, 268–290.
Reizer, A., Deutscher, J., Saier, M. H., Jr & Reizer, J. (1991).
Analysis of the gluconate (gnt) operon of Bacillus subtilis. Mol
Microbiol 5, 1081–1089.
Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular
Cloning : a Laboratory Manual, 2nd edn. Cold Spring Harbor,
NY : Cold Spring Harbor Laboratory.
Sangari, F. & Agu$ ero, J. (1991). Mutagenesis of Brucella abortus :
comparative efficiency of three transposon delivery systems.
Microb Pathog 11, 443–446.
Sangari, F. J. & Agu$ ero, J. (1994). Identification of Brucella
abortus B19 vaccine strain by the detection of DNA polymorphism at the ery locus. Vaccine 12, 435–438.
Sangari, F. J., Garcı! a-Lobo, J. M. & Agu$ ero, J. (1994). The Brucella
abortus vaccine strain B19 carries a deletion in the erythritol
catabolic genes. FEMS Microbiol Lett 121, 337–342.
Sangari, F. J., Agu$ ero, J. & Garcı! a-Lobo, J. M. (1996). Improvement
of the Brucella abortus B19 vaccine by its preparation in a glycerol
based medium. Vaccine 14, 274–276.
Sangari, F. J., Grillo, M. J., Jime! nez De Bagu$ e! s, M. P., Gonza! lezCarrero! , M. I., Garcı! a-Lobo, J. M., Blasco, J. M. & Agu$ ero, J.
(1998). The defect in the metabolism of erythritol of the Brucella
abortus B 19 vaccine strain is unrelated with its attenuated
virulence in mice. Vaccine 16, 1640–1645.
Sanger, F., Nicklen, S. & Coulson, A. R. (1977). DNA sequencing
with chain-terminating inhibitors. Proc Natl Acad Sci USA 74,
5463–5467.
494
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 14:53:47
Erythritol utilization operon of Brucella abortus
Simon, R., Priefer, U. & Pu$ hler, A. (1983). A broad host range
mobilization system for in vivo genetic engineering, transposon
mutagenesis in gram negative bacteria. Bio\Technology 1,
784–791.
Smith, H., Williams, A. E., Pearce, J. H., Keppie, J., Harris-Smith,
P. W., Fitzgeorge, R. B. & Witt, K. (1962). Foetal erythritol : a
cause of the localisation of Brucella abortus in bovine contagious
abortion. Nature 193, 4749.
Sperry, J. F. & Robertson, D. C. (1975a). Erythritol catabolism by
Brucella abortus. J Bacteriol 121, 619–630.
Sperry, J. F. & Robertson, D. C. (1975b). Inhibition of growth by
erythritol catabolism in Brucella abortus. J Bacteriol 124,
391–397.
Ubben, D. & Schmitt, R. (1986). Tn1721 derivatives for transposon
mutagenesis, restriction mapping and nucleotide sequence analysis. Gene 41, 145–152.
.................................................................................................................................................
Received 21 June 1999 ; revised 10 October 1999 ; accepted 1 November
1999.
495
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 14:53:47